text
stringlengths 87
880k
| pmid
stringlengths 1
8
| accession_id
stringlengths 9
10
| license
stringclasses 2
values | last_updated
stringlengths 19
19
| retracted
stringclasses 2
values | citation
stringlengths 22
94
| decoded_as
stringclasses 2
values | journal
stringlengths 3
48
| year
int32 1.95k
2.02k
| doi
stringlengths 3
61
| oa_subset
stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|
==== Front
BMC ImmunolBMC Immunology1471-2172BioMed Central London 1471-2172-6-111596704410.1186/1471-2172-6-11Research ArticleThe allergy adjuvant effect of particles – genetic factors influence antibody and cytokine responses Nygaard Unni Cecilie [email protected] Audun [email protected]øvik Martinus [email protected] Division of Environmental Medicine, Norwegian Institute of Public Health, P.O.Box 4404 Nydalen, NO-0403 Oslo, Norway2 Division of Infectious Disease Control, Norwegian Institute of Public Health, P.O.Box 4404 Nydalen, NO-0403 Oslo, Norway3 Institute of Cancer Research and Molecular Medicine, Faculty of Medicine, Norwegian University of Science and Technology, NO-7489 Trondheim, Norway2005 21 6 2005 6 11 11 12 11 2004 21 6 2005 Copyright © 2005 Nygaard 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
There is increasing epidemiological and experimental evidence for an aggravating effect of particulate air pollution on asthma and allergic symptoms and, to a lesser extent, on allergic sensitization. Genetic factors appear to influence not only the magnitude, but also the quality of the adjuvant effect of particles with respect to allergen-specific IgE (Th2-associated) and IgG2a (Th1-associated) responses. In the present study, we aimed to investigate how the genetic background influences the responses to the allergen and particles alone and in combination. We examined how polystyrene particles (PSP) affected the IgE and IgG2a responses against the model allergen ovalbumin (OVA), after subcutaneous injection into the footpad of BALB/cA, BALB/cJ, NIH and C3H/HeN mice, Further, ex vivo IL-4, IFN-γ and IL-10 cytokine secretion by Con A-stimulated cells from the draining popliteal lymph node (PLN) five days after injection of OVA and PSP separately or in combination was determined.
Results
PSP injected with OVA increased the levels of OVA-specific IgE antibodies in all strains examined. In contrast, the IgG2a levels were significantly increased only in NIH and C3H/HeN mice. PSP in the presence of OVA increased cell numbers and IL-4, IL-10 and IFN-γ levels in BALB/cA, NIH and C3H/HeN mice, with the exception of IFN-γ in NIH mice. However, each mouse strain had their unique pattern of response to OVA+PSP, OVA and PSP, and also their unique background cytokine response (i.e. the cytokine response in cells from mice injected with buffer only).
Conclusion
Genetic factors (i.e. the strain of mice) influenced the susceptibility to the adjuvant effect of PSP on both secondary antibody responses and primary cellular responses in the lymph node, as well as the cellular responses to both OVA and PSP given separately. Interestingly, PSP alone induced cytokine responses in the lymph node in some of the mouse strains. Furthermore, we found that the ex vivo cytokine patterns did not predict the in vivo Th2- and Th1-associated antibody response patterns in the different mouse strains. The results indicate that insoluble particles act by increasing the inherent response to the allergen, and that the genetic background may determine whether an additional Th1-associated component is added to the response.
==== Body
Background
Numerous epidemiological studies conducted in different parts of the world have demonstrated a consistent association between levels of ambient air particles and various health outcomes, including mortality, cardiopulmonary disease, reduced lung function, and exacerbation of asthma and allergy-related symptoms (reviewed in [1,2]). Some populations seem to be at particular risk, like the elderly, children or patients with diabetes, chronic obstructive pulmonary disease and asthma [1,3]. Furthermore, the genetic background has been associated with differences in human susceptibility to environmental agents including pesticides and infectious agents [4], and may be of importance also with regard to biological effects of particles.
Supporting a causal relationship between particles and increased allergic symptoms, particles have been shown to increase allergic responses both in humans and in animal models (reviewed in [5,6]). In rodent models, widely different particles have been shown to increase allergen-specific IgE levels both after inhalation, intratracheal, intranasal, intraperitoneal and subcutaneous exposure [7-12]. Particle characteristics like particle size, particle composition and chemicals adsorbed to the particle surface have been shown to influence the adjuvant effect on allergic sensitization and airway inflammation and hyperreactivity in mice [11-18]. For combustion particles, the insoluble particle core appears to play an important role in the adjuvant effect of particles in several mouse models [18-23]. Mouse strain differences in allergic airway inflammation and allergen-specific antibody levels have been reported after intratracheal instillation of diesel exhaust particles (DEP) and allergen [24,25]. Also with "clean" polystyrene particles (PSP, 0.1 μm diameter) as a surrogate of the insoluble combustion particle core, Granum et al. [26] concluded that the particle adjuvant effect differed qualitatively with regard to Th2- and Th1-associated antibody responses, depending on the genetic background or induced immune status of the mice.
The overall aim of the present study was to investigate how the genetic background influences the responses to the allergen and particles alone and in combination. We therefore examined the antibody and lymph node responses in various inbred strains of mice expected to display different responses to OVA and particles. BALB/cA mice are regarded as a high IgE responder strain to OVA [27,28]. NIH mice were included because they in an intraperitoneal model responded differently from BALB/cA mice to PSP and OVA [26]. Furthermore, we included C3H/HeN mice, which have been shown to be IgE low-responders to OVA [28]. BALB/cJ mice were included in the antibody experiments because this sub-strain has been reported to respond more strongly to OVA than BALB/cA mice in a subcutaneous model [29].
The lymph node is a key site in the priming of T lymphocytes and IgE production by B lymphocytes [11,30,31]. Therefore, in addition to serum antibody levels, we examined the primary cellular response in the draining popliteal lymph node (PLN) after footpad injection of polystyrene particles (PSP) and the allergen ovalbumin (OVA), alone and in combination. The PLN is the only lymph node draining the footpad [32], and is therefore well suited for studies of responses in the lymph node cells. We investigated how the selected inbred strains of mice differed qualitatively and quantitatively in their Th2- and Th1-associated serum antibody and PLN cellular response patterns to OVA+PSP. Furthermore, we investigated if the background (i.e. the ex vivo cytokine response in cells from mice injected with buffer only), OVA-, PSP- and OVA+PSP-induced cytokine patterns differed in the various mouse strains, and whether the cytokine patterns could explain the differences in the antibody responses to OVA+PSP. We found that the susceptibility to the adjuvant effect of particles on both secondary antibody responses and primary cellular responses was influenced by genetic factors (i.e. mouse strain). The ex vivo cytokine response patterns did not predict the in vivo antibody responses.
Results
The antibody adjuvant activity of PSP in different strains of mice
To examine the dose-response relationship of OVA and PSP with regard to antibody production, BALB/cA and NIH mice were immunized with various dose combinations of OVA (100, 50 and 10 μg) and PSP (100, 40, 10 and 0 μg). After a booster dose of OVA on day 21, the OVA-specific IgE, IgG1 and IgG2a antibody levels in serum were determined on day 26. In BALB/cA mice, PSP given with OVA strongly increased the levels of the Th2-associated IgE and IgG1 antibodies compared to OVA alone (Figure 1A, B). In contrast, the Th1-associated IgG2a levels did not differ between the groups given OVA in the presence or absence of PSP (Figure 1C). Overall, there was an inverse relationship between the OVA dose and both the IgE and IgG1 levels, with 10 μg of OVA giving the highest antibody levels for all particle doses. Analyses of the factorial design showed that the dose of 10 μg OVA, regardless of PSP dose, induced significantly higher IgE levels than the other OVA doses, and higher IgG1 levels than 100 μg OVA (p < 0.001). Irrespective of OVA dose, the IgG1 response increased significantly with each increase in PSP dose (p < 0.001). A dose-response trend for particles was less clear for IgE due to an absence of a difference between 40 and 100 μg PSP.
In NIH mice, PSP also exerted an adjuvant effect on the antibody responses to OVA, however NIH mice responded differently from BALB/cA mice in two ways. First, the NIH mice receiving the highest dose combinations of OVA and PSP experienced anaphylactic reactions and died within minutes after the OVA booster injection (indicated with crosses in Figure 1D, E, F). Secondly, also the OVA-specific IgG2a levels were increased by PSP in NIH mice, in contrast to BALB/cA mice. However, the responses in the surviving mice (i.e. 10 μg OVA) suggest that IgE and IgG1 responses were similar in NIH and BALB/cA mice with regard to the dose-response relationship for particles. Also the IgG2a levels were significantly higher for OVA with all particle doses than for OVA alone, and significantly higher antibody levels were observed with 40 μg PSP than with 10 μg PSP.
As 10 μg OVA and 40 μg PSP gave a medium but still pronounced response, these doses were used in further experiments. Next, we compared the allergen-specific antibody responses after immunization with OVA+PSP, OVA and PSP in BALB/cA, BALB/cJ, NIH and C3H/HeN mice. As shown in Figure 2, the response patterns for BALB/cA and NIH mice were similar to those shown in Figure 1, with increased IgE production in both strains, whereas IgG2a production was increased only in NIH mice. Furthermore, in our system, BALB/cJ mice had specific IgE and IgG2a responses similar to BALB/cA mice, however the IgG2a levels tended to be slightly increased by OVA+PSP in BALB/cJ mice (statistically significant in one of two experiments). Similar to the responses in NIH mice, PSP had an adjuvant effect on both IgE and IgG2a responses in C3H/HeN mice, although the responses appeared weaker in C3H/HeN than in NIH and BALB/c mice.
Immunization with OVA alone gave IgE responses in some individual BALB/cA, BALB/cJ and C3H/HeN mice, but never in NIH mice (Figure 1 and 2). As expected, HBSS and PSP without allergen did not affect the specific antibody levels in any of the strains.
Cell numbers and ex vivo cytokine responses in the draining lymph node during the primary response
To examine how the primary cellular response to OVA and PSP, alone or in combination, varied in BALB/cA, NIH and C3H/HeN mice, we studied the cell numbers and the ex vivo cytokine secretion by the popliteal lymph node (PLN) cells. Five days after injection of OVA+PSP, OVA, PSP and buffer (HBSS) into the footpad, the PLN cell numbers were determined, and the amount of IL-4, IFN-γ and IL-10 released from PLN cells after Con A stimulation ex vivo were determined.
For all three strains, the PLN cell numbers were significantly increased by OVA+PSP compared to all three control groups (Figure 3A). Interestingly, the cell number response to OVA alone and PSP alone (compared to HBSS) varied between the three strains. In BALB/cA mice, OVA significantly increased the cell numbers, whereas PSP significantly increased the cell numbers in NIH mice. C3H/HeN mice took an intermediate position, as both OVA and PSP weakly but significantly increased the cell numbers compared to HBSS.
With regard to the cytokines (Figure 3B, C and 3D), OVA+PSP strongly increased the levels of IL-4 and IL-10 compared to the three control groups in BALB/cA mice (p < 0.006), whereas the IFN-γ levels were only marginally affected (p < 0.035). Furthermore, OVA tended to weakly (n.s.) increase the IL-4 levels in BALB/cA mice. The IL-4 levels for the BALB/cA control groups were relatively low compared to levels in NIH mice. In NIH mice, OVA induced no changes in cytokines or cell numbers, whereas PSP significantly increased the IL-4 and IL-10 levels (p < 0.007), and OVA+PSP also increased these "responsive" cytokines (p < 0.005, compared to OVA). The IFN-γ levels in NIH mice were not altered by PSP. Again, the C3H/HeN mice responded to both OVA and PSP. All cytokines were significantly increased by OVA and further increased by OVA+PSP in C3H/HeN mice, although more pronounced for IFN-γ than for IL-4 and IL-10. PSP alone also significantly increased the levels of IFN-γ and IL-10 in C3H/HeN mice. The levels of IL-4 and IL-10 were low and the levels of IFN-γ were high in C3H/HeN mice compared to the levels in the other strains. The background levels (buffer treated animals) of IL-4 and IFN-γ were relatively low in BALB/cA and relatively high in NIH mice, whereas the levels in C3H/HeN mice appeared low for IL-4 and high for IFN-γ, respectively.
Discussion
The genetic background appears to strongly influence whether an individual develops allergic immune responses to an allergen [33]. In mice, this is illustrated by reports of strain differences in allergic responses to allergens [28,34-36]. In agreement with Granum et al. [26], our data indicate that also the antibody-enhancing capacity of the "clean" insoluble particle is influenced by the genetic background. This is consistent with reports of mouse strain differences in allergic airway inflammation and specific antibody levels after intratracheal instillation of DEP and allergen [24,25]. Also the effects of particles on airway inflammation, epithelial permeability and alveolar macrophage phagocytic activity have been shown to vary between mouse strains [37].
The dose-response relationship for particles appeared similar for BALB/cA and NIH mice, as the low OVA dose (10 μg) and the high particle doses (40 and 100 μg PSP) induced the strongest IgE and IgG1 antibody responses. However, the NIH mice immunized with a high concentration of both OVA and PSP were lost due to anaphylactic shock after the booster injection. High booster antigen concentrations have earlier been shown to induce anaphylactic shock in mice after subcutaneous priming with horse gamma globulin covalently bound to cellulose particles [38]. In mice, IgG(1) has been suggested to be as effective as IgE in triggering anaphylactic reactions [39,40]. The induced levels of OVA-specific IgG1 appeared to be about 10 times higher in the surviving NIH mice than in BALB/cA mice (Figure 1B and 1E). Thus, NIH mice may be more susceptible than BALB/cA mice to experience anaphylactic shock either because of a stronger IgG1 response, or because of an otherwise greater susceptibility to anaphylactic shock.
In the present study, the antibody adjuvant effect of PSP differed between the different strains of mice, both with regard to the magnitude and whether a Th2 or a mixed Th1- and Th2-associated antibody response was increased. Whereas the Th1-associated IgG2a response was markedly increased by PSP only in NIH and C3H/HeN mice, PSP enhanced the Th2-associated IgE response in all mouse strains examined. PSP seemed also to increase IgE responses in pilot studies with C57BL/6J and C3H/HeJ mice (data not shown). Also with regard to the adjuvant effect of PSP, the BALB/cA mice appeared to be a high IgE-responder, whereas the C3H/HeN mice seemed to be a relatively low responder for both OVA-specific IgE and IgG2a antibody levels. This responder status is in accordance with other studies of the IgE responsiveness to OVA in these strains of mice [27,28].
Also with regard to the primary cellular responses in the lymph node, PSP increased the response to OVA in both BALB/cA, NIH and C3H/HeN mice (Figure 3). However, while IL-4 and IL-10 cytokine responses were significantly increased by OVA+PSP in all three strains, in C3H/HeN mice the levels were low and the increases may not be biologically significant. On the other hand, OVA+PSP markedly increased the IFN-γ levels in C3H/HeN mice, and marginally in BALB/cA mice. IL-4 and IFN-γ cytokines are often used as indicators for a Th2 and Thl response, respectively [41]. In this regard, the cytokine responses to OVA+PSP could be characterized as Th2-, Th2- and Th1-biased for BALB/cA, NIH and C3H/HeN mice, respectively, which would not, however, predict the balance between Th2- and Th1-associated antibody responses (Figure 1 and 2, see below). Neither could the background (buffer control) responses for IL-4 and IFN-γ in the different strains explain the balance between Th2- and Th1- associated antibody responses to OVA+PSP.
The complexity in gene-environment interactions were illustrated by the highly varying responsiveness to OVA and PSP given separately in the three strains examined. BALB/cA mice responded with a non-significant but consistent increase in IL-4 cytokine release in OVA-treated mice, whereas NIH mice responded with a significantly increased Th2-biased (IL-4 and IL-10) cytokine release in PSP-treated mice. C3H/HeN mice responded with both Th1- and Th2-type cytokine secretion, although apparently Th1-skewed due to low IL-4 and IL-10 levels, after both OVA or PSP treatments. This suggests that the genetic background strongly influences the susceptibility to immunogens and irritative stimuli such as particles, also with regard to the Th1 and Th2 cytokine balance of the response. Interestingly, PSP alone increased the Th2-associated, but not the Th1-associated, cytokine responses in NIH mice. If particles by themselves are able to induce IL-4 production in the lymph nodes also in vivo, this might promote an allergic response to the allergen present.
Taken together, PSP exerted an adjuvant effect on OVA-induced IgE and IL-4 levels in all three strains, while there was no association between the Th1-associated parameters IgG2a and IFN-γ. Thus, for the strains of mice used, the primary ex vivo cytokine pattern could not predict the Th2- and Th1-associated antibody pattern. One reason could be that the Con A-stimulation in vitro was not an optimal way to describe the actual in vivo cytokine secretion by the PLN cells. However, in vitro OVA-stimulation of PLN cells from OVA+PSP treated BALB/cA mice gave similar IL-4 and IFN-γ response pattern as Con A-stimulation (data not shown). Our findings are consistent with previous studies showing that cytokine expression was only in part correlated with airway disease and IgE responses [35,42], and our data illustrate that interpretation and extrapolation of in vitro data should be done with caution. The description of the Th2/Th1 balance may explain parts of an immune response, however the immune regulation is more complex, involving other cytokines, cell surface molecules and regulatory T cells. IL-10 seems to play an important role in the development and function of certain regulatory T cells (reviewed in [43,44]). However, in the present study, IL-10 secretion by the pool of PLN cells was highly correlated with the secretion of IL-4 (rs = 0.86). Although contribution from regulatory T cells could not be excluded, this indicates that IL-10, which are secreted to a higher degree by Th2 than Th1 cells [45], in the PLN at this stage was produced mainly by Th2 cells.
In agreement with Randolph et al. [46], we have previously shown that both PSP, DEP and ambient air particles in the draining lymph node were located within acid phosphatase-, MHC class II- and CD86-positive cells, most probably antigen-presenting cells (APC) [23,47]. It is remarkable that PSP, in the absence of allergen, affected the lymph node cells in both NIH and C3H/HeN mice. In contrast, our present and previous data from BALB/cA mice indicate that although the particles synergistically enhanced the cellular responses to OVA, model particles (PSP, DEP) had no detectable effects neither on PLN cell numbers, expression of lymphocyte surface molecules nor on ex vivo cytokine secretion by the PLN cells [23,47]. Taken together, our data from BALB/cA mice therefore suggest a role for solid particles as allergen carriers [48] in the adjuvant effect on allergic sensitization. Our present data suggest that particles, in susceptible strains, additionally influence the immune response in the lymph node. In our model it appears that the particles which induced cellular responses in the absence of allergen (i.e. ambient air particles in BALB/cA mice [47] as well as PSP in NIH and C3H/HeN mice (present study)) enhanced both Th2- and Th1-associated antibody responses, whereas the particles enhancing a Th2-dominated antibody response (i.e. PSP and DEP in BALB/cA mice) did not induce detectable cellular changes themselves. This suggests that the insoluble particle core act by enhancing the inherent response to the allergen (acts like a "general motor" [6]), and that genetic background and particle properties determine whether an additional Th1-associated response is elicited. The presence of a Th1 response has been suggested to amplify Th2-driven inflammation [49-52]. Our results therefore suggest that particles may both enhance the IgE sensitization and, depending on genetically determined susceptibility, further aggravate allergic inflammation and thereby increase the allergic symptoms in some individuals.
Conclusion
PSP enhanced the OVA-specific IgE antibody responses in all strains of mice examined, whereas the IgG2a response was increased to varying degrees and only in some strains. Genetic factors (i.e. the strain of mice) influenced the responsiveness to the adjuvant effect of PSP on secondary antibody responses and the primary cellular responses to OVA. Furthermore, genetic factors also influenced the lymph node cell responsiveness to both OVA and PSP given separately. Interestingly, PSP alone induced different cytokine responses in the lymph node in some of the mouse strains. Furthermore, we found that the ex vivo cytokine patterns could not predict the in vivo Th2- and Th1-associated antibody response patterns in the different mouse strains. The results indicate that the insoluble particle core acts by increasing the inherent response to the allergen, and that the genetic background may determine whether an additional Th1-associated component is added.
Methods
Animals
Female inbred BALB/cABomTac and BALB/cJBomTac mice (Taconic M&B A/S, Ry, Denmark) and NIH/OlaHsd and C3H/HeNHsd mice (Harlan UK Ltd., Oxon, UK) were used. The mice were housed, 5–8 animals per cage, on BeeKay bedding (B&K Universal AS, Nittedal, Norway) in type III macrolon cages in filter cabinets (Scantainers). The animals were allowed to rest for at least one week before entering the experiments 6–7 weeks old. The mice were exposed to a 12-hr/12-hr light/dark cycle (30–60 lux in cages), room temperature (20 ± 2°C) and 40–60% relative humidity. The animals were given pelleted food (RM1, SDS, Essex, UK) and tap water ad libitum. The experiments were performed in conformity with the laws and regulations for experiments with live animals in Norway, and they were approved by the Experimental Animal Board under the Ministry of Agriculture in Norway.
Particles and allergen
As a surrogate for the insoluble particle core of combustion particles, polystyrene particles (PSP) with a diameter of 0.1 μm were used (Polybead Polystyrene Microspheres; Polysciences Europe GmbH, Eppelheim, Germany). As allergen, ovalbumin (OVA, Gal d 2; chicken egg albumin, grade VII, Sigma, St. Louis, MO, USA) was used. In the dose-response experiments, we used particle doses of 100, 40 and 10 μg per injection (14.9, 5.94 and 1.49 × 1010 particles, respectively), as well as OVA doses of 100, 50 and 10 μg per injection. For all further studies, the dose of 40 μg PSP (5.94 × 1010 particles) and 10 μg OVA per injection was used. All preparations were made in Hank's balanced salt solution (HBSS; PAA Laboratories GmbH, Linz, Austria). 20 μl of the different preparations were injected subcutaneously into the hind footpad (heal-toe direction) using a 100 μl Hamilton syringe (Hamilton Bonaduz AG, Switzerland) with a 30 G sterile needle (BD Medical Systems, Ireland).
Experimental design
Dose response experiments were performed in groups of eight BALB/cA or NIH mice, with all combinations of 100, 40, 10 and 0 μg PSP and 100, 50 and 10 μg OVA. The suspensions were injected into the right hind footpad on day 0, and a booster with the same OVA dose was given on day 21. On day 26, the mice were exsanguinated by heart puncture under CO2 anesthesia. Serum was stored at -20°C until analyzed. Serum levels of OVA-specific IgE, IgG1 and IgG2a were measured by enzyme-linked immunosorbent assay (ELISA). Next, to further examine the qualitative differences in the adjuvant effect of particles, antibody responses after injection of OVA+PSP, OVA, PSP and buffer (HBSS) were determined in the mouse strains NIH, BALB/cA and BALB/cJ, and in a separate experiment in C3H/HeN. These experiments were all performed twice, with eight mice per group.
In a model previously described [23], we studied the primary cellular response in the draining lymph node in BALB/cA, NIH and C3H/HeN mice (fifteen mice per group). The popliteal lymph nodes (PLN) were excised five days after injection into both hind footpads of OVA and PSP, separately or in combination. After pooling of the PLN from the same animal, the total cell numbers per PLN were determined (Coulter Counter). Thereafter, PLN cells from five animals were pooled to achieve sufficient cell numbers per well for ex vivo culture with Con A, and the cytokine concentrations in cell culture supernatants were determined by sandwich ELISAs. In one set of experiments, the responses of cells from OVA and OVA+PSP treated animals from all strains were determined, and untreated mice were studied in a separate experiment. A confirming set of experiments was then conducted, this time performed in separate experiments for each strain of mice (the results shown in Figure 3).
ELISA for detection of IgE, IgG1 and IgG2a anti-OVA antibodies
IgE, IgG1 and IgG2a anti-OVA antibodies were detected in a capture ELISAs as previously described [19,53], as modified in [54]. In short, microtiter plates coated with a monoclonal rat anti-mouse IgE, rat anti-mouse IgG1 or rat anti-mouse IgG2a antibody were incubated with the mouse sera in triplicates, diluted 1:10, 1:20 and 1:20, respectively. After incubation with biotinylated OVA, followed by incubation with pre-formed complexes of streptavidin and biotinylated alkaline phosphatase, p-nitrophenyl phosphate disodiumhexahydrat diluted in a diethanolamine buffer was added as substrate. Dilutions of a standard mouse OVA-immune serum pool were included on each plate for standard curve production. OD was measured at 405 nm on a MRX Microplate Reader (Dynatech Laboratories, Chantilly, VA, USA) connected to a PC using BioLinx software (Dynatech Laboratories), and the antibody concentrations are given in arbitrary units (AU) per ml.
PLN single cell suspensions and ex vivo cytokine secretion after Con A stimulation
Five days after injection, the popliteal lymph nodes were excised and single cell suspensions prepared as previously described [23]. After pooling of the two PLN cell suspensions from the same animal, the cell concentration was measured with a Coulter Counter Z1 (Beckman Coulter Incorporated, FL, USA), and presented as the total cell number (106) per PLN. Thereafter, as described elsewhere [23], the PLN cells were cultured with 5 μg/ml Concanavalin A (Con A, Sigma, MO, USA), and ex vivo IL-4, IFN-γ and IL-10 cytokine secretion into the culture supernatants was determined by Mouse DuoSets ELISA (R&D Systems Inc., MN, USA). The optimal incubation times for the measurement of IL-4, IFN-γ and IL-10 with BALB/cA mice were previously found to be 24, 48 and 24 hours, respectively [23], and data for these time points are presented.
Statistical analysis
Statistical analysis was performed with SigmaStat ® Statistical Analysis System for Windows Version 2.03 (Jandel Scientific, Erkrath, Germany). Antibody levels, cell numbers and cytokine levels were log10-transformed to achieve normally distributed data with equal variances. In the dose-response experiments in BALB/cA mice, the data were analyzed in a two way ANOVA, using "OVA dose" as factor 1 and "PSP dose" as factor 2. Due to lost animals in the experiment with NIH mice, similar analyses could not be performed, and a one way ANOVA was run only on the data with 10 μg OVA. In both strains, the pair-wise comparisons were performed using Tukey's post hoc test. In all other experiments, to assure equal criteria for all strains independent of experimental size/setup, the data from each strain of mice were analyzed separately, using a parametric one way ANOVA. If the ANOVAs showed statistical differences with a significance of p ≤ 0.05, pair-wise comparisons of all groups within each mouse strain were performed by Tukey's post hoc test. Experimental groups were considered statistically different if p ≤ 0.05, and differences found statistically significant in two experiments are indicated in the figures.
Authors' contributions
UCN carried out the experimental work, statistical analyses and wrote the manuscript. ML initiated the project and obtained funding. UCN, ML and AA participated in designing the study and the interpretation of the data, and edited the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Else-Carin Groeng, Rita Bente Leikvold, Åse Eikeset, Bodil Hasseltvedt and Berit A. Stensby for excellent technical assistance. This work was funded by The Norwegian Research Council, grant no. 129593/310, and received further financial support from The Norwegian Institute of Public Health.
Figures and Tables
Figure 1 Allergen and particle dose-response experiments in BALB/cA and NIH mice Levels of serum OVA-specific IgE (A, D), IgG1 (B, E) and IgG2a (C, F) on day 26 after injection into one hind footpad of 100 (open columns), 50 (gray) or 10 (black) μg OVA combined with 100, 40, 10 or 0 μg PSP, or buffer (HBSS (hatched)), in BALB/cA (A, B, C) and NIH (D, E, F) mice. On day 21 the mice were boostered with the same OVA dose in the footpad. The PSP-control ("PSP ctr") group was given 100 μg PSP on day 0, followed by HBSS injection on day 21 (diamonds). Values (arbitrary units, AU) for individual mice (circles) and median values (columns) for groups of eight mice are shown. Dotted lines indicate the lower or higher detection limits for the ELISA assays. Note different scales for the two strains of mice.
Figure 2 The adjuvant effect of PSP on OVA-specific IgE and IgG2a responses in different strains of mice Levels of serum OVA-specific IgE (A) and IgG2a (B) on day 26 after injection into one hind footpad of HBSS (open columns), 10 μg OVA (gray), 40 μg PSP (hatched) or 10 μg OVA + 40 μg PSP (black) in different mouse strains. On day 21 all mice were boostered with 10 μg OVA in the footpad. Vertical straight lines indicate separate experiments. Values (arbitrary units, AU) for individual mice (circles) and median values (columns) for groups of mice are shown. Dotted lines indicate the lower detection limits for the ELISA assays. Brackets indicate statistically significant differences between groups (p < 0.05). All experiments were performed twice with similar results, with eight mice per group.
Figure 3 The primary cellular response in the draining lymph node The primary cellular response in the draining PLN was determined five days after a single injection of HBSS (open columns), 10 μg OVA (gray), 40 μg PSP (hatched) or 10 μg OVA + 40 μg PSP (black) into both hind footpads of BALB/cA, NIH and C3H/HeN mice. The total lymph node cell numbers (A) and IL-4 (B), IFN-γ (C) and IL-10 (D) secreted from the PLN cells after ex vivo stimulation with Con A were determined. Values for individual samples (circles) and group median values (columns) are shown. Dotted lines indicate the lower detection limits for the cytokine ELISA assays, and brackets indicate statistically significant differences between groups (p < 0.05).
==== Refs
Pope CA Epidemiology of fine particulate air pollution and human health: biologic mechanisms and Who's at risk? Environ Health Perspect 2000 108 Suppl 4 713 723 10931790
Brunekreef B Sunyer J Asthma, rhinitis and air pollution: is traffic to blame? Eur Respir J 2003 21 913 915 12797480
Lippmann M Frampton M Schwartz J Dockery D Schlesinger R Koutrakis P Froines J Nel A Finkelstein J Godleski J Kaufman J Koenig J Larson T Luchtel D Liu LJ Oberdorster G Peters A Sarnat J Sioutas C Suh H Sullivan J Utell M Wichmann E Zelikoff J The U.S. Environmental Protection Agency Particulate Matter Health Effects Research Centers Program: a midcourse report of status, progress, and plans Environ Health Perspect 2003 111 1074 1092 12826479
Kleeberger SR Genetic aspects of susceptibility to air pollution Eur Respir J 2003 40 Suppl. 52 56 10.1183/09031936.03.00403003
Nel AE Diaz-Sanchez D Ng D Hiura T Saxon A Enhancement of allergic inflammation by the interaction between diesel exhaust particles and the immune system J Allergy Clin Immunol 1998 102 539 554 9802360
Granum B Lovik M The effect of particles on allergic immune responses Toxicol Sci 2002 65 7 17 11752680 10.1093/toxsci/65.1.7
Fujimaki H Saneyoshi K Shiraishi F Imai T Endo T Inhalation of diesel exhaust enhances antigen-specific IgE antibody production in mice Toxicology 1997 116 227 233 9020525 10.1016/S0300-483X(96)03539-1
Takafuji S Suzuki S Koizumi K Tadokoro K Ohashi H Muranaka M Miyamoto T Enhancing effect of suspended particulate matter on the IgE antibody production in mice Int Arch Allergy Appl Immunol 1989 90 1 7 2807574
Maejima K Tamura K Taniguchi Y Nagase S Tanaka H Comparison of the effects of various fine particles on IgE antibody production in mice inhaling Japanese cedar pollen allergens J Toxicol Environ Health 1997 52 231 248 9316645 10.1080/009841097159674
Granum B Gaarder PI Groeng E Leikvold R Namork E Lovik M Fine particles of widely different composition have an adjuvant effect on the production of allergen-specific antibodies Toxicol Lett 2001 118 171 181 11137324 10.1016/S0378-4274(00)00292-7
van Zijverden M van der Pijl A Bol M van Pinxteren FA de Haar C Penninks AH van Loveren H Pieters R Diesel exhaust, carbon black, and silica particles display distinct Th1/Th2 modulating activity Toxicol Appl Pharmacol 2000 168 131 139 11032768 10.1006/taap.2000.9013
Lambert AL Dong W Selgrade MK Gilmour MI Enhanced allergic sensitization by residual oil fly ash particles is mediated by soluble metal constituents Toxicol Appl Pharmacol 2000 165 84 93 10814556 10.1006/taap.2000.8932
Nygaard UC Samuelsen M Aase A Lovik M The Capacity of Particles to Increase Allergic Sensitization is Predicted by Particle Number and Surface Area, Not by Particle Mass Toxicol Sci 2004 82 515 524 15456925 10.1093/toxsci/kfh287
Lovdal T Groeng E Dybing E Lovik M Respiratory allergy and inflammation due to ambient particles (RAIAP), a European-wide assessment. Allergy screening. Toxicol Sci 2003 72 Suppl. 593
Steerenberg PA Withagen CE van Dalen WJ Dormans JA Heisterkamp SH van Loveren H F.R. C Dose dependency of adjuvant activity of particulate matter from five European sites in three different seasons in an ovalbumin mouse model Inhal Toxicol 2005 17 133 145 15788374 10.1080/08958370590904490
Fernvik E Scharnweber T Knopp D Niessner R Vargaftig BB Peltre G Effects of fractions of traffic particulate matter on TH2-cytokines, IgE levels, and bronchial hyperresponsiveness in mice J Toxicol Environ Health A 2002 65 1025 1045 12167217 10.1080/152873902760125200
Yoshino S Hayashi H Taneda S Takano H Sagai M Mori Y Effects of diesel exhaust particle extracts on Th1 and Th2 immune responses in mice Int J Immunopathol Pharmacol 2002 15 13 18 12593783
Heo Y Saxon A Hankinson O Effect of diesel exhaust particles and their components on the allergen- specific IgE and IgG1 response in mice Toxicology 2001 159 143 158 11223170 10.1016/S0300-483X(00)00418-2
Lovik M Hogseth AK Gaarder PI Hagemann R Eide I Diesel exhaust particles and carbon black have adjuvant activity on the local lymph node response and systemic IgE production to ovalbumin Toxicology 1997 121 165 178 9230448 10.1016/S0300-483X(97)00075-9
Al Humadi NH Siegel PD Lewis DM Barger MW Ma JY Weissman DN Ma JK The effect of diesel exhaust particles (DEP) and carbon black (CB) on thiol changes in pulmonary ovalbumin allergic sensitized Brown Norway rats Exp Lung Res 2002 28 333 349 12097228 10.1080/01902140290091976
Granum B Gaarder PI Lovik M IgE adjuvant effect caused by particles - immediate and delayed effects Toxicology 2001 156 149 159 11164617 10.1016/S0300-483X(00)00375-9
van Zijverden M de Haar C van Beelen A van Loveren H Penninks A Pieters R Coadministration of antigen and particles optimally stimulates the immune response in an intranasal administration model in mice Toxicol Appl Pharmacol 2001 177 174 178 11749116 10.1006/taap.2001.9306
Nygaard UC Ormstad H Aase A Lovik M The IgE adjuvant effect of particles: characterisation of the primary cellular response in the draining lymph node Toxicology 2005 206 181 193 15588912 10.1016/j.tox.2004.07.018
Ichinose T Takano H Miyabara Y Yanagisawa R Sagai M Murine strain differences in allergic airway inflammation and immunoglobulin production by a combination of antigen and diesel exhaust particles Toxicology 1997 122 183 192 9328218 10.1016/S0300-483X(97)00096-6
Ichinose T Takano H Sadakane K Yanagisawa R Yoshikawa T Sagai M Shibamoto T Mouse strain differences in eosinophilic airway inflammation caused by intratracheal instillation of mite allergen and diesel exhaust particles J Appl Toxicol 2004 24 69 76 14745849 10.1002/jat.949
Granum B Gaarder PI Eikeset A Stensby BA Lovik M The adjuvant effect of particles-importance of genetic background and pre-sensitisation Int Arch Allergy Immunol 2000 122 167 173 10899759 10.1159/000024393
Holt PG Rose AH Batty JE Turner KJ Induction of adjuvant-independent IgE responses in inbred mice: primary, secondary, and persistent IgE responses to ovalbumin and ovomucoid Int Arch Allergy Appl Immunol 1981 65 42 50 7216513
Brewer JP Kisselgof AB Martin TR Genetic variability in pulmonary physiological, cellular, and antibody responses to antigen in mice Am J Respir Crit Care Med 1999 160 1150 1156 10508801
Larsen ST Hansen JS Thygesen P Begtrup M Poulsen OM Nielsen GD Adjuvant and immuno-suppressive effect of six monophthalates in a subcutaneous injection model with BALB/c mice Toxicology 2001 169 37 51 11696408 10.1016/S0300-483X(01)00484-X
Toews GB Stockley RA Animal Models of Immune Defences Pulmonary Defences 1997 Chichester, John Wiley & Sons 197 215
Sedgwick JD Holt PG Induction of IgE-secreting cells and IgE isotype-specific suppressor T cells in the respiratory lymph nodes of rats in response to antigen inhalation Cell Immunol 1985 94 182 194 3160487 10.1016/0008-8749(85)90095-4
Cuq P Le système lymphatique de la Souris Recl Med Vet 1966 142 1211 1236
Skadhauge LR Christensen K Kyvik KO Sigsgaard T Genetic and environmental influence on asthma: a population-based study of 11,688 Danish twin pairs Eur Respir J 1999 13 8 14 10836316 10.1183/09031936.99.13100899
Okano M Nishizaki K Abe M Wang MM Yoshino T Satoskar AR Masuda Y Harn DAJ Strain-dependent induction of allergic rhinitis without adjuvant in mice Allergy 1999 54 593 601 10435473 10.1034/j.1398-9995.1999.00063.x
Whitehead GS Walker JK Berman KG Foster WM Schwartz DA Allergen-induced airway disease is mouse strain dependent Am J Physiol Lung Cell Mol Physiol 2003 285 32 42
Shinagawa K Kojima M Mouse model of airway remodeling: strain differences Am J Respir Crit Care Med 2003 168 959 967 12857720 10.1164/rccm.200210-1188OC
Ohtsuka Y Clarke RW Mitzner W Brunson K Jakab GJ Kleeberger SR Interstrain variation in murine susceptibility to inhaled acid-coated particles Am J Physiol Lung Cell Mol Physiol 2000 278 469 476
Gurvich AE Korukova A Induction of abundant antibody formation with a protein-cellulose complex in mice J Immunol Methods 1986 87 161 167 2419450 10.1016/0022-1759(86)90526-0
Miyajima I Dombrowicz D Martin TR Ravetch JV Kinet JP Galli SJ Systemic anaphylaxis in the mouse can be mediated largely through IgG1 and Fc gammaRIII. Assessment of the cardiopulmonary changes, mast cell degranulation, and death associated with active or IgE- or IgG1-dependent passive anaphylaxis J Clin Invest 1997 99 901 914 9062348
Strait RT Morris SC Yang M Qu XW Finkelman FD Pathways of anaphylaxis in the mouse J Allergy Clin Immunol 2002 109 658 668 11941316 10.1067/mai.2002.123302
Romagnani S Oppenheim JJ and Feldmann M TH1/TH2 Interleukines Cytokine Reference, Volume 1: Ligands Cytokine Reference 2001 London, Academic Press 99 108
Yang M Cao Y Mohapatra SS CD8+ T cells inhibit immunoglobulin E synthesis in low responder SJL/J mice Immunology 1998 93 230 237 9616373 10.1046/j.1365-2567.1998.00415.x
Cottrez F Groux H Specialization in tolerance: innate CD(4+)CD(25+) versus acquired TR1 and TH3 regulatory T cells Transplantation 2004 77 12 15 10.1097/01.TP.0000106471.23410.32
Curotto de Lafaille MA Lafaille JJ CD4(+) regulatory T cells in autoimmunity and allergy Curr Opin Immunol 2002 14 771 778 12413528 10.1016/S0952-7915(02)00408-9
Romagnani S Kay AB Regulation of IgE synthesis Allergy and Allergic Diseases 1997 London, Blackwell Sciences 96 112
Randolph GJ Inaba K Robbiani DF Steinman RM Muller WA Differentiation of phagocytic monocytes into lymph node dendritic cells in vivo Immunity 1999 11 753 761 10626897 10.1016/S1074-7613(00)80149-1
Nygaard UC Aase A Lovik M Ambient air particles from four European cities increase the primary cellular response to allergen in the draining lymph node Toxicology 2005 207 241 254 15596255 10.1016/j.tox.2004.09.016
Ormstad H Johansen BV Gaarder PI Airborne house dust particles and diesel exhaust particles as allergen carriers Clin Exp Allergy 1998 28 702 708 9677134 10.1046/j.1365-2222.1998.00302.x
Rowe J Heaton T Kusel M Suriyaarachchi D Serralha M Holt BJ de Klerk N Sly PD Holt PG High IFN-gamma production by CD8+ T cells and early sensitization among infants at high risk of atopy J Allergy Clin Immunol 2004 113 710 716 15100678 10.1016/j.jaci.2003.12.585
Medoff BD Sauty A Tager AM Maclean JA Smith RN Mathew A Dufour JH Luster AD IFN-gamma-inducible protein 10 (CXCL10) contributes to airway hyperreactivity and airway inflammation in a mouse model of asthma J Immunol 2002 168 5278 5286 11994485
Randolph DA Stephens R Carruthers CJ Chaplin DD Cooperation between Th1 and Th2 cells in a murine model of eosinophilic airway inflammation J Clin Invest 1999 104 1021 1029 10525040
Secord EA Rizzo LV Barroso EW Umetsu DT Thorbecke GJ DeKruyff RH Reconstitution of germinal center formation in nude mice with Th1 and Th2 clones Cell Immunol 1996 174 173 179 8954617 10.1006/cimm.1996.0307
Ormstad H Groeng EC Lovik M Hetland G The fungal cell wall component beta-1,3-glucan has an adjuvant effect on the allergic response to ovalbumin in mice J Toxicol Environ Health A 2000 61 55 67 10990163 10.1080/00984100050116780
Ormstad H Gaarder PI Johansen BV Lovik M Airborne house dust elicits a local lymph node reaction and has an adjuvant effect on specific IgE production in the mouse Toxicology 1998 129 227 236 9772100 10.1016/S0300-483X(98)00079-1
|
15967044
|
PMC1182367
|
CC BY
|
2021-01-04 16:03:31
|
no
|
BMC Immunol. 2005 Jun 21; 6:11
|
utf-8
|
BMC Immunol
| 2,005 |
10.1186/1471-2172-6-11
|
oa_comm
|
==== Front
BMC ImmunolBMC Immunology1471-2172BioMed Central London 1471-2172-6-141600198110.1186/1471-2172-6-14Research ArticleElevated ex vivo monocyte chemotactic protein-1 (CCL2) in pulmonary as compared with extra-pulmonary tuberculosis Hasan Zahra [email protected] Irfan [email protected] Bushra [email protected] M Aslam [email protected] Akbar [email protected] Rabia [email protected] Department of Pathology and Microbiology, The Aga Khan University, Karachi, Pakistan2 Department of Medicine, The Aga Khan University, Karachi, Pakistan2005 7 7 2005 6 14 14 12 4 2005 7 7 2005 Copyright © 2005 Hasan 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
Tuberculosis causes 3 million deaths annually. The most common site of tuberculosis is pulmonary however; extra-pulmonary forms of the disease also remain prevalent. Restriction of Mycobacterium tuberculosis depends on effective recruitment and subsequent activation of T lymphocytes, mononuclear and polymorphonuclear cells to the site of infection. Tumor necrosis factor (TNF)-α is essential for granuloma formation and is a potent activator of monocyte chemotactic protein (MCP-1, CCL2). CCL2 is essential for recruitment of monocytes and T cells and has been shown to play a role in protection against tuberculosis. Interleukin -8 (CXCL8) is a potent activator of neutrophils. Increased levels of CCL2, CXCL8 and TNFα are reported in tuberculosis but their significance in different forms of tuberculosis is as yet unclear. We have used an ex vivo assay to investigate differences in immune parameters in patients with either pulmonary or extra-pulmonary tuberculosis.
Methods
Serum levels of CCL2, CXCL8 and TNFα were measured in patients with pulmonary tuberculosis (N = 12), extra-pulmonary tuberculosis (N = 8) and BCG-vaccinated healthy volunteers (N = 12). Whole blood cells were stimulated with non-pathogenic Mycobacterium bovis bacille-Calmette Guerin (BCG) vaccine strain or bacterial lipopolysaccharide (LPS) and cyto/chemokines were monitored in supernatants.
Results
Circulating serum levels of CXCL8 and TNFα were raised in all tuberculosis patients, while CCL2 levels were not. There was no difference in spontaneous cytokine secretion from whole blood cells between patients and controls. M. bovis BCG-induced ex vivo CCL2 secretion was significantly greater in pulmonary as compared with both extra-pulmonary tuberculosis patients and endemic controls. In response to LPS stimulation, patients with pulmonary tuberculosis showed increased CCL2 and TNFα responses as compared with the extra-pulmonary group. BCG-, and LPS-induced CXCL8 secretion was comparable between patients and controls.
Conclusion
CCL2 is activated by TNFα and is essential for recruitment of monocytes and T cells to the site of mycobacterial infection. Increased CCL2 activation in pulmonary tuberculosis may result in a stronger cellular response as compared with extra-pulmonary tuberculosis patients, and this may contribute to the localization of infection to the pulmonary site.
==== Body
Background
Tuberculosis is a major cause of morbidity and mortality worldwide, causing approximately 3 millions deaths annually [1]. Mycobacterium tuberculosis, the causative agent of tuberculosis is a successful pathogen due to its ability to down regulate host immune responses. The primary site of infection for M. tuberculosis is the alveolar macrophage. In cases where pulmonary infection cannot be controlled the organism disseminates via blood to other sites. Pulmonary involvement is seen in the majority of tuberculosis cases however infections of extra-pulmonary sites such as lymph nodes, skeletal, abdominal and genito-urinary sites also remain common [2].
Local cellular responses required for restriction of infection with M. tuberculosis are fairly well understood. Granulomas which are the characteristic histopathologic lesions of tuberculosis can be formed in any infected tissue. Granulomas exhibit a characteristic mixture of macrophages and lymphocytes and the outcome of infection depends on the interplay between immune activating cytokines produced at this site. T cell interferon -gamma (IFN-γ) and macrophage activating tumor necrosis factor-alpha (TNFα) are critical for protection and play a central role in granuloma formation [3,4]. Recent studies have also indicated an important role for chemokines in granuloma formation [5,6].
Chemokines are small mass chemotactic cytokines produced by epithelial cells, mast cells, monocytes and neutrophils. Chemokines can be grouped into structurally different families including C-C chemokines; monocyte chemotactic protein-1 (MCP-1, CCL2), macrophage inflammatory protein-1α (MIP-1α, CCL3), MIP-1β (CCL4) and regulated-upon-activation, normally T-cell-expressed and-secreted (RANTES, CCL5) and C-X-C chemokines such as, IL-8 (CXCL8), MIG (CXCL9) and IP-10 (CXCL10). These are all activated by M. tuberculosis [7]. CCL2 is the most potent chemoattractant and activator for monocytes and attracts CD4 and γδ T cells [8]. CCL2 has also been shown to play a role in protection against murine tuberculosis [9-11]. CXCL8 is the most potent attractant and activating factor for neutrophils and is chemotactic for lymphocytes [12,13]. CXCL8 is required for stimulation of a pro-inflammatory response against M. tuberculosis and its components [14].
Previous studies have shown that in vivo (circulating) serum levels of IFNγ, IL-10 [15], TNFα [16], CXCL8 and IL-6 [17] are raised in pulmonary tuberculosis. Circulating levels of chemokines CXCL8, CCL2 and RANTES are raised in bronchoalveolar lavage fluid (BALF) and alveolar macrophages from pulmonary tuberculosis patients [18,19]. Increased CXCL8 and CCL2 activation is also observed in patients with tuberculous pleurisy [20] and meningitis [21,22]. In addition, the expression levels of cytokines and chemokines have been correlated with disease severity [15] and with subsequent recovery after treatment [23,24].
We have focused on immune responses in newly diagnosed untreated tuberculosis patients with either pulmonary, or extra-pulmonary disease. We observed that systemic levels of CXCL8 and TNFα were raised in all patients while CCL2 was not. Using an ex vivo whole blood assay we observed differential Mycobacterium-induced and LPS stimulated responses between patients with pulmonary and extra-pulmonary tuberculosis. Mycobacterium-induced CCL2 and TNFα, and LPS-induced TNFα responses were greater in pulmonary as compared with extra-pulmonary tuberculosis. However, both Mycobacterium-, and LPS-induced CXCL8 responses were comparable between patient groups.
Results
Patient characteristics
Hematological characteristics of tuberculosis patients and healthy endemic controls are shown in Table 1. Median values obtained for pulmonary (Pul-TB) and extra-pulmonary tuberculosis (EPul-TB) were not different from controls, although for total lymphocyte and neutrophil counts were higher in tuberculosis patients. The erythrocyte sedimentation rate was raised in both patient groups as compared with controls.
Serum CCL2, CXCL8 and TNFα in tuberculosis patients
We determined circulating levels of CCL2, CXCL8 and TNFα in sera of pulmonary and extra-pulmonary tuberculosis patients and controls (Figure 1). No difference was observed in serum CCL2 levels between the Pul-TB, EPul-TB and control (EC) groups (Fig. 1A). However, serum CXCL8 was significantly greater in both Pul-TB (P = 0.003) and EPul-TB (P = 0.002) as compared with controls (Fig. 1B). Serum TNFα was also significantly greater in both Pul-TB (P = 0.001) and in EPul-TB (P = 0.001) as compared with controls (Fig. 1C).
BCG-, and LPS -induced CCL2, CXCL8 and TNFα
A time course profile of M. bovis BCG – and LPS-induced CCL2, CXCL8 and TNFα secretion was determined in healthy controls, see Fig. 2. BCG infection induced CXCL8 secretion within 6 – 18 h post stimulation (Fig. 2B), followed by TNFα (Fig. 2C) and then by CCL2 (Fig. 2A). LPS induced higher levels of all three cytokines as compared with BCG. In addition, CXCL8, TNF α and CCL2 were secreted more rapidly by LPS. Peak secretion of chemo/cytokines was observed between 18 and 48 h post-stimulation, therefore BCG and LPS induced responses were subsequently studied at these time intervals.
CCL2, CXCL8 and TNFα secretion in whole blood cells from tuberculosis patients
Baseline (spontaneous) levels of CCL2, CXCL8 and TNFα were determined in Pul-TB, EPul-TB and control groups. Chemokine activity was measured in supernatants of un-stimulated whole blood cells after 18 h and 48 h of ex vivo culture. At 18 h, increased levels of CCL2 were observed in the Pul-TB group; 438 pg/ml (median) (range, 0–3000 pg/ml) as compared with EPul-TB; 35 pg/ml (0–774 pg/ml) and controls; 42 pg/ml (0–102 pg/ml). Spontaneous secretion of CXCL8 after 18 h was : Pul-TB; 284 pg/ml (15–1927 pg/ml), EPul-TB; 681 pg/ml (1–2076 pg/ml) and controls; 121 pg/ml (8–2546 pg/ml). Spontaneous TNFα secretion at 18 h was: Pul-TB; 5 pg/ml (0–55 pg/ml), E-Pul-TB; 1 pg/ml (0–585 pg/ml) and controls; 25 pg/ml (0–102 pg/ml). There were no significant differences between groups in their baseline levels of CCL2, CXCL8 or TNFα (P > 0.05). Spontaneous levels of cytokines measured after 48 h (data not shown) of in vitro culture showed a similar trend as observed at 18 h, with no difference between groups.
We subsequently investigated LPS-induced CCL2, CXCL8 and TNFα in tuberculosis patients. Table 2 illustrates chemokine activity measured at 18 and 48 h post LPS stimulation. Spontaneous chemokine secretion (indicated above) have been subtracted from the values indicated in the table. LPS-induced CCL2 activation in Pul-TB was significantly greater than in EPul-TB at 18 h (*1, P = 0.01) and at 48 h (*2, P = 0.02) post-stimulation. LPS-induced CXCL8 secretion was comparable between patients and endemic controls. At 18 h, TNFα was significantly greater in Pul-TB as compared with EPul-TB (*4, P = 0.003) and also as compared with controls (*3, P = 0.003), whilst no difference was observed between these groups at 48 h.
BCG-induced chemokine secretion in tuberculosis patients and controls at 18 h post-stimulation is illustrated in Fig. 3. Baseline levels of chemokine secretion have been subtracted and are described in the text above. BCG- induced CCL2 was significantly greater in pulmonary tuberculosis as compared with extra-pulmonary at both 18 h (P = 0.001) (Fig. 3A) and 48 h (P = 0.02, data not shown) post-stimulation. CCL2 secreted in whole blood cells from pulmonary patients was also greater as compared with controls at 18 h (P = 0.05) (Fig. 3A) and 48 h (P < 0.001, data not shown). There was no difference in BCG-induced CXCL8 between pulmonary or extra-pulmonary tuberculosis patients. Similarly, no difference was noted between patients and controls at both time points studied (18 h, Fig. 3B; 48 h, data not shown). In addition, we did not observe any differences in BCG-induced TNFα between controls and tuberculosis patients, P > 0.05 (Fig. 3C) at either time interval studied. Although the trend of TNFα secretion was lower in the extra-pulmonary group as compared with the pulmonary group, this difference was also not significant.
Discussion
We have used an ex vivo whole blood assay to study differences in immune parameters in pulmonary and extra-pulmonary tuberculosis. We observed that BCG-induced CCL2 responses in pulmonary tuberculosis were greater as compared to both extra-pulmonary tuberculosis and control groups. LPS stimulated CCL2 and TNFα responses were also raised in pulmonary tuberculosis as compared with extra-pulmonary tuberculosis patients.
We included both PPD skin test positive and negative individuals to be representative of our healthy endemic population, where a large proportion of BCG vaccinated individuals remain non-reactive to the PPD skin test antigen [25]. Our tuberculosis patient group was small because of stringent selection criteria. We excluded most of those patients who had received some form of anti-tuberculous treatment for any length of time. Due to poor therapeutic practices in Pakistan the majority of patients presenting to a tertiary care center such as AKUH have received variable course anti-tuberculous therapy [26]. There were three times as many female than male patients in our study and this ratio is representative of female preponderance in the disease pattern in Pakistan [27]. In congruence with previous studies, neutrophils and monocyte counts were not significantly altered in patient groups [28]. However, erythrocyte sedimentation was significantly higher in tuberculosis patients, reflective of active infection [28]. Two extra-pulmonary tuberculosis patients (abdominal and lymphnode respectively) showed lung involvement on chest X-ray. These were grouped based on their predominant extra-pulmonary disease site and found that their responses correlated with others in the group.
The raised circulating levels of CXCL8 we observe in tuberculosis patients correlate with previous reports [16,17]. Increased CXCL8 has been associated with increased neutrophils in broncho-alveolar lavage [18] and pleural fluid [29]. However, we did not find increased neutrophils or monocytes in peripheral circulation of our patients. We observed raised TNFα in sera of tuberculosis patients. This is contrast to reports by Juffermans et al. who found that TNFα was not raised in sera of tuberculosis patients [30], while additional reports by Olobo et al. indicate that serum TNFα is increased in healthy contacts [31]. We did not observe any increase in circulating CCL2, correlating with reports by Lee et al. [32] who compared sera from tuberculosis patients with healthy tuberculin reactive controls from a disease endemic region. This was in contrast with Juffermans et al. who observed raised CCL2 in sera of tuberculosis patients as compared with healthy individuals from a non-endemic region [16].
Ex vivo stimulation of whole blood leucocytes provides useful information on control of cytokine secretion, since it utilizes the host environment [33]. Whole blood assays have been shown to provide useful insight into the immunopathology of tuberculosis infection [34]. The time course profiles of CCL2, CXCL8 and TNFα secretion we observed in response to BCG and LPS corresponded to previous reports of Mycobacterium and LPS induced activation of purified monocytes and adherent cells [19,35,36]. Further, after LPS stimulation we observed greater cytokine activation than after mycobacterial stimulation as has been shown in the case of cytokines such as CXCL8 [37]. Therefore the data we obtained using our whole blood assay correlated with patterns observed using purified subsets of cells.
Increased CCL2 in pulmonary tuberculosis may be related to localization of infection to the lung and may subsequently prevent disease dissemination. We observed raised CCL2 in pulmonary tuberculosis compared with controls in response to BCG stimulation. Raised CCL2 activation in tuberculosis correlates with previous studies [19,32,38]. However, this contrasts with reports by Lee et al. who did not observe any increase in CCL2 secreted by adherent cells from pulmonary tuberculosis patients as compared with healthy tuberculin reactors – in response to PPD or M. tuberculosis 30 kD antigen [32]. One difference as compared to our study was that Lee et al. used healthy controls with a PPD skin test induration > 15 mm, while our control group comprised of either PPD negative or PPD positive individuals with induration > 10 mm. In addition, the use of live Mycobacteria and unprocessed whole blood is likely to give more physiologically relevant results for in vivo chemokine and cytokine activation as compared with recombinant antigen or subunit preparations. We also observed greater CCL2 activation in pulmonary tuberculosis as compared with extra-pulmonary disease in response to both BCG and LPS stimulation.
LPS-induced CXCL8 was not raised in whole blood cells from patients and this was paradoxical with increased levels in sera of these patients. Differences between in vivo (serum and plasma) and in vitro (PBMC-induced cytokine) activation have been shown previously [39]. Previously, we have shown increased spontaneous CXCL8 in disseminated mycobacterial (leprosy) infections while inducible CXCL8 was not raised in patients [40]. Therefore, increased levels of CXCL8 in sera may be indicative of pathology in these patients and may not contribute to the effective recruitment of cells.
TNFα levels were raised in pulmonary as compared with both extra-pulmonary tuberculosis and controls in response to LPS but not BCG stimulation. This difference in TNFα levels was evident at the earlier (18 h) but not later time interval. Probably, due to the reduced time course profile of TNFα by 48 h. Lack of difference in TNFα in response to BCG between groups, may also be due to reduced overall levels induced by mycobacterium as compared with the LPS.
Conclusion
Macrophages and other polymorphonuclear cells play essential roles in initiation and maintenance of immune responses to M. tuberculosis. Pro-inflammatory TNFα, CCL2 and CXCL8 are important in maintaining a balance between recruitment of other macrophages and also T cells to restrict infection. CCL2 contributes to anti-mycobacterial inflammatory response by attracting monocytes and T lymphocytes. Raised CCL2 in pulmonary tuberculosis may lead to greater monocyte mediated immunity. Reduced CCL2 concomitant with lowered TNFα in extra-pulmonary tuberculosis may lead to lowered protective cellular responses and increased dissemination in these patients.
Methods
Subject selection and diagnosis
Patients were recruited from the Aga Khan University (AKUH) and Masoomeen Hospitals in Karachi. Ethical Clearance was obtained from the AKUH Human Subjects Protection and Ethical Review Committee and samples were taken from patients with informed consent. All study subjects were examined and investigated by infectious diseases consultants. Patients with significant co-morbid conditions including diabetes mellitus, chronic renal failure, chronic liver disease, those on high dose corticosteroid therapy and those with HIV/AIDS were excluded from this analysis. All patients were untreated except for two who had been on anti-tuberculous therapy for 2 months each due to delays in diagnostic confirmation of extra-pulmonary tuberculosis. Patients were diagnosed by clinical examination, chest X-ray, sputum M. tuberculosis microscopy and culture. Chest X-rays were evaluated by one of the consulting physicians using the classification of the National Tuberculosis Association of the USA into minimal, moderate and advanced lung tissue involvement [41]. Microscopy was performed using Ziehl-Neelsen staining. Cultures were done by means of a set of Lowenstein Jensen slants with and without glycerol. Purified protein derivative (PPD) skin test was carried out on each study subject. PPD skin test negative were those with induration < 10 mm, with PPD positive with induration = 10 mm. Hematological parameters for all subjects were noted on the date of enrollment and shown in Table 1. The pulmonary tuberculosis (Pul-TB) group comprised of 10 patients (3 males and 7 females) with moderate disease and no other site involved except the lung. The extra-pulmonary tuberculosis (EPul-TB) group comprised of 8 patients (2 males and 6 females) with tuberculosis in the following sites; lymph nodes (N = 3), pericardial effusion (N = 2), meningitis (N = 1), abdomen (N = 2). Of the EPul-TB patients, 6/8 were confirmed on M. tuberculosis smear/culture and/or radiology, while 2/8 were diagnosed on the basis of clinical presentation and a favorable response to anti-tuberculosis drugs. Two of the EPul-TB patients also showed pulmonary involvement on Chest X-ray. Twelve BCG-vaccinated healthy endemic controls (EC) (9 males and 3 females) who were staff at the AKUH were also included in the study. PPD+ volunteers had normal chest X-Ray and no clinical evidence of active disease.
Mycobacterium culture
Mycobacterium bovis BCG (Montreal vaccine strain) was grown to logarithmic phase in 7H9 Middlebrook medium with supplements (DIFCO Laboratories, MI, USA). Mycobacteria were frozen in growth medium containing 20 % glycerol and stored at -70°C in single use aliquots for the assay as described previously [40]. Mycobacterial viability was confirmed prior to assays by fluorescent staining [42] and by colony counts on Middlebrook 7H11 agar.
Whole blood assays
Venous blood was diluted 1: 10 in RPMI-1640 medium each well containing 106 whole blood cells was set up as per protocol [34]. Freshly thawed aliquots of BCG were washed in PBS and diluted to infect cells as described previously [43]. Each was stimulated with either E. coli lipopolysaccharide (LPS) (Sigma, USA) 0.1 μg/ml or mycobacteria at 2 × 105 CFU/ml (at a ratio of 0.2 bacteria per cell). Supernatants were collected at 0, 1, 6, 18, 48 and 120 h (5 days) post infection for chemokine measurement, spun to collect cellular debris and stored at -70 C until tested. BCG innoculum was tested for LPS contamination using the E-Toxate Kit (Sigma, USA).
ELISAs
ELISA reagents for CCL2, CXCL8 and TNFα were from R&D Systems (USA). Assays were carried out according to the manufacturer's recommendation and as reported previously [43]. The lower limit of detection for CCL2 was 7.8 pg/ml, and 3.9 pg/ml for TNFα and CXCL8.
Statistical analysis
Data is expressed as Mean (SD) or Median as described in figures. Statistical analysis was carried out using Mann-Whitney U test for Pair-wise analysis of groups using the SPSS package, USA and also ANOVA analysis.
List of abbreviations used
BCG, bacille-Calmette Guerin vaccine strain; LPS – lipo polysaccharide; TNFα – tumor necrosis factor alpha; Pul-TB – pulmonary tuberculosis; EPul-TB – extra-pulmonary tuberculosis; EC – endemic controls; CCL2 – C-C chemokine ligand 2; CXCL8 – C-X-C chemokine ligand 8.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
This work was conceived and written up by ZH. IZ carried out the immunological assays and AK helped coordinate patient samples and ex vivo assays. BJ and MAK recruited patients for the study and gave input into the data analyses and interpretation. RH helped in preparation of the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
This study was supported by a Seed Money Grant, The Aga Khan University, Karachi, Pakistan. Thanks for help with patient recruitment to Dr. Ghaffar Dawood, Masoomeen Hospital, Karachi, Pakistan.
Figures and Tables
Figure 1 Raised CXCL8 and TNFα in sera of tuberculosis patients. Circulating levels of chemo/cytokines were determined in sera of endemic controls (EC, N = 12), pulmonary tuberculosis (Pul-TB, N = 10) and extra-pulmonary tuberculosis (EPul-TB, N = 8) patients. Values for individual subjects are shown, with the median value indicated by a horizontal line. A. CCL2, B. CXCL8 and C. TNFα. Statistical analysis was carried out using the Mann Whitney-U test with significant differences, P < 0.05, '*'.
Figure 2 Time-dependent secretion of Mycobacterium-, and LPS- induced CCL2, CXCL8 and TNFα. Whole blood cells (106) from healthy controls (N = 12) were stimulated with M. bovis BCG (2 × 105 CFU/ml, ratio of 0.2 bacilli per cell) and LPS at 0.1 μg/ml. Supernatants were harvested at 1, 6, 18, 48 and 120 h post-stimulation and tested for cytokine activity. Mean values are illustrated with SD indicated as 'y' error bars. A. CCL2, B. CXCL8, C. TNFα. Un-stimulated cells (spon) '◇', BCG stimulated '■' and LPS stimulated '▲'.
Figure 3 Increased BCG-induced CCL2 secretion in pulmonary tuberculosis. Whole blood cells were stimulated with M. bovis BCG as described in Fig. 2 and chemo/cytokine secretion measured at 18 h post-stimulation. Data for individual subjects; controls 'EC' (N = 12), pulmonary 'Pul-TB' (N = 10) and extra pulmonary 'EPul-TB' (N = 8) tuberculosis is illustrated. The median value is indicated by a horizontal line. Baseline levels of cytokines have been subtracted in each case 'Δ '. A. CCL2, B. CXCL8 and C. TNFα. Pair-wise differences were calculated using the Mann-Whitney U test with P < 0.05, '*'.
Table 1 Hematological characteristics of study subjects
Group Pul-TB, N = 10 Median (range) EPul-TB, N = 8 Median (range) Controls, N = 12 Median (range)
Age (y) 25 (15 – 50) 32 (15 – 58) 26 (18–29)
TLC (10 xE9/L) 9.8 (7 – 22.2) 9.4 (6.4 – 16.3) 7.2 (3 – 9)
Neutrophils (%) 64 (38 – 87) 70 (60 – 90 56 (38.1 – 70.6)
Monocytes (%) 6.7 (3.6 – 9.2) 7.4 (1.7 – 11.4) 7.2 (5.8 – 9.4)
ESR (mm/h) 60* (40–100) 50* (20–80) 0–3
M. tuberculosis+a 9/10 3 / 8 NA
%PPD+ (No. tested) 80 (8) 50 (4) 50 (12)
a Sputum/Biopsy/Culture positive for M. tuberculosis. Pulmonary tuberculosis (Pul-TB), extra-pulmonary tuberculosis (EPul-TB), purified protein derivative (PPD) skin test, total lymphocyte count (TLC), erythrocyte sedimentation rate (ESR), NA -not applicable. '*' P < 0.05 as compared with Controls.
Table 2 LPS-induced CCL2, CXCL8 and TNFα secretion in TB patients
Chemo/cytokines pg/ml Median (range)
Group Time (h) ΔCCL2 ΔCXCL8 ΔTNFα
EC (N = 12) 18 754 (17–3058) 1712 (788–2786) *3 1153 (0–1489)
48 1881 (713–7164) 1996 (1823–2977) 625 (211–858)
Pul (N = 10) 18 *1 1218 (765–2254) 1175 (318–2770) *3,4 2154 (1244–4852)
48 *2 3398 (1643–7012) 1190 (0–3249) 957 (74–1869)
E-Pul (N = 8) 18 *1 33 (0–818) 727 (0–1219) *4 577 (0–2180)
48 *2 32 (0–1194) 915 (0–1574) 73 (0–1821)
All values are 'Δ' after baseline subtraction of CCL2, CXCL8 or TNFα from each group. Pair-wise analysis was carried out to determine differences between groups using Mann-Whitney – U test '*' P ≤ 0.05. '1–4' defines pairs of values that were statistically significant from each other; Endemic controls (EC), Pul-TB (pulmonary tuberculosis), EPul-TB (extra-pulmonary tuberculosis).
==== Refs
WHO Global Tuberculosis Control 2003 99 101
Fanning A Tuberculosis: 6. Extrapulmonary disease C M A J 1999 160 1597 1603
Smith S Liggitt D Jeromsky E Tan X Skerrett SJ Wilson CB A local role for tumor necrosis factor alpha in the pulmonary inflammatory response to Mycobacterium tuberculosis infection Infect Immun 2002 70 2082 2089 11895974 10.1128/IAI.70.4.2082-2089.2002
Kindler V Sappino AP Grau GE Piguet PF Via LE Vassalli P The inducing role of tumor necrosis factor in the development of bactericidal granulomas during BCG infection. Cell 1989 56 731 740 2647299 10.1016/0092-8674(89)90676-4
Qiu B Frait KA Reich F Komuniecki E Chensue SW Chemokine expression dynamics in mycobacterial (Type-1) and schistosomal (Type-2) antigen-elicited pulmonary granuloma formation Am J Pathol 2001 158 1503 1515 11290568
Roach DR Bean AGD Demangel C France MP Briscoe H Britton WJ TNF regulates chemokine induction essential for cell recruitment, granuloma formation, and clearance of mycobacterial infection J Immunol 2002 168 4620 4627 11971010
Algood HMS Chan J Flynn JL Chemokines and tuberculosis Cytokine-Growth-Factor-Rev 2003 14 467 477 14563349 10.1016/S1359-6101(03)00054-6
Penido C Vieira-de-Abreu A Bozza MT Castro-Faria-Neto HC Bozza PT Role of monocyte chemotactic protein-1/CC chemokine ligand 2 on gd T lymphocyte trafficking during inflammation induced by lipopolysaccharide of Mycobacterium bovis bacille Calmette-Guerin J Immunol 2003 171 6788 6794 14662884
Kipnis A Basaraba RJ Orme IM Cooper AM Role of chemokine ligand 2 in the protective response to early murine pulmonary tuberculosis Immunology 2003 109 547 551 12871221 10.1046/j.1365-2567.2003.01680.x
Scott HM Flynn JL Mycobacterium tuberculosis in chemokine receptor 2-deficient mice: Influence of dose on disease progression Infect Immun 2002 70 5946 5954 12379669 10.1128/IAI.70.11.5946-5954.2002
Peters W Scott HM Chambers HF Flynn JL Charo IF Ernst JD Chemokine receptor 2 serves an early and essential role in resistance to Mycobacterium tuberculosis Proc-Natl-Acad-Sci-USA 2001 98 7958 7963 11438742 10.1073/pnas.131207398
Moser B Wolf M Walz A Loetscher P Chemokines: multiple levels of leukocyte migration control Trends Immunol 2004 25 75 84 15102366 10.1016/j.it.2003.12.005
Park JS Kim YS Jee YK Myong NH Lee KY Interleukin-8 production in tuberculous pleurisy: role of mesothelial cells stimulated by cytokine network involving tumour necrosis factor-a and interleukin-1b Scand J Immunol 2003 57 463 469 12753503 10.1046/j.1365-3083.2003.01201.x
Wilkinson PC Newman I Identification of IL-8 as locomotor attractant for activated human lymphocytes in mononuclear cell cultures with anti-CD3 or purified protein derivative of Mycobacterium tuberculosis J Immunol 1992 149 2689 2694 1401908
Dlugovitzky D Bay ML Rateni L Fiorenza G Vietti L Farroni MA Influence of disease severity on nitrite and cytokine production by peripheral blood mononuclear cells (PBMC) from patients with pulmonary tuberculosis (TB) Clin Exp Immunol 2000 122 343 349 11122239 10.1046/j.1365-2249.2000.01394.x
Juffermans NP Verbon A Deventer SJH Deutekom H Belisle JT Ellis ME Speelman P van der Poll T Elevated chemokine concentrations in sera of human immunodeficiency virus (HIV)-seropositive and HIV-seronegative patients with tuberculosis: a possible role for mycobacterial lipoarabinomannan Infect Immun 1999 67 4295 4297 10417209
Friedland JS Hartley JC Hartley CG Shattock RJ Griffin GE Inhibition of ex vivo proinflammatory cytokine secretion in fatal Mycobacterium tuberculosis infection Clin Exp Immunol 1995 100 233 238 7743661
Kurashima K Mukaida N Fujimura M Yssui M Nakazumi Y Matusda T Matsushima K Elevated chemokine levels in bronchalveolar lavage fluid of tuberculosis patients Am J Respir Crit Care Med 1997 155 1474 1477 9105097
Sadek MI Sada E Toosi Z Schwander SK Rich EA Chemokines induced by infection of mononuclear phagocytes with mycobacteria and present in lung alveoli during active pulmonary tuberculosis Am J Respir Cell Mol Biol 1998 19 513 21 9730880
Fietta AM Morosini M Meloni F Bianco AM Pozzi E Pharmacological analysis of signal transduction pathways required for Mycobacterium tuberculosis-induced IL-8 and MCP-1 production in human peripheral monocytes Cytokine 2002 19 242 249 12393171 10.1016/S1043-4666(02)91968-X
Mastroianni CM Lancella L Mengoni F Lichtner M Santopadre P Dagostino C Ticca F Vullo V Chemokine profiles in the cerebrospinal fluid (CSF) during the course of pyogenic and tuberculous meningitis Clin Exp Immunol 1998 114 210 214 9822278 10.1046/j.1365-2249.1998.00698.x
Thwaites GE Simmons CP Quyen NTH Chau TTH Mai PP Dung NT Phu NH White NP Hien TT Pathophysiology and prognosis in Vietnamese adults with tuberculous meningitis J Infect Dis 2003 188 1105 1115 14551879 10.1086/378642
Verbon A Juffermans NP Deventer SJH Speelman P Deutekom H van der Poll T Serum concentrations of cytokine in patients with active tuberculosis (TB) and after treatment Clin Exp Immunol 1999 115 110 113 9933428 10.1046/j.1365-2249.1999.00783.x
Riberio-Rodrifues R Co TR Johnson JL Riberio F Palaci M Sa RT Maciel EL Lima FEP Dettoni V Toosi Z Boom WH Dietze R Ellner J Hirsch CS Sputum cytokine levels in patients with pulmonary tuberculosis as early markers of mycobacterial clearance Clin Diagn Lab Immunol 2002 9 818 823 12093679 10.1128/CDLI.9.4.818-823.2002
Menzies R Vissandjee B Effect of bacille Calmette-Guerin vaccination on tuberculin reactivity Am Rev Resp Dis 1992 145 621 625 1546843
Khan J Malik A Hussain H Ali NK Akbani F Hussain SJ Kazi GN Hussain SF Tuberculosis diagnosis and treatment practices of private physicians in Karachi, Pakistan East Mediterr Health J 2003 9 769 775 15748073
Hussain SF Irfan M Abbasi M Anwer SS Davidson S Haqqee R Khan JA Islam M Clinical characteristics of 110 miliary tuberculosis patients from a low HIV prevalence country Int J Tuberc Lung Dis 2004 8 493 499 15141744
Hussain R Hasan R Khurshid M Sturm AW Ellner JJ Dawood G Pulmonary tuberculosis in a BCG vaccinated area: Relationship of disease severity with immunological and hematological parameters and drug resistance patterns Southeast Asian J Trop Med Public Health 1996 27 1 6
Barnes PF Cave MD Molecular epidemiology of tuberculosis New Engl J Med 2003 349 1149 1156 13679530 10.1056/NEJMra021964
Juffermans NP Verbon A vn Deventer SJH van Deutekom H Speelman P van der Poll T Tumor necrosis factor and interleukin-1 inhibitors as markers of disease activity of tuberculosis Am J Resp Crit Care Med 1998 157 1328 1331 9563758
Olobo JO Geletu M Demissie A Eguale T Hiwot K Aderaye G Britton S Circulating TNF-α, TNF-β and IL-10 in tuberculosis patients and healthy contacts Scan J Immunol 2001 53 85 91 10.1046/j.1365-3083.2001.00844.x
Lee JS Song CH Lim JH Kim HJ Park JK Paik TH Jo EK Monocyte chemotactic protein-1 production in patients with active pulmonary tuberculosis and tuberculous pleurisy Inflamm res 2003 52 297 304 12861395
DeGroote D Zangerle PF Gevaert Y Direct stimulation of cytokines (IL-1β, TNF-α, IL-6, IL-2, IFN-γ and GM-CSF) in whole blood Cytokine 1991 4 239 48 10.1016/1043-4666(92)90062-V
Hussain R Kaleem A Shahid F Dojki M Jamil B Mehmood M Dawood G Dockrell HM Cytokine profiles using whole-blood assays can discriminate between tuberculosis patients and healthy endemic controls in a BCG-vaccinated population J Immunol Methods 2002 264 95 108 12191514 10.1016/S0022-1759(02)00092-3
Kasahara K Tobe T Tomita M Mukaida N Shao-Bo S Matsushima K Yoshida.T Sugihara S Kobayashi K Selective expression of monocyte chemotactic and activating factor / monocyte chemotactic protein 1 in human blood monocytes by Mycobacterium tuberculosis J Infect Dis 1994 170 1238 1247 7963719
Faldt J Dahlgren C Ridell M Difference in neutrophil cytokine production induced by pathogenic and non-pathogenic mycobacteria Apmis 2002 110 593 600 12529011 10.1034/j.1600-0463.2002.1100901.x
Kasahara K Sato I Ogura K Takeuchi H Kobayashi K Adachi M Expression of chemokines and induction of rapid cell death in human blood neutrophils by Mycobacterium tuberculosis J Infect Dis 1998 178 127 137 9652432
Lin Y Gong J Zhang M Xue W Barnes PF Production of monocyte chemoattractant protein 1 in tuberculosis patients Infect Immun 1998 66 2319 2322 9573123
Vankayalapati R Wizel B Weis SE Klucar P Shams H Samten B Barnes PF Serum cytokine concentrations do not parallel Mycobacterium tuberculosis -induced cytokine production in patients with tuberculosis Clin Infect Dis 2003 36 24 28 12491197 10.1086/344903
Hasan Z Mahmood A Zafar S Khan AA Hussain R Leprosy patients with lepromatous disease have an up-regulated IL-8 response that is unlinked to TNFα responses Int J Lepr other Mycobact Dis 2004 72 35 44 15217317 10.1489/1544-581X(2004)072<0035:LPWLDH>2.0.CO;2
Crofton J Clinical features of tuberculosis Crofton and Douglas Respiratory Diseases 1990 London, Blackwell Scientific 395 421
Kvach JT Veras JR A fluorescent staining procedure for determining the viability of mycobacterial cells Int J Lepr other Mycobact Dis 1982 50 183 192 6180992
Hasan Z Shah BH Mahmood A Young DB Hussain R The effect of mycobacterial virulence and viability on MAP kinase signaling and TNFα production by human monocytes Tuberculosis 2003 83 299 309 12972343 10.1016/S1472-9792(03)00003-9
|
16001981
|
PMC1182368
|
CC BY
|
2021-01-04 16:03:32
|
no
|
BMC Immunol. 2005 Jul 7; 6:14
|
utf-8
|
BMC Immunol
| 2,005 |
10.1186/1471-2172-6-14
|
oa_comm
|
==== Front
BMC ImmunolBMC Immunology1471-2172BioMed Central London 1471-2172-6-151600198310.1186/1471-2172-6-15Research ArticleCloning and functional characterization of the rabbit C-C chemokine receptor 2 Lu Deshun [email protected] Xiu-juan [email protected] Robert J [email protected] Amy T [email protected] He [email protected] Eric W [email protected] Chafiq [email protected] Chandrasekar [email protected] Divisions of Atherosclerosis, Cancer, Integrative Biology and Discovery Chemistry, Lilly Research Laboratories, Lilly Corporate Center, Indianapolis, Indiana 46285, USA2005 7 7 2005 6 15 15 12 4 2005 7 7 2005 Copyright © 2005 Lu 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
CC-family chemokine receptor 2 (CCR2) is implicated in the trafficking of blood-borne monocytes to sites of inflammation and is implicated in the pathogenesis of several inflammatory diseases such as rheumatoid arthritis, multiple sclerosis and atherosclerosis. The major challenge in the development of small molecule chemokine receptor antagonists is the lack of cross-species activity to the receptor in the preclinical species. Rabbit models have been widely used to study the role of various inflammatory molecules in the development of inflammatory processes. Therefore, in this study, we report the cloning and characterization of rabbit CCR2. Data regarding the activity of the CCR2 antagonist will provide valuable tools to perform toxicology and efficacy studies in the rabbit model.
Results
Sequence alignment indicated that rabbit CCR2 shares 80 % identity to human CCR2b. Tissue distribution indicated that rabbit CCR2 is abundantly expressed in spleen and lung. Recombinant rabbit CCR2 expressed as stable transfectants in U-937 cells binds radiolabeled 125I-mouse JE (murine MCP-1) with a calculated Kd of 0.1 nM. In competition binding assays, binding of radiolabeled mouse JE to rabbit CCR2 is differentially competed by human MCP-1, -2, -3 and -4, but not by RANTES, MIP-1α or MIP-1β. U-937/rabbit CCR2 stable transfectants undergo chemotaxis in response to both human MCP-1 and mouse JE with potencies comparable to those reported for human CCR2b. Finally, TAK-779, a dual CCR2/CCR5 antagonist effectively inhibits the binding of 125I-mouse JE (IC50 = 2.3 nM) to rabbit CCR2 and effectively blocks CCR2-mediated chemotaxis.
Conclusion
In this study, we report the cloning of rabbit CCR2 and demonstrate that this receptor is a functional chemotactic receptor for MCP-1.
==== Body
Background
Chemokine receptors are seven-transmembrane G-protein coupled receptors that direct the migration of various immune cells to the sites of inflammation in addition to their pivotal role in maintaining immune cell homeostasis in various lymphoid compartments [1,2]. Chemokines (8–14 kDa molecular weight), the ligands of chemokine receptors, are broadly classified based on the positioning of the two conserved cysteines into the CC, CXC, C or CX3C family [3]. Binding of chemokines to their cognate receptors on immune cells triggers a signaling cascade primarily mediated by the Gαi family of G proteins leading to a variety of cell effector functions including chemotaxis, degranulation, and release of pro-inflammatory cytokines such as TNF α [4].
The CC family chemokine receptor 2 (CCR2) is primarily expressed in almost all circulating monocytes. CCR2 is believed to mediate extravasation of blood monocytes to the sites of inflammation [5-7] and is implicated in the pathogenesis of several inflammatory diseases such as rheumatoid arthritis, multiple sclerosis and atherosclerosis [8,9]. MCP subfamily members, namely MCP-1, -2, -3, and -4, bind to CCR2 with high affinity [7,10,11]. Several studies involving genetic deletion of CCR2 or MCP-1, and antibody neutralization studies of MCP-1 have clearly established the critical role of the CCR2/MCP-1 axis in mediating several key pathogenic events in animal models of multiple sclerosis and atherosclerosis.
The major challenge in the development of small molecule chemokine receptor antagonists is the lack of cross-species activity to the receptor in the preclinical species. For example, BX-471, a small molecule antagonist for CCR1, exhibits low nano-molar affinity to the human receptor but has sub-micromolar to high micromolar affinity to the rat and mouse CCR1 receptors, respectively [12]. Importantly, BX-471 had intermediate affinity to the rabbit CCR1 receptor [13]. This imposes significant challenges to compound testing in efficacy and toxicology studies in the preclinical models.
In recent years, rabbits have been used to study the pathogenesis of inflammatory disease model. It has been recently shown that the therapeutic potential of thiazolidinediones in rabbit model of balloon injury and re-endothelialization is associated with a decrease in MCP-1 expression [14]. In addition, the efficacy of a CXCR2 antagonist has been successfully demonstrated in a chronic antigen induced arthritis model [15]. Moreover, CCR1 antagonists have therapeutic benefit in a rabbit allograft rejection model [13]. In the present study we report the cloning and functional characterization of the rabbit CCR2. We also demonstrate that the recently reported CCR2/CCR5 dual antagonist, TAK-779, exhibits high level of potency against rabbit CCR2 in both binding and chemotaxis assays.
Results and Discussion
Cloning and expression analysis of rabbit CCR2
Full-length cDNA sequence of rabbit CCR2 was cloned from total spleen RNA of New Zealand white rabbits using a PCR based method. The deduced amino acid sequence of rabbit CCR2 encoded a protein of 369 amino acids (Fig. 1) that shared 80 % identity to the human CCR2b receptor. The sequence alignment with the mouse, rat and rhesus monkey CCR2 is also depicted in Fig. 2. Rat and mouse CCR2 contained additional 13 amino acids at the N-terminus compared to human, monkey and rabbit CCR2. In addition, one amino acid deletion in rabbit CCR2 was observed at position 18 (Fig. 2). Cysteines that were essential for the disulfide bond formation were also conserved in rabbit CCR2. Based on the SS and 2SS model [16], cysteine pairs (Cys-72/Cys-119 and Cys-31/Cys-276) in rabbit CCR2 were highly conserved compared to its counterparts from all known species. Expression levels of rabbit CCR2 were also determined in various rabbit tissues using Taqman analysis. Rabbit CCR2 was abundantly expressed in lungs and spleen compared to low levels of expression in brain, heart, liver and testis (Fig. 3).
Radioligand binding studies in rabbit CCR2 expressing cells
Stable cell lines expressing rabbit CCR2 in U-937 cells were generated to characterize the binding properties of radiolabeled human MCP-1 or mouse JE to rabbit CCR2. The binding affinity of radiolabeled human MCP-1 or mouse JE was measured using a competition binding assay of the respective cold ligands. As shown in Figs. 4, 125I-mouse JE showed high affinity with a calculated Kd of 0.95 ± 0.02 nM and Bmax of 18829 ± 1555 dpm. In contrast, 125I-human MCP-1 has less binding affinity to rabbit CCR2 with lack of saturable binding as high as 6 nM of hot ligand (data not shown). Detailed examination of the binding characteristics of radiolabeled human MCP-1 to rabbit CCR2 will be necessary to resolve this discrepancy. To address the point, future studies are necessary to understand rabbit CCR2 binding characteristics and pharmacology with 125I-rabbit- MCP-1.
Receptor pharmacology of rabbit CCR2 was studied using a variety of ligands that belonged to the MCP subfamily and irrelevant ligands such as RANTES, MIP-1α and MIP-1β. As expected, human MCP-1, -2, and -4 differentially competed the binding of radiolabeled mouse JE with Ki values of 8.4, 13.5 and 0.44 nM, respectively. In contrast, cold MCP-3 failed to compete JE binding to rabbit CCR2 when used at concentrations as high as 30 nM (Fig. 5A). In addition, irrelevant ligands that bind with high affinity to CCR1 or CCR5 were unable to compete binding of radiolabeled mouse JE to rabbit CCR2, while cold JE competed binding with a Ki value of 0.116 nM (Fig. 5B). TAK-779 is a well characterized small molecule antagonist of both CCR2 and CCR5, and has a 27 nM affinity against human CCR2b receptor [17]. We tested the activity of TAK-779 on rabbit CCR2 transfected cells using radiolabeled mouse JE in a standard competition binding assay. As shown in Fig. 5C, TAK-779 was also highly potent against rabbit CCR2 with an IC50 of 2.3 nM.
Functional characterization of Rabbit CCR2 stable cell line
In order to confirm whether rabbit CCR2 expressed in U-937 cells is functionally coupled to Gαi family of G proteins, chemotaxis assays were conducted using both parental U-937 and U-937/rabbit CCR2 stable transfectants. Cells were incubated in the top chamber in the presence of increasing concentrations of ligand in the lower chamber. Parental U-937 failed to migrate to either human MCP-1 or mouse JE. In contrast, U-937/rabbit CCR2 stable transfectants migrated effectively to either human MCP-1 or mouse JE (Fig. 6A and 6B). The maximal migration was observed at roughly 4 nM of ligand and ranged from 20–40 % total input compared to 0.5–1 % of cells migrating in the absence of any ligand. A typical bell-shaped curve was observed in the chemotaxis response with either human MCP-1 or mouse JE. As expected, both parental and rabbit CCR2 stable cell lines migrated well to the CXCR4 ligand SDF-1α (Fig. 6C). The antagonist effect of TAK-779 was also examined against rabbit CCR2 in the presence of defined concentration of either mouse JE or hMCP-1 (4 nM). Preincubation of rabbit CCR2 expressing cells with TAK-779 effectively blocked migration of either mouse JE or hMCP-1 with an IC50 of 98 nM and 4 nM, respectively (Fig. 6D and 6E). As expected, TAK-779 inhibited MCP-1 dependent migration of U-937/hCCR2b stable cells with an IC50 of 3.8 nM (Fig. 6F)
Conclusion
In summary, we have cloned the full-length rabbit CCR2 cDNA and proved that the receptor is functional by expressing rabbit CCR2 in U-937 cells. The sequence of rabbit CCR2 will provide additional insight into the molecular evolution of chemokine receptors in different species. More importantly, the data obtained on cross-species activity of the CCR2 antagonist will provide valuable tools to perform toxicology and efficacy studies in the rabbit model.
Methods
Materials
Recombinant human chemokines (MCP-1, -2, -3, -4, RANTES, MIP-1α, MIP-1β) were purchased from Pepro Tech (Rocky Hill, NJ). Mouse JE and human SDF-1α were obtained from R&D Systems (Minneapolis, MN). Radiolabeled ligands (125I-human MCP-1 and 125I-mouse JE) were purchased from Perkin Elmer (Boston, MA). RPMI 1640, bovine albumin serum (BSA) and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) were purchased from Sigma (Gaithersburgh, MD). Polymeric chain reaction (PCR) and Taqman primers were purchased from Qiagen (Valencia, CA). The CCR5/CCR2 dual antagonist TAK-779 was synthesized as described elsewhere [14].
Rabbit cDNA cloning
Total RNA was prepared from the spleen of New Zealand white rabbits by the Trizol method and the cDNA was transcribed by using SuperScript™ First-Strand Synthesis System for RT-PCR (Invitrogen, Carlson, CA). Then the cDNA was amplified by PCR by using conserved primers (primer 1: 5'-GTATTCATCTTTGGTTTTGTGGGCAACATG-3' and primer 2: 5'-CAAAGGTAACTGTCCTGGCTTTTAAAGCAA-3'). The resultant PCR fragment was sequenced and this information was used to design the rabbit CCR2 gene specific primers. Rabbit CCR2 gene specific primers were used to amplify two fragments of rabbit cDNA by 5'- and 3'- rapid amplification of cDNA ends (RACE) using SMART RACE cDNA Amplification Kit (Clonetech, Palo Alto, CA). The sequence information from 5'- and 3'- RACE products was further used to design two primers (5'-GGTTGCTGAGAAGCCTGACACGC-3', and 5'-CAGGTCTGTATTCTTCAACAAGCCCTCG-3') out of start and stop codons, and full-length rabbit CCR2 cDNA was cloned using PCR amplification. The final PCR product with corresponding size was cloned to PCRII-TAPO (Invitrogen) and sequenced.
Tissue distribution
Rabbit cDNA from brain, heart, liver, lung, spleen and testis were quantitated for expression levels of CCR2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) by Taqman method. TaqMan primers and probes were designed using the Primer Express Computer Program (Applied Biosystems, Foster City, CA). The forward and reverse primers and probe selected for rabbit CCR2 are: 5'-catgacacactgctgcatcaac-3', 5'-gagaggtagctccggaacttctc-3' and 5'-[6-FAM]-ccgtggtctacgccttcgtcgg-[TAMRA-6-FAM]-3'. The forward, reverse primers and probe selected for rabbit GAPDH are: 5'-ggatttggccgcattgg-3', 5'-caacatccactttgccagagttaa-3' and 5'-[6-FAM]-cgcctggtcaccagggctgct-[TAMRA-6-FAM]-3'. Amplification mixture contained a total volume of 10 μl with 6 μl of TaqMan universal PCR master mixture (300 nM forward primer, 300 nM reverse primer, and 900 nM probe) and 4 μl of diluted samples. Taqman amplification was conducted on a ABI PRISM 7900HT Sequence Detector System (Applied Biosystems, Foster City, CA) with the following thermal profile: 1 cycle each of 50°C for 2 minutes and 95°C for 10 minutes followed by 40 cycles each of 95°C for 15 seconds and 60°C for 1 minute. Data was analyzed using a built-in standard curve method.
Generation of retrovirus infected rabbit CCR2 stable cell line
Rabbit cDNA described above was amplified by PCR with primer 11: 5'-ggaggcagatctcgaacaggcagctggcgg-3' and primer 12: 5'-ttattggaattccaagccgacggctgtcca-3' (underlined nucleotides are the cleavage sites of Bgl II and EcoR I, respectively). The resulted PCR fragment was digested with Bgl II and EcoR I, and then subcloned into the corresponding site of pMSCVpuro retroviral expression vector (BD Biosciences, CA). Lipofectamine-mediated transfection was carried out in the AmphoPack 293 packaging cell line (BD Biosciences, CA) with a full-length cDNA for rabbit CCR2 in the retroviral expression vector pMSCVpuro, containing a puromycin resistance selection marker. Thirty six hours after transfection, the pool of viral supernatant was recovered from the packaging cells, centrifuged, and polybreen was added to a final concentration of 8 μg/ml. The viral supernatant was then added to parental U-937 cells in a tissue culture treated flask and the infection was allowed to proceed for 7 hours. At this time, additional medium was added to the packaging cells to generate more virus. After 7 hours, the second pool of viral supernatant was prepared as the first, except with 4 μg/ml polybreen. U-937 cells were resuspended in this viral supernatant and the infection was allowed to proceed overnight. Then the U-937 cells were placed into fresh medium without any selection. After 18 hours, the medium was changed to selection medium containing 0.4 μg/ml puromycin. The pool of stably transfected clones was propagated for binding and chemotaxis assays.
Equilibrium binding assays
Equilibrium binding assay was conducted in RPMI 1640 medium with 10 mM HEPES and 0.2% BSA in room temperature with constant shaking for 2 hours. Reaction mixture contained a total volume of 100 μl with various concentrations of 125I-labeled ligands (125I-human MCP-1 or 125I-mouse JE), U-937/rabbit CCR2 stable transfectants (105 cells /well for 125I-human MCP-1 and 104 cells /well for 125I-mouse JE) in the presence or absence of corresponding cold ligand. Nonspecific binding was determined by 100-fold excess of cold ligand. Reactions were stopped by separation using the glass fiber filter (Wallac Printed Filtermat A) on a Tomtec Harvester 96-2 (Hamden, CT). Filters were subsequently washed 5 times with wash buffer (10 mM HEPES, 500 mM NaCl, pH 7.4) to remove the unbound radioligand. Finally, bound radioligand was quantitated using a liquid scintillation counter (Wallac 1205 Betaplate, Perkin Elmer). Dissociation constant (Kd) and maximum binding (Bmax) were calculated by non-linear regression with Graphpad Prism 4.01 (San Diego, CA).
Competition binding assays
The assay was conducted under the identical conditions as saturation binding assay described above with the radiolabeled ligand concentration fixed (0.11 nM 125I-mouse JE) in the presence of various concentrations of unlabeled chemokines or TAK-779. Values of IC50 were calculated by fitting the competition curves with Graphpad Prism 4.01.
Chemotaxis assays
Chemotaxis assay was performed as described elsewhere [18] with a few modifications. Parental U-937 or U-937/rabbit CCR2 stable transfectants were harvested, counted and resuspended at a final cell density of 5 × 106 /ml in chemotaxis buffer (RPMI 1640 medium, containing 0.3% BSA and 1 mM HEPES). Chemokines were diluted to 1000 ng/ml stock and serially diluted (1:3) in migration buffer in a Costar 96 well plate (Cat #3790, Costar, Corning NY). Thirty microliters of each chemokine dilution were transferred to the corresponding well in the bottom chamber of a ChemoTX 96 well migration plate (5 μM pore size; Neuroprobe, Gaithersburg, MD). The ChemoTX filter unit was assembled and 50 μl of cells were added to the top chamber and incubated for 2.5 hours. Plates were then removed and cells were aspirated off the filter top. The cells that had migrated to the bottom chamber were pulsed with 5 μl of Aqueous One Solution (Promega, Madison, WI) and incubated for an additional 30–60 minutes at 37°C. Plates were shaken gently to permit uniform mixing and read at 490 nm in a Spectramax Plus reader (Molecular Dynamics, Piscataway, NJ). Cell standard curves were incorporated into the bottom of each plate with cell number ranges from 0–150,000 cells/well. The results were expressed as percent cell migration versus chemokine concentration. For compound testing, TAK-779 was diluted in DMSO at various concentrations and added to both the top and bottom chambers and chemotaxis assay was performed in the presence of 4 nM of mouse JE as described above.
Abbreviations
CCR2, CC chemokine receptor 2; MCP, monocyte chemoattactant protein; RANTES, regulated on activation-regulated chemokine; MIP, macrophage inflammatory protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RACE, rapid amplification of cDNA ends, HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; BSA, bovine serum albumin; PCR, polymerase chain reaction.
Authors' contributions
DL designed experiments, oversaw research and wrote the paper. X-JY conducted the experiments. RJE conducted the experiments. ATP conducted the experiments. HW and EWS participated in sequence alignments. CH synthesized the compound. CV conceived the study, designed experiments, oversaw research and wrote the paper.
Figures and Tables
Figure 1 Full-length rabbit CCR2 cDNA sequence. The complete cDNA sequence is shown with nucleotides numbered from the putative initiation site (underlined). The deduced amino acid sequence is on the top of nucleotide sequence. Cysteines that form the disulfide bonds are labeled in bold for Cys-112, Cys-189 (SS model) and Cys-31, Cys-276 (2SS model), respectively.
Figure 2 CCR2 peptide sequence alignment. The CCR2 peptide sequences from mouse, rat, monkey, human and rabbit are aligned and the consensus sequence is indicated in the bottom of the alignment. Transmembrane domains (TM) designated based on the human CCR2 sequence are also indicated.
Figure 3 Tissue distribution of rabbit CCR2. The mRNA levels from rabbit tissues were quantitated by Taqman technique as described under "Materials and Methods". CCR2 mRNA levels are expressed as the ratio of CCR2 and GAPDH. Error bars represent the mean ± standard deviation of quadruplicate values.
Figure 4 Saturation binding of mouse JE to rabbit CCR2. Various concentrations of 125I-mouse JE was added to U-937/rabbit CCR2 stable transfectants as described under "Materials and Methods". The nonspecific binding was determined by including 100-fold excess of unlabeled chemokine. Specific binding was calculated by subtracting nonspecific binding from total binding. The dissociation constant (Kd) was calculated using the nonlinear curve fit. Each point represents mean values of the duplicates from one of two representative assays. The scatchard plot is depicted in the insert.
Figure 5 Competition assays of various chemokines on U-937/rabbit CCR2 stable transfectants. (A) U-937/rabbit CCR2 stable cells (104 cells/well) were incubated with 125I-mouse JE (0.11 nM) for 2 hours in the presence of unlabeled chemokines (Human MCP-1, -2, -3, -4) at concentrations as high as 30 nM. (B). U-937/rabbit CCR2 stable cells (104 cells/well) were incubated with 125I-mouse JE (0.11 nM) for 2 hours in the presence of unlabeled chemokines (mouse JE, hMIP-1α, hMIP-1β and hRANTES) at concentrations as high as 30 nM. (C) Competition binding assay was performed on U-937/rabbit CCR2 stable cells (104 cells/well) that were incubated with 125I-mouse JE (0.11 nM) for 2 hours in the presence of various concentrations of TAK-779. Each point represents mean values ± standard deviations of the duplicates from one of two representative assays.
Figure 6 Chemotaxis assay in U-937/rabbit CCR2 stable transfectants. Chemotaxis assays were performed in parental U-937 or U-937/rabbit CCR2 stable transfectants using increasing concentrations of human MCP-1 (A), mouse JE (B) or human SDF-1α (C). The effect of TAK-779 on rabbit CCR2-induced chemotaxis was tested in the presence of 4 nM of mouse JE (D) or hMCP-1 (E). Effect of TAK-779 on hCCR2-induced chemotaxis in the presence of 4 nM of hMCP-1 (F). Each point represents means of the triplicate from one assay.
==== Refs
Rollins BJ Chemokines Blood 1997 90 909 928 9242519
Fernandez EJ Lolis E Structure, function, and inhibition of chemokines Annu Rev of Pharmacol Toxicol 2002 42 469 499 11807180 10.1146/annurev.pharmtox.42.091901.115838
Rossi D Zlotnik A The biology of chemokines and their receptors Annual Review of Immunology 2000 18 217 242 10837058 10.1146/annurev.immunol.18.1.217
Stievano LPE Amadori A C and CX3C chemokines: cell sources and physiopathological implications Crit Rev Immunol 2004 24 205 228 15482255 10.1615/CritRevImmunol.v24.i3.40
Wong L-M Myers SJ Tsou C-L Gosling J Arai H Charo IF Organization and differential expression of the human monocyte chemoattractant protein 1 receptor gene. Evidence for the role of the carboxyl-terminal tail in receptor trafficking J Biol Chem 1997 272 1038 1045 8995400 10.1074/jbc.272.2.1038
Yamagami S Tokuda Y Ishii K Tanaka H Endo N CDNA cloning and functional expression of a human monocyte chemoattractant protein 1 receptor Biochem and Biophys Res Commun 1994 202 1156 1162 8048929 10.1006/bbrc.1994.2049
Charo I Myers S Herman A Franci C Connolly A Coughlin S Molecular cloning and functional expression of two monocyte chemoattractant protein 1 receptors reveals alternative splicing of the carboxyl-terminal tails PNAS 1994 91 2752 2756 8146186
Gerard C Rollins BJ Chemokines and disease Nat Immunol 2001 2 108 115 11175802 10.1038/84209
Charo IF Taubman MB Chemokines in the pathogenesis of vascular disease Circ Res 2004 95 858 866 15514167 10.1161/01.RES.0000146672.10582.17
Franci C Wong L Van Damme J Proost P Charo I Monocyte chemoattractant protein-3, but not monocyte chemoattractant protein-2, is a functional ligand of the human monocyte chemoattractant protein-1 receptor J Immunol 1995 154 6511 6517 7759884
Gong X Gong W Kuhns DB Ben-Baruch A Howard OMZ Wang JM Monocyte chemotactic protein-2 (MCP-2) uses CCR1 and CCR2b as its functional receptors J Biol Chem 1997 272 11682 11685 9115216 10.1074/jbc.272.18.11682
Liang M Mallari C Rosser M Ng HP May K Monahan S Bauman JG Islam I Ghannam A Buckman B Shaw K Wei G-P Xu W Zhao Z Ho E Shen J Oanh H Subramanyam B Vergona R Taub D Dunning L Harvey S Snider RM Hesselgesser J Morrissey MM Perez HD Horuk R Identification and characterization of a potent, selective, and orally active antagonist of the CC chemokine receptor-1 J Biol Chem 2000 275 19000 19008 10748002 10.1074/jbc.M001222200
Horuk R Shurey S Ng HP May K Bauman JG Islam I Ghannam A Buckman B Wei GP Xu W CCR1-specific non-peptide antagonist: efficacy in a rabbit allograft rejection model Immunology Letters 2001 76 193 201 11306147 10.1016/S0165-2478(01)00172-9
Tanaka T Fukunaga Y Itoh H Doi K Yamashita J Chun T-H Inuoe M Masatsugu K Saito T Sawada N Sakaguchi S Arai H Nakao K Therapeutic potential of thiazolidinediones in activation of peroxisome proliferator-activated receptor γ for monocyte recruitment and endothelial regeneration Eur J Pharmacol 2005 508 255 265 15680279 10.1016/j.ejphar.2004.10.056
Podolin P Bolognese BJ Foley JJ Schmidt DB Buckley PT Widdowson KL Jin Q White JR Lee JM Goodman RB Hagen TR Kajikawa O Marshall LA Hay DWP Sarau HM A potent and selective nonpeptide anatagonist of CXCR2 inhibits acute and chronic models of arthritis in the rabbit J Immunol 2002 169 6435 6444 12444152
Shi X Liu S Xiangyu J Zhang J Liu S Liu C Structural analysis of human CCR2b and primate CCR2b by modeling and molecular dynamics simulation J Mol Model (online) 2002 8 217 222 12192431 10.1007/s00894-002-0089-6
Baba M Nishimura O Kanzaki N Okamoto M Sawada H Iizawa Y Shiraishi M Aramaki Y Okonogi K Ogawa Y Meguro K Fujino M A small-molecule, nonpeptide CCR5 antagonist with highly potent and selective anti-HIV-1 activity PNAS 1999 96 5698 5703 10318947 10.1073/pnas.96.10.5698
Chung CD Kuo F Kumer J Motani AS Lawrence CE Henderson WR JrVenkataraman C CCR8 Is not essential for the development of inflammation in a mouse model of allergic airway disease J Immunol 2003 170 581 587 12496446
|
16001983
|
PMC1182369
|
CC BY
|
2021-01-04 16:03:31
|
no
|
BMC Immunol. 2005 Jul 7; 6:15
|
utf-8
|
BMC Immunol
| 2,005 |
10.1186/1471-2172-6-15
|
oa_comm
|
==== Front
BMC Mol BiolBMC Molecular Biology1471-2199BioMed Central London 1471-2199-6-131594147010.1186/1471-2199-6-13Research ArticleNuclear distribution and chromatin association of DNA polymerase α-primase is affected by TEV protease cleavage of Cdc23 (Mcm10) in fission yeast Yang Xiaowen [email protected] Juraj [email protected] Karola [email protected] Hedi [email protected] Stephen E [email protected] Department of Zoology, University of Oxford, South Parks Road, Oxford OX13PS UK2 Current address: Structural Genomics Consortium, Nuffield Department of Clinical Medicine, Botnar Research Centre, University of Oxford, Oxford OX3 7LD, UK3 Current address: IMP, Dr. Bohr-Gasse 7, A-1030, Austria2005 7 6 2005 6 13 13 4 2 2005 7 6 2005 Copyright © 2005 Yang et al; licensee BioMed Central Ltd.2005Yang 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
Cdc23/Mcm10 is required for the initiation and elongation steps of DNA replication but its biochemical function is unclear. Here, we probe its function using a novel approach in fission yeast, involving Cdc23 cleavage by the TEV protease.
Results
Insertion of a TEV protease cleavage site into Cdc23 allows in vivo removal of the C-terminal 170 aa of the protein by TEV protease induction, resulting in an S phase arrest. This C-terminal fragment of Cdc23 is not retained in the nucleus after cleavage, showing that it lacks a nuclear localization signal and ability to bind to chromatin. Using an in situ chromatin binding procedure we have determined how the S phase chromatin association of DNA polymerase α-primase and the GINS (Sld5-Psf1-Psf2-Psf3) complex is affected by Cdc23 inactivation. The chromatin binding and sub-nuclear distribution of DNA primase catalytic subunit (Spp1) is affected by Cdc23 cleavage and also by inactivation of Cdc23 using a degron allele, implying that DNA polymerase α-primase function is dependent on Cdc23. In contrast to the effect on Spp1, the chromatin association of the Psf2 subunit of the GINS complex is not affected by Cdc23 inactivation.
Conclusion
An important function of Cdc23 in the elongation step of DNA replication may be to assist in the docking of DNA polymerase α-primase to chromatin.
==== Body
Background
Proteases play key roles in the regulation of many cellular processes, such as signal transduction, apoptosis, and the activation of chromosome disjunction in mitosis [1-4]. Artificial proteolytic regulation is possible using TEV protease, which recognizes a highly specific sequence [5] and is not deleterious when expressed in a variety of cell types [6-11]. Artificial cleavage of target proteins, engineered to contain a TEV cleavage sequence (Tcs), can thus be effected by TEV protease expression in vivo. TEV protease-mediated cleavage has been used in topological studies of protein location [12] and to study the role of regulatory proteolysis, such as in analysis of separase function [9]. Proteolytic cleavage can be used to determine how removal of specific domains of a protein affects its function [13] and, by removing a signal that targets the protein to a specific compartment, can effect a change in a protein's cellular localization [11].
We investigate here the utility of TEV-mediated cleavage of nuclear proteins in fission yeast. We demonstrate that a strain expressing nuclear-targeted TEV protease has no growth defects compared to a wild-type strain and shows efficient cleavage of nuclear proteins containing a Tcs. We use TEV protease-mediated cleavage to investigate the function of Cdc23 (homologous to Mcm10 in other organisms), which is an essential replication protein, required for the initiation and elongation steps of DNA replication ([14-17], see [18,19] for reviews). Cdc23/Mcm10 is chromatin associated, and in S. cerevisiae and vertebrate cells this association occurs during S phase [20-22], but its biochemical function is unclear. In S. cerevisiae, Mcm10 was originally implicated in pre-replicative complex formation, when Mcm2-7 proteins associate with ORC at replication origins [23]. Subsequent studies in both yeasts and Xenopus have shown that Mcm10 is required for the later step of replication initiation, as the protein is needed for chromatin association of Cdc45 [21,24,25]. In vitro, fission yeast Cdc23 enhances the ability of Hsk1 (Cdc7) protein kinase to phosphorylate Mcm2 [26], and binds to the catalytic subunit of DNA polymerase α-primase (pol α-primase), catalytically activating its DNA polymerase activity [27]. Very recently, Ricke and Bielinsky have shown that Mcm10 is required for the stability of the pol α catalytic subunit in S. cerevisiae [20].
In this paper, we use TEV protease cleavage of Cdc23 to show that removal of a 170 aa C-terminal domain of the protein, previously with no attributed functions, blocks DNA replication. Regulation of TEV protease under the thiamine-repressible nmt1 promoter in conjunction with a Tcs-containing allele of cdc23 can thus be used to inactivate conditionally the protein. We show in addition that inactivation of Cdc23 affects the chromatin binding and nuclear distribution of the DNA primase catalytic subunit (Spp1) of pol α-primase.
Results
Expression of TEV protease in S. pombe and cleavage of Mcm7-Tcs-GFP
To explore the utility of using TEV protease to construct conditional alleles in S. pombe, we constructed a strain where TEV protease, conjugated to nuclear localization signals (NLSs), is expressed under the control of the thiamine-regulatable nmt1 promoter. A strain expressing TEV protease is viable over a range of temperatures (25–37°C, Fig. 1A, B), and no effects on growth rate or DNA contents of cells as assessed by flow cytometry were detected (data not shown). To demonstrate that this TEV protease is capable of cleaving nuclear proteins, we modified a strain expressing Mcm7-GFP to contain a Tcs between the C-terminus of Mcm7 and GFP (Mcm7-GFPN760::Tcs). Mcm7 forms a complex with five other Mcm subunits (Mcm2-6), and probably provides the helicase function during the initiation and elongation steps of DNA replication (reviewed in [28,29]). Western blot analysis showed that expression of TEV protease efficiently removed the GFP tag from Mcm7 (Fig. 1D). Mcm7 is present in the nucleus throughout the cell cycle [14] (Fig. 1E, +T), in common with other Mcm subunits, but on expression of the TEV protease, the GFP fluorescence becomes delocalized over the entire cell (Fig. 1E), showing that the test substrate is efficiently cleaved on protease induction. The small size of the C-terminal GFP fragment presumably allows it to exit the nucleus by passive diffusion, consistent with earlier studies[30].
Inactivation of Cdc23 by TEV protease cleavage
A number of functions have been ascribed to the Cdc23/Mcm10 replication factor and some of these have been mapped to specific domains of the protein (Fig. 2A), such as in vitro stimulation of Hsk1 phosphorylation of Mcm2 [26] and stimulation of pol α-primase function [27]. Mcm10 contains a region encompassing a Zn-finger motif that is important for formation of high molecular weight Mcm10 homocomplexes [31] and a large segment of the protein has been reported to be needed for interaction with the Mcm2-7 complex. Most of these domains map to the N-terminal and central regions of the protein (corresponding to aa ca. 1–400 for the S. pombe protein, Fig 2A). All mapped cdc23 temperature-sensitive (ts) mutations also map to a central (232-269 aa) region. The function of the C-terminal portion of the protein, which is less well conserved than N-terminal and central domains, is unclear as a C-terminally truncated version of S. pombe Cdc23 (aa 1-423) can complement the cdc23-M36 mutant [26]. However, the C-terminal region of S. cerevisiae Mcm10 contains two NLSs, at least one of which is needed for function of the protein [32].
To determine if removal of a C-terminal domain of the Cdc23 replication factor affects the function of the protein, we inserted a Tcs just after serine 424 of a CFP-tagged version of Cdc23, 170 aa from the C terminus, thus generating Cdc23S424::Tcs, (Fig 2A). Amino acids 1-424 of Cdc23 contains most domains with ascribed functions and it was of interest to determine how cleavage after this region would affect the function of the protein. A poorly conserved region was chosen for the Tcs insertion to minimize the possibility that this modification alone would inactivate the protein. The expression level of Cdc23S424::Tcs is the same as the wild-type protein (Fig. 2C). A strain containing this insertion is viable at all temperatures in the absence of TEV protease, implying that the Tcs insertion does not affect Cdc23 function, but induction of the TEV protease prevents growth of this strain (Fig. 2B). Unexpectedly, this effect shows temperature-sensitivity, in that the strain grows at temperatures up to about 32°C, but is not viable at 37°C. The cleavage site used generates a C-terminal fragment with a stabilizing N-terminal amino acid (serine) as predicted by the N-end rule [33], and we could detect two fragments of the expected size after induction of the protease (Fig. 2C). The proportion of full-length protein is reduced at 37°C on TEV protease expression, consistent with the temperature-sensitivity of the strain. The effect is unlikely to be due to an increase in TEV protease activity, since the enzyme is more active at 32°C than 37°C [34] and we observed efficient cleavage of a different substrate (Mcm7-GFPN760::Tcs) at 32°C (Fig. 1). Instead, the Tcs in Cdc23 is presumably more accessible to the protease at higher temperatures, perhaps due to a change in conformation of Cdc23, allowing more efficient cleavage to occur.
Cdc23/Mcm10 appears to be required both for the initiation and elongation steps of DNA replication, although ts alleles do not cause an abrupt block to DNA replication in S. pombe, presumably reflecting leakiness of these mutations [14]. To determine the effect of Cdc23 cleavage on DNA replication, we arrested cells expressing Cdc23S424::Tcs in G1 by nitrogen starvation, and then released cells from the block under conditions where the TEV protease was expressed (Fig. 3A). Control cells, where the TEV protease was not induced, executed DNA replication at around 4 h after release (Fig. 3C), but in contrast expression of TEV protease prevented DNA replication (Fig. 3B), implying that the replication function of Cdc23 is inactivated by separation of the C-terminal segment of the protein. Under these conditions, expression of TEV protease in a cdc23+ background is not deleterious (Fig. 1A, data not shown). We also investigated how cleavage of the protein in a log phase culture affected DNA replication. Because de-repression of the nmt1 promoter by removal of thiamine is slow, taking 12–16 h, instead we first derepressed the promoter at 25°C, when the strain is viable. After TEV protease induction, cells were shifted to 37°C to effect Cdc23S424::Tcs cleavage, which caused cells to arrest with slightly less than 2C DNA content (Fig. 3D). Cells are not blocked at the start of S phase but appear to arrest without completing DNA replication. Cdc23 levels are higher in a log-phase cells compared to nitrogen-starved cells [25] which may make TEV protease inactivation of Cdc23 less efficient in this experiment.
Nuclear localization or chromatin binding of Cdc23 is not specified by the C-terminal region of the protein
The previous experiments establish that Cdc23's replication function is abolished when the C-terminal 170 aa domain is cleaved off. To determine whether the C-terminal region has a function in nuclear localization, as shown in S. cerevisiae [32], we followed the fate of the CFP-tagged C-terminal fragment after Cdc23S424::Tcs cleavage. Expression of TEV protease at 25°C had no effect on the nuclear localization of Cdc23 (Fig. 4), consistent with the viability of strain under these conditions and reduced cleavage of Cdc23S424::Tcs at lower temperatures (Fig. 2C, data not shown). In contrast TEV protease cleavage of Cdc23S424::Tcs at 37°C caused the CFP signal to become dispersed all over the cell, implying that the C-terminal fragment does not contain an NLS. Since the TEV protease is targeted to the nucleus, we assume that this is the main location of Cdc23S424::Tcs cleavage. Thus the lack of nuclear retention of the cleaved C-terminal fragment also implies that it is incapable of binding to chromatin, although it is possible that this region contributes to Cdc23 chromatin binding in the context of the intact protein. In contrast to the C-terminal NLSs of S. cerevisiae Mcm10 [32], Cdc23 has a bipartite NLS near the N-terminus of the protein (aa 17-33, [27]), and taken together these results suggest that the region of the protein responsible for nuclear localization has not been conserved in evolution.
Cleavage of Cdc23 affects the chromatin association and nuclear distribution of DNA polymerase α-primase
We used an in situ chromatin binding assay [35] to determine how inactivation of Cdc23 affects the chromatin association of pol α-primase and GINS (Sld5-Psf1-Psf2-Psf3) complexes, which are both required for the initiation and elongation steps of DNA replication. An earlier study has shown that the primase catalytic subunit, Spp1, is functional when C-terminally tagged with GFP and nuclear throughout the cell cycle [36]. However, cell cycle changes in chromatin association of Spp1-GFP are revealed by detergent extraction of permeabilized cells, using a low salt extraction buffer (Fig. 5A–D). Detergent washing removed Spp1 from uninucleate (G2) cells, but about 60% of binucleate (G1/S phase) cells retained Spp1 (Fig. 5A, B, D). Arrest of DNA replication with hydroxyurea (HU) caused Spp1 to become refractory to detergent extraction (Fig. 5C, E). Taken together, these results suggest that this assay can be used to monitor the G1/S-dependent binding of pol α primase to chromatin and is broadly consistent with western analysis of chromatin-enriched fractions [37,38].
Cleavage of the Cdc23S424::Tcs protein in a strain also expressing Spp1-GFP was effected by inducing the TEV protease at 25°C, and then shifting to 37°C. In a wild-type strain, this temperature shift has no effect on Spp1-GFP chromatin binding (data not shown). In contrast to the effect of HU arrest on Spp1 chromatin binding, Cdc23S424::Tcs cleavage caused a reduction in the proportion of cells retaining chromatin-associated Spp1 (from 14% to 3%; Fig. 5F). A reduction was also seen in thiamine-containing medium, probably reflecting leaky expression of TEV protease from the nmt1 promoter under repressing conditions, leading to some Cdc23 inactivation. We also observed a reduction in Spp1 chromatin binding when TEV protease was induced at 37°C in the absence of a temperature shift (data not shown). In addition to an effect on Spp1 chromatin binding, we observed a striking increase in the percentage of cells containing 1–2 bright foci of Spp1-GFP after Cdc23 inactivation (from 2% to 23%; Fig. 5F). These were generally nuclear or peri-nuclear but not necessarily co-localizing with DAPI-staining chromatin regions and may reflect formation of insoluble aggregates of Spp1 after Cdc23 inactivation. These foci were not observed during an HU-induced S phase arrest (Fig. 5C), so they appear to be a specific response to Cdc23 inactivation.
Chromatin association of pol α-primase has been reported to be dependent on Cdc45 [37,39]. Since Cdc23/Mcm10 can affect Cdc45 chromatin association [21,24,25], the effect on pol α-primase could be indirectly mediated by Cdc45. We therefore examined chromatin association of Cdc45-YFP after TEV protease inactivation of Cdc23S424::Tcs. In contrast to the effect on Spp1, Cdc45 chromatin association was not prevented by TEV protease cleavage of Cdc23 and in fact there was an increase in the percentage of cells with chromatin-bound Cdc45 (Fig. 5G). This is similar to what is observed in an HU arrest [25,40], presumably reflecting accumulation of cells blocked in S phase. This result suggests that the essential function of the C-terminal region of Cdc23 is not connected to Cdc45 chromatin association and may be more directly related to pol α-primase function.
To examine whether pol α-primase was also affected by a different conditional allele of Cdc23, we also constructed a strain expressing Spp1-GFP in the background of a degron cdc23tstd allele (this allele combines a ts degron [41] with the ts cdc23-1E2 allele, [25]). In contrast to the effect of HU, a reduction in Spp1 chromatin binding occurred on S phase arrest by Cdc23 inactivation (Fig. 6A, D). As with the experiment using Cdc23S424::Tcs cleavage, degron inactivation of Cdc23 also caused an increase in the proportion of cells showing foci of Spp1-GFP (Fig. 6D). Taken together these observations suggest that Cdc23 is needed for the normal chromatin association and subnuclear distribution of pol α-primase.
The GINS complex has been shown to be loaded onto chromatin at initiation in S. cerevisiae and Xenopus [42-45] and we examined whether this complex is also affected by disruption of Cdc23 function. We first determined whether GINS shows cell cycle dependent changes in detergent extractability. A strain where the Psf2 subunit of GINS is C-terminally tagged with YFP is viable and shows nuclear localization throughout the cell cycle (Fig. 6E). After detergent extraction, Psf2 is only retained in binucleate cells, and is resistant to detergent extraction in HU-arrested cells (Fig. 6B, E), suggesting that the protein is chromatin associated in S phase similar to results obtained with Spp1. After shifting the cdc23tstd strain to the non-permissive temperature, Psf2 becomes resistant to detergent extraction, similar to what is observed in an HU arrest (Fig. 6C, E). Similar results were obtained after TEV protease cleavage of Cdc23 (data not shown). This suggests Cdc23 is not required to maintain Psf2 chromatin association and execution of the Cdc23 function is needed for displacement of Psf2 in the course of S phase.
Discussion
In this paper we demonstrate that in vivo cleavage by TEV protease can be used to inactivate a modified version of the Cdc23 replication factor, engineered to contain a cleavage site for the protease. We infer that the C-terminal 170 aa region is essential for the DNA replication function of Cdc23. Since N- and C-terminal fragments appear to be stable after cleavage, the C-terminal region must require covalent linkage to the rest of the protein for its function. Although no clear sequence motifs have been identified in the C-terminal region, this part of the protein could have a discrete function that requires linkage to the N-terminal domain. Alternatively, interaction between the N and C-terminal regions could simply be necessary for the function of the protein in a manner which is affected by TEV protease cleavage. This function does not appear to involve nuclear localization or chromatin binding, since nuclear localization of the C-terminal fragment is lost after Cdc23 cleavage (Fig. 4).
We also show that inactivation of Cdc23, either by proteolytic cleavage or using a ts degron allele, affects the chromatin association and nuclear distribution of Spp1 primase subunit (Figs. 5, 6), suggesting that an essential role of Cdc23 is related to recruitment of pol α-primase to chromatin. Cdc45 has also been implicated in binding of pol α-primase to chromatin [37,39], but in this context it may not be relevant. Although complete inactivation of Cdc23/Mcm10 can block chromatin association of Cdc45 [21,24,25], TEV protease cleavage of Cdc23 does not prevent Cdc45 chromatin binding. One interpretation of these results is that Mcm10/Cdc23 has two functions, one related to pol α primase function, perhaps related to the C-terminal domain cleaved off by TEV protease, and the other related to Cdc45 chromatin binding.
These results are consistent with in vitro effects of S. pombe Cdc23 on pol α-primase activity [27], and evidence that S. cerevisiae Mcm10 promotes chromatin binding of the large primase subunit (Pri2) independently of effects on Cdc45, based on histone cross-linking experiments [20]. The observations we have made for Spp1-GFP may also apply to the three other subunits of pol α-primase (p49/Spp2, p70/B-subunit, p180/Pol1). However, given the requirement for S. cerevisiae Mcm10 for stability of the p180 polymerase catalytic subunit [20] and the constitutive chromatin binding of at least a fraction of S. pombe p70 [38], further work will be needed to determine the fate of other pol α primase subunits on Cdc23 inactivation.
In contrast to effects on Spp1, chromatin association of the Psf2 subunit of GINS is seen in cells arrested in S phase by using a degron allele of cdc23 (Fig. 6), similar to the result obtained by an HU arrest of DNA replication. This suggests that maintenance of GINS chromatin association is independent of Cdc23 function and that Psf2 displacement requires Cdc23 function in the elongation step of replication. At first glance, this result is inconsistent with previous work linking Cdc45 with GINS function. Specifically, chromatin association of Sld5 in Xenopus extracts is blocked by Cdc45 depletion [43], and in S. cerevisiae, the Sld3 partner of Cdc45 is essential for Psf1 origin association [42]. However, degradation of Cdc23 using the cdc23tstd allele used in our experiments is inefficient in cycling cells [25] and Cdc45 chromatin binding is not blocked under conditions of the Psf2 experiment shown in Fig. 6 (see Additional file: 1).
In this and a related study [13] we have demonstrated the utility of TEV protease for analysis of protein function in fission yeast. TEV protease can be used to determine the role of specific domains of the protein and this method can be used for topological analysis of protein function by restricting TEV protease expression to specific compartments. For this approach to be useful, the target protein must tolerate the insertion of a Tcs, which was an issue with our related study on Pol3 function [13] and the Tcs must be accessible to TEV protease in the folded protein. In this study, cleavage of Cdc23S424::Tcs was more efficient at higher temperatures, possibly reflecting a conformational change in Cdc23 which makes the Tcs more accessible to the protease. Alternatively, the interaction of Cdc23 with other factors might be changed by temperature in a manner which affects access to the Tcs.
In the approach used here, the Tcs generates a C-terminal fragment with serine as the N-terminal amino acid, which is predicted to generate a stable protein. However, TEV protease also cleaves sequences such as ENLYFQ∨R [5,46], which generates a C-terminal fragment with a destabilizing N-terminal amino acid according to the N-end rule. Such a module can be inserted at the N-terminus of the target protein, which may be more tolerated and accessible than insertion of the Tcs at internal sites. We are currently investigating whether this strategy provides a simpler approach to effect rapid inactivation of target proteins.
Conclusion
We show that TEV protease can be used in S. pombe to inactivate target proteins engineered to contain a protease cleavage site. Using this approach we show that the C-terminal domain of the Cdc23/Mcm10 replication factor, previously with no attributed functions, is essential for DNA replication. Inactivation of Cdc23/Mcm10 affects the chromatin binding and nuclear distribution of pol α-primase, suggesting it may help to attach the polymerase to the replication fork. In contrast, Cdc23/Mcm10 inactivation does not prevent maintenance of chromatin binding of the GINS elongation factor, as judged by analysis of the Psf2 subunit.
Methods
Fission yeast strains
Fission yeast strains used were constructed by standard genetic methods and are shown in Table 1. Strains were grown in rich medium (YE3S) or minimal medium (EMM). Repression of transcription from the nmt1 promoter was achieved by addition of thiamine (15 μg/ml) to EMM.
Construction of a plasmid expressing TEV protease under nmt1 regulation(pREP3X-TEV-NLS)
A fragment encoding NLS-myc9-TEV-NLS-NLS was amplified from plasmid YIp204 (GAL-NLS-myc9-TEV proteinase-NLS-NLS, [9]) using oligonucleotide primers 5'XhoITEV204 (TTTCTCGAGAGAAAAAACCCCGGATCTATGCCAA) and 3'SmaITEV204 (AAACCCGGGGCCAAGCTTGCATGCCTGCAGTTAATC) and cloned into XhoI and SmaI sites of the plasmid pREP3X [47], to generate the pREP3X-TEV-NLS plasmid.
Construction of mcm7 and cdc23 alleles containing TEV cleavage sites
To construct mcm7-GFPN760::Tcs, we first constructed a plasmid pSMUG2+mcm7 by subcloning the ApaI-SmaI fragment from pSMUC2+mcm7 [25] into pSMUG2+ [48]. The sense oligonucleotide 5'EspI-TEV-KasI-F (CCGGGTTACTTCTTTAGAAGTTGGTCGTCGCTTTTCTCCTGATGAAAATTTGTACTTCCAATCTGGCGC) and anti-sense oligonucleotide 5'KasI-TEV-EspI-R (CCGGGCGCCAGATTGGAAGTACAAATTTTCATCAGGAGAAAAGCGACGACCAACTTCTAAAGAAGTAAC) were annealed and cloned into the XmaI site of pSMUG2+mcm7. This was used to tag the endogenous mcm7+ gene by cutting with MluI and transforming into strain P138 to generate P1288. The sequence inserted after the terminal amino acid of Mcm7, before the GA linker-GFP module is RVTSLEVGRRFSPDENLYFQS (underlined sequence is the Tcs).
Two rounds of PCR were performed to construct the cdc23S424::Tcs allele. Initially, two PCR reactions were set up, using plasmid pSMUC2+Cdc23 as DNA template, using the primers: (i) 5'ApaIMCM10 (AAAGGGCCCGAACCA AGAAAGGAAGAGGAGCGATAA) and 3'MCM10tevAclI1010 (CCCGTAGAG AGGTTATTCGACTGGAAGTACAACGTTTCGGAAGCGCAACATCGC); (ii) 5' MCM10tevAclI1010 (GCGATGTTGCGCTTCCGAAACGTTGTACTTCCAGTC GAATAACCTCTCTACGGG) and 3' ATGCYfp-screen (ACAGCTCCTCGCCCTTGCTCACCAT). Products from the two PCR reactions were mixed and boiled and used for a secondary PCR reaction using primers 5'ApaIMCM10 and 3' ATGCYfp-screen. The PCR product was cloned into ApaI and HindIII sites of plasmid pSMUC2-Cdc23 [25]. This was used to tag endogenous cdc23+ gene by cutting with SpeI and transforming into strain P138 to generate P1322. All constructs were verified by sequencing.
Construction of strains expressing fluorescently-tagged Spp1 and Psf2
The catalytic subunit of DNA primase (Spp1) was tagged with GFP using the oligos p48F1 (TGTACTTTAAATCGTTCAGTAGCCAGCTTTTTAAAGAAACAGTAGGAAATAAAAGAAAACATGAGAATTTGGAATTTCGGATCCCCGGGT) and p48R1 (GATTTAACGGAAAACAATATGCTCCACGTAAATAGAAGTCAGAGCTTCCATCCTTTCATGAGTAAATGTATGCCCAAATGAATTCGAGCTCGTTTAAACT) using pFA6a-GFP-kanMX6 as template [49] to generate a PCR product that was used to tag the endogenous Spp1 gene by transforming strain P138. Psf2 was tagged with YFP using the oligos 5'ApaI-Psf2 (CTTGGGCCCATACGAGATGATGAATTGGAAAACG) and 3'XhoIPsf2 (TTTCTCGAGTTCTTCTTGGGAAACTTGAACAAT) which was cloned into pSMUY2+ [25]; cleavage of this plasmid with EcoRV was used to tag the endogenous psf2+ gene in strain P138 with YFP, generating strain P1411. All constructs were verified by sequencing. Strains containing Spp1-GFP or Psf2-YFP showed tagged proteins with the expected molecular weights when analyzed by western blotting, using anti-GFP antibody (data not shown).
Chromatin binding assay
Chromatin binding assay and image analysis was carried out as previously described for Psf2 [35,50]. For analysis of Spp1, a low-salt extraction buffer was used with the following composition: 20 mM Pipes-KOH pH 6.8, 0.4 M sorbitol, 10 mM KAc, 0.5 mM spermidine, 0.15 mM spermine, 1 mM EDTA. Extraction with higher salt buffers results in loss of Spp1 from S phase cells (data not shown). Data shown are the averages of at least two experiments. Nuclear retention of Spp1-GFP and Psf2-YFP after detergent extraction was abolished after digestion of chromatin with micrococcal nuclease (data not shown). Chromatin binding assays were performed at least twice and error bars show the statistical range. At least 100 cell were counted for each data point.
Flow cytometry
Flow cytometry was carried out using sytox green (1 μm) as previously described [25].
Western blotting
Protein extracts for Western blotting were made by trichloroacetic acid extraction as previously described [51]. For western blot analysis, antibodies against full-length Cdc23 (from S. Aves [25]), GFP (3E1 monoclonal), TEV protease (from M. Ehrmann), and α-tubulin (Sigma T5168) were used. Detection was performed using the chemiluminescence procedure.
Authors' contributions
XY carried out most of the experiments. JG devised conditions for the Spp1 chromatin binding assay. KL constructed strains. HY contributed to the experiments shown in Figs. 5, 6. SK designed the experiments and wrote the paper.
Supplementary Material
Additional File 1
Chromatin binding analysis of Cdc45-YFP in a degron cdc23td strain, in asynchronous culture. A. Scheme of experiment shown in (B). Cultures of P1083 (cdc45-YFP) and P1100 (cdc45-YFP cdc23tstd) were grown at 25°C to log phase. HU (12 mM) was added and the cultures were split; half the cells were shifted to 37°C. Chromatin binding analysis was carried after on the -HU 25°C cells, and on cells from the +HU cultures after 3 h. B. Analysis of cells with nuclear Cdc45 either with or without detergent extraction. The control strain shows an increase in chromatin-associated Cdc45 (i.e. nuclear Cdc45 after detergent extraction) during the HU arrest either at 25°C or 37°C as previously reported [25], as displacement of Cdc45 from chromatin at the end of S phase is prevented by the S phase arrest. In the cdc23tstd strain, a similar result is shown, indicating that the degron allele does not affect Cdc45-YFP chromatin association. This may reflect inefficient inactivation of Cdc23 under these conditions. In contrast to this result, Cdc45 chromatin association is affected when cdc23 is inactivated following G1 arrest by nitrogen starvation [25].
Click here for file
Acknowledgements
This work was supported by grants from the BBSRC (43/G15095) and Cancer Research-UK. We are grateful to Michael Ehrmann and Steve Aves for generous gifts of antibody and Frank Uhlmann for plasmids. We are grateful to Lynne Larkman for technical support and to Shao-Win Wang for comments.
Figures and Tables
Figure 1 Effects of TEV protease expression on wild-type and mcm7-GFPN760::Tcs S. pombe strains. A. Western analysis comparing levels of expression of TEV protease in strain P1292 in – thiamine medium (lane 2) and + thiamine medium (lane 3). Lane 1 is protein extract from a strain lacking the pREP3X-TEV-NLS plasmid (P138). B. Expression of TEV protease does not affect the viability of S. pombe. Serial dilutions of strain P1292 were spotted onto plus (+T) and minus (-T) medium and incubated at the temperatures shown. C. Design of Mcm7-GFPN760::Tcs. The Mcm7 gene was modified so that a TEV cleavage site, fused to GFP is expressed after the last codon of the Mcm7 reading frame. Cleavage by TEV protease liberates GFP with an N- terminal serine, which is stable according to the N-end rule. D. Western analysis of the mcm7-GFPN760::Tcs strains, using anti-GFP antibody. Lane 1: untagged control strain (P138); lane 2: mcm7-GFPN760::Tcs strain (P1288); lane 3: mcm7-GFPN760::Tcs containing pREP3X-TEV-NLS, +thiamine (P1171); lane 4: mcm7-GFPN760::Tcs strain containing pREP3X-TEV-NLS, -thiamine (P1171). All cultures were grown at 32°C. D. Fluorescence microscopy analysis of cells expressing Mcm7-GFPN760::Tcs in the absence (+thiamine) or presence (-thiamine) of TEV protease. Cells were fixed using methanol and acetone.
Figure 2 Construction and characterization of a strain expressing Cdc23-CFP with an internal TEV cleavage site (Cdc23S424::Tcs). A. Regions of Cdc23 with attributed functions (see text for details), and location of Tcs in Cdc23S424::Tcs. The grey bar superimposed on the black Zn-finger region corresponds to the domain of S. cerevisiae Mcm10 required for homocomplex assembly [31]. "ts" indicates the location of cdc23-1E2 and -M36 ts mutations. B. Expression of TEV protease causes loss of viability of a cdc23S424::Tcs strain (P1323) at 37°C. Serial dilutions of the P1323 strain were spotted onto minus or plus thiamine media and incubated at the temperatures shown. C. Western analysis of Cdc23 levels in cdc23S424::Tcs strains in presence or absence of pTEV (pREP3X-TEV-NLS) plasmid, grown in the presence or absence of thiamine. The blot was probed with antibodies against Cdc23, and α-tubulin is shown as a loading control. Cells were first grown at 25°C in the presence of thiamine to repress TEV protease expression, then washed and grown either in the presence or absence of thiamine for 24 h at the indicated temperatures, before extracting protein for western analysis. Cdc23S424::Tcs indicates the position of full length protein, N & C Cdc23 indicate the fragments (49 kDa (N) and 46 kDa (C)) after cleavage by TEV protease. Comparison of Cdc23 levels in strains 3 and 4 (wt) shows that the modification of Cdc23 does not affect protein levels under conditions where TEV protease is absent. Strains used were: (1) P1323-1; (2) P1323-2; (3); P1322 (4) P138.
Figure 3 Flow cytometry analysis of cdc23S424::Tcs strains following induction of TEV protease. A. Scheme for experiment shown in (B) & (C). B-C. Flow cytometric analysis of G1-arrested cdc23S424::Tcs strain (P1376), released from a G1 block in the absence (B) or presence (C) of thiamine, showing arrest of DNA replication when TEV protease is expressed. D. Flow cytometric analysis of cdc23S424::Tcs strain (P1376) expressing TEV protease (i.e. – thiamine). The strain was grown to log phase at 25°C in – thiamine medium and shifted to 37°C at t = 0. The top panel superimposes the t = 0 (grey) and t = 3 h (black) histograms.
Figure 4 The C-terminal fragment of Cdc23S424::Tcs is not localized to the nucleus after cleavage by TEV protease. A. cdc23S424::Tcs strain (P1323) was grown in plus (-TEV) or minus thiamine (+TEV) media at 25°C for 24 h, then shifted to 37°C for 4 h before fixation with methanol and acetone. Panels show Cdc23-CFP fluorescence and DAPI staining superimposed on phase images. Bar = 10 μm. B. Quantitation of experiment shown in A. 100 cells were counted for each time point.
Figure 5 Analysis of Spp1-GFP nuclear distribution after TEV protease cleavage of Cdc23S424::Tcs. A. Localization of Spp1-GFP after direct fixation (methanol/acetone) of an asynchronous culture (strain P903). B. Localization of Spp1-GFP (strain P903) after detergent extraction of cells; >50% binuclear cells show nuclear Spp1 (arrow) C. As in (B) except that cells were grown in 12 mM HU for 2 h before analysis. Bar = 10 μm. D. Quantitation of data in (B); 100 cells were counted for each data point and 100% represents the total number of binucleate or uninucleate cells counted. E. Quantitation of data in (C); data for other time points (with and without detergent extraction) are also shown. F. Chromatin binding analysis of Spp1-GFP after inactivation of Cdc23S424::Tcs by TEV protease cleavage. Strain P1460 was grown either in the absence or presence of thiamine for 24 h at 25°C, then shifted to 37°C for 2, 4 h before analysis using the chromatin binding assay. Grey bars show the percentage of total cells showing Spp1-GFP fluorescence coincident with chromatin (DAPI-staining region) after detergent extraction. Black bars show the percentage of cells showing foci of Spp1-GFP fluorescence in the absence of general nuclear fluorescence (example shown in right-hand panels). At least 300 cells were counted for each data point. G. Chromatin association of Cdc45-YFP is not prevented by cleavage of Cdc23S424::Tcs by TEV protease. Strain P1409 was grown either in the absence or presence of thiamine for 24 h at 25°C, then shifted to 37°C for 2, 4 h before analysis using the chromatin binding assay. Data points show the percentage of total cells showing nuclear retention of Cdc45-YFP after detergent extraction. At all time points, >95% of cells show nuclear Cdc45-YFP before detergent extraction.
Figure 6 Localization of Spp1-GFP and Psf2-YFP after inactivation of Cdc23 using a degron mutant. A. Strain P1205 containing Spp1-GFP in the background of a degron cdc23tstd allele [25] was initially grown at 25°C in the absence of thiamine. 3 h before the temperature shift to 37°C, thiamine was added to repress cdc23 transcription, and 2 h after the shift, cells were analyzed using the chromatin binding assay (focus of Spp1-GFP shown (arrow), bar = 10 μm.) Quantitative analysis is shown in (D). B. Analysis of wild-type Psf2-YFP strain (P1411) in log phase after detergent extraction, showing retention of Psf2 in >50% binucleate (G1 or S phase) cells but not uninucleate (G2) cells. Quantitative analysis is shown in (E). C. Strain P1453, containing Psf2-YFP in the background of a degron cdc23tstd mutant, was treated as in (A) and analyzed by the chromatin binding assay. Cells shown had been shifted to 37°C for 2 h to inactivate Cdc23 before processing using the chromatin binding assay. Quantitative analysis is shown in (E). D. Analysis of data in experiments shown in (A); at least 200 cells were counted for each data point in each experiment. White bars show the percentage of total cells showing Spp1-GFP fluorescence coincident with chromatin (DAPI-staining region) after detergent extraction. Black bars show the percentage of cells showing foci of Spp1-GFP fluorescence in the absence of general nuclear fluorescence. 100% represents total number of cells counted for each data point. E. Analysis of Psf2-YFP in nuclei of wild-type cells, and in a cdc23tstd degron mutant at permissive and non-permissive temperatures, either fixed directly (-detergent), or after detergent extraction (+detergent). For the experiment with the cdc23tstd mutant, the percentages of all cells (i.e. binucleate and uninucleate) are shown.
Table 1 S. pombe strains used
reference
P138 ade6-M210 leu1-32 ura4-D18 h- this work
P903 spp1-GFP::kanr ade6-M120 leu1-32 ura4-D18 h- [25]
P1083 cdc45(sna41)-YFP::ura4+ade6-M210 leu1-32 ura4-D18 h- [25]
P1100 cdc45(sna41)-YFP::ura4+nmt1*cdc23-1E2-td::kanr h- this work
P1171 mcm7-GFPN760::Tcs ::ura4+ade6-M210 leu1-32 ura4-D18 h- [pREP3X-TEV-NLS(LEU2)] this work
P1205 spp1-GFP::kanr nmt1*-cdc23-1E2-td::kanr leu1-32 this work
P1288 mcm7-GFPN760::Tcs ::ura4+ade6-M210 leu1-32 ura4-D18 h- this work
P1292 ade6-M210 leu1-32 ura4-D18 h- [pREP3X-TEV-NLS(LEU2)] this work
P1322 cdc23-CFPS424::Tcs::ura4+ade6-M210 leu1-32 ura4-D18 h- this work
P1323-1
P1323-2 cdc23-CFPS424::Tcs::ura4+ade6-M210 leu1-32 ura4-D18 h- [pREP3X-TEV-NLS(LEU2)] this work
P1376 cdc23-CFPS424::Tcs::ura4+leu1-32 ura4-D18 h- [pREP3X-TEV-NLS(LEU2)] this work
P1409 cdc45(sna41)-YFP::ura4+cdc23-CFPS424::Tcs::ura4+leu1-32 ura4-D18 ade6 [pREP3X-TEV-NLS(LEU2)] this work
P1411 psf2-YFP::kanrade6-M210 leu1-32 ura4-D18 h- this work
P1453 psf2-YFP::kanr nmt1*-cdc23-1E2-td::kanr this work
P1460 spp1-GFP::kanr cdc23-CFPS424::Tcs::ura4+leu1-32 ura4-D18 ade6 [pREP3X-TEV-NLS(LEU2)] this work
*medium-strength nmt1 promoter
==== Refs
Sullivan M Hornig NC Porstmann T Uhlmann F Studies on substrate recognition by the budding yeast separase J Biol Chem 2004 279 1191 1196 14585836 10.1074/jbc.M309761200
Hollenberg MD Protease-mediated signally: new paradigms for cell regulation and drug development. Trends Pharmacol Sci 1996 17 3 6 8789351 10.1016/0165-6147(96)81562-8
Porter AG Ng P Janicke RU Death substrates come alive Bioessays 1997 19 501 507 9204767 10.1002/bies.950190609
Ehrmann M Clausen T Proteolysis as a regulatory mechanism Annu Rev Genet 2004 38 709 724 15568990 10.1146/annurev.genet.38.072902.093416
Dougherty WG Cary SM Parks TD Molecular genetic analysis of a plant virus polyprotein cleavage site: a model Virology 1989 171 356 364 2669323 10.1016/0042-6822(89)90603-X
Mondigler M Ehrmann M Site-specific proteolysis of the Escherichia coli SecA protein in vivo J Bacteriol 1996 178 2986 2988 8631693
Ehrmann M Bolek P Mondigler M Boyd D Lange R TnTIN and TnTAP: mini-transposons for site-specific proteolysis in vivo Proc Natl Acad Sci U S A 1997 94 13111 13115 9371808 10.1073/pnas.94.24.13111
Smith TA Kohorn BD Direct selection for sequences encoding proteases of known specificity Proc Natl Acad Sci U S A 1991 88 5159 5162 2052595
Uhlmann F Wernic D Poupart MA Koonin EV Nasmyth K Cleavage of cohesin by the CD clan protease separin triggers anaphase in yeast Cell 2000 103 375 386 11081625 10.1016/S0092-8674(00)00130-6
Faber KN Kram AM Ehrmann M Veenhuis M A novel method to determine the topology of peroxisomal membrane proteins in vivo using the tobacco etch virus protease J Biol Chem 2001 276 36501 36507 11443138 10.1074/jbc.M105828200
Urabe M Kume A Takahashi T Serizawa N Tobita K Ozawa K A switching system regulating subcellular localization of nuclear proteins using a viral protease Biochem Biophys Res Commun 1999 266 92 96 10581171 10.1006/bbrc.1999.1788
Gouffi K Gerard F Santini CL Wu LF Dual topology of the Escherichia coli TatA protein J Biol Chem 2004 279 11608 11615 14701831 10.1074/jbc.M313187200
Garcia JS Ciufo L Yang X Kearsey SE MacNeill S The C-terminal zinc finger of the catalytic subunit of DNA polymerase delta is responsible for direct interaction with the B-subunit Nucleic Acids Res 2004 32 3005 3016 15173383 10.1093/nar/gkh709
Liang DT Forsburg SL Characterization of Schizosaccharomyces pombe mcm7+ and cdc23+ (MCM10) and interactions with replication checkpoints Genetics 2001 159 471 486 11606526
Nasmyth KA Nurse P Cell division mutants altered in DNA replication and mitosis in the fission yeast Schizosaccharomyces pombe Mol Gen Genet 1981 182 119 124 6943408 10.1007/BF00422777
Hart EA Bryant JA Moore K Aves SJ Fission yeast Cdc23 interactions with DNA replication initiation proteins Curr-Genet 2002 41 342 348 12185500 10.1007/s00294-002-0316-9
Aves SJ Tongue N Foster AJ Hart EA The essential schizosaccharomyces pombe cdc23 DNA replication gene shares structural and functional homology with the Saccharomyces cerevisiae DNA43 (MCM10) gene Curr-Genet 1998 34 164 171 9745018 10.1007/s002940050382
Kearsey SE Cotterill S Enigmatic variations: divergent modes of regulating eukaryotic DNA replication Molecular Cell 2003 12 1067 1075 14636567 10.1016/S1097-2765(03)00441-6
Bell SP Dutta A DNA replication in eukaryotic cells Ann Rev Biochem 2002 71 333 374 12045100 10.1146/annurev.biochem.71.110601.135425
Ricke RM Bielinsky AK Mcm10 regulates the stability and chromatin association of DNA polymerase-alpha Mol Cell 2004 16 173 185 15494305 10.1016/j.molcel.2004.09.017
Wohlschlegel JA Dhar SK Prokhorova TA Dutta A Walter JC Xenopus Mcm10 binds to origins of DNA replication after Mcm2-7 and stimulates origin binding of Cdc45 Mol Cell 2002 9 233 240 11864598 10.1016/S1097-2765(02)00456-2
Izumi M Yanagi K Mizuno T Yokoi M Kawasaki Y Moon KY Hurwitz J Yatagai F Hanaoka F The human homolog of Saccharomyces cerevisiae Mcm10 interacts with replication factors and dissociates from nuclease-resistant nuclear structures in G(2) phase Nucleic-Acids-Res 2000 28 4769 4777 11095689 10.1093/nar/28.23.4769
Homesley L Lei M Kawasaki Y Sawyer S Christensen T Tye BK Mcm10 and the MCM2-7 complex interact to initiate DNA synthesis and to release replication factors from origins Genes Dev 2000 14 913 926 10783164
Sawyer SL Cheng IH Chai W Tye BK Mcm10 and Cdc45 cooperate in origin activation in Saccharomyces cerevisiae J Mol Biol 2004 340 195 202 15201046 10.1016/j.jmb.2004.04.066
Gregan J Lindner K Brimage L Franklin R Namdar M Hart EA Aves SJ Kearsey SE Fission yeast Cdc23/Mcm10 functions after pre-replicative complex formation to promote Cdc45 chromatin binding Molecular Biology of the Cell 2003 14 3876 3887 12972571 10.1091/mbc.E03-02-0090
Lee JK Seo YS Hurwitz J The Cdc23 (Mcm10) protein is required for the phosphorylation of minichromosome maintenance complex by the Dfp1-Hsk1 kinase Proc Natl Acad Sci U S A 2003 100 2334 2339 12604790 10.1073/pnas.0237384100
Fien K Cho YS Lee JK Raychaudhuri S Tappin I Hurwitz J Primer utilization by DNA polymerase alpha-primase is influenced by its interaction with Mcm10p J Biol Chem 2004 279 16144 16153 14766746 10.1074/jbc.M400142200
Labib K Diffley JFX Is the MCM2-7 complex the eukaryotic DNA replication fork helicase? Curr Opin Gen Dev 2001 11 64 70 10.1016/S0959-437X(00)00158-1
Kearsey SE Labib K MCM proteins: evolution, properties and role in eukaryotic DNA replication BBA 1998 1398 113 136 9689912
Shulga N Mosammaparast N Wozniak R Goldfarb DS Yeast Nucleoporins Involved in Passive Nuclear Envelope Permeability 2000 1 1027–1038
Cook CR Kung G Peterson FC Volkman BF Lei M A novel zinc finger is required for Mcm10 homocomplex assembly J Biol Chem 2003 278 36051 36058 12844493 10.1074/jbc.M306049200
Burich R Lei M Two bipartite NLSs mediate constitutive nuclear localization of Mcm10 Curr Genet 2003 44 195 201 13680157 10.1007/s00294-003-0443-y
Varshavsky A The N-end rule pathway of protein degradation Genes Cells 1997 2 13 28 9112437 10.1046/j.1365-2443.1997.1020301.x
Nallamsetty S Kapust RB Tozser J Cherry S Tropea JE Copeland TD Waugh DS Efficient site-specific processing of fusion proteins by tobacco vein mottling virus protease in vivo and in vitro Protein Expr Purif 2004 38 108 115 15477088 10.1016/j.pep.2004.08.016
Kearsey SE Montgomery S Labib K Lindner K Chromatin binding of the fission yeast replication factor mcm4 occurs during anaphase and requires ORC and cdc18 EMBO J 2000 19 1681 1690 10747035 10.1093/emboj/19.7.1681
Griffiths DJ Liu VF Nurse P Wang TS Role of fission yeast primase catalytic subunit in the replication checkpoint Mol Biol Cell 2001 12 115 128 11160827
Uchiyama M Griffiths D Arai K Masai H Essential role of Sna41/Cdc45 in loading of DNA polymerase alpha onto minichromosome maintenance proteins in fission yeast J Biol Chem 2001 276 26189 26196 11344166 10.1074/jbc.M100007200
Uchiyama M Wang TS The B-subunit of DNA polymerase alpha-primase associates with the origin recognition complex for initiation of DNA replication Mol Cell Biol 2004 24 7419 7434 15314153 10.1128/MCB.24.17.7419-7434.2004
Mimura S Takisawa H Xenopus Cdc45-dependent loading of DNA polymerase alpha onto chromatin under the control of S-phase Cdk EMBO J 1998 17 5699 5707 9755170 10.1093/emboj/17.19.5699
Dolan WP Sherman DA Forsburg SL Schizosaccharomyces pombe replication protein Cdc45/San41 requires Hsk1/Cdc7 and Rad4/Cut5 for chromatin binding Chromosoma 2004 113 145 156 15338237 10.1007/s00412-004-0302-8
Dohman RJ Wu P Varshavsky A Heat-inducible degron: a method for constructing temperature-sensitive mutants Science 1994 263 1273 1276 8122109
Takayama Y Kamimura Y Okawa M Muramatsu S Sugino A Araki H GINS, a novel multiprotein complex required for chromosomal DNA replication in budding yeast Genes-Dev 2003 17 1153 1165 12730134 10.1101/gad.1065903
Kubota Y Takase Y Komori Y Hashimoto Y Arata T Kamimura Y Araki H Takisawa H A novel ring-like complex of Xenopus proteins essential for the initiation of DNA replication Genes-Dev 2003 17 1141 1152 12730133 10.1101/gad.1070003
Kanemaki M Sanchez Diaz A Gambus A Labib K Functional proteomic identification of DNA replication proteins by induced proteolysis in vivo Nature 2003 423 720 724 12768207 10.1038/nature01692
Gomez EB Angeles VT Forsburg SL A screen for S. pombe mutants defective in re-replication identifies new alleles of rad4+, cut9+ and psf2+. Genetics 2005
Dougherty WG Carrington JC Cary SM Parks TD Biochemical and mutational analysis of a plant virus polyprotein cleavage site EMBO J 1988 7 1281 1287 3409865
Forsburg SL Comparison of Schizosaccharomyces pombe expression systems Nucleic Acids Res 1993 21 2955 22956 8332516
Lindner K Gregan J Montgomery S Kearsey S Essential role of MCM proteins in pre-meiotic DNA replication Mol Biol Cell 2002 13 435 444 11854402 10.1091/mbc.01-11-0537
Bahler J Wu JQ Longtine MS Shah NG McKenzie A Steever AB Wach A Philippsen P Pringle JR Heterologous modules for efficient and versatile PCR-based gene targeting in Schizosaccharomyces pombe Yeast 1998 14 943 951 9717240 10.1002/(SICI)1097-0061(199807)14:10<943::AID-YEA292>3.3.CO;2-P
Kearsey SE Brimage L Namdar M Ralph E Yang X In situ assay for analyzing the chromatin binding of proteins in fission yeast Methods Mol Biol 2004 296 181 188 15576932
Grallert B Kearsey SE Lenhard M Carlson CR Nurse P Boye E Labib K A fission yeast general translation factor reveals links between protein synthesis and cell cycle controls J Cell Sci 2000 113 1447 1458 10725227
|
15941470
|
PMC1182370
|
CC BY
|
2021-01-04 16:22:25
|
no
|
BMC Mol Biol. 2005 Jun 7; 6:13
|
utf-8
|
BMC Mol Biol
| 2,005 |
10.1186/1471-2199-6-13
|
oa_comm
|
==== Front
BMC Mol BiolBMC Molecular Biology1471-2199BioMed Central London 1471-2199-6-151596323410.1186/1471-2199-6-15Research ArticleZinc-finger domains of the transcriptional repressor KLF15 bind multiple sites in rhodopsin and IRBP promoters including the CRS-1 and G-rich repressor elements Otteson Deborah C [email protected] Hong [email protected] Yuhui [email protected] Donald J [email protected] Guerrieri Center for Genetic Engineering and Molecular Ophthalmology at the Wilmer Eye Institute, Johns Hopkins University School of Medicine; 600 North Wolfe Street; Baltimore, MD 21287 USA2 Department of Ophthalmology, Johns Hopkins University School of Medicine; 600 North Wolfe Street; Baltimore, MD 21287 USA3 Departments of Neuroscience, and Molecular Biology and Genetics; Johns Hopkins University School of Medicine; 600 North Wolfe Street; Baltimore, MD 21287 USA4 College of Optometry, University of Houston; Houston, TX 77204 USA5 Department of Genetics, Stanford University School of Medicine; Stanford, CA 94305 USA2005 17 6 2005 6 15 15 26 4 2005 17 6 2005 Copyright © 2005 Otteson et al; licensee BioMed Central Ltd.2005Otteson 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 retina, many of the genes that encode components of the visual transduction cascade and retinoid recycling are exclusively expressed in photoreceptor cells and show highly stereotyped temporal and spatial expression patterns. Multiple transcriptional activators of photoreceptor-specific genes have been identified, but little is known about negative regulation of gene expression in the retina. We recently identified KLF15, a member of the Sp/Krüppel-like Factor family of zinc-finger containing transcription factors, as an in vitro repressor of the promoters of the photoreceptor-specific genes rhodopsin and IRBP/Rbp3. To gain further insight into the mechanism of KLF15-mediated regulation of gene expression, we have characterized the binding characteristics and specificity of KLF15's DNA binding domains and defined the KLF15 binding sites in the rhodopsin and IRBP promoters.
Results
In EMSA and DNAseI footprinting assays, a KLF15-GST fusion protein containing the C-terminal zinc-finger domains (123 amino acids) showed zinc-dependent and sequence-specific binding to a 9 bp consensus sequence containing a core CG/TCCCC. Both the bovine rhodopsin and IRBP promoters contained multiple KLF15 binding sites that included the previously identified CRS-1 and G-rich repressor elements. KLF15 binding sites were highly conserved between the bovine, human, chimp and dog rhodopsin promoters, but less conserved in rodents. KLF15 reduced luciferase expression by bRho130-luc (containing 4 KLF15 sites) and repressed promoter activation by CRX (cone rod homeobox) and/or NRL (neural retina leucine zipper), although the magnitude of the reduction was smaller than previously reported for a longer bRho225-luc (containing 6 KFL15 sites).
Conclusion
KLF15 binds to multiple 9 bp consensus sites in the Rhodospin and IRBP promoters including the CRS-1 and G-rich repressor elements. Based on the known expression pattern of KLF15 in non-photoreceptor cells, we hypothesize an in vivo role for KLF15 in repressing photoreceptor-specific gene expression in the inner retina.
==== Body
Background
Photoreceptors are the highly specialized sensory receptors of the retina and express a unique array of genes that enable them to convert light energy into a neural signal. Many of these genes, including those encoding components of the phototransduction cascade [e.g. rhodopsin (Rho), transducin, arrestin, α– and β-phosphodiesterase (PDE)] and those involved in retinoid recycling [e.g. interphotoreceptor retinoid binding protein (IRBP/RBP3)], are expressed only in photoreceptor cells in the retina and in a subset of cells in the pineal gland [1-5]. In addition, these genes have highly stereotyped temporal and spatial patterns of expression during retinal development [6-8] that are transcriptionally regulated [7,9,10]. We are interested in understanding the transcriptional networks that regulate photoreceptor-specific gene expression, not only as a model for cell-specific gene regulation, but to gain insight into the mechanisms that regulate photoreceptor differentiation and neuro-degenerative retinal disease.
The proximal promoter regions of both Rho (-225 to +70 base pairs relative to the transcription start site) [11-13] and IRBP (-123 to +18 bp) [6,14] are sufficient to drive photoreceptor-specific expression of reporter genes in transgenic mice. Using DNAse I protection assays, binding sites for both retina-enriched and more widely expressed proteins have been identified on these promoters [15-22] and include both positive and negative regulatory elements [17-19,23-29]. Among the positive regulatory elements, Ret-1, BAT-1 and PCE-1 each contain a core ATTA sequence that can be bound by homeodomain containing transcription factors including CRX [30], RX/RAX [31], ERX [32], QRX [33] and OTX2 [18,34]. Binding sites for the basic-leucine zipper transcription factor NRL have been identified in the promoters of rhodopsin [35,36] as well as other photoreceptor-specific genes including arrestin [26] and the rod-specific α- and β-PDE [37,38]. Combinatorial effects by CRX and NRL result in synergistic increases in Rho promoter activation in vitro that likely contribute to the high levels of expression in rod photoreceptors in vivo [29,30,39] Both Crx and Nrl play key roles in photoreceptor development and survival: in Crx knock-out mice, rod photoreceptors fail to elaborate outer segments and degenerate and the mice lack detectible visual function [40]; in Nrl knock-out mice, rod photoreceptors apparently fail to differentiate (although there is an increase in cone-like cells) and mice have severely reduced visual function [41]. In addition, mutations in both CRX and NRL have been identified in patients with a variety of inherited retinal degenerative diseases including autosomal dominant and recessive retinitis pigmentosa, Leber congenital amaurosis and cone-rod dystrophy [42-45].
Although transcriptional repression plays an important role in development and cell fate determination in the central nervous system and inappropriate de-repression of gene expression can contribute to disease [46,47], relatively little is known about negative regulation of photoreceptor-specific gene expression. Negative regulatory elements, including the CRS-1 site in the rhodopsin promoter [21] and the G-rich repressor in the IRBP promoter [20] are bound by nuclear proteins present in both retinal and non-retinal tissues, although the identities of the specific transcription factors have not been established. The zinc-finger transcription factor MOK2 binds to sites in the promoter and intron 1 of IRBP and represses transactivation of reporter constructs containing these elements in vitro [48]. Other proteins interact with CRX (e.g. ATXN-7 [49] and BAF [50]) or NRL (e.g. FIZ-1 [51]) and can reduce their ability to activate transcription, but these appear to repress via protein/protein interactions rather than binding directly to the promoter. We recently identified KLF15, a member of the Sp/KLF family of zinc-finger containing transcription factors, as a repressor of both the rhodopsin and IRBP promoters in vitro [52].
KLF15 is expressed in both retinal and non-retinal tissues and, like other members of the Sp/KLF family, is characterized by the presence of three Krüppel-type zinc-finger domains at the C-terminus [52-54]. Members of this gene family have been characterized as regulators of both tissue-specific and ubiquitous genes and can function as either transcriptional activators, repressors or both depending on promoter context [55,56] KLF15 appears to be bi-functional, as it represses the Rho and IRBP promoters [52] and the kidney-specific Clc-K1 and Clc-K2 promoters [53] but activates the Glut4 (glucose transporter 4) promoter [54]. KLF1/EKLF [57,58], KLF8 [59] and KLF15 [54] all bind a core CACCC site in vitro, although this is shorter than the 9 bp site predicted for a three zinc-finger containing protein and is likely to provide only limited specificity. To gain further insight into the mechanism of KLF15-mediated regulation of gene expression, we have characterized the binding characteristics and specificity of KLF15 and defined the KLF15 binding sites in the Rho and IRBP promoters.
Results
EMSA analysis of KLF15 binding
DNA binding properties of KLF15 were analyzed by electrophoretic mobility shift assays (EMSA) using GST-tagged, human KLF15 fusion protein. Constructs for the full-length KLF15 fusion protein expressed poorly in bacterial cells and yielded a highly degraded protein. Therefore, a recombinant fusion protein (KLF15-ZF) containing the C-terminal 123 amino acids including the three C-terminal zinc-finger domains fused to GST was expressed and purified from bacteria. Western blot of crude bacterial lysates using anti-GST antibodies showed minor degradation of the fusion protein prior to purification; however, the majority of the protein recovered following affinity purification consisted of intact KLF15-ZF fusion protein (Fig. 1A).
Comparison of the proximal promoter regions showed that the 29 bp fragment of the bovine Rho promoter (bRho29) used as bait in the yeast one-hybrid assay that identified KLF15 [52] shared high sequence identity with the equivalent element from the human rhodopsin promoter (hRho29), differing by only 5 bp. In EMSA, KLF15-ZF fusion protein shifted oligomers for both bRho29 (Fig. 1B) and hRho29 (Fig. 1C) and these could be supershifted by addition of antibodies against KLF15 (generous gift of S. Uchida [53]) (Fig. 1D) or GST (Fig. 1E). The lower molecular weight band that was supershifted by the anti-KLF15 but not by anti-GST antibodies most likely resulted from low levels of protein that had lost the N-terminal, GST-tag, but retained the C-terminus of the protein containing the DNA binding domain and the anti-KLF15 antibody recognition site.
Zinc dependence of KLF15 binding
The secondary structure of zinc-finger domains and as well as their ability to bind DNA is dependent on the chelation of Zn+2 ions by Cys2-His2 residues within each zinc finger [60,61]. To determine if DNA binding by KLF15-ZF was in fact Zn+2 dependent, aliquots of the fusion protein were diluted in EDTA to chelate divalent cations and subsequently added to binding reactions containing 32P-labeled bRho29 or hRho29 oligonucleotides. Increasing concentrations of EDTA resulted in a dose-dependent reduction in KLF15-ZF binding to bRho29 (Fig. 2A, B) or to hRho29 (data not shown), with binding completely eliminated when protein was prepared in 50 mM EDTA, resulting in 20 mM EDTA in the final binding reaction. DNA binding was reconstituted by addition of 5 mM ZnCl2 (Fig. 2A) or ZnAcetate (data not shown), but not MgCl2 (Fig. 2B) to the binding reactions. As an additional test of cation specificity, we tested KLF15 binding activity in the presence of EGTA, a chelator that preferentially binds Ca+2. No loss of KLF15-ZF binding was observed in the presence of up to 250 mM EGTA (Fig. 2C). It is interesting to note that the retina and other ocular tissues contain unusually high levels of zinc compared to other tissues and zinc deficiency has been implicated with various retinopathies including macular degeneration, night blindness and retinitis pigmentosa [62,63]. Chelation of intracellular zinc can modulate conformation and DNA binding activity of the zinc-finger transcription factor p53 in cultured cells [64,65]; therefore, physiological zinc deficiencies could result in loss of DNA binding activity of zinc-finger proteins in vivo.
Sequence specificity of KLF15 binding
Sequence specificity of KLF15-ZF binding was examined by EMSA using a series of unlabeled oligomers as competitors (Fig. 3). The addition of 250-fold excess of cold bRho29 or hRho29 effectively competed with the radiolabeled bRho29 probe for KLF15 binding. We also tested oligomers IRBP1 and IRBP2 that contained binding sites for the zinc-finger transcription factor MOK2 and are located in the promoter and intron 2 of IRBP respectively [48]. Neither of these heterologous oligomers competed with radiolabeled bRho29 (Fig. 3) or hRho29 probes (data not shown).
To identify the specific bases within bRho29 and hRho29 that were critical for binding, we generated a series of scanning oligomers, each containing a triplet of mutated bases (Fig. 4A). Because an individual zinc-finger domain typically binds to a DNA triplet, this approach was selected to maximize the likelihood that the mutations would disrupt protein/DNA interactions. Oligomers with 1 or more mutated bases within the central C-rich sequence (bovine: ACG CCC CCA; human ACA CCC CCA) had reduced ability to compete with the wild-type bRho29 (Fig. 4B) or hRho29 (Fig. 4C), whereas oligomers with mutations of any other triplet competed as well as wild-type.
To test if there was binding activity in retinal proteins that was specific to the KLF15 binding site, we performed EMSA using a total retinal extract with either the wild-type bRho29 oligo or the Δ10 oligo (see Fig. 4A) containing a mutated KLF15 binding site (Fig. 5). With both the wild-type and mutant oligos, there were multiple shifted bands observed, however mutation of the KLF15 binding site eliminated several of the major bands. Bands that were unique to the wild-type oligo were not abolished by the addition of up to 200-fold excess of the Δ10 oligo as a competitor, while those that were common to both were reduced in intensity to near background levels. This is consistent with our findings that Δ10 was unable to compete with the wild-type bRho29 for binding to the KLF15-ZF fusion protein in competitive EMSA (Fig. 4B). We observed that with the wild-type oligo, those bands that were dependent on the presence of an intact KLF15 binding site showed the greatest intensity; in contrast, when the KLF15 binding site was mutated, those bands that were not dependent on the presence of the KLF15 binding site were increased in intensity.
KLF15 binding sites in rhodopsin and IRBP promoters
DNAse I footprinting was used to identify KLF15 binding sites in the bovine Rho and IRBP promoters. Using two overlapping fragments of the bovine RPPR (-225 to + 70 bp and -315 to -31 bp) as templates, KLF15-ZF protected six distinct sites (designated KR-a to KR-f) that were generally similar in their boundaries on both the forward and reverse strands (Fig. 6 A–D). In the IRBP proximal promoter, KLF15 protected three sites (KI-a, KI-b, KI-c) on both forward and reverse strands (Fig. 7A, B). Examination of the sequences protected by KLF15-ZF showed that all contained extended C-rich or G-rich sequences (Fig. 10A), consistent with the site identified by competitive EMSA. To identify a consensus binding site, we analyzed the sequences of the C-rich strands of the nine KLF15 protected regions using Target Explorer software [66]. Assuming a 60% AT, 40% GC content in the genome and a 9 bp binding site, we identified the top 4 matrices (Additional file: 1). All of the matrices included the site identified by competitive EMSA and, when used to evaluate the oligos tested in competitive EMSA, could discriminate between the strong competitors and weak/non-competitors.
Matrix 2 yielded the greatest difference between the scores of competitors and non-competitors (Fig. 10). Possible scores ranged from -17.45 to 10.65, with the scores for the KLF15 protected sites ranging from 7.26 to 10.27. Oligos that were good competitors in EMSA all contained at least one binding site with a score > 5.68, whereas those that were poor competitors had maximal scores < 4.06. Using 1/2 of the difference between these scores as a cutoff (4.87), we analyzed the proximal promoters of rhodopsin (Fig. 8A) and IRBP (Fig. 8B) from human, chimp, dog, mouse and rat to identify potential KLF15 binding sites (Fig. 8; see also Additional file: 1). These analyses showed that the KR-a site was conserved across all mammalian rhodopsin promoters analyzed, and the KR-d site was present in all but the dog promoter, which contained an additional KLF15 binding site immediately 5' to KR-e. The other four KLF15 binding sites were conserved only in non-rodents, although one novel binding site located 3' to site KR-d was predicted in the mouse and rat promoters. In the IRBP promoters, KI-b and KI-c were conserved in bovine, human, chimp and dog, with an additional site predicted in the mouse promoter (5' to KI-a) and in bovine and dog (5' to KI-c, outside of the region analyzed by DNAse I footprinting).
Effects of KLF15 on a minimal bovine rhodopsin promoter
We previously reported that KLF15 repressed transactivation of bRho225-luc, a promoter-luciferase reporter construct containing -225 to +70 bp of the bovine rhodopsin promoter [52] that contained the six KLF15 binding sites identified in DNAse I footprinting analysis. A smaller fragment of the rhodopsin proximal promoter (-130 to +70 bp) containing only four KLF15 binding sites, but lacking the KR-e/CRS-1 and KR-f sites, is still sufficient to drive expression of a reporter gene in primary cultures retinal cells from chick [36]. Using a luciferase reporter construct containing this smaller fragment of the rhodopsin promoter (bRho130-luc), the results of transient transfections of 293 cells were qualitatively similar to those previously reported using the bRho225-luc construct: KLF15 alone or in co-transfections with CRX and/or NRL resulted in statistically significant decreases in luciferase expression (Fig. 9). In co-transfections with CRX, high concentrations of KLF15 were more effective at reducing luciferase expression (Fig. 9B); in contrast, there was a relatively concentration-independent reduction (>50%) in luciferase expression in co-transfections with NRL at all concentrations of KLF15 tested (Fig. 9C). We compared the results of promoter transactivation assays using bRho130-luc with those previously obtained using bRho225-luc [52] and found that luciferase expression in transfections with KLF15 was consistently lower with bRho225-luc than with bRho130-luc, with the differences within the 95% confidence interval for KFL15+CRX (p = 0.0355) and within 90% confidence interval for KLF15 alone (p = 0.0879), KLF15+NRL (p = 0.0594) and KLF15+CRX+NRL (p = 0.0894).
Discussion
We previously identified KLF15 as a transcriptional repressor of the rhodopsin and IRBP promoters in vitro and report here the characterization of KLF15's DNA binding properties. For these analyses, we analyzed protein binding specificity on bovine promoters using a truncated human KLF15-GST fusion protein containing the C-terminal DNA binding domain, but lacking the N-terminal repressor domain [52]. Since zinc-finger domains typically function as modular DNA binding motifs that can bind their target sequences independently of other protein domains [67,68], the binding specificity of the truncated KLF15 fusion protein is predicted to be similar, if not identical, to the full length protein. Previous studies showed that recombinant GST fusion proteins containing the full length human or rat KLF15 bound to an element in the CLC-K1 promoter [53] and we found that our truncated KLF15 fusion protein also bound the same element in EMSA (data not shown). The amino acid sequence of the zinc finger domains of bovine KLF15 differs from human, chimp, dog, mouse and rat at only a single amino acid residue (isoleucine to valine) and the flanking regions that were included in the KLF15-ZF fusion protein differ at only 11 amino acid residues between human and bovine KLF15. We would predict, therefore, that the DNA binding specificity would show only minor differences (if any) between these species. Supporting this prediction, the human KLF15-GST fusion protein shifted oligos containing either the bovine or human KLF15 consensus site. Using EMSA, we have confirmed that proteins present in retinal extracts can bind specifically to the KLF15 binding site. However, this result does not necessarily mean that the observed binding activity is attributable to KLF15. Other members of the KLF family recognize similar G/C rich binding motifs [54,57-59] and MAZ (myc-associated zinc-finger protein) can also recognize the KLF15 core binding site [53]. Thus, it remains a distinct possibility that, in addition to KLF15, there are other retinal proteins that can recognize this site.
Several of the KLF15 binding sites that we identified in DNAse I footprinting correspond to previously identified promoter elements. In EMSA, both retina-specific and ubiquitous nuclear proteins bind a 20 bp probe (-55 to -36) containing the Ret-4 element (including the KR-b site) in the rhodopsin promoter [24]. Interestingly, mutations in the G-rich sequence within the Ret-4 element that would be predicted to disrupt the KLF15 binding site eliminated binding by the ubiquitously expressed proteins [24]. In DNAse I footprinting assays of the bovine rhodopsin proximal promoter, nuclear extracts from bovine retina [21] and from Y79 and WERI retinoblastoma cells [69] protected the BAT-1 (-103 to -84) site. Two additional sites protected by bovine retinal proteins are Ret-1 (-138 to -126) and CRS-1 (-177 to -199) [16,20]. Three of these sites correspond to sites protected by KLF15 in our DNAse I footprinting, KR-6/Ret-4, KR-c/BAT-1 and KR-e/CRS-1. In addition to being protected by nuclear proteins from retinal extract, the CRS-1 site is protected by proteins present in nuclear extracts from non-retinal tissues [21]. KLF15 is expressed in the multiple tissues including the retina [52-54] and can repress rhodopsin promoter constructs in transient transfections [52]. Although we do not yet know the relative importance of these different KLF15 binding sites in regulation of target gene expression in vivo, our in vitro assays showed that KLF15 more effectively repressed a rhodopsin promoter construct (bRho225) containing six KLF15 binding sites (including KR-e/CRS-1) than a shorter promoter construct (bRho130) that contains only 4 (including the KR-b/Ret-4 and KR-c/BAT-1 sites), suggesting that both the proximal and distal sites influence KLF15's repressor activity.
The G-rich repressor element is located between -156 and -70 bp in IRBP promoter and overlaps the KI-c binding site for KLF15. Deletion of the G-rich element increases reporter gene expression in retinoblastoma cells [14,19,34] and results in inappropriate expression of GFP reporter constructs in non-retinal tissues in Xenopus laevis [28]. In DNAse I footprinting, this site is protected by proteins present in nuclear extracts from both retinoblastoma WERI-Rb1 and non-retinal HeLa cells [18]. Based on the ability of oligos containing the Sp1 consensus site to compete with the G-rich element in EMSA, it was proposed that Sp1 or Sp1-related proteins bind this site in vivo [19]. KLF15 represses transactivation of an IRBP promoter construct in vitro, is expressed in both retinal and non-retinal tissues, and is structurally and phylogenetically related to the Sp family of transcription factors [52], making it an attractive candidate to bind the G-rich repressor element in vivo.
Potential KLF15 binding sites corresponding to the CRS-1 and G-rich sites are present in the bovine, human, chimp and dog promoters. It was, however, somewhat surprising that no consensus KLF15 binding sites were identified at corresponding locations in the mouse or rat promoters. As the zinc-finger domains are identical between these species, it seems unlikely that the lack of identifiable binding sites at these locations reflects species-specific differences in KLF15's binding specificity. Although we cannot formally exclude this possibility, there are several possible alternative interpretations of this observation. The simplest explanation is that the location or number of KLF15 binding sites differs between species, possibly reflecting the accumulation of mutations during evolution. A comparison of well characterized promoters of 51 genes found substantial differences between human and mouse, with 32–40% of the binding sites identified in the human promoters containing sufficient sequence changes in the mouse promoters to render them non-functional [70]. In addition, despite our evidence showing that KLF15 binds the CRS-1 and G-rich repressor elements in vitro, there may be species-specific differences in the identity and/or sequence specificity of the transcription factor(s) that bind this site in vivo.
Changes in transcriptional regulation are thought to play a significant role in generating phenotypic differences between species [71] and there are clearly morphological and developmental differences between cows, primates, dogs and rodents. In the retina, differences include the time course of embryonic/fetal development and cellular differentiation, as well as the proportions and absolute numbers of different neuronal subtypes (e.g. rods vs. cones) present in the mature retina. Along with these morphological differences, there are species-specific variations in the regulation of photoreceptor-specific gene expression. In the bovine retina, transcripts for rhodopsin, IRBP, arrestin, rod α-transducin and rod α-PDE all begin to accumulate relatively late in retinal development, at the time when the first outer segments are detected [7]. In contrast, in mice and rats there is considerable variation in the developmental stage when different photoreceptor-specific genes are first expressed [8,9,72]. Interestingly, the CRS-1/KR-e site was initially proposed to play a role in repressing premature rhodopsin expression during retinal development, based on the observation that protection of the CRS-1 site by fetal bovine retinal proteins diminishes throughout development [21] in a temporal pattern that is coordinated with increases in rhodopsin transcription and mRNA accumulation [9]. Further emphasizing the extent of the differences between species, a large scale analysis of expressed genes in the mouse retina found that out of nearly 2000 known genes that could be assigned map locations in the mouse genome, 25% had no identifiable human ortholog [73]. Although the identification and characterization of conserved genes and regulatory mechanisms has provided significant contributions to our understanding of retinal development and disease, there are likely to be novel and potentially valuable insights that can be obtained through comparative analysis of the basis of species-specific differences.
Several KLF proteins have been shown to bind a core CACCC element [57-59] and Gray and colleagues [54] demonstrated that 5'-CACCC-3' is sufficient for KLF15 binding in EMSA. A single zinc finger typically makes contact with 3 adjacent bases [67] therefore the three zinc-fingers of KLF15 would be predicted to have a 9 bp binding site, consistent with our EMSA results. We found that mutations outside of the 9 bp consensus site had no effect on DNA binding in EMSA; however, when we used TargetExplorer to re-analyze the KLF15 protected sequences identified in DNAse I footprinting, making no a priori assumptions as to the actual size of the consensus binding site, a longer 12–13 bp consensus sequence was identified. The functional significance of the extended consensus site is unclear, but it is interesting to note that the KR-b, KR-c and KR-d sites respectively overlapped the Ret-4, BAT-1 and Ret-1 sites, all containing known binding sites for homeodomain containing proteins including CRX [30,31]. The KR-b site is located at the 3' end of the Ret-4 site within a G-rich sequence is a binding site for nuclear proteins present in both retinal and non-retinal tissues [24]. The highly conserved KR-d site is within a GC-rich sequence that functions to facilitate binding of retinal nuclear proteins to the adjacent Ret-1 site [15]. One interpretation is that KLF15 binding sites are preferentially located adjacent to binding sites for other transcription factors, possibly CRX or other OTX-related factors. In EMSA using total retinal proteins, mutation of the KLF15 binding site increased binding of some retinal proteins to the Δ10 oligo, as shown by the increased intensity of KLF15-independent bands (Fig. 5). As mutation of the KLF15 binding site does not alter the core homeodomain binding site present in these oligos, homeodomain proteins (e.g. CRX; RX/RAX; OTX2) present in retinal extracts likely bind to both the wild-type and mutant oligos. Although mutation of the KLF15 site may have inadvertently created a novel protein binding site, several of the bands showing increased intensity with the Δ10 oligo were also present using the wild-type bRho29 oligo, raising the possibility that occupancy of the KLF15 site can affect protein/DNA interactions of other DNA binding proteins. The proximity of KLF15 and CRX binding sites and our findings that (1) KLF15 can repress transactivation of the rhodopsin promoter by CRX in vitro and (2) both KFL15 and CRX are expressed in non-photoreceptor cells in the inner nuclear layer [52] are consistent with a possible in vivo role for KLF15 in repressing inappropriate activation of the rhodopsin promoter by CRX in non-photoreceptor cells.
A similar correlation between KLF15 and CRX binding sites was not observed in the IRBP promoter. However, the KI-c site/G-rich repressor element in the bovine IRBP promoter is adjacent to a CpG dinucleotide (-115) that becomes hypomethylated specifically in retinal cells at the time IRBP expression is first detected [74,75]. The G-rich repressor element is bound by proteins present in non-retinal tissues where the CpG dinucleotide remains methylated; therefore, it is tempting to speculate that KLF15 may participate in regulating methylation at this site. Other KLF genes are known to regulate chromatin remodeling. The transcriptional activator KLF1/EKLF recruits a SWI/SNF-related chromatin remodeling complex that is thought to be involved in regulation of the locus control region of the β-globin gene [76]. The transcriptional repressor KLF14/BTEB3 interacts with a co-repressor (mSin3A) and the histone deacetylase protein HDAC-1 though repressor domains within the N-terminus to mediate transcriptional repression through chromatin remodeling [77]. Interestingly, in our footprinting analysis of the rhodopsin promoter we observed several sites of increased DNAse I hypersensitivity adjacent to KLF15 protected regions, possibly reflecting alterations in DNA configuration and potentially lending additional support for involvement of KLF15 in chromatin remodeling.
Conclusion
The goal of these studies was to determine the binding characteristics and specificity of KLF15 and define the KLF15 binding sites in the Rho and IRBP promoters in order to gain further insight into the mechanism of KLF15-mediated regulation of retinal gene expression. We found that KLF15 binds to multiple 9 bp consensus sites in the rhodopsin and IRBP promoters including the previously identified CRS-1 and G-rich repressor elements. The presence of multiple KLF15 binding sites suggests that transcriptional repression is likely to play an important role in regulation of photoreceptor-specific genes. Based on the known expression of KLF15 in retinal and non-retinal tissues and the absence of KLF15 expression in photoreceptors, we hypothesize an in vivo role for KLF15 in repressing photoreceptor-specific gene expression in non-photoreceptor cells.
Methods
Protein isolation
To generate KLF15-ZF-GST, the 3' end of the human KLF15 coding region (369 bp) and the stop codon was amplified from pCRII-hKLF15 [52] using a combination of Taq (Invitrogen; Carlsbad, California) and PFU polymerase (Promega; Madison, WI), subcloned into the BamHI/EcoRI sites of pGEX (Amersham/Pharmacia; Piscataway, NJ) and confirmed by sequencing. KLF15-ZF fusion protein with an N-terminal, glutathione-S-transferase (GST) tag was purified from whole cell lysates of IPTG-induced (0.4 mM) E. coli (BL21S; Invitrogen; Carlsbad, California) using sepharose 4B (Amersham/Pharmacia; Piscataway, NJ) according to maunfacturer's instructions. Purified protein was quantified using bicinchoninic acid protein assay kit (BCA; Sigma; St. Louis, MO) and analyzed using standard 10% SDS-polyacrylamide gel electrophoresis and Western blot with anti-GST antibodies (Amersham/Pharmacia; Piscataway, NJ) diluted 1:8000.
Mice used for protein isolation were handled in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals and methods were approved by the Animal Care and Use Committee of Johns Hopkins University. For total protein extracts, retinas were dissected from adult mice and placed directly into ice cold PBS containing protease inhibitors (CompleteMini Tablets, EDTA-free; Roche, Indianapolis, IN), sonicated for 10–15 seconds (Setting 5, 50% Duty Cycle) using a Branson Model 250 Sonifier (Branson Ultrasonics Corp., Danbury, CT) with a micro-tip, quantified by BCA and stored at -80C.
Electrophoretic mobility shift assays (EMSA)
EMSA assays followed previously published protocols [24,78] and used a binding buffer optimized for the KLF15-ZF-GST fusion protein (4 mM HEPES, 5 mM EGTA; 100 mM KCl, 0.25 mM ZnCl2; 0.02% NP40; 0.1 M DTT; 4% glycerol). For analysis of Zn+2 specificity, ZnCl2 was eliminated from the binding buffer except as indicated. Probes with 5'-GG dinucleotide overhangs were generated by annealing single-stranded oligos with their reverse complement and end-labeled with 32P-dCTP using Klenow fragment of DNA polymerase (New England Biolabs, Beverly MA). Cold competitors were end-filled with Klenow using cold dCTP. Binding reactions were incubated for 90 to 120 minutes on ice; for supershift experiments antibodies against GST (Amersham/Pharmacia; Piscataway, NJ) or KLF15 C-terminal peptide [a generous gift of S. Uchida; [53]] were added after an initial 60 minute incubation. Following polyacrylamide gel electrophoresis on a 5% non-denaturing gel, dried gels were exposed to storage phosphor screens; peaks for each band and DLU (Digital Light Units) for the areas under the peaks were determined using Optiquant software (Packard Instruments). For each lane, the percent of oligo shifted was calculated as [(DLU shifted band/ total DLU) × 100]. For competition assays, "percent competed" was calculated as the difference between the percent oligo shifted without competitor and the percent oligo shifted with competitor.
DNAse I footprint analysis
DNAse I footprinting followed previously published procedures [24,78] using 33P-end-labeled promoter fragments generated by PCR, KLF15-ZF-GST or GST fusion proteins and EMSA binding buffer optimized for KLF15-ZF. For bovine Rho promoter fragments, the template was a plasmid containing the bovine upstream Rho promoter (-2174 to +32 bp) [36]; for bovine IRBP promoter fragment (-300 to +132 bp), the template was a plasmid containing the corresponding region of the bovine IRBP promoter [30]. Primers used for PCR amplification of promoter fragments were as follows: for bovine Rho (-225 to +70 bp): forward: 5'-AGGCCTCTGCTCTTTCCC-3'; reverse: 5'-CGCCGGCGGCGCGAACCCG-3'; for bovine Rho (-315 to -31 bp): forward: 5'-AGAGGGAAGTGGGCCTAGAG-3', reverse 5'- GAAGTGACCTCCCCTCCCTA-3'; for bovine IRBP: forward: 5'-CAGATGAGACCCCAACATAC-3'; reverse: 5'-ACAGAAGCTCTCTTGACACC-3'. The amount of fusion protein used varied from 16 to 250 ng per reaction and DNAse I digestions were carried out at room temperature for 1 minute.
Consensus sequence analysis
To identify consensus sequences for protein binding sites, sequences protected in DNAse I footprinting were analyzed using Target Explorer [66], a program that uses CONSENSUS (version 6c) and WCONSENSUS (version 5c) to determine consensus patterns in unaligned sequences using an algorithm based on a matrix representation of a consensus pattern. WCONSENSUS determines the width of the pattern being sought without a priori knowledge of the length of the consensus sequence.
Transient transfections
Promoter transactivation assays used transient transfections of human embryonic kidney cells (293) with expression vectors, hKLF15-pcDNA3.1 [52], Crx-pcDNA3.1 [30], Nrl-pED (gift of A. Swaroop [35]) and luciferase reporter plasmid bRho130-luc [36] following previously published methods [30,52]. The total amount (μg) of DNA in each plate was kept constant by addition of the corresponding expression vector(s) lacking a cDNA insert. Basal activity of each reporter construct in 293 cells was defined as the relative luciferase expression in control plates transfected with reporter constructs and "empty" expression vectors. For each experiment, at least two independent transfections were performed on separate days, and for each combination of expression vectors tested, at least 4 DNA precipitates prepared with each precipitate divided equally to transfect two plates for a total of≥8 transfections per condition.
Statistical analyses
For each condition the fold change in relative luciferase relative to control transfections was calculated and analyzed as previously described [52]. Statistical analysis used linear mixed models and comparisons of bRho130-luc and bRho225-luc reporter constructs used Wilcoxon Two Sample test (SAS; Cary, NC).
List of abbreviations
ATXN7, ataxin-7; BAF, barrier to autointegration factor 1; BTEB3, basic transcription element binding protein 3; Clc-K, kidney-specific chloride channel; CRX, cone-rod homeobox; DTT, dithiothreitol; EDTA, ethylenediaminetraacetic acid; EGTA, O,O'-bis(2-aminoethyl)ethyleneglycol-N,N,N',N'-tetraacetic acid; EKLF, erythroid-specific Krüppel-like factor; ERX, empty spiracles-related homeobox; FIZ-1, Flt3 interacting zinc finger protein 1; GST, glutathione-S-transferase; HEPES, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; IPTG, isopropyl-β-D-thiogalactopyranoside; KLF, Krüppel-like factor; NRL, neural retina leucine zipper; OTX, orthodenticle-related homeobox; QRX, Q50-type retinal homeobox; Rbp3, retinoid binding protein 3; RX/RAX, retina and anterior neural fold homeobox.
Authors' contributions
DCO participated in the experimental design, generated and purified fusion proteins, carried out EMSA, DNAse I footprinting, transient transfections, consensus sequence analysis, promoter alignments and drafted the manuscript. HL performed all statistical analyses and contributed to the interpretation of the transfection data; YL isolated the KLF15 clone and contributed to the conceptual basis of the study; DJZ conceived the study, participated in experimental design and data interpretation and helped to draft the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Supplemental Data Files (Otteson SupplementalData.doc) containing full results of analysis of KFL15 binding sites using Target Explorer. Included in this data set are: 1. Sequences from bovine rhodopsin and IRBP promoters protected by KLF15-gst fusion proteins in DNAseI footprint analysis used as "training set" for Target Explorer. 2. List of top matrices for KLF15_9bpsite_otteson binding site. (sorted by information content): Assume 9 bp binding site and 60% AT/40% GC in genome. 3. Analysis to determine cutoff score for identification of KLF15 binding sites. Sequences and scores for bRho29, hRho29 and bRho29-mutations oligonucleotides used in competitive EMSA. Lists of putative binding sites in these oligos identified using Matrices 1–4: cut-off score = 1 indicating good and poor competitors. 4. List of top matrices for KLF15: Assume unknown binding site length and 60% T/40% GC in genome. 5. Analysis to determine cutoff score for identification of KLF15 binding sites assuming unknown binding site length. Sequences and scores for bRho29, hRho29 and bRho29-mutations oligonucleotides used in competitive EMSA. Lists of putative binding sites in these oligos identified using Matrices 1–4: cut-off score = 1 indicating good and poor competitors. 6. Identification of potential KLF15 binding sites in rhodopsin and IRBP promoters using matrix 2 with 9 bp binding site (cut-off score 4.87). Promoter sequences analyzed. Predicted binding sites
Click here for file
Acknowledgements
Supported in part by grants F32EY13499 (DCO) and EY09769 (DJZ) from the NEI/NIH; the Foundation for Fighting Blindness Foundation and generous gifts from Marshall and Stevie Wishnack, and Robert and Clarice Smith; DJZ is the Guerrieri Professor of Genetic Engineering and Molecular Ophthalmology and is a recipient of the Senior Investigator Award from Research to Prevent Blindness.
Figures and Tables
Figure 1 EMSA analysis of KLF15 binding. (A) KLF15-ZF fusion protein. Lanes 1 and 2 show western blots of crude bacterial lysates at 1.5 and 2 hours following induction of fusion protein expression. Anti-GST antibodies detect a major band of the anticipated size and additional lower molecular weight bands. Lane 3, Coomassie stained gel of affinity purified KLF15-ZF-GST fusion protein showing enrichment of major band containing full length fusion protein. (B) EMSA using 32P-labeled oligomers (bRho29) containing a 29 bp fragment from bovine rhodopsin promoter (-94 to -66). Lane 1: no protein; Lane 2: 67.5 ng GST; Lane 3–5: 100, 50, 25 ng KLF15-ZF-GST. (C) Same as (B) except 32P-labeled oligonucleotide (hRho29) contained corresponding sequence from human rhodopsin promoter. (D) Supershift using 32P-bRho29 oligonucleotide and anti-KLF15 antibodies. Lane 1, no protein; Lane 2, 33 ng GST protein; Lane 3, 33 ng GST + anti-KLF15; Lane 4, 50 ng KLF15-ZF-GST protein; Lanes 5–7, 50 ng KLF15-ZF-GST protein plus increasing amounts of anti-KLF15. (E) Supershift using 32P-bRho29 oligonucleotide and anti-GST antibodies. Lane 1, 50 ng KLF15-ZF-GST protein; Lanes 2–4, 50 ng KLF15-ZF-GST protein plus increasing amounts of anti-GST. Large arrows, KLF15 shifted bands; small arrows, non-specific bands; arrowheads, supershifted bands.
Figure 2 Effects of zinc chelation on KLF15 Binding. A. EMSA using 32P-labeled oligomers (bRho29) containing -96 to -66 bp fragment from bovine rhodopsin promoter. Lane 1, GST protein; Lane 2, KLF15-ZF-GST protein, Lanes 3 -7 same as 2 with EDTA and ZnCl2 (mM) as indicated. B. Same as A, but with MgCl2. C. Same as A, but with EGTA and ZnCl2. The altered migration of the unbound oligonucleotides is attributable to the salt effects of high concentrations of EGTA.
Figure 3 Competitive EMSA using (A) 32P- bRho29 oligonuclotides (bovine). Lane 1, no protein; lane 2, GST protein control; lane 3, KLF15-ZF-GST without competitor; lanes 4 -11, same as lane 3 but with unlabeled competitors (50 ×, 250 × fold-excess) as indicated. Large arrows indicate specific bands; small arrow indicates non-specific band present in negative (no protein and GST only) controls; IRBP1, MOK2 site from bovine IRBP promoter (5'-GGACAGGATTAAAGGCTTACTGGAG-3'); IRBP2, MOK2 site from bovine IRBP intron1 (5'-GGACTTGTCAGGGCCTTTA-3')
Figure 4 Sequence specificity of KLF15 binding: mutational analysis of the bovine rhodopsin promoter element (bRho29). (A) Sequence of oligomers used in EMSA analysis showing mutations analyzed in B and C. (B) Graph summarizing results of competitive EMSA using 32P-labeled bRho29 and unlabeled mutant oligomers as indicated. B, bRho29 competitor; H, hRho20 competitor. Bars show % of shifted oligonucleotide competed by addition of cold competitor. Bars with negative value reflect an increase in the amount of oligonucleotide shifted following the addition of cold competitor. (C) Same as B, except 32P-labeled hRho29 oligonucleotide was used.
Figure 5 Binding of KLF15 site by retinal proteins in EMSA. All lanes contain equal amounts of total retinal extract from wild-type adult mouse retina. Lanes 1, 3, 4 and 5, 32P-labeled bRho29 oligo; lane 2, 32P-labeled Δ10 oligo containing mutated KLF15 binding site; cold Δ10 oligo as competitor was added to lane 4, at 20-fold excess and lane 5, at 200-fold excess. Open arrowheads, bands shifted with both wild-type and mutant oligos; solid arrowheads, bands observed only with wild-type oligos; large and small arrows respectively indicate major and minor bands not competed by addition of excess mutant oligo. Discontinuities in the image resulted from removal of irrelevant lanes.
Figure 6 DNAse I footprint analysis of KLF15 binding sites in the rhodopsin promoter. (A) A PCR fragment spanning -225 to +70 bp of the bovine RPPR was end labeled with 33P at the -225 bp end (forward strand) and incubated with 250, 125, 62.5 and 31.25 ng of purified KLF15-ZF fusion protein (lanes 2–5). Lanes 1 and 6 contained no fusion protein. Protected regions are designated by lines on the left of the image with the nucleotide position of each protected region indicated. Previously identified protein binding sites are indicated by brackets. Solid arrowheads indicate novel hypersensitive sites resulting from KLF15-ZF binding; small arrows indicate hypersensitive sites that are lost or altered by KLF15-ZF binding. Transcription start site (+1) is indicated by an open arrowhead. (B) Same as (A) except the -225 to +70 fragment was labeled at the +70 end (reverse strand). (C) Same as (A) except the PCR fragment used spanned -315 to -31 and was end-labeled with 33P at the -315 end (forward strand) and lanes 1 and 5 contained no protein. (D) Same as (C) except the -315 to -31 fragment was labeled at the -31 end (reverse strand).
Figure 7 DNAse I footprint analysis of KLF15 binding sites in the IRBP promoter. (A) A PCR fragment spanning -132 to +70 bp of the bovine IRBP promoter was end labeled with 33P at the -132 bp end (forward strand) and incubated with 250, 125 and 62.5 ng of purified KLF15-ZF fusion protein (lanes 2–4). Lanes 1 and 5 contained no fusion protein. Protected regions are designated by bars on the left of each panel with the nucleotide position of each protected region indicated. Previously identified protein binding sites are indicated by brackets. Transcription start site (+1) is indicated by an open arrowhead. (B) Same as (A) except the fragment was end-labeled with 33P at the +70 end (reverse strand). The bands present in KI-a (indicated by grey bar in Panel A) were also present in negative control lanes containing undigested probe (data not shown).
Figure 8 Alignment of (A) rhodopsin and (B) IRBP proximal promoters. KLF15 protected sites in bovine promoter are boxed; 9 bp KLF15 binding sites predicted using Target Explorer Matrix with a cutoff score of 4.86 (Table 1; see also Additional file: 1) are underlined and in bold; at locations where multiple overlapping binding sites were predicted, only the one with the highest score is marked. Previously identified protein binding sites/regulatory elements are indicated by heavy solid bars above sequence. Numbering is based on the bovine promoter sequence and an arrowhead indicates the transcriptional start site (+1). B, bovine; H, Human; C, chimp; D, dog; R, rat; M, mouse.
Figure 9 Effects of KLF15 on rhodopsin promoter. (A) 293 cells in 22 mm dishes were co-transfected with 2.5 μg of bovine rhodopsin (-130 to +70 bp)-luciferase fusion construct (bRho130-luc) together with the indicated amount (μg) of KLF15 expression vector. Luciferase activity (in relative light units) was corrected for transfection efficiency using β-GAL internal control. Bars show fold-change in relative luciferase expression compared to control transfections using expression vector lacking cDNA inserts. Error bars, 95% confidence interval. (B) same as (A) except 0.25 μg Crx expression vector was added. (C) Same as (A) except 0.25 μg Nrl expression vector was added. (D) Same as (A) except 0.25 μg Crx and 0.25 μg Nrl were added. Statistical differences from (A) control, (B) Crx alone; (C) Nrl alone; and (D) Crx+Nrl indicated by * p = 0.012, ** p < 0.008, *** p < 0.0001.
Figure 10 Analysis of KLF15 protected sequences in Rhodopsin and IRBP promoters using Target Explorer: Matrix 2: A. Alignment Matrix; B. Frequency Matrix; C. Scoring Matrix
==== Refs
Carter-Dawson L Alvarez RA Fong SL Liou GI Sperling HG Bridges CD Rhodopsin, 11-cis vitamin A, and interstitial retinol-binding protein (IRBP) during retinal development in normal and rd mutant mice Dev Biol 1986 116 431 438 3732615 10.1016/0012-1606(86)90144-2
Hauswirth WW Langerijt AV Timmers AM Adamus G Ulshafer RJ Early expression and localization of rhodopsin and interphotoreceptor retinoid-binding protein (IRBP) in the developing fetal bovine retina Exp Eye Res 1992 54 661 670 1623951 10.1016/0014-4835(92)90021-J
Thulin CD Howes K Driscoll CD The immunolocalization and divergent roles of phosducin and phosducin-like protein in the retina Mol Vis 1999 5 40 10617777
Zhu X Li A Brown B Weiss ER Osawa S Craft CM Mouse cone arrestin expression pattern: light induced translocation in cone photoreceptors Mol Vis 2002 8 462 12486395
Blackshaw S Snyder SH Developmental expression patterns of phototransduction components in mammalian pineal implies a light sensing function J Neurosci 1997 17 8074 8082 9334383
Liou GI Matragoon S Yang J Geng L Overbeek PA Ma DP Retina-specific expression from the IRBP promoter in transgenic mice is conferred by 212 bp of the 5'-flanking region Biochem Biophys Res Commun 1991 181 159 165 1958183 10.1016/S0006-291X(05)81395-6
Timmers AM Newton BR Hauswirth WW Synthesis and stability of retinal photoreceptor mRNAs are coordinately regulated during bovine fetal development Exp Eye Res 1993 56 257 265 8386100 10.1006/exer.1993.1034
Bibb LC Holt JK Tarttelin EE Temporal and spatial expression patterns of the CRX transcription factor and its downstream targets. Critical differences during human and mouse eye development Hum Mol Genet 2001 10 1571 1579 11468275 10.1093/hmg/10.15.1571
Treisman JE Morabito MA Barnstable CJ Opsin expression in the rat retina is developmentally regulated by transcriptional activation Mol Cell Biol 1988 8 1570 1579 2967911
DesJardin LE Timmers AM Hauswirth WW Transcription of photoreceptor genes during fetal retinal development. Evidence for positive and negative regulation J Biol Chem 1993 268 6953 6960 8463227
Zack DJ Bennett J Wang Y Unusual topography of bovine rhodopsin promoter-lacZ fusion gene expression in transgenic mouse retinas Neuron 1991 6 187 199 1899580 10.1016/0896-6273(91)90355-4
Gouras P Kjeldbye H Zack DJ Reporter gene expression in cones in transgenic mice carrying bovine rhodopsin promoter/lacZ transgenes Vis Neurosci 1994 11 1227 1231 7841129
Quiambao AB Peachey NS Mangini NJ Rohlich P Hollyfield JG al-Ubaidi MR A 221-bp fragment of the mouse opsin promoter directs expression specifically to the rod photoreceptors of transgenic mice Vis Neurosci 1997 14 617 625 9278991
Bobola N Hirsch E Albini A Altruda F Noonan D Ravazzolo R A single cis-acting element in a short promoter segment of the gene encoding the interphotoreceptor retinoid-binding protein confers tissue-specific expression J Biol Chem 1995 270 1289 1294 7836393 10.1074/jbc.270.3.1289
Morabito MA Yu X Barnstable CJ Characterization of developmentally regulated and retina-specific nuclear protein binding to a site in the upstream region of the rat opsin gene J Biol Chem 1991 266 9667 9672 1827795
Yu X Barnstable CJ Characterization and regulation of the protein binding to a cis-acting element, RET 1, in the rat opsin promoter J Mol Neurosci 1994 5 259 271 7577368
Sheshberadaran H Takahashi JS Characterization of the chicken rhodopsin promoter: identification of retina-specific and glass-like protein binding domains Mol Cell Neurosci 1994 5 309 318 7804600 10.1006/mcne.1994.1037
Fong SL Fong WB Elements regulating the transcription of human interstitial retinoid-binding protein (IRBP) gene in cultured retinoblastoma cells Curr Eye Res 1999 18 283 291 10372988 10.1076/ceyr.18.4.283.5360
Fei Y Matragoon S Smith SB Functional dissection of the promoter of the interphotoreceptor retinoid-binding protein gene: the cone-rod-homeobox element is essential for photoreceptor-specific expression in vivo J Biochem (Tokyo) 1999 125 1189 1199 10348924
Liou GI Matragoon S Overbeek PA Fei Y Identification of a retina-specific footprint within the retina-specific regulatory region of the human interphotoreceptor retinoid-binding protein gene Biochem Biophys Res Commun 1994 203 1875 1881 7945340 10.1006/bbrc.1994.2406
DesJardin LE Hauswirth WW Developmentally important DNA elements within the bovine opsin upstream region Invest Ophthalmol Vis Sci 1996 37 154 165 8550318
Borst DE Boatright JH Si JS Structural characterization and comparison of promoter activity of mouse and bovine interphotoreceptor retinoid-binding protein (IRBP) gene 5' flanking regions in WERI, Y79, chick retina cells, and transgenic mice Curr Eye Res 2001 23 20 32 11821982 10.1076/ceyr.23.1.20.5419
Yu X Leconte L Martinez JA Barnstable CJ Ret 1, a cis-acting element of the rat opsin promoter, can direct gene expression in rod photoreceptors J Neurochem 1996 67 2494 2504 8931483
Chen S Zack DJ Ret 4, a positive acting rhodopsin regulatory element identified using a bovine retina in vitro transcription system J Biol Chem 1996 271 28549 28557 8910484 10.1074/jbc.271.45.28549
Boatright JH Borst DE Peoples JW A major cis activator of the IRBP gene contains CRX-binding and Ret-1/PCE-I elements Mol Vis 1997 3 15 9479006
Mani SS Besharse JC Knox BE Immediate upstream sequence of arrestin directs rod-specific expression in Xenopus J Biol Chem 1999 274 15590 15597 10336455 10.1074/jbc.274.22.15590
Ma GC Wang TM Su CY Wang YL Chen S Tsai HJ Retina-specific cis-elements and binding nuclear proteins of carp rhodopsin gene FEBS Lett 2001 508 265 271 11718728 10.1016/S0014-5793(01)03058-7
Boatright JH Knox BE Jones KM Evidence of a tissue-restricting DNA regulatory element in the mouse IRBP promoter FEBS Lett 2001 504 27 30 11522290 10.1016/S0014-5793(01)02662-X
Whitaker SL Knox BE Conserved transcriptional activators of the xenopus rhodopsin gene J Biol Chem 2004 279 49010 49018 15347652 10.1074/jbc.M406080200
Chen S Wang QL Nie Z Crx, a novel Otx-like paired-homeodomain protein, binds to and transactivates photoreceptor cell-specific genes Neuron 1997 19 1017 1030 9390516 10.1016/S0896-6273(00)80394-3
Kimura A Singh D Wawrousek EF Kikuchi M Nakamura M Shinohara T Both PCE-1/RX and OTX/CRX interactions are necessary for photoreceptor- specific gene expression J Biol Chem 2000 275 1152 1160 10625658 10.1074/jbc.275.2.1152
Martinez JA Barnstable CJ Erx, a novel retina-specific homeodomain transcription factor, can interact with Ret 1/PCEI sites Biochem Biophys Res Commun 1998 250 175 180 9735352 10.1006/bbrc.1998.9261
Wang QL Chen S Esumi N QRX, a novel homeobox gene, modulates photoreceptor gene expression Hum Mol Genet 2004 13 1025 1040 15028672 10.1093/hmg/ddh117
Bobola N Briata P Ilengo C OTX2 homeodomain protein binds a DNA element necessary for interphotoreceptor retinoid binding protein gene expression Mech Dev 1999 82 165 169 10354480 10.1016/S0925-4773(98)00162-2
Rehemtulla A Warwar R Kumar R Ji X Zack DJ Swaroop A The basic motif-leucine zipper transcription factor Nrl can positively regulate rhodopsin gene expression Proc Natl Acad Sci USA 1996 93 191 195 8552602 10.1073/pnas.93.1.191
Kumar R Chen S Scheurer D The bZIP transcriptionfactor Nrl stimulates rhodopsin promoter activity in primary retinal cell cultures J Biol Chem 1996 271 29612 29618 8939891 10.1074/jbc.271.47.29612
Lerner LE Gribanova YE Ji M Knox BE Farber DB Nrl and Sp nuclear proteins mediate transcription of rod-specific cGMP-phosphodiesterase beta-subunit gene: involvement of multiple response elements J Biol Chem 2001 276 34999 35007 11438531 10.1074/jbc.M103301200
Pittler SJ Zhang Y Chen S Functional analysis of the rod photoreceptor cGMP phosphodiesterase alpha-subunit gene promoter: Nrl and Crx are required for full transcriptional activity J Biol Chem 2004 279 19800 19807 15001570 10.1074/jbc.M401864200
Mitton KP Swain PK Chen S Xu S Zack DJ Swaroop A The leucine zipper of NRL interacts with the CRX homeodomain. A possible mechanism of transcriptional synergy in rhodopsin regulation J Biol Chem 2000 275 29794 29799 10887186 10.1074/jbc.M003658200
Furukawa T Morrow EM Li T Davis FC Cepko CL Retinopathy and attenuated circadian entrainment in Crx-deficient mice Nat Genet 1999 23 466 470 10581037 10.1038/70591
Mears AJ Kondo M Swain PK Nrl is required for rod photoreceptor development Nat Genet 2001 29 447 452 11694879 10.1038/ng774
Sohocki MM Sullivan LS Mintz-Hittner HA A range of clinical phenotypes associated with mutations in CRX, a photoreceptor transcription-factor gene Am J Hum Genet 1998 63 1307 1315 9792858 10.1086/302101
Swain PK Chen S Wang QL Mutations in the cone-rod homeobox gene are associated with the cone-rod dystrophy photoreceptor degeneration Neuron 1997 19 1329 1336 9427255 10.1016/S0896-6273(00)80423-7
Freund CL Wang QL Chen S De novo mutations in the CRX homeobox gene associated with Leber congenital amaurosis Nat Genet 1998 18 311 312 9537410 10.1038/ng0498-311
Jacobson SG Cideciyan AV Huang Y Retinal degenerations with truncation mutations in the cone-rod homeobox (CRX) gene Invest Ophthalmol Vis Sci 1998 39 2417 2426 9804150
Simpson P Positive and negative regulators of neural fate Neuron 1995 15 739 742 7576621 10.1016/0896-6273(95)90163-9
Roopra A Huang Y Dingledine R Neurological disease: listening to gene silencers Mol Interv 2001 1 219 228 14993344
Arranz V Dreuillet C Crisanti P Tillit J Kress M Ernoult-Lange M The zinc finger transcription factor, MOK2, negatively modulates expression of the interphotoreceptor retinoid-binding protein gene, IRBP J Biol Chem 2001 276 11963 11969 11278819 10.1074/jbc.M011036200
La Spada AR Fu YH Sopher BL Polyglutamine-expanded Ataxin-7 antagonizes CRX function and induces cone-rod dystrophy in a mouse model of SCA7 Neuron 2001 31 913 927 11580893 10.1016/S0896-6273(01)00422-6
Wang X Xu S Rivolta C Barrier-to-autointegration factor (Baf) interacts with the cone-rod homeobox (Crx) and represses it transactivation function J Biol Chem 2002 277 43288 43300 12215455 10.1074/jbc.M207952200
Mitton KP Swain PK Khanna H Dowd M Apel IJ Swaroop A Interaction of retinal bZIP transcription factor NRL with Flt3-interacting zinc-finger protein Fiz1: possible role of Fiz1 as a transcriptional repressor Hum Mol Genet 2003 12 365 373 12566383 10.1093/hmg/ddg035
Otteson DC Liu Y Lai H Krüppel-like factor 15, a zinc-finger transcriptional regulator, represses the rhodopsin and interphotoreceptor retinoid-binding protein promoters Invest Ophthalmol Vis Sci 2004 45 2522 2530 15277472 10.1167/iovs.04-0072
Uchida S Tanaka Y Ito H Transcriptional regulation of the CLC-K1 promoter by myc-associated zinc finger protein and kidney-enriched Krüppel-like factor, a novel zinc finger repressor Mol Cell Biol 2000 20 7319 7331 10982849 10.1128/MCB.20.19.7319-7331.2000
Gray S Feinberg MW Hull S The Krüppel-like factor KLF15 regulates the insulin-sensitive glucose transporter GLUT4 J Biol Chem 2002 277 34322 34328 12097321 10.1074/jbc.M201304200
Cook T Gebelein B Urrutia R Sp1 and its likes: biochemical and functional predictions for a growing family of zinc finger transcription factors Ann NY Acad Sci 1999 880 94 102 10415854
Kaczynski J Cook T Urrutia R Sp1- and Krüppel-like transcription factors Genome Biol 2003 4 206 12620113 10.1186/gb-2003-4-2-206
Lee JS Ngo H Kim D Chung JH Erythroid Krüppel-like factor is recruited to the CACCC box in the beta-globin promoter but not to the CACCC box in the gamma-globin promoter: the role of the neighboring promoter elements Proc Natl Acad Sci USA 2000 97 2468 2473 10706605 10.1073/pnas.040476297
Mantovani R Malgaretti N Nicolis S An erythroid specific nuclear factor binding to the proximal CACCC box of the beta-globin gene promoter Nucleic Acids Res 1988 16 4299 4313 2837728
van Vliet J Turner J Crossley M Krüppel-like factor 8: a CACCC-box binding protein that associates with CtBP and represses transcription Nucleic Acids Res 2000 28 1955 1962 10756197 10.1093/nar/28.9.1955
Laity JH Lee BM Wright PE Zinc finger proteins: new insights into structural and functional diversity Curr Opin Struct Biol 2001 11 39 46 11179890 10.1016/S0959-440X(00)00167-6
Michael SF Kilfoil VJ Schmidt MH Amann BT Berg JM Metal binding and folding properties of a minimalist Cys2His2 zinc finger peptide Proc Natl Acad Sci USA 1992 89 4796 4800 1594580
Ugarte M Osborne NN Zinc in the retina Prog Neurobiol 2001 64 219 249 11240307 10.1016/S0301-0082(00)00057-5
Grahn BH Paterson PG Gottschall-Pass KT Zhang Z Zinc and the eye J Am Coll Nutr 2001 20 106 118 11349933
Meplan C Richard MJ Hainaut P Metalloregulation of the tumor suppressor protein p53: zinc mediates the renaturation of p53 after exposure to metal chelators in vitro and in intact cells Oncogene 2000 19 5227 5236 11077439 10.1038/sj.onc.1203907
Verhaegh GW Parat MO Richard MJ Hainaut P Modulation of p53 protein conformation and DNA-binding activity by intracellular chelation of zinc Mol Carcinog 1998 21 205 214 9537652 10.1002/(SICI)1098-2744(199803)21:3<205::AID-MC8>3.0.CO;2-K
Sosinsky A Bonin CP Mann RS Honig B Target Explorer: an automated tool for the identification of new target genes for a specified set of transcription factors Nucleic Acids Res 2003 31 3589 3592 12824372 10.1093/nar/gkg544
Wolfe SA Nekludova L Pabo CO DNA recognition by Cys2His2 zinc finger proteins Annu Rev Biophys Biomol Struct 2000 29 183 212 10940247 10.1146/annurev.biophys.29.1.183
Kang JS Kim JS Zinc finger proteins as designer transcription factors J Biol Chem 2000 275 8742 8748 10722717 10.1074/jbc.275.12.8742
Chau KY Munshi N Keane-Myers A The architectural transcription factor high mobility group I(Y) participates in photoreceptor-specific gene expression J Neurosci 2000 20 7317 7324 11007889
Dermitzakis ET Clark AG Evolution of transcription factor binding sites in mammalian gene regulatory regions: conservation and turnover Mol Biol Evol 2002 19 1114 1121 12082130
Wray GA Transcriptional regulation and the evolution of development Int J Dev Biol 2003 47 675 684 14756343
Bowes C van Veen T Farber DB Opsin, G-protein and 48-kDa protein in normal and rd mouse retinas: developmental expression of mRNAs and proteins and light/dark cycling of mRNAs Exp Eye Res 1988 47 369 390 2846333 10.1016/0014-4835(88)90049-8
Yu J Farjo R MacNee SP Baehr W Stambolian DE Swaroop A Annotation and analysis of 10,000 expressed sequence tags from developing mouse eye and adult retina Genome Biol 2003 4 R65 14519200 10.1186/gb-2003-4-10-r65
Liou GI Wang M Matragoon S Timing of interphotoreceptor retinoid-binding protein (IRBP) gene expression and hypomethylation in developing mouse retina Dev Biol 1994 161 345 356 8313988 10.1006/dbio.1994.1036
Boatright JH Nickerson JM Borst DE Site-specific DNA hypomethylation permits expression of the IRBP gene Brain Res 2000 887 211 221 11134609 10.1016/S0006-8993(00)02990-5
Zhang W Kadam S Emerson BM Bieker JJ Site-specific acetylation by p300 or CREB binding protein regulates erythroid Krüppel-like factor transcriptional activity via its interaction with the SWI-SNF complex Mol Cell Biol 2001 21 2413 2422 11259590 10.1128/MCB.21.7.2413-2422.2001
Kaczynski J Zhang JS Ellenrieder V The Sp1-like protein BTEB3 inhibits transcription via the basic transcription element box by interacting with mSin3A and HDAC-1 co-repressors and competing with Sp1 J Biol Chem 2001 276 36749 36756 11477107 10.1074/jbc.M105831200
Chen S Zack DJ Cloning and characterization of retinal transcription factors, using target site-based methodology Methods Enzymol 2000 316 590 610 10800704
|
15963234
|
PMC1182371
|
CC BY
|
2021-01-04 16:22:25
|
no
|
BMC Mol Biol. 2005 Jun 17; 6:15
|
utf-8
|
BMC Mol Biol
| 2,005 |
10.1186/1471-2199-6-15
|
oa_comm
|
==== Front
BMC Med EducBMC Medical Education1472-6920BioMed Central London 1472-6920-5-251601180010.1186/1472-6920-5-25Research ArticleUndergraduate nursing students' compatibility with the nursing profession Adib-Hajbaghery Mohsen [email protected] Mansur [email protected] Faculty of Nursing, Kashan University of Medical Sciences, Kashan, Iran2005 12 7 2005 5 25 25 11 1 2005 12 7 2005 Copyright © 2005 Adib-Hajbaghery and Dianati; licensee BioMed Central Ltd.2005Adib-Hajbaghery and Dianati; 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 high rate of attrition among nursing students has caused some nursing leaders to think about the necessity of considering students' personality during the process of admission into nursing schools. Due to the lack of studies on Iranian nursing students' personality traits, this study was designed to assess freshmen nursing students' personality characteristics and their compatibility with the demands of the nursing profession.
Methods
A descriptive study was conducted at Tehran and kashan medical universities and one of the branches of Azad University. Convenience sampling was used and 52 freshmen nursing students were assessed using Holland's Vocational Interests Inventory.
Results
From the total participants 63.5% were females and 36.5% were males. Based on the Holland's Vocational Interests Inventory 44% did not have appropriate personality characteristics for the nursing profession. 77% of the nursing students participating in the study reported that they lacked information about nursing.
Conclusion
It seems that personality tests can help to select the best students for nursing schools from those who show good academic capabilities. This would decrease the rate of attrition and could improve the quality of care.
==== Body
Background
Every profession calls for a special level of knowledge, skills and personal characteristics. If the correspondence between the applicants' individual characteristics and their intended profession is not adequately taken into account, their job compatibility will be hampered [1].
Tests are nowadays used for the selection of people to different professions. In Iran, university entrance exams are administered annually to assess students' suitability for admission into different programs and professions. Administration of such tests is apparently based on a number of assumptions. These assumptions include the following:
- Students' test scores indicate their suitability for the intended program or profession.
- Applicants with higher test score are suitable for more important and sensitive positions.
- Choice of a profession or program is the result of the applicant's informed selection.
- Applicants carefully consider their interests, abilities, and characteristics for selecting an appropriate program or profession.
- Since participation in tests and the choice of program are voluntary, it is assumed that the successful applicants will be highly motivated and will provide high quality services once they get into their profession.
Reports, unfortunately, indicate that this is not always the case. Many nursing students are not well motivated; nurses tend to lose their interest in their jobs early; and despite an increase in the number of nurses with graduate and postgraduate qualifications, the quality of nursing care seems to have declined instead of improving [2,3]. Alikhani (2000) reported that the problem seems to be caused by factors such as wrong selection criteria, teaching methods, content of courses, and methods of evaluation [4]. The purpose of test administration is to select the best candidates for a program or profession. This can be achieved when those who are selected have characteristics compatible with the features of their profession [5].
A review of admission procedures for nursing students in different countries shows that these issues have been largely neglected. Dornik and Vidmar reported that in Slovenia only a high school diploma is required for admission into nursing schools [6]. Similarly, Ehrenfeld et al. as well as the American National League for Nursing argue that admission to nursing schools in America is based on testes similar to the ones for other programs, such as aptitude tests or teacher-made achievement tests [3,7]. Similarly, in Iran, university entrance tests measure only the theoretical knowledge and the aptitude of the applicants; whereas, requirements of the nursing profession are much different and broader. This is why researchers have emphasized, in recent years, that applicants selected for nursing should have the appropriate psychological and personal characteristics in addition to their knowledge and aptitude. To account for this, personality tests have also been recommended to complement educational aptitude tests [7-10].
In his investigations of different professions and job conditions, Holland has determined personality features appropriate for the nursing profession. In his view, nursing requires social personality, even though some degrees of artistic and exploratory personality types may also be desirable [1].
Regarding the importance of correct selection criteria for the nursing profession, the present study was carried out to determine freshman nursing students' personality features and to investigate the compatibility of such features with studying and practicing nursing. It is hoped that the results of the study would help promote nursing services and increase job satisfaction in the profession.
Methods
This descriptive study was conducted in the nursing schools of Kashan and Tehran medical sciences universities and one of the branches of Azad University (free university), in Iran. The nursing schools of these three universities were selected to represent different types of universities in Iran. These universities admit applicants with varying scores on admission tests and thus the study sample could represent a wider population of students.
The convenience sampling method was used. There were not any exclusion criteria; so, all of the freshmen nursing students in the first semester of the year 2001 were considered as potential samples in each of the three nursing schools (N = 90, 50 from Tehran medical university, 20 from kashan medical university and also 20 from Azad university). After the approval of the schools' principals or the institutional review boards and with the prior consent of the instructors of each group, the potential participants were given an explanation of the study by the researchers during a scheduled session held during the first two month after admission. The questionnaire was introduced and the participants were instructed on how to complete it. Dominant personality features for each personality type-Realistic, Investigative, Artistic, Social, Enterprising, and Conventional – were also explored for the participants. Students who agreed to participate in the study were asked for written consent and then were given the questionnaire to complete in a private place. 75 students volunteered to complete the questionnaire and they were asked to return them within 24 hours.
Instrument
A two-part questionnaire was used:
A:) the first part consisted of a questionnaire on demographic features which also included an item on the participants self-evaluation of their personality type (among the six personality types of: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional), two open-ended questions on the level of familiarity with the nursing profession and their mentality about it at the time of application, and the purpose of choosing nursing as a field of study.
B): The second part consisted of Holland's Self-Directed Inventory for Career Development. The Self-Directed Search inventory is one of the most widely used interest inventories [11,12]. It is used widely to assess congruence or "fit" between individuals' personality types and their work or educational environments. The SDS was developed by Holland in 1971 and revised in 1977 and 1985. Internal consistency coefficient on the summary scale of the SDS ranged from 0.90 to 0.94. Test retest reliability correlation of the summery scales also ranged from 0.76 to 0.89 [12-14]. We used the farsi version of SDS that was translated and culturally adapted by Hoseinian and Yazdy, 1995 [1]. This scale consists of two sections. The first, the occupation classification section, includes 500 occupations classified according to six personality styles (i.e.: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional). The second, the Self-directed search (SDS), which comprises of six sub-scales including: occupational daydreams, activities, competencies, occupations, self-estimation, and response organization.
In occupational daydreams, the respondent is asked to name his/her occupational wishes in chronological order. Then a list of 66 activities is presented in the activities subscale.
Activities are presented in 6 sections congruent with the 6 personality types. So, 11 activities stand for each personality type and respondent would express his/her interest as "Like" or "Dislike" for those activities he/she would like or dislike to do.
Subsequently-at the Competencies subscale – respondents respond yes or no to a list of 66 activities that they can or cannot do well or competently. Here also 11 activities stand for each personality type. Then a list of 84 occupations is presented in the occupations sub-scale. These occupations are also presented in 6 sections in accordance with the personality types. So, 14 activities stand for each. The respondent would express his/her interest as "yes" or "no" for those occupations he/she would like or dislike. The fifth subscale is for self-estimation where the respondent would compare him/herself with other persons with the same age. The respondent ranks him/herself on a 1–7 scale in his skills and abilities of different types such as mechanical, scientific, artistic, educational, business, administrative, manual, mathematical, musical, and cooperative skills. This part is presented in two sections each including of 6 types of skills or abilities in accordance with the Holland's personality types.
The final subscale is called the response organization subscale. This is the section for analyzing the participant's responses to the different subscales. Here, scores related to each personality type are calculated. For doing this the "like" and "yes" responses the respondent have given to the subscales of occupation, activities and competencies being calculated for each personality type. The resulted score then will be added up to the person's self scoring in the self-estimation subscale. The possible range of score could be between 0–50 for each personality type. Finally the three personality types with the highest scores will chose to make a three-letter code to indicate the most prominent personality characteristics of person. For instance the following scores could represent a person who has essentially a social character, however she has some considerable degree of investigator and artistic characteristics (i.e. social = 43, investigative = 36, artistic = 29, conventional = 20, enterprising = 20, realistic = 18). So the code SIA will chose to refer to such a person. This not only indicates the most prominent personality characteristics of such individual but also it will be used as a criterion for determining the individual's compatibility with an occupation or a filed of study. In this research we used the personality type with the highest score (i.e. social in the above example) as the most prominent personality character for each student and then we used it as the criterion for determining the student's compatibility with the nursing as a profession or a field of study. So, students with social, artistic and/or investigative characters were considered compatible with the nursing profession.
In this study the participants completed the first questionnaires as well as Holland's Self-Directed Inventory (except for the Professional daydreams and the response organization sections). The researchers analyzed the participants' responses as described above. The collected data were manually analyzed, using descriptive statistics.
Responses to open-ended questions were analyzed through content analysis. These responses were read carefully and their main themes were coded and classified.
Results
Of 75 questionnaires delivered to the participants, 52 were returned completed. From the total respondents 63.5% were females and 36.5% were males. The mean age of sample was 22 ± 1.2, ranging from 18 to 25 years.
Graph 1 shows the distribution of personality types reported by the participants as well as their personality types according to the Holland's inventory. As the figure shows, 44% of the subjects evaluated themselves as realistic whereas the questionnaire showed only 25% as realistic. The figure also shows that 45% of the participants did not enjoy the personality types appropriate for the nursing profession.
Respondents had been asked how much they knew about nursing at the time of application. About 23% reported average and high levels of information, whereas, 77% knew little about nursing at the time of application. Most of participants did not have a correct picture of the nursing profession and associated it more with medicine. The respondents' presuppositions about nursing fell into four categories: a) Lack of information and no presupposition (30.8%) b) a picture like medicine (40.4%) c) little knowledge (17.3%) d) complete familiarity (11.5%).
For the purpose of entering the program, students' responses fell in the following categories: 1) just entering the university (21.2%), 2) continuing studies in medicine (21.2%), 3) Finding a job (19.2%), 4) lack of attention in choosing a program (15.4%), 5) Service to mankind (13.4%), and 6) Others' influences on course selection (9.6%).
Discussion
We used Holland's Self-Directed Search (SDS) for assessing Iranian freshmen nursing students' personality characteristics and their compatibility with the demands of the nursing profession. Many studies used SDS for assessing relationship between academic achievement and the students' personality type and confirmed its usefulness for better career planning [15-20]. However, limited publication is available related to using SDS in nursing students.
The present study showed that 45%of the sample lacked the appropriate dominant personality features for nursing. Horn and Holzemer, 1991, were also used SDS questionnaire to compare the personality type of nursing students to women studying engineering. However, they reported that the most of nursing students demonstrated Holland's "social" personality type and engineering students were more "realistic" or "investigative"[21]. However, our finding is similar to the ones reported by Long and Gordon; Fujita et al; Martin et al; and Zolfaghari and Adibi who proposed the use of personality tests for admitting students into nursing programs [10,22-24].
Personality types indicated by the questionnaire did not conform to those found through Holland's self-directed inventory. This maybe partly due to the fact that the subjects were all youngsters who were not personally matured and didn't know their personality well. Marlow has argued that at this stage peoples personality is still being shaped and is, therefore, not stable yet [25].
The results also indicated that most participants did not know the requirements of the profession very well at the time of applying. As Ghazi and Henshaw have claimed, many students enter nursing programs with common sense insights [2]. These people who enter the profession with inadequate knowledge will cease to perform with required standards and experience psychological pressures [5]. Therefore, administrative officials aught to take the necessary steps to help students get adequate information about their planned professions before taking part in the national-wide admission tests. Job councilors believe that successful job selection requires an understanding of individual characteristics, backgrounds, interests and job requirements on the part of both students and selection committees. Otherwise the resulting job-personality incompatibility can lead to poor performance and reduce people's satisfaction and security [5].
The nursing profession and its requirements call for persons with social, artistic and investigative personality types and with characteristics such as patience, tolerance, friendliness, love and sense of cooperation and responsibility [21]. These characteristics are not usually tested in current selection tests for the nursing profession. Ghazi and Henshaw as well as Ehrenfeld et al. have also reported this fact and have attributed to it such factors as lack of motivation during the study period, low achievement, high attrition rates, lack of interest in the job and quitting the job [2,7].
The purpose of the nation-wide university entrance exam in Iran is to evaluate the applicants' readiness for different educational programs and their corresponding professions and to select the best candidates for them. Yet, this annual exam has a fixed format for all disciplines and dose not include any components on the evaluation of the applicants' psychological traits and required personality features. Based on the findings of this study, it is recommended that personality and interest tests be used to supplement the existing academic and aptitude tests of admission into nursing programs. In other words applicants can be short-listed based on their academic capabilities. Once twice the number of eligible applicants for each nursing school has been short listed, interest and personality tests can be used to select the best applicants. This can greatly reduce students' attrition rates. It can also increase job satisfaction and would guarantee the quality of nursing care provided by the future nurses.
This was the first step in the investigation of job-personality compatibility among a limited sample of Iranian nursing students. So the results may not be necessarily be generalizable to all nursing students. The replication of this study on a larger sample is recommended.
Conclusion
The present study showed that many nursing students lacked the appropriate personality features for nursing. The results also indicated that most participants did not know the requirements of the profession at the time of applying. Therefore, administrative officials aught to take the necessary steps to help students get adequate information related to their planned professions before taking part in admission test. It seems that personality tests can help to select the best students for nursing schools from those who show good academic capabilities. This would decrease the rate of attrition and could improve the quality of care.
List of abbreviations used
Self-Directed Search (SDS).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MAH: Initiation and design of the research, collection and analysis of the data and writing the paper.
MD: helping in data gathering and revision of draft paper.
Figure 1 The distribution of personality types reported by the participants as well as their personality types according to the Holland's inventory.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Holland J Hoseinian S and yazdy M Tehran Making vocational choices: A theory of vocational personalities and work environments Farsi translation 1995
Ghazi F Henshaw L How to keep student nurses motivated Nursing Standard 1998 13 43 48 9923363
National League for nursing Official guide to undergraduate and graduate nursing schools 1999 Boston: Jones & Bartlet publishers
Aalikhani S How can motivate the nursing students? Journal of nursing school of Iranian army 2001 9 1 8
Ardebili J Occupational counseling, how to obtain and use the occupational information 1997 Tehran: Virayesh and Fahim publication
Dornik E Vidmar G Impact of nursing education in Slovenia on nurses' publishing in their professional journal Stud Health Technol Inform 2003 95 794 9 14664085
Ehrenfeld M Rotenberg A Sharon R Bergman R Reasons for student attrition on nursing courses Nursing Standard 1997 11 34 38 9096522
Stuart CC Assessment, supervision and support in clinical practice 2003 Edinburgh Churchill living stone
Sanjary K Rreorganizing the job in the human resourses management Tadbir 2002 118
Zolfaghari B Adibi P Criterion for admission in medicine Supplement of the Journal of Tehran Faculty of Medicine, Abstract book of the fourth congress on medical education 2001
Spokane AR Holland JL The Self-Directed Search: A family of self-guided career interventions Journal of Career Assessment 1995 3 373 390
Lumsden JA Sampson JP Reardon RC Lenz JG A Comparison Study of the Paper, Personal Computer (PC), and Internet Versions of Holland's Self-Directed Search: Technical Report No. 30 The Center for the Study of Technology in Counseling and Career Development 2002
Eberly C EDG 5720 Test evaluation outline Eastern luiziana university: college students affair
Holland JL Fritzsche BA Powell AB The Self-Directed Search technical manual 1994 Odesssa, FL: Psychological Assessment Resources, Inc
Henry P Relationship between academic achievement and measured career interest: examination of Holland's theory Psychol Rep 1989 64 35 40 2928452
Seibert SE Crant JM Kraimer ML Proactive personality and career success J Appl Psychol 1999 84 416 27 10380421 10.1037//0021-9010.84.3.416
Reardon R Bullock E Holland's Theory and Implications for Academic Advising and Career Counseling:Technical Report 38 2004 The Florida State University
Petrides KV McManus IC Mapping medical careers: Questionnaire assessment of career preferences in medical school applicants and final-year students BMC Medical Education 2004 4 18 15461786 10.1186/1472-6920-4-18
Bore M Munro D Kerridge I Powis D Selection of medical students according to their moral orientation Med Educ 2005 39 266 75 15733162 10.1111/j.1365-2929.2005.02088.x
Lumsden MA Bore M Millar K Jack R Powis D Assessment of personal qualities in relation to admission to medical school Med Educ 2005 39 258 65 15733161 10.1111/j.1365-2929.2005.02087.x
Horn H Holzemer WL Characteristics of Israeli women studying nursing compared to women studying education and engineering J Nurs Educ 1991 30 411 8 1663543
Long NR Gordon CJ Variables important in the selection of New Zealand nurses: implementation and evaluation of a multivariate selection technique Int J Nurs Stud 1981 18 227 235 6914324 10.1016/0020-7489(81)90036-5
Fujita M Uchino Y Ohara H Study on personality characteristics of nursing students (1). A comparison of student groups of various nursing courses at the time of school admission Kango Kyoiku 1987 28 167 171 3647163
Martin MJ Vila J Capellas R Studies of the personality of nursing students Rev Enferm 2000 23 643 646 11111680
Marlow DR Redding BA Textbook of pediatric nursing 1988 6 Philadelphia: WB Saunders
|
16011800
|
PMC1182372
|
CC BY
|
2021-01-04 16:30:56
|
no
|
BMC Med Educ. 2005 Jul 12; 5:25
|
utf-8
|
BMC Med Educ
| 2,005 |
10.1186/1472-6920-5-25
|
oa_comm
|
==== Front
BMC Med EthicsBMC Medical Ethics1472-6939BioMed Central London 1472-6939-6-71599626810.1186/1472-6939-6-7Research ArticleAction ethical dilemmas in surgery: an interview study of practicing surgeons Torjuul Kirsti [email protected] Ann [email protected]ørlie Venke [email protected] Sør-Trøndelag University College, Faculty of Nursing, Trondheim, Norway2 Centre for Medical Ethics, University of Oslo, Norway3 Institute of Nursing and Health Sciences, Faculty of Medicine, University of Oslo, Norway2005 4 7 2005 6 7 7 11 4 2005 4 7 2005 Copyright © 2005 Torjuul et al; licensee BioMed Central Ltd.2005Torjuul et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The aim of this study was to describe the kinds of ethical dilemmas surgeons face during practice.
Methods
Five male and five female surgeons at a University hospital in Norway were interviewed as part of a comprehensive investigation into the narratives of physicians and nurses about ethically difficult situations in surgical units. The transcribed interview texts were subjected to a phenomenological-hermeneutic interpretation.
Results
No gender differences were found in the kinds of ethical dilemmas identified among male and female surgeons. The main finding was that surgeons experienced ethical dilemmas in deciding the right treatment in different situations. The dilemmas included starting or withholding treatment, continuing or withdrawing treatment, overtreatment, respecting the patients and meeting patients' expectations. The main focus in the narratives was on ethical dilemmas concerning the patients' well-being, treatment and care. The surgeons narrated about whether they should act according to their own convictions or according to the opinions of principal colleagues or colleagues from other departments. Handling incompetent colleagues was also seen as an ethical dilemma. Prioritization of limited resources and following social laws and regulations represented ethical dilemmas when they contradicted what the surgeons considered was in the patients' best interests.
Conclusion
The surgeons seemed confident in their professional role although the many ethical dilemmas they experienced in trying to meet the expectations of patients, colleagues and society also made them professionally and personally vulnerable.
==== Body
Background
Surgeons are responsible for all activities related to patients' treatment and care in surgical units and it is therefore important for them to act in a the best and correct way towards patients, relatives, and colleagues. Studies have shown, however, that physicians often are in doubt about the best and correct actions to take for the patients in specific situations [1-3]. This question is not only a medical one, but can be understood in both action and relational ethical perspectives. A relational ethical perspective means reflecting on the challenges we encounter in our relationships with others and how to best fulfil our social roles and obligations – as a person, a surgeon, and a colleague. It tries to answer questions such as "How can I adequately meet the challenges that confront me in the relationships in which I am involved in this situation?" [4]. Qualities that make a person a good physician are not only individual traits but they are characteristics of the relationships [5].
Surgeons' responsibility and the imperative to save life often lead to their focusing on an action ethical perspective in their reasoning [1,4]. Reasoning according to this perspective means explaining our choices of actions in situations in which we are not sure what the right thing to do is. In this perspective, ethics often centers on difficult ethical dilemmas and decision-making, justifying our actions, and gives answers to the questions: What should I (we) do? Did I (we) do the right thing in this situation? [4]. Ethical dilemmas occur when physicians have to choose between at least two alternative and equally difficult courses of actions. Because neither of the alternatives have positive outcomes, physicians have to choose between two evils [4]. Ethical dilemmas can also be understood as conflicts between different courses of actions that result from following general and mutually exclusive ethical principles in medicine [6].
Action and relational ethical perspectives are not mutually exclusive, but rather complementary, as surgeons have a dual responsibility for their actions in specific situations as well as their way of being in their relationships [4,7]. Being a good surgeon presupposes both professional competencies based on scientific and clinical knowledge and skills, and being present and showing respect and compassion for patients [4,5,8].
Surgeons today are confronted with more ethical dilemmas than before due to the growth in scientific knowledge, an increase in the availability and efficiency of medical technology, a more equal relationship between patients and surgeons, and changes in the organizational arrangement and financing of the health care system [3,9,10]. The growth in scientific knowledge and technology has given surgeons new and better diagnostic equipment and treatment opportunities. The frequency of surgical treatment is expanding and surgeons are able to successfully operate older patients and patients with multiple and more serious diseases than before [3,11]. New treatment opportunities have increased the number of possible ethical dilemmas in surgical practice and put heavy pressure on the individual surgeon who has a personal responsibility for all decisions concerning the patients' treatment and care. Ethical considerations cannot be avoided when surgeons have to choose between what ought to be done among the many courses of action that are available for patients in particular situations [10]. Physicians and surgeons are said to experience a decrease in their autonomy at the same time because more external factors and stakeholders are influencing their decisions concerning patients' treatment [3,10,12].
The development of better anaesthetic methods and less invasive surgical techniques has made it possible to increase the frequency of surgery and perform rather extensive operations in out-patient facilities. The length of patients' stays in hospitals has also been reduced. The growth of new diagnostic and therapeutic opportunities in modern medicine have, in turn, created great demands on resources and made medicine a high cost endeavor. Economic factors are said to increasingly determine the patterns of clinical work, either directly or indirectly, and physicians frequently experience the ethical dilemma of allocating limited resources [13]. It is argued that surgeons have been put under heavy political and administrative pressure to reduce costs to a greater extent than other medical specialists, and they may experience dilemmas between promoting the patients' health interests and the economic interest of the hospital and of society [3,14].
Patients today are said to expect more from medical diagnostics and treatment than before, expectations that may be greater than physicians are able to provide [15]. They almost take for granted that everything can be treated and cured and are more willing to sue physicians for suboptimal results of treatment or deviation from perfect performance [9,16]. Surgeons often experience high expectations from patients, patients' relatives, colleagues and the media, and can even feel pressure to perform innovative and undocumented surgical operations [3,11,17-19]. The fear of being sued can lead to defensive medicine and reduced trust between physicians and their patients [20]. Health care is increasingly perceived as a commodity [21] and consumerism may lead physicians to spend more time attending to patients' wants than before, and this has made the physicians work more complicated [13,16,22,23].
Few empirical studies have been found that explore the ethical dilemmas in surgery from the surgeons' point of view and their experiences of ethically difficult situations in during practice. The present study is part of a comprehensive investigation into the narratives of physicians and nurses about ethically difficult situations and the meaning of being in such situations in surgical units.
The results of this interview study with male and female surgeons will be presented in two papers. The present study describes the kinds of ethical dilemmas surgeons face in their practice. The other paper describes the surgeons' experiences of being in ethically difficult situations in their relationships with patients, relatives and colleagues [24]. The results from the interviews with the registered nurses (RNs) working in surgical units are in progress and will be addressed in a third paper.
Methods
Participants and setting
Five male and five female surgeons working in surgical units at a university hospital in Norway participated in the study. All were experienced and had been working in health care from 9 to 31 years (median = 21.5), and in surgery between 5 to 21 years (median = 13). The surgeons worked full time and were on duty when the interviews were conducted. No individual characteristics will be disclosed in order to guarantee anonymity. The surgeons gave their informed consent to participate in the study, which was approved by the 5th Regional Ethics Committee in Norway.
Data collection
Interviews
The interviews were conducted by the first author and lasted from 35 to 75 minutes (median = 55). They were tape recorded and subsequently transcribed verbatim. The interviewees were asked to tell about one or more ethically difficult care situations that they had experienced in their work as surgeons. What constituted an ethically difficult situation was not defined, allowing the interviewees to determine what they considered ethically difficult themselves. The aim of the interviews was to obtain as many rich narratives as possible without interrupting the surgeons' narrative flow and reflection. If the surgeons did not spontaneously reflect on the events they talked about, their reflections were sought. Questions were asked when the interviewer wanted the interviewees to elaborate on their stories or had difficulty understanding the narration. These questions referred to the interviewees' thoughts, feelings, and actions [25]. Field notes were taken during the interview as aids to the interviewer's memory and in order to understand the interview text in relation to its context, e.g. arrangements and interruptions. Nonverbal communications that seemed relevant were also noted, such as laughter and long pauses. The transcribed text was compared with the field notes and adjusted if necessary.
Data analysis
Interpretations
The method of interpretation used was inspired by the French philosopher Paul Ricoeur's phenomenological hermeneutics [26], and developed at the University of Tromsø (Norway) and Umeå University (Sweden) and has previously been used by Lindseth et al., [27] Udén et al., [1,28] Søderberg and Norberg., [29] Sørlie et al., [18,30-33] and Nordam et al. [34]. This method is useful to elucidate the narratives of people's experiences. The method of interpretation proceeds through three phases, which constitute a dialectical movement between the whole and the parts of the text and between understanding and explanation [26].
Each interview was regarded as a text. First, a naïve reading was made of all the transcribed interviews as a whole to gain a first impression of the surgeons' experiences of ethical dilemmas in their clinical work. The repeated naïve reading was made as open-minded as possible, without any deliberate analysis of the text. The naïve reading shows the direction the structural analysis may take. Second, a structural analysis was performed in order to validate or refute the initial understanding obtained from the naïve reading and to explain what the text was saying. The interviews were divided into meaningful parts and patterns, i.e. one sentence, parts of a sentence, or a whole paragraph with a related meaning content. The meaning units were condensed and discussed among all the authors, and themes and subthemes were identified, and presented in 'Results'. Third, a comprehensive understanding was developed, taking into account the authors' pre-understanding, the naïve reading, and the structural analysis (results). The text was read as a whole and interpreted in relation to relevant theories of ethics and results from previous investigations into ethical dilemmas in surgery [35]. The comprehensive understanding is presented under the heading 'Discussion'.
The analysis was conducted by all the authors and the interpretative agreement was considered satisfactory and to be the most useful understanding of the surgeons' experiences of ethical dilemmas in their clinical work. The authors' interpretation was not shared or validated with the surgeons. A kind of validation is accomplished by the structural analysis as the objective part of the interpretation process [35].
Results
Several readings of the interview texts revealed that the surgeons narrated about ethical dilemmas they face in their practice. They also told about the ethical challenges that confront them in their relationships with patients and colleagues and their experiences of living with these challenges. The surgeons admitted that they were faced with several ethical dilemmas at the same time, but concluded that consideration for their patients was the main single factor in resolving these dilemmas. No gender differences were found in the kinds of ethical dilemmas identified among male and female surgeons. The results showed that each surgeon created many and detailed narratives. When the surgeons were asked to tell about the ethically difficult situations that they had experienced, they did not differentiate between action and relational perspectives in their narration. This is an analytical distinction made by the authors in order to structure the results. The authors therefore decided to separate the presentation of the results in two papers: one paper about the surgeons' experiences of being in ethically difficult situations from a relational ethical perspective [24] and the other paper about the surgeons' experiences of ethically difficult situations from an action ethical perspective according to the theory presented in the background [4]. This paper presents the kinds of ethical dilemmas surgeons are faced with during practice.
The themes and the subthemes from the structural analysis are shown in Table 1 and presented in the text below. Direct quotations from the interviewers are included to illuminate the results.
Table 1 Overview of themes and subthemes that emerged from the structural analysis of interviews with the surgeons about ethical dilemmas.
Themes Subthemes
Treatment Starting or withholding treatment
Continuing or withdrawing treatment
Overtreatment
Respecting the patients
Meeting patients expectations
Resolving differences in opinions Superior colleagues
Colleagues from other departments
Incompetent colleagues
Society Local limited resources
Laws and regulations
Global limited resources
Treatment
The interviews revealed that the surgeons experienced ethical dilemmas on almost a daily or weekly basis. They emphasized that decision making is strongly dependent on the context as each situation and patient is unique. Making existential decisions, like withholding or withdrawing treatment where patients' lives and quality of lives are at stake, were the main ethical dilemmas narrated by the surgeons.
Starting or withholding treatment
The surgeons related that they experience the dilemma of staring and withholding treatment especially when working with incurable cancer patients, very old patients with additional physical and mental diseases, and with emergency and trauma patients. In situations when the disease cannot be cured, palliative operations can cause complications, prolonged suffering and an earlier death. A newborn child with serious congenital malformations or deformities may also raise the dilemma of starting or withholding treatment. Although the malformations can be corrected so that the child's life is saved, the result may be a life with serious handicaps and a reduced quality of life for both the child and its parents.
The question of starting or withholding treatment was experienced as most dramatic when patients were admitted with life-threatening conditions like a ruptured aortic aneurysm which will lead to death without an operation, but where the outcome of surgery and the patients' quality of life afterwards are impossible to predict. "You know that if you do not do anything immediately, the patient will die. If you start you have a chance of making it. You give it a try". The surgeons said that it is more difficult to withhold treatment the younger the patients are, mainly because the emotional feelings surrounding the decisions are experienced as more difficult. Wondering about starting or withholding treatment means that the surgeons are uncertain about the outcome of surgery. The surgeons remembered cases of unexpected recovery and success against all odds. "We experience patients where everything looks hopeless and we go on doubting for a while. And then they recover from a seemingly hopeless condition and live well, even with severe complications. We have seen people with paralyzed legs, a stoma and permanent kidney failure, who are nevertheless content for having survived. This shows how difficult it is to assess a person's quality of life and how they will look at things afterwards".
The surgeons said that if there is the slightest possibility of survival, they usually try to operate in order to give the patients a chance, especially if the patients are conscious and strongly want the operation. They said that most patients are willing to undergo any treatment although their prognoses are poor. Withholding treatment was felt like "destroying all means of retreat".
Continuing or withdrawing treatment
The decision whether to withdraw active medical treatment other than palliative, or to continue treatment, was experienced as equally difficult by the surgeons. They said that it is more difficult to withdraw treatment the younger the patients are. Patients' expectations are important when the surgeons make their decisions. "If the patients are able to express that they do not want any more [treatment], then the decision is simple. But if they are unable to express themselves, it is very difficult to end an ongoing treatment". The surgeons said that they always consult the patients' relatives before a decision to withdraw treatment is made, but stressed that it is not the relatives that should make the final decision, but the physicians.
Overtreatment
Finding the right level of treatment was experienced as an ethical dilemma for severely ill and very old patients and patients with non-curable cancer. The surgeons said that they attempt to give every patient a chance to survive, and avoid the risk of termination of treatment too early may result in overtreatment, and prolonging the patients' pain and suffering instead of alleviating them. The surgeons said that overtreatment occurred because it is easier to act than refrain from action, and that it was difficult to refuse the patients' wishes.
Respecting patients
The surgeons emphasized that the patients' right to decide their own treatment created ethical dilemmas. Not being allowed by the patients to operate where there are possibilities of full recovery or substantial improvement was a difficult dilemma for the surgeons, e.g. patients with cancer. Members of Jehovah's Witnesses who refuse blood transfusion create a dilemma if the operation entails a high risk of life-threatening blood loss. "In a way they force me to participate in a situation where I can risk losing a patient who I know I could easily have saved". Another dilemma arose when parents ask for ritual circumcision of baby boys.
Respecting patients' right to decide their own treatment meant that the surgeons felt responsible for informing the patients about their diseases, the risks and benefits of surgery and other medical treatments available, and for giving advice to help patients make the 'right' decision. "I cannot remove any parts of a person's body unless they are accountable; unless they themselves comprehend that it is necessary". Deciding the right amount of information to give the patients was not always easy, according to the surgeons. Assessing the accurate risk and benefits of an operation was experienced as difficult in each individual case. Also presenting the risks and benefits of undergoing or not undergoing treatment in a neutral and unbiased way could also be difficult. The surgeons said that patients with unquestioning faith in the surgeon may give their consent to surgery without having understood the consequences.
It was experienced as difficult to assess patients' understanding of the surgeons' information about diagnostics and treatment options. "It can sometimes be difficult to give enough information. Even if you spend a lot of time providing information, it is still not certain that the patients understand". The surgeons said it was difficult to perform operations when they were in doubt about patients' capacity to make informed decisions, e.g. dementia sufferers, or psychiatric patients with reduced mental capacity. "Patients suffering from dementia will experience any operation as an encroachment because they cannot take part in a discussion".
The surgeons complained about having no separate room where they could deliver sensitive information and bad news to patients and relatives. To use the corridor or other places where they could be overheard was experienced as not respecting the patients. "I think that the greatest problem in this hospital is that they [the administrators] do not comprehend that this is something we do almost every week with patients and relatives".
Patients seeking alternative treatment, medicines and drugs where the effects are uncertain and impossible to control was experienced as a dilemma for the surgeons. Although the surgeons respected the patients' right to decide their own treatment, they feared that some alternative treatment may harm the patients' standardized medical treatment. In some cases, patients with incurable cancer even have alternative operations in private clinics. "The patients may use all their savings paying for the treatment and then it turns out to be of no benefit to them". The surgeons said that their patients often seek advice about alternative medicine and they experienced a dilemma when they felt compelled to advice against alternative medicine as they feared that the patients might lose hope.
Meeting patients' expectations
The surgeons revealed that they are willing to go to considerable lengths in order to fulfil their patients' expectations. They experienced that many patients have "unrealistic" expectations of surgery and do not accept that surgery cannot do everything to relieve their ailments and impairments. "Sometimes they have pains or other torments that they expect to disappear after the operation. And they cannot see the connection between their torments and what they can do themselves to get better".
The surgeons told that patients have a strong belief in the health care system and expect, and even demand, to have the best available diagnostics and treatments. The surgeons emphasized that they have to trust the patient and believe in their complaints. If, however, they found that the clinical picture did not correspond to the patients' complaints, then they experienced the dilemma of how far they should go to examine the patients to rule out possible serious diseases. They said this is also a question of the use of limited resources, as extensive examinations of one patient will reduce resources available for other patients.
The surgeons emphasized that when patients have expectations they cannot fulfil, they tried to meet the patients part-way by providing information and education. If the patients' demands exceeded what surgeons found medically advisable and justifiable, they negotiated with the patients to find workable solutions. They referred patients to other physicians in extreme cases. Refusing patients' demands for narcotics, medical service, sick leave, social insurance etc. were experienced as dilemmas by surgeons who are trained to believe in the patient's complaints. Some surgeons said they have been exposed to physical and verbal threats by patients.
Resolving differences in opinions
Superior colleagues
The interviews revealed that even though surgeons had different opinions, they usually reached agreement about the most suitable treatment for the patients. The surgeons experienced a dilemma whether they should act against or give in to the chief surgeons. "Sometimes you are asked to do something you do not think is right. You are for instance asked to do things in a particular way, and then you think that it is difficult to go against the decisions of those who are more experienced. It does not need to be anything strictly right or wrong, but minor things, like you would have chosen another type of drug". The surgeons felt they had to make decisions on their own on evening and night duty, because they found it difficult to wake the chief surgeon for advice. "Perhaps I make more decisions on my own then, because I feel that I have to sort it out on my own".
Colleagues from other departments
The surgeons experienced a dilemma if specialists from other departments ordered operations which the surgeons did not approve of themselves, for instance the feeding of severely ill, old or terminal patients with PEG-tubes. If the patients were not able to express what they want, the surgeons found it difficult to go against the specialists issuing the orders.
Incompetent colleagues
Correcting a colleague who is not competent or who has unacceptable behavior in relation to patients, was felt as a personal and a professional dilemma for the surgeons, who said that they felt responsible for both their patients and their colleagues. They said that the standard of surgical performance is high and it is difficult to determine fairly and objectively whether a colleague is performing adequately or not. "On the one hand you realize that surgery is complicated, and you cannot blame people for a single decision or a single action that is wrong". The chief surgeons in particular said they felt responsible towards the patients and society, to stop incompetent colleagues from practicing surgery.
Society
Local limited resources
The interviews revealed that surgeons had to consider prioritization of limited resources in their decision-making. The amount of resources spent on one patient affected other patients on the waiting lists for operations. The dilemma of making the right prioritization was always present, according to the surgeons. They also mentioned that it was an ethical dilemma whether economical considerations should influence decisions about withholding or withdrawing medical treatment and if so, how much. The surgeons had to prioritize between patients when all the beds in the intensive care unit were occupied. Intensive treatment of seriously ill patients implies spending a lot of resources on one patient over time. They said that by prolonging the intensive treatment of very old patients, other patients in need may not receive treatment, operations are delayed and patients waiting for operations risk having complications because of the delay. Time shortage may also harm the quality of medical care.
Long waiting lists for surgical treatment are a problem in the health care system. The surgeons were concerned that some patients receive treatment too late because they are omitted from the waiting lists for some reason or other, mainly because of the breakdown of routines.
Laws and regulations
The surgeons were upset because they are legally responsible for informing the patients about 'DNR' orders '(do not resuscitate') and document it in the patients' medical records. Although they stressed that every patient has a right to be given sufficient information about medical examination and treatment according to legalisation on patients' rights, they felt that this regulation was inhuman and removed the patients' spark of hope that enable them to fight the disease. They therefore said that they placed consideration for patients before laws and regulations. The surgeons also mentioned the regulation that parents have no economic right to be with their sick children unless they are dying, which they thought was preposterous, and they therefore tried to bend the rules.
In the recent years the surgeons have been ordered to document their activities in a code system based on diagnosis-related groups (DRGs). The patients' diagnoses and the sequences in which they are coded determine the amount of government reimbursement the hospital receives. The surgeons felt pressure to report the patients' diagnoses in such a way that reimbursement to the hospital increased, but stressed that they document in agreement with their assessment of the patients' ailments. They feared that the DRG system would influence their way of working in a negative way.
Global limited resources
The surgeons expressed a global perspective on their work, realizing that they have more available resources at their disposal than physicians in other parts of the world, and that the global distribution of health care resources is unfair. "For instance, in our part of the world we are using enormous resources in trying to help an 85-year-old to have one or two more years to live, while in other parts of the world they are striving to keep children alive". From a global perspective, the surgeons experienced a dilemma in that people in our part of the world do no realize that both local and global resources are limited, but insist having the best treatment whatever the cost. This dilemma was especially pronounced when they experienced increasing expectations and pressure from patients and society to do even more despite limited and unequally distributed global resources. However, they found it gratifying having enough resources "to do a conscientious and decent job" and "to be of use to society" under the present circumstances. They appreciated the recognition they received from patients and relatives.
Discussion
The surgeons in this study narrated about the many and complicated ethical dilemmas connected to the treatment of patients, differences of opinions between colleagues and prioritization of limited local and global resources. The main focus in the narratives was on ethical dilemmas concerning the patients' treatment and quality of life, patients' right to decide and meeting patients' expectations. Trying to find the ethically correct action that benefited their patients most pervaded the surgeons' narratives. Consideration for the patients was salient even when they told about resolving differences in opinions with colleagues and conforming to the laws and regulations in society. It seems that the ethical dilemmas the surgeons experienced originated mainly out of regard for their patients, and this was equally important in resolving the dilemmas, for instance by bending social laws and regulations in the patients' and relatives' favor.
The surgeons' ethical awareness is not highlighted in previous research. On the contrary, it is suggested that surgeons until recently have taken less interest in ethics than their colleagues in other medical areas [8,10,36-39]. The surgeons in this study however, reflected over and related many different ethical dilemmas in their everyday practice. They are aware of the many ethical dilemmas in their practice and are dedicated to the best possible outcome and care for each patient using their scientific knowledge and practical wisdom. Thus it seems reasonable to suggest that they take both their professional and ethical responsibility seriously. To be morally committed to patients is to be aware of the responsibility inherent in the patient-physician relationship. This ethical responsibility arises from the ethical demands with which they are faced in their meetings with patients. It is necessary for every person to act in a way that is determined by our understanding of a situation and related to the particular risk to which we are subjected, according to the Danish philosopher Løgstrup [40]. Our understanding of the situation presents us with a choice which in turn demands that we act and give of ourselves and thereby contribute to the realization of one or more other individuals' possibilities in life, which is at our disposal in the given situation. The physicians' challenge is to be open to the ethical appeal without being destroyed by an overwhelming responsibility [41].
The surgeons in this study showed a strong dedication to their patients' well-being, treatment and care in their relationships with patients [24] as well as when they face the ethical dilemmas in clinical practice. Other studies have found that physicians in ethically difficult situations were preoccupied with making decisions based on factual, scientific knowledge and that this knowledge was considered essential in order to give the patients qualitatively good treatment [1,28,30]. In this study, however, the surgeons did not mention scientific knowledge as a resource in resolving ethical dilemmas. They seemed to be preoccupied with performing surgery in a way that led to something beneficial for their patients. In the other paper based on this study we have described the surgeons' involvement in their patients' destinies and the way this affected their own lives [24]. This study seems to contradict the myth of surgeons as distant and indifferent technicians.
Patients expect to benefit from surgical care. They trust surgeons to exercise their knowledge and skills to the best of their ability, and assume that they will take all reasonable steps to ensure a favorable outcome. In this respect the surgeons and patients and relatives interests in most instances seems to correspond. Physicians usually want to respond to patients' expectations by taking action [42]. The surgeons in this study said they felt obliged to exert themselves to help patients for whom they are responsible and thereby meet their expectations. Giving the patient a chance means that the surgeons take personal and professional risks because they can never know the outcome of their decisions, thus the risk of overtreatment [18,24,27,32,43]. The surgeons deliberations can be understood in an Aristotelian perspective as an expression of clinical phronesis; a search for the middle course of action, which lies somewhere between the excess involved in "doing everything possible" or "doing nothing at all" [44,45]. The role of the physician is said to require both the ability and courage to act in uncertain and high-risk situations where the outcomes may be marginal [46]. It may seem that both physicians and patients want to force medicine to the edge of what medicine can accomplish and patients can endure.
Physicians assume personal responsibility for the way they manage their patients and do so; on the basis of their own clinical experience [42]. The role of a surgeon in our culture is filled with high expectations. Meeting patients' and relatives' expectations was experienced as a dilemma for the surgeons because these expectations in many instances were experienced as unrealistic and unlimited. Many other authors have found that physicians working in different parts of the health care system experience greater expectations from patients, relatives and the media than they can fulfil [3,11,15,18,20,23,27,47-49]. Pahuus [50] writes that when our expectations are without limits and countless, we tend to assume the object of our expectation as being perfect. Unlike the patients, the surgeons also have knowledge and experiences of the limitation of surgery [24]. Furthermore, it belongs to the surgeons' ethical responsibility to restrict their intervention to the cases where they have reasonable prospects to succeed [41]. This ethical dilemma facing surgeons will probably not decrease in the future due to the rising cost of modern medicine. As the surgeons' knowledge and skills exceed the available resources, they will experience the ethical dilemma of distributing health care to a greater degree [3,10,17].
The physician's role is said to have changed from paternalism to respecting patients' autonomy [15,22]. Several studies have found that physicians emphasize the importance of respecting the patients and acknowledging their autonomy and authority to decide their own treatment and judge their own quality of life [15,32,47,49]. Falkum & Førde [22] found in their study that surgeons had the lowest scores for valuing patients' autonomy, and tended to decide for their patients without wasting time on information and dialogue. In this study, the surgeons emphasized the importance of showing respect for patients' rights to decide to choose to undergo surgical treatment or not, and the importance of providing adequate information in a way that the patients understood. Studies show conflicting results as to whether the patients' wish to participate in decisions concerning their own treatments or not. Some studies seem to confirm the experiences of the surgeons in this study; that patients want to be well informed, they want more information then they usually get, but often prefer that the final decisions are made by physicians [15,51].
The surgeons in this study seemed to be content in performing their profession and of being in a position where they are able to cure people and alleviate suffering. They appreciated the recognition they received from patients and relatives. This recognition is based on their ability to act in a correct way as well as their attitude in their relationships with patients. It helped the surgeons that they remembered cases were they succeed against all odds although they found it difficult to withhold treatment with uncertain outcome. It is reasonable to suggest that the recognition they get from patients and especially from patients, who survive against all odds, and their relatives, gives them a social confirmation which is important for them [24,33]. This confirmation tells them that they are successful and proficient surgeons.
The surgeons felt personal responsibility for the treatment and care of their patients. They said that this responsibility is sometimes experienced as a heavy burden. According to Henriksen and Vetlesen [52] being responsible for another person's well being ought to be felt as a heavy burden in order to count as moral responsibility. In this study, the surgeons seem to accept the professional and moral challenges in their work in spite of their experiences with many different ethical dilemmas. They also seem content to be 'of use' both for their patients and not least, for society.
Everyone is vulnerable and it is reasonable to suggest that surgeons are especially vulnerable in their professional role as it often concerns questions about life and death, and issues that affect the patients' quality of life. Many authors have found that physicians accept their own vulnerability and fallibility when they are faced with difficult ethical dilemmas [15,26,28,33]. We found that the surgeons in this study seemed content to be of use both for their patients and for society. Thus, the gratifying aspects of their profession seemed to outweigh their vulnerability.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KT participated in the design of the study, carried out the interviews, participated in the analysis and completed the manuscript.
AN read the interviews, participated in the analysis and helped to draft the manuscript.
VS participated in the design of the study, read the interviews, participated in the analysis and helped to draft the manuscript.
All authors have read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by Sør-Trøndelag University College, Faculty of Nursing, Norway. This study is part of a larger ethical project which includes studies from different medical specialties. The authors are grateful to the participants in the study and to Jill Almvang and Stewart Clark for revising the English.
==== Refs
Udén G Norberg A Lindseth A Marhaug V Ethical reasoning in nurses' and physicians' stories about care episodes J Adv Nurs 1992 9 1028 1034
LaSalle DL Ethics in research and surgical practice Am J Surg 1997 6 589 591
McCullough LB Jones JW Brody BA (Eds) Surgical Ethics 1998 New York, Oxford: Oxford University Press
Lindseth A Udén G The role of caring in nursing ethics Quality development in nursing care From practice to science 1992 Linköping: University Health Service Studies 97 106
MacIntyre A After Virtue – A study in moral theory 2001 2 Cambridge, Massachusetts and London: Harvard University Press
Beauchamp TL Childress JF Principles of biomedical ethics 2001 5 Oxford: Oxford University Press
Nerheim H The ethical foundational experience From understanding rules to experiential knowledge 1991 Oslo: Universitetsforlaget AS (in Norwegian)
Peterson L Human values in the care of the surgical patient Arch Surg 2000 1 46 50 10636346 10.1001/archsurg.135.1.46
Hanlon CR Ethics in surgery J Am Coll Surg 1998 1 41 49 9449599 10.1016/S1072-7515(97)00157-9
Nandi PL Ethical aspects of clinical practice Arch Surg 2000 135 22 25 10636342 10.1001/archsurg.135.1.22
McKneally MF Ethical problems in surgery: Innovation leading to unforeseen complication World J Surg 1999 8 786 788 10415203 10.1007/s002689900580
Rothman DJ Strangers at the bedside 1991 USA: BasicBooks, HaperCollins publishers
Armstrong D Clinical autonomy, individual and collective: the problem of changing doctors behaviour Soc Sci Med 2002 10 1771 1777 12383461 10.1016/S0277-9536(01)00309-4
Axelrod D Goold SD Maintaining trust in the surgeon-patient relationship. Challenges for the new millennium Arch Surg 2000 1 55 61 10636348 10.1001/archsurg.135.1.55
Charles C Gafni A Whelan T Decision-making in the physician-patient encounter: revisiting the shared treatment decision-making model Soc Sci Med 1999 5 651 661 10452420 10.1016/S0277-9536(99)00145-8
Edwards N Kornacki MJ Silversin J Unhappy doctors: what are the causes and what can be done? BMJ 2002 7341 835 838 11934779 10.1136/bmj.324.7341.835
Moore FD Ethical problems special to surgery: surgical teaching, surgical innovation, and the surgeon in managed care Arch Sur 2000 1 14 16 10.1001/archsurg.135.1.14
Sørlie V Lindseth A Udén G Norberg A Women physicians' narratives about being in ethically difficult care situations in paediatrics Nurs Ethics 2000 1 47 62 10.1191/096973300676204891
Jones R Declining altruism in medicine BMJ 2002 7338 624 625 11895808 10.1136/bmj.324.7338.624
Hendrick HL Nelson RM Handling ethical conflicts in the clinical settings Semin Pediatr Surg 2001 4 192 197 10.1053/spsu.2001.26840
Arnetz BB Psychosocial challenges facing physicians of today Soc Sci Med 2001 2 203 213 11144776 10.1016/S0277-9536(00)00220-3
Falkum E Førde R Paternalism, patient autonomy, and moral deliberation in the physician – patient relationship Soc Sci Med 2000 2 239 248
Kälvemark S Höglund AT Hansson MG Westerholm P Arnetz B Living with conflicts – ethical dilemmas and moral distress in the health care system Soc Sci Med 2004 6 1075 1084 10.1016/S0277-9536(03)00279-X
Torjuul K Nordam A Sørlie V Ethical challenges in surgery as narrated by practicing surgeons BMC Medical Ethics 2005 6 2 10.1186/1472-6939-6-2
Mishler EG Research interviewing – Context and narrative 1986 Cambridge, Massachusetts and London: Harvard University Press
Ricoeur P Interpretation Theory Discourse and the surplus of meaning 1976 Texas: Texas University Press
Lindseth A Marhaug V Norberg A Udén G Registered nurses' and physicians' reflections on their narratives about ethically difficult care episodes J Adv Nurs 1994 2 245 250 7930141 10.1046/j.1365-2648.1994.20020245.x
Udén G Norberg A Norberg S The stories of physicians, registered nurses and enrolled nurses about ethically difficult care episodes in surgical care Scand J Caring Sci 1995 4 245 253
Söderberg A Norberg A Intensive care: situations of ethical difficulty J Adv Nurs 1993 12 2008 2014 10.1046/j.1365-2648.1993.18122008.x
Sørlie V Førde R Lindseth A Norberg A Male physicians' narratives about being in ethically difficult care situations in paediatrics Soc Sci Med 2001 5 657 667
Sørlie V Jansson L Norberg A The meaning of being in ethically difficult care situations in pediatric care as narrated by female registered nurses Scand J Caring Sci 2003 3 285 292 10.1046/j.1471-6712.2003.00229.x
Sørlie V Kihlgren AL Kihlgren M Meeting ethical challenges in acute care work as narrated by enrolled nurses Nurs Ethics 2004 2 179 188 10.1191/0969733004ne682oa
Sørlie V Kihlgren A Kihlgren M Meeting ethical challenges in acute nursing care as narrated by registered nurses Nurs Ethics 2005 2 134 142
Nordam A Sørlie V Førde R Integrity in the care of elderly people, as narrated by female physicians Nurs Ethics 2003 4 388 403 12875536 10.1191/0969733003ne589oa
Lindseth A Norberg A A phenomenological hermeneutical method for researching lived experience Scand J Caring Sci 2004 2 145 53 15147477 10.1111/j.1471-6712.2004.00258.x
Beachamp G General surgeons and clinical ethics: a survey Can J Surg 1998 6 451 454
Sade RM Williams TH Perlman DJ Haney CL Stroud MR Ethics gap in surgery Ann Thorac Surg 2000 2 326 329 10735656 10.1016/S0003-4975(99)01469-1
Tung T Organ CH Ethics in surgery. Historical perspective Arch Surg 2000 1 10 13 10636339 10.1001/archsurg.135.1.10
Little JM Ethics in surgical practice Br J Surg 2001 6 769 70 11412244 10.1046/j.0007-1323.2001.01791.x
Løgstrup KE Jensen TI The ethical demand (Dan Original 1956, Copenhagen, Gyldendal) 1997 Philadelphia: Fortress Press
Gadamer HG Jakobsson J The enigmatic health (German Original 2003, Frankfurt am Main, Suhrkamp Verlag) 2003 Falun: Dualis Förlag AB
McIntyre N Popper K The critical attitude in medicine: the need for a new ethics BMJ 1983 287 1919 1923 6418271
Førde R Aasland OG Steen PA Medical end-of-life-decisions in Norway Resuscitation 2002 3 235 240 10.1016/S0300-9572(02)00270-8
Aristotle The Nicomachean ethics The world's classics 1980 Oxford: Oxford University Press
Scultz DS Carnevale FA Engagement and suffering in responsible caregiving: on overcoming maleficience in health care Theor Med 1996 3 189 207 10.1007/BF00489445
Førde R Professional ethics and prioritization is not fire and water Tidsskr Nor Laegeforen 2004 6 958 959 (in Norwegian)
Gulbrandsen P Førde R Aasland OG What does it feel like to be a gatekeeper? Tidsskr Nor Laegeforen 2002 19 1874 1879 (in Norwegian, English summary) 12362710
Norberg A Udén G Gender differences in moral reasoning among physicians, registered nurses and enrolled nurses engaged in geriatric and surgical care Nurs Ethics 1995 3 233 242 7583429
Sullivan M The new subjective medicine: taking the patient's point of view on health care and health Soc Sci Med 2003 7 1595 1604 12614708 10.1016/S0277-9536(02)00159-4
Pahuus M Life itself – a life philosophical interpretation of Christian thoughts 1993 Århus, Denmark: Philosophia (in Danish)
Guadagnoli E Ward P Patient participation in decision-making Soc Sci Med 1998 3 329 339 9681902 10.1016/S0277-9536(98)00059-8
Henriksen JO Vetlesen AJ Closeness and Distance Foundation, values and ethical theories in working with humans beings 2000 Oslo: Gyldendal Akademiske (in Norwegian)
|
15996268
|
PMC1182373
|
CC BY
|
2021-01-04 16:31:58
|
no
|
BMC Med Ethics. 2005 Jul 4; 6:7
|
utf-8
|
BMC Med Ethics
| 2,005 |
10.1186/1472-6939-6-7
|
oa_comm
|
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-221601881510.1186/1472-6947-5-22SoftwareA software tool for creating simulated outbreaks to benchmark surveillance systems Cassa Christopher A [email protected] Karin [email protected] Karen L [email protected] Kenneth D [email protected] Children's Hospital Informatics Program, Informatics Program – Mandl Group, 1 Autumn St, #721, Children's Hospital Boston, Boston, MA 02115, USA2 Clinical Decision Making Group, Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139, USA3 Harvard Medical School, Boston, MA 02115, USA4 Harvard/MIT Division of Health Sciences and Technology, Cambridge, MA 02139, USA2005 14 7 2005 5 22 22 7 11 2004 14 7 2005 Copyright © 2005 Cassa et al; licensee BioMed Central Ltd.2005Cassa 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
Evaluating surveillance systems for the early detection of bioterrorism is particularly challenging when systems are designed to detect events for which there are few or no historical examples. One approach to benchmarking outbreak detection performance is to create semi-synthetic datasets containing authentic baseline patient data (noise) and injected artificial patient clusters, as signal.
Methods
We describe a software tool, the AEGIS Cluster Creation Tool (AEGIS-CCT), that enables users to create simulated clusters with controlled feature sets, varying the desired cluster radius, density, distance, relative location from a reference point, and temporal epidemiological growth pattern. AEGIS-CCT does not require the use of an external geographical information system program for cluster creation. The cluster creation tool is an open source program, implemented in Java and is freely available under the Lesser GNU Public License at its Sourceforge website. Cluster data are written to files or can be appended to existing files so that the resulting file will include both existing baseline and artificially added cases. Multiple cluster file creation is an automated process in which multiple cluster files are created by varying a single parameter within a user-specified range. To evaluate the output of this software tool, sets of test clusters were created and graphically rendered.
Results
Based on user-specified parameters describing the location, properties, and temporal pattern of simulated clusters, AEGIS-CCT created clusters accurately and uniformly.
Conclusion
AEGIS-CCT enables the ready creation of datasets for benchmarking outbreak detection systems. It may be useful for automating the testing and validation of spatial and temporal cluster detection algorithms.
==== Body
Background
The public health information infrastructure is yielding real-time access to health data, enabling new approaches to surveillance for infectious outbreaks. Prior to the laboratory confirmation or physician diagnosis of an infectious disease, ill persons may exhibit behavioral patterns, symptoms, signs, or laboratory findings that can be tracked through a variety of data sources. The process of monitoring these data is often referred to as syndromic surveillance [1-3].
Real time outbreak detection algorithms tend to focus on the temporal and spatial patterns of cases. Some detection routines, such as Cusum [4], look at just one type of pattern, while others, such as the Space-Time Scan Statistic [5] incorporate both. What these algorithms have in common is an underlying model of typical, or baseline patterns, and the goal of detection is to recognize perturbations from the baseline [1]. The outbreak detection performance of a surveillance system can be measured in terms of its ability to distill "signal" (a cluster of cases in time and/or space) from noisy baseline.
Benchmarking the performance of detection algorithms requires training and validation data. When real data are not available, simulated data are often used [6,7]. Simulated outbreaks must reflect the diversity of threats that a surveillance system is expected to encounter and detect, whether these outbreaks occur naturally or are man-made.
Our approach to validating detection algorithms is to use semi-synthetic data, that is, authentic baseline data injected with artificial signals [8]. These signals are defined by a controlled feature set of variable parameters such as the size, location, shape, and duration of simulated outbreaks. Here we describe a software tool, the AEGIS Cluster Creation Tool (AEGIS-CCT), that enables users to create simulated clusters with controlled feature sets, varying the desired cluster radius, density, distance, and relative location away from a central point. AEGIS-CCT does not require the use of an external geographic information system (GIS) program for cluster creation.
Implementation
Functional specifications
AEGIS-CCT can create single patient clusters as well as sets of patient clusters based on a simple geographical model. Cluster data points are outputted as comma separated variable (CSV) files, and can optionally be appended to an existing file, supplied by the user. The tool can also generate different sets of clusters that range in value over a single parameter to rigorously validate detection algorithms. To create individual clusters, the user can vary a number of relevant outbreak parameters (Table 1). AEGIS-CCT creates at least two output files each time it is executed, with file names specified by the user. One is a cluster data file that contains the artificial cluster data, and the other is a record file describing the session cluster parameters. The data file contains a cluster point identification number (assigned numerically from 0 to the number of points minus 1), the latitude and longitude of the cluster point, and the relative date of the cluster point. When generating a series of n clusters, the program automatically generates n files with appended identifiers, each as separate data files.
Table 1 Parameters that can be altered when creating a single cluster.
Parameter Description
Cluster ID Number User specified reference or identification number for each cluster
Number of Points in the cluster Number of patients or points in the generated cluster.
"Reference Point" GIS Location The latitude-longitude coordinates of a reference point, which could be a hospital or a primary care facility, for example.
Maximum cluster radius The distance between the outermost point in the cluster and the center of the cluster.
"Angle" from the reference point The angle of the cluster measured counter-clockwise from due east of the reference point as zero degrees, using unit circle convention.
Distance from the reference point The distance between the center point of the cluster and the reference point.
Numbers of Days the Cluster should span The number of days from when the first person shows symptoms to when the last person does.
Date Algorithm This specifies which of the three models of temporal progression to use. Additional models can be incorporated into the software.
Description and output filenames The user can specify where the cluster data and user-specified cluster description will be written.
The AEGIS-CCT is a Java package including a geospatial engine and a user interface created using the Swing toolkit. The source for the entire package is provided under the Lesser GNU Public License [9] on a sourceforge.net development site [10]. Full details and updates to AEGIS-CCT can be obtained online [11].
Geocoding and precision of location
Programmatic methods were implemented to assign latitude-longitude coordinates to simulated cluster points, taking into account physical earth surface distances and not relying on external GIS software. Inside an AEGIS-CCT GIS class, there are three primary methods to handle these conversions. The first is a method to find the distance between two locations, which uses the specific latitude-longitude of the reference point to create a ratio of degrees per meter for north-south latitude and east-west longitude. Artificial data points are created 0.05 degrees to the north or east, and the corresponding physical distances (x, y) are calculated using the Haversine Formula, described below. The ratio is then computed, dividing the artificial data point distance by the calculated physical distance in meters on the Earth's surface. A second method finds a point that is a specific physical distance, measured at a specified angle, from a reference point. The angle is measured from the Euclidian x-axis and increased in a counter-clockwise form. The output is a second GIS data point related to the reference point by the angle and distance specified by the user. The third method finds the number of degrees of latitude and longitude per unit of physical distance in each respective direction.
GIS precision
Placing patients on a map requires consideration of earth curvature and precise latitudes and longitudes. Spherical equations break down significantly at small distances, but the Haversine formula [12] provides computationally exact results in almost all circumstances. For this calculation, Earth has radius R, and the locations of two points in spherical coordinates (latitude and longitude) have names [lon1, lat1] and [lon2, lat2]. The Haversine Formula is calculated using the following code:
dlon = lon2 - lon1;
dlat = lat2 - lat1;
a = (sin(dlat/2)) ^2 + cos(lat1) * cos(lat2) * (sin(dlon/2)) ^2;
c = 2 * atan2(sqrt(a), sqrt(1-a));
d = R * c;
This implementation was quality tested for accuracy using a series of latitude-longitude pairs from a sample dataset, measuring distances between two data points. Those results were identical with distances calculated by commercially-available GIS software.
Geotemporal progression
Outbreaks vary in their temporal progressions or epidemic curves. Three such progressions were implemented in AEGIS-CCT as date algorithms to model the ways in which a disease might manifest in a population over time: a random, a linear, and an exponential growth spread. Additional epidemiological date algorithms can be added by other users to the AEGIS-CCT by implementing a method to assign a specific temporal distribution within an array in Java.
For the random algorithm, a random number is generated that falls within the range of the number of days in the cluster, producing a random distribution. For the linear distribution, the day value is divided by the total day values and multiplied by the total number of points to determine the fraction of the total points to be injected each day. Scaling by a multiplier can alter the rate of linear growth. Similarly for the exponential distribution, the numerical value of e(multiplier * day number) is divided by the sum of e(multiplier * day number), for all day values. This ratio is then multiplied by the total number of points to determine the number of points that occur each day. Examples of the linear and exponential growth modified probability distribution estimations for date distribution are presented in Figures 1 and 2.
Figure 1 An example of the linear date algorithm estimation for thirty points spanning three days. The x-axis represents the day number.
Figure 2 An example of the exponential date algorithm estimation for thirty points spanning three days. The x-axis represents the day number.
Results
Analysis of accuracy and uniformity of patient cluster data
Semi-synthetic datasets created by the cluster generator fall within a specified set of parameter-based boundaries. Cluster data points are created randomly within the domain defined by those parameters, so it is important to verify that the clusters are accurately created and are close to uniformly generated.
To measure the uniformity of generated clusters, 10 test clusters were created with 100 points in each cluster. The centroid of each set of cluster points was then calculated and compared to the specified center point of that cluster. In every case, the cluster centroid was within five percent of the specified cluster radius, in distance, from the specified center point. This result demonstrates that the datasets are uniform, within a small threshold, when they contain a sufficient number of points, as would be expected with a random distribution.
To measure the accuracy of the geocoding engine, 360 clusters were made, forming a circle, around a single center point, varying the angle evenly (one degree added per cluster,) and they each had a cluster center point that fell precisely along the circle defined by all points at the same radial distance from the original counterpoint. This same test was conducted at five randomly selected latitude-longitude locations and the same results were obtained.
Sample cluster parameters and output
Two sample scenarios are described below, and their parameters are listed in Table 2. The first example demonstrates the creation of a single artificial cluster and the second example demonstrates the creation of several clusters at various angles around a single reference point. The first cluster includes cases in a simulated outbreak spanning five days.
Table 2 Example of single cluster parameters and multiple cluster parameters.
Parameter Type of Cluster
Single Cluster Set of Clusters Varying Angle
Cluster ID 100 101 [1:4]
Number of Points in Cluster 30 30
Reference Point Latitude 42.35666 42.35666
Reference Point Longitude -71.09516 -71.09516
Cluster Radius 600 m 400 m
Angle from reference point 90 [varies, see below]
Distance from reference point 1600 m 3000 m
Number of Days 5 5
Time-growth Pattern Linear Linear
Cluster Description Linear time-growth cluster north of center point Varied angle around center point and created 4 clusters.
Number of Clusters N/A 4
Minimum Angle N/A 0
Maximum Angle N/A 270
The single linear time-growth cluster was placed approximately 1600 m due north of a center point at longitude -71.09516 and latitude 42.35666. AEGIS-CCT outputted a CSV file and partial results are listed in Table 3. There are two points on the first day, four on the second day, six on the third day, linearly increasing to include thirty patient points by day 5.
Table 3 Sample output to a comma separated value file from AEGIS-CCT. Note: Values are point identification number, longitude, latitude and day number.
0,-71.09600452536358,42.37455407329273,1
1,-71.10149672138236,42.365894560466806,1
2,-71.0954755413253,42.373890954435524,2
3,-71.08968377859539,42.37242100053542,2
4,-71.09281946336338,42.36955324904336,2
5,-71.09564524977307,42.371694560897,2
6,-71.09345472615571,42.370979504450304,3
7,-71.09983295495935,42.369683605959985,3
8,-71.09781606117451,42.37282397457113,3
9,-71.09685871099056,42.37540065852763,3
10,-71.0921214185705,42.37216701505921,3
... (to point with ClusterID 29)
The output CSV can be imported into a GIS analysis tool. A map made from AEGIS-CCT output using MapPoint 2002 (Microsoft Corporation, Redmond, WA) is presented in Figure 3. The temporal progression (linear-growth algorithm) is indicated by the shaded color of coordinate points for each simulated case as shown in the legend. The temporal growth pattern of a similar injected linear-growth cluster spanning 7 days (containing a total of 56 points) is graphed when appended to a low-volume week (120 visits, 07/15-21/2001) and a high-volume week (472 visits, 01/14-20/2001) in Figure 4.
Figure 3 A single linear time-growth cluster north of center point.
Figure 4 Artificially-Injected Temporal Cluster into Low and High-Volume Weeks of Children's Hospital Boston ED Visit Data: An artificially-generated cluster (dashed line at bottom) containing a total of 56 additional points, linearly increasing in magnitude over a 7 day span, was added to two separate weeks of Children's Hospital Boston temporal visit data. The first week of data was from a low-volume week, containing a total of 120 authentic patient visits with an additional 56 artificially-generated visits while the second series contains a total of 472 authentic visits with the same 56 artificially-generated visits appended. While growth in the low-volume week is visible by inspection, it is difficult to notice the artificially added visits in a higher-volume week.
In the second example, the angle around the center point was varied, creating a series of four clusters. The series cluster generator automatically created four files, and each file was imported into MapPoint, and charted in a different color, as shown in Figure 5.
Figure 5 Creation of a series of four clusters around the center point (with the angle varied.)
The CCT can create new data files or it can append cluster data to an existing file. The artificial portion of a semi-synthetic dataset can then be automatically combined with baseline patient distributions by specifying a file that contains the baseline data. In future versions, the CCT may provide an xml schema to which input baseline data files should conform.
Limitations
AEGIS-CCT does not yet have built-in procedures to generate more complicated time-series or spatial distributions. Extensibility was taken very seriously when creating AEGIS-CCT, and sufficient abstractions were made to allow for ease of adding additional models. The generator is also extensible in other ways so that additional parameters can be added if they are easily computable. Parameters can be added or deleted by updating the GUI and modifying the GenerateCluster method in the main geospatial class. Potential areas for expansion in future versions include log-linear and logarithmic time-series models as well as Gaussian spatial distributions. It will be necessary to determine what the most pertinent and physically realistic models of syndromic spread are. Once the most realistic scenarios are assessed and modeled, they can be programmatically implemented and incorporated into the cluster creation tool. As long as these distributions can be implemented using Java methods, it is possible to quickly add another temporal distribution to AEGIS-CCT.
Conclusion
Evaluation of surveillance systems for the early detection of outbreaks is particularly challenging [13] when the systems are designed to detect events for which there are a few or no historical examples. Some real-time surveillance systems are designed to provide early warning of a biological attack. Fortunately, few people have been infected with biological warfare agents, although there are notable exceptions. For example, residents of Sverdlovsk were exposed in 1979 during an accidental release of anthrax from a weapons plant [14] and there were eleven infections, resulting in five deaths in the Florida, New York and Washington DC mailed-anthrax attacks in 2001 [15]. In the absence of sufficient real outbreak data, measuring the detection performance of a system requires simulation. AEGIS-CCT enables the ready creation of datasets for benchmarking outbreak detection systems.
Availability and requirements
Lists the following
* Project name: AEGIS Cluster Creation Tool
* Project home page:
* Operating system(s): Platform independent
* Programming language: Java
* Other requirements: Java 1.3.1 or higher
* License: e.g. GNU LGPL
* Any restrictions to use by non-academics: none
List of abbreviations
GIS – Geographical Information Systems
GUI – Graphical User Interface
CSV – Comma-Separated Values
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CC implemented the GIS engine and dataset management, thoroughly tested of program, and was the primary author of the manuscript. KI helped implement the graphical user interface and the file read and write components. KO provided datasets and domain knowledge of specific issues relevant to the evaluation of detection performance. KM 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
This work was supported by the National Institutes of Health through R01LM007970-01 from the National Library of Medicine, and by grant 2002-12-1 from the Alfred P. Sloan Foundation. This software was presented in part at the 2003 National Syndromic Surveillance Conference in New York, NY.
==== Refs
Mandl KD Overhage JM Wagner MM Lober WB Sebastiani P Mostashari F Pavlin J Gesteland PH Treadwell T Koski E Hutwagner L Buckeridge DL Aller R Grannis S Implementing syndromic surveillance: a practical guide informed by the early experience J Am Med Inform Assoc 2004 11 141 150 [PrePrint published Nov 21, 2003; as doi:10.1197/jamia.M1356] 14633933 10.1197/jamia.M1356
Tsui FC Espino JU Dato VM Gesteland PH Hutman J Wagner MM Technical description of RODS: a real-time public health surveillance system J Am Med Inform Assoc 2003 10 399 408 12807803 10.1197/jamia.M1345
Lombardo J Burkom H Elbert E Magruder S Lewis SH Loschen W Sari J Sniegoski C Wojcik R Pavlin J A Systems Overview of the Electronic Surveillance System for the Early Notification of Community-Based Epidemics (ESSENCE II) J Urban Health 2003 80 I32 I42. 12791777
Hutwagner LC Maloney EK Bean NH Slutsker L Martin SM Using laboratory-based surveillance data for prevention: an algorithm for detecting Salmonella outbreaks Emerging Infectious Diseases 1997 3 395 400 9284390
Kulldorff M Athas WF Feurer EJ Miller BA Key CR Evaluating cluster alarms: a space-time scan statistic and brain cancer in Los Alamos, New Mexico Am J Public Health 1998 88 1377 1380 9736881
Reis BY Pagano M Mandl KD Using temporal context to improve biosurveillance Proc Natl Acad Sci U S A 2003 100 1961 1965 12574522 10.1073/pnas.0335026100
Goldenberg A Shmueli G Caruana RA Fienberg SE Early statistical detection of anthrax outbreaks by tracking over-the-counter medication sales Proc Natl Acad Sci U S A 2002 99 5237 5240 11959973 10.1073/pnas.042117499
Mandl KD Reis BY Cassa C Measuring outbreak detection performance using controlled feature set simulations MMWR 2004 National Syndromic Surveillance Conference Proceedings
Open Source Initiative GNU Lesser General Public License
SourceForge AEGIS-CCT
Sinnott RW Virtues of the Haversine Sky and Telescope 1984 68 159
Buehler JW Hopkins RS Overhage JM Sosin DM Tong V Framework for evaluating public health surveillance systems for early detection of outbreaks: recommendations from the CDC Working Group MMWR Recomm Rep 2004 53 1 11 15129191
Meselson M Guillemin J Hugh-Jones M Langmuir A Popova I Shelokov A Yampolskaya O The Sverdlovsk anthrax outbreak of 1979 Science 1994 266 1202 1208 7973702
Jernigan JA Stephens DS Ashford DA Omenaca C Topiel MS Galbraith M Tapper M Fisk TL Zaki S Popovic T Meyer RF Quinn CP Harper SA Fridkin SK Sejvar JJ Shepard CW McConnell M Guarner J Shieh WJ Malecki JM Gerberding JL Hughes JM Perkins BA Anthrax Bioterrorism Investigation T Bioterrorism-related inhalational anthrax: the first 10 cases reported in the United States Emerging Infectious Diseases 2001 7 933 944 11747719
|
16018815
|
PMC1182374
|
CC BY
|
2021-01-04 23:52:11
|
no
|
BMC Med Inform Decis Mak. 2005 Jul 14; 5:22
|
utf-8
|
BMC Med Inform Decis Mak
| 2,005 |
10.1186/1472-6947-5-22
|
oa_comm
|
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-291595524710.1186/1471-2474-6-29Research ArticleTransforming growth factor beta-1 (TGFB1) and peak bone mass: association between intragenic polymorphisms and quantitative ultrasound of the heel Tzakas Peter [email protected] Betty YL [email protected] Alexander G [email protected] Laurence A [email protected] David EC [email protected] Dept. of Laboratory Medicine and Pathobiology, University of Toronto, Toronto ON, Canada2 Dept. of Clinical Pathology, Sunnybrook and Women's College Health Sciences Centre, Toronto ON, Canada3 Prosserman Centre for Health Research, Mount Sinai Hospital, Toronto ON, Canada4 Dept. of Medicine, University of Toronto, Toronto ON, Canada5 Dept. of Rheumatology, St Michael's Hospital, Toronto ON, Canada6 Dept. of Paediatrics (Genetics), University of Toronto, Toronto ON, Canada2005 14 6 2005 6 29 29 15 1 2005 14 6 2005 Copyright © 2005 Tzakas et al; licensee BioMed Central Ltd.2005Tzakas 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
Variance of peak bone mass has a substantial genetic component, as has been shown with twin studies examining quantitative measures such as bone mineral density (BMD) and quantitative ultrasound (QUS). Evidence implicating single nucleotide polymorphisms (SNPs) of the transforming growth factor beta-1 (TGFB1) gene is steadily accumulating. However, a comprehensive look at multiple SNPs at this locus for their association with indices of peak bone mass has not been reported.
Methods
A cohort of 653 healthy Caucasian females 18 to 35 years old was genotyped for seven TGFB1 SNPs. Polymorphisms were detected by restriction endonuclease digestion of amplified DNA segments.
Results
The frequencies of the least common allele at G-800A, C-509T, codon 10 (L10P), codon 25 (R25P), codon 263 (T263I), C861-20T, and 713-8 delC loci were 0.07, 0.33, 0.41, 0.08, 0.04, 0.25 and 0.01, respectively. A significant association was seen between QUS Stiffness Index (QUS-SI) and the SNP at codon 10 and the linked promoter SNP, C-509T. This association remained significant after multiple regression was used to incorporate important clinical covariates – age, BMI, level of activity, family history, and caffeine intake – into the model.
Conclusion
The association of QUS-SI with -509T is consistent with a gene-dose effect, while only individuals homozygous for the codon 10P allele showed a significant increase. In this cohort of young healthy Caucasian females, the T allele at position -509 is associated with greater bone mass as measured by calcaneal ultrasound.
==== Body
Background
Osteoporosis is a common disorder of aging characterized by low bone mineral density (BMD), deteriorating bone microarchitecture, and increased fracture rate. Osteoporosis and BMD are both complex traits with a strong genetic component [1,2]. Among the genes that have been associated with BMD are those encoding the vitamin D receptor [3,4], the estrogen receptors [5,6], collagen Iα1 [7], apolipoprotein E [8], and TGFB1 [9].
Abundant in bone matrix [10], secreted TGFB1 is an important regulator of osteoblast proliferation and differentiation and directly affects bone formation in vivo [11]. Activating mutations in humans are associated with Camurati-Engelmann disease, a disorder of progressive diaphyseal dysplasia characterized by increased BMD [12,13]. Homozygous knock-out of the TGFB1 gene in mice is associated with an osteopenic phenotype [14]. It is not surprising, therefore, that the TGFB1 locus has emerged as a strong candidate gene in the study of osteoporosis genetics.
The TGF-β1 amino acid sequence is highly conserved across mammalian species, indicating a strong selection against variant forms of the protein [15]. Variable expression or activation of TGF-β1 may therefore be associated with altered bone remodeling and different overall BMD [6]. Thus, elevated serum TGF-β1 levels are associated with osteosclerosis, and conversely, decreased serum TGF-β1 with osteopenia.
Nine TGFB1 SNPs have been identified and studied. Three reside in the promoter region (C-988A, G-800A, C-509T). An insertion of a basepair is found in the 5' UTR at n.+72C. Two SNPs are in the signal sequence (L10P or n.T869C, and R25P or n.G915C), one in exon 5 (T263I or c.C788T) [16], and one in each of introns 4 and 5 (713-8delC [17] and C861-20T [18], respectively) (Figure 1).
Figure 1 Polymorphic loci of the TGFB1 gene. Each locus (A to G) is shown in relation to the promoter, transcription start site (arrow), the exons (rectangles), the pre-propeptide (white rectangles), propeptide (black rectangles) and mature sequence (grey rectangles). Synonyms for each SNP (named differently by different researchers depending on annotation of position 1) are given in the shaded boxes.
A number of studies have examined the association of TGFB1 polymorphisms with TGF-β1 levels. In post-menopausal British women, the -509T allele was associated with higher total serum TGF-β1 [19], but this association was not confirmed in post-menopausal Chinese [20] or Japanese cohorts [21]. Similarly, the L allele at codon 10 was found to be associated with higher serum TGF-β1 levels in middle aged European women [22], but these findings were not confirmed in elderly Australian [23] or Chinese [20] women.
Other studies have also examined these allelic variants with respect to effects on bone. In 256 postmenopausal Italian women, Bertoldo et al. found strong evidence of association between the 713-8delC SNP and: (i) femoral neck BMD (FN-BMD), (ii) prevalence of hip fractures, and (iii) high bone turnover [24]. In 123 osteoporotic cases and 131 matched normal controls, Langdahl et al. found that the same 713-8delC deletion was associated with low bone mass and increased bone turnover [17]. In 76 pre-menopausal women, Keen et al. reported that the 861-20C allele was associated with higher BMD at the femoral neck, but not at the lumbar spine or with quantitative ultrasound (QUS) at the heel [18]. In a subsequent publication by Langdahl et al. involving 1168 osteoporotic cases and controls, the T allele at T861-20C was associated with significantly higher LS and FN BMD, while the T allele at C-509T was significantly associated with higher FN BMD, and the Pro allele of L10P was associated with greater FN BMD [25]. A study of the L10P variant in 625 Japanese women showed significant association of the 10P allele with higher lumbar spine (LS) BMD and total BMD [21]. This study also examined the -509 site and concluded that -509 T allele, alone or in combination with 10P, is a genetic determinant for osteoprosis. However, the more common allele at codon 10 (leucine) was found to be associated with higher LS BMD and FN BMD in middle-aged European women [22] and with higher FN BMD in Chinese elderly [20]. In 1337 elderly Australian women, the 10P allele associated with lower hip BMD and heel QUS [23]. The codon 10 polymorphism was not seen to have any association to BMD in 3382 elderly Caucasian American females [26]. In 356 healthy Japanese adolescents, however, the -509 CC homozygotes were found to have 5 to 6% greater radial BMD, and higher bone mineral content than the corresponding TT group [27].
In none of these reports were multiple SNP sites examined simultaneously in relation to peak bone mass. We therefore undertook a study of seven TGFB1 polymorphisms in relation to BMD indices as well as clinical covariates in a group of healthy young women with well characterized clinical and laboratory phenotypes to determine their relationship to peak bone mass.
Methods
Subjects
The 653 healthy non-related Caucasian female subjects presented in this report were recruited by advertising for a study of bone and mineral metabolism in the Greater Toronto Area, as described previously [4,6]. In this study, data on 25 women identified as originating from the Indian sub-continent were excluded.
The study protocol was approved by the Institutional Review Board of the University of Toronto (Toronto, ON) and written consent was obtained from each individual at the onset of the study [4]. Each subject completed a standardized questionnaire about lifestyle factors, menstrual and reproductive history, past history of medical diseases, current and prior medication use, history of fractures, and family history of osteoporosis. Summary clinical characteristics (mean ± SD) include: age, 27.5 ± 4.5 yr (range 18 to 35 yr); weight, 63.3 ± 11.6 kg; height, 1.65 ± 0.06 m; and BMI 23.2 ± 3.9 kg/m2.
Assessment of BMD
BMD was measured at the hip (left femoral neck = FN) and lumbar spine (L2-L4 = LS) by dual-energy X-ray absorptiometry using a DPX-L absorptiometer (Lunar Corp., Madison, WI, U.S.A.). BMD was reported as grams per square centimeter. Quantitative Ultrasound (QUS) measurements were also conducted, consisting of broadband ultrasound attenuation (BUA, dB/MHz), speed of sound (SOS, m/s) and the calculated heel stiffness index (SI, % of age-matched controls), at the right heel (Lunar Achilles; Lunar Corp., Madison WI), as published previously [4]. Summary statistics for our study cohort were: LS-BMD, 1.19 ± 0.13 g/cm2; FN-BMD, 1.00 ± 0.12 g/cm2; and QUS-SI, 97.2 ± 14.1.
DNA isolation and genotyping
DNA was extracted from EDTA-anticoagulated blood as previously described [6]. Gene fragments including the SNPs of interest were amplified by polymerase chain reaction (PCR). Of the nine SNPs known in this region, two were not included in this study. SNPs are named by either nucleotide sequence or amino acid sequence, using the most commonly reported annotation in literature. The SNP +72C was not examined, since previous work [16] revealed it to be in complete linkage with the codon 25 SNP. An additional SNP in the promoter, C-988A, was excluded because the frequency was reported to be only 0.2% [16]. The relative positions of the remaining seven intragenic SNPs are shown in Figure 1. To genotype the C861-20T SNP, previously published methods were used [18]. Methods for genotyping the other six SNPS by PCR amplification and RFLP digestion are similar to those previously published [4]. Given their pairwise proximity, three amplicons were used for genotyping: one for the pair of promoter SNPs, -800 and -509, a second for codon 10 and codon 25, and a third for 713-8delC and codon 263. Primer pairs are given in Table 1. PCR conditions were: 94°C for 1 min (denaturing), 62°C for 40 secs (primer annealing), and 72°C for 1 min + 1 second/cycle (extension) for 35 cycles. Two μl of Q-solution (Qiagen Corp) were added to the codon 10/25 PCR mixture to facilitate amplification of this GC-rich fragment. Also listed in Table 1 are the restriction endonuclease enzymes used to detect one of the two alleles (as specified). Digested DNA products were run on pre-cast Clearose BG wide-mini S-50 gels (Elchrom Scientific, Zurich, Switzerland) and stained with ethidium bromide. All genotypes were sequenced to confirm restriction digestion results.
Table 1 Primer sequences used for the amplification of polymorphic TGFB1 sites.
Downstream Primer Upstream Primer Amplicon Size (bp) Enzymea Cleavageb
-800 Eco181I Wild-type
AGAACAGTTGGCACGGGCTT AACGGAAGAGAGTCAGGCT 583
-509 TaiI Wild-type
Codon10 MspA1I Wild-type
CTACCTTTTGCCGGGAGACC CACCAGCTCCATGTCGATAG 226
Codon25 CTTTCGCCTTAGCGCCCACT TAGTTGGTGTCCAGGGCTCG 343 Sau961 Wild-type
Codon 263 Fok1 Wild-type
713-8delC BseL1 Variant
aRestriction endonucleases (Enzyme) are shown with digestions conducted separately. bAllele recognition (and cleavage) for either the major (Wild-type) or minor (Variant) allele.
Statistical analysis
Haplotyping of -800 and -509 SNP loci was possible since both SNPs were amplified in one PCR product, and simultaneous digestion by Eco181 and TaiI produced a unique pattern of fragments for each possible haplotype [28,29]. Loci were tested for Hardy-Weinberg (HW) equilibrium of the distribution of the genotypes, and pairwise linkage disequilibrium (LD) coefficients calculated and normalized (Δ'), using methods of Weir et al. [30], and improved software as we have described [31]. Bivariate correlations between BMD, QUS-SI, and lifestyle factors, menstrual and reproductive factors, and TGFB1 genotypes were assessed using Spearman rank correlation coefficients, as reported previously [4,6]. Each TGFB1 SNP was added to the model without any clinical covariates and then with covariates used in our established model [6]. All data analyses were performed with SPSS for Windows 10.0.7 (SPSS Inc., Chicago, Illinois, U.S.A.).
Results
TGFB1 polymorphisms
Frequencies of the minor alleles were: codon 10 (41%), -509 (33%), C861-20T (25%), codon 25 (8%), -800 (7%), codon 263 (4%), and 713-8delC (1%). For the -800 and -509 haplotypes, 4.3% (n = 28) of the samples were doubly heterozygous, and all were in trans, as detected by sequential enzyme digestion (data not shown). Shown in Table 2 are the pair-wise standardized LD coefficients (Δ'). In this population the -509 is closely linked to codon 10 (Δ' = 0.88) and is in negative LD with 713-8delC (Δ' = -0.77). Codon 10 is also in negative LD with 713-8delC (Δ' = -0.80).
Table 2 TGFB1 genotype and variant allele frequencies, and pair-wise standardized LD coefficients (Δ')
Linkage Disequilibrium Coefficients (Δ')
SNP Genotype Frequency Variant Allele Frequency -800 -509 Codon 10 Codon 25 713- 8DelC Codon 263
GG 564
-800 GA 85 0.07
AA 4
CC 280
-509 CT 310 0.33 -0.53***
TT 63
Codon 10 TT 212
TC 337 0.41 -0.47*** 0.88***
CC 97
Codon 25 GG 555
GC 92 0.08 -0.66** -0.44*** 0.48***
CC 5
713- 8delC CC 626
C - 18 0.04 -0.61 -0.77** -0.80*** -1.00**
- - 0
Codon 263 CC 369
CT 242 0.01 -0.25 0.46*** 0.50*** -0.32 -1.00**
TT 41
C861-20T CC 608
CT 43 0.25 -0.26 * 0.007 -0.10* -0.37** 0.22* -0.49**
TT 2
* P <0.05, ** P < 0.01, *** P <0.001
Clinical characteristics by TGFB1 genotype
In comparing clinical parameters among genotype groups, no significant differences were observed for age, weight, height, calcium intake, caffeine intake, or BMI. Summary data grouped by -509 and codon 10 genotypes are shown in Table 3. Results for other genotypes are not shown, as there are no significant differences, and the other SNPs as explained below were not significant determinants of peak bone mass in this population.
Table 3 Clinical variables grouped by -509 and codon10 Genotype.*
-509 Codon10
CC §(280) CT (310) TT (63) LL (212) LP (337) PP (97)
Age (yr) 27.1 ± 4.3 27.9 ± 4.7 27.5 ± 4.8 27.3 ± 4.2 27.6 ± 4.6 27.7 ± 4.9
Weight (kg) 63 ± 13 64 ± 11 62 ± 9 63 ± 2 64 ±12 62 ± 9
Height (m) 1.65 ± 0.06 1.65 ± 0.07 1.66 ± 0.07 1.65 ± 0.06 1.65 ± 0.07 1.65 ± 0.07
Calcium intake (mg/day) 590 ± 380 550 ± 340 510 ± 310 580 ± 370 570 ± 370 510 ± 300
Caffeine intake (cups/day) 1.49 ± 1.21 1.62 ± 1.61 1.33 ± 1.18 1.49 ± 1.20 1.65 ± 1.57 1.25 ± 1.21
BMI (kg/m2) 23 ± 3 23 ± 3 23 ± 4 23 ± 3 23 ± 3 23 ± 3
*Data expressed as mean ± SD
§Genotype counts (n) given in parentheses.
TGFB1 SNPs and clinical characteristics
Correlations of QUS-SI, lumbar spine BMD, and femoral neck BMD with clinical, lifestyle and the TGFB1 genotypes are shown in Table 4. As previously reported for this cohort, LS BMD is positively correlated with age, height, weight, BMI and level of physical activity as an adolescent [4]. At the hip, significant positive correlations of BMD were found with height, weight, BMI, calcium intake, and the amount of physical activity reported at time of study and during adolescent years. Negative correlations were seen with age and the amount of caffeine consumption for the FN BMD. Heel QUS was positively correlated with the same factors as FN BMD, with the exception of calcium intake.
Table 4 Correlation coefficients (Spearman R) and their significance (p) are shown for QUS-SI and BMD in relation to clinical, lifestyle and TGFB1 genotypes.*
Regression Parameters Heel QUS-SI Lumbar Spine BMD Femoral Neck BMD
R p R p R p
Clinical
Age -0.106 0.007 0.124 0.001 -0.109 0.005
Height 0.096 0.015 0.337 <0.001 0.323 <0.001
Weight 0.195 <0.001 0.166 <0.001 0.242 <0.001
BMI 0.157 <0.001 0.287 <0.001 0.208 <0.001
Lifestyle
Calcium Intake 0.050 0.203 0.065 0.098 0.095 0.016
Caffeine Intake -0.125 0.002 -0.004 0.927 -0.082 0.037
Smoking -0.027 0.499 -0.017 0.660 0.027 0.498
Currently Active 0.104 0.008 0.039 0.318 0.083 0.033
Adolescent Activity 0.147 <0.001 0.108 0.006 0.112 0.004
SNPs
TGFB1 -509 0.078 0.038 0.39 0.321 0.002 0.958
TGFB1 codon 10 0.043 0.281 0.001 0.994 -0.02 0.614
*In bold are significant results (p < 0.05).
Association of TGFB1 genotypes with BMD and QUS
There was no significant correlation of any TGFB1 SNP with BMD at FN or LS (data not shown). Of the seven loci, only the -509 SNP was significantly correlated with QUS-SI. (r = 0.078, p = 0.038). Since this SNP is in strong linkage disequilibrium with codon 10 (Δ' = 0.88) and 713-8delC (Δ' = -0.77), all three of these SNPs were included in further analysis.
The number of T alleles (0, 1 or 2) at the -509 site shows significant association with QUS-SI (Figure 2). Mean QUS-SI was higher in TT homozygotes (101.2 ± 12.9) than in CT heterozygotes (97.42 ± 10.30, p = 0.04) or CC homozygotes (95.8 ± 9.31, p = 0.006) and in heterozygotes versus CC homozygotes (p < 0.05). A similar allele-dose effect was not observed with the codon10 SNP; homozygotes for the rare allele (PP group) had a significantly higher mean QUS-SI (100.6 ± 13.2) than heterozygotes (96.4 ± 14.1, p = 0.007) or LL homozygotes for the T allele (97.2 ± 14.2, p = 0.041). Adjusted QUS-SI means for the -509 SNP, after correction for age, caffeine intake and BMI (means of 27.5 years, 1.54 cups/day and 23.2 kg/m2, respectively), were: CC 93.6 ± 1.0, CT 95.0 ± 1.0, and TT 98.7 ± 1.9. For codon 10, adjusted means were: LL 94.8 ± 1.1, LP 93.7 ± 1.0 and PP 98.3 ± 1.5. For SNP 713-8delC, only 18 heterozygotes were found and no significant difference in QUS-SI means was observed (data not shown).
Figure 2 Mean right heel ultrasound stiffness for -509 and codon10 genotypes. Error bars represent standard error of means. P-values were obtained by Scheffé correction for multiple testing after routine ANOVA.
Multivariate analysis
Based on the model previously reported by Rubin et al. [4], the seven TGFB1 genotypes were each introduced as single additive variables (variant allele counted as 0, 1 or 2) to the original multivariate analysis to test for significance. The LS BMD model included the following clinical variables: weight, physical activity at time of study, and physical activity during adolescence, age, paternal history of osteoporosis, family history of osteoporosis, and past amenorrhea greater than 3 months. None of the TGFB1 SNPs were significant predictors of LS BMD, nor were results significant for femoral neck (data not shown).
In a previously reported model from our laboratory on this group of Caucasian females the estrogen receptor (ER1) SNPs were examined [6]. ER1 XbaI, PvuII restriction haplotypes and AIB1 genotypes were interactively significant in that model, which included the following clinical covariates: age, physical activity at the time of the study and physical activity in adolescent years, caffeine consumption, and BMI. When -509 and codon 10 SNPs were added as new variables to the clinical covariates model, they remained significant (Table 5). Thus, the -509 and codon 10 SNPs appear to make significant independent contributions to the explained variance of QUS-SI.
Table 5 Multiple linear regression analysis of genetic and clinical determinants of QUS-SI modelled to determine individual contribution of -509 and codon10.*
-509 Codon 10
Variables F-stat R2 p-value Variables F-stat R2 p-value
Age 6.21 0.072 0.0130 Age 5.68 0.069 0.0175
Caffeine Intake 12.74 0.084 0.0004 Caffeine Intake 9.54 0.077 0.0021
BMI 19.24 0.185 <0.0001 BMI 18.75 0.185 <0.0001
Active as adolescent 7.46 0.107 0.0065 Active as Adolescent 9.06 0.120 0.0027
Currently Active 8.51 0.082 0.0037 Currently active 7.43 0.078 0.0066
Relatives with Osteoporosis 1.98 0.008 0.1598 Relatives with Osteoporosis 1.95 0.007 0.1627
§-509 SNP 3.18 0.056 0.0422 §Codon 10 3.32 0.062 0.0369
*Shown for each dependent variable are the individual F-statistics (F-stat), the variance attributable to that variable (R2, based on Type III sum of squares), and the significance (p-value).
§ The -509 and codon10 SNPS are significant (P <0.05) when added to the previously reported model [6]
Discussion
TGF-β1 is a secreted factor that plays an important role in bone remodelling. It is a potent stimulator of osteoblastic bone formation, causing chemotaxis, proliferation and differentiation in committed osteoblasts [11]. Although in vitro experiments have led to conflicting reports about the effects of exogenous TGF-β1 in cultured osteoblast systems [32,33], secreted TGF-β1 leads to matrix growth and osteoblast stimulation in vivo, while it inhibits mineralization, osteoclast differentiation and the resorptive activity of mature osteoclasts [34,35]. Mouse knockouts for TGFB1 have skeletal defects, including shortened long bones and decreased tibial BMD. In humans, however, mutations found in the pro-peptide of TGFB1 are associated with Camurati-Engelman Disease (CED), an autosomal dominant disorder manifesting as periosteal and endosteal thickening of the long bone diaphyses [12,13]. Most of these mutations lie in the Latent Associated Peptide (LAP) that is cleaved from the mature TGF-β1 peptide. LAP subsequently binds to mature TGF-β1 to form an inactivated secretory complex. Presumably, mutations that interfere with binding of LAP to the mature peptide would lead to increased TGF-β1 activation, stimulating bone remodelling, net bone deposition, and resulting in a denser skeleton.
In our cohort, allele frequencies for the -800, -509, codon 10, codon 25 and codon 263 SNPs are in agreement with those published by Cambien et al. [16], and Syrris et al. [36], who examined French, Irish and UK populations. Allele frequencies in our cohort for 713-8delC and C861-20T, respectively, are comparable to those in the European Caucasians studied by Langdahl et al. and by Keen et al. [17,18]. Of the 7 SNPs examined for linkage, 713-8delC was in complete negative LD with codon 25 and codon 263. The 713-8delC SNP was in strong negative LD with -509 (Δ' = -0.77) and codon 10 (Δ' = -0.80), suggesting a single haplotype. Moreover, disequilibrium coefficients in our population were comparable to those reported by Cambien et al [16].
The two SNPs in the promoter, G-800A and C-509T, may theoretically alter the binding of RNA polymerase or other transcription factors. Luedecking et al. found that the transcriptional activity of the -509T variant allele of the TGFB1 gene was slightly greater than that of the common C allele [37]. Serum TGF-β1 concentrations are increased in a gene dose-dependent fashion, with differences in concentrations in TT homozygotes being twice that of TC individuals, when the CC genotype is taken as the referent [19]. Our results for the -509 polymorphism are consistent with this view that the T allele (high TGF-β1 producer) is associated with increased bone formation in young women, since heel QUS-SI is increased in rough proportion to the number of T alleles present. Langdahl reported a similar finding in a healthy Danish middle aged cohort, in whom higher FN BMD was associated with TT genotype [25]. In the Japanese population, however, the -509 CC genotype has been found associated with higher BMD [21]. This difference may be due to background genetic differences in the populations, ascertainment differences affecting the TGF-β1 genotype distribution in the sample set, or environmental factors. Differences in linkage disequilibrium may also contribute to this discordance, and shed light on parallel differences with codon 10 associations. In some studies, the 10Pro allele is associated with increased BMD [21,25], but not in others [20,22,23].
The TGFB1 codon 10 and 25 polymorphisms are located in the signal sequence, which is cleaved from the TGF-β1 precursor at codon 29 (Gly29-Leu30 peptide bond). In general, signal sequence mutations affect peptide export efficiency [38]. The replacement of leucine by proline at position 10 would be expected to disrupt the alpha-helical domain, while replacement of arginine by proline at position 25 would change the characteristic polarity of the carboxyl domain of the signal sequence. In either case, altered signal peptide regulation leading to differential trafficking or export would be the likely mechanism underlying genotype-dependent differences in TGF-β1 metabolism.
The physiologic evidence for such differential expression is not extensive. Awad et al. reported a significant correlation between codon 25 genotype and amounts of TGF-β1 secreted by cultured lymphocytes stimulated in vitro [39]. Significant association of the codon 10 genotype with plasma TGF-β1 concentrations, BMD at lumbar spine, and vertebral fracture frequency, has been reported in controls and 2 different Japanese populations of osteoporotic patients [40]. In both control and osteoporotic groups, the association with higher plasma TGF-β1 and the 10P allele was roughly additive [9]. In a study of Japanese adolescents, those homozygous for the 10P genotype had significantly higher BMD than those homozygous for the L10 allele [27]. These results suggest that increased skeletal mineralization during puberty may be related to the presence of a 10P allele. Our study of Caucasian pre-menopausal women found no difference in BMD at either FN or LS, but found a positive association between the Leu10 allele and calcaneal QUS-SI.
Of the remaining SNPs, Langdahl et al. reported that the 713-8delC was more frequent in osteoporotic women (6/123), than controls (2/131) but no correlation to bone mass was demonstrated in the controls [25]. It is possible that linkage of this SNP with other functional sites provides an explanation for variable but genuine association with parameters of bone strength and mineral density. Indeed, we observed that the 713-8delC is in negative disequilibrium with codon 10 and -509 (Δ' = -0.77 and -0.80, respectively) in our population. Thus, previous positive results with this genotype alone may represent association with a haplotype extending across all 3 loci.
None of the TGFB1 SNPs in this study was significantly associated with either spine or hip BMD. Although apparently negative in contrast to other studies, our findings may be explained in part by the demographic characteristics of our cohort. Ours was a population of healthy young adults, whereas previous research showing an association with codon 10 and 713-8delC was based on post-menopausal women diagnosed with osteoporosis [25]. It is generally understood that studies of post-menopausal women are predominantly investigations of the rate of bone loss, whereas those in young adults are identifying genetic determinants of peak bone mass accumulation. While BMD is a measure of the density of the mineralized bone, QUS-SI is a measure of bone architecture as well [41].
The skeletal site examined for bone mass and bone quality is important. Modelling of data in twins has indicated that there are both common and specific genetic factors acting on bone at different skeletal sites and on different aspects on bone quality [42,43]. Our study suggests that TGFB1 genetic variation does not seem to be a major factor in total bone mineralization per se, at least as measured by BMD at the hip and femur.
Significant linkage disequilibrium exists between -509 and codon10 alleles and the results in our cohort (Δ' = 0.88) are no exception. However, the LD is not complete, and it is therefore no surprise that both -509 and codon 10 show association to QUS-SI, even when entered as independent variables in the multiple regression analysis. When means are compared for each genotype at a single locus, only the -509 genotypes showed a gene dosage effect, much as Yamada et al. described in a cohort of Japanese adolescents [27]. Evidence indicates that the T allele of -509 may be the more important SNP leading to an increase rate of transcription, which in turn affects bone mass. For our cohort, the multiple regression model for QUS-SI, as the dependent variable, adjusted for clinical covariates (age, height, weight, BMI, caffeine and calcium intake) and therefore these factors are likely confounders of the genotype effects seen.
Our findings must be interpreted in the context of several potential limitations. Our Caucasian cohort was recruited from an ethnically diverse urban population and admixture effects cannot be excluded. Only women were recruited, and we cannot say whether a similar association would be observed in men. Recruitment was conducted by public advertisement, and there may be significant bias toward those believing they were at higher risk for a bone related disease, particularly because of family history. However, the parameters for LS and FN BMD and for QUS-SI are comparable to other cohorts of healthy young women [4], and the associations we describe are independent of family history in our multiple regression model.
Common variations in DNA sequence are often associated with mild phenotypic effects [44]. Thus even a single SNP could account for a significant variation in bone mass so as to potentially influence subsequent fracture risk. However, the contribution of genotypes determined by all intragenic loci within a gene must be evaluated with environmental factors, in order to generate a balanced picture of gene-environment interactions for a complex quantitative trait like peak bone mass.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PT carried out genetic tests, analysed data and drafted the manuscript.
DC conceived the study, and supervised the study design, statistical analysis, and drafting of manuscript
BW was involved in the design of the genotyping assays and sample preparation
AL was involved in study design and interpretation of the data
LR was involved in recruitment of study population, acquisition of samples, and auditing of the clinical database
All authors read and approved the final manuscript
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by an IOC Partnership Grant from the National Science and Engineering Research Council and Dairy Farmers of Canada.
==== Refs
Pocock NA Eisman JA Hopper JL Yeates MG Sambrook PN Eberl S Genetic determinants of bone mass in adults. A twin study J Clin Invest 1987 80 706 710 3624485
Patel MS Rubin LA Cole DE Henderson JE, Goltzman D Genetic determinants of peak bone mass The Osteoporosis Primer 2000 Cambridge: Cambridge University Press 131 146
Morrison NA Qi JC Tokita A Kelly PJ Crofts L Nguyen TV Sambrook PN Eisman JA Prediction of bone density from vitamin D receptor alleles Nature 1994 367 284 287 8161378 10.1038/367284a0
Rubin LA Hawker GA Peltekova VD Fielding LJ Ridout R Cole DE Determinants of peak bone mass: clinical and genetic analyses in a young female Canadian cohort J Bone Miner Res 1999 14 633 643 10234586
Kobayashi S Inoue S Hosoi T Ouchi Y Shiraki M Orimo H Association of bone mineral density with polymorphism of the estrogen receptor gene J Bone Miner Res 1996 11 306 311 8852941
Patel MS Cole DE Smith JD Hawker GA Wong B Trang H Vieth R Meltzer P Rubin LA Alleles of the estrogen receptor alpha-gene and an estrogen receptor cotranscriptional activator gene, amplified in breast cancer-1 (AIB1), are associated with quantitative calcaneal ultrasound J Bone Miner Res 2000 15 2231 2239 11092404
Grant SF Reid DM Blake G Herd R Fogelman I Ralston SH Reduced bone density and osteoporosis associated with a polymorphic Sp1 binding site in the collagen type I alpha 1 gene Nat Genet 1996 14 203 205 8841196 10.1038/ng1096-203
Shiraki M Shiraki Y Aoki C Hosoi T Inoue S Kaneki M Ouchi Y Association of bone mineral density with apolipoprotein E phenotype J Bone Miner Res 1997 12 1438 1445 9286760
Yamada Y Miyauchi A Goto J Takagi Y Okuizumi H Kanematsu M Hase M Takai H Harada A Ikeda K Association of a polymorphism of the transforming growth factor-beta1 gene with genetic susceptibility to osteoporosis in postmenopausal Japanese women J Bone Miner Res 1998 13 1569 1576 9783545
Seyedin SM Thompson AY Bentz H Rosen DM McPherson JM Conti A Siegel NR Galluppi GR Piez KA Cartilage-inducing factor-A. Apparent identity to transforming growth factor-beta J Biol Chem 1986 261 5693 5695 3754555
Cohen MM Jr TGF beta/Smad signaling system and its pathologic correlates Am J Med Genet A 2003 116 1 10 12476444 10.1002/ajmg.a.10750
Janssens K Gershoni-Baruch R Guanabens N Migone N Ralston S Bonduelle M Lissens W Van ML Vanhoenacker F Verbruggen L Mutations in the gene encoding the latency-associated peptide of TGF-beta 1 cause Camurati-Engelmann disease Nat Genet 2000 26 273 275 11062463 10.1038/81563
Kinoshita A Fukumaki Y Shirahama S Miyahara A Nishimura G Haga N Namba A Ueda H Hayashi H Ikegawa S TGFB1 mutations in four new families with Camurati-Engelmann disease: confirmation of independently arising LAP-domain-specific mutations Am J Med Genet A 2004 127 104 107 15103729 10.1002/ajmg.a.20671
Dickson MC Martin JS Cousins FM Kulkarni AB Karlsson S Akhurst RJ Defective hematopoiesis and vasculogenesis in transforming growth-factor-beta-1 knock out mice Development 1995 121 1845 1854 7600998
Derynck R Jarrett JA Chen EY Goeddel DV The murine transforming growth factor-beta precursor J Biol Chem 1986 261 4377 4379 3007454
Cambien F Ricard S Troesch A Mallet C Generenaz L Evans A Arveiler D Luc G Ruidavets JB Poirier O Polymorphisms of the transforming growth factor-beta 1 gene in relation to myocardial infarction and blood pressure. The Etude Cas-Temoin de l'Infarctus du Myocarde (ECTIM) Study Hypertension 1996 28 881 887 8901839
Langdahl BL Knudsen JY Jensen HK Gregersen N Eriksen EF A sequence variation: 713-8delC in the transforming growth factor-beta 1 gene has higher prevalence in osteoporotic women than in normal women and is associated with very low bone mass in osteoporotic women and increased bone turnover in both osteoporotic and normal women Bone 1997 20 289 294 9071481 10.1016/S8756-3282(96)00363-8
Keen RW Snieder H Molloy H Daniels J Chiano M Gibson F Fairbairn L Smith P MacGregor AJ Gewert D Evidence of association and linkage disequilibrium between a novel polymorphism in the transforming growth factor beta 1 gene and hip bone mineral density: a study of female twins Rheumatology (Oxford) 2001 40 48 54 11157141 10.1093/rheumatology/40.1.48
Grainger DJ Heathcote K Chiano M Snieder H Kemp PR Metcalfe JC Carter ND Spector TD Genetic control of the circulating concentration of transforming growth factor type beta1 Hum Mol Genet 1999 8 93 97 9887336 10.1093/hmg/8.1.93
Lau HH Ho AY Luk KD Kung AW Transforming growth factor-beta1 gene polymorphisms and bone turnover, bone mineral density and fracture risk in southern chinese women Calcif Tissue Int 2004 74 516 521 15354859 10.1007/s00223-004-0163-4
Yamada Y Miyauchi A Takagi Y Tanaka M Mizuno M Harada A Association of the C-509-->T polymorphism, alone of in combination with the T869-->C polymorphism, of the transforming growth factor-beta1 gene with bone mineral density and genetic susceptibility to osteoporosis in Japanese women J Mol Med 2001 79 149 156 11357939 10.1007/s001090100190
Hinke V Seck T Clanget C Scheidt-Nave C Ziegler R Pfeilschifter J Association of transforming growth factor-beta1 (TGFbeta1) T29 -> C gene polymorphism with bone mineral density (BMD), changes in BMD, and serum concentrations of TGF-beta1 in a population-based sample of postmenopausal german women Calcif Tissue Int 2001 69 315 320 11800227 10.1007/s002230020024
Dick IM Devine A Li S Dhaliwal SS Prince RL The T869C TGF beta polymorphism is associated with fracture, bone mineral density, and calcaneal quantitative ultrasound in elderly women Bone 2003 33 335 341 13678774 10.1016/S8756-3282(03)00158-3
Bertoldo F D'Agruma L Furlan F Colapietro F Lorenzi MT Maiorano N Iolascon A Zelante L Locascio V Gasparini P Transforming growth factor-beta1 gene polymorphism, bone turnover, and bone mass in Italian postmenopausal women J Bone Miner Res 2000 15 634 639 10780855
Langdahl BL Carstens M Stenkjaer L Eriksen EF Polymorphisms in the transforming growth factor beta 1 gene and osteoporosis Bone 2003 32 297 310 12667558 10.1016/S8756-3282(02)00971-7
Ziv E Kahn A Cauley J Morin P Saiz R Browner W No association between the TGF-beta 1 Leu10Pro polymorphism and osteoporosis among white women in the United States Am J Med 2003 114 227 231 12637138 10.1016/S0002-9343(02)01393-1
Yamada Y Hosoi T Makimoto F Tanaka H Seino Y Ikeda K Transforming growth factor beta-1 gene polymorphism and bone mineral density in Japanese adolescents Am J Med 1999 106 477 479 10225252 10.1016/S0002-9343(99)00043-1
Sawitzke AD Sawitzke AL Ward RH HLA-DPB typing using co-digestion of amplified fragments allows efficient identification of heterozygous genotypes Tissue Antigens 1992 40 175 181 1361686
Peltekova VD Cole DE Pavlova A Rubin LA Improved method for direct haplotyping at the vitamin D receptor gene locus Clin Biochem 1998 31 191 194 9629494 10.1016/S0009-9120(98)00005-8
Weir BS Genetic Data Analysis II 1996 Sunderland, Massachusetts: Sinauer Associates, Inc. Publishers
Hamilton DC Cole DE Standardizing a composite measure of linkage disequilibrium Ann Hum Genet 2004 68 234 239 15180703 10.1046/j.1529-8817.2004.00056.x
Pfeilschifter J Seyedin SM Mundy GR Transforming growth factor beta inhibits bone resorption in fetal rat long bone cultures J Clin Invest 1988 82 680 685 3165385
Fuller K Lean JM Bayley KE Wani MR Chambers TJ A role for TGFbeta(1) in osteoclast differentiation and survival J Cell Sci 2000 113 2445 2453 10852823
Antosz ME Bellows CG Aubin JE Effects of transforming growth factor beta and epidermal growth factor on cell proliferation and the formation of bone nodules in isolated fetal rat calvaria cells J Cell Physiol 1989 140 386 395 2787326 10.1002/jcp.1041400225
Dieudonne SC Semeins CM Goei SW Vukicevic S Nulend JK Sampath TK Helder M Burger EH Opposite effects of osteogenic protein and transforming growth factor beta on chondrogenesis in cultured long bone rudiments J Bone Miner Res 1994 9 771 780 8079653
Syrris P Carter ND Metcalfe JC Kemp PR Grainger DJ Kaski JC Crossman DC Francis SE Gunn J Jeffery S Transforming growth factor-beta1 gene polymorphisms and coronary artery disease Clin Sci 1998 95 659 667 9831690 10.1042/CS19980154
Luedecking EK DeKosky ST Mehdi H Ganguli M Kamboh MI Analysis of genetic polymorphisms in the transforming growth factor-beta1 gene and the risk of Alzheimer's disease Hum Genet 2000 106 565 569 10914688 10.1007/s004390050026
Walter P Johnson AE Signal sequence recognition and protein targeting to the endoplasmic reticulum membrane Annu Rev Cell Biol 1994 10 87 119 7888184 10.1146/annurev.cb.10.110194.000511
Awad MR El Gamel A Hasleton P Turner DM Sinnott PJ Hutchinson IV Genotypic variation in the transforming growth factor-beta1 gene: association with transforming growth factor-beta1 production, fibrotic lung disease, and graft fibrosis after lung transplantation Transplantation 1998 66 1014 1020 9808485 10.1097/00007890-199810270-00009
Yamada Y Association of polymorphisms of the transforming growth factor-beta1 gene with genetic susceptibility to osteoporosis Pharmacogenetics 2001 11 765 771 11740340 10.1097/00008571-200112000-00004
Njeh CF Boivin CM Langton CM The role of ultrasound in the assessment of osteoporosis: a review Osteoporos Int 1997 7 7 22 9102067 10.1007/BF01623454
Howard GM Nguyen TV Harris M Kelly PJ Eisman JA Genetic and environmental contributions to the association between quantitative ultrasound and bone mineral density measurements: a twin study J Bone Miner Res 1998 13 1318 1327 9718201
Beamer WG Donahue LR Rosen CJ Baylink DJ Genetic variability in adult bone density among inbred strains of mice Bone 1996 18 397 403 8739896 10.1016/8756-3282(96)00047-6
Chakravarti A Population genetics--making sense out of sequence Nat Genet 1999 21 56 60 9915503 10.1038/4482
|
15955247
|
PMC1182375
|
CC BY
|
2021-01-04 16:32:05
|
no
|
BMC Musculoskelet Disord. 2005 Jun 14; 6:29
|
utf-8
|
BMC Musculoskelet Disord
| 2,005 |
10.1186/1471-2474-6-29
|
oa_comm
|
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-391600883510.1186/1471-2474-6-39Research ArticleDoes Alendronate reduce the risk of fracture in men? A meta-analysis incorporating prior knowledge of anti-fracture efficacy in women Sawka Anna M [email protected] Alexandra [email protected] Jonathan D [email protected] Amiram [email protected] David A [email protected] Lehana [email protected] Department of Medicine, St. Joseph's Healthcare and McMaster University, Hamilton, Ontario, Canada2 Division of Endocrinology and Metabolism, McMaster University, Hamilton, Ontario, Canada3 Graduate Student in Health Research Methodology, McMaster University, Hamilton, Ontario, Canada4 Department of Medicine, Hamilton Health Sciences and McMaster University, Hamilton, Ontario, Canada5 Division of Geriatrics, Hamilton Health Sciences and McMaster University, Hamilton, Ontario, Canada6 Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada7 Division of Rheumatology, St. Joseph's Healthcare and McMaster University, Hamilton, Ontario, Canada8 Division of Endocrinology and Metabolism, Department of Medicine, University of Calgary, Calgary, Alberta, Canada9 Centre for Evaluation of Medicines, St. Joseph's Healthcare, Hamilton, Ontario, Canada2005 11 7 2005 6 39 39 27 9 2004 11 7 2005 Copyright © 2005 Sawka et al; licensee BioMed Central Ltd.2005Sawka 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
Alendronate has been found to reduce the risk of fractures in postmenopausal women as demonstrated in multiple randomized controlled trials enrolling thousands of women. Yet there is a paucity of such randomized controlled trials in osteoporotic men. Our objective was to systematically review the anti-fracture efficacy of alendronate in men with low bone mass or with a history of prevalent fracture(s) and incorporate prior knowledge of alendronate efficacy in women in the analysis.
Methods
We examined randomized controlled trials in men comparing the anti-fracture efficacy of alendronate to placebo or calcium or vitamin D, or any combination of these. Studies of men with secondary causes of osteoporosis other than hypogonadism were excluded. We searched the following electronic databases (without language restrictions) for potentially relevant citations: Medline, Medline in Process (1966-May 24/2004), and Embase (1996–2004). We also contacted the manufacturer of the drug in search of other relevant trials. Two reviewers independently identified two trials (including 375 men), which met all inclusion criteria. Data were abstracted by one reviewer and checked by another. Results of the male trials were pooled using Bayesian random effects models, incorporating prior information of anti-fracture efficacy from meta-analyses of women.
Results
The odds ratios of incident fractures in men (with 95% credibility intervals) with alendronate (10 mg daily) were: vertebral fractures, 0.44 (0.23, 0.83) and non-vertebral fractures, 0.60 (0.29, 1.44).
Conclusion
In conclusion, alendronate decreases the risk of vertebral fractures in men at risk. There is currently insufficient evidence of a statistically significant reduction of non-vertebral fractures, but the paucity of trials in men limit the statistical power to detect such an effect.
==== Body
Background
Osteoporosis increases the risk of fragility fracture in both genders. In a population-based study of Canadians age 50 years by the Canadian Multicenter Osteoporosis Study Group, the prevalence of vertebral fractures was found to be 23.5% in men and 21.5% in women [1]. Alendronate, a potent oral bisphosphonate, decreases the risk of fractures in postmenopausal women with low bone mass or prevalent fractures, as established in a recent meta-analyses examining outcomes in thousands of women [2-4]. Less is known about the effect of alendronate in men, due to a paucity of randomized controlled trials. However, osteoporotic fractures are common in aging men; in fact the lifetime risk of a fracture of the spine, hip or distal radius is 13% for white men older than 50 years [5]. Our objective was to determine whether alendronate decreases risk of vertebral and non-vertebral fractures in men.
Upon initiating our review, we were aware of the paucity of trials examining anti-fracture efficacy of bisphosphonates in men. However, given that bisphosphonates decrease osteoclastic resorption in a mechanism independent of sex steroid status [6], we believed that the effect of alendronate in men would be similar to that previously observed in women. Therefore, the anti-fracture efficacy of alendronate in women would be relevant prior information to be incorporated in assessing treatment effects in men. Classical, frequentist, statistical methods do not offer the flexibility to incorporate relevant prior knowledge or beliefs in analysis of data, thus we sought an alternative statistical approach to analyze the results of our systematic review and we turned to Bayesian methods.
Bayesian statistical methods can explicitly and quantitatively incorporate relevant prior evidence in health technology assessment [7]. The foundation of Bayesian statistical methodology is Bayes' theorem, which is "a formula that shows how existing beliefs, formally expressed as probability distributions, are modified by new information" [8]. In Bayesian methodology, the conclusions of the analysis (known as the "posterior" inferences) are a result of modification of the "prior" data (in this case, known anti-fracture efficacy of alendronate in post-menopausal women), by new data collected (known as the "likelihood function", in this case, the data collected in men) [9]. Bayesian methodology is similar to clinical practice, as typically a clinician has a strong "prior" belief of, for example, a diagnosis such as osteoporosis prior to diagnostic testing, based on the clinical profile of the patient (such as age, gender, risk factors) and the results of diagnostic testing (analogous to a "likelihood function"), are used to confirm or refute those clinical suspicions and formulate a final conclusion (analogous to a "posterior" inference). Thus, Bayesian approaches are clinically intuitive. Moreover, results of Bayesian analyses are also more easily clinically translated than those of frequentist analyses [8]. For example, a Bayesian result tells us how likely is the result (such as an odds ratio), given the data [8]; as opposed to a frequentist result, which tell us how likely are the data given the null hypothesis. For these reasons, Bayesian methodology may be incorporated in healthcare medical decision-making [9]. Justification for the use of Bayesian approach in this study is the ability to directly answer the clinically relevant question: how likely is an osteoporotic man treated with alendronate to be protected from fracture given the current evidence in men and prior evidence in women? A classical frequentist analysis does not allow the flexibility to incorporate prior relevant information in the analysis and all relevant data from women would be ignored in such an analysis. Thus, a Bayesian approach was chosen as the primary analysis method for our study.
Methods
Design of the systematic review, inclusion and exclusion criteria
A systematic review of randomized-controlled trials was performed to determine the efficacy of alendronate (≥ 5 mg daily) in preventing vertebral and non-vertebral fractures in adult men with low bone mass or prevalent fracture(s). As a requirement for inclusion for each study, at least half the study population was required to consist of men, and the required minimum trial follow-up period was one year. A control group could be treated with calcium or vitamin D (any formulation, any dose), but no other active anti-osteoporotic therapies. In the case of duplicate publications on the same study population, only the study with the largest population size and longest follow-up period was included in our review. Studies were excluded if fracture outcome data were not provided, or if they included patients with bone or mineral disorders other than primary osteoporosis, with the exception of osteoporosis secondary to hypogonadism. Our focus was primarily on intention-to-treat data as such analyses are thought to avoid selection bias [10].
Data sources and abstraction
One author searched the following electronic databases (without language restrictions): Medline, Medline-in-Process (1966-May 24/2004), Embase (1996–2004), and the Cochrane Central Register of Controlled Trials (1800 to May, 2004). The search terms used included the textwords "alendronate" or "fosamax", and the search was restricted to articles pertaining to adult male humans (age ≥ 19 years), clinical trials, controlled clinical trials, and randomized controlled trials (textwords "male" or "men" added in Embase) (Appendix). All citations were reviewed by two co-authors and citations which were deemed potentially relevant by either reviewer were examined in full by both investigators. Consensus was reached by both reviewers on all included studies. One co-author performed a PubMed search on the primary author of each included study to look for updated information on the included trials (such as updated follow-up data) or additional relevant trials and this search was restricted to studies including the text word "alendronate" or "fosamax" (up to May 24, 2004). The co-author who performed this secondary search examined the retrieved references and reviewed any potentially relevant trials with a second investigator. Consensus was again reached by both reviewers on whether any trial updates or new trials would be included. Merck-Frosst representatives at the Canadian and world-wide headquarters were contacted for information on any other potentially relevant trials not found by the electronic search. Male trial data were abstracted by one co-author and was checked by another. Fracture outcomes of individual trials in men were summarized using an intention-to-treat analysis principle. For randomized controlled trials meeting inclusion criteria in which fracture data were not published, primary authors were contacted by electronic mail for supplemental information.
Statistical analyses and assumptions
The Bayesian approach was chosen because it allows formal incorporation of prior information in the analysis [7]. Bayesian random effects models predicting the log odds of vertebral fracture and non-vertebral fracture in alendronate-treated men incorporated prior information from published meta-analyses in postmenopausal women [2]. The effect of alendronate on fracture outcomes was assumed to be similar in both genders. In estimating the priors, we noted that in published random effects meta-analyses of postmenopausal women, the relative risk of vertebral fracture with alendronate therapy (≥ 5 mg daily) was 0.52 (95% confidence intervals [CI], 0.43, 0.65) (9360 women in eight studies); whereas the relative risk of non-vertebral fracture (alendronate ≥ 10 mg daily) was 0.51 (95% CI, 0.38, 0.69) (3723 women in six studies) [2]. For the purpose of our analyses, each relative risk was assumed to approach an odds ratio (OR), based on the assumption that relative risks and odds ratios are similar if the risk of the disease is very small, which was the case given the large sample sizes and low event rates. For each prior we assumed the normal distribution on log (OR) scale. Thus, the mean log OR was assumed to be -0.65 with variance 0.19 (precision 5.2) for vertebral fractures and -0.67 with variance 0.24 (precision 4.1) for non-vertebral fractures, respectively for alendronate-treated women. For the study heterogeneity, we assumed a Gamma prior distribution, determined on the basis that the mean heterogeneity between k studies is expected to be equal to the degrees of freedom, k-1 [21]. Thus, in calculating priors for study heterogeneity, we determined that a = 3.5 and b = 0.5 with k-1 = 7 for the analysis of vertebral fracture data; furthermore, a = 2.5 and b = 0.5 with k-1 = 5 for the analysis of non-vertebral fracture data.
WinBUGS version 1.4 (MRC Biostatistics Unit, Cambridge, UK) was used for all Bayesian statistical analyses. The prior knowledge of treatment effects of alendronate in women ("prior distribution") was incorporated in a hierarchical model and then the "likelihood function" of data in men was incorporated in this model for the same outcome. In order to transform the "prior" and "likelihood function" data to a "posterior" inference (final result), simulations were performed using Markov chain Monte Carlo Methods [9]. For each outcome, we performed 20,000 simulations, with Gibbs sampling of results for posterior distributions started at 2,500, after convergence was achieved in all models. A posterior distribution of the treatment effect of the results of alendronate on the log (OR) and inverse log (OR) scales was obtained for the outcomes of vertebral and non-vertebral fractures, respectively. The results were expressed as the posterior mean with corresponding 95% credibility interval [CRI]) on the inverse log (OR) scale (the latter interpreted as analogous to the OR). Using the traditional definition of 95% credibility interval, there is a 95% probability that the true treatment effect, in this case, the odds ratio, lies within this interval [20]. Comparison of the Bayesian results with the classical or frequentist random effects results using Review manager 4.1 (Cochrane Collaboration, Oxford, UK) was performed. The frequentist analysis did not incorporate any prior assumptions based on anti-fracture efficacy of alendronate in women and odds ratios of treatment effects were expressed as OR with the corresponding 95% confidence interval (CI).
Results
Results of search for relevant studies
We obtained 103 unique references from our primary electronic searches of Medline/Medline-in-Process and Embase; no additional unique references were found using the Cochrane Central Register of Controlled Trials. No additional studies were revealed by Merck-Frosst. An additional six unique citations were examined upon searching under the name of the primary author of included trials. Nine of the references retrieved by our electronic searches were considered potentially relevant and reviewed in full-text form by two authors [11-19]. Only studies by Orwoll et al. [11] and Ringe et al. [12] (with information on numbers of patients with vertebral fractures from Orwoll documented in a subsequent paper [13]), met all inclusion criteria. The following studies, were excluded after review of full-text for the following reasons: Ringe 2002 (overlap of included study) [14], Gonnelli 2003 (lack of fracture outcomes, no additional information provided after attempting to contact primary author by electronic mail) [15], Finkelstein 2003 (active control of parathyroid hormone) [16], van der Poest 2002 (study of bone density changes in the setting of acute lower extremity fracture with immobilization) [17], Kushida 2002 (men comprised only 4% of the study population) [18], Ho 2000 (lack of control group) [19]. Details of the process of exclusion of trials are shown in Figure 1.
Figure 1 Process of exclusion of studies. The trials included at the final step are 1) Orwoll E et al. Alendronate for the treatment of osteoporosis in men. N Engl J Med 2000;343(9):604-10; and 2) Ringe JD., et al. Alendronate treatment of established primary osteoporosis in men: 3-year results of a prospective, comparative, two-arm study. Rheumatol Int 2004;24(2):110-3; with supplemental information on the Orwoll trial obtained from the review article, 3) Ringe JD, Orwoll E, et al. Treatment of male osteoporosis: recent advances with alendronate. Osteoporos Int 2002; 13(3):195-9. The two-year analysis of the Ringe trial was excluded because of overlap of patients with the 3-year analysis: Ringe JD, et al. Alendronate treatment of established primary osteoporosis in men: results of a 2-year prospective study. J Clin Endocrinol Metab 2001;86(11):5252–5255.
Summary of alendronate trials in men
Characteristics of the included randomized controlled studies in men are shown in Table 1. Hypogonadal men were included in the Orwoll study [11] but not the Ringe study [12]. Approximately half of the men in each study had a history of one or more prevalent vertebral fractures before enrollment in the trial [11,12]. In the Orwoll study, an incident vertebral fracture was defined using the semi-quantitative method of Genant [22], as assessed by a radiologist blinded to treatment assignment at the University of California in San Francisco [11]. In the Ringe study, a vertebral fracture was defined by a new decrease in vertebral height of at least 20%, as assessed by a radiologist blinded to treatment assignment [12]. The Orwoll study was two years in duration [11], whereas the Ringe study extended for three years [12]. In both trials, the active treatment group received 10 mg of alendronate as well as 500 mg supplemental calcium orally daily [11,12]. In the Orwoll study, men in the treatment and control group received 400–450 IU vitamin D daily [11], whereas in the Ringe study, only men in the control arm received supplemental vitamin D in the form of 1 μg alfacalcidiol daily [12].
Table 1 Characteristics of included studies
Study (reference) Sample (ALN*/ control) and duration Inclusion criteria Age (years) (SD‡) Percentage prevalent vertebral fractures (VF, %) Intervention/ Control Blinding, randomization Loss to follow-up or withdrew from study (n/N) (%)
Orwoll 2000 [11] 146/95 (2 years) Men with BMD† T-score = -2 at femoral neck and T-score = -1 at the lumbar spine;
OR Men with T-score = -1 at the femoral neck and at least one vertebral or osteoporotic fracture ALN:
Mean age 63 (13)
49% VF
Control: Mean age 63 (12)
52% VF ALN*: 10 mg + 500 mg Calcium + 400–450 IU Vitamin D
Control: 500 mg Calcium + 400–450 IU Vitamin D - Double-blind
- Radiologists reading vertebral x-rays blinded to intervention
-Method of randomization unclear 38/241 (15.8%)
Ringe 2004 [12] 68/66 (3 years) Men with BMD† T-score = -2.5 at femoral neck or lumbar spine, excluding hypogonadal men ALN*:
Mean age 52.7 (11.1)
54% VF
Control:
Mean age 53.3 (10.9)
53% VF ALN*: 10 mg + 500 mg Calcium Control: 500 mg Calcium + 1 μg alfacalcidiol - Open-label
- Radiologists reading vertebral x-rays blinded to intervention
-Method of randomization unclear 16/134 (11.9%)
*ALN, alendronate (daily dose)
\dagBMD, bone mineral density measurement by dual X-ray absorptiometry, compared to young adult male peak bone mass
\ddagSD, Standard deviation
In the Orwoll study, 4/146 alendronate-treated (2.7%) and 7/95 placebo-treated (7.4%) men sustained a vertebral fracture at two years [11,13]. In the Ringe study, 7/68 alendronate-treated men (10.3%) and 16/66 (24.2%) control participants sustained a vertebral fracture at three years [12]. Furthermore, In the Orwoll study, 6/146 alendronate-treated men (4.1%) and 5/95 men (5.3%) placebo-treated men sustained a non-vertebral fracture [11]. In the Ringe study, 6/68 men (8.8%) alendronate-treated men and 8/66 (12.1%) control participants sustained a non-vertebral fracture [12].
Results of pooled analyses
Upon incorporating the prior information from postmenopausal women with the male data in the Bayesian random effects model, the OR of vertebral fracture in alendronate-treated men was 0.44 (95% CRI 0.23, 0.83) and the OR of non-vertebral fracture was 0.60 (95% CRI 0.29, 1.44). Similar results were found using the frequentist random effects models of male data (vertebral fracture OR 0.36, 95% CI 0.17, 0.77; non-vertebral fracture OR 0.73, 95% CI, 0.32, 1.67), without incorporation of any of the data from women (Figure 2).
Figure 2 Pooled random effects meta-analyses of vertebral and non-vertebral fracture outcomes with alendronate treatment (frequentist models). OR, Odds Ratio; 95% CI Random, 95% confidence interval using a frequentist random effects model for pooling the data from studies
Discussion
Limitations of our systematic review include the paucity of trial data from men, the small sample sizes in trials, the variations in trial duration between studies, and the inconsistency of calcium and vitamin D formulations between trials (with the possibility that the alfacalcidiol used in the Ringe study could be considered a form of active therapy [23]). We were also unable to perform a per protocol sensitivity analysis of the anti-fracture efficacy of alendronate in men who were compliant with therapy as these data were not published in the primary trials. We did not include unpublished data. We also did not contact author to clarify details of the randomization procedures. Merits of our study include its systematic nature, including both Bayesian and traditional frequentist analyses of available data, and the examination of clinically relevant fracture outcomes. Of note, the numerical estimates of odds ratios and their respective credibility or confidence intervals were similar using Bayesian and frequentist analyses in this study. These findings are not surprising, given that the treatment effects of alendronate in women, from whom prior information was derived, were similar to those observed in men. The precision of our estimate was however slightly improved using a Bayesian approach for the outcome of non-vertebral fractures as seen by the slightly narrower credibility interval than confidence interval for that outcome. Credibility intervals can be narrower using a Bayesian approach than confidence intervals obtained using a frequentist approach because of additional data provided by the priors [9].
Given that the results of our Bayesian and frequentist meta-analyses were similar, was there any advantage to the Bayesian approach? The basic question posed using a frequentist approach is how likely are the data given the null hypothesis? In contrast, the question posed from a Bayesian perspective is, how likely is the odds ratio, given the data [8]? Furthermore, the interpretation of a traditional 95% confidence interval is that given a long series of such intervals, 95% of them should contain the true value of the odds ratio [9]. In contrast, in interpreting a 95% credibility interval, there is a 95% probability that the true value of the parameter lies within this interval [9]. Thus, the clinical question posed and the interpretation of the probability interval, are more clinically intuitive from a Bayesian perspective. Moreover, the Bayesian approach allowed the flexibility to incorporate clinically relevant evidence from trials in women in this meta-analysis of alendronate treatment in men.
In the future, more trials of active osteoporotic therapies and their effects on fracture outcomes need to be performed in both genders.
Conclusion
We conclude that alendronate (10 mg daily) decreases the risk of vertebral fractures in men with low bone mass or fractures, but there is currently insufficient evidence to prove a significant effect on non-vertebral fractures in this population. The paucity of osteoporosis therapy trials in men as well as the relative infrequency of non-vertebral fractures in the control groups may have restricted our statistical power to detect a significant reduction of non-vertebral fractures. Of note, the relative risk of vertebral fracture in alendronate-treated men was similar to that previously observed in a meta-analysis of data from postmenopausal women [2].
Appendix
Electronic search strategy for potentially relevant studies
For Medline and Medline-in-Process (1966 to May 24, 2004), Cochrane Central Register of Controlled Trials (1800 to May 24, 2004): "alendronate" or "fosamax" - restricted to adult male humans (age ≥ 19 years), clinical trials, controlled clinical trials, and randomized controlled trials.
For Embase 1996 to May 24, 2004: "alendronate" or "fosamax" AND "male" or "men"- restricted to adult male humans (age ≥ 19 years), clinical trials, controlled clinical trials, and randomized controlled trials.
For PubMed (up to May 24, 2004): name and initials of authors of included studies AND "alendronate" or "fosamax"
Competing interests
Dr. Sawka is a Skeletal Health Scholar funded, in part, by the Canadian Institutes of Health Research. Dr. Sawka is also a Fellow in Health Economics at McMaster University, partly funded by an unrestricted educational grant from Hoffmann-La Roche.
Dr. Papaioannou's competing interests include: Aventis, Eli Lilly, Merck, Novartis, and Procter and Gamble.
Dr. Adachi, Consultant to: Amgen, Astra Zeneca, Aventis, Eli Lilly, Glaxo Smith Kline, Merck, Novartis, Procter and Gamble, and Hoffman-La Roche.
Dr. Hanley's competing interests include consultancies with, honoraria for speaking from, or involvement in research with, the following companies or organizations : Amgen, Astra-Zeneca, Aventis, the Dairy Farmers of Canada, Eli Lilly, Merck, Novartis, NPS Pharmaceuticals, Pfizer, Procter and Gamble, Hoffman-La Roche, and Wyeth.
The other co-authors have no competing interests to declare.
Authors' contributions
All co-authors reviewed the manuscript and made suggestions for revisions. The project idea was conceived by Dr. Sawka. Analyses were performed by Dr. Sawka, with input from Dr. Thabane. The manuscript was written and revised by Dr. Sawka.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
A.M. Sawka is a Skeletal Health Scholar funded by the Canadian Institutes of Health Research and a Fellow in Health Economics at McMaster University partly sponsored by an unrestricted educational grant from Hoffmann-La Roche.
==== Refs
Jackson SA Tenenhouse A Robertson L Vertebral fracture definition from population-based data: preliminary results from the Canadian Multicenter Osteoporosis Study (CaMos) Osteoporos Int 2000 11 680 687 11095171
Cranney A Wells G Willan A Griffith L Zytaruk N Robinson V Black D Adachi J Shea B Tugwell P Guyatt G Meta-analyses of therapies for postmenopausal osteoporosis. II. Meta-analysis of alendronate for the treatment of postmenopausal women Endocr Rev 2002 23 508 16 12202465
Cranney A Tugwell P Wells G Guyatt G Meta-analyses of therapies for postmenopausal osteoporosis. I. Systematic reviews of randomized trials in osteoporosis: introduction and methodology Endocr Rev 2002 23 496 507 12202464
Cranney A Guyatt G Griffith L Wells G Tugwell P Rosen C Meta-analyses of therapies for postmenopausal osteoporosis. IX: Summary of meta-analyses of therapies for postmenopausal osteoporosis Endocr Rev 2002 23 570 578 12202472
Riggs BL Melton LJ III The worldwide problem of osteoporosis: insights afforded by epidemiology Bone 1995 17 505S 511S 8573428
Bell NH Johnson RH Bisphosphonates in the treatment of osteoporosis Endocrine 1997 6 203 206 9225137
Spiegelhalter DJ Myles JP Jones DR Abrams KR Bayesian methods in health technology assessment: a review Health Technol Assess 2000 4 1 130 11134920
Spiegelhalter DJ Myles JP Jones DR Abrams KR Methods in health service research: An introduction to bayesian methods in health technology assessment BMJ 1999 319 508 512 10454409
Spiegelhalter DJ Abrams KR Myles JP Senn S and Barnett V Bayesian Approaches to Clinical Trials and Health-Care Evaluation 2004 First West Sussex: John Wiley and Sons 1 391
Juni P Altman DG Egger M Egger M, Smith GD, Altman DG Assessing the Quality of Randomised Controlled Trials Systematic Reviews in Health Care: Meta-analysis in Context 2001 London, UK: BMJ Books 87 108
Orwoll E Ettinger M Weiss S Miller P Kendler D Graham J Adami S Weber K Lorenc R Pietschmann P Vandormael K Lombardi A Alendronate for the treatment of osteoporosis in men N Engl J Med 2000 343 604 10 10979796
Ringe JD Dorst A Faber H Ibach K Alendronate treatment of established primary osteoporosis in men: 3-year results of a prospective, comparative, two-arm study Rheumatol Int 2004 24 110 3 13680141
Ringe JD Orwoll E Daifotis A Lombardi A Treatment of male osteoporosis: recent advances with alendronate Osteoporos Int 2002 13 195 9 11991437
Ringe JD Faber H Dorst A Alendronate treatment of established primary osteoporosis in men: results of a 2-year prospective study J Clin Endocrinol Metab 2001 86 5252 5255 11701687
Gonnelli S Cepollaro C Montagnani A Bruni D Caffarelli C Breschi M Gennari L Gennari C Nuti R Alendronate treatment in men with primary osteoporosis: a three-year longitudinal study Calcif Tissue Int 2003 73 133 139 14565594
Finkelstein JS Hayes A Hunzelman JL Wyland JJ Lee H Neer RM The effects of parathyroid hormone, alendronate, or both in men with osteoporosis N Engl J Med 2003 349 1216 1226 14500805
van der Poest CE van Engeland M Ader H Roos JC Patka P Lips P Alendronate in the prevention of bone loss after a fracture of the lower leg J Bone Miner Res 2002 17 2247 2255 12469919
Kushida K Shiraki M Nakamura I Kishimoto H Morii H Yamamoto K Kaneda K Fukunaga M Inoue T Nakashima M Orimo H The efficacy of Alendronate in reducing the risk for vertebral fracture in Japanese patients with osteoporosis: A randomzed, double-blind, active-controlled, double-dummy trial Curr Ther Res Clin E 2002 63 606 620
Ho YV Frauman AG Thomson W Seeman E Effects of alendronate on bone density in men with primary and secondary osteoporosis Osteoporos Int 2000 11 98 101 10793867
Utts JM Utts JM Chapter 2 Seeing Through Statistics 1999 Second Pacific Grove, CA: Brooks-Cole/Duxbury Press 210
Rice JA Mathematical Statistics and Data Analysis 1995 Second Belmont, CA: Duxbury Press 178
Genant HK Assessment of vertebral fractures in osteoporosis research J Rheumatol 1997 24 1212 1214 9195538
Papadimitropoulos EA Wells G Gillespie W Weaver B Zytaruk N Cranney A Adachi J Tugwell P Josse R Greenwood C Guyatt G Meta-analyses of therapies for postmenopausal osteoporosis. VII: Meta-analysis of the efficacyof vitamin D treatment in preventing osteoporosis in postmenopausal women Endocr Rev 2002 23 560 569 12202471
|
16008835
|
PMC1182376
|
CC BY
|
2021-01-04 16:32:05
|
no
|
BMC Musculoskelet Disord. 2005 Jul 11; 6:39
|
utf-8
|
BMC Musculoskelet Disord
| 2,005 |
10.1186/1471-2474-6-39
|
oa_comm
|
==== Front
BMC Oral HealthBMC Oral Health1472-6831BioMed Central London 1472-6831-5-41598516610.1186/1472-6831-5-4Research ArticleOral clefts with associated anomalies: findings in the Hungarian Congenital Abnormality Registry Sárközi Andrea [email protected] Diego F [email protected] Andrew E [email protected] Foundation for the Community Control of Hereditary Diseases, Torokvesz lejto 32, 1026 Budapest, Hungary2 Departments of Medicine and of Epidemiology, Boston University School of Medicine, 715 Albany Street, L-320, Boston, MA 02118, USA2005 28 6 2005 5 4 4 3 2 2005 28 6 2005 Copyright © 2005 Sárközi et al; licensee BioMed Central Ltd.2005Sárközi 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 years, great efforts have been made to record the frequency of orofacial clefts in different populations. However, very few studies were able to account for the etiological and phenotypic heterogeneity of these conditions. Thus, data of cases with syndromic orofacial clefts from large population-based studies are infrequent.
Methods
Clinically recognized and notified syndromes and associations including cleft lip with or without cleft palate and other congenital anomalies were selected from the Hungarian Congenital Abnormality Registry (HCAR) between 1973 and 1982 and prevalence rates were calculated.
Results
Of 3,110 cases reported as having orofacial clefts, 653 had multiple congenital abnormalities. Of these, 60 (9.2%) had a known etiology (monogenic: 25 or 3.8%, chromosomal: 31 or 4.7%, teratogenic: 4 or 0.6%). Seventy-three subjects (11.2%) had schisis in addition to the oral cleft. Skeletal anomalies were the most common malformations among cases with cleft lip with/without cleft palate (CL/P) and cleft palate (CP). Disorders of the central nervous system and cardiovascular malformations were also frequently associated.
Conclusion
Surveillance systems, such as the HCAR, provide useful information about prevalence rates of congenital anomalies in a population. However, in a field where new syndromes are being discovered and classifications regularly updated, these rates should only be accepted as provisional.
==== Body
Background
It has been known for more than 80 years that cleft lip with or without cleft palate (CL/P) and isolated cleft palate (CP), collectively termed oral clefts (OCs), are frequently associated with congenital anomalies [1]. The prevalence of associated anomalies in subjects with OCs varies widely, ranging from 6% to 63%; however, when broken down by subtype, it is clear that they are much more frequent in patients with isolated CP (13–50%) than in patients with cleft lip (CL) (7%-13%) or patients with cleft lip and palate (CLP) (2%-26%) [1-7]. The sources of variation have been recently described by Wyszynski et al. [1] as 1) differences in case definition and inclusion/exclusion criteria, 2) how long after birth cases are examined, 3) variability of clinical expression of associated anomalies, 4) knowledge and technology available to produce syndrome delineation, 5) selection of patients, sources of ascertainment, and sample size, 6) true population differences and changes in frequency over time.
The evaluation of patients with multiple congenital anomalies (MCAs) is of critical importance because 1. all unbalanced autosomal chromosomal aberrations and most gene mutations and teratogens produce syndromes. Therefore, MCAs are sensitive indicators of germinal mutagens and teratogens [8]. 2. The delineation of an MCA-entity facilitates a better understanding of the phenotypic spectrum, prognosis, and origin of the condition. The latter may be of great importance in genetic counseling or to detect new teratogenic agents. 3. Gene mapping efforts for some of these conditions might become feasible after the identification of informative families.
The objective of this paper is to describe the cases with OCs and associated anomalies identified in a large population-based birth defects registry in Hungary.
Methods
Eligible cases were newborns with OCs and at least one other congenital anomaly identified from the records of the nation-based "Hungarian Congenital Abnormality Registry" (HCAR; [9]) and born between 1973 and 1982. No data were collected after 1982. Notification by physicians of cases with structural birth defects (i.e., congenital abnormalities) to the HCAR was mandatory during that time period. Most cases were notified by obstetricians, since in Hungary virtually all deliveries occur in inpatient obstetric clinics, or from pediatricians who were working in the neonatal units of inpatient obstetric clinics and various inpatient and outpatient pediatric clinics. During the study period, autopsy was required for all infant deaths, and was often practiced for stillborn fetuses. Pathologists sent a copy of each autopsy report to the HCAR if any birth defects were identified. The recorded total (birth + fetal) prevalence of cases with CL/P and with isolated CP was 1.01 and 0.35 per 1,000 newborns (liveborn infants, stillborn and malformed fetuses from electively-terminated pregnancies), respectively. About 95% of cases with CL/P and close to 90% of cases with isolated CP were reported to the HCAR during the 10 years of the study period [9]. Based on the available clinical notations, each case was assigned to one of three categories:
1) Unspecified multiple congenital anomalies. These cases had such limited information that it was not possible to differentiate the type of malformations present. New or supplementary information was requested from the clinicians. Of 6,641 total cases with multiple congenital anomalies reported to the HCAR, 131 subjects (2.0%) were in this category.
2) Unidentified (but specified) multiple congenital anomalies. These cases had information about the associated anomalies, but it was not possible to distinguish between syndrome, sequence, and association. Several attempts were made to remediate this situation: (a) a copy of the detailed necropsy records was requested from the pathologist when the case expired (stillborn or infant death). Pathology records were received in 88% of these cases; (b) cases that listed congenital dislocation of the hip (mainly hip dysplasia or Ortolani positivity), club-foot, congenital inguinal hernia, or some other mild congenital anomaly were also part of a study on postural congenital anomalies [10], which provided additional information; (c) other cases with multiple congenital anomalies who survived the infant period were officially referred to one of the eight Multiple Congenital Anomalies Examination Centers, which functioned as part of the HCAR. Each center had a circumscribed catchment area and was equipped with laboratory facilities suitable to carry out chromosome analysis, serological examinations (ie, for rubella and cytomegalovirus, toxoplasmosis, etc.), and certain biochemical tests. Clinical examination by a dysmorphologist was combined with multidisciplinary counseling concerning possible treatment and prognosis for the patient, as well as regarding the risk of recurrence in further pregnancies [11]; (d) occasionally, the records had sufficient information, but no diagnosis. Of 17 cases with OCs, four cases were eventually identified as having a syndrome (Ectrodactyly-Ectodermal Dysplasia-CL/P, Meckel, Mohr, and Roberts) and one had an association (schisis) [8]. Finally, (e) cases born between 1980 and 1982 were studied further in the context of a separate epidemiological study carried-out between 1982 and 1983. A structured questionnaire and an explanatory letter were mailed to the parents of these cases soliciting information on their children's condition. The information gathered made it possible to confirm or modify some diagnoses. Of 6,641 total cases with multiple congenital anomalies reported to the HCAR, 3,393 subjects (51%) had unidentified but specified MCAs.
3) Identified syndromes or associations. These cases were accepted without any further follow-up on the basis of the clinical records. Of 6,641 total cases with multiple congenital anomalies reported to the HCAR, 3,117 subjects (47%) were in this category.
Results
The dataset comprised 65,923 cases with congenital anomalies born between 1973 and 1982. The number of livebirths during the study period was 1,667,166, resulting in a prevalence of cases with congenital anomalies of 39.5 per 1,000 or ≈ 4% livebirths. The number of confirmed cases with more than one congenital anomaly was 6,641 (prevalence at birth: 3.98 per 1,000 or 10% of all anomalies). The stillbirth and infant death rates for cases with multiple congenital anomalies were 8.67% and 23.8%, respectively, which are about 10 times higher than the corresponding national figures for the study period.
Of the 6,641 cases with more than one anomaly, 2,341 had a syndrome, 776 an association, 131 were unspecified, and 3,393 were unidentified. Of the 3,110 cases with OCs, 2,457 or ≈ 80% were isolated OCs (nonsyndromic and sequences) and 653 had multiple congenital anomalies (syndromic and associations) (Table 1). In the latter group, 60 (9.2%) had a known etiology (monogenic: 25 or 3.8%, chromosomal: 31 or 4.7%, teratogenic: 4 or 0.6%) (Table 2). There were 73 cases (11.2%) with OCs associated to schisis. The remaining 520 cases with OCs and other anomalies (351 CL/P and 169 CP) were unidentified (verba, of unknown etiology).
Table 1 Cases with isolated oral clefts (OCs) and with OCs plus other congenital anomalies in the Hungarian Congenital Abnormality Registry (HCAR), 1973–1982.
Category Group Number Prevalence*
Isolated
Cleft lip with or without cleft palate (CL/P) 1,687 1.02
Cleft lip only 607 0.36
Cleft lip and palate 1,080 0.65
Posterior cleft palate only 632 0.38
Robin sequence 99 0.06
ADAM sequence (n = 31)
including atypical oral clefts 10 0.01
Holoprosencephaly (n = 38)
including orofacial clefts 12 0.01
Others (median, oblique, etc.) 17 0.01
Subtotal 2,457 1.47
Multiple Congenital Anomalies
CL/P in recognized entities 83 0.05
CL/P in unidentified entities 351 0.21
CP in recognized entities 48 0.03
CP in unidentified entities 169 0.10
Robin sequence in recognized syndrome 2 0.00
Total 653 0.39
All cases 3,110 1.87
*per 1,000 livebirths
Table 2 Etiology of recognized syndromes and associations in the entire HCAR datatset and among the subjects with OCs.
Etiological Entity Entire Dataset Subjects with OCs
Mendelian Syndromes
Stickler type I* 4 2
Faciogenitopopliteal 2 2
Ectrodactyly, ectodermal dysplasia, CL/P 6 6
Diastrophic displasia 2 1
Meckel 28 4
Orofaciodigital type II 5 5
Roberts 4 4
Orofaciodigital type I 9 1
Subtotal 60 25
Chromosomal
Trisomy 13 35 29
Trisomy 18 22 1
Deletions 25 1
Subtotal 82 31
Teratogenic
Hydantoin 4 4
Associations
Schisis 130 73
Total 276 133
*Stickler syndrome included Robin sequence
Most cases with unidentified syndromic orofacial clefts had a total of 2 malformations. There were 181 subjects with unidentified syndromic CL/P and only 1 other malformation (or 51.6% of all unidentified syndromic CL/P). Similarly, there were 81 subjects with unidentified syndromic CP and only 1 other malformation (or 47.9% of all unidentified syndromic CP) (Table 3). In unidentified syndromic CL/P cases, the most frequent combinations were with anomalies of the heart or circulatory system (n = 37, rate: 7.2/100,000 births), club foot (n = 21, rate: 3.1/100,000 births), congenital hydrocephaly (n = 22, rate: 3.0/100,000 births), and polydactyly in hand or foot (n = 18, rate: 3.0/100,000 births). The most common combinations in those with unidentified syndromic CP were anomalies of the heart or circulatory system (n = 24, rate: 4.5/100,000 births), club foot (n = 16, rate: 2.4/100,000 births), congenital anomalies of the ear (n = 4, rate: 1.4/100,000 births), and anomalies of the skeletal system, especially spine, ribs, and sternum (n = 3, rate: 1.1/100,000 births).
Table 3 Frequency (and percentage) of anomalies in cases with non-isolated CL/P and CP of unidentified etiology.
Total Number of Anomalies Unidentified Etiology CL/P Unidentified Etiology CP
2 181 51.6 81 47.9
3 81 23.1 31 18.3
4 40 11.4 28 16.6
5 21 6.0 17 10.1
6 16 4.5 8 4.7
7 or more 12 3.4 4 2.4
Total 351 100.0 169 100.0
There were 81 and 31 patients with a total of three congenital anomalies including syndromic unidentified CL/P and CP, respectively (Table 3). In these cases, the most common coexisting anomalies for CL/P were of the heart and circulatory system (n = 17), polydactyly (n = 13) and reduction of the limbs (n = 12), while only anomalies of the heart and circulatory system were frequent among those patients with CP (n = 11). One hundred and seventy one subjects with unidentified CL/P and 88 with CP had four or more anomalies; however, there were few repeated combinations:
2 cases with CL, congenital heart defect, polydactyly, and hydrocephaly,
2 cases with CP, polydactyly, anorectal atresia/stenosis, and branchial anomalies,
2 cases with CP, hydrocephaly, and anomalies of the skeletal and digestive systems, and
2 cases with CP, anomalies of the diaphragm and of the ear and renal agenesis/dysgenesis.
Table 4 presents the frequency of malformations by affected organ system in children with multiple congenital anomalies and orofacial clefts. Malformations of the skeletal system were the most common in both CL/P and CP subjects, followed by CNS and cardiovascular among the former and cardiovascular and CNS and urogenital in the latter.
Table 4 Frequency (and percentage of the total) of malformations by affected organ systems in subjects with multiple congenital anomalies.
CL/P (n = 436) CP (n = 217) Both (n = 653)
Central nervous system 158 (20.6) 44 (11.1) 202 (17.4)
Eye 36 (4.7) 7 (1.8) 44 (3.8)
Ear 32 (4.2) 23 (5.8) 55 (4.7)
Face-neck 16 (2.1) 22 (5.5) 38 (3.3)
Cardiovascular system 119 (15.6) 75 (18.9) 194 (16.7)
Respiratory system 4 (0.5) 4 (1.0) 8 (0.7)
Digestive system 51 (6.7) 30 (7.5) 81 (7.0)
Urogenital system 88 (11.5) 44 (11.1) 132 (11.3)
Skeletal (including limb deficiency) 191 (25.0) 121 (30.5) 312 (26.8)
Skin 0 (0.0) 1 (0.2) 1 (0.1)
Abdominal wall/diaphragma 46 (6.0) 15 (3.8) 61 (5.2)
Other† 24 (3.1) 11 (2.8) 35 (3.0)
Total number of malformations 765 397 1,162
†includes congenital inguinal hernia.
Discussion
The main objective of this article was to present prevalence and baseline characteristics of cases with syndromic OCs and associated anomalies in the HCAR dataset. The HCAR is an excellent source of cases because it is (1) population-based, (2) from an ethnically homogeneous, well-defined population, (3) with a high recorded birth prevalence of cases with congenital anomalies compared to other registries (approximately 4%) indicating a nearly complete ascertainment, and (4) the clinical diagnoses have a high level of precision [12]. However, the HCAR has two major weaknesses: First, the information on the cases is based on the notification made by several thousand medical doctors who have uneven experience with children with dysmorphic features. Second, the ascertainment covered a period when many of the craniofacial syndromes known today had not been yet delineated.
Of 653 cases with OCs and multiple congenital anomalies, only 133 (20.4%) were part of a known etiological entity (Table 2). Among the 60 Mendelian syndromic cases in the entire HCAR data set, 25 included OCs (41.7%). During the ascertainment period (1973–1982), over 100 syndromes including OCs had been reported in the literature [13]. However, many were not diagnosed and/or notified to the HCAR. Two examples are the Van der Woude syndrome, which is a relatively common autosomal dominant condition with CL/P [14] and is absent in the HCAR and the oculo-auriculo-vertebral spectrum (previously known as Goldenhar syndrome), frequently associated with orofacial clefts, also non-existent in the HCAR. In other cases, there seems to be an over-reporting of certain conditions. For example, there are 28 cases of Meckel syndrome in the HCAR, with 4 cases including OCs, a rare combination.
The proportion of reported chromosomal abnormalities to the HCAR is lower than expected and orofacial clefts are no exception. This might be due to any of the following three facts: (1) although chromosome analysis was recommended in all cases with multiple congenital anomalies, this advice was rarely followed, (2) cases with multiple congenital anomalies had a high perinatal mortality and karyotyping was seldom performed in these cases, and (3) recently developed sensitive chromosomal diagnostic techniques, such as fluorescent in situ hybridization (FISH), were not available during the study period. Therefore, the relatively common deletion 22q11.2 seen in patients with OCs (previously known as velocardiofacial syndrome or Di George syndrome) could not be identified. This might explain our finding of chromosomal anomalies in only 4.7% of the cases with MCAs, significantly less than other population-based studies (ie, Tolarova and Cervenka: 8.8% [7], Stoll et al.: 7.8% [15]).
All four cases caused by teratogens were identified as fetal hydantoin syndrome (Table 2). Fetal alcohol syndrome was notified very rarely to the HCAR. Among cases with diabetic embryopathy, orofacial clefts were not recorded and therefore are not included in this report. The collection and analysis of information on pregnancy history, including maternal drug use, would help identify known or new syndromes caused by teratogens. This was the main motivation for the establishment of the Case-Control Surveillance Program of Congenital Abnormalities in Hungary [9].
Of 130 cases with the schisis association, 55 (42.3%) had CL/P and 16 had CP (12.3%). This association had been originally described by one of us when it was noted that there were cases with neural tube defects (anencephaly, encephalocele, spina bifida cystica), OCs, omphalocele, and diaphragmatic hernia associated with one another far more frequently than at the expected random combination rates [16]. No other association with OCs was found.
The evaluation of these results suggests that the Hungarian registry probably under-reported chromosomal abnormalities. This might be due to the following: first, the HCAR is a cross-sectional registry; thus, its is limited in its ability to obtain detailed clinical descriptions of each infant with a syndrome, including X-rays, karyotypes, or screening for the 22q11.2 deletion. Second, HCAR ascertained cases between 1973 and 1982. The development of improved molecular and cytogenetic tools in the 1980s, which led to the identification of the etiology of many conditions of previously unidentified origin, might account for some misclassification.
Conclusion
The description of component anomalies in cases with multiple congenital anomalies may help identify recognizable entities and delineate new syndromes. This knowledge can be used to better understand the needs of the population (ie, diagnosis, prognosis, counseling) and to develop policies for health care. In order to be successful, birth defects surveillance systems must include experienced dysmorphologists up-to-date with the latest diagnostic tools and definitions [1]. Engagement with the large multinational registries, such as the International Collaborative Research on Craniofacial Anomalies Project supported by the World Health Organization (WHO) [17], would be of benefit to all as well. It behooves the readers to note that in the HCAR dataset, which contains close to 66,000 congenital anomalies (4% of the total live births), almost 10% of these had more than one anomaly and more than half of these could not be allocated to a particular syndrome or association. This in itself points to the need for a global effort to improve the sensitivity and specificity of diagnosis. It is our hope that the information displayed in this paper will contribute to increase that awareness.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AS, DFW, and AEC conceived the study and participated in its design and coordination and drafted the manuscript. AEC coordinated the overall collection of data. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Part of this work was funded by a contract from the Massachusetts Center for Birth Defects Research and Prevention of the Massachusetts Department of Public Health (to DFW) and by a Peer Foundation/Cleft Palate Foundation Etiology Grant (to DFW).
==== Refs
Wyszynski DF Sarkozi A Czeizel A Oral Clefts with Associated Anomalies: Methodological Issues Cleft Palate Craniofac J 2005 In Press
Shprintzen RJ Siegel-Sadewitz VL Amato J Goldberg RB Anomalies associated with cleft lip, cleft palate, or both Am J Med Genet 1985 20 585 595 3993684 10.1002/ajmg.1320200404
Shaw GM Croen LA Curry CJ Isolated oral cleft malformations: associations with maternal and infant characteristics in a California population Teratology 1991 43 225 228 2014485 10.1002/tera.1420430306
Lilius GP Clefts with associated anomalies and syndromes in Finland Scand J Plast Reconstr Surg Hand Surg 1992 26 185 196 1411347
Milerad J Larson O Ph DD Hagberg C Ideberg M Associated malformations in infants with cleft lip and palate: a prospective, population-based study Pediatrics 1997 100 180 186 9240796 10.1542/peds.100.2.180
Cohen Jr. MM Wyszynski DF Syndromes with orofacial clefting Cleft Lip and Palate: From Origin to Treatment 2002 New York , Oxford University Press 53 65
Tolarova MM Cervenka J Classification and birth prevalence of orofacial clefts Am J Med Genet 1998 75 126 137 9450872 10.1002/(SICI)1096-8628(19980113)75:2<126::AID-AJMG2>3.0.CO;2-R
Czeizel AETLTG Multiple congenital abnormalities 1988 Budapest , Akadémiai Kiadó
Czeizel AE Rockenbauer M Siffel C Varga E Description and mission evaluation of the Hungarian case-control surveillance of congenital abnormalities, 1980-1996 Teratology 2001 63 176 185 11320528 10.1002/tera.1032
Pazonyi I Kun A Czeizel A Congenital postural deformity association Acta Paediatr Acad Sci Hung 1982 23 431 445 7170954
Czeizel A Kovacs M Kiss P Mehes K Szabo L Olah E Kosztolanyi G Szemere G Kovacs H Fekete G A nationwide evaluation of multiple congenital abnormalities in Hungary Genet Epidemiol 1988 5 183 202 3262553 10.1002/gepi.1370050305
Czeizel AE Hirschberg J Orofacial clefts in Hungary. Epidemiological and genetic data, primary prevention Folia Phoniatr Logop 1997 49 111 116 9256533
Gorlin RJ Cervenka J Pruzansky S Facial clefting and its syndromes Birth Defects Orig Artic Ser 1971 7 3 49 4950925
Schinzel A Klausler M The Van der Woude syndrome (dominantly inherited lip pits and clefts) J Med Genet 1986 23 291 294 3746828
Stoll C Alembik Y Dott B Roth MP Associated malformations in cases with oral clefts Cleft Palate Craniofac J 2000 37 41 47 10670888 10.1597/1545-1569(2000)037<0041:AMICWO>2.3.CO;2
Czeizel A Schisis-association Am J Med Genet 1981 10 25 35 7294060 10.1002/ajmg.1320100105
WHO International Collaborative Research on Craniofacial Anomalies Project. World Health Organization.
|
15985166
|
PMC1182377
|
CC BY
|
2021-01-04 16:29:57
|
no
|
BMC Oral Health. 2005 Jun 28; 5:4
|
utf-8
|
BMC Oral Health
| 2,005 |
10.1186/1472-6831-5-4
|
oa_comm
|
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-221600461310.1186/1471-2431-5-22Research ArticleVariations in rates of nosocomial infection among Canadian neonatal intensive care units may be practice-related Aziz Khalid [email protected] Douglas D [email protected] Wayne [email protected] Margaret [email protected] Zhenguo [email protected] Stella [email protected] Shoo K [email protected] Canadian Neonatal Network [email protected] Discipline of Pediatrics, Memorial University of Newfoundland, St. John's NF, Canada2 Department of Pediatrics, Dalhousie University, Halifax NS, Canada3 Department of Pediatrics, University of British Columbia, Vancouver BC, Canada4 Centre for Healthcare Innovation and Improvement, Vancouver BC, Canada5 Canadian Neonatal Network Coordinating Center, Vancouver BC, Canada6 Discipline of Pediatrics, Memorial University of Newfoundland, St Johns NF, Canada2005 8 7 2005 5 22 22 20 12 2004 8 7 2005 Copyright © 2005 Aziz et al; licensee BioMed Central Ltd.2005Aziz 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
Nosocomial infection (NI), particularly with positive blood or cerebrospinal fluid bacterial cultures, is a major cause of morbidity in neonatal intensive care units (NICUs). Rates of NI appear to vary substantially between NICUs. The aim of this study was to determine risk factors for NI, as well as the risk-adjusted variations in NI rates among Canadian NICUs.
Methods
From January 1996 to October 1997, data on demographics, intervention, illness severity and NI rates were submitted from 17 Canadian NICUs. Infants admitted at <4 days of age were included. NI was defined as a positive blood or cerebrospinal fluid culture after > 48 hrs in hospital.
Results
765 (23.5%) of 3253 infants <1500 g and 328 (2.5%) of 13228 infants ≥1500 g developed at least one episode of NI. Over 95% of episodes were due to nosocomial bacteremia. Major morbidity was more common amongst those with NI versus those without. Mortality was more strongly associated with NI versus those without for infants ≥1500 g, but not for infants <1500 g. Multiple logistic regression analysis showed that for infants <1500 g, risk factors for NI included gestation <29 weeks, outborn status, increased acuity on day 1, mechanical ventilation and parenteral nutrition. When NICUs were compared for babies <1500 g, the odds ratios for NI ranged from 0.2 (95% confidence interval [CI] 0.1 to 0.4) to 8.6 (95% CI 4.1 to 18.2) when compared to a reference site. This trend persisted after adjustment for risk factors, and was also found in larger babies.
Conclusion
Rates of nosocomial infection in Canadian NICUs vary considerably, even after adjustment for known risk factors. The implication is that this variation is due to differences in clinical practices and therefore may be amenable to interventions that alter practice.
==== Body
Background
Hospital acquired (nosocomial) infection in neonatal intensive care units (NICUs) is a significant cause of mortality and morbidity, particularly in very low birth weight (VLBW, <1500 g) babies. VLBW infants when affected by these infections are at increased risk for death, have a longer hospital stay, and utilize more resources than non-affected controls [1]. The 15 participating centres in the National Institute of Child Health and Human Development Neonatal Research Network found that 11% to 32% (mean 21%) of VLBW infants have at least one episode of nosocomial sepsis [2]. Similar variations have been described in other datasets [3,4]. These studies also showed that valid comparisons of rates of nosocomial sepsis between centres require adjustment for risk factors, such as patient demographics and admission illness severity.
Risk-adjustment has also been used to look at variations in rates of both neonatal mortality and morbidity [5]. For example, it has been shown that, after correcting for a number of confounders, Canadian NICUs vary significantly in their rates of death, catheter-related sepsis and intraventricular hemorrhage [6-8]. Each of these authors concluded that differences in outcome may be partly attributable to practice variation. Given the complexity and intensity of care provided to VLBW infants, it is likely that the etiology of acquired infections in NICUs is multifactorial [9], with risk arising from the factors attributable to the host, the organism, the environment and clinical interventions. However, previous studies only examined risk factors for hospital-acquired infections in VLBW infants. It is unclear whether risk factors are different in larger infants with 1500 g or higher birth weights (HBW) admitted to the NICU.
The aims of this study were to examine (a) risk factors for culture-proven hospital acquired infections for VLBW and HBW infants admitted to the NICU, (b) the risks associated with specific clinical interventions in the NICU and (c) the risk-adjusted variations in hospital acquired infection rates among Canadian NICUs.
Methods
Study population
The study cohort included all infants admitted to 17 Canadian NICUs prior to 4 days of age. Neonates without a recorded birthweight were excluded. The study cohort was derived from a larger study of 9506 babies admitted to 17 tertiary NICUs in the Canadian Neonatal Network from January 1996 through to October 1997 [10]. These units represent approximately three-quarters of all the tertiary NICU beds in Canada, a nation with >350 000 births annually and almost 30 million people [11,12]. Information about these neonates was prospectively collected by trained abstractors with standardized definitions and operations as part of the larger study of Canadian NICUs [10].
Variable definitions
For the purpose of this study, nosocomial infection (or NI) was defined as one or more positive single organism blood or CSF culture sampled after 48 h of admission in an infant with clinical suspicion of infection. To differentiate between nosocomial and primary (maternal origin) infections, the infant blood culture isolates were required to be different from maternal isolates or to occur at least 7 days after a treated positive blood culture obtained during the first 48 h of life. CSF samples were obtained when indicated according to local practices and policies. The time period of 2 days was chosen to exclude any neonates born with primary infection. Absence of NI included those infants who were never cultured and those with negative blood and/or CSF cultures. An infection episode was defined as a positive culture occurring at least 7 days after a previous treated culture or if the culture isolates were different from the previous culture [13]. We defined nosocomial infections (NI) in this way because blood and CSF are normally sterile, thus a positive culture from one of these sites in an unwell baby is likely to be of serious clinical significance. Additionally, from a practical perspective, these data are well defined and readily abstracted from patient charts.
Other study definitions were in accordance with the Canadian Neonatal Network Data Abstractor's Manual [14]: gestational age was defined as the attending obstetrician's best estimate based on obstetric history and examination, and prenatal ultrasound, except where this was not available or the postnatal physical assessment disagreed with obstetrical estimate by greater than 2 weeks. In that case, the pediatric estimate of gestational age was used instead. Small for gestational age meant being born below the 3rd percentile for weight (corrected for gestation using growth charts developed by Whitfield et al [15]). Major congenital and chromosomal anomalies were specified in the Abstractor's Manual. Outborn infants were those born at a different hospital from the admitting NICU. The Score for Neonatal Acute Physiology, Version II (SNAP-II) was collected on day 1 as a measure of patient acuity on the day of admission. The SNAP-II is a validated score [16] using 6 empirically weighted physiological measurements made during the first 12 hours after admission to the NICU. To be consistent with previous reports, SNAP-II score was categorized as 0 to 9, 10 to 19, 20 to 29 and greater or equal to 30. Therapy related risk factors were extracted from day-1 and day-3 evaluations, using a subset of relevant variables from the Neonatal Therapeutic Intensity Scoring System (NTISS) [17]. The subset of variables included the use of assisted ventilation, peripheral intravenous (IV) access, central venous access, arterial line access, use of antibiotics (day 1 or 3), gavage feeding, intravenous amino acids, and intravenous fat emulsion.
Data analysis
Univariate and bivariate analyses using SPSS (Chicago, Illinois) were used to describe the study population and to explore the relationships between baseline population risks (gestational age, birth weight, gender, Apgar score at 5 minutes, small for gestational age, use of antenatal corticosteroids, multiple births, outborn status, caesarean section, maternal hypertension, congenital anomalies, admission illness severity) and NI. Chi-square test was used to compare mortality and morbidity among infants with and without NI, for infants with birth weight less than 1500 g and 1500 g or more.
We used a multivariable logistic regression model to develop a risk adjustment model for NI. The outcome variable was NI and independent variables were baseline population risks significantly correlated with NI on bivariate analysis. Correlation matrix was used to examine correlation between variables. The full model was created by adding population risk factors, then SNAP-II score, and finally therapy-related factors – variables were considered to be included if their p-value was <0.10. Backward and forward selections from the full model were applied to help determine the parsimonious model. Finally, centres were compared using adjusted odds ratios and 95% confidence intervals derived from the multivariable logistic regression analysis, using a reference centre (the hospital with the median incidence of nosocomial infection). Sensitivity analyses were conducted by excluding outlier sites with large confidence intervals.
Institutional review
This study was part of a larger study surveying morbidities and mortality in Canada, for which ethics approval for anonymous collation of data was obtained at each individual centre.
Results
Patient population and nosocomial infection episodes
16538 infants in the dataset met entry criteria (age less than 4 days on admission). 765 (23.5%) of 3253 VLBW infants, and 329 (2.5%) of 13244 HBW infants (41 were missing birth weight and were excluded), developed at least one episode of NI. Among VLBW infants who developed NI, 78.7% had only 1 episode, 16.2% had 2 episodes, and 5.1% had 3 or more episodes of NI, whereas among HBW infants, 87.9% had 1 episode, 9.3% had 2 episodes, and 2.8% had 3 or more episodes of NI. The median onset of the first infection from day of admission was at 19 days for VLBW infants and 15 days for HBW infants.
Organisms and antibiotics
Among infants with the first episode of NI, the incidence of CSF culture positive infections was 3.7% in VLBW infants, and 2.8% in HBW infants. The most common organism obtained from blood culture was coagulase negative Staphylococcus (72% of isolates in VLBW, and 57% in HBW). The same was true of CSF isolates (68% and 41% respectively). Gram-negative organisms were the next most common pathogen in VLBW, whereas Gram positives, particularly group B Streptococcus, was the next most prevalent in HBW.
Ninety percent of VLBW infants and 73.3% of HBW infants received at least one course of antibiotics in the NICU on either day 1 and / or day 3, as recorded by NTISS
Infant characteristics and therapy related factors
Among VLBW infants, bivariate analysis showed that infants with NI were more likely to be outborn, delivered by vaginal birth, and have 5 minute Apgar score <7, but less likely to be SGA (see Table 1). Infants with NI also had significantly lower birth weight and gestational age, and higher mean day 1 SNAP-II score, than those without NI. A number of therapy related risk factors were significantly associated with NI: these included use of assisted ventilation, peripheral IV access, central venous access, arterial line access, use of antibiotics (on day 1 and/or day 3), gavage feeding, IV amino acid, and IV fat. For HBW infants, NI was associated with similar infant characteristics, but also with congenital anomalies, 5 minute Apgar score <7, and higher SNAP-II score. Correlation between variables: NI has significant correlation with all variables identified except for antibiotic use. However, when stratified by birth weight, antibiotic use has significant correlation with NI for HBW infants. The Spearman correlation coefficient between CVL and parenteral nutrition is 0.26, which is significant. Furthermore, a cross-tabulation of NI by CVL controlling for IV parenteral nutrition (TPN) indicates significant correlation beween NI and CVL and a contingency coefficient of 0.1 for either TPN present or absent. Therefore, there is no evidence that the presence of TPN or CVL in the models renders the other redundant.
Table 1 Characteristics of infants with and without nosocomial infections (NI), for birth weights less than 1500 g and 1500 g or more (asterisks indicate significant difference at p < 0.05 level between infants with and without NI in each birthweight group)
Birth weight Less than 1500 g 1500 g or more
Characterististics With NI (n = 765) Without NI (n = 2488) With NI (n = 329) Without NI (n = 13244)
Gestation (mean weeks ± sd) 27.2 +/- 2.3 28.8 +/- 2.8* 35.8 +/- 3.6 36.3 +/- 3.2*
Birth weight (mean gm ± sd) 957 +/- 259 1010 +/- 267* 2611 +/- 806 2758 +/- 796*
Day 1 SNAP-II (mean ± sd) 18.2 +/- 12.5 14.1 +/- 13.0* 11.9 +/- 12.5 4.9 +/- 8.5*
Small for gestational age (%) 9.7 12.6* 2.8 2.7
Antenatal steroids (%) 71.4 68.2 21.1 20.4
Multiple birth (%) 26.7 28.5 7.9 12.5*
Outborn (%) 24.6 12.7* 59.9 23.1*
Male (%) 55.3 51.8 59.9 58.1
Cesarean delivery (%) 47.5 55.2* 38.9 36.3
Maternal hypertension (%) 18.1 19.4 10.4 11.6
Apgar @ 5 mins < 7 (%) 26.9 20.0* 21.4 11.9*
Congenital Anomalies (%) 10.6 7.9* 48.6 10.9*
Mortality and morbidity
Mortality was more strongly associated with the NI group compared to the group without NI for HBW infants, but not for VLBW infants. Major morbidities (necrotizing enterocolitis, chronic lung disease at 36 weeks post-menstrual age, severe intraventricular hemorrhage/periventricular lesions, and severe stages (3 and 4) of retinopathy of prematurity) were more common in the NI group (Table 2).
Table 2 Mortality and morbidity among infants with and without nosocomial infections, for infants with birth weight less than 1500 g and 1500 g or more (asterisks indicate significant difference at p < 0.05 level between infants with and without nosocomial infections within each birth weight group).
Birthweight Less than 1500 g 1500 g or more
Outcome (%) With NI (n = 765) Without NI (n = 2488) With NI (n = 329) Without NI (n = 12915)
Mortality 8.7 8.6 8.5 1.3*
Necrotizing enterocolitis 13.4 4.4* 8.8 0.4*
Chronic lung disease 38.7 18.5* 25.7 4.4*
Severe intraventricular hemorrhage 10.5 8.2* 5.3 2.8*
Severe retinopathy of prematurity 14.6 8.1* 0.0 0.0
Survival without major morbidity 46.3 69.5* 74.8 96.0*
Risk factors predictive of nosocomial infection on multivariable logistic regression
In VLBW infants, factors associated with NI in the multivariable logistic model were: gestational age, use of assisted ventilation and IV fat and, and outborn status (see table 3). Ninety-nine percent of infants receiving IV fat also received IV amino acid, so these interventions were combined. SNAP-II scores were combined into ≥10 and 0–9 (the reference) groups because the parameters for the higher score groups [10–19, 20–29, > = 30) appeared to be similar. Compared to infants with a gestational age of greater than 28 weeks, lower gestation was increasingly more likely to be associated with NI. SNAP-II score ≥10 was of borderline significance.
Table 3 Risk factors associated with nosocomial infection on multivariate analysis, for infants with birth weight less than 1500 g and 1500 g or more
Birth weight Less than 1500 g Odds Ratio (95% confidence intervals) 1500 g or more Odds Ratios (95% confidence intervals)
Gestation ≤24 weeks 2.7 (1.4 – 5.3) 4.0 (0.5 – 32.7)
Gestation 25–28 weeks 2.2 (1.2 – 4.1) 6.1 (1.6 – 23.1)
Gestation 29–32 weeks 1.1 (0.6 – 2.0) 1.9 (1.3 – 2.6)
SNAP II ≥10 1.2 (1.0 – 1.6) 1.0 (0.8 – 1.3)
Peripheral intravenous line 0.9 (0.7 – 1.1) 1.9 (1.1 – 3.2)
Central venous Line 0.9 (0.8 – 1.2) 1.6 (1.2 -2.1)
Assisted ventilation 1.5 (1.1 – 2.0) 2.9 (1.9 – 4.6)
Parenteral nutrition 3.9 (3.0 – 5.2) 5.1 (3.8 – 6.9)
Outborn status 2.0 (1.6 – 2.5) 1.9 (1.4 – 2.6)
Use of antibiotics 0.9 (0.4 – 2.3) 2.9 (1.6 – 5.0)
Congenital anomalies 1.2 (0.9 – 1.6) 2.9 (2.2 – 3.8)
In HBW infants, factors associated (p < 0.05) with NI in the multivariable logistic model were: gestational age, use of central venous access, peripheral intravenous access, and use of IV fat or IV amino acid, assisted ventilation, outborn status, use of antibiotics and presence of congenital anomalies. Again, IV fat and IV amino acid could be used interchangeably in the model. Infants who received parenteral nutrition were more likely to have NI. Central venous access was associated with higher risk for NI, as was use of peripheral IV access.
Variations in nosocomial infection rates among NICUs: site comparisons
For VLBW infants, cross-tabulation of site and NI showed significant variation between sites (chi-square = 248.7, df = 16, P < 0.001). Crude rates of NI ranged from 6.7% to 74.5% of infants having at least one episode of NI. The crude incidences of NI for 8 hospitals were significantly (p < 0.05) different from the reference hospital (site O) (Figure 1). Site variations persisted after adjustment for baseline patient risk factors (gestational age, admission day SNAP-II score, outborn status) using multivariate logistic regression analysis (Figure 1), with the highest odds ratio being 8.6 (95% confidence interval (CI) 4.1 to 18.2), and the lowest 0.2 (95% CI 0.1 to 0.4) when compared to the reference site (Figure 1).
Figure 1 Crude and risk adjusted odds ratios of nosocomial infection rates by hospital for infants <1500 g birth weight (point estimates and 95% confidence interval limits shown). Adjustment is for gestational age, admission day SNAP-II score, and outborn status. Site O, the site closest to the median risk, was used as the reference site.
For HBW infants, significant variation was also present between sites (chi-square = 360.0, df = 16, P < 0.001). Crude rates of NI ranged from 0.1% to 17.0 %. When only site variables were in the regression model, 6 hospitals had significantly different (p < 0.05) NI rates from the reference hospital (F). Significant site differences largely corrected after adjustment for baseline population risks (gestational age, admission day SNAP-II score, outborn status) (Figure 2).
Figure 2 Crude and risk adjusted odds ratios of nosocomial infection rates by hospital for infants ≥1500 g birth weight (point estimates and 95% confidence interval limits shown). Adjustment is for gestational age, admission day SNAP-II score, and outborn status. Site O, the site closest to the median risk, was used as the reference site.
Comparison of Figures 1 and 2 shows that hospitals with low and high incidences of NI among VLBW infants also tended to have lower and higher incidences of NI among larger infants. When sites P and N were excluded from the analysis, site A was also found to have significantly higher incidence of NI than the median hospital for both VLBW and HBW infants. (Figures 3 and 4)
Figure 3 Crude and risk adjusted odds ratios of nosocomial infection rates by hospital (with sites P and N excluded) for infants <1500 g birth weight (point estimates and 95% confidence interval limits shown). Adjustment is for gestational age, admission day SNAP-II score, and outborn status.
Figure 4 Crude and risk adjusted odds ratios of nosocomial infection rates by hospital (with sites P and N excluded) for infants ≥1500 g birth weight (point estimates and 95% confidence interval limits shown). Adjustment is for gestational age, admission day SNAP-II score, and outborn status.
Discussion
The above results represent data from 17 tertiary NICUs, making up approximately three-quarters of level III neonatal cots in Canada. Canada has publicly-funded health care with a highly regionalized neonatal-perinatal care system in which each tertiary NICU is the referral centre for its geographic region. Consequently, our results approximate a population-based study for three quarters of the Canadian population.
The bivariate analysis suggests that infants who experience at least one episode of nosocomial bacteremia or nosocomial bacterial meningitis are smaller, more preterm, and had higher illness severity on admission. They have lower 5-minute Apgar scores and are more likely to have been transported from another centre. VLBW infants with NI were more likely to have been delivered vaginally. Nosocomial infection rates will vary from institution to institution if these demographic factors vary from region to region. Adjusted data, to some extent, correct for these geographic variations.
Our data show that significant variations in nosocomial infection do occur between units, with crude rates ranging 6.7% to 74.5% among VLBW infants, and 0.1% to 17.0% among HBW infants. These variations persist after adjustment for gestation, outborn status and admission illness severity. These findings are not unique to hospital acquired infections: using data from the same population of infants, other authors have identified between-site variations in the crude and adjusted rates of both mortality and a number of major neonatal morbidities, including chronic lung disease, retinopathy of prematurity, catheter-related sepsis and intraventricular lesions [6-8,10,18]. There seems little doubt that these differences in outcome exist, the important question remains as to the underlying reasons for the differences.
Multivariate analysis suggests that, in addition to gestation and acuity, a number of therapeutic interventions are associated with nosocomial infection. These include mechanical ventilation, vascular access (central and peripheral), and parenteral nutrition. These factors are predictive in cohorts of both VLBW and HBW babies – supporting their place as clinically significant risks factors for NI. The preponderance of NI among HBW babies with congenital anomalies may reflect surgical interventions, or the possibility of immunocompromise. We recognize that there are likely to be other, unrecorded demographic factors or interventions, that may account for risk, such as feeding practices or ethnicity – these remain unaccounted for in our study.
There is little doubt from this study that nosocomial infection is associated with significant in-hospital morbidity, but the results on mortality are more subtle. Unadjusted mortality in VLBW infants was not increased by an episode of nosocomial sepsis. A possible explanation is that the effect size is small because most deaths among VLBW infants occur in the first few days of life from other causes, for example cardiorespiratory failure, whereas NI occurs more commonly at a later age. In addition, coagulase negative staphylococcal infections (the commonest isolate in NICU) are reported to have a relatively low mortality in VLBW infants (2% or less in three large cohort studies [19-21]), making a small contribution to overall mortality that may not achieve statistical significance.
Between-site analysis confirmed that inter-NICU variation in NI rates persisted after risk adjustment for birth weight, gestation and illness severity in VLBW infants. The 3 sites with lower than median incidence of NI for VLBW infants, also had a lower incidence of NI for HBW infants. Three out of 4 sites with higher than median incidence of NI for HBW infants also had higher incidence of NI for VLBW infants. The impression that overall site performance is consistent for infants at different birth weight groups supports the contention that factors other that those recorded and adjusted for in this study, probably related to interventions or the NICU environment, contribute to the risk of nosocomial infection – especially in VLBW infants, with whom most of the burden of NI lies.
We have seen that nosocomial infection rates vary across Canada and are significantly affected by clinical interventions. However, a number of important factors have not been adjusted for, particularly those related to organism virulence or infectivity, and population susceptibility. One could hypothesize that organism virulence or infectivity varies across the country, as does the susceptibility of population groups. There is some evidence to support variation in genetic susceptibility: Ahrens et al [22] reported that VLBW neonates with specific CD14 gene mutations had a greater rate of neonatal sepsis than those without. A more detailed study of population demographics and organism subtypes may reveal whether these factors impact on nosocomial infection rates. Similarly, environmental factors such as unit layout and clinical practices may affect nosocomial infection rates – the extensive list of potential confounders may require an alternative research methodology to discover association and/or causation. Bloom et al [23] reported decreased rates of nosocomial infections by sharing information about differences between NICUs with high and low nosocomial infection rates, and addressing these variations in clinical practice. The Canadian Neonatal Network is embarking on a multicentre study to examine the effect of multiple evidence-based practice interventions on nosocomial infection rates, which will hopefully answer some of these questions.
A limitation of our study is that it is possible that some positive blood or CSF cultures are in fact false positives. Struthers et al [24] estimated that 5% of positive blood cultures were false, suggesting that the vast majority of confirmed bacteremia is true. However, it might be argued that even false positive results represent a significant burden of illness, and including these episodes may fairly represent this burden. Another limitation of this type of observational study is the difficulty in differentiating cause and effect. The observational data in this study were presented using univariate, unadjusted and risk-adjusted multivariate models. Unfortunately, these analyses cannot account for the temporal relationships between events, nor do they adjust for risk factors that are not included in data collection. However, the strengths of the logistic regression model are its robustness and extensive use in similar studies. Future studies should take temporal relationships into account, permitting time-sensitive analysis, better inference of cause and effect, and assessment of attributable risk.
Conclusion
Rates of nosocomial infection in Canadian NICUs vary considerably, even after correction for known risk factors. The data suggest that this variation is due, in part, to differences in clinical practices, and therefore may be amenable to interventions that alter practice. More study, and alternative study designs, may be required to evaluate contributing factors and effect practice change.
Canadian Neonatal Network
Shoo K. Lee (Coordinator, Canadian Neonatal Network; Centre for Healthcare Innovation and Improvement, Vancouver, BC); Wayne Andrews (Charles A. Janeway Child Health Centre, St John's, NF); Ranjit Baboolal (North York Hospital, N. York, ON); Jill Boulton (St Joseph's Health Centre, London, ON; previously Mt Sinai Hospital, Toronto, ON); David Brabyn (Royal Columbian Hospital, New Westminster, BC); David S.C. Lee (St Joseph's Health Centre; London, ON); Derek Matthew (Victoria General Hospital (Victoria, BC); Douglas D. McMillan (IWK-Grace Health Centre for Women, Children and Families, Halifax, NS; formerly Foothills Hospital, Calgary, AB); Christine Newman (Hospital for Sick Children; Toronto, ON); Arne Ohlsson (Mt Sinai Hospital, Toronto, ON; formerly Women's College Hospital, Toronto, ON); Abraham Peliowski (Royal Alexandra Hospital, Edmonton, AB); Margaret Pendray (Children's & Women's Health Centre of British Columbia (Vancouver, BC); Koravangattu Sankaran (Royal University Hospital, Saskatoon, SK); Barbara Schmidt (Hamilton Health Sciences Corporation, Hamilton, ON); Mary Seshia (Health Sciences Centre, Winnipeg, MB); Anne Synnes (Children's and Women's Health Centre of British Columbia, Vancouver, BC; formerly Montreal Children's Hospital, Montreal, PQ); Paul Thiessen (Children's & Women's Health Centre of British Columbia, Vancouver, BC); Robin Walker (Children's Hospital of Eastern Ontario and Ottawa General Hospital, Ottawa, ON); Robin Whyte (IWK-Grace Health Centre for Women, Children and Families, Halifax, NS).
Coordinating centre staff
Holly Bavinton; Stella Karuri; Zhenguo Qiu; Sarka Lisonkova, Shawn Stewart.
Abbreviations
Nosocomial infection (NI)
Neonatal Intensive Care Unit (NICU)
Score for Neonatal Acute Physiology, Version II (SNAP-II)
Neonatal Therapeutic Intensity Scoring System (NTISS)
Very low birth weight (VLBW)
Higher birth weight (HBW)
Cerebrospinal fluid (CSF)
Standard deviation (SD)
Confidence interval (CI)
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KA interpreted data and drafted the manuscript. DDM, WA and MP were site investigators. ZQ performed statistical analysis and interpretation. SKL was the principal investigator and drafted the manuscript. All these individuals read and approved the final manuscript. The CNN represents all site investigators, and was responsible for organization and administration of the SNAP study, and subsequent data flow.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Ruth Little for editorial assistance and Li-Yin Chien for assistance with data analysis.
This study was supported by Grant 40503 and Grant 00152 from the Medical Research Council of Canada. Additional funding was provided by the B.C.'s Children's Hospital Foundation; Calgary Regional Health Authority; Dalhousie University Neonatal-Perinatal Research Fund; Division of Neonatology, Children's Hospital of Eastern Ontario; Child Health Program, Health Care Corporation of St John's; The Neonatology Program, Hospital for Sick Children; Lawson Research Institute; Midland Walwyn Capital Inc; Division of Neonatology, Hamilton Health Sciences Corporation; Mount Sinai Hospital; North York General Hospital Foundation; Saint Joseph's Health Centre; University of Saskatchewan Neonatal Research Fund; University of Western Ontario; Women's College Hospital.
==== Refs
Stoll BJ Gordon T Korones SB Shankaran S Tyson JE Bauer CR Fanaroff AA Lemons JA Donovan EF Oh W Stevenson DK Ehrenkranz RA Papile LA Verter J Wright LL Late-onset sepsis in very low birth weight neonates: a report from the National Institute of Child Health and Human Development Neonatal Research Network J Pediatr 1996 129 63 71 8757564
Stoll BJ Hansen N Fanaroff AA Wright L Carlo WA Ehrenkranz RA Lemons JA Donovan EF Stark AR Tyson JE Oh W Bauer CR Korones SB Shankaran S Laptook AR Stevenson DK Papile LA Poole WK Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network Pediatrics 2002 110 285 91 12165580 10.1542/peds.110.2.285
Stevenson DK Wright LL Lemons JA Oh W Korones SB Papile LA Bauer CR Stoll BJ Tyson JE Shankaran S Fanaroff AA Donovan EF Ehrenkranz RA Verter J Very low birth weight outcomes of the National Institute of Child Health and Human Development Neonatal Research Network, January 1993 through December 1994 Am J Obstet Gynecol 1998 179 1632 1639 9855609
Brodie SB Sands KE Gray JE Parker RA Goldmann DA Davis R Richardson DK Occurrence of nosocomial bloodstream infections in six neonatal intensive care units Pediatr Infect Dis J 2000 19 56 65 10643852 10.1097/00006454-200001000-00012
Horbar JD Badger J Lewit EM Rogowski J Shiono PH Hospital and patient characteristics associated with variation in 28-day mortality rates for very low birth weight infants Pediatrics 1997 99 149 156 9024438 10.1542/peds.99.2.149
Sankaran K Chien LY Walker R Seshia M Ohlsson A Lee SK The Canadian Neonatal Network Variations in mortality rates among Canadian neonatal intensive care units CMAJ 2002 166 173 178 11826939
Chien LY McNab Y Aziz K Andrews W McMillan DD Lee SK The Canadian Neonatal Network Variations in central venous catheter related infection risks among Canadian neonatal intensive care units Pediatr Infect Dis J 2002 21 505 511 12182373 10.1097/00006454-200206000-00006
Synnes A Chien LY Peliowski A Baboolal R Lee SK The Canadian Neonatal Network Variations in Intraventricular Hemorrhage Incidence Rates among Canadian Neonatal Intensive Care Units J Pediatr 2001 138 525 531 11295716 10.1067/mpd.2001.111822
Edwards WH Preventing nosocomial bloodstream infection in very low birth weight infants Semin Neonatol 2002 7 325 33 12401302
Lee SK McMillan D Ohlsson A Pendray M Synnes A Whyte R Chien LY Sale J the Canadian Neonatal Network Variations in practice and outcomes of the Canadian NICU Network : 1996–7 Pediatrics 2000 106 1070 9 11061777 10.1542/peds.106.5.1070
Statistics Canada. Births and birth rate Accessed February 23, 2001
Statistics Canada. Population Accessed February 23, 2001
Chien LY MacNab Y Aziz K Andrews W McMillan DD Lee SK the Canadian Neonatal Network Variations in central venous catheter-related infection risks among Canadian neonatal intensive care units Pediatr Infect Dis J 2002 21 505 12182373 10.1097/00006454-200206000-00006
The Canadian Neonatal Network™ The Canadian Neonatal Network™ Data Abstractor's Manual 1996 Vancouver, B.C
Whitfield M British Columbia Provincial Growth Chart 1996 Vancouver, Canada: British Columbia Children's Hospital
Richardson DK Corcoran JD Escobar GJ Lee SK SNAP-II and SNAPPE-II : Simplified newborn illness severity and mortality risk scores J Pediatr 2001 138 92 100 11148519 10.1067/mpd.2001.109608
Gray JE Richardson DK McCormick MC Workman-Daniels K Goldmann DA Neonatal therapeutic intervention scoring system: a therapy-based severity-of-illness index Pediatrics 1992 90 561 7 1408510
Ohlsson A Synnes AR Seshia MMK Newman CJ Schmidt B Chien LY Lee SK Variations in the Site Specific Rates of Bronchopulmonary Dysplasia (BPD) among Infants Admitted to 17 Neonatal Intensive Care Units (NICUs) in Canada – Implications for Bench Marking Pediatr Res 2000 47 371A
Isaacs D A ten year, multicentre study of coagulase negative staphylococcal infections in Australasian neonatal units Arch Dis Child Fetal Neonatal Ed 2003 88 F89 F93 12598493 10.1136/fn.88.2.F89
Karlowicz MG Buescher ES Surka AE Fulminant late-onset sepsis in a neonatal intensive care unit, 1988-and the impact of avoiding empiric vancomycin therapy Pediatrics 2000 106 1387 90 11099593 10.1542/peds.106.6.1387
Stoll BJ Hansen N Fanaroff AA Wright LL Carlo WA Ehrenkranz RA Lemons JA Donovan EF Stark AR Tyson JE Oh W Bauer CR Korones SB Shankaran S Laptook AR Stevenson DK Papile LA Poole WK Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network Pediatrics 2002 110 285 91 12165580 10.1542/peds.110.2.285
Ahrens P Kattner E Köhler B Härtel C Seidenberg J Segerer H Moller J Gopel W Genetic Factors in Neonatology Study Group. Mutations of genes involved in the innate immune system as predictors of sepsis in very low birth weight infants Pediatr Res 2004 55 652 656 14739370 10.1203/01.PDR.0000112100.61253.85
Bloom BT Craddock A Delmore PM Kurlinski JP Voeloker M Landfish N Rodriguez-Pierce M Swanton D Rossi J Ehlen J Harmon C Deterding J Houser F Reducing acquired infections in the NICU: Observing and implementing meaningful differences in process between high and low acquired infection rate centres J Perinatol 2003 23 489 492 13679937 10.1038/sj.jp.7210981
Struthers S Underhill H Albersheim S Greenberg D Dobson S A comparison of two versus one blood culture in the diagnosis and treatment of coagulase-negative staphylococcus in the neonatal intensive care unit J Perinatol 2002 22 547 549 12368970 10.1038/sj.jp.7210792
|
16004613
|
PMC1182378
|
CC BY
|
2021-01-04 16:31:05
|
no
|
BMC Pediatr. 2005 Jul 8; 5:22
|
utf-8
|
BMC Pediatr
| 2,005 |
10.1186/1471-2431-5-22
|
oa_comm
|
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-701596703010.1186/1471-2458-5-70DebateWashing our hands of the congenital cytomegalovirus disease epidemic Cannon Michael J [email protected] Katherine Finn [email protected] National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia2 Rollins School of Public Health, Emory University, Atlanta, Georgia, USA3 Nell Hodgson Woodruff School of Nursing, Emory University, Atlanta, Georgia, USA2005 20 6 2005 5 70 70 9 12 2004 20 6 2005 Copyright © 2005 Cannon and Davis; licensee BioMed Central Ltd.2005Cannon and Davis; 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
Each year in the United States, an estimated 40,000 children are born with congenital cytomegalovirus (CMV) infection, causing an estimated 400 deaths and leaving approximately 8000 children with permanent disabilities such as hearing or vision loss, or mental retardation. More children are affected by serious CMV-related disabilities than by several better-known childhood maladies, including Down syndrome, fetal alcohol syndrome, and spina bifida.
Discussion
Congenital CMV is a prime target for prevention not only because of its substantial disease burden but also because the biology and epidemiology of CMV suggest that there are ways to reduce viral transmission. Because exposure to the saliva or urine of young children is a major cause of CMV infection among pregnant women, it is likely that good personal hygiene, especially hand-washing, can reduce the risk of CMV acquisition. Experts agree that such measures are likely to be efficacious (i.e., they will work if consistently followed) and the American College of Obstetricians and Gynecologists recommends that physicians counsel pregnant women about preventing CMV acquisition through careful attention to hygiene. However, because of concerns about effectiveness (i.e., Will women consistently follow hygienic practices as the result of interventions?), the medical and public health communities appear reluctant to embrace primary CMV prevention via improved hygienic practices, and educational interventions are rare. Current data on the effectiveness of such measures in preventing CMV infection are promising, but limited. There is strong evidence, however, that educational interventions can prevent other infectious diseases with similar transmission modes, suggesting that effective interventions can also be found for CMV. Until a CMV vaccine becomes available, effective educational interventions are needed to inform women about congenital CMV prevention.
Summary
Perhaps no single cause of birth defects and developmental disabilities in the United States currently provides greater opportunity for improved outcomes in more children than congenital CMV. Given the present state of knowledge, women deserve to be informed about how they can reduce their risk of CMV infection during pregnancy, and trials are needed to identify effective educational interventions.
==== Body
Background
The history of public health in twentieth century America is replete with successes in the prevention of birth defects and childhood disabilities. Vaccines have virtually eliminated polio, congenital rubella syndrome, and Haemophilus influenzae meningitis [1-3]. Educational efforts aimed at preventing fetal alcohol syndrome have reduced maternal alcohol consumption during pregnancy [4]. Prenatal vitamins and folic acid fortification of cereals have lowered rates of neural tube defects [5], while antiretroviral treatments have caused the occurrence of mother-to-child human immunodeficiency virus (HIV) transmission to plummet [6]. Notably absent from the list of successes, however, is the prevention of congenital cytomegalovirus (CMV) disease [7].
Perhaps no single cause of birth defects and developmental disabilities in the United States currently provides greater opportunity for improved outcomes in more children than congenital CMV. Each year in the United States, an estimated 40,000 children are born with congenital CMV infection, causing an estimated 400 deaths and leaving approximately 8000 children with permanent disabilities such as hearing or vision loss, or mental retardation [8]. The direct annual economic costs of caring for these children are estimated at $1-$2 billion [8,9]. More children are adversely affected by congenital CMV disease than by several better-known childhood diseases or syndromes (Figure 1). Congenital CMV is a prime target for prevention not only because of its substantial disease burden but also because the biology and epidemiology of CMV suggest that there are ways to reduce viral transmission. Unfortunately, by missing prevention opportunities, we in the medical and public health communities are washing our hands of the congenital CMV disease epidemic.
Figure 1 Estimates of the annual burden of prominent childhood diseases and syndromes in the US [3, 5, 6, 8, 51–57]. Assumes 4 million live births per year and 20 million children <5 years of age. Childhood deaths were defined as those occurring <1 year after birth except for Haemophilus influenzae type B (Hib) (<5 years) and HIV/AIDS (<13 years). Where applicable, numbers represent means of published estimates. All estimates should be considered useful for rough comparisons only since surveillance methodology and diagnostic accuracy varied over different studies. CRS, congenital rubella syndrome.
CMV and congenital CMV disease
As with other human herpesviruses, initial infection by CMV (also known as primary infection) is followed by the establishment of lifelong latent infection, from which periodic reactivation is common [10,11]. Symptoms are usually absent during primary infection and reactivation, but CMV is shed in various bodily secretions, particularly urine and saliva [12]. CMV excretion can be continuous or intermittent, generally lasting several weeks in adults but often continuing for months or years in young children [13-15]. CMV infection is widespread, with estimates of CMV seroprevalence in the United States ranging from 40% to 80% [16-18].
CMV is transmitted person-to-person via close non-sexual contact, sexual activity, breastfeeding, blood transfusions, and organ transplantation [12]. CMV has not been shown to be transmitted via respiratory secretions or aerosolized virus. For the pregnant woman, the most likely source of infection may be contact with the urine or saliva of young children, especially her own children [19,20].
Congenital CMV disease is most likely to occur following a primary infection in the mother. Primary infections occur in 1%-4% of seronegative, pregnant women and lead to fetal infection in 40%-50% of these pregnancies. Maternal CMV reactivation or reinfection with a different CMV strain leads to fetal infection in about 1% of seropositive, pregnant women. Approximately 10% of congenitally infected infants are symptomatic at birth, and of the 90% who are asymptomatic, 10%-15% will develop symptoms over months or even years (Figure 2) [21]. Permanent sequelae can result from CMV infection of the fetus during any trimester, but infection during early fetal development is likely to be especially damaging [22,23]. Since few newborns are screened for CMV, the true impact of congenital CMV infection is underappreciated.
Figure 2 Transient and permanent outcomes among children with congenital CMV disease.
Discussion
Reducing the burden of congenital CMV disease
Of all the ways to fight congenital CMV disease, the development of a vaccine is viewed as the most promising. Considerable progress has been made over the last 30 years, but insufficient interest by vaccine manufacturers (Stanley Plotkin, personal communication) and technical challenges make it is uncertain when a vaccine will become available [24]. Avenues for improving outcomes in congenitally infected children have also been explored, including anti-CMV therapies (e.g., ganciclovir) for seriously infected infants [25,26] and supportive care, such as hearing screening, language therapy, and special education [27,28]. In contrast, insufficient emphasis has been given to preventing CMV infection in pregnant women. While women may be infected via several routes, the remainder of this article focuses on preventing transmission via the important child-to-mother route, by encouraging hygienic practices such as frequent hand washing.
A number of experts have suggested that women be educated about hygienic practices for preventing CMV transmission from young children, and there is little dispute over what the prevention guidelines should entail (Figure 3) [7,11,29-32]. This consensus is reflected in current American College of Obstetricians and Gynecologists guidelines, which recommend that physicians counsel pregnant women about preventing CMV acquisition through careful attention to hygiene [33]. Nevertheless, hygienic practices do not appear to be widely discussed by healthcare providers and prospective mothers are often unaware of both CMV disease and the potential benefits of hygienic practices. The virtual absence of a prevention message has been due, in part, to the low profile of congenital CMV. Infection is usually asymptomatic in both mother and infant, and when symptoms do occur, they are non-specific, so most CMV infections go undiagnosed. The prevention message has also been hindered by a sense that infection is unavoidable. For example, a number of authors have urged prevention education for women on the one hand but on the other hand, they have noted that "CMV is neither preventable nor treatable..."[34],"...it is not certain that infections in pregnant women can be prevented by avoiding exposure" [35], "...it is doubtful whether parents will comply with these [behavioral measures in nonstudy settings..." [36], "...there is very little evidence for the efficacy of these strategies and even less for their practical implementation...", and "The only effective prevention strategy relies upon the development of a vaccine." [37] Given the relative invisibility of CMV disease and these mixed messages about prevention education, it is not surprising that healthcare providers do not discuss CMV with their patients and that women are unaware of the risks of CMV infection.
Figure 3 Hygienic practices to reduce risk of CMV infection for women who are pregnant or planning to become pregnant. When interacting with young children, women should assume the children are secreting CMV in their urine and saliva.
Preventing CMV infection through hygienic practices
Why the ambivalence toward hygienic practices? Studies have shown that transmission of CMV via the urine and saliva of children is a major cause of infection among pregnant women [19,20]. In addition, more than 100 years of evidence conclusively demonstrates that hand washing reduces risk of infection for a wide range of pathogens [38]. Thus, nearly everyone would agree that, in theory, hand washing can prevent CMV infection because hands are an important vehicle for transmission. The concern, then, is not the efficacy of hygienic practices (i.e., Will they work if consistently followed?) but, instead, the effectiveness of interventions to promote them (i.e., Will women consistently follow hygienic practices as the result of interventions?).
It is important to recognize the implications of the consensus that hygienic practices are efficacious for preventing CMV transmission. Individual women have the right to know that, under ideal conditions, risk of child-to-mother CMV transmission can be reduced by proper hygienic practices. This is equivalent to the ethical obligation to inform individuals that, under ideal conditions, safer sexual practices will reduce the risk of acquiring HIV. This obligation is independent of whether any particular educational program or intervention is effective. All women of childbearing age, whether they are CMV seropositive or seronegative, carry some risk of new CMV infection during pregnancy and thus should be informed of hygienic practices that reduce that risk. As Revello and Gerna aptly remind us, "...withholding information on possible medical interventions is unethical (and legally risky)" [39]. The terrible burden of congenital CMV disease (Figure 1) should make the provision of such information a priority.
As there is consensus on the efficacy of hygienic practices in preventing CMV transmission, the next step is to evaluate the effectiveness of educational interventions in preventing CMV transmission. Current evidence of effectiveness is promising, but limited. In one study, after non-pregnant women were educated about CMV prevention, hygienic practices improved [40]. In a small study of Houston families, Demmler and colleagues found that behavioral changes prevented transmission of CMV (unpublished report described in Yow and Demmler [7]). Adler and colleagues studied the effectiveness of hygienic practices in a randomized, controlled trial of 39 seronegative, non-pregnant women with young children who were shedding CMV [31]. Although the study was underpowered to detect significant differences in infection rates between the intervention and control groups (and thus the intervention was deemed unsuccessful by some), seroconversion rates decreased as CMV education and support increased. Furthermore, in the same study, 14 pregnant women were educated regarding hygienic practices and then followed for comparison with the randomized groups; none of the women seroconverted – a significant difference compared with the randomized groups. A more recent study also reported that pregnant women who received an intervention involving hygienic practices were significantly less likely to acquire CMV infection than were non-pregnant women [41].
More conclusive evidence of effectiveness can be found in the literature on community-based interventions for the prevention of other infectious diseases with similar transmission modes. For example, a meta-analysis found that community intervention trials that encouraged washing hands with soap reduced the risk of diarrheal diseases by 47% [42]. Hand-washing programs reduced respiratory illness among military recruits [43] and children in daycare [44], and interventions involving hand sanitizers reduced absenteeism among elementary school teachers and children [45]. An in-depth review of the literature would be useful for determining the key factors associated with the success of these and other community interventions.
Although all women of childbearing age deserve to be informed about CMV, interventions for preventing CMV transmission are most likely to be effective for pregnant women, who tend to be highly motivated, often changing behavior to protect the health of their developing fetuses. As a case in point, 25% of low-income smokers spontaneously quit smoking during pregnancy [46]; this percentage is higher than that achieved by most smoking cessation programs [47]. Studies by Adler and colleagues suggest that motivation for avoiding CMV infection is considerably higher among pregnant than non-pregnant women [31,41].
In sum, the evidence to date gives every indication that effective interventions can be found for preventing CMV infection among pregnant women. Thus, the paradigm must shift from wondering whether such interventions will be effective to developing and evaluating interventions until effective ones are identified.
Next steps
Given the consequences of allowing the congenital CMV disease epidemic to continue unabated, it is imperative that women receive the educational message about congenital CMV disease prevention (see ) [48]. Promotion of this message could translate years of careful CMV research into an immediate public health benefit. Encouraging hygienic practices would be relatively inexpensive (requires no laboratory testing or additional doctor visits), ethically responsible (allows women to make informed decisions), and likely to prevent disabilities and save lives. Based on studies of the economic impact of future CMV vaccines, which estimate savings of $50,000 per quality adjusted life-year saved [8], it is likely that CMV education efforts would provide a highly favorable cost-benefit ratio as well. CMV educational messages should emphasize hygienic practices as a precaution for all women who are pregnant or planning to become pregnant, and reasonable but not extreme measures for minimizing risk during interactions with young children (Figure 3). In many instances, a CMV education message could build upon and strengthen other public health messages about infection prevention through improved hand hygiene. Once effective hand-hygiene messages are identified, more ambitious goals might also be considered, such as prevention of sexual transmission or transmission between children in daycare.
Enlisting the support of healthcare providers to convey the gravity of CMV infection and the importance of good hygienic practices is crucial. Healthcare providers have many opportunities to provide women with such information, such as during annual gynecological exams or well-baby visits as women accompany their children to the pediatrician's office. Other steps, such as community education by healthcare providers (including information sessions with daycare directors, daycare providers, and parents) and provision of information by the Public Health Service and other professional organizations, can supplement the healthcare provider's information [49]. Success in delivering the message will depend on the active involvement of all relevant healthcare professionals.
In addition, trials are needed to identify effective community-based interventions for preventing CMV transmission to pregnant women. These trials will be important for quantifying the effectiveness of the proposed hygienic practices and for assessing the proportion of CMV infections that result from child-to-mother transmission as opposed to other routes, such as sexual transmission. However, such trials should not delay nor hinder the educational effort, just as we would not wait to inform skydivers about the prudence of parachute use because the results of controlled trials are not yet in. "As with many interventions intended to prevent ill health, the effectiveness of parachutes has not been subjected to rigorous evaluation by using randomised controlled trials...[One option] is that we accept that, under exceptional circumstances, common sense might be applied when considering the potential risks and benefits of interventions [our emphasis]."[50] Common sense tells us that avoiding CMV-laden secretions will prevent transmission and that the potential benefits of educational interventions far outweigh the potential risks.
Although an effective CMV vaccine would be ideal and vaccine development deserves increased support, hope for a vaccine tomorrow should not stand in the way of a vigorous educational message today. Prevention through improved hygienic practices will not be easy, but washing our hands of the problem by staying the present course guarantees that each year in the United States, hundreds of children will die and thousands of others will swell the ranks of CMV-affected children growing up with serious disabilities. Just as the medical and public health communities are successfully meeting the challenges posed by polio, rubella, HIV, and neural tube defects, now is the time to meet the congenital CMV challenge head on by making awareness and prevention high priorities.
Summary
• Each year in the United States, congenital CMV
- causes an estimated 400 deaths
- leaves more than 8000 children with permanent disabilities such as hearing or vision loss, or mental retardation
• Exposure to the saliva or urine of young children is a major cause of CMV infection among pregnant women.
• Risk of CMV infection is likely to be reduced by careful attention to good personal hygiene, especially hand-washing.
• Women should be informed about how to reduce their risk of CMV infection during pregnancy.
• Trials are needed to identify educational interventions that are effective in preventing CMV infection.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MJC and KFD contributed equally to the drafting and critical revision of the manuscript. Both authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Larry Anderson, Pamela Buchalter, Gail Demmler, Sheila Dollard, Philip Pellett, Stanley Plotkin, Scott Schmid, Stephanie Staras, and Laura Wagner for helpful discussions and comments.
This research was supported in part by an appointment (KFD) to the Research Participation Program at the Centers for Disease Control and Prevention, National Center for Infectious Diseases, Division of Viral and Rickettsial Diseases administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the CDC.
==== Refs
Centers for Disease Control and Prevention Progress toward global eradication of poliomyelitis, 2002 MMWR Morb Mortal Wkly Rep 2003 52 366 369 12749477
Watson JC Hadler SC Dykewicz CA Reef S Phillips L Measles, mumps, and rubella--vaccine use and strategies for elimination of measles, rubella, and congenital rubella syndrome and control of mumps: recommendations of the Advisory Committee on Immunization Practices (ACIP) MMWR Recomm Rep 1998 47 1 57 9639369
Centers for Disease Control and Prevention Progress toward elimination of Haemophilus influenzae type b invasive disease among infants and children--United States, 1998-2000 MMWR Morb Mortal Wkly Rep 2002 51 234 237 11925021
Hankin JR Fetal alcohol syndrome prevention research Alcohol Res Health 2002 26 58 65 12154653
Centers for Disease Control and Prevention Spina bifida and anencephaly before and after folic acid mandate--United States, 1995-1996 and 1999-2000 MMWR Morb Mortal Wkly Rep 2004 53 362 365 15129193
Sullivan JL Luzuriaga K The changing face of pediatric HIV-1 infection N Engl J Med 2001 345 1568 1569 11794226 10.1056/NEJM200111223452111
Yow MD Demmler GJ Congenital cytomegalovirus disease--20 years is long enough N Engl J Med 1992 326 702 703 1310526
Institute of Medicine(U.S.) Committee to Study Priorities for Vaccine Development Vaccines for the 21st Century: A Tool for Decision making 2000 Washington D.C., National Academy Press 460
Dobbins JG Stewart JA Demmler GJ Surveillance of congenital cytomegalovirus disease, 1990-1991. Collaborating Registry Group Mor Mortal Wkly Rep CDC Surveill Summ 1992 41 35 39
Alford CA Stagno S Pass RF Britt WJ Congenital and perinatal cytomegalovirus infections Rev Infect Dis 1990 12 Suppl 7 S745 S753 2173104
Brown HL Abernathy MP Cytomegalovirus infection Semin Perinatol 1998 22 260 266 9738990
Stagno S Remington JS and Klein JO Cytomegalovirus Infectious diseases of the fetus and newborn infant 2001 Philadelphia, W.B. Saunders Company 389 424
Zanghellini F Boppana SB Emery VC Griffiths PD Pass RF Asymptomatic primary cytomegalovirus infection: virologic and immunologic features J Infect Dis 1999 180 702 707 10438357 10.1086/314939
Pass RF Stagno S Dworsky ME Smith RJ Alford CA Excretion of cytomegalovirus in mothers: observations after delivery of congenitally infected and normal infants J Infect Dis 1982 146 1 6 6282987
Noyola DE Demmler GJ Williamson WD Griesser C Sellers S Llorente A Littman T Williams S Jarrett L Yow MD Cytomegalovirus urinary excretion and long term outcome in children with congenital cytomegalovirus infection. Congenital CMV Longitudinal Study Group Pediatr Infect Dis J 2000 19 505 510 10877163 10.1097/00006454-200006000-00003
Montgomery JR Mason EOJ Williamson AP Desmond MM South MA Prospective study of congenital cytomegalovirus infection South Med J 1980 73 590 3, 595 6246642
Murph JR Souza IE Dawson JD Benson P Petheram SJ Pfab D Gregg A O'Neill ME Zimmerman B Bale JFJ Epidemiology of congenital cytomegalovirus infection: maternal risk factors and molecular analysis of cytomegalovirus strains Am J Epidemiol 1998 147 940 947 9596472
Stagno S Dworsky ME Torres J Mesa T Hirsh T Prevalence and importance of congenital cytomegalovirus infection in three different populations J Pediatr 1982 101 897 900 6292387
Pass RF Hutto C Ricks R Cloud GA Increased rate of cytomegalovirus infection among parents of children attending day-care centers N Engl J Med 1986 314 1414 1418 3010113
Adler SP Cytomegalovirus and child day care: risk factors for maternal infection Pediatr Infect Dis J 1991 10 590 594 1653939
Stagno S Whitley RJ Herpesvirus infection of pregnancy N Engl J Med 1985 313 1270 1274 2997607
Preece PM Blount JM Glover J Fletcher GM Peckham CS Griffiths PD The consequences of primary cytomegalovirus infection in pregnancy Arch Dis Child 1983 58 970 975 6197937
Stagno S Pass RF Cloud G Britt WJ Henderson RE Walton PD Veren DA Page F Alford CA Primary cytomegalovirus infection in pregnancy. Incidence, transmission to fetus, and clinical outcome JAMA 1986 256 1904 1908 3020264 10.1001/jama.256.14.1904
Plotkin SA Is there a formula for an effective CMV vaccine? J Clin Virol 2002 25 S13 - S21 12361753 10.1016/S1386-6532(02)00093-8
Kimberlin DW Lin CY Sanchez PJ Demmler GJ Dankner W Shelton M Jacobs RF Vaudry W Pass RF Kiell JM Soong SJ Whitley RJ Effect of ganciclovir therapy on hearing in symptomatic congenital cytomegalovirus disease involving the central nervous system: a randomized, controlled trial J Pediatr 2003 143 16 25 12915819 10.1016/S0022-3476(03)00192-6
Michaels MG Greenberg DP Sabo DL Wald ER Treatment of children with congenital cytomegalovirus infection with ganciclovir Pediatr Infect Dis J 2003 22 504 509 12799506 10.1097/00006454-200306000-00005
Fowler KB McCollister FP Dahle AJ Boppana S Britt WJ Pass RF Progressive and fluctuating sensorineural hearing loss in children with asymptomatic congenital cytomegalovirus infection J Pediatr 1997 130 624 630 9108862
Guralnick MJ Effectiveness of early intervention for vulnerable children: a developmental perspective Am J Ment Retard 1998 102 319 345 9475942 10.1352/0895-8017(1998)102<0319:EOEIFV>2.0.CO;2
Kinney JS Onorato IM Stewart JA Pass RF Stagno S Cheeseman SH Chin J Kumar ML Yaeger AS Herrmann KL . Cytomegaloviral infection and disease J Infect Dis 1985 151 772 774 2985708
Onorato IM Morens DM Martone WJ Stansfield SK Epidemiology of cytomegaloviral infections: recommendations for prevention and control Rev Infect Dis 1985 7 479 497 2994198
Adler SP Finney JW Manganello AM Best AM Prevention of child-to-mother transmission of cytomegalovirus by changing behaviors: a randomized controlled trial Pediatr Infect Dis J 1996 15 240 246 8852913 10.1097/00006454-199603000-00013
Demmler GJ Feigin RG and Cherry JD Cytomegalovirus Textbook of pediatric infectious disease 1998 4th Philadelphia, W.B. Saunders Company 1732 1751
American College of Obstetricians and Gynecologists Perinatal viral and parasitic infections ACOG Practice Bulletin 20 2000 20 Washington, DC, ACOG
Vielfaure MKM Milhalchuk D Cytomegalovirus: no prevention, no cure Can Nurse 1995 91 25 28
Pass RF Cytomegalovirus infection Pediatr Rev 2002 23 163 170 11986492
Gaytant MA Steegers EA Semmekrot BA Merkus HM Galama JM Congenital cytomegalovirus infection: review of the epidemiology and outcome Obstet Gynecol Surv 2002 57 245 256 11961482 10.1097/00006254-200204000-00024
Nigro G Anceschi MM Cosmi EV Clinical manifestations and abnormal laboratory findings in pregnant women with primary cytomegalovirus infection BJOG 2003 110 572 577 12798474
Larson E A causal link between handwashing and risk of infection? Examination of the evidence Infect Control 1988 9 28 36 3276640
Revello MG Gerna G Diagnosis and management of human cytomegalovirus infection in the mother, fetus, and newborn infant Clin Microbiol Rev 2002 15 680 715 12364375 10.1128/CMR.15.4.680-715.2002
Finney JW Miller KM Adler SP Changing protective and risky behaviors to prevent child-to-parent transmission of cytomegalovirus J Appl Behav Anal 1993 26 471 472 8307832
Adler SP Finney JW Manganello AM Best AM Prevention of child-to-mother transmission of cytomegalovirus among pregnant women J Pediatr 2004 145 485 491 15480372 10.1016/j.jpeds.2004.05.041
Curtis V Cairncross S Effect of washing hands with soap on diarrhoea risk in the community: a systematic review Lancet Infect Dis 2003 3 275 281 12726975 10.1016/S1473-3099(03)00606-6
Ryan MA Christian RS Wohlrabe J Handwashing and respiratory illness among young adults in military training Am J Prev Med 2001 21 79 83 11457626 10.1016/S0749-3797(01)00323-3
Roberts L Smith W Jorm L Patel M Douglas RM McGilchrist C Effect of infection control measures on the frequency of upper respiratory infection in child care: a randomized, controlled trial Pediatrics 2000 105 738 742 10742313 10.1542/peds.105.4.738
Hammond B Ali Y Fendler E Dolan M Donovan S Effect of hand sanitizer use on elementary school absenteeism Am J Infect Control 2000 28 340 346 11029132 10.1067/mic.2000.107276
Ockene J Ma Y Zapka J Pbert L Valentine GK Stoddard A Spontaneous cessation of smoking and alcohol use among low-income pregnant women Am J Prev Med 2002 23 150 159 12350446 10.1016/S0749-3797(02)00492-0
Lawrence D Graber JE Mills SL Meissner HI Warnecke R Smoking cessation interventions in U.S. racial/ethnic minority populations: an assessment of the literature Prev Med 2003 36 204 216 12590996 10.1016/S0091-7435(02)00023-3
Centers for Disease Control and Prevention CMV 2004
Osterholm MT Reves RR Murph JR Pickering LK Infectious diseases and child day care Pediatr Infect Dis J 1992 11 S31 - S41 1513610
Smith GC Pell JP Parachute use to prevent death and major trauma related to gravitational challenge: systematic review of randomised controlled trials BMJ 2003 327 1459 1461 14684649 10.1136/bmj.327.7429.1459
Centers for Disease Control and Prevention HIV/AIDS Surveillance Report 2001 13(No.2) 39
Groseclose SL Brathwaite WS Hall PA Connor FJ Sharp P Anderson WJ Fagan RF Aponte JJ Jones GF Nitschke DA Chang MH Doyle T Dhara R Jajosky RA Hatmaker JD Summary of notifiable diseases--United States, 2002 MMWR Morb Mortal Wkly Rep 2004 51 1 84 15123988
May PA Gossage JP Estimating the prevalence of fetal alcohol syndrome. A summary Alcohol Res Health 2001 25 159 167 11810953
Habbick BF Nanson JL Snyder RE Casey RE Mortality in foetal alcohol syndrome Can J Public Health 1997 88 181 183 9260359
Nembhard WN Waller DK Sever LE Canfield MA Patterns of first-year survival among infants with selected congenital anomalies in Texas, 1995-1997 Teratology 2001 64 267 275 11745833 10.1002/tera.1073
Vintzileos AM Ananth CV Smulian JC Day-Salvatore DL Beazoglou T Knuppel RA Cost-benefit analysis of prenatal diagnosis for Down syndrome using the British or the American approach Obstet Gynecol 2000 95 577 583 10725493 10.1016/S0029-7844(99)00613-4
Davidoff MJ Petrini J Damus K Russell RB Mattison D Neural tube defect-specific infant mortality in the United States Teratology 2002 66 Suppl 1 S17 S22 12239739 10.1002/tera.90005
|
15967030
|
PMC1182379
|
CC BY
|
2021-01-04 16:28:54
|
no
|
BMC Public Health. 2005 Jun 20; 5:70
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-70
|
oa_comm
|
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-111596975110.1186/1471-2229-5-11Research ArticleAnalysis of xylem sap proteins from Brassica napus Kehr Julia [email protected] Anja [email protected] Patrick [email protected] Department Lothar Willmitzer, Max-Planck-Institute of Molecular Plant Physiology, 14424 Potsdam, Germany2005 21 6 2005 5 11 11 23 11 2004 21 6 2005 Copyright © 2005 Kehr 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
Substance transport in higher land plants is mediated by vascular bundles, consisting of phloem and xylem strands that interconnect all plant organs.
While the phloem mainly allocates photoassimilates, the role of the xylem is the transport of water and inorganic nutrients from roots to all aerial plant parts. Only recently it was noticed that in addition to mineral salts, xylem sap contains organic nutrients and even proteins. Although these proteins might have important impact on the performance of above-ground organs, only a few of them have been identified so far and their physiological functions are still unclear.
Results
We used root-pressure xylem exudate, collected from cut Brassica napus stems, to extract total proteins. These protein preparations were then separated by high-resolution two-dimensional gel electrophoresis (2-DE). After individual tryptic digests of the most abundant coomassie-stained protein spots, partial peptide sequence information was deduced from tandem mass spectrometric (MS/MS) fragmentation spectra and subsequently used for protein identifications by database searches. This approach resulted in the identification of 69 proteins. These identifications include different proteins potentially involved in defence-related reactions and cell wall metabolism.
Conclusion
This study provides a comprehensive overview of the most abundant proteins present in xylem sap of Brassica napus. A number of 69 proteins could be identified from which many previously were not known to be localized to this compartment in any other plant species. Since Brassica napus, a close relative of the fully sequenced model plant Arabidopsis thaliana, was used as the experimental system, our results provide a large number of candidate proteins for directed molecular and biochemical analyses of the physiological functions of the xylem under different environmental and developmental conditions. This approach will allow exploiting many of the already established functional genomic resources, like i.e. the large mutant collections, that are available for Arabidopsis.
==== Body
Background
The higher plant body consists of functionally specialized organs such as leaves, stem, fruits, flowers, and roots. Because plants are immobile and have to cope with changes in their environment, interaction of different organs is essential to coordinate growth, development and defence reactions also between the most distant plant parts [1]. Transport of nutrients and information molecules over long distances is, in most instances, mediated by the vascular bundles that mainly consist of xylem and phloem. The xylem constitutes a channel system for water and inorganic nutrient transport from roots to above-ground plant organs. Xylem transport occurs through the dead and hollow xylem vessels that belong to the apoplastic space. In addition to inorganic salts, organic nutrients, such as amino acids, sugars, and organic acids are translocated through the xylem from roots to aerial organs [2-4]. The above-ground plant parts are dependent on the inorganic and organic compounds that are taken up or produced by the roots and distributed by the xylem network. A specific example of root-produced organic compounds that are translocated in xylem sap are plant hormones (i.e. cytokinin, abscisic acid, auxins, gibberellins), which are known to be important in the control of different aspects of plant development in above-ground organs [1]. For example, they are involved in the coordination between root and shoot differentiation, growth, and development [5-9].
Earlier reports have already described the presence of proteins in the xylem sap of numerous plants, like watermelon [10], apple, peach, pear [11], cucumber [12], squash [13], rice [14], and tomato [15] and recently, biochemical approaches have revealed the identities of a few of these xylem sap proteins. Peroxidases and chitinases [11,16,17], pathogenesis-related (PR) proteins [15], a glycine-rich protein [18], a cysteine-rich protein [19] and a 30 kD lectin [12] have been found. It is speculated that some of these proteins might exert specific physiological functions in aerial organs [13], although the biological significance and the regulation of these proteins are not fully understood [1]. It has been shown that xylem protein patterns change in response to infection by pathogenic fungi [15,19] and there are indications that interactions between proteins and pathogens within the xylem vessels, at least partly, determine the grade of resistance or susceptibility of tomato plants towards the vascular wilt pathogen Fusarium oxysporum [15]. Also after bacterial infection in rice, a xylem peroxidase was described to accumulate in xylem vessels [14]. However, further detailed evidence supporting the role of xylem sap proteins in plant defence reactions is so far missing.
Recent results indicate that the expression of xylem proteins can be highly regulated also by other factors than pathogen invasion. The root-specific expression of 30 kD xylem sap protein (XSP30), for example, is controlled by a circadian clock and shows diurnal fluctuations. This protein appears to be influenced additionally by unknown gibberellin-induced mediators that are produced by leaves and transported to roots to influence XSP30 expression [1].
Another important issue is the origin of xylem sap proteins, because xylem vessels are dead cells that are incapable of transcription and translation. Proteins may reach the xylem sap either specifically or they could originate from developing tracheary elements or flushed away from adjacent parenchyma cells [11] or the vessel cell walls. Currently, there is no data on the synthesis sites of most xylem sap proteins available. The few proteins analyzed so far appear to be expressed root-specifically in xylem parenchyma and pericycle cells and are supposed to be actively secreted into xylem sap by root cells, as has been shown previously for XSP30 [12] and two glycine-rich xylem sap proteins [20]. The secretion of proteins into xylem sap is, like for other apoplastic proteins, most likely mediated by an amino-terminal signal peptide [21,22], which has been detected in the sequences of most of the thus far known xylem sap proteins [12,16,18,19].
Most recent approaches to analyze xylem sap proteins have been performed by low resolution one-dimensional polyacrylamide gel electrophoresis (1-DE) and resulted in the identification of a limited number of xylem specific proteins from different plant species [15,16,19]. Based on this lack of comprehensive protein information of the xylem sap, the aim of the present study was to provide an overview of the proteins present in this plant specific transport fluid, by separating them on high-resolution two-dimensional (2-DE) polyacrylamide gels and subsequently identify a substantial number by tandem mass spectrometry (MS/MS).
Results and discussion
Xylem protein extraction, separation and identification
In this study, proteins from xylem sap of adult Brassica napus plants were precipitated by acetone, separated by 2-DE and partial amino acid sequences were determined by tandem mass spectrometry. As demonstrated before, the xylem sap collected with the employed method showed no detectable contaminations from phloem sap or other adjacent cellular compartments, if the cut stems are thoroughly rinsed with water before starting the sample collection [16]. In addition, the xylem sap protein patterns of Brassica napus plants were clearly distinct from 2-DE spot patterns derived from purified phloem sap or whole Brassica napus stem tissue protein extracts (unpublished data).
Coomassie staining of the 2-DE gels from this root pressure exudate allowed visualization of approximately 300 protein spots (Figures 1 &2). In the presented experiments the most intense spots, which could be reproducibly retrieved from several independent xylem sap protein extractions (Figure 1), were excised from 2-DE gels, digested in situ with the site-specific protease trypsin, before partial amino acid sequences were determined by tandem mass spectrometry, followed by database searches for protein identification. Using this approach, a number of 69 protein spots could be reliably identified from at least two independent gels by the high similarity of the determined partial amino acid sequences to plant proteins in the NCBI plant protein database (Figure 2, Table 1).
Most of these proteins, 64 protein spots, matched to proteins from the fully sequenced genome of the model plant Arabidopsis thaliana. This was not unexpected, since the genomes of Arabidopsis thaliana and Brassica napus show a very high degree of identity [23]. The other 5 proteins either directly corresponded to database entries from Brassica napus, or matched to database entries from Brassica rapa or rice (Table 1).
The observed molecular masses of the majority of the identified proteins matched well to the theoretical molecular weights predicted from the amino acid sequences of the database entries (Table 1). Only a few proteins showed significant differences between the theoretical and the observed masses, i.e. the protein spots for the curculin lectins (spots 35, 36, 38–41) showed significantly lower molecular weight on SDS PAGE gels than expected for the Arabidopsis homologues the identification was based on (Table 1). This could indicate on the one hand that some of the Brassica napus curculin lectins are smaller than their Arabidopsis homologues or that they, as has been shown for other proteins [24,25], show a higher mobility in SDS PAGE gels than predicted. Alternatively, these Brassica curculin lectins might be products of proteolytic processes trimming these gene products to a smaller size. Additionally it could be observed that most of the proteases identified from the xylem sap showed lower than predicted molecular weights on the 2-DE gels (spots 8, 11, 19, 34). In the case of the identified subtilases, two forms with different observed masses, namely one of about 56 kD (spots 55, 56, 59, 60, 63), which represents a smaller than predicted isoform and one of 80 kD (spots 65–67), which corresponds to the expected mass of the identified protein, were observed. These differences might be explained by the fact that proteases often show auto-proteolytic activity that results in molecules of different sizes, unprocessed large pro-proteases and proteolytically processed smaller proteins. It has been shown for one subtilases that this step is needed to activate the enzyme [26].
In contrast to the observed smaller than expected proteins, one of the identified germin-like proteins (spot 69) displayed a higher molecular mass than expected, which might be due to the reported observation that this class of proteins is known to occur as oligomers in vivo [27]. Alternatively, as has been shown for other proteins [28-30], this germin-like protein could show an abnormal, reduced mobility in the SDS PAGE separation.
Analyzing the amino acid sequences of the homologous proteins derived from the plant protein database searches, revealed a common characteristic that is probably essential for the apoplastic localization of these proteins: with the exception of 2 proteins (ubiquitin in spots 1 & 2 and beta-1,3-glucanases in spots 49 & 50), all identified proteins are likely targeted to the secretion pathway when analyzed by SignalP, a program that predicts N-terminal peptides and signal peptidase I cleavage sites [21,22]. This observation is in full agreement with previous results of xylem sap proteins from different plant species, underlining that xylem sap proteins belong to the class of secreted proteins [15,16].
While some of the identified proteins appeared to be present in only one single protein spot (PR1a, spot 4; thaumatin, spot 18; a cyclase family protein, spot 44; polygalacturonase inhibitor, spot 51), in some cases the partial sequences led to the identification of the same protein in two or more different protein spots of similar molecular weight, but with a variable isoelectric point (ubiquitin, spots 1 & 2; beta-glucanases, spots 49 & 50; pectin methylesterase inhibitor, spots 6 & 9; xyloglucan:xyloglycosyl transferase, spots 21 & 22; polygalacturonase, spots 58 & 64). The occurrence of similar or identical gene products in different spots is a common feature observed in 2-DE analysis and can be caused by post-translationally modifications of distinct amino acids of one single gene product [31]. Alternatively, since our MS analysis does not provide full sequence coverage of the identified proteins, these protein spots can represent products from highly similar, but distinct genes of a gene family. The appearance of multiple protein spots, derived from different but functionally or sequence-related protein families, could also be observed for the glycine-rich cell wall proteins (spots 3, 5, 13); the germin-like proteins (spots 17 & 69), the chitinases (spots 14, 15, 16, 20, 23, 27, 29); the lectins (spots 25, 26, 28, 35, 36, 38–41, 57, 61, 62); the peroxidases (spots 24, 30–33, 37, 42, 43, 45, 46–48, 52–54) and for some proteases (spots 7, 8, 11, 19, 34, 55, 56, 59, 60, 63, 65–68), indicating a certain degree of redundant expression for specific gene products.
Possible functions of the identified xylem sap proteins
Peroxidases
One protein family detected in our xylem 2-DE gels comprises peroxidases. This class of proteins has a multitude of possible functions [32], including the generation of reactive oxygen species and the regulation of H2O2 levels in planta. It has been shown that peroxidases can be involved in plant cell wall strengthening by different mechanisms, e.g. by cross-linking and polymerizing proline-rich proteins in the cell wall [33,34] or by catalyzing lignin deposition [35]. Interestingly, this family of proteins have been found in xylem sap from all plants that have been analyzed so far [11,16]. In the current analysis, peroxidases were the largest protein group, containing 6 different proteins, derived from a total of 15 different protein spots (spots 24, 30–33, 37, 42, 43, 45, 46–48 and 52–54). Apart from the previously mentioned cell wall dependent activities, a large number of additional possible functions were attributed to this ubiquitous protein family [34].
Proteases
Another large and functionally diverse group of proteins within the plant genome are the proteases. They represent the second largest functional group identified from our 2-DE gels, with 14 identified protein spots (7, 8, 11, 19, 34, 55, 56, 59, 60, 63, 65–68), representing 6 different, unrelated proteases. The eclectic mix of proteases identified from rape xylem sap represents a cross section of the large protease repertoire in plants. For example, coding sequences for a number of more than 550 different potential proteases, which are grouped in more than 50 different families, are found in the Arabidopsis genome [36]. The proteases identified from our analysis could be grouped into 5 different protease families: S8 (subtilisin-like serine proteases, spots 59, 60, 63, 65–67, 68), S10 (serine carboxypeptidase, spot 34), C1A (papain-like cysteine proteases, spots 7, 8), A1 (pepsin-like aspartic acid proteases, spot 11) and T3 (threonine proteases, spot 19). Only a few of the proteases identified in the current study have been previously characterized from other plant species. It was shown that a homologue of the cucumisin-like S8 subtilisin protease ARA12 (spot 68) is involved in actinorhizal nodule development in Alnus glutinosa (European alder) roots [37], while the Arabidopsis homologue seems to be expressed more ubiquitously, with a certain specificity in silique development [37]. The reports concerning other subtilisin-like proteases demonstrate their importance in the regulation of developmental processes, like the distribution and density of stomata on Arabidopsis thaliana leaves [26,38]. A function in xylem development could be associated with the C1A papain-like protease, XCP2, which was identified from 2 protein spots from our 2-DE gels (spots 7, 8). This protein and its close homologue XCP1 were shown to be expressed in xylem tissue of Arabidopsis [39] and these proteins are thought to be involved in xylem formation.
For the other proteases identified from the xylem sample nothing, except from the facts that they all contain specific protease domains and that they contain a secretion pathway sequence, explaining their presence in the xylem, is known.
Defence-related proteins
One group of proteins that has been closely associated with plant defence are the pathogenesis-related (PR) proteins [35]. In our 2-DE gels we detected a single, low MW protein spot (spot 4, Figure 2) that is similar to a protein belonging to the family of PR1 proteins. This result confirms the previously observed occurrence of a PR-1a-like protein in fungus-infected tomato xylem sap [15], although in our study the rape plants were not actively challenged with pathogens. The PR1 family proteins were the first identified PR proteins and show antimicrobial properties [40]. However, the molecular or biochemical basis by which these proteins provide this function remained up to date elusive.
Another class of proteins that is often associated to the PR proteins are beta-1, 3-glucanases (BGs) and chitinases, which are believed to mediate defence responses because of their potential to degrade fungal cell walls [41]. In our analysis we found 3 different putative endochitinases, belonging to the two different classes I and IV, in a total of 7 different protein spots (spots 14, 15, 16, 20, 23, 27 and 29 in Figure 2), while one BG protein could be identified within two different protein spots (spots 49 & 50), indicating a possible co-transcriptional or post-translational modification of these proteins. The occurrence of chitinases and chitinase activity in xylem sap of different plant species has been observed in earlier studies [16,17], while BG proteins were found thus far only in tomato [11,15].
Interestingly, chitinases and glucanases have been suggested to act in a synergistic manner with thaumatin-like proteins (spot 18) that can bind to β-1,3-glucans [42] and have not been described to occur in xylem sap of healthy, unchallenged plants before, while they were found after fungal infection [15].
Lectins
Lectins are carbohydrate-binding proteins that can bind glycans of glycoproteins, glycolipids, or polysaccharides with high affinity. It is assumed that lectins play fundamental biological roles in plants because they are found in many different species and in many different organs and tissues [43].
Legume lectins (spots 25, 26, 28) are one of the largest lectin families with more than 70 lectins reported [44]. Functionally these proteins specifically recognize diverse sugar structures and mediate a variety of biological processes such as cell-cell and host-pathogen interactions and innate immune responses [45]. Curculin lectins (spot 38–41, 57, 61, 62) are, similar to TLPs, sweet-tasting proteins, which often maintain possible mannose-binding sites. Nevertheless, earlier studies have shown that the three mannose-binding sites of curculin from Curculigo latifolia are devoid of mannose-binding activity [46] indicating that curculin lectins might have different additional thus far unknown functions.
Cell wall metabolism and remodelling
The other 20 protein spots identified from Brassica xylem sap have predicted functions probably connected to cell wall metabolism and remodelling.
The proteins found in our study represent a set of proteins discussed to be involved in cell wall stabilization and repair, like the glycine-rich proteins (spots 3, 5, 10) [20,47,48], and the multi-functional xyloglucan:xyloglycosyl transferases (spot 21 & 22), which have all been shown to modify cell wall structure during growth or stress responses [49].
In addition, polygalacturonases (spots 58 & 64) and polygalacturonase-as well as pectin methylesterase inhibitors (PGIPs, spot 51 & PMEIs, spots 6 & 9) have been detected in the xylem sap samples analyzed in this study. While there is not much known about a possible roles of PMEIs in plants, the widespread PGIPs have been thoroughly investigated in different plant species [50]. PGIPs are typically induced by pathogen infection and stress-related signals [50,51]. Usually they are effective only against fungal PGs and do not influence endogenous plant PGs [52].
Conclusion
The present study demonstrates that Brassica napus, due to its high gene sequence identity to the model plant Arabidopsis thaliana [23], provides an excellent source for the large-scale analysis of xylem sap proteins. In the course of our analysis, 69 abundant xylem sap proteins were successfully identified. Nearly all of these proteins contained a N-terminal sequence, targeting them to the secretion pathway [53], which correlates to the fact that the xylem is a part of the apoplastic space [5].
The mixture of rape xylem sap proteins identified in the present study is composed of proteins with various potential functions. In addition to a large number of peroxidases and proteases, different potentially defence-related proteins, lectins, and a number of proteins involved in cell wall modification, remodelling and strengthening could be detected. Further experimentation, employing biochemical and immuno-histochemical analysis of the identified proteins, in conjunction with enzyme assays, will be needed to dissect the precise physiological functions of these xylem sap proteins. This attempt should be largely facilitated by the close relatedness of Brassica napus to Arabidopsis, where a lot of functional genomic resources are already accessible.
Methods
Plant material
Brassica napus plants (cv. Drakkar, Serasem GIE, la Chapelle d'Armentiers, France) were grown in 19 cm pots containing steam-sterilized soil (Einheitserde® Typ T) in a greenhouse under controlled conditions (16 h light, 8 h dark, 25°C day, 20°C night, 55% relative air humidity). Plants were automatically watered thrice a day with tap water containing Hakaphos® spezial as a fertilizer.
Sample preparation
Xylem samples were obtained after cutting stems of flowering 12-week-old plants, approximately 5 cm above soil level. After thorough washing of the surface on the root side with distilled water, they were blotted dry with filter paper and the exuding fluid was then collected with a hand held pipette until sufficient sample volumes (usually 9–12 ml) were obtained. Aliquots of 3–4 ml xylem sap were each collected from 5 plants in parallel and immediately expelled into 7 ml of precipitation solution [90% (v/v) acetone, 10% (v/v) methanol, 10 mM DTT] provided in falcon tubes on ice and precipitated over night at -20°C. The collection of sufficient xylem sap usually took about 30 min. The precipitated proteins were collected by centrifugation for 15 min at 4000 g at 4°C, washed with acetone, the supernatant was discarded and the pellet was air-dried.
In parallel, protein concentrations were determined omitting acetone precipitation with the Bradford method (BioRad, Munich, Germany), using 100 μl of xylem samples.
Gel electrophoresis
For 2-DE, the precipitated xylem sap proteins (~120 μg) were re-suspended in 50 μl first dimension buffer [2 M thiourea, 7 M urea, 4% 3-[(3-chloramidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS) and 10 mM DTT]. The following protein separation was performed in a 2-DE system described previously [54,55] but in 16 cm times 1.5 mm tube gels for the isoelectric focusing and 18 × 16 × 0,1 cm SDS polyacrylamide gels in the second dimension.
After protein separation, the gels were stained with coomassie brilliant blue G250 according to [55]. Molecular weights were calculated from a run of the PrecisionPlus All Blue Protein Standard (BioRad) performed on a separate gel.
Mass spectrometry, database searches and sequence analysis
Sample pre-treatment, trypsin digests and determination of partial amino acid sequences was performed as described previously by Walz. et al [56]. In short, protein spots were excised, destained, dehydrated, and digested overnight with modified Trypsin (Roche Diagnostics, Mannheim, Germany). Proteolytic peptides were extracted by trifluoroacetic acid/acetonitrile. The extracts were then vacuum dried and desalted on C18 pipette tips (Millipore) before analysis by tandem MS in an electrospray ionization quadrupole time-of-flight tandem mass spectrometer (Q-TOFMS, Micromass, Altrincham, UK), resulting in peptide fragmentation spectra. Raw fragmentation spectra were recalculated with the MaxEnt3 algorithm and then partial sequences were manually deduced using the PepSeq software from Micromass (Altrincham, UK). Only peptides that appeared in digests of the same spot from at least two different gels were taken into account.
Database similarity searches with the derived sequence stretches were performed using the short sequence blast algorithm against the non-redundant protein database limited to green plants. If several partial amino acid sequences could be obtained, the combined sequences were used for an additional BLAST search to calculate BLAST E-values (the resulting E-values are listed in Table 1). Results with BLAST E-values < E-5 were regarded as being reliable protein identifications. Since E-values are not necessarily meaningful for short sequences, we additionally regarded proteins as identified, from which only a single partial sequence could be obtained that missed the E-value threshold, but that showed a stretch of at least 12 amino acids identical to the corresponding database sequence.
Cleavage sites for signal peptidase I, characterizing secreted proteins, were predicted with SignalP version 3.0 [21,22].
Authors' contributions
JK participated in the design and coordination of the study, in MS data processing, database searches and writing of the manuscript. AB carried out the mass spectrometric measurements. PG optimized and performed high-resolution 2-DE and participated in the design of the study as well as in manuscript preparation.
All authors read and approved the final version of the manuscript.
Acknowledgements
This work was financed by the MPG. We thank Professor Joachim Klose, Charité Berlin, for support with 2-DE.
Figures and Tables
Figure 1 Reproducibility of xylem sap 2-DE patterns. 2-DE patterns of xylem sap proteins from three independent sets of flowering, 12-week-old Brassica napus plants that were grown independently during a period of eight month. The left hand image shows the protein gel represented in Figure 2.
Figure 2 Protein identifications from a representative xylem sap 2-DE gel. Representative 2-DE gel of xylem sap proteins from 12-week-old Brassica napus plants. Proteins were identified by partial sequences determined by tandem mass spectrometry. Numbers indicate spots from which proteins could be identified. Sequences and identifications are listed in Table 1. Asterisks indicate spots that were analyzed but led to no or low quality mass spectra that did not allow protein identifications. Molecular masses in kD are indicated on the right hand side of the gel.
Table 1 List of identified xylem sap proteins from Brassica napus. Identifications of xylem sap proteins from the 2-DE gel shown in Figure 2. Sequences were determined from fragmentation spectra obtained by ESI-Q-TOF tandem MS. The resulting partial amino acid sequences were used for database searches with a BLAST algorithm optimized for short sequences. The isobaric amino acids isoleucine (I) and leucine (L), although not distinguishable by MS, are displayed as specified in the corresponding database sequences. BLAST E-values were determined using all non-redundant peptide sequences from one protein spot for another BLAST search. Bold letters in the partial sequences indicate amino acids identical to the database entry. Results of N-terminal secretion sequence predictions are indicated in the right column (Y= secretion sequence predicted, N= no secretion sequence predicted), * indicates that the observed molecular mass was lower and ** indicates that the observed mass was higher than expected.
Spot No Similarity to Organism Acc. No. Sequence identity E value expected Mr observed Mr aS
1 Ubiquitin Arabidopsis thaliana At2g47110 TITLEVESSDTIDNVK 16/16 (100%) 9.00E-08 8525 12064 N
2 Ubiquitin Arabidopsis thaliana At2g47110 TITLEVESSDTIDNVK 16/16 (100%) 9.00E-08 8525 11944 N
3 Glycine-rich cell wall protein Arabidopsis thaliana At4g30450 AGSSAGSFAGSR 12/12 (100%) 0.011 9530 13863 Y
4 Pathogenesis-related protein PR1a Brassica napus AAB09587 SPQDYVNAHNQAR 13/13 (100%) 7.00E-39 17532 18145 Y
LAAYAQNYADR 11/11 (100%)
QDSPQDYVNAHNQAR 15/15 (100%)
GPVQWDGTLAAYAQNYADR 19/19 (100%)
AGSSADFSGVSAVNLWVNEK 20/20 (100%)
5 Glycine-rich cell wall protein Oryza sativa BAD62284 GGGYGEGGGYGGG 15/16 (94%) 2.00E-10 14152 17822 Y
GGGYWGGGYGGGYGGG
6 Invertase/pectin methylesterase inhibitor Arabidopsis thaliana At1 g47960 GSSADTSGLALILVDK 15/16 (94%) 2.00E-32 22326 18297 Y
GVIDAGVEAAV 10/11 (91%)
ETPDFNLCVACPLDSDPR 15/18 (83%)
GSSADTSGLALILVDKIK 17/18 (94%)
ADVPEAIECCSK 10/12 (83%)
FGEDGVIDAGVEAAV 14/15 (93%)
7 Papain-type cysteine proteinase XCP2 Arabidopsis thaliana At4g35350 LDHGVAAVGYGSSK 14/14 (100%) 4.00E-18 39618 16,603* Y
ALAHQPVSVAIEASGR 16/16 (100%)
GVDLDHGVAAVGYGSSK 16/17 (94%)
8 Papain-type cysteine proteinase XCP2 Arabidopsis thaliana At4g35350 LDHGVAAVGYGSSK 14/14 (100%) 4.00E-18 39618 16,977* Y
ALAHQPVSVAIEASGR 16/16 (100%)
GVDLDHGVAAVGYGSSK 16/17 (94%)
9 Invertase/pectin methylesterase inhibitor Arabidopsis thaliana At1g47960 GSSADTSGLALILVDK 15/16 (94%) 9.00E-38 22326 18297 Y
ETPDFNLCVACPLDSDPR 15/18 (83%)
GSSADTSGLALILVDKIK 17/18 (94%)
ALDECASR 6/8 (75%)
ADVPEAIECCSK 10/12 (83%)
FGEDGVIDAGVEAAV 14/15 (93%)
10 Glycine-rich protein Arabidopsis thaliana At4g30460 GEGGGYGGGYGGGGD 14/15 (93%) 1.00E-04 13738 19195 Y
11 Aspartyl protease family protein Arabidopsis thaliana At5g07030 LPPSAIAFNPATGAGTIFDSGTVYTR 24/26 (92%) 3.00E-15 46919 21,384* Y
VLIDLPNSR 7/9 (77%)
12 Disease resistance response/ dirigent protein Arabidopsis thaliana At1g64160 VIFDDPVTLDK 8/12 (67%) 3.00E-13 20573 22711 Y
GCLNIMGADL 9/10 (90%)
DLSVVGGTGDFFFSR 14/15 (93%)
13 Disease resistance response/ dirigent protein Arabidopsis thaliana At4g23690 VIFDDPVTLDQNYLSKPVSR 17/20 (85%) 1.00E-11 21412 24544 Y
SVVGGTGDFFMSR 13/13 (100%)
14 Endochitinase class I, putative Arabidopsis thaliana At2g43620 RDTIANVVTLSVFNSIFSK 12/19 (63%) 2.00E-08 30378 25856 Y
QAFISAAAQSSDAYK 13/15 (87%)
FNGLPLLTDPDMVSR 13/15 (87%)
15 Endochitinase class I, putative Arabidopsis thaliana At2g43620 RDTIANVVTLSVFNSIFSK 12/19 (63%) 2.00E-08 30378 25969 Y
QAFISAAAQSSDAYK 13/15 (87%)
FNGLPLLTDPDMVSR 13/15 (87%)
16 Endochitinase class I, putative Arabidopsis thaliana At2g43620 RDTIANVVTLSVFNSIFSK 12/19 (63%) 7.00E-13 30378 25744 Y
QAFISAAAQSSDAYK 13/15 (87%)
FNGLPLLTDPDMVSR 13/15 (87%)
PVLDQGFGATTR 12/12 (100%)
17 Germin-like protein, subfamily 2 Arabidopsis thaliana At1 g02335 YDPDALQDLCVAADK 14/15 (93%) 8.00E-07 23441 26543 Y
KIPGLNTLSV 10/10 (100%)
GEVFVFPR 8/8 (100%)
18 Thaumatin-like protein/ osmotin Brassica rapa AAN23104 RAPNTLAEYALK 12/12 (100%) 4.00E-04 incomplete 32060
19 Gamma-glutamyltranspeptidase 1 Arabidopsis thaliana At4g39640 PPPAPANFIRPGK 13/13 (100%) 4.00E-D5 61190 26,312* Y
VPPPAPANFIRPGK 14/14 (100%)
QFIVQES 7/7 (100%)
20 Endochitinase class I, putative Arabidopsis thaliana At2g43620 TIANVVTLSVFNSIFSK 10/17 (59%) 2.00E-20 30378 28613 Y
QAFISAAAQSSDAYK 13/15 (87%)
FLGLPLLTDPDFVAR 12/15 (80%)
IQITSDNYNYGAAGK 12/17 (71%)
21 Xyloglucan:xyloglucosyl transferase Arabidopsis thaliana At5g57560 GQITNDGELLTLSLDK 14/16 (88%) 3.00E-36 32093 29928 Y
LVPGNSAGTVTTLYLK 16/16 (100%)
WFDPTWFHTYTILSDNPQR 16/19 (84%)
IFTVDGTPIR 10/10 (100%)
NFESVGTLFPNNKPFR 13/16 (81%)
22 Xyloglucan:xyloglucosyl transferase Arabidopsis thaliana At5g57560 GQITNDGELLTLSLDK 14/16 (88%) 3.00E-36 32093 30200 Y
LVPGNSAGTVTTLYLK 16/16 (100%)
WFDPTVVFHTYTILSDNPQR 16/19 (84%)
IFTVDGTPIR 10/10 (100%)
NFESVGTLFPNNKPFR 13/16 (81%)
23 Chitinase class IV, putative Arabidopsis thaliana At2g43570 PSFGSSISK 8/9 (89%) 2.00E-05 29775 31326 Y
AAANSYPSFGSSISK 14/15 (93%)
LNENLLASPEK 10/11 (91%)
VAQDPPSAFK 7/10 (70%)
24 Peroxidase 25 precursor Arabidopsis thaliana At2g41480 STVESHFDSDPTISPGLLR 19/19 (100%) 4.00E-09 35886 32665 Y
EASNLPSPLDSVAVQK 16/16 (100%)
VALDIGSPSNFDVSFFK 15/17 (88%)
25 Legume lectin family protein Arabidopsis thaliana At3g16530 LAVEFDTFQNK 10/11 (91%) 4.00E-06 30509 32360 Y
GYWVQTR 7/7 (100%)
Spot No Similarity to Organism Acc. No. Sequence identity E value expected Mr observed Mr aS
26 Legume lectin family protein Arabidopsis thaliana At3g15356 GDNLFFLGDAELG 11/13(85%) 7.00E-15 29749 32819 Y
EFDTFKNK 7/8 (88%)
HVGININSMTSNVAEK 15/16 (94%)
TITIAPENVK 10/10 (100%)
27 Chitinase class I Brassica napus CAA43708 DSFINAANTFPNFANSVTR 19/19 (100%) 5.00E-11 28733 30200 Y
NGGNSGAVNAR 11/11 (100%)
28 Legume lectin family protein Arabidopsis thaliana At3g16530 LAVEFDTFQNK 9/10 (90%) 5.00E-06 30509 31763 Y
AGYWVQTR 8/8 (100%)
29 Chitinase class I, putative Arabidopsis thaliana At4g01700 GFYPYEAFVEATR 13/13 (100%) 1.00E-08 31464 32819 Y
GPIQLSWNYNYGQAGR 16/16 (100%)
30 Peroxidase, putative Arabidopsis thaliana At4g36430 PGTVSCAADLLTLAAR 14/16 (88%) 4.00E-05 36164 34420 Y
QNLSVLDIVSAAK 10/13 (77%)
GNISPLTGSSGEIR 14/14 (100%)
31 Peroxidase, putative Arabidopsis thaliana At4g36430 DSSVLTGGPCWVVP 13/14 (93%) 7.00E-18 36164 36508 Y
QGLDITDLVALSGSHTIGFSR 21/21 (100%)
NSDQVLFSS 9/9 (100%)
GNISPLTGSSGEIR 14/14 (100%)
32 Peroxidase, putative Arabidopsis thaliana At1g49570 EAVVLTGGPFMAPVPLGR 16/18 (89%) 9.00E-06 38030 41164 Y
GQPDPNLAASSDLLSK 15/16 (94%)
GNIGVFTGSDGVIR 13/14 (93%)
33 Peroxidase ATP5a Arabidopsis thaliana At1g49570 EAVVLTGGPFMAPVPLGR 16/18 (89%) 7.00E-05 38030 40065 Y
GNIGVFTGSDGVIR 13/14 (93%)
Peroxidase 59 precursor Arabidopsis thaliana At5g19890 LVEAYSQSQSLFFR 13/14 (93%) 6.00E-05 35023 40065 Y
34 Serine carboxypeptidase S10 family protein Arabidopsis thaliana At3g17180 GFIVGNPLTDDEYDNK 16/16 (100%) 1.00E-08 54145 40,498* Y
35 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78830 PLVYGGDSPNHDFR 11/14 (79%) 2.00E-09 50343 39,640 Y
VVNNDEFGDYITEYDASYR 16/19 (84%)
GVTGFEILPNGNIVLHDK 17/18 (94%)
TGQPLVYGGDSPNHDFR 13/17 (76%)
36 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78830 PLVYGGDSPNHDFR 11/14 (79%) 2.00E-09 50343 36,326* Y
VVNNDEFGDYITEYDASYR 16/19 (84%)
GVTGFEILPNGNIVLHDK 17/18 (94%)
TGQPLVYGGDSPNHDFR 13/17 (76%)
37 Peroxidase, putative Arabidopsis thaliana At4g33420 PTLSAGLIR 8/9 (89%) 1.00E-17 35975 37062 Y
NDFDNAYFNALQR 13/13 (100%)
SGVLFSDQTLFNSPFTR 15/17 (88%)
38 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78830 PLVYGGDSPNHDFR 11/14 (79%) 3.00E-08 50343 37,631* Y
VVNNDEFGDYITEYDASYR 16/19 (84%)
ILPNGNIVLHDK 12/1 (100%)
TGQPLVYGGDSPNHDFR 13/17 (76%)
39 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78830 PLVYGGDSPNHDFR 11/14 (79%) 3.00E-08 50343 35,967* Y
VVNNDEFGDYITEYDASYR 16/19 (84%)
ILPNGNIVLHDK 12/1 (100%)
TGQPLVYGGDSPNHDFR 13/17 (76%)
40 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78830 PLVYGGDSPNHDFR 11/14 (79%) 3.00E-08 50343 37,824* Y
VVNNDEFGDYITEYDASYR 16/19 (84%)
ILPNGNIVLHDK 12/1 (100%)
TGQPLVYGGDSPNHDFR 13/17 (76%)
41 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78830 PLVYGGDSPNHDFR 11/14 (79%) 3.00E-08 50343 36,326* Y
VVNNDEFGDYITEYDASYR 16/19 (84%)
ILPNGNIVLHDK 12/1 (100%)
TGQPLVYGGDSPNHDFR 13/17 (76%)
42 Peroxidase 59 precursor Arabidopsis thaliana At5g19890 KNAIPNINSAR 10/11 (91%) 0.003 35023 40280 Y
EAYSQSQSLFFR 12/12 (100%)
43 Peroxidase, putative Arabidopsis thaliana At4g33420 PTLAAGLIR 9/9 (100%) 1.00E-09 35975 37824 Y
NTAEKDSPANLSLR 14/14 (100%)
NDFDNAYFNALQR 13/13 (100%)
44 Cyclase family protein Arabidopsis thaliana At4g34180 DYLSFAAFDESPATHK 15/16 (94%) 1.00E-28 28384 34586 Y
GLDYLSFAAFDESPATHK 17/18 (94%)
EFDSSFSGFFTDGAK 13/15 (87%)
LIGLDYLSFAAFDESPATHK 19/20 (95%)
GRDIIPVEALK 11/11 (100%)
45 Peroxidase 21 (ATP2a) Arabidopsis thaliana At2g37130 DASLLLETAR 10/10 (100%) 5.00E-18 36741 35614 Y
PTLDPDYALYLK 11/12 (92%)
ADNGYFHEQFSR 12/12 (100%)
LLSETNPLTGDQGEIR 16/16 (100%)
46 Peroxidase, putative Arabidopsis thaliana At4g33420 PTLAAGLIR 9/9 (100%) 8.00E-40 35975 38018 Y
GFTPQDVVALSGAHTLGVAR 20/20 (100%)
LTTPDSSFDSSFVNTLTK 13/18 (72%)
FDNAYFNALQR 10/11 (91%)
SGVLFSDQTLFNTPATR 17/17 (100%)
47 Peroxidase, putative Arabidopsis thaliana At4g33420 DNTAEKDSPANLSLR 15/15 (100%) 1.00E-25 35975 37062 Y
GFTPQDVVALSGAHTLGVAR 20/20 (100%)
NDFDNAYFNALQR 12/13 (92%)
LFSDQTLFNTP 11/11 (100%)
48 Peroxidase 21 (ATP2a) Arabidopsis thaliana At2g37130 DASLLLETAR 10/10 (100%) 8.00E-09 36741 36691 Y
LLSETNPLTGDQGEIR 16/16 (100%)
49 Beta-1,3-glucanase (BG3) Arabidopsis thaliana At3g57240 LLLDVPNPDLQR 11/12 (92%) 3.00E-1B 30712 36691 N
YISVGNEVQPS 11/11 (100%)
FVLPAMQNIDR 10/11 (91%)
TYVNNLIQTVK 11/11 (100%)
Spot No Similarity to Organism Acc. No. Sequence identity E value expected Mr observed Mr aS
50 Beta-1,3-glucanase (BG3) Arabidopsis thaliana At3g57240 LLLDVPNPDLQR 11/12(92%) 3.00E-18 30712 37250 N
YISVGNEVQPS 11/11 (100%)
FVLPAMQNIDR 10/11 (91%)
TYVNNLIQTVK 11/11 (100%)
51 Polygalacturonase inhibitor protein 2 Brassica napus AAM94869 NFLQFDLSR 9/9 (100%) 5.00E-05 37117 40280 Y
LQEFDTYSYFHNK 13/13 (100%)
52 Peroxidase 59 precursor Arabidopsis thaliana At5g19890 PYLVQIVR 7/8 (88%) 2.00E-06 35023 41850 Y
GFEVVDTIK 8/9 (89%)
LVEAYSQSQSLFFR 13/14 (93%)
53 Peroxidase 59 precursor Arabidopsis thaliana At5g19890 PYLVQIVR 7/8 (88%) 5.00E-04 35023 43793 Y
LVEAYSQSQSLFFR 13/14 (93%)
54 Peroxidase 59 precursor Arabidopsis thaliana At5g19890 PYLVQIVR 7/8 (88%) 2.00E-06 35023 43793 Y
GFEVVDTIK 8/9 (89%)
LVEAYSQSQSLFFR 13/14 (93%)
55 Subtilase family protein Arabidopsis thaliana At1g20160 AVASAYGTFPTTVIDSK 16/17 (94%) 1.00E-17 81480 56,025* Y
SILKPDITAPGVAILAAWTG 19/20 (95%)
LTYQVTVSATA 10/11 (91%)
56 Subtilase family protein Arabidopsis thaliana At1g20160 SPVYPLIHGK 10/10 (100%) 2.00E-05 81480 56,025* Y
ASAYGTFPTTVIDSK 14/15 (93%)
VSVETPAYQIQVTPEK 13/16 (63%)
VFGALT 6/6 (100%)
57 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78850 KVPVDEQFLVVNK 10/13 (77%) 8.00E-16 49051 50573 Y
TDYSPIEYNPDVR 13/13 (100%)
ILENGNFVIYDSSVK 13/15 (87%)
SVNTDGPYSLVFEDK 12/15 (80%)
TPKPIAVREGYEFFSK 12/16 (75%)
58 Polygalacturonase/ glycoside hydrolase 28 Arabidopsis thaliana At3g61490 DATLLAAQDLEEYPVLK 14/17 (82%) 1.00E-10 51939 56471 Y
DEYGIAFGFPTK 10/12 (83%)
IALGSEFSGGIEDVR 14/15 (93%)
59 Subtilase family protein Arabidopsis thaliana At1g20160 SPVYPLIHGK 10/10 (100%) 1.00E-10 81480 58,848* Y
ASAYGTFPTTVIDSK 14/15 (93%)
SLHPTWCPSAIR 10/12 (83%)
VSVETPAYQIQVTPEK 13/16(81%)
LTYQVTVSAADDDV 10/14(71%)
VFGALT 6/6 (100%)
60 Subtilase family protein Arabidopsis thaliana At1g20160 SPVYPLIHGK 10/10 (100%) 1.00E-10 81480 56,025* Y
ASAYGTFPTTVIDSK 14/15 (93%)
SLHPTWCPSAIR 10/12 (83%)
VSVETPAYQIQVTPEK 13/16(81%)
LTYQVTVSAADDDV 10/14(71%)
VFGALT 6/6 (100%)
61 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78850 KVPVDEQFLVVNK 10/13 (77%) 6.00E-11 49051 50573 Y
TDYSPIEYNPDVR 13/13 (100%)
ILENGNFVIYDSSVK 13/15 (87%)
SVNTDGPYSLVFEDK 12/15 (80%)
TPKPIAVREGYEFFSK 12/16 (75%)
62 Curculin-like (mannose-binding) lectin Arabidopsis thaliana At1g78850 KVPVDEQFLVVNK 10/13 (77%) 6.00E-11 49051 51999 Y
TDYSPIEYNPDVR 13/13 (100%)
ILENGNFVIYDSSVK 13/15 (87%)
SVNTDGPYSLVFEDK 12/15 (80%)
TPKPIAVREGYEFFSK 12/16 (75%)
63 Subtilase family protein Arabidopsis thaliana At1g20160 AVASAYGTFPTTVIDSK 16/17 (94%) 3.00E-07 81480 56,926* Y
TVTNVGGDRAVYK 10/13 (77%)
VFGALT 6/6 (100%)
64 Polygalacturonase/ glycoside hydrolase 28 Arabidopsis thaliana At3g61490 DEYGIAFGFPTK 10/12 (83%) 9.00E-08 51939 5786B Y
IALGSEFSGGIEDVR 14/15 (93%)
WTLLAAQDLEEYPVLK 14/16 (88%)
65 Subtilase family protein Arabidopsis thaliana At1g20160 ESDVVLGGSK 9/10 (90%) 5.00E-08 81480 81188 Y
SPVYPLIHGK 10/10 (100%)
AVASAYGTFPTTVIDSK 16/17 (94%)
SLHPTWCPSAIR 10/12 (83%)
TVTNVGGDRAVYK 10/13 (77%)
LTYQVTVSA 8/9 (89%)
66 Subtilase family protein Arabidopsis thaliana At1g20160 ESDVVLGGSK 9/10 (90%) 5.00E-08 81480 79849 Y
SPVYPLIHGK 10/10 (100%)
AVASAYGTFPTTVIDSK 16/17 (94%)
SLHPTWCPSAIR 10/12 (83%)
TVTNVGGDRAVYK 10/13 (77%)
LTYQVTVSA 8/9 (89%)
67 Subtilase family protein Arabidopsis thaliana At1g20160 ESDVVLGGSK 9/10 (90%) 6.00E-07 81480 81188 Y
AVASAYGTFPTTVIDSK 16/17 (94%)
TVTNVGGDRAVYK 10/13 (77%)
LTYQVTVSA 8/9 (89%)
VFGALT 6/6 (100%)
68 Cucumisin-like serine protease (ARA12) Arabidopsis thaliana At5g67360 VEGASLLGFASGTAR 14/15 (93%) 1.00E-06 79415 78587 Y
DGVAIGAFAAFER 12/13 (92%)
PALAILGNGK 10/10 (100%)
LNYPSFAVNVDGSGAYK 16/17 (94%)
TVTFTVDSSK 10/10 (100%)
69 Germin-like protein Arabidopsis thaliana At3g05950 FTSGLNIAGNTIN 13/13 (100%) 7.00E-05 24736 76,260** Y
AFQLDVNVVR 9/10 (90%)
==== Refs
Oda A Sakuta C Masuda S Mizoguchi T Kamada H Satoh S Possible involvement of leaf gibberellins in the clock-controlled expression of XSP30, a gene encoding a xylem sap lectin, in cucumber roots Plant Physiology 2003 133 1779 1790 14605217 10.1104/pp.103.030742
Schurr U Schulze ED The concentration of xylem sap constituents in root exudate, and in sap from intact, transpiring castor bean plants (Ricinus communis L.). Plant, Cell and Environment 1995 18 409 420
Zornoza P Gonzalez M Serrano S Carpena O Intervarietal differences in xylem exudate composition and growth under contrasting forms of N supply in cucumber. Plant and Soil 1996 178 311 317 10.1007/BF00011597
Satoh S Kuroha T Wakahoi T Inouye Y Inhibition of the formation of adventitious roots on cucumber hypocotyls by the fractions and methoxybenzylglutamine from xylem sap of squash root. Journal of Plant Research 1998 111 541 546
Sattelmacher B The apoplast and its significance for plant mineral nutrition New Phytol 2001 149 167 192 10.1046/j.1469-8137.2001.00034.x
Hartung W Sauter A Hose E Abscisic acid in the xylem: where does it come from, where does it go to? Journal of Experimental Botany 2002 53 27 32 11741037 10.1093/jexbot/53.366.27
Olszewski N Sun TP Gubler F Gibberellin signalling: biosynthesis, catabolism, and response pathways Plant Cell 2002 14 Suppl S61 80 12045270
Mok DW Mok MC Cytokinin metabolism and action Annu Rev Plant Physiol Plant Mol Biol 2001 52 89 118 11337393 10.1146/annurev.arplant.52.1.89
Finkelstein RR Gibson SI ABA and sugar interactions regulating development: cross-talk or voices in a crowd? Curr Opin Plant Biol 2002 5 26 32 11788304 10.1016/S1369-5266(01)00225-4
Biles CL Martyn RD Wilson HD Isozymes and general proteins from various watermelon cultivars and tissue types HortScience 1989 24 810 812
Biles CL Abeles FB Xylem sap proteins Plant Physiology 1991 96 597 601
Masuda S Sakuta C Satoh S cDNA cloning of a novel lectin-like xylem sap protein and its root-specific expression in cucumber Plant and Cell Physiology 1999 40 1177 1181 10635119
Satoh S Iizuka C Kikuchi A Nakamura N Fujii T Proteins and carbohydrates in xylem sap from squash root Plant and Cell Physiology 1992 33 841 847
Young SA Guo A Guikema JA White FF Leach JE Rice cationic peroxidase accumulates in xylem vessels during incompatible interactions with Xanthomonas oryzae pv oryzae Plant Physiology 1995 107 1333 1341 7770527 10.1104/pp.107.4.1333
Rep M Dekker HL Vossen JH de Boer AD Houterman PM Speijer D Back JW de Koster CG Cornelissen BJC Mass spectrometric identification of isoforms of PR proteins in xylem sap of fungus-infected tomato Plant Physiology 2002 130 904 917 12376655 10.1104/pp.007427
Buhtz A Kolasa A Arlt K Walz C Kehr J Xylem sap protein composition is conserved among different plant species Planta 2004 219 610 618 15064951 10.1007/s00425-004-1259-9
Masuda S Kamada H Satoh S Chitinase in cucumber xylem sap Bioscience Biotechnology and Biochemistry 2001 65 1883 1885 10.1271/bbb.65.1883
Sakuta C Oda C Yamakawa S Satoh S Root-specific expression of genes for novel glycine-rich proteins cloned by use of an antiserum against xylem sap proteins of cucumber Plant and Cell Physiology 1998 39 1330 1336 10050317
Rep M Dekker HL Vossen JH de Boer AD Houterman PM de Koster CG Cornelissen BJC A tomato xylem sap protein represents a new family of small cysteine-rich proteins with structural similarity to lipid transfer proteins FEBS Letters 2003 534 82 86 12527365 10.1016/S0014-5793(02)03788-2
Sakuta C Satoh S Vascular tissue-specific gene expression of xylem sap glycine-rich proteins in root and their localization in the walls of metaxylem vessels in cucumber Plant and Cell Physiology 2000 41 627 638 10929946
Nielsen H Engelbrecht J Brunak S von Heijne G Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites Protein Engineering 1997 10 1 6 9051728 10.1093/protein/10.1.1
Bendtsen JD Nielsen H Heijne G Brunak S Improved prediction of signal peptides: SignalP 3.0 Journal of Molecular Biology 2004 340 783 795 15223320 10.1016/j.jmb.2004.05.028
Cavell AC Lydiate DJ Parkin IAP Dean C Trick M Colinearity between a 30-centimorgan segment of Arabidopsis thaliana chromosome 4 and duplicated regions within the Brassica napus genome Genome 1998 41 62 69 9549059 10.1139/gen-41-1-62
Murray MT Schiller DL Franke WW Sequence analysis of cytoplasmic mRNA-binding proteins of Xenopus oocytes identifies a family of RNA-binding proteins Proc Natl Acad Sci USA 1992 89 11 15 1729676
Ohgiya S Ishizaki K Shinriki N Molecular cloning of guinea pig CYP1A1: complete primary structure and fast mobility of expressed protein on electrophoresis Biochim Biophys Acta 1993 1216 237 244 8241264
Berger D Altmann T A subtilisin-like serine protease involved in the regulation of stomatal density and distribution in Arabidopsis thaliana Genes Dev 2000 14 1119 1131 10809670
Segarra CI Casalongue CA Pinedo ML Ronchi VP Conde RD A germin-like protein of wheat leaf apoplast inhibits serine proteases J Exp Bot 2003 54 1335 1341 12709479 10.1093/jxb/erg139
Armstrong DJ Roman A The anomalous electrophoretic behavior of the human papillomavirus type 16 E7 protein is due to the high content of acidic amino acid residues Biochem Biophys Res Commun 1993 192 1380 1387 8389548 10.1006/bbrc.1993.1569
Billings PC Orf JW Palmer DK Talmage DA Pan CG Blumenfeld M Anomalous electrophoretic mobility of Drosophila phosphorylated H1 histone: is it related to the compaction of satellite DNA into heterochromatin? Nucleic Acids Res 1979 6 2151 2164 461184
Hu CC Ghabrial SA The conserved, hydrophilic and arginine-rich N-terminal domain of cucumovirus coat proteins contributes to their anomalous electrophoretic mobilities in sodium dodecylsulfate-polyacrylamide gels J Virol Methods 1995 55 367 379 8609202 10.1016/0166-0934(95)00085-1
Mann M Jensen ON Proteomic analysis of post-translational modifications Nat Biotech 2003 21 255 261 10.1038/nbt0303-255
Hiraga S Sasaki K Ito H Ohashi Y Matsui H A large family of class III plant peroxidases Plant and Cell Physiology 2001 42 462 468 11382811 10.1093/pcp/pce061
Bradley DJ Kjellbom P Lamb CJ Elicitor- and wound-induced oxidative cross-linking of a proline-rich plant cell wall protein: a novel, rapid defense response. Cell 1992 70 21 30 1623521 10.1016/0092-8674(92)90530-P
Passardi F Cosio C Penel C Dunand C Peroxidases have more functions than a Swiss army knife Plant Cell Reports 2005 DOI: 10.1007/s00299-005-0972-6
van Loon LC van Strien EA The families of pathogenesis-related proteins, their activities, and comparative analysis of PR-1 type proteins Physiological & Molecular Plant Pathology 1999 55 85 97 10.1006/pmpp.1999.0213
Beers EP Jones AM Dickerman AW The S8 serine, C1A cysteine and A1 aspartic protease families in Arabidopsis Phytochemistry 2004 65 43 58 14697270 10.1016/j.phytochem.2003.09.005
Ribeiro A Akkermans ADL van Kammen A Bisseling T Pawlowski K A nodule-specific gene encoding a subtilisin-like protease is expressed in early stages of actinorhizal nodule development The Plant Cell 1995 7 785 794 7647567 10.1105/tpc.7.6.785
von Groll U Berger D Altmann T The subtilisin-like serine protease SDD1 mediates cell-to-cell signaling during Arabidopsis stomatal development The Plant Cell 2002 14 1527 1539 12119372 10.1105/tpc.001016
Funk V Kositsup B Zhao C Beers EP The Arabidopsis xylem peptidase XCP1 is a tracheary element vacuolar protein that may be a papain ortholog. Plant Physiology 2002 128 84 94 11788755 10.1104/pp.128.1.84
Kitajima S Sato F Plant pathogenesis-related proteins: molecular mechanisms of gene expression and protein function Journal of Biochemistry 1999 125 1 8 9880788
Mauch F Staehelin LA Functional implications of the subcellular localization of ethylene-induced chitinase and beta 1,3-glucanase in bean leaves The Plant Cell 1989 1 447 457 12359894 10.1105/tpc.1.4.447
Trudel J Grenier J Potvin C Asselin A Several thaumatin-like proteins bind to beta-1,3-glucans Plant Physiology 1998 118 1431 1438 9847118 10.1104/pp.118.4.1431
Chrispeels MJ Raikhel NV Lectins, lectin genes, and their role in plant defense The Plant Cell 1991 3 1 9 1824332 10.1105/tpc.3.1.1
Sharon N Lis H History of lectins: from hemagglutinins to biological recognition molecules Glycobiology 2004 14 53R 62R 15229195 10.1093/glycob/cwh122
Vijayan M Chandra N Lectins. Current Opinion in Structural Biology 1999 9 707 714 10607664 10.1016/S0959-440X(99)00034-2
Barre A Van Damme EJM Peumans WJ Rougé P Curculin, a sweet-tasting and taste-modifying protein, is a non-functional mannose-binding lectin Plant Molecular Biology 1997 33 691 698 9132060 10.1023/A:1005704616565
Ringli C Keller B Ryser U Glycine-rich proteins as structural components of plant cell walls Cellular and Molecular Life Sciences 2001 58 1430 1441 11693524
Ryser U Schorderet M Zhao GF Studer D Ruel K Hauf G Keller B Structural cell-wall proteins in protoxylem development: evidence for a repair process mediated by a glycine-rich protein The Plant Journal 1997 12 97 111 9263454 10.1046/j.1365-313X.1997.12010097.x
Rose JKC Braam J Fry SC Nishitani K The XTH family of enzymes involved in xyloglucan endotransglucosylation and endohydrolysis: current perspectives and a new unifying nomenclature Plant Cell Physiol 2002 43 1421 1435 12514239 10.1093/pcp/pcf171
De Lorenzo G D' Ovidio R Cervone F The role of polygalacturonase-inhibiting proteins (PGIPs) in defense against pathogenic fungi Annual Review of Phytopathology 2001 39 313 335 11701868 10.1146/annurev.phyto.39.1.313
Li R Rimmer R Yu M Sharpe AG Seguin-Swartz G Lydiate D Hegedus DD Two Brassica napus polygalacturonase inhibitory protein genes are expressed at different levels in response to biotic and abiotic stresses. Planta 2003 217 299 308 12783338
Cervone F de Lorenzo G Pressey R Darvill AG Albersheim P Can Phaseolus PGIP inhibit pectic enzymes from microbes and plants? Phytochemistry 1990 29 447 449 10.1016/0031-9422(90)85094-V
Neumann U Brandizzi F Hawes C Protein transport in plant cells: in and out of the golgi Annals of Botany 2003 92 167 180 12876187 10.1093/aob/mcg134
Giavalisco P Nordhoff E Lehrach H Gobom J Klose J Extraction of proteins from plant tissues for two-dimensional electrophoresis analysis Electrophoresis 2003 24 207 216 12652593 10.1002/elps.200390016
Giavalisco P Nordhoff E Kreitler T Klöppel KD Lehrach H Klose J Gobom J Proteome analysis of Arabidopsis thaliana by two-dimensional gel electrophoresis and matrix-assisted laser desorption/ionisation-time of flight mass spectrometry Proteomics 2005 5 1902 1913 15815986 10.1002/pmic.200401062
Walz C Juenger M Schad M Kehr J Evidence for the presence and activity of a complete antioxidant defence system in mature sieve tubes The Plant Journal 2002 31 189 197 12121448 10.1046/j.1365-313X.2002.01348.x
|
15969751
|
PMC1182380
|
CC BY
|
2021-01-04 16:03:52
|
no
|
BMC Plant Biol. 2005 Jun 21; 5:11
|
utf-8
|
BMC Plant Biol
| 2,005 |
10.1186/1471-2229-5-11
|
oa_comm
|
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-121598517610.1186/1471-2229-5-12Research ArticleExpressed sequence tags (ESTs) and simple sequence repeat (SSR) markers from octoploid strawberry (Fragaria × ananassa) Folta Kevin M [email protected] Margaret [email protected] Philip J [email protected] Sook [email protected] Dawn H [email protected] Christopher [email protected] Dorrie [email protected] Plant Molecular and Cellular Biology Program and Horticultural Sciences Department, University of Florida, Gainesville, FL, USA2 Genetics, Biochemistry & Life Science Studies, Clemson University, Clemson, SC, USA2005 28 6 2005 5 12 12 6 4 2005 28 6 2005 Copyright © 2005 Folta 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
Cultivated strawberry (Fragaria × ananassa) represents one of the most valued fruit crops in the United States. Despite its economic importance, the octoploid genome presents a formidable barrier to efficient study of genome structure and molecular mechanisms that underlie agriculturally-relevant traits. Many potentially fruitful research avenues, especially large-scale gene expression surveys and development of molecular genetic markers have been limited by a lack of sequence information in public databases. As a first step to remedy this discrepancy a cDNA library has been developed from salicylate-treated, whole-plant tissues and over 1800 expressed sequence tags (EST's) have been sequenced and analyzed.
Results
A putative unigene set of 1304 sequences – 133 contigs and 1171 singlets – has been developed, and the transcripts have been functionally annotated. Homology searches indicate that 89.5% of sequences share significant similarity to known/putative proteins or Rosaceae ESTs. The ESTs have been functionally characterized and genes relevant to specific physiological processes of economic importance have been identified. A set of tools useful for SSR development and mapping is presented.
Conclusion
Sequences derived from this effort may be used to speed gene discovery efforts in Fragaria and the Rosaceae in general and also open avenues of comparative mapping. This report represents a first step in expanding molecular-genetic analyses in strawberry and demonstrates how computational tools can be used to optimally mine a large body of useful information from a relatively small data set.
==== Body
Background
Commercial strawberry has a value of 1.4 billion dollars in the United States, and represents a significant regional crop throughout the world. Despite its value, fewer than 100 annotated sequences existed in public databases in early 2004. The information discrepancy is a consequence of limited molecular study in the challenging octoploid cultivated varieties. The thin public informatics base hence represents a barrier to meaningful study of functional genomics, genetic mechanisms, as well as the molecular-systematic relationships between the octoploid strawberry, the Rosaceae and other species. The lack of basic sequence information hinders the development of transgenic technologies that would advance molecular-physiological studies and potentially benefit the grower and consumer. Overall, the dearth of sequence information has limited agile molecular resolution studies in this important crop plant.
To remedy this discrepancy ~1800 expressed sequence tags (ESTs) were sequenced from a whole-plant cDNA library derived from various tissues of the Strawberry Festival cultivar. This cultivar was chosen because of its east-coast and west-coast lineage as well as its range of favorable horticultural attributes. 'Strawberry Festival' produces large, uniform, firm fruit, and is resistant to Botrytis cinera, the causative agent behind gray mold [1]. It is a predominant cultivar grown in Florida, and has been well studied in many reports of fungicide use, disease resistance and post-harvest fruit quality. The study of an important commercial variety will provide tools to directly aid breeding and probe genetic mechanisms in these cultivars.
Strawberry has unrealized potential as a research model and tool, and the lack of molecular markers for breeding and the eventual need for genetic improvement of the current suite of cultivars makes sequence examination especially timely. Information gained from the octoploid will also translate to defining molecular markers to facilitate mapping in both the diploid species (eg. Fragaria vesca and Fragaria nubicola) as well as octoploid cultivars. A strong sequence database is the cornerstone of functional genomics studies, and this information will aid development of such tools in Fragaria and in the Rosaceae in general. Definition of expressed gene sequence variation in the octoploid may aid in the understanding of polyploid evolution and/or silencing of component genomes. Sequence information constitutes a basis for eventual reverse-genetic and activation-tag studies. Both the diploid and octoploid species are excellent candidates for such studies as they are efficiently transformed and regenerated [2-4], possess a diploid genome that is slightly larger than that of Arabidopsis thaliana [5], and can be rapidly propagated from seed (3–5 months) or runners [6]. Strawberry also may be an excellent candidate as a bioreactor, a system to manufacture specific compounds of interest. A presentation of the elements of the strawberry transcriptome facilitates the initiation of such studies.
Despite strawberry's crop value and potential as a research tool, a formal analysis of EST data has not been reported. In this report we identify over 1300 unique transcripts assembled from 1,847 ESTs derived from whole-plant vegetative tissues 24 h after salicylic acid treatment. The cDNA library was prepared from total RNA pooled from roots, petioles, stolons, leaves and meristems to generate a diverse set of transcripts with limited redundancy. Multiple analyses, such as developing a unigene set, annotation with putative function and identification of SSRs, opens additional paths that will speed research into strawberry physiology, evolution, genetics and genomics. This represents the first major EST report from Fragaria and can now serve as a baseline for these further studies.
Results
The Fragaria × ananassa EST library
The Lambda ZAP cDNA library was generated from whole-plant tissues from mature plants 24 after salicylic acid treatment. The details of salicylate treatment, plant materials and library construction are presented in Methods.
EST processing and assembly
A total of 1847 of ESTs were sequenced, resulting in 1505 high-quality trimmed sequences which were submitted to GenBank on August 6, 2004. Representing a success rate of 81.5%, these sequences have an average length of 613 bp and a PHRED quality value of 35. Assembly of the sequences into a unigene was performed in order to reduce redundancy of the sequences and identify those coding for the same protein (Methods, Assembly). The total unigene consists of 1171 singlets for a total of 1304 unigenes.
Contigs were assembled from EST sequences. The final unigene has 133 contigs, 120 comprised of two or three merged ESTs. Eight contigs were assembled from four individual ESTs. Contigs assembled from five or more ESTs may be useful to deconstruct in the interest of studying allelic diversity in the octoploid. In diploid species alleles represent heterozygousity at a given locus as well as gene duplication and subfunctionalization of a given coding region. Allelic diversity is potentially enriched in the octoploid, since the octoploid maintains the alleles maintained from at least three donor diploid genomes. Expression of specific alleles may be informative, as patterns may be traced back to the diploid genome contributors, allowing description of expression from within, or between, donor genomes.
For instance, Contig 23 represents psaL, a nuclear-encoded subunit of the photosystem 1 reaction center. The contig was assembled from five ESTs, two of which (4C07 and 6C09) are identical in sequence yet vary in length. The other members contain SNPs, especially 18C04, which maintains five unique base changes over a 540 bp alignment of all five ESTs. Others contain a single alteration in this relatively conserved gene sequence. Similar results were observed for contigs 32 and 99, which were assembled from seven and nine ESTs, respectively.
Other contigs have been assembled from many ESTs, such as Contigs 29 and 12. These contigs encode light-harvesting, chlorophyll-binding (Lhcb, formerly cab) proteins and a non-specific lipid transfer protein, respectively. The ESTs corresponding to these genes arise from small multigene families within a diploid genome in most species, making these constructs less useful for studying between-genome polymorphisms.
Functional annotation
Computational tools are now regularly used to infer function based upon significant sequence similarity toexperimentally verifiedproteins or putative proteins. These analyses implement FASTA and BLAST comparisons against non-redundant databases as well as GO annotation. The EST sequences were compared against known databases using these tools.
Protein homology searches were performed in order to identify the putative function of the ESTs (Methods, Functional Characterization). NCBI's non-redundant (nr) protein and Rosaceae EST databases searches were run on February 27, 2005 using the FASTX3.4 algorithm [7]. The nr database contained 2321663 protein amino acid sequences at the time of the search. Of the 1304 unigenes, 1105, or 84.74% of the set, had significant matches to this database (Table 1). A comparison against SWISS-PROT was performed on July 26, 2004, yielding a lower number of significant matches. SWISS-PROT is a curated, highly-annotated, smaller database of 153,871 proteins of demonstrated function. 714 of the unigenes (54.75%) had significant matches (Table 1). Only 191 of the unigenes did not match a protein in either of these two protein databases (Table 2). Upon close scrutiny the EST sequences did not contain significant open reading frames, suggesting that the EST sequence represents long untranslated regions, structural RNAs, or bona fide proteins, unique to Fragaria based on current comparisons.
Comparisons to Rosaceae ESTs
Table 1 also presents the results of comparison of the unigene against publicly-available Rosaceae ESTs in order to assess how Fragaria relates to the rest of the Rosaceae family at the gene sequence and content levels. The BLASTN algorithm was then used for EST homology searches against known Rosaceae ESTs. 227,250 Rosaceae ESTs were downloaded from dbEST. Of the 1304 unigenes, 835 (64.03%) had significant homology to other Rosaceae ESTs. Since this dataset is composed of public ESTs, it contains a large amount of redundancy. The majority of public ESTs have been sequenced from the 5' end, so ESTs generated from the 3' end may be less likely to find homologs in a search against public ESTs. Still, of the 191 ESTs that did not show significant homology with SWISS-PROT and Genbank nr (Table 2), 54 ESTs had homologs represented in the Rosaceae EST set. This leaves 137 transcripts that show no significant homology outside of Fragaria within the Rosaceae family. These ESTs were compared against the TIGR plant repeat databases to test if they may have originated from retroelement expression. None of the apparently Fragaria-specific transcripts exhibited significant homology with sequences within the repeat database.
Characterization by gene ontology
The Fragaria unigenes were further annotated by gene ontology (GO) assignment based on the single "best hit" match against the SWISS-PROT database. All 714 ESTs with hits to SWISS-PROT have matching GO-Terms (Figure 1). The three categories are function (Panel A), process (Panel B), and component (Panel C). For molecular function, the strawberry ESTs were assigned to eight categories. The majority (51%) of the ESTs were assigned to "Catalytic Activity" (GO:0003824). For biological process, the ESTs were assigned to four categories with the majority (77%) representing genes participating in metabolism (GO: 0008152). When grouped according to likely cellular component, the ESTs were assigned to six categories and 93% were covered by two GO terms: "Intracellular" (GO:0005622) and "Membrane" (GO:0016020). The full chart of the assignment of EST's to specific GO Term categories may be viewed on the GDR website [8].
Homology to mapped peach ESTs
Linkage relationships have been identified for many peach ESTs and have facilitated placement on the peach genetic map. Comparison of the Fragaria unigene to this set of ESTs presents a basis for developing linkage relationships between the established peach, and growing Fragaria, linkage maps. A series of peach ESTs have been conclusively anchored to genetic maps by sharing BACs with genetic markers previously used for BAC hybridization [9]. Of the 295 mapped peach ESTs 22 (7.04%) showed a significant match with the strawberry unigene (Table 3).
Computational analysis for SSRs and ORFs in the ESTs
Simple Sequence Repeats (SSRs) were identified in the strawberry unigene data set (Open Reading Frame and Microsatellite Analysis, Methods). In this study, SSRs are defined as dimers with at least 5 repeats, trimers with at least 4 repeats, tetramers with at least 3 repeats, and pentamers with at least 3 repeats. 190 unigene sequences (14%) were found to have one or more repeat, and 79 different motifs were identified within the set of clones. A total of 269 SSRs were found with trimers being the most common motif length (Table 4). The frequency of motifs for all the possible dimers and trimers is listed in Table 5. To examine the distribution of SSRs in the putative coding region and the UTR, we detected open reading frames in the unigenes using the FLIP program (Brossard 1997). When the longest open reading frame was selected as the putative coding region, 176 (65.4%) of these microsatellites were found inside putative coding regions. When filtered for the most optimal primer candidates (40–60% GC content) a total of 208 SSR-flanking sequences met the criteria (Table 6). These optimal candidates can be downloaded via the GDR ftp site [10].
Discussion
Fragaria × ananassa is complex polyploid, arising from a spontaneous cross between Fragaria virginiana and Fragaria chiloensis. The genome contains contributions from at least three diploid species [11,12]. Over the past century cultivation of octoploid strawberry has progressed solely on the careful efforts of breeders, physiologists and biochemists. This complex genome and coincident awkward genetics has slowed the development of molecular markers and other tools that would benefit breeding efforts and understanding of strawberry genomics. This report details a starting point to advance the traditional strawberry research avenues using modern molecular tools to forward structural-and functional-genomics studies in this important crop species. As important, it demonstrates that computational tools may be used to comprehensively mine large quantities of important data from a relatively small data set. As these tools become available as web-based applications, small sequencing efforts may extract valuable information that may shape research questions in under-represented crops like strawberry.
Recent efforts demonstrate the importance of sequence information as the basis of functional-genomics studies. Previous reports of gene expression in strawberry have been dependent on the discovery and characterization of specific genes of interest, such as an O-methyltransferase associated with flavor [13], enzymes that influence fruit firmness [14-17], as well as several others [18,19]. Current technologies have the capacity to assess genome-wide transcriptome changes associated with a given treatment or developmental process [20,21]. Recent studies in cultivated strawberry have implemented proprietary sequence information in a microarray format to unveil the transcriptome that coincides with fruit ripening [22-24]. These studies have identified critical regulators of fruit flavor. The first study identified strawberry alcohol acetyl transferase as a critical enzyme in the production of volatiles esters. The associated transcript increased during fruit ripening and the recombinant protein catalyzes the appropriate syntheses from a variety of substrates in E. coli [23]. A recent functional-genomics study characterized Nerolidol Synthase 1, the enzyme that catalyzes the formation of the flavor compounds linalool and nerolidol from geranyl diphosphase and farnesyl diphosphate, respectively. The enzyme is expressed in the receptacle of ripening fruit, not in leaves, and is highly expressed in cultivated species relative to wild species. The report concludes that selection of cultivated varieties for fruit flavor fixed mechanisms to express and localize terpene-associated enzymes that favorably affected flavor, while repressing those that make fruits less desirable [22]. Although the factors leading to fruit flavor in strawberry have been studied for decades, a transcriptome survey produced the most definitive results, owing again to the usefulness of a sequence database in Fragaria.
The transcripts characterized from this project will allow development of genomics resources for the study of other important physiological responses. A subset of these ESTs is shown in Table 7. These ESTs are relevant to processes of interest to the strawberry industry and may represent important molecular tools to researchers. The first set represents a series of ESTs with sequence homology to genes associated with the photoperiodic control of flowering. These include close homologs to CONSTANS (CO), a likely transcription factor that induces specific meristem identity genes under the appropriate photoperiod [25,26]. A homolog of a critical regulator of meristem identity AGL20/SUPPRESSOR OF CO OVEREXPRESSION was also identified. This gene encodes a MADS-box transcription factor that likely functions downstream of CO in conferring light signals to the promoters of meristem identity genes [27]. An EST representing VERNALIZATION INSENSITIVE 3 also was identified in this library. VIN3 is a protein shown to function downstream of CO in regulating seasonal flowering responses [28]. VIN3 is a chromatin-remodeling protein that represses FLC, a protein that negatively-regulates CO function [29] allowing the plant to appropriately time flowering relative to seasonal chilling.
Analysis of this dataset revealed a suite of likely homologs to pathogenesis-related (PR) genes, such as thionins, Ndr1, β 1-3-glucanase and chitinases, and LRR proteins. The prevalence of this family of proteins was not surprising as the plants were treated with salicylic acid 24 h before RNA harvest to enrich for PR genes in the library. These genes are of particular interest to plant scientists because of their potential to help define the mechanism(s) of disease resistance and susceptibility. It is possible that these genes may be especially useful targets for antisense or overexpression in unveiling these agriculturally-important traits, or possibly in the design of transgenic plants with heightened resistance to common plant pathogens. All of these facets are important, as strawberry cultivation requires copious application of fungicides and/or bacteriostatic compounds to ensure proper fruit set.
Of interest to this laboratory are homologs of genes associated with photomorphogenesis, such as Hy5 and Non-phototropic hypocotyl 3. These both play roles in early light development, yet HY5 also has been shown to influence downstream developmental processes such as fruit ripening and pigmentation [30] and also binds to the promoters of genes associated with circadian clock progression [31].
The information distilled from all of these analyses can now be used to design strawberry-specific probes to assess gene expression patterns and develop transgenics to directly test gene function. These important studies are underway and will facilitate comparisons between the biological sensory/response mechanisms in strawberry to those of model systems.
The apparent sequence conservation between Fragaria and other rosaceous tree crops suggests that cross-species microarray studies may be productive within the Rosaceae. This study demonstrates that less than 11% of the ESTs are unique to strawberry. This value is likely inflated, as ESTs by nature contain variable untranslated regions and other features that may preclude efficient identification of homologs. Of the 1305 ESTs, 835 have strong homology with other Rosaceae ESTs. Those featuring over 85% homology over 100 bases are between 86 and 100% identical to transcripts isolated from other Rosaceae, with an average identity of 91% (+/-0.001%). The high degree of similarity may be a useful platform for comparisons between molecular-mechanistic differences exhibited between diverse species with little sequence variation. Here, the diversity within the Rosaceae is likely due to variation in gene expression, and EST data and microarray technologies are an idea platform to study these patterns.
The relatively extensive genetic mapping in Prunus has delineated linkage associations between genes of the genus, those in select Rosaceae species and even Arabidopsis [32]. Physical maps have also been developed from transcript mapping [9]. The ESTs from this Fragaria collection were compared to the mapped peach genes, and 23 agreed with strong homology (Table 3). These relationships are important as they present the basis to study structural relationships between cousin species within the Rosaceae. Since these loci are mapped in peach, they represent excellent loci to also add to the growing diploid strawberry linkage map [33], and eventually map in the octoploid.
Mapping efforts may also be hastened from identification of SSRs. SSRs derived from ESTs provide a basis to assign linkage relationships to known gene products, and such studies have been initiated in diploid strawberry [33]. In the EST collection presented herein, a number of SSRs are present in transcripts correlating to putative allergens, regulators of the circadian clock, and general housekeeping genes. These transcripts can now be readily mapped in the diploid using existing populations, and such studies are currently underway. Furthermore, specific genes of interest can be studied for variation within diploid species or for intron-specific polymorphisms that will allow their assignment to the diploid strawberry linkage map. These studies will ultimately facilitate the generation of molecular markers to follow traits/genes of interest in the commercial cultivars, adding the resolution of molecular tools to complement conventional breeding strategies.
The general proportions of the different functional groups (Figure 1) reflect well the expected state of the mature plant transcriptome as reported in previous studies. Transcripts encoding enzymes associated with the cell cycle, cytoskeleton or cell walls are not abundant as mature plants are less reliant on processes governing greater cell number or cell size. Approximately half of the transcripts associated with photosynthesis are members of the chlorophyll a/b binding protein family; the other half typically contains plastid-encoded transcripts. As expected, the majority of transcripts detected represent enzymes of general metabolism.
Conclusion
Although a small EST set, the complete suite of analyses performed herein demonstrate that a finite transcriptome snapshot may provide ample resources to seed additional study. Here a relatively small number of ESTs has provided sufficient information to engage in further molecular, physiological and genetic studies. For instance, the pretreatment with salicylate likely enriched the expression of pathogenesis-related transcripts that can now be used to study disease progression in specific strawberry cultivars with large variations in sensitivity and resistance. Clearly, the development of a comprehensive SSR catalog allows characterization of these potential genetic markers in the progeny of polymorphic cultivars, in an important crop species virtually devoid of linkage associations. Unlike other markers, EST-derived SSRs by definition originate from a sequence that is expressed, adding functional resolution to linkage groups built on structural polymorphisms. More importantly, the same suite of tools used to perform these analyses will be soon available through a public interface at the GDR, making comparable analyses possible. These applications are an important rationale for sequencing and analysis of a limited EST set, as even a small research program may find sufficient resources to initiate molecular-genetic study of an under-represented crop species.
Methods
Library construction
Roots, leaves, petioles, stolons, meristems and new daughter plants were harvested from several individual chamber-grown strawberry (Fragaria × ananassa cultivar 'Strawberry Festival) plants 24 h after salicylic acid treatment (4 μm foliar spray, 1 μM drench). Tissues were washed briefly to remove soil and then were frozen in liquid nitrogen. Total RNA was extracted using the following method, a modification of protocols used in the extraction of RNA from pine cones [34]. Briefly, 1 g of tissue was ground in liquid nitrogen using a mortar and pestle, then incubated in extraction buffer (2% CTAB, 2% polyvinylpyrrolidone, 100 mM Tris-HCl (pH 8.0), 25 mM EDTA, 2.0 M NaCl, 0.5 g/ml spermidine, and 2.0% β-mercaptoethanol) at 65°C for 10 min. The samples were cooled to room temperature, an equal volume of chloroform:octonol (24:1) was added and the mixture was homogenized using a Polytron (T10-35 homogenizer) at 80–90% maximum speed. The organic and aqueous phases were separated by centrifugation at 5700 × g and the supernatant was vortexed with an equal volume of chloroform:octonol. The phases were again separated by centrifugation and the supernatant was transferred to a clean test tube, LiCl was added to a final concentration of 2.5 M and precipitated on ice overnight. RNA was then collected by centrifugation at 5700 × g. The pellet was resuspended in 500 μl SSTE (1 M NaCl, 0.5% SDS, 10 mM Tris-HCl (pH 8.0), 1 mM EDTA) and extracted with an equal volume of chloroform:octonol. The supernatant was precipitated with two volumes of ethanol, the pellet was washed with 76% ethanol containing 0.3 M sodium acetate, dried briefly in a Speed Vac, and resuspended in 50 μl 10 mM Tris-HCl (pH 8.0) 2.5 mM EDTA before quantitation by spectrophotometry.
For library construction mRNA was isolated from total RNA using the Oligotex Direct mRNA Mini Kit (Qiagen Inc., Valencia, CA) using 500 μg total RNA. The cDNA library was constructed from 5 μg mRNA using the Uni-ZAP XR Cloning Kit (Stratagene Inc, Carlsbad, CA) as per manufacturer's directions. The primary library consisted of 6.2 × 107 colony forming units with average insert size of 800 bp and 98% of clones containing inserts of ≥ 200 bp. Mass excision of filamentous phage was performed and phagemids were cloned to E. coli for sequencing.
Sequencing and sequence processing
A total of 1847 EST clones were sequenced from the 3' end at the University of Florida ICBR Core Facility using ET Terminator (Amersham Inc, Schaumburg, IL). These sequences were processed using publicly available software incorporated in a fully automated in-house script (ProcEST.pl) developed at Clemson University by the Genome Database for Rosaceae (GDR) bioinformatics team. Sequence trace files were converted into FASTA formatted sequence and quality score files using the PHRED [35] base-calling program. Vector and host contamination were identified and masked using the sequence comparison program CROSS_MATCH [36]. Vector trimming excised the longest non vector sequence and further trimming removed low quality bases (less than phred score 20) at both ends of a read. Sequences were discarded if they had greater than 5% ambiguous bases, more than 40 PolyA or Poly T bases or less than 100 high quality bases (minimum phred score of 20). Using this protocol, 81% of the sequences (1505) were considered high quality and submitted to the NCBI public EST repository dbEST [37]. To reduce redundancy and increase transcript length the high quality sequences were assembled using the contig assembly program CAP3 [38]. Various assemblies were performed using different CAP3 parameters to identify the build that required least manual editing. More stringent parameters (- p 90 -d 60) were used to prevent over assembly and help identify potential paralogs. The assembly was refined where possible using homology to the SwissProt database to indicate contig accuracy. Homology was determined by comparing the contigs and clones against the Swiss Prot database using the fastx3.4 algorithm [7] with EXP < 1e -6. Contigs whose clones showed difference in homology were deconstructed and contigs with the same homology to other contigs were joined using default CAP3 parameters. The unigene data set was derived by combining the contig and singleton data sets.
Functional characterization
Functional characterization of the unigene data set consisted of pairwise comparison of both the high quality clones and the contig consensus sequences against the NCBI nr [39] and SWISS-PROT [40] protein databases using the fastx3.4 algorithm [7]. The most significant matches (EXP < 1e -7 and EXP <1e-6 for the NCBI nr SWISS-PROT searches, respectively) for each contig and individual clones in the library were recorded. The Swiss-Prot matches were further classified by gene ontology [41]. Contigs or clones that did not have a significant match with either of these databases were searched against the InterPro protein families and domains database (Mulder et al, 2005) using InterProScan [42].
The unigene sequences were also characterized by comparison with the Genbank Rosaceae EST dataset (227250 as of February 14, 2005) and 256 peach mapped ESTs [43], downloaded from the Genome Database for Rosaceae (GDR). Using the BLASTN algorithm [44], sequences with > 85% similarity over an alignment length of 100 bp were considered significant matches.
Open reading frame and microsatellite analysis
Open reading frames (ORFs) were identified in the ESTs using the software program FLIP [45] and the longest ORF was recorded as the putative coding region. Simple Sequence Repeats (SSRs) were identified in the unigene data set using a modified version (CUGISSR) of a perl script SSRIT [46]. SSRs recorded for the final dataset include dimers with at least 5 repeats, trimers with at least 4 repeats, tetramers with at least 3 repeats, and pentamers with at least 3 repeats. SSR-containing sequences were identified as optimal candidates for primer development if they contained a GC content between 40% and 60% and a minimum of 20 base pairs of sequence on either side of the SSR. Using the FLIP output, CUGISSR reports the location of SSRs in the relation to the putative coding region.
Data storage and web interface
All sequence, assembly, homology, ORF and SSR data were uploaded to the Genome Database for Rosaceae (GDR) (Jung et al, 2004) as well as library, protocol, contact and publication information. GDR scripts were utilized to allow users to browse, query or download all the project data.
Public access and dissemination
The GDR website has a number of different EST project sections including the Fragaria EST dataset detailed here. These web pages are extensively linked such that users can easily access data of interest regardless of the navigation entry point. To access the project pages for this EST project, users can go to the project page which can be found by a link in the "About Us" drop down menu in the top navigation bar. This project is listed on the "Data Overview" page as "Folta – University of Florida" [47]. The sidebar for this project allows the user to view the project description, the library details, the processing protocol, a report on the successful clones, unigene details, gene homology pages, microsatellite analysis, contact information, and publication information. The cDNA phage library and individual clones generated in this study are available upon request.
For members of the Rosaceae community or of the public who are interested in searching the dataset, the EST search page allows users to search the Fragaria sequence set directly [48]. The ESTs and the unigene can be searched by name, by homology, and by features such as presence of a microsatellite or component of a contig. Once an EST or contig has been selected, the sidebar allows users to view all information relating to the sequence (or consensus sequence), the library details, the assembly information, the open reading frame and microsatellites, homology, and for contigs, the component ESTs.
Authors' contributions
KF prepared the RNA for, and generated the cDNA libraries, provided functional annotation and analysis and drafted the manuscript with MS. MS and CJ performed all computational analyses under the guidance of SJ and DM. PS and DB collected plant tissue, participated in RNA isolation and functional EST annotation. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by the Florida Agricultural Experiment Station, the NSF Plant Genome Research Program award #0320544 (DM), a grant from the North American Strawberry Growers Association (KMF) and funding from the Horticultural Sciences Department at the University of Florida (KMF). This work was approved for publication as Florida Agricultural Experiment Station Journal Series Number R-10920.
Figures and Tables
Figure 1 Strawberry EST characterization as derived from GO analyses. ESTs were assigned to GO categories based onA) Functional GO matches B) Process GO matches C) Component GO matches.
Table 1 Homology between the Fragaria × ananassa unigene sequences in public database as inferred through comparisons to Genbank nr, SWISSPROT, Rosaceae ESTs and mapped peach ESTs.
nr.pep.11 sprot.fas1 Rosaceae ESTs2 mapped peach ESTs2
database size 2321663 153871 227250 256
total # sequences 1304 1304 1304 1304
# sequences with matches 1105 714 835 22
% sequences with matches 84.74% 54.75% 64.03% 1.69%
# sequences w/o matches 199 590 469 1282
% sequences w/o matches 15.26% 45.25% 35.97% 98.31%
1 FASTX3.4 Algorithm with cut-off of < 1e-6
2 BLASTN Algorithm with identity >= 85% and overlap > = 100
Table 2 The summarized results of homology searches of the Fragaria × ananassa unigene sequences against SPROT, nr pep and Rosaceae ESTs.
Number Percent
Sequences with hits to sprot 714 54.75%
Sequences without hits to sprot 590 45.25%
Sequences without hits to sprot with hits to nr.pep 399 30.60%
Sequences without hits to sprot or nr.pep 191 14.64%
Sequences without hits to sprot or nr.pep but with hits to Rosaceae 54 4.14%
Sequences without hits to any database 137 10.51%
Table 3 Fragaria ESTs with homology to mapped peach BACs
Strawberry Unigene Peach EST GenBank Accession of Peach EST Homology E-Value from BLAST output
Contig 25 PP_LEa0003M24f BU039764 enolase 2 (allergen HEVb8) 1e-111
Contig 37 PP_LEa0003O13f BU039800 ubiquitin extension protein 6e-33
Contig 64 PP_LEa0003O13f BU039800 ubiquitin extension protein 3e-51
Contig 131 PP_LEa0003M11f BU039753 glyceraldehyde-3-phosphate dehydrogenase 0.0
FA_SEa0001D01r PP_LEa0009G24f BU041453 ion stress related protein 2e-39
FA_SEa0003G08r PP_LEa0027M15f BU046817 20S proteosome subunit PAF1 1e-166
FA_SEa0006D03r PP_LEa0011D12f BU042012 ribosomal protein L7 2e-34
FA_SEa0007B10r PP_LEa0003O13f BU039800 ubiquitin extension protein 8e-35
FA_SEa0008B09r PP_LEa0012A24f BU042298 expressed protein 1e-78
FA_SEa0008C05r PP_LEa0004I18f BU039998 cysteine protease 1e-81
FA_SEa0009F12r PP_LEa0035B03f BU048314 allergen PRU AV 1 4e-32
FA_SEa0011D12r PP_LEa0003O13f BU039800 ubiquitin extension protein 4e-66
FA_SEa0011E01r PP_LEa0009N05f BU041589 expressed protein 4e-63
FA_SEa0011F03r PP_LEa0011F03f BU042049 Centrin 1e-74
FA_SEa0012D04r PP_LEa0036C16f BU048550 putative gsh-dependent dehydroascorabate reductase 1 1e-144
FA_SEa0015A08r PP_LEa0006B19f BU040484 heat shock protein 70 0.0
FA_SEa0015F10r PP_LEa0003P11f BU039816 cytosolic aldolase 2e-30
FA_SEa0015H12r PP_LEa0013C06f BU042571 cell switch protein e-177
FA_SEa0016A10r PP_LEa0035H24f BU048407 putative ribosomal protein 9e-64
FA_SEa0018A02r PP_LEa0027P18f BU046858 ADP-ribosylation factor 0.0
FA_SEa0018D07r PP_LEa0036E14f BU048583 putative glyoxylase II 3e-47
FA_SEa0018F08r PP_LEa0035M02f BU048454 NADH-cytochrome b5 reductase 2e-61
Table 4 The frequency of simple sequence repeats in the Fragaria unigene.
Motif Length Frequency Percentage Frequency
2 bp 88 32.7%
3 bp 102 36.8%
4 bp 75 27.9%
5 bp 7 2.6%
Table 5 Frequency of different types of dinucleotide and trinucleotide repeats in the Fragaria unigene set
Motif Frequency Percentage Frequency
AT/TA 15 5.58%
AG/GA/CT/TC 65 24.16%
AC/CA/TG/GT 8 2.97%
GC/CG 0 0.00%
AAT/ATA/TAA/ATT/TTA/TAT 2 0.74%
AAG/AGA/GAA/CTT/TTC/TCT 35 13.01%
AAC/ACA/CAA/GTT/TTG/TGT 4 1.49%
ATG/TGA/GAT/CAT/ATC/TCA 6 2.23%
AGT/GTA/TAG/ACT/CTA/TAC 0 0.00%
AGG/GGA/GAG/CCT/CTC/TCC 21 7.81%
AGC/GCA/CAG/GCT/CTG/TGC 10 3.72%
ACG/CGA/GAC/CGT/GTC/TCG 7 2.60%
ACC/CCA/CAC/GGT/GTG/TGG 10 3.72%
GGC/GCG/CGG/GCC/CCG/CGC 7 2.60%
Table 6 Statistics for optimal primer candidates
Inside ORF Outside ORF
2 bp 35 (24.6%) 21 (31.8%)
3 bp 77 (54.2%) 14 (21.2%)
4 bp 29 (20.4%) 26 (39.4%)
5 bp 1 (0.7%) 5 (7.6%)
Total 142 (68.3%) 66 (31.7%)
Table 7 Transcripts corresponding to genes of described function in important physiological processes
EST Homolog E
Photoperiodic Control of Flowering Time
FA_Sea0007C05 B-box, zinc-finger protein CONSTANS 7.30E-21
FA_Sea0020G05 B-box, zinc-finger protein CONSTANS 3.40E-06
FA_Sea0016A05 MADS box protein AGL20/SUPPRESSOR OF CONSTANS 1.90E-15
FA_Sea0002H08 VIN3 – Vernalization insensitive 3 protein 1.50E-33
Disease Resistance
FA_Sea0004D05 Disease resistance protein (TIR-NBS-LRR class) 6.80E-18
FA_Sea0006F10 Enhanced Disease Susceptibility protein EDS5 7.10E-58
FA_Sea0007F04 Plant defensin PDF2.2 3.10E-22
FA_Sea0010B10 Pathogenesis-related thaumatin (PR5) 1.20E-21
FA_Sea0014H12 Putative thaumatin (PR5) 7.10E-16
FA_Sea0015A01 Harpin-induced protein 3.80E-26
FA_Sea0015D01 NDR1 family protein 7.00E-69
FA_Sea0017F09 Disease resistance protein (CC-NBS-LRR class) 2.10E-23
FA_SEa0020H01 Harpin-induced protein 3.40E-18
FA_Sea0010F01 glycosyl hydrolase family 17 p (PR2) 2.60E-12
FA_Sea0017H06 Osmotin-like protein (PR5) 5.00E-13
FA_SEa0001D03 Peroxidase PRXR1 (PR9) 8.20E-51
FA_SEa0019D07 Bet v 1 (PR10) 2.10E-35
FA_SEa0012C06 Lipid transfer protein LPT4 (PR14) 1.70E-19
Photomorphogenesis
FA_Sea0004E09 B-zip transcription factor HY5 4.60E-37
FA_Sea0001C09 NON-PHOTOTROPIC HYPOCOTYL 3 3.20E-31
FA_SEa0006H04 Far-red impaired / FAR1 3.30E-29
==== Refs
Chandler CK Legard DE Dunigan DD Crocker TE Sims TA 'Strawberry Festival' strawberry HortScience 2000 35 1366 1367
Passey AJ Barrett KJ James DJ Adventitious shoot regeneration from seven commercial strawberry cultivars (Fragaria x ananassa Duch.) using a range of explant types Plant Cell Rep 2003 21 397 401 12789440
Rugini E Orlando R High-Efficiency Shoot Regeneration from Calluses of Strawberry (Fragaria x ananassa-Duch) Stipules of In-vitro Shoot Cultures J Hortic Sci 1992 67 577 582
Alsheikh MK Suso HP Robson M Battey NH Wetten A Appropriate choice of antibiotic and Agrobacterium strain improves transformation of anti biotic-sensitive Fragaria vesca and F.v. semperflorens Plant Cell Rep 2002 20 1173 1180 10.1007/s00299-002-0453-0
Akiyama Y Yamamoto Y Ohmido N Oshima M Fukui K Estimation of the nuclear DNA content of strawberries (Fragaria spp.) compared with Arabidopsis thaliana by using dual-stem flow cytometry. Cytologia 2001 66 431 436
Darrow GM The Strawberry 1966 New York, Holt, Rinehart and Winston
Pearson WR Lipman DJ Improved tools for biological sequence comparison Proc Natl Acad Sci U S A 1988 85 2444 2448 3162770
Strawberry FA_SEa EST SwissProt Database GO-IDs: http://www.genome.clemson.edu/cgi-bin/DisplSwissProtTbl.cgi?map=Strawberry&type=EST&gocounts.
Horn R A L Garay L McCord P Howad W Chan H Georgi L Main D Jung S Jesudurai C Ramaswamy K Forrest S Mook J Zhebentyayeva T Callahan A Dandekar A Sosinski B Arus P Baird V Parfitt D Reighard G Scorza R Wing R Abbott A Development of a candidate gene database and a transcript map for peach, a model species for fruit trees. Theoretical and Applied Genetics 2005 110 1419 1428 15846479 10.1007/s00122-005-1968-x
Index of Fragaria sequence data: ftp://ftp.genome.clemson.edu/pub/fragaria
Senanayake YDA Bringhurst RS Origin of Fragaria Polyploids .I. Cytological Analysis Am J Bot Am J Bot 1967 54 221 228
Bringhurst RS Cytogenetics and Evolution in American Fragaria Hortscience 1990 25 879 881
Wein M Lavid N Lunkenbein S Lewinsohn E Schwab W Kaldenhoff R Isolation, cloning and expression of a multifunctional O-methyltransferase capable of forming 2,5-dimethyl-4-methoxy-3(2H)-furanone, one of the key aroma compounds in strawberry fruits Plant J 2002 31 755 765 12220266 10.1046/j.1365-313X.2002.01396.x
Castillejo C de la Fuente JI Iannetta P Botella MA Valpuesta V Pectin esterase gene family in strawberry fruit: study of FaPE1, a ripening-specific isoform J Exp Bot 2004 55 909 918 15020638 10.1093/jxb/erh102
Llop-Tous I Dominguez-Puigjaner E Palomer X Vendrell M Characterization of Two Divergent Endo-beta -1,4-Glucanase cDNA Clones Highly Expressed in the Nonclimacteric Strawberry Fruit Plant Physiol 1999 119 1415 1422 10198101 10.1104/pp.119.4.1415
Benitez-Burraco A Blanco-Portales R Redondo-Nevado J Bellido ML Moyano E Caballero JL Munoz-Blanco J Cloning and characterization of two ripening-related strawberry (Fragaria x ananassa cv. Chandler) pectate lyase genes J Exp Bot J Exp Bot 2003 54 633 645 10.1093/jxb/erg065
Jimenez-Bermudez S Redondo-Nevado J Munoz-Blanco J Caballero JL Lopez-Aranda JM Valpuesta V Pliego-Alfaro F Quesada MA Mercado JA Manipulation of strawberry fruit softening by antisense expression of a pectate lyase gene Plant Physiology 2002 128 751 759 11842178 10.1104/pp.128.2.751
Agius F Gonzalez-Lamothe R Caballero JL Munoz-Blanco J Botella MA Valpuesta V Engineering increased vitamin C levels in plants by overexpression of a D-galacturonic acid reductase Nature Biotechnology 2003 21 177 181 12524550 10.1038/nbt777
Blanco-Portales R Lopez-Raez JA Bellido ML Moyano E Dorado G Gonzalez-Reyes JA Caballero JL Munoz-Blanco J A strawberry fruit-specific and ripening-related gene codes for a HyPRP protein involved in polyphenol anchoring Plant Molecular Biology 2004 55 763 780 15604715
Devlin PF Yanovsky MJ Kay SA A genomic analysis of the shade avoidance response in Arabidopsis Plant Physiol 2003 133 1617 1629 14645734 10.1104/pp.103.034397
Folta KM Pontin MA Karlin-Neumann G Bottini R Spalding EP Genomic and physiological studies demonstrate roles for auxin and gibberellin in the early phase of cryptochrome 1 action in blue light Plant J 2003 36 203 214 14535885 10.1046/j.1365-313X.2003.01870.x
Aharoni A Giri AP Verstappen FW Bertea CM Sevenier R Sun Z Jongsma MA Schwab W Bouwmeester HJ Gain and loss of fruit flavor compounds produced by wild and cultivated strawberry species Plant Cell 2004 16 3110 3131 15522848 10.1105/tpc.104.023895
Aharoni A Keizer LC Bouwmeester HJ Sun Z Alvarez-Huerta M Verhoeven HA Blaas J van Houwelingen AM De Vos RC van der Voet H Jansen RC Guis M Mol J Davis RW Schena M van Tunen AJ O'Connell AP Identification of the SAAT gene involved in strawberry flavor biogenesis by use of DNA microarrays Plant Cell 2000 12 647 662 10810141 10.1105/tpc.12.5.647
Aharoni A O'Connell AP Gene expression analysis of strawberry achene and receptacle maturation using DNA microarrays J Exp Bot 2002 53 2073 2087 12324531 10.1093/jxb/erf026
Putterill J Robson F Lee K Simon R Coupland G The CONSTANS gene of Arabidopsis promotes flowering and encodes a protein showing similarities to zinc finger transcription factors Cell 1995 80 847 857 7697715 10.1016/0092-8674(95)90288-0
Valverde F Mouradov A Soppe W Ravenscroft D Samach A Coupland G Photoreceptor regulation of CONSTANS protein in photoperiodic flowering Science 2004 303 1003 1006 14963328 10.1126/science.1091761
Onouchi H Igeno MI Perilleux C Graves K Coupland G Mutagenesis of plants overexpressing CONSTANS demonstrates novel interactions among Arabidopsis flowering-time genes Plant Cell 2000 12 885 900 10852935 10.1105/tpc.12.6.885
Sung S Amasino RM Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3 Nature 2004 427 159 164 14712276 10.1038/nature02195
Michaels SD Amasino RM Loss of FLOWERING LOCUS C activity eliminates the late-flowering phenotype of FRIGIDA and autonomous pathway mutations but not responsiveness to vernalization Plant Cell 2001 13 935 941 11283346 10.1105/tpc.13.4.935
Liu Y Roof S Ye Z Barry C van Tuinen A Vrebalov J Bowler C Giovannoni J Manipulation of light signal transduction as a means of modifying fruit nutritional quality in tomato Proc Natl Acad Sci U S A 2004 101 9897 9902 15178762 10.1073/pnas.0400935101
Martinez-Garcia JF Huq E Quail PH Direct targeting of light signals to a promoter element-bound transcription factor Science 2000 288 859 863 10797009 10.1126/science.288.5467.859
Dirlewanger E Graziano E Joobeur T Garriga-Caldere F Cosson P Howad W Arus P Comparative mapping and marker-assisted selection in Rosaceae fruit crops Proc Natl Acad Sci U S A 2004 101 9891 9896 15159547 10.1073/pnas.0307937101
Sargent DJ Davis TM Tobutt KR Wilkinson MJ Battey NH Simpson DW A genetic linkage map of microsatellite, gene-specific and morphological markers in diploid Fragaria Theor Appl Genet 2004
Chang S Puryear J Cairney J A simple and efficient method for isolating RNA from pine trees Plant Molecular Biology Reporter 1993 11 113 116
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
Gordon D Abajian C Green P Consed: a graphical tool for sequence finishing Genome Res 1998 8 195 202 9521923
dbEST website: http://www.ncbi.nlm.nih.gov/dbEST
Huang X Madan A CAP3: A DNA sequence assembly program Genome Res 1999 9 868 877 10508846 10.1101/gr.9.9.868
Wheeler DL Barrett T Benson DA Bryant SH Canese K Church DM DiCuccio M Edgar R Federhen S Helmberg W Kenton DL Khovayko O Lipman DJ Madden TL Maglott DR Ostell J Pontius JU Pruitt KD Schuler GD Schriml LM Sequeira E Sherry ST Sirotkin K Starchenko G Suzek TO Tatusov R Tatusova TA Wagner L Yaschenko E Database resources of the National Center for Biotechnology Information Nucleic Acids Res 2005 33 Database Issue D39 45 15608222
Boeckmann B Bairoch A Apweiler R Blatter MC Estreicher A Gasteiger E Martin MJ Michoud K O'Donovan C Phan I Pilbout S Schneider M The SWISS-PROT protein knowledgebase and its supplement TrEMBL in 2003 Nucleic Acids Res 2003 31 365 370 12520024 10.1093/nar/gkg095
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 The Gene Ontology (GO) database and informatics resource Nucleic Acids Res 2004 32 Database issue D258 61 14681407
InterProScan website: http://www.ebi.ac.uk/InterProScan
Joobeur T Viruel MA de Vicente MC Jauregui B Ballester J Dettori MT Verde I Truco MJ Messeguer R Batlle I Quarta R Dirlewanger E Arus P Construction of a saturated linkage map for Prunus using an almond x peach F-2 progeny Theoretical and Applied Genetics 1998 97 1034 1041 10.1007/s001220050988
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
Bossard N FLIP: a Unix program used to find/translate ORFs 1997 , Bionet Software
Temnykh S DeClerck G Lukashova A Lipovich L Cartinhour S McCouch S Computational and experimental analysis of microsatellites in rice (Oryza sativa L.): Frequency, length variation, transposon associations, and genetic marker potential Genome Research 2001 11 1441 1452 11483586 10.1101/gr.184001
University of Florida Strawberry Project: http://www.mainlab.clemson.edu/gdr/projects/fragaria/folta/FA_SEa/index.shtml
BLAST search strawberry sequences: http://www.mainlab.clemson.edu/gdr/ESTsearchStrawberry.shtml
|
15985176
|
PMC1182381
|
CC BY
|
2021-01-04 16:03:52
|
no
|
BMC Plant Biol. 2005 Jun 28; 5:12
|
utf-8
|
BMC Plant Biol
| 2,005 |
10.1186/1471-2229-5-12
|
oa_comm
|
==== Front
BMC Struct BiolBMC Structural Biology1472-6807BioMed Central London 1472-6807-5-121601180310.1186/1472-6807-5-12DatabaseNh3D: A reference dataset of non-homologous protein structures Thiruv B [email protected] G [email protected] SA [email protected] B [email protected] Department of Biochemistry, University of Toronto, 1 Kings College Circle, Toronto, Ontario M5S 1A8, Canada2005 12 7 2005 5 12 12 11 2 2005 12 7 2005 Copyright © 2005 Thiruv 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 statistical analysis of protein structures requires datasets in which structural features can be considered independently distributed, i.e. not related through common ancestry, and that fulfil minimal requirements regarding the experimental quality of the structures it contains. However, non-redundant datasets based on sequence similarity invariably contain distantly related homologues. Here we provide a reference dataset of non-homologous protein domains, assuming that structural dissimilarity at the topology level is incompatible with recognizable common ancestry. The dataset is based on domains at the Topology level of the CATH database which hierarchically classifies all protein structures. It contains the best refined representatives of each Topology level, validates structural dissimilarity and removes internally duplicated fragments. The compilation of Nh3D is fully scripted.
Results
The current Nh3D list contains 570 domains with a total of 90780 residues. It covers more than 70% of folds at the Topology level of the CATH database and represents more than 90% of the structures in the PDB that have been classified by CATH. We observe that even though all protein pairs are structurally dissimilar, some pairwise sequence identities after global alignment are greater than 30%.
Conclusion
Nh3D is freely available as a reference dataset for the statistical analysis of sequence and structure features of proteins in the PDB. Regularly updated versions of Nh3D and the corresponding PDB-formatted coordinate sets are accessible from our Web site .
==== Body
Background
The number of structures in the Protein Data Bank (PDB) [1] has grown to over 30,000. Many of the proteins in the PDB are homologous, i.e. have descended from a common ancestor, conserving significant aspects of their structure, function, and sequence. For purposes such as a statistical analysis of protein structure features, a subset of the PDB is required in which structural features can be presumed to be independently distributed, i.e. unbiased with respect to evolutionary descent.
A number of PDB subsets have been proposed to address this need. The most often used subset of structures in the PDB is the PDBSELECT [2,3]. This provides lists of chains compiled at predefined maximum percent sequence identities. The most stringent cutoff employed is a maximum of 25% residue identity between chains. Other subsets culled at predefined sequence identity are available from the PDB itself. In recent years, servers such as the PDB-REPRDB [4,5] and PISCES [6] allow users to compile customized lists of protein chains based on structure quality and maximum mutual sequence identities. PISCES allows the user to define additional restriction parameters such as the minimum sequence length and maximum R-value. However, all of the above-mentioned lists use sequence similarity as the defining criterion for protein selection and include distantly related but recognizably homologous proteins.
In order to define a reference dataset of non-homologous protein structures, we have constructed a tool to extract coordinate subsets from the PDB, driven by CATH [7] domain topology information. The underlying hypothesis is that differences in the topology of protein secondary structure arrangements are incompatible with the notion of gradual, divergent evolution from a common ancestor. The experimentally best-defined representatives of domain sets at the CATH Topology level are selected, the structures are validated with respect to sequence/structure similarity that might be indicative of homology, and homologous internal repeats are purged to remove redundancies. Our reference database of non-homologous, dissimilar structures is versioned and automatically updated with every revision of the CATH database.
Construction and contents
The CATH database is a hierarchic classification of protein domains. At the Topology level of the hierarchy (similar to the Fold-level of SCOP [8]), the domains are globally dissimilar and thus provide a set of structures that are not recognizably homologous.
To create Nh3D (Figure 1), representative structures were selected for each of the Topology for classes 1–4 in the CATH hierarchy (Version 2.5.1: released January 2004). We have not considered proteins classified into classes 5–9, because these assignments are preliminary. Using the PDB resolution and compound index files, we have selected the best representative for each Topology, that minimally fulfill the following criteria:
1) X-ray or neutron-diffraction structure, resolved at 2.2 Å or better (since we consider NMR structures to be less reliable for the statistical analysis of detailed structural relationships);
2) Domain has a minimum length of 50 residues (since we consider shorter structured domains, frequently dominated by a large percentage of disulfide bonds, to be atypical for independent folding units);
3) Engineered structures are excluded, unless no alternatives are available (we exclude structures that contain the sub-strings {MUTAT | MUTAN | REPLACE} in their description line of the compound index file since engineering may have introduced structural strain that has not been resolved through evolutionary optimization of the protein.).
When two or more structures meet the selection criteria, the one with the better resolution is chosen as the representative. For structures that have the same resolution, the one with the better Ramachandran Z score (obtained from the PDBFINDERII [9] database) is chosen as the representative.
To ensure that these structurally dissimilar domains are indeed not recognizably homologous, two stages of validation and purification are employed (Figure 2).
1) Evaluation of structural and sequence similarity between the domains
We check for sequence similarities that correspond to globally similar structures, to guard against the inclusion of homologous subdomains that might be variably assigned to different Topologies due to imprecise definitions of structural domains. Exhaustive pairwise global alignments of all domains are performed using the Needle program from the EMBOSS suite [10] with the EBLOSUM62 similarity matrix and default gap and extension parameters of -10.0 and -0.5. The RMSD of aligned residues after optimal superposition between their backbone atoms is calculated by the Kabsch method [11]. Domains with greater than 40 residues aligned (a conservative limit for foldable and thus independently inheritable domains), possessing greater the 25% sequence identity (a threshold commonly used to confidently identify homologues through sequence analysis) and superimposable at less then 3.8 Å RMSD (the average Cα – Cα distance that defines the limits of a meaningful pairwise residue superposition) are considered homologous. The smaller of the two domains is removed from the dataset. None of the domain pairs in this version of the CATH database met the above criteria for exclusion; this emphasizes the reliability of CATH domain assignments.
2) Identification of repeat regions in domains
Internal repeats may arise due to internal duplication events and are known to occur more frequently after the first duplication [12]. The internal repeats in Nh3D are identified using the program RADAR [13]. The RMSD after optimal superposition between the RADAR aligned repeat residues is calculated by the Kabsch method. Repeats with greater than 25% sequence identity and less than 3.8Å RMSD are considered redundant and removed from the database.
We annotate the final set of protein structures with the following information:
a) secondary structure assignment using DSSP [14] and STRIDE [15] algorithms
b) relative side-chain solvent accessible surface areas using the Shrake and Rupley [16] approach and a reference dataset of modeled, extended GLY-XXX-GLY tripeptides
c) regions with high B-Factor (where the average B-factor of the backbone atoms of three consecutive residues is found to be greater than 60 Å 2are identified and flagged)
d) beta and gamma turn assignments using PROMOTIF [17]
Residues in the PDB coordinate files that have missing backbone residues are identified and removed from the dataset. Modified residues that appear as 'HETATM' records in the PDB files are treated as chain breaks; a distance check between bonded backbone atoms is implemented to check for additional discontinuities in a chain.
Nh3D is freely available in an easily parseable, fully specified flatfile format and as PDB-formatted coordinates. Differently formatted output for specific needs can be made available upon request to the authors.
Utility and discussion
Our dataset uses coordinates from 477 proteins consisting of 570 of 820 CATH Topologies and contains 90780 residues. The chosen toplogies represent over 90% of the structures in the PDB classified by CATH version 2.5.1.
Interestingly, we observe numerous domains with striking sequence similarity (> 30 % over the aligned residues), with completely dissimilar structure. A single likely homologue (CATH domains 1d0cA3 and 1nos03, 42.7 % identity in 89 aligned residues) has been identified in CATH Version 2.5.1. However, the domain 1nos03 was already excluded from the database in a previous step, due to the high residue B-factors over the length of its domain.
Significant internal repeats were removed from 38 domains. For example, repeats were found in representatives of Porin, Tachylectin-2, and MutS topologies. Other repeats, most notably those of TIM barrel proteins, did not meet our cutoff criteria for recognizable homology and hence were retained.
A comparison of our results with a sequence based PDB subset (PDBSELECT of October 2004) shows the large number of redundant, similar structures that cannot be eliminated based on sequence dissimilarity alone. The most redundant is the Rossman fold (CATH: 3.40.50, 105 families; Nh3D representative: 1GCI:1-275; PDBselect: 138 sequences), the second most redundant are immunoglobulin-like domains (CATH: 2.60.40, 52 families; Nh3D representative: 1OE1:A:2-152; PDBselect: 93 sequences). While it has been a matter of discussion whether such "superfolds" might be the result of convergent evolution [see e.g. [18]], we take a conservative approach with Nh3D. In constructing the reference dataset by selecting a single representative for every CATH Topology, the residual chance of inappropriately including homologues is minimized. Nh3D does not attempt to be exhaustive at this point. For example, our "immunoglobulin-like" topology representative is the human copper, zinc superoxide dismutase, 1MFM:A:1-153, however the metal-binding greek-key proteins and the immunoglobulin superfamily, both of which are included in this class, may indeed be examples of convergent evolution towards a common fold. Since Nh3D is a representative sample of detailed conformations in protein structures, completeness is not an issue, given that the number of excluded analogous folds is small relative to those folds for which we do not yet possess high-resolution structures. Nevertheless, we wish to emphasize a continuing need to further improve on the automated definition and classification of dissimilar folds.
Conclusion
Nh3D is freely accessible as a reference dataset for the unbiased statistical analysis of sequence and structure features of proteins in the PDB. While our own motivation was to construct a source dataset for an analysis of short peptide conformations, we anticipate many uses of Nh3D, including the calibration of algorithms for the detection of distant gene relationships, for the recognition of false positives in sequence or structure alignments, or for the prediction of protein structure.
Availability and requirements
The dataset, a list of best representatives for CATH Topologies that did not meet our criteria, documentation, format specifications and supplementary information can be downloaded from . Nh3D Version 2.0 based on CATH 2.6.0 has since been made available.
List of abbreviations
CATH-A hierarchical classification of protein domain structures
DSSP-Definition of Secondary Structure of Proteins
EMBOSS – European Molecular Biology Open Software Suite
NMR-Nuclear Magnetic Resonance
SCOP-Structural Classification of Proteins
STRIDE – STRuctural IDEntification method
PDB-Protein Data Bank
RADAR – Rapid Automatic Detection and Alignment of Repeats
RMSD – Root Mean Square Deviation
WWW-World Wide Web
Authors' contributions
BTG wrote the second version of the code to generate Nh3D and helped to draft the manuscript.
GQ wrote the first version of the code and helped to prepare the manuscript.
SAS participated in the design and helped to draft the manuscript.
BS conceived of the study, and participated in its design and coordination and drafted the manuscript.
Acknowledgements
Funding contributions are gratefully acknowledged from the Canadian Institutes of Health Research (operating grant MOP 93075) and from a sub-grant within the Genome Canada Competition II project "An Integrated & Distributed Bioinformatics Platform for Genome Canada".
Figures and Tables
Figure 1 Data Flow Diagram for Nh3D, from start to a list of domain sequences. For details see text.
Figure 2 Data Flow Diagram for Nh3D, from a list of domain sequences to the final reference dataset. For details see text.
==== Refs
Bernstein F Koetzle T Williams G Meyer EJ Brice M Rodgers J Kennard O Shimanouchi T Tasumi M The Protein Data Bank: a computer-based archival file for macromolecular structures J Mol Biol 1977 112 535 42 875032
Hobohm U Scharf M Schneider R Sander C Selection of representative protein data sets Protein Sci 1992 1 409 17 1304348
Hobohm U Sander C Enlarged representative set of protein structures Protein Sci 1994 3 522 24 8019422
Noguchi T Matsuda H Akiyama Y PDB-REPRDB: a database of representative protein chains from the Protein Data Bank (PDB) Nucl Acids Res 2001 29 219 20 11125096 10.1093/nar/29.1.219
Noguchi T Akiyama Y PDB-REPRDB: a database of representative protein chains from the Protein Data Bank (PDB) in 2003 Nucleic Acids Res 2003 1 492 93 10.1093/nar/gkg022
Wang G Dunbrack RL Jr PISCES: A protein sequence culling server Bioinformatics 2003 19 1589 91 12912846 10.1093/bioinformatics/btg224
Orengo CA Michie AD Jones S Jones DT Swindells MB Thornton JM CATH-A Hierarchic Classification of Protein Domain Structures Structure 1997 5 1093 1108 9309224 10.1016/S0969-2126(97)00260-8
Murzin AG Brenner SE Hubbard T Chothia C SCOP: a structural classification of proteins database for the investigation of sequences and structures J Mol Biol 1995 247 536 40 7723011 10.1006/jmbi.1995.0159
Krieger E Hooft RWW Nabuurs S Vriend G PDBFinderII – a database for protein structure analysis and prediction 2004
Rice P Longden I Bleasby A EMBOSS: The European Molecular Biology Open Software Suite Trends in Genetics 2000 16 276 77 10827456 10.1016/S0168-9525(00)02024-2
Kabsch W A discussion of the solution for the best rotation to relate two sets of vectors Acta Cryst 1978 A34 827 28
Marcotte EM Pellegrini M Yeates TO Eisenberg D A census of protein repeats J Mol Biol 1999 293 151 60 10512723 10.1006/jmbi.1999.3136
Heger A Holm L Rapid automatic detection and alignment of repeats in protein sequences Proteins 2000 41 224 37 10966575 10.1002/1097-0134(20001101)41:2<224::AID-PROT70>3.0.CO;2-Z
Kabsch W Sander C Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features Biopolymers 1983 22 2577 2637 6667333 10.1002/bip.360221211
Frishman D Argos P Knowledge-based protein secondary structure assignment Proteins 1995 23 566 79 8749853 10.1002/prot.340230412
Shrake A Rupley JA Environment and exposure to solvent of protein atoms. Lysozyme and insulin J Mol Biol 1973 79 351 71 4760134 10.1016/0022-2836(73)90011-9
Hutchinson EG Thornton JM PROMOTIF-a program to identify and analyze structural motifs in proteins Protein Sci 1996 5 212 20 8745398
Mirny LA Shakhnovich EI Universallly conserved positions in protein folds: reading evolutionary signals about stability, folding kinetics and function J Mol Biol 1999 291 177 96 10438614 10.1006/jmbi.1999.2911
|
16011803
|
PMC1182382
|
CC BY
|
2021-01-04 16:03:52
|
no
|
BMC Struct Biol. 2005 Jul 12; 5:12
|
utf-8
|
BMC Struct Biol
| 2,005 |
10.1186/1472-6807-5-12
|
oa_comm
|
==== Front
BMC SurgBMC Surgery1471-2482BioMed Central London 1471-2482-5-141596322710.1186/1471-2482-5-14Technical AdvanceCan bile duct injuries be prevented? "A new technique in laparoscopic cholecystectomy" Sari Yavuz Selim [email protected] Vahit [email protected] Kamer [email protected]öz Binnur [email protected]üneyİ Ayhan [email protected]öZ İbrahim [email protected] SSK İstanbul Training Hospital Department of General Surgery – Istanbul, Turkey2 Saint Georg Hospital Department of General Surgery, Hamburg, Austria2005 17 6 2005 5 14 14 11 2 2005 17 6 2005 Copyright © 2005 sari et al; licensee BioMed Central Ltd.2005sari 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, laparoscopic cholecystectomy has gained worldwide acceptance and considered to be as "gold standard" in the surgical management of symptomatic cholecystolithiasis. However, the incidence of bile duct injury in laparoscopic cholecystectomy is still two times greater compared to classic open surgery. The development of bile duct injury may result in biliary cirrhosis and increase in mortality rates. The mostly blamed causitive factor is the misidentification of the anatomy, especially by a surgeon who is at the beginning of his learning curve. Biliary tree injuries may be decreased by direct coloration of the cystic duct, ductus choledochus and even the gall bladder.
Methods
gall bladder fundus was punctured by Veress needle and all the bile was aspirated. The same amount of fifty percent methylene blue diluted by saline solution was injected into the gall bladder for coloration of biliary tree. The dissection of Calot triangle was much more safely performed after obtention of coloration of the gall bladder, cystic duct and choledocus.
Results
Between October 2003 and December 2004, overall 46 patients (of which 9 males) with a mean age of 47 (between 24 and 74) underwent laparoscopic cholecystectomy with methylene blue injection technique. The diagnosis of chronic cholecystitis (the thickness of the gall bladder wall was normal) confirmed by pre-operative abdominal ultrasonography in all patients. The diameters of the stones were greater than 1 centimeter in 32 patients and calcula of various sizes being smaller than 1 cm. were documented in 13 cases. One patient was operated for gall bladder polyp (our first case). Successful coloration of the gall bladder, cystic duct and ductus choledochus was possible in 43 patients, whereas only the gall bladder and proximal cystic duct were visualised in 3 cases. In these cases, ductus choledochus visibility was not possible. None of the patients developed bile duct injury.
Conclusion
The number of bile duct injuries related to anatomic misidentification can be decreased and even vanished by using intraoperative methylene blue injection technique into the gall bladder fundus intraoperatively.
==== Body
Background
Laparoscopic cholecystectomy (LC) is considered as the "golden standard" in the surgical management of symptomatic cholelithiasis. Short hospitalisation period and rapid return to normal activity, less post-operative pain, more acceptable cosmetic results and lesser morbidity and mortality rates, are the principle advantages of this technique. However, the incidence of bile duct injuries is two times greater when compared to open cholecystectomy [1-11]. Bile duct injury, either in classic open or laparoscopic cholecystectomy, may necessitate several consecutive operations and invasive procedures, causing fear and anxiety to all surgeons.
The development of bile duct injuries following LC is not common but a serious complication resulting in long-term morbidity [8,12]. When the literature is reviewed, the incidence of bile duct injuries in LC is between 0,3 – 0,6 % [4,6-10,13-15], which may be considered an acceptable percentage, may in fact result in secondary biliary cirrhosis with considerable financial burden [6,8,10]. Higher incidence of biliary tree injuries has also been reported [3]. In United States, 600 000 cases of laparoscopic cholecystectomies are performed annually. When this number is taken into consideration, it will be clearly understood that the economic problem caused by even small (0,3 – 0,6 %) rates of bile duct injuries, can not be underestimated[1,15].
Herein, we introduce a new technique, with the hope to reduce bile duct injuries during LC and we publish the results of 46 cases.
Methods
The patients were installed in French position. The trocards were placed as in French position. The gall bladder fundus was grasped and held tight towards the anterior abdominal wall with the help of two atraumatic pinces (or graspers) introduced via right anterior axillary and subxyphoid trocards. The gall bladder fundus was punctured by a Veress needle which was introduced via the abdominal wall in projection to this area.
All the bile in the gall bladder was aspirated and 50 percent diluted methylene blue equal to the amount of aspirated bile was injected slowly into the gall bladder (Figure 1). In order to prevent bile leakage, the gall bladder fundus was held tight anteriorly during the withdrawal of the Veress needle and a grasper introduced via the xyphoid trocard was applied immediately to the puncture site and was held so throughout the operation. During cholecystectomy, the gall bladder, cystic duct and ductus choledochus were visible with methylene blue dye and the dissection was performed more safely (Figure 2).
The gall bladder was removed from the abdominal cavity through the trocar inserted from lateral border of left rectus muscle. In order to minimize bile leakage into the abdominal cavity, the gall bladder was completely aspirated before removal from the abdominal cavity.
Results
Between October 2003 and December 2004, 46 patients underwent LC by "Methylene blue dye injection" technique. 37 patients were female (mean age 45) and 9 patients were male (mean age 52). Chronic cholecystitis was found in all patients in pre-operative ultrasonographic evaluation. (wall thickness of the gall bladder was normal). The diameter of the stones in 32 patients was more than 1 centimeter and multiple small stones were found in 13 patients. One patient was operated with the diagnosis of gall bladder polyp (first case). The gall bladder, cystic duct and ductus choledochus were painted with methylene blue in 43 cases but only the gall bladder and the proximal cystic duct were visualised in 3 cases.
In 5 cases operated by the residents, methylene blue leakage from the gall bladder was observed into the abdominal cavity during the removal procedure. The region was irrigated with saline solution. All patients were informed that they might pass blue urine in the early post-operative period. None of the patients developed any complication and all of them were discharged the day after the operation. We did not use this technique in acute cases.
In one case, the trajectory of the cystic duct was very close to the right hepatic duct and its opening to main hepatic duct was just distal to the bifurcation. A trifurcation was demonstrated in one case where the cystic duct was directly opening to the branch of 6th and 7th hepatic segments. These anatomic variations were clearly demonstrated with our technique.
Discussion
Many factors have been incriminated in occurance of bile duct injuries during LC. These are mainly anatomical misidentification of main hepatic duct, right hepatic ducts or of aberrant right hepatic duct as ductus cysticus, other anatomical variations or unidentifiable anatomy, surgeon's experience (a surgeon who is at the beginning of his learning curve), technical difficulties, poor visualization of the operative field, acute and chronic inflammation of the gall bladder and local factors such as excessive haemorrage and fat tissue [1-3,5-8,15-17]. On the other hand, the problems related to the equipment have been accused [8,9]. However, misidentification of the anatomy and surgeon's experience seem to be preliminary [1,3,5,9,11,12,14-17].
Bile duct injuries are associated with significant morbidity, prolonged hospitalization, increased financial burden, potential litigation and occasional mortality [1,5,6,10,11]. It is the third most commonly litigated general surgical complication in The United States and it has been also reported that on the average two procedures (between 1 to 8) are required for definitive repair of bile ducts [5]. Obviously, if bile duct injury is noticed peroperatively and repaired in the best way, morbidity and mortality rates would be significantly reduced.
Although the importance of intraoperative cholangiography (IOC) to prevent bile duct injury is stressed by a significant number of authors, its role still remains controversial [1,5,8,9,15,17,18]. Despite its potential benefits, routine use of IOC has not taken its place in the surgical era. [2,4,5]. It has been claimed that routine use of IOC does not have a significant practical advantage [1,2,5,15,17]. Additionally, the operation room conditions should be suitable for IOC. Other disadvantages of IOC are; the necessity of some disposable equipment, the need of surgical experience, the inevitable prolongation of the operation time and the need of interpretation by an experienced radiologist.
During medical education, in Textbooks of Anatomy, we have seen the arteries nicely colored in red, the veins in blue and the lymphatics in yellow. Later on, facing the truth in cadavers, we were all somewhat disappointed. The idea of using methylene blue dye intraoperatively to colorise the anatomic details, is in fact based on this simple truth. The basic principle is to minimize the probable injuries by painting the gall bladder, ductus cysticus and ductus choledochus peroperatively.
As explained in the technique, a few minutes after methylene blue injection into the gall bladder, the gall bladder, ductus cysticus, ductus choledochus and in most of the cases even the duodenum have been painted. So the dissection can easily be performed. The operation time has not been significantly increased by methylene blue injection. Also, it is not associated with increased cost. Additionally, the flow of methylene blue from the nasogastric tube (noted by the anaesthesiologist) and/or the coloration of the duodenum may lead to indirect conclusion that the bile duct flow is uninterrupted.
When the images obtained by IOC are to be evaluated in the operating room, with this technique, a comparison with the anatomic details of the operating field may also be possible.
No bile duct injuries were encountered in 46 cases. This result may certainly not be directly attributable to our technique, however, it is true that the dissection was performed much more safely. Since the boundaries of the gall bladder were significantly painted with methylene blue, the residents managed to remove the gall bladder from the liver bed without causing any perforation.
Obviously, overlooked bile duct injuries during LC appears to be another problem [6,12-14,19]. We believe that it is possible to notice the bile duct injury with this technique because of methylene blue leakage from the injured area. In this case, the decision for conversion to open surgery may be easily taken into consideration.
Because methylene blue is excreted by the kidneys, the patients should be informed of the possibility of blue urine in the early post-operative period.
LC is successfully used in acute cases [7,20]. We did not use this technique in acute cases.
Conclusion
Bile duct injury during cholecystectomy is always a possibility, no matter which technique is used. In fact, the incidence of bile duct injury during laparoscopic cholecystectomy is slightly elevated compared to classic open surgery. A bile duct injury may be the beginninig of a catastrophic sequence of a serious complication. We believe that the incidence of bile duct injury related to anatomic misidentification can be decreased or even totally suppressed by intraoperative injection of methylene blue into the gall bladder fundus and visualisation of the gall bladder, cystic duct and ductus choledochus.
Figure 1 The injection of methylene blue into the gall bladder.
Figure 2 The visualisation of cystic duct and ductus choledochus with methylene blue after removal of gall bladder.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Ahrendt SA Pitt HA Surgical therapy of iatrogenic lesions of biliary tract World J Surg 2001 25 1360 1365 11596904 10.1007/s00268-001-0124-2
Archer SB Brown DW Smith CD Branum GD Hunter JG Bile duct injury during laparoscopic cholecystectomy. Results of a national survey Ann Surg 2001 234 549 559 11573048 10.1097/00000658-200110000-00014
Calvete J Sabater L Camps B Verdu A Gomez-Portilla A Martin J Torrico MA Flor B Cassinello N LIedo S Bile duct injury during laparoscopic cholecystectomy. Myth or reality of the learning curve? Surg Endosc 2000 14 608 611 10948294 10.1007/s004640000103
Doctor N Dooley JS Dick R Watkinson A Rolles K Davidson BR Multidisciplinary approach to biliary complications of laparoscopic cholecystectomy Br J Surg 1998 85 627 632 9635808 10.1046/j.1365-2168.1998.00662.x
Flum DR Koepsell T Heagerty P Sinanan M Dellinger EP Common bile duct injury during laparoscopic cholecystectomy and the use of intraoperativ cholangiography. Advers outcome or prevetable error? Arch Surg 2001 136 1287 1292 11695975 10.1001/archsurg.136.11.1287
Johnson SR Koehler A Pennington LK Hanto DW Long-term results of surgical repair of bile duct injuries following laparoscopic cholecystectomy Surgery 2000 128 668 677 11015101 10.1067/msy.2000.108422
Kiviluoto T Siren J Luukkonen P Kivilaasko E Randomised trial of laparoscopic versus open cholecystectomy for acute and gangrenous cholecystitis The Lancet 1998 351 321 325 9652612 10.1016/S0140-6736(97)08447-X
Mirza DF Narsimhan KL Ferraz Neto BH Mayer AD McMaster P Buckels JAC Bile duct injury following laparoscopic cholecystectomy: referral pattern and manegement Br J Surg 1997 84 786 790 9189087 10.1046/j.1365-2168.1997.02666.x
Olsen D Bile duct injuries during laparoscopic cholecystectomy Surg Endosc 1997 11 133 138 9069144 10.1007/s004649900315
Savader SJ Lillemoe KD Prescott CA Winick AB Venbrux AC Lund GB Mitchell SE Cameron JL Osterman FA Laparoscopic cholecystectomy-related bile duct injuries. A healt and financial disaster Ann Surg 1997 225 268 273 9060582 10.1097/00000658-199703000-00005
Targarona EM Marco C Balagué C Rodriguez J Cugat E Hoyuela C Veloso E Trias M How, when, and why bile duct occors. A comparison between open and laparoscopic cholecystectomy Surg Endosc 1998 12 322 326 9543521 10.1007/s004649900662
Tsaalis KG Chritoforidis EC Dimitriadis CA Kalfadis SC Botsios DS Daduokis JD Management of bile duct injury during and after laparoscopic cholecystectomy Surg Endosc 2003 17 31 37 12384766 10.1007/s00464-001-9230-3
Bachellier P Nakano H Weber JC Lemarque P Oussoultzoglou E Candau C Wolf P Jaeck D Surgical repair after bile duct and vascular injuries during laparoscopic cholecystectomy: When and how? World J Surg 2001 25 1335 1345 11596900 10.1007/s00268-001-0120-6
Krahenbühl L Sclabas G Wente MN Schafer M Schlumpf R Büchler MW Incidance, risk factors, and prevention of biliary tract injuries during laparoscopic cholecystectomy in Switzerland World J Surg 2001 25 1325 1330 11596898 10.1007/s00268-001-0118-0
MacFadyen BV Vecchio R Ricardo AE Mathis CR Bile duct injury after laparoscopic cholecystectomy. The United States experience Surg Endosc 1998 12 315 321 9543520 10.1007/s004649900661
Kramling HJ Hüttl TP Heberer G Development of gallstone surgery in Germany Surg Endosc 1999 13 909 913 10449851 10.1007/s004649901132
Regöly-Mérei J Ihasz M Szeberin Z Sandor J Maté M Biliary tract complications in laparoscopic cholecystectomy. A multicenter study of 148 biliary tract injuries in 26.440 operations Surg Endosc 1998 12 294 300 9543516 10.1007/s004649900657
Arul GS Rooney PS Gregson R Steele RJC The Standard of laparoscopic intraoperative cholangiography: A quality control study Endoscopy 1999 31 248 252 10344430 10.1055/s-1999-13677
Robinson TN Stiegmann GV Durham D Johnson SI Wachs ME Sera AD Kumpe DA Management of major bile duct injury associated with laparoscopic cholecystectomy Surg Endosc 2001 15 1381 1385 11965450
Suziki M Akaishi S Rikiyama T Naitoh T Rahman MM Matsunu S Laparoscopic cholecystectomy, Calot's triangle, and variations in cystic arterial supply Surg Endosc 2000 14 141 144 10656947
|
15963227
|
PMC1182383
|
CC BY
|
2021-01-04 16:28:04
|
no
|
BMC Surg. 2005 Jun 17; 5:14
|
utf-8
|
BMC Surg
| 2,005 |
10.1186/1471-2482-5-14
|
oa_comm
|
==== Front
Front ZoolFrontiers in Zoology1742-9994BioMed Central London 1742-9994-2-121598750610.1186/1742-9994-2-12ResearchThe male genital system of the cellar spider Pholcus phalangioides (Fuesslin, 1775) (Pholcidae, Araneae): development of spermatozoa and seminal secretion Michalik Peter [email protected] Gabriele [email protected] Zoologisches Institut und Museum, Ernst-Moritz-Arndt-Universität, J.-S.-Bach-Straße 11/12, D-17489 Greifswald, Germany2 Institut für Zoologie, Universität Bonn, Endenicher Allee 11-13, D-53115 Bonn, Germany2005 29 6 2005 2 12 12 7 3 2005 29 6 2005 Copyright © 2005 Michalik and Uhl; licensee BioMed Central Ltd.2005Michalik and Uhl; 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 arthropods pass through several molting stages (instars) before reaching sexual maturity. In spiders, very little is known about the male genital system, its development and seminal secretions. For example, it is unknown whether spermatozoa exist prior to-, or only after the final molt. Likewise, it is unclear whether sperm are produced throughout male adulthood or only once in a lifetime, as is whether seminal secretions contain factors capable of manipulating female behavior. In order to shed light on these aspects of the reproductive biology of spiders, we investigated the male genital system of the common cellar spider Pholcus phalangioides, with special emphasis on its development and seminal secretions.
Results
Testes already display all stages of spermatogenesis in subadult males (about four weeks before the final molt). Their vasa deferentia possess proximally a very voluminous lumen containing dense seminal fluid and few spermatozoa, whereas the distal part is seemingly devoid of contents. Spermatoza of P. phalangioides are typical cleistospermia with individual secretion sheaths. In male stages approximately two weeks prior to the final molt, the lumina of the testes are wider and filled with a dense secretion. The wide, proximal portion of the vasa deferentia is filled with secretion and a large number of spermatozoa, and the narrow distal part also contains secretion. In adult males, the wide lumina of the testes are packed with spermatozoa and secretions. The latter are produced by the somatic cells that bear microvilli and contain many vesicles. The lumina of the vasa deferentia are narrow and filled with spermatozoa and secretions. We could identify a dense matrix of secretion consisting of mucosubstances and at least three types of secretion droplets, likely consisting of proteinaceous substances.
Conclusion
This study reveals that spermatogenesis begins weeks before maturity and takes place continuously in the long-lived males of P. phalangioides. Possible functions of the various types of secretion in the seminal fluid and previously investigated female secretions are discussed in the light of sexual selection.
==== Body
Background
When and for how long males produce sperm very likely depends on the mating system of the species in question. Long-lived species in which males can expect several matings probably continue to produce sperm throughout their lifetime. On the other hand, in species with a short reproductive time window or in species that show a considerable drop in female receptivity after a single mating, a male is expected to allocate all resources into sperm production for a single mating. For spiders, both evolutionary scenarios are conceivable: in sexually cannibalistic spiders, males load each of their copulatory organs, the pedipalps, only once and cease to produce sperm after the final molt, which seems to be a consequence of the high incidence of monogamy that is forced upon the male by female cannibalistic attacks [[1]; PM personal observation], whereas in most spider species, males can expect more than one mating. In our focal species, Pholcus phalangioides, adult males are especially long-lived, with life-spans of up to a year [GU personal observation]. In this group of spiders sperm should be produced throughout a male's lifetime.
During mating, males not only transfer sperm, but also seminal fluid [2-6]. The seminal fluid of Drosophila melanogaster contains over 80 proteins and peptides. The few substances that have been identified have marked effects on the reproductive success of males and females: seminal fluid proteins and peptides can decrease female receptivity, increase egg production, facilitate sperm storage, and are necessary for sperm transfer and success in sperm competition. Moreover, antimicrobial agents in the secretions ensure that the female reproductive tract is a hospitable environment during sperm transfer and storage. In some species, even noxious chemicals are transferred and incorporated into developing eggs to protect them from predators and pathogens [reviews: [4-7]]. Similar functions have been reported from other insects such as butterflies [8] and crickets [9]. In most insects, males produce secretions in separate accessory glands that can be dissected to characterize their products. However, male spiders have not been reported to possess separate accessory glands that are directly connected to the genital tract. The production of seminal secretion is thus very likely to occur within the testes. However, it is conceivable that the effects of the seminal secretions on the reproductive success of males and females may be similar in spiders and insects despite their different origin.
The shape of the male genital system in spiders differs enormously across spider taxa, but usually consists of two thick strands of testes, continuing in the thin convoluted vasa deferentia, which fuse distally to form the unpaired ductus ejaculatorius that opens into the genital opening located in the epigastric furrow [for entelegyne araneomorph spiders: [10-13]; but see [14] for Theraphosidae]. Generally, the male genital system in spiders is embedded in the midgut gland and the testes often extend deeply into the opisthosoma. Additional multicellular glands present in male spiders are the so-called epiandrous (=epigastric) glands, which seem to provide part of the sperm web and release their acinous substances near the genital opening through special spigots (=fusules) [15-17]. For some spider families unicellular glands that open into the epigastric furrow via individual ducts are described, but it is unknown whether the glandular secretion is added to the sperm mass [17,18]. Unpublished work supports the notion that there is a great diversity of secretions in the male genital tract of spiders, the origin and functions of which are unknown [[10]; PM, GU personal observation].
In the present study, we briefly describe the genital system of the haplogyne spider P. phalangioides. We investigate the organisation and development of the male genital system, the ultrastructure of the different parts, and give a first characterization of the secretions produced in the genital system. Finally, we present some additional findings on spermatogenesis and spermatozoa that complement previous investigations [19-21].
Results
The male genital system of P. phalangioides consists of one pair of massive testes and convoluted vasa deferentia which become thicker near the genital opening and fuse distally to form the ductus ejaculatorius (Fig. 1). The genital tract is located ventrally in the opisthosoma and the testes extend as far as to the spinning apparatus (Figs. 2c, 3c, 4a). Parts of the testes and the vasa deferentia are bordered ventrally by the ampullate silk glands (Figs. 2c, 3c, 4a). The genital system is surrounded by extensions of the midgut gland (Figs. 2c, 3c, 4a).
Figure 1 Schematic drawing of the male genital system of Pholcus phalangioides.
Figure 2 Male genital system of Pholcus phalangioides in subadult stage 1 (about four weeks before final molt). (a-b): Drawings of the palpal organ from medial (a) and lateral (b) view to characterize the age of the observed males. (c): Longitudinal section of the male genital system (anterior = left). The vas deferens and large part of the testis are bordered ventrally by silk glands. The proximal part of the lumen of the vas deferens is very voluminous and filled with some spermatozoa and dense secretion. (d): Longitudinal section in the region of the genital opening (arrow). In this part, the vas deferens has a very narrow lumen without recognizable content. (e): Transition between testis and vas deferens. The connection between these two parts has a valve-like appearance. The vas deferens possesses a thick epithelium and is filled with dense secretion and some spermatozoa. (f): Section of the testis. Note germ cells at different stages of spermatogenesis in cysts. The branched lumen is very narrow and a secretion is not yet recognizable. Mg, midgut gland; Sg, silk gland; Sp, spermatids; Te, testis; Vd, vas deferens.
Figure 3 Male genital system of Pholcus phalangioides in subadult stage 2 (about two weeks before final molt). (a-b): Drawings of the palpal organ from medial (a) and lateral (b) view to characterize the age of the observed males. (c): Longitudinal section of the male genital system. Note the different staining of cells compared with Fig. 2c resulting from a higher secretory activity (see Fig. 3e). (d): Longitudinal section of the vas deferens. The voluminous proximal part is completely filled with spermatozoa and secretion. Towards the distal part, the lumen becomes narrow but is also filled with seminal fluid (arrows). (e): Section of the testis. Between the cysts the narrow lumen contains homogenous secretion characterized by the blue-green color. Sg, silk gland; Sp, spermatids; Te, testis; Vd, vas deferens.
Figure 4 Male genital system of adult Pholcus phalangioides (one day after final molt). (a): Longitudinal section of the male genital system. The convoluted vas deferens possesses a thin tube-like shape. Within the testis the different stages of spermatogenesis, spermatogonia (dark) and later stages of spermatogenesis near the center (bright) are clearly visible. (b): Transition between testis and vas deferens. The thin vas deferens is filled with spermatozoa and different kinds of secretion (c.f. Fig. 6). (c): Section of the testis. The lumen is very wide and contains different secretions. Note the secretion droplets (asterisk) and the different cysts of spermatids. (d): Near the genital opening the vasa deferentia fuse to form the ductus ejaculatorius which accumulates large amounts of seminal fluid. (e): Longitudinal section through the genital opening, bordered by the cuticle. The epiandrous apparatus is located in front of the genital opening. Note the epiandrous gland and its secretion within the spigot. De, ductus ejaculatorius; Eg, epiandrous gland; Esp, epiandrous spigot; Go, genital opening; Mg, midgut gland; Sg, silk gland; Sp, spermatids; Te, testis; Vd, vas deferens.
Developmental Stages of the Male Genital System
Stage 1 (about four weeks before the final molt)
These young males are characterized by palpal organs that are bent in the joint of the tibia and patella (Figs. 2a,b). Dark sclerotized areas are present on the femur and tibia. The tarsus has a broad shape without any signs of internal structures or appendages as present in adult males (Figs. 2a,b).
The longitudinal section through the opisthosoma (Fig. 2c) shows that the dimensions of the testis are similar to the final dimensions in adult males (compare Fig. 4a), and all stages of spermatogenesis are observable (Fig. 2f). Spermatogenesis occurs in cysts containing spermatids of the same developmental stage (Fig. 2f, see also Figs. 3e, 4c). The secretory activity within the testis is very low as indicated by the absence of the dark green secretion visible in subadult stage 2 and adults (see below).
The most distal part of the testis becomes thinner and opens into the vas deferens, forming a valve-like structure (Fig. 2e). The proximal part of the vas deferens possesses a thick epithelium and has an extensive lumen that is filled with spermatozoa embedded in a bright green secretion (Figs. 2c,e). In the more distal parts the lumen, the vas deferens is very narrow. It contains neither spermatozoa nor any other recognizable substance (Fig. 2d).
Stage 2 (about two weeks before the final molt)
In this stage the palpal organs are only slightly bent. The tarsus is extended and the internal structures known from the palps of adult males are visible through the cuticle (Figs. 3a,b). Dark sclerotized areas are present on the distal and lateral side of the tarsus (Figs. 3a,b).
The section in Figure 3c displays a similar organization as that in subadult males of stage 1. The testis contains all stages of spermatogenesis (Fig. 3e). Lumina with dark green secretion and several spots with red, roundish secretion are present between the cysts of spermatid (Fig. 3e, arrows). This secretion is produced by the somatic cells of the testis (see below).
The extensive proximal part of the vas deferens is almost completely filled with spermatozoa and secretions (Figs. 3c,d). Distally, the vas deferens becomes thinner and is shaped as a convoluted tube (Figs. 3c,d). Within the lumen a similar material as in the proximal part is present (Fig. 3d, arrows).
Adult Stage
The fully developed male genital system is also characterized by tube-like testes and thin convoluted vasa deferentia (Figs. 1, 4a). Within the testes all stages of spermatogenesis are visible (Figs. 4a,b). The spermatozoa are coiled, which is typical of spiders (Fig. 5e). In comparison to the subadult stages, the lumen of the adult testis seems wider and is filled with a bright green secretion matrix and red, roundish secretion droplets of different sizes (Fig. 4c).
Figure 5 Ultrastructure of spermatozoa and testis of Pholcus phalangioides. (a): Early spermatid. The nucleus is surrounded by a manchette of microtubules with a peculiar regular pattern. The nuclear envelope possesses a wave-like outline. The cytoplasm in front of the acrosomal vacuole at the anterior pole of the nucleus is indented. Note the annulate lamellae near the nucleus. (b): Late spermatids. The chromatin of the nucleus is almost completely condensed. The nucleus possesses a very elongated shape with a conspicuous helical band whirling around it (arrows). (c): Detail of the nucleus. The helical band of the nucleus is very thin and bordered by microtubules (arrows). Between the band, the dense manchette of microtubules is conspicuous (arrowheads). (d): Scanning electron micrograph of a late spermatid. Note the extensive cytoplasm drop at the posterior end of the spermatid. The helical character of the nucleus is clearly visible. (e): Lumen of the testis with coiled spermatids. The coiling occurs at the end of spermiogenesis within the testis. Between the coiled spermatids several secretion droplets are visible. (f): The somatic cells possess many different vesicles and are bordered apically by microvilli. They continuously produce the large secretion droplets (arrows). AF, acrosomal filament; AV, acrosomal vacuole; Ax, axoneme; Cd, cytoplasm droplet; Fl, flagellum; Lu, lumen; Mi, mitochondria; Mv, microvilli; N, nucleus; peN, postcentriolar elongation of the nucleus; SC, somatic cell; Sec, secretion droplet; Sp, spermatids.
In contrast to subadult males, the vasa deferentia possess a narrow lumen filled with spermatozoa and different secretions over their entire length (Figs. 4b,d). Here, the secretion matrix has a bright green color similar to that of the testis (Figs. 4b,d). Furthermore, there are also different red, roundish secretions of varying sizes (Fig. 4d, compare also Fig. 6a).
Figure 6 Ultrastructure of the vas deferens of Pholcus phalangioides. (a): The spermatozoa are embedded in a dense secretion matrix. Between the spermatozoa three kinds of secretion droplets are visible. The secretion droplet type 3 is characterized by an irregular surface which sometimes includes other particles (arrows). (b): The epithelium of the vas deferens is very flat and bears microvilli. It is basally bordered by a thin muscle layer. Note the thick secretion sheath which surrounds each spermatozoon. (c): Scanning electron micrograph of the seminal fluid. Note the irregular surface of the secretion droplet type 3 and the differences in size of spermatozoa and secretions. Lu, lumen; Mu, muscles; Mv, microvilli; Sec1–3, secretions of droplet type 1–3; Sp, spermatozoon; SSh, secretion sheath.
Close to the genital opening the thin vasa deferentia fuse and form the ductus ejaculatorius, which is characterized by an extensive lumen with large amounts of spermatozoa and secretions (Fig. 4d). The epithelium near the genital opening and towards the ductus ejaculatorius is prismatic and apically bordered by cuticle (Fig. 4e).
Near the ductus ejaculatorius, epiandrous glands are present which open into small spigots anterior to the genital opening (Fig. 4e). These glands are characterized by large lumina and a prismatic epithelium (Fig. 4e). The secretion of the glands has a granular green appearance as seen in the duct of the epiandrous spigots (Fig. 4e). The lumen of the gland appears empty, though this may be an artifact.
Spermatogenesis and Ultrastructure of the Male Genital System
The spermatozoa of P. phalangioides are characterized by several unique features, as pointed out in detail by Alberti and Weinmann [20]. Apart from additional details, we provide information on the relationship between sperm and secretions in the male genital system.
Early stages of spermatogenesis are characterized by a large spherical nucleus surrounded by a manchette of microtubules, an acrosomal vacuole on its anterior pole and an axonemal basis, which migrates into an indentation at the posterior pole of the nucleus. In this early stage, the manchette of microtubules and the nuclear surface possesses a unique arrangement. As seen in cross-sections, the microtubules are not densely packed and the nuclear envelope shows a slightly wave-like outline (Fig. 5a). The quadrangular acrosomal vacuole is tightly connected to the cell membrane, which is indented in this region (Fig. 5a). Within the cytoplasm annulate lamellae are present (Fig. 5a).
During spermiogenesis the chromatin condenses and the nucleus strongly elongates (Fig. 5b). A thin helical band of nuclear material covered by microtubules surrounds the nucleus over its entire length (Figs. 5b,c). The acrosomal vacuole elongates during spermiogenesis, and finally possesses a cylindrical shape (Fig. 5d). A considerable amount of cytoplasm accumulates at the posterior end of the spermatids, forming a big droplet (Figs. 5b,d).
Within the lumen of the testis, only coiled spermatozoa are found embedded in a secretion matrix, which is likely produced by the extensive somatic cells (Figs. 5e,f). The extensions of the somatic cells surround the cysts of spermatids. Apically, the somatic cells bear long microvilli and contain many vesicles, indicative of strong secretory activity (Fig. 5f). Several secretion droplets are visible between the spermatozoa (Figs. 5e,f). The large secretion droplets seem to be produced through fusion of smaller droplets, as indicated by the irregular ring of loose material around the electron-dense center of the droplet (Fig. 5f, arrows).
Within the vas deferens the spermatozoa are densely packed and embedded in a homogenous secretion matrix (Fig. 6a). Each single spermatozoon is surrounded by a secretion sheath forming a so-called cleistosperm (Figs. 6a,b). The secretion sheath is produced within the vas deferens since it is absent in the testis (Fig. 5e). Within the secretion matrix three different kinds of secretions are present (Fig. 6). The first type of secretion consists of small droplets which are densely distributed and characterized by a bright center and a dark border. The second kind is a large electron-dense secretion droplet with a very irregular shape. These droplets often partially surround the spermatozoa (Fig. 6a). The third and rarest type of secretion is a droplet which appears less electron-dense. It is characterized by a heterogeneous content and an irregular surface to which particles may adhere or are partly embedded (Fig. 6c). The epithelium of the vas deferens is flat and bears microvilli in its apical region. A muscle layer surrounds the vas deferens (Fig. 6b).
Discussion
General Organization of the Male Genital System
The male genital system of P. phalangioides is characterized by thick tube-like testes and thin convoluted vasa deferentia which fuse distally to form the ductus ejaculatorius. This organization is reported from many species of different families, e.g., Cybaeidae, Theridiidae and Agelenidae, and seems to be the general condition in araneomorph spiders (see background). However, theraphosid spiders show no distinct separation between testis and vas deferens [[14]; PM personal observation] and in theraphosid as well as mesothelid spiders the testes themselves are convoluted [[14,22]; PM personal observation]. In some species, e.g., the cybaeid Argyroneta aquatica, the testes are curved but distinct from the vasa deferentia [11]. Petrunkevitch [23] suggested that the general organization of the male genital system is not useful for phylogenetic consideration, but the few data available indicate a potential for systematic interpretation. For example, in all theridiid spiders studied thus far, the vasa deferentia lead into a roundish vesicula seminalis (=ductus ejaculatorius?), the storage site of the mature spermatozoa, which is connected via an unpaired duct with the genital opening [[13]; PM personal observation]. Considering the information available to date, it is at least misleading to take the theraphosid type of male genital system as representative for spiders [24,25].
Development of the Male Genital System
In P. phalangioides, spermiogenesis starts several weeks before the last molt and continues in the adults. In the subadult males approximately four weeks before the final molt, the lumina of the testes are very thin and only a small amount of the dense secretion matrix is observable. The secretory activity increases within the last two weeks of the subadult stage. In adults, the lumina are wide and full of secretion and mature spermatozoa. The two different functions of the testis – the production of sperm cells and different kinds of secretion (see below), are evident. Several authors suggested that these secretions are products of degenerated spermatids and the cytoplasmic droplet discarded at the end of spermiogenesis [20,26,27]. However, rough endoplasmic reticulum, Golgi bodies and vesicles in the somatic cells demonstrate that the epithelium itself produces the secretions as was suggested by Alberti et al. [28].
In the subadult stages, the vasa deferentia possess a very extensive lumen in which the mature spermatozoa and secretion from the testes accumulate continuously. The thick epithelium of the vasa deferentia also shows secretory activity. It may be responsible for the formation of the secretion sheath surrounding each spermatozoon, as suggested for other spider species [20,28-32]. In adults, several kinds of secretion are present in the vasa deferentia, whereas in young subadult stages only a dense secretion matrix is present. Thus, it is evident that secretory activity increases in subadults when approaching the final molt. Furthermore, in adult male spiders the epithelium of the vasa deferentia is very thin and secretory activity is low as was stated for the cybaeid Argyroneta aquatica [11].
In adult males of P. phalangioides cleistospermia and secretions are transferred into the palpal copulatory organ soon after the final molt. Sperm uptake by the male can occur repeatedly after each mating and may require a continuous production of sperm cells. In fact, P. phalangioides males mate several times [33] and refill their pedipalp sperm stores after each mating [GU unpublished observation]. To do so successfully, cellar spider males begin sperm production in the subadult phase and continue production throughout adulthood, as demonstrated by our study. In contrast, studies on short lived spider males with only a single mating opportunity showed that males do not refill their sperm stores and probably cease to produce sperm after the final molt [1], instead using their resources for rigorous mate selection [Pauly A, Uhl G, Schneider JM unpublished data].
Secretory Products
Three different kinds of secretion droplets could be identified in the dense secretion matrices of the testes and vasa deferentia of P. phalangioides. According to Romeis [36], the Goldner staining method results in a green coloration of acidic mucosubstances and in red coloration of proteinaceous substances. It cannot be ruled out that the seemingly different kinds of secretion represent different steps of a single secretory pathway. However, since we also found different secretion droplets within the palpal organ, we conclude that different secretory pathways result in a variety of secretions with potentially different functions. Morphologically, the secretions are clearly different from those found in two other spider species, a mesothelid spider [30] and the related pholcid spider Holocnemus pluchei [34]. The obvious difference in seminal secretory products in closely related species suggests that there is rapid divergence of seminal fluid substances that may originate from the process of sexual selection. It has been assumed that the rapid evolution of reproductive proteins is a motor in the speciation process [35].
In spiders, sperm and the seminal secretions are not directly transferred into the female genital tract, but taken up into the palpal organ before mating. Within the pedipalps, glandular epithelia adjacent to the sperm storage site seem to be widespread and also occur in P. phalangioides and other spider species [[37,38]; Rose W, unpublished data]. It has been assumed that their products serve to release the sperm into the female genital tract during copulation or to serve as a protective matrix for sperm storage within the male palp [e.g., [37,39,40]]. During mating, sperm are transferred to the storage site within the female together with the secretions originating from both genital tract and pedipalp. Finally, inside the female cellar spider, sperm and male secretions encounter secretion that is produced by the female. The only non-morphological investigation existing to date identified proteinaceous substances and glyco- and lipoprotein components in the sperm storage site of virgin female P. phalangioides using gel-electrophoretic methods [41].
For spiders, a nutritive and protective function (desiccation and pathogens) has mainly been attributed to either male or female secretions since the sperm cells are often stored for extended periods of time [e.g., [41-45]], but secretions may also serve as lubricants and promote physical uptake or release of sperm inside the male palp (see above). Female secretory products have been suggested to control sperm activation [18,46-48]. However, male secretions may also play an important role in this process.
Spermatozoa and Spermatogenesis
It has been known since the study of Alberti and Weinmann [20] that P. phalangioides possesses very aberrant spermatozoa with several unique characters (e.g., an elongated proximal centriole and a central nuclear canal containing the acrosomal filament), but the most conspicuous character is the band of nuclear material spirally surrounding the nucleus. As shown in the present study, the first signs of this peculiar transformation of the helical band are present in very early spermatid stages: microtubules are loosely arranged around the nucleus resulting in a wave-like outline of the nuclear envelope. The functional implications of the band are still unknown. It may influence the mobility of the spermatozoa or play a role during fertilization. The phylogenetic value of the peculiar spermatozoa of P. phalangioides remains uncertain as well, as the sperm ultrastructure of only one further pholcid spider, H. pluchei, is known [34,49]. The spermatozoa of these two species however do markedly differ in accord with the present ideas on systematic relationships within the Pholcidae [50]. On the other hand, a helical band of nuclear material is also present in Spermophora senoculata [PM personal observation], a species closely related to P. phalangioides [50]. It therefore seems that sperm characteristics may be successfully applied successfully in phylogenetic considerations.
Conclusion
Our study demonstrates that sperm production in males of a long-lived spider species begins weeks before the final molt and continues throughout adulthood. Within the male genital tract, there are at least three different types of proteinaceous secretion droplets apart from a basic matrix secretion containing acidic mucosubstances. The different secretions may serve functions similar to those found in insects, such as oviposition stimulants or manipulators of female receptivity, and clearly warrant future investigation. Moreover, since secretory glands of the female are known to discharge their products into the sperm storage site, their effect on sperm storage, activation and sperm competition should also be investigated. Further, both male and female secretions may interact in as yet unexpected ways. If males have evolved manipulative substances that are against female interests, or if they offer a platform for cryptic female choice, we expect a similar divergence in female secretions and in their interactions with male derived substances.
Methods
Light Microscopy (LM)
Males of P. phalangioides were collected in houses and cellars within the city of Freiburg i.Br., Germany at an early juvenile stage and reared in the laboratory in individual containers on a diet of fruitflies. During the last intermolt interval, males were classified into two developmental stages according to the external morphology of their developing palpal organs. Young subadult males (stage 1, N = 5) were fixed approximately four weeks before the final molt and older subadult males (stage 2, N = 5) were fixed about two week before the final molt. Adult males (N = 5) were examined one to two days after their final molt.
Male specimens were fixed in Bouin solution according to Gregory [51], dehydrated in a graded series of ethanol or isopropanol and embedded in Paraplast. Sections (7 μm) were made with a Reichert 1150 Autocut microtome, stained according to Goldner [52], and mounted with Eukitt medium. Examination was performed with an Olympus BX 60 and pictures were taken with an Olympus DP 10.
Transmission Electron Microscopy (TEM)
Elder male specimens (N = 2) were dissected in 0.1 M phosphate buffer supplemented with 1.8% sucrose. The isolated genital system was fixed in 2.5% glutaraldehyde in the same buffer followed by postfixation in buffered 2% osmium tetroxide. After washing, the tissue pieces were dehydrated in graded ethanols and embedded in Spurr's resin [53]. Ultrathin sections (50 nm) were made with a Leica ultramicrotome and stained with uranyl acetate and lead citrate [54]. Examination was performed with a Zeiss EM 10A electron microscope.
Scanning Electron Microscopy (SEM)
The isolated genital system of elder males (N = 3) was split open in a droplet of phosphate buffer (see above) with thin needles on glass coverslips covered with 1% poly-L-lysin. After 10 min sedimentation the adhering material was fixed in 2.5% glutaraldehyde buffer for 1 h at 4°C. Samples were then rinsed in buffer and post-fixed in buffered 1% osmium tetroxide, dehydrated in ethanol, dried in a Balzers CPD 030 critical point dryer, coated with gold in a Balzers MED 010 sputtering device and examined in a Philips XL20 scanning electron microscope.
Authors' contributions
PM performed the ultrastructural studies and prepared the manuscript. GU performed the histological studies and helped prepare the manuscript. Both authors read and approved the final manuscript.
Acknowledgements
We are indebted to Gerd Alberti and Michael Schmitt for useful comments on the manuscript. For the help with the SEM, the authors are grateful to David Mercati, Fabiola Giusti and Eugenio Paccagnini from the lab of Romano Dallai (University of Siena, Italy). PM acknowledges a grant from the German National Merit Foundation and from the VIGONI-program (German Academic Exchange Service, DAAD). Furthermore, the financial support of the German Research Council (DFG, Al 138/11-1) is greatly appreciated. The histological investigations were performed by GU during her doctoral studies at the University of Freiburg, supported by the German National Merit Foundation. GU thanks Peter Weygoldt, Claudia Gack and Gerhard von der Emde for support.
==== Refs
Herberstein ME Gaskett AC Schneider JM Vella NGF Elgar MA Limits to male copulation frequency: sexual cannibalism and sterility in St Andrew's Cross spiders (Araneae, Araneidae) Ethology
Leopold R The role of male accessory glands in insect reproduction Ann Rev Entomol 1976 21 199 221 10.1146/annurev.en.21.010176.001215
Marchini D Del Bene G Cappelli L Dallai R Ultrastructure of the male reproductive accessory glands in the medfly Ceratitis capitata (Diptera: Tephritidae) and preliminary characterization of their secretions Arthropod Struct Dev 2003 31 313 327 10.1016/S1467-8039(03)00003-3
Gillot C Male accessory gland secretions: Modulators of female reproductive physiology and behavior Ann Rev Entomol 2003 48 163 184 12208817 10.1146/annurev.ento.48.091801.112657
Wolfner MF Tokens of love: functions and regulation of Drosophila male accessory gland products Insect Biochem Mol Biol 1997 27 179 192 9090115 10.1016/S0965-1748(96)00084-7
Wolfner MF The gifts that keep on giving: physiological functions and evolutionary dynamics of male seminal proteins in Drosophila Heredity 2002 88 85 93 11932766 10.1038/sj.hdy.6800017
Chapman T Davies SJ Functions and analysis of the seminal fluid proteins of male Drosophila melanogaster fruit flies Peptides 2004 25 1477 1490 15374649 10.1016/j.peptides.2003.10.023
Jin ZY Gong H Male accessory gland derived factors can stimulate oogenesis and enhance oviposition in Helicoverpa armigera (Lepidoptera: Noctuidae) Arch Insect Biochem Physiol 2001 46 175 185 11304751 10.1002/arch.1027
Wagner WE Harper CJ Female life span and fertility are increased by the ejaculates of preferred males Evolution 2003 57 2054 2066 14575327
Bertkau P Ueber den Generationsapparat der Araneiden Arch Naturgesch 1875 41 235 262
Crome W Die grobe Morphologie des männlichen Genitalapparates einiger Araneen Dtsch Zool Z 1951 1 169 186
Kim JK Kim TH Moon MJ Ultrastructure of the testis in the spider, Pardosa astrigera L. Koch Korean Arachnol 1993 9 43 53
Knoflach B Mating in Theridion varians Hahn and related species (Araneae: Theridiidae) J Nat Hist 1998 32 545 604
Melchers M Zur Biologie der Vogelspinnen (Fam. Aviculariidae) Z Morph Oekol Tiere 1964 52 517 536 10.1007/BF00408073
Marples BJ The spinnerets and epiandrous glands of spiders J Linn Soc Lond 1967 46 209 223
Lopez A Emerit M New data on the epigastric apparatus of male spiders Bull Br arachnol Soc 1988 7 220 224
Lopez A Glandes épigastrique et classification des araneides Bull Ass Anat Paris 1974 58 899 904
Lopez A Nentwig W Glandular aspects of sexual biology Ecophysiology of spiders 1987 Berlin: Springer 121 132
Rosati F Baccetti B Dallai R Baccetti B The spermatozoon of Arthropoda. X. Araneids and the lowest Myriapods Comparative Spermatology 1970 New York, London: Academic Press 247 254
Alberti G Weinmann C Fine structure of spermatozoa of some labidognath spiders (Filistatidae, Segestriidae, Dysderidae, Oonopidae, Scytodidae, Pholcidae; Araneae; Arachnida) with remarks on spermiogenesis J Morphol 1985 185 1 35 10.1002/jmor.1051850102
Dallai R Afzelius BA Witalinski W The axoneme of the spider spermatozoon Boll Zool 1995 62 335 338
Haupt J The Mesothelae – a monograph of an exceptional group of spiders (Araneae: Mesothelae) Zoologica 2003 154 1 102
Petrunkevitch A An inquiry into the natural classification of spiders, based on a study of their internal anatomy T Connecticut Acad Arts Sci 1933 31 299 389
Foelix RF Biology of Spiders 1996 New York: Oxford University Press
Felgenhauer BE Harrison FW, Foelix RF Araneae Microscopic anatomy of invertebrates 1999 8A New York: Wiley-Liss 223 266
Wagner J Einige Beobachtungen über die Spermatogenese bei den Spinnen Zool Anz 1896 19 188 190
Bösenberg H Beiträge zur Kenntnis der Spermatogenese bei den Arachnoiden Zool Jb Anat 1905 21 515 570
Alberti G Afzelius BA Lucas SM Ultrastructure of spermatozoa and spermatogenesis in bird spiders (Theraphosidae, Mygalomorphae, Araneae) J Submicrosc Cytol 1986 18 739 753
Michalik P Gray MR Alberti G Ultrastructural observations of spermatozoa and spermiogenesis in Wandella orana Gray, 1994 (Araneae: Filistatidae) with notes on their phylogenetic implications Tissue Cell 2003 35 325 337 14517100 10.1016/S0040-8166(03)00054-5
Michalik P Haupt J Alberti G On the occurrence of coenospermia in mesothelid spiders (Araneae: Heptathelidae) Arthropod Struct Dev 2004 33 173 181 10.1016/j.asd.2004.02.001
Michalik P Dallai R Giusti F Alberti G The ultrastructure of the peculiar synspermia of some Dysderidae (Araneae, Arachnida) Tissue Cell 2004 36 447 460 15533460 10.1016/j.tice.2004.08.001
Michalik P Sacher P Alberti G Ultrastructural observations of spermatozoa of some tetragnathid spiders and their phylogenetic implications (Araneae, Tetragnathidae) J Morphol 2005
Uhl G Mating behaviour and female sperm storage in Pholcus phalangioides (Fuesslin) (Araneae) Memoirs of the Queensland Museum 1993 33 667 674
Michalik P Dallai R Giusti F Mercati D Alberti G The spermatozoa and spermiogenesis of Holocnemus pluchei (Scopoli, 1763) (Pholcidae, Araneae)
Swanson WJ Vacquier VD Reproductive protein evolution Ann Rev Ecol Syst 2002 33 161 179 10.1146/annurev.ecolsys.33.010802.150439
Romeis B Mikroskopische Technik 1989 München: Urban und Schwarzenberg
Lopez A Juberthie-Jupeau L Ultrastructure compare du tube séminifére chez les males d'araignées Mém Biospéol 1985 12 97 109
Huber BA Evolutionary transformation from muscular to hydraulic movements in spider (Arachnida, Araneae) genitalia: a study based on histological serial sections J Morphol 2004 261 364 376 15281063 10.1002/jmor.10255
Kraus O Male Spider Genitalia: Evolutionary Changes in Structure and Function Verh Naturwissenschaftl Ver Hamburg 1984 27 373 382
Suhm M Thaler K Alberti G Glands in the male palpal organ and the origin of the mating plug in Amaurobius species (Araneae: Amaurobiidae) Zool Anz 1996 234 191 199
Uhl G Sperm storage secretion of female cellar spiders (Pholcus phalangioides: Araneae): Agel-electrophoretic analysis J Zool 1996 240 153 161
Engelhardt V Beiträge zur Kenntnis der weiblichen Kopulationsorgane einiger Spinnen Z Wiss Zool 1910 96 32 117
Coyle FA Harrison FW McGimsey WC Palmer JM Observation on the structure and function of spermathecae in haplogyne spiders Trans Am Microsc Soc 1983 102 272 280
Uhl G Sperm storage and repeated egg production in female Pholcus phalangioides Fuesslin (Araneae) Bull Soc Neuchâteloise Sci Nat 1993 116 245 252
Michalik P Reiher W Suhm-Tintelnot M Coyle FA Alberti G The female genital system of the folding-trapdoor spider Antrodiaetus unicolor (Hentz, 1842) (Antrodiaetidae, Araneae) – an ultrastructural study of form and function with notes on reproductive biology of spiders J Morphol 2005 263 284 309 15672391 10.1002/jmor.10309
Uhl G Genital morphology and sperm storage in Pholcus phalangioides (Fuesslin, 1775) (Pholcidae; Araneae) Acta Zool Stockholm 1994 75 1 12
Uhl G Ultrastructure of the accessory glands in female genitalia of Pholcus phalangioides (Fuesslin, 1771) (Pholcidae; Araneae) Acta Zool Stockholm 1994 75 13 25
Berendonck B Greven H Genital structures in the entelegyne widow spider Latrodectus revivensis (Arachnida; Araneae; Theridiidae) indicate a low ability for cryptic female choice by sperm manipulation J Morphol 2005 263 118 132 15562503 10.1002/jmor.10296
Lopez A Boissin L La spermatide d'Holocnemus pluchei (Scop.) (Arachnida, Araneida, Pholcidae): Étude ultrastructurale Bull Soc Zool France 1976 101 423 431
Huber BA High species diversity in one of the dominant groups of spiders in east african montane forests (Araneae: Pholcidae: Buitinga n. sp., Spermophora Hentz) Zool J Linn Soc 2003 137 555 619 10.1046/j.1096-3642.2003.00053.x
Gregory RE Alcoholic Bouin fixation of insect nervous systems for Bodian silver staining. II. Modified solutions Stain Technol 1980 55 151 160 6161444
Goldner J A modification of the Masson trichrome technique for routine laboratory purposes Am J Pathol 1938 14 247 249
Spurr AR A low-viscosity epoxy resin embedding medium for electron microscopy J Ultra Mol Struct Res 1969 26 31 43
Reynolds ES The use of lead citrate at high pH as an electron-opaque stain in electron microscopy J Cell Biol 1963 17 208 13986422 10.1083/jcb.17.1.208
|
15987506
|
PMC1182384
|
CC BY
|
2021-01-04 16:38:34
|
no
|
Front Zool. 2005 Jun 29; 2:12
|
utf-8
|
Front Zool
| 2,005 |
10.1186/1742-9994-2-12
|
oa_comm
|
==== Front
Genet Vaccines TherGenetic Vaccines and Therapy1479-0556BioMed Central London 1479-0556-3-41596975410.1186/1479-0556-3-4ResearchImmediate transfection of patient-derived leukemia: a novel source for generating cell-based vaccines Gershan Jill A [email protected] Bryon D [email protected] James [email protected] Dennis W [email protected] Natalia [email protected] Stephanie [email protected] Karen [email protected] Kelly W [email protected] Anne B [email protected] Rimas J [email protected] Department of Pediatrics, Medical College of Wisconsin and the Children's Research Institute, Children's Hospital of Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226, USA2005 21 6 2005 3 4 4 26 3 2005 21 6 2005 Copyright © 2005 Gershan et al; licensee BioMed Central Ltd.2005Gershan 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 production of cell-based cancer vaccines by gene vectors encoding proteins that stimulate the immune system has advanced rapidly in model systems. We sought to develop non-viral transfection methods that could transform patient tumor cells into cancer vaccines, paving the way for rapid production of autologous cell-based vaccines.
Methods
As the extended culture and expansion of most patient tumor cells is not possible, we sought to first evaluate a new technology that combines electroporation and chemical transfection in order to determine if plasmid-based gene vectors could be instantaneously delivered to the nucleus, and to determine if gene expression was possible in a cell-cycle independent manner. We tested cultured cell lines, a primary murine tumor, and primary human leukemia cells from diagnostic work-up for transgene expression, using both RFP and CD137L expression vectors.
Results
Combined electroporation-transfection directly delivered plasmid DNA to the nucleus of transfected cells, as demonstrated by confocal microscopy and real-time PCR analysis of isolated nuclei. Expression of protein from plasmid vectors could be detected as early as two hours post transfection. However, the kinetics of gene expression from plasmid-based vectors in tumor cell lines indicated that optimal gene expression was still dependent on cell division. We then tested to see if pediatric acute lymphocytic leukemia (ALL) would also display the rapid gene expression kinetics of tumor cells lines, determining gene expression 24 hours after transfection. Six of 12 specimens showed greater than 17% transgene expression, and all samples showed at least some transgene expression.
Conclusion
Given that transgene expression could be detected in a majority of primary tumor samples analyzed within hours, direct electroporation-based transfection of primary leukemia holds the potential to generate patient-specific cancer vaccines. Plasmid-based gene therapy represents a simple means to generate cell-based cancer vaccines and does not require the extensive infrastructure of a virus-based vector system.
==== Body
Background
The efficacy of cell-based tumor vaccines in murine models of malignancy is well established. Using tumor cells lines transfected with soluble immune stimulatory molecules such as IL-2 or IL-12, or cell surface co-stimulatory antigens including CD80, and CD137L, or even allogeneic MHC results in profound immune activation [1-5]. The advantage of working in model systems is that unlimited amounts of tumor are available to produce cell-based vaccines. The ability to produce cell-based vaccines from clinic-derived material, however, remains a challenge.
Cell-based vaccines from tumor-derived material have been prepared and administered in either an allogeneic or autologous fashion, recently reviewed by Mocellin, et al. [1]. An allogeneic vaccine usually features the expansion of a single tumor cell line that can grow well in culture, genetic transduction by the desired vector, and preparation of large vaccine stocks that can be qualified for clinical use. A vaccine for neuroblastoma featuring the expression of both a cytokine and a chemokine transgene (IL-2 and lymphotactin) by a single human neuroblastoma cell line is a recent example of this strategy [7]. The disadvantage of a single cell line approach is that each malignancy is to some degree unique, and perhaps the most immunogenic antigens, or the most relevant ones for a given patient, will fail to be expressed by the allogeneic vaccine.
Give these limitations, we propose that a cell-based vaccine could be produced in an autologous manner for patients with a high disease burden, such as those who present with significant bone marrow involvement. For example, the large amount of tumor material typically available from leukemia patients makes these cells accessible for autologous patient-derived vaccine production.
A major hurdle to be overcome in using primary cells is the need to culture tumor cells in vitro in order for transduction or transfection procedures to be carried out. Most malignancies will not survive in culture in large enough numbers to be utilized. However, if the time required for in vitro manipulation was minimized, for example to 8–24 hours, patient-derived leukemia cells could be isolated from blood or bone marrow, transfected, and then upon irradiation used as a cell-based vaccine. Here we report the application of a novel electroporation-based transfection methodology that holds the potential to immediately transform a patient tumor sample into a cell-based cancer vaccine. This process, termed nucleofection, was pursued in our laboratory because it is the most rapid method of gene vector introduction available. We demonstrate that even though delivery of a plasmid gene vector to the nucleus is immediate, short-term culture is still required, and that a single-round of cell division may be needed to reach optimal gene expression levels. Importantly the time for tumor vaccine preparation is now measured in terms of hours instead of days. Our findings confirm studies carried out by Schakowski et al., where 3 samples from acute myeloid leukemia (AML) patients were transfected with a GFP expression vector [8]. The large degree of transgene expression in the majority of patient-derived acute lymphoblastic leukemia (ALL) specimens that we present here suggests that a clinical trial using these procedures should be pursued.
Methods
Cell lines
The mouse neuroblastoma cell line AGN2a, an aggressive subclone of Neuro-2a, was cultured in Dulbecco's modified Eagle's medium (DMEM), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine and 10% heat inactivated fetal bovine serum (FBS), 1 mM MEM sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, 0.01 M HEPES buffer, 2 mM L-glutamine, 0.05 M 2-mercaptoethanol, and 0.069 M L-arginine HCl [4]. Primary murine tumor was generated by subcutaneous injection of 1 × 106 cultured AGN2a cells into strain A/J mice (Jackson Labs, Bar Harbor, ME). The human osteosarcoma cell line U2OS, kindly provided by Dr. Kent Wilcox, Medical College of Wisconsin, and the mouse squamous cell carcinoma cell line SCCVII, kindly provided by Dr. Scott Strome, Mayo Clinic, Rochester, MN, were cultured in DMEM as above. Mouse primary tumors were processed into single-cell suspensions by injection of 1–2 ml of 1 mg/ml collagenase D into the excised tumor mass (1 mg/ml in 10 mM HEPES, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2) and incubated at 37°C for 45 min followed by mechanical disruption through a sterile screen. Tumor cells were washed in DMEM and PBS and viable cells were separated by centrifugation over a Ficoll-Paque™ (Amersham Biosciences, Piscataway, NJ) density gradient.
Transfection of tumor cell lines
Tumor cells were transfected with either pcDNA3.1/Hygro(-) (Invitrogen, Carlsbad, CA) or pDSRed2-C1 (BD Biosciences, San Diego, CA) plasmid vectors using a cationic lipid-based transfection methodology (Novafection, VennNova, Inc., Pompano Beach, FL) or a proprietary electroporation method (Nucleofection, Amaxa, Köln, Germany). Cells were nucleofected with 0.5 μg plasmid per 106 cells or lipid transfected with 0.5 μg plasmid and 2 μg of NovaFECTOR reagent per 106 cells. Similarly, U2OS, SCCVII and AGN2a cells were nucleofected with 0.5 μg per 106 cells pDSRed2-C1. To determine expression levels, cells were stained with 7AAD (BD Biosciences) and the expression of red fluorescent protein (RFP) in live gated cells was analyzed by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ) at designated time points. U2OS and SCCVII cells were also nucleofected with pCI-neo (Promega, Madison, WI) encoding CD137L (4-1BBL) cDNA at a concentration of 1.5 μg per 106 cells [5]. The CD137L transfected cells were stained with phycoerythrin (PE) conjugated mouse anti-human CDw137 Ligand (BD Biosciences Pharmingen, San Diego, CA).
Patient Samples
Patient leukemia and lymphoma samples were obtained in accordance to the Helsinki Declaration from excess bone marrow or peripheral blood specimens submitted to the Cell Marker Lab, Children's Hospital of Wisconsin, for routine leukemia screening. Studies involving these samples were approved by the Medical College of Wisconsin and Children's Hospital of Wisconsin Institutional Review Boards. Informed consent was obtained from the parents or guardians of each child and each sample was assigned a unique identifier number to ensure confidentiality.
Transfection of primary acute lymphocytic leukemia cells
Leukocytes from bone marrow or peripheral blood patient samples were separated by centrifugation over a Ficoll-Paque™ density gradient. Cells were nucleofected with 1 μg pDSRed2C-1 (red fluorescent protein, RFP, expression vector) plasmid per 106 cells using a variety of Amaxa solutions and program parameters, cultured in RPMI-1640, 100 U/ml penicillin, 100 μg/ml streptomycin and 10% heat inactivated FBS for 24 hours then analyzed for RFP expression by flow cytometry (FACScan, Becton Dickinson). FACS acquisition and analysis was done using either propidium iodide (PI) or 7AAD to exclude dead cells. The leukemic blast population phenotype was determined by the flow cytometric and CD antigen expression profile as compared to normal cell populations. Both CD45+ and CD45- leukemic blasts could be gated when stained with anti-CD45 antibody and analyzed by flow cytometry for CD45 expression and side scatter properties. All antibodies utilized were clinical grade direct fluorochrome conjugates (Becton Dickinson).
Confocal microscopy
U2OS cells were nucleofected with 3 μg fluorescein labeled (Mirus Label IT® Tracker, Madison, WI) pUC19 plasmid or pDSRed2C-1 plasmid per 106 cells. Immediately, or 3 days following nucleofection, cells were washed in cDMEM and counted. Cells were fixed on a glass slide with 3.7% buffered formalin, washed, permeabilized with 0.5% Triton X-100 (Surfact-amp, Pierce, Rockford, IL) and washed again. Pearmeabilized cells were incubated with 2.4 nM TOTO3 (Molecular Probes, Eugene, OR) and washed. Vectashield (Vector Laboratories, Inc., Burlingame, CA) was added to the cells prior to sealing with a coverslip. Optical sectioning of cells was taken sequentially using argon (488 nm excitation) and helium/neon (633 nm excitation) lasers on a Leica SPT S2 confocal microscope with a 100x oil immersion lens.
Quantitative real-time PCR
U2OS cells were nucleofected with 0.5 μg pDSRed2C-1 plasmid per 106 cells. Cells were harvested and used for nuclear DNA isolation. Prior to DNA isolation, nuclei were washed in PBS and incubated with 0.5U DNase I (Ambion, Austin, TX) at 37° for 10 min and washed again twice in PBS. Nuclear DNA was isolated (Nuclei EZ prep, Sigma, Saint Louis, MO) from transfected cells at designated time-points. The plasmid encoded neomycin phosphotransferase gene (neo) was amplified with primers and TaqMan hydrolysis probe as described by Sanburn, et al. [9]. Nuclear DNA from each of three experimental and three parallel control samples (U2OS cells Nucleofected without plasmid) was amplified in triplicate in an Opticon™ 2 Cycler (MJ Research™, Inc., Waltham, MA) with the following cycling protocol: 50°C 2 min, 95°C 10 min, with 40 cycles of 95°C for 15 sec., and 62°C for 1 min. To normalize the number of cells/nuclei, human RNase P was amplified using the TaqMan® RNase P reagents kit (Applied Biosystems, Foster City, CA), or for mouse cells, mouse Apo B was amplified using the primers 5' CACGTGGGCTCCAGCATT 3'and 5' TCACCAGTCATTTCTGCCTTTG 3' and the TaqMan hydrolysis probe 5'(FAM) CCAATGGTCGGGCACTGCTCAATA (TAMRA) 3' (courtesy Renee Horner, qpcrlistserver, yahoo groups, yahoo.com). The neo gene copy number per cell was determined using a plasmid-based standard curve.
Analysis of cell division
Cells were suspended in PBS and incubated with CFDA SE (5-(and -6)-carboxyfluorescein diacetate succinimidyl ester, CFSE (Molecular Probes, Eugene, OR) at a final concentration of 0.35 μM per 4 × 106 cells, incubated for 10 min at 37°C, and washed x3 in DMEM, 10% FBS. CFSE expression was analyzed by flow cytometry to assess cell division.
Cell cycle blockade
At 50–60% confluency, 0.6 mM mimosine (Sigma, Saint Louis, MO) was added to U2OS cells (2). Both U2OS and U2OS cells treated with mimosine were incubated at 37°C for 48 hours at which time the cells were harvested, counted and nucleofected with 1.5 μg per 106 cells pCI-neo vector encoding human 4IBBL (CD137L) cDNA [5]. As a control, cells were also nucleofected without plasmid. Four hours post-nucleofection cells were harvested, counted, stained with phycoerytherin (PE) labeled anti-human CD137L (BD Biosciences) and 7AAD (BD Biosciences), and analyzed for CD137L-expression by flow cytometry. To determine DNA content prior to nucleofection, cells were washed in phosphate buffered saline (PBS), fixed with 4% paraformaldehyde, washed again in PBS, and 1 μl propidium iodide (BD Bioscience) at 50 ug/ml was added. The propidium iodide labeled cells were then analyzed by flow cytometry [11].
Results
Optimization of Transfection by Electroporation
To determine differences in the kinetics and strength of expression of a transfected reporter gene using either an electroporation method (nucleofection) or a cationic liposome-based methodology (novafection), U2OS human osteosarcoma cells were transfected with 0.5 μg pDSRed2C-1 (RFP) plasmid per 106 cells, and RFP protein expression measured over time by flow cytometry. Following nucleofection, RFP was expressed as early as 4 hours in U2OS cells, Figure 1A. By 12 hours, 60% of the nucleofected cells expressed RFP. When this rate was adjusted for the cell death associated with nucleofection, the transfection rate dropped to 18% of the total cells initially transfected now expressing RFP at 12 hours. On day 3, when the percent viability of the nucleofected cells returned to 100%, the transfection rates no longer needed correction and the reported rates are identical. This expression was maintained for several days and gradually diminished until day 14 when expression could no longer be detected. In contrast, using novafection, RFP expression required 12 hours of culture (as opposed to 4 hours for nucleofection) and did not approach peak expression levels until day 1, Figure 1B. The efficiency of transfection was determined by using identical amounts of plasmid gene vector in each method. Upon comparison of RFP expression levels at day 1, the superiority of electroporation-based transfection was evident. In the viable fraction of nucleofected cells, 60% of these cells expressed RFP at 24 hours, as opposed to 15% of the novafected cells. Even when corrected for cells lost to electroporation-associated cell death, the nucleofection expression rate is 26%. As a control, cells were also nucleofected with a non-RFP expressing plasmid, pcDNA3.1/Hygro(-), and the minimal autofluoresence of transfected cells (less than 2%) subtracted from the reported expression levels. The RFP expression levels in cells transfected with the liposome-based reagent could be increased by increasing the amount of plasmid DNA, therefore these comparisons are relative and not maximized for each method (not shown).
Figure 1 Kinetics of transgene expression in electroporated and cationic lipid transfected U2OS cells. (A) RFP expression over time in cultured U2OS cells when nucleofected with pDSRed2C-1 (RFP) plasmid. Black bars represent the percent cells expressing RFP corrected for the total number of cells transfected and gray bars represent the percent of viable cells expressing RFP. (B) RFP expression over time in cultured U2OS cells when novafected with pDSRed2C-1 (RFP) plasmid. Since there is no cell death associated with novafection, the gray and black bars both represent the percent cells expressing RFP from the total number of cells transfected. Autofluorescence (always <2%) detected from cells nucleofected or novafected with pcDNAHygro(-) control plasmid was subtracted from the experimental values. The black line in both A and B represents average cell number at each time point. All experiments were done in triplicate with the standard deviation indicated by the error bars.
We then tested a primary mouse-derived neuroblastoma mass for the ability to be transfected by these methodologies. AGN2a tumor cells were injected subcutaneously into host strain mice, A/J, and the resulting subcutaneous tumor cell mass was excised, processed into a single-cell suspension, transfected with 0.5 μg pDSRed2C-1 (RFP) plasmid per 106 cells, and RFP expression over time measured by flow cytometry. Nucleofected primary murine tumor cells began to express RFP earlier than novafected cells (5 hr versus 1 day) and expression levels peaked at day 2 in viable nucleofected cells and day 4 in novafected cells, Figure 2. This later peak is likely due to the longer duration of transgene expression in novafected cells. Both the kinetics and total RFP expression levels differed for the human U2OS cell line, Figure 1, and the primary mouse-derived tumor, AGN2a, Figure 2. The nucleofection rates were not as high for the nucleofected primary tumor, while the novafection rate improved. These are vastly different systems and the rapid cell division rate of the cultured U2OS, Figure 3, as opposed to uncultured tumor that was excised, processed into a single cell solution and then transfected, may partially explain this result.
Figure 2 Kinetics of transgene expression in electroporated and cationic liposome transfected mouse primary tumor. The percent AGN2a mouse primary tumor cells expressing RFP in nucleofected (black bars) and novafected (gray bars) cells is shown. Autofluorescence detected from cells nucleofected or novafected with pcDNAHygro(-) control plasmid was subtracted from the experimental values. All experiments were done in triplicate with the standard deviation indicated by the error bars.
Figure 3 Confocal images of FITC-labeled plasmid in U2OS cells immediately and 3 days following nucleofection. A) Confocal images captured immediately following nucleofection: a) fluorescence of FITC-labeled pUC19 plasmid, b) phase contrast image, c) TOTO3 stained nuclei (dark blue), d) overlay of the images in (a) and (c) showing FITC-labeled plasmid in the nucleus. B) Confocal images captured 3 days following nucleofection: a) fluorescence of FITC-labeled pDSRed2C-1 plasmid, b) fluorescence from expression of RFP encoded on the FITC-labeled pDSRed2C-1 plasmid, c) TOTO3 stained nuclei, d) an overlay of the images in a, b and c showing FITC-labeled plasmid, expression of RFP and nuclear localization. C) Average number of plasmid copies per cell as determined by real-time PCR of the plasmid encoded neo gene from nuclear DNA harvested immediately following nucleofection. D) Plasmid neo copies amplified from nuclear DNA harvested 3 days following nucleofection. In both C and D the gray bars represent neo amplification from U2OS cells nucleofected with 0.5 μg pDSRed2C-1 per 106 cells and the black bars represent neo amplification from non-nucleofected U2OS cells. Nuclear DNA was harvested and amplified from 3 separate samples, error bars indicate the standard deviation from the average of triplicate samples.
The primary limitation of electroporation-based transfection is cell death. Preliminary experiments confirmed that increasing the strength of the electric field corresponded to both a higher transfection rate, and increased cell death. The nucleofection setting that we found optimal resulted in 70% cell death, Figure 1A. Cell numbers gradually recovered post-nucleofection, beginning at 24 hours. In contrast, there was no decrease in cell number following novafection, Figure 1B.
Delivery of plasmid DNA to the nucleus by electroporation is rapid and short-lived
The inability to culture most primary human tumors led us to search for methods of transfection that would require minimal culture and processing time while allowing for efficient gene transfection. Given the rapid kinetics of expression using nucleofection, we sought to determine if this was due to direct delivery of plasmid DNA in to the nucleus. Confocal microscopy was used to visualize individual z-plane sections that represent internal nuclear layers of U2OS cells that had been nucleofected with 3 μg FITC-labeled pUC19 plasmid per 106 cells, immediately cytospun onto glass slides, and then prepared for microscopy. The nuclear and cytoplasmic boundaries of nucleofected cells were visualized by phase contrast microscopy, Figure 3A, panel b, or by staining with the nuclear dye TOTO3, Figure 3A, panel c. The nuclei are stained dark blue, with a lighter blue staining in the cytoplasmic compartment. The plasmid-associated fluorescein signal was present in both the cytoplasmic and nuclear compartments immediately following nucleofection, Figure 3A, panel d. Visual inspection reveals that most cells contained nuclear plasmid, Figure 3A, d (an overlay of the plasmid signal with the TOTO3 stain).
Using the same technique, we then sought to determine how long after nucleofection the plasmid vector remained in the nucleus. Three days after nucleofection of U2OS cells with 3 μg FITC-labeled pDSRed2C-1 (RFP) plasmid per 106 cells, the presence of plasmid vector DNA, was greatly diminished, Figure 3B, panel a. The presence of plasmid vector DNA, as detected by FITC fluorescence, was seen in a small minority of cells, and when present on day 3 it appeared to associate more with a punctate fluorescence in the cytoplasm, Figure 3B, a and d. Despite the loss of plasmid vector from the nucleus, intense red fluorescence was seen in many of the cells at this time, indicating the continued presence of red fluorescent protein, Figure 3B, panels b and d.
To further confirm the presence of plasmid in the nucleus, the copy number of plasmid vector per cell was calculated by real-time PCR amplification of the pDSRed2C-1 encoded neomycin phosphotransferase gene (neo). U2OS cells were nucleofected with 0.5 μg pDSRed2C-1 per 106 cells and immediately, or at day 3, nuclear DNA was isolated from the nuclear fraction of cell lysates and PCR amplified using neo primers and a neo-specific TaqMan probe. The total number of plasmid neo copies was calculated based on comparison to a standard curve generated with the same plasmid vector. The number of cells (or nuclei) analyzed was determined using a standard curve calibrated to genomic DNA mass and signal from the single copy gene RNAseP. Nuclei were purified on a sucrose cushion, washed with PBS, digested with DNAse in order to remove contaminating cytoplasmic plasmid DNA, and total DNA extracted. Immediately following nucleofection, there were 200 to 400 copies of plasmid per cell, Figure 3C. In agreement with microscopy data, copy number in U2OS cell nuclei decreased to 50 copies or less per cell by day 3 post-nucleofection, Figure 3D. Immediate localization of plasmid to the nucleus following nucleofection was also observed by real-time PCR in the AGN2a and SCCVII cell lines (data not shown).
Delivery of plasmid gene vectors to the nucleus requires cell division for optimal gene expression
The ideal cell-based cancer vaccine would allow recombinant gene expression in primary human tumor cells. The bottleneck is that primary human tumors do not grow and divide efficiently in culture. Therefore, it was essential to evaluate the association between cell division and expression of nucleofected genes. To determine the pattern of cell division rates in nucleofected tumor cell lines (U2OS, AGN2a and SCCVII), cultured tumor cell lines were incubated with carboxyfluorescein diacetate succinimidyl ester (CFSE). The intensity of CFSE fluorescence was analyzed by flow cytometry every two hours for 10 hours and then daily for 3 days after nucleofection, Figure 4A–C. According to the decrease in CFSE fluorescence intensity with each cell division (leftward shift of peaks), the U2OS cells underwent about 3 cell divisions in the first 10 hours post-nucleofection, whereas, the AGN2a and SCCVII cells divided approximately once in the first 10 hours post-nucleofection. CFSE stained cells were also nucleofected with RFP-encoding plasmid and there was no difference in the CFSE fluorescent pattern by flow cytometry in the nucleofected versus non-nucleofected cells (data not shown). As demonstrated previously, it was the rapidly dividing U2OS cells that exhibited transgene expression at 4 hours (20% of the viable expressed RFP), while RFP could not be detected in the slower dividing AGN2a and SCCVII until 12 hours post-nucleofection, Figure 4D.
Figure 4 Cell division and gene expression in U2OS, AGN2a and SCCVII cells. Tumor cells were labeled with CFSE to detect changes in fluorescence that occur with cell division post-nucloefection. (A) U2OS, (B) cultured AGN2a, (C) SCCVII cells were analyzed for CFSE fluorescence by flow cytometry immediately following nucleofection or cultured up to 3 days following nucleofection. The solid gray peak (far left of each panel) represents unstained cells. Peaks decreasing in fluorescence (right to left) indicate the loss of CFSE fluorescence over time as cells divide. Cells were analyzed at hour 2, 4, 6, 8, 10, day 1, day 2, and day 3. Only U2OS showed peaks indicating decrease in fluorescence prior to day one: at hour 2, and hours 4–10. D) In parallel experiments, the average percent of cells expressing RFP from 3 separate experiments at the indicated time points following nucleofection was determined in U2OS (■), SCCVII (▲) and AGN2a (◆) cell lines by flow cytometry. E) Average percentage of cells expressing CD137L at time 0, 2 hours, and 6 hours post-Nucleofection in U2OS and SCCVII cells. The error bars represent the standard deviation from 3 experiments. F) Average viable cell number of U2OS, SCCVII and AGN2a tumor cells post-nucleofection from 3 separate experiments, expressed as a percentage of the number of cells originally nucleofected.
To further explore this finding, we used an alternate plasmid-encoded transgene and compared the kinetics of expression in rapidly and slower dividing cell lines. Our laboratory has produced a number of immune co-stimulatory expression vectors, and we chose a human CD137L (4IBBL) expression vector for further study. Expression of this cell-surface antigen can be directly measured by flow cytometry using a labeled CD137L-specific antibody. One of our concerns with using RFP was that the DsRed2 protein we utilized requires approximately 6 hours to reach full fluorescence intensity due to a requirement for intracellular oxidation [12]. Therefore a CD137L-encoding plasmid, 1.5 μg per 106 cells, was nucleofected into the U2OS and SCCVII cells and expression compared. In the rapidly dividing U2OS, 40% of the cells expressed CD137L as early as 2 hours post-nucleofection. At 6 hours post-nucleofection, 80% of the U2OS cells expressed CD137L, Figure 4E. In contrast, only 10% of SCCVII cells expressed CD137L at 6 hours post-nucleofection. The nucleofection process also induced significant cell death, demonstrating that cell death was not an RFP-associated phenomenon, Figure 4F. All cells experienced 60 to 80% cell death upon nucleofection, however the SCCVII cells recovered much more rapidly than either the rapidly dividing U2OS cells or the less rapidly dividing AGN2a cells, indicating that factors other than cell division are involved in cellular recovery from the electroporation and transfection processes.
To directly assess the need for cell division, U2OS cells were incubated with 0.6mM mimosine, which blocks cell cycle progression, and CD137L expression of nucleofected cells compared to non-treated controls, Figure 5A. In non-mimosine treated U2OS cells, approximately 20% of cells expressed CD137L at 4 hours post-nucleofection. In contrast, less than 4% of mimosine treated cells expressed CD137L at 4 hours post-nucleofection. To assure that cell cycle block occurred in mimosine treated cells, cell cycle status was determined by propidium iodide staining. The majority of mimosine treated cells were in G1/G0 while the mimosine untreated cells had a greater proportion of cells in S, and G2/M phases of the cell cycle, Figure 5B. This finding indicates that cell cycle progression enhances transgene expression by nucleofection. This finding was unexpected as it implies that other mechanisms are at work besides the delivery of plasmid DNA to the nucleus.
Figure 5 Impact of cell cycle inhibition on gene vector expression. A) The average percentage of U2OS cells expressing vector-encoded CD137L 4 hours following nucleofection. The black bar represents cells nucleofected without plasmid, stippled bar represents untreated cells, and the gray bar represents mimosine treated cells. The error bars show the standard deviation from 3 separate experiments. B) Flow cytometric profile of propidium iodide stained mimosine treated U2OS cells (solid gray) and mimosine untreated U2OS cells (black line) prior to nucleofection of CD137L.
Application of electroporation-based transfection to human leukemias
To assess the ability of electroporation-based transfection to be utilized in the production of autologous tumor vaccines for patients, we first determined the optimal electrotransduction parameters. Fresh primary leukemia cells from 6 patient samples were nucleofected with 1 μg of pDSRed2C-1 plasmid per 106 cells, using three different Amaxa solutions (R, T, V) and several different electrotransduction programs (T20, U15, T17, T27, S04, O17, T16, T01, and O17). The expression of RFP protein was determined by flow cytometry 24 hours post-nucleofection. There was a wide range of RFP expression depending on the electroporation conditions, Table 1. Amaxa solution R with program T20 and solution T in combination with program U15 consistently provided the highest expression of RFP protein. The highest expression level seen was 62%, (sample 4 PB) and the lowest expression was 1.8% (sample 12 BM). The RFP expression of 12 total samples collected using solution R and program T20, along with available diagnostic phenotypic information, is presented in Table 2. The R/T20 settings were chosen based on a slightly better viability profile than the T/U15. Greater than 5% expression was seen in 9 of 12 samples and greater than 20% expression was seen in 5 of 12 samples. These patient leukemias were extensively characterized for standard childhood leukemia markers, and no phenotype was clearly associated with the ability to be transfected using the parameters we established. Although not reported, the number of leukemic cells obtained per patient was variable- as the samples obtained for this study were essentially diagnostic remains that were to be discarded. For two of the samples (5 and 6) we had both peripheral blood and bone marrow-derived matched samples. In both cases the peripheral blood cells showed a higher transfection rate. Our certainty that the transfected cells analyzed were the leukemia cells comes from the clinical diagnostic experience of the Cell Marker Laboratory of the Children's Hospital of Wisconsin, where the identical procedures used for clinical diagnosis of malignancy were used to analyze the nucleofected samples. The majority of the samples we analyzed were bone marrow aspirates. In patients with advanced disease, autologous bone marrow may prove to be both an accessible and abundant source of tumor cells that can be modified by plasmid-based gene vectors to produce cell-based vaccines.
Table 1 Comparison of nucleofection parameters in ALL patient samples. Nucleofection using an RFP expression vector was carried out using 6 different patient samples (patients 4,13,8,11,19,12) from one of two potential tissue sources, PB, peripheral blood, or BM, bone marrow. 1No NF, not nucleofected autofluorescence control. 2Nucleofection was carried out with one of three solutions R, T, or V, and the following electrical settings on the Amaxa nucleofection device: T20, U15, T17, T27, S04, O17, T16, T01, O17. 3The table lists the percentage of viable cells expressing the plasmid-encoded RFP at 24 hours post-nucleofection.
Electroporation Parameters
Patient Samples No NF1 R2/T20 T/U15 R/T17 R/T27 T/S04 T/O17 R/T16 V/T01 V/O17
4 PB 30.1 62.3 61.5 56.8 52.9 52.5 46.2 46.1 40.5 20.9
13 BM 1.0 45.0 45.6 42.9 40.2 32.4 36.1 38.5 21.3 36.1
8 PB 1.0 9.2 10.9 9.3 6.7 8.2 5.6 6.4 6.1 3.7
11 BM 1.2 17.1 25.2 13.2 13.4 9.5 5.7 10.0 5.7 7.1
10 BM 0.1 8.0 10.0 4.6 4.5 2.3 2.0 3.9 2.0 0.9
12 BM 1.3 2.9 3.6 2.4 2.3 2.6 3.5 1.8 2.1 3.6
Discussion
Cell-based autologous cancer vaccines hold great promise in the effort to shift the adaptive immune system from ignorance or anergy toward cell-mediated immune recognition of cancer. The generation of a Th1-type immune response with cell-based vaccines, and the resultant CTL-mediated killing of tumor, have offered the most effective anti-tumor responses in pre-clinical models, and also initiated clinical responses, as reviewed by Nitti et al., [13]. Ex vivo introduction of gene vectors encoding immune activating signals (such as co-stimulatory antigens, cytokines and adhesion molecules) into tumor cells with subsequent re-introduction of irradiated modified tumor cells into patients is currently being pursued with the aim of initiating a clinically-relevant immune response against tumor [14,15]. Because this type of therapy involves cell processing, culture procedures, as well as genetic transfer, any steps that streamline the process will make a significant contribution toward tumor vaccine development.
Newer generation transfection procedures based on electroporation or branched polyethyleneamine (PEI) have been reported to be independent of cell cycle effects [14]. Nevertheless, we found that cell cycle progression does enhance transgene expression and that certain tumor cell lines (such as U2OS) thought to be highly "transfectable" may be highly amenable to vector-derived transgene expression precisely because they have rapid rates of cell division. Most reports do not analyze transgene expression at the early times (2–8 hours) we report here, and unless proven otherwise, the contribution of cell cycle progression within the first 8 hours of transfection cannot be excluded. We examined this question in detail because our goal is to transfect primary human leukemias that do not persist in tissue culture and thus have limited in vitro proliferative capacity.
Having thus defined the problem, we examined in detail a transfection procedure that appears to be the most rapid. Nucleofection delivers plasmid vector DNA in to the nucleus immediately, Figure 3. However, as we demonstrate here, nucleofection still benefits from cell cycle progression and cell division, Figures 4,5. Using nucleofection we found that tumor cells harvested directly from a mouse bearing the AGN2a neuroblastoma could be nucleofected with high efficiency. Thus, the process of tumor disaggregation and processing did not render the neuroblastoma cells we studied resistant to nucleofection. Confident of these results, we initiated analysis of primary human tumors.
The tumor samples we analyzed were from patients undergoing diagnosis for suspected childhood leukemia. If samples from blood or bone marrow remained after diagnostic procedures were completed, they were then tested for the ability to be nucleofected. Nine of 12 specimens showed greater than 5% transgene expression, and half of these showed greater than 17% expression, Table 2. This means that for the majority of the patients we tested, nucleofection with transgenes that encode either immune co-stimulatory molecules or cytokines may be feasible. We did not analyze gene expression in nucleofected ALL specimens over time, as our long-term goal is to transfect and then use these cells as experimental vaccines soon thereafter. Studies with other cultured tumor cell lines in our lab have shown that transgene expression rarely persists beyond 9 days, as nucleofected plasmid vectors do not integrate into the genome. While clinical samples could be frozen, thawed, cultured overnight, and then nucleofected, the viability of the frozen and thawed clinical samples was low, between 10–25% (not shown). We propose that the highest nucleofection rates are likely to be seen using fresh leukemic cells, and that viability upon thawing may not be critical as cells will be irradiated prior to use as a vaccine. Future studies will be carried out to address this issue. Our report also highlights that the nucleofection of primary ALL was quite variable. Reasons for this could be due to a specific phenotype that could be identified by expression profiling. Building upon our in vitro data it may be the ability of one clinical sample to undergo cell division or even limited expansion in culture over another during the overnight culture period following nucleofection may allow for better transgene expression. Thus, inclusion of growth factors or cytokines may increase nucleofection rates above those we report here.
Table 2 Transfection of Patient Leukemia. Patient leukemia cells were nucleofected with solution R, setting T20, as described in Methods. Patient diagnosis and specimen type are listed (PB, peripheral blood, BM, bone marrow). Relevant diagnostic criteria with respect to phenotype is also listed. Transfection efficiency is based upon the percentage of viable cells expressing RFP at 24 hours post-nucleofection.
Patient Sample Diagnosis Specimen Phenotype Transfection Efficiency
4 ALL PB 45+/Tdt+/DR+/34+/10-/19+/20–24+/9+/56+ 62.3
13 ALL BM 45+/Tdt+/DR+/34-/10+/19+/20+/24+/9-/52+/22+ 45.0
14 ALL BM 45-/Tdt+/DR+/34–10+/19+/20dim/24+/9+ 44.3
9 ALL-CD56 BM 45-/Tdt+/DR-/34+/10+/19+/20-/24-9+/56+ 26.3
5 ALL-CD33 (a)PB
(b)BM 45+/Tdt+/DR+/34+/10+/19+/20–24+/9+/33+/117dim 23.0 (a)
13.8 (b)
11 ALL BM 45+/Tdt+/DR+/34+/10+/19+/20-/24+/9+/13-/33-/117-/52+/25+ 17.1
8 ALL PB 45+/Tdt+/DR+/34-/10+/19+/20het/24+/9+/μ-/κ-/λ- 9.2
10 ALL-CD13/33 BM 45-/Tdt+/DR+/34+/10+/19+/20-/24+/9+/13+/33+/117- 8.0
3 ALL-CD15 BM 45-/Tdt+/DR-/34+/10-/19+/20-/24-9+15+/56+ 5.9
7 ALL BM 45-/Tdt+/DR+/34+/10+/19+/20–24+/9+ 4.3
6 ALL (a)PB
(b)BM 45-/Tdt+/DR+/34+/10+/19+/20-/24-/9+ 3.4 (a)
1.0 (b)
12 T-ALL BM 45+/Tdt+/DR+/34+/10-/19-/20-/24-/9-/13-/33-/117-/52+/7+/5+/2+/3-/4+/8-/1a-/TCRab-/ TCRgd- 2.9
Conclusion
The process of nucloefection delivers plasmid DNA directly to the nucleus. Even though delivery to the nucleus is thought to circumvent dependence on cell division, we found that the highest and earliest levels of transgene expression from plasmid-based vectors occurred in rapidly dividing cells. We were also able to demonstrate that primary acute lymphocytic leukemia cells (ALL) from pediatric patients could also be nucleofected with plasmid-based vectors, thus opening the door to patient-specific cell manipulation. In light of the laboratory studies we present, transfection rates of clinical samples may be increased even further if some degree of cell division could be induced during the in vitro culture of these specimens. Even though this culture time is less than 24 hours, a single cell division could potently increase transgene expression levels, and thus immunogenicity of the vaccine preparation. Future work will include analysis of the correlation between the ability to be nucleofected, markers of cell cycle progression, and the induction of cell cycle progression post-nucleofection.
Declaration of competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JG, BJ, SB and JW carried out laboratory studies on established cells lines. JG carried out all confocal microscopy. DS analyzed marker expression on clinical samples and organized clinical data, NN was responsible for transfection of clinical samples, KB, KM, and AW were responsible for clinical care, patient consent, and pathological interpretation, RO was responsible for study design and execution.
Acknowledgements
We would like to thank Elizabeth C. Glaser for expert technical assistance, and Dr. James Casper and Dr. Bruce Camitta for helpful discussions and clinical leadership. This work was supported by grants from the Midwest Athletes Against Childhood Cancer, MACC Fund, Inc., and by an Internal Research Grant from the American Cancer Society, Medical College of Wisconsin Cancer Center.
==== Refs
Hock RA Reynolds BD Tucker-McClung CL Kwok WW Human class II major histocompatibility complex gene transfer into murine neuroblastoma leads to loss of tumorigenicity, immunity against subsequent tumor challenge, and elimination of microscopic preestablished tumors J Immunother 1995 17 12 8
Bausero MA Panoskaltsis-Mortari A Blazar BR Katsanis E Effective immunization against neuroblastoma using double-transduced tumor cells secreting GM-CSF and interferon-γ J Immunother 1996 19 113 24
Katsanis E Orchard PJ Bausero MA Gorden KB McIvor RS Blazar BR Interleukin-2 gene transfer into murine neuroblastoma decreases tumorigenicity and enhances systemic immunity causing regression of preestablished retroperitoneal tumors J Immunother 1997 15 81 90
Johnson BD Yan X Schauer DW Orentas RJ Dual expression of CD80 and CD86 produces a tumor vaccine superior to single expression of either molecule Cell Immunol 2003 222 15 26 12798304 10.1016/S0008-8749(03)00079-0
Yan X Johnson BD Orentas RJ Murine CD8 lymphocyte expansion in vitro by artificial antigen-presenting cells expressing CD137L (4-1BBL) is superior to CD28, and CD137L expressed on neuroblastoma expands CD8 tumor-reactive effector cells in vivo Immunol 2004 112 105 116 10.1111/j.1365-2567.2004.01853.x
Mocellin S Mandruzzato S Bronte V Lise M Nitti D Part I: Vaccines for solid tumors Lancet Oncol 2004 5 681 689 15522656 10.1016/S1470-2045(04)01610-9
Rousseau RF Haight AE Hirschmann-Jax C Local and systemic effects of an allogeneic tumor cell vaccine combining human lymphotactin with interleukin-2 in patients with advanced or refractory neuroblastoma Blood 2003 101 1718 1726 12406881 10.1182/blood-2002-08-2493
Schakowski F Buttgereit P Mazur M Novel non-viral method for transfection of primary leukemia cells and cell lines Genet Vaccines Ther 2004 2 1 11 14715084 10.1186/1479-0556-2-1
Sanburn N Cornetta K Rapid titer determination using quantitative real-time PCR Gene Ther 1999 6 1340 1345 10455446 10.1038/sj.gt.3300948
Perry C Sastry R Masrallah IM Stover PJ Mimosine attenuates serine hydroxymethyltransferase transcription by chelating zinc J Biol Chem 2005 280 396 400 15531579 10.1074/jbc.M412914200
Tlsty T Briot A Poulose B Analysis of cell cycle checkpoint status in mammalian cells Methods Enzymol 1995 254 125 133 8531681
Bevis BJ Glick BS Rapidly maturing variants of the Discosoma red fluorescent protein (DsRed) Nat Biotechnol 2002 20 83 87 11753367 10.1038/nbt0102-83
Mocellin S Rossi CR Nitti D Cancer vaccine development: on the way to break immune tolerance to malignant cells Exp Cell Res 2004 299 267 78 15350526 10.1016/j.yexcr.2004.06.017
Brunner S Fürtbauer E Sauer T Kursa M Wagner E Overcoming the nuclear barrier: cell cycle independent nonviral gene transfer with linear polyethylenimine or electroporation Mol Ther 2002 5 80 86 11786049 10.1006/mthe.2001.0509
Soiffer R Lynch T Mihm M Jung K Rhuda C Schmollinger JC Hodi FS Liebster L Lam P Mentzer S Singer S Tanabe KK Cosimi AB Duda R Sober A Bhan A Daley J Neuberg D Parry G Rokovich J Richards L Drayer J Berns A Clift S Cohen LK Mulligan RC Dranoff G Vaccination with irradiated autologous melanoma cells engineered to secrete human granulocyte-macrophage colony-stimulating factor generates potent antitumor immunity in patients with metastatic melanoma Proc Natl Acad Sci USA 1998 95 13141 13146 9789055 10.1073/pnas.95.22.13141
|
15969754
|
PMC1182385
|
CC BY
|
2021-01-04 16:39:08
|
no
|
Genet Vaccines Ther. 2005 Jun 21; 3:4
|
utf-8
|
Genet Vaccines Ther
| 2,005 |
10.1186/1479-0556-3-4
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-151598518110.1186/1743-0003-2-15ResearchA Dynamic Neuro-Fuzzy Model Providing Bio-State Estimation and Prognosis Prediction for Wearable Intelligent Assistants Wang Yu [email protected] Jack M [email protected] Department of Biomedical Engineering, Marquette University, Milwaukee, WI, USA2005 28 6 2005 2 15 15 10 2 2005 28 6 2005 Copyright © 2005 Wang and Winters; licensee BioMed Central Ltd.2005Wang and Winters; 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
Intelligent management of wearable applications in rehabilitation requires an understanding of the current context, which is constantly changing over the rehabilitation process because of changes in the person's status and environment. This paper presents a dynamic recurrent neuro-fuzzy system that implements expert-and evidence-based reasoning. It is intended to provide context-awareness for wearable intelligent agents/assistants (WIAs).
Methods
The model structure includes the following types of signals: inputs, states, outputs and outcomes. Inputs are facts or events which have effects on patients' physiological and rehabilitative states; different classes of inputs (e.g., facts, context, medication, therapy) have different nonlinear mappings to a fuzzy "effect." States are dimensionless linguistic fuzzy variables that change based on causal rules, as implemented by a fuzzy inference system (FIS). The FIS, with rules based on expertise and evidence, essentially defines the nonlinear state equations that are implemented by nuclei of dynamic neurons. Outputs, a function of weighing of states and effective inputs using conventional or fuzzy mapping, can perform actions, predict performance, or assist with decision-making. Outcomes are scalars to be extremized that are a function of outputs and states.
Results
The first example demonstrates setup and use for a large-scale stroke neurorehabilitation application (with 16 inputs, 12 states, 5 outputs and 3 outcomes), showing how this modelling tool can successfully capture causal dynamic change in context-relevant states (e.g., impairments, pain) as a function of input event patterns (e.g., medications). The second example demonstrates use of scientific evidence to develop rule-based dynamic models, here for predicting changes in muscle strength with short-term fatigue and long-term strength-training.
Conclusion
A neuro-fuzzy modelling framework is developed for estimating rehabilitative change that can be applied in any field of rehabilitation if sufficient evidence and/or expert knowledge are available. It is intended to provide context-awareness of changing status through state estimation, which is critical information for WIA's to be effective.
==== Body
Background
Emerging wearable technologies are expected to constitute an important component of the vision of user-centered, 21st-century rehabilitative healthcare [1-4]. Indeed, the consensus recommendations of a workshop on future homecare technologies envisioned intelligent wearable sensors as one of the top trends [1]. The top two knowledge gaps that were identified targeted the need for better [1,2]:
1. information reduction algorithms and sense-making tools, and
2. outcomes and functional assessment tools.
This project addresses these gaps in knowledge for the area of rehabilitative healthcare.
The first of these recognizes the challenge of effectively integrating and using the massive amount of sensor-based data that can be potentially be collected. It is well established in the intelligent systems community that a key barrier to intelligent use of information is context-awareness. With humans, this "context" is always changing as their state of health and their present environment or goals change. Relevant "states" of a person with disability can range from a degree of impairment (e.g., spasticity) to a perception of pain, and such states frequently change over the course of a day (e.g., due to medication). Thus a first goal is context-awareness , which for an intelligent wearable technology includes estimation of relevant states of the person. For instance, how a certain sensed event is interpreted can be influenced by the current "state" of person (e.g., degree of spasticity, pain), as well the history of past inputs (e.g., medications taken recently).
In response to the second of these, our original work on this project was motivated by the desire to create an intelligent system that was based on the mind-set of the rehabilitation practitioner. This led to the aim of designing a prognosis-prediction system that integrated the stages identified in clinical practise guidelines [5], a dynamic process that includes diagnosis (based on factual and context data), prognosis (prediction of outcomes, based on certain assumptions), a "clinical algorithm" of interventions (inputs to the human system), allocation of human resources (e.g., practitioner time), and outcomes measurement. While we started from the perspective of planning to use consensus expert experience to build models, a key trend in clinical rehabilitation has been a focus on evidence-based practice [2,5,6]. Also, we noted that the common goal of optimizing therapeutic interventions (e.g., movement therapy) over the continuum of care [6,7] bears striking similarity to classic engineering optimization problem [3].
The above concepts provided the core motivation for our Intelligent Telerehabilitation Assistant (ITA) project [1,3,8,9]. There are two core parts to our vision for mobile ITA technology [1]: i) a user-customized interface that supports multimedia teleconferencing and wireless communication, and collection of sensor-and user-based information that can be used to determine events; and ii) embedded intelligent "soft" computing, based on event-driven expert system modules. This paper addresses a part of the latter, which to us appears to be the greater challenge. Given this focus, perhaps a better term than ITA, at least for mobile applications, would be a wearable intelligent assistant/agent (WIA). Use of WIA emphasizes the need for context-awareness and prognosis prediction to a greater degree, with the focus on the person rather than on the connection. Aims of a WIA include: i) providing data within an ecologically valid setting, ii) improving timely assessment of health status, iii) identifying and predicting client outcomes (a running prognosis); and iv) assisting with intervention strategies.
Notice the inclusion of both "assistant" and "agent" for a WIA. The former is motivated by the disability community, and the latter by the intelligent systems community. An intelligent assistant is an assistive technology that directly interacts with and supports the user-client by providing strategic assistance (e.g., with completion of a certain task; providing reminders related to a certain assessment or therapeutic protocol; using performance monitoring to change settings during a therapeutic task). In contrast, an intelligent agent recognizes events and/or senses data on the user's behalf, and once triggered (normally by using a previously designed rule database), can perform certain actions (e.g., process and manage data, prompt a session between the client and a remote site, negotiate with other agents) while requiring minimal attentional resources by the user. We view ITAs and WIAs as falling into two categories [3]:
• Task-based, assistive modules that facilitate ease of use and implementation of evaluative and therapeutic protocols, and
• Decision-support modules that assist practitioners and consumers with outcomes assessment and with optimizing the rehab intervention strategy.
The present contribution can be viewed as an encapsulated, distributed intelligent processor that is used by a WIA, or more specifically as a resource for a WIA.
Importantly, it is designed in two stages. In the development stage, the designer possesses a suite of tools for creating the model. This model includes identification of:
• input events and facts,
• the bio-states of interest that are expected to change over time (and whose estimation provides context-awareness),
• performance outputs to be predicted by the model (and in some cases can be compared to sampled measures), and
• desired outcomes (optional capability).
All of these are represented as signals, and furthermore signals that change over time. Indeed, the aim of clinical rehabilitation is to cause change that is over-and-above spontaneous healing bioprocesses [3], and to study such processes one must also model intrinsic healing mechanisms. Thus what is needed is a dynamic model that captures change, and can furthermore predict future change (make a "prognosis") if assumptions are made on future inputs (e.g., a "clinical algorithm" of interventions is implemented). The need to model change in states such as "impairment" implies a model that includes differential equations, and the desire to "remodel" the system suggests adaptive control mechanisms. Yet the likely designer of the system is one with experience and knowledge of available evidence, i.e. a practitioner or a clinical researcher. This makes a strong case for using rule-based fuzzy inference , which is well-known for its ability to both capture expert reasoning and provide robust system performance [10,11]. It also suggests that any model development environment must have carefully-designed graphical user interface (GUI) windows that can help guide the designer through the process of defining linguistically-meaningful signals (inputs, states, outputs, outcomes) and using rules to establish how changes in states will happen in response to input events and current states. More broadly, it can be viewed as a bio-modelling tool for uses rules to generate nonlinear differential equations that can be used by stakeholders ranging from telepractitioners to basic scientists who are addressing healing and remodelling bioprocesses.
When formulated in this way, the structure bears direct similarity to the classic state and output vector equations of systems and control theory, only with the nonlinear state equations developed by fully linguistic and interactive procedures of a rule-based fuzzy inference system (FIS). In our case the equations are implemented via dynamic connectionist neural network (CNN) connections. We thus use "rules" as the bridge between human reasoning and the mathematical model [8-11]. Note that crisp logic can be viewed as a special case of fuzzy logic [11].
Such neuro-fuzzy approaches fall under the umbrella of "soft computing" technologies [10,11], but the approach described here appears to be unique in its focus associating rules with changes in state and thus nonlinear differential state equations created in a linguistic space. Such soft computing approaches have the dual advantages of a structure that can enable robust model behaviour (if designed well) that has made fuzzy controllers such an economic success story, plus use of a intelligent systems architecture that should make it interface well with WIAs decision-making modules. We have coined our general design system SoftBioME (Soft Bio-Modeling Environment, pronounced "soft-by-ohm").
Once designed and customized for a client, in the embedded "run" mode, the model must receive inputs (sensor-events, user-events) as a function of time. The job of the model is then to produce ongoing state estimation (for context-awareness) and useful outputs . There are three types of useful outputs: i) performance predictions (e.g., for comparison to actual performance, when measured); ii) specific actions that are a function of states and inputs (e.g., prompting/informing/reminding a client); iii) other value-added decision-support signals for a WIA. Note that it also allows "what if" use by the WIA or a user: it will predict future states, output and outcomes if assumptions are made on future input events.
Developed within the Microsoft .Net Framework using mostly C# code, the "run mode" code is designed to run on any Windows-base system ranging from desktop to PDA. It uses an object-oriented structure, it's support for XML should make it easy to interface with other modules or the web. However, when used in designer mode, it requires a monitor that is large enough to display interface window sizes that are normally intended for desktop/laptops.
Methods
The fuzzy system is implemented by a dynamic recurrent neural network that is composed of four layers of CNNs (Figure 1): input, rule-state, output and outcome. Collectively, it is defined by its structure, signals, and parameters (e.g., membership function describing parameters, weights, time constants). We define four roles for users, listed by level of security access:
Figure 1 Structural relation between the model and the real human system. The intervention plan drives both the real system and fuzzy model, with the sampled (measured) output signals feedback back as an error event signal, and outcome error signals available to mildly tune the adaptive state estimators and output and outcome predictors. Targeted parameters can include input or output mappings or rule weights. When used in a simulation mode, the model can be used to predict the consequences of alternative treatment/intervention plans, and thus help the user optimize the intervention strategy. CNN: connectionist neural network. Dashed line: Sampling. Dotted line: future adaptive CNN work.
• User-designers, who have access to all aspects of model creation and implementation, including defining and adding signals, rules and parameters.
• User-analysts, who have access to specifying inputs, to all graphics capabilities, and to using tools such as sensitivity analysis on any internal signals or parameters, but cannot add rules or permanently change parameters.
• User-practitioners, who have access to specifying inputs and storing "what-if" and sensitivity-analysis simulations, as well as full desktop graphics features.
• User-clients, who are often also patients, and have a simpler interface intended for a PDA that can specify inputs, receive outputs, and can obtain current state and output information and summary predictive information.
A given user may participate in (and thus have access to) multiple roles. For instance, an informed and highly engaged patient-client who is active in self-care may normally function in the role of user-client, but can log in to a desktop version where they have "user-practitioner" or "user-analyst" access. Similarly, an experienced practitioner may normally function in the role of user-practitioner, but periodically login as user-analyst and on occasion as user-designer so as to add a new rule or change a membership function or gain. The remainder of this section targets the capabilities of the system from the perspective of the user-designer.
Early versions of this model have been presented as conference papers [8,9]. In the process of using this model for research and for homework projects in rehabilitation courses, it became clear that there was a need to add a number of features:
i) to more fully delineate between and support key dynamic processes associated with different forms of inputs;
ii) to set up a rule structure that enables parametric time constant changes;
iii) to define and implement homeostatic states; and
iv) to support advanced sensitivity and optimization tools.
This paper presents this refined structure, with a special focus on two areas of special interest for WIAs: state estimation for context-awareness and outputs/outcomes prediction for prognosis updating. The model of Figure 1 is presented in a right-to-left progression, since a user-designer normally starts by identifying desired outcomes and outputs.
Outcomes Layer: Predicting Client Outcomes
Outcomes are defined as scalar signals that relate to what in engineering optimization are called performance sub-criteria or cost functions, and can be a function of fuzzy states and outputs (and if desired, also inputs). Outcomes are thus what a "clinical algorithm" seeks to maximize or minimize. Examples of rehabilitative outcomes are numerical representations of terms such as impairment, disability, independence, quality of life, satisfaction, and cost. An outcome is calculated as a weighted sum or a weighted sum of squares of dimensionless state signals (X ) and state expressions (e.g., result of "state is low", called M x ), and output (Y ) signals. Weights are selected by the user-designer from a menu table.
Output Layer: Converging Signals to Predict Performance
As with a conventional control system, outputs are linguistic variables that are function of states and inputs, and change value dynamically only as states and/or inputs change. A given output typically falls into one of three categories:
i) performs an action (e.g., prompt WIA or user-client, initiate communication, store data in an electronic record),
ii) predicts a performance metric, preferably of a quantity that can be sampled on occasion (e.g., a measure such as a clinical scale or biomechanical metric), or
iii) provide targeted decision-support information of use to the user.
The output of the ith output-neuron in this layer, y i , is a function of the states of the rules and the input events (see figure 2).
Figure 2 Layer structure of the model. Most of the neurons in the input layer detect the occurrence of events and mapping the events into fuzzy variables. Others are pre-processing neurons for certain types of inputs, such as performing as pharmacokinetic models to map the dose and/or regimen of one kind of medication into the effective concentration, or integration neurons to calculate the accumulative effect of interventions. For each state, there are generally five nuclei in the rule-state layer. The outputs of tonic rules nuclei determine the absolute value of the state, and the phasic rules nuclei brings the instant change to the state. (Specially, the nuclei connect the fact/context and the states as tonic rules and phasic rules, with neuronal leaky integrators defined by a time constant to describe how fast the caused change in states reaches its result value.) One nuclei functions as homeostasis mechanism, whose reference is given by the output of phasic rule for reference nuclei (see also Figure 3). The last nuclei works as a math model to relate the Type B interventions and the change of the state. The output of the integration neuron in the rule-state layer is the state X, which then along with inputs are mapped into output Y. The outcome J is a function of all inputs, states, and outputs.
y i = f (X , M U , M X ) (1)
where X are state signals, M U are the values of membership-neurons based on fuzzy input-MF mapping, M X are membership-neuron values for fuzzy state-MF mapping.
The function f can be a Sujeno fuzzy inference system [11] or a weighted sum, selected by the user-designer. Depending on the application and the user-designer's intent, the output can be treated as a fuzzy or crisp value.
When output predictions are of measures that can be experimentally sampled, the user can determine an error signal. Such sampled errors can be viewed as a form of corrective "context" input that can be used to help tune future states and outputs.
Rules and State Layer: Nuclei Generating Differential Equations
States in this model are fuzzy linguistic variables that are dynamic estimators of physical, physiological and/or psychological states of the human body, of body impairments and of risks. They are modelled as dimensionless signals that can change value as a function of time, based on rules designed within a fuzzy expert system that serve to set up the dynamic state equations that are implemented as a CNN. The rule-state layer consists of a nuclei (cluster of neurons) for each state (see Figure 2), with each nuclei essentially implementing a nonlinear differential equation for that state that can also include recurrent connections from all states, including self-connections.
The fuzzy inference ("expert") system (FIS) consists of a left-half side (LHS, also called "if" or "antecedent" side) and a right-half side (RHS, also called "then" or "consequence" side). As is conventional for a FIS [11], each linguistic state variable has one or more fuzzy sets (represented by a linguistic "value") that are characterized by associated membership functions (MFs) over the variable's Universe of Discourse, such that a state membership value (M X ) represents the "degree of membership" of the state variable x in a fuzzy set (linguistic value), or the "degree of truth" that "x is value." The result is a number on the interval <0,1>, where "1.0" is full membership. Each rule may include any combination of state memberships (M x ) and input memberships (M u ) on the LHS, and must include a state membership value calculation (M x ) on the RHS that indicates how the state would change. Classic fuzzy operations (AND, OR, NOT) and hedges (VERY, MORE-OR-LESS) are supported, and easily added to rules through an interactive GUI. The end result is that the LHS provides a "strength" of firing for the state-change operation(s) described on the RHS.
Of note is that while the logic of the FIS is a function of the states x and input effects u * occurring at the same time iteration and thus is a nonlinear static mapping, there are dynamic operations both after and often prior to this FIS operation. The form of the RHS determines the manner of desired change in the state. Rule consequents that target the absolute value of the affected state are implemented by tonic-neurons, while rule consequences that target a relative positive or negative change in state are implemented by phasic-neurons. The dynamic effect of the FES on a state is determined by which of two classes the state is associated with, as is now discussed.
1) Group I: Conventional Fuzzy States
Conventional states change over time based on one or more rules. For one state x s, normally the spontaneous recovery procedure is:
where x r is the new drive, based on weighted consideration of the current strength of rules associated for a given state, as implemented by the state's nuclei. The time constant τ represents first-order dynamics.
There is also a FIS associated with dynamically changing the time constant of the rules as a function of states and inputs on the LHS. This is a feature that needn't be part of the user-designer's strategy, but is really quite a powerful addition since it makes available a range of possibilities for state transition dynamics. For instance, the popular Michaelis-Menten kinetics [12] and various cell growth laws [13] can be mathematically viewed as state-dependent variable time constants (inverse of rate constants) that represent special cases of the menu of possibilities.
While all linguistic states can be treated as dimensionless fuzzy signals with first-order dynamics that use a variable time constant that can also be set by a fuzzy rule, based our experiences and those of students using versions of the model in courses, there is also a need for another class – homeostatic states, which are described next. Examples of states that are inherently non-homeostatic are pain, skill, balance and risks.
2) Group II: Homeostatic fuzzy states
While conventional fuzzy signals can always be used when evidence and/or expertise is available, our experience has been that many states are not well captured by first order dynamics because they are part of more involved internal body processes. Thus many physiologic and functional states of the body, including both measurable signals and linguistic variables, are part of inherent homeostatic systems. For instance, physiologic measures ranging from body core temperature to heart rate are regulated, and after a tissue injury there are intrinsic healing mechanisms that aim to minimize the degree of impairment. All these states are controlled by a negative feedback loop. Thus this class of states can include nearly all physical and physiological signals, from blood pressure to muscle strength.
In determining the modelling strategy for such states, it is important to recognize that the user-designer's experience is typically with the closed-loop system, with no real knowledge of open-loop dynamics. Thus a challenge is to extract closed-loop knowledge of temporal dynamics and reference state to implement elements within the framework of a "plant" and "controller," and a reference ("set-point") input that itself can change through an intrinsic remodelling process. The current algorithm for how the homeostatic states maintain their equilibrium under the effect of different kinds of inputs is demonstrated in Figure 3, and includes a PID (proportional-integral-derivative) controller to represent the real capabilities of neurons for neural differentiation (e.g., primary muscle affects) and neural integration (e.g., brain stem interneurons). For any homeostatic state, there are two values in this model: the reference and the actual dynamic state.
Figure 3 The structure of nuclei for reference and homeostasis. A fact event can changes the reference via its own FIS (Rule Type A), and the change will be added to the reference through a first order system with a certain time of delay. When a context event happens, it will affect the reference in the same way as fact events. When there is an intervention, its frequency at the point will be calculated based on the history by a frequency calculator. A user-defined mapping function will then be applied to calculate the change. The mapping function maps the frequency and intensity of the intervention and the initial reference value into the result change. Then the change will be added to the reference through a first order system with a certain time of delay. The mapping function is defined by the user as two tables. If the frequency or the intensity is not in the table, the result change will be calculated by interpolation. All the result changes on the reference of one state caused by different inputs will be summed together by fuzzy OR operation, and then applied to the reference value. Users are encouraged to change references slowly and conservatively. The homeostasis nuclei sense the state value and compare it to the reference. Its output is sent to the integration neuron in the rule-state layer. In homeostasis nuclei, each path in control part and nonlinear paths and the feedback path can be turned on/off by the user. The fuzzy OR operation is used to assure the stability.
The reference is the value that represents the homeostatic "ideal" for the human body. If, for any reason, the actual state value deviates from the reference, the controlling organs such as the nervous system and glands will, by sending control signals, try to drive the actual state value toward the reference. Homeostatic references may change under the effect of both internal and external factors. Internal factors include developmental growth and the aging process. External factors include trauma causing impairment and/or lifestyle changes. When intrinsic homeostatic recovery processes are not successful or lifestyle changes are sustained, certain states may gradually adapt to a new reference.
Often D-action is zero unless there is an initial sharp response to a sudden input effect. In such a case an initial closed-loop time constant provided by the user-designer relates primarily to P-action. There is often then a slower drift toward homeostasis and/or remodelling, which can be used to estimate I-action and slow (near-permanent) transitions in reference.
As seen in Figure 3 this model contains two parts: the subsystem for the actual state value and the subsystem for the reference, both of which work as a feedback control system. In the former, the human body senses the actual state value and compares it to the reference. The error between them is the input to the control part, which represents the neural system and glands. The fuzzy OR operation is used as summation because of the physical limitation of the control signal. After the first order plant, the model supports nonlinear paths to capture plant-based nonlinear characters such as time delay or saturation (e.g., a fact event of injury may cut off or activate some specific nonlinear rehabilitation pathway); at present this has not yet been used, and research on optimizing the homeostatic feedback process continues.
To summarize, users specifying "homeostatic states" need only provide general closed-loop temporal and steady-state behavior, and a reasonable but conservative homeostatic regulator is automatically implemented.
Pragmatic Consideration: Separate Use of the FIS for Other WIA Modules
While the rule structure in the model is set up for addressing changes in dynamic states within a FIS framework, static rules and crisp logic are just special cases where the post-FIS time constant is zero and MF's have a hard boundary, respectively. Thus a WIA could also use this model, for instance, to create a separate FIS module that uses simpler, conventional real-time crisp logic, where states-to-output mapping is trivial (states equal outputs) or serves to perform aggregation/defuzification.
Input Layer: Classification and Implementation
Operations within the input layer depend on the type, with inputs classified into facts, contexts, and interventions. This layer can be viewed as a collector and pre-conditioner of inputs, designed to help map them to fuzzy "input effects" that are used in the rules that determine the state equations. Options include pre-filters such as physical models (e.g., a pharmokinetic model for Intervention Type-A (medications)) that are implemented prior to mapping to the fuzzy linguistic world via MF's that are associated with the input's fuzzy values.
In general, MF's are defined by two parameters that define either Gaussian and boundary (sigmoidal) shapes (states also have a monotone option). While these shapes provide continuous derivatives (good for many CNN algorithms), the boundary option does support the special case of a hard (crisp) boundary.
Facts
FIS systems often call their inputs "facts." As used here, facts are linguistic variables with a universe of discourse (range) that can be turned on but not normally turned off. In rehabilitation and sport medicine, these are often associated with the patient healthcare record, and include demographic information (e.g., age, gender, education level) and the occurrence of some diseases and diagnosis information (e.g., severity and localization of an event such as a stroke; co-morbidities). Each fact variable has at least one associated fuzzy linguistic value (each with an associated MF on <0,1>).
The relations between inputs such as facts and states are represented within fuzzy rules in the FIS, as describe previously. However, before a fact-event is used in the FIS, it is first mapped within the early part of rule-state nuclei into a "fact-effect" by a first-order time constant selected by the rule-designer (with default value of zero). Since a fact-event provides a step change (and thus a fact-effect a first-order step response), if one fact-effect was the only input on the LHS (i.e., a "fact-effect is value" yielding a M u number), the overall state change would be up to a second-order (overdamped) step response (one time constant before the FIS calculation that maps the "input event" to an "input effect" and is associated with the rule, and one after that is associated with the state). Individual facts thus can trigger rules to fire and cause changes in values of certain states, and possibly changes in the state's time constant and/or the reference value if the state is a homeostatic state (see Figure 3).
Context Inputs
Contexts are inputs that can be turned on or off, and make event-based "context awareness" available to the FIS for state estimation [1]. Normally they relate to external environmental events that can have an impact on the state of the person, but there are no limitations placed on context inputs. For instance, in stroke rehabilitation the clinical prognosis is a function of factors such as the ongoing degree of supports (e.g., social, caregiver, family), the clients diet and other nutritional concerns, the location and type of rehabilitation that is available, the client's normal daily or weekly life events, variation in their degree of motivation or ability to achieve lifestyle modifications, assistive technologies that are available to support independence, and so on. All can be viewed as context inputs, as can some interventions as long as the user-designer doesn't desire to use the types of more sophisticated mappings discussed in the next parts of this section.
Context inputs are important for WIA's, and are often used in tandem with state estimates for WIA decision-making. To some extent, they can be viewed as "temporary facts" that help sculpt rules, often weakly but occasionally strongly. Often they help add robustness and integrated realism to the rules and thus state estimation.
The form of the relations between contexts and states are the same as that between facts and states, except that the effect is a pulse (rather than step) response. The change of the status of one context (from off to on, or from on to off) is treated as a context event, which in turn may cause rules to fire differently.
Interventions
Interventions are a purposeful procedures and techniques aimed at producing changes in the condition consistent with the diagnosis and prognosis. Interventions may occur regularly or irregularly. Relative to the temporal dynamics of adaptive change, interventions can usually be viewed as impulses to the system. While interventions can always be treated as context input events of short duration, it is useful to develop evidence-based customized approaches for dealing with certain classes of interventions that are common in rehabilitation.
Although one individual intervention often only brings an "impulse response" change to state values because of length of time required for adaptive remodeling, available evidence or professional expertise may be available that indicates that a series of one type of intervention – a treatment "dosing" plan such as three sessions per week – may gradually change the reference value since the human body is an adaptive system. Often scientific studies provide evidence of remodelling based on a global dosing algorithm that is maintained for weeks or months. Adaptation thus can be due to the integration of the responses of the body to each intervention, and to slower intrinsic changes in homeostatic reference values. Based on the mathematics used to mapping intervention inputs to the effect on states, interventions are currently classified into three types.
1) Type A: Medication
This type of intervention supports both oral and injected medications or special dietary measures. In order to describe the effect of a medication, pharmacokinetics (the study of the bodily absorption, distribution, metabolism, and excretion of drugs) and pharmacodynamics (the study of the time course of pharmacological effects of drugs) are included in this conventional (non-fuzzy) model that is implemented within the input layer. The common methods in pharmacokinetics, which are consequently used in this model, are compartment model and Michaelis-Menten kinetics [12]. There are several different mechanism-based pharmacodynamics models [14], each applicable in certain conditions. Essentially, pharmacodynamics is the mapping between the concentration of certain drug and its "effect" on the state. Therefore, fuzzy logic as a very powerful non-linear mapping tool is adopted to implement the pharmacodynamics in this model.
As shown in Figure 2, when there is an event of medication, at first it is mapped into a time series, which represents the concentration of that medication in the blood or other destination spots, through a pharmacokinetics model. If it is an oral medication, a compartment model with two compartments (gut and blood) and Michaelis-Menten (M-M) kinetics are used. The former describes how fast the drug transfers from gut to blood, and the latter calculates the consuming velocity of that drug in blood. Assuming the mass and concentration in the gut are m 1 and C 1 and in the blood are m 2 and C 2 , the diffusion constant between the gut and blood is K , and the constants of M-M kinetics are V max and K m , the equations are:
If injected, only the M-M kinetics equation is applied. As part of a collaborative project with a post-doctoral fellow (Nicole Sirota, D.O.), estimated values have been tabulated for over 40 medications commonly administered by rehabilitation physicians. The concentration is then an input to a Tsukamoto fuzzy inference system [11,15] to determine the dynamic effect on target states, for use in the rule-state layer.
2) Effective Pulse Energy
Possible inputs of Intervention Type B include exercise, language therapy, recreation therapy, etc. In this type of intervention, a patient and/or provider provides inputs of magnitude and duration that have associated "energy" that is partially or fully "consumed" – the "effective" input. If subsequent changes in the affected state exhibit temporal dynamics that are long in relation to the time duration of the intervention, the input can be viewed as an impulse with an effective impulse energy; otherwise it is a pulse with a changing "effective" magnitude over its duration. In either case, how much energy is consumed in one intervention relates to whether the pulse energy becomes greater than an accumulation threshold energy, after which it triggers a first-order history-dependent recovery/refractory/fatigue variable that subtracts from the input until full effectiveness is gradually restored. Additionally, if another intervention event of the same type happens during the period of time before full recovery, the effectiveness of that event on states will depreciate. This type of intervention is thus mapped to an input effect that is then used to determine its effect on changes in the affected states. Research in this area continues, and details are not provided here.
3) Anticipated Intervention Types C, etc
It is anticipated that there may be dynamic effects of other interventions not yet modelled, which may be defined by users if evidence suggests dynamic processes (e.g., physical lumped-parameter or compartmental models) prior to mapping for use of fuzzy inference capabilities (e.g., functional electrical stimulation).
Results
Example Model #1: State Estimator and Output Predictor for Neurorehab Using Medication & Activity Interventions
This first example demonstrates the model's use in providing ongoing context awareness of a person's state, which is a critical need for future WIAs. A secondary purpose is to predict performance outputs and outcomes prognosis. There are two steps to the interactive design process: setting up the model, and running simulations.
Table 1 describes the inputs, states, outputs and outcomes for a hypothetical client, defined by a problem statement. Design of the system usually proceeds with a right-to-left flow, starting by identifying desired outcomes and performance outputs, and then determining the internal states that ideally would be estimated to determine these measures. However, for the type of context-awareness needed by WIA's, the WIA user-designer may have a need for certain specific state estimates, and there is no requirement that every state map to an output or outcome.
Table 1 Signals for Example #1.
Female client with stroke-induced disability a large-scale model with 16 types of potential input events, 12 states to estimate, 5 outputs, and 3 outcomes.
Problem statement: An older woman presents with stroke-induced disability (4 months post-stroke) that includes mild functional limitations to gait and posture, and significant impairment of the right arm and hand and of speech production. She also presents with mild osteoarthritis that affects her hips and knees. Released from outpatient care and living alone, her current "prescriptions" include three types of medication doses (for general joint and skeletal health, for pain from arthritis, and for spasticity), and three types of activities suggested by her former therapist (walking/cycling, hand operation, and oral communication). She also has two important weekly events: a visit most Sundays from her daughter (who is a nurse), and a visit most Tuesday's to the local community center (transportation is provided). She regularly uses a PDA-cellphone and a desktop computer (both set up by her other daughter who is an engineer, but lives in another state), and prefers to use an IP videoconferencing package to tele-visit with either of her daughters. Thus she is a good candidate for an assistive WIA.
Inputs (and MF example) States (and MF example) Outputs Outcomes
Facts:
- Age (is old)
- Initial Stroke (is severe)
- Osteoarthritis (is mild)
Contexts:
- VisitDaughter (is full)
- VisitCommCenter (is full)
- LocationByGPS (is outside)
- TeleVisitDaught (is active)
- TimeOfDay (is morning)
- NovelEvent (is negative)
Interventions (Meds or Activity)
- PillsOsteo (is right-dose)
- PillsPain (is high-dose; conc)
- PillsSpast (is 2-pills; conc)
- Walking (is good)
- Cycling (is good-quality)
- Speech (is good-duration)
- Keyboard (is good-session) Degree of Impairment:
- Gait (is faster)
- Balance (is better)
- RightArm (is worse)
- RightHand (is better)
- Speech (is improved)
Physiologic:
- RestingHR (is higher)
- RestingBP (is higher)
- BoneJointHealth (is low )
Other ("Degree of ..."):
- Pain (is high)
- RiskFalling (is high)
- Motivation (is high)
- SleepAtNight (is restful) Communication [Φ (Speech, Pain)]
HandROM [Φ (Hand)]
FIM [Φ (Arm, Hand, Balance, Speech, Pain)]
RiskFracture [Φ (BJ-Health, Risk-Falling)]
Adherence [Φ (Motivation, Pain, Sleep)] GenHealth [Φ (all impairment \physiologic states)]
Participation [Φ (Communication., Gait]
QualityLife [Φ (Weekly-Pain, FIM, Speech, Gait, Adherence, Hand-ROM)]
Notes: while one MF value is shown for each input or state, typically there are additional ones. Use hedges such as "not" or "very" or "more-or-less" can lower the number of MFs (and thus parameters) associated with a linguistic variable.
Key abbreviations: MF: membership function; GPS: Global Positioning System; HR: heart rate; BP: blood pressure; FIM: Functional Independence Measure [21].
The desired outcomes are in this case to be maximized. Outputs are performance measures that are a function of several states (e.g., FIM score) and/or represent a predicted measurement based on a state (e.g., hand ROM is one measure of hand impairment). Dynamic state behavior is fully dependent on the rules that map current inputs and states (LHS) to changes in states (RHS).
Inputs are mostly pre-determined, based on practical considerations of available data and events that can be sensed or entered by the user. In this case of a WIA application for a "chronic" case that is customized to the specific user-client, it is likely that the user-designer implicitly assumes the effects of fact-events have already played out and can focus on rules involving context-inputs and intervention-inputs. In other scenarios or for more generic population-based rule sets, facts would be a part of rules.
To illustrate how rules creation by a user-designer works, a GUI for a representative subset of states from Table 1, and their associated rules, are provided in Figure 4 and Figure 5. Notice that for most rules there is one state-MF on the "then" (right) side, and more than one input-MF and/or state-MF on the "if" (left) side. State-MF's on either side can take express the linguistic expression of a "tonic" neuron "(state is high)" and/or a "phasic" neuron "(state is higher)"; the state "pain," for instance, uses both. Also notice that the rule operates on an "input effect" which is the input mapped through a gain and time constant (see also Figure 2); this allows a rule such as for an impairment state, where changes happen slowly, to integrate context inputs so their effect extends well beyond the time that they are actually on, for that particular rule. Figure 5 also shows that a state, such as pain, can be a function of several rules (e.g., one more related to context inputs, the other medications) that combine through fuzzy operations. Finally, the intrinsic time constant for each state-neuron, differs dramatically between states (e.g., higher value for impairment which changes gradually over weeks versus a measure such as "pain" that can change on the order of hours). These affect logic development. The user-designer needs to understand several features that affect rule design.
Figure 4 The interface the inputs, states and outputs in model #1. There are three facts (bottom left), ten contexts (up), three medications (bottom left two), twelve states (bottom right two) and five outputs (bottom right). User-designer can add or delete inputs, states, outputs or outcomes. For a selected variable, the user-designer is able to set the range (min and max), add/delete membership functions, define the membership functions, and see the graphics of the membership functions. If the variable is a state, the user-designer also has access to the reference, time constant, the negative feedback (on/off), and all of the control parameters.
Figure 5 Type C rules and type D rules in model #1. There are six types of rules (RA to RF) based on what kind of relation they represent between inputs and states. For example, RC (back window) describes how the facts/contexts change the states' values directly, and RD (front window) defines the relation between medication and affected states. The user-designer can add, delete and change rules, and also change the properties of rules such as rule name and rule weight. When working on "If" side or "Then" side, user-designer can add an element, or delete the right-most element, which is demonstrated helpful for designing rules. In the "If" part, several options are available to help define precise and flexible rules. These operations include: "AND" and "OR" operations, constraints (such as "NOT", "VERY", and "NOT VERY"). There is also a weight for each input element.
An example simulation, with a few weeks of inputs and with states initialized, is presented in Figure 6. Here we focus on the "context" (state estimates) based primarily on "context" inputs, and on a certain slice of time – the "present." Five conceptual points are emphasized here:
Five conceptual points are emphasized here:
1. State change often requires that a combination of input/state conditions occurs, and certain states can change suddenly (e.g., pain) while others only gradually (e.g., impairments, and homeostatic states in general).
2. As with most large-scale nonlinear systems, the "functioning" order of the system, and overall behavior, tends to be only a function of a small subset of the model parameters. As different events fire, different "subsystems" of changing states emerge and the collection of parameters with the highest sensitivity changes. Sensitivity analysis tools, embedded in the model, can be used to gain insight into what parameters matter most at any given time.
3. Many of the expressions on the left-half ("if") side of the rule may be designed to add robustness – they rarely affect the rule, but when they do they effectively drive or shut off firing.
4. While the primary need of WIAs is for context-awareness of state, the fact there is also prediction of the future can potentially be used by a WIA to consider the effects of alternative plans for future events. This may be especially useful for WIA decision-making while functioning in "assist" mode, as there is a fine balance between the benefit of providing a reminder or warning to the client versus the cost of overburdening the client; "what-if" prediction of the future can help in making this decision.
5. While not shown in Table 1 (but evident in Figure 1), of note is that "errors" between a predicted output and measured output (e.g., arm force, hand ROM, FIM and recent pain or adherence can measured by the daughter during weekly visits) can be fed back as context input events that can then be integrated into a rule for a state, helping gradually improve the state estimate.
Example Model #2: Muscle Force and Joint Strength Changes: Short-Term Fatigue and Long-Term Adaptation
This example illustrates use of the model by a user-designer who has expertise in a certain area plus access to scientific evidence, here demonstrated for muscle strength. One can easily envision an athlete or coach using a WIA to plan and implement an exercise program that has the desired outcome of maximizing muscle strength and tissue hypertrophy over a certain time window, and has estimates of relevant internal states during the process. Similarly, one can envision a musculoskeletal or neuromuscular rehabilitation program that seeks to regain muscle strength or minimize muscle atrophy. In either case, the estimated states and sampled performance output measures can help a WIA to provide a user-client with a suggested input intervention program (e.g., exercise regimen, diet).
This example also exposes another use for the model: by scientists who study bio-change, and in particular who desire to synthesize knowledge of macro-and micro-changes at the organ/tissue and cellular levels, to make model predictions that may be testable, and to bridge human macro-studies with animal micro-studies. Here the onus is on the expert to integrate experience and available evidence. One of us (JMW) has published extensively using neuromusculoskeletal models that include Hill muscle models [16-18]. Hill-based muscle models predict force as a function of muscle activation, length and velocity. In traditional use of such models, parameters are assumed constant for a given simulation. But we know that some parameters do change as a function of activity, and in recent years a growing body of evidence has amassed on how three key parameters [maximum isometric force (Fmax), maximum unloaded velocity (Vmax), a Michaelis-Menten kinetics parameter related to calcium deactivation] change as a function of: i) fatigue (a shorter-term reversible change in parameters over a time period of seconds to hours); and ii) true muscle adaptation (a more permanent change that occurs over time periods of days to months).
Table 2 focuses on a simple model structure for estimating one of these parameters: Fmax, which also directly correlates to muscle strength and size. It does so on two timeframes, using different models: i) for fatigue, the model runs for minutes, with rules structured on the assumption that an exercise "pulse" corresponds to the intensity (percent of maximum) and duration (number of repetitions) of a weight-training "set"; and ii) for adaptive change, the model runs for weeks or months, and an exercise "pulse" is the average intensity of a "workout" where a time of an hour is small relative to the dynamics of adaptive tissue change. In both cases these are "converging" models with many inputs; Table 2 keeps these inputs simple. Figure 7 provides an example simulation, here for a client with a sedentary lifestyle who makes a number of positive lifestyle changes but then, after nearly four weeks of training and some improvements in Fmax, gets injured.
Table 2 Example of a converging model designed to estimate states, outputs and outcomes.
Inputs States Outputs Outcome
Fatigue Model:
Facts:
• InitFiberComp Context:
• Motivation Interventions
• SetHiRep60%
• SetLowRep80% CalciumConc
Pi/H+-Conc1
PossibleReps@80%
PossibleReps@60% PredictedRepsAt60%
PredictedMax Fmax
Adaptive Model:
Facts:
• InitFiberComp Context:
• Diet
• Injury
• GenActivityLevel Interventions
• WeightSession
• AerobicSession
• PillsSteroids Hypertrophy
Atrophy
FiberComp
MuscMass PredictedStrength
PredictedPower Fmax
Vmax
1 Phosphate and pH concentration, which reflect muscle energetics and transient recovery dynamics; for Fmax, evidence suggests these are are similar in influence (if we added Vmax to the model, available scientific evidence would suggest we separate these states).
Figure 6 Example simulation result of model #1. The simulation period is from Feb. 09 2005 to Mar. 09 2005 (see top left). This figure shows the event train of TeleVisit in the input frame (up left), the status of Speech (bottom left), the rule firing rate of SpeechContext (up right), the output Communication (middle right) and the outcome Participation (bottom right). In the state frame, the blue line is the curve without medication and the red line the curve with medication. Comparing the event train of TeleVisit and the curve of the Speech, we can see clearly the effect of every TeleVisit on the Speech. The other protuberance on the Speech curve is caused by the visit to the local community center on every Tuesday.
Discussion
This paper develops a novel rule-based neuro-fuzzy dynamic model that is intended to provide continuous state estimation, to predict outputs, and to evaluate the effect of different intervention plans. It enables a user-designer who is an expert in a rehabilitative field, but not necessarily in mathematical modeling, to generate and use causal models that contain underlying nonlinear differential equations implemented with a CNN. To be effective they need to have a solid understanding of the concepts of a time constant, a weight (or gain), how a MF maps a variable, and how negative feedback works; the interactive GUIs can actually be used as a learning tool to help pick up these skills. When creating new inputs and states, default MFs for classic linguistic values such as "high" or "low" are automatically created for the user-designer, using either Gaussian or Boundary fits that are defined by two intuitive parameters – a "middle" and a "shape." These default MFs can be renamed, edited or deleted. The GUI for rule creation, shown in Figure 5, is similar to GUI's for other FIS implementations, only with some added capabilities that are unique to this model. Once the model structure and parameters are set, simulations are managed on a separate GUI that enables the creation of input trains, state initialisation, and other bookkeeping features. In addition to plotting state, output and outcome trajectories, a GUI is available for parameter sensitivity analysis. Once refined to the satisfaction of the user-designer, the model is ready to be used as an embedded application.
The first example shows the application of this neuro-fuzzy modelling framework as a WIA to estimate the current states values and predict key outputs and outcomes. There is no doubt that real-time monitoring of certain key states is valuable for some patients/clients in their daily lives. However, not all of these states can be directly measured by wireless sensors. SoftBioME provides the possibility for an integrated WIA in which wireless sensors measure the measurable events and states and send them to the intelligent agent, while the intelligent agent records these states and estimates un-measurable states, detects the occurrences of pre-defined events (e.g., a state is above or below certain value), and does some pre-processing. This example demonstrated that SoftBioME can provide an estimate of certain states' values at any time. If the user-designer defines crisp MFs and crisp rules, it can also serve as an event-detector. Since the model is designed based on expert knowledge (e.g., how the walking exercise affect the gait) and scientific evidence (e.g., the effect of medications on states), the estimation error shouldn't be beyond expectation. In addition, rules can include error signals on the LHS that are based on any difference between an estimated variable (output, outcome) and periodically measured signals (output, outcome), enabling state estimation to improve over time. Furthermore, by running the simulation repeatedly, an experienced user-designer may adjust the parameters to try to heuristically optimize a customized model before use for real time estimation. The CNN model structure is designed so that in the future a neuro-optimization toolset can be provided to improve the model performance for a certain client, i.e. to "learn" the client's behavior. All of the above promise an accurate-enough estimation for the type of context-awareness that is needed for effective WIAs.
Unlike all the use of macro-states in the first example, the second example contains both macro-states and micro-states. In this example, the macro-states depend on the micro-states and macro evidence from strength training and visa-versa, and that dependence can be described by fuzzy rules. The muscle force model also demonstrated that the model created in SoftBioME can not only estimate states, outputs and outcomes, but also focus on parameters changes. That's because one of the purposes of SoftBioME is to support both signal models and parameter models (e.g., longer-time remodelling models). The parameters in a signal model may simultaneously be the signal in a parameter model, with the two models operating on different time scales (e.g., seconds versus weeks). For example, for some exercise activity performed frequently, Fmax is the signal in the second example and is also a (now adaptive) parameter in the first example (e.g., neuromusculoskeletal model using Hill-based muscles). The ability to work at both signal dimension and parameter enables SoftBioME to deal with a variety of problems in a broad area in rehabilitation.
When designing an intelligent agent through SoftBioME, the most critical thing is to collect expert knowledge and/or scientific evidence. There are several ways to collect expert knowledge, such as Analytical Hierarchy Process (AHP) [19] and Delphi [20]. The latter is often used by doctors and nurses as decision making protocol, which makes it a good choice when creating a rehabilitation model. Published paper and textbooks are the main sources for scientific evidence. Given available expert knowledge and scientific evidence, how to transfer them into membership functions and fuzzy rules is the next challenge. Normally experts will help define MFs and fuzzy rules. If creating an evidence-based model, usually the evidence itself contains the rules implicitly (e,g, abstracts often summarize findings in the form of rules). Sensitivity analysis tools can help refine the MFs and evaluate the importance of rules, which assist the user-designer in improving the model.
Of note is that because of the natural tendency for signal and rule "soft saturation" when using fuzzy models with smooth MFs, the nonlinear differential equations tend to be inherently stable. Most biosignals also have soft saturation at the extremes of their operating range. It is as if the user can think in a more rule-based "linear" and causal manner, but end up with models that, if well designed, are robust over a larger region of the operating state space than for a linearized version of a bio-model.
Conclusion
A neuro-fuzzy modelling framework (SoftBioME) is developed for estimating changes of states in bio-systems as a function of input event patterns. If carefully designed with sufficient expert knowledge and/or scientific evidence, it can be applied in rehabilitation (e.g., predict intervention outcomes), sport medicine (e.g., evaluate the effect of a training plan), biology (e.g., adaptive changes in muscle), pharmacy (e.g., study the action or effect of drugs), and perhaps other fields whose subject is a dynamic system and adaptive change. It is able to make predictions or real-time estimation. The latter is intended to provide context-awareness of changing states, which is critical for WIA's to be effective.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Both YW and JMW were involved in all parts of this work, with YW responsible for model implementation and most simulations, and JMW responsible for most of the first draft of the manuscript.
Figure 7 Example simulation of the "adaptive model" of Table 2. For this simulation, the following four simple rules were used, one for each state: IF WeightSession is intense & (Diet is good & Hypertrophy is Low) THEN Hypertrophy is high & higher IF (GenActLevel is low & AerobicSess is not intense) or Injury is bad THEN Atrophy is high & higher IF WeightSession is intense & AerobicSess is not intense & FiberComp is low THEN Fibercomp is high & higher IF Hypertrophy is high & (WeightSession is not intense & Diet is not good & AerobicSess not intense THEN MuscMass is high & higher Since only one input, state, rule, etc can be shown in an image (user can easily toggle between them), others are described here. At the start the client has states that reflect a sedentary lifestyle. Inputs reflect that he gradually increases his general activity level (this is the input that happens to be shown), improves his diet, and starts a weight-training program. This continues for three weeks through the end of February, at which time he stops the weight training and starts an aerobic training program. However, on his fourth aerobic event, he gets injured and his activity decreases. The hypertrophy and atrophy states are viewed as bioprocesses that are always somewhat present, and compete with each other. Of the four states, the hypertrophy state is shown (lower left), and we see an initial rise and a subsequent mild effect of each weight training session. After these inputs stop the state falls a bit. The atrophy state follows the shape of the atrophy rule, which is shown (upper right). Notice that with increases in various activities, atrophy rule firing decreases until the injury occurs. The output (predicted strength) is assumed a weighted function of all states, and the "outcome" Fmax (which could have also been viewed as an output) is a weighted function of the predicted strength and some of the states. Both show increases with these lifestyle changes, then the start of a decrease after the injury.
Acknowledgements
The financial support from The Ralph and Marian Falk Medical Trust Foundation, The Whitaker Foundation, and the Rehabilitation Engineering Research Center on Telerehabilitation (U.S. Department of Education, NIDRR #H133A990008) is gratefully acknowledged. The opinions are those of the authors
==== Refs
Winters JM Winters JM, Robinson C, Simpson R, Vanderheiden G Emerging rehabilitative telehealthcare anywhere. Was the Homecare Technologies Workshop visionary? Emerging and Accessible Telecommunications, Information and Healthcare Technologies 2002 Arlington: RESNA Press 95 111
Winters JM Herman W Report of the Workshop on Home Care Technologies for the 21st Century Catholic University of America, Washington DC Tech Rep HCTR-10-v10 2000 96
Winters JM Telerehabilitation research: Emerging possibilities Annu Rev Biomed Eng 2002 4 287 320 12117760 10.1146/annurev.bioeng.4.112801.121923
Winters JM Rosen MJ, Lauderdale D Mobile telerehabilitative healthcare technologies Proceedings of the state of the science conference on telerehabiliation and application of virtual reality 2001 Washington DC: NRH Press 102 108
American Physical Therapy Association Guide to Physical Therapist Practice 1997 ch1
Law M Evidence-Based Rehabilitation: A Guide to Practice, Slack Incorporated 2002
Carr JH Shepherd RB Neurological rehabilitation: Optimizing motor performance Oxford: Butterworth-Heinemann 1998
Wang Y Winters JM An Event-Driven Dynamic Recurrent Neuro-Fuzzy System for Adaptive Prognosis in Rehabilitation Proc 25th EMBS Conference, Cancun, Mexico 2003 1256 9
Wang Y Winters JM An Input Classification Scheme for Use in Evidence-Based Dynamic Recurrent Neuro-Fuzzy Prognosis, Proc IEEE/EMBS, San Francisco 2004 3198 3201
Winters JM Lathan C Sukthankar S Pieters TM Rahman T Winters JM and Crago PE Human performance and rehabilitation technologies Biomechanics and neural control of posture and movement 2000 New York: Springer-Verlag 493 551
Jang JSR Sun CT Mizutani E Neuro-Fuzzy and Soft Computing. A Computational Approach to Learning and Machine Intelligence Prentice Hall, Upper Saddle River, NJ 1997
Ch 13
Palsson BO Bharia SN Tissue Engineering, Pearson Prentice Hall, Upper Saddle River, NJ 2004
Mager DE Wyska E Jusko WJ Diversity of Mechanism-based Pharmacodynamic Models, Drug Metabolism and Disposition 2003 31 510 518 12695336 10.1124/dmd.31.5.510
Tsukamoto T An Approach to Fuzzy Reasoning Method, Advances in Fuzzy Set Theory and Applications, North-Holland, Amsterdam 1979 137 49
Winters JM Stark L Analysis of Fundamental Movement Patterns Through the Use of In-Depth Antagonistic Muscle Models, IEEE Trans Biomed Engng, BME 1985 32 826 839
Winters JM Winters JM, Woo SY Hill-Based Muscle Models: A Systems Engineering Perspective, Chapter 5 in Multiple Muscle Systems: Biomech and Movem Organiz 1990 Springer-Verlag, New York 69 93
Winters JM An Improved Muscle-Reflex Actuator for Use in Large-Scale Neuromusculoskeletal Models, Annals of Biomed Engng 1995 23 359 374
Spires EE Using the Analytic Hierarchy Process to Analyze Multiattribute Decisions, Multivariate Behavioral Research 1991 26 345 361
Keeney S Hasson F McKenna HP A critical review of the Delphi technique as a research methodology for nursing, Int J Nurs Stud 2001 38 195 200 11223060 10.1016/S0020-7489(00)00044-4
Finch E Brooks D Stratford PW Mayo NE Physical Rehabilitation Outcome Measures: A Guide to Enhanced Clinical Decision Making, BC Decker 2002 144 148
|
15985181
|
PMC1182386
|
CC BY
|
2021-01-04 16:37:40
|
no
|
J Neuroengineering Rehabil. 2005 Jun 28; 2:15
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-15
|
oa_comm
|
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-131594903910.1186/1476-511X-4-13HypothesisDoes cholesterol act as a protector of cholinergic projections in Alzheimer's disease? Bohr Iwo J [email protected] University of Newcastle, Department of Neurology, Neurobiology and Psychiatry, Institute for Ageing and Health, Newcastle General Hospital, Westgate Road, Newcastle-upon-Tyne, NE4 6BE, UK2005 10 6 2005 4 13 13 12 5 2005 10 6 2005 Copyright © 2005 Bohr; licensee BioMed Central Ltd.2005Bohr; 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 relationship between Alzheimer's disease (AD) and progressive degeneration of the forebrain cholinergic system is very well established, whereas mechanisms linking this disease with cholesterol, apolipoprotein E (apoE) phenotype, and amyloid precursor protein (APP) metabolism have not been fully elucidated even though there is a plethora of publications separately on each of these issues. The intention of this hypothesis is to unify knowledge coming from all of these areas. It is based on an assumption that the process of APP hypermetabolism is a neuroprotective response for age-related cholinergic deterioration. In some individuals this initially positive process becomes highly overregulated by genetic or/and epigenetic risk factors and after many years of accumulations lead eventually to AD. I hypothesise that neuroprotective role of APP-hypermetabolism might be related to enrichment of neuronal membranes (lipid rafts in particular) in cholesterol in order to compensate for decrease in presynaptic cholinergic transmission and/or AD-related decrease in cholesterol levels. The above is consistent with findings indicating that activity of both muscarinic and nicotinic cholinergic receptors is correlated in a positive manner with cholesterol plasmalemmal content. Briefly – APP metabolism together with transport of cholesterol in apoE containing lipoproteins seem to play a key role in mobilising cholesterol into neuronal membranes.
==== Body
Background
The role of cholesterol in Alzheimer's disease (AD) is attracting increasing attention of researchers [1] and there are conflicting messages coming form a great deal of reports. Despite the fact that a wide-spread opinion about high levels of this lipid in the organism still remains negative, there is a growing body of evidence suggesting its beneficial role in the brain. It is for example corroborated by the study showing that high cholesterol blood levels correlate with a lower mortality index and a better outcome following a first stroke [2]. There was also a positive relationship between a hypercholesterolemic diet and improved preservation of cognitive functions in rats which previously underwent anoxic period [3]. The significance of the results obtained with the use of a dietetic paradigm has recently been confirmed by findings reporting a net flux of peripheral cholesterol through Blood-Brain Barrier in the form of 27-hydroxycholesterol [4]. It has also been found that patients suffering from AD have lower levels of cholesterol in cerebrospinal fluid[5] in the lipid fraction of brain membranes resulting in altered membrane physical properties [6] and recently in cholesterol-enriched lipid microdomains in plasmalemma – lipid rafts [7]. Moreover a relationship between AD and down-regulation of seladin-1; a protein involved in cholesterol synthesis was found, reviewed in [7] which may be due to a genetic disorder.
It appears that cholesterol has universal neuroprotective properties. However for the purpose of this article this activity will be described only in connection to AD.
The hypothesis combines within one unifying concept well established facts from the three following main streams of AD research:
- the prevalence of forebrain cholinergic system deficits in the disease development, – metabolism of amyloid precursor protein (APP) and
- the role of apolipoprotein E4 (apoE4) isoform as a risk factor in association with cholesterol metabolism in the brain.
These key issues will be briefly introduced before presenting the hypothesis.
The main type of neurons primarily affected by AD are these belonging to and innervated by the cholinergic forebrain projection. It is a neuromodulatory system related to high cognitive functions. Deficits in these commands are the first clinical manifestations of AD [8].
Forebrain cholinergic neurons and areas extensively innervated by them (hippocampus and neocortex) contain the largest amounts of senile plaques. According to amyloid cascade hypothesis, widely accepted by scientists, senile plaques are primary factors causing neuronal death in AD, whereas neurofibrillary tangles are secondary [9]. The major constituent of senile plaques are peptides called β-amyloids, products of enzymatic cleavage of APP. Formation of senile plaques is a result of many years of accumulation of β-amyloids and other peptides forming extracellular insoluble aggregates. However it appears that in shorter periods APP and its metabolites demonstrate neuroprotective activity. The increased deposition of APP is a relatively quick reaction to factors deteriorating brain functioning, see for example:[10]. Direct neuroprotective effects were shown in the rat hippocampus [11]. They also were reported to protect cognitive functions following application of anticholinergic agent [12]. Recently Koudinov and Berezov have reviewed evidence showing a positive role of β-amyloids in the brain [13].
Cholinergic neurons display particular vulnerability to any negative factors affecting brain function, see for example: [14-16]. Ageing is a major process causing chronic deterioration of brain functioning, the cholinergic system being particularly susceptible and inducing long lasting overproduction of β-amyloids. If this process is aggravated by genetic and/or epigenetic factors it may eventually lead to development of AD.
The importance of cholesterol in the brain functioning is suggestively reflected by the fact that the human brain making up only 2% of total body weight contains as much as 25% of the total pool of this lipid [17]. To a big extent it is concentrated in myelin sheath. However there are also considerable amounts also in neuronal plasmalemma and in lipid rafts in particular. Lipid rafts seem to play a key role in transmembrane signalling processes, including synaptic transmission [18]. Importantly, APP is suggested to occur in lipid rafts [19,20].
The hypothesis aims at explaining this positive role of cholesterol in the brain in association with geriatric cholinergic deficits and APP metabolism.
Presentation of the hypothesis
As suggested above age-related dysfunction in the forebrain cholinergic system results in APP hypermetabolism. However the mechanisms by which APP metabolites fulfil their neuroprotective functions despite some attempts, have not been fully elucidated. I hypothesise that these properties are due to the involvement of APP metabolites in the process of internalization of lipoproteins labelled with apoE, enriched with cholesterol. What could be the aim of this process ? There are data indicating a positive dependence of cholinergic receptors (both muscarinic and nicotinic) on cholesterol content in plasma membranes [21] and the mechanism of molecular interactions between cholesterol and nicotinic receptors have been proposed [22]. In contrast, receptors for monoaminergic agonists, seem to be negatively modulated by high membrane cholesterol [21,23]. In this respect increased uptake of cholesterol might be at least in part a process aiming at compensating cholinergic deficits and/or cholesterol deficits in AD caused for instance by genetic factors. Within the framework of this concept it is possible to explain the causal relationship between AD and a phenotype of apoE. The importance of apoE containing lipoproteins in neuroregenerative processes in connection to the role of cholesterol seems to be well established [20,24], although the role of APP and relationship of apoE-dependent transport with cholinergic transmission has not been fully clarified, despite some interesting proposals [24], which may be regarded as complementary to this hypothesis. There are three allele: ε2, ε3 and ε4 coding different isoforms of the protein. Expression of apoE4 isoform increases the risk of both sporadic and familial late onset of AD. Consequently one can assume that a higher demand of neurons for cholesterol results in higher production of β-amyloids which are engaged in internalization of apoE containing lipoproteins. Possibly interaction between the apoE4 isoform and β-amyloids in contrast to other isoforms is more prone to accumulation of insoluble aggregates and in addition might be less effective in cholesterol uptake as suggested in[20]. The relationship between levels of expression of APP metabolites, amount of apoE and cholesterol levels has been shown in several instances. An interesting example of such a relationship is provided by Howland et al [25] who carried out experiments exploring a mouse model of AD. In these mice exposure to a high cholesterol diet resulted in reduction of APP metabolites and concomitant increase of apoE. Similarly, it was reported that cells in culture exposed to high cholesterol reduced APP metabolism [26]. Reduction in APP metabolite production in conditions of high cholesterol in these publications may be explained by the negative feedback principle.
Testing the hypothesis
The best way to test the hypothesis might be by experiments combining all of its main constituents. For instance by inducing APP hypermetabolism, and then verify whether supplementation of high cholesterol would result in:
- lowering of APP hypermetabolism
- better survival of neurons in crucial areas like the forebrain cholinergic system, hippocampus and some neocortical areas
- better preservation of cognitive functions
Implications of the hypothesis
If the hypothesis is proved to be true it should first of all change negative attitude towards blood high cholesterol levels in clinical practice. In particular the use of statins in older subjects with neurological disorders should be revised. Recently there is an increasing number of reports indicating uncertainties related to this issue [20,27]. However this should be handled with care, since some authors even acknowledging the positive role of cholesterol in the brain do not exclude some beneficial actions of this group of drugs see e.g. [7]. Nevertheless, if the hypothesis is validated it may result in changes in some diet recommendations, especially while considering that definitely not in all cases high blood cholesterol must result in arteriosclerosis, this is supported by findings indicating homocysteine and not cholesterol as the primary vessel damaging factor in this disease [28], see also a strong criticism of the concept "blaming" cholesterol as a primary factor in the disease [29].
Moreover it would open up new alleys in brain function studies in general and AD in particular. It would imply the need of widening our understanding of the activity of cholesterol in the plasmalemma of neurons and mechanisms by which cholinergic receptors interact with plasmalemmal cholesterol.
==== Refs
Hartman T Cholesterol and Alzheimer's disease: statins, cholesterol depletion in APP processing and Abeta generation. Subcell Biochem 2005 38 365 380 15709489
Vauthey C de Freitas GR van Melle G Devuyst G Bogousslavsky J Better outcome after stroke with higher serum cholesterol levels. Neurology 2000 54 1944 1948 10822434
Bohr I Hypercholesterolemic diet applied to rat dams protects their offspring against cognitive deficits. Simulated neonatal anoxia model Physiol Behav 2004 82 703 711 15327920 10.1016/j.physbeh.2004.06.009
Heverin M Meaney S Lutjohann D Diczfalusy U Wahren J Bjorkhem I Crossing the barrier: net flux of 27-hydroxycholesterol into the human brain. J Lipid Res 2005 46 1047 1052 15741649 10.1194/jlr.M500024-JLR200
Demeester N Castrol G Desrumaux C De Geitere C Fruchart JC Santens P Mulleners E Engelborghs S De Deyn PP Vandekerckhove J Rosseneu M Labeur C Characterization and functional studies of lipoproteins, lipid transfer proteins, and lecithin : cholesterol acyltransferase in CSF of normal individuals and patients with Alzheimer's disease. J Lipid Res 2000 41 963 974 10828089
Mason RP Shoemaker WJ Shajenko L Chambers TE Herbette LG Evidence for changes in the Alzheimer's disease brain cortical membrane structure mediated by cholesterol. Neurobiol Aging 1992 13 413 420 1625771 10.1016/0197-4580(92)90116-F
Ledesma MD Dotti CG The conflicting role of brain cholesterol in Alzheimer's disease: lessons from the brain plasminogen system. Biochem Soc Symp 2005 72 129 138 15649137
Collerton D Cholinergic function and intellectual decline in Alzheimer's disease. Neuroscience 1986 19 1 28 3537837 10.1016/0306-4522(86)90002-3
Verdile G Fuller S Atwood CS Laws SM Gandy SE Martins RN The role of beta amyloid in Alzheimer's disease: still a cause of everything or the only one who got caught? Pharmacol Res 2004 50 397 340 15304237 10.1016/j.phrs.2003.12.028
Kalaria RN Bhatti SU Lust WD Perry G The amyloid precursor protein in ischemic brain injury and chronic hypoperfusion. Ann NY Acad Sci 1993 695 190 193 8239281
Smith-Swintosky VL Pettigrew LC Craddock SD Culwell AR Rydel RE Mattson MP Secreted forms of -amyloid precursor protein protect against ischemic brain injury. J Neurochem 1994 63 781 784 8035204
Meziane H Dodart JC Mathis C Little S Clemens J Paul SM Ungerer A Memory-enhancing effects of secreted forms of the -amyloid precursor protein in normal and amnesic mice. Proc Natl Acad Sci USA 1998 95 12683 12688 9770546 10.1073/pnas.95.21.12683
Koudinov AR Berezov TT Alzheimer's amyloid-beta (A beta) is an essential synaptic protein, not neurotoxic junk. Acta Neurobiol Exp (Wars) 2004 64 71 97 15190681
Nyakas CB Buwalda B Luiten PG Hypoxia and brain development. Prog Neurobiol 1996 49 1 51 8817697
Slotkin TA Cholinergic systems in brain development and disruption by neurotoxicants: nicotine, environmental tobacco smoke, organophosphates. Toxicol Appl Pharmacol 2004 198 132 151 15236950 10.1016/j.taap.2003.06.001
Shohami E Kaufer D Chen Y Seidman S Cohen O Ginzberg D Melamed-Book N Yirmiya R Soreq H Antisense prevention of neuronal damages following head injury in mice. J Mol Med 2000 78 228 236 10933585 10.1007/s001090000104
Dietschy JM Turley SD Cholesterol metabolism in the brain. Curr Opin Lipidol 2001 12 105 112 11264981 10.1097/00041433-200104000-00003
Tsui-Pierchala BA Encinas M Milbrandt J Johnson EM Jr Lipid rafts in neuronal signaling and function. Trends Neurosci 2002 25 412 417 12127758 10.1016/S0166-2236(02)02215-4
Hooper NM Roles of proteolysis and lipid rafts in the processing of the amyloid precursor protein and prion protein. Biochem Soc Trans 2005 33 335 338 15787600 10.1042/BST0330335
Lane RM Farlow MR Lipid homeostasis and apolipoprotein E in the development and progression of Alzheimer's disease. J Lipid Res 2005 46 949 968 15716586 10.1194/jlr.M400486-JLR200
Bastiaanse EM Hold KM Van der Laarse A The effect of membrane cholesterol content on ion transport processes in plasma membranes. Cardiovasc Res 1997 33 272 283 9074689 10.1016/S0008-6363(96)00193-9
Barrantes FJ Structural basis for lipid modulation of nicotinic acetylcholine receptor function. Brain Res Brain Res Rev 2004 47 71 95 15572164 10.1016/j.brainresrev.2004.06.008
Maguire PA Druse MJ The influence of cholesterol on synaptic fluidity, dopamine D1 binding and dopamine-stimulated adenylate cyclase. Brain Res Bull 1989 23 69 74 2529953 10.1016/0361-9230(89)90165-2
Poirier J Apolipoprotein E and Alzheimer's disease. A role in amyloid catabolism. Ann N Y Acad Sci 2000 924 81 90 11193807
Howland DS Trusko SP Savage MJ Reaume AG Lang DM Hirsch JD Maeda N Siman R Greenberg BD Scott RW Modulation of secreted -amyloid precursor protein and amyloid -peptide in brain by cholesterol. J Biol Chem 1998 273 16576 16582 9632729 10.1074/jbc.273.26.16576
Bodovitz S Klein WL Cholesterol modulates -secretase cleavage of amyloid precursor protein. J Biol Chem 1996 271 4436 4440 8626795 10.1074/jbc.271.8.4436
Eckert GP Wood WG Muller WE Statins: drugs for Alzheimer's disease? J Neural Transm 2005 in press 15682268
McCully KS Homocysteine, vitamins, and prevention of vascular disease. Mil Med 2004 169 325 329 15132238
Ravnskov U A hypothesis out-of-date. the diet-heart idea. J Clin Epidemiol 2002 55 1057 1063 12507667 10.1016/S0895-4356(02)00504-8
|
15949039
|
PMC1182387
|
CC BY
|
2021-01-04 16:39:18
|
no
|
Lipids Health Dis. 2005 Jun 10; 4:13
|
utf-8
|
Lipids Health Dis
| 2,005 |
10.1186/1476-511X-4-13
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-291602662310.1186/1475-2875-4-29ResearchIncreasing coverage of insecticide-treated nets in rural Nigeria: implications of consumer knowledge, preferences and expenditures for malaria prevention Onwujekwe Obinna [email protected] Benjamin [email protected] Nkoli [email protected] Elvis [email protected] Health Policy Research Unit, Department of Health Administration and Management, College of Medicine, University of Nigeria Enugu-Campus, Nigeria2 Department of Community Medicine, College of Medicine, University of Nigeria Enugu-Campus, Nigeria3 Department of Sociology and Anthropology, University of Nigeria, Nsukka, Nigeria4 Department of Pharmacology and Therapeutics, College of Medicine, University of Nigeria Enugu-Campus, Nigeria2005 18 7 2005 4 29 29 19 4 2005 18 7 2005 Copyright © 2005 Onwujekwe et al; licensee BioMed Central Ltd.2005Onwujekwe 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 coverage of insecticide-treated nets (ITNs) remains low despite existing distribution strategies, hence, it was important to assess consumers' preferences for distribution of ITNs, as well as their perceptions and expenditures for malaria prevention and to examine the implications for scaling-up ITNs in rural Nigeria.
Methods
Nine focus group discussions (FGDs) and questionnaires to 798 respondents from three malaria hyper-endemic villages from Enugu state, south-east Nigeria were the study tools.
Results
There was a broad spectrum of malaria preventive tools being used by people. The average monthly expenditure on malaria prevention per household was 55.55 Naira ($0.4). More than 80% of the respondent had never purchased any form of untreated mosquito net. People mostly preferred centralized community-based sales of the ITNS, with instalment payments.
Conclusion
People were knowledgeable about malaria and the beneficial effects of using nets to protect themselves from the disease. The mostly preferred community-based distribution of ITNs implies that the strategy is a potential untapped additional channel for scaling-up ITNs in Nigeria and possibly other parts of sub-Saharan Africa.
==== Body
Introduction
A current challenge that is facing many sub-Saharan African countries like Nigeria is how to achieve widespread distribution and use of insecticide-treated nets (ITNs) for the control of malaria. The Africa Malaria Report shows that many countries are quite far from reaching the target of 60% ITNs coverage in sub-Saharan African countries by the year 2005, which was set in Abuja by the African Heads of State for the provision of ITNs to children under five and to pregnant women [1]. Malaria is the number one public health problem in Nigeria [2,3]. By preventing malaria, ITNs reduce the need for treatment and the pressure on health services [2,3]. ITNs were recently added as a malaria control policy in Nigeria and the government wishes to scale-up the use of ITNs.
The determination of distribution mechanisms that will assure high coverage with the ITNs, especially in rural areas, remains a topical issue in Nigeria and in many sub-Saharan African countries (SSA). The public health care system was initially used to distribute ITNs in Nigeria, but the coverage was quite low. Currently, commercial sector distribution and social marketing of ITNs is being promoted in some states in Nigeria, but the coverage remains low. Community-based distribution has not been tried on a large scale, but it could possibly be added to existing strategies to successfully distribute and scale-up ITNs in rural areas. Similar strategies are being successfully used for the distribution of ivermectin for the control of onchocerciasis [4], filters for the control of guinea-worm [5] and home management of malaria [6].
Chavasse et al [7] identified four 'most common options' for ITN and insecticide distribution and these were: government systems; non-governmental organization systems; unassisted private sector; and assisted private sector (social marketing models). The current major strategies for scaling-up of ITNs in SSA are social marketing and use of the commercial sector [3,8-11]. However, it has been argued that the ITNs distribution approach chosen must depend primarily on local circumstances, as experience gives no clear general reason to prefer one option over another [8,12].
This paper examines the perceptions, expenditures and preferences of consumers for the prevention of malaria, as well as ITNs ownership and preferred strategies for distributing ITNs. Such information should precede the design of sustainable and effective locally relevant strategies for scaled-up distribution of ITNs. The information will also inform policy makers on the nature of information, education and communication campaigns that could be used to increase ITNs coverage. A successful social marketing strategy for ITNs in Tanzania was developed based on a similar assessment [13]. The information could also reveal the existing and potential demand for mosquito nets and provide channels for the optimal strategies for financing the ITNs.
Methods
Study area
The study was conducted in three villages in Achi community, Oji-river local government area (LGA) of Enugu State, south-east Nigeria. The villages were Ahani, Amaetiti and Enugwu-Akwu. Achi has a high malaria transmission rate, year round, with an average malaria incidence rate of 15%[14]. The major malaria vector in Achi is Anopheles gambiae, while Plasmodium falciparum causes more than 90% of all malaria infections. Achi has an estimated population of 45,000 people. Untreated mosquito nets and ITNs are not sold in these villages, but they are sold in urban areas like Enugu and Onitsha. It usually takes about one hour to reach Enugu and one hour and 20 minutes to reach Onitsha by bus.
Data collection
Qualitative data was collected through three focus group discussions (FGDs), with separate groups of men, women and youths. Hence, a total of three FGDs were held in each village (total of nine). The number of participants per FGD ranged from six to nine people and the participants were purposively selected with the help of the village heads, so that all sections of the villages were represented. The agegroups for men and women were from 35 to 60 years and for youths 20 to 34 years. Youth groups had equal numbers of male and female participants. A discussion guide was used to direct the discussions during the FGDs, which were moderated by a social scientist. Each lasted a maximum of two hours. The discussions focused on causes of malaria, how to prevent the disease, level of net ownership, preferences for distributing and paying for the nets within their villages.
Quantitative data was collected using a pre-tested interviewer-administered questionnaire that was administered to respondents from a total of 900 households (300 from each village). Adequate sample size was determined, using a power of 80%, 95% confidence level and a malaria incidence rate of 15%. The EPI-info software programme was used to calculate the sample size. The heads of households or their representatives (if the household head was not available) from the selected households were interviewed. The participants to the FGDs were excluded from the quantitative survey so as not to bias the results, since they already had better information than others. Data was collected on the socio-economic and demographic characteristics of the households, the level of their ownership of mosquito nets and their expenditure on the prevention of mosquito nuisance and malaria in the month prior to the survey. Expenditure included mosquito nets, coils, insecticide-sprays, drugs, body creams etc. The Ethics Committee of the University of Nigeria Teaching Hospital, Enugu, Nigeria approved the study.
Data analysis
The variables that were explored by both the FGDs and the questionnaire focused on determining the most common diseases in the community, local names for malaria, the different types of malaria, its symptoms and causes. Other points explored were the perception of mosquito nuisance, the subject' s health care seeking behaviour, mosquito control effort and solution to the malaria problem. Participants were also asked their preferences for paying and distributing ITNs. The possible major avenues for delivery of ITNs, such as through the commercial sector, the public health system, community-based distribution and social marketing were explored. The records of the FGDs were transcribed on the same day the FGDs were held and content analysis was used to categorize the responses into domains representing the common themes. The areas of consensus and divergence in the responses according to the groups and villages were determined for better identification of factors that influence health seeking behaviour. Tabulations and tests of statistically significant differences using non-parametric chi-squared tests were used to analyse the quantitative data.
Results
Qualitative data
Causes and prevention of malaria
The local terminology for presumptive malaria was "Iba", which denotes fever. Many causes of Iba were identified by people, though as a consensus, most of the participants in all the groups identified mosquitoes as its major cause. The methods given by the participants on mosquito nuisance control included insecticide sprays, keeping doors and windows closed in the night, use of electric fans, applying mosquito repellent cream on the exposed body at night, clearing weeds around the house, wearing long dresses, applying kerosene to the body and in the house, use of mosquito nets and physically killing mosquitoes. Other mosquito control techniques were draining stagnant water and covering water containers, especially during the rainy season, and burning or placing local leaves (osigbu) in and around the house. The leaves are good mosquito repellents because of their smell. Some participants said that preventive drugs such as paludrine were used, especially by pregnant women.
Mosquito net ownership
The discussions showed that there was low mosquito net ownership and use in these three villages. A few Amaetiti youths said that they had used bed-nets in secondary school boarding houses. All the Enugu-Akwu women said that they did not use mosquito nets due to the high cost, while one of them said that nursing mothers used nets to cover their babies. Though, window-nets were preferred to bed-nets, lack of money was the major constraints on net ownership, as people indicated that they would have bought nets, but for the high cost. Window-nets were preferred over bed-nets because they provided better aeration to the users and did not generate heat. A woman from Amaetiti said, "I will never use mosquito bed-nets even given free of charge. They make me feel as if I am being suffocated". Only one man from Enugu-Akwu had heard about ITNs through the radio. Many questioned their safety, since they felt that sleeping under nets treated with insecticides may be harmful. Most people were excited when told about ITNs and their mode of action and indicated their willingness to buy the nets when available, if affordable and not harmful.
Preferred distribution mechanisms for ITNs
Almost all groups preferred community-based distribution of the ITNs to sales by the commercial sector (e.g. patent medicine dealers), public health system and by health teams that occasionally visited the villages. The major reason for their preference was that community-based distribution would make access to the ITNs easier and the nets would not be too expensive, since the distributors would be resident in the villages and the nets would be sold at uniform prices that would be fixed by the government or the major suppliers. The majority of people felt that the profit motive could make the nets from the commercial sector very expensive. Also, they felt that teams coming to their villages to promote and sell the nets might come irregularly and, when they did, people might not have money to buy the nets. Most of the participants agreed that ITNs should be sold centrally within their communities. Some people also suggested the use of the door-to-door sales method for the nets, especially for people unable to go to the central location to buy the nets. However, the Ahani youths were against door-to-door sales. As one of them explained "If the ITNs are taken from house to house, people will value them less, because they will feel that the government was unable to sell them and is begging them to buy the nets".
Preferred mode of payments and pricing for ITNs
There was a group concensus that single full payment was the best payment system, if people could afford to pay the full price at once. However, in the event that the price of the nets is high, then instalment payments should be allowed. A one-off pre-payment system was not preferred, because income in rural areas fluctuates and even people in the formal wage-earning sectors, who get paid monthly, have competing needs, or their salaries may go unpaid for many months. Additionally, a one-off payment could prove to be a big burden, especially if a household purchases many nets. Some people from all the groups suggested that the nets should have a subsidized, but fixed price. In the words of a man from Amaetiti: "The price of the nets should be very low. They should be fixed prices and the ward leaders will tell all households about the price". Another stated that "Trained field workers should educate the people on the importance of the nets, and then seek their opinion on the prices for the net. Then, the average price should be used as what people should pay".
Quantitative data
The usable number of questionnaires for the analysis were 798. The respondents were mostly heads of households, middle-aged, females, married and farmers (Table 1). More than 87.1% of the respondents correctly identified mosquitoes as the cause of malaria, while 4.6% of them disagreed that mosquitoes cause malaria and the rest did not know whether they did (Table 2). A total of 76.2% of the respondents concurred that nets could prevent mosquito bites (this was for all kinds of nets), while 15.5%did not and the rest did not know. Though 71.4% of the respondents believed that insecticides could prevent mosquito bites, 20.9% did not believe that it could and the rest were not sure. About 83% of the respondents perceived that either they or other members of their household could contract malaria, while the rest did not. Chi-squared tests showed that all the differences in proportions were statistically significantly different (p < 0.0001).
Table 1 Socio-economic and demographic data of the respondents
Variables N = 798 n (%)
Head of household 528 (65.2%)
Number of household residents
Mean (Standard deviation) 4.62 (2.69)
The age of the respondent in years
Mean (Standard deviation) 48.77 (15.55)
Male respondents 369 (46.2%)
Years of formal education
Mean (standard deviation) 4.19 (4.76)
Whether the respondent ever married 703 (88.1%)
Occupation:
• Unemployed/unskilled labourers/housewives 11 (1.4%)
• Farmers 475 (59.5%)
• Skilled labourers/petty traders/pensioners 215 (29.6%)
• Formally employed/regular wage earners/medium business people 62 (7.8%)
• Professionals/big business people 33 (4.1%)
Table 2 Knowledge respondents had about the cause and preventive measures for malaria
Variables Yes n (%) No n (%) Do not know n (%)
Mosquitoes could cause malaria 695 (87.1%) 37 (4.6%) 66 (8.3%)
Nets could be used to prevent mosquito bites 608 (76.2%) 124 (15.5%) 66 (8.3%)
Insecticides could prevent mosquito bites 570 (71.4%) 167 (20.9%) 61 (7.6%)
Know where to buy a mosquito net 260 (32.6%) 538 (67.4%) 0
Know the current price of mosquito nets 46 (5.8%) 752 (94.2%) 0
Have ever heard of ITNs 91 (11.4%) 707 (88.6%) 0
n = 798
A small proportion (25.2%) of the households spent money on malaria prevention and the average monthly expenditure to prevent malaria was 55.55 Naira ($0.5) (Table 3). Conversely, the average monthly expenditure that people spent to treat malaria was 365.9 Naira ($3.05) with a 95% confidence interval of 306.8 Naira ($2.56) to 429.9 Naira ($3.58). People spent money on many preventive tools, including insecticide sprays, mosquito coils, window-nets, untreated bed-nets and drugs for chemoprophylaxis. More than 80% of the respondents had never purchased any form of untreated mosquito net. The nets that a few of the people had were window-nets. Only 32.6% claimed to know where to buy an untreated net, that is mainly the markets in Enugu and Onitsha. Just 5.8% of the respondents knew the current prices of untreated nets. There was very little prior knowledge about ITNs, as only 11.4% of respondents had ever heard of ITNs before the study. Nobody had ever purchased or re/treated an ITN in the three groups.
Table 3 Household expenditure to prevent malaria and level of acquisition of untreated nets
Variables Values n = 798 n (%)
Households that had bought untreated nets before the survey 122 (15.3%)
Number of households that actually spent money to prevent malaria 202 (25.2%)
Amount households spent the previous month to prevent malaria (Naira)
Mean (S.D.) 55.55 (175.6)
95% Confidence interval 43.45 – 67.84
Median 0.00
Note: • Chi-square of differences in numbers of people that bought untreated nets and those who did not = 198.0 (p < 0.0001)
• Chi-square of differences in numbers of households that spent money to prvent malaria and those who did not = 386.4 (p < 0.0001)
Discussion
People were knowledgeable about the mode of transmission of malaria and the benefits of using malaria preventive methods such as nets and insecticides, but very few households spent money on malaria preventive tools. The acquisition and usage of untreated mosquito nets was low and nil for ITNs and very few people had heard about ITNs. The results of this study are comparable to those of a marketing survey in Nigeria, where 10% of 5,000 households owned at least one net [15]. The results are also similar to findings in Mozambique, where only 3% of people had heard about ITNs and 9% used treated or ordinary nets [16]. The use of untreated bed-nets, though uncommon in households, could be found in secondary school boarding houses used by students.
The lack of substantial expenditure on malaria prevention means that getting people to pay for ITNs at the unit cost of 450 Naira ($3.8) per net, which is the average prevailing market price of ITNs in Nigeria, would be an uphill task. People would need to be convinced to increase their budget on prevention so as to cover the expense of using nets, especially in areas without a net usage culture. This depends, too, on whether they perceive ITNs as a complement to or a substitute for existing malaria preventive measures. Since, health care expenditures are the minimum amount or lower bound estimate of the amount that people are willing to pay for health care [17], the expenditures on malaria prevention could be taken as the lower bound the respondents would pay for ITNs.
An implication of these findings for scaling up of ITNs in rural areas is that malaria control programme managers should design and fine-tune how community-based distribution of ITNs could be added to existing distribution strategies so that ITNs could penetrate into the rural areas in large numbers. Community-based distribution in this context involves the recruitment and training of community residents to become ITNs community-based distributors (CBDs). The CBDs could be recruited in conjunction with the community leaders and trained by promoters of ITNs distribution such as the state and local government malaria control programme officers as well as other non-governmental promoters of ITNs. It is hoped that the CBDs would regularly obtain the ITNs from both public and non-governmental sources for sale to their community members. For sustainability of community-based distribution, the CBDs should either be paid a stipend directly by the body that supplies them the nets or allowed to slightly add a mark-up on the sale prices of the nets, which they would collect as their commission. The primary healthcare system would be expected to supervise, monitor and evaluate the CBDs as well as provide them with continuous re-training. They will also train new CBDs, when there is CBD attrition.
The malaria control programme managers should also consider how they would tackle the villagers' reluctance to prefer the commercial sector and vertical teams, so that these other strategies would also be useful in rural areas in the medium to long term basis. This is because the community-based distribution strategy would be used alongside the distribution of the ITNs through public and private healthcare facilities as well as through the commercial sector.
The payment mechanism within community-based distribution of ITNs should be designed to limit the occurrence of payment defaulters. The mostly preferred instalment payment before ITN acquisition ensures that people will get the ITNs whenever available, even if they have irregular availability of cash. Nonetheless, a mixture of instalment payment and one-off payment could be used in order to avoid the pitfalls of instalment payment, where people default in completing their payments after collecting the ITNs. The village heads could be made to act as guarantors for people who would be allowed to pay by instalment. However, subsidy mechanisms such as vouchers [9], subsidies and exemptions could be used to financially protect the poor and high risk groups.
A cost-effective use of resources is the promotion of multiple compatible distribution strategies in communities by government and other organizations, so that there would be an appreciable scaling-up of the supply and use of ITNs. The use of multiple distribution strategies within rural areas is in line with the pluralistic approach which has been advocated by the WHO Strategic Framework for Scaling-Up ITNs [18]. Using single distribution methods such as social marketing or the commercial sector alone might not lead to widespread supply and use of ITNs, especially if the consumers do not really prefer such strategies.
What is needed to scale up ITNs in rural areas is more effort to develop and implement consumer preferred distribution strategies, which will complement more widely used distribution strategies such as the commercial sector and social marketing. Motivational health education would encourage people to increase their current low level of expenditure on malaria prevention, so that they would be able to buy ITNs and re-treatment services. Examples of where community-based distribution did not work [19] and where it worked [20] should guide the development and implementation of community-based distribution of ITNs.
Authors' contributions
OO conceived and designed the study. All the authors participated in data collection and analysis. OO wrote the first draft and all the authors revised the drafts until the final draft was produced for publication.
Acknowledgements
We are grateful to Julia Fox-Rushby, Kara Hanson and Nkem Dike for their comments on an earlier draft of the paper. This study received financial support from the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical diseases.
==== Refs
World Health Organization (WHO) The African Malaria Report 2003 2003 WHO/CDS/MAL/2003.1093. Geneva:WHO/UNICEF
Onwujekwe OE Chima RI Okonkwo PO Economic burden of malaria illness versus that of a combination of all other illnesses: A study in five malaria holo-endemic communities Health Policy 2000 54 143 159 11094267 10.1016/S0168-8510(00)00105-6
Federal Ministry of Health National Strategic Plan for Roll Back Malaria 2001 Federal Ministry of Health, Abuja, Nigeria
Onwujekwe OE Chima RI Shu EN Okonkwo PO Community-directed treatment with Ivermectin in two Nigerian communities: an analysis of first year start-up processes, costs and consequences Health Policy 2002 62 31 51 12151133 10.1016/S0168-8510(01)00226-3
Brieger WR Ramakrishna J Adeniyi JD Community response to social marketing: filters for guinea worm control International Quarterly for Community Health Education 1989 19 17
Kidane G Morrow RH Teaching mothers to provide home treatment in Tigray, Ethiopia: a randomized trial Lancet 2000 356 550 555 10950232 10.1016/S0140-6736(00)02580-0
Chavasse D Reed C Attawell K Insecticide Treated Net Projects A Handbook for Managers 1999 London and Liverpool; Malaria Consortium
Armstrong-Schellenberg JRM Abdulla S Minja H Nathan R Mukasa O Marchant TJ Mponda H Kikumbih N Lyimo E Manchester T Tanner M Lengeler C KINET: a social marketing programme of treated nets and net treatment for malaria control in Tanzania, with evaluation of child health and long-term survival Trans R Soc Trop Med Hyg 1999 93 225 231 10492745 10.1016/S0035-9203(99)90001-9
Marchant T Schellenberg JA Edgar T Nathan R Abdulla S Mukasa O Mponda H Lengeler C Socially marketed insecticide-treated nets improve malaria and anaemia in pregnancy in southern Tanzania Trop Med Int Health 2002 7 149 158 11841705 10.1046/j.1365-3156.2002.00840.x
Nathan R Masanja H Mshinda H Schellenberg JA de Savigny D Lengeler C Tanner M Victora CG Mosquito nets and the poor: can social marketing redress inequities in access? Trop Med Int Health 2004 9 1121 1126 15482406 10.1111/j.1365-3156.2004.01309.x
Hanson K Goodman C Lines J Meek S Bradley D Mills A The economics of malaria control interventions 2004 Geneva: Global Forum for Health Research
Lines J Mosquito nets and insecticides for net treatment: a discussion of existing and potential distribution systems in Africa Trop Med Int Health 1996 1 616 632 8911446 10.1046/j.1365-3156.1996.d01-97.x
Minja H Schellenberg JA Mukasa O Nathan R Abdulla S Mponda H Tanner M Lengeler C Obrist B Introducing insecticide-treated nets in the Kilombero valley, Tanzania: the relevance of local knowledge and practice for an information, education and communication (IEC) campaign Trop Med Int Health 2001 6 614 623 11555427 10.1046/j.1365-3156.2001.00755.x
MVCU Prevalence of malaria morbidity and mortality in Enugu state, 1995 to 1999 2000 Malaria and Vectors Control Unit (MVCU). Ministry of Health, Enugu, Nigeria
Vyas S Hanson K Lines J Assessing the impact of ITN scaling-up activities: Consumer marketing surveys in Nigeria The Third MIM Pan-African Malaria Conference, Arusha, Tanzania 2002 246
Mbofana FS Use of insecticide-treated nets in Inharrime and Zavala districts, Mozambique: Knowledge, availability, affordability and acceptability The Third MIM Pan-African Malaria Conference, Arusha, Tanzania 2002 287
Weaver M Ndamobissi R Kornfield R Blewane C Sathe A Chapko M Bendje N Nguembi E Senwara-Defiobona J Willingness to pay for child survival: results of a national survey in a central African Republic Soc Sci Med 1996 43 985 998 8888468 10.1016/0277-9536(96)00015-9
WHO Scaling up insecticide-treated netting programmes in Africa: A strategic framework for co-ordinated national activity 2002 Geneva, WHO/RBM
Makemba AC Winch PJ Kamazima SR Makame VR Sengo F Lubega PB Minjas JN Shiff CJ Community-based sale, distribution and insecticide impregnation of mosquito nets in Bagamoyo District, Tanzania Health Policy Plann 1995 10 50 59
Kroeger A Meyer R Mancheno M Gonzalez M Pesse K Operational aspects of bednet impregnation for community-based malaria control in Nicaragua, Ecuador, Peru and Colombia Trop Med Int Health 1997 2 589 602 9236827 10.1046/j.1365-3156.1997.d01-319.x
|
16026623
|
PMC1182388
|
CC BY
|
2021-01-04 16:24:13
|
no
|
Malar J. 2005 Jul 18; 4:29
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-29
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-361604281510.1186/1475-2875-4-36ResearchBrands, costs and registration status of antimalarial drugs in the Kenyan retail sector Amin Abdinasir A [email protected] Robert W [email protected] Malaria Public Health & Epidemiology Group, Centre for Geographic Medicine, Kenya Medical Research Institute/Wellcome Trust Collaborative Programme, P.O. Box 43640, Nairobi, 00100 GPO, Kenya2 Centre for Tropical Medicine, University of Oxford, John Radcliffe Hospital, Headington, Oxford, OX3 9DU, UK2005 26 7 2005 4 36 36 19 5 2005 26 7 2005 Copyright © 2005 Amin and Snow; licensee BioMed Central Ltd.2005Amin and Snow; 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 an important source of treatment for fevers, little is known about the structure of the retail sector in Africa with regard to antimalarial drugs. This study aimed to assess the range, costs, sources and registration of antimalarial drugs in the Kenyan retail sector.
Methods
In 2002, antimalarial drug registration and trade prices were established by triangulating national registration lists, government gazettes and trade price indices. Data on registration status and trade prices were compared with similar data generated through a retail audit undertaken among 880 randomly sampled retailers in four districts of Kenya.
Results
Two hundred and eighteen antimalarial drugs were in circulation in Kenya in 2002. These included 65 "sulfur"-pyrimethamine (sulfadoxine-pyrimethamine and sulfalene-pyrimethamine (SP), the first-line recommended drug in 2002) and 33 amodiaquine (AQ, the second-line recommended drug) preparations. Only half of SP and AQ products were registered with the Pharmacy and Poisons Board. Of SP and AQ brands at district level, 40% and 44% were officially within legal registration requirements. 29% of retailers at district level stocked SP and 95% stocked AQ. The retail price of adult doses of SP and AQ were on average 0.38 and 0.76 US dollars, 100% and 347% higher than trade prices from manufacturers and importers. Artemether-lumefantrine, the newly announced first-line recommended antimalarial drug in 2004, was found in less than 1% of all retail outlets at a median cost of 7.6 US dollars.
Conclusion
There is a need to ensure that all antimalarial drugs are registered with the Pharmacy and Poisons Board to facilitate a more stringent post-marketing surveillance system to ensure drugs are safe and of good quality post-registration.
==== Body
Background
Drugs' retailers play an important role in the management of childhood fevers in Africa [1,2]. The extent of self-medication with proprietary drugs from retailers varies across the continent and has been reported to be as low as 19% in Guinea to as high as 94% in some parts of rural Ghana [3]. Reasons for reported use of the retail sector for fever management are diverse, ranging from ease of geographical access [4] to economic accessibility [5] and perceived failures of the formal health sector [3,6].
The World Health Organization (WHO) has recognised the role of the retail sector in helping to meet international targets on prompt access to antimalarial drugs [2,7]. Despite the renewed interest in home-based care, there is still a relatively poor understanding of the structure of the retail sector compared to our knowledge of formal service providers. While there has been a plethora of research on the use of the retail sector by communities in malaria endemic areas of Africa [8], complimented by a few studies on the knowledge of antimalarial drugs by service providers in the retail sector [9-12], to-date there has been only one study that has characterized the structure of the retail sector with respect to antimalarial drug ranges and sources, and how these relate to the national supply [13].
In Kenya, the retail sector is an important source of treatment for fevers [9,14-19] and the aim of this paper is to characterize the antimalarial drug products provided by this sector in terms of legal status, product ranges, costs and sources. The results focus on first and second-line treatment for uncomplicated malaria at the time of the studies, "sulfur"-pyrimethamine (sulfadoxine-pyrimethamine and sulfalene-pyrimethamine, SP) and amodiaquine (AQ), respectively. Issues related to the availability of other antimalarial products, including those that are no longer promoted as efficacious (chloroquine) and those that will serve as replacement therapies for SP and AQ in 2005 (the artemisinins) are highlighted.
Methods
National antimalarial audit
A national list of all registered antimalarial drugs was obtained from the Pharmacy and Poisons Board (PPB) of the Ministry of Health (MoH) [20,21] and updated through a series of reviews of minutes of the PPB and Kenya Gazette notices. Gazette notices are provided by the Government Printers and notify the public of registration of new products. The Committee for Drug Registration (CDR) of the PPB registers drugs based on their safety, quality and efficacy and unregistered drugs are considered illegal. Applicants pay 500 US dollars for Kenyan products and 1,000 US dollars for imported products per application. Registration is valid for five years, after which a re-registration is sought from the CDR for a further five years with a second fee of 300 US dollars and 500 US dollars for local and imported products respectively.
The official PPB list was augmented by two commercially available drugs and medical devices price lists [22,23], published yearly or bi-monthly, and serving as national trade indices. These lists provided information not available from the PPB sources. Information collected from all sources included brand names, dosage forms, strengths, manufacturers, trade packs, trade costs and product registration status and registration dates. The composite national antimalarial database was finalized on May 31, 2002 to enable comparison with subsequent retail audits at district level.
Retail audit survey procedures
A retail audit was undertaken between February and June 2002 in four malaria monitoring sentinel districts: Kwale in Coast Province, Makueni in Eastern Province, Greater Kisii in the western highlands and Bondo on the shores of Lake Victoria. These are described elsewhere in detail [4,19,24].
A national census of retailers was developed between 1999 and 2000 by a commercial market research organisation. Data for each of the four sentinel districts were purchased and used to identify outlets that stocked and sold antimalarial or antipyretic products at the time of the retail census. Data were then displayed in MapInfo (Version 6.0, 1985–2000) and physical addresses compared against coordinates of market centres obtained from topographic maps and GPS data from various sources [4,25]. Any errors in positioning the 1999–2000 data were then corrected and positions of outlets redefined to the market centre. Outlets were categorized into pharmacies, large retailers (defined crudely as stores with more than one person serving customers during normal working hours), and small retailers (defined as outlets with only one person serving customers during normal working hours). Outlets were then sampled based upon the estimates of the numbers of retail outlets in each district and the expected prevalence of SP stocks (assumed to be 50%) to achieve between 5–10% precision and 95% confidence in the parameters of interest. A minimum of 20 pharmacies in each district was targeted, large retailers were randomly sampled to achieve a minimum of 40 outlets per district, and a random sample of 160 small retailers per district was selected.
Between February and May 2002, districts were visited to a) confirm which outlets in the sampling frame were retailing antimalarial drugs (the main drugs of interest); b) establish precise geographical positions of the outlet using a hand-held GPS unit (Magellan GPS 315 or Garmin etrex); and c) obtain permission from shopkeepers for a more indepth interview at a later date. In June 2002, a retail audit was undertaken among retailers who consented to the capture of information on brands of antimalarial drugs, pharmacological groups, wholesale source and retail costs. Data were entered twice using MS-Access 2000 (Microsoft Corp., Redmond, USA) developed data-entry screens, verified and cleaned. Data were analysed using a combination of MS Excel 2000 (Microsoft Corp., Redmond, USA) and SPSS version 9.0 for Windows (SPSS Inc., Chicago, USA) and presented as proportions, medians and interquartile ranges.
Results
Range of products at the national level
One hundred and thirty five oral antimalarial products registered with the PPB were identified. Of these, registration dates were gazetted for 122 (90.4%), the remaining 13 (9.6%) were noted in the minutes of the CDR as approved for registration, but no registration dates or reference numbers were available. Of the 122 products for which details were available, only 40 (32.8%) were within their five-year registration period, and 82 (67.2%) were due for re-registration. It is possible that some products in the latter group had been granted marketing approval by the PPB, but were awaiting final gazettement and registration (and could, therefore, not be found on official lists) or that the manufacturers and dealers no longer marketed these products and had, therefore, allowed registration to lapse. Of those with an expired registration status, however, 39 (47.6%) were identifiable on the commercial trade indices, suggesting they were still marketed in Kenya.
83 products were identified that were not on the PPB list, but were available on commercial lists. It is estimated, therefore, that 218 oral antimalarial products were in circulation in 2002. Of these products, 92 (42.2%) were manufactured locally and 126 (57.8%) were products imported from overseas.
Table 1 shows the overall registration status of the 218 products according to the various antimalarial classes. From the national audit, 34/65 (52.3%) of SP products were registered, while 17/33 (51.5%) of AQ products were registered with the PPB. All artemisinin (ART) tablets, mefloquine (MEF) tablets and halofantrine (HAL) products were registered, while none of the ART suspensions were registered.
Table 1 National and retail audit of oral anti-malarial drugs available on the Kenyan market in 2002
National audit Retail audit
Generic Name # Brands identified (registered)* Median (IQR) cost (USD) of treatment course† # Brands identified (registered)* # Outlets stocking (%) Median (IQR) cost (USD) of treatment course†
SP tablets 49 (29) 0.19 [0.13, 0.32] 30 (16) 250 (28.5%) 0.38 [0.25, 0.65]
SP suspensions and drops‡ 16 (5) 0.39 [0.28, 0.53] 15 (2) 57 (6.5%) 0.44 [0.44, 0.56]
AQ tablets 22 (12) 0.17 [0.14, 0.55] 13 (6) 818 (93.4%) 0.76 [0.76, 0.76]
AQ suspensions 11 (5) 0.28 [0.15, 0.49] 12 (5) 71 (8.1%) 0.51 [0.39, 0.60]
CQ tablets 43 (33) 0.09 [0.08, 0.19] 12 (5) 132 (15.1%) 0.44 [0.25, 0.44]
CQ syrups 22 (10) 0.05 [0.03, 0.30] 9 (4) 12 (1.4%) 0.26 [0.06, 0.31]
QN tablets 25 (19) 2.73 [2.40, 3.16] 3 (2) 32 (3.7%) 3.20 [2.40, 4.00]
QN drops and mixtures 3 (1) 2.24 [1.87, 3.00] 4 (1) 49 (5.6%) 4.00 [3.27, 4.00]
ART tabs 11 (11) 5.34 [4.16, 5.56] 7 (7) 24 (2.7%) 7.11 [6.14, 7.96]
ART suspensions 1 (0) 3.86 [3.86, 3.86] 2 (0) 21 (2.4%) 5.00 [4.44, 5.13]
MEF tablets 4 (4) 5.04 [3.83, 9.53] 4 (4) 14 (1.6%) 7.61 [7.33, 7.87]
HAL tablets 1 (1) 7.96 [7.96, 7.96] 1 (1) 22 (2.5%) 9.90 [9.26, 10.25]
HAL suspensions 1 (1) 2.83 [2.83, 2.83] 1 (1) 18 (2.1%) 3.55 [3.20, 3.71]
Other tablets 9 (6) 3.62 [0.35, 13.20] 1 (1) 18 (2.1%) 17.51 [17.51, 17.51]
* Registration period covers up to and including May 31, 2002.
† For packaged commodities, the calculations were derived per tablet and per recommended dose for adults. Where possible, large, bulk packaging was selected for individual suppliers to provide the cheapest values for the national audit.
‡ Liquid dosage forms (suspensions, syrups, mixtures and paediatric drops) were all costed per dosage per child aged 1–5 years – not adult treatment courses. The mean dose per product was calculated as the mid-point between the Division of Malaria Control (DOMC) recommended dose for a 1 year old (lower limit of 10 kg) and a 5 year old (upper limit of 18 kg).
Range of antimalarial drug classes and formulations at district level
Eight hundred and eighty retailers were sampled, but four were excluded from analysis since the shops remained closed even after three visits (Table 1). Overall, SP was stocked by 28.7% of retailers and AQ by 94.7%. Chloroquine (CQ), which had been replaced by SP in 1998 as the first-line recommended drug, was still available in 15.4% of retail outlets. Other antimalarial drugs were available in less than 10% of retail outlets. SP and AQ tablets were the most widely stocked formulations (28.5% and 93.4%, respectively) and were available in pharmacies, large retailers, and small retailers. ART, HAL and MEF were sold exclusively in pharmacies.
Range, availability and registration of antimalarial brands at district level
Thirty brands of SP tablets were identified in the districts (Table 1), the two most widely stocked being Falcidin® (stocked by 19.9% of outlets, Cosmos Limited, Kenya) and Fansidar® (8.8%, L. Hoffmann La Roche, Switzerland). Sixteen SP brands (53.3%) were registered with the PPB. Zentakelfin®, Sudorin® and Lansidar®, all unregistered, were available in some district level pharmacies and not recorded during the national audit. Fifteen brands of SP suspensions were also found, yet only two were registered with the PPB.
For AQ tablets, 13 brands were identified of which six (46.2%) were registered with the PPB. Three brands (which were not registered) were identified in the districts and were not recorded during the national audit: Amowin®, Vanida®, and Maratab®. Malaratab® (Cosmos Limited, Kenya) was the most widely stocked AQ tablet, found in 87.9% of outlets. Twelve brands of AQ syrup were encountered with only five (41.7%) registered with the PPB.
Of interest was that all higher order antimalarial tablets, such as quinine (QN), HAL, MEF and ART class of drugs, were to a large extent all registered with the PPB and mostly available in pharmacies. The exception was one QN tablet formulation. Conversely, only 2/5 (40%) of the higher order antimalarial syrups were registered.
Wholesale sources of antimalarial drugs to district level retailers
Due to the wide range of products in pharmacies, it was not possible to ask the wholesale source of individual products. Respondents were asked the primary wholesale source of antimalarial drugs in stock at the time. For large and small retailers, the source of each product audited was established. The wholesale source of most products in stock was adopted as the primary source of antimalarial drugs for each outlet. Drug sources thus defined were classified in eight groups as shown in Table 2. Results show that overall pharmacies obtained their antimalarial drugs from pharmaceutical wholesalers outside the districts (67.1%). Large retailers obtained their drugs from general wholesalers outside the districts (39.8%) or inside the district (34.8%). Most small retailers (45.3%) obtained antimalarial drugs from general wholesalers within the districts. A substantial proportion of small retailers (23.3%) also obtained their antimalarial drugs from general wholesalers outside the districts. Mobile vendors supplied a good number of small retailers (14.0%) and large retailers (7.5%), but not pharmacies.
Table 2 Primary wholesale sources of antimalarial products to 876* retailers in the four study districts.
Totals
Pharmacies Large Shops Small Shops
Mobile vendors 0 12 (7.5%) 90 (14.0%)
General wholesalers-within district 0 56 (34.8%) 292 (45.3%)
General wholesaler-outside district 2 (2.9%) 64 (39.8%) 150 (23.3%)
Pharmaceutical wholesaler-within district 17 (24.3%) 11 (6.8%) 51 (7.9%)
Pharmaceutical wholesaler-outside district 47 (67.1%) 3 (1.9%) 13 (2.0%)
Pharmaceutical company 2 (2.9%) 3 (1.9%) 0
Drug representative 0 7 (4.3%) 5 (0.8%)
Unknown 2 (2.9%) 5 (3.1%) 44 (6.8%)
* Four small retailers were excluded from analysis since they remained closed even after three repeated visits.
Trade versus retail costs of antimalarial drugs
For product costs, standardized dose regimens were used to enable comparisons between the antimalarial classes. SP, AQ, CQ and QN doses were based on the malaria standard treatment guidelines set by the DOMC of the MoH [26]. For all other antimalarial drugs (which are not the subject of DOMC guidelines), the East African Pharmaceutical Loci, a regional formulary for healthcare professionals, was used [22]. Costs were calculated in US dollars (USD) based on the Central Bank of Kenya rates at the time of the survey (June 2002).
For trade prices, large bulk packaging was selected per supplier to provide the cheapest factory gate costs (Table 1). The trade price for an adult dose of SP was 0.19 USD (IQR 0.13, 0.32), while a standard dose for a paediatric patient on SP suspensions cost nearly twice as much. AQ tablet and suspension prices were 0.17 USD for an adult dose and 0.28 USD for a paediatric suspension. The trade price range for QN, MEF, HAL and ART products was between 2.24 and 7.96 USD (Table 1).
Retail prices at the district level were standardized to an adult and paediatric dose as described above and are shown in Table 1. The retail price for an adult dose of SP was 0.38 USD (IQR 0.25, 0.65), while a standard dose for a paediatric patient on SP suspension was 0.44 USD (IQR 0.44, 0.56). AQ tablet prices were 0.76 USD (IQR 0.76, 0.76), while AQ suspension cost a median of 0.51 USD (0.39, 0.60). The retail price range for QN, MEF, HAL and ART products was between 3.55 to 9.90 USD (Table 1).
The two most widely stocked brands of SP and AQ tablets (Falcidin® and Malaratab®, respectively) are used to demonstrate retail mark-ups on trade costs. The trade cost for an adult treatment course of Falcidin® in May 2002 was 0.15 USD, while that for Malaratab® was 0.55 USD. The equivalent median retail prices for the two products were 0.32 USD and 0.76 USD, respectively, representing a 113% and 38% mark-up on the trade price.
Discussion
There was a wide range of antimalarial products available in the Kenyan retail market in 2002. However, not all nationally marketed first- and second-line drugs were in circulation in the peripheral retail sector; 17 preparations of SP (26%) and five (15%) of AQ were not detected at the district level. Conversely, there were three brands of SP and three of AQ which were in retail circulation at the district level, but which had not been identified during the central audit and were unregistered by the PPB. One serious consequence of a wide range of products available to largely biomedically ill-informed, rural populations is brand confusion, which may lead to unintentional repeated doses of the same drug class and consequently, dose-dependent adverse effects [27]. Moreover, the availability of unregistered products poses a danger to malaria treatment since the safety, efficacy and quality of such drugs cannot be guaranteed [28].
The most widely available antimalarial in the retail sector was AQ, sold in 95% of outlets surveyed. AQ was a prescription-only medication (POM) and regarded as the second-line treatment at the time of the survey. In contrast, SP, the first-line drug in 2002, was available in only 29% of outlets surveyed. The situation was found to be different in neighbouring Tanzania when CQ was still the first-line drug for uncomplicated malaria. 33% of general retailers and 98% of pharmacies stocked CQ. Conversely SP, the second-line treatment, was sold in less than1% of general retailers and in 37% of pharmacies [13]. Although the apparent policy-practice disconnect in Kenya in drug scheduling and national malaria policy could be attributed to a more prolific and unregulated retail market in antimalarial drugs compared to Tanzania, a more plausible explanation is a lack of a concerted national effort to inform populations about changes in drug policy. In the absence of strong government action to back up new drug policies (e.g. mass communication, training of health workers, etc.), market forces will fill the gap and dictate the stocking of antimalarial drugs by retailers. Closer cooperation and consultation with local pharmaceutical manufacturers and importers of antimalarial drugs during national policy change and broad, high profile, community-wide communication strategies for drug policy change are critical for successful implementation.
Compared to the national trade price data, mark-ups are between 100–347% (using SP and AQ tablets as examples) when they reach the peripheral retail level. Retailers represent the last link in a chain of manufacturers/drug importers and wholesalers, and it is not clear from the results if the high price mark-ups on antimalarial drugs are transferred from these primary/middle level suppliers to the retailer or these represent retailer pricing structures. More formative research is required to understand the pricing structure for antimalarial drugs in the retail sector to identify opportunities to reduce costs and retail profit margins with new medicines such as artemisinin-based combination therapies (ACTs).
At the time of this study, the artemisinins (potential replacements to SP monotherapy) were only available in pharmacies in Kenya. More importantly, the proposed new first-line drug for Kenya, artemether-lumefantrine (registered in Kenya in July 1999 as Coartem®), was available in 8/70 (11%) of district level pharmacies (i.e. less than 1% of all retailers) at a median cost of 7.6 US dollars (interquartile range 6.2, 10.1). This represents a huge cost to households which invariably bear most of the cost of treating malaria [29].
Conclusion
The studies show that when drug classes become established treatments, many branded options soon become available on the retail market. In Kenya, the regulation of this plethora of brands is weak: many do not go through the regulatory pipeline and for those that do, there is little follow-up post-registration. In addition, antimalarial drug prices vary widely within the same drug class. During the era of ACTs, there is a need to ensure that these new drugs are registered and monitored post-registration to ensure their continued safety, quality and efficacy. Regulatory mechanisms and price controls must be strengthened and enforced to improve the use of drugs in the retail sector. In addition, nation-wide, high profile communication strategies should accompany drug policy changes to bring demand for antimalarial drugs in line with policy.
Authors' contributions
AAA was responsible for the study conception and design, data analysis, interpretation and writing of the manuscript. RWS supervised the studies overall and contributed substantially to redrafting the manuscript.
Acknowledgements
Funds were provided by the Department for International Development-Kenya (#031555054), the Wellcome Trust, UK (#058992), the Ministry of Health, Government of Kenya, and the Kenya Medical Research Institute. None of these institutions was involved in the study conception, data collection, analysis and interpretation of results. The authors are grateful to Abdisalan Noor, Judy Omumbo, Dejan Zurovac and Vicki Marsh for comments on an earlier draft. The authors also wish to acknowledge the support provided by: the Pharmacy & Poisons Board, the Ministry of Health, Lydia Mwangi and Lucy Muhunyo. RWS is a Senior Wellcome Trust Fellow. This paper is published with the permission of the Director, KEMRI.
==== Refs
McCombie SC Treatment seeking for malaria: A review of recent research Soc Sci Med 1996 43 933 945 8888463 10.1016/0277-9536(95)00446-7
RBM Roll Back Malaria Consultative Meeting on the Role of Medicine Sellers in the Management of Malaria: What worked and where do we go from here? Accra 26–27 May 2004 2004 Roll Back Malaria (RBM)
Foster SDF Treatment of malaria outside the formal health services J Trop Med Hyg 1995 98 29 34 7861477
Noor AM Zurovac D Hay SI Ochola SA Snow RW Defining equity in physical access to clinical services using geographic information systems as part of malaria planning and monitoring in Kenya Trop Med Int Health 2003 8 917 926 14516303 10.1046/j.1365-3156.2003.01112.x
Ettling M McFarland DA Schultz LJ Chitsulo L Economic impact of malaria in Malawian households Trop Med Parasitol 1994 45 74 79 8066390
Hassouna WA Reaching the people: A three-country study of health systems World Health Forum 1983 4 57 62
WHO Scaling up home-based management of malaria. From research to implementation 2003 Geneva: World Health Organization 1 94
McCombie SC Self-treatment for malaria: the evidence and methodological issues Health Policy Plan 2002 17 333 344 12424205 10.1093/heapol/17.4.333
Marsh VM Mutemi WM Muturi J Haaland A Watkins WM Otieno G Marsh K Changing home treatment of childhood fevers by training shopkeepers in rural Kenya Trop Med Int Health 1999 4 383 389 10402975 10.1046/j.1365-3156.1999.00403.x
Marsh VM Mutemi WM Willetts A Bayah K Were S Ross A Marsh K Improving malaria home treatment by training drug retailers in rural Kenya Trop Med Int Health 2004 9 451 460 15078263 10.1111/j.1365-3156.2004.01223.x
Tavrow P Shabahang J Makama S Vendor-to-vendor education to improve malaria treatment by private drug outlets in Bungoma district, Kenya Malar J 2003 2 10 12812525 10.1186/1475-2875-2-10
Oshiname FO Brieger WR Primary care training for patent medicine vendors in rural Nigeria Soc Sci Med 1992 35 1477 1484 1485195 10.1016/0277-9536(92)90050-Z
Goodman CA Kachur SP Abdulla S Mwageni E Nyoni J Schellenberg JA Mills A Bloland P Retail supply of malaria-related drugs in rural Tanzania: risks and opportunities Trop Med Int Health 2004 9 655 663 15189455 10.1111/j.1365-3156.2004.01245.x
Hamel MJ Odhacha A Roberts JM Deming MS Malaria control in Bungoma District, Kenya: a survey of home treatment of children with fever, bed net use and attendance at antenatal clinics Bull World Health Organ 2001 79 1014 1023 11731808
Mwenesi HA Harpham T Snow RW Child malaria treatment practices among mothers in Kenya Soc Sci Med 1995 40 1271 1277 7610432 10.1016/0277-9536(94)00250-W
Molyneux CS Mung'ala-Odera V Harpham T Snow RW Maternal responses to childhood fevers: a comparison of rural and urban residents in coastal Kenya Trop Med Int Health 1999 4 836 845 10632992 10.1046/j.1365-3156.1999.00489.x
Snow RW Peshu N Forster D Mwenesi HA Marsh K The role of shops in the treatment and prevention of childhood malaria on the coast of Kenya Trans R Soc Trop Med Hyg 1992 86 237 239 1412642 10.1016/0035-9203(92)90290-S
Ruebush TK Kern MK Campbell CC Oloo AJ Self-treatment of malaria in a rural area of western Kenya Bull World Health Organ 1995 73 229 236 7743595
Amin AA Marsh VM Noor AM Ochola SA Snow RW The use of formal and informal curative services in the management of paediatric fevers in four districts in Kenya Trop Med Int Health 2003 8 1143 1152 14641851 10.1046/j.1360-2276.2003.01140.x
PPB Registered Drugs 1982–June 1996, First Edition 1996 Nairobi: Pharmacy and Poisons Board (PPB), Ministry of Health, Republic of Kenya 1 205
PPB Registered Drugs 1st July 1996–December-2000, Second Edition 2001 Nairobi: Pharmacy and Poisons Board (PPB), Ministry of Health, Republic of Kenya 1 91
Kimotho JH Akumu PF Muriuki G Ombega JN (eds) East African Pharmaceutical Loci. A regional drug index for Healthcare Practitioners 2002 Nairobi: Pharmaceutical Loci Publishers
SMS Drugs & Chemists, 2001 5 Nairobi: Scientific Media Services
Noor AM Gikandi P Hay SI Muga RO Snow RW Creating spatially defined databases for equitable health service planning in low-income countries: the example of Kenya Acta Trop 2004 91 239 251 15246930 10.1016/j.actatropica.2004.05.003
CBS 1999 Population and Housing Census: Counting our People for Development: Population Distribution by Administrative Areas and Urban Centres 2001 1 Nairobi: Central Bureau of Statistics (CBS), Ministry of Finance and Planning, Government of Kenya 1 411
DOMC National guidelines for diagnosis, treatment and prevention of malaria for health workers 1998 Nairobi: Division of Malaria Control (DOMC), Ministry of Health 1 49
Osei L Beecham HA Drug use before clinic attendance Ghana Med J 1990 24 249 254
WHO Effective drug regulation: what can countries do? 1999 Geneva 1 53
Chima RI Goodman CA Mills A The economic impact of malaria in Africa: a critical review of the evidence Health Policy 2003 63 17 36 12468115 10.1016/S0168-8510(02)00036-2
|
16042815
|
PMC1182389
|
CC BY
|
2021-01-04 16:24:13
|
no
|
Malar J. 2005 Jul 26; 4:36
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-36
|
oa_comm
|
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-161592153710.1186/1475-2859-4-16ResearchOptimization of the Lactococcus lactis nisin-controlled gene expression system NICE for industrial applications Mierau Igor [email protected] Kees [email protected] James [email protected] Eddy J [email protected] NIZO food research, P.O. Box 20, 6710 BA EDE, The Netherlands2 Biosynexus Inc., 9119 Gaither Road, Gaithersburg, MD 20877, USA2005 30 5 2005 4 16 16 18 4 2005 30 5 2005 Copyright © 2005 Mierau et al; licensee BioMed Central Ltd.2005Mierau 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 nisin-controlled gene expression system NICE of Lactococcus lactis is one of the most widely used expression systems in Gram-positive bacteria. Despite its widespread use, no optimization of the culture conditions and nisin induction has been carried out to obtain maximum yields. As a model system induced production of lysostaphin, an antibacterial protein (mainly against Staphylococcus aureus) produced by S. simulans biovar. Staphylolyticus, was used. Three main areas need optimization for maximum yields: cell density, nisin-controlled induction and protein production, and parameters specific for the target-protein.
Results
In a series of pH-controlled fermentations the following parameters were optimized: pH of the culture, use of NaOH or NH4OH as neutralizing agent, the addition of zinc and phosphate, the fermentation temperature, the time point of induction (cell density of the culture), the amount of nisin added for induction and the amount of three basic medium components, i.e. yeast extract, peptone and lactose. For each culture growth and lysostaphin production was followed. Lysostaphin production yields depended on all parameters that were varied. In the course of the optimization a three-fold increase in lysostaphin yield was achieved from 100 mg/l to 300 mg/l.
Conclusion
Protein production with the NICE gene expression system in L. lactis strongly depends on the medium composition, the fermentation parameters and the amount of nisin added for induction. Careful optimization of key parameters lead to a significant increase in the yield of the target protein.
==== Body
Background
Lactococcus lactis is a Gram-positive lactic acid bacterium that is widely used in food fermentations, such as in cheese and butter production. In the last two decades the physiology and genetics of this bacterium have been thoroughly studied [1,2]. At present the genome sequences of several strains of L. lactis have been elucidated, leading to an acceleration and integration of our knowledge of these bacteria [3-6]. Because of its genetic accessibility and because it is easy to handle, L. lactis, in addition to its traditional applications, has been extensively developed and used for the expression of heterologous genes, and has become one of the most used Gram-positive gene expression hosts. Table 1 gives an overview of the wide range of applications that involve L. lactis as host bacterium.
Table 1 Overview of various applications of the NICE system
Application area Application Reference
Expression of homologous genes Aminopeptidase N [7]
ATPase system (8 gene operon) [28]
Aminoacylase [29]
Cluster of genes encoding folate biosynthesis [19]
Expression of heterologous genes of Gram+ and Gram- bacteria NADH oxidase of Streptococcus mutans [30]
Fructose bisphosphatase of Escherichia coli [31]
Green fluorescent protein of Aequoria victoria [32]
Lysostaphin of Staphylococcus simulans biovar. Staphylolyticus [15]
Protein secretion Lipase of Staphylococcus hyicus [33]
Bovine beta-lactoglobulin [34]
Membrane proteins: prokaryotic and eukaryotic Multidrug transporter of Lactococcus lactis
Xylidose transporter of Lactobacillus pentosus
ATP/ADP translocator of Rickettsia prowazekii Review [35]
KDEL receptor of Homo sapiens
Mitochondrial carriers of Saccharomyces cerevisiae
Cloning of toxic genes Lysis cassette of the virulent phage us3 of [26]
Lactococcus lactis
Autolysin gene of Leuconostoc citreum [36]
Bacterial antigens Antigen L7/12 of Brucella abortus [13, 37]
C subunit of tetanus toxin (TTFC) of Clostridium tetani [38]
Viral antigens Non-structural protein 4 of bovine rotavirus [13, 39]
Cytokines Interleukin 12 of Homo sapiens [40]
Industrial-scale application Lysostaphin of Staphylococcus simulans biovar. Staphylolyticus [15]
Regulated gene expression can be of critical importance in achieving high yields of proteins. This is either important for the expression of genes that form toxic products for the cell (see Table 1) or for the conservation of energy for producing biomass prior to directed overproduction of the protein of interest. The most commonly used regulated expression system of Gram positive bacteria is the NIsin Controlled gene Expression system NICE of L. lactis [7]. Sub-toxic amounts of nisin in the ng/mL range are sufficient to fully activate the otherwise tightly closed promoter [8]. In the natural situation nisin binds to the receptor NisK. Subsequently NisK activates NisR by phosphorylation and the activated NisR induces the nisin operon at the nisin A promoter [9]. To exploit this system for gene expression, the genes of the receptor protein and the response regulator – nisK and nisR – have been isolated and placed on the chromosome of a suitable host strain. Furthermore, the nisin A promoter has been isolated and placed on a plasmid. When a gene is cloned down-stream of this promoter and the construct is placed in a nisRK strain, expression can be activated by the addition of nisin (Figure 1) [7]. Depending on the presence of a signal sequence the product is either accumulated inside the cell or secreted into the medium (see Table 1).
Figure 1 Schematic overview of the NICE system, its components and its function. NisK and NisR are the sensor protein and the response regulator, respectively. The product of the expressed gene can either accumulate in the cell or be secreted into the extracellular medium depending on the presence of a signal sequence in the construct.
Escherichia coli is, at present, the dominant prokaryotic system for industrial gene expression. This is due to high yields, ease in genetic handling, long-term experience and extensive documentation with the US Food and Drug Administration and other regulatory bodies. However, there are also various disadvantages, such as the formation of endotoxins, the formation of inclusion bodies, the presence of two membranes, which hampers secretion, and the relatively complicated aerobic fermentation [10,11]. L. lactis, on the other hand, has a number of properties that make this bacterium an interesting alternative candidate for large-scale gene expression: the bacterium is food grade (used in food production for thousands of years), it is used at very large scales, and plasmid selection mechanisms are available that are food grade and self-cloning (e.g. growth on lactose) [12]. Furthermore, no endotoxins or inclusion bodies are formed and sophisticated genetic tools enable easy genetic handling [13,14]. Finally, simple, non-aerated fermentation makes direct scale-up from 1-L scale to 1000-L scale possible [2,14]. Recently, we have demonstrated in our laboratory that nisin controlled gene expression can be effectively used in 3000-L scale fermentations [15].
The NICE system has been used in a multitude of laboratory applications (Table 1), in which gene expression is often performed in acidifying cultures. The drawback of this culture type is its low final cell density due to medium acidification by lactic acid. pH-controlled fermentations, on the other hand, result in at least five-fold higher cell densities and thus higher biomass yields. However, no systematic optimization of the induction and expression of the NICE system under pH-regulated conditions has yet been performed. Three main areas need optimization for maximum yields: cell density of the culture, nisin-controlled induction and protein production, and parameters specific for the target-protein.
Lysostaphin is a 25 kD antibacterial protein produced by Staphylococcus simulans biovar. Staphylolyticus [16] and mainly used against multiple antibiotic resistant S. aureus [17,18]. In this paper we demonstrate, with lysostaphin as a model, that careful optimization of key parameters of the induction and production process can lead to an at least three-fold increase of the fermentation yield from 100 mg/L to 300 mg/L lysostaphin.
Results
Introduction
At present, protocols for nisin-induced gene expression only exist for acidifying batch cultures (see e.g. [7,19]) but not for pH-controlled batch cultures. Therefore, a protocol that was previously developed at our laboratory for large-scale lysostaphin production [15] was used as a starting point. The cultivation and induction conditions were as follows: pH controlled (using NaOH) growth at pH 6.5, growth temperature 30°C, inoculum 1%, induction at OD600 = 1 (= 0.3 g/L cell dry weight [20]) with 10 ng/mL nisin, and harvest after 6 h. The initial medium composition was: 5% lactose, 1.5 % soy peptone, 1% yeast extract, 1 mM MgSO4 and 0.1 mM MnSO4.
In a series of 58 1-L fermentations the following parameters were tested and optimized: pH, neutralization agent, addition of phosphate and zinc, fermentation temperature, the time point of induction, i.e. cell density, the concentration of peptone, yeast extract and lactose in the medium and the lysostaphin production time after induction.
pH and neutralizing agent
The main fermentation end product of L. lactis is lactate, which is toxic and retards growth above a certain concentration. The toxic effect of lactate depends on its undissociated form [21]. Therefore, any increase of the pH can lead to a prolongation of growth in the presence of higher concentrations of lactate. NaOH is a more alkaline neutralizing agent than NH4OH and could cause cell damage. Furthermore, NH4OH could contribute to the ammonium metabolism of the cell [22]. Figure 2 shows that neutralization with NH4OH has a small effect, however the combination of NH4OH with an increase of the fermentation pH to 7.0 leads to prolonged exponential growth (not shown) and to a higher final cell density. Increase of the pH to 7.5 leads to a retardation of the growth rate and has therefore not been further pursued.
Figure 2 Biomass production depending on pH and neutralizing agent. The grey bar indicates the use of NaOH for neutralization of the culture and black bars indicate the use of NH4OH for neutralization. Biomass production is expressed in optical density at 600 nm (path length = 1 cm) (the dry cell weight factor is 0.3 g/L/1 OD600). Higher biomass production may be beneficial for higher yield of the produced heterologous protein.
Addition of phosphate and zinc
0.01% of sodium phosphate (Na2HPO4 *2 H2O) was added to evaluate the effect on growth performance. Furthermore, zinc was specifically added, because lysostaphin is a metallo-enzyme with zinc as co-factor [16]. Elemental analysis of the basic medium had shown that it contained 1.2 mg/L Zn2+. This amount would be sufficient for a maximum of 500 mg/L lysostaphin if all the zinc were available. Therefore, 100 μM (16 mg/L) ZnSO4 was added, which would allow the production of a maximum of 2.5 g/L active lysostaphin. Addition of these components had no significant effect on growth under the initial fermentation and induction conditions (Figure 3). However, in a later stage when induction was performed at a higher cell density and with more nisin (induction at OD600 = 5 with 40 ng/ml nisin) considerably less lysostaphin was formed when phosphate was omitted (150 mg/L without phosphate versus 220 mg/L lysostaphin with phosphate).
Figure 3 Influence of zinc and phosphorus (Zn Pi) on growth, induction and lysostaphin production. Growth of the culture is indicated with continuous lines. Production of lysostaphin is indicated with broken lines. The cells are induced for lysostaphin (Lss) production at OD600 = 1 with 10 ng/mL nisin. [OD 1 10] or [Lss 1 10] indicate conditions for growth or lysostaphin production.
After these initial experiments, it was decided to set the following basic fermentation conditions: pH at 7.0, to use NH4OH as neutralizing agent, to add both 0.01 g/L sodium phosphate (Na2HPO4 *2 H2O) and 100 μM ZnSO4. To establish a baseline for lysostaphin production and to analyze the effect of phosphate or zinc on lysostaphin production, an induction experiment was carried out. Induction of lysostaphin production was initiated at OD600 = 1.0 with 10 ng/mL nisin. Figure 3 shows the basic pattern of growth and induction. After the addition of nisin, product formation begins immediately, and occurs parallel to growth. When lysostaphin production was induced, growth of the culture slowed down considerably about 30 min after induction. This is accompanied by a steep drop of the viable counts on lactose M17 agar plates of 4 orders of magnitude. However, despite the growth retardation, lysostaphin production was linear for 8 h before it abruptly ended (Figure 3) [15].
Fermentation temperature
Overproduction of proteins can trigger stress responses and thus e.g. degradation of the target protein in the cell [23]. One of the strategies employed to prevent or minimize this effect is to lower the fermentation temperature and thereby lower the threshold for the induction of the stress response, and also slow down protein production. Figure 4 shows both growth and lysostaphin production during fermentation at 30°C, 25°C and 20°C. Induction was carried out at a cell density of OD600 = 1 (0.3 g/L cell dry weight) with 10 ng/mL nisin. At 25°C and 20°C growth is, as expected, slower than at 30°C. Also lysostaphin production is slower and much lower at 20°C as compared to 30°C. However, these experiments also show that the NICE system works at these temperatures; especially at 25°C similar yields are obtained as at 30°C. However, because of the faster production and the slightly higher yields, 30°C was chosen as fermentation temperature for all subsequent experiments.
Figure 4 Influence of the temperature on growth, induction and lysostaphin production. Growth of the culture is indicated with continuous lines. Production of lysostaphin is indicated with broken lines. The cells are induced at the different growth temperatures for lysostaphin production at OD600 = 1 with 10 ng/mL nisin. [Lss 30°C], condition for lysostaphin production; [OD 1 10 30°C], condition for growth.
Combination of cell density at induction and the amount of added nisin
If induction could be performed at higher cell densities a higher yield per volume of culture fluid could be expected. Since at higher cell densities more nisin may also be needed for complete induction, these parameters were tested together. Initial experiments showed that induction could be performed at an OD600 of 3 or even 5. Figure 5 shows an experiment were induction was tested at OD600 values of 1, 5 and 7 (equal to a cell dry weight of 0.3 g/L, 1.5 g/L and 2.1 g/L, respectively) with nisin concentrations of up to 80 ng/mL. Clearly lysostaphin production can be increased when the culture is induced at higher cell densities. At the same time, more nisin needs to be added for maximum induction. When induction was performed at OD600 = 5, 160 mg/L lysostaphin was formed with 20 ng/mL nisin and 220 mg/L lysostaphin was produced when 40 ng/mL nisin was used for induction. This means that there was a clear correlation between the cell density at induction and the amount of nisin that is needed for maximal induction. Induction at even higher cell densities, OD600 = 7 with 40 and 80 ng/mL nisin, was also tried; however this did not lead to higher lysostaphin yields, indicating that there is a limit in the growth cycle beyond which induction becomes ineffective. Exponential growth without induction continues until about OD600 = 8, after which growth slows down as the culture slowly approaches the final cell density of about 15. This is due to the accumulating lactate that increasingly interferes with the energy metabolism of the cell [24,25]. The same mechanism also affects and limits the production of the recombinant protein. Induction at a cell density of OD600 = 5 with 40 ng/mL nisin resulted in a maximum yield of the recombinant protein.
Figure 5 Influence of the time point of induction and the amount of added nisin on the amount of produced lysostaphin 8 hours after induction. The time point of induction is indicated as the optical density of the culture [OD600] at which nisin was added.
Addition of extra nutrients
Although induction was successful at a five times higher cell density than the original induction conditions, we observed only about 2 – 2.5 times more product formation. This could be due to either a shortage in building blocks, i.e. amino acids or a limitation of the energy supply, i.e. the fermentable sugar. Therefore, the next step was to supply more nitrogen and carbohydrate sources. Initial experiments showed that the lysostaphin yield increases when either more yeast extract or more peptone was added and increases even more when both nutrients were increased simultaneously (results not shown). Figure 6 shows a series of fermentations in which the influence of a combination of nitrogen and carbon sources on product formation was investigated. Lysostaphin production was induced at OD600 = 5 (1.5 g/L cell dry weight) with 40 ng/mL nisin. The addition of extra carbon source, 7% lactose versus 5% lactose, did not significantly increase the production of lysostaphin. However, doubling of the carbon source did raise lysostaphin yield from 225 mg/L to 290 mg/L. An increase in the supply of the nitrogen sources peptone and yeast extract also raised the lysostaphin yield form 225 mg/L to 290 mg/L. In a final experiment we used the best medium (2.5% peptone, 2% yeast extract and 7% lactose) and looked again at the combination of cell density and the amount of nisin for induction (Figure 7). Clearly, either lower or higher cell densities (OD600 = 4 or 7) lead to slower production and lower yields. The same is true for the addition of more nisin (60 ng/mL), which at that concentration probably has a detrimental effect on the production of lysostaphin.
Figure 6 Influence of the medium composition on lysostaphin yield. Growth of the culture is indicated with continuous lines. Production of lysostaphin is indicated with broken lines. Lysostaphin production was induced at OD600 = 5 with 40 ng/mL nisin. Y, P, and L indicate the percentage of yeast extract, peptone and lactose, respectively, used in the culture medium. [Lss 5 40 1.5Y 2P 5L], conditions for lysostaphin production; [OD 5 40 1.5Y 2P 5L], conditions for growth.
Figure 7 Influence of induction conditions and the medium composition on lysostaphin yields. Growth of the culture is indicated with continuous lines. Production of lysostaphin is indicated with broken lines. The cultures were induced at an O600 = 4, 5 or 6 with 40 or 60 ng/mL nisin. Y, P, and L indicate the amount in % of yeast extract, peptone and lactose, respectively, used in the culture medium. [Lss 5 40 2Y 2.5P 7L], conditions for lysostaphin production; [OD 5 40 2Y 2.5P 7L], conditions for growth.
Summary of the optimization
The highest production of lysostaphin was observed under the following conditions: the pH is fixed at 7.0, the temperature is 30°C, the neutralizing agent is NH4OH, medium components are 7% lactose, 2.5% peptone, 2% yeast extract, 0.01% sodium phosphate (Na2PO4 *2 H2O), 100 μM ZnSO4, 1 mM MgSO4 and 0.1 mM MnSO4, cells are grown to an OD600 = 5 and induced with 40 ng/mL nisin. Lysostaphin production proceeded for 6 to 8 hours and then abruptly stopped. After that time there may be a small decrease of lysostaphin concentration over time. However, this depended on the medium composition and the fermentation conditions (Figures 4, 6 and 7).
Discussion
The nisin-controlled gene expression system NICE is widely used for a multitude of different applications. However, induction of this gene expression system has never been optimized for either laboratory conditions or for industrial-scale applications. In the present publication we show that the yield of the production of a heterologous protein can be increased at least three-fold by careful optimization of the fermentation and induction conditions. A number of general observations have been made:
(I) in a pH controlled culture, the point at which the pH is fixed and the neutralizing agent influence the general growth and biomass yield of the culture. A neutral pH and a mild neutralizing agent such as NH4OH was beneficial for growth of lactococci.
(II) Nisin induction works over a broad range of temperatures, even at 20°C. However, there is a clear correlation between the temperature of the fermentation and the speed and yield of the induction. The lower the temperature, the slower was the response to nisin and the lower was the yield of the heterologous protein. For lysostaphin production, a temperature of 30°C was found to be the optimum. Figure 4 may give a hint that it is worthwhile to also consider production at a lower temperature. At 25°C maximum lysostaphin production was slower and somewhat lower than at 30°C; however it seemed to be more stable in the cells, as can be seen at the end of the culture.
(III) Addition of extra and specific nutrients may be needed. Lysostaphin is a zinc-containing enzyme. Production of large amounts of active enzyme may be limited by the amount of zinc that is present in the different medium components. Therefore, extra zinc was added.
More generally, phosphate is needed for DNA biosynthesis and the energy metabolism of the cell. There are certain amounts of phosphorus compounds in the yeast extract and in the peptone, but growth at higher cell densities and product formation may require more phosphorus. Initially, little difference was seen after addition of phosphate, however, in an induction experiment at higher cell densities, strongly reduced product formation was observed (150 mg/L lysostaphin without extra phosphate versus 220 mg/L with phosphate). Zinc is an example of specific components that need to be selected and optimized for each target protein individually.
(IV) There is a strong correlation between the medium composition, the cell density of the culture at the moment of induction and the amount of nisin that is added. Induction is a dynamic process and needs growing cells both for the induction to occur and for the subsequent protein production to proceed as fast and as long as possible. In Figure 5 we saw increases in lysostaphin production after both the cell density for induction and the amount of nisin were increased. However, when the induction occurs too late in the growth cycle (e.g. OD600 = 7) no further product increase can be observed, even when more nisin was added. Similarly, if too much nisin is added (see Figure 7) it will become detrimental for product formation. Under the present conditions the maximum yield was reached at an OD600 around 5 with about 40 ng/mL nisin. These parameters can directly be used for the increased production of any other heterologous protein and as a starting point for further optimization of the respective process.
(V) As can be seen in Figure 6, lysostaphin production was still limited by the supply of nutrients in the fermentation medium. Addition of higher amounts of peptone, yeast extract and sugar significantly increased the production of lysostaphin. These parameters need to be carefully optimized, since they make up the highest costs of the growth medium.
The present publication outlines the optimization of controlled expression of heterologous genes in L. lactis using the NICE system. Careful optimization of a number of key parameters leads to a considerable increase in the overall yield of the target protein. The current outline gives a framework for this optimization. However, since every protein is different, a number of steps need to be checked and fine-tuned individually.
After the present round of optimization, the model protein lysostaphin could be produced up to 300 mg/L. Since lysostaphin has a growth-inhibiting effect on the host cells, and maybe also on its own production, a higher production capacity of the L. lactis host can be inferred for target proteins that do not have functionally detrimental effects on the host cells. Application of the described optimization to other target proteins will show the actual potential and limits of L. lactis for the industrial production of heterologous proteins.
Conclusion
The nisin-controlled gene expression system NICE of Lactococcus lactis is one of the most widely used Gram-positive gene expression systems. To date, no systematic study has been undertaken to optimize the system for maximum yields, especially for industrial scale applications. The present study shows that by careful optimization of growth, induction and target protein-specific parameters, an at least three-fold increase of the yield can be achieved.
Methods
Bacterial strains and growth media
Lactococcus lactis subsp. cremoris NZ3900 [26] carrying plasmid pNZ1710 (plasmid with gene for mature lysostaphin under control of the nisA promoter; [15]) was stored as frozen stock in M17 medium [27] containing 0.5% lactose as carbon source and plasmid selection agent. The basic fermentation medium contained 5% lactose (Lactochem, Borculo Domo Ingredients, Zwolle, The Netherlands), 1.0% yeast extract (Biospringer, Maisons-Alfort, France), 1.5% soy peptone (Merck, VWR International, Amsterdam, The Netherlands), 1 mM MgSO4 and 0.1 mM MnSO4. All components were dissolved in water and sterilized at 110°C for 20 min. Additional components such as Na2HPO4 and ZnSO4 were filter sterilized and added separately.
Nisin for induction was prepared as follows: 0.04% nisin powder (Sigma-Aldrich Chemie, Zwijndrecht, The Netherlands) was dissolved in 0.05% acetic acid and precipitated proteins were removed by centrifugation. Nisin was added for induction as indicated in the Results section.
Fermentations
The strain was taken from stock and sub-cultured twice before inoculation. 1-L Applicon fermenters were used coupled to an Applicon Biocontroller ADI 1030 (Applicon, Frederiksberg, Denmark) for pH and temperature control. The pH of the culture was controlled at the indicated values (see Results) with either 2.5 M NaOH or 2.5 M NH4OH. The cell density was measured in samples that were taken at regular intervals, by determining the absorbance at 600 nm with a path length of 1 cm [OD600]. For lysostaphin measurements, cells of appropriate samples were sedimented by centrifugation and stored at -20°C. After thawing, cell density was adjusted to OD600 = 10 (according to the initial cell density reading) and 1 ml of the cell suspension was mixed with 1 g glass beads (0.1 mm Zirconia/Silica beads from Biospec products, Bartlesville, OK., U.S.A.) in screw-cap Eppendorf tubes. Subsequently, cells were subjected to bead-beating in a FastPrep beadbeater (FP120, QBiogene, Irivne, CA, U.S.A.), adjustment 4, 4 × 30 s. After the beads were allowed to sediment, lysostaphin concentration was determined in the whole cell extract.
Lysostaphin measurement
Lysostaphin was quantified using SDS-capillary zone electrophoresis (SDS-CZE). The SDS sample buffer was prepared by dissolving in ca. 80 mL of water 606 mg of tris(hydroxymethyl)aminomethane (Tris), 1.00 g of sodium dodecylsulphate (SDS) and 37 mg of EDTA. Hydrochloric acid (0.1 M, 14.7 mL) was added and the solution was made up to 100 mL. Daily, 25 mg of DTT was added to 10 mL SDS sample buffer. This solution was used to dissolve cell extract samples (see below).
The SDS-CZE separation was performed using a Beckman Coulter P/ACE MDQ capillary electrophoresis system (CA, U.S.A.), equipped with a UV-detector operating at 214 nm using a 30 cm coated capillary and the separation buffer of the Beckman Coulter SDS 14–200 kit. The capillary temperature was set at 20°C.
Before each analysis, the 30 cm capillary was rinsed in the reverse direction for 1 min at 20 psi using 0.1 M HCl and subsequently for 3 min at 20 psi with SDS separation buffer. The sample solution was injected for 30 s at 1 psi, followed by forward rinse for 30 s at 0.5 psi of SDS sample buffer/water (1:1). The separation voltage was ramped in 1 min to 9 kV (ground at detector outlet) and held constant for 18 min.
A standard solution of lysostaphin was prepared as follows. From a solution of a known concentration of lysostaphin (ca. 1 mg/mL) 40 μL was pipetted into an Eppendorf vial of 0.5 mL and 120 μL of SDS sample buffer was added. The closed vial was incubated for 30 min at 80°C and subsequently rapidly cooled using ice water. From this solution 80 μL was transferred to the sample vial (of 0.2 mL Eppendorf vial).
For the preparation of samples of cell extracts, the same procedure as that for the standard solution of lysostaphin was followed, except that after cooling in ice water the solution was centrifuged for 5 min at 3000 g to remove any possible traces of precipitated proteins.
Element analysis
Analysis of minerals and trace elements in the basic medium was carried out by ICP-AES (Inductively Coupled Plasma – Atomic Emission Spectrometry) using the Vista Axial ICP of Varian (Palo Alto, CA, U.S.A.). The growth medium sample was prepared for ICP by dry-ashing, dissolved in nitric acid and measured against standards of Ca, Mg, Na, K, P, Fe, Zn, Mn and Cu.
Authors' contributions
IM and JM were the initiators and main supervisors of the experiments. KO developed the CE lysostaphin assay. IM and ES developed the optimization strategy.
Acknowledgements
We thank Bert van de Bunt and Fedde Kingma for excellent technical assistance. Furthermore, we thank Andy Lees for critical reading of the manuscript.
==== Refs
Teuber M Wood BJB and Holzapfel WH The Genus Lactococcus The Genera of Lactic Acid Bacteria 1995 2 London, Glasgow and others, Blackie Academic & Professional 173 1234
Konings WN Kok J Kuipers OP Poolman B Lactic acid bacteria: the bugs of the new millennium Current Opinion in Microbiology 2000 3 276 282 10851157 10.1016/S1369-5274(00)00089-8
Bolotin A Wincker P Mauger S Jaillon O Malarme K Weissenbach J Ehrlich SD Sorokin A The complete genome sequence of the lactic acid bacterium Lactococcus lactis ssp. lactis IL1403 Genome Res 2001 11 731 753 11337471 10.1101/gr.GR-1697R
Bolotin A Ehrlich SD Sorokin A Studies of genomes of dairy bacteria Lactococcus lactis Sci Aliment 2002 22 45 53
Guillot A Gitton C Anglade P Mistou MY Proteomic analysis of lactococcus lactis, a lactic acid bacterium Proteomics 2003 3 337 354 12627387 10.1002/pmic.200390047
de Vos WM Advances in genomics for microbial food fermentations and safety Current Opinion in Biotechnology 2001 12 493 498 11604327 10.1016/S0958-1669(00)00252-4
de Ruyter PG Kuipers OP Beerthuyzen MM Alen-Boerrigter I de Vos WM Functional analysis of promoters in the nisin gene cluster of Lactococcus lactis Journal of Bacteriology 1996 178 3434 3439 8655538
Kuipers OP de Ruyter PGGA Kleerebezem M de Vos WM Quorum sensing-controlled gene expression in lactic acid bacteria Journal of Biotechnology 1998 64 15 21 10.1016/S0168-1656(98)00100-X
Kleerebezem M Quadri LE Peptide pheromone-dependent regulation of antimicrobial peptide production in Gram-positive bacteria: a case of multicellular behavior Peptides 2001 22 1579 1596 11587786 10.1016/S0196-9781(01)00493-4
Makrides SC Strategies for achieving high-level expression of genes in Escherichia coli Microbiological Reviews 1996 60 512 538 8840785
Hannig G Makrides SC Strategies for optimizing heterologous protein expression in Escherichia coli Trends in Biotechnology 1998 16 54 60 9487731 10.1016/S0167-7799(97)01155-4
Platteeuw C van Alen-Boerrigter I van Schalkwijk S de Vos WM Food-grade cloning and expression system for Lactococcus lactis Appl Environ Microbiol 1996 62 1008 1013 8975595
Miyoshi A Poquet I Azevedo V Commissaire J Bermudez-Humaran L Domakova E Leloir Y Oliveira SC Gruss A Langella P Controlled production of stable heterologous proteins in Lactococcus lactis Appl Environ Microbiol 2002 68 3141 3146 12039780 10.1128/AEM.68.6.3141-3146.2002
Nouaille S Ribeiro LA Miyoshi A Pontes D Le Loir Y Oliveira SC Langella P Azevedo V Heterologous protein production and delivery systems for Lactococcus lactis Genet Mol Res 2003 2 102 111 12917806
Mierau I Leij P Swam I Blommestein B Floris E Mond J Smid EJ Industrial-Scale Production and Purification of a Heterologous Protein in Lactococcus lactis using the Nisin-Controlled Gene Expression System NICE: The Case of Lysostaphin Microbial Cell Factories 2005 submitted
Thumm G Götz F Studies on prolysostaphin processing and characterization of the lysostaphin immunity factor (Lif) of Staphylococcus simulans biovar staphylolyticus Molecular Microbiology 1997 23 1251 1265 9106216 10.1046/j.1365-2958.1997.2911657.x
von Eiff C Kokai-Kun JF Becker K Peters G In vitro activity of recombinant lysostaphin against Staphylococcus aureus isolates from anterior nares and blood Antimicrob Agents Chemother 2003 47 3613 3615 14576128 10.1128/AAC.47.11.3613-3615.2003
Wu JA Kusuma C Mond JJ Kokai-Kun JF Lysostaphin disrupts Staphylococcus aureus and Staphylococcus epidermidis biofilms on artificial surfaces Antimicrob Agents Chemother 2003 47 3407 3414 14576095 10.1128/AAC.47.11.3407-3414.2003
Sybesma W Starrenburg M Kleerebezem M Mierau I de Vos WM Hugenholtz J Increased production of folate by metabolic engineering of Lactococcus lactis Appl Environ Microbiol 2003 69 3069 3076 12788700 10.1128/AEM.69.6.3069-3076.2003
Pedersen MB Koebmann BJ Jensen PR Nilsson D Increasing acidification of nonreplicating Lactococcus lactis delta-thyA mutants by incorporating ATPase activity Appl Environ Microbiol 2002 68 5249 5257 12406711 10.1128/AEM.68.6.3010-3023.2002
Kashket ER Bioenergetics of lactic acid bacteria: cytoplasmic pH and osmotolerance FEMS Microbiology Reviews 1987 46 233 244 10.1016/0378-1097(87)90110-8
Vasiljevic T Jelen P Lactose hydrolysis in milk as affected by neutralizers used for the preparation of crude beta-galactosidase extracts from Lactobacillus bulgaricus 11842 Innovative Food Science and Emerging Technologies 2002 3 175 184 10.1016/S1466-8564(02)00016-4
Thumm G Götz F Studies on prolysostaphin processing and characterization of the lysostaphin immunity factor (Lif) of Staphylococcus simulans biovar Staphylolyticus Molecular Microbiology 1997 23 1251 1265 9106216 10.1046/j.1365-2958.1997.2911657.x
Gill RT DeLisa MP Valdes JJ Bentley WE Genomic analysis of high-cell-density recombinant Escherichia coli fermentation and "cell conditioning" for improved recombinant protein yield Biotechnol Bioeng 2001 72 85 95 11084598 10.1002/1097-0290(20010105)72:1<85::AID-BIT12>3.0.CO;2-H
ten Brink B Otto R Hansen UP Konings WN Energy recycling by lactate efflux in growing and nongrowing cells of Streptococcus cremoris Journal of Bacteriology 1985 162 383 390 2984179
Otto R Sonnenberg AS Veldkamp H Konings WN Generation of an electrochemical proton gradient in Streptococcus cremoris by lactate efflux Proc Nat Acad Sci U S A 1980 77 5502 5506
de Ruyter PG Kuipers OP Meijer WC de Vos WM Food-grade controlled lysis of Lactococcus lactis for accelerated cheese ripening Nature Biotechnology 1997 15 976 979 9335048 10.1038/nbt1097-976
Terzaghi BE Sandine WE Improved medium for lactic streptococci and their bacteriophages Applied Microbiology 1975 29 807 813 16350018
Koebmann BJ Nilsson D Kuipers OP Jensen PR The membrane-bound H(+)-ATPase complex is essential for growth of Lactococcus lactis Journal of Bacteriology 2000 182 4738 4743 10940012 10.1128/JB.182.17.4738-4743.2000
Curley P van Sinderen D Identification and characterisation of a gene encoding aminoacylase activity from Lactococcus lactis MG1363 FEMS Microbiology Letters 2000 183 177 182 10650223 10.1016/S0378-1097(99)00663-1
Lopez de Felipe F Kleerebezem M de Vos WM Hugenholtz J Cofactor engineering: a novel approach to metabolic engineering in Lactococcus lactis by controlled expression of NADH oxidase Journal of Bacteriology 1998 180 3804 3808 9683475
Looijesteijn PJ Boels IC Kleerebezem M Hugenholtz J Regulation of exopolysaccharide production by Lactococcus lactis subsp. cremoris by the sugar source Appl Environ Microbiol 1999 65 5003 5008 10543815
Geoffroy MC Guyard C Quatannens B Pavan S Lange M Mercenier A Use of green fluorescent protein to tag lactic acid bacterium strains under development as live vaccine vectors Appl Environ Microbiol 2000 66 383 391 10618252
Drouault S Juste C Marteau P Renault P Corthier G Oral treatment with Lactococcus lactis expressing Staphylococcus hyicus lipase enhances lipid digestion in pigs with induced pancreatic insufficiency Appl Environ Microbiol 2002 68 3166 3168 12039786 10.1128/AEM.68.6.3166-3168.2002
Chatel JM Langella P Adel-Patient K Commissaire J Wal JM Corthier G Induction of mucosal immune response after intranasal or oral inoculation of mice with Lactococcus lactis producing bovine beta-lactoglobulin Clin Diag Lab Immunol 2001 8 545 551 10.1128/CDLI.8.3.545-551.2001
Kunji ER Slotboom DJ Poolman B Lactococcus lactis as host for overproduction of functional membrane proteins Biochimica et Biophysica Acta 2003 1610 97 108 12586384
Cibik R Tailliez P Langella P Chapot-Chartier MP Identification of Mur, an atypical peptidoglycan hydrolase derived from Leuconostoc citreum Appl Environ Microbiol 2001 67 858 864 11157255 10.1128/AEM.67.2.858-864.2001
Ribeiro LA Azevedo VAC Le Loir Y Oliveira SC Dieye Y Piard JC Gruss A Langella P Production and targeting of the Brucella abortus immunodominant antigen L7/L12 by Lactococcus lactis: a first step towards new live vaccines against brucellosis Appl Environ Microbiol 2001
Pavan S Hols P Delcour J Geoffroy MC Grangette C Kleerebezem M Mercenier A Adaptation of the nisin-controlled expression system in Lactobacillus plantarum: a tool to study in vivo biological effects Appl Environ Microbiol 2000 66 4427 4432 11010894 10.1128/AEM.66.10.4427-4432.2000
Enouf V Langella P Commissaire J Cohen J Corthier G Bovine rotavirus nonstructural protein 4 produced by Lactococcus lactis is antigenic and immunogenic Appl Environ Microbiol 2001 67 1423 1428 11282586 10.1128/AEM.67.4.1423-1428.2001
Bermúdez-Humarán LG Langella P Cortes-Perez NG Gruss A Tamez-Guerra RS Oliveira SC Saucedo-Cardenas O Montes de Oca-Luna R Le Loir Y Intranasal immunization with recombinant Lactococcus lactis secreting murine interleukin-12 enhances antigen-specific Th1 cytokine production Infect Immun 2003 71 1887 1896 12654805 10.1128/IAI.71.4.1887-1896.2003
|
15921537
|
PMC1182390
|
CC BY
|
2021-01-04 16:24:35
|
no
|
Microb Cell Fact. 2005 May 30; 4:16
|
utf-8
|
Microb Cell Fact
| 2,005 |
10.1186/1475-2859-4-16
|
oa_comm
|
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-191598516210.1186/1475-2859-4-19ReviewBioinformatics in microbial biotechnology – a mini review Bansal Arvind K [email protected] Department of Computer Science, Kent State University, Kent, OH 44242, USA2005 28 6 2005 4 19 19 1 5 2005 28 6 2005 Copyright © 2005 Bansal; licensee BioMed Central Ltd.2005Bansal; 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 revolutionary growth in the computation speed and memory storage capability has fueled a new era in the analysis of biological data. Hundreds of microbial genomes and many eukaryotic genomes including a cleaner draft of human genome have been sequenced raising the expectation of better control of microorganisms. The goals are as lofty as the development of rational drugs and antimicrobial agents, development of new enhanced bacterial strains for bioremediation and pollution control, development of better and easy to administer vaccines, the development of protein biomarkers for various bacterial diseases, and better understanding of host-bacteria interaction to prevent bacterial infections. In the last decade the development of many new bioinformatics techniques and integrated databases has facilitated the realization of these goals. Current research in bioinformatics can be classified into: (i) genomics – sequencing and comparative study of genomes to identify gene and genome functionality, (ii) proteomics – identification and characterization of protein related properties and reconstruction of metabolic and regulatory pathways, (iii) cell visualization and simulation to study and model cell behavior, and (iv) application to the development of drugs and anti-microbial agents. In this article, we will focus on the techniques and their limitations in genomics and proteomics. Bioinformatics research can be classified under three major approaches: (1) analysis based upon the available experimental wet-lab data, (2) the use of mathematical modeling to derive new information, and (3) an integrated approach that integrates search techniques with mathematical modeling. The major impact of bioinformatics research has been to automate the genome sequencing, automated development of integrated genomics and proteomics databases, automated genome comparisons to identify the genome function, automated derivation of metabolic pathways, gene expression analysis to derive regulatory pathways, the development of statistical techniques, clustering techniques and data mining techniques to derive protein-protein and protein-DNA interactions, and modeling of 3D structure of proteins and 3D docking between proteins and biochemicals for rational drug design, difference analysis between pathogenic and non-pathogenic strains to identify candidate genes for vaccines and anti-microbial agents, and the whole genome comparison to understand the microbial evolution. The development of bioinformatics techniques has enhanced the pace of biological discovery by automated analysis of large number of microbial genomes. We are on the verge of using all this knowledge to understand cellular mechanisms at the systemic level. The developed bioinformatics techniques have potential to facilitate (i) the discovery of causes of diseases, (ii) vaccine and rational drug design, and (iii) improved cost effective agents for bioremediation by pruning out the dead ends. Despite the fast paced global effort, the current analysis is limited by the lack of available gene-functionality from the wet-lab data, the lack of computer algorithms to explore vast amount of data with unknown functionality, limited availability of protein-protein and protein-DNA interactions, and the lack of knowledge of temporal and transient behavior of genes and pathways.
==== Body
Background
In the last decade, the revolution in computer technology and memory storage capability has made it possible to model grand challenge problems such as large scale sequencing of genomes and management of large integrated databases over the Internet. This vastly improved computational capability integrated with large-scale miniaturization of biochemical techniques such as PCR, BAC, gel electrophoresis and microarray chips has delivered enormous amount of genomic and proteomic data to the researchers all over the world. This availability of data has led to an explosion of genome and proteome analysis leading to many new discoveries and tools that are not possible in wet-lab experiments.
The availability of genomic and proteomics data and improved bioinformatics and biochemical tools has raised the expectation of the humanity to be able to control the genetics by manipulating the existing microbes. The advantages are enormous such as better diagnosis of the diseases through the use of protein biomarkers, protection against diseases using cost effective vaccines [56,73] and rational drug design, improvement in agricultural quality and quantity, and the development of techniques that help us visualize and understand the detailed microbial machine at the systemic level.
Since the sequencing of the first complete microbial genome of Haemophilus influenzae in 1995 [29], hundreds of microbial genomes have been sequenced and archived for public research in GenBank through the concerted effort of federal health agencies such as NIH and DOE in USA, EMBL and EBI in Europe, DNA databank of Japan, national laboratories, academic universities, multiple drug development companies such as Celera and non-profit organizations such as TIGR, and companies involved in agricultural industry and bioremediation. The sequencing of human genome [68] and other relevant eukaryotic genomes has raised the expectation of understanding host pathogen interaction for the development of better vaccines and rational drugs to control the gene level and pathway level aberrations that are responsible for pathogenesis.
Except for the availability of bioinformatics techniques, the vast amount of data generated by genome sequencing projects would be unmanageable and would not be interpreted due to the lack of expert manpower and due to the prohibitive cost of sustaining such an effort. In the last decade bioinformatics has silently filled in the role of cost effective data analysis. This has quickened the pace of discoveries, the drug and vaccine design [56] and the design of anti-microbial agents [40]. In addition bioinformatics analysis has enhanced our understandings about the genome structure and the microorganism restructuring process.
Bioinformatics analysis will facilitate and quicken the analysis of systemic level behavior of cellular processes, and to understanding the cellular processes in order to treat and control microbial cells as factories. For the last decade, bioinformatics techniques have been developed to identify and analyze various components of cells such as gene and protein function, interactions, and metabolic and regulatory pathways. The next decade will belong to understanding cellular mechanism and cellular manipulation using the integration of bioinformatics, wet lab, and cell simulation techniques. More recently, researchers have started using these techniques for the production of recombinant proteins [48]. It is anticipated that in this decade, the semi-automated study of cellular behavior at systemic level will accelerate this capability.
Review
In the last decade, Bioinformatics has been used for the microbial biotechnology in many ways: computationally analyzing the wet-lab data, genome sequencing, identification of protein coding segments [6,24,41,64], and genome comparison to identify the gene function [4,5,11,25,35,46,53,70,71], the development of genomic and proteomics databases [8,9,16,21,33,49,62,63], and inference of phenotypes (higher level functions) from genotypes (gene level functions). In order to understand higher level functions four major studies have been undertaken: (i) automated reconstruction and comparison of metabolic pathways [12,14,18,38,49,58,59,65], (ii) study of protein-protein and protein-DNA interactions to understand regulatory pathways [2,7,15,27,28,30,42-44,47,55,60,61,66], (iii) modeling 2D and 3D structure of proteins [10,31,52,57,67], and (iv) modeling the docking of 3D models of proteins with drugs [34]. Understanding 3D structure of proteins has a major impact in understanding protein-protein interactions. Protein-protein and protein-DNA interactions will provide a good understanding of binding sites in signaling pathways; understanding the interactions between proteins and chemical compounds has already facilitated the development of drug design.
Three approaches have been used in bioinformatics: (i) use of computational search and alignment techniques [4,5,53,70] to compare new genome against the set of known genes to annotate the structure and function of genes in a newly sequence genome, (ii) the use of mathematical modeling techniques such as data mining, statistical analysis, neural networks, genetic algorithm, and graph matching techniques to identify common patterns, features and high level functions, and (iii) an integrated approach that integrates search techniques with mathematical modeling.
Genome sequencing
The major contribution of the bioinformatics in genome sequencing has been in the: (i) development of automated sequencing techniques that integrate the PCR or BAC based amplification, 2D gel electrophoresis and automated reading of nucleotides, (ii) joining the sequences of smaller fragments (contigs) together to form a complete genome sequence, and (iii) the prediction of promoters and protein coding regions of the genome.
PCR (Polymerase Chain Reaction) or BAC (Bacterial Artificial Chromosome) based amplification techniques derive limited size fragments of a genome. The available fragment sequences suffer from nucleotide reading errors, repeats – very small and very similar fragments that fit in two or more parts of a genome, and chimera – two different parts of the genome or artifacts caused by contamination that join end to end giving a artifactual fragment. Generating multiple copies of the fragments, aligning the fragments, and using the majority voting at the same nucleotide positions solve the nucleotide reading error problem. Multiple experimental copies are needed to establish repeats and chimeras. Chimeras and repeats are removed before the final assembly of the genome-fragments. The joining of the fragments is modeled as a mathematical weighted graph where nodes are fragments and the weights of edges are the number of overlapping nucleotides, and the fragments are joined based upon maximum overlap using a greedy algorithm [46,70]. In a greedy algorithm, most nodes having maximum (or minimum) scores are collapsed first. To join contigs, the fragments with larger nucleotide sequence overlap are joined first.
Automated identification of genes
After the contigs are joined, the next issue is to identify the protein coding regions or ORFs (open reading frames) in the genomes. The identification of ORFs can be done in three ways: (1) using Hidden Markov Model (HMM) based techniques such as GLIMMER [24] and GeneMark [41], (2) by searching the known database of genes such as GenBank to identify genes, and (3) the use of algorithms based on decision trees that identify start codons [64] and stop codons of the coding regions. HMM based techniques develop multiple probabilistic state machines each capable of identifying an ORF. Each machine predicts the next nucleotide character using a state transition with maximum probability and matches the predicted nucleotide character with the current nucleotide character in the actual sequence. Statistical training using known sample sequences derives the probability of state transition. In the case of microbial genomes, the HMM based software such as GLIMMER has provided 95% – 97% accuracy.
Identifying gene function: searching and alignment
After identifying the ORFs (open reading frames), the next step is to annotate the genes with proper structure and function. The function of the gene has been identified using popular sequence search and pair-wise gene alignment techniques. The four most popular algorithms used for functional annotation of the genes are BLAST [4] and its variations [5], dynamic programming technique Smith-Waterman alignment [70] and its variations, indexing based scheme FASTA [53] and its variations, and BLOCKS [35] that uses multiple sequence alignment of conserved domains to identify motifs – characterizing patterns of proteins.
BLAST search is based upon expanding multiple probable seed points (longer than four nucleotides) that match (with the help of scoring matrices such as BLOSUM or PAM [46,70]) to identify the largest matching nonrandom segment. Scoring matrices have positive match-value for the amino acids that have common biochemical or biophysical properties and negative match-values if the amino acids do not share biophysical or biochemical properties. Substitution matrices such as BLOSUM (BLOcks SUbstitution Matrix) have been derived by statistically comparing the frequency patterns of the amino acids occurring in conserved domains of protein families. Nucleotide sequences use a nucleotide matrix for scoring that penalizes non-matching positions. BLAST algorithm has near linear time complexity, and the current implementations are fast. However, in order to enhance computational efficiency, BLAST algorithm uses most probable combinations of nucleotide seeds to index the sequences in the database sacrificing some accuracy.
BLAST algorithm has gone through many improvements in heuristics to improve the execution speed, accuracy, and the dependence on predefined scoring matrices. Two major improvements are: (i) the use of two or more hits within a matching region before extending the high scoring segment, and the use of multiple iteration of matching to derive a position specific scoring matrix to be used in place of predefined biochemical matrix. PSI-BLAST (Position Specific Iterative BLAST) [5] is a popular implementation of BLAST that uses both these improvements. The use of two hits improves the execution efficiency in the segment extension, and the use of position specific matrix improves the search for weakly homologous sequences in evolutionary distant species. Position specific matrix is built by deriving multiple sequence alignment of the best matching segments and analyzing the frequency of the amino acid substitutions in the matching segments.
Dynamic algorithms such as Smith-Waterman [70] and other indexing schemes [53] are more accurate for pair-wise gene alignment. The alignment of gene-pairs using dynamic algorithms is based upon incremental matching by maximizing the sum of the score of the best alignment of the preceding subsequences and the score of matching the current amino-acid characters (or nucleotide characters). The mismatches in amino-acid sequences are penalized using scoring matrices such as BLOSUM or PAM [46,70]; the nucleotide sequences use a nucleotide matrix for scoring that penalizes non-matching positions. A gap is inserted to show the insertion and deletion of nucleotides (or amino-acids). Gaps are not part of a substitution matrix, and are provided as parameters by users. The presence of a gap also results into score penalty. There are two major types of protein (or gene) alignments using dynamic programming: global and local. In global alignment, the amino-acid (or nucleotide) characters are placed to maximize the overall score. In contrast local alignment finds the segment with the maximum score, and the segments with negative scores are ignored. For comparing amino acid sequences from evolutionary distant organisms, local alignment is preferred to take care of large-scale amino-acid variations. Global alignment fares well when small amount of random mutation is involved. Due to the pair-wise comparison of all characters in an amino acid sequence to identify best matching subsequence, all dynamic programming techniques have quadratic time complexity making them less suitable for large-scale pair-wise genome comparisons unless preprocessed by BLAST to remove dissimilar genes [11].
Multiple sequence alignment techniques [22] compare multiple homologous genes (genes that have similar sequences) to derive conserved segments and to derive evolutionary tree. The technique uses the integration of pair-wise alignment between two homologs and the notion of distance between two nucleotide sequences or between two amino-acid sequences. The notion of distance can be derived either as an edit distance – number of mismatches derived after pair-wise alignment of two sequences, or as the evolutionary distance between two microorganisms given by an evolutionary tree. The technique is based upon progressive pair-wise comparison to make intermediate alignments between nearest neighbors – homologs having shortest distance, and has been implemented as a greedy algorithm. ClustalX [22] is a popular multiple sequence alignment technique that has been used to identify conserved portions in a gene, and to develop new evolutionary trees [36].
A major source of problem in the above sequence comparison techniques is the assignment of user defined equal weight to indels (gaps) that undermines the importance of a specific amino-acid(s) or a group of amino-acid characters would have. Another minor problem is the presence of repeat characters in the sequences as the repeat characters only show the functional or structural separation of the component units within a gene, and can not be mixed with other amino-acid characters.
Multiple sequence comparison techniques such as BLOCK [35] have been used to identify the conserved subsequences in very similar gene sequences, and are good to derive motifs. Motifs – a set of unique subsequences characterizing a protein – have been found very useful to identify genes with the same functionality. Motifs are derived by identifying the conserved subsequences of the functionally equivalent genes from multiple organisms after aligning the sequences.
Protein domain is the basic unit of protein function and is associated with a unique pattern (possibly one) of folding (alpha helix, beta sheet or their variations) at the structure level. The researchers have used multiple sequence alignment and HMM to identify the regions that are individually homologous to each other in multiple homologous genes. These regions are probable domains. Currently there are many domain related databases such as PRODOM [21], Pfam [16] and SMART [39] (also see ). Pfam [16,63] (and ) is a database of multiple alignments of protein domains or conserved protein regions. The alignments represent some evolutionary conserved structure that has implications for the protein's function. Profile hidden Markov models (profile HMMs) built from the Pfam alignments are useful for automatically recognizing that a new protein belongs to an existing protein family, even under weak homology. Currently Pfam is derived automatically by cluster analysis of PRODOM database.
The sequence search based techniques assume that best sequence is sufficient to annotate the function. This assumption is generally true. However, in many cases best sequence match fails to identify the function due to: (1) function being localized to a specific area in the protein such as hydrophobic region, (2) the function being dependent on the presence of specific pattern of amino acids, or (3) function being dependent to a specific 3D conformational state in a multi domain protein. Sometimes mutation of few nucleotides alters the corresponding amino acids resulting into a different 3D conformation of a protein. Another limitation of best match techniques is that they cannot identify all possible functions of a multi-domain protein. A protein may have multiple domains, and may be multifunctional. The problem is more complex as there is no direct correlation between the number of domains in a protein and the number of functionality [32,37].
3D structure modeling and docking
A protein may live under one or more low free-energy conformational states depending upon its interaction with other proteins. Under a stable conformational state certain regions of the protein are exposed for protein-protein or protein-DNA interactions. Since function is also dependent upon exposed active sites, protein function can be predicted by matching the 3D structure of an unknown protein with the 3D structure of a known protein [10,71]. However, 3D structures from X-ray crystallography and NMR spectroscopy are limited. Thus there is a need for alternate mechanism to match genes. Generally there is close correspondence between gene sequence and 3D structure. In such cases sequence matching is sufficient for function annotation. However, many times multiple sequences map to the same 3D structure; the lack of matching of amino acid sequences does not exclude same 3D structure. In such cases matching 2D structure [57,66] – patterns of alpha helix and beta sheets – and matching 3D structures is needed to verify the function of the newly sequenced protein [71].
There are two major approaches to model 3D structure of a protein: (i) sequence homology based prediction and (ii) ab initio (or de novo) method. The sequence homology approach uses sequence alignment to identify the best matching 3D structure for different components: conserved portion, loop portion and side chains from the database, and threads them to predict the overall 3D structure. The ab initio method is based upon energy minimization principle, and predicts the structure from the sequence alone [10]. Recent advances in ab initio methods integrate the biochemical and biophysical properties such as folding of beta sheets and the information of hydrophobic regions to achieve better accuracy.
Docking is a term used to identify best matches between 3D structures of two molecules (receptor and ligand) that bind to each other by simulating interacting surfaces and free energy minimization at the domain level [34]. Docking problem requires modeling of surfaces using spheres (or grids) and identifying the best match that will fit two surfaces without excessive intersection. Many times biochemical information such as binding sites is provided. There are three major problems in docking: (i) for multi-domain proteins conformation may change during docking, (ii) docking algorithms have high computational overhead that makes large-scale modeling quite slow, and (iii) docking algorithms suffer from over prediction that results in a high number of false positives.
Pair-wise genome comparison
After the identification of gene-functions, a natural step is to perform pair-wise genome comparisons. Pair-wise genome comparison of a genome against itself provides the details of paralogous genes – duplicated genes that have similar sequence with some variation in function. Pair-wise genome comparisons of a genome against other genomes have been used to identify a wealth of information such as ortholologous genes – functionally equivalent genes diverged in two genomes due to speciation, different types of gene-groups – adjacent genes that are constrained to occur in close proximity due to their involvement in some common higher level function, lateral gene-transfer – gene transfer from a microorganism that is evolutionary distant, gene-fusion/gene-fission, gene-group duplication, gene-duplication, and difference analysis to identify genes specific to a group of genomes such as pathogens, and conserved genes [11,13].
To derive orthologs and sets of gene-groups, genomes are modeled as an ordered set of genes, and a pair of genomes is modeled as a bipartite graph where each node in one set is connected to homologous nodes – similar genes using pair-wise gene-alignment – in the second set. Orthologs are derived as the best matching homologs. To identify homologous gene-group, two neighboring genes in one genome that are homologous to two neighboring genes in the other genome are identified, a window consisting of neighboring genes is created in both the genomes and slided until the next gene in the first genome has no homologous gene in the corresponding neighborhood window in the second genome. After a non-matching gene is identified, the matching genes are collected as one gene-group.
The detailed comparative study [11,12,14] has shown that: (i) a large percentage of these gene-groups are co-transcribed or co-regulated [11,26], (ii) there are multiple types of gene-groups in a genome, (iii) the order of homologous genes in a gene-group is not always the same in two microorganisms, (iv) gene-groups are duplicated a lot, (v) all the genes in ordered gene-group are embedded in the same pathway, and unordered gene-groups occur at the junction points of adjacent pathways [12], (vi) larger genomes share more genes-groups despite not being evolutionary too close, (vii) gene-duplication and gene-insertion/gene-deletion are common means of genome restructuring, and horizontal gene-transfer and gene fusion are not uncommon, and (viii) gene duplication occurs mainly for the genes involved in cell surface interaction, nutrient transport, and sensor proteins. The rationale for duplication is a need to adapt under different external conditions and the use of similar mechanism for multiple sensors and transport proteins. The knowledge of genes specific to pathogens, genes inserted/deleted from pathways that are homologous to genes in the plasmids, and conserved genes are very useful to identify candidates for vaccine development and anti-microbial agents [11,56,73].
An interesting observation of pair-wise genome comparison studies has been that genome restructuring occurs by a combination of insertion/deletion, duplication, and fusion of domains as well as genes. However, the domain level comparative analysis tools are in the stage of infancy due to computational complexity and the limited availability of domain level functional information about various genes from the wet-labs.
Reconstructing metabolic pathways
Identification of gene functionality has started a new level of bioinformatics research: automated reconstruction and comparison of pathways of newly sequence organisms [12,14,18,38,49,50,58,59,65]. There have been many efforts and approaches related to pathway reconstruction. The three major approaches can be classified as: (i) global network of reactions catalyzed by enzymes, (2) network of gene-groups connected through the reactions catalyzed by enzymes embedded in the gene-groups, (3) global modeling of chemical reactions in the microbial cells.
The first approach [49] uses the knowledge of known biochemical pathways and enzymes [9,33], identifies the enzyme function of new genes in a newly sequenced genome using BLAST based search or using pair-wise genome comparison of evolutionary close genomes [65], and matches the product and substrate of chemical reactions catalyzed by enzymes to build the network of reactions [18]. This approach is quite powerful. However, it has many drawbacks: (i) it can not disambiguate the exact position in pathways for homologous genes, (ii) it does not take into account genes occurring in the same pathway due to gene-grouping and co-transcription, and (iii) it does not take into account the reaction rate.
The knowledge of gene-groups [11,26] has been used to develop an integrated approach for reconstructing metabolic pathways [12,14,65]. In this approach there are four steps: (i) identifying the enzymes and their functions in a newly sequenced genome using ortholog analysis, (ii) identifying the co-transcribed gene-groups – groups of genes sharing a common promoter – by analyzing the promoter region of the genes, (iii) deriving the gene-groups by pair-wise comparison of newly sequenced genome with multiple genomes, and (iv) using biochemical knowledge of existing pathways and enzymes [9,33] to connect network of gene-groups. The intergenic distance – distance between the stop codons of the preceding gene and the start codons of the following gene – in co-transcribed gene-groups (possible operons) is generally less than 75 nucleotides except for the leading gene. By computationally comparing the intergenic distance most of these possible co-transcribed gene-groups are identified. However, the knowledge of co-transcribed gene-groups in itself is insufficient to identify pathways since (i) co-transcribed gene-groups may have missing genes due to conservative estimate of cutoff threshold, (ii) multiple adjacent co-transcribed gene-groups in the same pathway may be separated due to gene insertion/deletion caused by genome restructuring, and (iii) some of the regulating genes that regulate pathways and are in close proximity are not picked up. These three problems are reduced by taking union of genes in the same gene-group derived from multiple pair-wise genome comparisons with the newly sequenced genome. The overall gene-groups are identified by merging the information derived from promoter based analysis and pair-wise genome comparison analysis [14]. Since gene-groups in a pathway are scattered across the genome, the gene-groups are networked to each other by matching the biochemical product and substrates in the reactions catalyzed by the enzymes embedded in the gene-groups using enzyme databases [9,33]. This scheme improves the computational efficiency, reduces the ambiguity of homologous genes, and includes many regulatory genes involved with a pathway. However, this scheme does not model cell level behavior as the notion of reaction rate is missing.
The third approach [50,58] is based on modeling the biochemical reactions globally involving products, byproducts and the effect of cofactors on the reaction rate [59]. The model is based upon representing the network of metabolic reaction as a set of vector of reactions called extreme pathways that correspond to study state flux distribution in a metabolic network needed to synthesize target products. In this technique the whole network of pathways is modeled as a matrix where the rows are extreme pathways and columns represent specific reactions. This technique is useful to model the overall metabolic behavior within a microbial cell.
Current metabolic pathway techniques are limited by the available gene-functions from wet-laboratories. Another issue is that the identification of metabolic pathways is not sufficient unless the reaction rates and the effect of stress response over the reaction rates are known. While there have been recent approaches to model the reaction rate of metabolic pathways [59], the complete picture cannot be verified largely due to unavailability of gene-functions from wet-labs.
Phenotype similarity and automated pathway comparisons
The next level of study that the researchers have taken is to compare the similar pathways to understand the effect of insertion and deletion of genes in various microorganisms and to understand the evolution at pathway level [38]. To compare two pathways, the genes in the pathway are aligned as follows. Two pathways match completely if every protein in the first pathway (or a gene-group within a pathway) has a corresponding homologous gene in another pathway (or the gene-group within the pathway). There is a gap if a homologous gene is deleted (inserted), and there is a mismatch if the corresponding homologous genes have a low similarity score. Based upon this modeling, comparison of H. pylori and yeast has shown many similar pathways. More importantly, a quantification mechanism has been found to compare two pathways.
Derivation of regulatory mechanism and pathways
The genomics and proteomics research front has progressively moved from metabolic pathway reconstruction to the identification of signaling pathways and promoter analysis to identify transcription factors for protein-DNA interactions.
There are four major approaches to study protein-DNA interactions: (i) micro-array analysis of gene-expressions under different stress conditions of cells, (ii) statistical analysis of promoter regions of orthologous genes (functionally equivalent genes in different organisms identified as best homologs), (iii) global analysis of frequency patterns of dimers in the intergenic region – promoter region occurring between adjacent protein coding regions – of a genome, and (iv) biochemical modeling at the atomic bond level to understand how a protein will bind to nucleotides. Only the microarray analysis technique is based upon experimental data, and other three approaches are based on mathematical modeling and sequence analysis.
Micro array analysis [69] measures the relative change in the gene-expressions for a stressed (or a stimulated) cell and a change in cellular expression pattern – differentiation, cellular cycle, tissue remodeling, sporulation etc – in response to change in stimuli using a two step process: (i) mapping all the genes in the same genome etched on a thin glass plate and hybridizing the genes of a healthy cell with etched genes to derive the regular gene expression under equilibrium condition, and (ii) hybridizing the affected cells with etched genes to derive the gene expression of affected cells under equilibrium condition. Comparative study of gene expressions under normal condition and under a stimulated (or stressed) condition provides the information about the affected genes. Under the assumption that auto regulation in a gene-group and any cyclic self-regulation is absent, the interaction between protein and transcription factors is responsible for the observed increase or decrease of gene-expressions. This gene-expression data is analyzed using (i) cluster analysis [69] to identify meaningful patterns of gene-expressions, or (ii) data mining techniques – a statistical technique that associates and correlates expressed genes and different stress conditions.
The second approach of statistical promoter analysis [30,43,44] first identifies the orthologous genes from evolutionary close microorganisms [11] with active pathways using pair-wise genome comparisons databases (see ) or using the knowledge of cluster of orthologs (COGS) – a group of genes in a super family archived at NCBI at NIH that has been derived by multiple genome comparisons. In the next step, the upstream region between two genes of the orthologs are identified and compared to identify statistically conserved patterns. Under the assumption that functionally equivalent genes in the very similar pathways of evolutionary close organisms will have similar regulation mechanism, the transcription factors – regions of promoters involved in enhancing or repressing the gene-expression of the associated gene – for protein-DNA interaction in the promoters of orthologous genes would also be very similar. This analysis has led to discovery of many transcription factors.
The third approach [47] has been to extract and statistically analyze the dimers in the intergenic region in a whole genome and plot the frequency of occurrence. The non-random dimers that occur more frequently are possibly involved in protein-DNA interactions.
The biochemical approach [42] studies the protein-DNA interactions at the atomic bond level by considering hydrogen bonds in amino-acid base interactions, Van der Wall forces at contacts and water mediated bonds at different levels of proximity of two molecules. Based upon the analysis of the bonds and the actual statistical results, it has been concluded that amino-acid base interaction plays a major role in binding, Van der Wall forces provide stabilization, and protein-DNA interactions are complex and biased: different amino-acids have preferences for certain types of bases. For example, arginine, lysine, histidine and serine have preference for guanine.
Currently no researcher has attempted a hybrid approach integrating biochemical approach with other four approaches. An integrated approach will give a better overall picture. Another complex problem is that a co-regulated gene may have more than one transcription factor; some of these transcription factors may be individually weak and may be correlated with other transcription factors. An approach to identify the weak transcription factor is a two step process: (i) first identify the strong related transcription factor using one of the previous approaches followed by (ii) a pattern search in the neighborhood of the strong pattern [27].
Figuring out the connectivity in protein-protein interactions to derive signaling pathway has been a long drawn challenge. Recently, in last two years, two approaches have emerged: (1) integration of microarray analysis and entropy based modeling to derive gene clustering of the genes involved in the same regulatory pathway [2,7], and (2) technique based upon random algorithms maximizing transition probability. The first approach computes the mutual information of all the gene-pairs, and clusters the protein groups having more mutual information above a threshold [20]. The mutual information is entropy based approach, and is derived by the cumulative sum of the frequency patterns of occurrence of gene-pairs. To derive entropy, gene-expressions are divided into discrete histograms, and the mutual information between every gene-pair is computed [20]. Higher mutual information means direct correlation of the genes. It has been statistically found that genes that belong to the same pathway tend to group together. Using this cluster analysis, many signaling pathways have been identified in yeast-based system [15]. The analysis is a general-purpose technique, and can be used both in prokaryotic as well as eukaryotic systems.
Even figuring out the connectivity will not be able to answer the transient temporal behavior of many genes involved in the regulation mechanism and auto-regulation mechanism of operons – co-transcribed gene-group within a pathway involved in a common functionality. The modeling of transient behavior of genes cannot be captured by hybridization based microarray analysis since the data corresponds to equilibrium state of reactions. To understand the malfunctioning cells and cells of pathogenic bacterial strains, the overall organization and behavior including transient behavior and stress responses have to be studied.
Microbial evolution revisited
Bioinformatics researchers have compared extensively multiple genomes to correlate and classify the genomes into various families and to study evolution. It has been established by many researchers that overall evolution is a combination of point based mutation giving rise to speciation and restructuring of genomes based upon gene duplications, gene insertion, gene deletion, gene-fusion/fission, horizontal gene transfer, and domain level restructuring [11,17,45].
The evolutionary study efforts can be classified into three approaches: (1) point based mutation approaches used to build traditional evolutionary tree using multiple sequence alignment of 16SrRNA [72], (2) study of genome restructuring based upon inversion and transposition at the gene level [17,45], and (3) the study based upon whole genome comparisons using gene identity of orthologous genes across multiple microbial genomes [13].
The 16SrRNA approach is rooted in the concept of point mutation of conserved genes due to their slow mutation rate, uses 16SrRNA database and multiple sequence alignment [22], and uses neighbor join algorithm [36] to build an evolutionary tree. Before microbial genomes were sequenced, this technique was considered quantitatively sound, and using 16SrRNA database three distinct domains – bacteria, archaea, and eukaryotes – were identified. Archea domain is hyperthermophilic, and its 16SrRNA is somewhat different from 16SrRNA of bacteria.
Since 1998, after the availability of multiple microbial genomes, the researchers have tried to build the evolutionary tree by comparing other highly conserved genes. The results have shown that the evolutionary tree varies a lot depending upon the choice of the conserved genes, and shows no clear distinction between archaea and bacteria. This observation combined with the knowledge of genome restructuring caused by domain level and gene level restructuring such as horizontal gene transfer has shaken up the traditional evolutionary trees based upon point mutations in 16SrRNA [54].
The second approach uses the genome rearrangement caused by gene shuffling as a measure for the genomic distance between the two organisms [17,45]. Gene shuffling is caused by inversion and transposition. This scheme is based upon the distance measure as the breakaway from the standard gene-order in two genomes. Under this scheme the breakaway distance for each orthologous gene is added to give a cumulative score for the genome. This score is used as a distance between two genomes. Building large scale evolutionary tree using this approach was cost prohibitive due to pair-wise comparison until recently when a new development in parallel algorithms made such an evolutionary tree possible [45]. Is this scheme horizontal transfer of genes do not play a role: insertion and deletion are not counted in the assumption, and duplications are mapped to a single gene. It has been shown that duplication, insertion, deletion of gene-domains and genes are a major component of evolution [11]. Specially duplicated genes are involved in multiple sensor and transportation pathways such as ABC transporters, and cannot be ignored.
The third approach [13] is based on comparing overall gene-content of functionally equivalent genes to identify the cumulative similarity of two genomes. The data is normalized to take care of different size of genomes. The major assumption in this scheme is that conserved genes are very few and do not give a consensus, and slow mutation rate only contributes to good multiple sequence alignment. Whole genome comparisons can balance out the error introduced by comparing a single conserved gene. The results show that the overall amino-acid composition in the microorganisms does not differ significantly between archaea and bacteria to give a separate domain status to archaea [13]. In addition, the composition of other hyperthermophilic bacteria cannot be distinguished from archaea.
Currently no proteomic level approach has been suggested to classify the genomes. In future, one such approach could be based upon comparative analysis and alignment of pathways of multiple genomes [38]. Under this scheme, after the pathways are aligned, a combination of the cumulative number insertion and deletion of genes in the pathways, gene duplication in the same pathway, and gene shuffling could be used to describe the distance between two genomes since all three factors are directly involved in the pathway variations. However, exact mechanism of combining these three components of pathway evolution has to be studied.
Conclusion
Despite being a young field, bioinformatics has helped both fundamental microbiology and biotechnology through the development of algorithms, tools, and discoveries refining the abstract model of microbial cell functioning. The major impact of the bioinformatics has been in automating the microbial genome sequencing, the development of integrated databases over the Internet, and analysis of genomes to understand gene and genome function. BLAST based database search and Smith-Waterman based gene-pair alignment algorithm and their variations are being used extensively in comparing genes and genomes, and have become the first steps to derive the gene-function and the functionality of genomes. Significant success has been achieved in comparative genome analysis to: (i) identify conserved function within a genome family, (ii) identify specific genes in a group of genomes, and (iii) model 3D structures of proteins and docking of biochemical compounds and receptors. These successes have direct impact in the development of anti microbial agents, vaccines, and rational drug design. By integrating the knowledge of orthologs and gene-functions, gene-grouping based upon the integration of pair-wise genome comparison, and co-transcribed gene-groups, and graph based matching of substrates and products catalyzed by enzymes metabolic pathways reconstruction has been nearly automated.
The current front has moved to the identification of regulatory pathways, identification of protein-protein interactions, protein-DNA interactions, protein-RNA interactions, and simulations of metabolic reactions to study the effect of reaction rates, and the analysis of experimental data available from micro-array data to study the correlation between the gene-expressions and stress conditions.
Most of the bioinformatics techniques are critically dependent upon the knowledge derived from wet laboratories and the available computational algorithms and tools. Unfortunately, both the resources have limited capability of handling a vast amount of data to interpret genomics and proteomics with so many unknowns. Since there is a limited set of gene-functions available from the wet lab data, there are many holes in the complete picture of gene functions in many newly sequenced genomes. A lack of integration of bioinformatics research with biochemical knowledge also contributes to the holes in the complete picture.
The mathematical modeling approaches are suitable for new discoveries to derive candidate genes for vaccine and rational drug design, metabolic pathways, metabolic pathway variations, and transcription factors for regulatory pathways. However modeling results contain many false positives and false negatives. These results need to be verified and cured by wet-lab experiments. However, complete verification is becoming humanly impossible due to the unavailability of experts, resources, and problems in co-ordination and ever changing bioinformatics databases caused by new analysis and discoveries [51].
With the availability of better cell visualization techniques and the abstract genomics models based upon current bioinformatics analysis and their integration with existing biochemical knowledge, the microbial wet lab experiments will become more focused in their goal. The progress in bioinformatics and wet-lab techniques has to remain interdependent and focused complementing each other for their own progress and for the progress of biotechnology in future. In future more and more focus would be to apply the techniques in an integrated way to manipulate the microbial cells at systemic level.
Authors' contributions
AB is the sole contributor of this original review article. The review is based upon the published current research in the area of microbial bioinformatics.
Acknowledgements
Due to the limited scope of this review, the author was restricted to acknowledge a small number of relevant publications contributing directly to microbial bioinformatics. The author acknowledges all the microbial bioinformatics researchers especially in algorithm development and eukaryotic research such as yeast-based models who have helped (directly or indirectly) microbial bioinformatics research. The author is also grateful to anonymous referees for their helpful insights during the review process.
==== Refs
Akutsu T Miyano S Kuhara S Algorithms for Identifying Boolean Networks and Related Biological Networks Based on Matrix Multiplication and Fingerprint Function Journal of Computational Biology 2000 7 331 343 11108466 10.1089/106652700750050817
Aloy P Russell RB Ten Thousand Interactions for Molecular Biologists Nature Biotechnology 2004 22 1317 1321 15470473 10.1038/nbt1018
Apweiler R Interpro – an integrated documentation resource for protein families, domains and functional sites Bioinformatics 16 1145 50 11159333 10.1093/bioinformatics/16.12.1145
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic alignment search tools Journal of Molecular Biology 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
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 Research 1997 15 3389 3402 9254694 10.1093/nar/25.17.3389
Azad RK Borodovsky M Probabilistic methods of identifying genes in prokaryotic genomes: connections to the HMM theory Brief Bioinformatics 2004 5 118 30 15260893
Bader JS Chaudhary A Rothberg JM Chant J Gaining confidence in high-throughput protein interaction networks Nature Biotechnology 2004 22 78 85 14704708 10.1038/nbt924
Bairoch A PROSITE: a dictionary of sites and patterns in proteins Nucleic Acids Research 1991 19 2241 2245 2041810
Bairoch A The ENZYME database in 2000 Nucleic Acids Research 2000 304 305 10592255 10.1093/nar/28.1.304
Baker D Sali A Protein structure prediction and structural genomics Science 2001 294 93 96 11588250 10.1126/science.1065659
Bansal AK An automated comparative analysis of seventeen complete microbial genomes Bioinformatics 1999 15 900 908 10743556 10.1093/bioinformatics/15.11.900
Bansal AK Woolverton C Applying automatically derived gene-groups to automatically predicts and refines microbial pathways IEEE Transactions of Knowledge and Data Engineering 2003 15 883 894 10.1109/TKDE.2003.1209006
Bansal AK Meyer TE Evolutionary analysis by whole genome comparisons Journal of Bacteriology 2002 184 2260 2272 11914358 10.1128/JB.184.8.2260-2272.2002
Bansal AK Integrating co-regulated gene-groups and pair-wise genome comparisons to automate reconstruction of microbial pathways IEEE International Symposium on Bioinformatics and Biomedical Engineering, Washington 2001 209 216
Bar-Joseph Z Gerber GK Lee TI Computational discovery of gene modules and regulatory networks Nature Biotechnology 2003 21 1337 1342 14555958 10.1038/nbt890
Bateman A Coin L Durbin R The Pfam protein families database Nucleic Acids Research 2004 32 D138 D141 14681378 10.1093/nar/gkh121
Blanchette M Boroque G Sankoff D Breakpoint phylogenies Genome Informatics 1997 25 34 11072302
Bono H Ogata H Goto S Kanehisa M Reconstruction of amino acid biosynthesis pathways from the complete genome sequence Genome Research 1998 8 203 210 9521924
Butcher EC Berg EL Kunkel EJ Systems biology in drug discovery Nature Biotechnology 2004 22 1253 1259 15470465 10.1038/nbt1017
Butte A Kohan I Mutual information relevance networks: functional genomics clustering using pair-wise entropy measurements Proceedings of the Pacific Biocomputing 2000 418 29
Catherine B Emmanuel C Sébastien C Yoann B Sandrine D Daniel K The ProDom database of protein domain families: more emphasis on 3D Nucleic Acids Research 2005 33 D212 D215 15608179
Chenna R Sugawara1 H Koike Lopez TR Gibson TJ Higgins DJ Thompson JD Multiple sequence alignment with the Clustal series of programs Nucleic Acids Research 2003 31 3497 3500 12824352 10.1093/nar/gkg500
Date SV Marcotte EM Discovery of uncharacterized cellular systems by genome-wide analysis of functional linkages Nature Biotechnology 2003 21 1055 1062 12923548 10.1038/nbt861
Delcher AL Harmon D Kasif S White O Salzberg SL Improved microbial gene identification with GLIMMER Nucleic Acids Research 1999 27 4636 41 10556321 10.1093/nar/27.23.4636
Delcher AL Phillipy A Carlton J Salzberg SL Fast algorithms for large scale genome alignment and comparison Nucleic Acids Research 2002 30 2478 83 12034836 10.1093/nar/30.11.2478
Durand D Sankoff D Tests for gene clustering Journal of Computational Biology 2003 10 453 482 12935338 10.1089/10665270360688129
Eskin E Pevzner PA Finding composite regulatory patterns in DNA sequences Bioinformatics 2002 18 354 363
Farre D Identification of patterns in biological sequences at ALGGEN server: PROMO and MALGEN Nucleic Acid Research 2003 31 3651 3653 10.1093/nar/gkg605
Fleischmann RD Whole-genome random sequencing and assembly of Haemophilus influenzae Rd Science 1995 269 496 512 7542800
Gelfand MS Novichkov PS Novichkova ES Mironov AA Comparative analysis of regulatory patterns in bacterial genomes Briefings in Bioinformatics 2000 1 357 371 11465053
Ginalski K Grishin NV Godzik A Rychlewski L Practical lessons from protein structure prediction Nucleic Acids Research 2005 33 1874 1891 15805122 10.1093/nar/gki327
Gomez A Domedel N Cedano J Penol J Querol E Do current sequence analysis algorithms disclose multifunctional (moonlighting) proteins? Bioinformatics 2003 19 895 896 12724303 10.1093/bioinformatics/btg111
Goto S Nishioka T Kanehisa M LIGAND: chemical database for enzyme reactions Bioinformatics 1998 14 591 599 9730924 10.1093/bioinformatics/14.7.591
Halperin I Ma B Wolfson H Nussinov R Principles of docking: An overview of search algorithms and a guide to scoring functions Proteins 2002 47 409 43 12001221 10.1002/prot.10115
Henikoff S Henikoff JG Alford WJ Pietrokovski S Automated construction and graphical presentation of protein blocks from unaligned sequences Gene 1995 163 GC17 26 7590261 10.1016/0378-1119(95)00486-P
Howe K Bateman A Durbin R Quick tree: building huge neighbor-joining trees of protein sequences Bioinformatics 2002 11 1546 1547 12424131 10.1093/bioinformatics/18.11.1546
Jeffrey CJ Moonlighting proteins: old proteins learning new tricks Trend Genet 19 415 417 10.1016/S0168-9525(03)00167-7
Kelley BP Sharan R Karp RM Sittler T Root DE Stockwell BR Ideker T Conserved pathways within bacteria and yeast as revealed by global protein network alignment PNAS 2003 100 11394 11399 14504397 10.1073/pnas.1534710100
Letunic I Copley RR Schmidt S Ciccarelli FD Doerks T Schultz J Ponting CP Bork P SMART 4.0: towards genomic data integration Nucleic Acids Research 2004 32 D142 D144 14681379 10.1093/nar/gkh088
Liu J Dehbi M Moeck G Antimicrobial drug discovery through bacteriophage genomics Nature Biotechnology 2004 22 185 191 14716317 10.1038/nbt932
Lukashin AV Borodovsky M GeneMark.hmm: new solutions for gene finding Nucleic Acids Research 1998 26 1107 15 9461475 10.1093/nar/26.4.1107
Luscombe NM Laskowski RA Thornton JM Amino acid base interactions: a three dimensional analysis of protein-DNA interactions at an atomic level Nucleic Acids Research 2001 29 2860 2874 11433033 10.1093/nar/29.13.2860
McGuire AM Hughes JD Church GM Conservation of DNA regulatory motifs and discovery of new motifs in microbial genomes Genome Research 2000 10 744 757 10854408 10.1101/gr.10.6.744
Mironov AA Koonin EV Roytberg MA Gelfand MS Computer analysis of transcription regulatory patterns in completely sequenced bacterial genomes Nucleic Acids Research 1999 27 2981 2989 10390542 10.1093/nar/27.14.2981
Moret BME Bader DA Warnow T High Performance Algorithm Engineering for Computational Phylogenetics The Journal of Supercomputing 2002 22 99 111 10.1023/A:1014362705613
Mount DW Bioinformatics Sequence and Genome Analysis 2000 Cold Spring Harbor Laboratory Press, Newyork
Mwangi MM Siggia ED Genome wide identification of regulatory motifs in Bacillus subtilis BMC Bioinformatics 2003 4 18, 20 pages
Nakashima N Mitani Y Tamura T Actinomycetes as host cells for production of recombinant proteins Microbial Cell Factories 2005 4 1 5 15629064
Ogata H Goto S Fujibuchi W Kanehisa M Computation with the KEGG pathway database Biosystems 47 119 28 9715755 10.1016/S0303-2647(98)00017-3
Papin JA Price ND Palsson BØ Extreme pathway lengths and reaction participation in genome-scale metabolic networks Genome Research 2002 12 1889 900 12466293 10.1101/gr.327702
Patterson SD Data analysis – the Achilles heel of proteomics Nature Biotechnology 2003 21 221 222 12610558 10.1038/nbt0303-221
Pawlowski K Rychlewski L Zhang B Godzik A Fold predictions for bacterial genomes Journal of Structural Biology 2001 134 219 31 11551181 10.1006/jsbi.2001.4394
Pearson WR Lipman DJ Improved tools for biological sequence comparison Proceedings National Academy of Science U S A 1988 85 2444 2448
Pennisi E Genome data shake tree of life Science 280 672 674 9599142 10.1126/science.280.5364.672
Qin ZS McCue LA Thompson W Identification of co-regulated genes through Bayesian Clustering of predicted regulatory binding sites Nature Biotechnology 2003 21 435 439 12627170 10.1038/nbt802
Robinson WH Fontoura P Lee BJ Protein microarrays guide tolerizing DNA vaccine treatment of autoimmune encephalomyelitis Nature Biotechnology 2003 21 1033 1039 12910246 10.1038/nbt859
Rost B Sander C Prediction of protein secondary structure at better than 70 percent accuracy Journal of Molecular Biology 1995 232 584 599 8345525 10.1006/jmbi.1993.1413
Schilling CH Covert MW Famili I Church GM Edwards JS Palsson BØ Genome-scale metabolic model of Helicobacter pylori 26695 Journal of Bacteriology 2002 184 4582 4593 12142428 10.1128/JB.184.16.4582-4593.2002
Schuster S Dandekar T Fell DA Detection of elementary flux modes in biochemical networks: a promising tool for pathway analysis and metabolic engineering Trends Biotechnology 1999 17 53 60 10.1016/S0167-7799(98)01290-6
Scott J Ideker T Karp RM Sharan R Efficient algorithms for detecting signaling pathways in protein interaction networks Proceedings of RECOMB 2005 1 13
Sidhu SS Bader GD Boone C Functional genomics of intracellular peptide domains with combinatorial biology methods Current opinion in Chemical Biology 2003 7 97 102 12547433 10.1016/S1367-5931(02)00011-X
Sigrist CJ Cerutti L Hulo N Gattiker A Falquet L Pagni M Bairoch A Bucher P PROSITE: a documented database using patterns and profiles as motif descriptors Briefings in Bioinformatics 2002 3 265 74 12230035
Sonnhammer EL Eddy SR Birney E Batman A Durbin R Pfam: multiple sequence alignments and HMM-profiles of protein domains Nucleic Acids Research 1998 26 320 322 9399864 10.1093/nar/26.1.320
Suzek BE Ermolaeva MD Schreiber M Salzberg SL A probabilistic method for identifying start codons in bacterial genomes Bioinformatics 2001 17 1123 30 11751220 10.1093/bioinformatics/17.12.1123
Tatusov RL Mushegian M Bork P Brown N Hayes WS Borodovsky M Rudd KE Koonin EV Metabolism and evolution of Haemophilius Influenzae deduced from a whole-genome comparison with Escherichia Coli Current Biology 1996 6 279 291 8805245 10.1016/S0960-9822(02)00478-5
Thomas DJ Casari G Sander C The prediction of protein contacts from multiple sequence alignments Protein Engineering 1996 9 941 948 8961347
Unger R Moult J Genetic algorithms for protein folding simulations Journal of Molecular Biology 1993 231 75 81 8496967 10.1006/jmbi.1993.1258
Venter JC The sequence of the human genome Science 2001 291 1304 1351 11181995 10.1126/science.1058040
Waddell PJ Kishino H Cluster inference methods and graphical models evaluated on NCI60 microarray gene expression data Genome Informatics 2000 11 129 140 11700594
Waterman MS Introduction to Computational Biology: Maps, Sequence, and Genomes 1995 Chapman & Hall, London
Whisstock JC Lesk AM Prediction of protein function from protein sequence and structure Q Rev Biophysics 2003 36 307 340 10.1017/S0033583503003901
Woese C The universal ancestor PNAS 1998 95 6854 6859 9618502 10.1073/pnas.95.12.6854
Zagursky RJ Olmsted SB Russell DP Wooters JL Bioinformatics: how it is being used to identify bacterial vaccine candidates Vaccines 2003 2 417 436 10.1586/14760584.2.3.417
|
15985162
|
PMC1182391
|
CC BY
|
2021-01-04 16:24:35
|
no
|
Microb Cell Fact. 2005 Jun 28; 4:19
|
utf-8
|
Microb Cell Fact
| 2,005 |
10.1186/1475-2859-4-19
|
oa_comm
|
==== Front
J NanobiotechnologyJournal of Nanobiotechnology1477-3155BioMed Central London 1477-3155-3-51598518410.1186/1477-3155-3-5ResearchChemically synthesized zinc finger molecules as nano-addressable probes for double-stranded DNAs von Nickisch-Rosenegk Markus [email protected] Eva [email protected] Rothin [email protected] Alexander [email protected] Frank F [email protected] Fraunhofer Institute of Biomedical Engineering (IBMT), Dpt. Molecular Bioanalytics and Bioelectronics, Arthur-Scheunert-Allee 114-116, 14558 Nuthetal, Germany2005 29 6 2005 3 5 5 14 1 2005 29 6 2005 Copyright © 2005 von Nickisch-Rosenegk et al; licensee BioMed Central Ltd.2005von Nickisch-Rosenegk 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 experiments describe an alternative method of dsDNA recognition using zinc finger (ZF) molecules which bind DNA specifically and with high affinity. Our aim was to develop zinc finger probes which are able to bind to dsDNA molecules at predetermined sites. In our basic approach we used pairs of complementary oligonucleotides to form dsDNAs, containing one of the three SP1-transcription factor motifs as a zinc finger recognition site. Two zinc finger probes of the SP1 motif were chemically synthesized and modified with a Dy-633 fluorophore. The SP1 peptides were folded into functional zinc fingers using zinc chloride. The addressable dsDNAs were immobilized on optical fibres, and the kinetics and binding rates of the artificial zinc finger probes were measured by a fluorescence detecting device (photomultiplying tube). The two zinc fingers and their corresponding DNA recognition sites served as specific probes and controls for the matching site and vice versa. Our experiments showed that a variety of dsDNA-binding probes may be created by modification of the amino acid sequence of natural zinc finger proteins. Our findings offer an alternative approach of addressing dsDNA molecules, for example for use in a nanoarray device.
==== Body
Background
Nanostructures based on DNA have been reported previously. Generally it is a prerequisite to generate single stranded DNA (ssDNA) molecules to open the possibility of probing with oligonucleotides. However, the production of long ssDNA may not be an easy task. ssDNAs are more sensitive to physical handling, tend to form stable secondary structures and require considerable effort to obtain sufficient quantities intact molecules. An alternative approach relies on the use of peptide-nucleic acids (PNAs), which are capable of recognizing dsDNAs. The recent reports have demonstrated that this approach is not always practical and PNA binding is not always reliable [1]. Another potential approach is to use existing cellular mechanisms involved in the control of the production of proteins on the level of nucleic acids. Transcription factors which bind to dsDNA target sequences are switching specific genes "on" or "off" by recognizing and binding to short specific DNA fragments. Zinc finger proteins are transcription factors which have concise and simple structure [2], can be made synthetically and might therefore provide alternative probes for dsDNA recognition [3].
Results and discussion
In our experiments we used one of the three zinc finger motifs of the SP1-zinc-finger as a template for our chemically synthesized zinc finger probes, which were of the C2H2 type (Figure 1). The DNA-binding region was altered in two places to enable the folded zinc finger to bind either at the "AAA"-DNA sequence site or at the "GCG"-DNA sequence site (see M & M, bold and underlined). The custom-made peptides were modified with a fluorophore (Dye633) at their amino-termini to allow quantitative detection. The double-stranded DNA targets were made of custom-made complementary oligonucleotides carrying the appropriate DNA recognition-sequence and a phosphate modification at the 5'-end of one of the complementary oligonucleotides for their immobilization.
Figure 1 Scheme of a C2H2 zinc finger motif. X denotes any amino acid; Z – hydrophobic amino acids [L, I, V, M, F, Y, W, C]; L is Leucine; C is Cysteine; H is Histidine. The binding interface of the α-Helix is formed by amino acids No -1, 1, 2, 3, L, 5 and 6.
Detection was done by fluorescence on a fibre-optical-device developed in our institute (Figure 2), which was driven by the in-house built software. The double-stranded DNA targets were immobilized on a silanized glass fibre and incubated with the zinc fingers in a flow-through chamber (Figure 2). The binding and dissociation events were measured in real time and binding curves (binding signal vs. time) were obtained (Figures 3 and 4). The curves were fitted using in-house built fitting routine. The resulting dissociation rate (kD) in a millimolar range were obtained, indicating high rates of clearance of the zinc finger molecules (see Figures 3 and 4). However, these were sufficient for our experiments. We have measured for the first time the specific binding of synthetic single-motif zinc finger peptides to their specific binding sites in real time in vitro.
Figure 2 Scheme of the used fibre optical device. A) complete work flow. B) Reaction and detection chamber (cross section) (derived from Schwonbeck 2004)
Figure 3 Real-time detection of the binding kinetics to the "GCG"-sequence immobilised on a glass fibre. Black curve corresponds to the interaction between matching pair of ZF1 and "GCG" motif (kD: 4,9·10-3 /s), the red curve corresponds to the ZF2 (kD: 7,3·10-3 /s) which should recognise the "AAA" sequence, but not the "GCG".
Figure 4 Real-time detection of the binding kinetics to the "AAA"-sequence immobilised on a glass fibre. Black curve corresponds to the interaction between matching pair of ZF2 and "AAA" motif (kD: 3,5·10-3 /s), the red curve corresponds to the ZF1 (kD: 4,6·10-3 /s) which should recognise the "GCG" sequence, but not the "AAA".
Figure 3 shows a typical result obtained using the fibre covered with DNA containing the "GCG" motif (the motif recognised by ZF1). The black curve represents ZF1 binding, which is characterised by fast association kinetics. A small amount of ZF1 remained associated with the DNA for over 1.5 hours into the dissociation stage. A subsequent incubation of the same fibre with ZF2 (Figure 3, red curve) showed faster dissociation kinetics. Similarly, the ZF2 motif bound the "AAA" sequence strongly and significant amount of ZF2 remained bound to the DNA irreversibly (Figure 4, black curve), whereas ZF1 dissociated completely (Figure 4, red curve).
Conclusion
In this work we have shown that synthetic zinc fingers bind specifically to their binding sites which consist of nucleic acid triplets. An additional advantage of the SP1 system is that it offers an opportunity to use multiple single-motif zinc fingers [4,5], similar to the native SP1 transcription factor motifs. Using multiple ZF domains should result in binding affinities in a nanomolar range and should also allow for more specific sequence recognition. We envisage that nano-addressable zinc finger motif-based probes for dsDNAs will employ multiple linked ZF motifs and possess binding affinities in the nanomolar range.
Materials and methods
Nucleic acids
Complementary oligonucleotides (Carl Roth, Karlsruhe) with specific DNA-zinc-finger-recognition sites were:
GCG-motif
5' -Phosphat- TgACTgACTgACTgACTgACgCgTgACTgACTgAC -3'(sense)
5' -gTCAgTCAgTCACgCgTCAgTCAgTCAgTCAgTCA-3'(antisense)
AAA-motif
5' -Phosphat-TgACTgACTgACTgACTgACAAATgACTgACTgAC-3'(sense)
5' – gTCAgTCAgTCATTTgTCAgTCAgTCAgTCAgTCA -3'(antisense)
Corresponding pairs of the sense and antisense oligonucleotides were annealed to yield dsDNA fragments, which were used for immobilisation (see below)
Synthetic peptides (zinc fingers in "one-letter-code")
The custom-made peptides were modified with a fluorophore (Dye633, Dyomics) at the amino-terminus for detection. The peptides were ZF1 (GCG-motif): N-Dy633-GFMCTWSYCGKRFTRSDELQRHKRTH-COOH (Biosynthan, Berlin) and ZF2 (AAA-motif):
N-Dy633-GFMCTWSYCGKRFTQSSNLQTHKRTH-COOH (peptides & elephants, Nuthetal).
Storage buffer contained 10 mM Tris/HCl (pH 7,5), 90 mM KCl, 1 mM MgCl2, 90 μM ZnCl2, 5 mM DTT; running buffer was made by adding acetylated BSA(Aurion, Wageningen NL) to the storage buffer to a final concentration of 1–2%. The peptides were dissolved in the running buffer (at the concentration of 0.5 μg peptide / 500 μl buffer) and incubated for 15 min to allow peptides to fold to form zinc fingers prior to incubation with DNA (Association stage). To measure the dissociation kinetics, the buffer was changed to the "running" buffer (without ZF).
Detection
Detection was using "in-house" built optical fibre device "Horst" (see Figure 2), driven by in-house software.
Silanized glass fibres
Glass fibres were cleaned with freshly prepared hot "Piranha"-solution (1:2 mixture of 30% hydrogen peroxide and 97% sulphuric acid) for 1 minute and thoroughly rinsed with distilled water. The fibres were further incubated in freshly prepared hot 10N NaOH for 2 h. The activated fibres were immersed in 10% aminopropyl-triethoxysilane (APTES) solution, pH adjusted to 3.5 and heated to 80°C in a water bath for 2 h. Finally the fibres were rinsed with HPLC grade water again and baked at 110°C for 1 h. The amino-modified surface is stable for several months.
Immobilization of DNA
A solution of 5'-phosphorylated oligonucleotide was diluted with 30 mM 1-methylimidazole to a final concentration of 5 μM. Then 10 mg/ml EDC was added to each oligonucleotide and mixed for 1 minute. The oligonucleotide solutions were used immediately to cover the activated parts of the glass fibres (see above).
Acknowledgements
The authors are grateful to Berit Umann and Alexandra Grychta for their excellent technical assistance. This work was supported by the German ministry of education and research (BMBF: Az. 0311842 and 0311842A) and the European Community (EU) within the 5th frame program (FP5, Nanocell: QLK3-CT-2001-00278)
==== Refs
Nielsen Peter Gene targeting using peptide nucleic acid Methods Mol Biol 2005 288 343 58 15333914
Sugiura Yukio Natural and artificial zinc finger proteins Riken Review 2001 35
Wolfe Scot A Nekludova Lena Pabo Carl O DNA Recognition by Cys2His2 Zinc Finger Proteins Biophys Biomol Struc 1999 3 183 212
Wu Herren Yang Wei-Ping Barras III Carlos F Building Zinc Fingers By Selection: Toward A Therapeutic Application Proc Natl Acad Sci 1995 92 344 348 7831288
Kwang-Hee Bae Young Do Kwon Hyun-Chul Shin Human zinc fingers as building blocks in the construction of artificial transcription factors Nat Biotechnol 2003 2 275 280
|
15985184
|
PMC1182392
|
CC BY
|
2021-01-04 16:38:05
|
no
|
J Nanobiotechnology. 2005 Jun 29; 3:5
|
utf-8
|
J Nanobiotechnology
| 2,005 |
10.1186/1477-3155-3-5
|
oa_comm
|
==== Front
Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central London 1743-7075-2-141595817110.1186/1743-7075-2-14ReviewContribution of anaerobic energy expenditure to whole body thermogenesis Scott Christopher B [email protected] Department of Sports Medicine, University of Southern Maine, 37 College Avenue, Gorham, ME 04038, USA2005 15 6 2005 2 14 14 24 1 2005 15 6 2005 Copyright © 2005 Scott; licensee BioMed Central Ltd.2005Scott; 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.
Heat production serves as the standard measurement for the determination of energy expenditure and efficiency in animals. Estimations of metabolic heat production have traditionally focused on gas exchange (oxygen uptake and carbon dioxide production) although direct heat measurements may include an anaerobic component particularly when carbohydrate is oxidized. Stoichiometric interpretations of the ratio of carbon dioxide production to oxygen uptake suggest that both anaerobic and aerobic heat production and, by inference, all energy expenditure – can be accounted for with a measurement of oxygen uptake as 21.1 kJ per liter of oxygen. This manuscript incorporates contemporary bioenergetic interpretations of anaerobic and aerobic ATP turnover to promote the independence of these disparate types of metabolic energy transfer: each has different reactants and products, uses dissimilar enzymes, involves different types of biochemical reactions, takes place in separate cellular compartments, exploits different types of gradients and ultimately each operates with distinct efficiency. The 21.1 kJ per liter of oxygen for carbohydrate oxidation includes a small anaerobic heat component as part of anaerobic energy transfer. Faster rates of ATP turnover that exceed mitochondrial respiration and that are supported by rapid glycolytic phosphorylation with lactate production result in heat production that is independent of oxygen uptake. Simultaneous direct and indirect calorimetry has revealed that this anaerobic heat does not disappear when lactate is later oxidized and so oxygen uptake does not adequately measure anaerobic efficiency or energy expenditure (as was suggested by the "oxygen debt" hypothesis). An estimate of anaerobic energy transfer supplements the measurement of oxygen uptake and may improve the interpretation of whole-body energy expenditure.
==== Body
Background
"...(animals) take up oxygen and complex compounds made by plants, discharge these compounds largely in the form of carbonic acid (CO2)and water as the products of combustion and partly as simpler reduced products, thus consuming a certain quantity of chemical potential energy, and generate thereby heat and mechanical energy" (H.L.F. Helmholtz, 1821-1894)
Measurements of heat loss and oxygen uptake are the two major methods for determining energy expenditure although they do not always provide equivalent results at equivalent time points [1-4]. The focus on oxygen uptake follows from the extensive involvement of mitochondria in ATP re-synthesis accompanied by concomitant heat production [5-8]. Sites of ATP hydrolysis (e.g. contracting muscle) represent another source of energy transfer and heat exchange. Non-steady state periods of rapid growth and development, disease, arousal from torpor, heavy/severe exercise and hypoxia, however, offer proof of how tenuous the relationship between heat loss and oxygen uptake can be [1,3,4,9-11]. In isolated mammalian cells, for example, the accelerated production of lactate has been shown to make a substantial contribution to heat production beyond mitochondrial (aerobic) involvement [12]. If heat serves as the standard measure of energy expenditure then anaerobic energy transfer, specifically rapid glycolysis and glycogenolysis with lactate production (i.e., rapid anaerobic ATP re-synthesis) has the potential to make significant contributions to cellular energy expenditure.
Glycolysis as a form of fermentation has been a part of life for an estimated three billion years [13]. It has been observed that anaerobic glycolysis and oxygen uptake often behave in a reciprocal manner. Pasteur, for example, demonstrated that glucose utilization in yeast was more rapid when oxygen was absent [14]. It was subsequently hypothesized that alterations in aerobic respiration influence glycolytic rate. Crabtree [15] described the suppression of oxygen uptake when an abundance of glucose was provided to tumor cells. More recently it has been shown that this "'Crabtree Effect" is not the result of altered respiratory function, but rather an induction of the glycolytic enzymes during cellular proliferation as lactate dehydrogenase (LDH) increased 10-fold and appeared to influence the subsequent routing of NAD+ to the cytoplasm and away from immediate mitochondrial respiration [16]. As it pertains to cellular metabolism then, a distinct trade-off between anaerobic and aerobic metabolic pathways can be seen; high rates of mitochondrial ATP re-synthesis have the potential to suppress anaerobic glycolysis and, conversely, rapid glycolytic ATP re-synthesis can suppress aerobic respiration. In an experiment with yeast, the relative contributions of anaerobic and aerobic processes to total ATP re-synthesis were genetically modified by increasing the glycolytic enzyme, phosphofructokinase (PFK). This modification resulted in yeast with enhanced anaerobic ATP re-synthesis – accompanied by a 36% lower oxygen uptake – but unchanged total ATP turnover compared to normal aerobic-respiring yeast [17]. It appears then, at least in single cell-types, anaerobic ATP re-synthesis has the potential to promote a discrepancy between energy expenditure (heat loss) and oxygen uptake. The question that remains is whether similar discrepancies are seen at the level of the whole-animal.
This review contains four sections. The first briefly describes thermodynamic and bioenergetics interpretations of energy transfer. The second section describes the traditional (stoichiometry and gas exchange) and contemporary (bioenergetic) interpretations behind metabolic heat production. The third section describes energy transfer as lactate production and lactate removal. In the fourth section examples are provided that suggest how an estimate of anaerobic energy transfer along with a separate measure of oxygen uptake may better influence the interpretation of whole-body efficiency and energy expenditure.
Energy transfer
The first law of thermodynamics states that energy can not be created or destroyed but can and does change form. The second law describes how energy is transferred from one form to another. For example heat, as an expression of energy, always flows in one direction – from hot to cold. Other ways of stating this are that energy flows "downhill" or, from a state of lower entropy to one of higher entropy. Entropy represents energy that is not available to perform work so that simply put, energy transfer is inefficient. Inefficiency also appears in the form of heat production that is usually discarded into the environment. In the late 1800's Josiah Gibbs acknowledged the importance of entropy and enthalpy in his explanations of chemical energy transfer. The Gibbs free energy is recognized as energy that is available to perform work at constant temperature and pressure and is the usual thermodynamic parameter for identifying spontaneity of chemical reactions. Thermodynamics was further developed in the context of a closed system where heat but not matter was exchanged with the environment (e.g. test tube reactions). This description was also applied to living cells. It is of interest that A.V. Hill shared the 1922 Nobel Prize in part for his recognition that muscle cells were not heat-to-mechanical motion converters as modeled by the steam engine, but could rather be understood as chemo-mechanical converters.
Biological energy transfer or bioenergetics is accurately described in the context of an open system where matter and energy are continuously exchanged between a cell and its immediate environment [18-20]. In open or in closed systems the Gibbs free energy 'drives' biochemical and chemical reactions, respectively. Closed systems have specific starting and ending points for the Gibbs free energy change during energy transfer. In an open system however, the Gibbs free energy availability may change as the rate of energy transfer and the ratio of product to reactant varies during the exchange (e.g., as the distance from equilibrium is altered) [2]. Within cells, heat and entropy production are the continuous result of energy transfer during ATP hydrolysis and re-synthesis, collectively known as ATP turnover. ATP undergoes hydrolysis to "fuel" a variety of cellular functions such as muscle contraction, the sodium-potassium pump within cell membranes and coupling to endergonic reactions. Aerobic and anaerobic metabolisms serve to re-synthesize ATP. Entropy production can not be directly measured. Heat loss can be quantified with direct calorimetry as a measure of energy expenditure (transient heat storage is not described here so that heat loss is equated with heat production). Heat production also can be estimated with measures of oxygen uptake and carbon dioxide as indirect calorimetry.
Gas exchange and energy expenditure
Lavoisier first described both biological respiration and combustion in terms of their equivalence of gas exchange and heat production. At the end of the nineteenth century experiments by Eduard Pfluger and others compared direct measurements of heat production with indirect measures of gas exchange. Pfluger utilized a stoichiometric analysis to uncover the relationship between the chemical compositions of different foodstuff and their oxidation. From this data, for example, glucose oxidation is described as,
C6H12O6 + 6 O2 → 6 CO2 + 6 H2O
The stoichiometric ratio of CO2: O2 as measured from the mouth became known as the respiratory exchange ratio (RER) and serves as a valuable means of interpreting substrate utilization and heat production. When an all-carbohydrate diet is being oxidized the RER is 1.00 (6 CO2: 6 O2) and heat production is estimated at 21.1 kJ per liter of oxygen (1 l O2 = 21.1 kJ). When fatty acids are the principle substrate oxidized, the RER is 0.70 (palm oil oxidation = 16 CO2: 23 O2) and one liter of oxygen uptake estimates heat production at 19.6 kilojoules (1 l O2 = 19.6 kJ).
The higher RER for carbohydrate oxidation has been interpreted to mean that fat oxidation requires more oxygen and results in less heat production than carbohydrate oxidation [21]; this does not signify that carbohydrate is the more efficient fuel source. Only 2-carbon intermediates (acetyl CoA) can enter into the Krebs cycle for complete aerobic oxidation and the product of anaerobic carbohydrate breakdown – pyruvate – must undergo de-carboxylation (i.e., carbon dioxide production) by the enzyme pyruvate dehydrogenase (PDH) before it can be oxidized aerobically. In comparison, fat is broken down by mitochondrial beta-oxidation enzymes into 2-carbon intermediates; no de-carboxylation takes place prior to entrance into the Krebs cycle. Per volume of ATP re-synthesized aerobically then, the complete oxidation of glucose and glycogen has additional relative carbon dioxide production, not less relative oxygen uptake, as compared to fat oxidation. The conversion of one liter of carbon dioxide into an estimate of heat production for glucose and fat oxidation reveals larger discrepancy in energy expenditure at 21.1 and 27.6 kJ, respectively [22]. If oxygen uptake better represents energy expenditure than carbon dioxide production, then it must be concluded that the ratio of CO2: O2 provides a poor explanation of energy transfer efficiency.
Heat measurements that are independent of carbon dioxide production reveal a strong linear relationship between oxygen uptake and the enthalpy of combustion of many organic compounds [22-24]. The calorimetric to respiratory (CR) ratio is similar for both combustion and respiration at -460 kJ. mol O2-1 (± 5%) because enthalpy production per electron equivalent approximates -115 kJ·mol O2-1 regardless of the carbon source (a carbon atom has four valences so that four electrons represent -460 kJ·mol O2-1 ± 5%). In this regard, differences in heat production per unit of oxygen among fat and carbohydrate oxidation are better interpreted by bioenergetic explanations of energy transfer as opposed to gas exchange stoichiometry.
Of the ~36 total ATP re-synthesized by complete glucose oxidation, 2 come from glycolysis (~6% of the total) and 34 come from mitochondrial respiration (~94% of the total). The slight 1.5 kJ increase in heat production per oxygen equivalent when carbohydrate is oxidized compared to fat (at 21.1 kJ vs.19.6 kJ) may be better attributed to the small but requisite energy transfer production of heat and entropy during anaerobic substrate level phosphorylation [25]. The anaerobic 1.5 kJ increase represents ~7% of the total heat production of complete glucose oxidation and is similar to the ~6% anaerobic ATP re-synthesis (2 of 36 ATP); like all energy transfer, glycolytic ATP re-synthesis (phosphorylation) is inefficient.
Bioenergetics and energy expenditure
Glycolytic phosphorylation and mitochondrial respiration represent separate and distinct acts of energy transfer. Glycolysis and glycogenolysis take place in the cytoplasm of cells, within and around the contractile apparatus of muscles for example. Glycolysis and glycogenolysis require multiple enzymes that catalyze proton and electron transfer. Moreover, glycolytic phosphorylation takes place where the useful energy within glucose and glycogen is converted to ATP. These reactions can be summarized as a series of phosphate transfers, phosphate shifts, isomerizations, dehydrations and aldol cleavages [26]. The inefficiency of glycolytic substrate level ATP re-synthesis is a result of heat and entropy production.
In comparison, the mitochondria are distinct double-membrane cellular organelles; these membranes create an effective compartment that is separated from the cellular cytoplasm. Within these membranes are a collection of further enzymes that continue to strip protons and electrons from substrate. Protons and electrons are subsequently delivered by carriers (e.g., NAD+) to the electron transport chain (ETC). Energy transfer in the aerobic re-synthesis of ATP is not directly related to enzymatic glycolytic phosphorylation. Instead, reduction of reduced carriers by oxygen is used to create a gradient of protons. Using the inner membrane as a barrier, protons are pumped to one side; the subsequent gradient of protons creates an uphill-downhill energy transfer scenario whereby specific membrane portals known as mitochondrial ATPases allow protons to pass through. The energy of this downhill flow is exploited to re-synthesize ATP [26]. Mitochondrial heat production has been traced largely to the flow of protons down this gradient [6].
Contemporary bioenergetic interpretations of anaerobic and aerobic metabolism recognize the energy transfer independence of anaerobic and aerobic ATP re-synthesis; each has different reactants and products, uses dissimilar enzymes, involves different types of biochemical reactions, takes place in separate cellular compartments, exploits different types of gradients and, ultimately, each operates with different efficiency [27]. Thus, the heat and entropy production of anaerobic metabolic energy transfer can not possibly be represented by mitochondrial respiration (or vice-versa for that matter). Dissimilar energy transfer formats and operational efficiency must both be kept soundly in mind when interpreting energy expenditure. Nonetheless, glycolytic phosphorylation can proceed aerobically whereby pyruvate is immediately and directly routed for mitochondrial respiration (within the Krebs cycle). When the rate of glycolytic phosphorylation (with 2 ATP; 1.5 kJ per l O2) matches the rate of mitochondrial respiration (with 34 ATP; 19.6 kJ per l O2) then the anaerobic and aerobic components of glucose and glycogen oxidation can be added together to interpret the collective ATP turnover with the energy expenditure conversion, 21.1 kJ per liter of O2 (~36 ATP).
Lactate production
Anaerobic glycolysis and glycogenolysis can proceed by the rapid reduction of pyruvate to form lactate (i.e., exceeding mitochondrial respiratory rates and regardless of oxygen availability). In an open system the rate of energy transfer and alterations in the product to reactant ratio can promote greater inefficiency [2,28]. When rapid glycolytic ATP re-synthesis exceeds mitochondrial rates, lactate and heat production ensues and a measure of oxygen uptake no longer accurately reflects the rate or the amount of ATP re-synthesis that takes place. Recall that the calorimetric to respiratory (CR) ratio during respiration is -460 kJ·mol O2-1 (± 5%). In cultured mammalian cells however, the ratio of heat production to oxygen uptake was found to vary from -490 to -800 kJ·mol O2-1 or more [12]. Gnaiger and Kemp found that the -30 kJ to -340 kJ·mol O2-1 increase was best related to the increase in lactate formation and presumably an increase in the anaerobic energy expenditure contribution to total ATP re-synthesis [12,17].
Lactate production in resting fully oxygenated cells is readily apparent [12,16,29,30]. In addition to providing ATP, rapid glycolytic phosphorylation has been suggested to maintain the redox potential within mammalian cells [31], to protect cells against oxidative stress [32], to promote the formation of biosynthetic precursors in growing cells [33] and as a mechanism of control in cellular growth [34]. Whatever its role, rapid glycolytic ATP re-synthesis with lactate production is associated with heat and entropy production and by definition inefficiency and energy expenditure. It appears that the most important step for heat production during rapid rates of glycolysis and glycogenolysis is the reduction of pyruvate to lactate at -63 to -80 kJ per mol of lactate (dependent on the immediate internal and external environments) [12,35]. This energy expenditure is irreversible.
Lactate removal
Removal of lactate involves conversion back to pyruvate. Pyruvate, in turn, can be converted into a variety of compounds that may include glucose within the liver (Cori Cycle), glycogen within cells (gluconeogenesis) or alanine (an amino acid). It is presumed that the ATP turnover that is required for these conversions comes from mitochondrial energy transfer (as 19.6 kJ per l O2) [22].
Lactate can also be removed via the complete aerobic oxidation of pyruvate [36]. The application of energy conservation as expressed in Hess's law (reactions that start and end with the same reactants and products produce the same amount of enthalpy regardless of path) led to the idea that anaerobic energy expenditure during exercise could be measured via subsequent oxygen uptake during the recovery from exercise, as part of the so-called "oxygen debt" [37]. This hypothesis proposes that all ATP re-synthesized via glycolytic phosphorylation is included in the net aerobic ATP yield when pyruvate undergoes subsequent aerobic oxidation (36 ATP; 21.1 kJ per l O2), even if it passes transiently through lactate. Gaesser and Brooks argued that the many fates of lactate and pyruvate removal in addition to complete aerobic oxidation indicate that the oxygen debt does not adequately represent anaerobic glycolytic energy expenditure [38]. Moreover, both aerobic and anaerobic biochemical reactions are often held far-from-equilibrium as part of an open system and this occurs at an irreversible expense [2,18,19].
Strict application of Hess's law to the in vitro exothermic reaction of pyruvate to lactate requires that the reverse reaction should consume an equivalent amount of heat. While this is true within closed systems it should not be the case within an in vivo open system. It is in the heat loss (calorimetric) to oxygen uptake (respiratory) ratio (kJ·mol O2-1) that this is most clearly revealed. We found that in cell preparations and cardiac muscle fibers that respire on externally supplied pyruvate or lactate, there is equivalent heat production when expressed per mol of oxygen uptake [20]. That is, heat is not consumed when lactate is converted back to pyruvate; the reaction is not thermodynamically reversible, energy transfer during mitochondrial respiration does not represent energy transfer in the form of rapid or accelerated anaerobic glycolytic ATP re-synthesis with lactate formation. It is therefore ironic that for most of the 20th century muscle cells were known to be chemo-mechanical converters as part of an open system yet energy transfer, as described by the oxygen debt hypothesis, continued to be explained from a traditional thermodynamic closed system standpoint.
Application and interpretations
Indirect calorimetry is a much simpler procedure than direct calorimetry accounting for its continued popularity in estimating biological heat production. When anaerobic energy expenditure contributions are large, however, whole-body energy expenditure may be significantly underestimated (figure 1). It is unfortunate that no valid measure of anaerobic heat production is available – this appears to be another reason for the hesitation to include an anaerobic component as separate from an oxygen-only interpretation of energy expenditure. The problem lies in the inherent difficulties of the collection of anaerobic metabolites from within active cells. Moreover, there are stores of ATP and phosphocreatine (PC) contained within muscle tissue that are utilized during heavy to severe exercise as anaerobic energy transfer but that are re-synthesized aerobically during the recovery from exercise as excess post-exercise oxygen consumption (EPOC). Thus one part of this ATP/ PC turnover is anaerobic, the other is aerobic [38-40]. In fact, heat measurements taken during brief intense exercise have revealed anaerobic metabolism to be more efficient than aerobic metabolism [27]. Such a finding must be considered carefully however as the heat loss during the oxygen deficit portion of exercise contains separate proportions of rapid glycolytic phosphorylation (that represents full ATP turnover) and stored ATP/ PC usage (but not ATP/PC re-synthesis) [40].
Figure 1 In the top figure, oxygen deficit (pink) represents the anaerobic energy expenditure component to exercise: rapid glycolytic ATP re-synthesis and the use of stored ATP/PC. In this bout of long duration, low to moderate intensity, steady state exercise, the rapid glycolytic component does not make a significant contribution to total energy expenditure. The bottom left figure reveals oxygen uptake measurements for brief, non-steady state, heavy to severe exercise (e.g., a single weight lifting exercise or a quick sprint up a steep hill); vertical lines mark the start and finish to the exercise. The question marks indicate that it is not possible to determine the rapid glycolytic ATP re-synthesis from oxygen-only measurements. The bottom right figure includes a (theoretical) estimate of rapid glycolytic ATP re-synthesis (pink area) and reveals a large absolute and relative anaerobic energy expenditure component to total energy expenditure (restoration of ATP/PC stores are represented in the EPOC measurement).
There are non-invasive methodologies that estimate only the anaerobic substrate level phosphorylation component of anaerobic energy expenditure (i.e., glycolysis and glycogenolysis without the ATP/ PC stores). One such estimate suggests that every millimole of blood lactate above resting levels equals an energy expenditure of 3 milliliters of oxygen uptake per kilogram of body weight [41]. For example, a 65 kg woman with a resting blood lactate level of 1.1 mmol engages in a 400 meter sprint to exhaustion. Peak lactate levels for her sprint are 12.1 mmol so that the change in blood lactate is 11.0 mmol, resulting in an anaerobic energy expenditure contribution of ~45 kJ (~11 kcals).
Because blood lactate concentrations provide at best an approximate description of muscle lactate levels and glycolytic ATP re-synthesis, it is clear that more research is needed to obtain a valid estimate of anaerobic energy expenditure (concentrations of lactate within active muscle are almost always higher than blood) [42]. On the other hand, the potential error of not including an estimate of anaerobic energy expenditure can result in further misinterpretation [20,43,44]. How important is it to include an estimate of anaerobic energy expenditure for the interpretation of whole-body thermogenesis? Below are a few examples where anaerobic energy expenditure contributions may be sufficiently large that their inclusion may improve current interpretations of whole-body energy expenditure.
Exercise energy expenditure
It has been concluded from exercise oxygen uptake-only measurements that a one-set circuit weight training regimen consisting of 8 exercises was 15 kcals short of meeting the energy expenditure criteria for a healthy lifestyle in men (i.e., 150-200 kcals per exercise session) [45]. However, these criteria would appear to have been met if an estimate of rapid glycolytic ATP re-synthesis were included with the exercise oxygen uptake measurements. Depending on the size of the exercising muscle mass, my students and I have found blood lactate contributions to a single bout of weight training exercise (i.e., 1 set) to range from 3 to 12 kcalories in men; a minimal contribution of 3 kcal per exercise would result in an increase in energy expenditure of almost 25 kcal for this weight training circuit. The use of both an anaerobic estimate and an aerobic measure of energy expenditure would provide support for regular circuit weight training as an effective method of obtaining a healthy lifestyle in men. The anaerobic energy expenditure component needs to be large to make a significant contribution to total energy expenditure and this is best seen during brief heavy to severe exercise (total energy expenditure includes exercise anaerobic and aerobic energy expenditure and an acute measure of EPOC) (Figure 1).
The effect of anaerobic energy expenditure on total energy expenditure can be seen in the observation that exercise duration and intensity in reptiles and humans have been shown to affect EPOC size [46-48]. It may be inferred that anaerobic and aerobic energy expenditure interact to promote a larger EPOC. In sprinting mice however, EPOC has been found to be independent of either exercise duration or intensity [49]. Mice are very aerobic and may have a limited anaerobic energy expenditure contribution to sprinting, explaining why EPOC volumes are limited in sprinting mice. Unfortunately anaerobic energy expenditure was not estimated in the mouse study. It is of interest to speculate whether, if energy transfer as rapid glycolytic ATP re-synthesis had originally been considered separate from oxygen uptake, the concept of oxygen debt would have been recognized as an interaction between aerobic and anaerobic energy expenditure (metabolism) rather than being interpreted as "repayment on a loan."
Exercise economy
Exercise economy is traditionally defined as the oxygen uptake required to perform a bout of work at a given rate (e.g., a specific running or cycling pace). During steady state light to moderate intensity exercise, oxygen uptake remains level and provides a sufficient measure of economy. However oxygen uptake steadily increases as heavy to severe steady-state work continues (with ultimate exhaustion) and this has been termed the "slow oxygen uptake component" [50]. This phenomenon remains, for the most part, unexplained yet it is thought that motor unit recruitment patterns may be altered resulting in "additional energy expenditure" [50]. The term "slow oxygen uptake component" implies an aerobic-only approach because the anaerobic glycolytic component is a well known part of heavy to severe exercise. Bioenergetic interpretations might suggest that "additional energy expenditure" is the result of the further dissipation of Gibbs free energy under cellular conditions where both anaerobic and aerobic energy expenditure contributions are changing [2,28].
Ramp-type stress tests, unlike steady-state exercise, utilize a continually increasing power output until the test is terminated at exhaustion (figure 2). At low to moderate workloads, oxygen uptake and power output are linear for slow and fast ramp testing, but this is not seen at heavy to severe workloads [51,52]. Slow ramps to exhaustion have gradual increases in power output so that the test can be lengthy, lasting many minutes. Toward the end of a slow ramping test, the ratio of oxygen uptake to power output begins to increase so that exercise oxygen uptake appears to contain a "slow oxygen uptake component"; a larger relative aerobic versus anaerobic energy expenditure component is found with slow ramping [52]. On the other hand, fast ramping utilizes rapidly increasing power outputs that promote fatigue quickly, resulting in brief test lengths. Toward the end of a fast ramping test to exhaustion the ratio of oxygen uptake to power output may decrease, the traditional interpretation being that this promotes larger relative anaerobic energy expenditure. An alternative explanation is that the decrease in the rate of oxygen uptake is caused by a faster rate of rapid glycolytic phosphorylation that results in a larger relative anaerobic energy expenditure contribution; that is, a whole-body "Crabtree effect" where a non-linear component to "additional energy expenditure" in the form of anaerobic energy transfer is found [53]. Measures of economy for all types of exercise testing would be improved by an estimate of anaerobic energy expenditure.
Figure 2 Continuously increasing ramp exercise tests to exhaustion. Resting oxygen uptake is seen until the start of exercise (vertical line). At low to moderate work rates the oxygen uptake to Watts ratio is similar and linear for both slow (e.g., 15 Watts·min-1) and fast (e.g., 60 Watts·min-1) ramping tests. As the exercise intensity becomes "heavy to severe", the oxygen uptake to Watts ratio increases for the slow ramp test (top line). The opposite is true for the fast ramp test to exhaustion where the oxygen uptake to Watt ratio decreases (bottom line). Notice that the peak Watts are significantly different but the VO2 maximum for the two tests is similar [51, 52]. Contributions of both anaerobic and aerobic energy transfer may explain these apparently disparate phenomena as described in the text.
Exothermic to endothermic transition
Mammals are avid consumers of oxygen and well known producers of heat. Mammalian cellular membranes have been shown to leak ions at a rate that is several-fold greater than those in reptiles; the result is an obligatory increase in ion pumping to maintain the electro-chemical membrane potential [54]. Stevens [55] has suggested that stimulation of the sodium pump was an important evolutionary development toward endothermy. Brisk activity of the sodium pump necessitates a rapid rate of ATP re-synthesis. If this is true then it is important to recognize that in some cells lactate with presumed heat production is better correlated with sodium and potassium pumping than is oxygen uptake [29]. The removal of lactate as provided by mitochondrial ATP re-synthesis further contributes to heat production (e.g., Cori cycle, gluconeogenesis, aerobic oxidation). Because resting lactate turnover in endotherms is as much as 1,500-fold higher than in a similar sized ectotherm, the potential for extensive anaerobic ATP re-synthesis needs to be considered as part of basal whole-body thermogenesis in mammals [56]. It seems logical to conclude that most mammalian energy expenditure does come from aerobic metabolism but the evolution of a metabolic acceleration with concomitant heat production comes from both anaerobic and aerobic pathways. The relative contributions of each pathway to whole-body thermogenesis are not known.
Arousal from torpor
Tucker [11] has shown that heat production in hibernating mice as estimated by oxygen uptake does not account for all of the temperature increases when mice arouse from their metabolic torpor. It is possible therefore that heat production can be accounted for in full when anaerobic energy expenditure is considered as an addition to oxygen-only measurements. Arousal from torpor often induces intense shivering that promotes rapid glycogen degradation accompanied by lactate production and perhaps, like heavy to severe exercise, additional heat production above oxygen uptake-only estimates [57]. The addition of an anaerobic-heat component to whole-body oxygen uptake would appear beneficial to thermogenesis during arousal. Lactate may later be re-converted back to glycogen, a process that may be fueled by mitochondrial fat oxidation to conserve glycogen stores. Such a "'futile cycle" of lactate turnover that is, rapid glycogenolysis (lactate appearance) coupled to gluconeogenesis (lactate disappearance to form glycogen) would be of importance to an obese hibernator who undergoes multiple arousal periods over the course of a winter and has limited access to carbohydrate but has substantial body fat reserves.
Synopsis
Metabolic energy transfer takes place in part as the oxidation of carbohydrate that includes an anaerobic (glycolysis) and aerobic (mitochondrial) component. Rapid glycolytic ATP re-synthesis with lactate production can exceed mitochondrial rates and under these conditions the efficiency of anaerobic energy transfer can not be interpreted using gas exchange stoichiometry. When rapid glycolytic ATP re-synthesis with concomitant heat production is extensive, the anaerobic contribution to energy expenditure can be significant both in cells and in whole-animals. The interpretation of efficiency and energy expenditure may be improved if a separate estimate of anaerobic ATP turnover is provided along with a measure of oxygen uptake.
==== Refs
Loehr KD Sayyadi P Lamprecht I Thermodynamic aspects of development for Tenebrio Molitor L Experientia 1976 32 1002 1003 954999 10.1007/BF01933936
Welch GR Some problems in the usage of Gibbs free energy inbiochemistry J Theor Biol 1985 114 433 446 4021504
Zotin AI Thermodynamic Processes of Biological Processes 1990 Walter de Gruyter, Berlin
Webb P The physiology of heat regulation Am J Appl Physiol 1995 268 R838 R850
Hulbert AJ Else PL Mammals: an allometric study of metabolism at tissue and mitochondrial level Am J Physiol 1985 248 R415 R421 3985183
Brand MD Couture P Else PL Withers KW Hulbert AJ Evolution of energy metabolism: proton permeability of the inner membrane of liver mitochondria is greater in a mammal than in a reptile Biochem J 1991 275 81 86 1850242
Porter RK Brand MD Cellular oxygen consumption depends on body mass Am J Physiol 1995 269 R226 R228 7631898
Rolfe DFS Brown GC Cellular energy utilization and molecular origin of standard metabolic rate in mammals Physiol Rev 1997 77 731 758 9234964
Carlson LD Hsieh ACL Control of Energy Exchange 1970 New York: MacMillan
Shick MJ Widdows J Gnaiger E Calorimetric studies of behavior, metabolism and energetics of sessile intertidal animals Amer Zool 1988 28 161 181
Tucker V The relation between the torpor cycle and heat exchange in the California Pocket Mouse Perognathus californicus J Cell Comp Physiol 1965 65 405 414 10.1002/jcp.1030650314
Gnaiger E Kemp RB Anaerobic metabolism in aerobic mammalian cells: Information from the ratio of calorimetric heat flux and respirometric oxygen flux Biochim Biophys Acta 1990 1016 328 332 2184896
Dyer BD Obar RA Tracing the History of Eukaryotic Cells 1994 Columbia University Press, New York
Racker E History of the Pasteur effect and its pathobiology Mol Cell Biochem 1974 5 17 23 4279327 10.1007/BF01874168
Crabtree HG Observations on the carbohydrate metabolism of tumours Biochem 1929 23 536 545
Greiner EF Guppy M Brand K Glucose is essential for proliferation and the glycolytic enzyme induction that provokes a transition to glycolytic energy production J Biol Chem 1994 269 31484 31490 7989314
Davies SEC Brindle KM Effects of overexpression of phosphofructokinase on glycolysis in the yeast Saccharomyces cerevisiae Biochem 1992 31 4729 4735 1533788 10.1021/bi00134a028
von Bertalanffy L The theory of open systems in physics and biology Science 1950 111 23 29 15398815
Toussaint O Schneider ED The thermodynamics and evolution of complexity in biological systems Comp Biochem Physiol A 1998 120 3 9
Scott CB Kemp RB Direct and indirect calorimetry of lactate oxidation: implications for whole-body energy expenditure J Sports Sciences 2005 23 15 19 10.1080/02640410410001716760
McArdle WD Katch FI Katch VL Exercise Physiology: Energy, Nutrition and Human 1986 2 Lea & Febiger: Philadelphia
Kleiber M Robert F. Kreiger, Malabar, FL The Fire of Life: An Introduction to Animal Energetics 1987
Huggett C Estimation of rate of heat release by means of oxygen consumption measurements Fire Mater 1980 4 61 65 10.1002/fam.810040202
Thornton WM The relation of oxygen to the heat of combustion of organic compounds Philos Mag Ser 1917 33 196 203
Scott CB Energy expenditure for heavy to severe exercise and recovery J Theor Biol 2000 207 293 297 11034835 10.1006/jtbi.2000.2174
Stryer L Biochemistry 1988 3 W.H. Freemon, New York
Krustrup P Ferguson RA Kjaer M Bangsbo J ATP and heat production in human skeletal muscle during dynamic exercise: higher efficiency of anaerobic than aerobic ATP resynthesis J Physiol 2003 549 255 269 12651917 10.1113/jphysiol.2002.035089
Gnaiger E Hochachka. P. W Efficiency and power strategies under hypoxia. Is low efficiency at high glycolytic ATP production a paradox? Surviving Hypoxia: Mechanisms of Control and Adaptation 1993 CRCPress, Boca Raton,
Connett RJ Gayeski TEJ Honig CR Lactate efflux is unrelated to intracellular PO2 in a working red muscle in situ J Appl Physiol 1986 61 402 408 3745033
Paul RJ Functional compartmentalization of oxidative and glycolytic metabolism in vascular smooth muscle Am J Physiol 1983 244 C399 C409 6846528
Sparks ED The purpose of glycolysis Science 1997 277 459 460 9254411 10.1126/science.277.5325.459b
Brand K Aerobic glycolysis by proliferating cells – protection against oxidative stress atthe expense of energy yield J Bioenerg Biomemb 1997 29 355 364 10.1023/A:1022498714522
Kemp RB Guan Y Heat flux and the calorimetric-respirometric ratio as measures of catabolic heat flux in mammalian cells Thermochim Acta 1997 300 199 211 10.1016/S0040-6031(96)03125-5
Newsholme EA Start C Regulation in Metabolism 1973 John Wiley & Sons, London
Minikami S de Verdier CH Calorimetric study on human erythrocyte glycolysis Eur J Biochem 1976 65 451 460 7451 10.1111/j.1432-1033.1976.tb10360.x
Brooks GA Dubouchard H Brown M Sicurello JP Butz CR Role of mitochondrial lactate dehydrogenase and lactate oxidation in the intracellular lactate shuttle Proc Natl Acad Sci 1999 96 1129 1134 9927705 10.1073/pnas.96.3.1129
Hill AV Long CNH Lupton H Muscular exercise, lactic acid and the supply and utilization of oxygen, Parts I-III Proc Roy Soc B 1924 96 438 475
Gaesser GA Brooks GA Metabolic bases of excess post-exercise oxygen consumption: a review Med Sci Sports Exerc 1984 16 29 43 6369064
Bangsbo J Gollnick PD Graham TE Juel C Kiens B Mizuno M Saltin B Anaerobic energy production and O2 deficit-debt relationships during exhaustive exercise in humans J Physiol 1990 422 539 559 2352192
Pahud P Ravussin E Jequier E Energy expended during oxygen deficit period of submaximal exercise in man J Appl Physiol 1980 48 770 775 7451284
di Prampero PE Ferretti G The energetics of anaerobic muscle metabolism: a reappraisal of older and recent concepts Resp Physiol 1999 118 103 115 10.1016/S0034-5687(99)00083-3
Tesch PA Daniels WL Sharp DS Lactate accumulation in muscle and blood during submaximal exercise Acta Physiol Scand 1982 114 441 446 7136774
Scott CB Interpreting anaerobic and total energy expenditure for brief non-exhaustive exercise and recovery Med Sci Sports Exerc 2004 36 S278 10.1097/00005768-200405001-01336
Scott CB Interpreting anaerobic energy expenditure for anaerobic exercise and recovery: an anaerobic hypothesis J Sports Med Phys Fit 1997 37 18 23
Phillips WT Ziuraitis JR Energy cost of the ACSM single-set resistance training protocol J Strength Cond Res 2003 17 350 355 12741877 10.1519/1533-4287(2003)017<0350:ECOTAS>2.0.CO;2
Gleeson TT Hancock TV Metabolic implications of a run now, pay later strategy in lizards: an analysis of post-exercise oxygen consumption Comp Biochem Physiol A 267 259 267
Gore CJ Withers RT The effect of exercise intensity and duration on the oxygen deficit and excess post-exercise oxygen consumption Eur J Appl Physiol 1990 60 169 174 10.1007/BF00839153
Hagberg JM Mullin JP Nagle FJ Effect of work intensity and duration on recovery O2 J Appl Physiol 1980 48 540 544 7372525
Baker EJ Gleeson TT EPOC and the energetics of brief locomoter activity in Mus domesticus J Exp Zool 1988 280 114 120 9433798 10.1002/(SICI)1097-010X(19980201)280:2<114::AID-JEZ2>3.0.CO;2-R
Gaesser G Poole DC Holloszy JO The slow component of oxygen uptake kinetics in humans Exercise and Sport Science Reviews 1996 24 Williams & Wilkins, Baltimore 35 70
Hansen JE Casaburi R Cooper DM Wasserman K Oxygen uptake as related to work rate increment during cycle ergometer Eur J Appl Physiol 1988 57 140 145 10.1007/BF00640653
Scott CB Bogdanffy GM Aerobic and anaerobic energy expenditure during exhaustive ramp exercise Int J Sports Med 1998 19 277 280 9657369
Scott CB Oxygen deficit and slow oxygen component relationships between intermittent and continuous exercise J Sports Sci 1999 17 951 956 10622355 10.1080/026404199365344
Else PL Hulbert AJ Evolution of mammalian endothermic metabolism: "leaky" membranes as a source of heat Am J Physiol 1987 253 R1 R7 3605374
Stevens ED The evolution of endothermy J Theor Biol 1973 38 597 611 4692354 10.1016/0022-5193(73)90260-9
Guppy M Bradshaw SD Fergusson B Hansen IA Atwood C Metabolism in lizards: low lactate turnover and advantages of heterothermy Am J Physiol 1987 253 R77 R82 3605392
Mokrasch LC Grady HG Grisolia Thermogenic and adaptive mechanisms in hibernation and arousal from hibernation Amer J Physiol 1960 199 945 949 13771765
|
15958171
|
PMC1182393
|
CC BY
|
2021-01-04 16:37:46
|
no
|
Nutr Metab (Lond). 2005 Jun 15; 2:14
|
utf-8
|
Nutr Metab (Lond)
| 2,005 |
10.1186/1743-7075-2-14
|
oa_comm
|
==== Front
Proteome SciProteome Science1477-5956BioMed Central London 1477-5956-3-61594149110.1186/1477-5956-3-6MethodologyMolecular weight assessment of proteins in total proteome profiles using 1D-PAGE and LC/MS/MS Ahmad Q Rushdy [email protected] Dat H [email protected] Mark A [email protected] George M [email protected] Martin A [email protected] Dept. of Genetics and Genomics, Boston University School of Medicine, Boston University, 715 Albany St., E639, Boston MA, 02118, USA2 Dept. of Genetics, Harvard Medical School, 200 Longwood Ave., Boston, MA 02115, USA3 The Lipper Center for Computational Genetics. Harvard Medical School, 200 Longwood Ave., Boston, MA 02115, USA4 Dept. of Biomedical Engineering, Boston University, 44 Cummington St., Boston, MA 02215, USA2005 8 6 2005 3 6 6 17 1 2005 8 6 2005 Copyright © 2005 Ahmad et al; licensee BioMed Central Ltd.2005Ahmad 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 observed molecular weight of a protein on a 1D polyacrylamide gel can provide meaningful insight into its biological function. Differences between a protein's observed molecular weight and that predicted by its full length amino acid sequence can be the result of different types of post-translational events, such as alternative splicing (AS), endoproteolytic processing (EPP), and post-translational modifications (PTMs). The characterization of these events is one of the important goals of total proteome profiling (TPP). LC/MS/MS has emerged as one of the primary tools for TPP, but since this method identifies tryptic fragments of proteins, it has not generally been used for large-scale determination of the molecular weight of intact proteins in complex mixtures.
Results
We have developed a set of computational tools for extracting molecular weight information of intact proteins from total proteome profiles in a high throughput manner using 1D-PAGE and LC/MS/MS. We have applied this technology to the proteome profile of a human lymphoblastoid cell line under standard culture conditions. From a total of 1 × 107 cells, we identified 821 proteins by at least two tryptic peptides. Additionally, these 821 proteins are well-localized on the 1D-SDS gel. 656 proteins (80%) occur in gel slices in which the observed molecular weight of the protein is consistent with its predicted full-length sequence. A total of 165 proteins (20%) are observed to have molecular weights that differ from their predicted full-length sequence. We explore these molecular-weight differences based on existing protein annotation.
Conclusion
We demonstrate that the determination of intact protein molecular weight can be achieved in a high-throughput manner using 1D-PAGE and LC/MS/MS. The ability to determine the molecular weight of intact proteins represents a further step in our ability to characterize gene expression at the protein level. The identification of 165 proteins whose observed molecular weight differs from the molecular weight of the predicted full-length sequence provides another entry point into the high-throughput characterization of protein modification.
==== Body
Background
One of the challenges of the post-genome era is the development of technologies and methodologies for the complete characterization of a cell's proteome [1]. This task includes the determination of all protein identities, their amounts, the complexes that they form, their splice forms, and their post-translational modifications. Significant progress has been made on nearly all of these fronts. For instance, protein identities are determined efficiently using 2D-LC/MS/MS [2], or MudPIT [3], or 2DE coupled with MALDI [4]. For the determination of protein quantities, ICAT [5], SILAC [6], and AQUA [7] have made significant contributions. Protein complexes have been characterized in high-throughput fashion using epitope tagging [8,9]. PTMs, in particular phosphorylation, can be targeted using IMAC [10] and other methods [11-13]. Comparatively, there has been relatively little progress with regards to high-throughput characterization of protein splice- or isoforms.
DNA microarray technology revolutionized the field of mRNA profiling [14]. Although mRNA profiling can lend insight into transcriptional control and RNA degradation, it does not directly address translational control of expression, does not characterize PTMs, nor generally identify alternatively spliced transcripts. It is also insensitive to cleavages or chemical modifications of proteins. Since, existing methods for total proteome profiling can, in principle, address many of these issues, there is now a growing need for new tools that can aid in the characterization of these biological processes.
There have been a number of attempts at combining 1D-SDS PAGE with LC/MS/MS for total proteome profiling [15,16]. And there have also been many efforts in which the observed molecular weight of spots on 2D gels are compared to the predicted molecular weight [17,18]. This approach is straightforward and depends on comparison to an external molecular weight marker. While 2D SDS-PAGE is capable of resolving thousands of protein spots, 1D-SDS PAGE offers a number of attractive features, including excellent mass resolution, superior protein solubilization, can accommodate large amounts of protein, and has good run-to-run reproducibility.
In this paper we describe an approach for the automated cataloguing of intact protein molecular weights using 1D-SDS PAGE and LC/MS/MS. This method uses proteins identified in a common gel slice to act as internal standards for each other for the determination of molecular weight of proteins found in that gel slice. We have applied our method to the total proteome profile of lymphoblastoid cells grown on RPI medium.
Results
Sample preparation and analysis by mass spectrometry
Lymphoblastoid cells grown in suspension were collected, pelleted and washed, and then lysed by the direct addition of SDS. The total cell lysate was separated on a 16 cm 4–20% gel and stained with Coomassie blue. The entire gel lane was then sliced into 50 fractions, and each was digested manually with trypsin [19]. Peptides were extracted, dried and resuspended in 0.1% formic acid. The fractions were sequentially run on a C18 column with two-hour gradients. Raw data files were analysed with SEQUEST [20]. Fully tryptic peptides which had Xcorr scores that exceeded a threshold (1.75, 2.5, 3.5 for charge states +1,+2,+3, DelCn > 0.1) were compiled.
This procedure identified 1982 proteins (excluding keratins) from 5972 tryptic peptides (see Additional File 1) which differ in their amino acid sequence (hereafter referred to as unique-sequence peptides). We then created a subset of that data, requiring that a protein be identified by at least 2 of the above peptides in a single gel-slice fraction. This process did not include those proteins that were identified by two unique-sequence peptides if they were from different gel-slice fractions. This subset of data contained a total of 850 proteins and 4256 unique-sequence peptides, eliminating a total of 1132 proteins and 1716 peptides. All further analyses were performed on the 850 proteins that were identified by at least two unique-sequence peptides in at least one gel slice.
Method for identification of well-localized proteins
In order to calculate the average molecular weight of proteins within a gel slice, we identified those proteins that migrated as a single well-resolved band in the gel. This was necessary, as we frequently observe that very abundant proteins "smear" along the gel and can be found in all regions of the gel. For example, the worst offender, alpha actin (NP_001091), was observed by at least two unique-sequence peptides in 39 of the 50 gel slices. If actin were included it would distort the average molecular weight calculation in many of the gel slices.
We developed a custom algorithm, called MWFilter [21], to assign a gel localization score, LScore, to each of the 850 proteins. Proteins which migrate as a single well-resolved band have low LScores, and proteins which are smeared out into many fractions have high LScores. LScores are calculated by utilizing the peptide distribution for a given protein, and is the normalized sum of all distances from a peptide hit to the peak of the peptide hit distribution. So, if the jth protein has peptide hits in n gel slices and the peak of the peptide hit distribution is given by the coordinates (xp, yp) then its localization score is given by the following equation:
If a protein has all its identified peptides in only one fraction then this protein's LScore = 0. For a protein which has peptides in multiple fractions, the algorithm selects the fraction with the greatest number of peptides for that protein, and then calculates the "distance" of all other peptides from that fraction. As another example, actin has an LScore = 45.8. The distribution of LScores for the 850 proteins is shown in Figure 1.
Figure 1 LScores for observed proteins. LScores are calculated for each protein based on the gel-slice fractions in which its peptides are observed. A protein that is well-localized and only has peptides in a small number of fractions has a low LScore.
Next, we chose an LScore cut-off of one standard deviation away from the mean LScore. This value is 4.25, and separates the 850 proteins into a well-localized group (821 proteins) and a poorly localized group (29 proteins – Figure 2). MWFilter allows the user to specify alternative Lscore cutoff values. We manually inspected the 29 proteins and established that they did appear in multiple fractions spread across the gel.
Figure 2 Summary of total proteome profiling results. 850 proteins were identified by at least two tryptic peptides within one gel-slice fraction. Of these, 821 proteins were localized in the poly-acrylamide gel, while 29 were broadly distributed throughout the gel. These proteins tended to be ones that are generally considered highly-abundant proteins.
Calculation of Average Molecular Weight for each gel slice
The 821 proteins that are well-localized and are identified by at least two peptides in a single gel slice are used to calculate the average molecular weight of proteins within each individual gel slice (MWFilter allows the user to specify the number of peptides required for inclusion in this calculation. If instead the inclusion criteria is three peptides in a gel slice, the calculations are essentially unchanged for this dataset [data not published]). The average molecular weight calculation is performed in two steps. An initial molecular weight distribution is calculated as a means of identifying outliers, which are then removed, and the molecular weight distribution is recalculated in a second step. This sequence of steps was found to be necessary to properly account for modified proteins, and is treated in greater detail in the Discussion section below. Predicted masses for each observed protein were based on unmodified full-length sequences as found in RefSeq. For all proteins observed in a gel-slice fraction, we calculated the average molecular weight (AvgMW) and the standard deviation (StdDev). For the removal of outliers at this stage of the calculation, we removed those proteins whose predicted molecular weight was more than 1 standard deviation from the mean. After removal of the outliers, the AvgMW and StdDev were recalculated, and the results are shown in Fig 3.
Figure 3 Molecular weight distribution by gel slice. The average MW for each gel slice calculated for each gel slice fraction with MWFilter based on the predicted unmodified full-length sequence and plotted in red. The blue bars represent ± 2SD of the molecular weight distribution of the proteins from that fraction (i.e. they are not error bars, per se). The inset highlights the low MW region of the gel.
Next, for each protein observed in a gel slice, the algorithm compares the predicted full-length molecular weight with the range of molecular weights defined by: AvgMW +/- 2StdDev. If the predicted MW falls within this range, then the protein is scored as being in agreement. If it is outside this range, then the protein is flagged as having a significant molecular weight modification. If a protein, which has already been scored as being well-localized, has at least two peptides in multiple gel slices and is found to match its predicted MW in at least one of these slices, then the protein is considered to be within range. We found for the 821 well-localized proteins, that a total of 656 (80%) proteins showed agreement between their predicted MW and the average MW for that gel slice, and a total of 165 proteins [20%] which had a significant difference between their predicted full-length MW and their location on the gel (Figure 3).
Discussion
We have developed a software tool for the high-throughput characterization of molecular weights of intact proteins using 1D-PAGE and LC/MS/MS. An observed molecular weight is calculated for a protein based on its location on the gel and the proteins with which it co-migrates. Such an approach is attractive in that it does not require reference to an external standard, or uniform cutting of the gel from one gel to the next. Because of the inevitability of cutting protein bands into multiple gel slices when processing a lane, we devised a score that allows for peptides to be in multiple fractions, while still allowing one to exclude those, primarily abundant, proteins which smear over the entire length of the gel lane. Proteins that are well-localized on the gel and identified by at least two unique-sequence peptides in a given gel-slice fraction act as internal standards for the other proteins in that slice.
The observed molecular weight of a protein can differ from its predicted molecular weight for a number of systematic biological reasons. The mass of a protein can be increased by post-translational modifications, such as glycosylation, ubiquitination, and sumoylation, among others, while the mass can be decreased by alternative splicing and endoproteolytic cleavage. Additionally, there are reports of altered migration for some subsets of proteins, including highly acidic [22], highly basic [23], and arginine-rich proteins [24]. The detailed characterization of these protein-modifying events is one of the goals towards which our MWFilter algorithm strives, yet it also presents a challenge for any algorithm that is in essence a "voting" or "majority rules" type of algorithm. If the majority of proteins in a cell had their molecular weight systematically altered by any mechanism, an average molecular weight of a gel slice calculated from full-length sequences would not be meaningful. However, several lines of evidence indicate that this is not the case, at least in this example. First, as can be seen in figure 2, the majority of proteins, 656 (80%), have observed molecular weights that agree with their predicted molecular weight, based on their unmodified full-length sequence. Secondly, if proteins were significantly modified, it is unlikely that the calculated average molecular weights of each gel slice would be monotonically increasing, as is very nearly the case observed in Figure 3. In this sense, each slice acts as a standard for all other slices. Lastly, calculated molecular weights agree with external standards (data not shown).
In this experiment, we identified 821 proteins that migrate as localized, single bands on a 1D gel. 165 of these proteins, or 20%, have molecular weights that do not fall into the range specified by our algorithm and the proteins with which it co-migrates. 88 of the 165 proteins are observed at lower MW than predicted by the full-length sequence. These proteins are potential candidates for having alternatively spliced transcripts or may be cleaved endoproteolytically. Many proteins in this group are annotated as having signal or transit peptides. If one subtracts the mass due to the signal/transit peptides from the full-length sequence, one observes good agreement between observed and predicted MW (last column, Table 1). Additionally, we observed a total of 77 proteins that have an observed MW that is greater than that predicted by their sequence. PTMs such as glycosylation, ubiquitination and sumoylation can account for reduced migration on gels in principle, but these possibilities need to be investigated by other means.
Table 1 Proteins which are potential candidates for endoproteolytic cleavage events.
Protein ID Protein Name Predicted MW Observed MW Observed Difference Length of Transit/Signal peptide Predicted MW after cleaving of Transit or Signal peptide MW Difference for protein with cleaved leader
NP_001852 Cytochrome c oxidase subunit IV isoform 1 precursor 19577 16996 -2582 22 AA 17200 -205
NP_002483 NADH dehydrogenase 21750 16666 -5084 46 AA 17000 -334
NP_000088 Coproporphyrinogen oxidase 50175 36543 -13632 131 AA 36900 -357
NP_002114 Major histocompatibility complex, class II, DQ beta 1 precursor 29733 25896 -3837 32 AA 26300 -404
NP_004541 NADH dehydrogenase (ubiquinone) Fe-S protein 2 52545 48185 -4360 33 AA 49000 -815
NP_004083 Mitochondrial short-chain enoyl-coenzyme A hydratase 1 31371 27499 -3872 27 AA 28400 -901
NP_000933 peptidylprolyl isomerase B (cyclophilin B) 23742 19360 -4382 25 AA 20300 -940
A future goal is to extend this method to greater resolution. While 50 fractions per lane represents a practical limit for hand-digestion of gel slices, robots which perform in-gel digestion (e.g. Intavis, Cologne, Germany) can extend this number into the hundreds. It is expected that increasing the number of gel-slice fractions will reduce the spread of MW within a slice, thereby allowing the detection of smaller MW changes. These observations will be most useful when comparing a series of related conditions, where "mobility-shifts" of a protein across conditions will highlight functionally relevant changes of a protein's state. Proteins suspected of being alternatively spliced in several conditions can be easily interrogated with RT-PCR, and proteins which are not well-localized only under certain conditions can be examined for the simultaneous presence of multiple isoforms [25]. Additionally, as the analysis of protein complexes using mass spectrometry is an area of increasing interest [2,8,9], this method may be applied to protein complexes separated by native gels.
Conclusion
We have developed a set of computational tools for extracting molecular weight information of intact proteins in total proteome profiles in a high throughput manner using 1D-PAGE and LC/MS/MS, and applied this method to proteins identified from lymphoblastoid cells. The ability to characterize the molecular weight of intact proteins represents a further step in our ability to characterize gene expression at the protein level. All 50 gel slices in our experiment were assigned an average MW and corresponding StdDev, which were then used to determine the observed MW of a given protein. We identified 165 proteins (20%) that have molecular weights that differ from their predicted full-length sequence. These 165 proteins are likely to be enriched for proteins whose MW has been altered by an interesting biological process, such as alternative splicing, endoproteolytic processing, and post-translational modifications. As such, MWFilter provides a convenient entry point for the discovery and characterization of protein processing events.
Methods
Sample Preparation
Cells were grown in suspension to early stationary phase in Iscove's media containing 10% fetal calf serum and pen-strep in 5% CO2 at 37°. Cells were pelleted in a 50 ml conical tube, washed three times with PBS, and lysed by the direct addition of gel-loading buffer containing 2% SDS. The sample was sonicated to reduce viscosity. Proteins were separated on a 16 cm, 4–20% polyacrylamide gel (Jules Inc., Milford, CT) and visualized by Coomassie staining. The entire gel lane was manually cut into 50 sections, and subjected to in-gel tryptic digestion [19].
Mass spectrometry
An aliquot of each fraction was injected onto a C18 reverse phase column using a ThermoAS autosampler with Surveyor pumps (ThermoFinnigan, San Jose, CA). Nanospray columns were constructed by packing a 10 cm bed of MAGIC C18 AQ reverse phase bulk media (Michrom Inc.; Auburn, CA) into pulled, fritless 75 micron ID fused silica capillaries under pressure. Gradients were from 0%-30% B buffer in 90 minutes, followed by 30%-90% B in 10 minutes (Buffer A: 0.1% formic acid; Buffer B: 0.1% formic acid in acetonitrile). The nanospray column was directly interfaced to the orifice of an LTQ ProteomeX ion trap mass spectrometer (ThermoFinnigan) and mass spectra were recorded. From a single parent scan (MS) spectrum, the ten most abundant ions were selected for collision-induced dissociation (CID). MS2 spectra were collected for each of these top ten ions. If a particular parent ion was observed more than 3 times in a 2 minute span, it was excluded from analysis for the subsequent 3 minutes (dynamic exclusion). Mass spectra were analyzed by SEQUEST [20]. Fully tryptic peptides with a SEQUEST XCorr score of > 1.75 (Z = 1), 2.5 (Z = 2), and 3.5 (Z = 3), and DeltaCn >0.1 were queried against RefSeq entries that have index numbers of the form NP_XXXXXX.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
QRA performed sample preparation, analysis and wrote software. DN aided in algorithm development. MAW assisted in mass spec analysis. MAS and GMC participated in the design and coordination of the study.
Supplementary Material
Additional File 1
MultiConsensus Data file
Click here for file
Acknowledgements
We thank Heather Arruda, Jessica Rumpf and Myrienne Guerrier for assistance with cell culture, and Jake Jaffe for valuable assistance with mass spectrometry. D.H.N. acknowledges support from Alfred P. Sloan and U.S. Department of Energy Postdoctoral Fellowship in Computational Molecular Biology and Bioinformatics through the Office of Science (BER), U.S. Department of Energy. GMC acknowledges support from the Genomes to Life program of the U.S. Department of Energy. M.S. thanks the Whitaker Foundation Leadership Award to Boston University for support.
==== Refs
Aebersold R Mann M Mass spectrometry-based proteomics Nature 2003 422 198 207 12634793 10.1038/nature01511
Link AJ Eng J Schieltz DM Carmack E Mize GJ Morris DR Garvik BM Yates JR Direct analysis of protein complexes using mass spectrometry Nat Biotechnol 1999 17 676 682 10404161 10.1038/10890
Washburn MP Wolters D Yates JR Large-scale analysis of the yeast proteome by multidimensional protein identification technology Nat Biotechnol 2001 19 242 247 11231557 10.1038/85686
Eckerskorn C Strupat K Schleuder D Hochstrasser D Sanchez JC Lottspeich F Hillenkamp F Analysis of proteins by direct-scanning infrared-MALDI mass spectrometry after 2D-PAGE separation and electroblotting Anal Chem 1997 69 2888 2892 9253242 10.1021/ac970077e
Gygi SP Rist B Gerber SA Turecek F Gelb MH Aebersold R Quantitative analysis of complex protein mixtures using isotope-coded affinity tags Nat Biotechnol 1999 17 994 999 10504701 10.1038/13690
Ong SE Blagoev B Kratchmarova I Kristensen DB Steen H Pandey A Mann M Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics Mol Cell Proteomics 2002 1 376 386 12118079 10.1074/mcp.M200025-MCP200
Gerber SA Rush J Stemman O Kirschner MW Gygi SP Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS Proc Natl Acad Sci U S A 2003 100 6940 6945 12771378 10.1073/pnas.0832254100
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
Neville DC Rozanas CR Price EM Gruis DB Verkman AS Townsend RR Evidence for phosphorylation of serine 753 in CFTR using a novel metal-ion affinity resin and matrix-assisted laser desorption mass spectrometry Protein Sci 1997 6 2436 2445 9385646
Ficarro SB McCleland ML Stukenberg PT Burke DJ Ross MM Shabanowitz J Hunt DF White FM Phosphoproteome analysis by mass spectrometry and its application to Saccharomyces cerevisiae Nat Biotechnol 2002 20 301 305 11875433 10.1038/nbt0302-301
Beausoleil SA Jedrychowski M Schwartz D Elias JE Villen J Li J Cohn MA Cantley LC Gygi SP Large-scale characterization of HeLa cell nuclear phosphoproteins Proc Natl Acad Sci U S A 2004 101 12130 12135 15302935 10.1073/pnas.0404720101
Rush J Moritz A Lee KA Guo A Goss VL Spek EJ Zhang H Zha XM Polakiewicz RD Comb MJ Immunoaffinity profiling of tyrosine phosphorylation in cancer cells Nat Biotechnol 2005 23 94 101 15592455 10.1038/nbt1046
Lockhart DJ Winzeler EA Genomics, gene expression and DNA arrays Nature 2000 405 827 836 10866209 10.1038/35015701
Lasonder E Ishihama Y Andersen JS Vermunt AM Pain A Sauerwein RW Eling WM Hall N Waters AP Stunnenberg HG Mann M Analysis of the Plasmodium falciparum proteome by high-accuracy mass spectrometry Nature 2002 419 537 542 12368870 10.1038/nature01111
Schirle M Heurtier MA Kuster B Profiling Core Proteomes of Human Cell Lines by One-dimensional PAGE and Liquid Chromatography-Tandem Mass Spectrometry Mol Cell Proteomics 2003 2 1297 1305 14532353 10.1074/mcp.M300087-MCP200
Link AJ Robison K Church GM Comparing the predicted and observed properties of proteins encoded in the genome of Escherichia coli K-12 Electrophoresis 1997 18 1259 1313 9298646 10.1002/elps.1150180807
Caron M Imam-Sghiouar N Poirier F Le Caer JP Labas V Joubert-Caron R Proteomic map and database of lymphoblastoid proteins J Chromatogr B Analyt Technol Biomed Life Sci 2002 771 197 209 12015999 10.1016/S1570-0232(02)00040-5
Shevchenko A Wilm M Vorm O Mann M Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels Anal Chem 1996 68 850 858 8779443 10.1021/ac950914h
Eng J McCormack AL Yates JR An Approach to Correlate Tandem Mass Spectral Data of Peptides with Amino Acid Sequences in a Protein Database. J Am Soc Mass Spectrom 1994 5 976 10.1016/1044-0305(94)80016-2
[http://gg.bu.edu/proteomics/software]
Georges E Mushynski WE Chemical modification of charged amino acid moieties alters the electrophoretic mobilities of neurofilament subunits on SDS/polyacrylamide gels Eur J Biochem 1987 165 281 287 3595591 10.1111/j.1432-1033.1987.tb11439.x
Panyim S Chalkley R The molecular weights of vertebrate histones exploiting a modified sodium dodecyl sulfate electrophoretic method J Biol Chem 1971 246 7557 7560 5135315
Hu CC Ghabrial SA The conserved, hydrophilic and arginine-rich N-terminal domain of cucumovirus coat proteins contributes to their anomalous electrophoretic mobilities in sodium dodecylsulfate-polyacrylamide gels J Virol Methods 1995 55 367 379 8609202 10.1016/0166-0934(95)00085-1
Zhu J Shendure J Mitra RD Church GM Single molecule profiling of alternative pre-mRNA splicing Science 2003 301 836 838 12907803 10.1126/science.1085792
|
15941491
|
PMC1182394
|
CC BY
|
2021-01-04 16:37:15
|
no
|
Proteome Sci. 2005 Jun 8; 3:6
|
utf-8
|
Proteome Sci
| 2,005 |
10.1186/1477-5956-3-6
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-261597209810.1186/1477-7827-3-26ResearchIn vivo effect of interleukin-1beta and interleukin-1RA on oocyte cytoplasmic maturation, ovulation, and early embryonic development in the mare Caillaud Maud [email protected] Guy [email protected]érard Nadine [email protected] Physiologie de la Reproduction et des Comportements. INRA-CNRS-Université de Tours-Haras Nationaux, IFR135, 37380 Nouzilly France2005 22 6 2005 3 26 26 21 2 2005 22 6 2005 Copyright © 2005 Caillaud et al; licensee BioMed Central Ltd.2005Caillaud 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 growing body of evidence suggests that the interleukin-1 system is involved in periovulatory events. Previous work from our lab demonstrated that in the mare, interleukin-1beta (IL-1beta) increases the ovulatory rate of metaphase II oocytes. The present study was conducted to analyze in vivo the effect of IL-1 on oocyte cytoplasmic maturation, ovulation and pregnancy rate. In the present work, IL-1beta (experiment 1, n = 13; experiment 2, n = 25) and interleukin-1RA (IL-1RA; experiment 1, n = 25) were injected intrafollicularly by using the transvaginal ultrasound-guided injection method. Injections were performed on cyclic mares when the diameter of the growing dominant follicle reached 30–34 mm. In experiment 1, mares were inseminated the day of the treatment and all the other day until ovulation. The time of ovulation was determined and a pregnancy diagnosis was performed 14 days after ovulation of the injected follicle. In experiment 2, the cumulus-oocyte complex from each injected follicle was collected by transvaginal ultrasound-guided aspiration 38 h after the intrafollicular injection. Oocyte nuclear stage and oocyte cytoplasmic maturation were assessed by analyzing chromatin configuration, cortical granules migration and mitochondria distribution under a confocal microscope. The results from experiment 1 confirm that an intrafollicular injection of 1 microgram IL-1beta induces ovulation in the mare whereas IL-1RA has no effect at the dose used in the present study. Furthemore, we demonstrated, that in our experimental conditions, IL-1beta and IL-1RA induced a decrease in embryo development. Experiment 2 leads us to observe that IL-1beta is unable to induce cortical granules migration and remodelling of mitochondria, that commonly occurs during oocyte maturation, whereas it acts on nuclear maturation. This result may explain the decrease in embryo development we observed after IL-1beta intrafollicular injection. In conclusion, the present study tends to demonstrate that IL-1beta plays a role in the ovulatory process and may acts on oocyte maturation in the mare, but additional factors are required to complete equine oocyte cytoplasmic maturation to allow embryo development.
==== Body
Background
In recent years, evidence has accumulated to suggest that cytokines are important regulators of the ovarian function. Interleukin-1 (IL-1), one major cytokine, as well as receptors (IL-1R1 and IL-1R2) and antagonist (IL-1RA) have been demonstrated to be produced in the ovary of several species [1, for review]. Recently, we demonstrated the presence of IL-1β mRNA in equine cumulus-oocyte complexes and granulosa cells, whereas immunoreactive IL-1β has been observed in equine follicular fluids [2,3].
In vitro studies have shown that IL-1β regulates some cellular activities of granulosa and theca cells, such as steroidogenesis [4,5], as well as synthesis of proteases [6-8] and prostaglandins [9,10]. Moreover, it has been demonstrated that IL-1β promotes the ovulation process in the rat [11], the rabbit [12] and the mare [13]. IL-1β increases in vitro the germinal vesicle breakdown of oocytes in the rabbit model [12], as well as in vivo in the mare [13], demonstrating a beneficial role of IL-1β in oocyte nuclear maturation. These observations led to conclude that IL-1 may be a paracrine factor that is involved in ovulation and in oocyte nuclear maturation. In 1994, Takehara et al., demonstrated that IL-1β can facilitate in vitro fertilization in the rabbit, suggesting that IL-1 may be involved in oocyte cytoplasmic maturation and could improve fertilization. In the mare, no data is available on the effect of IL-1β on oocyte cytoplasmic maturation or fertilization rates. Nevertheless, a better knowledge of regulating factors involved in these processes would help to improve the embryo production in this species.
In this context, the objectives of the present study were to investigate in vivo in the mare, the potential role of IL-1β on oocyte cytoplasmic maturation, ovulation and embryo development.
Materials and methods
Animals and Treatments
All animal procedures were approved by the Agricultural Agency and Scientific Research Agency (approval number A37801) and conducted in accordance with guidelines for Care and Use of Agricultural Animals in Agricultural Research and Teaching.
Adult cyclic pony mares in good body condition, kept indoors and fed with concentrates, were used. Mares received a prostaglandin F2α analogue (Cloprostenol [estrumate], 125 μg/mare i.m.; Scherring-Plough, Levallois-Peret, France) during the midluteal phase to induce luteolysis. Ovarian activity was then assessed by routine rectal ultrasound scanning. Intrafollicular injections were performed in the dominant follicle, between 30–34 mm in diameter, at the end of the follicular phase (see below). Before intrafollicular injections or punctures, the mares were sedated with detomidine (0.6 mg/100 kg body weight [BW] i.v., Domosedan; Pfizer, Amboise, France) and propantheline bromide (3 mg/100 kg BW i.v.; Sigma, St. Louis, MO) was administered to achieve rectal relaxation. After intrafollicular injection or punctures, mares were treated with an antibiotic (Intramicine, 1 600 000 IU penicillin/100 kg BW and 1.3 g dihydrostreptomycin/100 kg BW i.m.; Sanofi, Libourne, France).
Crude Equine Gonadotropin (CEG) was prepared from horse pituitaries that were supplied commercially. The resulting preparation of CEG contained approximately 6% horse LH and 2–4% horse FSH. An i.v. injection of 15 mg CEG was performed when the diameter of the dominant follicle reached 30–34 mm in order to induce ovulation. Ovulation occurs between 35–40 hours after the injection [14]
Intrafollicular Injection Procedure
The injection was performed using a transvaginal ultrasound-guided system. A 60-cm-long single lumen needle of 1.1-mm outer diameter was inserted into the dominant follicle the day it reached 30–34 mm in diameter. Follicular fluid (2.5 ml) was aspirated in a syringe directly connected to the needle. The studied molecule was diluted in 2 ml of PBS (Dulbecco 'A'; Unipath, Dardilly, France) and maintained at 37°C. The 2 ml were then injected in the follicle by using a syringe directly connected to the needle. Follicular fluid (0.5 ml) from the first syringe was then injected back into the follicle to rince the needle and reestablish the initial follicular volume.
Experimental Designs
In experiment 1, 61 cyclic mares were used. They were divided into five groups. Mares from the first group (negative control group; n = 12) received an intrafollicular injection (i.f.) of 2 ml of PBS and an jugular injection (i.v.) of 5 ml of saline. Mares from the second group (IL-1RA group; n = 12) received an i.f. injection of 2 ml of recombinant human interleukin-1RA (2.5 μg/ml in PBS, rhIL-1RA; R&D System, Abingdon, UK) and an i.v. injection of 5 ml of saline. Mares from the third group (IL-1β group; n = 13) received an i.f. injection of 2 ml of recombinant human IL-1β (0.5 μg/ml in PBS, rhIL-1β; R&D System) and an i.v. injection of 5 ml of saline. Mares from the fourth group (positive control group; n = 12) received an i.f. injection of 2 ml of PBS and an i.v. injection of 15 mg CEG in 5 ml of saline. Mares from the last group (IL-1RA/CEG group; n = 13) received an i.f. injection of 2 ml of recombinant human interleukin-1RA (2.5 μg/ml in PBS, rhIL-1RA) and an i.v. injection of 15 mg CEG in 5 ml of saline.
In experiment 2, 63 cyclic mares were used. They were divided into three groups. Mares from the first group (positive control group; n = 23) received an i.f. injection of 2 ml of PBS and an i.v. injection of 15 mg CEG in 5 ml of saline. Mares from the second group (IL-1β group; n = 25) received an i.f. injection of 2 ml of recombinant human IL-1β (0.5 μg/ml in PBS, rhIL-1β; R&D System) and an i.v. injection of 5 ml of saline. Mares from the last group (negative control group; n = 15) received an i.f. injection of 2 ml of PBS and an i.v. injection of 5 ml of saline.
Artificial insemination, Detection of Ovulation and Pregnancy Diagnosis
Mares from experiment 1 were inseminated with fresh semen collected from a fertile stallion, using 200 × 106 spermatozoa per insemination. Inseminations were performed on the day of the intrafollicular injection, on the day after and every 2 days until ovulation of the injected follicle. Ovulation was assessed ultrasonographically twice a day from 8 h after the intrafollicular injection until ovulation (i.e., absence of a large dominant follicle and presence of a corpus luteum). Fourteen days after ovulation, pregnancy diagnosis was performed by routine ultrasonography of the uterus.
Cumulus Oocyte Complex (COC) Recovery and Analysis
COCs were collected from injected follicles of experiment 2. Transvaginal ultrasound-guided aspiration was used 38 h after the intrafollicular injection, as previously described [15]. Briefly, a single lumen needle (1.8-mm outer diameter) was used to aspirate follicular fluid. The follicle was then flushed several times with PBS containing heparin (2.5 IU/ml; LEO S.A., St Quentin Yvelines, France) at 37°C and scraped with the needle in order to pick up the oocyte. All aspirated fluids were examined with a stereomicroscope for oocyte recovery. Cumulus cells were removed from the oocytes by repeated pipetting and subsequent treatment with 1% hyaluronidase (type III, 875IU/mg; Sigma, La Verpillère, France). Oocytes were then incubated for 30 min at 37°C in PBS containing 3% BSA and 280 nM MitoTracker Orange CMTM Ros (Molecular Probes, Oregon, USA) for mitochondria detection. Then, oocytes were washed three times in pre-warmed PBS without BSA. The oocytes were then fixed for 20 min at 37°C using freshly prepared 2.5 % paraformaldehyde in PBS. After fixation, the oocytes were washed three times in PBS. The oocytes were then permeabilized in Triton X-100 (Sigma), 0.1% in PBS, for 5 min at room temperature and washed in PBS. Oocytes were then incubated for 30 min at room temperature in 100 μg/ml of fluorescein isothiocyanate-conjugated peanut agglutinin in washing solution (PBS containing 0.05 % NaN3 and 1 mM PMSF) to detect distribution of cortical granules. After rinsing, oocytes were stained with 1 μg/ml bis-benzimide (Hoechst 33342; Sigma) in PBS for 6 min at room temperature for DNA detection. They were then mounted between slide and cover slide in a mixture of Moviol V4–88 (133 mg/ml; Hoechst, Frankfurt, Germany) and n-propyl gallate (5 mg/ml; Sigma). The slides were kept at 4°C in darkness until observation. The oocytes were observed under a confocal laser scanning microscope (Olympus, France). Mitochondria and cortical granules labelling were visualized in the confocal mode, whereas Hoechst staining was detected by conventional epifluorescence. For fluorescein, an argon ion laser adjusted at 525-nm emission wave length was used; for MitoTracker Orange CMTM Ros, a helium-neon ion laser adjusted at 576-nm was used.
Statistical analysis
The nonparametric tests of Kruskal-Wallis and Wilcoxon-Mann-Withney were performed using StatXact 5 software (CYTEL, Cambridge, MA ) in order to compare time of ovulation and percentage of embryos. The corrected chi-square test was used to compare oocyte in vivo maturation rates.
Results
Experiment 1: Effect of IL-1β and IL-1RA intrafollicular injection on ovulation and embryo development
During this experiment, 61 follicles were injected. Twelve to thirteen mares were used per group, in 5 groups.
Time of ovulation
Time of ovulation after intrafollicular injection of PBS, IL-1β or IL-1RA was determined by ultrasonography. We determined three intervals during which the mares ovulated: between 20 and 31 h, between 31 and 47 h, and more than 47 h after the intrafollicular injection. Figure 1 illustrates the ovulation profiles after intrafollicular injection of PBS (2 ml), IL-1β (1 μg/2 ml), or IL-1RA (5 μg/2 ml). In the positive control group (PBS i.f., CEG i.v.), mares ovulated mainly between 31 and 47 h after the intrafollicular injection (7/12). Mares from the IL-1β group (IL-1β i.f., saline i.v.) ovulated mainly in the same interval (9/13). The distribution of the time of ovulation was not significantly different between these two groups. In the negative control group (PBS i.f., saline i.v.), most of the mares ovulated more than 47 h after the i.f. injection (11/12). The IL-1RA group (IL-1RA i.f., saline i.v.) displayed a similar distribution of time of ovulation (11/12 ovulations after 47 h). This distribution was significantly different from that of the positive control group (p < 0.002) and IL-1β group (p < 0.001). In the IL-1RA/CEG group (IL-1RA i.f., CEG i.v.), the length of time from injection to ovulation was heterogeneous. Half of the mares (6/13) ovulated between 31 and 47 h after the i.f. injection, whereas 2/13 ovulated during the first interval (20–31 h) and 5/13 ovulated more than 47 h after the i.f. injection. This distribution was not significantly different from that of the positive control group and IL-1β group but differed significantly from the negative control group and the IL-1RA group (p < 0.01 each).
Figure 1 Length of time from injection to ovulation after intrafollicular administration of IL-1β or IL-1RA in equine dominant follicles. For each group, histograms represent the percentage of mares that ovulated during each time interval. The proportion of mares is indicated on the top of each histogram. The distribution of the time of ovulation was not significantly different between the positive control group and the IL-1β group. In the negative control group and in the IL-1RA group, the distribution of the time of ovulation was significantly different from that in the positive control group and in the IL-1β group. In the IL-1RA/CEG group, this distribution was not significantly different from that in the positive control group and the IL-1β group but differed significantly from that in the negative control group and in the IL-1RA group.
Embryo development
Presence or absence of an embryo was determined in each mare by ultrasonography 14 days after ovulation. Table 1 illustrates the percentage of embryos observed after intrafollicular injection of PBS (2 ml), IL-1β (1 μg/2 ml), or IL-1RA (5 μg/2 ml) coupled to an intraveneous injection of either saline or CEG. The percentage of embryos was the same in the two control groups (67%). In contrast, the percentage of embryos was lower in the three other groups (25%, 23%, 23%) which received IL-1β or IL-1RA intrafollicularly. These lower percentages are not statistically different from one to the other but are significantly different from that of both control groups (p < 0.05).
Table 1 Percentage of embryos after intrafollicular administration of IL-1β or IL-1RA in equine dominant follicles.
Number of injected follicles Percent of embryos
PBS i.f., saline i.v. (negative control group) 12 67 (8/12)a
IL-1RA i.f., saline i.v. (IL-1RA group) 12 25 (3/12)b
IL-1β i.f., saline i.v. (IL-1β group) 13 23 (3/13)b
PBS i.f., CEG i.v. (positive control group) 12 67 (8/12)a
IL-1RA i.f., CEG i.v. (IL-1RA/CEG group) 13 23 (3/13)b
PBS, Phosphate buffer saline; CEG, crude equine gonadotropin; IL-1β, interleukin 1 beta; IL-1RA, interleukin 1 receptor antagonist; i.f., intrafollicular; i.v., intravenous. a and b are significantly different (p < 0.05).
Experiment 2: Effect of IL-1β on oocyte nuclear and cytoplasmic maturation
Cumulus Oocyte Complex recovery and Oocyte nuclear stage at recovery
In this experiment, 3 groups were constituted: positive control group, IL-1β group, negative control group. Twenty three to 25 mares were used in the 2 first groups and 15 mares were used in the third group. Out of the sixty three injected follicles that were punctured, 52 oocytes were collected and analyzed. This corresponds to 82.5% of collection rate. The collection rate was similar in each group (table 2). All aspirated COCs had expanded cumuli.
As shown in Table 2, the rate of oocytes that display resumption of meiosis was significantly higher in the positive control group (PBS i.f., CEG i.v.) than in the negative control group (PBS i.f., saline i.v.; p < 0.04). The other oocytes were either immature or degenerated. Moreover, 60% of oocytes from the IL-1β group display resumption of meiosis, that tends (p < 0,1) to be higher than that in the negative control group.
Table 2 Number of punctured follicles, percentage of recovered cumulus-oocyte complexes, and percentage of degenerated, immature, metaphase I and metaphase II oocytes in each group.
COCs Oocytes
Number of injected-punctured follicles Percent of recovery Percent of DEG Percent of immature Resumption of meiosis
PBS i.f., CEG i.v. 23 91 (21/23)a 19 (4/21) 0 (0/21) 81 (17/21)a
IL-1β i.f., saline i.v. 25 80 (20/25)a 20 (4/20) 20 (4/20) 60 (12/20)ab
PBS i.f., saline i.v. 15 73 (11/15)a 27 (3/11) 27 (3/11) 45 (5/11)b
PBS, Phosphate buffer saline; CEG, crude equine gonadotropin; IL-1β, interleukin 1 beta; i.f., intrafollicular; i.v., intravenous; DEG, degenerated; MI metaphase I; MII, metaphase II. a and b are significantly different (P < 0.05).
Cortical granules localization and Mitochondria distribution
Mature (MII) oocytes were analyzed in order to visualize cortical granules and mitochondria. Thirteen matures oocytes from the positive control group were analyzed. Cortical granules migration was achieved or in progress in most oocytes from this group (8/13), with cortical granules lining the oolemma (Figure 2A). In the 5 other oocytes from this group, cortical granules were located in the medullary zone. In 3 to 5 mature oocytes analyzed from the IL-1β group, cortical granules had an uniform distribution between the oolemma and the medullary zone (Figure 2B), and only 2 oocytes displayed cortical granules migration in progress.
In most oocytes from the positive control group (9/13), the distribution of mitochondria was heterogeneous i.e. labelled mitochondria were distributed unevenly within the cytoplasm (data not shown). Mitochondria distribution in oocytes from the IL-1β group was more homogeneous with labelled mitochondria distributed evenly throughout the cytoplasm (data not shown).
Figure 2 Cortical granules localization on serial optical sections in equine oocytes after intrafollicular injection of PBS (A) or IL-1β (B) and intravenous injection of CEG (A) or saline (B). 1–6: different sections from the top of the oocyte to the bottom of the oocyte A) Cortical granules migration achieved: the majority of the cortical granules are lining the oolemma. B) No cortical granules migration: the cortical granules are located in the medullary zone, and no cortical granules are lining the oolemma; X40
Discussion
The aim of the present work was to analyze in vivo the effect of IL-1β and IL-1RA, after their intrafollicular injection into the preovulatory follicle. In experiment 1, ovulation and embryo development were studied. In experiment 2, the effect of IL-1β on oocyte nuclear and cytoplasmic maturation was investigated.
The intrafollicular injection technique is an interesting alternative to study the effect of a factor on oocyte and follicle maturation. It has been used in different species like rabbit [16], ewe [17], cow [18], rhesus monkey [19], and mares [13,15,20,21]. In the mare, this approach needs no surgery because the large size of the preovulatory follicle and the presence of a tunica albuginea, which allows injecting follicles with limited leakage by using the transvaginal ultrasound echo-guided method. In the present work, 103 preovulatory follicles were successfully injected and only 5 suffered serious damage.
In our study, 1 μg IL-1β was injected in the follicle. This dose was chosen regarding a dose-dependent study that had been performed in vitro [22], and tested in vivo [13]. Similarly, the dose of IL-1RA we used in the present study (5 μg) was chosen in accordance to a previous in vivo study [13] where 1 μg IL-1RA had no effect.
We demonstrated in the present work, that the intrafollicular injection of IL-1β induced synchronized ovulations. Actually, IL-1β gave the same result, in terms of ovulation distribution, as an i.v. injection of CEG [14]. This observation is in accordance with the recent work performed by Martoriati et al. [13]. CEG is a pituitary extract used to induce ovulation in the mare. By its LH activity, CEG is used to give a precise timing of ovulation. It is preferentially used in this specie instead of hCG, which is known to induce immunoreaction after successive injections. Although, the mechanisms by which IL-1 regulates follicular maturation is unclear, our observations suggest that IL-1 system could be involved in periovulatory events. Thus, the mechanism of action of IL-1β remains to be elucidated, but we can hypothesize that the increase in IL-1β intrafollicular level after injection may mimics the local preovulatory events that precede ovulation.
That hypothesis explains why we constituted the IL-1RA-CEG group, in which the intrafollicular injection of IL-1RA would be able to block the CEG effect. After such treatment (IL-1RA i.f. and CEG i.v.), we observed that half of mares ovulated more than 47 h after injections. In such mares, we can conclude either that CEG had no effect (that is unlikely at this rate), or that the intrafollicular injection of IL-1RA at the dose we used, inhibited the ovulatory effect of CEG. Other mares from this group ovulated between 31 and 47 h after injection. In these mares and in our experimental conditions, IL-1RA did not inhibit the CEG ovulatory effect. These mares would probably had the process of ovulation engaged (i.e. increase in the endogenous plasmatic concentration of LH) before IL-1RA and CEG injections. The data we obtained from the IL-1RA-CEG group are in accordance with a previous study performed in the rat, that showed a time-dependent inhibition of hCG-induced ovulation by IL-1RA [23].
As demonstrated in the present work, 60% of the oocytes collected 38 h after IL-1β intrafollicular injection displayed resumption of meiosis. This rate only tends to be higher than that of the negative control group (p < 0,1). This result is consistent with our previous result obtained in the mare (57%) [13]. Taken together, both studies imply a positive effect of IL-1β on equine oocyte nuclear maturation.
Considering this result, it was then questionable if ovulated mature equine oocytes obtained after an intrafollicular injection of IL-1β were able to be fertilized. The very low rate of embryo development that we observed in our study after IL-1β intrafollicular injection (25%) leads us to hypothesize that IL-1β may have a negative effect on oocyte quality.
We tested this hypothesis by analyzing metaphase II (MII) oocytes for cortical granules localization and mitochondria distribution as criteria for cytoplasmic maturation. To our knowledge, this is the first study of the effect of IL-1β on oocyte cytoplasmic maturation that has been performed in vivo in any species. In several mammals, it has been reported that cortical granules are mainly lining the oolemma of oocytes at the time of ovulation (in sheep [24] ; in bovine [25-27] ; in equine [28]). In our study, cortical granules migration was achieved or in progress in most MII oocytes from the positive control group (CEG group). In contrast, cortical granules did not line the oolemma in most MII oocytes from IL-1β injected follicle, but remained homogeneously distributed in the cytoplasm. This data suggests that in our study, IL-1β was unable to induce cortical granules migration, and thus, to be fully involved in oocyte cytoplasmic maturation. Additional factors may be required for cortical granules migration.
The distribution of mitochondria in MII oocytes was studied as a second criterion of cytoplasmic maturation. Two mitochondrial distribution patterns were observed in equine oocytes. The oocytes from the positive control group (CEG group) displayed a clustering of mitochondria around cytoplasmic vacuoles (heterogeneous pattern), which observation is in accordance with previous studies (bovine [29] ; human [30,33,34] ; porcine [31,32]). The second mitochondrial distribution pattern that we observed mainly concerned oocytes from the IL-1β group. In these oocytes, mitochondria displayed a homogeneous distribution. Such a homogeneous pattern of mitochondria had previously been observed in immature porcine oocytes [32]. This data suggests that in our experimental conditions, IL-1β is unable to induce the remodelling of mitochondria that occurs during oocyte maturation.
All in all, whereas IL-1β seems to be able to act on resumption of meiosis, it is not sufficient for complete oocyte cytoplasmic maturation, at least in the mare.
To be fertilized, an oocyte has to be mature up to his nucleus and his cytoplasm. Thus, we could propose that the low number of embryos conceived after IL-1β intrafollicular injection could be probably due to an abnormal cytoplasmic maturation of oocytes. This result seems at odds with a previous study performed on perfused rabbit ovary, in which IL-1β facilitated fertilization [12]. However, these authors reported that most of fertilized oocytes were arrested at the four-cell stage. Regarding this findings, two hypothesis can be proposed to explain our results. The first one is that IL-1β is able to act on resumption of meiosis, but cannot bring about cytoplasmic maturation. The consequence is that oocytes are not suitable for fertilization. The second hypothesis is that IL-1β may exert a cytotoxic effect on oocytes, that would allow fertilization but prevent subsequent embryonic development. In our study, pregnancy diagnosis was performed 14 days after ovulation, that do not allow to verify one of these two hypothesis. To alleviate this problem, in vitro fertilization (IVF) could be performed on IL-1β-induced matured oocyte.
Conclusion
In conclusion, the mechanism by which IL-1β acts at the ovarian level is unclear, and the present data tend to demonstrate that IL-1β alone is not able to promote cytoplasmic maturation of equine oocyte, and thus to allow embryo development. Nevertheless, IL-1β may play an essential role in the physiology of equine oocytes by acting on meiosis resumption as well as in the ovarian function by inducing ovulation.
Authors' contributions
MC participated to the conception of the study, coordinated and participated to the experimental procedures, analyzed results and drafted the manuscript. GD carried out and coordinated the experimental procedures. NG conceived the study, participated in its design and to the experimental procedures. NG participated to the final analysis of results and corrected the manuscript. All the authors read and approved the final manuscript.
Financial support
This work was supported by grants from the Institut National de la Recherche Agronomique, France, and the Haras Nationaux, France. Maud Caillaud was supported by a fellowship from the Région Centre, France and the Haras Nationaux, France.
Acknowledgements
The authors wish to thank Cécile Lahuec, Carine Rouchez, Ophélie Dhumez, Bernard Bruneau and the staff of the experimental stud for technical assistance. We are grateful to Jean-Marie Yvon and Dr. Michèle Magistrini for sperm preparation for artificial insemination, to Mr Pierre-Yves Sizaret, who taught us manipulation of the confocal microscope, to Dr H.Torner (from Dummerstorf, Germany), who gave us the protocol for mitochondria labelling.
==== Refs
Gérard N Caillaud M Martoriati A Goudet G Lalmanach AC The interleukin-1 system and female reproduction J Endocrinol 2004 180 203 212 14765973 10.1677/joe.0.1800203
Martoriati A Lalmanach AC Goudet G Gérard N Expression of interleukin-1 (IL-1) system genes in equine cumulus-oocyte complexes and influence of IL-1β during in vitro maturation Biol Reprod 2002 67 630 636 12135907
Martoriati A Gérard N Interleukin-1 (IL-1) system gene expression in granulosa cells: kinetics during terminal preovulatory follicle maturation in the mare Reprod Biol Endocrinol 2003 16 42 12803652 10.1186/1477-7827-1-42
Nakamura Y Kato H Terranova PF Interleukin-1α increases thecal progesterone production of preovulatory follicles in cyclic hamsters Biol Reprod 1990 43 169 173 2378930
Best CL Hill JA Interleukin-1α and -β modulation of luteinized human granulosa cell oestrogen and progesterone biosynthesis Hum Reprod 1995 10 3206 3210 8822445
Hurwitz A Dushnik M Solomon H Ben-Chetrit A Finci-Yeheskel Z Milwidsky A Mayer M Adashi EY Yagel S Cytokine-mediated regulation of rat ovarian function: interleukin-1 stimulates the accumulation of a 92-kilodalton gelatinase Endocrinology 1993 132 2709 2714 8389288 10.1210/en.132.6.2709
Bonello NP Norman RJ Brännström M Interleukin-1β inhibits luteinizing hormone-induced plasminogen activator activity in rat preovulatory follicles in vitro Endocrine 1995 3 49 54
Karakji EG Tsang BK Regulation of rat granulosa cell plasminogen activator system: influence of interleukin-1β and ovarian follicular development Biol Reprod 1995 53 1302 1310 8562685
Kokia E Hurwitz A Ricciarelli E Tedeschi C Resnick CE Mitchell MD Adashi EY Interleukin-1 stimulates ovarian prostaglandin biosynthesis: evidence for heterologous contact-independent cell-cell interaction Endocrinology 1992 130 3095 3097 1572315 10.1210/en.130.5.3095
Brännström M Wang L Norman RJ Effects of cytokines on prostaglandin production and steroidogenesis of incubated preovulatory follicles of the rat Biol Reprod 1993 48 165 171 8418905
Brännström M Wang L Norman RJ Ovulatory effect of interleukin-1 beta on the perfused rat ovary Endocrinology 1993 132 399 404 8419137 10.1210/en.132.1.399
Takehara Y Dharmarajan AM Kaufman G Wallach EE Effect of interleukin-1 beta on ovulation in the in vitro perfused rabbit ovary Endocrinology 1994 134 1788 1793 8137743 10.1210/en.134.4.1788
Martoriati A Duchamp G Gérard N In vivo effect of epidermal growth factor, interleukin-1β, and interleukin-1RA on equine preovulatory follicles Biol Reprod 2003 68 1748 1754 12606330 10.1095/biolreprod.102.012138
Duchamp G Bour B Combarnous Y Palmer E Alternative solutions to hCG induction of ovulation in the mare J Reprod Fertil 1987 221 228
Gastal EL Kot K Ginther OJ Ultrasound-guided intrafollicular treatment in mares Theriogenology 1995 44 1027 1037 16727797 10.1016/0093-691X(95)00289-K
Jones EE Nalbandov AV Effects of intrafollicular injection of gonadotrophins on ovulation or luteinization of ovarian follicle Biol Reprod 1972 7 87 93 5050153
Murdoch WJ De Silva M Dunn TG Luteal phase insufficiency in the ewe as a consequence of premature induction of ovulation by intrafollicular injection of gonadotropins J Anim Sci 1983 57 1507 1511 6425254
Kot K Gibbons JR Ginther OJ A technique for intrafollicular injection in cattle: effect of hCG Theriogenology 1995 44 41 50 10.1016/0093-691X(95)00146-Y
Hazzard TM Rohan RM Molskness TA Fanton JW D'Amato RJ Stouffer RL Injection of antiangiogenic agents into the macaque preovulatory follicle: disruption of corpus luteum development and function Endocrine 2002 17 199 206 12108520 10.1385/ENDO:17:3:199
Watson ED Sertich PL Concentrations of arachidonate metabolites, steroids and histamine in preovulatory horse follicles after administration of human chorionic gonadotrophin and the effect of intrafollicular injection of indomethacin J Endocrinol 1991 129 131 139 2030323
Gérard N Duchamp G Goudet G Bézard J Magistrini M Palmer E A high-molecular-weight preovulatory stage-related protein in equine follicular fluid and granulosa cells Biol Reprod 1998 58 551 557 9475413
Martoriati M Caillaud M Goudet G Gérard N Inhibition of in vitro maturation of equine oocytes by interleukin 1β via specific IL-1 receptors Reproduction 2003 126 509 515 14525533 10.1530/rep.0.1260509
Simon C Tsafriri A Chun SY Piquette GN Dang W Polan ML Interleukin-1 receptor antagonist suppresses human chorionic gonadotropin-induced ovulation in the rat Biol Reprod 1994 51 662 667 7819447
Cran DG Moor RM Hay MF Fine structure of the sheep oocyte during antral follicle development J Reprod Fertil 1980 59 125 132 7190607
Kruip ThAM Cran DG Beneden Th van Dieleman SJ Structural changes in bovine oocytes during final maturation in vivo Gamete Res 1983 8 9 47 10.1002/mrd.1120080105
Hyttel P Callesen H Greve T Ultrastructural features of preovulatory oocyte maturation in superovulated cattle J Reprod Fertil 1986 76 645 656 3084771
Hyttel P Greve T Callesen H Ultrastructural aspects of oocyte maturation and fertilization in cattle J Reprod Fertil 1989 35 47
Goudet G Bézard J Duchamp G Gérard N Palmer E Equine oocyte competence for nuclear and cytoplasmic in vitro maturation: effect of follicle size and hormonal environment Biol Reprod 1997 57 232 245 9241036
Stojkovic M Machado SA Stojkovic P Zakhartchenko V Hutzler P Goncalves PB Wolf E Mitochondrial distribution and adenosine triphosphate content of bovine oocytes before and after in vitro maturation: correlation with morphological criteria and developmental capacity after in vitro fertilization and culture Biol Reprod 2001 64 904 909 11207207
Motta PM Nottola SA Makabe S Hyen R Mitochondrial morphology in human fetal and adult female germ cells Hum Reprod 2000 15 129 147 11041520
Sun QY Wu GM Lai L Park KW Cabot R Cheong HT Day BN Prather RS Schatten H Translocation of active mitochondria during pig oocyte maturation, fertilization and early embryo development in vitro Reproduction 2001 122 155 163 11425340 10.1530/rep.0.1220155
Torner H Brüssow KP Alm H Ratky J Pöhland R Tuchscherer A Kanitz W Mitochondrial aggregation patterns and activity in porcine oocytes and apoptosis in surrounding cumulus cells depends on the stage of pre-ovulatory maturation Theriogenology 2004 61 1675 1689 15019463 10.1016/j.theriogenology.2003.09.013
Sathananthan AH Ultrastructure of the human egg Hum Cell 1997 10 21 38 9234062
Sathananthan AH Trounson AO Mitochondrial morphology during pre-implantional human embryogenesis Hum Repro 2000 15 148 159
|
15972098
|
PMC1182395
|
CC BY
|
2021-01-04 16:37:11
|
no
|
Reprod Biol Endocrinol. 2005 Jun 22; 3:26
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-26
|
oa_comm
|
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-231597812510.1186/1742-4682-2-23ResearchPredicting transcription factor activities from combined analysis of microarray and ChIP data: a partial least squares approach Boulesteix Anne-Laure [email protected] Korbinian [email protected] Department of Statistics, University of Munich, Ludwigstr. 33, D-80539 Munich, Germany2005 24 6 2005 2 23 23 15 4 2005 24 6 2005 Copyright © 2005 Boulesteix and Strimmer; licensee BioMed Central Ltd.2005Boulesteix and Strimmer; 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 study of the network between transcription factors and their targets is important for understanding the complex regulatory mechanisms in a cell. Unfortunately, with standard microarray experiments it is not possible to measure the transcription factor activities (TFAs) directly, as their own transcription levels are subject to post-translational modifications.
Results
Here we propose a statistical approach based on partial least squares (PLS) regression to infer the true TFAs from a combination of mRNA expression and DNA-protein binding measurements. This method is also statistically sound for small samples and allows the detection of functional interactions among the transcription factors via the notion of "meta"-transcription factors. In addition, it enables false positives to be identified in ChIP data and activation and suppression activities to be distinguished.
Conclusion
The proposed method performs very well both for simulated data and for real expression and ChIP data from yeast and E. Coli experiments. It overcomes the limitations of previously used approaches to estimating TFAs. The estimated profiles may also serve as input for further studies, such as tests of periodicity or differential regulation. An R package "plsgenomics" implementing the proposed methods is available for download from the CRAN archive.
==== Body
Background
The transcription of genes is regulated by DNA binding proteins that attach to specific DNA promoter regions. These proteins are known as transcriptional regulators or transcription factors and recruit chromatin-modifying complexes and the transcription apparatus to initiate RNA synthesis [1,2].
In the last few years, considerable efforts have been made by both experimental and computational biologists to identify transcription factors, their target genes and the sensitivity of the regulation mechanism to changes in environment [3-5]. An important technique for the identification of target genes bound in vivo by known transcription factors is the combination of a modified chromatin immunoprecipitation (ChIP) assay with microarray technology, as proposed by Ren et al. [1]. For instance, in the budding yeast Saccharomyces cerevisiae, ChIP experiments have been utilized to elucidate the binding interactions between 6270 genes and 113 preselected transcription factors [2]. However, as physical binding of transcription factors is a necessary but not a sufficient condition for transcription initiation, ChIP data typically suffer from a large proportion of false positives.
Several attempts have also been made to recover the network structure between transcription factors and their targets using only the gene expression levels of both the transcription factors and the targets, either with [6] or without [7] assuming a subset of putative regulators. Such approaches implicitly assume that the measured gene expression levels of the transcription factors reflect their actual activity. However, owing to various complex post-translational modifications as well as to interactions among transcription factors themselves, regulator transcription levels are generally inappropriate proxies for transcription factor activities (TFA).
In a few recent papers, integrative analysis of gene expression data and ChIP connectivity data has been suggested as a way of overcoming these difficulties [8]. Most prominently, Liao and coworkers have developed the technique of "network component analysis" (NCA) [9,10], a dimension reduction approach to inferring the true regulatory activities. In NCA one can also incorporate further a priori qualitative knowledge about gene-transcription factor interactions [11]. Unfortunately, a major drawback of the original NCA method is that for identifiable reasons it imposes very strong restrictions on the network topologies allowed, which renders application of classic NCA difficult in many practical cases. Alter and Golub [12] introduced an approach for integrating ChIP and microarray data using pseudo-inverse projection. Like NCA, this method is based on algebraic matrix decomposition (in this case singular value decomposition). However, this ignores measurement and biological errors present in both connectivity and gene expression data. Kato et al. [13] proposed yet another integrative approach consisting of several steps combining sequence data, ChIP data and gene expression data. However, here gene expression is used only to check the coherence of expression profiles of genes with common sequence motifs, and not to estimate transcription factor activities. Finally, Gao et al. [14] suggested the "MA-Networker" algorithm, which employs multivariate regression to estimate TFAs and backward variable selection to identify the active transcription factors. Unlike the other approaches, it takes full account of stochastic error. However, for classical regression theory to be valid it is necessary not only that the number of gene targets is much greater than both the number of samples and the number of transcription factors, but also that the transcription factors are independent of each other. The latter condition in particular is clearly not generally satisfied with genome data.
Here, we suggest an alternative statistical framework to tackle the problem of network component and regulator analysis. Our approach centers around multivariate partial least squares (PLS) regression, a well-known analysis tool for high-dimensional data with many continuous response variables that has been widely applied, especially to chemometric data [15-17]. Using PLS we are able not only both to integrate and generalize previous NCA approaches, but also to overcome their respective limitations. In particular, PLS-based network component analysis offers a computationally highly efficient and statistically sound way to infer true TFAs for any given connectivity matrix. In addition, it allows statistical assessment of the available connectivity information, and also the discovery of interactions and natural groupings among regulatory genes (corresponding to "meta"-transcription factors).
Results
Network model
Suppose gene expression data for n genes and m samples (= arrays, tissue types, time points etc.) are collected in a n × m data matrix . Furthermore, let denote the so-called connectivity matrix with n rows and p columns. Each column in describes the strength of interaction between one of p transcription factors and the n considered gene targets. The entries of can either be binary (0–1) or numeric (e.g. ChIP data), with a zero value indicating no physical binding between a transcription factor and a target.
In order to relate expression to connectivity data we consider the linear model
where A is a n × m constant matrix, is a p × m matrix of regression coefficients and E is a n × m matrix containing error terms. A contains the m different offsets, and may be interpreted as the matrix of the true transcription factor activities (TFAs) of the p transcription factors for each of the m samples.
It is worth noting that in this setting, unlike in most other gene expression analysis studies, the number of genes n is considered as the number of cases rather than the number of variables. In the present case the latter corresponds to the number of transcription factors p (hence, in general, p <n).
NCA and MA-Networker algorithms
The above model linking TFAs both with gene expression of the regulated genes and external connectivity information has been the subject of a series of recent studies.
In the classic network component analysis approach [9,10] the offset matrix A is set to zero and the remainder of Eq. 1 is interpreted as a dimension reduction that projects the output layer with m samples on to a "hidden" layer of p <m transcription factors. In the original NCA algorithm the coefficients are obtained via a novel matrix decomposition that respects the zero pattern constraint given in the connectivity matrix . Unfortunately, this also imposes rather strict identifiability conditions. As a consequence, classic NCA may only be employed with certain classes of "NCA compatible" [9].
In contrast, the "MA-Networker" algorithm by Gao et al. [14] employs standard multiple least-squares regression in conjunction with step-wise variable selection to estimate the true transcription factor activities . This requires that the number of target genes is much larger than both the number of transcription factors and the number of samples. More important, however, is that the step-wise model selection procedure employed is only poorly suited if the regulator genes are themselves interacting with each other. This is a major drawback as it is biologically well-known that transcription factors often work in conjunction with other regulators, and rarely act independently.
Partial least squares regression
Here we propose to employ the method of partial least squares regression [15] to infer true TFAs and the functional interactions of regulators.
PLS is a well-known analysis tool for high-dimensional data with many continuous response variables that has been widely applied, especially to chemometric data [17]. PLS is particularly suited to the case of non-independent predictors and for small-sample regression settings [16,18-20]. It is computationally highly efficient, it does not necessitate variable selection, and it additionally infers meaningful structural components.
For these reasons PLS is now being adopted as a standard tool for multivariate microarray data analysis, particularly in classification problems [21-24]. We believe that PLS also provides an excellent framework for integrative network analysis, as it combines dimension reduction with regression and variable selection, the two key elements from both the NCA and the MA-Networker approaches.
In a nutshell, the PLS algorithm consists of the following consecutive steps:
1. First, the data matrices and are centered to column mean zero, resulting in matrices X and Y, in order to estimate and to remove the offset A. In addition, it is common practice in PLS analysis (and also recommended here) to scale the input matrices to unit variance.
2. Second, using the linear dimension reduction T = XR, the p predictors in X are mapped onto c ≤ rank(X) ≤ min(p, n) latent components in T (an n × c matrix). See the section "SIMPLS algorithm" below for the precise procedure employed in this paper. The important key idea in PLS is that the weights R (a p × c matrix) are chosen with the response Y explicitly taken into account, so that the predictive performance is maximal even for small c.
3. Next, assuming the model Y = TQ' + E, Y is regressed by ordinary least squares against the latent components T (also known as X-scores) to obtain the loadings Q (a m × c matrix), i.e. Q = Y'T(T'T)-1.
4. Subsequently, the PLS estimate of the coefficients B in Y = XB + E is computed from estimates of the weight matrix R and the Y-loadings Q via B = RQ'.
5. Finally, the coefficients for the original Eq. 1 are computed by rescaling B.
Note that it is step 2 that mostly distinguishes PLS from related bilinear regression approaches such as principal and independent components regression (PCR/ICR) and the pseudo-inverse-based method of Alter and Golub [12]. In the latter approaches the scores T are computed solely on the basis of the data matrix X without considering the response Y [16].
Other quantities often considered in PLS include, e.g., the X-loadings P that are obtained by regressing X against T, i.e. X = TP' + F and P = X'T(T'T)-1.
SIMPLS algorithm
PLS aims to find latent variables T that simultaneously explain both the predictors X and the response Y. The original ideas motivating the PLS decomposition were entirely heuristic. As a result, a broad variety of different, but in terms of predictive power equivalent, PLS algorithms have emerged – for an overview see e.g. Martens [17].
For the present application to infer true TFAs, we suggest using the SIMPLS ("Statistically Inspired Modification of PLS") algorithm, which has the following appealing properties [18-20]:
• it produces orthogonal, i.e. empirically uncorrelated, latent components;
• it allows for a multivariate response; and
• it optimizes a simple statistical criterion.
A further added advantage of SIMPLS is that it is also one of the most computationally efficient PLS algorithms.
We note that other PLS variants described in the literature have predictive power comparable to SIMPLS. However, these either provide orthogonal loadings rather than orthogonal latent components T (Martens' PLS), or they do not elegantly extend from 1-dimensional to m-dimensional responses Y in terms of their optimized objective function (NIPALS).
In SIMPLS, the latent components t1, t2,..., tc of the columns in T are inferred by sequentially estimating the column vectors r1,..., rc of R according to the following criterion [20]:
1. r1 is the unit vector (with |r1| = 1) maximizing the length |Y' Xr1| of the m × 1 covariance vector cov(Y, t1).
2. For all j = 2,...,c, rj are the unit vectors (with |rj| = 1) maximizing the length |Y' Xrj| of the vector cov(Y, tj) subject to the orthogonality constraint for all i = 1,...,j - 1.
In the actual SIMPLS procedure, the weights R and the derived quantities T and Q are obtained by a Gram-Schmidt-type algorithm [18].
On a practical note, we would like to mention that in many implementations of SIMPLS (e.g. in the "pls.pcr" R package by Ron Wehrens, University of Nijmegen), conventions different from the above are used. In particular, the X-scores T* returned will often be orthonormal (rather than orthogonal) and consequently the weights R* will not have unit norm as in our case. For conversion, define M = diag(||,...,||,) and set T = T*M-1, R = R*M-1, Q = Q*M, and P = P*M. This provides orthogonal scores and unit-norm weights as assumed in our description of SIMPLS.
The resulting estimates of the matrices B, T, and R are now straightforward to interpret in terms of transcriptional regulation. B (and ) give the inferred activities of the p transcription factors in each of the m experiments. The inferred latent components T describe "meta"-transcription factors that combine related groups of transcription factors. R reflects the involvement of each of the p regulators in the c meta-factors.
Determining the number of PLS components
A remaining aspect of PLS regression analysis is the optimal choice of the number c of latent components. If the maximal value cmax = rank(X) is chosen, then PLS becomes equivalent to principal components regression (PCR) with the same number of components, and if additionally n >p both PLS and PCR turn into ordinary least-squares multiple regression.
Hence, with PLS it is desirable to choose as small a value of c as possible without sacrificing too much predictive power. One straightforward statistical procedure to estimate this minimum value cmin is the method of cross-validation, which proceeds as follows (cf. also refs. [25] and [26]):
1. Split the set of n genes randomly into 2 sets: a learning set containing 2/3 of the genes and a test set containing the remaining genes.
2. Use the learning set to determine the matrix of regression coefficients B for different values c = 1, 2,...,cmax.
3. Predict the gene expression of the n/3 genes from the test set using B with the different values of c.
4. Repeat steps 1–3 K = 100 times and compute the mean squared prediction error for each c.
Subsequently, the value of c yielding the smallest mean squared prediction error is selected.
Figure 1 Comparison of true (top row) and estimated (bottom row) spectra, as obtained by multivariate PLS regression from the validation data set.
Figure 2 Top row: Mean sum of squared prediction error for E. Coli and yeast data sets over 100 cross-validation runs. Bottom row: maximized objective criterion for each PLS component.
Alternatively, the optimal number of components may also be determined by considering the value of the criterion Zi = |Y'ti| for a given latent component ti. If Zi falls below an a priori specified threshold then cmin = i is reached.
Discussion
Data sets
Next, we illustrate the versatility of the proposed PLS approach to network component analysis by analyzing several real biomolecular data sets.
First, in order to validate the linear regression approach (Eq. 1) we reanalyzed hemoglobin data from Liao et al. [9]. Second, we analyzed two different S. Cerevisiae gene expression data sets in conjunction with a regulator-target connectivity matrix from the large-scale ChIP experiment of Lee et al. [2]. The yeast expression data investigated comprise a time series experiment from Spellman et al. [27] and a compilation of yeast stress response experiments from Gasch et al. [6,28]. Finally, we analyzed expression and connectivity data for an E. Coli regulatory network containing 100 genes and 16 transcription factors from Kao et al. [10]. The general characteristics of these four data sets are summarized in Table 1.
Table 1 Characteristics of the analyzed data sets.
Data Reference n p m cmin
Hemoglobin [9] 7 3 321 3
S. cerevisiae [27] 3638 113 24 5
S. cerevisiae [6, 28] 1993 113 173 8
E. coli [10] 100 16 23 2
Abbreviations: n, number of genes; p, number of transcription factors; m, number of arrays resp. measurements.
The data investigated were preprocessed as follows. The yeast ChIP data set [2] contains protein-DNA interaction data for 6270 genes and 113 transcription factors. It includes missing values that correspond to non-interacting gene-transcription factor pairs. Although ChIP data are essentially continuous, it is common practice to dichotomize them according to the p-values into discrete levels of interaction (0 or 1). In this study, we used data obtained at a p-value threshold of 0.001, as suggested by Lee et al. [2]. However, note that in contrast to the NCA method, dichotomization of the ChIP data is optional in our approach.
The Spellman et al. [27] microarray data originally contained the gene expression of 4289 genes at 24 time points during the cell-cycle. From these genes, a subset of 3638 are also contained in the Lee et al. [2] ChIP data set. Our analysis is based on these 3638 genes. Similarly, the Gasch expression data set [6,28] contains the expression of 2292 genes for 173 arrays corresponding to different stress conditions (e.g. heat shock, amino acid starvation, nitrogen depletion). Of these 2292 genes, a subset of 1993 overlap with the genes considered in the ChIP data.
The connectivity matrix for the E. coli data was compiled mainly by Kao et al. [10] from the RegulonDB [11] database. In addition, they incorporated a few corrections using literature data. The temporal E. coli expression data for 100 genes across 25 time points was also introduced in Kao et al. [10] and is publicly available at .
Validation of the regression approach
The hemoglobin data used in Liao et al. [9] for validation of the classic NCA approach have the advantage that the true coefficients of the network model in Eq. 1 are known, and therefore can be directly compared with the inferred values.
Reanalyzing these data, it is straightforward to show (see Figure 1) that the true regression coefficients can be recovered exactly by multivariate regression (of which PLS is a special case). According to Liao et al. [9], this is also true for classic NCA but not for PCA and ICA interpretations of Eq. 1. This discrepancy can be explained by the fact the neither PCA nor ICA explicitly takes account of the response Y, whereas NCA and PLS do.
PLS components and Y-loadings
Subsequently, we determined the minimum number of PLS components for the yeast and E. coli data sets using cross-validation. The results are plotted in Figure 2 (top) after normalization (the mean cross-validation error with one PLS component is set to one). As can be seen from Figure 2, the minimal mean cross-validation error is obtained with 5 PLS components for the Spellman data, 8 PLS components for the Gasch data and 2 PLS components for the E. coli data. For comparison, the (normalized) objective criterion |Y'ti| of the SIMPLS algorithm is also represented on Figure 2 (bottom) for different numbers of PLS components. These results are in good agreement with the cross-validation error: it increases when PLS components with a low objective criterion are added.
The Y-loadings contained in the m × c matrix Q give the projection of the c "meta"-transcription factors for each of the m experiments. As can be seen from Figure 3 for the Spellman data, both the first and the third meta-factors explain the periodic part of the expression data, but with different phases. The second meta-factor corresponds to small oscillations with very short period, whereas the fourth and fifth meta-factors reflect long-time trends (slow and step-wise increasing, respectively). Using Fisher's g-test as proposed in Wichert et al. [29], we detected statistically relevant periodicity for the four first meta-factors. In Figure 3, the Y-loadings are also represented for the E. coli data. Whereas the projection of the first meta-factor is approximately constant over time, the projection of the second meta-factor increases strongly and (almost) uniformly. Thus, in both data sets, the PLS algorithm allows us to extract meta-factors from the data corresponding to distinct latent trends.
Figure 3 Y-loadings for the E. Coli (top and middle row) and Spellman (bottom row) data sets.
For the Gasch data, the m experiments do not correspond to different time points but to 13 different stress conditions (see Gasch et al. [28] for further details, and Table 2 for the list of conditions). In this case the Y-loadings may be interestingly analyzed using Wilcoxon's rank sum test. For each condition k and each meta-factor j, we tested the H0 hypothesis that the median of the projection of the j-th meta-factor is the same in condition k as in all the other conditions ({1,..., k - 1, k + 1,..., 13}). In this situation, Wilcoxon's rank sum test is preferable to the well-known two-sample t-test, because some of the conditions include only a very small number of experiments. The results obtained with a p-value threshold of 0.05 are displayed in Table 2. The entries 1 and 0 correspond to significant and non-significant (FDR adjusted) p-values, respectively. As can be seen from Table 2, each PLS component carries a particular pattern of associated significant conditions, indicating that the meta-factors capture a distinct direction of the data.
Table 2 Significant conditions for the first 8 PLS components of the Gasch yeast data set.
Condition \ PLS Component 1 2 3 4 5 6 7 8 Arrays
Heat shock 0 0 0 0 0 0 0 0 1–9,12–15
Variable temperature shocks 0 0 1 0 1 0 0 0 21–25
Hydrogen peroxide 0 0 0 0 0 1 0 0 36–45
Menadione 0 1 0 0 1 1 0 0 46–54
DTT 0 0 0 0 0 0 0 0 55–69
Diamide 1 1 1 0 0 0 1 1 70–77
Sorbitol osmotic shock 0 0 0 0 0 0 0 0 78–89
Amino acid starvation 0 0 1 1 1 0 1 1 91–95
Nitrogen depletion 0 0 1 0 0 1 1 1 96–105
Diauxic shift 0 0 1 0 0 0 1 0 106–112
Stationary phase 1 1 0 1 1 1 1 0 113–134
Continuous carbon sources 1 0 0 0 0 1 0 1 148–160
Continuous temperatures 1 0 0 0 0 0 1 0 161–173
Inferred transcription factor activities
One of the main objectives of our PLS-based approach is to estimate the true transcription factor activities (TFAs). Although all the TFAs can be estimated in the same way for the three data sets, we display only the evolution over time of a few interesting TFAs for the two time series data sets (i.e. the Spellman and the E. coli data).
The TFAs (top) and expression profiles (bottom) of 4 well-known cell-cycle regulators are depicted in Figure 4 for the Spellman data. The TFAs of MCM1, SWI4, SWI5 and ACE2 show highly periodic patterns, which is consistent with common biological knowledge. In contrast, the expression profiles of MCM1 and SWI4 are not periodic (this can be confirmed by Fisher's g-test [29]). On the other hand, the expression profiles of SWI5 and ACE2 are periodic, though not with the same phase as the inferred TFAs. This may indicate either an inhibiting or a phase-shift effect of the transcription factors on the regulated genes.
Figure 4 Time profiles of the TFAs (top row) of four well-known cell-cycle transcription factors from the Spellman data compared to the respective gene expression measurements (bottom row).
The remainder of the TFAs and the regulated genes were also tested for periodicity using the g-test [29]. After FDR adjustment of the p-values, we found that 62 of the 113 transcription factors (= 55%) in the Spellman/Lee data have significantly periodic TFAs at the level 0.05. In contrast, only 804 of the 4289 genes (= 19%) exhibit significantly periodic expression profiles.
For the E. coli data the time profiles of the estimated TFAs of the 16 transcription factors are represented in Figure 5. The TFAs of ArcA, GatR, Lrp, PhoB, PurR, RpoS decrease over time, those of CRP, CysB, FadR, IcIR, NarL, RpoE, TrpR and TyrR remain approximately constant and those of FruR and LeuO increase strongly. This is consistent with previous results obtained by NCA [10]. We point out, however, that unlike NCA our approach may be applied to any arbitrary network topology, whereas the present E. coli network was chosen specifically to meet the NCA compatibility criteria [9].
Figure 5 Time profiles of the 16 estimated TFAs (E. Coli data).
As can be seen already from the few examples depicted in Figure 4, the TFAs do not always correlate with the respective expression profiles. We tested this for all the transcription factors of which the expression profiles were also included in the data sets. For the Gasch data, we found that only 63 from the 90 available transcription factors exhibit expression profiles that are correlated with TFAs (at the level 0.05 with FDR p-value adjustment). For the Spellman time series data, none of the 78 available TFA-expression profile pairs are correlated. These results clearly indicate that methods investigating transcriptional regulation with expression data as their sole basis are likely to miss potentially important regulation activities.
Gene-regulator coupling factors
Another topic of interest is the identification of false positives in ChIP data. Following Gao et al. [14] we investigate this problem using Pearson's correlation test. For each supposed gene-transcription factor pair (according to the dichotomized ChIP data) we test if the inferred TFA is significantly correlated with the expression profile of the regulated gene. For the Gasch data, we find that 73% of the 1495 gene-transcription factor pairs are correct (i.e. the TFA is significantly correlated with the expression profile at the 0.05 level with FDR p-value adjustment). The concordance with the ChIP connectivity information is much worse for the Spellman data, where only 32% of the 2535 gene-transcription factor pairs are significantly correlated.
We should like to add as a note of caution that the lack of correlation between TFA and target gene needs to be viewed as specific to the microarray study investigated. Other expression experiments may activate different pathways and thus produce different patterns of correlation in conjunction with the ChIP connectivity information.
Conclusion
Network component analysis combines microarray data with ChIP data with the aim of enhancing the estimation of regulator activities and of connectivity strengths. In this paper we have presented an approach to NCA based on partial least squares, a computationally efficient statistical regression tool.
Our PLS framework allows several drawbacks, inherent both in the classic NCA methods based on matrix decomposition and in the MA-Networker algorithm, to be overcome. Its simplicity (no iterative step, no variable selection, no stochastic search) and its flexibility (no distributional assumptions, no topological constraints, no conditions on the dimensions) compared to competing approaches make it particularly attractive as an integrative method for analyzing complex regulatory networks. Moreover, the PLS algorithm not only extracts information on gene-regulator and on TFA-expression profile pairs but also identifies coherent meta-factors reflecting the main directions of variation of the data, taking account both of the expression () and the connectivity information ().
Our analysis of biological data shows the versatility of our PLS approach and at the same time dramatically confirms the need for a combined expression-ChIP analysis for inferring regulation. Particularly striking are the sometimes drastic differences between the measured transcription levels and the PLS-inferred transcription activities. According to Segal et al. [6], some transcription factors may also not be active in all conditions. Note that this assumption is also automatically taken into account by our approach.
NCA in general, and the present PLS-based variant in particular, may be criticized for relying on a simple linear model - see Buchler et al. [30] and Setty et al. [31] for counter-examples. Therefore, more elaborate regression approaches such as generalized linear models (GLMs) or generalized additive models (GAMs) may be required to further enhance our current understanding of how best to model the complex structures governing genetic networks.
[Note added in proof: See Yang et al. [32] for a related study in the sister journal BMC Genomics.]
Authors' contributions
A.-L.B. performed all the data analysis and simulations. Both authors jointly developed the methodology, wrote the manuscript, and approved of the final version.
Appendix: Computer program
All algorithms have been implemented in the R language [33]. A corresponding R package "plsgenomics" developed by the authors is available for download from the CRAN archive .
Acknowledgements
We thank Eran Segal and James Liao for kindly providing the Saccharomyces cerevisiae data and the hemoglobin data, respectively. We also thank the anonymous referees for useful comments. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) through an Emmy-Noether research grant to K.S. and the Sonderforschungsbereich 386.
==== Refs
Ren B Robert F Wyrick JJ Aparicio O Jennings EG Simon I Zeitlinger J Schreiber J Hannett N Kanin E Volkert TL Wilson CJ Bell SP Young RA Genome-wide location and function of DNA binding proteins Science 2000 290 2306 2309 11125145 10.1126/science.290.5500.2306
Lee TI Rinaldi NJ Robert F Odom DT Bar-Joseph Z Gerber GK Hannett NM Harbison CT Thompson CM Simon I Zeitlinger J Jennings EG Murray HL Gordon DB Ren B Wyrick JJ Tagne JB Volkert TL Fraenkel E Gifford DK Young RA Transcriptional Regulatory Networks in Saccharomyces cerevisiae Science 2002 298 799 804 12399584 10.1126/science.1075090
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
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
Harbison CT Gordon DB Lee TI Rinaldi NJ Macisaac KD Danford TW Hannett NM Tagne JB Reynolds DB Yoo J Jennings EG Zeitlinger J Pokholok DK Kellis M Rolfe PA Takusagawa KT Lander ES Gifford DK Fraenkel E Young RA Transcriptional regulatory code of a eukaryotic genome Nature 2004 431 99 104 15343339 10.1038/nature02800
Segal E Shapira M Regev A Pe'er D Botstein D Koller D Friedman N Module networks: identifing regulatory modules and their condition-specific regulators from gene expression data Nature Genetics 2003 34 166 176 12740579
Xiong M Li J Fang X Identification of genetic networks Genetics 2004 166 1037 1052 15020486 10.1534/genetics.166.2.1037
Li Z Chan C Extracting novel information from gene expression data Trends Biotechnol 2004 22 381 383 15283978 10.1016/j.tibtech.2004.06.007
Liao JC Boscolo R Yang YL Tran LM Sabatti C Roychowdhury VP Network component analysis: reconstruction of regulatory signals in biological systems Proc Natl Acad Sci USA 2003 100 15522 15527 14673099 10.1073/pnas.2136632100
Kao KC Yang YL Boscolo R Sabatti C Roychowdhury V Liao JC Transcriptome-based determination of multiple transcription regulator activities in Escherichia coli by using network component analysis Proc Natl Acad Sci USA 2004 101 641 646 14694202 10.1073/pnas.0305287101
Salgado H Santos-Zavaleta A Gama-Castro S Millan-Zarate D Diaz-Peredo E Sanchez-Solano F Perez-Rueda E Bonavides-Martinez C Collado-Vides J RegulonDB (version 3.2): transcriptional regulation and operon organization in Escherichia coli K-12 Nucleic Acids Res 2001 29 72 74 11125053 10.1093/nar/29.1.72
Alter O Golub GH Integrative analysis of genome-scale data by using pseudoinverse projection predicts novel correlation between DNA replication and RNA transcription Proc Natl Acad Sci USA 2004 101 16577 16582 15545604 10.1073/pnas.0406767101
Kato M Hata N Banerjee N Futcher B Zhang MQ Identifying combinatorial regulation of transcription factors and binding motifs Genome Biology 2004 5 R56 15287978 10.1186/gb-2004-5-8-r56
Gao F Foat BC Bussemaker HJ Defining transcriptional networks through integrative modeling of mRNA expression and trasncription factor binding data BMC Bioinformatics 2004 5 31 15113405 10.1186/1471-2105-5-31
Wold S Martens H Wold H Ruhe A, Kagstrom B The multivariate calibration method in chemistry solved by the PLS method Proc Conf Matrix Pencils, Lecture Notes in Mathematics 1983 Heidelberg: Springer Verlag 286 293
Frank IE Friedman JH A statistical view of some chemometrics regression tools Technometrics 1993 35 109 135
Martens H Reliable and relevant modelling of real world data: a personal account of the development of PLS regression Chemom Intell Lab Syst 2001 58 85 95 10.1016/S0169-7439(01)00153-8
de Jong S SIMPLS: An alternative approach to partial least squares regression Chemom Intell Lab Syst 1993 18 251 253 10.1016/0169-7439(93)85002-X
Braak CJFT de Jong S The objective function of partial least squares regression J Chemometrics 1998 12 41 54 10.1002/(SICI)1099-128X(199801/02)12:1<41::AID-CEM500>3.0.CO;2-F
de Jong S Wise BM Ricker NL Canonical partial least squares and continuum power regression J Chemometrics 2001 15 85 100 10.1002/1099-128X(200102)15:2<85::AID-CEM601>3.0.CO;2-9
Datta S Exploring relationships in gene expressions: a partial least squares approach Gene Expression 2001 9 249 255 11763996
Nguyen D Rocke DM Tumor classification by partial least squares using microarray gene expression data Bioinformatics 2002 18 39 50 11836210 10.1093/bioinformatics/18.1.39
Nguyen D Rocke DM Partial least squares proportional hazard regression for application to DNA microarray survival data Bioinformatics 2002 18 1625 1632 12490447 10.1093/bioinformatics/18.12.1625
Boulesteix AL PLS dimension reduction for classification with microarray data SAGMB 2004 3 33
Dudoit S Fridlyand J Speed TP Comparison of discrimination methods for the classification of tumors using gene expression data J Amer Statist Assoc 2002 97 77 87 10.1198/016214502753479248
Braga-Neto U Dougherty ER Is cross-validation valid for small-sample microarray classification? Bioinformatics 2004 20 374 380 14960464 10.1093/bioinformatics/btg419
Spellman PT Sherlock G Zhang MQ Iyer VR Anders K Eisen MB Brown PO Botstein D Futcher B Comprehensive identification of cell cycle-regulated genes of the yeast Saccharomyces cerevisiae by microarray hybridization Mol Biol Cell 1998 9 3273 3297 9843569
Gasch AP Spellman PT Kao CM Carmel-Harel O Eisen MB Storz G Botstein D Brown PO Genomic expression programs in the response of yeast cells to environmental changes Mol Biol Cell 2000 11 4241 4257 11102521
Wichert S Fokianos K Strimmer K Indentifying periodically expressed transcripts in microarray time series data Bioinformatics 2004 20 5 20 14693803 10.1093/bioinformatics/btg364
Buchler NE Gerland U Hwa T On schemes of combinatorial transcription logic Proc Natl Acad Sci USA 2003 100 5136 5141 12702751 10.1073/pnas.0930314100
Setty Y Mayo AE Surette MG Alon U Detailed map of a cis-regulatory input function Proc Natl Acad Sci USA 2003 100 7702 7707 12805558 10.1073/pnas.1230759100
Yang YL Suen J Brynildsen M Galbraith S Liao JC Inferring yeast cell cycle regulators and interactions using transcription factor activities BMC Genomics 2005 6 90 15949038 10.1186/1471-2164-6-90
R Development Core Team R: A language and environment for statistical computing 2004 R Foundation for Statistical Computing, Vienna, Austria [ISBN 3-900051-07-0].
|
15978125
|
PMC1182396
|
CC BY
|
2021-01-04 16:39:26
|
no
|
Theor Biol Med Model. 2005 Jun 24; 2:23
|
utf-8
|
Theor Biol Med Model
| 2,005 |
10.1186/1742-4682-2-23
|
oa_comm
|
==== Front
J Transl MedJournal of Translational Medicine1479-5876BioMed Central London 1479-5876-3-251596974810.1186/1479-5876-3-25ReviewOvarian cancer, the coagulation pathway, and inflammation Wang Xipeng [email protected] Ena [email protected] John J [email protected] Ralph S [email protected] Department of Gynecologic Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA2 Department of Transfusion Medicine, National Institutes of Health, Bethesda, MD, USA3 Department of Gynecologic Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA2005 21 6 2005 3 25 25 10 3 2005 21 6 2005 Copyright © 2005 Wang et al; licensee BioMed Central Ltd.2005Wang 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.
Epithelial ovarian cancer (EOC) represents the most frequent cause of death in the United States from a cancer involving the female genital tract. Contributing to the overall poor outcome in EOC patients, are the metastases to the peritoneum and stroma that are common in this cancer. In one study, cDNA microarray analysis was performed on fresh tissue to profile gene expression in patients with EOC. This study showed a number of genes with significantly altered expression in the pelvic peritoneum and stroma, and in the vicinity of EOC implants. These genes included those encoding coagulation factors and regulatory proteins in the coagulation cascade and genes encoding proteins associated with inflammatory responses. In addition to promoting the formation of blood clots, coagulation factors exhibit many other biologic functions as well as tumorigenic functions, the later including tumor cell proliferation, angiogenesis, invasion, and metastasis. Coagulation pathway proteins involved in tumorigenesis consist of factor II (thrombin), thrombin receptor (protease-activated receptors), factor III (tissue factor), factor VII, factor X and factor I (fibrinogen), and fibrin and factor XIII. In a recent study we conducted, we found that factor XII, factor XI, and several coagulation regulatory proteins, including heparin cofactor-II and epithelial protein C receptor (EPCR), were also upregulated in the peritoneum of EOC.
In this review, we summarize evidence in support of a role for these factors in promoting tumor cell progression and the formation of ascites. We also discuss the different roles of coagulation factor pathways in the tumor and peritumoral microenvironments as they relate to angiogenesis, proliferation, invasion, and metastasis. . Since inflammatory responses are another characteristic of the peritoneum in EOC, we also discuss the linkage between the coagulation cascade and the cytokines/chemokines involved in inflammation. Interleukin-8, which is considered an important chemokine associated with tumor progression, appears to be a linkage point for coagulation and inflammation in malignancy. Lastly, we review findings regarding the inflammatory process yielded by certain clinical trials of agents that target members of the coagulation cascade in the treatment of cancer. Current data suggest that disrupting certain elements of the coagulation and inflammation processes in the tumor microenvironment could be a new biologic approach to cancer therapeutics.
==== Body
Introduction
An estimated 22,220 new cases of and 16,210 deaths from epithelial ovarian cancer (EOC) will occur in 2005 [1]. EOC is insidious: 75% of patients present at an advanced stage, in which metastasis to the peritoneal cavity lining is characteristic. Peritoneal and serosal involvement is associated overall with a poor patient outcome. The mechanism of peritoneal seeding, spreading, and progression is poorly understood. Direct contact of free-floating ovarian tumor cells in the peritoneal cavity may place the peritoneum at risk for metastatic spread. Other evidence of metastatic spread points to a surface epithelium origin of EOC [2] or genetic instability of the stroma [3].
Evidence is mounting that an inflammatory process contributes to tumor growth and metastasis to the peritoneum in EOC [4,5]. This process is facilitated by cross talk between tumor cells and the surrounding cellular stroma [6]. Surgical findings from EOC frequently include enhanced tumor vascularity with or without malignant ascites. Microscopic examination may reveal an inflammatory infiltrate comprising different leukocyte populations [6]. The stroma associated with tumors may resemble granulation tissue formed during wound healing [7]. The interaction of cancer cells with their surroundings might influence phenotypic alterations in the tumor. For example, the presence of human breast tumor stromal explants in mice may determine the formation of breast cancer-like lesions [8].
In a recent study, we compared the transcriptional profile of normal-appearing peritoneum and its adherent and subjacent stroma in EOC patients with that of patients with pathologically benign ovarian disease [3]. In our study, we identified 402 genes differentially expressed between malignant and benign peritoneum and 663 genes differentially expressed between malignant and benign stroma. No significant differences in gene expression between malignant peritoneum and malignant stroma were observed, which suggests that the transcriptional changes induced by the malignant condition overlapped at least partly. Our preliminary analysis indicated that the gene profile of peritoneal structures might be affected by the presence of EOC cells [3]. We found it interesting that the genes included those linked to inflammatory processes and a number were linked to the coagulation pathway. An examination of these two processes at the gene transcript level suggests a number of possible links between coagulation and inflammation. In this paper, we discuss the different roles of coagulation factor pathways as they relate to angiogenesis, proliferation, invasion, and metastasis in the tumor and peritumoral microenvironment.
Coagulation and cancer development
The history of a known association between coagulation and cancer dates back to 1865, when Armand Trousseau observed that patients who presented with idiopathic venous thromboembolism frequently had an underlying occult cancer and vice versa [9]. The association has only recently become more apparent. Previously, researchers focused on how tumors activate blood coagulation and how to overcome it. However, the underlying mechanism by which coagulation factors promote tumor cell growth, invasion, metastasis, and angiogenesis has become a hot topic in the field of cancer research. Some coagulation factors that displaya role in tumor progression have been reviewed [10]. The most frequent reports on coagulant proteins and cancer interactions include factor III (tissue factor; TF) [11], TF-factor VIIa [12,13], factor Xa [14], factor IIa (thrombin)-factorII receptors (also called proease-activated receptors (PARs) [15], and factor XIIIa-factor Ia (fibrin) (Table 1) [16]. Two cascades, one intrinsic and the other extrinsic, lead to the formation of a fibrin clot. Although they are initiated by distinct events, the two cascades converge on a common pathway that includes thrombin, PARs, and fibrin and lead to clot formation. The intrinsic cascade, which is initiated when contact is made between blood and exposed endothelial cell (EC) surface, requires the clotting factors VIII, IX, XI, and XII. Also required are the proteins prekallikrein and high-molecular-weight kininogen and Ca2+ and phospholipids secreted from platelets. The extrinsic coagulation cascade is initiated at the site of injury in response to the release of TF. Activated factor X is the site at which the intrinsic and extrinsic coagulation cascades converge. TF and factor VIIa contribute to the extrinsic cascade and possibly to the development of cancer. Other factors from the intrinsic pathway, such as factors XI and XII, have not yet been directly implicated in cancer progression. Our recent analysis of the gene expression profile in EOC has revealed increased levels of factor XI and factor XII transcripts in the peritoneum of EOC patients [3]. These findings suggest that the intrinsic cascade might promote ovarian cancer cell metastasis in the peritoneal cavity. We also showed that the serine proteinase inhibitor D1 (SERPIND1), which inhibits plasmin, tissue kallikrein, and factor XIa, were downregulated [3].
Table 1 Coagulation factors and the associated regulatory proteins in EOC
Coagulation factors Effects
Tissue factor (TF) ↑ Promoting angiogenesis by activation of MAPK[73] and protein C kinase C-dependent signaling[76]; TF-PAR2 selectively synergizes with PDGF-BB to enhance to metastasis in lymphnodes[78].
Promoting invasion and metastasis by the activation of P21Ras and P42/P44 MAPK pathway to inhibit apoptosis[14]; overexpressing growth factors and chemokines (i.e. IL-8)[13].
TF-VII-PAR2 ↑ Promoting angiogenesis, invasion, and metastasis by clotting-independent mechanism[77] in presence of inflammatory cytokines
Factor X ↑ Forming complex with TF-VIIa to promote tumor angiogenesis and metastasis[14]
Thrombin/PAR1 ↑ Promoting angiogenesis by inhibiting EC migration to collagen type IV or to laminin[99]; upregulating VEGF expression[100].
Promoting invasion and metastasis depended on at least 6 mechanisms (text)
Fibrinogen/fibrin ↑ Stimulating angiogenesis; the fibrin gel matrix facilitating tumor metastasis; increasing plasma exudates to form ascites[129, 130, 132].
Factor XII/XI ↑ Positive feedback on human kallikreins system
Factor XIII ↑ Form stable fibrin
Regulatory proteins
Heparin cofactor II ↑ Produce chemoattractant peptide for MAs migration.
Endothelial protein C receptor ↑ Intensifying APC-PAR1 signal transduction [205] and contributing to antiapoptosis in tumor.
Tissue factor pathway inhibitor ↓ Loss of control of tumor growth and metastasis by activating Factor Xa and increasing Factor Xa-PAR2 signaling[81].
Tissue factor pathway inhibitor-2 ↓ Loss of Inhibiting TF-VIIa complex and various protease but not Factor Xa;
Loss of antiangiogenesis and antimetastasis.
Blood coagulation cascades can be activated by different mechanisms and to different levels in cancer patients. The alterations range from subtle abnormalities in laboratory tests to clinically overt thrombosis and disseminated intravascular coagulation [17]. Up to 50% of all cancer patients and 90% of those with metastases exhibit hemostatic abnormalities [17]. These abnormalities may be reflected in the dominance of the tumor cell-associated procoagulant pathway, which leads to thrombin generation and hypercoagulation. Similar observations were made using in vitro ovarian cancer cells for the coagulation process. Ovarian cancer cells may also express TFs and the other coagulation components that generate local thrombin, as indicated by the conversion of fibrinogen to fibrin [18]. Fibrin has been found on ECs and the surface of ovarian tumor cells and nodules [18]. Coagulation activation in malignancy may be triggered by direct or indirect mechanisms. Direct activation of blood coagulation by the induction of thrombin may occur through the activity of tumor cell procoagulation, whereas indirect activation may occur through the production of tumor-associated cytokines that trigger TF production by host macrophages (MAs) or ECs. The coagulation pathway components may contribute to tumor cell proliferation, invasion, and metastasis [10], although these alterations could also be a consequence of advanced disease [19].
Factor XII and the kallikrein family (positive feedback loop)
Factor XII is a procoagulant protein that participates directly or indirectly in activating the intrinsic clotting pathway. The structure of factor XII, which has been inferred from its DNA and amino acid sequences, include two epidermal growth factor (EGF) homologous domains in the amino-terminal region [20]. This type of structure suggests that factor XII might mimic EGF biological characteristics and act as a growth factor. One study showed that factor XII had mitogenic effects on the HepG2 human hepatocellular carcinoma cell line and, similar to EGF, stimulated tumor cell proliferation [20].
Assembly of contact phase components results in the conversion of prekallikrein to kallikrein, which in turn activates factor XII to factor XIIa. Factor XIIa then hydrolyzes more prekallikrein to kallikrein, thereby establishing a reciprocal activation loop. Factor XIIa also activates factor XI to factor XIa and leads to the release of bradykinin, a potent vasodilator, from high-molecular-weight kininogen. The human kallikreins (hK) are a subfamily of the serine protease enzyme family [21], which consists of proteolytic enzymes important to various physiologic processes, such as digestion, coagulation, fibrinolysis, apoptosis, cell migration, tissue remodeling, and inflammation [21]. The hK gene family, which includes 15 members (hK1-hK15) clustered in a 300-kb region on chromosome 19q13.4, could be altered in cancer. A number of studies have reported increased amounts of kallikrein transcripts or proteins in cancer cells, particularly adenocarcinomas derived from steroid hormone-regulated tissues [21]. At least 11 kallikrein genes or proteins have been found to be overexpressed in EOC; they are hK4 [22], hK5 [23,24], hK6 [25,26], hK7 [27], hK8 [28,29], hK9 [30], hK10 [31-33], hK11 [34-37], hK13 [38,39], hK14 [40,41], and hK15 [42] mRNA or protein in ovarian cancer tissue, cell lines, serum, and tumor ascites fluid (Table 2). A recent gene microarray analysis on ovarian cancer tissue found overexpression of hK2 and hK3 [43]. Overexpression of hK8 [28,29], hK9 [30], hK11 [34,35], hK13 [38,39], and hK14 [41] have been reported as independent variables associated with longer progression-free and overall survival. Another study found hK11 to be an unfavorable factor for EOC [36]. The other hKs have all been linked to progression in EOC (Table 2). Although many hKs are overexpressed in EOC tissues and in ascites, it is unknown whether this phenomenon reflects increased proteolytic activity because of the heterogeneity of hK forms and the presence of active enzymes found in the extracelluar matrix (ECM). This apparent paradox in which the clinical outcome differs with different hKs might be due to different biologic roles of hKs in the tumor progression process, either stimulating or inhibiting the development of the tumor and its microenvironment. hK-mediated pericellular proteolysis in the ECM might help regulate tumor cell growth, angiogenesis, invasion, and metastasis, which could contribute to tumor progression [21]. Some hKs are under steroid hormone regulation and might represent downstream targets through which hormones affect the initiation or progression of EOC.
Table 2 Expression and clinical features of hK members in ovarian cancer
hK family member Location Expression level and site Clinical feature Prognosis
hK4 Tumor cells Increased in EOC tissue Predictive marker for paclitaxel resistance[22] Unfavorable
hK5 Serum, ascites, and tumor extracts [23, 24] Increased on tumor cells and in serum and ascites Potential biomarker for diagnosis Unfavorable
hK6 Tumor cell[26] and serum[25] High in early-stage and low-grade tumor tissue and in EOC serum Overexpression is an early phenomenon in the development of ECO; serum level could be used as a biomarker Unfavorable
hK7 Tumor tissue[27] High in late-stage EOC A potential biomarker for diagnosis Unfavorable
hK8 Tumor extract, serum, and ascites[29] High in serum and ascites High level is associated with good prognosis[28, 29] Favorable
hK9 Tumor cells[30] High in tumor tissue early-stage and optimal debulking patients Associated with longer progression-free and overall survival times Favorable
hK10 Serum[33] and tumor cells[31,32] High in serum and on tumor cells High serum level is associated with increased risk for relapse and death; a potential biomarker for diagnosis Unfavorable
hK11 Serum, ascites[37], and tumor extract [34-36] Increased in tumor samples High level is an independent factor for favorable prognosis and is associated with long progression-free and overall survival times and with slower disease progression[34, 35]; however, high level is also associated with poor survival rate[36]. Depends
hK13 Tumor tissue[38] verexpressed in tumor tissue Associated with longer progression-free and overall survival times Favorable
hK14 Tumor cells[41] and serum[40] Increased in tumor tissue and serum Associated with longer progression-free and overall survival times; an independent prognostic factor Favorable
hK15 Tumor extract[42] Increased in tumor tissue Associated with short progression-free and overall survival times; an independent prognostic factor Unfavorable
Other data have shown that hK5-hK7 can contribute to ECM remodeling through fibrinogen, collagen types I and IV, laminin, and fibronectin and, in this fashion, facilitate tumor angiogenesis, invasion, and metastasis [21]. Kallikreins might also promote angiogenesis by disrupting ECM barriers. hK2 [44], hK3 [45], hK6 [46], hK7 [27], and hK14 [41] directly catalyze the hydrolysis of certain ECM proteins, enabling both EC and tumor cell migration and invasion. The ECM may also be remodeled through activation of the uPA-PAR system [21]. For example, hK2 and hK4 may remodel certain ECM components through plasmin and release or activate certain pro-angiogenic growth factors such as vascular endothelial growth factors (VEGFs) and pro-matrix metalloproteins (MMPs) [47]. It could be assumed that factor XII and kallikrein family members might also participate together in the formation of ascites and peritoneal implants in EOC. Furthermore, hK1 is expressed in ECs and hK1-generated kinins, which are multifunctional, biologically active peptides released from low-molecular-weight kininogen that can stimulate angiogenesis [48]. The remodeling effect of hKs on ECs has been demonstrated in Matrigel in vitro invasion assays. For example, invasion of MDA-MB-231 human breast cancer cells into Matrigel was suppressed by a synthetic hK1 inhibitor [49], and invasion of LNCaP human prostate cancer cells through Matrigel was attenuated by hK3-neutralizing antibodies [45]. Conversely, hK3 has been shown to have antimetastatic properties in mice [45]. The antimetastatic effect might be due to an antiangiogenic effect, which is independent of serine protease action. hKs might also be able to promote cancer cell invasion through PAR signaling in a manner similar to thrombin [50].
According to a study conducted by Rehault, hK2 and hK3 could represent important regulators of insulin-like growth factors (IGFs) in the proliferation of prostate cancer cells [51]; however, similar findings have not been published on EOC. IGFs are mitogenic peptides that help regulate normal and malignant cellular proliferation, differentiation, apoptosis, and transformation. IGFs have to be released from IGF-binding proteins (IGFBPs) before their activation. hK1, hK2, and hK3 are IGFBP proteases that collectively degrade IGFBPs 2–5 and could abrogate their affinity for IGF1. In our microarray analysis of the peritoneum of EOC patients, we detected overexpression of IGF1 and decreased expression of SPINT2. These findings are important because SPINT2 inhibits the production of tissue kallikreins [3]; therefore, it could be deduced that a high expression of hKs might result. In the meantime, IGF1 can stimulate the growth of normal, stromal, and malignant prostate cells in vitro [51,52]. The IGF-IGFBP systems can operate in many organs, suggesting roles for them in cancer growth [53].
The blood coagulation and kallikrein-kinin systems consist of plasmatic proteolytic cascades. These cascades participate in the initial host response against foreign components and in wound healing. The increased expression of factor XII in the peritoneum tissue of EOC patients [3] suggests that the intrinsic coagulation cascade is being activated and may be involved in ovarian carcinoma cell progression. At least two mechanisms may be responsible for activating the cascade: the molecular structure of factor XII mimics the EGF motif and, thus, may have a growth factor-like role on cancer cells, and because of the positive feedback loop, the kallikrein system is capable of promoting ovarian cancer cell growth, angiogenesis, metastasis, and invasion.
Factor XI
Factor XI is a component of the intrinsic coagulation pathway and is activated after factor XII activation. Factor XI is the zymogene (i.e., the gene associated with a proteolytic enzyme) precursor of a plasma serine protease produced primarily in the liver that contributes to blood coagulation by activating factor IX by limited proteolysis [54]. Factor XI mRNA has been detected in human platelets [55], but a link between factor XI and the stimulation or inhibition of cancer cell proliferation and metastasis has not been established. In our microarray analysis of the peritoneum of EOC patients, we found a higher expression of factor XI mRNA in peritoneum and stroma tissue in the general vicinity of ovarian cancer implants [3]. Because factor XII expression also increased, factor XI might be induced by the activation of factor XII. In addition, factor XI might have an indirect role in the pathogenesis of cancer by facilitating downstream factors in the coagulation cascade, such as thrombin and its receptors. These are discussed later.
Tissue factor
There is abundant evidence that TF plays an important role in the pathogenesis of cancer. Human TF molecule is a single-chain, 263-amino-acid, 47-kDa transmembrane glycoprotein whose primary sequence indicates structural similarity to members of the cytokine receptor family [56]. TF is the principal surface receptor and cofactor for the activated coagulation protease factor VIIa and is a receptor for its zymogen precursor. The extrinsic coagulation cascade is triggered by the binding of factor VIIa to TF, which creates a complex for the activation of the protease factor X and its conversion to factor Xa. Generation of factor Xa and the TF-factor VIIa complex triggers the proteolytic conversion of factor II (prothrombin) to thrombin.
TF is expressed constitutively on the adventitia of uninjured blood vessels and other extravascular tissues [57]. Upregulation of TF gene expression occurs in malignant cells and in normal host cells that respond to inflammatory or remodeling signals, which might arise from ECs, MAs, and fibroblasts [58]. Therefore, cytokines and growth factors, such as platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), transforming growth factor-β (TGF-β), and EGF [58], produced by inflammatory and malignant cells might induce the expression of TF in fibroblasts and ECs. Experimental evidence has also shown that expression of TF by ECs can be under the control of VEGF, which can be mediated by the VEGF receptor fms-like tyrosine kinase (flt-1) and flt-1/kinase insert domain-containing receptor [59]. Two downstream components of the coagulation cascade, thrombin [60] and fibrin [61], can also regulate TF expression by ECs. Both downregulation and inhibition of TF expression can modulate tumor cell procoagulant activity. Recent evidence suggests that TF expression might alter the cancer cell phenotype and possibly contribute to angiogenesis, proliferation, and metastasis [62-69]. Aberrant TF expression has been detected in various human tumors, including EOC [62-64], breast cancer [65], non-small cell lung cancer [66], colon cancer [67], and pancreatic cancer [68] but is not usually found in normal tissues from these sites. Elevated expression of TF in tumors has been associated with certain unfavorable prognostic indicators, such as angiogenesis, metastasis, advanced disease stage, and multidrug resistance [69].
TF and angiogenesis
The formation of new cellular masses at an organ site creates an obvious extra demand for oxygen, growth factors, and metabolites and increases the need for tumor-associated blood vessels. The stimulators and inhibitors of angiogenesis that can regulate new vessel formation include VEGF, acidic and basic FGF, hepatocyte growth factor, and TGF. Angiostatin, endostatin, and antiangiogenic antithrombin might function as inhibitors of angiogenesis [70].
Angiogenesis can be mediated by either coagulation-dependent or-independent activation pathways [71]. Coagulation-dependent pathways always involve activation of the TF receptor to its ligand, followed by production of thrombin and ensuing clot formation. Coagulation-independent pathways appear to involve phosphorylation of the cytoplasmic domain of the TF receptor and subsequent downstream signaling events that occur independently of thrombin production or clot formation. Both pathways can contribute to angiogenesis and tumor progression. Increased expression of TF in tumors may contribute to angiogenesis in part by increasing VEGF protein expression and downregulating the expression of the antiangiogenic protein thrombospondin [11]. For example, TF-positive colorectal tumors have higher levels of microvessel density and VEGF expression than TF-negative colorectal tumors do [67]. Colocalization of TF and VEGF mRNA and protein has been demonstrated in breast cancer, malignant glioma, and adenocarcinoma of the lung [72]. In the coagulation-dependent pathway, the interaction of TF and factor VIIa induces Ca2+ oscillations and changes in gene expression, and the formation of the TF-factor VIIa complex leads to the activation of the mitogen-activated protein kinase (MAPK) pathway [73], which is a major inducer of VEGF expression. The catalytic activity of the TF-factor VIIa complex contributes to the activation of factor X to factor Xa and subsequently to thrombin. Factor Xa and thrombin) can also exert signaling activities through PAR-2 and PAR-1, respectively. Factor Xa activated by tumor cells may trigger PAR-2 expression as well as formation of TF-factor VIIa complexes on ECs and may induce intracellular signals; however, factor Xa exerts the most critical effect in the TF-factor VIIa-factor Xa complex [74].
In the clotting-independent pathway, the cytoplasmic domain of TF activates protein kinase C-dependent signaling, in contrast to the ligand-binding extraplasmatic tail of TF, which upregulates the synthesis of VEGF in response to different stimuli [75,76]. The cytoplasmic tail of TF appears to regulate non-clotting-dependent mechanisms such as cytoskeletal reorganization, vascular remodeling, angiogenesis, and cellular metastasis. Results from a recent study provided an alternative explanation whereby TF can promote angiogenesis by a clotting-independent mechanism [77]: the genetic deletion of the TF cytoplasm domain involves the loss of negative regulatory control with PAR-2 upregulation. Expression of TF and PAR-2 and the function of TF-PAR-2 signaling pathway require the presence of both angiogenic growth factors and inflammatory cytokines. Inflammatory cytokines produced by monocytes (MO) are important for angiogenesis and collateral growth vessels. Colocalization of upregulated PAR-2 with phosphorylated TF occurs only in neovessels. The phosphorylation of TF appears to be the mechanism that switches off the negative regulatory control that promotes pathologic PAR-2-dependent angiogenesis [77].
However, TF-PAR-2 signaling selectively synergizes with the PDGF isoform BB but not with VEGF, basic FGF, and the PDGF isoform AA. PDGF-BB is available through release from activated platelets in the context of local coagulation or by synthesis from sprouting ECs. PDGF-BB also has a role in lymphoangiogenesis. PDGF-BB stimulated MAPK activity and cell motility of isolated lymphatic ECs in vitro and potently induced the growth of lymphatic vessels in vivo [78]. Expression of PDGF-BB in murine fibrosarcoma cells induced tumor lymphoangiogenesis, leading to enhanced metastasis in lymph nodes [78].
The cascade of TF-factor VIIa to factor Xa and subsequently to thrombin can be inhibited by TF pathway inhibitor (TFPI), which occurs when the TF-factor VIIa complex is combined with factor Xa and cell-bound TFPI. Once this stable quaternary complex is formed, it is translocated to and internalized in caveolae (i.e., small vesicles of the plasma membrane) [79]. Produced by ECs and tumor cells, most TFPI is expressed on the cell surface, although it can be detected peripherally in plasma [80]. In tumor cells, the diminished expression of TFPI could result in activated factor Xa and increase factor Xa-PAR-2 signaling [81]. Results from in vitro and in vivo studies have suggested that therapeutic strategies that target an increase in the expression of TFPI could inhibit tumor angiogenesis [82], growth [83], and metastasis [53,84].
Another inhibitor of the TF-dependent pathway of blood coagulation that was recently identified is TFPI-2, which, in contrast to TFPI, inhibits the TF-factor VIIa complex but not factor Xa. TFPI-2 also inhibits various proteases [85]. Its expression was lower in laryngeal, breast, gastric, colon, pancreatic, renal, and endometrial tumor tissues and glial neoplasms than in normal counterpart tissue [85]. Some studies have provided evidence of the antiangiogenic and antimetastatic activity of TFPI-2 [86]. Epigenetic inactivation of TFPI-2 also contributes to the progression of pancreatic ductal adenocarcinoma. TFPI-2 mRNA was undetectable in many pancreatic cancer cell lines and in primary pancreatic ductal neoplasms. Hypermethylation of its promoter CpG island decreased TFPI-2 expression. However, expression of the TFPI-2 gene in pancreatic cancer cells has resulted in markedly suppressed proliferation, migration, and invasive potential in vitro [86]. TFPI-2 exerts antitumor effects through antiangiogenesis and antimetastatic mechanisms [87,88].
TF and metastasis
The role of TF-factor VIIa in metastasis has been extensively investigated. Some recent studies have revealed that TF can promote the invasion and metastasis of the A2780 human ovarian cancer cell line through the TF-factor VIIa pathway. In turn, TF-factor VIIa can upregulate the transcription of uPA receptor expression and enhance tumor invasion and metastasis [63,64]. Overexpression of TF of up to a 1,000-fold greater level than normal is characteristic of metastatic tumor cells relative to nonmetastatic cells [89]. Inhibition of TF through antibodies or active site-blocked factor XIIa interferes with experimental metastasis [90], but how TF on tumor cells contributes to tumor metastasis is unclear. TF-induced metastasis requires the participation of the cytoplasmic tail of TF and the assembly of an active TF-factor VIIa complex, which is thought to target G protein-coupled PAR-2. Factor VIIa induces the activation of p21Ras and p42/p44 MAPK, which further induces the activation of the protein kinase B pathway in TF-expressing cells and increases protein synthesis [91]. Both the p42/p44 MAPK and protein kinase B pathways are able to inhibit apoptosis. Inhibition of apoptosis may be related to adhesion-independent survival and thus may be linked to metastasis. In one study, factor VIIa inhibited cell death and caspase 3 activation in Baby Hamster Kidney cells induced by serum-deprivation and loss of adhesion (lack of integrin signaling) when transfected with TF genes [14]. The effects of factor VIIa on caspase 3 activity are sensitive to inhibitors of the phosphatidylinositol 3'-kinase and P42/P44 MAPK pathways; thus, factor VIIa appears to be a survival factor for TF-expressing cells and seems to mediate the survival of tumor cells during metastasis [14,92].
Another mechanism that links TF effects to metastasis is mediated by the overexpression of growth factors and proteins related to cellular reorganization. Interleukin (IL)-8, a member of the CXC chemokine (CXCL8) and a chemoattractant for neutrophils and lymphocytes, can act multifunctionally to induce tumor growth and metastasis, and IL-8 overexpression can occur through TF-FVIIa stimulation. In contrast, FXa and thrombin cannot upregulate IL-8 expression. IL-8 leads to increasing cell migration and invasion, and these effects are attenuated when binding of factor VIIa to TF is prevented [13]. In ovarian cancer studies, IL-8 has been suggested to be involved in the formation of ascites [93,94]. In our peritoneal gene transcript profile, we showed that IL-8 connects a number of pathways that might be linked to the pathogenesis of cancer by cellspace analysis (Figure 1) [3]. IL-8 may induce other immune cell-related modulators, such as the IL-13 receptor, vascular cell adhesion molecule 1, chemokine receptor 1, and other molecules associated with inflammatory function [95]. It is known that the inflammatory response is involved in the progression of tumors and that IL-8 might amplify inflammation in the tumor microenvironment to promote tumor cell proliferation, angiogenesis, and metastasis.
Figure 1 Genes associated with MO/MA activation. This figure was made by cellspace software according to an analysis of the functional genes involved in inflammatory response in peritoneal cavity with EOC[3]. The connected lines with different colors show different levels of association among genes. The warmer colors (yellow to red) represent stronger correlations.
Thrombin and thrombin receptor
Generation of thrombin is the central step in blood coagulation. Expression of thrombin has been detected in various tumor types, including ovarian cancer [18]. Thrombin's cellular effect is triggered by its binding to PARs, which are G protein-coupled receptors, of which four subtypes have been identified in human tissues. PAR-1, PAR-3, and PAR-4 are activated mainly by thrombin, whereas PAR-2 is activated by trypsin, tryptase, factor VIIa, and factor Xa. PAR-1, the major prototype of the family, is a receptor with seven transmembrane domains and is activated by cleavage of a specific site at the N-terminus of its extracellular domain. The presence of PAR family members has been demonstrated on many tumor cells, and it appears to be positively correlated with metastasis [96]. Studies of proteins at the cellular level (in situ hybridization and immunostaining) and studies of gene transcripts (reverse-transcriptase polymerase chain reaction) have demonstrated the differential expression pattern of PAR-1 gene in low-grade and high-grade EOC, but PAR-1 expression has not been shown on normal ovarian tissue even though the gene was overexpressed in the peritoneum of EOC patients [3,97]; thus PAR-1 signaling could be involved in the initiation and pathogenesis of EOC.
Thrombin-PAR system and angiogenesis
Thrombin is a potent activator of angiogenesis and could be the underlying mechanism for the promotion of tumor progression after thrombosis. However, evidence suggests that angiogenesis induced via thrombin is mediated via PAR and is independent of fibrin formation, which can be modulated without interfering with blood coagulation [98]. Thrombin also stimulates the chemotaxis of inflammatory cells and thus possibly promotes tumor angiogenesis indirectly [98]. Furthermore, thrombin could induce several cellular effects on ECs, thereby contributing to the angiogenic cascade.
There are two possible mechanisms by which thrombin may promote angiogenesis: (1) Thrombin at physiologic concentrations can inhibit ECs' ability to adhere to collagen type IV or to laminin [99], even after short exposure of ECs to thrombin. The result is reversible and is mediated specifically through the thrombin receptor. The result is detachment and migration of ECs from basement membrane components, the initial step in the activation of normally quiescent ECs. (2) Thrombin upregulates VEGF expression via PAR activation on ECs or tumor cells [100] and through release of platelets [101]. Thrombin antagonizes any blood vessel-stabilizing effects of angiopoietin 1 by production of angiopoietin 2 [102,103]. The angiogenic effects of thrombin include causing the release of basic FGF from ECs and the ECM [104]. Thrombin can also markedly stimulate expression of the VEGF receptor (kinase insert domain-containing receptor and flt-1). The activation of protein C kinase and MAPK signaling pathways was confirmed in the process of VEGF and VEGF receptor production [102]. Thrombin may thus contribute to new vessel formation by providing the pathologic microenvironment for tumor cell implants and proliferation; VEGF production through thrombin activation could facilitate this process. VEGF also increases the permeability of existing and new capillary vessels and results in plasma protein leakage and the development of a proangiogenic matrix [105]. A high expression level of VEGF allows the formation of malignant ascites. Thus, the thrombin-PAR system could play a role in the progression of ascites in EOC patients by inducing VEGF production.
Thrombin-PAR system in invasion and metastasis
In our peritoneal transcript profile study, we found that PAR-1 was upregulated in peritoneal tissue, which suggested that this receptor might be involved in EOC pathogenesis, invasion, and metastasis [3]. Overexpression of PAR-1 has been reported in malignant invasive melanoma [106] and breast cancer in vivo [96] and in breast cancer cell lines [107]. The increase in growth and metastatic potential of tumor cells may be partly attributable to the proangiogenic effects of thrombin. By mobilizing adhesion molecules, such as αIIβ3 integrin [108], P-selectin [109], and CD40 ligand [110], to the cell surface, thrombin enhances adhesion between tumor cells, platelets, ECs, and ECM and contributes to tumor progression. Thrombin also triggers the release of growth factors, chemokines, and ECM proteins that could promote the proliferation and migration of tumor cells. Facilitation of the metastatic activity of the thrombin-PAR complex has been demonstrated in vivo with experimental lung cancer [111], colon cancer [112], and breast cancer [15,96]. The principal thrombin receptor, PAR-1, has been implicated in promoting these effects. Most of the cellular effects elicited by thrombin are mediated through the activation and subsequent signal transduction cascades of members of the PAR family. Likewise, thrombin and PAR inhibitors, such as argatroban [113], activating protein 1 [112], and small interfering RNAs for PAR-1 [114], have been shown to prevent tumor migration and metastasis. Thrombin can also facilitate trans-basement membrane migration by ECM remodeling after activation of collagen type IV-degrading enzyme, MMP-2, and αvβ3 integrin. However, an immunostaining study on EOC found that focal adhesion kinase (FAK), a major tyrosine phosphorylated protein of adherent cells, was abundantly expressed in invasive EOC but not in normal ovarian tissues, which is consistent with a high level of PAR1 expression [97]. Possibly, PAR1 in malignant ovarian carcinoma transmits signals leading to the phosphorylation of FAK and thereby to alterations in integrin expression. Increased expression of αvβ3 integrin was also detected in our peritoneal profile study [3]. Anticoagulants such as heparin and warfarin have been shown to significantly decrease experimentally induced pulmonary metastasis in vivo [115], and a similar decrease in tumor cell metastasis was observed in vivo with hirudin, a highly specific inhibitor for thrombin [15]. In another study, hirudin inhibited tumor implantation and growth of four different tumor cell lines (MDA-MB-231 human breast cancer, 4T1 human breast cancer, A549 human non-small cell lung cancer, and B16F10 murine melanoma) in nude and syngeneic mice by inhibiting spontaneous seeding of tumors through the blood to susceptible organs, including the lungs, liver, spleen, and lymph nodes [15]. Overall, the interaction of endogenously generated thrombin and host thrombin receptors enhances tumor angiogenesis by enhancing the expression of some genes that may be associated with an invasive profile, and this interaction involves activation of PAR and other tumor genes.
At least six signal transduction mechanisms based on thrombin-PAR systems have been proposed to have a role in invasion and metastasis: (1) PAR may induce phosphorylation of FAK and integrin expression [97,116], thereby contributing to the modulation of cytoskeletal reorganization. Changes in cell shape result from changes in actin stress fibers and focal adhesion and are mediated through the Rho family of GTPases [117]. Rho-Ras proteins downstream of PAR-1 are intermediates of the extracellular signal-regulated kinases 1 and 2 (ERK1/2)-MAPK pathway and are implicated in cancer progression [118]. (2) Loss of activator protein 2 (AP-2) results in the increased expression of PAR-1, which subsequently contributes to the metastatic phenotype of cancer by upregulating the expression of angiogenic molecules, proteases, and adhesion molecules, which are involved in tumor invasion and metastasis [119]. This upregulation is based on MAPK signaling. AP-2 is a 52-kDa protein transcription factor mapped to the short term of chromosome 6, which is a critical regulator of gene expression during mammalian development, differentiation, and carcinogenesis. Loss of AP-2 may contribute to the development and progression of many different types of cancers, including melanoma and prostate, breast, and colorectal carcinoma [119]. AP-2 is a negative regulator for PAR-1 signaling, and functional AP-2-binding elements have been identified in some genes, including c-erbB2, VEGF, TGF-β, IGFBP-2, E-cadherin, and c-myc [119]. (3) Overexpression of flotillin-2, a highly conserved caveolae/lipid raft-associated protein, is associated with upregulation of PAR-1. Binding of flotillin-2 to PAR-1 might stabilize it by blocking inactivation or degradation of an activated PAR-1. Cytoplasmic flotillin-2 might also modulate PAR-1 trafficking of the cell or stabilize PAR-1 mRNA, which could promote PAR-1-induced constitutive signaling through the MAPK or other signal transduction pathways [120]. (4) PAR-1 cooperates with αvβ5 integrin to promote cytoskeletal reorganization. Activation of PAR-1 increased the phosphorylation of FAK and paxillin and induced the formation of focal contact complexes. The activation of PAR-1 may foster tumor cell invasion via a mechanism involving cooperation with the αvβ5 integrin [116]. (5) The increased and persistent expression of PAR-1 is induced by deregulated defective receptor trafficking, which is correlated with the failure to efficiently downregulate activated PAR-1 by internalization and lysosomal sorting [121]. (6) Coexpression of PAR-1 and PAR-2 has been observed on some tumor cells and the cells of the tumor microenvironment, such as fibroblasts, platelets, and ECs. Simultaneous activation of PAR-1 and PAR-2 by thrombin is required to cause the chemokinetic effect. Thrombin-induced activation of PAR-2 could play a subtle role in signaling induced by the thrombin-PAR-1 complex and stimulate the motility of metastatic tumor cells [122].
Platelets can act as bridges between ECs and lymphocytes or MOs/MAs, and thrombin activation of platelets can enhance hematogenous metastasis [123]. In addition, platelet activation and fibrin formation are both important mechanisms by which tumor cell-related "procoagulant activity" promotes metastasis. Platelet activation induced by thrombin is another possible mechanism for metastasis, though some studies have shown that knockout of the PAR-1 and PAR-2 gene of mice does not affect thrombin signaling in mouse platelets but might attenuate signals in mouse ECs [124,125].
Factor XIII and fibrin
Factor XIII, fibrinogen, and fibrin contribute to the final step in the coagulation cascade. They are involved in some pathologic states, including solid tumor growth and the formation of ascites. Fibrin is formed by thrombin cleavage of fibrinopeptides A and B from fibrinogen, exposure of cryptic polymerization sites, and finally the cleavage of plasma fibrinogen to fibrin by thrombin. A more stable clot is formed when active factor XIII, a transglutaminase, is covalently cross-linked to fibrin α and γ chains. Factor XIII binds to integrin and might also exert antiapoptotic effects on ECs [16]. In our peritoneal transcript profile study, the factor XIII gene was upregulated in the peritoneum and stroma in the vicinity of the tumor [3]. The factor XIIIa subunit supports the adhesion of platelets mediated by αIIbβ3 integrin [126] and fibroblasts mediated by αvβ3 integrin and β1-containing integrins [127]. The αvβ3 integrin, one of several integrins upregulated in the peritoneal transcript profile study [3], is present on vascular ECs and is a marker for angiogenesis. Factor XIII might also participate in angiogenesis. Factor XIIIa was shown to be a mediator of cell adhesion and an inhibitor of fibrin capillary tube formation in a dose-dependent manner [16].
Fibrin accumulation within tumors might result from increased permeability of tumor microvessels, which would lead to the leakage of plasma proteins, such as fibrinogen and plasminogen, and extravascular clotting and the cross-linking of fibrin to factor XIII [7]. In EOC ascites, high amounts of cross-linked fibrin degradation products have been identified [128]. VEGF could contribute to extravascular fibrin clotting in a tumor environment because it encourages leakage of plasma fibrinogen into extravascular spaces, where the fibrinogen clots [128]. Fibrin in the tumor or elsewhere in the peritoneum and serosa might help stimulate the new ingrowth of blood vessels. This fibrin gel matrix might also help provide the matrix that facilitates the ingrowth of MAs and fibroblasts and reorganizes the stroma, preparing it for tumor metastasis to the abdominal cavity [129,130]. Ascitic tumor cells as well as tumor cells derived from peritoneal and serosal surfaces can retain their viability in suspension cultures [131]. It is interesting that plasma exudates that contribute to ascites generally remain fluid and do not form an insoluble fibrin gel until removed from the patient [131]. Earlier studies showed that hyperpermeability of peritoneal lining vessels is required for ascites development in EOC tumors grown in murine models [132,133]. Cross-linked fibrin staining has not been observed in the peritoneal wall, diaphragm, mesentery, or bowel serosa of normal controls, but it has been detected in the peritoneum of ascitic tumor-bearing animals. Immunostaining revealed that adherent tumor cells were enmeshed in the fibrin meshwork binding them to each other and to peritoneal surfaces [129]. We also found that expression of phospholipase glutaminase A2, which mediates the first steps in the eicosanoid pathway, was increased in the peritoneum [3]. Leukotriene products on this pathway might contribute to capillary permeability and ascites formation.
Fibrin and angiogenesis
Fibrin-induced neovasculation is based on clotting-related mechanisms that involve platelet activation and fibrin deposition. Cross-linked fibrin has been found in different human malignant tumors. It is present in the endothelium of angiogenic vessels of invasive cancer specimens but not in vessels of benign tumors [134]. Fibrin can bind to inflammatory cells or to tumor cells and is deposited around tumor cells as scaffolding that promotes further angiogenesis. Fibrinogen and fibrin fragments, such as fragment E, have been shown to stimulate angiogenesis both in vitro and in vivo [135]. The presence of the fibrin degradation product D-dimer is significantly associated with a poor prognosis in some cancer patients [136,137].
The binding of ECs to fibrin with the involvement of the adhesion molecule vascular endothelial cadherin may be necessary for capillary tube formation, a critical step in angiogenesis. The fibrin matrix that develops around tumors provides a provisional proangiogenic environment that supports vessel formation and stimulates EC proliferation and migration [138]. The fibrin matrix can promote a proangiogenic response by upregulating the expression of αvβ3 integrin receptor to facilitate EC migration and capillary formation [139]. The αvβ3 integrin provides survival signals to ECs during their interaction with fibrin. Tumor-containing tissue has revealed increased deposition of fibrin stimulated by VEGF-induced vessel leakiness to sustain the proangiogeneic environment [7]. The fibrin matrix also stimulates the production of VEGF, basic FGF, IGF1, and IL-8 to promote an autocrine procoagulant loop by inducing TF expression in ECs [61]. The expression of IGF1 and αvβ3 integrin genes is upregulated in the peritoneum of EOC; therefore, fibrin may stimulate the production of these genes [3].
Regulatory proteins in the coagulation cascade
Two important regulatory proteins in the coagulation cascade, serine proteinase inhibitor D1 (Serpins, also called heparin cofactor II – HCII) and endothelial protein C receptor (EPCR), were first detected as an upexpression in a study we conducted [3], which suggests that there is value in studying them further. Serpins are a protein superfamily of which many members possess potent activity as serine proteinase inhibitors [140]. Another very important family of serpins is antithrombins. Both HCII and antithrombins can inhibit the blood coagulation proteinase thrombin.
HCII is a 480-amino-acid, single-chain glycoprotein with a molecular weight of about 66 kDa that is synthesized by the liver and circulates in the plasma. The human HCII gene is located on chromosome 22q11, spans 15.8 kb, and includes 5 exons [141]. When HCII interacts with heparin and dermatan sulfate, the potential inhibition of thrombin can be increased by more than 1,000-fold in vivo [142]. HCII inactivates thrombin by forming a stable, bimolecular complex in plasma. HCII does not inhibit other proteases involved in coagulation or fibrinolysis. It can also participate in wound healing by regulating the mitogenic and chemotactic activities of thrombin [143]. Two recent clinical research studies demonstrated that HCII is a very effective factor in combating heart and vascular disease, which it does by inhibiting the action of thrombin [144,145]. HCII is inactivated after cleavage of its reactive site (Leu-444 through Ser-445) by elastase from leukocytes during the inflammation process without forming the stable thrombin-HCII complex [146].
HCII has a physiological role in the inflammation response. An active peptide from the amino-terminal region (corresponding to Asp-39 through Ile-66) with chemoattractant action for both neutrophils and MOs is proteolyzed by neutrophil elastase, and another leukocyte migration peptide is derived from dodecapeptide from Asp-49 through Tyr-60 of HCII. HCII-neutrophil proteinase products appear to have a role in the local inflammatory response [147]. Leukocytes and MOs/MAs are important cellular components of tissue injury, wound healing, and the microenvironment of EOC tumors [6]. The detection of upregulated HCII in the peritoneum and stroma of EOC patients is interesting because the degradation of HCII can produce a chemoattractant peptide for MAs [3]. MOs/MAs have multifunctional roles in EOC progression by releasing cytokines and chemokines, and HCII may help induce MA migration from the endothelium to extravascular tissue.
EPCR
EPCR is a 46-kDa type 1 transmembrane glycoprotein with two domains in the extracellular region that are homologous to the α1 and α2 domains of CD1/major histocompatibility complex class I molecules [148]. In humans, the EPCR gene is located on chromosome 20q11.2. Most EPCRs are expressed on the surface of ECs of vessels. A 43-kDa soluble EPCR has recently been found in the plasma of humans [149].
Downstream of coagulation, thrombin stimulates platelet aggregation, promotes coagulation by cleavage of fibrinogen, and fosters fibrin formation and activation of factor XIII. Thrombin is inhibited via the activated protein C (APC)-EPCR system. When thrombin binds to thrombomodulin on the surface of ECs, it activates protein C through EPCR to prevent coagulation. APC could inactivate factors V and VIII. The catalytic reaction for protein C activation by the thrombin-thrombomodulin complex is inefficient; only after combining with EPCR is protein C fully activated and does it acquire an anticoagulant role [150]. In addition to its anticoagulation role, APC inhibits thrombin-induced proinflammatory activity, such as platelet activation, cytokine-induced chemotaxis for MOs and neutrophils, and upregulation of leukocyte adhesion molecules. APC may prevent inflammation by downregulating TF and tumor necrosis factor-α (TNF-α), nuclear factor κB translocation, and cytokine signaling from MAs [151] and by inhibiting TNF-α-induced upregulation of cell surface leukocyte adhesion molecules [152]. APC may also protect the vasculature by blocking p53-mediated apoptosis in ischemic cerebral vasculature [153]. Because of its anti-inflammatory properties, APC is important in controlling serious infections. The apoptotic function of APC is independent of its anticoagulant function and requires EPCR as a cofactor but is mediated via PAR-1 [154]. When complexed with APC, membrane EPCR, which is expressed on the surface of ECs, could play a role in anticoagulation, inhibition of inflammation, and stimulation of cell proliferation (i.e., antiapoptosis) [155]. In contrast, soluble EPCR combines with protein C and APC with the same affinity but inhibits the anticoagulant activity of APC and interferes with the protective role of membrane EPCR and APC. Soluble EPCR may bind to and interfere with the function of neutrophil integrins, such as CD11b/CD18, which facilitates neutrophil adhesion to activated ECs and extravasation into extravascular tissue [156]. Soluble EPCR seems to amplify the inflammatory response [157,158]. High levels of soluble EPCR have been detected in sepsis [159] and in autoimmune disease [159]. Results from an animal model study showed that EPCR expressed on endothelium had a protective role in the cardiovascular system [160].
Only two studies relating to EPCR and cancer have been published [161,162]. Both studies were performed on cell lines. The membrane protein LMR42 was upregulated in most multidrug-resistant tumor cells. Expression cloning and sequence analysis showed LMR42 to be identical to EPCR. Elevated EPCR expression occurred in 47% of the primary tumor cell lines, including melanomas and renal and colon carcinomas. The authors of one of the studies concluded that elevated expression of EPCR may have a role in the resistant phenotype of multidrug-resistant tumor cells [161]. EPCR was also expressed on glioblastoma, leukemia, and most breast cancer cell lines; increased levels of APC activation were observed in tumor cells that express both EPCR and thrombomodulin [162]. Although the role of EPCR in cancer biology is poorly understood, it appears that EPCR may contribute to tumor progression. Our peritoneal transcript profile study showed high expression of EPCR on EOC peritoneum and its stroma [3]. PAR-1 is known to mediate a response in ECs, including the production of platelet-activating factor and chemokines (e.g., IL-8) and upregulation of adhesion molecules for neutrophils and platelets, which promote inflammation, thrombosis, and tumor progression [163]. However, the interaction of APC and PAR-1 is mediated by EPCR. At physiologic levels, APC is 10,000 times less potent than thrombin in this PAR-1-mediated pathway. When EPCR expression is upregulated, APC-PAR-1 signal transduction is significantly intensified, although it is still less potent than thrombin in this system [164].
The APC-EPCR interaction provides protection for ECs through its antiapoptotic effects. Because high EPCR expression occurs on tumors, EPCR might contribute to antiapoptosis in tumor cells. Finally, APC can stimulate PAR-1, whose activation depends on the EPCR concentration. A PAR-1-associated cellular response might contribute to tumor progression, possibly by producing IL-8, which stimulates tumor cell proliferation and metastasis.
The connection between coagulation and inflammation in the EOC peritoneum
There is substantial evidence of infiltrating immune cells in EOC and the peritoneal environment. In a previous study, we showed that about 70% of T cells within EOC tumors were mononuclear cells [165]. Results from other researchers' experiments have suggested that the presence of these T cells is associated with an antigen-driven immune response [166-168]. However, there is little evidence to suggest there is a chronic cellular adaptive immune response in vivo because interferon γ message is absent or present at low levels [169]. Expression of the CD3ζ chain was absent or poor in another study [170]. A recent study showed longer overall and progression-free survival among EOC patients with intratumoral T cells than among those lacking these cells [168]. In a more recent study, CD4+CD25+ regulatory T cells present in EOC tumors appeared to be associated with poor patient survival. Regulatory T cells preferentially move to and accumulate in tumors and ascites but rarely enter the draining lymph nodes in later cancer stages [171]. Large numbers of MOs/MAs are also present in ascites, where they may make up 50% or more of mononuclear leukocytes, whereas the proportion that consists of T lymphocytes is usually less than 40% [165]. In recent preliminary studies, we found that pelvic peritoneal tissue biopsied from patients with advanced-stage EOC had a higher proportion of MOs/MAs than T cells even in the absence of tumor involvement (unpublished data).
Cytokines, chemokines, adhesion molecules, and components of the ECM may contribute to a tissue environment that supports tumor proliferation and invasion. Chemokines and certain cytokines may facilitate the migration of immune cells, including T cells and MAs, into the tumor environment (Tables 3 and 4). A network of CC (cysteines with no intervening amino acid) and CXC chemokines has been found in solid tumors and in ascites [172]. For example, CCL2, one of the ligands for CCR2, localizes to epithelial areas of the tumor [173], and its expression appears to correlate with the numbers of lymphocytes and MAs at this site [174]. CCL5 localizes with tumor-infiltrating leukocytes, and CCL5 concentration may be associated with CD8+ T cell infiltration. Chemokines CCL2, CC8, CCL4, CCL5, CCL8, CCL22, CXCL2, and CXCL12 have all been found in increased amounts in ascitic fluid from EOC patients, and the presence of CCL5 in ascitic fluid has been associated with the number of T cells in ascites [175].
Table 3 Cytokines involved in inflammation in ovarian cancer
Cytokine Effect
TGF-β Stimulates tumor cell attachment and invasion by upregulating plasminogen activator inhibitor type 1
Deregulates expression of the major histocompatibility complex
Deregulates costimulatory antigen expression by dendritic cells
Suppresses Th1-Th2 cells and the conversion of pro-cytotoxic T lymphocytes to cytotoxic T lymphocytes
Suppresses the proliferation response to antigen-presenting cells
Inhibits natural killer and MA activation
IL-6 Upregulates tumor cell attachment
Interferes with macrophage maturation in dendritic cells
Inhibits cell proliferation through PI3K (phosphatidylinositol3-kinase)
Suppresses Th1-Th2 transformation
IL-10 Deregulates expression of major compatibility complex on T cells
Deregulates costimulatory antigen expression
Suppresses cytotoxic T lymphocyte activation and Th1-Th2 transformation
Inhibits interferon production
Inhibits T cell production
VEGF Increases neoangiogenesis, invasion, and metastasis
Increases prevalence of ascites
FGF-2 Increases tumor cell invasion and metastasis
TNF-α Increases adhesion molecule expression
Induces tumor cell apoptosis
Increases tissue factor expression
Decreases the expression of thrombomodulin and EPCR
Table 4 Chemokines involved in ovarian pathogenesis
Chemokine Receptor Cells targeted Comments
CCL22 CCR4 CD25+CD4+ regulatory T cells Forster tumor immune escape
CCL2 CCR2 Activated T cells, monocytes, and DC Deregulate CD8+ T cells and CD68+ macrophage
CCL3 CCR2 Activated T, NK, MO, Eosinophils Inflammatory cells migration
CCL4 CCR5 DC, MO, NK Inflammatory cells migration
CCL5 CCR2 Activated T, NK, MO, Eosinophils Inflammatory cells migration
CCL7 CCR2 Activated T, NK, MO, Eosinophils Inflammatory cells migration
CCL18 Unknown MOs/MAs MA produced but not induced in EOC
CXCL8 CXCR1, CXCR2 Neutrophils and resting T cells Angiogenesis, metastasis
CXCL12 CXCR4 Neutrophils, resting T cells, activated T and B cells, and macrophages CXCR4 is preferentially expressed on EOC cells
Adhesion molecules may also be involved in the implantation of cancers cells onto the peritoneal lining. For example, E-cadherins and P-cadherins are expressed during EOC progression and may facilitate peritoneal invasion because mesothelial cells express cadherin-binding catenins [4]. Upregulation of discoidin domain receptor 1, claudin 3, and epithelial cell adhesion molecule occurs early in the development of EOC [176]. Soluble intracellular adhesion molecule 1 concentrations are elevated in patients with ovarian malignant tumors but do not correlate with clinical status[177].
Depending on their degree of expression in the tissue microenvironment, coagulation and inflammation might influence each other. Furthermore, components generated from both the coagulation cascade and an inflammatory response might be involved in heart disease, as recently reviewed [178-181]. Normal hemostasis is initiated when the blood vessels rupture, allowing blood cells to interact with extravascular cells and the ECM [182]. In one study, endotoxin, TNF-α, and IL-1α induced TF expression, primarily on MO/MA [58]. Blood clotting might thus be initiated when inflammatory cytokines and endotoxin induce the synthesis of TF on migrating leukocytes [183]. In this way, immune cells could help initiate the coagulation cascade through damaged tissue or cytokine production [184] (Table 5).
Table 5 Procoagulant effects of inflammation products
Product Procoagulant effect
Leukocytes or MAs Activate platelets
Induce microvascular occlusion
Release neutrophil elastase
Induce thrombomodulin release from the endothelium
Inactivate antithrombin
Complements Increase procoagulant factors production
Provide membrane surfaces
Augment TF expression
Inflammatory Cytokines Increase TF production
Decrease thrombin, EPCR, and protein S levels
Increase C-reactive protein production
Increase complement activation
Increase platelet count and reactivity
Chemokines Activate platelet aggregation and adhesion
Attract more leukocytes
The complement system may also contribute to hemostasis by activating the membrane attack complement pathway, thereby contributing to phosphatidylserine expression on the outer membrane of cells [185]. Expression of phosphatidylserine on the outer cell membrane is necessary for the effective initiation and amplification of the coagulation cascade. Our microarray analysis of the peritoneum of EOC patients showed increased expression of the C2 component [3]. In contrast, expression of complement HF1, which blocks the membrane attack complement pathway, was increased and might have shifted the complement activation pathway toward a proinflammatory response with the recruitment of MO/MA. MO/MA are also an important source of complement components.
Chemokines can influence coagulation by activating platelets indirectly. Three main chemokines are involved in this process, CXCL12 (the ligand for CXCR4), CCL17 (the ligand for CCR4), and CCR22 (the ligand for CCR4) [186]. Platelets are an important link in the inflammation and coagulation processes. Inflammatory mediators, cytokines, and chemokines do not increase platelet production, but the platelets generated are more thrombogenic and thus are more sensitive to platelet agonists such as thrombin. Platelets contain high concentrations of the proinflammatory mediator CD40 [187] and certain chemokines [188]. When CD40 and these chemokines are released, they can induce TF synthesis and increase the production of inflammatory cytokines such as IL-6 and IL-8. Of the naturally occurring anticoagulants, protein C is the most negatively influenced by inflammation. Thrombomodulin and EPCR are both downregulated by inflammatory cytokines such as TNF-α [189] and IL-1 [190].
Thrombin has a variety of effects on cells that can enhance the inflammatory response. For example, it augments leukocyte adhesion and activation, which amplifies the inflammatory response, and induces endothelium and platelet activation. The activation of platelets can produce cytokines (IL-6) and chemokines (IL-8), which are involved in inflammatory responses [186]. Thrombin also acts as a direct mitogen for fibroblasts, inducing PDGF and TGF from platelets and ECs. PDGF and TGF-β are important cytokines for stimulating neovessel formation, and TGF-β may contribute to an environment conducive to tumor escape from immune system surveillance. Notably, PDGF and TGF-β 3 receptor expression were increased in our peritoneal transcript profile study [3]; this finding provides additional evidence that thrombin could foster tumor cell growth and proliferation via PDGF and TGF release from platelets and ECs.
Anticancer therapy based on targeting the coagulation cascade
Coagulation factors have a profound effect on tumor cell behavior in both in vivo and in vitro studies. These factors could enhance tumor cell proliferation, invasion, angiogenesis, and metastasis. Hence, targeting activated coagulation factors might provide a viable cancer treatment strategy.
Heparins are the most extensively used anticoagulants in clinics. In blood coagulation, unfractionated heparin and low-molecular-weight heparins potentiate the activity of antithrombin III, thus inhibiting the activation of coagulation factors II and X [191]. They also release TFPI, a physiologic inhibitor of the TF pathway that prevents pulmonary embolism and is used to treat deep vein thrombosis [192]. Retrospective and meta-analytic studies of deep vein thrombosis treatment have shown longer survival among cancer patients with thrombosis who were treated with unfractionated heparin and low-molecular-weight heparins than among patients treated without heparin [193-198]. Thus, the use of anticoagulants might allow them to live longer.
There is a higher incidence of venous thromboembolism in patients with melanoma or small cell lung cancer. In 2003, a pilot study of enoxaprin, a low-molecular-weight heparin for the treatment of advanced melanoma was reported [199], but the clinical outcome was not clear. Then, a phase II trial of combination chemotherapy and anticoagulant therapy (docetaxel plus enoxaparin) was performed in chemotherapy-naive patients with metastatic non-small cell lung cancer [200]. The median time to progression was 5 months, and the median survival time was 11 months. The most frequent toxic effects were neutropenia and asthenia; no clinically significant bleeding or thrombotic events were observed. Treatment was well tolerated in patients with advanced small cell lung cancer, and the results suggested that enoxaparin could prolong the time to disease progression [200]. Both the oral anticoagulant warfarin and unfractionated heparin have been shown to prolong survival time among patients with small cell carcinoma [201].
In 2004, a prospective, randomized, controlled trial enrolled 385 cancer patients into two arms, placebo and dalteparin (another low-molecular-weight heparin). Types of cancer included breast, colorectal, ovarian, and pancreatic [202]. Thirty-four percent of the dalteparin group and 31% of the placebo group received chemotherapy alone; 8% of each group received radiation therapy alone. Estimated overall survival at 1, 2, and 3 years did not differ significantly between groups overall. However, the estimated overall survival among patients with a better prognosis at enrollment (55 patients in the dalteparin group and 47 patients in the placebo group) was significantly longer in the dalteparin group at 2 years (78% and 60%; P = 0.03) and at 3 years (55% and 36%; P = 0.03). In 2005, another study confirmed these antineoplastic effects of dalteparin [203]. During the 12-month follow-up period of the study, 602 patients with solid tumors and venous thromboembolism were randomly assigned to a dalteparin or coumarin-derivative treatment group. Among patients without metastatic disease, the probability of death at 12 months was 20% in the dalteparin group compared with 36% in the oral anticoagulant group (P = .03). In patients with metastatic cancer, no difference was observed in mortality between the treatment groups (72% and 69%, P = .46). However, the observed effects of dalteparin on survival were statistically significantly different between patients with and without metastatic disease (P = .02). Clinical trials are warranted to investigate these findings. The exact mechanism by which heparin mediates antitumor or antimetastatic activity is unknown but also merits further study.
Conclusion
Factors from the extrinsic and intrinsic coagulation cascades play complex and important roles in cancer progression by promoting blood clotting; in the microenvironment of cancer (e.g., ECs, platelets, fibroblasts, leukocytes, and the ECM), altered gene expression affects key intracellular signaling events. Certain signaling pathways may facilitate thrombus formation in the peritumoral environment and promote localized angiogenesis. In our study of transcript profiles of the peritoneum, several factors that are part of the intrinsic coagulation cascade were overexpressed at the transcript level in the peritoneum in the vicinity of but beyond the periphery of the tumor [3]. Overall, the extrinsic and intrinsic pathways favor clot formation. Thus, products of the peritoneal environment, which include chemokines, cytokines, and coagulation factors or their receptors, are evolving as potential targets for biologic therapy or in vivo diagnostic tools. The number of humanized monoclonal antibodies applicable to cancer treatment is increasing as more targets are discovered. Other approaches to treatment might include inhibitory oligonucleotides packaged to avoid degradation and small molecules that block specific pathways. Retrospective studies have shown that antithrombotic therapy may be associated with a lower incidence of certain types of cancer [204]. Clinical trials need to be expanded to target the production of certain coagulation factors or receptors in the coagulation cascade of advanced ovarian cancer. Such studies may reveal new approaches for chemoprevention or combination therapy that include chemotherapy to control tumor proliferation, angiogenesis, and metastasis.
==== Refs
Jemal A Murray T Ward E Samuels A Tiwari RC Ghafoor A Feuer EJ Thun MJ Cancer statistics, 2005 CA Cancer J Clin 2005 55 10 30 15661684
Feeley KM Wells M Precursor lesions of ovarian epithelial malignancy Histopathology 2001 38 87 95 11207821 10.1046/j.1365-2559.2001.01042.x
Wang E Ngalame Y Panelli MC Nguyen-Jackson H Deavers M Mueller P Hu W Savary CA Kobayashi R Freedman RS Marincola FM Peritoneal and Subperitoneal Stroma May Facilitate Regional Spread of Ovarian Cancer Clin Cancer Res 2005 11 113 122 15671535
Patel IS Madan P Getsios S Bertrand MA MacCalman CD Cadherin switching in ovarian cancer progression Int J Cancer 2003 106 172 177 12800191 10.1002/ijc.11086
Roskelley CD Bissell MJ The dominance of the microenvironment in breast and ovarian cancer Semin Cancer Biol 2002 12 97 104 12027581 10.1006/scbi.2001.0417
Freedman RS Deavers M Liu J Wang E Peritoneal inflammation - A microenvironment for Epithelial Ovarian Cancer (EOC) J Transl Med 2004 2 23 15219235 10.1186/1479-5876-2-23
Dvorak HF Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing N Engl J Med 1986 315 1650 1659 3537791
Kuperwasser C Chavarria T Wu M Magrane G Gray JW Carey L Richardson A Weinberg RA Reconstruction of functionally normal and malignant human breast tissues in mice Proc Natl Acad Sci U S A 2004 101 4966 4971 15051869 10.1073/pnas.0401064101
A. T Phlegmasia alba dolens: Clinique medicale de l'Hotel-Dieu de Paris. 2nd ed. paris, Fance: JBBalliere et Fils 1865 654 712
Sampson MT Kakkar AK Coagulation proteases and human cancer Biochem Soc Trans 2002 30 201 207 12023851 10.1042/BST0300201
Rickles FR Levine M Edwards RL Hemostatic alterations in cancer patients Cancer Metastasis Rev 1992 11 237 248 1423816 10.1007/BF01307180
Zacharski LR Memoli VA Ornstein DL Rousseau SM Kisiel W Kudryk BJ Tumor cell procoagulant and urokinase expression in carcinoma of the ovary J Natl Cancer Inst 1993 85 1225 1230 8331683
Naschitz JE Yeshurun D Lev LM Thromboembolism in cancer. Changing trends Cancer 1993 71 1384 1390 8435814
Schmeidler-Sapiro KT Ratnoff OD Gordon EM Mitogenic Effects of Coagulation Factor XII and Factor XIIa on HepG2 Cells PNAS 1991 88 4382 4385 1852005
Borgono CA Diamandis EP The emerging roles of human tissue kallikreins in cancer Nat Rev Cancer 2004 4 876 890 15516960 10.1038/nrc1474
Xi Z Kaern J Davidson B Klokk TI Risberg B Trope C Saatcioglu F Kallikrein 4 is associated with paclitaxel resistance in ovarian cancer Gynecol Oncol 2004 94 80 85 15262123 10.1016/j.ygyno.2004.03.044
Yousef GM Polymeris ME Grass L Soosaipillai A Chan PC Scorilas A Borgono C Harbeck N Schmalfeldt B Dorn J Schmitt M Diamandis EP Human kallikrein 5: a potential novel serum biomarker for breast and ovarian cancer Cancer Res 2003 63 3958 3965 12873991
Yousef GM Polymeris ME Yacoub GM Scorilas A Soosaipillai A Popalis C Fracchioli S Katsaros D Diamandis EP Parallel overexpression of seven kallikrein genes in ovarian cancer Cancer Res 2003 63 2223 2227 12727843
Diamandis EP Scorilas A Fracchioli S Van Gramberen M De Bruijn H Henrik A Soosaipillai A Grass L Yousef GM Stenman UH Massobrio M Van Der Zee AG Vergote I Katsaros D Human kallikrein 6 (hK6): a new potential serum biomarker for diagnosis and prognosis of ovarian carcinoma J Clin Oncol 2003 21 1035 1043 12637468 10.1200/JCO.2003.02.022
Ni X Zhang W Huang KC Wang Y Ng SK Mok SC Berkowitz RS Ng SW Characterisation of human kallikrein 6/protease M expression in ovarian cancer Br J Cancer 2004 91 725 731 15305183
Dong Y Kaushal A Brattsand M Nicklin J Clements JA Differential splicing of KLK5 and KLK7 in epithelial ovarian cancer produces novel variants with potential as cancer biomarkers Clin Cancer Res 2003 9 1710 1720 12738725
Magklara A Scorilas A Katsaros D Massobrio M Yousef GM Fracchioli S Danese S Diamandis EP The human KLK8 (neuropsin/ovasin) gene: identification of two novel splice variants and its prognostic value in ovarian cancer Clin Cancer Res 2001 7 806 811 11309326
Kishi T Grass L Soosaipillai A Scorilas A Harbeck N Schmalfeldt B Dorn J Mysliwiec M Schmitt M Diamandis EP Human kallikrein 8, a novel biomarker for ovarian carcinoma Cancer Res 2003 63 2771 2774 12782581
Yousef GM Kyriakopoulou LG Scorilas A Fracchioli S Ghiringhello B Zarghooni M Chang A Diamandis M Giardina G Hartwick WJ Richiardi G Massobrio M Diamandis EP Katsaros D Quantitative expression of the human kallikrein gene 9 (KLK9) in ovarian cancer: a new independent and favorable prognostic marker Cancer Res 2001 61 7811 7818 11691797
Luo LY Katsaros D Scorilas A Fracchioli S Piccinno R Rigault de la Longrais IA Howarth DJ Diamandis EP Prognostic value of human kallikrein 10 expression in epithelial ovarian carcinoma Clin Cancer Res 2001 7 2372 2379 11489815
Shvartsman HS Lu KH Lee J Lillie J Deavers MT Clifford S Wolf JK Mills GB Bast RCJ Gershenson DM Schmandt R Overexpression of kallikrein 10 in epithelial ovarian carcinomas Gynecol Oncol 2003 90 44 50 12821340 10.1016/S0090-8258(03)00257-9
Luo LY Katsaros D Scorilas A Fracchioli S Bellino R van Gramberen M de Bruijn H Henrik A Stenman UH Massobrio M van der Zee AG Vergote I Diamandis EP The serum concentration of human kallikrein 10 represents a novel biomarker for ovarian cancer diagnosis and prognosis Cancer Res 2003 63 807 811 12591730
Borgono CA Fracchioli S Yousef GM Rigault de la Longrais IA Luo LY Soosaipillai A Puopolo M Grass L Scorilas A Diamandis EP Katsaros D Favorable prognostic value of tissue human kallikrein 11 (hK11) in patients with ovarian carcinoma Int J Cancer 2003 106 605 610 12845660 10.1002/ijc.11296
Diamandis EP Borgono CA Scorilas A Harbeck N Dorn J Schmitt M Human kallikrein 11: an indicator of favorable prognosis in ovarian cancer patients Clin Biochem 2004 37 823 829 15329323 10.1016/j.clinbiochem.2004.04.009
Shigemasa K Gu L Tanimoto H O'Brien TJ Ohama K Human kallikrein gene 11 (KLK11) mRNA overexpression is associated with poor prognosis in patients with epithelial ovarian cancer Clin Cancer Res 2004 10 2766 2770 15102682
Diamandis EP Okui A Mitsui S Luo LY Soosaipillai A Grass L Nakamura T Howarth DJ Yamaguchi N Human kallikrein 11: a new biomarker of prostate and ovarian carcinoma Cancer Res 2002 62 295 300 11782391
Scorilas A Borgono CA Harbeck N Dorn J Schmalfeldt B Schmitt M Diamandis EP Human kallikrein 13 protein in ovarian cancer cytosols: a new favorable prognostic marker J Clin Oncol 2004 22 678 685 14966091 10.1200/JCO.2004.05.144
Kapadia C Ghosh MC Grass L Diamandis EP Human kallikrein 13 involvement in extracellular matrix degradation Biochem Biophys Res Commun 2004 323 1084 1090 15381110 10.1016/j.bbrc.2004.08.206
Borgono CA Grass L Soosaipillai A Yousef GM Petraki CD Howarth DH Fracchioli S Katsaros D Diamandis EP Human kallikrein 14: a new potential biomarker for ovarian and breast cancer Cancer Res 2003 63 9032 9041 14695222
Yousef GM Fracchioli S Scorilas A Borgono CA Iskander L Puopolo M Massobrio M Diamandis EP Katsaros D Steroid hormone regulation and prognostic value of the human kallikrein gene 14 in ovarian cancer Am J Clin Pathol 2003 119 346 355 12645335
Yousef GM Scorilas A Katsaros D Fracchioli S Iskander L Borgono C Rigault de la Longrais IA Puopolo M Massobrio M Diamandis EP Prognostic value of the human kallikrein gene 15 expression in ovarian cancer J Clin Oncol 2003 21 3119 3126 12915603 10.1200/JCO.2003.09.111
Adib TR Henderson S Perrett C Hewitt D Bourmpoulia D Ledermann J Boshoff C Predicting biomarkers for ovarian cancer using gene-expression microarrays Br J Cancer 2004 90 686 692 14760385 10.1038/sj.bjc.6601603
Deperthes D Frenette G Brillard-Bourdet M Bourgeois L Gauthier F Tremblay RR Dube JY Potential involvement of kallikrein hK2 in the hydrolysis of the human seminal vesicle proteins after ejaculation J Androl 1996 17 659 665 9016396
Webber MM Waghray A Bello D Prostate-specific antigen, a serine protease, facilitates human prostate cancer cell invasion Clin Cancer Res 1995 1 1089 1094 9815898
Magklara A Mellati AA Wasney GA Little SP Sotiropoulou G Becker GW Diamandis EP Characterization of the enzymatic activity of human kallikrein 6: Autoactivation, substrate specificity, and regulation by inhibitors Biochem Biophys Res Commun 2003 307 948 955 12878203 10.1016/S0006-291X(03)01271-3
Mikolajczyk SD Millar LS Kumar A Saedi MS Prostatic human kallikrein 2 inactivates and complexes with plasminogen activator inhibitor-1 Int J Cancer 1999 81 438 442 10209959 10.1002/(SICI)1097-0215(19990505)81:3<438::AID-IJC18>3.0.CO;2-U
Emanueli C Minasi A Zacheo A Chao J Chao L Salis MB Straino S Tozzi MG Smith R Gaspa L Bianchini G Stillo F Capogrossi MC Madeddu P Local delivery of human tissue kallikrein gene accelerates spontaneous angiogenesis in mouse model of hindlimb ischemia Circulation 2001 103 125 132 11136697
Wolf WC Evans DM Chao L Chao J A Synthetic Tissue Kallikrein Inhibitor Suppresses Cancer Cell Invasiveness Am J Pathol 2001 159 1797 1805 11696440
Kamath L Meydani A Foss F Kuliopulos A Signaling from protease-activated receptor-1 inhibits migration and invasion of breast cancer cells Cancer Res 2001 61 5933 5940 11479236
Rehault S Monget P Mazerbourg S Tremblay R Gutman N Gauthier F Moreau T Insulin-like growth factor binding proteins (IGFBPs) as potential physiological substrates for human kallikreins hK2 and hK3 Eur J Biochem 2001 268 2960 2968 11358513 10.1046/j.1432-1327.2001.02185.x
Sutkowski DM Goode RL Baniel J Teater C Cohen P McNulty AM Hsiung HM Becker GW Neubauer BL Growth regulation of prostatic stromal cells by prostate-specific antigen J Natl Cancer Inst 1999 91 1663 1669 10511594 10.1093/jnci/91.19.1663
D'Andrea MR Derian CK Santulli RJ Andrade-Gordon P Differential Expression of Protease-Activated Receptors-1 and -2 in Stromal Fibroblasts of Normal, Benign, and Malignant Human Tissues Am J Pathol 2001 158 2031 2041 11395381
Seligsohn U Factor XI deficiency Thromb Haemost 1993 70 68 71 8236117
Podmore A Smith M Savidge G Alhaq A Real-time quantitative PCR analysis of factor XI mRNA variants in human platelets J Thromb Haemost 2004 2 1713 1719 15456480 10.1111/j.1538-7836.2004.00924.x
Bazan JF Structural design and molecular evolution of a cytokine receptor superfamily Proc Natl Acad Sci U S A 1990 87 6934 6938 2169613
Carson SD Bromberg ME Tissue factor encryption/de-encryption is not altered in the absence of the cytoplasmic domain Thromb Haemost 2000 84 657 663 11057866
Edgington TS Mackman N Brand K Ruf W The structural biology of expression and function of tissue factor Thromb Haemost 1991 66 67 79 1833852
Clauss M Weich H Breier G Knies U Rockl W Waltenberger J Risau W The vascular endothelial growth factor receptor Flt-1 mediates biological activities. Implications for a functional role of placenta growth factor in monocyte activation and chemotaxis J Biol Chem 1996 271 17629 17634 8663424 10.1074/jbc.271.30.17629
Bartha K Brisson C Archipoff G de la Salle C Lanza F Cazenave JP Beretz A Thrombin regulates tissue factor and thrombomodulin mRNA levels and activities in human saphenous vein endothelial cells by distinct mechanisms J Biol Chem 1993 268 421 429 7678000
Contrino J Goralnick S Qi J Hair G Rickles FR Kreutzer DL Fibrin induction of tissue factor expression in human vascular endothelial cells Circulation 1997 96 605 613 9244233
Fang J Wei WN Liu ZP Song SJ [Cloning of expression vector of human tissue factor gene and its expression in human ovarian cancer cell line] Zhongguo Shi Yan Xue Ye Xue Za Zhi 2003 11 579 582 14706138
Fang J Wei WN Xia LH Song SJ [The effects of tissue factor/activated factor VII complex on the invasion and metastasis of human ovarian cancer.] Zhonghua Xue Ye Xue Za Zhi 2004 25 523 527 15569528
Fang J Wei WN Xia LH Song SJ [Study of the effect of TF/FVIIa complex on the expression of u-PAR mRNA in human ovarian cancer] Zhonghua Xue Ye Xue Za Zhi 2004 25 143 146 15182581
Ueno T Toi M Koike M Nakamura S Tominaga T Tissue factor expression in breast cancer tissues: its correlation with prognosis and plasma concentration Br J Cancer 2000 83 164 170 10901365 10.1054/bjoc.2000.1272
Sawada M Miyake S Ohdama S Matsubara O Masuda S Yakumaru K Yoshizawa Y Expression of tissue factor in non-small-cell lung cancers and its relationship to metastasis Br J Cancer 1999 79 472 477 10027315 10.1038/sj.bjc.6690073
Nakasaki T Wada H Shigemori C Miki C Gabazza EC Nobori T Nakamura S Shiku H Expression of tissue factor and vascular endothelial growth factor is associated with angiogenesis in colorectal cancer Am J Hematol 2002 69 247 254 11921018 10.1002/ajh.10061
Ueda C Hirohata Y Kihara Y Nakamura H Abe S Akahane K Okamoto K Itoh H Otsuki M Pancreatic cancer complicated by disseminated intravascular coagulation associated with production of tissue factor J Gastroenterol 2001 36 848 850 11777214 10.1007/s005350170008
Lwaleed BA Cooper AJ Tissue factor expression and multidrug resistance in cancer: two aspects of a common cellular response to a hostile milieu Med Hypotheses 2000 55 470 473 11090292 10.1054/mehy.2000.1093
Hanahan D Folkman J Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis Cell 1996 86 353 364 8756718 10.1016/S0092-8674(00)80108-7
Fernandez PM Rickles FR Tissue factor and angiogenesis in cancer Curr Opin Hematol 2002 9 401 406 12172458 10.1097/00062752-200209000-00003
Zhang Y Deng Y Luther T Muller M Ziegler R Waldherr R Stern DM Nawroth PP Tissue factor controls the balance of angiogenic and antiangiogenic properties of tumor cells in mice J Clin Invest 1994 94 1320 1327 7521887
Shoji M Abe K Nawroth PP Rickles FR Molecular Mechanisms Linking Thrombosis and Angiogenesis in Cancer Trends in Cardiovascular Medicine 1997 7 52 59 10.1016/S1050-1738(96)00142-9
Sorensen BB Freskgard PO Nielsen LS Rao LV Ezban M Petersen LC Factor VIIa-induced p44/42 mitogen-activated protein kinase activation requires the proteolytic activity of factor VIIa and is independent of the tissue factor cytoplasmic domain J Biol Chem 1999 274 21349 21354 10409695 10.1074/jbc.274.30.21349
Camerer E Huang W Coughlin SR Tissue factor- and factor X-dependent activation of protease-activated receptor 2 by factor VIIa Proc Natl Acad Sci U S A 2000 97 5255 5260 10805786 10.1073/pnas.97.10.5255
Rottingen JA Enden T Camerer E Iversen JG Prydz H Binding of human factor VIIa to tissue factor induces cytosolic Ca2+ signals in J82 cells, transfected COS-1 cells, Madin-Darby canine kidney cells and in human endothelial cells induced to synthesize tissue factor J Biol Chem 1995 270 4650 4660 7876236 10.1074/jbc.270.9.4650
Camerer E Rottingen JA Iversen JG Prydz H Coagulation factors VII and X induce Ca2+ oscillations in Madin-Darby canine kidney cells only when proteolytically active J Biol Chem 1996 271 29034 29042 8910556 10.1074/jbc.271.46.29034
Belting M Dorrell MI Sandgren S Aguilar E Ahamed J Dorfleutner A Carmeliet P Mueller BM Friedlander M Ruf W Regulation of angiogenesis by tissue factor cytoplasmic domain signaling Nat Med 2004 10 502 509 15098027 10.1038/nm1037
Cao R Bjorndahl MA Religa P Clasper S Garvin S Galter D Meister B Ikomi F Tritsaris K Dissing S Ohhashi T Jackson DG Cao Y PDGF-BB induces intratumoral lymphangiogenesis and promotes lymphatic metastasis Cancer Cell 2004 6 333 345 15488757 10.1016/j.ccr.2004.08.034
Sevinsky JR Rao LV Ruf W Ligand-induced protease receptor translocation into caveolae: a mechanism for regulating cell surface proteolysis of the tissue factor-dependent coagulation pathway J Cell Biol 1996 133 293 304 8609163 10.1083/jcb.133.2.293
Ott I Miyagi Y Miyazaki K Heeb MJ Mueller BM Rao LV Ruf W Reversible regulation of tissue factor-induced coagulation by glycosyl phosphatidylinositol-anchored tissue factor pathway inhibitor Arterioscler Thromb Vasc Biol 2000 20 874 882 10712416
Werling RW Zacharski LR Kisiel W Bajaj SP Memoli VA Rousseau SM Distribution of tissue factor pathway inhibitor in normal and malignant human tissues Thromb Haemost 1993 69 366 369 8497849
Ahamed J Belting M Ruf W Regulation of tissue factor-induced signaling by endogenous and recombinant tissue factor pathway inhibitor-1 Blood 2004
Hembrough TA Swartz GM Papathanassiu A Vlasuk GP Rote WE Green SJ Pribluda VS Tissue factor/factor VIIa inhibitors block angiogenesis and tumor growth through a nonhemostatic mechanism Cancer Res 2003 63 2997 3000 12782609
Amirkhosravi A Meyer T Chang JY Amaya M Siddiqui F Desai H Francis JL Tissue factor pathway inhibitor reduces experimental lung metastasis of B16 melanoma Thromb Haemost 2002 87 930 936 12083498
Sprecher CA Kisiel W Mathewes S Foster DC Molecular cloning, expression, and partial characterization of a second human tissue-factor-pathway inhibitor Proc Natl Acad Sci U S A 1994 91 3353 3357 8159751
Sato N Parker AR Fukushima N Miyagi Y Iacobuzio-Donahue CA Eshleman JR Goggins M Epigenetic inactivation of TFPI-2 as a common mechanism associated with growth and invasion of pancreatic ductal adenocarcinoma Oncogene 2004
Chand HS Du X Ma D Inzunza HD Kamei S Foster D Brodie S Kisiel W The effect of human tissue factor pathway inhibitor-2 on the growth and metastasis of fibrosarcoma tumors in athymic mice Blood 2004 103 1069 1077 14525759 10.1182/blood-2003-06-1930
Rao CN Lakka SS Kin Y Konduri SD Fuller GN Mohanam S Rao JS Expression of tissue factor pathway inhibitor 2 inversely correlates during the progression of human gliomas Clin Cancer Res 2001 7 570 576 11297250
Mueller BM Reisfeld RA Edgington TS Ruf W Expression of tissue factor by melanoma cells promotes efficient hematogenous metastasis Proc Natl Acad Sci U S A 1992 89 11832 11836 1465406
Mueller BM Ruf W Requirement for binding of catalytically active factor VIIa in tissue factor-dependent experimental metastasis J Clin Invest 1998 101 1372 1378 9525979
Versteeg HH Sorensen BB Slofstra SH Van den Brande JH Stam JC van Bergen en Henegouwen PM Richel DJ Petersen LC Peppelenbosch MP VIIa/tissue factor interaction results in a tissue factor cytoplasmic domain-independent activation of protein synthesis, p70, and p90 S6 kinase phosphorylation J Biol Chem 2002 277 27065 27072 12019261 10.1074/jbc.M110325200
Versteeg HH Spek CA Richel DJ Peppelenbosch MP Coagulation factors VIIa and Xa inhibit apoptosis and anoikis Oncogene 2004 23 410 417 14724569 10.1038/sj.onc.1207066
Sorensen BB Rao LV Tornehave D Gammeltoft S Petersen LC Antiapoptotic effect of coagulation factor VIIa Blood 2003 102 1708 1715 12738672 10.1182/blood-2003-01-0157
Hjortoe GM Petersen LC Albrektsen T Sorensen BB Norby PL Mandal SK Pendurthi UR Rao LV Tissue factor-factor VIIa-specific up-regulation of IL-8 expression in MDA-MB-231 cells is mediated by PAR-2 and results in increased cell migration Blood 2004 103 3029 3037 15070680 10.1182/blood-2003-10-3417
Davidson B Reich R Kopolovic J Berner A Nesland JM Kristensen GB Trope CG Bryne M Risberg B van de Putte G Goldberg I Interleukin-8 and vascular endothelial growth factor mRNA and protein levels are down-regulated in ovarian carcinoma cells in serous effusions Clin Exp Metastasis 2002 19 135 144 11964077 10.1023/A:1014582911680
Gawrychowski K Skopinska-Rozewska E Barcz E Sommer E Szaniawska B Roszkowska-Purska K Janik P Zielinski J Angiogenic activity and interleukin-8 content of human ovarian cancer ascites Eur J Gynaecol Oncol 1998 19 262 264 9641227
Mukaida N Pathophysiological roles of interleukin-8/CXCL8 in pulmonary diseases Am J Physiol Lung Cell Mol Physiol 2003 284 L566 77 12618418
Even-Ram S Uziely B Cohen P Grisaru-Granovsky S Maoz M Ginzburg Y Reich R Vlodavsky I Bar-Shavit R Thrombin receptor overexpression in malignant and physiological invasion processes Nat Med 1998 4 909 914 9701242 10.1038/nm0898-909
Grisaru-Granovsky S Salah Z Maoz M Pruss D Beller U Bar-Shavit R Differential expression of Protease activated receptor 1 (Par1) and pY397FAK in benign and malignant human ovarian tissue samples Int J Cancer 2005 113 372 378 15455382 10.1002/ijc.20607
Tsopanoglou NE Pipili-Synetos E Maragoudakis ME Thrombin promotes angiogenesis by a mechanism independent of fibrin formation Am J Physiol 1993 264 C1302 7 7684562
Tsopanoglou NE Maragoudakis ME On the mechanism of thrombin-induced angiogenesis: inhibition of attachment of endothelial cells on basement membrane components Angiogenesis 1998 1 192 200 14517385 10.1023/A:1018381822011
Yamahata H Takeshima H Kuratsu J Sarker KP Tanioka K Wakimaru N Nakata M Kitajima I Maruyama I The role of thrombin in the neo-vascularization of malignant gliomas: an intrinsic modulator for the up-regulation of vascular endothelial growth factor Int J Oncol 2002 20 921 928 11956584
Mohle R Green D Moore MA Nachman RL Rafii S Constitutive production and thrombin-induced release of vascular endothelial growth factor by human megakaryocytes and platelets Proc Natl Acad Sci U S A 1997 94 663 668 9012841 10.1073/pnas.94.2.663
Tsopanoglou NE Maragoudakis ME On the mechanism of thrombin-induced angiogenesis. Potentiation of vascular endothelial growth factor activity on endothelial cells by up-regulation of its receptors J Biol Chem 1999 274 23969 23976 10446165 10.1074/jbc.274.34.23969
Huang YQ Li JJ Hu L Lee M Karpatkin S Thrombin induces increased expression and secretion of angiopoietin-2 from human umbilical vein endothelial cells Blood 2002 99 1646 1650 11861279 10.1182/blood.V99.5.1646
Herbert JM Dupuy E Laplace MC Zini JM Bar Shavit R Tobelem G Thrombin induces endothelial cell growth via both a proteolytic and a non-proteolytic pathway Biochem J 1994 303 ( Pt 1) 227 231 7945245
Coughlin SR Thrombin signalling and protease-activated receptors Nature 2000 407 258 264 11001069 10.1038/35025229
Nierodzik ML Chen K Takeshita K Li JJ Huang YQ Feng XS D'Andrea MR Andrade-Gordon P Karpatkin S Protease-activated receptor 1 (PAR-1) is required and rate-limiting for thrombin-enhanced experimental pulmonary metastasis Blood 1998 92 3694 3700 9808563
Henrikson KP Salazar SL Fenton JW Pentecost BT Role of thrombin receptor in breast cancer invasiveness Br J Cancer 1999 79 401 406 10027305 10.1038/sj.bjc.6690063
Hughes PE Pfaff M Integrin affinity modulation Trends Cell Biol 1998 8 359 364 9728397 10.1016/S0962-8924(98)01339-7
Stenberg PE McEver RP Shuman MA Jacques YV Bainton DF A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation J Cell Biol 1985 101 880 886 2411738 10.1083/jcb.101.3.880
Henn V Slupsky JR Grafe M Anagnostopoulos I Forster R Muller-Berghaus G Kroczek RA CD40 ligand on activated platelets triggers an inflammatory reaction of endothelial cells Nature 1998 391 591 594 9468137 10.1038/35393
Nierodzik ML Plotkin A Kajumo F Karpatkin S Thrombin stimulates tumor-platelet adhesion in vitro and metastasis in vivo J Clin Invest 1991 87 229 236 1845869
Darmoul D Gratio V Devaud H Lehy T Laburthe M Aberrant expression and activation of the thrombin receptor protease-activated receptor-1 induces cell proliferation and motility in human colon cancer cells Am J Pathol 2003 162 1503 1513 12707033
Hu L Lee M Campbell W Perez-Soler R Karpatkin S Role of endogenous thrombin in tumor implantation, seeding, and spontaneous metastasis Blood 2004 104 2746 2751 15265791 10.1182/blood-2004-03-1047
Asanuma K Wakabayashi H Hayashi T Okuyama N Seto M Matsumine A Kusuzaki K Suzuki K Uchida A Thrombin inhibitor, argatroban, prevents tumor cell migration and bone metastasis Oncology 2004 67 166 173 15539922 10.1159/000081004
Liu J Schuff-Werner P Steiner M Double transfection improves small interfering RNA-induced thrombin receptor (PAR-1) gene silencing in DU 145 prostate cancer cells FEBS Lett 2004 577 175 180 15527781 10.1016/j.febslet.2004.09.079
Walz DA Fenton JW The role of thrombin in tumor cell metastasis Invasion Metastasis 1994 14 303 308 7657523
Even-Ram SC Maoz M Pokroy E Reich R Katz BZ Gutwein P Altevogt P Bar-Shavit R Tumor cell invasion is promoted by activation of protease activated receptor-1 in cooperation with the alpha vbeta 5 integrin J Biol Chem 2001 276 10952 10962 11278329 10.1074/jbc.M007027200
Vial E Sahai E Marshall CJ ERK-MAPK signaling coordinately regulates activity of Rac1 and RhoA for tumor cell motility Cancer Cell 2003 4 67 79 12892714 10.1016/S1535-6108(03)00162-4
Conner SR Scott G Aplin AE Adhesion-dependent activation of the ERK1/2 cascade is by-passed in melanoma cells J Biol Chem 2003 278 34548 34554 12821662 10.1074/jbc.M305797200
Tellez C Bar-Eli M Role and regulation of the thrombin receptor (PAR-1) in human melanoma Oncogene 2003 22 3130 3137 12789289 10.1038/sj.onc.1206453
Hazarika P McCarty MF Prieto VG George S Babu D Koul D Bar-Eli M Duvic M Up-regulation of Flotillin-2 is associated with melanoma progression and modulates expression of the thrombin receptor protease activated receptor 1 Cancer Res 2004 64 7361 7369 15492257
Booden MA Eckert LB Der CJ Trejo J Persistent signaling by dysregulated thrombin receptor trafficking promotes breast carcinoma cell invasion Mol Cell Biol 2004 24 1990 1999 14966279 10.1128/MCB.24.5.1990-1999.2004
Shi X Gangadharan B Brass LF Ruf W Mueller BM Protease-activated receptors (PAR1 and PAR2) contribute to tumor cell motility and metastasis Mol Cancer Res 2004 2 395 402 15280447
Camerer E Qazi AA Duong DN Cornelissen I Advincula R Coughlin SR Platelets, protease-activated receptors, and fibrinogen in hematogenous metastasis Blood 2004 104 397 401 15031212 10.1182/blood-2004-02-0434
Camerer E Kataoka H Kahn M Lease K Coughlin SR Genetic evidence that protease-activated receptors mediate factor Xa signaling in endothelial cells J Biol Chem 2002 277 16081 16087 11850418 10.1074/jbc.M108555200
Kataoka H Hamilton JR McKemy DD Camerer E Zheng YW Cheng A Griffin C Coughlin SR Protease-activated receptors 1 and 4 mediate thrombin signaling in endothelial cells Blood 2003 102 3224 3231 12869501 10.1182/blood-2003-04-1130
Dallabrida SM Falls LA Farrell DH Factor XIIIa supports microvascular endothelial cell adhesion and inhibits capillary tube formation in fibrin Blood 2000 95 2586 2592 10753838
Cox AD Devine DV Factor XIIIa binding to activated platelets is mediated through activation of glycoprotein IIb-IIIa Blood 1994 83 1006 1016 7906563
Barry EL Mosher DF Binding and degradation of blood coagulation factor XIII by cultured fibroblasts J Biol Chem 1990 265 9302 9307 1971626
Wilhelm O Hafter R Henschen A Schmitt M Graeff H Role of plasmin in the degradation of the stroma-derived fibrin in human ovarian carcinoma Blood 1990 75 1673 1678 2139349
Nagy JA Meyers MS Masse EM Herzberg KT Dvorak HF Pathogenesis of ascites tumor growth: fibrinogen influx and fibrin accumulation in tissues lining the peritoneal cavity Cancer Res 1995 55 369 375 7812970
Nagy JA Morgan ES Herzberg KT Manseau EJ Dvorak AM Dvorak HF Pathogenesis of ascites tumor growth: angiogenesis, vascular remodeling, and stroma formation in the peritoneal lining Cancer Res 1995 55 376 385 7529135
van de Molengraft F Jap P Poels L Croes H Vooijs P Tumor cell settling and early invasion of the peritoneum Clin Exp Metastasis 1989 7 187 199 2465862 10.1007/BF01787023
Nagy JA Masse EM Herzberg KT Meyers MS Yeo KT Yeo TK Sioussat TM Dvorak HF Pathogenesis of ascites tumor growth: vascular permeability factor, vascular hyperpermeability, and ascites fluid accumulation Cancer Res 1995 55 360 368 7812969
Nagy JA Herzberg KT Masse EM Zientara GP Dvorak HF Exchange of macromolecules between plasma and peritoneal cavity in ascites tumor-bearing, normal, and serotonin-injected mice Cancer Res 1989 49 5448 5458 2475250
Contrino J Hair G Kreutzer DL Rickles FR In situ detection of tissue factor in vascular endothelial cells: correlation with the malignant phenotype of human breast disease Nat Med 1996 2 209 215 8574967 10.1038/nm0296-209
Bootle-Wilbraham CA Tazzyman S Marshall JM Lewis CE Fibrinogen E-fragment inhibits the migration and tubule formation of human dermal microvascular endothelial cells in vitro Cancer Res 2000 60 4719 4724 10987275
Dirix LY Salgado R Weytjens R Colpaert C Benoy I Huget P van Dam P Prove A Lemmens J Vermeulen P Plasma fibrin D-dimer levels correlate with tumour volume, progression rate and survival in patients with metastatic breast cancer Br J Cancer 2002 86 389 395 11875705 10.1038/sj.bjc.6600069
Blackwell K Haroon Z Broadwater G Berry D Harris L Iglehart JD Dewhirst M Greenberg C Plasma D-dimer levels in operable breast cancer patients correlate with clinical stage and axillary lymph node status J Clin Oncol 2000 18 600 608 10653875
van Hinsbergh VW Collen A Koolwijk P Role of fibrin matrix in angiogenesis Ann N Y Acad Sci 2001 936 426 437 11460496
Clark RA Tonnesen MG Gailit J Cheresh DA Transient functional expression of alphaVbeta 3 on vascular cells during wound repair Am J Pathol 1996 148 1407 1421 8623913
Maekawa H Sato H Tollefsen DM Thrombin inhibition by HCII in the presence of elastase-cleaved HCII and thrombin-HCII complex Thromb Res 2000 100 443 451 11150588 10.1016/S0049-3848(00)00350-9
Herzog R Lutz S Blin N Marasa JC Blinder MA Tollefsen DM Complete nucleotide sequence of the gene for human heparin cofactor II and mapping to chromosomal band 22q11 Biochemistry 1991 30 1350 1357 1671335 10.1021/bi00219a027
Liu L Dewar L Song Y Kulczycky M Blajchman MA Fenton JW Andrew M Delorme M Ginsberg J Preissner KT Inhibition of thrombin by antithrombin III and heparin cofactor II in vivo Thromb Haemost 1995 73 405 412 7545318
He L Vicente CP Westrick RJ Eitzman DT Tollefsen DM Heparin cofactor II inhibits arterial thrombosis after endothelial injury J Clin Invest 2002 109 213 219 11805133 10.1172/JCI200213432
Takamori N Azuma H Kato M Hashizume S Aihara K Akaike M Tamura K Matsumoto T High plasma heparin cofactor II activity is associated with reduced incidence of in-stent restenosis after percutaneous coronary intervention Circulation 2004 109 481 486 14744972 10.1161/01.CIR.0000109695.39671.37
Aihara K Azuma H Takamori N Kanagawa Y Akaike M Fujimura M Yoshida T Hashizume S Kato M Yamaguchi H Kato S Ikeda Y Arase T Kondo A Matsumoto T Heparin cofactor II is a novel protective factor against carotid atherosclerosis in elderly individuals Circulation 2004 109 2761 2765 15148272 10.1161/01.CIR.0000129968.46095.F3
Pratt CW Tobin RB Church FC Interaction of heparin cofactor II with neutrophil elastase and cathepsin G J Biol Chem 1990 265 6092 6097 2318847
Church FC Pratt CW Hoffman M Leukocyte chemoattractant peptides from the serpin heparin cofactor II J Biol Chem 1991 266 704 709 1985958
Simmonds RE Lane DA Structural and functional implications of the intron/exon organization of the human endothelial cell protein C/activated protein C receptor (EPCR) gene: comparison with the structure of CD1/major histocompatibility complex alpha1 and alpha2 domains Blood 1999 94 632 641 10397730
Kurosawa S Stearns-Kurosawa DJ Hidari N Esmon CT Identification of functional endothelial protein C receptor in human plasma J Clin Invest 1997 100 411 418 9218519
Fukudome K Ye X Tsuneyoshi N Tokunaga O Sugawara K Mizokami H Kimoto M Activation mechanism of anticoagulant protein C in large blood vessels involving the endothelial cell protein C receptor J Exp Med 1998 187 1029 1035 9529319 10.1084/jem.187.7.1029
Joyce DE Gelbert L Ciaccia A DeHoff B Grinnell BW Gene expression profile of antithrombotic protein c defines new mechanisms modulating inflammation and apoptosis J Biol Chem 2001 276 11199 11203 11278252 10.1074/jbc.C100017200
Van de Wouwer M Collen D Conway EM Thrombomodulin-protein C-EPCR system: integrated to regulate coagulation and inflammation Arterioscler Thromb Vasc Biol 2004 24 1374 1383 15178554 10.1161/01.ATV.0000134298.25489.92
Cheng T Liu D Griffin JH Fernandez JA Castellino F Rosen ED Fukudome K Zlokovic BV Activated protein C blocks p53-mediated apoptosis in ischemic human brain endothelium and is neuroprotective Nat Med 2003 9 338 342 12563316 10.1038/nm826
Riewald M Petrovan RJ Donner A Mueller BM Ruf W Activation of endothelial cell protease activated receptor 1 by the protein C pathway Science 2002 296 1880 1882 12052963 10.1126/science.1071699
Esmon CT Structure and functions of the endothelial cell protein C receptor Crit Care Med 2004 32 S298 301 15118534 10.1097/01.CCM.0000126128.64614.81
Kurosawa S Esmon CT Stearns-Kurosawa DJ The soluble endothelial protein C receptor binds to activated neutrophils: involvement of proteinase-3 and CD11b/CD18 J Immunol 2000 165 4697 4703 11035113
Saposnik B Reny JL Gaussem P Emmerich J Aiach M Gandrille S A haplotype of the EPCR gene is associated with increased plasma levels of sEPCR and is a candidate risk factor for thrombosis Blood 2004 103 1311 1318 14576048 10.1182/blood-2003-07-2520
Medina P Navarro S Estelles A Vaya A Woodhams B Mira Y Villa P Migaud-Fressart M Ferrando F Aznar J Bertina RM Espana F Contribution of polymorphisms in the endothelial protein C receptor gene to soluble endothelial protein C receptor and circulating activated protein C levels, and thrombotic risk Thromb Haemost 2004 91 905 911 15116250
Kurosawa S Stearns-Kurosawa DJ Carson CW D'Angelo A Della Valle P Esmon CT Plasma levels of endothelial cell protein C receptor are elevated in patients with sepsis and systemic lupus erythematosus: lack of correlation with thrombomodulin suggests involvement of different pathological processes Blood 1998 91 725 727 9427734
Iwaki T Cruz DT Martin JA Castellino FJ A cardioprotective role for the endothelial protein C receptor in lipopolysaccharide-induced endotoxemia in the mouse Blood 2004
Scheffer GL Flens MJ Hageman S Izquierdo MA Shoemaker RH Scheper RJ Expression of the vascular endothelial cell protein C receptor in epithelial tumour cells Eur J Cancer 2002 38 1535 1542 12110501 10.1016/S0959-8049(02)00108-9
Tsuneyoshi N Fukudome K Horiguchi S Ye X Matsuzaki M Toi M Suzuki K Kimoto M Expression and anticoagulant function of the endothelial cell protein C receptor (EPCR) in cancer cell lines Thromb Haemost 2001 85 356 361 11246560
Johnson K Choi Y DeGroot E Samuels I Creasey A Aarden L Potential mechanisms for a proinflammatory vascular cytokine response to coagulation activation J Immunol 1998 160 5130 5135 9590265
Ludeman MJ Kataoka H Srinivasan Y Esmon N Esmon CT Coughlin SR PAR1 cleavage and signaling in response to activated protein C and thrombin J Biol Chem 2005
Freedman RS Kudelka AP Kavanagh JJ Verschraegen C Edwards CL Nash M Levy L Atkinson EN Zhang HZ Melichar B Patenia R Templin S Scott W Platsoucas CD Clinical and biological effects of intraperitoneal injections of recombinant interferon-gamma and recombinant interleukin 2 with or without tumor-infiltrating lymphocytes in patients with ovarian or peritoneal carcinoma Clin Cancer Res 2000 6 2268 2278 10873077
Freedman RS Edwards CL Kavanagh JJ Kudelka AP Katz RL Carrasco CH Atkinson EN Scott W Tomasovic B Templin S Intraperitoneal adoptive immunotherapy of ovarian carcinoma with tumor-infiltrating lymphocytes and low-dose recombinant interleukin-2: a pilot trial J Immunother Emphasis Tumor Immunol 1994 16 198 210 7834119
Freedman RS Tomasovic B Templin S Atkinson EN Kudelka A Edwards CL Platsoucas CD Large-scale expansion in interleukin-2 of tumor-infiltrating lymphocytes from patients with ovarian carcinoma for adoptive immunotherapy J Immunol Methods 1994 167 145 160 8308273 10.1016/0022-1759(94)90084-1
Zhang L Conejo-Garcia JR Katsaros D Gimotty PA Massobrio M Regnani G Makrigiannakis A Gray H Schlienger K Liebman MN Rubin SC Coukos G Intratumoral T cells, recurrence, and survival in epithelial ovarian cancer N Engl J Med 2003 348 203 213 12529460 10.1056/NEJMoa020177
Nash MA Lenzi R Edwards CL Kavanagh JJ Kudelka AP Verschraegen CF Platsoucas CD Freedman RS Differential expression of cytokine transcripts in human epithelial ovarian carcinoma by solid tumour specimens, peritoneal exudate cells containing tumour, tumour-infiltrating lymphocyte (TIL)-derived T cell lines and established tumour cell lines Clin Exp Immunol 1998 112 172 180 9649178 10.1046/j.1365-2249.1998.00576.x
Rabinowich H Suminami Y Reichert TE Crowley-Nowick P Bell M Edwards R Whiteside TL Expression of cytokine genes or proteins and signaling molecules in lymphocytes associated with human ovarian carcinoma Int J Cancer 1996 68 276 284 8903466 10.1002/(SICI)1097-0215(19961104)68:3<276::AID-IJC2>3.0.CO;2-Z
Curiel TJ Coukos G Zou L Alvarez X Cheng P Mottram P Evdemon-Hogan M Conejo-Garcia JR Zhang L Burow M Zhu Y Wei S Kryczek I Daniel B Gordon A Myers L Lackner A Disis ML Knutson KL Chen L Zou W Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival Nat Med 2004 10 942 949 15322536 10.1038/nm1093
Schutyser E Struyf S Proost P Opdenakker G Laureys G Verhasselt B Peperstraete L Van de Putte I Saccani A Allavena P Mantovani A Van Damme J Identification of biologically active chemokine isoforms from ascitic fluid and elevated levels of CCL18/pulmonary and activation-regulated chemokine in ovarian carcinoma J Biol Chem 2002 277 24584 24593 11978786 10.1074/jbc.M112275200
Negus RP Stamp GW Relf MG Burke F Malik ST Bernasconi S Allavena P Sozzani S Mantovani A Balkwill FR The detection and localization of monocyte chemoattractant protein-1 (MCP-1) in human ovarian cancer J Clin Invest 1995 95 2391 2396 7738202
Negus RP Stamp GW Hadley J Balkwill FR Quantitative assessment of the leukocyte infiltrate in ovarian cancer and its relationship to the expression of C-C chemokines Am J Pathol 1997 150 1723 1734 9137096
Milliken D Scotton C Raju S Balkwill F Wilson J Analysis of chemokines and chemokine receptor expression in ovarian cancer ascites Clin Cancer Res 2002 8 1108 1114 11948121
Heinzelmann-Schwarz VA Gardiner-Garden M Henshall SM Scurry J Scolyer RA Davies MJ Heinzelmann M Kalish LH Bali A Kench JG Edwards LS Vanden Bergh PM Hacker NF Sutherland RL O'Brien PM Overexpression of the cell adhesion molecules DDR1, Claudin 3, and Ep-CAM in metaplastic ovarian epithelium and ovarian cancer Clin Cancer Res 2004 10 4427 4436 15240533
Opala T Drews K Rzymski P Wozniak J Sajdak S Evaluation of soluble intracellular adhesion molecule-1 (sICAM-1) in benign and malignant ovarian masses Eur J Gynaecol Oncol 2003 24 255 257 12807235
Esmon CT Inflammation and thrombosis J Thromb Haemost 2003 1 1343 1348 12871267 10.1046/j.1538-7836.2003.00261.x
Esmon CT Coagulation and inflammation J Endotoxin Res 2003 9 192 198 12831462 10.1179/096805103125001603
Esmon CT Interactions between the innate immune and blood coagulation systems Trends Immunol 2004 25 536 542 15364056 10.1016/j.it.2004.08.003
Esmon CT Crosstalk between inflammation and thrombosis Maturitas 2004 47 305 314 15063484 10.1016/j.maturitas.2003.10.015
Giesen PL Rauch U Bohrmann B Kling D Roque M Fallon JT Badimon JJ Himber J Riederer MA Nemerson Y Blood-borne tissue factor: another view of thrombosis Proc Natl Acad Sci U S A 1999 96 2311 2315 10051638 10.1073/pnas.96.5.2311
Rauch U Bonderman D Bohrmann B Badimon JJ Himber J Riederer MA Nemerson Y Transfer of tissue factor from leukocytes to platelets is mediated by CD15 and tissue factor Blood 2000 96 170 175 10891447
Esmon CT The impact of the inflammatory response on coagulation Thromb Res 2004 114 321 327 15507261 10.1016/j.thromres.2004.06.028
Sims PJ Faioni EM Wiedmer T Shattil SJ Complement proteins C5b-9 cause release of membrane vesicles from the platelet surface that are enriched in the membrane receptor for coagulation factor Va and express prothrombinase activity J Biol Chem 1988 263 18205 18212 2848029
Gear AR Camerini D Platelet chemokines and chemokine receptors: linking hemostasis, inflammation, and host defense Microcirculation 2003 10 335 350 12851650 10.1038/sj.mn.7800198
Miller DL Yaron R Yellin MJ CD40L-CD40 interactions regulate endothelial cell surface tissue factor and thrombomodulin expression J Leukoc Biol 1998 63 373 379 9500526
Fujisawa T Fujisawa R Kato Y Nakayama T Morita A Katsumata H Nishimori H Iguchi K Kamiya H Gray PW Chantry D Suzuki R Yoshie O Presence of high contents of thymus and activation-regulated chemokine in platelets and elevated plasma levels of thymus and activation-regulated chemokine and macrophage-derived chemokine in patients with atopic dermatitis J Allergy Clin Immunol 2002 110 139 146 12110833 10.1067/mai.2002.126079
Fukudome K Esmon CT Identification, cloning, and regulation of a novel endothelial cell protein C/activated protein C receptor J Biol Chem 1994 269 26486 26491 7929370
Conway EM Rosenberg RD Tumor necrosis factor suppresses transcription of the thrombomodulin gene in endothelial cells Mol Cell Biol 1988 8 5588 5592 2854203
Gould MK Dembitzer AD Doyle RL Hastie TJ Garber AM Low-molecular-weight heparins compared with unfractionated heparin for treatment of acute deep venous thrombosis. A meta-analysis of randomized, controlled trials Ann Intern Med 1999 130 800 809 10366369
Abildgaard U Heparin/low molecular weight heparin and tissue factor pathway inhibitor Haemostasis 1993 23 Suppl 1 103 106 8388347
Green D Hull RD Brant R Pineo GF Lower mortality in cancer patients treated with low-molecular-weight versus standard heparin Lancet 1992 339 1476 1351142 10.1016/0140-6736(92)92064-M
Siragusa S Cosmi B Piovella F Hirsh J Ginsberg JS Low-molecular-weight heparins and unfractionated heparin in the treatment of patients with acute venous thromboembolism: results of a meta-analysis Am J Med 1996 100 269 277 8629671 10.1016/S0002-9343(97)89484-3
Hettiarachchi RJ Smorenburg SM Ginsberg J Levine M Prins MH Buller HR Do heparins do more than just treat thrombosis? The influence of heparins on cancer spread Thromb Haemost 1999 82 947 952 10605808
Cosgrove RH Zacharski LR Racine E Andersen JC Improved cancer mortality with low-molecular-weight heparin treatment: a review of the evidence Semin Thromb Hemost 2002 28 79 87 11885028 10.1055/s-2002-20566
Kakkar AK Williamson RC Antithrombotic therapy in cancer Bmj 1999 318 1571 1572 10364099
Zacharski LR Ornstein DL Mamourian AC Low-molecular-weight heparin and cancer Semin Thromb Hemost 2000 26 Suppl 1 69 77 11011810 10.1055/s-2000-9499
Wojtukiewicz MZ Kozlowski L Ostrowska K Dmitruk A Zacharski LR Low molecular weight heparin treatment for malignant melanoma: a pilot clinical trial Thromb Haemost 2003 89 405 407 12574822
Robert F Busby E Marques MB Reynolds RE Carey DE Phase II study of docetaxel plus enoxaparin in chemotherapy-naive patients with metastatic non-small cell lung cancer: preliminary results Lung Cancer 2003 42 237 245 14568692 10.1016/S0169-5002(03)00354-4
Zacharski LR Henderson WG Rickles FR Forman WB Cornell CJJ Forcier RJ Edwards RL Headley E Kim SH O'Donnell JF Effect of warfarin anticoagulation on survival in carcinoma of the lung, colon, head and neck, and prostate. Final report of VA Cooperative Study #75 Cancer 1984 53 2046 2052 6322957
Kakkar AK Levine MN Kadziola Z Lemoine NR Low V Patel HK Rustin G Thomas M Quigley M Williamson RCN Low Molecular Weight Heparin, Therapy With Dalteparin, and Survival in Advanced Cancer: The Fragmin Advanced Malignancy Outcome Study (FAMOUS) J Clin Oncol 2004 22 1944 1948 15143088 10.1200/JCO.2004.10.002
Lee AYY Rickles FR Julian JA Gent M Baker RI Bowden C Kakkar AK Prins M Levine MN Randomized Comparison of Low Molecular Weight Heparin and Coumarin Derivatives on the Survival of Patients With Cancer and Venous Thromboembolism J Clin Oncol 2005 JCO.2005.03.133
Schulman S Lindmarker P Incidence of cancer after prophylaxis with warfarin against recurrent venous thromboembolism. Duration of Anticoagulation Trial N Engl J Med 2000 342 1953 1958 10874063 10.1056/NEJM200006293422604
|
15969748
|
PMC1182397
|
CC BY
|
2021-01-04 16:39:28
|
no
|
J Transl Med. 2005 Jun 21; 3:25
|
utf-8
|
J Transl Med
| 2,005 |
10.1186/1479-5876-3-25
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-361596703810.1186/1477-7819-3-36ReviewGoblet cell carcinoid of the appendix Pahlavan Payam S [email protected] Rani [email protected] Department of Physiology and Pathophysiology, University of Heidelberg, Heidelberg, Germany2 Department of Pathology, University of Saskatchewan, Saskatoon, Canada2005 20 6 2005 3 36 36 16 2 2005 20 6 2005 Copyright © 2005 Pahlavan and Kanthan; licensee BioMed Central Ltd.2005Pahlavan and Kanthan; 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
Goblet cell carcinoid (GCC) of the appendix is a rare neoplasm that share histological features of both adenocarcinoma and carcinoid tumor. While its malignant potential remains unclear, GCC's are more aggressive than conventional carcinoid. The clinical presentations of this neoplasm are also varied. This review summarizes the published literature on GCC of the appendix. The focus is on its diagnosis, histopathological aspects, clinical manifestations, and management.
Methods
Published studies in the English language between 1966 to 2004 were identified through Medline keyword search utilizing terms "goblet cell carcinoid," "adenocarcinoid", "mucinous carcinoid" and "crypt cell carcinoma" of the appendix.
Results
Based on the review of 57 published papers encompassing nearly 600 diagnosed patients, the mean age of presentation for GCC of the appendix was 58.89 years with equal representation in both males and females. Accurate diagnosis of this neoplasm requires astute observations within an acutely inflamed appendix as this neoplasm has a prominent pattern of submucosal growth and usually lacks the formation of a well-defined tumor mass. The mesoappendix was involved in 21.64% followed by perineural involvement in 2.06%. The most common clinical presentations in order of frequency were acute appendicitis in 22.5%; asymptomatic in 5.4%; non-localized abdominal pain in 5.15% and an appendicular mass in 3.09%. The most common surgical treatment of choice was appendectomy with right hemicolectomy in 34.70% followed by simple appendectomy in 24.57%. Concomitant distant metastasis at diagnosis was present in 11.16% of patients with the ovaries being the most common site in 3.60% followed by disseminated abdominal carcinomatosis in 1.03%. Local lymph node involvement was seen in 8.76% of patients at the time of diagnosis. The reported 5-year survival ranges from 60 % to 84%. GCC's of the appendix remains a neoplasm of unpredictable biological behavior and thus warrants lifelong surveillance for recurrence of the disease upon diagnosis and successful surgical extirpation.
Conclusion
GCC of the appendix is a rare neoplasm. Due to its wide range of presentation, this tumor should be considered as a possible diagnosis in many varied situations leading to abdominal surgery. Histopathological features such as increased number of Paneth cells, increased amount of mucin secretion and presence of pancreatic polypeptide may predict a more aggressive behavior. The advocated plan of management recommended for patients with tumors that involve the adjacent caecum or with high-grade tumors with histological features such as an increased mitotic rate involve initial appendectomy with completion right hemicolectomy due to the high possibility of local recurrence with intraperitoneal seeding prior to lymph node involvement and a 20% risk of metastatic behavior. In female patients with GCC of the appendix regardless of age, bilateral salpingo-oophorectomy is advocated. In cases with obvious spread of the disease chemotherapy, mostly with 5-FU and leucovorin is advised. Cytoreductive surgery with adjuvant intraperitoneal chemotherapy can offer improved survival in cases with advanced peritoneal dissemination.
==== Body
Background
Cancer of the appendix is an uncommon disease [1]. Appendiceal carcinoids however are found in 1 out of every 300 appendectomies. Since more than 30 years ago, a new variant type of epithelial tumor of appendix has been recognized [2,3]. This tumor is reported under different names including goblet cell carcinoid (GCC), adenocarcinoid, mucinous carcinoid, intermediate type of carcinoid, crypt cell carcinoma, amphicrine (endo-exocrine) neoplasia, composite tumor and microglandular carcinoma [2,4,5]. All names except GCC have been omitted from the current World Health Organization (WHO) classification [6]. In general, GCC (International Classification of Disease for Oncology, ICD-O 8243/3) is a rare neuroendocrine tumor of the vermiform appendix, comprising approximately 6% of appendiceal carcinoids and usually occurs in the pure form [7-9]. Some characteristics including its intramural position in the appendix, the continuity with the basiglandular cells of the mucus membrane, and it's generally well-differentiated suggesting its relation to carcinoid [10]. Nevertheless, some authors say that this tumor is far more closely related to intestinal crypts than to carcinoids and insist on not considering this tumor as a variant of carcinoids [11]. GCC has both endocrine and glandular differentiation [9]. This tumor appears to combine features of epithelial and carcinoid neoplasms and in addition the surface mucosal epithelium is not neoplastic [12]. Furthermore, in comparison to the other carcinoid tumors, GCC is usually larger in size [12].
Subbuswamy et al described the first report of GCC in 1974 [10]. This type of neoplasm predominantly occurs in the appendix as a distinct pathological entity [13,14] and has uncertain distinct biological behavior [15]. The histology and biology appears to be intermediate between carcinoid tumors and adenocarcinomas [16]. Some authors believe that this tumor usually presents as a low-grade malignancy and metastases are uncommon [17]. Yet, in general, it is believed that it has a more aggressive natural history than classic appendiceal carcinoids with variable malignant potential; thus, it requires a different surgical approach than is usually advocated for classic carcinoids [3]. The tumor most commonly infiltrates the entire appendix transmurally [6]. Anatomically, GCC's are usually located near the tip of the appendix [18]. Preoperative diagnosis is rare and patients usually present with acute appendicitis [2]. The GCC cells are probably APUD (amine precursor uptake and decarboxylation) derived cells with both mucinous and entrochromafin granules [19]. The number of patients who were diagnosed with this neoplasm based on the Medline search is less than 600 worldwide. The purpose of this study is to review the literature regarding this rare neoplasm to identify diagnostic, clinical, histological features, treatment guidelines and understand the prognosis that is somehow different from the usual appendiceal carcinoid.
History of GCC of the appendix
Appendiceal tumors that share the histological characteristics of carcinoids and adenocarcinomas have been recognized since the 1960s, with early reports by Gibbs, Bates and Butler, Rosai and Rodriguez [5]. In 1969, Gagne et al described 3 unusual tumors of appendix with common features of 1) association of nests of entrochromafin cells with mucus secreting glandular structures; 2) integrity of the appendiceal mucosa, and 3) an infiltrative pattern similar to carcinoid tumors but with the propensity to invade nerves [17]. In 1974, Klein reported three cases of mucinous carcinoid tumor of appendix [20] but the first reported cases under the name of GCC were by Subbuswamy et al in 1974 [10]. After that Abt et al reported microscopical demonstration of the cells of the other known cases with GCC of the appendix [21]. DiPaola et al reported the first case of appendico-vesical fistula due to appendix abscess with GCC in 1976 [22]. In 1978 Warkel et al reported 39 cases of GCC and revealed more aspects about this neoplasm [23]. In 1979, Warner et al described the origin of these cells as being endodermal in origin and explained that these cells arise from crypt base stem cells [19]. In 1980, Ratzenhofer et al explained this neoplasm as amphicrine (endo-exocrine) neoplasia [4]. Recent reports are more focused on the genetic aspects and molecular levels of the pathogenesis of this neoplasm [16,7,24]. Due to the paucity of reported cases and non-specific presentations e.g., acute appendicitis, diagnosis of this neoplasm requires a high degree of astute observations in an otherwise inflamed appendix. This is primarily attributed to its predominant submucosal growth and lack of forming a well-defined mass within the appendix.
Epidemiologic characteristics and etiopathogenetic factors
GCC accounts for less than 5% of primary tumors of the appendix [25]. Based on a review on 227 patients with GCC of the appendix, GCC occurs equally in men and women and with higher occurrence in the white race [1]. Aizawa et al explained that the mean age for GCC is 58.8 years, whereas that for carcinoid tumor is 35.9 years [26]. The mean age at diagnosis of GCC is between malignant carcinoid (38 years) and mucinous adenocarcinoma (60 years) [1]. The mean age for those presenting with more localized tumor is significantly lower, compared to the mean age of those with spread beyond the appendix [27]. In addition, the mean age of goblet cell type is significantly higher than the tubular carcinoid of the appendix [28]. In contrast to appendiceal carcinoids that metastasize in 2 to 5% of cases, GCCs metastasize in 15 to 30% of cases [25].
GCCs of the appendix are amphicrine tumors arising from a pluripotent cell with divergent neuroendocrine and mucinous differentiation and with dysregulation of the cell cycle [15]. There is a large reserve G0/G1 pool of pluripotent cells that remain noninvasive until further genetic instability is acquired by the cell [15]. Some studies demonstrated that p53 mutations play a major role in pathogenesis of some GCCs and most of these mutations are G:C to A:T transitions [16]. However, other studies failed to demonstrate such alterations [8]. Allelic loss of chromosomes 11q, 16q, and 18q is frequent in GCC and ileal carcinoids and maybe has a role in tumorigenesis [8]. There were also three reported cases of GCC of the appendix with concomitant appendiceal epithelial neoplasm. Therefore it is assumed that their maybe a common etiologic factor for both GCC and epithelial neoplasm in these cases with two components and this may be attributed to a divergent differentiation from a single stem cell [12]. If this tumor were a true member of the carcinoid family, then its co-existence with mucinous cysadenoma in two reported rare cases would support the unitary stem cell hypothesis of its origin. If this tumor is a carcinoma of crypt cell origin rather than a carcinoid, then combination with cystadenocarcinoma is an example of adenoma-carcinoma sequence that is extremely rare [29].
In conclusion, based on the review of all published papers for nearly 600 diagnosed patients, the mean age for diagnosing GCC was 58.89 years with equal representation in both genders with male to female ratio of 1:0.84. GCC of the appendix occurs later in life in comparison to conventional carcinoid. The etiopathogenic factor that most authors agree is the occurrence of p53 mutations and this is confirmed in most of the reported analyzed cases.
Histopathological aspects
Although GCC was previously considered a variant of carcinoid tumor, current evidence suggests that GCC is a distinct tumor with a histogenesis different from that of typical carcinoid [8]. Appendiceal neoplasms that exhibit aggressive behavior contain the typical nests of goblet cells that are the essential elements for inclusion in the category of GCC [23]. As immunohistochemical studies have shown that GCC cells contain lysozyme, secretory component (SC) and immunoglobulin A, it suggests that these tumors are derived from cells similar to the lysozyme-producing cells of the small intestinal crypts [27]. The GCC cells are rather larger than normal goblet cells [10] and arise from a pluripotent cell (stem cell) as intestinal APUD cell with endodermal origin with divergent neuroendocrine and mucinous differentiation [3].
GCCs represent a composite biphasic neoplasm of two morphologically differentiated cells, each with a common endodermal embryogenic origin [2]. In early reports, the GCCs were subdivided into two types: the goblet cell type and the tubular type [23]. Goblet cell type is more common than tubular type [30]. Goblet cell type is composed of smooth-bordered nests, usually 4 or more cells. The tubular type is composed of small, discrete acini or tubules lined by a single layer of cuboidal or columnar cells. Both types possess the essential criteria for classification as GCC [23]. The tubular type is more closely related than goblet cell type to the usual appendiceal carcinoid in both histological and prognostic characteristics and is probably an evolutionary stage in the development of the goblet cell type [23]. These cells are located below the crypts of Lieberkühn and usually do not involve the mucosa [3]. Cells are usually uniform with pattern of infiltration being nests, rosettes, or clumps without a distinct lumen [2]. As the cells are distended with mucin, they have crescentric nuclei arranged around the periphery of the clumps, resembling goblet cells or signet-ring cells [2]. A few cells are multinucleated [23]. A high cellular proliferation rate with dysregulation of the cell cycle and minimal cytologic atypia is reported in most studies [15,16,29,19]. Some studies do mention that the mitotic rate is low [10,23].
The two types of granules (neuroendocrine and mucinous) that normally exist in GCC cells are not mixed with one another in the same cell and are completely separate [31,21] though in one type, the granules are membrane-bound and sometimes fused (confluent) with adjacent granules, constituting a loose reticular matrix [32]. Electron microscopy shows pleomorphic granules including mucinous vacuoles of varying sizes and membrane-bound neuroendocrine molecules [15,19]. Because of copious production of mucin, which is predominantly acid in type, the cells are strongly positive for mucicarmine, perodic acid-Schiff diastase, and alician blue [15,22]. This mucin is a nonsulfated acid mucopolysaccaride and resistant to sialidase digestion [23]. The degree of mucin production ranges from cells distended by coalescent mucin granules to cells with variable numbers of smaller, dark granules [33].
Because GCC is considered to be derived from crypt cells, there are other immunohistochemical techniques for their diagnosis including Cytokeratin 20 (CK20), CK7 (with the positivity of CK20 more than CK7), neuron-specific enolase (NSE), chromogranin A, serotonin, lysozyme, PGP 9.5, IgA and vimentin [7,14,24,27,33,34]. In general, because of some neuroendocrine granules in the cytoplasm, the cells are reactive for NSE, chromogranin A, Synaptophysin, Gerimelius stain, Fontana-Masson stain, serotonin, substance P, peptide YY, glucagon alpha-chain and S-100 protein [26,35]. All GCCs display extensive SC (secretory component) reactivity [35]. It is important to mention that synthesis of SC and transport of IgA are not functions of APUD cells and probably the source of this IgA is from circulating dimeric IgA synthesized by lamina propria plasma cells [36]. In one study, some tumors showed weakly positive staining with antiserum to somatostatin [27]. The presence of transitional or mixed endo/exocrine cells can be proved only in semi thin sections stained with a combination of the immunocytochemical reaction for 5-HT and alician-blue [35]. It is also important to mention that the reaction for NSE is somewhat weaker in GCC than in typical carcinoids [35]. Some eosinophilic (and phloxinophilic) granules, similar in size to basal granules of crypt base cells, i.e., Kulchitsky cells, are present in basal aspects of some cells [19].
Carcinoembryonic antigen (CEA) is one of the markers that are diffusely positive on the plasma membrane of tumor cells in GCC [7]. Höfler et al mentioned that SC-, lyzozyme- and CEA-reactivity was strictly limited to the exocrine tumor areas and was not found in endocrine cells of the normal mucosa [35]. In one case with bilateral ovarian metastasis, tumor cells were strongly positive for CEA and CA19-9 [37]. Tumor cells of GCC of the appendix show a preserved expression of beta-catenin and a reduced expression of E-cadherin, compared to the expression of these proteins in normal mucosa [7]. Most cells are positive for synaptophysin [15] and anti p53 antibody [38]. Sometimes expression of cell proliferation markers p53, cyclin D1 and p21 are increased while others such as p16 is down regulated. [15].
The characteristics of GCCs that somewhat simulate ordinary carcinoids are: 1) intramural position; 2) continuity with basiglandular cells of the mucous membrane; and 3) generally well-differentiated nature as well as the presence of goblet cells, argentaffin or argyrophil cells and Paneth cells in varying proportions [2]. Normally goblet cells develop independently of Paneth cells but in special circumstances, Paneth cells might represent a reserve from which goblet cells could develop [2]. Paneth cells can be seen in between one third and one half of cases [33]. Paneth cell differentiation is predominant in tumors that metastasize to the ovaries [16] and both argyrophil positive cells and Paneth cells can be seen in lymph nodes and small bowel metastases [33] but only one study mentioned that Paneth cells were not found in the metastatic deposits [23]. In addition, Paneth cells are located at the periphery of cell nests [31]. Paneth cells are not the major cells in gastrointestinal mucin-producing adenocarcinoma but they are prominent in GCC of the appendix [18]. One study mentioned that the presence of Paneth cells, like APUD cells can be as part of the neoplastic process or because of the migration [36]. Argentaffin cells are one of the cell types in the GCC cellular mass that in addition to goblet cells proliferate during tumor growth and both cell types are invasive [19]. Argyrophil cells are found in nearly 88% of GCCs and argentaffin cells in nearly 50% of GCCs [23]. The demonstration of argentaffin or argyrophil cells in a high percentage supports the classification of GCC within the spectrum of carcinoids. However, the mere presence of these cells in a tumor is not pathognomonic for carcinoid. Argentaffin and argyrophil cells are also seen in gastric and intestinal adenocarcinoma and cystadenocarcinoma of the appendix, although these cells are not as numerous as GCCs [23]. The Hamperl-Masson and diazo reactions for argentaffin granules are positive predominantly in supranuclear position in the nonvacuolated cells [21]. In addition, the argentaffin cells also react positively to the Sevier-Munger argyrophil stain [20]. The Churukian-Schenk and Grimelius staining have been used for argyrophil granules too [39,5]. In general, approximately three-fourths of the neoplastic cell nests have sharply defined cytoplasmic areas that contain highly electron-dense, pleomorphic enterochromaffin granules (EG]. Enterochromaffin cells are situated basally in the cell nest and have processes that extend either circumferentially or, less commonly, a short distance toward the center of the cell nest. Because of the processes, the number of these cells per cell nest may be overestimated when viewed by light microscopy and some nests do not contain enterochromaffin cells [31].
In summary, GCCs are derived from undifferentiated stem cells and are different from typical carcinoids that originate from endocrine cells in the mucosal stroma [35] and the degree of integration of the goblet cells versus APUD cells exists across a spectrum varying from pure GCC to pure carcinoid tumor [32]. GCC cells have two type of granules which are not mixed and are mainly acid mucinous and can be recognized by different histochemical staining.
When microscopically evaluated, atypical cells show a prominent submucosal growth pattern with diffuse infiltration from the submucosal layer to the subserosal layer forming a microlumen structure. The growth pattern of atypical cells is within the lamina propria and extends through the muscularis propria into the subsequent serosa without destroying the appendiceal structure [26]. Generally, GCC affects the base of the glands and infiltrates the lamina propria and other layers [18]. In addition, in electron microscopy, there are equal number of elongated cells containing many mitochondria and electron dense granules compressed between goblet cells. The cell membrane of both goblet cells and elongated cells shows complex interlocking of villous processes 90–100 nm in width [19]. In general, elongated cell nests are oriented parallel to the muscle that they invade and some branching can occasionally being observed. Many of these cells are invested by variable layers of fibroblasts [19]. In the mucosa, single endocrine cells can be observed incorporated in the exocrine mucus producing tissue. They are located at the periphery of tumor cell clusters, making long cell processes [35]. Höfler et al explained features of GCC as 1) isomorphous tumor structure, 2) the presence of adenoid tumor cell groups in the intestinal wall, 3) intact mucosa overlying the tumor, and 4) the presence of endocrine cells [35].
Macroscopically, in most cases, there is no well-defined mass. There maybe areas of indurations in the wall of the appendix or stenosis of the lumen with diffuse fibrous thickening of the appendix, which draw attention to the possibility of tumor [10,40]. The depth of bowel wall invasion is defined as follows: T1, tumor invading the submucosa; T2, tumor invading the muscularis propria; T3, tumor invading through the muscularis propria into the subserosa; and T4, tumor perforating the visceral peritoneum or directly invading other organs or structures [41]. The hallmark of GCC is the presence of individual glands separated by smooth muscle or stroma and the lining cells contain intracytoplasmic mucin [28]. This tumor is characterized by its tendency to arise within the lower lamina propria and to infiltrate extensively through the appendiceal wall without much tissue destruction [40]. Architecturally, GCC arises from the base of the glands and spares the luminal mucosa. Unless perforation has occurred, the outer muscular coats of the appendix are relatively well preserved [5].
Regarding the anatomical location of the tumor, one study mentioned that the tumor was present in 70% in the distal segment of the appendix [42]. Regarding peri-appendiceal involvement another study mentioned that more than 50% serosa or mesoappendix was involved, and in more than 60% the lymph nodes were not involved [1]. Generally, diffuse infiltration into the peri-appendiceal fat and perineural invasion is seen in most cases [15]. The tumor cells are present in the muscle layers and serosa of the appendix, in deeper layers of the appendix, the tumors are more distended, with mucin causing a ballooning effect and sometimes there are pools of extracellular mucin in the muscular coat [10]. The tumor with the most differentiated histological pattern is closest to the lumen in all cases [10]. In some cases the only gross feature is a localized yellow to grayish-white, soft or gelatinous appendiceal thickening [23]. The growth of this tumor may be retarded by fibrosis and muscle spasm that impairs the blood supply with some degeneration of tumor islands [19]. Exclusive of tumors that metastasize, the lesions for the most parts are uniform in size and shape [23].
The majority of the metastasizing GCCs extend to the serosa of the appendix though serosal involvement is not necessarily an ominous finding [23]. Metastases usually do not occur from any GCC's that does not extend beyond the muscularis propria [23]. In some cases where metastasis occurs, immunoreactivity for serotonin is seen only in the primary tumor, but not in the metastasis, and therefore it is hypothesized that that the mucinous component might have a selective ability to metastasize [43].
Review of all diagnosed patients has shown that the mesoappendix was involved in 21.64% of cases and perineural involvement as the second most common tumoral extension was seen in 2.06% of all patients.
Clinical manifestations of GCC
Clinical diagnosis of GCC is seldom made preoperatively. Most of the patients present with signs and symptoms of an acute appendicitis due to luminal obstruction [2,6]. The tumor cells proliferate sparsely and do not form nodules. The appendiceal wall thickens diffusely with fibrous proliferation, leading to contraction of the appendiceal lumen, which is the cause of the appendicitis [26]. Other manifestations include asymptomatic patients, intussusception, a palpable mass, gastrointestinal bleeding, increasing abdominal girth, chronic intermittent lower abdominal pain, and secondary genitourinary complications [6,45]. The duration of symptoms in symptomatic patients is in the range of 12 hours to 90 days [2]. In a number of patients the tumor is identified incidentally during an operation performed for some unrelated entities [2]. In women, they may also present as Krukenberg tumors to the ovaries [46].
In those patients where the tumor metastasizes to the ovaries, half of them primarily presented as an ovarian mass [47]. Some rare instances of clinical presentations include mesenteric adenitis, intestinal obstruction (caused by tumor infiltration of the terminal ileum) and anemia (caused by extensive tumor infiltration of the ceacum with mucosal ulceration) [27]. Höfler et al reported one GCC with mucosal ulceration [35]. Although serotonin can be detected immunohistochemically in these tumors, carcinoid syndrome at any stage is rare [27]. Carcinoid syndrome or Cushing's-like syndrome, resulting from the systemic effect of peptide secretion, indicates liver metastasis [48]. Sometimes mucocele may be seen as a parallel pathological finding around the involved GCC of appendix [17]. A very rare complication of GCC of the appendix is appendico-vesical fistula. In this report, the patient clinically presented with urological symptoms. After laparotomy the abscess and granulation tissue involving the appendix and bladder wall was detected and GCC was confirmed by histopathological examination [22].
Review of the literature has shown that the most common clinical presentations for GCC of the appendix in order of frequency are as follows: acute appendicitis 22.5%, asymptomatic 5.4%, non-localized abdominal pain 5.15% and abdominal mass in 3.09%.
Diagnosis of GCC
GCC is included under the general heading of mixed (composite) glandular-endocrine-cell carcinomas [49]. The diagnosis is almost always made postoperatively [3]. GCCs often present clinically as a diffusely inflamed appendix, with the diagnosis made by histological examination after surgery [15]. GCCs do not form discrete tumors and are not suspected during operation or at gross examination of the appendix. So, the surgical resection margins should be examined microscopically in all appendectomy specimens, especially in older patients, in whom these tumors most often occur [28].
GCC should be sharply distinguished from ordinary carcinoid tumors of the appendix, since they could behave biologically like carcinoma [2]. The mode of growth with concentration of the tumor elements within the basal part of the mucosal lamina propria, the pattern of invasion, growing diffusely through the layers of the wall without causing necrosis or much tissue destruction are more compatible with a carcinoid than with an adenocarcinoma [30]. GCCs can be easily diagnosed when they are of pure type [28]. The distinction between GCC and conventional carcinoid is based on a quantitative estimation of mucin production, since many carcinoids contain a little mucin [40]. In addition, in contrast to conventional carcinoids, GCCs contain few APUD cells and show positive staining of tumor cells for lysozyme, SC, and IgA [36]. For discriminating carcinoids and GCCs of the appendix from other gastrointestinal carcinomas, the predominance of signet-ring cells is found in sites often involved by metastatic signet-ring cell carcinomas of the gastrointestinal tract, whereas predominance of the carcinoid-cell type is found in sites most often involved by metastasis of conventional carcinoid tumors of appendix and GCC [43].
The resemblance of some areas in GCCs to Brunner's glands has been noted. However, the mucin in the latter is essentially neutral, whereas it is predominantly acid in GCC of the appendix [10]. In addition, in the face of excessive mucus production and an infiltrative pattern, differentiation from a mucinous adenocarcinoma remains a real challenge [17]. Histological transformation to ordinary adenocarcinoma has been observed occasionally in metastatic lesions of GCC, but most adenocarcinomatous lesions reported so far were still accompanied by varied numbers of argyrophilic tumor cells [37]. So the number of argyrophilic or argentaffin cells, as a tumor component is allegedly an indicator of the malignant potential in GCC's of the appendix with nonmetastasizing tumors commonly containing numerous neuroendocrine-type cells [37]. GCC of the appendix can be regarded as a peculiar type of adenocarcinoma, particularly when the number of neuroendocrine-type cells is limited [37]. In histological aspects for differential diagnosis, mucinous carcinomas differ from GCCs by the presence of glandular fusion and lack of glandular lumina within the mucin lakes [28]. Distinction between mucinous cystadenocarcinoma of appendix and GCC should also be considered. In the former, the mucosa lining the lumen is itself neoplastic, unlike GCC where the overlying epithelium appears normal with neoplastic transformation occurring from the deeper layers of the mucosa [17].
In most reports of metastasizing GCCs, the patients present with spread beyond the appendix at initial diagnosis [28]. In several reported cases, the ovarian masses were detected before the appendiceal tumor was recognized. In this clinical setting, the danger of misdiagnosing either a primary ovarian neoplasm or a metastasis from another site is high [5]. When mucin producing solid ovarian tumor is found without an extra-ovarian primary tumor during operation for another reason, then a diagnostic appendectomy should also be performed concomitantly [48]. When cystic mucinous tumors of the ovary showing bilateral involvement and/or stromal invasion by signet ring-type cells are encountered, the possibility of metastasis from an appendiceal GCC should be considered, irrespective of the presence or absence of argyrophilic tumor cells [37]. In-labeled octreotide scintigraphy is useful for diagnosis and staging of metastasis.
Paraclinical aspects of GCC
The white blood cell count is generally elevated with predominantly polymorphonuclear leukocytes and a "shift to the left," especially in patients presenting with acute symptoms [2,20]. However, it can be within normal range in some patients [2]. Generally, c-reactive protein is slightly elevated [7]. In one reported case with metastasis to ovaries, serum level of CEA was elevated [37] and in another report, plasma serotonin level was elevated preoperatively but these do not occur frequently [5]. Urinary 5-hydroxyindoleacetic acid (5-HIAA) level is estimated to be mostly within normal limits [27]. In cases with liver metastasis, liver enzymes and serum bilirubin will be elevated [43]. One of the rare complications of GCC of the appendix and all mucinous tumors of the appendix and or ovary is the occurrence of pseudomyxoma peritonei. This is a neoplastic disease of MUC2-expressing goblet cells, which collectively overexpress MUC2 because of their increase in number. Goblet cells of the appendix express both MUC2 and MUC5AC mucins but primary ovarian mucinous tumors only express MUC5AC. So measuring MUC2 is a specific marker for diagnosis of appendiceal induced pseudomyxoma peritonei [49,51].
X-ray of the abdomen shows dilated small bowel loops with fluid levels when obstruction occurs [45]. Cross-sectional imaging with CT is the technique of choice for any suspected appendiceal mass to rule out a tumor. In this technique the findings for GCC reflects the infiltrative nature of the tumor, with mild but diffuse mural thickening [6].
Surgical therapy of GCC
The site of the tumor within the appendix is not important for the choice of treatment [51]. In general, therapy is surgical. Even though GCC has a more aggressive phenotype than benign carcinoid tumors, the prognosis is generally good in a patient who is treated by simple appendectomy. Appendectomy remains the treatment of choice in the majority of patients. In some patients however a more radical procedure is indicated especially in the presence of diffuse appendiceal involvement [38,2,3]. Recently Varisco et al emphasized that in patients with no concomitant caecal involvement and low-grade tumor's histology; a simple appendectomy is enough [52].
Some studies recommend that because GCCs are among the semi high-grade malignant carcinoids with unpredictable behavior, in addition to simple appendectomy a right hemicolectomy should also be performed [45,42]. The general consensus remains that a right hemicolectomy should be performed when one or more of the following criteria are noted: 1), cellular undifferentiation; 2), increased mitotic activity; 3), involvement of the base of the appendix with caecal wall infiltration; 4), lymph node metastasis; and 5), tumor size greater than 2 cm. Invasion only into the mesoappendix or histological findings of intralymphatic or perineural invasion does not necessarily, by itself, require a more radical procedure [2]. However, if it cannot be confirmed that the base of the appendix is tumor-free, a right hemicolectomy is indicated due to the frequency of regional nodal metastases [27]. One study explained the indication of right hemicolectomy based on: incomplete excision of the tumor, foci of dedifferentiation and two or more mitotic figures per ten high power fields [25].
In addition to right hemicolectomy, some authors also advocate bilateral oophorectomy in female patients because of the possibility of ovarian metastasis [46,48]. Conversely, due to the tendency of GCCs affinity of spread to the ovaries, it is recommended that an appendectomy be routinely performed in patients with Krukenberg's tumor of the ovary where no gross obvious primary neoplasm can be found [53]. So, if an intraoperative histological evaluation reveals metastatic ovarian tumor, and at the same time, an obvious primary lesion is apparently absent, thorough exploration of the appendix is mandatory and routine appendectomy is recommended even if the appendix appears macroscopically normal [46]. In disseminated disease of the appendix in the abdomen, radical surgery is not recommended because it is doubtful that the course of the disease will be altered [48]. Thus, until better prognostic markers can be identified, an elective right hemicolectomy offers the best chance of controlling the disease and preventing recurrence in such patients [54,42]. As the route of metastatic spread of GCC's is through lymphatic and venous drainage, one should check the areas on the anterior surface of the third portion of the duodenum and above the origin of the ileocolic artery for metastatic deposits [55].
Even though the GCC of the appendix is not a complete high-grade malignant tumor, it can recur and metastasize so complete removal is imperative [10]. It is emphasized that recognition of this disease during the appendectomy from clues such as an abnormally thickened appendiceal wall, or the retention of mucinous material in the lumen should be encouraged [7].
Review of all reported GCCs showed that simple appendectomy was performed in nearly 24.57% of cases especially in the early reports. The most common surgical treatment of choice was appendectomy with right hemicolectomy (RHC) especially in the recent reports (34.7%). Oophorectomy was also performed in nearly 11.39% of female patients. In conclusion, because of the moderately aggressive behavior of the tumor, it is recommended that in all patients in addition to appendectomy a right hemicolectomy is also performed. As the mean age of the patients is over 50 and ovaries are the most common sites of metastasis, and bilateral involvement of the ovaries is the commonest metastatic behavior, bilateral salpingo-oophorectomy with total abdominal hysterectomy (TAH-BSO) is also advocated in selected female patients. Additional file: 1 summarizes the management of all diagnosed GCCs based on chronological reporting.
Adjuvant therapy of GCC
Peritoneal carcinomatosis from GCC origin is as invasive as peritoneal surface malignancy from colorectal adenocarcinoma. Patients in whom complete or near-complete surgical removal is possible should be considered for cytoreduction in combination with intraperitoneal chemotherapy [56].
As GCC's have this capacity to seed diffusely into the peritoneal cavity, innovative chemotherapy regimens similar to gastric cancer chemotherapy is advocated though in one study chemotherapy was administered without successful response [3]. One study of GCC with subserosal involvement advocated postoperative chemotherapy with 5-FU, Leucovorin, and CPT-11 [8]. Some authors mention that even small tumors early in the course of the disease will cause appendix obstruction and perforation which results in release of tumor cells into the free peritoneal cavity and this almost occurs before lymph node metastasis and therefore advocate the routine usage of interaperitoneal chemotherapy. There is no added benefit using either combination of mitomycin C plus 5-FU versus cisplatin plus Adriamycin [57]. There is also one report of a patient with GCC of the appendix and bilateral ovarian metastasis receiving X-ray radiation therapy as an adjuvant and after 8 months follow-up, no recurrence was detected [55]. Wojcik et al reported the first case of GCC on peritoneal fluid cytology in a woman with a prior history of carcinoid tumor of the appendix and the patient underwent 11 courses of Adriamycin and alpha-interferon with unknown documented follow-up [49]. A regimen of 5-FU and streptozocin was recommended in a case with appendiceal GCC and bilateral Krukenberg's of the ovaries and the patient was alive after 7 months follow-up [53]. There is also another report of GCC with Krukenberg's tumor, which advocated that cisplatin-based chemotherapy has some therapeutic benefits [46]. In general, Krukenberg's tumors subgroup due to appendiceal GCC has a survival advantage in patients receiving surgery and adjuvant therapy compared to surgery alone [34]. In one reported case with bilateral ovarian metastases patient was treated with carboplatin, cyclophosphamide and pirarubicin and followed by 9 months outpatient tegafur uracil (UTF) [37]. A recent study mentioned that a folfox chemotherapy that includes oxaliplatin combined with 5-FU and leucovorin is helpful in complete and persistent remission of a GCC case with metastasis to ovaries, peritoneal cavity and lymph nodes [58]. A summary of all reported GCCs and their management including chemotherapy is shown in Additional file: 1.
Even though it was not prescribed routinely, most protocols include 5-FU. As peritoneal seeding can occur before lymph node metastasis and there is proven advantage of chemotherapy in metastatic patients, 5-FU is therefore highly recommended in all GCC patients. Radiotherapy was only ordered in 4 cases without any specific consequence.
Follow-up of patients with GCC
Metastasis occurs in nearly 20% of the cases [16]. In general, metastatic ability is calculated in a range from 0 to 50 % [40]. The most common route of metastasis is through lymphatic vessels, trans-celomic and intraperitoneal invasion [27,33]. However, hematogenous metastasis to the liver or other distant organs is rarely mentioned [33]. Metastasis beyond the appendix is more frequent in older patients [27]. Metastasis from GCCs of the appendix has been described to the ovaries, pelvis, abdominal cavity, rib, vertebra, and lymph nodes [2,52]. The ovary is the most common target site of metastasis of GCC and metastatic lesions sometimes show a histological picture of mucin-producing adenocarcinoma [37]. In general, when metastatic spread occurs, it shows a predilection for the peritoneum and ovaries, with peritoneal carcinoidosis [43]. GCC's have a propensity to metastasize to the ovaries indistinguishable from the typical Krukenberg's tumor [46]. Primary ovarian carcinoids are invariably unilateral, whereas metastatic carcinoids to the ovaries are almost always bilateral or unilateral with extraovarian involvement [34]. In female patients with metastatic GCCs, ovarian involvement is seen in almost 90% of cases and is the source for further tumor dissemination [16]. Liver metastasis with abnormal liver function tests and hepatomegaly was also seen occasionally [27]. In metastatic deposits to organs including ovaries, fallopian tubes, uterus, and colon the histology is different and mostly composed by signet-ring cells without typical carcinoid cells but the metastases to liver and lymph nodes show a pattern of typical midgut carcinoids with signet-ring cells centrally and carcinoid cell type peripherally [43]. Also in rare cases only the mucinous component may spread and metastasize [39].
In-labeled octreotide scintigraphy is the most sensitive imaging modality in the diagnosis and staging of metastatic disease. Plasma chromogranin A is the most important blood marker available with levels being raised in 80–100 percent of patients with neuroendocrine tumors including GCC. Chromogranin A levels correspond to tumor load and levels above 5000 μg/l predict a poor outcome. Therefore, all patients with GCCs warrant investigation with plasma chromogranin A concentration, 24-h urinary levels of 5-HIAA, CT and In-labeled octreotide scintigraphy [45]. Where appendectomy has been done as the only treatment, prolonged follow-up is recommended and may involve laparoscopy in view of the risk of trans-celomic spread [27]. One study mentioned that pancreatic polypeptide might be useful as a tumor marker for follow-up purposes [27].
Review of the literature has shown that 11.2% of the patients were diagnosed with concomitant distant metastasis with ovaries being the most common site (3.60%) followed by abdominal carcinomatosis (1.03%). In addition, local lymph node involvement was seen in 8.7% of patients at the time of diagnosis. After surgical treatment, metastasis occurred in 4.1% of patients most commonly with abdominal carcinomatosis in 1.37% followed by ovaries in 0.85%. In conclusion, when GCC is suspected, both the ovaries and the peritoneum should be examined both during surgical extirpation and during follow up.
Prognosis of GCC
The prognosis of these tumors is estimated to be somewhere between carcinoids and well-differentiated adenocarcinomas of the appendix [3]. Prognosis is mostly favorable except in those patients who present with the more virulent form, delayed local recurrence, lung metastases or dual neoplasia e.g., GCC/epithelial neoplasia. In dual neoplasia prognosis depends on the stage and grade of the other component [3]. The association of GCC of the appendix with other cancers also clearly worsens the prognosis of the patient [2]. The associated carcinomas are most commonly carcinoma of the urinary bladder, prostate, ovaries, stomach and breast [2]. Other study however suggested that the prognosis of dual appendiceal GCC/epithelial neoplasia would be no worse than for either of the components alone [12]. There are also two reported rare cases of GCC and mucinous cystadenoma of the appendix with no information on their prognosis and outcome [29]. Involvement of the base of the appendix with only limited caecal wall involvement may still carry a favorable prognosis. Extension of the tumor beyond the appendix to the ovaries or the abdominal cavity, however, greatly worsens the prognosis [2]. In general, when tumor metastasis to the ovary is present, the prognosis is poor with a mean survival of 7 to 9 months [47]. Widespread intraabdominal dissemination can be fatal even in a short time [48].
Histological features such as nuclear pleomorphism and mitotic activity show no significant correlation with staging or prognosis [27]. On the other hand, some authors say the vast majority of GCCs that metastasize have spread beyond the serosa, suggesting that stage is another important prognostic factor [16]. Tumor characteristics that predict aggressive behavior include size, histological subtype and mesoappendiceal involvement [45]. Discussion regarding tumor size is controversial. Some studies say that size is not a good indicator for prognosis because these lesions do not always form discrete masses [3]. Other studies say the tumor size is only related to the metastatic potential and tumors larger than 2 cm are associated with a poorer prognosis. However, even smaller tumors do occasionally metastasize [48]. Angioinvasion should be regarded as a feature of more malignant behavior. In general, serosal involvement and mesoappendiceal extension are more predictive of outcome than lymph node status [45]. One study reported that those GCC tumors that show positive staining for pancreatic polypeptide behave in an aggressive fashion [27].
As a rule, survival for GCC of appendix is not significantly worse than that for patients with malignant carcinoid when adjusted for age and extent of disease at presentation, and is similar to survival after treatment of a low-risk adenocarcinoma [45]. One report referred to increasing age and evidence of tumor spread as poor prognostic factors [27]. In addition, reduced E-cadherin staining and overexpression of p21 need further assessment as future prognostic follow-up markers [59]. Another study mentioned that 20% of appendiceal GCCs that metastasize result in death [16]. The reported 5-year survival of GCC is between 60% and 84% [61,15,12,60,8,53] and 16% of all cases have recurrence of the disease [54].
Generally, the behavior of GCC is unpredictable and cannot be anticipated from the tumor's diameter, as in the case with carcinoid tumors [25]. Average survival time is 4.6 years with a range of 6 months to 20 years [25].
Conclusion
GCC of the appendix is a rare neoplasm that has varied clinical presentations ranging from an asymptomatic patient that is recognized during investigations for another problem, to acute appendicitis, the most common presentation, and toward the other end of the spectrum the metastatic presentation like Krukenberg's tumor or widespread peritoneal metastatic disease. So with this wide range of presentation, GCC should be always considered as a differential diagnosis in patients with "suspected appendicitis" with or without a mass leading to abdominal surgery. After the tumor is histologically confirmed, some factors need to be evaluated histologically as predictive factors for the biological behavior, e.g., the increased number of Paneth cells, increased mucin secretion and positivity for pancreatic polypeptide show a more aggressive and metastatic behavior and such cases should be considered for long-term follow-up. As nearly 20% show metastatic behavior and high possibility of intraperitoneal seeding before lymph node involvement, appendectomy with completion right hemicolectomy is recommended for all patients with caecal involvement and or high-grade tumors. Moreover, as the most common sites of metastasis are ovaries, in female patients, regardless of age and involvement, bilateral salpingo-oophorectomy is also recommended. In the presence of metastasis, combination chemotherapy regimes include 5-FU and leucovorin. It should be kept in mind that despite different terminology the biological difference between adenocarcinoid and carcinoma is not significant. Peritoneal carcinomatosis from adenocarcinoid of appendiceal origin is as invasive as peritoneal surface malignancy from colorectal adenocarcinoma. Patients in whom complete or near-complete surgical removal is possible should be considered for cytoreduction in combination with intraperitoneal chemotherapy. Based on the tumor histological predictors, long-term follow-up should be routinely advocated in all cases for possible metastasis. In this context, In-labeled octreotide scintigraphy is recommended for diagnosis and long-term follow-up of this neoplasm.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PSP: Is the first and the corresponding author, Conception and design, Literature review with references, Drafting the article and revising, provided statistical analysis of the reviewed data, preparation of table.
RK: helping in the preparation and drafting the manuscript, literature update, supervision, proof-reading and revising the manuscript
Supplementary Material
Additional file 1
Click here for file
Acknowledgements
the authors would like to thank Prof. Wolfgang Kuschinsky, the director of the department of physiology and pathophysiology, University of Heidelberg, Germany. This study was started and finished at his department.
==== Refs
McCusker ME Cote TR Clegg LX Sobin LH Primary malignant neoplasms of the appendix: a population-based study from the surveillance, epidemiology and end-results program, 1973–1998 Cancer 2002 94 3307 3312 12115365 10.1002/cncr.10589
Berardi RS Lee SS Chen HP Goblet cell carcinoids of the appendix Surg Gynecol Obstet 1988 167 81 86 3289135
Butler JA Houshiar A Lin F Wilson SE Goblet cell carcinoid of the appendix Am J Surg 1994 168 685 687 7978019
Ratzenhofer M Aubock L The amphicrine (endo-exocrine) cells in the human gut, with a short reference to amphicrine neoplasias Acta Morphol Acad Sci Hung 1980 28 37 58 7446221
Edmonds P Merino MJ LiVolsi VA Duray PH Adenocarcinoid (mucinous carcinoid) of the appendix Gastroenterology 1984 86 302 309 6690357
Pickhardt PJ Levy AD Rohrmann CA JrKende AI Primary neoplasms of the appendix: radiologic spectrum of disease with pathologic correlation Radiographics 2003 23 645 662 12740466
Kuroda N Mizushima S Guo L Jin Y Tao L Miyazaki E Toi M Naruse K Hiroi M Enzan H Goblet cell carcinoid of the appendix: Investigation of the expression of beta-catenin and E-cadherin Pathol Int 2001 51 283 287 11350611 10.1046/j.1440-1827.2001.01190.x
Stancu M Wu TT Wallace C Houlihan PS Hamilton SR Rashid A Genetic alterations in goblet cell carcinoids of the vermiform appendix and comparison with gastrointestinal carcinoid tumors Mod Pathol 2003 16 1189 1198 14681318 10.1097/01.MP.0000097362.10330.B1
Lancaster KJ Liang CY Myers JC McCabe KM Goblet cell carcinoid arising in a mature teratoma of the mediastinum Am J Surg Pathol 1997 21 109 113 8990147 10.1097/00000478-199701000-00012
Subbuswamy SG Gibbs NM Ross CF Morson BC Goblet cell carcinoid of the appendix Cancer 1974 34 338 344 4852178
Wolff M Crypt cell carcinoma Am J Surg Pathol 1982 6 188 189 7102895
Carr NJ Remotti H Sobin LH Dual carcinoid/epithelial neoplasia of the appendix Histopathology 1995 27 557 562 8838336
Li CC Hirowaka M Qian ZR Xu B Sano T Expression of E-cadherin, b-catenin, and Ki-67 in goblet cell carcinoids of the appendix: an immunohistochemical study with clinical correlation Endocr Pathol 2002 13 47 58 12114750 10.1385/EP:13:1:47
Zea-Iriarte WL Ito M Naito S Nakayama T Itsuno M Fujii H Furukawa M Makiyama K Sekine I Goblet cell carcinoid of the appendix Intern Med 1994 33 422 426 7949643
Kanthan R Saxena A Kanthan SC Goblet cell carcinoids of the appendix: immunophenotype and ultrastructural study Arch Pathol Lab Med 2001 125 386 390 11231488
Ramnani DM Wistuba II Behrens C Gazdar AF Sobin LH Albores-Saavedra J K-ras and p53 mutations in the pathogenesis of classical and goblet cell carcinoids of the appendix Cancer 1999 86 14 21 10391558 10.1002/(SICI)1097-0142(19990701)86:1<14::AID-CNCR4>3.0.CO;2-X
Chen V Qizilbash AH Goblet cell carcinoid tumor of the appendix. Report of five cases and review of the literature Arch Pathol Lab Med 1979 103 180 182 218519
Haqqani MT Williams G Mucin producing carcinoid tumors of the vermiform appendix J Clin Pathol 1977 30 473 480 864014
Warner TF SEo IS Goblet cell carcinoid of appendix: ultrastructural features and histogenetic aspects Cancer 1979 44 1700 1706 498041
Klein HZ Mucinous carcinoid tumor of the vermiform appendix Cancer 1974 33 770 777 4815580
Abt AB Carter SL Goblet cell carcinoid of the appendix. An ultrastructural and histochemical study Arch Pathol Lab Med 1976 100 301 306 946758
Di Paola M Stockwell WS Appendico-vesical fistula due to an appendix abscess with associated goblet cell carcinoid of the appendix Br J Urol 1976 48 436 188508
Warkel RL Cooper PH Helwig EB Adenocarcinoid, a mucin-producing carcinoid tumor of the appendix: a study of 39 cases Cancer 1978 42 2781 2793 728874
Kende AI Carr NJ Sobin LH Expression of cytokeratins 7 and 20 in carcinomas of the gastrointestinal tract Histopathology 2003 42 137 140 12558745 10.1046/j.1365-2559.2003.01545.x
Gallegos NC Milroy C Linehan IP Boulos PB Crypt cell carcinoma of the appendix Eur J Surg Oncol 1992 18 386 387 1521632
Aizawa M Watanabe O Naritaka Y Katsube T Imamura H Kinoshita J Shimakawa T Kobayashi S Asaka S Haga S Ogawa K Aiba M Kajiwara T Adenocarcinoid of the appendix: report of two cases Surg Today 2003 33 375 378 12734735 10.1007/s005950300085
Anderson NH Somerville JE Johnston CF Hayes DM Buchanan KD Sloan JM Appendiceal goblet cell carcinoids: a clinico-pathological and immunohistochemical study Histopathology 1991 18 61 65 1672861
Burke AP Sobin LH Federspiel BH Shekitka KM Helwig EB Goblet cell carcinoids and related tumors of the vermiform appendix Am J Clin Pathol 1990 94 27 35 2163192
AL-Talib RK Mason CH Theaker JM Combined goblet cell carcinoid and mucinous cystadenoma of the appendix J Clin Pathol 1995 48 869 870 7490325
Olsson B Ljungberg O Adenocarcinoid of the vermiform appendix Virchows Arch A Pathol Anat Histol 1980 386 201 210 7368563 10.1007/BF00427232
Cooper PH Warkel RL Ultrastructure of the goblet cell type of adenocarcinoid of the appendix Cancer 1978 42 2687 2695 728868
Rodriguez FH JrSarma DP Lunseth JH Goblet cell carcinoid of the appendix Hum Pathol 1982 13 286 288 7076213
Watson PH Alguacil-Garcia A Mixed crypt cell carcinoma. A clinicopathological study of the so-called 'goblet cell carcinoid' Virchows Arch A Pathol Anat Histopathol 1987 412 175 182 3122418 10.1007/BF00716191
Klein EA Rosen MH Bilateral Krukenberg tumors due to appendiceal mucinous carcinoid Int J Gynecol Pathol 1996 15 85 88 8852453
Hofler H Kloppel G Heitz PU Combined production of mucus, amines and peptides by goblet-cell carcinoids of the appendix and ileum Pathol Res Pract 1984 178 555 561 6483683
Isaacson P Crypt cell carcinoma of the appendix (so-called adenocarcinoid tumor) Am J Surg Pathol 1981 5 213 224 7235117
Ikeda E Tsutsumi Y Yoshida H Yanagi K Goblet cell carcinoid of the vermiform appendix with ovarian metastasis mimicking mucinous cystadenocarcinoma Acta Pathol Jpn 1991 41 455 460 1950561
Horiuchi S Endo T Shimoji H Takahashi H Mitsuuchi M Yawata A Mita H Yoshida M Arimura Y Sakamoto H Itoh F Hinoda Y Imai K Sasaki K Sato M Goblet cell carcinoid of the appendix endoscopically diagnosed and examined with p53 immunostaining J Gastroenterol 1998 33 582 587 9719248 10.1007/s005350050138
Miller RT Sarikaya Jenison EL Adenocarcinoid tumor of appendix presenting as unilateral Krukenberg tumor J Surg Oncol 1988 37 65 71 2826924
Bak M Asschenfeldt P Adenocarcinoid of the vermiform appendix. A clinicopathologic study of 20 cases Dis Colon Rectum 1988 31 605 612 3402286
Shebani KO Souba WW Finkelstein DM Stark PC Elgadi KM Tanabe KK Ott MJ Prognosis and survival in patients with gastrointestinal tract carcinoid tumors Ann Surg 1999 229 815 821 10363895 10.1097/00000658-199906000-00008
Park K Blessing K Kerr K Chetty U Gilmour H Goblet cell carcinoid of the appendix Gut 1990 31 322 324 2323597
Bak M Jorgensen LJ Adenocarcinoid of the appendix presenting with metastases to the liver Dis Colon Rectum 1987 30 112 115 3803115
Torres-melero J Garcia-Calvo M Blanco S Navarro MD Sanchez-Pernaute A Suarez B Ortega L Balibrea JL An adenocarcinoid tumor of the vermiform appendix Rev Esp Enfem Dig 1994 86 771 773
Goede AC Caplin ME Winslet MC Carcinoid tumor of the appendix Br J Surg 2003 90 1317 1322 14598408 10.1002/bjs.4375
Mandai M Konishi I Tsuruta Y Suginami N Kusakari T Iwasaki T Fujii S Krukenberg tumor from an occult appendiceal adenocarcinoid: a case report and review of the literature Eur J Obstet Gynecol Reprod Biol 2001 97 90 95 11435017 10.1016/S0301-2115(00)00503-0
Baker PM Oliva E Young RH Talerman A Scully RE Ovarian mucinous carcinoids including some with a carcinomatous component: a report of 17 cases Am J Surg Pathol 2001 25 557 568 11342766 10.1097/00000478-200105000-00001
Tjalma WA Schatteman E Goovaerts G Verkinderen L Van-den Borre F Keersmaekers G Adenocarcinoid of the appendix presenting as a disseminated ovarian carcinoma: report of a case Surg Today 2000 30 78 81 10648090
Wojcik EM Selvaggi SM Goblet-cell carcinoid tumor in peritoneal fluid: a case report Diagn Cytopathol 1991 7 155 157 1648469
O'Connell JT Tomlinson JS Roberts AA McGonigle KF Barsky SH Pseudomyxoma peritonei is a disease of MUC2-expressing goblet cells Am J Pathol 2002 161 551 564 12163380
Larsen SG Nilssen A Helseth A Bohler P Giercksky KE Invagination of the appendix with carcinoid tumor Eur J Surg 1999 165 993 997 10574110 10.1080/110241599750008152
Varisco B McAlvin B Dias J Franga D Adenocarcinoid of the appendix: is right hemicolectomy necessary? A meta-analysis of retrospective chart reviews Am Surg 2004 70 593 599 15279181
Hirschfield LS Kahn LB Winkler B Bochner RZ Gibstein AA Adenocarcinoid of the appendix presenting as bilateral Krukenberg's tumor of the ovaries. Immunohistochemical and ultrastructural studies and literature review Arch Pathol Lab Med 1985 109 930 933 2994592
Smith JA Goblet cell carcinoid of the appendix Gut 1990 31 840 2370022
Heisterberg L Wahlin A Nielsen KS Two cases of goblet cell carcinoid tumor of the appendix with bilateral ovarian metastases Acta Obstet Gynecol Scand 1982 61 153 156 7113691
Mahteme H Sugarbaker PH Treatment of peritoneal carcinomatosis from adenocarcinoid of appendiceal origin Br J Surg 2004 91 1168 1173 15449269 10.1002/bjs.4609
Sugarbaker PH Management of peritoneal surface malignancy: Appendix cancer and pseudomyxoma peritonei, colon cancer, gastric cancer, abdominopelvic sarcoma, and primary peritoneal malignancy, clinical results of treatment Assessed on June 6, 2005
Grain L Corbinais S Boucher E Blanchot J Le Guilcher P Raoul JL Adenocarcinoid of the appendix vermiformis: complete and persistent remission after chemotherapy (folfox) of a metastatic case Dig Dis Sci 2002 47 2760 2762 12498298 10.1023/A:1021065407822
Kawahara M Kammori M Kanauchi H Noguchi C Kuramoto S Kaminishi M Endo H Takubo K Immunohistochemical prognostic indicators of gastrointestinal carcinoid tumours Eur J Surg Oncol 2002 28 140 146 11884049 10.1053/ejso.2001.1229
Ludtke-Handjery A Pickartz H Strietzel S Goblet cell carcinoid of the vermiform appendix Leber Magen Darm 1991 21 226, 229 230 1758238
Berardi RS Chen H Goblet cell carcinoid of the appendix Int Surg 1989 74 109 110 2753618
Argani P van Hoeven KH Artymyshyn RL Diagnosis of metastatic appendiceal adenocarcinoid in liver by fine-needle aspiration cytology Diagn Cytopathol 1995 12 59 61 7789249
Zirkin RM Brown S Hertz M Adenocarcinoid of appendix presenting as bilateral ovarian tumors. A case report with histochemical and ultrastructural studies Diagn Gynecol Obstet 1980 2 269 274 7215111
|
15967038
|
PMC1182398
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Jun 20; 3:36
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-36
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-401597812810.1186/1477-7819-3-40ResearchTranshiatal and transthoracic resection in adenocarcinoma of the esophagus: Does the operative approach have an influence on the long-term prognosis? Gockel Ines [email protected] Sina [email protected] Claudia M [email protected] Werner [email protected] Theodor [email protected] Department of General and Abdominal Surgery, Johannes Gutenberg-University, Mainz, Germany2 Institute of Biostatistics and Documentation, Johannes Gutenberg-University, Mainz, Germany2005 24 6 2005 3 40 40 1 1 2005 24 6 2005 Copyright © 2005 Gockel et al; licensee BioMed Central Ltd.2005Gockel 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 goal of the present analysis was to investigate the long-term prognosis for adenocarcinoma of the esophagus treated with either the transhiatal (TH) or the transthoracic (TT) operative approach.
Methods
Between September 1985 and March 2004, esophageal resection due to carcinoma was performed on a total of 424 patients. This manuscript takes into account the 150 patients suffering from adenocarcinoma of the esophagus in whom a transhiatal resection of the esophagus was performed. In the event of transmural tumor growth and a justifiable risk of surgery, the transthoracic resection was selected. An extended mediastinal lymph node dissection, however, was only carried out in the course of the transthoracic approach.
Results
The transthoracic resection of the esophagus demonstrated a higher rate of general complications (p = 0.011) as well as a higher mortality rate (p = 0.011). The mediastinal dissection of the lymph nodes, however, revealed no prognostic influence. Considering all of the 150 patients with adenocarcinoma, as well as only those patients who had undergone curative resections (R0), the transhiatal approach was seen to demonstrate a better five-year survival rate of 32.1% versus 35.1%, with a median survival time of 24 versus 28 months, as compared with those who had undergone a transthoracic approach with a five-year survival rate of 13.6% (all patients) versus 17.7% (R0 resection) with a median survival time of 16 versus 17 months (p < 0.05).
Conclusion
The prognosis in patients with adenocarcinoma of the esophagus is influenced by the depth of the tumor (pT) and the pM-category, as shown in the multivariate analysis. The present analysis did not demonstrate a relevant difference in survival for patients with N0 and N1 stages undergoing transhiatal or transthoracic esophagectomy. It is questionable, if a more extensive mediastinal lymph node dissection, in addition to the clearance of abdominal lymph nodes, offers prognostic advantages in adenocarcinoma of the esophagus. However, the morbidity and mortality associated with the transthoracic approach is higher.
==== Body
Background
The adequate operative procedure for adenocarcinoma of the esophagus is currently being discussed with a great deal of controversy. The question is whether or not the prognosis can be improved through an extended lymph node dissection by the transthoracic approach as compared with the use of the transhiatal procedure combined with removal of the posterior, lower mediastinal lymph nodes. A randomized study comparing both operative procedures revealed a higher morbidity following the transthoracic approach, but no significant improvement in the prognosis even when including tumors of the esophagogastric junction [1]. Our own approach includes the transhiatal resection as a routine procedure and demonstrates a preference for the transthoracic approach in the event of a suspected tumor infiltration of intrathoracic or of an affliction of the mediastinal lymph nodes.
An analysis of 150 consecutive patients with adenocarcinoma of the esophagus, who had undergone esophageal resection and were evaluated prospectively, was carried out in order to investigate the significance of the surgical procedure on the prognosis.
Patients and methods
In 424 prospectively-evaluated patients, esophageal resections had been performed between September 1985 and March 2004 due to the existence of malignant tumors. The current data analysis, however, involves 150 patients with histologically verified adenocarcinoma of the esophagus, of whom 135 were Barrett's carcinoma. Adenocarcinoma enrolled in this study only included Siewert type I tumors. Type II (tumors of the cardia) and type III (subcardial tumors with infiltration of the cardia) were strictly not taken into consideration.
For preoperative staging, EUS (endoscopic ultrasound), CT (computed tomography) of the neck, chest and abdomen and PET (positron emission tomography) were routinely carried out.
Aside from the selected surgical procedure, the following parameters were documented:
The duration of surgery (incision-to-suture time), intraoperative blood loss as well as the intraoperative substitution of blood, time spent in the intensive care unit, tumor staging according to the UICC classification [2], number of tumor-free and tumor-infiltrated abdominal as well as mediastinal lymph nodes, surgical complications (anastomotic insufficiency, transplant necrosis, postoperative bleeding, tracheal fistula, recurrent laryngeal nerve paralysis, chylothorax), general complications (pneumonia, ARDS = adult respiratory distress syndrome, pulmonary embolism, decompensated cardiac insufficiency, myocardial infarction, transitional syndrome, renal insufficiency), duration of the postoperative hospital stay, mortality (30-day and hospital mortality rate) as well as the survival time have all been recorded. Until April 30, 2004, the end of the observation period for this study, the long-term follow-up was recorded and evaluated. It was possible to follow-up the post-stationary course of development in each of these 150 patients.
For adenocarcinoma of the esophagus, a transhiatal resection of the esophagus was carried out as a routine procedure. Patients with transmural tumor spread and suspected intrathoracic lymph nodes involvement and a justifiable risk of surgery, underwent a surgical procedure employing the transthoracic approach. All together, 103 patients underwent transhiatal and 47 patients transthoracic resection. The transthoracic surgical approach was performed via a right dorso-lateral thoracotomy, and the esophageal substitute consisted of a gastric pull-up located in the anatomical esophageal bed prevertebrally. During transthoracic esophageal resection, the anastomosis was usually done intrathoracically with a stapler (EEA 25 mm with an additional suture line by hand), whereas patients undergoing the transhiatal procedure had a cervical, hand-sewn anastomosis (end-to-side esophagogastrostomy with two suture lines).
The transhiatal procedure was carried out with an abdominal lymph node dissection (including the paracardial nodes, the left gastric artery nodes along with the lymph nodes of the lesser curvature of the stomach, the celiac trunk, the common hepatic artery and in selected cases – as macroscopic tumor involvement – the splenic artery), as well as with an excision of the lymph nodes extending as far as the carina of the trachea, and to those lymph nodes which could be reached in the lower, posterior mediastinum. The transthoracic technique involved an abdominal (as described above) and a more extensive mediastinal lymphadenectomy in the sense of a two-field dissection. The specimen here included the lower and middle mediastinal, subcarinal, and right-sided paratracheal lymph nodes (en bloc dissection). Paratracheal and bifurcal nodes were only removed on both sides in case of clinical suspicion of bilateral involvement. The aortopulmonary – window nodes were dissected separately. About 90% of the patients with adenocarcinomas (n = 133) demonstrated a tumor localization in the distal third of the esophagus.
Statistical analysis
In order to make a statistically-evaluative comparison between the operative procedures, a minimum of two, non-parametric statistical tests were applied to the measured parameters, making use of cross-tabulation analyses (in association with Fisher's exact test), which were employed for all categorized parameters involving (a) parameters that are considered to be nominal variables or (b) variables with a specific, ordinal scale value demonstrating only two intensities or categories.
Additionally, the Mann-Whitney U test (M-W test or U-test) was used for all of the parameters which could be classified according to an ordinal scale and which also demonstrated more than two intensities or categories.
The life-table and death-table analyses were primarily performed in order to calculate the three-and five-year survival rate for subgroups, following either one of the two surgical procedures. The median survival time, the specific survival curves of the various random samples, were determined by Kaplan-Meier analyses (K-M analyses). Differences in survival between the groups were assessed with the log-rank test. Multivariate survival analysis was performed using the Cox proportional hazard model. As no adjustment for multiple testing was performed, p-values should be considered as descriptive.
Median, minimum and maximum values, as well as the arithmetic mean and standard deviation, were also collected for the descriptive statistics, accompanied by the respective number of valid cases.
Results
With regard to age, ASA classification, tumor localization, metastasis as well as tumor grading, there was no statistically significant difference between patients undergoing the transhiatal or the transthoracic approach.
For those patients who had undergone transhiatal resections, individuals demonstrating both pT-category 1 and 2 were seen to be more frequent, compared to pT3-category. These patients also had fewer lymph node metastases (59% versus 78.3%, p = 0.027). Consequently, UICC stages I and II were seen to be more prevalent in those patients who had undergone transhiatal procedures (p < 0.0005) (Table 1). Tumor-free margins proximal, distal as well as circumferential (R0 resections) were observed more frequently following a transhiatal resection (92.2%) as compared with those treated using a transthoracic approach (78.7%) (p = 0.029) (Table 1).
Table 1 Clinicopathological features
Clinicopathological features
total transhiatal (TH) transthoracic (TT) p-value
(n = 150) (n = 103) (n = 47)
age (years)
median (range) 61 (28–78) 63 (28–78) 60 (35–75) 0.199
U-test
ASA classification
ASA II 62 145 42.8% 44 99 44.4% 18 46 39.1% 0.592
ASA III-IV 83 145 57.2% 55 99 55.6% 28 46 60.9%
Fisher's exact test
R classification
RO 131 149 87.9% 94 102 92.2% 37 47 78.7% 0.029
R1, R2 18 149 12.1% 8 102 7.8% 10 47 21.3%
Fisher's exact test
tumor localization
upper third 1 149 0.7% 0 103 0.0% 1 46 2.2%
middle third 15 149 10.1% 8 103 7.8% 7 46 15.2% 0.155
lower third 133 149 89.3% 95 103 92.2% 38 46 82.6%
Chi-Square (Pearson)
pT category
pT1 28 150 18.7% 27 103 26.2% 1 47 2.1%
pT2 39 150 26.0% 34 103 33.0% 5 47 10.6% <0.0005
pT3 77 150 51.3% 41 103 39.8% 36 47 76.6%
pT4 6 150 4.0% 1 103 0.9% 5 47 10.6%
Chi-Square (Pearson)
pN category
pN0 51 146 34.9% 41 100 41.0% 10 46 21.7% 0.027
pN1-2 95 146 65.1% 59 100 59.0% 36 46 78.3%
Fisher's exact test
pM category
pM0 105 149 70.5% 75 102 73.5% 30 47 63.8% 0.250
pM1 44 149 29.5% 27 102 26.5% 17 47 36.2%
Fisher's exact test
UICC stage
UICC I-II 62 149 41.6% 53 102 52.0% 9 47 19.2% <0.0005
UICC III-IV 87 149 58.4% 49 102 48.0% 38 47 80.9%
Chi-Square (Pearson)
tumor grading
G1 10 146 6.9% 8 100 8.0% 2 46 4.4% 0.167
G2 43 146 29.5% 32 100 32.0% 11 46 23.9%
G3 76 146 52.1% 46 100 46.0% 30 46 65.2%
G4 17 146 11.6% 14 100 14.0% 3 46 6.5%
Chi-Square (Pearson)
tumor length (grouped)
0–4 cm 69 143 48.3% 52 98 53.1% 17 45 37.8% 0.225
>4–8 cm 62 143 43.4% 39 98 39.8% 23 45 51.1%
>8 cm 12 143 8.4% 7 98 7.1% 5 45 11.1%
Chi-Square (Pearson)
Morbidity and mortality
The median time taken duration surgery for patients undergoing transhiatal resection was 254 minutes as compared with 335 minutes (without including the time required to change the position of the patient) for those undergoing a transthoracic resection (p < 0.0005).
The intraoperative blood loss was seen to be 800 ml (under the transhiatal procedure) in contrast to 1000 ml (using the transthoracic approach) (p = 0.002). The intraoperative blood substitution was a median of 0 units of packed red blood cells (PRBC) in the course of the transhiatal versus 1 unit during the transthoracic resection (p = 0.128) (Table 2).
Table 2 Morbidity and mortality
transhiatal (TH) transthoracic (TT) p-value
duration of surgery median (range) 254 (160–485) 335 (240–540) <0.0005
(min) U-test
blood loss median (range) 800 (0–7500) 1000 (200–5000) 0.002
(ml) intraoperatively U-test
blood substitution median (range) 0 (0–6) 1 (0–14) 0.128
(PRBC) intraoperatively U-test
surgical complications 44 103 42.7% 7 47 14.9% 0.001
Fisher
general complications 22 102 21.6% 20 47 42.6% 0.011
Fisher
total complications 62 103 60.2% 25 47 53.2% 0.477
Fisher
duration of postoperative intensive care unit stay
(days) median (range) 8 (2–107) 8 (1–97) 0.566
U-test
duration of postoperative in-hospital stay
(days) median (range) 21 (14–126) 20 (13–107) 0.646
hospital mortality 3 103 2.9% 7 47 14.9% 0.011
Fisher
30-day mortality 2 100 2.0% 7 47 14.9% 0.005
Fisher
The rate of surgical complications, with 42.7% (44/103) for the transhiatal approach, was higher than that observed in the course of a transthoracic resection (14.9%) (7/47) (p = 0.001), primarily as a result of an insufficiency of the cervical anastomosis, which could nevertheless be treated conservatively in all cases. The rate of general complications, with 42.6% (20/47) versus 21.6% (22/102) (p = 0.011), was higher following the transthoracic approach, which was associated with a higher rate of pneumonia. The median duration spent in the intensive care unit (8 days after both procedures; p = 0.566), as well as the median postoperative in-hospital stay (21 versus 20 days; p = 0.646), did not differ between the two types of resection (Table 2).
The hospital mortality for the entire population of patients was seen to be 6.7% (10/150). Following the transhiatal procedure, the 30-day (p = 0.005) and the hospital mortality (p = 0.011) were lower than that observed following the transthoracic approach (2.0 and 2.9% versus 14.9%) (Table 2). Causes of death were related to pulmonary sepsis in one patient of the TT group only (the others were due to cardiovascular problems (n = 5) and anastomotic insufficiency with sepsis (n = 1)). Causes of deaths following the TH-resection were pulmonary sepsis (n = 1), hemorrhagic shock (n = 1) and acute rupture of the mitral valve tendon (n = 1).
Lymph node status
The total number of lymph nodes dissected in the course of transhiatal resection were lower (19; range 0–55) than that observed during the transthoracic procedure (31, range 2–94) (p < 0.0005) (Table 3). The number of afflicted mediastinal lymph nodes dissected in the course of transhiatal resection, correspondingly, was lower as well (2 versus 4; p = 0.030). The lower number of excised lymph nodes could be based on the lower number of mediastinal lymph nodes which were seen to have been removed in the course of the transhiatal approach (3 versus 9, p < 0.0005) (Table 3). Correspondingly, the number of afflicted, mediastinal lymph nodes dissected during transhiatal resection was lower (0 versus 2; p < 0.0005) (Table 3). The number of abdominal lymph nodes which were excised was higher under the transhiatal in contrast to the transthoracic approach (14 versus 11, p = 0.017). The number of afflicted, abdominal lymph nodes, however, revealed no difference between the two operative procedures. The median lymph node-ratio (quotient between the number of tumor-infiltrated lymph nodes and the total number of lymph nodes dissected) during transthoracic resection was seen to be 0.10 versus 0 for the transhiatal as compared with 0.154 versus 0.111 for the transthoracic procedure, respectively (p > 0.05) (Table 3).
Table 3 Lymph node (LN) dissection
transhiatal (TH) transthoracic (TT)
median range mean median range mean U-test
total LN dissected 19 (0–55) 21.3 31 (2–94) 32.4 <0.0005
positive LN dissected 2 (0–33) 5.4 4 (0–76) 7.9 0.030
total abdominal LN 14 (0–48) 16.4 11 (0–51) 12.0 0.017
positive abdominal LN 1 (0–28) 3.4 1 (0–17) 3.1 0.802
total thoracic LN 3 (0–42) 5.1 19 (2–83) 20.2 <0.0005
positive thoracic LN 0 (0–30) 1.5 2 (0–67) 4.8 <0.0005
LN ratio total 0.11 (0–0.86) 0.20 0.17 (0–0.81) 0.24 0.160
LK ratio abdominal 0.10 (0–0.89) 0.17 0.15 (0–1) 0.25 0.197
LK ratio thoracic 0.00 (0–1) 0.22 0.11 (0–1) 0.24 0.156
Prognosis
The median and the five-year survival rates for the total patient population were markedly better following transhiatal resection than subsequent to the transthoracic procedure (16 months and 32.1% as compared with 16 months and 13.6%, respectively; p = 0.018) (Table 4). For patients with curative (R0) resection, the prognosis was also seen to be better for those patients who underwent transhiatal esophagectomy (28 months versus 17 months; p = 0.045) (Figure 1). In patients without lymph node metastases (pN0), the median survival time following an R0 resection and a transhiatal approach was 67 months (n = 40) as compared to 27 months for those who had undergone transthoracic esophagectomy (n = 9) (p > 0.05). For the patients who were seen to be nodal-positive (pN1), those who had undergone a transhiatal resection (n = 51) demonstrated a median survival time of 16 months contrasted to 14 months for those treated using a transthoracic approach (n = 26) (p > 0.05). An expansion in the number of mediastinal lymph nodes which had been dissected to include more than 19 lymph nodes (=median), however, was not seen to have any positive effect in this group (Table 4). In the scatter diagram (Figure 2), there was no correlation seen between the long-term prognosis of adenocarcinomas of the esophagus and the number of dissected mediastinal lymph nodes.
Table 4 Prognosis
operative approach n median survival (months) 5-year survival rate p-value
total TT 46 16 13.6% 0.018
TH 103 24 32.1%
curative (R0) TT 36 17 17.6% 0.045
TH 94 28 35.1%
palliative (R1) TT 10 9 - 0.778
TH 8 8 -
R0 pN0 TT 9 27 22.2% 0.072
TH 40 67 54.1%
R0 pN1 TT 26 14 14.3% 0.430 a
<19 thor. LN 12 17 16.7% 0.829 b
>/ = 19 thor. LN 14 8 -
TH 51 16 22,4% 0.703 c
R0 pN1 T1-2 TT 4 14 25.0% 0.357
TH 21 28 33.3%
R0 pN1 T3-4 TT 22 9 12.2% 0.793
TH 30 10 -
R0 pN1 ASA 1–2 TT 11 16 20.0% 0.345
TH 23 19 33.3%
R0pN1 ASA 3–4 TT 15 9 10.3% 0.999
TH 25 10 -
R0 middle third TT 4 5 25% 0.931
(tumor localization) TH 8 24 15%
R0 lower third TT 32 17 17.1% 0.044
(tumor localization) TH 86 30 37.9%
TT: Transthoracic, TH: Transhiatal
a = (TT vs. TH)
b= (TT intern)
c = (TT thor. LN vs. TH)
Figure 1 Kaplan-Meier survival curves for patients with adenocarcinoma and R0 resection undergoing transhiatal (TH) and transthoracic (TT) esophagectomy.
Figure 2 Correlation between long-term survival and the number of dissected thoracic lymph nodes (LN) for curative (R0) resected, nodal positive (pN1) patients (n = 77) with adenocarcinoma of the esophagus.
A relevant advantage to the transhiatal in comparison with the transthoracic operative procedure (n = 32; median survival time: 17 months; p = 0.044) was also to be seen for tumor localizations in the distal third of the esophagus (n = 86; median survival time: 30 months). For the minority of patients with a tumor localization in the middle third, however, there was no significant difference between these two surgical approaches (TT : 4, TH : 8).
In the multivariate survival analysis, the simultaneous influence of the surgical approach, R-classification, pT-, pN-, pM-category, tumor grading, ASA classification, and age was assessed. Only the pM-(p = 0.020) and pT-category (p = 0.009) had a relevant influence on survival (Table 5). In a forward variable selection pT and pM were the only variables that were selected.
Table 5 Multivariate analysis of prognostic factors for survival
95%-confidence interval
variable hazard ratio lower limit upper limit p-value
R-classification
-R1 and R2 vs. R0 1.368 0.772 0.2424 0.284
pT-category 0.009
-pT2 vs. pT1 2.418 0.952 6.144 0.063
-pT3 and pT4 vs. pT1 3.930 1.569 9.847 0.003
pN-category
-pN1, 2 vs. pN0 1.398 0.776 2.526 0.267
tumor grading 0.659
G2 vs. G1 0.731 0.264 2.021 0.545
G3 vs. G1 0.993 0.370 2.666 0.989
G4 vs. G1 1.082 0.348 3.360 0.892
ASA-classification 0.777
ASA III vs. ASA II 1.100 0.680 1.779 0.699
ASA IV vs. ASA II 1.391 0.544 3.558 0.491
pM-category 1.782 1.096 2.898 0.020
surgical approach
TT vs. TH 0.937 0.592 1.482 0.781
age 0.997 0.972 1.023 0.839
5-year survival rates for R0-resections according to the different UICC stages for the whole group were: stage I (n = 20): 56.3%; stage IIa (n = 20): 26.5%; stage IIb (n = 16): 16.5%; stage III (n = 34): 34.4%; and stage IV (n = 31): 14.8%. Multivariate analysis of three-and five-year survival including surgical approach and UICC staging showed no significant influence of the surgical approach, neither with respect to the three-nor to the five-year survival (3 years: p = 0.289; 5 years: p = 0.685) (Table 6, 7).
Table 6 Multivariate analysis of three-year survival
95%-confidence interval
variable odds ratio lower limit upper limit p-value
surgical approach
TT vs. TH 1.931 0.573 6.514 0.289
UICC 0.001
-IIa vs. I 1.866 0.459 7.583 0.383
-IIb vs. I 20.350 2.132 194.254 0.009
-III vs. I 5.967 1.400 25.439 0.016
-IV vs. I 46.422 5.071 424.994 0.001
Odds ratios are to be interpreted as odds for death so that they are interpretable in the same way as the hazard ratios in Table 5.
Table 7 Multivariate analysis of five-year survival
95%-confidence interval
variable odds ratio lower limit upper limit p-value
surgical approach
TT vs. TH 0.716 0.143 3.594 0.685
UICC 0.020
-IIa vs. I 5.483 0.796 37.788 0.084
-IIb vs. I 8.575 0.850 86.488 0.068
-III vs. I 12.133 1.626 90.526 0.015
-IV vs. I 28.124 2.773 285.211 0.005
Odds ratios are to be interpreted as odds for death so that they are interpretable in the same way as the hazard ratios in Table 5.
Discussion
The present investigation is based on 150 consecutive patients who had undergone surgery for adenocarcinoma of the esophagus and whose peri-and postoperative course of was evaluated prospectively. The transhiatal approach with the posterior excision of the lower mediastinal lymph nodes was employed as a routine procedure. In the event that a tumor was to be found in the vicinity of the tracheal bifurcation, or should there be suspicion of an infiltration of intrathoracic organs or mediastinal lymph node metastases, a transthoracic approach with two-field lymphadenectomy was selected.
The study verified the varying perioperative risks of the transthoracic and transhiatal procedures. With the same age distribution and the same risk factors, measured according to the ASA classification, the 30-day and hospital mortality following the transhiatal approach were seen to be markedly reduced as compared with the transthoracic procedure. This operative technique, however, primarily led to general complications, especially of pulmonary kind, which served to bring about an unfavorable course of development. The relatively high mortality rate after the transthoracic procedure reflects patients seen during the whole study period including the earlier interval. Recently, probably due to improved surgical techniques and advanced intensive care therapy, the rate was seen to be much lower compared to the first study period. Following transhiatal esophagectomy, the surgical complications predominated, especially the insufficiency of the cervical esophagogastrostomy, a condition which was nevertheless seen to close in all cases subsequent to a conservative therapy. The reduced general risk of the transhiatal procedure has been confirmed in most of the available studies [3-6]. On the other hand, though exerting a higher operative risk, transthoracic resection intends to improve long-term survival by wide excision of the tumor and peritumoral tissue with extended en bloc mediastinal lymphadenectomy [7-11].
The influence of the operative procedure on the prognosis in adenocarcinoma of the esophagus has been investigated in a number of retrospective and in one prospective, randomized study [1]. The disadvantage of the blunt transhiatal resection is the reduced transthoracic lymph node dissection, which is unfortunately limited to the lower posterior, mediastinal lymph nodes [12-15]. In the present investigation, a higher number of mediastinal lymph nodes were consequently seen to be excised in the course of transthoracic resections, a situation which has also resulted in an increase in the number of involved lymph nodes. The number of afflicted, abdominal lymph nodes was seen to be similar following both of the two types of operative procedures.
Our data show that through transhiatal esophagectomy, a five-year survival rate of 54.1% was seen in patients without (pN0) and of 22.4% in those with lymph node involvement (pN1). Patients undergoing transthoracic resection without afflicted lymph nodes demonstrated no improvement in prognosis (22.2%) as a result of this procedure, although only a smaller number of such cases were observed here. Findings derived from studies investigating the spread of lymph nodes in cases of adenocarcinoma allow one to conclude that the lymph node metastases associated with distal adenocarcinomas are initially seen to metastasize into the lymph nodes in the vicinity of the tumor and only later into the lymph nodes of the upper mediastinal region [16]. A counter argument is a likelihood of cervical lymph node involvement as high as 3.5% reported for T3 adenocarcinoma of the distal esophagus and the gastro-esophageal junction [17]. Thus, the transhiatal procedure must consequently be considered to be adequate at least for patients without any lymph node involvement.
In cases with lymph node involvement, the prognosis of our own patient population was independent of the surgical procedure and, consequently, also of the extent of lymph node dissection. The transhiatal esophagectomy, because of the reduced morbidity and mortality and the presence of more early stage tumors with correspondingly more frequent R0 resections associated with this approach, however, was seen to result in a better long-term prognosis. Thus, the present study was not randomized and the survival benefit seen after transhiatal esophagectomy might therefore be based on a more favorable patient selection. This needs to be checked in a randomized controlled trial.
The higher number of advanced stages in the transthoracic group is certainly explained by the marked difference in the pT-category (p < 0.0005) and also possibly by stage migration due to the more extended lymph node dissection (pN-category: p = 0.027) (Table 1).
In the event of existing lymph node metastases, this indicates that a generalized disease is present which can no longer be influenced by local, surgical measures. In the multivariate analysis of potential factors with an influence on the long-term survival rate of our patient population, the pT-and pM-category were seen to represent independent, prognostic parameters. In accordance with this, the findings of other authors must also be expounded upon, which demonstrated, following extensive lymph node dissection in cases of adenocarcinoma, that the most favorable results were to be observed in the event of lacking or only minimal lymph node involvement [1].
The question of whether or not a neoadjuvant (radio-) chemotherapy can lead to an improvement in the prognosis, as verified in one study, but not confirmed in another, cannot presently be answered conclusively [18,19].
The prognosis following an R1 resection, subsequent to both – transthoracic as well as transhiatal – approaches, with survival times of 8 or 9 months, respectively, must be considered unfavorable, so that this situation must be avoided whenever possible.
For the operative procedure involving distal adenocarcinomas of the esophagus, the present results permit one to make the following conclusions: The transhiatal procedure together with a posterior, lower mediastinal lymph node dissection is associated with a comparably reduced perioperative risk and, for patients in whom a radical tumor excision is possible, represents an oncologically adequate method. The prognosis following R0 resection is defined by the T-and N-categories. In the event of an N1 situation, the prognosis must be considered unfavorable, independent of the operative procedure selected. According to the present investigation including a lack of randomization, it is questionable, if the long-term survival can be improved through the expansion of a lymph node dissection using the transthoracic procedure. This approach, however, can then be indicated when a complete resection of the tumor is not to be achieved by transhiatal esophagectomy, for instance when the tumor is seen to demonstrate a close relationship to the tracheal bifurcation, the tumor is suspected to involve an infiltration of intrathoracic structures or in the event of suspected lymph node metastases in the upper mediastinal region.
Conclusion
The present analysis did not demonstrate a relevant difference in survival for patients with N0 and N1 stages undergoing transhiatal or transthoracic esophagectomy. It is questionable, if an extensive mediastinal lymph node dissection in addition to the clearance of abdominal lymph nodes offers any prognostic advantages in adenocarcinoma of the esophagus also probably due to the increased morbidity and mortality associated with the transthoracic approach.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
IG: study design, collection of data, statistical analysis, sequence alignment, draft of manuscript
SH: collection of data, statistical analysis
CMM: statistical analysis
WK: study design, collection of data, statistical analysis
ThJ: conceived of the study, design and coordination of the study, draft and revision of the manuscript
==== Refs
Hulscher JB van Sandick JW de Boer AG Wijnhoven BP Tijssen JG Fockens P Stalmeier PF Ten Kate FJ van Dekken H Obertop H Tilanus HW van Lanschot JJ Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus N Engl J Med 2002 347 1662 1669 12444180 10.1056/NEJMoa022343
Wittekind CH Wagner G TNM-Klassifikation maligner Tumoren 1997 Springer, Berlin Heidelberg New York 123
Orringer MB Marshall B Stirling MC Transhiatal esophagectomy for benign and malignant disease J Thorac Cardiovasc Surg 1993 105 265 276 8429654
Pac M Basoglu A Kocak H Yekeler I Yediyildiz S Aydin NE Yilm A Okcu N Keles M Transhiatal versus transthoracic esophagectomy for esophageal cancer J Thorac Cardiovasc Surg 1993 106 205 209 8341062
Fok M Siu KF Wong J A comparison of transhiatal and transthoracic resection for carcinoma of the thoracic esophagus Am J Surg 1989 158 414 419 2817222 10.1016/0002-9610(89)90275-4
Boyle MJ Franceschi D Livingstone AS Transhiatal versus transthoracic esophagectomy: complication and survival rates Am Surg 1999 65 1137 1141 10597061
Müller JM Erasmi H Stelzner M Zieren U Pichlmaier H Surgical therapy of esophageal carcinoma Br J Surg 1990 77 845 857 2203505
Chu KM Law SY Fok M Wong J A prospective randomized comparison of transhiatal and transthoracic resection for lower-third esophageal carcinoma Am J Surg 1997 174 320 324 9324146 10.1016/S0002-9610(97)00105-0
Goldminc M Maddern G Le Prise E Meunier B Campion JP Launois B Oesophagectomy by transhiatal approach or thoracotomy: a prospective randomized trial Br J Surg 1993 80 367 370 8472154
Jacobi CA Zieren HU Müller JM Pichlmaier H Surgical therapy of esophageal carcinoma: the influence of surgical approach and esophageal resection on cardiopulmonary function Eur J Cardiothorac Surg 1997 11 32 37 9030787 10.1016/S1010-7940(96)01106-2
Hulscher JBF Tijssen JGP Obertop H van Lanschot JJB Transthoracic versus transhiatal resection for carcinoma of the esophagus: a meta-analysis Ann Thorac Surg 2001 72 306 313 11465217 10.1016/S0003-4975(00)02570-4
Akiyama H Tsurumaru M Udagawa H Kajiyama Y Radical lymph node dissection for cancer of the thoracic esophagus Ann Surg 1994 220 364 372 8092902
Roder JD Busch R Stein HJ Fink U Siewert JR Ratio of invaded to removed lymph nodes as a predictor of survival in squamous cell carcinoma of the oesophagus Br J Surg 1994 81 410 413 8173915
Siewert JR Stein HJ Lymph node dissection in squamous cell esophageal cancer – who benefits? Langenbecks Arch Surg 1999 384 141 148 10328167 10.1007/s004230050184
Tsurumaru M Kajiyama Y Udagawa H Akiyama H Outcomes of extended lymph node dissection for squamous cell carcinoma of the thoracic esophagus Ann Thorac Cardiovasc Surg 2001 7 325 329 11888470
Tachimori Y Kato H Watanabe H Sasako M Kinoshita T Maruyama K Difference between carcinoma of the lower esophagus and the cardia World J Surg 1996 20 507 510 8662143 10.1007/s002689900080
Van de Ven C De Leyn P Coosemans W Van Raemdonck D Lerut T Three-field lymphadenectomy and pattern of lymph node spread in T3 adenocarcinoma of the distal esophagus and the gastro-esophageal junction Eur J Cardiothorac Surg 1999 15 769 773 10431857 10.1016/S1010-7940(99)00122-0
Medical Research Council Oesophageal Cancer Working Group Surgical resection with or without preoperative chemotherapy in esophageal cancer. A randomized controlled trial Lancet 2002 359 1727 1733 12049861 10.1016/S0140-6736(02)08651-8
Samel S Hofheinz R Hundt A Sturm J Knoll MR Wenz F Queißer W Post S Neoadjuvant radio-chemotherapy of adenocarcinoma of the oesophagogastric junction Onkologie 2001 24 278 282 11455222 10.1159/000055092
|
15978128
|
PMC1182399
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Jun 24; 3:40
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-40
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-441602273610.1186/1477-7819-3-44Case ReportOccult solitary submucosal jejunal metastasis from esophageal carcinoma Lindenmann Joerg [email protected] Franz [email protected] Veronika [email protected] Christian [email protected] Alfred [email protected] Freyja Maria [email protected] Department of Thoracic and Hyperbaric Surgery, University Medical School, Auenbruggerplatz 29, Graz, Austria2 Department of Pathology, University Medical School, Auenbruggerplatz 35, Graz, Austria2005 16 7 2005 3 44 44 27 2 2005 16 7 2005 Copyright © 2005 Lindenmann et al; licensee BioMed Central Ltd.2005Lindenmann 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
Metastatic tumors of the intestinal tract from extra-abdominal sites are rare. In esophageal cancer, the liver, lung and the bones are the most common sites of metastases. Metastasis to intestines are very rare.
Case presentation
A 54-year old male was admitted with esophageal squamous cell carcinoma (SCC) associated with dysphagia II-III and weight loss of 20 kg. Preoperative routine staging failed to detect any metastases. A transthoracic esophagectomy and orthotopic gastric pull-up with collar esophago-gastrostomy, associated with 2-field lymphadenectomy was perfromed. During the digital placement of the naso-jejunal feeding catheter a submucosal jejunal nodule with a diameter of 1 cm, about 40 cm distal to the duodeno-jejunal fold was detected which was completely resected by jejunotomy. Histopathology of jejunal nodule showed metastasis from esophageal squamous cell carcinoma.
Conclusion
Because of the extensic esophageal lymphatic system, an occult widespread dissemination of the tumor cells into the abdominal cavity is possible. Additional intraoperative evaluation of the small intestine and the complete abdominal cavity should be performed in every operation of esophageal carcinoma to detect possible occult intraabdominal metastases.
==== Body
Background
Metastatic tumors of the intestinal tract from extra-abdominal sites are rare. In esophageal cancer, the liver, lung and the bones are the most common sites of metastases. Metastases to the small intestine are very rare. We describe the case of a 54-year-old man suffering from esophageal carcinoma, who underwent transthoracic esophagectomy and 2-field lymphadenectomy, the reconstruction was done by orthotopic gastric pull-up. During the operation, a solitary submucosal jejunal metastasis of the esophageal carcinoma was detected and excised.
Case presentation
A 54-year-old male, alcohol and tobacco user, was admitted with squamous cell carcinoma of the esophagus associated with the clinical symptoms of high-grade dysphagia and weight loss of 20 kg. Routine staging was done by computerized tomography (CT)-scan of the thorax and abdomen, ultrasonography, positron emission tomography (PET) scan, esophago-gastro-duodenoscopy and bronchoscopy. Functional evaluation was done by electrocardiogram (ECG), cardiac ultrasonography and spiroergometry.
The CT-scan of the thorax and the mediastinum showed a tumor of the middle third of the esophagus with a suspicion of tumor infiltration of the thoracic aorta and the left main bronchus. However, infiltration of the main bronchus could be excluded by bronchoscopy. Infiltration of the thoracic aorta could also be excluded by MRI-angiography. The CT-scan of the abdomen and the abdominal ultrasound showed no signs of tumor spread. The PET-scan demonstrated a pathological tracer uptake at the level of the middle third of the esophagus with no signs of distal metastasis. Esophago-gastro-duodenoscopy showed a tumor stenosis of 7 cm length, from 30 to 37 cm, and a diameter of 4 – 6 mm, further there was an intraluminal obstruction about more than 50 % which was additional confirmed by an esophagogram.
Transthoracic esophagectomy and 2-field lymphadenectomy was carried out as reported in literature [1,2]. Intraoperative frozen section showed no residual tumor at the lateral, distal and proximal margins. The mediastinal lymph nodes showed no signs of tumor infiltration. Reconstruction was done by orthotopic gastric pull-up and hand-sewn side to end collar esophago-gastrostomy. In order to perform postoperative early feeding a naso-jejunal catheter was introduced and placed distal to the duodeno-jejunal fold.
During the digital placement of the naso-jejunal feeding catheter a submucosal jejunal nodule with a diameter of 1 cm was detected about 40 cm distal to the duodeno-jejunal fold. The nodule was completely resected by jejunotomy. Intraoperative frozen section showed a submucosal metastasis of the esophagus. Further evaluation of the small intestine and the complete abdominal cavity showed no signs of metastasis.
Finally histopathological work-up of the specimens confirmed the diagnosis of squamous cell carcinoma of the esophagus (Figure 1) and a solitary submucosal jejunal metastasis. Furthermore, the submucosal jejunal metastasis was associated with local submucosal venous and lymphatic infiltration (Figure 2). In immunhistochemical tests the tumor cells showed reaction to CEA and CK 5–6. The definitive staging was T3, N0, M1, L1, G3, R0.
Figure 1 Photomicrograph of specimen showing the esophageal SCC (hematoxylin and eosin × 25).
Figure 2 Submucosal jejunal metastasis of the esophageal squamous cell carcinoma (SCC). Photomicrograph of specimen showing the metastasis, associated with local submucosal venous and lymphatic infiltration. Neither the jejunal serosa nor the mucosa are affected by the carcinoma (hematoxylin and eosin × 50).
The further course of the patient was uneventful and he could be discharged on the 14th postoperative day. As this was stage IV squamous cell carcinoma of the esophagus adjuvant chemotherapy was not offered to the patient [3]. However, a follow-up protocol was initiated with CT-scans every 3-month for the first year.
Discussion
Metastatic tumors of the intestinal tract from extra- abdominal sites are rare [4,5]. In esophageal cancer, the liver, lung and the bones are the most common sites of metastases. Metastases to the small intestine are very rare [6].
Squamous cell carcinoma of the esophagus is characterized by an extensive lymphatic dissemination via the longitudinal lymphatic system along the esophagus, which drains in to a large number of widespread lymph nodes of the thorax and the abdomen depending on the location of the tumor [2]. Retrograde spread to the cervical, mesenteric and iliac lymph nodes are very unusual, but have been reported in literature [7,8]. The intraabdominal region can be reached via the vertebral venous system [9,10]. This huge and widespread lymphatic and hematogenous system connecting the esophagus with the intraabdominal region may be sufficient to explain, why this small solitary metastasis of the jejunum was discovered for the first time during the operation and could probably be responsible for this extraordinary bad prognosis in patients with esophageal squamous cell carcinoma caused by the occult widespread dissemination of the tumor cells.
In this case, the operation was absolutely indicated to prevent the patient from further progression of his dysphagia, to avoid the development of a malignant esophagotracheal fistula with its typical complication, mainly bronchopneumonia due to nocturnal aspiration and to avoid the development of an jejunal ileus caused by intraluminal obstruction by that solitary jejunal metastasis. A quite interesting aspect is, why this solitary metastasis could not be discovered by the PET-scan, which showed a pathological tracer uptake at the level of the middle third of the esophagus, further an unspecific tracer uptake due to inflammatory disease in the subhepatic region of the right upper abdomen was seen but there was no suspicion of further distal metastases. The main reason for this failure of the PET-scan could be the high incidence of artifacts caused by many different small intraabdominal focuses of inflammation which make the correct assessment of the abdominal cavity quite difficult.
Conclusion
Additional intraoperative evaluation of the small intestine and the complete abdominal cavity should be performed in every laparotomy for esophageal carcinoma to detect possible occult intraabdominal metastases. PET is not sensitive enough to pick small metastasis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JL: Preparation of draft manuscript, Surgical management, Revision of manuscript and preparation of final manuscript.
VM: Preparation of manuscript, surgical management
AM, Surgical management, Revision of manuscript and preparation of final manuscript
CP, Surgical management
FG: Histopathological work-up, immunhistochemical tests, photomicrograph of specimen.
FMSJ: Revision of manuscript and preparation of final manuscript.
All authors read and approved the final version of the manuscript.
Acknowledgements
Written consent was taken from the patient for publishing his clinical details and histopathological photomicrographs.
==== Refs
Dutkowski P Kneist W Sultanow F Junginger T Adenocarcinoma of the esophagus: prognostic comparison between transthoracic esophageal resection with expanded 2-field lymph node dissection and trans-hiatal esophageal dissection with abdominal lymph node excision Kongressbd Dtsch Ges Chir Kongr 2002 119 333 338 12704894
Monig SP Baldus SE Zirbes TK Collet PH Schroder W Schneider PM Dienes HP Holscher AH Topographical distribution of lymph node metastasis in adenocarcinoma of the gastroesophageal junction Hepatogastroenterology 2002 49 419 422 11995464
Heroor A Fujita H Sueyoshi S Tanaka T Toh U Mine T Sasahara H Sudo T Matono S Yamana H Shirouzu K Adjuvant chemotherapy after radical resection of squamous cell carcinoma in the thoracic esophagus: who benefits? A retrospective study Dig Surg 2003 20 229 235 discussion 236-237 12759503 10.1159/000070390
Caramella E Bruneton JN Roux P Aubanel D Leconte P Metastases of the digestive tract. Report of 77 cases and review of the literature Eur J Radiol 1983 3 331 338 6653567
Wang M Patel J Casey TT Kieffer R Dunn GD Metastatic squamous cell carcinoma from the esophagus occurring as small bowel obstruction South Med J 1985 78 884 886 3925568
Yamada T Yagi S Tatsuzawa Y Fujioka S Sato H Kitagawa S Nakagawa M Kurumaya H Small intestinal metastasis from esophageal carcinoma associated with small intestinal obstruction: report of a case Surg Today 1996 26 800 802 8897678 10.1007/BF00311639
Thorp MA Carrie S Neck abscess: an unusual presentation of a thoracic malignancy J Laryngol Otol 1998 112 891 892 9876387
Matthew RM Singh S Viswanathan PN Faith RV Esophageal carcinoma with spread to mesenteric and iliac lymph nodes Indian J Gastroenterol 1999 18 125 10407570
Kolbusz R Reyes CV Hakky M Gradini R Asymptomatic esophageal squamous cell carcinoma masquerading as a rare primary panceatic carcinoma. Diagnosis by percutaneous fine needle aspiration Acta Cytol 1988 32 399 402 3376708
Allen HA Bush JE Midesophageal carcinoma metastatic to the stomach: its unusual appearance on an upper gastrointestinal series South Med J 1983 76 1049 1051 6879274
|
16022736
|
PMC1182400
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Jul 16; 3:44
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-44
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-461602662810.1186/1477-7819-3-46ReviewGastrointestinal autonomic nerve tumours – report of a case and review of literature Mulchandani Manoj H [email protected] Dipankar [email protected] John O [email protected] Vickram B [email protected] Department of Surgery, South Tyneside District Hospital, Harton Lane, South Shields, Tyne and Wear, NE34 OPL, UK2 Department of Pathology, South Tyneside District Hospital, Harton Lane, South Shields, Tyne and Wear, NE34 OPL, UK2005 19 7 2005 3 46 46 6 3 2005 19 7 2005 Copyright © 2005 Mulchandani et al; licensee BioMed Central Ltd.2005Mulchandani 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
Gastrointestinal autonomic nerve tumours are uncommon stromal tumours of the intestinal tract. They can involve any part of the gastrointestinal system, but are very rarely seen in the rectum.
Case presentation
We report a unique case of rectal schwannoma with associated synchronous adenocarcinoma of the splenic flexure and adenoma of the descending colon. A 70-year-old patient was admitted with complaint of bleeding per rectum and investigations revealed the presence of a large submucosal rectal lesion in addition to the colonic pathologies. Following panproctocolectomy with permanent spout ileostomy, histopathology and immunohistochemistry confirmed the rectal lesion to be a schwannoma.
Conclusion
Literature review of the few reported cases has suggested radical surgical excision to be the best approach. Prognosis tends to be favourable after resection.
==== Body
Background
The gastrointestinal autonomic nerve tumours (GANTs) were first described and defined by Herrera et al, in 1984 [1]. GANTs are uncommon stromal tumours accounting for 0.1% of benign tumours of the gastrointestinal tract [2]. It is a subgroup of gastrointestinal stromal tumours (GISTs) with specific ultrastructural differences; suggesting its origin from the myenteric plexus [3]. Schwannomas belong to this group and may develop practically in any anatomic region [2].
Without immunohistochemical studies, Schwannomas are often misdiagnosed as leiomyomas or leiomyosarcomas and generally present as an asymptomatic mass and/or with non-specific symptoms of fatigue and prolonged pain as well as signs of low grade pyrexia, anaemia and haemorrhage. Conventional pathological techniques are usually not diagnostic [2]; electron microscopy often being required to establish the diagnosis of GANTs and to exclude them from other gastrointestinal tumours [4].
Common sites for GANTs include the stomach, duodenum, jejunum, ileum [5] and to date literature search has revealed only twenty cases of colonic schwannomas and only four reported cases of rectal schwannomas [2,6-9]. We report a rare case of rectal schwannoma who also had incidental adenocarcinoma and adenoma of the colon.
Case presentation
A 70-year-old gentleman presented with a two-month history of bleeding per rectum and altered bowel habits. There was no history of tenesmus but sensation of incomplete evacuation following defecation as well as a feeling of a lump in the anal canal was present. His appetite and weight had been stable.
Clinically he looked well and abdominal examination revealed no abnormality. Per rectal examination revealed a mass in the posterior wall of the rectum with intact mucosa, which had significantly narrowed the lumen. Colonoscopy revealed the presence of a mass in the posterolateral wall of the rectum extending from 4 to 14 cm from the anal verge, with normal overlying mucosa. In addition, there was also a large sessile polyp in the descending colon and an ulceroproliferative growth at the splenic flexure; histopathology confirmed adenomatous polyp and adenocarcinoma respectively.
Computerised tomography (CT) of the pelvis showed thickening of the distal rectum with an obvious low density poorly enhancing lesion with possible calcification in its wall. Tissue plane between this lesion and prostate was not clearly seen but there was no infiltration of the ischiorectal fat (Figure 1a). Magnetic resonance imaging (MRI) showed a very large well-defined mass in the left presacral / ischio-anal fossa contiguous with rectal wall and thus almost certainly arising from one of the elements of the rectal wall (Figure 1b). Barium enema showed an apple core lesion in the splenic flexure (Figure 2a) as well as narrowing of the distal rectal lumen (Figure 2b).
Figure 1 Imaging of the large bowel 1a) CT scan showing the large lesion in the posterior and left lateral wall of the rectum and 1b) MRI imaging: the distal rectal wall lesion (GANT) is shown to be intramural and free from adjoining structures.
Figure 2 Barium enema a) the apple core lesion at the splenic flexure and b) the extrinsic compression in the distal rectum caused by schwannoma.
The case was fully discussed at the colorectal multidisciplinary team meeting and all the surgical options including left hemicolectomy combined with rectosigmoidectomy plus transverse colostomy were fully evaluated. However, the consensus was that the additional presence of an adenocarcinoma as well as an adenoma higher up in the colon combined with the size of rectal schwannoma (12 cms) made panproctocolectomy with permanent ileostomy the best viable option. Furthermore, this was also the preferred choice of the patient as he opted to have an ileostomy rather than a colostomy. Surgery was performed without complications and the postoperative recovery was uneventful. At 6 months follow-up, the patient remains asymptomatic.
Macroscopic examination of the resected specimen confirmed the malignant tumour in the splenic flexure, the polyp in the descending colon and a 10 × 12 cm tumour with a well-defined pseudocapsule in the postero-lateral wall of the mid and low rectum (Figure 3). The surface of the latter lesion was fusiform with bluish-grey colour and its cut surface was irregularly lobulated with occasional cystic and necrotic areas.
Figure 3 Resected specimen showing malignant tumour at the splenic flexure, polyp in the descending colon and rectal schwannoma.
Histopathology of the splenic flexure lesion showed an invasive moderately differentiated adenocarcinoma (G1; Dukes B; Stage2). The polyp further down the colon was a tubulovillous adenoma with moderate focal dysplasia. The rectal tumour showed irregularly running fascicles of spindly to plump mesenchymal cells. The tumour was fairly vascularised with presence of a number of dilated and engorged vessels. The mitotic count was 4 per high power field. Some parts of the tumour were quite cellular while few areas were hypocellular reminiscent of the Antoni A and Antoni B patterns respectively (Figures 4a and 4b). Immunostaining revealed strong reactivity for neuron specific enolase (NSE), vimentin, CD34 and focal positivity for S100 (Figures 5a, 5b, 5c and 5d). Smooth muscle markers actin and desmin were negative.
Figure 4 Photomicrograph showing a) cellular (Antoni A) and b) myxoid (Antoni B) areas (hematoxylin and eosin ×40).
Figure 5 Immunohistochemical staining of the rectal lesion with a) CD34, b) S100, c) Vimentin and d) NSE (original magnification ×40).
Discussion
Gastrointestinal autonomic nerve tumours represent a distinct but rare subcategory of gastrointestinal stromal tumours accounting for 1% of all malignant gastrointestinal tumours. Initially described as plexosarcomas, these tumours have been reported to be more common in males and have a wide spectrum of age range [5]. Literature review has shown rare association of GANTs with neurofibromatosis [10-12] and adrenal ganglioneuroma [13]. Schwannomas are types of gastrointestinal tract autonomic tumours and amongst this group, rectal schwannomas are very rare [13,14]. Furthermore, the case described here, was associated with adenocarcinoma and adenoma of the colon (Figure 3) and to date, there has been no reports of a similar association. Chance or an epigenetic event could explain such association of the different synchronous lesions.
Conventional imaging modalities such as barium enema, colonoscopy, computerised tomography and MRI have been used to investigate these patients but there are no definite radiological criteria to differentiate benign from malignant stromal tumour [15]. As shown in Figures 1a and 1b, CT and MRI scans located the large encapsulated lesion as arising from the rectal musculature, but yielded no additional information which allow for the distinction of benign from malignant potential. Recently Levy et al, [16] have reported that CT imaging may help to differentiate GANTs from gastrointestinal stromal tumours (GISTs). Low attenuation features (indicative of haemorrhage, necrosis and degeneration typically found in the centre of the GISTs) was not seen in our case (Figure 1a); thus favouring GANTs. Endoscopic ultrasound has been suggested to be reliable in predicting malignancy and the predictive features being irregular margins, depth of penetration, cystic spaces and lymph nodes with a malignant pattern [17]. However, it does not differentiate GANTs from the other stromal tumours [15]. Endoscopic ultrasound guided fine needle aspiration with immunohistochemical analysis may be useful in the preoperative diagnosis of GIST [18] but in our patient, this was not done, as it would not have altered the treatment, due to the presence of synchronous tumours in the colon and also the size of the lesion.
Although GANTs do exhibit a variety of specific histological features, Lauwers et al, [10] believe that neither individual cell characteristics nor the growth pattern of these tumours allow distinction from GISTs. The differentiation of schwannomas from other stromal tumours is important because the latter group has high-risk of malignant behaviour [19]. Furthermore, whilst some authors disagree [5], others have suggested the presence of both Antoni A / Verocay bodies (cells forming a typical palisade arrangement in a well-organised pattern) and Antoni B (small lacunar foci with loss of palisade architecture) areas to be very specific for schwannoma [20]. Such patterns have been clearly demonstrated in our case (Figures 4a and 4b).
Immunohistochemical studies of GANTs have usually demonstrated positivity to vimentin, CD34 and CD117 [5,10,21]. Positive reactivity with neuron specific enolase (NSE), S-100 protein, synaptophysin, and chromogranin A (proteins expressed by neurons from the autonomic nerve plexus) have also been reported and this support the histogenesis of GANTs from the autonomic plexus of Meissner or Auerbach [5]. Furthermore, positive immunoreactivity to the S-100 protein and Leu7 antigen tend to indicate the Schwannian nature of the tumour [22], whereas positivity to the glial fibrillary acidic protein point towards a myenteric plexus origin [3]. GANT is an ultrastructural variant of GIST, based on its consistent CD117 positivity and the presence of GIST-specific c-kit gene mutations in a significant number of cases [23]. The published GANT tumours have been variably and inconsistently S-100 protein positive. In our patient, immunohistochemistry revealed reactivity for vimentin, NSE, CD34 and for S-100 (Figure 5). Desmin and actin were negative.
Whilst immunohistochemistry provides important diagnostic differentiation of GANTs from other stromal tumours, definitive diagnosis can only be based on ultrastructural studies [5]. The ultrastructural criteria that suggest origin from myenteric plexus are neuron-like cells with long cytoplasmic processes containing microtubules, bulbous synapse like structures with dense core neurosecretory type granules and empty vesicles [10]. Ultrastructural examination is not available at most pathological units (including ours); this could explain the paucity of these cases and it is suggested that representative samples should be sent to centres having electron microscopic facilities.
The exact biological behaviour of GANTs is not yet fully elucidated due to the limited number of reported cases and as a result, determination of malignancy poses a difficult challenge, which cannot always be resolved by conventional histopathology. Various parameters have been studied in relation to tumour behaviour and to date; no single one is fully predictive of malignancy. Mitotic activity (counts >5 mitoses per high power fields) and tumour size (> 5 cm) tend to be associated with a high risk of metastasis or recurrence [24]. Although GANTs are generally considered benign, Lauwers et al, reported that 30% of these patients developed local recurrence [12] and as a result, radical surgery is the optimal treatment [5]. In our case, the additional presence of an adenocarcinoma and adenoma higher up in the colon combined with the size of the rectal schwannoma (>12 cms) made pan-proctocolectomy with permanent ileostomy the better option.
The recent finding of CD117 receptors in these tumours combined with technological advances has led to the development of anti-tumour agents. Treatment of such CD117 positive GANTs with tyrosine kinase inhibitors have been shown to be beneficial [25] and could in future, represent an appropriate form of palliative therapy in those patients with unresectable as well as metastatic tumours [5].
Conclusion
Further studies are required not only to fully characterise the molecular biology of these tumours but also their aggressiveness as there is an inherent difficulty in making a benign diagnosis in GANTs. This is clearly important since it may result in a poor cross over in the assessment of the responsiveness of GANTs to anti-CD117 treatment.
Competing interests
The author(s) declare that they have no competing interest.
Authors' contributions
MMH collected the information, did literature search and wrote the manuscript.
DC assisted in writing the manuscript.
OJO assisted in pathological view and the microscopic pictures.
JBV helped in preparing the manuscript and edited the final version.
All authors have read and approved the final version.
Acknowledgements
Written consent was obtained from the patient for the publication of case report.
==== Refs
Herrera GA DeMoraes HP Grizzle WE Malignant small bowel neoplasm of enteric plexus derivation (plexosarcoma): Light and electron microscopic study confirming the origin of the neoplasm Dig Dis Sci 1984 29 275 284 6321118 10.1007/BF01296263
Maciejewski A Lange D Wloch J Case report of schwannoma of the rectum – clinical and pathological contribution Med Sci Monit 2000 6 779 782 11208409
Daimaru Y Kido K Hashimoto H Ejoji BenignM Schwannoma of the gastrointestinal tract: a clinicopathologic and immunohistochemical study Hum Pthol 1988 19 257 264
Eyden B Chorneyko KA Shanks JII Menasce LP Banerjee SS Contribution of electron microscopy to understanding cellular differentiation in mesenchymal tumours of the gastrointestinal tract: a study of 82 tumours Ultrastruct Pathol 2002 26 269 285 12396237 10.1080/01913120290104548
Stift A Friedl J Gnant M Herbst F Jakesz R Wenzl E Gastrointestinal autonomic nerve tumours: A surgical point of view World J Gastoenterol 2004 10 2447 2451
Bhardwaj K BAL MS Kumar P Rectal schwannoma Indian J Gastroenterol 2002 21 116 117 12118926
Catania G Puleo C Cardi F Iuppa A Buffone A Malignant schwannoma of the rectum: a clinical and pathological contribution Chir Ital 2001 53 873 877 11824066
Murakami N Tanaka T Ohmuri Y Shirouzu Y Ishibashi S Harada Y Yatsuka K Jimi A Shirouzu K A case report of rectal schwannoma Kurume Med J 1996 43 101 106 8709552
Miettinen M Shekitka KM Sobin LH Schwannomas in the colon and rectum: a clinicopathologic and immunohistochemical study of 20 cases Am J Surg Pathol 2001 25 846 855 11420455
Lauwers GY Erlandson RA Casper ES Brennan MF Woodroff JM Gastrointestinal autonomic nerve tumours. A clinicopathological, immunohistochemical and ultrastructuralstudy of 12 cases Am J Surg Pathol 1993 17 887 897 8394653
Dhimes P Lopez-Carreira M Ortega-Serano MP Garcia Munoz II Martinez-Gonzalez MA Ballestin C Gastrointestinal autonomic nerve tumours and their separation from other gastrointestinal stromal tumours: an ultrastructural and immunohistochemical study of seven cases Virchows Arch 1995 426 27 35 7704320 10.1007/BF00194695
Sakaguchi M Sano K Ito M Baba T Fukuzawa M Hotchi M A case of von Recklinghausen's disease with bilateral pheochromocytoma-malignant peripheral nerve sheath tumours of the adrenal and gastrointestinal autonomic nerve tumours Am J Surg Pathol 1996 20 889 897 8669538 10.1097/00000478-199607000-00013
Erlandson RA Klimstra DS Woodruff JM Subclassification of gastrointestinal stromal tumours based on evaluation by electron microscopy and immunohistochemistry Ultrastruct Pathol 1996 20 373 393 8837346
Lev D Kariv Y Messer GY Isakov J Gutman M Gastrointestinal autonomic nerve (GAN) tumour of the rectum J Clin Gastroenterol 2000 30 138 140 10.1097/00004836-200006000-00018
Rueda O Escrlbano J Vicente JM Garcia F Villeta R Gastrointestinal autonomic nerve tumours (plexosarcomas). Is a radiological diagnosis possible? Eur Rad 1998 8 458 460 10.1007/s003300050413
Levy AD Quiles AM Miettinen M Sobin LH Gastrointestinal schwannomas: CT features with clinicopathologic correlation AJR Am J Roentgenol 2005 184 797 802 15728600
Palazzo L Landi B Cellier C Cuillerier C Roseau G Barbier J-P Endoscopic features predictive of benign and malignant gastrointestinal stromal cell tumours Gut 2000 46 88 92 10601061 10.1136/gut.46.1.88
Nobuhiro A Hidemi G Yasumasa N Yoshiki H Naoki O Tetsuo N Tetsuo H The diagnosis of stromal tumours with EUS-guided fine needle aspiration with immunohistochemical analysis Gastrointest Endosc 2002 55 37 43 11756912 10.1067/mge.2002.120323
Kwon MS Lee SS Ahn GH Schwannomas of the gastrointestinal tract: clinicopathological features of 12 cases including a case of oesophageal tumour compared with those of gastrointestinal stromal tumours and leiomyomas of the gastrointestinal tract Pathol Res Pract 2002 198 605 613 12440783
Das Gupta TK Brasfield RD Strong ET Hajdn SI Benign solitary schwannomas (neurilemmomas) Cancer 1969 24 355 366 5796779
Lee JR Joshi V Griffin JW JrLasota J Miettinen M Gastrointestinal autonomic nerve tumour: Immunohistochemical and molecular identity with gastrointestinal stromal tumour Am J Surg Pathol 2001 25 979 987 11474281 10.1097/00000478-200108000-00001
Arai T Sugimura H Suzuki M Benign schwannoma of the oesophagus: Report of two cases with immunohistochemical and ultrastructural studies Pathol Int 1994 44 460 465 8055113
Gibson PC Cooper K CD117 (KIT) A diverse protein with selective applications in surgical pathology Adv Anat Pathol 2002 9 65 69 11756760 10.1097/00125480-200201000-00007
Davila RE Faigel DO GI stromal tumors Gastrointest Endosc 2003 58 80 88 12838226 10.1067/mge.2003.317
Joensuu H Fletcher C Dimitrijevic S Silberman S Roberts P Demetri G Management of malignant gastrointestinal stromal tumours Lancet Oncol 2002 3 655 664 12424067 10.1016/S1470-2045(02)00899-9
|
16026628
|
PMC1182401
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Jul 19; 3:46
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-46
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1614984910.1371/journal.pbio.0030282Research ArticleDevelopmentEcologyEvolutionGenetics/Genomics/Gene TherapyZoologyMolluscsSpeciation and Gene Flow between Snails of Opposite Chirality Chirality and Speciation in SnailsDavison Angus [email protected]
1
2
3
Chiba Satoshi
1
Barton Nicholas H
2
4
Clarke Bryan
3
1Graduate School of Life Sciences, Tohoku University, Aramaki-Aza-Aoba, Aoba-ku, Japan,2Institute of Evolutionary Biology, School of Biological Sciences, University of Edinburgh, Edinburgh, United Kingdom,3Institute of Genetics, School of Biology, University of Nottingham, Nottingham, United Kingdom,4Department of Genetics, University of Cambridge, Cambridge, United KingdomButlin Roger Academic EditorUniversity of SheffieldUnited Kingdom9 2005 9 8 2005 9 8 2005 3 9 e28229 10 2003 10 6 2005 Copyright: © 2005 Davison 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.
Speciation Begins, but Doesn't End, with the Twist of a Shell
Left-right asymmetry in snails is intriguing because individuals of opposite chirality are either unable to mate or can only mate with difficulty, so could be reproductively isolated from each other. We have therefore investigated chiral evolution in the Japanese land snail genus Euhadra to understand whether changes in chirality have promoted speciation. In particular, we aimed to understand the effect of the maternal inheritance of chirality on reproductive isolation and gene flow. We found that the mitochondrial DNA phylogeny of Euhadra is consistent with a single, relatively ancient evolution of sinistral species and suggests either recent “single-gene speciation” or gene flow between chiral morphs that are unable to mate. To clarify the conditions under which new chiral morphs might evolve and whether single-gene speciation can occur, we developed a mathematical model that is relevant to any maternal-effect gene. The model shows that reproductive character displacement can promote the evolution of new chiral morphs, tending to counteract the positive frequency-dependent selection that would otherwise drive the more common chiral morph to fixation. This therefore suggests a general mechanism as to how chiral variation arises in snails. In populations that contain both chiral morphs, two different situations are then possible. In the first, gene flow is substantial between morphs even without interchiral mating, because of the maternal inheritance of chirality. In the second, reproductive isolation is possible but unstable, and will also lead to gene flow if intrachiral matings occasionally produce offspring with the opposite chirality. Together, the results imply that speciation by chiral reversal is only meaningful in the context of a complex biogeographical process, and so must usually involve other factors. In order to understand the roles of reproductive character displacement and gene flow in the chiral evolution of Euhadra, it will be necessary to investigate populations in which both chiral morphs coexist.
Is single gene speciation possible in snails? Although feasible, the authors reveal that reproductive isolation between chiral morphs is possible, but unstable, and will lead almost inevitably to gene flow.
==== Body
Introduction
Left-right asymmetry is an integral part of the establishment of a body plan that may ultimately be traced back to a much deeper molecular asymmetry [1,2]. Although substantial progress has been made recently towards understanding the establishment of asymmetry in several organisms [3–5], almost nothing is known about the first stages in any embryo. Snails may be crucial tools for studying left-right asymmetry, because their chirality is determined at a very early cleavage, and because several species have morphological variation, so that the gene(s) involved can be mapped [6]. It is intriguing that molluscan asymmetry is determined by the effects on the developing embryo of a maternal “chirality” gene (or series of closely linked loci). The genotype of the mother determines the phenotype of the offspring [7–11].
Some recent work has focussed on the issue of why snails are almost always invariant in chirality. One approach has been to study the exceptions, species that are chirally dimorphic, but even in these most populations are fixed for a particular type [11–14] (with the exception of Amphidromus [15]). This is because positive frequency-dependent selection tends to drive the chiral majority to fixation; rarer snails of opposite chirality are less likely to find a mate and successfully copulate [13]. In the best-studied case, the Polynesian tree snail Partula suturalis, dimorphic populations occurred in narrow clines between areas of dextral or sinistral snails [12,13]. In “no choice” experiments, interchiral matings occurred only 20% as frequently as intrachiral matings, and fewer young were produced by the mixed pairs [13]. Since dextral populations of P. suturalis tended to coincide geographically with closely related sinistral species, it was suggested that dextral P. suturalis became established because of reproductive character displacement: Dextral individuals were favoured even when rare, because they were less likely to waste time (or gametes) mating with sinistral species [12,13,16].
More general aspects of shell morphology in land snails have been of interest since the observation by Cain that members of widely separate and taxonomically distinct land snail faunas tend to be either high- or low-spired, with relatively few globular forms in between [17]. Following the suggestion of Gittenberger that there is an association between shell shape and the degree of variation in asymmetry [18], Asami et al. went on to report that reciprocal mating between dimorphic low-spired snails is not usually possible, because the genitalia of a sinistral individual cannot engage with those of a dextral snail [14]. In contrast, high-spired dimorphic snail species such as Partula are able to mate, albeit with some behavioural adjustments [14,19,20].
Interchiral mating in snails, in conjunction with the maternal inheritance of the gene, has attracted attention because it has been thought capable of producing “single-gene speciation” in sympatry ([18,21–24], but see [25]). Simulations have been used to establish whether chirally reversed snails can reach a sufficiently high density to overcome the effects of positive frequency-dependent selection. Johnson et al. considered that sympatric speciation due to coil reversal was unlikely in Partula, because new variant snails would have difficulty finding partners of the same chiral morph, and because the maternal inheritance allowed free gene flow between the morphs [25]. However, simulations by Orr [23], van Batenburg and Gittenberger [26], and Stone and Björklund [27] suggested that in some other circumstances single-gene speciation could occur. It is interesting that the establishment of a new chiral morph in snails has a parallel in Müllerian mimetic species, where new models can only become established when they reach a threshold density [28,29]. The positive frequency-dependent selection that creates sharp clines in the frequencies of P. suturalis chiral morphs could similarly maintain the narrowness of hybrid zones between geographic races of Heliconius butterflies [28].
We chose to investigate the Japanese land snail genus Euhadra (Bradybaenidae) (Figure 1), precisely because it is the best candidate for single-gene speciation. Five of the 22 Euhadra species are sinistral (Figures 1 and 2), a situation that is particularly unusual because most other low-spired snail genera are invariant for their chirality. If single-gene speciation can ever occur, then it is most likely to occur in Euhadra, because interchiral mating is either difficult or impossible [14]. We were interested to know how many times sinistral Euhadra have evolved, and whether chiral variation has promoted speciation. We tested these ideas using a phylogeny of mitochondrial DNA gene fragments from 19 of the species, including several representatives from each of the sinistral taxa (Tables 1 and S1). As it was necessary to understand the conditions under which genes might flow between different chiral morphs (i.e., whether reproductive isolation and single-gene speciation is theoretically possible), and to make predictions of the genealogical patterns that we expect to see, we have developed a mathematical model of the population genetics of chirality (Protocol S1).
Figure 1 Sinistral Species
(A) E. quaesita from Sendai and (B) E. murayamai from Myojo-san (in cave); dextral (C) E.
senck.
senckenbergiana from Imajyo, (D) E. senck.
amoriensis from Tamayama, (E) E. senck.
amoriensis from Iide-san, and (F) E. senck. ibukicola from Mt. Fujiwara. [Photos: AD and SC]
Figure 2 Distributions of Four Euhadra Species on the Main Japanese Island of Honshu
E. senck.
aomoriensis specimens that were used for DNA analysis were collected from Tamayama, Tsugaru, and Iide-san. E. murayamai is confined to a single mountain (Myojo-san). The remaining marked sites refer to sites of snails in Figure 1, as well as to others that are specifically referred to in the main text. Sinistral species are shown in red and dextral species in blue.
Table 1 Species and Collection Localities of Samples Used for DNA Analysis
Table 1 Continued
Results
Phylogenetic Analysis of Gene Sequences
As expected from earlier studies [30–32], mitochondrial variation was extreme both within and between species. Only the 16S rRNA fragment had a level of variation suitable for resolving both the deeper relationships and those within species (881 bp when aligned, of which 639 bp were used for the phylogeny; for GenBank references, see Table 1). The variation within 12S rRNA was confined to several hypervariable regions. Cytochrome oxidase c synonymous nucleotide positions were saturated and its amino acids barely variable. There was also not enough variation within the nuclear sequence of ITS2. These latter three gene fragments were therefore not used further (GenBank AY445024–AY445027, AY251858–AY251871, and AY251820–AY251837).
The sinistral species Euhadra quaesita (Figure 1A), E. decorata,
E. grata,
E. scaevola, and E. murayamai (Figure 1B) were confined to a single clade in the 16S rRNA neighbour-joining phylogeny (Figure 3), which also contained the dextral species E. eoa,
E. senckenbergiana (E. senck.; see Figure 1C–1F), and E. latispira. Bootstrap support for the branch that defined this group is 76%, so the simplest explanation is that sinistral Euhadra evolved once only. The maximum likelihood phylogeny, using representative sequences from each clade, had the same overall structure, and the same as that discovered by Ueshima and Asami using a different mitochondrial gene fragment [24].
Figure 3 Neighbour-Joining Phylogeny
16S rRNA rate-corrected neighbour-joining phylogeny, rooted using Nesiohelix bipyramidalis, showing the relationship between mitochondrial DNA lineages from dextral and sinistral Euhadra. The most parsimonious explanation is a single, relatively ancient evolution of sinistral snails. Lineages within the box are shown in detail in Figure 4. Bootstrap support of more than 70% is shown below the node. Shape parameter = 0.30.
The relationships within the clade that contains the sinistral species were surprisingly complicated (Figures 3 and 4). Six lineages were found: individual E. decorata,
E. eoa,
E. grata, and E. scaevola lineages, plus two multi-species lineages, one comprising two species, E. senck. and E. latispira, and the other three species, E. quaesita,
E. senck., and E. murayamai. The latter lineage group divides into five subgroups (Figure 4), two that have not been discovered before (QUAI and II), and three already reported (QUAIII, IV, and V) [33]. The geographic distributions of these subgroups are principally Tohoku and North East Chubu (QUAI and II), Kanto and Miura peninsula (QUAIII), Kanto and Tohoku (QUAIV), and Kanto and Izu peninsula (QUAV). All the major lineages within the clade that contains the sinistral species were strongly supported by bootstrapping, but their relationship to each other was not resolved (Figures 3 and 4).
Figure 4 Neighbour-Joining Phylogeny Subtree
16S rRNA rate-corrected neighbour-joining phylogeny (subtree from Figure 3) showing the relationship between E. quaesita,
E. senck. aomoriensis, and E. murayamai mitochondrial lineages. All unlabelled tips are E. quaesita. Both E. senck. aomoriensis and E. murayamai are polyphyletic. The polyphyly of E. senck. aomoriensis can be explained by either single-gene speciation or mitochondrial introgression. Bootstrap support shown for important nodes only.
Mitochondrial lineages of sinistral E. murayamai from Myojo-san, as well as dextral E. senck.
aomoriensis from Iide-san, Tamayama, and Tsugaru were within the E. quaesita,
E. senck., and E. murayamai mitochondrial clade (Figure 4; DNA sample sites and distributions are shown in Figure 2, with further details in Figure S1 and Table S1). E. murayamai is found on a single mountain, yet its mitochondrial haplotypes were found in QUAI and II, nested with haplotypes from E. quaesita. Individual haplotypes of dextral E. senck.
aomoriensis were found in QUAI, II, and IV (n = 6), three clades of which are predominantly found in Tohoku, but also North East Chubu and Kanto. The other subspecies of E. senck. grouped with E. latispira. Within the QUAII lineages (Figure 4), snails were therefore either sinistral E. murayamai from Myojo-san (North East Chubu), sinistral E. quaesita from Haguro-san (central Tohoku), or dextral E. senck.
aomoriensis from Tsugaru and Tamayama (northern Tohoku).
Analysis of Morphological Characters
The morphological analysis showed that both E. senck. aomoriensis and E. senck. notoensis have genitalia similar to those in E. quaesita and E. murayamai (Table S2). Specifically, ten of 18 E. senck. aomoriensis genital characters were identical to those of E. quaesita and different from those of E. senck. senckenbergiana, whereas only one of 18 was similar to E. senck. senckenbergiana and different from E. quaesita (Table S2). The morphology of the genitalia in the other E. senck. subspecies (E. senck. senckenbergiana,
E. senck. minoensis, and E. senck. ibukicola) was similar to that of E. latispira. E. senck. aomoriensis,
E. senck. notoensis, E. quaesita, and E. murayamai grouped together in the genital character maximum parsimony tree, distinct from the other E. senck. subspecies as well as from E. latispira (Figure 5A). The analysis of shell characters showed that there is considerable variation between populations (Figure 5B; Table S3). Consequently, different populations of the same species group closely with other species in the parsimony analysis of shell characters (Figure 5B).
Figure 5 Maximum Parsimony Tree of Two Characters
The tree is based on (A) genital characters and (B) shell characters. The number after each species is the sample location (see also Tables S1–S3). Sinistral species are shown in red and dextral species in blue.
In summary, E. senck. aomoriensis and E. quaesita specimens grouped together in the mitochondrial DNA phylogeny and have similar genitalia. In contrast, while E. senck. notoensis samples have similar genitalia to E. senck. aomoriensis and E quaesita, their mitochondrial DNA sequences grouped with the other E. senck. subspecies (Figure 5B).
Analysis and Interpretation of the Model
Because the chirality gene is maternally expressed, an individual's phenotype may not correspond to its genotype. This means that we needed to follow the proportions of five kinds of snail: Assuming for the moment that the sinistral allele (S) is dominant over the dextral allele (D), sinistral snails may contain SS, SD, or DD genotypes, whereas dextral snails may be SD or DD (Table S4). No dextral snails can have a SS genotype, because all snails with the SS genotype inherit an S allele from their mother, and so are sinistral. The proportions of the five kinds of snail tend to a characteristic equilibrium that can be thought of as analogous to the Hardy-Weinberg proportion (Table S5). Full details of our model are described in the Protocol S1. Here, an overview is given with an emphasis on findings that are directly relevant to Euhadra.
Briefly, the model was used to find the equilibrium frequencies of the chirality gene in each chiral morph, depending upon whether there is random or nonrandom mating (i.e., either interchiral mating with no assortment between chiral morphs, α = 0, or complete assortment with no interchiral mating, α = 1, respectively). The model is powerful because it predicts the frequency of the sinistral offspring from sinistral morphs as a function of the frequency of sinistral morphs in the population (Figure 6). We expected that the model would be useful in understanding the mitochondrial phylogeny because, by extension, if chirality genes are able to flow freely between chiral morphs, then so will any other gene regardless of whether it is physically linked (on the same chromosome) to the chirality gene or not. Throughout the analysis, we assumed that the allele for sinistral coiling is dominant, as in Partula, so the interpretation would have to be modified if dextral is dominant in Euhadra.
Figure 6 Proportion of Sinistral Snails Amongst Offspring of Sinistrals Compared with Frequency of Sinistrals in the Population
The curves are the boundaries defined by extreme values of α, the parameter that describes the degree of interchiral mating (lower curve: random mating between chiral morphs, α = 0; upper curve: no interchiral mating, α = 1). The space between the two curves represents intermediate values of α. The intercepts illustrate the predictive nature of the model; e.g., since interchiral mating is not possible in Euhadra (α = 1), then if a mixed population of Euhadra contains 40% sinistrals, the model predicts that 80% of the offspring of sinistral snails should be sinistral. For a species with no barriers to interchiral mating (α = 0), the model predicts that 65% of the offspring should be sinistral. Finally, if the mixed population is actually two separate species that might have formed by single-gene speciation, then the prediction is that sinistral snails should always give birth to sinistral offspring (y = 1, indicated by the dotted line; this is the second, unstable equilibrium that is referred to in the main text). Equilibria are calculated from Equation 5 in Protocol S1.
We were able to account for a variety of other factors as well as those determining whether mating is random or not. They included frequency-dependent selection, in which rare chiral morphs are less likely to find a mate, and selection against hybrid mating, which might occur if snails are less likely to transfer sperm in interchiral mating. We also considered the effects of reproductive character displacement; i.e., a sinistral species would be at a disadvantage if it risked mating with another sympatric sinistral species when hybrid matings between the two species are sterile or when hybrid offspring are less fit. If a new dextral morph arose within a sinistral species, then while it might initially be at a disadvantage (it would have difficulty finding morphs of the same chirality to mate with), the disadvantage could be counterbalanced if the dextral individuals did not waste time mating with the other species. This explanation for the origin of new chiral morphs has been suggested for Partula [12], but has not been tested theoretically.
We first investigated the extent of gene flow between different chiral morphs, depending upon their degree of interchiral mating (Section 2 in Protocol S1), using a parameter, α, to define the degree of interchiral or assortative mating. Two equilibria are possible in a chirally mixed population. In the first, gene flow between chiral morphs is substantial, sinistral mothers produce a large proportion of dextral offspring (and dextral mothers produce a large proportion of dextrals), and the population rapidly moves to a balanced equilibrium, even if there is absolutely no interchiral mating (α = 1). This high degree of gene flow is a direct consequence of the delayed (maternal) expression of chirality. The implication for Euhadra is that gene flow between different chiral morphs is inevitable, even if the morphs are themselves unable to mate.
We have illustrated this first equilibrium in Figure 6. In this figure, the axes represent the proportion of sinistrals in the population (x-axis) and the proportion of sinistral offspring from sinistral mothers (y-axis). These two parameters were used to illustrate the model because they represent data that can be gathered in the field, and because they have predictive value. For example, in Euhadra, interchiral mating is either not possible (α = 1, upper line in Figure 6) or very difficult (for example, when α ≈ 0.99). At sites where dextral and sinistral Euhadra coexist, we can measure the frequencies of each. If sinistrals are 40% of all snails collected, and different chiral morphs are not able to mate (α = 1), then the model predicts that 80% of the offspring of sinistral snails will be sinistral. In contrast, if sinistral and dextral chiral morphs were able to mate freely (α = 0; unlikely in Euhadra, but possible in other species), then the model predicts that 65% of the offspring of sinistral snails will be sinistral. Finally, if there is no gene flow, then sinistral snails will not give birth to dextral individuals (and vice versa). This is the second equilibrium (the upper dotted line in Figure 6), which corresponds to single-gene speciation, and is discussed below.
The second equilibrium is a special case because it is the only equilibrium in which there is no gene flow between the two morphs, so that they are reproductively isolated. It can only occur if the different morphs are completely unable to mate (α = 1). However, this equilibrium is unstable and will return to the first equilibrium (free gene flow between chiral morphs) if any sinistral snails have recessive dextral alleles (or alternatively, dextral snails have dominant sinistral alleles), or if there is a very low frequency of interchiral mating (e.g., α ≈ 0.99). The only requirement, therefore, to enable gene flow between different chiral morphs in this second equilibrium is that intrachiral matings occasionally produce offspring of the opposite asymmetry to the mother, or else that interchiral mating occasionally occurs. Intrachiral matings could produce offspring of the opposite asymmetry by either de novo gene mutation or accidents of development, which can happen when embryos develop in extreme environments. At present we do not know the frequency of interchiral mating in species such as Euhadra. Although no interchiral matings have been observed, they may be possible (i.e., α ≤ 1).
We next considered the effects of a lower frequency or fertility of mating in the rarer chiral morph, giving the more-common morph an advantage (Figure 7; section 3.i in Protocol S1). To simplify the equations, we also assumed that differences between selection on “males” and “females” are not of significant effect, which is reasonable since most snails are hermaphrodites. We imagine that fertility is reduced in interchiral mating because sperm is less likely to be transferred. Perhaps more importantly, an individual of the rare chirality might produce fewer offspring, because it will waste time searching for a suitable mate.
Figure 7 Proportion of Sinistral Snails Amongst Offspring of Sinistrals Compared with Frequency of Sinistrals in the Population, with Reduced Mating Frequency or Fertility for the Rare Chiral Morph
We assumed a strong frequency-dependent mating disadvantage against the rare morph, no interchiral mating (α = 1) and equal sexual selection on the two sexes. The graph shows eight runs that start close to the border separating two alternative outcomes, which runs approximately vertically down the middle of x-axis of the graph (when the frequency of sinistrals is about 0.5). As there was no interchiral mating, then if the starting condition is two distinct populations (sinistral snails have only sinistral alleles and vice versa), then the two chiral morphs stay separate: the points along the top horizontal line represent this, indicating competition between two reproductively isolated types, or single-gene speciation (y = 1; as in Figure 6, this is also the second, unstable equilibrium that is referred to in the main text). In all other circumstances, the model shows that the population moves to equilibrium, so that there is extensive gene flow between different chiral morphs. For further information, see Results or section 3.i in Protocol S1.
Although the model and its equations became more complicated when the above selective factors were included (section 3.i in Protocol S1), the results were remarkably similar to the first set of circumstances (described above). With random mating (no barriers to interchiral mating, α = 0), the population moves close to an equilibrium level of gene flow within two generations, even when selection is close to its maximum value. With complete assortment (no interchiral mating, α = 1), the population moves to equilibrium rather more slowly, since gene flow between morphs is somewhat restricted, but eventually reaches a similar position (Figure 6).
We used the same axes as in Figure 6 to illustrate the equilibria that are reached when selection is included (Figure 7). As the description of the outcome is somewhat more complicated, we used the figure to illustrate the extreme condition only, where no interchiral mating is possible (α = 1). The graph shows eight trajectories that start close to the boundary separating two alternative outcomes. As there was no interchiral mating, if the initial condition is two separate subpopulations (i.e., sinistral snails contain only sinistral alleles and vice versa), then the two types stay distinct: the population moves along the top horizontal (dotted) line, with competition between two reproductively isolated types. As before, this situation is equivalent to single-gene speciation. However, if there are any chiral morphs that contain the opposite allele, then the population will move towards an apparent equilibrium in which there is extensive gene flow between the morphs (dotted line running from bottom left to top right in Figure 7), irrespective of the starting frequency of chiral phenotypes and genotypes. Once near this quasi-equilibrium, frequency-dependent selection moves the population towards fixation of one or other morph. Under these conditions, the opportunity for single-gene speciation must be very limited, and even more so if interchiral mating is possible (α < 1). Moreover, if single-gene speciation does occur, then as before, it is unstable because of the introduction of new chiral alleles or accidental reversals of chirality.
Another consideration was that selection might act against hybrids between different chiral morphs (section 3.ii in Protocol S1), because genetic differences leading to postmating isolation might be held in disequilibrium with the chirality locus itself. While the chirality gene and genes that determine whether hybrids are less fit (or not) are not likely to be physically linked (on the same chromosome), they may be associated because of coincident selection against the physical act of interchiral mating and because the offspring from interchiral matings tend to be less fit. While this seems unlikely because gene flow should be extensive between different chiral morphs (recombination will break down the disequilibrium), selection against hybrids could occur if the different chiral morphs originally evolved in distinct populations and then came together into a single population. In the most extreme circumstance, all heterozygotes would die, interchiral matings would not produce viable offspring, and the system would be stable.
We also considered a more realistic model, in which heterozygotes have their fertility or survival reduced by a fixed proportion (section 3.ii in Protocol S1). As in the former models, except when there is an almost complete lack of interchiral mating (α > 0.99), the model shows that there should be high gene flow between morphs, so that heterozygotes are abundant, and selection causes fixation of the more-common allele relatively quickly. When assortment is almost complete (α ≈ 1), there is a narrow range of parameters with two alternative ways for the population to reach fixation: either rapidly, with high gene flow between morphs, or much more slowly, with little gene flow. Nonetheless, the outcome in all circumstances (except when all heterozygotes die) is the same: extensive gene flow between chiral morphs.
Up to this point, we assumed a single population, whereas the more-common circumstance may be that most populations are fixed in their chirality, with narrow clines between them, as in Partula [12]. The model suggests that regardless of whether mating is random or there is no interchiral mating at all, the position of the cline will move in favour of the dominant allele (section 4.i in Protocol S1) because the dominant allele tends to increase the frequency of its corresponding phenotype following interchiral matings. This conclusion was reached independently by Johnson et al. [25] and Mallet (“dominance drive”) [34]. It is interesting that the shape of the cline will be almost identical regardless of the degree of assortative mating, and in all cases long tails of introgression are expected on either side of the centre of the cline.
A final use of the model was to try to understand the conditions under which new chiral morphs can become established in the first place (section 5 in Protocol S1). If the population were initially entirely sinistral, then positive frequency-dependent selection would act to prevent the establishment of dextral morphs. This is because new chiral morphs would not be able to find a mating partner. However, if a proportion of all matings would otherwise be with another sinistral species, and these matings did not give viable offspring (or the hybrids were less fit), then the mechanical isolation that leads to lower fitness in attempted interspecific interchiral matings would give the new chiral morphs an advantage, because of reproductive character displacement. The model shows that the dextral morph would be at an advantage, even when rare, if at least one-third of matings would otherwise be with another species. Therefore, under conditions where reproductive character displacement is favoured, it is considerably more likely that a new morph will become established. If the new chiral morph is due to a dominant mutation, then it will move forwards in any cline that is created, due to the “dominance drive” mentioned above.
Discussion
A Single Evolution of Sinistral Snails
The mitochondrial phylogeny is consistent with a single and relatively ancient origin of sinistral Euhadra. After the new sinistral morph became established, five sinistral species evolved, now distributed through central and northern parts of the main Japanese island of Honshu (see Figures 1 and 2). An intriguing question is how the sinistral morph spread and then speciated, despite positive frequency-dependent selection against the rare chiral morph. The focus of most simulations to date has been to understand the relative contributions of genetic drift and gene flow in small populations [23,25–27]. An alternative that could assist the establishment of a new morph is reproductive character displacement. Although this explanation was first suggested by Clarke and Murray in relation to Partula [12], it has not been tested theoretically. Reproductive character displacement could be a general explanation for many instances of chiral reversal, including Euhadra, the clausilid Isabellaria, and Diplommatina [35–37]. It could also explain the observation that chirally reversed species are more often found on isolated islands, because they are often inhabited by large clades of closely related species.
In Partula, character displacement probably acted so that dextral morphs of otherwise sinistral species such as P. suturalis were less likely to hybridize with sympatric, sinistral congeners. In a similar manner, a newly arising sinistral morph in a dextral Euhadra species would be less likely to hybridize with other dextral Euhadra species. Thus, a common explanation of how a new chiral morph was established in Partula, Diplommatina, Euhadra, and perhaps other species is that initially rare morphs were stabilized against positive frequency-dependent selection by reproductive character displacement. While detailed observations are presently available only for Partula, we used the mathematical model to investigate what might be a general mode of divergence. We found that the new morph would be at an advantage even when rare if at least a third of all matings would otherwise be with another species. In these conditions, the new morph would be fixed and begin to spread out from the region in which it originated.
If the new morph went on to establish a population beyond the region of overlap with the other chiral morph, a cline might develop, as presumably happened in P.
suturalis [13]. Moreover, if the allele for the new morph were dominant, then it would tend to move forward due to “dominance drive” [25,34], although movement might be impeded by local barriers. Since the dominance relations in Euhadra are unknown, it is difficult to make predictions that are more specific, but in P. suturalis the dextral gene is recessive, so dominance drive could have acted against the establishment of a new dextral morph.
An effect of population density could aid the establishment of new morphs. In Heliconius butterflies, the interaction between frequency and density is often overlooked as a factor that could assist the establishment of new wing-pattern morphs [29]. At low density, purifying selection against new morphs is strong because predators remain naive, but when the population density is high, selection is weak over a broad range of intermediate frequencies [29]. In Euhadra, the parallel situation is that new morphs are less likely to find a mate at low density than at high density. New chiral morphs are therefore more likely to become established in populations at high density, because mates of the same chirality are more easily found.
Single-Gene Speciation in Sympatry or Mitochondrial DNA Introgression?
The phylogeny did not resolve the relationship between the six major groups within the sinistral clade, as it contains two exclusively dextral groups, three sinistral groups, and one mostly sinistral group that also includes dextral E. senck. aomoriensis (see Figures 3 and 4). Recently, Ueshima and Asami found a similar phylogeny using a different set of mitochondrial gene fragments [24]. One interpretation of these phylogenies is that the polyphyly of E. senck.
aomoriensis within the sinistral mitochondrial DNA E. quaesita clade is suggestive of recent gene flow between sinistral E. quaesita and dextral E. senck. aomoriensis. An alternative, as concluded by Ueshima and Asami, is that the polyphyletic E. senck.
aomoriensis lineages are in fact a different (i.e., new) species that has been derived from E. quaesita by single-gene speciation [24].
As in the original evolution of sinistral Euhadra that was discussed earlier, initially rare dextral morphs of E. quaesita could have been favoured in their spread by reproductive character displacement from another species. The most widespread sinistral species with which E. quaesita might have hybridized is E. decorata, which is found in the northern part of Tohoku. This explanation is reasonable because E. quaesita has a restricted distribution in northern Tohoku, whereas E. senck. aomoriensis and E. decorata are quite widespread there (see Figure 2). Thus, dextral E. senck. aomoriensis may have evolved from sinistral E. quaesita, because otherwise E. quaesita would tend to hybridize with sympatric, sinistral E. decorata. If so, this process could have been quite recent and in situ, because northern Tohoku is considered to have been unvegetated during the last glaciation (Figure 8).
Figure 8 Map of the Periglacial Region in Japan at the Height of the Last Glaciation
The last glaciation peaked about 20,000 years ago. Most of northern Tohoku must have been unvegetated (white shading), so Euhadra probably colonised from the south in the early Holocene. Brown shading shows regions above sea level, blue is sea, and the black outline indicates present-day Japan. The distribution of the periglacial region in earlier epochs is more uncertain. Data compiled from [56–58].
However, it is presently not possible to resolve whether hybridization or single-gene speciation best explains the mitochondrial DNA phylogeny, especially since the dominance of the chirality gene is unknown, and the morphological analysis complicates the interpretation (see Figure 5). We therefore used our model to compare the two competing explanations. We found that when there is even a slight degree of interchiral mating (incomplete assortment; α < 1), there is a single feasible equilibrium that is always stable. This equilibrium, which is analogous to a Hardy-Weinberg equilibrium while accounting for the maternal inheritance of the chirality gene, is one of substantial gene flow between the chiral morphs (see Figure 6).
When there is no interchiral mating (complete assortment; α = 1), two equilibria become possible. As before, in the first equilibrium there is substantial gene flow between morphs. This is interesting because it means that even when different chiral morphs are unable to mate (as in Euhadra), gene flow will still be substantial because of the delayed (maternal) inheritance. This observation was first pointed out by Johnson et al. [25]. In the second equilibrium, there is complete reproductive isolation so it amounts to single-gene speciation, but is only possible when sinistral snails are entirely sinistral homozygotes, and dextral snails are entirely dextral homozygotes. The latter equilibrium is unstable to the introduction of sinistral snails carrying dextral alleles or vice versa, and to a low frequency of interchiral mating. The empirical studies that have been carried out in Partula and also Achatinella tend to support the predictions from the model of high gene flow between morphs [38–40]. While assortative mating should be much greater in Euhadra because interchiral mating is not possible (α ≈ 1) [14], the prediction is that gene flow still should be substantial.
Another factor that could reduce gene flow between chiral morphs is selection. First, the more-common morph could be at an advantage because it spends less time searching for its own morph to mate with, and may well have more offspring as a consequence. Second, snails are probably less likely to transfer sperm successfully in interchiral matings. There is direct evidence for reduced fertility in Partula, since fewer young are produced by mixed pairs [13]. When we incorporated this mode of selection into the model, with random mating (inter- and intrachiral mating equally likely; α = 0) the population moves rapidly to equilibrium gene flow. Even with a complete lack of interchiral mating (α = 1), there is still substantial gene flow. The main difference is that the population moves towards equilibrium rather more slowly (see Figure 7).
Although there are no data for Euhadra, it is possible that a selection operates against hybrids between chiral morphs because genetic differences leading to postmating isolation are held in association (disequilibrium) with the chiral gene itself, preventing introgression. Evidence from the model suggests that such disequilibrium is unlikely to build up in sympatry. However, if two different morphs met after a period in allopatry, then the population could fall into an equilibrium in which gene flow is reduced, and so maintain the initial association. Even in this circumstance, the model shows that for all but very strong assortment (α > 0.99), there still should be high gene flow between morphs, so that heterozygotes would be abundant and selection would cause fixation of the more-common viability allele relatively quickly.
The results from the model have important implications for the interpretation of the mitochondrial DNA phylogeny, and can help to make suggestions for future research. Reproductive character displacement can explain how new chiral morphs arise and become the majority. However, even with reproductive character displacement there will still be a significant flow of the chirality gene (and other genes) between morphs. In these circumstances, speciation by chiral reversal is meaningful only in the context of a complex biogeographical process, which must usually involve other factors. Allopatry, perhaps combined with ecological selection, will be required for genetic differences and reproductive isolation to build up between chiral morphs. As this must have happened in Euhadra and several other species, it will be interesting to disentangle the relative contributions of each towards reproductive isolation.
The immediate question is whether sinistral E. quaesita reverted back to a dextral species, producing E. senck. aomoriensis, or whether there has been gene flow between E. quaesita and E. senck. aomoriensis. To resolve this issue, a priority should be to investigate the phylogeny of nuclear gene sequences and compare them with the mitochondrial DNA phylogeny. If dextral E. senck. aomoriensis arose by single-gene speciation, then most nuclear loci should place these specimens within the E. quaesita clade. Alternatively, if the polyphyletic position of dextral E. senck. aomoriensis in the mitochondrial DNA phylogeny is due to hybridization and introgression, then the genomes of the two populations should be a mosaic in their extent of differentiation [41–43]. However, even with a complete lack of interchiral mating, introgression and fixation of lineages will occur rapidly, so that introgression may be difficult to distinguish from single-gene speciation.
Another priority should be to establish the dominance relations of the chirality genes in Euhadra. This is because a prediction of the model and previous simulations [25,26,34] is that morphs are more likely to be established by single-gene speciation if the mutation involved is dominant. One means to establish dominance would be to attempt breeding experiments in the laboratory. However, there are no known dimorphic species, and interchiral mating has not been observed, so this may not be possible. A solution would be to discover a rare coiling variant, as has been done in Bradybaena similaris [24]. The final priority is to finely map the distributions of the different morphs, especially sinistral E. quaesita, sinistral E. decorata, and dextral E. senck. aomoriensis, to test whether the patterns are consistent with reproductive character displacement as an explanation. Individuals from zones of contact could be used in captive breeding experiments to determine the dominance relations, the proportions of each chirality allele, and the extent of gene flow (if any).
Wider Significance
The interpretation of the model could impact upon the understanding of the evolution of any gene with an indirect genetic effect, especially maternal-effect genes [44]. Maternal-effect genes are important in development because maternally produced molecules (e.g., mRNAs) in the egg affect early developmental processes such as axis formation, and often have later pleiotropic effects. Maternal-effect genes have been implicated in causing infertility [45], with a few also directly implicated in causing hybrid sterility [46], although generally it has been realized that their contribution to postzygotic isolation might be limited [47]. Finally, sex determination in many species involves interactions between maternal and zygote genes, so it has been proposed that conflicting selective pressures between maternally and zygotically expressed sex-determining loci can in turn have a role in shaping the evolution of sex determining systems [48]. However, the detailed discussion of the relevance of our model to these and other issues is beyond the scope of this paper.
Conclusions
The concept of single-gene speciation in snails has attracted attention because it is an exception to the accepted view that at least two pairs of interacting genes are required for reproductive isolation [21,22,24]. We have shown that reproductive isolation between chiral morphs is possible, but unstable, and will lead almost inevitably to gene flow. While chiral morphs may be a step towards distinct species, complete reproductive isolation must require other factors, such as divergent selection for habitat use or geographic separation of populations. Assuming that the differences in genitalia are not due to dominance or epistasis, chiral morphs with distinct genitalia should be considered “good” species (even if not biological species; [49]), because the morphological distinctness must have been maintained despite gene flow.
Materials and Methods
Samples
The 22 species of Euhadra (Bradybaenidae) are distributed throughout Japan and the neighbouring Korean island of Jeju [50]. We sampled all five sinistral species for morphological and molecular analysis, in addition to 14 of the 17 dextral species (Figure S1; Tables 1 and S1). The remaining dextral species, E. sadoensis,
E. nachicola, and E. sigeonis have restricted distributions. The distributions of four species are shown in Figure 2, using data compiled from [50,51] and the Web site http://www.biodic.go.jp/site_map/site_map.html (in Japanese). We were also able to use an extensive set of 16S rRNA sequences from E. peliomphala [52]) and E. quaesita [33]. For the outgroup, Nesiohelix
bipyramidalis was used because a prior analysis has suggested that it is suitable [31].
To enable a clear comparison between our results and those of Ueshima and Asami [24], we refer to subspecies where necessary, especially with regard to E. senck. Note that Ueshima and Asami refer to E. senck. aomoriensis as “E. aomoriensis” and E. murayamai as “E. quaesita murayamai” [24].
Morphological analysis
Morphological characters were measured in a subset of the species, primarily those from the “sinistral” clade, and then used to make a maximum parsimony tree. In the first analysis of genital shape and structure, the characters were classified as follows: (1) wings on the basal part of dart: 0, absent; 1, small; and 2, large (ordered). (2) Wings on the middle part of dart: 0, absent; 1, small; and 2, large (ordered). (3) Curve of wings on the front part: 0, absent; 1, present. (4) Shape of wings: 0, absent; 1, straight; and 2, curved (ordered). (5) Maximum number of wings: 0, absent; 1, two; and 2, three (ordered). (6) Shape of cross-section of dart: 0, round or oval; 2, triangular; and 1, crescent (ordered). (7) Flatness of the front section of the dart: 0, round; 1, flat; and 2, very flat (ordered). (8) Flatness of the middle section of dart: 0, round; 1, flat; and 2, very flat (ordered). (9) Flatness of the basal section of dart: 0, round; and 1, flat. (10) Shape of the front part of dart-sac: 0, round; 1, oval; and 2, long-oval (ordered). (11) Shape of the mid-dorsal part of dart-sac: 0, curved; and 1, straight. (12) Shape of the dart-sac near atrium: 0, round; and 1, pointed. (13) Length of the middle part relative to the front part of dart-sac: 0, short; and 1, long. (14) Size of the sheath at the basal part of dart-sac: 0, small; and 1, large. (15) Relative size of sub-dart sac: 0, large; 1, medium; and 2, small (ordered). (16) Shape of sub-dart sac: 0, round; and 1, flat. (17) Number of mucus glands: 0, > 10; 1, from 5 to 10; and 2, < 5 (ordered). (18) Region between sub-dart sac and dart sac: 0, not contracted; and 1, contracted.
In the second analysis, that of shell shape and structure, the characters were classified as follows (all sizes in millimetres): (1) size of shell (diameter × height): 0, < 50; 1, 50–60; and 2, > 60 (ordered). (2) Relative height (height/diameter): 0, < 0.5; 1, 0.5–0.6; 2, 0.6–0.7; and 3, > 0.7 (ordered). (3) Relative aperture size ([aperture height + aperture width]/diameter): 0, < 1.0; and 1, > 1.0 (ordered). (4) Uppermost band: 0, missing; and 1, present. (5) Peripheral band: 0, missing; and 1, present. (6) Background colour: 0, pale; and 1, dark. (7) Narrow flare-like colour pattern along the growth line: 0, missing; and 1, present. (8) Wide flare-like colour pattern along spiral line: 0, missing; and 1, present. (9) Growth line: 0, weak; 1, medium; and 2, strong (ordered). (10) Interval of growth line: 0, narrow; and 1, wide. (11) Fine spiral lines: 0, unclear; and 1, clear. (12) Gloss of surface: 0, absent; and 1, present. (13) Small dents irregularly located on the surface: 0, missing; and 1, present. (14) Umbilicus: 0, deeply open; and 1, shallow. (15) Umbilicus: 0, < 1/3 covered by basal lip; and 1, > 1/3 covered by basal lip. (16) Basal lip: 0, attached inside of umbilicus; and 1, attached outside of umbilicus. (17) Shoulder of body whorl: 0, weak; and 1, strong. (18) Peripheral angularity: 0, missing; and 1, present. (19) Inflation of upper part of body whorl: 0, weak; and 1, strong. Shell characters are relatively uniform within Euhadra populations, but where necessary, polymorphic characters were coded as an extra character (5).
DNA extraction and PCR amplification
Genomic DNA was isolated using previously described methods [53]. Primers for PCR amplification of an approximately 900-bp fragment of 16S rRNA are described in [31]. Variation in the mitochondrial 12S rRNA and cytochrome oxidase c subunit I genes, and nuclear ITS2 region was also investigated, using primers described in [31,53,54]. All PCR reactions used Takara rTaq (Takara Biomedicals, Tokyo, Japan) and buffers, with annealing temperatures of 50 °C. Cycle sequencing was carried out with both forward and reverse primers, using about 80–100 ng of PCR product in the reaction and the BigDye Terminator v3.0 Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, California, United States). DNA sequences were electrophoresed on a 310 Genetic Analyzer (Applied Biosystems).
Phylogenetic analyses
Sequences were aligned using the ClustalX software, and then checked manually. All insertion and deletion sites (indels), as well as several difficult-to-align regions, were removed before phylogenetic analysis. Phylogenetic relationships were analyzed using two methods: neighbour-joining and maximum likelihood, both with PAUP*4.0b8 [55]. Multiple hits were corrected using the general time reversible (GTR) model. The rate matrix, base frequencies, and shape parameter (α) of the gamma distribution (based on 16 rate categories) were estimated using likelihood, by iteration from an initial neighbour-joining tree. Parameters estimated from the initial tree were used to make a new neighbour-joining tree. The parameters were then re-estimated, and the process was repeated until there was no further improvement in likelihood. Bootstrap values were then calculated using 1,000 replicates. Maximum likelihood methods used a heuristic procedure with tree-bisection-reconnection, and 100 bootstrap replicates.
Assumptions
In the snails that have been investigated, asymmetry (or chirality) is under control of one or a few linked loci, where the phenotype of the offspring is controlled by the maternal genotype (maternal inheritance). Dominance is variable, even within genera: dextral is dominant in Lymnaea peregra, L. stagnalis,
B. similaris, and P. mirabilis [5,7,8,10,11,24], but sinistral is dominant in P. suturalis and Laciniaria [9,25]. Reversed asymmetry can sometimes be agenetic, when development is disrupted due to environmental extremes, usually heat or cold shock. We assumed that asymmetry is also maternally inherited in Euhadra. In Partula, mating is possible between different morphs (although it occurs at reduced frequency) because they are high-spired and mate by shell-mounting [19]. There is no such detailed information for Euhadra, but evidence suggests that interchiral mating is difficult and perhaps impossible, because they mate “face-to-face” [14].
Mathematical model
We developed a mathematical model to understand gene flow between chiral morphs, as well as the route by which new chiral morphs can become established. The methods are described in detail in Protocol S1.
Supporting Information
Protocol S1 The Mathematical Model
(89 KB DOC).
Click here for additional data file.
Figure S1 Sample Sites Used in This Study
The sample sites used in this study, highlighting in particular the sites where E. quaesita, E. murayamai and E. senckenbergiana were collected (underlined if used for DNA methods). m, E. murayamai; q, E. quaesita; sa, E. senck. aomoriensis; si, E. senck. ibukicola; sm, E. senck. minoensis; sn, E. senck. notoensis; ss E. senck. senck.
(662 KB TIF).
Click here for additional data file.
Table S1 Species Collected at Each Site
(24 KB DOC).
Click here for additional data file.
Table S2 Genital Characters
See Materials and Methods for definitions.
(24 KB DOC).
Click here for additional data file.
Table S3 Shell Characters
See Materials and Methods for definitions.
(27 KB DOC).
Click here for additional data file.
Table S4 The Five Classes of Snail and Their Frequencies
The formulas used in calculating frequencies were Q + P = 1, and d + s = 1.
(33 KB DOC).
Click here for additional data file.
Table S5 The Contribution of Each of the Four Phenotypic Mating Combinations to the Next Generation
Note that α describes the strength of assortative mating, β describes the mating advantage of dextral over sinistral snails (averaged over the sexes), and Δ describes the difference in mating advantage between the reciprocal crosses. Necessarily, α < 1−|Δ|/2.
(25 KB DOC).
Click here for additional data file.
We are grateful to Yuji Watanabe and Morito Hayashi for providing some of the 16S rRNA sequences. In addition, Jonathan Hobley, Osamu Takahashi, Jun Yokoyama, Norio Wakayama, Osamu Miura, and Lazaro Echenique Diaz helped collect snails. We thank Sara Goodacre, John Brookfield, Chris Wade, Adele Grindon, and Kester Jarvis for helpful comments; and Menno Schilthuizen, James Mallet, and three further anonymous referees for improvements to the manuscript. AD was supported by a Royal Society “2+2” fellowship. NB was supported by a Natural Environment Research Council postdoctoral fellowship.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. AD, SC, NHB, and BC conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, and wrote the paper.
Citation: Davison A, Chiba S, Barton NH, Clarke B (2005) Speciation and gene flow between snails of opposite chirality. PLoS Biol 3(9): e282.
Abbreviations
Ddextral allele
Ssinistral allele
==== Refs
References
Brown NA Wolpert L The development of handedness in left/right asymmetry Development 1990 109 1 9 2209459
McManus C Right hand left hand. The origins of asymmetry in brains, bodies, atoms and cultures 2002 London Weidenfeld and Nicolson 412
Nonaka S Shiratori H Saijoh Y Hamada H Determination of left-right patterning of the mouse embryo by artificial nodal flow Nature 2002 418 96 99 12097914
Bergmann DC Lee M Robertson B Tsou MFB Rose LS Embryonic handedness choice in C. elegans involves the G alpha protein GPA-16 Development 2003 130 5731 5740 14534142
Shibazaki Y Shimizu M Kuroda R Body handedness is directed by genetically determined cytoskeletal dynamics in the early embryo Curr Biol 2004 14 1462 1467 15324662
Hosoiri Y Harada Y Kuroda R Construction of a backcross progeny collection of dextral and sinistral individuals of a freshwater gastropod, Lymnaea stagnalis
Dev Genes Evol 2003 213 193 198 12690455
Boycott AE Diver C On the inheritance of sinistrality in Limnaea peregra
Proc R Soc Lond B Biol Sci 1923 95 207 213
Sturtevant AH Inheritance of direction of coiling in Limnaea
Science 1923 58 269 270 17837785
Degner E Der erbgang der inversion bei Laciniaria biplicata MTG (Gastr Pulm) Mitt Hamb Zool Mus Inst 1952 51 3 61
Murray J Clarke B Supergenes in polymorphic land snails 2. Partula suturalis
Heredity 1976 37 271 282 1068113
Freeman G Lundelius JW The developmental genetics of dextrality and sinistrality in the gastropod Lymnaea peregra
Roux Arch Dev Biol 1982 191 69 83
Clarke B Murray J Ecological genetics and speciation in land snails of the genus Partula
Biol J Linn Soc 1969 1 31 42
Johnson MS Polymorphism for direction of coil in Partula suturalis —Behavioral isolation and positive frequency-dependent selection Heredity 1982 49 145 151
Asami T Cowie RH Ohbayashi K Evolution of mirror images by sexually asymmetric mating behavior in hermaphroditic snails Am Nat 1998 152 225 236 18811387
Schilthuizen M Scott BJ Cabanban AS Craze PG Population structure and coil dimorphism in a tropical land snail Heredity 2005 In press
Murray J Clarke B The genus Partula on Moorea: Speciation in progress Proc R Soc Lond B Biol Sci 1980 211 83 117
Cain AJ Variation in spire index of some coiled gastropod shells and its evolutionary significance Philos Trans R Soc Lond B Biol Sci 1977 277 377 428 16299
Gittenberger E Sympatric speciation in snail—A largely neglected model Evolution 1988 42 826 828
Lipton CS Murray J Courtship of lands snails of the genus Partula
Malacologia 1979 19 129 146
Asami T Genetic variation and evolution of coiling chirality in snails Forma 1993 8 263 276
Dobzhansky T Genetics and the origin of species 1937 New York Columbia University Press 364
Muller HJ Isolating mechanisms, evolution and temperature Biol Symp 1942 6 71 125
Orr HA Is single-gene speciation possible? Evolution 1991 45 764 769
Ueshima R Asami T Single-gene speciation by left-right reversal—A land-snail species of polyphyletic origin results from chirality constraints on mating Nature 2003 425 679 679 14562091
Johnson MS Clarke B Murray J The coil polymorphism in Partula suturalis does not favor sympatric speciation Evolution 1990 44 459 464
van Batenburg FHD Gittenberger E Ease of fixation of a change in coiling: Computer experiments on chirality in snails Heredity 1996 76 278 286
Stone J Björklund M Delayed prezygotic isolating mechanisms: Evolution with a twist Proc R Soc Lond B Biol Sci 2002 269 861 865
Mallet J Singer MC Individual selection, kin selection, and the shifting balance in the evolution of warning colors—The evidence from butterflies Biol J Linn Soc Lond 1987 32 337 350
Mallet J Joron M Evolution of diversity in warning color and mimicry: Polymorphisms, shifting balance, and speciation Annu Rev Ecol Syst 1999 30 201 233
Thomaz D Guiller A Clarke B Extreme divergence of mitochondrial DNA within species of pulmonate land snails Proc R Soc Lond B Biol Sci 1996 263 363 368
Chiba S Accelerated evolution of land snails Mandarina in the oceanic Bonin Islands: Evidence from mitochondrial DNA sequences Evolution 1999 53 460 471
Davison A Land snails as a model to understand the role of history and selection in the origins of biodiversity Popul Ecol 2002 44 129 136
Watanabe Y Chiba S High within-population mitochondrial DNA variation due to microvicariance and population mixing in the land snail Euhadra quaesita (Pulmonata : Bradybaenidae) Mol Ecol 2001 10 2635 2645 11883878
Mallet J Hybrid zones of Heliconius butterflies in Panama and the stability and movement of warning color clines Heredity 1986 56 191 202
Peake J Species isolation in sympatric populations of the genus Diplommatina (Gastropoda, Prosobranchia, Cyclophoridae, Diplommatininae) Malacologia 1973 14 303 312
Uit de Weerd DR Groenenberg DSJ Schilthuizen M Gittenberger E Reproductive character displacement by inversion of coiling in clausiliid snails (Gastropoda, Pulmonata) Biol J Linn Soc Lond 2005 In press
Schilthuizen M Davison A The convoluted evolution of snail chirality Naturwissenschaften 2005 In press
Johnson MS Murray J Clarke B Independence of genetic subdivision and variation for coil in Partula suturalis
Heredity 1987 58 307 313
Goodacre SL Population structure, history and gene flow in a group of closely related land snails: Genetic variation in Partula from the Society Islands of the Pacific Mol Ecol 2002 11 55 68 11903904
Holland BS Hadfield MG Islands within an island: Phylogeography and conservation genetics of the endangered Hawaiian tree snail Achatinella mustelina
Mol Ecol 2002 11 365 375 11918776
Abernethy K The establishment of a hybrid zone between red and sika deer (Genus Cervus ) Mol Ecol 1994 3 551 562 7834107
Goodman SJ Barton NH Swanson G Abernethy K Pemberton JM Introgression through rare hybridization: A genetic study of a hybrid zone between red and sika deer (genus Cervus ) in Argyll, Scotland Genetics 1999 152 355 371 10224266
Wu CI The genic view of the process of speciation J Evol Biol 2001 14 851 865
Wolf JB Brodie ED Cheverud JM Moore AJ Wade MJ Evolutionary consequences of indirect genetic effects Trends Ecol Evol 1998 13 64 69 21238202
Boswell RE Prout ME Steichen JC Mutations in a newly identified Drosophila melanogaster gene, mago nashi, disrupt germ cell formation and result in the formation of mirror image symmetrical double abdomen embryos Development 1991 113 373 384 1765008
Beeman RW Friesen KS Properties and natural occurrence of maternal-effect selfish genes (“Medea” factors) in the red flour beetle Tribolium castaneum
Heredity 1999 82 529 534 10383673
Hurst GDD Schilthuizen M Selfish genetic elements and speciation Heredity 1998 80 2 8
Werren JH Hatcher MJ Godfray HCJ Maternal-offspring conflict leads to the evolution of dominant zygotic sex determination Heredity 2002 88 102 111 11932768
Mayr E 1963 Mayr E (1963) Animal species and evolution Harvard University Press
Azuma M [Coloured illustrations of the land snails of Japan] 1995 Osaka Hoikusya 359
Anonymous The national survey on the natural environment: Report on the distributional survey of Japanese animals (land and fresh water mollusca) 2002 Tokyo Biodiversity Center of Japan, Ministry of the Environment 1317
Hayashi M Chiba S Intraspecific diversity of mitochondrial DNA in the land snail Euhadra peliomphala (Bradybaenidae) Biol J Linn Soc Lond 2000 70 391 401
Teshima H Davison A Kuwahara Y Yokoyama J Chiba S The evolution of extreme shell shape variation in the land snail Ainohelix editha : a phylogeny and hybrid zone analysis Mol Ecol 2003 12 1869 1878 12803638
Folmer O Black M Hoeh W Lutz R Vrijenhoek R DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates Mol Mar Biol Biotech 1994 3 294 299
Swofford DL PAUP* 4.b10 1999 Sunderland, MA Sinauer Associates
Kaizuka S Geology and geomorphology of the Bonin Islands Bulletin of Ogasawara Research 1977 1 29 34
Yoshikawa T Kaizuka S Sakaguchi Y Sugimura S Ohta Y Geomorphology of Japan 1973 Tokyo University of Tokyo Press 415
Kaizuka S Chinzei K Nature of Japan: Mountains in Japan 1995 Tokyo Iwanami-Shoten 270
|
16149849
|
PMC1182688
|
CC BY
|
2021-01-05 08:21:25
|
no
|
PLoS Biol. 2005 Sep 9; 3(9):e282
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030282
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1607624110.1371/journal.pbio.0030285Research ArticleEvolutionGenetics/Genomics/Gene TherapyAnophelesGenomic Islands of Speciation in Anopheles gambiae
Speciation IslandsTurner Thomas L [email protected]
1
Hahn Matthew W
1
¤Nuzhdin Sergey V
1
1Center for Population Biology, University of California, Davis, California, United States of AmericaBarton Nick Academic EditorUniversity of EdinburghUnited Kingdom9 2005 9 8 2005 9 8 2005 3 9 e28518 1 2005 14 6 2005 Copyright: © 2005 Turner 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.
Islands in the Genome Promote Speciation
The African malaria mosquito, Anopheles gambiae sensu stricto (A. gambiae), provides a unique opportunity to study the evolution of reproductive isolation because it is divided into two sympatric, partially isolated subtaxa known as M form and S form. With the annotated genome of this species now available, high-throughput techniques can be applied to locate and characterize the genomic regions contributing to reproductive isolation. In order to quantify patterns of differentiation within A. gambiae, we hybridized population samples of genomic DNA from each form to Affymetrix GeneChip microarrays. We found that three regions, together encompassing less than 2.8 Mb, are the only locations where the M and S forms are significantly differentiated. Two of these regions are adjacent to centromeres, on Chromosomes 2L and X, and contain 50 and 12 predicted genes, respectively. Sequenced loci in these regions contain fixed differences between forms and no shared polymorphisms, while no fixed differences were found at nearby control loci. The third region, on Chromosome 2R, contains only five predicted genes; fixed differences in this region were also verified by direct sequencing. These “speciation islands” remain differentiated despite considerable gene flow, and are therefore expected to contain the genes responsible for reproductive isolation. Much effort has recently been applied to locating the genes and genetic changes responsible for reproductive isolation between species. Though much can be inferred about speciation by studying taxa that have diverged for millions of years, studying differentiation between taxa that are in the early stages of isolation will lead to a clearer view of the number and size of regions involved in the genetics of speciation. Despite appreciable levels of gene flow between the M and S forms of A. gambiae, we were able to isolate three small regions of differentiation where genes responsible for ecological and behavioral isolation are likely to be located. We expect reproductive isolation to be due to changes at a small number of loci, as these regions together contain only 67 predicted genes. Concentrating future mapping experiments on these regions should reveal the genes responsible for reproductive isolation between forms.
Using DNA microarrays, the authors identify 3 small regions of the genome that differ between two forms of hybridizing mosquitoes; regions that are likely to contain the genes responsible for reproductive isolation.
==== Body
Introduction
Uncovering the genetic basis for reproductive isolation is a key to understanding how biological diversity is generated. Many researchers have used quantitative trait locus (QTL) mapping experiments to find the number and size of regions involved in both pre- and post-mating isolation between species (e.g., [1–6]). Although QTL mapping experiments are a powerful method for mapping large regions of the genome responsible for isolation traits, large numbers of recombinant offspring or advanced genetic tools are needed to fine-map the genes underlying QTLs (e.g., [7–9]). By contrast, studies of genomic differentiation between naturally hybridizing taxa make it possible to take advantage of the many recombination events that occur between backcrossed hybrid individuals in order to map the regions responsible for reproductive isolation [10–14].
Anopheles gambiae sensu stricto (A. gambiae), is the type species of the Anopheles gambiae sensu lato complex: a group of seven closely related African species morphologically indistinguishable as adults [15] and incompletely reproductively isolated from one another (hybrid females are fertile) [15,16]. The observation that some species blood-feed exclusively on humans and breed in artificial environments (which have been available for less than 10,000 y) further suggests that this species complex is the result of recent radiation [17]. In addition to these seven recognized taxa, A. gambiae is further subdivided into two partially isolated taxa known as the M form and S form [18]. These forms were originally delineated based on several tightly linked single nucleotide polymorphisms (SNPs) in the rDNA of the X chromosome that are rarely found as heterozygotes [19]. Subsequent studies of reproductive isolation in nature revealed that these forms mate assortatively, with 98.8% of wild-caught gravid females (within an area of sympatry) having mated within their own form [20,21]. When forms are crossed in the lab, no intrinsic fitness reductions are found [22], suggesting that the observed heterozygotic deficiencies are due to nonrandom mating and/or ecologically dependent postzygotic isolation. Studies of gene flow using microsatellite markers have repeatedly found no appreciable genetic differentiation outside the centromeric end of the X chromosome (except between inversions that are not fixed between forms [23]), but these studies have only genotyped 10–25 loci [23–27].
To better examine the genetic basis for the maintenance of reproductive isolation between the M and S forms of A. gambiae and to delineate the number and size of regions that do not introgress—which may contain genes involved in reproductive isolation—we hybridized DNA of single mosquitoes from population samples of M and S forms to Affymetrix GeneChip microarrays. Recent studies in Saccharomyces cerevisiae [28] and Arabidopsis thaliana [29] have shown that hybridizing genomic DNA to Affymetrix arrays, which are printed with 25 base pair (bp) oligonucleotides (“probes”), allows precise mapping of DNA polymorphisms between samples. Because the hybridization intensity between probes on the array and DNA in the sample depends on sequence identity, polymorphisms located within these probes (either single nucleotide differences or small indels) can be quantified. Borevitz et al. [29] used this technique to genotype two Arabidopsis thaliana inbred strains and their recombinant inbred line, and were able to precisely delineate which regions of the recombinant line came from either parental line. Our goals were (1) to locate regions of the genome where M and S forms differed; (2) to test whether the observed pattern of differentiation resulted from selection against gene flow or could be explained by genetic drift or other processes; and (3) to determine what this genomic pattern could tell us about speciation and adaptive radiation.
Results
We used seven samples of M form and seven samples of S form mosquitoes from areas of Cameroon where they are sympatric (see Materials and Methods), and where gene flow at microsatellites is known to be high [26]. These samples were chosen to avoid the confounding factor of several segregating inversions found within A. gambiae [23,30]. In Cameroon both M and S forms possess the same (standard) karyotype [26].
We remapped each 25-bp probe on the Affymetrix Plasmodium/Anopheles GeneChip array to the most recent A. gambiae genome assembly and removed probes with multiple exact matches, which generated a marker map of 142,065 unique probes (see Materials and Methods). Whole genomic DNA from single female mosquitoes was hybridized to each array, with seven individuals hybridized per form (a total of 14 arrays). In addition, one mosquito DNA isolate was labeled and hybridized a second time as a technical replicate. As expected, all samples were very similar, with the technical replicates more highly correlated than any of the biological replicates, indicating high reproducibility (average Spearman M vs. M correlation = 0.954, average Spearman M vs. S correlation = 0.942, Spearman technical replicate correlation = 0.989).
Differentiation between forms is shown in Figure 1. To predict which probes contained polymorphisms between forms, we calculated t-test p-values for each probe and considered probes with p < 0.01 to be candidate single-feature polymorphisms (SFPs; [29]). We also directly sequenced several candidate SFPs to verify these differences (see below and Materials and Methods). The number of probes with differences between forms within a window of 300 probes was then tested against the number expected to appear if sequence differences were distributed randomly across each chromosome (via a χ2 test). The null hypothesis in this analysis, a random distribution of SFPs, could be violated simply because of linkage disequilibrium of probes within a gene. We permuted probesets—preserving the association of probes within a gene—to test for this effect (see Materials and Methods). After correcting for multiple tests, four regions were found to be significantly differentiated in the initial sliding window analysis: the region proximal to the centromere on the left arm of Chromosome 2 (2L), the centromeric end of the telocentric X chromosome, and two regions on the right arm of Chromosome 2 (2R). We also searched for differentiated regions using a hidden Markov model (HMM), which recovered the regions on Chromosomes 2L and X, and one of the regions on Chromosome 2R (see Materials and Methods). The 2L and X regions remained highly significant after permutation testing (2L, p = 0.002; X, p < 0.001), but the 2R regions were not significant. For the 2R region detected in both analyses, nonsignificance may be due simply to its small size in relation to the size of sliding windows: seven of 11 SFPs in this window fell within four probesets, spanning only 40 kilobases (kb). Overall, the significance of the differentiated regions on Chromosomes 2L and X is strongly supported by all analyses, and the small region on 2R is suggestive: these three regions are our candidate “speciation islands.” Using the HMM, we estimated the sizes of these regions to be 2,160 kb, 566 kb, and 37 kb for Chromosomes 2L, X, and 2R, respectively. We expect these values to be underestimates for the regions on 2L and X because some heterochromatic portions of the neighboring centromeres have not been assembled. The number of predicted genes in each chromosomal region is 50 in 2L, 12 in X, and five in 2R.
Figure 1 Differentiation between Forms
The significance threshold shown is p = 0.05, Bonferroni-corrected for the number of windows tested per chromosome. The centromeres of Chromosomes 2 and 3 are located between the right and left arms, i.e., between each pair of graphs; the centromere of Chromosome X is located at the right end of the graph. Grey areas are divergent regions identified by our HMM. The grey region at the tip of Chromosome X appears to lie outside of the final window because the chromosomal position given for each window is the location of the central probe in that window; the final window on Chromosome X spans a large region because of low gene density. Sequenced loci are shown with red triangles; overlapping triangles on Chromosomes 3R and 2R obscure multiple sequenced loci (see text for details). 3L, left arm of Chromosome 3; 3R, right arm of Chromosome 3.
We directly assayed sequence variation from the islands in a larger sample to verify the contrast between these regions and the rest of the genome. Divergent regions were compared both to sequences from nearby loci that did not fall within the island of differentiation and to additional control sequences on Chromosome 3R. Sequenced loci are indicated in Figure 1, and sample sizes and sequence statistics are listed for all loci in Table 1. Fully supporting inferences from the whole-genome analysis, sequences from differentiated regions on Chromosomes 2L and X contained fixed differences and no shared polymorphisms, while adjacent control loci contained shared polymorphisms and no fixed differences (Table 1). This disparity between fixed and shared differences is highly significant at both regions (Chromosome 2L: seven fixed, none shared within the differentiated region, none fixed, ten shared outside the differentiated region, Fisher's exact test, p = 0.00005; Chromosome X: five fixed, none shared within the differentiated region, none fixed, four shared outside the differentiated region, Fisher's exact test, p = 0.008). The intron of the P450–2 gene, in the divergent Chromosome X region, also had a 51-bp indel fixed between forms. The level of polymorphism within differentiated regions on Chromosomes 2L and X was low within each form (π ≤ 0.001 for all four loci), as would be expected if these regions had low rates of recombination because of proximity to centromeres. The nearby control region on X also had low polymorphism (Table 1), but showed no fixed differences.
Table 1 DNA Variation and Differentiation at Sequenced Loci
On Chromosome 2R we sequenced five loci: three loci inside the 37-kb region that was detected in both of our whole-genome analyses and two control loci adjacent to this region (one on either side; see Figure 1). Both control loci showed only shared polymorphisms and equal levels of nucleotide diversity between forms (Table 1). We found a single gene within the island that showed fixed differences and no shared polymorphisms between forms, similar to the loci on Chromosomes 2L and X. This gene (denoted UNK1) has no known function, and a BLASTn search yielded significant similarity only to an adjacent gene in A. gambiae that is also within the differentiated region but was not sequenced in our study. The two other genes within the island had no fixed differences, but all three loci had highly unequal levels of diversity between forms; S form mosquitoes showed up to 20 times the levels of nucleotide polymorphism as M form individuals. Gene UNK1 showed the greatest asymmetry in polymorphism, with 21 SNPs in the S form sample and two in the M form sample. There is no evidence from tests of selection [31] that a recent sweep has occurred (Table 1), leaving the reasons for this asymmetry unclear.
Additional control sequences from Chromosome 3R showed shared polymorphisms between forms and no fixed differences, despite low levels of variability at some loci (Table 1). This highlights one possible complication of our analysis: reduced variability within forms may lead to fixed differences between forms in the absence of selection against hybrids [32]. Areas of low recombination that are exposed to repeated linked selection can show differentiation between populations, even in the face of high levels of gene flow [32,33]. Though there is little information on genomic recombination rates in A. gambiae, it is likely, as in Drosophila, that areas adjacent to centromeres have reduced recombination; these regions have high numbers of DNA repeats and low gene density [34]. The lack of fixed differences in our control regions adjacent to centromeres and the fixed differences we found in the middle of Chromosome 2R both refute the idea that recombination alone is responsible for the observed pattern of differentiation. We further tested for the effect of reduced recombination in causing the observed pattern through coalescent simulations. We used the program MS [35] to generate coalescent genealogies under a Wright–Fisher symmetric island model of gene flow between two subpopulations (cf. [33]). Estimates of the migration rate between forms in Cameroon were taken from the study by Wondji et al. [26] (we used the most conservative estimate, 4Nem = 10). We simulated two types of loci: one with levels of nucleotide polymorphism and recombination typical of most of our control regions, and one with 10-fold reductions in effective population size and recombination rate typical of most of our differentiated regions (see Materials and Methods). After generating 10,000 coalescent genealogies we did not observe a single instance of fixed differences in the absence of shared polymorphisms between subpopulations for our simulated control loci. Conversely, for our simulated differentiated loci we observed 32 instances where this occurred (p = 0.0032). Conservatively considering the two sequenced genes within each of our Chromosome 2L and Chromosome X islands as single loci, the probability of sequencing loci from both regions and observing fixed differences without shared polymorphisms—even with a much reduced effective population size—is p < 0.0001. It is therefore unlikely that decreased variability alone is responsible for the observed differences between forms.
Discussion
Our genome-wide array analysis and our analysis of sequence polymorphism clearly show that differentiation between the M and S forms of A. gambiae is only present in a few regions of the genome. Although some of this similarity could be due to ancestral polymorphism, several lines of evidence support the hypothesis of substantial current gene flow between A. gambiae M form and A. gambiae S form. Previous studies have documented between-form mating of A. gambiae in nature (1.2% of scorable matings) [20,21], and hybrid M/S genotypes have been found (1.1% of larvae and 0.3% of adults in a population where M and S are sympatric) [36]. A previous survey using microsatellite loci of M and S forms in Cameroon found Fst values between forms that were consistent with substantial migration rates (calculating 4Nem from the average Fst reported in [26] yields 10 < 4Nem < 47). Microsatellite Fst values are currently being calculated for the mosquitoes used in our study, with comparable results (F. Tripet, unpublished data).
In contrast to the low levels of differentiation found throughout most of the genome, our array experiments revealed three small regions to be significantly differentiated between forms. Sequences from islands on Chromosomes 2L and X contain 13 fixed differences and no shared polymorphisms. One gene within the third island, a 37-kb region on Chromosome 2R, also shows only fixed differences between forms.
In the early stages of divergence, above-average differentiation is expected between regions of low recombination. We conducted coalescent simulations to test whether the observed differentiation on Chromosomes 2L and X could plausibly result from neutral scenarios. Rejection of this neutral hypothesis (p < 0.0001) suggests that differentiation in these regions is due to selection against hybrid genotypes during backcrossing. This conclusion supports the prediction that when gene flow is present, differentiation between incipient species can be limited to small regions surrounding isolating genes [37]. No intrinsic postzygotic isolation has been found between forms, so we expect that the genes in these speciation islands are responsible for the observed prezygotic isolation or cause postzygotic isolation in nature (e.g., through ecological maladaptation, sexual isolation of hybrids from both parent species, or other causes).
Recombination between genes responsible for assortative mating and postzygotic isolation is thought to be a major barrier to speciation in the presence of gene flow [38]. Recent work has shown that inversions may facilitate speciation by creating linkage disequilibrium between these genes [39–41]. Although no direct information on recombination rates in these islands is available, the low polymorphism, high repeat density [34], and low gene density [34] within them suggest that regions near the centromeres have reduced recombination. This reduced recombination may create linkage disequilibrium between isolating factors, although the speciation islands we detected are so small that there are few genes within them to link together. Further investigation is necessary to determine whether our speciation islands contain co-adapted gene complexes, or whether they contain single loci experiencing divergent selection.
In the A. gambiae sensu lato species complex, overlapping distributions of partially isolated taxa are the rule and not the exception. A. gambiae M form and A. gambiae S form are broadly sympatric, and they are also sympatric with their two closest sibling species, A. merus and A. arabiensis [42], which are only partially isolated from the A. gambiae forms [16]. Mapping genomic differentiation between other members of this species complex will inform the study of speciation and adaptive radiation by showing whether there are consistent patterns of genomic differentiation between these species, how these speciation islands may change in size or number with time, and what genes within them are responsible for reproductive isolation.
Materials and Methods
DNA labeling and microarray hybridization
Mosquitoes were collected in the towns of Buea (one M form, one S form), Mutanguene (one M form, two S forms), and Tiko (five M forms, four S forms), in Cameroon in 2003 by the lab of G. C. Lanzaro (Department of Entomology, University of California, Davis, California, United States), who graciously shared samples for this study. DNA was extracted following Post et al. [43], and standard PCR diagnostics were used to differentiate A. gambiae from A. arabiensis [44] and to differentiate A. gambiae M and S forms [18]. Polytene chromosome preparation and analysis were as described in Hunt et al. [45]. Extracted DNA was labeled with an Invitrogen Bioprime Labling Kit (Invitrogen, Carlsbad, California, United States), following Borevitz et al. [29]. Isolated DNA in water (1,200 ng) was added on ice to 120 μl of random primers to a total volume of 200 μl. This solution was denatured in a 95 °C water bath, cooled on ice, then added to Klenow polymerase (6 μl), and dNTPs (30 μl). Incubation at 25 °C overnight resulted in small biotinylated oligos of approximately 50 bp. DNA was precipitated by adding 20 μl of 3 M NaOH and 400 μl of cold 100% EtOH. This solution was frozen at −80 °C for 15 min and centrifuged at 15,000g for 10 min, and the pelletized DNA was dried and resuspended in 50 μl of ddI H2O. Microarrays were hybridized by the University of California at Davis School of Medicine Microarray Core Facility using genomic DNA in place of cDNA in the standard Affymetrix protocol (Affymetrix, Santa Clara, California, United States).
Microarray analysis
Raw hybridization intensities were normalized in R via RMA [46] using the Bioconductor Affy package [47] (http://www.bioconductor.org). Each probe on the array was blasted against the most recent A. gambiae genome assembly in order to remove probes with more than one exact match and to remap all probes onto the current assembly. We considered probes with two tailed t-test p-values less than 0.01 to contain SNPs between forms (1,577 total probes; this expectation was based on previous studies that validated the method [28,29], and was verified for our samples by sequence data discussed below). To test for regional differentiation, the chromosomal positions of probes with p < 0.01 were considered. To detect regional clustering of these probes, a sliding window of 300 probes was moved 30 probes at each step, and a χ2 test was performed to test whether the number of significant probes in each window was more than that expected by chance. Four regions were significant in this analysis at p < 0.05 after Bonferroni correction for the number of windows tested per chromosome. Three of these regions were robust to varying window sizes (data not shown). These analyses were carried out on each chromosome individually; the average window size was approximately 500 kb, with a range of approximately 60–2,000 kb, depending on the density of probes in each region. Significant probes could be more clustered than random because of the shared history of linked probes, so the sliding window analysis was repeated on 1,000 permuted datasets to test the effect of short-range linkage disequilibrium on our conclusions. For each permutation, probesets were reshuffled, but probes within a probeset remained associated. The 300-probe window with the highest number of significant probes in each permutation was recorded, and significance was assigned based on the number of permutations containing a window with differentiation equal to the observed value. As an additional test, we constructed a HMM [48] to segment each chromosome into introgressed and differentiated regions. Transition and emission probabilities of the HMM were estimated by expectation maximization; hidden states were then inferred using the Viterbi algorithm in Matlab (The MathWorks, Natick, Massachusetts, United States). The three divergent regions together are estimated to be approximately 2,740 kb, which is 1.2% of the assembled genome. All significant overlapping probes were combined in both analyses to control for nonindependence (two overlapping probes with p < 0.01 counted as one observation). Nonindependence could also arise because of deletions covering whole probesets; deletions were characterized by low p-values throughout a probeset, and were collapsed into one observation for the analysis. The data shown in Figure 1 were corrected for both overlapping probes and deletions in candidate regions. Excel files of normalized data for each chromosome are available from the authors upon request.
Sequencing
All products were sequenced in both directions; see Table 1 for sequence lengths and sample sizes for each sequence. All individuals used in the whole-genome analysis were included, and sample sizes were increased to include additional samples collected in the same locations at the same time. Sequence traces were assembled and edited in CodonCode (http://www.codoncode.com, which uses ABI quality scores and Phred/Phrap to call bases, find heterozygous SNPs, and correct for heterozygous indels. Analysis of polymorphism and divergence was done in DNAsp (http://www.ub.es/dnasp/). The regions we sequenced covered nine probes with uncorrected p-values less than 0.01. The expected nucleotide difference was found in seven probes (78%), and one of the remaining two probes overlapped with a probe that contained the expected difference, highlighting the nonindependence of overlapping probes (which we corrected for in our whole-genome analysis). Because our samples were potentially heterozygous, we expected probes with low p-values to be either fixed differences or nucleotides with highly differentiated frequencies between forms. Sequencing of these six probes verified this expectation, as the detected nucleotide difference was either fixed between forms, or nearly fixed with either one or two heterozygous individuals for the rare allele.
Coalescent simulations
Using the program MS [35], we generated coalescent genealogies with samples drawn from two subpopulations exchanging migrants. As with the structure of most of our sequenced loci, we simulated 26 alleles sampled from one population and 24 alleles sampled from the other. We conditioned the simulation of our control loci on the number of segregating sites observed at the Chromosome 2L control locus (S = 27) and the differentiated loci on the 2L loci within the island (S = 12). We estimated migration to be 4Nem = 10 for the control loci and 4Nem = 1 for the differentiated loci [26]. Because recombination rates are not known, we used the typical value of r = 1 × 10−8 for recombination per site in control regions (4Ner = 2; r = 1 × 10–8; Ne = 100,000; 500 bases per locus) and r = 1 × 10–9 for recombination per site in differentiated regions (4Ner = 0.04; r = 1 × 10−9; Ne = 10,000; 1,000 bases per locus). All simulations were run 10,000 times, where the output of MS was parsed to count the number of fixed and shared polymorphisms for each run. To obtain the p-value associated with observing two loci with fixed differences and no shared polymorphisms, we sampled two loci for each of the 10,000 iterations and counted the number of times both showed this pattern.
Supporting Information
Table S1 Additional Information on Sequenced Loci
(23 KB XLS).
Click here for additional data file.
Accession Numbers
The GenBank (http://www.ncbi.nlm.nih.gov/Genbank) accession numbers for the gene sequences discussed in this paper areAY825543–AY825922 and DQ080663–DQ080909.
We are grateful to G. C. Lanzaro and F. Tripet for providing DNA from karyotyped mosquitoes for this work, and to D. Begun, A. Kern, L. Moyle, F. Tripet, R. Glor, M. Slotman, W. Stephan, M. Turelli, and many others for advice and support. This work was supported by the Center for Population Biology at the University of California, Davis (TLT), by a National Science Foundation (NSF) Interdisciplinary Informatics postdoctoral fellowship (MWH), by National Institutes of Health grant RO1 GM61773–01, and by NSF grant DEB-0316513 (SVN).
Competing Interests. The authors have declared that no competing interests exist.
Author Contributions. TLT, MWH, and SVN conceived and designed the experiments. TLT performed the experiments. TLT and MWH analyzed the data. TLT wrote the paper with editorial suggestions from MWH and SVN.
¤ Current address: Department of Biology and School of Informatics, Indiana University, Bloomington, Indiana, United States of America
Citation: Turner TL, Hahn MW, Nuzhdin SV (2005) Genomic islands of speciation in Anopheles gambiae. PLoS Biol 3(9): e285.
Abbreviations
2Lleft arm of Chromosome 2
2Rright arm of Chromosome 2
A. gambiae
Anopheles Gambiae sensu strictu
bpbase pair(s)
HMMhidden Markov model
kbkilobase(s)
SFPsingle-feature polymorphism
SNPsingle nucleotide polymorphism
==== Refs
References
Coyne JA Crittenden AP Mah K Genetics of a pheromonal difference contributing to reproductive isolation in Drosophila
Science 1994 265 1461 1464 8073292
Bradshaw HD Wilbert SM Otto KG Schemske DW Genetic mapping of floral traits associated with reproductive isolation in Monkeyflowers (Mimulus)
Nature 1995 376 762 765
Jones CD The genetics of adaptation in Drosophila sechellia
Genetica 2005 123 137 145 15881686
Fishman L Kelly AJ Willis JH Minor quantitative trait loci underlie floral traits associated with mating system divergence in Mimulus
Evolution 2002 56 2138 2155 12487345
Moyle LC Graham EB Genetics of hybrid incompatibility between Lycopersicon esculentum and L. hirsutum
Genetics 2005 169 355 373 15466436
Ortiz-Barrientos D Counterman BA Noor MAF The genetics of speciation by reinforcement PLoS Biol 2004 2 e416 10.1371/journal.pbio.0020416 15550988
Ting CT Tsaur SC Wu ML Wu CI A rapidly evolving homeobox at the site of a hybrid sterility gene Science 1998 282 1501 1504 9822383
Barbash DA Siino DF Tarone AM Roote J A rapidly evolving MYB-related protein causes species isolation in Drosophila
Proc Natl Acad Sci U S A 2003 100 5302 5307 12695567
Presgraves DC Balagopalan L Abmayr SM Orr HA Adaptive evolution drives divergence of a hybrid inviability gene between two species of Drosophila
Nature 2003 423 715 719 12802326
Barton NH Gale KS Harrison RG Genetic analysis of hybrid zones Hybrid zones and the evolutionary process 1993 New York Oxford University Press 13 45
Payseur BA Krenz JG Nachman MW Differential patterns of introgression across the X chromosome in a hybrid zone between two species of house mice Evolution Int J Org Evolution 2004 58 2064 2078
Emelianov I Marec F Mallet J Genomic evidence for divergence with gene flow in host races of the larch budmoth Proc R Soc Lond B Biol Sci 2004 271 97 105
Machado CA Kliman RM Markert JA Hey J Inferring the history of speciation from multilocus DNA sequence data: The case of Drosophila
pseudoobscura and close relatives Mol Biol Evol 2002 19 472 488 11919289
Wilding CS Butlin RK Grahame J Differential gene exchange between parapatric morphs of Littorina saxitalis detected using AFLP markers J Evol Biol 2001 14 611 619
White G
Anopheles gambiae complex and disease transmission in Africa Trans R Soc Trop Med Hyg 1974 68 278 301 4420769
Hunt RH Coetzee M Fettene M The Anopheles gambiae complex: A new species from Ethiopia Trans R Soc Trop Med Hyg 1998 92 231 235 9764342
Coluzzi M Sabatini A della Torre A Di Deco MA Petrarca V A polytene chromosome analysis of the Anopheles gambiae species complex Science 2002 298 1415 1418 12364623
Gentile G Slotman M Ketmaier V Powell JR Caccone A Attempts to molecularly distinguish cryptic taxa in Anopheles gambiae s.s Insect Mol Biol 2001 10 25 32 11240634
Favia G Lanfrancotti A Spanos L Siden-Kiamos I Louis C Molecular characterization of ribosomal DNA polymorphisms discriminating among chromosomal forms of Anopheles gambiae s.s Insect Mol Biol 2001 10 19 23 11240633
Tripet F Toure YT Taylor CE Norris DE Dolo G DNA analysis of transferred sperm reveals significant levels of gene flow between molecular forms of Anopheles gambiae
Mol Ecol 2001 10 1725 1732 11472539
Tripet F Toure YT Dolo G Lanzaro GC Frequency of multiple inseminations in field-collected Anopheles gambiae females revealed by DNA analysis of transferred sperm Am J Trop Med Hyg 2003 68 1 5
Di Deco MA Petrarca V Villani F Coluzzi M Polimorfismo cromosomico da inversioni paracentiche ed eccesso delgi eterocariotipi in ceppi di Anopheles allevati in laboratorio Parassitologia 1980 22 304 306
Lanzaro GC Toure YT Carnahan J Zheng LB Dolo G Complexities in the genetic structure of Anopheles gambiae populations in West Africa as revealed by microsatellite DNA analysis Proc Natl Acad Sci U S A 1998 95 14260 14265 9826688
Wang R Zheng LB Toure YT Dandekar T Kafatos FC When genetic distance matters: Measuring genetic differentiation at microsatellite loci in whole-genome scans of recent and incipient mosquito species Proc Natl Acad Sci U S A 2001 98 10769 10774 11553812
Stump AD Schoener JA Constantini C Sagnon NF Besansky NJ Sex-linked differentiation between incipient species of Anopheles gambiae
Genetics 2005 169 1509 1519 15654109
Wondji C Simard F Fontenille D Evidence for genetic differentiation between the molecular forms M and S within the Forest chromosomal form of Anopheles gambiae in an area of sympatry Insect Mol Biol 2002 11 11 19 11841498
Lehmann T Licht M Elissa N Maega TA Chimumbwa JM Population structure of Anopheles gambiae in Africa J Hered 2003 94 133 147 12721225
Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DR Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 163 79 89 12586698
Borevitz J Liang D Plouffe D Chang HS Zhu T Large scale identification of single-feature polymorphisms in complex genomes Genome Res 2003 13 513 523 12618383
Touré YT Petrarca V Traoré SF Coulibaly A Maiga HM The distribution and inversion polymorphism of chromosomally recognized taxa of the Anopheles gambiae complex in Mali, West Africa Parassitologia 1998 40 477 511 10645562
Tajima F Statistical-method for testing the neutral mutation hypothesis by DNA polymorphism Genetics 1989 123 585 595 2513255
Charlesworth B Measures of divergence between populations and the effect of forces that reduce variability Mol Biol Evol 1998 15 538 543 9580982
Stephan W Xing L Kirby DA Braverman JM A test of the background selection hypothesis based on nucleotide data from Drosophila ananassae
Proc Natl Acad Sci U S A 1998 95 5649 5654 9576938
Holt RA Subramanian GM Halpern A Sutton GG Charlab R The genome sequence of the malaria mosquito Anopheles gambiae
Science 2002 298 129 149 12364791
Hudson RR Generating samples under a Wright-Fisher neutral model of genetic variation Bioinformatics 2002 18 337 338 11847089
Gentile G della Torre A Maegga B Powell JR Caccone A Genetic differentiation in the African malaria vector, Anopheles gambiae s.s., and the problem of taxonomic status Genetics 2002 161 1561 1578 12196401
Wu CI Ting CT Genes and speciation Nat Rev Genet 2004 5 114 122 14735122
Felsenstein J Skepticism towards Santa Rosalia, or why are there so few kinds of animals Evolution Int J Org Evolution 1981 35 124 138
Noor MAF Grams KL Bertucci LA Reiland J Chromosomal inversions and the reproductive isolation of species Proc Natl Acad Sci U S A 2001 98 12084 12088 11593019
Rieseberg LH Chromosomal rearrangements and speciation Trends Ecol Evol 2001 16 351 358 11403867
Navarro A Barton NH Accumulating postzygotic isolation genes in parapatry: A new twist on chromosomal speciation Evolution Int J Org Evolution 2003 57 447 459
Coetzee M Craig M le Sueur D Distribution of African malaria mosquitoes belonging to the Anopheles gambiae complex Parasitol Today 2000 16 74 77 10652493
Post RJ Flook PK Millest AL Methods for the preservation of insects for DNA studies Biochem Syst Ecol 1993 21 85 92
Scott JA Brogdon WG Collins FH Identification of single specimens of the Anopheles gambiae complex by the polymerase chain reaction Am J Trop Med Hyg 1993 49 520 529 8214283
Hunt RH A cytological technique for the study of Anopheles gambiae complex Parassitologia 1973 15 137 139 4788355
Wu Z Irizarry RA Gentleman R Murillo FM Spencer F A model-based background adjustment for oligonucleotide expression arrays Johns Hopkins University Department of Biostatistics Working Papers 2004 1 1 26
Gautier L Cope L Bolstad BM Irizarry RA affy—Analysis of Affymetrix GeneChip data at the probe level Bioinformatics 2004 20 307 315 14960456
Durbin R Eddy S Krogh A Mitchison G Biological sequence analysis: Probabilistic models of proteins and nucleic acids 1998 Cambridge Cambridge University Press 368
|
16076241
|
PMC1182689
|
CC BY
|
2021-01-05 08:21:25
|
no
|
PLoS Biol. 2005 Sep 9; 3(9):e285
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030285
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1607624010.1371/journal.pbio.0030298Research ArticleAnimal BehaviorEvolutionInfectious DiseasesEpidemiology/Public HealthAnophelesPlasmodiumMalaria Infection Increases Attractiveness of Humans to Mosquitoes Malaria Gametocytes Attract MosquitoesLacroix Renaud
1
Mukabana Wolfgang R
2
Gouagna Louis Clement
3
¤Koella Jacob C [email protected]
1
1Laboratoire de Parasitologie Evolutive, Université P. and M. Curie, Paris, France,2Department of Zoology, University of Nairobi, Nairobi, Kenya,3Mbita Point Research and Training Centre, International Centre of Insect Physiology and Ecology, Mbita Point, KenyaSmith Tom Academic EditorSwiss Tropical InstituteSwitzerland9 2005 9 8 2005 9 8 2005 3 9 e29810 3 2005 24 6 2005 Copyright: © 2005 Lacroix 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.
To Mosquitoes, People with Malaria Smell Like Dinner
Do malaria parasites enhance the attractiveness of humans to the parasite's vector? As such manipulation would have important implications for the epidemiology of the disease, the question has been debated for many years. To investigate the issue in a semi-natural situation, we assayed the attractiveness of 12 groups of three western Kenyan children to the main African malaria vector, the mosquito Anopheles gambiae. In each group, one child was uninfected, one was naturally infected with the asexual (non-infective) stage of Plasmodium falciparum, and one harboured the parasite's gametocytes (the stage transmissible to mosquitoes). The children harbouring gametocytes attracted about twice as many mosquitoes as the two other classes of children. In a second assay of the same children, when the parasites had been cleared with anti-malarial treatment, the attractiveness was similar between the three classes of children. In particular, the children who had previously harboured gametocytes, but had now cleared the parasite, were not more attractive than other children. This ruled out the possibility of a bias due to differential intrinsic attractiveness of the children to mosquitoes and strongly suggests that gametocytes increase the attractiveness of the children.
A study of children in Kenya reveals that those individuals harboring the transmissible gametocyte stage of malaria parasites are more attractive to the mosquito vectors that spread this disease.
==== Body
Introduction
Malaria remains one of the most important causes of human mortality [1]. This epidemiological success of the parasite is at least partly due to its very high intensity of transmission, leading, on average, to several hundred infections per year in humans living in areas with the most intense transmission [2]. It became clear with the earliest models of the epidemiology of malaria that this intense transmission is largely due to the part of the parasite's life-cycle that takes place within the mosquito vector [3], and in particular to the interaction between the life-span of the mosquito, the duration of the parasite's development, and the mosquito's biting rate. More recently, ideas have moved from the epidemiological description of these interactions to the evolutionary idea that the parasite might manipulate the mosquito's behaviour to enhance its transmission.
Thus, malaria parasites manipulate various aspects of their mosquito vector's biting behaviour in ways that should increase their transmission success [4]. When the parasites have completed their development within the mosquito and have thus developed into the sporozoite stage, which can be transmitted by a bite on a human host, they increase the biting frequency, and thus the rate of contact between mosquitoes and humans and, consequently, the rate of transmission [5]. In contrast, at an earlier developmental stage (i.e., as oocysts, the non-transmissible developmental stage of the parasite), they decrease the biting rate by decreasing the mosquito's motivation to bite [6,7]. Thus, as biting is risky, the probability that the mosquito survives the parasite's development to the transmissible sporozoite stage is increased during early pre-sporozoite development.
During their development within the human, malaria parasites should be expected to manipulate the mosquito's biting behaviour by making infectious humans (i.e., those harbouring gametocytes) more attractive to mosquitoes. This would not only increase the parasite's transmission, but would also have a strong effect on the epidemiology of malaria [8]. However, while increased feeding on infected hosts has been observed in some studies [9,10], others do not show increased attractiveness [11]; the question of behavioural manipulation by gametocytes to increase the host's attractiveness to mosquitoes therefore remains a controversial issue in the biology of malaria.
One of the reasons for the lack of conclusive studies is that individual people vary considerably in their intrinsic attractiveness to mosquitoes [12,13]. Attractiveness depends on, for example, human sweat components [14], body temperature and moisture [15,16], or body odour and breath [13,17]. This background variation makes it difficult to find an effect of manipulation and, more importantly, to be certain that any observed difference is due to infection rather than to a difference in intrinsic attractiveness. Indeed, without a proper control, any observation of increased attractiveness by people infected with gametocytes might mean that harbouring gametocytes is a consequence of being very attractive to mosquitoes (and thus being infected frequently) rather than vice versa. However, the variability of intrinsic attractiveness to mosquitoes could be accounted for with a comparison of two measures of the attractiveness of individuals: the first when they are infected and the second when they have cleared the parasite.
We used this approach in our study, in which we estimated, in semi-natural conditions, the extent to which malaria gametocytes enhance the attractiveness of asymptomatic humans to mosquitoes in the area around Mbita, Nyanza Province, Kenya between April and July 2004 (with the agreement of the Kenya National Ethical Review Committee). We considered 12 groups of three children who were between 3 and 15 y old. This age-class constitutes the main reservoir of gametocytes in the area of Mbita [18]. After we had obtained informed consent by the parents or guardians, apparently healthy children from local primary schools were screened for malaria in the morning. Each day, we chose three children to participate in the study: one who was uninfected, one infected with the asexual (non-infective) stage of the malaria parasite Plasmodium falciparum, and one harbouring the parasite's gametocytes (the stage infective to mosquitoes). Note that symptomatic children (i.e., with increased temperature or other symptoms of disease) were immediately referred to the health clinic for proper treatment and therefore did not participate in the study. The attractiveness of the children in each group was measured with a three-way olfactometer consisting of a central chamber attached with PVC tubes to three tents where the three children of the group were resting or sleeping [12] (Figure 1). Mosquitoes were released in the central chamber and given the opportunity to follow their preferred odour. For further details of the set-up, see Materials and Methods.
Figure 1 Experimental Set-Up
(A)–(C) Top (A), cross-sectional (B), and three-dimensional (C) views of the olfactometer used. The fan (a) draws air (∼130 l/min per tent) from the three tents (b) to the outside environment via PVC pipes (c), trap chambers (d), and a central chamber (e). Each trap chamber contains a collecting cage (f) into which an exit trap opens (g). The fan pipe and release cup (h) are fitted to the top and bottom of the central chamber, respectively. Diagrams are not shown to scale; all dimensions are in centimetres. Source: [12].
To account for the variability of intrinsic attractiveness to mosquitoes, we treated all children with detectable parasitaemia with Fansidar and assayed the same children 2 wk after their first assay. Although it is not clear whether Fansidar kills gametocytes, subsequent microscopic examinations confirmed, in each of the 12 groups, that the children had cleared the parasite before the second assay. Thus, the first assay showed the combined effects of intrinsic and parasite-induced attractiveness, while the second assay gave an indication of the intrinsic variation of attractiveness among the children; the difference between the two assays could be used to analyse the effect of infection above the background variation.
Results/Discussion
The total number of mosquitoes attracted to the set of children in a group was similar before (20.6 mosquitoes per triplet) and after (18.6 per triplet) treatment with Fansidar (Figure 2A). Before treatment, mosquitoes preferred the gametocyte carriers, so that, on average, 10.2 mosquitoes were attracted by the gametocyte carriers, 5.3 by uninfected children, and 5.4 by children harbouring asexual stages. That the strong attraction of the gametocyte carriers was indeed due to the presence of the gametocytes is suggested by the fact that, after treatment, the children who had previously harboured gametocytes attracted a similar number of mosquitoes (4.4) to both the children who had previously harboured asexual stages (5.8) and the previously uninfected children (8.1). This pattern was confirmed with a statistical analysis that considered the difference between the proportions of the mosquitoes (relative to the 100 or so that were added to the olfactometer) attracted to gametocyte carriers before and after treatment. Indeed, after treatment, mosquitoes were approximately 14% less likely to be attracted to the former gametocyte carriers than before treatment (t-test: t = 2.23, df = 11, p = 0.048; Wilcoxon signed-rank test: p = 0.042).
Figure 2 Graphical Representation of Results
(A) Number of mosquitoes attracted to each class of children. Points show means of 12 groups; vertical lines show standard errors of the means. Solid diamonds show data of children before treatment; open diamonds denote children after treatment.
(B) Proportion of the responsive mosquitoes (i.e., the ones that were attracted to any of the children within a group) attracted to the children who harboured gametocytes (before treatment; dark bars) and to the children who had cleared their gametocytes (after treatment; light bars). The horizontal line shows the proportion expected if the mosquitoes showed no preference.
As a considerable number of mosquitoes did not respond to the odours of the three children (which is often the case when mosquitoes are used in experimental set-ups), we also considered only the mosquitoes that responded to the children's odours and compared the proportion of the mosquitoes within a replicate attracted to the gametocyte carrier before treatment and after treatment (Figure 2B). The figure shows that, in most replicates, the gametocyte carrier attracted more than a third of the mosquitoes before he or she had been treated, but attracted a smaller proportion of the mosquitoes after treatment. This result suggests, again, that the gametocytes enhance the attractiveness of their host. This visual analysis was confirmed by a statistical one, which showed that the difference between the proportion of mosquitoes attracted to the gametocyte carrier before and after the treatment is greater than 0 (df = 11, p = 0.039). A Wilcoxon signed-rank test gave a similar result with p = 0.052. (Note that the statistically significant finding, despite a small number of groups, emphasises the role of the gametocytes in determining attractiveness to mosquitoes above the intrinsic variation of attractiveness among individuals.)
Our analysis shows that increased attractiveness was not due to an intrinsic attractiveness of gametocyte carriers but to the infection status associated with the presence of gametocytes. The mechanism underlying this manipulation is unknown, but it is likely that the parasites change the infected individual's breath or body odour, as these are involved in attracting mosquitoes at the distances involved in our experiment [15,17]. While transpiration and body temperature also attract mosquitoes at these distances [15], these factors are less likely to be involved in the manipulation as the infection was asymptomatic in all of the children involved in our study.
A striking aspect of our results is that former gametocyte carriers (i.e., after treatment) seem to repel mosquitoes, as only 22%, i.e., less than one third, of the responding mosquitoes prefer these children (t-test: t = −2.203, df = 11, p = 0.050; Wilcoxon signed-rank test: p = 0.040). An explanation for this result could be based on the slight anaemia in previously infected children. Mosquitoes might sense this anaemia and prefer those children with a higher concentration of red blood cells as it is these that the mosquitoes require. This would indicate a remarkable adaptation by the mosquitoes. The interpretation, however, would predict that the children previously infected with the asexual stage should also repel mosquitoes—but this was not observed. Indeed, as 47% of the mosquitoes preferred these children, there was a tendency for the opposite effect (t-test: t = 1.65, df = 11, p = 0.127; Wilcoxon signed-rank test: p = 0.233).
In conclusion, our data suggest that mosquitoes are more attracted to humans infected by the transmissible gametocyte stage of malaria parasites than to uninfected individuals or individuals infected with asexual, non-transmissible stages. Previous studies have shown that malaria also manipulates the mosquito's biting behaviour at the oocyst stage (when infection decreases the motivation to bite and thereby increases the probability that the mosquito survives the parasite's development) [6,7] and at the sporozoite stage (when infection increases both the motivation to bite and the biting frequency) [5,7]. Thus this study completes the picture by demonstrating that the gametocyte, the only transmission stage from the human to the mosquito, manipulates the biting behaviour of its vector to enhance its transmission to the vector.
Apart from the diversity of mechanisms employed by the parasite to manipulate its vector and increase its transmission, what is striking about the set of demonstrations of behavioural manipulation is that the parasite appears to influence the parameters that are most critical for transmission and thus for the parasite's fitness [3]: the biting rate of the mosquitoes when they become infected and when they infect humans and, consequently, the mosquito's mortality during the parasite's development. Such manipulation therefore has a profound effect on the epidemiology of disease and, if it is not considered, can lead to severe biases in our estimates of the intensity of malaria transmission.
Materials and Methods
Screening for malaria was carried out with thick blood smears, stained with 10% Giemsa and examined microscopically with (100x) oil-immersion lens for the presence of sexual and asexual parasites. As we did not use molecular techniques, infections with low densities of parasites may have been missed. Thus our results compare, strictly speaking, individuals with no parasites or a low level of parasitaemia with individuals with more intense infections, rather than comparing uninfected with infected individuals. In addition, microscopy and Giemsa staining detect only those gametocytes circulating in the blood, so that we missed the sequestered gametocytes. However, as mosquitoes do not pick up sequestered gametocytes, missing these parasites in an assay is indeed desired.
Attractiveness of the children in each group to mosquitoes was measured with a three-way olfactometer consisting of a central chamber attached to three tents [12] (see Figure 1). At sunset, the volunteer children entered the tents to rest or sleep. The infection status of the children was changed randomly among tents among the 12 replicates, so that any residual odour in the olfactometer could not bias the results (but rather, at most, increase random variation). The tents were connected with PVC tubes to a central cage, where a fan was placed to draw air from the tents. About 100 mosquitoes were released into the central cage, and were given 30 min to follow their preferred odour into the entry traps placed between the PVC tubes and tents. The mosquitoes were uninfected females from a colony of Anopheles gambiae s.s. that had been established from specimens collected in Njage village, Kilobero District, Tanzania in 1996 and is maintained at the ICIPE laboratory. We used females 5 d after emergence that had had no prior access to blood, had been maintained on glucose solution (6%) provided with a cotton wick, and had been starved (i.e., provided with water only) for 6 h before the experiment.
This project was supported by the PAL+ program of the French Ministry of Education and Research.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. RL, WRM, and JCK conceived and designed the experiments. RL and WRM performed the experiments. RL and JCK analysed the data. WRM and LCG contributed reagents, materials, and analysis tools. JCK wrote the paper.
¤Current address: Instiut de Recherche pour le Développement (IRD), Département Société et Santé–UR 016, Montpellier, France
Citation: Lacroix R, Mukabana WR, Gouagna LC, Koella JC (2005) Malaria infection increases attractiveness of humans to mosquitoes. PLoS Biol. 3(9): e298.
==== Refs
References
World Health Organization The world malaria report 2005 2005 Geneva World Health Organization Available: http://rbm.who.int/wmr2005/ . Accessed 6 July 2005
Smith T Charlwood JD Kihonda J Mwankusye S Billingsley P Absence of seasonal variation in malaria parasitaemia in an area of intense seasonal transmission Acta Trop 1993 54 55 72 8103627
Macdonald G The epidemiology and control of malaria 1957 London Oxford University Press
Koella JC An evolutionary view of the interactions between anopheline mosquitoes and malaria parasites Microbes Infect 1999 1 303 308 10602664
Koella JC Sørensen FL Anderson R The malaria parasite Plasmodium falciparum increases the frequency of multiple feeding of its mosquito vector Anopheles gambiae
Proc R Soc Lond B Biol Sci 1998 265 763 768
Anderson RA Koella JC Hurd H The effect of Plasmodium yoelii nigeriensis infection on the feeding persistence of Anopheles stephensi Liston throughout the sporogonic cycle Proc R Soc Lond B Biol Sci 1999 266 1729 1733
Koella JC Rieu L Paul REL Stage-specific manipulation of a mosquito's host-seeking behavior by the malaria parasite Plasmodium gallinaceum
Behav Ecol 2002 13 816 820
Kingsolver JG Mosquito host choice and the epidemiology of malaria Am Nat 1987 130 811 827
Day JF Ebert KM Edman JD Feeding patterns of mosquitoes (Diptera: Culicidae) simultaneously exposed to malarious and healthy mice, including a method for separating blood meals from conspecific hosts J Med Entomol 1983 20 120 127 6341591
Ferguson HM Read AF Mosquito appetite for blood is stimulated by Plasmodium chabaudi infections in themselves and their vertebrate hosts Malaria J 2004 3 12
Burkot TR Narara A Paru R Graves PM Garner P Human host selection by anophelines: No evidence for preferential selection of malaria or microfilariae-infected individuals in a hyperendemic area Parasitology 1989 98 337 342 2771444
Mukabana WR Takken W Coe R Knols BCJ Host-specific cues cause differential attractiveness of Kenyan men to the African malaria vector Anopheles gambiae
Malaria J 2002 1 17
Knols BGJ De Jong R Takken W Differential attractiveness of isolated humans to mosquitoes in Tanzania Trans R Soc Trop Med Hyg 1995 89 604 606 8594668
Skinner W Tong H Pearson T Strauss W Maibach H Human sweat components attractive to mosquitoes Nature 1965 207 661 662 5883658
Gillies MT The role of carbon dioxide in host-finding by mosquitoes (Diptera: Culicidae): A review Bull Entomol Res 1980 70 525 532
Takken W Knols BGJ Otten H Interactions between physical and olfactory cues in the host seeking behavior of mosquitoes: The role of relative humidity Ann Trop Med Parasitol 1997 91 Suppl 1 119 120 9093438
Mukabana WR Takken W Killeen GF Knols BGJ Allomonal effect of breath contributes to differential attractiveness of humans to the African malaria vector Anopheles gambiae
Malaria J 2004 3 1
Bousema JT Gougna LC Drakeley CJ Meutstege AM Okech BA
Plasmodium falciparum gametocyte carriage in asymptomatic children in western Kenya Malaria J 2004 3 18
|
16076240
|
PMC1182690
|
CC BY
|
2021-01-05 08:21:26
|
no
|
PLoS Biol. 2005 Sep 9; 3(9):e298
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030298
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030306SynopsisAnimal BehaviorEvolutionInfectious DiseasesEpidemiology/Public HealthAnophelesPlasmodiumTo Mosquitoes, People with Malaria Smell Like Dinner Synopsis9 2005 9 8 2005 9 8 2005 3 9 e306Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Malaria Infection Increases Attractiveness of Humans to Mosquitoes
==== Body
Malaria is a misnomer. People used to believe that poisoned or “bad air,” the translation of the Italian phrase “mal aria,” caused disease. In the 19th century, when parasitologists figured out that single-celled parasites cause malaria, they didn't bother to change the disease's name. Experimenters proved that these parasites need a host organism to survive—so they can't be transmitted through air—and that the hosts, mosquitoes, carry the parasite to humans. Researchers were optimistic that if they could find a disease's cause, they could also find the cure. Kill the mosquitoes and eradicate malaria. And with the advent of DDT and less environmentally harmful insecticides, potent anti-malarial drugs, and international funding in the late 20th century, eradication of malaria seemed imminent.
But that expectation underestimated the flexibility of living creatures. Mosquitoes acquired resistance to insecticides while the parasites acquired resistance to anti-malarial drugs. Worse, the aggressive eradication campaign skipped over vast regions of the globe, especially sub-Saharan Africa.
Malaria remains a devastating problem in Africa for several reasons. Environmental conditions provide an amenable atmosphere for both Plasmodium falciparum, the most dangerous form of the parasite, and the Anopheles gambiae mosquito, the most effective vector. Also, many countries in sub-Saharan Africa lack the infrastructure to protect their citizens from malaria. Given the overwhelming scope of malarial infection in Africa, new understanding of the disease will help epidemiologists devise targeted anti-malarial strategies.
Mosquitoes are most attracted to children infected with malarial parasites in the gametocyte stage (pictured above). The Anopheles mosquito ingests gametocytes during its blood meal. (CDC/ Dr. Mae Melvin)
A new study conducted in Western Kenya by Jacob Koella and colleagues analyzed mosquito behavior to discover how it facilitates the transmission of malaria. The research determined that mosquitoes are more attracted to people infected with transmittable malaria than to either people infected with non-transmittable forms of the disease or uninfected people. To measure the attraction of the mosquitoes, the researchers set up a chamber of infected mosquitoes surrounded by tents containing the study participants. A device called an olfactometer wafted the odors of each participant toward the mosquitoes. Researchers measured which smell most attracted the hungry bugs.
This question had long stalled scientists because of contradictory and indirect evidence. Sweat, breath odor, and high body temperature all increase mosquitoes' blood lust, and no previous study had isolated the variable of malarial infection.
To control for the natural variation in how attractive mosquitoes found each participant, Koella et al. compared the number of mosquitoes that were attracted to infected people to the number of mosquitoes that were attracted to those same people after they were no longer infected. The researchers found that in general, an individual attracted more mosquitoes when infected with transmittable malaria. This demonstrates that malaria, in addition to causing fever, vomiting, headache, and sometimes death, causes more mosquito bites. The biting mosquitoes will then pick up the parasite and spread it to other people.
As another control, the researchers compared infection with a non-transmittable form of the parasite to infection with the transmittable form and to no infection. A mosquito can pick up the malaria parasite only when in its sexually reproductive stage. The transmittable parasite, known as a gametocyte, multiplies in the mosquito's belly before traveling to the mosquito's salivary glands and, eventually, to the blood of the next human victim. But the malaria parasite has a complicated life cycle that also includes non-transmittable asexual stages. Koella and colleagues found that these parasitic forms, unlike the sexually reproductive form, did not make humans more attractive to mosquitoes.
Previous to the recent study, malaria researchers had proved that mosquito biting rates greatly influence the spread of malaria. Koella and colleagues showed that the parasite itself increases these biting rates when it is ready for a new host.
|
0
|
PMC1182691
|
CC BY
|
2021-01-05 08:21:25
|
no
|
PLoS Biol. 2005 Sep 9; 3(9):e306
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030306
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030318SynopsisEvolutionGenetics/Genomics/Gene TherapyAnophelesIslands in the Genome Promote Speciation Synopsis9 2005 9 8 2005 9 8 2005 3 9 e318Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Genomic Islands of Speciation in Anopheles gambiae
Using DNA microarrays, the authors identify three small regions of the genome that differ between two forms of hybridizing mosquitoes-regions that are likely to contain the genes responsible for reproductive isolation.
==== Body
Have you ever wondered how the myriad insect forms—beetles, flies, dragonflies, mosquitoes, grasshoppers, ants, wasps, bees, and countless others—evolved? Insects make up 75% of all species known. The large number of insect species is probably a result of a combination of one or more factors: a high rate of formation of new species, or speciation, an ability to adapt to new environments and exploit new ecological niches, and a lower rate of extinction. Speciation, adaptation, and extinction are all controlled by the interplay between genetic and environmental factors. Understanding the genetic changes that lead to the formation of new species is an important area of research in evolutionary biology.
Two forms of the Anopheles mosquito, which transmits the malarial parasite, shed light on the genetic changes that prevent related species from producing fertile offspring— a condition of speciation. (CDC)
In a new study, Thomas Turner, Matthew Hahn, and Sergey Nuzhdin worked with the malaria mosquito Anopheles gambiae to uncover genes that may be driving speciation. A. gambiae exists in multiple forms that may be in the early stages of differentiating into separate species; on the other hand, they may be partially differentiated, co-existing races that could give us valuable information on genes responsible for racial differences in mosquitoes. Turner and colleagues focused on two forms, A. gambiae M and A. gambiae S, that sometimes mate and create hybrid forms in nature. While it's unclear whether the forms can produce fertile hybrid offspring in the wild, the progeny of lab matings appear to have no problems with fertility. This suggests that individuals either naturally prefer to mate with others of their own form, or that there must be environmental and/or genetic conditions that are not favorable for the survival of hybrid progeny in nature.
To study the genetic underpinnings of speciation, the researchers used DNA microarrays to identify global differences in the mosquito genomes. Using a combination of gene chips, statistics, and computational biology, Turner and colleagues found that the M and S genomes differ at just three regions. The researchers suggested that genes present here may be responsible for early speciation. These three “speciation islands” in the genome contain 67 predicted genes. In a preliminary analysis of seven of these genes, Turner and colleagues identified five that are different between the two Anopheles forms; these include genes that play a role in a range of cellular processes, including energy metabolism, response to sudden increases in temperature (heat shock), and ion transport across cell membranes. Future work focusing on the 67 genes hypothesized to reside in the divergent regions should yield interesting clues to the identity of genes that drive speciation, and the mechanism by which they do so.
This is a significant finding in the field of speciation research: in terms of methodology, this study shows that DNA microarrays can be used to identify regions of the genome that are different between two diverging species, allowing researchers to home in on potentially interesting genes. This study also shows that in spite of possible cross-flow of genetic material (natural hybrids between the two forms are found at a low frequency) between two populations, the populations can still be accumulating differences in their genomes—differences that could eventually lead to the formation of new species. Comparing results in Anopheles and the well-studied insect model Drosophila, in which scientists have also started identifying “speciation genes,” should tell us if similar genes are employed repeatedly in different genera during the formation of new species.
|
0
|
PMC1182692
|
CC BY
|
2021-01-05 08:21:26
|
no
|
PLoS Biol. 2005 Sep 9; 3(9):e318
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030318
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1607624410.1371/journal.pbio.0030319Research ArticleEcologyPlantsResponses of Grassland Production to Single and Multiple Global Environmental Changes Grassland Responses to Global ChangeDukes Jeffrey S [email protected]
1
¤Chiariello Nona R
2
Cleland Elsa E
2
Moore Lisa A
1
Shaw M. Rebecca
1
Thayer Susan
3
Tobeck Todd
1
Mooney Harold A
2
Field Christopher B
1
1Department of Global Ecology, Carnegie Institution of Washington, Stanford, California, United States of America,2Department of Biological Sciences, Stanford University, Stanford, California, United States of America,3Department of Plant Biology, Carnegie Institution of Washington, Stanford, California, United States of AmericaLoreau Michel Academic EditorEcole Normale SuperieureFrance10 2005 9 8 2005 9 8 2005 3 10 e31927 9 2004 13 7 2005 Copyright: © 2005 Dukes 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.
Field Tested: Grasslands Won't Help Buffer Climate Change as CO2 Levels Rise
In this century, increasing concentrations of carbon dioxide (CO2) and other greenhouse gases in the Earth's atmosphere are expected to cause warmer surface temperatures and changes in precipitation patterns. At the same time, reactive nitrogen is entering natural systems at unprecedented rates. These global environmental changes have consequences for the functioning of natural ecosystems, and responses of these systems may feed back to affect climate and atmospheric composition. Here, we report plant growth responses of an ecosystem exposed to factorial combinations of four expected global environmental changes. We exposed California grassland to elevated CO2, temperature, precipitation, and nitrogen deposition for five years. Root and shoot production did not respond to elevated CO2 or modest warming. Supplemental precipitation led to increases in shoot production and offsetting decreases in root production. Supplemental nitrate deposition increased total production by an average of 26%, primarily by stimulating shoot growth. Interactions among the main treatments were rare. Together, these results suggest that production in this grassland will respond minimally to changes in CO2 and winter precipitation, and to small amounts of warming. Increased nitrate deposition would have stronger effects on the grassland. Aside from this nitrate response, expectations that a changing atmosphere and climate would promote carbon storage by increasing plant growth appear unlikely to be realized in this system.
The first five years of the Jasper Ridge Global Change Experiment show that production responses of the grassland to changes in climate and CO2 concentration are unlikely to lead to increased productivity on their own.
==== Body
Introduction
Since the start of the Industrial Revolution, human activities have changed the composition of the atmosphere at an accelerating rate, with increasingly recognized consequences for Earth's climate and biogeochemical cycles [1–3]. Ecosystem responses to these changes may further affect climate and biogeochemical cycling [3,4], and alter the character of ecosystem services provided to society [5]. During the past two decades, researchers have studied ecosystem responses to changes in climate, nitrogen (N) deposition, and atmospheric carbon dioxide (CO2) [6]. In some natural systems, responses of plant growth and resource use to one of these global changes have been extensively quantified. However, few studies have examined responses of ecosystems to the simultaneous and interacting global changes likely to be seen later this century. Even fewer studies have observed these responses over many years.
Production responses to single environmental changes vary widely among systems, and by year.
First, doubled atmospheric CO2 increased aboveground biomass production by an average of 14% across nine herbaceous systems [6]. However, CO2 enrichment suppressed production in some systems, while increasing it in others by as much as 85%. Some grasslands responded more positively in dry years than wet years [7–9], possibly because plants narrow their stomatal openings under elevated CO2, which leads to water savings.
Second, observed patterns of plant growth across natural gradients of precipitation and across years within locations suggest that increases in precipitation have the most positive effect on plant growth in systems with the lowest annual inputs [10]. Where precipitation exceeds about 3,000 mm per year, additional precipitation may suppress growth [11].
Third, warming increases aboveground biomass production in many systems, with the strongest effects in colder climates. Across 20 experimental warming sites in tundra, grassland, and forest, increases in aboveground productivity averaged 19% [12]. Across natural systems, production tends to increase with increasing mean annual temperature [11]. Within some productive systems, aboveground growth is correlated with maximum growing season temperature [10].
Fourth, responses to N additions are generally positive across temperate, boreal, and arctic systems [13,14].
While all terrestrial systems are experiencing a fairly uniform increase in CO2, the character of other global changes varies from one region to the next. Thus, the mix of global changes impacting a given region will depend on both space and time. Understanding the responses of ecosystems to potentially interacting global changes is critical to predicting ecosystem feedbacks to climate and biogeochemical cycles. In particular, the response of carbon (C) storage in ecosystems is dependent on (and proportionally related to) two ecosystem processes: C inputs from primary production, and the residence time of C in the system [15].
Several previous studies have examined interactions between N availability and other global change factors [16–18], and some have examined interactions between CO2 and changes in water availability [8,9], climate [19–21], or loss of biodiversity [22,23]. However, we are still developing a conceptual framework to describe the conditions under which a given interaction is most important. For instance, mineral element availability may progressively limit positive CO2 responses in some systems, but other systems are unlikely to develop such an interaction [24]. Similarly, where elevated CO2 leads to important soil moisture savings [25], increases in precipitation might negate any CO2 effect. Temperature and CO2 responses are frequently assumed to be additive, although few ecosystem-scale experiments exist [26]. No previous studies, to our knowledge, have simultaneously tested responses to enhanced CO2, warming, increased precipitation, and increased N deposition.
Since 1998, the Jasper Ridge Global Change Experiment (JRGCE) has exposed a moderately fertile grassland to atmospheric and climate conditions expected later this century, and to enhanced nitrate deposition. Because small-statured, annual species dominate California grasslands, this ecosystem is well suited for the study of responses to global changes. Thousands of individual plants can be examined within a small area, and changes in the chemistry of plants and plant litter quickly reach the soil as the plants die. Additionally, the plants complete one generation each year, so competition and selection can “tune” the performance of the grassland to new environmental conditions more quickly than would occur in systems with longer-lived species. While systems dominated by larger, longer-lived organisms might adjust to a step change in CO2 or N deposition over a span of decades, annual grassland can be expected to reach a steady, “representative” response more quickly.
With a wide range of treatments and treatment combinations, the JRGCE provides a foundation for characterizing how ecosystems may perform in the future in a range of possible scenarios. Of particular interest is determining whether ecosystem responses to individual factors are additive. How reliably can we predict ecosystem responses to many concurrent environmental changes based on responses to individual changes? Previously, Shaw et al. [27] focused on CO2 responses in this grassland and found an unexpected result: elevated CO2 suppressed positive production responses to other global changes during the third year of the JRGCE. Here, we present a comprehensive description of the responses of grassland production to all four global changes over the first 5 y of experimental treatments, and discuss these responses in the context of natural, as well as experimental, climate variation. With this expanded dataset, we are able to put the results from Shaw et al. [27] in a larger context, and determine whether there have been consistent changes in grassland net primary production (NPP) that could directly affect the amount of C stored in this ecosystem.
Results
Production Responses to Global Changes
Mean production (NPP) of the control treatment varied from 577 to 933 g m−2 across the 1999–2003 growing seasons (see related data in Figure 1). The four main treatments differed in their average effects on NPP (Figures 2 and 3). Only nitrate deposition consistently affected NPP, causing increases of 21%–42% in all years but 2000. Shoots generally responded more positively to N addition than roots did, leading to decreases in root-to-shoot ratios (Figure 2; Tables S1 and S2). Increased precipitation had little effect on NPP, as negative root responses largely counteracted positive shoot responses. Neither warming nor elevated CO2 significantly affected shoot, root, or total production in any year. Increased precipitation and nitrate deposition frequently decreased root-to-shoot ratios, while heat and CO2 did not affect allocation (Figure 2; Tables S1 and S2).
Figure 1 Biomass Production, Cumulative Precipitation, and Cumulative Temperature from 1998 (the Year before Treatments Began) to 2003
Bars for shoot biomass (grey) and root biomass (black) represent mean values (±SE, n = 8) for quadrants in the “infrastructure control” treatment, which experienced ambient conditions (see Materials and Methods). Cumulative precipitation includes the germinating rain event (defined here as the event that brought cumulative rainfall after October 1 above 12.5 mm) and subsequent precipitation until the last harvest of the year. Cumulative temperature is the sum of average temperatures (in °C) over all days from germination until the final harvest. Shoot biomass (g m−2) was not related to growing season precipitation (mm; linear regression: p = 0.87) or cumulative temperature (°C; p = 0.27), but there was a weak relationship between shoot growth and fall precipitation (linear regression: p = 0.06, slope = 0.646, r
2 = 0.627).
Figure 2 Proportional Responses of NPP (Measured as Root + Shoot Biomass), Shoot Biomass, Root Biomass, and Root-to-Shoot Ratio to the Four Global Change Treatments
Each line represents the response over time to a single global change factor, and each data point represents the sum of eight elevated treatment averages divided by the sum of eight ambient treatment averages. Elevated CO2, C (gray dashed line and filled diamonds); increased temperature, T (thin red solid line); increased rainfall, R (thick blue dashed line and triangles); nitrate deposition, N (thick solid green line).
Figure 3 NPP (Measured as Root + Shoot Biomass) in Individual and Combined Global Change Treatments
Note that in the similar Figure 2 of Shaw et al. [27], the reference treatments were not the plots with all global change factors at ambient levels, but the average of all of the plots with the global-change factors identified under each bar at ambient levels. Note also that while Shaw et al. eliminated two blocks heavily invaded by non-native perennials in 2000–2001, the current analysis uses data from all of the blocks from every year.
Paired bars depict mean values under ambient (open) and elevated CO2 (grey). A line representing mean biomass of the ambient treatment is drawn across each panel to facilitate comparisons. Shaded yellow areas similarly mark zones within one standard error of the ambient treatment. Panels are shaded blue behind treatments receiving N to highlight plant responses to this treatment. Letters in each panel identify treatment effects (α < 0.05; C denotes elevated CO2) from mixed model analyses of NPP data (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). Interactions are presented as multiple letters. Ambient, amb; increased temperature, T; increased rainfall, R; nitrate deposition, N. Error bars denote one standard error.
Nitrate strongly increased NPP in all years but 2000, when a four-way interaction was significant (Figure 3; Table S3). Aboveground biomass responses drove the NPP results, as nitrate increased shoot production in all years but 2000 (Figure 4; Table S4). In 2001, nitrate also affected shoot responses to rainfall and CO2; precipitation responses were more positive under increased N, but only in ambient CO2 levels. In this year only, elevated CO2 suppressed the shoot response to combined N and precipitation (Figure 4). Roots responded positively to nitrate deposition in 1999, but not in subsequent years. In 2000, 2001, and 2003, increased rainfall suppressed root production (Figure 5; Table S5).
Figure 4 Mean Shoot Biomass in Individual and Combined Global Change Treatments
Error bars denote one standard error. Shading, labels, and statistics as in Figure 3.
Figure 5 Mean Root Biomass in Individual and Combined Global Change Treatments
Error bars denote one standard error. Shading, labels, and statistics as in Figure 3, with one addition (†, 0.10 > p > 0.05).
Across all 5 y of the experiment, nitrate deposition strongly increased shoot biomass and NPP, and slightly increased root biomass (repeated measures analysis; Figure 6). The only other treatment to affect biomass production across years was precipitation, which increased shoot growth but suppressed root growth, leading to no effect on NPP (Figure 6).
Figure 6 Mean Total Biomass, Shoot Biomass, and Root Biomass in Individual and Combined Global Change Treatments during the 1999–2003 Period
Shading and labels as in Figure 3. Letters identify treatment effects (α < 0.05) from a repeated measures mixed model analysis (*, p < 0.05; **, p < 0.01; ****, p < 0.0001). Error bars denote one standard error.
How Are Responses to Global Changes Affected by the Background Climate?
In most cases, grassland responses to the global change treatments did not depend on climatic factors as measured by regressions against accumulated degree-days or total precipitation (p > 0.05). The exception was the response of shoot growth to temperature, which increased in warmer years (p < 0.05, slope = 0.001, r
2 = 0.10). While the total precipitation and degree-day sums facilitate simple comparisons of responses across years and climate treatments, these metrics do not necessarily capture the critical aspects of climate during the growing season. The sensitivity of grassland production to interannual variation in weather is further discussed below.
Progressive Effects
The pattern from 5 y of treatments in the JRGCE hints at the possibility of progressive or cumulative effects but is rarely definitive. For NPP, aboveground production, and belowground production, no regression of treatment effect on time is significant. Across all of the single-factor treatments and treatment combinations, treatment effects appeared to progressively decrease root production (p = 0.09). The strongest evidence for progressive effects comes from changes in allocation patterns. The effect of nitrate on root-to-shoot ratios became increasingly negative over time (p = 0.02), and the response of root-to-shoot ratios to elevated CO2 shifted from positive to negative (p = 0.07).
Discussion
Over the first 5 y of the JRGCE, production responded minimally to elevated temperature and CO2, positively to increased nitrate deposition, and in a context-dependent manner to supplemental precipitation. The response of production to elevated CO2 is negative in some but not all treatments and years [27]. As a consequence, NPP did not respond to elevated CO2 over the 5-y period (means: ambient CO2, 925 g m−2; elevated CO2, 887 g m−2; p = 0.62). The lack of response to CO2 enrichment indicates that this and similar grasslands are unlikely to experience strong C sinks from CO2 fertilization over the next century. Neither is the response to any future precipitation increase likely to lead to increased C storage. Future increases in nitrate deposition could increase NPP and perhaps C storage, while the experimental warming was probably too slight to interpret as an analog for the end of the century.
Responses to Global Environmental Changes
Except in groups of treatments [27], production did not respond significantly to elevated CO2 (p > 0.1). This suggests that our system exhibits one of the lower grassland NPP responses [9], but is consistent with results from an earlier open-top chamber study of CO2 effects on a neighboring patch of California grassland. In that experiment, production also was not affected by CO2 enrichment [28]. In the same study, NPP of an adjacent patch of serpentine grassland increased in response to elevated CO2 (averages: +12% and +29% in 2 y), largely as a result of increased growth by a few summer-active species that take advantage of wetter soils in late spring. In the JRGCE, such late-flowering forbs were rare. They were absent from all harvested areas in 1999 and constituted 0.37% of the aboveground biomass harvest in 2003. Our peak biomass harvests took place before these forbs reached full size, but the species were so rare that their responses could have impacted overall NPP only if they were massively sensitive to the observed water savings from elevated CO2 [29].
Warming did not affect grassland production, but it did affect the phenology of many species ([30]; Chiariello et al., unpublished data). Zavaleta et al. [29] found that warming led to earlier senescence of many of the dominant species, leaving additional water in soils over the summer. In a setting with responsive plant species, this water savings could combine with that from elevated CO2 to increase the production and/or establishment of late-season annuals, shrubs, and trees [31].
Precipitation affected plant growth more strongly than warming or CO2, with positive effects on shoots and negative effects on roots across years (Figure 6), although analyses of individual years indicated that these effects were not always significant (see Figures 2 and 5). Averaged across treatments, these counteracting shoot and root effects led to no effect on NPP (see Figures 3 and 6). Why did supplemental precipitation decrease root growth? It is possible that allocation to roots decreased as soil resources became more available. In this case, root growth could have been downregulated by increases in water availability or availability of nutrients that are mobile in water. A leading candidate for such a nutrient would be nitrate, which consistently decreased root-to-shoot ratio, with significant effects in two of the 5 y. It is also possible that root growth is affected by small changes in water availability, or that soils in the precipitation treatment occasionally became waterlogged, suppressing root respiration.
Of the four global changes, nitrate deposition had the most consistent, positive effects on plant production. Nitrate increased shoot growth more consistently and by a greater amount than root growth, leading to lower root-to-shoot ratios. Several previous studies have found similarly positive responses of California grasslands to various forms of N (e.g., [32–35]).
Explaining the CO2 Response
The biomass responses presented here include the results from the 2000–2001 growing season discussed by Shaw et al. [27]. The analysis by Shaw et al. primarily focused on CO2 responses, demonstrating that elevated CO2 partially suppressed NPP increases in response to warming, extra precipitation, and nitrate deposition. Across the 5-y dataset, as in the analysis of Shaw et al., there was not a significant, experiment-wide CO2 effect over the entire dataset or in any individual year.
Why doesn't elevated CO2 increase production in this grassland? Several lines of evidence suggest phosphorus (P) limitation could play a role. Both N deposition and elevated CO2 decrease plant P concentration, and, at least in some treatment combinations, elevated CO2 reduces total plant P uptake [30]. In some years, elevated CO2 appears to favor P uptake by microbes over plants [30]. A comparison of grassland production responses after a summer wildfire at the JRGCE also supports the P limitation hypothesis (H. Henry et al., unpublished data). In this study, elevated CO2 suppressed production in unburned grassland, but not in burned areas. Plant growth also responded more strongly to N deposition in burned areas. Ratios of N to P in shoots of the dominant annual grasses were lower in the burned area, suggesting that the fire may have made available more P in ash deposits. The burn alone did not increase plant growth, indicating P limitation may be triggered by elevated CO2 or N deposition. Henry and colleagues cannot definitively separate effects of P availability from changes in microclimate in the burned area, but further studies on the role of P in this grassland are under way. Ongoing research in the JRGCE is also exploring how changes in herbivory, phenology, allocation, and other factors may prevent the grassland from responding positively to CO2.
The Role of Weather
Rangeland scientists have developed many equations to predict shoot growth (forage yields) in annual grasslands based on weather variables [36–40]. In general, measures of heat (as degree-days) and/or precipitation predict annual shoot growth with reasonable accuracy [39,40]. Warmer years and wetter fall and spring seasons usually lead to greater shoot growth. Why, then, did increased temperatures and precipitation not consistently increase shoot growth in this experiment? Most of the precipitation additions in this experiment were associated with rain events, and most of these events occurred in the colder months of winter. In contrast to precipitation in the fall and spring, winter rainfall is not significantly related to shoot production in California grasslands ([36]; J. Dukes, unpublished data). Supplemental precipitation could have its greatest effect by advancing the start of the growing season, eliminating occasional mid-season droughts, or delaying the end of the growing season in years with dry spring months. Our precipitation treatments never advanced the start of growing seasons, but occasionally reduced drought severity. The most positive effect of precipitation was in 2001 (see Figure 2), when a dry period occurred at the end of the growing season and the added precipitation extended the growing period in addition to eliminating the drought.
Despite predictions that CO2 responses would be most positive in drier years [7], this was not the pattern in the JRGCE. Across the range of annual precipitation that the ambient and watered treatments experienced from 1999 through 2003, there was no trend in CO2 response.
Progressive Effects
Responses of NPP to global changes could be progressive for a number of reasons. Changes in community structure could lead to increases in the abundance of unusually responsive (or unresponsive) species. Continuing additions may directly affect the availability of a potentially persistent resource (e.g., nitrate). Or there might be feedbacks through the quantity or quality of soil organic matter [24]. The strengthening effect of N deposition on root-to-shoot allocation patterns could result from a progressive increase in N availability as a consequence of an increasing ecosystem stock. It may also reflect stimulated N mineralization resulting from increased rates of decomposition [41].
In the JRGCE, the evidence for progressive effects is limited in the results to date. Tests to quantify the magnitude, direction, and persistence of progressive effects are a central goal of continuing studies.
Other Considerations
The change in the harvest strategy from one to two harvests between 2000 and 2001 may account for subtle differences between the responses in the first two and the last three years of the dataset. In some cases, responses to treatments were stronger at the second harvest than in the first. High variability in harvested root biomass complicates the task of quantifying treatment effects on root production. Frequently, variation within a treatment was extreme. For instance, in 2003 the control treatment averaged 278 g m−2 root biomass, with a range of 122–552 g m−2.
Our biomass-based root production results underestimated actual production for two reasons. First, annual root production exceeds peak live biomass by approximately 50% in this grassland [42]. Second, we used root data from soil cores taken to 15 cm, a depth that captures 80%–90% of total root biomass (L. Moore, unpublished data). To assess the impacts of these simplifications, we recalculated our data using correction factors. First, we estimated root biomass to 30 cm. For three of the five years, we had measured root biomass from soil cores to 30 cm depth. In the other 2 y, we estimated peak root biomass to 30 cm based on the ratio of roots to 15 cm and 30 cm in the years for which we had measurements. Second, we multiplied the biomass to 30 cm by 1.54, to account for turnover [42]. These corrections had little effect on patterns of NPP responses to the treatments, and altered the significance of the results in only two cases (the temperature × nitrate interaction became significant [p = 0.047] in 2000, and the nitrate effect became only marginally significant [p = 0.086] in 2002).
Implications
What are the implications of these changes in grassland production for C storage and other ecosystem services? Production increases are a likely requirement for greater C storage, although the relationship between NPP and storage depends on several other factors that affect C in soils. Shoot biomass provides forage for livestock and wildlife, and influences fire behavior in wildlands and urban/wildland interfaces [43]. Results from the first 5 y of the JRGCE suggest that the rising atmospheric CO2 concentration will have small and year-dependent effects on production. Across all treatments and years, total grassland production did not respond to CO2 enrichment. This overall lack of sensitivity suggests that, over the long term, effects of CO2 on ecosystem services are more likely to occur through secondary responses such as changes in tissue chemistry, soil moisture, and species composition than directly through changes in production. In individual years, however, the CO2 response may alter production in a meaningful way [27]. Our warming treatment was quite modest, and the lack of a production response to this small temperature change does not suggest that the system will be insensitive to the greater warming likely to occur by 2100. Additionally, effects of our warming treatment on the phenology of grassland species could have important consequences for forage availability and the duration of the wildfire season ([30]; Chiariello et al., unpublished data). Responses of the fungal community to warming [44] may also be important for C storage. Increases in growing season precipitation led to small changes in shoot production, but acted to strongly decrease root production. As with the other factors, whether precipitation leads to greater C storage under any scenario will depend on the responses of other processes, such as microbial respiration. Of the four global change treatments, nitrate deposition had the most consistent and most positive effects on shoot biomass and total biomass production. These production increases, which averaged 37% and 26%, respectively, have obvious potential to meaningfully alter ecosystem services.
The first 5 y of the JRGCE show that production responses of the grassland to changes in climate and CO2 concentration are unlikely to lead to increased productivity on their own. While interactions among changes in climate and CO2 may influence biomass production in specific years, the consequences of these interactions appear limited when averaged over longer time scales.
The JRGCE is one of the most comprehensive global-change experiments to date. It is one of relatively few ecosystem scale experiments to use naturally occurring—as opposed to artificially constructed—ecosystems. We see no reason to think that the kinds of responses observed in the JRGCE are not quite general, at least for natural communities in temperate climates on soils of moderate nutrient availability. Comparing the single-factor responses in the JRGCE with single-factor responses in other ecosystems should provide an efficient approach for assessing the generality of the JRGCE responses to simulated global changes.
Materials and Methods
Study site and system
The JRGCE is located in Jasper Ridge Biological Preserve, near Woodside, California, United States (37°24′N, 122°14′W, 120 m elevation). This region experiences a Mediterranean climate, with cool, wet winters and warm, dry summers. The experiment was conducted in 36 plots dispersed across ∼0.75 ha of natural grassland. Each plot was circular, 2 m in diameter, and divided into four equal-sized quadrants. The dominant species in this location are typically annual grasses (Bromus hordeaceus, Avena barbata, A. fatua) and annual forbs (Geranium dissectum, Erodium botrys). Perennial grasses and forbs are common but rarely dominant. A few of the ∼35 herbaceous species present in the study area increased in dominance over the course of the experiment, most prominently the perennial grass Danthonia californica and the biennial forb Crepis vesicaria ssp. taraxifolia.
Plots were established in the summer of 1997, with the 1997–1998 growing season used as a pretreatment year. The exceptional rainfall and warmth of the 1997–1998 El Niño (see Figure 1), however, made this year somewhat unusual. From 1974–2003, the site received an annual average of 655 mm precipitation (as measured by a weather station located within 1 km of the experimental area).
Global change treatments
In a complete factorial design, we exposed the grassland to ambient and elevated levels of four factors: atmospheric CO2, temperature, precipitation, and nitrate deposition. Experimental treatments were imposed during every growing season (roughly November to June) starting in fall 1998. We added heat and CO2 at the whole-plot level. Free-air CO2 enrichment (FACE; [45]) was used to elevate atmospheric CO2 to ∼680 μmol mol−1. Single 250-W infrared heaters, suspended 1 m over the center of each warming treatment plot, operated continuously during the growing season from 1998–2002. In the 2002–2003 growing season, the central infrared heater was replaced by four 60-W heaters, each centered over one quadrant. In both cases, the heaters elevated canopy/air temperatures by approximately 1 °C. Unheated plots were equipped with “dummy” heaters to reproduce any shading or other effects of the heaters. Precipitation and nitrate deposition were supplemented factorially at the quadrant level. Quadrants in the increased precipitation treatment received 150% of the annual rainfall, with precipitation supplemented via drip tubing (1998–1999) or overhead sprinklers (1999–2003) shortly after each rain event. In addition, two rain events (20 mm each) were applied at the end of each growing season to extend the rainy season by 3 wk. Nitrate was applied twice per year as Ca(NO3)2. Early in the growing season (November), 2 g N m−2 was added in solution to mimic the flush of accumulated dry deposition N that enters the system with the first rains. Later in the season (January–February), 5g N m−2 was added as slow-release pellets (Nutricote 12–0–0, Agrivert, Riverside, California, United States). Fiberglass barriers (0.5 m deep) separated soil in neighboring quadrants, and kept soil in the plot separate from soil in the surrounding grassland. Above the surface, vertical sections of netting discouraged plants, seeds, and plant litter from crossing quadrant boundaries. This netting had a minimal effect on light, reducing the intensity of incoming photosynthetically active radiation by less than 5% (Sunfleck Ceptometer, Decagon Devices, Pullman, Washington, United States).
A separate set of four 2-m diameter “infrastructure control” plots were demarcated in the field but were not equipped with soil partitions, netting, heaters, or CO2 and water distribution tubes. These plots experienced ambient conditions throughout the experiment. Plant production was measured as described below in two quadrants of each infrastructure control plot.
Treatments vs. predicted global changes
The treatments in this experiment were selected not only to simulate conditions predicted to occur in the region within the next century, but also to allow a more comprehensive understanding of mechanisms driving the grassland responses, their relevance to other ecosystems, and their likely limits. Modulation of CO2 sensitivity by availability of other resources is a dominant but unresolved theme in global change research [6]. Variation in resource availability is a central motif across a broad range of global change studies, and along with climate and species composition, may prove useful in generalizing responses across ecosystems. Consistent patterns of response in relation to resource availability could lay the foundation for extending results to other locations, time periods, or management regimes.
To place our treatment levels in the context of recent predictions, atmospheric CO2 concentrations could reach 680 ppm as early as 2070 [2]. In California, climate is expected to warm by 2.3–5.8 °C within this century [46]. Our warming treatment is therefore quite conservative, and will likely be exceeded by mid-century or earlier [46]. Our precipitation treatment simulates slightly smaller increases than predicted by the Hadley Climate Model, version 2, for 2080–2099 [47,48], although projections with newer models tend to be somewhat drier (cf. [46]). Our N deposition treatment was designed primarily to help determine whether N availability limits grassland responses to other global changes. Jasper Ridge currently receives approximately 0.5 g N m−2 yr−1 [49], but other areas of the world receive rates of deposition approaching or exceeding our treatment [50]. Taken together, the treatments used here provide a broad range of combinations of conditions that may occur both locally and in diverse sites, within this century.
Production measurements
We estimated NPP by summing measurements of live and senesced shoot and live root biomass within each quadrant. Because annual plants dominate the grassland, measurements made at the time of peak biomass (mid April to late May, depending on the treatment and year) provide reasonable estimates of aboveground and belowground production, though without including the transfer of C to mycorrhizae, root exudation, and root turnover during the growing season [42]. In 1998 (the pretreatment year), 1999, and 2000, we harvested aboveground biomass once. Beginning in 2001, we conducted two aboveground biomass harvests approximately 1 mo apart, to increase our chances of capturing the maximum value for the season in each quadrant. For years with two aboveground harvests, we used the maximum value from each quadrant to estimate NPP. During each aboveground harvest, we collected all aboveground plant matter in a 141 cm2 area, and separated the current year's production from that of previous years. Biomass from the first (or only) harvest was separated by species before weighing. Starting in 1999, root biomass was determined by separating live roots out of soil cores (15 cm depth) taken in the area of the first aboveground biomass harvest, shortly after the harvest. All biomass was oven-dried (70 °C) before weighing. We refer to data from each growing season by the year in which the harvests occurred.
Statistical analysis
Treatment effects for our experimental design were analyzed with a full factorial, split-plot model using the PROC MIXED method of maximum likelihood estimation in SAS (SAS v.8, Cary, North Carolina, United States). Warming and elevated CO2 treatments were included as whole-plot effects, and precipitation and nitrate deposition treatments were included as split-plot effects. We tested for treatment effects with the restricted maximum likelihood method, using the containment method for determining degrees of freedom and using ordinary least squares starting values where necessary. For analyses of the full 5 y of NPP data, we used PROC MIXED to run a repeated measures version of the split-plot model, in which we used default starting values, unstructured covariance, and the Kenward-Roger technique for determining denominator degrees of freedom. Unless otherwise indicated, values were untransformed (in these cases, analyses with log-transformed data provided virtually identical results). Other statistical techniques are discussed with the results. We excluded data from 13 quadrants in the 1999 dataset, due to errors in the application of the nitrate treatment and inconsistencies in the output of one heater. In all other years, all 128 quadrants were analyzed.
To examine whether grassland responses to global change treatments were dependent on climatic factors, we regressed proportional responses of shoots, roots, and total biomass to the global change treatments against accumulated degree-days (°C) and growing season precipitation (mm). Proportional response values were calculated using treatment means. Means from treatments with elevated levels of a factor were divided by means of the corresponding treatments with ambient levels of that factor. For instance, when regressing the N response against precipitation, one proportional response value was the mean of the CTN (elevated CO2, temperature, and nitrogen) treatment divided by the mean of the CT (elevated CO2 and temperature) treatment. The rainfall and heating treatments caused four data points to fall on each of two values on the independent axis (precipitation or temperature) each year. For each regression, n = 40.
Supporting Information
Table S1 Results (p-Values) from Mixed Model Analyses of Treatment Effects on Root-to-Shoot Ratio (ln-Transformed)
Because these analyses used the containment method to determine denominator degrees of freedom (DDF), all treatments in a given year had the same number of DDF. Values for DDF from 1999–2003 were 23, 28, 28, 28, and 27, respectively. Elevated CO2, C; increased temperature, T; increased rainfall, R; nitrate deposition, N. Numerator degrees of freedom: 1 (all years).
(43 KB DOC).
Click here for additional data file.
Table S2 Results (p-Values) from Mixed Model Repeated Measures Analysis of Treatment Effects on Root-to-Shoot Ratio (ln-Transformed)
Treatment labels as in Table S1.
(36 KB DOC).
Click here for additional data file.
Table S3 Results (p-Values) from Mixed Model Analyses of Treatment Effects on NPP
Treatment labels as in Table S1. Numerator degrees of freedom: 1 (all years). Denominator degrees of freedom (1999–2003): 23, 28, 28, 28, 27, respectively.
(42 KB DOC).
Click here for additional data file.
Table S4 Results (p-Values) from Mixed Model Analyses of Treatment Effects on Aboveground Production
Treatment labels as in Table S1. Numerator degrees of freedom: 1 (all years). Denominator degrees of freedom (1999–2003): 23, 28, 28, 28, 28, respectively.
(42 KB DOC).
Click here for additional data file.
Table S5 Results (p-Values) from Mixed Model Analyses of Treatment Effects on Belowground Production
Treatment labels as in Table S1. Numerator degrees of freedom: 1 (all years). Denominator degrees of freedom (1999–2003): 23, 28, 28, 28, 27, respectively.
(43 KB DOC).
Click here for additional data file.
We thank the dozens of researchers, technicians, and assistants who have helped out over the course of this project, including: A. Ferreira, B. Mortimer, B. Poulter, M. Rillig, B. Thomas, E. Zavaleta, B. Bohannan, H. Henry, D. Khatry, H. Peters, K. Amatangelo, J. Ayers, E. Ferreira, H. Fields, S. Finlayson, L. Forwand, A. Gallego, S. Gere, W. Gomez, J. Juarez, C. Lund, T. Mahe, M. Milne, W. Nott, S. Robinson, V. Schoung, J. Silva, J. Silvis, J. Thayer, P. Yelton, and many volunteers. We appreciate the comments of three anonymous reviewers. The JRGCE has been supported by grants from the National Science Foundation, the Morgan Family Foundation, the David and Lucile Packard Foundation, Jasper Ridge Biological Preserve, and the Carnegie Institution of Washington. EEC was supported by a US Department of Energy Graduate Research Environmental Fellowship.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. CBF, HAM, NRC, and MRS conceived and designed the experiments. JSD analyzed the data and wrote the paper. JSD, NRC, EEC, LAM, MRS, ST, TT, and CBF contributed substantially to the maintenance of the experiment and the collection, organization, and verification of data relevant to this paper.
¤ Current address: Department of Biology, University of Massachusetts, Boston, Massachusetts, United States of America
Citation: Dukes JS, Chiariello NR, Cleland EE, Moore LA, Shaw MR, et al. (2005) Responses of grassland production to single and multiple global environmental changes. PLoS Biol 3(10): e319
Abbreviations
Ccarbon
CO2carbon dioxide
JRGCEJasper Ridge Global Change Experiment
Nnitrogen
NPPnet primary production
Pphosphorus
==== Refs
References
Walker B Steffen W Canadell J Ingram J The terrestrial biosphere and global change 1999 Cambridge (United Kingdom) Cambridge University Press 439
IPCC Climate change 2001: The scientific basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change 2001 Cambridge (United Kingdom) Cambridge University Press 881
IPCC Climate change 2001: Impacts, adaptation, and vulnerability. Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change 2001 Cambridge (United Kingdom) Cambridge University Press 1032
Gruber N Friedlingstein P Field CB Valentini R Heimann M Field CB Raupach MR The vulnerability of the carbon cycle in the 21st century: An assessment of carbon-climate-human interactions The global carbon cycle: Integrating humans, climate, and the natural world 2004 Washington Island Press 45 76
Gitay H Brown S Easterling W Jallow B Antle J McCarthy JJ Canziani OF Leary NA Dokken DJ White KS Ecosystems and their goods and services Climate change 2001: Impacts, adaptation, and vulnerability. Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change 2001 Cambridge (United Kingdom) Cambridge University Press 235 342
Mooney HA Canadell J Chapin FS Ehleringer JR Körner C Walker B Steffen W Canadell J Ingram J Ecosystem physiology responses to global change The terrestrial biosphere and global change 1999 Cambridge (United Kingdom) Cambridge University Press 141 189
Körner C Biosphere responses to CO2 enrichment Ecol Appl 2000 10 1590 1619
Morgan JA Pataki DE Clark H Del Grosso SJ Grünzweig JM The role of water relations in grassland and desert ecosystem responses to rising atmospheric CO2
Oecologia 2004 140 11 25 15156395
Nowak RS Smith SD Ellsworth DS Plant and ecosystem responses to elevated atmospheric CO2 —Do data from FACE experiments support early predictions? New Phytol 2004 162 253 280
Huxman TE Smith MD Fay PA Knapp AK Shaw MR Convergence across biomes to a common rain-use efficiency Nature 2004 429 651 654 15190350
Schuur EAG Productivity and global climate revisited: The sensitivity of tropical forest growth to precipitation Ecology 2003 84 1165 1170
Rustad LE Campbell JL Marion GM Norby RJ Mitchell MJ A meta-analysis of the response of soil respiration, net nitrogen mineralization, and aboveground plant growth to experimental ecosystem warming Oecologia 2001 126 543 562
Matson P Lohse KA Hall SJ The globalization of nitrogen deposition: Consequences for terrestrial ecosystems Ambio 2002 31 113 119 12077999
Shaver GR Chapin FS Effect of fertilizer on production and biomass of tussock tundra, Alaska, USA Arc Alp Res 1986 18 261 268
Thompson MV Randerson JT Malmstrom CM Field CB Change in net primary production and heterotrophic respiration: How much is necessary to sustain the terrestrial carbon sink? Global Biogeochem Cycles 1996 10 711 726
Chapin FS Shaver GR Giblin AE Nadelhoffer KJ Laundre JA Responses of arctic tundra to experimental and observed changes in climate Ecology 1995 76 694 711
Chiariello NR Field CB Körner C Bazzaz FA Annual grassland responses to elevated CO2 in long-term community microcosms Carbon dioxide, populations, and communities 1996 San Diego Academic Press 139 157
Oren R Ellsworth DS Johnsen KH Phillips N Ewers BE Soil fertility limits carbon sequestration by forest ecosystems in a CO2 -enriched atmosphere Nature 2001 411 469 472 11373677
Oechel WC Cowles S Grulke N Hastings SJ Lawrence B Transient nature of CO2 fertilization in arctic tundra Nature 1994 371 500 503
Peltola H Kilpeläinen A Kellomäki S Diameter growth of Scots pine (Pinus sylvestris) trees grown at elevated temperature and carbon dioxide concentration under boreal conditions Tree Physiol 2002 22 963 972 12359523
Tingey DT McKane RB Olszyk DM Johnson MG Rygiewicz PT Elevated CO2 and temperature alter nitrogen allocation in Douglas-fir Global Change Biol 2003 9 1038 1050
Niklaus PA Leadley PW Schmid B Korner C A long-term field study on biodiversity × elevated CO2 interactions in grassland Ecol Monogr 2001 71 341 356
Reich PB Knops J Tilman D Craine J Ellsworth D Plant diversity enhances ecosystem responses to elevated CO2 and nitrogen deposition Nature 2001 410 809 812 11298447
Luo Y Su B Currie WS Dukes JS Finzi A Progressive nitrogen limitation of ecosystem responses to rising atmospheric carbon dioxide BioScience 2004 54 731 739
Volk M Niklaus PA Körner C Soil moisture effects determine CO2 responses of grassland species Oecologia 2000 125 380 388
Norby RJ Luo YQ Evaluating ecosystem responses to rising atmospheric CO2 and global warming in a multi-factor world New Phytol 2004 162 281 293
Shaw MR Zavaleta ES Chiariello NR Cleland EE Mooney HA Grassland responses to global environmental changes suppressed by elevated CO2
Science 2002 298 1987 1990 12471257
Field CB Chapin FS Chiariello NR Holland EA Mooney HA Koch GW Mooney HA The Jasper Ridge CO2 experiment: Design and motivation Ecosystem responses to elevated CO2 1996 London Academic Press 121 145
Zavaleta ES Thomas BD Chiariello NR Asner GP Shaw MR Plants reverse warming effect on ecosystem water balance Proc Natl Acad Sci U S A 2003 100 9892 9893 12907704
Cleland EE The influence of multiple interacting global changes on the structure and function of a California annual grassland ecosystem [thesis] 2005 Stanford (California) Stanford University 107
Williams K Hobbs RJ Control of shrub establishment by springtime soil water availability in an annual grassland Oecologia 1989 81 62 66
Jones MB Yield, percent nitrogen, and total nitrogen uptake of various California annual grassland species fertilized with increasing rates of nitrogen Agron J 1963 55 254 257
Jones MB Forage and nitrogen production by subclover-grass and nitrogen-fertilized California grassland Agron J 1967 59 209 214
Luebs RE Laag AE Brown MJ Effect of site class and rainfall on annual range response to nitrogen and phosphorus J Range Manag 1971 24 366 370
Hull JC Muller CH Responses of California annual grassland species to variations in moisture and fertilization J Range Manag 1976 29 49 52
Murphy AH Predicted forage yield based on fall precipitation in California annual grasslands J Range Manag 1970 23 363 365
Duncan DA Woodmansee RG Forecasting forage yield from precipitation in California's annual rangeland J Range Manag 1975 28 327 329
Pitt MD Heady HF Responses of annual vegetation to temperature and rainfall patterns in northern California Ecology 1978 59 336 350
George MR Raguse CA Clawson WJ Wilson CB Willoughby RL Correlation of degree-days with annual herbage yields and livestock gains J Range Manag 1988 41 193 197
George MR Williams WA McDougald NK Clawson WJ Murphy AH Predicting peak standing crop on annual range using weather variables J Range Manag 1989 42 508 513
Chapin FS The mineral nutrition of wild plants Annu Rev Ecol Syst 1980 11 233 260
Higgins PAT Jackson RB Des Rosiers JM Field CB Root production and demography in a California annual grassland under elevated carbon dioxide Global Change Biol 2002 8 841 850
Rothermel RC A mathematical model for predicting fire spread in wildland fuels 1972 Ogden (Utah) USDA Forest Service 40 Publication INT-115
Rillig MC Wright SF Shaw MR Field CB Artificial climate warming positively affects arbuscular mycorrhizae but decreases soil aggregate water stability in an annual grassland Oikos 2002 97 52 58
Allen LH Field techniques for exposure of plants and ecosystems to elevated CO2 and other trace gases CRC Crit Rev Plant Sci 1992 11 85 119
Hayhoe K Cayan D Field CB Frumhoff PC Maurer EP Emissions pathways, climate change, and impacts on California Proc Natl Acad Sci U S A 2004 101 12422 12427 15314227
Johns TC Carnell RE Crossley JF Gregory JM Mitchell JFB The second Hadley Centre coupled ocean-atmosphere GCM: Model description, spinup and validation Climate Dyn 1997 13 103 134
Wilson T Williams L Smith J Mendelsohn R Global climate change and California: Potential implications for ecosystems, health, and the economy 2003 Sacramento California Energy Commission 121
Weiss SB Cars, cows, and checkerspot butterflies: Nitrogen deposition and management of nutrient poor grasslands for a threatened species Conserv Biol 2000 13 1476 1486
Galloway JN Cowling EB Reactive nitrogen and the world: 200 years of change Ambio 2002 31 64 71 12078011
|
16076244
|
PMC1182693
|
CC BY
|
2021-01-05 08:21:26
|
no
|
PLoS Biol. 2005 Oct 9; 3(10):e319
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030319
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030329SynopsisEcologyPlantsField Tested: Grasslands Won’t Help Buffer Climate Change as CO2 Levels Rise Synopsis10 2005 9 8 2005 9 8 2005 3 10 e329Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Responses of Grassland Production to Single and Multiple Global Environmental Changes
==== Body
Serious gardeners know a healthy harvest depends on the right mix of ingredients at the right time. Nitrogen, which spurs leafy growth at the expense of flowers, works best soon after plants emerge. Phosphate, which promotes blossoms, should come later. Indoor growers often gas their plants with CO2 to boost growth, development, and budding. The really serious indoor gardener can pay $599 for a microclimate controller that automatically regulates temperature, humidity, CO2, lighting, and irrigation.
These components are just as important to natural ecosystems as they are to the well-tended garden. But their composition in the atmosphere and soil is changing at an unprecedented rate. Rising concentrations of CO2 and other greenhouse gases—which have increased apace with global agricultural and industrial development—are linked to warmer temperatures and changes in precipitation patterns. The by-products of industrial and agricultural operations have also increased nitrogen concentrations, saturating systems in highly polluted areas.
Plants use CO2 (and water and light) during photosynthesis to make sugar and grow. A plant absorbs CO2 through its leaves and takes up nitrogen (often in the form of nitrate) and other soil nutrients through its roots. Increasing CO2 concentrations can facilitate plant growth by increasing the rate of photosynthesis, though this effect varies depending on the photosynthetic pathway the plant uses. Trees, for example, use a different pathway than most tropical grasses. Higher carbon inputs can also trigger more efficient nitrogen use. Because plants are primary producers, at the base of the food chain, major shifts in the global cycles that support them will have wide-ranging consequences.
To observe how a natural system might respond to these changes over the long term, a team led by researchers from the Carnegie Institution of Washington and Stanford University created an experimental system in their Northern California backyard. Jeffrey Dukes, Christopher Field, and their colleagues treated grassland plots to every possible combination of current or increased levels of four environmental factors—CO2, temperature, precipitation, and nitrogen influx—to simulate likely regional changes over the next 100 years. Previous studies have tested natural systems’ responses to one or two of these changes, but none has tested the long-term, simultaneous impacts of each.
Researchers work on some of the 36 plots that lie below four infrared heaters in the Jasper Ridge Global Change Experiment. Total plant growth in grassland plots like these rarely responded to changes in climate or atmospheric CO2 concentration
The strongest effects on grassland production came from elevated levels of nitrogen (which typically reaches a fertilization limit). Elevated temperature, precipitation, and, surprisingly, CO2, had minimal impacts. These results suggest, the authors argue, that California grasslands, and ecosystems that respond similarly, are not likely to help buffer the rate of climate change by acting as a carbon “sink”—slowing the rise of CO2 levels by storing more carbon in new growth. It’s thought that ocean and terrestrial ecosystems have stored nearly half the carbon emissions produced by humans since the industrial revolution. If it turns out that other natural systems also fail to sequester as much carbon as scientists once thought, atmospheric CO2 concentrations will rise even faster than expected—with serious implications for future climate change.
The experiments were part of the Jasper Ridge Global Change Experiment (JRGCE), which started on Stanford’s 1,200-acre biological preserve in 1997. Since 1998, this grassland ecosystem has been outfitted with an ecologist’s version of a microclimate controller (complete with CO2 pumps, heaters, and irrigation tubing) and subjected to experimentally controlled atmospheric, climatic, and nutrient conditions. (This study examines the experiment’s first five years.) To quantify the grassland response to these treatments, the authors estimated net primary production, or NPP (the amount of carbon left over after cellular respiration) by measuring shoot and root growth in 36 circular plots scattered across roughly two acres. Four control plots experienced the natural variations of California’s Mediterranean climate.
Overall, increased rainfall, warming, and elevated CO2 had little effect on NPP. (More rain triggered shoot growth but stunted root growth, so NPP wasn’t affected.) In some experimental treatments and years, elevated CO2 actually reduced grassland production. Increased nitrate, on the other hand, led to shoot and root growth imbalance, with shoots growing faster than roots. And this added nitrogen “strongly increased” NPP in every year but one. These results suggest that increasing concentrations of atmospheric CO2 are not likely to increase growth of the roots and leaves of plants in this grassland. Why not?
One possibility involves phosphorus. High levels of CO2 and nitrogen can reduce phosphorus concentrations or limit its uptake in these plants. Ongoing JRGCE experiments are exploring how this and other factors—such as grazing or shifts in seasonal events—might limit the growth effects of CO2.
Because grasslands and forests operate in complex feedback loops with both the atmosphere and soil, understanding how ecosystems respond to global changes in climate and element cycling is critical to predicting the range of global environmental changes—and attendant ecosystem responses—likely to occur. Ecosystem responses may well vary according to their composition and location. By studying natural systems over multiple years on an ecosystem scale, experiments like JRGCE offer a valuable tool for simulating predicted global changes and assessing their likely impacts, region by region.
|
0
|
PMC1182694
|
CC BY
|
2021-01-05 08:21:25
|
no
|
PLoS Biol. 2005 Oct 9; 3(10):e329
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030329
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030330SynopsisDevelopmentEcologyEvolutionGenetics/Genomics/Gene TherapyZoologyMolluscsSpeciation Begins, but Doesn't End, with the Twist of a Shell Synopsis9 2005 9 8 2005 9 8 2005 3 9 e330Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Speciation and Gene Flow between Snails of Opposite Chirality
==== Body
The coil of a snail shell can be either right-handed (dextral) or left-handed (sinistral), based on whether the shell spirals out clockwise or counterclockwise when viewed from above. Most species are composed entirely of individuals that are one or the other type; in exceptional cases, populations may differ in their handedness, or chirality, but within a single population, all individuals tend to be alike. This makes sense, since the mechanics of reproduction are harder between two individuals of opposite chirality (their genitalia are also reversed), reducing the likelihood that they will successfully mate and produce offspring. Over time, therefore, the rarer type will become rarer and rarer until it goes extinct.
This poses the interesting evolutionary question of how a species of one chirality can give rise to another of opposite chirality. If the rarer types are less likely to reproduce, then how do they ever establish themselves beyond a threshold frequency? If they are able to establish themselves, then is a change in chirality—which is caused by a single gene—enough to isolate them so that they are a new species? A study in this issue by Angus Davison et al. sheds light on the complex interplay of factors that influence evolution in the snail Euhadra. Although a single gene does cause a change in chirality, and snails with different chirality are able to mate only with great difficulty, there is nevertheless almost free gene flow between them. Other factors must ultimately become involved to cause speciation.
The 22 species of Euhadra are land-dwelling natives of Japan, and include five sinistral and 17 dextral species. Using mitochondrial DNA analysis to construct a family tree, the authors showed that the sinistral species compose a distinct branch, indicating that this feature arose only once in the history of the genus. How did the first sinistral shell types arise, and why didn't they gradually evaporate from the gene pool? One possibility is “reproductive character displacement,” in which a new feature that directly affects mating, such as sinistral shell chirality, decreases the likelihood that its owner will mate with snails of other, closely related, species that live nearby. While their dextral brothers or sisters waste valuable resources in such unsuccessful interspecific pairings, the few sinistral individuals engage in fewer, but more productive, matings exclusively with their own kind, thus increasing their numbers despite the odds stacked against them.
Even though snails with reversed shell spirals have trouble mating (their genitalia are also reversed), gene exchange occurs freely between two forms in a population, suggesting that speciation requires other factors
To test this hypothesis, Davison et al. constructed a model that took into account a variety of factors, including population density, the proximity of other species, and the maternal inheritance pattern of shell chirality (the direction of a snail's shell is determined so early in development that it is governed not by its own genes, but by its mother's). The surprising conclusion is that the last factor, the unusual mode of inheritance, allows for near free gene flow between the two forms within a population, even if the two forms are themselves almost unable to mate. The reason is that the offspring of a sinistral mother could itself be sinistral, even if it contains entirely dextral genes. Its offspring, though, might include dextral snails, because its own dextral genes determined their shell chirality.
Their model indicated that new chiral types are able to arise, in spite of there being fewer suitable mates, if there is reproductive character displacement. They cannot be considered new species, however, because of the gene flow between them. Reproductive character displacement can account for the speciation of sinistral Euhadra only under a complex set of conditions. Interspecific mating would need to be common among the dextral snails. High population density helps, since it allows those with the rare new form to find each other more easily. But gene flow between left and right forms would preserve the population as a single species, unless other factors, such as difference in habitat use or geographic separation, increased the isolation of the two forms. This argues against so-called “single-gene speciation,” and shows that the creation of a new species requires more than a simple twist of fate.
|
0
|
PMC1182695
|
CC BY
|
2021-01-05 08:21:26
|
no
|
PLoS Biol. 2005 Sep 9; 3(9):e330
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030330
|
oa_comm
|
==== Front
Chiropr OsteopatChiropractic & Osteopathy1746-1340BioMed Central London 1746-1340-13-81597210410.1186/1746-1340-13-8ReviewAre chronic low back pain outcomes improved with co-management of concurrent depression? Middleton Peter [email protected] Henry [email protected] Health Equilibrium, 539 Galston Road, Dural, 2158, NSW, Australia2 Macquarie Injury Management Group, Department Health and Chiropractic, Macquarie University, 2109, Sydney, Australia2005 22 6 2005 13 8 8 11 4 2005 22 6 2005 Copyright © 2005 Middleton and Pollard; licensee BioMed Central Ltd.2005Middleton and Pollard; 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.
Objective
To discuss the role of depression in chronic lower back pain and comment on appropriate methods of screening and co-management.
Data Sources
The current scientific literature was investigated using the online web databases CINAHL, Medline/PUBMED, Proquest, Meditext and from manual library searches.
Data Extraction
Databases were searched from 1980 to the present (2005). Articles were searched with the key words "depression" and "low back pain". Over three hundred articles were sourced and articles were then selected on their relevance to the chronic spinal pain states that present to manual therapy practitioners.
Data synthesis
Pain is a subjective awareness of peripheral nociceptive stimulation, projected from the thalamus to the cerebral cortex with each individual's pain experience being mediated by his or her psychological state. Thus a psychological component will often be associated with any painful experience. A number of studies suggest (among other things) that the incidence of depression predicts chronicity in lower back pain syndromes but that chronic lower back pain does not have the reciprocal action to predict depression.
Conclusion
The aetiology of chronic pain is multifactorial. There is sufficient evidence in the literature to demonstrate a requirement to draw treatment options from many sources in order to achieve a favourable pain relief outcome. The treatment should be multimodal, including mental and emotional support, counseling and herbal advice. While a strong correlation between depression and chronic low back pain can be demonstrated, an apparent paucity of literature that specifically addresses the patient response to chiropractic treatment and concurrent psychotherapy identifies the need for prospective studies of this nature to be undertaken. It is likely that multimodal/multidisciplinary treatment approaches should be encouraged to deal with these chronic lower back pain syndromes.
Chiropracticdepressionpsychosocialchronic low back pain.
==== Body
Introduction
Specific causes for acute back pain, such as infections, tumours, osteoporosis, spondyloarthropathies, and trauma actually represent a minority of pain syndromes requiring specific therapeutic approaches [1]. Chronic pain, by definition is pain "that persists for a month beyond that usual course of an acute illness or a reasonable duration for an injury to heal; that is associated with a chronic pathologic process, (and is) reocurrent at intervals for months or years"[2].
It is also important to recognise that chronic pain can occur in the absence of any pathological process (International Association for the Study of Pain {IASP} Task Force on Taxonomy 1994)[3]. The IASP describe pain as: "an unpleasant sensory and emotional experience associated with actual or potential tissue damage or described in terms of such damage". This definition reflects the view that pain is multifactorial. Engel first heralded this view in 1977 when he proposed the biopsychosocial model of pain, a model that recognises pain to involve variables such as biological, psychological, social, biomechanical dysfunction, physical deconditioning, and entrenched disability. Engel describes these variables as all being important to the generation and maintenance of chronic pain states [4].
Low back pain is defined as pain and discomfort localised below the costal margin to the inferior gluteals folds with or without leg pain as viewed from the rear [5]. Low back pain is common in western cultures with a lifetime adult population prevalence of about 70% [6], and a one-year prevalence of between 15–45%. Peak prevalence is said to occur between 35 and 55 years [7]. Much of the lower back pain is self-limiting with only 2–7% developing chronicity. Reocurrent and chronic back pain account for 75–85% of all costs associated with lower back pain [8,9]. The cause of low back pain is non-specific in most cases and serious conditions are relatively rare [10,11]. These serious conditions are usually marked by "red flag" factors that include: age of onset <20 and >50 years, recent history of violent trauma, constant progressive non mechanical pain (no relief with bed rest), thoracic pain, past medical condition of malignancy, prolonged use of corticosteroids, drug abuse, immunosuppression, HIV, systemically unwell, unexplained weight loss, widespread neurological symptoms, cauda equina syndrome, structural deformity and fever [12].
"Yellow flag" factors are those psychosocial conditions that are associated with an increased risk of developing or perpetuating chronic pain and long-term disability [13]. Yellow flag conditions include: inappropriate attitudes and beliefs about back pain (for example belief that back pain is harmful or potentially severely disabling or expectation of passive treatments rather than a belief that active participation will help), inappropriate pain behaviour (e.g. fear avoidance behaviour and reduced activity levels), work related problems or compensation issues (for example, poor work satisfaction), emotional problems (such as depression, anxiety, stress), tendency towards a low mood and withdrawal from social interaction [5].
The term acute (includes sub-acute) low back pain is defined as pain that has duration of less than three months [14]. Chronic pain is that pain which lasts for more than three months [15]. The subsequent conversion, in the absence of appropriate effective interventions, of acute back pain to chronic back pain has been found to be at times iatrogenic. This is especially so if no specific tissue can be isolated as being the cause of the pain and practitioner attempts to alleviate it prove to be only partially effectual [1]. Repeated failed treatments and various explanations of causation add to the feelings of impotence leading to catastrophising and fear avoidance behaviours (symptoms, pathology and radiological appearances are often poorly correlated) [6,16].
There often exists a strong functional overlay of psychosocial factors or yellow flags that influence this change [1]. It is recognised that there is a relationship between chronic pain and depression [17,18]. It is reported that between 50 and 65 percent of chronic pain patients also have a diagnosis for depression [19]. The treatment implications for chronic pain with the co-occurrence of depression are generally negative, with non-depressed pain patients tending to benefit from treatment more than depressed patients [20]. The relationship is complex and multifactorial, including a lower tolerance for pain in people with depression [21]. Also, an avoidance of activities that may be directly or indirectly associated with the effectiveness of the therapeutic process [22].
Evidence suggesting either a positive or negative pain outcome when psychosocial aspects of a patient's clinical presentation are addressed appears varied. This review discusses the role of depression in chronic lower back pain and comments on appropriate methods of screening and co-management.
Data Sources
The current scientific literature was investigated using online search engines to examine the web based databases CINAHL, Medline/ PUBMED, Proquest and Meditext and from manual library searches.
Study Period Selection
The current scientific literature was investigated from 1980 to the present (2005).
Data Extraction
Articles were searched with the key words "depression" and "low back pain". Over three hundred articles were sourced and articles were then selected on their relevance to the chronic spinal pain states that present to manual therapy practitioners. Approximately 60 references were selected for this review.
Data synthesis
A majority of studies suggest (among other things) that the incidence of depression predicts chronicity in pain syndromes but that chronic pain does not have the reciprocal action to predict depression.
Results and Discussion
Pain is a subjective awareness of peripheral nociceptive stimulation, projected from the thalamus to the cerebral cortex with each individual's pain experience being mediated by his or her psychological state. Thus a psychological component will often be associated with any painful experience. But, is depression predictive of chronic low back pain? It has been suggested that the presence of depressive symptoms predicts future musculoskeletal disorders, but not vice versa [23]. Another study investigating musculoskeletal pain syndromes found that depressed patients were more likely than those without depression to report chronic pain [24]. Demonstration of psychological distress promoting low-back pain has been made [25,26]. In the study by Deyo and Diehl, a group of 1638 subjects without back pain were observed to determine the relationship between psychological distress and low-back pain. The results indicated that symptoms of psychological distress could predict the onset of new episodes of back pain. The psychological factors included depression and anxiety, which, it was stated, were involved in 16% of new episodes of low-back pain in the general population [26].
The results of chiropractic treatment of 526 patients with chronic low back pain with radiation below the knees were recorded in a prospective, longitudinal observational study [27]. The study concluded that patient outcomes were significantly better (using a Visual Analog Scale score [28]) for pain at periods of 6 and 12 months compared to those recorded in a group of 309 patients treated by medical practitioners. Depression notably, was a predictor of a poorer outcome within both groups [27].
Other literature however is equivocal on this point with some authors suggesting that addressing the issue of depression with cognitive behavioural therapy aimed at increasing patient coping strategies gives a poor prognosis towards regaining normal functional capacity [29-31].
They propose that a causal relationship exists whereby disability caused by chronic pain affecting activities of daily living leads to depressive illness. It is implied that a successful therapeutic intervention that targets low back pain could have beneficial effects on depression [32], however, such outcomes have not been conclusively demonstrated in manual therapy groups.
Despite this apparent disparity, it is worthy of note that chiropractors, as primary contact health care practitioners, should look for signs of the psychosocial aspects of chronic pain. Practitioners need to be mindful of the possibility of further exacerbating the pattern of pain by only addressing the musculoskeletal aspects of the problem [33]. Multimodal treatment approaches should be considered and implemented [34-36].
Somatisation and its association with pain perception and depression
Somatisation is a disorder that takes the form of an expression of distress characterised by clinically significant physical symptoms that cannot be explained fully by a physical disorder. It is stated that somatisation is one of the most common of the psychiatric phenomena seen in general medical practice [37] and as such, is a presentation that chiropractors should be aware of. It is usually accompanied by degrees of depression and anxiety. These patients often hold a strong belief in the somatic symptoms they are experiencing, despite an absence of objective measures of physical disease. They can be frustrating patients for clinicians.
Patients with this somatoform disorder and a habit of 'doctor shopping' have been shown to be at a higher risk of a poor outcome after treatment for pain. This is particularly seen with quality of life issues [38,39].
BenDebba and coworkers [40] have examined the stability of the relationship established between the perception of pain and psychological distress after treatment of low back pain. Their findings suggest that the strength of the relationship between chronic pain perception and distress is related to both aspects of the patient's personality and characteristics of their illness and interestingly not to the duration of their complaint [40]. Practitioners who focus on treating somatic structures, such as chiropractors, osteopaths and physiotherapists, may tend to minimise the importance of these psychological factors in the promotion of pain [41].
Screening for pain disability and depression
Identification of the underlying or contributing issue of depression is one that requires appropriate screening tools. Signs that may lead a practitioner to suspect that a patient requires further specific screening for an underlying psychological problem are self reported symptoms that reproduce a pain pattern inconsistent with known anatomical structures and neurology. Variable test results may similarly alert the practitioner and lead to the same conclusion.
Depression associated with low back pain and other pain populations is often different to the classical signs and symptoms of "clinical depression" [42]. Much of the emotional distress in patients with chronic pain does not include the common cognitive characteristics associated with clinical depression. These include feelings of shame, guilt and emotions of anxiety and anger. This is despite the fact that patients are often hostile toward the various medical profession(s) for not resolving their low back pain [43].
Pincus & Williams suggest that, instead of searching for a direct causal path (depression as vulnerability from developing chronic pain or chronic pain leading to development of depression) we accept that this solution does not describe the experience of most of our patients [42]. They suggest that affect and sensory information are processed in parallel, and even if one of the processing channels is more dominant the relationship is most likely to be cyclical. They conclude that we should focus on who is more vulnerable to negative affect and stress as that may allow us to help patients more effectively.
Banks and Kerns concluded in their review that "there is growing empirical evidence to suggest that depression is most commonly secondary to chronic pain" [44]. However, Pincus & Williams conclude that such a conclusion is highly compromised if the measures of depression are contaminated by somatic symptoms reflecting the effect of pain itself or its effect on illness behaviour over time. They further suggest that modern measures should attempt to integrate the disability with pain affect and relate both to psychosocial variables in order to appropriately apply a biopsychosocial model to the management of chronic low back pain conditions.
Commonly used indices and questionnaires such as the Revised Owestry Low Back Pain Disability Questionnaire (ODI)[45] and the Roland Morris Low Back Pain and Disability Questionnaire [46] are objective measures of back specific function and have high clinical utility for the recording of painful disability [47,48]. They do not however, adequately identify the possibility of psychosomatic issues in their matrices. It has been demonstrated that, of the comorbidities most adversely impacting the ODI scores, depression was ranked highly along with osteoporosis, osteoarthritis, blood disorders and headaches [49].
When investigating correlation of pain intensity measured by a Visual Analog Score, the social and anxiety/depression dimensions of the ODI do not appear to be responsive [50].
As such, the ODI should be used as a whole instrument rather than attempt to use subscale components. While the ODI has a demonstrated sensitivity and ability to measure changes in low back pain disability for the purpose of evaluating clinical progress, the lack of sensitivity to identification of possible underlying depressive states demonstrate a need to use a more appropriate instrument.
Screening with a depression specific tool such as the Beck Depression Inventory (BDI) [51] may, in those incidences of high suspicion of an underlying depressive state, be appropriate. This and other questionnaires are frequently used to identify the disability associated with the depression rather than the psychosocial factors associated with them. Therefore, care must be taken in the use of these scales. Other more recently developed questionnaires should be considered to determine such psychosocial variables. An example of such a questionnaire is the depression, anxiety, and positive outlook scale (DAPOS) [52].
The BDI is a 21 item, self-report inventory, with each item consisting of 4 statements rank-ordered in terms of increasing severity for a particular depressive symptom. In use, subjects identify the degree to which each item statement describes the way they have been feeling over the past week. Higher scores indicate greater depressive symptomatology. The BDI has been demonstrated as a sensitive measure of depression in chronic pain patients [18,53].
The fear avoidance beliefs questionnaire (FABQ) was developed to determine if patients' beliefs in physical activity and work affected their low back pain. Research suggests that specific fear-avoidance beliefs about work are strongly related to work loss due to low back pain. These findings are incorporated into a biopsychosocial model of the cognitive, affective and behavioural influences in low back pain and disability. Researchers have recommended that fear-avoidance beliefs should be considered in the management of low back pain and disability [54].
The Distress and Risk Assessment Method (DRAM) questionnaire was developed in an attempt to integrate the physical and psychological assessment of the patient. It is derived from a simple set of scales that were developed for use with low back pain patients. It can distinguish between patients with no psychological distress; those at risk of developing major psychological overlay and those that are distressed [55]. Other measures (the Spielberger Trait Anxiety Inventory, Zung Depression Scale, Modified Somatic Perception Questionnaire, and Cook-Medley Hostility Scale) have been used to predict poor outcome at surgery for lumbar surgical procedures [56].
Treatment Options
The prevalence of major depression in patients with chronic low back pain is approximately three to four times greater than that reported in the general population [57]. Within a chiropractic patient population, Jamison demonstrated the incidence of psychological disease occurring concurrently on initial presentation to be as high as 30%. This suggests that where there is an index of suspicion evaluation of the patients' psychoemotional status needs to be considered [58].
It is argued that, in patients with clinical levels of depression, treatment modalities, including cognitive behavioural therapy and administration of anti-depressant medication, specifically targeting depressive symptomatology deserve serious consideration as an integral component of pain management programs [59].
In a multimodal treatment program of 90 patients with chronic low back pain admitted to an 8-week program of functional restoration and behavioural support the combined functional and psychological treatment resulted in significant improvements among patients by the end of the program [60]. The program consisted of 3 weeks of education, stretching and calisthenic exercises, an intensive treatment period of aerobic, functional strength and endurance exercises, education, cognitive behavioral group therapy and relaxation training in an outpatient program. The targets of psychological intervention were to alter maladaptive perceptions such as somatisation and to counteract depressive symptoms. Reduction of pain or coping with pain, were not primary targets of the program, but changing of negative perceptions and improving coping behaviour were the focus. It was found that the perception of pain was altered favourably and that coping mechanisms aimed at improving functional capacity were similarly improved [59].
A chiropractor needs to consider applying a management program with components that address the psychological aspects of the perception of chronic low back pain. All appropriate patients need to be reassured and given information that explains why they need to become active participants in their treatment. Literature designed to provide this content has been shown to assist positive outcomes in chronic lower back pain patients [60]. Collectively, these interventions may help patients become more confident and less prone to anxiety and depression. Further studies are required to determine if reassurance can alter the levels of anxiety, stress and depression in chronic low back patients.
Simple stress management advice such as yoga, relaxation techniques, an exercise regime and herbal remedies such as skullcap and valerian [61] may be a beneficial (yet unproven) adjunct to musculoskeletal treatment programs. Additionally, St. Johns Wort has been shown to beneficially effect depression [62,63], although others disagree [64]. These measures may at times have insufficient power to be of profound benefit or may only be effective in a hitherto undefined subset of patients.
While some techniques remain unproven psychological cognitive behavioural therapeutic (CBT) options have been demonstrated to have clearly advantageous outcomes with regard to decreasing the pain and distress of chronic pain syndromes [65,66].
Thus, it is likely that multiprofessional rehabilitation will evolve to provide the component parts of the management programs required to maximise outcomes in patients with chronic low back pain [34-36].
Further research
A randomised controlled trial to examine the treatment outcomes of patients presenting for chiropractic treatment with a history of chronic back pain requires the participation of a clinical psychologist/psychiatrist employing psychometric testing. This testing would determine the presence of underlying depression and the need to administer specific treatment modalities. Groups could be broken down into control, manual therapy, manual and cognitive therapy and cognitive therapy to determine which form of therapy demonstrated the best outcomes.
Recent studies have attempted to investigate psychological outcomes in manual therapy based trials. These studies are relatively recent and provide a clear path for future research in the field [16,67,68].
Conclusion
The aetiology of chronic pain is multifactorial. There is sufficient evidence in the literature to demonstrate a requirement to draw treatment options from many sources in order to achieve a favourable pain relief outcome. A requirement for chiropractors to adopt a broader scope of both practice and case management is suggested. Treatments administered should be multimodal with a need to include mental and emotional support, counseling and natural remedy advice (in particular St. John's Wort and possibly Valerian).
While a strong correlation between depression and chronic low back pain can be demonstrated, an apparent paucity of literature that specifically addresses the patient response to chiropractic treatment and concurrent psychotherapy identifies the need for prospective studies of this nature to be undertaken. It is likely that different modes of therapy (exercise, manipulation, mobilisation or combinations of therapy) will have different outcomes. Future studies should focus on effectiveness and the dose response characteristics of these interventions in isolation and in combination.
==== Refs
Ehrlich G Back pain J Rheumatology 2003 26 31
Portenoy R Kanner R Portenoy RK, Kanner RM Definition and assessment of pain Pain management: theory and practice 1996 Philadelphia: FA Davis 3 18
IASP Task Force on Taxonomy Merskey H, Bogduk N Pain terms: a current list with definitions and notes on usage Classification of Chronic Pain 1994 2 IASP Press, Seattle 207 213
Engel G The need for a new medical model: a challenge for biomedicine Science 1977 196 129 136 847460
van Tulder M Becker A Bekkering T Breen A Gil del Real Hutchinson A Vleeming A, Mooney V, Hodges P, Lee D, McGill S, Ostgaard HC European guidelines for the management of active non-specific low back pain in primary care Proceedings of the 5th Interdisciplinary World Congress on Low back and pelvic pain 2004 Melbourne, Australia 56 79 Nov.10–13 2004
Deyo R Rainville J Kent D What can the history and physical examination tell us about low back pain? JAMA 1992 268 760 5 1386391 10.1001/jama.268.6.760
Andersson G Frymoyer JW The epidemiology of spinal disorders The adult spine: Principles and practice 1997 2 New York: Raven Press 93 141
Andersson GB Low back pain J Rehabil Res Dev 1997 34 ix x 9323640
Frymoyer J Back pain and sciatica N Engl J Med 1988 318 291 300 2961994
Suarez-Almazor ME Belseck E Russell AS Mackel JV Use of lumbar radiographs for the early diagnosis of low back pain. Proposed guidelines would increase utilization JAMA 1997 11 1782 6 9178791 10.1001/jama.277.22.1782
Hollingworth W Todd CJ King H Males T Dixon AK Karia KR Kinmonth AL Primary care referrals for lumbar spine radiography: diagnostic yield and clinical guidelines Br J Gen Pract 2002 52 475 80 12051212
Royal College of General Practitioners Clinical guidelines for the management of acute low back pain 1999 London, Royal College of General Practitioners,
Kendall N Linton S Main C Guide to assessing psychosocial yellow flags in acute low back pain: risk factors for long term disability and work loss 1997 Accident Rehabilitation & Compensation Insurance Commission of New Zealand and the National Health Committee, Wellington, New Zealand,
Mersky H Pain terms: a list with definitions and notes on usage recommendated by the IASP Subcommittee on Taxonomy Pain 1979 6 249 252 460932
Mersky H Bogduk N eds Classification of Chronic Pain, 1994 2 IASP Press, Seattle 164 6
Moseley GL Evidence for a direct relationship between cognitive and physical change during an education intervention in people with chronic low back pain Eur J Pain 2004 8 39 45 14690673 10.1016/S1090-3801(03)00063-6
Atkinson J Slater M Patterson T Grant I Garfin S Prevalence, onset and risk of psychiatric disorders in men with chronic low back pain: a controlled study Pain 1991 45 111 21 1831555 10.1016/0304-3959(91)90175-W
Romano J Turner J Chronic pain and depression: does the evidence support a relationship? Psychology Bulletin 1985 97 18 34 10.1037//0033-2909.97.1.18
Kramlinger K Swanson D Maruta T Are patients with chronic pain depressed? Am J Psychiatry 1983 140 747 9 6846634
Dworkin R Richlin D Handlin D Brand C Predicting treatment response in depressed and non-depressed chronic pain patients Pain 1986 24 343 53 3960575 10.1016/0304-3959(86)90120-X
Haley W Turner L Romano L Depression in chronic pain patients: relation to pain, activity and sex differences Pain 1985 46 177 84
Weikgenant A Slater M Patterson T Atkinson J Grant I Garfin S Coping activities in chronic low back pain: relationship with depression Pain 1993 53 95 103 8316396 10.1016/0304-3959(93)90061-S
Aaum P Leino I Magni G Depressive and distress symptoms as predictors of low back pain, neck-shoulder pain, and other musculoskeletal morbidity: a 10-year follow-up of metal industry employees Pain 1993 53 89 94 8316395 10.1016/0304-3959(93)90060-3
Magni G Moreschi C Rigatti-Luchini S Merskey H Prospective study on the relationship between depressive symptoms and chronic musculoskeletal pain Pain 1994 56 289 297 8022622 10.1016/0304-3959(94)90167-8
Croft P Papageorgiou A Ferry S Thomas E Jayson M Silman A Psychologic distress and low back pain Spine 1996 20 2731 7 8747252
Deyo R Diehl A Psychological predictors of disability in patients with low back pain J Rheumatology 1988 27 483 9
Nyiendo J Haas M Goldberg B Sexton G Pain, disability, and satisfaction outcomes and predictors of outcomes: A practice-based study of chronic low back pain patients attending primary care and chiropractic physicians J Manipulative Physiol Ther 2001 24 433 9 11562650 10.1067/mmt.2001.117091
Price D McGrath P Rafii A Buckingham B The validation of visual analogue scales as ratio scale measures for chronic and experimental pain Pain 1983 17 45 56 6226917 10.1016/0304-3959(83)90126-4
Newton-John TR Spence SH Schotte D Cognitive-behavioural therapy versus EMG biofeedback in the treatment of chronic low back pain Behav Res Ther 1995 33 691 7 7654161 10.1016/0005-7967(95)00008-L
Turner JA Jensen MP Efficacy of cognitive therapy for chronic low back pain Pain 1993 52 169 77 8455964 10.1016/0304-3959(93)90128-C
Vlaeyen JW Haazen IW Schuerman JA Kole-Snijders AM van Eek H Behavioural rehabilitation of chronic low back pain: comparison of an operant treatment, an operant-cognitive treatment and an operant-respondent treatment Br J Clin Psychol 1995 34 95 118 7757046
Tait R Chibnall I Margolis R Pain extent: relations with psychological state, pain severity, pain history and disability Pain 1990 41 295 301 2388768 10.1016/0304-3959(90)90006-Y
Jayson M Back pain Bri Med J 1996 313 355 8
Spinhoven P Ter Kuile M Kole-Snijders AM Hutten Mansfeld M Den Ouden DJ Vlaeyen JW Catastrophizing and internal pain control as mediators of outcome in the multidisciplinary treatment of chronic low back pain Eur J Pain 2004 8 211 9 15109971 10.1016/j.ejpain.2003.08.003
Olason M Outcome of an interdisciplinary pain management program in a rehabilitation clinic Work 2004 22 9 15 14757900
McCracken LM Gross RT Eccleston C Multimethod assessment of treatment process in chronic low back pain: comparison of reported pain-related anxiety with directly measured physical capacity Behav Res Ther 2002 40 585 94 12038650 10.1016/S0005-7967(01)00074-2
Clarke D Smith G Somatisation: what is it? Aust Family Physician 2000 29 109 13
Nickel R Egle U Rompe J Eysel P Hoffmann S Somatisation predicts the outcome of treatment in patients with low back pain J Bone Joint Surgery 2002 84 189 95 10.1302/0301-620X.84B2.12235
Lipowski Z Somatisation The concept and its clinical application Am J Psychiatry 1988 145 1358 68 3056044
BenDebba M Torgerson W Long D Personality traits, pain duration and severity, functional impairment and psychological distress in patients with persistent low back pain Pain 1997 72 115 25 9272795 10.1016/S0304-3959(97)00020-1
Seaman D Cleveland C Spinal pain syndromes Nociceptive, neuropathic, and psychologic mechanisms J Manipulative Physiol Ther 1999 22 458 72 10519563
Pincus T Williams A Models and measurements of depression in chronic pain J Psychosom Res 1999 47 211 9 10576470 10.1016/S0022-3999(99)00045-8
Fernandez E Turk DC The scope and significance of anger in the experience of chronic pain Pain 1995 61 165 175 7659426 10.1016/0304-3959(95)00192-U
Banks SM Kerns RD Explaining high rates of depression in chronic pain: a diathesis-stress framework Psychol Bull 1996 199 95 110
Fairbank J Couper J Davies J O'Brian J The Oswestry Low Back Pain Disability Questionnaire Physiotherapy 1980 66 271 3 6450426
Roland M Morris R A study of the natural history of low back pain. Part 1. Development of a reliable and sensitive measure of disability in low back pain Spine 1983 8 141 4 6222486
Fairbank J Pynsent P The Oswestry Disability Index Spine 2000 25 2940 53 11074683 10.1097/00007632-200011150-00017
Maire J Evaluating the effectiveness of care in neck and back pain. Pain and functional status as outcome measures Australasian Chiropractic Osteopathy 2002 10 16 20
Slover J Abdu W Lurie J Hanscom B Weinstein J General versus spine-specific health surveys, do comorbidities influence baseline results? Spine J 2002 2 63 14588290 10.1016/S1529-9430(02)00302-9
Haas M Jacobs G Raphael R Petzing K Low back pain outcome measurement assessment in chiropractic teaching clinics: responsiveness and applicability of two functional disability questionnaires J Manipulative Physiol Ther 1995 18 79 87 7790787
Beck A Steer R Garbin N Psychometric properties of the Beck Depression Inventory: twenty-five years of evaluation Clin Psychol Rev 1988 8 77 100 10.1016/0272-7358(88)90050-5
Pincus T Williams AC Vogel S Field A The development and testing of the depression, anxiety, and positive outlook scale (DAPOS) Pain 2004 109 181 8 15082140 10.1016/j.pain.2004.02.004
Pedersen P Prognostic indicators in low back pain J Coll Gen Pract 1981 31 209 16
Waddell G Newton M Henderson I Somerville D Main CJ A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability Pain 1993 52 157 68 8455963 10.1016/0304-3959(93)90127-B
Main CJ Wood PL Hollis S Spanswick CC Waddell G The Distress and Risk Assessment Method. A simple patient classification to identify distress and evaluate the risk of poor outcome Spine 1992 17 42 52 1531554
Trief PM Grant W Fredrickson B A prospective study of psychological predictors of lumbar surgery outcome Spine 2000 25 2616 21 11034646 10.1097/00007632-200010150-00012
Sullivan M Reesorb K Mikailc S Fisherd R The treatment of depression in chronic low back pain: review and recommendations Pain 1992 50 5 13 1387469 10.1016/0304-3959(92)90107-M
Jamison J Stress: the chiropractic patients' self-perceptions J Manipulative Physiol Ther 1999 22 395 8 10478772
Pfingsten M Hildebrandt J Leibing E Franz C Saur P Effectiveness of a multimodal treatment program for chronic low-back pain Pain 1997 73 77 85 9414059 10.1016/S0304-3959(97)00083-3
Burton AK Waddell G Tillotson KM Summerton N Information and advice to patients with back pain can have a positive effect. A randomized controlled trial of a novel educational booklet in primary care Spine 1999 24 2484 91 10626311 10.1097/00007632-199912010-00010
Metagenics Practitioner Technical Data Manual Nervine relaxant herbs 2000 Health World Ltd 127 30
Linde K Mulrow CD St John's wort for depression Cochrane Database Syst Rev 2000 CD000448 10796719
Muller D Pfeil T von den Driesch V Treating depression comorbid with anxiety – results of an open, practice-oriented study with St John's wort WS 5572 and valerian extract in high doses Phytomedicine 2003 10 25 30 12807339
Shelton RC Keller MB Gelenberg A Dunner DL Hirschfeld R Thase ME Effectiveness of St John's wort in major depression: a randomized controlled trial JAMA 2001 285 1978 86 11308434 10.1001/jama.285.15.1978
Kroenke K Cognitive behavioural therapy for somatatization and symptom syndromes: a critical review of controlled clinical trials Psychotherapy Psychosom 2000 69 205 15 10.1159/000012395
Hoffmann B Back to basics: Confronting psychological issues in patients with low back pain Topics Clinical Chiropractic 1999 6 1 7
Licciardone JC Stoll ST Fulda KG Russo DP Siu J Winn W Swift J Jr Osteopathic manipulative treatment for chronic low back pain: a randomized controlled trial Spine 2003 28 1355 62 12838090 10.1097/00007632-200307010-00002
Burton AK McClune TD Clarke RD Main CJ Long-term follow-up of patients with low back pain attending for manipulative care: outcomes and predictors Man Ther 2004 9 30 5 14723859 10.1016/S1356-689X(03)00052-3
|
15972104
|
PMC1182702
|
CC BY
|
2021-01-04 16:38:23
|
no
|
Chiropr Osteopat. 2005 Jun 22; 13:8
|
utf-8
|
Chiropr Osteopat
| 2,005 |
10.1186/1746-1340-13-8
|
oa_comm
|
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-121587174110.1186/1472-6947-5-12SoftwareThe tissue microarray data exchange specification: A document type definition to validate and enhance XML data Nohle David G [email protected] Leona W [email protected] The Mid-Region AIDS and Cancer Specimen Resource (ACSR), Department of Pathology, The Ohio State University, Columbus, OH USA2005 4 5 2005 5 12 12 22 2 2005 4 5 2005 Copyright © 2005 Nohle and Ayers; licensee BioMed Central Ltd.2005Nohle and Ayers; 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 Association for Pathology Informatics (API) Extensible Mark-up Language (XML) TMA Data Exchange Specification (TMA DES) proposed in April 2003 provides a community-based, open source tool for sharing tissue microarray (TMA) data in a common format. Each tissue core within an array has separate data including digital images; therefore an organized, common approach to produce, navigate and publish such data facilitates viewing, sharing and merging TMA data from different laboratories. The AIDS and Cancer Specimen Resource (ACSR) is a HIV/AIDS tissue bank consortium sponsored by the National Cancer Institute (NCI) Division of Cancer Treatment and Diagnosis (DCTD). The ACSR offers HIV-related malignancies and uninfected control tissues in microarrays (TMA) accompanied by de-identified clinical data to approved researchers. Exporting our TMA data into the proposed API specified format offers an opportunity to evaluate the API specification in an applied setting and to explore its usefulness.
Results
A document type definition (DTD) that governs the allowed common data elements (CDE) in TMA DES export XML files was written, tested and evolved and is in routine use by the ACSR. This DTD defines TMA DES CDEs which are implemented in an external file that can be supplemented by internal DTD extensions for locally defined TMA data elements (LDE).
Conclusion
ACSR implementation of the TMA DES demonstrated the utility of the specification and allowed application of a DTD to validate the language of the API specified XML elements and to identify possible enhancements within our TMA data management application. Improvements to the specification have additionally been suggested by our experience in importing other institution's exported TMA data. Enhancements to TMA DES to remove ambiguous situations and clarify the data should be considered. Better specified identifiers and hierarchical relationships will make automatic use of the data possible. Our tool can be used to reorder data and add identifiers; upgrading data for changes in the specification can be automatically accomplished. Using a DTD (optionally reflecting our proposed enhancements) can provide stronger validation of exported TMA data.
==== Body
Background
ACSR
The National Cancer Institute's (NCI) Division of Cancer Treatment and Diagnosis (DCTD) founded the AIDS Malignancy Bank (AMB) in 1994. Now funded as the AIDS and Cancer Specimen Resource (ACSR), a wide variety of biological samples from HIV/AIDS related malignancies are offered to approved researchers [1]. Tissue microarrays (TMA) with hundreds of HIV infected tissue cores [2] are constructed and offered for distribution by the ACSR. To facilitate researchers' review of available TMAs, digital images of tissue cores and accompanying de-identified data are displayed on the ACSR Mid-Region web site .
API TMA DES
The April 2003 Association for Pathology Informatics (API) TMA Data Exchange Specification (TMA DES) proposes a community-based, open source tool in a common XML format for improved portability of TMA data. A validator (written in Perl) is also provided with the specification (the workshop participants did not approve a DTD and emphasized that users must be allowed to add their own tags) [3]. The usefulness of TMAs to the scientific community is greatly increased by an organized, common approach to producing, navigating and publishing the large quantity of generated data. Exporting TMA data into this API specification format offers an opportunity to evaluate the specification and the validator. Importing and using TMA data from a variety of other institutions also offers an opportunity to further evaluate the utility and functionality of the proposed TMA DES.
XML
"XML, the Extensible Mark-up Language, is a W3C-endorsed standard for document mark-up. XML is a metamark-up language for text documents. Data is included in XML documents as strings of text. The data is surrounded by text mark-up that describes the data [4]." The tag delimiters that will be used as mark-up are defined in a particular application and comprise a particular language.
The API specification defines a mark-up language for the TMA data representation application. Our identifier for a tissue microarray block is TA00-050, a data text string. The API TMA DES tag delimiter for such a block identifier is block_identifier [6]. The marked up data is an XML element:
<block_identifier>
TA00-050
</block_identifier>
XML data may be well formed or valid. Well-formed data has matching end tags that are properly nested and follows generic XML rules. It may not necessarily be valid, i.e. follow the rules of a particular language. To assure that data is valid, it should be validated more rigorously using a DTD. Although a Perl script validator was published with the TMA DES, XML validation must involve a DTD [5].
DTD
A document type definition (DTD) is used to specify a valid language of XML elements. The DTD can be used with various validators [6,7] to check that XML documents written in this language are valid. Although a Schema is an alternative way to specify the properties of an XML document, the DTD syntax was part of early XML standards and has long been supported by various XML processors.
In the DTD language, elements are the tag names that will appear between '<' and '>' in XML documents. Entities are temporary names that will not be used as tags but stand for the text in their definition. They may be referred to preceding the name with '&' and following it with ';'. Elements and/or entities are separated by ',' which means followed by or '|' which means or. Parentheses are used to group elements and entities. A suffix can follow groups, elements or entities: '?' means zero or one, '+' means one or many, and '*' means zero or many. Absence of a suffix indicates that a single one is mandated.
This document type declaration (not to be called a DTD although it may contain definitions between the square brackets) defines a block element that must contain a single block_identifier element and may be followed by a single block_description element and any number of core elements:
<!DOCTYPE block [
<!ELEMENT block_identifier (#PCDATA)>
<!ELEMENT block_description (#PCDATA)>
<!ELEMENT core (#PCDATA)>
<!ELEMENT block (block_identifier,
block_description?, core*)>
]>
The #PCDATA (parsed character data) type allows raw text (but not tags or child elements) to be included in an element. The definition portion could appear at the beginning of an XML file (an internal DTD). An external DTD file, block.dtd, containing such a definition could be referred to with the following declaration in the beginning of an XML file:
<!DOCTYPE block SYSTEM "block.dtd">
When an external DTD file is referred to and an internal DTD subset is present in an XML file, the two parts collectively are the DTD as shown in Figure 1.
Figure 1 Document type definition subsets A DTD can have either an external portion (or subset) or an internal portion or both. Together, these subsets are regarded as the complete DTD. The external DTD subset is in an XML file (or group of files) by itself. The internal DTD subset is placed in the beginning of the XML file within the document type declaration before the root element start tag. The document type declaration refers to the file containing the external DTD subset file name (block.dtd in this case) and contains the DTD name (block in this case) which must be the top level element in the XML file.
The DTD syntax allows definitions to be placed in a file and included into the internal portion as well. In the following example, block.dtd contains the external portion of the DTD and myBlockTags.dtd contains most of the internal portion:
<!DOCTYPE block SYSTEM "block.dtd" [
<!ELEMENT % myBlockTagFile SYSTEM "myBlockTags.dtd">
%myBlockTagFile;
]>
The first definition of a given element or entity encountered is used. Internal DTD definitions are encountered before external ones.
Implementation
The ACSR has been improving the organization of generated TMA data for several years [8-10]. A sample TMA export illustrates the TMA DES format [See TA00-050.xml in Additional file 1]. Compare this to an improved version for the same TMA during the discussion below [See TA00-050_recordered.xml in Additional file 1].
Designing the DTD
The XML document type definition (DTD) for the TMA DES must represent the rules given in the specification and the 80 CDEs defined in the associated Tissue MicroArray Common Data Elements document. This DTD [See tmades.dtd in Additional file 1] accommodates the following special circumstances:
• The TMA DES permits locally needed data element (LDE) definitions to extend those defined in it. A thorough DTD must specify every element. When such elements are added, definitions must be added to the DTD locally. The next section, Extending the DTD, explains this further.
• The order of elements within a parent element is never constrained while the presence of at least one of certain CDEs is mandated. The DTD language will not necessarily allow specification of a certain number of elements when the order of elements is unconstrained. The DTD must be deterministic, i.e. a parser looking at successive elements must have only one path forward through the rules. The section called Mandating CDEs discusses this further.
• The TMA DES specifies a hierarchical nesting arrangement for the CDEs defined in it, it does not define the specific values for them. The content of elements is therefore specified as parsed character data (#PCDATA).
Extending the DTD
Our DTD provides for extensions to the specification to add data elements not in the specification. We made DTD extensions to implement those needed at the Mid-Region ACSR, which serve as examples of how to implement the TMA DES.
Our design uses a single external DTD file with selectable modes to contain definitions that implement the API proposed TMA DES and our proposed improvements and additions to it. We defined three place holder entities (block_LOCAL-TAGS, core_LOCAL-TAGS, slide_LOCAL-TAGS) that can be redefined in the internal portion to extend the specification with local defined elements (LDE) [See myLDEs.dtd in Additional file 1]. We illustrate how to place these in a separate file and include it into multiple XML files to reduce maintenance.
Suppose that age and gender are tracked for cores in TMA blocks at an institution. There are no CDEs in the specification to hold this data. These internal LDE and entity definitions can be shared by several TMA blocks because they are placed in a file called myCoreTags.dtd:
<!ELEMENT core_AGE (#PCDATA)>
<!ELEMENT core_GENDER (#PCDATA)>
<!ENTITY % core_LOCAL-TAGS (core_AGE | core_GENDER)>
This external DTD file fragment defines core_LOCAL-TAGS as a place-holder:
<!ELEMENT core_array-id (#PCDATA)>
<!ELEMENT core (#PCDATA)>
<!ENTITY % core_LOCAL-TAGS (place-holder)>
<!ELEMENT core (core_array-id, (%core_STD-TAGS; |
%core_LOCAL-TAGS;)*>
Multiple XML files, as in the following example, can reference the public external DTD file as well as the locally shared internal DTD file eliminating maintenance of multiple copies. (Usually a public URI is specified: . The file name is used here to simplify the example.)
<!DOCTYPE histo SYSTEM "tmades.dtd" [
<!ENTITY % noLDEs 'IGNORE' >
<!ENTITY % myCoreTagFile SYSTEM "myCoreTags.dtd">
%myCoreTagFile;
]>
The first (internal) definition encountered is used. The external definition of block_LOCAL-TAGS as a place-holder is thus ignored when an internal one is provided. A similar arrangement is shown in Figure 2.
Figure 2 Implementation A DTD governs the allowed data elements in each type (TMA block list, TMA LDE, TMA DES export & style) of XML file. The TMA block list uses BrowseTMA to process each listed block file and style. BrowseTMA produces a single HTML TMA block list file, a TMA block HTML file for each specified block and an HTML TMA style file for each style used.
TMA DES export (& style) DTD is always tmades.dtd, TMA LDE DTD could be myCoreTags.dtd and TMA DES export XML would then contain the above document type declaration.
Mandating CDEs
In the following examples, assume there is an entity called header_unlimited_items that contains all of the elements in the header except filename as follows:
<!ENTITY % header_unlimited_items "(
Title |
Creator |
Subject |
Keywords |
Description |
Publisher |
Contributer |
Date |
Resource_Type |
Format |
Resource_Identifier |
Source |
Language |
Relation |
Coverage |
Rights_Management)">
Here is an example of a problem situation. There is a list of CDEs that may appear inside the header CDE. The filename CDE is in this list and has a maximum occurrence of one. All other CDEs in this list have a maximum occurrence that is unlimited. The order of elements is not limited and no CDEs are required. Here is a straightforward way to describe this in a DTD:
<!ELEMENT header (
(%header_unlimited_items;)*,
filename?,
(%header_unlimited_items;)*)>
Stated in English: a header is zero or many of the non-filename elements followed by a single optional filename followed by perhaps more non-filename elements. Although this faithfully describes our situation, a parser will decide this definition is non deterministic because for some input there are multiple paths through the definition (ex. when there are two Title elements and no others, the term before and/or after the filename can be used for them).
The following definition faithfully represents the allowable header contents and is deterministic but is not as easy to understand:
<!ELEMENT header (
(filename, ((%header_unlimited_items;)*)) |
(((%header_unlimited_items;)+),
((filename, ((%header_unlimited_items;)*)))?))>
Stated in English: a header is either:
• a filename followed by zero or many of the non-filename elements or
• this sequence:
◦ one or many of the non-filename elements followed by
◦ zero or one of this sequence:
▪ a filename followed by
▪ perhaps more non-filename elements.
If the language is changed slightly to say that any number of all header elements is allowed a simple definition is possible:
<!ELEMENT header (((%header_unlimited_items;)*))>
Likewise, if a variety of numbers of header elements is necessary (with some restricted differently than others) but it is acceptable to mandate an order, a simple definition is possible:
<!ELEMENT header (filename, Title*, Creator+, Subject?, ...)>
There are other elements that have a situation similar to the header:
• A tma is required to have one or more headers and one or more blocks but the order is unconstrained.
• A block is required to have one or more slide elements, one or more core elements and optionally any number of (non-parent) block elements in an unconstrained order. (In fact, the requirement for at least one slide and one core may be only for the entire file and not per block. A very convoluted DTD would be required to support a per file restriction.)
A DTD that facilitates validation can also have a role in communicating the details of what is acceptable in the XML application language. While allowing as much flexibility as possible so as to not restrict the style that TMA users might like to employ is desirable, simplifications that make the language easier to understand are also desirable and at times may be involved in a trade-off.
Improving the DTD
While the DTD implements the specification as initially proposed, some improvements were added for conditional use based on the suggestions in [10]. The external DTD can be used in one of two modes by defining one to INCLUDE and the other to IGNORE:
unimproved (default) mode – Enforces the TMA DES rules as proposed, notably:
• Allows multiple header elements within a tma which can be interspersed with block elements within that tma. Ambiguous situations are possible: As each header can have a filename element, what would it mean to have multiple filenames for a single file? Should we associate a certain header with a certain block? How could we tell which?
• Allows at most one filename element to be anywhere within header, if present. The DTD is more difficult to encode and understand unless the filename must be first (or must be an attribute).
• None of these identifiers are required: filename in header, block_identifier in block, slide_identifier in slide, or core_array-id in core. This leaves no certain way to refer to files, blocks, slides and cores. The absence of core_array-id leaves no certain way to locate the core in the array.
• Enforces that at least one block, one slide, and one core element are required. These elements may be empty and serve no purpose. For example, when a TMA block is constructed and no slides have been made from it and it is being provided with exported data to another institution, a slide element must be in the export although no slide exists. Enforcing the presence of these elements does not assure better use of the specification and that data is indeed provided.
improved mode – Enforces the TMA DES rules with the following changes:
• Enforces that there can be at most a single header element within a tma which must precede all block elements within that tma.
• Enforces that filename is the first element within header, if present.
• Enforces that a single identifier is present as the first element within each parent (block_identifier in block, slide_identifier in slide, core_array-id in core). Human readability is improved if an identifier is at the beginning of each element.
• No block, slide, or core elements are required. It is expected that every block, slide and core for which data is to be provided will have a corresponding element containing that data. This expectation does improve usage of the specification.
A program, BrowseTMA Reorder XSLT script [see BrowseTMAReorder.xsl in Additional file 1], was added to BrowseTMA that can convert unimproved TMA DES XML data to improved data by reordering elements and, if needed, adding consecutively numbered identifiers. Several batch and java scripts are used to invoke the XSLT script [see BrowseTMAReorder.bat in Additional file 1], repair the XML file [see fixordered.js in Additional file 1] and invoke the Microsoft XSLT Parser [see xsltTest.js in Additional file 1].
The same DTD also contains the TMA style definitions. By default they are ignored; defining addStyles as INCLUDE will cause them to be available.
Results
Verifying XML data
We used our DTD with the Internet Explorer MSXML2 4.0 parser (set to verify) and at two public web sites to verify a few TMA DES block files.
1. Internet Explorer MSXML2 4.0 parser – No errors or warnings were reported in either mode.
2. Brown University Scholarly Technology Group's XML Validation Form [6] – Reported "Document validates OK" in either mode.
3. Richard Tobin's XML well-formedness checker and validator [7] – Reported that the document appears to be well-formed and listed no validity or namespace errors in either mode.
Conclusion
A user extensible DTD has been written for the API TMA DES and is in routine use in our ACSR program. Improvements to the TMA DES have been suggested by our experience in importing other institution's exported TMA data. Enhancements to the TMA DES are suggested. Better specified identifiers and hierarchical relationships among blocks, slides and cores will facilitate automatic use of the data. Proposed improvements will remove the potential for some ambiguous situations and strengthen the ability to understand data using this format. Our tool can be used to reorder data and add missing identifiers; upgrading data from the original specification to the improved can be automatically accomplished. Using a DTD (optionally reflecting our proposed enhancements) can provide stronger validation of exported TMA data. We hope interested parties will continue to participate in the evolution of the API TMA DES.
Availability and requirements
Example TMA DTD and XML data files and all source code for the BrowseTMA tool and related software are available at the public ACSR Mid-Region web site . Personal computers running Microsoft Windows 98, 2000, NT and XP have been used with Internet Explorer 6.0, Word 2002, and FrontPage 2002 in this work.
List of abbreviations used
ACSR – AIDS and Cancer Specimen Resource
AIDS – Acquired Immunodeficiency Syndrome
AMB – AIDS Malignancy Bank
API – Association for Pathology Informatics
CDE – Common Data Elements
DCTD – Division of Cancer Treatment and Diagnosis
DES – Data Exchange Specification
DTD – Document Type Definition
HIV – Human Immunodeficiency Virus
HTML – Hypertext Markup Language
LDE – locally defined data elements
NCI – National Cancer Institute
OSU – Ohio State University
TMA – Tissue Microarray
TMA DES – Tissue Microarray Data Exchange Specification
URI – Universal Resource Identifier
W3C – World Wide Web Consortium
XML – Extensible Mark-up Language
XSL – Extensible Stylesheet Language
XSLT – XSL Transformation
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DGN developed TMA DES DTD and wrote the first draft of the manuscript. LWA is the principle investigator for this project and participated in writing the manuscript. Both authors reviewed and commented on successive drafts of the manuscript and versions of the software and have approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
All additional files are available at .
This .zip file contains the following eight files which may be extracted using standard unzip software.
Sample specification TMA export: TA00-050.xml: This TMA block contains 35 cores in a 5 by 7 array. This XML file can be viewed with Internet Explorer or another browser or editor.
Sample improved TMA export: TA00-050_reordered.xml: This XML file is a modified version of TA00-050.xml. Certain identifier CDEs (filename, block_identifier, slide_identifier, core_array-id) have been moved to the first position inside the parent CDE. Identifiers have been added where not present. This file can be used with the DTD in improved mode.
TMA DES export DTD: tmades.dtd: This DTD defines the TMA DES CDEs (tags) and allowed structure of API TMA DES exports. Also present are elements to define the allowed structure of TMA styles in TMA style files. It can be viewed with any text editor.
TMA local data element file DTD: myLDEs.dtd: This file defines the LDEs (tags) and their allowed structure. It can be viewed with any text editor.
BrowseTMA Reorder XSLT script: BrowseTMAReorder.xsl: This XSLT file reads in a TMA block XML file and produces another version . In the resulting XML file, certain identifier CDEs (filename, block_identifier, slide_identifier, core_array-id) have been moved to the first position inside the parent CDE. Identifiers have been added where not present. The resulting file will use the DTD in the improved mode.
BrowseTMA WSH batch file: BrowseTMAReorder.bat: This WSH batch file runs the BrowseTMAReorder.xsl XSLT script and then the fixOrdered.js for a TMA block. The resulting file will use the DTD in the improved mode.
Fix ordered XML JScript: fixOrdered.js: This JScript program reads a passed XML file and makes a repaired version.
XSLT test JScript: xsltTest.js: This JScript program runs the Microsoft XSLT Parser, MSXML2 4.0, on a passed XML file and XSLT file.
Click here for file
Acknowledgements
This work was supported by a grant from the National Cancer Institute for support of the AIDS and Cancer Specimen Resource: U01 CA66531. The authors are funding recipients of this grant. Dr. Jules Berman, of the National Cancer Institute, is acknowledged for assistance with the TMA DES specification.
==== Refs
AIDS and Cancer Specimen Resource web site
Kononen J Bubendorf L Kallioniemi A Barlund M Schraml P Leighton S Torhorst J Mihatsch MJ Sauter G Kallioniemi OP Tissue microarrays for high-throughput molecular profiling of tumor specimens Nat Med 1998 4 844 847 9662379 10.1038/nm0798-844
Berman JJ Edgerton ME Friedman BA The tissue microarray data exchange specification: A community-based, open source tool for sharing tissue microarray data BMC Med Inform Decis Mak 3 5 2003 May 23 12769826 10.1186/1472-6947-3-5
Harold ER Means WS XML In A Nutshell 2002 2 O'Reilly & Associates, Inc
Bray T Paoli J Sperberg-McQueen CM Maler E Yergeau F Extensible Markup Language (XML) 1.0 (Third Edition) W3C Recommendation 04 February 2004 W3C Recommendation 2004 February 4
Brown University Scholarly Technology Group's XML Validation Form at
Richard Tobin's XML well-formedness checker and validator at
Nohle DG Hackman BA Ayers LW Web-based virtual tissue microarray slides with clinical data for researchers in HTML, Excel® and API standard XML data exchange specification formats produced with Microsoft Office® applications [abstract] Proceedings of the Advancing Pathology Informatics Imaging and the Internet 8th annual conference: 8–10 October 2003; Pittsburgh, Pennsylvania inpress
Nohle DG Hackman BA Ayers LW Tissue microarray data TMA standard and public tool for linked legend and clinical details at acsr.mid-region.org [abstract] Proceedings of the 8th International Conference on Malignancies in AIDS and Other Immunodeficiencies (ICMAOI): Basic, Epidemiologic and Clinical Research: 98 29–30 April 2004,
Nohle DG Hackman BA Ayers LW A document type definition for the API standard tissue microarray XML data exchange specification [abstract] Proceedings of the Advancing Practice, Instruction and Innovation through Informatics 9th annual conference: 6–8 October 2004; Pittsburgh, Pennsylvania inpress
|
15871741
|
PMC1183103
|
CC BY
|
2021-01-04 23:52:11
|
no
|
BMC Med Inform Decis Mak. 2005 May 4; 5:12
|
utf-8
|
BMC Med Inform Decis Mak
| 2,005 |
10.1186/1472-6947-5-12
|
oa_comm
|
==== Front
Behav Brain FunctBehavioral and brain functions : BBF1744-9081BioMed Central London 1744-9081-1-101603364410.1186/1744-9081-1-10ResearchMore to ADHD than meets the eye: Observable abnormalities in search behaviour do not account for performance deficits on a discrimination task Sonuga-Barke Edmund JS [email protected] Sarah [email protected] Martin [email protected] Developmental Brain & Behaviour Unit, school of Psychology, University of Southampton, Highfield, Southampton, SO17 1BJ, UK2005 20 7 2005 1 10 10 18 1 2005 20 7 2005 Copyright © 2005 Sonuga-Barke et al; licensee BioMed Central Ltd.2005Sonuga-Barke 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.
Children with Attention Deficit/Hyperactivity Disorder (ADHD) often perform poorly on tasks requiring sustained and systematic attention to stimuli for extended periods of time. The current paper tested the hypothesis that such deficits are the result of observable abnormalities in search behaviour (e.g., attention-onset, -duration and -sequencing), and therefore can be explained without reference to deficits in non-observable (i.e., cognitive) processes. Forty boys (20 ADHD and 20 controls) performed a computer-based complex discrimination task adapted from the Matching Familiar Figures Task with four different fixed search interval lengths (5-, 10-, 15- and 20-s). Children with ADHD identified fewer targets than controls (p < 0.001), initiated searches later, spent less time attending to stimuli, and searched in a less intensive and less systematic way (p's < 0.05). There were significant univariate associations between ADHD, task performance and search behaviour. However, there was no support for the hypothesis that abnormalities in search carried the effect of ADHD on performance. The pattern of results in fact suggested that abnormal attending during testing is a statistical marker, rather than a mediator, of ADHD performance deficits. The results confirm the importance of examining covert processes, as well as behavioural abnormalities when trying to understand the psychopathophyiology of ADHD.
==== Body
Attention Deficit/Hyperactivity Disorder (ADHD; [1]) is a disorder of childhood and adolescence characterised by a pattern of extreme, pervasive, persistent and debilitating inattention, overactivity and impulsiveness. Children with ADHD are more likely than their peers to experience educational under-achievement, social isolation and anti-social behaviour during the school years and to go on to have significant difficulties in the post-school years. Children with ADHD often perform poorly on tasks requiring the sustained and systematic allocation of attention over periods of extended time [2]. This appears to be true of tasks that require vigilance for rare targets amongst consecutively presented distractors (e.g., Continuous Performance Task; CPT – [3]). It is also true of more complex tasks that require self-directed and controlled search for targets amongst multiple concurrently presented distractors (e.g., Matching Familiar Figures Task; [4]).
In trying to explain the causes of this commonly observed pattern of performance deficit a range of different mechanisms operating at different levels of analysis have been invoked. For instance, cognitive accounts link deficits in performance to impairments in covert processes such as information encoding and retrieval as well as the 'holding in mind' of targets and their systematic comparison to distractors [5-7]. Such analyses fall firmly within the domain occupied by contemporary models of ADHD which emphasise its cognitive character [[8-10] for a discussion] and appear to receive considerable support from experimental studies of cognitive performance: ADHD children perform poorly on tasks thought to tap a range of executive and non-executive cognitive skills such as working and spatial memory, planning, attentional flexibility and inhibition [11-17].
Despite this strong body of evidence for the existence of cognitive deficits and the compelling nature of the cognitive deficit account, performance on complex discrimination tasks such as those described above can, in fact, be explained much more straightforwardly without invoking deficits in non-observable cognitive processes. This is because effective performance on these tasks requires the provision, protection and, systematic and skilled use of available processing time. This means one could account for the poor performance of ADHD children on such tasks purely in terms of their tendency to (i) start to attend later, and to terminate searches earlier – so producing a shorter duration of attention than controls (i.e., quantitative aspects of attending), and/or (ii) employ less systematic sequencing of attention to individual stimuli and to look at a smaller proportion of stimuli before trying to identify a target (i.e., qualitative aspects of search). Importantly if these quantitative and qualitative abnormalities in observed search behaviour exist they could affect the performance of ADHD children whether, or not, underlying, unobservable cognitive abilities are intact. Furthermore the theoretical significance of such a finding would be considerable. This is because it would in principle offer an explanation of poor performance on many laboratory tests; even those that have been used to index the non-observable cognitive processes discussed above. This in turn would caste doubt on the essentially cognitive nature of ADHD.
Given the potential significance of the search-based account of task performance it is surprising that there has not been more study of the quantitative and qualitative characteristics of search behaviour in ADHD. There is some evidence that children with ADHD fail to provide and exploit sufficient processing time during search. This appears to be partly because of difficulty either sustaining attention or modulating attentional fluctuations that occur over time or in protecting such time from interference by extraneous stimuli [3,18-20]. Furthermore Karatekin and Asarnow, [21] found that ADHD children initiated searches later than controls and fixated for shorter periods of time on more demanding tasks. Frank et al., [22] suggested that ADHD children's searches are self-terminating rather than exhaustive. However, there has been almost no enquiry into the structure of ADHD children's search or its level of organisation. Furthermore there has been no formal analysis of whether any abnormalities in the quantitative aspects of search described above in fact account for, or in statistical terms mediate, performance deficits in ADHD.
The study presented here compares the quality and timing of aspects of ADHD and control children's search behaviour while they are performing a computer-based fixed-trial version of the MFFT. The key questions were; (i) whether ADHD children differ from controls in these aspects of search behaviour and (ii) whether these differences in search behaviour, if identified, would mediate the association between ADHD and poor task performance. According to the statistical concept of mediation support for this hypothesis requires four predictions be confirmed.
I – that there is an association between ADHD and performance with children with ADHD performing more poorly than controls.
II – that there is an association between ADHD and search style with children with ADHD engaging in less systematic or efficient search behaviour than controls
III – that search-related style and behaviour are related to performance in the sample as a whole.
IV – that the association between ADHD and performance is lost when the association between ADHD and search style and behaviour is controlled.
In the current study we set out to test these predictions using 4 different fixed trial conditions in which children had 5-, 10-, 15- or 20 s to identify the target from amongst foils. These different time intervals were used to examine the extent to which ADHD and control children adapt their search style to different temporal constraints and to see whether the impact of restricted search behaviours becomes more important at particular search intervals.
In this study we used a novel approach to measuring search behaviour that did not employ eye gaze measurements. It was based on an analysis of different aspects of the timing and order in which initially 'covered' individual targets on the MFFT were accessed for viewing. The assumption in this study, therefore, is that targets accessed for inspection were actually inspected. This assumption could be tested in future studies using more traditional eye-tracking techniques.
Results
Do AD/HD children perform less well than controls?
The number of correct responses at each inspection interval is reported in Table 1 along with the F statistics and significance from a two-way ANOVA with group (ADHD v controls) as the between subjects factor and inspection interval (5-, 10-, 15-, 20-s) as the within subjects factor. There were effects of both inspection interval and group but no interaction between these factors. Controls outperformed children with ADHD at each level although these differences only reached significance at 5-, t(38) = -2.85; p < 0.01, 10-, t(38) = -3.22; p < 0.005 and 15-s, t(38) = -2.44; p < 0.05, but not at 20-s, t(39) = -1.18; p > 0.20. The significance of the quadratic term suggested that as one might expect children's performance demonstrated a diminishing level of gain for each additional 5 s of processing time. Introducing conduct or emotional problems as covariates into the analysis had no effect on this pattern of significance, F(1,36) = 13.84; p < 0.001.
Table 1 Performance and search characteristics of the ADHD and control groups.
Mean Performance & Search Style (standard deviation) F Statistics
5-s 10-s 15-s 20-s Group Interval GxI
correct responses Control 4.15 (2.21) 6.10 (1.68) 6.30 (2.20) 5.94 (1.84) 13.71* 13.61** 10.25** 0.6
AD/HD 2.56 (1.13) 4.31 (1.80) 4.60 (2.26) 5.15 (2.36)
number boxes open Control 1.54 (0.37) 3.31 (0.50) 4.55 (1.04) 4.22 (1.14) 7.09* 138.42** 36.82** 5.12**
AD/HD 1.30 (0.37) 2.51 (0.79) 3.42 (1.10) 4.14 (1.27)
look time per trial (s) Control 3.03 (0.24) 6.58 (0.68) 10.10 (1.50) 14.38 (1.47) 6.10* 548.9** 0.71 1.54
AD/HD 2.67 (0.47) 5.91 (1.11) 9.96 (1.45) 12.99 (2.71)
search Initiation (s) Control 1.49 (0.29) 1.58 (0.43) 1.61 (0.45) 1.65 (0.42) 4.85* 3.64* 1.77 1.07
AD/HD 1.59 (0.51) 2.02 (0.91) 1.99 (0.75) 2.00 (0.77)
systematic searches Control 5.00 (2.61) 7.95 (1.43) 7.46 (1.23) 7.62 (1.72) 25.76** 39.25** 36.07** 1.89
AD/HD 2.35 (1.98) 5.20 (1.57) 6.15 (2.10) 5.52 (2.11)
Note: AD/HD = Attention Deficit/Hyperactivity Disorder. Figures in parentheses are standard deviations. * p < 0.05; ** p < 0.01. Italicised figures in F column relate to quadratic contrasts.
Was ADHD search behaviour less systematic, intensive and sustained than that of controls?
There was an effect of interval on all four search measures with the number of boxes opened, the length of time the boxes were looked at, the number of systematic searches and time at which searches were initiated all increasing as function of interval length. There was also an effect of group on all measures with ADHD children tending to begin searches later, open fewer boxes, look for less time overall and search in a less systematic way than controls. The effect of group persisted when conduct and emotional problems were entered as covariates; F(1,36) > 5.79; p < 0.05. The effect of group and interval interacted significantly only in the case of the number of boxes opened. This interaction was associated with a levelling off of the number searches made by children in the control group between the 15- and 20-s intervals. However, this interaction was much reduced when conduct and emotional problems were controlled, F (3,34) = 3.41; p < 0.05.
Was AD/HD search style associated with poorer performance in the whole sample?
In order to address this question in an easily comprehensible way search style and performance variables were collapsed across search intervals (5-, 10-, 15- and 20-s). This was justified on the grounds that in general there appeared to be no interaction between inspection interval and group. This suggested that differences in search style and performance demonstrated between the groups were not systematically influenced by search interval (except for number of boxes opened). Furthermore interval specific measures of each variable showed good internal consistency. Cronbach's α for number of systematic searches (.81), number of boxes opened (.75), search initiation (.77), total look time (.55) and number of correct (.63) were all in the acceptable range. These aggregate scores for performance was significantly correlated with systematic searches (.30), number of boxes opened (.32), search initiation (-.42) and average look time (.34).
Did search behaviour mediate the association between ADHD and poor performance?
For this analysis the number of search related-variables was reduced in order to simplify the test of the mediational hypothesis. The four measures were submitted to Principal Components Analysis with Oblimin rotation. The use of this technique for reducing the number of variables is justified on the basis of the current ratio of 1 item or search-related variable for every 10 subjects (i.e., 4 search variables and 40 subjects) although it needs to recognise that the solution may be unstable given the absolute size of the sample. Two factors with eigen values greater than 1 were extracted. These had the following loadings: Factor I (intensive and systematic search; 59.0 percent of the variance) had loadings of .83 for number of systematic searches, .98 number of boxes opened, -.43 for search initiation and -.14 for average look time. The loadings for Factor II (late start – short look; 27.6 percent of the variance) were -.17 for number of systematic searches, .15 number of boxes opened, .70 for search initiation and -.99 for average look time.
The mediational hypothesis was tested using a standard three stage procedure using single and multiple regressions for each search factor separately and with both search factors together. In stage one ADHD status was regressed onto the number of correct responses and the standardised β coefficient derived indicating the strength of the association. In stage 2 the completeness of the chain of associations linking ADHD to performance via search factors was tested by regressing ADHD status onto search behaviour and search behaviour onto the number of correct responses. In stage 3 the search factor was introduced alongside ADHD into a multiple regression analysis to test whether the pathway between ADHD and performance remained significant once search behaviour was controlled. ADHD status was entered in step 1 on the model and search behaviours in step 2. The results are presented diagrammatically in figure 1 along with the regression statistics. As indicated by the preceding analyses ADHD is related to performance and search behaviour. Also search behaviour is related to performance for both search factors and their combination. However, the association between ADHD and performance, although slightly but not significantly reduced, remained clearly significant when search behaviour and ADHD were introduced alongside each other into the analysis. In fact in each case it was the link between search behaviour and performance that was reduced to non-significant levels. This suggests that this association was partly the result of a common effect of ADHD status on both search behaviour and performance. These effects were not affected by entering conduct and emotional problems into the analysis alongside ADHD status into the multiple regression analyses.
Figure 1 Diagrammatic representation of the three steps of the mediational analysis of the effects of search style on performance. In each case step one involves the regression on ADHD group status on to the number of correct responses: Step two involved regressing ADHD into the search factor and then the search factor score on the number of correct responses: Step three involved simultaneously regressing ADHD status and search factor scores onto the number of correct responses. ISS = intense and systematic search; LSSL = Late start short look.
Discussion
The results of the current are consistent with the previous evidence that ADHD children perform poorly on complex tasks requiring systematic search. They also provide new evidence about the abnormalities in their style of search. However, they give no support to the assertion that their performance on such tasks is the result of abnormalities in observable characteristics of their search behaviour; either in terms of the reduced amount of time allocated to search, because of a late onset and a premature off-set of searching, or the more chaotic and sluggish search style. Even taking into account these factors ADHD children still performed more poorly than controls. While it is possible that aspects of search-related behaviour other than those observed and coded might have been the constraining element it is difficult to identify what these additional characteristics might have been. It is also possible the approach taken to measuring search behaviour was insensitive to more subtle differences between the groups that might become obvious if more fine grained approaches to measuring search such as eye-tracking were used. However, one possibility is that deficits in performance displayed by the children with ADHD were the result of deficits in non-observable processes that either were unrelated to search behaviour or at least did not have their effect on performance via an impact on search behaviour. In this sense the current data add support to the idea that ADHD is a disorder with distinct cognitive and behavioural elements which will probably affect different domains of functioning in different ways. From a practical point of view this suggests that interventions targeted at improving attending or search behaviour rather than improving underlying processes are unlikely to be successful in improving task performance.
The performance data is interesting in itself for a number of reasons. First, it does not provide support for accounts of ADHD performance on tasks of extended duration that stress the role of premature task disengagement and the existence of a deficit in sustained attention [23]. Such accounts would have predicted an interaction between group and inspection interval with errors increasing across interval duration at a greater rate in the ADHD, than the control group. In contrast to this prediction the size of the deficit displayed by children with ADHD, relative to controls, was largely independent of interval length. In fact deficits were most marked on the 5-, 10- and 15-s intervals, the opposite of what would be predicted by a task disengagement account. At first glance, this seems to suggest that, relative to that of controls, ADHD performance improved on the 20-s trial. A closer inspection of the results reveals that the narrowing gap between the two groups on this interval is due as much to a slight decline in the control performance as it is to improvements in the ADHD group. Second, the results provide no support for the state regulation account of ADHD [24]. According to this account ADHD children's difficulties arise because of problems modulating their energetic state to meet the changing demands of different settings [25]. A key prediction from this account is that the performance of children with ADHD will be disrupted in settings and tasks with either high (over-arousing) or low (under-activating) rates of stimulus presentation. Support for the state regulation account recently came from a paper reporting the results of two experiments using a similar version of the MFFT to that employed in the current study [26]. In these studies a quadratic interaction between ADHD group status and inspection interval was reported (5-, 10-, and 15-s) with ADHD children performing less well than controls on 5- (short) and 15-s (long) intervals but as well as them on 10-s intervals. The failure of the current study to replicate this pattern may be due to differences between the studies in terms of task difficulty. For instance, in the task version used in the Sonuga-Barke [26] paper all test stimuli were visible throughout every trial as in the original MFFT. In the version used in the current only one stimulus was visible at any one time. The current task was therefore more likely to tap working memory capacities more directly as participants had to hold one stimulus in mind if they wanted to compare it directly with another. The task used in the current paper was therefore likely to have a higher cognitive load and be considerably more demanding than the task used previously. This suggestion is confirmed by a direct comparison of performance on the two tasks. The average error rate in the current study was 54 percent while in the previous studies it was 43 and 44 percent respectively (F(1,78) = 8.85; p < 0.001). It is possible that in the current study the increased levels of task difficulty masked cognitive-energetic effects associated with interval duration. A direct comparison of the two versions of the MFFT is required to test this hypothesis. There was no evidence to support the view that children with AD/HD produce better than expected performance on shorter trials (i.e., 5-s intervals) because they employ compensatory strategies developed in response to the impact of their own impulsive cognitive style on processing opportunities [26]. Deficits were as great on the 5-s, as they were on the 10-, 15- and 20-s, intervals.
The data on search behaviour are also interesting in its own right. ADHD children took longer to initiate their search and spent less time attending to stimuli during the task. Second, they tended to be less systematic and slower during searches when they were actively on task. These findings are consistent with the impression given by previous studies of ADHD children search-related performance [21,22,27]. However, they do not fit well with accounts of ADHD that emphasize the impulsive and disinhibited nature of ADHD children's cognitive performance [28]. On the current task ADHD children's performance was marked by a slow start to searching, a longer time to look at each stimuli and a slower passage from one stimulus to another. This is consistent with the view that ADHD children process information more slowly, rather than more quickly, than controls [29]. However, it must be born in mind that the task used in this study was externally paced, rather than, self-paced. Trial length could not be shortened by 'impulsive responding', a factor that appears to be crucial in determining the extent to which children with ADHD are willing to trade accuracy for speed [30].
The difference between the groups in terms of the ISS factor was considerable for the 5-, 10-, and 15-s intervals. However, there was no difference between the groups at the 20-s interval value. This was related to a drop in the levels of ISS displayed by controls rather than an increase by children with ADHD. This decrease in ISS by the controls was largely due to a slowing down of the average rate of search rather than a drop in other elements of the ISS construct (e.g., numbers of systematic searches). Such a change is perhaps not surprising given the fact that rapid search becomes less important on longer trials. In fact slowing search down could be regarded as a sign that controls were able to adapt their search behaviour to the existing time constraints.
Method
Participants
Forty boys between the ages of eight and twelve (20 ADHD and 20 typically developing controls) took part in the study. The ADHD children were recruited through National Health Service child psychiatric and paediatric outpatient clinics. All had a diagnosis of Hyperkinetic Disorder [31] the equivalent of severe combined type ADHD which affects approximately only 30 percent of the most severe ADHD cases [32]. Diagnosis was made against research criteria on the basis of a thorough clinical investigation that involved parental and child interviews, parent and teacher rating scales and direct observation of the child. In addition, children were only included in the ADHD group if they scored above the clinical cut-off (seven or more out of ten) on the hyperactivity scale of both the parent and teacher versions of the Strengths and Difficulties Questionnaire [SDQ; [33]].
The SDQ is a development of the Rutter questionnaire [34]. It has an extended hyperactivity scale and scales measuring more general conduct and emotional problems [35]. The psychometric properties and norms of the SDQ have been extensively examined in large epidemiological studies within the pre-school age group [36]. Together these criteria would place the ADHD children included in the current sample within the top two or three percent of the childhood population in terms of severity of ADHD symptoms and impairment. No child in the ADHD group received a diagnosis of Hyperkinetic Conduct Disorder although many showed some signs of conduct problems (see table 2). Children were identified from a region of predominantly Caucasian ethnic composition that had a mixed socio-economic background. Seventy-five boys with a diagnosis were originally identified, 35 of these agreed to take part in the study. Fifteen children from the original group were excluded because they did not meet entry criteria for the study (i.e., they had an IQ below 80 and/or were outside the required age range). Children refrained from taking any medication prescribed for ADHD in the 24 hours prior to testing. Controls were selected at random from local schools that reflected the ethnic and socio-economic composition of the region in general. They were children who did not meet either parent or teacher borderline cut-offs for hyperactivity on the SDQ (i.e., less than six on both parent and teacher scales; [33]). Of the 50 control children who were initially contacted 26 agreed to participate. Four of these children were excluded because they failed to meet the inclusion criteria for controls. Two children were excluded because they failed to co-operate during testing.
Table 2 Age, IQ and behaviour ratings for the ADHD and control children.
CONTROLS AD/HD
Mean age 9.9 (0.7) 10.5 (1.5) t = 1.65
Mean IQ 110.3(17.8) 106.7 (13.1) t = 0.71
% conduct problems 0 35 χ2 = 8.48**
% emotional problems 0 20 χ2 = 4.44*
Note: AD/HD = Attention Deficit/ Hyperactivity Disorder; * p < 0.05; ** p < 0.01; % behaviour problems based on children meeting SDQ criteria for presence of difficulties at home and school. Figures in parentheses are standard deviations.
IQ was measured using four sub-scales of the Wechsler Intelligence Scale for Children (WISC-IIIR; [37]); similarities, vocabulary, block design, object assembly. The four sub-tests were pro-rated and an estimate of the child's full scale IQ was derived. Table 2 shows the mean IQ scores and ages for members of the two groups as well as the proportion of children who met SDQ cut-offs for borderline conduct or emotional problems. While groups did not differ in terms of either age or IQ the ADHD group had a higher proportion of children with conduct and emotional problems. Analyses were performed both with and without behavioural and emotional problems as covariates.
Procedure
Children took part in one session consisting of four blocks of 10 trials. At the start of each trial the target stimulus was presented in a box measuring 5.5 × 6.5 cm in the centre at the top of the computer monitor screen. Six other boxes of the same dimensions were presented in two rows of three under the target stimuli. These contained the six test stimuli (five foils and a second copy of the target stimulus). The position of the target copy was varied randomly amongst the foils. At the start of each trial all six boxes containing test stimuli were 'covered' by a white square equal in size to the surrounding box. The target stimulus was visible throughout the inspection period of each trial. To 'open' a box to view one of the six test stimuli participants were required to click on the area within the box with a mouse controlled cursor. When participants had finished inspecting the stimuli in the open box a second click closed the box and re-covered the stimuli. Participants could not open a second or subsequent box until they had closed the box they were inspecting. In block one, children had 5-s in which to inspect the stimuli on each trial, in block two the viewing time was 10-s, in block three it was 15-s and in block four it was 20-s. The order of presentation of these blocks was completely randomised across subjects. At the end of the inspection period (5-, 10-, 15- or 20-s) the squares 'covering' the stimuli were removed so that all stimuli were visible. At the same time the target was covered in order to remove the opportunity for children to continue to search after the end of the viewing period. The participants were prompted to identify the copy of the target by the words "choose now" written on the computer screen. This was achieved by moving the cursor over the chosen stimulus and pressing the button on the top of the mouse. Children were allowed one attempt to select the target on each trial. The experiment was presented on, and data collected using, a Pentium 75 computer connected to a mouse. Variables collected included; time to initiation of search; number of boxes 'opened' during each trial; order in which boxes were opened and length of time which boxes were opened. Two higher order variables were derived: The number of exhaustive searches (i.e., the number of times all boxes were opened during a trial) and the number of systematic searches. A search was deemed to be systematic when boxes were opened in a systematic order (i.e., going across columns or up or down rows) with no box being opened twice until all other boxes had been opened. Ratings of these variables were extremely reliable with 100 percent agreement between two raters. All sessions were video taped and periods of on- and off-task behaviour was recorded. Computer generated, synchronised, real-time information about search behaviour (i.e., if boxes were open or not at a particular point in time) was printed on to the video tape of each session so that the time open boxes were being attended to could be estimated. Inter-rater reliability for these estimates was high with 98 percent agreement between raters.
==== Refs
American Psychiatric Association Diagnostic and statistical manual of mental disorders, 4th ed 1994 Washington: American Psychiatric Association
Manly T Anderson V Nimmo-Smith I Turner A Watson P Robertson IH The differential assessment of children's attention: The Test of Everyday Attention for Children (TEA-Ch), normative sample and ADHD performance J Child Psychol Psychiatry 2001 42 1065 1081 11806689 10.1111/1469-7610.00806
Epstein JN Erkanli A Conners CK Klaric J Costello JE Angold A Relations between continuous performance test performance measures and ADHD behaviours J Abnorm Child Psychol 2003 31 543 554 14561061 10.1023/A:1025405216339
Sonuga-Barke EJS Williams E Hall M Saxton T Hyperactivity and Delay Aversion III; The effect on cognitive style of imposing delay after errors J Child Psychol Psychiatry 1996 37 189 194 8682898
Karatekin C A test of the integrity of the components of Baddeley's model of working memory in attention-deficit/hyperactivity disorder (ADHD) J Child Psychol Psychiatry 2004 45 912 926 15225335 10.1111/j.1469-7610.2004.t01-1-00285.x
Kourakis IE Katachanakis CN Vlahonikolis IG Paritsis NK Examination of verbal memory and recall time in children with attention deficit hyperactivity disorder Dev Neuropsychol 2004 26 565 570 15456684 10.1207/s15326942dn2602_2
Brocki KC Bohlin G Executive functions in children aged 6 to 13: A dimensional and developmental study Dev Neuropsychol 2004 26 571 593 15456685 10.1207/s15326942dn2602_3
Mirsky AF Pascualvaca DM Duncan CC French LM A model of attention and its relation to ADHD Mental Retardation Dev Disabil Res Rev 1999 5 169 176 10.1002/(SICI)1098-2779(1999)5:3<169::AID-MRDD2>3.0.CO;2-K
Barkley RA Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD Psycholl Bull 1997 121 65 94 10.1037//0033-2909.121.1.65
Sonuga-Barke EJS (2002b) Psychological Heterogeneity in AD/HD; A dual pathway model of behaviour and cognition Behav Brain Res 2002 130 29 36 11864715 10.1016/S0166-4328(01)00432-6
Willcutt EG Gozal D, Molfese D The neuropsychology of ADHD: Validity of the executive function hypothesis, Attention deficit hyperactivity disorder: from genes to practice in press Humana Press: Totowa, NJ
Cairney S Maruff P Vance A Barnett R Luk E Currie J Contextual abnormalities of saccadic inhibition in children with attention deficit hyperactivity disorder Exper Brain Res 2001 141 507 518 11810144 10.1007/s002210100890
Barnett R Maruff P Vance A Luk ESL Costin J Wood C Pantelis C Abnormal executive function in AD/HD: the effect of stimulant medication and age on spatial working memory Psychol Med 2001 31 1107 1115 11513378 10.1017/S0033291701004172
Booth R Charlton R Hughes C Happe F Disentangling weak coherence and executive dysfunction: planning drawing in autism and attention-deficit/hyperactivity disorder Phil Trans Royal Soc London Series B-Biol Sci 2003 358 387 392 10.1098/rstb.2002.1204
Nigg JT Blaskey LG Huang-Pollock CL Rappley MD Neuropsychological executive functions and DSM-IV ADHD subtypes J Amer Acad Child Adoles Psychiat 2002 41 59 66 10.1097/00004583-200201000-00012
Shallice T Marzocchi GM Coser S Del Savio M Meuter RF Rumiati RI Executive function profile of children with attention deficit hyperactivity disorder Dev Neuropsychol 2002 21 43 71 12058835 10.1207/S15326942DN2101_3
Williams D Stott CM Goodyer IM Sahakian BJ Specific language impairment with or without hyperactivity: neuropsychological evidence for fronto-striatal dysfunction Dev Med Child Neurol 2000 42 368 375 10875521 10.1017/S0012162200000682
Conte R Kinsbourne M Swanson J Zirk H Samuels M Presentation rate effects on pair associate learning by attention deficit disordered children Child Dev 1986 57 681 687 3720397
Dalby JT Kinsbourne M Swanson JM Self-paced learning in children with attention deficit disorder with hyperactivity J Abnorm Child Psychol 1989 17 269 275 2666475 10.1007/BF00917398
Lawrence V Houghton S Tannock R Douglas G Durkin K Whiting K ADHD outside the laboratory: Boys' executive function performance on tasks in videogame play and on a visit to the zoo J Abnorm Child Psychol 2002 30 447 462 12403149 10.1023/A:1019812829706
Karatekin C Asarnow RF Working memory in childhood onset schizophrenia and attention deficit/hyperactivity disorder Psychiatry Research 1998 80 165 176 9754696 10.1016/S0165-1781(98)00061-4
Frank Y Seiden J Napolitano B Visual event related potentials and reaction time in normal adults, normal children, and children with attention deficit hyperactivity disorder: Differences in short-term memory processing Int J Neurosci 1996 88 109 124 9003969
Chee P Logan G Schachar R Lindsey P Wachsmuth R Effects of event rate and display time on sustained attention in hyperactive, normal and control children J Abnorm Child Psychol 1989 17 371 391 2794253 10.1007/BF00915033
Sergeant J The cognitive-energetic model: an empirical approach to Attention-Deficit Hyperactivity Disorder Neurosci Biobehav Rev 2000 24 7 12 10654654 10.1016/S0149-7634(99)00060-3
Van der Meere J Stemerdink N Gunning B Effects of presentation rate of stimuli on response-inhibition in AD/HD children with and without tics Percep Motor Skills 1995 81 259 262
Sonuga-Barke EJS Interval length and time-use by children with AD/HD: A comparison of four models J Abnorm Child Psychol 2002 30 257 264 12041711 10.1023/A:1015154829796
Weiler MD Bernstein JH Bellinger D Waber DP Information processing deficits in children with attention-deficit/hyperactivity disorder, inattentive type, and children with reading disability J Learn Disabil 2002 35 448 461 15490541
Koschack J Kunert HJ Derichs G Weniger G Irle E Impaired and enhanced attentional function in children with attention deficit/hyperactivity disorder Psychol Med 2003 33 481 489 12701668 10.1017/S0033291702007067
Landau YE Auerbach JG Gross-Tsur V Shalev RS Speed of performance of children with developmental right hemisphere syndrome and with attention-deficit hyperactivity disorder J Child Neurol 2003 18 264 268 12760429
Sonuga-Barke EJS Houlberg K Hall M When is impulsiveness not impulsiveness?; The case of children's cognitive style J Child PsycholPsychiatry 1994 35 1247 1253
World Health Organisation International classification of Diseases: Tenth Edition 1993 Chapter 5 Mental and behavioural disorders Diagnostic criteria for research Geneva: Author
Swanson J Sergeant J Taylor E Sonuga-Barke EJS Jensen P Cantwell D Attention Deficit Hyperactivity Disorder and Hyperkinetic Disorder The Lancet 1998 351 429 433 9482319 10.1016/S0140-6736(97)11450-7
Goodman R The strengths and difficulties questionnaire: A research note J Child Psychol Psychiatry 1997 38 581 586 9255702
Rutter M A childrens' behaviour questionnaire for completion by teachers: Preliminary findings J Child Psychol Psychiatry 1967 8 1 11 6033260
Goodman R Scott S Comparing the Strengths and Difficulties Questionnaire and the Child Behavior Checklist: Is small beautiful? J Abnorm Child Psychol 1999 27 17 24 10197403 10.1023/A:1022658222914
Goodman R Psychometric properties of the Strengths and Difficulties Questionnaire (SDQ) J Amer Acad Child Adoles Psychiat 2001 40 1337 10.1097/00004583-200111000-00015
Weschler D Manual for the Weschler Pre-school and Primary Scale Intelligence Test for Children, New York: The Psychological Corporation 2001
|
16033644
|
PMC1183186
|
CC BY
|
2021-01-04 16:39:21
|
no
|
Behav Brain Funct. 2005 Jul 20; 1:10
|
utf-8
|
Behav Brain Funct
| 2,005 |
10.1186/1744-9081-1-10
|
oa_comm
|
==== Front
Behav Brain FunctBehavioral and brain functions : BBF1744-9081BioMed Central London 1744-9081-1-81598241310.1186/1744-9081-1-8ResearchExecutive and motivational processes in adolescents with Attention-Deficit-Hyperactivity Disorder (ADHD) Toplak Maggie E [email protected] Umesh [email protected] Rosemary [email protected] Brain and Behaviour Research Program, Research Institute, The Hospital for Sick Children, Toronto, Canada2 York University, Toronto, Canada3 Centre for Addiction and Mental Health, Toronto, Canada4 Centre for Advanced Study at the Norwegian Academy of Science and Letters 2004–20052005 27 6 2005 1 8 8 10 3 2005 27 6 2005 Copyright © 2005 Toplak et al; licensee BioMed Central Ltd.2005Toplak 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 the current study was to examine performance and correlates of performance on a decision-making card task involving risky choices (Iowa Gambling Task) in adolescents with ADHD and comparison controls. Forty-four participants with ADHD and 34 controls were administered measures of estimated intellectual ability, working memory, and the card task. Also, behavioural ratings were obtained from parents and teachers.
Results
Adolescents with ADHD scored lower on the measures of intellectual ability, working memory, and made less advantageous selections on the card task compared to controls. Performance on measures of intellectual ability and working memory were unrelated to card task performance in both the ADHD and control samples. Parent ratings of hyperactivity/impulsivity were significantly associated with card task performance in the adolescents with ADHD, but not in controls.
Conclusion
These findings demonstrate impaired decision-making in adolescents with ADHD, and the separability of motivational and executive function processes, supporting current dual pathway models of ADHD.
==== Body
Background
Individuals with Attention-Deficit-Hyperactivity Disorder (ADHD) are more likely than their peers to make poor real-life decisions, as these individuals are described as impulsive [1], engage in more risky activities than controls [2,3], and tend to exhibit a preference for immediate rather than delayed rewards [4-7]. Much of the recent work over the last 20 years in ADHD has focused on the cognitive features of ADHD, defined as executive functions [8-10]. However, recent theories of ADHD have included both executive processes and motivational style characterized by delay aversion as two important pathways in ADHD [11,12]. The purpose of this study was to elaborate and extend this conceptualization by examining performance on a risky-choice decision-making card task, also known as the Iowa Gambling Task, in a sample of adolescents with ADHD and comparison controls. We refer to this task as the card task for simplicity. In addition, we were also interested in examining associations between performance on the card task and measures of intellectual ability and working memory.
The most recent development in the field has been an emphasis on multiple pathways for explaining impairment in ADHD [13-16,12]. Sonuga-Barke's [12] most recent model of ADHD argues for a dual pathway model of ADHD, highlighting both the executive and motivational, delay aversion aspects of ADHD. Specifically, the executive pathway involves a dysregulation of thought and action that is primarily characterized by a core deficit in inhibitory control [11]. Alternatively, the motivational pathway is hypothesized to mediate a link between behavioural symptoms, task engagement, and a biologically embedded alteration in reward mechanisms [11]. One important change reflected in this dual-pathway conceptualization is that the executive and motivational pathways are not regarded as competing theories, rather that deficits in both processes are thought to give rise to the manifestation of ADHD [12]. Sonuga-Barke [11] argues that these two pathways likely give rise to an ADHD diagnosis of the Combined subtype, and that the executive pathway is more likely associated with more severe and generalized cognitive impairment. This idea of dual pathways maps well onto clinical and research characterizations of ADHD, as the Inattentive subtype has been associated with executive dysfunction [17] and the Hyperactive/Impulsive subtype has been associated with the impulsiveness feature of ADHD [18]. Sonuga-Barke, Dalen, & Remington [19] reported that executive function and delay aversion made significant, independent contributions to ADHD symptoms in a sample of children. Similar findings have also been reported by Crone, Jennings, and van der Molen [20].
In addition to separability at a neuropsychological level, it has been hypothesized that the executive and motivational pathways are rooted in conceptually similar, but functionally segregated brain circuits [12]. It has been hypothesized that the executive pathway receives inputs from the dorsolateral prefrontal cortex to the dorsal portion of the neostriatum, as well as reciprocal connections from subcortical regions, including the dorso-medial sections of the thalamus. Alternatively, the motivational pathway centers on the reward circuits of the ventral striatum, specifically the nucleus accumbens, with connections from frontal regions, including the anterior cingulate and orbitofrontal cortex, and the amygdala. Importantly, this dual pathway model offers a theoretical account of interactions between cortical and subcortical pathways in the regulation of executive processes and motivation. The notion of separable pathways has also been theorized at a neurobiological level as well. Sagvolden et al.'s [21] dynamic developmental model of ADHD suggests that altered dopamine branches give rise to the different ADHD symptomotology, specifically, a hypofunctioning mesolimbic dopamine branch gives rise to delay aversion, a hypofunctioning mesocortical dopamine branch gives risk to poor executive functions, and a hypofunctioning nigrostriatal dopamine branch gives rise to other motor symptoms. This neurobiological account is consistent with the idea of separable pathways in ADHD, specifically the executive and delay aversion pathways. It is important to understand and test the relative contribution of these different pathways to the clinical manifestation of ADHD at both behavioural and neurobiological levels of analysis.
An important behavioural measure, called the Iowa Gambling Task (or card task), was designed to simulate the uncertainties of real-life decision-making, necessitating the weighting of rewards and penalties. In this task, participants are asked to select cards from four decks, which unbeknownst to these participants vary on expected outcome [two decks are composed of quick high gains and high losses (disadvantageous decks) or low gains and low losses (advantageous decks)] and in frequency of penalties (two decks have frequent, smaller penalties, while the other two decks have infrequent, large penalties). Expected outcome and frequency of penalties are crossed, creating four different conditions with these four decks. What has been particularly striking and important about these studies on Iowa Gambling Task performance with patients who have ventromedial cortex lesions is not simply the link between brain function and higher cognitive processes, but also the possibility that this form of reasoning may be fairly modular and localized, and separable from other cognitive abilities, such as intelligence [22].
The card task was originally used to study patients with ventromedial cortex lesions [22]. The ventromedial cortex, which may include the orbitofrontal cortex, but some have argued for subtle distinctions between these areas [23,24]. For present purposes, the orbitofrontal cortex and ventromedial areas have generally been implicated in the emotional experience associated with gains and losses in decision processes, which is to be differentiated from other frontal processes and regions, including the dorsolateral prefrontal cortex and anterior cingulate [25].
Patients with ventromedial prefrontal cortex lesions displayed less optimal performance than normal controls, making considerably more selections from the disadvantageous decks than from the advantageous decks [26]. This link between the orbitofrontal cortex and impulsivity has been described previously by others, such as, Newman et al. [18], who reported on brain lesion studies with rats which suggested an association between disinhibitory syndromes in humans (including, psychopathy, addictions, and ADHD) and lesions in the orbitofrontal cortex. Other clinical samples have reportedly displayed less optimal performance on the card task, including high school students with multiple suspensions [27], heroin addicts [28], and individuals with antisocial behaviour and psychopathic tendencies [29]. Sex differences have also been reported on card task performance, with females tending to make more selections from disadvantageous decks than males [30].
Performance on the card task has also been examined in a small group of adults with and without ADHD [31] and in a sample of adolescents with disruptive behaviour disorders, including ADHD and conduct problems [32]. In the study which included adolescents with disruptive behaviour disorders [32], adolescents with disruptive behaviour disorders displayed less optimal performance on this task than controls. In the study using adults with ADHD [31], no behavioural differences were observed on decision-making performance. However, PET scans revealed that brain activation in the ADHD group was less extended than in the control group. Specifically, control participants recruited hippocampal and insular regions more than adults with ADHD, and the adults with ADHD engaged the caudal part of the right anterior cingulated more than the controls. These results, therefore, suggest that this task activates different processes and brain regions. The behavioural significance of these differential activations is unclear at this time, and suggests that more work needs to be done on elaborating our understanding of the motivational pathway [14]. The card task and its links with patients with ventromedial lesions makes it an important task to further study and understand the motivational pathway in ADHD.
As recent models of ADHD suggest two potential pathways supporting the motivational and executive processes in ADHD [12], we must also investigate the possible associations and/or dissociations between these processes. One study investigated card task performance and working memory in normal controls, patients with lesions in the ventromedial prefrontal cortex, and patients with lesions of the dorsolatoral/mesial region of the prefrontal cortex [33]. Bechara et al. [33] reported a cognitive and anatomical double dissociation between card task performance and working memory. Others have also reported a dissociation between impulsivity and working memory in the orbitofrontal and dorsolateral regions of the prefrontal cortex [34]. However, Ernst et al. [32] found that IQ was a significant predictor of card task performance in their sample of adults with substance abuse disorders and adolescents with behaviour disorders. Hinson, Jameson, & Whitney [35] reported that increasing working memory load resulted in poorer performance on the card task. Therefore, mixed results have been reported on the behavioural associations between card task performance and working memory. These findings suggest that lesions in the ventromedial cortex may not uniquely or independently explain risky decision-making, and that contributions from other regions, such as the frontostriatal dopamine system [25], and other processes, such as working memory [35,36], may also play a role in decision-making involving gains and losses. Therefore, we also examined the relationship between card task performance and intellectual ability, and working memory in our sample of adolescents with ADHD.
The purpose of the present study was to examine how adolescents with ADHD weight risks and benefits in the card task compared to comparison controls. Impaired performance on the card task would importantly implicate ventromedial prefrontal regions for study in ADHD, which would importantly extend the motivational pathway of the dual pathway model of ADHD [11,12] by suggesting a link between subcortical structures, like the nucleus accumbens, and cortical structures, like the ventromedial cortex. In addition, associations between card task performance, intellectual ability, and working memory were also examined. A dissociation between performance on the card task and intellectual ability and working memory (our executive tasks) was predicted, which would be consistent with dual pathway models of ADHD.
Results
Data Analysis and Statistical Methods
Card Task Analysis. We conducted an analysis of covariance (ANCOVA), with outcome (advantageous vs. disadvantageous) and frequency (frequent vs. infrequent penalties) as within-subject factors and group (ADHD vs. control) as a between subject factor. Gender, FSIQ, auditory working memory, and visual-spatial working memory were examined as covariates. All posthoc analyses were a priori planned comparisons, and were conducted using the Bonferroni correction [37]. Then, correlational analyses were conducted separately and together in the ADHD and control samples to examine whether estimated FSIQ, or working memory were significantly correlated with card task performance; this provided a further, converging examination of the relationship between card task performance, FSIQ, and working memory. A repeated measures analysis was conducted to examine learning across trials on the card task. An ANCOVA analysis was conducted to examine the impact of subtype in the ADHD sample, covarying for gender. We also examined correlations between card task performance and behavioural ratings by parents and teachers.
Group Differences on Standardized Measures and Behaviour Ratings
Table 1 displays the standardized measures and clinical characteristics of the adolescents with ADHD and comparison controls. Overall, adolescents with ADHD displayed significantly lower scores, albeit in the normal range, on the measures of FSIQ, and auditory and visual-spatial working memory than comparison controls.
Table 1 Diagnostic Characteristics of the ADHD and Comparison Control Groups
ADHD n Controls n F (1, 77)
Age 15.6 (1.4) 44 15.4 (1.5) 34 0.24
Standardized Measures
Estimated VIQ 105.5 (10.6) 44 109.9 (10.0) 34 3.48
Estimated PIQ 101.6 (11.2) 44 106.4 (11.3) 34 3.45
Estimated FSIQ Score 104.1 (10.1) 44 109.4 (10.2) 34 5.20*
Auditory Working Memory Scaled Score 9.6 (3.5) 44 11.9 (2.5) 34 10.35**
Visual-Spatial Working Memory Scaled Score 9.2 (3.4) 44 11.3 (2.7) 34 8.41**
Diagnostic Characteristics from Conners Ratings
Parent
Inattention T-score 73.2 (10.1) 43 49.3 (6.8) 32 134.25***
Hyperactivity/Impulsivity T-score 72.7 (12.0) 43 51.6 (7.8) 32 74.85***
Teacher
Inattention T-score 70.9 (14.8) 37 49.7 (9.9) 28 42.97***
Hyperactivity/Impulsivity T-score 65.4 (16.3) 37 51.1 (9.2) 28 17.29***
*** p < .001, **p < .01, *p < .05
Group Differences on Card Task
The means and standard deviations for all 100 trials of the card task for each deck are presented in Table 3. Using an analysis of covariance (ANCOVA), we found that outcome was significant, F(1, 76) = 12.45, p = .001, indicating that more cards were selected overall from the disadvantageous than the advantageous decks by all participants. Frequency was also significant, F(1, 76) = 33.94, p = .0001, indicating that all participants selected more cards from the infrequent penalty decks than the frequent penalty decks. The outcome by group and frequency by group interactions were non-significant. The outcome by frequency interaction was significant, F(1, 76) = 4.47, p = .038, indicating that most cards were selected from deck B, p = .0001. The outcome by frequency by group interaction was significant, F(1, 76) = 5.76, p = .021. Posthoc analyses indicated that participants with ADHD selected more cards from deck B, p = .045, and significantly fewer cards from deck D, p = .018, than comparison controls. The group main effect was not significant. Sex, FSIQ, auditory working memory, and visual-spatial working memory did not enter as significant covariates.
Table 3 Mean scores of the adolescents with ADHD and comparison controls for the last 100 selections on the card task (standard deviations in parentheses)
Deck ADHD n Control n F(1,77) Cohen's d
Deck A – Disadvantageous 23.3 (9.7) 44 24.4 (7.3) 34 0.34 -0.13
Deck B – Disadvantageous 34.1 (10.0) 44 29.4 (10.4) 34 4.17* 0.46
Deck C – Advantageous 19.9 (8.3) 44 20.3 (8.8) 34 0.06 -0.05
Deck D – Advantageous 21.6 (7.0) 44 26.9 (12.2) 34 5.81* -0.53
Monetary Outcome -10.0 (7.0) 44 -6.4 (11.4) 34 2.99† -0.38
* p < .05, † p < .10
Correlational analyses with decks B and D were also conducted for a convergent analysis. Correlational analyses were conducted separately within the ADHD and control groups, as it is important to determine whether these variables co-vary differently in the ADHD and control groups. There were no significant associations between deck selections on the card task and intellectual ability or working memory in the ADHD or control samples. When this same analysis was conducted within the entire sample, the same results were obtained. Effect sizes in Table 3 indicate moderate effects for selections from decks B and D. Participants with ADHD also tended to lose more money than controls, an effect which was marginally significant between groups.
Figures 1 and 2 display a visual breakdown of the mean number of cards that were selected from each deck at every ten card interval. The first 50 trials for both ADHD and control groups do not suggest any pattern in card selections, suggesting a sampling and learning phase in both groups. As evidenced in the group differences, the trends differ for decks B and D in the ADHD and control groups for the last 50 card selections. Clearly, the ADHD group continues to select more cards from deck B right until the end of the game, whereas controls display a clear preference for selecting cards from deck D. Selections from deck C slightly mirror those of deck D, but there were obviously no differences evident in the selections from deck A. We conducted analyses of the last 50 card selections for decks B and D. Using a 2 (group: ADHD vs. controls) X 5 (Block: B51-60, B61-70, B71-80, B81-90, B91-100) repeated measures MANOVA to examine the last 50 card selections from deck B, we obtained a significant group by block interaction, F(1, 76), 4.33, p = .041, indicating that adolescents with ADHD picked progressively more cards from deck B than controls. When this same analysis was conducted with deck D, the interaction did not reach significance.
Figure 1 Pattern of Selections on the Card Task from Decks A and B.
Figure 2 Pattern of Selections on the Card Task from Decks C and D.
The open-ended question about the content of the decks was coded for whether participants recognized that some decks were more advantageous than others. It was found that 58% (n = 23) and 86% (n = 36) of the adolescents with ADHD recognized that decks A and B were disadvantageous, and 89% (n = 34) and 81% (n = 34) recognized that decks C and D were advantageous. Of the controls, 79% (n = 23) and 91% (n = 30) recognized that decks A and B were disadvantageous, and 83% (n = 20) and 87% (n = 27) recognized that decks C and D were advantageous. One participant with ADHD (2 %) ended the game with a net gain, and 10 control participants (29 %) ended the game with a net gain. Overall, these data indicate that participants in both groups generally understood the task and recognized the advantageous and disadvantageous decks. One exception to this is that only 58% of the adolescents with ADHD recognized that A was a disadvantageous deck, supporting the idea that the content of this deck was more difficult to track.
Subtype Effects and Correlational Analyses Between The Card Task and Behavioural Measures
Twenty-seven of our adolescents with ADHD met criteria for the Combined subtype, and 17 met criteria for the Inattentive subtype. We examined the effect of subtype on card task performance within the ADHD group using a multivariate ANOVA, with subtype (Inattentive vs. Combined) as a between-subject factor and outcome (advantageous vs. disadvantageous decks) and frequency (frequent vs. infrequent penalties) as within-subject factors. The subtype by frequency interaction was significant, F(1, 42) = 4.41, p = .042. Posthoc analyses indicated that participants with ADHD of the Combined subtype, M = 57.9, SD = 9.1, tended to select more cards from decks B and D than participants with ADHD of the Inattentive subtype, M = 52.2, SD = 8.3, p = .043. Then, participants with ADHD of the Inattentive subtype, M = 47.8, SD = 8.3, tended to select more cards from decks A and C than participants with ADHD of the Combined subtype, M = 42.0, SD = 9.1, p = .041. We examined sex as a covariate of subtype, but no significant associations were obtained. Five of the females with ADHD met criteria for the Combined subtype, whereas only one met criteria for the Inattentive subtype.
With the behavioural measures, two associations between hyperactive/impulsive symptoms on the K-SADS-PL (that is, DSM-IV-TR symptoms), r = .31, p = .05, and Conners' parent rating of hyperactivity/impulsivity, r = .31, p = .04, correlated significantly with the total number of card selections from disadvantageous deck B in the ADHD sample, however, there were no such associations within the control sample. The associations between inattentive symptoms and deck selections did not reach significance in the ADHD or control samples. These correlations are displayed in Table 4. When the same correlations were conducted with the entire sample, the significant association between deck B selections and parent report of hyperactivity/impulsivity on the Conners'scales remained, r = .25, p = .031.
Table 4 Correlations between Card Task Selections and Behaviour Rating Measures in Adolescents with ADHD and Controls
Deck B (100 Trials) Deck D (100 Trials)
Behaviour Rating Measures
K-SADS – Number of Inattentive Symptoms Control (n = 0) -- --
ADHD (n = 42) -.22 .23
K-SADS – Number of Hyperactive/Impulsive Symptoms Control (n = 0) -- --
ADHD (n = 44) .31* .02
Conners Parent Inattention T-Score Control (n = 34) .26 .00
ADHD (n = 43) -.10 .29
Conners Parent Hyperactivity/Impulsivity T-Score Control (n = 34) -.22 .12
ADHD (n = 43) .31* .26
Conners Teacher Inattention T-Score Control (n = 28) -.07 .28
ADHD (n = 37) -.28 -.06
Conners Teacher Hyperactivity/Impulsivity T-Score Control (n = 28) -.13 .09
ADHD (n = 37) -.11 .14
*p < .05
Discussion
The results of the present study indicated that the adolescents with ADHD made less optimal selections on the card task than controls by selecting more cards from the disadvantageous deck B and fewer cards from the advantageous deck D than controls. Adolescents with ADHD also scored lower on estimated FSIQ, and auditory and visual-spatial working memory than controls, but these variables were not significant covariates of card task performance. In the correlational analyses with the behavioural ratings, it was parent report of hyperactivity and impulsivity, not estimated FSIQ or working memory, which was associated with performance on the card task in the adolescents with ADHD. Sex did not enter as a significant covariate of card task performance.
Consistent with previous studies, our participants with ADHD displayed a pattern of performance similar to adolescents with behaviour disorders [32]. In particular, the group differences in the current study emerged on only two of the decks; importantly, these two particular decks (decks B and D) necessitated less tracking of expected value as penalties were less frequent. This finding suggests that the frequency of dispensing rewards and penalties should be considered as a potential variable in this task, and must be examined systematically. It may be the case that requiring participants to track more carefully may result in the recruitment of executive processes, such as working memory [35]. Overall, the group differences on the card task highlight poor behaviour regulation and impulsivity as critical features in the profile of ADHD, which have both been described as pervasive features of ADHD [1].
In addition, we did not find any association between selections on the card task and on our measures of estimated FSIQ or working memory, suggesting that these processes are behaviourally separable. While our adolescents with ADHD displayed lower estimated FSIQ and working memory scores than controls, our ANCOVA and correlational analyses suggest that their less optimal performance on the card task is not attributable to limitations in FSIQ or working memory, which is consistent with the findings of Bechara et al. [33] and Berlin et al. [34], but not with Ernst et al. [32] or Hinson et al. [35]. Importantly, Ernst et al. [32] used a more heterogeneous sample than we used, including adolescents with conduct problems and adults with substance abuse problems. Then, as Hinson et al. [35] increased memory load on the card task, it would be expected that further taxing of working memory during administration of the card task would negatively affect performance. Notably performance on both estimated FSIQ and working memory in our study were not associated or correlated with card task performance. Cognitive scientists understand tests of cognitive and intellectual ability as general indicators of cognitive efficiency [44,47-49], which is consistent with the general concept of executive functions [8]. Indeed, correlations between estimated FSIQ and the working memory measures within the ADHD and control samples ranged from r = .35, p = .045 to r = .57, p = .0001. Therefore, estimated FSIQ and our working memory measures provide a reasonable index of executive, dorsolateral types of processes.
These findings are consistent with separable pathway models of ADHD [15,11,19,21], which includes the separable contributions of both executive processes and behavioural regulation variables. The motivational pathway highlights the impulsive tendencies [1], delay aversion [5-7], and sensitivity to rewards and response costs [50-52] in ADHD. In addition, the current study also extends the conceptualization of the motivational pathway to include cortically-based processes in ADHD, particularly the ventromedial prefrontal cortex [19], suggesting that more research is needed to better parse how frontal and subcortical mechanisms function and interact in the clinical manifestation of ADHD. These findings are also consistent with cognitive developmental models of normative development which include the separable contributions of "hot" affective decision-making versus "cool" executive function processes [53].
Studies have consistently replicated the role of executive function deficits in ADHD [8-10], however, further elaboration and understanding are needed at a behavioural level of analysis for what has been termed as the motivational pathway in ADHD. Drawing from the work by Damasio's interpretation of card task performance [54-56], one possibility worthy of further investigation for less optimal card selections in the ADHD group is dysregulation of somatic markers. It may be the case that individuals with ADHD have weaker somatic or physiological cues to guide risky choices, which would be consistent with Damasio's [54-56] somatic marker hypothesis. Somatic markers, or emotions, assist by constraining the decision-making space, giving various alternatives preferential availability over other alternatives [57], and serve an adaptive evolutionary human function, consistent with cognitive science perspectives on the role of emotion [58]. Damasio [54] argues that his patients with ventromedial lesions lack the physiological cues needed to signal risky choices, as evidenced by skin conductance studies performed in his lab [26,59]. At least two studies have been conducted investigating the physiological reactions of children with ADHD and controls in the presence of reward and extinction conditions [60]. It was reported that children with ADHD displayed a faster heart rate habituation to reward and less of a galvanic skin response during habituation than controls. A study by Crone, Jennings, and van der Molen [20] found that children with ADHD had lower heart rate responses to immediate reward feedback than comparison controls, but no group differences were observed in skin conductance responses. These studies suggest that individuals with ADHD may experience a different physiological reaction than controls in the presence of rewards, suggesting that these cues may give way to different somatic markers, affecting decision-making in these individuals. This is a viable hypothesis in individuals with ADHD that deserves further study and consideration for understanding the motivational pathway. Notably, any difficulties in the physiological and/or affective regulation are likely to be more subtle in individuals with ADHD than in patients with ventromedial lesions, which is one reason why the statistical analyses may not have yielded such strong findings. Other limitations of the present study include sample heterogeneity, the gender imbalance between groups, and a lack of a full psychiatric diagnostic assessment in the control participants. Therefore, future research and methodologies should take these variables into account.
Another critical finding in this study was that card task performance was correlated with the behavioural and diagnostic symptoms of hyperactivity and impulsivity, but not with inattention, estimated FSIQ, or working memory. This correlation is consistent with Dinn, Robbins, and Harris [61], who found that adults with ADHD – Combined Subtype performed worse on orbitofrontal tasks, whereas adults with ADHD-Inattentive subtype performed worse on dorsolateral tasks. Motivational regulation difficulties seem to be associated with the hyperactive and impulsive features of ADHD [18], and the executive function deficits are associated with the inattentive features of ADHD [17], which is consistent with the current conceptualization. While subtype did not differentiate performance on advantageous or disadvantageous card selections, this may have been due to the fact that we did not have any participants with only the Hyperactive/Impulsive subtype. The next step will be to understand the relationships between other cognitive processes thought to be deficient in ADHD. For example, the relationship between inhibitory control [9], other components of inhibition [62], and time perception [63], with other executive and motivational processes as defined in the dual pathway model [12], and to examine how they interact and give rise to the clinical presentation of ADHD.
Conclusion
The results of the current study demonstrated that adolescents with ADHD displayed impaired performance on the Iowa Gambling card task as well as on measures of intellectual ability and working memory in comparison to adolescent controls. Notably, performance on the card task and the measures of intellectual ability and working memory were not associated in the ADHD and control samples. Non-optimal card selections were associated with hyperactive and impulsive symptoms in the adolescents with ADHD. These findings provide support for the separable motivational and executive pathways in current models of ADHD [11,12]. Further research must elaborate what processes constitute the motivational pathway in ADHD, and how these processes may interact with executive processes and give rise to the clinical presentation of ADHD.
Method
Participants
Two groups of adolescents participated: 44 adolescents (86% male) with a confirmed clinical diagnosis of ADHD based on DSM-IV criteria and 34 comparison adolescents (41% male). Of our ADHD sample, 45% (n = 20) met criteria for Predominantly Inattentive subtype, and 55% (n = 24) met criteria for Predominantly Combined subtype. All adolescents were between the ages of 13 and 18 years of age (M = 15.5; SD = 1.5). Adolescents with ADHD were recruited from the YEARS (Youth, Education, and Assessment Research Service) Program. Adolescents in the control comparison group were recruited through hospital staff and community resources. All adolescents participating in the study were native English speakers. Adolescents were excluded if they had evidence of psychosis, pervasive developmental disorder, a serious medical condition, or an estimated Full-Scale IQ (FSIQ) below 80.
ADHD Sample. All adolescents had a DSM-IV diagnosis of ADHD confirmed by a systematic and comprehensive clinical diagnostic assessment conducted at the time of the study. The assessment comprised a semi-structured clinical diagnostic interview [Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version; K-SADS-PL; [38]]. Also, parents and teachers completed the Conners' Rating Scales-Revised [39] to obtain standardized ratings of behaviour. Diagnosis of ADHD in adolescents was based on the following algorithm: 1) met DSM-IV criteria according to the clinician summary based on the K-SADS-PL interviews; and 2) met the clinical cut-offs for inattentive or hyperactive/impulsive symptoms on the Conners teacher questionnaires (t-score > 70) in order to ensure pervasiveness of symptoms across settings. The K-SADS was conducted separately with the adolescent and parent, and the clinician summarized the information from both informants. The K-SADS interview was the primary source for diagnosis, and the Conners' scales did not always reach the threshold t-score, particularly on the teacher reports. If parents reported a history of ADHD symptoms (both at home and from school reports) and evidence of pervasiveness across settings during the K-SADS interview, this information was adequate for a diagnosis of ADHD.
Subtypes were determined by counting the number of inattentive and/or hyperactive/impulsive symptoms on the K-SADS, and comparing with the Conners scales for convergence. Participants needed a total of six symptoms of inattention endorsed to be identified as the Inattentive subtype, six symptoms of hyperactivity or impulsivity to be identified as the Hyperactive/Impulsive subtype, or six of each to be identified as the Combined subtype. A reading or math learning disability was defined as a score below the 25th percentile on a measure of reading or math achievement. Many of the adolescents with ADHD had comorbid disorders: 16 (36%) had a Learning Disability, 12 (27%) had Oppositional Defiant Disorder, four had an Anxiety Disorder, three had Depression, and one had Conduct Disorder. Diagnostic characteristics of the sample are presented in Table 1.
Fourteen of our participants used stimulant medication (30%), eight had previously used stimulant medication (17%), five used a non-stimulant medication, and 18 had never used psychoactive medication (43%). All adolescents were asked to stop taking any medication for six half-lives prior to assessment, except for the two participants who used antidepressants.
Comparison control sample. The Conners' questionnaires were given to parents and adolescents to screen for any mental health concerns. Any participants who obtained scores above a t-score of 60 were interviewed further with a complete K-SADS-PL interview to rule out any diagnosis of ADHD. A total of four K-SADS-PL were conducted to follow up on issues raised on the Conners questionnaires in the comparison control group. Otherwise, a psychiatric assessment with control participants was not completed.
Standardized Tasks
Intellectual Ability. The Wechsler Abbreviated Scale of Intelligence [WASI; [40]] comprised four subtests (Vocabulary, Block Design, Similarities, and Matrices), and was used to provide an estimate of verbal (VIQ), nonverbal (PIQ), and Full-Scale (FSIQ) intellectual ability.
Auditory and Visual-Spatial Working Memory. We used two different tasks to measure memory performance, one auditory-verbal and one visual-spatial. Our measure of auditory memory was the Digit Span subtest (Forwards and Backwards components) from the Wechsler Intelligence Scale for Children – Third Edition [WISC-III; [40]] and the Wechsler Adult Intelligence Scale – Third Edition [WAIS-III; [42]]. Our measure of visual-spatial memory was the Spatial Span subtest (Forwards and Backwards components) from the WISC-III Processing Instrument [WISC-III-PI, [43]]. The combination of the forwards and backwards components of each memory measure provide a composite of maintenance and working memory [44]; for simplicity, we have termed these as our working memory measures.
Experimental Task
Card Task. This task was designed after the Iowa gambling task by Bechara et al. [26]. The materials required for this task included: four decks of cards designed by the researcher, monopoly play money, and a hand counter to track the number of trials during the task. Four decks of cards were placed in front of the participant, labeled as deck A, B, C, or D. On the back of each card, there was either a reward or a reward and penalty indicated. Each deck varied on expected outcome [two decks are composed of quick high gains and high losses (disadvantageous decks) or low gains and low losses (advantageous decks)] and in frequency of penalties (two decks have frequent, smaller penalties, while the other two decks have infrequent, large penalties). Expected outcome and frequency of penalties are crossed, creating four different conditions with these four decks. Therefore, deck A was a disadvantageous deck with frequent, lower penalties, deck B was a disadvantageous deck with infrequent, high penalties, deck C was an advantageous deck with frequent, lower penalties, and deck D was an advantageous deck with infrequent, high penalties. The cards were organized in the exact same order for each and every participant. There were 50 cards in each deck. The reward-penalty schedules that were used for this task are displayed in Table 2. For decks A and B, participants received $1.00 with each card selection AND the penalty indicated for each respective deck in Table 2, and for decks C and D, participants received $0.50 with each card selection AND the penalty indicated for each respective deck in Table 2. Bechara et al. [26] used rewards and penalties with a base of 10, whereas we used a base of .10, however, the relative size of the rewards and penalties were the same as those used by Bechara et al. [26].
Table 2 The schedule of rewards and penalties in the four decks of the card task
Card Number Deck A (+$1.00)‡ Deck B (+$1.00)‡ Deck C (+$.50)‡ Deck D (+$.50)‡
1
2
3 -$1.50 -$.25
4
5 -$3.00 -$.75
6
7 -$2.00 -$.25
8
9 -$2.50 -$12.50 -$.75
10 -$3.50 -$.50 -$2.50
11
12 -$3.50 -$.25
13 -$.75
14 -$2.50 -$12.50
15 -$2.00 -$2.50
16
17 -$3.00 -$.25
18 -$1.50 -$.75
19
20 -$.50
21 -$12.50 -$2.50
22 -$3.00
23
24 -$3.50 -$.50
25 -$.25
26 -$2.00 -$.50
27 -$2.50
28 -$1.50
29 -$.75
30 -$.50
31 -$3.50
32 -$2.50 -$12.50 -$2.50
33 -$2.50
34 -$.25
35 -$.25
36
37 -$1.50 -$.75
38 -$3.00
39 -$.50
40 -$.25
41 -$12.50 -$.50 -$2.50
42 -$3.00
43
44 -$3.50 -$.50
45 -$.25
46 -$2.00 -$.50
47 -$2.50
48 -$1.50
49 -$.75
50 -$.50
‡ Note: Only the penalty amounts varied in each deck, which are indicated in the Table. The reward amounts were consistent for each deck, which were $1.00 for decks A and B and $.50 for decks C and D with each card selection
Participants were told that they would play a card game in which they had to select 100 cards, and that the purpose of the game was to maximize the amount of money they could win. They were told that on the back of each card in each deck, there was a reward or a reward and a penalty. They were to select cards one by one, and the examiner would give them the reward and collect the penalty after each card pull. Participants were loaned $20.00 at the beginning of the game. Participants were not given any cues or signals during the game about the content of each deck. Participant selections were recorded on a score sheet following the task. As play money was used, participants were told that they could win a $10 gift certificate to a popular local bookstore if they made a net gain in the game as an added incentive, as participants did not receive the actual amount of play money in the game.
The dependent measures were the total number of cards participants selected from each deck and the net amount of money participants had at the end of the game. At the end of the game, participants were asked about how the four decks differed to determine whether the participants realized the reward and penalty structure of the advantageous and disadvantageous decks. Specifically, they were asked: "Did you notice anything about the content of each of the decks?"
Authors' Contributions
MT was involved in the theoretical conceptualization, data collection and analysis, and writing of the present manuscript. UJ was involved in the clinical assessment of the participants and feedback on the manuscript. RT was involved in the theoretical conceptualization, and provided ongoing input and feedback on the analysis and writing of this manuscript.
Acknowledgements
This research was supported in part by research grants from the National Institute of Mental Health (Grant R21-MH066393), HSC Psychiatry Endowment Fund, and a Canadian Institutes of Health Research (CIHR) Fellowship granted to M. Toplak. This article is part of the international and interdisciplinary project "ADHD: From genes to therapy" (Project Leader: Terje Sagvolden) conducted at the Centre for Advanced Study (CAS) in Oslo, Norway (2004–2005), in which RT participated.
We thank Rebecca Stewart for assistance with data collection and scoring. We also thank Ruth Barton at CAMH for clinical support, Karen Ghelani for assistance with interviewing, Colleen Dockstader for comments on the manuscript, and Paul Voorn for assistance with references.
==== Refs
American Psychiatric Association Diagnostic and statistical manual of mental disorders, Text Revision (DSM-IV-TR) 2000 4 Washington, DC: American Psychiatric Association
Barkley RA ADHD and the nature of self-control 1997 New York: Guilford Press
Barkley RA Attention-deficit hyperactivity disorder 1998 2 New York: The Guilford Press
Barkley RA Edwards G Laneri M Fletcher K Metevia L Executive functioning, temporal discounting, and sense of time in adolescents with Attention Deficit Hyperactivity Disorder (ADHD) and Oppositional Defiant Disorder (ODD) J Abnorm Child Psychol 2001 29 541 556 11761287 10.1023/A:1012233310098
Sonuga-Barke EJS Taylor E Heptinstall E Hyperactivity and delay aversion – II. The effect of self versus externally imposed stimulus presentation periods on memory J Child Psychol Psychiatry 1992 33 399 409 1564082
Sonuga-Barke EJS Taylor E Sembi S Smith J Hyperactivity and delay version – I. The effect of delay on choice J Child Psychol Psychiatry 1992 33 387 398 1564081
Sonuga-Barke EJS Williams E Hall M Saxton T Hyperactivity and delay aversion – III. The effect on cognitive style of imposing delay after errors J Child Psychol Psychiatry 1996 37 189 194 8682898
Pennington BF Ozonoff S Executive functions and developmental psychopathology J Child Psychol Psychiatry 1996 37 51 87 8655658
Schachar R Mota VL Logan GD Tannock R Klim P Confirmation of an inhibitory control deficit in attention-deficit/hyperactivity disorder J Abnorm Child Psychol 2000 28 227 235 10885681 10.1023/A:1005140103162
Sergeant JA Geurts H Oosterlaan J How specific is a deficit of executive functioning for Attention-Deficit/Hyperactivity Disorder? Behav Brain Res 2002 130 3 28 11864714 10.1016/S0166-4328(01)00430-2
Sonuga-Barke EJS Psychological heterogeneity in AD/HD – A dual pathway model of behaviour and cognition Behav Brain Res 2002 130 29 36 11864715 10.1016/S0166-4328(01)00432-6
Sonuga-Barke EJS The dual pathway model of AD/HD: An elaboration of neurodevelopmental characteristics Neurosci Biobehav Rev 2003 27 593 604 14624804 10.1016/j.neubiorev.2003.08.005
Castellanos FX Tannock R Neuroscience of Attention-Deficit/Hyperactivity Disorder: The search for endophenotypes Nat Rev Neurosci 2002 3 1 12 10.1038/nrn721
Hinshaw SP Impulsivity, emotion regulation, and developmental psychopathology: Specificity versus generality of linkages Ann N Y Acad Sci 2003 1008 149 159 14998881 10.1196/annals.1301.016
Nigg JT Is ADHD a disinhibitory disorder? Psychol Bull 2001 127 571 598 11548968 10.1037//0033-2909.127.5.571
Nigg JT Response inhibition and disruptive behaviors: Toward a multiprocess conception of etiological heterogeneity for ADHD combined type and conduct disorder early-onset type Ann N Y Acad Sci 2003 1008 170 182 14998883 10.1196/annals.1301.018
Chhabildas N Pennington BF Willcutt EG A comparison of the neuropsychological profiles of the DSM-IV subtypes of ADHD J Abnorm Child Psychol 2001 29 529 540 11761286 10.1023/A:1012281226028
Newman JP Gorenstein EE Kelsey JE Failure to delay gratification following septal lesions in rats: Implications for an animal model of disinhibitory psychopathology Pers Individ Dif 1983 4 147 156 10.1016/0191-8869(83)90014-4
Sonuga-Barke EJS Dalen L Remington B Do executive deficits and delay aversion make independent contributions to preschool attention-deficit/hyperactivity disorder symptoms? J Am Acad Child Psychiatry 2003 42 1335 1342 10.1097/01.chi.0000087564.34977.21
Crone EA Jennings R van der Molen MW Sensitivity to interference and response contingencies in Attention-Deficit/Hyperactivity Disorder J Child Psychol Psychiatry 2003 44 214 226 12587858 10.1111/1469-7610.00115
Sagvolden T Johansen EB Aase H Russell VA A dynamic developmental theory of Attention-Deficit/Hyperactivity Disorder (ADHD) predominantly hyperactive/impulsive and combined subtypes Behav Brain Sci 2005
Bechara A Damasio AR Damasio H Anderson SW Insensitivity to future consequences following damage to human prefrontal cortex Cogn 1994 50 7 15 10.1016/0010-0277(94)90018-3
Bechara A Tranel D Damasio H Characterization of the decision making deficit of patients with ventromedial prefrontal cortex lesions Brain 2000 123 1289 2202 10.1093/brain/123.11.2189
Rolls ET The orbitofrontal cortex and reward Cereb Cortex 2000 20 284 294 10731223 10.1093/cercor/10.3.284
Krawczyk DC Contributions of the prefrontal cortex to the neural basis of human decision making Neurosci Biobehav Rev 2002 26 631 664 12479840 10.1016/S0149-7634(02)00021-0
Bechara A Tranel D Damasio H Damasio AR Failure to respond automatically to anticipated future outcomes following damage to prefrontal cortex Cereb Cortex 1996 6 215 225 8670652
Stanovich KE Grunewald M West RF Cost-benefit reasoning in students with multiple secondary school suspensions Pers Individ Dif 2003 35 1061 1072 10.1016/S0191-8869(02)00318-5
Petry NM Bickel WK Arnett M Shortened time horizons and insensitivity to future consequences in heroin addicts Addiction 1998 93 729 738 9692271 10.1046/j.1360-0443.1998.9357298.x
Blair RJR Colledge E Mitchell DGV Somatic markers and response reversal: Is there orbitofrontal cortex dysfunction in Boys with psychopathic tendencies? J Abnorm Child Psychol 2001 29 499 511 11761284 10.1023/A:1012277125119
Overman WH Sex differences in early childhood, adolescence, and adulthood on cognitive tasks that rely on orbital prefrontal cortex Brain Cogn 55 134 147 15134848 10.1016/S0278-2626(03)00279-3
Ernst M Kimes AS London ED Matochik JA Eldreth D Tata S Contoreggi C Leff M Bolla K Neural substrates of decision making in adults with Attention Deficit Hyperactivity Disorder Am J Psychiatry 2003 160 1061 1070 12777263 10.1176/appi.ajp.160.6.1061
Ernst M Grant SJ London ED Contoreggi CS Kimes AS Spurgeon L Decision making in adolescents with behavior disorders and adults with substance abuse Am J Psychiatry 2003 160 33 40 12505799 10.1176/appi.ajp.160.1.33
Bechara A Damasio H Tranel D Anderson SW Dissociation of working memory from decision-making within the human prefrontal cortex J Neurosci 1998 18 428 437 9412519
Berlin HA Rolls ET Kischka U Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions Brain 2004 127 1108 1126 14985269 10.1093/brain/awh135
Hinson JM Jameson TL Whitney P Somatic markers, working memory, and decision-making Cogn Affect Behav Neurosci 2002 2 341 353 12641178
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
Tabachnick BG Fidell LS Using Multivariate Statistics 2001 4 Boston: Allyn and Bacon
Kaufman J Birmaher B Brent D Schedule for affective disorders and schizophrenia for school-age children – Present and Lifetime Version (K-SADS-PL): Initial reliability and validity data J Am Acad Child Psychiatry 1997 36 980 988 10.1097/00004583-199707000-00021
Conners CK Conners' Rating Scales-Revised Technical Manual 1997 New York: Multi-Health Systems Inc
Wechsler D Wechsler Abbreviated Scale of Intelligence (WASI) 1999 USA: The Psychological Corporation
Wechsler D Wechsler Intelligence Scale for Children 1991 Third USA: The Psychological Corporation
Wechsler D Wechsler Adult Intelligence Scale 1997 Third USA: The Psychological Corporation
Kaplan E Fein D Kramer J Delis D Morris R WISC-III-PI Manual 1999 U.S.A: The Psychological Corporation
Engle RW Working memory capacity as executive attention Curr Dir Psychol Sci 2002 11 19 23 10.1111/1467-8721.00160
Cohen J A power primer Psychol Bull 1992 112 155 159 10.1037//0033-2909.112.1.155
Mintzer MZ Stitzer ML Cognitive impairment in methadone maintenance patients Drug and Alcoh Depend 2002 67 41 51 10.1016/S0376-8716(02)00013-3
Pinker S How the mind works 1997 New York: Norton & Company
Stanovich KE Who is rational? Studies of individual differences in reasoning 1999 New Jersey: Lawrence Erlbaum Associates
Stanovich KE West RF Individual differences in reasoning: Implications for the rationality debate? Behav Brain Sci 2000 23 645 665 11301544 10.1017/S0140525X00003435
Douglas VI Bloomingdale L The response of ADD children to reinforcement: Theoretical and clinical implications Attention deficit disorder: Identification, course and treatment rationale 1985 New York: Spectrum Press 49 66
Douglas VI Bloomingdale LM, Sergeant J Cognitive deficits in children with attention deficit disorder with hyperactivity Attention deficit disorder: Criteria, cognition, and intervention 1988 Elmsford, NY: Pergamon Press 65 81
Iaboni F Douglas VI Baker AG Effects of reward and response costs on inhibition in ADHD children J Abnorm Psychol 1995 104 232 240 7897047 10.1037//0021-843X.104.1.232
Kerr A Zelazo PD Development of "hot" executive function: The children's gambling task Brain and Cogn 2004 55 148 157 10.1016/S0278-2626(03)00275-6
Damasio AR Descartes' error: Emotion, reason, and the human brain 1994 New York: Avon Books
Damasio AR The somatic marker hypothesis and the possible functions of the prefrontal cortex Philos Trans R Soc Lond B Biol Sci 1996 B-351 1413 1420 8941953
Damasio A The feeling of what happens: Body and emotion in the making of consciousness 1999 San Diego: Harcourt, Inc
Oatley K Wilson RA, Keil FC Emotions The MIT encyclopedia of the cognitive sciences 1999 Cambridge, MA: The MIT Press
Johnson-Laird P Oatley K Basic emotions, rationality, and folk theory Cogn Emotion 1992 6 201 223
Bechara A Damasio H Tranel D Damasio AR Deciding advantageously before knowing the advantageous strategy Science 1997 272 1293 1295 9036851 10.1126/science.275.5304.1293
Iaboni F Douglas VI Ditto B Psychophysiological response of ADHD children to reward and extinction Psychophysiology 1997 34 116 123 9009815
Dinn WM Robbins NC Harris CL Adult attention-deficit/hyperactivity disorder: Neuropsychological correlates and clinical presentation Brain Cogn 2001 46 114 121 11527308
Friedman NP Miyake A The relations among inhibition and interference control functions: A latent-variable analysis J Exp Psychol: Gen 2004 133 101 135 14979754 10.1037/0096-3445.133.1.101
Toplak ME Rucklidge JJ Hetherington R John SCF Tannock R Time perception deficits in Attention-Deficit/Hyperactivity Disorder and Comorbid Reading Difficulties in child and adolescent samples J Child Psychol Psychiatry 2003 44 888 903 12959497 10.1111/1469-7610.00173
|
15982413
|
PMC1183187
|
CC BY
|
2021-01-04 16:39:21
|
no
|
Behav Brain Funct. 2005 Jun 27; 1:8
|
utf-8
|
Behav Brain Funct
| 2,005 |
10.1186/1744-9081-1-8
|
oa_comm
|
==== Front
BMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 1472-6882-5-141596976810.1186/1472-6882-5-14Research ArticleAntihyperlipidemic and antiperoxidative effect of Diasulin, a polyherbal formulation in alloxan induced hyperglycemic rats Saravanan Ramalingam [email protected] Leelavinothan [email protected] Department of Biochemistry, Faculty of Science, Annamalai University, Annamalai Nagar, Tamil Nadu – 608 002, India2005 22 6 2005 5 14 14 10 12 2004 22 6 2005 Copyright © 2005 Saravanan and Pari; licensee BioMed Central Ltd.2005Saravanan and Pari; 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 was undertaken to investigation the effect of Diasulin, a poly herbal drug composed of ethanolic extract of ten medicinal plants on blood glucose, plasma insulin, tissue lipid profile, and lipidperoxidation in alloxan induced diabetes.
Methods
Ethanolic extract of Diasulin a, poly herbal drug was administered orally (200 mg/kg body weight) for 30 days. The different doses of Diasulin on blood glucose and plasma insulin in diabetic rats were studied and the levels of lipid peroxides [TBARS, and Hydroperoxide] and tissue lipids [cholesterol, triglyceride, phospholipides and free fatty acids] were also estimated in alloxan induced diabetic rats. The effects were compared with glibenclamide.
Result
Treatment with Diasulin and glibenclamide resulted in a significant reduction of blood glucose and increase in plasma insulin. Diasulin also resulted in a significant decrease in tissue lipids and lipid peroxide formation. The effect produced by Diasulin was comparable with that of glibenclamide.
Conclusion
The decreased lipid peroxides and tissue lipids clearly showed the antihyperlipidemic and antiperoxidative effect of Diasulin apart from its antidiabetic effect.
==== Body
Background
Diabetes mellitus is syndrome, initially characterized by a loss of glucose homeostasis resulting from defects in insulin secretion, insulin action both resulting impaired metabolism of glucose and other energy- yielding fuels such as lipids and protein [1]. Experimental diabetes in animals has provided considerable insight into the physiologic and biochemical derangement of the diabetic state. Many of the derangement have been characterized in hyperglycemic animals. Significant changes in lipid metabolism and structure also occur in diabetes [2]. In these cases the structural changes are clearly oxidative in nature and are associated with development of vascular disease in diabetes [3]. In diabetic rats, increased lipidperoxidation was also associated with hyperlipidemia [4]. Liver, an insulin dependent tissue that plays a pivotal role in glucose and lipid homeostasis and it is severely affected during diabetes [5]. Liver and kidney participates in the uptake, oxidation and metabolic conversion of free fatty acids, synthesis of cholesterol, phospholipids, and triglycerides. During diabetes, a profound alteration in the concentration and composition of lipids occurs.
Despite the great strides that have been made in the understanding and management of diabetes, the disease and disease related complications are increasing unabated [6]. Inspite of the presence of known antidiabetic medicine in the pharmaceutical market, remedies from medicinal plants are used with success to treat this disease [7]. Many traditional plant treatments for diabetes are used throughout the world. Plant drugs [8] and herbal formulation [9-11] are frequently considered to be less toxic and more free from side effects than synthetic one. Based on the WHO recommendations hypoglycemic agents of plant origin used in traditional medicine are important [12]. The attributed antihyperglycemic effects of these plants is due to their ability to restore the function of pancreatic tissues by causing an increase in insulin output or inhibit the intestinal absorption of glucose or to the facilitation of metabolites in insulin dependent processes. Hence treatment with herbal drugs has an effect on protecting β-cells and smoothing out fluctuation in glucose levels [13,14]. In general, there is very little biological knowledge on the specific modes of action in the treatment of diabetes, but most of the plants have been found to contain substances like glycosides, alkaloids, terpenoids, flavonoids etc., that are frequently implicated as having antidiabetic effects [15]. In the traditional system of Indian medicine, plant formulation and combined extracts of plants are used as drug of choice rather than individual. Various herbal formulations such as diamed [16], coagent db [17] and hyponidd [18], are well known for their antidiabetic effects. Diasulin is a poly herbal drug composed of ten medicinal plants (Table 1), which are already reported for their antidiabetic activity. In our previous study, we have demonstrated the antidiabetic effect of Diasulin in alloxan diabetic rats [19]. The present investigation was undertaken to study the effect of the potential antidiabetic herbal formulation, Diasulin on lipidperoxidation and tissue lipid profile in alloxan diabetic rats. The effects produced by this drug on different parameters were compared with glibenclamide, a reference drug.
Table 1 Diasulin (Composition and concentration)
SI. No Botanical name Common name Family Part used Concentration (mg/dL)
1 Cassia auriculata Tanner's cassia Ceasalpinaceae Flower 40
2 Coccinia indica Little gourd Cucurbitaceae Fruit 40
3 Curcuma longa Turmeric Zingiberaceae Rhizome 40
4 Emblica officinalis Indian gooseberry Euphorbiaceae Fruit 20
5 Gymnema sylvestre Ram's horn Asclepiadaceae Leaves 20
6 Momordica charantia Bitter gourd Cucurbitaceae Fruit 30
7 Scoparia dulcis Sweet broom weed Scrophulariaceae Whole Plant 40
8 Syzigium cumini Jamun Myrtaceae Seed 20
9 Tinospora cardifolia Gulancha tinospora Menispermaceae Root 20
10 Trigonella foenum graecum Fenugreek Fabaceae Seed 50
Methods
Experimental animals
Male Wistar rats of body wt. 180–200 g were obtained from central Animal House, Raja Muthiah Medical College, Annamalai University. The animals were fed on standard pellet diet (Hindustan Lever, Mumbai, India) and water ad libitum. The rats used in the present study were maintained in accordance with guidelines of the National Institute of Nutrition, Indian Council of Medical Research, Hyderbad, India and the study approved by the ethical committee (Vide No: 88, 2002).
Drug and chemicals
Ethanolic extract of ten medicinal plants, which are involved in preparation of Diasulin, was a gift from herbal remedies, Pondicherry, India.
The residual extracts of the antidiabetic plants were mixed and named as Diasulin was prepared (Table 1) on the basis of an ayurvedic antidiabetic formulation proposed by Pandey et al. (1995) [20]. 500 g of each plant (chopped into small pieces) was extracted individually were, soaked overnight in 1.5 litres of 95% ethanol. This suspension was filtered and the residue was resuspended in an equal volume of 95% ethanol for 48 h and filtered again. The two filtrates were pooled and the solvents were evaporated in a rotavapor at 40° – 50°C under reduced pressure and lyophilized. Alloxan monohydrate was purchased from BDH Chemicals, Poole, England. Enzyme linked immunosorbant assay (ELISA) kit for insulin assay was purchased from Boehringer Mannheim, Germany. All other biochemicals used in this experiment were purchased from Sigma Chemical Company Inc., St Louis, Mo, and USA. The chemicals were analytical grade.
Drug administration
Diasulin was suspended in distilled water and administered orally through intragastric tube at the following doses of 50, 100 and 200-mg/kg body weight.
Experimental induction of diabetes in rats
The rats were injected intraperitoneally with alloxan monohydrate dissolved in sterile normal saline at a dose of 150 mg/kg body wt [21]. After 2 weeks, rats with moderate diabetes having glycosuria (indicated by Benedict's qualitative test) and hyperglycemia (i.e. with a blood glucose of 200–300 mg/dl) were used for the experiment.
Experimental design
In the experiment, a total of 42 rats (30 diabetic surviving rats, 12 normal rats) were used. The rats were divided into seven groups of six rats each after the induction of alloxan diabetes. Group1: Normal treated rats. Group2: Normal rats given aqueous solution of Diasulin (200 mg/kg body weight) daily using an intragastric tube for 30 days. Group 3: Diabetic control rats. Group 4: Diabetic rats given aqueous solution of Diasulin (50 mg /kg body weight) daily using an intragastric tube for 30 days. Group 5: Diabetic rats given aqueous solution of Diasulin (100 mg /kg body weight) daily using an intragastric tube for 30 days. Group 6: Diabetic rats given aqueous solution of Diasulin (200 mg /kg body weight) daily using an intragastric tube for 30 days. Group 7: Diabetic rats given aqueous solution of glibenclamide (600 μg/kg body weight) daily using an intragastric tube for 30 days.
At the end of 30 days, all the rats were killed by decapitation under pentobarbitone sodium (60 mg/kg) anaesthesia. Blood was collected in tubes containing potassium oxalate and sodium fluoride solution for the estimation of blood glucose and plasma was separated for the assay of insulin. Liver and kidney were immediately dissected out, washed in ice-cold saline to remove the blood. The tissues were weighed and 10% tissue homogenate was prepared with 0.025 M Tris – Hcl buffer, pH 7.5. After centrifugation at 200 rpm for 10 min, the clear supernatant was used to measure thiobarbituric acid reactive substances (TBARS) and hydroperoxides. For the determinations of lipids the liver and kidney tissues were weighed and lipids were extracted from tissues by the method of Folch et al [22] using chloroform – methanol mixture (CHCl3: MeOH)(2:1 v/v).
Biochemical analysis
Estimation of blood glucose and plasma insulin
Blood glucose was determined by the O-toluidine method [23]. Plasma insulin was assayed by ELISA, using Boeheringer- Mannheim Kit with a Boeheringer analyser ES300.
Estimation of lipid peroxidation
Lipid peroxidation in liver and kidney were estimated colorimetrically by thiobarbituric acid reactive substances TBARS and hydroperoxides by the method of Niehius and Samuelsson [24] and Jiang et al. [25], respectively. In brief, 0.1 ml of tissue homogenate (Tris-Hcl buffer, pH 7.5) was treated with 2 ml of (1:1:1 ratio) TBA-TCA-HCl reagent (thiobarbituric acid 0.37%, 0.25 N HCl and 15% TCA) and placed in water bath for 15 min, cooled. The absorbance of clear supernatant was measured against reference blank at 535 nm.
0.1 ml of tissue homogenate was treated with 0.9 ml of Fox reagent (88 mg butylated hydroxytoluene (BHT), 7.6 mg xylenol orange and 9.8 mg ammonium ion sulphate were added to 90 ml of methanol and 10 ml 250 mM sulphuric acid) and incubated at 37°C for 30 min. The color developed was read at 560 nm colorimetrically. Hydroperoxides was expressed as mM/100 g tissue.
Estimation of lipids
Lipids were extracted from tissues by the method of Folch et al [22] using chloroform – methanol mixture (CHCl3: MeOH) (2:1 v/v). The total cholesterol was estimated by the method of zlatkis et al [26]. To 0.1 ml of the lipid extract, 9.9 ml of ferric chloride-acetic acid reagent was added and allowed to stand for 15 min and then centrifuged. To 5 ml of the supernatant, add 3 ml of Conc. H2So4. The colour developed was read after 20 min at 560 nm against a reagent blank. Values were expressed as mg/100 g tissue.
Triglycerides were estimated by the method of foster and Dunn [27]. To an aliquot of lipid extract, evaporated to dryness. 0.1 ml of methanol was added followed by 4 ml of isopropanol. 0.4 g of alumina was added to all the tubes and shaken well for 15 min. Centrifuged and then 2 ml of the supernatant was transferred to labeled tubes. The tubes were placed in a water bath at 65°C for 15 min for saponification after adding 0.6 ml of the saponification reagent followed by 0.5 ml of acetyl acetone reagent. After mixing, the tubes were kept in a water bath at 65°C for 1 h, the contents were cooled and absorbance was read at 420 nm. The triglyceride content was expressed as mg/100 g tissue.
Phospholipid content was determined by the method of Zilversmit et al [28]. To 0.1 ml of lipid extract, added 1 ml of 5 N H2SO4 and 1 ml of concentrated nitric acid and digested to a colourless solution. The phosphorus content in the extract was determined by the method of Fiske and Subba Row [29]. The values were expressed as g/100 g tissue.
Free fatty acids were estimated by the method of Falholt et al [30]. 0.1 ml of lipid extract was evaporated to dryness. 1 ml of phosphate buffer, 6 ml of extraction solvent and 2.5 ml of copper reagent were added. All the tubes were shaken vigorously. 200 mg of activated silicic acid was added and left aside for 30 min. The tubes were centrifuged and 3 ml of the copper layer was transferred to another tube containing 0.5 ml of diphenyl carbazide and mixed carefully. The absorbance was read at 550 nm immediately. The amount of free fatty acids was expressed as mg/100 g tissue.
Statistical analysis
The data for various biochemical parameters were analysed using analysis of variance (ANOVA) and the group means were compared by Duncan's Multiple Range Test (DMRT). Values were considered statistically significant when at p < 0.05 [31].
Results
Table 2 shows the level of blood glucose and plasma insulin in control and experimental animals. There was a significant elevation in blood glucose level with significant decrease in plasma insulin levels in alloxan diabetic rats, compared with normal rats. Administration of Diasulin and glibenclamide tended to bring blood glucose and plasma insulin towards near normal levels. The effect of Diasulin at 200 mg /kg was significantly better than 50 and 100 mg/kg, therefore the higher dose was used for further biochemical studies. The administration of Diasulin and glibenclamide to normal rats showed a significant effect in lowering blood glucose and increasing plasma insulin.
Table 2 Changes in blood glucose and plasma insulin levels of control and experimental animals
Groups fasting blood glucose (mg/dL) Plasma insulin (μU / mL)
Normal 81.58 ± 2.40a 11.18 ± 0.74 a
Normal 81.58 ± 2.40a 11.18 ± 0.74 a
Diabetic control 265.00 ± 24.40b 3.55 ± 0.38 c
Diabetic + Diasulin (50 mg/kg) 210.80 ± 14.26c 5.61 ± 0.48d
Diabetic + Diasulin (100 mg/kg) 158.33 ± 11.00d 6.03 ± 0.41d
Diabetic + Diasulin (200 mg/kg) 104.16 ± 6.70e 7.05 ± 0.64e
Diabetic + Glibenclamide (600 μg/kg) 111.60 ± 9.86e 6.32 ± 0.55de
Values are given as mean ± S.D for 6 rats in each group.
Values not sharing a common superscript letter differ significantly at p < 0.05 (DMRT).
Duncan procedure, Range for the level 2.89, 3.03, 3.13, 3.20, 3.25.
Diabetic control was compared with normal, † p < 0.001.
Experimental groups were compared with diabetic control, * p < 0.001.+ + +, > 2% sugar, ++; -2% sugar; +, -1%
Table 3 represents the concentration of TBARS and hydroperoxides in tissues of control and experimental animals. There was a significant elevation in tissue TBARS and hydroperoxides during diabetes, when compared to the corresponding control group. Administration of Diasulin and glibenclamide tends to bring the values to near normal.
Table 3 Changes in levels of Tbars and hydroperoxides in liver and kidney of control and experimental animals
Groups TBARS (mM / 100 g tissue) Hydro peroxide (mM / 100 g tissue)
Liver Kidney Liver Kidney
Normal 0.73 ± 0.066a 0.845 ± 0.07a 71.42 ± 4.61a 53.56 ± 4.61a
Normal + Diasulin (200 mg/ kg) 0.72 ± 0.079a 0.785 ± 0.065a 69.63 ± 6.09a 52.40 ± 3.90a
Diabetic control 1.77 ± 0.06b 1.565 ± 0.105b 102.37 ± 6.41b 75.35 ± 4.46b
Diabetic + Diasulin (200 mg/kg) 0.88 ± 0.053c 0.970 ± 0.072c 78.56 ± 5.83c 61.89 ± 6.07c
Diabetic + Glibenclamide (600 μg/kg) 0.93 ± 0.088c 1.00 ± 0.086c 80.90 ± 5.32c 63.16 ± 6.30c
Values are given as mean ± S.D for 6 rats in each group.
Values not sharing a common superscript letter differ significantly at p < 0.05 (DMRT).
Duncan procedure, Range for the level 2.95, 3.09, 3.20.
Table 4 and 5 shows the levels of cholesterol, triglycerides, free fatty acids and phospholipids in liver and kidney of control and experimental rats respectively. Liver and kidney of diabetic rats showed significantly increased levels of cholesterol, triglycerides, free fatty acids and phospholipids, when compared with normal rats. In rats treated with Diasulin and glibenclamide there was a significant decrease in the content of cholesterol, triglycerides, free fatty acids and phospholipids in both the tissues, when compared with diabetic control rats.
Table 4 Changes in levels of cholesterol, free fatty acids, triglycerides and phospholipids in liver of control and experimental animals
Groups Cholesterol (mg/100 g wet tissue) Free fatty acids (mg/100 g wet tissue) Triglycerides (mg/100 g wet tissue) Phospholipids (mg/100 g wet tissue)
Normal 328.90 ± 18.90a 606.10 ± 21.76a 344.50 ± 23.20a 1598.00 ± 19.30a
Normal + Diasulin (200 mg/ kg) 321.25 ± 5.74a 601.93 ± 19.00a 341.10 ± 21.00a 1593.10 ± 24.10a
Diabetic control 517.33 ± 13.10b 921.60 ± 44.60b 622.50± 18.80b 1858.60 ± 18.70b
Diabetic + Diasulin (200 mg/kg) 421.75 ± 17.54c 769.16± 13.30c 456.25 ± 17.30c 1718.80 ± 14.40c
Diabetic + Glibenclamide(600 μg/kg) 445.00 ± 9.79c 802.80 ± 53.77c 530.80 ± 35.70d 1769.30 ± 17.60c
Values are given as mean ± S.D for 6 rats in each group.
Values not sharing a common superscript letter differ significantly at p < 0.05 (DMRT).
Duncan procedure, Range for the level 2.95, 3.09, 3.20.
Table 5 Changes in levels of cholesterol, free fatty acids, triglycerides and phospholipids in kidney of control and experimental animals
Groups Cholesterol (mg/100 g wet tissue) Free fatty acids (mg/100 g wet tissue) Triglycerides (mg/100 g wet tissue) Phospholipids (mg/100 g wet tissue)
Normal 375.90 ± 16.58a 434.01 ± 11.50a 282.50 ± 14.70a 1447.60 ± 26.90a
Normal + Diasulin (200 mg/ kg) 371.08 ± 8.28a 432.51 ± 11.60a 278.75 ± 14.60a 1442.50 ± 43.30a
Diabetic control 546.90 ± 23.80b 743.00 ± 25.70b 501.10 ± 34.10b 2041.50 ± 33.69b
Diabetic + Diasulin (200 mg/kg) 435.20 ± 12.18c 556.80 ± 38.50c 382.90 ± 9.28c 1684.00 ± 28.80c
Diabetic + Glibenclamide (600 μg/kg) 449.90 ± 13.49c 600.30 ± 33.40d 438.66 ± 39.30d 1819.30 ± 34.70d
Values are given as mean ± S.D for 6 rats in each group.
Values not sharing a common superscript letter differ significantly at p < 0.05 (DMRT).
Duncan procedure, Range for the level 2.95, 3.09, 3.20.
Discussion
Diabetes mellitus is one of the most common chronic disease and is associated with hyperlipidemia and co-morbidities such as obesity, hypertension. Hyperlipidemia is a metabolic complications of both clinical and experimental diabetes [32]. Alloxan, a beta cytotoxin, induces "chemical diabetes" (alloxan diabetes) in a wide variety of animal species by damaging the insulin secreting pancreatic -cell, resulting in a decrease in endogenous insulin release, which paves the ways for the decreased utilization of glucose by the tissues [33]. In our study, we have observed that Diasulin decreases blood glucose in alloxan diabetic rats. The possible mechanism of action of extract could be correlated with the reminiscent effect of the hypoglycemic sulphonylureas that promote insulin secretion by closure of K+-ATP channels, membrane depolarization and stimulation of Ca2+ influx, an initial key step in insulin secretion. In this context, number of other plants has also been reported to have antihyperglycemic and insulin stimulatory effects [34,35]. Like the plant extract, glibenclamide also produced significant reduction in blood glucose levels of alloxan diabetic rats. Since alloxan is known to destroy pancreatic β-cells, the present findings appear to be in consonance with the earlier suggestion of Jackson and Bressler [36] that sulphonylureas have extra- pancreatic antihyperglycemic mechanism of action secondary to their insulin secreting effect and the attendant glucose uptake into, and utilization by, the tissues.
Apart from the regulation of carbohydrate metabolism, insulin also plays an important role in the metabolism of lipids. Insulin is potent inhibitor of lipolysis. Since it inhibits the activity of the hormone sensitive lipases in adipose tissue and suppresses the release of free fatty acids [37]. During diabetes, enhanced activity of this enzyme increases lipolysis and releases more free fatty acids in to the circulation [38]. Increased fatty acids concentration also increases the β-oxidation of fatty acids, producing more acetyl CoA and cholesterol during diabetes.
In normal condition, insulin increases the receptor-mediated removal of LDL-cholesterol and decreased activity of insulin during diabetes causes hypercholestrolemia. Hypercholestrolemia and hypertriglycridemia have been reported to occur in diabetic rats [39]. The increased concentration of cholesterol could result in a relative molecular ordering of the residual phospholipids resulting in a decrease in membrane fluidity [40]. The increased concentration of free fatty acids in liver and kidney may be due to lipid breakdown and this may cause increased generation of NADPH, which results in the activation of NADPH dependent microsomal lipid peroxidation. Liver and kidney phospholipids were increased in diabetic control rats. Phospholipids is present in cell membrane and make up vast majority of the surface lipoprotein forming a lipid bilayer that acts as an interface with both polar plasma environment and non-polar lipoprotein of lipoprotein core [41]. Phospholipids are vital part of biomembrane rich in PUFA, which are susceptible substrate for free radicals such as O2•- and OH• radicals [42]. Increased phospholipids levels in tissues were reported by [43,44] in streptozotocin diabetic rats. Administration of Diasulin decreased the levels of tissue free fatty acids and phospholipids.
Accumulation of triglycerides is one of the risk factors in Coronary Heart Disease (CHD). The significant increase in the level of triglycerides in liver and kidney of diabetic control rats may be due to the lack of insulin. Since under normal condition, insulin activates the enzyme lipoprotein lipase and hydrolysis triglycerides [45]. Diasulin reduces triglycerides in tissues of alloxan-induced diabetic rats and may prevent the progression of CHD.
The results show increased lipid peroxidation in the tissues (liver and kidney) of diabetic control group. Previous studies have reported that there was an increased lipid peroxidation in liver, kidney and brain of diabetic rats [46,47]. This may be because the tissues contain relatively high concentration of easily peroxidizable fatty acids. Liver during diabetes, showed a relatively severe impairment in antioxidant capacity than kidney. The kidney exhibits a characteristic pattern of changes during diabetes [48]. The increase in oxygen free radicals in diabetes could be primarily due to increase in blood glucose levels, which upon autoxidation generate free radicals and secondarily due to the effects of diabetogenic agent alloxan [49]. In diabetes, hypoinsulinaemia increases the activity of the enzyme, fatty acyl coenzyme, coenzyme A oxidase, which initiates β-oxidation of fatty acids resulting in lipid peroxidation [50,51] Increased lipid peroxidation impairs membrane functions by decreasing membrane fluidity, and changing the activity of membrane-bound enzymes [51]. Its products (lipid radicals and lipid peroxide) are harmful to the cells in the body and are associated with atherosclerosis and brain damage [51]. Administration of Diasulin and glibenclamide reduced the lipid peroxidative markers in liver and kidney tissues of diabetic rats. This indicates that Diasulin inhibit oxidative damage due to the antiperoxidative effect of ingredients present in Diasulin. This could be correlated with previous study that reported that Cassia auriculata [52,53], Syzigium cumini [54,55]Tinospora cardifolia [56], and Scoparia dulcis [57] (ingredients of Diasulin) have antiperoxidative and antihyperlipidemic effect of diabetic animals.
Antidiabetic and antihyperlipidemic effect of Diasulin may be due to the effect of active constituents of different plants, viz, alkaloid and pectins from Coccinia indica [58] alkaloids from Tinospora cordifolia [59], emlicanin A and B from Emblica officinalis [60], trigonelline and scopoltin from Trigonella foenum graecum [61], alkaloid-6-methoxybenzoxazolinone and terpenoids such as scoparic acids A,B,C and scopadulcic acid A and B from 'scoparia dulcis' [62], which may be responsible for scavenging free radicals liberated by alloxan in diabetic rats.
On the basis of above results, it could be concluded that Diasulin, a combination of ten herbal plants exert a significant antihyperlipidemic and antiperoxidative effect. This could be due to different types of active principles, each with a single or a diverse range of biological activities, which serves as a good adjuvant in the present armamentarium of antidiabetic drug.
Conclusion
The liver and kidney exhibits numerous morphological and functional alterations during diabetes. Since both diabetes and hyperlipidemia are considered to be major risk factors for the premature atherosclerosis and essentially all the cholesterol in atherosclerotic plaques is derived from that of circulatory cholesterol. The antihyperlipidemic and antiperoxidative effect of Diasulin in particular could be considered as of possible therapeutic value.
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
LP – supervised the design and co-ordination of the study
RS – Practically conducted the design of the study and drafted the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Scheen JA Drug treatment of non- insulin dependent diabetes mellitus in the 1990s. Achievements and future development Drug 1997 54 355 368
Sochar M Baquer NZ Mclean P Glucose under utilization in diabetes. Comparative studies on the changes in the activities of enzymes of glucose metabolism in rat kidney and liver Mol Physiol 1985 7 51 68
Baynes JW Thrope SR Role of oxidative stress in diabetic complications Diabetes 1999 48 1 4 9892215
Morel DW Chisolm GM Antioxidant treatment of diabetic rats inhibits lipoprotein oxidation and cytotoxicity J Lipid Res 1989 30 1827 1834 2621411
Seifter S England S Arias I, Papper M, Schacter D Energy metabolism The Liver; Biology and Pathology 1982 New York, Reven Press 219 249
Tiwari AK Madhusudana Rao J Diabetes mellitus and multiple therapeutic approaches of phytochemicals: Present status and future prospects Curr Sci 2002 83 30 38
Bhattaram VA Ceraefe M Kohlest C Vest M Deundorf H Pharmacokinetics and bioavailabitlity of herbal medicinal products Phytomed 2002 9 1 36
Bailey CJ Day C Traditional treatments for diabetes Diabetes Care 1989 12 553 564 2673695
Mitra SK Gopumadhavan S Muralidhar TS Anturlikar SD Sujatha MB Effect of a herbomineral preparation D-400 in streptozotocin induced diabetic rats J Ethnopharmacol 1996 54 41 46 8941867 10.1016/0378-8741(96)01439-0
Annapurna A Kanaka Mahalakshmi D Murali Krishna K Antidiabetic activity of a polyherbal preparation (tincture of punchparna) in normal and diabetic rats Indian J Exp Biol 2001 39 500 502 11510141
Bhattacharya SK Satyan KS Chakrbarti A Effect of Trasina, an Ayurvedic herbal formulation, on pancreatic islet superoxide dismutase activity in hyperglycaemic rats Indian J Exp Biol 1997 35 297 299 9332177
The WHO Expert Committee on Diabetes Mellitus Technical Report Series 646, Geneva, and World Health Organisation 1980
Jia W Gao WY Xiao PG Antidaibetic drugs of plant origin used in China: Composition, pharmacology and hypoglycemic mechanisms Zhongguo Zhong Yao Za Zhi 2003 28 108 113 15015278
Elder C Ayurveda for diabetes mellitus: a review of the biomedical literature Altern Ther Health Med 2004 10 44 50 14727499
Loew D Kaszkin M Approaching the problem of bioequivalence of Herbal Medicinal Products Phytother Res 2002 16 705 711 12458470 10.1002/ptr.1248
Pari L Ramakrishnan R Venkateswaran S Antihyperglycaemic effect of Diamed, an herbal formulation in experimental diabetes in rats J Pharm Pharmacol 2001 53 1139 1143 11518024 10.1211/0022357011776397
Pari L Saravanan G Antidiabetic effect of coagent db, a herbal drug in alloxan induced diabetes mellitus Comp Biochem Physiol Part C 2002 131 19 25 10.1016/S1532-0456(01)00259-9
Subash Babu P Prince PSM Antihyperglycemic and antioxidant effect of hyponidd an ayurvedic herbomineral formulation in streptozotocin induced diabetic rats J Pharm Pharmacol 2004 56 1435 1442 15525451 10.1211/0022357044607
Pari L Saravanan R Antidiabetic effect of Diasulin, an herbal drug, on blood glucose, plasma insulin and hepatic enzymes of glucose metabolism in hyperglycemic rats Diabetes Obes Metab 2004 6 286 292 15171753 10.1111/j.1462-8902.2004.0349.x
Pandey VN Rajagopalan SS Chowdhorry DP An effective ayurvedic hypoglycemic formulation J Res Ayur Sid 1995 16 1 14
Katsumata K Katsumata y Ozawa T Katsumata K Potentiating effects of combined usage of three sulfonylurea drugs on the occurrence of alloxan diabetic rats Horm Metab Res 1999 25 125 126 8458609
Folch J Lees M Solane SGH A simple method for isolation and purification of total lipids from animal tissues J Biol Chem 1957 226 497 509 13428781
Sasaki T Matsy S Sonae A Effect of acetic acid concentration on the colour reaction in the O-toluidine boric acid method for blood glucose estimation Rinshbokagaku 1972 1 346 353
Niehius WG Samuelsson D Formation of Malondialdehyde from phospholipid arachidonate during microsomal lipidperoxidation Eur J Biochem 1968 6 126 130 4387188 10.1111/j.1432-1033.1968.tb00428.x
Jiang ZY Hunt JV Wolft SD Ferrous ion oxidation in the presence of xylenol orange for detection of lipid hydroperoxide in low density lipoprotein Anal Biochem 1992 202 384 389 1519766 10.1016/0003-2697(92)90122-N
Zlatkis A Zak B Boyle GJ A method for the determination of serum cholesterol J Clin Med 1953 41 486 492
Foster JB Dunn RT Stable reagents for determination of serum triglyceride by colorimetric condensation method Clin Chim Acta 1973 19 338 340
Zilversmit BB Davis AK Micro determination of plasma phospholipids by TCA precipitation J Lab Clin Med 1950 35 155 160 15400638
Fiske CH Subbarow J The colorimetric determination of Phosphorus J Biol Chem 1925 66 375 400
Falholt K Falholt W Lund B An assay colorimetric method for routine determination of free fatty acids in plasma Clin Chim Acta 1973 46 105 111 4745354 10.1016/0009-8981(73)90016-8
Duncan BD Multiple range test for correlated and heteroscedastic means Biometrics 1957 13 359 364
Bierman EL Amaral JAP Balknap BH hyperlipidemia and diabetes mellitus Diabetes 1975 25 509 515
Omamoto H Ucgigata Y Hiroskitckan STZ and alloxan induces DNA strand breaks and poly (ADPribose) synthatase in pancreatic islets Nature 1981 294 284 286 6272129 10.1038/294284a0
Venkateswaran S Pari L Effect of Coccinia indica on blood glucose, insulin and hepatic key enzymes in experimental diabetes Pharm Biol 2002 40 165 170
Latha M Pari L Antihyperglycaemic effect of Cassia auriculata in experimental diabetes and its effects on key metabolic enzymes involved in carbohydrate metabolism Clin Exp Pharmacol Physiol 2003 30 38 43 12542451 10.1046/j.1440-1681.2003.03785.x
Jackson JE Bressler R Clinical pharmacology of sulphonylurea hypoglycemic agents: part 1 Drugs 1981 22 211 245 7021124
Loci AS Shaabha M Khazraji AL Husain A Twaija A Hypoglycemic effect of a valuable extract of artemicisia herb Alba II. Effect of a valuable extract on some blood parameters in diabetic animals J Ethnopharmacol 1994 43 167 171 7990489 10.1016/0378-8741(94)90038-8
Agardh CD Bjorgell P Nilson EP The effect of tolbutamide on lipoproteins, and lipoproteinlipase and hormone sensitive lipase Diabetes Res Clin Pract 1999 46 99 108 10724087 10.1016/S0168-8227(99)00088-1
Bopanna KN Kannan J Sushma G Balaraman R Rathod SP Antidiabetic and antihyperlipidemic effect of neem seed, kernal powder on alloxan diabetic rabbits Ind J Pharmacol 1997 29 162 167
Dario G Antonio C Giuseppe P Oxidative stress and diabetic complications Diabetes Care 1996 19 257 267 8742574
Cohn RM Roth KS Lipid and lipoprotein metabolism Biochemistry and disease 1996 Williams and Wilkins Publishers, Baltimore 280
Ahmed M Shikha HA Sadhu SK Rahman MT Datta BK Diuretic, and anti-inflammatory principle from Scoparia dulcis Pharmazie 2001 56 657 660 11534346
Venkateswaran S Pari L Saravanan G Effect of Phaseolus vulgaris on circulatory antioxidants and lipids in streptozotocin-induced diabetic rats J Med Food 2002 5 97 104 12487757 10.1089/109662002760178186
Pari L Amarnath satheesh M Antidiabetic Effect of Boerhavia diffusa : Effect on Serum and Tissue Lipids in Experimental Diabetes J Med Food 2004 7 472 476 15671692 10.1089/jmf.2004.7.472
Frayn KN Insulin resistance and lipid metabolism Curr Opin Lipidol 1993 4 197 204
Latha M Pari L Preventive effects of Cassia auriculata L. flowers on brain lipid peroxidation in rats treated with streptozotocin Mol Cell Biochem 2003 243 23 28 12619885 10.1023/A:1021697311150
Ananthan R Latha M Ramkumar KM Pari L Baskar C Narmatha Bai V Antidiabetic Effect of Gymnema montanum Leaves: Effect on lipid peroxidation induced Oxidative stress in experimental diabetes Nutrition 2004 6 379 386
Aragno M Tamagno E Gatto V Brignardello E Parola S Danni O Boccuzzi G Dehydroepiandrosterone protects tissues of streptozotocin-treated rats against oxidative stress Free Rad Biol Med 1999 26 1467 1474 10401610 10.1016/S0891-5849(99)00012-X
Szkudelski T The mechanism of alloxan and streptozotocin action in b cells of the rat pancreas Physiol Res 2001 50 536 546
Oberley LW Free radicals and diabetes Free Radic Biol Med 1988 5 113 124 3075947 10.1016/0891-5849(88)90036-6
Baynes JW Ioannides C, Flatt PR Reactive oxygen in the aetiology and complications of diabetes Drug, Diet and Disease, Mechanistic Approach to Diabetes 1995 2 Hertfordshive: Ellis Horwood limited 230 31
Pari L Latha M Antidiabetic activity of Cassia auriculata flowers: Effect on lipidperoxidation in streptozotocin diabeties rats Pharm Biol 2002 40 512 517
Pari L Latha M Effect of Cassia auriculata flowers on blood sugar levels, serum and tissue lipids in Streptoszotocin diabetic rats Sing Med J 2002 43 617 621
Prince PSM Menon VP Hypoglycemic activity of Syzigium cumini seeds; effect on lipid peroxidation in alloxan diabetic rats J Ethnopharmcol 1998 61 1 7 10.1016/S0378-8741(98)00002-6
Prince PSM Kamalakkannan N Menon VP Antidiabetic and antihyperlipidaemic effect of alcoholic Syzigium cumini seeds in alloxan induced diabetic albino rats J Ethnopharmacol 2004 91 209 13 15120440 10.1016/j.jep.2003.11.001
Prince PSM Menon VP Gunasekaran G Hypolipidimic action of Tinnospora coradifolia root extract in alloxan diabetic rats J Ethnopharmacol 1999 14 4 16
Latha M Pari L Modulatorry effect of Scoparia dulcis in oxidative stress-induced lipidperoxidation in streptozotocin diabetic rats J Med Food 2003 6 379 386 14977448 10.1089/109662003772519958
Hossain MZ Shibib BA Rahman R Hypoglycemic effect of coccinia indica; inhibition if key gluconeogenic enzyme, glugose-6-phosphate Indian J Biol 1992 30 418 420
Prince PSM Menon VP Pari L Antioxidant action of Tinosfora cordifolia root extract in alloxan diabetic rats Phytother Res 2002 15 213 218
Aruna Bhattacharya chatterjee A hosal S Sail K Antioxidant activity of active tannoid principles of Emblica officinalis (amala) Indian J Exp Biol 1999 l37 676 680
Jachak SanjaM Herbal drugs as antidiabetics An overview CRIPS 2002 3 2
Pari L Latha L Protective role of scopari dulcis plant extract on brain antioxidant status and lipidperoxidation in STZ diabetic male wistar rats BMC Compliment Altern Med 2004 4 16 10.1186/1472-6882-4-16
|
15969768
|
PMC1183188
|
CC BY
|
2021-01-04 16:31:46
|
no
|
BMC Complement Altern Med. 2005 Jun 22; 5:14
|
utf-8
|
BMC Complement Altern Med
| 2,005 |
10.1186/1472-6882-5-14
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1441594148810.1186/1471-2105-6-144Methodology ArticlePAGE: Parametric Analysis of Gene Set Enrichment Kim Seon-Young [email protected] David J [email protected] Molecular Virology Division, St. Luke's-Roosevelt Hospital Center and Columbia University, New York, NY 10019, USA2 Human Genomics Laboratory, Genome Research Center, Korea Research, Institute of Biosciences and Biotechnology 52 Eoeun-dong Yuseong-gu, Daejon, 305-333, Korea3 Molecular Virology Division, St. Luke's-Roosevelt Hospital Center 432 West 58th, Street Antenucci Building, Room 709 New York, NY 10019, USA2005 8 6 2005 6 144 144 20 12 2004 8 6 2005 Copyright © 2005 Kim and Volsky; licensee BioMed Central Ltd.2005Kim and Volsky; 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
Gene set enrichment analysis (GSEA) is a microarray data analysis method that uses predefined gene sets and ranks of genes to identify significant biological changes in microarray data sets. GSEA is especially useful when gene expression changes in a given microarray data set is minimal or moderate.
Results
We developed a modified gene set enrichment analysis method based on a parametric statistical analysis model. Compared with GSEA, the parametric analysis of gene set enrichment (PAGE) detected a larger number of significantly altered gene sets and their p-values were lower than the corresponding p-values calculated by GSEA. Because PAGE uses normal distribution for statistical inference, it requires less computation than GSEA, which needs repeated computation of the permutated data set. PAGE was able to detect significantly changed gene sets from microarray data irrespective of different Affymetrix probe level analysis methods or different microarray platforms. Comparison of two aged muscle microarray data sets at gene set level using PAGE revealed common biological themes better than comparison at individual gene level.
Conclusion
PAGE was statistically more sensitive and required much less computational effort than GSEA, it could identify significantly changed biological themes from microarray data irrespective of analysis methods or microarray platforms, and it was useful in comparison of multiple microarray data sets. We offer PAGE as a useful microarray analysis method.
==== Body
Background
High-throughput technologies such as DNA microarrays and proteomics are revolutionizing biology and medicine. Global gene expression profiling using microarrays monitors changes in expression of thousands of genes simultaneously. At the data acquisition level, gene expression profiles from a given system should be reproducible and yield statistically significant changes in gene expression [1]. The large amounts of data acquired must then be reduced or "translated" to a smaller set of genes representing meaningful biological differences between control and test systems and validated in an experimental or clinical setting [2]. Since inception of the microarray technology, significant technological and analytical improvements have been introduced to meet these challenges, from experimental design [1], probe-level analysis of oligonucleotide chips [3,4], data normalization [5], statistical analysis [6], clustering techniques [7-9], to various data mining tools [10-12]. A large number of studies used microarrays successfully to discern changes in gene expression patterns either in well defined cellular populations responding to a specific stimulus in vitro [13,14] or in complex clinical settings such as cancer and neurological diseases [15-17].
While individual microarray studies can be highly informative, it is generally difficult to compare independently obtained data sets addressing the same biological problem [18], regardless of whether the same or different microarray platform was used [19-23]. The poor congruence of cross-study comparisons was attributed to incorrectly annotated probes, non-sequence overlapping but Unigene-matching probes, variation in experimental conditions, and actual biological variations among different clinical or experimental materials used [19-23]. Several new bioinformatics programs have attempted to circumvent the variability between different published data sets at the data acquisition level by comparing gene expression results for coordinate changes in biological themes [12], for similarity of significance values for each gene obtained through meta-analysis methods [15], and for reproducible gene expression patterns revealed by integrative correlation statistics [24]. Each of the methods revealed some congruency among the data sets analyzed [12,15,24] indicating that results from various transcriptional profiling studies can eventually be integrated for better general definition of normal and disease-related processes.
Another challenge of microarray data analysis is that the majority of genes in any genome-wide transcriptional profile data set are excluded from consideration because they show only subtle changes in expression. This problem was recently addressed in a gene expression profiling study of human diabetic muscles, in which no single gene (out of over 20,000) showed significant difference in expression between control and patients groups [25]. Assuming that gene expression changes can be detected at the level of co-regulated gene sets rather than individual genes, the authors devised a new analytical tool, GSEA, that tested predefined gene sets for association with disease phenotypes [25]. GSEA successfully detected oxidative phosphorylation as a biological theme that coordinately changed in diabetic muscles [25]. Although subsequent analysis suggested that the statistical tools used in GSEA may be biased toward assigning higher enrichment scores to gene sets of large size [26], the program significantly expands the potential for discovery of important process or disease related genes in a given microarray data set.
In the present work we describe a modified gene set enrichment analysis strategy that improves analysis of minimally changed gene expression profiles. PAGE employs fold change between experimental groups or other parametric data to calculate Z scores of predefined gene sets and use normal distribution to infer statistical significance of gene sets. We show here that PAGE has several advantages over GSEA and is useful in comparison of multiple microarray data sets.
Results
Statistical model and selection of minimal gene set size
According to the Central Limit Theorem in statistics, the distribution of the average of randomly sampled n observations tends to follow normal distribution as the sampling size n becomes larger, even when the parent distribution from which the average is calculated is not normal. The distribution of the average of randomly sampled observations has the same mean as the parent distribution and its variance is equal to the variance of the parent divided by the sampling size [27]. In other words, when the mean and variance of the parent distribution (whether it is normally distributed or not) are μ and σ2 the average of n observations from the parent distribution will follow a normal distribution of mean μ and variance σ2/n when the sampling size n is large enough. In PAGE, the parent distribution is a distribution of any numerical values (also termed parameters here) that describe differential expression of genes among samples in a microarray data set. Usually, the values are a fold change for an individual gene between two experimental groups or they can be a correlation coefficient between clinical indices and individual gene expression values in a microarray data set. In most cases, the distribution of a parameter, i.e., a fold change values for all genes in a gene set between two experimental groups, is not normally distributed. However, as the Central Limit Theorem states, when we sample n observations from the parent distribution of a parameter, the average of the sampled observations tends to follow the normal distribution as our sampling size n becomes larger. Here, we define sampled observations as expression values for randomly chosen individual genes within pre-defined gene sets, which may be any randomly chosen groups of genes, groups of genes representing close family members with similar functions, genes in the same biological pathway, and so on. If we define a gene set of sufficiently large size, we can use the normal distribution to test the statistical significance of that gene set.
To determine the minimal gene set size m, we first examined the distribution pattern of several microarray data sets. We used fold change between two experimental groups as a parameter and observed the distribution of fold change values in a microarray data set. As an example, we show the distribution of fold change values from microarray data set that compared gene expression of diabetic muscles with that of normal control muscles [25] (Fig. 1). The histogram of fold change values (Fig. 1A) and quantile-quantile plot of fold change values against standard normal distribution (Fig. 1B) suggested that fold change values were not normally distributed. The null hypothesis that the distribution of fold change values was normal was rejected by Kolmogorov-Smirnov normality test (D = 0.08, p-value < 2.2e-16).
Figure 1 Distribution pattern of fold change values in a microarray data and determination of minimal gene set size in PAGE. A and B. A Histogram (A) and a quantile-quantile (Q-Q) plot against standard normal distribution (B) of fold change values from microarray data set. The diabetic muscle microarray data set [25] was analyzed as described in Methods section. The fold change values between normal and patient groups were calculated and used to draw histogram (A) and Q-Q plot (B). C and D. A Histrogram (C) and a Q-Q plot (D) of an average of 10 randomly sampled values from fold change values of diabetic muscle microarray data. Kolmogorov-Smirnov normality test was performed with a null hypothesis that distribution is normal. For the distribution of fold change values (A and B), the null hypothesis was rejected (D = 0.08, p-value < 2.2e-16). For the distribution of an average of 10 randomly sampled values from fold change values (C and D), the null hypothesis was not rejected (D = 0.0239, p-value = 0.1783).
Subsequently, we sampled m genes from the parent population of fold change values, calculated the average of the sampled m observations, and observed the distribution pattern of the average values. We began from sampling size two, incremented by one until the sampling size m became 50. As expected, as the sampling size m increased, we found that the distribution of the average of sampled observations became closer to normal. As an example, we show the distribution of the average with sampling size of 10 (Fig. 1C and 1D). Both histogram (Fig. 1C) and quantile-quantile plot against standard normal distribution (Fig. 1D) suggested that the distribution of the average of 10 observations was close to normal and the null hypothesis that this distribution was normal was not rejected by Kolmogorovo-Smirnov normality test (D = 0.0239, p-value = 0.1783). Based on these observations, we set the minimal gene set size as 10. We performed the same analysis with several microarray data sets using diverse data processing procedures and found similar results (data not shown).
Comparison of PAGE with GSEA
GSEA calculates an enrichment score (ES) for a given gene set using rank of genes and infers statistical significance of each ES against ES background distribution calculated by permutation of the original data set. In contrast, PAGE calculates a Z score for a given gene set from a parameter such as fold change value between two experimental groups and infers statistical significance of the Z score against standard normal distribution.
To compare the statistical sensitivity of PAGE with that of GSEA, we analyzed human diabetic muscle microarray data sets used in the initial description of GSEA [25]. We calculated fold change values between diabetic muscles and control muscles as described above, calculated Z scores and corresponding p-values for each gene set, and compared these parameters with enrichment scores and corresponding p-values available as supplementary information (Table 1). We found that both PAGE and GSEA detected OXPHOS_HG-U133A as the most significantly changed gene set; however, the statistical significance of PAGE detection of this gene set was much greater, p = 1 × 10-11 versus p = 0.003. PAGE and GSEA ranked the next three gene sets, human_mitoDB_6_2002_HG-U133A, mitochondr_HG-U133A, and MAP00190_Oxidative_Phosphorylation, as the second through fourth significant gene sets and again, the statistical power of PAGE analysis (the p-value) was greater than that of GSEA (Table 1). Overall, PAGE detected seven gene sets as statistically significant at p < 0.05 in this data set whereas GSEA detected only one gene set at this significance. For all gene sets, p-values obtained by PAGE analysis PAGE were generally smaller than p-values of corresponding gene sets obtained by GSEA.
Table 1 Comparison of PAGE with GSEA
PAGE GSEA
Gene Set Z score p-value Gene Set ES p-value
OXPHOS_HG-U133A -10.5835 <1.0E-11 OXPHOS_HG-U133A 346.8827 0.003
human_mitoDB_6_2002_HG-U133A -6.7213 1.81E-11 human_mitoDB_6_2002_HG-U133A 215.9424 0.091
mitochondr_HG-U133A -6.4761 9.46E-11 mitochondr_HG-U133A 207.9381 0.087
MAP00190_Oxidative_phosphorylation -4.5745 4.78E-05 c20_U133 181.1569 0.062
c20_U133 -3.7461 0.0002 MAP00190_Oxidative_phosphorylation 148.9061 0.084
c25_U133 -2.7617 0.0058 c22_U133 142.9006 0.028
c21_U133 -2.1116 0.0347 c29_U133 131.4732 0.026
We extended the comparison of PAGE and GSEA to additional data sets including gene expression profiles of young and aged muscles from males (GDS 287) and females (GDS 472), or dermal fibroblasts subjected or not subjected to oxidative stress (GDS 963n1). Unlike the analysis shown in Table 1, we applied GSEA here as a multiple comparison testing tool [28] to directly compare the ability of both programs to detect multiple significant gene sets. The results of these analyses are shown in Additional files 1, 2, 3, 4. In data set GDS 472, PAGE detected 15 out of the first 30 gene sets as significantly up-regulated versus 1 out of 30 for GSEA (see Additional file 1). The corresponding numbers in GDS 287 were 26 of 30 for PAGE and 12 of 30 for GSEA (see Additional file 2). In data set GDS 963n1, both PAGE and GSEA detected 14 significant gene set out of 32 (see Additional file 3). It should be noted that the results in Additional files 1, 2, 3 were ranked by the GSEA NE scores. Although in many cases the same gene sets were identified as significant by both programs, PAGE generally detected a larger number of significant gene sets than GSEA across the entire range of GSEA NES. This is illustrated by expressing the results of the analysis shown in Additional file 3 with gene set rankings by the PAGE Z score (see Additional file 4). These results further demonstrate the utility of PAGE for sensitive detection of significantly altered biological pathways in various publicly available microarray data sets.
To further compare statistical robustness of PAGE and GSEA for detection of minimally changed gene sets we performed a simulation study using 10 hypothetical "experimental" and 10 "control" data sets, each containing expression values for 2,000 genes that were randomly chosen from standard normal distribution curve (see Additional file 5). Of the 2000 genes, 20 were designated as a hypothetical gene set of interest. The method and parameters of this simulation are described in the legend to Additional file 5. The simulation indicates that PAGE is more statistically sensitive than GSEA, being able to detect the test gene set as significant when the mean difference between the "experiment" and "control" was as small as 0.25 (see Additional file 5).
Robustness of PAGE across different microarray probe analysis methods or different microarray platforms
For a given microarray data set, use of different methods of data preprocessing, array normalization, and statistical inference can lead to different end results [3,4,29]. We tested the general applicability of PAGE to different Affymetrix probe analysis programs using the Duchenne muscular dystrophy (DMD) data set GDS 563 which contains 29 CEL files available from Gene Expression Omnibus website. Starting with 11 control and 23 DMD sets, we calculated expression values by MAS5 [30], MBEI [3], and RMA [4] programs, logarithm transformed expression values calculated by MAS5 and MBEI by base two, determined fold changes between two groups, and performed PAGE with pathway gene sets. With a cut-off p-value of < 0.05, PAGE identified eight significantly impaired gene sets with MAS5 and RMA platforms and six suppressed gene sets with MBEI platform, although the next three gene sets with lower significance derived from MBEI analysis were the same as those identified with MAS5 or RMA platforms (Table 2). For significantly induced gene sets, all three methods identified identical five gene sets (Table 2).
Table 2 Application of PAGE to different Affymetrix probe level analysis methods
MAS5 MBEI RMA
Gene Set Z score p-value Gene Set Z score p-value Gene Set Z score p-value
Inflammatory Response Pathway 7.5051 <1.0E-12 Inflammatory Responses 7.5195 5.51E-14 Inflammatory Responses 7.3487 2.00E-13
Eicosanoid Synthesis 6.5925 <1.0E-12 Eicosanoid Synthesis 3.8957 9.79E-05 Eicosanoid Synthesis 4.1557 3.24E-05
Complement Activation Classical 3.0382 0.0024 Complement Activation 3.5487 0.0004 TGF-β Signaling Pathway 3.0438 0.0023
Nucleotide Metabolism 2.0536 0.0400 Nucleotide Metabolism 2.4207 0.0155 Complement Activation Classical 2.9402 0.0033
TGF-β Signaling Pathway 1.9758 0.0482 TGF-β Signaling Pathway 2.2379 0.0250 Nucleotide Metabolism 2.5867 0.0097
MAPK Cascade -2.2529 0.0243 Glutamate Metabolism -1.7249 0.0846 GPCRs Class A Rhodopsin-like -1.9907 0.0465
Translation Factors -2.3124 0.0208 MAPK Cascade -1.8685 0.0617 MAPK Cascade -2.1306 0.0331
Krebs-TCA Cycle -3.2551 0.0011 Proteasome Degradation -1.9605 0.0499 Krebs-TCA Cycle -2.6076 0.0091
Glycogen Metabolism -3.3488 0.0008 Krebs-TCA Cycle -2.1182 0.0342 Proteasome Degradation -2.7822 0.0054
Proteasome Degradation -3.7468 0.0002 Glycogen Metabolism -2.8330 0.0046 Glycogen Metabolism -2.9328 0.0034
Fatty Acid Degradation -3.8286 0.0001 Fatty Acid Degradation -2.8570 0.0043 Fatty Acid Degradation -3.4024 0.0007
Nuclear Receptors -4.2579 2.06E-05 Nuclear Receptors -3.4686 0.0005 Nuclear Receptors -4.1600 3.18E-05
Electron Transport Chain -6.3789 1.78E-10 Electron Transport Chain -4.2009 2.66E-05 Electron Transport Chain -5.1177 3.09E-07
We then tested whether PAGE could detect common biological themes from microarray data sets produced using different microarray platforms. We analyzed breast cancer cell line experiment GSE 1299 in which three different arrays (U133A, U95A, and Agilent Human cDNA) were employed with the same RNA to analyze platform dependency of microarray data [20]. We programmed PAGE to identify pathway gene sets in each of the data sets obtained by three microarray platforms used in the experiment. Among significantly down-modulated pathways, two pathways (gap_junction proteins-connexins and cholesterol_biosynthesis) were identified by PAGE as common to all three microarray platforms (Table 3) and for up-regulated gene sets, Krebs_TCA was detected with all three microarray platforms (Table 3).
Table 3 Comparison of PAGE results from data sets produced using different microarray platforms
U95A U133A Agilent
Gene Set Z score p-value Gene Set Z score p-value Gene Set Z score p-value
Complement Activation Classical 3.0477 0.0023 Krebs TCA Cycle 5.0505 4.41E-07 tRNA Synthetases 3.1714 0.0015
Krebs-TCA Cycle 2.6771 0.0074 Cell Cycle 3.4088 0.0007 Krebs TCA Cycle 3.1291 0.0018
Nuclear Receptors 2.6562 0.0079 Translation Factors 2.8213 0.0048 Proteasome Degradation 1.9894 0.0467
Calcium Channels 1.6956 0.0900 Nuclear Receptors 2.3404 0.0193 Glycolysis and Gluconeogenesis 1.5679 0.1169
Apoptosis 1.4085 0.1590 Complement Activation Classical 2.2861 0.0223 Steroid Biosynthesis 1.5109 0.1308
TGF Beta Signaling Pathway -1.8580 0.0632 Matrix Metalloproteinases -2.2837 0.0224 Ribosomal Proteins -2.1275 0.0334
Inflammatory Response Pathway -2.2590 0.0239 TGF Beta Signaling Pathway -3.3762 0.0007 Glycogen Metabolism -2.4904 0.0128
Glycogen Metabolism -2.6981 0.0070 Inflammatory Response Pathway -3.5668 0.0004 TGF Beta Signaling Pathway -2.6910 0.0071
Cholesterol Biosynthesis -4.6988 2.62E-06 Cholesterol Biosynthesis -5.8165 6.01E-09 Cholesterol Biosynthesis -4.1642 3.12E-05
Gap Junction Proteins-Connexins -5.7792 7.51E-09 Gap Junction Proteins-Connexins -6.2689 3.64E-10 Gap Junction Proteins-Connexins -6.6162 3.69E-11
Application of PAGE to comparing different microarray data sets
Comparison of different microarray data sets dealing with similar biological questions often poses a problem of poor congruency among data sets when compared at gene level. We tested whether comparing at gene set level was better than at individual gene level to reveal congruency among different microarray data sets.
We compared two microarray data sets, GDS 287 and GDS 472, produced by the same authors using the same microarray platform Affymetrix U133A (Fig. 2 and Table 4). The data set GDS 287 records differential gene expression of muscles of young and old aged males, and GDS 472 records differential gene expression of muscles of young and old aged females. We analyzed both data sets in the same manner, selected significantly changed genes (|fold change| > 1.5 and t-test p-value < 0.05), and found the percentage of genes common to both data sets. Only 12.4% of significantly up-regulated genes occurred in both data sets and 4.4% of significantly down-regulated genes occurred in both data sets (Fig. 2A). In contrast, when we compared both data sets at gene set level, 62.5% of significantly up-regulated gene sets occurred in both data sets and 49.6% of significantly down-regulated gene sets occurred in both data sets (Fig. 2B). Actually, gene set level comparison correctly pointed out that energy metabolism such as electron transport, tricarboxylic acid cycle, and glycolysis was impaired and genes involved in mRNA processing and cell cycle regulation were up-regulated in both old aged male (GDS 287) and old aged female (GDS 472) data sets (Table 4).
Table 4 Comparison of two microarray data sets at gene set level
GDS 287 GDS 472
Gene Set Z score p-value Z score p-value
mRNA processing 4.0322 0.0001 3.2288 0.0012
cell cycle 3.0715 0.0021 3.1127 0.0019
mRNA catabolism 2.2233 0.0262 2.9762 0.0029
mRNA splicing 5.9613 3.00E-09 2.8041 0.0050
nuclear mRNA splicing_via spliceosome 6.3123 2.75E-10 2.6642 0.0077
regulation of cyclin dependent protein kinase activity 2.1343 0.0328 2.5703 0.0102
G1 phase of mitotic cell cycle 2.5482 0.0108 2.5361 0.0112
negative regulation of cell proliferation 3.0116 0.0026 2.0812 0.0374
cholesterol metabolism 2.2758 0.0229 2.0312 0.0422
blood coagulation -2.6705 0.0076 -2.0119 0.0442
protein folding -2.1739 0.0297 -2.1045 0.0353
regulation of blood pressure -3.0984 0.0019 -2.3543 0.0186
carboxylic acid transport -2.9366 0.0033 -2.8088 0.0050
signal transduction -2.5917 0.0095 -3.0115 0.0026
glycolysis -4.0631 4.84E-05 -5.3157 1.06E-07
tricarboxylic acid cycle -2.2914 0.0219 -6.3019 2.94E-10
electron transport -3.9150 0.0001 -6.9442 3.81E-12
Figure 2 Comparison of different microarray data sets at gene set level shows better congruence than comparison at gene level. A. Comparison of two different microarray data sets at gene level. Two microarray data sets, GDS 287 (Muscle function and aging-Male) and GDS 472 (Muscle function and aging-Female) were analyzed, significantly changed genes (|fold change| > 1.5 and t-test p < 0.05) from each data set were selected, and the percentage of common gene lists for both data sets was calculated. B. Comparison at gene set level. We first performed PAGE on the two microarray data sets, selected significant gene sets (p < 0.05), and calculated percentage of common gene sets for both data sets.
Discussion
Our initial objectives in developing PAGE were to increase the statistical power of the existing gene set enrichment program for analysis of subtle changes in microarray data and to simplify the laborious computational process involved. We designed PAGE as a parametric statistical test that uses normal distribution to infer the statistical significance of Z scores calculated from actual numerical parameters such as fold change between two experimental groups. Distribution-free, non-parametric methods such as the ones used in GSEA [25,30] make no assumptions about variability or the form of the population distribution and are useful when the population distribution is not normal or unknown. However, because non-parametric tests use ranks instead of measured values, they tend to be less powerful, informative, and flexible than corresponding parametric tests [31].
As described earlier, the theoretical basis for using normal distribution in PAGE is Central Limit Theorem, which states that when sampling size n is large enough, distribution of an average of sampled observations is normal regardless of the nature of parent distribution. In statistics, sampling size of 30 is generally sufficient, although the actual sampling size fulfilling Central Limit Theorem is dictated by how parent distribution is close to normal distribution. In our case, sampled observations were fold changes in expression of randomly chosen genes in a microarray data set grouped into pre-defined randomly chosen gene sets. We found that sampling size of 10 was sufficient for demonstrating close to normal distribution of averages of fold changes of constituent genes in gene sets as inferred by several normality tests including Kolmogorov-Smirnov (Fig. 1C), Anderson-Darling and Cramer-von Mises (data not shown). In our opinion, the reason that normality analysis of microarray data sets can be performed with a much smaller sampling size (10) than generally required is because parent population of parameters, i.e., fold changes of all genes being compared in microarray data sets, is already somewhat close to normal. Indeed, most fold change values lie in the center position of the distribution and the proportion of significantly changed genes decreased along the axis to both directions (Fig. 1A).
It is clear that the statistical tools used in PAGE direct the program to analysis of pre-defined gene sets in microarray data sets rather than individual genes. This design was intentional. Regardless of the experimental paradigm, the majority of the cellular transcripts analyzed for differential expression on genome-wide microarray chips such as the Affymetrix 133A/B show statistically insignificant changes. For example, our analysis of gene expression profile of HIV-1-infected astrocytes on U133A/B detected about 740 different transcripts with fold changes of > 2 or < -2 and p ≤ 0.05 [32]. This result also means that the signals obtained with over 40,000 other probes on the chips in these experiments were not considered as significant. Thus, many potentially relevant but subtle changes in biological systems may not be readily detectable by individual gene analysis of differentially expressed gene lists. PAGE, like GSEA [25], attempts to resolve this problem by utilizing the phenomenon of gene co-regulation. In complex biological systems, many genes belonging to the same family and performing similar functions or genes acting in the same biological pathway are co-regulated. Conversely, in disease states, these genes may be coordinately dysregulated. Characterization of gene co-regulation (or co-dysregulation) under different physiological and pathological conditions is an important research problem that can now be approached by bioinformatics tools [33]. The assumption behind the gene set enrichment concept is that the statistical significance of coordinated changes in a set of co-regulated genes will be greater than that for individual genes in the set. This assumption was at least in part validated by applications of GSEA and seems to be borne out for PAGE as well (this work). In fact, we consider PAGE (as well as GSEA) not only as a program for detecting correlations between experimental conditions and changes in behavior of known gene sets containing co-regulated genes, but also as a tool for intentional search for novel gene co-regulation (or co-dysregulation) in microarray data sets as part of testable hypotheses.
It may be considered paradoxical to apply a statistical test based on normal distribution to an explicit goal of detecting sets of co-regulated, that is, interdependent genes. The normal distribution paradigm requires that sampled observations are independent and identically distributed, or IID. However, we would like to argue that gene dependency caused by co-regulation in a given microarray data set should be regarded as rare, and thus statistically significant. In developing this program, we started with a basic assumption, a null-hypothesis, that all genes in a given microarray data set are independent of each other and identically distributed, that is, they are not co-regulated. With given gene sets as testable hypotheses, we then tested whether there is a significant shift of behavior of genes as a group. When we observed a significant change in a given gene set, we rejected the null-hypothesis and concluded that those genes in a gene set are co-regulated and dependent on each other. We found that with the statistical tools we used, we matched and in most cases exceeded the ability of GSEA to detect co-regulated genes.
In a direct comparison of PAGE and GSEA using published data bases, the p-values of PAGE were lower than the respective p values obtained by GSEA and as a result, the number of gene sets that can be considered significantly changed was larger (Table 1 and see Additional files 1, 2, 3, 4). Similar results were obtained in an extensive simulation study (see Additional file 5). This confirms that similar to other applications [31] the parametric statistical test is more powerful than the non-parametric method when applied to gene set enrichment analysis Two other features of PAGE facilitate the computational process involved in running the program. First, because PAGE uses standard normal distribution as a background distribution, there is no need for the preceding permutation step required for this calculation in GSEA [25]. This reduces computation time at least 1,000 times when one performs 1,000 permutation of data set to get a background distribution. Secondly, the Z score of PAGE is two-tailed, showing gene sets of both increased and decreased expression in a single analysis. In one-tailed programs [11,12,25], the entire process from ranking gene lists to class permutation and statistical inference must be repeated after analysis in one direction. Thus PAGE is a statistically powerful gene set enrichment analysis tool with features that decrease the computational burden of such programs and increase the amount of information obtained per one analysis.
With wider availability of the gene microarray technology, there is an exponential increase in publicly accessible microarray data bases obtained on different platforms, by different laboratories, and addressing a variety of biological questions. A number of increasingly advanced data analysis tools have been developed to begin to compare and integrate this diverse and often incompatible information, including programs to identify biological themes instead of differentially expressed gene lists [12,25] or programs which identify significant genes displaying consistent changes across biologically different systems [15,24]. Each approach was shown to lead to better congruency among diverse data sets than would be achieved by direct comparison of data sets, whether in demonstrating common transcriptional profiles of prostate cancer [34], common molecular markers of lung cancer [15], or a common biological theme in diabetic muscle [25].
We have found that PAGE also can be applied to integrative data analysis across various microarray platforms and biological systems. As with GSEA [25] or EASE [12], the key to PAGE utility for this purpose is the ability of the program to compare microarray data sets for gene sets rather than individual genes. Our results indicate that PAGE works well with different probe level analysis methods (Table 2) and different microarray platforms (Table 3), in each case being able to identify several common biological pathways in the same starting material tested irrespective of the platform or primary analytical method used. Gene set analysis by PAGE was also far more discriminatory than individual gene analysis in finding common biological pathway changes in different microarray data sets generated to address the same biological question, the difference between young and aged muscles (Fig. 2 and Table 4). Another feature of PAGE that is useful for comparison of multiple microarray data sets is the Z score. Z score is a normalized and linear-scale value which is microarray platform independent and which is convenient to use as an input for subsequent analysis. It is possible to generate a data matrix containing Z scores of pre-defined gene sets of multiple data sets obtained on different microarray platforms and then perform cluster analysis to identify gene sets of specific interest or to identify relationships among data sets. We applied this approach recently to cluster analysis of multiple microarray data sets of macrophages infected with bacteria, protozoa, HIV-1, or treated with cytokines, and identified gene sets that were specifically changed in HIV-1-infected cells (S.-Y. Kim and M.J. Potash, unpublished), suggesting that the Z score system of PAGE will be useful for asking broad biological questions.
Conclusion
The increasing use of microarrays comparing a carefully selected baseline to transformed tissue, differentiated tissue, or pathogen-infected tissue among others is creating a truly global database of gene expression profiles. Integrative analysis of this immense amount of information by programs such as EASE [12], meta-analysis of microarrays [15], or GSEA [25] begins to discern general patterns governing fundamental and disease-related biological processes. The program described here, parametric analysis of gene set enrichment analysis or PAGE is statistically sensitive and requires less computation than many other programs. PAGE identified significantly changed biological themes from microarray data set irrespective of microarray data analysis methods or microarray platforms. PAGE identified more common biological themes in different microarray data sets addressing the same biological problem than analysis of individual gene level. Finally, the Z score of PAGE is a normalized, linear scale value that can be used in subsequent meta-analysis. We offer PAGE as a useful microarray analysis tool.
Methods
Data sets
The microarray data set, GSEA results for NGT versus DM2, and probe sets corresponding to gene sets of Mootha et al. [25] were downloaded from the authors' website [35]. Microarray data sets of the muscle function and aging studies (GDS 287 and GDS 472) [36,37], Duchenne muscular dystrophy (DMD) study (GDS 563) and breast cancer cell line experiment (GSE 1299) [20] were downloaded from Gene Expression Omnibus [38] website.
Analysis of Microarray Data
Each Affymetrix microarray data set was normalized to have mean expression value of 1,000. Expression values below 100 were floored to 100 and then all expression values were transformed by logarithm base two. Fold change between two experimental groups was measured to identify differentially expressed genes and unpaired t-test was used to infer statistical significance of differentially expressed genes.
Parametric Analysis of Gene Set Enrichment (PAGE)
We created predefined gene sets for major mammalian Affymetrix platforms (human: U133A, U95A, HuFL6800; mouse: M74Av2; and rat: U34) from corresponding Affymetrix annotation files [39] with gene ontology (GO) biological processes, GO cellular components, GO molecular functions, and pathways as the main categories.
Z score for each gene set was calculated as follows. First, from input data containing fold change values for each genes between two experimental groups, mean of total fold change values (μ) and standard deviation of total fold change values (δ) of a given microarray data set were calculated. Then, when the mean of fold change values of genes for a given gene set was Sm and the size of a given gene set was m, the Z score was calculated as Z = (Sm – μ)*m1/2 / δ
The statistical tools within the Microsoft Excel program were used to calculate p-values from Z scores. We used an R package nortest [40] to test whether a given distribution is different from standard normal distribution or not. The R statistical programming language [40] was used for general statistical analysis and computing.
Implementation of PAGE
PAGE was written in the freely available Python programming language [41] applicable to most computer platforms and operating systems including Windows, Macintosh, and LINUX/UNIX. We also prepared stand-alone Windows-executable PAGE program using py2exe tool [42].
Availability and requirements
• Project name: PAGE
• Project home page: In construction
• Operating system: Platform independent
• Programming language: Python, Windows-executable through py2exe [42]
• Other requirements: None
• License: None
• Any restrictions to use by non-academics: None
• The Python script PAGE is available from SYK upon request
• The beta version of GSEA is now available at [43]
Authors' contributions
SYK designed and implemented PAGE method, performed the bioinformatics analyses, and drafted the manuscript. DJV supervised the development of PAGE method and wrote the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Comparison of GDS 472 by PAGE and GSEA: Ranking by GSEA.
Click here for file
Additional File 2
Comparison of GDS 287 by PAGE and GSEA: Ranking by GSEA.
Click here for file
Additional File 3
Comparison of GDS 963n1 by PAGE and GSEA: Ranking by GSEA.
Click here for file
Additional File 4
Comparison of GDS 963n1 by PAGE and GSEA: Ranking by PAGE.
Click here for file
Additional File 5
Simulation study comparing statistical sensitivity PAGE and GSEA.
Click here for file
Acknowledgements
The authors thank Drs. Mary Jane Potash and Andrew I. Brooks for critical reading of the manuscript and Mrs. Ilene M. Totillo for secretarial assistance. We also thank anonymous reviewers for critical and helpful comments to improve our manuscript. This work was supported by grants NS31492 and DA17618 from the National Institutes of Health.
==== Refs
Churchill GA Fundamentals of experimental design for cDNA microarrays Nat Genet 2002 32 490 495 12454643 10.1038/ng1031
Rhodes DR Chinnaiyan AM Bioinformatics strategies for translating genome-wide expression analyses into clinically useful cancer markers Ann NY Acad Sci 2004 1020 32 40 15208181 10.1196/annals.1310.005
Li C Wong WH Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection Proc Natl Acad Sci USA 2001 98 31 36 11134512 10.1073/pnas.011404098
Irizarry RA Bolstad BM Collin F Cope LM Hobbs B Speed TP Summaries of Affymetrix GeneChip probe level data Nucleic Acids Res 2003 31 e15 12582260 10.1093/nar/gng015
Quackenbush J Microarray data normalization and transformation Nat Genet 2002 32 496 501 12454644 10.1038/ng1032
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
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
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 USA 1999 96 2907 2912 10077610 10.1073/pnas.96.6.2907
Tavazoie S Hughes JD Campbell MJ Cho RJ Church GM Systematic determination of genetic network architecture Nat Genet 1999 22 281 285 10391217 10.1038/10343
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
Doniger SW Salomonis N Dahlquist KD Vranizan K Lawlor SC Conklin BR MAPPFinder: using Gene Ontology and GenMAPP to create a global gene-expression profile from microarray data Genome Biol 2003 4 R7 12540299 10.1186/gb-2003-4-1-r7
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
Izmailova E Bertley FMN Huang Q Makori N Miller CJ Young RA Aldovini A HIV-1 Tat reprograms immature dendritic cells to express chemoattractants for activated T cells and macrophages Nat Med 2003 9 191 197 12539042 10.1038/nm822
Nau GJ Richmond JFL Schlesinger A Jennings EG Lander ES Young RA Human macrophage activation programs induced by bacterial pathogens Proc Natl Acad Sci USA 2002 99 1503 1508 11805289 10.1073/pnas.022649799
Rhodes DR Yu J Shanker K Deshpande N Varambally R Ghosh D Barrette T Pandey A Chinnaiyan AM Large-scale meta-analysis of cancer microarray data identifies common transcriptional profiles of neoplastic transformation and progression Proc Natl Acad Sci USA 2004 101 9309 9314 15184677 10.1073/pnas.0401994101
Sanchez-Carbayo M Use of high-throughput DNA microarrays to identify biomarkers for bladder cancer Clin Chem 2003 49 23 31 12507957 10.1373/49.1.23
Ebert BL Golub TR Genomic approaches to hematologic malignancies Blood 2004 104 923 932 15155462 10.1182/blood-2004-01-0274
Sorlie T Tibshirani R Parker J Hastie T Marron JS Nobel A Deng S Johnsen H Pesich R Geisler S Repeated observation of breast tumor subtypes in independent gene expression data sets Proc Natl Acad Sci USA 2003 100 8418 8423 12829800 10.1073/pnas.0932692100
Kuo WP Jenssen TK Butte AJ Ohno-Machado L Kohane IS Analysis of matched mRNA measurements from two different microarray technologies Bioinformatics 2002 18 405 412 11934739 10.1093/bioinformatics/18.3.405
Mecham BH Klus GT Strovel J Augustus M Byrne D Bozso P Wetmore DZ Mariani TJ Kohane IS Szallasi Z Sequence-matched probes produce increased cross-platform consistency and more reproducible biological results in microarray-based gene expression measurements Nucleic Acids Res 2004 32 e74 15161944 10.1093/nar/gnh071
Mecham BH Wetmore DZ Szallasi Z Sadovsky Y Kohane I Mariani TJ Increased measurement accuracy for sequence-verified microarray probes Physiol Genomics 2004 18 308 315 15161964 10.1152/physiolgenomics.00066.2004
Tan PK Downey TJ Spitznagel EL JrXu P Fu D Dimitrov DS Lempicki RA Raaka BM Cam MC Evaluation of gene expression measurements from commercial microarray platforms Nucleic Acids Res 2003 31 5676 5684 14500831 10.1093/nar/gkg763
Jarvinen AK Hautaniemi S Edgren H Auvinen P Saarela J Kallioniemi OP Monni O Are data from different gene expression microarray platforms comparable? Genomics 2004 83 1164 1168 15177569 10.1016/j.ygeno.2004.01.004
Parmigiani G Garrett-Mayer ES Anbazhagan R Gabrielson E A cross-study comparison of gene expression studies for the molecular classification of lung cancer Clin Cancer Res 2004 10 2922 2927 15131026
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
Damian D Gorfine M Statistical concerns about the GSEA procedure Nat Genet 2004 36 663 15226741 10.1038/ng0704-663a
Grinstead CM Snell JL Introduction to Probability 1997 2 American Mathmatical Society
Sweet-Cordero A Mukherjee S Subramanian A You H Roix JJ Ladd-Acosta C Mesirov J Golub TR Jacks T An oncogenic KRAS2 expression signature identified by cross-species gene-expression analysis Nat Genet 2005 37 48 55 15608639
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
Affymetrix Microarray Suite User Guide, 2001 5 Santa Clara, CA: Affymetrix
Zolman JF Experimental Design and Statistical Inference 1993 New York: Oxford University Press
Kim S-Y Li J Bentsman G Brooks AI Volsky DJ Microarray analysis of changes in cellular gene expression induced by productive infection of primary human astrocytes: implications for HAD J Neuroimmunol 2004 157 17 26 15579276 10.1016/j.jneuroim.2004.08.032
Zeeberg BR Feng W Wang G Wang MD Fojo AT Sunshine M Narasimhan S Kane DW Reinhold WC Lababidi S GoMiner: a resource for biological interpretation of genomic and proteomic data Genome Biol 2003 4 R28 12702209 10.1186/gb-2003-4-4-r28
Rhodes DR Barrette TR Rubin MA Ghosh D Chinnaiyan AM Meta-analysis of microarrays: interstudy validation of gene expression profiles reveals pathway dysregulation in prostate cancer Cancer Res 2002 62 4427 4433 12154050
Cancer Program Data Sets of Broad Institute
Welle S Brooks AI Delehanty JM Needler N Bhatt K Shah B Thornton CA Skeletal muscle gene expression profiles in 20–29 year old and 65–71 year old women Exp Gerontol 2004 39 369 377 15036396 10.1016/j.exger.2003.11.011
Welle S Brooks A Thornton CA Senescence-related changes in gene expression in muscle: similarities and differences between mice and men Physiol Genomics 2001 5 67 73 11242590
Gene Expression Omnibus (GEO)
Affymetrix – NetAffyx Analysis Center
The Comprehensive R Archive Network
Python Programming Language
py2exe
Gene Set Enrichment Analysis – Broad
|
15941488
|
PMC1183189
|
CC BY
|
2021-01-04 16:27:24
|
no
|
BMC Bioinformatics. 2005 Jun 8; 6:144
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-144
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1491595817210.1186/1471-2105-6-149Research ArticleThesaurus-based disambiguation of gene symbols Schijvenaars Bob JA [email protected] Barend [email protected] Marc [email protected] Martijn J [email protected] Mulligen Erik M [email protected] Hester M [email protected] Jan A [email protected] Department of Medical Informatics, Erasmus University Medical Center Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands2 HUGO Gene Nomenclature Committee, Department of Biology, University College London, Wolfson House, 4 Stephenson Way, London NW1 2HE, UK2005 16 6 2005 6 149 149 22 11 2004 16 6 2005 Copyright © 2005 Schijvenaars et al; licensee BioMed Central Ltd.2005Schijvenaars 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
Massive text mining of the biological literature holds great promise of relating disparate information and discovering new knowledge. However, disambiguation of gene symbols is a major bottleneck.
Results
We developed a simple thesaurus-based disambiguation algorithm that can operate with very little training data. The thesaurus comprises the information from five human genetic databases and MeSH. The extent of the homonym problem for human gene symbols is shown to be substantial (33% of the genes in our combined thesaurus had one or more ambiguous symbols), not only because one symbol can refer to multiple genes, but also because a gene symbol can have many non-gene meanings. A test set of 52,529 Medline abstracts, containing 690 ambiguous human gene symbols taken from OMIM, was automatically generated. Overall accuracy of the disambiguation algorithm was up to 92.7% on the test set.
Conclusion
The ambiguity of human gene symbols is substantial, not only because one symbol may denote multiple genes but particularly because many symbols have other, non-gene meanings. The proposed disambiguation approach resolves most ambiguities in our test set with high accuracy, including the important gene/not a gene decisions. The algorithm is fast and scalable, enabling gene-symbol disambiguation in massive text mining applications.
==== Body
Background
The amount of information in the life sciences is staggering and growing exponentially. One of the largest biomedical resources of textual scientific information, the Medline database, currently contains over 14 million abstracts, with an estimated increase in size of more than one article per minute. Scientists are faced with an overload of information, which is particularly pressing in the biological field where high-throughput experiments in genomics and proteomics generate new data at an unprecedented rate. More often than not, interpretation of these data requires the digestion and integration of information contained in many thousands of articles and other information sources, a daunting task clearly beyond the capacity of human reading and comprehension.
Recently, a number of information retrieval systems have been proposed to extract and relate pertinent biological information from large corpora of text [1-9]. These systems even hold promise for the discovery of new, "tacit" knowledge that is hidden in the literature. The term "conceptual biology" has already been coined to distinguish this emerging field of research as a branch of biological research in its own right [10]. There are however several issues that limit the practical utility of current text-mining tools [11]. One problem is the highly-variable use of gene nomenclature in the literature [12,13], producing multiple symbols and names for one and the same gene. This complicates relating information in different documents that deal with the same gene but use different symbols. One approach to deal with this synonym problem is to make use of the information about genes and their aliases that is available in existing genetic databases.
A second, probably more intricate, problem is that a single gene symbol may refer to multiple genes, or may also be the abbreviation of terms with completely different, non-gene meanings. When building gene networks from the literature [1], for example, one would not want to contaminate the network on prostate specific antigen (PSA) with puromycin-sensitive aminopeptidase, psoriatric arthritis, pig serum albumin, or one of the more than 100 other meanings of PSA that can be found in the literature [14].
The extent of this ambiguity or homonym problem has been further subject of two recent studies. Tuason et al. [15] compared gene symbols of four organisms (not including human) and showed that up to 20% of the gene symbols of an individual organism were ambiguous with the other three organisms. In another study by the same group, Chen et al. [16] found that 85% of correctly retrieved mouse genes in a set of 45,000 abstracts were ambiguous with gene names from 20 other organisms, while ignoring gene names that were also English words. When the latter were included, 233% additional "gene" instances were retrieved, most of which were false positives. In several other studies [17-19], it was also suggested that solving this ambiguity problem is an important requirement for large-scale application of text-mining tools in the biomedical field.
General word-sense disambiguation has been studied extensively in the field of natural language processing. A wide variety of approaches has been proposed (see [20,21] for excellent reviews), including dictionary-based approaches and the use of supervised learning techniques to build classifiers that assign the proper sense to an ambiguous term. Typically, these methods use the words in a window around the ambiguous term, or information derived from this context window, such as part-of-speech or collocation.
Recently, several studies have explored the use of disambiguation techniques in the biological field. Hatzivassiloglou et al. [22] applied machine learning methods to classify symbols into one of three categories: genes, proteins, and mRNA. No attempt was made to resolve homonyms with two or more senses within one group, or with a sense outside of these three groups, and performance results were rather moderate, although still better than human interpretation. The same problem was recently tackled by Ginter et al. [23], who proposed a new classifier design and were able to slightly improve on the best method used by Hatzivassiloglou [22]. In a series of articles [17,24,25], Liu and co-workers investigated the effect of different supervised learning techniques, feature representations, and context window sizes on disambiguation performance. They obtained excellent results on a small number of ambiguous biomedical abbreviations [17,24], but for training they typically needed dozens of examples for each of the possible senses. In practice, these numbers may often be difficult to obtain. Widdows et al. [26] compared several methods for disambiguating ambiguous concepts from the Medical Subject Headings (MeSH) thesaurus [27] on a set of 70 ambiguous terms. Their most successful method achieved 74% precision and utilized existing MeSH-term co-occurrence data, which were derived from the MeSH annotations by human annotators. However, their method would not work well for gene symbols, which are poorly covered by MeSH.
Recently, Podowski et al. [19] used Bayesian classifier models to disambiguate gene symbols found in LocusLink [28]. Interestingly, their system can distinguish between gene and non-gene meanings of a symbol, acknowledging the fact that many gene symbols are abbreviations of terms with non-gene meanings. They validated their system on two manually curated test sets of 66 gene symbols, and found that the accuracy of the system is mostly over 90% when more than 20 abstracts per gene sense were available for training.
Although several of these approaches produced very good disambiguation results, they require substantial amounts of training data, typically tens of instances per sense. For gene symbol disambiguation, these numbers may be difficult to acquire. Given the extent of the homonym problem for gene symbols, manual curation of training data would be extremely laborious. Any practical disambiguation system should be trained with data that are gathered automatically, but even then the required numbers are unlikely to be available for many of the ambiguous symbols.
Here we present a disambiguation method for gene symbols, which maintains excellent performance when trained with sparse data. At the basis of our approach lies a thesaurus that is used to find biomedical concepts, including gene symbols, in text. Focusing on human genes, we first quantify the ambiguity problem for gene symbols, particularly paying attention to ambiguity arising from non-gene meanings of gene symbols. We then describe our disambiguation approach and assess the performance of the disambiguation algorithm on a large test set of documents.
Results
Thesaurus construction and ambiguity of human gene symbols
We extracted human gene symbols and aliases, gene names, and identification numbers from five publicly available databases [29]: Genew, the Genome Database (GDB), LocusLink, Online Mendelian Inheritance in Man (OMIM), and Swiss-Prot. Genes from the different databases were matched based on identification numbers and overlap in gene symbols or names, and the corresponding information was merged into a new gene thesaurus. The resulting thesaurus contained information on 26,367 human genes, with a total of 63,148 gene symbols. The percentage of ambiguous gene symbols was 4.9% (2,911 of 59,604 distinct gene symbols), whereas the percentage of genes affected by homonymy was 17.5% (4,606/26,367).
Gene symbols may not only denote multiple genes, but may also have other, non-gene meanings. To further gauge the extent of ambiguity, we searched 13 years of Medline abstracts for abbreviations (or short forms) and their expansions (or long forms) using an abbreviation expansion algorithm [30]. A total of 10,398 unique, case-sensitive short forms were found that matched a gene symbol from our gene thesaurus, with 146,198 long forms. Of these, 117,149 long forms (corresponding with 5,639 symbols) did not match, partially or completely, any of the gene names associated with that symbol, and were assumed to have a non-gene meaning. The number of different long forms per symbol with a non-gene meaning varies widely (Figure 1), up to 734 (for the short form "PC", which for example can denote "pachyonychia congenital", "prefrontal cortex", and "protective clothing"). The short forms with at least one non-gene meaning affected 26.9% of the genes and 9.5% of the gene symbols in our combined gene thesaurus. Overall, taking into account both gene and non-gene meanings, 32.7% of the genes in our combined thesaurus had one or more homonymous symbols, and 12.6% of the gene symbols in the thesaurus were ambiguous.
Figure 1 Number of non-gene meanings for gene symbols. Dots indicate the number of human gene symbols (on the vertical axis) and, for each of these symbols, the number of corresponding long forms with a non-gene meaning (horizontal axis). It should be noted that spelling variations may yield different long forms for the same non-gene meaning. To reduce the effect of these variations, long forms were stemmed.
Disambiguation of gene symbol senses
The algorithm to disambiguate homonymous gene symbols operates as follows. For each of the possible genes that the symbol can denote, a reference description is assumed to be available. Given an ambiguous symbol, the textual context in which it occurs, say, a Medline abstract, is matched with each reference description, yielding a set of matching scores. The gene corresponding with the reference description that best matches the context is then taken to indicate the symbol's meaning. However, the symbol is assumed to have a non-gene meaning if the context does not match well with any of the reference descriptions and the matching score stays below a homonym-dependent threshold, as determined by a leave-one-out procedure (see Additional file: 1 for an example of the disambiguation process). For the textual context, we used the title, abstract, and MeSH terms that had been assigned to the Medline abstract. As reference descriptions, we selected either gene annotations that were culled from OMIM, or one or more (up to five) Medline abstracts about a particular gene.
For training and testing purposes, we automatically compiled a set of annotations and abstracts for 690 ambiguous symbols, having 974 different gene meanings; 528 of the symbols had at least one non-gene meaning. All abstracts and annotations were sought for concepts from MeSH and the gene thesaurus with indexing software from Collexis (Geldermalsen, The Netherlands) [31]. For each document this yielded a list of biomedical concepts with attached relevance scores (a "concept fingerprint", or CFP), which was used for subsequent processing.
For each gene sense of a symbol, five randomly chosen abstracts were set aside for generating different reference CFPs; the remaining abstracts were used for testing. The test set contained 52,529 abstracts. The matching score between textual context and reference description was defined as the normalized cosine-vector score [32] between the CFPs of these two texts.
Overall accuracy of the disambiguation algorithm, defined as the percentage of abstracts in our test set in which the correct meaning of the homonym was chosen, was 88.9% when OMIM annotations were used as reference description. This was comparable to using one abstract as the reference (87.6%), while accuracy increased to 92.7% when a CFP combination of five abstracts was used as the reference (Figure 2). For comparison, a simple majority rule (for each symbol always select the sense that occurs most often in the test set) resulted in a baseline accuracy of 72.4%.
We made a breakdown of the errors when a combination of five abstracts was used as the reference description. As shown in Table 1, symbols indicating a gene were assigned a non-gene meaning in 6.5% of the cases; symbols with a non-gene meaning were misclassified as a gene even less frequently (4.5%). For gene symbols with multiple gene meanings, 9.9% of the symbols were assigned to the incorrect genes (993 out of 10,054 abstracts that contain an in-thesaurus homonym).
Table 1 Disambiguation of gene vs. non-gene senses. Table entries show the number of abstracts in the test set with gene symbols that were correctly or incorrectly classified by the disambiguation algorithm as having a gene or non-gene sense. The percentages indicate the correctly and incorrectly classified symbols relative to the row totals. Reference fingerprints per gene symbol sense were derived from a combination of five Medline abstracts, not being part of the test set.
Algorithm
Reference Gene Non-gene
Gene 24243 (93.5%) 1666 (6.5%)
Non-gene 1197 (4.5%) 25323 (95.5%)
Figure 2 Performance of the disambiguation algorithm. Total accuracies of the disambiguation algorithm were determined on the test set of 52,529 Medline abstracts for reference fingerprints derived from different reference descriptions: OMIM annotations or a varying number of Medline abstracts. When two or more abstracts were used, the fingerprints of the individual abstracts were averaged to yield the final reference fingerprint.
Discussion
Ambiguity of gene symbols in free text is an impediment for the massive application of text mining and literature-based discovery methods. Our assessment of gene symbol ambiguity indicates that the homonym problem cannot be ignored when text mining in the biological field is performed, corroborating findings of previous studies [15]. We found up to 33% of the genes in our thesaurus being affected by homonymy, and even this high figure is underestimating the problem because we limited ourselves to human genes only, not considering other organisms and gene products.
Disambiguation would be a trivial task if each ambiguous symbol in an abstract were accompanied by its corresponding long form at least once. Unfortunately, this approach is of limited practical value. We recently checked 3,901 Nature Genetics and BioMed Central articles and found that only 30% of the gene symbols in the abstracts are accompanied by a matching long form [33]. For an additional 8% of the symbols in the abstracts, the long form could be found in the full text. Of all gene symbols mentioned in the full-text articles, only 18% were accompanied by a long form.
We compared two sources of reference descriptions, gene annotations and abstracts about a particular gene. OMIM annotations did not perform better as a reference description than randomly chosen abstracts about a gene. The performance increased when the reference CFP was constructed from the information of several abstracts combined, but the improvement appeared marginal when more than three abstracts were used (Figure 2). This suggests that excellent disambiguation results can be obtained with relatively simple reference descriptions, and offers a viable way for massive acquisition of such descriptions from literature links in genetic databases, which in view of the extent of the homonym problem should be automatic for all practical purposes.
Our disambiguation algorithm could be used as part of a gene identification module in an information extraction application. In a recent review article on term identification, Krauthammer and Nenadic [34] distinguish between two types of disambiguation: at the broader level of pinpointing the type of a concept (e.g., distinguishing between genes and non-genes) and at the specific level of resolving different meanings of a term within a term class (e.g., distinguishing between homonymous genes within a gene thesaurus). Our approach addresses both types of disambiguation. While the algorithm was developed and tested for disambiguation of gene symbols, the approach is general and easy to apply to other ambiguous entities as well, provided adequate reference descriptions are available.
We would like to emphasize the practicality of our approach in terms of processing speed, scalability, and accuracy. The initial indexing process is the most time-consuming step, but in principle has to be done only once. For our whole test set of 52,529 abstracts, indexing currently takes about two hours on a standard Pentium IV computer. Once the context and reference fingerprints are available, the disambiguation process itself is very fast, taking about two minutes for the whole test set.
In practical applications, reference descriptions will be needed for many more than the almost 700 homonymous gene symbols that we used in this study. Considering that even a single abstract about a particular gene can provide an adequate reference description, our approach can easily be scaled up, for instance by taking the abstracts that are referenced with each of the gene descriptions in LocusLink. In our download of LocusLink, 13,811 genes had at least one reference, and 7,587 had three or more.
Automatic determination of a gene/non-gene threshold may be more difficult, as our approach presently requires the availability of examples of non-gene meanings of a symbol. We are currently investigating automatic threshold setting based on a general set of abstracts with non-gene meanings, obviating the need to acquire non-gene examples of each specific symbol.
The accuracy of any disambiguation algorithm must be very high in order to be of practical value in massive literature mining. In this respect, gene symbols that are assigned a non-gene meaning (6.5% in our test set) may be less of a problem than the other way around (4.5%), or than being assigned the wrong gene meaning (9.9%). It should be remarked that these results pertain to our test set, which is large but still limited in scope because we only selected gene symbols that occur in OMIM and had six or more abstracts per gene sense. This may have favored selection of relatively well-known genes. It is conceivable that some abstracts with gene symbols that were not selected provide a less focussed context that would perform less well.
In our data selection, we focused on human genes. The abstracts in our test set with ambiguous symbols that indicated a gene, were taken from two sources: OMIM, which is a database about human genes and diseases, or the SF/LF data set, if both short form and long form in the abstract exactly matched an entry in our human gene thesaurus. However, previous investigations [15,16] showed that substantial ambiguity of gene symbols exists across species, and suggested that most of this ambiguity was attributable to homologous genes. In a random sample of 100 abstracts from our test set with symbols that had a gene meaning, 31 symbols referred to non-human genes, mostly from mouse or rat (data not shown). All of these were apparently homologous to the human genes with identical names. In this study, we did not attempt to distinguish between homologous genes. Model organisms are often used to understand the biology of human genes and the distinction between homologous genes in text is often difficult to make, or not useful. If disambiguation of homologous genes is important, though, our approach could be extended by including reference descriptions of genes from other species.
Finally, the current performance is based on reference descriptions that were acquired fully automatically. Manual curation of these descriptions or their corresponding fingerprints for low-scoring symbols may further add to the algorithm's performance.
Conclusion
The ambiguity of gene symbols is substantial, not only because one symbol may denote multiple genes but particularly because many symbols have other, non-gene meanings. A simple, thesaurus-based disambiguation approach can resolve most ambiguities in our test set with high accuracy, including the important gene/not a gene decisions. The proposed method is fast and scalable, enabling gene-symbol disambiguation in massive text mining and information extraction applications.
Methods
Construction of the gene thesaurus
We downloaded (January 2004) information about human genes from five curated databases: Genew [35], GDB [36], LocusLink [37], OMIM [38], and Swiss-Prot [39]. For each human gene in a database, gene symbols (including aliases), gene names, and gene identification codes were extracted. Since gene name fields in the databases often contain more descriptive statements rather than gene names proper, we excluded gene names that could not be matched with one of the corresponding gene symbols according to the abbreviation expansion algorithm described by Schwartz and Hearst [30].
The number of identification codes per gene varied from database to database. Each database maintains its own set of gene identification codes, but also included cross-references to one or more of the other databases; also gene identification codes from Unigene and RefSeq were extracted if available. The original databases, including Unigene and RefSeq, were searched for information about obsolete identification codes and their possible replacements, and the extracted codes were corrected or excluded as appropriate.
To find corresponding genes from the different databases, genes with any matching identification code, gene symbol, or gene name were grouped. Within each group, subgroups of genes without conflicting identification codes were generated. If there was only one subgroup, the genes in this subgroup were taken to represent one and the same gene and all gene symbols and names of the separate genes were merged. If there was more than one subgroup, an iterative procedure was entered in which the number of similar and disparate identification codes as well as the overlap in gene symbols and names were determined for all bigroup comparisons of subgroups. A scoring rule was then used to decide whether two subgroups represented the same gene and should be merged.
The new gene thesaurus contained information on 26,367 human genes, with a total of 63,148 gene symbols. The overlap with the original databases is most substantial for LocusLink, which covers 98.1% of the genes and 92.3% of the gene symbols in the new thesaurus; Genew, maintained by the HUGO Gene Nomenclature Committee, covers 67.0% of the genes and 54.3% of the symbols. The average number of symbols per gene in the original databases varies from 1.68 (in OMIM) to 2.25 (in LocusLink); the combined gene thesaurus has an average of 2.39 symbols per gene.
Text indexing
Text documents were indexed with Collexis (Geldermalsen, The Netherlands) indexing software. For a given text, frequently occurring non-informative words are removed and the remaining terms are stemmed (using the LVG software which is part of the UMLS lexical tools [40]), i.e., brought into a standard, canonical form.
Subsequently, the document is searched for biomedical terms that occur in MeSH or in our gene thesaurus. Each found term is mapped to a unique identification code that denotes the preferred term, or concept, ti and is assigned a relevance score or weight wi that equals the Term Frequency TF (the number of occurrences fi of the concept ti (i.e., the term or its synonyms) in the document) multiplied by the Inverse Document Frequency IDF (a correction factor for the number of documents Ni containing ti in a given set of N documents; we used 10 years of Medline) [32]. We applied a commonly-used variant of the IDF that normalizes for the total number of documents [41]:
A document is then represented by an M-dimensional vector W = (w1,w2,...,wM), where M is the number of distinct concepts in the thesaurus, and wi = 0 if ti is not in the document. This weight vector W will subsequently be called the "concept fingerprint" (CFP) of the document, and is used for subsequent processing.
Document test sets
The various processing steps related to the construction of our test sets are summarized in Figure 3 and will be described below. We automatically generated two sets of abstracts containing symbols with known meaning. First, we searched all abstracts from Medline 1990–2002 for abbreviations, or short forms, and their corresponding expansions, or long forms, with the abbreviation expansion algorithm of Schwartz and Hearst [30]. For each homonymous short form, i.e., a short form with at least two different long forms, the gene thesaurus was mined for genes with a symbol and name that fully matched one of the short form/long form pairs. The abstracts in which these pairs occurred were labeled as containing the gene. If the long form only partially matched the names of a particular gene, we excluded the short form/long form pair and related abstracts from further consideration in order to guard against a non-gene meaning being assigned to a gene, or a gene meaning to an incorrect other gene. The remaining pairs, which could not be matched against the gene thesaurus, were lumped together in a non-gene meaning of the short form, and the corresponding abstracts were labeled accordingly. Thus, a test set of abstracts was collected containing short forms that represent either gene symbols or other meanings, excluding abstracts with uncertain meanings (short form/long form test set, 350,501 abstracts for 961 homonymous gene symbols) [14]. This approach to automatically create a sense-tagged set of abbreviations was originally proposed by [17], although they did not match against a thesaurus to focus on gene symbols and non-gene meanings.
Second, we extracted from our gene thesaurus all homonymous genes that had an OMIM identification code and a symbol that referred to multiple genes in the thesaurus or to a long form with non-gene meaning. For each gene, we culled the corresponding annotation from the OMIM database, including the PubMed citations that were given as a reference for the gene. It should be noted that if the Medline abstract contained a synonym of the homonymous gene symbol, we replaced the synonym with the homonymous symbol. All these abstracts together formed a second test set (OMIM test set, 23,678 abstracts for 1,376 gene symbols). The short form/long form set is much larger than the OMIM set because it contains many abstracts with symbols that have non-gene meanings.
In order to compare the performance of the disambiguation algorithm when using either annotations or abstracts as the reference descriptions (see below), symbols were selected that occurred in both our test sets and had at least six abstracts for each of its gene senses. A total of 690 symbols qualified, having 974 different gene meanings; 528 of the symbols had at least one non-gene meaning. For each gene sense of a symbol, five randomly chosen abstracts were set aside for generating a reference CFP; the remaining abstracts, with a maximum of 100 abstracts per sense, were used for testing. The test set contained 52,529 abstracts, 25,809 with symbols having a gene meaning and 26,720 with symbols having a non-gene meaning.
Figure 3 Steps involved in the construction of test and reference fingerprints. Two sets of abstracts containing symbols with known gene or non-gene meaning were constructed. One set consisted of abstracts with short-form/long-form combinations culled from Medline, the other set consisted of abstracts that were mentioned in OMIM annotations of genes. The two sets were merged by selecting symbols that occurred in both sets and had at least six abstracts for each of their gene senses. The OMIM annotations for the genes in the merged set were stored separately. A reference set was generated by randomly selecting five abstracts per gene sense from the merged set; the remaining abstracts were used for testing. All abstracts in the test and reference set as well as the OMIM annotations were indexed using the combined gene thesaurus, and the resulting "concept fingerprints" were used for reference fingerprint construction and testing of the disambiguation algorithm.
Homonym disambiguation
The disambiguation algorithm compares the textual context of a homonym in a document with a reference description of each of the genes that the homonym may possibly denote. For the textual context, we took the CFP of the abstract (including title and MeSH terms) in which the homonym occurs, setting the weight of the homonym itself to 0. For the reference descriptions, two approaches were studied. In one approach, the CFP of the OMIM annotation of a gene served as the reference. In the other approach, an averaged reference CFP was derived from the CFPs of abstracts (up to five) with a known sense of the homonym, by summing the term frequencies of the individual abstracts and computing the weights. In either case, the context CFP Wc and the reference CFP Wr were then compared by computing a normalized cosine-vector score [32]:
where wci and wri are the relevance scores of concept ti in the context CFP and the reference CFP, respectively, and |Wc| and |Wr| are the lengths of these CFPs. The cosine score varies between 1 (identical CFPs) and 0 (no overlap between CFPs). For each gene sense of the homonym, a score was determined and the sense with the highest score was assigned to the term, unless the maximum score was lower than a homonym-dependent threshold value, in which case the homonym was taken to have a non-gene meaning. Error rates were determined with a leave-one-out procedure: for each of the test abstracts of a particular gene symbol (containing both gene and non-gene meanings), scores between the context CFP and each of the reference CFPs for that symbol were determined and the highest score selected. (If there was only one reference CFP for the symbol, there was obviously only one score per abstract, which was selected.). From this set of scores, one score was left out and the threshold that minimized the error rate on the remaining scores was determined. The symbol in the abstract associated with the removed score was then classified as having a gene or non-gene meaning, depending on whether the removed score was higher or lower than this threshold value. This was done for each score in turn, yielding an overall error rate for the ambiguous symbol. The threshold that minimized the error rate for the whole sample of scores was then taken as the final homonym-dependent threshold.
Authors' contributions
BJAS designed the study, developed and implemented the disambiguation algorithm and performed most of the data analysis. BM participated in the design of the study and helped to draft the manuscript. MW participated in building the gene thesaurus and generating the training and test data, and helped to draft the manuscript. MJS participated in the design of the study and in the data analysis. EMVM was involved in text indexing. HMW participated in the gene thesaurus construction. JAK conceived of the study, participated in its design and coordination and drafted the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
This file illustrates the disambiguation process by a specific example.
Click here for file
Acknowledgements
This research was supported in part by the European Commission under the ORIEL project, contract no. IST-2001-32688.
==== Refs
Jenssen TK Laegreid A Komorowski J Hovig E A literature network of human genes for high-throughput analysis of gene expression Nat Genet 2001 28 21 28 11326270 10.1038/88213
Masys DR Welsh JB Lynn Fink J Gribskov M Klacansky I Corbeil J Use of keyword hierarchies to interpret gene expression patterns Bioinformatics 2001 17 319 326 11301300 10.1093/bioinformatics/17.4.319
Shatkay H Edwards S Wilbur WJ Boguski M Genes, themes and microarrays: using information retrieval for large- scale gene analysis Proc Int Conf Intell Syst Mol Biol 2000 8 317 328 10977093
Friedman C Kra P Yu H Krauthammer M Rzhetsky A GENIES: a natural-language processing system for the extraction of molecular pathways from journal articles Bioinformatics 2001 17 S74 82 11472995
Blaschke C Valencia A The potential use of SUISEKI as a protein interaction discovery tool Genome Inform Ser Workshop Genome Inform 2001 12 123 134 11791231
Andrade MA Bork P Automated extraction of information in molecular biology FEBS Lett 2000 476 12 17 10878241 10.1016/S0014-5793(00)01661-6
Shatkay H Feldman R Mining the biomedical literature in the genomic era: an overview J Comput Biol 2003 10 821 855 14980013 10.1089/106652703322756104
Wren JD Bekeredjian R Stewart JA Shohet RV Garner HR Knowledge discovery by automated identification and ranking of implicit relationships Bioinformatics 2004 20 389 398 14960466 10.1093/bioinformatics/btg421
Hoffmann R Valencia A A gene network for navigating the literature Nat Genet 2004 36 664 15226743 10.1038/ng0704-664
Blagosklonny MV Pardee AB Conceptual biology: unearthing the gems Nature 2002 416 373 11919607 10.1038/416373a
Masys DR Linking microarray data to the literature Nat Genet 2001 28 9 10 11326264 10.1038/88324
Obstacles of nomenclature Nature 1997 389 1 9288949
Wain H White J Povey S The changing challenges of nomenclature Cytogenet Cell Genet 1999 86 162 164 10545710 10.1159/000015372
Weeber M Schijvenaars BJ Van Mulligen EM Mons B Jelier R Van der Eijk CC Kors JA Ambiguity of human gene symbols in LocusLink and MEDLINE: creating an inventory and a disambiguation test collection Proc AMIA Symp 2003 704 708 14728264
Tuason O Chen L Liu H Blake JA Friedman C Biological nomenclatures: a source of lexical knowledge and ambiguity Pac Symp Biocomput 2004 238 249 14992507
Chen L Liu H Friedman C Gene name ambiguity of eukaryotic nomenclatures Bioinformatics 2005 21 248 256 15333458 10.1093/bioinformatics/bth496
Liu H Lussier YA Friedman C Disambiguating ambiguous biomedical terms in biomedical narrative text: an unsupervised method J Biomed Inform 2001 34 249 261 11977807 10.1006/jbin.2001.1023
Cohen KB Dolbey AE Acquaah-Mensah GK Hunter L Contrast and variability in gene names Proceedings of the Workshop on Natural Language Processing in the Biomedical Domain 2002 Philadelphia 14 20
Podowski RM Cleary JG Goncharoff NT Amoutzias G Hayes WS AZuRe, a scalable system for automated term disambiguation of gene and protein names Proceedings of the Computational Systems Bioinformatics Conference 2004 Stanford
Ide N Véronis J Introduction to the special issue on word sense disambiguation: the state of the art Computational Linguistics 1998 24 1 40
Stevenson M Wilks Y The interaction of knowledge sources in word sense disambiguation Computational Linguistics 2001 27 321 349 10.1162/089120101317066104
Hatzivassiloglou V Duboue PA Rzhetsky A Disambiguating proteins, genes, and RNA in text: a machine learning approach Bioinformatics 2001 17 S97 106 11472998
Ginter F Boberg J Järvinen J Salakosi T New techniques for disambiguation in natural language and their application to biological text J Machine Learning Res 2004 5 605 621
Liu H Johnson SB Friedman C Automatic resolution of ambiguous terms based on machine learning and conceptual relations in the UMLS J Am Med Inform Assoc 2002 9 621 636 12386113 10.1197/jamia.M1101
Liu H Teller V Friedman C A multi-aspect comparison study of supervised word sense disambiguation J Am Med Inform Assoc 2004 11 320 331 15064284 10.1197/jamia.M1533
Widdows D Peters S Cederberg S Chan C Steffen D Buitelaar P Unsupervised monolingual and bilingual word-sense disambiguation of medical documents using UMLS Natural Language Processing in Biomedicine, ACL 2003 Workshop 2003 Sapporo 9 16
Lowe HJ Barnett GO Understanding and using the medical subject headings (MeSH) vocabulary to perform literature searches Jama 1994 271 1103 1108 8151853 10.1001/jama.271.14.1103
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
Galperin MY The Molecular Biology Database Collection: 2004 update Nucleic Acids Res 2004 32 D3 22 14681349 10.1093/nar/gkh143
Schwartz AS Hearst MA A simple algorithm for identifying abbreviation definitions in biomedical text Pac Symp Biocomput 2003 451 462 12603049
van Mulligen EM Diwersy M Schmidt M Buurman H Mons B Facilitating networks of information Proc AMIA Symp 2000 868 872 11080008
Salton G Introduction to modern information retrieval 1983 New York: McGraw-Hill
Schuemie MJ Weeber M Schijvenaars BJ van Mulligen EM van der Eijk CC Jelier R Mons B Kors JA Distribution of information in biomedical abstracts and full-text publications Bioinformatics 2004 20 2597 2604 15130936 10.1093/bioinformatics/bth291
Krauthammer M Nenadic G Term identification in the biomedical literature J Biomed Inform 2004 37 512 526 15542023 10.1016/j.jbi.2004.08.004
Genew download
GDB download
LocusLink download
OMIM download
Swiss-Prot download
UMLS lexical tools
Hersh WR Information retrieval: a health and biomedical perspective 2003 New York: Springer-Verlag
|
15958172
|
PMC1183190
|
CC BY
|
2021-01-04 16:27:24
|
no
|
BMC Bioinformatics. 2005 Jun 16; 6:149
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-149
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1881604279910.1186/1471-2105-6-188SoftwareCGH-Profiler: Data mining based on genomic aberration profiles Schubert Falk [email protected] Bernhard [email protected] Stefan [email protected] Roland [email protected] Theoretical Bioinformatics, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany2 Molecular Genetics, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany3 Current address: Research Group Information Systems and Semantic Web, Institute for Computer Science, University of Koblenz-Landau, Universitätsstraße 1 56070 Koblenz, Germany2005 25 7 2005 6 188 188 14 4 2005 25 7 2005 Copyright © 2005 Schubert et al; licensee BioMed Central Ltd.2005Schubert 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
CGH-Profiler is a program that supports the analysis of genomic aberrations measured by Comparative Genomic Hybridisation (CGH). Comparative genomic hybridisation (CGH) is a well-established, molecular cytogenetic method that allows the detection of chromosomal imbalances in entire genomes. This technique is widely used in routine molecular diagnostics. Typically, chromosomal imbalances are described in a complex syntax based on the International Standard for Cytogenetic Nomenclature (ISCN). This semantic description of chromosomal imbalances hinders a large-scale statistical analysis across different experiments, e.g. for finding aberration patterns associated with a particular disease type or state.
Results
CGH-Profiler circumvents the semantic ISCN description by importing data from different CGH system vendors and by directly transferring the data into a table format that is readily accessible for subsequent statistical analysis. CGH-profiler comes with different consistency checks, calculates various statistics and automatically assigns a median copy number ratio to each chromosomal band.
Import of CGH profiles from different CGH system vendors is already supported; its extension to other systems can be readily achieved through Perl scripts.
CGH profiler can also be used to analyse comparative expressed sequence hybridisation (CESH) data. CESH reveals gene expression patterns according to chromosomal locations in a similar manner as CGH detects chromosomal imbalances.
Conclusion
CGH-Profiler is a useful tool for processing of CGH and CESH data.
==== Body
Background
CGH (comparative genomic hybridisation) is a molecular cytogenetic method to detect chromosomal imbalances [1,2]. This technology has been widely used to study genomic imbalances with prognostic and therapeutic relevance in a variety of different diseases including cancer and mental disorders (e.g., [3,4]).
For CGH, test DNA (e.g. from tumours) and reference (normal) DNA are labelled with different fluorochromes and co-hybridised onto metaphase chromosomes from normal cells. Test and reference DNA compete for binding sites, with binding probabilities depending on the abundance of the respective DNA. When hybridising the two differently labelled DNA to a normal metaphase spread, imbalances can be detected as colour changes of the chromosomes. Quantitative measurements of the colour ratio profiles along each chromosome yields the DNA copy number differences between sample and reference DNA. Digital image processing and analysis of profiles is usually performed within commercially available CGH analysis software.
Traditionally, CGH profiles have been classified according to the International System for Human Cytogenetic Nomenclature (ISCN) [5]. ISCN is a formal language for describing DNA copy number changes, amongst others. It covers low level gains (rev ish enh), high level gain (rev ish ampl) and losses (rev ish dim). A loss of the chromosomal band 4p16 is e.g. specified as "rev ish dim(4p16)". The transformation of CGH profiles into the ISCN nomenclature is a tedious process that requires a trained molecular cytogeneticist.
Here, we describe our programme CGH-profiler, which circumvents the ISCN nomenclature by automatically assigning a median copy number ratio to each chromosomal band thus allowing for an automatic detection of losses, gains and high-level gains.
The program can also be used to analyse data from the more recently introduced technique called comparative expressed sequence hybridisation (CESH) [6-9]. CESH reveals gene expression patterns according to chromosomal locations in a similar manner as CGH detects copy number changes. In brief, reverse transcribed test and reference RNA are differentially labelled and co-hybridised to normal metaphase chromosomes. The resolution of CESH is low compared to microarray gene expression arrays but no prior sequence information of genes or cloning is required. Furthermore, CESH can be performed by using existing CGH / fluorescence in situ hybridisation expertise, equipment and software. Thus it appears that the CESH data format is appropriate for CGH-profiler.
Implementation
The program includes the following processing steps:
• Conversion of the CGH profile values to a meta format independent of the used CGH system
• Profile cleansing, consistency check of the chromosomal length
• Interpolation to a given length (adjustable, e.g., 128 points per chromosome) by using cubic (Akima) splines
• Calculation of the median copy number change from all metaphases of a given case
• User defined exclusion of certain regions (centromeres, telomeres, tumour specific bands)
• Assignment of median copy numbers to chromosomal bands
Profile transformation to a meta format
The CGH profile values have to be exported from a commercial CGH system. Irrespective of the CGH system, the exported profile values are then transformed to a meta format using a Perl script. A resulting meta format file includes all metaphase profiles for all chromosomes of all cases. Each metaphase profile can have a different length.
The parsing and transformation of CGH profile values from two popular CGH systems, namely CytoVision (Applied imaging, [10]) and Isis CGH (Metasystems, [11]), are supported. Profiles from other CGH system vendors may be integrated using adapted Perl scripts.
Consistency check
A consistency check of all metaphases is performed to exclude wrongly assigned metaphases. We exclude metaphases from further analysis if the difference between its length and the mean length of the respective type of chromosome is larger than a user defined threshold (e.g. 15%). The consistency check may be switched off (by assigning 100%).
Interpolation
The remaining metaphases are interpolated to a given length. This is a prerequisite for a consistent merge of all measurements. We used cubic Akima and Fritsch/Carlson splines (polynomials of degree 2) for this interpolation implemented in the matpack library [12]. The number of interpolation points can be defined by the user, the predefined value is 128. The predefined value of 128 is especially useful for applying a wavelet transform.
Merging
From all metaphases of a given case and chromosome we calculate the median or mean copy number at each interpolation point. The choice of median or mean is optional to the user.
Exclusion of bands
The CGH measurements of some chromosomal regions (e.g. those containing a large number of highly repetitive sequences) are not reliable [13], especially after PCR amplification of the probes. The measurements of certain regions should therefore not be used for an automatic analysis. We excluded all centromeres, some telomeric regions, chromosome 19 and the sex chromosomes. However, the user can specify all critical regions in a configuration file. The ratios of all excluded regions can be marked as NA or balanced.
Assignment of median copy numbers to chromosomal bands
The mean or median profile of each case and chromosome can be mapped to an ISCN-400-ideogram without subbands [5] so that a single mean value is assigned to each chromosomal band. According to the definition of the ideogram the profile values are combined to a mean value for each chromosomal band. The predefined mapping file is based on the ISCN-400-ideogramm and a resolution of 128 interpolation points. E.g., band 1p36 is located from 1/128 to 13/128 on an ideogram. The mean value of this band is therefore the mean of the profile values 1,..,13. This data representation is the starting point for a further analysis. Using threshold values the median copy numbers can be readily translated into semantic expressions, namely losses (threshold <0.75), gains (threshold >1.25), high level gains (threshold >2) and balanced.
Results and discussion
CGH is a well-established and still widely used method. More than 150 new CGH studies were published in 2004 and referenced in Medline.
Here, we presented the program CGH-profiler that allows the input of profile values from different commercial CGH systems and transforms these values into a format that can be readily used for a quantitative analysis. Notably, our program circumvents the widely used semantic notation in the ISCN standard. Thus, it provides a basis for a more accurate and reproducible interpretation of data from large-scale genomic aberration screens.
We compared losses and gains automatically detected by CGH-profiler with those described by conventional CGH analysis (encoded in ISCN) for two data sets (data not shown) and found a high degree of accordance. Notably, conventional CGH evaluation often characterises large regions as gain or loss whereas the ratio value as determined by the programme CGH-profiler is only altered in part of the entire region.
Data mining of CGH profiles requires a matrix representation of CGH profiles. An alternative to our approach is an ISCN-to-matrix parser [14]. This is useful for large repositories of CGH studies (e.g., Progenetix [15], providing more than 10818 cases from 383 publications, SKY [16], or Charite CGH database [17]). However, a direct transformation of profile values to a matrix representation is more efficient.
The program CGH-profiler has only been used for CGH-analysis in humans so far. An extension to CGH profiles in other species can be easily achieved by adopting the mapping file used for band assignment. CGH profiler can also be used to analyse comparative expressed sequence hybridisation (CESH) data.
Conclusion
CGH-Profiler assigns to each chromosomal band a median copy number ratio by importing and processing data from different CGH system vendors. Data analysis of these continuous variables is much more efficient compared to the semantic descriptions defined by ISCN. CGH-Profiler supports therefore the data mining process of CGH and CESH data and enhances the use of mathematical functions (e.g., wavelets) on CGH and CESH profiles.
Abbreviations
CGH-Comparative genomic hybridisation
CESH comparative expressed sequence hybridisation
Authors' contributions
SJ and FS developed the method. FS and RE drafted the paper. BT and FS implemented the method. RE supervised the study. All authors prepared, read and approved the final manuscript.
Availability and requirements
The program CGH-profiler is written in C++ while the preceding data transformation is done in Perl. It is available under the GPL-license for non-commercial purposes.
-Project name: CGH-profiler
-Project home page:
-Operating systems: Unix
-Programming language: C++, Perl
-License: GNU GPL
-Restrictions to use by non-academics: on request
Acknowledgements
We thank Ian Wood for proof-reading. This study was supported by the National Genome Research network (NGFN 01GR0101).
==== Refs
Kallioniemi A Kallioniemi OP Sudar D Rutovitz D Gray JW Waldman F Pinkel D Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors Science 1992 258 818 821 1359641
du Manoir S Speicher MR Joos S Schrock E Popp S Dohner H Kovacs G Robert-Nicoud M Lichter P Cremer T Detection of complete and partial chromosome gains and losses by comparative genomic in situ hybridization Hum Genet 1993 90 590 610 8444465
Monni O Hyman E Mousses S Barlund M Kallioniemi A Kallioniemi OP From chromosomal alterations to target genes for therapy: integrating cytogenetic and functional genomic views of the breast cancer genome Semin Cancer Biol 2001 11 395 401 11562182 10.1006/scbi.2001.0395
Lichter P Joos S Bentz M Lampel S Comparative genomic hybridization: uses and limitations Semin Hematol 2000 37 348 357 11071357 10.1053/shem.2000.16594
Mitelman F Ed ISCN 1995: An International System for Human Cytogenetic Nomenclature (1995) 1995 Basel: Karger
Lu YJ Williamson D Clark J Wang R Tiffin N Skelton L Gordon T Williams R Allan B Jackman A Comparative expressed sequence hybridization to chromosomes for tumor classification and identification of genomic regions of differential gene expression Proc Natl Acad Sci U S A 2001 98 9197 9202 11481483 10.1073/pnas.161272798
Lu YJ Williamson D Wang R Summersgill B Rodriguez S Rogers S Pritchard-Jones K Campbell C Shipley J Expression profiling targeting chromosomes for tumor classification and prediction of clinical behavior Genes Chromosomes Cancer 2003 38 207 214 14506694 10.1002/gcc.10276
Gruszka-Westwood AM Horsley SW Martinez-Ramirez A Harrison CJ Kempski H Moorman AV Ross FM Griffiths M Greaves MF Kearney L Comparative expressed sequence hybridization studies of high-hyperdiploid childhood acute lymphoblastic leukemia Genes Chromosomes Cancer 2004 41 191 202 15334542 10.1002/gcc.20085
Vanhentenrijk V De Wolf-Peeters C Wlodarska I Comparative expressed sequence hybridization studies of hairy cell leukemia show uniform expression profile and imprint of spleen signature Blood 2004 104 250 255 15016649 10.1182/blood-2004-01-0181
Applied imaging
Metasystems
Matpack library
Kallioniemi OP Kallioniemi A Piper J Isola J Waldman FM Gray JW Pinkel D Optimizing comparative genomic hybridization for analysis of DNA sequence copy number changes in solid tumors Genes Chromosomes Cancer 1994 10 231 243 7522536
Baudis M Cleary ML Progenetix.net: an online repository for molecular cytogenetic aberration data Bioinformatics 2001 17 1228 1229 11751233 10.1093/bioinformatics/17.12.1228
Progenetix CGH database
SKY CGH database
Charite CGH database
|
16042799
|
PMC1183191
|
CC BY
|
2021-01-04 16:27:24
|
no
|
BMC Bioinformatics. 2005 Jul 25; 6:188
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-188
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1901604281410.1186/1471-2105-6-190SoftwareInformation theory-based algorithm for in silico prediction of PCR products with whole genomic sequences as templates Cao Youfang [email protected] Lianjie [email protected] Kexue [email protected] Chunhai [email protected] Yulei [email protected] Guifang [email protected] Junjian [email protected] Yunfang [email protected] Liping [email protected] School of Life Science and Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China2 Department of Food Science and Engineering, Northwest Institute of Light Industry, Xianyang 712081, Shaanxi, China3 Department of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, China4 Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China2005 26 7 2005 6 190 190 18 3 2005 26 7 2005 Copyright © 2005 Cao et al; licensee BioMed Central Ltd.2005Cao et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A new algorithm for assessing similarity between primer and template has been developed based on the hypothesis that annealing of primer to template is an information transfer process.
Results
Primer sequence is converted to a vector of the full potential hydrogen numbers (3 for G or C, 2 for A or T), while template sequence is converted to a vector of the actual hydrogen bond numbers formed after primer annealing. The former is considered as source information and the latter destination information. An information coefficient is calculated as a measure for fidelity of this information transfer process and thus a measure of similarity between primer and potential annealing site on template.
Conclusion
Successful prediction of PCR products from whole genomic sequences with a computer program based on the algorithm demonstrated the potential of this new algorithm in areas like in silico PCR and gene finding.
==== Body
Background
PCR technique is widely used in the molecular biology laboratory. The key step to successful PCR is primer design. Usually, software for PCR primer design only gives a subset of candidates, and we must choose the workable primers empirically. With more organisms' genomes sequenced and data freely available, we may predict PCR products with computer programs and evaluate primer candidates against the target organisms' genomic sequence prior to performing the PCR in the laboratory.
The most important thing for PCR product prediction is finding appropriate primer annealing sites on the template. Many attempts have been made to develop computer programs based on different algorithms for this purpose. Such programs include Amplify [1], simPCR [2], PCRAna [3], PUNS, and Virtual PCR [4]. PUNS [5] is a web-based program for PCR prediction, however, it does not deal with degenerate PCR primer. Some algorithms for the selection of probes or DNA oligos are also related to this field, such as longest common factor approach proposed by Rahmann [6], hybridization free energy based method used by Li [7], Kaderali's work [8] based on an extended nearest neighbor model, and Lexa's PRIMEX [9].
In our research, we designed a new algorithm based on information theory. One information source was obtained by converting the primer sequences to numeric vectors of the potential full hydrogen bond numbers, and the second was created as a vector of the actual hydrogen bond numbers formed between the primer and its potential binding site on the template. An information coefficient was computed for determining the similarity between the two information sources as a criterion to locate primer-annealing sites, and predict products. A computer program, SPCR (Simulated PCR), based on this algorithm was developed to predict PCR products, and its performance was evaluated by replicating 4 cases of laboratory PCR experiments in silico, and performing comparisons between the predicting results of our program and VPCR.
Implementation
The algorithm is based on base digitalization followed by calculation of information coefficient. Information coefficient we used in this research is a formula based on Shannon's information theory. Shannon's information theory was ever used in other research concerned on primer design, e.g. Purohit et al. [10] used entropy measure to identify conserved regions in aligned sequences for primer design.
The first step of this algorithm is to find out appropriate annealing sites from template sequences. Firstly, let the upstream and downstream primer sequences slide along template sequences respectively, one base per step. At each position where 3'-end base of primers match with the base of templates, a DNA fragment can be acquired from template sequences which length is the same as primer length, which we name as Candidates of Annealing Sites (CAS). The primer sequences and CAS fragments are all converted into numeric vectors. We name the numeric vectors of two primer sequences as PU and PD, and fragments of template as TU and TD. They are defined as below.
PU = [p1, p2,..., pm], PD = [p1, p2,..., pn] (1)
TU = [t1, t2,..., tm], TD = [t1, t2,..., tn] (2)
Variables m and n are lengths of upstream and downstream primers. The pi and ti are defined as below.
According to equation (1), (2), (3) and (4), we can transform the DNA sequences of primers and CASs into numeric vectors. Then we can perform the next step, computing the information coefficient (I) for each primer-CAS pair. The formula for the calculation of information coefficient (I) is as equation (5).
Only those sites where the similarity is higher than a preset threshold were selected as annealing sites if the last digit in the vector of T was not 0 (a requirement for perfect match at the 3'-end).
The formula for information coefficient (I) calculation is as follows:
Here
The value field of information coefficient (I) is (0,1], when primer sequence match with template completely, i.e. P = T, I = 1, and the higher the affinity between primer and template, the greater the value of information coefficient. Information coefficients formed by upstream primer and CAS of template are represented as a set Iup; and accordingly, information coefficients of downstream primer and CAS are Idn.
For each probable product, a successful amplification was determined by 5 parameters: upstream information coefficient Iup, downstream information coefficient Idn, estimated limits for product maximum and minimum length, and product amplification coefficient (Pa) which equals to an average of Iup and Idn. There are two kinds of average methods provided in SPCR program, the arithmetic average, i.e. (Iup+Idn)/2, and the geometric average, (Iup*Idn). We discuss only the geometric average here. Although the two method of average are different in computation, the values are close, and will not change the result of prediction significantly. If, and only if, the values of Iup, Idn, and Pa are all greater than the preset thresholds, and the length of predicted product lies in the preset length limit, did SPCR generate a product between upstream and downstream primer annealing sites within the product length range.
SPCR was implemented as a Win32 application and written in C++ language. It comprises an executable program that can be run directly without the need for installation. The user inputs the primer sequences, sets the thresholds, and provides locally one or more template sequence files, and push the button "Start PCR" to begin the prediction. SPCR can recognize degenerate primers encoded with the IUPAC nucleotide codes. Degenerated base are allowed in primer sequences. Template sequences can be available genomic sequence of the target organism, which must be Single- or Multi-FASTA format. The output of SPCR is a text file of a list of all predicted PCR products.
SPCR saves all produced data, including predicted products and all parameters, into a user-specified result file in pure text format. The output file consists of four parts. The first part is a table in which all of predicted products are listed, including their Pa value, product length, template it comes from, direction of amplification, position of beginning and end, Iup and Idn value, and the upstream and downstream primer sequences. The second part is a digit indicating the number of predicted products. The third part is the detailed nucleotide sequence data in FASTA format of all predicted products in the same order as in the table. The last part includes all parameters set before SPCR running. The time complexity of this algorithm is O(n), where n is the aggregate length of all the template sequences. In addition, SPCR provides a function to simulate agarose gel patterns of output data.
Results
To test the performance of the SPCR program, we first used SPCR to simulate the PCR experiments presented in the Virtual PCR (VPCR) paper [4].
The SPCR prediction results for ARR5 and ARR7 genes were identical to the laboratory PCR results, and no unspecific products were produced (Fig. 1). VPCR gave one unspecific product for the ARR7 gene and 8 unspecific products for ARR5. The SPCR prediction for GEN12 and GEN13 families with degenerate primers gave a predicted agarose profile similar to the laboratory results. More than 80% of the predicted products had counterparts in the laboratory PCR results. VPCR predicted only 2 products for each of the gene families and gave more than 8 unspecific products in each case. Although Lexa et al. has released the 2.0 version of VPCR , the predictions of GEN12 and GEN13 are still not as accurate as SPCR.
Figure 1 The comparison of SPCR prediction results with real experimental results and Virtual PCR results. Use of specific primers ARR5 (ARR5a, ARR5b), ARR7 (ARR7a, ARR7b), and degenerate primers (GEN12 and GEN13) to predict PCR products from the Arabidopsis genome. (A) Prediction results of SPCR, lanes ARR7 and ARR5 are ARR7 and ARR5 gene, respectively, amplified from the genome with the specific primers, lanes GEN12 and GEN13 are predicted results with the degenerate primers; (B) the laboratory PCR results; (C) prediction results of VPCR. Arrows show bands present on gels of actual PCR products. (B and C are adopted from Lexa et al. (2001) with permission.)
We used the universal primer pair P0 and P6 [11] to predict 16S rRNA genes from 59 completed bacterial genomes. We correctly located all copies of 16S rRNA genes from 52 of the tested genomes with all threshold parameters at 0.85. Predictions for three strains yielded 1 or 2 unspecific products, which were eliminated when the threshold parameters were increased to 0.86 or 0.9. Discrepancies between SPCR prediction and GenBank data for three strains were found to be a result of incorrect annotation of the data. In only one strain was an unspecific product found that was longer than the expected gene copy.
SOX proteins contain a conserved HMG-related DNA binding domain, which shares at least 50% identity with that of SRY. Cremazy [12] designed a pair of highly degenerate primers, which were capable of amplifying a broad spectrum of SOX HMG sequences. We predicted SOX HMG sequences from Homo sapiens genome with the primer pair P5-2 and P3-1 and verified with the BLAST (blastn) program [13] that the top 10 products were all known SOX genes.
Two primer pairs for amplification of conserved regions in polymerase coronavirus genes were used to predict PCR products from 7 complete coronavirus genomes, including SARS-CoV and 14 other coronavirus species. All predictions yielded the expected products from all templates, 453 bp for primer pair IN-2(+)/IN-4(-) [14] and 251 bp for 2 Bp/4Bm [15]. Phylogenetic trees based on the predicted products were in agreement with the current taxonomy of coronaviruses.
Besides above cases, we also tested SPCR with successful prediction of primer pairs for genome-specific amplification of environmental bacteria. Unspecific PCR primers against bacterial genomes for random amplification of bacterial community samples also yielded satisfactory results.
The threshold value of 3 parameters Iup, Idn and Pa are selected empirically. Increase of the parameter values can lead to reduction of unspecific products. This corresponds to increase of annealing temperature for reducing unspecific products in actual PCR. Usually, the 3 parameters should be greater than 0.8, if a specific PCR is predicted. If parameters higher than 9.0 still give unspecific products we suggest change of primer pairs if specificity is a major concern.
Discussion
Amplification of non-targeted products is a common problem in PCR experiments, especially when using complex templates, such as whole genomic DNA or mixtures of genomes. It has been suggested that mismatch tolerance during primer annealing to template is the most important reason for unspecific PCR products, followed by primer length, template size, and product size limits [2]. The current PCR prediction methods are mainly based on probabilistic theories for similarity analysis between primer and template, while some other methods are also used in this field, such as statistical thermodynamics [16] and string comparisons.
A successful computer program for PCR product prediction should be able to identify all potential annealing sites. Sequence similarity between primer and template is the primary factor for selecting annealing sites. In this work we developed a new algorithm to assess the similarity between primer and template after the base sequences were converted into vectors of hydrogen bond numbers. We consider annealing of primer to template a means of information transfer. The hydrogen bond number vector for the primer is the source information, while the vector for template is the target information. The difference between the two information sources is a reflection of the fidelity of this information transfer process. Since hydrogen bonds are formed as a result of specific base recognition between the two DNA strands and the total number of hydrogen bonds is also a major force holding the two strands together, the similarity between these two information sources may be a good estimate of the probability of both annealing site selection and annealed structure stability. The information coefficient calculated in this work is a measure of similarity between two information sources. The value varies between 0–1, reflecting complete difference to 100% identity. In the SPCR program, the threshold can be increased to reduce expected products, which is comparable to increasing annealing temperature to reduce unspecific products in laboratory PCR. In contrast to BLAST, this algorithm tolerates any type of mismatch between primer and template. The successful prediction of all copies of 16S rRNA genes in complete genomes demonstrates the potential of using this algorithm for gene prediction of newly sequenced genomes.
Conclusion
In our evaluating cases, the SPCR program is reasonably good for predicting all potential PCR products with complex templates. This can help the user choose the primer pair that gives the least possible non-targeted products. However, the prediction for random PCR products with SPCR is not satisfactory. When the template is too big, as the case with the human genome, the running time can reach 48 hours.
In current version, SPCR do not consider the situation of insertions and deletions of template sequences, these situations should be considered in future versions. Some refinements for the algorithm can also be done in future, for example, considering the effect of base stacking and alternative penalty for mispairs may improve the accuracy of prediction.
Availability and requirements
The SPCR program and supplemental materials, including details of all the prediction experiments in this paper are freely available at our website: , also see [Additional file 1]. SPCR program was developed with C++ under Win32 and Linux, so it can be run under both platforms. There is no restriction for using the SPCR program.
Authors' contributions
YC developed the original SPCR program, tested it and analyzed the results of SPCR predictions. LW put up with the concept, which uses hydrogen bonds number to represent basepairs. KX put forward the information coefficient formula, which was used in this paper. CK, YZ and JH improved the SPCR program. GW and YW carried out experimental verifications against the result of SPCR program. LZ is responsible for guiding the whole project. All authors have read and approved this final manuscript.
Supplementary Material
Additional File 1
Zip compressed file, included 4 files in it: SPCR.exe, Help.chm, SimGel.exe, and Cases used to test the SPCR program.mht. SPCR.exe is the main program of SPCR; Help.chm is the help information; SimGel.exe is the assistant program for simulating Gel images; Cases used to test the SPCR program.mht is an html format file, which includes most of detailed data of all cases carried out with SPCR.
Click here for file
Acknowledgements
This work was supported by a grant from National Science Foundation of China (grant No. 30470061)
==== Refs
Engels WR Contributing software to the internet: the Amplify program Trends Biochem Sci 1993 18 448 450 8291093 10.1016/0968-0004(93)90148-G
Rubin E Levy AA A mathematical model and a computerized simulation of PCR using complex templates Nucleic Acids Res 1996 24 3538 3545 8836180 10.1093/nar/24.18.3538
Nishigaki K Saito A Takashi H Naimuddin M Whole genome sequence-enabled prediction of sequences performed for random PCR products of Escherichia coli Nucleic Acids Res 2000 28 1879 1884 10756186 10.1093/nar/28.9.1879
Lexa M Horak J Brzobohaty B Virtual PCR Bioinformatics 2001 17 192 193 11238077 10.1093/bioinformatics/17.2.192
Boutros PC Okey AB PUNS: transcriptomic- and genomic-in silico PCR for enhanced primer design Bioinformatics 2004 20 2399 2400 15073008 10.1093/bioinformatics/bth257
Rahmann S Algorithms for Probe Selection and DNA Microarray Design 2004 Doctorate Berlin , Free University of Berlin 208
Li F Stormo GD Selection of optimal DNA oligos for gene expression arrays Bioinformatics 2001 17 1067 1076 11724738 10.1093/bioinformatics/17.11.1067
Kaderali L Schliep A Selecting signature oligonucleotides to identify organisms using DNA arrays Bioinformatics 2002 18 1340 1349 12376378 10.1093/bioinformatics/18.10.1340
Lexa M Valle G PRIMEX: rapid identification of oligonucleotide matches in whole genomes Bioinformatics 2003 19 2486 2488 14668240 10.1093/bioinformatics/btg350
Purohit HJ Raje DV Kapley A Identification of signature and primers specific to genus Pseudomonas using mismatched patterns of 16S rDNA sequences BMC Bioinformatics 2003 4 19 12769821 10.1186/1471-2105-4-19
Grifoni A Bazzicalupo M Di Serio C Fancelli S Fani R Identification of Azospirillum strains by restriction fragment length polymorphism of the 16S rDNA and of the histidine operon FEMS Microbiol Lett 1995 127 85 91 7737487 10.1016/0378-1097(95)00042-4
Cremazy F Soullier S Berta P Jay P Further complexity of the human SOX gene family revealed by the combined use of highly degenerate primers and nested PCR FEBS Lett 1998 438 311 314 9827568 10.1016/S0014-5793(98)01294-0
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
Ksiazek TG Erdman D Goldsmith CS Zaki SR Peret T Emery S Tong S Urbani C Comer JA Lim W Rollin PE Dowell SF Ling AE Humphrey CD Shieh WJ Guarner J Paddock CD Rota P Fields B DeRisi J Yang JY Cox N Hughes JM LeDuc JW Bellini WJ Anderson LJ A novel coronavirus associated with severe acute respiratory syndrome N Engl J Med 2003 348 1953 1966 12690092 10.1056/NEJMoa030781
Stephensen CB Casebolt DB Gangopadhyay NN Phylogenetic analysis of a highly conserved region of the polymerase gene from 11 coronaviruses and development of a consensus polymerase chain reaction assay Virus Res 1999 60 181 189 10392726 10.1016/S0168-1702(99)00017-9
Gilson MK Given JA Bush BL McCammon JA The statistical-thermodynamic basis for computation of binding affinities: a critical review Biophys J 1997 72 1047 1069 9138555
|
16042814
|
PMC1183192
|
CC BY
|
2021-01-04 16:27:23
|
no
|
BMC Bioinformatics. 2005 Jul 26; 6:190
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-190
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-711600016710.1186/1471-2407-5-71Research ArticleConcurrent chemoradiotherapy with low dose weekly gemcitabine in stage III non-small cell lung cancer Abacioglu Ufuk [email protected] Perran F [email protected] Hale [email protected] Meric [email protected] Nazim S [email protected] Department of Radiation Oncology, Marmara University Medical School Hospital, Istanbul, Turkey2 Division of Medical Oncology, Marmara University Medical School Hospital, Istanbul, Turkey2005 6 7 2005 5 71 71 5 1 2005 6 7 2005 Copyright © 2005 Abacioglu 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
Combined chemoradiotherapy (CRT) is the treatment of choice for stage III NSCLC. Gemcitabine (G) is a novel deoxycitidine analogue that has been proven to be a potent radiosensitizer. Twenty-two consecutive patients were treated with concurrent CRT to demonstrate the tolerability and efficacy of low dose G given weekly as radiosensitizer in stage III NSCLC.
Methods
Patients with KPS ≥70, adequate bone marrow reserve, with no prior radiotherapy (RT) and surgery were included. Eighteen patients had received prior induction chemotherapy (CT). G (75 mg/m2/week) was infused over 1 hour for 6 weeks. Thoracic RT was given two hours later over 6 weeks at 1.8 Gy/day fractions (total dose of 61.2 Gy). Pulmonary toxicity was evaluated with computed tomography scans in 6 weeks.
Results
Median age was 60 years (range, 48–75), median follow-up was 15 months (range, 2–40). Sixty-eight percent of patients were male and median KPS score was 90. Conformal 3D-RT planning was used in 64% of patients. G was given for a median of 5 weeks (range 1–9). Twelve patients (54.6%) received all planned CT. G was stopped because of intolerance in 6 and death in 2 patients. Seven patients (31.8%) had radiation pneumonitis. Twenty patients were evaluated for overall response, 1 patient (4.5%) had clinical CR, 81.8% had PR while 9.5% had SD. Median overall survival (OS) was 14 ± 5 months (95% CI 3–25). One- and 2-year OS rates were 55% and 38%. Sixteen patients died of disease-related events (6 with progression of primary tumor, 8 due to metastatic disease), 2 patients died of other causes. One- and 2-year progression-free survival and local control rates were 56%, 27% and 79%, 51%, respectively.
Conclusion
G might be used as radiosensitizer for patients with stage III NSCLC who could not receive full doses CT with concurrent RT.
==== Body
Background
Lung cancer is the leading cause of cancer related deaths all around the world. About 80% of all lung cancer patients are non-small cell lung cancer (NSCLC) and 20–30% of these patients present with locally advanced disease. Although 1-year survival rate of patients with local disease is more than 70%, this rate is about 40% for patients with stage III disease [1]. Chemoradiotherapy (CRT) is considered the gold standard treatment of these patients after significant survival advantage and increase in local control had been shown by 3 meta-analyses [2-4]. Sequential versus concurrent CRT has been further investigated and concurrent CRT has been found to be superior [5-7]. Despite advances in treatment modalities overall 5-year survival rate for locally advanced NSCLC remains less than 15% [8].
Gemcitabine (G) is a deoxycytidine analogue that demonstrated activity in NSCLC as single agent and in combination also has been proven to be a potent radiosensitizer. Mechanism of radiosensitizing activity of G is not well understood. Possible mechanisms that have been shown in preclinical studies are reduction of apoptotic threshold for radiation due to intracellular dATP pool depletion with simultaneous redistribution of cells into the S phase [9]. As a result of these factors DNA damage caused by radiation cannot be properly repaired, apoptosis was induced and cell death increased. The average enhancement ratio of G is ≥1.5 and sensitization persists for at least 72 hours [10]. Preclinical data supports that more frequent G dosing schedule with radiation could yield local control advantages in the setting of a clinical trial [11].
In this study it is aimed to demonstrate the efficacy and tolerability of low dose G given weekly as radiosensitizer in stage III NSCLC patients.
Methods
Consecutive patients between the ages 18–75 years with histological or cytological confirmed stage III NSCLC who had Karnofsky Performance Status (KPS) ≥70 in Medical and Radiation Oncology Clinics of Marmara University between December 1999 and July 2002 were considered for this study. Other patient selection criteria were white blood cell count ≥3000/mm3, hemoglobin level ≥10 gr/dl, platelet count ≥100000/mm3, no prior RT or lung resection accept open lung biopsy or mediastinoscopy for diagnostic purposes. All combinations and cycles of induction CT prior to planned concurrent CRT were eligible as the ones who did not receive any induction CT. All patients started CRT within 1 month after completion of induction CT. There was not a limitation to the lung volume planned to be irradiated.
The overall dose of RT was 60–61.2 Gy administered over 6 weeks with daily fractions of 2 Gy administered for 5 days each week. Planning computed tomography scan (CTS) was performed for each patient in supine position. The planning target volume (PTV) consisted of radiological visible primary tumor with a margin of 2 cm in each diameter. Elective nodal irradiation field included the PTV for the first 6 patients. After 6 patients the treatment planning and volume was amended to multiple field arrangements with custom blocking to all fields and involved hilar and mediastinal lymph nodes. Irradiation was administered with megavoltage photons of at least 6 MV. The dose of 60 Gy was specified at the isocenter of the central axis and was corrected for pulmonary heterogeneity. The spinal cord was not allowed to receive RT more than 45 Gy.
G (75 mg/m2/week) was infused each Monday by i.v. route over 1 hour in 500 cc of 0.9% normal saline for 6 weeks. RT was administered 2 hours after the G infusion. 5-HT antagonists were given 30 minutes prior to G infusion.
Patients were followed weekly with routine physical examination and complete blood count (CBC). Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria (CTC), version 2.0 [12]. If the patient had WBC ≤2500/mm3, platelets ≤50000/mm3 or grade 3 or 4 hematological toxicity, G was not administered that week. G was stopped if CBC was as low as mentioned above for 2 consecutive weeks or grade ≥3 non-hematological toxicities were seen during the treatment period.
Pulmonary toxicity was evaluated with CTS of the thorax 6 weeks after the completion of RT. Patients having ground glass appearance or alveolar consolidation on CTS were started on corticosteroids (40 mg of fluocortolon daily, which is the standard treatment protocol of the centre) for 2–4 weeks even if they were asymptomatic. Responses were routinely evaluated in 3 months. Standard World Health Organization (WHO) criteria were used to determine response [13]. Tumor volume at the beginning of CRT was compared to the volume at the end of this treatment and this difference was used to calculate the response rate. Survival rates were calculated by using the Kaplan-Meier method.
Results
Total of 22 consecutive patients were treated with this treatment protocol. Patient characteristics are given in Table 1. Median age was 60 years (range, 48–75) and 68.2% were male. Median KPS score was 90 (range, 70–100). This group had a median 43 pack-years smoking history (range, 0–90). Fifty percent of the patients had stage IIIA disease. Twelve patients (54.5%) had squamous cell histology. Eighteen patients had induction CT and median number of cycles was 3 (range, 2–8). One patient received 8 cycles, 3 patients received 6 cycles, 3 patients received 4 cycles of induction CT. After induction CT, 9 patients (40.9%) had partial response (PR), 5 (22.7%) had stable disease (SD) and 4 (18.2%) had progressive disease (PD). The patient who had 8 cycles of induction CT received 4 cycles of two different CT regimens and had PD after both. Two of the other three patients who received 6 cycles of induction CT had SD and 1 had PR.
Conformal 3D-RT planning was used in 63.6% of the patients. Median RT dose administered was 60 Gy (range, 32–66 Gy) with 1.8 Gy (range, 1.8–2 Gy) daily fraction in a median of 46 days (range, 29–64 days) over 7 weeks (range, 4–9 weeks). Elective nodal irradiation was given in 6 patients (27.3%). Rest of the patients (72.7%) received RT only to tumor and involved hilar and mediastinal lymph nodes (Table 2).
G was given for a median of 5 weeks (range, 1–9). Twelve patients (54.6%) received all planned CT. G was stopped in 8 patients, because of intolerance (grade 2–3 fatigue, nausea, vomiting, esophagitis and low CBC for 2 consecutive weeks) in 6 and death during treatment in 2. One patient who died of pneumonia had asymptomatic congestive heart failure and 52.5 pack-years smoking history. Another patient died with acute onset of loss of consciousness, hypotension, breathing depression, and this was considered to be secondary to cerebrovascular accident. Both deaths were seen on the 5th week of treatment. In 2 patients G was held in the middle of the treatment for 1 week because of grade 2 hematological toxicity. Neutropenia and anemia were the most significant toxicities observed (Table 2 and 3). There were only 4 patients (18.2%) who had acute grade 3 esophagitis, while no other grade 3 acute toxicity was seen. These 3 patients had recovered completely after completion of RT. Seven patients (31.8 %) had radiation pneumonitis; four (18.2%) of these had grade 1, two (9.1 %) had grade 2 and one patient had grade 5 (4.5%) (Table 3). One other patient was not assessed because of new brain metastasis was seen right after the completion of her CRT. All patients with signs of radiation pneumonitis on their CTS received corticosteroid treatment, according to our centers' protocol.
One patient (4.5%) had clinical CR, 18 (81.8%) had PR, while 2 (9.5%) had SD after CRT. Two patients who had PR were operated (1 had lobectomy, the other pneumonectomy). Two other patients received adjuvant CT after completion of CRT (Table 3).
Median follow-up was 15 months (range, 2–40 months). Median overall survival (OS) was 14 months (Figure 1). One- and 2-year OS rates were 54.5% and 37.5%. Sixteen patients (72.7%) died, 6 (27.3%) with progression of their primary NSCLC, 8 due to metastatic disease (brain metastasis in 6, contralateral lung metastasis in 2), 1 with pneumonia and another with cerebrovascular event. Median disease free survival (DFS) was 14 months, while 1- and 2-year DFS rates were 55.5% and 26.5%, respectively. Seven patients (31.8%) progressed locally, while 8 other (36.4%) had distant metastasis during follow-up after CRT. Another patient had both local progression and distant metastasis. Five of the patients with local progression had stage IIIB disease initially. First site of metastasis seen in this group was brain in 4, contralateral lung in 3, skin in 1 and neck in another patient. Local control was achieved in 63.6% of the patients. One- and 2-year local control rates were 78.6% and 51%, respectively (Table 4).
Discussion
For many years the mainstay of treatment for non-resectable NSCLC was RT alone. The results of several phase II studies provided support for the addition of CT to RT and a landmark trial was reported by Dillman et al in 1990 [14]. He demonstrated increased rates of three-year and long-term survival [15]. Two large meta-analyses have provided support for the benefit of CRT combinations [3,4].
The initial phase I and II clinical experience in NSCLC had demonstrated that weekly G (600–1000 mg/m2) with concurrent standard thoracic RT was effective, but it had significant toxicity (pulmonary fibrosis, esophagitis, thrombocytopenia and neutropenia) [16-21]. Doses of G used in these preliminary phase II studies were not derived from well-planned phase I studies and were very high. Also significant number of patients had dose reductions or they missed doses during the treatment. After these unclear results dose escalating trials of G as a radiosensitizer with concurrent RT were started [22,23]. In one of the first studies Gonzalez et al recommended that G can be used at 300 mg/m2/week dose with 50 Gy of RT, but the dose of RT was lower than standard in this trial [16]. McMullen et al reported in their phase I/II trial (CALGB 89805) that 70 mg/m2/week was the recommended dose of G with 60–74 Gy RT [23]. Trodella et al recommended weekly G dose of 350 mg/m2 for 5 weeks concurrent with 50.4 Gy RT after their phase I trial [24].
Our rationale for evaluating weekly 75 mg/m2 dosing was based upon a phase I study of Blackstock et al [22]. In that study they suggested that thoracic RT and G given with a dose of weekly 70 mg/m2 in two days appeared tolerable in patients with advanced NSCLC. Although this was a phase I study, overall response rate for 16 patients was 88% and median survival for all 17 patients was 16 months. Dose limiting toxicities were pulmonary (pneumonitis, pulmonary fibrosis), esophageal (esophagitis, esophageal stricture, ulceration), and hematological (anemia, thrombocytopenia) toxicities in this CALGB trial [22]. They delivered almost all the planned G doses (range, 7–12). Several preclinical [25-27] and clinical [28,29] trials have also shown G to be an effective RT enhancer, even when low doses are administered.
In this study the RT planning was changed from 2D conventional to 3D conformal multiple fields planning after the first 6 patients. G given concurrently with 3D-RT is much better tolerated than with 2D-RT and a reduction in grade 3 esophagitis was seen, as it was reported by Zinner et al [30]. The maximal tolerated dose of G given concurrently 7 weeks on row with RT as a radiosensitizer was 190 mg/m2/week with 3D planning while 125 mg/m2/week could be administered with 2D-RT in that study. In our study the differences in the rates of grade 3 or 4 esophagitis and pneumonia between the two treatment schedules were not significant. This might be due to having small number of patients in our study. We also changed elective nodal irradiation to involved nodal irradiation after the first 6 patients. Emami et al concluded in their analysis of the RTOG data that elective irradiation of uninvolved supraclavicular and mediastinal nodes may not be necessary in the treatment of unresectable NSCLC [31].
Blackstock et al also reported that there appeared to be a relationship between volume of lung irradiated, pre-treatment pulmonary function and pulmonary fibrosis, although the data was limited in their phase I trial [22]. Pre-treatment volumes of the tissue irradiated was not calculated in our patient population, but they were much larger than 2000 cm3 which was reported as a cut off in inclusion criteria in the trial reported by Price et al. [28]. Therefore we favored to administer a much lower dose of G then what was used in the above trial.
G and cisplatin have also been studied with concurrent RT. Phase II data have been reported during the last years. Vokes et al recommended 600 mg/m2/day, day 1 and 8 of G and 80 mg/m2/day of cisplatin repeated every 21 days × 2 courses during RT (66 Gy) and this was the dose used in CALGB 9431 [32]. There was a marked increase in the toxicities of this schedule, especially in hematological toxicities and esophagitis. Although this was a phase II study, median survival was reported as 18.3 months which is 1–2 months longer than other concurrent CRT studies. In several randomized phase III trials comparing chemotherapy (either sequential or concomitant) with RT alone, the median survival was found between 8–18 months [3]. In our study median survival was 14 months. Although the survival data's extracted from phase II studies should not be used for comparison, this survival difference between our pilot study and other phase II data might be secondary to using lower dose of G in our study, having 2 very early deaths, lower dose of RT administered, not using cisplatin, treating patients with more tumor load or nature of the disease. But the median survival in our study was similar to the phase III data [3].
In our study there might be two reasons for the high incidence of radiation pneumonitis. The first reason might be the high number of cycles of induction CT. Three of the four patients receiving 6 or more cycles of induction had grade 1 or 2 pneumonitis. Although chemotherapeautic agents were found to induce radiation pneumonitis either alone or with radiation therapy in previous studies this was not seen in our patients [33-35]. The other reason might be the screening schedule of pneumonitis in our centre. All of the patients were considered for radiation pneumonitis during the first CTS examination for the tumor response. The patients received corticosteroid therapy for radiation pneumonitis according to the CTS images but most of them were asymptomatic.
The patients could not receive all planned G dose concurrent with RT. Heavily pre-treated patients with induction CT before CRT might have not tolerated weekly planned G schedule. Eighteen of the patients had induction CT, the median number of cycles were 3 and seven of the patients received ≥4 cycles. The reasons for this were that patients either progressed during induction CT and received second line CT regimens or longer induction CT was their physicians decision before referring them to our centre for RT.
Conclusion
G might be used as radiosensitizer for patients with stage III NSCLC who could not receive full doses CT with concurrent RT. In order to decrease the therapy associated toxicities smaller irradiation volumes used with 3D conformal techniques might be better. We need prospective randomized trials with increased number of subjects in order to figure out where this regimen exactly stands in the standard treatment of this group of patients.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
UA, PFY, HC, MS, NST designed the study, followed the patients, collected the data, performed the statistical analysis and drafted the manuscript. 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 Overall survival plot for all patients.
Table 1 Patient characteristics
Characteristics Number (n) Percentage (%)
Gender
Male 15 68.2
Female 7 31.8
Stage
IIIA 11 50
IIIB 11 50
KPS
70 4 18.2
80 5 22.7
90 12 54.5
100 1 4.5
Histology
Adenocarcinoma 6 27.3
Squamous cell carcinoma 12 54.5
Adenosquamous cell carcinoma 1 4.5
Unclassified 3 13.6
ICT
Yes 18 81.8
No 4 18.2
ICT Regimens
PE 4 22.2
NP or NC 6 33.3
GP 3 16.7
INC 2 11.1
TC 2 11.1
INC + TC 1 5.6
Number of ICT Cycles
2–3 11 61.1
≥4 7 38.9
Response to ICT
PR 9 50
SD 5 27.8
PD 4 22.2
ICT: Induction Chemotherapy, PE: Cisplatin-Etoposide, NP: Vinorelbine-Cisplatin, NC: Vinorelbine-Carboplatin, GP: Gemcitabine-Cisplatin, INC: Ifosfamide-Navelbine-Carboplatin, TC: Paclitaxel-Carboplatin, PR: Partial Response, SD: Stable Disease, PD: Progressive Disease
Table 2 Treatment characteristics of the patients
Variable Number (n) Percentage (%)
RT Planning
Conventional 8 36.4
Conformal 14 63.6
RT Dose/Fraction
1.8 Gy 13 59.1
2 Gy 9 40.9
Number of RT Treatment Weeks
4 1 4.5
5–7 16 72.7
≥8 5 22.7
Number of G Treatment Weeks
≤4 7 31.8
5–7 13 59.1
≥8 2 9.1
RT: Radiotherapy, G: Gemcitabine
Table 3 Non-hematological and hematological toxicities for CRT and responses
Variable Number (n) Percentage (%)
Pneumonia 7 31.8
Grade 1 4 18.2
Grade 2 2 9.1
Grade 5 1 4.5
Nausea/Vomiting 4 18.2
Grade 1 2 9.1
Grade 2 2 9.1
Esophagitis 16 72.8
Grade 1 8 36.4
Grade 2 4 18.2
Grade 3 4 18.2
Hematological toxicity 5 22.7
Grade 1 3 13.6
Grade 2 2 9.1
Response to CRT
Clinical CR 1 4.5
PR 18 81.8
SD 2 9.1
Unknown 1 4.5
CRT: Chemoradiotherapy, CR: Complete Response, PR: Partial Response, SD: Stable disease
Table 4 Overall survival, progression-free survival and local control data
Months ± SD (95% Confidence Interval)
Median OS 14 ± 5 (3–25)
Median PFS 14 ± 3 (8–20)
1-year (%) 2-year (%)
OS Rates 55 38
PFS Rates 56 27
Local Control Rates 79 51
OS: Overall survival, PFS: Progression-free survival
==== Refs
Greene FL and others Lung American Joint Committee on Cancer Cancer Staging Handbook 2002 6 New York: Springer-Verlag 189 203
Non-small Cell Lung Cancer Collaborative Group Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomized clinical trials BMJ 1995 311 899 909 7580546
Pritchard RS Anthony SP Chemotherapy plus radiotherapy compared with radiotherapy alone in the treatment of locally advanced, unresectable, non-small-cell lung cancer: A meta-analysis Ann Intern Med 1996 125 723 729 8929005
Marino P Preatoni A Cantoni A Randomized trials of radiotherapy alone versus combined chemotherapy and radiotherapy in stages IIIA and IIIB Nonsmall cell lung cancer: A meta-analysis Cancer 1995 76 593 601 8625152
Furuse K Fukuoka M Kawahara M Nishikawa H Takada Y Kudoh S Katagami N Ariyoshi Y Phase III study of concurrent versus sequential thoracic radiotherapy in combination with mitomycin, vindesine, and cisplatin in unresectable stage III non-small-cell lung cancer J Clin Oncol 1999 17 2692 2699 10561343
Curran WJ Scoot C Langer C Komaki R Lee JS Hauser S Movsas B Wasserman T Rosenthal S Byhardt R Sause W Cox J Phase III comparison of sequential vs. concurrent chemoradiation for patients with unresected stage III non-small cell lung cancer. Initial report of RTOG 9410 [abstract] Proc Am Soc Clin Oncol 2000 19 A1891
Schaake-Koning C Van Den Bogaert W Dalesio O Festen J Hoogenhout J Van Houtte P Kirkpatrick A Koolen M Maat B Nijs A Effects of concomitant cisplatin and radiotherapy on inoperable non-small-cell lung cancer N Engl J Med 1992 326 524 530 1310160
Jemal A Murray T Samuels A Ghafoor A Ward E Thun MJ Cancer statistics, 2003 CA Cancer J Clin 2003 53 5 26 12568441
Choy H Combination chemoradiotherapy with gemcitabine: Potential applications Oncology (Williston Park) 2000 14 20 25 10960941
Fields MT Eisbruch A Normolle D Orfali A Davis MA Pu AT Lawrence TS Radiosensitization produced in vivo by once- vs. twice-weekly 2'2'-difluoro-2'-deoxycytidine (gemcitabine) Int J Radiat Oncol Biol Phys 2000 47 785 791 10837965 10.1016/S0360-3016(00)00447-8
Milas L Fujii T Hunter N Elshaikh M Mason K Plunkett W Ang KK Hittelman W Enhancement of tumor radioresponse in vivo by gemcitabine Cancer Res 1999 59 107 114 9892194
National Cancer Institute Common Toxicity Criteria 1988 Bethesda, MD
Miller AB Hogestraeten B Staquet M Winkler A Reporting results of cancer treatment Cancer 1981 47 207 214 7459811
Dillman RO Seagren SL Propert KJ Guerra J Eaton WL Perry MC Carey RW Frei EF 3rdGreen MR A randomized trial of induction chemotherapy plus high dose radiation versus radiation alone in stage III NSCLC N Engl J Med 1990 323 940 945 2169587
Dillman RO Herndon J Seagren SL Eaton WL JrGreen MR Improved survival in stage III NSCLC: Seven year follow up of Cancer and Leukemia Group B (CALGB) 8433 trial J Natl Cancer Inst 1996 88 1210 1215 8780630
Gonzalez E Sanchez-Rovira P Jaen A Porras I Medina B Duenas R Mohedano N Fernandez M Lozano A Phase I study concomitant gemcitabine (G) with radiotherapy (RT) in stage III NSCLC [abstract] Eur J Cancer 1999 35 A1054
Zinner R Fossella FV Komaki R Delclos M Jung M Bonnie G Jin L Pisters K Jhingran A Waun H Phase I trial of concurrent gemcitabine (G) plus chest radiotherapy (XRT), followed by consolidation systemic chemotherapy with G plus cisplatin (CDDP), for patients with stage III and medically inoperable stage II non-small cell lung cancer (NSCLC) [abstract] Proc Am Soc Clin Oncol 1999 18 A1798
Gregor A Price A Van Der Leest AHD Scalliet P Wets M Jungnelius U Lemmens JP Groen HJ Phase I study of gemcitabine and radiotherapy in stage III non-small cell lung cancer [abstract] Proc Am Soc Clin Oncol 1999 18 A1953
Athanasiou E Cyrgias G Boinis K Samantas E Christodoulidou M Tamvaki F Panoussaki E Concurrent chemoradiotherapy with Gemcitabine for locally advanced non-small cell lung cancer (NSCLC): Preliminary results of a phase II study [abstract] Ann Oncol 1998 9 A468
Scalliet P Goor C Galdermans D Van Meerbeek J Groen HJ Van de Leest AHD Westerink H Jungnelius U Turrisi A Gemzar (Gemcitabine) with thoracic radiotherapy- A phase II pilot study in chemonaive patients with advanced non-small cell lung cancer (NSCLC) [abstract] Proc Am Soc Clin Oncol 1998 17 A1923
Vokes EE Leopold KA Herndon JE Crawford J Perry MC Miller AA Green MR A randomized phase II study of gemcitabine or paclitaxel or vinorelbine with cisplatin as induction chemotherapy (Ind CT) and concomitant chemoradiotherapy (XRT) for unresectable stage III non-small cell lung cancer (NSCLC) (CALGB Study 9431) [abstract] Proc Am Soc Clin Oncol 1999 18 A1771
Blackstock AW Lesser GJ Fletcher-Steede J Case LD Tucker RW Russo SM White DR Miller A Phase I study of twice-weekly gemcitabine and concurrent thoracic radiation for patients with locally advanced non-small-cell lung cancer Int J Radiat Oncol Biol Phys 2001 51 1281 1289 11728688 10.1016/S0360-3016(01)01732-1
McMullen KP Miller AA Case LD Lesser G Culbreth M Patton S Tucker R Blackstock AW Thoracic radiation and concurrent gemcitabine: A phase I/II study in patients with advanced non-small cell lung cancer [abstract] Int J Radiat Oncol Biol Phys 2002 54 A2150
Trodella L Granone P Valente S Turriziani A Macis G Corbo GM Margaritora S Cesario A D'Angelillo RM Gualano G Ramella S Galetta D Cellini N Phase I trial of weekly gemcitabine and concurrent radiotherapy in patients with inoperable non-small-cell lung cancer J Clin Oncol 2002 20 804 810 11821464 10.1200/JCO.20.3.804
Shewach DS Lawrence TS Radiosensitization of human tumor cells by gemcitabine in vitro Semin Oncol 1995 22 68 71 7481848
Blackstock AW Kwock L Hess SM Leadon SA Tepper JE Gemcitabine: In vitro and in vivo evidence of its radiation sensitizing activity and studies using fluorine-19 magnetic resonance spectroscopy to determine an optimal dosing schedule – preclinical observations relevant to gemcitabine clinical trials [abstract] Int J Radiat Oncol Biol Phys 1997 39 A141
Lawrence TS Eisbruch A Shewach DS Gemcitabine mediated radiosensitization Semin Oncol 1997 24 S24 S28 9122731
Groen H Gregor A Van Putten J Van Der Leest A Little F Jungnelius U Wets M Price A Phase I of gemcitabine (G) and high dose thoracic radiation (RT) in stage III non-small lung cancer (NSCLC) [abstract] Proc Am Soc Clin Oncol 2000 19 A2123
Blackstock AW Bernard SA Richards F Eagle KS Case LD Poole ME Savage PD Tepper JE Phase I trial of twice weekly gemcitabine and concurrent radiation in patients with advanced pancreatic cancer J Clin Oncol 1999 17 2208 2212 10561277
Zinner R Fossella F Komaki R Jung M Jhingran A Glisson B Pisters K Stevens C Khuri F Kurie J Lee J Herbst R Liao Z Honget W Gemcitabine with concurrent 3D conformal radiation followed by consolidation gemcitabine plus cisplatin; a phase I trial for patients with stage III non-small cell lung cancer (NSCLC) [abstract] Proc Am Soc Clin Oncol 2000 19 A1996
Emami B Mirkovic N Scott C Byhardt R Graham MV James Andras E John M Herskovic A Urtasun R Asbell SO Perez CA Cox J The impact of regional nodal radiotherapy (dose/volume) on regional progression and survival in unresectable non-small cell lung cancer: an analysis of RTOG data Lung Cancer 2003 41 207 214 12871784 10.1016/S0169-5002(03)00228-9
Vokes EE Herndon JE 2ndCrawford J Leopold KA Perry MC Miller AA Green MR Randomized phase II study of cisplatin with gemcitabine or paclitaxel or vinorelbine as induction chemotherapy followed by concomitant chemoradiotherapy for stage IIIB non-small-cell lung cancer: Cancer and Leukemia Group B Study 9431 J Clin Oncol 2002 20 4191 4198 12377962 10.1200/JCO.2002.03.054
Roychowdhury DF Cassidy CA Peterson P Arning M A report on serious pulmonary toxicity associated with gemcitabine-based therapy Invest New Drugs 2002 20 311 315 12201493 10.1023/A:1016214032272
Choy H Safran H Akerley W Graziano SL Bogart JA Cole BF Phase II trial of weekly paclitaxel and concurrent radiation therapy for locally advanced non-small cell lung cancer Clin Cancer Res 1998 4 1931 1936 9717821
Thomas AL Cox G Sharma RA Steward WP Shields F Jeyapalan K Muller S O'Byrne KJ Gemcitabine and paclitaxel associated pneumonitis in non-small cell lung cancer: report of a phase I/II dose-escalating study Eur J Cancer 2000 36 2329 2334 11094306 10.1016/S0959-8049(00)00341-5
|
16000167
|
PMC1183193
|
CC BY
|
2021-01-04 16:03:07
|
no
|
BMC Cancer. 2005 Jul 6; 5:71
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-71
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-741600197310.1186/1471-2407-5-74Research ArticleResponse rate of fibrosarcoma cells to cytotoxic drugs on the expression level correlates to the therapeutic response rate of fibrosarcomas and is mediated by regulation of apoptotic pathways Lehnhardt Marcus [email protected] Ludger [email protected] Cornelius [email protected] Heinz Herbert [email protected] Adrien [email protected] Hans Ulrich [email protected] Sonja [email protected] Laura [email protected] Lars [email protected] Oliver [email protected] Department of Plastic Surgery, Burn Center, Hand surgery, Sarcoma Reference Center, BG University Hospital Bergmannsheil, Ruhr University Bochum, Bürkle de la Camp Platz 1, 44789 Bochum, Germany2 Institute of Cell Biology (Tumor Research), IFZ, University of Essen, Virchowstr. 173, 45122 Essen, Germany3 Institute of Pathology, BG University Hospital Bergmannsheil, Ruhr University Bochum, Bürkle de la Camp Platz 1, 44789 Bochum, Germany4 Tumor Genetics Group, Max-Planck-Institut für molekulare Physiologie, Otto Hahnstr. 11, 44227 Dortmund, Germany2005 7 7 2005 5 74 74 27 12 2004 7 7 2005 Copyright © 2005 Lehnhardt 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
Because of the high resistance rate of fibrosarcomas against cytotoxic agents clinical chemotherapy of these tumors is not established. A better understanding of the diverse modes of tumor cell death following cytotoxic therapies will provide a molecular basis for new chemotherapeutic strategies. In this study we elucidated the response of a fibrosarcoma cell line to clinically used cytostatic agents on the level of gene expression.
Methods
HT1080 fibrosarcoma cells were exposed to the chemotherapeutic agents doxorubicin, actinomycin D or vincristine. Total RNA was isolated and the gene expression patterns were analyzed by microarray analysis. Expression levels for 46 selected candidate genes were validated by quantitative real-time PCR.
Results
The analysis of the microarray data resulted in 3.309 (actinomycin D), 1.019 (doxorubicin) and 134 (vincristine) probesets that showed significant expression changes. For the RNA synthesis blocker actinomycin D, 99.4% of all differentially expressed probesets were under-represented. In comparison, probesets down-regulated by doxorubicin comprised only 37.4% of all genes effected by this agent. Closer analysis of the differentially regulated genes revealed that doxorubicin induced cell death of HT1080 fibrosarcoma cells mainly by regulating the abundance of factors mediating the mitochondrial (intrinsic) apoptosis pathway. Furthermore doxorubicin influences other pathways and crosstalk to other pathways (including to the death receptor pathway) at multiple levels. We found increased levels of cytochrome c, APAF-1 and members of the STAT-family (STAT1, STAT3), while Bcl-2 expression was decreased. Caspase-1, -3, -6, -8, and -9 were increased indicating that these proteases are key factors in the execution of doxorubicin mediated apoptosis.
Conclusion
This study demonstrates that chemotherapy regulates the expression of apoptosis-related factors in fibrosarcoma cells. The number and the specific pattern of the genes depend on the used cytotoxic drug. The response rates on the gene expression level, i.e. the number of genes regulated by the drugs actinomycin D, doxorubicin and vincristine, correlate to the clinical effectiveness of the drugs. Doxorubicin seems to exert its cytotoxic mechanism by regulating genes, which are involved in several different apoptosis regulating pathways. The exact knowledge of the genes affected by the drugs will help to understand the diverse modes of soft tissue sarcoma cell death in response to cytotoxic therapies.
==== Body
Background
With only 1% of all solid malignancies and more than 50 sub entities soft tissue sarcomas are rare and heterogeneous [1,2]. Adequate complete surgical resection in combination with radiotherapy is the mainstay of therapy resulting in a 50% to 80% 5-year survival rate. Recent chemotherapy studies revealed a high fraction of resistant soft tissue sarcoma. Response rates above 15% [3-5] within all variant histological subtypes were reported only for doxorubicin (adriamycin), actinomycin D and ifosfamide [6,7].
Generally, two main classes of drug resistance can be distinguished. Tumor cells are either primarily resistant to chemotherapeutic drugs (intrinsic resistance), or some of the cells respond to chemotherapy at the first treatment but the remaining cells recur later to form a multidrug-resistant tumor (acquired resistance) [8]. A single mechanism or a distinct pathway cannot explain the effectiveness of a cancer drug. In carcinomas multiple mechanisms of drug resistance have been characterized on the molecular level [9,10]. These include the overexpression of the genes p53 [11-14], MDR1 (multidrug resistance gene 1) [14-16], MRP1 (multidrug resistance-associated protein), or the induction of DNA repair [14]. In addition various regulatory genes targeted for genetic alterations during tumorigenesis may also influence cellular sensitivity to chemotherapeutic drugs. These genetic alterations involve tumor suppressor genes, oncogenes, cell cycle regulators, transcription factors, growth factor receptors, DNA repair factors and cell death regulators. Only little is known about the molecular basis of drug resistance in soft tissue sarcomas [17-20]. Comprehensive knowledge of the expression changes induced by cytostatic drugs should be useful for examining the molecular basis of drug resistance. The specific expression and response profiles of a cell line established from a distinct tumor may ultimately allow to design improved therapeutic regimes with the aim to circumvent drug resistance [20]. In this study, we used Affymetrix microarrays to monitor mRNA expression changes in an established fibrosarcoma cell line. The cells were treated with the two widely used cytotoxic drugs doxorubicin and actinomycin D. In a parallel analysis, cells were treated with vincristine as a cytostatic drug of very low response in human soft tissue sarcomas for comparison.
Methods
Cell Culture and RNA-preparation
HT-1080 (human fibrosarcoma cells, cell line CCl 121 from ATCC) were cultured in modified Eagle's medium supplemented with 10% FCS in 15 cm Petri dishes. Semi-confluent cultures were treated with 0,5 μg/ml doxorubicin for 6 h or 24 h, 0.1 μg/ml actinomycin D or 0.4 μg/ml vincristine for 24 h. Total RNA was purified from the cells using Trizol reagent (Life Technologies), as specified by the manufacturer. RNA integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies).
Oligonucleotide microarray analysis
For microarray analyses we used the Affymetrix Gene Chip platform employing a standard protocol for sample preparation and microarray hybridization that has been described in detail previously [20,21]. Briefly, total RNA was converted into double-stranded cDNA using an oligo-deoxythymidine primer containing the T7 RNA polymerase binding site (5'- GCATTAGCGGCCGCGAAATTAATACGACTCACTATAGGGAGA – (dT)21V-3') (MWG Biotech) for first strand synthesis. After generation of double-stranded cDNA from the first-strand cDNA, biotinylated cRNA was synthesized by in vitro transcription using the BioArray High Yield RNA Transcript Labeling Kit (Enzo Diagnostics). Labeled cRNA was purified on RNeasy columns (Qiagen), fragmented and hybridized to HG-U133A microarrays (Affymetrix). The arrays were washed and stained according to the manufacturer's recommendation and finally scanned in a GeneArray scanner 2500 (Agilent).
Array images were processed to determine signals and detection calls (Present, Absent, Marginal) for each probeset using the Affymetrix Microarray Suite 5.0 software (MAS 5.0; statistical algorithm). To determine expression changes induced by doxorubicin and vincristine, arrays were scaled across all probesets to an average intensity of 1000 to compensate for variations in the amount and quality of the cRNA samples and other experimental variables of non-biological origin. To determine target genes of actinomycin D, which induced rather strong alterations in gene expression, scaling to externally added polyA+ spike controls proved to be more approriate. Pairwise comparisons of treated versus control samples were carried out with MAS 5.0, which calculates the significance (change p-value) of each change in gene expression based on a Wilcoxon ranking test. To limit the number of false positives, we restricted further target identification to those probesets, which received at least one present detection call in the treated/control pair. Probesets exhibiting a signal log2 ratio >1.32 and a change p-value <0.001 or a signal log2 ratio <-1.32 and a change p-value >0.999 (corresponding to 2.5 fold up- or down-regulation) were identified by filtering using the Affymetrix Data Mining Tool 3.0 (Table 1). The normalized raw data set of this analysis can be found at the Global Environment Outlook portal at http://www.ncbi.nlm.nih.gov/geo (GEO accession code: GSE2238).
Real-time RT-PCR and data processing
Real-time RT-PCR was performed using Taqman low density arrays on a ABI PRISMR 7900HT Sequence Detection System (Applied Biosystems). cDNA was generated with random primers using the High Capacity cDNA Archive Kit (Applied Biosystems) using 5 μg total RNA in a 20 μl reaction volume. Approximately 400 ng cDNA was used per port (48 wells) representing 24 different Assays-on-Demand in two replicates. Transcript levels were determined in duplicate reactions by the ΔΔCt method using 5 transcripts as an internal reference [see Additional file 1].
GO and pathway analysis
Statistically overrepresented Gene Ontology (GO) terms within target gene lists were identified by the hypergeometric distribution implemented in the GO Browser of Spotfire DecisionSite for Functional Genomics (Spotfire). Specific signaling pathways were identified using the software package Pathway Assist™ (Stratagene).
Results
Microarray analysis revealed the differential gene expression patterns of fibrosarcoma cells treated with cytostatic drugs. Based on the Affymetrix statistical change and detection algorithm, a stringent combination of filter criteria was used to identify significantly changed genes, while keeping the false postive rate at a minimum (Table 1; see Material and Methods for details). As indicated in Figure 1, the three different drugs proved to affect remarkably different numbers of probesets. After actinomycin D treatment for 24 hours, 3309 probesets showed differential expression levels. Of these, a vast majority of 99.4% were under-represented, confirming the general inhibitory activity of actinomycin D on RNA polymerization. In comparison, treatment with doxorubicin for 24 hours affected a lower number of probesets (1019) than actinomycin D, while 62.6 % of these probesets were up-regulated. Treatment with doxorubicin for 6 hours resulted in just 3 differentially expressed targets, indicating that a shorter incubation period leads to a minor response. Vincristine treatment for 24 hours led to the identification of 134 differentially regulated probesets. Of these, 87 probesets were increased and 47 decreased. Interestingly, the numbers of probesets, which were regulated by the different drugs, correlate with the clinical effectiveness of these three cytostatics. Vincristine, the agent with the lowest impact on gene expression has the weakest effect on the growth of fibrosarcomas in-vivo [22], while actinomycin D and especially doxorubicin appeared to have a much stronger effect on cell growth in soft tissue sarcoma [23,24].
The Venn diagram shows that there is only a limited overlap of the genes affected by the three different drugs (Figure 1). Thirteen probesets representing 12 genes were common targets of all three drugs; however, while these targets were all down-regulated by actinomycin D, doxorubicin and vincristine consistently up-regulated 4 of those transcripts. Analyzing the up- and down-regulated transcripts in more detail (Figure 1B), it is evident that 43 out of the 44 probesets common to doxorubicin and vincristine target lists are affected in the same direction (29 up, 14 down). Vincristine and actinomycin D targets lists share 33 targets, but concordant regulation was restricted to 18 down-regulated probesets. Similarly, doxorubicin and actinomycin D shared 264 common targets, while only 3 up- and 208 down-regulated transcripts were affected in the same direction. The scatter plots and moreover the dendogramm illustrate these quantitative findings (Figures 3, 4).
To get an overview of the biological processes affected by the three different drugs, we classified the regulated probesets to the various categories as defined by the gene ontology annotations (Table 2) [25]. Analyzing the distribution of the targets to the various categories, we found a significant overrepresentation in several categories that depended on the drug used.
The biochemical functions of the genes in the expression profile of the actinomycin D-treated fibrosarcoma cells are diverse and include cell proliferation, cell cycle, metabolism and others. Altered expression of the metabolism genes especially include nucleotide metabolism (Table 2, column 1). Increased expression of cell proliferation and cell cycle genes were found after doxorubicin treatment (column 2) while cells treated with vincristine significantly overexpressed metabolism (especially nucleotide metabolism) genes (column 3).
Figure 4 represents pathways analyses based on decreased and increased genes detected after 24 h treatment with doxorubicin. Because killing of cancer cells by cytotoxic therapies is predominantly mediated by triggering apotosis in target cells, we focused on differential expressed genes involved in cell death pathways.
There are two known cellular pathways, which lead to the initiation of apoptosis [9,10]. The 'extrinsic/exogenous' pathway (activated by death receptors, a subgroup of the TNF receptor super family) and the 'intrinsic/endogenous' pathway (a mitochondrial pathway controlled by members of the Bcl-2 family) [26]. Most cytotoxic drugs, which activate apoptosis, lead to the activation of the mitochondrial pathway. The expression of Bcl-2 may be responsible for resistance to apoptosis induced by cytotoxic drugs, and down-regulation of Bcl-2 sensitizes cells to doxorubicin [27]. Doxorubicin-induced cell death is mediated by the tumor suppressor p53. Remarkably, none of the drugs tested in this study changed the expression level of p53. Nevertheless, the drugs showed an influence on the expression levels of other important genes, which are involved in the intrinsic pathway (Figures 4, 5, 6). Cytochrome c expression showed an increased level, and Bcl-2 showed a decreased level after doxorubicin treatment in the fibrosarcoma cell line (Figure 4).
The apoptotic protease activating factor 1 (APAF1) was up-regulated. APAF1 plays a central role in the cell death machinery and is an activator of the caspases-cascade, especially of caspase-9. Several members of the protease family of caspases (cysteine dependent aspartate-specific proteases) play a major role in apoptosis [10]. We found up-regulated expression of caspase-1, caspase-3, caspase-6, caspase-8 and caspase-9 (Figures 4, 5, 6). Increased expression levels were found for two members of the STAT-family (signal transducer and activator of transcription): STAT1 and STAT3, too. These proteins play roles in several different apoptotic pathways (Figures 3A, D). Increased expression of STAT1 triggers tumor progression and leads to down-regulation of apoptosis [28]. STAT3 plays a key role in signal transduction and promotion of apoptosis [29]. In addition the protein is involved in interleukin signaling. Altered expression of the interleukin-family includes up-regulation of IL-6, a mediator of anti-apoptotic effects and drug-resistance mechanisms. In addition to the regulation of apoptosis activating pathways doxorubicin regulates an anti-apoptotic pathway. The insulin-like growth factor 1 receptor (IGF-1R) is an important signaling molecule in cancer cells. IGF-1R was prominently targeted for up-regulation by doxorubicin treatment.
In addition we found expression changes in genes, which are involved in the extrinsic pathway. Fas/APO-1/CD95 is a member of the tumor necrosis factor (TNF) receptor super family. We found altered expression level of different FAS family mediators. PTPN13 (increased) activates directly caspase-8. The Fas-associated factor 1 (FAF1) was down-regulated. FAS and the tumor necrosis factor (TNF) super family of receptors induce apoptosis by recruiting adaptor molecules through death domain interactions. The central adaptor molecule for these receptors is the death domain-containing protein FADD. FADD binds a death domain on a receptor or an additional adaptor and recruits caspases to the activated receptor [30,31]. FADD was up-regulated in this experiment. In contrast we found decreased expressions of FAIM and FAIM2, Fas-associated apoptosis inhibitory molecules.
Looking at non-apoptotic pathways, we found an influence of doxorubicin on the pathway of skeletal myogenesis. PCAF (p300/CBP-associated factor), a co-activator of the tumor suppressor p53, was increased. This finding is in line with the results of others [32-34]. Furthermore the transcription factor MEF2 (Myocyte enhancer factor-2), that controls critical cellular processes including proliferation, differentiation and survival, was increased.
Analyzing the proliferation activating pathways controlled by EGF and Ras we found altered expression of the GTPase activating protein Rasa-1 (Ras p21 protein activator-1). Rasa-1 is involved in the signal transduction pathway of several growth factors and in the down-regulating of different proteins of the Ras GTPase super family [34]. In addition we found the up-regulated expression of SOS1, which encodes for a Ras-specific exchange factor. Furthermore H-RAS was up-regulated, which is a member of the Ras GTPase super family.
To confirm the changes observed by microarray, we measured the expression of 46 representative apoptosis genes that were in part differently expressed in these experiments by quantitative RT-PCR using Taqman low density arrays (for details see Material and Methods). Overall the success rate of validation by RT-PCR was 62%. For securely detected probestes (pp-call) in the doxorubicin-experiment the success rate of validation was 78% [see Additional file 1].
Discussion
Soft tissue sarcomas are very heterogeneous regarding their presentation, growth behavior and response to therapeutic interference. The presentation and behavior of these tumors depend on location and histological characteristics [24]. Standard therapy consists of complete surgical resection in combination with adjuvant radiotherapy. Because of low response rates chemotherapy is largely restricted to clinical trials. Hence, novel chemotherapy protocols that overcome resistance would be highly valuable [3].
In this study we analyzed the differential gene expression pattern of fibrosarcoma cells after treatment with the three cytostatics doxorubicin, actinomycin D and vincristine. First of all we found that the number of significantly altered probe sets directly correlates with the known clinical response rates of the three cytostatics. Treatment with vincristine shows a very low response rate in soft tissue sarcomas (134 altered probe sets). Doxorubicin (1.109 altered probe sets) and actinomycin D (3.309 altered probe sets) are more effective chemotherapeutic drugs with a significantly higher response rate in these tumors (Figure 1) [35]. This correlation suggests that a strong impact on gene expression leads to a high clinical effectiveness. Thus the microarray analysis of the differential gene expression might be a valuable tool to judge the clinical effectiveness of newly developed drug candidates.
The alkaloid vincristine inhibits mitosis by inhibiting tubulin polymerization and thus by preventing metaphase spindle formation. In addition there are hints that vincristine interferes with DNA and RNA synthesis (Figure 1). Because of response rates lower than 5% vincristine is not well established in the clinical treatment of soft tissue sarcomas [36,37]. Vincristine seems to influence the expression of several genes of different pathways rather than of distinct genes of specific pathways.
With response rates between 15% and 20% actinomycin D is one of the most widely used chemotherapeutic agents in the treatment of soft tissue sarcomas [38]. Actinomycin D is known to inhibit cell proliferation in a rather non-specific way by forming a stable complex with double-stranded DNA and thus inhibiting DNA-primed RNA synthesis [39-42]. In conformance with this inhibitory effect on transcription we found 3.289 probe sets decreased and correspondingly only 20 probe sets increased in actinomycin D treated cells (Figure 1). The biochemical functions of these down-regulated genes include cell proliferation, cell cycle, programmed cell death and metabolism (Figure 2). These findings provide a first clue of the complex biological processes affected by the drug both on a quantitative and on a qualitative basis.
Doxorubicin achieves clinical response rates of 20% to 26% and represents one of the most effective chemotherapeutic agents among the available drugs, which are used in first line single drug therapies [43]. Doxorubicin inhibits topoisomerase II and causes DNA damage, at least partially in a p53-dependent manner [44]. There was a marked count of probesets affected by doxorubicin. The regulated sequences include genes involved in cell proliferation and cell cycle (Figure 2). This finding is in line with the results of others and may represent the signature profile of doxorubicin induced gene expression [8].
Understanding the exact molecular events and the resistance mechanisms induced by anticancer therapy would provide new opportunities for pathway-based rational therapy and for drug development [9,10]. To elucidate the mechanisms and the degree of the molecular response we analyzed the intracellular pathways, which are influenced by doxorubicin in a fibrosarcoma cell line. We identified five different pathways, which include at least three genes with altered expression levels. Four of these pathways regulate apoptosis (Figure 4).
Apoptosis in response to cancer therapy proceeds through activation of the core apoptotic machinery including the receptor and the mitochondrial signaling pathway [45-47]. The relative contribution of the receptor and mitochondrial pathways to drug-induced apoptosis has been a subject of controversial discussion. While a number of initial studies suggested that cancer therapy-triggered apoptosis involves activation of the CD95 receptor/ligand system, compelling evidence subsequently indicated that the majority of cytotoxic drugs initiate cell death by triggering the mitochondrial signaling pathway [48]. The data of our study support either possibilities. Doxorubicin showed influences on genes involved in the extrinsic as well as on genes involved in the intrinsic pathway.
One major extrinsic apoptogenic pathway is the cytochrome c/Apaf-1/caspase-9 dependent pathway (Figure 4). We found cytochrome c, Apaf-1, caspase-3, caspase-6, caspase-8 and caspase-9 up-regulated by doxorubicin. The release of mitochondrial apoptogenic proteins, such as cytochrome c, into the cytoplasm leads to Apaf-1 dependent activation of caspase-9, which is a key event in this pathway. Especially Apaf-1, a member of the NOD-family (nucleotide binding oligomerization domain) is known to play a central role in the cell death machinery. Its expression has been demonstrated to be involved in several cell death pathways. Furthermore Apaf-1 is an activator of caspases [49,50]. Apaf-1 inactivation leads to chemoresistance for example in malignant melanoma cells [51]. Additionally, the anti-apoptotic factor Bcl-2, a direct antagonist of p53 and cytochrome c, was down-regulated. Various pro- and anti-apoptotic Bcl-2 family proteins regulate the permeability of the mitochondrial outer membrane, and DNA damage triggers apoptosis through the down-regulation of anti-apoptotic Bcl-2. Yeh and Pusztai demonstrated that Bcl-2 could be responsible for the resistance to apoptosis induced by cytotoxic drugs. Thus the down-regulation of Bcl-2 should sensitize cells to doxorubicin [52-54]. These results might indicate that doxorubicin initiates cell death of HT1080 cells by activating the cytochrome c/Apaf-1/caspase-9 dependent pathway. Our results confirm the results of others showing that doxorubicin triggers Bcl-2 down-regulation, cytochrome c release from mitochondria and the activation of caspases-3 and 9 [55-57]. Doxorubicin changed the expression of several caspases. These findings are in agreement with recently published data about different caspases acting as key factors in doxorubicin mediated cell death [58-62]. Caspases are the key effector molecules in most forms of apoptotic cell death and represent the level, where the intrinsic and the extrinsic apoptotic pathway converge [62]. Additionally they interact with other non-caspases apoptotic pathways [63]. Caspases are categorized into initiator (caspase-2, -8, -9, -10) and effector caspases (caspase-3, -6, -7) [64]. Doxorubicin induced the expression of caspases-1, -3, -6, -8 and -9. Caspase-6 cleavage may feed back to the receptor pathway by cleaving caspase-8 [65,66]. Within the caspases pathway we found two additional genes down-regulated, LMNB1 and ARHGDIB (Figure 5). Little is known about the function of these proteins. They might play a role in a number of processes related to metastases [67,68]. LMNB1 and ARHGDIB are effectors of the caspases-3 and -6, respectively, and are thus also involved in the apoptotic Fas signaling pathway (Figure 6). There are hints that the extrinsic pathway is connected with the sensitivity and with the resistance of tumor cells towards cytotoxic treatment. Chemotherapeutic treatment led to an increase in CD95L expression, which triggered the receptor pathway. In support of this concept, increased levels of CD95L were observed in many different tumor cell lines [9,32,69-71]. Our results indicated that the pathway is influenced by doxorubicin. In addition to the expression of the two genes LMNB1 and ARHGDIB, the expression of the death domain containing protein FADD, PTPN13 (protein tyrosine phosphatase non-receptor type 13) and the expression of the caspases-8, -3 and -6 were regulated. Furthermore we found decreased expression levels of the Fas-associated apoptosis inhibitory molecules FAIM and FAIM2, what is another indication for the activation of the Fas pathway by doxorubicin.
In the intrinsic pathway, various apoptosis inducing signals directly or indirectly change mitochondria membrane permeability and cause release of mitochondrial intermembrane proteins, including cytochrome c [72]. In the cytoplasm, pro-caspase-9, Apaf-1 and cytochrome-c assemble to form the apoptosome [73,74]. In an ATP-dependent process, procaspase-9 is cleaved in the apoptosome. Activated caspase-9 then cleaves and activates down stream caspases, such as caspase-3. Down-stream of the caspase-8 or caspase-9 activation cascade the intrinsic and extrinsic apoptosis pathways converge in the executioner caspases-3, -6 and -7 [75,76]. Once activated, the executioner caspases are responsible for the proteolytic cleavage of a broad range of cellular proteins resulting in cell death [73,74].
In addition to the activation of several different apoptosis regulating pathways doxorubicin activates also pathways that activate proliferation, differentiation and inhibition of apoptosis. The EGF signaling pathway regulates cell proliferation and transformation. Analyzing this pathway we found altered expression of the genes Rasa1 (decreased levels), STAT1 and STAT3 (increased levels). STAT1 has been implicated in modulating pro- and anti-apoptotic genes following several stress-induced responses and therefore is able to act as a tumor suppressor [77]. STAT3 suppresses expression of pro-inflammatory mediators in tumor cells. Blocking STAT3 in tumor cells induces the expression of pro-inflammatory cytokines and chemokines that activate innate immunity and dendritic cells [78]. For example STAT3 protects prostatic cancer cells from apoptosis induced by T lymphocytes [79]. In human neuroblastoma cells doxorubicin leads to STAT3 activation that results in apoptosis inhibition [80].
The pleiotropic cytokine IL-6 exhibits many physiological functions, including host defense, bone metabolism and acute phase responses. A variety of malignant tumors express IL-6. In multiple myeloma cells IL-6 inhibits apoptosis induced by serum starvation, dexamethasone, and Fas [81]. In pre-B cells, however, IL-6 is important for differentiation and apoptosis. A further study demonstrated that IL-6 is responsible for drug resistance and anti-apoptotic effects in prostatic cancer cells. The anti-apoptotic ability of IL-6 is associated with expression of the Bcl-2 family proteins [82]. Interestingly, the IL-6 mediated inhibition of apoptosis is associated with up-regulation of Bcl-2, which is independent of STAT3 [83]. We found increased IL-6 levels after doxorubicin treatment. However there were increased levels of STAT3 in combination with decreased levels of Bcl-2 at the same time. In conclusion the pathway-analyses offer doxorubicin mediated apoptotic and anti-apoptotic mechanisms side by side.
The concept that anticancer therapies primarily act by triggering apoptosis has been challenged, since a consistent link between the ability of tumor cells to undergo apoptosis in vitro and their susceptibility to anticancer therapy in vivo is not essential. Therefore, pathways leading to non-apoptotic cell death, e.g. necrosis or yet unclassified forms of cell death, may mediate the response to cytotoxic therapy [84,85]. Also, caspase independent apoptosis has been found to be induced by anticancer drugs in some cells [86-88]. Furthermore, recent findings have shown that caspase inhibitors do not protect cancer cells from death by cytotoxic agents, but may switch drug-induced apoptosis to an alternative default death e.g. sensescence [89].
Understanding these events and the diverse modes of tumor cell death may provide new opportunities for pathway-based rational therapy and for drug development [90].
The drug response is of course dependent on their concentration. Further studies including different sarcoma cells, drug concentrations and time responses will strengthened the knowledge about the chemotherapy resistance in the very heterogeneous soft tissue sarcomas.
Conclusion
The analysis of gene expression patterns and regulated pathways, which result from treatment with cytostatic agents, could provide a new basis for tumor therapy. This study demonstrates that HT1080-fibrosarcoma cells reveal a specific response pattern of apoptosis-related factors on mRNA level, which depends on the used cytostatic drug. The response rates on the gene expression level, i.e. the number of genes regulated by the drugs, correlate to the clinical effectiveness of the drugs. In addition we found a large number of differentially expressed genes whose biochemical functions include cell proliferation, cell cycle regulators, cell growth, metabolism and others. These findings provide an insight into the complex biological processes affected by the drug and may represent the signature profiles of drug induced gene expression. The knowledge of the genes affected by the drugs in HT1080 fibrosarcoma cells will help to understand the diverse modes of cell death in other soft tissue sarcomas in response to cytotoxic therapies.
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
ML: coordinated the work, developed the study design and prepared the manuscriptLKH: carried out statistical analyses, prepared the manuscript and figures, have given substantial contribution to conception and designCK: prepared the manuscript and figures
HH: carried out cell culture, RNA isolation, microarray and PCR
AD: participated in the study design, prepared the figures and was helpful in preparing the manuscript
HUS: developed the idea, have given final approval of the version to be published
LS: carried out RNA isolation, microarray and PCR
SR: carried out the microarrays
LS: carried out RNA isolation and PCR
OM: developed the idea, study design and conceived the work
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Validation of 46 selected candidate genes by RT-PCR signal log ratio values of 46 selected genes. Microarray and PCR data are displayed for comparison
Click here for file
Figures and Tables
Figure 1 A+B: Patterns of genes affected by Actinomycin D, Doxorubicin and Vincristin. (A) Venn diagram showing the overlap of probesets regulated by the three different drugs in HT1080 cells after 24 h of treatment. Numbers in brackets refer to the total number of affected probesets. (B) Cross table showing the number of concordant probesets up- or down-regulated by the three different compounds.
Figure 2 A-D: Induction of HT1080 cells by doxorubicin, actinomycin D and vincristine. HT1080 cells were treated with 0,5 μg/ml doxorubicin for 6 h or 24 h (A, B), 0.1 μg/ml actinomycin D (C) or 0.4 μg/ml vincristine (D) for 24 h. Expression profiles of the untreated control (x-axis) and the drug-induced (y-axis) HT1080 cells are shown as bivariate scatterplots of 14,500 genes from the microarray. The values are corrected intensities representing levels of expression for the DNA elements of the microarray. Colour legend: red: P-P = present in baseline (x-axis) and present in experiment (y-axis) black: P-A = present in baseline but absent in experiment green: A-P = absent in baseline but present in experiment yellow: A-A, A-M, M-A, M-M = absent or marginal in baseline or experiment
Figure 3 Overall expression patterns of 4137 genes in fibrosarcoma cells treated with doxorubicin (6 h and 24 h), actinomycin D and vincristine (each 24 h). Horizontal rows represent individual genes; vertical columns represent individual samples. Colour range: Brightest red: Signal Log Ratio (SLR)> = 2 (indicates expression level above untreated cells); brightest green: SLR< = 2 (indicates expression level below untreated cells); black: SLR = 0 (indicates unchanged expression). The dendogram at the top of the matrix indicates the degree of similarity between tumor samples; the dendogram at the left side indicates the degree of similarity among the selected genes according to their expression patterns.
Figure 4 Pathway-analyses based on decreased (green) and increased (red) genes detected in fibrosarcoma cells after 24 h treatment with doxorubicin. Results shown for apoptosis. Signaling pathway maps were identified using the software package Pathway Assist. A minimum of 3 interacting genes is the required criteria for creation of a map (Software: Pathway Assist™ Stratagene).
Figure 5 Pathway-analyses based on decreased (green) and increased (red) genes detected in fibrosarcoma cells after 24 h treatment with doxorubicin. Results shown for caspases. Signaling pathway maps were identified using the software package Pathway Assist. A minimum of 3 interacting genes is the required criteria for creation of a map (Software: Pathway Assist™ Stratagene).
Figure 6 Pathway-analyses based on decreased (green) and increased (red) genes detected in fibrosarcoma cells after 24 h treatment with doxorubicin. Results shown for FAS-signalling. Signaling pathway maps were identified using the software package Pathway Assist. A minimum of 3 interacting genes is the required criteria for creation of a map (Software: Pathway Assist™ Stratagene).
Table 1 Hierarchical filter analysis for selection of substantially altered expression levels
Filter 1 change: increased – decreased
Filter 1 change: increased – decreased
Filter 3 SLR*<1.32 and Change p-value >0.999 and at least one Present (P) Detection call in control or treated array
SLR*>1.32 and Change p-value <0.001 and at least one Present (P) Detection call in control or treated array
* SLR: Signal log Ratio
Table 2 Distribution of the probesets affected by the various drugs to categories of the GO biological categories.
actinomycin D doxorubicin vincristine HG-U133A
Total number of probesets in target list 3309 1019 134 22215
Total number of probesets with annotated biological process 2158 702 87 13704
Physiological process
Cell growth and/or maintenance 827*** 262* 40* 4701
Cell proliferation 367*** 123*** 14 1576
Cell cycle 290*** 87*** 10 1085
Death 151*** 39 5 682
Metabolism 1507*** 429 50 8247
Regulation of biological process
Regulation of programmed cell death 87*** 22 2 400
Anti-apoptosis 41*** 5 2 152
Positive regulation of apoptosis 50* 10 0 231
Regulation of signal transduction 49* 13 3 242
I-κB kinase/NF-κB cascade 29* 7 2 123
Regulation of metabolism 509*** 121 22* 2398
Protein biosynthesis 54*** 6 1 156
Nucleobase, nucleoside, nucleotide synthesis 455*** 111 21* 2232
Negative regulation of nucleobase, nucleoside, nucleotide and nucleid acid metabolism 49*** 8 2 180
DNA metabolism 15** 1 0 50
Based on the 'Gene Ontology Standard Vocabulary' the probesets affected by actinomycin D, doxorubicin and vincristine were sorted according to the 'Biological Process' into the indicated major categories. The number of probesets with annotated biological processes is given and corresponds to approximately 66% of the probesets affected. The two major categories investigated are shown in bold and subcategories of the previous process are indicated by offset. As a reference the distribution of all annotated probesets of the HG-U133A array to the same categories are listed. Genes significantly overrepresented in the target lists were identified using the statistics (hypergeometric distribution) implemented in the 'GO browser' of Spotfire DecisionSite for Functional Genomics. *, ** and *** indicate p-values of p < 0.05, p < 0.01, and p < 0.001, respectively.
==== Refs
Singer S Demetri GD Baldini EH Fletcher CD Management of soft-tissue sarcomas: an overview and update Lancet Oncol 2000 1 75 85 11905672 10.1016/S1470-2045(00)00016-4
Chao C McMasters KM Edwards MJ Advances in the treatment of soft-tissue sarcomas J Ky Med Assoc 2002 100 10 16 11831094
Donato Di Paola E Nielsen OS The EORTC soft tissue and bone sarcoma group Eur J Cancer 2002 38 Suppl 4 138 141 10.1016/S0959-8049(01)00444-0
Budach V Budach W Scheulen M Czeglarski G Primary chemosensitivity and secondary drug resistance of xenografted soft part sarcomas Strahlenther Onkol 1989 165 535 536 2473534
Frustaci S Gherlinzoni F De Paoli A Bonetti M Azzarelli A Comandone A Olmi P Buonadonna A Pignatti G Barbieri E Apice G Zmerly H Serraino D Picci P Adjuvant chemotherapy for adult soft tissue sarcomas of the extremities and girdles: results of the Italian randomized cooperative trial J Clin Oncol 2001 19 1238 1247 11230464
Bui BN Tabrizi R Dagada C Trufflandier N St ckle E Coindre JM [Update on soft tissue sarcomas] Bull Cancer 2002 89 100 107 11847031
Frustaci S De Paoli A Bidoli E La Mura N Berretta M Buonadonna A Boz G Gherlinzoni F Ifosfamide in the adjuvant therapy of soft tissue sarcomas Oncology 2003 65 80 84 14586155 10.1159/000073366
Kudoh K Ramanna M Ravatn R Elkahloun AG Bittner ML Meltzer PS Trent JM Dalton WS Chin KV Monitoring the expression profiles of doxorubicin-induced and doxorubicin-resistant cancer cells by cDNA microarray Cancer Res 2000 60 4161 4166 10945624
Debatin KM Krammer PH Death receptors in chemotherapy and cancer Oncogene 2004 23 2950 2966 15077156 10.1038/sj.onc.1207558
Debatin KM Apoptosis pathways in cancer and cancer therapy Cancer Immunol Immunother 2004 53 153 159 14749900 10.1007/s00262-003-0474-8
Soengas MS Alarcon RM Yoshida H Giaccia AJ Hakem R Mak TW Lowe SW Apaf-1 and caspase-9 in p53-dependent apoptosis and tumor inhibition Science 1999 284 156 159 10102818 10.1126/science.284.5411.156
Hostanska K Vuong V Rocha S Soengas MS Glanzmann C Saller R Bodis S Pruschy M Recombinant mistletoe lectin induces p53-independent apoptosis in tumour cells and cooperates with ionising radiation Br J Cancer 2003 88 1785 1792 12771996 10.1038/sj.bjc.6600982
Zhan M Yu D Lang A Li L Pollock RE Wild type p53 sensitizes soft tissue sarcoma cells to doxorubicin by down-regulating multidrug resistance-1 expression Cancer 2001 92 1556 1566 11745235 10.1002/1097-0142(20010915)92:6<1556::AID-CNCR1482>3.0.CO;2-S
Coley HM Drug resistance studies using fresh human ovarian carcinoma and soft tissue sarcoma samples Keio J Med 1997 46 142 147 9339643
Toffoli G Frustaci S Tumiotto L Talamini R Gherlinzoni F Picci P Boiocchi M Expression of MDR1 and GST-pi in human soft tissue sarcomas: relation to drug resistance and biological aggressiveness Ann Oncol 1992 3 63 69 1606072
Komdeur R Molenaar WM Zwart N Hoekstra HJ van den Berg E van der Graaf WT Multidrug resistance proteins in primary and metastatic soft-tissue sarcomas: down-regulation of P-glycoprotein during metastatic progression Anticancer Res 2004 24 291 295 15015610
Kuhnen C Muller KM Steinau HU Lehnhardt M [Therapy-induced tumor regression in adult soft tissue sarcomas-morphological findings] Pathologe 2004 25 437 44 15449026 10.1007/s00292-004-0720-7
Skubitz KM Skubitz AP Differential gene expression in leiomyosarcoma Cancer 2003 98 1029 1038 12942572 10.1002/cncr.11586
Hwang RF Hunt KK Experimental approaches to treatment of soft tissue sarcoma Surg Oncol Clin N Am 2003 12 499 521 12916466 10.1016/S1055-3207(03)00006-1
Schofield D Triche TJ cDNA microarray analysis of global gene expression in sarcomas Curr Opin Oncol 2002 14 406 411 12130925 10.1097/00001622-200207000-00007
Nielsen TO West RB Linn SC Alter O Knowling MA O'Connell JX Zhu S Fero M Sherlock G Pollack JR Brown PO Botstein D van de Rijn M Molecular characterisation of soft tissue tumours: a gene expression study Lancet 2002 359 1301 1307 11965276 10.1016/S0140-6736(02)08270-3
Yap BS Baker LH Sinkovics JG Rivkin SE Bottomley R Thigpen T Burgess MA Benjamin RS Bodey GP Cyclophosphamide, vincristine, adriamycin, and DTIC (CYVADIC) combination chemotherapy for the treatment of advanced sarcomas Cancer Treat Rep 1980 64 93 98 7379060
Sauer H [Adjuvant chemotherapy in early soft tissue sarcoma and palliative chemotherapy in advanced soft tissue sarcoma in adults] Schweiz Rundsch Med Prax 1998 87 1066 1071 9757790
Reichardt P High-dose chemotherapy in adult soft tissue sarcoma Crit Rev Oncol Hematol 2002 41 157 167 11856592
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
Tsujimoto Y Cell death regulation by the Bcl-2 protein family in the mitochondria J Cell Physiol 2003 195 158 167 12652643 10.1002/jcp.10254
Sprick MR Walczak H The interplay between the Bcl-2 family and death receptor-mediated apoptosis Biochim Biophys Acta 2004 1644 125 132 14996497 10.1016/j.bbamcr.2003.11.002
Stephanou A Latchman DS STAT-1: a novel regulator of apoptosis Int J Exp Pathol 2003 84 239 244 14748743 10.1111/j.0959-9673.2003.00363.x
Amin HM McDonnell TJ Ma Y Lin Q Fujio Y Kunisada K Leventaki V Das P Rassidakis GZ Cutler C Medeiros LJ Lai R Selective inhibition of STAT3 induces apoptosis and G(1) cell cycle arrest in ALK-positive anaplastic large cell lymphoma Oncogene 2004 23 5426 5434 15184887 10.1038/sj.onc.1207703
Thomas LR Johnson RL Reed JC Thorburn A The C-terminal tails of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) and Fas receptors have opposing functions in Fas associated death domain (FADD) recruitment and can regulate agonist-specific mechanisms of receptor activation J Biol Chem 2004 279 52479 86 15452120 10.1074/jbc.M409578200
Thomas LR Henson A Reed JC Salsbury FR Thorburn A Direct binding of Fas-associated death domain (FADD) to the tumor necrosis factor-related apoptosis-inducing ligand receptor DR5 is regulated by the death effector domain of FADD J Biol Chem 2004 279 32780 32785 15173180 10.1074/jbc.M401680200
Watts GS Oshiro MM Junk DJ Wozniak RJ Watterson S Domann FE Futscher BW The acetyltransferase p300/CBP-associated factor is a p53 target gene in breast tumor cells Neoplasia 2004 6 187 194 15153330
Dhaka A Costa RM Hu H Irvin DK Patel A Kornblum HI Silva AJ O'Dell TJ Colicelli J The RAS effector RIN1 modulates the formation of aversive memories J Neurosci 2003 23 748 757 12574403
Colicelli J Human RAS Superfamily Proteins and Related GTPases Sci STKE 2004 2004 RE13 15367757 10.1126/stke.2502004re13
Koscielniak E Morgan M Treuner J Soft tissue sarcoma in children: prognosis and management Paediatr Drugs 2002 4 21 28 11817983
Phillips KA Toner GC Chemotherapy for soft tissue sarcomas. Indications and advances Acta Orthop Scand Suppl 1997 273 133 138 9057604
Valle AA Kraybill WG Management of soft tissue sarcomas of the extremity in adults J Surg Oncol 1996 63 271 279 8982374 10.1002/(SICI)1096-9098(199612)63:4<271::AID-JSO11>3.0.CO;2-6
Ruymann FB Grovas AC Progress in the diagnosis and treatment of rhabdomyosarcoma and related soft tissue sarcomas Cancer Invest 2000 18 223 241 10754991
Ikenaga M [Cytotoxic action of alkylating agents in human tumor cells and its relationship to apoptosis] Gan To Kagaku Ryoho 1994 21 596 601 8154884
Orlova TI [Mechanism of actinomycin biosynthesis] Antibiot Khimioter 1991 36 47 51 1723587
Friche E Skovsgaard T Nissen NI Anthracycline resistance Acta Oncol 1989 28 877 881 2692650
Hiipakka RA Liao S Steroid receptor recycling and interaction of receptor with RNA Am J Clin Oncol 1988 11 Suppl 2 S18 22 2468270
Bramwell VH Anderson D Charette ML Doxorubicin-based chemotherapy for the palliative treatment of adult patients with locally advanced or metastatic soft tissue sarcoma Cochrane Database Syst Rev 2003 CD003293 12917960
Tang HJ Qian D Sondak VK Stachura S Lin J A modified p53 enhances apoptosis in sarcoma cell lines mediated by doxorubicin Br J Cancer 2004 90 1285 1292 15026814 10.1038/sj.bjc.6601653
Herr I Debatin KM Cellular stress response and apoptosis in cancer therapy Blood 2001 98 2603 2614 11675328 10.1182/blood.V98.9.2603
Kaufmann SH Vaux DL Alterations in the apoptotic machinery and their potential role in anticancer drug resistance Oncogene 2003 22 7414 7430 14576849 10.1038/sj.onc.1206945
Kaufmann SH Hengartner MO Programmed cell death: alive and well in the new millennium Trends Cell Biol 2001 11 526 534 11719060 10.1016/S0962-8924(01)02173-0
Debatin KM Poncet D Kroemer G Chemotherapy: targeting the mitochondrial cell death pathway Oncogene 2002 21 8786 8803 12483532 10.1038/sj.onc.1206039
Philchenkov A Zavelevich M Kroczak TJ Los M Caspases and cancer: mechanisms of inactivation and new treatment modalities Exp Oncol 2004 26 82 97 15273659
Ekert PG Read SH Silke J Marsden VS Kaufmann H Hawkins CJ Gerl R Kumar S Vaux DL Apaf-1 and caspase-9 accelerate apoptosis, but do not determine whether factor-deprived or drug-treated cells die J Cell Biol 2004 165 835 842 15210730 10.1083/jcb.200312031
Soengas MS Capodieci P Polsky D Mora J Esteller M Opitz-Araya X McCombie R Herman JG Gerald WL Lazebnik YA Cordon-Cardo C Lowe SW Inactivation of the apoptosis effector Apaf-1 in malignant melanoma Nature 2001 409 207 211 11196646 10.1038/35051606
Yeh PY Chuang SE Yeh KH Song YC Chang LL Cheng AL Phosphorylation of p53 on Thr55 by ERK2 is necessary for doxorubicin-induced p53 activation and cell death Oncogene 2004 23 3580 3588 15116093 10.1038/sj.onc.1207426
Pusztai L Perspectives and challenges of clinical pharmacogenomics in cancer Pharmacogenomics 2004 5 451 454 15212580 10.1517/14622416.5.5.451
Pusztai L Krishnamurti S Perez Cardona J Sneige N Esteva FJ Volchenok M Breitenfelder P Kau SW Takayama S Krajewski S Reed JC Bast RCJ Hortobagyi GN Expression of BAG-1 and BcL-2 proteins before and after neoadjuvant chemotherapy of locally advanced breast cancer Cancer Invest 2004 22 248 256 15199607 10.1081/CNV-120030213
Okamura S Arakawa H Tanaka T Nakanishi H Ng CC Taya Y Monden M Nakamura Y p53DINP1, a p53-inducible gene, regulates p53-dependent apoptosis Mol Cell 2001 8 85 94 11511362 10.1016/S1097-2765(01)00284-2
Li J Wang WL Yang XK Yu XX Hou YD Zhang J Inducible overexpression of Bak sensitizes HCC-9204 cells to apoptosis induced by doxorubicin Acta Pharmacol Sin 2000 21 769 776 11501155
Takahashi N Li WW Banerjee D Scotto KW Bertino JR Sequence-dependent enhancement of cytotoxicity produced by ecteinascidin 743 (ET-743) with doxorubicin or paclitaxel in soft tissue sarcoma cells Clin Cancer Res 2001 7 3251 3257 11595721
Wu XX Ogawa O Kakehi Y TRAIL and chemotherapeutic drugs in cancer therapy Vitam Horm 2004 67 365 383 15110186
Kotamraju S Kalivendi SV Konorev E Chitambar CR Joseph J Kalyanaraman B Oxidant-induced iron signaling in Doxorubicin-mediated apoptosis Methods Enzymol 2004 378 362 382 15038980
Kumar D Lou H Singal PK Oxidative stress and apoptosis in heart dysfunction Herz 2002 27 662 668 12439637 10.1007/s00059-002-2430-3
Friesen C Fulda S Debatin KM Cytotoxic drugs and the CD95 pathway Leukemia 1999 13 1854 1858 10557062 10.1038/sj/leu/2401333
Utz PJ Anderson P Life and death decisions: regulation of apoptosis by proteolysis of signaling molecules Cell Death Differ 2000 7 589 602 10889504 10.1038/sj.cdd.4400696
Yuan J Molecular control of life and death Curr Opin Cell Biol 1995 7 211 214 7612273 10.1016/0955-0674(95)80030-1
Salvesen GS Dixit VM Caspase activation: the induced-proximity model Proc Natl Acad Sci U S A 1999 96 10964 10967 10500109 10.1073/pnas.96.20.10964
Slee EA Adrain C Martin SJ Serial killers: ordering caspase activation events in apoptosis Cell Death Differ 1999 6 1067 1074 10578175 10.1038/sj.cdd.4400601
Slee EA Harte MT Kluck RM Wolf BB Casiano CA Newmeyer DD Wang HG Reed JC Nicholson DW Alnemri ES Green DR Martin SJ Ordering the cytochrome c-initiated caspase cascade: hierarchical activation of caspases-2, -3, -6, -7, -8, and -10 in a caspase-9-dependent manner J Cell Biol 1999 144 281 292 9922454 10.1083/jcb.144.2.281
de Angelis PM Fjell B Kravik KL Haug T Tunheim SH Reichelt W Beigi M Clausen OP Galteland E Stokke T Molecular characterizations of derivatives of HCT116 colorectal cancer cells that are resistant to the chemotherapeutic agent 5-fluorouracil Int J Oncol 2004 24 1279 1288 15067352
Ota T Maeda M Suto S Tatsuka M LyGDI functions in cancer metastasis by anchoring Rho proteins to the cell membrane Mol Carcinog 2004 39 206 220 15057873 10.1002/mc.20006
Bian X Giordano TD Lin HJ Solomon G Castle VP Opipari AWJ Chemotherapy-induced apoptosis of S-type neuroblastoma cells requires caspase-9 and is augmented by CD95/Fas stimulation J Biol Chem 2004 279 4663 4669 14617634 10.1074/jbc.M306905200
Mitsiades CS Poulaki V Mitsiades N The role of apoptosis-inducing receptors of the tumor necrosis factor family in thyroid cancer J Endocrinol 2003 178 205 216 12904168 10.1677/joe.0.1780205
Crowe DL Yoon E A common pathway for chemotherapy-induced apoptosis in human squamous cell carcinoma lines distinct from that of receptor-mediated cell death Anticancer Res 2003 23 2321 2328 12894510
Parone PA James D Martinou JC Mitochondria: regulating the inevitable Biochimie 2002 84 105 111 12022941 10.1016/S0300-9084(02)01380-9
Belmokhtar CA Hillion J Dudognon C Fiorentino S Flexor M Lanotte M Segal-Bendirdjian E Apoptosome-independent pathway for apoptosis. Biochemical analysis of APAF-1 defects and biological outcomes J Biol Chem 2003 278 29571 29580 12773531 10.1074/jbc.M302924200
Haga A Funasaka T Niinaka Y Raz A Nagase H Autocrine motility factor signaling induces tumor apoptotic resistance by regulations Apaf-1 and Caspase-9 apoptosome expression Int J Cancer 2003 107 707 714 14566819 10.1002/ijc.11449
Fischer B Coelho D Dufour P Bergerat JP Denis JM Gueulette J Bischoff P Caspase 8-mediated cleavage of the pro-apoptotic BCL-2 family member BID in p53-dependent apoptosis Biochem Biophys Res Commun 2003 306 516 522 12804595 10.1016/S0006-291X(03)01004-0
Fischer U Janicke RU Schulze-Osthoff K Many cuts to ruin: a comprehensive update of caspase substrates Cell Death Differ 2003 10 76 100 12655297 10.1038/sj.cdd.4401160
Wang T Niu G Kortylewski M Burdelya L Shain K Zhang S Bhattacharya R Gabrilovich D Heller R Coppola D Dalton W Jove R Pardoll D Yu H Regulation of the innate and adaptive immune responses by Stat-3 signaling in tumor cells Nat Med 2004 10 48 54 14702634 10.1038/nm976
Lee SO Lou W Qureshi KM Mehraein-Ghomi F Trump DL Gao AC RNA interference targeting Stat3 inhibits growth and induces apoptosis of human prostate cancer cells Prostate 2004 60 303 309 15264241 10.1002/pros.20072
Rebbaa A Chou PM Mirkin BL Factors secreted by human neuroblastoma mediated doxorubicin resistance by activating STAT3 and inhibiting apoptosis Mol Med 2001 7 393 400 11474132
Kovalovich K Li W DeAngelis R Greenbaum LE Ciliberto G Taub R Interleukin-6 protects against Fas-mediated death by establishing a critical level of anti-apoptotic hepatic proteins FLIP, Bcl-2, and Bcl-xL J Biol Chem 2001 276 26605 26613 11349125 10.1074/jbc.M100740200
Pu YS Hour TC Chuang SE Cheng AL Lai MK Kuo ML Interleukin-6 is responsible for drug resistance and anti-apoptotic effects in prostatic cancer cells Prostate 2004 60 120 129 15162378 10.1002/pros.20057
Takeda T Kurachi H Yamamoto T Nishio Y Nakatsuji Y Morishige K Miyake A Murata Y Crosstalk between the interleukin-6 (IL-6)-JAK-STAT and the glucocorticoid-nuclear receptor pathway: synergistic activation of IL-6 response element by IL-6 and glucocorticoid J Endocrinol 1998 159 323 330 9795374 10.1677/joe.0.1590323
Sperandio S de Belle I Bredesen DE An alternative, nonapoptotic form of programmed cell death Proc Natl Acad Sci U S A 2000 97 14376 14381 11121041 10.1073/pnas.97.26.14376
Golstein P Wyllie AH T cell death and transforming growth factor beta1 J Exp Med 2001 194 F19 22 11514612 10.1084/jem.194.4.F19
Borner C Monney L Apoptosis without caspases: an inefficient molecular guillotine? Cell Death Differ 1999 6 497 507 10381652 10.1038/sj.cdd.4400525
Johnson VL Ko SC Holmstrom TH Eriksson JE Chow SC Effector caspases are dispensable for the early nuclear morphological changes during chemical-induced apoptosis J Cell Sci 2000 113 ( Pt 17) 2941 2953 10934034
Leist M Jaattela M Four deaths and a funeral: from caspases to alternative mechanisms Nat Rev Mol Cell Biol 2001 2 589 598 11483992 10.1038/35085008
Rebbaa A Zheng X Chou PM Mirkin BL Caspase inhibition switches doxorubicin-induced apoptosis to senescence Oncogene 2003 22 2805 2811 12743603 10.1038/sj.onc.1206366
Zheng X Chou PM Mirkin BL Rebbaa A Emran MA Emran M Senescence-initiated reversal of drug resistance: specific role of cathepsin L
Caspase inhibition switches doxorubicin-induced apoptosis to senescence
Doxorubicin resistant neuroblastoma cells secrete factors that activate AKT and attenuate cytotoxicity in drug-sensitive cells
Doxorubicin-induced apoptosis in caspase-8-deficient neuroblastoma cells is mediated through direct action on mitochondria
Factors secreted by human neuroblastoma mediated doxorubicin resistance by activating STAT3 and inhibiting apoptosis Cancer Res 2004 64 1773 1780 14996739
|
16001973
|
PMC1183194
|
CC BY
|
2021-01-04 16:03:06
|
no
|
BMC Cancer. 2005 Jul 7; 5:74
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-74
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-831602951610.1186/1471-2407-5-83Research ArticleApplication of serum SELDI proteomic patterns in diagnosis of lung cancer Yang Shuan-ying [email protected] Xue-yuan [email protected] Wang-gang [email protected] Li-juan [email protected] Wei [email protected] Bin [email protected] Guoan [email protected] Da-cheng [email protected] Department of Respiratory Medicine, Second Hospital of Xi'an Jiaotong University, Xi'an, 710004, China2 Key Laboratory for Cell Proliferation and Regulation Biology Ministry of Education, Bejing Normal University, Bejing, 100875, China3 General Thoracic Surgery, Second Hospital of Xi'an Jiaotong University, Xi'an, 710004, China4 Department of Surgery, University of Michigan Medical School, MI, 48109, USA2005 20 7 2005 5 83 83 17 3 2005 20 7 2005 Copyright © 2005 Yang et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Currently, no satisfactory biomarkers are available to screen for lung cancer. Surface-Enhanced Laser Desorption/ionization Time-of- Flight Mass Spectrometry ProteinChip system (SELDI-TOF-MS) is one of the currently used techniques to identify biomarkers for cancers. The aim of this study is to explore the application of serum SELDI proteomic patterns to distinguish lung cancer patients from healthy individuals.
Methods
A total of 208 serum samples, including 158 lung cancer patients and 50 healthy individuals, were randomly divided into a training set (including 11 sera from patients with stages I/II lung cancer, 63 from patients with stages III/IV lung cancer and 20 from healthy controls) and a blinded test set (including 43 sera from patients with stages I/II lung cancer, 41 from patients with stages III/IV lung cancer and 30 from healthy controls). All samples were analyzed by SELDI technology. The spectra were generated on weak cation exchange (WCX2) chips, and protein peaks clustering and classification analyses were made using Ciphergen Biomarker Wizard and Biomarker Pattern software, respectively. We additionally determined Cyfra21-1 and NSE in the 208 serum samples included in this study using an electrochemiluminescent immunoassay.
Results
Five protein peaks at 11493, 6429, 8245, 5335 and 2538 Da were automatically chosen as a biomarker pattern in the training set. When the SELDI marker pattern was tested with the blinded test set, it yielded a sensitivity of 86.9%, a specificity of 80.0% and a positive predictive value of 92.4%. The sensitivities provided by Cyfra21-1 and NSE used individually or in combination were significantly lower than that of the SELDI marker pattern (P < 0.005 or 0.05, respectively). Based on the results of the test set, we found that the SELDI marker pattern showed a sensitivity of 91.4% in the detection of non-small cell lung cancers (NSCLC), which was significantly higher than that in the detection of small cell lung cancers (P < 0.05); The pattern also had a sensitivity of 79.1% in the detection of lung cancers in stages I/II.
Conclusion
These results suggest that serum SELDI protein profiling can distinguish lung cancer patients, especially NSCLC patients, from normal subjects with relatively high sensitivity and specificity, and the SELDI-TOF-MS is a potential tool for the screening of lung cancer.
==== Body
Background
Lung cancer is, at present, the most common malignancy in the world and its overall 5-year survival rate is only 14% [1]. The poor prognosis is due largely to lack of sufficient screening and early diagnostic tools to physicians. Currently in clinic the screening and early diagnosis of lung cancer relies mainly on chest X-ray, low-dose computed tomography, bronchoscopy, sputum cytology, and tumor markers including carcinoembryonic antigen (CEA), cytokeratin-19 fragments (Cyfra21-1), carbohydrate antigen 19-9 (CA19-9), squamous cell carcinoma antigen (SCCAg) and neuron-specific enolase (NSE), etc [2]. All these methods, however, lack adequate sensitivity and/or specificity [3-6]. Thus, it is urgent to search for better methods which provide more valuable information for screening and early diagnosis of lung cancer. Because of the marked heterogeneity of lung cancer [5], a panel of biomarkers for screening and diagnosis would be most appropriate. Surface-Enhanced Laser Desorption /ionization Time-of-Flight Mass Spectrometry (SELDI-TOF-MS), an innovative proteomic technology introduced by Hutchens and Yip [7], has overcome many of the limitations of two-dimensional electrophoresis and Matrix-Assisted Laser Desorption/ionization Time-of-Flight Mass Spectrometry (MALDI-TOF-MS) [8,9]. This is a high through-put technique for analysis of complex biological specimens such as serum. It can detect multiple protein changes simultaneously with high sensitivity and specificity [10,11]. Recently, SELDI has been successfully used to distinguish pancreatic, ovarian and prostate cancer patients from controls [9,12,13], and detect markers of bladder cancer in urine [14].
The aim of the current study was to investigate the application of serum SELDI protein profiling to distinguish lung cancer patients from a healthy population.
Methods
Patients
A total of 208 serum samples including 158 pathologically confirmed lung cancer patients and 50 healthy subjects were collected from the Department of Respiratory and Thoracic Surgery of the Second Hospital of Xi'an Jiaotong University. Informed consent was obtained from every subject prior to the study. All patients with lung cancer were found to have no evidence of other disease. The distribution of clinical stages (UICC, 1997) was as follows: 13 cases were at stage I, 41 stage II, 58 stage III, 46 stage IV. Among these patients, 68 patients suffered from squamous cell carcinomas, 53 from adenocarcinomas, 35 from small cell cancers and 2 from bronchioloalveolar carcinomas. The average age of the patients (101 males, 57 females ranging from 28 to 79 years) was 56.8. The healthy controls (31 males, 19 females ranging from 30 to72 years) came from general physical examinations, and had an average age of 54.5. The two groups were matched for age, sex and smoking history. Two milliliters of whole blood were collected during fasting and stored within one hour at 4°C. The blood was later centrifuged for 20 min at 4000 rpm, distributed into 100 μl aliquots, and stored at -80°C until used.
SELDI protein profiling
Five μL of 10 mM HCl was applied to a weak cation exchange (WCX2) chip and placed at room temperature for 10 min. Chips were rinsed with deionized water in a conical tube and then put into a bioprocessor and washed with binding buffer (100 mM NaAc, pH4) with gentle shaking twice for 5 min. Five μL of each serum and 10 μL of 9 mol/L urea were combined and vortexed on ice. 5 μL of this mixture was added to 60 μL of binding buffer. 50 μL of the serum mixture was applied to each spot and incubated on a shaker for 60 min. Chips were washed again with binding buffer with slight shaking 3 times. 200 μL of 1 mM HEPES pH7.0 was added to each well. Wells were quickly rinsed and then removed and let dry. Once dry, 0.5 μL of sinapinic acid (SPA) was applied to each spot twice. The arrays were allowed to air-dry and then stored in the dark at RT until SELDI analysis.
Data analysis
Before analysis, the data were randomly divided into two sets as follows: the training set consisted of 11 patients with stages I/II lung cancer, 63 patients with stages III/IV lung cancer and 20 healthy controls. The blinded test (in which the disease status was unrevealed) set consisted of 43 patients with stages I/II lung cancer, 41 patients with stages III/IV lung cancer and 30 healthy controls. The chips were placed in the Protein Biological system II-C mass spectrometer reader (Ciphergen Biosystems, Inc.) and TOF spectra were generated by averaging 128 laser shots with an intensity of 215 and a detector sensitivity of 9. The optimization range was from 3,000 to 50,000 Da, and a maximum of 200,000 Da. External calibration of the instrument was performed using the All-in-one peptide molecular mass standard (Ciphergen Biosystems, Inc.). We achieved a mass accuracy of 0.1% with this system.
Peak detection
Peak detection using Ciphergen Biomarker Wizard software 3.0.2 identified an average of 72 peaks/spectrum. Of the 72 peaks, 64 common peaks or clusters were generated from the training set. Eighteen of these proteins were found to have statistically differential expression levels between lung cancer and normal control sera (P < 10-4). Peak detections involved baseline subtraction, mass accuracy calibration, and automatic peak detection. The settings used for our work were as follows: for peak detection the signal-to-noise ratio was 3, minimum peak threshold was 10%; for cluster completion, the cluster mass was 0.5% and the signal-to-noise ratio for the second pass was 1.
Decision tree classification
Construction of the decision tree classification algorithm was performed by Ciphergen Biomarker Pattern software version 5.0. Classification tree, selected Gini, split the data into two nodes using one rule at a time in the form of peak intensity. The splitting decisions in this case were based on the normalized intensity levels of peaks from SELDI protein expression profile. The process of splitting was continued until terminal nodes were created. After V-fold cross validation 50, the accuracy of each classification tree was then challenged with the blinded test set.
Detection of serum Cyfra21-1 and NSE
The two markers, Crfra21-1 and NSE, were measured in the 208 sera included in this study using an electrochemiluminescent immunoassay (ECLIA, Elecsys 2010 system, Roche Diagnostics, Switzerland). The cutoff values for Crfra21-1 and NSE, recommended by the manufacturers, were 3.3 ng ml-1and 16.3 ng ml-1, respectively.
Statistical analysis
Comparison of relative peak intensity levels between groups was made using the Student's t test and in all cases P < 10-4 was considered statistically significant. Comparison of rates between groups was conducted using the χ2 test and P < 0.05 was regarded as a significant difference.
Results
Reproducibility
The reproducibility of each SELDI proteinchip assay was determined by SELDI profiling of 10 aliquots of pooled normal serum. The average coefficient of variance (CV) based on 10 pooled normal human sera for intensities of 22 randomly chosen peaks was less than 20%. Little variation with day-to-day sampling and instrumentation or chip variations was found.
Serum SELDI profiles of lung cancers versus healthy controls
Using Ciphergen Biomarker pattern software to analyze the data derived from Ciphergen Biomarker wizard software, approximately 64 peaks per spectrum identified in the training set were determined with masses ranging from 3–30 kDa. We found that no single peak could adequately discriminate lung cancer sera from normal sera. Using all 64 peaks, a decision tree classification algorithm was built and five protein peaks at 11493, 8245, 5335, 6429 and 2538 Da were automatically selected as splitters. The 11,493 Da peak was used as the root node in the classification tree to divide the 94 samples into two groups (Fig. 1): the left node (node 2) included cases with peak intensity < 2.018. The right node (node 6) contained the remaining with peak intensity = 2.018. The cases in each branch node were then reclassified at the next layer following the same process with 6429, 5335, 2538 and 8245 Da as splitters. This splitting process stops if terminal nodes for further splitting have no gain. Finally, all 94 cases in the training set were classified in the 7 terminal nodes, and a classification tree was obtained (Fig. 1). The tree correctly classified 95.9% of the lung cancer sera in the training set (Table 1). The validity of this classification tree algorithm was then challenged with the test set and a total of 80.0% of controls and 86.9% of lung cancer samples were correctly identified (Table 1). Based on the results of the test set we calculated the sensitivity of the SELDI marker pattern in the detection of lung cancers with different stages and pathological types (Table 2). The peaks at 11493 and 5335 Da are shown in Fig. 2. Aside from the 11493 Da peak, any of the other 5 peaks could have been used as the first node in the classification trees in the same way as 11493 Da, but their performance scores were inferior to the 11493Da peak.
Discriminatory power of serum Cyfra21-1 and NSE
Table 3 provides the results of sensitivities and specificities of Cyfra21-1 and NSE used individually and combined. We compared the diagnostic capacities of the SELDI marker pattern with Cyfra21-1 and NSE individually and combined (Table 3).
Discussion
Currently, there are no satisfactory screening and early diagnostic strategies for lung cancer. SELDI is a high through-put technique used to generate protein expression profiles which, in combination with bioinformatics tools to extract information for biomarker discovery, has been essential in identifying novel protein biomarkers. Indeed, application of this technology has shown great potential for the early detection of ovarian and prostate cancers [10,12].
Proteomic studies of lung cancer are still scarce [15]. Recently, Xiao, et al [16] reported that a proteomic panel consisting of three protein peaks yielded a sensitivity of 93.3% and specificity of 96.7% in distinguishing lung cancer patients from healthy controls. This study was, however, based on only 45 tumor samples. In the present study, we examined 158 serum samples from lung cancer patients and 50 from healthy individuals using the SELDI technique with the WCX2 proteinchip. The classification tree was constructed to distinguish lung cancer cases from healthy individuals using 5 protein peaks at 11493, 6429, 8245, 5335 and 2538 Da as a marker pattern. When the model was tested with the blinded test set, it yielded a sensitivity of 86.9%, specificity of 80.0%, and positive predictive value of 92.4% (73/79). For comparison, Cyfra21-1 and NSE were measured using ECLIA in our study. Although there is no statistical differences between the specificities of Cyfra21-1, NSE and the SELDI marker pattern, the sensitivity achieved by Cyfra21-1, NSE individually or in combination were significantly lower than that of the SELDI pattern. These results indicate that the SELDI pattern is distinctly superior to Cyfra21-1 and NSE individually or combined in distinguishing lung cancer patients from healthy individuals.
Based on the results of the blinded test set, we found that the sensitivity of the SELDI marker pattern for NSCLCs was significantly higher than for SCLCs, indicating that the pattern may be more effective in discriminating NSCLC patients from healthy controls than SCLC patients. Similarly, the pattern also had a sensitivity of 79.1% in the detection of lung cancers with stages I/II, suggesting that the pattern might be better for early detection of lung cancer than any other single or panel of biomarkers currently used in clinic [17,18].
To develop a broad biomarker panel for screening a diverse, high-risk population, both NSCLC and SCLC patients were chosen for our study. Due to the relatively fewer healthy control samples and the subgroup of patients with SCLC, our results require more samples to broaden and improve its diagnostic value. Furthermore, the five proteins included in the SELDI marker pattern will be identified by MALDI-MS-MS.
Conclusion
We have found that serum SELDI protein profiling can distinguish lung cancer patients, especially NSCLC patients, from healthy controls with relatively high sensitivity and specificity. The SELDI-TOF-MS is a potential tool for the screening of lung cancer.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SY was responsible for the conception and design of this study, providing samples and clinical data, drafting and revising the article. XX contributed to the design of this study, performed statistical analysis and interpretation of the data. LZ provided technical support and some experiments. WZ and BZ provided in part study materials and medical aspects of the work. GC has been involved in discussion and revising this article. DH contributed to the conception of this 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 work was supported by grants from the Shaanxi Science and Technology Development Program [No.2004K13-G3 (3)]
Figures and Tables
Figure 1 Classification of lung cancer vs. normal samples by the decision tree algorithm. The left branch node after the first layer is the cases of peak intensity under 2.018, the right one is over or equal to 2.018. The cutoff points for 8245, 6429 and 2538 Da were 1.574, 49.64 and 13.01, respectively. The cutoff points of mass 5335 Da were 0.288 (left) and 2.163 (right). N represents the number of samples. M represents the molecular weight.
Figure 2 Differential expressions of the SELDI peaks at 11493 (group A) and 5335Da (group B) in the comparisons of lung cancer and healthy control sera. a-c: squamous cell carcinomas; d-f: adenocarcinomas; g-i: small cell carcinomas; j-l, healthy controls. X-axis was molecular weight of peak; Y-axis was intensity of peak.
Table 1 Classification tree analysis of the lung cancer training and blinded test sets.
Sets Groups Percentage correct Percentage misclassified
Training set Normal (n = 20) 90.0 % (18/20) 10.0% (2/20)
Cancer (74) 95.9% (71/74) 4.1% (3/74)
Blinded test set Normal (30) 80.0% (24/30) 20.0% (6/30)
Cancer (84) 86.9% (73/84) 13.1% (11/84)
Table 2 The sensitivity of the SELDI marker pattern in lung cancer patients with different stages and pathological types.
Variables n Correct cases Error cases Sensitivity (%) P value
Clinical stages
I/II 43 34 9 79.1
III/IV 41 39 2 95.1 <0.05
Pathological types
NSCLC 70 64 6 91.4
SCLC 14 9 5 64.3 <0.05
NSCLC: non-small cell lung cancer; SCLC: small cell lung cancer.
Table 3 The sensitivities and specificities of the SELDI marker pattern, Cyfra21-1 and NSE.
Items Pattern Cyfra21-1 NSE Combination
Blinded test set
Sensitivity (%) 86.9 (73/84) 44.0(37/84)a 36.9(31/84)a 63.1(53/84)a
Specificity (%) 80.0 (24/30) 70.0(21/30) 73.3(22/30) 53.3(16/30) b
Whole set
Sensitivity (%) 46.2 (73/158) 35.4 (56/158) 62.0 (98/158)
Specificity (%) 72.0 (36/50) 74.0 (37/50) 56.0 (28/50)
Pattern: SELDI marker pattern; Combination: combined use of Cyfra21-1 and NSE. a P < 0.005 compared with the SELDI marker pattern; b P < 0.05 compared with the SELDI marker pattern.
==== Refs
Spira A Ettinger DS Multidisciplinary management of lung cancer N Engl J Med 2004 350 379 392 14736930 10.1056/NEJMra035536
Stieber P Aronsson AC Bialk P Tumor markers in lung cancer: EGTM recommendations Anticancer Res 1999 19 2817 2819 10470248
Swensen SJ Jett JR Hartman TE Midthun DE Sloan JA Sykes AM Aughenbaugh GL Clemens MA Lung cancer screening with CT: Mayo clinic experience Radiology 2003 226 756 761 12601181
Kulpa J Wojcik E Reinfuss M Kolodziejski L Carcinoembryonic antigen, squamous cell carcinoma antigen, CYFRA21-1, and neuro-specific enolase in squamous cell lung cancer patients Clin Chem 2002 48 1931 1937 12406978
Zhong L Peng X Hidalgo GE Doherty DE Stromberg AJ Hirschowitz EA Identification of circulating antibodies to tumor-associated proteins for combined use as markers of non-small cell lung cancer Proteomics 2004 4 1216 1225 15049001 10.1002/pmic.200200679
Lam S Kennedy T Unger M Miller YE Gelmont D Rusch V Gipe B Howard D LeRiche JC Coldman A Gazdar AF Localization of bronchial intraepithelial neoplastic lesions by fluorescence bronchoscopy Chest 1998 113 696 702 9515845
Hutchen TW Yip TT New desorption strategies for the mass spectrometric analysis of macromolecules Rapid Commun Mass Spectrom 1993 7 576 580 10.1002/rcm.1290070703
Wadsworth JT Somers KD Cazares LH Malik G Adam BL Stack BC JrWright GL JrSemmes OJ Serum protein profiles to identify head and neck cancer Clin Cancer Res 2004 10 1625 1632 15014013
Rosty C Christa L Kuzdzal S Baldwin WM Zahurak ML Carnot F Chan DW Canto M Lillemoe KD Cameron JL Yeo CJ Hruban RH Goggins M Identification of hepatocarcinoma-intestine-pancreas/pancreatitis – associated protein I as a biomarker for pancreatic ductal adenocarcinoma by protein biochip technology Cancer Res 2002 62 1868 1875 11912167
Cazares LH Adam BL Ward MD Nasim S Schellhammer PF Semmes OJ Wright GL Jr Normal benign, preneoplastic, and malignant prostate cells have distinct protein expression profiles resolved by Surface Enhanced Laser Desorption/Ionization Mass Spectrometry Clin Cancer Res 2002 8 2541 2552 12171882
Koopmann J Zhang Z White N Rosenzweig J Fedarko N Jagannath S Canto MI Yeo CJ Chan DW Goggins M Serum diagnosis of pancreatic adenocarcinoma using Surface Enhanced Laser Desorption/Ionization Mass Spectrometry Clin Cancer Res 2004 10 860 868 14871961
Petricoin EF Ardekani AM Hitt BA Levine PJ Fusaro VA Steinberg SM Mills GB Simone C Fishman DA Kohn EC Liotta LA Use of proteomic patterns in serum to identify ovarian cancer Lancet 2002 359 572 577 11867112 10.1016/S0140-6736(02)07746-2
Adam BL Qu Y Davis JW Ward MD Clements MA Cazares LH Semmes OJ Schellhammer PF Yasui Y Feng Z Wright GL Jr Serum protein fingerprinting coupled with a pattern-matching algorithm distinguishes prostate cancer from benign prostate hyperplasia and healthy men Cancer Res 2002 62 3609 3614 12097261
Vlahou A Schellhammer PF Mendrinos S Patel K Kondylis FI Gong L Nasim S Wright GL Jr Development of a novel proteomic approach for the detection of transitional cell carcinoma of the bladder in urine Am J Pathol 2001 158 1491 1502 11290567
Alfonso P Catala M Rico-Morales ML Durante-Rodriguez G Moro-Rodriguez E Fernandez-Garcia H Escribano JM Alvarez-Fernandez E Garcia-Poblete E Proteomic analysis of lung biopsies: Differential protein expression profile between peritumoral and tumoral tissue Proteomics 2004 4 442 447 14760715 10.1002/pmic.200300647
Xiao X Liu D Tang Y Guo F Xia L Liu J He D Development of proteomic patterns for detecting lung cancer Dis Markers 2003 19 33 39 14757945
Luo S Wang M Li Y Clinical significance in diagnosing lung cancer with the combined determination of serum tumor markers J tumor Marker Oncology 2004 19 90 10.1159/000074964
Gao Q Li Y Omoli H Clinical value of a cancer marker cytokeratin 19 for diagnosis of lung cancer J Chin Med Univ 1999 28 293 294
|
16029516
|
PMC1183195
|
CC BY
|
2021-01-04 16:03:07
|
no
|
BMC Cancer. 2005 Jul 20; 5:83
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-83
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-871604278610.1186/1471-2407-5-87Research ArticlePhase I dose-escalating study of docetaxel in combination with 5-day continuous infusion of 5-fluorouracil in patients with advanced gastric cancer Park Se Hoon [email protected] Soo-Mee [email protected] Eun Kyung [email protected] Dong Bok [email protected] Jae Hoon [email protected] Woon KI [email protected] Min [email protected] Division of Hematology and Oncology, Department of Internal Medicine, Gachon Medical School Gil Medical Center, Incheon 405-760, Korea2 Department of General Surgery, Gachon Medical School Gil Medical Center, Incheon 405-760, Korea2005 22 7 2005 5 87 87 31 1 2005 22 7 2005 Copyright © 2005 Park 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
Published data suggests that docetaxel combined with 5-fluorouracil (5-FU) may have synergistic activity in treating advanced gastric cancer. We performed a phase I study of docetaxel and 5-FU to determine the maximum tolerated dose (MTD), the recommended dose for phase II studies, and the safety of this combination.
Methods
Eligible patients had recurrent and/or metastatic advanced gastric cancer with normal cardiac, renal and hepatic function. Traditional phase I methodology was employed in assessing dose-limiting toxicity (DLT) and MTD. On day 1 every 3 weeks, docetaxel 75 mg/m2 (fixed dose) was infused over 1-h, followed immediately by 5-FU as a 5-day continuous infusion.
Results
Dose escalation schema was as follows: dose level (DL) 1 (5-FU 250 mg/m2/day), 2 (500), 3 (750), and 4 (1000). Three patients were enrolled on DL1, without DLT. On DL2, 1 DLT (grade 3 stomatitis) was developed in first 3 patients, and this cohort was expanded to 6 patients. Three patients had been enrolled on DL3. Because two out of 3 patients had DLTs, the MTD was reached at DL3.
Conclusion
The recommended phase II dose of this combination is 75 mg/m2 docetaxel on day 1 immediately followed by a 5-day continuous infusion of 5-FU 500 mg/m2/day.
==== Body
Background
Gastric cancer remains the most common cause of cancer-related death in Korea [1]. For patients with unresectable, locally advanced, or metastatic disease, chemotherapy can provide significant palliation of symptoms [2,3]. When used as single agents, 5-fluorouracil (5-FU), doxorubicin, cisplatin, and mitomycin C are considered active in gastric cancer, producing response rates in up to 20% of patients [4]. Of these, 5-FU is an central cytostatic antimetabolite with a broad range of antitumor activity in breast, gastrointestinal, head and neck, and ovarian cancers. When given as a prolonged continuous infusion, stomatitis and diarrhea are the principal toxicities, whereas myelosuppression is more commonly observed with intravenous bolus injections [5,6]. Because 5-FU has demonstrated synergistic interaction with many antineoplastic agents, it is currently most often administered in the setting of combination chemotherapy regimens. However, most trials with different combinations of these drugs provided similar survivals ranging from 6 to 10 months [7,8]. No regimen has been universally recognized as standard.
Among the new agents, docetaxel is a novel semisynthetic taxane with significant antitumor activity and manageable toxicity consisting primarily of myelosuppression [9]. Results of docetaxel-containing regimen in the treatment of gastric cancer are encouraging. Phase I and II clinical trials have confirmed that docetaxel is effective in patients with advanced gastric cancer when used as monotherapy, yielding response rates of 20% [10,11]. Docetaxel and 5-FU are highly synergistic [12], and clinical evidence suggests lack of cross-resistance.
When this study was initiated, most previous studies used docetaxel 60–85 mg/m2 on day 1, followed by a 5-day continuous infusion of 5-FU 500–1000 mg/m2/day, every 3 weeks [12-15]. However, we had experienced severe toxicities such as febrile neutropenia and grade 3 or 4 stomatitis in treating patients with advanced gastric cancer as a practice with docetaxel 75 mg/m2 and 5-FU 1000 mg/m2/day for 5 days. Therefore, we performed a phase I study of docetaxel and 5-FU to determine the maximum tolerated dose (MTD) and the safety of this combination.
Methods
Patients
This was a phase I dose-escalating study conducted between Mar 2002 and Aug 2002. The study was conducted according to the principles stated in the latest version of the Declaration of Helsinki. This study protocol was reviewed and approved by the Gil Medical Center institutional review board, and signed informed consent was obtained from all patients prior to their enrollment.
Patients enrolled into this study had to be at least 18 years of age and have a histologically confirmed diagnosis of metastatic or recurrent gastric adenocarcinoma that was previously treated with, though not necessarily resistant to, cytotoxic chemotherapy. The patients were required to have a Eastern Cooperative Oncology Group (ECOG) performance status ≤2, normal blood counts, normal renal and hepatic functions, no history of anaphylaxis and no peripheral neuropathy of any origin. Patients had to have received at least one chemotherapy regimen, either as adjuvant treatment or for metastatic disease. Patients could have received prior 5-FU, provided it was administered in intravenous bolus or in oral form, but not with paclitaxel or docetaxel. Patients had to have fully recovered from the toxic effects of previous chemotherapy except for alopecia.
Treatment plan
The treatment consisted of docetaxel 75 mg/m2 on day 1 in a 1-h infusion followed by 5-FU in continuous infusion from day 1 to day 5, according to the escalating dose levels. The starting dose of 5-FU was 250 mg/m2/day for 5 days. In the absence of DLT, dose escalation in additional cohorts continued at a dose increase of 250 mg/m2/dose. Traditional phase I methodology was employed in assessing dose-limiting toxicity (DLT) and MTD. Patients were to be treated at the same dose level in groups of three. If no DLT, defined as febrile neutropenia and/or grade 3/4 toxicity of any other kind apart from alopecia, occurred, the next 3 patients were treated at the next higher dose level. If one DLT occurred, 3 additional patients had to be treated at the same dose level. If 2 or more DLTs occurred at a given dose level, the MTD would be considered to be reached and the dose escalation had to be stopped. The dose just below would be considered to be the recommended dose for further evaluation in phase II trials.
Only the first cycle of treatment was evaluated to determine DLT. However, to adequately determine to safety of this combination, the recruitment of further patients at the recommended dose level was planned in a subsequent phase II study [16]. All patients received a standard supportive regimen consisting of dexamethasone and 5-HT3 inhibitors. The use of hematological growth factors was not allowed during the first cycle of treatment, but was permitted thereafter among patients who had one episode of febrile neutropenia or infection according to the guideline provided by American Society of Clinical Oncology [17].
Patient evaluation
Toxicity was graded according to National Cancer Institute Common Toxicity Criteria for adverse events (NCI-CTC) [18]. Dose modifications were planned for toxicity. The primary objective of the study was to define the MTD of the regimen under investigation. Evaluation at baseline and during the study included a medical history, data on toxicity and physical examination every courses.
Results
Patient characteristics
Sixteen patients received a total of 45 cycles of treatment (median 3, range 1–6). Patient characteristics are listed in Table 1. All patients previously received one cytotoxic chemotherapy regimen, with 88% having received prior fluoropyrimidines. Previous chemotherapy was commonly administered as adjuvant treatment (4 patients, 25%) or for metastatic disease (12 patients; 75%). Twelve patients (75%) had no measurable lesions to evaluate treatment efficacy.
Toxicity
All patients were evaluable for toxicity. Of three patients enrolled on dose level 1 (5-FU 250 mg/m2/day for 5 days), none experienced DLT. On dose level 2 (5-FU 500 mg/m2/day for 5 days), a 63-year-old male with peritoneal dissemination developed grade 3 stomatitis and fatigue. Three additional patients were enrolled at this dose level did not experience DLT. Dose escalation thus proceeded to dose level 3 (5-FU 750 mg/m2/day for 5 days). One of the 3 initial patients at this level, a 57-year-old male with peritoneal dissemination developed grade 3 stomatitis and another patient, a 60-year-old male with liver metastasis had febrile neutropenia. Because two out of 3 patients had DLTs, the MTD was reached at dose level 3. Dose level 2 was therefore the recommended regimen with docetaxel 75 mg/m2 on day 1 and 5-FU 750 mg/m2/day in a 5-day continuous infusion.
Toxicity analysis is based on 16 patients and 45 cycles of treatment. Table 2 summarizes DLTs per patients and per cycles. No treatment-related death was observed.
Discussion
This phase I trial determined the recommended phase II dose of 5-FU to be 500 mg/m2/day as a 5-day continuous infusion when combined with 75 mg/m2 docetaxel every 3 weeks. DLTs included stomatitis, neutropenia, diarrhea, and asthenia. Toxicity in our study did not differ significantly from that seen in other phase I studies that evaluated docetaxel combined with 5-FU continuous infusion. A phase I trial showed that docetaxel at 85 mg/m2 with 5-FU given in continuous infusion over 5 days at 750 mg/m2/day was tolerable without any major mucosal toxicity or any substantial increase in docetaxel-induced myelotoxicity [12]. Ando et al. [13] determined the MTDs of docetaxel and 5-FU on this schedule to be 50 mg/m2 and 500 mg/m2/day, respectively, and neutropenia to be the primary DLT, with diarrhea also dose limiting. Using a similar schedule, Van den Neste et al. [14] reported the MTDs of docetaxel and 5-FU to be 85 mg/m2 and 750 mg/m2/day, with diarrhea, stomatitis and neutropenia as DLTs.
Because of the relatively high incidence of hematologic toxicity induced by docetaxel, it was felt that only a moderately myelotoxic agent could be added as a combination. 5-FU given in continuous infusion could be an interesting option since it is known to induce very little myelotoxicity, if any.
With the addition of 5-FU, the main changes in the non-hematologic toxicity were observed in the occurrence of stomatitis and diarrhea, and in the appearance of hand-foot syndrome. All three toxicities are known to be associated with 5-FU. Hawkins et al. [15] compared docetaxel with the addition of irinotecan or 5-FU in a randomized phase II study. Both docetaxel/irinotecan and docetaxel/5-FU were well-tolerated with acceptable toxicity profiles. Three-drug combination of docetaxel, 5-FU and cisplatin for patients with gastric cancer was also investigated by some authors. Roth et al. [19] administered combination chemotherapy with docetaxel 85 mg/m2, cisplatin 75 mg/m2, and 5-FU 300 mg/m2/day on days 1–14, resulting in a 51% objective response. The major toxicity was neutropenia which reached grade 3 or 4 in 79% of 52 patients. Ajani et al. [20] evaluated a combination of docetaxel, cisplatin, and 5-FU (DCF) for chemotherapy-naïve patients with advanced gastric cancer. They achieved a response rate of 39% and median survival of 10.2 months, and toxicity was manageable. It is apparent that the more complex a chemotherapy regimen, the more toxic and difficult for patients to tolerate. Although cisplatin is often used in combination with other agents, it is well known that cisplatin is associated with significant toxicity and usually requires high level of clinical monitoring and supportive care including intensive intravenous hydration. Thuss-Patience et al. [21] recently reported that docetaxel/5-FU had similar efficacy to epirubicin/cisplatin/5-FU (ECF) and even to DCF.
Since nearly 80% of the patients did not have measurable disease, an efficacy assessment could not be performed in this patient population. Our randomized phase II trial with docetaxel/5-FU versus paclitaxel/5-FU is currently ongoing [16].
Conclusion
In conclusion, the recommended dose of this combination is 75 mg/m2 docetaxel on day 1 immediately followed by a 5-day continuous infusion of 5-FU 500 mg/m2/day. Further evaluation of efficacy and safety is currently underway in a randomized phase II trial of 5-FU combined with docetaxel versus paclitaxel.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SHP collected the data, performed the statistical analysis and drafted the manuscript. SB, EKC, JHL, WKL and MC followed the patients. DBS designed the study and helped with the manuscript. 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
Table 1 Patient characteristics
Number of patients N = 16
Age
Median 59
Range 36–73
Male:female 14:2
ECOG performance status
0 6 37%
1 8 50%
2 2 13%
Disease status
Primary metastatic 12 75%
Locally-advanced 4 25%
Prior chemotherapy
Uracil-tegafur 4 25%
5-FU bolus + folinic acid 1 6%
5-FU + cisplatin 4 25%
Epirubicin + cisplatin + capecitabine 5 31%
Epirubicin + doxorubicin + cisplatin 2 13%
Table 2 Dose-limiting toxicities (DLTs) according to dose level.
Dose level (5-FU/day) Pts/cycles After first cycle After all courses (pts/cycles)
Neutropenia FN1 stomatitis diarrhea asthenia
Level 1 (250 mg) 3/4 None 1/1 1/1
Level 2 (500 mg) 10/35 Stomatitis (1) 4/5 6/11 1/1 4/7
Level 3 (750 mg) 3/4 Stomatitis (1)
FN (1) 2/2 1/1 2/2 1/1 1/1
1 FN denotes febrile neutropenia.
==== Refs
Bae JM Won YJ Jung KW Park JG Annual report of the Korean central cancer registry program 2000 Cancer Res Treat 2002 34 77 83
Pyrhonen S Kuitunen T Nyandoto P Kouri M Randomised comparison of fluorouracil, epidoxorubicin and methotrexate (FEMTX) plus supportive care with supportive care alone in patients with non-resectable gastric cancer Br J Cancer 1995 71 587 591 7533517
Glimelius B Ekstrom K Hoffman K Graf W Sjoden PO Haglund U Svensson C Enander LK Linne T Sellstrom H Heuman R Randomized comparison between chemotherapy plus best supportive care with best supportive care in advanced gastric cancer Ann Oncol 1997 8 163 168 9093725 10.1023/A:1008243606668
Murad AM Chemotherapy for Advanced Gastric Cancer: Focus on New Agents and Combinations Cancer Control 1999 6 361 368 10758568
Seifert P Baker LH Reed ML Vaitkevicius VK Comparison of continuously infused 5-fluorouracil with bolus injection in treatment of patients with colorectal adenocarcinoma Cancer 1975 36 123 128 1203840
Lokich JJ Ahlgren JD Gullo JJ Philips JA Fryer JG A prospective randomized comparison of continuous infusion fluorouracil with a conventional bolus schedule in metastatic colorectal carcinoma: a Mid-Atlantic Oncology Program Study J Clin Oncol 1989 7 425 432 2926468
Cullinan SA Moertel CG Wieand HS O'Connell MJ Poon MA Krook JE Mailliard JA Tschetter LK Controlled evaluation of three drug combination regimens versus fluorouracil alone for the therapy of advanced gastric cancer. North Central Cancer Treatment Group J Clin Oncol 1994 12 412 416 8113849
Ohtsu A Shimada Y Shirao K Boku N Hyodo I Saito H Yamamichi N Miyata Y Ikeda N Yamamoto S Fukuda H Yoshida S Randomized phase III trial of fluorouracil alone versus fluorouracil plus cisplatin versus uracil and tegafur plus mitomycin in patients with unresectable, advanced gastric cancer: The Japan Clinical Oncology Group Study (JCOG9205) J Clin Oncol 2003 21 54 59 12506170 10.1200/JCO.2003.04.130
Ringel I Horwitz SB Studies with RP 56976 (taxotere): a semisynthetic analogue of taxol J Natl Cancer Inst 1991 83 288 291 1671606
Sulkes A Smyth J Sessa C Dirix LY Vermorken JB Kaye S Wanders J Franklin H LeBail N Verweij J Docetaxel (Taxotere) in advanced gastric cancer: results of a phase II clinical trial. EORTC Early Clinical Trials Group Br J Cancer 1994 70 380 383 7914428
Einzig AI Neuberg D Remick SC Karp DD O'Dwyer PJ Stewart JA Benson AB Phase II trial of docetaxel (Taxotere) in patients with adenocarcinoma of the upper gastrointestinal tract previously untreated with cytotoxic chemotherapy: the Eastern Cooperative Oncology Group (ECOG) results of protocol E1293 Med Oncol 1996 13 87 93 9013471
Burris HA 3rd Docetaxel in combination with fluorouracil for advanced solid tumors Oncology (Williston Park) 1997 11 50 52 9364544
Ando M Watanabe T Sasaki Y Ying DF Omuro Y Katsumata N Narabayashi M Tokue Y Fujii H Igarashi T Wakita H Ohtsu T Itoh K Adachi I Taguchi T A phase I trial of docetaxel and 5-day continuous infusion of 5-fluorouracil in patients with advanced or recurrent breast cancer Br J Cancer 1998 77 1937 1943 9667671
Van Den Neste E de Valeriola D Kerger J Bleiberg H Kusenda Z Brassinne C Bartholomeus S Selleslags J Hennebert P Wythouck H Cazenave I Lefresne-Soulas F Piccart M A phase I and pharmacokinetic study of docetaxel administered in combination with continuous intravenous infusion of 5-fluorouracil in patients with advanced solid tumors Clin Cancer Res 2000 6 64 71 10656433
Hawkins R Cunningham D Soerbye H Adenis A Canon JL Lopez-Vivanco G Jacques C Stenger P Zuber E Misset JL Randomized phase II trial of docetaxel plus irinotecan versus docetaxel plus 5-fluorouracil (5FU) in patients with untreated advanced gastric adenocarcinoma (AGAC) Proc Am Soc Clin Oncol 2003 22 257 (abstr 1032)
Park SH Kim MJ Chung M Lee WK Bang SM Cho EK Shin DB Lee JH Interim analysis from a prospective randomized trial of taxanes plus 5-FU in advanced gastric cancer Proc Am Soc Clin Oncol 2004 22 361 (abstr 4191)
Ozer H Armitage JO Bennett CL Crawford J Demetri GD Pizzo PA Schiffer CA Smith TJ Somlo G Wade JC Wade JL Winn RJ Wozniak AJ Somerfield MR 2000 update of recommendations for the use of hematopoietic colony-stimulating factors: evidence-based, clinical practice guidelines. American Society of Clinical Oncology Growth Factors Expert Panel J Clin Oncol 2000 18 3558 3585 11032599
Ajani JA Welch SR Raber MN Fields WS Krakoff IH Comprehensive criteria for assessing therapy-induced toxicity Cancer Invest 1990 8 147 159 2400936
Roth AD Maibach R Fazio N Sessa C Stupp R Morant R Herrmann R Borner MM Goldhirsch A de Braud F 5-Fluorouracil as protracted continuous intravenous infusion can be added to full-dose docetaxel (Taxotere)-cisplatin in advanced gastric carcinoma: a phase I-II trial Ann Oncol 2004 15 759 764 15111343 10.1093/annonc/mdh187
Ajani JA Van Cutsem E Moiseyenko V Tjulandin S Fodor M Majlis A Boni C Zuber E Blattmann A Docetaxel (D), cisplatin, 5-fluorouracil compare to cisplatin (C) and 5-fluorouracil (F) for chemotherapy-naive patients with metastatic or locally recurrent, unresectable gastric carcinoma (MGC): Interim results of a randomized phase III trial (V325) Proc Am Soc Clin Oncol 2003 22 249 (abstr 999)
Thuss-Patience PC Kretzschmar A Repp M Kingreen D Hennesser D Micheel S Pink D Scholz C Dorken B Reichardt P Docetaxel and Continuous-Infusion Fluorouracil Versus Epirubicin, Cisplatin, and Fluorouracil for Advanced Gastric Adenocarcinoma: A Randomized Phase II Study J Clin Oncol 2005 23 494 501 15659494 10.1200/JCO.2005.02.163
|
16042786
|
PMC1183196
|
CC BY
|
2021-01-04 16:03:05
|
no
|
BMC Cancer. 2005 Jul 22; 5:87
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-87
|
oa_comm
|
==== Front
BMC Cardiovasc DisordBMC Cardiovascular Disorders1471-2261BioMed Central London 1471-2261-5-141597209910.1186/1471-2261-5-14Research ArticleThe morbidity and mortality following a diagnosis of peripheral arterial disease: Long-term follow-up of a large database Caro Jaime [email protected] Kristen [email protected] Khajak J [email protected] Irina [email protected] Caro Research Institute, Concord, MA, USA2 Division of General Internal Medicine, McGill University, Montreal, Quebec, Canada3 Caro Research, Montreal, Quebec, Canada2005 22 6 2005 5 14 14 4 3 2005 22 6 2005 Copyright © 2005 Caro et al; licensee BioMed Central Ltd.2005Caro 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
Awareness of the significance of peripheral arterial disease is increasing, but quantitative estimates of the ensuing burden and the impact of other risk factors remains limited. The objective of this study was to fill this need.
Methods
Morbidity and mortality were examined in 16,440 index patients diagnosed with peripheral arterial disease in Saskatchewan, Canada between 1985 and 1995. Medical history and patient characteristics were available retrospectively to January 1980 and follow-up was complete to March 1998. Crude and adjusted event rates were calculated and Kaplan-Meier survival curves estimated. Cox proportional hazards analyses were conducted to examine the effect of risk factors on these rates. Patients suffering a myocardial infarction or ischemic stroke in Saskatchewan provided two reference populations.
Results
Half of the index patients were male; the majority was over age 65; 73% had at least one additional risk factor at index diagnosis; 10% suffered a subsequent stroke, another 10% a myocardial infarction, and 49% died within the mean follow-up of 5.9 years. Annual mortality (8.2%) was higher among patients with PAD than after a myocardial infarction (6.3%) but slightly lower than that in patients suffering a stroke (11.3%). Index patients with comorbid disease (e.g., diabetes) were at highest risk of death and other events.
Conclusion
A diagnosis of peripheral arterial disease is critical evidence of more widespread atherothrombotic disease, with substantial risks of subsequent cardiovascular events and death. Given that the majority has additional comorbidities, these risks are further increased.
==== Body
Background
Atherothrombosis is a chronic affliction of the arterial vascular tree. While its coronary and intracerebral manifestations are well recognized, its effects on the peripheral vasculature and the ensuing complications are less so. This condition, peripheral arterial disease, is best known for the ischemic pain it produces, but significant impairment can exist before it is symptomatic. Estimates of the prevalence of peripheral arterial disease vary widely, from 4.3% to 57%, depending on how the disease is identified, the age and risk factor distributions [1-6]. Advancing age, male gender, a history of diabetes and smoking are associated with a higher risk of peripheral arterial disease [7,8]. Approximately one-quarter to more than one-half of patients diagnosed with peripheral arterial disease experience the frequently painful symptoms associated with this diagnosis. Rates of newly diagnosed disease are in the range of 7% to 13% per year [1,7-10].
Awareness and understanding of the significance of a peripheral arterial disease diagnosis is increasing, but precise quantitative estimates of the burden and our understanding of the prognosis and impact of other risk factors remains limited [11]. Documentation of the burden of this disease can provide the means to better understand its course and, in turn, to make more effective decisions in the management of patients. In this paper, we report the results of analyses of the morbidity and mortality experienced by a large cohort of patients diagnosed with peripheral arterial disease.
Methods
Study population
The health care databases maintained by Saskatchewan Health were the source of data for this study. Saskatchewan Health is a provincial government department that administers a universal public health insurance program for approximately one million residents. Its administrative databases (e.g., formulary outpatient prescriptions, physician services, hospitalizations, and vital statistics) can be linked electronically using unique patient identifiers [12].
Residents of Saskatchewan, Canada, covered by Saskatchewan Health and eligible for outpatient prescription drug benefits (the prescription drug plan does not capture the claims of Registered Canadians of First Nations), diagnosed with peripheral arterial disease between January 1, 1985 and December31, 1995, during a hospitalization or physician consultation were eligible if at least 21 years of age at the time of diagnosis; no other exclusion criteria were specified. Patients with peripheral arterial disease were identified using the International Classification of Diseases (ICD-9) codes 440, 440.2, or 443.9. The broader three-digit code 443 was allowed if it coincided with documentation of a prescription for pentoxifylline on the assumption that all patients who received this prescription were diagnosed with peripheral arterial disease [13]. Patient history was available from January 1, 1980 and follow-up through March 1998 or until a patient could no longer be followed due to emigration or death. The date of first diagnosis, either hospital admission or physician visit depending on where the index diagnosis was recorded, was taken as the date of entry into the study.
Reference populations
Analyses were also carried out in reference populations comprised separately of Saskatchewan resident's age 21 years and older, without preexisting peripheral arterial disease, who suffered a myocardial infarction (ICD-9 code 410) or stroke (ICD-9 codes 433, 434, 436, and 362.3) between January 1, 1990 and December31, 1995. The same selection criteria used to identify the subjects with peripheral arterial disease were applied to these populations. The first diagnosis or event on record within the study dates was used to define the populations. For example, a subject with a diagnosis of peripheral arterial disease in October 1988 and a subsequent myocardial infarction recorded in January 1990 was assigned to the peripheral arterial disease cohort.
Outcomes and risk factors
Hospitalization records subsequent to the index diagnosis were used to identify outcome events, namely angina, transient ischemic attack, myocardial infarction, or stroke. To estimate the event rates, each event was considered independent of the others. The hospital admission date was taken as the time the outcome took place. Deaths were determined from the vital statistics file.
For subjects whose index event was a myocardial infarction, subsequent hospitalizations for myocardial infarction were only considered if they occurred at least seven days after the index event. Similarly, recurrence of stroke had to be at least seven days after an index stroke. For the outcome cluster consisting of angina, transient ischemic attack, myocardial infarction, stroke or cardiovascular death, the first occurrence of any of these events was taken.
Event-free time and survival were measured from the date of the index event (index date) to the time of hospitalization for the outcome or the date of death, as appropriate. If no outcome event occurred, then the subject's time was considered right-censored and the study end date (or date of emigration) was taken as the end of the period.
The patients' medical history prior to index event (available back to January 1980) was examined to identify other diseases documented prior to the qualifying diagnosis. Risk factors of interest included: male gender, age greater than 65 years at index diagnosis, prior myocardial infarction, prior stroke, diabetes, hypercholesterolemia, congestive heart failure, angina, atrialfibrillation, hypertension, and transient ischemic attack. The presence of each risk factor was determined based on ICD-9 coding of diagnoses recorded at hospitalizations or doctor visits prior to the index diagnosis. Given the relatively long inclusion period, we also considered the possibility of a calendar trend by including an indicator of the time of diagnosis of peripheral arterial disease, prior to January 1, 1990 or afterwards.
Analyses
Descriptive analyses were carried out to characterize the study and reference populations in terms of age, gender, setting of diagnosis (hospital versus outpatient), and medical history. The occurrence of each outcome event was independently estimated as a hazard (events/patient years). The Kaplan-Meier approach was used to determine the cumulative proportion (%) of patients suffering any one of the outcome events over time. To assess the impact of individual risk factors on the rate of outcome events, Cox proportional hazards analyses were conducted. Next, the number of risk factors present at diagnosis was tallied and the average hazard rates of myocardial infarction, stroke, and death in each group were compared to those of the patients with no risk factors at diagnosis. Comparisons were also made by gender.
Mortality among patients with peripheral arterial disease was evaluated using the Kaplan-Meier method and through estimation of average hazard rates. These were compared to post-acute mortality in each reference population (myocardial infarction and stroke) – deaths recorded in the first 30 days following those acute index events were excluded to accord better with the chronic nature of the diagnoses. The patients' medical history prior to index event (available back to January 1980) was examined to identify other diseases documented prior to the qualifying diagnosis.
Results
Population characteristics
The study population consisted of 16,440 patients diagnosed with peripheral arterial disease between 1985 and 1995 with an average follow-up of 5.9 (± 3.6) years. Most patients (90%) were identified on a physician visit and a majority presented with at least one comorbid condition at the time of diagnosis, as displayed in Table 1. Reference populations were comprised of 15,590 patients who suffered a myocardial infarction, 2,700 of whom died during the first 30 days, and 18,704 who suffered a stroke, of whom 1,877 died during the first 30 days. One-third of patients with an index stroke and one-half of patients with an index myocardial infarction were identified in hospital (Table 1).
Table 1 Patient populations. Patient characteristics and medical history for each population at the time of the index diagnosis. Age is reported as mean ± standard deviation. All other values are reported as number (%).
Reference Populations
Item PAD
N = 16,440 MI
N = 15,590 Stroke
N = 18,704
Age (years) 67.3 ± 9.2 66.9 ± 11.1 70.5 ± 9.9
Male 9,030 (55) 9,955 (64) 9,723 (52)
Qualifying diagnosis in hospital 1,659 (10) 8,954 (56) 6,733 (36)
Medical History
Atrial Fibrillation 862 (5) 905 (6) 1,812 (10)
Angina 5,446 (33) 6,435 (41) 4,849 (26)
Congestive Heart Failure 4,178 (25) 3,902 (25) 5,324 (29)
Diabetes Mellitus 3,172 (19) 3,693 (24) 4,487 (24)
Hypercholesterolemia 1,128 (7) 1,498 (10) 1,183 (6)
Hypertension 9,554 (58) 9,502 (61) 12,603 (67)
Ischemic Stroke 2,111 (13) 1,001 (6) 4,408 (24)
Myocardial Infarction 2,868 (18) 3,672 (24) 2,211 (12)
Transient Ischemic Attack 2,217 (14) 1,372 (9) 5,151 (28)
Outcome events
Subsequent myocardial infarction, stroke, and angina each occurred in approximately 10% of the peripheral arterial disease population at rates varying between 16.8 and 17.1 per 1,000 patient years (Table 2). Transient ischemic attack occurred in only 3.7% of these patients at a rate of 6.3 per 1,000 patient years.
Table 2 Event risk. First cardiovascular events of each type and all-cause deaths subsequent to a diagnosis of peripheral arterial disease. Event rates are reported per 1,000 person-years.
All Patients Males Females
Outcome Events (%) PY Event Rate Events (%) PY Event Rate Events (%) PY Event Rate
Angina 1,555 (9.5) 90,734 17.1 1,012 (11.2) 48,363 20.9 543 (7.3) 42,371 12.8
TIA 601 (3.7) 94,768 6.3 306 (3.4) 51,526 5.9 295 (4.0) 43,242 6.8
MI 1,563 (9.5) 93,160 16.8 1,012 (11.2) 49,956 20.3 551 (7.4) 43,204 12.8
Stroke 1,589 (9.7) 93,066 17.1 917 (10.2) 50,309 18.2 672 (9.1) 42,757 15.7
Death 7,973 (48.5) 96,771 82.4 4,519 (50.0) 52,503 86.1 3,454 (46.6) 44,269 78.0
Only half of all patients remained alive at the end of follow-up. The crude five-year death rate among patients diagnosed with peripheral arterial disease was 33.2% – a rate of 82.4 deaths per 1,000 patient years (Figure 1). When all outcome events and cardiovascular deaths are evaluated together, it is estimated that 16.0% of patients had an event by the end of the first year after diagnosis and this rose to about one-third by three years. In comparison, 13.5% of patients in the myocardial infarction reference population were re-hospitalized for a recurrent myocardial infarction, at a rate of 41.2 per 1,000 person-years, and 12.0% with angina (36.3 per 1,000 person years). Less than 5% suffered a stroke (13.0 per 1,000 person years) or a transient ischemic attack (4.6 per 1,000 person years). Among patients in the stroke reference group, less than 5% suffered a myocardial infarction (12.4 per 1,000 person years), transient ischemic attack (11.1 per 1,000 person years) or angina (7.3 per 1,000 person years), whereas, 16.6% had a recurrent stroke (54.7 per 1,000 person years). Approximately two-thirds (59.6%) of patients with a myocardial infarction and one-half (51.3%) of those with a stroke were alive at the end of follow-up.
Figure 1 Patient survival. Survival among patients with peripheral arterial disease compared to patients suffering a stroke or myocardial infarction.
Mortality
The crude five-year death rate among patients diagnosed with peripheral arterial disease was 33.2%. When adjusted for duration of follow-up, the rate is 82.4 deaths per 1,000 patient years. Compared to the reference populations, mortality is lower than among patients who suffer an index stroke (41.8% crude rate; 113.1 per 1,000 patient years adjusted) but higher than that among patients who suffer an index myocardial infarction (26.6%; 63.2 per 1,000 patient years). Survival in each group, excluding deaths during the acute period (first 30 days) for index myocardial infarction and index stroke, is displayed in Figure 1.
Impact of risk factors
On average, women were older than the men (68.2 ± 8.9 vs 66.5 ± 9.4). Over the course of an average follow-up of six years, men were considerably more likely to manifest further atherothrombotic complications (Table 2): angina (63% more in men), myocardial infarction (59% more), stroke (16% more), and death (10% more). Transient ischemic attack, however, occurred less often in men (13% less).
As might be expected, prior myocardial infarction (hazard ratio = 1.53, 95% CI 1.35–1.73) was a strong risk factor for a subsequent myocardial infarction. A prior stroke was also a predictor of myocardial infarction (hazard ratio = 1.26, 1.08–1.46) and, more strongly, of subsequent stroke (hazardratio = 2.23, 1.96–2.53). In addition, a transient ischemic attack (hazard ratio = 1.35, 1.18–1.54) was a powerful predictor of stroke. Even amongst patients with no history of atherothrombotic disease (myocardial infarction or stroke), however, 7.5% suffered a myocardial infarction and 8.0% a stroke – rates of 13.5 and 14.0 events per 1,000 patient years, respectively.
Cox proportional hazards analyses conducted to determine the effect of various potential risk factors show that the majority of the eleven considered are significant predictors of future stroke, myocardial infarction and death. Apart from prior myocardial infarction or stroke, the risk of myocardial infarction was significantly increased with male gender (hazard ratio = 1.53, 1.38–1.71), age over 65 years (hazard ratio = 1.50, 1.34–1.69), angina (hazard ratio = 1.47, 1.31–1.64), diabetes mellitus (hazard ratio = 1.50, 1.33–1.68), heart failure (hazard ratio = 1.15, 1.02–1.30), and hypertension (hazard ratio = 1.31, 1.18–1.46). Many of these same conditions also significantly increased the risk of stroke: male gender (hazard ratio = 1.26, 1.13–1.39), age over 65 years (hazard ratio = 1.98, 1.73–2.26), atrial fibrillation (hazard ratio = 1.64, 1.36–1.98), diabetes mellitus (hazard ratio = 1.50, 1.34–1.69), heart failure (hazard ratio = 1.20, 1.07–1.35), and hypertension (hazard ratio = 1.51, 1.35–1.69). Increased risk of death was associated with male gender (hazard ratio = 1.27, 1.21–1.32), age over 65 years (hazardratio = 3.65, 3.39–3.92), atrial fibrillation (hazard ratio = 1.25, 1.15–1.35), diabetes mellitus (hazard ratio = 1.40, 1.32–1.47), heart failure (hazard ratio = 2.06, 1.97–2.17), and prior stroke(hazard ratio = 1.53, 1.44–1.62).
When myocardial infarction, stroke, angina, transient ischemic attack, and cardiovascular death are evaluated as a cluster endpoint, 10 of the 11 risk factors were significantly predictive (p ≤ 0.0006) of future events: male gender (hazard ratio = 1.29, 1.23–1.36), age over 65 years (hazard ratio = 1.78, 1.67–1.89), atrial fibrillation (hazard ratio = 1.24, 1.13–1.36), angina (hazard ratio = 1.25, 1.19–1.32), diabetes mellitus (hazard ratio = 1.35, 1.28–1.43), heart failure (hazard ratio = 1.59, 1.51–1.68), hypertension (hazard ratio = 1.11, 1.06–1.17), prior myocardial infarction (hazard ratio = 1.27, 1.19–1.35), prior stroke (hazardratio = 1.51, 1.42–1.62), and prior transient ischemic attack (hazard ratio = 1.12, 1.05–1.20).
The extent to which the risk increases depends on the number of risk factors present at diagnosis of peripheral arterial disease (Figure 2). Compared to patients with no risk factors at diagnosis (n = 516), presence of one or more risk factors increases the risk of myocardial infarction by 4.08%, of stroke by 5.04% and of death by 5.82%.
Figure 2 Relative event risk. Risk increase in events for various levels of risk factors, compared to no risk factor.
To further examine the impact of a prior diagnosis of myocardial infarction or stroke, the 27.7% of patients with at least one of these conditions were evaluated separately. Even in this group, occurrence of a subsequent myocardial infarction was significantly increased with male gender (hazard ratio = 1.57, 1.30–1.90), age over 65 years (hazard ratio = 1.20, 1.00–1.44), angina (hazard ratio = 1.44, 1.21–1.72), diabetes mellitus (hazard ratio = 1.28, 1.06–1.54), and hypertension (hazard ratio = 1.30, 1.09–1.55).
As the study spans more than a decade, the effect of calendar time was examined. Patients diagnosed after January 1, 1990 had an 11.1 % lower risk of subsequent myocardial infarction but there was no effect on stroke or death. Amongst patients with a history of myocardial infarction or stroke, the risk of myocardial infarction was reduced (by 22.3%) for patients with their peripheral arterial disease diagnosed in the 1990s.
Discussion
Although awareness of peripheral arterial disease and the factors associated with its onset has improved in recent years, to our knowledge, this study provides the first large-scale, long-term comprehensive analyses of the prognosis of patients diagnosed with peripheral arterial diseasetogether with a comparison to those diagnosed first with myocardial infarction or stroke [3-11,14-18]. These analyses expand on current knowledge by providing quantitative estimates of the risks faced by patients diagnosed with peripheral arterial disease and the impact of comorbidities.
A diagnosis of peripheral arterial disease emerged from our analyses as critical evidence of more widespread atherothrombotic disease. The risk of subsequent cardiovascular events is substantial among patients diagnosed with peripheral arterial disease even when prior atherothrombotic disease (myocardial infarction and prior ischemic stroke) is considered. The majority of these patients have additional cardiovascular risk factors at the time of diagnosis, thus contributing to these increased risks.
In comparisons with patients suffering an acute stroke or myocardial infarction, comparable proportions of patients with peripheral arterial disease suffered subsequent events and death. Generally, risks among patients with peripheral arterial disease were observed to be lower than those suffering a stroke and higher than those experiencing a myocardial infarction. Exceptions included higher rates of subsequent myocardial infarction and angina among patients in the myocardial infarction group and lower rates of myocardial infarction and transient ischemic attack among patients in the stroke group. These results suggest that patients diagnosed with peripheral arterial disease are at comparable risks for experiencing subsequent events.
These findings are consistent with those of patients in a six-year follow-up study who were identified as having disease on non-invasive screening measurement of the ankle-brachial index. Approximately half suffered a major acute event (stroke or myocardial infarction) or underwent major surgery (amputation or vascular surgery) within six years of referral to a specialty clinic [17]. Another one quarter of patients died during this period. Over ten years, this estimate is notably higher with 62% of men and 33% of women expiring within this time-more than two times the rate observed in men without a diagnosis of peripheral arterial disease and approximately twice the rate among women [18]. Concomitant diagnoses, such as diabetes mellitus and hypertension were associated with significant risk of morbidity and mortality. As well, atrial fibrillation was linked to an increase in acute emergency admissions and in-hospital mortality in a recent study of patients with symptomatic peripheral arterial disease [19]. Evaluation of the medical treatment received by patients in this cohort was beyond the scope of this study and was not considered in our analyses.
Analyses conducted using administrative data such as those we used from Saskatchewan Health have some limitations. Definition of the populations and identification of events is dependent on accuracy of the ICD-9 codes submitted to Saskatchewan Health. Validation work has indicated, however, a low error rate for cardiovascular diagnoses [12]. Another problem with administrative data is that they do not contain clinical information such as results of laboratory tests, smoking, family history, blood pressure, and so on. Thus, these potentially relevant factors could not be incorporated in the analyses.
Conclusion
Despite prior data in favor of treating peripheral arterial disease as evidence of disseminated atherothrombosis, the management of patients with this diagnosis remains less aggressive than for patients suffering cardiovascular events [3,20]. The results of this study reinforce the call for attention to peripheral arterial disease as a serious diagnosis with important implications for patient management. In particular, identification of other risk factors and risk reduction through pharmacologic or other management strategies ought to be aggressively pursued in this population, as is already common practice among patients who suffer a myocardial infarction or stroke.
Abbreviations
CI: Confidence Interval
ICD-9: International Classification of Diseases, version 9
MI: Myocardial Infarction
PAD: Peripheral Arterial Disease
PY: Person-Years
RF: Risk Factors
TIA: Transient Ischemic Attack
Competing interests
This research was supported in part by an unrestricted grant from Sanofi-Aventis & Bristol-Myers Squibb. Neither funding agency had any role in the design, execution of the analyses, their interpretation, or the writing of the paper. The authors had sole responsibility for all methodological decisions.
Authors' contributions
JC designed the study, participated in the data analyses and writing of the paper. KMW participated in the study design, data analyses and in writing the paper. KJI. led the analyses and participated in writing the paper. IP conducted the analyses and helped write the paper.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study is based in part on de-identified data provided by the Saskatchewan Department of Health. The interpretation and conclusions contained herein do not necessarily represent those of the Government of Saskatchewan or the Saskatchewan Department of Health.
==== Refs
Stoffers HE Rinkens RP Kester AD Kaiser V Knottnerus JA The prevalence of asymptomatic and unrecognized peripheral arterial occlusive disease Int J Epidemiol 1996 25 282 290 9119553
Selvin E Erlinger TP Prevalence and risk factors for peripheral arterial disease in the United States. Results from the National Health and Nutrition Examination Survey, 1999–2000 Circulation 2004 110 738 743 15262830
Hirsch AT Criqui MH Treat-Jacobson D Regensteiner JG Creager MA Olin JW Krook SH Hunninghake DB Comerota AJ Walsh ME McDermott MM Hiatt WR Peripheral arterial disease detection, awareness, and treatment in primary care JAMA 2001 286 1317 1324 11560536
Criqui MH Systematic atherosclerosis risk and the mandate for intervention in atherosclerotic peripheral arterial disease Am J Cardiol 2001 88 43 47
Weitz JI Byrne J Clagett P Farkouh ME Porter JM Sackett DL Strandness DE Taylor LM Diagnosis and treatment of chronic arterial insufficiency of the lower extremities: A critical review Circulation 1996 94 3026 3049 8941154
Criqui MH Peripheral arterial disease – epidemiological aspects Vasc Med 2001 3 7 11789963
Price PF Mowbray PI Lee AJ Rumley A Lowe GD Fowkes FG Relationship between smoking and cardiovascular risk factors in the development of peripheral arterial disease and coronary artery disease Eur Heart J 1999 20 344 353 10206381
McDermott MM Greenland P Liu K Guralnik JM Crique MH Dolan NC Chan C Celic L Pearce WH Schneider JR Sharma L Clark E Gibson D Martin GJ Leg symptoms in peripheral arterial disease: Associated clinical characteristics and functional impairment JAMA 2001 286 1599 1606 11585483
Hooi JD Stoffers HE Kester AD Rinkens PE Kaiser V van Ree JW Knottnerus JA Risk factors and cardiovascular diseases associated with asymptomatic peripheral arterial occlusive disease. The Limburg PAOD Study. Peripheral Arterial Occlusive Disease Scand J Prim Health Care 1998 16 177 182 9800232
Fowkes F Housley E Cawood EH MacIntyre CC Ruckley CV Prescott RJ Edinburgh Artery Study: Prevalence of asymptomatic and symptomatic peripheral arterial disease in the general population Int J Epidemiol 1991 20 384 392 1917239
Newman AB Peripheral arterial disease: Insights from population studies of older adults J Am Geriatr Soc 2000 48 1157 1162 10983919
Downey W Beck P McNutt M Stang MR Osei W Nichol J Strom BL Health Databases in Saskatchewan Pharmacoepidemiology 2000 Third Chichester: Wiley 325 345
Stergachis A Sheingold S Luce BR Psaty BM Revicki DA Medical care and cost outcomes after pentoxifylline treatment for peripheral arterial disease Arch Intern Med 1992 152 1220 1224 1599350
Meijer WT Grobbee DE Hunink MGM Hofman A Hoes AW Determinants of peripheral arterial disease in the elderly: The Rotterdam Study Arch Intern Med 2000 160 2934 2938 11041900
Bowlin SJ Medalie JH Flocke SA Zyzanski SJ Yaari S Goldbourt U Intermittent claudication in 8343 men and 21-year specific mortality follow-up Ann Epidemiol 1997 7 180 187 9141640
Hiatt WR Medical treatment of peripheral arterial disease and claudication N Engl J Med 2001 344 1608 1621 11372014
Howell MA Colgan MP Seeger RW Ramsey DE Sumner DS Relationship of severity of lower limb peripheral vascular disease to mortality and morbidity: A six-year follow-up study J Vasc Surg 1989 9 691 697 2724457
Criqui MH Langer RD Fronek A Feigelson HS Klauber MR McCann TJ Browner D Mortality over a period of 10 years in patients with peripheral arterial disease N Engl J Med 1992 326 381 386 1729621
Conway DSG Lip GYH Comparison of outcomes of patients with symptomatic peripheral artery disease with and without atrial fibrillation (the West Birmingham Atrial Fibrillation Project) Am J Cardiol 2004 93 1422 1425 15165931
McDermott MM Mehta S Ahn H Greenland P Atherosclerotic risk factors are less intensively treated in patients with peripheral arterial disease than in patients with coronary artery disease J Gen Intern Med 1997 12 209 215 9127224
|
15972099
|
PMC1183197
|
CC BY
|
2021-01-04 16:30:06
|
no
|
BMC Cardiovasc Disord. 2005 Jun 22; 5:14
|
utf-8
|
BMC Cardiovasc Disord
| 2,005 |
10.1186/1471-2261-5-14
|
oa_comm
|
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-271598241210.1186/1471-2121-6-27Research ArticleDermal fibroblasts in Hutchinson-Gilford progeria syndrome with the lamin A G608G mutation have dysmorphic nuclei and are hypersensitive to heat stress Paradisi Mauro [email protected] Dayle [email protected] Revekka L [email protected] Christina [email protected] Howard J [email protected] Karima [email protected] VII Divisione, Dermatologia Pediatrica, Istituto Dermopatico Dell'Immacolata IRCCS, Rome, Italy2 Department of Dermatology, Columbia University, College of Physicians & Surgeons, New York, New York, USA3 Department of Medicine and Department of Anatomy and Cell Biology, Columbia University, College of Physicians & Surgeons, New York, New York, USA2005 27 6 2005 6 27 27 30 11 2004 27 6 2005 Copyright © 2005 Paradisi et al; licensee BioMed Central Ltd.2005Paradisi 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
Hutchinson-Gilford progeria syndrome (HGPS, OMIM 176670) is a rare sporadic disorder with an incidence of approximately 1 per 8 million live births. The phenotypic appearance consists of short stature, sculptured nose, alopecia, prominent scalp veins, small face, loss of subcutaneous fat, faint mid-facial cyanosis, and dystrophic nails. HGPS is caused by mutations in LMNA, the gene that encodes nuclear lamins A and C. The most common mutation in subjects with HGPS is a de novo single-base pair substitution, G608G (GGC>GGT), within exon 11 of LMNA. This creates an abnormal splice donor site, leading to expression of a truncated protein.
Results
We studied a new case of a 5 year-old girl with HGPS and found a heterozygous point mutation, G608G, in LMNA. Complementary DNA sequencing of RNA showed that this mutation resulted in the deletion of 50 amino acids in the carboxyl-terminal tail domain of prelamin A. We characterized a primary dermal fibroblast cell line derived from the subject's skin. These cells expressed the mutant protein and exhibited a normal growth rate at early passage in primary culture but showed alterations in nuclear morphology. Expression levels and overall distributions of nuclear lamins and emerin, an integral protein of the inner nuclear membrane, were not dramatically altered. Ultrastructural analysis of the nuclear envelope using electron microscopy showed that chromatin is in close association to the nuclear lamina, even in areas with abnormal nuclear envelope morphology. The fibroblasts were hypersensitive to heat shock, and demonstrated a delayed response to heat stress.
Conclusion
Dermal fibroblasts from a subject with HGPS expressing a mutant truncated lamin A have dysmorphic nuclei, hypersensitivity to heat shock, and delayed response to heat stress. This suggests that the mutant protein, even when expressed at low levels, causes defective cell stability, which may be responsible for phenotypic abnormalities in the disease.
==== Body
Background
Hutchinson-Gilford progeria syndrome (HGPS, OMIM 176670) is a rare sporadic disorder with an incidence of 1 per 8 million live births. Birth weight and appearance are usually normal, but growth typically becomes retarded at the age of 1 year. Phenotypic features include short stature, sculptured nose, alopecia, prominent scalp veins, small face, subcutaneous fat loss, faint mid-facial cyanosis, and dystrophic nails. Features occurring in the skin during late adulthood of normal individuals, such as hair greying, hair loss, and skin thinning, occur in the first few years of life in subjects with HGPS [1-3]. Most subjects die in their teenage years from cardiac complications of coronary artery disease or stroke due to widespread arteriosclerosis [2].
The diagnosis of HGPS was formerly based on the criteria of growth retardation and prematurely aged phenotype in children. In 2003, however, mutations in the LMNA gene that encodes nuclear lamins A and C were identified as responsible for this syndrome [4-6]. As such, HGPS belongs to the group of diseases caused by mutations in LMNA, sometimes referred to as "laminopathies," which also includes disorders of striated muscle, peripheral nerve and partial lipodystrophy syndromes [7,8]. The LMNA mutation present in the majority of subjects with HGPS is a de novo heterozygous base change (GGC>GGT) within exon 11 of the LMNA gene, which does not cause an amino acid substitution (G608G) but creates an abnormal splice donor site [4,5].
Nuclear lamins are members of the intermediate filament protein superfamily [9,10]. They are the building blocks of the nuclear lamina, a fibrous proteinaceous meshwork underlying the inner nuclear membrane [11]. Nuclear lamins have an extra 42 amino acids (six heptads) in coil 1B compared to cytoplasmic intermediate filament proteins [12,13]. Nuclear lamins also contain two unique sequences: a nuclear localization signal in the tail domain [14], and, except for lamin C, a carboxyl-terminal CAAX box (cysteine-aliphatic-aliphatic-any amino acid), a target for isoprenylation [15-18]. In the human genome, 3 distinct loci encode lamins. LMNA is located on chromosome 1q21.2 [19,20] and encodes 4 lamins by alternative RNA splicing: A, C, AΔ10, and C2 [21]. Lamin A is synthesized as a precursor, prelamin A, from which 17 amino acids are removed from the carboxyl-terminal by endoproteolysis after isoprenylation. There are 2 B-type lamin genes: LMNB1 on chromosome 5q23-q31.1 that encodes lamin B1 [20,22], and LMNB2 on chromosome 19p13.3 [23] that encodes lamin B2 [24] and lamin B3, an alternatively spliced isoform expressed in germ cells [25].
We recently identified a new female subject with HGPS from Italy. We now show that she has the most common heterozygous point mutation, G608G, in LMNA resulting in expression of the prelamin A mutant with 50 amino acids deleted from the carboxyl-terminal tail domain. We analyzed the nuclear morphology and growth characteristics of these fibroblasts, and for the first time demonstrate that cells from a subject with HGPS exhibit increased susceptibility to heat stress.
Results
Clinical description of a new subject with HGPS
The female subject is the second child from consanguineous parents (second cousins) (Fig. 1A). The subject's mother and an uncle (Fig. 1A, red) were affected with pseudoxanthoma elasticum, an inherited disorder of connective tissue. At birth, the subject had cutaneous xerosis and mild skin indurations of the lower limbs. At 1 month, she developed more severe skin stiffening on the lower limbs, trunk, and extensor areas on the forearms, accompanied by functional limitation in leg extension. Mild perioral cyanosis was also visible. A skin biopsy was performed at 2 years of age and histological analysis revealed fibrous thickening of the lower dermis, subcutaneous septa, and fascia, accompanied by a few mucinous deposits. Weigert staining showed pronounced rarefactions of elastic fibers. A dermal fibroblast culture was established from the biopsy sample. Physical examination of the subject at age 5 years revealed loss of subcutaneous tissue, especially on the face and limbs, and thickening of the skin that appeared shiny and taut in most areas. She also had a small face with a recessed chin, thin beaked nose, small ears, prominent eyes, prominent scalp veins, and alopecia.
Figure 1 Characterization of a new subject with HGPS. (A) Pedigree of the subject with HGPS. Filled black circle indicates the proband; red circle and square respectively represent the female (mother of proband) and male (uncle of proband) with pseudoxanthoma elasticum (PXE). A diagonal line indicates a deceased individual. Double lines indicate consanguinity. (B) Short portion of the LMNA sequence within exon 11. The paternal sequence (Father) corresponds to the wild type LMNA sequence. The affected subject (HGPS Patient) has a heterozygous C to T transition (indicated by a star) at nucleotide 1824 in exon 11 of the LMNA sequence.
Detection of a LMNA mutation encoding a truncated prelamin A
We sequenced all 12 exons of LMNA that comprise the lamin A/C coding region and splice junctions of the subject's DNA using previously described oligonucleotide primers [26]. We found a heterozygous C to T transition at nucleotide 1824 in exon 11 of LMNA, which created a silent point mutation at codon 608 (GGC>GGT; G608G) (Fig. 1B). We also sequenced exon 11 from the father (Fig. 1B) and the mother (data not shown) and found no mutation in either.
Amplification of cDNA fragments by RT-PCR corresponding to nucleotides 1561 to 2010 of prelamin A mRNA showed that the wild type fragment was 449 nucleotides and the mutant fragment 299 nucleotides. The mutant sequence was 150 nucleotides shorter than the wild type, corresponding to a deletion of 50 amino acids. Figure 2A shows the sequence of the mutant lamin A cDNA from nucleotide 1804 (in the codon for amino acid 602) to the last amino acid codon. The deducted amino acid sequence of the mutant from residue 602 to the carboxyl-terminus is GSGAQSPQNCSIM (Fig. 2A).
Figure 2 Sequencing of the mutant prelamin A cDNA and fibroblast morphology in the subject with HGPS. (A) Sequence of cDNA encoding mutant prelamin A in the subject with HGPS. Mutant prelamin A cDNA sequence from nucleotide 1804 to the end of the coding region and the corresponding deduced amino acid sequence are shown. (B) Confocal microscopic analysis of dermal fibroblast in primary culture from a control (a and b) and the subject with HGPS (c and d). Labelling was performed with anti-lamin A/C antibodies. Note the presence of irregularly shaped nuclear envelopes in many of the subject's fibroblasts. (C) Double label immunofluorescence microscopic analysis of fibroblasts from an unaffected control (a to c) and subject with HGPS (d to f) labelled with anti-lamin B1 (red)) and anti-vimentin (green) antibodies. Bars, 10 μm.
Nuclear envelope morphology in HGPS fibroblasts
We examined the nuclear envelope morphology of dermal fibroblasts from control individuals and from the subject with HGPS after 10 or less passages in primary cultures. Cells were fixed and immunostained with anti-lamin A/C antibodies and examined by confocal microscopy. Nuclei of control fibroblasts were mostly regular in size and shape, and appeared generally round or ovoid (Fig. 2B, a and b). In contrast, the size of nuclei in the subject's fibroblasts was more variable, and the nuclear envelopes had convex "blebs" or herniations projecting towards the cytoplasm (Fig. 2B, c and d). By direct count of 1,000 cells in 3 different cultures from the subject and controls, we determined that 19% of nuclei in the subject's fibroblasts had irregularities in nuclear envelope shape compared to only 4% of control cells, in which the dysmorphic nuclei exhibited less severe abnormalities. Despite abnormalities in nuclear envelope size and shape in fibroblasts from the subject with HGPS, there were no gross abnormalities in the localization of lamins A/C (Fig. 2B, c and d) or lamin B1 (Fig. 2C, d). Furthermore, the vimentin network appeared normal (Fig. 2C, e). Staining with 4',6-diamidino-2-phenylindole (DAPI) did not reveal obvious defects in chromatin density (data not shown).
We further examined the localizations of A-type lamins, B-type lamins, and emerin, an integral protein of the inner nuclear membrane, using double labelling immunofluorescence microscopy. We examined the most abnormal appearing nuclei in the subject's cells, carefully analyzing the localization of lamin B1 (Fig. 3b) and emerin (Fig. 3f) compared to A-type lamins (Fig. 3a and 3e). At the nuclear periphery of the subject's most dysmorphic nuclei, labelling of A-type and B-type lamins is super-imposable (Fig. 3c). Even in sites where the nuclear envelope contained "blebs," both A-type and B-type lamins were colocalized. Emerin also colocalized with A-type lamins in the same distribution pattern (Fig. 3g). When intranuclear lamin A/C labelling was present, emerin was found in the same structures, suggesting that they represented invaginations of the nuclear envelope. Concurrent DAPI staining did not demonstrate a gross detachment between the nuclear lamina (A-type and B-type lamin labelling) and chromatin, even in the most abnormally shaped nuclei (Fig. 3 a to d). Lamin A/C, B1 and emerin labelling were in close apposition to the DNA labelling, suggesting that the nuclear envelope and lamina remained in close approximation to chromatin. DNA labelling detected in the herniated areas of nuclei also was associated with labelling by anti-lamin A/C, anti-lamin B1 and anti-emerin antibodies (Fig. 3d and 3h).
Figure 3 Comparison of lamin A/C localization with lamin B1, emerin and chromatin in dysmorphic nuclei of fibroblasts from the subject with HGPS. Upper panel shows micrographs of fibroblasts from the subject with HGPS labelled with anti-lamin A/C antibodies (a), anti-lamin B1 antibodies (b), the merged signal (c) and DAPI (d). Lower panel shows micrographs of fibroblasts from the same subject labelled with anti-lamin A/C antibodies (e), anti-emerin antibodies (f), the merged signal (g), and DAPI (h). Bar, 10 μm.
We carried out an ultrastructural analysis of cultured fibroblasts from the subject with HGPS using transmission electron microscopy. The electron micrographs clearly showed that chromatin remained in contact with the nuclear envelope (Fig. 4a to 4c). Analysis of a dysmorphic nucleus with 3 blebs (Fig. 4a) revealed that the chromatin also remained attached to the nuclear envelope at those sites (Fig. 4b and 4c).
Figure 4 Ultrastructural analysis of the nuclear envelope in fibroblasts from the subject with HGPS. Low magnification transmission electron microscopic image of a passage 10 PT001 nucleus showed several herniations (a). Two higher-magnification images of the same nucleus at sites of blebs (b and c) showed a close apposition of the chromatin to the nuclear envelope. In a, b, and c the nucleus is to the left. Scale bars correspond to 2 μm in panel a, and 500 nm in panels b and c.
Analysis of lamin expression by Western blotting
We performed Western blot analysis on total protein extractions from cultured fibroblasts at passage number 10. Anti-lamin A/C antibodies that recognize a sequence predicted to be within the truncated G608G prelamin A mutant, in addition to lamin A and C bands, detected the mutant protein in the subject's cells migrating above the band corresponding to lamin C (Fig. 5, panel Lamin A/C, lanes 3 and 4, arrow). This protein was not detected in extracts of control fibroblasts (Fig. 5, panel Lamin A/C, lanes 1 and 2). In Western blots containing equal amounts of cell extracts, the lamin A and lamin C signals were approximately the same for subject and control cells (Fig. 5; panel Lamin A/C). This was confirmed by densitometric analysis of autoradiograms of 4 separate Western blots from subject and control cells. In addition, the signals on Western blots obtained using anti-lamin B1 and anti-emerin antibodies showed no obvious differences between the subject and unaffected individuals (Fig. 5, panels Lamin B1 and Emerin). Hence, at low passage number in primary culture, fibroblasts from a subject with HGPS express a mutant truncated prelamin A, but lamin A, lamin C, lamin B1, and emerin are expressed at approximately normal levels.
Figure 5 Detection of lamins A and C, lamin B1 and emerin by Western blotting in fibroblasts from the subject with HGPS. Fibroblasts from unaffected controls (lanes 1 and 2) and the subject with HGPS (lanes 3 and 4) were lysed in Laemmli buffer, and whole-cell extracts corresponding to 0.5 × 106 cells (lanes 1 and 3) and 0.25 × 106 (lanes 2 and 4) were analyzed by immunoblotting using anti-lamin A/C (lane: Lamin A/C), anti-lamin B1 (lane: Lamin B1) and anti-emerin (lane: Emerin) antibodies. Arrowhead points to the position of the prelamin A mutant in lanes 3 and 4.
Growth rate of fibroblasts from a subject with HGPS in early passage primary culture
We examined the growth rate of cultured primary dermal fibroblasts from the subject with HGPS. At each cellular passage, the total number of cells from control and proband was assessed by direct count prior to plating. The mean number of harvested HGPS fibroblasts was comparable to the number obtained with two control fibroblast lines derived from unaffected individuals at passage numbers in culture below passage 10. Dermal fibroblasts from the subject with HGPS therefore had a similar growth rate to the control counterparts and did not exhibit gross cell cycle defects at low passage numbers in culture.
Dermal fibroblasts from a subject with HGPS are hypersensitive to heat stress
To evaluate the resistance to stress of the nuclear envelope, fibroblasts from the subject with HGPS and control individuals were heat shocked for 30 minutes at 45°C. Cells were then either fixed immediately (time 0) or incubated at 37°C for 24 and 48 hours to allow for recovery. Nuclear shape was examined by microscopy after labelling with anti-lamin A/C or anti-lamin B1 antibodies at time 0, 24 hours and 48 hours after heat shock (Fig. 6). In control cells, nuclear shape and distribution of A-type and B-type lamins were not altered by heat shock (Fig. 6A a to c and Fig. 6B g to i, respectively). In contrast, extensive nuclear deformations appeared in the cells from the subject with HGPS (Fig. 6A d to f and Fig. 6B m to o). The subject's fibroblasts showed an increase in the number of cells with altered nuclei immediately after heat shock. Nuclear envelopes were deformed, many had a ruffled appearance, and some had "pleats" or "folds" (Fig. 6, d and m, time 0). The number of irregular nuclei increased to nearly 70 % after 24 hours, as determined by direct count of 1,000 nuclei in cells on different coverslips. Some nuclei were more severely affected than others and showed "blebbing" or nuclear lobe formation 24 hours after heat shock (Fig. 6, e and n, time 24). This type of nuclear damage was never observed in the control cells (Fig. 6; a, b, g, and h). At 48 hours after heat shock, nuclei with invaginations were no longer observed in the fibroblasts from the subject with HGPS (Fig. 6; f and o), suggesting that the cells with severely dysmorphic nuclei died and detached from the coverslips during processing for immunofluorescence microscopy. Fibroblasts remaining on the coverslips appeared to have recovered from the stress, as their nuclei showed a less ruffled appearance and the irregularities resembled those observed prior to the heat shock (Fig. 6; f and o, time 48). There was no apparent rearrangement of the vimentin cytoplasmic intermediate filament network after heat shock and 24 and 48 hours after recovery in the subject's and control fibroblasts (Fig. 6; p to r).
Figure 6 Confocal analysis of dermal fibroblasts after heat shock stress. Cells were incubated at 45°C for 30 minutes, then either immediately fixed (time 0) or allowed to recover for 24 or 48 hours at 37°C. Cells were processed for indirect immunofluorescence labelling using anti-lamin A/C antibodies (panel A, a to f), anti-lamin B1 antibodies (panel B, g to r) and anti-vimentin antibodies (panel B j to i and p to r). Control and HGPS fibroblasts are indicated. (A) Fibroblasts were immunostained with anti-lamin A/C antibodies. Note the increased number of dysmorphic nuclei in HGPS fibroblasts compared to control 24 hours after recovery from heat shock. (B) Fibroblasts were immunostained with anti-lamin B1 and vimentin at times indicated after heat shock. Bar, 10 μm.
Survival rates of fibroblasts subjected to heat shock were evaluated and compared to unheated cells (Table 1). At 24 hours after recovery, 25% of the subject's cells were lost compared to the number of cells prior to treatment. In contrast, the total number of control fibroblasts increased by 39%. This suggests that growth of control cells was not significantly affected by heat shock, as they continued to divide. After 48 hours, the subject's fibroblasts appeared to have recovered from stress, as the number of cells increased, reaching approximately the same number of cells per dish as before heat shock. From 24 hours to 48 hours after heat shock, the number of cells from the subject with HGPS increased by 30%. Control cells continued to increase by an average of 36% at 48 hours after heat shock.
Table 1 Percentage of cells recovered after heat shock treatment. Values shown are mean percentages (plus or minus standard errors, n = 3) of fibroblasts collected from 10 cm culture dishes prior to and 24 hours and 48 hours after heat shock for a period of 30 minutes at 45°C.
% of fibroblasts recovered after heat treatment
Before heat shock Hours after heat shock
0 24 48
Control 100 97.6 ± 0.9 139.2 ± 1.5 175 ± 2.9
HGPS 100 95.8 ± 1.5 74.6 ± 5.1 104.9 ± 5.8
These experiments indicate that control cells recovered rapidly from heat shock, since no changes were apparent in their growth rate 24 and 48 hours later. However, fibroblasts from the subject with HGPS were hypersensitive to heat shock, and the number with dysmorphic nuclei was high after treatment. Their recovery was only observed 48 hours after heat shock, when cell numbers started to increase, suggesting that surviving cells had begun to divide again.
Discussion
We report another subject with HGPS who has the LMNA G608G mutation. So far, 21 out of 25 reported cases of HGPS genetically analyzed have this point mutation [4,5,27], making it a molecular signature for HGPS. Sequencing of parental DNA, when available, has shown that none of them carried this mutation, indicating that it is a de novo mutation. These findings also show that HGPS is an autosomal dominant disease possibly resulting from germinal mosaicism [4].
The LMNA G608G mutation creates an abnormal splice donor site producing mRNA with 150 nucleotides deleted from the prelamin A coding sequence. The encoded truncated protein is predicted to be missing residues 607 to 656 of prelamin A and to retain the CAAX box for the prenylation at its carboxyl-terminus. As the "upstream" endoproteolytic cleavage site of wild type prelamin A is deleted [15], the truncated prelamin A in HGPS may remain prenylated; however, this remains to be shown experimentally.
To understand the cellular basis of HGPS, we analyzed the morphology dermal fibroblasts from our subject, which express the mutant protein. We studied the distribution of A-type lamins in primary fibroblast cultures at early cellular passage numbers (less than or equal to 10) to better reflect the in vivo distribution of proteins. We observed nuclear abnormalities in a subpopulation of HGPS fibroblasts, where dysmorphic nuclei had "blebs" surrounded by A-type lamins. We did not observe defects in the lamin A/C meshwork in "blebs" of the most dysmorphic nuclei, and chromatin distribution appears normal even within the 'blebs." The fibroblasts did not have gross anomalies of the localization of lamin B1 or emerin, even when the lamina network appeared ruffled.
Only a subpopulation of cultured fibroblasts from our subject with HGPS had dysmorphic nuclei. Furthermore, the nuclear abnormalities observed in these fibroblasts may have been less dramatic than those reported in cell lines from other individuals with HGPS [28]. Nuclear abnormalities in fibroblasts from our subject with HGPS also differed from those reported in fibroblasts from subjects with Dunnigan-type familial partial lipodystrophy, a condition caused by different autosomal dominant LMNA mutations. Cells from these subjects have defects in lamin A/C distribution in nuclear envelope "blebs," and there is also a loss of lamin B1 staining around the nuclear envelope "blebs" and sometimes at the nuclear pores [29]. Despite these subtle differences between cells from subjects with HGPS and subjects with other laminopathies, generally similar morphological alterations invariably occur in some portion of cultured fibroblasts from subjects with different lamin A/C mutations [21,29,30]. Grossly similar alterations are also observed in fibroblasts from Lmna knockout mice [31,32]. Based on these nuclear envelope abnormalities common to all laminopathies, it appears that A-type lamins play a crucial role in the maintenance of the size and shape of the nucleus. These alterations in nuclear morphology could potentially lead to abnormalities in cell growth or structural stability.
We did not observe differences in the proliferation of fibroblasts from our subject with HGPS compared to fibroblasts from unaffected controls at passage numbers below 10 in primary culture. A recent report suggested that, after a certain number of passages, fibroblasts from subjects with HGPS were no longer able to proliferate at a similar rate as fibroblasts from unaffected controls [28]. Growth inhibition was recently reported for other cell lines from subjects with HGPS, where after a certain number of doublings, the cells rapidly entered a senescence phase [33,34]. Fibroblasts from Lmna knockout mice are prone to apoptosis when derived from newborns, while fibroblasts derived from knockout embryos can grow in culture for a restricted number of passages [30].
With regards to cell stability, we observed that fibroblasts from our subject with HGPS cells were more susceptible to damage from heat shock than controls. The subject's cells in culture had an increased number of dysmorphic nuclei and enhanced cell death within the first 24 hours after the heat shock. Fibroblasts from the subject with HGPS appeared to be able to eventually recover from heat stress, since they started to grow again after a delay of 24 hours. These findings suggest that fibroblasts form subjects with HGPS are fragile and more readily damaged or killed after stress. Increased nuclear fragility to mechanical stress and head shock has also been reported in fibroblasts from mice lacking lamins A and C and human subjects with Dunnigan-type familial partial lipodystrophy [29,35].
Conclusion
Cells from a subject with HGPS and the lamin A G608G mutation have dysmorphic nuclei and an increased susceptibility to damage by heat shock. Our findings suggest that these cells may also be sensitive to other types of stress, such as metabolic or mechanical. Improving the cellular response to stress may be one possible way of helping individuals with HGPS and other laminopathies.
Methods
Clinical material
After obtaining informed consent, we obtained blood from a subject with a clinical diagnosis of HGPS, her farther, and her mother. Genomic DNA was isolated from peripheral blood according to standard techniques [36]. A skin biopsy had been performed from the right leg when the subject was 2 years of age for histological examination, and a dermal fibroblast culture was concomitantly established from the same tissue sample. Two control dermal fibroblast cultures were established from neonatal foreskin and skin of a 9 year-old who underwent surgery without any known diseases. Primary dermal fibroblasts were maintained in DMEM containing 15 % fetal calf serum, 2 mM glutamine, 100 IU/ml penicillin, and 100 mg/ml streptomycin for a maximum of 12 passages. The Columbia University Medical Center Institutional Review Board approved the use of cells from human subjects.
LMNA sequencing
All exons of the lamin A/C coding region of LMNA and splice junctions were amplified by PCR from genomic DNA using primers described previously [26]. Amplified DNA was sequenced directly using an ABI Prism 310 Automated Sequencer and the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit (PE Applied Biosystems) following purification in a Centriflex Gel Filtration Cartridges (Edge Biosystems). Mutations were identified by visual inspection and comparison with sequences generated from unrelated, unaffected individuals, and the GeneBank reference sequence.
Prelamin A cDNA sequencing
Total RNA was extracted from cultured dermal fibroblasts using Trizol reagent according to the manufacturer (Invitrogen). Using the forward primer (5' GGCTGCGGGAACAGC 3') and the reverse primer (5' CTGGCAGGTCCC 3'), we amplified by RT-PCR the cDNA spanning nucleotides 1561 to 2010 of the prelamin A coding sequence. Amplified DNA products were purified and subcloned into the pGBT9 (Invitrogen). Clones were sequenced using the forward and reverse primers described above. The mutant sequence was verified by sequencing 3 independent clones generated from 3 different RT-PCRs.
Indirect immunofluorescence microscopy
Primary cultures of dermal fibroblasts were grown on glass coverslips, washed with PBS, fixed in methanol at -20°C, and then processed for indirect immunofluorescence as described previously [37,38]. Rabbit antibodies directed against A-type and B-type lamins were kindly provided by Dr. N. Chaudhary and have been previously described [39,40]. Secondary antibodies were affinity purified Alexa Fluor 488 goat anti-rabbit or anti-mouse IgG antibodies (Molecular probes) and Cy3-conjugated goat anti-rabbit or anti-mouse (Jackson ImmunoResearch Laboratories). For double immunofluorescence microscopy with the rabbit anti-A-type lamin antibodies, we used a mouse monoclonal anti-emerin clone 4G5 (Novacastra Laboratories) or a previously described human anti-lamin B antiserum [41], which was kindly provided by Dr. J.-C. Courvalin. Cells were also examined with anti-vimentin antibodies [42], kindly provided by Dr. S. D. Georgatos, and appropriate secondary antibodies. All samples were also counterstained with DAPI (Sigma-Aldrich) and slides were examined using a confocal microscope.
Transmission Electron Microscopy
Cells were fixed with 2.5% glutaraldehyde in 0.1 M Sorenson's buffer (PH 7.2) for an hour, and postfixed with 1% OsO4 in Sorenson's buffer for one hour. Enblock staining was performed using 1% tannic acid. After dehydration cells were embedded in a mixture of Lx-112 (Ladd Research Industries, Inc.) and Embed-812 (EMS, Fortwashington, PA). Thin sections were cut on a MT-7000 ultramicrotome. Sections were stained with uranyl acetate and lead citrate, and examined under a JEOL JEM-1200 EXII electron microscope.
Western blot analysis
Total cellular protein extracts were isolated from dermal fibroblasts in primary cultures at passage number 10. Total cell number was determined for each culture, and cell pellets were extracted directly in Laemmli sample buffer [43]. Cells were lysed and equal amounts of extracts (corresponding to the same number of cells) were loaded in parallel on a 6% polyacrylamide gel. After separation by electrophoresis, proteins were transferred to nitrocellulose membranes and incubated with blocking buffer for several hours as described previously [44]. Membranes were incubated with primary antibodies, washed, and then incubated with the corresponding secondary antibody coupled to horseradish peroxidase (Jackson ImmunoResearch Laboratories). Proteins were visualized using the enhanced chemiluminescence's detection system (Amersham Pharmacia Biotech). Signals obtained on the autoradiograms were analyzed by densitometry using Quantity One 1-D analysis software (BioRad Laboratories) on the scanned images.
Heat shock treatment
Cells were grown in 10 cm diameter dishes in complete medium and transferred from 37°C to 45°C for 30 minutes. Viability was determined prior to heat shock, and at time 0 (after 30 minutes stress), 24 hours, and 48 hours after recovery at 37°C. Attached cells were trypsinized at each time point, collected and counted. For morphological analysis, fibroblasts were grown on coverslips and treated as above. Unheated and heated samples were fixed in methanol at -20°C, and processed for indirect immunofluorescence as described above.
Authors' contributions
MP identified the new subject with HGPS and provided the clinical data and materials. CP established the HGPS dermal fibroblast primary culture. DM performed the heat shock experiments, and immunohistochemistry experiments. RLB participated in the subcloning and sequencing of the mutant lamin A cDNA. HJW was involved in the subcloning, sequencing of the mutant lamin A cDNA and in manuscript preparation. KD performed the screening of the LMNA gene and was responsible for designing and carrying out this work and for preparing the manuscript.
Acknowledgements
We thank Dr. Angela M. Christiano (Columbia University) for referring the child with HGPS to us and for discussions and advice. We also thank the child's family for allowing us to use the cells in this study. We thank Dr. Amalia Martinez-Mir for stimulating discussions and Naira Souleymanian for technical assistance. We are grateful to Dr. N. Chaudhary, Dr. J.-C. Courvalin and Dr. S. D. Georgatos for generous gifts of antibodies. This work was supported by Progeria Research Foundation, NIH Grants R03AR 47641 and KO1AR048594 (to K.D.) and in part by NIH Grant AR048997 (to H. J. W.).
==== Refs
DeBusk FL The Hutchinson-Gilford progeria syndrome Journal of Pediatric 1972 90 697 724
Brown WT Kieras FJ Houck GE Dutkovski R Jenkins A A comparison of adult and childhood progerias: Werner syndrome and Hutchinson Gilford syndrome Advance Experimental Medical Biology 1992 190 229 244
Shinton RA Sarkar PK Hutchinson-Gilford progeria syndrome Postgrad Medical Journal 2001 77 312 317 10.1136/pmj.77.907.312
Eriksson M Brown WT Gordon LB Glynn MW Singer J Scott L Erdos MR Robbins CM Moses TY Berglund P Dutra A Pak E Durkin S Csoka AB Boehnke M Glover TW Collins FS Recurrent de novo point mutations in lamin A cause Hutchinson-Gilford progeria syndrome Nature (London) 2003 423 293 298 12714972 10.1038/nature01629
De Sandre-Giovannoli A Bernard R Cau P Navarro C Amiel J Boccaccio I Lyonnet S Stewart CL Munnich A Le Merrer M Levy N Lamin A truncation in Hutchinson-Gilford progeria Science 2003 300 2055 12702809 10.1126/science.1084125
Cao H Hegele RA LMNA is mutated in Hutchinson-Gilford progeria (MIM 176670) but not in Wiedemann-Rautenstrauch progeria syndrome (MIM 264090) Journal of Human Genetic 2003 48 271 274 10.1007/s10038-003-0025-3
Mounkes LC Kozlov S Burke B Stewart CL The laminopathies: nuclear structure meets disease Current Opinion in Cell Biology 2003 13 223 2230
Worman HJ Courvalin JC How does mutations in lamin A and C cause disease? Journal of Clinical Investigation 2004 113 349 351 14755330 10.1172/JCI200420832
Fuchs E Weber K Intermediate filaments: structure, dynamics, function, disease Annual Review of Biochemistry 1994 63 345 382 7979242
Steinert PM Steven AC Roop DR The molecular biology of intermediate filaments Cell 1985 42 411 419 2411418 10.1016/0092-8674(85)90098-4
Aebi U Cohn JB Gerace LL The nuclear lamina is a meshwork of intermediate-type filaments. Nature (London) 1986 323 560 564 3762708 10.1038/323560a0
Fisher D Chaudhary N Blobel G cDNA sequencing of nuclear lamins A and C reveals primary and secondary structure homology to intermediate filaments proteins Proceedings of the National Academy of Sciences 1986 83 6450 6454
Riemer D Karabinos A Weber K Analysis of eight cDNAs and six genes for intermediate filament protein in the cephalochordate Brachiostoma reveals differences in the multigene families of lower chordates and vertebrates Gene 1998 211 361 373 9602172 10.1016/S0378-1119(98)00151-6
Frangioni J Neel B Use of a general purpose mammalian expression vector for studying intracellular protein targeting: identification of critical residues in the nuclear lamin A/C nuclear localization sequence Journal of Cell Science 1993 105 481 488 8408279
Sinensky M McLain T. Fantle K Expression of prelamin A but not mature lamin A confers sensitivity of DNA biosynthesis to Iovastatin on F9 teratocarcinoma cells Journal of Cell Science 1994 107 2215 2218 7983180
Holt I Clements L Manilal S Brown SC Morris GE The R428Q lamin A/C mutation that causes lipodistrophy does not prevent nuclear targeting of lamin A in adipocytes or its interaction with emerin European Journal of Human Genetic 2001 9 204 208 10.1038/sj.ejhg.5200609
Krohne G Waizenegger I Hoger TH The conserved carboxy-terminal cysteine of nuclear lamins is essential for lamin association with the nuclear envelope Journal of Cell Biology 1989 109 2003 2011 2808518 10.1083/jcb.109.5.2003
Vorburger K Kitten GT Nigg EA Modification of nuclear lamin proteins by a mevalonic acid derivative occurs in reticulocyte lysates and requires the cysteine residue of the C-terminal CXXM motif EMBO Journal 1989 8 4007 4013 2686979
Lin F Worman HJ Structural organization of the human gene encoding nuclear lamin A and nuclear lamin C Journal of Biological Chemistry 1993 268 16321 16326 8344919
Wydner KL McNeil JA Lin F Worman HJ Lawrence JB Chromosomal assignment of human nuclear envelope protein genes LMNA, LMNB1, and LBR by fluorescence in situ hybridization Genomics 1996 32 474 478 8838815 10.1006/geno.1996.0146
Worman HJ Courvalin JC The nuclear lamina and inherited disease Trends in Cell Biology 2002 12 591 597 12495848 10.1016/S0962-8924(02)02401-7
Lin F Worman HJ Structural organization of the human gene (LMNB1) encoding nuclear lamin B1 Genomics 1995 27 230 236 7557986 10.1006/geno.1995.1036
Biamonti G Giacca M Perini G Contreas G Zentilin L Weighardt F Guerra M Della Valle G Saccone S Riva S A. F The gene for a novel human lamin maps at a highly transcribed locus of chromosome 19 which replicates at the onset of S-phase Molecular and Cellular Biology 1992 12 3499 3506 1630457
Rober RA Weber K Osborn M Differential timing of nuclear lamin A/C expression in the various organs of the mouse embryo and the young animal: a developmental study Development 1989 105 365 378 2680424
Furukawa K Hotta Y cDNA cloning of a germ cell specific lamin B3 from mouse spermatocytes and analysis of its function by ectopic expression in somatic cells EMBO Journal 1993 12 97 106 8094052
De Sandre-Giovannoli A Chaouch M Kozlov S Vallat JM Tazir M Kassouri N Szepetowski P Hammadouche T Vandenberghe A Stewart CL Grid D Levy J Homozygous defects in LMNA, encoding lamin A/C nuclear-envelope proteins, cause autosomal recessive axonal neuropathy in human (Charcot-Marie-Tooth disorder type 2) and mouse American Journal Of Human Genetic 2002 70 726 736 10.1086/339274
Fukuchi K Katsuya T Sugimoto K Kuremura M Kim HD Li L Ogihara T LMNA mutation in a 45 year old Japanese subject with Hutchinson-Gilford progeria syndrome Journal of Medical Genetic 2004 41 e67 10.1136/jmg.2003.014688
Goldman RD Shumaker DK Erdos MR Eriksson M Goldman AE Gordon LB Gruenbaum Y Khuon S Mendez M Varga R Collins FS Accumulation of mutant lamin A causes progressive changes in nuclear architecture in Hutchinson-Gilford progeria syndrome Proceedings of the National Academy of Sciences 2004 101 8963 8968 10.1073/pnas.0402943101
Vigouroux C Auclair M Dubosclard E Pouchelet M Capeau J Courvalin JC Buendia B Nuclear envelope disorganization in fibroblasts from lipodystrophic patients with heterozygous R482Q/W mutations in the lamin A/C gene Journal of Cell Science 2001 114 4459 4468 11792811
Mounkes LC Stewart CL Aging and nuclear organization: lamins and progeria Current Opinion in Cell Biology 2004 16 322 327 15145358 10.1016/j.ceb.2004.03.009
Sullivan T Escalante-Alcalde D Bhatt H Anver M Bhat N Nagashima K Stewart CL Burke B Loss of A-type lamin expression compromises nuclear envelope integrity leading to muscular dystrophy Journal of Cell Biology 1999 29 913 920 10579712 10.1083/jcb.147.5.913
Burke B Stewart CL Life at the edge: the nuclear envelope and human disease Nature Reviews Molecular cell Biology 2002 3 575 585 12154369 10.1038/nrm879
Bridger JM Kill IR Aging of Hutchinson-Gilford progeria syndrome fibroblasts is characterized by hyperproliferation and increase apoptosis Experimental Gerontology 2004 39 717 724 15130666 10.1016/j.exger.2004.02.002
Wallis CV Sheerin AN Green MHL Jones CJ Kipling D Faragher RGA Fibroblast clones from patients with Hutchinson-Gilford progeria can senesce despite the presence of telomerase Experimental Gerontology 2004 39 461 467 15050279 10.1016/j.exger.2003.12.015
Lammerding J Schulze PC Takahashi T Kozlov S Sullivan AJ Kamm RD Stewart CL Lee CP Lamin A/C deficiency causes defective nuclear mechanics and mechanotransduction Journal of Clinical Investigation 2004 113 370 378 14755334 10.1172/JCI200419670
Sambrook J Fritsch EF Maniatis T Laboratory CSH Molecular cloning: a laboratory manual 1989 2nd Ed., New York, Cold Spring Harbor Laboratory
Djabali K de Néchaud B Landon F Portier MM aB-crystallin interacts with intermediate filaments in response to stress Journal of Cell Science 1997 110 2759 2769 9427392
Djabali K Aita VM Christiano AM Hairless is translocated to the nucleus via a novel bipartite nuclear localization signal and is associated with the nuclear matrix Journal of Cell Science 2001 114 367 376 11148138
Chaudhary N Courvalin JC Stepwise reassembly of the nuclear envelope at the end of mitosis Journal of Cell Biology 1993 122 295 306 8391536 10.1083/jcb.122.2.295
Cance WG Chaudhary N Worman HJ Blobel G Cordon-Cardo C Expression of the nuclear lamins in normal and neoplastic human tissues Journal of Experimental Clinical and Cancer Research 1992 11 233 246
Lassoued K Andre C Danon F Modogliani R Dhumeaux D Clauvel JP Brouet JC Courvalin JC Characterization of two human monoclonal IgM antibodies that recognize nuclear lamins European Journal of Immunology 1992 22 1547 1551 1601041
Maison C Pyrpasopoulou A Georgatos SD Vimentin-associated mitotic vesicles interact with chromosomes in a lamin B- and phosphorylation-dependent manner EMBO Journal 1995 14 3311 3324 7628433
Laemmli UK Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature (London) 1970 227 680 685 5432063 10.1038/227680a0
Djabali K Portier MM Gros F Blobel G Georgatos SD Network antibodies identify nuclear lamin B as a physiological attachment site for peripherin intermediate filaments Cell 1991 64 109 121 1986862 10.1016/0092-8674(91)90213-I
|
15982412
|
PMC1183198
|
CC BY
|
2021-01-04 16:31:30
|
no
|
BMC Cell Biol. 2005 Jun 27; 6:27
|
utf-8
|
BMC Cell Biol
| 2,005 |
10.1186/1471-2121-6-27
|
oa_comm
|
==== Front
BMC Dev BiolBMC Developmental Biology1471-213XBioMed Central London 1471-213X-5-121598517510.1186/1471-213X-5-12Research ArticleMorphogenesis of the anterior segment in the zebrafish eye Soules Kelly A [email protected] Brian A [email protected] Department of Cell Biology, Neurobiology and Anatomy, Medical College of Wisconsin, Milwaukee, Wisconsin, USA2005 28 6 2005 5 12 12 27 4 2005 28 6 2005 Copyright © 2005 Soules and Link; licensee BioMed Central Ltd.2005Soules and Link; 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 ocular anterior segment is critical for focusing incoming light onto the neural retina and for regulating intraocular pressure. It is comprised of the cornea, lens, iris, ciliary body, and highly specialized tissue at the iridocorneal angle. During development, cells from diverse embryonic lineages interact to form the anterior segment. Abnormal migration, proliferation, differentiation, or survival of these cells contribute to diseases of the anterior segment such as corneal dystrophy, lens cataract, and glaucoma. Zebrafish represent a powerful model organism for investigating the genetics and cell biology of development and disease. To lay the foundation for genetic studies of anterior segment development, we have described the morphogenesis of this structure in zebrafish.
Results
As in other vertebrates, the zebrafish anterior segment derives from diverse origins including surface ectoderm, periocular mesenchyme, and neuroepithelium. Similarly, the relative timing of tissue differentiation in the anterior segment is also conserved with other vertebrates. However, several morphogenic features of the zebrafish anterior segment differ with those of higher vertebrates. These include lens delamination as opposed to invagination, lack of iris muscles and ciliary folds, and altered organization in the iridocorneal angle. In addition, substantial dorsal-ventral differences exist within the zebrafish anterior segment.
Conclusion
Cumulatively, our anatomical findings provide a reference point to utilize zebrafish for genetic studies into the mechanisms of development and maintenance of the anterior segment.
==== Body
Background
The anterior segment of the vertebrate eye is comprised of the cornea, lens, iris, ciliary body, and highly specialized tissue at the iridocorneal angle. Two main functions are ascribed to the ocular anterior segment. The first is to focus incoming light onto the neural retina and the second is to regulate intraocular pressure. For mammals and other higher vertebrates, refraction of light entering the eye is accomplished by both the transparent cornea and lens. In many aquatic vertebrates, including fish, the lens is solely responsible for focusing incoming light [1,2]. In all vertebrates, intraocular pressure is maintained by the balance between aqueous humor production and outflow [3]. The dynamics of aqueous humor have been best characterized in mammals where ciliary epithelial cells produce the clear ocular fluid while the trabecular meshwork, which is situated at the iridocorneal angle overlying Schlemm's canal, regulates drainage.
The structures of the anterior segment arise from diverse embryonic lineages and there is exquisite coordination among the different compartments during development. Studies in avian and mammalian species have shown that tissues of the anterior segment derive from surface ectoderm, head mesoderm, neural crest and neuroectoderm [4-7]. Development of the anterior segment initiates with the invagination of the lens from surface ectoderm. With establishment of the lens vesicle, head mesoderm and neural crest cells migrate into a periocular location and eventually move into the anterior segment of the rudimentary eye between the surface ectoderm and the neural retina and lens. These mesenchymal cells differentiate into the corneal endoderm, structures at the iridocorneal angle, and iris and ciliary body stroma. The non-pigmented and pigmented epithelium of the iris and ciliary epithelium derive from the peripheral edge of the retinal neuroepithelium and retinal pigmented epithelium, respectively. Developmental anatomy of the anterior segment for many higher vertebrates has been well characterized and excellent reviews exist [8,9]. However, the relevant cellular interactions between various structures of the anterior segment and the molecular basis of development is just beginning to be understood.
A detailed understanding of the mechanisms of development of the anterior segment can provide general insights into questions such as tissue induction, cell type fate determination, and the regulation of cellular morphogenesis. In addition, an understanding of ontogeny of the anterior segment has significance to several human diseases. Primarily, several forms of glaucoma are associated with anterior segment dysgenesis and genes which are essential for formation of this part of the eye can promote glaucoma [9,10]. Corneal dystrophies and lens cataracts are additional examples of anterior segment disease. The zebrafish has many experimental advantages for studying both development and disease phenotypes, including those for glaucoma [11]. These include the ability to conduct genetic screens for complex traits owing to high fecundity and genomic infrastructure. In addition, ease of transgenesis for target gene functional analysis, coupled with rapid and transparent initial development, facilitates cell behavior characterization via time-lapse microscopy. However, an overview of the development for the zebrafish anterior segment has not been described and the extrapolation of ocular anatomy from other teleost species is not favorable due the high morphological diversification of the bony fish [1]. In this study we report the characterization by light and transmission electron microscopy (TEM) the morphogenesis of the zebrafish anterior segment. We find that while there is rapid initial establishment of the anterior segment, occurring within the first three days of embryogenesis, significant growth and morphogenesis continues until approximately 1 month when the mature morphology is attained. There are also significant differences in the elaboration of dorsal versus ventral regions within the zebrafish anterior segment. Importantly, both similarities and differences exist between the anatomy of the zebrafish ocular anterior segment and that of mammalian eyes.
Results
Establishment of the anterior chamber
The zebrafish anterior segment is established very rapidly and rudimentary structures of the anterior segment are present by 3 days post fertilization (dpf) when visually evoked behaviors are first observed [12]. The anterior chamber forms with the detachment of the lens vesicle from the surface ectoderm which occurs by 26 hours post fertilization (hpf). Beginning at approximately 24 hpf, mesenchymal cells migrate from periocular locations into the anterior segment (Figure 1; [12,13]). Within the anterior segment, a single layer of flattened mesenchymal cells associated with the posterior portion of the cornea can be seen by 36 hpf (Figure 1B). At this time within the peripheral angles of the anterior chamber, undifferentiated mesenchymal cells accumulate. Hyaloid vasculature is associated with the posterior region of the lens and ciliary vessels, which circumvent the anterior rim of the eye, and enter and exit at the embryonic fissure (Figure 1A, B). By 48 hpf, differentiation among mesenchymal cells can be detected within the angle region where the cornea and prospective iris meet (Figure 1C). By 3 dpf, multiple types of pigment cells, as well as less differentiated non-pigmented cells, are present at the angle (Figure 2A, B). Tissue extension from the margins of the neural retina has established the iris anlagen. The ellipsoid shape of the eye and large spherical lens make the newly formed anterior chamber relatively shallow at the center as compared to the periphery. While the rudimentary anterior segment is formed by day 3, extensive growth and morphogenesis occurs until 1 month when the eye reaches its mature form (Figure 2). Specifically there is extensive stratification within the cornea, differentiation and morphogenesis of the iris stroma and ciliary epithelial zones, specialization of angle structures, and elaboration of dorsal versus ventral differences. Interestingly, following initial establishment of the anterior segment, growth and morphogenesis appear to be coupled and independent of the age of the fish. For example, 17 dpf sibling zebrafish reared under the same conditions in the same tank can have measurable body length differences. Both eye size and anterior segment differentiation correlate with body length, but not the absolute age of the fish (data not show). For this reason, body length measurements were recorded for each specimen examined. Following 1 month, continued eye growth and subtle refinement of the zebrafish anterior segment persists past sexual maturity (~ 3 months).
Figure 1 Establishment of the anterior chamber. Histology of a 24 hpf (A), 36 hpf (B), and 48 hpf (C) zebrafish eye. Periocular mesenchyme is present at 24 hpf, but more prominent at 36 and 48 hpf (arrows). Hyaloid vasculature is indicated with asterisk.
Figure 2 Comparison of embryonic and adult zebrafish eyes. Diagram of embryonic (A) and adult (C) zebrafish eyes. Histology of 3 dpf embryonic (B) and 1 month adult (D) eyes.
The cornea
In all vertebrates studied, the cornea develops from both the surface ectoderm and periocular mesenchyme. Following detachment of the lens vesicle from the surface ectoderm, mesenchymal cells migrate into the anterior chamber along the surface ectoderm. The surface ectoderm and periocular mesenchyme differentiate into the corneal epithelium and endothelium, respectively. In mammals and avians, the corneal stroma is then formed from an additional wave of immigrating periocular mesenchymal cells [14]. The corneal stroma further stratifies through cellular differentiation and the deposition of highly ordered collagen fibrils and other extracellular matrix components. Histological observations support a similar mode of corneal establishment in zebrafish (Figure 1). Up to 60 hpf, the zebrafish cornea consists of relatively undifferentiated, multi-stratified cells, although distinct layers are not obvious. Differentiation within the cornea, however, is readily observable by 3 dpf as stratification becomes apparent (Figure 3). At this time, the surface epithelium, which is still continuous with the skin of the embryo, is composed of hexagonally shaped cells that appear scalloped is cross-section. Electron microscopy revealed the presence of numerous electron dense adherens junctions between surface epithelial cells, but few such cell junctions were found between the 2–3 layers of subepithelial cells (Figure 3C, E). During development, the corneal surface epithelium can also be distinguished ultrastructurally from subepithelial cells by the type of inclusion bodies. Surface epithelium shows numerous dark-staining small inclusion bodies, while subepithelia contain less numerous, but larger inclusion bodies (Figure 3C). A thin lamellar stroma with orthogonally arrayed collagen fibrils is visible by 3 days (Figure 3D). Within the periphery, the corneal stroma is continuous with the sclera. In the 3–5 dpf embryo, the thin collagen stroma directly overlays flattened endothelial cells (Figure 3C).
Figure 3 Cornea morphogenesis. Diagram of embryonic (A) and adult (F) zebrafish corneas. Morphology of embryonic (B-E) and adult corneas (G-K). Histology of cornea and anterior lens of 3 dpf embryo (B). TEM micrograph showing embryonic cornea epithelium, CE; cornea stroma, CS; and cornea endothelium, CN (C). In this panel, the lens epithelium (Lens EP) is indicated by blue shading. High magnification insets of developing cornea stroma (D) and "scalloped" cornea epithelium of 3 dpf embryo (E). Histology of cornea and anterior lens of 1 month adult (G). TEM micrograph showing 1 month cornea epithelium, cornea stroma, and cornea endothelium (H). Extensive interdigitation between epithelial cells is indicated with arrows (H). High magnification insets of the mature cornea epithelial cells (I), cornea sub-epithelial cells (J), and cornea stroma and endothelium (K). Note the increase of electron dense cell junctions in cornea sub-epithelial cells (J) and the establishment of a flattened layer of cells, indicated with asterisk, between the stroma and endothelium (K). In both the developing and mature cornea stroma, orthogonally arranged collagen fibrils are clearly visible and the stromal thickness has increased with development (D and K).
By one month, the cornea has further stratified (Figure 3G, H). The epithelium has lost its scalloped appearance and is no longer continuous with the outer epidermis. As compared to mammals, the corneal epithelium is relatively thick containing 3–4 layers of interdigitated cells (arrows, Figure 3H). An increase in the number of adherens juctions was noted between surface epithelial cells and subepithelial cells (Figure 3I, J). A thin extracellular deposition (Bowman's membrane) separates the subepithelial cells from the the corneal stroma layer. The stroma, although thin compared to mammals, has increased in thickness with development and maintains the orthogonal array of collagen fibrils (Figure 3K). Sparsely distributed, flattened cells can be seen upon electron microscopy within the stroma, particularly at the peripheral edges. The number of these keratocytes, as well as the thickness of the stroma, increases with age past 1 month. At the peripheral limbal region, the stroma splits into multiple layers. Just beneath the stroma is a single layer of cells overlying a second, thinner stromal layer with similar collagen organization (Figure 3K, [15]). Subjacent to these cells, an additional collagen-rich extracellular layer can be seen with electron microscopy (Descemet's membrane). The endothelium is also thin compared to mammals and is comprised of a single layer of flattened cells which extend over the iridocorneal angle covering the surface of the annular ligament. Mucus secreting goblet cells were frequently observed at the peripheral edges of the cornea (data not shown).
The lens
The zebrafish lens is relatively large and spherical, unlike mammalian and avian lenses which are more ellipsoid in shape. The zebrafish lens shows typical vertebrate composition with two cell types-lens epithelial cells and lens fiber cells. The establishment of the lens begins with the contact between the evaginating optic vesicle and surface ectoderm at 16 hpf (14–15 somite stage). This interaction results in a visible thickening of the lens placode [13]. Unlike the mammal, the lens of the zebrafish forms by delamination of lens placodal cells and not through invagination. This results in a solid spherical mass as opposed to a hollow lens vesicle (Figure 1A; [12,13]). Detachment of the solid lens vesicle of zebrafish at 24–26 hpf is accomplished in part by apoptosis, similar to mammals [16]. By 30 hpf, lens epithelial cells are discernible. Primary fiber cell elongation occurs in a circular fashion within the center of the newly formed lens. A transition zone is present at 36 hpf and is located more posteriorly as compared to mammals. This region of the lens is common among vertebrates and is where secondary fibers cells differentiate from proliferative epithelial cells. Loss of organelles within lens fiber cells is apparent by 72 hpf. We have focused our analysis on three diverse regions of the lens: the anterior epithelium, the transition zone, and the posterior lens (Figure 4). In the developing lens, anterior epithelial cells are cuboidal with flat plasma membranes (Figure 4E). Some interdigitations are present laterally between these cells. Large, irregularly shaped nuclei reside in the center of the lens epithelial cytoplasm. A thin extracellular capsule is present by 72 hpf. At the transition zone, epithelial nuclei elongate with the lengthening and differentiation into fiber cells (Figure 4B). The more centrally located, differentiated fiber cells show dark staining intracellular spheres, perhaps related to the disassembly of organelles. These dark staining bodies are also observed within the posterior lens (Figure 4D). Newly generated fiber cells show extensive interdigitations, but have yet to achieve either "ball and socket" or suture organization (Figure 4G, F) At this stage of development, very little vitreous exists and the lens is often observed in direct contact with ganglion cells of the neural retina as well as vascular hyaloid cells (Figure 4H).
Figure 4 Lens morphogenesis. Diagram of the embryonic (A) and adult (I) zebrafish lens. Insets in (A) indicate corresponding histological sections through the 3 dpf lens transition zone (B), lens epithelium (C) and posterior lens (D). TEM micrographs (E-H) show developing lens epithelium (E), interdigitating lens fiber cells (arrows, G), establishment of rudimentary lens sutures (arrow, F), and a vascular endothelial cell, EN, associated with the posterior lens, PL (H). Insets in (I) indicate corresponding histological sections through the 1 month lens transition zone (J), lens epithelium (K) and posterior lens (L). TEM micrographs (M-P) show the mature lens epithelium (M), "ball and socket" interdigitations between fiber cells (arrows, O), loose fiber cell organization at the periphery (N), and mature suture connections between the ends of lens fiber cells (arrows, P).
The lens of the one month old zebrafish retains spherical symmetry. The anterior epithelial cells and their nuclei appear more flattened than at younger ages (Figure 4K and 4M). The capsule is now more thick and prominent. The anterior epithelium covers lens fiber cells which appear hexagonal in cross-section. The packing density of the fiber cells increase in a peripheral to central manner. At the peripheral edges, fiber cells frequently show separations following histological processing due to the lower packing density (Figure 4M, N). Centrally located fiber cells now show "ball and socket" intracellular connections (Figure 4O). At the posterior lens, fiber cell ends meet to form sutures (Figure 4P). At one month, the lens is suspended by zonules attaching the lens capsule at the equator to the non-pigmented epithelium of the ciliary zone (Figure 5). The zonules appear as modified or "toughened" vitreous, similar to those described for other lower vertebrates [1]. In zebrafish, dorsal zonules are thickened as compared to ventral zonules. In both dorsal and ventral regions, these fibers delimit the vitreous body-aqueous humor boundary.
Figure 5 Lens zonules. Diagram for location of mature lens zonules (A). Histological sections through the dorsal (B) and ventral (C) anterior segment. Lens zonules are indicated by black arrows. Large ciliary circumferential arteries are indicated with white arrowheads. Iris stroma vasculature in indicated with asterisks.
The iris and ciliary zone
In zebrafish, the iris, ciliary epithelium and anterior segment angle specializations differentiate relatively late and in a more protracted fashion as compared to the cornea and lens. The iris and ciliary zone develop from the anterior margin of the retinal neuroepithelium as in other vertebrates. Iris stroma is derived from periocular mesenchyme. Morphological differentiation within the iris stroma is apparent by 3 dpf (Figure 6). Ordered layering of pigment cells within this iris stroma is visible at this early time (Figure 6C, F). The pigment epithelium contains dense, black melanosomes. At 3 dpf, the non-pigmented epithelial cell layer (ciliary or iris) is absent, as retinal margin neuroepithelial cells have yet to differentiate anteriorly. Iridophores, containing rod-shaped iridosome organelles are found directly overlying the pigment epithelial layer. Adjacent and superficial to iridophores are xanthopohores. These pigment cells contain foamy appearing pteranosome organelles. In whole view, iridophores appear silver in color, while xanthophores are gold. Interspersed within the developing iris stroma are non-pigmented, undifferentiated mesenchymal cells, and developing vasculature. In the ventral iris, differentiation of pigment cells appears delayed compared to the dorsal iris as both melanosomes and pteranosomes are reduced in number (Figure 6B–G).
Figure 6 Morphogenesis of the iris and ciliary zone. Diagram showing location of insets for the developing (A) and mature (H) iris and ciliary zone. Histology of the 3 dpf developing dorsal (B) and ventral (E) iris. TEM micrographs from 3 dpf show pigment cell stratification in dorsal (C) and ventral (F) regions. Note the reduced differentiation as indicated by pigmentation within the ventral region. Corresponding drawings indicate the principle cell types found within the developing iris (D, G). Color shadings indicate cells of similar embryonic origins. Histology of the 1 month old zebrafish dorsal (I) and ventral (L) iris. TEM micrographs show differences in the ultrastructure of non-pigmented ciliary epithelial cells between dorsal (J) and ventral (M) regions. Insets show the membranous infolding and secretory appearance within the dorsal ciliary zone (K) and the lack of this morphology in the ventral non-pigmented ciliary epithelium (N). RBC, red blood cell; UN, undifferentiated cell; Xa, xanthophore; Me, melanocyte; Ir, iridophore; CE, corneal epithelium; CN, corneal endothelium; NR, neural retina.
Inspection of the iris, ciliary zone, and anterior segment angle in the mature zebrafish highlights the dramatic differentiation that has occurred from embryonic stages (Figure 6H–N). To investigate the differentiation process within the ocular angle, we analyzed histological sections from dorsal and ventral regions at various timepoints between embryonic and adult stages (Figure 7 and 8). During this protracted development, the iris has grown extensively and a defined non-pigment epithelium, continuous with the retinal neuroepithelium, is present by 20 dpf. The zebrafish iris stroma is not contractile and lacks muscle cells. In addition, zebrafish do not show ciliary processes and instead we refer to the region underlying the base of the iris as the ciliary zone. The iris stoma remains stratified throughout development, with the same ordering of pigment cells as in the embryo. However, in addition to the posterior iridophores and superficial xanthophores, melanophores are interspersed with non-pigmented cells throughout the stroma. These cells appear at approximately 5–7 dpf. In the adult, the iridophore layer can be further sub-divided with superficial cells having iridosomes arranged perpendicular to iridosomes from cells closer to the pigmented epithelium (Figure 7F, 8F). The iris stroma also contains non-uniform clusters of vascular vessels that are more dense at the base where the iris is thicker. Similar to mammals, the superficial surface of the iris stroma is devoid of an epithelial covering.
Figure 7 Morphogenesis of the dorsal iridocorneal angle. Histological sections through the dorsal iridocorneal angle region of 3 dpf (A), 5 dpf (B), 10 dpf (C), 17 dpf (D), 25 dpf (E) and 1 month (F) zebrafish. Body length is indicated in parenthesis (upper left of each image) as differentiation in the iridocorneal angle is more dependant on the size of the fish than the absolute age.
Figure 8 Morphogenesis of the ventral iridocorneal angle. Histological sections through the ventral iridocorneal angle region of 3 dpf (A), 5 dpf (B), 10 dpf (C), 17 dpf (D), 25 dpf (E) and 1 month (F) zebrafish. Body length is indicated in parenthesis (upper left of each image) as differentiation in the iridocorneal angle is more dependant on the size of the fish than the absolute age.
Figure 9 Morphology of the annular ligament. Diagram of an adult zebrafish eye shows the location of insets for morphological analysis of the annular ligament (A). Histological section through the dorsal iridocorneal angle with extensive annular ligament (B). TEM micrograph of annular ligament cells (C). Glycoprotein aggregates, Gly, nuclei, N, and cell cytoplasm, Cyt, are indicated. Arrow points to cytoplasmic inclusion, In, within the glycoprotein aggregates.
The ciliary zone provides a site for lens zonule attachment, and in mammals and avians, this is also the site of aqueous humor production. In zebrafish, electron microscopy revealed differences in ciliary zone non-pigmented epithelial cells between the dorsal versus ventral regions. Non-pigmented epithelium in the dorsal ciliary zone is thicker than that in the ventral region (Figure 6J, K, M, N). Dorsal non-pigented ciliary epithelium is highly convoluted and secretory in appearance (Figure 6J, K). These cells are enriched in cytoplasmic extensions, golgi apparatus, and intracellular vesicles. The dorsal ciliary epithelium is ultrastructurally very similar to the non-pigmented epithelium lining ciliary processes of mammals [17]. In the ventral region, the non-pigmented epithelium is also secretory in appearance, but contains less cytoplasmic extensions and intracellular vesicles (Figure 6M, N). In the extreme ventral region, where there are canalicular specializations at the angle, the non-pigmented epithelium is absent (Figure 10, 11).
Figure 10 Serial section analysis through the ventral angle. Diagram of an embryonic (A) and mature (G) zebrafish eye shows the location of serial histological sections. Serial histological sections through a ~10 micron region of the 5 dpf embryonic ventral angle (B-F) and a ~60 micron region of the 1 month mature ventral angle. Note the elaboration of a canalicular sinus network. Arrows indicate the central region through the ciliary canal (D, J).
Figure 11 Ultrastructure of ventral iridocorneal angle specializations. Diagram of a mature (A) zebrafish eye shows the location of histological inset. Histology of ventral iridocorneal angle specializations in a 1 month old zebrafish (B). Low magnification TEM micrograph of endothelium lining annular ligament (C). High magnification TEM micrograph of smooth endothelium lining the central annular ligament (D). High magnification TEM micrograph of involuted and absorptive-appearing cells lining the iridocorneal angle (E). Low magnification TEM micrograph of mesothelium lining the iridocorneal canal (F). High magnification TEM micrograph of involuted and absorptive-appearing mesothelium lining the iridocorneal canal (G). High magnification TEM micrograph showing electron dense, thick filaments within myoepithelial cells of the ciliary "thumb" projection (arrows, H).
Anterior chamber angle and annular ligament
A striking difference between the embryonic angle and that of the adult zebrafish, is the presence of the annular ligament. The annular ligament is a prominent, but highly variable feature between species of bony fish [1,18]. The annular ligament is named for its ligament-like, fibrous meshwork appearance in aldehyde-fixed preparations. This meshwork fills much of the iridocorneal angle and runs circumferentially (thus annular) throughout the anterior segment. In zebrafish, this structure is visible by histology at approximately 17 dpf (Figure 7 and 8). It appears to differentiate from the mass of mesenchymal cells that are present during early developmental stages in the angle region. The shape and extent of the annular ligament varies greatly between dorsal and ventral regions of the eye. In the dorsal region, the inner surface of the annular ligament forms a deep "U" shape (Figure 7F). In the ventral region, the extent of the annular ligament is reduced and forms a "funnel" shape (Figure 8F). As part of this shape difference of the anterior segment angle, the iris projects slightly forward in the dorsal region, while in the ventral half, the iris bends posteriorly (Figure 9A). By histology, the annular ligament appears porous and devoid of nuclei. However, electron microscopy revealed that small, irregularly shaped nuclei are present in annular ligament cells (Figure 9C). Ultrastructurally these cells contain very large accumulations of non-membrane bound aggregates of what appears to be glycoprotein. In other teleosts, these cells have been described as glycogen aggregates [18], but in zebrafish, the fibrilar nature of the aggregated material appears to be glycoproteinatious in nature. Interspersed within the glycoprotein aggregates are islands of darker staining cytoplasm. The surface of the annular ligament facing the anterior chamber is lined by an endothelium. The endothelium, at the base of the annular ligament, becomes more villous with numerous cytoplasmic convolutions. The annular ligament itself thins at the edges and extends onto both the posterior surface of the cornea and the anterior surface of the iris at its base.
Ventral angle specializations
Serial histological sectioning in both embryos and adults revealed specializations at the iridocorneal angle within a narrow region of the ventral anterior segment (Figure 10). These specializations include a network of canals and openings that lead to episcleral vasculature. Within this region, the iris and ciliary zone dramatically change in morphology. The tip of the iris forms a nodule and bends sharply towards the vitreous (Figure 10I–L). Another feature of this ventral specialization is the loss of the non-pigmented ciliary epithelium. At the ciliary zone, the iris thickens and forms a thumb-like structure that projects into the vitreous and contains melanophores and non-pigmented myoepithelial cells (Figure 10I, 11B). By electron microscopy, the non-pigmented myoepithelial cells show thickened, electron dense intracellular filaments and may have contractile function to regulate lens position via the zonules (Figure 11H). A ventral canal is present as early as 3 dpf and matures significantly into a canalicular network as the embryo grows. This complex structure appears to arise as a specialization of the embryonic fissure.
Within this highly differentiated ventral region, the iris stroma has an expanded layer of non-pigmented mesothelial cells that form a band of tissue separating the annular ligament from the iris surface (Figure 11B, inset 11F). In young specimens this region is prone to separation during histological preparation, suggesting a loose extracellular matrix (for example Figure 12B). By 17 dpf, a distinct boundary is visible within the central part of this strip of mesenchymal cells. In the adult, this boundary develops into a channel that connects to an angular aqueous plexus and represents a major branch of the canalicular network (Figure 11B). The angular aqueous plexus is external to the iridocorneal angle, adjacent to the sclera, and in close proximity with connections to the vascular choroid. The mesothelial cells that line this canalicular network at the iridocorneal angle are highly involuted with whirls of internal membranes (Figure 11G). Ultrastructurally, cross-sections of these cells resemble absorptive intestinal epithelia (Figure 11E).
Figure 12 Ultrastructure of ventral ciliary canal. Diagrams of embryonic (A) and adult (E) zebrafish eye showing the locations of histological insets. Histological section through the central portion of the ventral ciliary canal (B). Low magnification TEM micrograph showing differentiating endothelial cells associated with the sinus (C). High magnification TEM micrograph of the undifferentiated endothelial cell, EN (D). Histological section through the central portion of the ventral ciliary canal (F). Low magnification TEM micrograph showing mature endothelial cells overlying the sinus (G). High magnification TEM micrograph of the mature endothelial cells (H). Note the cytoplasmic extensions and presence of large intracellular vacuoles.
Within this defined ventral region at the ciliary zone, there is another major branch of the canal that penetrates through the base of the iris and also connects to the angular aqueous plexus (Figure 12). Similar to the canal at the iridocorneal angle, this passage is also endothelial lined. At 5 dpf, flattened endothelial cells line the narrow opening which occasionally contains undifferentiated and red blood cells (Figure 12C, D). In the mature fish, the endothelial cells that line this ciliary canal have further differentiated (Figure 12G). These cells line the opening to the canal and show large vacuoles and extensive cyctoplasmic extensions (Figure 12H). The dorsal angle completely lacks this canalicular network and does not have either the band of absorptive-looking tissue at the iridocorneal angle or the subjacent ciliary passage.
Discussion
Zebrafish as a model for anterior segment development and disease
Zebrafish have several experimental advantages that have made this a valuable model organism for analysis of mechanisms of development and disease. Experimental advantages include the ability to efficiently conduct mutational analyses, including those designed to reveal complex genetic interactions. Ease of transgenesis facilitates gain-of-function and cell labeling studies. Cell labeling coupled with rapid, transparent, and external embryogenesis enable in vivo cell behavior studies during development. In our analysis, we have characterized the morphogenesis of the zebrafish anterior segment of the eye in order to establish an overview reference for future studies on the mechanisms of development and diseases of this structure. Ontogeny of the zebrafish anterior segment is protracted occurring over the first month of development. Some structures of the anterior segment such as the lens and cornea differentiate relatively early and show only peripheral growth. Other structures such as the iris, ciliary zone, and iridocorneal angle undergo dramatic morphological changes during this time. Our analysis of zebrafish anterior segment development has also identified differences between the dorsal and ventral regions. In particular, structures specialized for regulating aqueous humor dynamics are localized in a dorsoventral specific manner.
Comparing anterior segment development in zebrafish and mammals
Overall morphogenesis of the anterior segment in zebrafish is similar to mammals and other vertebrates but there are some differences. For example the lens, although derived from surface ectoderm, delaminates as opposed to invaginating. Periocular mesenchyme appears to contribute to the corneal endothelium, iris stroma and angle specializations like other vertebrates. However, there appears to be a proportionally smaller bolus of immigrating mesenchymal cells and migration does not occur in obvious waves as in the chick [4,19]. Local proliferation appears to be a driving force for growth of the anterior segment in zebrafish, similar to that in the mouse [6,20]. The relative timing of differentiation for the various components of the zebrafish anterior segment is like that for other vertebrates, suggesting conservation in tissue interaction and inductive events [19,21].
While hallmarks of anterior segment development in zebrafish appear to be largely conserved with other vertebrates, anatomical specializations exist. For example, the iris stroma is non-contractile and devoid of muscle cells and the ciliary "body" does not contain processes or a circumferential band of muscle. For this reason, we refer to the region adjacent and posterior to the iris as the ciliary zone, and not the ciliary body. The ciliary zone, however, is expanded and has a morphology consistent with contraction in a small area of the ventral-most region. The ciliary epithelium also displays dorsoventral differences. The ciliary epithelium in the dorsal eye appears specialized for secretion of aqueous humor while the ciliary epithelium in the ventral most region does not show extensive ultrastructural specializations. Instead, the ventral ciliary zone appears specialized for aqueous humor drainage. Other anatomical differences with mammals can be found within the iris. Within the iris stroma, in the place of muscle cells are lamina of pigment cells of various types. These include melanophores, iridophores, and xanthophores. The exact function of these chromophores in the iris are unknown, but they likely serve to prevent light from entering the eye in regions outside of the pupil. Finally, similar to the mammalian iris, the zebrafish iris is highly vascularized (Figure 5 and 6).
Morphological specializations at the iridocorneal angle
Another difference with mammals is found at the iridocorneal angle. In mammals, the angle is lined circumferentially with a trabecular meshwork, a complex structure of extracellular "trabecular" beams and a "meshwork" of absorptive endothelial cells. The trabecular meshwork forms a resistant barrier for aqueous humor outflow and covers Schlemm's canal. Schlemm's canal is a network of small canalicular passages leading to episcleral vessels where aqueous humor is ultimately drained. In zebrafish, the iridocorneal angle is lined with the annular ligament. This structure does not appear to be functionally analogous to the trabecular meshwork of mammals. However, within the ventral region of the angle, a specialized canalicular network of endothelial-lined openings does appear to be functionally analogous to the aqueous humor outflow structures of mammals. The circumferential ring of villous endothelial cells overlying the annular ligament at the iridocorneal angle may also function in absorption of metabolic byproducts, ions, or proteins in the aqueous humor. The anatomy of the mature zebrafish anterior segment suggests that in this species aqueous humor is produced primarily by dorsal ciliary epithelial cells, flows around the iris and into the anterior chamber, and then exits via a canalicular network in the ventral-most region. Additional experiments will be required to investigate this possibility further.
Conclusion
Cumulatively, our anatomical findings provide a reference point for utilizing zebrafish in genetic studies of the mechanisms of development and maintenance of the anterior segment. We find that while the morphological details of anterior segment structures in the zebrafish differ from higher vertebrates in many aspects, these structures show an overall conservation in anatomy and develop from similar cellular lineages. As in other vertebrates, surface ectoderm, periocular mesenchyme, and the anterior margin of the retinal neuroectoderm contribute to structures of the anterior segment. The timing of morphogenesis is also conserved among vertebrates. For example the lens and cornea are first to mature, while the iridocorneal angle does not reach adult morphology until later stages. In addition to providing a baseline for genetic studies of anterior segment development, our observations coupled with previous physiological approaches [22], provide insights into how aqueous humor is regulated in zebrafish and support the use of this species for investigation of anterior segment development and disease.
Methods
Specimens
Wild type zebrafish (Danio rerio) of the AB/AB and TL/TL backgrounds were reared under standard conditions with a light cycle of 14 h light/10 h dark [23]. No differences in anterior segment anatomy were observed between these two stains. Specimens were collected at various times of development from 24 hours to 2 months post fertilization. All individuals were photographed to document the body length prior to histological analysis. Prior to fixation, fish were anesthetized in 0.2 mg per ml of ethyl 3-aminobenzoate methanesulfonate (tricane). All experiments were performed in compliance with the ARVO statement for use of animals in vision research.
Light microscopy
Fish were fixed in primary fixative [2% paraformaldehyde, 2.5% glutaraldehyde, 3% sucrose, 0.06% phosphate buffer (pH 7.4)] at 4°C for at least 24 hours. Fish were washed in 0.1 M phosphate-buffered saline (PBS), dehydrated through an ethanol series and propylene oxide and then infiltrated with EMbed-812/Araldyte resin mixture. Traverse, semi-thin (1 μm), plastic sections were cut with a glass knife on a JB4 microtome. Serial sections were collected from the central retina, defined as maximum lens diameter at the optic nerve. Following heat fixation to glass slides, sections were stained with 1% Toluidine Blue in 1% Borax buffer. Images were captured using a Nikon coolpix 995 digital color digital camera mounted on a Nikon E800 compound microscope with a 60X oil-emersion objective.
Electron microscopy
Fish were fixed in primary fixative and washed as for LM. Specimens were then post-fixed with 1% Osmium Tetroxide on ice for 1 hour to preserve membranes. Fish were dehydrated through a methanol series and acetonitrile and infiltrated with EMbed-812/Araldyte resin mixture. Ultrathin sections (60–70 nm) were collected on coated grids and stained with uranyl acetate and lead citrate for contrast. Images were captured using Hitachi H600 TEM.
List of abreviations
TEM: transmission electron microscopy
hpf: hours post fertilization
dpf: days post fertilization
Authors' contributions
KS collected and processed all specimens for morphological analysis and helped to draft the manuscript. BL conceived of the study, participated in its design, coordination, and data interpretation, and helped to draft the manuscript. Both authors read and approved the final manuscript.
Acknowledgements
We greatly appreciate the technical expertise of Clive Wells for Electron Microscopy and Michael Cliff in maintaining the zebrafish stocks. We also thank Joseph Besharse, Simon John, Richard Smith, Elena Semina, and Sally Twinning for invaluable discussions. This work was supported by a Pilot Initiative grant from The Glaucoma Foundation (BL) and NIH grant R01EY16060 (BL).
==== Refs
Walls GL The vertebrate eye and its adaptive radiation 1942 New York, New York, Hafner Publishing Company
Leonard DW Meek KM Refractive indices of the collagen fibrils and extrafibrillar material of the corneal stroma Biophys J 1997 72 1382 1387 9138583
Gabelt BT Kaufman PL Kaufman PL and Alm A Aqueous humor hydrodynamics Adler's Physiology of the Eye 1997 10th St. Louis, MO, Mosby
Johnston MC Noden DM Hazelton RD Coulombre JL Coulombre AJ Origins of avian ocular and periocular tissues Exp Eye Res 1979 29 27 43 510425 10.1016/0014-4835(79)90164-7
Trainor PA Tam PP Cranial paraxial mesoderm and neural crest cells of the mouse embryo: co-distribution in the craniofacial mesenchyme but distinct segregation in branchial arches Development 1995 121 2569 2582 7671820
Smith RS Zabaleta A Savinova OV John SW The mouse anterior chamber angle and trabecular meshwork develop without cell death BMC Dev Biol 2001 1 3 11228591 10.1186/1471-213X-1-3
Gage PJ Pruckha SK Rhoades W Fate maps of neural crest and mesoderm in the mammalian eye Invest Ophthalmol Vis Sci 2003 44S 1085
Cvekl A Tamm ER Anterior eye development and ocular mesenchyme: new insights from mouse models and human diseases Bioessays 2004 26 374 386 15057935 10.1002/bies.20009
Gould DB Smith RS John SW Anterior segment development relevant to glaucoma Int J Dev Biol 2004 48 1015 1029 15558492 10.1387/ijdb.041865dg
Walter MA PITs and FOXes in ocular genetics: the Cogan lecture Invest Ophthalmol Vis Sci 2003 44 1402 1405 12657570 10.1167/iovs.02-0618
McMahon C Semina EV Link BA Using zebrafish to study the complex genetics of glaucoma Comp Biochem Physiol C Toxical Pharmacol 2004 138 343 350 10.1016/j.cca.2004.03.003
Easter SSJ Nicola GN The development of vision in the zebrafish (Danio rerio) Dev Biol 1996 180 646 663 8954734 10.1006/dbio.1996.0335
Schmitt EA Dowling JE Early eye morphogenesis in the zebrafish, Brachydanio rerio Journal of Comparative Neurology 1994 344 532 542 7929890 10.1002/cne.903440404
Hay ED Development of the vertebrate cornea Int Rev Cytol 1980 63 263 322 395131
Swamynathan SK Crawford MA Robison WGJ Kanungo J Piatigorsky J Adaptive differences in the structure and macromolecular compositions of the air and water corneas of the "four-eyed" fish (Anableps anableps) Faseb J 2003 17 1996 2005 14597669 10.1096/fj.03-0122com
Glass AS Dahm R The zebrafish as a model organism for eye development Ophthalmic Res 2004 36 4 24 15007235 10.1159/000076105
Fine BS Yanoff M Ocular histology 1979 2nd Hagerstown, MD, Harper and Row, Publishers, Inc.
Tripathi RC Davson H and Graham LT Comparative physiology and anatomy of the outflow pathway Comparative Physiology, The Eye 1974 5 New York, New York, Academic Press
Beebe DC Coats JM The lens organizes the anterior segment: specification of neural crest cell differentiation in the avian eye Dev Biol 2000 220 424 431 10753528 10.1006/dbio.2000.9638
Napier HR Kidson SH Proliferation and cell shape changes during ciliary body morphogenesis in the mouse Dev Dyn 2005 233 213 223 15759268 10.1002/dvdy.20302
Reneker LW Silversides DW Xu L Overbeek PA Formation of corneal endothelium is essential for anterior segment development - a transgenic mouse model of anterior segment dysgenesis Development 2000 127 533 542 10631174
Link BA Gray MP Smith RS John SW Intraocular pressure in zebrafish: comparison of inbred strains and identification of a reduced melanin mutant with raised IOP Invest Ophthalmol Vis Sci 2004 45 4415 4422 15557450 10.1167/iovs.04-0557
Westerfield M The Zebrafish Book 1995 Eugene, Oregon, University of Oregon Press
|
15985175
|
PMC1183199
|
CC BY
|
2021-01-04 16:30:31
|
no
|
BMC Dev Biol. 2005 Jun 28; 5:12
|
utf-8
|
BMC Dev Biol
| 2,005 |
10.1186/1471-213X-5-12
|
oa_comm
|
==== Front
BMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 1471-230X-5-211596974410.1186/1471-230X-5-21Research ArticleClinical presentation of abdominal tuberculosis in HIV seronegative adults Bolukbas Cengiz [email protected] Fusun F [email protected] Tulin [email protected] Remzi A [email protected] Nihat [email protected] Mehmet H [email protected] Ali T [email protected] Mithat [email protected] Erkan [email protected] Guray [email protected] Oya [email protected] Department of Internal Medicine, Gastroenterology Division, Faculty of Medicine, Harran University, Sanliurfa, Turkey2 Gastroenterology Clinic, Haydarpasa Numune Training and Research Hospital, Istanbul, Turkey3 Gastroenterology Clinic, Sisli Etfal Training and Research Hospital, Istanbul, Turkey4 General Surgery Clinic, Haydarpasa Numune Training and Research Hospital, Istanbul, Turkey5 Department of Chest Diseases, Faculty of Medicine, Harran University, Sanliurfa, Turkey6 Pathology Unit, Haydarpasa Numune Training and Research Hospital, Istanbul, Turkey2005 21 6 2005 5 21 21 23 2 2005 21 6 2005 Copyright © 2005 Bolukbas 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 accurate diagnosis of abdominal tuberculosis usually takes a long time and requires a high index of suspicion in clinic practice. Eighty-eight immune-competent patients with abdominal tuberculosis were grouped according to symptoms at presentation and followed prospectively in order to investigate the effect of symptomatic presentation on clinical diagnosis and prognosis.
Methods
Based upon the clinical presentation, the patients were divided into groups such as non-specific abdominal pain & less prominent in bowel habit, ascites, alteration in bowel habit, acute abdomen and others. Demographic, clinical and laboratory features, coexistence of pulmonary tuberculosis, diagnostic procedures, definitive diagnostic tests, need for surgical therapy, and response to treatment were assessed in each group.
Results
According to clinical presentation, five groups were constituted as non-specific abdominal pain (n = 24), ascites (n = 24), bowel habit alteration (n = 22), acute abdomen (n = 9) and others (n = 9). Patients presenting with acute abdomen had significantly higher white blood cell counts (p = 0.002) and abnormalities in abdominal plain radiographs (p = 0.014). Patients presenting with alteration in bowel habit were younger (p = 0.048). The frequency of colonoscopic abnormalities (7.5%), and need for therapeutic surgery (12.5%) were lower in patients with ascites, (p = 0.04) and (p = 0.001), respectively. There was no difference in gender, disease duration, diagnostic modalities, response to treatment, period to initial response, and mortality between groups (p > 0.05). Gastrointestinal tract alone was the most frequently involved part (38.5%), and this was associated with acid-fast bacteria in the sputum (p = 0.003).
Conclusion
Gastrointestinal tract involvement is frequent in patients with active pulmonary tuberculosis. Although different clinical presentations of patients with abdominal tuberculosis determine diagnostic work up and need for therapeutic surgery, evidence based diagnosis and consequences of the disease does not change.
==== Body
Background
Tuberculosis (TB) is rarely seen in western countries, but it is not an uncommon disease in the developing world [1]. Its reappearance has increased in association with the acquired immunodeficiency syndrome (AIDS) and immigration to the United States [2-4]. TB in its various forms remains an important cause of morbidity and mortality in developing countries and in patients with AIDS [5].
Recently, along with the increased incidence of TB, extrapulmonary TB incidence has also increased [6,7]. The occurrence of abdominal TB is independent of pulmonary disease in most patients, with a reported incidence of coexisting disease varying from 5 to 36% [8]. In patients with abdominal TB, the highest incidence of disease was noted in the gastrointestinal tract and in the peritoneum, followed by the mesenteric lymph nodes. Within the gastrointestinal tract, the ileocecal area is the most common site of involvement [9].
In general, upper gastrointestinal (GI), hepatic and pancreatic involvements are rare in autopsy series and case reports [10-12].
Because of chronic and nonspecific clinical and radiological findings mimicking several diseases, such as Crohn's disease, carcinoma, sarcoma [13], amebiasis, yersinia infection, gastrointestinal histoplasmosis, and periappendiceal abscess [9], the diagnosis of the abdominal TB requires a high index of suspicion and, despite using newly developed diagnostic tools, it usually takes a long time to get accurate diagnosis in clinic practice. The patients' subjective complaints on admission should guide the diagnostic procedures [14].
In this report, the effect of different clinical presentations of abdominal TB on clinical course, diagnosis and outcome of the disease in 88 immune-competent patients who were grouped according to the gastrointestinal symptoms at admission and followed prospectively between 1995 and 2004.
Methods
Patients, who were suspected to have abdominal tuberculosis through symptoms and / or operative findings, were prospectively evaluated. In addition to physical examination, basic laboratory studies, and anti-human immunodeficiency virus (HIV) serology, patients admitted with abdominal symptoms including pain, distension, nausea, vomiting, altered bowel habit, weight loss were evaluated for tuberculosis. Previous and family histories of TB were asked; chest X-ray, abdominal plain graph, tuberculin skin test and abdominal ultrasound (US) were routinely obtained during a nine-year period between 1995–2004. In case of suggestion of pulmonary TB in chest x-ray (infiltration or consolidation, nodularity, calcification, cavitary lesion, and fibrocalcific scar in lung parenchyma, pleural effusion, hilar or mediastinal lymphadenopathy), acid-fast bacilli (AFB) was searched in sputum and/or gastric juice. Sputum culture, and computed tomography (CT) of thorax were obtained. In case of negative results, bronchoscopy, and bronchoalveolar lavage plus biopsy were performed. If there was mediastinal lymphadenopathy (LAP) and the exact diagnosis could not be established by other means, mediastinoscopy with LAP biopsy was performed.
Based upon the clinical presentation, the patients were divided into five groups to guide diagnostic evaluations.
Group I: Patients with non-specific abdominal pain
Group II: Patients with ascites
Group III: Alteration in bowel habit (diarrhea, constipation or together)
Group IV: Acute Abdomen
Group V: Others (jaundice, dysphagia, fistula, fever of unknown origin, mass, weight loss)
Group I patients: abdominal CT, and according to the findings of CT, fine needle aspiration (FNA) biopsy or laparoscopy, upper and lower GI system endoscopy, and endoscopic biopsy for microbiologic and histopathologic diagnosis if a lesion seen.
Group II patients: ascitic cell count and type of cells, biochemistry analysis, smear for AFB, culture, IS6110-based in-house polymerase chain reaction (PCR) protocol for mycobacterium tuberculosis, and cytologic examination of ascitic fluid and abdominal CT.
Group III patients: colonoscopy and small bowel series, stool culture and AFB in stool if diarrhea present. Endoscopic biopsy for histopathologic examination and tissue culture and PCR for M. tuberculosis. Abdominal CT, and US. If ascites found, the exaimations in group II.
Group IV patients: histopathologic examination of surgical specimen if laparotomy performed. Abdominal CT, and colonoscopy if the patient is conservatively observed.
Group V patients: Upper endoscopy and thoraco-abdominal CT if dysphagia is present. Liver biopsy after excluding other causes if jaundice is present. Abdominal CT and if necessary MR and FNA in cases with abdominal mass were done. In patients with fistula only, AFB, culture and PCR of flux material; fistulography, CT, barium series or endoscopic examinations are performed to find out fistula tract.
Intraabdominal fluid (free or loculated; clear or complex with septae or debri), inter loop ascitis such as 'club sandwich' or 'sliced bread' signs which are due to localized fluid between radially oriented bowel loops, due to local exudation from the inflammed bowel, lymphadenopathy (discrete or conglomerated), bowel wall thickening and pseudo kidney sign were attributed to abdominal TB in abdominal US [15,16]. CT findings include adenopathy with low-density in centers, splenomegaly and hepatomegaly with nodules, ascites with complex features, bowel involvement, pleural effusion, intrasplenic, intrahepatic and intrapancreatic masses [8]. The colonoscopic features were defined as ulcers, nodules, deformed cecum and ileocecal valve, strictures, multiple fibrous bands and polypoid lesions [17].
Definitive diagnosis was based on one of the following methods with certain criteria: First of all, diagnostic priority for TB was given to show AFB microbiologically by culture or by both smear and PCR positivities, if that was not possible, the presence of caseating granulomatous lesions in the tissue was accepted as histopathological diagnosis of TB. If the diagnosis of TB was not obtained by using both of these methods, we achieved a clinical diagnosis of TB after a successful empiric therapeutic trial [9,18]. Empiric therapeutic trial was conducted for at least for 3 months with standard four drugs regiment (isonicotinic acid hydrazide, rifampin, pyrazinamide and ethambutol or streptomycin).
The following features were determined within each patient group: age, gender, disease duration, symptom frequency, TB presence in patient's history or environment, coexistence with pulmonary TB, tuberculin skin test status, laboratory findings, diseased segment within gastrointestinal segment, necessary diagnostic work-up, exact diagnosis methods, need for surgery except acute abdomen, response to therapy. The patients were followed up for adverse effects of medical treatment and surgical complications.
Numeric values were determined as percent or mean ± SD. Non-parametric Kruskal-Wallis test was used for quantitative results and chi-square test was used for qualitative results between groups. P value of less than 0.05 was considered to indicate the statistical significance.
Results
After extensive clinical, endoscopic, radiologic, microbiologic and histopathologic investigation, we identified and followed 88 abdominal TB cases with gastrointestinal involvement. Age, gender, and symptoms on admission are summarized in Table 1; laboratory, chest X-ray, abdominal plain film and US findings are summarized in Table 2.
The most frequent symptom was abdominal pain (28.4%). TB was present in patient history and environment in 17 and 9.1 percent, respectively. Ascites was the most frequent physical finding (35.2%). Other physical examination findings were hepatomegaly (22.7%), splenomegaly (17%), abdominal mass (17%), fistula (12.5%), peripheral LAP (11.4%) and acute abdomen (10.2%). Most of the patients were significantly anemic. Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) were mostly elevated in the study group. ESR was normal only in 4 cases (4.5%). Pulmonary active lesion or lesion consistent with prior TB were detected in 27.3 and 17 percent of the patients, respectively. There was significant association between AFB in sputum and GIS tract disease (p = 0.003). Ascites was detected in 40.9% of patients on US. On abdominal CT, besides ascites, gut wall thickness (frequently ileocecal region), LAP, abscess, and organomegaly etc. were detected in 81.8%.
According to clinical presentation, five groups were constituted as non-specific abdominal pain (n = 24), ascites (n = 24), alteration in bowel habit (n = 22), acute abdomen (n = 9), and others (n = 9). Age, gender, symptoms, need for surgery, diagnostic tests, how the diagnosis established (tissue and culture based against empiric), response to therapy according to the groups on admission are summarized in Table 3 and Table 4. With excluding grouping symptoms, mean white blood cell count (WBC) and lesion frequency in plain abdomen graphy were higher in patients admitting with acute abdomen (p = 0.002 and p = 0.014, respectively). Patients with alteration in bowel habit were statistically younger (p = 0.048). Abnormal finding frequency on abdominal CT was equal between groups (p = 0.196). CT guided FNA biopsy was preformed in 10 patients and its diagnostic yield was 70% in these patients. Colonoscopic lesion frequency was the lowest (7.5%) in colonoscopy-performed patients with ascites (p = 0.04). In our patients who had colonoscopic abnormalities, tissue culture and PCR positivity for TB was 11.1% and 19.2%, respectively. Although need for surgery rates was not different between patient groups, therapeutic surgery rate was significantly low (12.5%) in patients with ascites (p = 0.001). Gender, disease duration, associated symptoms, definitive diagnostic methods, response to therapy, time to response to therapy, and mortality were not different between groups (p > 0.05).
The localization and the type of disease were shown in Table 5. Gastrointestinal tract alone was the most frequent diseased part (38.5%) and iloececal region was the preferred area (53.8%) in the tract. Peritoneal and GI tract involvement proportions were approximately equal in patients admitted with abdominal pain. Isolated involvement of peritoneum was 71% in patients with ascites. The involvement rate of GI tract alone was 72.7% in patients with alteration in bowel habit. All patients with acute abdomen had GI tract involvement. Solitary mesenteric lymphadenitis (0.5%), liver (0.5%), pancreas (0.5%) and esophagus (0.5%) involvements were established.
In the whole group, tissue and culture based and empiric therapeutic trial diagnosis rates were 75.4 and 14.6 percent, respectively.
Peroperative mortality in 1 case, fistula in 3 cases, short bowel in 1 case, and chronic pancreatic insufficiency in 1 patient developed as complications.
Discussion
Patients with abdominal TB may have many symptoms and mimic many diseases, therefore if it is not clinically suspected, it may result in important morbidity and mortality. In abdominal TB series GI tract and peritoneum are reported as the most frequent sites of involvement [9]. Four major pathophysiologic mechanisms are proposed for abdominal TB: hematogenous spread, swallowing of infected sputum, ingestion of contaminated milk or food, and contiguous spread from adjacent organs [19].
In our patients, GI tract alone, peritoneum alone, and GI tract with peritoneum involvement (multi-site) were most frequent and detected in 38.5, 28, and 20.4 percent, respectively. Mesenteric lymph node involvement alone was rare, liver involvement was uncommon, but even esophageal and pancreatic involvement mimicking tumor was present.
The ileocecal area is the most frequent site of involvement in the GI tract [9,20]. Physiologic stasis, absorption of digested materials, and augmented lymphoid tissue at ileocecal area may be the causes of this property [21,22]. In our patients, GI tract and ileocecal region in particular, was also most frequently diseased part.
Gastroenterological symptoms of abdominal TB depend on the organ or tissue involved. Abdominal pain, tenseness, and diarrhea, as our patients demonstrated, are frequent symptoms. Also, as reported in the literature, active TB lesion or lesion consistent with prior TB in the lungs are detected in 27.3 and 17 percent of the whole group of patients, respectively [9,18]. Diagnosis is difficult in the absence of pulmonary involvement. As a result, clinical and radiologic signs of pulmonary TB must be searched for in every patient even in the absence of pulmonary symptoms. Meantime, absence of pulmonary findings does not rule out abdominal TB.
As reported previously [14], subjective complaints of the patients could guide the diagnostic procedures. For this reason, to group the patients according to their symptoms on admission may cause the diagnostic work up to be more cost effective. When our patients are grouped according to symptoms, disease duration, gender, weight loss, fever, and previous or family history of TB were not different between groups. There was no significant difference in frequency of radiologic pulmonary lesions, tuberculin skin test status, and ESR or CRP elevation between groups. This observation has not been reported before.
Although WBC, and abdominal plain film abnormalities were higher in patients with acute abdomens, these findings were nonspecific for TB diagnosis. On the other hand, macroscopic findings and histopathologic and microbiologic investigations of the tissues obtained during therapeutic surgery made abdominal TB diagnosis easier and faster in this specific patient group. Abdominal CT abnormalities regarding TB were observed in all patients with acute abdominal findings. This feature has given an opportunity to follow these patients, excluding those with perforation, with conservative medical treatment [8,23]. In this subgroup of patients colonoscopy during follow-up revealed abnormality in 75%, and provided histopathologic and microbiologic TB diagnosis in 50%.
Although both abdominal US and CT are equally sensitive in detecting ascites, CT is superior to US in detecting LAP, gut wall thickness, omental cake, and other abnormalities [24,25]. Abnormal abdominal CT findings in our series were higher than US findings and were equal in each group. CT shows localization and type of the lesion, but it is not specific for TB diagnosis [26,27]. CT guides the following diagnostic methods such as FNA, endoscopy, laparoscopy or laparotomy [16,27].
Therapeutic and diagnostic surgery rates were similar between our patient groups. This may be due to two reasons. Firstly, ascites examination, FNA biopsy, and endoscopic methods may be inadequate for definitive diagnosis. Secondly, patients with abdominal pain or altered bowel habit groups may develop complications requiring surgery. Diagnostic laparotomy was needed most often in patients with ascites. Infrequent detection of AFB in ascitic fluid, late results of ascites cultures, and uncommon culture yield of M. tuberculosis are reasons for increased laparoscopy/laparotomy need [28,29]. Diagnostic operation was lower in patients with alteration in bowel habit and the group in which GI tract was not directly involved. Endoscopic examinations increased diagnostic efficiency and decreased diagnostic surgical interventions in patients with alteration in bowel habit. Patients with acute abdominal symptoms frequently have GI tract and particularly ileocolonic involvement. Colonoscopy is effective to establish TB diagnosis in patients that could be managed medically in this group. Colonoscopy provides tissue for culture, PCR and histopathologic examination [30]. In our patients who had colonoscopic abnormalities, tissue culture and PCR positivity for TB rate was lower than or at most equal to previously reported series [17,31,32]. Low colonoscopic biopsy count and sample size may affect this result. Stool examination for AFB and culture for TB are proposed in the literature, but the association of GI TB with AFB in the sputum may cause misinterpretation [19].
AFB is searched for in the sputum and when necessary in gastric juice in our patients with pulmonary involvement. There was a positive association between AFB positivity in sputum and GI tract involvement. This association can be explained with swallowing infected sputum [19]. AFB in sputum rate was not different between groups; this can be explained by existence of GI tract involved patients in groups formed with patient symptomatology. Patients with alteration in bowel habit were younger as reported in the literature. Wig K and et al. reported that the maximum age incidence of intestinal tuberculosis was also 15 to 24 years [33]. The reason of the high incidence rate of intestinal tuberculosis in young population could be related to Peyer's patches which play a major role in intestinal immunity and are portals of entry for significant pathogens. Peyer's patches in the small intestine peaks at ages 15–25 and then declines with age [34].
There was no significant difference between evidence based diagnosis and response to therapy diagnosis in all groups.
Our diagnosis rate based on response to therapy was similar to the rates reported before [18]. There was no difference in evidence based and response to treatment diagnosis between our groups. There was also no difference in first response time to therapy between two diagnostic methods. Empiric therapeutic trial appears to be a useful method for rapid presumptive diagnosis and treatment of tuberculosis [35].
It is reported that there is no difference in response or relapse rates between 9 month and 18 month treatment times [36]. All our patients received therapy for 9 months and had a high cure rate. High cure and low relapse rates may be caused by immune competence of our patients and infrequency of primary drug resistance in this group. High adherence to therapy rate and low adverse effects of drugs to lower dose rate may influence secondary drug resistance.
Fistula, short bowel syndrome, and anastomosis leak are frequent complications of surgery performed in abdominal TB [37]. Our patients had similar complications.
Conclusion
Gastrointestinal tract involvement is frequent in patients with active pulmonary tuberculosis. Different clinical presentation of patients with abdominal TB involving GI system may determine the necessity of diagnostic or therapeutic surgical interventions, but evidence based diagnosis and prognosis are not influenced.
Abbreviations
TB, tuberculosis; AIDS, acquired immunodeficiency syndrome; GI, gastrointestinal; HIV, human immunodeficiency virus; US, ultrasound; AFB, acid-fast bacilli; CT, computed tomography; LAP, lymphadenopathy; PCR, polymerase chain reaction; FNA, fine needle aspiration; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; WBC, white blood cell count;
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CB conceived of the study, and participated in its design and coordination. FFB, TK, NA, HMS, ATI, MG, EC and GK collected the samples and clinical data, and carried out the laboratory analysis. ARD and FFB conceived of the study and participated in the sequence alignment and drafted the manuscript. OO participated in study design and coordination and revised the manuscript critically for important intellectual content. 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
Table 1 Demographic and clinic features in whole group.
Patients, (n) 88
Age (year) 31.4 ± 15
Gender (F/M) 48/40
Disease duration (month) 10.4 ± 19
Main symptom (n/%)
Abdominal pain 25 (28.4%)
Abdominal distension 23 (26.1%)
Diarrhea 15 (17%)
Fever 4 (4.5%)
Weight loss 4 (4.5%)
Ileus 4 (4.5%)
Perforation 4 (4.5%)
Others 4 (4.5%)
Dysphagia 2 (2.3%)
Icter 2 (2.3%)
Dyspepsia 1 (1.1%)
Previous history for TB 15 (17%)
Family history for TB 8 (9.1%)
F, Female; M, Male; TB, Tuberculosis.
Table 2 The results of the diagnostic methods in whole group.
ESR (mm/h) 68 ± 24
CRP (mg/L) 53 ± 39
WBC (/mm3 × 1000) 9.2 ± 3.6
Hct % 32.3 ± 5.2
Hb (g/dL) 10.6 ± 1.9
Chest X-ray %
Active lesion 24 (27.3%)
Lesion consistent with prior TB 15 (17%)
Normal 49 (55.7%)
AFB positivity in sputum % 11 (12.6%)
Lesion frequency in abdominal plain graphs % 21 (23.9%)
Abdominal US abnormality* 78 (88.6%)
*Including hepatomegaly-splenomegaly
ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; WBC, white blood cell count;
Hct, haemotocrit; Hb, haemoglobin; AFB, acid-fast bacteria; US, ultrasonography.
Table 3 Demographic and clinical features in divided groups based upon the clinical presentation.
Groups* I (n = 24) II (n = 24) III (n = 22) IV (n = 9) V (n = 9) P
Age (yr) 36.5 ± 20 30.4 ± 13 23.5 ± 8 32.7 ± 12 38 ± 16 0.048
Gender (F/M) 12/12 13/11 14/8 2/7 5/4 0.34
Disease Duration (mo) 11.3 ± 23 6 ± 8 9.8 ± 14 25 ± 41 6.6 ± 5 0.869
Previous history of TB % 16.7 8.3 18.2 22.2 33.3 0.530
Family history for TB % 8.3 4.2 18.2 11.1 0 0.425
Weight loss (kg) 8.2 ± 7 6 ± 4.5 8.5 ± 6 8.7 ± 5 8.1 ± 7.2 0.496
Temperature (°C) 37.5 ± 1 37.4 ± 0.8 37.5 ± 1 37.9 ± 0.9 37.2 ± 1 0.615
Fistula (%)* 8.3 0 31.8 0 22.2 0.07
Therapeutic Operation % 33.3 12.5 27.3 66.6 22.2 0.001
Response to therapy % 91.7 95.8 90.9 77.7 100 0.354
First response to therapy (mo) 2.5 ± 1.5 2.1 ± 1.6 1.9 ± 1.3 2.6 ± 1.5 2.8 ± 2 0.298
Death % 8.3 4.2 4.5 11.1 0 0.825
*Group I, non-specific abdominal pain; Group II, ascites; Group III, alteration in bowel habit (diarrhea, constipation or together); Group IV, acute abdomen; Group V, others (jaundice, dysphagia, fistula, fever of unknown origin, mass, weight loss);
Yr, year; F, female; M, male; Mo, month; TB, tuberculosis.
P value of less than 0.05 was considered to indicate the statistical significance in between groups.
Table 4 The results of the diagnostic laboratory methods, the rate of abnormality in imaging and invasive methods, and definitive diagnosis in divided groups based upon the clinical presentation.
Groups* I (n = 24) II (n = 24) III (n = 22) IV (n = 9) V (n = 9) P
WBC (/mm3 × 1000) 10 ± 4 8.7 ± 2.6 7.7 ± 1.8 14 ± 3.7 7.3 ± 2.8 0.002
ESR(mm/h) 67 ± 21 65 ± 31 71 ± 20 69 ± 21 72 ± 26 0.842
Tuberculin skin test (mm) 17 ± 7 18 ± 6.5 13 ± 7 18 ± 11 16 ± 9 0.119
AFB in sputum % 12.5 4.5 18.2 25 11 0.602
Chest X-ray % 54.2 27.5 40.9 66.6 22.2 0.448
APG % 20.8 12.5 31.8 66.7 0 0.014
Abdominal CT % 91.7 100 90.9 100 77.8 0.196
** Colonoscopy % 62.5 (5/8) 7.5 (2/15) 90.9 (20/22) 75 (3/4) 33.3 (1/3) 0.04
Operation % 54.2 54.2 33.3 75 22.2 0.156
Diagnosis %
Evidence based 79.2 83.3 86.4 88.9 88.9
Histopathologic 62.5 50 45.5 44.4 55.6
Microbiologic 4.2 12.5 9.1 33.3 11.1 0.933
H+M 12.5 20.8 31.8 11.1 22.2
ETT% 20.8 16.7 13.6 11.1 11.1
*Group I, non-specific abdominal pain; Group II, ascites; Group III, alteration in bowel habit (diarrhea or constipation or together); Group IV, acute abdomen; Group V, others (jaundice, dysphagia, fistula, fever of unknown origin, mass, weight loss);
** The percent in patients with colonoscopy performed WBC, white blood cell count; ESR, erythrocyte sedimentation rate; AFB, acid-fast bacteria; APG, abdominal plain graphy; CT, computed tomography; H, histopathologic; M, microbiologic; ETT, empiric therapeutic trial. P value of less than 0.05 was considered to indicate the statistical significance in between groups.
Table 5 The distribution of involvement area in the groups*.
Localization Group I (n = 24) Group II (n = 24) Group III (n = 22) Group IV (n = 9) Group V (n = 9) Total (n)
Multiple (n) (Peritoneum, SB, colon, pericardium) 2 7 6 1 2 18
GI Tract (n) 34
SB 3 2 1
Ileocecal 6 7 5
Colon 6 1
Stomach 1
Esophagus 2
Liver 3 1 4
Peritoneum 8 17 25
Mesenteric LAP 3 2 5
Pancreas and biliary system 1 1 2
Total (n) 24 24 22 9 9 88
*Group I, non-specific abdominal pain; Group II, ascites; Group III, alteration in bowel habit (diarrhea or constipation or together); Group IV, acute abdomen; Group V, others (jaundice, dysphagia, fistula, fever of unknown origin, mass, weight loss);
SB, small bowel; GI, gastrointestinal; LAP, lymphadenopathy
==== Refs
Shukla HS Hughes LE Abdominal tuberculosis in 1970s: a continuing problem Br J Surg 1978 65 403 405 656757
Guth AA Kim U The reappearance of abdominal tuberculosis Surg Gynecol Obstet 1991 172 432 436 2035131
Snider DE JrRoper WL The new tuberculosis N Engl J Med 1992 326 703 705 1736110
Probert CS Jayanthi V Wicks AC Carr-Locke DL Garner P Mayberry JF Epidemiological study of abdominal tuberculosis among Indian migrants and the indigenous population of Leicester, 1972–1989 Gut 1992 33 1085 1088 1398232
Aliyu MH Salihu HM Tuberculosis and HIV disease: two decades of a dual epidemic Wien Klin Wochenschr 2003 115 685 697 14650943
Das P Shukla HS Clinical diagnosis of abdominal tuberculosis Br J Surg 1976 63 941 946 1009343
Louie E Rice LB Holzman RS Tuberculosis in non-Haitian patients with acquired immunodeficiency syndrome Chest 1986 90 542 545 3757564
Hulnick DH Megibow AJ Naidich DP Hilton S Cho KC Balthazar EJ Abdominal tuberculosis: CT evaluation Radiology 1985 157 199 204 4034967
Marshall JB Tuberculosis of the gastrointestinal tract and peritoneum Am J Gastroenterol 1993 88 989 999 8317433
Auerbach O Acute generalized military tuberculosis Am J Pathol 1944 20 121 136
Leder RA Low VH Tuberculosis of the abdomen Radiol Clin North Am 1995 33 691 705 7610239
Amarapurkar DN Patel ND Amarapurkar AD Primary gastric tuberculosis – report of 5 cases BMC Gastroenterol 2003 3 6 12703983 10.1186/1471-230X-3-6
Howell JS Knapton PJ Ileo-caecal tuberculosis Gut 1964 19 524 529 14244026
Ebdrup L Storgaard M Jensen-Fangel S Obel N Ten years of extrapulmonary tuberculosis in a Danish university clinic Scand J Infect Dis 2003 35 244 246 12839152 10.1080/00365540310000274
Kedar RP Shah PP Shivde RS Malde HM Sonographic findings in gastrointestinal and peritoneal tuberculosis Clin Radiol 1994 49 24 29 8299328
Sharma MP Bhatia V Abdominal tuberculosis Indian J Med Res 2004 120 305 315 15520484
Misra SP Misra V Dwivedi M Gupta SC Colonic tuberculosis: clinical features, endoscopic appearance and management J Gastroenterol Hepatol 1999 14 723 729 10440219 10.1046/j.1440-1746.1999.01940.x
Khan MR Khan IR Pal KM Diagnostic issues in abdominal tuberculosis J Pak Med Assoc 2001 51 138 142 11759494
Horvath KD Whelan RL Intestinal tuberculosis: Return of an old disease Am J Gastroenterol 1998 93 692 696 9625110 10.1111/j.1572-0241.1998.207_a.x
Wadhwa N Agarwal S Mishra K Reappraisal of abdominal tuberculosis J Indian Med Assoc 2004 102 31 42 15195856
Patel N Amarapurkar D Agal S Baijal R Kulshrestha P Pramanik S Gupte P Gastrointestinal luminal tuberculosis: establishing the diagnosis J Gastroenterol Hepatol 2004 19 1240 1246 15482529 10.1111/j.1440-1746.2004.03485.x
Moss JD Knauer CM Tuberculous enteritis: A report of three patients Gastroenterology 1973 65 959 966 4201822
Anand BS Nanda R Sachdev GK Response of tuberculous stricture to antituberculous treatment Gut 1988 29 62 9 3343015
Balthazar EJ Gordon R Hulnick D Ileocecal tuberculosis: CT and radiologic evaluation AJR Am J Roentgenol 1990 154 499 503 2106212
Suri S Gupta S Suri R Computed tomography in abdominal tuberculosis Br J Radiol 1999 72 92 98 10341698
Koh DM Burn PR Mathews G Nelson M Healy JC Abdominal computed tomographic findings of Mycobacterium tuberculosis and Mycobacterium avium intracellulare infection in HIV seropositive patients Can Assoc Radiol J 2003 54 45 50 12625084
Buxi TB Vohra RB Sujatha Y Chawla D Byotra SP Gupta PS Dewan VK Kanwar C CT appearances in macronodular hepatosplenic tuberculosis: a review with five additional new cases Comput Med Imaging Graph 1992 16 381 387 1468072 10.1016/0895-6111(92)90056-F
Demir K Okten A Kaymakoglu S Dincer D Besisik F Cevikbas U Ozdil S Bostas G Mungan Z Cakaloglu Y Tuberculous peritonitis – reports of 26 cases, detailing diagnostic and therapeutic problems Eur J Gastroenterol Hepatol 2001 13 581 585 11396540 10.1097/00042737-200105000-00019
Ibrarullah M Mohan A Sarkari A Srinivas M Mishra A Sundar TS Abdominal tuberculosis: diagnosis by laparoscopy and colonoscopy Trop Gastroenterol 2002 23 150 153 12693163
Pulimod AB Ramakrishna BS Kurian G Peter S Patra S Mathan VI Mathan MM Endoscopic mucosal biopsies are useful in distinguishing granulomatous colitis due to Crohn's disease from tuberculosis Gut 1999 45 537 541 10486361
Singh V Kumar P Kamal J Prakash V Vaiphei K Singh K Clinicocolonoscopic profile of colonic tuberculosis Am J Gastroenterol 1996 91 565 568 8633510
Pfaller MA Application of new technology to the detection, identification, and antimicrobial susceptibility testing of mycobacteria Am J Clin Pathol 1994 101 329 37 7510928
Wig KL Chitkara NL Gupta SP Kishore K Manchada RL Ileocecal tuberculosis with particular reference to isolation of Mycobacterium tuberculosis. With a note on its relation to regional ileitis (Crohn's disease) Am Rev Resp Dis 1961 84 169 178 13785163
Van Kruiningen HJ West AB Freda BJ Holmes KA Distribution of Peyer's patches in the distal ileum Inflamm Bowel Dis 2002 8 180 185 11979138 10.1097/00054725-200205000-00004
Anglaret X Saba J Perronne C Lacassin F Longuet P Leport C Vilde JL Empiric antituberculosis treatment: benefits for earlier diagnosis and treatment of tuberculosis Tuber Lung Dis 1994 75 334 340 7841426 10.1016/0962-8479(94)90078-7
Kim SG Kim JS Jung HC Song IS Is a 9-month treatment sufficient in tuberculous enterocolitis? A prospective, randomized, single-centre study Aliment Pharmacol Ther 2003 18 85 91 12848629 10.1046/j.1365-2036.2003.01599.x
Ha HK Ko GY Yu ES Yoon K Hong WS Kim HR Jung HY Yang SK Jee KN Min YI Auh YH Intestinal tuberculosis with abdominal complications: radiologic and pathologic features Abdom Imaging 1999 24 32 38 9933670 10.1007/s002619900436
|
15969744
|
PMC1183200
|
CC BY
|
2021-01-04 16:03:27
|
no
|
BMC Gastroenterol. 2005 Jun 21; 5:21
|
utf-8
|
BMC Gastroenterol
| 2,005 |
10.1186/1471-230X-5-21
|
oa_comm
|
==== Front
BMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 1471-230X-5-231600461610.1186/1471-230X-5-23Research ArticleThe impact of illness in patients with moderate to severe gastro-esophageal reflux disease El-Dika Samer [email protected] Gordon H [email protected] David [email protected]'innocenti Alessio [email protected] Ingela [email protected] Carlo A [email protected] Lisa [email protected] Zanten Sander Veldhuyzen [email protected] Diane [email protected] Peter [email protected] Naoki [email protected] Alan N [email protected] Peggy [email protected]ünemann Holger J [email protected] Division of Gastroenterology, Veterans affairs medical center, Salem, Virginia, USA2 Department of Clinical Epidemiology and Biostatistics' McMaster University, Hamilton, Ontario, Canada3 Department of Medicine, McMaster University, Hamilton, Ontario, Canada4 AstraZeneca R&D, Clinical Science, Mölndal, Sweden5 Division of Gastroenterology, McGill University Health Center, Montreal, Quebec, Canada6 AstraZeneca R&D, Canada, Mississauga, Ontario, Canada7 Division of Gastroenterology, Dalhousie University, Halifax, Nova Scotia, Canada8 Surrey GI Clinic/Research, Guelph, Ontario, Canada9 Department of Medicine, School of Medicine and Biomedical Sciences, University at Buffalo, State University of New York, Buffalo, New York, USA10 Division of Clinical Research Development and Information Translation/INFORMA, Italian National Cancer Institute, Rome/Istituto Regina Elena, Rome, Italy2005 10 7 2005 5 23 23 16 2 2005 10 7 2005 Copyright © 2005 El-Dika 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
Gastro-esophageal reflux disease (GERD) is a common disease. It impairs health related quality of life (HRQL). However, the impact on utility scores and work productivity in patients with moderate to severe GERD is not well known.
Methods
We analyzed data from 217 patients with moderate to severe GERD (mean age 50, SD 13.7) across 17 Canadian centers. Patients completed three utility instruments – the standard gamble (SG), the feeling thermometer (FT), and the Health Utilities Index 3 (HUI 3) – and several HRQL instruments, including Quality of Life in Reflux and Dyspepsia (QOLRAD) and the Medical Outcomes Short Form-36 (SF-36). All patients received a proton pump inhibitor, esomeprazole 40 mg daily, for four to six weeks.
Results
The mean scores on a scale from 0 (dead) to 1 (full health) obtained for the FT, SG, and HUI 3 were 0.67 (95% CI, 0.64 to 0.70), 0.76 (95% CI, 0.75 to 0.80), and 0.80 (95% CI, 0.77 to 0.82) respectively. The mean scores on the SF-36 were lower than the previously reported Canadian and US general population mean scores and work productivity was impaired.
Conclusion
GERD has significant impact on utility scores, HRQL, and work productivity in patients with moderate to severe disease. Furthermore, the FT and HUI 3 provide more valid measurements of HRQL in GERD than the SG. After treatment with esomeprazole, patients showed improved HRQL.
==== Body
Background
The global prevalence of gastro-esophageal reflux disease (GERD), defined as heartburn once daily, is estimated to range from 5 to 7% but varies widely and depends on the definition used [1]. For example, 25% of the adult population in Belgium [2], nearly 18% in Australia [3], 20% in the United States and 9 % in Canada reported GERD symptoms once a week or more [4,5]. While heartburn is the leading symptom in GERD, the disease is associated with a broad range of esophageal problems, including acid regurgitation, epigastric pain, esophageal erosions and complications such as Barrett's esophagus, esophageal adenocarcinoma and esophageal stricture. GERD is also associated with a variety of extra-esophageal problems, including sleep disturbances, noncardiac chest pain, asthma, chronic cough and hoarseness [6].
Clinicians and health services researchers are becoming more aware of the importance of patient-reported outcomes (PROs), including health related quality of life (HRQL) in understanding the burden of disease and the outcome of medical treatment. While physiologic measures provide information to clinicians, these outcomes are often not important to patients [7] and they correlate poorly with functional status or well-being, the areas in which patients are mostly interested. For example, the majority of patients with typical symptoms of GERD do not have endoscopic evidence of esophagitis [8]. PROs assist in providing a better understanding of treatment outcomes from the patient's perspective by translating clinical improvement into patient-important outcomes. Moreover, HRQL assessment is important for measuring quality of care, clinical effectiveness, and in reimbursement decisions [9,10].
The use of validated questionnaires is appropriate for measuring PROs in clinical trials [11,12]. There are two categories of HRQL measures; disease-specific and generic HRQL instruments [13,14]. Disease-specific instruments are used to describe the burden of disease and treatment outcomes in patients with a specific disease, and generic instruments measure the overall HRQL of patients, including physical, emotional, and social function, as well as their level of general performance at work and in daily life across different diseases [15].
Utility measures, one type of generic health status measures, are based on economic and decision theory [16]. These instruments measure patient preferences and generate preference or utility scores for respondents' health states on a 0 to 1.0 scale where 0 typically equals dead and 1.0 is full health [17]. In this manuscript we will use the term "utility" for all scores generated with preference based instruments although. The standard gamble (SG) is regarded as the reference standard for utility measurement [18-20]. Another utility instrument is the feeling thermometer (FT), a visual analogue scale presented in the form of a thermometer [18]. When completing this instrument, patients choose the score on the thermometer that represents the value they place on their health state. It is far simpler than the SG and has shown good responsiveness and validity in several studies [21-24]. The health utility index 3 (HUI 3) is a multi attribute utility instrument designed to classify a patient's health status based on rating of a set of defined items [25,26].
The aim of this analysis was to address the impact of GERD on utility scores, HRQL and work productivity in patients with moderate to severe GERD and to evaluate the construct validity of utility instruments as measures of HRQL.
Methods
Patients
We enrolled 249 uninvestigated outpatients with a clinical diagnosis of moderate to severe GERD in 13 specialty centers and 4 general practices across Canada between March 2002 and March 2003. Table 1 lists the definitions of symptoms severity. To evaluate the aims described in the introduction of this manuscript, we utilized data from a study that had as primary aim the comparison of two different formats of administering the SG and the FT [27]. Therefore, the current publication describes the data from the baseline visit of the 217 patients (87%) who completed the study and the responsiveness of the QOLRAD and the four symptoms questionnaire after four to six weeks of treatment with a proton pump inhibitor (PPI), esomeprazole. The main inclusion criteria were a clinical diagnosis of GERD, main symptom of heartburn, age greater 18 years, symptomatic for three months or longer, and off PPI for the 2 weeks prior to questionnaire administration. GERD was defined as a burning feeling, rising from the stomach or lower part of the chest up towards the neck. We describe the detailed eligibility criteria elsewhere [27].
Study design
Patients completed several utility and HRQL instruments during a clinic visit and provided demographic and clinical data. The visit lasted approximately 80 minutes. An experienced research coordinator from the method center trained all site interviewers in a daylong session in HRQL instrument administration. Ethic review boards at all study sites approved the study protocol and all patients signed an informed consent form prior to enrollment in the study.
Utility measures
The feeling thermometer (FT) is a visual analogue scale shown as a thermometer in which the best state is full health (equal to a score of 100) and the worst state is dead (a score of 0) [18]. It has shown good responsiveness and validity in several studies [21-24]. In this trial, we used a self-administered form of the FT [27,28].
The SG [18-20] offers patients two options from which a choice must be made: Choice A is the certain outcome that the patient will stay in a health state (their own health state, or a marker state) for t years until death. We varied t depending on the patient's age as follows: patients aged more than 80 years, t = the rest of the patient's lifetime; age 76 – 80 years, t = 10 years; age 66 – 75 years, t = 15 years; age 56 – 65, t = 25 years; age 46 – 55 years, t = 30 years; age 36 – 45 years, t = 35 years; age 26 – 35 years, t = 40 years; age 18 – 25 years, t = 45 years. Specifying the duration of remaining life means that patients use the same time frame as other patients of the same age, and reduces the random error that might result from patients inferring different time frames. Varying time frame by age minimizes an additional lack of realism that could arise if one chose a single time frame and either young patients have an unrealistically short duration of remaining life, or old patients have an unrealistically long duration. The alternative (choice B) is a hypothetical treatment with 2 outcomes: 1) returning to full health (probability p) for t years, at the end of which the patient dies or 2) immediate death (probability 1-p). Interviewers used a chance board with a ping-pong approach varying the probability p in steps of 0.05 to obtain the value, p*, where the patient considered choice A equal to choice B. This indifference probability, p*, is the utility value for the patient's own health in choice A in the interval from dead (= 0) to full health (= 1). The greater the respondent's willingness to accept the risk of a worse outcome (e.g dead) to avoid the health state in choice A, then the lower is the utility of the state in choice A to them.
The HUI 3 is a 15 item self-administered questionnaire. It has 8 attributes that include vision, hearing, speech, ambulation, dexterity, emotion, cognition and pain. We calculated a utility score on a 0 to 1.0 scale where 0 represents dead and 1.0 represents full health [26]. HUI has been shown to be a reliable, responsive and valid measure in a wide variety of clinical studies [25].
In addition we used the four symptoms questionnaire that comprises a series of four questions on which patients rate how they felt for the past week using a seven-point Likert scale ranging from no problem to very severe problem. The four symptoms questionnaire evaluates heartburn, acid reflux, belching, and stomach ache.
Disease-specific HRQL
The quality of life in reflux and dyspepsia (QOLRAD) consists of 25 items across five dimensions: emotional distress, sleep dysfunction, vitality, food/drink problems, and physical/social functioning. Patients provide answers on a seven-point Likert-type scale. The lower the value, the more severe is the impact on daily functioning. The QOLRAD is reliable, valid and responsive [29-31].
GERD-specific work productivity and activity impairment questionnaire
The GERD-specific Work Productivity and Activity Impairment questionnaire (WPAI-GERD) contains 8 items (Table 2) that uses a one-week recall period and measures absence from work, reduced productivity while at work and reduced productivity while doing regular daily activities other than work [32].
Health profile measure
The Medical Outcomes Study Short Form-36 (SF-36) is a generic instrument that assesses a wide range of health problems, including GERD [34]. It consists of 8 domains including physical functioning, role limitations-physical, bodily pain, general health, vitality, social functioning, role limitations-emotional, and mental health. The SF-36 scores range from 0 to 100, with higher scores indicating better functioning and well-being.
HRQL and utilities for other health states
We compared the SF-36 scores for the Canadian general population reported by Hopman et al. [33], and the US general population, depression, hypertension, diabetes extracted from the Medical Outcome Study and reported by Revicki et al. [34-37] to the SF-36 scores of our study patients. In addition, we compared the utility scores of patients enrolled in this study with those of other patients reported previously, as utility ratings are comparable across conditions because they provide scores between 0 (dead) and 1 (full health) that are not disease specific [38,39]. A priori, we determined that we would focus on patients with common diseases for whom utility assessments are available.
Statistical analysis
We report baseline instrument scores as means with 95% confidence intervals (CI). The QOLRAD and the four symptoms questionnaire responsiveness to treatment are shown as mean change scores with 95% CI. We also evaluated the cross-sectional construct validity of the utility instruments by calculating Pearson's correlation coefficients of the scores on the FT, SG and HUI 3 with the validation instruments the QOLRAD, the SF-36, and the four symptoms questionnaire. We assumed that higher correlations with the validation instruments would indicate greater construct validity. For interpretation of the correlations we considered correlations of less than 0.2 as very weak, from 0.2 to 0.35 as weak, from 0.35 to 0.5 as moderate and of more than 0.5 as strong. We compared mean physical and mental component summary scores on the SF-36 of our study population and the previously reported scores of the Canadian population, the US population, clinical depression, diabetes, and hypertension using Student's t tests [33-37].
Results
Baseline characteristics and burden of disease
Table 3 shows the baseline characteristics of the patients. The mean age was 50 years (range 20 to 82), 53% were females, 69%% were employed, and the mean number of months since diagnosis was 86 (range 1 to 504). Table 4 presents mean QOLRAD scores and the mean scores for the four symptoms questionnaire along with the mean scores change after 4 weeks of PPI treatment. GERD has the greatest impact on the vitality and food/drink domains of the QOLRAD (scores of 4.3 and 3.8 respectively). The most severe symptoms were heartburn and acid reflux with scores of 4.5 and 4.1 respectively. PPI treatment significantly improved the scores of the QOLRAD and the four symptoms.
WPAI-GERD
A total of 153 (71%) patients were employed. The percentage of overall work impairment secondary to GERD that included absence from work plus time lost due to reduced productivity was 16% (95% CI, 12.9 to 18.8), which corresponds to 6.7 hours lost per week due to GERD symptoms. Furthermore, the reduced productivity during activities other than work in 216 patients was 21% (95% CI, 18.0 to 24.0).
Construct validity of the utility instruments
The correlations of the FT and HUI 3 with the QOLRAD and SF-36 domains were moderate. However, the SG showed lower correlations than the FT and the HUI 3. The FT had the highest, albeit weak correlation with the four symptoms questionnaire (Table 5).
Comparison of utility scores with other diseases
The systematic review by Morimoto et al. of utility measures reported SG weighted means for asthma, chronic renal failure, and angina pectoris of 0.88 (range 0.82–0.91), 0.52 (range 0.49–0.55), and 0.76 (range 0.64–0.97), respectively [38]. The systematic review by Post et al. revealed time trade off (TTO) and SG in survivors of minor stroke to be 0.72 (range 0.71–0.81), and 0.89 (range 0.81–0.95) respectively [39]. We previously reported baseline scores for the FT and SG in patients with moderate to severe COPD to be 0.60 (SD 0.18), and 0.66 (SD 0.27) respectively [24]. In this study, the utilities obtained for the FT, SG, and HUI 3 were 0.67 (95% CI, 0.64–0.70), 0.78 (95% CI, 0.75–0.80), and 0.80 (95% CI, 0.77–0.82) respectively.
Comparison of the SF-36 scores to other diseases and the general population
The mean scores on the SF-36 range from 42.8 to 47.7 across the different SF-36 domains (Table 6). These scores are significantly lower than the Canadian and US general population mean scores on the physical and mental component summaries (table 7) [33,35]. Table 7 also demonstrates the comparison of the SF-36 scores of our patients to other groups of patients included in the Medical Outcomes Study and reported by Revicki et al. [35,36,40].
Discussion
We determined the impact of GERD on utility scores, HRQL and work productivity in patients with moderate to severe GERD and the cross-sectional construct validity of three utility instruments (FT, SG, and HUI 3). Although the comparisons we made are indirect, the results of this study indicate that GERD causes important reductions in HRQL and utility when compared to the Canadian and US general population and to those of patients with a range of other chronic conditions. Heartburn and acid reflux were reported by our patients to have the worst impact on symptoms using the four symptoms questionnaire. The QOLRAD as well as the symptom scores improved after esomeprazole treatment although this study did not include a placebo group. In addition, our data indicate that GERD causes a considerable loss in work productivity.
The strengths of this study include the use of several utility and HRQL instruments that allow comparison to other chronic conditions. By using several quality of life instruments, we have studied GERD patients more comprehensively than previous studies [31,32,35]; however, our results are confined to GERD patients with moderate to severe symptoms who were participating in a clinical study. In addition, our comparison with population data is based on historical data. Despite the impressive responsive of the QOLRAD scores to esomeprazole treatment, one has to keep in mind that this study was not a randomized controlled trial. In regards to evaluation of validity, another limitation of our study is that we did not generate a priori predictions regarding correlations between the utility instruments and other measures. Had we generated such a priori predictions our conclusion about the validity of the instruments might be stronger.
Data on utility measures in GERD patients are sparse. We observed important disutility measured with the FT, SG, and HUI 3. The results suggest that the FT and HUI 3 are valid tools for the assessment of HRQL in patients with GERD. We were specifically interested in exploring the relative validity of utility instruments in patients with GERD.
In general, the correlations with other HRQL instruments were moderate. In contrast, the SG showed poor construct validity. Moreover, the FT shows better correlation with the four symptoms questionnaire than the SG and the HUI 3. Thus, our findings suggest that the FT and the HUI 3 are more appropriate indicators of HRQL impairment than the SG in these patients. Studies that use the FT, SG, HUI and SF-36 simultaneously are rare. We have previously observed a similar pattern of correlations in patients with chronic obstructive pulmonary disease (COPD) [24]. The correlations of the FT and HUI 3 with the SF-36 were higher compared to those of the SG with the SF-36. Thus, there is external evidence that the FT and HUI 3 show greater validity for the assessment of HRQL than the SG.
The data also indicate that the disutility in patients with moderate to severe GERD is similar to that of moderate to severe COPD [24]. The utility scores obtained with the SG are lower than what was previously reported for SG scores in patients with asthma [38] and comparable to those of minor stroke survivors [39].
We also found reductions in the SF-36 scores on all 8 domains in GERD patients. Compared to the Canadian and US general population, patients in this study had significantly reduced scores in the SF-36 MCS and PCS [33-36]. The SF-36 PCS scores are comparable to patients with clinical depression and hypertension. On the other hand, GERD patients have significantly worse SF-36 MCS scores than patients with diabetes mellitus and depression.
The QOLRAD results suggest that GERD has the greatest impact on the vitality and food/drink domains. These results are similar to those reported by Wiklund et al. [31] confirming the negative impact of GERD on the daily functioning of affected patients.
Health administrators and payers are interested in the magnitude of work productivity loss due to GERD [41]. Wahlqvist et al. showed that patients with GERD symptoms report 23% reduced productivity while at work, and 30% reduced productivity while doing regular daily activities in a Swedish population [32]. Our study supports the finding of impaired productivity in patients with moderate to severe GERD, but the estimates of work loss are somewhat lower demonstrating that about 16% of the work time is lost due to the illness. Since GERD affects approximately 9% of the Canadian population [5], impaired productivity has important economic consequences on society if it is not treated effectively.
Conclusion
In summary, we found that GERD has significant impact on utilities scores, HRQL and work productivity in patients with moderate to severe illness. The impact of the disease is similar in magnitude to other chronic conditions that are less responsive to treatment. In addition, utility instruments such as the FT and HUI 3 provide valid measurements of the impact of GERD on HRQL.
Competing interests
This study was funded by a grant from AstraZeneca, Sweden. Several authors of this manuscript (ADI, IW, LT, PW) are employees of the sponsor. Several authors have received research funding and/or honoraria from the sponsor (HJS, GHG, DA, NC, AB, CF, SVZ). NC and SVZ received honoraria for speaking about related conditions such as GERD. SVZ is a member of an AstraZeneca advisory board. AB is a consultant to AstraZeneca, Canada. The honoraria of Dr. Holger Schünemann and Dr. Gordon Guyatt were deposited into research accounts at the University at Buffalo and McMaster University.
Authors' contributions
Samer El-Dika was the project leader for this manuscript interpreted the data, performed data analysis and drafted the final version of the manuscript as well as early versions. Gordon H Guyatt and Holger Schünemann were the principal investigators of the study, wrote the clinical protocol and grant application, are responsible for the study protocol, interpreted data and participated in writing the final as well as early versions of this manuscript. Ingela Wiklund contributed to the study protocol, interpreted the data and edited the manuscript. Diane Heels-Ansdell was responsible for the statistical analysis and edited the final manuscript as well as early versions. David Armstrong was co-principal investigator of the study, revised the clinical protocol, assessed patients, interpreted data and edited the final manuscript as well as early versions. Peter Wahlqvist interpreted the WPAI-GERD data and edited the manuscript. Carlo A Fallone, Naoki Chiba and Sander Veldhuyzen van Zanten revised the clinical protocol, assessed patients, interpreted data and edited the final manuscript as well as early versions. Alessio Degli'Innocenti, Alan N Barkun, and Peggy Austin revised the clinical protocol, interpreted data and edited the final manuscript as well as early versions. Peggy Austin also co-ordinated the study. Lisa Tanser contributed to co-ordination of the study. All authors approved the final manuscript. The AstraZeneca global publications group reviewed the manuscript and made minor wording suggestions to clarify methodological aspects.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Participating investigators and affiliation: Dr. Iain Murray, Quest Clinical Trials, Markham, Ontario; Dr. Daniel Sadowski, Hys Medical Centre, Edmonton, Alberta; Dr. Alan Barkun & Dr. Serge Mayrand, Montreal General Hospital, Montreal, Quebec; Dr. Ford Bursey, St. John General Hospital, St. John's, New Foundland; Dr. Naoki Chiba, Surrey GI Clinic/Research, Guelph, Ontario; Dr. Lawrence Cohen, Sunnybrook & Women's College, Toronto, Ontario; Dr. Carlo Fallone, Royal Victoria Hospital, Montreal, Quebec; Dr. Francis Joanes, Port Arthur Clinic, Thunder Bay; Dr. David Morgan, Hamilton Health Sciences Center, Hamilton, Ontario; Dr. Marc Bradette, L'Hotel-Dieu de Quebec, Quebec; Dr. David Armstrong, McMaster University Medical Centre, Hamilton, Ontario; Dr. Sander Veldhuyzen van Zanten, Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia; Dr. Pierre Pare, Hospital St. Sacrement, Quebec, Quebec; Dr. W. Olsheski, Albany Medical Clinic, Toronto, Ontario; Dr. Ivor Teitelbaum, Yorkview Medical Centre, North York, Ontario; Dr. Subodh Kanani, Lakeshore West Medical Professional Centre, Toronto, Ontario; Dr. Paul Braude, Markham Research, Thornhill, Ontario. This work was supported by a grant from AstraZeneca Pharmaceuticals Inc, Mississauga, Canada and Lund, Sweden.
Figures and Tables
Table 1 Definition of symptom severity
Moderate problem: Cannot be ignored but does not influence my daily life.
Moderately severe problem: Cannot be ignored and occasionally limits my daily activities.
Severe problem: Cannot be ignored and often limits my concentration on daily activities.
Very severe problem: Cannot be ignored and markedly limits my daily activities and often requires rest.
Table 2 GERD-specific work productivity and activity impairment questionnaire (WPAI-GERD)
1. Are you currently employed?
2. During the past 7 days, how many days did you work?
3. How many of those days did you have reflux symptoms while working?
4. During the past 7 days, how many hours were you absent from work because of problems associated with your reflux symptoms?
5. During the past 7 days, how many hours were you absent from work for any other reason than problems associated with your reflux symptoms?
6. During the past 7 days, how many hours did you actually work?
7. During the past 7 days, how much did your reflux symptoms on average affect your normal productivity while you were working?
8. During the past 7 days, how much did your reflux symptoms on average affect your normal productivity while you were doing your regular activities, other than working?
Table 3 Baseline characteristics of patients with moderate or severe GERD mean age 50 years, range 20 to 82
Frequency (Percentage) (N 217)
Female 114 (52.5)
Single 33 (15.2)
Lives alone 23 (10.6)
Employed 153 (71.0)
Never smoked 94 (43.5)
Ever smoked 122 (56.5)
Caucasian 191 (88.0)
Other ethnic groups 26 (12.0)
Moderate GERD* 112 (51.6)
Moderately severe GERD* 74 (34.1)
Severe GERD* 27 (12.5)
Very severe GERD* 4 (1.8)
*GERD: Gastroesophageal reflux disease
Table 4 QOLRAD and four symptoms questionnaire baseline scores and mean change scores after 4 week treatment with PPI
Instrument Mean baseline score (95% CI) Mean change score (95% CI)
QOLRAD emotional Distress 4.5 (4.3, 4.7) (N = 217) 2.0 (1.9, 2.2) (N = 217)
QOLRAD sleep Disturbance 4.5 (4.3, 4.7) (N = 217) 2.1 (1.9, 2.3) (N = 217)
QOLRAD food/drink problems 3.8 (3.7, 4.0) (N = 217) 2.5 (2.3, 2.6) (N = 217)
QOLRAD physical/social 5.5 (5.3, 5.6) (N = 217) 1.3 (1.1, 1.4) (N = 217)
QOLRAD vitality 4.3 (4.1, 4.5) (N = 217) 2.1 (1.9, 2.3) (N = 217)
Stomach pain 3.9 (3.7, 4.1) (N = 217) -1.9 (-2.2, -1.7) (N = 217)
Heartburn 4.5 (4.3, 4.7) (N = 217) -2.9 (-3.1, -2.7) (N = 216)
Belching 3.6 (3.4, 3.8) (N = 216) -1.6 (-1.8, -1.4) (N = 216)
Acid reflux 4.1 (3.9, 4.3) (N = 215) -2.4 (-2.6, -2.2) (N = 212)
QOLRAD: Quality of life in reflux and dyspepsia, PPI: Proton pump inhibitor
Table 5 Correlation of the FT, SG and HUI 3 with the validation instruments (P-values)
Instrument FT SG HUI 3
SF-36 PCS (N = 212) 0.40 (<0.001) 0.31 (<0.001) 0.46 (<0.001)
SF-36 MCS (N = 212) 0.34 (<0.001) 0.22 (<0.001) 0.54 (<0.001)
QOLRAD emotional Distress 0.48 (<0.001) 0.27 (<0.001) 0.38 (<0.001)
QOLRAD sleep disturbance 0.39 (<0.001) 0.25 (<0.001) 0.23 (<0.001)
QOLRAD food/drink problem 0.37 (<0.001) 0.28 (<0.001) 0.25 (<0.001)
QOLRAD physical/social 0.38 (<0.001) 0.26 (<0.001) 0.41 (<0.001)
QOLRAD vitality 0.48 (<0.001) 0.30 (<0.001) 0.44 (<0.001)
Stomach pain 0.32 (<0.001) 0.15 (0.032) 0.26 (<0.001)
Heartburn 0.32 (<0.001) 0.13 (0.059) 0.09 (0.174)
Belching (N = 216) 0.20 (0.003) -0.09 (0.193) 0.09 (0.189)
Acid reflux (N = 215) 0.34 (<0.001) 0.15 (0.015) 0.14 (0.046)
FT: Feeling thermometer; SG: Standard gamble; HUI 3: Health utility index 3; SF-36: Short form 36; PCS: Physical component summary; MCS: Mental component summary; QOLRAD: Quality of life in reflux and dyspepsia. N = 217 unless indicated otherwise.
Table 6 SF-36 mean scores
Instrument Mean 95% CI
SF-36 physical functioning 46.6 (N = 213) 45.4, 47.8
SF-36 physical-Role 45.5 (N = 216) 44.0, 47.0
SF-36 bodily pain 42.8 (N = 217) 41.6, 44.1
SF-36 general health 46.2 (N = 216) 44.9, 47.5
SF-36 vitality 45.9 (N = 217) 44.6, 47.3
SF-36 social functioning 47.7 (N = 217) 46.3, 49.1
SF-36 role emotional 46.5 (N = 216) 44.9, 48.1
SF-36 mental health 46.9 (N = 217) 45.5, 48.3
SF-36: Short form 36
Table 7 Baseline mean (standard deviation) SF-36 physical component summary (PCS) and mental component summary (MCS) scores for patients with gastroesophageal reflux disease compared with data reported by Revicki et al. [35] and Hopman et al. [33]
SF-36 scale GERD (this study) Canadian population§ (N = 9,423) US Population† (N = 2,474) Clinical depression† (N = 502) Diabetes† (N = 541) Hypertension† (N = 2,089)
PCS 45.1 (n = 212) (8.7) 51.4* (8.5) 50.0* (10.0) 45.0 (12.1) 41.5* (11.3) 44.3 (10.8)
MCS 47.6 (n = 212) (11.0) 52.6* (8.5) 50.0* (10.0) 34.8* (12.2) 51.9* (9.6) 52.2* (9.3)
*P < 0.001 compared to our GERD patients.
§ Data from Hopman et al.
† Data from Revicki et al.
==== Refs
International Foundation for Functional Gastrointestinal Disorders Access date February 14, 2003
Louis E DeLooze D Deprez P Hiele M Urbain D Pelckmans P Deviere J Deltenre M Heartburn in Belgium: prevalence, impact on daily life, and utilization of medical resources Eur J Gastroenterol Hepatol 2002 14 279 284 11953693 10.1097/00042737-200203000-00012
Talley NJ Boyce P Jones M Identification of distinct upper and lower gastrointestinal symptom groupings in an urban population Gut 1998 42 690 695 9659166
Locke GR Talley NJ Fett SL Zinsmeister AR Melton LJ Prevalence and clinical spectrum of gastroesophageal reflux: a population-based study in Olmsted County, Minnesota Gastroenterology 1997 112 1448 1456 9136821
Frank L Kleinman L Ganoczy D McQuaid K Sloan S Eggleston A Tougas G Farup C Upper gastrointestinal symptoms in North America: prevalence and relationship to healthcare utilization and quality of life Dig Dis Sci 2000 45 809 818 10759254 10.1023/A:1005468332122
Wiklund I Talley NJ Update on health-related quality of life in patients with gastroesophageal reflux disease Expert Review of Pharmacoeconomics & Outcomes Research 2003 3 341 350 10.1586/14737167.3.3.341
Guyatt G Montori V Devereaux P Schünemann H Bhandari M Putting the patient first: In our practice, and in our use of language ACP Journal Club 2004 140 A11 14711297
Fass F Fennerty B Vakil N Nonerosive reflux disease-current concepts and dilemmas The Am J Gastroenterol 2001 96 303 314 11232668
Kartman B Utility and willingness to pay measurements among patients with gastroesophageal reflux disease Am J Gastroenterol 2001 96 S38 S43 11510769 10.1016/S0002-9270(01)02581-3
Ofman JJ Willingness to pay for improved outcomes in GERD: Response to Bernt Kartman Am J Gastroenterol 2001 96 S44 S45 11510770 10.1016/S0002-9270(01)02582-5
Guyatt GH Feeny DH Patrick DL Measuring health-related quality of life Ann Intern Med 1993 118 622 629 8452328
Kirshner B Guyatt G A methodological framework for assessing health indices J Chronic Dis 1985 38 27 36 3972947 10.1016/0021-9681(85)90005-0
Guyatt GH Veldhuyzen Van Zanten SJ Feeny DH Patrick DL Measuring quality of life in clinical trials: a taxonomy and review CMAJ 1989 140 1441 1448 2655856
Wiklund I Karlberg J Evaluation of quality of life in clinical trials. Selecting quality-of-life measures Control Clin Trials 1991 12 204S 216S 1663856
Fitzpatrick R Ziebland S Jenkinson C Mowat A Importance of sensitivity to change as a criterion for selecting health status measures Qual Health Care 1992 1 89 93 10136848
Torrance GW Feeny D Utilities and quality-adjusted life years Int J Technol Assess Health Care 1989 5 559 575 2634630
Boyle MH Torrance GW Sinclair JC Horwood SP Economic evaluation of neonatal intensive care of very-low-birth-weight infants N Engl J Med 1983 308 1330 1337 6405272
Bennet K Torrance GW Spilker B Measuring health state preferences and utilities: rating scale, time trade-off, and standard gamble techniques Quality of life and pharmacoeconomics in clinical trials 1996 Lippincott-Raven, Philadelphia 259
Feeny DH Torrance GW Incorporating utility-based quality-of-life assessment measures in clinical trials. Two examples Med Care 1989 27 S190 204 2522159
Green C Brazier J Deverill M Valuing health-related quality of life. A review of health state valuation techniques Pharmacoeconomics 2000 17 151 65 10947338
Bakker C Rutten M van Santen-Hoeufft M Bolwijn P van Doorslaer E Bennett K van der Linden S Patient utilities in fibromyalgia and the association with other outcome measures J Rheumatol 1995 22 1536 1543 7473480
Fries JF Ramey DR "Arthritis specific" global health analog scales assess "generic" health related quality-of-life in patients with rheumatoid arthritis J Rheumatol 1997 24 1697 1702 9292790
Juniper EF Guyatt GH Feeny DH Griffith LE Ferrie PJ Minimum skills required by children to complete health-related quality of life instruments for asthma: comparison of measurement properties Eur Respir J 1996 10 2285 2294 10.1183/09031936.97.10102285
Schünemann HJ Griffith L Stubbing D Goldstein R Guyatt GH A clinical trial to evaluate the measurement properties of 2 direct preference instruments administered with and without hypothetical marker states Med Decis Making 2003 23 140 149 12693876 10.1177/0272989X03251243
Furlong WJ Feeny DH Torrance GW Barr RD The Health Utilities Index (HUI) system for assessing health-related quality of life in clinical studies Ann Med 2001 33 375 384 11491197
Torrance GW Furlong W Feeny D Boyle M Multi-attribute preference functions. Health Utilities Index Pharmacoeconomics 1995 7 503 20 10155336
Schünemann HJ Armstrong D Degli'innocenti A Wiklund I Fallone C Tanser L Heels-Ansell D El-Dika S Austin P Barkun A Chiba N Guyatt G A randomized multi-Center trial to evaluate simple utility elicitation techniques in patients with gastroesophageal reflux disease Med Care 2004 42 1132 1142 15586841 10.1097/00005650-200411000-00013
Schünemann HJ Griffith L Jaeschke R Goldstein R Stubbing D Guyatt GH Evaluation of the minimal important difference for the feeling thermometer and the St. George's Respiratory Questionnaire in patients with chronic airflow obstruction J Clin Epidemiol 2003 56 1170 1176 14680667 10.1016/S0895-4356(03)00115-X
Kulich KR Wiklund I Junghard O Factor structure of the Quality of Life in Reflux and Dyspepsia (QOLRAD) questionnaire evaluated in patients with heartburn predominant reflux disease Qual Life Res 2003 12 699 708 14516179 10.1023/A:1025192100450
Talley NJ Fullerton S Junghard O Wiklund I Quality of life in patients with endoscopy-negative heartburn: reliability and sensitivity of disease-specific instruments Am J Gastroenterol 2001 96 1998 2004 11467624 10.1111/j.1572-0241.2001.03932.x
Wiklund IK Junghard O Grace E Talley NJ Kamm M Veldhuyzen van Zanten S Pare P Chiba N Leddin DS Bigard MA Colin R Schoenfeld P Quality of Life in Reflux and Dyspepsia patients. Psychometric documentation of a new disease-specific questionnaire (QOLRAD) Eur J Surg Suppl 1998 41 9 10027672
Wahlqvist P Carlsson J Stalhammar NO Wiklund I Validity of a Work Productivity and Activity Impairment questionnaire for patients with symptoms of gastro-esophageal reflux disease (WPAI-GERD) – results from a cross-sectional study Value Health 2002 5 106 113 11918826 10.1046/j.1524-4733.2002.52101.x
Hopman W Towheed T Anastassiades T Tenenhouse A Poliquin S Berger C Joseph L Brown J Murray T Adachi J Hanley D Papadimitropoulos E Canadian normative data for the SF-36 health survey CMAJ 2000 163 265 271 10951722
Ware JE JrSherbourne CD The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection Med Care 1992 30 473 483 1593914
Revicki DA Wood M Maton PN Sorensen S The impact of gastroesophageal reflux disease on health-related quality of life Am J Med 1998 104 252 258 9552088 10.1016/S0002-9343(97)00354-9
Stewart AL Greenfield S Hays RD Wells K Rogers WH Berry SD McGlynn EA Ware JE Jr Functional status and well-being of patients with chronic conditions. Results from the Medical Outcomes Study JAMA 1989 262 907 913 2754790 10.1001/jama.262.7.907
Stewart AL Hays RD Ware JE Jr The MOS short-form general health survey. Reliability and validity in a patient population Med Care 1988 26 724 735 3393032
Morimoto T Fukui T Utilities measured by rating scale, time trade-off, and standard gamble: review and reference for health care professionals J Epidemiol 2002 12 160 178 12033527
Post PN Stiggelbout AN Wakker PP The utility of health states after stroke: A systematic review of the literature Stroke 2001 32 1425 1429 11387509
Ware JE JrKosinski M Bayliss MS McHorney CA Rogers WH Raczek A Comparison of methods for the scoring and statistical analysis of SF-36 health profile and summary measures: summary of results from the Medical Outcomes Study Med Care 1995 33 AS264 279 7723455
Borchardt PJ Employee productivity and gastroesophageal reflux disease: the payer's viewpoint Am J Gastroenterol 2001 96 S62 63 11510774 10.1016/S0002-9270(01)02585-0
|
16004616
|
PMC1183201
|
CC BY
|
2021-01-04 16:03:27
|
no
|
BMC Gastroenterol. 2005 Jul 10; 5:23
|
utf-8
|
BMC Gastroenterol
| 2,005 |
10.1186/1471-230X-5-23
|
oa_comm
|
==== Front
BMC GenetBMC Genetics1471-2156BioMed Central London 1471-2156-6-341596085310.1186/1471-2156-6-34Research ArticlePolymorphisms within the canine MLPH gene are associated with dilute coat color in dogs Philipp Ute [email protected] Henning [email protected] Lars [email protected] Seiji [email protected] Emmanuel [email protected]ünzel-Apel Anne-Rose [email protected] Sheila M [email protected] Tosso [email protected] Institute of Animal Breeding and Genetics, University of Veterinary Medicine Hannover, Bünteweg 17p, 30559 Hannover, Germany2 Department of Pathobiology, College of Veterinary Medicine, Texas A&M University, College Station, TX 77843-4467, USA3 Center of Narcolepsy Department of Psychiatry Stanford University School of Medicine, 701 Welch road B, Palo Alto CA 94304-5742, USA4 Institute for Reproductive Medicine, University of Veterinary Medicine Hannover, Bünteweg 15, 30559 Hannover, Germany5 Department of Animal and Poultry Science, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N 5A82005 16 6 2005 6 34 34 23 12 2004 16 6 2005 Copyright © 2005 Philipp et al; licensee BioMed Central Ltd.2005Philipp 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
Pinschers and other dogs with coat color dilution show a characteristic pigmentation phenotype. The fur colors are a lighter shade, e.g. silvery grey (blue) instead of black and a sandy color (Isabella fawn) instead of red or brown. In some dogs the coat color dilution is sometimes accompanied by hair loss and recurrent skin inflammation, the so called color dilution alopecia (CDA) or black hair follicular dysplasia (BHFD). In humans and mice a comparable pigmentation phenotype without any documented hair loss is caused by mutations within the melanophilin gene (MLPH).
Results
We sequenced the canine MLPH gene and performed a mutation analysis of the MLPH exons in 6 Doberman Pinschers and 5 German Pinschers. A total of 48 sequence variations was identified within and between the breeds. Three families of dogs showed co-segregation for at least one polymorphism in an MLPH exon and the dilute phenotype. No single polymorphism was identified in the coding sequences or at splice sites that is likely to be causative for the dilute phenotype of all dogs examined. In 18 German Pinschers a mutation in exon 7 (R199H) was consistently associated with the dilute phenotype. However, as this mutation was present in homozygous state in four dogs of other breeds with wildtype pigmentation, it seems unlikely that this mutation is truly causative for coat color dilution. In Doberman Pinschers as well as in Large Munsterlanders with BHFD, a set of single nucleotide polymorphisms (SNPs) around exon 2 was identified that show a highly significant association to the dilute phenotype.
Conclusion
This study provides evidence that coat color dilution is caused by one or more mutations within or near the MLPH gene in several dog breeds. The data on polymorphisms that are strongly associated with the dilute phenotype will allow the genetic testing of Pinschers to facilitate the breeding of dogs with defined coat colors and to select against Large Munsterlanders carrying BHFD.
==== Body
Background
Coat color dilution leads to the so-called blue pigmentation phenotype in black-and-tan Pinschers (Doberman Pinschers, German Pinschers, Miniature Pinschers), characterized by a silver-blue shade of the black fur areas (Figure 1). Similarly, coat color dilution is responsible for the Isabella fawn phenotype in brown-and-tan or tan Pinschers. Color dilution in Pinschers is inherited as a Mendelian autosomal recessive trait. Although there are no severe impairments known, this pigmentation variation is of clinical relevance as Pinschers with coat color dilution show an increased prevalence of color dilution alopecia (CDA) also called Blue Doberman syndrome. CDA is characterized by a progressive loss of hair, which is sometimes accompanied by recurrent bacterial infections of the hair follicles (folliculitis). Melanosome clumping occurs within melanocytes of the epidermis and hair follicles, resulting in macromelanosomes in hair shafts that subsequently fracture when emerging from the skin. The exposed skin of CDA affected dogs is often dry and scaly as well as sensitive to sunburn or extreme cold [1]. Black hair follicular dysplasia (BHFD), a form of alopecia in various breeds where only the black coat areas are affected, is phenotypically similar to CDA [2-4].
Figure 1 Blue Doberman Pinscher und black-and-tan Doberman Pinscher. Blue Doberman Pinscher (A) and black-and-tan Doberman Pinscher (B). Note the coat color differences between the two animals. The black and reddish fur parts of the black-and-tan Doberman Pinscher are changed to paler coloring in the blue dog. Classical genetics states that the blue dog is homozygous for the recessive dilute allele (d).
In human and mouse, genes are already known which lead to phenotypically similar coat color variations. In mice the mutants dilute, ashen and leaden are well characterized [5-7]. These mutants correspond to the human Griscelli syndromes (GS) 1 to 3 [8-10]. Griscelli syndromes 1–3 as well as the above mentioned mouse mutants are all inherited as Mendelian autosomal recessive traits.
Mutations in three different genes, i.e. myosin Va (MYO5A), RAB27A, and melanophilin (MLPH), are responsible for these phenotypes. The proteins which are encoded by these genes are part of the melanosome transport complex. Therefore, in the melanocytes of affected individuals an accumulation of melanosomes around the nucleus is observed as well as large clumps of pigments in the hair shaft.
In the dilute mouse mutant the Myo5a gene is mutated [5], while mutations in the Rab27A gene lead to the ashen mouse mutant [6]. The phenotypes of these mutants are close to their human counterparts of GS1 or GS2 affected patients, respectively. Individuals carrying a mutation in one of these two genes usually develop severe neurological (GS1) or rather immunological disorders (GS2) in addition to their skin and hair color dilution [8,9]. One human case report describes that the deletion of the MYO5A gene exon F, which is only expressed in melanocytes, leads to hypopigmentation without further disorders [10]. In contrast to MYO5A and RAB27A mutations, which normally cause complex phenotypes, mutations in the MLPH gene are responsible for color dilution without any further impairment in human GS3 patients or leaden mice [7,10]. Therefore the MLPH gene seemed to be the most suitable candidate gene for coat color dilution in dogs and we report here the analysis of this gene in several dog breeds with an emphasis on Doberman Pinschers and German Pinschers.
Results
Characterization of the canine MLPH gene
A human MLPH cDNA probe was used to retrieve a canine genomic clone (RP81-203J24) from a Doberman Pinscher BAC library. A draft sequence of this 198 kb BAC clone was determined. In order to finish this draft sequence additional public whole genome shotgun sequences from a Boxer were used. The BAC clone contained the complete collagen type VI alpha 3 gene (COL6A3) as well as the exons 1 to 10 of the 16 exon MLPH gene. To obtain the missing 3'-end of the canine MLPH gene, Boxer whole genome shotgun sequences were assembled and joined to the sequence of the BAC clone resulting in one large contiguous sequence of 212,696 bp. Comparison of these sequences revealed a number of polymorphisms between Doberman Pinscher and Boxer DNA.
The canine MLPH gene spans approximately 48 kb of genomic sequence compared to 67 kb for the human MLPH gene. The genomic organization of the canine MLPH gene was inferred by comparison of the genomic dog sequence with an experimentally derived canine cDNA sequence (Figure 2). The genomic structure of the MLPH gene was not entirely conserved between human and dog. All but one of the 16 human MLPH exons could be identified in the canine MLPH sequence. No dog exon homologous to the human exon 9 could be identified; however, this exon is not used constitutively in all human transcripts. On the other hand, the canine MLPH gene contains a fifth exon of 39 bp that is not present in the human or murine MLPH genes. The MLPH gene has a very high GC-content of about 59.5%, which is significantly above the mammalian average of 41%. Consistent with the high GC-content a CpG island is located upstream of exon 1 in the dog sequence in addition to numerous CpG islands within the gene. A canonical polyadenylation signal AATAAA was identified approximately 3.1 kb downstream of the stop codon but 3'-RACE experiments indicated that in dog polyadenylation occurs only ~ 300 bp downstream of the stop codon following a sequence motif ATTGAA that weakly resembles the canonical polyadenylation signal.
Figure 2 Architecture of the canine MLPH gene. (A) The 48 polymorphisms that were identified in Doberman Pinschers and/or German Pinschers are indicated. PCR products spanning each of the MLPH exons with adjacent flanking sequences were sequenced. The eight SNPs around exon 2 show strong association with the dilute allele in Doberman Pinschers, however they are monomorphic in German Pinschers. The R199H mutation shows strong association with the dilute allele in German Pinschers and in some Doberman Pinschers of European origin. (B) Genomic organization of the canine MLPH gene. Exons are denoted as boxes. Solid boxes represent coding sequence while open boxes contain the untranslated regions. The genomic section corresponds to a 50 kb interval in the analyzed sequence of 212.696 bp (positions 156.001 – 206.000). (C) Illustration of the unusually high GC-content of the canine MLPH gene. The GC-content was calculated using a 300 bp window. CpG island criteria were: GC > 0.5, CpGobs/CpGexp > 0.6, and length > 200 bp.
The canine MLPH mRNA contains an open reading frame of 1746 nt encoding a protein of 581 amino acids. The canine MLPH protein was predicted to have a molecular weight of 62.7 kDa, a pI of 5.7, and shows 62% identity to the orthologous human protein (human MLPH isoform lacking the amino acids encoded by exon 9).
Mutation analysis of the canine MLPH gene and association with dilute phenotype
Comparative sequencing of the exons and adjacent sequences of 6 animals from one Doberman Pinscher and 5 from one German Pinscher family revealed 43 sequence differences within these closely related breeds and an additional five variations between the breeds (Table 1). Within the Doberman Pinscher family 39 polymorphisms were observed while only 7 sequence variations were found in the German Pinscher family members. Only 3 variations segregated in both families and none of them was in the coding sequence.
Table 1 Polymorphisms within the canine MLPH gene
position1 cDNA position2 Doberman Pinscher German Pinscher
157354 (exon 1) -139 (5'-UTR) G/T G
157471 (exon 1) -22 (5'-UTR) A/G A/G3
157486 (intron 1) A/C3 A
163882 (intron 1) C/T T
163889 (intron 1) A/C C
163936 (intron 1) A/G A
163983 (intron 1) A/G A
164012 (intron 1) A/G A
164049 (intron 1) A/G A
164075 (intron 1) C/T C
164258 (exon 2) +106 C/T (silent) C
164397 (intron 2) A/G A
180908 (exon 3) +223 A/G (75M/75V) A
180994 (exon 3) +309 C/T (silent) C
180998 (exon 3) +313 A/G (105N/105D) A
183047 (intron 4) C/G G
183254 (exon 5) +461 C/T (154T/154M) C/T3 (154T/154M)
183336 (intron 5) C/T C/T3
184650 (intron 5) C/T T
184751 (exon 6) +553 A/G (185S/185G) G
184818 (intron 6) A/G G
186072 (intron 6) C T
186091 (intron 6) C T
186184 (exon 7) +596 A/G3 (199H/199R) A/G (199H/199R)
186329 (intron 7) A G
187449..54 (ex. 8)4 +775 – +780 indel GAGGAT +/- (indel 259E260D) indel GAGGAG +/-3 (indel 259E260E)
189560 (intron 8) A/G G
189575 (intron 8) A A/C
189719 (exon 9) +1032 C/T (silent) C
189728 (exon 9) +1041 G A/G (silent)
195548 (intron 9) A/G G
195720 (exon 10) +1148 G (383G) A (383D)
195808 (exon 10) +1236 A (silent) G (silent)
195888 (intron 10) A/G G
195891 (intron 10) C/G G
195900 (intron 10) A/G A/G
195901 (intron 10) indel G +/- indel G -
195932 (intron 10) A/G G
196097 (exon 11) +1263 A/G (silent) G
196354 (intron 11) C/T C
197343 (intron 11) C/T T
197510 (intron 11) C/T C
197673 (intron 12) C/T C
197730 (intron 12) A/G G
198080 (intron 12) A A/G
198126 (intron 12) A/G A/G
199847 (intron 13) A/G A
202837 (exon 16) +1801 (3'-UTR) C/T C/T
1 Positions refer to the entire determined genomic sequence of 212.696 bp [EMBL:BN000728].
2 +1 corresponds to the adenosine of the translation start codon in the MLPH cDNA.
3 These polymorphisms were not present in the 11 samples of the initial mutation analysis. They were identified in additionally sequenced Pinscher samples.
4 Note that there are three alleles at this site (GAGGAG, GAGGAT, : : : : : :).
Most polymorphisms were SNPs (46), only two indel polymorphisms were observed. Of the 48 observed polymorphisms 33 were located in introns. The remaining 15 polymorphisms were in exons, of which 7 led to amino acid exchanges in the MLPH protein (Figure 3).
Figure 3 Alignment of MLPH proteins from different species. The MLPH protein sequences were translated from nucleotide database accessions [EMBL:AJ920333] (dog), [Genbank:AK022207] (human), [Genbank:AF384098] (mouse), and [Genbank:BC081894] (rat), respectively. The three major predicted protein domains of MLPH are indicated in accordance with [24]. Note the 13 additional amino acids in the dog MLPH protein encoded by dog exon 5, which is not conserved in other species. Another big difference between the sequences is caused by the fact that dog is lacking a homologous exon to human exon 9. In human this exon is not used constitutively and for the alignment a protein isoform without the amino acids encoded by this alternative exon was used. Polymorphisms that affect the amino acid sequence of the dog MLPH protein are indicated with arrows. None of the observed protein polymorphisms has a segregation pattern in the investigated families that would be compatible with a causative mutation for dilute.
Although an indel in exon 8 was the most striking variation observed, dogs of normal color and homozygous for each of the possible variants were encountered at this polymorphism. A G→ A transition in exon 7 causing an exchange from arginine to histidine at position 199 of the MLPH protein showed perfect co-segregation with the dilute and wildtype phenotypes in the German Pinscher family (Figure 4). We therefore established a HhaI RFLP assay (Figure 5) and analyzed this mutation in 341 dogs. The histidine variant was homozygous in all dilute German Pinschers (18), Beagles (2), and Large Munsterlanders with BHFD (4). Although this mutation showed strong association with the dilute phenotype in German Pinschers, it must be noted, that we observed four dogs with wildtype color from other breeds (one Large Munsterlander and three Doberman Pinschers) that were also homozygous for the 199H allele. Therefore, the R199H mutation is a tightly linked marker for the d allele in German Pinschers but it seems unlikely that it represents a loss-of-function mutation that could cause dilute coat color. The allele distribution of the R199H mutation in Doberman Pinschers turned out to be very interesting. In samples collected form Doberman Pinschers in North America the 199H allele was very rare and not obviously associated with the dilute phenotype. However, in samples from European Doberman Pinschers we observed a strong albeit not perfect association of the 199H allele with the dilute phenotype (Table 2). The available genotyping data for the amino acid changing mutations are summarized in Table 3. These data show that homozygous animals with wildtype color exist for every single amino acid replacement that we found in dilute animals.
Figure 4 Selected families and MLPH genotyping data. (A) Doberman Pinscher family of American origin that was used in the initial mutation analysis. Dilute animals (dd) are indicated as solid black symbols. Animals 1 and 2 are obligate heterozygotes for dilute as they were black-and-tan with blue offspring. Animals 4 and 6 were classified DD and Dd based on their MLPH exon 2 genotypes. Genotypes for the seven polymorphic amino acid positions in the MLPH protein and the silent C/T SNP in exon 2 of the MLPH gene are shown. Three different marker haplotypes are color-coded. (B) Animals 8–12 of the depicted German Pinscher family were used for the initial mutation analysis. Animals 7 and 8 are obligate heterozygotes for dilute as they were black-and-tan with blue offspring. Animals 10 and 12 were classified as Dd based on their genotypes with respect to the R199H mutation. (C) Large Munsterlander family used in this study. Black symbols indicate dogs with dilute coat color and BHFD. Animals 14 and 15 are obligate heterozygotes for dilute and BHFD as they were normal with BHFD offspring. Classification of the animals 13 and 16–20, respectively, was done based on their MLPH exon 2 genotypes.
Figure 5 Genotyping of the R199H mutation. (A) Schematic diagram of the HhaI RFLP used for genotyping the R199H mutation. (B) Genotyping of the R199H mutation in Doberman Pinschers and German Pinschers. Numbers of the animals correspond to the numbers in Fig. 3A and 3B. Note that the Doberman Pinschers are homozygous for the presumed wildtype allele (199R) while in the studied German Pinscher family the R199H mutation cosegregates with the d allele.
Table 2 Genotype frequency of two MLPH polymorphisms in different breeds
Breed and phenotype No. of animals dilute genotype1 Exon 22 Exon 7
CC3 CT TT 199R199R 199R199H 199H199H
Doberman Pinscher (all) 140 50 (36%) 69 (49%) 21 (15%) 82 (59%) 47 (34%) 11 (8%)
wildtype coat color 98 D. 50 (51%) 48 (49%) - 62 (63%) 33 (34%) 3 (3%)
wildtype coat color 21 Dd - 21 (100%) - 9 (43%) 12 (57%) -
dilute coat color 21 dd - - 21 (100%) 11 (52%) 2 (10%) 8 (38%)
Doberman Pinscher (American origin) 38 6 (16%) 21 (55%) 11 (29%) 35 (92%) 3 (8%) -
wildtype coat color 19 D. 6 (32%) 13 (68%) - 16 (84%) 3 (16%) -
wildtype coat color 8 Dd - 8 (100%) - 8 (100%) - -
dilute coat color 11 dd - - 11 (100%) 11 (100%) - -
Doberman Pinscher (European origin) 102 44 (43%) 48 (47%) 10 (10%)
wildtype coat color 79 D. 44 (56%) 35 (44%) - 46 (58%) 30 (38%) 3 (4%)
wildtype coat color 13 Dd - 13 (100%) - 1 (8%) 12 (92%) -
dilute coat color 10 dd - - 10 (100%) - 2 (20%) 8 (80%)
German Pinscher 143 143 (100%) - - 64 (45%) 61 (43%) 18 (13%)
wildtype coat color 117 D. 117 (100%) - - 64 (55%) 53 (45%) -
wildtype coat color 8 Dd 8 (100%) - - - 8 (100%) -
dilute coat color 18 dd 18 (100%) - - - - 18 (100%)
Beagle 6 1 3 2 1 3 2
wildtype coat color 2 D. 1 1 1 1
wildtype coat color 2 Dd 2 2
dilute coat color 2 dd 2 2
Large Munsterlander 12 - 8 4 - 7 5
wildtype coat color 8 Dd - 8 - - 7 1
dilute coat color & BHFD 4 dd - - 4 - - 4
Weimeraner (dilute coat color) 1 dd 1 1
Am. Staffordshire (dilute col. & CDA) 1 dd 1 1
Mountain Dogs 26 D. 25 1 25 1
Breeds unspecified 12 D. 10 2 9 3
total 341 229 (67%) 85 (25%) 27 (8%) 181 (53%) 124 (36%) 36 (11%)
1 A dog with wildtype coat color has the dilute genotype DD or Dd. If the animal has never produced any dilute (dd) offspring, the genotype can not be deduced unambiguously from the phenotype and the genotype is then denoted "D.". Among the Pinschers with wildtype coat colors there were also obligate carriers of the dilute allele (Dd), as these animals had dilute (dd) progeny.
2 This polymorphism corresponds to the silent C/T SNP in exon 2 at position 164258 in the reference sequence [EMBL:BN000728].
3 The following associations were highly significant in Pinscher breeds under Fisher's Exact Test (p < 0.001): Exon 2 for all Doberman Pinschers, American Doberman Pinschers, and European Doberman Pinschers; exon 7 for European Doberman Pinschers and for German Pinschers. Non-Pinscher breeds were not evaluated statistically because of their small sample numbers.
Table 3 Genotype data of amino acid changing polymorphisms
Large Munsterlander Doberman Pinschers German Pinschers
Polymorphism wild type dilute wild type dilute wild type dilute
Exon 3 (V75M)
VV - - 2 5 - -
VM - - 11 - 4 -
MM - - 4 - 21 9
Exon 3 (D105N)
DD - - 2 5 - -
DN - - 11 - 4 -
NN - - 4 - 21 9
Exon 5 (M154T)
MM - - 5 11 2 -
MT - - 22 2 7 -
TT 6 4 34 6 45 14
Exon 6 (G185S)
GG - - 6 7 17 9
GS - - 5 - - -
SS - - - - - -
Exon 7 (R199H)
RR - - 71 11 64 -
RH 7 - 45 2 61 -
HH 1 4 3 8 - 18
Exon 8 (259–260)
del/del - - 1 2 - -
del/ED - - 3 - - -
del/EE - - - - 1 -
ED/ED - - - - - -
ED/EE - - - - - -
EE/EE 7 4 - - 3 2
Exon 10 (D383G)
DD - - - - 3 2
DG - - - - - -
GG - - 4 2 - -
A set of eight SNPs around exon 2 showed perfect association with the dilute allele in all 140 Doberman Pinschers that were analyzed. From the available genotyping data four haplotypes could be reconstructed (Table 4). A single haplotype (termed haplotype 2) was associated with the d allele in all Doberman Pinschers as well as in one Beagle family and the Large Munsterlander family segregating for BHFD. However, all German Pinschers under investigation were monomorphic around exon 2 and had the common haplotype 3. A PCR-RFLP assay was developed for the silent C/T SNP in exon 2 (position +106 in the MLPH cDNA sequence) since the presence of a T at this position was unique to haplotype 2.
Table 4 Haplotype frequencies of the eight SNPs around the MLPH exon 2
Wildtype color, 305 animals Dilute color, 45 animals
Haplotype Alleles1 N % N %
Haplotype 1 AAAAACCG 34 5.6 - -
Haplotype 2 AGGGGTTG 80 13.1 522 57.8
Haplotype 3 CAAAACCA 493 80.8 383 42.2
Haplotype 4 AAAAGCCG 3 0.5 - -
1 The alleles correspond to the eight polymorphic positions between 163889 and 164397 in the genomic reference sequence.
2 Each of the 22 analyzed dilute Doberman Pinschers was homozygous for haplotype 2.
3 Each of the 18 analyzed dilute German Pinschers was homozygous for haplotype 3.
While polymorphisms at the 5'-end of MLPH are tightly associated with dilute in American Doberman Pinschers and polymorphisms at the 3'-end are tightly associated with dilute in German Pinschers, a large group of dogs including European Doberman Pinschers, Large Munsterlanders and Beagles show strong association of dilute with markers across the entire MLPH gene. Detailed inspection of the available dilute chromosomes across different breeds revealed that all dilute chromosomes belonged to three different MLPH marker haplotypes. Each of the three families shown in Figure 4 carries one of these dilute haplotypes. A detailed comparison of the three dilute haplotypes is given in Table 5. The three different dilute haplotypes do not share extended haplotype blocks within the coding region of the MLPH gene. However, they do share the first three marker alleles from the region around exon 1. Thus it is possible that a single ancestral founder mutation within the promoter of the MLPH gene followed by subsequent recombinations is responsible for the observed diversity of dilute haplotypes.
Table 5 dilute haplotypes within the canine MLPH gene
position1 cDNA position2 Doberman Pinscher/ Large Munsterlander/ Beagle Doberman Pinscher (American origin) German Pinscher
157354 (exon 1) -139 (5'-UTR) G G G
157471 (exon 1) -22 (5'-UTR) A A A
157486 (intron 1) A A A
163882 (intron 1) C C T
163889 (intron 1) A A C
163936 (intron 1) G G A
163983 (intron 1) G G A
164012 (intron 1) G G A
164049 (intron 1) G G A
164075 (intron 1) T T C
164258 (exon 2) +106 T T C
164397 (intron 2) G G A
180908 (exon 3) +223 G G A
180994 (exon 3) +309 C T C
180998 (exon 3) +313 G G A
183047 (intron 4) C G G
183254 (exon 5) +461 C T C
183336 (intron 5) C T C
184650 (intron 5) C T C
184751 (exon 6) +553 G G G
184818 (intron 6) G A G
186072 (intron 6) T C T
186091 (intron 6) T C T
186184 (exon 7) +596 A G A
186329 (intron 7) G A G
187449..54 (ex. 8)4 +775 – +780 GAGGAG del GAGGAG
189560 (intron 8) G G G
189575 (intron 8) C A C
189719 (exon 9) +1032 C C C
189728 (exon 9) +1041 n.d. G A
195548 (intron 9) n.d. G G
195720 (exon 10) +1148 n.d. G A
195808 (exon 10) +1236 n.d. A G
195888 (intron 10) n.d. G G
195891 (intron 10) n.d. G G
195900 (intron 10) n.d. G G
195901 (intron 10) n.d. G del
195932 (intron 10) n.d. G G
196097 (exon 11) +1263 n.d. G G
196354 (intron 11) n.d. T C
197343 (intron 11) n.d. T T
197510 (intron 11) n.d. C C
197673 (intron 12) n.d. C C
197730 (intron 12) n.d. G G
198080 (intron 12) n.d. A n.d.
198126 (intron 12) n.d. A n.d.
199847 (intron 13) n.d. A A
202837 (exon 16) +1801 (3'-UTR) n.d. T C
1 Positions refer to the entire determined genomic sequence of 212.696 bp [EMBL:BN000728].
2 +1 corresponds to the adenosine of the translation start codon in the MLPH cDNA.
In order to rule out potential splicing aberrations we isolated skin RNA from a heterozygous Large Munsterlander (#15 in Figure 4), a dilute Beagle and a Beagle with wildtype color. We amplified the coding part of the MLPH cDNA by RT-PCR. Agarose gel electrophoresis gave no evidence for splicing aberrations or transcriptional silencing because the bands of the normal and dilute dogs were of the same sizes and comparable intensities. Sequencing of the RT-PCR products confirmed the MLPH polymorphisms previously obtained by comparative sequencing of genomic PCR products.
Discussion
Pinschers affected by coat color dilution have a phenotype comparable to the leaden mouse mutant (Mlphln). Therefore analyzing the canine ortholog of the Mlph gene causing this mutant in mice seemed a logical approach to elucidate the molecular basis for coat color dilution in dogs. The assignment of the canine MLPH gene to CF25q24 is in accordance with the location of the human and murine orthologous genes and with the known synteny data of the integrated canine map [11,12]. The orientation of the MLPH and COL6A3 genes to each other is also consistent with their orientation on the human map. The genomic structure of the MLPH gene is similar but not identical in dog, human, and mouse. Differences were observed with respect to the dog exon 5, which is lacking from other species and the human/mouse/rat exon 9 that could not be identified within the genomic dog sequence by sequence comparisons. All the experimental canine cDNA sequences obtained in this study lacked a corresponding sequence. As there are known splice variants in human lacking exon 9 (e.g. accession AK022207) it might be possible that this alternative exon is not conserved in the canine gene. An alternative explanation would be that the homology between the human and canine exon 9 is very low, so that it can not be identified by cross-species sequence comparison.
The Pinscher breeds are considered as closely related and sometimes Doberman Pinschers are still interbred with German Pinschers in order to modulate the size of the animals. In support of this, we observed 48 sequence polymorphisms within and between the two related Pinscher breeds, of which only five variations seemed to be breed specific. Taking into account the limited number of animals used in the mutation analysis it is quite likely that there are even less or no breed-specific polymorphisms at all in these breeds. Generally, German Pinscher sequences showed less variation than those of Doberman Pinschers. This had to be expected because the German Pinscher breed experienced a severe bottleneck after the second world war (7 founders in Germany, personal communication by breeders).
We identified a set of eight SNPs including a silent C to T change in exon 2, which are in linkage disequilibrium with the dilute phenotype in some breeds. In Doberman Pinschers, Large Munsterlanders, and in Beagles one haplotype co-segregated with the dilute phenotype.
The R199H mutation is in linkage disequilibrium with the dilute phenotype in German Pinschers. The R199H mutation also showed perfect association with dilute in the Beagle family and was strongly associated with the d allele in Doberman Pinschers from Europe but not from North America.
A Large Munsterlander family with pups affected with BHFD was included in this study. The phenotype of the BHFD affected animals is very similar to CDA affected Pinschers [2,3]. Histological analysis of skin biopsies of BHFD affected dogs showed the typical perinuclear clumping of melanosomes within melanocytes of the hair matrix, which is also observed in leaden mice and human GS3 patients. Since the same haplotype as in the dilute Doberman Pinschers cosegregated with BHFD in the Large Munsterlander family, this result supports the idea that CDA and BHFD are indeed the same disorder.
The data clearly imply that mutations in or near the MLPH gene are causing dilute coat color in dogs. The fact that the observed linkage disequilibrium between marker alleles and dilute is strongest around exon 2 in Doberman Pinschers and around exon 7 in German Pinschers suggests that there may be different mutations causing coat color dilution in dogs.
The newly identified polymorphisms in exon 2 of the MLPH gene should be suitable DNA markers for coat color dilution in Doberman Pinschers and for the BHFD allele in Large Munsterlanders. In German Pinschers the exon 7 polymorphisms can be used as a diagnostic test for the dilute allele. For Beagles a larger sample should be analyzed to confirm whether these polymorphisms are appropriate DNA markers for the coat color dilution in Beagles as well.
The data clearly imply that MLPH is the causative gene for dilute coat color in several dog breeds. However, the causal mutation has not yet been conclusively identified. The data are compatible with two alternative hypotheses: Dilute coat color could be caused by a single founder mutation in all investigated dog breeds. Under this scenario the observed haplotypes suggest a location of the causal mutation within the MLPH gene upstream of exon 2. The alternative hypothesis, which can not be ruled out at this time, would be that different independent MLPH mutations cause coat color dilution in dogs.
Although no single polymorphism affecting amino acids was associated with all dilute phenotypes, it is possible that multiple mutations in this gene or its promoter region are responsible as is true for the brown phenotype in dogs [13]. At this time a functional significance of a synonymous mutation, such as the C/T change in exon 2, can not be completely excluded as it has been reported that such synonymous polymorphisms may influence mRNA folding and stability thereby mediating functional effects. Such mutations can either act in isolation or in combination with other mutations in the same transcript [14]. All dogs with the TT genotype were of dilute coat color, even though the inverse was not true.
In order to fully explore the possibility of different functional MLPH mutations, mRNA from dogs of dilute phenotype of several breeds such as Great Dane, Newfoundland, Shar-Pei, Beagle, Doberman Pinscher and Large Munsterlander are being collected. If multiple mutations occurred or if the same mutation occurred at several points in time, the haplotypes would not be consistent in all individuals with dilute coat color. This trait is under selection in some breeds in which dilute coat color has minimal to no health associated problems and is under strong negative selection in other breeds such as Large Munsterlanders where the trait is typically associated with severe hair loss. This large variation in pleiotropic effects also suggests that multiple mutations may be involved.
In the mouse an independent gene termed suppressor of dilute (Dsu) is known that is able to suppress the effects of Mlph mutations. Mice carrying loss of function alleles at the Mlph and the Dsu loci have a coat color closely resembling the wildtype coat color [15]. So far, no equivalent DSU mutations have been reported in the dog. However, it seems possible that unrecognized DSU mutations might confound our analysis, which is based on the assumption of a strictly monogenic autosomal recessive inheritance of coat color dilution in dogs.
Conclusion
We characterized the canine MLPH gene and identified 48 polymorphisms of this gene that occur in Doberman Pinschers and/or German Pinschers. Eight of these 48 polymorphisms located around exon 2 are in strong linkage disequilibrium with coat color dilution in Doberman Pinschers. A R199H mutation is in strong linkage disequilibrium with the dilute phenotype in German Pinschers. These results indicate that mutations in or near the MLPH gene are responsible for the coat color dilution in Pinschers. The reported polymorphisms will allow genetic testing of Doberman and German Pinschers to facilitate the breeding of dogs with specific coat colors.
Methods
Cloning and sequencing the MLPH gene
For the isolation of a canine BAC clone with the MLPH gene the Doberman Pinscher RPCI-81 BAC library [16] was screened with a 32P-labeled PCR fragment derived from the conserved 5'-end of the human MLPH cDNA. The donor animal for the RPCI-81 BAC library was a black-and-tan Doberman Pinscher named Grumpy. Grumpy was heterozygous at the dilute locus (Dd) as he had blue offspring. The probe sequence was amplified from the IMAGE cDNA clone IRAKp961H0816 provided by the Resource Center/Primary Database of the German Human Genome Project [17]. Hybridization was performed according the RPCI protocols [18]. End sequences of the positive BAC clone RP81-203J24 were determined on a LI-COR 4200L-2 automated sequencer (MWG, Biotech, Ebersberg, Germany) using the Thermo Sequenase Primer Cycle Sequencing Kit (Amersham Biosciences, Freiburg, Germany) and comparatively mapped to the human genome. Additionally, the size of the BAC clone was determined by pulsed field gel electrophoresis (PFGE).
For sequencing the BAC insert plasmid subclones were produced using the TOPO Shotgun Cloning Kit (Invitrogen, Karlsruhe, Germany). Plasmid DNA was isolated with the Montage Plasmid Miniprep96 Kit (Millipore, Eschborn, Germany). Sequencing was done on a MegaBACE capillary sequencing machine (Amersham Biosciences, Freiburg, Germany) using the Dyenamic™ Terminator Cycle Sequencing Kit (Amersham Biosciences, Freiburg, Germany) or on a LI-COR 4200L-2 automated sequencer using the Thermo Sequenase Primer Cycle Sequencing Kit. Shotgun sequences were collected until eight-fold coverage of the BAC clone was achieved. The sequences were assembled with Sequencher 4.2 (GeneCodes, Ann Arbor, MI, USA). Whole genome shotgun sequences from a Boxer were retrieved from the trace archive to fill gaps in the BAC clone sequence as well as for assembling the 3'-end of the canine MLPH gene that was not contained on the BAC clone [19]. Our experimental genomic sequences were deposited under accession [EMBL:AJ920047] in the EMBL database. The entire 212.696 bp contig consisting of an assembly from our experimental sequence reads as well as public WGS reads was also deposited in the EMBL database [EMBL:BN000728]. The exon/intron boundaries were determined by comparative alignment of the canine genomic sequence versus a canine cDNA sequence using LALIGN [20] and BLAST [21]. GC content and CpG islands were calculated with CpG plot [22]. The protein translation and calculation of protein molecular weight and pI was done with DNASTAR software (GATC, Konstanz, Germany).
Sequencing the MLPH cDNA
Fresh skin biopsies (4 or 6 mm diameter ~ 30 to 60 mg) were either frozen in liquid nitrogen and stored at -80 °C or stored in RNAlater (Qiagen, Hilden, Germany) at -20°C. RNA could be isolated and yielded cDNA from both storage methods. RNA of skin was isolated using the Trizol™ reagent (Invitrogen, Karlsruhe, Germany) or the the Qiagen RNAeasy 96 Universal Tissue Kit (Qiagen, Hilden, Germany). cDNA synthesis was performed using oligo-dT and the SuperScript™III reverse transcriptase (Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions. For the subsequent PCR 2–3 μl of the cDNA were used in 50 μl reactions containing 20 pmol of each PCR primer, 200 μM dNTPs and 2.5 units of Taq DNA polymerase (Qiagen, Hilden, Germany). The entire coding sequence with the exception of the last two codons of the MLPH cDNA was amplified as four overlapping fragments. Two rounds of semi-nested PCR had to be performed in order to generate enough cDNA for DNA sequencing. The primers and conditions for the RT-PCRs are given in Table 6. The RT-PCR products were purified from agarose gels using QiaExII (Qiagen, Hilden, Germany) and directly sequenced with the Dyenamic™ Terminator Cycle Sequencing Kit and a MegaBACE capillary sequencer. The cDNA sequence of the canine MLPH gene was submitted to the EMBL nucleotide database [EMBL:AJ920333].
Mutation analysis
DNA from approximately 350 dogs was available for various aspects of this study (140 Doberman Pinschers, 143 German Pinschers, 12 Large Munsterlanders, 6 Beagles, and ~50 dogs from other breeds or crossings). Doberman Pinscher and German Pinscher samples were collected from European and North American dogs. As the Doberman Pinschers showed some genetic differences with regard to their origin, these samples were divided into Doberman Pinschers of American origin (38 animals) and European origin (102 animals). The coat color of the dogs was recorded based on their pedigree certificates. In Pinschers there were four colors, black-and-tan, brown or red, blue, and Isabella fawn, respectively. Black-and-tan and brown or red were classified as wildtype colors, whereas blue and Isabella fawn were classified as dilute colors. Genomic DNA was isolated from blood using the QiaAmp 96 DNA Kit (Qiagen, Hilden, Germany) or the Nucleon BACC2 kit (Amersham Biosciences, Freiburg, Germany). Genomic DNAs of tissue samples from Doberman Pinschers were isolated using the Puregene Kit (Gentra, Minneapolis, MN, USA). All kits were used according to the manufacturers' instructions.
The exons of the canine MLPH were sequenced in 11 dog samples from Pinscher families with segregating coat color phenotypes. Six samples belonged to a Doberman Pinscher family (Fig. 4A). The other five samples belonged to a German Pinscher family (Fig. 4B, animals 8–12). Exons 2–15 were individually amplified with specific PCR primer pairs (Table 6). For the first exon two rounds of semi-nested PCR amplification were necessary to generate enough product for DNA sequencing. PCR was carried out in 20 μl reactions containing 10 ng genomic DNA according to the standard protocol advised by the manufacturer of the Taq DNA polymerase (GLtaq, Bremen, Germany). The subsequent sequencing of the PCR products was performed using the Dyenamic™ Terminator Cycle Sequencing Kit. The products were analyzed on a MegaBACE capillary sequencing machine. A set of eight SNPs around exon 2 was genotyped by DNA sequencing in most available DNA samples and statistically evaluated. The R199H mutation was either genotyped by DNA sequencing of the exon 7 PCR product or by RFLP analysis of this PCR product with HhaI on 1.5% agarose gels. The commercial use of MLPH based genotyping for diagnostic purposes in dogs is protected by the international patent EP04106291.
Table 6 PCR Primers used for the MLPH cDNA amplification and/or genomic MLPH mutation analysis
primer name primer sequence product size TM comments
MLPH cDNA, first round amplification
Mlph_cDNA_Ex1_F1 CCT TCC TTC CCC TGT AGG AC 513 bp 52°C
Mlph_cDNA_Ex4_R GGA TCA CCT TGG CAC TCC
Mlph_cDNA_Ex4_F1 GTG AAG ATC GGC TCG GTA G 686 bp 52°C
Mlph_cDNA_Ex9_R GGA TGC TGA GAG GTG GTG
Mlph_cDNA_Ex7_F1 CTT GGA CTT TGA GGC AGA C 985 bp 52°C
Mlph_cDNA_Ex14_R ACT GAA CTT CCT TCT GAG G
Mlph_cDNA_Ex10_F AGA GAA GAG GAG ACC CTC 651 bp 52°C
Mlph_cDNA_Ex16_R GCT GGG TCA TCA CAG GC
MLPH cDNA, second round amplification
Mlph_cDNA_Ex1_F2 CTG TAG GAC CGG AGA GAG C 502 bp 52°C
Mlph_cDNA_Ex4_R GGA TCA CCT TGG CAC TCC
Mlph_cDNA_Ex4_F2 AGA TCG GCT CGG TAG AGT G 679 bp 52°C
Mlph_cDNA_Ex9_R GGA TGC TGA GAG GTG GTG
Mlph_cDNA_Ex7_F2 CTC TGA CGA CTC CAC TTG 967 bp 52°C
Mlph_cDNA_Ex14_R ACT GAA CTT CCT TCT GAG G
Mlph_cDNA_Ex10_F AGA GAA GAG GAG ACC CTC 651 bp 52°C same primers as in first round amplification
Mlph_cDNA_Ex16_R GCT GGG TCA TCA CAG GC
MLPH genomic DNA
P_Mlph_Ex1_for1 AGT GGC CTC AAG CCC TG 591 bp 56°C
P_Mlph_Ex1_rev ATG AGC TCC CTG AGA ACC
P_Mlph_Ex1_for2 CTC CTC CCG AGG CTC CTG 563 bp 56°C
P_Mlph_Ex1_rev ATG AGC TCC CTG AGA ACC
Mlph_Ex2_for GTC ACC GAC ATT ATG TCA CAG 598 bp 58°C
Mlph_Ex2_rev CAG GCT GGA AGG TCA GAT C
Mlph_Ex3&4_for GAG ACC CAG GAT CGA GTC 725 bp 56°C
Mlph_Ex3&4_rev CAC TCA CAT TCC AAA GGA TC
Mlph_Ex5_for AGA AGT GAT GGA GGT CAC TG 471 bp 56°C
Mlph_Ex5_rev ACC GAA CAG GAA CAG GAG
Mlph_Ex6_for AG CTT GCC TGG ATG GAT G 582 bp 53°C
Mlph_Ex6_rev CTG CAG CCT CTG CCA AC
Mlph_Ex7_for GTC CTC AGC ACT TCT GAG 568 bp 53°C
Mlph_Ex7_rev GTG AGA AGC TTC TGG ACC
Mlph_Ex8_for CAG CGG GAT TTC TGA AAG C 483 bp 56°C
Mlph_Ex8_rev GTG TTG GAC AGT CAG AGT G
Mlph_Ex9_for CCC GCC TTT GCC TTA AGC 416 bp 56°C
Mlph_Ex9_rev GCT GCA AGG AGG AGC TTC
Mlph_Ex10_for CAG AGC CTG GCT CCT GAG 600 bp 56°C
Mlph_Ex10_rev CAC CTG GGA CAG GGA AGC
Mlph_Ex11_for AGG CGT CCA GAG GTT GTG 550 bp 55°C
Mlph_Ex11_rev GCC TGC TTC TGG AGA AGC
Mlph_Ex12_for CGC CGA CCA AGT CTT TGC 623 bp 55°C
Mlph_Ex12_rev TGG ACT TGA GGC CGT GTG
Mlph_Ex13_for CCT GTC TCC CCA GAT TCG 539 bp 55°C
Mlph_Ex13_rev GGG TTT GCC AAG GCT GAG
Mlph_Ex14_for CTC CTT TAT GCT CTG GCA C 591 bp 55°C
Mlph_Ex14_rev TTG TCA CAC GGA GAC ACA G
Mlph_Ex15_for CTC AGG CAG GCA GAC AAG 461 bp 55°C
Mlph_Ex15_rev TGC TCT GGG GTC CTA ACG
P_MLph_Ex16_for GCT TCA GAG CCT GAA ATT CT 469 bp 53°C
P_Mlph_Ex16_rev GAC AAA AGA TCA GGC TGG AG
Statistical methods
A set of eight SNPs around exon 2 was analyzed in 350 dogs of several breeds using Haplotyper 1.0 This software for haplotype inference uses the Bayesian algorithm [23]. Allele frequencies were tested for significant associations with the dilute phenotype using Fisher's Exact Test.
Authors' contributions
UP performed the molecular genetic analyses and drafted the manuscript. HH performed the statistical analyses. LM provided initial ideas for candidate genes as well as dog samples. SN and EM established an experimental Doberman cross with segregating coat colors and provided samples from this colony. ARGA established an experimental Beagle cross with segregating coat colors and provided samples from these animals. SMS provided the BHFD samples from an experimental cross of Large Munsterlanders, performed some confirmative genotyping, and contributed to the writing of the manuscript. TL conceived the study, participated in the MLPH gene characterization, and finalized the manuscript.
Acknowledgements
The authors would like to thank Heike Klippert-Hasberg, Diana Seinige, and Stefan Neander for expert technical assistance. The authors would also like to thank Beat Indermaur and Sabine Schindler for the pictures of Doberman Pinschers and blood samples. Finally the tremendous support of numerous veterinarians as well as dog breeders and owners who donated samples is gratefully acknowledged.
==== Refs
Laukner A Coat color in dogs. 2: Clinical significance Tierärztl Prax Ausg K Kleintiere Heimtiere 1998 26 124 128
Schmutz SM Moker JS Clark EG Shewfelt R Black hair follicular dysplasia, an autosomal recessive condition in dogs Can Vet J 1998 39 644 646 9789677
Carlotti DN Canine hereditary black hair follicular dysplasia and color mutant alopecia: Clinical and histopathological aspects Adv Vet Dermatol 1990 1 43 46
Laffort-Dassot C Beco L Carlotti DN Follicular dysplasia in five Weimaraners Vet Dermatol 2002 13 253 260 12358609 10.1046/j.1365-3164.2002.00302.x
Mercer JA Seperack PK Strobel MC Copeland NG Jenkins NA Novel myosins heavy chain encoded by murine dilute coat colour locus Nature 1991 349 709 713 1996138 10.1038/349709a0
Wilson SM Yip R Swing DA O'Sullivan TN Zhang Y Novak EK Swank RT Russel LB Copeland NG Jenkins NA A mutation in Rab27A causes the vesicle transport defects observed in ashen mice Proc Natl Acad Sci, USA 2000 57 7933 7938 10859366 10.1073/pnas.140212797
Matesic LE Yip R Reuss AE Swing DA O'Sullivan TN Fletcher CF Copeland NG Jenkins NA Mutations in Mlph, encoding a member of the Rab effector family, cause the melanosome transport defects observed in leaden mice Proc Natl Acad Sci USA 2001 98 10238 10243 11504925 10.1073/pnas.181336698
Pastural E Barrat FJ Dufourq-Lagelouse R Gertain S Sanal O Seger R Griscelli C Fischer A de Saint Basile G Griscelli disease maps to chromosome 15q21 and is associated with mutations in the myosin-VA gene Nat Genet 1997 16 289 292 9207796 10.1038/ng0797-289
Menasche G Pastural E Feldmann J Gertain S Ersoy F Dupuis S Wulfrat N Bianchi D Le Deist F de Saint Basile G Mutations in Rab 27A cause Griscelli Syndrome associated with hemaphagocytic syndrome Nature Genet 2000 25 173 176 10835631 10.1038/76024
Menasche G Ho CH Ozden S Feldmann J Tezcan I Ersoy F Houdusse A Fischer A de Saint Basile G Griscelli syndrome restricted to hypopigmentation results from melanophilin defect (GS3) or a Myo5A F-exon deletion (GS1) J Clin Invest 2003 112 450 456 12897212 10.1172/JCI200318264
Philipp U Quignon P Scott A André C Breen M Leeb T Chromosomal assignment of the canine melanophilin gene (MLPH): A candidate gene for coat color dilution in Pinschers J Hered 2005:
Breen M Jouquand S Renier C Mellersh CS Hitte C Holmes NG Cheron A Suter N Vignaux F Bristow AE Priat C McCann E André C Boundy S Gitsham P Thomas R Bridge WL Spriggs HF Ryder EJ Curson A Sampson J Ostrander EA Binns MM Galibert F Chromosome-specific single-locus FISH probes allow anchorage of an 1800-marker integrated radiation-hybrid /linkage map of the domestic dog genome to all chromosmes Genome Res 2001 11 1784 1795 11591656 10.1101/gr.189401
Schmutz SM Berryere TG Goldfinch AD TYRP1 and MC1r genotypes and their effects on coat color in dogs Mamm Genome 2002 13 380 387 12140685 10.1007/s00335-001-2147-2
Duan J Wainwright MS Comeron JM Saitou N Sanders AR Gelernter J Gejman PV Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor Hum Mol Genet 2003 12 205 216 12554675 10.1093/hmg/ddg055
O'Sullivan TN Wu XS Rachel RA Huang JD Swing DA Matesic LE Hammer JA 3rdCopeland NG Jenkins NA dsu functions in a MYO5A-independent pathway to suppress the coat color of dilute mice Proc Natl Acad Sci USA 2004 101 16831 16836 15550542 10.1073/pnas.0407339101
Li R Mignot E Faraco J Kadotani H Cantanese J Zhao B Lin X Hinton L Ostrander EA Patterson EF de Jong PJ Construction and characterization of an eightfold redundant dog genome bacterial artificial chromosome library Genomics 1999 58 9 19 10331940 10.1006/geno.1999.5772
German Resource Center/Primary Database
BACBAC Resources
NCBI trace archive
LAlign
NCBI BLAST home page
CpG plot
Niu T Oin ZS Xu X Liu JS Bayesian haplotype inference for multiple linked single-nucleotide polymorphisms Am J Hum Genet 2002 70 157 169 11741196 10.1086/338446
Kuroda TS Ariga H Fukuda M The actin-binding domain of Slac2-a/melanophilin is required for melanosome distribution in melanocytes Mol Cell Biol 2003 15 5245 5255 12861011 10.1128/MCB.23.15.5245-5255.2003
|
15960853
|
PMC1183202
|
CC BY
|
2021-01-04 16:30:39
|
no
|
BMC Genet. 2005 Jun 16; 6:34
|
utf-8
|
BMC Genet
| 2,005 |
10.1186/1471-2156-6-34
|
oa_comm
|
==== Front
BMC GenetBMC Genetics1471-2156BioMed Central London 1471-2156-6-361596702510.1186/1471-2156-6-36Research ArticleInbred mouse strains C57BL/6J and DBA/2J vary in sensitivity to a subset of bitter stimuli Boughter John D [email protected] Sandeep [email protected] Theodore M [email protected] Steven D [email protected] Anatomy and Neurobiology, University of Tennessee Health Science Center, Memphis, TN 38163 USA2 Anatomy and Neurobiology, University of Maryland School of Medicine, Baltimore, MD 21201, USA3 Program in Neuroscience, University of Maryland School of Medicine, Baltimore, MD 21201, USA2005 20 6 2005 6 36 36 4 2 2005 20 6 2005 Copyright © 2005 Boughter et al; licensee BioMed Central Ltd.2005Boughter 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
Common inbred mouse strains are genotypically diverse, but it is still poorly understood how this diversity relates to specific differences in behavior. To identify quantitative trait genes that influence taste behavior differences, it is critical to utilize assays that exclusively measure the contribution of orosensory cues. With a few exceptions, previous characterizations of behavioral taste sensitivity in inbred mouse strains have generally measured consumption, which can be confounded by post-ingestive effects. Here, we used a taste-salient brief-access procedure to measure taste sensitivity to eight stimuli characterized as bitter or aversive in C57BL/6J (B6) and DBA/2J (D2) mice.
Results
B6 mice were more sensitive than D2 mice to a subset of bitter stimuli, including quinine hydrochloride (QHCl), 6-n-propylthiouracil (PROP), and MgCl2. D2 mice were more sensitive than B6 mice to the bitter stimulus raffinose undecaacetate (RUA). These strains did not differ in sensitivity to cycloheximide (CYX), denatonium benzoate (DB), KCl or HCl.
Conclusion
B6-D2 taste sensitivity differences indicate that differences in consumption of QHCl, PROP, MgCl2 and RUA are based on immediate orosensory cues, not post-ingestive effects. The absence of a strain difference for CYX suggests that polymorphisms in a T2R-type taste receptor shown to be differentially sensitive to CYX in vitro are unlikely to differentially contribute to the CYX behavioral response in vivo. The results of these studies point to the utility of these common mouse strains and their associated resources for investigation into the genetic mechanisms of taste.
==== Body
Background
The majority of heritable traits in humans and other species are complex in nature, determined by interactions among multiple genes and environmental factors. Mouse genetic models have been critical in identifying genes that determine complex behavioral traits (e.g., [1-3]). The standard inbred strains C57BL/6J (B6) and DBA/2J (D2) have played a key role in mouse genetics, and they are among the strains included in the public and private genome sequencing projects. BXD/Ty (BXD) recombinant inbred (RI) mice, created from B6 and D2 progenitors, have been used to identify and map quantitative trait loci (QTLs) that influence diverse phenotypes such as addictive behavior (e.g., [4-6]), lifespan [7], central nervous system anatomy [8-10], and solution consumption [11-13]. Recently, the BXD set has been expanded to about 80 strains, which makes it the largest mouse RI mapping panel and a useful resource for QTL analysis [14].
The characterization of behavioral taste sensitivity in B6 and D2 mice would facilitate the use of BXD mice in mapping QTLs for taste sensitivity (56). B6 mice have been the most common inbred strain used in gustatory research, and as such have been characterized in terms of one- and two-bottle intake, brief-access taste sensitivity, operant taste detection tasks, taste discrimination, gustatory nerve physiology, taste receptor cell physiology, and taste cell-specific gene expression (e.g., [15-23]. D2 mice have not been as thoroughly characterized (e.g., [23-25]. However, it has long been appreciated that significant differences between B6 and D2 mice exist for consumption of sweet- and bitter-tasting stimuli, ethanol, and sodium chloride [26-29].
For stimuli characterized by humans as possessing a bitter taste, B6 and D2 mice vary in level of responsiveness when queried with intake tests: concentration-dependent strain differences have been measured for the typical bitter stimulus quinine hydrochloride (QHCl; [22,28]), acetylated sugars such as raffinose undecaacetate (RUA; [11,30]), and copper glycinate [13]. Such two-bottle intake procedures have been manageable for testing the large numbers of mice required for quantitative analysis. However, it is questionable whether these tests provide valid indicators of an animal's ability to recognize or discriminate a substance based on gustatory cues. Bitter stimuli comprise an exceptionally diverse set of chemical compounds that vary greatly in toxicity (e.g., [31,32]). Our recent study [33] demonstrates that post-ingestive effects related to bitter stimulus toxicity directly influence results from two-bottle tests and that these effects are considerably minimized in brief-access tests. Here we describe the use of a taste-salient brief-access procedure (e.g., [16,33,34] to characterize taste sensitivity in B6 and D2 mice to eight stimuli: six which are perceived by humans as predominantly bitter-tasting (QHCl, 6-n-propylthiouracil (PROP), MgCl2, RUA, denatonium benzoate (DB), and cycloheximide (CYX)), one with a complex salt/bitter taste (KCl), and an acid (HCl). These studies demonstrate that B6 and D2 mice differ in taste sensitivity to some, but not all, bitter or aversive stimuli, and suggest that they will be a useful resource for characterizing the genetic basis of bitter taste.
Results
Response to bitter and acid stimuli
Thirty B6 and 30 D2 mice were tested with six concentrations each of three different stimuli, such that a total of ~ 10 mice of each strain were tested for each of eight stimuli (Table 1; QHCl, PROP, MgCl2, RUA, DB, CYX, KCl, and HCl). Concentration-response functions were created for individual mice (for all compounds) and were fitted with two-parameter logistic functions, so that the concentration evoking half-maximal avoidance (c) could be determined. Examples of such individual functions for QHCl in B6 and D2 mice are shown in Figures 1 and 2, respectively. All B6 mice displayed concentration-dependent avoidance of QHCl (Fig. 1). Estimated half-maximal avoidance (c) ranged from 0.09 to 1.0 mM QHCl among individual mice, but there was not a significant difference between the six mice tested with QHCl as the first stimulus versus the four that were tested with QHCl as the third stimulus. Estimated half-maximal avoidance of QHCl among individual D2 mice ranged from 0.27 to 3.07 mM. For all but one D2 mouse (D105; Fig. 2), c > 1.15 mM. Some of the D2 mice such as D98, D97 and D82 showed relatively little avoidance of the higher concentrations of QHCl. As was the case for B6 mice, c did not vary significantly among individual mice as a function of whether the QHCL was presented as the first or last stimulus in the test series (D98 was not included in this comparison, because c could not be accurately estimated. However, all B6 and D2 mice were used for repeated measures comparisons; see below). The average half-maximal avoidance was 0.41 mM for B6 mice and 1.75 mM for D2 mice [t(17) = 4.33; p < 0.001]. Notably, comparisons of c within strain for all eight stimuli did not reveal significant effects of test group (i.e. whether the stimulus was presented first, second or third in series; see Table 1) with a single exception, noted below. Data collected for each compound were therefore combined for analysis of potential strain and gender effects.
Table 1 Test compounds and test series for 30 B6 and 30 D2 mice. Each mouse was tested with three compounds (2 consecutive test sessions per compound) over a two-week period. Stimuli were KCl, cycloheximide (CYX), raffinose undecaacetate (RUA), 6-n-propylthiouracil, quinine hydrochloride (QHCl), denatonium benzoate (DB), and HCl. A total of 10 mice were tested from each strain (5 males, 5 females) for KCl, CYX, PROP, MgCl2and QHCl. A total of 10 mice from each strain (5 males, 5 females for B6; 7 females, 3 males for D2) were tested with DB and HCl. A total of 11 B6 mice (6 females, 5 males) and 9 D2 mice (5 females, 4 males) were tested with RUA.
Number of Mice Tested
Squad Compounds B6 D2
A KCl, CYX1 5 5
B CYX, RUA, KCl 5 5
C PROP, MgCl2, QHCl 2 3
D MgCl2, PROP, QHCl 2 2
E DB, HCl2 4 6
F QHCl, MgCl2, PROP 6 5
G HCl, DB, RUA 6 4
Total Mice 30 30
1,2In squads A and E additional data on a third aversive stimulus was collected but was treated as pilot data and not included as part of the current experiment.
Figure 1 Concentration-response functions quinine in 10 individual B6 mice. Data points for each mouse represent average ratios across two days of testing; these means were fitted with two-parameter logistic functions and the concentration evoking half-maximal avoidance, c, was estimated. Italicized mice (B103, B96, B93, B67) were given QHCl as the last of three stimuli, as opposed to the others, which received QHCl as the first stimulus.
Figure 2 Concentration-response functions for QHCl in 10 individual D2 mice. Data points for each mouse represent average ratios across two days of testing; these means were fitted with two-parameter logistic functions and the concentration evoking half-maximal avoidance, c, was estimated. For one mouse (D98) this parameter could not accurately estimated, although this mouse had a mean lick ratio of 0.52 for 1 mM QHCl. Italicized mice (D81, D82, D53, D57, D58) were given QHCl as the last of three stimuli, as opposed to the others which received QHCl as the first stimulus.
Strain differences in taste sensitivity were found for four of eight compounds: QHCl, PROP, MgCl2, and RUA (Figure 3). For QHCl, D2 mice had significantly higher lick ratios than B6 mice across most of the concentration range, indicating decreased aversion. A strain x gender x concentration ANOVA revealed a main effect of strain [F(1,16) = 16.64, p < 0.001] but not gender. There was a significant strain x concentration interaction [F(5,80) = 3.74, p < 0.01]. The strain x gender, or strain x gender x concentration interactions were not significant for QHCl (or for any of the 7 other stimuli). Planned comparisons (Least Squares means) between strain at each concentration revealed that D2 mice were significantly less sensitive (p < 0.01) to 0.3 – 3 mM QHCl.
Figure 3 Lick ratios (mean ± SE) for B6 and D2 mice to concentration series of QHCl (A), PROP (B), MgCl2 (C), and RUA (D). The dotted lines on each graph represent a ratio score of 1.0, which indicates a lick rate equal to that of water. Asterisks identify significant strain effects at particular concentrations, as indicated by planned comparisons (p < 0.01). Lick ratios for each strain generally decreased with increasing concentration; B6 made fewer licks than D2 mice to high concentrations of QHCl (A) and PROP (B), and to both low and concentrations of MgCl2 (C). D2 mice made fewer licks than B6 mice at almost all concentrations of RUA (D).
Similarly, D2 mice were less sensitive to PROP: a strain x gender x concentration ANOVA revealed a main effect of strain [F(1,16) = 10.5, p < 0.01] but not gender. There was a significant strain x concentration interaction [F(5,80) = 7.01, p < 0.0001]. Planned comparisons indicated D2 mice possessed higher lick ratios at 3 and 10 mM. The mean half-maximal avoidance was 1.46 mM for B6 mice and for 4.97 mM for D2 mice [t(17) = 4.05; p < 0.001].
For MgCl2, there was a main effect of strain [F(1,16) = 12.6, p < 0.01] but not gender. Planned comparisons indicated D2 mice possessed significantly higher lick ratios at 0.003, 0.1 and 0.3 mM. A modest curve shift across the entire concentration range was not quite significant: mean half-maximal avoidance was 0.02 M for B6 mice and 0.05 M for D2 mice [t(17) = 2.34; p = 0.03].
For RUA, the direction of the strain difference was reversed. Notably, B6 mice did not display strong aversion to any concentration, whereas D2 mice avoided RUA in a concentration-dependent manner. There was a main effect of strain [F(1,16) = 39.56, p < 0.0001] but not gender. A significant interaction was found for concentration x strain [F(5,80) = 3.84, p < 0.01]. Planned comparisons indicated D2 possessed greater avoidance than B6 mice at 0.005, and at 0.03–0.3 mM. D2 mice actually possessed a greater level of avoidance to the lowest concentration of RUA (0.005 mM) relative to the next two higher concentrations (0.01 and 0.03). This tendency was evident in eight of nine individual D2 mice (data not shown), although it is not clear why such an effect was found. The mean curve shift between strains was not examined for RUA due to the lack of concentration-dependent avoidance in B6 mice, and the resulting inability to estimate the c-parameter.
We did not detect strain or gender differences in sensitivity to CYX, DB, KCl, or HCl (Figure 4). Both strains avoided higher concentrations of each of these stimuli in a concentration-dependent fashion. An effect on half-maximal avoidance based on test series was found for CYX: D2 mice that received CYX as the first stimulus tended to have a lower c value (mean = 0.53 μM) than those that received it as the second stimulus (mean = 2.66 μM; p < 0.01). A similar effect was found in the B6 mice, although not quite significant (p = 0.03). Testing each of these subgroups for significance with ANOVA (with n = 5 / strain) showed that B6 and D2 still did not differ in level of aversion (p > 0.08; however, small sample sizes in this comparison should be noted). Between conditions, lick ratios for both strains tended to differ modestly at 0.3 and 1 μM, but not at the higher concentrations (3–100 μM). The cause of the within-strain curve shifts is unclear, although B6 and D2 mice did not differ significantly in sensitivity to this stimulus.
Figure 4 Lick ratios (mean ± SE) for B6 and D2 mice to concentration series of CYX (A), DB (B), KCl (C), and HCl (D). The dotted lines on each graph represent a ratio score of 1.0, which indicates a lick rate equal to that of water. Lick ratios for each strain decreased with increasing concentration. The strains did not differ significantly for any of these compounds.
Baseline licking, performance and latency
Taste data were reported as lick ratios in order to standardize for possible strain differences in water lick rate [16,33]. We compared these "baseline" rates of water licking between B6 and D2 mice collapsed across all days of stimulus testing. B6 mice licked water at an average rate of 33.81 licks / 5 s whereas D2 licked at an average rate of 35.76 licks / 5 s; this difference was not significant [F(1,58) = 2.68; p > 0.1]. In addition to lick rate, we were also interested in examining other aspects of behavior in the task that are thought to be non-gustatory in origin. Performance was reported as percent trials completed per test session, per mouse. Table 2 lists mean performance rates for each strain, by gender, for each of three consecutive test sessions. In general, B6 mice of either gender tended to complete a slightly larger percentage of trials than did D2 mice [F(1,52) = 17.10, p < 0.01]. In order to account for possible effects of each unique stimulus on performance, we also examined this performance for each of the eight stimuli: There was still a significant main effect of strain [F(1,147) = 20.9; p < 0.001], although the effect for stimulus was not significant.
Table 2 Mean % trials completed by stimulus order. A trial was considered to be complete as long as a mouse took a single lick. After 120 s had elapsed without a lick, a given trial was considered over, and the next trial began. B6 mice completed significantly more trials (out of a possible 24 per session) than did D2 mice across all three stimuli with which those mice were tested.
n Stim 1 Stim 2 Stim 3
Strain Gender
B6 Female 15 96.9 94.6 96.0
Male 15 93.1 91.2 90.6
D2 Female 17 96.6 93.6 97.2
Male 13 93.4 88.6 94.6
We asked if olfactory cues might contribute to avoidance in the brief-access test. We measured the latency to initiate trials for each compound, because concentration-dependent changes in latency have been suggested previously to indicate olfactory contribution [35]. Latencies for all stimuli are shown in Figure 5. A significant main effect of strain was found for only KCl [F(1,160) = 9.96, p < 0.01], whereas effects of concentration were significant for DB and HCl [F(6,96) > 3.4, p < 0.01]. Only a single significant interaction was detected: Concentration x strain for RUA [F(6,96 = 3.15; p < 0.01]. In general, greater latency was often observed for the highest concentration of a given stimulus. Latency to lick may increase as a function of olfactory cues, but it is important to consider that other strain differences such as overall activity levels may also affect this measure. Overall, reliable effects on latency to lick bitter stimuli in mice have not been reported [33,34,36].
Figure 5 Strain comparison of latency to initiate taste trials of all eight stimuli. W = water trials. Latencies are means (± SE) of the median latencies for individual mice. A significant strain difference was found only for KCl. Effects of concentration were found for DB and HCl.
Discussion
Previous studies comparing the consumption of bitter stimuli in two or more standard inbred strains of mice have fostered genetic and molecular approaches towards identifying taste transduction mechanisms (e.g., [11,12,28,37-39]). Given the potential for which post-ingestive factors affect levels of intake of particular bitter stimuli [33], we were eager to assess taste sensitivity in the commonly used inbred strains B6 and D2 with a taste-salient brief-access assay. We demonstrated significant concentration-dependent differences in taste sensitivity to a subset of aversive compounds, including the bitter stimuli QHCl, PROP, MgCl2, and RUA. Strain differences were not found for CYX, KCl, DB, or HCl. These strains did not differ in baseline levels of water licking. Additionally, we did not find effects of gender on taste sensitivity.
The B6-D2 taste sensitivity differences demonstrated for QHCl and RUA indicate that differences in consumption of these stimuli between these strains [22,28] are based on immediate orosensory cues. Our results are also consistent with a potential relationship between quinine and PROP aversion [37]. D2 mice displayed significantly less aversion than B6 mice to both QHCl and PROP; the strain differences were similar in magnitude (Fig. 4). The relationship between quinine and PROP taste sensitivity is surprising when one considers that, in contrast to the situation in humans, taste sensitivity to PROP in mice is not correlated with its structural analogue phenylthiocarbamide (PTC) [33].
Linkage studies have postulated a major locus controlling quinine intake, qui, closely linked to Prp2 and Prh1 (which encode two proline-rich salivary proteins) and the microsatellite marker D6Mit13 on distal mouse chromosome 6 [21,37,40]. Harder and Whitney [37] further showed, using BXH RI mice (bred from progenitor strains B6 and C3HeB/HeJ), that PROP intake was also under Chr 6 control. The linkage of putative bitter taste receptor genes (Tas2rs) to these loci predicts that polymorphisms in particular Tas2rs underlie strain differences to these stimuli. However, the ligand specificity of these receptors is difficult to predict, given that only a handful of mouse or human bitter taste receptors have been functionally characterized, and that evidence for narrow or broad tuning is equivocal [41-43]. There is strong evidence for polygenic control of both QHCl and PROP intake based on quantitative genetic analyses [15,37,44,45], and Tas2r-independent mechanisms for QHCl taste have been suggested (e.g., [46,47]). A genetic analysis using the brief-access assay may help to determine whether non-Tas2r genes contribute to quinine sensitivity based on immediate sensory cues (56).
We also found that D2 mice were less sensitive for MgCl2, a compound not previously investigated in these strains. Interestingly, D2 mice licked the lowest concentration of MgCl2 (0.003 mM) at a rate greater than distilled water. We previously reported that C3HeB/FeJ inbred mice prefer 0.01 and 0.03 mM MgCl2 to water in a two-bottle intake test [33]. It is possible that water-deprived mice, as used in this study, will lick water at a maximal rate, making an appetitive response difficult to discern. However, Dotson and Spector [25] demonstrated that water-restricted mice of some strains will lick certain concentrations of appetitive stimuli (e.g., sucrose and glycine) at a higher rate than water. In our study, water licking rates during the MgCl2 test session did not significantly differ for those collected during testing with other stimuli. At higher concentrations (0.3–1.0 mM), MgCl2 was strongly avoided by both strains. This stimulus is perceived as having both a salty and bitter taste to humans [48]; in neural recordings with macaques it generally correlates with other bitters [49]. It is possible that at lower concentrations it is preferred over distilled water by some strains of mice for its salt taste component. MgCl2 is an intriguing candidate for which to determine possible Tas2r linkage; it evokes a strong anterior tongue-related neural response [50], which is a region that does not include strong T2R expression.
RUA is a non-toxic acetylated sugar; substantial variation in intake levels among inbred mouse strains have been shown for RUA as well another acetylated sugar, sucrose octaacetate (SOA) [11,51,52]. Sensitivity to SOA is determined by allelic variation at Soa, a locus linked to Prp and the Tas2r cluster (e.g., [38,53]). A separate locus, Rua, was proposed to control aversion to RUA 11. In view of the close chemical similarity of RUA and SOA, and given identical inbred strain distribution patterns (SDPs), a more parsimonious explanation is that Soa and Rua are identical loci [13,30]. In our experiment, RUA taste sensitivity provides an example where D2 mice display greater levels of avoidance than B6 mice, demonstrating the specific nature of the bitter avoidance response in these two strains.
B6 and D2 mice did not significantly differ in taste sensitivity to DB, KCl, HCl and CYX. The non-difference for CYX is especially interesting in that there are several coding-region polymorphisms in the Tas2r105 receptor (initially called T2R5) gene between these strains [41]; 56); this receptor is activated by CYX in diverse in vitro assays, and a shift in the concentration-response function indicates that the B6 isoform is less sensitive [41]. This finding has led to the erroneous conclusion that B6 mice "cannot detect cycloheximide" (e.g., [54]). In fact, our finding that D2 and B6 mice do not differ significantly in taste-based aversion of CYX (Fig. 4) supports the previous finding that these strains do not differ in consumption [12]; see Fig. 3B and 3C). Additionally, CYX evokes a peripheral (glossopharyngeal) nerve response in B6 mice comparable to that evoked by PROP [18]. It is likely that other T2Rs, or T2R-independent mechanisms, contribute to the behavioral response in B6 mice and mask any potential difference produced by the polymorphic receptor 55. Alternatively, it is possible that a subtle, yet significant, strain difference might be detected with the testing of a larger sample of mice, or with the use of a different procedure such as threshold detection (e.g., [17]).
Our results point to the utility of these common inbred strains and their associated genetic resources, such as the BXD strains, for investigation into the genetic mechanisms of taste. Although behavioral sensitivity to bitter-tasting stimuli has been linked to the Tas2r family of bitter taste receptor genes, both by positional or functional studies, compelling questions remain about the peripheral and central organization of bitter taste and the relative contributions of individual genes to specific taste sensitivities.
Methods
Mice
A total of 60 male and female mice (Mus musculus) from inbred strains were used in these experiments: 30 from each inbred strain (B6 and D2). Mice were tested at an average age of 3.5 months, and were age-matched between strains in each test group. Table 1 lists the number of mice tested in each experiment. Roughly equal numbers of B6 and D2 mice were tested with up to 3 different stimuli over a two-week period. Mice were obtained directly from the Jackson Laboratory (Bar Harbor, ME), or were derived (first generation) from mice obtained from Jackson Laboratory. Mice were housed in plastic home cages (28 × 17.5 × 13 cm) with stainless steel wire lids. Food (Teklad 8640 rodent diet) and water were available ad lib. Immediately prior to water deprivation, the mean weight of B6 mice was 20.1 g (females) and 27.4 g (males); the mean weight of D2 mice was 21.6 g (females) and 26.2 g (males).
Apparatus
Mice were tested daily in the Davis MS-160 automated gustometer (DiLog Instruments, Inc., Tallahassee, FL). The test chamber consisted of a plastic rectangular cage (30 × 14.5 × 18 cm) with a wire mesh floor; an oval opening centered in the front wall allows access to water or taste solutions contained in leak-proof sipper tubes. Fluid access is restricted by a computer-operated shutter.
Trials began with the opening of the shutter and ended 5 s after the mouse made its first lick on the drinking spout (see Procedure). Licks were counted with a high-frequency AC contact circuit. Failure to initiate a lick within 120 s also ended a trial, although such "zero lick" trials were ignored in analyses of lick rate as the failure to initiate licking could not be ascribed to orosensory factors. In between trials, a platform upon which the stimulus bottles were mounted was advanced to a new position. The inter-trial interval was held constant at 10 s. The test session ended after the completion of 29 trials.
Solutions
The taste stimuli used in these experiments were made from reagent-grade chemicals: 6-n-propylthiouracil (PROP), cycloheximide (CYX), raffinose undecaacetate (RUA), magnesium chloride (MgCl2), potassium chloride (KCl), quinine hydrochloride (QHCl), denatonium benzoate (DB) and hydrochloric acid (HCl) (Sigma Aldrich Corp.; St. Louis, MO). Multiple concentrations of each solution were made fresh daily using distilled water, and all taste stimuli were presented at room temperature (concentrations are listed in Table 1).
Brief-access tests
Water-deprived mice were trained to lick water in the gustometer and subsequently tested with a six-concentration series of a taste stimulus. Mice were water-deprived for 24 h prior to the first day of training, and from that point on were restricted to water consumed during the training or testing session (approximately 1.5 ml per session). On the first training day, mice were placed in the test chamber and given access to distilled water for 20 min. Most mice took at least 100 licks from the drinking spout in this first session. On the second training day, access was restricted to 5 s trials. Water was delivered at random from one of four water tubes, and mice had the opportunity to initiate up to 16 trials. Testing with the first bitter stimulus occurred on days 3 and 4. Six concentrations of the stimulus plus water were delivered using a randomized block design. Twenty-four trials were divided into 3 blocks of 8; within each block, each concentration of the stimulus plus two water trials were presented in a random order. Although we have previously determined that most mice are sated after 24 5-s trials of both stimulus and water, a final block of five consecutive water trials were given in order to allow the thirstiest mice to rehydrate. Licking in these additional trials was not analyzed. In sum, each test session provided three possible data points per stimulus concentration, and six for water trials. The order of all trials was randomized anew for each mouse and the position of bottles on the gustometer was randomized each day. Finally, mice of both strains were tested in a random order each day.
After testing on day 4, mice received ad lib water in their home cages for 48 hours (over the weekend), followed by a second 24-h water deprivation prior to a single training session and consecutive two-day tests with two additional stimuli, conducted as described above. We tested 7 "squads" of mice this way, until testing of 10 mice from either strain for each of eight stimuli was completed (see Table 1).
Data analysis of behavioral tests
The number of licks for each stimulus trial (each concentration being presented twice per mouse per session), plus water test trials, were averaged across the two test days for each individual mouse. These data were then reported as lick ratios (LR: average number of licks to stimulus / average number of licks during water test trial) in order to standardize for possible strain differences in water lick rate. Lick ratios thus range from a hypothetical zero (complete avoidance) to 1.0 (or greater). A ratio equal to zero was not possible because zero lick trials were not included in this analysis. Concentration-response functions were fit with a two-parameter logistic function:
Where x is the concentration of stimulus, c is the concentration evoking half-maximal avoidance (i.e. lick ratio = 0.5) and b is the slope. Fitting such curves provides a single parameter (c) that is sensitive to shifts in the concentration-response function, as potentially resulting from strain differences. For lick ratios in response to MgCl2, a three-parameter function was used; the additional parameter a was used to calculate an asymptotic maximum > 1.0, since lick ratios to the lowest concentration (0.003 M) of this tended to be greater than ~1.0, especially in D2 mice (see Figure 1). For group comparisons, c values were log transformed; strain values presented are therefore geometric means.
All relevant variables were analyzed using a general linear model: repeated measures (concentration) with between-subjects factors (strain, gender) and planned comparisons (LSD) at single concentrations based on the expectation of strain differences (Statistica software, StatSoft, inc., Tulsa, Oklahoma). Latency data (median latency) was log transformed for ANOVA. The statistical rejection criterion (α) for all tests was set a priori at the 0.01 level for main effects.
Authors' contributions
JB and SR collected and analyzed the behavioral data. JB and SM drafted the manuscript. JB, TN, and SM conceived of the study, participated in its design, and edited the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Aldan Shank for technical assistance with the behavioral studies and David V. Smith for comments on an earlier draft of the manuscript. This study was supported by grants from the National Institute on Deafness and Other Communication Disorders (DC05786 to SM, DC04935 to JB), a Program Enrichment Fellowship from the University of Maryland, Baltimore (TN) and by the University of Tennessee Health Sciences Center (JB).
==== Refs
Chesler EJ Lu L Wang J Williams RW Manly KF WebQTL: rapid exploratory analysis of gene expression and genetic networks for brain and behavior Nat Neurosci 2004 7 485 486 15114364 10.1038/nn0504-485
Glazier AM Nadeau JH Aitman TJ Finding genes that underlie complex traits Science 2002 298 2345 2349 12493905 10.1126/science.1076641
Tecott LH The genes and brains of mice and men Am J Psychiatry 2003 160 646 656 12668350 10.1176/appi.ajp.160.4.646
Shirley RL Walter NA Reilly MT Fehr C Buck KJ Mpdz is a quantitative trait gene for drug withdrawal seizures Nat Neurosci 2004 7 699 700 15208631 10.1038/nn1271
Hitzemann R Hitzemann B Rivera S Gatley J Thanos P Shou LL Williams RW Dopamine D2 receptor binding, Drd2 expression and the number of dopamine neurons in the BXD recombinant inbred series: genetic relationships to alcohol and other drug associated phenotypes Alcohol Clin Exp Res 2003 27 1 11 12543998
Kirstein SL Davidson KL Ehringer MA Sikela JM Erwin VG Tabakoff B Quantitative trait loci affecting initial sensitivity and acute functional tolerance to ethanol-induced ataxia and brain cAMP signaling in BXD recombinant inbred mice J Pharmacol Exp Ther 2002 302 1238 1245 12183685 10.1124/jpet.302.3.1238
Van Zant G de Haan G Genetic control of lifespan: studies from animal models Expert Rev Mol Med 1999 1999 1 12 14585117 10.1017/S1462399499001441
Williams RW Gu J Qi S Lu L The genetic structure of recombinant inbred mice: high-resolution consensus maps for complex trait analysis Genome Biol 2001 2 RESEARCH0046 11737945 10.1186/gb-2001-2-11-research0046
Williams RW Holdener BC Angel JM Oakey R Hunter KW Encyclopedia of the mouse genome VII. Mouse chromosome 7 Mamm Genome 1998 8 Spec No S136 59 9662624 10.1007/s003359900652
Lu L Airey DC Williams RW Complex trait analysis of the hippocampus: mapping and biometric analysis of two novel gene loci with specific effects on hippocampal structure in mice J Neurosci 2001 21 3503 3514 11331379
Lush IE The genetics of tasting in mice IV. The acetates of raffinose, galactose and beta-lactose Genet Res 1986 42 117 23 3754827
Lush IE Holland G The genetics of tasting in mice. V. Glycine and cycloheximide Genet Res 1988 52 207 212 3243425
Lush IE Hornigold N King P Stoye JP The genetics of tasting in mice. VII. Glycine revisited, and the chromosomal location of Sac and Soa Genet Res 1995 66 167 174 8522158
Peirce JL Lu L Gu J Silver LM Williams RW A new set of BXD recombinant inbred lines from advanced intercross populations in mice BMC Genet 2004 5 7 15117419 10.1186/1471-2156-5-7
Bachmanov AA Reed DR Tordoff MG Price RA Beauchamp GK Intake of ethanol, sodium chloride, sucrose, citric acid, and quinine hydrochloride solutions by mice: a genetic analysis Behav Genet 1996 26 563 573 8990535
Glendinning JI Gresack J Spector AC A high-throughput screening procedure for identifying mice with aberrant taste and oromotor function Chem Senses 2002 27 461 474 12052783 10.1093/chemse/27.5.461
Eylam S Spector AC The effect of amiloride on operantly conditioned performance in an NaCl taste detection task and NaCl preference in C57BL/6J mice Behav Neurosci 2002 116 149 159 11895177 10.1037//0735-7044.116.1.149
Danilova V Hellekant G Comparison of the responses of the chorda tympani and glossopharyngeal nerves to taste stimuli in C57BL/6J mice BMC Neurosci 2003 4 5 12617752 10.1186/1471-2202-4-5
Kusakabe Y Miura H Hashimoto R Sugiyama C Ninomiya Y Hino A The neural differentiation gene Mash-1 has a distinct pattern of expression from the taste reception-related genes gustducin and T1R2 in the taste buds Chem Senses 2002 27 445 451 12052781 10.1093/chemse/27.5.445
Frank ME Blizard DA Chorda tympani responses in two inbred strains of mice with different taste preferences Physiol Behav 1999 67 287 297 10477061 10.1016/S0031-9384(99)00071-2
Blizard DA Kotlus B Frank ME Quantitative trait loci associated with short-term intake of sucrose, saccharin and quinine solutions in laboratory mice Chem Senses 1999 24 373 385 10480673 10.1093/chemse/24.4.373
Kotlus BS Blizard DA Measuring gustatory variation in mice: a short-term fluid-intake test Physiol Behav 1998 64 37 47 9661980 10.1016/S0031-9384(98)00016-X
Eylam S Spector AC Taste discrimination between NaCl and KCl is disrupted by amiloride in inbred mice with amiloride-insensitive chorda tympani nerves Am J Physiol Regul Integr Comp Physiol 2005 288 R1361 8 15821286
Eylam S Spector AC Stimulus processing of glycine is dissociable from that of sucrose and glucose based on behaviorally measured taste signal detection in Sac 'taster' and 'non-taster' mice Chem Senses 2004 29 639 649 15337688 10.1093/chemse/bjh068
Dotson CD Spector AC The relative affective potency of glycine, L-serine and sucrose as assessed by a brief-access taste test in inbred strains of mice Chem Senses 2004 29 489 498 15269121 10.1093/chemse/bjh051
McClearn GERDA Differences in alcohol preference among inbred strains of mice. J Stud Alcohol 1959 20 691 695
Fuller JL Single-locus control of saccharin preference in mice J Hered 1974 65 33 36 4847746
Lush IE The genetics of tasting in mice. III. Quinine Genet Res 1984 44 151 160 6510710
Ninomiya Y Sako N Funakoshi M Strain differences in amiloride inhibition of NaCl responses in mice, Mus musculus J Comp Physiol [A] 1989 166 1 5 2600883
Harder DB Capeless CG Maggio JC Boughter JJ Gannon KS Whitney G Azen EA Intermediate sucrose octaacetate sensitivity suggests a third allele at mouse bitter taste locus Soa and Soa-Rua identity. Chem Senses 1992 17 391 401
Spielman AI Huque T Whitney G Brand JG The diversity of bitter taste signal transduction mechanisms Soc Gen Physiol Ser 1992 47 307 324 1285441
Glendinning JI Is the bitter rejection response always adaptive? Physiol Behav 1994 56 1217 1227 7878094 10.1016/0031-9384(94)90369-7
Nelson TM Munger SD Boughter JDJ Taste sensitivities to PROP and PTC vary independently in mice Chem Senses 2003 28 695 704 14627538 10.1093/chemse/bjg062
Boughter JDJ St John SJ Noel DT Ndubuizu O Smith DV A brief-access test for bitter taste in mice Chem Senses 2002 27 133 142 11839611 10.1093/chemse/27.2.133
Rhinehart-Doty JA Schumm J Smith JC Smith GP A non-taste cue of sucrose in short-term taste tests in rats Chem Senses 1994 19 425 431 7881974
St John SJ Boughter JDJ The contribution of taste bud populations to bitter avoidance in mouse strains differentially sensitive to sucrose octa-acetate and quinine Chem Senses 2004 29 775 787 15574813 10.1093/chemse/bjh082
Harder DB Whitney G A common polygenic basis for quinine and PROP avoidance in mice Chem Senses 1998 23 327 332 9669045
Adler E Hoon MA Mueller KL Chandrashekar J Ryba NJ Zuker CS A novel family of mammalian taste receptors Cell 2000 100 693 702 10761934 10.1016/S0092-8674(00)80705-9
Matsunami H Montmayeur JP Buck LB A family of candidate taste receptors in human and mouse Nature 2000 404 601 604 10766242 10.1038/35007072
Azen EA Davisson MT Cherry M Taylor BA Prp (proline-rich protein) genes linked to markers Es-12 (esterase-12), Ea-10 (erythrocyte alloantigen), and loci on distal mouse chromosome 6 Genomics 1989 5 415 422 2613230 10.1016/0888-7543(89)90004-9
Chandrashekar J Mueller KL Hoon MA Adler E Feng L Guo W Zuker CS Ryba NJ T2Rs function as bitter taste receptors Cell 2000 100 703 711 10761935 10.1016/S0092-8674(00)80706-0
Behrens M Brockhoff A Kuhn C Bufe B Winnig M Meyerhof W The human taste receptor hTAS2R14 responds to a variety of different bitter compounds Biochem Biophys Res Commun 2004 319 479 485 15178431 10.1016/j.bbrc.2004.05.019
Bufe B Hofmann T Krautwurst D Raguse JD Meyerhof W The human TAS2R16 receptor mediates bitter taste in response to beta-glucopyranosides Nat Genet 2002 32 397 401 12379855 10.1038/ng1014
Boughter JDJ Harder DB Capeless CG Whitney G Polygenic determination of quinine aversion among mice Chem Senses 1992 427 434
Boughter JDJ Whitney G Behavioral specificity of the bitter taste gene Soa Physiol Behav 1997 63 101 108 9402622 10.1016/S0031-9384(97)00398-3
Chen Y Herness MS Electrophysiological actions of quinine on voltage-dependent currents in dissociated rat taste cells Pflugers Arch 1997 434 215 226 9178618 10.1007/s004240050388
Caicedo A Pereira E Margolskee RF Roper SD Role of the G-protein subunit alpha-gustducin in taste cell responses to bitter stimuli J Neurosci 2003 23 9947 9952 14586025
Delwiche JF Buletic Z Breslin PA Covariation in individuals' sensitivities to bitter compounds: evidence supporting multiple receptor/transduction mechanisms Percept Psychophys 2001 63 761 776 11521845
Scott TR Giza BK Yan J Gustatory neural coding in the cortex of the alert cynomolgus macaque: the quality of bitterness J Neurophysiol 1999 81 60 71 9914267
Dahl M Erickson RP Simon SA Neural responses to bitter compounds in rats Brain Res 1997 756 22 34 9187310 10.1016/S0006-8993(97)00131-5
Warren RP Lewis RC Taste polymorphism in mice involving a bitter sugar derivative Nature 1970 227 77 78 5422630 10.1038/227077a0
Lush IE The genetics of tasting in mice. I. Sucrose octaacetate Genet Res 1981 38 93 95 7274665
Capeless CG Whitney G Azen EA Chromosome mapping of Soa, a gene influencing gustatory sensitivity to sucrose octaacetate in mice Behav Genet 1992 22 655 663 1290451 10.1007/BF01066636
Scott K The sweet and the bitter of mammalian taste Curr Opin Neurobiol 2004 14 423 427 15321062 10.1016/j.conb.2004.06.003
Mueller KL Moon MA Erlenbach I Chandrashekar J Zuker CS Ryba NJ The receptors and coding logic for better taste Nature 2005 434 225 9 15759003 10.1038/nature03352
Nelson TM Munger SD Boughter JD Jr Haplotypes at the Tas2r locus on distal chromosome 6 vary with quinine taste sensitivity in inbred mice BMC Genet 2005 6 32 15938754 10.1186/1471-2156-6-32
|
15967025
|
PMC1183203
|
CC BY
|
2021-01-04 16:30:38
|
no
|
BMC Genet. 2005 Jun 20; 6:36
|
utf-8
|
BMC Genet
| 2,005 |
10.1186/1471-2156-6-36
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-951596322510.1186/1471-2164-6-95Research ArticleA generic approach for the design of whole-genome oligoarrays, validated for genomotyping, deletion mapping and gene expression analysis on Staphylococcus aureus Charbonnier Yvan [email protected] Brian [email protected]çois Patrice [email protected] Manuela [email protected] Adriana [email protected] Pierre [email protected] Werner [email protected] Jacques [email protected] Genomic Research Laboratory, University Hospitals of Geneva, rue Micheli-du-Crest 24, CH-1211 Geneva 14, Switzerland2 Fondation pour Recherches Médicales, University of Geneva, avenue Roseraie 64, CH-1211 Geneva 4, Switzerland3 Service of Infection Diseases, University Hospitals of Geneva, rue Micheli-du-Crest 24, CH-1211 Geneva 14, Switzerland4 Clinical Microbiology Laboratory, University Hospitals of Geneva, rue Micheli-du-Crest 24, CH-1211 Geneva 14, Switzerland2005 17 6 2005 6 95 95 8 2 2005 17 6 2005 Copyright © 2005 Charbonnier et al; licensee BioMed Central Ltd.2005Charbonnier 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
DNA microarray technology is widely used to determine the expression levels of thousands of genes in a single experiment, for a broad range of organisms. Optimal design of immobilized nucleic acids has a direct impact on the reliability of microarray results. However, despite small genome size and complexity, prokaryotic organisms are not frequently studied to validate selected bioinformatics approaches. Relying on parameters shown to affect the hybridization of nucleic acids, we designed freely available software and validated experimentally its performance on the bacterial pathogen Staphylococcus aureus.
Results
We describe an efficient procedure for selecting 40–60 mer oligonucleotide probes combining optimal thermodynamic properties with high target specificity, suitable for genomic studies of microbial species. The algorithm for filtering probes from extensive oligonucleotides libraries fitting standard thermodynamic criteria includes positional information of predicted target-probe binding regions. This algorithm efficiently selected probes recognizing homologous gene targets across three different sequenced genomes of Staphylococcus aureus. BLAST analysis of the final selection of 5,427 probes yielded >97%, 93%, and 81% of Staphylococcus aureus genome coverage in strains N315, Mu50, and COL, respectively. A manufactured oligoarray including a subset of control Escherichia coli probes was validated for applications in the fields of comparative genomics and molecular epidemiology, mapping of deletion mutations and transcription profiling.
Conclusion
This generic chip-design process merging sequence information from several related genomes improves genome coverage even in conserved regions.
==== Body
Background
Current hybridization technologies allow assaying thousands of nucleic acid sequences in a single reaction on a solid substrate. Such massively parallel systems offer unprecedented opportunities for basic research and diagnostic applications, including gene sequencing [1], detection of genetic polymorphisms [2], genome-composition analysis [3,4] and measurement of gene expression profiles in prokaryotes [5,6] or cancer cells [7]. Oligonucleotide probes (up to 70-mer) offer more flexibility than cDNA probes since they can be tailored according to optimal in silico physico-chemical and specificity properties, and applied to any sequence data.
Early available probe design software identified sets of probes sharing homogeneous thermodynamic properties for probe-target hybridization [8]. More elaborated software tools include cross-homology testing of probes against a reference database by BLAST (Basic Local Alignment Search Tool) [9,10] or prediction of secondary structures into the thermodynamically-based approach [11-14]. A frequent drawback of some of these algorithms is to yield an excessive number of unprocessed BLAST outputs that complicates final selection of the most specific probes. Furthermore, these approaches do not take into consideration probe interaction with microarray surface, in particular the impact of mismatches position between the target and probes, as shown by Hughes et al [15]. Designing reliable oligonucleotide probes with available software is quite difficult for bacterial genomes with low GC content [16], low complexity in sequence composition, or frequent conserved repeats leading to erroneous target identification by cross-hybridization.
The reported method (OliCheck) implements an algorithm for filtering oligonucleotide probes libraries sharing homogeneous thermodynamic properties by using positional information of predicted target-probe binding regions. An additional characteristic of OliCheck is to annotate probes recognizing highly conserved targets shared by different genomes. Staphylococcus aureus (S. aureus) was selected as a model organism for implementing and experimentally validating this approach. The choice of this clinically important pathogen for fundamental and applied genomic studies is prompted by the availability of several fully or partially sequenced strain genomes [16-18]. A set of feature elements was designed by OliCheck to yield an extensive S. aureus genome coverage. This S. aureus specific probe set together with control probes were used to manufacture an oligoarray that was extensively validated for comparative genomics, molecular epidemiology, mapping of deletion mutations, and transcription profiling applications. The specificity, signal-response linearity, and influence of hybridization temperatures for transcript profiling are also described. Further genomic oligoarrays of several distinct microbial species have been successfully designed using this generic methodological approach.
Results
Design of a S. aureus oligoarray
The major genomic component of S. aureus is a 2.8 million base pairs (bp) circular chromosome with a low GC content (32.8%) which represents 2,595 ORFs in strain N315. The median length of strain N315 ORFs is 768 bp and the 25th–75th percentile ranges from 444-1,152 bp [16]. A probe design software (i.e. ArrayDesigner™) generated a comprehensive list of candidate 40–60 mer probes for each ORF of strain N315 according to the preset thermodynamic hybridization parameters (Fig. 1, step A). A similar process was performed with strain Mu50 and COL. This step yielded 417,776, 607,461 and 321,286 candidate probes for strains N315, Mu50 and COL, respectively.
Figure 1 Schematic representation of StaphChip probe selection. All ORFs of N315 were loaded into ArrayDesigner™ (a) to select oligonucleotides according to their thermodynamical properties (Step A). The 417,776 N315-derived probes were filtered for target specificity using BLAST against N315 genome (Step B). Each probe should recognize a single target yielding a defined signal intensity threshold, i.e. outside the green box (b), otherwise it is rejected (c). During Step C, all accepted probes are aligned against heterologous S. aureus genomes (i.e. Mu50 or COL) to annotate probes common to the other genomes (d, e). Each probe should recognize a single target yielding a defined signal intensity threshold, i.e. inside the red box (d), with no other signals outside the green box; otherwise it is ignored (e). The process is repeated from Step A to C with the two other strain databases. Final selection by spreadsheet analysis (Step D) yielded a total number of 5,427 probes hybridizing with one or more S. aureus genomes.
Further selection of candidate oligonucleotides to sort out cross-reactive probes within the genome of each strain was achieved by OliCheck. This specificity filtering test is a mathematical transposition of Hughes et al observations [15] on the impact of mismatches position between the target and probe with respect to the microarray surface. The occurrence of mismatch(es) in the distal portion (solution end) of the probe leads to a strong decrease in signal intensity, as opposed to the proximal portion (surface end). The results of the specificity filtering test, when performed separately on each strain, yielded a list of 48,415, 48,510, and 34,303 probes for N315, Mu50 and COL, respectively (Fig. 1, step B). In contrast to OliCheck-filtered probes that are expected to be devoid of cross-hybridization, >85% of the probes selected by ArrayDesigner™ alone displayed significant cross-hybridization with multiple ORFs.
For each strain, all accepted probes were further annotated by OliCheck against heterologous S. aureus genomes to identify probes common to the different genomes (Fig. 1, step C).
To fulfill the manufacturing requirements of a S. aureus genome-wide oligoarray, a further probe selection was performed. This selection used a spreadsheet program to rank probes according to optimal strain coverage and thermodynamic criteria, for providing one or more non-overlapping probes per target ORF (Fig. 1, step D).
In silico properties of the S. aureus oligoarray and manufacturing of StaphChip
The final set of 5,335 S. aureus OliCheck-filtered probes recognized 97.5, 93.0, and 81.0% of N315, Mu50, and COL ORFs, respectively. The low residual percentage of ORFs (2.5% for strain N315 and 7.0% for Mu50) that escaped recognition by our final probe set mostly included (51/65 ORFs for N315) mobile genetic elements, located either on prophage or transposon elements. A likely explanation for exclusion of these targets by OliCheck is the occurrence of highly homologous (>98% nucleotide identity) sequence repeats found in other ORFs. Accordingly, 92 residual probes covering these ORFs had to be selected relying on step A (Fig. 1) only.
To manufacture StaphChip, a total of 5,427 S. aureus probes were synthesized on the array together with a subset of control probes designed by OliCheck against E. coli K12 genome [19] (see methods).
Comparison of OliCheck with a currently used method
To validate the properties of OliCheck by comparison with an established method, we generated 60-mer oligonucleotides probes with homogenous thermodynamic criteria for N315 genome by the recently published ArrayOligoSelector [13] tool. The final set of probes generated according to the ArrayOligoSelector procedure [13] (n = 2,592) was further tested for cross-homology by using OliCheck. A large percentage of these probes (83%) were sorted out by OliCheck because they failed specificity criteria defined by Hughes et al [15].
Comparison of in silico-predicted with experimentally detected hybridization signals
Among the total number of 5,427 S. aureus probes, 4,812 (89%) recognized targets common to all three strains (Fig. 2), thus affording StaphChip valuable properties for comparative genomic and transcriptomic studies on S. aureus. The finding of a significant number of probes (n = 401) commonly identified in N315 and Mu50 but not COL may be explained by their closely-related genomic contents [16].
Figure 2 In silico specificity of selected probes. Venn diagram showing probes recognized by all strains or strain-specific. Whereas the vast majority of probes recognized target in all three strains (4,812 probes), other recognized only 1 or 2 strains (n = 523).
The number of probes predicted to detect genomic elements from each S. aureus strain was compared to those experimentally recorded using StaphChip (Table 1). For each strain, >99.5% of the in silico-predicted hybridization signals were indeed detected by hybridization of genomic DNA on StaphChip probes. Most of the false-positive signals were recorded on probes that were generated using ArrayDesigner™ only (n = 92/104). In silico analysis determined that these false-positive signals did not originate from recognition of intergenic regions (data not shown). Thus, such false-positive frequency is not transferable to the whole array.
Table 1 Number of negative and positive signals predicted by in silico analysis and recorded by comparative genome hybridization on StaphChip for N315, Mu50, or COL targets. Differences between expected and recorded signals are also shown.
S. aureus genome
N315 Mu50 COL
Positive Signal Predicted 5,219 5,307 4,851
Recorded 5,216 5,304 4,819
Missed 3 3 32
Negative Signal Predicted* 116 28 484
Recorded 107 20 397
Differences 9 8 87
*See Figure 2.
Use of StaphChip for deletion mapping and genomotyping applications
To evaluate the accuracy of StaphChip for deletion mapping, the Cy-3 labelled DNA from the SA113ica deletion mutant [20] was competitively hybridized with the Cy-5 labelled DNA from its isogenic parent. Figure 3 maps the ica-related signals which are missing in the ica mutant, as opposed to the unique positive signal generated by the tetracycline resistance marker used for the construction and selection of strain SA113ica. Fluorescent intensities of all other signals except ica-related genes were equivalent for both strains.
Figure 3 Mapping of a deleted gene region by StaphChip. Cy-5 labelled DNA of parental strain SA113 was co-hybridized with Cy-3 labelled DNA from its isogenic ica deletion mutant. Colored bars indicate the position of each probe used to map ica-related and adjacent ORFs. Background signals (green) were recorded from probes recognizing the ica-region known to be deleted in strain SA113ica (arrows), as opposed to positive red signals recorded in the wild-type strain. The tetracycline resistance marker used for the construction and selection of strain SA113ica is recorded in the green channel only. Dye swap experiments provided similar results (not shown). Data are raw signal intensities; background level is indicated by a dotted line.
The potential of StaphChip for epidemiological typing was analysed by comparative genomic hybridization (CGH) of S. aureus strains from various epidemiological origins. Genomic DNA from each individual S. aureus strain labelled with Cy3 was co-hybridized with equivalent amounts of Cy5-labelled control genomic DNA (pooled from N315, Mu50 and COL) and analyzed by two-way hierarchical clustering (Fig. 4).
Figure 4 Comparative genome hybridization using clustering analysis. Genomic DNA of each individual S. aureus strain was labelled with Cy3 and co-hybridized with equivalent amounts of Cy5-labelled genomic DNA pooled from N315, Mu50 and COL. Background-subtracted data were expressed as Log10 ratios and analyzed by two-way clustering using GeneSpring 6.1. Probes yielding positive and negative signals are shown in blue and yellow, respectively. The significance of black bars (a, b, and c) is indicated in the text. Note that the figure resolution does not allow visualising single probe differences but only clusters of probes.
The recently sequenced community-acquired MRSA strain MW2 [17], not included in StaphChip probe design, revealed important differences with strains N315, Mu50, and COL, as shown by a major yellow region on Figure 4. The set of probes yielding negative signals with MW2 DNA (Fig. 4, cluster a) corresponds to ORFs coding for antibiotic and heavy metal resistance determinants, bacterial adhesins and DNA modification enzymes lacking in this epidemic strain [17]. In contrast, extensively conserved genomic regions (Fig. 4, cluster b) are composed of house-keeping genes contributing to cell-wall, DNA synthesis, as well as essential metabolic enzymes.
Further analysis of strain SA113 and SA113ica compared to the other strains, showed no hybridization signals (Fig. 4, cluster c) for a number of genes (e.g. exotoxins and antibiotic resistance determinants) known to be missing in pathogenicity islands I-II-III of the NCTC8325 family [21].
Application of StaphChip for expression profiling studies
Control experiments were performed to study the dose-response of labelled cDNA and influence of hybridization temperature on the linearity and intensities of fluorescent signals. Two portions of N315 cDNA were labelled during reverse-transcription, with either Cy-3 or Cy-5 and competitively hybridized on StaphChip. Characteristics of fluorescent signals obtained on N315 specific probes were compared to those recorded on control E. coli oligonucleotide probes, at either 55 or 60°C. The median level of fluorescence intensities were approximately 5–10 fold higher for S. aureus probes compared to E. coli probes.
On S. aureus capture elements, log-transformed signal intensities recorded with equivalent input amounts (5 or 10 μg) of Cy-3 and Cy-5 cDNA were highly correlated (R > 0.95). Linear regression of Cy-3 versus Cy-5 scatter plots yielded slopes from 0.94 to 1.02 at 55 or 60°C (Table 2). When 5 μg of Cy-3 was competitively hybridized with 10 μg of Cy-5 labelled cDNA, slopes from 0.48 to 0.53 were recorded at 55 or 60°C, as expected (Table 2). Since log-transformed signal intensities remained highly correlated (R>0.85), this assessed the robustness of the recorded signals. Furthermore, the median intensity of fluorescent signals was marginally altered by temperature changes (not shown).
Table 2 Comparison of fluorescent signals obtained with S. aureus transcripts hybridized on specific (S. aureus) versus non-specific (E. coli) oligonucleotide probes as a function of target amounts and hybridization temperatures.
S. aureus probes E. coli probes
RCy3/Cy5 Slope RCy3/Cy5 Slope
Temp. 55°C Cy3/Cy5: 5 μg/5 μg 0.95 1.02 0.95 1.2
Temp. 55°C Cy3/Cy5: 10 μg/10 μg 0.99 0.94 0.95 1.2
Temp. 55°C Cy3/Cy5: 5 μg/10 μg 0.86 0.48 0.38 0.39
Temp. 60°C Cy3/Cy5: 5 μg/5 μg 0.98 0.94 0.96 1.78
Temp. 60°C Cy3/Cy5: 10 μg/10 μg 0.98 0.94 0.98 1.25
Temp. 60°C Cy3/Cy5: 5 μg/10 μg 0.85 0.53 0.61 0.72
In contrast, signal intensities from E. coli control probes were highly disturbed by temperature changes or by altering the ratios of fluorescently-labelled S. aureus cDNA (Table 2). Furthermore, the median intensity of fluorescent signals decreased by >2 fold at 60 compared to 55°C (not shown). Dye-swap experiments yielded equivalent results (not shown).
Reproducibility of StaphChip for expression profiling
To evaluate the reproducibility of fluorescent signals (Fig. 5), eight independent hybridization experiments were performed under identical conditions using Cy-3 labelled N315 cDNA, derived from overnight cultures. Maximal relative errors [(average-min)/average] of fluorescent probe signal intensities were more widely scattered on E. coli (Fig. 5B) compared to S. aureus (Fig. 5A) capture elements, thus reflecting the instability of poorly specific interactions with S. aureus targets and E. coli probes. The 90th percentile of maximal relative errors from S. aureus probes represented <25% of average signal intensities. In contrast, the same percentile of maximal relative errors recorded from E. coli probes represented >100% of average signal intensities. Furthermore, the 50th percentile on E. coli probes was superior to the 90th percentile of maximal relative errors on S. aureus probes.
Figure 5 Reproducibility of fluorescent signals in replicate experiments. Signals generated on StaphChip by 10 μg Cy-3 labelled N315 cDNA hybridized at 60°C. Average fluorescence intensities from replicate experiments (n = 8) and their maximal relative errors on S. aureus (A) or E. coli (B) capture probe elements are presented as scatter plots. The cumulative distribution of maximal relative errors is shown for S. aureus (C) or E. coli (D).
Finally, we compared signal intensities generated from individual gene transcripts covered by two or more adjacent but non-overlapping S. aureus probes. Figure 6 shows average fluorescence intensities and their maximal relative errors (A) and the cumulative distribution of maximal relative errors (B) For nearly 90% of the gene transcripts (n = 2,269), maximal relative errors of ORF-related signals were <60%. This yields evidence that multiple probes recognizing the same transcripts provide reproducible signals and that StaphChip provides reliable and robust determinations of genome-wide transcripts.
Figure 6 Reproducibility of fluorescent signals recorded from multiple non-overlapping capture elements for common transcripts. 10 μg Cy-3 labelled N315 cDNA were hybridized at 60°C on StaphChip. For 2,269 selected transcripts detected by two or multiple probes (n = 5,079), average fluorescence intensities and their maximal relative errors are presented in panel (A), and the cumulative distribution of maximal relative errors in panel (B).
Validation of relative transcript levels with RT-qPCR
Figure 7 shows the fold changes estimated by quantitative RT-PCR for 18 transcripts found to be either up regulated, down regulated or unchanged on the StaphChip. A good fitting was observed between both methods, with a correlation coefficient of 0.91 obtained from linear fitting. Of note, the quantitative changes recorded with RT-qPCR tended to be higher than those quantified with the microarray. This finding is supported by previous observations that the dynamic range of RT qPCR is higher than that of microarray [22,23], that only reach 2–3 orders of magnitude. These data validate the use of our method for quantitative gene-expression analysis.
Figure 7 Comparison of gene expression changes by real-time quantitative PCR and microarray analysis. Fold changes of gene expression estimated by either technique are shown for a set of 18 genes of S. aureus tested in two metabolic conditions. Data represent average values ± standard error of the mean of three independent experiments performed in duplicates.
Discussion
Several recent studies have shown the usefulness of microarrays for genomic and global transcriptomic studies of microbial pathogens [23-25]. Each step of microarray experiments needs to be optimized and validated, from array design and manufacture to data collection and analysis. Among critical technical parameters that need to be controlled are microarray surface chemistry, probe sequence, probe deposition process, and hybridization conditions. Accuracy of microarray-generated data can be improved by using multiple replicates, dye swaps and statistical data analysis. Compared to cDNA microarrays, oligoarrays provide a flexible design and are considered more reliable in terms of sensitivity and specificity [26-29]. Reported drawbacks of cDNA or PCR probes include: i) unpredicted secondary structures, ii) uncontrolled cross-hybridization occurring on repeats or partially homologous regions of PCR probes, and iii) varying amounts and purity of spotted products [27,30]. Recent software development allows genome-wide selection of sub-sequences that uniquely identify genes. Ideally, these approaches should amplify fragments of constant length, thus minimizing the differences in PCR amplification efficiency as well as in hybridization kinetics [31]. However, the extent of cross-hybridization has rarely been evaluated and reported, and thus may lead to severe errors in higher level data analysis, such as clustering [32], genome composition analysis and genotyping for molecular epidemiology [4].
Most oligoarray applications dedicated to prokaryotes were developed by companies using proprietary algorithms [33,34] whose detailed properties are rarely available. Furthermore, the lack of published validation data prevents adequate comparison of those short-probe oligoarrays with investigator-designed oligonucleotide arrays.
To date, several strategies for oligoarray design have been described, but their experimental validation is often limited [35] or absent [36]. A drawback of these approaches is to apply thermodynamics laws on probe/target interaction as determined in solution [37], but ignoring effects related to oligonucleotide immobilization on a microarray surface [15,38].
To address this issue, OliCheck considered the influence of predicted probe/target binding with respect to its position along the immobilized probe, as demonstrated by Hughes et al [15]. This tool allowed selecting probes from large oligonucleotide libraries in order to provide multiple genome coverage, suitable for epidemiological and transcriptomic studies.
As a specific application, OliCheck was used to design StaphChip, an oligoarray dedicated to genomic studies on S. aureus, a clinically important pathogen with a low GC content and numerous sequence repeats. The 5,427 feature elements were selected to ideally cover all ORFs of three S. aureus genomes with two non-overlapping probes, as validated under experimental conditions. A particular achievement of this strategy is to yield cross-annotations between the designed probes and homologous ORF regions conserved across several genomes. Any new genome sequence can be screened by OliCheck to identify probes that can specifically detect homologous ORFs. Cross-annotations on the recently released S. aureus MW2 genome [17] yielded 78% gene coverage. It should be mentioned that the recently published S. aureus COL genome [39] composition and annotation have changed significantly since the early release by TIGR in 2003.
The properties of StaphChip design were confirmed by comparative genome hybridization and global gene expression studies. Work in progress assesses the reliability of StaphChip for monitoring S. aureus transcription changes during biofilm formation, endocytic stage, or expression of antibiotic resistance. Another achievement (unpublished data) was the design of oligonucleotide probes for the genomes of Neisseria meningitidis A and B, E. coli K12, Erwinia carotovora and E. chrysanthemi; having GC contents ranging from 32.8 to 52%. Furthermore, OliCheck design expands oligoarray use for the study of host-pathogen interactions by potentially preventing cross-hybridization between bacterial probes and contaminating host nucleic acids.
Conclusion
In summary, this work describes a validated approach to select optimal oligoarray capture elements for S. aureus expression analysis and comparative genome hybridization studies. This generic approach will enable researchers to develop customized oligoarrays for genomic studies of any sequenced microorganism.
Methods
Design of specific oligonucleotide probes
Step A
An initial set of candidate oligonucleotide probes was generated by ArrayDesigner™ 1.17 (Premier Biosoft Intl) using the following probe parameters: (i) 40–60 bp probe length, (ii) 65 ± 10°C target Tm, (iii) <-5.0 kcal/mol for hairpins, (iv) <-8.0 kcal/mol for self-dimers, and (v) dinucleotide repeats shorter than 5 bp (Supplementary material provides OliCheck input format description, for using other probe design software or probe list). The program tested separately each open reading frame (ORF) of the different S. aureus genomes freely available at NCBI (S. aureus N315 [Genbank# BA000018], Mu50 [Genbank#BA000017]), and TIGR (S. aureus COL [TIGR unfinished microbial genome, released in March 27, 2003]). A comprehensive list of all possible probes ranked according to thermodynamic criteria was provided for each genome (Fig. 1a).
Further selection of specific oligonucleotide probes was performed by the design of an original program called OliCheck. This approach is derived from the experimental findings of Hughes et al. that microarray hybridization signals are mostly influenced by mismatches in the solution-end (distal part) rather than surface-end (proximal part) portion of the oligonucleotide probe [15]. OliCheck queries the locally available S. aureus genome databases by performing a BLAST (BLASTN for Windows, version 2.2.2) [9] analysis for each probe.
Step B
A first probe selection is performed by aligning each probe against its own genome, e.g. N315 candidate probes against N315 genome (Fig. 1b, and 1c).
Each BLAST output is analyzed to extract alignment information. The best scored target-probe alignment (perfect match) is considered as the target of the probe, therefore exempted from further testing. To avoid cross-hybridization, all other alignments are tested using the specificity filtering test. For this test, any 60-mer probe that would align to the target with >11 matched nucleotides by its distal third (solution-end), or >31 matched nucleotides by its proximal two-thirds (surface-end) is rejected. An additional condition is the exclusion of irrelevant target-probe pairs with >18 consecutive nucleotide matches.
Step C
Further probe selection is achieved by aligning each candidate probe from each genome (e.g. N315) against the genomes of the other S. aureus strains (e.g. Mu50 and COL) (Fig. 1d, and 1e). This process allows annotating probes detecting homologous targets in the other genomes.
Each BLAST output is analyzed to extract alignment information. The best scored target-probe alignment is tested to predict a high hybridization signal. This efficiency test requires the absence of any mismatch in the distal half and <20 mismatches in the proximal half of the probe. If those requirements are not fulfilled, the probe is rejected; otherwise the alignment is considered appropriate for detecting a homologous target and the process is continued. To avoid cross-hybridization with targets from other genomes, further alignments obtained with that probe are checked by the specificity filtering test, as defined in step B. Each probe fulfilling these requirements is annotated as detecting a unique homologous sequence target.
Step D
Using a spreadsheet program, the probes were sorted by their ORF target names and their best thermodynamic criteria. Probes showing the best combination of thermodynamic criteria and strain coverage were selected for microarray manufacturing (Fig. 1f). In addition to the selected S. aureus probes, OliCheck was used with the same parameters to design 2,873 probes specific for Escherichia coli K12 (E. coli K12 [Genbank# U00096]).
A compiled version of OliCheck compatible with Windows 2000 or XP and written in the Delphi programming language (Delphi 7, Borland) is freely downloadable for non-profit use at Genomic Research Laboratory website [40].
In silico comparison of our algorithm with ArrayOligoSelector
Three probes set of 60-mer probes with homogenous thermodynamic criteria (Tm = 60°C) were generated using default parameters for N315 genome by: i) ArrayDesigner ii) ArrayOligoSelector for all candidate probes, iii) ArrayOligoSelector for the best candidate probes (one per ORF). The output lists generated by ArrayOligoSelector were reformatted to match OliCheck input file format. The list of probes generated by either software was further processed by OliCheck for cross-homology filtering. The sets of probes selected by each method were further compared for homology using BLAST. Alignment showing E-value <1E-20 were considered as homologous.
Microarray manufacturing
The StaphChip microarray was manufactured by in situ synthesis of 8,455 long oligonucleotide probes (Agilent). It consists of 5,427 S. aureus and 2,873 E. coli specific probes, together with A. thaliana control probes for spiked controls.
Preparation of the labelled nucleic acids
For comparative genome hybridization, each S. aureus strain was grown overnight in 2 ml Mueller-Hinton broth (MHB) and total DNA was extracted and purified using DNeasy columns (QIagen) following manufacturer's instructions. DNA purity and concentration were assayed by spectrophotometer. 2 μg DNA were labelled by the Klenow fragment of DNA polymerase I (BioPrime, Invitrogen) with Cyanine-3 or Cyanine-5 coupled dCTP (NEN) for 2 hours at 37°C, then stopped by the addition of 5 μl 0.5 M EDTA. Labelled DNA was purified on QiaQuick columns (Qiagen).
For gene expression analysis, total RNA was extracted from 2 ml exponential or overnight cultures using the Rneasy kit (Qiagen) as previously described [41]. Batches of 5 or 10 μg of total S. aureus RNA were spiked with increasing amounts of different Arabidopsis thaliana mRNAs (SpotReport, Strategene), used as external calibrators. The RNA mixture was labelled by Cy-3 dCTP or Cy-5 dCTP, using the SuperScript II (Invitrogen) following manufacturer instructions. Labelled cDNA was then purified onto QiaQuick columns.
Hybridization and scanning parameters
Unless specified, equivalent amounts of cDNA (or genomic DNA) labelled with Cyanine-3 or Cyanine-5, were diluted in 250 μl Agilent hybridization buffer, and hybridized at a temperature of 60°C for 17 hours in a dedicated hybridization oven (Robbins Scientific). For comparative genome hybridization, genomic DNA from each individual S. aureus strain was labelled with Cy3 and co-hybridized with equivalent amounts of Cy5-labelled genomic DNA pooled from N315, Mu50 and COL [42]. Slides were washed, dried under Nitrogen flow and scanned (Agilent) using 100% PMT power for both wavelengths. Data were extracted and processed using Feature Extraction™ software (version 5.0, Agilent).
For gene expression analysis, saturated spots were excluded from subsequent analysis. Local background-subtracted signals were corrected for unequal dye incorporation or unequal load of labelled product. The algorithm consisted of a rank consistency filter and a curve fit using the default LOWESS (locally weighted linear regression) method. Spots showing a reference signal lower than background plus two standard deviations were also excluded from subsequent analyses.
For comparative genome hybridization, local background-subtracted data were expressed as Log10 ratios and analyzed by two-way clustering using GeneSpring 6.1 (SiliconGenetics).
Real time, quantitative PCR analysis
The expression of 18 genes involved in metabolic pathways was quantitatively assayed using by 1 step RT-qPCR using a SDS7700 (Applied Biosystems, Framingham, MA). Primers and probes were identified by scanning each gene sequence using the software Primer Express 2.0(Applied Biosystems). All identified sequences were further aligned on the whole genome of sequenced strains to ensure gene specificity and conservation of the target sequence between strains. Optimal concentration of primers and Taqman probes (labelled with FAM in 3' and coupled to TAMRA in 5' as quencher and purchased from Eurogentec, Seraing, Belgium), determined accordingly to manufacturer's instructions were 200 nM and 100 nM per reaction, respectively. Primers and probes were mixed in Platinum qRT-PCR Thermoscript kit (Invitrogen) with 0.4 ng of total purified RNA. Fold changes were calculated after normalization with the expression level of the 16s rRNA gene as previously described [41].
Authors' contributions
YC performed Olicheck implementation, designed microarray and wrote the manuscript. BG performed Olicheck implementation, designed microarray and helped writing the manuscript. PF participated in the design of the study, designed experiment protocols, performed microarray analysis. MB contributed to design experiment protocols and performed microarray experiments. AR has been involved in the CGH project. PV has been involved in drafting the article and revising it critically for intellectual content. WS has made substantial contributions to conception and design. JS initiated the study and helped writing the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by grants from the Swiss National Science Foundation PP00B–103002/1 (JS), 4049-63250 and 3200BO-103951 (PV), National Research Program 49 4049-63250 (PV and JS) and the Fondation pour Recherches Médicales (Gene Expression in Health and Diseases Program). We thank Prof K. Hiramatsu and J. Etienne for sending strains N315, Mu50, and MW2 as well as Prof F. Götz and P. Moreillon for providing strains SA113, SA113ica, and COL. We also thank Prof D. Lew for his constructive discussion throughout the work.
==== Refs
Yershov G Barsky V Belgovskiy A Kirillov E Kreindlin E Ivanov I Parinov S Guschin D Drobishev A Dubiley S Mirzabekov A DNA analysis and diagnostics on oligonucleotide microchips Proc Natl Acad Sci U S A 1996 93 4913 4918 8643503 10.1073/pnas.93.10.4913
Dubiley S Kirillov E Mirzabekov A Polymorphism analysis and gene detection by minisequencing on an array of gel-immobilized primers Nucleic Acids Res 1999 27 e19 10471749 10.1093/nar/27.18.e19
Kim CC Joyce EA Chan K Falkow S Improved analytical methods for microarray-based genome-composition analysis Genome Biol 2002 3 RESEARCH0065 12429064 10.1186/gb-2002-3-11-research0065
Lucchini S Thompson A Hinton JC Microarrays for microbiologists Microbiology 2001 147 1403 1414 11390672
Eriksson S Lucchini S Thompson A Rhen M Hinton JC Unravelling the biology of macrophage infection by gene expression profiling of intracellular Salmonella enterica Mol Microbiol 2003 47 103 118 12492857 10.1046/j.1365-2958.2003.03313.x
Staudinger BJ Oberdoerster MA Lewis PJ Rosen H mRNA expression profiles for Escherichia coli ingested by normal and phagocyte oxidase-deficient human neutrophils J Clin Invest 2002 110 1151 1163 12393851 10.1172/JCI200215268
Golub TR Slonim DK Tamayo P Huard C Gaasenbeek M Mesirov JP Coller H Loh ML Downing JR Caligiuri MA Bloomfield CD Lander ES Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999 286 531 537 10521349 10.1126/science.286.5439.531
Array Designer 3.01 http://www premierbiosoft com/dnamicroarray/index html 2004
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
Talla E Tekaia F Brino L Dujon B A novel design of whole-genome microarray probes for Saccharomyces cerevisiae which minimizes cross-hybridization BMC Genomics 2003 4 38 14499002 10.1186/1471-2164-4-38
Rouillard JM Herbert CJ Zuker M OligoArray: genome-scale oligonucleotide design for microarrays Bioinformatics 2002 18 486 487 11934750 10.1093/bioinformatics/18.3.486
Wright MA Church GM An open-source oligomicroarray standard for human and mouse Nat Biotechnol 2002 20 1082 1083 12410248 10.1038/nbt1102-1082
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
Nielsen HB Wernersson R Knudsen S Design of oligonucleotides for microarrays and perspectives for design of multi-transcriptome arrays Nucleic Acids Res 2003 31 3491 3496 12824351 10.1093/nar/gkg622
Hughes TR Mao M Jones AR Burchard J Marton MJ Shannon KW Ziman M Meyer MR Kobayashi S Dai H He YD Stephaniants SB Cavet G Walker WL West A Coffey E Shoemaker DD Stoughton R Blanchard AP Friend SH Linsley PS Expression profiling using microarrays fabricated by an ink-jet oligonucleotide synthesizer Nat Biotechnol 2001 19 342 347 11283592 10.1038/86730
Kuroda M Ohta T Uchiyama I Baba T Yuzawa H Kobayashi I Cui L Oguchi A Aoki K Nagai Y Lian J Ito T Kanamori M Matsumaru H Maruyama A Murakami H Hosoyama A Mizutani-Ui Y Takahashi NK Sawano T Inoue R Kaito C Sekimizu K Hirakawa H Kuhara S Goto S Yabuzaki J Kanehisa M Yamashita A Oshima K Furuya K Yoshino C Shiba T Hattori M Ogasawara N Hayashi H Hiramatsu K Whole genome sequencing of meticillin-resistant Staphylococcus aureus Lancet 2001 357 1225 1240 11418146 10.1016/S0140-6736(00)04403-2
Baba T Takeuchi F Kuroda M Yuzawa H Aoki K Oguchi A Nagai Y Iwama N Asano K Naimi T Kuroda H Cui L Yamamoto K Hiramatsu K Genome and virulence determinants of high virulence community-acquired MRSA Lancet 2002 359 1819 1827 12044378 10.1016/S0140-6736(02)08713-5
Gill SR Staphylococcus aureus COL genome http://www tigr org/tigr-scripts/CMR2/GenomePage3 spl?database=gsa 2004
Blattner FR Plunkett GIII Bloch CA Perna NT Burland V Riley M Collado-Vides J Glasner JD Rode CK Mayhew GF Gregor J Davis NW Kirkpatrick HA Goeden MA Rose DJ Mau B Shao Y The complete genome sequence of Escherichia coli K-12 Science 1997 277 1453 1474 9278503 10.1126/science.277.5331.1453
Cramton SE Gerke C Schnell NF Nichols WW Gotz F The intercellular adhesion (ica) locus is present in Staphylococcus aureus and is required for biofilm formation Infect Immun 1999 67 5427 5433 10496925
Novick RP Schlievert P Ruzin A Pathogenicity and resistance islands of staphylococci Microbes Infect 2001 3 585 594 11418332 10.1016/S1286-4579(01)01414-9
Ramakrishnan R Dorris D Lublinsky A Nguyen A Domanus M Prokhorova A Gieser L Touma E Lockner R Tata M Zhu X Patterson M Shippy R Sendera TJ Mazumder A An assessment of Motorola CodeLink microarray performance for gene expression profiling applications Nucleic Acids Res 2002 30 e30 11917036 10.1093/nar/30.7.e30
Nuwaysir EF Huang W Albert TJ Singh J Nuwaysir K Pitas A Richmond T Gorski T Berg JP Ballin J McCormick M Norton J Pollock T Sumwalt T Butcher L Porter D Molla M Hall C Blattner F Sussman MR Wallace RL Cerrina F Green RD Gene expression analysis using oligonucleotide arrays produced by maskless photolithography Genome Res 2002 12 1749 1755 12421762 10.1101/gr.362402
Wang D Coscoy L Zylberberg M Avila PC Boushey HA Ganem D DeRisi JL Microarray-based detection and genotyping of viral pathogens Proc Natl Acad Sci U S A 2002 99 15687 15692 12429852 10.1073/pnas.242579699
Okamoto T Suzuki T Yamamoto N Microarray fabrication with covalent attachment of DNA using bubble jet technology Nat Biotechnol 2000 18 438 441 10748527 10.1038/74507
Blanchard AP Friend SH Cheap DNA arrays-it's not all smoke and mirrors Nat Biotechnol 1999 17 953 10504690 10.1038/13644
Barczak A Rodriguez MW Hanspers K Koth LL Tai YC Bolstad BM Speed TP Erle DJ Spotted long oligonucleotide arrays for human gene expression analysis Genome Res 2003 13 1775 1785 12805270 10.1101/gr.1048803
Kothapalli R Yoder SJ Mane S Loughran TPJ Microarray results: how accurate are they? BMC Bioinformatics 2002 3 22 12194703 10.1186/1471-2105-3-22
Li J Pankratz M Johnson JA Differential gene expression patterns revealed by oligonucleotide versus long cDNA arrays Toxicol Sci 2002 69 383 390 12377987 10.1093/toxsci/69.2.383
Yang IV Chen E Hasseman JP Liang W Frank BC Wang S Sharov V Saeed AI White J Li J Lee NH Yeatman TJ Quackenbush J Within the fold: assessing differential expression measures and reproducibility in microarray assays Genome Biol 2002 3 research0062 12429061
Haas SA Hild M Wright AP Hain T Talibi D Vingron M Genome-scale design of PCR primers and long oligomers for DNA microarrays Nucleic Acids Res 2003 31 5576 5581 14500820 10.1093/nar/gkg752
Quackenbush J Microarray data normalization and transformation Nat Genet 2002 32 Suppl 496 501 12454644 10.1038/ng1032
Rosenow C Saxena RM Durst M Gingeras TR Prokaryotic RNA preparation methods useful for high density array analysis: comparison of two approaches Nucleic Acids Research 2001 29 U35 U42 10.1093/nar/29.22.e112
Dunman PM Murphy E Hanney S Palacio D Tucker-Kellogg G Wu S Brown EL Zagurski RJ Shlaes D Projan SJ Transcription Profiling-Based Identification of Staphylococcus aureus Genes Regulated by the agr and/or sarA Loci J Bacteriol 2001 183 7341 7353 11717293 10.1128/JB.183.24.7341-7353.2001
Tolstrup N Nielsen PS Kolberg JG Frankel AM Vissing H Kauppinen S OligoDesign: Optimal design of LNA (locked nucleic acid) oligonucleotide capture probes for gene expression profiling Nucleic Acids Res 2003 31 3758 3762 12824412 10.1093/nar/gkg580
Rouillard JM Zuker M Gulari E OligoArray 2.0: design of oligonucleotide probes for DNA microarrays using a thermodynamic approach Nucleic Acids Res 2003 31 3057 3062 12799432 10.1093/nar/gkg426
SantaLucia JJ A unified view of polymer, dumbbell, and oligonucleotide DNA nearest-neighbor thermodynamics. Proc Natl Acad Sci U S A 1998 95 1460 1465 9465037 10.1073/pnas.95.4.1460
Ketomaki K Hakala H Kuronen O Lonnberg H Hybridization properties of support-bound oligonucleotides: the effect of the site of immobilization on the stability and selectivity of duplex formation Bioconjug Chem 2003 14 811 816 12862435 10.1021/bc0340058
Gill SR Fouts DE Archer GL Mongodin EF Deboy RT Ravel J Paulsen IT Kolonay JF Brinkac L Beanan M Dodson RJ Daugherty SC Madupu R Angiuoli SV Durkin AS Haft DH Vamathevan J Khouri H Utterback T Lee C Dimitrov G Jiang L Qin H Weidman J Tran K Kang K Hance IR Nelson KE Fraser CM Insights on evolution of virulence and resistance from the complete genome analysis of an early methicillin-resistant Staphylococcus aureus strain and a biofilm-producing methicillin-resistant Staphylococcus epidermidis strain J Bacteriol 2005 187 2426 2438 15774886 10.1128/JB.187.7.2426-2438.2005
Genomic Research Laboratory http://www genomic ch/software php 2004
Vaudaux P Francois P Bisognano C Kelley WL Lew DP Schrenzel J Proctor RA McNamara PJ Peters G von Eiff C Increased expression of clumping factor and fibronectin-binding proteins by hemB mutants of Staphylococcus aureus expressing small colony variant phenotypes Infect Immun 2002 70 5428 5437 12228267 10.1128/IAI.70.10.5428-5437.2002
Talaat AM Howard ST Hale W Lyons R Garner H Johnston SA Genomic DNA standards for gene expression profiling in Mycobacterium tuberculosis Nucleic Acids Res 2002 30 e104 12384606 10.1093/nar/gnf103
|
15963225
|
PMC1183204
|
CC BY
|
2021-01-04 16:32:47
|
no
|
BMC Genomics. 2005 Jun 17; 6:95
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-95
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-991604278310.1186/1471-2164-6-99Methodology ArticleComparing independent microarray studies: the case of human embryonic stem cells Suárez-Fariñas Mayte [email protected] Scott [email protected] Michael [email protected] Ali [email protected] Marcelo O [email protected] Center for Studies in Physics and Biology, The Rockefeller University. 1230 York Ave, Box 212, New York, NY 10021, U.S.A2 Department Laboratory of Molecular Vertebrate Embryology, Camridge, The Rockefeller University. 1230 York Ave, Box 32, New York, NY 10021, U.S.A2005 22 7 2005 6 99 99 26 4 2005 22 7 2005 Copyright © 2005 Suárez-Fariñas et al; licensee BioMed Central Ltd.2005Suárez-Fariñas 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 studies of the same phenomenon in different labs often appear at variance because the published lists of regulated transcripts have disproportionately small intersections. We demonstrate that comparing studies by intersecting lists in this manner is methodologically flawed by reanalyzing three studies of the molecular signature of "stemness" in human embryonic stem cells. There are only 7 genes common to all three published lists, suggesting disagreement.
Results
Carefully reanalyzing all three together from the raw data we detect 111 genes upregulated and 95 downregulated in all three studies. The upregulated list was subject to rtRTPCR analysis and 75% of the genes were confirmed.
Conclusion
Our findings show that the three studies have a substantial core of common genes, which is missed if only the published lists are examined. Combined analysis of multiple experiments can be a powerful way to distil coherent conclusions.
==== Body
Background
A grave concern in the use of microarray technology is the apparently little agreement between different studies of the same phenomenon carried out in different laboratories, or even in the same lab using different platforms. Particularly evident is the "small intersection" problem: when several competing studies each conclude with a large list of "statistically significant" genes – yet the intersection between the published lists is ridiculously small. This problem undermines confidence in the technology[1] or, potentially worse, may misdirect the researchers into suspecting strange biological processes. When the subject matter is itself under contention then controversies may erupt [2-6]. The repeatability of microarray experiments across labs and platforms has become a hot current topic of discussion in the microarray community, with some studies suggesting that reproducibility across platforms is poor [7-9] while other studies indicate that biological and laboratory variability are larger sources of discrepancy than platforms or technologies [10-13], and with other studies pointing out annotations can be a source of problems[14].
We set out to study whether indeed different studies cannot be reconciled or whether the small intersection problem is artifactual, for one concrete problem of great scientific and extra-scientific importance: human embryonic stem cells (HESC). Isolation of HESC lines has generated the exciting possibility of both access to the basic science of human development as well as the possibility of new hope for cell-based therapy in the clinic[15]. The realization of these clinical goals depends on an understanding of the molecular basis of signaling pathways that maintain ESCs in an undifferentiated, pluripotent state, and microarray studies of HESCs should provide a sound basis for discovery and exploration of these pathways.
We shall consider 3 microarray studies of HESCs: Battacharya[16], Sperger[17] and Sato[18]. These studies had considerable differences, both in platforms (See Table 1) as well as biological. The Bhattacharya and Sperger studies have none or few replicates, hybridizing instead different cell lines to each chip, and contrast to different control samples of pools of differentiated tissues. The Sato study analyzed triplicates of a single cell line compared against "nonlineage induced differentiation" of that same cell line. Maintenance conditions also varied; see Methods for more details.
Table 1 Summary of platforms
Bhattacharya Sperger Sato
Platform cDNA - cDNA – 12 × 4 Affy – HGU133a
Chip design 8 × 4, 23 × 23 12 × 4, 30 × 30 650 × 650
Scanner GenePix 4000B GenePix 4000 GeneArray
Image Analysis Software GenePix 3.0 Gene Pix 3.0 MAS Suite
Spot/probes (per chip) 16 928 43200 / 43008 22 283
UniGene 12041 25 400 12 441
Empty spots per chip 238 531 -
Each study concluded with a list of genes which are up-regulated in stem cells. These lists are important because they should show a good view of all pathways in the cell for which there might be important stem-cell-specific functions, not just developmental signaling pathways, but also potential stem-cell-specific "housekeeping" genes. Most studies have a tendency to focus on developmentally important pathways, such as those that are sufficient for self-renewal; a genome-wide search giving us an unbiased list of up-regulated transcripts is supposed to give us a wider view into the state of stemness.
But the three lists of significantly up-regulated genes, as published, are quite different from each other. Their intersection is shown in Figure 1: seven genes appear in all three studies out of 2226 total genes in the union of the three published lists. This is particularly troublesome since all three studies appear to be technically reliable and each study has good reproducibility between replicates (see Methods: "Within platform variations"). The intersection highlights mainly metabolic or housekeeping genes; significantly, few of these have been implicated in any of the nine major developmentally important signaling pathways.
Figure 1 Common published genes. Intersection (per UniGene) between the published lists of up-regulated genes for each study.
The small intersection is a big problem: it is generally taken to mean that the results are not reproducible from laboratory to laboratory and should not be believed. We shall show that this is not the case, that comparing the lists in this manner is methodologically flawed and that the experiments do have a large core of common genes.
No experimental scientist would be surprised when results change upon a change of protocol; the analysis of microarray data, from image analysis to the statistical procedure, has as many parameters as a biochemical protocol, and hence there's little surprise that lists obtained by widely different "statistical protocols" are incompatible, so we shall now set about the task of defining and applying a single such statistical protocol to these studies.
Results
In an experiment addressing differential gene expression, a common criterion to choose a p-value is the largest tolerable amount of false positives; then the power of the test to detect the true positives will depend on the difference between condition and control, the replication variability, the number of replicates and other parameters. In order to be at the intersection between three lists, a gene must have passed such a test three independent times, which it does with a probability equal to the product of the probabilities of each test. Therefore the intersection's p-value is the cube of the lists' p-value. This strongly throws off the balance between sensitivity and specificity, with the result that the intersection becomes insensitive and carries few genes (See section D1 of Additional file 1 – Supplementary Discussion and Methods.). The lists which are good for publishing are inadequate for intersecting; in order to generate the best intersection the lists have to be recomputed.
We procured the raw data from all three studies, and re-analyzed it using consistent criteria for spot quality, normalization and summarization, in order to obtain the expression measures for the three studies. Since all three platforms are different (see Methods: "description of studies") they have many genes which are not in common and so could not possibly be at the intersection. We carefully screened UniGene numbers to obtain the set of 7373 genes common to all 3 microarray platforms, which is the "universe" of our study. Most of these genes show variations between stem lineage and differentiated control which are no larger than replication errors.
Eliminating them increases the power of our analysis and thus we kept for further study only those 2463 genes in this set that displayed variations across the samples, since only these could be differentially expressed genes. It should be emphasized that our universe does not include all of the genes possibly involved in stemness. For example some genes, like TDGF1 (Hs385870), Nanog (Hs329296), DNMT3B (Hs252613), FOXD3 (Hs424212), OTX2 (Hs288655), MyosinX (Hs481720), HEY2 (Hs144287), FGF4 (Hs1755), Rex1 (Hs335787) and Nodal (Hs370414), most of which were highlighted as up-regulated in HESCs in various studies, were not included in our "universe" because they are missing from at least one of the three different chip designs.
We then analyzed the genes by Integrated Correlation Analysis[19] (see Methods), which was introduced to validate the agreement among studies and to select genes that exhibit a coherent behavior across different studies. The idea is that while studying the same system, co-regulated genes should exhibit correlated expression profiles, and these correlations should be maintained across studies. This quality of moving together with other genes we call "coherence". Conversely, when extraneous factors affect a small set of genes in a particular study, the correlations between those genes and the rest shall not be maintained across studies.
Figure 2 illustrates how "coherence" of genes between two studies is quantified. For a pair of genes, the correlations between their expression profiles are calculated within each study and the correlation in one study is plotted against the other; points near the identity line represent pairs of genes which maintain their correlation values across studies. The coherence score of a gene g is the correlation coefficient of the points corresponding to all pairs containing g. Using all gene pairs we obtain the integrated correlation score, an overall measurement of agreement between the two studies. We have changed the nomenclature from [19], where coherence is called "consistency", because we want to reserve the notion of consistency to mean moving in the same direction from condition to control in all studies (i.e., up-regulated in all studies). Notice that no information about the condition and control is introduced in the definition of coherence; if the genes are both up-regulated in study 1 and both down-regulated in study 2 they are coherent because they behave the same way with respect to each other within each study, even though they are both inconsistent. (See Methods: Coherence and Consistency)
Figure 2 Coherence Scores. Given the expression profile for four genes A, B, C, and D in two studies 1 and 2 shown in panel (a), we can calculate the correlation of the expression profiles for every pair of genes in each of the two studies. In panel (b) those pairwise correlations are plotted for study 1 against study 2. For example, pair AB is positively correlated in both studies and pair BD is negatively correlated in study 1 and positively in study 2. Correlations for pairs AB, AC, AD lie approximately on a line of positive slope, so gene A is called "coherent". The coherence score of A is the correlation coefficient of those points, which is positive. Gene D has a negative correlation coefficient and so is incoherent. Study 3 in panel (c) is study 2 where condition and control have been swapped; all genes are perfectly coherent for studies 2 and 3, yet each gene which is up-regulated in study 2 is down-regulated in study 3.
Previous use of this analysis applied to classification of cancer [19] yielded unimodal histograms of the gene-coherence scores (there called "gene-reproducibility score"); in our study, though, we obtain a strongly bimodal histogram, shown in Figure 3, indicating that there is a number of genes in strong agreement among the three studies and a number of genes in strong disagreement. (Further details, including bivariate density plots, can be found in Additional file 1, section D4). There are a number of potential reasons why genes could be strongly incoherent as shown in these histograms discussed later.
Figure 3 Coherence Scores Distribution. The histograms of the coherence scores are bimodal. (a) Pairwise comparisons, and averaged score over all three comparisons. This implies that in these studies the genes can be divided into two distinct categories, coherent and incoherent. When "erratic" genes are discarded, there is a marked improvement in the agreement between studies. b) Integrated Correlation and Correlation of M-values calculated using the genes in the top percentiles of Coherence Score, red: Bhattacharya-Sperger, blue: Bhattacharya-Sato, black: Sperger-Sato
Eliminating incoherent genes improves enormously the general agreement between the studies. The improvement in the integrated correlation score and the correlation between log2-fold changes (M-values) is illustrated in Figure 3b. We decided to keep for further analysis the 739 genes in the top 30% of the gene-coherence score distribution. (Little variation of our final results was observed for percentiles 40% and 50% since they only include positive coherence scores; we present the stricter criterion). Not so obviously, the correlation between the M-values between studies also markedly improved from 0.35 to 0.76, 0.68, and 0.66 respectively. (More detailed results about Integrated Correlations results can be found in Additional file 1, section D4)
Once we restrict ourselves to this set of coherent genes, we study those genes that are up- or down- regulated in stem cells vs. their differentiated controls in each one of the studies. We emphasize that exactly the same statistical tests and criteria were applied to all three studies, with a strict cutoff value selection based both on a p-value and a positive lods [20]. We used the moderated t-statistics as proposed by [21] and the FDR (BH) procedure was used to adjust the p-values for multiple hypothesis [22]. This is important because the set of coherent genes is enriched in genes that are not statistically independent. The BH procedure [23] controls FDR under certain general assumptions (positive regression dependence) and simulations shows its adequacy to control FDR in more general dependence structures [24], while the BY procedure [25] is assumption-free but is more conservative that BH. However, in our case both procedures led to practically the same results, because the proportion of differentially expressed genes is higher in the coherent set; we report the results from the BH. P-value cutoff was set in 0.01, which implies than the probability of error is 10-4 in the pairwise comparison and 10-6 when the three studies are considered.
The intersection between the lists is now quite larger and statistically significant, as shown in Figure 4; the 111 up-regulated and 95 down-regulated genes common to all three studies (vs. 3 expected by chance intersections) are listed in Additional file 2 and Additional file 3 online. Notice that the 111 up-regulated genes in this list are not necessarily the "most" up-regulated for any individual study; yet they are significantly up-regulated for each study.
Figure 4 Intersection of significant genes. Improved intersection of significant genes. After our screening process, the number of transcripts which are up (a) and down (b) regulated in stem cells for the three studies. Notice that now most genes in each study also appear at least in one other study (81% Bhattacharya, 85% Sperger, 94% Sato, about 11% expected by chance), with a very important fraction common to all three (about 3 genes expected by chance at the intersection).
We performed real-time RT-PCR analysis of the up-regulated genes to validate our findings. In an initial experiment, H1 HESCs were differentiated for 30 days with RNA samples taken before and after differentiation. We succeeded in analyzing 106 genes out of the 111 intersecting HESC enriched genes in triplicate for expression level changes upon differentiation by real-time RT-PCR (See Additional file 4 for primers used in real-time RT-PCR). After screening for significant differences (within this experiment) between the mean normalized Ct values (Student's t-test, p > 0.05, n = 3), between differentiated and undifferentiated samples, 87 genes (82%) had differences in expression. Of these, 77 (89%) were enriched in undifferentiated HESCs and 10 (11%) were enriched in differentiated HESCs. Overall, 88 of the 106 tested genes (83%) were enriched in undifferentiated HESCs (p = 1.5 × 10-12 in the exact binomial test); 77 of them also significant by RT-PCR, for a 73% of success probability (p = 1.7 × 10-6). 9% were enriched in the differentiated sample and 18% were either not regulated or not significant in this experiment. (See Additional file 5 for the raw data of real-time RT-PCR results)
A comparison between the fold changes found through real-time RT-PCR and those obtained in the three studies is shown in Figure 5. Our real-time RT-PCR test was carried out under identical cell line, culture and differentiation conditions to the Sato study, so we present comparisons against the mean log-fold-change of the three studies and against Sato's study separately, which has for evident reasons better fit. Positive slopes and correlation coefficients are obtained in all comparisons; correlations are higher when fitting only to genes with real-time RT-PCR log2 fold-changes bigger than 0.5, becoming 0.6 when comparing only to Sato's study. These slopes and correlation coefficients are in line with general results in the literature comparing different microarray platforms and real-time RT-PCR results for the same RNA sample.
Figure 5 RT-PCR results. Comparison between the real-time RT-PCR results and the microarray results. The blue lines are linear fits (without intercept) through all the 106 genes, while the magenta lines fit only the 67 genes with a fold change bigger than 0.5 (log2-scale) in RT-PCR analysis (magenta points). (a) log2-fold change of RT-PCR vs. mean of all microarray studies (R2 = 0.5, r = 0.32, p < 10-16 for all genes, R2 = 0.83, r = 0.51, for the 67 top genes); (b) log2-fold change RT-PCR vs. Sato study alone, which had identical conditions to our study (R2 = 0.53, r = 0.41, p < 10-16 for all genes, R2 = 0.85, r = 0.59 for the 67 top genes)
Discussion
How can we reconcile the agreement in Figure 4 with the disagreement in Figure 1? Figure 4 displays an incorrect comparison, since intersecting lists in this manner is flawed methodology. The three lists were created from different sets of genes using different statistical criteria for ranking the transcripts; they really are apples and oranges and should not be compared so lightly. To mention but three exemplars of problems with such comparison:
- there are more than 40000 genes in Sperger's study, yet only 7373 are common to all three arrays, so about 4/5 of the genes studied by Sperger were not studied in some other study; in fact, more than 650 genes in the published list could not possibly have been at the intersection.
- the size of the intersection cannot be larger than the smallest list and is thus controlled by it; in this case the Bhattacharya study had far more stringent criteria and reported the fewest genes; notice that about 2/3rds of that list were in at least one of the other studies.
- different versions of annotation databases were used to create the lists, so some genes have identifiers that changed over time and are missed in an automated comparison; in fact, when a current version of the Affymetrix probeset annotations is used for the Sato study, the intersection increases more than twofold to 16 due to several genes having changed identifiers
These are just exemplars of the general problems of differences in gene universe (a), in gene identity (c), and in statistical criteria (b), which include not only the significance threshold used, but which particular test was employed to assess it (e.g. "moderated t-statistics with fdr corrections for multiple hypothesis"). Other general problems affecting list intersection also include the preprocessing steps, which have been shown to have a substantial impact in the agreement between platforms[11,14]. So, how should different experiments be compared? Treating them on an equal footing goes most of the way: this means procuring from the authors the raw data, analyzing all studies with exactly the same statistical methods and cutoffs, and working only on the universe of common genes. This still does not make the studies all apples, but it significantly approaches this goal. We further refined this by using the notion of gene coherence; in doing so we discovered that a number of genes are strongly incoherent, behaving erratically across the studies.
The incoherent genes may behave so due to environmental interactions, differences in cell culture system or in the differentiated states being compared. Gene expression is the nervous system of cells, imprinted by anything in the environment. A large number of genes are involved in environmental interaction, leading to variability between labs which obscure the common signal we are trying to detect. Among the gene categories that classify the top 50 incoherent genes (see Additional file 6), i.e. those that exhibited a significant fold change in respective studies but that were regulated in opposite directions among the studies, the top classifications were mainly involved with metabolic activity, such as genes involved in the response to oxidative stress and reactive oxygen specific genes. The incoherent genes may also reflect differences in the conditions that the labs use to grow cells. The HESC culture system has many poorly understood variables, such as batch to batch variation in the MEFs used to support the HESCs and the use of serum in some culture systems. These are both examples for which there are, as yet, no good predictors for the ability to support "stemness" and can currently only be screened by their ability to support HESC maintenance based on few markers. However the list also includes developmentally important genes, such as MID1(Hs27695), a possible left-right determination factor [26], and FGF9(Hs111), implicated in several developmental decisions. Also included were several members of the Wnt pathway, including APC(Hs158932), DIXDC1(Hs116796), and TLE4(Hs444213), so the incoherent genes may also reflect differences in the control samples used in the studies. The Sato group uses HESCs differentiated by withdrawal of CM, which may result in biases in the differentiated cell types produced, while the other groups use pooled panels of RNA from a variety of somatic tissues.
The application of microarray technology to the study of HESC biology has the potential to provide a robust foundation for the exploration of molecular networks underlying the state of "stemness" in human embryonic stem cells. In addition to the insight into their utility in clinical applications provided by this exploration, elucidation of these networks in HESCs shall be an important step towards defining the molecular activity driving development in early human embryos.
Our list of confirmed transcripts intersects three experiments carried out on different cell lines, maintained in the stemness state by different protocols, and compared against different differentiated states. That we do get a confirmed list under these circumstances bears witness that there is a well-defined set of molecular circuits involved in the state of stemness that can be studied regardless of variations in protocol or cell line. This analysis more than octuples the list of confirmed molecular markers of the stemness state in HESCs [27].
A balanced global assay of the molecular state of "stemness" should reflect all activities of the cell including those pathways that are necessary as well as those that are sufficient for the functions of an embryonic stem cell. Therefore, we should expect to find activities that might include, for example, those required for the defense of chromosomal or genetic stability as well as those required for self-renewal in an undifferentiated state. Indeed, the two most represented GO Biological Process in a GO/EASE analysis of the enriched genes within our universe are "Traversing start control point of mitotic cell cycle", indicating a prominent role of cell cycle regulators, and "transmembrane receptor protein serine/threonine kinase signaling", indicating a significant role for genes involved in sensing and responding to the developmental environment. Further, "transforming growth factor-beta receptor activity" was the most represented GO Molecular Function classification, indicating a prominent role for TGF-beta signaling in undifferentiated HESCs (See Additional file 7 and Additional file 8 for the complete GO analysis). However, because many of the genes that are important for these and other activities may be excluded from the universe analyzed here, we do not present the results as a complete or unbiased picture of these activities. Instead, this study addresses the validity of using microarray technology for building this foundation by applying consistent analysis to evaluate reproducibility.
Conclusion
Our study supports concluding that the three studies are compatible and repeatable. The method we've used demonstrates that we can harness the power of several labs to give weight to the intersection. In fact, we may conclude that a list of regulated transcripts from a single lab obtained under a restricted number of maintenance and differentiation protocols should be considered with some reserve, for we only know about coherent behavior of genes when we have several studies to draw from. Our results indicate that publication of the raw data is far more valuable than publication of the analyzed data, and further suggest that the field shall move forth only upon agreement to set standards of backgrounds and statistical methods.
Methods
(see Additional file 1, section Methods for a full description)
Description of the studies
The Bhattacharya study has 6 chips. Different HESC lines were hybridized to the red channel (Cy5) of the arrays; 5 of them were lines BG01, BG02, GE01, GE09, TE06, and the sixth sample was a pool of GE01, GE07, and GE09. The control sample, hybridized to the green channel was "total human universal RNA (huURNA) isolated from a collection of adult human tissues to represent a broad range of expressed genes from both male and female donors (BD Biosciences, Palo Alto, CA)". No replicates were performed for individual lines. The Sperger study used a similar design, hybridizing lines H1, H7, H13, H14 and two samples of H9. The control samples were "a common reference pool of mRNA". The Sato study had 6 Affymetryix HGU133A chips, 3 replicates of H1 cells (in Matrigel/Conditioned Medium) and 3 replicates of "nonlineage-directed differentiation" (Matrigel/non-CM). Table 1 summarizes the architecture of chips and the number of spot/genes involved in the three studies.
Language and packages
The statistical analysis was carried out in the R language version 2.0 , and packages were from the Bioconductor project . Gene Ontology analysis was carried out using EASE 2.0 software [28].
Raw data
Data from the Bhattacharya and Sato studies were obtained directly from the authors. The Sperger data were obtained through the Stanford Microarray Database [29]. cDNA array data were output files from GenePix 3.0. Affymetrix raw data files were .CEL files.
We used the same image analysis criteria in all CDNA arrays to exclude low quality spots (the criteria were different in the published studies).
Expression measures
The marray package from the Bioconductor suite was used for cDNA arrays. Normalization was executed in two steps, first within-print-tip-group location-dependent intensity normalization followed by within-print-tip group scale normalization using median absolute deviation. Single-channel normalization of two-color cDNA was done as proposed by [30], using quantile normalization. The GCRMA algorithm was used to summarize Affymetrix data as proposed in [31]. This algorithm improves the widely used RMA [32] by including an extra step to adjust for non-specific binding, and computing the sequence-specific affinities between probes as described [33].
Within-platform variability
In order to assess the quality of the data replications the within-platform variations were analyzed. The within-study reproducibility is overall fairly good in all the studies, even noting that Bhattacharya's and Sperger's design contain different lines of HESC rather than true replicates of a single line. See further details in Additional file 1.
Annotations
For both Bhattacharya and Sperger studies, annotations were obtained from SOURCE from the Stanford microarray data homepage[34]. For Affymetrix data, annotations packages from Bioconductor were used. The IMAGE clone IDs and the Affymetrix probes were matched using UniGene Cluster Annotation. Genes with no UniGene number were eliminated from the study. Expression values from spots or probesets with duplicated UniGene identifiers were averaged together.
Common genes
Annotations were obtained with the raw data from each study. Genes without UniGene identifier were eliminated and duplicated probes/spots were averaged together. After this process there are 7373 genes common to all 3 studies as shown in Supplementary Discussion Figure 8a. We filtered for evidence of variation across samples, reducing our set of interesting genes to those showed in Supplementary Discussion Figure 8b. For the cDNA arrays, we select genes where the M-values was bigger than 0.3 in at least 4 arrays and in Affymetrix experiment we keep genes whose expression profile had range bigger than 0.5.
Coherent genes: the integrated correlation approach
Integrated correlation analysis was introduced in [19]. For each study s, let us define xg the expression profile for a gene g, and the correlation for the pair of genes p = (g1, g2) in the study. Based on we can asset both overall coherence between studies and gene-specific coherence. The integrated correlation, defined as: quantified the coherence between studies. If this expression is calculated considering only the pairs containing a gene g, then we have a measure of the gene-specific coherence between two studies: , where p = (g, j) When more that two studies are involved, the average over all s and s' is used as a Coherence Score for a gene g, . Confidence Interval for the correlation scores were obtained by bootstrapping.
Coherence and consistency
In [19], the coherence score defined above is called reproducibility score and coherent genes (genes with high values of the score) are called "consistent". However we once again stress that this score bears no direct logical relationship to the notion of reproducibility or consistency in the sense of consistent up- or down- regulation in both studies. A simple counterexample makes the point: create a fake Study 3 which is Study 2 with the values of the condition and control swapped; then by the above definitions, all genes are perfectly coherent for studies 2 and 3, having coherence (reproducibility) scores equal to 1; yet each gene which is up-regulated in study 2 is down-regulated by the same amount in study 3, so all genes are inconsistent. A relationship between the coherence scores and consistent behaviour is predicated on the counter-reciprocal: if a pair of genes is incoherent then both genes cannot be consistent, and hence if a gene has a negative coherence (reproducibility) score it is "likely" (though by no means sure) to be inconsistent by being the "odd one out".
Differentially expression criteria
Statistical analysis to determine which genes are differentially expressed was carried out using the package Limma from the Bioconductor project. For assessing differential expression the moderated t-statistics was used as proposed by [21] in all the 3 studies. To do so, Limma uses an empirical Bayes method to moderate the standard errors of the estimated log2-fold changes. This results in more stable inference and improved power, especially for experiments with small number of arrays. The p-values of the moderated t-test were adjusted for multiple hypothesis testing, controlling the false discovery rate (fdr) as proposed [22]. We use a strict cut off criterion for selectivity of the genes based on both the p-values and lods ratio, as proposed in [20]. The lods (or B-statistic) is the log of the odds that the gene is differentially expressed. For the set of coherent genes we selected here, the cutoff of p = 0.01 and positive lods leads to 206 (111 up, 92 down) genes while 244 (139 up 105 down) is obtained considered only the p-value cut-off. In Supplementary Discussion and Methods the reader can find how the number of selected genes depends on the criteria for different set of coherent genes.
Real-time RT-PCR verification of gene expression level in HESCs
Relative expression levels of the 106 (95%) of 111 genes in the intersection were analyzed in H1 HESCs maintained in CM or differentiated by withdrawal of CM for 30 days by real-time RT-PCR. Raw data were normalized to Ubiquitin-C expression and relative expression levels were determined using PCR efficiency-adjusted ratios [35].
Authors' contributions
MSF and MOM are the analytical members of the team while SN, MH, AHB are the biological team. Conception of the work came from join discussion between MSF and AHB. MSF and MOM are responsible for analysis and methodology. MH designed the primers for the RTPCR experiment, carried out by SN and AHB; the latter are also responsible for the biological assessment of all analytical results.
Supplementary Material
Additional File 1
Supplementary Discussion and Methods.
Click here for file
Additional File 2
Up-regulated genes in the intersection. List of up-regulated genes in the intersection of the 3 studies. In html format, including annotations and links.
Click here for file
Additional File 3
Down-regulated genes in the intersection. List of down-regulated genes in the intersection of the 3 studies. In html format, including annotations and links.
Click here for file
Additional File 4
RT-PCR primer sequences used. Primer sequences used in real-time RT-PCR.
Click here for file
Additional File 5
real RT-PCR results.
Click here for file
Additional File 6
Top 50 incoherent genes. List of the top 50 genes with the most negative coherence score.
Click here for file
Additional File 7
Up-regulated genes – EASE analysis. EASE analysis for the up-regulated genes in the intersection of the three studies
Click here for file
Additional File 8
Down-regulated genes – EASE analysis. EASE analysis for the up-regulated genes in the intersection of the three studies
Click here for file
Acknowledgements
M.S.F. acknowledges a Woman in Science fellowship from Rockefeller University.
We would like to acknowledge WiCell, the provider of the stem cell line H1.
==== Refs
Marshall E Getting the noise out of gene arrays SCIENCE 2004 306 630 631 15499004 10.1126/science.306.5696.630
Evsikov AV Solter D Comment on " 'Stemness': Transcriptional Profiling of Embryonic and Adult Stem Cells" and "A Stem Cell Molecular Signature" (II) SCIENCE 2003 302 393c 14563991 10.1126/science.1082380
Fortunel NO Otu HH Ng H-H Chen J Mu X Chevassut T Li X Joseph M Bailey C Hatzfeld JA Hatzfeld A Usta F Vega VB Long PM Libermann TA Lim B Comment on " 'Stemness': Transcriptional Profiling of Embryonic and Adult Stem Cells" and "A Stem Cell Molecular Signature" (I) SCIENCE 2003 302 393b 14563990 10.1126/science.1086384
Ivanova N Dimos J Schaniel C Hackney J Moore K Lemischka I A stem cell molecular signature SCIENCE 2002 298 601 604 12228721 10.1126/science.1073823
Ivanova NB Dimos JT Schaniel C Hackney JA Moore KA Ramalho-Santos M Yoon S Matsuzaki Y Mulligan RC Melton DA Lemischka IR Response to Comments on " 'Stemness': Transcriptional Profiling of Embryonic and Adult Stem Cells" and "A Stem Cell Molecular Signature" SCIENCE 2003 302 393d 10.1126/science.1088249
Ramalho-Santos M Yoon S Matsuzaki Y Mulligan R Melton D "Stemness": Transcriptional profiling of embryonic and adult stem cells SCIENCE 2002 298 597 600 12228720 10.1126/science.1072530
Kothapalli R Yoder S Mane S Loughran T Microarray results: how accurate are they? BMC BIOINFORMATICS 2002 3 22 12194703 10.1186/1471-2105-3-22
Kuo W Jenssen T Butte A Ohno-Machado L Kohane I Analysis of matched mRNA measurements from two different microarray technologies BIOINFORMATICS 2002 18 405 412 11934739 10.1093/bioinformatics/18.3.405
Li J Pankratz M Johnson J Differential gene expression patterns revealed by oligonucleotide versus long cDNA arrays TOXICOL SCI 2002 69 383 390 12377987 10.1093/toxsci/69.2.383
Barczak A Rodriguez M Hanspers K Koth L Tai Y Bolstad B Speed T Erle D Spotted long oligonucleotide arrays for human gene expression analysis GENOME RES 2003 13 1775 1785 12805270 10.1101/gr.1048803
Irizarry RA Warren D Spencer F Kim IF Biswal S Frank BC Gabrielson E Garcia JGN Geoghegan J Germino G Griffin C Hilmer SC Hoffman E Jedlicka AE Kawasaki E Martinez-Murillo F Morsberger L Lee H Petersen D Quackenbush J Scott A Wilson M Yang YQ Ye SQ Yu W Multiple-laboratory comparison of microarray platforms (vol 2, pg 345, 2005) Nature Methods 2005 2 477 477 10.1038/nmeth0605-477b
Yauk C Berndt M Williams A Douglas G Comprehensive comparison of six microarray technologies NUCLEIC ACIDS RES 2004 32 e124 15333675 10.1093/nar/gnh123
Yuen T Wurmbach E Pfeffer R Ebersole B Sealfon S Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays NUCLEIC ACIDS RES 2002 30 e48 12000853 10.1093/nar/30.10.e48
Mecham B Klus G Strovel J Augustus M Byrne D Bozso P Wetmore D Mariani T Kohane I Szallasi Z Sequence-matched probes produce increased cross-platform consistency and more reproducible biological results in microarray-based gene expression measurements NUCLEIC ACIDS RES 2004 32 e74 15161944 10.1093/nar/gnh071
Thomson JA Itskovitz-Eldor J Shapiro SS Waknitz MA Swiergiel JJ Marshall VS Jones JM Embryonic stem cell lines derived from human blastocysts Science 1998 282 1145 1147 9804556 10.1126/science.282.5391.1145
Bhattacharya B Miura T Brandenberger R Mejido J Luo Y Yang A Joshi B Ginis I Thies R Amit M Lyons I Condie B Itskovitz-Eldor J Rao M Puri R Gene expression in human embryonic stem cell lines: unique molecular signature BLOOD 2004 103 2956 2964 15070671 10.1182/blood-2003-09-3314
Sperger JM Chen X Draper JS Antosiewicz JE Chon CH Jones SB Brooks JD Andrews PW Brown PO Thomson JA Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors Proc Natl Acad Sci USA 2003 100 13350 13355 14595015 10.1073/pnas.2235735100
Sato N Sanjuan I Heke M Uchida M Naef F Brivanlou A Molecular signature of human embryonic stem cells and its comparison with the mouse DEV BIOL 2003 260 404 413 12921741 10.1016/S0012-1606(03)00256-2
Parmigiani G Garrett-Mayer E Anbazhagan R Gabrielson E A cross-study comparison of gene expression studies for the molecular classification of lung cancer CLINICAL CANCER RESEARCH 2004 10 2922 2927 15131026
Lonnstedt I Speed T Replicated microarray data STATISTICA SINICA 2002 12 31 46
Smyth GK Linear models and empirical Bayes methods for assessing differential expression in microarray experiments Statistical Applications in Genetics and Molecular Biology 2004 3 Article 3.
Dudoit S Shaffer J Boldrick J Multiple hypothesis testing in microarray experiments Statistical Science 2003 18 71 103 10.1214/ss/1056397487
Benjamini Y Hochberg Y Controlling the False Discovery Rate – a Practical and Powerful Approach to Multiple Testing Journal of the Royal Statistical Society Series B-Methodological 1995 57 289 300
Reiner A Yekutieli D Benjamini Y Identifying differentially expressed genes using false discovery rate controlling procedures BIOINFORMATICS 2003 19 368 375 12584122 10.1093/bioinformatics/btf877
Benjamini Y Yekutieli D The control of the false discovery rate in multiple testing under dependency Annals of Statistics 2001 29 1165 1188 10.1214/aos/1013699998
Granata A Quaderi N The Opitz syndrome gene MID1 is essential for establishing asymmetric gene expression in Hensen's node DEV BIOL 2003 258 397 405 12798296 10.1016/S0012-1606(03)00131-3
Brivanlou AH Gage FH Jaenisch R Jessell T Melton D Rossant J STEM CELLS: Enhanced: Setting Standards for Human Embryonic Stem Cells SCIENCE 2003 300 913 916 12738841 10.1126/science.1082940
Hosack D Dennis G Sherman B Lane H Lempicki R Identifying biological themes within lists of genes with EASE Genome Biol 2003 4 R70 14519205 10.1186/gb-2003-4-10-r70
Website
Yang YH Thorne N Goldstein DR Normalization for Two-color cDNA Microarray Data Science and Statistics: A Festschrift for Terry Speed 2003 40 403 418 Series ILN-M (Series Editor): Science and Statistics: A Festschrift for Terry Speed
Wu Z Irizarry R Gentleman R Martinez Murillo F Spencer F A Model Based Background Adjustement for Oligonucleotide Expression Arrays Journal of American Statistical Association 2004 99 909 917 10.1198/016214504000000683
Irizarry RA Bolstad BM Collin F Cope LM Hobbs B Speed TP Summaries of Affymetrix GeneChip probe level data Nucl Acids Res 2003 31 e15 12582260 10.1093/nar/gng015
Naef F Magnasco M Solving the riddle of the bright mismatches: Labeling and effective binding in oligonucleotide arrays PHYSICAL REVIEW E 2003 68
Website
Pfaffl MW A new mathematical model for relative quantification in real-time RT-PCR Nucleic Acids Res 2001 29 e45 11328886 10.1093/nar/29.9.e45
|
16042783
|
PMC1183205
|
CC BY
|
2021-01-04 16:32:48
|
no
|
BMC Genomics. 2005 Jul 22; 6:99
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-99
|
oa_comm
|
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-501601181110.1186/1472-6963-5-50Research ArticleWhat do we know about how to do audit and feedback? Pitfalls in applying evidence from a systematic review Foy R [email protected] MP [email protected] G [email protected] J [email protected] JM [email protected] R [email protected] Centre for Health Services Research, University of Newcastle upon Tyne, Newcastle upon Tyne, United Kingdom2 Department for Health Technology Assessment, Reviews and Dissemination, Norwegian Health Services Research Centre, Oslo Norway3 Surgical Outcomes Research Centre, Central Sydney Area Health Service and University of Sydney, Royal Prince Alfred Hospital, Camperdown Australia4 Clinical Epidemiology Programme, Ottawa Health Research Institute, Ottawa Canada5 Department of Health Sciences, University of Leicester Leicester, UK2005 13 7 2005 5 50 50 19 4 2005 13 7 2005 Copyright © 2005 Foy et al; licensee BioMed Central Ltd.2005Foy 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
Improving the quality of health care requires a range of evidence-based activities. Audit and feedback is commonly used as a quality improvement tool in the UK National Health Service [NHS]. We set out to assess whether current guidance and systematic review evidence can sufficiently inform practical decisions about how to use audit and feedback to improve quality of care.
Methods
We selected an important chronic disease encountered in primary care: diabetes mellitus. We identified recommendations from National Institute for Clinical Excellence (NICE) guidance on conducting audit and generated questions which would be relevant to any attempt to operationalise audit and feedback in a healthcare service setting. We explored the extent to which a systematic review of audit and feedback could provide practical guidance about whether audit and feedback should be used to improve quality of diabetes care and, if so, how audit and feedback could be optimised.
Results
National guidance suggests the importance of securing the right organisational conditions and processes. Review evidence suggests that audit and feedback can be effective in changing healthcare professional practice. However, the available evidence says relatively little about the detail of how to use audit and feedback most efficiently.
Conclusion
Audit and feedback will continue to be an unreliable approach to quality improvement until we learn how and when it works best. Conceptualising audit and feedback within a theoretical framework offers a way forward.
==== Body
Background
A range of strategies exist to promote the uptake of clinical research findings into the routine care of patients. They seek to change the behaviour of healthcare professionals and thereby improve the quality of patient care (Table 1). For each of these strategies a number of trials of their effectiveness have been drawn together within systematic reviews.[1,2] By examining interventions in a range of settings and circumstances such reviews aim to produce generalisable messages about the effectiveness or otherwise of these interventions.
Table 1 Examples of interventions to promote professional behaviour change.
Educational outreach visits A personal visit by a trained person to a health care provider in his or her own setting
Reminders (manual or computerised] Prompts performance of a patient specific clinical action
Interactive educational meetings Participation of health care providers in workshops that include discussion or practice
Audit and feedback Any summary of clinical performance over a specified period of time
Local opinion leaders Health professionals nominated by their colleagues as being educationally influential
Local consensus process Inclusion of professionals in discussions to agreed the approach to managing a clinical problem that they have selected as important
Patient mediated interventions Specific information sought from or given to patients
Educational materials Distribution of recommendations for clinical care (such as clinical practice guidelines, audio-visual materials, electronic publications).
Didactic educational meetings Lectures with minimal participant interaction
Financial incentives payments directly rewarding health care providers for specified behaviours
Multifaceted interventions A combination of two or more interventions
All healthcare systems are concerned with improving the quality of care that they deliver as demonstrated by their establishment of structures (such as the UK NHS National Institute for Clinical Excellence (NICE], the Australian National Institute for Clinical Studies) and high profile reports.[3] Across countries clinical audit (hereafter referred to as audit and feedback) is commonly used to both monitor and improve quality of care. [4,5]
The strategies in Table 1 vary considerably in their resource requirements and cost effectiveness and any healthcare system will have finite resources to commit to quality improvement activities. Therefore to make the best use of health service resources, interventions to change professional behaviour should be evidence-based, selected on the basis of their known effectiveness and efficiency, and should be directed towards important clinical conditions.
While rising prevalence and changing patterns of service delivery diabetes mellitus increasingly contributes to the primary care workload [6] and there is evidence of fragmented and variable provision of care.[7] This paper explores the utility of current systematic review evidence to support healthcare system decisions about how to provide evidence based audit and feedback to improve the quality of care by considering it in the context of a common chronic condition and setting – diabetes mellitus in primary care. We aimed to find out whether we could operationalise audit and feedback from existing review data.
Methods
Topic selection
The UK NHS has produced a framework and set of measurable criteria by which to judge the quality of care for patients with diabetes mellitus. The National Service Framework (NSF) for Diabetes was launched in 2002.[8] It suggests performance targets for primary care organisations, responsible for the commissioning and provision of health care for defined populations. Some of these targets have been incorporated into the revised contract for UK primary care doctors (GPs) reflecting disease monitoring (e.g. HbA1c measurement) or secondary prevention (e.g. proportion of patients with HbA1c under 7.5%).[9] Therefore, diabetes represents an appropriate condition with which to explore the utility of audit: it is a common condition with important consequences, effective interventions are available, measurable outcomes have been defined, and there is potential for improvement in the quality of care.
How best to conduct audit and feedback?
We informed the study with two definitions of audit and feedback (Table 2). The systematic review [5] offers a narrower definition than the National Institute for Clinical Excellence, Principles for Best Practice in Clinical Audit [4] which offers a broader definition and stresses the importance of integrating audit within an overall quality improvement framework. The latter sets out practical considerations for five stages of the audit and feedback process: preparing for audit, selecting criteria, measuring performance, making improvements, and sustaining improvement (Table 3). Much emphasis is given to creating the right organisational structures and culture for success, as well as taking account of local knowledge, experience and skills. Both are relevant to quality improvement at an organisational as well as individual level. However, neither describes in detail the manner in which audit and feedback should be conducted.
Table 2 Definitions of audit
Definition of audit endorsed by the National Institute for Clinical Excellence [4]
A quality improvement process that seeks to improve patient care and outcomes through systematic review of care against explicit criteria and the implementation of change. Aspects of the structure, processes, and outcomes of care are selected and systematically evaluated against explicit criteria. Where indicated, changes are implemented at an individual, team, or service level and further monitoring is used to confirm improvement in healthcare delivery.
Definition of audit used by the Cochrane systematic review [5]
The provision of any summary of clinical performance over a specified period of time. The summary may include data on processes of care (e.g. number of diagnostic tests ordered), clinical endpoints (e.g. blood pressure readings), and clinical practice recommendations (proportion of patients managed in line with a recommendation).
Table 3 Guiding principles for clinical audit.[4]
Stage Recommendations Addressed within Cochrane Review?
Preparing for audit Securing stake-holder interest and involvement (e.g. professionals, patients or carers) No
Selection of appropriate topic, according to whether:
• Topic concerned is of high cost, volume, or risk to staff or users No
• Evidence of a serious quality problem Yes: effects greater if low baseline
• Good evidence available to inform quality standards No
• Amenability of problem to change No
• Potential for involvement in a national audit project No
• Topic is pertinent to national policy initiatives No
• Topic is a priority for the organisation No
Clear definition of purpose of audit, e.g. to improve or ensure the quality of care No
Provision of necessary support structures, i.e.
• Structured audit programme (committee structure, feedback mechanisms, and regular audit meetings) No
• Sufficient funding (audit staff, time of clinical staff, data collection, feedback) No
Identification of skills and people needed to carry out the audit No
Selecting criteria Definition of criteria (structure, process and outcome) No
Validity and potential to lead to improvements in care
• Evidence based No
• Related to important aspects of care No
• Measurable Yes (implicitly)
Measuring level of performance Planning data collection
• Definition of user group (and exceptions) Can't tell
• Definition of healthcare professionals involved Yes (implicitly)
• Definition of time period over which criteria apply Yes (implicitly)
Making improvements Identification of barriers to change No
Implementing change
• Establishing the right environment (at individual, team and organisational levels) No
• Considering external relationships (e.g. with patients or other agencies) No
• Use of other supporting interventions (e.g. educational outreach, reminders) and / or multifaceted interventions Yes: not supported by evidence
Sustaining improvement Monitoring and evaluating changes, e.g. continuing audit cycle, use of performance indicators No
• Appropriate organisational development (e.g. cultural change, adequate training) No
• Use of existing strategic, organisational or clinical frameworks No
• Leadership No
We explored the extent to which a systematic review of audit and feedback could provide practical guidance about whether audit and feedback should be used to improve quality of diabetes care and, if so, how audit and feedback could be optimised. Based upon discussions with those responsible for conducting audit and feedback at a local level as well as our own experiences of doing so, we identified several questions which would be relevant to any attempt to operationalise audit and feedback in a healthcare service setting.
• Does audit and feedback work for this condition and setting, specifically improving the care of patients with a chronic disease – diabetes mellitus – in primary care?
• Does it work equally across all dimensions of care – from simple recording of cardiovascular risk factors to more complex areas of care such as glycaemic control? The latter requires a greater number of actions to achieve which include measuring blood glucose levels, reviewing the patient, checking compliance with drug and dietary therapies and checking patients' understanding of the condition.
• How should it be prepared? Should data be comparative and if so, what should the comparator group be? Should data be anonymised?
• How intensive should feedback be? Intuitively, providing more and personalised feedback on a recurrent and regular basis should have a greater impact on practice than a one-off report of (say) PCT-level aggregated data. However, it is uncertain whether the extra time and costs of ongoing data collection and preparing more frequent feedback would be matched by additional benefits.
• How should it be delivered – by post or by a messenger in person? And if by a messenger who should this be? Professionals might be more convinced by a message delivered by a colleague with a recognised interest in diabetes care rather than a non-clinical facilitator.
• What activities, if any, should accompany feedback? The likely costs and possible benefits of (say) educational meetings or outreach visits need to be weighed up against providing feedback via paper or computerised formats alone.
• What should be done about the poorest performers detected by the audit? Targeting such practices may help close the gap between the poorest and best performers. Alternatively, spreading effort to improve quality more equally amongst all practices may improve average performance for the whole PCT.
Results
The evidence from the systematic review
We identified a systematic review of audit and feedback that identified and appraised 85 randomised trial.[5] Audit and feedback was used for a wide range of clinical topics and problems. The review conclusions were:
• audit and feedback can improve professional practice, although the effects are generally small to moderate
• effectiveness varies substantially among different studies
• variation may be related to different methods of providing feedback or contextual factors, such as targeted behaviours and professionals
The review identified only five direct (head-to-head) comparisons of different methods of providing feedback (Table 4). One comparison suggested that feedback by a peer was more effective than that by a neutral observer [10]; another that feedback from a peer physician was no more effective than that from a nurse.[11] The other three comparisons found no effects related to recipients (group or individual) or content of feedback. None of these studies reported an economic evaluation.
Table 4 Evidence for questions addressed by the Cochrane Review.
Questions Most relevant analyses from Cochrane Review Evidence from all trials reviewed (n = 85) Evidence from chronic disease management trials (n = 15) Evidence from trials of diabetes care (n = 4)
Does audit and feedback work? Any intervention involving audit and feedback versus no intervention +/- educational materials 83 comparisons: for dichotomous outcomes, median adjusted relative risk (RR) of non-compliance was 0.85 [Interquartile range (IQR) 0.74 to 0.96]* Small to moderate effects in 11 of 19 comparisons Moderate to large effects in two comparisons [12;13]
Audit and feedback versus other interventions Five comparisons: two show audit and feedback more effective than reminders; one that local opinion leaders more effective; one no effect over patient education; one no effect of audit and feedback with educational meetings over educational meetings alone Small effect of audit and feedback over reminders from one comparison None
Does it work equally across all dimensions of care? No direct comparisons; exploration of heterogeneity No heterogeneity explained by complexity of the targeted behaviour None None
How should it be prepared? Should data be comparative and if so, what should the comparator group be? Should data be anonymised? Content. Patient information, such as blood pressure or test results, compliance with a standard or guideline, or peer comparison; versus information about costs or numbers of tests ordered or prescriptions Two comparisons: no difference between peer comparison and individual feedback without peer comparison; nor between feedback on medication and feedback on performance No difference between feedback on medication versus feedback on performance in one comparison None
How intensive should feedback be? Recipients. Individual or group No difference between individual versus group feedback in one comparison None None
Frequency. Once only or more frequent feedback None None None
Length. Once only feedback versus audit and feedback over a period of time None None None
Short term effects compared to longer term effects after audit and feedback stops Mixed results from 11 comparisons No difference from one comparison [14] No difference from one comparison [14]
Exploration of heterogeneity No heterogeneity explained by intensity of audit and feedback
Questions Most relevant analyses from Cochrane Review Evidence from all trials reviewed (n = 85) Evidence from chronic disease management trials (n = 15) Evidence from trials of diabetes care (n = 4)
How should it be delivered – by post or by a messenger in person? And if by a messenger who should this be? Format. Verbal, written or both None None None
Source. Influential source [seen to be credible and trustworthy by the professional] or feedback from any other source Two comparisons: peer feedback better than non-physician observer feedback; no difference between peer physician versus nurse feedback No difference between peer physician versus nurse feedback in one comparison [11] No difference between peer physician versus nurse feedback in one comparison [11]
What activities, if any, should accompany feedback? Audit and feedback with complementary interventions versus audit and feedback alone No clear effect of complementary interventions from 14 studies including various comparisons except for small effect of audit and feedback combined with educational outreach. Lower baseline compliance associated with larger effect sizes. Small or mixed effects in two out of four comparisons Outreach by peer or nurse more effective than feedback alone [11]
What should be done about the poorest performers detected by the audit? None None None None
*Relative risk [RR] is given for non-compliance. Therefore a lower RR is equivalent to greater effect size.
The review also evaluated 14 direct comparisons of audit and feedback alone compared to audit and feedback combined with other interventions (multifaceted interventions). There was no evidence that multifaceted interventions worked better than audit and feedback alone. A multivariate analysis explored potential causes of heterogeneity in the results (study quality, whether audit and feedback was combined with other interventions, intensity of feedback, complexity of the targeted behaviour, and level of baseline compliance). Only low baseline compliance was associated with greater effect sizes for multifaceted interventions. There was no evidence of larger effects with increasing intensity of feedback.
The evidence for chronic disease management
Fifteen studies relate to chronic disease management (hypertension, diabetes, cholesterol control, depression, asthma and end-stage renal failure). Just over half of comparisons indicated that audit and feedback was more effective than doing nothing (Table 4). Using multifaceted interventions or modifying feedback methods did not enhance effectiveness.
The evidence for diabetes care
Four studies evaluated audit and feedback in diabetes care, three set in primary care. Two comparisons addressed one of our key questions (Does audit and feedback work for this condition and setting?) and showed that audit and feedback, with or without other interventions, was more effective than doing nothing.[12,13] A UK primary care study.[12] showed that a multifaceted intervention incorporating low intensity audit and feedback moderately improved practice, specifically recording of key variables (e.g. glycaemic control, smoking habit). Audit and feedback also moderately increased US primary care physician compliance with guidelines.[13]
Two studies partially addressed three of our key questions about how to conduct audit and feedback (How intensive should feedback be? What activities, if any, should accompany feedback? How should feedback be delivered?). In US secondary care, there was no difference between continuing feedback against withdrawal of feedback in the accuracy of capillary blood glucose monitoring.[14] An Australian study of GPs found a small benefit of feedback given by a doctor or nurse compared with feedback alone, although it is difficult to judge whether the benefits of this approach outweighed the additional costs.[11] There was no difference in effect size between doctor and nurse feedback in this comparison. There was no relationship between study effect size and feedback intensity, co-intervention use or complexity of targeted behaviour across the four studies.
Discussion
The review evidence was of limited use in informing the operationalisation of evidence based audit and feedback. A number of issues contributed to this – the heterogeneity of the studies in the overall review, the problems of interpreting sub-groups of studies within the larger review, and the lack of direct evidence (particularly from head-to-head comparisons) to answer key questions.
It is unclear how to use the review to extract generalisable lessons about how audit and feedback achieves its effects. For example, individual level feedback could reasonably be assumed to be more personally relevant and persuasive and thus more effective than feedback at a group level; there are no such direct comparisons available. Four out of 10 studies using individual feedback for chronic disease management reported no effect whilst both studies using group feedback reported positive effects. Therefore group feedback might be more effective than individual feedback, possibly by promoting peer pressure, consensus and subsequent action. Unfortunately this must all remain conjecture given the paucity of data to test different hypotheses about the causal mechanisms that make audit and feedback work.
Based upon a limited number of comparisons, audit and feedback appears to work better for diabetes than for other conditions. It is unclear whether this is because there is something intrinsically different about diabetes (or the audit methods used in diabetes) compared to other conditions, or whether this is an unreliable sub-group analysis of four studies selected from the 85 available. It is hard to have confidence in the findings of the diabetes studies in the absence of good, preferably a priori, reasons as to why these studies should be examined separately from others in the review.
Similar pitfalls exist in judging the relative effectiveness of different feedback methods. This is mainly because of the limited number of head-to-head comparisons comparing audit and feedback alone against combined interventions or variations in providing feedback. Across all studies, audit and feedback alone appears similarly effective to multifaceted strategies. However, the lack of difference in effect size could have occurred because multifaceted strategies were used in situations where investigators judged them necessary to overcome greater obstacles to improving care. In the absence of primary studies, the review cannot address some of the key questions such as whether intensive feedback would improve more complex outcomes (gylcaemic control) at an acceptable cost.
Mapped back onto the principles for good clinical audit, the evidence only supports doing audit if there is low baseline compliance (Table 3). This evidence relates to situations where there is low mean baseline compliance across all study physicians rather than relating solely to a selected “low compliance” group. Thus it is of no direct relevance to the key question of whether or not audit and feedback can promote change in poorly performing individuals. However, a baseline audit of multiple aspects of diabetes care would enable targeting of implementation activities at areas of low compliance.
The issues of external validity of randomised controlled trials (and by inference, systematic reviews) have been aired in the context of clinical studies.[15,16]. However, what we have had to deal with here is more to do with inadequate description of the interventions in the primary studies and an inadequate understanding of the causal mechanisms by which the intervention or its variants might exert their effects. Thus this lack of fundamental understanding accounts for the impossibility of assessing a behaviour change interventions' applicability to a particular service setting. We are a long way from being able to do what is now commonplace with clinical studies in terms of assessing the applicability of a clinical study to an individual patient.[17]
A rational approach to this situation is to develop a conceptual framework within which to describe common elements of settings, individuals, targeted behaviours and interventions [18-20]. This would enable the identification of features that systematically influence the effectiveness of interventions. For example, the effectiveness of audit and feedback may be influenced by factors such as health professionals' motivation to change or perceived peer pressure – generalisable concepts that can be used across different contexts. Behavioural theory can identify potentially modifiable factors underlying professional behaviour in order to identify those processes to target with an intervention. Hence, if perceived peer pressure was predictive of adherence to good practice criteria, feedback incorporating peer comparison might enhance effectiveness. This approach potentially offers a method for more effective selection and development of interventions to improve practice. The longer term possibility is to establish a theoretically grounded basis for selecting or tailoring interventions given specific barriers and circumstances. This would apply to all behaviour change strategies, not just audit and feedback.
Conclusion
Review evidence was of limited use in informing the operationalisation of evidence based audit and feedback. This is mainly because of the heterogeneity of the studies in the overall review, the problems of interpreting sub-groups of studies within the larger review, and the lack of head-to-head comparisons to answer key questions. Audit and feedback will continue to be an unreliable approach to quality improvement until we learn how and when it works best. Conceptualising audit and feedback within a theoretical framework offers a way forward.
Abbreviations
UK NHS United Kingdom National Health Service
NSF National Service Framework
GP General Practitioner
NICE National Institute for Clinical Excellence
HbA1c Glycosylated haemoglobin
PCT Primary Care Trust
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GJ and JY undertook the Cochrane Review of audit and feedback. RB co-authored Principles for Best Practice in Clinical Audit. ME suggested the idea for this paper. RF wrote the first draft and is guarantor. All other authors helped draft the manuscript and have approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Grimshaw JM Thomas RE MacLennan G Fraser C Ramsay CR Vale L Whitty P Eccles MP Matowe L Shirran L Wensing M Dijkstra R Donaldson C Effectiveness and efficiency of guideline dissemination and implementation strategies Health Technol Assess 2004 8 1 72 14960256
NHS Centre for Reviews and Dissemination Getting evidence into practice Effective Health Care 1999 University of York
Crossing the quality chasm: the IOM Health Care Quality Initiative Institute of Medicine 2001
National Institute for Clinical Excellence Principles for Best Practice in Clinical Audit 2002 Abingdon: Radcliffe Medical Press
Jamtvedt G Young JM Kristoffersen DT O'Brien MA Oxman AD Audit and feedback: effects on professional practice and health care outcomes The Cochrane Library 2004 Oxford: Update Software
National Service Framework for Diabetes 2001 London, Department of Health
Audit Commission Testing Times: A Review of Diabetes Services in England and Wales 2000 London, Audit Commission
Department of Health National Service Framework for Diabetes. Delivery Strategy 2002 London, Department of Health
The NHS Confederation & British Medical Association 2003 New GMS Contract. London
Hombergh Pvd Grol R Hoogen HJMvd Bosch WJHNvd Practice visits as a tool in quality improvement: mutual visits and feedback by peers compared with visits and feedback by non-physician observers Quality in Health Care 1999 161 166 10847872
Ward A Kamien M Mansfield F Fatovich B Educational feedback in management of diabetes in general practice Education for General Practice 1996 7 142 150
Feder G Griffiths C Highton C Eldridge S Spence M Southgate L Do clinical guidelines introduced with practice based education improve care of asthmatic and diabetic patients? A randomised controlled trial in general practice BMJ 1995 311 1473 1478 8520339
Lobach DF Electronically distributed, computer-generated, individualized feedback enhances the use of a computerized practice guideline Proc AMIA Annu Fall Symp 1996 493 497 8947715
Jones HE Cleave B Zinman B Szalai JP Nichol HL Hoffman BR Efficacy of feedback from quarterly laboratory comparison in maintaining quality of a hospital capillary blood glucose monitoring program Diabetes Care 1996 19 168 170 8718440
Black N Why we need observational studies to evaluate the effectiveness of health care BMJ 1996 312 1215 1218 BMJ 1996; 312:1215-1218. 8634569
McKee M Britton A Black N McPherson K Sanderson C Bain C Methods in health services research: Interpreting the evidence: choosing between randomised and non-randomised studies BMJ 1999 319 312 315 10426754
Dans AL Dans LF Guyatt GH Richardson S Users' guides to the medical literature: XIV. How to decide on the applicability of clinical trial results to your patient. Evidence-Based Medicine Working Group JAMA 1998 279 545 549 9480367 10.1001/jama.279.7.545
Foy R Eccles M Grimshaw J Why does primary care need more implementation research? Family Practice 2001 18 353 355 11477039 10.1093/fampra/18.4.353
Walker AE Grimshaw J Johnston M Pitts N Steen N Eccles M PRIME - PRocess modelling in ImpleMEntation research: selecting a theoretical basis for interventions to change clinical practice BMC Health Serv Res 2003 3 22 14683530 10.1186/1472-6963-3-22
Eccles M Grimshaw J Walker A Johnston M Pitts N Changing the behaviour of healthcare professionals: the use of theory in promoting the uptake of research findings J Clin Epidemiol 2005 58 107 112 15680740 10.1016/j.jclinepi.2004.09.002
|
16011811
|
PMC1183206
|
CC BY
|
2021-01-04 16:31:52
|
no
|
BMC Health Serv Res. 2005 Jul 13; 5:50
|
utf-8
|
BMC Health Serv Res
| 2,005 |
10.1186/1472-6963-5-50
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-501596974710.1186/1471-2334-5-50Research ArticleDiagnostic and prognostic utility of an inexpensive rapid on site malaria diagnostic test (ParaHIT f) among ethnic tribal population in areas of high, low and no transmission in central India Singh Neeru [email protected] AK [email protected] MM [email protected] SK [email protected] Praveen Kumar [email protected] Malaria Research Centre (ICMR) RMRCT Campus, Nagpur Road Jabalpur (M.P.), 482003 India2005 21 6 2005 5 50 50 11 2 2005 21 6 2005 Copyright © 2005 Singh et al; licensee BioMed Central Ltd.2005Singh 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
Malaria presents a diagnostic challenge in most tropical countries. Rapid detection of the malaria parasite and early treatment of infection still remain the most important goals of disease management. Therefore, performance characteristics of the new indigenous ParaHIT f test (Span diagnostic Ltd, Surat, India) was determined among ethnic tribal population in four districts of different transmission potential in central India to assess whether this rapid diagnostic test (RDT) could be widely applied as a diagnostic tool to control malaria. Beyond diagnosis, the logical utilization of RDTs is to monitor treatment outcome.
Methods
A finger prick blood sample was collected from each clinically suspected case of malaria to prepare blood smear and for testing with the RDT after taking informed consent. The blood smears were read by an experienced technician blinded to the RDT results and clinical status of the subjects. The figures for specificity, sensitivity, accuracy and predictive values were calculated using microscopy as gold standard.
Results
The prevalence of malaria infection estimated by RDT in parallel with microscopy provide evidence of the type of high, low or no transmission in the study area. Analysis revealed (pooled data of all four epidemiological settings) that overall sensitivity, specificity and accuracy of the RDT were >90% in areas of different endemicity. While, RDT is useful to confirm the diagnosis of new symptomatic cases of suspected P. falciparum infection, the persistence of parasite antigen leading to false positives even after clearance of asexual parasitaemia has limited its utility as a prognostic tool.
Conclusion
The study showed that the ParaHIT f test was easy to use, reliable and cheap. Thus this RDT is an appropriate test for the use in the field by paramedical staff when laboratory facilities are not available and thus likely to contribute greatly to an effective control of malaria in resource poor countries.
==== Body
Background
Malaria presents a diagnostic challenge in most tropical countries [1] and diagnosis of malaria still relies predominantly upon clinical presentation and the century old technique of microscopic examination of blood smears. Diagnosis by clinical symptoms alone is highly unreliable [2]. Microscopy is labour intensive, requires significant skills and time, which caused therapeutic delays. These diagnostic limitations affects the medical care provided, as Malaria is a potentially fatal disease, usually curable if diagnosed quickly [3,4]. The urgency and importance of obtaining results quickly from the examination of blood samples from patients with suspected acute malaria is now made possible with the introduction of rapid malaria diagnostic tests (RDTS) [2]. Very few studies have evaluated a rapid test in different epidemiological settings. This is important since a test may be valid but expensive or valid and reliable in laboratory/ hospital clinic but with several inconveniences to use in the field or very easy but with an unacceptably low validity [5]. A new indigenous test, the cheapest available so far, ParaHIT f has recently been introduced in the national programme for malaria control in remote and inaccessible villages of Central India. We conducted a study to assess if this RDT is reliable, simple and practical in varied epidemiological situation in hospital and field. Our objective was to determine whether this RDT could be widely applied as a diagnostic tool to effective management and control of malaria.
Methods
The ParaHIT f dipstick test (Code No. 25977, Span diagnostic Ltd, Surat, India) was evaluated as a diagnostic test in 4 districts of Central India (Fig. 1) with different transmission potential namely Jabalpur (Medical College hospital, a tertiary facility), Dindori (high transmission area), Mandla (very low/ no transmission due to intensive malaria control measures) and Korea district in Chhattisgarh (meso-endemic). The performance of the ParaHIT f for detecting P. falciparum infections was evaluated against microscopical examination of thick blood smears in field / hospital during April-August 2004.
Study sites
Baigachak (Dindori district)
Dindori is highly malarious district contributing highest number of malaria cases (12%) in the state (6). The baiga tribe, living in an extremely difficult hilly terrain, with numerous hillocks (MSL. 980m). The population of 10 study villages ranged from 50–800. The villages are surrounded by reserved forests of Sal (Shorea robusta). The inhabitants are barely clothed and live in extremely primitive conditions. Medical facilities are non-existent. Anopheles culicifacies and An. fluviatilis, are two efficient vectors in the area. Seasonal transmission of P. vivax and P. falciparum are common. Chloroquine resistance against P. falciparum is prevalent. The villages are under indoor residual spray with synthetic pyrethroid (deltamethrin @20 mg/m2). All patients with fever or history of fever were screened for malaria in a mobile field clinic on a first come first serve basis.
Korea district (Chhattisgarh state)
District Korea is situated on the border of Madhya Pradesh in Chhattisgarh state. It also has dense forest and the terrain is undulating (MSL750m). The field clinic traveled 10 villages for screening malaria patients (gond ethnic tribe). The area is under 2 rounds of DDT spray (1 gm/m2). Other characteristics (vectors and transmission pattern) are the same as in Dindori.
Mocha (Mandla district)
Surveillance of fever cases was carried out in 10 villages (MSL – 475m). Transmission is seasonal and mainly due to P. falciparum. Other characteristics are the same as in Dindori.
Medical College hospital (Jabalpur district)
The subjects of the hospital based study were the suspected cases of malaria who presented with fever or history of fever, at the malaria clinic run by the Malaria Research Centre (MRC) at Medicine department of Government Medical College hospital. This hospital is the largest medical facility in Jabalpur district (MSL 394m) and serves both as a general hospital for the local people and as a referral hospital for the adjoining districts.
Sampling
A finger-prick blood sample was collected from each case, after verbal informed consent was obtained. This sample was used to prepare thick and thin smears and for testing with the ParaHIT f test. There were four teams of MRC, Field Station Jabalpur, each consisting of two field workers and one Research Scientist. A one hour workshop, including training in blood collection from finger prick, performance and interpretation of rapid test was conducted at MRC laboratory. All the team members had attended the workshop in the use of the RDTs. All the samples were tested in field/hospital unsupervised. The thick blood smears were also examined in the field. The smears were checked by a microscopist blinded to the clinical status of the subjects and to the results of the diagnostic tests. Parasitaemia was determined from the thick films by counting the number of parasites against 200 leucocytes and assuming that each subject had 8000 leucocytes/μl.
The ParaHIT f test was performed as per manufacturer's instructions [5]. The standard reading time is between 15–30 minutes. All adult subjects with P. falciparum were administered the standard oral dose of chloroquine (1500 mg chloroquine in 3 days) followed by primaquine (45 mg as a single dose). Children were given proportionally lower doses. Infants and pregnant women were not given primaquine.
To check that HRP-2 antigenaemia was cleared by treatment, finger prick blood samples were collected on day 10 post-treatment and tested using the RDT in parallel with blood smears. To minimize variability, one microscopist examined all the smears and one technician interpreted all the results of the rapid tests. A quality control procedure was put in place. All discordant results (between the RDT and the slide), all slides where only P. falciparum gametocytes were detected and random sample of 20% of the remaining slides were checked blind by an independent trained laboratory technician at MRC laboratory.
Ethical clearance and data analysis
The study protocol was approved by the ethics committee of the Malaria Research Centre (ICMR) Delhi.
The data were recorded and analyzed using statistical software (SPSS version 10.0, SPSS inc. Chicago 12). Once all the samples had been tested, specificity, sensitivity, predictive values and accuracy to the ParaHIT f were estimated using microscopy as gold standard [7,8]. Briefly, sensitivity was calculated as TP/ (TP+FN), specificity as TN/ (TN+FP), positive predictive value (PPV) as TP/ (TP+FP), negative predictive value (NPV) as TN/ (TN+FN) and accuracy as (TP+TN)/ number of all tests. The J-index i.e., the over all measure of reliability of a diagnostic test calculated as (TPxTN)-(FPxFN)/ (TP+FN) (TN+FP). The mixed infection of P. vivax and P. falciparum was treated as P. falciparum for the purpose of analysis.
Results
The rates of fever and parasitaemia varied by study sites. The performance indices of RDT at each site is shown in Table 1.
Baigachak, Dindori district (high transmission area)
Two hundred five of 467 patient's (43.9%) had positive thick films, with P. falciparum comprising 191, (40.9%) and P. vivax 14 (3.0%). The asexual parasite density ranged from 640–280,000 parasites/μl (GMPD 9840 ± 7.38parasites/μl). The RDT detected malaria infections in 171 (89.5%) of 191 patients. The sensitivity of the rapid test was 90.2% (95%CI, 85–94) and specificity 83% (95%CI, 78–87%) compared to microscopy. The PPV and NPV were respectively 77% (95%CI, 71–83) and 93% (95%CI, 89–96).
In all 18 patients were positive by microscopy but negative by RDT and the parasitaemia in these patients ranged from 100–36,920 parasites/μl (GMPD 3715.35 ± 5.37). Only 94 cases were followed up on 10th day, out of which 39 (41.5%) test were still positive post treatment. In 18 of the subjects (19%), the blood smears were also positive for P. falciparum. However, the RDT was positive only in 16 matching cases. 23 subjects were found false positive and 2 false negative with RDT. The sensitivity, specificity, PPV and NPV were respectively 84% (95% CI, 60.4–96.6), 69% (95% CI, 57.6–79.5), 41% (95% CI, 25.6–57.9) and 94.5% (95% CI, 84.9–98.9). The accuracy of test was 72.3% and J index 0.53 as compared to microscopy.
Korea district (high to low transmission area, mero-endemic)
Of the 440 individuals screened, microscopic examination of the thick film detected 64 P. falciparum, 26 P. vivax and 1 mixed infection. The GMPD for P. falciparum was 323.59 ± 2.51 parasites/μl (range 80–7840 parasites/μl). The RDT detected malaria parasite in 57 subjects. The sensitivity and specificity of the test were 97% (95%CI, 88.5–99.6) and 95% (95% CI, 92–97) respectively with PPV 75% (95% CI, 64–85) and an NPV of 99.4% (95%CI, 98–100). When compared with microscopy, the accuracy of the test was 95.2% and J-index 0.92.
Medical College hospital, Jabalpur (low transmission area)
Of the 208 patients screened, only 19 were infected with malaria. 7 P. vivax and 12 P. falciparum, of which 7 were cerebral malaria (CM) patients, comatose with various complications. All seven P. falciparum (CM) were also positive by RDT. However, results of the RDT were not used to guide treatment and antimalarial therapy was prescribed by the physician on the basis of clinical severity. The asexual parasitaemia ranged from 5000–100000 parasites/μl. The sensitivity of the test was 70% (95% CI, 35–93) with 93% specificity (95% CI, 88.4–96). The PPV and NPV were 33% (95%CI, 15–57) and 98.4% (95%CI, 95.4–99.7) respectively. The accuracy of the test was 92% and J-index 0.63 in comparison with gold standard.
Mocha, Mandla district (no transmission area)
Only one hundred and two blood samples were screened by both the methods. Out of which only one was found RDT positive that too was smear negative.
Overall from 1217 subjects, 83 (6.8%) false positive were recorded. A prolonged re-examination of the slides from these 83 subjects, who were positive by the test, but apparently had negative slides, did not reveal the presence of P. falciparum trophozoites. 60 out of 83 subjects claimed to have taken antimalarials drugs in the preceding days. P. falciparum gametocytes only were seen in 14 cases of which 9 were RDT positive. Out of 47 subjects with P. vivax infections as detected by microscopy, only 5 were found positive for P. falciparum by the RDT.
Discussion
Rapid, accurate diagnosis is necessary to effective management and control of malaria [2,9,10]. The strength of the present study is that it was performed in different epidemiological settings i.e. in one of the most difficult field conditions of baigachak and within the routine of tertiary Medical College hospital receiving persons of all ages with varying clinical severity and allowing the evaluation of the performance of the RDTs in all persons suspected to have malaria what ever their history i.e. recent malaria attack or anti-malaria drug intake in the previous days and clinical status (fever or not, severe or mild malaria). Therefore, the results obtained reflect the performance of the tests in a real situation.
From a malaria transmission perspective in baigachak, the RDT can play a key role in rapid diagnosis and prompt treatment of malaria in high transmission areas (where resistance to chloroquine also necessitates the use of more expensive alternate therapy). As RDT can be conducted immediately in the field clinic while the patient is present, the most important point for the villagers is the knowledge that they are infected with malaria parasite as well as dispersing any information for preventive measures of the disease. On the contrary, the delay in the results of microscopic diagnosis is a serious obstacles for the operation of a malaria control programme in remote areas where health staff operating the control programme have to visit several times to treat the positive subjects as per the results of microscopic examination.
Of clinical concern is the persistence of parasite antigen leading to false positives even after the clinical symptoms of malaria has disappeared and the parasites have apparently been cleared from the host as recorded earlier from central India [11,12] and world wide in a variety of settings [1]. In this study also we detected persistent positivity of RDT in 24.4% of treated patients without asexual parasitaemia on day 10. Thus the test has limited utility as a prognostic tool. Because of the relatively high rate of persistent false positive tests in recently treated cases, the value of predictability of a test band may be limited to new, untreated cases. However, the high NPV allow us to confidently diagnose negative test patients as non malaria patients in all epidemiological settings.
An additional clinical concern is the occasional failure of rapid test to detect high parasite densities as recorded earlier [10]. Pf HRP-2 deleted mutants have been reported but their prevalence is not known [13]. Unless a solution is implemented for these rare but recurrent diagnostic failure, clinicians using a Pf HRP-2 based rapid test must keep in mind that a negative test greatly reduces the probability of, but does not rule out a parasitaemia of greater than 35,000 parasites/μl. The results of more field tests may help to explain the apparent discrepancies.
In Mocha, an area free from malaria transmission, RDT may be useful to exclude the falciparum infection and indicate the need to give treatment for other causes of fever. RDT thus would improve the accuracy of the diagnosis and consequently decrease the use of antimalarials. In the medical college hospital the RDT also appeared useful although the test sensitivity was not high as in field clinics. However, RDT could facilitate the early diagnosis and an appropriate therapy in patients with cerebral malaria thereby reducing mortality as recorded earlier [7,9]. The value of the RDT specificity observed in the present study is consistent with the result of the study carried out in Uganda [5].
In addition to good sensitivity and specificity, the criteria for the rapid test should also include affordability and user friendliness. The cost per test is Rs. 28 i.e. US$ 0.50, the cheapest available so far to the Ministry of Health and the institutes buying large numbers of test kits. Besides, where the number of people to be tested is very low in small villages in remote areas, where health-facility coverage is low, and the risk of contracting malaria is high, the test may be more cost effective than microscopy as the time factor may be in favour of the rapid test.
Despite some limitations, a post study survey of all the personnel directly involved in the performance of the RDT were of opinion that the test was easy to use, decreased stress and potential delay in the diagnosis of falciparum malaria in field. If the present observations are validated in larger multicentre, clinical trial, the test may prove to be a useful alternative to microscopy, particularly in places where the facilities for microscopy are poor or non existent.
Conclusion
RDTs in conjunction with microscopy should improve diagnosis of malaria. However, RDTs are more suited to investigators/health workers in situations where health services are deficient or absent. Therefore, it is reasonable to consider future use of the RDTs as an epidemiological tool for the rapid screening of malaria.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Neeru Singh: Participated in design of the study and writing the manuscript.
MM Shukla: Involved in rapid analysis in Medical College Hospital, Jabalpur.
AK Mishra: Performed the rapid test in Korea district and help in analysis.
SK Chand: Involved in data collection and analysis in Mocha, Mandla.
Praveen Kumar Bharti: Performed the test in baigachak, Dindori.
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 Prof. (Dr.) A.P. Dash, Director, Malaria Research Centre in New Delhi for his help in various way.
==== Refs
Moody A Rapid diagnostic tests for malaria parasites Clin Microbiol Ref 2002 15 66 78 10.1128/CMR.15.1.66-78.2002
World Health Organization Malaria diagnostics, New Perspectives WHO/MAL 2000 1091 4 29
Marsh K English M Peshu N Crawley J Snow R Clinical algorithm for malaria in Africa (letter) Lancet 1996 347 1327 8622515 10.1016/S0140-6736(96)90972-1
White NJ The treatment of malaria N Engl J Med 1996 335 800 806 8703186 10.1056/NEJM199609123351107
Guthmann JP Ruiz A Protto G Kiguli J Bonte L Legros D Validity, reliability and ease of use in the field of five rapid tests for the diagnosis of Plasmodium falciparum malaria in Uganda Trans R Soc Trop Med Hyg 2002 96 254 257 12174772 10.1016/S0035-9203(02)90091-X
Anonymous Annual report National Anti Malaria Programme Directorate of Health Services, Bhopal, Madhya Pradesh, India 2003
Singh N Saxena A Usefulness of rapid on site Plasmodium falciparum diagnosis (Paracheck® Pf) in forest migrants and among indigenous population at the site of their occupational activities in central India Am J Trop Med Hyg 2005 72 26 29 15728862
Singh N Valecha N Evaluation of a rapid diagnostic test 'Determine™ Malaria pf', in epidemic-prone forest villages of central India (Madhya Pradesh) Annal of Trop Med & Paraasitol 2000 94 421 427
Singh N Nagpal AC Performance of OptiMAL dipstick test for management of severe and complicated malaria cases in a tertiary hospital, central India J infect 2004 48 364 365 15066340 10.1016/j.jinf.2004.01.007
Jelinek T Grobusch MP Schwenke S Steidl S Sonnenburg FN Nothdurft HD Klein E Loscher T Sensitivity and specificity of dipstick tests for rapid diagnosis of malaria in non immune travelers J Clin Microbiol 1999 37 721 723 9986839
Singh N Valecha N Sharma VP Malaria diagnosis by field workers using an immunochromatographic tests Trans R Soc Trop Med Hyg 1997 91 396 397 9373631 10.1016/S0035-9203(97)90254-6
Singh N Singh MP Sharma VP The use of a dipstick antigen capture assay for the diagnosis of Plasmodium falciparum infection in a remote forested area of central India Am J Trop Med Hyg 1997 56 188 191 9080879
Traore I Koita O Doumbo O Kassambara L Ouattara A Diakite M Sagara I Diallo M Krogstad DJ Field studies of the ParaSight™ F test in a malaria endemic area: cost, feasibility, sensitivity, specificity, predictive value and detection of HRP2 gene among wild type Plasmodium falciparum in Mali Am J Trop Med Hyg 1997 57 272 9311635
|
15969747
|
PMC1183207
|
CC BY
|
2021-01-04 16:28:15
|
no
|
BMC Infect Dis. 2005 Jun 21; 5:50
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-50
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-561600197810.1186/1471-2334-5-56Research ArticleSeroprevalence of hepatitis A infection in a low endemicity country: a systematic review Pham Ba' [email protected] Bernard [email protected] Serres Gaston [email protected] Vladimir [email protected] Andrea C [email protected] Jan [email protected] David W [email protected] BioMedical Data Sciences, GlaxoSmithKline, Ontario, Canada2 Chalmers Research Group, Children's Hospital of Eastern Ontario Research Institute, Ontario, Canada3 Centre De Recherche du CHUQ, Institut Nationale Sante Publique du Quebec, Quebec, Canada4 Centre de Recherche du CHUQ, Quebec, Canada5 Vaccine Evaluation Centre, University of British Columbia, British Columbia, Canada2005 7 7 2005 5 56 56 11 5 2005 7 7 2005 Copyright © 2005 Pham et al; licensee BioMed Central Ltd.2005Pham 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 Canada – a low endemicity country, vaccines for hepatitis A virus (HAV) are currently recommended to individuals at increased risk for infection or its complications. Applying these recommendations is difficult because the epidemiology of HAV infection is poorly defined, complex, and changing. This systematic review aimed to 1) estimate age-specific prevalence of HAV antibody in Canada and 2) evaluate infection-associated risk factors.
Methods
MEDLINE (1966–2005) and EMBASE (1980–2005) were searched to identify relevant studies for the systematic review. Archives for the Canada Diseases Weekly Report (1975–1991) and Canada Communicable Disease Report (1992–2005) were searched for relevant public health reports. Data were abstracted for study and participants' characteristics, age-specific prevalence, and risk factors.
Results
A total of 36 reports describing 34 unique studies were included.
The seroprevalence in Canadian-born children was approximately 1% in ages 8–13, 1–6% in 20–24, 10% in 25–29, 17% in 30–39, and increased subsequently. In age groups below 20 and 20–29, age-specific seroprevalence generally remained constant for studies conducted across geographic areas and over time.
Compared to Canadian-born individuals, subjects born outside Canada were approximately 6 times more likely to be seropositive (relative risk: 5.7 [95% CI 3.6, 9.0]). Travel to high risk areas in individuals aged 20–39 was associated with a significant increase in anti-HAV seropositivity (RR 2.8 [1.4, 5.5]). Compared to heterosexuals, men having sex with men were only at a marginally higher risk (adjusted odds ratio 2.4 [0.9, 6.1]). High risk for seropositivity was also observed for Canadian First Nations and Inuit populations.
Conclusion
Results from the current systematic review show that in this low endemicity country, disease acquisition occurs in adulthood rather than childhood. The burden of disease is high; approximately 1 in 10 Canadians had been infected by ages 24–29. The increase in prevalence in young adults coincides with disease importation and increasing frequency of risk factors, most likely behavioral-related ones.
Gaps in seroprevalence data were identified rendering the application of current immunization recommendations difficult. A nationwide prevalence survey for all Canadians is needed. This is essential to quantify the effectiveness of current recommendations and conduct cost-effectiveness evaluations of alternative immunization programs, if necessary.
==== Body
Background
Hepatitis A virus (HAV) is prominent in many areas of the world [1-3]. In North America, infection rates have declined with better hygiene practice and public sanitation but remain heterogeneous across geographic and socioeconomic strata [4-6]. Further decline is possible with HAV vaccines which provide consistent, long-lasting protection and have been available since the mid-1990s [7,8]. In the United States, universal vaccination of children and youth has been in place for about 6 years in high endemicity areas [9], leading to historically low rates nationally in recent years [10]. In Canada, the current national immunization guide recommends HAV vaccines for individuals at increased risk of infection or its complications. The guide also states that a universal immunization program should be considered, but further discussion is needed nationally [11].
Applying the Canadian recommendations is difficult because the epidemiology of HAV infection in Canada is poorly defined, complex, and changing [8]. Reported rates differ substantially by province, gender, and age [12]. The rates show repeated peaks and troughs [13] and the last peak occurred in mid-1990 [12,13]. It was during the subsequent period of decline that vaccines were used as a tool to enhance HAV control [8].
Evaluating the impact of the current recommendations is also difficult. Data are needed to distinguish between a cyclical decline and a further decline associated with the recommendations [10]. Such assessment is important to inform future immunization policies. A combination of timely case-notification data, prevalence data, and risk factor data is required for both the application and evaluation of the recommendations.
Case-notification data is of limited use due to under-detection of sub-clinical infection and under-reporting of confirmed cases [8,14]. A useful means that circumvent these limitations is to measure the prevalence of HAV antibody [15]. Following an acute infection, antibody to HAV develops in virtually every instance, remaining detectable for decades, and providing a reliable marker of past infection. In the United States, countrywide seroprevalence surveys and sentinel surveillance have been conducted to provide insight into HA epidemiology, and to rationalize and evaluate immunization programs [10,15,16]. In Canada, similar surveys exist but are limited in scope and comprehensiveness [17-19]. Consequently, risk factor data are also limited and fragmented [12,17]. The current systematic review aimed to 1) estimate age-specific prevalence of hepatitis A antibody in Canada and 2) evaluate infection-associated risk factors.
Methods
MEDLINE (Jan. 1966 – Mar. 2005) and EMBASE (Jan.1980 – Mar. 2005) were searched to identify citations of potentially relevant studies for the systematic review (MeSH terms: "hepatitis" exploded AND "Canada" exploded). A study report was included if it contained prevalence data of HAV-antibody (detected through sera or saliva samples, hereafter referred to as seroprevalence) for a Canadian population. Reasons for exclusion were categorized and reported. Citations were screened independently by two reviewers. Full-text study reports from citations deemed relevant by one reviewer were obtained. Archives for the Canada Diseases Weekly Report (Jan. 1975 – Dec. 1991, the last year of reporting) and Canada Communicable Disease Report (Jan. 1992 – Mar. 2005) were also searched for potentially relevant public health reports [20]. Two reviewers independently reviewed both published and public health reports. Discrepancies were resolved through discussion. Back referencing and author searches of all included studies were conducted. Other potentially relevant reports were obtained by contacting HA experts and related public health units.
From the included reports, data were independently abstracted for study and participants' characteristics. Age-specific prevalence of HAV antibody data were extracted for Canadian-born participants, all Canadians including individuals born outside the country, and participants with known risk factors [11]. Seroprevalence estimates and 95% confidence intervals (CI) were derived assuming a binomial distribution for the number of seropositive individuals from the total numbers of tested individuals. Participants who reported receiving the HA vaccine were excluded from the prevalence estimates as they were likely to have vaccine-induced antibody.
If available, adjusted odds ratios (AOR) for seropositivity of both demographics and risk factors were extracted, together with the baseline risk of HAV seropositivity (i.e., population, location and timing of the survey). When the AOR of a variable was not reported, HAV antibody data stratified by the variable were obtained to derive the unadjusted relative risk for seropositivity (URR). Relative risk was used as it is a better risk estimate than the odds-ratio in the range of seroprevalence observed in this systematic review [21]. If appropriate, a random effects model was used to combine URRs across studies, together with an assessment for heterogeneity (i.e., chi-squared test).
Information related to the risk of seropositivity was summarized for the following risk categories [11]: 1) travelers to high endemicity areas; 2) groups with high risk activities such as men who have sex with men (MSM), illicit drug users, and street people; 3) First Nations and Inuit populations; and 4) others (e.g., individuals with chronic hepatitis, household contacts, infected food handlers, etc.).
Results
Literature search
A total of 36 reports describing 34 unique studies were included in the systematic review (Table 1) [18,19,22-55]. These were obtained from screening 413 potentially relevant citations and reviewing 66 full-text study reports and 25 public health reports (Figure 1). Common reasons for exclusion at the screening stage included studies of hepatitis B virus (n = 95), hepatitis C virus (n = 64), commentaries (n = 24), and others (n = 121; Figure 1). Common reasons for exclusion at the full-text review stage included general review of HAV (n = 9), no seroprevalence data (n = 17), and other viral hepatitis (n = 16; Figure 1).
Table 1 Study and participant characteristics
Study Study Design Study Year n Population Location Age in years
PUBLISHED LITERATURE Mean ± SD or (range)
Ochnio 2005* [22] P 2000–1 811 Grade 9 students British Columbia (14–15)
Muecke 2004* [23] CC 2001 492 Day-care educators Montreal 37
Minuk 2003 [24] P 1999 315 First Nations Manitoba 34 ± 15
Ochnio 2001 [25] P 1998 494 Street youth, IDU, MSM Vancouver 19, 35, 34
Smieja 2001 [26] CC 1997–8 179 IHD patients Hamilton 61 (38–81)
Kiefer 2000 [27] R 1997 343 Hepatitis C patients Edmonton 40 (0–95)
Allard 2001 [28] P 1995–97 353 Gay men Montreal 37
Moses 2002 [29] P 1995–6 533 Street people Winnipeg 26 (11–65)
Roy 2002 [30] R 1995–6 427 Street youth Montreal (14–25)
Ochnio 1997 [31] P 1995–6 224 Grade 6 students Vancouver (10–12)
De Serres 1997 [32] P 1995 85 Sewer workers Quebec 36 ± 7
De Serres 1995 [33] CC 1995 228 Sewer workers Quebec 41 (28–64)
Smieja 2003 [34] R 1993–5 3127 CV or high risk diabetes Canada 65
Payment 1991 [18] P 1988–9 617 French-Canadian Montreal (9–79)
Embil 1989 [35] P 1981–3 2036
1922 1/CF recruits
2/CF males Nova Scotia, Quebec, Posted abroad 1/(15–25)
2/26 (17–53)
Nicolle 1986 [36], Minuk 1985 [37] P 1982 172 Chesterfield Inlet Northwest Territories 0 – 78
Crewe 1983 [38] P 1981–2 304 Children attending outpatient clinic Halifax (0.5–16)
Minuk 1982 [39]
Minuk 1982 [40] P 1980 720 Inuit Northwest Territories (0.3–86)
McFarlane 1980 [41] P, R 1980 243
152
293 & 282 1/STD clinic patients
2/Student nurses
3/2 groups of blood donors Nova Scotia 1/(16–26)
2/(18–24)
3/(16–26) & (51–65)
Buchner 1980 [42] R 1980 5097 Blood donors Toronto <21, >60
Richer 1982 [55] R 1970–79 447 Samples of acute viral hepatitis Montreal Not reported
Minuk 1994 [43] P 1974–8 42 Household transmission Winnipeg 27 ± 12
Minuk 2003 [44] SR 1980–2000 1706 Inuit and First Nations Various locations 0–60+
McFarlane 1982 [45] P NR 154 Institutions Nova Scotia (13–28)
McFarlane 1981 [46] P NR 130 Patients with hematological malignancy Nova Scotia (4–76)
GREY LITERATURE
Duval 2005* [48] P 2003 1057 Canadian aged 8–13 Canada (8–13)
Wu 2005 [47] R 1992–9 NR Subjects tested for HAV infection Manitoba Not reported
Ochnio 2004 [49] P 2003 585 Young adults Vancouver (20–39)
Cook 2000 [19] R 2000 1206 Women of child-bearing age British Columbia (15–44)
Harb 2000 [50] P 1999 172 First Nations British Columbia (0–40+)
Levy 2001 [51] P 1997 1000 University students Toronto 25 ± 5
Ford-Jones 1995 [52] P 1993 122 Day-care providers Toronto Not reported
Ochnio 1995 [53] P 1994–5 1019 Clients to travel clinic Vancouver (2–69+)
Kocuipchyk 1995 [54] P 1991–2 505 Individuals attending travel clinic Edmonton (16–60+)
Notes:*Study reported seroprevalence data for individuals with or without HAV vaccination.
Abbreviations: Study design: P prospective data acquisition, R retrospective data acquisition, CC case control, SR systematic review. Population: IDU injection drug users, MSM men who have sex with men, IHD ischemic heart disease, CV cardiovascular, CF Canadian Forces, STD sexually transmitted disease.
Figure 1 Results of the literature search.
Overall, 74% (n = 25) of the included studies were reported in peer-reviewed journals while 26% (n = 9) were grey literature [20], including 9% (n = 3) public health reports, 14% (n = 5) abstracts, and 3% (n = 1) unpublished study (Table 2). HAV antibody was detected using serum samples in 28 studies and saliva samples in 6. The median sample size was 427 and 793 for the published and grey literature studies, respectively. Only 21% (7/34) of all studies reported prevalence among Canadian-born participants and 29% (10/34) reported prevalence data of all participants including foreign-born individuals. The majority of these studies (27/34) reported prevalence data of participants with known risk factors.
Table 2 Study characteristics
Published Literature (n = 25) Grey Literature (n = 9)
Peer-reviewed study report (n = 25) Public health report (n = 3)
Abstract (n = 5)
Unpublished report (n = 1)
Study Design
Case-control 2 0
Prospective (P) data acquisition 16 7
Retrospective (R) data acquisition 5 2
P & R data acquisition 1 0
Systematic review 1 0
Sample Size
>1000 4 3
100 – 1000 19 6
<100 2 0
Median [1st, 3rd Quartile] 427 [224, 720] 793 [422, 1029]
Mean (Min, Max) 877 (42, 5097) 708 (122, 1206)
Timing of data collection
2000 – 2004 2 3
1990 – 1999 12 6
1980 – 1989 7 0
1970 – 1979 2 0
Not reported 2 0
Populations with prevalence data
Canadian born subjects 4 3
All Canadians¶ 6 4
Participants with known risk factors 21 7
Seropositivity test
Serum samples 22 6
Saliva samples 3 3
Notes: ¶including subjects born outside of Canada
Age-specific seroprevalence
The seroprevalence in Canadian-born children aged 8–13 was 1% [95% CI: 0.5–2%] according to a national survey conducted in 2003 [48]. The seroprevalence was 1–6% in ages 20–24, approximately 10% in 25–29, 17% in 30–39, and increased subsequently (Figure 2). In age groups below 20 and 20–29, age-specific seroprevalence generally remained constant for studies conducted across geographic areas in 1980, 1988, 1997, and 2003. This remained so despite differences in study methodology.
Figure 2 Seropositivity rate (95% confidence interval) among Canadian-born study participants.
There was no association between seropositivity and gender based on 9 population comparisons from 5 studies (n = 4158, URR: 1.0 [95% CI: 0.9, 1.1]) [22,31,35,41,54], which was consistent with results of 3 other studies reporting adjusted risk estimates (Table 3) [25,34,48]. Two studies in the early 1980's suggested that individuals living in urban areas were 30% more likely to have HAV antibody compared to those in rural areas (n = 647, URR: 1.3 [1.2, 1.5]) [41,46].
Table 3 Assessment of risk factors
Risk factor n Risk Measure Risk Estimate (95% CI) Population, Location, Timing of Data Acquisition Age in years Study
DEMOGRAPHICS Mean ± SD or (range)
Female versus Male 1003 AOR 2.2 (0.8, 6.25) School-aged children, Canada, 2003 (8–13) [48]
Female versus Male 3128 AOR 0.8 (0.6, 0.96) CV or high risk diabetes, Canada, 1993–5 65 [34]
Female versus Male 494 AOR 1.3 (0.8, 2.3) SY, MSM, IDU, Vancouver, 1998 32 ± 11 [25]
Female versus Male 4158 URR 1.0 (0.9. 1.1) [p = 0.30]* 9 population comparisons from 5 studies (8–65+) [31,35,41,49,54]
Urban versus Rural 647 URR 1.3 (1.2, 1.5) [p = 0.59]* 3 population comparisons from 2 studies (16–76) [41,46]
Born in high risk country versus born in Canada 494 AOR 2.9 (1.1, 7.6) SY, MSM, IDU, Vancouver, 1998 32 ± 11 [25]
Born in endemic country versus born in Canada 1003 AOR 22.3 (6.6, 75.0) School-aged children, Canada, 2003 (8–13) [48]
Foreign-born versus Canadian-born 353 AOR 6.2 (2.6, 15.0) Gay men, Montreal, 1995–97 36 [28]
Born in a high-income country versus moderate to low† 492 AOR 20.8 (9.4, 46.0) Day-care educators, Montreal, 2001 37 [23]
Foreign-born versus Canadian-born 3008 URR 5.7 (3.6, 9.0) [p < 0.01]* 5 population comparisons from 5 studies (2–69+) [30,31,49,51,53]
TRAVEL TO HIGH RISK AREA
Travel to high risk area versus otherwise 1003 AOR 1.4 (0.4, 4.8) School-aged children, Canada, 2003 (8–13) [48]
Travel to high risk areas versus otherwise 407 URR 2.8 (1.4, 5.5) Canadian-born adults, Vancouver, 2003 (20–39) [49]
Ever travelled to a developing country‡ 492 AOR 2.4 (1.3, 4.2) Day-care educators, Montreal, 2001 37 [23]
HIGH RISK ACTIVITIES
MSM versus heterosexuals 494 AOR 2.4 (0.9, 6.1) SY, MSM, IDU, Vancouver, 1998 (25–34) [25]
Sexual partners with VH history versus otherwise 420 AOR 13.8 (4.2, 45.2) Street youths, Montreal, 1995–6 (14–25) [30]
Insertive anal penetration versus otherwise 420 AOR 5.1 (1.6, 16.7) Street youths, Montreal, 1995–6 (14–25) [30]
History of STD versus no history 500 AOR 2.0 (1.2, 3.3) Street people, Winnipeg, 1995–6 26 (11–65) [29]
History of IDU versus no history 494 AOR 6.5 (1.6, 26.3) SY, MSM, IDU, Vancouver, 1998 (25–34) [25]
History of IDU versus no history 500 AOR 1.6 (0.99, 2.7) Street people, Winnipeg, 1995–6 26 (11–65) [29]
FIRST NATIONS AND INUIT
Native versus Non-native 1003 AOR 5.2 (1.0, 26.0) School-aged children, Canada, 2003 (8–13) [48]
Aboriginal versus Non-aboriginal 500 AOR 6.6 (3.8, 11.5) Street people, Winnipeg, 1995–6 26 (11–65) [29]
Inuit versus white in NWT 708 URR 4.5 (2.4, 8.5) Inuits, Baker Lake, NWT, 1980 (0.3–86) [39,40]
4+ versus 1–3 household occupants 635 URR 1.1 (0.98, 1.3) Canadian Inuit, Baker Lake, 1980 (0.3–86) [39,40]
OTHERS
Years working in day-care, 5-year groups§ 339 AOR 1.3 (1.0, 1.8) Canadian-born day-care educators, Montreal, 2001 34 [23]
History of daycare versus no history 1278 URR 1.2 [0.7, 2.2] [p = 0.30]* 2 population comparisons from 2 studies 8–13 [31,48]
Sewer workers versus controls|| 228 URR 1.1 (0.8, 1.4) Sewer workers, Quebec City, 1993 41 (28–64) [33]
3+ versus 0–3 siblings 502 URR 2.0 (1.7, 2.5) Travel clinic, Edmonton, 1991–2 (16–60+) [54]
Current household income <20,000/yr¶ 153 AOR 5.3 (1.2, 24.2) Foreign-born day-care educators, Montreal, 2001 39.7 [23]
Annual family income <30,000 vs ≥30,000 1057 URR 0.7 (0.3, 2.0) School-aged children, Canada, 2003 (8–13) [48]
Abbreviations: AOR adjusted odds ratio. CV cardiovascular. SY street youth. MSM men who have sex with men. IDU injectable drug user. URR unadjusted relative risk. STD sexually transmitted disease. NWT North West Territory. G6 grade 6. yr year
Notes: *Meta-analytical estimates (95% CI) [p-value from a test of homogeneity] from random-effects models. † The day-care educator study included 492 participants, including 339 Canadian-born individuals and 153 foreign-born. The odds-ratio for "born in a high-income country versus moderate to low" was 20.8 (95% CI 9.4, 46.0) for all 492 participants, not reported for Canadian-born, and 4.6 (1.7, 12.2) for foreign-born. ‡ The odds-ratio for "ever traveled to a developing country" in the day-care educators study was 2.4 (1.3, 4.2) for all 492 participants, not significant for Canadian-born (estimated OR not available), and 8.1 (2.3, 29.0) for foreign-born. § The odds-ratio for "years working in day-care, 5-year groups" was not significant for all 492 participants (estimated OR not available), 1.3 (1.0, 1.8) for Canadian-born, and not significant (estimated OR not available) for foreign-born. || Control subjects were outpatients undergoing lipid testing,. ¶ The odds-ratio for "current household income <20,000/yr" in the day-care educators study was not significant (estimated OR not available) for all 492 participants and Canadian-born and was 5.3 (1.2, 24.2) for foreign-born.
Compared to Canadian-born individuals, subjects born outside Canada were approximately 6 times more likely to be seropositive (n = 3008, URR: 5.7 [3.6, 9.0], Table 3) [30,31,49,51,53], which most likely occurred in their birth country. However, the possibility that infection occurred in Canada could not be ruled out. Age-specific seroprevalence estimates including these individuals varied substantially and could only be used to infer the level of immunity in the population. For example, seroprevalence among all Canadians aged <20 ranged from 2–16% [18,31,38,48,51] for which 3–25% of the sampled populations were individuals born outside the country [25,48,53]. This contrasted the 1% seropositivity for Canadian-born participants reported above.
Risk factors
Travel-related data were available in 6 studies (Table 1) [23,35,48,49,53,54]. HAV antibody prevalence for Canadian-born individuals visiting a travel clinic was 2.3% in ages 20–25 and 4.3% in ages 25–28 [53]. The prevalence of seropositivity in these individuals was comparable to that reported above for Canadian-born individuals. The risk related to travel among Canadian-born was also not significant in a study of day-care educators [23]. Two population-based surveys reported travel-related risk [48,49]. In one study, travel to high risk areas by Canadian-born individuals aged 20–39 (approximately 12% of study participants) was associated with a significant increase in seropositivity (n = 407, URR: 2.8 [1.4, 5.5]) [49]. In a national survey of children aged 8–13, the prevalence was 1.9% in Canadian-born non-vaccinated travelers and 1.3% in non-travelers; the association was again not significant (Table 3) [48].
Two studies evaluated HAV infection among MSM in two different cities [25,28]. MSM participants on average had 3 sexual partners over the preceding 6 months, according to one study [28]. Also, 18% of these individuals were food handlers. Compared to heterosexuals, MSM were only at a marginally higher risk for seropositivity (n = 494, AOR 2.4 [0.9, 6.1]), according to the second study [25]. However, the study sample was highly heterogeneous and included MSM, injection drug users (IDU), and street youth.
Data on street-involved populations were available in three studies (Table 3) [25,29,30]. Seropositivity was approximately 5% in street youth aged 14–25 in Vancouver and Montreal [25,30]. In the Montreal study [30], the outbreak in MSM (n = 376 cases from December 1994 to February 1998 [28]) seemed to have little effect on the prevalence of anti-HAV among street youth, measured during the same period. Significant behavioral risk factors for seropositivity were reported for street-involved individuals. These included IDU, history of sexually-transmitted disease, and high HAV-risk sexual activities (Table 3).
Seroprevalence in Canadian First Nations and Inuit populations were reported in four studies and summarized in a systematic review [44]. The prevalence ranged from 75–95% and was approximately three times that of non-Aboriginal Canadians residing in the same communities across all ages [44]. For example, Minuk and colleagues reported on a seroprevalence survey of 720 inhabitants of an Inuit community (n = 850). Approximately 27% of this community were aged 0–9, 30% aged 10–19, 32% aged 20–49, and 11% aged 50 or above [40].
Among Canadian-born day-care educators, there was a borderline significant association between risk of HAV positivity and years of employment (Table 3) [23]. A history of daycare attendance among grade 6 students was not associated with seropositivity [31]. Also, the seropositivity was 1.3–1.6 times higher in children aged 8–13 who attended day-care, but no statistical difference was evident in all participants, non-vaccinated participants, or those without known risk factors [48].
Other potential risk factors were also examined. From a sample of 343 individuals who tested positive for hepatitis C, 30% of those aged 20–29 were seropositive for HAV [27]. One prospective cohort followed 62 household contacts and 20 index cases over 6 months; the risk of infection was 52% among other susceptible household members [43]. Working in a sewage plant was not associated with seropositivity [33].
Discussion
The seroprevalence data consolidated in this systematic review had many limitations. Except for one national survey in ages 8–13 [48], other studies were generally not representative of the general population. Substantial variation across studies was observed with respect to study population, timing, sample size, and location. Reporting of data was inconsistent with respect to age stratification and definition of risk factors. Some studies conducted after the introduction of the vaccine around 1997 did not take vaccine-induced HAV antibody into account. This was, however, rectified in more recent studies [23,48,49]. For example, the seroprevalence in a national survey of children aged 8–13 was 2.7% overall and 2.0% after the exclusion of self-reported vaccinees [48]. The corresponding figures in a survey of young adults aged 20–39 were 22% and 16%, respectively [49]. Given these limitations, improvements in the reporting of future HAV prevalence studies are required. Most importantly, prevalence data should be stratified by participants' birthplace and account for vaccine-induced antibody.
Results from the current systematic review show that disease acquisition occurs in adulthood rather than childhood [14]. In Canada, the increase in prevalence in young adults coincides with disease importation and increasing frequency of behavioral risk factors, such as risk activities among MSM and street-involved populations. Even in this low endemicity country, approximately 1 in 10 Canadians had been infected by ages 24–29.
A low level of HAV immunity in Canada is evident from this systematic review. Over 90% of Canadian-born individuals aged 20–29, and over 80% of those aged 30–39 remained unprotected. Canadians born outside the country generally have a higher prevalence of HAV antibody, yet including these individuals did not significantly improve the percentage of protected Canadians. This low level of immunity and persistent risk of exposure to HAV suggest that outbreaks are possible in the future [14,47]. For example, unprotected clients exposed to an infected food handler led to mass immunizations in the early 2000's (Toronto 2002 [56], n = 19,208; London 2002 [57], n = 16,320; Vancouver 2002 [58] n = 6,000).
Clarifications are required to better understand the epidemiology of HAV in Canada, especially the inter-relation between timely case-notification data, seroprevalence data, and risk factor data. In a study examining national case-notification data from 1990–1999, estimated incidence of reported cases decreased while the average age of exposure and subsequent infection increased [12]. Given the low seroprevalence in Canadian youth, the current results suggest that the average age of HA exposure is above 24 and is increasing. While infection in children is often sub-clinical or mild, infected adults often experience more severe symptoms [1,59].
Results of the current systematic review are consistent with low HAV endemicity patterns in developed countries [60-62]. A substantial burden of infection was observed in young Canadians and this did not decrease among successive generations over the past 20 years. Similar observations were reported elsewhere [63]. In these low endemicity countries, outbreaks are common [64,65]. Sources of outbreaks that are common in these countries include infected food handlers [56,57], contaminated food importation [66,67], and unprotected immigrants who visit friends and relatives in their original countries [68].
In order to apply the current immunization recommendations, substantial information pertaining to groups at increased risk of HA infection or its complications is required. Results from this systematic review suggest that the risk of HA infection in these target groups was not well documented. With the exception of a few population-based surveys [48,49], most studies enrolled participants with known risk factors and failed to include a control group. In addition, some used residual sera obtained for other tests with virtually no risk factor data.
Conclusion
Results from the current systematic review show that in this low endemicity country, disease acquisition occurs in adulthood rather than childhood. The burden of disease is high; approximately 1 in 10 Canadians had been infected by ages 24–29. The increase in prevalence in young adults coincides with disease importation and increasing frequency of risk factors, most likely behavioral-related ones.
Gaps in seroprevalence data were also identified in this systematic review, rendering the application of current recommendations difficult. A nationwide prevalence survey for all Canadians is needed. This is essential to quantify the effectiveness of current recommendations [10] and conduct cost-effectiveness evaluations of alternative immunization programs, if necessary [69].
Abbreviations
G6 students grade 6 students
Competing interests
Funding for this systematic review was provided by GlaxoSmithKline Canada. BP and ACT are employed by GlaxoSmithKline. BD, GDS, VG, JO, and DS have received research funding from GlaxoSmithKline.
Authors' contributions
BP contributed to the development of the research question and methodology, project management, systematic review, data management and data analysis, derivation of charts and tables, and interpretation of the results. BD, GDS, DS contributed to the development of the research question and methodology, acquisition of unpublished data, and interpretation of the results. JO and VG contributed to the development of the research methodology, acquisition of unpublished data, derivation of charts and tables and interpretation of the results. ACT contributed to the development of the research methodology, project management, systematic review, derivation charts and tables and interpretation of the results. All of us contributed to the manuscript writing and approved the final version of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We would like to thank Perica Sever for her continued support. We would like to also thank Dr. Jun Wu for his guidance on how hepatitis A cases were reported as a notifiable disease in Canada. Additionally, we would also like to thank Richard Foty for compiling the data and Yang Liu for her technical assistance.
==== Refs
Cuthbert JA Hepatitis A: old and new Clin Microbiol Rev 2001 14 38 58 11148002 10.1128/CMR.14.1.38-58.2001
Gust ID Epidemiological patterns of hepatitis A in different parts of the world Vaccine 1992 10 Suppl 1 S56 S58 1335660 10.1016/0264-410X(92)90544-T
Tapia-Conyer R Santos JI Cavalcanti AM Urdaneta E Rivera L Manterola A Potin M Ruttiman R Tanaka KJ Hepatitis A in Latin America: a changing epidemiologic pattern Am J Trop Med Hyg 1999 61 825 829 10586919
Bell BP Shapiro CN Alter MJ Moyer LA Judson FN Mottram K Fleenor M Ryder PL Margolis HS The diverse patterns of hepatitis A epidemiology in the United States-implications for vaccination strategies J Infect Dis 1998 178 1579 1584 9815207 10.1086/314518
Notifiable disease annual summary 1997 Can Commun Dis Rep 1999 25 119 121
British Columbia Center for Disease Control BC reportable communicable diseases, annual summary 1998 1999 23 25
Furesz J Scheifele DW Palkonyay L Safety and effectiveness of the new inactivated hepatitis A virus vaccine CMAJ 1995 152 343 348 7828098
Scheifele DW Ochnio J Hepatitis A vaccine: is it being used to best advantage? CMAJ 2002 167 44 45 12137079
Practices ACI Prevention of hepatitis A through active or passive immunization MMWR Morb Mort Wkly Rep 1999 48 1 37
Samandari T Bell BP Armstrong GL Quantifying the impact of hepatitis A immunization in the United States, 1995-2001 Vaccine 2004 22 4342 4350 15474727 10.1016/j.vaccine.2004.04.014
Immunization NAC Canadian Immunization Guide 2002 6 Ottawa, Canadian Medical Association
Wu J Zou S Giulivi A Current hepatitis A status in Canada Can J Infect Dis Med Microbiol 2001 12 341 344
Population and Public Health Branch HC Notifiable disease incidence for years 1987-1999 2004
Scheifele DW Hepatitis A vaccines: the growing case for universal immunisation of children Expert Opin Pharmacother 2005 6 157 164 15757414 10.1517/14656566.6.2.157
Koff RS Seroepidemiology of hepatitis A in the United States J Infect Dis 1995 171 Suppl 1 S19 S23 7876642
Armstrong GL Bell BP Hepatitis A virus infections in the United States: model-based estimates and implications for childhood immunization Pediatrics 2002 109 839 845 11986444 10.1542/peds.109.5.839
elSaadany S Gully P Giulivi A Hepatitis A, B, and C in Canada. Results from the National Sentinel Health Unit Surveillance System, 1993-1995 Can J Public Health 2002 93 435 438 12448866
Payment P Antibody Levels to Selected Enteric Viruses in a French-Canadian Population in the Province of Quebec Immunol Infect Dis 1991 1 317 322
Cook D Wilton L D P Zou S Sherman M Krajden M Prevalence of antibodies to hepatitis A virus in a cohort of women of child-bearing age 68th Canadian Association for Clinical Microbiology and Infectious Disease, Nov 5-8, 2000 - Ottawa 2000
McAuley L Pham B Tugwell P Moher D Does the inclusion of grey literature influence estimates of intervention effectiveness reported in meta-analyses? Lancet 2000 356 1228 1231 11072941 10.1016/S0140-6736(00)02786-0
Deeks JJ Issues in the selection of a summary statistic for meta-analysis of clinical trials with binary outcomes Stat Med 2002 21 1575 1600 12111921 10.1002/sim.1188
Ochnio JJ Scheifele DW Fyfe M Bigham M Bowering D Martiquet P Ho M Talling DN The prevalence of hepatitis A in children in British Columbia. Can J Infect Dis Med Microbiol 2005 16 175 179
Muecke CJ Beliveau C Rahme E Soto JC Gyorkos TW Hepatitis A seroprevalence and risk factors among day-care educators Clin Invest Med 2004 27 259 264 15559862
Minuk GY Zhang M Wong SG Uhanova J Bernstein CN Martin B Dawood MR Vardy L Giulvi A Viral hepatitis in a Canadian First Nations community Can J Gastroenterol 2003 17 593 596 14571297
Ochnio JJ Patrick D Ho M Talling DN Dobson SR Past infection with hepatitis A virus among Vancouver street youth, injection drug users and men who have sex with men: implications for vaccination programs CMAJ 2001 165 293 297 11517645
Smieja M Cronin L Levine M Goldsmith CH Yusuf S Mahony JB Previous exposure to Chlamydia pneumoniae, Helicobacter pylori and other infections in Canadian patients with ischemic heart disease Can J Cardiol 2001 17 270 276 11264559
Kiefer LA Honish A Predy G Talbot JA The seroprevalence of hepatitis A and B in people testing positive for hepatitis C CMAJ 2000 162 207 208 10674054
Allard R Beauchemin J Bedard L Dion R Tremblay M Carsley J Hepatitis A vaccination during an outbreak among gay men in Montreal, Canada, 1995-1997 J Epidemiol Community Health 2001 55 251 256 11238580 10.1136/jech.55.4.251
Moses S Mestery K Kaita KDE Minuk GY Viral hepatitis in a Canadian street-involved population Can J Public Health 2002 93 123 128 11963516
Roy E Haley N Leclerc P Cedras L Bedard L Allard R Seroprevalence and risk factors for hepatitis A among Montreal street youth Can J Public Health 2002 93 52 53 11925701
Ochnio JJ Scheifele DW Ho M Hepatitis A virus infections in urban children--are preventive opportunities being missed? J Infect Dis 1997 176 1610 1613 9395375
De Serres G Laliberte D Hepatitis A among workers from a waste water treatment plant during a small community outbreak Occup Environ Med 1997 54 60 62 9072036
De Serres G Levesque B Higgins R Major M Laliberte D Boulianne N Duval B Need for vaccination of sewer workers against leptospirosis and hepatitis A Occup Environ Med 1995 52 505 507 7663634
Smieja M Gnarpe J Lonn E Gnarpe H Olsson G Yi Q Dzavik V McQueen M Yusuf S Multiple infections and subsequent cardiovascular events in the Heart Outcomes Prevention Evaluation (HOPE) Study Circulation 2003 107 251 257 12538424 10.1161/01.CIR.0000044940.65226.1F
Embil JA Manley K White LA Hepatitis A: a serological study in the Canadian Armed Forces Mil Med 1989 154 461 465 2552354
Nicolle LE Minuk GY Postl B Ling N Madden DL Hoofnagle JH Cross-sectional seroepidemiologic study of the prevalence of cytomegalovirus and herpes simplex virus infection in a Canadian Inuit (Eskimo) community Scand J Infect Dis 1986 18 19 23 3008308
Minuk GY Ling N Postl B Waggoner JG Nicolle LE Hoofnagle JH The changing epidemiology of hepatitis B virus infection in the Canadian north Am J Epidemiol 1985 121 598 604 4014149
Crewe MD Embil JA Garner JB Prevalence of antibodies to hepatitis A virus in Nova Scotia children Can Med Assoc J 1983 128 1195 1197 6301669
Minuk GY Waggoner JG Jernigan R Nicolle LE Postl B Hoofnagle JH Prevalence of antibody to hepatitis A virus in a Canadian Inuit community Can Med Assoc J 1982 127 850 852 7139503
Minuk GY Nicolle LE Postl B Waggoner JG Hoofnagle JH Hepatitis virus infection in an isolated Canadian Inuit (Eskimo) population J Med Virol 1982 10 255 264 7161596
McFarlane ES Embil JA Manuel FR Gorelick M Prevalence of antibodies to hepatitis A antigen in patients attending a clinic for treatment of sexually transmitted diseases Sex Transm Dis 1980 7 87 89 6893088
Buchner BK Staszak M Duravetz J Moore BP Hepatitis markers detected in a Canadian blood donor population Acta Haematol Pol 1980 11 69 71 6252746
Minuk GY Ding LX Hannon C Sekla L The risks of transmission of acute hepatitis A and B virus infection in an urban centre J Hepatol 1994 21 118 121 7963411 10.1016/S0168-8278(94)80147-9
Minuk GY Uhanova J Viral hepatitis in the Canadian Inuit and First Nations populations Can J Gastroenterol 2003 17 707 712 14679425
McFarlane ES Embil JA Antibody to hepatitis A virus in young inmates J Infect Dis 1982 146 442 7108284
McFarlane ES Embil JA Prevalence of antibodies to Hepatitis A antigen in sera from patients with haematological malignancies J Med Virol 1981 8 49 53 6271921
Wu H Zaniewski G Hong Z Wu J Dawood MR Smart G Geodhuis N Giulivi A Changes in the epidemiology of hepatitis A infection in Manitoba 1992-2003: Implications for vaccination strategies Can J Infect Dis Med Microbiol 2004 15 356 (Abstract P25)
Duval B De Serres G Gilca V Ochnio J Scheifele DW Nationwide Canadian study of hepatitis A antibody prevalence among children eight to thirteen years old Pediatr Infect Dis J 2005 24 514 519 15933561 10.1097/01.inf.0000164705.74498.86
Ochnio JJ Scheifele DW Marion SA Bigham M Ho M Hepatitis A virus (HAV) infections in young adults in British Columbia: feasibility of a novel mail-based approach to population sampling Proceedings of the 7th Annual Conference on Vaccine Research Arlington, VA, USA 2004 Abstract P4
Harb J Hepatitis A in the Northern Interior of British Columbia: An Outbreak Among Members of a First Nations Community Can Commun Dis Rep 2000 26 157 161 11057007
Levy I Chen D Sherman M Smith D Krajden M Hepatitis A virus seroprevalence in 1,000 university students in Toronto. Can Commun Dis Rep 2001 27 93 96 11416942
EL FJ I K M C M P R N N D R G Seroprevalence of hepatitis A (HAV), hepatitis B (HBV), human parvovirus B19 (HPV B19) and toxoplasmosis (Toxo) in Toronto infant-toddler day care providers 35th Interscience Conference on Antimicrobial Agents and Chemotherapy, Sept 17-20, 1995 1995 San Francisco, CA. Washington: American Society for Microbiology, 1995 (Abstract K154)
Ochnio JJ Traveler Seroprevalence Data in Vancouver 1995
Kocuipchyk FR Seroprevalence of hepatitis A antibodies in travellers at the Edmonton Travellers' Health Clinic--Alberta Can Commun Dis Rep 1995 21 65 71 7787817
Richer G Chen YY Huet PM Incidence of hepatitis non-A, non-B compared with types A and B in hospital patients Can Med Assoc J 1982 127 384 386 6809306
Basrur V Toronto Staff Report: Hepatitis A at a Toronto grocery store Toronto Public Health 2002
Pollett GL Middlesex-London Health Unit Report No. 128-02: Hepatitis A in a foodhandler 2002
Authority VCH Hepatitis A in a Capers grocery store in Vancouver Infectious Diseases News Brief 2002
Jacobsen KH Koopman JS Declining hepatitis A seroprevalence: a global review and analysis Epidemiol Infect 2004 132 1005 1022 15635957 10.1017/S0950268804002857
Beutels M Van Damme P Vranckx R Meheus A The shift in prevalence of hepatitis A immunity in Flanders, Belgium Acta Gastroenterol Belg 1998 61 4 7 9629763
Morris-Cunnington MC Edmunds WJ Miller E Brown DW A population-based seroprevalence study of hepatitis A virus using oral fluid in England and Wales Am J Epidemiol 2004 159 786 794 15051588 10.1093/aje/kwh107
Prodinger WM Larcher C Solder BM Geissler D Dierich MP Hepatitis A in Western Austria--the epidemiological situation before the introduction of active immunisation Infection 1994 22 53 55 8181844 10.1007/BF01780769
Bottiger M Christenson B Grillner L Hepatitis A immunity in the Swedish population. A study of the prevalence of markers in the Swedish population Scand J Infect Dis 1997 29 99 102 9181642
Hollinger FB LSMMHS Global impact of hepatitis A virus infection: changing patterns Viral Hepatitis and Liver Disease 2005 Baltimore, Williams and Wilkins 14 20
Hadler SC cala de MM Bensabath G Martinez DM Schatz G Fields HA Epidemiology of hepatitis delta virus infection in less developed countries Prog Clin Biol Res 1991 364 21 31 2020698
Hutin YJ Pool V Cramer EH Nainan OV Weth J Williams IT Goldstein ST Gensheimer KF Bell BP Shapiro CN Alter MJ Margolis HS A multistate, foodborne outbreak of hepatitis A. National Hepatitis A Investigation Team N Engl J Med 1999 340 595 602 10029643 10.1056/NEJM199902253400802
Rosenblum LS Mirkin IR Allen DT Safford S Hadler SC A multifocal outbreak of hepatitis A traced to commercially distributed lettuce Am J Public Health 1990 80 1075 1079 2382744
Fulford M Keystone JS Health Risks Associated with Visiting Friends and Relatives in Developing Countries Curr Infect Dis Rep 2005 7 48 53 15610671
Krahn M Remis R Hogg R Weber A Wu J Sherman M Shimian Z Universal hepatitis A vaccination for adolescents and children in Canada: a cost effectiveness analysis 2005
|
16001978
|
PMC1183208
|
CC BY
|
2021-01-04 16:28:14
|
no
|
BMC Infect Dis. 2005 Jul 7; 5:56
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-56
|
oa_comm
|
==== Front
BMC Med EducBMC Medical Education1472-6920BioMed Central London 1472-6920-5-211597210010.1186/1472-6920-5-21Research ArticleWell-being in residency training: a survey examining resident physician satisfaction both within and outside of residency training and mental health in Alberta Cohen Jordan S [email protected] Scott [email protected] Department of Psychiatry, University of Calgary Faculty of Medicine Foothills Medical Center (Special Services Building) 1403-29th St. NW Calgary, Alberta, Canada T2N 2T92 Departments of Community Health Sciences and Psychiatry, University of Calgary. 1403-29th St. NW Calgary, Alberta, Canada T2N 2T92005 22 6 2005 5 21 21 5 2 2005 22 6 2005 Copyright © 2005 Cohen and Patten; licensee BioMed Central Ltd.2005Cohen and Patten; 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
Despite the critical importance of well-being during residency training, only a few Canadian studies have examined stress in residency and none have examined well-being resources. No recent studies have reported any significant concerns with respect to perceived stress levels in residency. We investigated the level of perceived stress, mental health and understanding and need for well-being resources among resident physicians in training programs in Alberta, Canada.
Methods
A mail questionnaire was distributed to the entire resident membership of PARA during 2003 academic year. PARA represents each of the two medical schools in the province of Alberta.
Results
In total 415 (51 %) residents participated in the study. Thirty-four percent of residents who responded to the survey reported their life as being stressful. Females reported stress more frequently than males (40% vs. 27%, p < 0.02). Time pressure was reported as the number one factor contributing to stress (44% of males and 57% of females). A considerable proportion of residents would change their specialty program (14%) and even more would not pursue medicine (22%) if given the opportunity to relive their career. Up to 55% of residents reported experiencing intimidation and harassment. Intimidation and harassment was strongly related to gender (12% of males and 38% of females). Many residents (17%) rated their mental health as fair or poor. This was more than double the amount reported in the Canadian Community Health Survey from the province (8%) or the country (7%).
Residents highly valued their colleagues (67%), program directors (60%) and external psychiatrist/psychologist (49%) as well-being resources. Over one third of residents wished to have a career counselor (39%) and financial counselor (38%).
Conclusion
Many Albertan residents experience significant stressors and emotional and mental health problems. Some of which differ among genders. This study can serve as a basis for future resource application, research and advocacy for overall improvements to well-being during residency training.
Residencyphysicianpost-graduate well-beingstressintimidationharassmentand resources
==== Body
Background
Studies have shown medical school training to be a source of significant stress. The expectations and responsibilities only increase during residency training [1]. Although today's residents no longer live in their respective training hospitals, the pressures of residency are still extremely high. Residents are expected to be proficient clinicians, educators, researchers and administrators by the time they have completed their training. Improvements have been implemented in North America to identify, manage, and reduce resident stress [2]. Still, many residents report psychological symptoms during residency, feelings of becoming less humanistic and more cynical and "burning out" [3,4].
Researchers have attempted to examine and categorize stressors experienced by residents, which are experienced both within residency and in their personal life [5]. Still others have tried to predict levels of stress [6]. Resident responses to stress that have been described in the literature include: depression, burnout, anger/irritability, anxiety and substance abuse [5]. Resident stress has even been demonstrated at the biological level using cortisol measurements [7]. Others have reported changes in mood patterns from enthusiasm and depression, to anger and fatigue [8]. Sleep deprivation alone, has been shown to predispose residents towards more medical errors, injuries, increased alcohol and drug use, and increased conflict with other healthcare staff [9]. In the worst cases, resident suicides have occurred, forcing researchers to more closely examine residency stresses [10]. National wellness programs are attempting to improve resources for all physicians in Canada and changes are being made to individual residency programs [11,12]. In order to negotiate the stressors of residency, residents must become proficient "jugglers of the mind," maintaining the balance of all the complex biological, psychological and social interactions that occur during their training.
Differences have been reported in how males and females deal with stress [13,14]. Other groups have found no gender differences in stress response [15]. In some instances, the prevalence of symptoms that occur as reactions to stress is greater for residents than for the general population [16]. Many stress responses do not differ between these groups [8] However, residents are a unique group that are in many ways responsible for the health of the broader population and thus one might believe they should be more carefully scrutinized for such disorders as substance abuse. Involvement in stress management can lead to an overall more positive experience in medical residency [17].
The only Canadian study to date that has examined stress in residency training was performed by Toews and associates [12]. This study examined four of the medical schools in the country and did not identify any concerns with residents dealing with stress. To the best of our knowledge, our survey, known to PARA as "the Happy Doc study," is the first in Canada that will examine both stressors and well-being resources in residency. This paper presents findings from the "Happy Doc Study," referred to subsequently in this manuscript as "the study." The study was performed in each of two medical schools in Alberta during the 2003 academic year. It was our hope that by surveying members of the resident association we could increase our knowledge of current stressors effecting residents and determine what well-being resources are required, to improve residency training on a national level.
Methods
Survey composition
The study was conducted during the 2003 academic year among members of the Professional Association of Residents of Alberta (PARA). PARA represents all the resident physicians employed in the province of Alberta. The membership list from PARA thereby served as a sampling frame for this survey. The entire population indexed in the sampling frame was considered eligible for inclusion. This included residents in all levels of academic training. These residents were undergoing postgraduate training at both the University of Alberta and the University of Calgary. This was a voluntary study that PARA believed necessary to deal with the future well-being of its membership. The survey was distributed and collected via the PARA board members primarily during academic days, local mailboxes and e-mail. Surveys were distributed both by representatives within each specialty group and non-responders were followed up by e-mail versions of the survey. Unique identifiers were used in order to maintain the confidentiality of all participants. All of the data were collected over a six month period (January through June).
The survey required approximately twenty minutes to complete. The survey needed to be brief in order to maximize the response rate. Because the goals of the study were broad-based and descriptive, a decision was made to cover a large number of variables using single items and small modules, rather than including a restricted set of detailed "gold standard" measures. The questionnaire was divided into five sections: demographics, stress, intimidation and harassment, well-being and resources. Survey questions included qualitative rating scales, multiple responses and yes/no questions. To minimize a bias in rating scales responses (response acquiescence bias); the survey included a mixture of positively and negatively stated items [18]. Where numbers were large enough for valid statistical inference, groups were stratified by gender. The stress section contained questions regarding sources of stress as well as methods for dealing with stress (e.g. what would you say is the most important thing contributing to feelings of stress you may have?; thinking about the ways you deal with stress, how often do you do each of the following?). The term "stress" was not formally defined in the survey. The purpose of this was to measure "perceived stress" which might vary both quantitatively and qualitatively among individuals. For similar reasons we did not define the term intimidation and harassment as this is also perceived differently among individuals. We also decided to keep the two terms (intimidation and harassment) linked to avoid any confusion among resident when completing the survey. Percentages reported for intimidation and harassment among subgroups (e.g. inappropriate verbal comments) reflect overall percent of residents completing the entire survey and not those that reported intimidation and harassment, except where directly quoted in text (e.g. n = 65/170).
In certain instances group data was collapsed to increase numbers within response categories. Well-being questions (section D of results) were derived from the Canadian Community Health Survey (CCHS) so that results could be compared with members of the general Canadian population [19]. As a part of the study we have also included mental illness screening questions from the CCHS. Individuals that were positively screened for these mental disorders then underwent in depth questioning in the CCHS regarding diagnoses of these disorders. We have used these screening questions just to identify possible symptoms in residents. To avoid any stigmatization and thus decreased response rate, we did not further screen these individuals for psychiatric diagnoses.
Resources questions focused on knowledge of current resources (list provided of all resources and residents were asked which resource they were aware of prior to the survey), perceived need for future resources, and barriers and limitations towards residents seeking aid (e.g. If you were in a situation where you were experiencing an emotional or mental health problem, how would you deal with it?). In addition, questions were asked regarding career decisions (e.g. If possible, I would consider changing my residency program? [Yes/No]) and family physicians utilization.
Statistics
Descriptive statistics were used to give an overview of the data as well as for comparison with results from a Canadian national survey, the Canadian Community Health Survey (CCHS) conducted by Statistics Canada . In cases where not all residents responded to individual questions, percentagesfall short of 100%. Confidence intervals, Chi-squares, and Fisher exact tests were used to compare differences between groups. p-values less then 0.05 were interpreted as indicating statistical difference. All percentages reported in this paper were rounded to the nearest whole number. In addition, decimal points were rounded to 2, except in cases where this would make the statistic difficult to comprehend (i.e. p-values < 0.001).
Results
(A) Demographics
The response rate for the survey was 51% (415/800). Of those residents who completed the survey, 47% (n = 195) were male and 52% (n = 217) were female. The median age was 29 years, with a range from 24–49 years. The marital status among residents revealed that 55% of residents surveyed were either married or in common-law relationships, 42% were never married, and 2% were widowed, separated or divorced. Among the residents in post-graduate programs, 40% graduated from an Alberta medical school, 45% graduated from a medical school within another Canadian province and 14% received their medical training from outside of Canada. Among all the levels of residency, 35% were from first year of post-graduate training, 30% were in second year, 15% were in third year, 8% were in fourth year, 10% were in fifth year, 2% were in sixth year and 0.2% were in their seventh year of training. The average number of hours worked per week among residents was 75 ± 16 hours. The number of hours worked ranged from 30 to 126 hours per week. There were no significant differences between genders with respect to marital status, residency specialty program, location of medical school training or mean number of hours worked per week.
(B) Stress
The amount of stress in life on most days was determined using a five point scale (0 = not at all stressful, and 5 = extremely stressful). Many residents (34%) rated most days of their life as 4 to 5 on this scale. In the last 12 months of residency 43% of residents found this period of time to be "quite a bit" up to "extremely" stressful. Overall both genders reported a good degree of ability to handle unexpected and difficult stress. Fourteen percent reported their abilities as excellent, 56% good, 27% fair and only 3% of residents reported poor abilities to handle unexpected and difficult problems. Most residents reported an ability to handle day to day life demands (97% reported fair to excellent ability), with no differences between genders.
Residents were asked to rate the degree to which they agreed with how certain stresses relate directly to their residency program in the past 12 months. There were no significant differences between genders in these questions. Almost a third of residents (31%) either disagreed or strongly disagreed that their residency program allowed them freedom to decide how they did their job. Many of the residents (62% either agreed or strongly agreed) felt that their residency was very hectic. Residents disagreed or strongly disagreed (61%) that residency was free of conflicting demands that others made. More than one quarter of residents (27%) disagreed or strongly disagreed with having a lot to say about what happened in their residency. Over half of the residents (56%) either agreed or strongly agreed that there was pressure on examinations and evaluations. Two thirds of residents (66%) agreed or strongly agreed that there was insufficient sleep and frequent call. Many reported significant pressure due to their clinical workload (60% either agreed or strongly agreed). Many residents (17%) felt that there was stress due to high rates of death among patients (agreed or strongly agreed).
Contentment with career
When asked about their residency program, 14% of residence would consider changing their training program. In addition, more than one fifth (22%) of residents who participated in the survey reported that they would pursue another career if they had it to do all over again. When asked if they would change their career if they could live their lives over again, more males than females considered this change in program (18% vs. 11%, p < 0.02).
Gender differences
When categories were collapsed, females reported more significant stress than males (40% vs. 27%, p < 0.02). They also reported a higher degree of stress in the last 12 months (50 vs. 36%, p = 0.019). Overall males rated their abilities to handle unexpected and difficult problems better than females (p = 0.02).
When asked to think about stresses in day to day life, residents rated different conditions as a source of stress (see Table 1). Categories were collapsed to a 3 point scale (from original 7 point scale), from conditions causing less stress to more stress. Gender differences existed in time pressure, and caring for own children stress condition categories. Woman reported time pressure as being more stressful than males (79% vs. 64%, respectively, p < 0.001). However, more males reported caring for their own children as being either moderately or more stressful (11% and 8%, respectively) compared to females (4% and 9%, for moderate and more stress, respectively, p = 0.02).
Table 1 Rating of amount of stress in day to day life due to various conditions, separated by gender (actual numbers reported in brackets below percent responses).
Possible sources of stress in day to day life Less stress (%) Moderate stress(%) More stress (%)
Male Female Total Male Female Total Male Female Total
Time pressure/not enough time a 18%
(36) 6%
(14) 12%
(50) 17%
(33) 13%
(29) 15%
(62) 64%
(124) 79%
(171) 71%
(295)
Own physical health problem or condition 77
(150) 78
(169) 77
(319) 11
(22) 8
(17) 10
(39) 11
(22) 14
(31) 13
(53)
Own emotional or mental health problem/condition 76
(148) 73 (160) 74
(308) 14
(28) 13
(29) 14
(57) 9
(17) 12
(25) 10
(42)
Financial situation 36
(70) 50
(108) 43
(178) 15
(29) 12
(26) 14
(55) 47
(92) 36
(78) 41
(170)
Own work situation 34 (67) 27 (59) 31 (126) 20 (39) 21 (46) 20 (85) 42 (83) 50 (109) 46 (192)
Residency program 40 (78) 43 (93) 41 (171) 24 (46) 25 (54) 24 (100) 35 (69) 31 (68) 33 (137)
Employment status 67 (131) 73 (158) 70 (289) 16 (31) 17 (36) 16 (67) 14 (28) 9 (20) 12 (48)
Caring for own children b 38
(74) 40
(86) 39
(160) 11
(22) 4
(8) 7
(30) 8
(16) 9
(20) 9
(36)
Caring for others 66
(129) 67
(145) 66
(274) 19
(37) 11
(24) 15
(61) 8
(15) 8
(18) 8
(33)
Other personal or family responsibilities 56
(109) 57
(124) 56
(233) 18
(36) 17
(37) 18
(73) 24
(46) 24
(51) 24
(97)
Personal relationships 53
(103) 56
(122) 54
(225) 25
(48) 21
(45) 22
(93) 21
(40) 23
(49) 22
(89)
Discrimination 81
(157) 83
(181) 82
(338) 8
(16) 9
(19) 8
(35) 8
(15) 6
(13) 7
(28)
Personal and family safety 84
(163) 88
(192) 86
(355) 8
(16) 6
(14) 7
(30) 5
(9) 5
(10) 5
(19)
a Females more stressed by time pressure (p < 0.001)
b Males more stressed by caring for own children (p < 0.020)
In addition, respondents had the opportunity to report any other conditions that contributed to their daily stress. These conditions included: the travel time required in rural medicine, large amount of information to learn, hospital politics between services, living separated from spouse, marital relationship, motivation, responsibilities at work, examinations, balancing learning and clinical service components of residency and illness in family members. When asked to rank the most important thing contributing to feelings of stress, both males and females ranked time pressure as their number one choice (see Table 3). Other highly ranked contributors to stress included financial situation, own work situation, residency program, residency program and personal relationships.
Table 3 Top resident ranked contributors to feelings of stress during residency between genders.
Contributor to stress Males Females Totals
Percent (%) Rank Percent (%) Rank Percent (%) Rank
Time pressure 44% 1 57% 1 51% 1
Financial situation 25 2 11 3 18 2
Own work situation 11 3 12 2 12 3
Residency program 7 4 9 4 8 4
Personal relationship 7 5 6 5 6 5
Dealing with stress
The questionnaire included a battery of questions about ways residents deal with stress. This was rated on a 4 point scale from 4 = never to 1 = often, which was collapsed to a 3 point scale for analysis (see Table 2). There were many residents that used positive approaches in dealing with stress (e.g. by talking to others). However, a significant amount of residents dealt with stress in less productive ways (e.g. by avoiding others).
Table 2 Frequency of the ways residents reported dealing with stress, separated by gender (actual numbers reported in brackets below percent responses).
Frequency of occurrence (%): Often Sometim es Rarely/ never
Male Female Total Male Female Total Male Female Total
Talk to others 48%a
(94) 80%
(172) 64%
(266) 44%b
(84) 18%
(38) 29%
(122) 8%
(15) 3%
(5) 5%
(20)
Avoid being with people 10
(19) 5
(11) 7
(30) 46
(89) 48
(105) 47
(194) 44
(85) 46
(100) 44
(185)
Sleep more than usual 9
(18) 12
(25) 10
(43) 33
(64) 37
(81) 35
(145) 57
(112) 51
(111) 54
(223)
Try to feel better by eating more or less c 5
(10) 18
(39) 12
(49) 31
(60) 39
(83) 35
(143) 64
(123) 43
(92) 52
(215)
Try to feel better by smoking more 2
(3) 0.46
(1) 1
(4) 2
(5) 1
(2) 2
(7) 25
(49) 21
(46) 23
(95)
Try to feel better by drinking alcohol d 3
(5) 0.46
(1) 1
(6) 18
(35) 12
(26) 15
(61) 79
(154) 87
(188) 83
(342)
Try to feel better by using drugs/meds e 0.49
(2) 3
(14) 95
(391)
Try to look on the bright side of things 44
(85) 53
(115) 48
(200) 47
(91) 38
(81) 42
(172) 8
(16) 9
(19) 8
(35)
Exercise 47
(91) 43
(94) 45
(185) 33
(65) 36
(78) 35
(143) 19
(37) 21
(45) 20
(82)
Pray or seek spiritual help 20
(39) 24(51) 22
(90) 28
54) 21
(46) 24
(100) 52
(101) 55
(119) 53
(220)
Relax by doing something enjoyable 47
(92) 49
(106) 48
(198) 43
(84) 42
(91) 42
(175) 9
(18) 8
(17) 8
(35)
Blame yourself 8
(15) 12
(25) 10
(40) 32 f
(62) 52
(113) 42
(175) 60
(96) 36
(77) 46
(173)
Wish the situation would go away 19
(37) 24
(52) 22
(89) 45
(87) 56
(121) 50
(208) 35 g
(69) 20
(43) 27
(112)
a Males talk to often others less than females (95% CI, 0.41–0.56 vs. 0.74–0.85).
b Males talk sometimes to others more than females (95% CI, 0.37–0.51 vs. 0.13–0.23).
c Females more often change eating habits compared to males (p < 0.001).
d There was a significant difference between those who drink often/sometimes and those who drink rarely/never (95% CI, 0.13–0.20 vs. 0.80–0.87).
e Numbers were too small to stratify by gender.
f Males blame themselves sometimes less than females (95% CI, 0.26–0.39 vs. 0.46–0.59).
g More males rarely wish the situation will go away compared to females (95% CI, 0.21–0.34 vs. 0.11–0.21).
(C) Intimidation and harassment
More than two-thirds of residents responding to the survey (n = 302; 73%) reported experiencing intimidation and harassment. Residents reported intimidation and harassment from many members of the healthcare team. The highest degree of reported intimidation and harassment was experienced from nurses (n = 166; 55%) (Fig. 1). In addition to our suggested categories, residents also had the opportunity to mention other sources of intimidation and harassment. Forty percent of residents reported intimidation and harassment from families of patients (n = 121). Many residents (n = 42; 14%) reported intimidation and harassment from secretaries/unit clerks.
Figure 1 *Sources of intimidation and harassment as reported by residents in form of percent response. *Note: respondents had the opportunity for multiple responses, accounting for the total percent response exceeding 100%.
Over two thirds (n = 273; 66%) of residents reported intimidation and harassment in the form of inappropriate verbal comments. Inappropriate or unwanted physical contact was experienced by (n = 21; 5%) of residents. Sexual harassment was reported by (n = 18; 4%) of residents. Work given as punishment was reported as a form of intimidation and harassment by (n = 39; 9%) of residents. Other residents reported that privileges/opportunities were taken away as their form of intimidation and harassment (n = 26; 6%). Finally, 3% (n = 10) of those reporting intimidation and harassment reported it occurring in the form of recrimination for reporting these incidences. Residents were questioned on what was the basis for the reported intimidation and harassment (Figure 2). Similar to Figure 1, multiple responses were permissible. The primary reported basis for intimidation and harassment was gender (n = 17/123, 12% of males; and n = 65/105, 38% of females). Due to the smaller sample sizes we were unable to look for statistical differences between genders in responses to these questions. Of the residents that reported intimidation and harassment 51% (n = 211) stated that this had occurred more than once. Only 52% (n = 215) of residents who reported intimidation and harassment were aware of the process to address this issue and only 44% (n = 181) of residents felt this process was adequate, fair and independent.
Figure 2 *Percentage distribution of the perceived bases for resident intimidation and harassment.* The Y-axis has been collapsed for better distinguishing differences among groups, but it should be noted that this only reflects 20% of the residents responding to the survey. As in Figure 1, multiple responses were possible from each resident.
(D) Well-being
The results for life satisfaction and self-rated mental health used in this survey were compared to the CCHS estimates for the province's (Alberta) population in Tables 4 and 5. CCHS estimates are for people in the 25 to 64 year age group. Most residents (78%; 74% of males, 82% of females) reported their satisfaction with life in general as either satisfactory of very satisfactory. This was lower than CCHS results for both the province of Alberta (85%) and for the entire country (85%). Seven percent of residents reported being dissatisfied or very dissatisfied with their life in general (7% of males, 6% of females). This was slightly higher for men than the CCHS results for Alberta (6% for males and 6% of females) and the country (5% of males and 5% of females).
Table 4 Comparison of relative frequencies of life satisfaction results between residents from the study and the provincial population from the CCHS.
Canada (%) Very Satisfied Satisfied Neither satisfied nor dissatisfied Dissatisfied or very dissatisfied
Alberta TS CCHS TS CCHS TS CCHS TS CCHS
Males 24% 29% 50% 55% 18% 10% 7% 6%
Females 21 35 61 52 12 8 6 6
Totals 22 32 56 54 15 9 6 6
CCHS = Canadian Community Health Survey, TS = The Study
Table 5 Comparison of relative frequencies of self-rated mental health results between residents from the study and the provincial population from the CCHS.
Canada (%) Excellent Very good Good Fair or poor
Alberta TS CCHS TS CCHS TS CCHS TS CCHS
Males 21% 27% 30% 35% 32% 31% 16% 7%
Females 14 23 30 36 38 32 18 9
Totals 17 25 30 35 35 32 17 8
CCHS = Canadian Community Health Survey, TS = The Study
Overall residents rated their mental health lower than CCHS data for both Alberta and the country. More male residents rated their mental health as excellent (20%) than females (14%). A similar discrepancy between genders occurred in the CCHS (26% of males and 23% of females in Alberta; and 30% of males and 26% of females throughout the country). Many more residents reported their mental health as fair or poor (17%) compared to CCHS data from the province (8% for Alberta) or the country (7%).
Table 6 show the percentage of residents that would require further screening for mental disorders based upon the results of symptoms screening questions from the CCHS for specific mental disorders. It should be noted that being positively screened does not infer a resident will have the disorder. The proportions screening positive was much smaller than the national CCHS data.
Table 6 Percentage of Residents that would require further screening for mental disorders based upon the results of symptoms screening questions from the CCHS* for specific mental disorders.
Type of mental disorder % Males screening positive for disorder % Females screening positive for disorder
"The Study" CCHS 1.2* "The Study" CCHS 1.2
Depression 30% 49% 35% 55%
Panic disorder 9 40 16 49
Bipolar Disorder 7 20 12 19
Generalized anxiety 5 40 12 48
Social Phobia 12 20 15 22
Agoraphobia 2 13 3 20
*CCHS = Canadian Community Health Survey. This project was a component of the World Mental Health 2000 project. The estimates are based on a national sample 15 years of age or over, weighted to national household resident demographics (n = 36,984), for more information, see .
Over half of residents reported having a family physician (55%). More female residents have a family doctor compared to their male counterparts (68% vs. 41%, p < 0.001). Of those who reported having a family doctor, 68% have had an appointment in the last 12 months. Again, more female residents (76%) than males (54%) reported visiting their family physician (p < 0.001).
(E) Resources
There was a large variation in resident awareness of the well-being resources available to them in Alberta (see Table 7). The majority of residents (86%) identified their program director as a possible resource. This was the only resident resource that over 80% of residents were aware of. Over one third of residents wished to have a career counselor (39%) and financial counselor (38%). Approximately one fourth of residents would like a program ombudsman (26%) and resident support group (24%). Residents were asked to rate well-being resources on a seven point, scale (1 = not important, 7 = extremely important). Based on this scale, resident colleagues (67%), program directors (60%) and external psychiatrist/psychologist (49%) were rated as extremely important most frequently. When asked to rank the top three well-being resources a resident would use if they were in a situation where they were experiencing an emotional or mental health problem, resident colleagues were ranked as the number one choice, followed by program director and external psychiatrists and psychologists.
Table 7 Resident awareness of well-being resources within the province of Alberta.
Resource Number Percent (%)
Telephone hotline 258 63%
Physician and family support group 294 71
Resident advocate 264 64
External psychiatrist or psychologist 156 38
Emergency consultation service 118 29
Program director * 354 86
Chief resident 218 53
Resident colleague 280 68
Health region 59 14
University 59 14
* Identified as most recognized well-being resource by residents completing the survey.
Unfortunately, 9% stated they would not seek help for emotional or mental health concerns. Reasons for not seeking help included (summarized into categories): the ability and desire to handle these situations on their own or with assistance exclusively from family or friends, fear of repercussion or recrimination, stigma/embarrassment/social stigma, accessibility (time, money), and issues of confidentiality. Estimates from the CCHS for those who responded that they had not sought help for an emotional or mental health issue in the past 12 months because of acceptability issues (i.e. competing demands on time, attitudes towards illness, health care providers or the health care system) indicated that 4% of were in this category.
Residents were asked to choose one or more ways that they would deal with a fellow resident who was experiencing an emotional health problem. The majority of residents would suggest that the resident get help (85%). Many residents would offer to go with the resident for help (76%). A quarter of residents would contact their program director (25%). Less than one in ten residents would notify the Post-Graduate Medical Education office (2%), the provincial resident association (8%), the Royal College of Physicians and Surgeons (1%), and the Alberta Medical Association (3%). Four percent of residents responded that they would do nothing.
Discussion
While the rate of return (51%) was disappointing, it is comparable to other studies of this type [20]. More than two-thirds of responders (67%) to the survey were in first or second year of training. Thus, the results are more representative of residents who are less experienced in their training. Less experience might lead to increased stress. This possible response bias may have lead to an increased reporting of stress. However, residents that are more senior in training have other stresses that may be equally concerning (e.g. final examinations, higher expectations). The number of hours worked per week (75 ±16) was consistent with previous surveys of the PARA membership. In addition, since the survey was performed over a period of six months (January to June, 2004) it is possible that responses were biased based on the higher prevalence of stressors, such as depressed mood during the winter months.
Compared to previous Canadian studies of resident stress in which self-reported stress was not stated as a concern [14], residents in the study reported most days of their life as stressful (34%). Similar to the research done by Toews and associates [14], we also found that females reported a higher degree of stress than males (40% vs. 27%, p < 0.02). This may be due to stresses that are unique to the female gender [21]. One must also wonder whether the results are due to a reporting bias, in that females tend to be more open about their stress than their male counterparts.
Unfortunately, residents also had difficulties dealing with stress and resorted to more troublesome behaviors. A significant amount of residents reported often turning to alcohol to deal with stress and just less than 5% reported using drugs or medication to feel better (sometimes or often). These numbers are not significantly different than the population, but are quite concerning when considering the responsibilities of this group of professionals [8].
A large portion of the residents surveyed would consider changing their programs if given the opportunity. This speaks to the need for post-graduate medical education to ensure there is increased flexibility in residency, by taking measures such as increasing the amount of re-entry positions. Even more concerning was that over one fifth of residents reported that they would pursue another career if they had it to do all over again. Clearly, this speaks to the need to improving resident well-being in training.
Many residents reported experiencing intimidation and harassment. This result is consistent with the studies examining resident bullying [20]. The main form of this intimidation and harassment was in inappropriate verbal comments (66%). These results seem quite different from a previous study of psychiatry residents in Edmonton, Alberta, which concluded that intimidation in the psychiatric educational environment was not a significant issue [23]. However, due to the setup of our study we did not choose to stratify results from individual programs and therefore cannot directly comment on the psychiatry residents results. The primary basis reported for the intimidation and harassment was gender (12% of males and 38% of females), a difference that did not attain statistical significance, possibly because of the small sample size. However, it is quite possible that intimidation and harassment is one of the reasons that females in our survey reported more stress. This difference is consistent with recent publications that revealed increased female reporting of resident bullying [22,24]. Intimidation and harassment occurred often multiple times (more than once in 52% of those responding to the study) in both genders. Over half of the residents felt that the process to deal with it was not adequate, fair and independent. This speaks to the need for further educating all individuals in the healthcare system on resident well-being.
It appears that although the majority of residents are quite resilient to all of life's stresses during training. However, there is a significant group that seems to be having difficulty with their own well-being during this period of their lives (i.e. fair or poor life satisfaction and rated mental health), possibly to an extent below the levels in the general population.
Due to our study's design, we cannot predict what proportion of individuals had a mental illness, nor compare rates to the normal population in the province or country. However, ratings obtained with the CCHS screening questions did not suggest a higher prevalence of disorders. It would not be surprising if the prevalence of specific disorders were lower in this highly selected professional group than in the general population. The results suggest, however, that non-illness-related issues represent the main difference between residents and the general population. Another well-being concern was that many residents still do not have a family physician and a significant amount did not use them (no appointment in the twelve months prior to being surveyed). With all the stresses or residency and the potential for decreased physical health and well-being, the need for more residents to acquire a family physician to be available for dealing with such issues is crucial [25]. The Canadian Psychiatric association has position statements both on the treatment of mentally ill physicians and on trainee safety [25,26]. Some of the recommendations include: 1) that any physician with a possible psychiatric illness should receive an assessment quickly, ideally by a psychiatrist who is not a colleague or friend; 2) that the treating psychiatrist must urge the physician-patient to obtain a family physician as soon as possible and aid in this process as necessary; and 3) all provinces should have psychiatrists who serve on, or consult to, their physician well-being committees. In addition, the trainee safety position statement recommends that minimum standards exist for resident safety and that there should not be any coercion of trainees to see potentially violent patients.
Based on the wide variation of awareness of many well-being resources more education should be applied to this area. Resident career and financial counseling were the highest ranked well-being resource. This was likely rooted in the fact, that many residents's reported high reported stress due to financial situation and the dissatisfaction with residency training and the medical profession. Resident colleagues, program directors and psychiatrist/psychologist(s), were the top resources residents preferred in times of emotional or mental health need. There is a definite need to properly train and educate program directors and all residents in how to deal with well-being concerns. Adequate psychiatric/psychological aid to residents in the province must be an important priority [25]. The majority of residents reported that they would intervene to aid a colleague having emotional difficulties. Most often by suggesting they go for help (85%) or by offering to go with them for help (76%). Only a small portion of the residents would inform any medical organization. This may suggest that while residents want to help their peers, they prefer to do so in ways that do don't involve notifying external guarding bodies.
Conclusion
It is clear that there are significant stressors incurred during residency. Some of which have differences between genders. Intimidation and harassment occurs among many residents. It is also important to recognize that a significant amount of residents are vulnerable to emotional and mental health concerns. Residents need to be better informed about well-being resources. It is not clear as to whether the resources resident's perceived to be important are the necessary aids to deal with well-being concerns during training. However, ensuring the education of other healthcare professionals in the area of well-being is needed, so that resident's who ask for help will be directed the correct sources.
Now that this pilot study is complete we are working on administering the happy doc survey throughout the entire country. Once this data is collected comparisons can be made inter-provincially. Results of this collaboration can be used to help identify key stressors in residency and develop well-being resources for the future.
Abbreviations
AMA: Alberta Medical Association
CCHS: Canadian Community Health Survey
PARA: The Professional Association of Residents of Alberta
TS: The Study (referring to the survey presented in this paper).
Authors' contributions
JC conceived the study, and participated in its design and coordination and drafted the manuscript. He is a physician in residency training in Psychiatry at the University of Calgary. He is the past president of PARA and past member on the executive of the Canadian Association of Interns and Residents (CAIR).
SP participated in the design of the study, read and approved all stages of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
The complete version of the resident physician questionnaire utilized for this study - can be referenced in text in the methodology as soon as any ref. made to survey/questionnaire.
Click here for file
Acknowledgements
We would like to thank Jeanne Williams who provided her research assistance and statistical analysis skills to this paper. We would like to acknowledge the Alberta Medical Association (AMA) and the Professional Association of Residents of Alberta (PARA) for providing funding for this research.
==== Refs
Woloschuk W Harasym P Mandin H Implementing a clinical presentation curriculum: impact on student stress and workload Teach Learn Med 1998 10 44 50 10.1207/S15328015TLM1001_8
Collier VU McCue JD Markus A Smith L Stress in medical residency: status quo after a decade of reform? Ann Intern Med 2002 136 384 390 11874311
Hillhouse JJ Adler CM Walters DN A simple model of stress, burnout and symptomatology in medical residents: a longitudinal study Psych Health Med 2000 5 63 73 10.1080/135485000106016
Thomas NK Resident burnout JAMA 2004 292 2880 2889 15598920 10.1001/jama.292.23.2880
Levey RE Sources of stress for residents and recommendations for programs to assist them Acad Med 2001 70 142 150 11158832
Daly MG Wilcock SM Examining stress and responses to stress in medical students and new medical graduates MJA 2002 177 S14 21 12088494
Coeck C Jorens PG Vandevivere J Mahler C ACTH and cortisol levels during residency training N Engl J Med 1991 325 738 1651455
Hurwitz TA Beiser M Nichol H Patrick I Kozak J Impaired interns and residents Can J Psychiatry 1987 32 165 169 3567830
Baldwin DC JrDaugherty SR Sleep deprivation and fatigue in residency training: results of a national survey of first- and second-year residents Sleep 2004 27 371 372 15164886
Williams LS Manitoba suicides force consideration of stresses facing medical residents Can Med Assoc J 1997 156 1599 1602 9176428
MacDonald NE Davidson S The wellness program for medical faculty at the University of Ottawa: a work in progress CMAJ 2000 163 735 738 11022590
Tennant CC A student mental health and welfare program in medical faculty MJA 2002 177 S9 11 12088491
Hendrie HC Clair DK Brittain HM Fadul PE A study of anxiety/depressive symptoms of medical students, house staff, and their spouses/partners J Nerv Ment Dis 1990 178 204 207 2152447
Toews JA Lockyer JM Dobson DJ Simpson E Brownell AK Brenneis F others Analysis of stress levels among medical students, residents, and graduate students at four Canadian schools of medicine Acad Med 1997 72 997 1002 9387825
Tyssen R Vaglum P Grenvold NT Ekeberg O The impact of job stress and working conditions of mental health problems among junior house officers. A nationwide Norwegian prospective cohort study Medical Education 2000 34 374 384 10760123 10.1046/j.1365-2923.2000.00540.x
Hsu K Marshall V Prevalence of depression and distress in a large sample of Canadian residents, interns, and fellows Am J Psychiatry 1987 144 1561 1566 3688279
Shapiro SL Shapiro DE Schwartz GER Stress management in medical education: a review of the literature Acad Med 2000 75 748 759 10926029
Guyatt GH Cook DJ King D Norman GR Kane SC van Ineveld C Effect of the framing of questionnaire items regarding satisfaction with training on resident responses Acad Med 1999 74 192 194 10065060
Statistics Canada, Canadian Community Health Survey, Mental Health and Well-being, Canada 2002
Paice E Rutter H Wetherell M Winder B McManus IC Stressful incidents, stress and coping strategies in the pre-registration house officer year Med Educ 2002 36 56 65 11849525 10.1046/j.1365-2923.2002.01101.x
Gautam M Females in medicine: stresses and solutions West J Med 2001 174 37 41 11154666 10.1136/ewjm.174.1.37
Quine L Workplace bullying in junior doctors: questionnaire survey BMJ 2002 324 878 879 11950736 10.1136/bmj.324.7342.878
Tibbo P de Gara CJ Blake TM Steinberg C Stonehocker B Perceptions of intimidation in the psychiatric educational environment in Edmonton, Alberta Can J Psychiatry 2002 47 562 567 12211885
Paice E Aitken M Houghton A Firth-Cozens J Bullying among doctors in training: cross sectional questionnaire survey BMJ 2004 329 658 659 15256417 10.1136/bmj.38133.502569.AE
Myers MF Canadian Psychiatric Association: Position paper: treatment of the mentally ill physician 1996
Moscovitch A Chaimowitz GA Patterson PGR Canadian Psychiatric Association: Position paper: trainee safety in psychiatric units and facilities 1990
|
15972100
|
PMC1183209
|
CC BY
|
2021-01-04 16:30:56
|
no
|
BMC Med Educ. 2005 Jun 22; 5:21
|
utf-8
|
BMC Med Educ
| 2,005 |
10.1186/1472-6920-5-21
|
oa_comm
|
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-351595523910.1186/1471-2180-5-35Research ArticleA census of membrane-bound and intracellular signal transduction proteins in bacteria: Bacterial IQ, extroverts and introverts Galperin Michael Y [email protected] National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD 20894, USA2005 14 6 2005 5 35 35 18 4 2005 14 6 2005 Copyright © 2005 Galperin; 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 complete microbial genomes showed that intracellular parasites and other microorganisms that inhabit stable ecological niches encode relatively primitive signaling systems, whereas environmental microorganisms typically have sophisticated systems of environmental sensing and signal transduction.
Results
This paper presents results of a comprehensive census of signal transduction proteins – histidine kinases, methyl-accepting chemotaxis receptors, Ser/Thr/Tyr protein kinases, adenylate and diguanylate cyclases and c-di-GMP phosphodiesterases – encoded in 167 bacterial and archaeal genomes, sequenced by the end of 2004. The data have been manually checked to avoid false-negative and false-positive hits that commonly arise during large-scale automated analyses and compared against other available resources. The census data show uneven distribution of most signaling proteins among bacterial and archaeal phyla. The total number of signal transduction proteins grows approximately as a square of genome size. While histidine kinases are found in representatives of all phyla and are distributed according to the power law, other signal transducers are abundant in certain phylogenetic groups but virtually absent in others.
Conclusion
The complexity of signaling systems differs even among closely related organisms. Still, it usually can be correlated with the phylogenetic position of the organism, its lifestyle, and typical environmental challenges it encounters. The number of encoded signal transducers (or their fraction in the total protein set) can be used as a measure of the organism's ability to adapt to diverse conditions, the 'bacterial IQ', while the ratio of transmembrane receptors to intracellular sensors can be used to define whether the organism is an 'extrovert', actively sensing the environmental parameters, or an 'introvert', more concerned about its internal homeostasis. Some of the microorganisms with the highest IQ, including the current leader Wolinella succinogenes, are found among the poorly studied beta-, delta- and epsilon-proteobacteria. Among all bacterial phyla, only cyanobacteria appear to be true introverts, probably due to their capacity to conduct oxygenic photosynthesis, using a complex system of intracellular membranes. The census data, available at , can be used to get an insight into metabolic and behavioral propensities of each given organism and improve prediction of the organism's properties based solely on its genome sequence.
==== Body
Background
All living organisms adjust their metabolism and behavior in response to the changes in their environment. For unicellular microorganisms, knowing themselves, i.e. constantly monitoring a variety of environmental and intracellular parameters, is a necessary condition of survival. Mechanisms of some adjustments can be as simple as those in the lac operon – the presence of a substrate induces expression of the genes that are necessary for assimilation of that substrate (although even lac operon has a complex high-level regulation through catabolite repression and inducer exclusion, see [1] and references therein). More complex regulatory mechanisms include transmission of an external signal across the cytoplasmic membrane, followed by intracellular signal transduction to the appropriate genes (operons), metabolic enzymes, or to such organelles as bacterial flagella. Given that all these mechanisms have to be encoded in the organism's genome, the complexity of the signaling systems correlates with the genome size and the range of environmental challenges it normally encounters. Bacterial parasites that inhabit relatively stable host environments typically encode few, if any, signaling proteins (see [2-4]).
Analysis of the first three sequenced microbial genomes revealed very few signaling systems: four histidine kinases (HKs), five response regulators (RRs) and no methyl-accepting chemotaxis proteins (MCPs) in Haemophilus influenzae, none of these in Mycoplasma genitalium or Methanococcus (recently renamed Methanocaldococcus) jannaschii. Analysis of the fourth sequenced organism, the freshwater cyanobacterium Synechocystis sp. PCC 6803, revealed 42 HKs and 38 RRs [5], whereas the fifth, Mycoplasma pneumoniae, again had none. The list of signaling proteins encoded in microbial genomes grew by leaps and bounds ever since, generally following the exponential increase in the number of completely sequenced genomes and the total number of proteins that they encode (Figure 1). Given the importance of two-component signal transduction in bacteria [6,7], the numbers of HKs and RRs were routinely reported in many genome descriptions. However, due to the limitations of employed algorithms and arbitrarily high cut-off values in most sequence comparison protocols, certain HK variants were often missed, for example, the HKs of the LytS family (family HPK8 in the classification of Grebe and Stock [8,9]). Some HKs of the recently described HWE family [10] have not been recognized as HKs even when compared against SMART [11,12] and Pfam [13,14] domain databases [15]. Because of that, HKs were systematically undercounted: the number of HKs in E. coli, first reported to be 28 [16], was then revised upwards to 29 [2,17] and now stands at 30 [18]; [see Additional file 1]). Likewise, the number of HKs encoded by Synechocystis sp. PCC 6803, originally estimated to be 42 [5], has been revised to 46 [2]. As a result, most estimates of the HK numbers published in previous years are unreliable. Besides, listings of signal transduction proteins typically did not take into account Ser/Thr/Tyr-specific protein kinases (STYKs) and protein phosphatases, which, as we now know, were encoded in the H. influenzae, M. genitalium, and M. jannaschii genomes [see Additional file 1], see [19,20]. Further, cross-genome comparisons revealed entirely new classes of signaling molecules with GGDEF and EAL domains, involved in the turnover of the c-di-GMP, a novel secondary messenger [21,22]. Although genetic data and sequence considerations have long pointed to the diguanylate cyclase (c-di-GMP synthetase) activity of the GGDEF domain and the phosphodiesterase (c-di-GMP hydrolase) activity of the EAL domain, direct biochemical proof that this is indeed the case has became available only in the past year [23-25], reviewed in [21,22]. Predicted phosphodiesterase activity of the HD-GYP domain [26] has never been experimentally verified. Finally, although participation of cellular adenylate cyclases (ACs) in signal transduction was never in question, class 3 enzymes (AC3s) were recognized as legitimate environmental sensors only last year, when they were shown to function as light receptors modulating motility in cyanobacteria [27,28]. Adenylate cyclases of class 1 and class 2, represented by experimentally characterized proteins from E. coli (AC1, [29]) and Aeromonas hydrophila (AC2, [30]), are cytoplasmic enzymes of relatively narrow phylogenetic distribution [see Additional file 1] and are not known to function as environmental sensors.
The diversity of the signal transduction systems made careful accounting for all of them a daunting task, further complicated by the paucity of the data on the cellular targets for STYKs [31] and virtual absence of any data on the mechanisms of c-di-GMP-mediated regulation [21,22]. Hence, most signaling protein surveys focused exclusively on certain classes of membrane receptors (HKs and/or MCPs) and RRs [5,16,17,32-34], or on certain organisms, mostly cyanobacteria and actinobacteria [35-38]. Shi, Kennelly and Potts performed a comprehensive survey of STYKs and protein phosphatases [19,20,39], but have not looked at other signaling proteins. Galperin and colleagues [2,26] performed a census of HKs, GGDEF, and EAL domains but never considered STYKs or ACs. Surveys of the MCP and AC3 distribution in complete microbial genomes by Zhulin [40] and Shenoy and Visweswariah [41], respectively, were limited to these protein domains. The information on signaling systems is poorly represented in public databases. While HKs and RRs are covered in the KEGG database [42,43], other signaling systems are not. The SENTRA [44,45]), SMART [11,12] and COG [46,47] databases have a good coverage of the first sequenced genomes but have not been updated in a while, whereas data in other databases, such as Pfam [13,14] or PEDANT [48,49] are generated mostly by automatic means and therefore prone to the biases described above (and also in the Results section).
While preparing recent reviews on signal transduction in bacteria [3,22], the need for comprehensive and reliable data on the distribution of specific signaling systems among different phylogenetic lineages became all too obvious. Since signal transduction systems grow in number and complexity with the genome size and play increasingly important roles in environmental bacteria [3,4], it has become clear that comparative analysis of such systems could provide a useful insight into bacterial behavior [50]. Here I present a comprehensive census of HKs, MCPs, STYKs and ACs, as well as GGDEF, EAL, and HD-GYP domains encoded in complete genomes of 167 bacterial and archaeal species, sequenced by the end of 2004. I hope that availability of these data on a public web site [51], which will be updated as needed, will stimulate further analysis of microbial signal transduction and will lead to a better understanding of microbial behavior in various ecological niches.
Results
Scope of the study
Bacterial signaling mechanisms are extremely diverse, ranging from simplest two-domain transcription regulators, such as AraC or LacI, to multi-component signaling cascades that regulate sporulation, flagellar biosynthesis or biofilm formation. Until recently, the term 'signal transduction' has been typically reserved for the two-component systems consisting of a sensor histidine kinase (HK) and a response regulator (RR). In keeping with this tradition, I did not include in this survey single-component transcriptional regulators, whether of AraC type [52] or much more complex NorR type [53] and considered only dedicated signaling systems that consist of more than two individual components. In addition to HKs, these included Ser/Thr protein kinases, adenylate and diguanylate cyclases and two types of predicted c-di-GMP phosphodiesterases, containing, respectively, EAL or HD-GYP domains. Other enzymatic output domains as well as DNA- or RNA-binding response regulators have not been considered here but could be added to the list in the future. Because of the previously noted parallelism between the domain architectures of intracellular signaling proteins (e.g. PAS-GGDEF-EAL) and respective response regulators (e.g. CheY-PAS-GGDEF-EAL) [3], no attempt has been made to distinguish such proteins; they were counted both in the GGDEF and EAL columns. Naturally, such proteins were counted only once to obtain the total number of signaling proteins encoded in any given genome.
The data set included complete bacterial and archaeal genomes sequenced by the end of 2004. While Archaea and Bacteria are generally considered separate domains of life in the prokaryotic world, there are indications that many signal transduction systems in archaea have been acquired from bacteria through lateral gene transfer [2,32]. Hence, for the purposes of this study, domain Archaea was treated as just another bacterial phylum. Owing to the redundancy of the current genome list, only one representative genome per species was used in the analysis, typically the first one to be publicly released. Exceptions included two strains of Escherichia coli, K12 and O157:H7 [54,55], and three serovars of Salmonella enterica, Typhi, Typhimurium, and Paratyphi [56-58].
Data validation
The total numbers of copies of each signaling domain encoded in each given genome were estimated in iterative PSI-BLAST [59] searches, using the strict inclusion threshold expect values of 10-7–10-4, adjusting as necessary. Potential false-positive hits were checked at every step of PSI-BLAST using the CDD Domain viewer [60] and manually removed (unselected) from the hit list for the next iteration of PSI-BLAST. The most typical sources of the false-positive hits were as follows.
Histidine kinases consist of two separate domains, (i) a well-conserved ATPase domain of the GHKL family [61,62], referred to as HATPase_c domain [Pfam:PF02518 ] in the Pfam database [14], and (ii) a less-conserved phosphoacceptor (dimerization) domain, carrying the phosphorylatable His residue [7,63]. The dimerization domains are quite diverse in their sequence and comprise the His Kinase A (phosphoacceptor) domain clan in Pfam, which unifies four individual domain families: HisKA [Pfam:PF00512 ], HisKA_2 [Pfam:PF07568 ], HisKA_3 [Pfam:PF07730 , and HWE_HK [Pfam:PF07536 ]. Due to the great variability of the HisKA domains, the results of PSI-BLAST search are largely determined by the presence of the HATPase_c domain and often include other members of the GHKL family, such as DNA gyrase B and DNA repair protein MutL, as well as anti-sigma F factors (SpoIIAB-like Ser/Thr kinases). Due to the presence of long α-helices in the phosphoacceptor domains, they sometimes show spurious low-complexity hits.
Methyl-accepting protein (MCP) domain (PF00015) [Pfam:PF00015 ] contains long α-helices, which also attract low-complexity hits. However, the extremely high conservation of the (LI)LALNAAIEAARAGExGRGFAVVAxEVR sequence pattern allows a relatively easy recognition of false-positive hits.
Ser/Thr/Tyr kinase (STYK) domain (PF00069) [Pfam:PF00069 ] belongs to the Protein kinase superfamily clan in Pfam [14]. Other members of this clan, such as kinases of kanamycin, streptomycin, methylthioribose, homoserine, choline, and 3-deoxy-D-manno-octulosonic acid (KDO), are often retrieved in PSI-BLAST searches. In fact, the latter enzyme, KDO kinase (product of the waaP gene, PF06293 [Pfam:PF06293 ]) often gives much better BLAST scores than certain divergent Ser/Thr kinases. Most of the discrepancies between the data presented here and those in the KinG database [64,65] could be attributed to those false-positive hits. The most common false-negative hits were the putative protein kinases of ABC1/AarF family (PF03109 [Pfam:PF03109 ] or COG0661 []), which are somehow involved in ubiquinone biosynthesis, most likely by regulating this pathway [66]. It should be noted that although members of the ABC1 (activity of bc1) family are sometimes misannotated as ABC transporters or even ABC transporter substrate binding proteins, this appears to be due to a simple misunderstanding, which I have ignored and counted these proteins as protein kinases.
GGDEF domains (PF00990 [] from diverse bacteria have diguanylate cyclase activity [23,24] and are structurally related to the eukaryotic adenylate cyclase (AC3) domains [67]. While PSI-BLAST searches of GGDEF domains rarely produced any false positive hits, many GGDEF-related domains appeared to be inactivated, some were clearly truncated. The latter ones were excluded from the total count. The most interesting example included a conserved family of proteins (COG3887 []) comprising a fusion of a modified (likely inactivated) GGDEF domain and the DHH-family (PF01368 [Pfam:PF01368 ], [68]) phosphoesterase domain. Members of this family are encoded in genomes of most Firmicutes, including tiny genomes of some Mycoplasma spp., but their function remains unknown.
EAL, AC1, AC2, or AC3 domains (corresponding to the Pfam entries PF00563 [Pfam:PF00583 ], PF01295 [Pfam:PF01295 ], PF01928 [Pfam:PF01928 ], and PF00211 [Pfam:PF00211 ], respectively) did not return any false-positive hits in PSI-BLAST searches.
HD-GYP domain is a variant of the widespread HD-type phosphohydrolase (PF01966 [Pfam:PF01966 ], [69]) domain that contains a C-terminal subdomain with extra conserved residues [26]. Classical HD domains without the second subdomain often showed up as false-positive hits; these were filtered based on the total length of the BLAST alignment.
Whenever possible, the domain and protein counts were compared to the published data and all discrepancies were manually verified. Thus, this census has identified 92 HKs in Bradyrhizobium japonicum, 62 HKs in Mesorhizobium loti, and 48 HKs in Sinorhizobium meliloti [see Additional file 1], which was much more than 80, 47 and 40 HKs, respectively, recognized in these bacteria in a recent survey [34]. A comparison of the two sets revealed that most of the proteins missing from the HK list by Hagiwara et al.[34] comprise a conserved family (COG3920 []) with an unusual HisKA_2 (PF07568 []) dimerization domain, which, however, still contains a conserved His residue, confirming that these proteins are true HKs. This and other comparisons showed that, in most cases, different authors correctly identified the core sets of signaling proteins and most discrepancies could be attributed to the different ways of treating divergent, inactivated and truncated sequences. The approach adopted here was to take a middle ground, not counting clearly truncated and highly diverged sequences but keeping in the list full-length domains that might have had inactivating point mutations. For example, although Gly?Ala and Glu?Ala changes in the GGEE motif of the GGDEF domain have been shown to abrogate its diguanylate cyclase activity, sequences with such changes were still counted as diguanylate cyclases, while the truncated sequences in Methanococcus kandleri protein MK0296 [UniProt:Q8TYK1 ], Aeropyrum pernix protein APE1864 [UniProt:Q9YAS9 , or in COG3887 [] proteins (see above) were not. Likewise, Archaeoglobus fulgidus encodes a family of proteins that have a typical HK domain architecture but lack the HATPase domain. Such truncated sequences were not included in the total count [see Additional file 1] but still listed (marked with asterisks) in the supporting files. Since the signaling protein count was based on the domain count, monster multidomain proteins, combining various output domains, such as the hybrid HK-STYK [UniProt:O32393 ] described in Spirulina platensis [70] or the HK-GGDEF combination, found in Geobacter sulfurreducens protein GSU3350 [UniProt:Q747B7 ], have been counted more than once.
General trends
The census of signal transduction proteins encoded in complete microbial genomes [see Additional file 1] revealed several interesting trends. It has largely confirmed previous observations [2,4,71] that the total number of regulatory proteins encoded by each given organism genome positively correlates with the genome size (Figure 2a) and the total number of encoded proteins (Figure 2b): microbes with complex life styles generally have larger genomes and encode more sophisticated and diverse regulatory systems than parasites with their largely degraded genomes.
While small genome size (and the correspondingly low number of signaling systems) is often associated with pathogenicity, there are numerous pathogens with relatively large genomes (e.g. Bordetella parapertussis, Mycobacterium tuberculosis), as well as free-living organisms with very small genome sizes (e.g. Thermoplasma acidophilum, Aquifex aeolicus, see Figure 2a). Many free-living archaea encode very few, if any, proteins involved in signal transduction. For example, among 2977 proteins, encoded by the extreme thermoacidophile Sulfolobus solfataricus, only 9 are signaling (8 STYKs and an AC2). A similar picture has been reported in marine cyanobacteria [72] and is seen in the recently sequenced genome of the ruminal bacterium Mannheimia succiniproducens, which encodes just 5 HKs, an AC, and a STYK [see Additional file 1]. Apparently, the constant and nutrient-rich ruminal environment does not require much signal transduction. These data indicate that organisms inhabiting stable environments can get away with relatively simple signal transduction systems. In contrast, organisms that survive in diverse ecological niches, including facultative pathogens, such as the spirochetes Leptospira interrogans and Treponema denticola, typically encode complex sensory systems. Of course, sophisticated bacteria can also be found in simple and stable environments: Wolinella succinogenes, another ruminal inhabitant [73], encodes many more signal transduction proteins than other bacteria with similar genome sizes (Table 1, see below).
Bacterial IQ
The total number of signaling proteins encoded in a given genome (or, rather, the fraction of such proteins among all encoded in the genome) can be used as a measure of the adaptive potential of an organism, some kind of 'bacterial IQ'. The slope of the best-fit line on Figure 2a is 2.03, meaning that the total number of signal transduction proteins grows approximately as a square of the genome size. The organisms whose genomes deviate most from this trend can be considered particularly 'smart' or 'dumb' compared to their relatives. There could be different ways to evaluate the relative abundance of signal transduction proteins at the given genome size; the data in Table 1 were calculated using the following formula:
IQ = 5 × 104 (n-5)1/2 L-1,
where n is the total number of signal transduction proteins, L is the complete genome size in kb (even counting plasmids, it is a more consistent measure than the number of predicted proteins), 5 × 104 and 5 are arbitrarily chosen empirical coefficients, so that IQ = 100 corresponds to 9 signal transducers in a 1000 kb genome and to 105 transducers in a 5000 kb genome. Accordingly, the IQ value is not defined for organisms with less than 6 signal transduction proteins.
With one exception, all organisms listed in Table 1 are environmental gram-negative bacteria (most gram-positive bacteria and archaea scored much lower) that are highly motile and are known to use a wide variety of electron donors and electron acceptors [73-76]. Such versatile organisms as Chromobacterium violaceum, Desulfovibrio vulgaris, Geobacter sulfurreducens, Vibrio vulnificus, and Wolinella succinogenes are also repeatedly found among the leaders in individual categories (Table 2), both in terms of absolute number of signal transduction proteins and of their fraction among all encoded proteins. Remarkably, most of the winners come from the relatively poorly characterized beta-, delta- and epsilon- subdivisions of Proteobacteria. This illustrates the limitations of relying just on Escherichia coli and Bacillus subtilis as model organisms for studying signaling transduction in environmental organisms. The recent efforts on the post-genomic analysis of the versatile gamma-proteobacterium Shewanella oneidensis [77], which encodes a decent set of 46 HKs, 26 MCPs, 7 STYKs, 3 ACs and 52 GGDEF, 28 EAL, and 9 HD-GYP domains [see Additional file 1] might be a step in the right direction. In contrast, E. coli appears to have a relatively low IQ. Although its 30 HKs, 19 GGDEF and 17 EAL domains at first seemed like a high number [16,26], E. coli, as well as Salmonella spp. and Yersinia spp., other members of Enterobacteriaceae, looks pretty 'dumb' compared to the representatives of Pseudomonadaceae, Vibrionaceae, or Xanthomonadaceae, particularly with respect to chemotaxis: any sequenced member of the three latter families encodes many more MCPs than the meager 5 MCPs in E. coli. The deep-sea bacterium Idiomarina loihiensis, which belongs to yet another gamma-proteobacterial lineage and whose protein set is just 62% of that of E. coli [78], encodes more diguanylate cyclases and 3 times more MCPs than E. coli. The delta-proteobacterium Bdellovibrio bacteriovorus, a predator that infects an E. coli cell and grows in its periplasmic space, also turned out to have a higher IQ: it has a smaller genome than E. coli but encodes almost twice as many HKs and four times more MCPs.
Phylogenetic distribution of signaling systems
Histidine kinases are by far the predominant type of sensory proteins (Figure 1), whose distribution in all sequenced organisms generally follows the power law (Figures 3a and 4a). The relative abundance of other types of receptors, however, varies widely among organisms of different phylogenetic lineages (Figure 3b–f). Still, the distribution of their total number also follows power law (Figure 4b). These observations will be analyzed in detail elsewhere. The following is just a brief listing of several unexpected trends:
1. Archaea do not encode AC1- or AC3-type adenylate cyclases, diguanylate cyclases or c-di-GMP-specific phosphodiesterases (with the exception of several highly diverged and probably inactive ORFs), but encode a fair amount of STYKs. In 11 of 20 archaeal genomes, STYKs and class 2 ACs are the only recognizable proteins involved in signal transduction. More than a half of all sequenced archaeal genomes do not encode any MCPs, others encode from 2 to 5 and only the two halophilic species have a large number of MCPs (17 each, Figure 3b).
2. Actinobacteria do not encode MCPs or, for that matter, any other chemotaxis or flagellar proteins (the only one that does, Symbiobacterium thermophilum, probably does not belong to the actinobacterial lineage [79]). Instead, actinobacteria encode relatively large numbers of HKs and STYKs (Figure 3a,c). As noted previously, Mycobacterium tuberculosis encodes a relatively high number of AC3s [80], as do two other mycobacteria, M. bovis and M. avium, but not M. leprae (16, 16, 12, and 4, respectively). The regulators of these AC3s remain unknown, although some ACs have been implicated in sensing of the bicarbonate level [81]. The dramatically lower number of signaling proteins in M. leprae, compared to other mycobacteria, is in line with the general picture of genome decay in this organism [82].
3. Cyanobacteria encode large numbers of HKs and STYKs, but very few MCPs (e.g. 134, 52 and 3, respectively, in Nostoc PCC 7120 [see Additional file 1]). These data are consistent with previous observations that cyanobacteria encode just several highly conserved MCPs [37] and regulate their motility using HKs (phytochromes) [83,84] and ACs [27,28].
4. There is great variation between different subdivisions of Proteobacteria with very few common trends. Proteobacteria generally encode few, if any, STYKs, but a large number of MCPs and diguanylate cyclases. The number of ACs is relatively low, except for representatives of the alpha-subdivision. While gamma-proteobacteria typically encode a single AC1 and no more than one AC3, in Pseudomonas aeruginosa this sole AC3 is important for virulence [85].
5. Several bacterial phyla that currently have only a handful of sequenced representatives show highly biased patterns of signal transducer distribution. For example, four sequenced members of the Bacteroidetes (formerly the CFB group) encode a relatively large number of HKs (85 in Bacteroides thetaiotaomicron), but few or no STYKs and no MCPs, ACs or diguanylate cyclases. It would be interesting to see if this trend holds when more genomes of this lineage become available.
Variation in IQ between close relatives
The recent genomic data revealed substantial differences in gene content among different strains that, judging by the level of 16S rRNA identity, belong to the same bacterial species [86,87]. It is therefore not surprising to see dramatic differences in signaling protein content among different species of the same genus. Still, different members of the Bacillus genus show very similar distributions of signaling proteins [see Additional file 1]. In contrast, three sequenced genomes of Clostridium spp. encode dramatically different numbers of MCPs (38 in C. acetobutylicum, 20 in C. tetani and 0 in C. perfringens) and HD-GYP domains (9, 1, and 1, respectively), whereas the content of other signaling proteins is more or less in line with the genome sizes. Accordingly, C. acetobutylicum makes it into the winners list in both MCP and HD-GYP categories (Table 2).
Although not seen in the current data set, domains that are missing in one strain were sometimes found in a different strain of the same species. Thus, although this domain census shows the absence of HD-GYP domains in Yersinia pestis strain CO92 and in Bacillus cereus strain ATCC 14579 [see Additional file 1], this domain is encoded in Y. pestis strain KIM and B. cereus strain ZK. These differences indicate that signaling proteins can be easily acquired and lost, so all observations on the presence or absence of certain signaling system in a certain organism are only as good as the current genome set.
Transmembrane and intracellular sensors: Extroverts and introverts
Analysis of complete microbial genomes revealed complex systems of intracellular monitoring that included PAS- and GAF-containing proteins with a variety of output domains [3]. The fraction of membrane-bound proteins among all signal transduction proteins encoded in each given genome was evaluated here using three different methods for predicting transmembrane (TM) segments, followed by manual analysis of the outputs. The census showed that while the great majority of HKs and MCPs were membrane-bound, as much as one-third of all HKs and one-sixth of all MCPs did not contain a single TM segment (Figure 1, Additional file 1). In contrast, only about a half of all adenylate and diguanylate cyclases and c-di-GMP phosphodiesterases were membrane-bound; a majority of STYKs and HD-GYP domains were soluble (Figure 1).
It must be noted that not every membrane-bound signal transduction protein is necessarily a sensor of the environmental parameters. An obvious example among HKs is the turgor sensor KdpD, where TM segments serve solely as anchors [88]. Aer, the energy-sensing MCP, presents a similar case [89]. Conversely, some cytoplasmic sensors might actually sense extracellular signals, e.g. when the sensing domains are present on separate transmembrane polypeptides, as is the case with CheA, the chemotaxis HK. Furthermore, many cytoplasmic sensors respond to signals that are membrane-permeable, such as light, oxygen, H2O2; NH3, and should not be considered purely external or internal. Keeping in mind all these caveats, the predominance of extracellular or intracellular transducers can be used to distinguish organisms that are concerned primarily with sensing environmental parameters ("extroverts") from those more closely monitoring the intracellular milieu ("introverts").
In obligately parasitic bacteria that encode only a handful of signal transduction proteins, most of these proteins are membrane-bound [see Additional file 1]. However, Figure 5a shows that once the total number of signal transduction proteins goes beyond a dozen, the fraction of them that are membrane-bound stabilizes at about 60%, approximately the same in representatives of all phyla, except for Cyanobacteria and Archaea. In the latter two groups, the fraction of membrane-bound signal transduction proteins is close to 30% and also shows very little variance (Figure 5a). Although, as mentioned above, cyanobacteria encode very few MCPs, this fact alone cannot account for the difference between them and all other bacteria. A comparison of other types of signaling proteins encoded in cyanobacteria and proteobacteria (Figure 5b) shows the prevalence of soluble proteins among cyanobacterial HKs, STYKs and GGDEF domains, compared to the prevalence of TM proteins among the same groups of proteins in proteobacteria. The difference between cyanobacterial and archaeal proteins on one hand and proteins from other lineages is most clearly seen in the comparison of HKs (Figure 5c). Remarkably, actinobacteria and firmicutes turn out to be firm extroverts with relatively few intracellular HKs; some of the latter, however, are known to participate in regulation of sporulation [6]. This schism between cyanobacteria and all other bacteria with completely sequenced genomes is likely to be due to the much more complex organization of the cyanobacterial cell, which contains intracellular membranes harboring the photosynthetic reaction centers. Among other autotrophic prokaryotes, prevalence of intracellular proteins is seen in methanogenic archaea and in the green sulfur bacterium Chlorobium tepidum, although not in the phototrophic alpha-proteobacterium Rhodopseudomonas palustris [see Additional file 1]. Since archaeal heterotrophs also show low amounts of TM signaling proteins, there does not seem to be a direct connection between 'introvertness' and autotrophic metabolism.
Discussion
This paper has grown out of a survey of signal transduction systems in several alpha- and gamma-proteobacteria prepared for a recent review (Table 1 in ref [3]). It turned out that mere 'counting the senses' could help understand bacterial behavior. For example, as discussed earlier, genomes of two alpha-proteobacteria, Caulobacter crescentus and Mesorhizobium loti, encode the same number of HKs but the former one encodes 19 MCPs compared to just one in M. loti [see Additional file 1]. In contrast, M. loti encodes 13 copies of AC3, compared to just two of them in C. crescentus ([3], [see Additional file 1]). Such observations could provide a useful insight into the physiology of many obscure bacteria whose genomes have been sequenced in the last several years or will be sequenced in the near future. I have therefore updated our previous listing of signal transduction proteins encoded in microbial genomes [2] to cover the genomes sequenced in the past five years.
Defining the set of signaling proteins
For the purposes of this study, the set of surveyed signal transduction proteins has been limited to just 7 classes of proteins: histidine kinases, methyl-accepting chemotaxis receptors, Ser/Thr protein kinases, adenylate and diguanylate cyclases, c-di-GMP phosphodiesterases with the EAL domain and predicted phosphodiesterases with the HD-GYP domain [see Additional file 1]. Certainly, this list is far from being complete. In a general sense, any cellular protein that participates in cellular adaptation to the changing environment can be considered part of the signaling machinery. Thus, AraC-type transcription regulator, whose DNA-binding properties are modulated by arabinose binding to its N-terminal domain [52], could also be treated as an intracellular signal transducer. According to a recent study by Ulrich, Koonin, and Zhulin, such 'one-component' signalers comprise a majority of signal transduction systems and were the first to arise in evolution [90]. More sophisticated mechanisms of signal transduction include two-component (HK and RR) signal transduction systems and a variety of other signaling systems that have been described only in the past several years (see [2,3,21,22,39] for reviews).
This census considered only dedicated signaling systems that consist of more than two individual components. Therefore, transcriptional regulators, even those of complex domain architecture, were left out (for a comprehensive survey of helix-turn-helix-type (HTH) transcriptional regulators, see [91]). I have also left out response regulators, which are typically considered together with HKs. One of the reasons for that was the frequent confusion between three classes of response regulators: (i) the single-domain chemotaxis response regulator CheY that transmits the signal through protein-protein interactions; (ii) the DNA-binding response regulators of the CheY-HTH domain architecture, and (iii) the response regulators with CheY-AC, CheY-GGDEF or CheY-GGDEF-EAL domain architectures, which produce secondary messengers, cAMP and c-di-GMP. Here, various proteins containing AC, GGDEF, EAL or HD-GYP domains have been lumped together, just as the chemotaxis signal transduction kinase CheA is typically treated as sensor kinase, despite being just a transmitter in the signaling cascade going from MCPs to the flagellar motor. This approach differed from that of Ulrich et al. [90], who included diguanylate cyclases and c-di-GMP phosphodiesterases (GGDEF and EAL domains, respectively) into the 'one-component' set.
Another important omission in this survey are Ser/Thr protein phosphatases, which can dephosphorylate STYKs, modulating their activity, and should also be able to dephosphorylate the cellular targets of STYKs. However, several surveys of these enzymes have been published recently [19,39,92], and more are apparently on the way. Due to the difficulties in separating true protein phosphatases from phosphatases of other specificities that often produce false-positive hits I have chosen to exclude them from this survey. Several other systems of the bacterial signal transduction machinery have also been left out. These include (i) Ser/Thr kinases of the bacterial (GHKL) type that regulate the activity of the RNA polymerase sigma subunit; (ii) HPrSer kinase/phosphorylase and other components of the bacterial PEP-dependent phosphotransferase systems, which regulate chemotaxis, membrane transport (inducer exclusion), and catabolite repression; (iii) the systems that regulate RNA and protein degradation; and many others. A census of each of these systems could be an interesting project in its own right.
The limited scope of this survey, which did not include the sophisticated sporulation machinery of the firmicutes and certain unique (potentially signaling) archaeal domains, could be a reason why representatives of these two groups have generally scored low in the IQ category. Including those proteins into a future version of this census might partly correct that bias, although that would increase the degree of 'introvertness' among archaea even further.
Caveats of automated domain counting
Even within the limited scope of this survey, there is a lot of space for controversy. There are no clear criteria to decide which proteins should be considered HKs or STYKs and which should be not. Thus, the discrepancies of the results presented here and in the papers by the Mizuno group [5,16,34] can all be attributed to their more conservative approach to defining HKs. The survey by Kim and Forst [17] shows a similar undercount of non-canonical HKs. In contrast, counting STYKs in the KinG database [64] used more permissive criteria than those employed here, which resulted in KDO kinases and other related kinases being counted as STYKs. For other signaling domains, there was much less room for disagreement. The counts of MCPs and ACs, presented here, are very similar to those reported, respectively, by Zhulin [40] and Shenoy and Visweswariah [41]. All our data with supporting information are available on a public web site [51], which should provide an easy way to analyze any discrepancies and, if necessary, correct the final count.
Do numbers really matter?
It is well known that growth in bacterial genome size is accompanied by accumulation of paralogous protein families, which can be easily seen in lineage-specific expansions of transcriptional regulators, metabolic enzymes, and/or surface proteins [93-95]. It can be argued therefore that the sheer number of signal transduction proteins encoded in a bacterial genome is hardly a good measure of its IQ, as many of these proteins are closely related paralogs. It would seem, however, that lineage-specific expansions that have been fixed in evolution must be of some value to the host organism. Among metabolic enzymes, there are indications of functional diversification even among close paralogs [96]. As for signaling proteins, Valley Stewart and colleagues have shown that NarQ and NarX, two paralogous HKs in E. coli, have similar but non-identical functions in modulating cellular response to nitrate and nitrite [97,98]. Likewise, out of 12 GGDEF domain-containing proteins – potential diguanylate cyclases – encoded in Salmonella Typhimurium genome, one, AdrA, was found to be primarily responsible for regulating biofilm formation in a complex medium, whereas another, STM1987, was critical for biofilm formation in the nutrient-poor medium [99,100]. These data show that we should be very careful in assigning the same function even to closely related paralogs. Differential regulation of expression and activity of paralogous signal transduction proteins could be yet another sophisticated mechanism allowing the bacterial cell to fine-tune its response to environmental changes. Therefore, until there is clear evidence that functions of paralogous signal transduction proteins are truly identical, the total number of such proteins remains the best measure of the bacterial IQ.
Intracellular signaling
One of the most significant insights to emerge from comparative genome analysis was the recognition of the vast system of intracellular signaling in bacteria. It became clear that many bacteria encode complex systems of intracellular monitoring whose domain organization is very similar to that used in transmembrane signaling: a sensor domain (typically, PAS and/or GAF), followed by HK, AC, GGDEF or EAL output domains [3]. In certain cases, soluble HKs, MCPs, and ACs have been experimentally characterized and shown to be involved in monitoring levels of intracellular ATP, oxygen, CO, bicarbonate, nitrate, reactive nitrogen species, and other metabolites and modulating the cellular response to the changes in these parameters [101-105]. Some intracellular sensors appeared to be specifically geared towards unusual substrates used by the particular bacterium, such as methanol and formaldehyde in Paracoccus denitrificans and Methylobacterium organophilum [106,107]. In the recently sequenced genome of Dehalococcoides ethenogenes, a major detoxifier of chlorinated organic pollutants, many soluble HKs were found encoded in close proximity to the genes for reductive halogenases, the enzymes that catalyze the dechlorination reactions [108]. It was proposed that these HKs respond to intracellular rather than extracellular stimuli, stimulating the expression of reductive halogenases in response to the presence of their chlorinated substrates [108].
This census shows that intracellular signal transduction proteins comprise a significant fraction of all signal transducers encoded in almost any bacterial genome. However, most of them are still uncharacterized and have yet to be recognized as legitimate members of the bacterial signaling network. The finding that these proteins are abundant in many pathogenic as well as free-living bacteria should help focus the attention of the research community on these novel components of the signal transduction network.
The predominance of intracellular signal transduction proteins in cyanobacteria is in stark contrast with the far smaller proportion of such proteins in other bacterial lineages. There could be several possible reasons for this 'introvertness', all linked to the ability of cyanobacteria to conduct oxygenic photosynthesis. Firstly, cyanobacteria harbor a complex system of intracellular membranes carrying the photosynthetic reaction centers. Intracellular signaling proteins could be needed to control formation and functioning of the photosynthetic system, as well as the transition from phototrophic to heterotrophic metabolism and back. The compartmentalization of the cellular interior probably requires a sophisticated system of monitoring conditions within the individual compartments. Last but not the least, cyanobacteria are unique among (known) prokaryotes in that their cells generate oxygen, which other bacteria try to keep outside the cell. The presence of oxygen affects the redox balance in the cytoplasm and leads to oxidative damage of numerous cellular compounds, including ATP, methionine, cysteine, and many others. It is very likely that numerous intracellular HKs that contain PAS domains are involved in maintaining the constant level of the redox potential in the cyanobacterial cell. Surprisingly, Rhodopseudomonas palustris, an alpha-proteobacterium that is also capable of transition between autotrophic and heterotrophic metabolism, does not appear to be an 'introvert' [see Additional file 1]. Hence, it seems that the trend of autotrophic bacteria and archaea being more of 'introverts' and heterotrophic bacteria being more of 'extroverts' might be biased by the current selection of the completely sequenced genomes. It would be interesting to see whether this trend holds when more genomes of bacterial photo- and chemolitotrophs become available.
Phylum-specific bias and evolution of signal transduction
The knowledge of the phylogenetic distribution of signal transduction systems allows a better understanding of their evolution. Previous analysis of HKs and RRs by Koretke and colleagues led to the conclusion that two-component systems originated in bacteria and radiated into two other domains of life through multiple events of horizontal gene transfer [32]. HKs and STYKs appear to be the principal signal transduction proteins in archaea, suggesting that these two classes of proteins could be already present in the last common ancestor of all living organisms (LUCA, [92,109,110]). The absence of AC3-type adenylate cyclases, diguanylate cyclases and c-di-GMP phosphodiesterases in any of the sequenced archaeal genomes is quite remarkable. In fact, the only full-size archaeal AC3 domain known to date has been found in an uncultivated psychrophilic crenarchaeote that exhibited numerous cases of horizontal gene transfer [111]. Most archaea, however, encode ACs of class 2 (COG1437 []), which are found in only a handful of organisms outside Archaea [30]. These data show that although cAMP is a truly universal second messenger, different domains of life utilize different enzymes for its production and probably employ entirely different mechanisms of cAMP-dependent signaling.
Another remarkable example is the diversity of outputs of the chemotaxis machinery. Although all MCPs counted in this work are very similar, it has been noted [112] that chemotactic signals in diverse bacteria and archaea are being transduced to at least three different motility apparata: the bacterial flagellum, the archaeal flagellum that is unrelated to the bacterial one [113], and to the type IV pili, which are responsible for gliding motility of cyanobacteria and certain other bacteria [84,114].
In general, variability of signal transduction protein content in closely related bacteria, uneven distribution of these proteins among well-established phylogenetic lineages, and the presence in many genomes of tight clusters of closely related paralogs indicate that signaling proteins can be easily acquired and lost. Lineage-specific gene duplication and gene loss and lateral gene transfer probably play a key role in shaping the signaling protein repertoire of each given organism. Why, then, would the total number of signal transduction proteins grow as a square of the genome size (Figure 2a) across a wide variety of microorganisms with diverse lifestyles, phylogenetic affinities, and metabolic capabilities? It is tempting to suggest that there must be an underlying mechanism supporting this correlation. For example, the power-law distribution of HKs (Figure 4a) might stem from the simple fact that the number of binary interactions grows as a square of the number of interacting components [115], so that the number of sensory proteins that manage the linearly growing number of metabolic enzymes has to grow as a square of that number. This explanation is somewhat similar to the one offered by van Nimwegen to explain his observation that the number of transcriptional regulators in bacteria also grows as a square of the genome size [71], although his analysis did not include two-component systems. This was also the rationale behind the decision to measure bacterial IQ as a square root, rather than a linear function, of the total number of encoded signal transduction proteins (see the Results section). However, HKs comprise only but a half of all signal transduction proteins counted in this work [see Additional file 1]. The distribution of other types of signal transducers is even more fascinating: while distribution of each individual class of proteins seems almost random (Figures 3b–f), their total number still grows approximately as a square of genome size (Figure 4b). One could speculate that this quadratic dependence determines a near-optimal number of signal transducers at a given genome size. This would mean that during their adaptation to different ecological niches, bacteria evolve to rely primarily on certain types of signal transduction, while other types of transducers can be lost (or not fixed in the genome when acquired by lateral gene transfer). For example, during the reductive evolution of chlamydia, HKs and STYKs were retained, while all other transducers and were lost [see Additional file 1]. In contrast, spirochetes held on to their chemotaxis transducers but mostly lost their STYKs. The recent evidence for non-canonical roles of signal transduction proteins, e.g. regulation of gene expression by the chemotaxis system [116] and regulation of chemotaxis by adenylate cyclases [28], suggests that there is certain flexibility in functions of different transducers that could be used by bacterial evolution to generate even greater diversity of signal transduction mechanisms.
Future developments
The goal of genome analysis is to predict the organism's physiology and behavior based solely on the genomic sequence. There has been great progress in predicting metabolic pathways [110,117,118]; deciphering signaling pathways so far has lagged behind. Accumulation of complete genome sequences has led to the delineation of many new signaling and signal transduction domains and caused a revolution on our understanding of bacterial regulatory networks [2,3,20,119].
I believe that, despite all its limitations, this census would be useful for microbiologists, at least by highlighting still unresolved problems in prokaryotic signal transduction. This work should be complemented by surveys of other components of the signal transduction machinery, including various response regulators, Ser/Thr protein phosphatases, PTS proteins, and many others. Genomes of several environmental microorganisms, including 9-Mb genomes of Myxococcus xanthus, Rhodococcus sp., and Gemmata obscuriglobus, have been completed and are expected to be publicly released in the near future. Owing to their sheer size, these genomes are likely to bring new signaling domains and illuminate even more regulatory relations. Myxococcus xanthus, which reportedly encodes close to 200 HKs and many STYKs, would probably become a leader in both these categories.
The example of M. xanthus exposes certain flaws in the IQ calculation method used in this work. This bacterium has extremely complex behavioral patterns [114], but, at 9.1 Mb, it would need to encode more than 550 signal transduction proteins just to make it into the winners' list (Table 1). Certainly, better ways to evaluate bacterial IQ are needed, but that should be subject of a future work. Still, I believe that in the era of 'systems biology' when cellular metabolic pathways are being routinely modeled on a whole-genome level [50,120] and the cell itself is treated more as a machine with a number of interacting parts [121,122], it is important to keep in mind the real complexity of the signal network encoded in each given prokaryotic genome and have an easy measure of this complexity.
I also hope that this census will help us get a better understanding of the microbial diversity and the unique ways that bacteria use to adapt to changing environment. Such understanding is becoming increasingly important as our earlier methods of controlling bacterial growth with one-size-fits-all wide-spectrum antibiotics show progressively diminishing results.
Conclusion
Careful accounting of diverse proteins participating in prokaryotic signal transduction shows that the complexity of signaling mechanisms correlates well with the organism's genome size and the size of its proteome. The total number of proteins involved in signal transduction, the number of histidine kinases, and the total number of signal transduction proteins other than histidine kinases all grow as square of the genome size. At the same time, the fractions of the latter proteins – MCPs, STYKs, adenylate and diguanylate cyclases and phosphodiesterases – in the total set vary widely depending on the organism's ecology, metabolic properties, and phylogenetic position. The results of this census are freely available to the public and will be updated and corrected as necessary. The availability of this resource, as well as introduction of the concepts of bacterial IQ, introverts and extroverts among the prokaryotes, should help in achieving a better understanding of the microbial behavior and forces that shape microbial genome evolution.
Methods
Data sources
Complete genome sequences of 167 bacterial and archaeal species, sequenced by the end of 2004, were downloaded from the NCBI's Genomes database [123] or searched directly through the NCBI web site. Only one representative genome per species was used, usually the first one to be publicly released, according to the NCBI Genomes database listing. Exceptions were made for Escherichia coli, represented by two strains, K12 [GenBank:U00096] and O157:H7 [GenBank:BA000007], and Salmonella enterica, represented by three serovars, Paratyphi [GenBank:CP000026], Typhi [GenBank:AL513382], and Typhimurium [GenBank:AE006468]. For Prochlorococcus marinus, strain CCMP1375 [GenBank:AE017126] genome was used, the middle-sized one of the three. Among other simultaneously released genomes, Staphylococcus aureus N315 [GenBank:BA000018], Streptococcus thermophilus CNRZ1066 [GenBank:CP000024], and Thermus thermophilus HB27 [GenBank:AE017221] genomes were used.
A census of histidine kinases
The complete list of histidine kinases was compiled separately for each particular phylum of bacteria from the results of BLAST searches against selected genomes using the NCBI's Genomic BLAST tool [124], followed by iterative PSI-BLAST searches [59]. Typically, the searches used as the query sequence the C-terminal fragment (residues 301–579) of the well-characterized histidine kinase PhoR [UniProt:P23545 ] from Bacillus subtilis, which contains both HisKA and HATPase domains [125], and a position-specific scoring matrix (PSSM) derived from an alignment of well-characterized histidine kinases (both available as Supplementary Material). Additional searches against the NCBI's Reference Sequence (RefSeq) database [126,127] were performed through the NCBI BLAST web interface by limiting the search space to the given phylum (e.g. Actinobacteria [orgn]) and excluding reference sequences of incomplete genomes (srcdb_refseq [prop] NOT srcdb_refseq_model [prop]). The PSI-BLAST searches used strict inclusion threshold expect values of 10-5–10-7 (adjusting as necessary) and were iterated until no newly retrieved sequences belonged to HKs. The total numbers of copies of each signaling domain encoded in each given genome were estimated using the "Taxonomy Report" option in the BLAST output. Potential false-positive hits were checked at every step of PSI-BLAST using the CDD Domain viewer and manually removed (unselected) from the hit list for the next iteration of PSI-BLAST. In each case where the HATPase domain was easily recognized but HisKA domain was not, a BLAST2sequences [128] search was performed to check whether the HATPase domain was preceded by a conserved region carrying a conserved His residue. The presence of such His-containing regions would indicate that those questionable proteins (e.g., mlr1749 [UniProt:Q98JW4 ] and other members of COG3920 []) comprise legitimate HKs, contrary to the view of Hagiwara et al. [34].
Alternatively, PSI-BLAST searches were run against a local copy of the RefSeq database, using the same query sequence and search parameters with additional filtering against sequences translated from unfinished genomes (ZP_xxxxxxxx entries). The resulting hits were compared against the NCBI Taxonomy database to ensure that they all came from a single organism (only one genome of each bacterial species, usually the first one to be sequenced, was used in this analysis). Similar protocol was used to search for histidine kinases in other bacterial phyla.
Counting other signaling domains
Owing to the relatively high sequence conservation of the MCP, ACyc, GGDEF, and EAL domains, manual checking of the PSI-BLAST outputs revealed very few false-positive hits. In the case of the two latter domains, many low-scoring proteins had numerous amino acid changes, including ones in the likely active sites (see [2,22,67]). No attempt has been made to sort these domains into active and inactive ones. For the HD-GYP domain, which comprises a typical HD superfamily phosphoesterase domain with a number of additional conserved residues, high-scoring BLAST hits to the standard HD domains were filtered based on the shorter length of those hits.
Identification of transmembrane receptors
Transmembrane (TM) segments in verified sets of signal transduction proteins from various phylogenetic lineages were predicted using PHDhtm [129] and TMHMM [130] programs. The results were sorted into three bins: TM proteins (≥ 2 TM segments), 1 TM proteins, and soluble proteins, and the discrepancies between predictions of the two programs were manually inspected. Comparison of the results revealed many false-negative assignments, so that prediction of a TM segment by either program typically turned out to be justified. Questionable cases were also checked using the HMMTop [131] program, which, however, produced both false-negative and false-positive predictions of TM segments. Therefore, HMMTop assignments were considered only when supported by either PHDhtm or TMHMM results.
List of Abbreviations
AC, adenylate cyclase;
AC1, adenylate cyclase class 1;
AC2, adenylate cyclase class 2;
AC3, adenylate cyclase class 3;
c-di-GMP, cyclic dimeric (3',5'-guanosine monophosphate);
EAL, conserved protein domain with the Glu-Ala-Leu sequence motif and c-di-GMP-specific phosphodiesterase activity;
GGDEF, conserved protein domain with the Gly-Gly-(Asp/Glu)-Glu-Phe sequence motif and diguanylate cyclase activity;
HD-GYP, conserved protein domain of the HD phosphohydrolase superfamily with additional highly conserved residues, predicted phosphodiesterase;
HK or HisK, histidine kinase;
MCP, methyl-accepting chemotaxis protein;
STYK, Ser/Thr/Tyr-specific protein kinase
TM, transmembrane.
Authors' contributions
MYG conceived the study, performed all the calculations and wrote the manuscript.
Supplementary Material
Additional File 1
Results of the census of membrane-bound and intracellular signal transduction proteins in bacteria in HTML format
Click here for file
Additional File 2
Table 1 in HTML format
Click here for file
Additional File 3
Table 1 in HTML format
Click here for file
Acknowledgements
I thank Yuri Wolf and Darren Natale for valuable advice, Mark Gomelsky, Eugene Koonin, Armen Mulkidjanian, and Igor Zhulin for helpful comments, and many other colleagues for suggestions.
Figures and Tables
Figure 1 Growth in the number of signal transduction proteins encoded in complete microbial genomes. A. Histidine kinases (circles), methyl-accepting chemotaxis proteins (MCPs, squares) and Ser/Thr/Tyr protein kinases (STYKs, triangles). B. Diguanylate cyclases (GGDEF domains, diamonds), c-di-GMP-specific phosphodiesterases (EAL domains, triangles), and adenylate cyclases (circles). Open symbols indicate the total number of proteins with the corresponding domains, closed symbols – the number of membrane-bound proteins of each kind.
Figure 2 The total number of encoded signal transducers proteins grows with genome size. A. Distribution of signal transduction proteins among free-living bacteria and archaea (squares) and obligate pathogens (closed squares). Organisms with other status (symbionts and other commensals) are indicated by triangles. The best-fit line represents data from all species. B. Distribution of signal transduction proteins among organisms of different phylogenetic lineages. The symbols indicate members of the following phyla: Actinobacteria, black circles; Cyanobacteria, open circles (light blue); Alpha-proteobacteria, closed diamonds (dark brown); Beta-, Delta-, and Epsilon-proteobacteria, open diamonds (yellow); Gamma-proteobacteria, closed squares (dark blue); Firmicutes, open squares (magenta); members of other bacterial phyla (Aquificales, Bacteroidetes, Chlamydiae, Deinococcus-Thermus, Planctomycetes, Spirochetes, Thermotoga), closed triangles (red); Archaea, open triangles (yellow).
Figure 3 Phylogenetic distribution of certain types of signal transducers. A. Histidine kinases. B. Methyl-accepting chemotaxis proteins. C. Ser/Thr/Tyr-protein kinases. D. GGDEF domains (active and inactive diguanylate cyclases). E. EAL domains (active and inactive c-di-GMP phosphodiesterases). F. Adenylate cyclases. The symbols for various phyla are shown at the bottom and are the same as in Fig. 2b.
Figure 4 Distribution of the signal transduction proteins follows the power law. A. Distribution of histidine kinases. B. Distribution of the total number of all signal transduction proteins except for histidine kinases encoded in a given genome. The symbols for various phyla are as in Fig. 2b.
Figure 5 Phylogenetic distribution of membrane-bound signal transduction proteins. A. Phylogenetic distribution of the total number of transmembrane signal transducers. The best-fit lines are shown for proteins from gamma-bacteria (dark blue) and cyanobacteria (cyan). The symbols for various phyla are as in Fig. 2b. B. Transmembrane histidine kinases (squares), Ser/Thr kinases (circles) and diguanylate cyclases (triangles) in all proteobacteria (open symbols) and cyanobacteria (closed symbols). The best-fit lines are shown for proteobacterial (dark blue) and cyanobacterial (cyan) histidine kinases. C. Phylogenetic distribution of transmembrane histidine kinases. The best-fit lines are shown for actinobacterial (black) and cyanobacterial (cyan) histidine kinases.
Table 1 Bacteria with the highest adaptability index ("highest IQ")
Organism Phyluma Signal transducers Genome size, kb IQ
Wolinella succinogenes Epsilon 99 2,110 230
Geobacter sulfurreducens Delta 165 3,814 166
Idiomarina loihiensis Gamma 80 2,839 153
Desulfovibrio vulgaris Delta 135 3,773 151
Vibrio cholerae Gamma 152 4,033 150
Thermotoga maritima Other 34 1,861 145
Borrelia garinii Spiro 13 987 143
Vibrio vulnificus Gamma 200 5,127 136
Chromobacterium violaceum Beta 160 4,751 131
Thermosynechococcus elongatus Cyano 51 2,594 131
a – Beta, Gamma, Delta and Epsilon indicate the corresponding subdivisions of Proteobacteria; Cyano indicates cyanobacteria; Spiro indicates Spirochetes.
Table 2 Bacteria and archaea with the highest proportion of encoded signaling proteins of each type
Organism Phylum No. proteins (%total)
Histidine kinases
Nostoc sp. PCC 7120 Cyano 134 (2.2%)
Geobacter sulfurreducens Delta 92 (2.7%)
Bacteroides thetaiotaomicron Other 85 (1.8%)
Rhodopseudomonas palustris Alpha 66 (1.4%)
Desulfovibrio vulgaris Delta 64 (1.8%)
Wolinella succinogenes Epsilon 39 (1.9%)
Haloarcula marismortui Archaea 59 (1.4%)
MCPs
Vibrio vulnificus Gamma 52 (1.2%)
Pseudomonas syringae Gamma 48 (0.9%)
Vibrio cholerae Gamma 45 (1.2%)
Chromobacterium violaceum Beta 42 (1.0%)
Clostridium acetobutylicum Firmicutes 38 (1.0%)
Wolinella succinogenes Epsilon 31 (1.5%)
Halobacterium salinarium Archaea 17 (0.6%)
Ser/Thr protein kinases
Rhodopirellula baltica Other 60 (0.8%)
Nostoc sp. PCC 7120 Cyano 52 (0.9%)
Streptomyces coelicolor Actino 37 (0.4%)
Streptomyces avermitilis Actino 35 (0.4%)
Gloeobacter violaceus Cyano 20 (0.4%)
.Thermosynechococcus elongatus Cyano 17 (0.7%)
Sulfolobus tokodaii Archaea 12 (0.4%)
Diguanylate cyclases
Vibrio vulnificus Gamma 66 (1.5%)
Shewanella oneidensis Gamma 52 (1.2%)
Vibrio parahaemolyticus Gamma 44 (0.9%)
Chromobacterium violaceum Beta 43 (1.0%)
Vibrio cholerae Gamma 41 (1.1%)
Idiomarina loihiensis Gamma 33 (1.3%)
Adenylate cyclases
Bradyrhizobium japonicum Alpha 37 (0.4%)
Sinorhizobium meliloti Alpha 28 (0.5%)
Leptospira interrogans Spiro 18 (0.4%)
Mycobacterium bovis Actino 16 (0.4%)
Mycobacterium tuberculosis Actino 16 (0.4%)
Treponema denticola Spiro 9 (0.3%)
HD-GYP domains
Desulfovibrio vulgaris Delta 14 (0.4%)
Vibrio vulnificus Gamma 13 (0.3%)
Chromobacterium violaceum Beta 11 (0.2%)
Thermotoga maritima Other 10 (0.5%)
Desulfotalea psychrophila Delta 10 (0.3%)
Geobacter sulfurreducens Delta 10 (0.3%)
a – Beta, Gamma, Delta and Epsilon indicate the corresponding subdivisions of Proteobacteria; Cyano indicates cyanobacteria; Spiro indicates Spirochetes.
==== Refs
Sondej M Weinglass AB Peterkofsky A Kaback HR Binding of enzyme IIAGlc, a component of the phosphoenolpyruvate:sugar phosphotransferase system, to the Escherichia coli lactose permease Biochemistry 2002 41 5556 5565 11969416
Galperin MY Nikolskaya AN Koonin EV Novel domains of the prokaryotic two-component signal transduction systems FEMS Microbiol Lett 2001 203 11 21 11557134
Galperin MY Bacterial signal transduction network in a genomic perspective Environ Microbiol 2004 6 552 567 15142243
Konstantinidis KT Tiedje JM Trends between gene content and genome size in prokaryotic species with larger genomes Proc Natl Acad Sci U S A 2004 101 3160 3165 14973198
Mizuno T Kaneko T Tabata S Compilation of all genes encoding bacterial two-component signal transducers in the genome of the cyanobacterium, Synechocystis sp. strain PCC 6803 DNA Res 1996 3 407 414 9097043
Fabret C Feher VA Hoch JA Two-component signal transduction in Bacillus subtilis: how one organism sees its world J Bacteriol 1999 181 1975 1983 10094672
Wolanin PM Thomason PA Stock JB Histidine protein kinases: key signal transducers outside the animal kingdom Genome Biol 2002 3 REVIEWS3013. 12372152
Grebe TW Stock JB The histidine protein kinase superfamily Adv Microb Physiol 1999 41 139 227 10500846
Classification of Histididine Protein Kinases and Response Regulators
Karniol B Vierstra RD The HWE histidine kinases, a new family of bacterial two-component sensor kinases with potentially diverse roles in environmental signaling J Bacteriol 2004 186 445 453 14702314
Letunic I Copley RR Schmidt S Ciccarelli FD Doerks T Schultz J Ponting CP Bork P SMART 4.0: towards genomic data integration Nucleic Acids Res 2004 32 D142 D144. 14681379
Simple Modular Architecture Research Tool
Bateman A Coin L Durbin R Finn RD Hollich V Griffiths-Jones S Khanna A Marshall M Moxon S Sonnhammer EL Studholme DJ Yeats C Eddy SR The Pfam protein families database Nucleic Acids Res 2004 32 D138 D141. 14681378
Pfam - Protein families database of alignments and HMMs
Zhulin IB Digging with experimental pick and computational shovel: a new addition to the histidine kinase superfamily J Bacteriol 2004 186 267 269 14702293
Mizuno T Compilation of all genes encoding two-component phosphotransfer signal transducers in the genome of Escherichia coli DNA Res 1997 4 161 168 9205844
Kim D Forst S Genomic analysis of the histidine kinase family in bacteria and archaea Microbiology 2001 147 1197 1212 11320123
Mizuno T His-Asp phosphotransfer signal transduction J Biochem (Tokyo) 1998 123 555 563 9538242
Shi L Potts M Kennelly PJ The serine, threonine, and/or tyrosine-specific protein kinases and protein phosphatases of prokaryotic organisms: a family portrait FEMS Microbiol Rev 1998 22 229 253 9862122
Kennelly PJ Protein kinases and protein phosphatases in prokaryotes: a genomic perspective FEMS Microbiol Lett 2002 206 1 8 11786249
Jenal U Cyclic di-guanosine-monophosphate comes of age: a novel secondary messenger involved in modulating cell surface structures in bacteria? Curr Opin Microbiol 2004 7 185 191 15063857
Römling U Gomelsky M Galperin MY C-di-GMP: The dawning of a novel bacterial signalling system Mol Microbiol 2005
Paul R Weiser S Amiot NC Chan C Schirmer T Giese B Jenal U Cell cycle-dependent dynamic localization of a bacterial response regulator with a novel di-guanylate cyclase output domain Genes Dev 2004 18 715 727 15075296
Ryjenkov DA Tarutina M Moskvin OM Gomelsky M Cyclic diguanylate is a ubiquitous signaling molecule in Bacteria: Insights into biochemistry of the GGDEF protein domain. J Bacteriol 2005 187 1792 1798 15716451
Bobrov AG Kirillina O Perry RD The phosphodiesterase activity of the HmsP EAL domain is required for negative regulation of biofilm formation in Yersinia pestis FEMS Microbiol Lett 2005 247 123 130 15935569
Galperin MY Natale DA Aravind L Koonin EV A specialized version of the HD hydrolase domain implicated in signal transduction J Mol Microbiol Biotechnol 1999 1 303 305 10943560
Ohmori M Okamoto S Photoresponsive cAMP signal transduction in cyanobacteria Photochem Photobiol Sci 2004 3 503 511 15170478
Terauchi K Ohmori M Blue light stimulates cyanobacterial motility via a cAMP signal transduction system Mol Microbiol 2004 52 303 309 15049828
Barzu O Danchin A Adenylyl cyclases: a heterogeneous class of ATP-utilizing enzymes Prog Nucleic Acid Res Mol Biol 1994 49 241 283 7863008
Sismeiro O Trotot P Biville F Vivares C Danchin A Aeromonas hydrophila adenylyl cyclase 2: a new class of adenylyl cyclases with thermophilic properties and sequence similarities to proteins from hyperthermophilic archaebacteria J Bacteriol 1998 180 3339 3344 9642185
Novakova L Saskova L Pallova P Janecek J Novotna J Ulrych A Echenique J Trombe MC Branny P Characterization of a eukaryotic type serine/threonine protein kinase and protein phosphatase of Streptococcus pneumoniae and identification of kinase substrates FEBS J 2005 272 1243 1254 15720398
Koretke KK Lupas AN Warren PV Rosenberg M Brown JR Evolution of two-component signal transduction Mol Biol Evol 2000 17 1956 1970 11110912
Ashby MK Survey of the number of two-component response regulator genes in the complete and annotated genome sequences of prokaryotes FEMS Microbiol Lett 2004 231 277 281 14987775
Hagiwara D Yamashino T Mizuno T Genome-wide comparison of the His-to-Asp phosphorelay signaling components of three symbiotic genera of Rhizobia DNA Res 2004 11 57 65 15141946
Ohmori M Ikeuchi M Sato N Wolk P Kaneko T Ogawa T Kanehisa M Goto S Kawashima S Okamoto S Yoshimura H Katoh H Fujisawa T Ehira S Kamei A Yoshihara S Narikawa R Tabata S Characterization of genes encoding multi-domain proteins in the genome of the filamentous nitrogen-fixing cyanobacterium Anabaena sp. strain PCC 7120 DNA Res 2001 8 271 284 11858227
Wang L Sun YP Chen WL Li JH Zhang CC Genomic analysis of protein kinases, protein phosphatases and two-component regulatory systems of the cyanobacterium Anabaena sp. strain PCC 7120 FEMS Microbiol Lett 2002 217 155 165 12480098
Wuichet K Zhulin IB Molecular evolution of sensory domains in cyanobacterial chemoreceptors Trends Microbiol 2003 11 200 203 12781518
Shi L Zhang W Comparative analysis of eukaryotic-type protein phosphatases in two streptomycete genomes Microbiology 2004 150 2247 2256 15256567
Shi L Manganese-dependent protein O-phosphatases in prokaryotes and their biological functions Front Biosci 2004 9 1382 1397 14977554
Zhulin IB The superfamily of chemotaxis transducers: from physiology to genomics and back Adv Microb Physiol 2001 45 157 198 11450109
Shenoy AR Visweswariah SS Class III nucleotide cyclases in bacteria and archaebacteria: lineage-spefic expansion of adenylyl cyclases and a dearth of guanylyl cyclases FEBS Lett 2004 561 11 21 15043055
Kanehisa M Goto S Kawashima S Okuno Y Hattori M The KEGG resource for deciphering the genome Nucleic Acids Res 2004 32 D277 D280 14681412
KEGG: Kyoto Encyclopedia of Genes and Genomes. Two-component system - Escherichia coli K-12 MG1655
Maltsev N Marland E Yu GX Bhatnagar S Lusk R Sentra, a database of signal transduction proteins Nucleic Acids Res 2002 30 349 350 11752334
SENTRA, A Database of Signal Transduction Proteins for Prokaryotes
Tatusov RL Fedorova ND Jackson JD Jacobs AR Kiryutin B Koonin EV Krylov DM Mazumder R Mekhedov SL Nikolskaya AN Rao BS Smirnov S Sverdlov AV Vasudevan S Wolf YI Yin JJ Natale DA The COG database: an updated version includes eukaryotes BMC Bioinformatics 2003 4 41 12969510
Clusters of Orthologous Groups of proteins database
Riley ML Schmidt T Wagner C Mewes HW Frishman D The PEDANT genome database in 2005 Nucleic Acids Res 2005 33 D308 D310 15608204
PEDANT - Protein Extraction, Description and ANalysis Tool
Hellingwerf KJ Bacterial observations: a rudimentary form of intelligence? Trends Microbiol 2005 13 152 158 15817384
Census of bacterial signal transduction proteins
Schleif R AraC protein: a love-hate relationship Bioessays 2003 25 274 282 12596232
Mukhopadhyay P Zheng M Bedzyk LA LaRossa RA Storz G Prominent roles of the NorR and Fur regulators in the Escherichia coli transcriptional response to reactive nitrogen species Proc Natl Acad Sci USA 2004 101 745 750 14718666
Blattner FR Plunkett G Bloch CA Perna NT Burland V Riley M Collado-Vides J Glasner JD Rode CK Mayhew GF Gregor J Davis NW Kirkpatrick HA Goeden MA Rose DJ Mau B Shao Y The complete genome sequence of Escherichia coli K-12 Science 1997 277 1453 1474 9278503
Hayashi T Makino K Ohnishi M Kurokawa K Ishii K Yokoyama K Han CG Ohtsubo E Nakayama K Murata T Tanaka M Tobe T Iida T Takami H Honda T Sasakawa C Ogasawara N Yasunaga T Kuhara S Shiba T Hattori M Shinagawa H Complete genome sequence of enterohemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain K-12 DNA Res 2001 8 11 22 11258796
Parkhill J Dougan G James KD Thomson NR Pickard D Wain J Churcher C Mungall KL Bentley SD Holden MT Sebaihia M Baker S Basham D Brooks K Chillingworth T Connerton P Cronin A Davis P Davies RM Dowd L White N Farrar J Feltwell T Hamlin N Haque A Hien TT Holroyd S Jagels K Krogh A Larsen TS Leather S Moule S O'Gaora P Parry C Quail M Rutherford K Simmonds M Skelton J Stevens K Whitehead S Barrell BG Complete genome sequence of a multiple drug resistant Salmonella enterica serovar Typhi CT18 Nature 2001 413 848 852 11677608
McClelland M Sanderson KE Spieth J Clifton SW Latreille P Courtney L Porwollik S Ali J Dante M Du F Hou S Layman D Leonard S Nguyen C Scott K Holmes A Grewal N Mulvaney E Ryan E Sun H Florea L Miller W Stoneking T Nhan M Waterston R Wilson RK Complete genome sequence of Salmonella enterica serovar Typhimurium LT2 Nature 2001 413 852 856 11677609
McClelland M Sanderson KE Clifton SW Latreille P Porwollik S Sabo A Meyer R Bieri T Ozersky P McLellan M Harkins CR Wang C Nguyen C Berghoff A Elliott G Kohlberg S Strong C Du F Carter J Kremizki C Layman D Leonard S Sun H Fulton L Nash W Miner T Minx P Delehaunty K Fronick C Magrini V Nhan M Warren W Florea L Spieth J Wilson RK Comparison of genome degradation in Paratyphi A and Typhi, human-restricted serovars of Salmonella enterica that cause typhoid Nat Genet 2004 36 1268 1274 15531882
Altschul SF Madden TL Schaffer AA Zhang J Zheng 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
Marchler-Bauer A Bryant SH CD-Search: protein domain annotations on the fly Nucleic Acids Res 2004 32 W327 W331 15215404
Ban C Yang W Crystal structure and ATPase activity of MutL: implications for DNA repair and mutagenesis Cell 1998 95 541 552 9827806
Dutta R Inouye M GHKL, an emergent ATPase/kinase superfamily Trends Biochem Sci 2000 25 24 28 10637609
Stock AM Robinson VL Goudreau PN Two-component signal transduction Annu Rev Biochem 2000 69 183 215 10966457
Krupa A Abhinandan KR Srinivasan N KinG: a database of protein kinases in genomes Nucleic Acids Res 2004 32 D153 D155 14681382
KinG database
Poon WW Davis DE Ha HT Jonassen T Rather PN Clarke CF Identification of Escherichia coli ubiB, a gene required for the first monooxygenase step in ubiquinone biosynthesis J Bacteriol 2000 182 5139 5146 10960098
Chan C Paul R Samoray D Amiot NC Giese B Jenal U Schirmer T Structural basis of activity and allosteric control of diguanylate cyclase Proc Natl Acad Sci U S A 2004 101 17084 17089 15569936
Aravind L Koonin EV A novel family of predicted phosphoesterases includes Drosophila prune protein and bacterial RecJ exonuclease Trends Biochem Sci 1998 23 17 19 9478130
Aravind L Koonin EV The HD domain defines a new superfamily of metal-dependent phosphohydrolases Trends Biochem Sci 1998 23 469 472 9868367
Kasahara M Ohmori M Activation of a cyanobacterial adenylate cyclase, CyaC, by autophosphorylation and a subsequent phosphotransfer reaction J Biol Chem 1999 274 15167 15172 10329724
van Nimwegen E Scaling laws in the functional content of genomes Trends Genet 2003 19 479 484 12957540
Dufresne A Salanoubat M Partensky F Artiguenave F Axmann IM Barbe V Duprat S Galperin MY Koonin EV Le Gall F Makarova KS Ostrowski M Oztas S Robert C Rogozin IB Scanlan DJ Tandeau de Marsac N Weissenbach J Wincker P Wolf YI Hess WR Genome sequence of the cyanobacterium Prochlorococcus marinus SS120, a nearly minimal oxyphototrophic genome Proc Natl Acad Sci USA 2003 100 10020 10025 12917486
Baar C Eppinger M Raddatz G Simon J Lanz C Klimmek O Nandakumar R Gross R Rosinus A Keller H Jagtap P Linke B Meyer F Lederer H Schuster SC Complete genome sequence and analysis of Wolinella succinogenes Proc Natl Acad Sci USA 2003 100 11690 11695 14500908
Methe BA Nelson KE Eisen JA Paulsen IT Nelson W Heidelberg JF Wu D Wu M Ward N Beanan MJ Dodson RJ Madupu R Brinkac LM Daugherty SC DeBoy RT Durkin AS Gwinn M Kolonay JF Sullivan SA Haft DH Selengut J Davidsen TM Zafar N White O Tran B Romero C Forberger HA Weidman J Khouri H Feldblyum TV Utterback TR Van Aken SE Lovley DR Fraser CM Genome of Geobacter sulfurreducens: metal reduction in subsurface environments Science 2003 302 1967 1969 14671304
Consortium BNGP The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability Proc Natl Acad Sci USA 2003 100 11660 11665 14500782
Heidelberg JF Seshadri R Haveman SA Hemme CL Paulsen IT Kolonay JF Eisen JA Ward N Methe B Brinkac LM Daugherty SC Deboy RT Dodson RJ Durkin AS Madupu R Nelson WC Sullivan SA Fouts D Haft DH Selengut J Peterson JD Davidsen TM Zafar N Zhou L Radune D Dimitrov G Hance M Tran K Khouri H Gill J Utterback TR Feldblyum TV Wall JD Voordouw G Fraser CM The genome sequence of the anaerobic, sulfate-reducing bacterium Desulfovibrio vulgaris Hildenborough Nat Biotechnol 2004 22 554 559 15077118
Kolker E Picone AF Galperin MY Romine MF Higdon R Makarova KS Kolker N Anderson GA Qiu X Auberry KJ Babnigg G Beliaev AS Edlefsen P Elias DA Gorby YA Holzman T Klappenbach JA Konstantinidis KT Land ML Lipton MS McCue LA Monroe M Pasa-Tolic L Pinchuk G Purvine S Serres MH Tsapin S Zakrajsek BA Zhu W Zhou J Larimer FW Lawrence CE Riley M Collart FR Yates JR Smith RD Giometti CS Nealson KH Fredrickson JK Tiedje JM Global profiling of Shewanella oneidensis MR-1: expression of hypothetical genes and improved functional annotations Proc Natl Acad Sci USA 2005 102 2099 2104 15684069
Hou S Saw JH Lee KS Freitas TA Belisle C Kawarabayasi Y Donachie SP Pikina A Galperin MY Koonin EV Makarova KS Omelchenko MV Sorokin A Wolf YI Li QX Keum YS Campbell S Denery J Aizawa S Shibata S Malahoff A Alam M Genome sequence of the deep-sea gamma-proteobacterium Idiomarina loihiensis reveals amino acid fermentation as a source of carbon and energy Proc Natl Acad Sci USA 2004 101 18036 18041 15596722
Ueda K Yamashita A Ishikawa J Shimada M Watsuji TO Morimura K Ikeda H Hattori M Beppu T Genome sequence of Symbiobacterium thermophilum, an uncultivable bacterium that depends on microbial commensalism Nucleic Acids Res 2004 32 4937 4944 15383646
McCue LA McDonough KA Lawrence CE Functional classification of cNMP-binding proteins and nucleotide cyclases with implications for novel regulatory pathways in Mycobacterium tuberculosis Genome Res 2000 10 204 219 10673278
Cann MJ Hammer A Zhou J Kanacher T A defined subset of adenylyl cyclases is regulated by bicarbonate ion J Biol Chem 2003 278 35033 35038 12829712
Cole ST Eiglmeier K Parkhill J James KD Thomson NR Wheeler PR Honore N Garnier T Churcher C Harris D Mungall K Basham D Brown D Chillingworth T Connor R Davies RM Devlin K Duthoy S Feltwell T Fraser A Hamlin N Holroyd S Hornsby T Jagels K Lacroix C Maclean J Moule S Murphy L Oliver K Quail MA Rajandream MA Rutherford KM Rutter S Seeger K Simon S Simmonds M Skelton J Squares R Squares S Stevens K Taylor K Whitehead S Woodward JR Barrell BG Massive gene decay in the leprosy bacillus Nature 2001 409 1007 1011 11234002
Wilde A Fiedler B Borner T The cyanobacterial phytochrome Cph2 inhibits phototaxis towards blue light Mol Microbiol 2002 44 981 988 12010493
Bhaya D Light matters: phototaxis and signal transduction in unicellular cyanobacteria Mol Microbiol 2004 53 745 754 15255889
Lory S Wolfgang M Lee V Smith R The multi-talented bacterial adenylate cyclases Int J Med Microbiol 2004 293 479 482 15149021
Perna NT Plunkett G Burland V Mau B Glasner JD Rose DJ Mayhew GF Evans PS Gregor J Kirkpatrick HA Posfai G Hackett J Klink S Boutin A Shao Y Miller L Grotbeck EJ Davis NW Lim A Dimalanta ET Potamousis KD Apodaca J Anantharaman TS Lin J Yen G Schwartz DC Welch RA Blattner FR Genome sequence of enterohaemorrhagic Escherichia coli O157:H7 Nature 2001 409 529 533 11206551
Hess WR Genome analysis of marine photosynthetic microbes and their global role Curr Opin Biotechnol 2004 15 191 198 15193326
Heermann R Fohrmann A Altendorf K Jung K The transmembrane domains of the sensor kinase KdpD of Escherichia coli are not essential for sensing K+ limitation Mol Microbiol 2003 47 839 848 12535080
Watts KJ Ma Q Johnson MS Taylor BL Interactions between the PAS and HAMP domains of the Escherichia coli aerotaxis receptor Aer J Bacteriol 2004 186 7440 7449 15489456
Ulrich LE Koonin EV Zhulin IB One-component systems dominate signal transduction in prokaryotes Trends Microbiol 2005 13 52 56 15680762
Aravind L Anantharaman V Balaji S Babu MM Iyer LM The many faces of the helix-turn-helix domain: Transcription regulation and beyond FEMS Microbiol Rev 2005 29 231 262 15808743
Kennelly PJ Archaeal protein kinases and protein phosphatases: insights from genomics and biochemistry Biochem J 2003 370 373 389 12444920
Jordan IK Makarova KS Spouge JL Wolf YI Koonin EV Lineage-specific gene expansions in bacterial and archaeal genomes Genome Res 2001 11 555 565 11282971
Hooper SD Berg OG On the nature of gene innovation: duplication patterns in microbial genomes Mol Biol Evol 2003 20 945 954 12716994
Gevers D Vandepoele K Simillon C Van de Peer Y Gene duplication and biased functional retention of paralogs in bacterial genomes Trends Microbiol 2004 12 148 154 15116722
Gerlt JA Babbitt PC Rayment I Divergent evolution in the enolase superfamily: the interplay of mechanism and specificity Arch Biochem Biophys 2005 433 59 70 15581566
Stewart V Chen LL Wu HC Response to culture aeration mediated by the nitrate and nitrite sensor NarQ of Escherichia coli K-12 Mol Microbiol 2003 50 1391 1399 14622424
Stewart V Nitrate- and nitrite-responsive sensors NarX and NarQ of proteobacteria Biochem Soc Trans 2003 31 1 10 12546643
Römling U Rohde M Olsen A Normark S Reinkoster J AgfD, the checkpoint of multicellular and aggregative behaviour in Salmonella typhimurium regulates at least two independent pathways Mol Microbiol 2000 36 10 23 10760159
Garcia B Latasa C Solano C Garcia-del Portillo F Gamazo C Lasa I Role of the GGDEF protein family in Salmonella cellulose biosynthesis and biofilm formation Mol Microbiol 2004 54 264 277 15458421
Stephenson K Hoch JA PAS-A domain of phosphorelay sensor kinase A: a catalytic ATP-binding domain involved in the initiation of development in Bacillus subtilis Proc Natl Acad Sci USA 2001 98 15251 15256 11734624
Brooun A Zhang W Alam M Primary structure and functional analysis of the soluble transducer protein HtrXI in the archaeon Halobacterium salinarium J Bacteriol 1997 179 2963 2968 9139915
Hou S Larsen RW Boudko D Riley CW Karatan E Zimmer M Ordal GW Alam M Myoglobin-like aerotaxis transducers in Archaea and Bacteria Nature 2000 403 540 544 10676961
Delgado-Nixon VM Gonzalez G Gilles-Gonzalez MA Dos, a heme-binding PAS protein from Escherichia coli, is a direct oxygen sensor Biochemistry 2000 39 2685 2691 10704219
Chen Y Cann MJ Litvin TN Iourgenko V Sinclair ML Levin LR Buck J Soluble adenylyl cyclase as an evolutionarily conserved bicarbonate sensor Science 2000 289 625 628 10915626
Harms N Reijnders WN Anazawa H van der Palen CJ van Spanning RJ Oltmann LF Stouthamer AH Identification of a two-component regulatory system controlling methanol dehydrogenase synthesis in Paracoccus denitrificans Mol Microbiol 1993 8 457 470 8392137
Xu HH Janka JJ Viebahn M Hanson RS Nucleotide sequence of the mxcQ and mxcE genes, required for methanol dehydrogenase synthesis in Methylobacterium organophilum XX: a two-component regulatory system Microbiology 1995 141 2543 2551 7582014
Seshadri R Adrian L Fouts DE Eisen JA Phillippy AM Methe BA Ward NL Nelson WC Deboy RT Khouri HM Kolonay JF Dodson RJ Daugherty SC Brinkac LM Sullivan SA Madupu R Nelson KE Kang KH Impraim M Tran K Robinson JM Forberger HA Fraser CM Zinder SH Heidelberg JF Genome sequence of the PCE-dechlorinating bacterium Dehalococcoides ethenogenes Science 2005 307 105 108 15637277
Forterre P Philippe H The last universal common ancestor (LUCA), simple or complex? Biol Bull 1999 196 373 377 11536914
Koonin EV Galperin MY Sequence - Evolution - Function. Computational Approaches in Comparative Genomics 2002 Boston , Kluwer Academic Publishers
Lopez-Garcia P Brochier C Moreira D Rodriguez-Valera F Comparative analysis of a genome fragment of an uncultivated mesopelagic crenarchaeote reveals multiple horizontal gene transfers Environ Microbiol 2004 6 19 34 14686938
Szurmant H Ordal GW Diversity in chemotaxis mechanisms among the bacteria and archaea Microbiol Mol Biol Rev 2004 68 301 319 15187186
Bardy SL Ng SY Jarrell KF Recent advances in the structure and assembly of the archaeal flagellum J Mol Microbiol Biotechnol 2004 7 41 51 15170402
Sogaard-Andersen L Cell polarity, intercellular signalling and morphogenetic cell movements in Myxococcus xanthus Curr Opin Microbiol 2004 7 587 593 15556030
Karev GP Wolf YI Rzhetsky AY Berezovskaya FS Koonin EV Birth and death of protein domains: A simple model of evolution explains power law behavior BMC Evol Biol 2002 2 18 12379152
Kirby JR Zusman DR Chemosensory regulation of developmental gene expression in Myxococcus xanthus Proc Natl Acad Sci USA 2003 100 2008 2013 12566562
Osterman A Overbeek R Missing genes in metabolic pathways: a comparative genomics approach Curr Opin Chem Biol 2003 7 238 251 12714058
Krieger CJ Zhang P Mueller LA Wang A Paley S Arnaud M Pick J Rhee SY Karp PD MetaCyc: a multiorganism database of metabolic pathways and enzymes Nucleic Acids Res 2004 32 D438 42 14681452
Aravind L Anantharaman V Iyer LM Evolutionary connections between bacterial and eukaryotic signaling systems: a genomic perspective Curr Opin Microbiol 2003 6 490 497 14572542
Raghunathan A Price ND Galperin MY Makarova KS Purvine S Picone AF Cherny T Xie T Reilly TJ Munson RJ Tyler RE Akerley BJ Smith AL Palsson BO Kolker E In silico metabolic model and protein expression of Haemophilus influenzae strain Rd KW20 in rich medium OMICS 2004 8 25 41 15107235
Barabasi AL Oltvai ZN Network biology: understanding the cell's functional organization Nat Rev Genet 2004 5 101 113 14735121
Balazsi G Barabasi AL Oltvai ZN Topological units of environmental signal processing in the transcriptional regulatory network of Escherichia coli Proc Natl Acad Sci USA 2005
Entrez Genome Project
Cummings L Riley L Black L Souvorov A Resenchuk S Dondoshansky I Tatusova T Genomic BLAST: custom-defined virtual databases for complete and unfinished genomes FEMS Microbiol Lett 2002 216 133 138 12435493
Shi L Hulett FM The cytoplasmic kinase domain of PhoR is sufficient for the low phosphate-inducible expression of pho regulon genes in Bacillus subtilis Mol Microbiol 1999 31 211 222 9987123
Pruitt KD Tatusova T Maglott DR NCBI Reference Sequence (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins Nucleic Acids Res 2005 33 D501 D504 15608248
The NCBI Reference Sequence collection
Tatusova TA Madden TL BLAST 2 Sequences, a new tool for comparing protein and nucleotide sequences FEMS Microbiol Lett 1999 174 247 250 10339815
Rost B Fariselli P Casadio R Topology prediction for helical transmembrane proteins at 86% accuracy Protein Sci 1996 5 1704 1718 8844859
Krogh A Larsson B von Heijne G Sonnhammer EL Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes J Mol Biol 2001 305 567 580 11152613
Tusnady GE Simon I The HMMTOP transmembrane topology prediction server Bioinformatics 2001 17 849 850 11590105
|
15955239
|
PMC1183210
|
CC BY
|
2021-01-04 16:03:39
|
no
|
BMC Microbiol. 2005 Jun 14; 5:35
|
utf-8
|
BMC Microbiol
| 2,005 |
10.1186/1471-2180-5-35
|
oa_comm
|
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-371597514210.1186/1471-2180-5-37Research ArticleEvaluation of the HOOF-Print assay for typing Brucella abortus strains isolated from cattle in the United States: results with four performance criteria Bricker Betsy J [email protected] Darla R [email protected] Bacterial Diseases of Livestock Research Unit, National Animal Disease Center, Agricultural Research Service, United States Department of Agriculture, 2300 Dayton Rd, Ames, IA, 50010, USA2 Diagnostic Bacteriology Laboratory, National Veterinary Services Laboratories, Animal and Plant Health Inspection Service, United States Department of Agriculture, 1800 Dayton Rd, Ames, IA, 50010, USA2005 23 6 2005 5 37 37 12 10 2004 23 6 2005 Copyright © 2005 Bricker and Ewalt; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A fundamental question that arises during epidemiological investigations of bacterial disease outbreaks is whether the outbreak strain is genetically related to a proposed index strain. Highly discriminating genetic markers for characterizing bacterial strains can help in clarifying the genetic relationships among strains. Under the auspices of the European Society of Clinical Microbiology and Infectious Diseases, the European Study Group for Epidemiological Markers (ESGEM) established guidelines for evaluating the performance of typing systems based of a number of criteria. Recently, HOOF-Print genotype analysis, a new method for typing Brucella abortus strains based on hypervariability at eight tandem repeat loci, was described. This paper evaluates the HOOF-Print assay by four of the criteria set out by the ESGEM: typeability, reproducibility, power of discrimination, and concordance with other typing methods.
Results
The HOOF-Print Assay was evaluated with a test population composed of 97 unrelated field isolates and 6 common laboratory strains of B. abortus. Both typeability and reproducibility of the assay were excellent. Allele diversity and frequency varied widely among the eight loci, ranging from 1 to 13 alleles. The power of discrimination, measured by the Hunter-Gaston discrimination index (HGDI), varied by locus ranging from 0 to 0.89, where a maximal value of 1.0 indicates discrimination of all strains. The HGDI values calculated for subgroups sorted by biovar were similar to the values determined for the whole population. None of the individual loci achieved the recommended HGDI threshold of 0.95, but the HGDI of the composite profiles was 0.99 (93 unique genotypes from 97 field strains evaluated), well above the recommended threshold. By comparison, the HGDI value for biovar typing was 0.61 in a test population biased with disproportionate numbers of the less common biovars. Cluster analysis based on HOOF-Print genotypes assembled the strains into hierarchical groups with no apparent association with the time or location of strain isolation. Likewise, these hierarchical groups were not homogeneous with regard to biotype. In one extreme case, two field isolates with identical fingerprints were identified as different biovars by conventional methods.
Conclusion
The main purpose of this study was to assess the ability of HOOF-Print genotyping to discriminate unrelated field strains of B. abortus, and whether the assay met established requirements for bacterial strain typing methods. The discriminatory power of the assay was remarkable, considering the genetic homogeneity found among species within the genus. The assay met or exceeded all of the recommended levels for the performance criteria of typeability, reproducibility, and power of discrimination, however some inconsistencies with conventional biovar typing were observed. Nevertheless, the results indicate that with cautious interpretation, multilocus genotyping of polymorphic tandem repeats by HOOF-Print analysis could be a valuable complement to routine epidemiological investigations into localized B. abortus outbreaks.
==== Body
Background
During the epidemiological inquiry into a disease outbreak, investigators try to trace the outbreak strain back to the original source of infection. If a potential index strain is identified, the strains are compared to establish a genetic connection between them. To help in this endeavor, a variety of approaches to characterize and classify disease strains have been developed that exploit genotypic and/or phenotypic markers. However, it can be difficult to determine which approaches are practical and informative for routine investigation of the disease agent of choice. Furthermore, direct comparisons of published methods may not be possible because different types of data are generated. At the request of the European Society of Clinical Microbiology and Infectious Diseases, these questions and others were discussed by a panel of international experts, resulting in the formation of the European Study Group on Epidemiological Markers (ESGEM) in 1994. The result of the meeting was a published set of guidelines for the evaluation of epidemiological typing systems [1]. The guidelines describe the performance criteria that should be considered when evaluating a typing system for widespread use. These criteria include: typeability, reproducibility, stability, discriminatory power, epidemiologic concordance, and typing system concordance.
Recently, a new approach has been developed for genetic typing that exploits the greater than normal amount of polymorphism observed within genomic regions containing short (2 to a few tens of base-pairs), tandemly repeated DNA sequences [2-4]. The accelerated mutation rates associated with the repeated sequences are thought to be due to slip-strand mispairing (SSM) [5] and recombination (most likely in the form of gene conversion arising from double-strand break repair [6]), resulting in rapid micro-evolution within the locus. SSM occurs when DNA polymerase misreads the number of repeats on the template strand, typically causing step-wise mutations (the addition or loss of single repeat units). Recombination mechanisms can cause more dramatic expansions or contractions of repeat strings, especially within the very large repeat strings found in eukaryotes [6]. Each variable number tandem repeat (VNTR) locus mutates independently and at an individual rate determined by a number of factors including: the repeat sequence, the size and number of repeat units, the flanking sequence, DNA secondary structure and sequence function [7]. Characterization of these continuously evolving targets has facilitated differentiation of bacterial strains [8,9]. Examination of multiple tandem repeat loci enhances subtype characterization in two ways: firstly, the capacity for genetic discrimination increases when multiple VNTR loci are examined; secondly, the effects of homoplasy (identical alleles arising independently through convergence, reversal or parallelism) may be diminished. Multilocus VNTR analysis (MLVA) has become an effective technique that can discriminate many difficult-to-type bacteria, including many human pathogens such as Haemophilus influenzae [10]; Bacillus anthracis [11]; Yersinia pestis [12]; Francisella tularensis [13] and Mycobacterium tuberculosis [14].
Brucellosis is an economically important zoonotic disease found throughout many regions of the world, and until recently, throughout the U.S. One causative agent, Brucella abortus, is predominantly pathogenic for its natural hosts, cattle and bison (Bovidae family); but it is also pathogenic for humans and several incidental animal species including elk. The disease causes reproductive failure in the host species and chronic health problems in humans.
B. abortus is a member of a highly homogenous genus, exhibiting ~98.5% nucleotide sequence homology among species [15]. Conventional subtyping of Brucella strains into biovars, for epidemiological trace-back, relies on a large array of tests including phage susceptibility, metabolic, biochemical, and serological characterization [16]. Often the differences are subtle. Few biovars are recognized within most Brucella species and some species cannot be subtyped at all. Worldwide, B. abortus has seven biovars; only biovars 1, 2 and 4 occur in the U.S. Differences among these three biovars are minor; discrimination is based on serology and the ability to grow in culture media with certain dyes [16]. Historically, about 85% of U.S. strains were typed as biovar-1.
Previously, we reported the discovery of a reiterated 8-bp sequence, arrayed in tandemly ordered strings [17], that is present in at least eight loci within the sequenced genomes of three Brucella species [15,18,19]. A protocol was developed for assessing the number of repeats at each of the eight loci by PCR amplification. The resulting amplicons, containing the entire array of tandem repeats and a small amount of flanking sequence, are sized and the number of repeats is deduced from the length. Since all eight loci have the same 8-bp repeat sequence, the technique was named "HOOF-Prints", an acronym for hypervariable octameric oligonucleotide fingerprints[17].
This paper evaluates the HOOF-Prints technique as it is applied to a diverse collection of B. abortus field strains representing all three biovar subtypes isolated from throughout the U.S. Our primary goal was to assess how well HOOF-Print genotyping can discriminate among unrelated field strains. Genotyping performance is compared to conventional subtyping into biovars. Individual fingerprint patterns and allelic diversity are presented. The HOOF-Prints technique is also assessed by four of the performance criteria recommended by the ESGEM: typeability, reproducibility, power of discrimination, and concordance with other typing methods. Recommendations for test population criteria are also addressed.
Results and discussion
Selection of the test group
An important consideration for evaluating an epidemiological typing system is the selection of a suitable test population. The test population should be large (N > 100), consisting of a diverse collection of unrelated strains that is representative of the natural population in which the test is intended to be used [1]. The test group assembled for this study is listed in tabular form in Additional file 1. Information about biovar type, herd location and year of isolation is also included.
All of the strains tested, except for the reference strains, were randomly chosen from diagnostic specimens that had been cultured, positively identified as B. abortus, biovar typed, and archived by the diagnostic laboratory at the National Veterinary Services Laboratories (APHIS, USDA). As suggested by the ESGEM, the test population was diverse based on the location and date of collection. The field strains originated from 97 different cattle herds, including both dairy and market cattle. Most of the isolates (n = 95) were from the US, but one isolate from Mexico and one isolate from El Salvador were also included. Over half of the isolates (n = 54) were collected in 1991 when brucellosis was still widely disseminated over the country, but on the decline due to the success of the national brucellosis eradication program. In subsequent years, there was a dramatic decrease in outbreaks, so fewer isolates were available. Currently, reports of new outbreaks in the U.S. are uncommon. The primary source of new outbreaks has shifted from domestic cattle to wildlife reservoirs.
To measure the assay's power of discrimination, only one isolate from each herd was included in the test group to prevent statistical over-representation of genotypes. When possible, the infected herds originated from different cities, and in all cases the herds had different owners. However, information regarding possible epidemiological links among the herds was unavailable. A disproportionate number of B. abortus biovar-2 and biovar-4 isolates were selected to assess the level of natural genotypic diversity within these subtypes. These two biovars are relatively more common in wildlife reservoirs and have recently been on the increase in cattle, due to more cattle outbreaks originating from wildlife. Six common laboratory strains of B. abortus including the reference strains of B. abortus biovars 1, 2, and 4, and the two major vaccine strains were also included in this study. Only B. abortus biovars 1, 2 and 4 were included in the test group, since these are the biovars that occur naturally in the United States.
HOOF-Print analysis
Eight loci containing octameric tandem repeats were characterized for each of the 103 isolates in the test population. Tandem repeat loci were amplified by PCR with primers directed to the conserved sequences flanking the repeat regions and the total number of repeat units at each locus was deduced from the size of the corresponding amplicon [17]. The results are presented in Additional file 1. The HOOF-Print profiles of some of the strains used in this study have been previously reported in earlier studies (see Additional file 1) [17,20].
At each locus, the alleles are named for the calculated number of repeat units at that locus, such that Allele-4 contains 4 repeat units and Allele-8 contains 8 repeat units. The HOOF-Print (genotypic fingerprint) for an isolate was generated from the allelic profile at all eight loci of that isolate.
Typeability
An important feature of any typing method is the ability of that method to conclusively classify every sample to a specific type defined by the test parameters. The ESGEM recommends that "T" be as close to 1.0 as possible [1]. In the case of the HOOF-Print assay, all isolates produced an amplified product at each of the eight loci. Taq DNA Polymerase adds a non-template nucleotide resulting in 2 products differing by 1-bp. With this protocol, the majority of amplicon products contain the extra nucleotide, but a significant portion of amplicons without the non-template nucleotide are also produced, resulting in the resolution of 2 peaks, 1-bp apart, by capillary electrophoresis. Nevertheless, since the repeat units increased in increments of 8-bp, allele assignment was clear-cut and an allele was assigned for all eight loci (see Additional file 1), giving the method a typeability index of 1.0. It should be noted that some strains were assigned an allele called M for Locus-6 and also for Locus-1. These isolates repeatedly produced single amplicons that were outside the predicted size range. For Locus-6, all M amplicons were the same size. Sequence analysis of the amplicons from several of these isolates showed a specific deletion in the flanking sequence region. All of the sequenced mutants produced the same aberrant sequence. Sequence analysis of the M allele at Locus-1 also showed a deletion in the flanking region of the DNA. Although these isolates did not fall in the normal expected range of alleles, they gave consistent results and could be assigned to a new allele that was called M for mutant.
Reproducibility
Like typeability, it is critical that a typing method reliably produce the same result for a given sample. The ESGEM recommends a reproducibility index of R ≥ 0.95. This feature was continuously tested during the study since every isolate was independently tested twice. However, in a more structured evaluation, 20 isolates were analyzed in triplicate in a randomized sequence. In a blinded fashion, one individual was responsible for preparing the assay, performing the assay, and interpreting the results. The HOOF-Print Assay had an R = 0.998 at the locus level and an R = 0.983 at the composite fingerprint level. The reason that R was less than 1.0 was because in one test, Locus-6 was incorrectly recorded as containing Allele-2 instead of Allele-5. This was an obvious clerical error, not an experimental failure, since the raw data clearly showed Allele-5 as the only form present in the sample. Nevertheless, the assay easily met the suggested limit for reproducibility.
Allelic diversity among the HOOF-Print loci
The B. abortus strains exhibited extensive variability in the number and range of alleles at each locus (Table 1). Similar levels of locus diversity were observed among each of the three biovars individually and within the test population as a whole. For example, Loci 5 and 8 had very little diversity, regardless of the biovar designation. By contrast, Locus-7 is highly diverse in all three biovars, with the same alleles occurring in multiple biovars.
Among the loci with multiple alleles, the distribution of alleles by size resembled an asymmetrical bell shaped curve skewed towards smaller repeat numbers (data not shown). This pattern of distribution is consistent with step-wise mutations resulting from SSM. The overall trend in allele frequency is toward short strings of repeats ranging from 2 to 5 repeat units per locus.
Power of discrimination
A fundamental question addressed in epidemiological investigations is whether the outbreak strain is derived from or genetically related to the proposed index strain. The answer requires a method for differentiation of genetically related and unrelated strains. The more genetic markers that are available to define isolates, the easier it becomes to discriminate among related and unrelated strains. Therefore, the discriminatory power of a test indicates how successful the test will be in identifying genetic relationships among strains. The ESGEM recommends evaluating a large group (≥ 100 samples) of genetically diverse samples by the Hunter-Gaston Discrimination Index (HGDI) [21], and proposes as a limit that HGDI is ≥ 0.95. In other words, for any two randomly-chosen, unrelated isolates, there is a 95% or greater probability that they will be placed in separate groups. Since the HGDI is heavily influenced by the level of genetic diversity within the test population, our test population consisted of only one isolate per herd to avoid over-representation of related fingerprint profiles. To assess the demonstrable diversity within the less common biovars, a disproportionate number of randomly selected B. abortus biovar-2 and biovar-4 isolates was included. Despite all attempts to use genetically unrelated isolates, the level of genetic relatedness among outbreaks from the different herds is not known, and some isolates may be epidemiologically linked.
The discriminatory power of the conventional typing method for Brucella, biovar typing, was examined. The calculated HGDI for the total test population based solely on biovar typing was 0.60. It should be noted that this value is artificially high due to disproportionate representation of biovar-2 and biovar-4 isolates in the test population. Historically, about 85% of B. abortus infections in the U.S. were caused by biovar 1, which would result in an HGDI ≈ 0.2. Even with the disproportionate biovar representation, the discrimination index is well below the recommended power of discrimination, ≥ 0.95, but for many years it has been the only subtyping method available.
The discriminatory power of HOOF-Print genotyping was determined for each of the three biovar subpopulations and for the entire test population (n = 103). When the HGDI was calculated on a locus-by-locus basis within the biovar specific subgroups and within the total test population, the results were variable, ranging from 0 (no discrimination; Locus-8, all populations) to 0.93 (Locus-7; biovar-2 subgroup) as seen in Table 1. Although the locus-specific HGDI values varied slightly among the biovar subgroups, the trends were similar, indicating that HOOF-Print genotyping is equally discriminating for each biovar subgroup. On a single locus basis, none of the loci met the minimal discrimination level of ≥ 0.95.
When the alleles from all eight loci were combined into a composite HOOF-Print genotype, the test group of 103 samples exhibited 98 different fingerprint patterns (HGDI = 0.99). This is well above the recommended limit set by ESGEM. In only four cases did two isolates have the same fingerprint: isolates #16 and #17 [strains 1–2361 and 1–2385]; isolates #46 and #47 [strains 1–2421 and 1–2040]; isolates #79 and #96 [strains 98-0689 and 4-0213]; and isolates #102 and #103 [strains 6-0242 and 4-1303]; (see Additional file 1, highlighted in colored pairs). One field isolate, #25 [strain1-2052] matched the fingerprint pattern for the vaccine strain S19. Independent analysis of phenotypic and biochemical characteristics also identified this isolate as vaccine strain S19. Another field isolate, #31 [strain 1-2050], was previously characterized by biovar typing as vaccine strain S19. While the fingerprint pattern for this isolate was not an exact match for the fingerprint patterns for the other S19 samples, it differed by only 1 repeat unit at Locus-1. This type of difference is consistent with micro-evolution from a step-wise mutation event.
Cluster analysis of HOOF-Print genotypes from the test populations
The HOOF-Print assay was designed to complement epidemiological investigations. Genotypic similarities are assumed to demonstrate genetic linkage among related lineages, and conversely, related lineages would be expected to have genotypic similarities. However, since the mutation rates for the selected loci among Brucella strains have yet to be determined, the evolutionary distance defined by HOOF-Print genotyping is unclear. We used cluster analysis to see if the HOOF-Print genotypes could be used to infer long and short term evolutionary history. Pairwise genetic distances within the total 103 strain test population were calculated from the absolute difference in repeat units at each of the eight loci, consistent with the stepwise model of mutation. A dendrogram was created by the neighbor-joining method (Figure 1). The clusters appear to be independent of the time or location of strain isolation (e.g. isolates #18 and #19 that were isolated in 1999 and 1994, respectively). This is not surprising since the test population is presumably composed of unrelated strains. As expected, the few strains known to be genetically related did cluster appropriately (e.g. field isolates of the vaccine strain S19; RB51 and its parental strain 2308). Unfortunately, epidemiological information and comprehensive histories are not routinely submitted with diagnostic samples and so specific information about the selected test isolates was not available, making it impossible to assess the validity of the cluster results for these data. If VNTR polymorphism in Brucella is generated by a mechanism other than the stepwise mutation model, then the genetic relationships proposed in Figure 1 could be invalid.
Distribution of HOOF-Print genotypes and alleles by geographic region
We wanted to see if there was a connection between geographic region and the HOOF-Print genotypes or alleles. The multilocus genotype clusters shown in the dendrogram in (Figure 1) do not correspond to specific states or regions. However, because multilocus genotyping is so highly discriminating, nearly every HOOF-Print genotype that was identified is unique. It is possible that the rapid evolution of the most variable loci could potentially mask regional influences in genotype composition. To detect regional relationships at the locus level, the alleles for each locus were examined for asymmetrical geographic distribution. To simplify the data and assure that sufficient data was available for statistical analysis, the US was divided into 2 regions: the east and the west, separated by the Mississippi River. For loci that contain a large number of alleles, the alleles were grouped so that no more that 4 or 5 groups were compared. Only data from B. abortus field strains from the US were included (n = 95). The data for each locus was cross tabulated and analyzed by the chi-square test statistic to see if a statistically significant association between alleles and geographic region could be demonstrated. The data are presented in Table 2 and Additional file 2. Note that for Locus-6, Fisher's exact test was used instead of the chi-square test, because of the small number of strains carrying alleles with more than 2 repeat units. The data indicate that within most loci there are no significant associations between alleles and geographic region (p values ranging from 0.1739 to 0.8563). For Locus 1, there may be a slight imbalance in distribution, with smaller alleles (7 or less repeat units) appearing more often in the west and larger alleles (8 or more repeat units) appearing more often in the east (p = 0.0400).
While it is reasonable to expect a relationship between genotypes/alleles and geographic region, several factors can affect geographic distributions. In the case of brucellosis, the government sponsored eradication program is drawing to a close. As a result, the incidence of Brucella has decreased dramatically to nearly complete eradication. Many states have been disease free for decades. Therefore, the initial source of a new infection is often found outside of the local area and involves the importation of diseased cattle from elsewhere. In a few areas, transmission from wildlife to cattle has become the leading source of disease outbreaks. Only a very small number of residual endemic bovine infections continue to be identified and eradicated. As the number of infected herds decrease, each outbreak is more isolated and can be quickly contained so that subsequent regional spread of the associated genotype is less likely. Therefore, it is likely that with so the few infected herds, the remaining B. abortus genotypes in the US are no longer representative of the natural distribution of genotypes that were present prior to the eradication program.
Finally, the selection criteria for the test population used in this study were designed to maximize the number of unrelated strains. Strains were selected from a large number of different states, and when possible, only one isolate from each city was included. Thus, possible links between alleles and geographic region may not be represented in the test population.
Comparison of biovar typing and HOOF-Print genotyping
The fourth performance criterion recommended by the ESGEM is concordance of the typing results with the results of other typing methods. When HOOF-Print genotyping was compared with conventional biovar typing, the most obvious difference was the high discriminatory power of the genotyping method and low discriminatory power of the biovar typing method. This complicates the direct comparison of typing results because the numbers of subtypes generated vary greatly between the two methods. In the dendrogram shown in Figure 1, all 103 strains were clustered by genotypic similarity. However, the hierarchical groups assembled by genotype were not homogeneous for biovar type. Most genotype clusters contain an assortment of biovars (shown by color coding in Figure 1), although some small clusters of biovars were found (e.g. biovar-4 isolates #89, 90, 91, 92 and 93). It is important to keep in mind that the dendrogram in Figure 1 is only one of many possible arrangements of the data. Furthermore, the method used for clustering the data was based on the step-wise model of mutation. If a different mechanism or a combination of mechanisms is involved, then the clustering parameters used in this study do not apply and the clusters are invalid.
In the most extreme case, two isolates with identical fingerprints were identified as belonging to different biovars (see Additional file 1, isolate #79 – biovar 2 and isolate #96 – biovar 4, highlighted in yellow). This result was somewhat surprising, considering the clonal nature of Brucella, but there are at least two possible explanations for this observation. One possible cause is homoplasy, resulting from convergent evolution among unrelated strains. The hypermutability of VNTR loci results in continuous micro-evolution, and HOOF-Print genotyping reveals the most recent mutation events in a strain's genetic history. Even among the most variable loci in this study, a limited repertoire of alleles was found. Although the total number of allelic combinations is large, random convergence of genotype patterns between genetically unrelated strains is likely to occur occasionally. Incorporation of additional polymorphic loci into the assay may help resolve inconsistencies caused by homoplasy.
Alternatively, the strain could have typed as a different biovar type due to mutation(s) in genes that affect biovar phenotype. Only two differences distinguish B. abortus biovar 2 and biovar 4: biovar 2 strains are A-antigen dominant/fuchsin dye sensitive while biovar 4 strains are M-antigen dominant/fuchsin dye insensitive [16]. Both traits are associated with the bacterial surface and may require common genes. Although spontaneous mutations are much less common than VNTR polymorphisms, atypical biovar profiles are occasionally found among field isolates. Thus, this argument cannot be ruled out without further study.
Conclusion
The HOOF-Print Assay is a rapid, easy to perform technique for subtyping B. abortus strains. When the assay was evaluated with a test population consisting of a large number of unrelated, naturally occurring field isolates, the selected tandem repeat loci displayed considerable polymorphism as evidenced by the large number of alleles found at many of the loci. The study demonstrates that the HOOF-Print assay meets or exceeds the minimum limits recommended by the ESGEM for epidemiological typing, based on the performance criteria: typeability, reproducibility, and power of discrimination. When compared to conventional biovar typing, HOOF-Print genotyping is considerably more discriminating. However, some inconsistencies with conventional biovar typing were observed. Caution will be needed in interpreting the data, to prevent drawing incorrect conclusions from artificial genotypic similarity caused by homoplasy. Incorporation of additional polymorphic loci may help identify convergent evolution among unrelated strains. In practice, HOOF-Print genotyping should be utilized as a complement to conventional epidemiological investigations into the short-term history of localized brucellosis outbreaks. In the future, we plan to look at genotypic variability among related isolates to better understand how HOOF-Print genotypes evolve and to experimentally measure the VNTR mutation rates among the polymorphic B. abortus loci.
Methods
Disclaimer: Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture.
Isolation and characterization of the bacterial strains used in this study
The bacterial field strains used in this study (see Additional file 1) were originally isolated in the Bacterial Diagnostic Laboratory at the USDA National Veterinary Services Laboratories, Ames, IA, as part of the USDA Brucellosis Eradication Program. The suspect bacteria were subcultured, identified by conventional microbiological tests and biotyped by additional biochemical and phenotypic characteristics [16] prior to archiving at -70°C. The type strains were originally obtained from the American Type Culture Collection bank and stored at -70°C until propagated.
The selected strains were thawed, grown, harvested and preserved in 66% methanol. Many of the isolates were retrieved and prepared specifically for this study while other strains had been preserved in methanol for a number of years with no apparent DNA degradation.
HOOF-Print analysis
The HOOF-Print technique was performed as previously described [17] with minor modifications as follows: PCR amplification was performed with 1 unit of FastStart Taq polymerase (cat no. 2-032-902; Roche Biochemicals, Indianapolis, IN) in 50-mM Tris, 10-mM KCl, 1.5-mM MgCl2, and 5-mM (NH4)2SO4, pH 8.3, with the addition of dGTP, dATP, dCTP and dTTP at 250-μM each, primers at 250-nM each, and GC-Rich Resolution Solution (included with the polymerase) at 1X concentration, per manufacturer's instructions. Primer pairs [17] consisted of one primer fluorescently labeled with HEX, FAM or NED and one unlabeled primer. Both primers were designed to anneal to the conserved sequences flanking the repeat region. Amplification was done one locus at a time (8 assays per sample) or multiplexed with up to three primer sets per reaction (with a different dye for each primer set). Cycling parameters were as follows: 95°C for 5 min to activate the modified polymerase, followed by 37 cycles of 95°C for 30 sec, 53°C for 30 sec and 75°C for 45 sec; and ending with a 60-min incubation at 75°C to promote maximal addition of the non-templated nucleotide to the amplicons. The higher extension temperature increased amplicon yield, possibly by denaturing short hairpin formations in the target sequence. The amplified samples were maintained at 4°C prior to analysis.
The degree of amplification success was monitored by gel electrophoresis through 4% agarose. Before size analysis, the fluorescent amplicons were diluted in water, usually in a 1:100 or 1:200 ratio, then separated by capillary electrophoresis on an ABI Prism 3100 Genetic Analyzer with GeneScan-500 [ROX] size markers (cat. no. 401734; Applied BioSystems, Foster City, CA). Data were collected and the amplicon sizes were determined with GeneScan Data Analysis Software, ver 3.7 (Applied BioSystems). The number of repeat copies was deduced from the amplicon size.
For some isolates, more than one product was reproducibly synthesized for a given locus, typically differing by a single repeat unit. This phenomenon, which was especially common in reference strains, is thought to be the result of micro-evolution. Minor heterogeneity at some loci was not unexpected since many of these samples were primary isolates that had not been propagated clonally. In these cases the dominant allele was used for analysis.
Reproducibility analysis
A reproducibility evaluation was integrated into the protocol by analyzing every sample twice in independent assays. To formally evaluate the reproducibility of the HOOF-Print technique an additional blinded survey was performed. Twenty isolates were prepared in triplicate and randomized in order. A different individual prepared the reaction mixes, performed the assay and analyzed the data, unaware of the identities of the samples.
Statistical analyses
The following statistical analyses were performed as suggested by the ESGEM [1].
Typeability is the success in determining an unambiguous type for each isolate, calculated by the formula where T is the typeability, Nt is the number of strains characterized by a complete array of alleles, and N is the size of the test population. If a complete 8-allele fingerprint can be determined for all strains in the test population, then T = 1.
Reproducibility is the ability to assign the same type to an isolate in independently repeated tests. It is calculated from the equation where R is the reproducibility index, Nr is the number of strains that are repeatedly assigned a single type and N is the total number in the test sample.
Allele frequency is the relative proportion of each allele, given as a percentage by the formula where nj is the number of strains with the allele "j"; and N is the size of the test population.
The Hunter-Gaston Discrimination Index [21] measures the power of discrimination for a typing method by calculating the probability that two unrelated strains will be correctly assigned to different types. The equation used is
where DI is the index of discrimination; N is the size of the
test population; s is the total number of alleles per locus or fingerprint patterns in the population, and nj is the number of isolates with the allele "j".
Cluster and molecular evolutionary analyses were conducted using MEGA version 2.1 [22]. A distance matrix was created using pairwise comparison of HOOF-Print genotypes from all of the 103 isolates in the test population. The genetic distance was calculated from the absolute difference in repeat units at each locus, assuming that each incremental change in repeat number represents an equal and independent mutation event. The distance matrix was used to make an unrooted tree using the neighbor-joining method [23].
Statistical analysis of the distribution of alleles by geographic region was done by creating a cross tabulation of each locus into two regions, defined as east or west of the Mississippi River. For the loci that contain a large number of alleles, the alleles were grouped so that no more than four or five groups were compared to make sure that enough data for each locus would be compared (see Additional file 2). Differences among the data were analyzed using the chi-square test statistic or, in the case of Locus-6, the Fisher's exact test (Table 2). Statistical analyses were performed with the SAS system version 8 (SAS Institute, Cary, N.C.) using the FREQ Procedure based on a sample size of 95. A P value < 0.05 was considered to be significant.
List of abbreviations
PCR – polymerase chain reaction; bp – base pair; SSM – slip-strand mispairing; DI – discrimination index; HGDI – Hunter Gaston discrimination index; ESGEM – European Study Group on Epidemiological Markers; VNTR – variable number tandem repeats; MLVA – multilocus VNTR analysis; SSR – short sequence repeats.
Authors' contributions
DRE isolated, cultivated, harvested and processed most of the bacterial strains used in this study; typed each strain by conventional biotyping methods and compiled the available background information for each isolate. BJB performed the HOOF-Print assay on all strains, determined the genotypes, performed most of the statistical analyses of the data, and wrote the manuscript. Both authors contributed to discussions about the data, read and approved the final manuscript.
Supplementary Material
Additional File 1
Brucella strains and HOOF-Print genotypes. List of all of the bacterial isolates used in this study, including the biovar designation of the isolate, the location of the infected herd, the year of isolation, and the HOOF-Print genotype.
Click here for file
Additional File 2
Cross-tabulation of allelic distribution within the variable HOOF-Print loci by geographic region. Statistical data from each of the six most variable loci are presented in cross-tabulations displaying the distribution of alleles between the eastern and western regions of the US.
Click here for file
Acknowledgements
We would like to thank Dr. Bruce Wagner of the Centers for Epidemiology and Animal Health, Veterinary Services, Animal and Plant Health Inspection Service, USDA, for his statistical assistance; and Nancy Koster for her technical assistance in this project.
Figures and Tables
Figure 1 Neighbor-joining dendrogram of clustered HOOF-Print genotypes. Genotyped were clustered into groups based on the differences in the numbers of repeat units at the eight VNTR loci. Biovars are differentiated by colored text: black = biovar-1; green = biovar-2 and red = biovar-4.
Table 1 Allelic diversity among B. abortus VNTR loci
No. of Alleles Range of Repeats Discrimination Indexa
Biovar 1 strains
Locus-1 14 2 – 14, Mc 0.90
Locus-2 4 3 – 6 0.66
Locus-3 9 3 – 16 0.78
Locus-4 7 2 – 10 0.77
Locus-5 2 2 – 3 0.07
Locus-6 3 2 – 3, M 0.17
Locus-7 12 3 – 14 0.87
Locus-8 1 2 0.00
Biovar 2 strains
Locus-1 8 2 – 10 0.80
Locus-2 5 3 – 7 0.71
Locus-3 8 2 – 13 0.86
Locus-4 6 2 – 7 0.78
Locus-5 2 2 – 5 0.14
Locus-6 4 2 – 4, M 0.39
Locus-7 12 3 – 15 0.93
Locus-8 1 2 0.00
Biovar 4 strains
Locus-1 10 2 – 12 0.92
Locus-2 2 4 – 5 0.53
Locus-3 5 1 – 6 0.74
Locus-4 5 2 – 6 0.77
Locus-5 1 2 0.00
Locus-6 2 2, M 0.43
Locus-7 9 3 – 12 0.90
Locus-8 1 2 0.00
All strainsb
Locus-1 14 2 – 14, M 0.87
Locus-2 5 3 – 7 0.67
Locus-3 11 1 – 16 0.82
Locus-4 7 2 – 13 0.78
Locus-5 3 2 – 5 0.04
Locus-6 5 2 – 5, M 0.40
Locus-7 13 3 – 15 0.89
Locus-8 1 2 0.00
a-The Discrimination Index was calculated for each locus by the method of Hunter and Gaston as described in the Materials and methods section; b -These numbers were calculated from the entire 103-sample test population. c – "M" indicates an atypical (mutant) allele.
Table 2 Statistical analysis of the allelic distributions within each VNTR locus by geographic regiona
VNTR Locus P value
Locus-1 0.0400
Locus-2 0.1739
Locus-3 0.8365
Locus-4 0.4147
Locus-6 0.2824
Locus-7 0.8563
a – See Additional file 2 for the data that was used in the analysis.
==== Refs
Struelens MJ Consensus guidelines for appropriate use and evaluation of microbial epidemiologic typing systems Clin Microbiol Infect 1996 2 2 11 11866804
Field D Wills C Abundant microsatellite polymorphism in Saccharomyces cerevisiae, and the different distributions of microsatellites in eight prokaryotes and S. cerevisiae, result from strong mutation pressures and a variety of selective forces Proc Natl Acad Sci USA 1998 95 1647 1652 9465070 10.1073/pnas.95.4.1647
van Belkum A Scherer S van Alphen L Verbrugh H Short-sequence DNA repeats in prokaryotic genomes Microbiol Mol Biol Rev 1998 62 275 293 9618442
Le Fleche P Hauck Y Onteniente L Prieur A Denoeud F Ramisse V Sylvestre P Benson G Ramisse F Vergnaud G A tandem repeats database for bacterial genomes: application to the genotyping of Yersinia pestis and Bacillus anthracis BMC Microbiol 2001 1 2 Epub 2001 Mar 30 11299044 10.1186/1471-2180-1-2
Levinson G Gutman GA Slipped-strand mispairing: A major mechanism for DNA sequence evolution Mol Biol Evol 1987 4 203 221 3328815
Paques F Leung W-Y Haber JE Expansions and contractions in a tandem repeat induced by double-strand break repair Mol Cell Biol 1998 18 2045 2054 9528777
Brinkmann B Klintschar M Neuhuber F Huhne J Rolf B Mutation rate in human microsatellites: influence of the structure and length of the tandem repeat Am J Hum Genet 1998 62 1408 1415 9585597 10.1086/301869
van Belkum A The role of short sequence repeats in epidemiologic typing Curr Opin Microbiol 1999 2 306 311 10383858 10.1016/S1369-5274(99)80053-8
Keim P Klevytska AM Price LB Schupp JM Zinser G Smith KL Hugh-Jones ME Okinaka R Hill KK Jackson PJ Molecular diversity in Bacillus anthracis J Appl Microbiol 1999 87 215 217 10475952 10.1046/j.1365-2672.1999.00873.x
van Belkum A Scherer S van Leeuwen W Willemse D van Alphen L Verbrugh H Variable number of tandem repeats in clinical strains of Haemophilus influenzae Infect Immun 1997 65 5017 5027 9393791
Keim P Price LB Klevytska AM Smith KL Schupp JM Okinaka R Jackson PJ Hugh-Jones ME Multiple-locus variable-number tandem repeat analysis reveals genetic relationships within Bacillus anthracis J Bacteriol 2000 182 2928 2936 10781564 10.1128/JB.182.10.2928-2936.2000
Klevytska AM Price LB Schupp JM Worsham PL Wong J Keim P Identification and characterization of variable-number tandem repeats in the Yersinia pestis genome J Clin Microbiol 2001 39 3179 3185 11526147 10.1128/JCM.39.9.3179-3185.2001
Johansson A Goransson I Larsson P Sjostedt A Extensive allelic variation among Francisella tularensis strains in a short-sequence tandem repeat region J Clin Microbiol 2001 39 3140 3146 11526142 10.1128/JCM.39.9.3140-3146.2001
Le Fleche P Fabre M Denoeud F Koeck JL Vergnaud G High resolution, on-line identification of strains from the Mycobacterium tuberculosis complex based on tandem repeat typing BMC Microbiol 2002 2 37 12456266 10.1186/1471-2180-2-37
Paulsen IT Seshadri R Nelson KE Eisen JA Heidelberg JF Read TD Dodson RJ Umayam L Brinkac LM Beanan MJ Daugherty SC Deboy RT Durkin AS Kolonay JF Madupu R Nelson WC Ayodeji B Kraul M Shetty J Malek J Van Aken SE Riedmuller S Tettelin H Gill SR White O Salzberg SL Hoover DL Lindler LE Halling SM Boyle SM Fraser CM The Brucella suis genome reveals fundamental similarities between animal and plant pathogens and symbionts Proc Natl Acad Sci USA 2002 99 13148 13153 12271122 10.1073/pnas.192319099
Alton GG Jones LM Pietz DE Laboratory Techniques in Brucellosis Monograph Series No 55 1975 2 Geneva, Switzerland: World Health Organization 34 59
Bricker BJ Ewalt DR Halling SM Brucella 'HOOF-Prints': strain typing by multi-locus analysis of variable number tandem repeats (VNTRs) BMC Microbiol 2003 3 15 12857351 10.1186/1471-2180-3-15
DelVecchio VG Kapatral V Redkar RJ Patra G Mujer C Los T Ivanova N Anderson I Bhattacharyya A Lykidis A Reznik G Jablonski L Larsen N D'Souza M Bernal A Mazur M Goltsman E Selkov E Elzer PH Hagius S O'Callaghan D Letesson JJ Haselkorn R Kyrpides N Overbeek R The genome sequence of the facultative intracellular pathogen Brucella melitensis Proc Natl Acad Sci USA 2002 99 443 448 11756688 10.1073/pnas.221575398
Halling SM Peterson-Burch BD Bricker BJ Zuerner RL Qing Z Li L-L Kapur V Alt DP Olsen SC Completion of the genome sequence of Brucella abortus and comparison to the highly similar genomes of B. melitensis and B. suis J Bacteriol 2005 187 2715 2726 15805518 10.1128/JB.187.8.2715-2726.2005
Bricker BJ Lopez-Goni I, Moriyon, I Molecular diagnostics of animal brucellosis: A review of PCR-based assays and approaches Brucella molecular and cellular biology 2004 Wymondham, UK: Horizon Bioscience 25 51
Hunter PR Gaston MA Numerical index of the discriminatory ability of typing systems: an application of Simpson's index of diversity J Clin Microbiol 1988 26 2465 2466 3069867
Kumar S Tamura K Jakobsen IB Nei M MEGA2: molecular evolutionary genetics analysis software Bioinformatics 2001 17 1244 1245 11751241 10.1093/bioinformatics/17.12.1244
Saitou N Nei M The neighbor-joining method: A new method for reconstructing phylogenetic trees Molec Biol Evol 1987 4 406 425 3447015
|
15975142
|
PMC1183211
|
CC BY
|
2021-01-04 16:03:40
|
no
|
BMC Microbiol. 2005 Jun 23; 5:37
|
utf-8
|
BMC Microbiol
| 2,005 |
10.1186/1471-2180-5-37
|
oa_comm
|
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-431603365810.1186/1471-2180-5-43SoftwareCharacterization of Pseudomonas aeruginosa isolated from chronically infected children with cystic fibrosis in India Agarwal Gunjan [email protected] Arti [email protected] Susheel Kumar [email protected] Bimal Kumar [email protected] Sada Nand [email protected] Department of Microbiology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi-110029, INDIA2 Department of Pediatrics, All India Institute of Medical Sciences, Ansari Nagar, New Delhi-110029, INDIA3 Department of Biostatistics, All India Institute of Medical Sciences, Ansari Nagar, New Delhi-110029, INDIA2005 21 7 2005 5 43 43 3 4 2005 21 7 2005 Copyright © 2005 Agarwal 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
Pseudomonas aeruginosa is the leading cause of morbidity and mortality in patients with cystic fibrosis (CF). With chronicity of infection, the organism resides as a biofilm, shows multi-drug resistance, diversifies its colony morphology and becomes auxotrophic. The patients have been found to be colonized with multiple genotypes. The present work was carried out to characterize P. aeruginosa isolated from children with cystic fibrosis using phenotypic and genotypic methods.
Results
We studied 56 patients with CF attending the Pediatric Chest clinic at All India Institute of Medical Sciences, New Delhi, India during August 1998-August 2001. These patients were regularly followed up at the clinic. Out of 56 patients, 27 were culture positive for P. aeruginosa where 8 were chronically infected (Group1) and 19 were intermittently colonized with the organism (Group2). Patients under Group1 had significantly higher rates of hospitalization, death and colonization with different colony morphotypes (p < 0.05). The isolates from Group1 patients were the positive producers of extended spectrum beta lactamase. A total of 5 auxotrophs were recovered from 2 patients where one was chronically infected with P. aeruginosa and the other was a recently enrolled patient. The auxotrophs had the specific requirement for methionine and arginine. Molecular typing revealed 33 ERIC-PCR (E1-E33) and 5 PCR-ribotyping (P1-P5) patterns. By ERIC-PCR, 4 patients were colonized with 2–4 genotypes and the remaining 23 patients were colonized with the single genotype.
Conclusion
With chronicity of infection, P. aeruginosa becomes multidrug resistant, diversifies its colony morphology, acquires mucoidity and shows auxotrophy for amino acids. The chronically infected patients can be colonized with multiple genotypes. Thus in a particular clinical set up, high index of suspicion should be there for diagnosis of CF patients so as to prevent the delay in diagnosis and management of CF patients.
==== Body
Background
Cystic fibrosis (CF) is the recessive genetic disorder of Caucasian race. CF in Asians and Indians was thought to be almost nonexistent in the past. However, recent reports have suggested CF to be more common in Indian population than previously thought. For proper management, it is important to diagnose the infection in the patients with CF [1-3].
The mutations in the cystic fibrosis transmembrane regulator gene in CF patients lead to recurrent and chronic respiratory tract infections, which serves as a major cause of morbidity and mortality [4]. If appropriate antibiotics are used, the quality of life improves in CF patients. Thus, the knowledge of the etiological agents and their antimicrobial susceptibility can help in planning the appropriate antibiotic therapy to be instituted in CF patients.
Infection due to Pseudomonas aeruginosa (P. aeruginosa) has been implicated as a major cause of morbidity and mortality in patients with CF [5-8]. Once P. aeruginosa colonizes the respiratory passages, it is rarely possible to eradicate it. Prolonged and repeated administration of drugs lead to the development of multidrug resistance in P. aeruginosa strains. To detect emergence of the resistant strains, it is important to monitor the antibiotic susceptibility pattern of P. aeruginosa isolates. Increased resistance to gentamicin, quinolone and cephalosporins has been reported in chronically infected patients [9,10].
With severity of pulmonary infection, the organism undergoes phenotypic modification, which helps in its persistence under adverse environment of the respiratory tract. The organism diversifies into different colony morphotypes [11,12] and develops antibiotic resistance for its survival. This may be due to an adaptation of the organism to the host defense mechanisms and antibiotic selective pressure [13]. Under recurrent and chronic endobronchial P. aeruginosa infections, the initial colonization is with nonmucoid forms of an organism but with progressive lung tissue damage, it gets converted to the mucoid form. The mucoidity helps an organism to grow as biofilm, which serves as a protective niche and helps to evade the host immune response and action of antibiotics. The organism also becomes auxotrophic for amino acid due to high amino acid concentration in the respiratory passages. The most commonly required amino acid reported in literature is methionine while other amino acids-leucine, arginine and ornithine have been required by a minority of P. aeruginosa strains [14].
During the course of infection, the organism may undergo genotypic changes. The chronically infected patients have been found to have long-term colonization by one predominant genotype but variation in genotypes within a single patient has also been observed [13,15].
Since no data is available on the characterization of P. aeruginosa from CF patients in India, the present study was undertaken to study phenotypic and genotypic variations of P. aeruginosa isolated from CF patients attending the Pediatric Chest Clinic at All India Institute Of Medical Sciences (AIIMS), New Delhi, INDIA.
Results
Subjects
A total of 56 patients with CF regularly attended the Pediatric Chest Clinic between August 1998-August 2001. These patients were followed monthly at the chest clinic. At the time of enrollment, all the patients were below 15 yrs of age (mean age (yrs) ± standard deviation was 5.33 ± 4.3 and median age was 4.5). The male: female ratio was 2:1.
Out of 56 patients, culture positive infection was observed in 33(58.9%) patients of whom 27(81.8%) were culture positive for P. aeruginosa and 6(18.2%) patients were culture positive for organism other than P. aeruginosa where one patient each was colonized with Citrobacter freundii, Staphylococcus aureus, Escherchia coli, Enterobacter sp and two patients were colonized with Klebsiella pneumoniae. The rate of infection with P. aeruginosa was significantly higher than the rate of infection with organism other than P. aeruginosa in CF patients (p < 0.05). Out of 27 culture positive patients, 8 had chronic infection with the organism and were included in Group1. The remaining 19 patients were intermittently colonized with the organism and were included in Group2.
During the study period, 9 patients required hospitalization and 4 have died. Of the 4 deaths, 3 occurred during the hospitalization. All the patients who required hospitalization were culture positive for P. aeruginosa and were followed daily. The mean duration of hospitalization was 12 days (range was 7–20 days). Out of 9 patients, 6 were included in Group1 and 3 in Group 2. All 4 patients who died, had pulmonary exacerbations and were chronically infected with P. aeruginosa. The mortality related to P. aeruginosa infection in the study was 14.8%. The frequency of hospitalization and death in chronically infected patients was significantly higher as compared to intermittently colonized patients with P. aeruginosa (p < 0.05).
Antimicrobial susceptibility
Three hundred and fifty P. aeruginosa isolates were recovered from 27 patients where 94(26.9%) showed the mucoid morphotype (Type5) and the rest 256 (73.1%) isolates showed nonmucoid morphotype(Type1–4 and Type 6). The antimicrobial susceptibility was determined by agar dilution method based on NCCLS guidelines and antibiograms were further constructed (Table 1). One hundred and forty three (41%) isolates were multidrug resistant and showed A3 and A5-A8 antibiogram. All the isolates were recovered from chronically infected patients.
We further characterized all 29 P. aeruginosa isolates that showed resistance to ceftazidime and were the positive producers of extended spectrum beta lactamase (ESBL). These isolates were detected as ESBL positive by disk diffusion according to NCCLS guidelines (Fig 1A) and further confirmation was done by Etest method (Fig. 1B). Twenty-four of the 29 (82.8%) ESBL positive isolates were multidrug resistant (A6 antibiogram). All 29 isolates were found sensitive to imipenam and meropenam.
The ESBL positive isolates were recovered from 3 patients (PCC.No. 180/95, 160/98 and 499/00) where 2 (PCC.No. 180/95 and 160/98) were chronically infected and one patient (PCC.No.499/00) was intermittently colonized. Out of 29 ESBL positive isolates, 27 were recovered from the patient (PCC.No. 180/95) during his different follow-ups past two years and 1 isolate each was recovered from 2 patients (PCC.No. 160/98 and 499/00).
Phenotypic characterization of P. aeruginosa isolates
Colony morphotypes
Out of 27 patients infected with P. aeruginosa, 7 harbored different colony morphotypes in a single sputum sample while the rest 20 patients harbored Type 1 colony morphotype. All the 7 patients were chronically infected with the organism(Group 1). The rate of colonization with different colony morphotypes of P. aeruginosa was significantly higher in Group 1 patients (p < 0.05).
Six (85.7%) of 7 patients harbored mucoid P. aeruginosa in their respiratory samples and all were chronically infected. The rate of colonization with mucoid P. aeruginosa was significantly higher (75%) in chronically infected patients (p < 0.05). Of the 6 patients who were positive for mucoid P. aeruginosa, 3 have died.
Out of 350 P. aeruginosa isolates, 194(55.4%) showed Type 1 morphotype. The distribution of antibiogram among colony morphotypes is shown in Table 2.
Auxotrophy
Out of 350 P. aeruginosa isolates, 5 were auxotrophs, which did not grow on minimal agar medium (MAM) and required Mueller Hinton agar (MHA) for their growth. Of the 5 auxotrophs, four (isolate No.2–5) were isolated from the chronically infected patient (PCC.No.180/95) who was diagnosed as CF six years ago. One auxotroph (isolate No.1) was isolated from a recently enrolled patient (PCC.No.499/00). The respiratory samples collected during their subsequent follow-ups were found negative for the presence of auxotrophs. These auxotrophs were recovered when these patients presented with pulmonary exacerbation.
The detection of specific amino acid requirements of auxotrophs revealed that 3(isolate No.1–3) had the specific requirement for arginine (Fig. 2) and 2(isolate No. 4,5) for methionine. The reproducible results were obtained on repeating the experiment thrice.
The resistance against different antibiotics was compared between auxotrophs and prototrophs. Although auxotrophs were small in number, they showed significantly higher resistance against ceftazidime (60%) as compared to prototrophs (p < 0.05). Three auxotrsophs (isolate No. 1–3) were multidrug resistant (A6 antibiogram) and the producers of ESBL. The reproducible results were obtained on repeating the experiment thrice.
Molecular characterization of P. aeruginosa isolates
Molecular characterization of P. aeruginosa isolates (N = 350) recovered from 27 patients infected with P. aeruginosa was achieved by Enterobacterial Repetitive Intergenic Consensus Polymerase Chain Reaction (ERIC-PCR) and PCR-ribotyping.
ERIC-PCR
A total of 33 ERIC-PCR (E1-E33) patterns were constructed (Table 3). All 27 patients infected with P. aeruginosa harbored different genotypes with their unique ERIC-PCR pattern except one patient (PCC.No. 499/00) who was colonized with the similar genotype (E1). Out of 8 patients who were chronically infected with P. aeruginosa, 4 patients (PCC.No.180/95, 179/93, 160/98 and 10/99) were colonized with 2–4 genotypes (Fig. 3A&3B). The remaining 23 patients (Group1 = 4 and Group 2 = 19 patients) were colonized with a single genotype.
PCR-ribotyping
A total of 5 PCR ribotyping patterns (P1-P5) were obtained (Fig. 4A &4B). The molecular weight (bp) of each PCR-ribotyping pattern was P1:570, P2: 730, P3: 1500 &610, P4: 600 and P5: 570&130. Twenty four of 27 patients harbored P. aeruginosa with a single PCR-ribotyping Pattern P1 and 3 patients (PCC.No. 180/95, 160/98 and 179/93) were colonized with P. aeruginosa with 2 or more PCR-ribotyping patterns. All 3 patients were chronically infected with the organism. Isolates from patient PCC.No 180/95 Pattern P1-P4, PCC.No. 160/98 Pattern P1 and P2 and patient PCC.No. 179/93 Pattern P1 and P5.
By ERIC-PCR, P. aeruginosa isolates from 26/27 patients revealed different genotypes (E1-E33) while by PCR-ribotyping, 24/27 patients revealed a single PCR-ribotyping pattern (Pattern P1).
Table 3 shows the compilation of results on the patient basis including the phenotypic and genotypic characteristics of each patient.
Discussion
Cystic fibrosis (CF) is the most common inherited fatal disease in Caucasian population. This generalized disorder of salt and water transport of exocrine glands is caused by mutations in cystic fibrosis transmembrane regulator gene in the airway epithelial cells and submucosal glands. These patients are very prone to bronchopulmonary infection due to lack of the innate defensive mechanisms. Pulmonary infection is the most common cause of morbidity and mortality in CF patients [16].
CF was thought to be extremely rare in India. However, recent reports from Indian subcontinent suggested CF to be more common in Indian population than what was previously thought [1]. In our study, we have followed all 56 patients attending the Pediatric Chest Clinic at All India Institute Of Medical Sciences, New Delhi, INDIA between August 1998-August 2001. The median age of diagnosis of patients was 4.5 years, which was much higher than what was earlier reported in the literature where the median age was found as 6 and 12 months among Caucasian American and Indian American children respectively [17]. This may be due to delayed diagnosis of CF patients in India.
P. aeruginosa has been the predominant bacterium associated with pulmonary infection in patients with CF [18-22]. In our study, P. aeruginosa was also found as an implicated organism in 82% culture positive patients and its rate of infection was significantly higher as compared to the other organisms in CF patients. The colonization with P. aeruginosa has been linked with pulmonary deterioration of CF patient resulting in an increased risk of hospitalization [21]. In the study, all the patients (n = 9) who required hospitalization were chronically infected with P. aeruginosa.
Besides morbidity, P. aeruginosa has been associated with increased mortality [18-22]. In the study all the 4 patients who died were chronically infected with P. aeruginosa. Thus significant higher rates of hospitalization and death were observed among patients chronically infected with P. aeruginosa.
Although prolonged and frequent administration of antibiotics has been useful in the treatment of pulmonary exacerbations, there is risk of development of multiple drug resistance in P. aeruginosa strains. In the study, 41% isolates were multidrug resistant and were recovered from chronically infected patients, indicating the possible role of antibiotic pressure. Similar findings have been reported elsewhere [23-25].
All three patients infected with ESBL positive isolates were found sensitive to meropenam and imipenam. The similar finding has been reported earlier where carbapenams have been found the most effective and reliable against ESBL positive isolates [26].
During chronic infection with the organism, there is progressive anatomical deterioration of the CF lung, which provides a highly spatial structured environment leading to diversification of the organism into six different colony morphological types-Type1-Type 6[12]. This diversification enhances the capacity of an organism to survive in lower respiratory tract of CF patients [13]. In the present study, 7/8(87.5%) chronically infected patients harbored different colony morphological types in the sputum samples.
Once P. aeruginosa has colonized the respiratory passages of CF patients, it leads to the induction of mucoidity, resulting in persistent infections. In the study, 6/8(75%) patients who were chronically infected with P. aeruginosa harbored mucoid P. aeruginosa in their respiratory samples. Our finding in agreement with the earlier reports where an increased association of mucoidity with chronicity of infection was observed [4,27,28]. Out of 4 patients who died, 3(75%) harbored mucoid colonies in their respiratory samples signifying the association of mucoidy with high mortality in CF patients. Our finding was in accordance with the earlier study where association of mucoidity with high risk of death was observed [26].
The antimicrobial susceptibility among mucoid and nonmucoid isolates was compared. The mucoid isolates showed increased resistance against gentamicin and lower resistance against beta lactam (7.4% to piperacillin and no resistance to ceftazidime). The results were analogy to the studies where high resistance to aminoglycosides [29,30] and low resistance to beta lactam [31,32] was observed among mucoid isolates.
With chronicity of infection, there is an increase in concentration of amino acids in bronchial secretions, which is due to active epithelial transport, plasma exudation and protease activity in lung tissue and surface fluid. This results in development of auxotrophy for certain amino acids required for growth. In the present study, auxotrophic P. aeruginosa was found in 2/27(7.4%) patients in contrary to an earlier report where auxotrophy of P. aeruginosa was found in 86% patients infected with the organism [33]. The reason for low percentage of auxotrophs among our patients may be due to lower number of chronic patients in this study.
Furthermore in the study, 3/5 auxotrophs had arginine requirement and the remaining 2 were auxotrophic for methionine. This was in agreement with the earlier reports where methionine and arginine were considered as the most commonly required amino acids by auxotrophs [14,33,34]. As these strains could be more resistant to antibiotics, it is important to study auxotrophy of isolates in CF patients with severe P. aeruginosa infection in order to plan out appropriate therapy for the management of CF patients.
Bacterial typing schemes are based on phenotypic or genotypic analysis of multiple isolates within a particular species. Such analysis helps in investigation of outbreaks and determines the recurrence or relapse of an infection. The epidemiology of P. aeruginosa infection in CF patients has been hampered since the conventional phenotyping methods-phage and pyocin typing have poor discriminatory power and poor reproducibility respectively. Genetic typing methods have shown to be more discriminatory than phenotypic methods for typing P. aeruginosa CF isolates [35]. Various genotypic typing methods- restriction fragment length polymorphism (RFLP) analysis of tox A I and pil A loci, pulse field gel electrophoresis (PFGE), random amplified DNA polymorphism (RAPD) PCR have been used to identify the number of strains colonizing the CF patient [13].
PCR-ribotyping is based on the uniform spacer region in rRNA operons of P. aeruginosa [36] resulting in low degree of heterogeneity of rRNA operons. In the study, majority of isolates showed the single PCR-ribotyping Pattern P1, probably due to the above reason. Hence for the study of dynamic of P. aeruginosa strain variation, ERIC-PCR may be a useful tool.
By ERIC-PCR, all 27 culture positive patients except one were found colonized with the unique genotype (E1-E33). This indicates that cross-colonization among unrelated CF patients was uncommon. The results were concordant to the earlier studies where cross-infection between unrelated patients was considered unusual [37,38]. All 27 patients except 4 were colonized with the single persistent genotype. These 4 patients were chronically infected and were colonized with 2 or more genotypes. Our results were in agreement with the previous studies where majority of CF patients were colonized with a predominant genotype for longer period of time but high degree of genotypic variability within patients have been observed suggesting that the colonizing strain may occasionally be replaced [13,39-41].
Thus we have found that CF is a problem in our country but delayed diagnosis and low index of suspicion have delayed the appropriate management of these patients. Although all the patients attending the Pediatric chest clinic were included in the study, the number of patients, especially the chronic ones, was limited in India, which further limited the number of chronic patients. Although the small number of patients in our study is not sufficient to answer all the epidemiological questions, these results indicate that a larger scale CF study should be conducted in India.
Conclusion
P. aeruginosa is the major colonizer of the respiratory passages of CF patients. The patients with chronic infection showed significantly higher rates of hospitalization and death. With chronicity of infection, the organism undergoes phenotypic and genotypic variations. Among phenotypic variations, the organism becomes multidrug resistant, diversifies its colony morphology, acquires mucoidity and becomes auxotrophic for various amino acids. The multiple genotypes have been found to colonize the respiratory passages of CF patients.
Thus in a particular clinical set up, high index of suspicion should be there for diagnosis of CF patients so as to prevent the delay in diagnosis and management of CF patients. The routine microbiological examination of respiratory samples from CF patients are advised to study the characteristics of the infecting agents of which P. aeruginosa is the most common.
Methods
Subjects
All patients with CF (age equal to or less than 15 years) attending the Pediatric Chest Clinic at All India Institute Of Medical Sciences between August 1998-August 2001 were included in the study. There was a regular monthly follow-up of each patient to the clinic. Patients who required hospitalization were admitted to the Pediatric ward and were followed daily. The diagnosis of CF was based on both clinical and laboratory parameters [42]. For analysis, patients were divided into 2 groups -Group 1 and Group 2 based on the chronicity of infection with P. aeruginosa [12]. Patients under Group 1 were chronically infected with P. aeruginosa for the past two years and Group 2 included patients who were intermittently colonized with the organism.
Sample collection
During each visit of a patient to the clinic and during the period of hospitalization, a representative sample – sputum and/or throat swab after physiotherapy was collected [41]. The sputum sample was collected directly in a sterile vial after coughing. Throat swab was considered as an alternative respiratory sample and was collected from children who were either below 2 years of age or were not able to expectorate sputum [43].
Isolation and characterization of P. aeruginosa strains
The sputum was vortexed (for complete homogenization) and cultured on Chocolate agar, Blood agar containing 5% sheep blood and Mac Conkey agar (Hi Media, Mumbai). The plates were incubated at 37°C for 18 hrs. P. aeruginosa was identified by standard bacteriological methods [44] and its different colony morphotypes were recorded [12].
Antibiotic susceptibility
Antibiogram
The single representative from each colony morphotype was taken and tested for antimicrobial susceptibility by agar dilution method as per NCCLS guidelines [45]. Of the six colony morphotypes, Type 5 was the mucoid morphotype and the rest morphotypes(Type 1–4 and Type 6) were the nonmucoid morphotypes. Depending on antimicrobial susceptibility of an organism, antibiograms were constructed where intermediate and resistant categories were clubbed. The antibiotics tested were gentamicin, ciprofloxacin, piperacillin and ceftazidime. All these antibiotics were procured from Hi Media, Mumbai. P. aeruginosa ATCC 27853 and Escherchia coli ATCC 25922(ATCC, USA) served as the control strains. Those isolates, which showed resistance to two or more than two antibiotics were considered multidrug resistant [10].
Production of Extended spectrum beta lactamase (ESBL)
Those isolates, which showed resistance to ceftazidime (CAZ) were tested for ESBL production by Disc Diffusion according to NCCLS guidelines [46] and by Etest method (AB BIODISK, Sweden). All ESBL positive isolates were further tested for carbapenam-imipenam and carbapenam susceptibilty by disk diffusion method [46].
Phenotypic characterization of P. aeruginosa isolates
Different phenotypic features of P. aeruginosa were studied in CF patients. The phenotypic features included colony morphotypes and auxotrophy.
Colony morphotypes
Based on colony morphology of the organism, there are six different colony morphotypes of P. aeruginosa [12]. The presence of six different morphotypes was observed in respiratory samples of CF patients. Type 5 morphotype showed mucoidy, which is one of the virulence factor of the organism.
Auxotrophy
All P. aeruginosa isolates from CF patients were tested for auxotrophy. The inoculum was adjusted to 0.5 Mc Farland and diluted to have the final concentration of 104cells/μl. One microlitre of the inoculum was spotted on Mueller Hinton agar (MHA) and Minimal agar medium (MAM) where MHA served as the complex medium and MAM as nutrient deficient medium. The plates were incubated at 37°C for 48 hrs. Those isolates, which did not grow on MAM plate but grew on MHA were considered auxotrophs. Those isolates, which were able to grow on both MHA and MAM plates were considered prototrophs. P. aeruginosa NCTC 50184 (meth -) and P. aeruginosa ATCC 27853 were included as the known auxotrophic and prototrophic strain respectively. The specific amino acid requirements were identified.
Identification of specific amino acid requirements
The stock solution of all 23 L-amino acids including cysteine, proline, lysine, arginine (Sigma, USA), cystine, methionine, serine (CDH), glutamine, glutamic acid, glycine, histidine, hydroxy-proline, isoleucine, leucine, ornithine, phenylalanine, alanine, asparagine, aspartic acid, threonine, tryptophan, tyrosine and valine (HiMedia, Mumbai) were prepared at the concentration of 2 mg/ml in either sterile distilled water or the appropriate solvent. All the amino acids were filter sterilized by 0.45 μm Millipore filter (Millipore, USA). The specific amino acid/acids requirement was determined by two methods [34].
The first method helps in detection of single or multiple amino acids. In this method, a single amino acid at the concentration of 20 μg/ml was added to each MAM plate. The plates were inoculated and incubated at 37°C for 48 hrs. The growth on specific amino acid containing plates indicated the requirement for single/multiple amino acids while no growth in any of the MAM plate supplemented with amino acid indicated either the requirement for two or more amino acids in combination or some other growth factor.
The second method helps in detection of combination of amino acids. All amino acids (at the concentration of 20 μg/ml) less one were added to molten MAM. The plates were inoculated and incubated as described above. The specific requirement for an amino acid/combination of amino acids was evident, if the isolate failed to grow on agar not containing that particular amino acid.
Antibiotic resistance
The antibiotic susceptibility of auxotrophs and prototrophs was determined by agar dilution method and the antibiotic resistance was compared between them. The antibiotics tested were gentamicin, ciprofloxacin, piperacillin and ceftazidime.
Molecular characterization of P. aeruginosa isolates
Molecular typing of P. aeruginosa isolates from all culture positive patients was done by ERIC-PCR and PCR-ribotyping. Both the molecular methods were carried out as described earlier [12].
Statistical analysis
To see the association between the categorical variables, Chi square test (with Yates correction)/Fischer's exact test was used. The result was considered significant at 5% level of significance (p < 0.05). SAS 8.0 statistical package was used for analysis.
Authors' contributions
GA: Technical procedures, study design and manuscript preparation.
AK: Infrastructure, time-to-time guidance and participated in microbiological experiments.
BKD: Involved in the molecular studies.
SKK: Helped in collection of the samples. Provided case histories of CF patients and helped in evaluation of clinical data.
SND: Performed the statistical analysis.
Acknowledgements
This work was supported by the Indian Council of Medical Research, New Delhi, INDIA
Figures and Tables
Figure 1 Detection of Extended spectrum β lactamase in P. aeruginosa isolate from a patient (180/95) by:A. Disc Diffusion method (Amika: Amikacin; Cipro: Ciprofloxacin; Pipra: Piperacillin; Genta: Gentamicin; CAZ: Ceftazidime). B. Etest method (AB Biodisk, Sweden).
Figure 2 Detection of Auxotrophy of P. aeruginosa. Isolate No: Patient PCC.No. 1: PCC.No. 499/2000; 2–5: PCC.No. 180/95, A: P. aeruginosa ATCC 27853; N: P. aeruginosa NCTC 50184(meth-). Plate A: Growth of P. aeruginosa on MAM supplemented with arginine only, Plate B: Growth of P. aeruginosa on MAM supplemented with all 22 amino acids Plate C: Growth of P. aeruginosa on MAM supplemented with all 22 amino acids except arginine,
Figure 3 ERIC-PCR patterns of P. aeruginosa isolates from the patient A. PCC. No. 180/95. • Lane M : 100 bp ladder, • Lane 1 : Pattern E1, • Lane 2 : Pattern E3, • Lane 3–5 : Pattern E1, • Lane 6–11 : Pattern E2. B. PCC.No. 180/95 and 179/93 • Lane M : 100 bp ladder, • Lane 1–7 (PCC.No.180/95) : Pattern E4, • Lane 8 (PCC. No. 179/93) : Pattern E5, • Lane 9 (PCC.No.180/95) : Pattern E4, • Lane 10 : P. aeruginosa ATCC 27853 • Lane 11 : Negative control
Figure 4 PCR-ribotyping of P. aeruginosa isolates from the patient A. PCC.No. 180/95 • Lane M : 100 bp ladder, • Lane1, 3, 7 and 9 : Pattern P1, • Lane 2, 4 and 8 : Pattern P2, • Lane 5, 10–14 : Pattern P3, • Lane 6 : Pattern P4B. PCC.No. 180/95 and 179/93 • Lane M : 100 bp ladder, • Lane 1,3 and10 (PCC.No.180/95) : Pattern P2, • Lane 2,4 (PCC.No.180/95) : Pattern P4, • Lane 5–7 and 9 (PCC.No.180/95) : Pattern P1, • Lane 8 (PCC.No.179/93 : Pattern P5, • Lane 11 : Negative control
Table 1 Antibiograms
Antibiogram Susceptibility to antibiotics Total isolates (%)
A1 S to all 102(29.1)
A2 R to Genta only 81(23.1)
A3 R to Genta+Cipro 85(24.3)
A4 R to Cipro only 24(6.9)
A5 R to Genta+Pipra 23(6.6)
A6 R to Genta+Cipro+CAZ 24(6.9)
A7 R to Genta+CAZ 5(1.4)
A8 R to Genta+Cipro+Pipra 6(1.7)
Genta: Gentamicin; Cipro: Ciprofloxacin; Pipra: Piperacillin; CAZ: Ceftazidime CAZ:
S: Sensitive; R: Resistant
+: Indicates "and"
Table 2 Distribution of eight antibiograms (A1-A8) among six colony morphological types of P. aeruginosa.
Colony Morphotype (N) Total Number of isolate (%) showing respective Antibiograms
A1 A2 A3 A4 A5 A6 A7 A8
Type 1 (194) 102 (52.5) 23 (11.8) 27 (10.3) 4 (1.5) 15 (7.7) 14 (7.2) 3 (1.5) 6(3)
Type2 (26) 5 (19.2) 0 10(38.5) 2(7.7) 1(3.8) 7(27) 1(3.8) 0
Type3 (23) 9 (39.1) 0 6(26.1) 5(21.7) 0 3(13.0) 0 0
Type4 (5) 2 (40) 0 2 (40) 1(20) 0 0 0 0
Type5 (94) 0 39 (41.5) 48 (51.1) 0 7(7.4) 0 0 0
Type6 (8) 0 5 (62.5) 1 (12.5) 0 0 1(12.5) 0 1(12.5)
N = Total Number of isolates
Table 3 Phenotypic and genotypic characteristics of P. aeruginosa isolated from cystic fibrosis patients (Total No. of patients infected with P. aeruginosa = 27)
S.No. PCC.No. Total isolates Phenotypic features Genotypic pattern
Colony morphotype Antibiogram ESBL Auxotrophy ERIC-PCR PCR-ribotyping
Group1
1 180/95 146 Type 1–6 A1-A4, A6-A8 + + E1-E4 P1-P4
2 179/93 59 Type 1,2,5 A1-A5, A7 - - E5-E7 P1, P5
3 174/97 22 Type 1 A1 - - E8 P1
4 308/98 22 Type1–3, 5 A1, A2, A5 - - E9 P1
5 10/99 22 Type1, 3 A1, A2, A4, A5 - - E10, E11 P1
6 160/98 19 Type1, 3,5 A3-A5, A7, A8 + - E12, E13 P1, P2
7 121/99 19 Type1–3, 5 A1, A2 - - E14 P1
8 308/00 4 Type 2,3,5 A1, A4 - - E15 P1
Group2
9 167/98 4 Type1 A1 - - E16 P1
10 250/99 4 Type1 A1 - - E17 P1
11 5/00 4 Type1 A1 - - E18 P1
12 256/99 3 Type1 A1, A2 - - E19 P1
13 337/98 3 Type1 A1 - - E20 P1
14 38/95 2 Type1 A1 - - E21 P1
15 202/98 2 Type1 A1, A3 - - E22 P1
16 312/99 2 Type1 A1 - - E23 P1
17 170/99 2 Type1 A1 - - E24 P1
18 499/00 2 Type1 A6 + + E1 P1
19 302/96 1 Type1 A1 - - E25 P1
20 95/96 1 Type1 A1 - - E26 P1
21 82/96 1 Type1 A1 - - E27 P1
22 8/97 1 Type1 A1 - - E28 P1
23 203/99 1 Type1 A1 - - E29 P1
24 133/99 1 Type1 A1 - - E30 P1
25 124/00 1 Type1 A1 - - E31 P1
26 452/00 1 Type1 A1 - - E32 P1
27 76/01 1 Type1 A1 - - E33 P1
Genta: Gentamicin; Cipro: Ciprofloxacin; Pipra: Piperacillin; CAZ: Ceftazidime CAZ: S: Sensitive; R: Resistant +: Indicates "and"
==== Refs
Ahuja AS Kabra SK Cystic fibrosis: Indian experience Ind Pediatr 2002 39 813 818
Singh M Prasad R Kumar L Cystic fibrosis in north Indian children Ind J Pediatr 2002 69 627 629
Kabra SK Kabra M Lodha R Shastri S Ghosh M pandey RM Kapil A Aggarwal G Kapoor V Clinical profile and frequency of delta 508 mutation in Indian children with cystic fibrosis Ind Pediatr 2003 40 612 619
Koch C Hoiby N Pathogenesis of cystic fibrosis Lancet 1993 341 1065 1069 7682274 10.1016/0140-6736(93)92422-P
Kulczycki LL Murphy TM Bellanti JA Pseudomonas colonization in cystic fibrosis JAMA 1978 240 30 34 96277 10.1001/jama.240.1.30
Friend PA Pulmonary infection in cystic fibrosis J Infect 1986 13 55 72 3090153
Govan JR Deretic V Microbial pathogenesis in cystic fibrosis:mucoid Pseudomonas aeruginosa and Burkholderia cepacia Microbiol Rev 1996 60 539 74 8840786
Mearns MB Hunt GH Rushworth R Bacterial flora of respiratory tract in patient with cystic fibrosis Arch Dis Child 1971 47 902 07
Cheng K Smith R Govan JRW Catherine D Winstanley Denning N Spread of beta lactam resistant Pseudomonas aeruginosa in a cystic fibrosis clinic Lancet 1996 348 639 42 8782753 10.1016/S0140-6736(96)05169-0
Saiman L Mehar F Niu W Neu HC Shaw KJ Millar G Prince A Antibiotic susceptibility of multiply resistant Pseudomonas aeruginosa isolated from CF patients, including transplant candidates Clin Infect Dis 1996 23 532 537 8879776
Oliver A Canton R Baquero F Blazquez J High frequency of Hypermutable P. aeruginosa in cystic fibrosis lung infection Science 2000 288 1251 1253 10818002 10.1126/science.288.5469.1251
Agarwal G Kapil A Kabra SK Chandra R Das B; Dwivedi SN Phenotypic and Genotypic variants of Pseudomonas aeruginosa isolated from Cystic Fibrosis Children in India Ind J Med Res 2002 116 73 81
Mahenthiralingam E Campbell ME Foster J Lam JS Speert DP Random Amplified DNA Typing of Pseudomonas aeruginosa Isolates Recovered from Patients with Cystic Fibrosis J Clin Microbiol 1996 34 1129 1135 8727889
Barth AL Pitt TL The high amino acid content of sputum from cystic fibrosis patients promotes the growth of auxotrophic P. aeruginosa J Med Microbiol 1996 45 110 119 8683546
Belkum AV Renders NHM 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 Immuno Med Microbiol 2000 27 51 57 10.1016/S0928-8244(99)00161-3
Cystic Fibrosis Foundation Patient Registry 1998 annual data repot 1999 Bethesda MD: Cystic fibrosis Foundation
Power CA Potter EM Wessel HU Lloyd Still JD Cystic fibrosis in Asian Indians Arch Pediatr Adolesc Med 1996 150 554 555 8620244
FitzSimmons S The changing epidemiology of cystic fibrosis J Pediatr 1993 122 1 9 8419592
Govan JRW Nelson JW Microbiology of lung infections in cystic fibrosis Br Med Bull 1992 48 912 930 1281036
Gilligan PH Microbiology of airway disease in patients with cystic fibrosis Clin Microbiol Rev 1991 4 35 51 1900735
Fick RB Jr Pathogenesis of Pseudomonas aeruginosa lung lesion in cystic fibrosis Chest 1989 96 158 164 2500307
Barth AL Pitt TL Microbial pathogens associated with Cystic fibrosis: special focus on P. aeruginosa Braz J Infect Dis 1998 2 43 61 11101911
Burns JL Ramsey BW Smith AL Clinical manifestations and treatment of pulmonary infections in cystic fibrosis Adv Pediatr Infect Dis 1993 8 53 66 8217005
Mouton JW Hollander JGD Horrevorts AM Emergence of antibiotic resistance amongst P. aeruginosa isolates from patients with cystic fibrosis J Antimicrob Chemother 1993 31 919 926 8360129
Ogle JW Reller LB Vasil ML Development of resistance in P. aeruginosa to imipenam, norfloxacin and ciprofloxacin during therapy: proof provided by typing with a DNA probe J Infect Dis 1988 157 743 8 3126247
Kfoury JNS Araj GF Recent developments in β lactamases and extended spectrum β lactamases BMJ 2003 327 1209 13 14630759 10.1136/bmj.327.7425.1209
Pederson SS Lung infection with alginate producing, mucoid Pseudomonas aeruginosa in cystic fibrosis APMIS 1992 28 1 79
Thomassen MJ Demko CA Boxerbaum B Stern RC Kuchenbrod PJ Multiple isolates of Pseudomonas aeruginosa with differing antimicrobial susceptibility patterns from patients with cystic fibrosis J Infect Dis 1979 140 873 80 120383
Govan JRW Fyfe JAM Mucoid Pseudomonas aeruginosa and cystic fibrosis: resitance of the mucoid form to carbenicillin, flucloxacillin, and tobramycin and the isolation of mucoid variants in vitro J Antimicrob Chemother 1978 4 233 240 97259
Demko CA Thomassen Effect of mucoid property on antibiotic susceptibility of P. aeruginosa Curr Microbiol 1980 4 69 73
Tannenbaum CS Hastie AT Higgins ML Kueppers F Weinbaum G Inability of purified Pseudomonas aeruginosa exopolysaccharide to bind selected antibiotics Antimicrob Agents Chemother 1984 25 673 5 6430224
Nicholas WW Evans MJ Slack MPE Walmsley HL The penetration of antibiotics into aggregrates of mucoid and nonmucoid P. aeruginosa J gen Microbiol 1989 135 1291 1303 2516117
Taylor RFH Hodson ME Pitt TL Adult cystic fibrosis association of acute pulmonary exacerbation and increasing severity of lung disease with auxotrophic mutants of P. aeruginosa Thorax 1993 48 1002 5 8256228
Barth AL Pitt TL Auxotrophic variants of Pseudomonas aeruginosa are selected from prototrophic wild type strains in respiratory infections in patients with cystic fibrosis J Clin Microbiol 1995 33 37 40 7699062
The International P. aeruginosa Typing Study Group A multicenter comparison of methods for typing strains of P. aeruginosa predominantly from patients with cystic fibrosis J Infect Dis 1994 169 134 142 7903973
Severino P Darini AL Magalhaes VD The discriminatory power of ribo-PCR compared to conventional ribotyping for epidemiological purposes APMIS 1999 107 1079 84 10660137
Speert DP Campbell ME Hospital epidemiology of Pseudomonass aeruginosa from patients with cystic fibrosis J Hosp Infect 1987 9 11 21 2880893 10.1016/0195-6701(87)90089-2
Speert DP Lawton D Damm S Communicability of Pseudomonas aeruginosa in a cystic fibrosis summer camp J Pediatr 1982 101 227 229 6808106
Renders NHM Sijmons MAF Van Belkum A Overbeek E Mouton JW Verbrugh HA Exchange of Pseudomonas aeruginosa strains among cystic fibrosis siblings Res Microbiol 1997 148 447 454 9765823 10.1016/S0923-2508(97)83875-2
Silbert S Barth AL Sadar HS Heterogeneity of P. aeruginosa in Brazilian Cystic Fibrosis Patients J Clin Microbiol 2001 39 3976 3981 11682517 10.1128/JCM.39.11.3976-3981.2001
Burns JL Gibson RL McNamara S Yuin D Emerson J Rosenfeld M Longitudinal assessment of Pseudomonas aeruginosa in young children with cystic fibrosis J Infect Dis 2001 183 444 52 11133376 10.1086/318075
Boat TF Wesh MJ Beaudet A Scriver CR, Beaudet AL, Sly WS, Valle D, Stansbury JB, Wyngaorden JB Cystic Fibrosis The metabolic basis of inherited diseases 1989 New York: Mc GrawHill 2649 2680
Kabra SK Alok A Kapil A Aggarwal G Kabra M Lodha R Can Throat Swab after physiotherapy replace sputum for identification of microbial pathogens in children with cystic fibrosis? Ind J Ped 2004 71 21 23
Govan JRW Pseudomonas, Stenotrophomonas, Burkholderia Practical Medical Microbiology 1996 14 Collee JG, Fraser AG, Marmion BP and Simmons A. New York: Churchill Livingstone 413 424
National Committee for Clinical Laboratory Standards MIC testing-supplement tables. Approved standard M100-S10(M7) 2000 National Committee for Clinical Laboratory Standards, Wayne, Pa
National Committee for Clinical Laboratory Standards Performance standards for antimicrobial disk susceptibility tests. Approved standard M2-A7. NCCLS Document M2-A7 2000 7 National Committee for Clinical Laboratory Standards, Wayne, Pa
|
16033658
|
PMC1183212
|
CC BY
|
2021-01-04 16:03:39
|
no
|
BMC Microbiol. 2005 Jul 21; 5:43
|
utf-8
|
BMC Microbiol
| 2,005 |
10.1186/1471-2180-5-43
|
oa_comm
|
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-201596976510.1186/1472-6947-5-20Study ProtocolAn overview of the design and methods for retrieving high-quality studies for clinical care Wilczynski Nancy L [email protected] Douglas [email protected] R Brian [email protected] Hedges Team [email protected] Health Information Research Unit, Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada2 Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada3 Michael G. DeGroote School of Medicine, McMaster University, Hamilton, Ontario, Canada2005 21 6 2005 5 20 20 14 12 2004 21 6 2005 Copyright © 2005 Wilczynski et al; licensee BioMed Central Ltd.2005Wilczynski 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
With the information explosion, the retrieval of the best clinical evidence from large, general purpose, bibliographic databases such as MEDLINE can be difficult. Both researchers conducting systematic reviews and clinicians faced with a patient care question are confronted with the daunting task of searching for the best medical literature in electronic databases. Many have advocated the use of search filters or "hedges" to assist with the searching process. The purpose of this report is to describe the design and methods of a study that set out to develop optimal search strategies for retrieving sound clinical studies of health disorders in large electronics databases.
Objective
To describe the design and methods of a study that set out to develop optimal search strategies for retrieving sound clinical studies of health disorders in large electronic databases.
Design
An analytic survey comparing hand searches of 170 journals in the year 2000 with retrievals from MEDLINE, EMBASE, CINAHL, and PsycINFO for candidate search terms and combinations. The sensitivity, specificity, precision, and accuracy of unique search terms and combinations of search terms were calculated.
Conclusion
A study design modeled after a diagnostic testing procedure with a gold standard (the hand search of the literature) and a test (the search terms) is an effective way of developing, testing, and validating search strategies for use in large electronic databases.
==== Body
Background
The Clinical Hedges Study was designed with the objective of developing optimal search strategies to improve the retrieval of clinically relevant and scientifically sound study reports from large, general purpose, biomedical research bibliographic databases including MEDLINE, EMBASE, CINAHL, and PsycINFO. The search strategies were developed to 1) assist health care providers to do their own searches; 2) help reviewers of published evidence concerning health care interventions retrieve all relevant citations; 3) provide resources for librarians to help health care providers construct their own searches; and 4) provide input to the database producers about their indexing processes and the organization of their databases.
Data for the Clinical Hedges Study was collected throughout the year 2000 and continued into the year 2001. Database construction and analyses are ongoing with some of our pre-study results (calibration of the hand search) appearing in conference proceedings as early as 2001 [1]. Search strategies developed for use in MEDLINE have been published [2-9] but the study design and methods have not been fully detailed. We are in the process of publishing the search strategies developed for use in EMBASE [10], CINAHL, and PsycINFO. The current report provides full disclosure of the design and methods of the Clinical Hedges Study, including many details that could not be accommodated in other reports, to facilitate replication studies and comparisons with other approaches to bibliographic information retrieval.
Methods/design
Study design including data organization and programming
The methods and design of the Clinical Hedges Study are outlined in this paper and are illustrated in the Figure. The study design used to address the above mentioned objective was an analytic survey comparing hand searches of 170 journals in the year 2000 with retrievals from MEDLINE (161 of the 170 were indexed in MEDLINE in the year 2000), EMBASE (135 were indexed in EMBASE), CINAHL (75 were indexed in CINAHL), and PsycINFO (64 were indexed in PsycINFO) through the Ovid web gateway for candidate search terms and combinations. Candidate search terms and combinations were tested in the four electronic databases by treating the search terms as "diagnostic tests" for sound studies. Due to the size of these four electronic databases, it is not feasible to determine the total number of relevant citations in each database for a given search. The hand search of the literature for the 170 journals circumvents this problem as it provides a "gold standard" for a segment of the literature in the electronic database file, and an approximation for the operating characteristics of search strategies (i.e., sensitivity, specificity, precision, and accuracy).
There are two primary sources of data in the Clinical Hedges Study. The first source of data was generated from a hand search of the literature (the "gold standard"). The second source of data was generated from downloads of search terms and citation information from the four electronic databases (the "test"). After these two data sources were obtained, we created a database that contained the matched merged content from these sources. This merged database was used for the development and validation of search strategies. In the following sections we outline this process.
Hand search data
To generate the hand search data six research assistants reviewed 170 journals titles for the year 2000. The 170 journal titles reviewed were chosen over several years in an iterative process based on hand search review of over 400 journals recommended by clinicians and librarians, Science Citation Index Impact Factors provided by the Institute for Scientific Information, recommendations by editors and publishers, and ongoing assessment of their yield of studies and reviews of scientific merit and clinical relevance. These journals include content for the disciplines of internal medicine (e.g., Annals of Internal Medicine), general medical practice (e.g., BMJ, JAMA, Lancet), mental health (e.g., Archives of General Psychiatry, British Journal of Psychiatry), and general nursing practice (e.g., Nursing Research) (list of journals provided by the authors upon request).
Each item (e.g., article, editorial, letter) in each issue of the 170 journals for the year 2000 was classified for article format (Table 1) and whether the content was of interest to human health care (Table 2). Original and review articles that were of interest to human health care were additionally classified for type of data presentation if a review article (Table 3), age of study participants (Table 4), purpose of the article (i.e., what question [s] is [are] the investigation addressing [Table 5]), and methodologic rigor for each of the purpose categories except for cost and qualitative studies and those articles classified as "something else" (Table 6). The methodologic criteria outlined in Table 6 are the same as those used for critically appraising articles for inclusion in 4 evidence-based medicine journals that our research group produced in 2000 (i.e., ACP Journal Club, Evidence-Based Medicine, Evidence-Based Nursing, and Evidence-Based Mental Health). Research staff were rigorously calibrated before the hand search of the 170 journal titles and inter-rater agreement for identifying the purpose of articles was 81% beyond chance (kappa statistic, 95% confidence interval [CI] 0.79–0.84). Inter-rater agreement for which articles met all scientific criteria was 89% (CI 0.78 to 0.99) beyond chance [1].
Table 1 Format categories
Format type Definition
Original study Any full text article in which the authors report first-hand observations.
Review article Any full text article that was bannered 'review, overview, or meta-analysis' in the title or in a section heading, or it was indicated in the text of the article that the intention was to review, summarize, or highlight the literature on a particular topic.
General article A general or philosophical discussion of a topic without original observation and without a statement that the purpose was to review a body of knowledge.
Case report An original study or report that presented only individualized data.
Table 2 Interest to human health care
Of interest Definition
Yes Concerned with the understanding of health care in humans; anything that will have an effect on the patient/subject.
No Not concerned with the understanding of health care in humans; anything that will not have an effect on the patient/subject (e.g., studies that describe the normal development of people; basic science; studies involving animals; gender and equality studies in the health profession; or studies looking at research methodology issues).
Table 3 Categories of data presentation in review articles
Type of data presentation Definition
Individual patient data Individual patient data was used in a meta-analysis.
Meta-analysis The reported summary data were pooled from relevant studies.
Overview A general discussion of the reviewed studies with no attempt to quantitatively combine the results.
Table 4 Age categories of ≥ 50% of study participants
Category Definition
Fetus Fetus
Newborn Birth to 1 month
Infant > 1 month to < 24 months
Preschool 2 years to < 6 years
Child 6 years to < 13 years
Adolescent 13 years to < 19 years
Adult 19 years to < 45 years
Middle age 45 years to < 65 years
Aged 65 years to < 80 years
Aged 80 ≥ 80 years
ND Age of study participants was non-discernible
Table 5 Purpose categories
Purpose type Definition
Etiology Content pertained directly to determining if there was an association between an exposure and a disease or condition. The question is "What causes people to get a disease or condition?"
Prognosis Content pertains directly to the prediction of the clinical course or the natural history of a disease or condition with the disease or condition existing at the beginning of the study.
Diagnosis Content pertains directly to using a tool to arrive at a diagnosis of a disease or condition.
Treatment Content pertains directly to an intervention for therapy (including adverse effects studies), prevention, rehabilitation, quality improvement, or continuing medical education.
Cost Content pertains directly to the costs or financing or economics of a health care issue.
Economics Content pertains directly to the economics of a health care issue.
Clinical Prediction Guide Content pertains directly to the prediction of some aspect of a disease or condition.
Qualitative Content relates to how people feel or experience certain situations, and data collection methods and analyses are appropriate for qualitative data.
Something Else Content of the study does not fit any of the above definitions.
Table 6 Methodologic rigor
Purpose category Methodologic rigor
Etiology Observations concerned with the relationship between exposures and putative clinical outcomes;
Data collection is prospective;
Clearly identified comparison group(s);
Blinding of observers of outcome to exposure.
Prognosis Inception cohort of individuals all initially free of the outcome of interest;
Follow-up of ≥ 80% of patients until the occurrence of a major study end point or to the end of the study;
Analysis consistent with study design.
Diagnosis Inclusion of a spectrum of participants;
Objective diagnostic ("gold") standard OR current clinical standard for diagnosis;
Participants received both the new test and some form of the diagnostic standard;
Interpretation of diagnostic standard without knowledge of test result and vice versa;
Analysis consistent with study design.
Treatment Random allocation of participants to comparison groups;
Outcome assessment of at least 80% of those entering the investigation accounted for in 1 major analysis at any given follow up assessment;
Analysis consistent with study design.
Economics Question is a comparison of alternatives;
Alternative services or activities compared on outcomes produced (effectiveness) and resources consumed (costs);
Evidence of effectiveness must be from a study of real patients that meets the above-noted criteria for diagnosis, treatment, quality improvement, or a systematic review article;
Effectiveness and cost estimates based on individual patient data (micro-economics);
Results presented in terms of the incremental or additional costs and outcomes of one intervention over another;
Sensitivity analysis if there is uncertainty.
Clinical Prediction Guide Guide is generated in one or more sets of real patients (training set);
Guide is validated in another set of real patients (test set).
Review articles Statement of the clinical topic;
Explicit statement of the inclusion and exclusion criteria;
Description of the methods;
≥ 1 article must meet the above noted criteria.
Hand search data were recorded on a paper based data collection form that was compatible with an optical mark and character recognition system called Teleform (Cardiff Software Publishing, Bozeman MT). The data collection form was designed using Teleform Designer. On each form data entry fields were available for the journal name, volume, issue and publication date, as well as a tabular data collection area to record the classification of individual items in the journal (e.g., article, letter, news item – 16 records per sheet). Multiple pages of the form could apply to one issue of a journal. Alphanumeric data fields were encoded in Teleform "comb" fields, where each hand printed block character had to be drawn within a fixed assigned space. A variety of strategies were used to maximize the reliability of the optical character/mark recognition abilities of Teleform. As many of the data fields as possible were either multiple choice or numeric. For fields containing text, research staff were trained in optimal letter formation and were required to use designated pre-tested pens. A mechanism was also established for correcting data entry errors.
All data collection forms were scanned with a Hewlett Packard 610cxi Scanjet scanner using Teleform Reader and the interpretation of the form was verified using Teleform Verifier. Scripts were written in a limited form of Visual Basic for Applications to perform basic validation on the incoming data while being interpreted by Teleform Reader. Data entry staff performed a data verification step by comparing the scanned image of the form alongside the Teleform interpretation. The staff were given the opportunity to accept the interpretation or to correct it. The original data collection sheets and/or the researcher who classified the material may have been consulted to make corrections. Tests to determine data entry error rates were conducted and an overall error rate of 0.01% was found.
After data entry and data verification using Teleform, hand search data were exported to a Microsoft (MS) Access database. The hand search data was split into two MS Access tables, one containing journal information and the other containing article information. The two tables were linked on a key field.
The hand search database volume was extensive with 11 data fields recorded for a collection of 60,330 articles in 170 publications.
On-line data
The data acquired from the on-line database were matching information (i.e., journal name, issue, and volume; first author's last name; title of the article; and first page number of the article), and the results of executing search terms in the on-line database. To generate the second source of data, the on-line data, it was necessary to construct a comprehensive set of search terms. We began a list of index terms and textwords for each of the four electronic databases, MEDLINE, EMBASE, CINAHL, and PsycINFO, and then sought input from clinicians and librarians in the United States and Canada through interviews of known searchers, requests at meetings and conferences, and requests to the National Library of Medicine. Individuals were asked what terms or phrases they used when searching for studies of causation, prognosis, diagnosis, treatment, economics, clinical prediction guides, reviews, costs, and of a qualitative nature when using these databases. For instance, for MEDLINE, terms could be from Medical Subject Headings (MeSH), including publication types (pt), and subheadings (sh), or could be textwords (tw) denoting methodology in titles and abstracts of articles. We compiled a list of 5,345 terms for MEDLINE of which 4,862 were unique and 3,870 returned results (list of terms tested provided by the authors upon request). For EMBASE we compiled a list of 5,385 terms of which 4,843 were unique and 3,524 returned results (list of terms tested provided by the authors upon request). For CINAHL we compiled a list of 5,020 unique search terms of which 3,110 returned results (list of terms tested provided by the authors upon request). For PsycINFO we compiled a list of 4,985 unique search terms of which 2,583 returned results (list of terms tested provided by the authors upon request). Index terms varied by electronic database whereas the same list of textwords were tested in each of the electronic databases.
Since the primary goal of the Clinical Hedges Study was to deliver search strategies which could be used by clinicians and researchers to locate the best quality published research specific to their interests it was required that the data be gathered via the same user interface that end users would use. Thus, raw performance data for individual search terms were downloaded via Ovid. Because of the volume of terms and their combinations, we automated the submission of terms using a telnet connection.
Ovid provides a simulated Graphical User Interface (GUI) through telnet using a simulated Dec VT-100 terminal interface. To handle a series of exchanges an open loop automation scheme was used. This consisted of a script-reading program written in Visual Basic for Applications, which passed keystroke data that mimicked what a user would enter into a commercially available telnet program. The script reader retrieved specific details from internally derived reference tables such as journal names (the 170 journals that were hand searched) and search terms (text of all search terms compiled for each of the four electronic databases) and inserted this information into the script through a parameter substitution scheme. To retrieve search term data for the Clinical Hedges Study, we "ANDed" each search term with a strategy saved on Ovid, which comprised our reading list of 170 journals published in the year 2000.
We used MS Outlook to recover the requested data via e-mail. Filters and dedicated "pst" files were set up to handle e-mail retrieval from a research project e-mail account. An "autosave" program saved the e-mail messages in individual text files for automated processing. A "downloads" program read the saved text files and entered the data into the appropriate MS Access tables. These scripts were stored in an MS Access database along with the program, and were organized in tables where keystroke sequences formed individual commands, which could be timed. The sequences of commands were grouped according to their function in the process of connecting to the on-line service (Ovid) or gathering data from it.
Matching hand search and on-line data
The approach to matching the hand search data records to the on-line data was undertaken in a two-stage process. First, a minimal set of information was retrieved from the on-line source, organized by journal title, which was used to link hand search data. At this stage, the linking software only attempted to match the hand-coded journal, volume, and issue information from the hand search data to a similarly hand-coded field retrieved from the on-line source. Later, a more complete set of matching information was retrieved from the on-line source, including a unique identifier for the index journals, article titles, authors, abstracts, indexing terms, and publication types. These data were organized by journal and index.
The matching algorithm was initially conservative requiring 100% certainty to establish a match. An "unmatched" report was subsequently generated which was used to refine the algorithm. For each of the four electronic databases approximately 95% of the records could be matched through an automated process. The remaining unmatched records were processed manually.
It should be noted here that there was not a one-to-one relationship between individual items entered in the hand search database and items recovered from the on-line database because an individual article could be determined to serve more than one purpose in the context of the clinical HEDGES study. Thus, an article could be a "review" (one format) about "diagnosis" and "therapy", which were additional purposes.
After extensive attempts, a small fraction of the hand-search items failed to be matched to citations in each of the four electronic databases and a small number of citations downloaded from each of the four electronic databases failed to be match to the hand-search data. As a conservative approach, unmatched citations that were detected by a given search strategy were included in cell 'b' of the analysis table (Table 7) leading to slight underestimates of the precision, specificity, and accuracy of the search strategy. Similarly, unmatched citations that were not detected by a search strategy were included in cell 'd' of the table (Table 7), leading to slight overestimates of the specificity and accuracy of the strategy.
Table 7 Formulae for calculating the sensitivity, specificity, precision, and accuracy of searches for detecting sound clinical studies
Manual Review (Hand search)
Meets Criteria Does Not Meet Criteria
Search Terms Detected a b
Not detected c d
a + c b + d
a = true positives, articles found by the search term meet the criteria for purpose category (e.g., treatment) and methodologic rigor (i.e., "pass")
b = false positives, articles found by the search term do not meet the criteria for purpose category and methodologic rigor
c = false negatives, articles not found by the search term but did meet the criteria for purpose category and methodologic rigor
d = true negatives, articles not found by the search term and did not meet the criteria for purpose category and methodologic rigor
Sensitivity = a/(a+c); Precision = a/(a+b); Specificity = d/(b+d);
Accuracy = (a+d)/(a+b+c+d).
All articles classified during the manual review of the literature, n = (a+b+c+d).
Computations
After the hand search and on-line data were matched the merged data file was prepared for deriving computations. At this point, other than the unique identifiers from both data sources (hand search and on-line data), none of the other matching information (journal name, journal issue, etc) was relevant to the computations. This extraneous data was, therefore, removed from the tables that were used to compute the performance of the search terms, as this information was already stored in a journal table within MS Access.
We determined the sensitivity, specificity, precision, and accuracy of each single term and combinations of terms using an automated process. The formulae for calculating these statistics are shown in Table 7. Sensitivity for a given topic (e.g., articles that are classified as original treatment studies that "pass" methodologic rigor) is defined as the proportion of high quality articles for that topic that are retrieved; specificity is the proportion of low quality articles not retrieved; precision is the proportion of retrieved articles that are of high quality; and accuracy is the proportion of all articles that are correctly classified.
Once the performance parameters of individual search terms were computed, it was possible to select individual terms for the construction of search strategies. In the Clinical Hedges Study, we had a collection of over 4,800 unique search terms for MEDLINE and EMBASE, with up to 1,200 of them returning results for a particular purpose category (e.g., treatment, prognosis). A rather simple estimation of computation time required, based on single term calculations and some preliminary two-term calculation runs, indicated that the time to compute search strategies using all of the available terms and an arbitrary limit on the number of terms in combinations, could amount to literally hundreds of years of computing time with equipment available at the time. Many of the combinations of search terms could be predicted to perform poorly as they contained individual terms that had poor performance in terms of returning very few true positives, or returning far too many false positives (for definition of terms see Table 7). Thus, for MEDLINE and EMBASE, only individual search terms with sensitivity > 25% and specificity > 75% for a given purpose category were incorporated into the development of search strategies that included 2 or more terms. All combinations of terms used the Boolean OR, for example, "predict.tw. OR survival.sh.". For the development of multiple-term search strategies to either optimize sensitivity or specificity, we tested all 2-term search strategies with sensitivity at least 75% and specificity at least 50%. For optimizing accuracy, 2-term search strategies with accuracy > 75% were considered for multiple-term development. These criteria were relaxed somewhat for CINAHL and PsycINFO since the number of terms returning results were fewer. Individual search terms with sensitivity ≥ 10% and specificity ≥ 10% for a given purpose category were incorporated into the development of search strategies that included 2 or more terms. All possible "ORed" two-and three-term combination of terms for each of the purpose categories were derived and tested through an automated iterative process. Our target through combining search terms was to optimize each of sensitivity, specificity, and accuracy.
In addition to developing search strategies using the Boolean approach described above, we also evaluated the potential for improving performance using logistic regression. Two approaches were taken. First, we took the top performing Boolean search strategies and ORed additional terms to these base strategies using stepwise logistic regression. The level of significance for entering and removing search terms from the model was 0.05. Adding terms to the model stopped when the increase in the area under the ROC curve was < 1%. Second, we developed search strategies from scratch with stepwise logistic regression using these same cut off values. Both logistic regression approaches were compared with the Boolean approach to search strategy development when developing strategies for treatment articles and prognostic articles for MEDLINE. Treatment and prognosis were chosen because they represented the best and the worst cases for MEDLINE search strategy performance. For both purpose categories the logistic regression approaches to developing search strategies did not improve performance compared with search strategies developed using the Boolean approach described above. We also found that when strategies were developed in 60% of the database and validated in the remaining 40% (a random allocation method was used to assign individual items to the development or validation datasets) there were no statistical differences in performance. Thus, for subsequent purpose categories and databases, the Boolean approach was used for search strategy development and search strategies were developed using all records in the database.
Search strategies developed for use in MEDLINE have been translated for use in PubMed by staff of the National Library of Medicine, and compared for performance by the senior author (RBH).
Discussion
A study design modeled after a diagnostic testing procedure with a gold standard (the hand search of the literature) and a test (the search terms) is an effective way of developing, testing, and validating search strategies for use in large electronic databases.
Additional research is underway in search strategy development including testing the strategies developed through this research, when combined with disease content terms, and when combined with terms using the Boolean "AND" and/or "NOT".
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
NLW and RBH prepared grant submissions in relation to this project and supplied intellectual content to the collection and analysis of the data. NLW participated in the data collection and all authors were involved in data analysis. All authors drafted, commented on and approved the final manuscript.
The Hedges Team
Includes Angela Eady, Brian Haynes, Chris Cotoi, Susan Marks, Ann McKibbon, Doug Morgan, Cindy Walker-Dilks, Stephen Walter, Stephen Werre, Nancy Wilczynski, and Sharon Wong.
Figure 1 Steps in data collection.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This research was funded by the National Library of Medicine, USA.
==== Refs
Wilczynski NL McKibbon KA Haynes RB Enhancing retrieval of best evidence for health care from bibliographic databases: calibration of the hand search of the literature Medinfo 2001 10 390 393 11604770
Wilczynski NL Haynes RB Developing optimal search strategies for detecting clinically sound causation studies in MEDLINE Proc AMIA Symp 2003 719 23 14728267
Wong SS Wilczynski NL Haynes RB Ramkissoonsingh R Developing optimal search strategies for detecting sound clinical prediction studies in MEDLINE Proc AMIA Symp 2003 728 32 14728269
Haynes RB Wilczynski NL Optimal search strategies for retrieving scientifically strong studies of diagnosis from MEDLINE: analytical survey BMJ 2004 328 1040 1042 15073027 10.1136/bmj.38068.557998.EE
Wilczynski NL Haynes RB The Hedges Team. Developing optimal search strategies for detecting clinically sound prognostic studies in MEDLINE BMC Medicine 2004 2 23 5 pages 15189561 10.1186/1741-7015-2-23
Wong SS Wilczynski NL Haynes RB Developing Optimal Search Strategies for Detecting Clinically Relevant Qualitative Studies in MEDLINE Medinfo 2004 311 316 15360825
Montori VM Wilczynski NL Morgan D Haynes RB Hedges Team Optimal search strategies for retrieving systematic reviews from MEDLINE: analytical survey BMJ 2005 330 68 6 pages 15619601 10.1136/bmj.38336.804167.47
Haynes RB McKibbon KA Wilczynski NL Walters SD Werre SR Hedges Team Optimal search strategies for retrieving scientifically strong studies of treatment from MEDLINE: an analytical survey BMJ 2005 330 1179 6 pages 15894554 10.1136/bmj.38446.498542.8F
Wilczynski NL Haynes RB Lavis JN Ramkissoonsingh R Arnold-Oatley AE HSR Hedges team Optimal search strategies for detecting health services research studies in MEDLINE CMAJ 2004 171 1179 1185 15534310
Haynes RB Kastner M Wilczynski NL Hedges Team Developing optimal search strategies for detecting clinically sound and relevant causation studies in EMBASE BMC Med Inform Decis Mak 2005 5 8 7 pages 15784134 10.1186/1472-6947-5-8
|
15969765
|
PMC1183213
|
CC BY
|
2021-01-04 23:52:11
|
no
|
BMC Med Inform Decis Mak. 2005 Jun 21; 5:20
|
utf-8
|
BMC Med Inform Decis Mak
| 2,005 |
10.1186/1472-6947-5-20
|
oa_comm
|
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-231604278910.1186/1471-2288-5-23Research ArticleA simplified search strategy for identifying randomised controlled trials for systematic reviews of health care interventions: a comparison with more exhaustive strategies Royle Pamela [email protected] Norman [email protected] Department of Public Health, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, Scotland2005 23 7 2005 5 23 23 5 2 2005 23 7 2005 Copyright © 2005 Royle and Waugh; licensee BioMed Central Ltd.2005Royle and Waugh; 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 is generally believed that exhaustive searches of bibliographic databases are needed for systematic reviews of health care interventions. The CENTRAL database of controlled trials (RCTs) has been built up by exhaustive searching. The CONSORT statement aims to encourage better reporting, and hence indexing, of RCTs. Our aim was to assess whether developments in the CENTRAL database, and the CONSORT statement, mean that a simplified RCT search strategy for identifying RCTs now suffices for systematic reviews of health care interventions.
Methods
RCTs used in the Cochrane reviews were identified. A brief RCT search strategy (BRSS), consisting of a search of CENTRAL, and then for variants of the word random across all fields (random$.af.) in MEDLINE and EMBASE, was devised and run. Any trials included in the meta-analyses, but missed by the BRSS, were identified. The meta-analyses were then re-run, with and without the missed RCTs, and the differences quantified. The proportion of trials with variants of the word random in the title or abstract was calculated for each year. The number of RCTs retrieved by searching with "random$.af." was compared to the highly sensitive search strategy (HSSS).
Results
The BRSS had a sensitivity of 94%. It found all journal RCTs in 47 of the 57 reviews. The missing RCTs made some significant differences to a small proportion of the total outcomes in only five reviews, but no important differences in conclusions resulted. In the post-CONSORT years, 1997–2003, the percentage of RCTs with random in the title or abstract was 85%, a mean increase of 17% compared to the seven years pre-CONSORT (95% CI, 8.3% to 25.9%). The search using random$.af. reduced the MEDLINE retrieval by 84%, compared to the HSSS, thereby reducing the workload of checking retrievals.
Conclusion
A brief RCT search strategy is now sufficient to locate RCTs for systematic reviews in most cases. Exhaustive searching is no longer cost-effective, because in effect it has already been done for CENTRAL.
==== Body
Background
Literature searching for systematic reviews of interventions in health care has been largely based on finding all randomized controlled trials (RCTs), as this study design is considered the gold standard. However RCTs make up only a very small proportion of all the articles included in bibliographic databases, and so the problem for systematic reviewers has been to devise a search strategy which is sensitive enough to find all the RCTs, but specific enough not to bury them in a large number of other unwanted retrievals needing to be manually excluded.
Two developments have facilitated searching for RCTs. The first is CENTRAL (the Cochrane Central Register of Controlled Trials), the world's most comprehensive database consisting exclusively of controlled clinical trials. It currently contains over 425,000 citations. The majority of the trials in CENTRAL have been identified through systematic searches of MEDLINE and EMBASE. The first two phases of a three phase highly sensitive search strategy (HSSS) have been used to search MEDLINE [1]. CENTRAL also includes citations to reports of controlled trials that are not indexed in MEDLINE or EMBASE, derived through searches of other bibliographic databases, and extensive hand searching.
Because identification has relied solely on the titles and, where available, the abstracts, some relevant trials may not have been identified. Therefore, the Cochrane Handbook says "it may still be worthwhile for reviewers to search MEDLINE using the Cochrane highly sensitive search strategy and to obtain and check the full reports of possibly relevant citations" [2]. However, this strategy is complicated to run, and may require time-consuming screening of abstracts, and perhaps of full articles.
The second development is the CONSORT statement, introduced in 1996 [3], and since revised [4]. CONSORT comprises a 22 item checklist and a flow diagram to help improve the quality of reports of RCTs, and has been endorsed by prominent medical journals There is evidence from a comparative before-and-after evaluation that there has been an increase over time in the number of CONSORT checklist items included in the reports of RCTs [5].
Item 1 on the CONSORT checklist recommends that the method in which participants were allocated to interventions (e.g., "random allocation", "randomized" or "randomly assigned") is described in the title and abstract. This allows instant identification of RCTs, and should help ensure that a study is appropriately indexed as an RCT in bibliographic databases.
As the Cochrane Collaboration has already done exhaustive work to ensure that CENTRAL is as complete as possible, we wanted to examine whether this eliminates the need for individual reviewers to run the HSSS, and the effectiveness of replacing this with a simplified search strategy.
The aims were
1. To determine the effect on the results of Cochrane reviews of using a brief RCT search strategy (BRSS).
2. To examine the change in use of variants of the word random in the title or abstract of journal articles, pre- and post-CONSORT.
Methods
All reviews new to the Cochrane Database of Systematic Reviews in the Cochrane Library 2004 issue 2 were selected. Those that stated that they were considering only RCTs, and which found at least one RCT, were selected.
All trials included in each review were identified from the section 'References to studies included in this review'. Each trial was checked to determine whether it was indexed in CENTRAL (on Cochrane Library 2004, issue 2), and then MEDLINE. If not in either of these databases, EMBASE was checked.
The full bibliographic records of all trials that were in either MEDLINE or EMBASE (using the OVID interface) were examined to determine whether random$ was in any field. (Random$.af. means a search of variants of random in all fields, where $ is truncation symbol).
The full papers of any trials that were: 1) not in CENTRAL, or 2) did not have random$ in any field in the bibliographic record, were obtained and checked to see whether they were actually RCTs. All non-English articles were translated, apart from those in Japanese and Chinese, as resources were not available.
The impact of omitting the RCTs not found with the BRSS was quantified, using Review Manager 4.2.7. The forest plots for the meta-analyses were reproduced, both with and without data from the missing trials, and the results compared to see if omission would make any important difference. The differences could theoretically include;
a) There might be less, or no data left, for some outcomes.
b) There could be a different result; no benefit over comparator, or vice versa.
c) There could be the same result, but with a different effect size.
d) There could be the same result and effect size, but with a wider confidence interval, and possibly loss of statistical significance.
In summary, the BRSS would involve: 1) searching CENTRAL, and 2) supplementing that with a search of MEDLINE and EMBASE, using a search of 'random$.af.' to pick up trials not in CENTRAL.
Results
There were 78 new reviews, of which 57 met the inclusion criteria. They cited a total of 920 trials; 79% (725) were journal articles. The remaining 21% were from the grey literature, 80% of these being conference abstracts).
Twenty one reviews were excluded from our study; 14 because they did not find any RCTs that met their inclusion criteria, and seven because they included other study designs in addition to RCTs.
Determination of the proportion of journal articles found using the BRSS
It was found that 93.3% (677) of the 725 journal articles included in the systematic reviews were in CENTRAL. It was assumed that these were all RCTs, but this was not checked due to the large numbers involved (and it was not relevant to the aims of the study). This left 48 journal articles not indexed in CENTRAL.
The full texts of all 48 of these articles (apart from the four in Japanese and five in Chinese) were translated and checked. It was found that 40 were RCTs, and the remaining eight used non-RCT study designs. (It was assumed that the untranslated articles were RCTs). Eleven of the 40 RCTs were found by searching MEDLINE or EMBASE with 'random$.af'. Therefore, 29 (4%) of RCTs would not have been found with the BRSS.
Details of the 29 RCTs not found by the BRSS
The 29 trials were distributed over 12 reviews. Ten were in MEDLINE, 11 in EMBASE only. Twelve were non-English language.
In two reviews [6,7], each with one missing trial, the data used in the meta-analyses were available in two other papers; both were in CENTRAL, and included in the reviews. Zhang 1990 [8] was confirmed (by authors of the review) to be the same trial as Chang 1996 [9]. The data in the Stensrud trial [10] was also reported in another paper by Stensrud [11].
Therefore, this leaves 27 missing journal articles, in 10 separate reviews that contain at least one trial with data not found by the BRSS. Table 1 shows the detail of these trials and the impact of excluding them from the reviews.
Table 1 Summary of impact of excluding trials not found with brief RCT search.
Topic No RCTs included Details of RCTs not found with BRSS Effect on review of excluding the RCTs not found
Lymphoedema of the limbs [20] 4 1. Kasseroller (1996) [21] Kasseroller study involved in 2 of 17 meta-analyses; exclusion makes no significant difference to outcome of either meta-analysis.
Ventricular pacemakers [22] 34 1. Davis (1985) [23] Davis study included in 1 of 12 meta-analyses; exclusion makes no significant difference.
Early intervention for psychosis [13] 7 1. Linszen (1998) trial comprised 5 papers; 2 not found with brief search.
a) Linszen (1998a) [24]
b) Linszen (1998b) [25] Linszen study included in 1 of 12 meta-analyses. Excluding Linszen would mean that there would be no data for this outcome.
Interventions for impetigo [15] 60 1. Arata (1989a) [26]
2. Arata (1989b) [27]
3. Bass (1997) [28]
4. Koranyi (1976) [29]
5. Moraes Barbosa (1986) [30]
6. Park (1993) [31]
7. Pruksachat. (1993) [32]
8. Sutton (1992) [33]
9. Tamayo (1991) [34] The 9 trials were included in 8 of the 19 meta-analyses in this review.
Outcome 1: excluding Sutton, Bass and Tamayo studies individually made no significant difference. The removal of all 3 together causes meta-analysis to lose statistical significance.
Outcome 2: excluding Bass, Park, Koranyi; no significant change
Outcome 3: excluding Park; no significant change
Outcome 4: excluding Arata 1989a; no significant change
Outcome 5: excluding Pruksachat.; no significant change
Outcome 6: excluding Arata 1989b; only trial providing data to this outcome, so no data now available.
Outcomes 7 & 8: Moraes-Barbosa 1986 is the only trial providing data for both outcomes, so no data now available.
Adherence to treatment in patients with high blood pressure [12] 38 1. Gabriel (1977) [35]
2. Hamilton (1993) [36]
3. Kerr (1985) [37]
4. Morisky (1985) [38]
5. Rehder (1980) [39] 58 different interventions were tested on 15519 patients. Review did not do meta-analysis due to heterogeneity between studies in terms of interventions and the methods used to measure adherence. Missing trials were all small studies of poor methodological quality. Their exclusion would have little impact on the final conclusions.
Preventing infection in nephrotic syndrome [14] 5 1. Dou (2000) [40]
2. Li (2000) [41]
3. Zhang (2000) [42]
4. Tong (1998) [43] The 4 missing trials were included in 3 of 4 of the meta-analyses.
Outcome 1: Excluding Dou and Tong; no significant change.
Outcome 2: Tong was the only trial providing data, so no data now available.
Outcome 3: Zhang was the only trial providing data, so no data now available.
Outcome 4: Li was the only trial providing data, so no data now available.
Probiotics for treating infectious diarrhoea [17] 25 Sugita (1994) [44] Sugita included in 7 of 27 meta-analyses in review. Exclusion of Sugita makes no significant difference to any of the clinical outcomes. The only outcome to lose statistical significance is not a clinical outcome, but a sensitivity analysis based on a methodological characteristic (blinding).
Psychological interventions for coronary heart disease [16] 56 1.Gallacher (1997) [45]
2. Mitsibounas (1992) [46] 1. Gallacher was included in 8 of 28 meta-analyses. In six, omission makes no difference.
In 2, confidence intervals are much wider; in the one of these statistical significance is lost. 2. Mitsibounas was included in 5 of 28 meta-analyses. Excluding Mitsibounas makes no significant difference in four outcomes. In the other, it cause causes the meta-analysis to lose statistical significance
Insulin analogues versus human insulin [47] 43 Iwamoto (2001) [48] Iwamoto in 4 of the 8 meta-analyses. Exclusion of Iwamoto makes no significant difference to any of the outcomes.
Tramadol for neuropathic pain [49] 5 Leppert (2001) [50] Leppert not used in either of 2 meta-analyses in the review. The review says: 'Given the small (even if unknown) number of subjects and the non-blinded nature of the trial it is probably not possible to draw and conclusions about their relative efficacy from this study.
Only one review did not do a meta-analysis[12] The nine remaining reviews did a total of 129 meta-analyses of various outcomes. In five out of the 10 reviews the missing trials made no significant difference, in that there were no clinically significant changes in effect size, nor any change in whether results were statistically significant or not. Hence, there was some difference in only five reviews.
The impact of the 'missing trials' being excluded from the reviews
Two consequences of missing data were found:
1) In three reviews [13-15], the missing trials were the only ones providing data for seven (of a total 35) outcomes, so no data were available for these seven outcomes. In a fourth review [16], omission of a trial would lose 92% of patients for two outcomes out of 28, but this did not change the significance or direction of the result.
2) In three reviews [15-17], the meta-analyses lost statistical significance due to the wider confidence intervals, for three of 74 outcomes, but one of these was not a clinical outcome.
Comparison of retrieval of HSSS and random$.af. in MEDLINE
For the period 1966 to October, 2004, the HSSS search strategy retrieved 2,505,742 records in MEDLINE, compared to 'random$.af.', which retrieved 399,208 records. Therefore, only 16% of the number of records were found using 'random$.af.', compared to HSSS.
CONSORT and the change over time in the proportion of RCTs with random$ in title or abstract
Using the 725 journal articles in our sample, we compared the frequency of random$ in the title or abstract between pre-CONSORT (published up to 1996) and post-CONSORT trials (published from 1997 onwards). Figure 1 shows the change in the proportion of RCTs with 'random$' in the title or abstract (with three year moving average trendline added).
Figure 1 Change in the proportion of RCTs with 'random$' in the title or abstract (with three year moving average trendline added).
In the post-CONSORT years, 1997–2003, the term appeared in 85% of the titles or abstracts, compared to 68% in the seven years before CONSORT (1990–96); mean difference was 17% (95% CI 8.3 to 25.9%; p = 0.001). If a longer pre-CONSORT period is used, 1980 to 1996, the proportion is similar at 62%.
Discussion
Using the BRSS, rather than the much more exhaustive HSSS, to retrieve RCTs in journal articles for Cochrane reviews, affected only a very small percentage of total outcomes of a few reviews, and made no important difference to the conclusions of these reviews. The most affected review had four of the five included trials published in Chinese [14].
The CONSORT statement appears to be associated with a significant increase in the frequency of random in the titles and abstracts of journal articles. Whether this is directly due to CONSORT, or whether CONSORT simply accelerated a pre-existing trend, and raised the general awareness amongst authors and editors for better reporting of RCTs, is uncertain. This improved description of RCTs by authors should mean that all RCTs are indexed with the appropriate publication type in MEDLINE, and also result in more sensitive retrieval of RCTs using the BRSS.
The BRSS had a sensitivity of 96% for RCTs in journal articles. Compared to the HSSS, the BRSS reduced the MEDLINE retrieval by 84%. This would represent a major time and cost saving in manual screening.
The strengths of this study include firstly that we used Cochrane reviews, as they approximate the 'gold standard' in searching, due to the requirement for exhaustive searches. Hence we can be fairly certain that we were starting with as comprehensive set of included trials as possible. Secondly, we quantified the impact of omitting trials not found with the simplified search.
A weakness of this study was that we could not check whether the RCTs would have been in CENTRAL at the time the searchers did their searching, and hence eliminate the possibility that some RCTs were first identified by the reviewers, and then passed to CENTRAL. However, given the extensive searching routinely done for CENTRAL, it is highly likely most RCTs would be identified sooner or later, and therefore be included in a subsequent update of the review.
A range of subject areas was included, which helps with generalisability, though it may decrease power in any one subject area. However, a recent study on identifying quality RCTs in pain relief gives general support to our findings [18]. It investigated the efficiency of the search strategy DBRCT.af., ("double-blind," "random," or variations of these terms) in MEDLINE and EMBASE, and was found have a sensitivity of 97%.
This study focused only on a simplified search strategy for retrieving RCTs in journal articles, since these made up the vast majority of the references used in Cochrane reviews, and are most important in terms of the quality and quantity of assessable data available. By contrast, most grey literature (the vast majority of which is meeting abstracts) gives very limited data. However, CENTRAL includes many grey literature trials, so these will be identified with the BRSS.
There are currently over 2200 Cochrane reviews, and these will need maintaining in the future. Authors may be encouraged to update their reviews, if they can be confident they can identify RCTs comprehensively with a simple search. The simplified search may also be useful for those doing reviews in a tight timescale, or by clinicians who just want a rapid but reliable answer to a question. There is a case for the 'not quite perfect but rapid, easy and almost complete' search'.
In practice, some of the few trials missed by the BRSS were small or of poor quality, and as Egger and colleagues have reported, the last few studies found by exhaustive searching could introduce bias, by being of poor quality [19].
The issue is whether the marginal benefits of exhaustive searching justify the extra costs. When the inclusion criteria demand only RCTS, this study suggests that exhaustive searching is now, in the era of CENTRAL and CONSORT, no longer cost-effective.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PR conceived and designed the study, and analysed the data. NW helped with interpretation and with drafting the paper. Both authors approved the final version
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Mark Deakin and Lynda Bain for help with data extraction.
==== Refs
Dickersin K Manheimer E Wieland S Robinson KA Lefebvre C McDonald S Development of the Cochrane Collaboration's CENTRAL Register of controlled clinical trials Eval Health Prof 2002 25 38 64 11868444 10.1177/0163278702025001004
Alderson P Green S Higgins JP editors Cochrane Reviewers' Handbook 4.2.2 [updated March 2004] The Cochrane Library, Issue 2, 2004; Section 5 2004 Chichester, UK, John Wiley & Sons, Ltd
Begg C Cho M Eastwood S Horton R Moher D Olkin I Pitkin R Rennie D Schulz KF Simel D Stroup DF Improving the quality of reporting of randomized controlled trials. The CONSORT statement JAMA 1996 276 637 639 8773637 10.1001/jama.276.8.637
Moher D Schulz KF Altman D CONSORT Group (Consolidated Standards of Reporting Trials) The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials.[see comment]. [Review] [30 refs] JAMA 2001 285 1987 1991 11308435 10.1001/jama.285.15.1987
Moher D Jones A Lepage L Use of the CONSORT statement and quality of reports of randomized trials: a comparative before-and-after evaluation JAMA 2001 285 1992 1995 11308436 10.1001/jama.285.15.1992
Linde K Rossnagel K Propranolol for migraine prophylaxis Cochrane Database Syst Rev 2004 2 CD003225 15106196
Badger C Preston N Seers K Mortimer P Benzo-pyrones for reducing and controlling lymphoedema of the limbs Cochrane Database Syst Rev 2004 2 CD003140 15106192
Zhang D Benzo-pyrones in the treatment of chronic lymphoedema of the arms and legs Zhonghua Yi Xue Za Zhi 1990 70 655 657 1963364
Chang TS Gan JL Fu KD Huang WY The use of 5,6 benzo-[alpha]-pyrone (coumarin) and heating by microwaves in the treatment of chronic lymphedema of the legs Lymphology 1996 29 106 111 8897354
Stensrud P Sjaastad O Comparative trial of Tenormin (atenol) and Inderal (propranolol) in migraine Headache 1980 20 204 207 6993420 10.1111/j.1526-4610.1980.h2004006.x
Stensrud P Sjaastad O Comparative trial of Tenormin (atenolol) and Inderal (propranolol) in migraine Ups J Med Sci Suppl 1980 31 37 40 7006182
Schroeder K Fahey T Ebrahim S Interventions for improving adherence to treatment in patients with high blood pressure in ambulatory settings Cochrane Database Syst Rev 2004 2 CD004804 15106262
Marshall M Lockwood A Early Intervention for psychosis Cochrane Database Syst Rev 2004 2 CD004718 15106257
Wu HM Tang JL Sha ZH Cao L Li YP Interventions for preventing infection in nephrotic syndrome Cochrane Database Syst Rev 2004 2 CD003964 15106229
Koning S Verhagen AP Suijlekom-Smit LW Morris A Butler CC van der Wouden JC Interventions for impetigo Cochrane Database Syst Rev 2004 2 CD003261 15106198
Rees K Bennett P West R Davey SG Ebrahim S Psychological interventions for coronary heart disease Cochrane Database Syst Rev 2004 2 CD002902 15106183
Allen SJ Okoko B Martinez E Gregorio G Dans LF Probiotics for treating infectious diarrhoea Cochrane Database Syst Rev 2004 2 CD003048 15106189
Chow TK To E Goodchild CS McNeil JJ A simple, fast, easy method to identify the evidence base in pain-relief research: validation of a computer search strategy used alone to identify quality randomized controlled trials Anesth Analg 2004 98 1557 65, table 15155305 10.1213/01.ANE.0000114071.78448.2D
Egger M Juni P Bartlett C Holenstein F Sterne J How important are comprehensive literature searches and the assessment of trial quality in systematic reviews? Empirical study Health Technol Assess 2003 7 1 76 12583822
Badger C Seers K Preston N Mortimer P Antibiotics / anti-inflammatories for reducing acute inflammatory episodes in lymphoedema of the limbs Cochrane Database Syst Rev 2004 CD003143 15106193
Kasseroller R Erysipelprophylaxe beim sekundaren lymphoedem mit selen Der Allgemeinartzt 1996 3 244 247
Dretzke J Toff WD Lip GY Raftery J Fry-Smith A Taylor R Dual chamber versus single chamber ventricular pacemakers for sick sinus syndrome and atrioventricular block Cochrane Database Syst Rev 2004 2 CD003710 15106214
Davis MJ Mundin HA Mews GC Cope GD Functional benefits of physiologic compared with ventricular pacing in complete heart block Clin Prog Electrophysiol Pacing 1985 3 457 460
Linszen D Dingemans PM Lenior ME Scholte WF Goldstein M Early family and individual interventions and relapse in recent-onset schizophrenia and related disorders Ital J Psychiatry Behav Sci 1998 8 77 84
Linszen DH Dingemans PMAJ Scholte WF Lenior ME Goldstein M Early recognition, intensive intervention and other protective and risk factors for psychotic relapse in patients with first psychotic episodes in schizophrenia Int Clin Psychopharmacol 1998 13 S7 S12
Arata J Kanzaki H Kanamoto A Okawara A Kato N Kumakiri M Shimizu T Ishibashi Y Nogita T Iozumi K Harada S Nakanishi H Hyang YS Okouchi H Takano S Ohara K Ota T Shishiba T Nakabayashi Y A double blind comparative study of cefdinir and cefaclor in skin and skin structure infections Chemotherapy 1989 37 1016 1042
Arata J Yamamoto Y Tamaki H Ookawara A Fukaya T Ishibashi Y Shimozuma M Iozumi T Kukita A Kimura Y Takahashi H Sasaki J Nishiwaki M Urushibata O Tomizawa T Eto H Kurihara S Morohashi M Seki T Double-blind study of lomefloxacin vs. norfloxacin in the treatment of skin and soft tissue infections Chemotherapy 1989 37 482 503
Bass JW Chan DS Creamer KM Thompson MW Malone FJ Becker TM Marks SN Comparison of oral cephalexin, topical mupirocin and topical bacitracin for treatment of impetigo Pediatr Infect Dis J 1997 16 708 710 9239775 10.1097/00006454-199707000-00013
Koranyi KI Burech DL Haynes RE Evaluation of bacitracin ointment in the treatment of impetigo Ohio State Medical Journal 1976 72 368 370 958658
Moraes-Barbosa AD Comparative study between topical 2% sodium fusidate and oral association of chloranphenicol/neomycin/bacitracin in the treatment of staphylococcic impetigo in new-born Arq Bras Med 1986 60 509 511
Park SW Wang HY Sung HS A study for the isolation of the causative organism, antimicrobial susceptibility tests and therapeutic aspects in patients with impetigo Korean J Dermatol 1993 31 312 319
Pruksachatkunakorn C Vaniyapongs T Pruksakorn S Impetigo: an assessment of etiology and appropriate therapy in infants and children J Med Assoc Thai 1993 76 222 229 8113643
Sutton JB Efficacy and acceptability of fusidic acid cream and mupirocin ointment in facial impetigo Curr Ther Res 1992 51 673 678
Tamayo L De la Luz OM Sosa de Martinez MC Rifamycin and mupirocin in the treatment of impetigo Dermatol Rev Mex 1991 35 99 103
Gabriel M Gagnon JP Bryan CK Improved patients compliance through use of a daily drug reminder chart Am J Public Health 1977 67 968 969 911007
Hamilton GA Roberts SJ Johnson JM Tropp JR Anthony-Odgren D Johnson BF Increasing adherence in patients with primary hypertension: an intervention Health Value 1993 17 3 11
Kerr JA Adherence and self-care Heart Lung 1985 14 24 31 3844002
Morisky DE DeMuth NM Field-Fass M Green LW Levine DM Evaluation of family health education to build social support for long-terms control of high blood pressure Health Education Quarterly 1985 12 35 50 3980239
Rehder TL McCoy LK Blackwell B Whitehead W Robinson A Improving medication compliance by counseling and special prescription container Am J Hosp Pharm 1980 37 379 385 7369220
Dou ZY Wang JY Liu YP Preventive efficiency of low-dose IVIgc on infection in nephrotic synrome Chin J Biologicals 2000 13 160
Li RH Peng ZP Wei YL Liu CH Clinical observation on Chinese medicinal herbs combined with predisone for reducing the risks of infection in children with nephrotic syndrome Inf J Chin Med 2000 7 60 61
Zhang YJ Wang Y Yang ZW Li XT Clinical investigation of thymosin for preventing infection in children with primary nephrotic syndrome Chin J Contemp Pediatr 2000 2 197 198
Tong LZ Mi LZ Preventive efficiency of IVIgG on secondary nosocomial in nephrotic syndrome Mod Rehabil 1998 2 236
Sugita T Togawa M Efficacy of Lactobacillus preparation biolactis powder in children with rotavirus enteritis Jpn J Pediatr 1994 47 2755 2762
Gallacher JE Hopkinson CA Bennett P Burr ML Elwood PC Effect of stress management on angina Psychol Health 1997 12 523 532
Mitsibounas DN Tsouma-Hadjis ED Rotas VR Sideris DA Effect of group psychosocial intervention on coronary risk factors Psychother Psychosom 1992 58 97 102 1484925
Siebenhofer A Plank J Berghold A Narath M Gfrerer R Pieber TR Short acting insulin analogues versus regular human insulin in patients with diabetes mellitus Cochrane Database Syst Rev 2004 2 CD003287 15106199
Iwamoto Y Akanuma Y Niimi H Sasaki N Tajima N Kawamori R Comparison between insulin aspart and soluble human insulin in type 1 diabetes (IDDM) patients treated with basal-bolus insulin therapy - Phase III clinical trial in Japan J Japan Diab Soc 2001 44 799 811
Duhmke RM Cornblath DD Hollingshead JR Tramadol for neuropathic pain Cochrane Database Syst Rev 2004 2 CD003726 15106216
Leppert W Analgesic efficacy and side effects of oral tramadol and morphine administered orally in the treatment of cancer pain Nowotwory 2001 51 257 266
|
16042789
|
PMC1183214
|
CC BY
|
2021-01-04 16:32:51
|
no
|
BMC Med Res Methodol. 2005 Jul 23; 5:23
|
utf-8
|
BMC Med Res Methodol
| 2,005 |
10.1186/1471-2288-5-23
|
oa_comm
|
==== Front
BMC NeurolBMC Neurology1471-2377BioMed Central London 1471-2377-5-111596703210.1186/1471-2377-5-11Research ArticleLack of evidence for a genetic association between FGF20 and Parkinson's disease in Finnish and Greek patients Clarimon Jordi [email protected] Georgia [email protected] Johanna [email protected] Vanesa [email protected]öm Olli [email protected] Euthimios [email protected] Terhi [email protected] Alexandros [email protected] George M [email protected] Pentti J [email protected] Andrew B [email protected] Molecular Genetics Unit, Laboratory of Neurogenetics. National Institute on Aging, National Institutes of Health, Building 35 Room 1A1000. Bethesda, MD 20892 USA2 Neurogenetics Unit, Department of Neurology, Medical School, University of Thessaly, Larissa, Greece3 Department of Neurology, Helsinki University Central Hospital and University of Helsinki Biomedicum-Helsinki, Neuroscience Programme, Finland4 Department of Neurology, Seinäjoki Central Hospital, Finland2005 20 6 2005 5 11 11 6 4 2005 20 6 2005 Copyright © 2005 Clarimon et al; licensee BioMed Central Ltd.2005Clarimon 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
Fibroblast growth factor 20 (FGF20) is a neurotrophic factor preferentially expressed in the substantia nigra of rat brain and could be involved in dopaminergic neurons survival. Recently, a strong genetic association has been found between FGF20 gene and the risk of suffering from Parkinson's disease (PD). Our aim was to replicate this association in two independent populations.
Methods
Allelic, genotypic, and haplotype frequencies of four biallelic polymorphisms were assessed in 151 sporadic PD cases and 186 controls from Greece, and 144 sporadic PD patients and 135 controls from Finland.
Results
No association was found in any of the populations studied.
Conclusion
Taken together, these findings suggest that common genetic variants in FGF20 are not a risk factor for PD in, at least, some European populations.
==== Body
Background
Parkinson's disease (PD), with a prevalence of nearly 2% in those older than 65 years of age, is one of the most common neurodegenerative disorders [1]. The cause of its clinical features (bradykinesia, rigidity, resting tremor, and postural instability) is the selective degeneration of dopaminergic neurons in the substantia nigra, leading to a deficiency of dopamine in their striatal projections areas. Although the etiology of PD is still unclear, increasing evidence suggest that genetic factors are likely to contribute [2,3]. To date, mutations in seven genes have been related to mendelian forms of the disease, and several loci across the genome have also been identified to be influencing not only familial, but also sporadic forms of PD [4-6]. One of the extant linkages to late onset PD is on chromosome 8p [7]. Analysis of candidate genes located within the region through the use of the pedigree disequilibrium test in 644 families, demonstrated a significant association between several polymorphisms at the fibroblast growth factor 20 (FGF20) and PD [8]. FGF20 is a neurotrophic factor widely expressed in dopaminergic neurons of the substantia nigra pars compacta [9], and enhances the survival of midbrain dopaminergic neurons [10].
We have attempted to replicate the previously described association by performing a case-control analysis in two different series: one from Finland and the other from Greece.
Methods
Patients and healthy individuals
The Finnish series is composed of 144 sporadic PD patients (mean age 67.2 years, range 38 to 88, 41% women) and 135 neurologically normal healthy control subjects (mean age 65.8 years, range 37 to 80, 64% women). Eleven patients had an onset before 45 years old. All individuals were recruited from the Neurological outpatient clinics of the Helsinki University Central Hospital and Seinäjoki Central Hospital. Patients were diagnosed according to PD Society Brain Bank criteria [11] and were either followed for at least 4 years or, alternatively, clinical follow-up for at least 2 years plus 123I-β-CIT-SPECT findings supporting idiopathic PD. Patients with dementia or patients who reported first-degree relatives with parkinsonism were excluded. Written informed consent was obtained from all participants. The Ethical Review Committee of the Department of Neurology at the Helsinki University Central Hospital approved the study.
The Greek series is composed of 151 sporadic PD patients (mean age 70.3, range 40 to 95, 46% women) and 186 age-, gender-and ethnicity-matched neurologically healthy controls. Three patients had an onset before 45 years old. All patients were residents of Thessaly (Central Greece) and were identified prospectively during a 3-year period (2001–2004) in the outpatient clinic for movement disorders in Larissa University Hospital and were followed up for at least one year and up to 3 years. The diagnosis of PD was based on established criteria [4].
After approval from Hospital Scientific Committee and informed consent, blood samples were drawn for DNA extraction from patients and controls.
Genotyping
Taqman® Assays-by-DesignSM SNP Genotyping (Applied Biosystems) based assays were employed for allelic discrimination of the two intronic SNPs (rs1989756 and rs1989754) as well as the 3'UTR SNP rs1721100. Thermal cycling and end-point PCR analysis was performed on an ABI PRISM® 7900HT Sequence Detection System and analyzed with SDS software (Applied Biosystems). Another SNP at the 3'UTR (rs20399075), which is located 99 base pairs (bp) from rs1721100, was genotyped using a restriction fragment length polymorphism (RFLP) approach. In brief, a PCR was conducted using the following primers: forward, 5'-AGTCTATTTTCTACTGAGAG-3'; and reverse, 5'-TAAGGACTAATCCAATGTATC-3'. The 139 bp amplified product was digested with 10 units of BspHI restriction enzyme and fragments were resolved on a 2% agarose gel stained with ethidium bromide. Two fragments of 80 bp and 59 bp appeared when a PCR fragment containing a cytosine in the polymorphic site was present.
Statistical analysis
Odds ratios and Chi square tests were calculated using SPSS v11.0 software for windows. Linkage disequilibrium was assessed by means of a genotype association test, implemented in the Arlqeuin ver.2.000 program, and haplotype frequencies were estimated with an expectation-maximization (EM) algorithm, also implemented in the Arlequin ver.2.000 program [12]. Comparison of FGF20 haplotype frequencies between cases and controls was performed using Fisher's exact test. Power calculations were performed with the epidemiological program EpiInfo version 6. Pomelo Tool software was used for multiple testing corrections.
Results
Distribution of genotype frequencies are presented in Table 1. All polymorphisms were in Hardy-Weinberg equilibrium for PD patients and controls in both the Greek and the Finnish series. No statistically significant differences in genotype frequency were found between cases and controls in the Greek series. Allele frequencies also showed no significant association with disease in this series (data not shown). When we focused on the Finnish individuals, a slightly but not significant overrepresentation of the GG genotype in the two intronic polymorphisms (rs1989756 and rs1989754) was identified in patients compared to controls (P = 0.087 and 0.089 respectively). There was significant increase in the rs1989754 G allele in patients compared to control Finnish individuals (52% vs. 42%, P = 0.030). Comparison of the rs1989756, rs1721100, and ss20399075 allele frequencies did not reveal any significant differences (P = 0.099, P = 0.239, and P = 0.424 respectively). Repeating the analyses after the exclusion of all early onset cases did not result in any significant changes (data not shown).
Table 1 FGF20 genotype frequencies in cases and controls
Finnish series Greek series
rs1989756 PD cases (%) Controls (%) PD cases (%) Controls (%)
GG 130 (90.3) 110 (83.3) 124 (82.1) 153 (83.6)
GA 14 (9.7) 22 (16.7) 25 (16.6) 29 (15.8)
AA 0 0 2 (1.3) 1 (0.5)
χ2 (2 df); 2.928 P = 0.087 χ2 (2 df); 0.605 P = 0.739
rs1989754
GG 39 (27.1) 27 (20.1) 30 (20.0) 34 (19.2)
GC 71 (49.3) 60 (44.8) 63 (42.0) 85 (48.0)
CC 34 (23.6) 47 (35.1) 57 (38.0) 58 (32.8)
χ2 (2 df); 4.838 P = 0.089 χ2 (2 df); 1.309 P = 0.520
rs1721100
CC 74 (51.4) 68 (51.9) 73 (50.3) 90 (50.3)
CG 60 (41.7) 48 (36.6) 57 (39.3) 75 (41.9)
GG 10 (6.9) 15 (11.5) 15 (10.3) 14 (7.8)
χ2 (2 df); 1.977 P = 0.372 χ2 (2 df); 0.702 P = 0.704
ss20399075
TT 8 (6.6) 4 (3.3) 2 (1.4) 0
TC 36 (29.5) 36 (29.5) 32 (21.8) 39 (21.7)
CC 78 (63.9) 82 (67.2) 113 (76.9) 141 (78.3)
χ2 (2 df); 1.433 P = 0.488 χ2 (2 df); 2.472 P = 0.291
Linkage disequilibrium (LD) analysis revealed a strong LD between all the neighboring markers (P values < 0.01 in all comparisons). Only the neighboring SNPs rs1721100 and ss20399075 did not show significant LD in the Finnish control group.
No significant difference in the distribution of four marker haplotypes between cases and controls was found in either of the series (see Table 2). Just a slightly decrease of the GGCC haplotype in Greek PD patients was found compared to control individuals (P = 0.073).
Table 2 FGF20 haplotype frequencies distributions
Finnish sample a Greek sample b
PD cases Controls PD cases Controls
G G C C 42.6% 33.8% 33.8% 40.4%
G C G C 17.9% 23.4% 17.0% 19.4%
G C C C 16.1% 15.7% 21.4% 19.4%
G G C T 8.6% 8.2% 2.6% 1.9%
G C G T 8.2% 5.6% 9.2% 8.0%
A C C C 2.4% 8.3% 9.6% 8.2%
* 4.8% 5.0% 6.4% 2.7%
a Fisher's exact test, P = 0.113.
b Fisher's exact test, P = 0.073
* Other haplotypes with frequencies lower than 5%.
The order of SNPs are: rs1989756, rs1989754, rs1721100, and ss20399075.
Discussion
Among all the comparisons that we have performed; only polymorphism rs1989754 G allele frequency was significantly higher in patients than controls in the Finnish sample. However, when we adjusted for multiple testing in order to control for a type I error by means of a permutation test, the P value was not longer significant.
Taking into account the allele frequencies, our study has a power of detecting risks from 1.7 to 3.6 in the Finnish sample and from 1.6 to 2.1 in the Greek series with 80% power (95% confidence interval). These results, thus, suggest that FGF20 is not a major risk factor for sporadic PD in the Finnish and Greek population.
The present work does not replicate the recently reported association between polymorphisms at the FGF20 gene and PD [8]. It is important to note that we have genotyped the same polymorphisms that they previously performed except the one located upstream the exon 1 (ss20399076) because this was not found in the SNP database and we could not localize this SNP based on the published report. Nevertheless, because no association was reported between this SNP and PD, the absence of this polymorphism in our work should not be a major limitation.
Several reasons could explain the discrepancy between our results and the one by Van der Walt and collaborators [8]: first, the reported positive association could be the result of a type I error, due to possible admixture. Population stratification due to ethnic admixture could have been the result of using US based sample, which was not fully detailed by the authors. Another possible explanation for the lack of replication could be the result of using different epidemiological approaches. We believe that the association reported by Walt et al. should be strong enough to be consistent when other genetic epidemiology methods are employed. Since the reported p values are in the order of 1*10E-3, one could expect a similar pattern of association, independently of the method utilized. Therefore, it seems very improbable that the lack of replication is due to the use of another methodology. A third interpretation could be that the association is real but the effect is smaller in Greek and Finnish individuals. If this was true, our series may not be large enough to assess variants that contribute to complex traits and are likely to have modest effects. Fourth and finally, a real association could exist only in groups of a specific ethnic origin and, presumably, a similar genetic background.
Conclusion
In order to test whether the previously described association between FGF20 and PD is consistent in other populations, we have performed a case-control study in which we compare and analyze four biallelic polymorphisms within FGF20 gene in two different populations from Europe: one from Greece and one from Finland. Our results clearly indicate that genetic variations in FGF20 do not influence the risk of PD in these two populations. As far as we know, this is the only study that has tried to replicate the previous reported association between FGF20 and PD.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JC performed the genotyping, statistical analysis, and drafted the manuscript. GX, JE, VG, OH, ED, TP, AP, GMH, and PJT, contributed to collecting materials and participated in its design and coordination together with drafting the manuscript. ABS conceived the study and participated in the conceptual design and coordination of this work, together with drafting the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Supported in part by the Helsinki University Central Hospital, The Finnish Cultural Foundation.
==== Refs
de Rijk MC Tzourio C Breteler MM Dartigues JF Amaducci L Lopez-Pousa S Manubens-Bertran JM Alperovitch A Rocca WA Prevalence of parkinsonism and Parkinson's disease in Europe: the EUROPARKINSON Collaborative Study. European Community Concerted Action on the Epidemiology of Parkinson's disease J Neurol Neurosurg Psychiatry 1997 62 10 15 9010393
Sveinbjornsdottir S Hicks AA Jonsson T Petursson H Gugmundsson G Frigge ML Kong A Gulcher JR Stefansson K Familial aggregation of Parkinson's disease in Iceland N Engl J Med 2000 343 1765 1770 11114315 10.1056/NEJM200012143432404
Payami H Zareparsi S James D Nutt J Familial aggregation of Parkinson disease: a comparative study of early-onset and late-onset disease Arch Neurol 2002 59 848 850 12020270 10.1001/archneur.59.5.848
Hardy J Cookson MR Singleton A Genes and parkinsonism Lancet Neurol 2003 2 221 228 12849210 10.1016/S1474-4422(03)00350-8
Gwinn-Hardy K Genetics of parkinsonism Mov Disord 2002 17 645 656 12210852 10.1002/mds.10173
Paisan-Ruiz C Jain S Evans EW Gilks WP Simon J van der Brug M de Munain AL Aparicio S Gil AM Khan N Johnson J Martinez JR Nicholl D Carrera IM Pena AS de Silva R Lees A Marti-Masso JF Perez-Tur J Wood NW Singleton AB Cloning of the gene containing mutations that cause PARK8-linked Parkinson's disease Neuron 2004 44 595 600 15541308 10.1016/j.neuron.2004.10.023
Scott WK Nance MA Watts RL Hubble JP Koller WC Lyons K Pahwa R Stern MB Colcher A Hiner BC Jankovic J Ondo WG Allen FHJ Goetz CG Small GW Masterman D Mastaglia F Laing NG Stajich JM Slotterbeck B Booze MW Ribble RC Rampersaud E West SG Gibson RA Middleton LT Roses AD Haines JL Scott BL Vance JM Pericak-Vance MA Complete genomic screen in Parkinson disease: evidence for multiple genes Jama 2001 286 2239 2244 11710888 10.1001/jama.286.18.2239
van der Walt JM Noureddine MA Kittappa R Hauser MA Scott WK McKay R Zhang F Stajich JM Fujiwara K Scott BL Pericak-Vance MA Vance JM Martin ER Fibroblast growth factor 20 polymorphisms and haplotypes strongly influence risk of Parkinson disease Am J Hum Genet 2004 74 1121 1127 15122513 10.1086/421052
Ohmachi S Watanabe Y Mikami T Kusu N Ibi T Akaike A Itoh N FGF-20, a novel neurotrophic factor, preferentially expressed in the substantia nigra pars compacta of rat brain Biochem Biophys Res Commun 2000 277 355 360 11032730 10.1006/bbrc.2000.3675
Ohmachi S Mikami T Konishi M Miyake A Itoh N Preferential neurotrophic activity of fibroblast growth factor-20 for dopaminergic neurons through fibroblast growth factor receptor-1c J Neurosci Res 2003 72 436 443 12704805 10.1002/jnr.10592
Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases J Neurol Neurosurg Psychiatry 1992 55 181 184 1564476
|
15967032
|
PMC1183215
|
CC BY
|
2021-01-04 16:28:53
|
no
|
BMC Neurol. 2005 Jun 20; 5:11
|
utf-8
|
BMC Neurol
| 2,005 |
10.1186/1471-2377-5-11
|
oa_comm
|
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-471602949810.1186/1471-2202-6-47Research ArticleStimulus-dependent spatial patterns of response in SI cortex Chiu Joannellyn S [email protected] Mark [email protected] Barry L [email protected] Oleg V [email protected] Departments of Biomedical Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA2 Cell and Molecular Physiology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA2005 19 7 2005 6 47 47 9 2 2005 19 7 2005 Copyright © 2005 Chiu 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
Recently we reported that vibrotactile flutter stimulation of a skin locus at different amplitudes evokes an optical response confined to the same local region of the primary somatosensory cortex (SI), where its overall magnitude varies proportionally to the flutter amplitude. In this report, we characterize the impact of the flutter amplitude on the spatial patterns of activity evoked within the responding SI region.
Results
In order to characterize the spatial pattern of activity within the responding SI region, images of the flutter-evoked SI optical response were segmented and analyzed with spatial frequency analysis. The analysis revealed that: (1) dominant spatial frequencies in the optical intrinsic signal emerge within the responding SI region within 3–5 sec of stimulus onset; (2) the stimulus-evoked activity is spatially organized in a form of several roughly parallel, anterior-posteriorly extended waves, spaced 0.4–0.5 mm apart; (3) the waves themselves exhibit spatial periodicities along their long axis; and (4) depending on the flutter stimulus amplitude, these periodicities can range from fine 0.15 mm "ripples" at 50 μm amplitude to well-developed 0.5 mm fluctuations at the amplitude of 400 μm.
Conclusion
The observed spatiointensive fractionation on a sub-macrocolumnar scale of the SI response to skin stimulation might be the product of local competitive interactions within the stimulus-activated SI region and may be a feature that could yield novel insights into the functional interactions that take place in SI cortex.
==== Body
Background
Afferent projections from skin to primary somatosensory cortex (SI) are well known to form a fine map of the body surface in SI. In this map, a skin locus provides afferent input to an extensive cortical region in SI [1,2]. In particular, the direct connectivity between somatosensory thalamus and SI cortex is now recognized to be much more spatially distributed than previously believed (e.g., in primates the ventrobasal thalamic region which receives its input from a single digit projects to an extensive, 20 mm2 sector of SI cortex – [3,4]). The intrinsic SI excitatory connections link not only neighboring but also widely separated regions of somatosensory cortex [5]. These connections ensure that many members of widely distributed neuronal populations interact extensively within milliseconds after the onset of stimulus-evoked thalamocortical drive. Thus it is not surprising to find that the processing of even a very local skin stimulus is associated with SI activation over several sq. millimeters of cortical area, as revealed, for example, with optical imaging techniques [6-10].
Such spatially extensive cortical regions are not functionally homogeneous. For example, using Optical Intrinsic Signal (OIS) imaging in near-infrared (830 nm) range, we find that in squirrel monkeys a small-diameter stimulus probe oscillating on the skin at 25 Hz activates more than 3 mm2 of cortical territory in area 3b of SI [6,8,11]. Such a territory can contain as many as 20 place-defined cortical columns ("segregates"; [12]) organized into 4–6 alternating rapidly- and slowly-adapting submodality bands [13]. Chen et al. [10] reported that the relative magnitudes of optical response in local, 0.2–0.4 mm wide, SI regions changes when the frequency of the stimulating probe is changed from simple taps to 25 Hz to 200 Hz (thus preferentially activating different submodalities of skin mechanoreceptors). And on even finer spatial scale SI might be organized in ~50 μm-diameter functionally distinct minicolumns [1,12,14-18].
Together these considerations suggest that the spatial pattern of activity evoked in SI by even the smallest stimuli might be structurally more complex than a typically envisioned basic bell-shaped pattern. A closer inspection of such patterns might reveal certain spatial formations within them with significant functional implications. Recently, we investigated the response of SI cortex to varying amplitudes of flutter stimulation. Regardless of the amplitude of stimulation (in the range of 50 to 400 μm), we found that the spatial extent of the response of SI cortex remained the same [19]. Instead, the actuated cortical region exhibits increases in its magnitude of neuronal response proportional to the intensity of stimulation [19-21]. One feature of particular interest in our study was that the activity patterns evoked within these spatially delineated regions, when viewed at high resolution, appeared to develop in an orderly manner dependent on stimulus amplitude. The purpose of this study was to determine if those patterns are indeed systematic, and if so, to characterize them quantitatively.
Results
Figure 1 illustrates the basic method that we used to examine and compare the absorbance patterns evoked by different amplitudes of flutter stimulation. Panel A shows the cortical field (SI cortex of squirrel monkey) that was imaged. Panels B and C are the light absorbance images evoked in this subject by low-amplitude (50 μm) and high-amplitude (400 μm) flutter stimulation of the same spot on the thenar eminence. The responses to both stimuli occupy the same approximately circular, 2 mm-diameter cortical region. The stimulus-evoked activity in the central 2 × 2 mm sector of this responding region was plotted as a 3-dimensional surface map (Panels D and E) to facilitate the view of the evoked patterns of response. In order to quantify this pattern, the pattern was segmented along the cortical anterior-posterior axis and the medial-lateral axis (segmentation orientations indicated by Panels F and G) and generated spatial histograms (Panels H, I, L and M). These histograms hint at a spatial pattern that could have some underlying spatial frequency. To observe their spatial frequency composition, the power spectra of these spatial histograms are plotted in Panels J, K, N and O. The periodograms in these panels reveal major differences in the spatial organization of the SI responses to the 50 μm and 400 μm stimuli when segmented in the anterior-posterior dimension. That is, the SI response to the higher-amplitude stimulus is greatly dominated by lower spatial frequencies between 1.5–2.5 cycles/mm, whereas in the SI response to the weaker stimulus the relative power in these frequencies is greatly reduced and greater power is present at spatial frequencies between 5.5–9 cycles/mm. In contrast, when segmented in the medial-lateral dimension (Panels N and O), there is no shift in the power spectra between the low and high amplitude responses. The medial-lateral power spectra are clearly dominated by the low spatial frequencies around 1.5–2.5 cycles/mm.
Comparable results were obtained from the 4 other subjects, (3 subjects shown Fig. 2). Viewing the 3-D activity maps (Figure 2, second-left column), it is evident that the low amplitude stimulus evokes a pattern with a much higher spatial frequency than that evoked by the higher amplitude stimulus. The third column also shows that in each subject there was a shift of the most prominent frequency band of the OIS power spectrum from ~7.5 cycles/mm to ~2 cycles/mm as the stimulus amplitude was increased from 50 μm to 400 μm, when the stimulus-evoked activity was sampled in the anterior-posterior dimension. In the medial-lateral dimension (Figure 2, right column), no such shift in the power spectra was observed. Low frequencies around 2–2.5 cycles/mm dominate all of these spectra.
Figure 3 illustrates results obtained in a more detailed examination of the effect of increased amplitude of flutter stimulation on the spatial organization of the OIS response pattern. Figure 3A shows the SI optical responses, obtained in the same experiment, to five different stimulus amplitudes ranging from 0 (control) to 400 μm. 3-D surface maps of the central 2 × 2 mm activated cortical region are displayed in Figure 3B. The power spectra of the responses sampled in the anterior-posterior dimension (Fig. 3C) show that in the absence of flutter (0 μm stimulus amplitude) the distribution of power at different spatial frequencies is relatively uniform, with the most dominant frequency at 6.5 cycles/mm. With increments in stimulus amplitude, however, the relative power in the spectral distribution shifts towards lower spatial frequencies. In the case of the medial-lateral dimension (Figure 3D), the power spectra remain essentially constant, with an exception of the no-stimulus control, and there does not appear to be a shift in the distribution of the power with increasing the amplitude of stimulation.
This effect of stimulus amplitude on the distribution of spatial frequencies in the OIS response along the anterior-posterior dimension is highly reproducible across all subjects. Figure 4A shows the average across-subject (n = 5) power spectra, obtained in response to flutter stimuli delivered at 5 different amplitudes. The 6–9.5 cycles/mm frequency band, which is most prominent in the absence of stimulation (control; 0 μm amplitude), loses relative power as stimulus amplitude is increased (Fig. 4B top). In contrast, relative power at 1.5–3 cycles/mm grows as stimulus amplitude is increased, and at the highest amplitude used (400 μm) it greatly dominates the power spectrum (Fig. 4B bottom). Overall, the OIS power spectrum appears to respond to an increase in stimulus amplitude by a shift of the relative power towards lower frequencies. This tendency is expressed more clearly in Figure 5, where the highest-power frequency (Fig. 5A; or, inversely, in Fig. 5B, the highest-power period) is plotted as a function of the stimulus amplitude.
Stimulus duration appears to alter the spatial organization of SI optical response to flutter in a manner similar to the alteration that accompanies an increase in stimulus amplitude. Figure 6 displays the temporal evolution of the responses evoked by four different amplitudes. In each case, the pattern of absorbance evoked by the flutter stimulus appears to become more organized and periodic with time after the stimulus onset. In other words, with increasing stimulus duration, the local aggregates of above background absorbance tend to form larger clusters – which would lead to higher periodic values (lower spatial frequencies). The spatial frequency changes with stimulus duration were quantified in a manner similar to those that were used to quantify the spatial frequency characteristics that changed with stimulus amplitude. To give a representative example, Figure 7 shows the temporal evolution of the SI response of a subject to a 400 μm-amplitude flutter stimulus. The images in Figure 7A were obtained 1 sec prior to stimulus onset (control), as well as at 1, 3, 5, and 7 sec after the onset of continuous skin flutter stimulation. The power spectra obtained from these images (Fig. 7B) show a systematic leftward shift of the dominant frequencies with increasing time after stimulus onset, from ~6.5 cycles/mm prior to stimulation, to ~2 cycles/mm after 7 sec of continuous stimulation. The plots in Figure 7C show that the 6–9.5 cycles/mm frequency band, which is dominant in the resting state, loses relative power after onset of stimulation, and during this same time relative power within 1.5–3 cycles/mm frequency band also becomes maximal.
Figure 8 summarizes the temporal development of OIS response patterns by plotting the highest-power spatial frequency in the anterior-posterior dimension (averaged across 5 subjects) at different times after stimulus onset. Two plots are shown for comparison: the first for 400 μm-amplitude stimuli, the second for 50 μm-amplitude stimuli. These plots make it evident that during continuous flutter stimulation: (1) the spectral power of the response migrates towards lower frequencies with increasing time after stimulus onset – indicating that SI optical response pattern undergoes a gradual spatial reorganization; and (2) the higher the stimulus amplitude, the faster the shift of spectral power towards lower frequencies.
Figure 8C plots the magnitude of OIS response to the 50 μm- and 400 μm-amplitude stimuli as a function of time after the stimulus onset, showing that during the time when the dominant spatial frequency migrates across the power spectrum, the OIS also grows overall in its magnitude. The concurrency of these changes raises a parsimonious possibility that OIS periodicity is a direct function of the OIS magnitude (and thereby only an indirect function – via their control of the OIS magnitude – of the stimulus strength and duration). To evaluate this possibility, Figure 8D plots the highest-power spatial frequency at any given time (taken from Figure 8A plot) as a function of the overall OIS magnitude at that time (taken from the Figure 8C plot). As the Figure 8D plot shows, the relationship between OIS magnitude and periodicity obtained with the 50 μm-amplitude stimuli appears to be different from the relationship obtained with the 400 μm stimuli, suggesting that OIS periodicity cannot be explained simply by the overall OIS magnitude.
Discussion
Spatial frequency analysis revealed that the SI response to flutter stimulation produces systematic, spatially periodic fluctuations in the magnitude of the OIS within the responding SI region. In the absence of stimulation – the control condition – the power spectra have a small prevalence of spatial periodicities in the range of 0.1–0.14 mm. In contrast, in the presence of a prominent flutter stimulus spatial periodicities in the range of 0.4–0.6 mm dominate the stimulus-activated SI region.
Visual inspection of the 3-D activity plots (Figures 1, 2, 3, 6), together with the results of spectral analyses performed along the anterior-posterior and medial-lateral cortical dimensions, point to a substantial tendency of the stimulus-evoked activity to take a form of a pattern of roughly parallel elongated waves oriented in the anterior-posterior direction. These waves have a medio-lateral spacing of 0.4–0.5 mm between their crests. This tendency is already well developed in response to a relatively weak, 50 μm-amplitude flutter. At such weaker stimuli, each anterior-posteriorly oriented wave has a pattern of fine "ripples" along its long axis, with the dominant spatial periodicities in the neighborhood of 0.15-0.2 mm – values close to those characteristic of the OIS in the no-stimulus condition. Stronger stimuli lead to an emergence of larger spatial periodicities in the anterior-posterior dimension, up to 0.6 mm. Such a stimulus amplitude-dependent shift towards larger spatial periodicities suggests that the spatial pattern of SI response to flutter stimulation undergoes substantial reorganization in response to changes in stimulus amplitude/intensity. That is, the stronger (200–400 μm amplitude) stimuli have an effect of restructuring the anterior-posteriorly oriented waves by replacing the high-frequency ripples along their crests with more prominent ca. 0.5 mm periodic fluctuations.
Such a reorganization of stimulus-evoked OIS spatial patterns apparently also occurs across time. That is, flutter stimulus-evoked OIS patterns in the anterior-posterior dimension develop gradually from the resting state – dominated by high spatial frequencies – through a series of states characterized by progressively lower spatial frequencies. Thus, the spatial organization of the patterned response evoked in SI by a flutter stimulus varies with both stimulus strength and stimulus duration.
In interpreting the outcomes of our spatial frequency analyses, it is important that it be recognized that they were performed not on the SI neuroelectrical responses to flutter stimulation, but on SI optical responses to such stimulation. It now is well established that cortical neuroelectrical activity is positively and strongly correlated with the local increase in cortical tissue light absorbance [8,11,22-27]. Although highly correlated with neuroelectrical activity, however, the OIS is not a direct reflection of either neuronal spike discharge activity or, more generally, the local voltage changes evoked by a sensory stimulus. Notably, it has a much slower onset and decay than the neuroelectrical responses of cortical cells. The intrinsic signal detected using the near-infrared light ("IR imaging") is relatively independent of changes in blood flow [28,29]. The optical signal observed under near-infrared light reflects a variety of factors but, most significantly, change (shrinkage) in the volume of the extracellular fluid compartment attributable to the glial swelling due to stimulus-evoked changes in extracellular [K+] and/or neurotransmitter release [23,30,31]. In view of these complex origins of the near-infrared OIS, it remains to be determined whether the observed stimulus-dependent periodicities reflect the spatial organization of stimulus-evoked neuroelectrical activity. Alternatively, the observed OIS periodicities might reflect the spatial organization of the cortical glial reaction to local neuroelectrical activity, or SI microvascular responses [32], which become more prominent with increases in both stimulus strength and duration.
If the observed stimulus-dependent OIS periodicities were to reflect neuroelectrical activity patterns, how well would they fit with the known features of SI functional organization? In particular, a macrocolumnar functional organization has been well documented in SI cortex. Receptive field-mapping techniques have revealed that SI cortex is partitioned into ~0.5 mm-wide submodality- and place-defined columns [1,2,12,13,33,34]. These topographic entities, repeating every 0.5 mm, might indeed be responsible for the prominent 0.5 mm periodicity in SI OIS stimulus-evoked activity patterns. If the spatial OIS periodicities were due to selective modulation of entire macrocolumns (in other words, due to SI response fractionation on the macrocolumnar scale), then the observed OIS periodicities would be at least double the size of macrocolumns. Since this is not the case, and the OIS periodicities approximate the size of macrocolumns, the fractionation of the SI response would appear to take place on the submacrocolumnar scale, with flutter stimuli preferentially activating only a subsector of each macrocolumn in the SI region engaged by the stimulus. Such preferentially activated subsectors would then form a roughly periodic pattern across the responding SI region. In support of this proposal, Chen et al. [10] have reported that single skin taps, 25 Hz flutter, and 200 Hz vibration (stimulus conditions that allow preferential activation of the different classes of skin mechanoreceptors) were associated with preferential activation of different 0.25 mm-diameter regions within the responding SI territory. The small size of these regions suggests that they should occupy only a subsector of a typical submodality column, rather than a whole column [13].
Whether or not the 0.5 mm periodicities in OIS stimulus-response patterns in SI owe their existence to submodality- and/or place-defined macrocolumns, the fact remains that a flutter stimulus activates multiple cortical loci, separated from each other by less active loci. Such a locally selective distribution of activity in SI is suggestive of the presence in SI functional architecture of a mapping factor in addition to submodality and place on the skin – a factor specific to some yet to be explored attribute(s) of mechanical skin stimulation. Similarly, Bruno and colleagues recently demonstrated that individual barrels in rat SI contain minicolumns of neurons preferring the same whisker deflection angle and that these angular tuning domains could be the result of convergent inputs from thalamocortical cells with corresponding angular preferences [35]. It is possible that upon further investigation, the spatial activity patterns evoked in SI cortex, such as the amplitude-dependent patterns described in this report, will also be found to be submodality-dependent as well.
Finally, considering that spatial periodicities in SI OIS response patterns emerge gradually, and evolve with time after the stimulus onset, the fine sculpting of the SI response might be the result of a network-level neurocomputational process that involves competitive and cooperative interactions among local neuronal aggregates. That is, SI sensitivity to the stimulus attribute(s) responsible for the observed fractionation of activity within the responding SI region might be an emergent property of the SI network (i.e., a product of network-level computation), rather than a simple outcome of selective convergence of thalamocortical afferents on SI neurons.
Conclusion
Observations of the spatial patterns of SI cortical response within an activated region, such as those evoked by flutter stimulation of the skin, suggest that evoked cortical activity within such a territory is not evenly distributed. Furthermore, the cortical activity patterns change in a manner that appears to be dependent upon stimulus conditions. The observed spatiointensive fractionation on a sub-macrocolumnar scale of the SI response to skin stimulation might be the product of local competitive interactions within the stimulus-activated SI region, and as such can lead to new insights about the functional interactions that take place in the SI cortex.
Methods
All methods and procedures are consistent with USPHS policies and guidelines on animal care and welfare in biomedical research, and were reviewed and approved in advance by an institutional animal use committee (IACUC). Experiments were conducted in 5 squirrel monkeys. Surgical procedures were carried out under deep general anesthesia (1–4% halothane in a 50/50 mixture of oxygen and nitrous oxide). After induction of anesthesia the trachea was intubated to facilitate positive pressure ventilation and delivery of the gaseous general anesthetic. A catheter was inserted into a branch of the femoral vein of the hindlimb ipsilateral to the hemisphere to be imaged, allowing intravenous (IV) administration of drugs and fluids (5% dextrose and 0.9% NaCl). Methylprednisolone sodium succinate (20 mg/kg) and gentamicin sulfate (2.5 mg/kg) were injected intramuscularly to lessen the probability of halothane-induced cerebral edema and prevent bacterial septicemia, respectively.
A 1.5 cm diameter opening in the skull exposed the forelimb region of SI cortex. A recording chamber was positioned over the opening and cemented to the skull with dental acrylic. The chamber was filled with artificial cerebrospinal fluid, the dura mater overlying SI cortex incised and removed, and all wound margins outside the chamber dressed with long-lasting local anesthetic in oil (Cetacaine). All skin and muscle incisions were closed with sutures and bandaged.
After the completion of all surgical procedures subjects were immobilized with Norcuron (vercuronium bromide; 0.5 mg/kg loading dose; 0. 25-0.5 mg/kg/hr maintenance dose) and ventilated with a gas mixture (a 50/50 mix of oxygen and nitrous oxide; supplemented with 0.5–1.0% halothane). At these concentrations and under normocapnic conditions, halothane has no effect on brain energy metabolism [36,37], and only minor effects on cerebrovascular regulation [38]. Ventilator rate and volume were adjusted to maintain end-tidal CO2 between 3.0–4.0%. EEG and cardiovascular signs (EEG slow wave content; EKG and heart rate) were monitored continuously, and the anesthetic gas mixture adjusted intermittently to maintain values and reactivity to skin stimuli consistent with light general anesthesia. Rectal temperature was maintained (using a heating pad) at 37.5°C.
The recording chamber was filled with artificial cerebrospinal fluid and hydraulically sealed using a clear glass plate. Vibrotactile stimuli were delivered to selected loci on the hand using a servocontrolled vibrotactile stimulator [39], capable of delivering precisely controlled sinusoidal vertical skin displacement stimuli. The stimulator made contact with the skin via a cylindrical 2 mm-diameter Delrin probe. All sinusoidal vibrotactile stimuli were superimposed on a static displacement ("pedestal") of 500 μm. Identical parameters of stimulation were used at each skin site that was studied: frequency of vibration 25 Hz (in the flutter range), stimulus duration 7 sec, and interstimulus interval 60 sec. Different peak-to-peak amplitudes of flutter stimulation (0, 50, 100, 200 and 400 μm) were interleaved on a trial-by-trial basis.
The optical imaging system consisted of a computer-interfaced CCD camera, light source, guide and filters required for near-infrared (830 nm) illumination of the cortical surface, a focusing device, and a recording chamber capped by an optical window (for additional methodological details see [7,40]). Images of the exposed cortical surface were acquired 200 ms before stimulus onset ("reference" or "prestimulus" images) and continuously thereafter ("poststimulus" images; at a resolution of one image/s) for 15s following stimulus onset. Exposure time was 200 ms. Light absorbance images were generated by subtracting each prestimulus (reference) image from a poststimulus image and subsequently dividing by the reference image. Absorbance images obtained in this way typically show regions in which light absorption increases and other regions in which absorption decreases in response to skin stimulation. These regions, respectively, have been shown to correspond to regions in which neuronal activity increases and decreases in response to sensory stimulation ([6,27,41-43]; for review see [44]). Stimulus-evoked OIS responses of SI were displayed as either grayscale images or 3-D surface plots. To reduce amount of noise in these displays, the images were smoothed using a 3 × 3 pixel boxcar filter.
Cortical images were taken at light/time exposures that place the region of interest in the middle of the range of the recorded pixel values. Histogram analysis was used during experimental setup to make sure to avoid any nonlinearities that may arise from overexposure. In some of the experiments the camera was rotated by 90° relative to the SI orientation to better capture the responding cortical field. These rotations did not have any noticeable effect on spectral power distributions along the anterior-posterior and medial-lateral cortical dimensions.
The spatial organization of the stimulus-related light absorbance changes in SI was evaluated using linear image segmentation. This involved segmentation of the relevant region of the image into a linear series of bins and computation of the average absorbance value of the pixels in each bin. The sequence of average absorbance values obtained in this way was plotted as a function of distance (mm) along the cortical path traced by the central points of the series of bins – yielding an absorbance vs. distance plot (thus forming a spatial histogram). Power spectra of the spatial histograms were then computed using Discrete Fourier Transform (DFT) algorithm and plotted as a periodogram. Fourier analysis was always performed on raw, unfiltered images.
At the end of the experiment the subject was euthanized by overdose of pentobarbital (50 mg/kg/IV), followed by intracardial perfusion with saline and 10% formalin.
Authors' contributions
JC participated in acquisition of optical data, performed analysis of the data and drafted the manuscript. OF and BW participated in the design of the study, the conduct of the experiments and the drafting of the manuscript. MT participated in the design, conduct, and analysis of the experiments, and in the manuscript preparation.
Acknowledgements
This work was supported, in part, by US Army Research Office grant P43077-LS (M. Tommerdahl, P.I.), NIH NS050587 (M. Tommerdahl, P.I.) and NIH NS35222 (B. Whitsel, P.I.).
Figures and Tables
Figure 1 Comparison of SI cortical responses evoked by low- and high-amplitude 25 Hz flutter stimuli. A: View of the somatosensory cortex with the lateral end of the central sulcus (CS) at the top. Stimulus location is shown on the hand figurine. B: Optical response of the SI to the low-amplitude (50 μm peak-to-peak) flutter. Shown is the light absorbance image ([poststimulus image – prestimulus image]/prestimulus image) averaged across 20 stimulus trials, and in each trial across 3 poststimulus images taken during the 5–7 sec interval of continuous stimulation. Increased light absorbance is indicated by darker shading. C: Optical response of SI to the high-amplitude (400 μm) flutter. D & E: Magnified 3-D surface maps of OIS magnitude in the responding cortical region. F & G: Orientation of segmentation in the M-L dimension (F) and the A-P dimension (G). H, I, L & M: Spatiointensive histograms generated via segmentation of areas indicated in F & G for two different stimulus conditions. J, K, N & O: Periodograms showing the spatial frequency composition of the spatiointensive histograms. Note change in the power spectra with the change in stimulus amplitude in data sampled in the A-P orientation, but not in the M-L orientation.
Figure 2 Comparison of SI cortical responses evoked by low- and high-amplitude 25 Hz flutter stimuli in three additional subjects. Shown for each subject (A-C) are: (1) the stimulus location on the hand, (2) the average light absorbance images of the cortical optical response to the 50 μm- and 400 μm-amplitude stimuli, (3) the magnified 3-D activity maps of the region sampled, and (4) the power spectra of the OIS responses. All subjects exhibited similar differences between their low and high stimulus amplitude power spectra.
Figure 3 Comparison of SI optical responses evoked at different stimulus amplitudes (0, 50, 100, 200, and 400 μm) in an exemplary experiment. A: Average light absorbance images for each of the stimulus amplitudes (20 trials, 3 poststimulus images per trial). B: 3-D activity maps of the responding cortical region. C & D: Power spectra measured for each stimulus amplitude in the two dimensions sampled. With an increase in stimulus amplitude, the spectral power shifts to lower spatial frequencies for the data sampled in the A-P dimension, but not in the M-L dimension.
Figure 4 Across-subject reproducibility of the effect of flutter stimulus amplitude on the power spectra of the data sampled in the A-P dimension. A: Average periodograms of the OIS responses of 5 subjects to 5 different stimulus amplitudes. B: Average relative power in the 6–9.5 cycles/mm and 1.5–3 cycles/mm frequency bands plotted as a function of the stimulus amplitude. As the stimulus amplitude increases, the spectral power shifts progressively to lower spatial frequencies.
Figure 5 Dependence of the position of the largest spectral peak in OIS periodograms on the stimulus amplitude. The frequency (A) and the period (B) of the largest spectral peak (average of 5 subjects – see Figure 4) are plotted as a function of the stimulus amplitude. The peak frequency/period shifts across the power spectrum with a change of stimulus amplitude.
Figure 6 Comparison of temporal development of cortical patterns of absorbance evoked by different amplitudes of flutter vibration. 3-D activity maps of the same cortical region (of the same subject) are plotted at -1 (control), 1, 3 and 5 secs after the stimulus onset for amplitudes of 50, 100, 200 and 400 μm (Panels A, B, C, & D respectively; in all panels, color-scales are normalized by their maxima). Note the difference in the progression of the pattern development evoked by the different stimulus amplitudes.
Figure 7 Temporal development of power spectra in response to a 400 μm-amplitude flutter stimulus. A: Light absorbance images obtained at selected times before and during flutter stimulation. B: Power spectra of OIS activity sampled in the anterior-posterior dimension across the activated cortical region. With increasing stimulus duration, the most prominent frequency band in the power spectra shifts to the left relative to the pre-stimulus condition. C: Temporal shift in relative power in the 6–9.5 cycles/mm band (top) and in the 1.5–3 cycles/mm band (bottom). As stimulus duration increases, the relative power decreases in the 6–9.5 cycles/mm band, but increases in the 1.5–3 cycles/mm band.
Figure 8 Temporal shift of the position of the largest spectral peak in the anterior-posterior OIS periodograms and its dependence on stimulus amplitude. The frequency (A) and the period (B) of the largest spectral peak (average of 5 subjects) are plotted as a function of time since the stimulus onset. Two plots are shown, for 50 μm-amplitude and 400 μm-amplitude stimuli. With increasing stimulus duration, the peak frequency/period shifts across the power spectrum, at a rate dependent on the stimulus amplitude. C: Average magnitude of OIS in the segmented region as a function of time after stimulus onset. The dominance of lower spatial frequencies develops in time in parallel with an overall growth of the OIS response to the stimulus. D: The frequency of the largest spectral peak plotted as a function of the average magnitude of OIS in the sequential region. The relationship between the largest spectral peak and overall OIS magnitude varies depending on the stimulus amplitude.
==== Refs
Favorov OV Whitsel BL Spatial organization of the peripheral input to area 1 cell columns: I. The detection of "segregates." Brain Res 1988 472 25 42 3342334
Favorov OV Whitsel BL Spatial organization of the peripheral input to area 1 cell columns: II. The forelimb representation achieved by a mosaic of segregates Brain Res 1988 472 43 56 3342335
Rausell E Bickford L Manger PR Woods TM Jones EG Extensive divergence and convergence in the thalamocortical projection to monkey somatosensory cortex J Neurosci 1998 18 4216 4232 9592100
Jones EG Cortical and subcortical contributions to activity-dependent plasticity in primate somatosensory cortex Annu Rev Neurosci 2000 23 1 37 10845057 10.1146/annurev.neuro.23.1.1
Burton H Fabri M Ipsilateral intracortical connections of physiologically defined cutaneous representations in areas 3b and 1 of macaque monkeys: projections in the vicinity of the central sulcus Journal of Comparative Neurology 1995 355 508 538 7636029 10.1002/cne.903550404
Tommerdahl M Delemos KA Favorov OV Metz CB Whitsel BL Response of anterior parietal cortex to different modes of same-site skin stimulation Journal of Neurophysiology 1998 80 3272 3283 9862921
Tommerdahl M Delemos KA Whitsel BL Favorov OV Metz CB Response of anterior parietal cortex to cutaneous flutter versus vibration Journal of Neurophysiology 1999 82 16 33 10400931
Tommerdahl M Favorov OV Whitsel BL Optical imaging of intrinsic signals in somatosensory cortex Behav Brain Res 2002 135 83 91 12356438 10.1016/S0166-4328(02)00159-6
Shoham D Grinvald A The cortical representation of the hand in macaque and human area S-I: high resolution optical imaging Journal of Neuroscience 2001 21 6820 6825 11517270
Chen LM Friedman RM Ramsden BM LaMotte RH Roe AW Fine-scale organization of SI (Area 3b) in the squirrel monkey revealed with intrinsic optical imaging Journal of Neurophysiology 2001 86 3011 3029 11731557
Tommerdahl M Delemos KA Whitsel BL Favorov OV Metz CB The response of anterior parietal cortex to cutaneous flutter and vibration J Neurophysiol 1999 82 16 33 10400931
Favorov OV Diamond ME Demonstration of discrete place-defined columns – segregates – in the cat SI Journal of Comparative Neurology 1990 298 97 112 2212100 10.1002/cne.902980108
Sur M Wall JT Kaas JH Modular distribution of neurons with slowly adapting and rapidly adapting responses in area 3b of somatosensory cortex in monkeys J Neurophysiol 1984 51 724 744 6716121
Tommerdahl M Whitsel BL Cox EG Diamond ME Kelly DG Analysis of the periodicities in somatosensory cortical activity patterns Society for Neuroscience Abstracts 1987 13 470
Tommerdahl M Favorov OV Whitsel BL Nakhle B Gonchar YA Minicolumnar activation patterns in cat and monkey SI cortex Cerebral Cortex 1993 3 399 411 8260808
Tommerdahl M Favorov O Chiu J Whitsel B Optical intrinsic signal imaging of ipsilateral, contralateral, and bilateral forelimb inputs to cat SII Society for Neuroscience Abstracts 2004 642 612
Buxhoeveden DP Switala AE Roy E Casanova MF Quantitative analysis of cell columns in the cerebral cortex Journal of Neuroscience Methods 2000 97 7 17 10771070 10.1016/S0165-0270(99)00192-2
Kohn A Pinheiro A Tommerdahl MA Whitsel BL Optical imaging in vitro provides evidence for the minicolumnar nature of cortical response Neuroreport 1997 8 3513 3518 9427317
Simons SP Tannan V Chiu J Favorov OV Whitsel BL Tommerdahl M Amplitude-Dependent Response of SI Cortex to vibrotactile stimulation Society for Neuroscience Abstracts 2004 643 613
Chen LM Friedman RM Roe Aw Optical imaging of a tactile illusion in area 3b of the primary somatosensory cortex Science 2003 302 881 885 14500850 10.1126/science.1087846
Sheth BR Moore CI Sur M Temporal modulation of spatial borders in rat barrel cortex Journal of Neurophysiology 1998 79 464 470 9425214
Grinvald A Frostig RD Lieke EE Hildescheim R Optical imaging of neuronal activity Physiol Reviews 1988 68 1285 1365
Lieke EE Frostig RD Arieli A Ts'o DY Hildesheim R Grinvald A Optical imaging of cortical activity: real-time imaging using extrinsic dye-signals and high resolution imaging based on slow intrinsic signals Annu Rev Physiol 1989 51 543 559 2653196 10.1146/annurev.ph.51.030189.002551
Bonhoeffer T Grinvald A Toga AW, Mazziotia JC Optical imaging based on intrinsic signals: The methodology Brain Mapping: The Methods Volume 1996 Academic Press, Inc. New York 55 97
Shmuel A Grinvald A Functional organization for direction of motion and its relationship to orientation maps in cat area 18 Journal of Neuroscience 1996 16 6945 6964 8824332
Grinvald A Shoham D Glaser I Vanzetta A Shyoterman E Slovin H Wijnbergen C Hildescheim RAA Windhorst U, Johansson H In-Vivo optical imaging of cortical architecture and dynamics Modern Techniques in Neuroscience Research 1999 New York: Springer 893 969
Tommerdahl M Delemos K Vierck CJ Favorov O Whitsel B Anterior parietal cortical response to tactile and skin-heating stimuli applied to the same skin site Journal of Neurophysiology 1996 75 2662 2670 8793772
Haglund M Ojemann G Blasdel G Optical imaging of eliptiform and functional activity in human cerebral cortex Nature 1993 358 668 671 10.1038/358668a0
Ba A Guiou M Pouration N Muthialu A Rex D Cannestra A Chen J Toga A Multiwavelength Optical Intrinsic Signal Imaging of Cortical Spreading Depression J Neurophysiol 2002 88 2726 2735 12424307
Cohen LB Changes in neuron structure during action potential propagation and synaptic transmission Physiol Reviews 1973 53 373 418
MacVicar BA Hochman D Imaging of synaptically evoked intrinsic optical signals in hippocampal slices Journal of Neuroscience 1991 11 1458 1469 1851222
Woolsey TA Rovainen CM Cox SB Henegar MH Liang GE Moskalenko YE Sui J Wei L Neuronal units linked to microvascular modules in cerebral cortex: response elements for imaging the brain Cereb Cortex 1996 6 647 660 8921201
Powell TPS Mountcastle VB Some aspects of the functional organization of the cortex of the postcentral gyrus of the monkey: A correlation of findings obtained in a single unit analysis with cytoarchitecture Bulletin of the Johns Hopkins Hospital 1959 105 133 162 14434571
Sretavan D Dykes RW The organization of two cutaneous submodalities in the forearm region of area 3b of cat somatosensory cortex Journal of Comparative Neurology 1983 213 381 398 6300198 10.1002/cne.902130403
Bruno RM Khatri VL Land PW Simons DJ Thalamocortical angular tuning domains within individual barrels of rat somatosensory cortex Journal of Neuroscience 2003 23 9565 9574 14573536
Hess A Stiller D Kaulisch T Heil P Scheich H New insights into the hemodynamic blood oxygenation level-dependent response through combination of functional magnetic resonance imaging and optical recording in gerbil barrel cortex Journal of Neuroscience 2000 20 3328 3338 10777796
Fujibayashi T Sugiura M Yanagimoto M Harada J Goto Y Brain energy metabolism and blood flow during sevoflurane and halothane anesthesia: Effects of hypocapnia and blood pressure fluctuations Acta Anaesthesiol Scand 1994 38 413 418 8067234
Burdett NG Menon DK Carpenter TA Jones JG Hall LD Visualization of changes in regional cerebral blood flow (rCBF) produced by ketamine using long TE gradient-echo sequences: Preliminary results Magnetic Resonance Imaging 1995 13 549 553 7674850 10.1016/0730-725X(95)00010-E
Chubbuck JG Small-motion biological stimulator APL Technical Digest 1966 May-Jun 18 23
Tommerdahl M Whitsel B Favorov O Metz C BL O'Quinn Responses of contralateral SI and SII in cat to same site cutaneous flutter versus vibration Journal of Neurophysiology 1999 82 1982 1992 10515988
Grinvald A Real-time optical mapping of neuronal activity: from single growth cones to the intact mammalian brain Annual Review of Neuroscience 1985 8 263 305 3885828 10.1146/annurev.ne.08.030185.001403
Grinvald A Bonhoeffer T Malonek D Shoham D Bartfeld E Arierli A Hildesheim R Ratzlaff E Squire L, Weinberger N, Lynch G, McGaugh JNY Optical imaging of architecture and function in the living brain Memory Organization and Locus of Change 1991 Oxford Univ. Press 49 85
Grinvald A Lieke E Frostig R Hildesheim R Cortical point-spread function and long-range lateral interactions revealed by real-time optical imaging of macaque monkey primary visual cortex Journal of Neuroscience 1994 14 2545 2568 8182427
Ebner T Chen G Use of voltage-sensitive dyes and optical recordings in the central nervous system Progress in Neurobiology 1995 46 463 506 8532849 10.1016/0301-0082(95)00010-S
|
16029498
|
PMC1183216
|
CC BY
|
2021-01-04 16:03:48
|
no
|
BMC Neurosci. 2005 Jul 19; 6:47
|
utf-8
|
BMC Neurosci
| 2,005 |
10.1186/1471-2202-6-47
|
oa_comm
|
==== Front
BMC OphthalmolBMC Ophthalmology1471-2415BioMed Central London 1471-2415-5-141596976310.1186/1471-2415-5-14Research ArticlePalm is expressed in both developing and adult mouse lens and retina Castellini Meryl [email protected] Louise V [email protected] Bharesh K [email protected] Deni S [email protected] Manfred W [email protected] Ales [email protected] Melinda K [email protected] Department of Biological Sciences, University of Delaware, Newark, DE 19716 USA2 Depts. of Ophthalmology and Visual Sciences and Molecular Genetics, Albert Einstein College of Medicine, Bronx, NY 10461 USA3 Department of Cell and Molecular Biology, Uppsala University, S-75124 Uppsala Sweden4 Department of Pathology, Anatomy, and Cell Biology, Thomas Jefferson University, Philadelphia, PA 19107 USA5 Developmental Biology Division and Department of Ophthalmology, Children's Hospital Research Foundation, 3333 Burnet Avenue, Cincinnati, OH 45229 USA2005 21 6 2005 5 14 14 1 3 2005 21 6 2005 Copyright © 2005 Castellini et al; licensee BioMed Central Ltd.2005Castellini 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
Paralemmin (Palm) is a prenyl-palmitoyl anchored membrane protein that can drive membrane and process formation in neurons. Earlier studies have shown brain preferred Palm expression, although this protein is a major water insoluble protein in chicken lens fiber cells and the Palm gene may be regulated by Pax6.
Methods
The expression profile of Palm protein in the embryonic, newborn and adult mouse eye as well as dissociated retinal neurons was determined by confocal immunofluorescence. The relative mRNA levels of Palm, Palmdelphin (PalmD) and paralemmin2 (Palm2) in the lens and retina were determined by real time rt-PCR.
Results
In the lens, Palm is already expressed at 9.5 dpc in the lens placode, and this expression is maintained in the lens vesicle throughout the formation of the adult lens. Palm is largely absent from the optic vesicle but is detectable at 10.5 dpc in the optic cup. In the developing retina, Palm expression transiently upregulates during the formation of optic nerve as well as in the formation of both the inner and outer plexiform layers. In short term dissociated chick retinal cultures, Palm protein is easily detectable, but the levels appear to reduce sharply as the cultures age. Palm mRNA was found at much higher levels relative to Palm2 or PalmD in both the retina and lens.
Conclusion
Palm is the major paralemmin family member expressed in the retina and lens and its expression in the retina transiently upregulates during active neurite outgrowth. The expression pattern of Palm in the eye is consistent with it being a Pax6 responsive gene. Since Palm is known to be able to drive membrane formation in brain neurons, it is possible that this molecule is crucial for the increase in membrane formation during lens fiber cell differentiation.
==== Body
Background
The retina and lens form from the neural tube and head ectoderm respectively. Despite these different origins, the development of the mature eye requires mutually inductive interactions between these two cell layers [1]. Further, in many cases, the lens and retina express the same developmentally important transcription factors [2-6]. In addition, a number of studies have identified the expression of proteins with known roles in neuronal function in the lens [7-12] and proteins important in lens function in the retina [13,14]. This may partially be due to the need of both retinal neurons and lens fiber cells to develop elaborated plasma membranes for their function [15-17].
Pax6 is a paired and homeodomain containing transcription factor that is required for the formation of the lens placode from the head ectoderm [18]. Specific loss of Pax6 expression from retinal progenitor cells results in the conversion of all retinal cell types to amacrine interneurons [19] and lens epithelial cells heterozygous for a Pax6 mutation preferentially differentiate into lens fiber cells [20]. Overexpression of the canonical form of Pax6 in lens fiber cells (Pax6 con transgenics) results in cataracts typified by incomplete lens fiber cell elongation and denucleation, instability of the transcription factor c-Maf and a drastic downregulation of βB1-crystallin expression [21] while overexpression of the Pax6 (5a) splice form also results in cataracts without the changes in cMaf stability [22]. Microarray analysis was previously performed on lenses from both Pax6 (con) transgenics and mice heterozygous for a Pax6 null allele and 13 genes were found to be upregulated in the transgenics and downregulated in the heterozygous knockout mice [23].
One of these genes, paralemmin (Palm), encodes a protein present at the plasma membrane in axons, dendrites and perikarya of differentiating neuronal cell lines, and at high levels in the processes of the cerebellar molecular layer [24]. Further, this gene is downregulated in lenses overexpressing the Pax6(5a) splice variant [25] and the protein is detected in lens cells from both mice and chickens [25,26]. Overexpression of Palm in both neuronal and non-neuronal cell lines initiates the expansion of the plasma membrane and the development of extended processes and microspikes which is dependent on Palm targeting to the cytoplasmic face of the plasma membrane via a palmitoyl group covalently linked near the protein's C-terminus [24,27].
Here we investigate the distribution of Palm in the developing lens and retina, and compare its mRNA levels with two other members of the paralemmin family, paralemmin-2 (Palm-2) and palmdelphin/paralemmin-like (PalmD) [28,29].
Methods
Animals
All experiments using animals were approved by the both the University of Delaware and Albert Einstein College of Medicine Institutional Animal Care Committees and conform to the ARVO statement for the Use of Animals in Ophthalmic and Vision Research. C57Bl/6 mice were generated in-house from breeding stock obtained from Harlan Sprague Dawley (Indianapolis, IN). CD-1 mice were obtained directly from Charles River Laboratories (Wilmington, MA). Embryonic mice were staged by designating noon of the day on which a semen plug was observed in the dam as 0.5 days post-coitum (dpc). Postnatal mice were staged by designating the day of birth as 1 day postnatal (DPN). All mice were maintained in a 12-hour light/dark cycle at 21–24°C and were given food and water ad libitum.
Immunofluorescent detection of Palm in tissue sections
Palm was detected by indirect immunofluorescence following the protocol previously described [30]. Briefly, tissue or embryos were excised from C57Bl/6 mice, embedded in tissue freezing media (TFM, Triangle Biomedical Sciences, Durham, NC) and sectioned at 16 μM on a Leica CM 3050 S Cryostat (Leica, Deerfield, IL). Sections were mounted on Colorfrost-plus™ slides (Fisher Scientific; Pittsburgh, PA), fixed in ice-cold acetone:methanol (1:1 vol/vol) for 15 minutes, dried and blocked with 1% BSA in phosphate buffered saline (PBS), pH 7.4. The blocking solution was removed and the sections incubated with a 1:150 dilution of rabbit polyclonal anti-Palm antibody [24] in 1% BSA-PBS for one hour at room temperature. The bound primary antibody was detected with AlexaFluor 568 goat anti-rabbit IgG (Molecular Probes, Inc. Eugene, OR) and cell nuclei were detected by counter-staining with TO-PRO-3 (1:3000 dilution in 1% BSA-PBS; Molecular Probes, Inc). Negative controls consisted of parallel staining experiments that omitted the primary antibody. Images were captured on a Zeiss LSM 510 Confocal Microscope configured with an Argon/Krypton laser (488 nm and 568 nm excitation lines) and Helium Neon laser (633 nm excitation line)(Carl Zeiss Inc, Göttingen, Germany).
Transfections and reporter assays
Four copies of the Pax6-binding site previously identified in the human PALM promoter [25] were cloned into E4TATA-pGL3 [31] using a synthetic double stranded oligonucleotide 5'-ctagGGCTACTTTCACTCTGCGATGGCAGAGCAGGGCTACTTTCACTCTGCGATGGCAGAGCA-3'. Nucleotides containing Pax6-binding sites are in bold and nucleotides used for subcloning are indicated by lower case letters. Transient transfection assays were performed in 293T cells, which do not express endogenous Pax6 proteins, as described earlier [32].
Immunofluorescent detection of Palm in cultured chick retina
Fertile White Leghorn eggs were obtained from the Department of Animal and Food Sciences at the University of Delaware and kept in a humidified, forced-draft incubator until embryonic day (E) 7. Retinas were dissected in calcium and magnesium-free saline solution (CMF). The neural retina was separated from the pigmented epithelium with fine forceps. The neural retina was minced with fine scissors and incubated in 0.25% trypsin in CMF for 20 minutes at 37°C. Retinas were dissociated into single cells by trituration with a Pasteur pipet in a 0.3 mg/ml soybean trypsin inhibitor/ 0.03 mg/ml DNaseI in Medium 199 (Cellgro, Herndon, Virginia). Cells were plated at a density of 5 × 105 cells / 12 mm diameter round glass coverslip in wells of a 24-well plate in one milliliter of Medium 199 (Cellgro) supplemented with 10% fetal bovine serum. Retina cultures were kept in a standard humidified culture incubator with 5% CO2.
Two days or one week after plating, cultures were fixed in 1% paraformaldehyde in PBS pH 7.4 for 30 minutes and then rinsed in PBS. Cells were then incubated for approximately 1 hour in a mixture of 1:200 rabbit polyclonal anti-chicken Palm [26] and 1:2 mouse monoclonal anti-neurofilament (RT-97) hybridoma supernatant (Developmental Studies Hybridoma Bank, Iowa City, IA; [33,34]) in PBS supplemented with 5% normal goat serum (NGS) and 0.03% Triton X-100 (TX-100). Cultures were rinsed in PBS and then incubated for approximately 1 hour in a mixture of 1:200 Alexa 488-goat anti-rabbit and 1:200 Alexa 594-goat anti-mouse secondary antibodies (both from Molecular Probes, Inc., Eugene OR) in the PBS/NGS/TX-100 mixture. Cultures were then rinsed in PBS and coverslips were mounted on glass slides in a buffered glycerol mounting medium containing ρ-phenylenediamine to retard photo-bleaching. Cultures were observed and photographed using a Nikon Microphot FX epifluorescence microscope equipped with a Nikon DXM-1200 CCD camera. Red and green channel images were merged using Adobe Photoshop.
Real Time RT-PCR
Tissue microdissected from the lens, cerebellum and telencephalon was stored in RNA later (Qiagen, Valencia, California). Total RNA from the lens, cerebellum and forebrain of newborn CD-1 mice was isolated using the RNeasy Protect Mini Kit (Qiagen). Retinal P0, P4 and P22 RNA was kindly provided by Drs. Mike Dorrel and Kenneth Mitton, respectively. DNaseI digestion was performed during RNA isolation with RNase-Free DNase Set (Qiagen). The RNA was quantified with an Agilent 2100 Bioanalyzer and first strand cDNA was then synthesized using 5 μg of RNA, Oligo(dT)12–18 primer and Superscript II RT (Invitrogen, Carlsbad, California) as per manufacturer's instructions. The cDNA was diluted 1:10 and PCR reactions were conducted using 2 μl of cDNA, 50 nm of forward and reverse primers, and 2X SYBR Green PCR Master Mix (Applied Biosystems, Foster City, California). Amplification of the cDNA was performed using a 7900 HP Applied Biosystems Real Time PCR machine. The cDNA was initially denatured at 94°C for 5 minutes, followed by 45 cycles of 94°C for 10 seconds, annealing at 60°C for 20 seconds, and extension at 72°C for 30 seconds. A final extension at 72°C for 5 minutes was then conducted. Each gene was amplified nine times (three times as triplicate experiments). The primers used with Ensembl or NCBI accession numbers follow: Palm (ENSMUSG00000035863) (5' -AGCAGGCAGAGATTGAGAGC-3' and 5' -AGCCAGCGTTCCCTCAGT-3'); Palm2 (NM 172868) (5' -CGCAGGCAGTCTGAAGAAG-3' and 5' -TTTCGAGCGCTTGTATTTCC-3'); PalmD (ENSMUSG00000033377) (5' -AGTAGCTGGAGACGGGACTG-3' and 5' -CACGGCTCTCAGATCACCTT-3'). The housekeeping genes β2-microglobulin, B2M (ENSMUSG00000033376) (5' -TGGTGCTTGTCTCACTGACC-3' and 5' -TATGTTCGGCTTCCCATTCT-3'); Hypoxanthine-guanine phosphoribosyltransferase, HPRT (ENSMUSG00000025630) (5' -GTTGTTGGATATGCCCTTGA-3' and 5' -GGCTTTGTATTTGGCTTTTCC-3'): and succinate dehydrogenase, SDHA (ENSMUSG00000021577) (5'-GAGGAAGCACACCCTCTCATA-3' and 5' -GCACAGTCAGCCTCATTCAA-3') were used for normalization of gene expression levels. Each primer set was designed using Primer3 [35] and specificity verified by NCBI Blast [36]. Standard PCR was then performed to verify amplification of a single PCR product bearing the correct size. The dissociation curve of each PCR amplicon was analyzed using ABI PRISM SDS 2.0 and revealed a single peak, indicating specific PCR amplification [37].
The mRNA levels were normalized to the internal housekeeping gene, B2M and the change in Ct values for each gene (ΔCt) were determined according to the standard method [38,39]. The standard deviation calculated for each sample was less than 5% and was therefore not shown in Figure 5. The primers used had similar efficiencies for amplification as determined by serial dilution experiments [38].
Results and discussion
Previously, we determined that Palm gene expression is downregulated in lenses from mice lacking one copy of the Pax6 gene [25] and upregulated in lenses overexpressing Pax6 [23]. Since potential Pax6 binding sites were identified upstream of the transcriptional start site of Palm [25], Palm may be a direct Pax6 target gene. Thus, we undertook a developmental expression study of Palm in the eye to assess the extent that its expression overlaps that of Pax6.
At 9.5 dpc, Palm immunoreactivity is prominent in the head ectoderm overlying the optic vesicle that is fated to give rise to the lens and corneal epithelium with much lower, but detectable, levels of expression in the optic vesicle (Figure 1A,B,C). At 10.5 dpc, Palm protein is detected at relatively similar levels in the presumptive neural retina, retinal pigmented epithelium (RPE), corneal epithelium and lens vesicle (Figure 1D,E,F). This overlaps well with Pax6 expression in both the optic vesicle and developing lens placode/ vesicle in mice [18,40]. By 11.5 dpc, relative Palm levels have decreased in the presumptive RPE although staining is still detected in both the periocular mesenchyme and presumptive neural retina (data not shown). At 12.5 dpc, intense Palm immunoreactivity is detected at the vitreal surface of the neural retina (Figure 1G, H, I), corresponding to the formation of the ganglion cell processes that will migrate down the optic stalk to form the neural component of the optic nerve [41,42]. This only partially corresponds with Pax6 expression at this stage, since Pax6 expression has been reported in the developing RPE of 13 dpc embryos [43], although the RPE can produce pigment without Pax6 [4]. The presence of Palm in ganglion cell processes at this stage is interesting since Pax6 expression is noted in mature ganglion cells of the adult retina although Pax6 is detected in only a subset of 13 dpc neural retinal precursors [43].
Figure 1 Localization of Palm protein during early mouse eye development. A-C, 9.5 dpc; D-F, 10.5 dpc; G-I, 12.5 dpc; A,D,G Palm; B,E,H cell nuclei stained with ToPro3, C,F,I, merge; Abbreviations- lp, lens placode; ov, optic vesicle; he, head ectoderm; di- lumen of the diencephalon; pce- presumptive corneal epithelium; lv- lens vesicle; pnr- presumptive neural retina; ppe- presumptive retinal pigmented epithelium; L- lens; nr- neural retina; ce- corneal epithelium; os- optic stalk. Arrowheads denote staining in developing neuronal processes that will grow through optic stalk to form the optic nerve. All scale bars are 77 μm. red- Palm; blue-ToPro3 DNA stain.
In the developing mouse lens, Palm expression is seen at both epithelial and fiber cell membranes from 11.5 dpc and is maintained in these cells throughout adulthood (Figure 1D–I; Figure 2). The presence of Palm in all lens cells early in development correlates well with the reported expression pattern of Pax6 in the embryonic lens [18]. During lens maturation, Pax6 expression decreases in lens fiber cells relative to the lens epithelium [21,22,44]. However, newborn rat lens fiber cells still maintain 12% of the levels seen in lens epithelium [45] although Pax6 mRNA levels are 95 fold lower in aged human lens fibers [46]. Since Palm mRNA levels are decreased in lenses from Pax6 heterozygous mice [25] and upregulated in lenses overexpressing Pax6 in lens fiber cells [23], it is plausible that Palm expression is either directly responsive to Pax6 or controlled by genes in the same pathway.
Figure 2 Localization of paralemmin protein in the mouse lens A-C 14.5 dpc; D 14.5 dpc negative control E-G One week post natal; H- 25 weeks postnatal. I- 25 weeks postnatal negative control A,E- paralemmin; B,F- cell nuclei stained with ToPro3; C, D, G, H, I- merge; Scale bars- A-C, 154 μm; D-G, 77 μm; red- paralemmin; blue-ToPro3 DNA stain.
In order to test this proposition functionally, we cloned the Pax6 binding site previously identified in the 5'-flanking region of human PALM [25] in front of a basal promoter and performed transient transfections in 293T cells which lack endogenous Pax6 proteins [32]. Co-transfection of this reporter construct with Pax6 and Pax6(5a) expression vectors activated this artificial promoter 3.4- and 2.1-fold, respectively while addition of both expression vectors simultaneously yielded a reporter activation similar to that of the Pax6 expression vector alone (Figure 3). These levels of Pax6 mediated activation are comparable to those typically obtained in transient transfections with Pax6 responsive promoters [44,47,48]. From these data, it appears likely that the human PALM promoter contains a Pax6-binding site functionally able to interact with both Pax6 and Pax6(5a) consistent with the upregulation of PALM expression in transgenic mice overexpressing Pax6 in the lens and reduced expression of Palm in Pax6 heterozygous lenses [23,25,49]. However, the functional significance of this Pax6 site in the context of the PALM gene is more difficult to ascertain since neither the transcriptional start site nor the functional minimal promoter of PALM have been experimentally investigated. Further studies of PALM/Palm promoters are necessary to fully establish their direct regulation by Pax6 proteins.
Figure 3 Pax6 proteins activate expression from a reporter consisting of four copies of a PAX6-binding site found in the putative 5' flanking sequence of the human PALM gene cloned upstream of the E4 basal promoter. (A) An alignment between the PAX6 site found in the PALM gene and a consensus paired domain Pax6-binding site, P6CON. Non-conserved nucleotides are shown in lower case letters. (B) Results of co-transfections in 293T cells. 200 ng of Pax6 and 25 ng of Pax6(5a) expression plasmids were used as indicated per experiment. The data were normalized using Renilla luciferase [31] and are expressed as a relative ratio of promoter activity in the presence of Pax6 compared to the presence of empty vector, pKW10.
In the developing retina, the intense Palm staining seen in elongating ganglion cell axons at 12.5 dpc downregulated markedly by 14.5 dpc as the development of these processes completes [41] (data not shown). At 16.5 dpc, Palm immunoreactivity is maintained at moderate levels in the cell bodies of both differentiating ganglion cells and undifferentiated neural precursors, but appears slightly stronger in the first morphologically distinguishable axons of the developing inner plexiform layer (ipl) which is composed of cell processes of the neurons of the inner nuclear layer and ganglion cells [50](Figure 4A,B,C). At birth, Palm levels are upregulated in the developing inner plexiform layer (inl) which is in the process of rapid expansion (Figure 4D,E,F). As the development of ipl proceeds, the intensity of Palm staining in this layer drops to that seen in the cell bodies of the ganglion cell and inl (Figure 4G–L). While not as dramatic, localized expression is seen in the developing outer plexiform layer (opl) processes at 1 week pn (Figure 4G–I), although both at that time and in the adult (Figure 4J–L), much less Palm staining is seen on the photoreceptor cell bodies of the outer nuclear layer then in any other retinal layer. Notable, Pax6 expression persists in both retinal ganglion cells and the inner nuclear layer into adulthood, correlating well with the expression pattern of Palm in this tissue [43].
Figure 4 Localization of paralemmin protein during mouse retinal development A-C, 16.5 dpc, arrowheads- emerging inner plexiform layer; D-F 1 day pn; G-I 1 week pn; J-L 2 week pn; A,D,G,J- paralemmin; B,E,H,K- cell nuclei stained with ToPro3; C,F,I,L- merge; Abbreviations- unp- undifferentiated retinal precursors; gc- ganglion cell; ipl- inner plexiform layer; s*- background staining in the sclera; inl- inner nuclear layer; opl- outer plexiform layer; onl- outer nuclear layer. All scale bars are 77 μm. red- paralemmin; blue-ToPro3 DNA stain.
Figure 5 Expression and localization of Palm in chick retinal cultures. Cultures were immunostained with polyclonal anti-Palm (A, D) and RT-97 anti-neurofilament (B, E) antibodies after 2 (A-C) or 7 (D-F) days in culture. For each pair, the merged images are shown in C and F. After 2 days in culture, Palm is present on most cells at cell borders as well as intracellular puncta (A). Fine processes resembling axons (arrows) that are sometimes positive for RT-97 (B) are also labeled. After 7 days in culture, Palm staining appears punctate but more diffuse (D), and does not appear to be localized on the numerous long processes stained with RT-97 (F). Bar in D, 25μm. Green- Palm; Red-RT-97;
In neuronal cell lines, Palm was previously detected at the cell membrane of the cell body and developing axons as well as in a granular localization intracellularly. In vivo, Palm co-purifies with chick brain synaptic plasma membranes consistent with its palmitoylation [24]. While the staining pattern of Palm in the developing mouse retina is consistent with this membrane localization, we wanted to confirm this in dissociated retinal cultures. The neural retina of the E7 chick is at a period of extensive neurogenesis, migration, and process formation in vivo, especially of ganglion cells [51-53]. This ability to extend neurites is also manifest in cultures made from this age retinal tissue [54,55]. Chick retinas were dissociated, plated and stained for Palm either 2 days or 7 days after plating. After 2 days in culture (Figure 5A–C), Palm appears expressed by most cells and is evident at the plasma membrane and as intracellular puncta. Fine processes resembling axons (arrows) that sometimes stain with the anti-neurofilament antibody RT-97 [33] are also positive for Palm immunoreactivity. After 7 days in culture (Figure 5D–F), Palm staining appears punctate but more diffuse in the cell body, and does not appear to be localized on the numerous long processes stained for neurofilament. Thus, like in the mouse retina in vivo, Palm is detected in retinal cultures undergoing active process formation while it is less evident in mature cells, which are undergoing less process extension.
Palm is a member of a multigene family consisting of two other family members, paralemmin-2 (Palm-2) and palmdelphin/paralemmin-like (PalmD/PalmL) [28,29]. Palm2 shares 37% amino acid identity with Palm and like Palm has a C-terminal CaaX motif that could potentially be prenylated. However, the Palm2 gene is alternatively spliced and not all variants contain the prenylation motif. PalmD is 23% identical to Palm but generally lacks a C-terminal prenylation motif although rare splice variants have an alternative C-terminus containing a prenylation motif similar to Palm. Experimentally, the majority of PalmD is cytoplasmic and does not co-purify with plasma membrane fractions [28,29]. Since Palm is potentially able to modulate plasma membrane growth in the lens, retina and brain, while Palm2 and PalmD are of related sequence, we performed quantitative rt-PCR to compare the relative expression levels of all three paralemmin family members in the lens, retina, cerebellum and forebrain.
The ratio between the housekeeping genes tested, B2M, HPRT and SDHA, in the different tissues analyzed was found to range between 0.98–1.04. Since the ratio of an ideal internal control between various tissues would be 1 and the variability of each of our internal normalizing genes between the various tissues assayed was low, we normalized our data to one housekeeping gene, B2M [39,56].
In the lens, cerebellum, forebrain and retina, Palm transcripts are significantly more abundant relative to B2M than those of either Palm2 or PalmD (Figure 6). Notably, Palm mRNA is more abundant in retinas isolated shortly after birth compared to the adult retina, correlating well with the expression of Palm protein detected by immunohistochemistry. Palm is alternatively spliced, and previous western blot analysis of mouse lens protein detected the 60 kDa form of paralemmin [25] which translates from mRNA lacking exon 8 [24]. Parallel qt-PCR analyses of the lens and retina for Palm transcripts harboring exon 8 only detected low levels of this splice variant in all cases (data not shown) which would be translated into a 80 kDa protein. In the lens and forebrain, appreciable Palm2 expression was detected (Ct values of about 21.5) while Palm2 levels are relatively low in all post natal retinal samples tested (Ct values of about 28.5). PalmD transcripts were usually present at low levels in the tissues examined with Ct values of about 26. Co-expression of Palm with Palm2 in tissues examined will aid to the interpretation of gene targeting studies of this family of genes.
Figure 6 Relative levels of Palm, Palm2 and PalmD transcripts in the lens, cerebellum forebrain and retina. All data are expressed as a relative to the amount of B2M in the sample.
Conclusion
The lens and retina express paralemmin during development with its transient upregulation during the formation of optic nerve and formation of both plexiform layers. Further, the putative PALM promoter contains a functional Pax6 binding site and the developmental expression pattern of Palm in the eye generally correlates well with that reported for Pax6, leading credence to the idea that Palm is a Pax6 directly-regulated gene.
Abbreviations
dpc, days post coitum; PBS- phosphate buffered saline; inl, inner nuclear layer; onl; outer nuclear layer; opl, outer plexiform layer; ipl; inner plexiform layer; pn, post-natal; rpe, retinal pigmented epithelium.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MC carried out all of the immunohistochemical studies on tissue and was involved in the initial drafting of the manuscript. LVW carried out all of the quantitative rt-PCR assays and BKC performed the transfection assays. DSG analyzed PALM expression in chick retinal cultures and MWK was involved in the experimental design and its interpretation. AC conceived of the molecular experiments and participated in their design and interpretation. MKD imaged all of the immunohistochemical data, was involved in its interpretation and drafted the manuscript at all stages of the submission process.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank the staff of the Albert Einstein College of Medicine Biotechnology Center for the qPCR analysis, Dr. Jean Hebert of AECOM for helpful suggestions. Dr. Harry Maisel for the anti-chick PALM antibody, Drs. M. Busslinger and R. Maas respectively for the Pax6 and PAX6(5a) expression vectors and Dr. Kirk Czymmek of the University of Delaware Core Imaging Facility for confocal microscopy support. This work was funded by National Eye Institute grants EY015279 and EY012221 to MKD and EY12200 and EY14237 to AC; National Institute of Neurological Diseases and Stroke grant NS40317 to DSG and INBRE program grant P20 RR16472 supporting the University of Delaware Core Imaging facility.
==== Refs
Chow RL Lang RA Early eye development in vertebrates Annu Rev Cell Dev Biol 2001 17 255 296 11687490 10.1146/annurev.cellbio.17.1.255
Ochi H Sakagami K Ishii A Morita N Nishiuchi M Ogino H Yasuda K Temporal expression of L-Maf and RaxL in developing chicken retina are arranged into mosaic pattern Gene Expr Patterns 2004 4 489 494 15261825 10.1016/j.modgep.2004.03.005
Lecoin L Sii-Felice K Pouponnot C Eychene A Felder-Schmittbuhl MP Comparison of maf gene expression patterns during chick embryo development Gene Expr Patterns 2004 4 35 46 14678826 10.1016/S1567-133X(03)00152-2
Collinson JM Quinn JC Hill RE West JD The roles of Pax6 in the cornea, retina, and olfactory epithelium of the developing mouse embryo Dev Biol 2003 255 303 312 12648492 10.1016/S0012-1606(02)00095-7
Dyer MA Livesey FJ Cepko CL Oliver G Prox1 function controls progenitor cell proliferation and horizontal cell genesis in the mammalian retina Nat Genet 2003 34 53 58 12692551 10.1038/ng1144
Duncan MK Cui W Oh DJ Tomarev SI Prox1 is differentially localized during lens development Mech Dev 2002 112 195 198 11850194 10.1016/S0925-4773(01)00645-1
Li X Ma W Barker JL Piatigorsky J Transient expression of glutamate decarboxylase and gamma-amino butyric acid in embryonic lens fibers of the rat Dev Dyn 1995 203 448 455 7496036
Giger RJ Wolfer DP De Wit GM Verhaagen J Anatomy of rat semaphorin III/collapsin-1 mRNA expression and relationship to developing nerve tracts during neuroembryogenesis J Comp Neurol 1996 375 378 392 8915837 10.1002/(SICI)1096-9861(19961118)375:3<378::AID-CNE3>3.0.CO;2-#
Baechner D Liehr T Hameister H Altenberger H Grehl H Suter U Rautenstrauss B Widespread expression of the peripheral myelin protein-22 gene (PMP22) in neural and non-neural tissues during murine development J Neurosci Res 1995 42 733 741 8847735 10.1002/jnr.490420602
Claudio JO Lutchman M Rouleau GA Widespread but cell type-specific expression of the mouse neurofibromatosis type 2 gene Neuroreport 1995 6 1942 1946 8547603
Tasheva ES Ke A Deng Y Jun C Takemoto LJ Koester A Conrad GW Differentially expressed genes in the lens of mimecan-null mice Mol Vis 2004 10 403 416 15215744
Frederikse PH Yun E Kao HT Zigler JSJ Sun Q Qazi AS Synapsin and synaptic vesicle protein expression during embryonic and post-natal lens fiber cell differentiation Mol Vis 2004 10 794 804 15529118
Xi J Farjo R Yoshida S Kern TS Swaroop A Andley UP A comprehensive analysis of the expression of crystallins in mouse retina Mol Vis 2003 9 410 419 12949468
Magabo KS Horwitz J Piatigorsky J Kantorow M Expression of betaB(2)-crystallin mRNA and protein in retina, brain, and testis Invest Ophthalmol Vis Sci 2000 41 3056 3060 10967064
Zampighi GA Eskandari S Kreman M Epithelial organization of the mammalian lens Exp Eye Res 2000 71 415 435 10995562 10.1006/exer.2000.0895
Kuszak JR The ultrastructure of epithelial and fiber cells in the crystalline lens Int Rev Cytol 1995 163 305 350 8522422
Valtorta F Leoni C Molecular mechanisms of neurite extension Philos Trans R Soc Lond B Biol Sci 1999 354 387 394 10212488 10.1098/rstb.1999.0391
Grindley JC Davidson DR Hill RE The role of Pax-6 in eye and nasal development Development 1995 121 1433 1442 7789273
Marquardt T Ashery-Padan R Andrejewski N Scardigli R Guillemot F Gruss P Pax6 is required for the multipotent state of retinal progenitor cells Cell 2001 105 43 55 11301001 10.1016/S0092-8674(01)00295-1
Collinson JM Quinn JC Buchanan MA Kaufman MH Wedden SE West JD Hill RE Primary defects in the lens underlie complex anterior segment abnormalities of the Pax6 heterozygous eye Proc Natl Acad Sci U S A 2001 98 9688 9693 11481423 10.1073/pnas.161144098
Duncan MK Xie L David LL Robinson ML Taube JR Cui W Reneker LW Ectopic Pax6 expression disturbs lens fiber cell differentiation Invest Ophthalmol Vis Sci 2004 45 3589 3598 15452066 10.1167/iovs.04-0151
Duncan MK Kozmik Z Cveklova K Piatigorsky J Cvekl A Overexpression of Pax-6 (5a) in lens fiber cells results in cataract and upregulation of a5b1 integrin expression. J Cell Sci 2000 113 3173 3185 10954416
Chauhan BK Reed NA Yang Y Cermak L Reneker L Duncan MK Cvekl A A comparative cDNA microarray analysis reveals a spectrum of genes regulated by Pax6 in mouse lens Genes Cells 2002 7 1267 1283 12485166 10.1046/j.1365-2443.2002.00602.x
Kutzleb C Sanders G Yamamoto R Wang X Lichte B Petrasch-Parwez E Kilimann MW Paralemmin, a prenyl-palmitoyl-anchored phosphoprotein abundant in neurons and implicated in plasma membrane dynamics and cell process formation J Cell Biol 1998 143 795 813 9813098 10.1083/jcb.143.3.795
Chauhan BK Reed NA Zhang W Duncan MK Kilimann MW Cvekl A Identification of genes downstream of Pax6 in the mouse lens using cDNA microarrays J Biol Chem 2002 277 11539 11548 11790784 10.1074/jbc.M110531200
Bagchi M Katar M Lo WK Maisel H Paralemnin of the lens J Cell Biochem 2003 89 917 921 12874826 10.1002/jcb.10574
Gauthier-Campbell C Bredt DS Murphy TH El-Husseini Ael D Regulation of dendritic branching and filopodia formation in hippocampal neurons by specific acylated protein motifs Mol Biol Cell 2004 15 2205 2217 14978216 10.1091/mbc.E03-07-0493
Andreu N Escarceller M Feather S Devriendt K Wolf AS Estivill X Sumoy L PALML, a novel paralemmin-related gene mapping on human chromosome 1p21 Gene 2001 278 33 40 11707320 10.1016/S0378-1119(01)00719-3
Hu B Copeland NG Gilbert DJ Jenkins NA Kilimann MW The paralemmin protein family: identification of paralemmin-2, an isoform differentially spliced to AKAP2/AKAP-KL, and of palmdelphin, a more distant cytosolic relative Biochem Biophys Res Commun 2001 285 1369 1376 11478809 10.1006/bbrc.2001.5329
Reed NA Oh D Czymmek KJ Duncan MK An immunohistochemical method for the detection of proteins in the vertebrate lens J Immunol Methods 2001 253 243 252 11384685 10.1016/S0022-1759(01)00374-X
Chauhan BK Yang Y Cveklova K Cvekl A Functional interactions between alternatively spliced forms of Pax6 in crystallin gene regulation and in haploinsufficiency Nucleic Acids Res 2004 32 1696 1709 15020706 10.1093/nar/gkh334
Yang Y Chauhan BK Cveklova K Cvekl A Transcriptional regulation of mouse alphaB- and gammaF-crystallin genes in lens: opposite promoter-specific interactions between Pax6 and large Maf transcription factors J Mol Biol 2004 344 351 368 15522290 10.1016/j.jmb.2004.07.102
Shin DH Lee KS Lee E Cho SS Kim J Kim JW Kwon BS Lee HW Lee WJ The correspondence between the labeling patterns of antibody RT97, neurofilaments, microtubule associated protein 1B and tau varies with cell types and development stages of chicken retina Neurosci Lett 2003 342 167 170 12757891 10.1016/S0304-3940(03)00276-3
Anderton BH Breinburg D Downes MJ Green PJ Tomlinson BE Ulrich J Wood JN Kahn J Monoclonal antibodies show that neurofibrillary tangles and neurofilaments share antigenic determinants Nature 1982 298 84 86 6178036 10.1038/298084a0
Primer 3
NCBI Blast
Schmittgen TD Zakrajsek BA Mills AG Gorn V Singer MJ Reed MW Quantitative reverse transcription-polymerase chain reaction to study mRNA decay: comparison of endpoint and real-time methods Anal Biochem 2000 285 194 204 11017702 10.1006/abio.2000.4753
Pfaffl MW A new mathematical model for relative quantification in real-time RT-PCR Nucleic Acids Res 2001 29 e45 11328886 10.1093/nar/29.9.e45
Microbiology and Immunology On-Line: Real Time PCR
Ashery-Padan R Marquardt T Zhou X Gruss P Pax6 activity in the lens primordium is required for lens formation and for correct placement of a single retina in the eye Genes Dev 2000 14 2701 2711 11069887 10.1101/gad.184000
Pei YF Rhodin JAG The prenatal development of the mouse eye Anat Rec 1970 168 105 126 5469558 10.1002/ar.1091680109
Barnstable CJ A molecular view of vertebrate retinal development Mol Neurobiol 1987 1 9 46 3077059
de Melo J Qiu X Du G Cristante L Eisenstat DD Dlx1, Dlx2, Pax6, Brn3b, and Chx10 homeobox gene expression defines the retinal ganglion and inner nuclear layers of the developing and adult mouse retina J Comp Neurol 2003 461 187 204 12724837 10.1002/cne.10674
Duncan MK HaynesII JI Cvekl A Piatigorsky J Dual roles for Pax-6: a transcriptional repressor of lens fiber-cell specific b-crystallin genes Mol Cell Biol 1998 18 5579 5586 9710641
Yang Y Cvekl A Tissue-specific regulation of the mouse or A-crystallin gene in lens via recruitment of Pax6 and c-Maf to its promoter. J Mol Biol 2005
Zhang W Cveklova K Oppermann B Kantorow M Cvekl A Quantitation of PAX6 and PAX6(5a) transcript levels in adult human lens, cornea, and monkey retina Mol Vis 2001 7 1 5 11172136
Chauhan BK Yang Y Cveklova K Cvekl A Functional properties of natural human PAX6 and PAX6(5a) mutants Invest Ophthalmol Vis Sci 2004 45 385 392 14744876 10.1167/iovs.03-0968
Goudreau G Petrou P Reneker LW Graw J Loster J Gruss P Mutually regulated expression of Pax6 and Six3 and its implications for the Pax6 haploinsufficient lens phenotype Proc Natl Acad Sci U S A 2002 99 8719 8724 12072567 10.1073/pnas.132195699
Calderwood DA Tuckwell DS Eble J Kuhn K Humphries MJ The integrin a1 A-domain is a ligand binding site for collagens and laminin J Biol Chem 1997 272 12311 12317 9139675 10.1074/jbc.272.19.12311
Finlay BL Sengelaub DR Blakemore C Development of the vertebrate retina Perspectives in vision research 1989 New York , Plenum Press
Coulombre AJ Correlations of structural and biochemical changes in the developing retina of the chick Am J Anat 1955 96 153 189 14361308 10.1002/aja.1000960106
Meller K Tetzlaff W Scanning electron microscopic studies on the development of the chick retina Cell Tissue Res 1976 170 145 159 954051 10.1007/BF00224296
Prada C Puga J Perez-Mendez L Lopez R Ramirez G Spatial and Temporal Patterns of Neurogenesis in the Chick Retina Eur J Neurosci 1991 3 559 569 12106488
Thompson JM Lu AH Ruch S Developmental regulation of chick embryo retina neurite extension by extrinsic factors Brain Res Bull 1987 18 479 484 3607520 10.1016/0361-9230(87)90112-2
Linser PJ Perkins MS Regulatory aspects of the in vitro development of retinal Muller glial cells Cell Differ 1987 20 189 196 3568138 10.1016/0045-6039(87)90433-7
Vandesompele J De Preter K Pattyn F Poppe B Van Roy N De Paepe A Speleman F Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes Genome Biol 2002 3 RESEARCH0034 12184808 10.1186/gb-2002-3-7-research0034
|
15969763
|
PMC1183217
|
CC BY
|
2021-01-04 16:36:32
|
no
|
BMC Ophthalmol. 2005 Jun 21; 5:14
|
utf-8
|
BMC Ophthalmol
| 2,005 |
10.1186/1471-2415-5-14
|
oa_comm
|
==== Front
BMC OphthalmolBMC Ophthalmology1471-2415BioMed Central London 1471-2415-5-151597210610.1186/1471-2415-5-15Research ArticleThe combination of intravitreal triamcinolone and phacoemulsification surgery in patients with diabeticfoveal oedema and cataract Habib Maged S [email protected] Paul S [email protected] David HW [email protected] Sunderland Eye Infirmary, Queen Alexandra Road, SR2 9HP, UK2005 22 6 2005 5 15 15 18 2 2005 22 6 2005 Copyright © 2005 Habib 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 management of diabetic patients with refractory macular oedema or patients with no adequate pre-operative view to administer laser treatment provide a challenge to the ophthalmologist. We wished to assess the use, safety and effect of intravitreal triamcinolone injection at the time of cataract surgery in patients with diabetic foveal oedema and sight limiting lens opacities.
Method
This was a longitudinal non-randomised prospective pilot study in 18 eyes (12 patients). All patients had visually significant lens opacities and either persistent diabetic foveal oedema unresponsive to laser treatment-group A, or foveal oedema with no adequate pre-operative view for laser treatment- group B. The cataract surgery was carried out under full aseptic technique using a self-sealing temporal incision and a foldable acrylic lens. Intravitreal triamcinolone was given infratemporally pars plana at the completion of the cataract surgery. The patients were reviewed at day 5, 2 weeks, 2 months and then every 3 months as required. The Wilcoxin matched-pairs test was used to assess the significance of the improvement in visual acuity at 2 months.
Results
Twelve patients with a total of 18 eyes were included in the study. There were 10 patients (15 eyes) in group A and 3 patients (3 eyes) in group B. Preoperatively 16 of the 18 eyes had a visual acuity of 6/24 or worse. Postoperatively 83% of patients had completely dry foveae at 2 weeks. Best-corrected visual acuities at two months review ranged from 6/6 to CF with 9 eyes (50%) achieving 6/12 or better (7 eyes (47%) in group A and 2 eyes (67%) in group B). Three eyes had no recorded improvement in visual acuity, but no eyes had deterioration in acuity. The improvement in visual acuity was significant at p = 0.001. There were no significant sight threatening complications.
Conclusion
Intravitreal triamcinolone has been shown to lead to an improvement in macular oedema and visual improvement in diabetic patients not undergoing cataract surgery but has not, to our knowledge, been previously used in a study like this one.
We suggest that intravitreal injection at the time of cataract surgery could be carried out safely with encouraging visual outcomes in patients with diabetic foveal oedema and cataract.
==== Body
Background
Diabetes mellitus is the most common predisposing risk factor for cataracts in the developed world with a three to four fold increased prevalence in people under 65 [1]. It is an important cause of reduced vision [2] and it has been estimated that up to 20% of all cataract surgery is performed on diabetic patients [3].
Cataract surgery in patients with diabetic retinopathy is associated with an increased risk of a number of problems including uveitis [4], posterior capsule opacity [5], and anterior capsule phimosis [6]. Some studies have found an increased risk of retinopathy progression and in particular macular oedema exacerbation with cataract surgery [7-12]. Cataracts can also impair the recognition of sight threatening retinopathy and obstruct treatment of maculopathy.
It is thought that the incidence of these complications and the visual outcome is related to the severity of the retinopathy and its activity at the time of surgery as well as glycaemic control [12,13]. Patients with no retinopathy have an excellent prognosis [14,15]. However, patients with macular oedema have a poor visual prognosis [14-21]. Macular oedema presenting de novo immediately following cataract surgery can spontaneously resolve with a good result, and patients with previously successfully treated macular oedema do well with cataract surgery [22]. However persistent macular oedema despite laser or macular oedema either undetected or untreatable prior to surgery does not resolve spontaneously. These patients have a substantially impaired prognosis [14-21].
Although laser photocoagulation can be applied post-operatively it can be difficult to perform immediately following surgery, because of problems such as photophobia, contact lens intolerance, poor mydriasis, intraocular lens deposits and edge effects, posterior capsule opacity or vitreous haemorrhage. Intra-operative pan retinal photocoagulation (PRP) can be applied with an indirect ophthalmoscope, however intra-operative focal laser applied in this way is imprecise. It is recognised that the increased blood ocular barrier breakdown associated with pan retinal photocoagulation can potentially exacerbate macular oedema [23].
Intravitreal triamcinolone has been successfully used to treat patients with refractory diffuse diabetic macular oedema [24-26] and patients with refractory pseudophakic macular oedema [27,28]. It has also been shown to reduce blood ocular barrier break down following PRP [29].
We designed an initial pilot study to assess the ease of use, safety and effect of intravitreal triamcinolone injection at the time of cataract surgery in patients with sight limiting cataract and diabetic foveal macular oedema.
Methods
We carried out a longitudinal pilot study of 12 patients (18 eyes) undergoing phacoemulsification cataract surgery. Inclusion criteria was visually significant cataract with either:
a) Pre-existing and persistent diabetic foveal oedema unresponsive to laser treatment as per ETDRS recommendations [30] – Group A.
b) Foveal oedema with no adequate view to perform pre-operative laser – Group B.
Exclusion criteria included a known increased intraocular pressure response to steroids and glaucoma. Fully informed consent was taken. Patients were examined immediately prior to surgery by recording the best-corrected visual acuity, intraocular pressure and slit lamp biomicroscopy using a 66 dioptre fundal lens. Wisconsin retinopathy grade [31] was recorded at each visit.
Surgical technique included full asepsis with povidone-iodine washout of conjunctival sac pre-operatively and subconjunctival cefuroxime at the completion of surgery. All surgeries were completed under topical anaesthesia with a self-sealing temporal corneal tunnel and a 6 mm folding acrylic lens (SA60- Alcon).
At the completion of cataract surgery all patients had 4 mgs in 0.1 ml of triamcinolone acetate injected via the inferotemporal pars plana, using a 27-gauge needle. The injection was directed into the inferior vitreous cavity to reduce the incidence of visually disturbing floaters post-operatively and was given inferior to the temporal corneal tunnel to avoid anterior chamber shallowing during scleral perforation of the needle. Patients were seen on day five, two weeks, two months and then three monthly depending on clinical assessment. Best-corrected visual acuities were obtained at two months post-operatively with refraction.
Results
There were 12 patients with a total of 18 eyes. There were 10 patients (15 eyes) in group A and 3 patients (3 eyes) in group B. There was one patient with one eye in group A and one eye in group B. There were 8 females and 4 males with a mean age of 69 years. Eleven of the twelve patients were type two diabetics of which three were taking insulin. Mean Haemoglobin A1C at the time of surgery was 8.7% (range 7.2–11.4%) and mean duration of diabetes was 9 years (ranges 2–25 years). All of the 11 type two diabetic patients were being treated for hypertension.
Pre-operative visual acuities ranged from 6/12 to hand movements secondary to a combination of cataract and maculopathy, with 16 of the 18 eyes having a visual acuity of 6/24 or worse. Two patients in group B had intra-operative pan retinal photocoagulation (but not focal laser) applied during phacoemulsification surgery. [Table 1]
Best-corrected visual acuities at two months post-operatively ranged from 6/6 to CF with 9 eyes (50%) achieving 6/12 or better (7 eyes (47%) in group A and 2 eyes (67%) in group B). No eyes had deterioration in visual acuity but 3 eyes (patients 4 and 6) had no recorded improvement in acuity despite the fovea being clinically dry in two of these cases (patient 6). [Table 2] The improvement in visual acuity was significant at p = 0.001 using a Wilcoxon matched-pairs test.
Fifteen eyes (83%) had clinically dry foveae at two weeks post surgery – 12 (80%) in group A and all 3 (100%) in group B. Last recorded visual acuities at a mean time of 8.5 months showed a deterioration from best-recorded visual acuity in 5 (27%) patients because of recurrent macular oedema – all in group A. [Table 2]
Six (33%) eyes; all in group A; showed a one-step or more deterioration in retinopathy grade at last follow up (mean follow up 8.5 months, range 4–15 months) compared with pre-operative levels. All cases of progression occurred five or more months post-operatively. [Table 3]
The triamcinolone was quick and easy to give and there were no immediate operative complications from its administration such as haemorrhage, capsule tears or passage of the triamcinolone into the anterior chamber. Post-operatively 4 patients had transiently raised intraocular pressure requiring treatment with ocular hypotensive drops. None of these patients had pre -existing raised intraocular pressure or were known steroid responders. All 4 patients had normal pressures at the first visit but 2 developed raised pressure at two weeks(IOPs 34 and 29 mmHg) and the other 2 by 2 months (IOPs 31 and 24 mmHg). [Table 3] Three patients (patients 2, 4 and 8) were treated successfully with topical beta blockers as a monotherapy and 1 (patient 5) required combination therapy with a topical beta blocker and a topical carbonic anhydrase inhibitor. At 6 months post-operatively all pressures were normal off treatment. Interestingly 2 patients (patients 2 and 4) who had both eyes operated upon, developed high pressure in their second eyes after having had no pressure rise in their first eye.
Three patients complained of visual floaters post-operatively which settled within a few days. The retinal view was unimpeded and accurate clinical examination was possible. There were no cases of endophthalmitis, uveitis, significant capsule phimosis or posterior capsule opacity in the follow-up period. [Table 3]
Discussion
This pilot study suggests that intravitreal triamcinolone can be given safely and easily at the time of phacoemulsification surgery.
We took care to inject the triamcinolone inferotemporally away from the visual axis and anterior chamber to avoid visually troublesome floaters and transit of triamcinolone into the anterior chamber. The retinal view was unimpaired post-operatively allowing accurate retinal assessment and further laser treatment if needed. Similar to other studies [32,33] we found only 22% of eyes developed increased intraocular pressure and all were treated successfully with topical ocular hypotensive agents with spontaneous improvement with time.
Intravitreal triamcinolone has been shown to lead to an improvement in macular oedema and visual improvement in diabetic patients not undergoing cataract surgery [24-26] but has not; to our knowledge; been previously studied in a series such as this.
Combining cataract surgery with triamcinolone rather than giving triamcinolone before surgery as a separate procedure avoided the potential for progression of lens opacities associated with intraocular steroids [34,35] which could have further interfered with retinopathy assessment. We had no cases of endophthalmitis, which can occur with triamcinolone injections [36,37]. Combining the two procedures reduces the patient's potential risk of endophthalmitis from two separate intraocular episodes to one, whilst at the same time offering improved patient convenience. The technique was simple adding very little time to the procedure and in this series there was no significant ocular morbidity associated with the triamcinolone. We choose to combine the procedure with our standard clear corneal temporal incision phacoemulsification technique. There is debate regarding the possibility of an increased risk of endophthalmitis with temporal clear corneal incisions [38]. However this has not been our experience. We have had no increase in our rate of endophthalmitis over the last five years during which a clear corneal temporal approach has been adopted by all surgeons at our unit. Incidence of endophthalmitis in our unit for 1999 was 0.11% and for 2004 was 0.09%, based on approximately 6000 cases /year [39]. It is possible that other factors such as wound construction and lid draping and preparation are more important than incision position itself [38]. All the cases in this series were done in theatre with full asepsis, topical povidone pre-operatively, careful lid draping and carefully constructed wounds, which were watertight at the close of surgery. These are important factors in avoiding infective complications.
We had no cases of posterior capsule rupture in this series. Triamcinolone has been used to help visualise vitreous during posterior capsule rupture and anterior vitrectomy [40] and anecdotally in cases of phacoemulsification surgery with posterior capsule rupture to reduce the incidence of post operative cystoid macular oedema and postoperative inflammation [41]. Potentially therefore intravitreal triamcinolone administration could be considered even if posterior capsule rupture was to occur although we have no experience of this.
The natural history of patients with foveal oedema at the time of cataract surgery is recognised as being poor and the patients in the study had a number of other features associated with a particularly poor prognosis after cataract surgery- increasing age, female sex, poor glycaemic control with high Haemoglobin A 1C (%) at the time of surgery and moderate to severe background retinopathy changes, have all been associated with a poor prognosis in other studies [12,13,16,18]. All the patients in group A had chronic macular oedema prior to cataract surgery, which had been unresponsive to treatment. Indeed patients such as these with chronic unresponsive macular oedema, particularly if there is only a moderate degree of cataract, are often declined surgery on the basis that the maculopathy would limit the underlying visual acuity, which can also deteriorate with surgery. Despite this approximately 50% of these patients achieved 6/12 vision.
Patients who present with dense cataracts and significant retinopathy, especially maculopathy, which is untreatable pre-operatively because of the lens opacities, pose a difficult clinical scenario. Laser can be performed post-operatively but this can be difficult, for reasons previously stated and surgically induced inflammation. There were three patients in the study in this group – group B. Two of these patients had very severe non proliferative retinopathy, in addition to maculopathy, and were treated with intra-operative PRP which can also exacerbate macular oedema. Despite these difficulties and risk factors for maculopathy exacerbation all three patients had complete macular oedema resolution at the two week examination without any macular laser having been applied at that stage.
Overall fifteen (83%) of the patients had complete resolution of their macular oedema at two weeks follow up. Recurrence of macular oedema occurred in some patients in this study as would be expected from the known short term effect of intravitreal triamcinolone and other studies with intravitreal triamcinolone and diabetic macular oedema. However the triamcinolone clearly prevented the short term exacerbation of macular oedema that can be associated with blood ocular breakdown due to intraocular surgery and PRP [23,29,42]. It seems logical to use a drug, albeit with a known short-lived effect, in this way to potentially improve visual outcome until longer lasting alternatives are produced.
Retinopathy progression occurred during follow up in only six patients (none within 5 months of surgery and 4 out of 6 at more than 10 months following surgery) and it may be that triamcinolone has a role in reducing the deterioration of retinopathy that has been reported following cataract surgery especially in patients with more advanced retinopathy. It may also have a role in the inhibition of retinal neovascularisation [43,44]. Numbers in this uncontrolled pilot study are too limited to draw any definitive conclusions.
Triamcinolone was only injected in those eyes with pre-existing macular oedema at the time of cataract surgery. It was not used in eyes thought to be at risk of developing macular oedema after surgery in those with clinically dry foveae at the time of surgery. The natural history of patients with dry maculae at the time of surgery who develop macular oedema following surgery is relatively good [22] and it was felt that the risks associated with triamcinolone in that group would outweigh the benefits. At present we are not recommending triamcinolone in that group of patients, although this may merit further investigation.
Conclusion
This pilot study suggests that intravitreal triamcinolone can be given safely and easily at the time of phacoemulsification surgery in patients with visually significant cataract and diabetic foveal macular oedema. Eighty-three percent of the eyes in this study had complete resolution of macular oedema at two weeks post-operatively and none experienced any exacerbation of their maculopathy. Further controlled studies are needed to demonstrate the potential visual advantages and long-term benefit of this treatment approach.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DS conceived of the study, participated in the design of the study and helped to draft the manuscript. MH participated in the design and the coordination of the study. PC participated in the coordination of the study 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 Pre-operative clinical characteristics of patients.
Group Patient number Age (years) DM type Pre-op PRP Pre-op Va Wisconsin Retinopathy Grade Pre-op
A 1 65 II No 6/60 Moderate
No CF Moderate
A 2 77 II No 6/60 Moderate
Yes CF LRPDR
A 3 73 II No 6/24 Mild
A 4 71 II No 6/60 Moderate
No CF Severe
A 5 67 II Yes HM HRPDR
A 6 72 II No 6/36 Moderate
No 6/24 Moderate
A 7 76 II No 6/24 Mild
A 8 68 II No 6/12 Mild
No 6/18 Moderate
A 9 71 II No 6/60 Severe
A 10 67 II No 6/36 Moderate
B 10 No (intra-op) HM Very Severe
B 11 69 II No (intra-op) HM Very Severe
B 12 49 I No 6/36 Mild
Mean age = 68.75
DM = Diabetes Mellitus; I = type one diabetes; II = type two diabetes; Pre-op = pre-operatively; PRP= pan retinal photocoagulation; Va = visual acuity; intra-op = intra-operatively; CF = counting fingers; HM = hand movements; LRPDR = low risk proliferative diabetic retinopathy; HRPDR= high risk proliferative diabetic retinopathy
Table 2 The post-operative clinical and visual outcomes of patients.
Group Patient Number Refracted Va two months post surgery Last Va Macula clinically dry at two weeks Macula clinically dry at last FU
A 1 6/9 6/18 Yes Yes
6/9 6/12 Yes Yes
A 2 6/12 6/12 Yes Yes
6/36 6/36 No No
A 3 6/18 6/18 Yes Yes
A 4 6/9 6/12 Yes No
CF CF No No
A 5 6/60 6/60 No No
A 6 6/36 6/36 Yes No
6/24 6/36 Yes No
A 7 6/6 6/6 Yes Yes
A 8 6/9 6/9 Yes Yes
6/9 6/12 Yes Yes
A 9 6/24 6/24 Yes No
A 10 6/18 6/18 Yes Yes
B 10 6/18 6/18 Yes No
B 11 6/12 6/12 Yes Yes
B 12 6/6 6/6 Yes Yes
Va = visual acuity; FU = follow up; CF = counting fingers
Table 3 Post-operative complications and deterioration in retinopathy grade on follow-up.
Group Patient Number Follow up (months) Raised IOP Stepwise deterioration in retinopathy grade at last FU (ETDRS)
A 1 15 3
14 1
A 2 6 0
8 Yes: 34 mmHg 0
A 3 9 0
A 4 10 0
12 Yes: 29 mm Hg 1
A 5 11 Yes: 31 mm Hg 0
A 6 10 1
7 0
A 7 12 0
A 8 5 Yes: 24 mmHg 1
5 0
A 9 7 1
A 10 6 0
B 10 4 0
B 11 8 0
B 12 4 0
Mean FU = 8.5
IOP = intraocular pressure; FU = follow-up; ETDRS = Early Treatment Diabetic Retinopathy Study
==== Refs
Ederer F Hiller R Taylor HR Senile lens changes and diabetes in two population studies Am J Ophthalmol 1981 91 381 395 7211996
Klein R Klein BE Moss SE Visual impairment in diabetes Ophthalmology 1984 91 1 9 6709312
Hamilton AMP Ulbig MW Polkinghorne P eds Epidemiology of diabetic retinopathy Management of diabetic retinopathy 1996 London: BMJ Publishing Group 1 15
Hykin PG Gregson RM Hamilton AM Extracapsular cataract extraction in diabetics with rubeosis iridis Eye 1992 6 296 299 1446764
Ionides A Dowler JG Hykin PG Rosen PH Hamilton AM Posterior capsule opacification following diabetic extracapsular cataract extraction Eye 1994 8 5353 537
Kato S Oshika T Numaga J Anterior capsular contraction after cataract surgery in eyes of diabetic patients Br J Ophthalmol 2001 85 21 3 11133706 10.1136/bjo.85.1.21
Sadiq SA Sleep T Amoaku WM The visual results and changes in retinopathy in diabetic patients following cataract surgery Eur J Ophthalmol 1999 9 14 20 10230587
Chung J Kim MY Kim HS Yoo JS Lee YC Effect of cataract surgery on the progression of diabetic retinopathy J Cataract Refract Surg 2002 28 626 630 11955902 10.1016/S0886-3350(01)01142-7
Royal College of Ophthalmologists Guidelines for diabetic retinopathy 2002 London: RCO
Cunliffe IA Flanagan DW George NDL Extra capsular cataract surgery with lens implantation in diabetics with and without proliferative retinopathy Br J Ophthalmol 1991 75 9 12 1991094
Pollack A Dotan S Oliver M Progression of diabetic retinopathy after cataract extraction Br J Ophthalmol 1991 75 547 551 1911658
Henricsson M Heijl A Janzon L Diabetic retinopathy before and after cataract surgery Br J Ophthalomol 1996 80 789 793
Jaffe GJ Burton TC Progression of non proliferative diabetic retinopathy following cataract extraction Arch Ophthalmol 1988 106 745 749 3369997
Squirrell D Bhola R Bush J Winder S Talbot JF A prospective, case controlled study of the natural history of diabetic retinopathy and maculopathy after uncomplicated phacoemulsification cataract surgery in patients with type 2 diabetes Br J Ophthalmol 2002 86 565 571 11973256 10.1136/bjo.86.5.565
Dowler JGF Hykin PG Lightman SL Visual acuity following extracapsular cataract extraction in diabetes: a meta-analysis Eye 1995 9 313 317 7556739
Zaczek A Olivestedt G Zetterstrom C Visual outcome after phacoemulsification and IOL implantation in diabetic patients Br J Ophthalmol 1999 83 1036 1041 10460771
Krepler K Biowski R Schrey S Jandrasits K Wedrich A Cataract surgery in patients with diabetic retinopathy: visual outcome, progression of diabetic retinopathy, and incidence of diabetic macular oedema Graefes Arch Clin Exp Ophthalmol 2002 240 735 738 12271370
Chew E Benson WE Remaley N Results after lens extraction in patients with diabetic retinopathy. ETDRS report no 25 Arch Ophthalmol 1999 117 1600 1606 10604663
Chiu D Meusemann RA Kaufman DV Visual outcome and progression of retinopathy after cataract surgery in diabetic patients Aust NZ J Ophthalmol 1998 26 129 133
Benson WE Brown GC Tasman W Extracapsular cataract extraction with placement of a posterior chamber lens in patients with diabetic retinopathy Ophthalmology 1993 100 730 738 8493017
Ancliffe RJ Poulson A Flanagan DW Phacoemulsification in diabetics Eye 1996 10 737 741 9091373
Dowler JGF Sehmi KS Hykin PG The natural history of macular oedema after cataract surgery in diabetes Ophthalmology 1999 106 663 668 10201584 10.1016/S0161-6420(99)90148-3
Zweng EC Little HL Hammond AM Complications of argon laser photocoagulation Trans Am Acad Ophthalmol Otolaryngol 1974 78 195 204
Jonas JB Kreissig I Sofker A Degenring RF Intravitreal injection of triamcinolone for diffuse diabetic macular edema Arch Ophthalmol 2003 121 57 61 12523885 10.1001/archopht.121.5.729
Martidis A Duker JS Greenberg PB Rogers AH Puliafito CA Reichel E Baumal C Intravitreal triamcinolone for refractory diabetic macular edema Ophthalmology 2002 109 920 7 11986098 10.1016/S0161-6420(02)00975-2
Massin P Audren F Haouchine B Erginay A Bergmann JF Benosman R Caulin C Gaudric A Intravitreal triamcinolone acetonide for diabetic diffuse macular edema Ophthalmology 2004 111 218 225 15019365 10.1016/j.ophtha.2003.05.037
Conway MD Canakis C Livir-Rallatos C Peyman GA Intravitreal triamcinolone acetonide for refractory chronic pseudophakic cystoid macular edema J Cataract Refract Surg 2003 29 27 33 12551663 10.1016/S0886-3350(02)01441-4
Benhamou N Massin P Haouchine B Audren F Tadayoni R Gaudric A Intravitreal triamcinolone for refractory pseudophakic macular edema Am J Ophthalmol 2003 135 246 9 12566041 10.1016/S0002-9394(02)01938-4
Wilson CA Berkowitz BA Sato Y Ando N Handa JT de Juan E Jr Treatment with intravitreal steroid reduces blood-retinal barrier breakdown due to retinal photocoagulation Arch Ophthalmol 1992 110 1155 9 1497531
Early Treatment Diabetic Retinopathy Study Research Group Photocoagulation for diabetic macular oedema. ETDRS report number 1 Arch Ophthalmol 1985 103 1796 1806 2866759
Klein R Klein BE Moss SE The Wisconsin Epidemiologic Study of Diabetic retinopathy. IX. Four year incidence and progression of retinopathy when age at diagnosis is less than 30 years Arch Ophthalmol 1989 107 237 43 2916977
Wingate RJ Beaumont PE Intravitreal triamcinolone and elevated intraocular pressure Aust N Z J Ophthalmol 1999 27 431 2 10641903 10.1046/j.1440-1606.1999.00238.x
Jonas JB Kreissig I Degenring R Intraocular pressure after intravitreal injection of triamcinolone acetonide Br J Ophthalmol 2003 87 24 7 12488256 10.1136/bjo.87.1.24
Jaissle GB Szurman P Bartz-Schmidt KU Ocular side effects and complications of intravitreal triamcinolone acetonide injection Ophthalmologe 2004 101 121 8 14991307 10.1007/s00347-003-0975-z
Challa JK Gillies MC Penfold PL Exudative macular degeneration and intravitreal triamcinolone after 18 month follow up Aust N Z J Ophthalmol 1998 26 277 81 9843254 10.1046/j.1440-1606.1998.00078.x
Moshfeghi DM Kaiser PK Scott IU endophthalmitis following Intravitreal triamcinolone acetonide injection Am J Ophthalmol 2003 136 791 6 14597028 10.1016/S0002-9394(03)00483-5
Sakamoto T Enaida H Kubota T Incidence of acute endophthalmitis after triamcinolone-assisted pars plana vitrectomy Am J Ophthalmol 2004 138 137 8 15234294 10.1016/j.ajo.2004.02.072
Masket S Is there a relationship between clear corneal cataract incisions and endophthalmitis? J Cataract Refract Surg 2005 31 643 5 15899418 10.1016/j.jcrs.2005.03.019
Morgan SJ Consultant Ophthalmic Surgeon, Sunderland Eye Infirmary Personal Communication 2005
Burk SE Da Mata AP Snyder ME Visualizing vitreous using Kenalog suspension J Cataract Refract Surg 2003 29 645 651 12686230 10.1016/S0886-3350(03)00016-6
Lavin M Managing posterior capsule rupture during phacoemulsification Refractive Eye News 2004 3 20 1
Sakamato T Miyazaki M Hisatomi T Triamcinolone – assisted pars plana vitrectomy improves the surgical procedures and decreases the post operative blood-ocular barrier breakdown Graefes Arch Clin Exp Ophthalmol 2002 240 423 9 12107507 10.1007/s00417-002-0454-2
Antoszyk AN Gottlieb JL Machemer R Hatchell DL The effects of intravitreal triamcinolone acetonide on experimental pre-retinal neovascularization Graefes Arch Clin Exp Ophthalmol 1993 231 34 40 8428678 10.1007/BF01681698
Danis RP Bingaman DP Yang Y Ladd B Inhibition of preretinal and optic nerve head neovascularization in pigs by intravitreal triamcinolone acetonide Ophthalmology 1996 103 2099 104 9003344
|
15972106
|
PMC1183218
|
CC BY
|
2021-01-04 16:03:50
|
no
|
BMC Ophthalmol. 2005 Jun 22; 5:15
|
utf-8
|
BMC Ophthalmol
| 2,005 |
10.1186/1471-2415-5-15
|
oa_comm
|
==== Front
BMC Oral HealthBMC Oral Health1472-6831BioMed Central London 1472-6831-5-31597514010.1186/1472-6831-5-3Research ArticleA randomised controlled trial to explore attitudes to routine scale and polish and compare manual versus ultrasonic scaling in the general dental service in Scotland [ISRCTN99609795] Bonner Brian C [email protected] Linda [email protected] Patricia A [email protected] Wendy [email protected] Jan E [email protected] The Dental Health Services Research Unit, Mackenzie Building, Kirsty Semple Way, Dundee DD2 4BF, UK2 Postgraduate Dental Office, Eastern Region Dental Postgraduate Centre, Dundee Dental Hospital and School, Park Place, Dundee DD1 4HN, UK3 Scottish Dental Practice Based Research Network, The Lister Postgraduate Institute, Edinburgh EH8 9DR, UK2005 23 6 2005 5 3 3 31 1 2005 23 6 2005 Copyright © 2005 Bonner et al; licensee BioMed Central Ltd.2005Bonner 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
To investigate, within general dental practice, patients' and vocational dental practitioners' (VDP) attitudes towards the benefits and costs of a simple scale and polish and to compare the experience of using manual versus ultrasonic instruments to scale teeth.
Methods
28 VDPs and 420 patients participated. Patients were randomly allocated to either group. Patients' and VDPs' attitudes towards, and experience of, the scale and polish were elicited by means of self-administered questionnaires.
Results
The majority of patients (99%) believed a scale and polish was beneficial. VDPs considered ultrasonic treatment to be appropriate on significantly more occasions than they did for manual scale and polish (P < 0.001). Patient discomfort: with ultrasonic scaling 69.2% felt 'a little uncomfortable' or worse compared with 60% of those undergoing manual treatment (P = 0.072). VDPs considered treatment charges were appropriate for 77% of patients.
Conclusion
Routine scaling and polishing is considered beneficial by both patients and vocational trainees. The majority of patients, regardless of treatment method, experience some degree of discomfort when undergoing a scale and polish. VDPs showed a preference for the ultrasonic treatment method.
==== Body
Background
The Scottish Dental Practice Based Research Network's (SDPBRN) Vocational Dental Practitioner (VDP) Practice Based Research Programme is a new initiative in which VDPs are invited to take part in practice-based research studies [1]. A key aim of the programme is to encourage the development of an interest in the link between improvements in primary dental care and the findings of good quality practice-based research. The randomised controlled trial (RCT) reported here, which was carried out in the North and North-East of Scotland, was the first study in this programme. As such, it was a pilot trial: the programme having subsequently been extended Scotland-wide.
The need for practice-based research in primary dental care is widely recognised, but significant barriers do exist [2]. One such is the perceived difficulty of conducting studies, such as RCTs, without disrupting clinical work and patient care, and this can lead to a general reluctance to become involved [3]. Despite this reservation, if the evidence base in dental primary care is to be improved and if general dental practitioners (GDPs) are to be able to assess the generalisability of research results to their own practice, then high quality practice-based RCTs must be conducted within primary dental care.
During the year ending March 2002, GDPs within the General Dental Service in Scotland carried out 1,376,500 simple scale and polishes (Item 10a; Statement of Dental Remuneration) at a total cost, to health boards and patients, of £13,874,109. In contrast, only 107,042 scales and polishes requiring a minimum of two visits (Item 10b) and 2,037 intensive scales including periodontal charting (Item 10c) were provided. This typifies one aspect of the existing long-term pattern of primary care treatment, where the majority of periodontal treatment currently carried out in general dental practice consists solely of simple scaling and polishing [4].
In the 1988 UK Adult Dental Health Survey, 60% of Scottish Adults (72%; UK) were found to have visible plaque on their teeth, 62% calculus (73%; UK), and 47% periodontal pocketing of 4 mm or more (54%; UK) [5]. A lower proportion of those individuals who had visited a GDP regularly had visible plaque and calculus but this was also the case with people who reported good, self-administered dental hygiene. In recent years, the clinical need for the large numbers of simple scales and polishes carried out in the GDS has been questioned. The 2002 Audit Commission Report on the primary dental care services in England and Wales [6] suggests that in excess of half the simple scale and polish treatments prescribed may be unnecessary and may not lead to any health gain.
Prior research has concentrated on the effects of scale and polish on periodontal health [7-11]. Little research has been carried out into the attitudes of patients or GDPs towards this treatment. This trial was designed to address this gap in the knowledge base by investigating, within general dental practice vocational training GDPVT), both patients' and VDPs' attitudes towards routine scale and polish, and by comparing the experience, again from both patients' and VDPs' viewpoints, of using either manual or ultrasonic techniques.
Methods
The trial protocol was developed in collaboration with the regional dental vocational training (DVT) adviser for the East and North East of Scotland and the GDPVT adviser for each participating DVT scheme. A key consideration, when developing the protocol, was to limit disruption of the normal routine of the surgery as much as possible. Therefore, a pragmatic approach was taken, with one potential advantage of this approach being the generalisability of the results to 'real world' general dental practice.
Participants
The trial was conducted from April to June 2001. All 28 VDPs in the Aberdeen, Dundee, and Perth DVT schemes were invited to participate. Each VDP was asked to recruit 16 patients. All adult patients who were dentate, generally fit and well, attending for a routine check-up appointment, and who, in the VDP's clinical opinion, required a simple scale and polish were eligible for inclusion in the study. This treatment is defined as: 'non-surgical treatment involving scaling, polishing, and simple periodontal treatment including oral hygiene instruction, requiring only one visit'. A patient's eligibility was determined only after examination by the VDP. If the patient's proposed treatment plan did not include a simple scale and polish, the patient was not invited to participate. No attempt was made to influence this decision.
Randomisation
Consenting patients were allocated to either the manual or ultrasonic scaling group. To improve the balance of the trial arms a computer generated block randomisation sequence was used. In this, scaling allocations were generated in groups of eight with four participants allocated, in a random sequence, to each intervention group. Scaling group allocation was concealed in an opaque envelope, which was not opened until the patient agreed to take part. At the behest of either the patient or the clinician, the patient could be transferred to the other treatment method and any such changes were noted.
Materials and interventions
All patients invited to participate were encouraged to read a laminated patient information sheet, accompanied by a verbal explanation of the study. Patient characteristics, age, sex, and number of teeth, were recorded on the patient recruitment form. When patients declined to take part, the reason was noted.
Individual patient and VDP trial questionnaires, consisting chiefly of closed single or multiple response questions, were developed following a review of the literature and in collaboration with the DVT regional and scheme advisers. A pilot study was carried out with the help of a number of GDPs and their patients and several small changes were made to both questionnaires before the trial commenced.
Questionnaires were self-administered and investigated reasons for carrying out the scale and polish and attitudes towards this treatment from both the patient and VDP viewpoints.
Both parties filled out questionnaires once the treatment had been completed and then concealed them in opaque envelopes.
Power calculation
When deciding how many subjects to include in each arm of the trial, a pragmatic approach was adopted, as no information was available from previous studies to indicate the proportions of patients who would select each option for the questions they were to be asked. Advice from regional and GPVT advisers indicated that it was reasonable for each VDP to recruit 16 patients, totalling 448 patient participants. This number of subjects is sufficient to detect differences of approximately 15% between proportions with a power of 80% (α = 0.05) [12].
Ethical approval
The authors were advised, by the Chairperson of the Local Ethics Committee, that as this was a study of routine treatment that had been prescribed for the patient solely on the grounds of clinical need, with positive consent for inclusion in the study of all participants, formal ethical approval was not required.
Results
Participant flow
In total, 420 patients (out of 442 invited) agreed to participate (Male n = 178, mean age 41, range 17–76; Female n = 240, mean age 38, range 16–80; both male and female had a mean number of 24 teeth, range 5–32; missing = 2). Reasons for refusal were varied and included insufficient time and no interest in participating, with one patient concerned that ultrasonic scaling would damage his restorations. Nineteen VDPs recruited 16 patients, four recruited 15 patients, two recruited 13 patients, two recruited 12 patients, and one recruited 6 patients.
Figure 1 summarises the flow of participants through the study. Two hundred and nine patients were allocated to the manual group and 211 patients to the ultrasonic instrument group. The scaling method was changed from one group to the other on 69 occasions. The number of patients who crossed over from one arm of the trial to the other was higher than anticipated, leading to the formation of a third group, named the crossover group (n = 69). These patients were excluded where comparisons were made between the manual (n = 166) and ultrasonic (n = 185) treatment, but included where appropriate.
Figure 1 Flow of patients through the randomised controlled trial.
Benefits of a routine scale and polish
Patients were offered six options to elicit their reasons for having a scale and polish (multiple responses allowed). In descending order the response to, "Why did you have a scale and polish?" were tartar 191 patients (46%), stained teeth 150 patients (36%), bleeding gums 82 patients (20%), mouth felt unclean 65 patients (16%), and gum disease 55 patients (13%). An alternative response, "always have one with a check up" was selected by 183 patients (44%). For 91 (22%) patients, this was the only reason chosen.
Figure 2 shows, for each patient, the reasons why the VDP prescribed the scale and polish. The predominant reason given was that calculus was present 333 patients (79%). Other reasons included staining 264 patients (63%), supragingival plaque 262 patients (62%), bleeding gums 255 patients (61%), periodontal disease 154 patients (37%), appearance 125 patients (30%), and to increase diagnostic ability 97 patients (23%).
Figure 2 "Why did you provide a scale and polish for this patient?"
Questions concerning patient's perceptions of how their teeth looked and felt following the scale and polish revealed that there were no significant differences between the manual and ultrasonic groups in their responses (table 1).
Table 1 Showing patients perception of how their teeth looked and felt after they had received a scale and polish.
Manual % (n = 166) Ultrasonic % (n = 185) P Crossover (%) (n = 69)
Feel cleaner 88 89 0.001 0.971 87
Feel smother 72 65 1.62 0.203 73
Feel the same 2 3 * 0.727 3
Feel worse 0 2 * 0.250 3
Look better 49 55 0.75 0.386 61
Look the same 4 7 0.493 0.483 4
Look worse 0.006 0.005 * > 0.999 0
*: Fisher's Exact Test
Related closed multiple response questions, each with the same five possible response options, were posed to the patient and VDP regarding the patient's perception of the benefits of a scale and polish. The patient was asked "What do you think is the benefit to you of having a scale and polish?" and, to elicit the VDP's appreciation of the patient's perceptions, the VDP was asked "What do you think the patient will feel they benefited from having this scale and polish?". Four patients (1%) believed that a scale and polish was 'of no benefit' to them, 195 patients (46%) believed a scale and polish 'improves appearance', and 329 patients (78%) believed a scale and polish would keep their 'gums healthy'. Whilst anticipating these responses, VDPs did not expect that as many as 243 patients (58%) would believe that scaling and polishing was instrumental in arresting tooth decay, and over-estimated the number of patients who felt it made their mouths 'feel good' (Figure 3).
Figure 3 Perceived benefit of scale and polish. Patients: "What do you think is the benefit to you of having a scale and polish?" Dental vocational practitioners: "How do you think the patient will feel they benefited from this scale and polish?"
Costs of a routine scale and polish
The patient's responses to the question "How much did the scale and polish cost? (from a choice of £5, £8, £10, £15, or more)" indicates that the majority of patients were unaware of the price paid for individual items of treatment. While 33% of patients correctly answered that a scale and polish cost £8, 25% gave an incorrect response and 38% answered 'don't know'. The remaining four percent declined to answer (Figure 4).
Figure 4 Patient question: "How much did the scale and polish cost?"
Patients were also asked "How much would you be prepared to pay for a scale and polish?" (from a choice of £5, £8, £10, £15, or more). Those receiving a manual scale indicated a willingness to pay slightly more and this difference became more apparent on comparing the proportion of each group who were prepared to pay £10 or more: 69% (109/158) from the manual group compared with 59% (101/170) from the ultrasonic group, though the difference was not statistically significant (, 2.857, P = 0.091). Twenty three patients (manual n = 8; ultrasonic n = 15) did not respond to this question.
VDP's considered that the patient charge for treatment was appropriate for 82% (132/161) of manual treatments and for 78% (142/182) of ultrasonic treatments ( = 0.836, P = 0.361; appropriate for 77% overall when the cross-over group was included). When the charge was considered inappropriate, the suggested alternative charge was higher than £8 in 90% of cases, regardless of treatment group.
Experience of manual versus ultrasonic treatment
VDPs considered that the allocated method of scaling was the most appropriate for a significantly higher proportion of the ultrasonic group (189/210) than for the manual group (111/166; = 40.3, P < 0.001).
Crossover was significantly less from the ultrasonic to the manual group (26/202) than vice versa (43/202), ( = 4.47, P = 0.03). VDPs gave a variety of reasons for their decision to alter the scaling method. Generally, increased speed and efficiency of stain removal was used to justify change to ultrasonic scaling and tight contact points and pain were the principal reasons for crossover from ultrasonic to manual scaling. When the change was patient requested, the reason given, regardless of the direction of crossover, was that the allocated scaling method was painful.
Patients were asked "What did the scale and polish feel like?" (choosing from comfortable, a little uncomfortable, very uncomfortable, and painful). In general, patients felt greater discomfort with ultrasonic scaling with 69% (126/182) of patients feeling 'a little uncomfortable' or worse compared with 60% (99/165) of those undergoing manual scaling, but this was not statistically significant ( = 3.24, P = 0.072).
Discussion
This pilot trial was designed to compare the attitudes of VDPs and their patients towards two alternative, randomly allocated, simple scale and polish procedures provided at general dental practices in Scotland, under normal day-to-day conditions.
For pragmatic reasons, the patients included in this study were not stratified in any way and no weighting to reflect proportions of local populations examined was considered. Nevertheless, the patients recruited to the trial were seen to be comparable, in general characteristics, with patients attending for National Health Service (NHS) dental treatment in Scotland with an average of 24 teeth (24; Adult Dental Health Survey, 1998) [5], mean age 40 years (41 Adult Dental Health Survey, 1998) [5], and male to female ratio of 1:1.34 (1:1.36; General Household Survey, 1991) [13].
The VDPs gave a variety of reasons why they thought it necessary for their patients to receive a scale and polish chief amongst which were calculus, staining, supragingival plaque, and bleeding gums. In contrast to the Audit Commission Report suggesting that more than half of the simple scales and polishes carried out in the GDS are unnecessary and confer no health benefits [6], the patients surveyed in this trial had all been judged by their clinician to require a scale and polish. The patients, including the 91 patients whose only reason for having a scale and polish was because they "always have one with a check up", also generally believed that they would benefit from a scale and polish, with only 1% of patients believing a scale and polish was of no benefit. The evidence supporting the benefit of frequent scale and polish has, however, been questioned [14]. A high proportion of patients (58%) expected the scaling and polishing to stop tooth decay: an opinion not always appreciated by their dentist as only in one fifth of cases were patients predicted to give this response.
VDPs appeared to show a preference for the ultrasonic treatment and were more likely to transfer the patient from hand to ultrasonic treatment than vice versa. They suggested that the ultrasonic treatment was more efficient in terms of speed of treatment and more effective where contact points were tight. A recent review of a number of, mostly hospital-based, comparisons between these two techniques did note a moderate time saving [15]. The comfort of patients during scaling should be considered, as many nervous dental patients apparently find dental hygiene treatment contributes greatly to their anxiety towards visits for dental treatment [16]. The spread of the patients' experience between comfortable and in pain was similar for the two treatments and similar small numbers asked to be transferred to the other treatment because of pain.
In the UK, dental care is provide either privately or within the NHS at a fixed cost that includes a patient contribution of 80% (up to a maximum charge for a course of treatment). Around 49% of the population are registered with an NHS dentist at any one time [4] but, as registration lapses if a visit is not made within 15 months, a greater proportion of the population is believed to receive care within the NHS dental service than this figure might suggest. Although the number of dental treatments provided outside the NHS scheme is increasing in the UK in general, in Scotland an estimated 81% of treatments are currently provided under the NHS system [5] and it is likely that the majority of the population will continue to seek NHS dental care. All the patient participants in this study were receiving a scale and polish under the NHS system, for which a charge of £8.08 was in place at the time of the study. Although the majority of patients were unaware of the price paid for individual items of service, many patients did seem to be aware of the approximate cost of their treatment, with 49% of the subjects suggesting either £8 or £10 as the cost of a scale and polish. Slightly more subjects in the manual treatment group said they would be prepared to pay more than in the ultrasonic group possibly indicating a feeling that this treatment involved more activity by the VDP and was therefore 'worth' more.
Undoubtedly, there are difficulties in conducting studies in general dental practice, including time pressures to both patients and dentists and the need to fit in with the priority of providing good patient care. However, this pilot trial has shown that primary care-focused studies can be successfully carried out and VDPs taking part ended up with a more positive view of the concept of undertaking research in the dental surgery (findings of focus groups: not detailed here). Useful information has emerged on the attitudes and beliefs of GDS patients and newly qualified dentists towards simple scaling and polishing and the co-ordinators of the SDPBRN VDP Practice Based Research Programme have been sufficiently encouraged to commit to further studies in this series on a Scotland-wide basis.
Conclusion
The results have demonstrated that routine scaling and polishing is considered to be beneficial by both patients and VDPs and that the majority of patients, regardless of whether they received ultrasonic or manual treatment, experience some degree of discomfort. VDPs showed a preference for the ultrasonic treatment method.
This study has also demonstrated that it is possible, with careful choice of research topic and a pragmatic approach, to carry out meaningful research in a primary care setting.
Competing interests
The author(s) declare that they have no completing interests.
Authors' contributions
BCB and LY managed the day-to-day running of the study, production and distribution of questionnaires, and analysis of results; PAS was responsible for suggestions towards questionnaires and conduction of de-briefing interviews with VDPs; WM is the regional training co-coordinator and liaised with and recruited VDPs to the study; and JEC was the principle investigator with overall responsibility for 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
The encouragement and enthusiastic co-operation of Stephen Rafferty, Colin Yule, Derek Harper, and the trainers and trainees in the initiation and execution on this study is gratefully acknowledged as is the financial support of the Scottish Dental Practice Based Research Network (SDPBRN) and the Scottish Executive Health Department's Chief Scientist Office.
The views expressed in this paper are those of the authors and are not necessarily shared by other members of the SDPBRN, the Dental Health Services Research Unit, or the Scottish Executive. Full details of questionnaires can be obtained on request.
==== Refs
Vocational Dental Practitioner Trials, Scottish Dental Practice-Based Research Network, avaliable from the Mackenzie Building, Dundee, DD2 4BF 2003
Cannavina CD Cannavina G Walsh TF Effects of evidence-based treatment and consent on professional autonomy British Dental Journal 2000 188 302 306 10800236 10.1038/sj.bdj.4800463a
Bonner BC Clarkson JE McCombes W General dental practitioners views on the pursuit and practice of evidence-based dentistry: The results of a questionnaire 2001
Scottish Dental Practice Board Annual Report 2001/02 2002 Edinburgh, UK
Adult Dental Health Survey 1998 Office for National Statistics, London, UK
Audit Commission National Health Report Dentistry: Primary dental care services in England and Wales 2002 Audit Commission for local authorities and the National Health Service in England and Wales, London
Elley K Gold L Burls A Gray M Scale and polish for chronic periodontal disease A West Midlands Development and Evaluation Service Report 2001 University of Birmingham, Birmingham
Stamm JW Epidemiology of gingivitis J Clin Periodontol 1986 13 360 366 3522645
Ismail AI Lewis DW Dingle JL Prevention of periodontal disease Canadian Task Force on the Periodic Health Examination 1994 Canadian Guide to Clinical Preventive Health Care. Ottawa: Health Canada 420 431
Glavind L Zeuner E Attström R Oral cleanliness and gingival health following oral hygiene instruction by self-educational programs J of Clin Periodont 1984 11 262 273
Beltrami M Bickel M Baehni PC The effect of supragingival plaque control on the composition of the subgingival microflora in human periodontitis J of Clin Periodont 1987 14 161 164
Armitage P Berry G Matthews JN Sample-size determination Statistical methods in medical research 2002 4 Blackwell Science Ltd., Oxford OX2 OEL
General Household Survey 1991 Office of Population Censuses and Surveys, HMSO, London
Bierne P Forgie A Worthington HV Clarkson JE Routine scale and polish for periodontal health in adults Cochrane Database Syst Rev 2005 15674957
Tunkel J Flemming TF A systematic review of efficacy of machine-driven and manual subgingival debridement in the treatment of chronic periodontitis J Clin Periodont 2002 29 72 81 10.1034/j.1600-051X.29.s3.4.x
de Jongh A Stouthard ME Anxiety about dental hygienist treatment Community Dent Oral Epidemiol 1993 21 91 95 8485976
|
15975140
|
PMC1183219
|
CC BY
|
2021-01-04 16:29:57
|
no
|
BMC Oral Health. 2005 Jun 23; 5:3
|
utf-8
|
BMC Oral Health
| 2,005 |
10.1186/1472-6831-5-3
|
oa_comm
|
==== Front
BMC Oral HealthBMC Oral Health1472-6831BioMed Central London 1472-6831-5-51600198410.1186/1472-6831-5-5Research ArticleSodium channel Nav1.8 immunoreactivity in painful human dental pulp Renton T [email protected] Y [email protected] C [email protected] S [email protected] C [email protected] P [email protected] Department of Oral & Maxillofacial Surgery, Dental Institute, Queen Mary's College, London University, Whitechapel, London UK2 Peripheral Neuropathy Unit, Imperial College, Hammersmith Hospital Campus, Du Cane Road, London W12 ONN, UK3 GlaxoSmithKline, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK4 Neurology-CEDD, GlaxoSmithKline, Third Avenue, Harlow CM19 5AW, UK2005 7 7 2005 5 5 5 24 12 2004 7 7 2005 Copyright © 2005 Renton et al; licensee BioMed Central Ltd.2005Renton 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 tetrodotoxin-resistant voltage-gated sodium channel Nav1.8 (SNS1/PN3) is expressed by nociceptors and may play a role in pain states.
Methods
Using specific antibodies for immunohistochemistry, we studied Nav1.8 – immunoreactivity in human dental pulp in relation to the neuronal marker neurofilament. Human tooth pulp was extracted from teeth harvested from a total of twenty-two patients (fourteen without dental pain, eight patients with dental pain).
Results
Fibres immunoreactive for Nav1.8, were significantly increased on image analysis in the painful group: median (range) Nav1.8 to Neurofilament % area ratio, non-painful 0.059 (0.006–0.24), painful 0.265 (0.13–0.5), P = 0.0019.
Conclusion
Nav1.8 sodium channels may thus represent a therapeutic target in trigeminal nerve pain states.
==== Body
Background
Pain is the most common symptom of diseased tooth pulp, often a result of coronal caries of the tooth, affecting up to 80% of the western population during their lives. The mature human dental pulp is densely innervated with over 900 axons entering the average human premolar tooth [1] that originate from the trigeminal ganglion. The normal pulp seems insensitive to exteroceptive stimuli; however, in pathological states, electrical, thermal, mechanical and chemical stimuli all produce a nociceptive response [2]. Primary and permanent tooth pulps contain 70–90% C-fibres [3], myelinated fibres mostly of the A delta category [3], with few myelinated fibres of the A beta group. The majority of nerve fibres terminate in the coronal region of the pulp, forming a subodontoblast plexus, with 40% terminating in the dentinal tubules close to the odontoblast processes [3]. Strong correlations have been reported between the afferent discharge frequency of human pulp nociceptors and pain levels [4]. Many suggestions have been made for the origin of pulpal pain e.g. pulp inflammation involving several mediators located within the pulp (cholinergic and adrenergic neurotransmitters, prostaglandins and cyclic AMP). However, thus far, no correlation has been established between pain characteristics and histology of the pulp [5,6].
Voltage-gated sodium channels play key roles in the pathophysiology of pain and are distinguished according to their sensitivity to the neurotoxin tetrodotoxin (TTX) as fast-activating TTX-sensitive (TTX-S) channels, or slow-inactivating TTX-resistant channels (TTX-R). The distribution and pathophysiology of these channels, particularly Nav1.8, (SNS/PN3) have been the focus of research in pain mechanisms [7]. Recently, antisense treatment blocking this channel reduced neuropathic pain [8]. We have previously described the temporal and spatial distribution of Nav1.8 in human sensory neurones [9]; the channels were decreased acutely in sensory cell bodies after spinal cord root avulsion but accumulated in fibres proximal to the site of injury in brachial plexus trunks, and in neuromas. Based on the presence of Nav1.8 in predominantly small medium sized neurons in human DRG, it is likely that Nav1.8 is present in both A delta and C fibres [9,10].
This study aimed to assess if pulpal pain associated with caries was associated with any change of Nav1.8-immunoreactivity within tooth pulp nerve fibres.
Methods
Patients scheduled for dental extraction at Guy's Dental Institute, London, were included in the study, subsequent to providing consent in accordance with the local research ethics committee.
22 permanent molar teeth about to be extracted were tested, 1 hour prior to extraction, for vitality using an electric pulp tester (analytic technology constant current at the mid-buccal surface of the tooth) and with ethyl chloride to confirm the neural vitality of the dental pulp. A pain history was also collected (existing pain and duration). The patients were divided into two groups, those with existing pain from the tooth (n = 8 patients age range: 40.3 ± 4.0 years) and those with no history of or existing pain (n = 14 patient age range: 37.3 ± 14.6 years). The gender distribution of the groups was M:F 1:1. All the dental pain in this study was attributable to pulpitis due to extensive dental caries of the molar tooth, the duration of pain was 2.9 weeks (range 0.5–8), and the indication for extraction of the non-painful teeth was pericoronitis. All the teeth were removed by standard buccal approach under local or general anaesthesia. Subsequent to the extraction process (lasting less than 5 min), the teeth were sectioned vertically with a water-cooled drill and the pulp lifted out, and specimens immediately snap-frozen at -70°C. Intentional examination of the coronal pulp, including the densely innervated subontoblastic layer, was assisted by careful orientation of the pulp on a marked sterile card. Based on the number of inflammatory cells present, the inflammation was graded in accordance with local histological standard scales, which were mild, moderate or severe.
Immunohistochemistry
Frozen pulp or nerve were embedded in OCT medium (RA Lamb, London, UK) and sections of 12 μm thaw-mounted onto glass slides pre-coated with poly-L-lysine. Sections were immersion-fixed in fresh 4% paraformaldehyde in phosphate buffered saline (PBS) for 30 min, then endogenous peroxidases blocked by incubation with alcoholic 0.3% hydrogen peroxide for a further 30 min. Sections were incubated overnight with a monoclonal antibody to the neuronal marker neurofilament (Clone 2F11, Dako, Cambridge, U.K., used at a final titre of 1:10, 000) and a polyclonal antibody against the Nav1.8 (K107), whose specificity has been described by us previously [9]. Sites of primary antibody attachment were revealed using avidin-biotin peroxidase method (Vector Elite ABC method, Vectastain, Novacastra, Newcastle, UK). Preparations were counterstained in 1% w/v aqueous neutral red to visualise nuclei and photographed with an Olympus photomicroscope. Specificity studies of the Nav1.8 antibody (K107), showing positive staining in human DRG neurons, and no staining in pre-absorption experiments using Nav1.8 peptide in tooth pulp sections, were performed as previously described [9].
Image analysis
Nav1.8 and neurofilament immunoreactivity in fibres were quantified using computerized image analysis (Seescan Cambridge, UK). Quantification of the data was performed using 12-micron thick serial sections. Images were captured via video link to an Olympus BX50 microscope (×40, objective) and scanned by computer. Setting grey-level detection limits at threshold, highlighted positive immunostaining, and the area of highlighted fibres was obtained as % area of the field scanned. Scanning was performed for a minimum of 5 fields at random per tissue section orientated longitudinally, assessed in a blind fashion; 3 tissue sections from each pulp were analysed, and the mean value from each patient obtained. Results are expressed as the average percentage ratio of the mean Nav1.8 to neurofilament reactive fibres in 5 fields.
Analysis
The Mann Whitney test was used to compare ratios between groups; P values less than 0.05 were considered significant.
Results
Figure 1 shows Neurofilament staining of the pulpal horn of a molar tooth highlighting the relationship of the subodontoblastic plexus (boxed area) to the deeper relatively poorer innervated coronal pulp at low magnification. Immunostaining demonstrated the presence of large numbers of nerve fibres within human tooth pulp that were immunoreactive for neurofilament (Fig. 2A and B). A subset of nerve fibres was also immunostained with the Nav1.8 antibody (Fig. 2C and D) in both non-painful and painful pulp groups. Specificity of the Nav1.8 antibody was demonstrated by methods described for this antibody by us in human DRG [9], including omission of primary antibody and pre-incubation with peptide.
Figure 1 Photomicrograph illustrates the pulpal horn of a molar tooth, highlighting the relationship of the subodontoblastic plexus (boxed area) to the deeper coronal pulp. Magnification × 10
Figure 2 Immunoreactive nerve fibres in non-painful (left column) and painful (right column) human tooth pulp sections, in pulp from painful teeth within the subontoblastic plexus region 'boxed' in Figure 1. Staining with antibodies to Nav1.8 (A and C) and neurofilament (B and D). Arrows indicate Nav1.8 immunoreactive nerve fibres. Magnification × 40
By image analysis, neurofilament fibre median % area (range) in non-painful tooth pulp was 13.94 (3.05–22.05), and in painful tooth pulp was 18.21 (9.11–27.88); there was trend for an increase, but this was not statistically significant. There was a significant change of the corresponding Nav1.8 % area in non-painful 0.68 (0.13–2.90) compared to painful tooth pulp 5.855 (1.3–7.52), P < 0.005. There were also significantly more fibres immunostaining for Nav1.8 in relationship to neurofilament positive fibres in the painful pulp, compared with those without pain (Fig 3). The median Nav1.8 to Neurofilament % area ratios were for non-painful 0.059 (0.006–0.24), and for painful 0.265 (0.13–0.5), P < 0.005. A degree of inflammation was seen in all painful pulp samples.
Figure 3 Scattergram of Nav1.8 to Neurofilament ratio of the % area in control and painful tooth pulp. The median value is indicated * P < 0.005.
Discussion
Neurofilament positive fibres have previously been reported in human dental pulp, including fine unmyelinated fibres in the sub odontoblastic layer [11,12]. However, thus far there have been few investigations of the expression of ion channels in the human dental pulp. We have demonstrated for the first time that numerous Nav1.8-immunoreactive nerve fibres are present in human dental pulp in the subodontoblastic layer, and the Nav1.8-immunoreactive fibres were increased in the presence of caries-induced painful pulpitis. While there was a trend for increased neurofilament fibres in painful dental pulp, this was not statistically significant. An increase of nerve fibres within inflamed dental pulp, possibly due to nerve sprouting, has been previously reported in the rat dental pulp after dentine injury [13], and other studies report an increased neuropeptide expression and sprouting in the human infected dental pulp [14,15]. In contrast to our findings for Nav1.8 in this study, our previous study of TRPV1 and P2X3 receptors showed no significant change in painful dental human dental pulp [16].
The presence of Nav1.8-immunoreactive neurons identified in the subodontoblastic layer implies that these receptors may be involved in signal transduction at the pulp-dentine junction. It is known that sensory neurones and odontoblasts exist in close proximity, but no synaptic or electrical connections have been identified [2,12] postulated that the odontoblast-neuron connection may be neurochemical. Infection or injury to the pulpal tissues may result in inflammation, resulting in increased expression of substance P, CGRP and collateral nerve sprouting, which are regulated by nerve growth factor (NGF), which also regulates Nav1.8 expression by sensory neurons [7]; NGF is itself increased in inflamed pulpal tissues [17]. The sampling of healthy non-painful pulp from partially erupted third molars, though developmentally mature, may not be representative of fully erupted molar pulps. Fibre numbers and receptor expression may change after eruption of the tooth [2]; it remains unknown whether Nav1.8-immunoreactivity varies with eruption or maturity of teeth.
Although our cross reactivity and pre-absorption studies showed specificity of the antibodies used, the data should be interpreted with caution. Immunostaining of nerves appears to be axoplasmic in this study, and not at the nodes of Ranvier. Recently [18] Henry and colleagues have reported Nav1.8-immunoreactivity associated with the nodes of Ranvier in the radicular human tooth pulp. The difference in localisation between the studies could reflect characteristics of the antibodies and immunostaining methods, and further studies, including different techniques and functional assays, are necessary. The relative expression of Nav1.8-immunoreactive fibres to neurofilament positive fibres is low – this may also be due to the affinity of the antibody used in this study, or reflect the expression of this ion channel in a small sub-set of nerve fibres. The specific type of fibre expressing Nav1.8, the distribution of Nav1.8 throughout the human dental pulp, and the longitudinal changes in the Nav1.8-immunoreactivity caused by pulpal inflammation, all require further study.
Conclusion
In conclusion, nerve fibres in dental pulp from patients with dental pain showed significantly more Nav1.8-fibres as a proportion of neurofilament positive fibres. As Nav1.8 has been implicated in neuropathic pain, its expression by nerve fibres within human tooth pulp may contribute to the pathophysiology of dental pain. Further studies of the time-course of the disease, and severity of pain and/or inflammation, are necessary to elucidate the role and regulation of Nav1.8 ion channels in the pathophysiology of trigeminal pain. Nav1.8 represents a target for novel analgesic agents.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TR performed all the surgical procedures, extracted the tooth pulp and helped write the paper. YY participated in the immunohistochemistry, analysis of data and drafted the manuscript. CP, ST and CB were responsible for the design and production of the NaV1.8 antibodies used, help with interpretation of the data, and writing the manuscript. CB participated in the conception of the study, development of antibodies, and interpreting the data. PA conceived the study and participated in its design and coordination, interpretation and completion of the manuscript. All authors read and approved the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Eddie Odell and his team, Department of Oral Pathology, GKT Dental Institute, King's College London. We also thank Kevin Coward for help with pilot studies.
==== Refs
Reader A Foreman DW An ultrastructural quantitative investigation of human intradental innervation J Endod 1981 7 493 499 6946179
Hildebrand C Fried K Tuisku F Johansson CS Teeth and tooth nerves Prog Neurobiol 1995 45 165 222 7777672 10.1016/0301-0082(94)00045-J
Fried K Hildebrand C Pulpal axons in developing, mature, and aging feline permanent incisors. A study by electron microscopy J Comp Neurol 1981 203 23 36 7309916 10.1002/cne.902030104
Narhi M The neurophysiology of the teeth Dent Clin North Am 1990 34 439 448 2197120
Tyldesley WR Mumford JM Dental pain and the histological condition of the pulp Dent Pract Dent Rec 1970 20 333 336 5272575
Seltzer S Hypothetic mechanisms for dentine sensitivity Oral Surg Oral Med Oral Pathol 1971 31 388 399 5277391
Waxman SG The molecular pathophysiology of pain: abnormal expression of sodium channel genes and its contributions to hyperexcitability of primary sensory neurons Pain 1999 Suppl 6 S133 40 10491982 10.1016/S0304-3959(99)00147-5
Lai J Gold MS Kim CS Bian D Ossipov MH Hunter JC Porreca F Inhibition of neuropathic pain by decreased expression of the tetrodotoxin-resistant sodium channel, NaV1.8 Pain 2002 95 143 152 11790477 10.1016/S0304-3959(01)00391-8
Coward K Plumpton C Facer P Birch R Carlstedt T Tate S Bountra C Anand P Immunolocalization of SNS/PN3 and NaN/SNS2 sodium channels in human pain states Pain 2000 85 41 50 10692601 10.1016/S0304-3959(99)00251-1
Akopian AN Sivilotti L Wood JN A tetrodotoxin-resistant voltage-gated sodium channel expressed by sensory neurons Nature 1996 379 257 262 8538791 10.1038/379257a0
Maeda T Iwanaga T Fujita T Kobayashi S Immunohistochemical demonstration of nerves in the predentin and dentin of human third molars with the use of an antiserum against neurofilament protein (NFP) Cell Tissue Res 1986 243 469 475 2420457 10.1007/BF00218053
Alavi AM Dubyak GR Burnstock G Immunohistochemical evidence for ATP receptors in human dental pulp J Dent Res 2001 80 476 483 11332536
Taylor PE Byers MR Redd PE Sprouting of CGRP nerve fibers in response to dentin injury in rat molars Brain Res 1988 461 371 376 3263169 10.1016/0006-8993(88)90270-3
Rodd HD Boissonade FM Substance P expression in human tooth pulp in relation to caries and pain experience Eur J Oral Sci 2000 108 467 474 11153921 10.1034/j.1600-0722.2000.00924.x
Awawdeh L Lundy FT Shaw C Lamey PJ Linden GJ Kennedy JG Quantitative analysis of substance P, neurokinin A and calcitonin gene-related peptide in pulp tissue from painful and healthy human teeth Int Endod J 2002 35 30 36 11853236 10.1046/j.1365-2591.2002.00451.x
Renton T Yiangou Y Baecker PA Ford AP Anand P Capsaicin receptor VR1 and ATP purinoceptor P2X3 in painful and nonpainful human tooth pulp J Orofac Pain 2003 17 245 250 14520770
Doubleday B Robinson PP Nerve growth factor depletion reduces collateral sprouting of cutaneous mechanoreceptive and tooth-pulp axons in ferrets J Physiol 1994 481 ( Pt 3) 709 718 7707237
Henry MA Sorensen HJ Johnson LR Levinson SR Localization of the Na(v)1.8 sodium channel isoform at nodes of Ranvier in normal human radicular tooth pulp Neurosci Lett 2005 380 32 36 15854746 10.1016/j.neulet.2005.01.017
|
16001984
|
PMC1183220
|
CC BY
|
2021-01-04 16:29:57
|
no
|
BMC Oral Health. 2005 Jul 7; 5:5
|
utf-8
|
BMC Oral Health
| 2,005 |
10.1186/1472-6831-5-5
|
oa_comm
|
==== Front
BMC Oral HealthBMC Oral Health1472-6831BioMed Central London 1472-6831-5-61604278210.1186/1472-6831-5-6Research ArticleChildren's acceptance of milk with xylitol or sorbitol for dental caries prevention Castillo Jorge L [email protected] Peter [email protected] Susan E [email protected] Ramon [email protected] Rocio [email protected] Departamento Academico de Estomatología del Niño y del Adolescente, Universidad Peruana Cayetano Heredia, Honorio Delgado 430, Lima 34, Peru2 Dental Public Health Sciences, Box 357475, University of Washington, Seattle, USA 98195-74753 Dental Public Health Sciences, Box 357475, University of Washington, Seattle, USA 98195-74754 Departamento Academico de Estomatología del Niño y del Adolescente, Universidad Peruana Cayetano Heredia, Honorio Delgado 430, Lima 34, Peru5 Departamento Academico de Estomatología del Niño y del Adolescente, Universidad Peruana Cayetano Heredia, Honorio Delgado 430, Lima 34, Peru2005 22 7 2005 5 6 6 16 4 2005 22 7 2005 Copyright © 2005 Castillo et al; licensee BioMed Central Ltd.2005Castillo 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
Xylitol, a polyol sugar, has been shown to reduce dental caries when mixed with food or chewing gum. This study examines the taste acceptability of xylitol in milk as a first step toward measuring the effectiveness of xylitol in milk for the reduction of dental caries in a public health program.
Methods
Three different types of milk (Ultra High Temperature (UHT), powder and evaporated) were tested for acceptability by 75 Peruvian children (25 per milk group, ages 4 to 7 years). Each group evaluated xylitol and sorbitol in one type of milk. In the first phase, each child was presented with a tray of four plastic cups containing 50 ml of milk with 0.021 g/ml xylitol, 0.042 g/ml xylitol, 0.042 g/ml sorbitol or no sugar. Each child was asked to taste the samples in a self-selected order. After tasting each sample, the child placed the milk cup in front of one of three cartoon faces (smile, frown or neutral) representing the child's response to the taste of each sample. In the second phase, the child was asked to rank order the milk samples within each category (smile, frown or neutral). Ranks within categories were then combined to obtain a rank ordering for all the test samples.
Results
The ranking from best to worst for the samples across categories (UHT, powder, evaporated) was xylitol (0.0.042 g/ml), sorbitol (0.042 g/ml), xylitol (0.021 g/ml) and milk alone (Friedman's ANOVA). Xylitol and sorbitol were preferred over milk alone, and xylitol (0.042 g/ml) was preferred to sorbitol (0.042 g/ml)(p < .05 sign test).
Conclusion
Milk sweetened with xylitol is well accepted by Peruvian children ages 4–7 years.
==== Body
Background
Xylitol is a non-nutritive sweetener that has demonstrated effectiveness for preventing dental caries [1]. It has been introduced in different foods for children including gum, candies, gelatin, sorbets, syrups and other products including multivitamins, lozenges, toothpaste, and oral rinses. Studies have demonstrated that the daily ingestion of 5 to 10 g of xylitol in different vehicles can reduce the levels of dental caries up to 82% [2-6]. This reduction can be explained by the effect of xylitol on cariogenic bacteria [7].
Xylitol has the same sweetness and bulk of sucrose but with one third fewer calories, approximately 2.4 cal/g [8]. Snacks made with xylitol are generally well accepted in preschool children and its application may be suitable for public health programs [9].
The perception of flavors in milk is one of the human infant's earliest sensory experiences, and there is support for the idea that this early experience with flavors has an effect on milk intake and on later food acceptance [10]. There have been attempts to use milk as a vehicle for caries prevention. While milk per se may have some protective action against dental caries [11], evidence suggests its impact is negligible to low if consumed in normal amounts [12]. Milk supplemented with fluoride has reduced dental caries rates in studies in Chile [13] and Hungary [14]. However, adding fluoride to milk in underdeveloped countries is potentially a problem because of the risk of overdose and fluorosis.
In Peru, chronic malnutrition is around 22 percent in children younger than five years old, but in children living in extreme poverty, it increases as high as 40 percent [15]. The Food and Agriculture Organization of the United Nations minimum recommended daily caloric intake for children is 2492 cal, but in Peru the part of the population living in extreme poverty consumes, on average, around 1658 cal. [16] In order to overcome this problem, there is a public health program called El Vaso De Leche (The Glass of Milk). This program offers a daily ration of food and milk to vulnerable populations, in order to improve their nutrition. The community actively participates in this program, always supported by the municipalities or the central government. Each child receives a daily glass of milk as part of the breakfast for the five weekdays. Among those who receive some type of food aid from the government, seventy eight percent belong to the El Vaso De Leche program [15].
Milk supplemented with xylitol has not been studied as a public health measure. If it can be demonstrated that xylitol added to milk has a beneficial anti-caries effect, it can be an important preventive measure in the young population, especially for children from low-income families who receive a daily glass of milk as part of the social programs in Peru. The purpose of this study is to examine the taste acceptability of xylitol in milk as a first step prior to measuring the effectiveness of xylitol in milk for the reduction of dental caries in a public health program.
Methods
Subjects
The parents of 108 out of 450 children (ages 4 to 7 years) enrolled in a primary school in Lima, Peru were approached in person and asked for permission to have their child participate in the study. Of the 108 parents contacted 101 (95%) agreed to allow their child to participate and 75 were included. Written, informed consent was obtained from these parents. Oral assent was also obtained from all the participating children prior to the initiation of study procedures. Study procedures were reviewed and approved by the human subjects committee at the Universidad Peruana Cayetano Heredia.
The study was carried out in three replications, one for each of the different types of milk (Ultra High Temperature (UHT), evaporated, or powdered) available in Peru. Twenty-five children participated in each replication, and no child participated in the study more than once. Ages of the children were matched across replications.
Testing procedure
Taste testing was based on the two-part procedure of Birch and Sullivan [17]. This ranking procedure has been demonstrated to be reliable and valid in children three years of age and older [18,19].
Children were tested individually in the dental room of the school. Each child was presented with a tray containing four 120 ml transparent plastic cups filled with 50 ml of milk at room temperature. Location of each of the four choices on the tray was randomly determined. Cups were labeled with three-digit codes on the bottom of the cup in order to allow the experimenter to identify each item. The milk in each cup contained one of four formulations of milk, plain milk, milk with 0.021 g/ml xylitol (low xylitol), milk with 0.042 g/ml xylitol (high xylitol) or milk with 0.042 g/ml sorbitol. The doses of xylitol were chosen based on the amounts used in previous studies and assuming a daily consumption volume of 240 ml of milk with either 5 mgs or 10 mgs of xylitol or sorbitol in it (5 mg/240 ml = 0.021 mg/ml, 10 mg/240 ml = 0.042 mg/ml) [2,4,7].
Hedonic ratings for each milk formulation were obtained in the first phase of testing. Each child was asked to taste each of the four formulations of milk in a self-selected order. After tasting each milk sample, the child was asked to place the cup in front of one of three cartoon faces. The faces were a smile (like), a frown (dislike), or neutral expression. These faces were introduced to the children ahead of time in a practice session as indicating foods they like a lot (for example, a fruit juice), feel neutral about (water) or dislike (sugarless and cold tea). The hedonic category in which the child placed each milk sample was recorded.
Rankings of each of the four milk formulations were obtained in the second phase of testing. All the milks placed in the like category were presented to the child, and he/she was asked to taste them again and pick which one he/she liked the best. This milk was then removed, and the child was asked to taste the remaining items and asked again which one he/she liked the best. This was repeated with all remaining items in the like category. The procedure was then repeated with the other two hedonic categories. In this way a ranking of the four milks was achieved in a manner that was easy for young children. Rankings of the milk (first, second, third and fourth) were recorded by the investigators.
Analysis
Friedman's ANOVA by ranks followed by post hoc sign tests were used to assess whether children's ranking of milks differed between formulations. Analyses were conducted separately for each study replication (UHT, evaporated, and powdered milk).
Results
In all three milk types (UHT, powdered and evaporated), the 0.042 g/ml xylitol formulations were preferred to the other milks tested (χ2's > 30, p < 0.00001, Z's > 2.0, p's < 0.05, Tables 2,3,4). Furthermore in all three milk types, xylitol 0.042 g/ml and sorbitol 0.042 g/ml were preferred to plain milk, and xylitol 0.042 g/ml was preferred to sorbitol 0.042 g/ml (Z's > 2.0, p's < 0.05, Tables 2,3,4).
Table 2 Children's ranking of UHT milks collapsed across preference categories
FOOD ITEM 1ST 2ND 3RD 4TH MEDIAN RANK OF FOODS
XYLITOL 0.042 g/mla 19 3 2 1 1
XYLITOL 0.021 g/mlb 3 8 10 4 3
SORBITOL 0.042 g/mlc 2 12 5 6 2
MILK ALONEd 1 2 8 14 4
Friedman ANOVA, N = 25, df = 3, Chi Square = 30.6, p < .00001. Food items labeled with different letters differ significantly from each other by sign test (p < 0.05).
Table 3 Children's ranking of powdered milks collapsed across preference categories
FOOD ITEM 1ST 2ND 3RD 4TH MEDIAN RANK OF FOODS
XYLITOL 0.042 g/mla 18 3 1 3 1
XYLITOL 0.021 g/mlb 1 8 13 3 3
SORBITOL 0.042 g/mlc 6 12 6 1 2
MILK ALONEd 0 2 5 18 4
Friedman ANOVA, N = 25, df = 3, Chi Square = 36.1, p < .000001. Food items labeled with different letters differ significantly from each other by sign test (p < 0.05).
Table 4 Children's ranking of evaporated milks across preference categories
FOOD ITEM 1ST 2ND 3RD 4TH MEDIAN RANK OF FOODS
XYLITOL 0.042 g/mla 21 2 0 2 1
XYLITOL 0.021 g/mlb,c 1 8 8 8 3
SORBITOL 0.042 g/mlb 2 12 9 2 2
MILK ALONEc 1 3 8 13 4
Friedman ANOVA, N = 25, df = 3, Chi Square = 33.7, p < .000001. Food items labeled with different letters differ significantly from each other by sign test (p < 0.05).
Discussion
The purpose of this study was to determine if children would accept milk sweetened with xylitol. We used three different types of milk (UHT, powder and evaporated) because there is variation in the milk given to children in the social programs in Peru. All children completed the experiment, including four (2 UHT, 1 evaporated and 1 powder) who stated before the experiment that they did not like milk. Because excluding these children could have reduced the practical value of our the results of our study, they were included [20]. Moreover, a dislike of milk altogether is not likely to have had much influence on the results because the primary dependent measure of interest was the relative ranking of each milk type.
Most previous studies that have investigated the addition of xylitol to different types of food have demonstrated that five to 10 mg of xylitol affects S. mutans and reduces dental caries levels [4,6,21-23]. The doses we used (0.042 g/ml and 0.021 g/ml), resulted in either five or 10 mg of xylitol per glass.
Children preferred the milk with xylitol 0.042 mg/ml over milk with sorbitol 0.042 mg/ml. We informally asked some of the children what they found different between the milks. Some of the children stated that sorbitol tasted more artificial, and that xylitol was sweeter. Xylitol is the only sugar alcohol that has the same sweetness as sucrose [24]. Sorbitol is 40 percent less sweet that xylitol and sucrose [24]. Maltitol is the next closest in sweetness to sucrose (20 percent less sweet) [24].
This project is the first step towards a more comprehensive study that will look at the effects of the ingestion of milk with xylitol on cariogenic bacteria and on dental caries levels. One issue for a larger study is whether the properties of xylitol make it suitable to place it in milk. Xylitol forms very loose complexes with calcium under some conditions in vitro and delays precipitation of calcium; however, these features should have no impact on the bioavailability of the xylitol in milk [25]. Also, previous studies using other xylitol vehicles have had daily frequencies up to five. Studies are needed to ascertain whether single dose per day is substantive enough to modify the bacteria flora.
Overall, milk with xylitol is well accepted by children four to seven years old. If we can demonstrate that there is a relevant effect of xylitol in milk on cariogenic bacteria and levels of dental caries, as xylitol does in chewing gum, puddings, and other meals, we will have another public health tool to be applied in populations with high levels of the disease, and who have the availability of at least one glass of milk at school. In Peru, with programs like El Vaso De Leche that are directed to improve the nutrition of populations living under conditions of poverty, children may have an additional benefit if we can demonstrate that xylitol in milk has an anti-caries effect.
Conclusion
This preliminary study found that xylitol in milk is acceptable to Peruvian children and that both xylitol and sorbitol in milk at 0.042 g/ml are preferred to plain milk. Xylitol is preferred over sorbitol.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Dr. Castillo organized and directed the study. He served as the liaison to the schools and also wrote the initial draft of the manuscript. Drs. Milgrom and Coldwell contributed to the design, statistical analysis, and writing of the subsequent drafts. Drs. Castillo and Lazo carried out the field aspects of the study. All authors read and approved the final manuscript.
Table 1 Design of the milk groups presented to groups of children
GROUP 1 UHT N = 25 GROUP 2 POWDER N = 25 GROUP 3 EVAPORATED N = 25
0.021 g/ml XYLITOL 0.021 g/ml XYLITOL 0.021 g/ml XYLITOL
0.042 g/ml XYLITOL 0.042 g/ml XYLITOL 0.042 g/ml XYLITOL
MILK ALONE MILK ALONE MILK ALONE
0.042 g/ml SORBITOL 0.042 g/ml SORBITOL 0.042 g/ml SORBITOL
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors acknowledge the assistance of Ross Craig and Danisco Sweeteners.
==== Refs
Hayes C The effect of non-cariogenic sweeteners on the prevention of dental caries: a review of the evidence J Dent Educ 2001 65 1106 1109 11699985
Isokangas P Alanen P Tiekso J Makinen KK Xylitol chewing gum in caries prevention. A field study in children at caries-active ages J Am Dent Assoc 1988 117 315 320 3166474
Isokangas P Alanen P Tiekso J Makinen KK Long-term effect of xylitol chewing gum on dental caries Community Dent Oral Epidemiol 1989 17 200 2758793
Makinen KK Bennett CA Hujoel PP Isokangas PJ Isotupa KP Pape HR JrMakinen PL Xylitol chewing gums and caries rates: A 40-month study J Dent Res 1995 74 1904 1913 8600188
Hujoel PP Makinen KK Bennett CA Isotupa KP Isokangas PJ Allen P Makinen PL The optimum time to initiate habitual xylitol gum-chewing for obtaining long-term caries prevention J Dent Res 1999 78 797 803 10096456
Alanen P Isokangas P Gutmann K Xylitol candies in caries prevention: results of a field study in Estonian children Community Dent Oral Epidemiol 2000 28 218 224 10830649 10.1034/j.1600-0528.2000.280308.x
Roberts MC Riedy CA Coldwell SE Nagahama S Judge K Lam M Kaakko T Castillo JL Milgrom P How xylitol-containing products affect cariogenic bacteria J Am Dent Assoc 2002 133 435 441 11991460
Lindley MG Birch GG Khan R Sweetness of sucrose and xylitol. Structural considerations J Sci Fd Agric 1976 27 140 144
Lam M Riedy CA Milgrom P Coldwell SE Craig R Children's acceptance of xylitol-based foods Community Dent Oral Epidemiol 2000 28 97 101 10730717 10.1034/j.1600-0528.2000.028002097.x
Birch LL Fisher JO Development of eating behaviors among children and adolescents Pediatrics 1998 101 539 549 12224660
Bowen WH Pearson SK Effect of milk on cariogenesis Caries Res 1993 27 461 466 8281559
Grenby TH Andrews AT Mistry M Williams RJ Dental caries-protective agents in milk and milk products: investigations in vitro J Dent 2001 29 83 92 11239581 10.1016/S0300-5712(00)00061-0
Marino R Villa A Guerrero S A community trial of fluoridated powdered milk in Chile Community Dent Oral Epidemiol 2001 29 435 442 11784286 10.1034/j.1600-0528.2001.290604.x
Scheinin A Banoczy J Szoke J Esztari I Pienihakkinen K Scheinin U Tiekso J Zimmermann P Hadas E Collaborative WHO xylitol field studies in Hungary. I. Three-year caries activity in institutionalized children Acta Odontol Scand 1985 43 327 347 3879082
Gajate G Inurritegui M El impacto del Vaso de Leche sobre el nivel de nutrición infantil Economia y Sociedad 2003 50 63 70
Suarez MA Caracterización del programa del Vaso de Leche
Birch LL Sullivan SA Measuring children's food preferences J School Health 1991 61 212 214 1943045
Birch LL Dimensions of preschool children's food preferences J Nutr Educ 1979 11 77 80
Birch LL Preschool children's food preferences and consumption patterns J Nutr Educ 1979 11 189 192
Chapman KW Boor KJ Acceptance of 2% ultra-pasteurized milk by consumers, 6 to 11 years old J Dairy Sci 2001 84 951 995 11352172
Soderling E Makinen KK Chen CY Pape HR Loesche W Makinen PL Effect of sorbitol, xylitol and xylitol/sorbitol chewing gums on dental plaque Caries Res 1989 23 378 384 2766327
Makinen KK Soderling E Isokangas P Tenuovo J Tiekso J Oral biochemical status and depression of Streptococcus mutans in children during 24- to 36- month use of xylitol chewing gum Caries Res 1989 23 261 267 2790861
Isotupa KP Gunn S Chen C-Y Lopatin D Makinen KK Effect of polyol gums on dental plaque in orthodontic patients Am J Orthod Dentofacial Orthop 1995 107 497 504 7733059
Bar A Nabors LO, Gelardi RC Xylitol Alternative Sweeteners 1992 2 New York: M Dekker 349 379
Kennefick S Cashman KD Investigation of an in vitro model for predicting the effect of food components on calcium availability from meals Int J Food Sci Nutr 2000 51 45 54 10746104 10.1080/096374800100895
|
16042782
|
PMC1183221
|
CC BY
|
2021-01-04 16:29:57
|
no
|
BMC Oral Health. 2005 Jul 22; 5:6
|
utf-8
|
BMC Oral Health
| 2,005 |
10.1186/1472-6831-5-6
|
oa_comm
|
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-281604276810.1186/1471-2431-5-28Research ArticleRisk of cardio-respiratory abnormalities in preterm infants placed in car seats: a cross-sectional study Ojadi Vallier C [email protected] Anna [email protected] Rajeev [email protected] Thomas [email protected] Department of Pediatrics, Division of Neonatal Medicine, Robert Wood Johnson Medical School / University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey, U.S.A2005 21 7 2005 5 28 28 28 12 2004 21 7 2005 Copyright © 2005 Ojadi et al; licensee BioMed Central Ltd.2005Ojadi 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 factors that predispose to the occurrence and severity of cardio-respiratory symptoms during the placement of a prematurely born infant in a car seat. The impact of gestational age, weight at discharge and infant's pre-existing cardio-respiratory status (in the supine position) on cardio-respiratory function during pre-discharge testing in a car seat (semi-upright position) has not been investigated.
Methods
The cardio-respiratory function of 42 preterm neonates with gestational age 24 to 35 weeks and discharge weight 1790 to 2570 grams were monitored for 45 minutes before, during, and after placement in a car seat. The occurrence of periodic breathing, apnea, bradycardia, or decreased oxygen saturation (SaO2) was analyzed.
Results
Prior to the car seat testing, 15 (35.7%) infants displayed one or more abnormalities of cardio-respiratory function. During the car seat testing, 25 (59.6%) infants had periodic breathing, 33 (78.2%) had oxygen saturation <90%, 14 (33.3%) had bradycardia less than 80 beats per minute, and 35 (83.3%) had a combination of these symptoms. Infants, both with and without pre-existing cardio-respiratory abnormalities, had an almost equal probability (80% vs. 83.3%) for the development of cardio-respiratory symptoms during placement in the car seat. Weight at discharge ([less than or equal to] 2,000 grams) but not the gestational age (<28 weeks or [greater than or equal to] 28<37 weeks), was associated with either increased episodes of oxygen desaturation or the combination of cardio-respiratory symptoms that were seen during the placement of these infants in the car seat. Repositioning from the car seat to the supine position showed normalization of cardio-respiratory function in the majority (83%) of the tested infants. None of the tested clinical factors were associated with the severity of the cardio-respiratory symptoms.
Conclusion
Pre-discharge testing of the cardio-respiratory function of preterm infants during placement in a car seat is important for the prevention of cardio-respiratory symptoms during their transportation. However, the high risk for developing cardio-respiratory symptoms will require the consideration of an alternative mode of safe home transportation for preterm infants; especially those with a discharge weight less than 2,000 grams.
==== Body
Background
The American Academy of Pediatrics (AAP) recommends pre-discharge testing of cardio-respiratory function during placement in a car seat for all infants less than 37 weeks in order to identify the safe mode of home transportation [1]. Several studies have shown that cardio-respiratory function is compromised in 18.4% to 30.0% of preterm infants tested in a car seat [2-6] possibly due to excessive head flexion leading to restriction of the upper airway [7]. Preterm infants are particularly prone to head flexion when they are placed in an upright position and therefore more vulnerable to hypoxia and apnea when their neck is flexed [8].
Despite this improved knowledge, many questions regarding the safe transport of preterm infants remain unresolved. There is wide inter-hospital variation in the duration of car seat testing and the recommendations for infants who fail the test [9]. The risk factors that contribute to the development and severity of cardio-respiratory symptoms in preterm neonates during the car seat testing are still poorly defined. Only pre-existing apnea of prematurity is clearly recognized as a risk factor [5] and few others have been systematically studied.
Identification of neonates at high risk for intolerance to car seat testing is important in order to develop appropriate recommendations for a safe mode of home transportation. We hypothesized that the infant's maturity and cardio-respiratory status in the supine position would influence the risk as well as the severity of cardio-respiratory symptoms during pre-discharge car seat testing. The measurements recommended by the AAP [1] for the assessment of cardio-respiratory function of an infant placed in a car seat were included in the protocol. Additionally, the infant was monitored for periodic breathing because of its possible association with oxygen desaturation [10,11] as well as cyclical desaturation and reoxygenation of cerebral blood [12] in the supine position.
Methods
A convenience sample of 42 preterm infants with gestational age ranging from 24 to 35 weeks (30.4 ± 3.3 weeks), birth weight of 550 to 2570 grams (1486 ± 523 grams), and discharge weight of 1790 to 2570 grams (2039 ± 215 grams) was studied 24–48 hours prior to discharge from the Neonatal Intensive Care Unit (NICU) at Saint Peter's University Hospital. The inclusion of 42 preterm infants would allow the detection of a thirty percent relative difference in the proportion of cardio-respiratory symptoms during placement in the car seat as compared to the supine position with a statistical power of 0.95 and Alpha <0.05 (two-sided).
Of the 42 sampled infants, 57.1% (n = 24) were white, 19.1% (n = 8) were black, 14.3 (n = 6) were Hispanic, and 9.5% (n = 4) were of other racial or ethnic groups; 4 (9.5%) were 24 weeks, 12 (28.6%) were 25 to 28 weeks, and 26 (61.9%) were 29–34 weeks gestation. The demographic characteristics of the study groups corresponded with the neonatal population discharged from the NICU at Saint Peter's University Hospital [13]. The study participants did not have any congenital anomalies, chronic cardiac, neurological or respiratory dysfunction, and at the time of testing none of them required oxygen supplementation or caffeine treatment for apnea of prematurity or any medications for gartroesophageal reflux. During the study period three of the discharged infants were not asked to participate in this study because one of the above mentioned conditions had been diagnosed during the NICU stay.
The study was approved by the Institutional Review Board (IRB) at the Robert Wood Johnson Medical School and the Saint Peter's University Hospital Committee for the Protection of Human Subjects in Research. Participation in the study required an informed consent to be signed by the parents. None of the mothers refused their baby's participation in the study.
Study procedures
The study procedure included automatic monitoring of heart rate (HR), respiratory rate (RR), and oxygen saturation (O2 Sat) for 45 minutes in each of the three positions: regular bassinet (supine), followed by the car seat (semi-upright), and after return to the bassinet (supine). Infants were studied between feedings and left undisturbed during the testing. As in a previous study [3], the Cosco-Peterson First Ride car seat for infants (Cosco, Columbus, IN, USA) was used to construct the semi-upright position. ECG electrodes connected to a Health Dyne Smart 1600 monitor (Healthdyne Technologies, Marietta, GA, USA) interfaced with an Oxford Pneumocardiogram recorder (Oxford Medilog Systems, Abingdon, UK)) were used for the monitoring of the HR and RR. Oxygen saturation (%) was measured with a Nellcor Pulse Oximeter (Nellcor, Hayward CA, USA).
The infant's tolerance for the car-seat position was assessed by monitoring for apnea, bradycardia and decreased oxygen saturation that was based on accepted definitions as meeting at least 1 of the following criteria: apnea ≥ 20 seconds [6], bradycardia ≤ 80 beats per minute [14], or oxygen saturation <90% [15]. In addition, the monitoring and analysis of periodic breathing (defined as repeated ventilatory cycles with a clear respiratory pause of at least 2 seconds between cycles) was included in the study [16,10]. Severe cardio-respiratory symptoms were defined as a fall in oxygen saturation to below 80% [17] and a fall in heart rate to less than 70 beats per minute [18] respectively.
Data analysis
The cardio-respiratory status prior to placement in the car seat (without or with any of the recorded abnormalities), the gestational age at birth (<28 vs. 28 to <37 weeks), and weight at discharge (< = 2,000 vs. >2,000 grams) were used to analyze the data. Postmenstrual age (PMA) at discharge was nearly identical across the gestational age groups and therefore was not used for the analysis. Chi-square and one-way ANOVA were used to compare categorical and continuous data, respectively. A conditional regression analysis was performed to evaluate the impact of gestational age, weight at discharge and the presence of cardio-respiratory symptoms in the supine position, on the infant's risk for developing cardio-respiratory symptoms during placement in the car seat. Statistical analysis was performed using Statistica 6.0 (StatSoft, Tulsa, OK, USA) and StatsDirect 2.4.4 (StatsDirect, Cheshire, UK). All data in the text are presented as mean ± standard deviation for continuous variables, as percentiles to identify the proportion, and crude as well as adjusted Odd Ratios (ORC and ORA) with 95% Confidence Interval (95% CI). P value <0.05 was considered statistically significant.
Results
During testing in the supine position for 45 minutes before placement in the car seat, 35.7% (15/42) of the infants displayed one or more abnormalities (Table 1). The symptoms were generally mild and solitary except in the few infants (3/15) who had 2–3 documented episodes. The infants with cardio-respiratory symptoms recorded before placement in the car seat had been born at a lower gestational age and birth weight (28.0 ± 2.5 weeks and 1715 ± 500 grams, respectively) as compared to infants without cardio-respiratory symptoms (31.8 ± 2.8 weeks and 1105 ± 293 grams, respectively), P < 0.0001, but the weight at discharge was similar (1974 ± 175 vs. 2078 ± 230, P>0.05).
Table 1 Comparison of number of preterm neonates with cardio-respiratory symptoms before, during, and after placement in a car seat (n, %)
Placement in the car seat
Cardio-Respiratory Symptom Before (1) (n = 42) During (2) (n = 42) After (3) (n = 42) P1–2 P2–3
Periodic Breathing 9 (21.4%) 25 (59.5%) 2 (4.8%) 0.01 0.001
Apnea 0 2 (4.8%) 0 NS NS
Bradycardia (HR ≤80/min) 4 (%) 14 (33.3%) 0 0.001 NS
O2 Saturation<90% 9 (21.4%) 33 (78.6%) 7 (16.7%) 0.01 0.01
Combination* 10 (23.8%) 35 (83.3%) 3 (7.1%) 0.001 0.001
* Presence of any cardio-respiratory symptoms
As shown in Table 1, cardio-respiratory symptoms were recorded with a greater frequency during placement in the car seat, and the majority of infants (83.3%, P < 0.001) showed a combination of cardio-respiratory symptoms. Placement of infants back in the supine position generally resulted in the resolution of the cardio-respiratory symptoms. During the car seat testing, oxygen desaturation <80% was observed in 13 (31.0%) of the tested infants and occurred more than 3 times in 11 (26.2%) infants. Bradycardia <70 beats per minutes was recorded in 4 (9.5%) of the infants, and 5 (11.9%) had multiple episodes. There was a significant association between periodic breathing and falls in oxygen saturation to less than 80% (P = 0.02, Figure 1).
Figure 1 Periodic breathing associated Oxygen Desaturation (O2 Sat, %).
As shown in Table 2, before placement in the car seat the cardio-respiratory symptoms were more frequent among infants with gestational age <28 weeks and pre-discharge weight < = 2,000 grams, but during the car seat testing cardio-respiratory events occurred more frequently in neonates with pre-discharge weight < = 2,000 grams. Conditional regression analysis showed that infants with a pre-discharge weight < = 2,000 grams were at greater risk for the occurrence of cardio-respiratory symptoms (ORC 4.23, 95%CI 1.02, 25.0). The risk of cardio-respiratory symptoms increased when gestational age and cardio-respiratory symptoms in the supine position were taken into account but did not reach statistical significance (ORA 11.6, 95%CI 0.94, 141.3). The infant's weight at discharge < = 2,000 grams did not affect the severity of the recorded cardio-respiratory symptoms during the car seat testing. Approximately one-third of the infants in both the discharge weight groups developed severe oxygen desaturation (<80%) and had multiple episodes (more than 3). Overall, only 2 infants with gestational age <28 weeks, pre-discharge weight < = 2,000 grams and no evidence of cardio-respiratory symptoms before placement in the car seat, developed severe apnea during their first 20 minutes of testing.
Table 2 The differences in cardio-respiratory symptoms before and during car seat testing in association with symptoms in supine position, gestational age and pre-discharge weight (n, %).
Cardio-respiratory symptom Car seat testing Symptoms while supine Gestational age (weeks) Pre-discharge weight (grams)
Symptoms (+) (n = 15) Symptoms(-) (n = 27) <28 (n = 16) 28–34 (n = 26) < = 2,000 (n = 24) >2,000 (n = 18)
Periodic Breathing before 9 (60.0%) 0 7 (43.8%) 2 (7.7%) † 5 (20.8%) 4 (22.2%)
during 9 (60.0%) 16 (59.3%) 9 (56.3%) 16 (61.5%) 16 (66.7%) 9 (50.0%)
Bradycardia (HR ≤80/min) before 4 (26.7%) 0 2 (12.5%) 2 (7.7%) 3 (12.5%) 1 (5.6%)
during 5 (33.3%) 9 (3.3%) 5 (31.3%) 9 (34.6%) 9 (37.5%) 5 (27.8%)
O2 Saturation <90% before 9 (60.0%) 0 5 (31.3%) 4 (15.4%) † 8 (33.3%) 1 (5.6%) †
during 13 (85.2%) 20 (74.1%) 12 (75.0%) 21 (80.8%) 20 (83.3%) 13 (72.2%)
Combination * before 10 (66.7%) 0 6 (37.5%) 4 (15.4%) † 9 (37.5%) 1 (5.5%) †
during 12 (80.0%) 23 (85.2%) 14 (87.5%) 21 (80.8%) 22 (91.7%) 13 (72.2%) ††
*Presence of any cardio-respiratory symptoms
† (P < 0.05) differences in frequency of cardio-respiratory symptoms in supine position in association with gestational age and pre-discharge weight.
†† (P < 0.05) differences in frequency of cardio-respiratory symptoms during car seat testing in association with symptoms in supine position, gestational age, and pre-discharge weight.
Discussion
The following conclusions can be made from this study: (i) during the 45 minutes of pre-discharge car seat testing, the majority of preterm neonates are at risk for the development of mild and infrequent episodes of cardio-respiratory symptoms such as oxygen desaturation and periodic breathing; (ii) there is a higher risk for the occurrence of cardio-respiratory symptoms during car seat testing in infants with a pre-discharge weight < = 2,000 grams that seemed to be related to lower gestational age at birth and an elevated rate of symptoms when supine; (iii) the presence of periodic breathing significantly increases the risk for the development of severe hypoxic events (oxygen saturation <80%) in preterm neonates placed in the car seat; (iv) repositioning from the car seat to the supine position eliminates the cardio-respiratory symptoms in the vast majority of tested infants.
Our findings are in accord with other investigators who showed a high rate of cardio-respiratory abnormalities, especially the development of oxygen desaturation during car seat testing [2-4,19,20]. For the first time we report that severe oxygen desaturation in infants during car seat testing is associated with episodes of periodic breathing. Because of the transient immaturity of respiratory control [21], periodic breathing occurs frequently in preterm infants and has been shown to be associated with oxygen desaturation even in the supine position [10,11]. The association between periodic breathing and cyclical desaturation as well as reoxygenation of the cerebral blood in term infants has been previously reported [12]. Yamamoto et all showed that during apneic attacks in preterm infants, oxygen saturation <85% was associated with a reduction in the cerebral blood flow and cerebral oxygenation, measured indirectly using near infrared spectroscopy [22]. Although monitoring for periodic breathing was not included in the AAP recommendations for pre-discharge car seat testing of preterm infants [1], our results strongly suggest that it may be of considerable value in identifying infants at high risk for oxygen desaturation during home transport.
The following aspects of our study may limit the interpretation of the results: (i) infants with chronic cardiologic, neurological, or respiratory dysfunction were not included the study population, however, the vast majority of prematurely born infants are discharged from the NICU in a normal condition; therefore, the number of infants with such complications at discharge would be very small; (ii) the duration of the cardio-respiratory events was not included in the results, because a wide variation was observed in the range (between 10–20 seconds) for both bradycardia and oxygen desaturation; and (iii) the stage and duration of sleep was not tested. A study showed that the frequency of adverse cardio-respiratory events increases during sleep [23]. There is delayed maturation and poor organization of sleep states in preterm infants [24-26]. Therefore, if preterm babies are left asleep in a car seat they are likely to be at an even greater risk for the occurrence of cardio-respiratory symptoms and ensuing episodes of oxygen desaturation. This is of particular concern since there is evidence that a high percentage of episodes of apnea and bradycardia are not detected clinically [27].
Our results may have important implications for the standardization of pre-discharge car seat testing protocols for preterm born infants. The findings of periodic breathing accompanied by oxygen desaturation during the pre-discharge car seat testing of preterm born infants should lead to the consideration of an alternative mode of home transportation. There is always the option of supine transport in a car bed. Nevertheless, newer approaches for the safe transportation of preterm born infants need to be explored. One possible approach has been described by Tonkin and his colleagues [19] who modified an infant car seat such that it allowed the infant's head to rest in the neutral position on the trunk and therefore prevented narrowing of the upper airway, which was associated with reduced frequency of oxygen desaturation in preterm infants restrained in car seats.
Conclusion
Pre-discharge testing of the cardio-respiratory function of preterm infants during placement in a car seat is important for the prevention of cardio-respiratory symptoms during their home transportation. The findings of periodic breathing accompanied by oxygen desaturation during the pre-discharge car seat testing of preterm born infants should lead to the consideration of an alternative mode of home transportation; especially those with a discharge weight less than 2,000 grams.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
VO participated in the study design, data collection, and carried out the study procedure. AP performed the statistical analysis and drafted the manuscript. RM participated in the coordination, interpretation of the data and drafting of the manuscript. TH conceived the study and participated in its design. 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 would like to acknowledge the contribution of Catherine Diala, RN, BSN who assisted with the study procedures.
==== Refs
Safe transportation of premature and low birth weight infants : American Academy of Pediatrics Committee on Injury and Poison Prevention and Committee on Fetus and Newborn Pediatrics 1996 97 758 760 8628626
Bass JL Mehta KA Camara J Monitoring premature infants in car seats implementing the AAP policy in the community Pediatrics 1993 91 1137 1141 8502516
Willett LD Leuschen MP Nelson LS Nelson RM Ventilatory changes in convalescent infants positioned in car seats J Pediatr 1989 115 451 455 2769505
Bull MJ Weber K Automotive restraint systems for premature infants J Pediatr 1988 112 385 388 3346776
Willet LD Leuschen MP Nelson RM Risk of hypoventilation in premature infants in car seats J Pediatr 1986 109 245 248 3734961
Bass JL Mehta KA Oxygen saturation of selected term infants in car seats Pediatrics 1995 96 288 290 7630686
Tonkin SL Bennet L Vogel S Gunn AJ Apparent life-threatening events associated with positional airways obstruction in infancy Pediatr Res 2000 47 120A
Stark AR Thatch BT Mechanisms of airway obstruction leading to apnea in newborn infants J Pediatr 1976 89 982 985 993928
Williams LE Martin JF Car seat challenges: where are we in implementation of these programs? J Perinat Neonatal Nurs 2003 17 158 163 12822703
Miller MJ Carlo WA DiFiore JM Martin RJ Airway obstruction during periodic breathing in premature infants J Appl Physiol 1988 64 2496 2500 3403433
Poets CF Southall DP Patterns of oxygenation during periodic breathing in preterm infants Early Hum Dev 1991 26 1 12 1914983 10.1016/0378-3782(91)90038-5
Urlesberger B Pichler G Gradnitzer E Reiterer F Zobel G Muller W Changes in cerebral blood volume and cerebral oxygenation during periodic breathing in term infants Neuropediatrics 2000 31 75 81 10832581 10.1055/s-2000-7477
Petrova A Mehta R Anwar M Hiatt M Hegyi T Impact of race and ethnicity on the outcome of preterm infants below 32 weeks gestation J Perinatol 2003 23 404 408 12847537 10.1038/sj.jp.7210934
Martin RM Abu-Shaweesh Baird MT Apnea of Prematurity Pediatr Resp Rev 2004 5 S377 S382 10.1016/S1526-0542(04)90067-X
Barrington KJ Finer NN Periodic breathing and apnea in preterm infants Pediatr Res 1990 27 118 121 2314939
Graff M Soriano C Hiatt M Hegyi T Undetected Apnea and bradycardia in preterm infants Pediatr Pulmonol 1991 11 195 197 1758738
Poets CF Stebbens VA Alexander JR Arrowsmith WA Salfield SA Southall DP Arterial oxygen saturation in preterm infants at discharge from the hospital and six weeks later J Pediatr 1992 120 447 454 1538297
Southall DP Richards JM Rhoden KJ Alexander JR Shinebourne EA Arrowsmith WA Cree JE Fleming PJ Goncalves A Orme RL Prolonged apnea and cardiac arrhythmias in infants discharged from neonatal intensive care units: failure to predict an increased risk for sudden infant death syndrome Pediatrics 1982 70 844 851 7145536
Tonkin SL McIntosh CG Hadden W Dakin C Rowley S Gunn AJ Simple car seat insert to prevent upper airway narrowing in preterm infants: A pilot study Pediatrics 2003 112 907 913 14523185 10.1542/peds.112.4.907
Hertz G Aggarwal R Rosenfeld WN Greensher J Premature infants in car seats: effect of sleep state on breathing J Sleep Res 1994 3 186 190 10607125
Glotzbach SF Ariagno RL Beckerman RC, Brouillette RT, Hunt CE Periodic breathing: Respiratory Control Disorders in Infancy and Children 1992 Baltimore, MD: Williams & Wilkins 142 160
Yamamoto A Yokoyama N Yonetani M Uetani Y Nakamura H Nakao H Evaluation of change of cerebral circulation by SpO2 in preterm infants with apneic episodes using near infrared spectroscopy Pediatr Int 2003 45 661 664 14651537 10.1111/j.1442-200X.2003.01803.x
Horne RSC Sly DS Cranage TM Effects of prematurity on arousal from sleep in the newborn infant Ped Res 2000 47 468 474
Nagase H Yonetani H Uetani Y Nakamura H Effects of child seats on the cardio-respiratory function of newborns Pediatr Intern 2002 44 60 63 10.1046/j.1442-200X.2002.01507.x
Simakajornboon N Beckerman RC Mack C Sharon D Gozal D Effect of supplemental oxygen on sleep architecture and cardiorespiratory events in preterm infants Pediatrics 2002 110 884 888 12415025 10.1542/peds.110.5.884
Scher MS Steppe DA Dahl RE Asthana S Guthrie RD Comparison of EEG sleep measures in healthy full-term and preterm infants at matched conceptional ages Sleep 1992 15 442 448 1455128
Razi NM Humphreys J Pandit PB Stahl GE Predischarge monitoring of preterm infants Pediatr Pulmonol 1999 27 113 116 10088934 10.1002/(SICI)1099-0496(199902)27:2<113::AID-PPUL7>3.0.CO;2-O
|
16042768
|
PMC1183222
|
CC BY
|
2021-01-04 16:31:06
|
no
|
BMC Pediatr. 2005 Jul 21; 5:28
|
utf-8
|
BMC Pediatr
| 2,005 |
10.1186/1471-2431-5-28
|
oa_comm
|
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-781604279610.1186/1471-2458-5-78Research ArticlePsychophysiological effects of a web-based stress management system: A prospective, randomized controlled intervention study of IT and media workers [ISRCTN54254861] Hasson Dan [email protected] Ulla Maria [email protected] Töres [email protected] Bengt B [email protected] Uppsala University, Department of Public Health and Caring Sciences, Section for Social Medicine/CEOS, Uppsala Science Park, SE-751 85 Uppsala, Sweden2 IPM – The National Swedish Institute for Psychosocial Medicine, Granits väg 8, SE-171 77 Stockholm, Sweden2005 25 7 2005 5 78 78 11 2 2005 25 7 2005 Copyright © 2005 Hasson et al; licensee BioMed Central Ltd.2005Hasson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The aim of the present study was to assess possible effects on mental and physical well-being and stress-related biological markers of a web-based health promotion tool.
Methods
A randomized, prospectively controlled study was conducted with before and after measurements, involving 303 employees (187 men and 116 women, age 23–64) from four information technology and two media companies. Half of the participants were offered web-based health promotion and stress management training (intervention) lasting for six months. All other participants constituted the reference group. Different biological markers were measured to detect possible physiological changes.
Results
After six months the intervention group had improved statistically significantly compared to the reference group on ratings of ability to manage stress, sleep quality, mental energy, concentration ability and social support. The anabolic hormone dehydroepiandosterone sulphate (DHEA-S) decreased significantly in the reference group as compared to unchanged levels in the intervention group. Neuropeptide Y (NPY) increased significantly in the intervention group compared to the reference group. Chromogranin A (CgA) decreased significantly in the intervention group as compared to the reference group. Tumour necrosis factor α (TNFα) decreased significantly in the reference group compared to the intervention group. Logistic regression analysis revealed that group (intervention vs. reference) remained a significant factor in five out of nine predictive models.
Conclusion
The results indicate that an automatic web-based system might have short-term beneficial physiological and psychological effects and thus might be an opportunity in counteracting some clinically relevant and common stress and health issues of today.
==== Body
Background
Stress-related disorders are major public health issues in many industrialized countries and are expected to become increasingly common in the coming decades [1,2]. Such disorders have a negative economic impact, disrupt work and home life and might even increase suicide risk [3]. Numerous web-based health sites and tools are being offered to the public for stress management and treatment of stress-related mental conditions such as depression and anxiety. More and more people are rapidly using these sites and tools. The majority of the "health seekers" rely on search engines and seldom check the source and date of online health information [4,5]. According to Fox & Fallows (2003), about 93 million Americans (half of American adults) have searched online for health information. Few prospectively controlled intervention studies have been published on the efficacy of these health sites and web-based tools. Even fewer of them have assessed both psychological and physiological effects. There are increasing indications, however, that Internet-based intervention programs have beneficial effects on various psychological conditions and other desired outcomes [6-8]. Moreover, prospectively controlled studies without physiological evaluation have indicated beneficial effects from computer- or web-based tools on, for example, headache [9], distress related to tinnitus [10], depression and anxiety [11-13], stress management [14], physical activity [15] and insomnia [16].
Considering these indications from previous studies, web-based interventions for stress management and health promotion may offer promising opportunities. Some possible advantages with web-based interventions compared to more traditional alternatives, such as books, coaches and therapists may be the 24-hour accessibility, possibility for interaction, instant feedback and support. Moreover, the scalability and potential reach of web-based interventions may further be an advantage in economical terms for individuals, corporations and the society. One major disadvantage and risk, however, is the lack of quality assurance of web-based health sites and interventions. There are no international agreed upon guidelines for assessment, and users of these services may receive misleading or incorrect information that may potentially be harmful to health and wellbeing [17].
The present study was conducted during a stressful period for the IT and media companies in Sweden. For the information technology companies, there was downsizing after the dot-com bubble burst. The media companies worked intensively in covering the election campaign to the Swedish parliament. The study population, however, had been chosen to be representative for a future that employees will face more and more frequently: increasing pace of changes, shorter status quo time, "project work" and other challenges that were new to knowledge workers. This implies that the results from the present study may be applicable more generally to employees in the future: How they will react in a stressful situation that directly affects the basic conditions of one's organization and workplace.
A multitude of biological and physiological markers have proved to be related to stress, health and recovery. Some markers have been thoroughly investigated in various studies, whereas there is limited information on others. Moreover, relationships between biological markers and stress, health and recovery seem to be complex since many factors may affect their patterns of secretion, including negative feedback and secretion of other related hormones [18]. For some hormones, such as cortisol, immune markers and sex hormones, it is also essential to consider seasonal changes and natural variation in daily cycles [19]. Clearly, more knowledge is needed regarding longitudinal relationships between biological markers, stress, health and recovery. Most of the biological and physiological markers to be analyzed in the present study have been assessed in relation to short- and long-term stress in previous studies. Some of the markers, such as blood status, were to be routinely sampled for overall health matters or general profiling that could indicate alterations in plasma volume.
The aim of the present study was to assess the possible effects on mental and physical well-being and biological stress markers from a web-based stress management and health promotion tool. It was expected that decreased levels in indicators of catabolism and increased levels in indicators of anabolism would be found in the intervention group compared to the reference group at the end of the study. To our knowledge, this is the first completely web-based assessment, where self-ratings are complemented with biologically relevant outcome data.
The web-based system was developed and adjusted with the aim of making it useful as a tool for everyday life, including usage on a daily or regular basis. Therefore it had to be easy to use, time-efficient and accessible through the work place and at home 24 hours a day. The participants were recruited at worksites since the study included some organizational aspects, such as health economics, not presented in this paper.
Methods
Participants and study groups
Flow of participants throughout the study is depicted in Figure 1. In collaboration with a White-Collar Union (Sif) and a Swedish Employers' Association (Almega), ten companies insured by the study's source of funding Alecta (an occupational pension plan company) were asked as to their interest in participating. The asked companies were selected and contacted by employees at Alecta, by mail and phone. The management departments of six out of the ten asked companies were interested. Informed of the basic inclusion criteria, i.e. minimum group of ten individuals and access to economic production data, 2–4 departments within each company were chosen and asked by the company management as to their interest in participating. The managers of the selected departments in turn asked their employees whether they were interested in participating. No incentives were offered to the participants, with exception of the extensive blood sampling including feedback of the results, which seemed to be a motivator for many participants.
Figure 1 Flow of participants. The figure illustrates the flow of participants through each stage of the randomized trial. Additionally, the numbers of excluded participants and drop-out rates, including generalized reasons for these actions are depicted.
For some companies the departments were located in different cities and for some in the same buildings or city. With the exception for one of the media companies where a whole department with five units enrolled, there was no "natural" connection between the participating departments. Consequently, there was only occasional risk of contamination of the extended intervention to the reference group. Moreover, the design of the websites for the intervention and reference group respectively was similar in appearance (see more information on the interventions below). The intervention group website only had two additional buttons, which makes it hard to notice any difference if the site would be exposed to a participant of the reference group.
The participating departments were, within each company, randomized by lottery to either the intervention or reference group. Thus each company had at least one intervention and one reference group. All departments received a 30-minute information session including 10 minutes for questions and answers. These information sessions included the aim of and general information about the study as well as general information on stress and health. Finally, after oral information each participant received written information about the project and consent forms. All the participants were informed, orally as well as in writing, that participation was voluntary and withdrawal was possible at any time.
There is no information on the exact number of employees that were asked to participate in the study. An exception was one of the media companies where 95 out of 100 possible participants chose to participate. In general there was also a great interest from the other departments and similar participation rates is therefore estimated. Altogether, 317 participants from 22 departments/units in four information technology and two media companies enrolled in the study. Fourteen participants were excluded because of communication-related problems (n = 7), change of mind to participate or quit their job before initiation of the study (n = 7). Thus 303 persons finally participated in the study, out of which 26 participants (8.6%) dropped out. The reasons for dropping out were job termination (n = 7), change of workplace (n = 2), foreign service or moving abroad (n = 6) or other reasons (n = 11). There were no significant differences in dropout rates between the groups (6.9% in the intervention group vs. 9.8% in the reference group, p between groups = n.s.). Nor were there any significant differences between the intervention and reference groups in socioeconomic background or psychophysiological measures at baseline. Regarding the participants, there was no information about possible mental or somatic disorders or medication.
The participants had professions such as IT technicians, programmers, system developers as well as journalists/reporters, news presenters, sound technicians and photographers. The main type of work-site was open plan offices. Many participants from the IT-companies were partly located in the work sites of their customers for longer or shorter periods. For the media companies, some participants, such as photographers and reporters, were partially ambulatory and worked in different locations. The common feature for all participants was regular and daily computer usage at work.
The web-based tool
Table 1 provides a detailed description of the web-based tool and illustrates similarities and differences in the features that were offered to the intervention and reference group respectively. A web-based tool for health promotion and stress management was developed and offered all participants real-time monitoring of perceived current health and stress status, a diary and information about stress and health (Table 1). In addition, participants in the intervention group were offered web-based cognitive exercises, aimed at decreasing unwanted stress and promoting health and recovery through health promotion initiatives. The exercises included techniques for relaxation, time management, cognitive reframing and a chat. Thus, the only things that distinguished the groups were the addition of the cognitive exercises and the chat in the intervention group. The web-based tool was developed by the researchers and most techniques are commonly utilized techniques in cognitive and behavioral psychology and stress management. These techniques were modified so that they could become more or less self-instructing to be used for self-help purposes.
Table 1 The web-based tools. The table depicts the different features included in the web-based tool for the study groups respectively. The only things that distinguished the groups were the addition of the cognitive exercises and the chat in the intervention group.
Feature Intervention group Reference group
Monitoring tool for stress and health levels with instant feedback; graphs illustrating current and retrospective ratings and an option to compare results with other groups with the same socioeconomic profile, within the same department/company and all the respondents in the data base. The questionnaire was compiled by a ten-item questionnaire for regular or daily usage. YES YES
Diary connected to the monitoring tool so that ratings and notes could be compared and examined retrospectively. The diary could be used as stress management but also as a tool for improving self-knowledge and how different events affect health and well-being. YES YES
Popular scientific information on stress and health compiled by various Swedish researchers. YES YES
Self-help in the form of classical stress management exercises for; relaxation and sleep improvement, cognitive reframing, time-management, emotional control and self-knowledge, strengthening self-esteem, life reflection, dissociation. YES NO
Chat YES NO
The web-based tool and the exercises were not pilot tested before the study. However, the tool as well as exercises were chosen and adjusted on the basis that they had to be time efficient in order to be utilized. It was hypothesized that basic demands for regular usage were instant feedback on the questionnaire and that the measurement and exercises could be used rapidly. Consequently, it was decided that regular or daily monitoring should not take more than 20–40 seconds. Moreover, every exercise was labeled with information of time for accomplishment (time span 1–60 minutes). Some of the cognitive exercises, e.g. improving self-confidence, were designed such that they consumed 5–10 minutes when learning and then could be conducted in a matter of seconds when utilizing.
Most exercises were presented in three different modes; on the web-site as plain text, as a downloadable PDF-file (sometimes including descriptive images), and as a flash animation, guiding the participant with image and sound through the exercise. Since the intervention for both groups was completely web-based it could only be accessed online. All information was however printable, which made it possible for the participants to print material of interest and thus intervene elsewhere. Exposure to the intervention for both groups could only be logged via the number of logins to the website.
Questionnaire
A questionnaire was compiled and included about 100 questions concerning socioeconomic status, consumption of caffeine drinks, expectations about the research project, self-rated health (SRH), stress and wellbeing at work as well as during leisure time, health economics and performance at work (Table 2). Most of the questions were presented as Visual Analogue Scales (VAS) and some, concerning health economy, work time, basic daily functioning and symptoms of ill health, were presented as multiple-choice questions. Most of the newly constructed single VAS questions were based on previously validated Likert-based items or indices [20-25]. Participants filled out the questionnaire online at baseline (before the initiation of the study) and at the end of the six-month intervention.
Table 2 Questionnaire. The table illustrates theoretical models, items and topics covered by the questionnaire. Most items were presented as "straight forward" VAS, e.g. How is your overall sleep quality (Very poor – Very good).
Models Topics – generalized self-ratings
Socioeconomic and background factors Age, sex, annual income and self-rated financial situation, educational level, marital status, possession of children, work role (co-worker, middle-manager, manager), amount of customer contact, duration of current working position, smoking habits, satisfaction with eating habits, consumption of coffee, tea, soft drinks and energy drinks. Expectations of the possible effects of the research project on stress and health level.
Lifestyle, health promoting and compromising behaviours, cognitive function, sense of coherence and wellbeing Self-rated health (last year, right now and future expectations), sleep quality, memory, concentration ability, ache in various body parts, physical exercise habits, mental energy, frequency and source (home, work or combination) of stress, stress management ability, satisfaction with leisure-time, life goals, communication ability with others, meaningful life, future optimism/pessimism, flexibility, daily computer, phone and cellular phone usage, social support, reflection on health improvement.
Work-related factors, demand/control, effort/reward Work satisfaction, efficiency, competence (sufficiency, development, usage), meaningful work, work atmosphere, work intensity, number of breaks during a regular working day, average working hours and distribution over the week (actual and desired), flexibility of work, general mood on the way to work (sad – happy), working effort, work reward, influence on work situation, work stress, work confidence, support from managers, collegial support, work-place goal clarity and realism, work-place efficiency, reflection on efficiency improvement, priority between health and achievement, time perspectives on decisions at work, existence of serious considerations to quit job, number of sick-leave days, health-economic aspects.
Blood sampling
The complete list of biological markers analyzed in the current study is presented in Table 3. More biological markers of general nature, such as blood status, were collected for overall health matters or all-purpose profiling. These markers were not analyzed in the present study. Furthermore, P-substance P, S-IL-1beta and P-endothelin were also collected. However, in the first measurement there was not enough blood collected to render the exact results needed for more sensitive analyses of these variables, resulting in the decision to not include them in the present study. Thus, the biological markers analyzed in the present study were only the ones that could be related to various stress-related hypotheses.
Table 3 Blood sampling and physiological measures. The table illustrates the biological markers and physiological measures sampled at baseline and after the six months intervention.
Categories Physiological marker
Cardiovascular system and lifestyle Blood pressure, pulse, waist-hip ratio, BMI, P-BNP (brain natriuretic peptide), P-PAI-1 (plasminogen activator inhibitor 1), S-insulin, B-HbA1C, S-triglycerides, S-cholesterol, S-HDL, S-LDL, P-fibrinogen, B-trombocytes.
Stress-related (HPA-axis, catabolic) S-prolactin, P-ACTH (adreno corticotropic hormone), S-cortisol, S-TSH (thyroid stimulating hormone), S-T3, S-T4 (free), S-urate.
Recovery-related (anabolic) S-GH (growth hormone), S-IGF-1, S-DHEAS-S (dehydroepiandosterone sulphate), S-estradiol, S-testosterone, S-SHBG (sexual hormone binding globulin).
Immune markers and neuropeptides S-TNFα (tumour necrosis factor alpha), high sensitive S-CRP (c-reactive protein), P-NPY, P-CgA (chromogranin A).
Blood samples were collected from study participants between 7.00–11.30 am at each specific worksite (or nearby). Unfortunately, it was impossible for practical reasons to sample the blood within more narrow time limits. Questionnaires were filled out during the same time period (usually same day or week) in order for the outcome of the blood and questionnaire data to be as comparable as possible. The exact time for blood sampling was recorded for each participant at baseline and at the end of the study so that the blood could be collected at the same time (± 15 minutes). Participants were instructed not to eat or drink (except water), nor use nicotinic substances at least ten hours before blood sampling. The blood samples were analyzed by the Karolinska University Hospital laboratory that is qualified by SWEDAC (Swedish Board for Accreditation and Conformity Assessment) that accredits laboratories in the medical sector according to the standard ISO/IEC 17025. Intra-assay and inter-assay coefficients of variation can be obtained from the laboratory [email protected]) or by e-mailing the authors [email protected].
Statistical analyses
The program SPSS 11.5 for windows was used for statistical analyses and an intention-to-treat approach was utilized. This means that all subjects in both groups were included in the follow-up regardless of how much they participated in the intervention programs. And the evaluation is based upon the assumption that everybody – even those who did not participate at all – in the intervention group were compared with everybody in the control group.
Initially, all variables were assessed for normality using Kolmogorov-Smirnov test. Changes over time (time, group and group × time) were assessed using two-way analysis of covariance (ANCOVA). ANCOVA adjusts for initial differences so that the results more precisely reflect possible intervention effects, and thus permits a more sensitive analysis compared to regular analysis of variance (ANOVA). The increase in sensitivity arises from the fact that the covariance reduces the error term (within-group variability) against which intervention effects are compared. Furthermore, ANCOVA is not very sensitive to small deviations from a normal distribution [26]. In the present study, baseline values of the assessed variables were used as covariates. Analyses were in some cases, such as sex hormones stratified with a break up by gender. Finally, to adjust all results of the ANCOVA analyses for the possible effect of multiple comparisons (mass-significance), Bonferroni correction was utilized for each analysis.
Since VAS can be treated as an interval or ordinal scale [27-29] and all variables were not normally distributed, both parametric and non-parametric tests were used where statistically significant differences between the groups were detected in the ANCOVA. Thus, it was decided that changes over time and differences between the groups would only be considered in cases where both parametric and non-parametric tests unanimously were statistically significant. For the non-parametric analyses, new variables (so called Δ variables) based on change between the first and second measurements were constructed. Differences between the groups were then assessed using a Mann-Whitney U test.
Logistic regression was used to model the probability of improvement in the significant Δ variables (dependent variables). Factors such as socioeconomic status, marital status and gender are known to be associated with outcomes in health and stress and were therefore included as covariates in the first step of the regression analysis. Also group was included as a factor in the first step to adjust for possible study group effects. The dependent and independent VAS variables used in the logistic regression were divided by quartile split into high (top quartile) and low (remaining quartiles) categories. Similarly the number of logins was dichotomized by quartile split. The independent VAS variables were selected for two subsequent steps in the regression analysis. The rationale for the second step was to adjust for work related factors that might disturb the relationships, i.e. working hours per week, working atmosphere, work intensity and number of breaks during a working day. The third step included all the remaining dependent variables. In order for the physiological markers to render comparable odds ratio they were dichotomized by quartile split. The fourth and final step included the number of logins to adjust for possible effects of high vs. low frequency of logins to the website.
The specific hypothesis tested in the present study was that the intervention group would improve compared to reference group on biological stress markers and health- and recovery-related ratings captured by the questionnaire. Decreased levels in indicators of catabolism and increased levels in indicators of anabolism were expected in the intervention group compared to the reference group. Since both groups received an intervention, some beneficial changes in the reference group, e.g. in SRH, might be expected as well.
The ethics committees of Uppsala University (Dnr 01–188) and Karolinska Institute (Dnr 01355) approved the research project. A modified version of the web-based tool used in the present study can be found at .
Role of the funding source
The funding source had no involvement in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.
Results
Baseline socioeconomic characteristics of the study participants are described in Table 4. Exposure to the interventions, i.e. number of logins to the website, revealed that the intervention group used the website statistically significantly more compared to the reference group (Figure 2; t-test p < .001, 2-tailed; Mann-Whitney U test p < .0001). For the whole sample, the frequency of logins for the lowest quartile was < 10, median 36 and the top quartile >71 logins.
Table 4 Baseline socioeconomic characteristics. The table depicts the socioeconomic characteristics age, sex, education, annual income and marital status of the participants (n = 303) from the enrolling IT and media companies.
Characteristic Intervention group n = 129 % Reference group n = 174 %
Age
≤ 30 31 24 46 27
31–45 44 34 72 41
≥ 46 54 42 56 32
Sex
Male 75 58 112 64
Female 54 42 62 36
Education*
Compulsory school/High school 54 42 89 51
Academic degree 73 57 83 48
Annual income*
< 25,000 USD 24 18 39 22
25,000 – 40,000 USD 76 59 106 61
> 40,000 USD 27 21 27 16
Marital status*
Married/co-inhabiting/liveapart 102 79 134 77
Single 25 19 38 22
* 4 missing values (two in intervention and reference group respectively).
Figure 2 Website login frequency. This figure illustrates the number of logins on the website made by the intervention group (median 48 logins) and reference group (median 26 logins) respectively, during the study period of six months (p < .0001, 2-tailed).
At the end of the 6-month intervention period, the intervention group had improved significantly as compared to the reference group on ratings of perceived ability to manage stress, sleep quality, mental energy, concentration ability, social support and competence usage at work (2-way ANCOVA, p < .05 time × group effect). With the exception for competence usage at work, all these changes and differences between the groups remained significant when applying the non-parametric Mann-Whitney U test (p < .05, two-tailed). Figures 3a–i illustrate changes in self-rated measures and biological markers over time between the intervention and reference groups, respectively. Results shown are covariated for baseline scores of the depicted outcome variable. SRH increased significantly in both groups, with no differences between the groups (2-way ANCOVA, p < .0001 time effect; time × group effect non-significant). The results of the gender stratified variables were not different from the ones obtained when analyzing the non-stratified data.
Figure 3 a-i Two-way ANCOVAs. The figures illustrate the results of the two-way ANCOVA on the significant outcome measures: a) Stress management ability, b) Sleep quality, c) Mental energy, d) Concentration ability, e) Social support, f) DHEA-S, g) NPY, h) CgA and i) TNFα. All measures are covariated for their own baseline levels.
Concerning the analyzed blood samples, the levels of the sulphated metabolite of the hormone dehydroepiandosterone (DHEA-S) decreased significantly in the reference group, with no changes in the intervention group. The levels of Neuropeptide Y (NPY) increased significantly in the intervention group compared to the reference group. CgA (chromogranin A) and ACTH (adrenocorticotropic hormone) decreased significantly in the intervention group as compared to the reference group. The levels of the immune marker TNFα decreased significantly in the reference group as compared to the intervention group (2-way ANCOVA, p < .05 time × group effect). With exception for ACTH, all these changes in biological markers and differences between the groups remained significant when applying the non-parametric Mann-Whitney U test (p < .05, two-tailed).
Tables 5 and 6 depict the results of the logistic regression analyses, which were utilized to predict the quartile exhibiting most improvement or beneficial change. The regression models correctly predicted 72.5–80.3% of changes of the various outcome measures. Improvement or beneficial changes in stress management (OR 2.364, 95% CI 1.220–4.578), mental energy (OR 2.194, 95% CI 1.107–4.346), social support (OR 2.752, 95% CI 1.432–5.287), NPY (OR 1.934, 95% CI 1.032–3.623) and TNFα (OR 3.185, 95% CI 1.637–6.196) were significantly predicted in the intervention group compared to the reference group. Thus, the intervention group was approximately two to three times more likely to exhibit the highest improvement quartile in stress management, mental energy, social support, NPY and TNFα. These predictions remained significant even after adjustment for age, gender, annual income, education, marital status, and work related factors that might disturb the relationships, i.e. working hours per week, working atmosphere, work intensity and number of breaks during a working day and all the remaining dependent variables. Beneficial changes in sleep quality, concentration ability, DHEA-S and CgA were not significantly predicted by group in the logistic regression analysis. The frequency of logins to the web was not a significant predictor of changes in any of the dependent variables.
Table 5 Logistic regression analyses of self-ratings. The table illustrates the final regression models predicting changes (Δ) in stress management ability, sleep quality, mental energy, concentration ability and social support.
Predictors of Δ Stress management ability OR 95,0% CI for OR Lower – Upper Predictors of Δ Sleep Quality OR 95,0% CI for OR Lower – Upper
Ageb 1.041 .655 – 1.654 Ageb .718 .457 – 1,126
Genderb 1.144 .580 – 2.259 Genderb 1.391 .717 – 2.702
Marital statusb 1.255 .561 – 2.808 Marital statusb 1.175 .536 – 2.572
Educational levelb 1.261 .645 – 2.462 Educational levelb .826 .428 – 1.596
Annual incomeb 1.363 .766 – 2.424 Annual incomeb 1.532 .856 – 2.740
Groupa, b 2.364 1.220 – 4.578 Groupa, b 1.638 .854 – 3.141
Δ Concentration abilityf 2.754 1.389 – 5.461 Δ Mental energye 2.343 1.151 – 4.766
Δ CgA 2.563 1.223 – 5.256 Δ Concentration abilityf 2.259 1.125 – 4.533
Constant .002 Constant .012
Predictors of Δ Mental energy OR 95,0% CI for OR Lower – Upper Predictors of Δ Concentration ability OR 95,0% CI for OR Lower – Upper
Ageb 1.152 .713 – 1,863 Ageb 1.348 .844 – 2.152
Genderb 1.319 .659 – 2.638 Genderb 1.250 .633 – 2.469
Marital statusb .998 .429 – 2.320 Marital statusb 1.280 .569 – 2.877
Educational levelb .861 .430 – 1.722 Educational levelb 1.033 .529 – 2.018
Annual incomeb .858 .471 – 1.562 Annual incomeb .784 .434 – 1.417
Groupa.b 2.194 1.107 – 4.346 Groupa, b .900 .454 – 1.783
Δ Sleep qualityd 2.350 1.156 – 4.775 Δ Sleep qualityd 2.145 1.052 – 4.376
Δ Concentration abilityf 3.831 1.910 – 7.683 Δ Mental energye 3.638 1.791 – 7.390
Constant .005 Δ Stress managementc 2.171 1.061 – 4.442
Constant .005
Predictors of Δ Social support OR 95,0% CI for OR Lower – Upper
Ageb 1.456 .920 – 2.304
Genderb 1.428 .734 – 2.779
Marital statusb 1.822 .836 – 3.970
Educational levelb .647 .335 – 1.248
Annual incomeb .978 .557 – 1.717
Groupa.b 2.752 1.432 – 5.287
Constant .019
a Reference group: code 1, intervention group: code 2.
b Baseline values.
c Change in stress management ability. Can you manage your stress in general? (VAS, Not at all – Very well).
d Change in sleep quality. How is your quality of sleep in general? (VAS, Very poor – Very good).
e Change in mental energy. How is your energy level in general? (VAS, Empty on energy – Full of energy).
f Change in concentration ability. How is your concentration ability in general? (VAS, Very poor – Very good).
Table 6 Logistic regression analyses of biological markers. The table illustrates the final regression models predicting changes (Δ) in the biological markers DHEA-S, NPY, CgA and TNFα.
Predictors of Δ DHEA OR 95,0% CI for OR Lower – Upper Predictors of Δ NPY OR 95,0% CI for OR Lower – Upper
Ageb .863 .547 – 1.363 Ageb .723 .466 – 1.121
Genderb .527 261 – 1.066 Genderb .533 .273 – 1.042
Marital statusb 1.982 .919 – 4.275 Marital statusb 1.510 .710 – 3.215
Educational levelb 1.368 .700 – 2.672 Educational levelb 1.294 .685 – 2.442
Annual incomeb .740 .415 – 1.321 Annual incomeb 1.404 .807 – 2.442
Groupa, b 1.409 .726 – 2.735 Groupa.b 1.934 1.032 – 3.623
Constant .277 Constant .136
Predictors of Δ CgA OR 95,0% CI for OR Lower – Upper Predictors of Δ TNFα OR 95,0% CI for OR Lower – Upper
Ageb .643 .411 – 1.006 Ageb .915 .593 – 1.412
Genderb 1.449 .749 – 2.803 Genderb 1.357 .698 – 2.638
Marital statusb .767 .339 – 1.737 Marital statusb .458 .182 – 1.154
Educational levelb .732 .380 – 1.409 Educational levelb 1.508 .778 – 2.925
Annual incomeb .891 .502 – 1.581 Annual incomeb 1.228 .695 – 2.169
Groupa, b .629 .324 – 1.221 Groupa.b 3.185 1.637 – 6.196
Δ Stress managementc 2.343 1.157 – 4.744 Δ Stress managementc .404 .182 – .895
Constant .887 Constant .130
a Reference group: code 1, intervention group: code 2.
b Baseline values.
c Change in stress management ability. Can you manage your stress in general? (VAS, Not at all – Very well).
Discussion
In the present study we evaluated whether or not a web-based tool, designed for health promotion and stress management, reduces stress and increases physiological markers and psychological ratings of health, recovery and general well-being. At the end of the 6-month intervention period, the intervention group had improved significantly as compared to the reference group on self-ratings of perceived ability to manage stress, sleep quality, mental energy, concentration ability and social support. SRH increased significantly in both groups, with no differences between the groups.
Questionnaire
A striking finding is that that ratings of sleep quality improve in the intervention group vs. reference group together with related systematic findings in biological markers and other self-ratings. There is emerging evidence suggesting that sleep alterations can modulate the stress-health relationship. Acute and chronic stressors are associated with subjective and objective measures of sleep disturbances [30]. Thus, improvements in sleep quality might mediate some of the stress protective and health promoting effects found in the intervention group.
Since all participants received some kind of intervention, some beneficial changes were expected in both groups. As a matter of fact, there were several health-related statistically significant improvements for both groups over time (time effect). To mention some, ratings of SRH, eating habits, memory, physical activity, self-esteem and work joy improved as well as levels of cortisol and cholesterol that decreased. However, as the groups did not differ (time × group effect was ns) it is not certain that these effects can be attributed to the web-based tool although they might have resulted from it. To draw such a conclusion a third, passive reference group would have been needed, which was not possible for budget reasons.
The findings of the present study are in line with previous computer-based intervention studies with cognitive exercises that have shown beneficial effects on affective states, such as depression and anxiety [11-13], stress management [14] and insomnia [16]. Results of the present study are further confirmed in a prospective non-controlled study, in which a web-based intervention was found to decrease ratings of loneliness and depression, whereas perceived social support and self-esteem increased [31].
Biological markers
In the present study, DHEA-S decreased significantly in the reference group but remained unchanged in the intervention group. DHEA-S is a steroid hormone that has anabolic as well as neuroprotective effects. DHEA-S has also been found to counteract the effects of corticosteroids, such as cortisol, and to be inversely related to both stress and cortisol [32,33]. Thus, the DHEA-S decrease in the reference group may indeed be a consequence of physiological stress caused by the turbulence that occurred in connection with the study period. This indicates that the intervention program might be protective against stress and facilitate recovery, since DHEA-S remained unaltered in the same stressful period in the intervention group. Furthermore, a number of studies have suggested that DHEA-S can have beneficial effects on cognition, metabolism, wellbeing, and vascular and immune function [32-34]. Considering such prior knowledge, it is of interest that we in the present study found concurrent improvements in DHEA-S and a range of cognitive functions, such as improved concentration ability and increased mental energy in the intervention group.
NPY increased significantly in the intervention group as compared to the reference group. NPY is a hormone that has been reported to have a soothing, anxiolytic as well as antidepressive effect in the central nervous system [35]. The anxiolytic effects of NPY are probably mediated by Y1 receptors in the amygdala and involve inhibition of corticotrophin-releasing hormone (CRH). Moreover, NPY inhibits hypothalamus-pituitary-adrenal (HPA) activity and is thereby effective in reducing secretion of CRH, adrenocorticotropic hormone (ACTH) and cortisol. Finally, NPY has been found to promote and improve sleep [35]. Consequently, the increase in NPY found in the present study may partly explain the beneficial effects, including sleep improvement, found in the intervention group. Some of the findings of the previous literature however, are hard to apply to the present study since they are based upon pharmacological doses of NPY.
Chromogranin A (CgA) decreased in both groups, but significantly more in the intervention group as compared to the reference group. CgA is stored in the core of catecholamine vesicles and is often, but not always co-released with catecholamines. Secretion occurs only during marked activation of the sympathochromaffin system and only stimuli strong enough to induce catecholamine secretion are associated with CgA release. However, CgA also shows ultradian variation, which does not appear to be linked to modifications of catecholamine release [36,37]. It has been suggested that in situations of mild mental stress CgA is stable and slow to respond [38]. The decrease in CgA in the intervention group might indicate a lesser activation of the HPA-axis and a higher activation in the reference group, perhaps combined with a reduction in activity related to seasonal variation. Cortisol production, for instance, usually declines during the autumn. This might explain why CgA decreased more in the intervention group.
One of the major inflammatory cytokines, TNFα, decreased significantly in the reference group compared to the intervention group. TNFα is one of many markers of the immune system, and the production increases during immunological, inflammatory and stress responses [39]. It has been suggested that cytokines are involved in the regulation of HPA-axis activity [40]. For example, TNFα increases the secretion of CRH (corticotrophin releasing hormone) from the hypothalamus, which in turn results in an increased secretion of ACTH. In turn, ACTH stimulates the release of glucocorticoids from the adrenal cortex. During chronic stress, however, there seems to be a poor relationship between ACTH plasma concentrations and the release of glucocorticoids [39,41]. Thus, at present there is insufficient information concerning the relative effects of acute and chronic stressors on cytokine activity [42]. It has however been shown that TNFα as well as other hormones, such as cortisol and DHEA-S, exhibit seasonal variation in production. The production of these biological markers seems to be elevated in the spring time and reduced during the autumn [19,43]. This seasonal variation, i.e. reduction in the autumn, in combination with long-term stress might explain why TNFα decreases in the reference group compared to the intervention group. This possibly stress-related reduction might partly have been counteracted in the intervention group that showed improved self-rated as well as physiological stress management abilities. Consequently, since DHEA-S also remained stable in the intervention group as compared to the reference group that decreased this might be regarded as a systematic finding indicating better stress management and/or decreased stress level in the intervention group as compared to the reference group.
Methodological considerations
Everything was completely web-based from the start in the present study. It means that the stress management tool was utilized and assessed via the same medium that was used for collecting self-ratings and other relevant background data. This automated, interactive self-help approach differs from previous studies of web-based interventions. Most commonly, other studies have been more similar with face-to-face counseling, where in addition to a website an active counterpart, often a psychologist, issues assignments and evaluations via e-mail. Consequently, the results of the present study might not be completely comparable with other assessments of web-based intervention studies.
The intervention and reference group were treated in the same way concerning blood sampling, advice, web-based questionnaires, etc. The only thing that distinguished the groups was the addition of the interactive cognitive exercises and a chat for the intervention group, which indicate that the complete web-based health promotion and stress management system contributed to the beneficial effects on health, well-being and recovery.
A multitude of items (57) and physiological markers (30) were analyzed in the present study, which makes it relevant to discuss the possible problem of mass-significance. To clarify this issue, percentages of significant findings out of the total number of analyses are presented. Altogether 87 parametric analyses were conducted on relevant VAS items and physiological markers, out of which 11 (13%) significant results were obtained and 9 (10%) remained significant when non-parametric tests were used. Additionally, for several of the biological markers there may be systematic variations in levels during the sampling interval (7.15–11.30 am). For instance, serum cortisol may start decreasing. Most of this variation takes place between morning and evening however. The circadian variation during the morning hours may introduce additional random error in our results. Since there was no systematic difference in sampling hours between the two groups, no systematic bias is likely to have arisen due to this source of error.
The study period of six months might not be enough to cover long-term effects. Accordingly, the beneficial effects found in the intervention group compared to the reference group might be attenuated or continue to improve on a longer term perspective. Therefore, a post intervention follow-up was conducted six months after termination of the study, i.e. twelve months after initiation of the study. The result of this post intervention long-term follow-up will be presented in a future article. However, there are indications that some of the beneficial improvements found in the present study are attenuated after 12 months. Another aspect is that an intervention that focuses solely on individuals might have less ability to produce a lasting effect compared to interventions that also consider organizational aspects. Such multidimensional interventions could perhaps increase the possibilities for the participants to pursue beneficial effects of the individually focused intervention.
Finally, it was mentioned in the introduction that the study was conducted during a high stress period. Therefore, the general health status and occurrence of stress-related problems among the study participants might be discussed. However, apparently the participants were healthy enough to be at work and at baseline there were no participants on sick leave. Furthermore, in the case of participants going on sick-leave they could register these changes in the "profile" section on the website.
There were some weaknesses with this study, e.g. incorrect e-mail addresses to some participants complicated or made communication impossible. Furthermore, we have no exact number of potential participants in the study. This fact might bias the results considering that the sample might not be representative for all the employees. However, in general there was a great interest among the employees of the enrolling departments to participate, as for instance in one of the departments where we have the total number of potential participants, 95% enrolled. Similar participation rates are therefore estimated for the other departments. In any case, based on approximation of the total potential number of employees at each department, enrollment rate was most probably not less than 80% in the worst case scenario.
Implications and future directions
The results of the present study imply some short-term beneficial effects from a web-based tool for stress management. However, initial analyses from a long-term post intervention follow-up indicate a reversion for some of these beneficial effects. Future studies would benefit from pilot testing the web-based tools and thereto related functions to reduce risks of computer-based malfunctioning. Furthermore, logging of usage patterns may contribute with knowledge about how web-based interventions could be improved. More studies assessing psychological as well as physiological effects are needed before more firm conclusions could be drawn.
Conclusion
In summary, the current study suggest that an automated web-based system for self-assessments and real-time feedback of scorings, combined with cognitive exercises, might be beneficial to counteract unwanted stress and improve mental and physiological indicators of health at least during a six-month intervention period. Thus, such a web-based system may be an opportunity in counteracting some clinically relevant and common stress and health issues of today.
List of abbreviations
ACTH Adrenocorticotropic hormone
ANCOVA Covariated analysis of variance
ANOVA Analysis of variance
CgA Chromogranin A
CRH Corticotropin releasing hormone
DHEA Dehydroepiandosterone
HPA Hypothalamus-pituitary-adrenal
IT Information technology
NPY Neuropeptide Y
NS Non significant
REM Rapid eye movement
SD Standard deviation
S.E.M Standard error of the mean
SRH Self-rated health
TNFα Tumour necrosis factor α
VAS Visual analogue scale
Competing interests
Following the termination of this study, BA and DH have commercialized the web-based health promotion and stress management tool.
Authors' contributions
BA and DH have designed the study and compiled the questionnaire. DH was responsible for overall project management. DH conducted the statistical analyses and drafted the manuscript, with substantial and essential input from TT, UMA and BA. All authors have read and approved the final version of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study has been made possible by funding from the insurance company Alecta. We would like to thank the White-Collar Union Sif and Swedish Employers' Association Almega that have played an important part in recruiting organizations and enrolled participants and companies. We also thank Camilla Sundström from Alecta who has been active in designing the study, compiling the questionnaire and invaluable help in running some of the study logistics. We are also grateful to Jens Pettersson who developed the website used in the present study and patiently provided participants and researchers with invaluable help and technical support.
==== Refs
Lecrubier Y The burden of depression and anxiety in general medicine J Clin Psychiatry 2001 62 Suppl 8 4 9; discussion 10-1 12108821
Murray CJL Lopez AD World Bank World Health Organization Harvard School of Public Health The global burden of disease : a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020 Global burden of disease and injury series, 1 1996 Cambridge, Mass. , published by the Harvard School of Public Health on behalf of the World Health Organization and the World Bank distributed by Harvard Univ. Press 990
Kalia M Assessing the economic impact of stress--the modern day hidden epidemic Metabolism 2002 51 49 53 12040542 10.1053/meta.2002.33193
Risk A Dzenowagis J Review of internet health information quality initiatives J Med Internet Res 2001 3 E28 11772543 10.2196/jmir.3.4.e28
Fox S Fallows D Internet Health Resources: Health searches and email have become more commonplace, but there is room for improvement in searches and overall Internet access 2003 Washington DC , PEW INTERNET & AMERICAN LIFE PROJECT 42
Cavanagh K Shapiro DA Computer treatment for common mental health problems J Clin Psychol 2004 60 239 251 14981789 10.1002/jclp.10261
Marks IM Mataix-Cols D Kenwright M Cameron R Hirsch S Gega L Pragmatic evaluation of computer-aided self-help for anxiety and depression Br J Psychiatry 2003 183 57 65 12835245 10.1192/bjp.183.1.57
Tate DF Zabinski MF Computer and Internet applications for psychological treatment: update for clinicians J Clin Psychol 2004 60 209 220 14724928 10.1002/jclp.10247
Strom L Pettersson R Andersson G A controlled trial of self-help treatment of recurrent headache conducted via the Internet J Consult Clin Psychol 2000 68 722 727 10965647 10.1037//0022-006X.68.4.722
Andersson G Stromgren T Strom L Lyttkens L Randomized controlled trial of internet-based cognitive behavior therapy for distress associated with tinnitus Psychosom Med 2002 64 810 816 12271112 10.1097/01.PSY.0000031577.42041.F8
Proudfoot J Goldberg D Mann A Everitt B Marks I Gray JA Computerized, interactive, multimedia cognitive-behavioural program for anxiety and depression in general practice Psychol Med 2003 33 217 227 12622301 10.1017/S0033291702007225
Christensen H Griffiths KM Jorm AF Delivering interventions for depression by using the internet: randomised controlled trial Bmj 2004 328 265 14742346 10.1136/bmj.37945.566632.EE
Kenardy J McCafferty K Rosa V Internet-deliverered indicated prevention for anxiety disorders: A randomized controlled trial Behavioural and Cognitive Psychotherapy 2003 31 279 289 10.1017/S1352465803003047
Zetterqvist K Maanmies J Ström L Andersson G Randomized Controlled Trial of Internet-Based Stress Management Cognitive Behaviour Therapy 2003 32 Brunner-Routledge, part of the Taylor & Francis Group 151 160 16291546 10.1080/16506070302316
Napolitano MA Fotheringham M Tate D Sciamanna C Leslie E Owen N Bauman A Marcus B Evaluation of an internet-based physical activity intervention: a preliminary investigation Ann Behav Med 2003 25 92 99 12704010 10.1207/S15324796ABM2502_04
Strom L Pettersson R Andersson G Internet-based treatment for insomnia: a controlled evaluation J Consult Clin Psychol 2004 72 113 120 14756620 10.1037/0022-006X.72.1.113
Eysenbach G Diepgen TL Towards quality management of medical information on the internet: evaluation, labelling, and filtering of information Bmj 1998 317 1496 1500 9831581
Fink G Encyclopedia of stress 2000 London , Academic Press 3 vol. (760 ; 811 ; 357 )
Hasselhorn HM Theorell T Vingård E Acute musculoskeletal disease and psychophysiological parameters Stress research reports (number 289) 2002 Stockholm , WHO psychosocial centre, Laboratory for clinical stress research, National centre for suicide research and prevention of mental ill health, Department for stress research (Karolinska Institute), The National Swedish Institute for Psychosocial Medicine
Antonovsky A The structure and properties of the sense of coherence scale Soc Sci Med 1993 36 725 733 8480217 10.1016/0277-9536(93)90033-Z
Arnetz BB Techno-stress: a prospective psychophysiological study of the impact of a controlled stress-reduction program in advanced telecommunication systems design work J Occup Environ Med 1996 38 53 65 8871332 10.1097/00043764-199601000-00017
Arnetz BB Staff perception of the impact of health care transformation on quality of care International Journal for Quality in Health Care 1999 11 345 351 10501605 10.1093/intqhc/11.4.345
Bernin P Theorell T Sandberg CG Biological correlates of social support and pressure at work in managers Integr Physiol Behav Sci 2001 36 121 136 11666041
Landsbergis P Theorell T Schwartz J Greiner BA Krause N Measurement of psychosocial workplace exposure variables Occup Med 2000 15 163 188 10620790
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
Polit DF Beck CT Nursing research : principles and methods 2004 7. Philadelphia , Lippincott Williams & Wilkins 758
McCormack HM Horne DJ Sheather S Clinical applications of visual analogue scales: a critical review Psychol Med 1988 18 1007 1019 3078045
Svensson E Construction of a single global scale for multi-item assessments of the same variable Stat Med 2001 20 3831 3846 11782037 10.1002/sim.1148
Svensson E Guidelines to statistical evaluation of data from rating scales and questionnaires J Rehabil Med 2001 33 47 48 11480471 10.1080/165019701300006542
Marshall GD Rossio JL Fink G Cytokines Encyclopedia of stress 2000 1 London , Academic Press 626 633
Shaw LH Gant LM In defense of the internet: the relationship between Internet communication and depression, loneliness, self-esteem, and perceived social support Cyberpsychol Behav 2002 5 157 171 12025883 10.1089/109493102753770552
Kroboth PD Salek FS Pittenger AL Fabian TJ Frye RF DHEA and DHEA-S: a review J Clin Pharmacol 1999 39 327 348 10197292 10.1177/00912709922007903
Buvat J Androgen therapy with dehydroepiandrosterone World J Urol 2003 21 346 355 14551720 10.1007/s00345-003-0367-7
van Niekerk JK Huppert FA Herbert J Salivary cortisol and DHEA: association with measures of cognition and well-being in normal older men, and effects of three months of DHEA supplementation Psychoneuroendocrinology 2001 26 591 612 11403980 10.1016/S0306-4530(01)00014-2
Antonijevic IA Murck H Bohlhalter S Frieboes RM Holsboer F Steiger A Neuropeptide Y promotes sleep and inhibits ACTH and cortisol release in young men Neuropharmacology 2000 39 1474 1481 10818263 10.1016/S0028-3908(00)00057-5
Giampaolo B Angelica M Antonio S Chromogranin 'A' in normal subjects, essential hypertensives and adrenalectomized patients Clin Endocrinol (Oxf) 2002 57 41 50 12100068 10.1046/j.1365-2265.2002.01557.x
Parmer RJ Mahata M Gong Y Mahata SK Jiang Q O'Connor DT Xi XP Miles LA Processing of chromogranin A by plasmin provides a novel mechanism for regulating catecholamine secretion J Clin Invest 2000 106 907 915 11018079
Dimsdale JE O'Connor D Ziegler M Mills P Does chromogranin a respond to short-term mild physiologic challenge? Neuropsychopharmacology 1989 2 237 240 2789663 10.1016/0893-133X(89)90027-4
Judd AM Call GB Barney M McIlmoil CJ Balls AG Adams A Oliveira GK Possible function of IL-6 and TNF as intraadrenal factors in the regulation of adrenal steroid secretion Ann N Y Acad Sci 2000 917 628 637 11268391
Turnbull AV Rivier CL Regulation of the hypothalamic-pituitary-adrenal axis by cytokines: actions and mechanisms of action Physiol Rev 1999 79 1 71 9922367
Pignatelli D Magalhaes MM Magalhaes MC Direct effects of stress on adrenocortical function Horm Metab Res 1998 30 464 474 9694580
Anisman H Hayley S Turrin N Merali Z Cytokines as a stressor: implications for depressive illness Int J Neuropsychopharmacol 2002 5 357 373 12466035 10.1017/S1461145702003097
Myrianthefs P Karatzas S Venetsanou K Grouzi E Evagelopoulou P Boutzouka E Fildissis G Spiliotopoulou I Baltopoulos G Seasonal variation in whole blood cytokine production after LPS stimulation in normal individuals Cytokine 2003 24 286 292 14609570 10.1016/j.cyto.2003.08.005
|
16042796
|
PMC1183223
|
CC BY
|
2021-01-04 16:28:54
|
no
|
BMC Public Health. 2005 Jul 25; 5:78
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-78
|
oa_comm
|
==== Front
BMC PhysiolBMC Physiology1472-6793BioMed Central London 1472-6793-5-101596704210.1186/1472-6793-5-10Research ArticleAdenosine infusion increases plasma levels of VEGF in humans Adair Thomas H [email protected] Reid [email protected] Jian-Wei [email protected] Janelle S [email protected] Kenneth R [email protected] Michael R [email protected] Preston [email protected] Jean-Pierre [email protected] Angiogenesis Research LaboratoriesCenter for Excellence in Cardiovascular-Renal Research2 Department of Physiology & Biophysics University of Mississippi Medical Center Jackson, MS 39216, USA3 Department of Medicine University of Mississippi Medical Center Jackson, MS 39216, USA4 Institute of PhysiologyUniversity of Fribourg, 1700 Fribourg, Switzerland2005 20 6 2005 5 10 10 9 3 2005 20 6 2005 Copyright © 2005 Adair et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Many in vitro studies have shown that adenosine (Ado) can induce vascular endothelial growth factor (VEGF) mRNA and protein expression and stimulate endothelial proliferation. In the present study, we seek to determine whether Ado can increase circulating levels of VEGF protein in the intact human.
Methods
Five outpatients 49.3 ± 6.7 years of age and weighing 88.2 ± 8.5 kg were selected. They were given a 6 min intravenous infusion of Ado (0.14 mg kg-1 min-1) in conjunction with sestamibi myocardial perfusion scans. Mean blood pressure (MBP, calculated from systolic and diastolic values) and heart rate (HR) were determined before Ado infusion and every 2 min for the next 10 min. Plasma VEGF concentrations (ELISA) were determined immediately before Ado infusion and 1 h, 2 h, and 8 h after the infusion.
Results
Plasma VEGF concentration averaged 20.3 ± 2.0 pg ml-1 prior to Ado infusion, and increased to 62.7 ± 18.1 pg ml-1 at 1 h post- infusion (p < 0.01). VEGF plasma concentration returned to basal levels 2 h after infusion (23.3 ± 3.4 pg ml-1). MBP averaged 116 ± 7 mmHg and heart rate averaged 70 ± 7 prior to Ado infusion. MBP decreased by a maximum of ~22% and HR increased by a maximum of ~17% during the infusion.
Conclusion
We conclude from these preliminary findings that intravenous infusion of adenosine can increase plasma levels of VEGF in humans.
==== Body
Background
It is well-established that adenosine is an important physiological regulator for maintaining oxygen homeostasis in the heart (Berne, 1980; Ely & Berne, 1992). Adenosine formation increases when the oxygen requirements of the myocardium exceed the oxygen supply, which leads to an increase in adenosine levels in the tissues (Sparks & Bardenheuer, 1986; Duessen et al. 1988). The adenosine released under these conditions is thought to play an important role in reestablishing a balance between oxygen supply and demand (Schrader, 1990) by virtue of its potent vasodilatory (Stepp et al. 1996) and anti-adrenergic effects (Neumann et al. 1995; Dobson et al. 1986).
Mounting evidence suggests that adenosine may also have a long-term role to increase tissue oxygenation by stimulating angiogenesis (Teuscher & Weildich, 1985; Adair et al. 1990; Ethier et al. 1993; Marshall, 2001; Adair, 2005). Adenosine can stimulate endothelial proliferation in vitro (Meininger et al. 1988; Ethier & Dobson, 1997; Grant et al. 1999; Gu et al. 1999) as well as angioge nesis in a variety of in vivo models (Adair et al. 1989; Adair et al. 1991; Adair, 2005). Although the angiogenic effects of adenosine are poorly understood, it is known that adenosine, mediated by way of adenosine A2 receptors can upregulate VEGF mRNA and protein expression in many different cell types (Hashimoto et al. 1994; Takagi et al. 1996; Grant et al. 1999; Gu et al. 1999; Gu et al. 2000; Feoktistov et al. 2002; Leibovich et al. 2002; Feoktistov et al. 2003; Gardener & Olah, 2003). Numerous investigators have postulated that this adenosine-mediated expression of VEGF might have a role in mediating adenosine-induced angiogenesis ; however, it is still not clear whether adenosine can induce VEGF expression in the intact organism.
In this study, we sought to determine whether adenosine can increase circulating VEGF protein in humans. The results show that a short-term intravenous infusion of adenosine caused the plasma levels of VEGF protein to increase by about threefold at one hour after the infusion
Results
The intravenous infusion of Ado (0.14 mg kg-1 min-1 for 6 min) caused the plasma levels of VEGF protein to increase (P1–P5) one hour after the infusion (Figure 1a). However, the VEGF response to Ado varied greatly among the subjects ranging from a 21% increase (P1, 15.6 pg ml-1 to 18.9 pg ml-1) to a 329% increase (P5, 25.6 pg ml-1 to 110.1 pg ml-1), as shown in Figure 1b. Plasma VEGF protein levels averaged 20.3 ± 2.0 pg ml-1 before the infusion, and increased to an average of 62.7 ± 18.1 pg ml-1 (p < 0.05) one hour after the infusion. VEGF protein levels returned to control values by 2 h after the Ado infusion (23.3 ± 3.4 pg ml-1) and remained at control levels at 8 h post- infusion (24.3 ± 7.0 pg ml-1), as shown in Figure 1a.
The infusion of Ado caused MBP to decrease in most of the subjects (Figure 2), ranging from a 1 mmHg decrease in P1 (control MBP, 94 mmHg) to a 59 mmHg decrease in P5 (control MBP, 128 mmHg). This maximum decrease in MBP occurred after 2 minutes of Ado infusion in two subjects (P2, P3) and after 6 min of infusion in three subjects (P1, P4, P5). Note on Figure 2 that MBP did not decrease below 69 mmHg in any of the subjects.
The decrease in MBP was associated with decreases in both systolic (SBP) and diastolic (DBP) pressures (Table 1). Note also on Table 1 that adenosine increased heart rate (HR) significantly during the period of infusion. This increase in HR is typically seen in human subjects infused with adenosine and is thought to be a reflex response to the fall in blood pressure.
Discussion
The results have shown that adenosine can increase plasma levels of VEGF protein in humans. The intravenous infusion of adenosine caused a three-fold increase in plasma VEGF protein levels one hour after the infusion (Figure 1a). Previous studies have shown that exogenous administration of adenosine or adenosine agonists (Hashimoto et al. 1994; Fischer et al. 1995; Takagi et al. 1996; Pueyo et al. 1998; Grant et al. 1999; Gu et al. 1999; Grant et al. 2001; Leibovich et al. 2002; Gardener & Olah, 2003; Feoktistov et al. 2003) or upregulation of endogenous adenosine (Gu et al. 2000) can enhance VEGF protein and/or mRNA expression in a variety of cultured cells (for review, see Adair, 2005). However, the present study is the first, to our knowledge, that demonstrates that the infusion of adenosine can increase circulating levels of VEGF in vivo.
The known cardiovascular effects of adenosine are mediated by the A1 and A2 (subtypes A2A and A2B) receptors (for review see Shryock & Belardinelli, 1997; Ralevic & Burnstock 1998; Rongen et al. 1999). Responses to activation of the A1 receptor include slowing of the heart and slowing of impulse conduction through the atrioventricular node, reduction of atrial contractility, inhibition of beta-adrenergic effects, and vasodilatation (Bryan & Marshall, 1999). Activation of A2 receptors causes (a) vascular smooth muscle relaxation (Belardinelli et al. 1998; Ngai et al. 2001; Hinschen et al. 2003), which can lead to decreases in blood pressure, and (b) induction of VEGF expression from a variety of cell types (Hashimoto et al. 1994; Pueyo et al. 1998; Grant et al. 1999; Gu et al. 1999; Leibovich et al. 2002; Gardener & Olah, 2003; Feoktistov et al. 2003). Recent development of selective agonists and antagonists for A2 receptor subtypes suggests that vasodilatation is mediated by A2A (Shryock & Belardinelli, 1997; Belardinelli et al. 1998; Rongen et al. 1999) and perhaps A2B receptors (Ralevic & Burnstock 1998; Ngai et al. 2001; Hinschen et al. 2003). The A2B receptor subtype appears to mediate the induction of VEGF protein and mRNA in various human cell types. Grant et al. (1999) have shown that activation of the adenosine A2B receptor (but not the A2A receptor) can increase VEGF mRNA and protein expression in cultured human retinal endothelial cells. Feoktistov et al. (2003) established that adenosine acting by way of the A2B receptor can stimulate VEGF expression in human mast cells. These latter investigators (Feoktistov et al. 2002) also showed that adenosine can induce VEGF expression in human microvascular endothelial cells, where A2B receptors are abundant, but not in human umbilical vein endothelial cells, which preferentially express A2A receptors. Therefore, the available data suggest that adenosine induced vasodilatation and VEGF expression could be mediated to some extent by similar receptors.
It is not clear from the present study whether the infusion of adenosine actually induced the transcription and translation of VEGF ; however, it is known that both VEGF mRNA and protein can be induced within a one hour period of time. Breen et al. (1996) found a fourfold increase in VEGF mRNA at the termination of a one hour bout of treadmill exercise in rats. Gu and Adair (1997) showed that exposing dog myocardial vascular smooth muscle cells to a hypoxic environment caused a twofold increase in VEGF protein concentration in the media after only one hour of exposure to the hypoxic environment. Therefore, it is conceivable that adenosine could induce VEGF transcription and translation resulting in elevated plasma levels of VEGF at one hour post infusion. However, additional studies will be required to determine whether the increase in circulating VEGF caused by adenosine did in fact result from adenosine-induced production of VEGF.
A possibility that must be considered is that the transient decrease in blood pressure caused by adenosine leads to transient hypoxia in the tissues, and that the hypoxia, in turn, leads to the induction of VEGF expression. Hypoxia is a well-known stimulus for the induction of VEGF expression both in vitro and in vivo (for review see Ferrara & Davis-Smyth, 1997; Ferrara, 2001). However, it seems unlikely that this indirect induction of VEGF by adenosine could explain entirely the results of the present study for the following reasons: (a) adenosine can induce VEGF mRNA and protein expression in vitro where blood pressure is not an issue, as discussed previously; (b) adenosine increased the plasma levels of VEGF in two patients in whom the blood pressure had decreased by less than 15 mmHg, a decrease that should not cause significant perfusion problems or hypoxia in the tissues; (c) mean blood pressure did not fall below 69 mmHg in any of the subjects; and (d) none of the subjects experienced symptoms suggestive of cerebral hypoxia during the infusion of adenosine. All subjects were supine during the infusion of adenosine to minimize potential perfusion problems. Furthermore, adenosine induced vasodilatation would be expected to improve the distribution of oxygen within a tissue, also arguing against tissue hypoxia. Therefore, it is unlikely that the adenosine induced decrease in blood pressure contributed significantly to the rise in VEGF.
Conclusion
In conclusion, this study provides preliminary data showing that an intravenous infusion of adenosine can increase plasma levels of VEGF in humans. Additional studies will be required to determine whether the increase in plasma VEGF caused by adenosine results from the transcription and translation of VEGF.
Methods
Subject demographics
The study protocol was approved by the University of Mississippi Medical Center Internal Review Board. The subjects underwent routine sestamibi myocardial perfusion scans because of suspected coronary artery disease. They consisted of one man and four women; mean age was 49.3 ± 6.7 years and mean weight was 88.2 ± 8.5 kg. None of the subjects had taken caffeine prior to the test, had a history of asthma or myocardial infarction, or were being treated with blood pressure medications. Their mean blood pressure (MBP, auscultatory sphygmomanometry) prior to testing was 116 ± 7 mmHg (systolic, 159 ± 12 mmHg; diastolic, 94 ± 5 mmHg). MBP was calculated as the diastolic pressure plus one third of the pulse pressure. Mean heart rate was 70 ± 3 beats per minute; one patient (P4) was paced at 66 beats per minute. Also, one individual (P3) had a history of multiple myeloma and was being evaluated for kidney transplantation.
Experimental protocol
Adenosine (Adenoscan®, Medco Research, Inc.) was infused intravenously at 0.14 mg kg-1 min-1 for six minutes (i.e., total dose was 0.84 mg kg-1). MBP was measured before the adenosine infusion and then every two minutes for the next ten minutes. Technetium-99 (sestamibi) was administered three minutes into the adenosine infusion and a nuclear scan was performed. The infusion of Ado was well tolerated by all subjects. There were no significant ECG abnormalities during the test and no myocardial perfusion abnormalities were found. None of the subjects were diagnosed with cardiovascular disease. Three mL of blood were collected in EDTA coated tubes before the adenosine infusion and at 1, 2, and 8 h after the adenosine infusion. The blood samples were centrifuged at 700 g for 5 minutes at 4°C, and plasma was stored at -80°C.
Measurement of VEGF protein
Plasma levels of VEGF were measured in duplicate by sandwich ELISA (R&D Systems, Minneapolis). In these assays, the test sample is sandwiched between a monoclonal antibody against human recombinant VEGF. A second polyclonal antibody against VEGF is conjugated to horseradish peroxidase and then added to the mixture. Color develops by addition of hydrogen peroxide and chromagen tetramethylbenzidine, and the intensity is measured at 450 nm. VEGF plasma protein levels (pg ml-1) were calculated using an equation generated from the regression analysis of standard VEGF concentrations.
Statistical analyses
The results are expressed as means ± standard errors of the mean (S.E.M.). Differences between the groups (e.g., plasma VEGF at various times) were compared using analysis of variance for repeated measurements and Dunnett's multiple comparison test (InStat, GraphPad Software, Inc.).
Abbreviations
Ado, adenosine
VEGF, vascular endothelial growth factor
HR, heart rate
MBP, mean blood pressure
SBP, systolic blood pressure
DBP, diastolic blood pressure
Authors' contributions
Thomas H. Adair, Jean-Pierre Montani, and Reid Cotten designed the study. Reid Cotten, Kenneth R. Bennett, and Michael R. McMullan performed the myocardial perfusion scans and collected the blood samples. Thomas H. Adair, Jian-Wei Gu, Janelle S. Pryor, and Preston B. McDonnell performed the various analyses. All authors contributed to writing the manuscript.
Acknowledgements
The work was supported by a grant from the National Institutes of Health HL51971.
Figures and Tables
Figure 1 Effect of adenosine on plasma VEGF protein concentration. Adenosine (0.14 mg kg-1 min-1) was infused intravenously for 6 minutes at 0 h. Panel A: Adenosine caused plasma VEGF protein concentration to increase by an average of ~3-fold at 1 h post-infusion. Plasma VEGF protein levels at 0, 1, 2, and 8 h were 20.3 ± 2.0, 62.7 ± 18.1, 23.3 ± 3.4, and 24.3 ± 7.0 pg ml-1, respectively. Panel B: Adenosine caused plasma levels of VEGF protein to increase in all 5 subjects one h after the infusion; however, the response was relatively small in two subjects. Bars are means ± S.E.M. (n = 5 for each bar). Asterisks indicate p < 0.01.
Figure 2 Mean blood pressure (MBP) was calculated from systolic and diastolicpressures for the 5 patients (P1–P5). MBP was measured before Ado infusion (time 0), and then every 2 min for the next 10 min, except in P3 in which a 10 min MBP was not obtained. Ado infusion caused MBP to decrease at some point during the infusion in all subjects, but the pressure did not decrease below 69 mmHg in any of the subjects.
Table 1 Acute cardiovascular effects of adenosine infusion.
0 min 2 min 4 min 6 min 8 min 10 min
HR 70 ± 3 77 ± 4 81 ± 5* 82 ± 5* 78 ± 4 77 ± 5
SBP 159 ± 12 153 ± 9 144 ± 11 135 ± 9* 155 ± 12 145 ± 7
DBP 94 ± 5 80 ± 5 72 ± 8* 68 ± 9** 81 ± 7 72 ± 6
MBP 116 ± 7 104 ± 6 96 ± 8* 90 ± 9** 106 ± 8 96 ± 3
Heart rate (HR, beats per minute), systolic blood pressure (SBP, mmHg), diastolic blood pressure (DBP, mmHg) and mean blood pressure (MBP, mmHg) were measured before Ado infusion (0 min) and then every 2 min for the next 10 min. MBP was calculated from SBP and DBP. Values are means ± S.E.M. (n = 5 at all times, except at 10 min where n = 4). * indicate p < 0.05. ** indicate p < 0.01.
==== Refs
Adair TH Growth regulation of the vascular system: an emerging role for adenosine Am J Physiol Regal Integr Comp Physiol 2005 289 R283 R296
Adair TH Gay WJ Hester RL Montani JP Imai S, Nakazawa M Does adenosine have a regulatory role in the growth of blood vessels? Role of adenosine and adenine nucleotides in the biological system 1991 Elsevier Science Publishers BV, Amsterdam 443 455
Adair TH Gay WJ Montani JP Growth regulation of the vascular system: evidence for a metabolic hypothesis Am J Physiol 1990 259 R393 404 1697737
Adair TH Montani JP Strick DM Guyton AC Vascular development in chick embryos: a possible role for adenosine Am J Physiol 1989 256 H240 246 2463773
Belardinelli L Shryock JC Snowdy S Zhang Y Monopoli A Lozza G Ongini E Olsson RA Dennis DM The A2a adenosine receptor mediates coronary vasodilation J Pharmacol Exp Ther 1998 284 1066 1073 9495868
Berne RM The role of adenosine in the regulation of coronary blood flow Circ Res 1980 47 807 813 6254686
Breen EC Johnson EC Wagner H Tseng HM Sung LA Wagner PD Angiogenic growth factor mRNA responses in muscle to a single bout of exercise J Appl Physiol 1996 81 355 361 8828685
Bryan PT Marshall JM Adenosine receptor subtypes and vasodilatation in rat skeletal muscle during systemic hypoxia: a role for A1 receptors J Physiol 1999 514.1 151 162 9831723 10.1111/j.1469-7793.1999.151af.x
Deussen A Borst M Kroll K Schrader J Formation of S-adenosylhomocysteine in the heart. II: a sensitive index for regional myocardial underperfusion Circ Res 1988 63 250 261 3383379
Dobson JG JrOrdway RW Fenton RA Endogenous adenosine inhibits catecholamine contractile responses in normoxic hearts Am J Physiol 1986 251 H455 462 3740298
Ely SW Berne RM Protective effects of adenosine in myocardial ischemia Circulation 1992 85 893 904 1537125
Ethier MF Chander V Dobson JG Jr Adenosine stimulates proliferation human endothelial cells in culture Am J Physiol 1993 265 H131 138 8342624
Ethier MF Dobson JG Adenosine stimulation of DNA synthesis in human endothelial cells Am J Physiol 1997 272 H1470 1479 9087626
Feoktistov I Goldstein AE Ryzhov S Zeng D Belardinelli L Voyno-Yasenetskaya T Biaggioni I Differential expression of adenosine receptors in human endothelial cells: role of A2B receptors in angiogenic factor regulation Circ Res 2002 90 531 538 11909816 10.1161/01.RES.0000012203.21416.14
Feoktistov I Ryzhov S Goldstein AE Biaggioni I Mast cell-mediated stimulation of angiogenesis: cooperative interaction between A2B and A3 adenosine receptors Circ Res 2003 92 485 492 12600879 10.1161/01.RES.0000061572.10929.2D
Ferrara N Role of vascular endothelial growth factor in regulation of physiological angiogenesis Am J Physiol 2001 280 C1358 1366
Ferrara N Davis-Smyth T The biology of vascular endothelial growth factor EndocrRev 1997 18 4 25
Fischer S Sharma HS Karliczek GF Schaper W Expression of vascular permeability factor / vascular endothelial growth factor in pig cerebral microvascular endothelial cells and its upregulation by adenosine Brain Res Mol Brain Res 1995 28 141 148 7707868 10.1016/0169-328X(94)00193-I
Gardner AM Olah ME Distinct protein kinase C isoforms mediate regulation of vascular endothelial growth factor expression by A2A adenosine receptor activation and phorbol esters in pheochromocytoma PC12 cells J Biol Chem 2003 278 15421 15428 12590138 10.1074/jbc.M208366200
Grant MB Davis MI Caballero S Feoktistov I Biaggioni I Belardinelli L Proliferation, migration, and ERK activation in human retinal endothelial cells through A2B adenosine receptor stimulation Invest Ophthalmol Vis Sci 2001 42 2068 2073 11481274
Grant MB Tarnuzzer RW Caballero S Ozeck MJ Davis MI Spoerri PE Feoktistov I Biaggioni I Shryock JC Belardinelli L Adenosine receptor activation induces vascular endothelial growth factor in human retinal endothelial cells Circ Res 1999 85 699 706 10521243
Gu JW Adair TH Hypoxia-induced expression of VEGF is reversible in myocardial vascular smooth muscle cells Am J Physiol 1997 273 H628 633 9277477
Gu JW Brady AL Anand V Moore MC Kelly WC Adair TH Adenosine upregulates VEGF Expression in cultured myocardial vascular smooth muscle cells Am J Physiol 1999 277 H595 602 10444484
Gu JW Ito BR Sartin A Frascogna N Moore M Adair TH Inhibition of adenosine kinase induces expression of VEGF mRNA and protein in myocardial myoblasts Am J Physiol 2000 279 H2116 2123
Hashimoto E Kage K Ogita T Nakaoka T Matsuoka R Kira Y Adenosine as an endogenous mediator of hypoxia for induction of vascular endothelial growth factor mRNA in U-937 cells Biochem Biophys Res Commun 1994 204 318 324 7945378 10.1006/bbrc.1994.2462
Hinschen AK Rose'Meyer RB Headrick JP Adenosine receptor subtypes mediating coronary vasodilation in rat hearts J Cardiovasc Pharmacol 2003 41 73 80 12500024 10.1097/00005344-200301000-00010
Leibovich SJ Chen JF Pinhal-Enfield G Belem PC Elson G Rosania A Ramanathan M Montesinos C Jacobson M Schwarzschild MA Fink JS Cronstein B Synergistic up-regulation of vascular endothelial growth factor expression in murine macrophages by adenosine A2A receptor agonists and endotoxin Am J Pathol 2002 160 2231 2244 12057925
Marshall JM Roles of adenosine and nitric oxide in skeletal muscle in acute and chronic hypoxia Adv Exp Med Biol 2001 502 349 363 11950149
Meininger CJ Granger HJ Mechanisms leading to adenosine-stimulated proliferation of microvascular endothelial cells Am J Physiol 1990 258 H198 206 2154131
Meininger CJ Schelling ME Granger HJ Adenosine and hypoxia stimulate proliferation and migration of endothelial cells Am J Physiol 1988 255 H554 562 3414822
Neumann J Gupta RC Bodor GS Bartel S Kruse EG Pask HT Schmitz W Scholz H Watanabe AM Interaction of beta-adrenoceptor and adenosine receptor agonists on phosphorylation: identification of target proteins in mammalian ventricles J Mol Cell Cardiol 1995 27 1655 1667 8523428 10.1016/S0022-2828(95)90689-4
Ngai AC Coyne EF Meno JR West GA Winn HR Receptor subtypes mediating adenosine-induced dilation of cerebral arterioles Am J Physiol 2001 280 H2329 2335
Pueyo ME Chen Y D'Angelo G Michel JB Regulation of vascular endothelial growth factor expression by cAMP in rat aortic smooth muscle cells Exp Cell Res 1998 238 354 358 9473343 10.1006/excr.1997.3864
Ralevic V Burnstock G Receptors for purines and pyrimidines Pharmacol Rev 1998 50 413 492 9755289
Richardson RS Wagner H Mudaliar SR Henry R Noyszewski EA Wagner PD Human VEGF gene expression in skeletal muscle: effect of acute normoxic and hypoxic exercise Am J Physiol 1999 277 H2247 2252 10600843
Rongen GA Brooks SC Pollard MJ Ando S Dajani HR Notarius CF Floras JS Effect of adenosine on heart rate variability in humans Clin Sci (Lond) 1999 96 597 604 10334965
Shryock JC Belardinelli L Adenosine and adenosine receptors in the cardiovascular system: biochemistry, physiology, and pharmacology Am J Cardiol 1997 79 2 10 9223356 10.1016/S0002-9149(97)00256-7
Schrader J Adenosine: a homeostatic metabolite in cardiac energy metabolism Circulation 1990 81 389 391 2297843
Sparks HV JrBardenheuer H Regulation of adenosine formation by the heart Circ Res 1986 58 193 201 3004778
Stepp DW Van Bibber R Kroll K Feigl EO Quantitative relation between interstitial adenosine concentration and coronary blood flow Circ Res 1996 79 601 610 8781493
Takagi H King GL Robinson GS Ferrara N Aiello LP Adenosine mediates hypoxic induction of vascular endothelial growth factor in retinal pericytes and endothelial cells Invest Ophthalmol Vis Sci 1996 37 2165 2176 8843903
Teuscher E Weidlich W Adenosine nucleotides, adenosine and adenine as angiogenesis factors Biomed Biochim Acta 1985 44 493 495 4004845
|
15967042
|
PMC1183224
|
CC BY
|
2021-01-04 16:03:51
|
no
|
BMC Physiol. 2005 Jun 20; 5:10
|
utf-8
|
BMC Physiol
| 2,005 |
10.1186/1472-6793-5-10
|
oa_comm
|
==== Front
BMC PhysiolBMC Physiology1472-6793BioMed Central London 1472-6793-5-91595816410.1186/1472-6793-5-9Research ArticlePI3K and PKC contribute to membrane depolarization mediated by α2-adrenoceptors in the canine isolated mesenteric vein Yamboliev Ilia A [email protected] Violeta N [email protected] Department of Pharmacology and Center of Biomedical Research Excellence, University of Nevada School of Medicine, Reno, Nevada 89557, USA2 Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, Nevada 89557, USA2005 15 6 2005 5 9 9 9 10 2004 15 6 2005 Copyright © 2005 Yamboliev and Mutafova-Yambolieva; 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
Norepinephrine (NE), a classic neurotransmitter in the sympathetic nervous system, induces vasoconstriction of canine isolated mesenteric vein that is accompanied by a sustained membrane depolarization. The mechanisms underlying the NE-elicited membrane depolarization remain undefined. In the present study we hypothesized that phosphatidylinositol 3-kinase (PI3K) and protein kinase C (PKC) are involved in the electrical field stimulation (EFS)-induced slow membrane depolarization (SMD) in canine isolated mesenteric vein. EFS (0.1–2 Hz, 0.1 ms, 15V, 10 s)-induced changes in the membrane potential were recorded with a conventional intracellular microelectrode technique and evaluated in the absence and presence of inhibitors of neuronal activity, α-adrenoceptors, membrane ion channels, PI3K, inositol 1,4,5-triphosphate (InsP3) receptors, and PKC. Activation of PI3Kγ and PKCζ in response to exogenous NE and clonidine in the absence and presence of receptor and kinase inhibitors were also determined.
Results
Contractile responses to NE and clonidine (0.05 – 10 μM) were significantly diminished in the presence of yohimbine (0.1 μM). Exogenous NE (0.1 μM) and clonidine (1 μM) elicited SMD. The resting membrane potential of canine mesenteric vein smooth muscle cells was -68.8 ± 0.8 mV. EFS elicited a biphasic depolarization comprised of excitatory junction potentials and SMD that are purinergic and adrenergic in nature, respectively. The magnitude of the SMD in response to EFS at 0.5 Hz was 9.4 ± 0.7 mV. This response was reduced by 65–98% by the fast Na+ channel inhibitor tetrodotoxin (1 μM), by the inhibitor of N-type Ca2+ channels ω-conotoxin GVIA (5 nM), the non-selective α-adrenoceptor blocker phentolamine (1 μM), the selective α2-adrenoceptor blocker yohimbine (0.1 μM), the ion channel inhibitors niflumic acid (NFA, 100 μM), 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB, 30 μM), 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS, 200 μM), and Gd3+ (30 μM), and the PI3K inhibitors wortmannin (100 nM) and LY-294002 (10 μM). The SMD remained unchanged in the presence of the L-type Ca2+ channel blocker nicardipine (1 μM) and the InsP3 receptor blockers 2-aminoethoxydiphenylborate (2APB, 50 μM) and xestospongin C (3 μM). The inhibitor of PKC chelerythrine (1 μM), but not calphostin C (10 μM), diminished the SMD. Exogenous NE and clonidine (1 μM each) activated both PI3Kγ and PKCζ, and the activation of these kinases was abolished by preincubation of tissue with the α2-adrenoceptor blocker yohimbine.
Conclusion
Neuronally-released NE stimulates smooth muscle α2-adrenoceptors and activates PI3K and atypical PKC in the canine mesenteric vein. Events downstream of PKC lead to SMD and vasoconstriction. This represents a novel pathway for NE-induced membrane depolarization in a vascular smooth muscle preparation.
==== Body
Background
Norepinephrine (NE), a classic neurotransmitter in the sympathetic nervous system, is released from adrenergic varicosities of stimulated postganglionic nerve terminals, activates postjunctional α-adrenoceptors and gives rise to a slow membrane depolarization (SMD) and contraction [1,29]. The NE-induced SMD represents an important mechanism of excitation-contraction coupling in blood vessels however the signaling pathways underlying the NE-elicited SMD in vascular smooth muscle remain undefined.
One well-documented pathway downstream of activated G-protein coupled receptors (GPCRs) includes dissociation of Gαβγ trimers and production of Gα monomer and Gβγ dimer, and involvement of the latter proteins in signal transduction events downstream of α-adrenoceptors. For example, Gα mediates activation of phospholipase C (PLC), hydrolysis of phosphatidylinositol 4,5-bisphosphate (PI4,5P2), and generation of second messengers including inositol 1,4,5-triphosphate (InsP3) and diacylglycerol, DAG [20]. These second messengers then mediate signal transduction events leading to activation of ion channels. InsP3 has the capacity to release cytosolic Ca2+ from intracellular stores, which then activates Ca2+-activated Cl- channels (ClCCa) and membrane depolarization, required for opening of voltage-operated calcium channels (VOCC) and Ca2+ influx. DAG, on the other hand, activates non-selective cation channels (NSCC) in rabbit portal vein [17]. In addition, it becomes increasingly clear that Gβγ dimers can initiate intracellular signal transduction events as well. Phosphatidylinositol 3-kinase-γ (PI3Kγ), a member of class IB PI3Ks, was identified as a major effector of Gβγ in various cell and tissue preparations [13,18]. Lipid products of the PI3Ks, phosphatidylinositol 3,4-bisphosphate (PI3,4P2) and phosphatidylinositol 3,4,5-trisphosphate (PI3,4,5P3), function as second messengers and can directly affect the activity of the membrane ion channels CFTR [12] and voltage-gated potassium channels [19]. Alternatively, PI3,4P2 and PI3,4,5P3 can modulate membrane ion channels via activation of PKC isozymes [6,25]. For example, Gβγ, PI3Kγ, and atypical PKC were shown to link activation of G-protein coupled M2-muscarinic receptors to metabotropic Ca2+ and voltage-independent Cl- channels in Xenopus oocytes [31]. It was also demonstrated that PI3K mediates activation of L-type Ca2+ channels upon stimulation of M2-muscarinic receptors in rabbit portal vein myocytes [3] and α2-adrenoceptor induced vasoconstriction in porcine palmar lateral vein [27]. These studies imply that activation of GPCRs could activate membrane ion channels and SMD via PI3K-dependent mechanisms. To our knowledge, however, coupling of α-adrenoceptors to PI3Kγ and membrane depolarization in vascular smooth muscles has not yet been reported.
We used canine isolated mesenteric vein to test the hypothesis that EFS-induced SMD is mediated by PI3K and PKC. Our results demonstrate both nerve stimulation and exogenous NE-mediated activation of α2-adrenoceptors, PI3K and PKC, and suggest a role for these kinases for the activation of membrane ion channels (e.g., ClCCa and/or NSCC) and development of SMD.
Results
α2-Adrenoceptors mediate vasoconstriction and membrane depolarization in canine isolated mesenteric vein
Cumulative application of exogenous NE and clonidine (0.05 μM-10 μM) resulted in concentration dependent contractile responses (Fig. 1A and 1B). In the presence of the selective α2-adrenoceptor antagonist yohimbine (0.1 μM) the contractile responses to 0.05–1 μM NE were virtually abolished, whereas the responses to 5 and 10 μM NE were significantly reduced (Fig. 1A). Yohimbine abolished the mechanical responses to all clonidine concentrations (Fig. 1B). These results indicate that α2-adrenoceptors are particularly important for NE-mediated vasoconstriction in this blood vessel. Moreover, exogenous application of either NE (0.1 μM, Fig. 1C) or clonidine (1 μM, Fig. 1D) elicited SMD, suggesting that α2-adrenoreceptors may be involved in the membrane potential changes in response to NE that is released upon EFS.
EFS gives rise to a biphasic membrane depolarization
The resting membrane potential of isolated canine mesenteric vein segments averaged -68.8 ± 0.8 mV (n = 78). EFS (0.1–2 Hz, 0.1 ms, 15 V, 10 s) gave rise to a characteristic biphasic depolarization of the cell membrane (Fig. 2A–E), composed of a fast excitation junction potential (EJP) of presumed purinergic nature, and a SMD of adrenergic origin. The amplitude of the SMD increased with the frequency of EFS (Fig. 2A–F). To better understand the processes that govern membrane depolarization, we exposed blood vessels to the fast Na+ channel blocker tetrodotoxin (TTX, 1 μM) or to the neuronal N-type Ca2+ channel blocker ω-conotoxin GVIA (ω-CtX GVIA, 5 nM). Both drugs virtually abolished the electrical responses to EFS (Fig. 3A, B), suggesting that the EFS-elicited SMD requires activation of postganglionic nerve terminals and release of a neurotransmitter substance. Phentolamine (1 μM) significantly reduced the SMD in response to 0.5 Hz EFS to 2.1 ± 0.5 mV (n = 3, P < 0.05, not shown), consistent with the possibility that NE-mediated activation of α-adrenoceptors is the primary cause for the SMD in blood vessels [14,29]. Furthermore, the selective α2-adrenoceptor antagonist yohimbine (0.1 μM, Fig. 3C), but not the selective α1-adrenoceptor antagonist prazosin (1 μM; 8.9 ± 0.9, n = 3, not shown), significantly reduced the SMD in response to 0.5 Hz EFS from 9.4 ± 0.7 (n = 14) to 0.8 ± 0.5 mV (n = 3, P < 0.05), indicating that the EFS-evoked SMD is mediated primarily by α2-adrenoceptors. Nicardipine (1 μM) abolished the contraction elicited by 70 mM KCl (not shown), indicating that L-type Ca2+ channels are present in this blood vessel. However, the EFS-induced SMD remained unchanged in tissues preincubated with nicardipine (Fig. 3D), suggesting that opening of L-type Ca2+ channels is not required for the membrane depolarization.
The EFS-induced slow membrane depolarization is sensitive to NFA, NPPB, DIDS and Gd3+
To test possible contribution of ion channels in the EFS-induced SMD, we incubated veins with the commonly used non-selective cation channel (NSCC) and Cl- channel inhibitors NFA (100 μM), NPPB (30 μM), and DIDS (200 μM) [16,26]. These inhibitors reduced the 0.5 Hz EFS-elicited SMD by 73% (Fig. 4A and 4E), 92% (Fig. 4B and 4E) and 67% (Fig. 4C and 4E, n = 3–14), respectively. DIDS also reduced the excitation junction potentials (Fig. 4C), which may be due to blockade of P2X receptors, as suggested previously [2]. The NSCC inhibitor Gd3+ reduced the SMD by 60% (Fig. 4D and 4E, n = 3). Together, our results indicate that the SMD in canine mesenteric vein depends on activation of membrane ion channels (e.g. ClCCa and/or NSCC).
InsP3 receptors are not required for the EFS-elicited slow depolarization
To determine whether EFS-induced SMD is mediated by InsP3 receptors, veins were incubated with 2APB (50 μM), a selective inhibitor of InsP3 receptors and hence of GPCR-PLC-InsP3-mediated Ca2+ release from sarcoplasmic reticulum. Surprisingly, SMD to 0.5 Hz EFS remained unaffected by 2APB (Fig. 5A and 5C). To test the possibility that the lack of effect of 2APB was due to poor membrane permeability, we carried out mechanical experiments with mesenteric veins, incubated with 2APB (50 μM) prior to incubation with increasing concentrations of the α1-adrenoceptor agonist methoxamine (0.1–100 μM). 2APB produced a significant rightward shift of the concentration-response curve of methoxamine (not shown) and a decrease of the EC50 from 6.2 ± 0.3 μM (n = 10) to 5.1 ± 0.2 μM (n = 6). These results suggest that activation of InsP3 receptors may be unnecessary for the SMD. This possibility was also confirmed by the inability of another inhibitor of InsP3 receptors, xestospongin C (3 μM), to reduce the EFS-evoked SMD (Fig. 5B and 5C, n = 3). Therefore, the EFS-mediated SMD may not require release of Ca2+ from InsP3 receptor-operated stores.
PI3K blockers reduce the EFS-elicited slow depolarization
To test whether PI3Ks are involved in the EFS-induced SMD, we incubated tissue strips with the PI3K inhibitors wortmannin (100 nM) and LY-294002 (10 μM). Both inhibitors significantly reduced the SMD elicited by 0.5 Hz EFS from 9.4 ± 0.7 mV in controls to 1.0 ± 1.1 and 2.5 ± 1.6 mV, respectively (Fig. 6), indicating that activation of PI3Ks is required for the EFS-induced activation of ion channels and SMD. Incubation of veins with exogenous NE (100 nM) induced 14.1 ± 3.0 mV SMD, which was reduced to 3.0 ± 2.0 mV (n = 3) in the presence of wortmannin (100 nM, not shown), suggesting that both EFS and exogenous NE-induced SMD are mediated by PI3K.
Exogenous NE and clonidine activate PI3Ks
To determine whether NE directly activates PI3Ks, we measured phosphorylation of Akt in control vein segments and in tissues incubated with exogenous NE (1 μM, 3 min). Incubation with NE increased phosphorylation of Akt about 2-fold the basal (Fig. 7A). When smooth muscle strips were preincubated with wortmannin (100 nM, 1 h) or LY-294002 (10 μM, 1 h), NE failed to increase phosphorylation of Akt. These results indicate that NE-elicited phosphorylation of Akt in mesenteric vein requires activation of PI3Ks.
As shown in Fig. 1 and Fig 3, incubation of veins with yohimbine significantly reduced both the NE-induced and clonidine-induced contractions as well as the SMD elicited by EFS. These observations raise the possibility that the EFS and exogenous NE-induced activation of PI3Ks may be mediated by α2-adrenoceptors. To test this possibility, we assayed phosphorylation of Akt (Fig. 7C) in tissues incubated with clonidine (1 μM, 3 min) in the absence or presence of yohimbine (0.1 μM, 1 h). Clonidine alone increased phosphorylation of Akt to 2.3-fold the basal phosphorylation and this increase was abolished by inhibition of α2-adrenoceptors with yohimbine and by inhibition of PI3Ks with LY-294002. These data indicate that α2-adrenoceptors are coupled to PI3Ks and their downstream target Akt.
Previous experiments have shown that PI3Kγ mediates signaling downstream of GPCRs [18,33]. To test whether NE activates PI3Kγ, we incubated mesenteric vein with exogenous NE, then immunoprecipitated PI3Kγ and used the immunoprecipitated kinase to phosphorylate phosphatidylinositol (PI) in vitro. As shown in Fig. 7B, stimulation with NE increased the amount of PI3P (reaction product), suggesting that NE activated PI3Kγ. NE-mediated phosphorylation of PI was reversed in strips preincubated with LY-294002, indicating that it was a PI3K-dependent event. To test whether α2-adrenoceptors are linked to activation of PI3Kγ, we analyzed the activation of PI3Kγ in veins incubated with clonidine (1 μM) in the absence or presence of yohimbine (0.1 μM). As shown in Fig. 7D, clonidine activated PI3Kγ to 1.8-fold the basal, while yohimbine eliminated this activation. Thus, the activation of PI3Ks, and more specifically PI3Kγ, by exogenous NE and clonidine suggests that α2-adrenoceptors are coupled to PI3Kγ via activation of G-proteins.
Stimulation of atypical PKC(s) contributes to the EFS-elicited SMD
PI3Kγ may be linked to membrane ion channels (i.e., ClCCa and/or NSCC) via isozymes of the multifunctional PKC family. To test this possibility, we incubated mesenteric veins with the broad-spectrum PKC blocker chelerythrine (1 μM, 1 hour). Although chelerythrine reduced the response to 0.5 Hz EFS from 9.4 ± 0.7 mV to 2.1 ± 1.5 mV, the residual SMD remained unchanged in veins, incubated simultaneously with LY-294002 and chelerythrine prior to EFS (2.6 ± 1.3). These results not only implicate the existence of a PKC-dependent component in the EFS-induced SMD (Fig. 8A and 8C), but also suggest that PI3Kγ and PKC may signal to ion channels in a linear fashion. To narrow down the PKCs that function downstream of PI3Kγ, we used calphostin C, an inhibitor of classical and novel PKCs [32]. One-hour incubation of mesenteric veins with calphostin C (10 μM) failed to inhibit EFS-stimulated SMD (Fig. 8B and 8C). However, calphostin C significantly reduced the contractile responses to PMA (100 nM, 30 min incubation), which is mediated by classical and novel PKCs (not shown). Together, these results suggest that activation of atypical, rather than classical or novel PKC isozymes is involved in the SMD.
Exogenous NE and clonidine activate PKCζ
Since exogenous NE and clonidine activated PI3Kγ, we tested whether these agents also activate PKCζ, a specific substrate of PI3Kγ [5]. Vein segments incubated with NE (1 μM, 3 min) and clonidine (1 μM, 3 min) caused activation of for a synthetic PKCζ peptide substrate in vitro to 2.11 and 1.89-fold the basal kinase activity of non-stimulated controls (Fig. 9A and 9B). The NE-induced activation of PKCζ was significantly reduced in tissues incubated with the PI3K inhibitor LY-294002 (10 μM) and was eliminated by a specific myristoylated peptide inhibitor of PKCζ (PKCζI, 50 μM) in vitro (Fig. 9A). Similarly, the clonidine-induced activation of PKCζ was eliminated in tissues preincubated with yohimbine and LY-294002 (Fig. 9B). These results indicate that stimulation of canine mesenteric veins with NE and clonidine is associated with activation of PI3Ks and a subsequent activation of PKCζ.
Ion channel blockers and PI3K blockers do not inhibit release of NE
Reduction of the SMD could be the result of suppressed NE release from sympathetic nerve terminals in veins, incubated with protein kinase and ion channel blockers. To test this possibility, we assayed the EFS-evoked release of NE in superfusates collected during EFS in control veins and in tissues preincubated with each of the aforementioned blockers. The average EFS -evoked overflow of NE in tissue controls (16 Hz, 0.1 ms) was 122 ± 27 fmol/mg (n = 17). In preincubated tissues, the overflow of NE changed to (fmol/mg tissue) 236 ± 40 in the presence of NFA (n = 6), 480 ± 122 by NPPB (n = 6), 142 ± 2 by DIDS (n = 6), 144 ± 5 by wortmannin (n = 5), and 214 ± 38 by LY-294002 (n = 7). Therefore, none of these agents reduced the EFS-evoked overflow of NE, indicating that their effects on the EFS-induced SMD are not due to inhibited NE release but to activation of postjunctional (i.e., smooth muscle) mechanisms.
Discussion
The NE-induced membrane depolarization is an essential requirement for opening VOCC, Ca2+ entry and smooth muscle contraction, and hence it represents an important mechanism of autonomic neurovascular control. Previous studies [4] as well as the present work indicate that α2-adrenoceptors are involved in the SMD and the vasoconstriction of mesenteric vein. However, the downstream mechanisms that couple α2-adronoceptors to SMD remain undefined. For example, NE-induced activation of ClCCa plays a key role in the associated vasoconstriction and presumed membrane depolarization in various vascular networks [11,16], but NSCC may also participate in the NE-induced vasoconstriction [17]. In the present study, we have expanded upon these previous works by directly measuring membrane potential in response to EFS of intact canine isolated mesenteric veins. We found that the SMD in response to EFS is frequency-dependent, and is sensitive to the fast Na+ channel blocker TTX, to the inhibitor of neuronal N-type Ca2+ channels ω-conotoxin GVIA [24], and to the selective antagonist of α2-adrenoceptors yohimbine. Consistent with previous works, therefore, our results indicate that EFS gives rise to a SMD, which is mediated by smooth muscle α2-adrenoceptors and hence is primarily mediated by NE, released upon action potential. Furthermore, the SMD of the vascular smooth muscle cell membrane appears to be mediated by channels sensitive to NFA, NPPB, and DIDS. Although these inhibitors target transporters with presumed high preference for Cl- [8], they may also affect other ion channels, such as the inhibition of NSCC by DIDS [7]. Because incubation of mesenteric veins with the NSCC blocker Gd3+ partially reduced the EFS-evoked SMD, our results suggest that activation of NSCC may indeed contribute to the EFS-elicited SMD in this blood vessel. None of the channel blockers reduced the release of NE thus demonstrating that the decrease of SMD is mediated by postjunctional smooth muscle mechanisms. The EFS-induced SMD remained unaffected by the InsP3 receptor inhibitors 2APB and xestospongin C, ruling out a substantial role of InsP3 receptor-mediated Ca2+ release in this response. Therefore, smooth muscle cell chloride channels and/or NCSS appear to be the primary activities that mediate the SMD in the canine mesenteric vein.
Furthermore, we identified a novel signal transduction mechanism governing the slow membrane depolarization, which involves enzymes of the PKC family. First, we used two PKC inhibitors with distinct PKC isozymes selectivity. Chelerythrine, for example, acts on the conserved catalytic domain of PKC as a competitive inhibitor with respect to the phosphate acceptor and a non-competitive inhibitor with respect to ATP [32]. Therefore, chelerythrine can inhibit PKCs of all classes and effects mediated by them, including the slow depolarization in mesenteric veins. In contrast, calphostin C interacts with the regulatory domain of PKC by competing for the binding site of diacylglycerol and phorbol esters, but not Ca2+ and phospholipids. Thus, calphostin C is more specific inhibitor for classical and novel, and less specific for atypical PKC [9,32]. In the present study, chelerythrine reduced, while calphostin C had no effect on the SMD. The lack of effect of calphostin C indicates that the EFS-induced membrane depolarization is most likely mediated by atypical PKC(s). We provided experimental support of this possibility by showing a distinct activation of the atypical PKCζ in mesenteric veins incubated with exogenous NE and clonidine. Although further studies are needed to identify the precise PKC isoform(s) contributing to the α2-adrenoceptor mediated SMD, the present study clearly suggests a PKC isozyme-specific regulation of the EFS- (and hence NE-) induced slow membrane depolarization in canine mesenteric vein.
The observation that atypical PKCs are involved in the SMC allowed us to speculate about the identity of some specific signaling molecules that function upstream of PKCζ, and could mediate the SMD as well. PI3Ks, and particularly PI3Kγ, was an obvious candidate for several reasons. Firstly, PI3Kγ is essential for activation of various ion transporters including L-type Ca2+ channels [30] and metabotropic nonselective cation channels [31]. Secondly, PI3Kγ activates the phosphoinositide-dependent protein kinase 1 (PDK1), and although the latter can modulate classical (α and βI) and novel (δ and ε) PKC isoform, activation of atypical (ζ/λ/τ) PKCs, and particularly PKCζ, is a highly specific event downstream of PI3Kγ-PDK1 [28]. We experimentally supported the hypothesized role of PI3Ks by showing that the PI3K inhibitors wortmannin and LY-294002 prevent the NE and clonidine-dependent activation of PKCζ and the SMD. These data are consistent with a role of PI3K in the regulation of voltage-independent Cl- channels, as well. PI3Kγ in particular is activated by Gβγ dimers, which are released upon dissociation of Gαβγ complexes following activation of GPCRs [18]. Our experimental data demonstrate activation of PI3Kγ in veins incubated with the G-protein coupled α2-adrenoceptor agonists NE and clonidine. Furthermore, since the NE and clonidine-mediated activation of PI3Kγ, PKCζ and of the SMD was prevented by yohimbine, our results demonstrate that activation of α2-adrenoceptors is required for activation of PI3Kγ and PKCζ, and possibly for activation of membrane Cl- and/or NCSS channels and SMD in canine mesenteric vein. While the molecular identity of the membrane ion channels involved in these effects is presently elusive, our electrophysiological and biochemical data provide support to the possibility that activation of α2-adrenoceptors, PI3K and atypical PKC (possibly PKCζ) is essential for the regulation of the autonomic nervous system and vascular smooth muscle tone.
Conclusion
In this study we provide functional and biochemical evidence that NE, released from postganglionic nerve terminals, activates postjunctional α2-adrenoceptors, PI3Ks and atypical PKCs in canine isolated mesenteric vein. Our results further suggest that specific isoforms of the PI3K and PKC families, i.e. PI3Kγ and PKCζ respectively, may participate in the signal transduction pathway that couples α2-adrenoceptors to membrane ion channels. This signal transduction pathway mediates slow membrane depolarization and vasoconstriction of canine mesenteric vein.
Methods
Tissue preparation
Seventy-four mongrel dogs of either sex (averaging 15 kg) were obtained from vendors licensed by the United States Department of Agriculture. The use of dogs was approved by the University of Nevada's Animal Care and Use Committee. The animals were sacrificed with an overdose of pentobarbitone sodium (100 mg/kg intravenously), as recommended by the Panel on Euthanasia of the American Veterinary Medical Association. Experiments were conducted with second and third order branches of the inferior mesenteric vein (0.7–1.2 mm in diameter), dissected and denuded of endothelium as outlined previously [21].
Intracellular recording of membrane potential
Ring segments (7–10 mm long; 700–800 μM external diameter) were pinned out on the sylgard bottom of a 2 ml recording chamber perfused with Krebs solution (3 ml/min; 37°C; aerated with 95% O2/5% CO2) with the following composition (mM): 150 NaCl, 4.6 KCl, 1.2 MgCl2, 2.5 CaCl2, 24.8 NaHCO3, 1.2 KH2PO4 and 5.6 dextrose. Intracellular measurements were made through the adventitia of the vessel with fiber-containing borosilicate electrodes filled with 3 M KCl (70–100 MΩ resistance), as described previously [22]. EFS at supramaximal voltage with trains of square-wave pulses (0.1 ms pulse width) was applied at 0.1–2 Hz for 10 s by means of two parallel platinum electrodes on both sides of the vessel connected to a Grass S48 stimulator. Once well-defined and reproducible slow depolarizations were obtained, various drugs were applied to the superfusion solution according to the experimental protocol. The maximum depolarization was evaluated.
The effects of NFA (100 μM), NPPB (30 μM), DIDS (200 μM), 2 APB (50 μM), and xestospongin C (3 μM) were examined by recording membrane potential in response to EFS before and 30-min after the tissue was superfused with the drug. To assess the role of PI3K and PKC or the NSCC, tissues were superfused with wortmannin (100 nM), LY 294002 (10 μM), chelerythrine (1 μM), calphostin C (10 μM) or Gd3+ (30 μM) for 60 min prior to the EFS. To produce appropriate time controls, EFS-mediated SMD were recorded periodically for 2–4 h in tissues not incubated with inhibitors. In some experiments tissues were perfused with NE (100 nM) for 30 s to induce SMD and changes in this response were monitored in the absence and presence of wortmannin (100 nM). In some experiments tissues were perfused with clonidine (1 μM) and changes in the membrane potential were monitored.
Transmitter release experiments and HPLC assay of NE
Segments of endothelium-denuded mesenteric veins (52.5 ± 3.2 mg wet weight, n = 30) were placed in 200-μl BRANDEL superfusion chambers as previously described [23]. After 45 min equilibration, the tissues were subjected to a 30-seconds "conditioning" stimulation with a train of square wave pulses of 0.3 ms duration and a frequency of 4 Hz. Thirty minutes later the blood vessels were subjected to EFS for 60 s with a train of suprathreshold pulses of 0.1 ms duration at 16 Hz. Samples of the superfusion solution were collected before the electrical stimulation (resting overflow) and during the electrical stimulation (electrically evoked overflow) in ice-cold test tubes. Samples were analyzed for NE content by high performance liquid chromatography (HPLC) technique in conjunction with electrochemical detection [23].
After the equilibration period the tissues were superfused either with wortmannin (100 nM), or LY-294002 (10 μM), or chelerythrine (1 μM), or calphostin C (10 μM) for 60 min prior to EFS. In some experiments tissues were superfused with NFA (100 μM), NPPB (30 μM), or DIDS (200 μM) for 30 min prior to EFS. Only one drug was tested in each tissue.
Mechanical responses
Ring preparations (3 mm long) were mounted in 3 ml organ baths by inserting two stainless steel triangle mounts into the lumen, and force displacements were further investigated as described previously [21]. The baths contained Krebs solution, which routinely contained indomethacin (1 μM) and Nω-nitro-L-arginine (l-NNA, 100 μM) to block potential residual effects of the endothelium and to eliminate possible time dependent effects due to activation of inducible nitric oxide synthase (iNOS) and/or cyclooxygenase. Thus, the contractile responses to KCl and nerve stimulation were reproducible over many hours when these blockers were included in the bathing solution. A resting force of 0.5 g was applied to the vein segments. This was found to stretch vessels to near the optimum length for tension development. After equilibrating the tissues [21], concentration-response relationships were obtained by cumulative addition of increasing concentrations of the α1-adrenoceptor agonist methoxamine in the absence and presence of 2APB (50 μM, 30 min pretreatment). Contractile responses elicited by methoxamine, NE and clonidine were expressed as a percentage of the contractile response produced with 70 mM KCl. The 70 mM KCl solution had the following composition (mM): 84.6 NaCl, 70 KCl, 1.2 MgCl2, 2.5 CaCl2, 24.8 NaHCO3, 1.2 KH2PO4 and 5.6 dextrose. Control experiments were carried out to determine the consistency of the contractile response to repeated applications of KCl (70 mM) over the duration of an average treatment protocol. The relative potency of methoxamine was determined by the concentration producing half-maximal effect. In some experiments, the effect of KCl (70 mM) was tested in the presence of nicardipine (1 μM). In other experiments, tissues were contracted with phorbol 12-myristate 13-acetate (PMA, 100 nM) in the absence or presence of chelerythrine (1 μM) or calphostin C (10 μM). In some experiments contractile responses to either exogenous NE or clonidine (both 0.05–10 μM) were monitored in the absence or presence of yohimbine (0.1 μM for 30 min).
Preparation of tissue homogenates and cytosolic fractions
Total protein extracts were prepared by glass-glass homogenization of mesenteric veins, pulverized under liquid nitrogen, with a buffer composed of (mM): 10 Tris-HCl (pH 7.4), 5 EDTA, 5 EGTA, 10 sodium pyrophosphate, 10 NaF, 1 sodium orthovanadate, 0.1 AEBSF and 0.001 leupeptin. Insoluble material was pelleted by centrifugation at 3,000 g (S3 supernatants) for 5 min at 4°C. S3 supernatants were transferred into clean tubes and centrifuged at 120,000 g for 60 min at 4°C to obtain S120 supernatants, which were used for assay of PI3K and PKC activity.
PI3K activation assay
Activation of PI3K was assayed by the phosphorylation of the downstream protein kinase Akt. Equal amounts of total supernatant protein (30 μg) were resolved by SDS-PAGE, transferred onto nitrocellulose membranes, and total and phospho-Ser473-Akt were assayed by immunoblotting, using a rabbit polyclonal or mouse monoclonal antibodies, respectively. Blots were scanned to obtain images and the immunoreactive bands were analyzed by densitometry, using the Quantity One software (BioRad). Changes in protein phosphorylation were calculated by normalizing the band density phospho-Akt to total Akt, and then were presented relative to the untreated group controls.
PI3Kγ activation assay
PI3Kγ was immunoprecipitated from S120 supernatants with a mouse monoclonal antibody, immobilized on Protein A/G agarose plus beads (Santa Cruz Biotechnology, Inc.). Kinase activity was assayed by phosphorylation of phosphatidylinositol (PI) in vitro, as described previously [34].
PKCζ activation assay
PKCζ activation was assayed by in vitro phosphorylation of a synthetic peptide substrate (PKCζPS), ERMRPRKRQGSVRRRV [15]. S120 fractions from control and treated tissues were used as enzyme sources. The phosphorylation reactions were stopped by cooling on ice, and 10 μl reactions were spotted on P81 filter strips. Excess radioactivity was removed from the filters by 3 washes with 75 mM ortho-phosphoric acid and a final wash with ethanol. The filters were air-dried prior to radiography and spots were quantified by densitometry, using a BioRad Model 525 Molecular Imager. In order to verify PKCζ-specific phosphorylation of the substrate, in some phosphorylation reactions we added a specific myristoylated PKCζ peptide inhibitor (PKCζI, 50 μM, Calbiochem).
Drugs
NFA, NPPB, wortmannin, LY 294002, DIDS, nicardipine, norepinephrine hydrochloride, tetrodotoxin, phentolamine, xestospongin C, gadolinium chloride, PMA (Sigma-Aldrich), 2 APB (Tocris), PKCζPS and PKCζI (Calbiochem), chelerythrine, calphostin C (Biomol), general and phospho-specific anti-Akt antibodies (Cell Signaling Technology, Inc.), phosphatidylinositol (PI, Avanti). Gadolinium solutions were prepared fresh each day, essentially as described elsewhere [10]. Wortmannin, LY-294002, chelerythrine, calphostin C, PMA were dissolved in dimethylsulfoxide (DMSO) and diluted in double distilled water or Krebs solution. The final concentration of DMSO was less than 0.1 % DMSO.
Statistical analysis
Data are presented as means ± SEM. Means were compared by analysis of variance (one-way ANOVA) (GraphPadPrism v. 3, GraphPad Software, Inc.). A probability value of less than 0.05 was considered significant. In the intracellular recording and mechanical response experiments, n refers to the number of rings, and hence dogs, included in each experimental group.
Authors' contributions
Author IAY participated in the PI3K activation, PI3Kγ activation, and PKCζ activation assays. Author VM-Y participated in the membrane potential, NE release and mechanical activity experiments and designed and coordinated the study. All authors participated in drafting the manuscript.
All authors read and approved the final manuscript.
Acknowledgements
This work was supported by NCRR grant P 20RR 15581 to IAY and US Public Service Grant HL60031 to VM-Y.
Figures and Tables
Figure 1 The α2-adrenoceptor blocker yohimbine reduces the contractile responses of canine isolated mesenteric vein to norepinephrine (NE) and clonidine. Contractile responses of control and yohimbine-incubated tissues (0.1 μM, 30 min) to exogenous NE (A) and clonidine (0.05–10 μM each). Contractile responses were expressed as percentage of the contractile response to 70 mM KCl. Data are presented as mean ± SEM, n = 4–8. *, P < 0.05. Representative traces of intracellular recordings of membrane potential changes elicited by NE (0.1 μM, 30 s, panel C) or clonidine (1 μM, 30 s, panel D). Scale bars apply to traces C and D.
Figure 2 Typical biphasic membrane depolarization of canine isolated mesenteric vein. Representative intracellular recordings depicting EFS-induced (0.1 ms, 13V, 10 s, 0.1–2 Hz) biphasic membrane depolarizations, consisting of excitatory junction potentials and slow membrane depolarization (A-E). The amplitude of the slow membrane depolarization increases with the increase of stimulation frequencies (F). Mean ± SEM, n = 4–12. Scale bars apply to traces A-E.
Figure 3 Tetrodotoxin (TTX), ω-conotoxin GVIA (Ctx GVIA) and yohimbine, but not nicardipine, inhibit the EFS-induces slow membrane depolarization of canine isolated mesenteric vein. Intracellular recordings of membrane potential in response to EFS (0.1 ms, 13 V, 10 s) in canine isolated mesenteric veins preincubated for 30 min prior to and throughout the experiment with 1 μM TTX (A), 5 nM ω-conotoxin GVIA (Ctx GVIA, B), 0.1 μM yohimbine (C) and 1 μM nicardipine (D). Scale bars apply to all traces.
Figure 4 The EFS-induced slow membrane depolarization of canine isolated mesenteric vein is reduced by various ion channel inhibitors. Intracellular recordings of membrane potential in response to EFS (0.1 ms, 13 V, 10 s, 0.5 Hz) in canine isolated mesenteric veins preincubated for 60 min prior to and throughout the experiment with 100 μM NFA (A), 30 μM NPPB (B), 200 μM DIDS (C) and 30 μM Gd3+ (D). The experimental data are presented as mean ± SEM, n = 3–4 (E). *, P < 0.05 versus control. Scale bars apply to traces A-D.
Figure 5 2APB and xestospongin C have no effect on the EFS-induced slow membrane depolarization of canine isolated mesenteric vein. Intracellular recordings of membrane potential in response to EFS (0.1 ms, 13 V, 10 s, 0.5 Hz) in canine isolated mesenteric veins preincubated for 60 min prior to and throughout the experiment with 50 μM 2APB (A) or 30 μM xestospongin C (B). The experimental data are presented as mean ± SEM, n = 3–4 (C). Scale bars apply to traces A and B.
Figure 6 The PI3K inhibitors wortmannin and LY-294002 inhibit the EFS-induced slow membrane depolarization in canine isolated mesenteric vein. Intracellular recordings of membrane potential in response to EFS (0.1 ms, 13 V, 10 s) in canine isolated mesenteric veins preincubated for 60 min prior to and throughout the EFS with 100 nM wortmannin (A) or 10 μM LY-294002 (B). The experimental data are presented as mean ± SEM, n = 3–4 (C). *, P < 0.05 vs control. Scale bars apply to traces A and B.
Figure 7 Exogenous NE and clonidine activate PI3Ks in canine isolated mesenteric vein. Immunoblots of Akt and phospho-Akt (P-Akt) in untreated control tissues (CT), and in tissues treated with NE (1 μM, 3 min) alone or after 60-min preincubation with 100 nM wortmannin, Wm, or 10 μM LY-294002, LY (A). Immune blots of P-Akt in veins incubated with clonidine (1 μM, 3 min) in the absence or after 60-min preincubation with 0.1 μM yohimbine, Yoh (C). The phospho-Akt (P-Akt) band densities were normalized to the respective Akt band densities, and the P-Akt/Akt ratios of treated tissues were presented relative to the ratios of non-treated tissue controls (bar graphs in A and C, n = 3). Thin-layer chromatography (TLC) separation of PI3P, obtained upon in vitro phosphorylation of PI by immunoprecipitated PI3Kγ. PI3Kγ activation was assessed in untreated (control) tissue (CT) and in tissues incubated for 3 min with 1 μM NE (B) or 1 μM clonidine (D), without or after 1-h preincubation with LY-294002 (LY, 10 μM) or yohimbine (Yoh, 0.1 μM). Phosphorylated PIP3 was quantified by radiography and densitometry, and PI3Kγ activation in treated tissues was expressed relative to non-treated tissue controls (bar graph, n = 3). *, P < 0.05 versus controls, #, P < 0.05 versus veins treated with NE or clonidine.
Figure 8 The PKC inhibitor chelerythrine, but not calphostin C, reduces the EFS-evoked slow membrane depolarization in canine isolated mesenteric vein. Intracellular recordings of membrane potential in response to EFS (0.1 ms, 13V, 10 s, 0.5 Hz) in control mesenteric veins (Control) or in tissues preincubated for 60 min with 1 μM chelerythrine (A) or 10 μM calphostin C (Calph C, B). The experimental data are presented as mean ± SEM, n = 3. *, P < 0.05 versus control. Scale bars apply to traces A and B.
Figure 9 NE and clonidine activate PKCζ in canine mesenteric vein. Phosphorylation of a specific PKCζ peptide substrate (PKCζPS) in vitro by total protein extracts obtained from untreated controls (CT) and tissues incubates for 3 min with 1 μM NE (A) or 1 μM clonidine (B), without or after 60-min preincubation with LY-294002 (10 μM), myristoylated PKCζ peptide inhibitor (PKCζI, 50 μM) or yohimbine (Yoh, 0.1 μM). Phosphorylation of the PKCζ PS (i.e. density of the phospho-PS dots) was quantitated by radiography and densitometry (insets). Phosphorylation of PKCζSP in treated tissues is presented relative to non-treated tissue controls (bar graph, n = 3). *, P < 0.05 versus control; #, P < 0.05 versus veins treated with NE or clonidine.
==== Refs
Bulbring E Tomita T Catecholamine action on smooth muscle Pharmacol Rev 1987 39 49 96 3033708
Bultmann R Starke K Blockade by 4,4'-diisothiocyanatostilbene-2,2'-disulphonate (DIDS) of P2X-purinoceptors in rat vas deferens Br J Pharmacol 1994 112 690 694 8075889
Callaghan B Koh SD Keef KD Muscarinic M2 Receptor Stimulation of Cav1.2b Requires Phosphatidylinositol 3-Kinase, Protein Kinase C, and c-Src Circ Res 2004 94 626 633 14739158 10.1161/01.RES.0000118248.17466.B7
Cheung DW An electrophysiological study of alpha-adrenoceptor mediated excitation-contraction coupling in the smooth muscle cells of the rat saphenous vein Br J Pharmacol 1985 84 265 271 2858233
Chou MM Hou W Johnson J Graham LK Lee MH Chen CS Newton AC Schaffhausen BS Toker A Regulation of protein kinase C zeta by PI 3-kinase and PDK-1 Curr Biol 1998 8 1069 1077 9768361 10.1016/S0960-9822(98)70444-0
Derman MP Toker A Hartwig JH Spokes K Falck JR Chen CS Cantley LC Cantley LG The lipid products of phosphoinositide 3-kinase increase cell motility through protein kinase C J Biol Chem 1997 272 6465 6470 9045671 10.1074/jbc.272.10.6465
Diakov A Koch JP Ducoudret O Muller-Berger S Fromter E The disulfonic stilbene DIDS and the marine poison maitotoxin activate the same two types of endogenous cation conductance in the cell membrane of Xenopus laevis oocytes Pflugers Arch 2001 442 700 708 11512026 10.1007/s004240100593
Fuller CM Ji HL Tousson A Elble RC Pauli BU Benos DJ Ca(2+)-activated Cl(-) channels: a newly emerging anion transport family Pflugers Arch 2001 443 S107 S110 11845314 10.1007/s004240100655
Gordge PC Ryves WJ Inhibitors of protein kinase C Cellular Signalling 1994 6 871 882 7718407 10.1016/0898-6568(94)90020-5
Hajduczok G Chapleau MW Ferlic RJ Mao HZ Abboud FM Gadolinium inhibits mechanoelectrical transduction in rabbit carotid baroreceptors. Implication of stretch-activated channels J Clin Invest 1994 94 2392 2396 7527431
He Y Tabrizchi R Effects of niflumic acid on alpha1-adrenoceptor-induced vasoconstriction in mesenteric artery in vitro and in vivo in two-kidney one-clip hypertensive rats Eur J Pharmacol 1997 328 191 199 9218701 10.1016/S0014-2999(97)83045-2
Himmel B Nagel G Protein kinase-independent activation of CFTR by phosphatidylinositol phosphates EMBO Rep 2004 5 85 90 14710196 10.1038/sj.embor.7400034
Hirsch E Katanaev VL Garlanda C Azzolino O Pirola L Silengo L Sozzani S Mantovani A Altruda F Wymann MP Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation Science 2000 287 1049 1053 10669418 10.1126/science.287.5455.1049
Hottenstein OD Kreulen DL Comparison of the frequency dependence of venous and arterial responses to sympathetic nerve stimulation in guinea-pigs J Physiol (Lond) 1987 384 153 167 2821237
Kazanietz MG Areces LB Bahador A Mischak H Goodnight J Mushinski JF Blumberg PM Characterization of ligand and substrate specificity for the calcium-dependent and calcium-independent protein kinase C isozymes Mol Pharmacol 1993 44 298 307 8355667
Lamb FS Barna TJ Chloride ion currents contribute functionally to norepinephrine-induced vascular contraction Am J Physiol 1998 275 H151 H160 9688908
Large WA Receptor-operated Ca2(+)-permeable nonselective cation channels in vascular smooth muscle: a physiologic perspective J Cardiovasc Electrophysiol 2002 13 493 501 12030534 10.1046/j.1540-8167.2002.00493.x
Leopoldt D Hanck T Exner T Maier U Wetzker R Nurnberg B Gbetagamma stimulates phosphoinositide 3-kinase-gamma by direct interaction with two domains of the catalytic p110 subunit J Biol Chem 1998 273 7024 7029 9507010 10.1074/jbc.273.12.7024
Loussouarn G Park KH Bellocq C Baro I Charpentier F Escande D Phosphatidylinositol-4,5-bisphosphate, PIP2, controls KCNQ1/KCNE1 voltage-gated potassium channels: a functional homology between voltage-gated and inward rectifier K+ channels EMBO J 2003 22 5412 5421 14532114 10.1093/emboj/cdg526
Mironneau J Macrez-Lepretre N Modulation of Ca2+ channels by alpha 1A- and alpha 2A-adrenoceptors in vascular myocytes: involvement of different transduction pathways Cell Signal 1995 7 471 479 8562308 10.1016/0898-6568(95)00014-G
Mutafova-Yambolieva VN Carolan BM Harden TK Keef KD Multiple P2Y receptors mediate contraction in guinea pig mesenteric vein Gen Pharmacol 2000 34 127 136 10974420 10.1016/S0306-3623(00)00054-9
Mutafova-Yambolieva VN Keef KD Frequency dependent alpha(2)-adrenoceptor mediated modulation of excitatory junction potentials in guinea-pig mesenteric artery Eur J Pharmacol 2001 411 123 127 11137866 10.1016/S0014-2999(00)00905-5
Mutafova-Yambolieva VN Smyth L Bobalova J Involvement of cyclic AMP-mediated pathway in neural release of noradrenaline in canine isolated mesenteric artery and vein Cardiovasc Res 2003 57 217 224 12504831 10.1016/S0008-6363(02)00648-X
Nielsen KJ Schroeder T Lewis R Structure-activity relationships of omega-conotoxins at N-type voltage-sensitive calcium channels J Mol Recognit 2000 13 55 70 10822250 10.1002/(SICI)1099-1352(200003/04)13:2<55::AID-JMR488>3.0.CO;2-O
Rameh LE Cantley LC The role of phosphoinositide 3-kinase lipid products in cell function J Biol Chem 1999 274 8347 8350 10085060 10.1074/jbc.274.13.8347
Remillard CV Lupien MA Crepeau V Leblanc N Role of Ca2+- and swelling-activated Cl- channels in alpha1-adrenoceptor-mediated tone in pressurized rabbit mesenteric arterioles Cardiovasc Res 2000 46 557 568 10912466 10.1016/S0008-6363(00)00021-3
Roberts RE Alpha 2 adrenoceptor-mediated vasoconstriction in porcine palmar lateral vein: role of phosphatidylinositol 3-kinase and EGF receptor transactivation Br J Pharmacol 2003 138 107 116 12522079 10.1038/sj.bjp.0705014
Ron D Kazanietz MG New insights into the regulation of protein kinase C and novel phorbol ester receptors FASEB J 1999 13 1658 1676 10506570
Suzuki H Adrenergic transmission in the dog mesenteric vein and its modulation by alpha-adrenoceptor antagonists Br J Pharmacol 1984 81 479 489 6141830
Viard P Exner T Maier U Mironneau J Nurnberg B Macrez N Gbetagamma dimers stimulate vascular L-type Ca2+ channels via phosphoinositide 3-kinase FASEB J 1999 13 685 694 10094929
Wang YX Dhulipala PD Li L Benovic JL Kotlikoff MI Coupling of M2 muscarinic receptors to membrane ion channels via phosphoinositide 3-kinase gamma and atypical protein kinase C J Biol Chem 1999 274 13859 13864 10318793 10.1074/jbc.274.20.13859
Way KJ Chou E King GL Identification of PKC-isoform-specific biological actions using pharmacological approaches Trends Pharmacol Sci 2000 21 181 187 10785652 10.1016/S0165-6147(00)01468-1
Wymann MP Pirola L Structure and function of phosphoinositide 3-kinases Biochim Biophys Acta 1998 1436 127 150 9838078
Yamboliev IA Wiesmann KM Singer CA Hedges JC Gerthoffer WT Phosphatidylinositol 3-kinases regulate ERK and p38 MAP kinases in canine colonic smooth muscle Am J Physiol Cell Physiol 2000 279 C352 C360 10913001
|
15958164
|
PMC1183225
|
CC BY
|
2021-01-04 16:03:51
|
no
|
BMC Physiol. 2005 Jun 15; 5:9
|
utf-8
|
BMC Physiol
| 2,005 |
10.1186/1472-6793-5-9
|
oa_comm
|
==== Front
BMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 1471-244X-5-281596703910.1186/1471-244X-5-28Research ArticleThe level of recognition of physical symptoms in patients with a major depression episode in the outpatient psychiatric practice in Puerto Rico: An observational study Tamayo Jorge M [email protected]án Karis [email protected] Juan J [email protected] María [email protected] Lilly Research Laboratories, Eli Lilly and Company, San Juan, Puerto Rico2 Department of Psychiatry, Medical Sciences Campus, University of Puerto Rico, San Juan, Puerto Rico3 Department of Biostatistics & Epidemiology, Medical Sciences Campus, University of Puerto Rico, San Juan, Puerto Rico4 Department of Psychiatry, San Juan Bautista University, Caguas, Puerto Rico5 Department of Internal Medicine, Medical Sciences Campus, University of Puerto Rico, San Juan Puerto Rico2005 20 6 2005 5 28 28 27 1 2005 20 6 2005 Copyright © 2005 Tamayo et al; licensee BioMed Central Ltd.2005Tamayo 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 was designed to evaluate the psychiatrists' level of recognition of somatic symptoms associated to a major depressive episode (MDE) (DSM-IV-TR criteria) and the impact of those somatic symptoms on the treatment effectiveness.
Methods
This non-interventional study was conducted in 25 medical offices in Puerto Rico from February to December 2003. It had 2 visits separated by 8 weeks. The level of recognition was determined by: the correlation between the physician clinical evaluation and their patients' self-evaluations through different validated instruments using kappa statistics. Chi-square test was used to evaluate the impact of somatic symptoms on treatment antidepressants' effectiveness.
Results
All the 145 recruited patients reported the presence of at least one somatic symptom associated with their current MDE. In the two visits covered by the study, a fair agreement between the psychiatrists' and the patients' reports was noted for headache, abdominal pain and upper limb pains (0.4003 ≤ κ ≥ 0.6594). For other painful symptoms and painless somatic symptoms, the Kappa values obtained were non-significant. Slight but significant reductions in depression and painful symptoms severity were observed after 8 weeks of treatment. A proportional relationship between the pain and depression severity was observed (p < 0.0001).
Conclusion
The study results show that somatic symptoms: are very common in depressed Puerto Rican patients; are significant under-reported by psychiatrists; and have a significant impact on the antidepressant effectiveness.
==== Body
Background
Although recent epidemiological studies have not demonstrated any differences in the prevalence rate of major depressive disorders in the Latino and Caucasian populations living in the United States [1], the Hispanic Health and Nutrition Examination Survey (Hispanic HANES), when compared to the Epidemiological Catchment Area (ECA) Study, showed that Puerto Ricans living in the United States have the highest prevalence of depressive symptomatology, followed by Caucasians, Mexican-Americans and Cuban-Americans [2-4]. Unfortunately, little is known about the prevalence of major depression in adults living in the island of Puerto Rico. Data suggest lower lifetime prevalence rates (4.3% – 4.6%) [5,6] in comparison to the US mainland (16.2%) [1].
Traditionally, diagnostic classification systems have focused on the emotional symptoms of a major depressive episode (MDE), such as depressed mood, markedly diminished interest or pleasure in almost all activities, and feelings of worthlessness, among others [7]. Nevertheless, the importance of somatic symptoms, also known as physical symptoms, in depressed patients has been well established. It is estimated that 69% to 92% of the patients have somatic symptoms [8-11]. Studies conducted in primary care settings [12-18] have shown significant association between major depression and painless somatic symptoms such as: vague and exaggerated multiple somatic complaints (usually more than three), fatigue, weakness, non-specific and painless musculoskeletal problems, sensations of heaviness or lightness in at least one part of the body, gastrointestinal dysfunction, shortness of breath, palpitations, dizziness, double vision, changes in sleep patterns and appetite, and polyuria. Similarly, significant association between major depression and painful somatic symptoms such as joint pains, lumbar pain and headache has also been reported [12-18]. In some cases, these somatic symptoms constitute the principal reason for consultation rather than the emotional symptoms [19]. Moreover, some studies have shown that the presence of somatic symptoms contributes greatly to the recurrence of another new depressive episode several years later [15,20-24]. In addition, this association apparently does not depend on social or economic factors such as gender, income level, education or age, but it is rather inherent to the condition just as emotional symptoms are [25-27].
Somatic symptoms have been described as part of a cultural language of affective disorders that, if misinterpreted by the clinician, can lead to unnecessary diagnostic procedures or to inadequate treatment [28]. Studies with depressed patients treated in a primary care setting show that somatic and anxiety symptoms are often overlooked in the diagnosis of depression despite such symptoms contributing significantly to detecting the disease, according to a recent logistic regression analysis [29]. In a study with primary care treated outpatients (n = 1, 456), 70% of whom were Latinos, the prevalence of somatization was 22%. Of the sub-sample with somatization, 35% had a Major Depressive Disorder (MDD) compared to only 13.7% for those without somatization (p < 0.0001) [30]. Studies comparing Latino populations with MDD residing in Colombia, Peru and Puerto Rico versus non-Latino populations with MDD in the United States, show generally higher indices of somatization despite similar rates of depressive symptoms [25,31,32].
However, very little is known about the role of somatic symptoms in populations of depressed patients treated by psychiatrists. It is possible that this group of physicians fails to recognize, value, or else underestimates the presence of these type of symptoms as an integral part of major depression and, therefore, does not monitor their evolution during the course of treatment. Some unanswered questions are: prevalence of somatic symptoms in depressed patients evaluated by psychiatrists, level of recognition of such symptoms by psychiatrists, clinical relevance of somatic symptoms in a psychiatric practice, impact of somatic symptoms on recovery and remission in depressed patients treated with different antidepressants, differences between available antidepressant medications in the improvement of painful or painless somatic symptoms.
In an effort to clarify some of these questions, we conducted an observational study in Puerto Rico. Our research method allowed for the collection of data from a representative group of psychiatrists on their level of recognition of painful or painless somatic symptoms, compared to the reports of their patients with a MDE without interfering with their usual clinical practice.
Methods
Selection of patients
The recruiting phase took place between February and December 2003 in 25 private outpatient psychiatry practices geographically distributed throughout Puerto Rico. Male and female subjects over 21 years of age diagnosed with MDE according to the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR) were eligible for the study. Patients were excluded from the study if they had participated in clinical studies thirty (30) days prior to the first study visit. They were also excluded if they were hospitalized or met the diagnostic criteria for: refractory depression, defined as poor response to two or more appropriate antidepressant treatments for at least 12 to 16 weeks, according to the guidelines published by Souery et al. [33]; bipolar disorders; psychotic disorders; dementia; secondary depression; or painful or painless somatic symptoms of known etiology.
Study design
Before beginning the study, the protocol was reviewed and approved by an institutional review board (IRB). Before being subjected to any study procedure, the study subjects signed an IRB approved informed consent form with a detailed explanation of the procedures and risks involved in the study.
Since this study was non-interventional, it did not contain requirements or recommendations on antidepressant or other treatment. The participating psychiatrists were free to choose the type and course of treatment for each of their patients to the best of their own clinical judgement and in accordance to their usual practice. There were no restrictions concerning the use of other therapies concomitantly with the antidepressant medication. Because the investigation sites were the psychiatrists' own private medical offices, research tools that interfered minimally with the usual clinical practice were utilized to collect the study data.
Primary and secondary measures
The presence, severity, and number of somatic symptoms were recorded during two study visits separated by an 8 week interval, by means of the trained psychiatrists own reports and a case report form with a table to list: all the somatic symptoms asked to or reported by the patients during the last 4 weeks, and the level of severity of each one of them (1 = mild; 2 = moderate; 3 = marked; 4 = severe). The patients, on the other hand, used a self-evaluation scale: the Somatic Symptom Inventory (SSI) [34]. The SSI is a 26-item questionnaire. In this inventory, the patients' degree of discomfort for each symptom is rated from 1 to 5 (1 = absent; 3 = moderate; 5 = a great deal). All the investigators received a structured training before the first patient visit. This training included a review of the protocol, the case report forms, and the SSI scale. With respect to the SSI, the psychiatrists were instructed to use the scale as a general reference during the patient interview and somatic symptom recording but not to follow the scale's items or language literally because this could affect the patient's spontaneous response to the self-administered SSI at the end of the visit.
We categorized total SSI scores as minimal (= ≤ 52) and moderate to high (= ≥ 52) according to the patients' degree of somatic symptom discomfort. This post-hoc cut-off point was arbitrarily determined based on the high number (mean of 14) of reports by patients of somatic symptoms causing at least some degree of discomfort. Internal analysis supported this cut-off point to statistically differentiate between two groups of patients in terms of depression severity and other variables discussed below. We also conducted subscale analyses for both painless (SSI items 1, 4–8, 10–13, 15–18, 20–26) and painful (SSI items 2,3,9,14,19) somatic symptoms (Table 4). Because the SSI that we used only included 5 painful somatic symptoms, limiting the options to detect any difference between the patients' and physicians' reports, we decided to ask the patients with painful symptoms to specify the location of their symptoms with the help of human silhouettes and to rate the impact of the treatments received on the magnitude of the pain using a Visual Analogue Scale (VAS).
Table 4 Somatic symptoms (SSI) scores: degree of discomfort analysis (V2 vs V1) (n = 88)*
SSI Symptom SSI Score (Visit 1) SSI Score (Visit 2) ‡ p-Value
1. Nausea and vomiting 1.8 1.6 NS
2. Muscles soreness 2.3 2.1 <0.05
3. Pains or cramps in your abdomen 2.0 1.8 NS
4. Feeling faint or dizzy 2.4 2.0 <0.05
5. Trouble with your vision 2.2 2.1 NS
6. Muscles twitching or jumping 3.1 2.6 <0.05
7. Feeling fatigued, weak, or tired all over 3.4 2.9 <0.05
8. A fullness in your head or nose 2.6 2.3 NS
9. Pain in your lower back 3.2 2.6 <0.05
10. Constipation 2.4 2.2 NS
11. Trouble catching your breath 2.1 1.9 NS
12. Hot or cold spells 2.4 2.3 NS
13. A ringing or buzzing in your ears 2.0 1.8 NS
14. Pains in your heart or chest 2.3 1.9 <0.05
15. Difficulty keeping your balance while walking 2.2 1.9 NS
16. Indigestion, upset stomach, or acid stomach 2.7 2.4 NS
17. The feeling that you are not in as good physical health as most of your friends 3.4 2.9 <0.05
18. Numbness, tingling, or burning in parts o your body 2.7 2.2 <0.05
19. Headaches 3.3 2.6 <0.05
20. A lump in your throat 2.2 2.0 NS
21. Feeling weak in parts of your body 2.9 2.4 <0.05
22. Not feeling well most of the time in the past few years 3.4 2.9 <0.05
23. Heavy feelings in your arms or legs 2.9 2.6 NS
24. Your heart pounding, turning over or missing a beat 2.4 2.0 NS
25. Your hands and feet not feeling warm enough 2.2 1.8 NS
26. The sense that your hearing is not as good as it used to be 2.4 2.1 NS
Painless subscale† 51.6 52.5 NS
Painful subscale 13.0 11.0 <0.05
Total SSI 63.4 56.8 <0.05
*SSI – Somatic Symptom Inventory (degree of discomfort: 1 = absent; 2 = a little bit; 3 = moderate; 4 = quite a bit; 5 = a great deal); ‡ 8 weeks after Visit 1; † Somatic painless subscale correspond to the SSI items: 1, 4–8, 10–13, 15–18, 20–26.
NS = Non significant
The effects on the emotional and somatic symptoms of antidepressant therapy, as selected by the psychiatrists, were evaluated after the 8-week interval between the two study visits. The Clinical Global Impression – Severity (CGI-S) scale was used by the psychiatrists to evaluate any changes in the severity of the MDE. Also, characteristic major depressive episode emotional symptoms such as depressed mood, guilt-related thoughts, feelings of worthlessness, anhedonia, psychomotor agitation or retardation, loss of concentration, anxiety, psychotic symptoms, and suicide behavior or thoughts, were each analyzed individually according to their clinically rated severity (1 = absent; 3 = moderate; 5 = severe). Patients evaluated also the impact of antidepressant therapy in their MDE using the Patient Global Impression – Improvement (PGI-I) and a VAS for assessment of depression. Additional information on the employment status, drug abuse or dependency, concomitant therapies used to manage somatic symptoms, and the efficacy of these therapies was also collected.
Statistical methods
The study was designed with a power of 90% to detect average differences in the level of recognition of at least one somatic (physical) symptom greater or equal to 23% by both groups analyzed, psychiatrists and patients, according to the estimate obtained using the least squares method, and based on previous data from studies on the recognition of symptoms in primary care patients.
Correlation analyses used to determine the level of agreement between the psychiatrists' evaluations and patient self-evaluations were conducted using Kappa statistics. Unanswered questions were not considered in the analysis. Symptoms were classified using a sensitivity analysis, considering the patients' report as point of comparison: >0.75, high degree of agreement beyond chance; 0.40 to 0.75, fair agreement beyond chance and <0.40, low degree of agreement.
The comparison of the number of symptoms reported by the psychiatrists versus the patients was performed using the statistical t test for independent samples. Secondary analyses were conducted to evaluate any significant statistical differences in clinical and demographic variables between genders, age groups, employment status, type of antidepressant medications, and the severity of the depression. The chi-square test for variable categories was used for this analysis.
Lastly, other secondary analyses were conducted: possible relationships between emotional and physical symptoms reported by patients and clinicians, the level of recognition of emotional symptoms by psychiatrists as compared to their patients' reports, and associations between different symptoms and response levels to the different antidepressants. The level of significance was set at P = 0.05. In order to describe the demographics of the study population, a univariate analysis was conducted by calculating the following descriptive statistics: median, variance and standard deviation.
Results
Patient characteristics
A total of 145 patients were evaluated in the initial visit. Table 1 presents the demographic values of the study population. Of the 145 patients that entered the study, 129 completed the procedures for visit 1 and 42 of them withdrew from the study before visit 2. The most common causes for withdrawing were: consent withdrawal (13.2%), visit 2 outside the stipulated protocol window (9.3%), researcher's decision (2.3%), and patients' change in provider (1.6%). None of the patients withdrew from the study because of adverse events or reasons related to the medications prescribed by their physicians. This report includes all the patients who completed at least the initial visit. For the analysis comparing results at visit 2 versus visit 1 we used the data of the 87 completers.
Table 1 Demographic variables of the study population (initial visit)
Variable N Value or % (SD)
Average Age (st. dev.) 145 44.5 (11.77)
Gender: Female 113 77.93%
Race and Ethnicity
Hispanic (Latino) 145 100%
Caucasian 0
African-Caribbean 0
Other 0
Employment Status
Employed 49 33.8%
Housewife 6 4.1%
Student 29 20%
Unemployed 1 0.01%
Unable to work 60 41.4%
For almost half the patients in the study, the clinicians prescribed concomitant medications in addition to the antidepressants they had selected. The concomitant medications most frequently reported were anxiolytics (37.9% of the patients), antipsychotics (10.6%), tricyclic antidepressants (13.6%) and analgesics (9.1%). At baseline, alcohol consumption rates were reported for 28 of the 144 patients (19.4%). None of the patients admitted any consumption of controlled substances.
Clinical findings
As mentioned in the Methodology section, patient responses were used as points of reference for different comparisons. Somatic symptom SSI patient self-reports were correlated to psychiatrists' somatic symptom reports, which the physician recorded during the patient interview by using the SSI as a general reference. All the patients included in the analysis reported at least one somatic symptom in the SSI. A low agreement between the psychiatrist report and his patients for somatic painless symptoms was observed in both visits. The Kappa values obtained for each painless somatic symptom were less than 0.1159 (low degree of agreement) with variable sensitivity ranged from 1.27% for walk balance difficulty to 18.81% for muscle twitches. In other words, the probability of psychiatrists reporting the presence of any somatic painless symptom in their patients with a MDE, which reported having this symptom, was less than 19% in this sample. In other way, a variable percentage of false negative were also reported, ranged from 81% for muscle twitches to 100% for lump in throat, meaning that near all painless somatic symptoms were erroneously interpreted by the psychiatrist as present in some cases, but they were not reported by the patient (Table 2). The total average number of painless symptoms reported by the psychiatrists was only an eighth of the number reported by the patients (2 vs. 16; p < 0.001). The number of painless somatic symptoms reported by the patient from baseline to endpoint visit was similar, showing consistency and no improvement despite treatment of their MDE (p = 0.999). No significant differences in the number of painless somatic symptoms were observed when data was adjusted for gender or age.
Table 2 Relationship between the treating psychiatrists and the patients reports of painless somatic symptoms
SSI Symptom Kappa Coefficient* N Sensitivity (%)† False negative (%)†
Nausea/Vomiting 0.1159 131 12.70 87
Constipation 0.0594 128 6.85 93
Dizziness 0.0521 133 10.75 89
Muscle twitches 0.0512 133 18.81 81
Breathlessness 0.0418 133 6.58 93
Fatigue 0.0261 133 9.48 90
Weakness 0.0225 133 4.85 95
Palpitations 0.0201 131 6.33 93
Lump in throat 0.0151 133 NA 100
Walk balance difficulty 0.0105 135 1.27 98
Head/nose fullness 0.0085 133 1.92 98
Body Numbness 0.0042 134 3.19 96
Other Symptoms 0.0000 132 NA NA
*Kappa Coefficient: <0.75 = high degree of agreement beyond chance; 0.40 to 0.75 = fair agreement beyond chance; <0.40 = low degree of agreement. †For major explanation regarding sensitivity and false negative items please refer to the principal text.
NA = Non applicable.
With regards to the painful somatic symptoms, contrary to the painless ones, a fair agreement between the psychiatrist report and his patients was observed for upper limbs joint pains, abdominal pain, back pain and headache (0.4008 < κ >0.5788). Variable sensitivity ranged from 46.7% for shoulder pain to 77.2% for headache, this being one of the symptoms with higher sensitivity. Other painful symptoms, including pain in the lower limbs and joints, showed a low degree of agreement with lower sensitivity values. With regards to the percentage of false negatives, these ranged from 22% for headache to 66% for lower limb pain (Table 3). The number of painful symptoms reported by psychiatrists was half that reported by their patients (1.5 vs. 3; p < 0.001). The number of painful symptoms reported by patients was similar in both visits showing little change despite treatment of their MDE (p = 0.1453). No significant differences in the number of painful symptoms were observed when data was adjusted for gender or age.
Table 3 Relationship between the treating psychiatrists and the patients reports of painful somatic symptoms
Type of Pain Kappa Coefficient* N Sensitivity (%)† False negative (%) †
Hand Pain 0.5788 125 64.52 55
Abdominal Pain 0.5710 125 57.14 42
Shoulder Pain 0.5303 125 46.67 53
Back Pain 0.4180 125 62.71 37
Headache 0.4008 125 77.22 22
Knee pain 0.4034 125 39.13 60
Lower limb pain 0.3826 125 33.33 66
Foot Pain 0.3736 125 44.44 55
Other Pains 0.00 – 0.32 125 6.8 – 28.5 71 – 95
*Kappa Coefficient: <0.75 = high degree of agreement beyond chance; 0.40 to 0.75 = fair agreement beyond chance; <0.40 = low degree of agreement. †For major explanation regarding sensitivity and false negative items please refer to the principal text.
In spite of minimal changes in the number of somatic symptoms between both study visits, we observed some changes in the patients' degree of discomfort according to the SSI scores. Some symptoms exhibited a little but significant score reduction, although not enough to be considered absent after the 8 weeks of treatment (Table 4).
The data obtained from the SSI in the initial visit allowed us to observe certain characteristics common to our population sample. For example, one hundred percent of the patients reported some type of somatic symptom associated with their major depressive episode. This percentage did not vary after an average of 8 weeks of treatment with different antidepressants selected by the treating clinicians. Age, female gender, and unemployment were variables significantly associated with the presence of somatic symptoms (p < 0.05). The severity of the depression according to the patient self-evaluation was greater in those patients with moderate to high degree of discomfort with their somatic symptoms (p < 0.001) (Table 5).
Table 5 Sample characteristics at initial visit according to the total SSI score*
Variable Minimum degree of discomfort (n = 34)¶ Moderate/High degree of discomfort (n = 87)¶ p-Value
Average SSI Score (st.dev.) 41.0 (7.7) 76.06 (16.65) <0.001
Average Age (st.dev.) 40.69 (13.23) 45.5 (10.00) 0.0193
Gender:
Female (77.9%) 29.25 70.75 0.034
Employment Status:
Employed (58.5%) § 41.30 58.70 0.013
Antidepressants:
SSRIs (61%)† T 34.88 65.12 0.553
Other antidepressants 29.03 70.97
(26.5%) †
Total VAS Score ‡ 49.8 70.8 <0.001
*SSI – Somatic Symptom Inventory; §Includes housewives and students; T Includes fluoxetine, setraline, paroxetine, citalopram; † Includes venlafaxine and bupropion; ‡ Self-evaluation analogue visual scale of the depression scored by the patients; ¶Minimum degree of discomfort due to somatic symptoms (SSI ≤ 52 – 26 items); Moderate/high degree of discomfort due to somatic symptoms (SSI > 52 – 26 items).
With respect to emotional symptoms, in the first visit only 17.5% of the patients were classified with a mild clinical depression by the psychiatrists, and the average score of the total CGI-S was 4.3. This suggests a moderate to high level of severity of depression for the study sample in general. Likewise, all individual emotional symptoms were reported as moderate to high severity. Slightly more than 35% of the patients reported moderate to severe feelings of guilt and/or hopelessness according to his psychiatrist. Thirty-eight percent reported moderate to severe anhedonia, 34.3% psychomotor changes, 42.4% loss of concentration, 31.9% significant levels of anxiety, 36.1% psychotic symptoms, and 31.9% suicidal thoughts.
At the end of 8 weeks of treatment with various antidepressants, there was a significant reduction in the basal score of the total CGI-S and in each of the depressive symptoms analyzed (p < 0.0001). Nevertheless, the average score in visit 2 was 3.2 and 38.1% of patients maintained moderate to high levels of severity. Only 13% of the patients achieved a score of ≤ 2 on the CGI-S (borderline depression or normal). Patients also reported feeling better with antidepressant treatment during visit 2 as reflected by a reduction in the total average PGI-I score (3.22 to 2.73; p < 0.0019) and the VAS score (from 0.63 to 0.49 p < 0.0001). No significant statistical differences in visit 2 VAS scores were noted between patients who were already on antidepressant therapy in visit 1 (87.5%) and those who were not, but this finding could be due to the sample size.
In general, although depressive symptoms decreased significantly in visit 2 compared to visit 1, a similar trend was not detected for somatic symptoms. No improvement was noted in the painless somatic symptoms as reflected by visits 1 and 2 SSI subscale scores (51.6 and 52.5 respectively; p = NS). Slight reduction in the degree of discomfort due to the following somatic symptoms was noted: dizziness, fatigue, muscle twitches, poor physical health, body numbness, weakness and "feeling bad". However, there were no appreciable changes in gastrointestinal, sensory and cardiopulmonary symptoms during the 8 weeks of treatment. Finally, slight but significant changes were noted in visits 1 and 2 SSI subscale scores for painful symptoms (13 and 11, respectively; p < 0.05) and total SSI scores (63.4 and 56.8 respectively; p < 0.05).
The relation between the patients' degree of discomfort with their somatic symptoms and the severity of their emotional symptoms was also analyzed. This study did not reveal any relation between the patients' level of response to antidepressant treatment and their degree of discomfort due to somatic symptoms (painless or painful) in visit 1. Patients with minimal as well as those with moderate to high degree of discomfort due to somatic symptoms demonstrated significant reductions on the CGI-S at the end of 8 weeks of treatment (Figure 1). Slight but significant reductions in depression and painful symptoms severity according with the Pain relief Visual Analogue Scale and the PGI-I – both of which were rated by the patients – were observed after 8 weeks of treatment. A proportional relationship between the pain and depression severity was observed (p < 0.0001) (Figure 2).
Figure 1 Analysis of the SSI total score related to the mean CGI-S score. The patients were divided into two groups according to their degree of discomfort with their somatic symptoms: minimum (≤ 52) (n = 29) and moderate/high (>52) (n = 50). CGI-S is measured from 1 to 7, where "1" corresponds to the absence of depressive symptoms and "7" to the greatest possible severity of depressive symptoms. Abbreviations: SSI – Somatic Symptom Inventory; CGI-S – Clinical Global Impression of Severity. *p < .001
Figure 2 Relationship between pain severity and depression improvement (patient report). Regression analysis shows an inversely proportionate relationship between the severity of the pain reported by the patients and the level of improvement from the major depressive episode according to the patients' global impression (PGI-I) (n = 87). Abbreviations: VAS – Visual Analogue Scale; PGI-I – Patient Global Impression of Improvement.
Discussion
Somatization is common to all cultures and social groups studied. However, there are differences in the styles of expression and attribution of symptoms according to the beliefs and health practices of each culture [35-38]. Somatization has been more frequently associated with Latino populations than with other ethnic groups [31]. It has been mentioned that societies that promote individualism with clear limits in interpersonal relations seem to value the direct expression of unpleasant feelings. On the other hand, collective societies with more flexible levels of relationships tend to place greater value on the indirect expression of feelings (idiom of distress hypothesis) [39]. The Psychological Problems in the General Health Care study (PPGHC) of the World Health Organization (WHO) was conducted in 14 cities from the same number of countries on four continents. It showed that adult patients treated in primary outpatient care centers were characterized by a high prevalence of reports of somatic symptoms in the two participating Latino cities (Rio de Janeiro in Brazil and Santiago in Chile), compared to the total sample (32% and 36.5% vs. 19.7%), according to the SSI scale. These results did not depend on the level of development, education, or gender of the cities evaluated [40]. The WHO study did not demonstrate culture to have a determining role in the manifestation of somatic symptoms except for the two Latin American cities evaluated.
Cultural differences can even be observed within different Latino populations in the prevalence of somatic symptoms or chronic pain. A comparative study of Mexican-Americans, natives living in Puerto Rico, and Caucasian-Americans used an analysis of 5 clusters of symptoms and demonstrated that the Puerto Ricans presented the highest levels of somatization [25,41]. In other publications, the same authors concluded that since the high rates of somatization in Puerto Ricans were not accompanied by higher rates of prevalence of depressive disorders compared with the US, Puerto Rican physicians were likely attributing somatic symptoms as "psychogenic" and not considering them manifestations of a MDE [42,43]. Another study of depressed patients living in the United States reflected prevalence rates of chronic abdominal pain that differed significantly between the three Latino groups evaluated: 4.6% in Mexican-Americans; 5.8% in Cuban-Americans; and 8.3% in Puerto Ricans. Logistic regression analyses showed a close relation between depression and chronic abdominal pain, female gender, and single marital status [44].
In our study, a direct relationship between moderate to high degree of discomfort due to somatic symptoms and age, female gender and unemployment was also observed (Table 5). In this study of Puerto Rican patients with a MDE, 100% of patients reported somatic symptoms using the SSI scale. Although this percentage is very high, previous depression trials in primary care settings have also shown that 69% to 92% of MDE patients experience somatic symptoms [8-11]. Somatization has been reported to be particularly more prevalent in Puerto Rican depressed patients than in other Hispanic populations. However, we cannot exclude the possibility that the use of an inventory such as SSI led to a higher number of somatic symptom reports than would have been spontaneously reported by patients. We could have avoided this potential bias by using a validated scale to collect information on somatic symptoms as a primary research tool. However, this would have precluded us from achieving the primary objective of our study.
Communication between patients and their physicians is influenced by various cultural factors such as the patient's perception of her emotional and physical symptoms, her report of these symptoms to the physician, and the physician's interpretation of the symptoms. These steps described by Leff [45] in the communication of unpleasant emotions could affect the final evaluation of the physician with respect to the clinical status of the patient. Since 1960, significant differences have been shown between patient self-evaluations of their depressive symptomatology and the evaluations conducted by their physicians. Carrol et al. [46] suggest that these differences are due to discrepancies between the instruments used by the patients and clinicians. They proved that the use of instruments with similar structure and matching items allow higher correlation coefficients. Corruble et al. [47], using structured instruments with similar items, demonstrated in their study with 64 hospitalized depressed patients that there was significant agreement between the patients and the psychiatrists' report of the severity of the depression. Nevertheless, depressed patients with high levels of somatization and anxiety exhibited a tendency to overestimate their symptomatology compared to what was reported by their physicians. To date, few studies have explored the level of agreement in the recognition of somatic symptoms between patients and physicians. A study with primary care physicians showed that these clinicians and their patients can be more comfortable with somatic symptoms than with emotional symptoms, which leads to an under-reporting of depression in those patients with associated physical symptoms [48]. Contrary to this, a review focused on the recognition of somatoform disorders by psychiatrists indicated that these specialists are more concerned with severe mental disorders than with the recognition of somatic symptoms. Psychiatrists attempt to "normalize" the physical symptoms expressed by their patients and tend to refer them to other specialists [49].
In this study, despite similar SSI mean scores for painful and painless symptoms (2.6 vs. 2.5, respectively) (Table 4), when compared to painful symptoms of the torso, upper limbs and head, painful symptoms of the lower limbs and painless somatic symptoms showed a lack of correlation between the psychiatrists' and the patients' reports. The marked difference between psychiatrists' and patients' somatic symptoms reports may be explained by the different methods used to document somatic symptoms and degree of discomfort caused by them, or may be due to a tendency by psychiatrists to dismiss certain type of pain or painless somatic symptoms in their usual clinical practice. This study was purposefully designed to avoid any intervention and allow for a naturalistic observation of usual clinical psychiatric care in Puerto Rico. The use of a structured symptom checklist for psychiatrists would not have been representative of the usual clinical practice and would not have allowed us to document the degree of recognition of somatic symptoms in MDE patients in this clinical setting.
Additionally, we observed that those somatic symptoms with lower sensitivity percentages usually have higher false negative percentages than those with fair agreement levels (Tables 2 &3). This suggest that psychiatrists may not only omit several somatic symptoms in their patients with MDE but may also detect and document other symptoms unreported by their patients. This observation may be a result of the study design, which in itself could have increased the psychiatrists' motivation to report somatic symptoms and may not represent necessarily their usual clinical practice.
Failure to detect somatic symptoms in depressed patients may have significant implications in the cost of treatment. Reid et al. [50] compared the health-care utilization patterns of patients with medically unexplained symptoms with those of other frequently referred patients. "Somatising patients" had at least two medical consultation visits for unexplained physical symptoms, a greater number of referrals for secondary care and were more likely to undergo additional clinical work-up and tests. In other trials, rates of health resources utilization for MDD patients with somatic symptoms were nine times higher than for the general population and three times higher than for all depressed patients [51,52]. Moreover, individuals with MDD consistently exhibit the highest rates of loss of productive time (LPT) when they concomitantly present symptoms like pain, weakness, fatigue, gastrointestinal discomfort or sensory changes [53]. Given the duration and size of the sample, this study did not show differences in certain variables related to the use of health resources such as consumption of analgesics or other medications. However, it did reflect a higher rate of unemployment in patients with moderate to high severe somatic symptoms.
Additionally, failure to detect somatic symptoms in depressed patients has significant impact in the response to treatment. The presence of physical symptoms in MDD patients has been associated with a poor level of response to antidepressant treatment [54]. Appropriate treatment to control non-emotional symptoms is essential in order to achieve proper compliance with therapy [55], to decrease the risk of recurrences [56], to maximize earlier initiation of the antidepressant action [57], and to increase the opportunities for complete remission of the depressive episode [58]. In this study, although there was a significant reduction in the average CGI-S depression score, only 13% of the patients reached a level of remission defined post-hoc as CGI ≤ 2.
Several studies with Selective Serotonin Reuptake Inhibitors (SSRIs) have demonstrated that depressed patients with somatic symptoms exhibit lower rates of response than those without somatic symptoms [54,59,60]. Additionally, somatic symptoms respond less to SSRIs than non-somatic symptoms [61]. According to several authors [7,62,63], antidepressants that act dually on the noradrenaline and serotonin pathways could be better candidates for treating both depression and concomitant somatic symptoms. In this study, 61% of the patients were treated with a SSRI and 13% with low doses of venlafaxine (average 56 mg/day). Clinicians did not report differences, according to the CGI-S, in the level of antidepressant response between patients with minimal and those with moderate to high degree of discomfort with their somatic symptoms. Nevertheless, it was possible to determine that those patients who reported less relief of painful symptoms exhibited lower levels of reduction in the PGI-I for depression. In other words, although psychiatrists were not able to observe significant differences in the level of response to treatment, patients who had less relief of painful symptoms reported a lower response to antidepressant treatment (Figure 2). This difference in the reporting of the response to treatment could be due to the emphasis placed by psychiatrists on detecting and controlling emotional symptoms rather than somatic symptoms. It also suggests that inappropriate identification of somatic symptoms may lead to an erroneous perception of appropriate antidepressant treatment response vis-à-vis the experience of the patient.
Our study has a number of strengths. Through a confirmatory analysis, we explicitly tested the level of recognition of somatic symptoms in patients with a MDE by their psychiatrists. This analysis was conducted for all the somatic symptoms as a group and separately according the presence of pain or not. We studied a random sample of the population, and we used valid and reliable measures.
One potential limitation in our study is the use of the SSI, which is a valid tool to detect the presence and degree of discomfort due to somatic symptoms. However, the SSI is probably more predictive in the case of anxiety disorders than in depression and has a reduced sensitivity in the detection of symptom changes due to treatment [64]. Another possible limitation is our use of a spontaneous report measure for somatic symptoms by psychiatrists. A structured psychiatric interview may have yielded different results. However, spontaneous reports were the only way not to bias the researchers' responses. In addition, patient reports constituted the gold standard for the analysis of data despite the absence of scales or figures to guide the patient in his response, potentially yielding an incomplete report of symptoms.
The type of sampling of patients used could also be a limiting factor. Although the study attempted to reproduce as faithfully as possible the usual psychiatry practice in Puerto Rico with appropriate representation throughout the Island, exclusion and inclusion criteria limited access to hospitalized patients with refractory depression, suicidal behavior and associated medical conditions. These are well-known factors that increase the presence of physical symptoms in patients with major depression [14,17]. Only Latino patients living in Puerto Rico were evaluated. Given the importance of cultural factors in measuring the objectives set forth in this study, it will be difficult to generalize the data to other Latino populations and even less so to other ethnic groups or cultures.
Another potential limitation is the size of the sample. On the one hand, it made possible the detection of significant differences between psychiatrists and their patients with a MDE in the reporting of physical symptoms, but it did not allow confirmation of the impact of antidepressants in the control of physical symptoms. A dichotomous classification of the clinical response of emotional and physical symptoms rather than a continuous measurement of their evolution was used because of the limited sample size.
Finally, the short period of time between the two visits (8 weeks), the absence of intermediate visits and the lack of treatment compliance measurements may limit conclusions on the response to antidepressant therapy.
Conclusion
In summary, these results confirm reports from previous clinical trials about the high rate of somatic symptoms in Puerto Rican patients with a MDE [25,41,44]. Collectively, our data reflect a significant difference in the report of several somatic symptoms – pain in lower limbs and joints and painless symptoms – by the psychiatrists and their patients. Age, female gender, and unemployment are variables significantly associated with the presence of somatic symptoms. The severity of the depression according to the patient self-evaluation is greater in patients with moderate to severe discomfort due to somatic symptoms.
Although depressive symptoms in general reflected a significant reduction in visit 2 compared to visit 1, 38.1% of patients still had moderate to high severity depression and only 13% had CGI-S scores of ≤ 2 (borderline or normal depression). Slight but significant reductions in depression and painful symptoms severity were observed according to the patient evaluations after 8 weeks of treatment. However, a proportional inverse relationship between the pain relief and depression severity was observed. This finding suggests that physical symptoms must be appropriately recognized by psychiatrists because they can interfere with the physicians' assessment of the magnitude of response to antidepressant treatment.
Competing interests
This work was sponsored by Eli Lilly and Company, San Juan, Puerto Rico.
Dr. Tamayo & Rivas are full-time employees and hold shares of Eli Lilly & Co. (San Juan, Puerto Rico). Mrs. Roman was a statistician contractor of Eli Lilly & Co. (San Juan, Puerto Rico). Dr. Fumero has received research support and/or honoraria from Abbott, Astra Zeneca, Bristol-Myers Squibb, Eli Lilly, Glaxo Smith Klaine, Janssen, and Wyeth-Ayers.
Authors' contributions
JT conceived and designed the study and the study clinical report forms, was involved in the drafting the article and the interpretation of the data. KR participated in the study design, made the statistical analyses, helped to draft the manuscript, and was involved in the interpretation of the data. JF participated in the acquisition of the data and revised the article for intellectual content. MR made substantial contributions to the conception of the article and revised the article for intellectual content. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Participating investigators in the Somatic and Pain Recognition in Depression (SaPRiD) project include: Andrea González, MD; Antonio A Milland, MD; Bárbara Díaz, MD; Carlos A Cabán, MD; Carmen Z Sanz, MD; Damaris Mangual, MD; Dessie L Vega, MD; Edelmiro Rodríguez, MD; Fabio H Lugo, MD; Francisco S Guzmán, MD; Ileana M Fumero, MD; Ivette Franceschi, MD; Jaime Marchena, MD; José A Alonso, MD; Juanita C León, MD; Lisa A Principe, MD; Luis M Dorta, MD; Manuel A Brignoni, MD; Mayra M Calderín, MD; Nelsa O Rodríguez, MD; Norberto Pellot, MD; Rafael M Báez, MD; Ricardo A Fumero, MD; Zulma Rodríguez, MD
Other contributors: Madelaine M Wohlreich, MD; Anne C Andorn, MD; Joel Raskin, MD; and Juan D Velásquez, MD.
==== Refs
Kessler RC Berglund P Demler O Jin R Koretz D Merikangas KR Rush AJ Walters EE Wang PS National Comorbidity Survey Replication The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R) JAMA 2003 289 3095 3105 12813115 10.1001/jama.289.23.3095
Moscicki EK Lokce BZ Rae DS Boyd JH Depressive symptoms among Mexican Americans: the Hispanic Health and Nutrition Examination Survey Am J Epidemiol 1989 130 348 360 2750730
Narrow WE Rae DS Moscicki EK Locke BZ Regier DA Depression among Cuban Americans: the Hispanic Health and Nutrition Examination Survey Soc Psychiatry Psychiatr Epidemiolog 1990 25 260 268 10.1007/BF00788647
Vera M Alegria M Freeman D Robles RR Rios R Rios CF Depressive symptoms among Puerto Ricans: island poor compared with residents of the New York City area Am J Epidemiol 1991 134 502 510 1897506
Canino GJ Bird HR Shrout PE Rubio-Stipec M Bravo M Martinez R Sesman M Guevara LM The prevalence of specific psychiatric disorders in Puerto Rico Arch Gen Psychiatry 1987 44 727 735 3498456
Weissman MM Bland RC Canino GJ Faravelli C Greenwald S Hwu H-G Joyce PR Karam EG Lee C-K Lellouch J Lepine J-P Newman S Rubio-Stipec M Wells JE Wickramaratne PJ Wittchen H-U Yeh E-K Cross-National Epidemiology of Major Depression and Bipolar Disorder JAMA 1996 276 293 299 8656541 10.1001/jama.276.4.293
Fava M Mallinckrodt CH Detke MJ Watkin JG Wohlreich MM The effect of duloxetine on painful physical symptoms in depressed patients: do improvements in these symptoms result in higher remission rates? J Clin Psychiatry 2004 65 521 530 15119915
Kroenke K Price RK Symptoms in the community: prevalence, classification, and psychiatry comorbidity Arch Int Med 1993 153 2474 2480 8215752 10.1001/archinte.153.21.2474
Ebert D Martus P Somatization as a core of melancholic type depression: evidence from a cross-cultural study J Affect Disord 1994 32 253 256 7897089 10.1016/0165-0327(94)90089-2
Simon GE VonKorff M Piccinelli M Fullerton C Ormel J An international study of the relation between somatic symptoms and depression NEJM 1999 341 1329 1335 10536124 10.1056/NEJM199910283411801
Corruble E Guelfi JD Pain complaints in depressed inpatients Psychopathology 2000 33 307 309 11060514 10.1159/000029163
Mathew NT Reuveni U Perez F Transformed or evolutive migraine Headache 1987 27 102 106 3570762 10.1111/j.1526-4610.1987.hed2702102.x
Katon W Von Korff M Lin E Lipscomb P Russo J Wagner E Polk E Distressed high utilizers of medical care. DSM-III-R diagnoses and treatment needs Gen Hosp Psychiatry 1990 12 355 362 2245919 10.1016/0163-8343(90)90002-T
Gerber PD Barrett JE Barrett JA Oxman TE Manheimer E Smith R Whiting RD The relationship of presenting physical complaints to depressive symptoms in primary care patients J Gen Intern Med 1992 7 170 173 1487765
Kroenke K Spitzer RL Williams JB Linzer M Hahn SR deGruy FV 3rdBrody D Physical symptoms in primary care. Predictors of psychiatric disorders and functional impairment Arch Fam Med 1994 3 774 779 7987511 10.1001/archfami.3.9.774
Juang K-D Wang S-J Fuh J-L Lu S-R Su T-P Comorbidity of depressive and anxiety disorders in chronic daily headache and its subtypes Headache 2000 40 818 823 11135026 10.1046/j.1526-4610.2000.00148.x
Ohayon MM Schatzberg AF Using chronic pain to predict depressive morbidity in the general population Arch Gen Psychiatry 2003 60 39 47 12511171
Posse M Hallstrom T Depressive disorders among somatizing patients in primary health care Acta Psychiatr Scand 1998 98 187 192 9761404
Simon GE VonKorff M Recognition, management, and outcomes of depression in primary care Arch Fam Med 1995 4 99 105 7842160 10.1001/archfami.4.2.99
Dworkin SF Von Korff M Le Resche L Multiple pains and psychiatric disturbance: An epidemiological investigation Arch Gen Psychiatry 1990 47 239 244 2306165
Simon G VonKorff M Somatization and psychiatric disorders in the NIMH Epidemiological Catchment Area Study Am J Psychiatry 1991 148 1494 1500 1928462
Zwaigenbaum L Szatmari P Boyle MH Offord DR Highly somatizing young adolescents and the risk of depression Pediatrics 1999 103 1203 1209 10353930 10.1542/peds.103.6.1203
Terre L Poston WS Foreyt J St Jeor ST Do somatic complaints predict subsequent symptoms of depression? Psychother Psychosom 2003 72 261 267 12920330 10.1159/000071897
Ising M Lauer CJ Holsboer F Modell S The Munich vulnerability study on affective disorders: premorbid psychometric profile of affected individuals Acta Psychiatr Scand 2004 109 332 338 15049769 10.1111/j.1600-0447.2004.00309.x
Canino IA Rubio-Stipec M Canino G Escobar JI Functional somatic symptoms: a cross-ethnic comparison Am J Orthopsychiatry 1992 62 605 612 1443069
Piccinelli M Simon G Gender and cross-cultural differences in somatic symptoms associated with emotional distress. An international study in primary care Psychol Med 1997 27 433 444 9089835 10.1017/S0033291796004539
Larson SL Clark MR Eaton WW Depressive disorder as a long-term antecedent risk factor for incident back pain: a 13-year follow-up study from the Baltimore Epidemiological Catchment Area sample Psychol Med 2004 34 211 219 14982127 10.1017/S0033291703001041
Kirmayer LJ Cultural variations in the clinical presentation of depression and anxiety: implications for diagnosis and treatment J Clin Psychiatry 2001 22 28 11434415
Barkow K Heun R Ustun TB Berger M Bermejo I Gaebel W Harter M Schneider F Stieglitz RD Maier W Identification of somatic and anxiety symptoms which contribute to the detection of depression in primary health care Eur Psychiatry 2004 19 250 257 15276656 10.1016/j.eurpsy.2004.04.015
Allen LA Gara MA Escobar JI Waitzkin H Silver RC Somatization: A debilitating syndrome in primary care Psychosomatics 2001 42 63 67 11161123 10.1176/appi.psy.42.1.63
Mezzich JE Raab ES Depressive symptomatology across the Americas Arch Gen Psychiatry 1980 37 818 823 7396661
Escobar JI Gomez J Tuason VB Depressive phenomenology in North and South American patients Am J Psychiatry 1983 140 47 51 6847984
Souery D Amsterdam J de Montigny C Lecrubier Y Montgomery S Lipp O Racagni G Zohar J Mendlewicz J Treatment resistant depression: methodological overview and operational criteria Eur Neuropsychopharmacol 1999 9 83 91 10082232 10.1016/S0924-977X(98)00004-2
Barsky AJ Wyshak G Hypochondriasis and somatosensory amplification Br J Psychiatry 1990 157 404 409 2245272
Kirmayer LJ Young A Robins JM Symptom attribution in cultural perspective Can J Psychiatry 1994 39 584 595 7828110
Isaac M Janca A Orley J Somatization: A cultur-bound or universal syndrome? J Ment Health 1996 5 219 222 10.1080/09638239650036884
Kirmayer LJ Young A Culture and somatization: Clinical, epidemiological, and ethnographic perspectives Psychosomatic Med 1998 60 420 430
Keyes CLM Ryff CD Somatization and mental health: A comparative study of the idiom of distress hypothesis Social Science & Medicine 2003 57 1833 1845 14499509 10.1016/S0277-9536(03)00017-0
Bhatt A Tomenson B Benjamin S Transcultural patterns of somatization in primary care: a preliminary report J Psychosom Res 1989 33 671 680 2621670 10.1016/0022-3999(89)90082-2
Gureje O What can we learn from a cross-national study of somatic distress? J Psychosom Res 2004 56 409 412 15094024 10.1016/S0022-3999(03)00623-8
Shrout PE Canino GJ Bird HR Rubio-Stipec M Bravo M Burnam MA Mental health status among Puerto Ricans, Mexican Americans, and non-Hispanic whites Am J Community Psychol 1992 20 729 752 1302447
Canino G Bird H Rubio-Stipec M Bravo M The epidemiology of mental disorders in the adult population of Puerto Rico P R Health Sci J 1997 16 117 124 9285988
Canino IA Canino G Lopez A, Carrillo E Puerto Ricans The Latino Psychiatric Patient: Assessment and Treatment 2001 Washington DC: American Psychiatric Publishing Inc 141
Magni G Rossi MR Rigatti-Luchini S Merskey H Chronic abdominal pain and depression. Epidemiologic findings in the United States. Hispanic Health and Nutrition Examination Survey Pain 1992 49 77 85 1594284 10.1016/0304-3959(92)90191-D
Leff JP Psychiatrists' versus patients' concepts of unpleasant emotions Br J Psychiatry 1978 133 306 313 709003
Carrol BJ Feinberg M Smouse PE Rawson SG Greden JF The Carrol Rating Scale for Depression: development, reliability and validation Br J Psychiatry 1981 138 194 200 7272609
Corruble E Legrand JM Zvenigorowski H Duret C Guelfi JD Concordance between self-report and clinician's assessment of depression J Psychiatr Res 1999 33 457 465 10504014 10.1016/S0022-3956(99)00011-4
Goldman LS Nielsen NH Champion HC Awareness, diagnosis, and treatment of depression J Gen Intern Med 1999 14 569 580 10491249 10.1046/j.1525-1497.1999.03478.x
Bass C Peveler R House A Somatoform disorders: severe psychiatric illnesses neglected by psychiatrists Br J Psychiatry 2001 179 11 14 11435262 10.1192/bjp.179.1.11
Reid S Wessely S Crayford T Hotopf M Frequent attenders with medically unexplained symptoms: service use and costs in secondary care Br J Psychiatry 2002 180 248 253 11872517 10.1192/bjp.180.3.248
Smith GR The course of somatization and its effects on utilization of health care resources Psychosomatics 1994 35 263 267 8036255
Katon W Schullberg HC Epidemiology of depression in primary care Gen Hosp Psychiatry 1992 14 237 247 1505745 10.1016/0163-8343(92)90094-Q
Stewart WF Ricci JA Chee E Hahn SR Morganstein D Cost of Lost Productive Work Time Among US Workers With Depression JAMA 2003 289 3135 3144 12813119 10.1001/jama.289.23.3135
Papakostas GI Petersen T Denninger J Sonawalla SB Mahal Y Alpert JE Nierenberg AA Fava M Somatic symptoms in treatment-resistant depression Psychiatry Res 2003 118 39 45 12759160 10.1016/S0165-1781(03)00063-5
Agosti V Quitkin FM Stewart JW McGrath PJ Somatization as a predictor of medication discontinuation due to adverse events Int Clin Psychopharmacol 2002 17 311 314 12409685 10.1097/00004850-200211000-00007
Judd LL Paulus M Schettler P Akiskal HS Endicott J Leon AC Maser JD Mueller T Solomon DA Keller MB Does incomplete recovery from first lifetime major depressive episode herald a chronic course of illness? Am J Psychiatry 2000 157 1501 1504 10964869 10.1176/appi.ajp.157.9.1501
Papakostas GI Petersen TJ Iofescu DV Summergrad P Sklarsky KG Alpert JE Nierenberg AA Fava M Somatic symptoms as predictors of time to onset of response to fluoxetine in major depressive disorder J Clin Psychiatry 2004 65 543 546 15119918
Paykel ES Ramana R Cooper Z Hayhurst H Kerr J Barocka A Residual symptoms after partial remission: an important outcome in depression Psychol Med 1995 25 1171 1180 8637947
Denninger J Merens W Mahal Y Alpert JE Nierenberg AA Fava M The relationship between somatic symptoms and depression 1995
Bair MJ Robinson RL Eckert GJ Stang PE Croghan TW Kroenke K Impact of Pain on Depression Treatment Response in Primary Care Psychosom Med 2004 66 17 22 14747633 10.1097/01.PSY.0000106883.94059.C5
Greco T Eckert G Kroenke K The outcome of physical symptoms with treatment of depression J Gen Intern Med 2004 19 813 818 15242465 10.1111/j.1525-1497.2004.30531.x
Blier P Abbott FV Putative mechanisms of action of antidepressant drugs in affective and anxiety disorders and pain J Psychiatry Neurosci 2001 26 37 43 11212592
Stahl SM Does depression hurt? [BRAINSTORMS] J Clin Psychiatry 2002 63 273 274 12000200
Weinstein MC Berwick DM Goldman PA Murphy JM Barsky AJ A comparison of three psychiatric screening tests using receiver operating characteristic (ROC) analysis Med Care 1989 27 593 607 2725088
|
15967039
|
PMC1183226
|
CC BY
|
2021-01-04 16:33:03
|
no
|
BMC Psychiatry. 2005 Jun 20; 5:28
|
utf-8
|
BMC Psychiatry
| 2,005 |
10.1186/1471-244X-5-28
|
oa_comm
|
==== Front
BMC SurgBMC Surgery1471-2482BioMed Central London 1471-2482-5-161598968810.1186/1471-2482-5-16Case ReportBenign gastro-bronchial fistula – an uncommon complication of esophagectomy: case report Devbhandari Mohan P [email protected] Rohit [email protected] Simon [email protected] Piotr [email protected] Department of Cardiothoracic surgery South Manchester University Hospital, NHS Trust, Southmoor Road, Wythenshawe, Manchester, M23 9LT, UK2 Department of General Surgery South Manchester University Hospital, NHS Trust, Southmoor Road, Wythenshawe, Manchester, M23 9LT, UK2005 30 6 2005 5 16 16 2 3 2005 30 6 2005 Copyright © 2005 Devbhandari et al; licensee BioMed Central Ltd.2005Devbhandari 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
Gastro-bronchial fistula (GBF) is a rare and devastating complication following esophagectomy. Making the correct diagnosis is difficult and there is no agreement on the treatment for this rare condition.
Case presentation
We report the case of a 56-year-old man who presented with features of repeated aspiration and chest infections six years following an esophagectomy for Barrett's esophagus. Despite extensive investigations the cause of symptoms was difficult to determine. The correct diagnosis of fistula from stomach to right main stem bronchus was made at bronchoscopy under general anesthesia. After ruling out local recurrence of cancer, a successful primary repair was carried out by resection of fistula and direct repair of gastric conduit and bronchus. He is well after 6 months of treatment.
Conclusion
Late development of gastro-bronchial fistula is a rare complication of esophageal resection that may be difficult to diagnose.
Surgical resection and direct closure is the treatment of choice, although the method of treatment should be tailored according to the anatomy of the fistula and the patient's condition.
==== Body
Background
Gastrobronchial fistula (GBF) following esophagectomy for malignant disease is a recognized complication. However benign GBF are rare and the literature on this complication is limited consisting of a handful of case reports. Their mode of presentation and management are varied and there is no single treatment method, which is widely accepted. We report a case, which was successfully treated by direct primary repair. A brief review of literature is presented.
Case presentation
A 67-year-old man with past history of Ivor Lewis esophagectomy 6 years previously for a Barrett's esophagus presented with dysphagia, recurrent cough, regurgitation and weight-loss. Endoscopy suggested a benign anastomotic stricture, for which he underwent dilatation. However his symptoms worsened and he continued to lose weight in spite of supervised dietary therapy. Aspiration was suspected and a barium swallow was performed. This demonstrated pharyngeal in-coordination, as the probable cause of aspiration. A repeat endoscopy was non-contributory. A feeding jejunostomy was inserted to prevent continued weight loss. In spite of this, he continued to have repeated chest infections resulting in extensive bronchiectatic changes in the right lower lobe.
A tracheostomy was planned at this stage to prevent aspiration. Whilst he was being intubated for the procedure, air was noted to escape from the right bronchus. A bronchoscopy revealed the presence of the GBF, with the bronchial opening in the right main stem bronchus located about 1 cm below the tracheal bifurcation and the gastric opening midway down the lesser curvature of the gastric conduit. The temporal sequence of events suggests that fistulation had occurred at the time of anastomotic dilatation, and thus the worsening of symptoms following the procedure. The fistula was however, difficult to diagnose in spite of extensive investigations leading to unexpected finding in the anaesthetic room.
Given his debilitated condition and poor respiratory function stenting was considered to occlude the fistula. However as the fistula was positioned in the middle of the gastric conduit (Figure 1) it was not deemed amenable to stenting due to poor fixation of the stent and risk of migration. And similarly satisfactory stenting of right main stem bronchus was difficult to achieve due to short length of this bronchus and consequent tendency to migrate. Though associated with greater risk of morbidity and mortality, a surgical approach was deemed the best approach for this problem.
Figure 1 CT scan of chest of showing communication between the right main bronchus and the gastric conduit.
At operation a right-sided re-thoracotomy was performed and careful dissection carried out to display the anatomy of the GBF (Figure 2). The fistula was excised with a cuff of healthy gastric conduit and bronchial tissue. The gastric and bronchial openings were closed directly with absorbable stitches with interposition of bovine pericardium in between the two suture lines. The patient made an uneventful recovery from operation. A contrast study one-week later confirmed absence of leakage from the gastric conduit and feeding was started. Histopathological examination confirmed the presence of benign GBF in the resected specimen.
Figure 2 Intra-operative picture showing the fistula (arrowed) between two forceps.
The patient was discharged home after two weeks and remains well six months later.
Discussion
GBF is a rare complication in patients with an intra-thoracic oesophagogastric anastomosis after esophageal resection. Recurrent malignancy is the commonest cause of this complication. Non-malignant cases may occur early or late. In the early postoperative period GBF usually arises as a result of the extensive dissection, ischemia [1] or surgical staples [2]. It may however present late as this case, several years after the operation. In such cases the causes are chronic peptic ulcer in the gastric conduit, traumatic anastomotic dilatation [3,4] or infection.
The usual modes of presentation are cough associated with eating, dyspnoea, haemoptysis and recurrent chest infections. The fistula may connect at any site in the respiratory tract, from trachea down to lobar bronchus. The fistula may be difficult to locate and may not be visualized at endoscopy of respiratory or upper GI tract. In such cases a methylene blue dye test may show bluish sputum when the patient is asked to swallow the dye. CT scan or upper GI contrast radiology is also helpful in arriving at the correct diagnosis. Multiple biopsies help to rule out recurrence of malignancy as the cause. Left untreated, GBF is usually fatal due to chronic pulmonary sepsis and one should not rely on conservative treatment [5].
Patients often present in poor general condition with malnutrition and chronic pulmonary infection. Prompt institution of broad-spectrum antibiotic cover, gastric drainage, attention to fluid and electrolyte balance, nutritional support and chest physiotherapy are essential steps in preparation for definitive surgery. In acute cases with the presence of gross mediastinal sepsis cervical oesophagostomy and return of the stomach tube to abdomen with débridement is performed. The continuity can be restored at a later date with an alternative conduit e.g. colonic interposition.
The ideal operation consists of re-thoracotomy and resection of fistula with direct closure of the openings in the esophagus and the respiratory tree, preferably with an intervening viable tissue. A variety of tissue and pedicles of muscles like pectoralis major, intercostal, latissimus dorsi [4] and sternocleidomastoid [3] have been used to interpose between the two repaired tubes. Where required the membranous portion of airway can be substituted with fascia lata, autologous pericardium or bovine pericardium to close the defect. A defect in the lower lobe bronchus is most often managed by lower lobectomy where as a defect in the main-stem bronchus may be resected and repaired [5]. GBF arising around the level of the oesophagogastric anastomosis may not be amenable to resection and direct closure. A tension free well-vascularized closure is crucial in the success of this procedure and an alternative conduit may be required. No one single procedure is suitable for all the patients and the surgeon should be aware of the options available.
Late development of GBF is a rare complication of esophageal surgery that may be difficult to diagnose. Surgical resection and direct closure is the treatment of choice. However where patients are debilitated or if there is insufficient length of airway available for resectional surgery, alternative, less invasive endoscopic stenting procedures may be used [6].
Conclusion
Late development of gastro-bronchial fistula is a rare complication of esophageal resection that may be difficult to diagnose.
Surgical resection and direct closure is the treatment of choice, although the method of treatment should be tailored according to the anatomy of the fistula and the patient's condition.
List of abbreviations used
GBF : Gastrobronchial Fistula
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MPD collated the information, searched literature and wrote the manuscript.
RJ assisted in literature search, writing of the manuscript, editing the final script and doing the final submission.
SG and PK assisted in providing a critical appraisal and review of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Written consent was obtained from the patient for publication of the case report.
==== Refs
Kalamar K Molnar TF Horvath OP Two cases of benign tracheo-gastric fistula following esophagectomy for cancer Acta Chir Hung 1999 38 261 267 10935135
Pramesh CS Sharma S Saklani AP Sanghvi BV Broncho-gastric fistula complicating transthoracic esophagectomy Dis Esophagus 2001 14 271 3 11869338 10.1046/j.1442-2050.2001.00201.x
Sakamoto K Ogawa M Yamamoto S Mugita N Saishoji T Azuma AS Hayashida K Closure of gastric tube-tracheal fistula by transposition of a pedicled sternocleidomastoid muscle flap Surg today 1997 27 181 185 9018001
Aguilo Espases R Lozano R Navarro AC Regueiro F Tejero E Salinas JC Gastrobronchial fistula and anastomotic esophagogastric stenosis after esophagectomy for esophageal carcinoma J Thorac Cardiovasc Surg 2004 127 297 9 14752457 10.1016/j.jtcvs.2003.07.025
Brega Massone PP Infante M Valente M Conti B Carboni U Cataldo I Gastrobronchial fistula repair followed by esophageal leak – rescue by transesophageal drainage of the pleural cavity Thorac Cardiovasc Surg 2002 50 113 116 11981718 10.1055/s-2002-26685
Bennie MJ Sabharwal T Dussek J Adam A Bronchogastric fistula successfully treated with the insertion of a covered bronchial stent Eur Radiol 2003 13 2222 5 12928968 10.1007/s00330-002-1698-2
|
15989688
|
PMC1183227
|
CC BY
|
2021-01-04 16:28:03
|
no
|
BMC Surg. 2005 Jun 30; 5:16
|
utf-8
|
BMC Surg
| 2,005 |
10.1186/1471-2482-5-16
|
oa_comm
|
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-221604279210.1186/1475-2867-5-22Primary ResearchMultidrug Resistance-Associated Protein 1 (MRP1) mediated vincristine resistance: effects of N-acetylcysteine and Buthionine Sulfoximine Akan Ilhan [email protected] Selma [email protected] Hakan [email protected] Burhan [email protected] Tomris [email protected] Akdeniz University, Faculty of Medicine, Department of Biochemistry, 07070 Antalya, Turkey2 Pamukkale University, Faculty of Art&Science, Department of Biology, Denizli, Turkey3 Akdeniz University, Faculty of Medicine, Department of Internal Medicine, Division of Oncology, 07070 Antalya, Turkey2005 24 7 2005 5 22 22 5 5 2005 24 7 2005 Copyright © 2005 Akan et al; licensee BioMed Central Ltd.2005Akan 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
Multidrug resistance mediated by the multidrug resistance-associated protein 1 (MRP1) decreases cellular drug accumulation. The exact mechanism of MRP1 involved multidrug resistance has not been clarified yet, though glutathione (GSH) is likely to have a role for the resistance to occur. N-acetylcysteine (NAC) is a pro-glutathione drug. DL-Buthionine (S,R)-sulfoximine (BSO) is an inhibitor of GSH synthesis. The aim of our study was to investigate the effect of NAC and BSO on MRP1-mediated vincristine resistance in Human Embryonic Kidney (HEK293) and its MRP1 transfected 293MRP cells. Human Embryonic Kidney (HEK293) cells were transfected with a plasmid encoding whole MRP1 gene. Both cells were incubated with vincristine in the presence or absence of NAC and/or BSO. The viability of both cells was determined under different incubation conditions. GSH, Glutathione S-Transferase (GST) and glutathione peroxidase (GPx) levels were measured in the cell extracts obtained from both cells incubated with different drugs.
Results
N-acetylcysteine increased the resistance of both cells against vincristine and BSO decreased NAC-enhanced MRP1-mediated vincristine resistance, indicating that induction of MRP1-mediated vincristine resistance depends on GSH. Vincristine decreased cellular GSH concentration and increased GPx activity. Glutathione S-Transferase activity was decreased by NAC.
Conclusion
Our results demonstrate that NAC and BSO have opposite effects in MRP1 mediated vincristine resistance and BSO seems a promising chemotherapy improving agent in MRP1 overexpressing tumor cells.
MRP1vincristineHEK293N-acetylcysteineBSOGSH
==== Body
Background
The acquisition of resistance to anticancer agents used in chemotherapy is the main cause of treatment failure in malignant disorders, provoking tumours to become resistant during treatment, although they initially respond to it [1-4]. Resistance of cancer cells to a single drug is usually accompanied by resistance to other drugs with different structures and cellular targets [3,4]. Identifying the mechanisms leading to intrinsic or acquired multidrug resistance (MDR) is important in developing more effective therapies. At least, two proteins are well-known for causing MDR. Both proteins, the MDR1 gene encoded-Pgp and MRP1 are members of the ATP binding cassette transporter superfamily. Despite their common involvement in MDR, there are clear differences in function and substrate specifity of Pgp and MRP1 [5]. Pgp transports neutral, or positively charged, hydrophobic compounds [5]. In contrast, MRP1 extrudes conjugated organic anions from cells and is known as multispecific aniontransporter (MOAT) [4,6,7]. The exact mechanism of MRP1 involved multidrug resistance remains unknown, although GSH is likely to have a role for the resistance to occur. Thus, clarifying the mechanism of action of MRP1 in cell lines ortumors overexpressing MRP1 and the search for inhibitors of drug transport can give new insights in future experiments and therapies.
Multidrug resistance protein (MRP1) mediated drug resistance occurs against a broad spectrum of natural product drugs like vincristine, although the mechanisms have not been exactly understood and it has not been possible to demonstrate that MRP1 can actively transport unmodified forms of vincristine [8]. Vincristine is a vinca alcaloid type drug and a widely used chemotherapeutic agent for the treatment of acute leukemia and solid tumors [9]. Efflux of hydrophobic natural product anticancer drugs agents such as vincristine from cells expressing MRP1 is thought to require GSH [10,11]. The nature of the involvement of GSH is not fully clarified, though co-transport of GSH is now believed to take place [8,10,12]. GSH is the most abundant non-protein intracellular thiol containing compound that is a key molecule in MRP1-mediated MDR [3,13]. It was shown that ATP-dependent uptake of vincristine by MRP-enriched, inside-out membrane vesicles could be stimulated by physiological concentrations of GSH [14]. It is suggested that increased MRP1 expression without an increase in GSH biosynthesis would not cause any drug resistance in tumor cells, but would result in cell death [15]. GSH conjugates with drugs catalyzed by the enzyme GST and causes their subsequent removal from the cells [15]. BSO inhibits GSH synthesis by irreversible inhibition of γ-glutamyl cysteine synthase and has no other known effect on cells [3,11,16]. N-acetylcysteine is a thiol antioxidant and cysteine source for GSH synthesis [17]. The study aimed to define the mechanism of action of vincristine and the effects of NAC and BSO on MRP1-mediated vincristine resistance in Human Embryonic Kidney (HEK293) and its MRP1 transfected 293MRP cells. For this purpose, HEK293 and 293MRP cells were incubated with vincristine in the presence or absence of NAC and/or BSO. Vincristine cytotoxicity, cell viability and the effect of vincristine on cellular GSH levels, GST and GPx enzyme activities were determined in both cell groups in the presence or absence of NAC at two different concentrations.
Results
Western Blot analysis using monoclonal QCRL-1 anti-MRP1 antibody demonstrated MRP1 expression in 293MRP cells, unlike HEK293 cells (Fig 1).
Figure 1 Western Blot Detection of MRP1 in Human Embryonal Kidney Cell Line Transfected with the MRP1 gene.
Cytotoxic Activity of Vincristine
The experiments were repeated 3 times and the results obtained from these repetitions were averaged. The cytotoxic effect of different concentrations of vincristine on HEK293 cells was shown in Figure 2. The lethal concentration (LD50) of vincristine was found as 0.156 μg/ml on HEK293 cells using crystal violet method (Fig 2). This concentration of vincristine was applied for incubation of the cells.
Figure 2 Cell viability of MRP1 and HEK293 cells against different concentrations of vincristine.
Effects of NAC on vincristine cytotoxicity
The viability of HEK293 and 293MRP cells treated with vincristine was significantly lower than the respective untreated control cells (11.4 ± 2.3% and 52.4 ± 5.2% respectively, p < 0.5) (Fig 3). 293MRP cells were more resistant to vincristine than HEK293 cells. The lowest level in cell viability was observed in HEK293 cells. Both cells were incubated with 1 or 5 mM NAC in the presence of vincristine. N-acetylcysteine supplementation at both concentrations enhanced significantly the resistance of 293MRP and HEK293 cells against vincristine cytotoxicity compared to their respective untreated control cells (p < 0.05). The viability of HEK293 cells increased significantly (p < 0.05) from 11.4 ± 2.3% to 31.1 ± 4.1% with 1 mM NAC and to 37.5 ± 4.7% with 5 mM NAC. The viability of 293MRP cells increased significantly (p < 0.05) from 52.4 ± 5.2% to 57.2 ± 5.4% with 1 mM NAC and to 70.1 ± 6.2% with 5 mM NAC. There was no significant difference between the viability of HEK293 cells treated with two different concentrations of NAC, but 5 mM NAC was more effective in 293MRP cells compared to the 1 mM NAC (p < 0.05).
Figure 3 Effect of N-acetylcysteine (NAC) on vincristine cytotoxicity in human embryonic kidney (HEK293) and 293MRP Cells. * p < 0.05 vs untreated control HEK293 cells ** p < 0.05 vs untreated control 293MRP cells *** p < 0.05 vs HEK293 cells treated with Vinc † p < 0.05 vs HEK293 cells treated with Vinc + 1 mM NAC †† p < 0.05 vs HEK293 cells treated with Vinc + 5 mM NAC ††† p < 0.05 vs 293MRP cells treated with Vinc ‡ p < 0.05 vs 293MRP cells treated with Vinc + 1 mM NAC.
Effect of BSO on vincristine cytotoxicity and survival promoting action of NAC
Cells were pretreated with 100 μM BSO for 24 hour before drug treatments. The viability of 293MRP and HEK293 cells pretreated with BSO was not different significantly (85 ± 5.3% and 93 ± 6.1%, respectively. p > 0.05) (Fig 4). After inhibition of GSH synthesis with BSO, 293MRP cells lost their vincristine resistance significantly from 52.4 ± 5.2% to 19.0 ± 1.9% (p < 0.05) (Fig 4). Pretreatment with BSO didnot affect the viability of HEK293 cells treated with vincristine (11.4 ± 2.3% vs 10.2 ± 1.2%). N-acetylcysteine at both concentrations increased significantly the viability of 293 MRP cells pretreated with BSO against vincristine (from 19.0 ± 1.9% to 33.6 ± 5.4 with 1 mM NAC and to 40.5 ± 6.2% with 5 mM NAC). Similar increase was observed in HEK293 cells under the same conditions (from 10.2 ± 1.2% to 19.2 ± 2.4 with 1 mM NAC and to 29.9 ± 3.2% with 5 mM NAC). Pretreatment with BSO antagonized partly the increases in the viability of both cells caused by treatment with NAC compared to the increases caused by NAC alone (Fig 3 and Fig 4). In other words, NAC increased less the viability of both cells pretreated with BSO than the cells treated with only NAC against vincristine.
Figure 4 Effect of DL-Buthionine (S,R)-sulfoximine (BSO) on vincristine cytotoxicity and survival promoting effect of N-acetylcysteine (NAC) in human embryonic kidney (HEK293) and 293MRP Cells. * p < 0.05 vs HEK293 cells pretreated with BSO ** p < 0.05 vs 293MRP cells pretreated with BSO *** p < 0.05 vs HEK293 cells treated with Vinc † p < 0.05 vs 293MRP cells treated with Vinc †† p < 0.05 vs HEK293 cells treated with BSO+Vinc and HEK293 cells treated with Vinc ††† p < 0.05 vs 293MRP cells treated with BSO+Vinc ‡ p < 0.05 vs 293MRP cells treated with BSO+Vinc+1 mM NAC ‡‡ p < 0.05 vs HEK293 cells treated with BSO+Vinc+1 mM NAC.
Effect of vincristine and NAC on cellular GSH concentrations
Cellular GSH levels were measured after 24 and 48 hour vincristine treatments in the presence or absence of NAC at two different concentrations. Cellular GSH concentrations were not different in untreated control HEK293 and 293MRP cells (80.2 ± 4.6 μg mg-1 protein and 84.6 ± 4.9 μg mg-1 protein, respectively) (Fig 5). Glutathione levels decreased not significantly after 24- and 48-h vincristine treatments in HEK293 cells from 80.2 ± 4.6 μg mg-1 protein to 75.9 ± 3.8 μg mg-1 protein and to 72.2 ± 3.5 μg mg-1 protein (p > 0.05) and in 293MRP cells from 86 ± 4.9 μg mg-1 protein to 72.4 ± 3.4 μg mg-1 protein and to 64.9 ± 3.2 μg mg-1 protein (p < 0.05). N-acetylcysteine at both concentrations caused a significant increase in GSH concentrations in both cells treated with vincristine for both incubation times, in comparison to untreated control cell lines and cells treated with only vincristine (p < 0.05) (Fig 5).
Figure 5 Effects of Vincristine and N-acetylcysteine (NAC) on intracellular glutathione (GSH) levels in human embryonic kidney (HEK293) and 293MRP Cells. * p < 0.05 vs untreated control 293MRP cells ** p < 0.05 vs untreated control HEK293 cells and HEK293 cells treated with Vinc for 24 and 48 hour *** p < 0.05 vs untreated control 293MRP cells and 293MRP cells treated with Vinc for 24 and 48 hour † p < 0.05 vs HEK293 cells treated with Vinc+1 mM NAC for 24 and 48 hour †† p < 0.05 vs 293MRP cells treated with Vinc+1 mM NAC for 24 and 48 hour.
Effect of vincristine and NAC on the activity of cellular Glutathione S-Transferase and Glutathione Peroxidase
Glutathione S-Transferase activity in both untreated control cell lines was not significantly different. 48 hour, but not 24 hour incubation with vincristine significantly increased the GST activity in both cell lines comparing to the corresponding untreated control cells (Fig 6). N-acetylcysteine (1 mM) for 48 hour caused a significant decrease in GST activity in 293MRP and HEK293 cells compared with nontreated control cells and cells treated with only vincristine for both incubation times. 5 mM NAC at both incubation times caused a significant decrease in the GST activity in both cell lines compared to nontreated control cells, cells treated with vincristine only and cells treated with vincristine + NAC (1 mM) for both incubation times (Fig 6).
Figure 6 Effects of Vincristine and N-acetylcysteine (NAC) on Glutathione S-transferase Activity (GST) in human embryonic kidney (HEK293) and 293MRP Cells. * p < 0.05 vs untreated control HEK293 cells ** p < 0.05 vs untreated control 293MRP cells *** p < 0.05 vs untreated control HEK293 cells and HEK293 cells treated with Vinc for 24 and 48 hour † p < 0.05 vs untreated control 293MRP cells and 293MRP cells treated with Vinc for 24 and 48 hour †† p < 0.05 vs untreated control HEK293 cells and HEK293 cells treated with Vinc and HEK293 cells treated with Vinc+1 mM NAC for 24 and 48 hour ††† p < 0.05 vs untreated control 293MRP cells and 293MRP cells treated with Vinc and 293MRP cells treated with Vinc+1 mM NAC for 24 and 48 hour.
There was no significant difference in GPx activity between HEK293 and 293MRP cells (2.4 ± 0.2 IU mg-1 protein and 2.2 ± 0.1 IU mg-1 protein, respectively) (Fig 7). Non-significant increases in GPx activity were observed after vincristine treatment for both incubation times in HEK293 (2.8 ± 0.3 IU mg-1 protein for 24 h vincristine incubation and 3.1 ± 0.3 IU mg-1 protein for 48 h vincristine incubation) and 293MRP cells (2.6 ± 0.2 IU mg-1 protein for 24 h vincristine incubation and 3.0 ± 0.3 IU mg-1 protein for 48 h vincristine incubation) compared to untreated control cells (p > 0.05). N-acetylcysteine (1 mM) incubation for 24 and 48 hours increased GPx activity in both cell lines compared to untreated control cells. 5 mM NAC incubation for 24 and 48 hours increased the GPx activity significantly in HEK293 and 293MRP cells compared to the other cell groups (p < 0.05) (Fig 7).
Figure 7 Effects of Vincristineand N-acetylcysteine (NAC) on Glutathione Peroxidase (GPx) Activity in human embryonic kidney (HEK293) and 293MRP Cells. * p < 0.05 vs untreated control HEK293 cells ** p < 0.05 vs untreated control 293MRP cells *** p < 0.05 vs untreated control HEK293 cells and HEK293 cells treated with Vinc and HEK293 cells treated with Vinc+1 mM NAC for 24 and 48 hour † p < 0.05 vs untreated control 293MRP cells and 293MRP cells treated with Vinc and 293MRP cells treated with Vinc+1 mM NAC for 24 and 48 hour.
Discussion
In our experiments, the viability of MRP1 transfected cells (293MRP) treated with vincristine was higher than human embryonic kidney (HEK293). Our results are in accordance with O'Brien and co-workers who reported that MRP1 confers resistance to doxorubicin, etoposide, and vincristine in NIH 3T3 fibroblast cell line [18]. Our experiments with N-acetylcysteine (NAC) and DL-Buthionine (S,R)-sulfoximine (BSO) showed that MRP1 mediated vincristine resistance largely depends on GSH and this is in accordance with previous data [3,18,19]. N-acetylcysteine supplementation at both concentrations increased the survival rate of vincristine treated HEK293 and 293MRP cells which had increased GSH levels confirming that the viability depends on the level of GSH (Fig 3 and Fig 4). After inhibition of GSH synthesis with BSO, 293MRP cells lost their vincristine resistance (Fig 4). Similar results were described previously in different cell lines overexpressing MRP1 [3,16].
We compared the viability of both cells treated with vincristine and NAC in the presence and absence of BSO. N-acetylcysteine at both concentrations increased significantly the viability of 293MRP and HEK293 cells pretreated with BSO against vincristine, but these increases were lower in comparison to the corresponding cells untreated with BSO (Fig 3 and Fig 4). Pretreatment with BSO antagonized partly the increases in the viability of both cells caused by treatment with NAC compared to the increases caused by NAC alone. In other words, NAC increased less the viability of both cells pretreated with BSO than the cells treated with only NAC against vincristine. This might be explained that BSO counterbalances the effect of NAC as a precursor of GSH. This is another proof that survival promoting action of NAC depends on GSH synthesis and is in accordance with our previous findings with doxorubicin [1].
In our experiments, cellular GSH concentration decreased after vincristine treatment which might be due to GSH efflux (Fig 5). Enhanced GSH efflux has been reported in MRP1 expressing cells and this enhanced efflux can be inhibited by indomethacin and probenecid [10]. They suggested that changes in the concentrations of GSH and its oxidised form GS-SG inside cells may each influence MRP1-mediated anion transport. Furthermore, hypoxia or oxidative stress may cause depletion of glutathione (GSH). Increased oxidative stress has been reported to associate tumorigenesis [20,21] and this may play a role in GSH depletion [22,23]. which in turn may affect efflux of drugs. The higher GPx activity in vincristine treated cells might be a compensatory effect of cells against depletion of GSH (Fig 6).
It has been hypothesized that vincristine resistance of myeloblasts is related to its degradation by myeloperoxidase (MPO) [9]. Myeloperoxidase (MPO) catalyzes the formation of HOCl from H2O2 and chloride ion. It was shown oxidation by HOCl is the final step in vincristine degradation in both a cell free system and in cultures leukemic cell lines. Oxidation of anti-neoplastic drugs may cause a reduction in efficacy or an increase in toxicity. This could lead to a decrease in the therapeutic index. Inhibition of MPO in these different disease states could eliminate this intra- and extracellular oxidation pathway and could effectively increase the therapeutic index.
The identification of MRP1 as an important glutathione-conjugate efflux pump raises the possibility that MRP1 and GST may act in synergy to confer cellular resistance to some of these compounds [3,14,24,25]. It is not clear yet if glutathione is either co-transported as a GS-conjugate with vincristine or activates MRP1 for vincristine transport [4]. Studies so far showed conjugation with GSH and extrusion are not the major pathway [4]. Co-expression with MRP1 of any of the human GST isozymes A1-1, M1-1, or P1-1 failed to augment MRP1-associated resistance to drugs including doxorubicin, vincristine, etoposide, and mitoxantrone [4,24]. This might be an evidence that vincristine is not conjugated with GSH, but co-transported with GSH in MRP1 mediated drug resistance.
In our study, NAC supplementation decreased GST activity level in both cell lines (Fig 7). This might be explained that NAC may spontaneously form conjugates with vincristine, therefore decreasing the need for GST activity for conjugation. Although, it is not clear whether NAC spontaneously conjugates with vincristine, it is known that mercapturic acids (N-acetylcysteine S-conjugates) are spontaneously formed, released into the circulation and delivered to the kidney for excretion in urine [26-28]. Similarly, Weigand et al reported attenuation of GST activity after NAC supplementation [29].
Conclusion
Our results demonstrate that NAC enhances MRP1-mediated vincristine resistance and this effect depends on GSH synthesis. DL-Buthionine (S,R)-sulfoximine seems a promising chemotherapy improving agent in MRP1 overexpressing tumor cells. This finding might be relevant and have an implication in cancer patients undergoing chemotherapy.
Methods
Materials
Dulbeccos' Modified Eagles' Medium (DMEM), NAC, BSO, geneticin, Feotal Bovine Serum (FBS), and other chemicals were purchased from Sigma-Aldrich Corp. St. Louis, MO, USA. Vincristine was obtained from Oncology Department of Akdeniz University Hospital. The plasmid (pcDNA3.1/MRPK) encoding the whole MRP1 gene was kindly provided by Dr. Susan Cole from Oueen's University, Ontario Canada. Protein Assay Kit was purchased from Bio-Rad Laboratories Ltd., Herts, UK. Monoclonal anti-MRP1 QCRL-1 antibody was obtained from Centocor Inc., Malvern, PA, USA. Horseradish peroxidase (HRP) conjugated seconder goat-anti mouse antibody was purchased from Santa-Cruz Biotechnology Inc., Santa Cruz, CA, USA.
Cell lines
Human embryonic kidney cell line, HEK293, was grown in DMEM, supplemented with 10% heat inactivated FBS, 2 mM L-glutamin, and 1% antibiotic-antimycotic solution. Cell cultures were kept at 37°C in a humid atmosphere containing 5% CO2.
Transfection
Cells (1 × 106 cells in 100 mm dish) were transfected with the plasmid (pcDNA3.1/MRPK) encoding the whole MRP1 gene. The transfection was made according to the calcium phosphate transfection method [30]. Sixteen hours after the transfection, the cells were feeded with DMEM supplemented with 400 μg/ml geneticin.
Preparation of Membrane Enriched Fractions and Immunoblotting
For immunoblotting of the 293MRP and HEK293 cells, membrane enriched fractions were prepared according to Grant et al [31]. Briefly, cell pellet was resuspended in the collection buffer (10 mM Tris-HCL, pH 7.4, 10 mM KCl, 1.5 mM MgCl2 and protease inhibitors), homogenized on ice in a Potter-Elvejhem tissue homogenizer. The intact cells and nuclei were removed by centrifugation at 800 g at +4°C, and the supernatant was further centrifuged at 100 000 g at +4°C for 20 minutes to prepare the membrane enriched fractions. The pellet was resuspended in buffer (10 mM Tris-HCl pH 7.4, 125 mM sucrose, and protease inhibitors). The protein suspension was mixed with solubilizing buffer (4 M urea, 0.5% Sodiumdodecylsulphate (SDS), and 50 mM dithiotreitol) and equal amounts of proteins were subjected to SDS-PAGE (SDS-Polyacrylamide gel electrophoresis) on 7% polyacrylamide gels, then transferred onto nitrocellulose sheet for overnight at 40 volt, and analysed by immunoblotting with anti-MRP1 monoclonal QCRL-1 antibody.
Cell Viability Assays
Cell viability was assayed using the crystal violet method [32]. 3 × 104 cells were seeded in a 96 well microplate. After 24 hours, cells were incubated with 0.06–1000 μg/ml vincristine for 72 hours. LD50 for vincristine was determined to be 0.156 μg/ml and this dose of vincristine was used in the rest of the experiments. Both HEK293 and 293MRP cells were incubated with vincristine (0.156 μg/ml) in the presence or absence of NAC (1 and 5 mM) for 72 hours at 37°C in a humid atmosphere containing 5% CO2. At the end of incubation period, the medium was replaced by 0.5% crystal-violet (w/v; in 50% methanol) solution. Plates were incubated for 10 min at room temperature, washed with water and adsorbed dye was eluted out with Na-citrate (0.1 M Na-citrate in 50% ethanol, pH 4.2). Absorbance, which was proportional to cell viability, was measured at a wavelength of 600 nm. Cell viability was monitored as the percentage of viable cells comparing to control, untreated cells. For BSO experiments, cells were pretreated with 100 μM BSO for 24 hours before incubating them with vincristine with or without NAC for 72 hours as described above.
Preparation of Cell Extract for GSH and Enzyme Measurements
Cell extracts were prepared as described by Bravard et al with a slight modification [33]. 106 cells were seeded in a cell culture dish and incubated for 24 hours. After incubation with vincristine in the presence or absence of NAC for 24 or 48 hours, medium was discharged and cells were washed with phosphate buffered saline (PBS), collected in potassium phosphate buffer (50 mM, pH 7.4) with cell scraper and repeatedly freezed and thawed in liquid nitrogen for four times, and then centrifuged at 10 000 g for 10 minutes at 4°C. The supernatant was used for enzyme activities and GSH measurements. Protein concentrations were determined using Bio-Rad protein assay kit. All measurements were adjusted by dividing with the protein content of each sample.
Reduced Glutathione Assay
Cellular GSH concentrations were determined as described by Virgil et al [34]. Briefly, the supernatant was deproteinized and GSH content was monitored spectrophotometrically with 5-5' dithiobis(2-nitrobenzoic acid) (DTNB) at a wavelength of 412 nm. The GSH concentration was evaluated using a standart curve of known amounts of GSH. Results are expressed as μg/mg protein.
Glutathione S-Transferase Activity Assay
Glutathione S-Transferase (GST) activity was measured at 340 nm wavelength in the presence of 1-cloro-2,4-dinitrobenzene (CDNB), GSH and sodium phosphate buffer (pH 6.5) at 30°C for 6 minutes [33,35]. Results are expressed as ΔOD / mg protein.
Glutathione Peroxidase Activity Assay
Glutathione Peroxidase (GPx) activity was determined using a modification of the method of Paglia and Valentine [36]. In a cuvette kept at 37°C, GPx activity was monitored at 340 nm by the absorbance of nicotinamide adenine dinucleotide phosphate (NADPH) for 3 minutes in the presence of glutathione reductase (0.5 IU), EDTA (0.3 mM) and t-buthyl hydroperoxide (0.4 mM). Results are expressed as IU/mg protein.
Statistical Analysis
Statistical analysis was performed using Anova test with SPSS packed program for Windows version 10.0 (SPSS Inc., Chicago, IL, USA). All the experiments were repeated three times. Mean values and standard deviations (mean ± S.D.) were calculated for every variable in each cell group and were compared between the groups. p < 0.05 was selected as statistically significant.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
IA, SA and HA carried out the cell culture studies, transfection, immunoblotting, viability assays, GSH and enzyme measurements in the cell extracts. BS and TO participated in the design of the study and performed the statistical analysis, coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This study was supported by Akdeniz University Research Fund.
==== Refs
Akan I Akan S Akca H Savas B Ozben T N-acetylcysteine enhances multidrug resistance-associated protein 1 mediated doxorubicin resistance Eur J Clin Invest 2004 34 683 689 15473893 10.1111/j.1365-2362.2004.01411.x
Ruiz-Gomez MJ Souviron A Martinez-Morillo M Gil L P-glycoprotein, glutathione and glutathione S-transferase increase in a colon carcinoma cell line by colchicine J Physiol Biochem 2000 56 307 312 11321524
Muller M Meijer C Zaman GJ Borst P Scheper RJ Mulder NH de Vries EG Jansen PL Overexpression of the gene encoding the multidrug resistance-associated protein results in increased ATP-dependent glutathione S-conjugate transport Proc Natl Acad Sci USA 1994 91 13033 13037 7809167
Hipfner DR Deeley RG Cole SP Structural, mechanistic and clinical aspects of MRP1 Biochim Biophys Acta 1999 1461 359 376 10581367
de Jong MC Slootstra JW Scheffer GL Schroeijers AB Puijk WC Dinkelberg R Kool M Broxterman HJ Meloen RH Scheper RJ Peptide transport by the multidrug resistance protein MRP1 Cancer Res 2001 61 2552 2557 11289130
Chuman Y Sumizawa T Takebayashi Y Niwa K Yamada K Haraguchi M Furukawa T Akiyama S Aikou T Expression of the multidrug-resistance-associated protein (MRP) gene in human colorectal, gastric and non-small-cell lung carcinomas Int J Cancer 1996 66 274 279 8603824 10.1002/(SICI)1097-0215(19960410)66:2<274::AID-IJC23>3.0.CO;2-B
Nooter K Westerman AM Flens MJ Zaman GJ Scheper RJ van Wingerden KE Burger H Oostrum R Boersma T Sonneveld P Expression of the multidrug resistance-associated protein (MRP) gene in human cancers Clin Cancer Res 1995 1 1301 1310 9815925
Loe DW Deeley RG Cole SP Characterization of vincristine transport by the M(r) 190,000 multidrug resistance protein (MRP): evidence for cotransport with reduced glutathione Cancer Res 1998 58 5130 5136 9823323
Ozgen U Turkoz Y Stout M Ozugurlu F Pelik F Bulut Y Aslan M Ravindranath Y Savasan S Degradation of vincristine by myeloperoxidase and hypochlorous acid in children with acute lymphoblastic leukemia Leuk Res 2003 27 1109 1113 12921949 10.1016/S0145-2126(03)00098-5
Bagrij T Klokouzas A Hladky SB Barrand MA Influences of glutathione on anionic substrate efflux in tumour cells expressing the multidrug resistance-associated protein, MRP1 Biochem Pharmacol 2001 62 199 206 11389878 10.1016/S0006-2952(01)00660-8
Zaman GJ Lankelma J van Tellingen O Beijnen J Dekker H Paulusma C Oude Elferink RP Baas F Borst P Role of glutathione in the export of compounds from cells by the multidrug-resistance-associated protein Proc Natl Acad Sci USA 1995 92 7690 7694 7644478
Rappa G Lorico A Flavell RA Sartorelli AC Evidence that the multidrug resistance protein (MRP) functions as a co-transporter of glutathione and natural product toxins Cancer Res 1997 57 5232 5237 9393740
Kearns PR Hall AG Glutathione and the response of malignant cells to chemotherapy Drug Discovery Today 1998 3 113 121 10.1016/S1359-6446(97)01156-2
Loe DW Almquist KC Deeley RG Cole SP Multidrug resistance protein (MRP)-mediated transport of leukotriene C4 and chemotherapeutic agents in membrane vesicles. Demonstration of glutathione-dependent vincristine transport J Biol Chem 1996 271 9675 9682 8621643 10.1074/jbc.271.44.27782
Yan CY Ferrari G Greene LA N-acetylcysteine-promoted survival of PC12 cells is glutathione-independent but transcription-dependent J Biol Chem 1995 270 26827 26832 7592924 10.1074/jbc.270.45.26827
Rappa G Gamcsik MP Mitina RL Baum C Fodstad O Lorico A Retroviral transfer of MRP1 and gamma-glutamyl cysteine synthetase modulates cell sensitivity to L-buthionine-S,R-sulphoximine (BSO): new rationale for the use of BSO in cancer therapy Eur J Cancer 2003 39 120 128 12504668 10.1016/S0959-8049(02)00447-1
Erkkila K Hirvonen V Wuokko E Parvinen M Dunkel L N-acetyl-L-cysteine inhibits apoptosis in human male germ cells in vitro J Clin Endocrinol Metab 1998 83 2523 2531 9661638 10.1210/jc.83.7.2523
O'Brien ML Vulevic B Freer S Boyd J Shen H Tew KD Glutathione peptidomimetic drug modulator of multidrug resistance-associated protein J Pharmacol Exp Ther 1999 291 1348 1355 10565860
Zaman GJ Flens MJ van Leusden MR de Haas M Mulder HS Lankelma J Pinedo HM Scheper RJ Baas F Broxterman HJ The human multidrug resistance-associated protein MRP is a plasma membrane drug-efflux pump Proc Natl Acad Sci USA 1994 91 8822 8826 7916458
Mori K Shibanuma M Nose K Invasive potential induced under long-term oxidative stress in mammary epithelial cells Cancer Res 2004 64 7464 7472 15492271
Ishii T Yasuda K Akatsuka A Hino O Hartman PS Ishii N A mutation in the SDHC gene of complex II increases oxidative stress, resulting in apoptosis and tumorigenesis Cancer Res 2005 65 203 209 15665296
Kalivendi SV Kotamraju S Zhao H Joseph J Kalyanaraman B Doxorubicin-induced apoptosis is associated with increased transcription of endothelial nitric-oxide synthase. Effect of antiapoptotic antioxidants and calcium J Biol Chem 2001 276 47266 47276 11579094 10.1074/jbc.M106829200
Gouaze V Mirault ME Carpentier S Salvayre R Levade T Andrieu Abadie N Glutathione peroxidase-1 overexpression prevents ceramide production and partially inhibits apoptosis in doxorubicin-treated human breast carcinoma cells Mol Pharmacol 2001 60 488 496 11502879
Morrow CS Smitherman PK Diah SK Schneider E Townsend AJ Coordinated action of glutathione S-transferases (GSTs) and multidrug resistance protein 1 (MRP1) in antineoplastic drug detoxification. Mechanism of GST A1-1- and MRP1-associated resistance to chlorambucil in MCF7 breast carcinoma cells J Biol Chem 1998 273 20114 20120 9685354 10.1074/jbc.273.32.20114
Morrow CS Diah S Smitherman PK Schneider E Townsend AJ Multidrug resistance protein and glutathione S-transferase P1-1 act in synergy to confer protection from 4-nitroquinoline 1-oxide toxicity Carcinogenesis 1998 19 109 115 9472701 10.1093/carcin/19.1.109
Aizawa S Ookawa K Kudo T Asano J Hayakari M Tsuchida S Characterization of cell death induced by ethacrynic acid in a human colon cancer cell line DLD-1 and suppression by N-acetyl-L-cysteine Cancer Sci 2003 94 886 893 14556662
De Flora S D'Agostini F Masiello L Giunciuglio D Albini A Synergism between N-acetylcysteine and doxorubicin in the prevention of tumorigenicity and metastasis in murine models Int J Cancer 1996 67 842 848 8824557 10.1002/(SICI)1097-0215(19960917)67:6<842::AID-IJC14>3.0.CO;2-3
James LP Mayeux PR Hinson JA Acetaminophen-induced hepatotoxicity Drug Metab Dispos 2003 31 1499 1506 14625346 10.1124/dmd.31.12.1499
Weigand MA Plachky J Thies JC Spies-Martin D Otto G Martin E Bardenheuer HJ N-acetylcysteine attenuates the increase in alpha-glutathione S-transferase and circulating ICAM-1 and VCAM-1 after reperfusion in humans undergoing liver transplantation Transplantation 2001 72 694 698 11544433 10.1097/00007890-200108270-00023
Chen C Okayama H High-efficiency transformation of mammalian cells by plasmid DNA Mol Cell Biol 1987 7 2745 2752 3670292
Grant CE Valdimarsson G Hipfner DR Almquist KC Cole SP Deeley RG Overexpression of multidrug resistance-associated protein (MRP) increases resistance to natural product drugs Cancer Res 1994 54 357 361 8275468
Akca H Akan SY Yanikoglu A Ozes ON Suppression of TNF-alpha mediated apoptosis by EGF in TNF-alpha sensitive human cervical carcinoma cell line Growth Factors 2003 21 31 39 12795334 10.1080/0897719031000115369
Bravard A Petridis F Luccioni C Modulation of antioxidant enzymes p21WAF1 and p53 expression during proliferation and differentiation of human melanoma cell lines Free Radic Biol Med 1999 26 1027 1033 10232848 10.1016/S0891-5849(98)00299-8
Virgil F Fairbanks VF Klee GG Tietz NW Biochemical aspects of hematology Textbook of clinical chemistry 3 Philadelphia: WB eds Saunders Company 1642 1711
Cohen AJ Smith JN Tubert H Comparative doxication. The enzymatic conjugation of chloro compounds with glutathione in locusts and other insects Biochem J 1964 90 457 464 5833356
Paglia DE Valentine WN Studies on the quantitative and qualitative characterization of erythrocyte glutathione peroxidase J Lab Clin Med 1967 70 158 169 6066618
|
16042792
|
PMC1183228
|
CC BY
|
2021-01-04 16:24:16
|
no
|
Cancer Cell Int. 2005 Jul 24; 5:22
|
utf-8
|
Cancer Cell Int
| 2,005 |
10.1186/1475-2867-5-22
|
oa_comm
|
==== Front
Cardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 1475-2840-4-101599847110.1186/1475-2840-4-10Original InvestigationSerum resistin is associated with C-reactive protein and LDL- cholesterol in type 2 diabetes and coronary artery disease in a Saudi population Al-Daghri N [email protected] R [email protected] PG [email protected] K [email protected] O [email protected] AF [email protected] S [email protected] King Saud University College of Science, Biochemistry Department, Riyadh, Saudi Arabia2 Birmingham Heartland Hospital, Clinical Biochemistry, Birmingham B9 5SS, UK3 King Saud University, College of Medicine, Medicine Department, Riyadh, Saudi Arabia4 University of Warwick, Warwick Medical School, Diabetes & Metabolism Unit, Coventry, CV4 7AL, UK2005 5 7 2005 4 10 10 18 2 2005 5 7 2005 Copyright © 2005 Al-Daghri et al; licensee BioMed Central Ltd.2005Al-Daghri 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.
Aims
Resistin is an adipocyte-derived factor implicated in obesity-associated type 2 diabetes (T2DM). This study examines the association between human serum resistin, T2DM and coronary heart disease.
Methods
One hundred and fourteen Saudi Arabian patients (male: female ratio 46:68; age 51.4 (mean ± SD)11.7 years; median and range: 45.59 (11.7) years and BMI: 27.1 (mean ± SD) 8.1 Kgm2 median and range: 30.3 (6.3) were studied. Serum resistin and C-reactive protein (CRP), a marker of inflammation CRP levels, were measured in all subjects. (35 patients had type 2 diabetes mellitus (T2DM); 22 patients had coronary heart disease (CHD).
Results
Serum resistin levels were 1.2-fold higher in type 2 diabetes and 1.3-fold higher in CHD than in controls (p = 0.01). In addition, CRP was significantly increased in both T2DM and CHD patients (p = 0.007 and p = 0.002 respectively). The use of regression analysis also determined that serum resistin correlated with CRP levels (p = 0.04, R2 0.045).
Conclusion
The findings from this study further implicate resistin as a circulating protein associated with T2DM and CHD. In addition this study also demonstrates an association between resistin and CRP, a marker of inflammation in type 2 diabetic patients.
type 2 diabetesCoronary Artery diseaseresistinC-reactive protein
==== Body
Introduction
Chronic sub-clinical inflammation has been identified as an important mechanism in the pathogenesis of T2DM and associated cardiovascular complications [1]. Obesity, a major risk factor for T2DM and cardiovascular disease, is now also established as a state of chronic inflammation [2]. For this reason, adipose tissue may represent an important site for the production of pro-inflammatory cytokines and acute phase reactants [3]. Such pro-inflammatory factors include tumour necrosis factor alpha (TNF-α), interleukin 6 (IL-6) and leptin, which are secreted by adipocytes [4-7] and are elevated in obesity and insulin resistant states. This may therefore suggest the adipocyte as a site also for an initial inflammatory response, therefore increasing the overlap between metabolic and inflammatory signalling pathways.
The adipocytokine resistin which belongs to a family of cysteine-rich C-terminal proteins known as resistin-like molecules (RELM; RELMα/FIZZ 1 and RELMβ/FIZZ 2) of FIZZ (found in inflammatory zone) are thought to be involved in inflammatory processes [8-11]. Previous studies have however highlighted that in mice, resistin impairs glucose tolerance and insulin action. In addition, resistin also inhibits adipogenesis in murine 3T3-L1 cells [8,11-14]. Therefore resistin has also been proposed as an adipocyte-secreted factor thought to link obesity and T2DM [9], although subsequent rodent studies have reported contrasting findings to this [15]. In addition, the role of resistin in human obesity has also produced conflicting reports, and at present resistin still remains controversial as a potential mediator in the pathogenesis of T2DM and the impact of resistin on T2DM status [16-21].
Our previous studies have highlighted that serum resistin levels are increased in Caucasian T2DM subjects and reduced with modest weight loss [18,22] and these factors are also associated with changes in C-reactive protein (CRP) levels, CRP represents one of the acute phase proteins which increase during systemic inflammation. CRP is a known inflammatory marker for T2DM and coronary heart disease (CHD) [23]. Our previous cohort studies have also determined that with T2DM patients CRP represents an independent predictor of serum resistin levels [18]. Furthermore, it is apparent that ethnicity can affect the relative risk of T2DM and cardiovascular complications with South Asians, Arabs and Afro-Asians, with T2DM having a higher adverse risk profile [24-29]. However, to date the relationship between serum resistin levels, and T2DM or CHD has not been investigated in a Saudi Arabian population. Therefore the aims of this study were to investigate for the first time whether serum resistin levels are increased in Saudi Arabians with T2DM and if this is affected by CHD. Furthermore these studies seek to ascertain whether CRP is correlated with serum resistin levels in a Saudi Arabian population.
Materials and methods
Subjects
The type 2 diabetic patients for this study were randomly selected from the Diabetes Center and non-diabetic case control patients were obtained from out-patients of King Abdulaziz University Hospital at King Saud University in Riyadh, Saudi Arabia. Coronary heart disease (CHD) patients were selected from the Coronary Care units at King Khalid University Hospital and Prince Sultan Cardiac Units in Riyadh. From these, 114 Saudi Arabian patients were studied (Table 1). CHD patients (n = 22) were classified as having coronary heart disease (CHD) if they had suffered any of the following: a myocardial infarction, angina with a positive exercise tolerance test, a resting ECG consistent with CHD or angiographically proven coronary artery disease. T2DM patients (n = 35), diagnosed according to WHO criteria, but with no clinical evidence of CHD were screened with ECG, clinical symptoms and ± thallium cardiac scanning or cathertization. All patients with T2DM were treated by diet with or without oral hypoglycemic agents, mainly glibenclamide and metformin. Finally 50 Saudi individuals with no prior history of CHD or T2DM were recruited as healthy case controls from general practice. Ethical approval was obtained from the local institution's review committee.
Table 1 Clinical Characteristics of all patients
Case Control T2DM CHD
N 50 35 22
Age (years) 42.5(10.0) 53.6(11.7)a 47.6(11.9)
Diabetes duration (years) 0 6.8(6.3) 0
Systolic BP (mm/Hg) 117.7(16.7) 124.1(25.2) 135.0(20.7)
Diastolic BP (mm/Hg) 76.8(11.6) 78.8(14.5) 78.8(14.6)
BMI (KG/M2) 29.7(6.3) 27.5(6.9) 33.7(15.9)
Waist (cm) 94.1(15.0) 92.7(9.5) 91.4(12.2)
Hips (cm) 105.7(17.1) 95.0(15.6) 98.6(12.2)
Data are presented as mean ± (SD). Clinical characteristics for case control versus either type 2 diabetic or CHD patients were assessed (p-value: a = p < 0.01).
All patients underwent a full physical examination and completed a general questionnaire. Body mass index (BMI), blood pressure, smoking habits and age were recorded. Blood samples were collected after a 12-hr fast for the determination of total cholesterol, HDL, triglycerides, insulin, apolipoprotein AI (Apo AI) and apolipoprotein AII (Apo AII), leptin and serum resistin concentrations. These studies were conducted with the approval of the King Abdulaziz Diabetic Center with approval obtained from the ethics committee and the individual consent of each patient.
Assays
Plasma samples were stored at -70°C prior to analysis. Measurements of serum cholesterol, HDL, and triglycerides, Apo AI and ApoAII were performed using routine laboratory methods. Insulin was analysed by a solid phase enzyme amplified sensitivity immunoassay (Medgenix INS-ELISA, Biosource, Belgium). Insulin resistance (HOMA IR) and β-cell function index were derived using the HOMA equation [27]. Leptin concentrations were measured by radio-immunoassay (Linco Research, St. Charles, MO). CRP was analysed by immunoturbidimetric method (Roche/ Hitachi analyser). Serum resistin was measured using the human resistin ELISA assay from Phoenix Pharmaceuticals (Belmont, California, USA; intra-assay variability <5%). The commercially available resistin ELISA had been validated as part of previous studies [15].
Statistical Analysis
Data were stored and analysed using the SPSS version 10 package (SPSS, Evanston, IL, USA) for Windows. Biochemical parameters not normally distributed were analysed after being logarithmically transformed. A student's unpaired t-test or one-way ANOVA exposed the differences between the groups. Simple and partial correlation coefficients between the variables were determined and multiple regression analysis was performed to determine the relationships between the variables of interest. Data were expressed as mean and standard deviation (SD) or median and range; statistical significance was accepted at p < 0.05.
Results
Clinical characteristics
The full clinical and biochemical characteristics for the three groups are set out in Tables 1 and 2 respectively. The control subjects were noted to be younger than the T2DM patient counterparts (p < 0.01). However, there were no significant differences in age or blood pressure, Body mass index (BMI) or waist circumstance between case controls and CHD patients.
Metabolic Characteristics
Serum resistin was significantly higher in T2DM 22.6(± 4.5)ng/mL and CHD patients 24.2(± 3.6) ng/mL compared with case controls 18.9(± 3.4) ng/mL (p-value <0.0001). CRP was significantly increased in T2DM 8.2 (0.7–27.1) μg/mL median (range) and CHD patients 7.3 (1.8–18.3) μg/mL(p = 0.007 and p = 0.002 respectively) compared with case controls 4.9(0.2–18.9) μg/mL (Table 2). Regression analysis determined that serum resistin was associated with serum CRP levels (Figure 1).
Table 2 Metabolic Characteristics of all patients
Case Control T2DM CHD
N 50 35 22
Fasting Glucose mmol/L 5.4(1.4) 15.4(5.4)a 5.6(0.6)
Cholesterol mmol/L 5.3(1.3) 6.9(1.9)b 7.7(2.7)c
HDL mmol/L 0.9(0.4) 0.88(0.3) 0.92(0.4)
Triglyceride mmol/L 1.8(0.6) 2.5(1.1) 2.1(1.2)
LDL mmol/L 3.9(1.4) 6.1(2.9)d 4.9(1.9)e
Insulin (Range) μmol/L* 7.5(5–12) 9.0(9–16) 7.0(1–49)
Leptin (Range)ng/mL* 7.2(2–30) 5.8(1.9–28) 3.7(0.7–8.9)
Apo1 mg/dL 80.1(15.6) 70.8.(23.4)c 74.1(24.5)
Apo11 mg/dL 48.7(13.5) 45.9(13.6)f 46.3(18.9)c
HOMAIR (Range)* 1.8(1.3–3.2) 8.2(0.8–49.7) 2.0(0.5–5.6)
Resistin ng/mL 18.9(3.4) 22.6(4.5)e 24.2(3.2)d
CRP (Range) μg/mL* 4.9(0.2–18.9) 8.2(0.7–27.1)g 7.3(1.8–18.3)f
Data are presented as means (SD) or as median (interquartile range) for insulin, leptin HOMIR and CRP data. P-values shown are comparison of metabolic characteristics for case control versus either type 2 diabetic (T2DM) or coronary heart disease (CHD) patients as highlighted: a = p < 0.0001, b = p < 0.05, c = p < 0.003, d = p < 0.017, e = p < 0.01, f = p < 0.002, g = p < 0.007.
Figure 1 The correlation of Serum resistin levels (ng/mL) with C-reactive protein (mg/mL) using univariate analysis across all Saudi subjects, case controls, T2DM and CHD subjects. Black denotes the actual values with the line of best fit shown in square (R2 0.045 p-value = 0.04).
Fasting plasma glucose 15.4(5.4) serum cholesterol 15.4(5.4) mmol/L, LDL6.1(2.9) mmol/L and Apo-AI 70.8 (23.4)mg/dL were higher in T2DM patients (p < 0.0001, p < 0.05, p < 0.017, p < 0.003, respectively) compared with case control subjects, while the Apo-AII 45.9 (13.6)mg/dL was lower in the T2DM than in the controls, where p < 0.002 (Table 2). In CHD, cholesterol 7.7 (2.7), and LDL 4.9(1.9) mmol/L were significantly higher than in the case controls (p = 0.003, p < 0.01 respectively; while the Apo-AII 46.3(18.9) mg/dL in CHD patients as lower than in the case controls, where p = 0.003 (Table 2).
Univariate analysis
Univariate analysis for all groups showed strong positive associations between resistin and fasting glucose (p = 0.01, Figure 2), and negative association with hip circumference (R2 0.6, p = 0.004, p = 0.03 respectively) but in CHD patients, resistin had a strong positive association with systolic and diastolic blood pressure and Apo-AII (R2 0.82, p < 0.0017, R2 = 0.45 p < 0.002 respectively). There was also a strong correlation with Apo-AII (p < 0.003), whereas, in T2DM patients, resistin was correlated with LDL (R2 = 0.79, p = 0.009).
Figure 2 The correlation of Serum resistin levels (ng/mL) with glucose (mmol/L) across Saudi subjects with univariate analysis across all case controls, T2DM and CHD subjects. Blue denotes the actual values with the line of best fit shown in pink (R2 = 0.03, p-value = 0.01).
Discussion
While it is apparent there are conflicting reports as to the role of serum resistin and its role in obesity mediated T2DM, no study to date has examined this in relation to a Saudi population with and without T2DM and/or CHD. This study determined that serum resistin levels were increased in T2DM and CHD in a Saudi population, when compared with case control subjects.
In addition it was demonstrated that a correlation exists between serum resistin and CRP levels in Saudi subjects, confirming our studies in Caucasian non-diabetic and T2DM subjects [18]. These data further suggest a potential role for resistin as a marker associated with inflammation in both T2DM and CHD disease. Studies have already shown pro-inflammatory cytokines to be important inducers of CRP, such as IL-6, TNF-α, and leptin [31], and resistin may also play a role as an inducer of CRP.
Other studies have examined the relationship between BMI and serum resistin levels. However, while some studies have shown such results as correlation [16], this study and our previous study [18] failed to determine such an effect. The relationship of resistin with increasing adiposity was not a primary determinant of this study and as such the range of BMI across this cohort may not be sufficient to identify a correlation. However, a negative correlation of serum resistin with hip circumference was noted, supporting a possible link between resistin and obesity.
Our earlier and present studies also investigated whether there was any correlation between serum resistin levels and fasted glucose levels across all the subjects. A significant correlation between fasting glucose and serum resistin was observed. This supports the previous finding by Lazar and his co-workers from their rodent model studies, where fasted blood glucose was higher in resistin-transgenic mice than in their non-transgenic littermates, and glucose tolerance was impaired in the hyper-resistinemic mice [32].
However in vivo human analysis of the relationship between serum resistin and glucose has produced conflicting reports [17,18,33,34]. No correlations were identified in serum resistin and glucose amongst non-obese, obese and obese T2DM subjects by examining the glucose disposal rate during a hyperinsulinaemic glucose clamp across groups [17]. In conjunction with this, an in vivo analysis by Azuma et al. also noted no significant correlation in serum resistin and glucose with cross-sectional analysis; but longitudinal analysis in the same cohort revealed changes in serum resistin to be positively correlated with glucose as well as other factors such as BMI, and insulin [34]. Hence, these studies may suggest that cross sectional analysis of serum resistin may also be inadequate to determine correlations. Further, the specificity of the resistin ELISA assay emerges as an important factor in these and other studies; to date, however, only one commercial ELISA appears to have been validated for cross-reactivity with different members of the RELM family and this may affect the findings in other studies [18].
Additional in vitro studies in human adipose cells supports the role of resistin in reducing glucose uptake, with the effect of revealing a potential functional role for resistin in in vivo metabolism [18].
In conclusion, these present findings suggest that resistin may have a dual role involved in sub-clinical inflammation as well as in altering glucose metabolism leading to the progression of T2DM in this Saudi cohort. As such, this study may add to the current speculation over the roles and actions of resistin, since studies continue to suggest a role for resistin in affecting glucose, insulin sensitivity and sub-clinical inflammation.
Acknowledgements
We are grateful to Mr Ziad Al-onaze and the Diabetic Center nurses in King Abdulaziz Hospital for their assistance.
==== Refs
Pickup JC Inflammation and activated innate immunity in the pathogenesis of type 2 diabetes Diabetes Care 2004 27 813 23 14988310
Dandona PA Aljada Inflammation : the lind between insulin resistance, obesity and diabetes Trends Immunol 2004 25 4 7 14698276 10.1016/j.it.2003.10.013
Hotamisligil GS Inflammatory pathways and insulin action Int J Obes Relat Metab Disord 2003 27 S53 5 14704746 10.1038/sj.ijo.0802502
Nuamah MA Yura S Sagawa N Itoh H Mise H Korita D kakui k Takemura M Ogawa Y Nakao K Fujii S Significant increase in maternal plasma leptin concentration in induce delivery: a possible contribution of pro-inflammatory cytokines to placental leptin secretion Endocr J 2004 51 177 87 15118268 10.1507/endocrj.51.177
Sitaraman S Liu X Charrier L Gu LH Ziegler TR Gewirtz A Merlin D Colonic leptin: source of a novel proinflammatory cytokine involved in IBD FASEB J 2004 18 696 8 Epub 2004 Feb 20. 14977884
Gaetke LM Oz HS Frederich RC McClain CJ Anti-TNF-alpha antibody normalizes serum leptin in IL-2 deficient mice J Am Coll Nutr 2003 22 415 20 14559934
Rajala MW Scherer PE Minireview: The adipocyte at the crossroads of energy homeostasis, inflammation, and atherosclerosis Endocrinology 2003 144 3765 73 12933646 10.1210/en.2003-0580
Holcomb IN Kabakoff RC Chan B Baker TW Gurney A Henzel W FIZZI, a novel cysteine-rich secreted protein associated with pulmonary inflammation, defines new gene family Embo J 2000 19 4046 55 10921885 10.1093/emboj/19.15.4046
Steppan CM Bailey ST Bhat S Brown EJ Banerjee RR Wright CM The hormone resistin links obesity to diabetes Nature 2001 409 307 12 11201732 10.1038/35053000
Janke J Engeli S Gorzelniak K Luft FC Sharma AM Resistin gene expression in human adipocytes is not related to insulin resistance Obes Res 2002 10 1 5 11786595
Ternan CL McTernan PG Harte AL Levick PL Barnett AH Kumar S Resistin, central obesity, and type 2 diabetes Lancet 2002 359 46 7 11809189 10.1016/S0140-6736(02)07281-1
Patel L Buckels AC Kinghorn IJ Murdock PR Holbrook JD Plumpton C Resistin is expressed in human macrophages and directly regulated by PPARagnna activators Biochem Biophys Res Commun 2003 300 472 6 12504108 10.1016/S0006-291X(02)02841-3
Donovan JA Koretzky GA CD45 and the immune response J Am Soc Nephrol 1993 4 976 85 8286719
Gomez-Ambrosi J Fruhbeck G Do resistin and resistin-linke molecules also link obesity to inflammatory diseases? Ann Intern Med 2001 135 306 7 11511160
Villena JA Kim KH Sul HS Pref-1 and ADSF/resistin: two secreted factors inhibiting adipose tissue development Horm Metab Res 2002 34 664 7 12660879 10.1055/s-2002-38244
Degawa-Yamauchi M Bovenkerk JE Juliar BE Waston W Kerr K Jones R Zhu Q Considine RV Serum resistin (FIZZ3) protein is increased in obese humans J Clin Endocrinol Metab 2003 88 5452 5 14602788 10.1210/jc.2002-021808
Heilbronn LK Rood J Relationship between serum resistin concentration and insulin resistance in nonobese, obese, and obese diabetic subjects J Clin Endocrinol Metab 2004 89 1844 8 15070954 10.1210/jc.2003-031410
Mc Ternan PG Fisher FM Resistin and type 2 diabetes: regulation of resistin expression by insulin and rosiglitazone and the effects of recombinant resistin on lipid and glucose metabolism in human differentiated adipocytes J Clin Endocrinol Metab 2003 88 6098 106 14671216 10.1210/jc.2003-030898
Volarova de courten B Degawa-Yamauchi M Considin RV Tatarnni PA High serum resistin is associated with an increase in adiposity but not a worsening of insulin resistance in Pima Indian Diabetes 2004 53 1279 84 15111497
Lee JH Chan JL Yiannakouris N Kontogianni M Eatrada E Seip R Orlova C Mantzoros CS Circulating resistin levels are not associated with obesity or insulin resistance in humans and not regulated by fasting or leptin administration: cross-sectional and interventional studies in normal, insulin-resistant, and diabetic subjects J Clin EndocrinolMetab 2003 88 4848 56 10.1210/jc.2003-030519
Fujinami A Obayashi H Ohta K Ichimura T Nishimura M Matsui H Kawahara Y Yamazaki M Ogata M Hasegawa G Nakamura N Yoshikawa T Nakano K Ohta M Enzyme-linked immunosorbent assay for circulating human resistin: resistin concentrations in normal subjects and patients with type 2 diabetes Clin Chim Acta 2004 339 57 63 14687894 10.1016/j.cccn.2003.09.009
Valsamakis G Mc Ternan PG Chetty R Al Daghri N Field A Hanif W Barnett AH Kumar S Modest weight loss and reduction in waist circumference after medical treatment are associated with favorable changes in serum adipocytokines Metabolism 2004 53 430 4 15045687 10.1016/j.metabol.2003.11.022
Wilson PW Assessing coronary heart disease risk with traditional and novel risk factors Clin Cardiol 2004 27 III7 11 15239485
Mather HM Keen H The South all Diabetes Survey: prevalence of known diabetes in Asians and Europeans Br Med J 1985 291 1081 1084 3931804
Jaber LA Slaughter RL Grunberger G Diabetes and related metabolic risk factors among Arab Americans Ann Pharmacother 1995 29 573 6 7663027
Kadiki OA Roaded RB Epidemiological and clinical patterns of diabetes mellitus in Benghazi, Libyan Arab Jamahiriya East Mediterr Health J 1999 5 6 13 10793775
Abdella NA Khogali MM Salman AD Ghuneimi SA Bajaj JS Pattern of non-insulin dependent diabetes mellitus in Kuwait Diabetes Res Clin Pract 1995 29 129 36 8591700 10.1016/0168-8227(95)01128-5
McNeely MJ Boyko EJ Type 2 diabetes prevalence in Asian Americans: results of a national health survey Diabetes Care 2004 27 66 9 14693968
Bajaj M Banerji MA Type 2 diabetes in South Asians: a pathophysiologic focus on the Asian-Indian epidemic Curr Diab Rep 2004 4 213 8 15132888
Matthews DRHJ Rudenski AS Naylor BA Treacher DF Turner RL Homeostasis model assessment: insulin resistance and b-cell function from fasting plasma glucose and insulin concentration in man Diabetologia 1985 28 412 419 3899825 10.1007/BF00280883
Maachi M Pieroni L Bruckert E Jardel C Fellahi S Hainque B Capeau J Bastard JP Systemic low-grade inflammation is related to both circulating and adipose tissue TNFalpha, leptin and IL-6 levels in obese women Int J Obes Relat Metab Disord 2004 Jun 22
Rangwala SM Rich AS Rhoades B Shapiro JS Obici S Rossetti L Lazar MA Abnormal Glucose Homeostasis due to Chronic Hyperresistinemia Diabetes 2004 Jun 9
Silha JV Krsek M Skrha JV Sucharda P Nyomba BL Murphy LJ Plasma resistin, adiponectin and leptin levels in lean and obese subjects: correlation with insulin resistance Eur J Endocrinol 2003 149 331 5 14514348 10.1530/eje.0.1490331
Azuma K Katsukawa F Oguchi S Murata M Yamazaki H Shimada A Saruta T Correlation between serum resistin level and adiposity in obese individuals Obes Res 2003 11 997 1001 12917505
|
15998471
|
PMC1183229
|
CC BY
|
2021-01-04 16:25:04
|
no
|
Cardiovasc Diabetol. 2005 Jul 5; 4:10
|
utf-8
|
Cardiovasc Diabetol
| 2,005 |
10.1186/1475-2840-4-10
|
oa_comm
|
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-171596323610.1186/1476-7120-3-17ReviewEchocardiography-based left ventricular mass estimation. How should we define hypertrophy? Foppa Murilo [email protected] Bruce B [email protected] Luis EP [email protected] Graduate Studies Program in Cardiology. School of Medicine. Federal University of Rio Grande do Sul. Porto Alegre – RS. Brazil2005 17 6 2005 3 17 17 3 5 2005 17 6 2005 Copyright © 2005 Foppa et al; licensee BioMed Central Ltd.2005Foppa 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.
Left ventricular hypertrophy is an important risk factor in cardiovascular disease and echocardiography has been widely used for diagnosis. Although an adequate methodologic standardization exists currently, differences in measurement and interpreting data is present in most of the older clinical studies. Variability in border limits criteria, left ventricular mass formulas, body size indexing and other adjustments affects the comparability among these studies and may influence both the clinical and epidemiologic use of echocardiography in the investigation of the left ventricular structure. We are going to review the most common measures that have been employed in left ventricular hypertrophy evaluation in the light of some recent population based echocardiographic studies, intending to show that echocardiography will remain a relatively inexpensive and accurate tool diagnostic tool.
cardiovascular diseaseechocardiographyepidemiologygeometric patternshypertrophyleft ventricularmalerisk factorsultrasonography
==== Body
Introduction
The diagnosis of left ventricular hypertrophy (LVH) has been incorporated in the clinical practice as an important marker of cardiovascular disease. Its prevalence depends on classification criteria and specific population characteristics, ranging from 3% in normotensive community-based samples [1] to about three-quarters of hypertensive patients [2]. Irrespective of other risk factors, those in the upper distribution of left ventricular mass have their risk at least duplicated for future cardiovascular morbidity and mortality, as summarized in one metanalysis [3].
Echocardiography has been clinically employed for more than 30 years, becoming one of the most important non-invasive imaging methods in the evaluation of cardiac morphology and dynamics. However, the apparent simplicity in LVH evaluation by echocardiography conceals several intrinsic and usually unrecognized critical steps that may limit its clinical validity.
This manuscript introduces the basic principles of left ventricular mass (LV mass) estimation by echocardiography, focusing on the potential limitations and discrepancies of such measurement, in order to provide the appropriate background for understanding the rather complex issue of defining the cut-off values for LVH diagnosis in populations. Although advances in echocardiographic techniques has minimized the impact of many of the methodological details discussed here, an understanding of these topics is important, since most of the clinical and epidemiologic studies currently published are based on the echocardiographic criteria here described. We also intend to give a critical evaluation of the diagnostic performance and clinical validation of this imaging method.
Left Ventricular Measurement
Left ventricular mass is generally calculated as the difference between the epicardium delimited volume and the left ventricular chamber volume multiplied by an estimate of myocardial density. Following this principle, several methodologies have been used to calculate left ventricular mass and to define hypertrophy with its own flaws and strengths on each step (Table 1), resulting in a wide range of values. Probably, the most significant echocardiographic limitation is related to inadequate quality imaging. Population-based studies are not able to obtain complete imaging in almost a quarter of screened patients [4,5]. mainly due to inappropriate acoustic windows.
Table 1 Critical Steps in Determining and Interpreting Left Ventricular Hypertrophy using Echocardiography
1. Imaging – Mode and Acquisition
2. Estimating Left Ventricular Volume
3. Defining Border Limits – Conventions of Layer Measurements
4. Calculating Mass – LV Mass Formulas
5. Indexing for Body Size
6. Determining Cut-off Points
a. Using a reference sample (normality/statistical criteria)
b. Using prognostic data (driven by clinical endpoint)
7. Evaluation of Left Ventricular Structure
8. Role of Additional Factors in LVM Determination
9. Clinical Correlates Associated with LVH
Imaging – Mode and Acquisition
Both M-mode and two-dimensional imaging can be employed to calculate left ventricular mass. M-mode imaging allows better endocardial border definition as it has greater resolution due to higher frame-rate, as long as adequate ultrasound beam positioning is ensured and ventricle shape approaches normality. Two-dimensional imaging, on the other hand, depicts the "real" ventricular shape and identifies regional motion abnormalities. However, the quality of two-dimensional imaging may be limited due to both lower lateral resolution and frame-rate. Additionally this option is more time consuming, limiting its use in epidemiological studies. Two-dimensional images are usually acquired both in paraesternal and apical views, depending on the geometrical formulas that are used.
Technological advances have joined both methods and partially minimized their limitations There are standardized and validated recommendations for the clinical use of two-dimensional determination of LV mass [6], However, two-dimensionally oriented M-mode, obtaining images perpendicularly from the longitudinal axis slightly above the papillary muscle level is widely employed in clinical practice and is accepted as an adequate alternative in epidemiological studies. Digital imaging has also made it possible to reconstruct diverse M-mode planes from two-dimensional images (so called anatomical M-mode), allowing better positioning, although the final image resolution is that of two-dimensional images. Although accurate [7], LV mass estimation using anatomical M-mode has not been adequately validated in clinical studies. Two-dimensional mode has also improved due to refined imaging processing technology, particularly second harmonic imaging. Also, built-in software for automatic border detection has been developed allowing calculation of real-time volumes. Although most of these newer technologies are widely available in commercially available ultrasound equipment, their potential for providing additional accuracy in the evaluation of LVH remains poorly characterized.
A relevant degree of variability in LV mass determination could be attributed to online measurement inaccuracies due to lower imaging resolution of older equipment, which could reach 10% of parietal thickness. Nowadays, variability due to on-line or off-line analysis of digitalized images calculations is of considerably smaller magnitude [8].
We will focus our discussion on M-mode estimation of LV mass, since most epidemiological reports use this imaging modality. Preference for M-mode is based on its technical feasibility and availability at the time when most studies were performed. However, despite adequate correlation with two-dimensional measurements [9], M-mode has been suggested to averagely underestimate LV mass in about 20 g [10]. Theoretically, two-dimensional imaging would be more adequate in samples of patients with cardiovascular disease, where LV shape assumptions play a critical role in LV mass estimation.
Estimating Left Ventricular Volume
Determination of left ventricular volumes is accomplished using formulas that fit ventricular shape to primary geometric figures. Ellipse, cylinder, cone, and truncated polyhedrons have been employed and validated in normally shaped ventricles, although most studies have limited sample sizes [11]. Some geometrical assumptions are best fit using two-dimensional images while others can be performed assuming geometric forms from M-mode imaging.
If a two-dimensional approach is used, both area-length and truncated-ellipsoid models are feasible and reasonably accurate, with validated formulas [6]. Among other formulas for volume calculation, the modified Simpson's Discs Rule, which is greatly facilitated nowadays by built-in software, offers flexibility that allows accurate estimation of left ventricular volumes even in greatly distorted ventricles. Even though volumes determined by Simpson's Rule are frequently used for ventricular function determination, they are not routinely used to calculate LV mass, probably due to limitations resulting from poor epicardial delimitation in some patients. Other geometrical models have been proposed in the past [12,13], with variable use in clinical practice.
The method of cubed formulas, which incorporate only one dimension of the left ventricular cavity and assume an ellipsoid geometry, is the most widely used to calculate left ventricular volumes and mass in noninvasive laboratories worldwide. The broad acceptance of M-mode derived estimation relies on the fact that it was the first method that was validated and also on its technical simplicity. The one-dimensional approach, however, imposes more strict geometrical assumptions and amplifies the risk of inaccuracy, as measurement errors are cubed.
Defining Border Limits – Conventions of Layer Measurements
Ultrasound signals are reinforced where surfaces change density, allowing definition of limits between surface layers. The inclusion or exclusion of these echoes from interfaces of the left ventricular cavity or myocardial wall can cause significant discrepancies in the overall measurements [14]. Initial M-mode standard recommended inclusion of the edges as part of interventricular septum thickness, but exclusion of the posterior wall epicardial edge [15]. Investigators of the University of Pennsylvania developed a criteria (The Penn Convention) in which all edges are not included in parietal thickness measurements, but are considered as part of the ventricle cavity [16]. This approach underestimate LV mass when compared to the M-mode convention, proposed by the American Society of Echocardiography (ASE). This latter convention (ASE) is the most accepted border definition criteria, becoming the standard recommendation for M-mode estimations, and uses the leading edge of each layer [17](Figure 1). Employing Penn and ASE convention with the same volume formulas may originate LV mass discrepancies in the range of 15% in men and 18% in women [18]. Measurement convention must be acknowledged and adequately corrected for in comparisons of clinical studies of LVH.
Figure 1 Comparison between M-mode border measurement conventions. The Standard convention measures from leading to trailing edge in the septum and from leading to leading edge of the posterior wall. Penn criteria excludes echoes from parietal walls while ASE criteria measure leading to leading edge. (LVDd: Left Ventricular Diameter in Diastole).
In the past few years, spatial resolution of transthoracic echocardiography has greatly enhanced, leading to major improvements in image quality. Most of this progress can be attributed to second harmonic imaging, that significantly increases signal-to-noise ratio by receiving only harmonic frequencies. Although this technology is now widely available in most ultrasound equipment, its potential accuracy to evaluate LVH remains poorly characterized. A recent study suggests that LV mass estimations using second harmonic imaging can cause as much as a 26% increment in mean LV mass index corrected for body surface area, when compared to standard fundamental imaging using similar formulas and conventions [19]. This is probably due to an increase in border refringency.
Calculating Mass (Left Ventricular Mass Formulas)
The most commonly used formulas to estimate LV mass are all variations of the same mathematical principle, based in the volume formulas stated above. Original calculations from Troy and coworkers were the first to be recommended as standard to estimate LV mass from M-mode measurements (Formula 1) [15].
Formula 1: LV mass(Troy) = 1.05 ([LVIDD + PWTD + IVSTD]3- [LVIDD]3) g.
Where: LVIDD = Left Ventricular Internal Diameter in Diastole
PWTD = Posterior Wall Thickness in Diastole
IVSTD = Interventricular Septum Thickness in Diastole
Subsequently, Devereux and colleagues suggested a slightly modified regression equation, using the Penn convention as the border definition criteria (Formula 2). Their prediction equation in this pivotal study was derived from necropsy findings of 34 patients [16].
Formula 2: LV mass(Penn) = 1.04 ([LVIDD + PWTD + IVSTD]3- [LVIDD]3) -13,6 g.
As depicted above, each regression equation was derived based on a specific border limits convention, an issue that is source of great confusion when interpreting different studies [20]. As expected, LV mass calculations derived from both formulas are linearly correlated, but final crude estimations may differ by more than 20%. Devereux and colleagues proposed a new adjusted equation, validated on necropsy findings of 52 individuals [21], using the ASE convention and accounting for this discrepancy (Formula 3).
Formula3: LVmass(ASE): 0.8 (1.04 ([LVIDD + PWTD + IVSTD]3- [LVIDD]3))+ 0,6 g.
Some critical aspects must be acknowledged regarding LV mass formulas. First, all necropsy validation studies have limited sample sizes and evaluate heterogeneous ventricular configurations. Second, these formulas may not perform adequately in distorted ventricles, where a two-dimensional approach is preferred. Different formulas may yield distinct cut point values, as demonstrated by Levy and coworkers in the Framingham cohort [18]. Finally, other post-mortem study showed only moderate correlation between echocardiographic and autopsy LV mass estimations (correlation coefficients ranging from 0.58 to 0.67) [22].
Indexing for Body Size
Both body size and body habitus are clearly associated with LV dimensions and mass. Diverse normalization and indexes were created and tested to adjust for three different sources of physiologic variation in LV mass: lean body mass, obesity, and gender. However, the interdependence of such associations should be carefully understood to allow an adequate correction of LV mass without distorting its association with cardiovascular disease.
Several indexes for body size correction have been proposed, such as height, diverse allometric height adjustments, weight, body surface area, body mass index, and free-fat mass. The best way for normalization of LV mass is still controversial and another source of confusion. Different body-size adjustment criteria and their standard cut points result in different prevalence of patients with LVH [18]. Not surprisingly, those with higher LV mass are more frequently classified as hypertrophic by different classifications simultaneously [23].
The body surface area correction, using the Dubois formula [24], reduces variability due to body size and gender [25], but this index underestimates LV mass in the upper range of the body surface area distribution [26]. A correction based on height alone would allow evaluation of the separate role of obesity in LVH as proposed by Levy and coworkers [18].
Adjustment of LV mass with body surface area would imply that obese patients are expected to have higher LV mass estimations per se. In this scenario, height-based adjustments can more accurately estimate LV mass and the resulting cardiovascular risk associated with LVH in the obese. This turns to be particularly relevant in risk stratification due to the frequent clustering of hypertension, obesity and LHV. Different allometric height-based adjustments have been used. Height2.7, derived from regression models in normal samples from De Simone and coworkers [26], appears to offer the most accurate estimation of LV hypertrophy and risk factors for pathologic changes in the heart structure, particularly in obese subjects. Zoccali and colleagues found LVH indexed by height2.7 to be a better predictor of cardiovascular events than LVH indexed using body surface area in a group of patients under dialysis [27]. Liao and colleagues [28] studied 988 patients and identified progressive increments in death rates with both body surface area and with height2.7 indexing criteria. Subjects simultaneously classified as LVH with body surface area and height2.7 criteria had increased average LV mass and a 3-fold increase in death rates, while those classified as LVH only when indexed by height indexes had no increase in future cardiovascular events. In summary, it appears prudent to favor indexes that do not adjust for obesity, such as height, and height2.7, particularly in studies in which the independent impact of obesity is in question. Body surface area indexing permits adequate classification of most of patients in clinical practice, incorporating in LVH determination some of the risk associated with obesity.
Finally, men have increased LV mass and at least part of this effect can be attributed to body size differences. Gender differences in LV mass are first noticed around puberty and can be minimized although not eliminated by adequate indexing of body size [29]. Due to this difference in LV mass, some criteria for body size adjustments use gender-specific cut points for normality as will be seen below.
Determining Cut-off Points
The determination of cut points in biological variables to define abnormality is frequently a source of controversy, and can be driven by different strategies. The definition of what constitutes an abnormal LV mass is no exception to this rule.
Left Ventricular hypertrophy diagnosis defined by deviation from mean
LV mass as most biological variables are statistically distributed in normal or skewed curves. One can consider the diagnosis of LV hypertrophy in those who are in the extreme right tail of a "Gaussian" distribution, such as beyond two standard deviations of a reference sample of normal individuals. Identification of a "normal representative sample" is not trivial and most studies use relatively small samples. In the late 80s, Levy and coworkers. [18] published a landmark paper evaluating a subset of individuals without known cardiovascular risk factors in the Framingham Cohort. These authors calculated LV mass both with the ASE convention and Troy equation (Formula 1) and with the Penn Convention and Devereux equation (Formula 2) to estimate LV mass, and proposed normal limits for LV mass for men and women, based on cut points at two standard deviations above the mean and using several indexes for body size correction (Table 2). These criteria are widely used in clinical practice and research, despite limitations in representativeness, as they may not perform well in non-white populations.
Table 2 Left ventricular hypertrophy cut points (Healthy reference group from The Framingham cohort).
Men Women
Mean Mean + 2sd Mean Mean + 2sd
LVM(ASE) (g) 208 294 145 198
LVM(Penn) (g) 177 259 118 166
LVM/BSA(ASE) (g/m2) 109 150 89 120
LVM/BSA(Penn) (g/m2) 92 131 72 100
LVM/Ht(ASE) (g/m) 117 163 89 121
LVM/Ht(Penn) (g/m) 99 143 73 102
Adapted from[18].
Left Ventricular hypertrophy defined by prediction of clinical disease
Increase in LV mass has been shown to be an independent prognostic factor for intermediate endpoints [30] and clinical outcomes such as major cardiovascular events and mortality [31,32]., total mortality [27,28,33] and sudden death [34]. However, the risk associated with increases in several "physiological" variables is mostly linear over a great range of variation. This behavior has already been suggested for blood pressure [35] and cholesterol levels [36], leading to aggressive management strategies ("the lower the better"). In fact, Levy and colleagues demonstrated a progressive increase in risk associated to LV mass, even at levels not considered as "hypertrophic". Cardiovascular disease and death rates had a 1.5-fold increase for each 50 g/m of LV mass indexed by height [31]. In a subset of hypertensive patients [32], cardiovascular disease increased monotonically with more than a 4-fold increase in risk between the lowest and highest LV mass quintiles. In this study, clinically relevant increment in risk was identified in patients with LV mass below the limits usually employed for LVH definition.
These findings suggest that traditional cut-off limits may ignore cardiovascular risk associated with increased LV mass in the "normal" range based on statistical assumptions. De Simone and coworkers introduced the concept of inappropriate LV mass increase [37] to use LVH within the context of risk prediction employing multiple factors. It is considered a clinical relevant LV mass increase values above 128% of a predicted LV mass based on gender, estimated stroke volume and height2.7 [38,39]
Moreover, LVH regression has been used in clinical trials as a favorable prognostic marker. A metanalysis has shown in hypertensive patients that the regression of LVH predicts a reduction of more than 50% in cardiovascular events[40]
Evaluation of Left Ventricular Structure
Alternative concepts to LVH in the determination of left ventricular adaptive processes that take place in the overloaded ventricle assesses the fundamental components used in LV mass estimations, namely wall thickness and diastolic chamber dimension. The expected pathophysiological response of each of these components is theoretically distinct, as pressure overload leads to increased wall thickness and volume overload leads to chamber dilation. These differences cannot be assessed solely by LV mass calculations.
Relative Wall Thickness
Parietal thickness and its relation to LV chamber size have been recognized as measures of hypertrophy for more than 30 years [41]. Relative wall thickness (RWT) is measured in clinical studies both as: 2 * posterior wall thickness divided by LV diastolic diameter or, septal wall thickness + posterior wall thickness divided by LV diastolic diameter. Even thought these measures have been used interchangeably by some investigators, septal asymmetry (IVSTD/PWTD > 1.3) was present in about 5% of Framingham subjects [42] and can lead to an underestimation of relative thickness when only posterior wall thickness is used. The reference cut point value for increased relative wall thickness derived from upper limits of normal samples is usually 0.44 [43] or 0.45 [42], irrespective of which formula is used. RWT provides information regarding LV geometry independent of other calculations [44], precluding the requirement of most corrections. Nevertheless, significant LVH can occur without major changes in RWT, particularly when simultaneous pressure and volume overload are present.
Geometric Patterns
Attempts have been made to evaluate separately adaptive responses in parietal thickness increase and in dilation. Initially, Savage and coworkers [42] stratified Framingham patients with LVH in subgroups as: disproportionate septal LVH; concentric LVH; eccentric-dilated LVH, and eccentric non-dilated LVH. They identified in the 3 last categories increasing levels of systolic blood pressure, utilizing retrospective blood pressure data from 30-years of the cohort follow-up, suggesting a progressive character of adaptive mechanisms. A later approach defined 4 distinct geometric patterns: normal geometry, concentric remodeling, concentric hypertrophy and eccentric hypertrophy (Figure 2). Ganau and coworkers [43], using echocardiographic hemodynamic estimates, reinforced the impression that the geometric patterns parallels progressive hemodynamic changes. Hypertensive target organ disease measured by fundoscopic alterations are also more frequent in hypertrophic geometric patterns [30].
Figure 2 Geometric Patterns.
Koren and coworkers [45] used cut points of 125 g/m2 for LVH and 0.45 for RWT in a sample of hypertensive patients and found a 10-year incidence of cardiovascular events of 31% in those with concentric hypertrophy compared to 11% in those with normal geometry. In 1995, two cohorts studies were simultaneously published evaluating geometric patterns impact in the incidence of cardiovascular events. Verdecchia and colleagues [46] studying 694 patients with body surface area indexed LV mass lower than 125 g/m2, without additional adjustment for obesity and other metabolic risk factors, found a relative risk of 2.6 in the 272 patients with concentric remodeling compared to normal geometry patients. Krumholz and coworkers [47], studied 3209 from The Framingham study, indexed LVM by height using cut points of 143 g/m in men and 102 g/m in women and adjusted the models for obesity and other relevant covariates. Their analysis showed a relative risk of 2.1 for all cause mortality with concentric hypertrophy, but not additional risk in those classified as concentric remodeling. Relative risk became nonsignificant when a correction for LV mass was included in the models. Verdecchia and colleagues [48], afterwards could not demonstrate additional risk associated with increased relative wall thickness in those classified as hypertrophic. These data may suggest a smaller independent risk associated with increased wall thickness in hypertensive patients without LVH criteria.
Even though the additional prognostic role of geometric patterns over LVH may be lesser than initially supposed, this classification permits identification of determined adaptive processes. Concentric remodeling may be related to specific pathophysiological adaptations, particularly related to glucose and insulin metabolism [49-51] and studies in contemporary cohorts have also shown an association of concentric forms with diabetes [52,53].
We believe that geometric classification, with adequate body size indexing and clearly defined standardization, may be an alternative and informative strategy to evaluate adaptive responses, providing information beyond that provided by classification with respect to left ventricular hypertrophy.
Role of additional factors in left ventricular mass and hypertrophy determination
Gender and body size are clearly identified as predictors of LV mass and LVH Definitions are usually corrected and/or stratified for these factors, as seen above. Many others constitutional factors and exposures may lead to changes in LV mass. Some of these factors are pathophysiologically involved in LVH and, moreover, interact among themselves, limiting the interpretation of the independent role of each one.
Gender
Differences in LV mass due to gender, independent of questions related to body size, may have pathophysiological implications. Women have been shown to have an increased parietal hypertrophic responses to pressure overload [54,55]., even after body size correction. This adaptive pattern was demonstrated also in animal models [56]. The unfavorable prognostic implications of this hypertrophic response are suggested by the findings of Liao and coworkers of a 5-fold greater risk of death associated with LV hypertrophy indexed by BSA in woman compared to the risk associated to LVH in men. However, despite using gender specific cut-offs for LVH, additional adjustment for obesity was not performed. Employment of height2.7 indexing allowed to use a unique cut-point of 51 g/m2.7 for both genders [26], reducing the impact of gender in LVH inference, at least in African-Americans [57].
Obesity
Although the best strategy to adjust LVM for obesity is a matter of debate, obesity is increasingly recognized as an independent predictor of cardiovascular morbidity and mortality [58,59]. The increase in LV mass related to obesity is probably more than a mere physiologic adaptation.
Obesity has been shown to be independently associated to LVH [60], particularly in populations with a high prevalence of hypertension and other metabolic risk factors [61,62]. Despite this association, the impact of obesity on LVH may be less than expected [63], as Iacobellis and colleagues [64] have demonstrated that "uncomplicated obesity" was not a risk factor for LVH when indexed by either body surface area or height 2.7. As obesity, however, causes complications, it is frequently accompanied by additional risk factors. Adjusting by height 2.7 minimizes the interference of obesity in LV mass estimates (See above: Body-size Indexing).
Age
LV mass progressively increases during aging [65], particularly parietal thickness [4], which was seen in both normotensive and hypertensive patients [66]. Heart size increases during infancy and adolescence due to body size enlargement and, at this stage, the gender differences become prominent [67]. The rate of LV mass increase due to age changes in magnitude [29], weakening its independent role at older individuals, when other risk factors play a greater role [63]. Dannenbeg and coworkers [68] demonstrated that LV mass did not increase with age in a healthy sub-sample of The Framingham study, suggesting that most of the supposed physiological increase is caused by other determinants. These results are reinforced by studies in younger subjects where the age-associated increase in LV mass is partially explained by body size and blood pressure changes [67]. Nevertheless, it appears prudent to adjust for age in epidemiological investigations related to LV mass and hypertrophy.
Ethnicity
LVH is particularly prevalent in African-Americans [5,62,69-72]. In these analyses, two particular aspects deserve consideration. An increased crude prevalence of LVH in African-Americans and Hispanics is more evident using height-indexed LV mass than with body surface area-indexed LV mass [71], suggesting that obesity may partially explain the reported ethnic differences. Furthermore, adaptive response to hypertension may differ across ethnic groups. Hypertensive African-Americans, in comparison with hypertensive whites, have increased relative wall thickness, resulting in an increased frequency of concentric remodeling, given equivalent LV mass estimates [72,73]. However, Afro-American ancestry has been identified as an independent risk factor for LVH [74].
Clinical correlates of left ventricular hypertrophy
Several factors have been shown repeatedly in epidemiologic studies to associate with LVH. Investigation and prognostication based on LVH should take these factors in the account.
Blood Pressure and Hypertension
Numerous population based studies have unequivocally shown an association between hypertension and LVH [4,5,65,75]. Other reports usually stratify their analysis by or restrict to those with hypertension to allow better evaluation of additional risk factors [45,73,76,77]. It is interesting that even within the normal range, increases in blood pressure is related to an increased LV mass [67]. This increment may be attributed to the classical pathophysiological concept of hypertrophic response to increased overload, although neuro-humoral and genetic factors have been also implicated [78]. LVH association with hypertension is so evident that it is recognized as target organ damage in hypertensive disease by several clinical practice guidelines, representing an intermediate unfavorable prognostic marker [79,80].
Diabetes and The Metabolic Syndrome
Together with obesity and hypertension, diabetes has been implicated as an important determinant of left ventricular mass in most population-based studies [5,52,62,81,82]. Myocardial and systemic mechanisms, as an increased extra-cellular matrix, vascular hypertrophy and vasoconstriction [83], have been attributed to this hypertrophic response.
An adaptive response has been shown to diverse degrees of altered carbohydrate metabolism, as in Cardiovascular Health Study [82] and in The Strong Heart Study cohort, where diabetes [52], impaired glucose tolerance [84] and insulin levels[85] where associated with increased LV mass. Although associated with an increase in left ventricular mass, hyperinsulinemia [49] and insulin resistance [50] show a stronger association with concentric remodeling. Concentric hypertrophy is more pronounced in diabetes presenting with microalbuminuria [51,86]., which could imply a progressive adaptive process.
A gender difference in the left ventricular response to diabetes, with an increase in parietal thickening, rather than hypertrophy, being prominent in women has been suggested [81,85,87].
LV mass increase is also seen in individuals with other known risk factors, as in those linked to the metabolic syndrome [62,88]., where pathophysiological aspects related to this syndrome may directly affect ventricular adaptive mechanisms.
Other Risk Factors
A multitude of other factors have been shown to be independently related to LV mass. It should be emphasized that estimates of the relative magnitude of these factors varies according to the degree of adjustment for other known risk factors in statistical modeling. Primary valvular and myocardial disease are clearly related to LV mass increase but will not be subject of our review.
Environmental exposures such as alcohol consumption [89], salt intake [90], smoking [4,89]. and increased leisure-time physical activity in men [91] have been associated to increased LV mass. Other factors such as blood lipids, pulmonary function, the heart rate and hematocrit have also been implicated but with some inconsistency among different studies [4,75,92,93]. Also, low weight at 1 year-old has been suggested as LV hypertrophy risk factor, concordant with Barker's Theory of the fetal and early life origin of chronic disease[94].
Clinical validity and impact of such factors is controversial, but it may be important to consider them as relevant potential confounders in epidemiological studies investigating the role of novel risk factors in LVH and the role of LVH in disease prediction.
Reproducibility
Each step in LV mass measurement is a potential source of variability. In M-mode measurement, differences of approximately 5% may translate into differences in LV mass between 8% and 15% [95], which can represent about 50 g. This variability can be attributed particularly to the measurement of wall thicknesses and border layer definition [96-98]. Reproducibility is slightly better using the ASE rather than the Penn convention [98]. Additional smaller differences in left ventricular volume determinations can also be attributed to changes in body position or circulatory loading conditions [99].
Intraobserver M-mode measurements may vary about 5% between echocardiographic studies, while interobserver variability may reach 15%. Some trials retesting patients found differences of up to 30 g between tests [100,101]. When all sources of variability are taken into account, differences in the estimates are not small, since they approach a difference in LV mass values that is associated with a clinically important increased cardiovascular risk. Strategies such as core laboratory reading, strict protocols and regular training may keep this variability in an acceptable range for clinical and epidemiologic studies.
Comparison with other Imaging Methods
Autopsy is classically employed as the gold standard in heart hypertrophy studies, because it objectively measures LV mass. However, use of reported data is complicated by the fact that macroscopic LVH definition criteria are usually more varied than those used in non-invasive testing [102-105], as well as by the fact that different studies have applied different indexing techniques for body size.
From a histological point of view, myocytes hyperplasia is uncommon in adults. Pathologic studies and animal models suggest evaluating hypertrophy suggest that myocytes keep their integrity and functionality until an increase up to 50 -70% above normal [106]. This is concordant with the Linzbach's critical level of LVH, above which cytopathological changes occur with disruption of myocardial tissue integrity and functioning [107].
LV mass can also be calculated from angiography. Although diverse formulas have been employed and validated [13], correlation with echocardiographic calculated LV mass is fair to moderate, with correlation coefficients of between 0.50 and 0.70 [108].
Radioisotopic gated myocardial perfusion imaging with 99mTc-Sestamibi has been employed to estimate LV mass. Its accuracy is limited by image construction and processing variability, resulting in a limited correlation with echocardiographic LV mass [109]. However, Maruyama and coworkers [110] found a good correlation coefficient (r = 0.96) between gated 99mTc-tetrofosmin myocardial perfusion and echocardiographic based LV mass estimations, using an automated quantitative software.
Newer imaging methods have been employed in LV mass determination. Computed tomography has a good correlation with necropsy findings (r = 0.97). In vivo intrareader variability was estimated to be equivalent to 19 g and intereader 28 g [111]. Magnetic resonance imaging (MRI) has emerged as a highly reproducible and accurate imaging methodology in the evaluation of LV geometry and mass [112-115]. As a result, it is of great value in evaluating distorted ventricles and its high accuracy may partially counterbalance its costs, due to the smaller samples needed. However, echocardiography costs are considerably lower in most of the countries, there is no significant radiation exposure [116], and just a few-population based studies have used these costly and less available newer imaging techniques.
Real time three-dimensional echocardiography is still experimental, but has incorporated technical advantages in image acquisition and processing. This method may permit accurate real time LV mass measurement without the caveats of geometrical assumptions. Preliminary data suggest that real time three-dimensional echocardiography is at least as accurate and reproducible as MRI calculations [117]. Contrast echo with microbubbles also permits increasing accuracy, particularly in those with inadequate acoustic windows or with distorted ventricles [118].
Conclusion
LV mass estimation and LVH diagnosis role in cardiovascular disease management is based on epidemiological research and also on clinical grounds. Despite more than 30 years of use echocardiography-based LVH calculation and definition are still variable among ultrasound technicians and laboratories around the world, leading to inconsistency among epidemiological studies and possibly limiting its clinical application. Several technical aspects of the echocardiographic exam can generate substantial errors in LV estimations, some of them equivalent in size to those expected to result from pathophysiological processes and therapeutic strategies. Also, adequate indexing for body size seems to be a critical point in defining pathological hypertrophy. LV mass is closely related to the other known cardiovascular risk factors, that must be taken into account concomitantly. Finally, since the risk associated to LV mass appears to be progressive, without a clear threshold, additional input can be added at different baseline risks, defined by the prevalence of other known cardiovascular risk factors. The addition of multiple newer markers, however, leads to a small increment in risk stratification capacity over formulas applying only classical risk factors [119]. Despite these limitations, the role of echocardiography in LV mass determination is of great clinical value.
Considering all the aspects reviewed, use of echocardiography in clinical studies must be standardized applying already defined criteria. In delineating a study, if two-dimensional is impractical, then two-dimensional guided M-mode, using ASE criteria and Devereux modified formula, will allow estimation of LV mass with an acceptable level of accuracy. Additionally, adequate adjustment of related covariates must be undertaken. LHV/Ht2.7greater than 51g/m2.7 appears to be a reliable criteria to define LVH, and the inclusion of a measurement of relative wall thickness, individually or classified as geometric patterns improves the identification of the adaptive mechanisms involved.
Echocardiography is widely available all over the world and major technical improvements have been achieved in the last two decades. Given careful attention with respect to the technical aspects appraised in this review, echocardiography will remain a safe, inexpensive and accurate tool for both the clinical diagnosis and epidemiologic investigation of left ventricular hypertrophy
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors contributed equally to this work, read and approved the final manuscript.
==== Refs
Antoniucci D Seccareccia F Menotti A Dovellini EV Prati PL Rovelli F Fazzini PF Prevalence and correlates of echocardiographic determined left ventricular hypertrophy in 2318 asymptomatic middle-aged men: the ECCIS project. Epidemiolgia e Clinica della Cardiopatia Ischemica Silente G Ital Cardiol 1997 27 363 369 9199953
Coca A Gabriel R de la FM Lopez-Sendon JL Fernandez R Sagastagoitia JD Garcia JJ Barajas R The impact of different echocardiographic diagnostic criteria on the prevalence of left ventricular hypertrophy in essential hypertension: the VITAE study. Ventriculo Izquierdo Tension Arterial Espana J Hypertens 1999 17 1471 1480 10526909 10.1097/00004872-199917100-00016
Vakili BA Okin PM Devereux RB Prognostic implications of left ventricular hypertrophy Am Heart J 2001 141 334 341 11231428 10.1067/mhj.2001.113218
Gardin JM Arnold A Gottdiener JS Wong ND Fried LP Klopfenstein HS O'Leary DH Tracy R Kronmal R Left ventricular mass in the elderly. The Cardiovascular Health Study Hypertension 1997 29 1095 1103 9149672
Skelton TN Andrew ME Arnett DK Burchfiel CM Garrison RJ Samdarshi TE Taylor HA Hutchinson RG Echocardiographic left ventricular mass in African-Americans: the Jackson cohort of the Atherosclerosis Risk in Communities Study Echocardiography 2003 20 111 120 12848675 10.1046/j.1540-8175.2003.03000.x
Schiller NB Shah PM Crawford M DeMaria A Devereux R Feigenbaum H Gutgesell H Reichek N Sahn D Schnittger I Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms J Am Soc Echocardiogr 1989 2 358 367 2698218
Katsuki K Nakatani S Kanzaki H Hanatani A Tanaka N Nakasone I Masuda Y Yamagishi M Miyatake K [Clinical validation of accuracy of anatomical M-mode measurements: effect of harmonic imaging] J Cardiol 2001 37 35 42 11200654
Arnett DK Skelton TN Liebson PR Benjamin E Hutchinson RG Comparison of m-mode echocardiographic left ventricular mass measured using digital and strip chart readings: The Atherosclerosis Risk in Communities (ARIC) study Cardiovasc Ultrasound 2003 1 8 12857352 10.1186/1476-7120-1-8
Park SH Shub C Nobrega TP Bailey KR Seward JB Two-dimensional echocardiographic calculation of left ventricular mass as recommended by the American Society of Echocardiography: correlation with autopsy and M-mode echocardiography J Am Soc Echocardiogr 1996 9 119 128 8849607
Paczek A Gardin JM Hardin JM Anton-Culver H Kurosaki T Hsu C Bild DE Comparison of M-mode and two-dimensional echocardiographic algorithms used to estimate left ventricular mass: the Coronary Artery Risk Development in Young Adults Study J Am Soc Echocardiogr 1995 8 780 792 8611278
Vuille C Weyman A Weyman A Left Ventricle I: General Considerations, assessment of chamber size and function Principles and Practice of Echocardiography 1994 2 Malvern, Pennsylvania: Lea & Febinger 575 624
Teichholz LE Kreulen T Herman MV Gorlin R Problems in echocardiographic volume determinations: echocardiographic-angiographic correlations in the presence of absence of asynergy Am J Cardiol 1976 37 7 11 1244736 10.1016/0002-9149(76)90491-4
Tortoledo FA Quinones MA Fernandez GC Waggoner AD Winters WL Jr Quantification of left ventricular volumes by two-dimensional echocardiography: a simplified and accurate approach Circulation 1983 67 579 584 6821900
Reichek N Standardization in the measurement of left ventricular wall mass. M-mode echocardiography Hypertension 1987 9 II27 II29 2948909
Troy BL Pombo J Rackley CE Measurement of left ventricular wall thickness and mass by echocardiography Circulation 1972 45 602 611 4258936
Devereux RB Reichek N Echocardiographic determination of left ventricular mass in man. Anatomic validation of the method Circulation 1977 55 613 618 138494
Sahn DJ DeMaria A Kisslo J Weyman A The committee on M-mode Standardization of the American Society of Echocardiography: recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic methods Circulation 1978 58 1072 1083 709763
Levy D Savage DD Garrison RJ Anderson KM Kannel WB Castelli WP Echocardiographic criteria for left ventricular hypertrophy: the Framingham Heart Study Am J Cardiol 1987 59 956 960 2952002 10.1016/0002-9149(87)91133-7
McGavigan AD Dunn FG Goodfield NE Secondary harmonic imaging overestimates left ventricular mass compared to fundamental echocardiography Eur J Echocardiogr 2003 4 178 181 12928020 10.1016/S1525-2167(02)00164-6
Devereux RB Detection of left ventricular hypertrophy by M-mode echocardiography. Anatomic validation, standardization, and comparison to other methods Hypertension 1987 9 II9 II26
Devereux RB Alonso DR Lutas EM Gottlieb GJ Campo E Sachs I Reichek N Echocardiographic assessment of left ventricular hypertrophy: comparison to necropsy findings Am J Cardiol 1986 57 450 458 2936235 10.1016/0002-9149(86)90771-X
Bachenberg TC Shub C Hauck AJ Edwards WD Can anatomical left ventricular mass be estimated reliably by M-mode echocardiography? A clinicopathological study of ninety-three patients Echocardiography 1991 8 9 15 10150003
Wachtell K Bella JN Liebson PR Gerdts E Dahlof B Aalto T Roman MJ Papademetriou V Ibsen H Rokkedal J Impact of different partition values on prevalences of left ventricular hypertrophy and concentric geometry in a large hypertensive population : the LIFE study Hypertension 2000 35 6 12 10642267
Dubois D Dubois EF A formula to estimate the aproximate surface area if height and weight be known Arch Intern Med 1916 17 863 871
Devereux RB Lutas EM Casale PN Kligfield P Eisenberg RR Hammond IW Miller DH Reis G Alderman MH Laragh JH Standardization of M-mode echocardiographic left ventricular anatomic measurements J Am Coll Cardiol 1984 4 1222 1230 6238987
De Simone G Daniels SR Devereux RB Meyer RA Roman MJ de Divitiis O Alderman MH Left ventricular mass and body size in normotensive children and adults: assessment of allometric relations and impact of overweight J Am Coll Cardiol 1992 20 1251 1260 1401629
Zoccali C Benedetto FA Mallamaci F Tripepi G Giacone G Cataliotti A Seminara G Stancanelli B Malatino LS Prognostic impact of the indexation of left ventricular mass in patients undergoing dialysis J Am Soc Nephrol 2001 12 2768 2774 11729247
Liao Y Cooper RS Durazo-Arvizu R Mensah GA Ghali JK Prediction of mortality risk by different methods of indexation for left ventricular mass J Am Coll Cardiol 1997 29 641 647 9060905 10.1016/S0735-1097(96)00552-9
De Simone G Devereux RB Daniels SR Meyer RA Gender differences in left ventricular growth Hypertension 1995 26 979 983 7490158
Shigematsu Y Hamada M Mukai M Matsuoka H Sumimoto T Hiwada K Clinical evidence for an association between left ventricular geometric adaptation and extracardiac target organ damage in essential hypertension J Hypertens 1995 13 155 160 7759846
Levy D Garrison RJ Savage DD Kannel WB Castelli WP Prognostic implications of echocardiographically determined left ventricular mass in the Framingham Heart Study N Engl J Med 1990 322 1561 1566 2139921
Schillaci G Verdecchia P Porcellati C Cuccurullo O Cosco C Perticone F Continuous relation between left ventricular mass and cardiovascular risk in essential hypertension Hypertension 2000 35 580 586 10679501
Verdecchia P Carini G Circo A Dovellini E Giovannini E Lombardo M Solinas P Gorini M Maggioni AP Left ventricular mass and cardiovascular morbidity in essential hypertension: the MAVI study J Am Coll Cardiol 2001 38 1829 1835 11738281 10.1016/S0735-1097(01)01663-1
Haider AW Larson MG Benjamin EJ Levy D Increased left ventricular mass and hypertrophy are associated with increased risk for sudden death J Am Coll Cardiol 1998 32 1454 1459 9809962 10.1016/S0735-1097(98)00407-0
MacMahon S Peto R Cutler J Collins R Sorlie P Neaton J Abbott R Godwin J Dyer A Stamler J Blood pressure, stroke, and coronary heart disease. Part 1, prolonged differences in blood pressure: prospective observational studies corrected for the regression dilutional bias Lancet 1990 335 765 774 1969518 10.1016/0140-6736(90)90878-9
Mortality rates after 10.5 years for participants in the Multiple Risk Factor Intervention Trial. Findings related to a priori hypotheses of the trial. The Multiple Risk Factor Intervention Trial Research Group JAMA 1990 263 1795 1801 2179590 10.1001/jama.263.13.1795
De Simone G Pasanisi F Contaldo F Link of nonhemodynamic factors to hemodynamic determinants of left ventricular hypertrophy Hypertension 2001 38 13 18 11463753
De Simone G Palmieri V Koren MJ Mensah GA Roman MJ Devereux RB Prognostic implications of the compensatory nature of left ventricular mass in arterial hypertension J Hypertens 2001 19 119 125 11204291 10.1097/00004872-200101000-00016
De Simone G Verdecchia P Pede S Gorini M Maggioni AP Prognosis of inappropriate left ventricular mass in hypertension: the MAVI Study Hypertension 2002 40 470 476 12364349 10.1161/01.HYP.0000034740.99323.8A
Verdecchia P Angeli F Borgioni C Gattobigio R De Simone G Devereux RB Porcellati C Changes in cardiovascular risk by reduction of left ventricular mass in hypertension: a meta-analysis Am J Hypertens 2003 16 895 899 14573325 10.1016/S0895-7061(03)01018-5
Sjogren AL Left ventricular wall thickness determined by ultrasound in 100 subjects without heart disease Chest 1971 60 341 346 5115855
Savage DD Garrison RJ Kannel WB Levy D Anderson SJ Stokes J IIIFeinleib M Castelli WP The spectrum of left ventricular hypertrophy in a general population sample: the Framingham Study Circulation 1987 75 I26 I33 2947749
Ganau A Devereux RB Roman MJ De Simone G Pickering TG Saba PS Vargiu P Simongini I Laragh JH Patterns of left ventricular hypertrophy and geometric remodeling in essential hypertension J Am Coll Cardiol 1992 19 1550 1558 1534335
Li L Shigematsu Y Hamada M Hiwada K Relative wall thickness is an independent predictor of left ventricular systolic and diastolic dysfunctions in essential hypertension Hypertens Res 2001 24 493 499 11675942 10.1291/hypres.24.493
Koren MJ Devereux RB Casale PN Savage DD Laragh JH Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension Ann Intern Med 1991 114 345 352 1825164
Verdecchia P Schillaci G Borgioni C Ciucci A Battistelli M Bartoccini C Santucci A Santucci C Reboldi G Porcellati C Adverse prognostic significance of concentric remodeling of the left ventricle in hypertensive patients with normal left ventricular mass J Am Coll Cardiol 1995 25 871 878 7884090 10.1016/0735-1097(94)00424-O
Krumholz HM Larson M Levy D Prognosis of left ventricular geometric patterns in the Framingham Heart Study J Am Coll Cardiol 1995 25 879 884 7884091 10.1016/0735-1097(94)00473-4
Verdecchia P Schillaci G Borgioni C Ciucci A Gattobigio R Zampi I Santucci A Santucci C Reboldi G Porcellati C Prognostic value of left ventricular mass and geometry in systemic hypertension with left ventricular hypertrophy Am J Cardiol 1996 78 197 202 8712142 10.1016/S0002-9149(96)90395-1
Ohya Y Abe I Fujii K Ohmori S Onaka U Kobayashi K Fujishima M Hyperinsulinemia and left ventricular geometry in a work-site population in Japan Hypertension 1996 27 729 734 8613232
Sundstrom J Lind L Nystrom N Zethelius B Andren B Hales CN Lithell HO Left ventricular concentric remodeling rather than left ventricular hypertrophy is related to the insulin resistance syndrome in elderly men Circulation 2000 101 2595 2600 10840010
Picca M Agozzino F Pelosi G Influence of microalbuminuria on left ventricular geometry and function in hypertensive patients with type 2 diabetes mellitus Ital Heart J 2003 4 48 52 12690921
Devereux RB Roman MJ Paranicas M O'Grady MJ Lee ET Welty TK Fabsitz RR Robbins D Rhoades ER Howard BV Impact of diabetes on cardiac structure and function: the strong heart study Circulation 2000 101 2271 2276 10811594
Palmieri V Bella JN Arnett DK Liu JE Oberman A Schuck MY Kitzman DW Hopkins PN Morgan D Rao DC Effect of type 2 diabetes mellitus on left ventricular geometry and systolic function in hypertensive subjects: Hypertension Genetic Epidemiology Network (HyperGEN) study Circulation 2001 103 102 107 11136693
Carroll JD Carroll EP Feldman T Ward DM Lang RM McGaughey D Karp RB Sex-associated differences in left ventricular function in aortic stenosis of the elderly Circulation 1992 86 1099 1107 1394918
Rohde LE Zhi G Aranki SF Beckel NE Lee RT Reimold SC Gender-associated differences in left ventricular geometry in patients with aortic valve disease and effect of distinct overload subsets Am J Cardiol 1997 80 475 480 9285661 10.1016/S0002-9149(97)00398-6
Weinberg EO Thienelt CD Katz SE Bartunek J Tajima M Rohrbach S Douglas PS Lorell BH Gender differences in molecular remodeling in pressure overload hypertrophy J Am Coll Cardiol 1999 34 264 273 10400020 10.1016/S0735-1097(99)00165-5
Nunez E Arnett DK Benjamin EJ Liebson PR Skelton TN Taylor H Andrew M Optimal threshold value for left venticular hypertrophy in blacks: the A therosclerosis risk in communities study Hypertension 2005 45 58 63 15569859
Seidell JC Visscher TL Hoogeveen RT Overweight and obesity in the mortality rate data: current evidence and research issues Med Sci Sports Exerc 1999 31 S597 S601 10593534 10.1097/00005768-199911001-00018
Zhou BF Effect of body mass index on all-cause mortality and incidence of cardiovascular diseases – report for meta-analysis of prospective studies open optimal cut-off points of body mass index in Chinese adults Biomed Environ Sci 2002 15 245 252 12500665
Kuch B Hense HW Gneiting B Doring A Muscholl M Brockel U Schunkert H Body composition and prevalence of left ventricular hypertrophy Circulation 2000 102 405 410 10908212
Kuperstein R Hanly P Niroumand M Sasson Z The importance of age and obesity on the relation between diabetes and left ventricular mass J Am Coll Cardiol 2001 37 1957 1962 11401138 10.1016/S0735-1097(01)01242-6
De Simone G Palmieri V Bella JN Celentano A Hong Y Oberman A Kitzman DW Hopkins PN Arnett DK Devereux RB Association of left ventricular hypertrophy with metabolic risk factors: the HyperGEN study J Hypertens 2002 20 323 331 11821719 10.1097/00004872-200202000-00024
Gardin JM Siscovick D Anton-Culver H Lynch JC Smith VE Klopfenstein HS Bommer WJ Fried L O'Leary D Manolio TA Sex, age, and disease affect echocardiographic left ventricular mass and systolic function in the free-living elderly. The Cardiovascular Health Study Circulation 1995 91 1739 1748 7882482
Iacobellis G Ribaudo MC Leto G Zappaterreno A Vecci E Di Mario U Leonetti F Influence of excess fat on cardiac morphology and function: study in uncomplicated obesity Obes Res 2002 10 767 773 12181385
Levy D Echocardiographically detected left ventricular hypertrophy: prevalence and risk factors. The Framingham Heart Study Ann Intern Med 1988 108 7 13 2962527
De Simone G Daniels SR Kimball TR Roman MJ Romano C Chinali M Galderisi M Devereux RB Evaluation of concentric left ventricular geometry in humans: evidence for age-related systematic underestimation Hypertension 2005 45 64 68 15557389
Burke GL ARCWWLCYBG Blood pressure and echocardiographic measures in children: the Bogalusa Heart Study Circulation 1987 75 106 114 2947739
Dannenberg AL Levy D Garrison RJ Impact of age on echocardiographic left ventricular mass in a healthy population (the Framingham Study) Am J Cardiol 1989 64 1066 1068 2530879 10.1016/0002-9149(89)90816-3
Arnett DK Rautaharju P Crow R Folsom AR Ekelund LG Hutchinson R Tyroler HA Heiss G Black-white differences in electrocardiographic left ventricular mass and its association with blood pressure (the ARIC study). Atherosclerosis Risk in Communities Am J Cardiol 1994 74 247 252 8037129 10.1016/0002-9149(94)90365-4
Koren MJ Mensah GA Blake J Laragh JH Devereux RB Comparison of left ventricular mass and geometry in black and white patients with essential hypertension Am J Hypertens 1993 6 815 823 8267936
Zabalgoitia M Ur Rahman SN Haley WE Oneschuk L Yunis C Lucas C Yarows S Krause L Amerena J Impact of ethnicity on left ventricular mass and relative wall thickness in essential hypertension Am J Cardiol 1998 81 412 417 9485129 10.1016/S0002-9149(97)00925-9
Liebson PR Grandits G Prineas R Dianzumba S Flack JM Cutler JA Grimm R Stamler J Echocardiographic correlates of left ventricular structure among 844 mildly hypertensive men and women in the Treatment of Mild Hypertension Study (TOMHS) Circulation 1993 87 476 486 8425295
Chaturvedi N Athanassopoulos G McKeigue PM Marmot MG Nihoyannopoulos P Echocardiographic measures of left ventricular structure and their relation with rest and ambulatory blood pressure in blacks and whites in the United Kingdom J Am Coll Cardiol 1994 24 1499 1505 7930282
Kizer JR Arnett DK Bella JN Paranicas M Rao DC Province MA Oberman A Kitzman DW Hopkins PN Liu JE Differences in left ventricular structure between black and white hypertensive adults: the Hypertension Genetic Epidemiology Network study Hypertension 2004 43 1182 1188 15123573 10.1161/01.HYP.0000128738.94190.9f
Gardin JM Wagenknecht LE Anton-Culver H Flack J Gidding S Kurosaki T Wong ND Manolio TA Relationship of cardiovascular risk factors to echocardiographic left ventricular mass in healthy young black and white adult men and women. The CARDIA Study Circulation 1995 92 380 387 7634452
Gottdiener JS Reda DJ Materson BJ Massie BM Notargiacomo A Hamburger RJ Williams DW Henderson WG Importance of obesity, race and age to the cardiac structural and functional effects of hypertension. The Department of Veterans Affairs Cooperative Study Group on Antihypertensive Agents J Am Coll Cardiol 1994 24 1492 1498 7930281
De Simone G Devereux RB Roman MJ Alderman MH Laragh JH Relation of obesity and gender to left ventricular hypertrophy in normotensive and hypertensive adults Hypertension 1994 23 600 606 8175168
Lips DJ deWindt LJ van Kraaij DJ Doevendans PA Molecular determinants of myocardial hypertrophy and failure: alternative pathways for beneficial and maladaptive hypertrophy Eur Heart J 2003 24 883 896 12714020 10.1016/S0195-668X(02)00829-1
Chobanian AV Bakris GL Black HR Cushman WC Green LA Izzo JL JrJones DW Materson BJ Oparil S Wright JT Jr The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report JAMA 2003 289 2560 2572 12748199 10.1001/jama.289.19.2560
2003 European Society of Hypertension-European Society of Cardiology guidelines for the management of arterial hypertension J Hypertens 2003 21 1011 1053 12777938 10.1097/00004872-200306000-00001
Galderisi M Anderson KM Wilson PWF Levy D Echocardiographic evidence for a existence of a distinct cardiomyopathy (The Framingham Heart Study) Am J Cardiol 1991 68 85 89 2058564 10.1016/0002-9149(91)90716-X
Lee M Gardin JM Lynch JC Smith V-E Tracy RP Savage PJ Szklo M Ward BJ Diabetes mellitus and echocardiographic left ventricular function in free-living elderly men and women: The Cardiovascular Health Study Am Heart J 1997 133 36 43 9006288
Feener EP King GL Vascular dysfunction in diabetes mellitus Lancet 1997 350 SI9 13 9250277
Ilercil A Devereux RB Roman MJ Paranicas M O'Grady MJ Welty TK Robbins DC Fabsitz RR Howard BV Lee ET Relationship of impaired glucose tolerance to left ventricular structure and function: The Strong Heart Study Am Heart J 2001 141 992 998 11376315 10.1067/mhj.2001.115302
Ilercil A Devereux RB Roman MJ Paranicas M O'Grady MJ Lee ET Welty TK Fabsitz RR Howard BV Associations of insulin levels with left ventricular structure and function in American Indians: the strong heart study Diabetes 2002 51 1543 1547 11978654
Wachtell K Palmieri V Olsen MH Bella JN Aalto T Dahlof B Gerdts E Wright JT JrPapademetriou V Mogensen CE Urine albumin/creatinine ratio and echocardiographic left ventricular structure and function in hypertensive patients with electrocardiographic left ventricular hypertrophy: the LIFE study. Losartan Intervention for Endpoint Reduction Am Heart J 2002 143 319 326 11835038 10.1067/mhj.2002.119895
Rutter MK Parise H Benjamin EJ Levy D Larson MG Meigs JB Nesto RW Wilson PW Vasan RS Impact of glucose intolerance and insulin resistance on cardiac structure and function: sex-related differences in the Framingham Heart Study Circulation 2003 107 448 454 12551870 10.1161/01.CIR.0000045671.62860.98
Chinali M Devereux RB Howard BV Roman MJ Bella JN Liu JE Resnick HE Lee ET Best LG De Simone G Comparison of cardiac structure and function in American Indians with and without the metabolic syndrome (the Strong Heart Study) Am J Cardiol 2004 93 40 44 14697463 10.1016/j.amjcard.2003.09.009
Manolio TA Levy D Relation of alcohol intake to left ventricular mass: The Framingham Study J Am Coll Cardiol 1991 17 717 721 1825213
Schmieder RE Messerli FH Garavaglia GE Nunez BD Dietary salt intake. A determinant of cardiac involvement in essential hypertension Circulation 1988 78 951 956 2971474
Savage DD Levy D Dannenberg AL Garrison RJ Castelli WP Association of echocardiographic left ventricular mass with body size, blood pressure and physical activity (the Framingham Study) Am J Cardiol 1990 65 371 376 2137280 10.1016/0002-9149(90)90304-J
Jullien V Gosse P Ansoborlo P Lemetayer P Clementy J Relationship between left ventricular mass and serum cholesterol level in the untreated hypertensive J Hypertens 1998 16 1043 1047 9794746 10.1097/00004872-199816070-00019
Bella JN Devereux RB Roman MJ O'Grady MJ Welty TK Lee ET Fabsitz RR Howard BV Relations of left ventricular mass to fat-free and adipose body mass: the strong heart study. The Strong Heart Study Investigators Circulation 1998 98 2538 2544 9843460
Vijayakumar M Birth weight, weight at one year, and left ventricular mass in adult life Br Heart J 1995 73 363 367 7756071
Stollberger C Hollander I Dimitrov L Slany J Influence of measurement inaccuracies on determination of left ventricular mass by M mode echocardiography Heart 1996 75 312 313 8800999
de Leonardis V Cinelli P Evidence of no interobserver variability in M-mode echocardiography Clin Cardiol 1986 9 324 326 3731555
Collins HW Kronenberg MW Byrd BF III Reproducibility of left ventricular mass measurements by two-dimensional and M-mode echocardiography J Am Coll Cardiol 1989 14 672 676 2768716
Fast JH Limits of reproducibility of left ventricular wall thickness and mass by M-mode echocardiography Neth J Med 1989 34 297 301 2770940
Lantelme P Bouchayer D Gayet C Lievre M Gessek J Milon H Influence of a rapid change of left ventricular dimensions on the echocardiographic measurement of left ventricular mass by the Penn convention J Hypertens 1999 17 1323 1328 10489111 10.1097/00004872-199917090-00013
Palmieri V Dahlof B DeQuattro V Sharpe N Bella JN De Simone G Paranicas M Fishman D Devereux RB Reliability of echocardiographic assessment of left ventricular structure and function: the PRESERVE study. Prospective Randomized Study Evaluating Regression of Ventricular Enlargement J Am Coll Cardiol 1999 34 1625 1632 10551715 10.1016/S0735-1097(99)00396-4
De Simone G Muiesan ML Ganau A Longhini C Verdecchia P Palmieri V Agabiti-Rosei E Mancia G Reliability and limitations of echocardiographic measurement of left ventricular mass for risk stratification and follow-up in single patients: the RES trial. Working Group on Heart and Hypertension of the Italian Society of Hypertension. Reliability of M-mode Echocardiographic Studies J Hypertens 1999 17 1955 1963 10703895 10.1097/00004872-199917121-00027
Dadgar SK Tyagi SP Singh RP Hameed S Factors influencing the normal heart weight – a study of 140 hearts Jpn Circ J 1979 43 77 82 449022
Hangartner JR Marley NJ Whitehead A Thomas AC Davies MJ The assessment of cardiac hypertrophy at autopsy Histopathology 1985 9 1295 1306 2937709
Murphy ML White HJ Meade J Straub KD The relationship between hypertrophy and dilatation in the postmortem heart Clin Cardiol 1988 11 297 302 2968221
de la Grandmaison GL Clairand I Durigon M Organ weight in 684 adult autopsies: new tables for a Caucasoid population Forensic Sci Int 2001 119 149 154 11376980 10.1016/S0379-0738(00)00401-1
Broughton A Korner PI Left ventricular pump function in renal hypertensive dogs with cardiac hypertrophy Am J Physiol 1986 251 H1260 H1266 2947477
Linzbach AJ Hypertrophy, hyperplasia and structural dilatation of the human heart Adv Cardiol 1976 18 1 14 136171
Hradec J Aschermann M Vancura J Petrasek J Drechslerova J Michaljanic A Jezek V Comparison of M-mode echocardiography and angiography in the evaluation of left ventricular hypertrophy Cor Vasa 1986 28 341 349 2947778
Williams KA Lang RM Reba RC Taillon LA Comparison of technetium-99m sestamibi-gated tomographic perfusion imaging with echocardiography and electrocardiography for determination of left ventricular mass Am J Cardiol 1996 77 750 755 8651128 10.1016/S0002-9149(97)89211-9
Maruyama K Hasegawa S Nakatani D Paul AK Kusuoka H Nishimura T Hori M Hatazawa J Left ventricular mass index measured by quantitative gated myocardial SPECT with 99mTc-tetrofosmin: a comparison with echocardiography Ann Nucl Med 2003 17 31 39 12691128
Mousseaux E Beygui F Fornes P Chatellier G Hagege A Desnos M Lecomte D Gaux JC Determination of left ventricular mass with electron beam computed tomography in deformed, hypertrophic human hearts Eur Heart J 1994 15 832 841 8088273
Missouris CG Forbat SM Singer DR Markandu ND Underwood R MacGregor GA Echocardiography overestimates left ventricular mass: a comparative study with magnetic resonance imaging in patients with hypertension J Hypertens 1996 14 1005 1010 8884556
Pons LG Carreras F Borras X Llauger J Palmer J [Comparative study of echocardiography and magnetic resonance imaging in the assessment of left ventricular mass] Rev Esp Cardiol 2001 54 22 28
Grothues F Smith GC Moon JC Bellenger NG Collins P Klein HU Pennell DJ Comparison of interstudy reproducibility of cardiovascular magnetic resonance with two-dimensional echocardiography in normal subjects and in patients with heart failure or left ventricular hypertrophy Am J Cardiol 2002 90 29 34 12088775 10.1016/S0002-9149(02)02381-0
Bottini PB Carr AA Prisant LM Flickinger FW Allison JD Gottdiener JS Magnetic resonance imaging compared to echocardiography to assess left ventricular mass in the hypertensive patient Am J Hypertens 1995 8 221 228 7794570 10.1016/0895-7061(94)00178-E
Picano E Economic and biological costs of cardiac imaging Cardiovasc Ultrasound 2005 3 13 15916702 10.1186/1476-7120-3-13
Chuang ML Beaudin RA Riley MF Mooney MG Mannin WJ Douglas PS Hibberd MG Three-dimensional echocardiographic measurement of left ventricular mass: comparison with magnetic resonance imaging and two-dimensional echocardiographic determinations in man Int J Card Imaging 2000 16 347 357 11215919 10.1023/A:1026540809758
Thomson HL Basmadjian AJ Rainbird AJ Razavi M Avierinos JF Pellikka PA Bailey KR Breen JF Enriquez-Sarano M Contrast echocardiography improves the accuracy and reproducibility of left ventricular remodeling measurements: a prospective, randomly assigned, blinded study J Am Coll Cardiol 2001 38 867 875 11527647 10.1016/S0735-1097(01)01416-4
Folsom AR Chambless LE Duncan BB Gilbert AC Pankow JS Prediction of coronary heart disease in middle-aged adults with diabetes Diabetes Care 2003 26 2777 2784 14514579
|
15963236
|
PMC1183230
|
CC BY
|
2021-01-04 16:38:31
|
no
|
Cardiovasc Ultrasound. 2005 Jun 17; 3:17
|
utf-8
|
Cardiovasc Ultrasound
| 2,005 |
10.1186/1476-7120-3-17
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-111595525310.1186/1476-069X-4-11ResearchSpatial analysis of lung, colorectal, and breast cancer on Cape Cod: An application of generalized additive models to case-control data Vieira Verónica [email protected] Thomas [email protected] Janice [email protected] Ann [email protected] David [email protected] Department of Environmental Health, Boston University School of Public Health, 715 Albany Street, Boston, MA 02118, USA2 Department of Biostatistics, Boston University School of Public Health, 715 Albany Street, Boston, MA 02118, USA3 Department of Epidemiology, Boston University School of Public Health, 715 Albany Street, Boston, MA 02118, USA2005 14 6 2005 4 11 11 23 11 2004 14 6 2005 Copyright © 2005 Vieira et al; licensee BioMed Central Ltd.2005Vieira 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 availability of geographic information from cancer and birth defect registries has increased public demands for investigation of perceived disease clusters. Many neighborhood-level cluster investigations are methodologically problematic, while maps made from registry data often ignore latency and many known risk factors. Population-based case-control and cohort studies provide a stronger foundation for spatial epidemiology because potential confounders and disease latency can be addressed.
Methods
We investigated the association between residence and colorectal, lung, and breast cancer on upper Cape Cod, Massachusetts (USA) using extensive data on covariates and residential history from two case-control studies for 1983–1993. We generated maps using generalized additive models, smoothing on longitude and latitude while adjusting for covariates. The resulting continuous surface estimates disease rates relative to the whole study area. We used permutation tests to examine the overall importance of location in the model and identify areas of increased and decreased risk.
Results
Maps of colorectal cancer were relatively flat. Assuming 15 years of latency, lung cancer was significantly elevated just northeast of the Massachusetts Military Reservation, although the result did not hold when we restricted to residences of longest duration. Earlier non-spatial epidemiology had found a weak association between lung cancer and proximity to gun and mortar positions on the reservation. Breast cancer hot spots tended to increase in magnitude as we increased latency and adjusted for covariates, indicating that confounders were partly hiding these areas. Significant breast cancer hot spots were located near known groundwater plumes and the Massachusetts Military Reservation.
Discussion
Spatial epidemiology of population-based case-control studies addresses many methodological criticisms of cluster studies and generates new exposure hypotheses. Our results provide evidence for spatial clustering of breast cancer on upper Cape Cod. The analysis suggests further investigation of the potential association between breast cancer and pollution plumes based on detailed exposure modeling.
==== Body
Background
Local disease mapping ("cluster") investigations are often desired by concerned communities, but many epidemiologists resist the pressure to search for environmental causes of clusters. Critics argue that such studies are unproductive and flawed because they often combine unrelated diseases, apply arbitrary or even "gerrymandered" boundaries, contain insufficient numbers of cases, and ignore population density, latency, and known risk factors [1]. Data based on cancer registries are generally mapped by town of diagnosis (or other geographic unit) and contain limited data on covariates. This results in poor spatial resolution, potential spatial confounding, and the inability to consider latency. Spatial confounding occurs when risk factors for a disease are not evenly distributed, e.g., a cluster of lung cancer may be due to an increased density of smokers. Since cancer typically takes many years to develop, residence at diagnosis is likely to be a poor measure of exposure. Maps that ignore latency may tend to be flatter if population movement is random with respect to disease status [2]. Nevertheless, cluster investigations can be an important part of responding to public concerns, even if no new etiologic knowledge is gained [3,4].
In 1988, an elevated cancer incidence in the Upper Cape Cod region of Massachusetts (Figure 1) prompted a series of epidemiological studies to investigate possible environmental risk factors, including air and water pollution associated with the Massachusetts Military Reservation (MMR), pesticide applications to cranberry bogs, particulate air pollution from a large electric power plant, and tetrachloroethylene-contaminated drinking water from vinyl-lined asbestos cement distribution pipes [5-15]. Positive associations were observed, but the environmental exposures explained only a portion of the excess cancer incidence. These studies provide an invaluable data set for spatial analysis. Population-based case-control studies can provide detailed information on individual-level covariates and residential history. Cases are identified using cancer registries while controls provide an estimate of the underlying population density. Subjects or next-of-kin are interviewed to obtain relevant data on covariates and residential history. Geocoding of this information produces a rich, point-based data set that can be analyzed with the help of geographical information systems (GIS).
Figure 1 Geographic location of the upper Cape Cod study area. Cape Cod is located in Massachusetts in the northeast United States.
Methods for mapping point-based epidemiologic data have received less attention than mapping areal data [16]. Generalized additive models (GAMs), a type of statistical model that combines smoothing with the ability to analyze binary outcome data and adjust for covariates, provide a useful framework for examining such point data [17-19], Webster et al. submitted. Using individual-level information and location in a generalized additive model, we calculated the crude and adjusted odds ratios for lung, colorectal, and breast cancers on Upper Cape Cod assuming different latency periods. These analyses have several objectives: i) to test if the disease maps are flat, ii) to determine if areas of increased or decreased risk are due to spatial confounding, iii) to examine the effect on the maps of increasing latency, iv) to suggest exposure hypotheses for further investigation, and v) to demonstrate spatial epidemiology using generalized additive models.
Methods
Study Population
We investigated the association between residence and breast, lung and colorectal cancer on Upper Cape Cod, Massachusetts (USA) using data from population-based case-control studies [10-12]. The Massachusetts Cancer Registry was used to identify incident breast cancer cases diagnosed from 1983–1993 and incident cases of lung and colorectal cancers diagnosed from 1983–1986. Participants were restricted to permanent residents of the upper Cape region with complete residential histories. A total of 638 breast cancer cases, 243 lung cancer cases, and 309 colorectal cancer cases were included.
Controls were chosen to represent the underlying population that gave rise to the cases, i.e., permanent residents of the same towns during the same time period. Controls were frequency matched to cases on age, gender, and vital status. Because many of the cases were deceased or elderly, three different sources of controls were used: (1) random digit dialing for living controls less than 65 years of age; (2) Centers for Medicare and Medicaid Services (formerly the Health Care Financing Administration) for the living population 65 years of age or older; and (3) death certificates for controls who had died from 1983 onward. There were 842 breast cancer controls, 1205 lung cancer controls, and 1138 colorectal cancer controls.
Participants or their next-of-kin completed an extensive interview, providing information on demographics (age, sex, marital status, and education), a forty-year residential history, and potential confounders. "Index years" were randomly assigned to controls in a distribution similar to that of diagnosis years for cases. We used index years to estimate length and time of environmental exposure for controls in a fashion comparable to that of cases. See earlier papers [10-12] for a detailed description of the methods used to define the study population, including the rationale for the method of control selection. The Institutional Review Board of Boston University Medical Center approved the research.
Geographical Information System (GIS)
All residential addresses reported by participants in the upper Cape Cod area over the forty-year period prior to the diagnosis or index year were eligible for spatial analysis. We excluded all addresses where residency time began after diagnosis date for cases or index date for controls. The breast cancer data set included 638 cases representing 1061 residential locations and 842 controls representing 1371 locations. The lung cancer data set included 243 cases representing 385 residential locations and 1205 controls representing 1927 residential locations. The colorectal cancer data set included 309 cases representing 469 residential locations and 1138 controls representing 1791 residential locations. Thus, individual participants may have contributed more than one address.
Locations of the participant residences were geocoded using the Massachusetts State Plane Coordinate System with North American Datum 1983 (NAD1983) and linked to the participant's interview data. Geocoding, the process where longitude and latitude are determined for each street address, was done without knowledge of case status, and the final data were checked for accuracy. GIS allows us to map the coordinates of participants and link the location to additional individual and environmental information. Figure 2 shows the distribution of lung, colorectal, and breast cancer cases and controls in the study area. To preserve confidentiality, the figure was created by randomly placing residences within a small grid that includes the actual location. Actual locations were used in the analysis.
Figure 2 Distribution of cases and controls for lung, colorectal, and breast cancer. Each point represents the residence of one participant. Locations have been geographically altered to preserve confidentiality.
Statistical methods for mapping point-based epidemiologic data
Statistical methods for mapping area-based epidemiologic data, e.g., disease rates by town or county, are well advanced [20]. Mapping such data often has two main components: adjusting for covariates, often via standardization, and contending with geographically varying degrees of precision, often by smoothing. Methods for point-based data are less well developed [16]. One approach uses kernel methods to estimate the density of cases and the density of the population giving rise to the cases [21,22]. Their ratio provides an estimate of the rate. Alternatively, one estimates the density of the population using controls. When controls are appropriately sampled from the population of a geographic area, the case/control ratio – disease odds – in a sub-area should be proportional to the disease incidence rate in that sub-area. Unfortunately, the density ratio approach provides no easy method to adjust for covariates [23]. Other multi-step methods have been suggested [24,25].
At least two methods provide unified frameworks for mapping point-based epidemiologic data, adjusting for covariates, and hypothesis testing: generalized linear mixed model formulations of kriging [15,23,26,27] and generalized additive models (GAMs) using bivariate kernel or loess smoothers [18,19,23]. Both are promising but relatively untried methods in spatial epidemiology. For example, kernel-based GAMs have been used to map risks of lung cancer [18], biliary cirrhosis [28], and infant mortality [29].
Mapping via generalized additive models (GAMs)
We estimated local disease odds using generalized additive models, a form of non-parametric or semi-parametric regression with the ability to analyze binary outcome data while adjusting for covariates [17]. We modeled location, a potential proxy measure of exposure, using a bivariate smooth (S) of latitude (x1) and longitude (x2)
logit [p(x1,x2)] = S(x1,x2) + γ 'z (1)
where the left-hand side is the log of the disease odds at location (x1,x2), z is a vector of covariates, and γ is a vector of parameters. The model is semiparametric because it has both nonparametric and parametric components. Without the smooth function, S(x1,x2), the model becomes an ordinary logistic regression on the covariates. Omitting the covariates produces a crude (unadjusted) map. We used a loess smooth which adapts to changes in population density [17]. The amount of smoothing depends on the percentage of the data points in the neighborhood, referred to as the span size. GAMs also allow selection of "optimal" span size and hypothesis testing. Webster et al. (submitted) provides a detailed discussion of the statistical methods, analyses using synthetic data, and a comparison with the kernel method of Kelsall and Diggle [18]. We used S-Plus [30] to perform the generalized additive modeling and ArcView [31] to map the results of our analyses. Program code is available on request.
We determined the optimal amount of smoothing for each map by minimizing the Akaike's Information Criterion (AIC). Small span sizes produce bumpier surfaces and larger span sizes produce smoother surfaces. As the span size increases, the amount of bias in the fit increases and the variance decreases [17]. We created a rectangular grid covering the study area using the minimum and maximum latitude and longitude coordinates from the original data set. Grid points lying outside the outline map of the study area were clipped, as were areas where people cannot live (e.g., ocean or wildlife refuges). We estimate the crude and adjusted log odds at each location on the grid using the S-Plus function predict.gam. As this function defines neighborhoods based on a combination of the data points and the grid, it can produce discrepancies from predictions based on the original data alone [32,33]. We therefore checked all maps and found that any discrepancies were minor, not changing our conclusions.
We converted from log odds to odds ratios (ORs) using the whole study area as the reference, dividing the odds at each grid point by the odds calculated by the reduced model omitting the location smoothing term. The odds ratio estimates the rate ratio and relative risk. In order to make maps visually comparable, we mapped all results using the same dark blue to dark red continuous (unclassified) color scale and range of odds ratios, 0.25–2.50. This range covers most but not all of the ORs observed in our analyses, preventing maps from being washed out by an area of extremely high or low ORs. We used a linear scale for odds ratios; although a log scale is a good option, it may be more difficult for many people to interpret. As odds ratios near unity appear as a light green, this scale is close to divergent, an effective way to communicate deviations of the map from flatness in both directions. Spectral scales, such as ours, are useful when there is a clear central value – here, an odds ratio of one – from which divergence is important (See Brewer et al [34] for a useful discussion of color schemes). Cancer maps have historically used blue and red for areas of low and high rates [35]. Blue and red are commonly associated with cold and hot, aiding interpretability of areas with decreased or increased risk.
We determined the presence of spatial confounding by visually comparing crude and adjusted maps. If their optimal span sizes differ, we also compared maps using a common span, allowing us to distinguish between changes due to adjustment and changes due to span.
GAMs also provide a framework for testing hypotheses. There are a number of ways to test the global null hypothesis that disease status does not depend on location, i.e., that the map is flat. Similar to analysis of variance in ordinary linear regression, we examined the overall significance of location using the difference in deviance of the complete model (equation 1) and the reduced model omitting the smoothing term. The S-Plus software provides an approximate p-value for this statistic assuming a chi square distribution. Because the latter assumption is in general not true for GAMs [17], we calculated the p-value using a permutation test. To test the null hypothesis of no association between case/control status and location, we randomly reassigned individuals to the eligible residences. This relabeling procedure preserves the number of cases and controls and the relationship between case/control status and covariates, but any deviation from a flat map is due to chance. We sampled from the null permutation distribution 999 times in addition to the original. For each permutation, we ran the GAM using the optimal span of the original data and computed the deviance statistic. We divided the rank of the observed value by 1000 to obtain the approximate permutation p-value. For comparison, we also computed a permutation p-value for the global statistic used by Kelsall and Diggle [18].
If the deviance global statistic indicated that location was significant at the 0.05 level, we conducted pointwise permutation tests to identify areas with significantly increased or decreased risk. We obtained a distribution of the log odds at every point using the same set of permutations we used for calculating the global statistics. The areas of significantly decreased risk ("cold spots") include all points that rank in the lower 2.5% of the pointwise distributions. Areas of significantly elevated risk ("hot spots") include all points that rank in the upper 2.5% of the pointwise distributions. By drawing the 2.5% and 97.5% contour lines, we mapped areas of significantly decreased and increased risk.
Covariates and Missing Data
A group of core confounders, chosen a priori based on the current scientific literature or study design, was included in all adjusted analyses of breast cancer: time period of case ascertainment and vital status at interview, age at diagnosis or index year, family history of breast cancer, personal history of breast cancer (before current diagnosis or index year), age at first live birth or stillbirth, and occupational exposure to solvents. A number of other covariates were retained because they changed the appearance of the map: history of benign breast cancer, race, body mass index, history of radiation exposure, and alcohol use. We dropped other covariates from the model because they did not change the appearance of the map, including past use of diethylstilbestrol (DES), oral contraceptives and menopausal hormones, history of cigarette smoking, marital status, religion, education level, exposure to tetrachloroethylene from distribution pipes, and physical activity level. Lung cancer data were adjusted for age at diagnosis or index year, sex, vital status at interview, smoking (cigarettes, pipe, and cigars), living with a smoker, occupational exposure to lung carcinogens (jobs with arsenic, asbestos, chromium, coal tar pitch exposure), and exposure to radiation. Dropped covariates included alcohol history, use of pesticides/herbicides in the garden, exposure to tetrachloroethylene from distribution pipes, and whether the residence had been treated for termites. For colorectal cancer, we adjusted for age at diagnosis or index year, sex, vital status at interview, history of inflammatory bowel disease, and occupational history associated with colorectal cancer (jobs with asbestos or solvent exposure). History of alcohol use and radiation exposure did not affect the appearance of the maps.
We restricted analysis to subjects with complete residential histories. In our initial analyses, subjects missing data for other covariates were included in the analyses but those variables were coded as missing using an indicator variable [36]. While this method is often adequate in our experience, it can theoretically lead to bias [37]. Therefore, to ensure that positive results were not biased by the use of the indicator method, we used multiple imputation for variables with over 10% missing data. The amount of missing data was less than 10% per variable for lung cancer (15 year latency analysis) and colorectal cancer (no latency analysis). Most breast cancer covariates (20 year latency analysis) had less than 10% missing data. Exceptions were family history of breast cancer (10%), personal history of benign breast cancer (10%), history of oral contraceptive use (11%), history of radiation exposure (13%), menopausal hormone treatment (19%) and past use of DES (20%). For breast cancer (20 year latency), we imputed six complete data sets, and then ran the GAM model and statistics on each. We combined the six maps by pointwise averaging of odds ratios prior to exponentiation.
Residential History
Our initial, no-latency analyses included all eligible residences, i.e., exposures occurring up to diagnosis were assumed to contribute to the risk of disease. However, cancers initiated by exposure to environmental carcinogens typically take more than a decade to develop. We therefore performed a fifteen-year latency analysis by restricting inclusion to the residences occupied by participants at least fifteen years prior to the diagnosis or index year (Residences within the fifteen year window were excluded because geographical location within that window was assumed not relevant to outcome). Between 46% and 48% of residences remained depending on the outcome. In addition, because breast cancer cases were obtained over a ten-year period, there were sufficient cases to perform a twenty-year latency analysis; 37% of residences remained eligible.
Some participants lived at more than one location on Cape Cod or more than once at a single location if they moved away and later returned. To determine the effects multiple residences may have had on maps, we also performed analyses that included for each individual only the residence of longest duration that met latency assumptions. Since the resulting data set is smaller, the optimal span chosen by the AIC is often, but not always, larger. As maps can change due to different span sizes, we also analyzed the reduced data set using the optimal span of the original data.
Results
Breast Cancer
Assuming no latency, location was not statistically significant at the 0.05 level (Table 1 and Figure 3a). Assuming 15 years produced a statistically significant, although still relatively flat map (Figure 3b). Assuming 20 years of latency increased the magnitude of the hot and cold spots (Figure 3c) and the overall significance of the map (Table 1). The adjusted map (Figure 3c) had more pronounced hot and cold spots than the crude map (Figure 3d). Spatial confounding was thus partially masking differences in the crude analysis. Race was the single most important variable responsible for this difference; at the time, there was a large population of Native Americans living in upper Cape Cod. The point-wise tests of significance showed a large hot spot that spans the towns of Falmouth, Mashpee, and southern Sandwich (Figure 3e). Other hot spots were identified in southeastern Barnstable and northwestern Bourne. Disease odds in certain areas were five times higher than the study area as a whole. Areas of significantly decreased risk relative to the whole study area were scattered along the southern coast of Falmouth and Mashpee and through the center of Barnstable. For individuals in the breast cancer analysis assuming 20 years of latency, 66% had only one eligible residence, 22% had two, 8% had three, and 4% had four or more.
Table 1 Summary of breast cancer models
Analysis Latency (yrs) Spana Cases/ Controls S-Plus p-valuec Deviance p-valuec Kelsall/ Diggle p-valuec Figure #
Adjusted All Residences 0 0.50 1061/ 1371 0.053 0.101 0.198 3a
Adjusted All Residences 15 0.35 528/ 650 0.004 0.010 0.016 3b
Crude All Residences 20 0.35 391/ 509 0.0008 0.003 0.003 ---
Crude All Residences 20 0.15b 391/ 509 ---d ---d ---d 3d
Adjusted All Residences 20 0.15 391/ 509 5.6E-6 0.001 0.006 3c,e 5a, 9
Adjusted Longest Duration 20 0.15b 248/ 341 ---d ---d ---d 4a
Adjusted Longest Duration 20 0.45 248/ 341 0.008 0.020 0.029 4b
a Optimal span obtained by using the Akaike's Information Criterion (AIC) unless otherwise noted.
b Same span as for adjusted, all residences.
c Null hypothesis is that the map is flat. S-Plus p-value is only approximate.
d The p-values were omitted because the maps were created using a non-optimal span.
Figure 3 Breast Cancer Results. Odds ratios are relative to the whole study area. a) Adjusted, no latency. b) Adjusted. Assuming 15 years of latency somewhat increases spatial variation. c) Adjusted, 20 years of latency. Further increasing latency increases magnitude of hot and cold spots. d) Crude, 20 years of latency, created using the optimal span (0.15) of the adjusted map. Difference from the adjusted map indicates spatial confounding. e) Adjusted, 20 years of latency. Black contour lines denote areas of significantly increased and decreased risk at the 0.05 level.
We next restricted the adjusted 20-year latency analysis to residences of longest duration. One third lived at their residence of longest duration for less than 20 years, 37% for 20–29 years and 30% for 30 or more years. Using the same span size as before (0.15), Figure 4a shows that although cluster size and shape has changed, the overall spatial pattern remained quite similar. The optimal span for the longest duration analysis was 0.45, and using the larger span size results in a smoother surface (Figure 4b). Hot spots and cold spots coalesced, reducing their magnitudes, but they remain statistically significant.
Figure 4 Breast Cancer Results, Restricted to Longest Duration Residences. a) Adjusted, 20 years of latency. Restriction to residences of longest duration has little effect when the same span (0.15) is used as for all residences (Figure 3e). b) Use of the optimal span (0.45) for the restricted analysis increases the smoothness of the map.
Compared with the original map produced using indicator variables for missing data (5a), multiple imputation had only minor effects on the appearance of the 20 year latency analysis for all residences (Figure 5b); as all six imputed maps (and their average) look virtually identical, we show only one. However, the optimal span was 0.35 for the imputed maps, larger than the span of 0.15 for the original map. At this higher span, the imputed map appears smoother (Fig. 5c). Comparison of the AIC curves for the original and imputed maps (virtually identical for the six imputed data sets) indicates that both have two local minima at spans 0.15 and 0.35 (Fig. 6a, 6b). Although quite similar in magnitude, the AIC at span 0.15 is slightly smaller for the original data set while the AIC at span 0.35 is slightly smaller for the imputed data sets. From a statistical point of view, both span sizes appear to be appropriate. However, because of the low population density around the military base, use of the larger span size tends to merge two "hot spots" in the center and the northwest corner of the map (Fig. 5b, 5c). The global statistics for the imputed maps were highly significant regardless of span size.
Figure 5 Multiple Imputation of Missing Data had Little Effect on Breast Cancer Results. Adjusted, 20 years of Latency. a) Breast cancer map estimated using indicator variables to signify missing covariate data (Fig. 3c). b) We imputed missing data for covariates missing 10% or more of values. We generated six data sets, applying the GAM model to each. All maps (and their average) looked virtually identical; only one is shown, drawn using the same span (0.15) as the non-imputed map in a. c) Imputed map drawn using its optimal span of 0.35. Since the span is larger, it appears smoother than in b. The global statistics for all imputed maps were highly significant, regardless of span size. Black contour lines denote areas of significantly increased and decreased risk at the 0.05 level.
Figure 6 AIC Curves for the Imputed and Non-Imputed Breast Cancer Maps. Adjusted, 20 years of Latency. a) AIC curve for the imputed map. b) AIC curve for the non-imputed map. Both curves have local minima at span sizes of 0.15 and 0.35. Although quite similar in magnitude, the AIC value at 0.35 is slightly smaller than the value at 0.15 for the imputed map; the reverse is true for the non-imputed map. From a statistical point of view, both span sizes appear appropriate.
Lung Cancer
Hot and cold spots became apparent as we increased latency from 0 to 15 years (Table 2 and Figures 7a, b). Adjusting for covariates increased odds ratios in the northern part of the map (Compare Figures 7b, c). Location was statistically significant for 15 years of latency. The northern region of upper Cape Cod and southern Barnstable were areas of significant increased risk relative to the entire study area, while small areas of Falmouth had areas of significant decreased risk (Figure 7d). For individuals in the lung cancer analysis assuming 15 years of latency, 61% had only one eligible residence, 22% had two, 11% had three, and 5% had four or more.
Table 2 Summary of lung cancer models
Analysis Latency (yrs) Spana Cases/ Controls S-Plus p-valuec Deviance p-valuec Kelsall/ Diggle p-valuec Figure #
Adjusted All Residences 0 0.95 385/ 1927 0.056 0.072 0.080 7a
Crude All Residences 15 0.30 182/ 871 0.005 0.016 0.101 7c
Adjusted All Residences 15 0.30 182/ 871 0.004 0.004 0.029 7b,d
Adjusted Longest Duration 15 0.30b 100/ 485 ---d ---d ---d 7e
Adjusted Longest Duration 15 0.95 100/ 485 0.227 0.351 0.381 7f
a Optimal span obtained by using the Akaike's Information Criterion (AIC) unless otherwise noted.
b Same span as for all residences.
c Null hypothesis is that the map is flat. S-Plus p-value is only approximate.
d The p-values were omitted because the maps were created using a non-optimal span.
Figure 7 Lung Cancer Results. Odds ratios are relative to the whole study area. a) Adjusted, no latency. b) Adjusted, 15 years of latency. Increasing latency increases magnitude of hot and cold spots. c) Crude, 15 years of latency. Difference from the adjusted map indicates spatial confounding. The crude and adjusted maps have the same optimal span. d) Adjusted, 15 years of latency. Black contour lines denote areas of significantly increased and decreased risk at the 0.05 level. e) Adjusted, 15 years of latency. Restriction to residences of longest duration greatly changes the map compared to results for all residences even when the same span is used (0.3). f) Use of the optimal span (0.95) for the restricted analysis produces a very flat map.
For the adjusted 15-year latency analysis restricted to residences of longest duration, 23% of the subjects lived at their residence for less than 20 years, 18% for 20–29 years and 30% for 30 or more years. Restricting the 15-year latency analysis to residences of longest duration changed the map, eliminating the significant hot spots and the global significance of location (Figure 7e, drawn using the same span as Figure 7d). This result may imply that inclusion of multiple residences biased the non-restricted analysis. The optimum span for the longest duration analysis increased from 0.30 to 0.95, producing a map rather flat in appearance (Figure 7f). The increased span size may be due in part to the decreased amount of data.
Colorectal Cancer
The maps for colorectal cancer showed less variation in odds ratios than those for breast and lung cancer. Location was not statistically significant except at the less plausible assumption of no latency (Table 3). Little change was seen in the odds ratios when latency was increased from 0 to 15 years (Figure 8). Neither adjusting for covariates nor restricting to residences of longest duration had much effect on the 15-year latency analysis (maps not shown).
Table 3 Summary of colorectal cancer models.
Analysis Latency (yrs) Spana Cases/ Controls S-Plus p-valuec Deviance p-valuec Kelsall/ Diggle p-valuec Figure #
Adjusted All Residences 0 0.35 469/ 1791 0.002 0.006 0.013 8a
Crude All Residences 15 0.60 203/ 854 0.090 0.162 0.116 ---
Adjusted All Residences 15 0.60 203/ 854 0.067 0.120 0.157 8b
Adjusted Longest Duration 15 0.60b 112/ 488 ---d ---d ---d ---
Adjusted Longest Duration 15 0.95 112/ 488 0.239 0.360 0.158 ---
a Optimal span obtained by using the Akaike's Information Criterion (AIC) unless otherwise noted.
b Same span as for all residences.
c Null hypothesis is that the map is flat. S-Plus p-value is only approximate.
d The p-values were omitted because the maps were created using a non-optimal span.
Figure 8 Colorectal Cancer Results. Increasing latency from 0 years (a) to 15 years (b) shows little effect on adjusted odds ratios. Odds ratios are relative to the whole study area.
Comparison of Global Statistics
As shown in Tables 1, 2, 3, the p-values computed for the permutation-based deviance test and the Kelsall-Diggle statistic were typically similar, with the former usually slightly smaller. The p-value provided by S-Plus for the deviance statistic using a chi square assumption was smaller, sometimes much smaller, than the permutation-based deviance test. While the chi square approximation provides a rough approximation, we recommend use of the permutation-based approach.
Discussion
In our analyses, the maps of lung and breast cancer on upper Cape Cod displayed more variation when we controlled for covariates and increased latency. Location also became a statistically more important part of the model. Rather than causing disease clusters as is often assumed, spatial confounding was partially hiding areas of increased risk. If population movement is random with respect to disease status, ignoring latency should cause nondifferential exposure misclassification and tend to make maps flatter. Areas of increased and decreased breast cancer risk became more pronounced, and the maps became more statistically significant, when we increased latency from 0 to 15 to 20 years. The trend towards greater spatial variation in both breast and lung cancer with increased latency is consistent with misclassification of geographically associated risk factors, including environmental exposures. Alternatively, people who lived on the Cape for many years may have personal risk factors that we did not control. A recent non-spatial analysis also found that breast cancer risk was associated with long residence on Cape Cod [13]. In contrast, the maps of colorectal cancer were relatively flat.
A number of epidemiology studies have examined cancer and environmental exposures on Cape Cod [5-15]. Brody et al. [14] recently reported no association between breast cancer and wide-area application of pesticides, assessed using historical records and GIS. Modest increased risks were associated with aerial application of persistent pesticides to cranberry bogs and use of less persistent pesticides for agriculture and tree pests. Previous studies investigated the association between breast, lung and colorectal cancer and tetrachloroethylene in drinking water from vinyl-lined asbestos cement distribution pipes [10-12]. Moderately increased risks were found for breast and lung cancer in the most exposed individuals. In our analysis of breast cancer with twenty years of latency, 12 out of 900 residences were exposed to tetrachloroethylene from pipes, only one within a significant hot spot. For lung cancer assuming 15 years of latency, 8 out of 1053 residences were exposed, only two within a significant hot spot. Adding tetrachloroethylene to the models had no effect on the appearance of either map.
Our analysis located a significant lung cancer "hot spot" north of the Massachusetts Military Reservation (Compare Figures 1 and 7d). Earlier research had found a modest increased risk of lung cancer within 3 km of gun and mortar training sites on the military base [7]. We also found a significant breast cancer hot spot on the southeastern edge of the MMR. French and Wand [15] reported an area of increased risk for prostate cancer southeast of the MMR. Others found suggestions of a link between low birth weight and proximity to the base [27].
Overlaying maps of odds ratios with maps of pollution sources can generate hypotheses about exposure. Caution is needed, however, because many geographic features may overlap. To generate hypotheses for further investigation, we looked in a Massachusetts online repository of geographically coded features for shape files potentially related to environmental exposure [38]. Groundwater plumes were of particular interest because of earlier hypotheses that breast cancer might be related to pollution of drinking water. With no prior knowledge of any geographic relationship to breast cancer, we compared the two data sets (Figure 9), and found a suggestive overlap between the three significant breast cancer hot spots and ground water plumes, some from the MMR. Since the plumes likely did not have the same positions during the exposure period (assuming latency) and subjects variously used private wells or public water, this concordance does not establish exposure. However, this hypothesis could be tested by identifying participants' drinking water sources and comparing years of residency to the years of possible contamination.
Figure 9 Groundwater plumes, the Massachusetts Military Reservation (MMR), and significant breast cancer hot spots. Adjusted, 20 years of Latency. Odds ratios are relative to the whole study area. a) Breast cancer map estimated using indicator variables to signify missing covariate data with an optimal span of 0.15 (Fig. 3c). b) Imputed map drawn using its optimal span of 0.35 (Fig. 5c). c) Location of the MMR and groundwater plumes from the MMR and other sources such as landfills. From a statistical point of view, both span sizes appear to be appropriate. However, because of the low population density around the military base, use of the larger span size tends to merge two "hot spots" in the center and the northwest corner of the map.
Case-control studies are one of the standard epidemiologic tools for investigating associations between disease and exposure. By combining such data with advanced statistical techniques, we were able to address many criticisms of spatial studies. Relatively large numbers of cancer cases were ascertained from a registry and cancer types were studied separately. Point-based data from a region were used, avoiding aggregation within arbitrary political boundaries. Controls provided an estimate of the underlying, non-uniform population density. We were able to control for many covariates not available in studies that rely on registry data alone. Residential history information allowed us to take latency into account, potentially quite important for diseases like cancer.
Nevertheless, our results have a number of potential limitations. Residential locations do not account for daily movement of individuals. For breast cancer, there is a possibility that areas of elevated disease risk are due to screening bias: Women in the underlying population may have had greater opportunity for screening in these areas. We therefore examined the association between location and whether controls had undergone mammography, adjusting for age and family history of breast cancer (Mammography data were only available for the non-proxy controls). The resulting map was relatively flat and different in appearance from the breast cancer maps, suggesting no spatial screening bias (map not shown, p-value for global test = 0.18). Our use of residential history allowed us to take latency into account but produced multiple residences, a potential source of bias. Since we analyzed residences, an apparent cluster may actually be caused by a few people moving within a small area. To examine the effect of multiple residences, we restricted our analyses to residences of longest duration. Although the spatial pattern of risk was similar for breast cancer, there were differences in the location and magnitude of hot and cold spots in the lung cancer analysis. This may indicate that the inclusion of multiple residences biased the lung cancer analyses. Improved methods for analyzing data with multiple residences are needed; weighting by residence time has been suggested [39]. While missing covariate data are a potential source of bias, multiple imputation suggested little effect on the results for breast cancer with 20 years of latency. Spatial methods for analyzing data with missing covariates are underdeveloped; the recent paper by French and Wand offers another possible option [15]. We adjusted for many individual level risk factors, but some authors argue for the inclusion of group-level contextual variables, e.g. [40]. By linking residential location to census data, one could test the importance of these variables relative to individual-level covariates. No information was available on some individual-level risk factors, e.g., genetic predispositions. While areas of increased or decreased risk may theoretically be caused by non-uniform control selection, sampling of controls within the study area did not depend on geography. We computed global and pointwise p-values, but many epidemiologists prefer confidence intervals when evaluating the precision of point estimates [41]. It should be possible to compute variance bands (also known as confidence bands) for our maps [17]. We identified areas with significantly increased or decreased risk using pointwise hypothesis tests. By making these multiple comparisons we increase the likelihood of finding significant hot or cold spots by chance alone. Although we make no adjustment for multiplicity, we only conducted pointwise tests if the global deviance test indicated that the map was unlikely to be flat. The location of significant hot and cold spots should be considered as exploratory.
Since several areas of elevated risk are near the coast, edge effects must be considered; GAMs may exhibit biased behavior at the edges of the data. However, loess may be less susceptible to this problem than many smoothers [17] and preliminary work with synthetic data found little bias on edges analyzed using our method [Webster et al. submitted].
Semiparametric studies of air pollution commonly employ GAMs. The effect of interest is modeled parametrically and several covariates are modeled with smoothers. Dominici et al. [42,43] reported that S-Plus may produce a biased parametric regression coefficient with inflated standard error. Ramsay et al. [44,45] warned that stricter convergence criteria alone are not sufficient for eliminating these issues: concurvity, a nonparametric counterpart to multicollinearity, is also responsible. We used our semiparametric model differently, modeling "exposure" (location) with a smoother and covariates parametrically, and statistically testing spatial variation with permutation methods. Inflation of software-provided standard errors is thus not an issue, but bias of the smooth is not ruled out. As an initial check, we modeled synthetic data using both default and more stringent convergence parameters; the maps were very similar and simple covariates were adequately controlled [Webster et al. submitted]. Additional work is needed on this issue.
Choice of bandwidth is one of the most important issues in smoothing [17]. We used the Akaike Information Criterion, a computationally feasible method for choosing an "optimal" bandwidth based on the tradeoff between bias and variance of the smooth. There are, however, problems with automatic bandwidth selection procedures. Selecting the span that optimizes the bias-variance tradeoff is not necessarily the same as understanding the importance of map features. The optimal span tends to be larger for smaller data sets, resulting in a smoother surface. Thus certain features in the data may not be captured in the analysis (e.g., compare Figures 4a, b). Furthermore, the AIC curves for breast cancer suggest two reasonable choices of bandwidths (Figures 6a, 6b). Rather than using a single bandwidth, there may be important aspects of the data at different scales. New methods are needed to address this issue, e.g., [46].
Statistical methods for mapping adjusted, point-based epidemiologic data are still fairly novel. It would be useful in the future to compare the results of generalized additive models and generalized linear mixed models.
Conclusion
Using generalized additive models and GIS, we generated maps of breast, lung and colorectal cancer risk. Our analyses showed little or no association between geographical location and colorectal cancer on upper Cape Cod. We observed an area of significantly elevated lung cancer risk north of the Massachusetts Military Reservation, similar to earlier research linking lung cancer to proximity to the military base. However, this result did not hold when we restricted analysis to residences of longest duration. Our results provide evidence for spatial clustering of breast cancer on upper Cape Cod. Areas of increased and decreased risk of breast cancer were not explained by covariates and became more extreme as we increased latency, findings consistent with geographical exposures. We identified three significant hot spots of breast cancer that coincide with groundwater plumes, an exposure hypothesis that warrants further investigation. We showed that spatial confounding can occur in maps, but in our analyses it tended to obscure rather than create clusters. Spatial epidemiology of population-based case-control studies addresses many methodological criticisms of cluster studies and generates new exposure hypotheses. Generalized additive models provide a relatively straightforward way to perform such analyses using standard software.
Abbreviations
AIC, Akaike's Information Criterion
DES, diethylstilbestrol
GAM, generalized additive model
GIS, geographical information systems
MMR, Massachusetts Military Reservation
OR, odds ratio
Competing Interests
The author(s) declare that they have no competing interests.
Authors' Contributions
VV conducted the spatial analyses and drafted the manuscript. TW directed the study, collaborated on all analytical decisions and wrote the second draft. JW provided statistical support and consulted on analytical and editorial issues. AA provided the data and assisted in epidemiologic analysis and editing. DO participated in the design of the study and the editing of the manuscript. The first two authors contributed equally. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by Superfund Basic Research Program Grant 5P42ES 07381. We also received support from the Silent Spring Institute Cape Cod Breast Cancer and Environment Study, which was funded by an appropriation of the Massachusetts Legislature administered by the Massachusetts Department of Public Health.
==== Refs
Rothman K A sobering start for the cluster busters' conference Am J Epidemiol 1990 132 S6 S13 2356837
Polissar L The effect of migration on comparison of disease rates in geographic studies in the United States Am J Epidemiol 1980 111 175 182 7355880
Neutra RR Counterpoint from a cluster buster Am J Epidemiol 1990 132 1 8 2356803
Wartenberg D Investigating disease clusters: Why, when, how? J Royal Statist Soc A 2001 164 13 22 10.1111/1467-985X.00181
Aschengrau A Ozonoff D Upper Cape Cancer Incidence Study Final Report 1992 Boston: Massachusetts Department of Public Health
Aschengrau A Ozonoff D Paulu C Coogan P Vezina R Heeren T Zhang Y Cancer risk and tetrachloroethylene-contaminated drinking water in Massachusetts Arch Environ Health 1993 48 284 292 8215591
Ozonoff D Aschengrau A Coogan P Cancer in the vicinity of a Department of Defense superfund site in Massachusetts Toxicol Indust Health 1994 10 119 141
Aschengrau A Ozonoff D Coogan P Vezina R Heeren T Zhang Y Cancer risk and residential proximity to cranberry cultivation in Massachusetts Am J Public Health 1996 86 1289 1296 8806382
Silent Spring Institute Cape Cod Breast Cancer and Environment Study: Final report 1997 Newton, MA
Aschengrau A Paulu C Ozonoff D Tetrachloroethylene-contaminated drinking water and the risk of breast cancer Environ Health Perspect 1998 106 947 53 9703477
Paulu C Aschengrau A Ozonoff D Tetrachloroethylene-contaminated drinking water and the risk of colon-rectum, lung, and other cancers Environ Health Perspect 1999 107 265 271 10090704
Aschengrau A Rogers S Ozonoff D Perchloroethylene-contaminated drinking water and the risk of breast cancer: Additional results from Cape Cod, Massachusetts, USA Environ Health Perspect 2003 111 167 173 12573900
McKelvey W Brody JG Aschengrau A Swartz CH Association between residence on Cape Cod, Massachusetts, and breast cancer Ann Epidemiol 2004 14 89 94 15018880 10.1016/S1047-2797(03)00120-0
Brody J Aschengrau A McKelvey W Rudel R Swartz C Kennedy T Breast cancer risk and historical exposure to pesticides from wide-area applications assessed with GIS Environ Health Perspect 2004 112 889 897 15175178
French JL Wand MP Generalized additive models for cancer mapping with incomplete covariates Biostatistics 2004 5 177 191 15054024 10.1093/biostatistics/5.2.177
Bithell JF A classification of disease mapping methods Statist Med 2000 19 2203 2215 10.1002/1097-0258(20000915/30)19:17/18<2203::AID-SIM564>3.0.CO;2-U
Hastie T Tibshirani R Generalized Additive Models 1990 London: Chapman and Hall
Kelsall J Diggle P Spatial variation in risk of disease: a nonparametric binary regression approach J Roy Stat Soc C-App 1998 47 559 573 10.1111/1467-9876.00128
Webster T Vieira V Weinberg J Aschengrau A Jarup L Spatial analysis of case-control data using generalized additive models Proceedings from EUROHEIS/SAHSU Conference: 30–31 March 2003; Östersund, Sweden 2003 London: Small Area Health Statistics Unit, Imperial College 56 59
Elliot E Wakefield JC Best NG Briggs DJ Eds Spatial Epidemiology: Methods and Applications 2000 Oxford: Oxford University Press
Bithell JF An application of density-estimation to geographical epidemiology Statist Med 1990 9 691 701
Kelsall JE Diggle PJ Nonparametric-estimation of spatial variation in relative risk Statist Med 1995 14 2335 2342
Diggle PJ Elliot E, Wakefield JC, Best NG, Briggs DJ Overview of statistical methods for disease mapping and its relationship to cluster detection Spatial Epidemiology: Methods and Applications 2000 Oxford: Oxford University Press
Paulu C Aschengrau A Ozonoff D Exploring associations between residential location and breast cancer incidence in a case-control study Environ Health Perspect 2002 110 471 478 12003750
James L Matthews I Nix B Spatial contouring of risk: a tool for environmental epidemiology Epidemiology 2004 15 287 292 15097008 10.1097/01.ede.0000121379.57583.84
Diggle PJ Tawn JA Moyeed RA Model-based geostatistics J Roy Stat Stat Soc C-App 1998 47 299 326 10.1111/1467-9876.00113
Kammann EE Wand MP Geoadditive models J Roy Stat Stat Soc C-App 2003 52 1 18 10.1111/1467-9876.00385
Prince MI Chetwynd A Diggle P Jarner M Metcalf JV James OFW The geographical distribution of primary biliary cirrhosis in a well-defined cohort Hepatology 2001 34 1083 1088 11731995 10.1053/jhep.2001.29760
Shimakura SE Carvalho MS Aerts DR Flores R Distribuição espacial do risco: modelagem da mortalidade infantil em Porto Alegre, Rio Grande do Sul, Brasil. [Spatial risk distribution: modeling infant mortality in Porto Alegre, Rio Grande do Sul State, Brazil] Cad Saúde Pública 2001 17 1251 61 [Article in Portuguese]
Insightful S-Plus
ESRI GIS and Mapping Software: ArcView
Chambers JM and Hastie TJ Statistical Models in S 1993 New York: Chapman & Hall
Insightful Corporation S-Plus 6 for Windows Guide to Statistics 2001 1 Seattle: Insightful Corporation 420 423
Brewer CA MacEachren AM Pickle LW Herrman D Mapping mortality: Evaluating color schemes for choropleth maps Ann Assoc Am Geogr 1997 87 411 438 10.1111/1467-8306.00061
Wiggins L (Ed) Using Geographic Information Systems Technology in the Collection, Analysis, and Presentation of Cancer Registry Data: A Handbook of Basic Practices 2002 Springfield (IL): North American Association of Central Cancer Registries
Miettinen OS Theoretical Epidemiology: Principles of Occurrence Research in Medicine 1985 Albany (NY): Delmar Publishers
Greenland S Finkle WD A critical look at methods for handling missing covariates in epidemiologic regression analyses Am J Epidemiol 1995 142 1255 64 7503045
Commonwealth of Massachusetts Massachusetts Geographic Information System (MassGIS). Cape Cod Commission Datalayers 1995
Sabel CE Gatrell AC Loytonen M Maasilta P Jokelainen M Modelling exposure opportunities: Estimating relative risk for motor neurone disease in Finland Soc Sci Med 2000 50 1121 1137 10714932 10.1016/S0277-9536(99)00360-3
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? Am J Epidemiol 2002 156 471 482 12196317 10.1093/aje/kwf068
Rothman KJ Greenland S Modern Epidemiology 1998 2 Philadelphia: Lippincott-Raven
Dominici F McDermott A Zeger S Samet J On the use of generalized additive models in time-series studies of air pollution and health Am J Epidemiol 2002 156 193 203 12142253 10.1093/aje/kwf062
Dominici F McDermott A Hastie T Issues in semi-parametric regression with applications in time-series models for air pollution and mortality Preprint 2003 Johns Hopkins University
Ramsay T Burnett Krewski D The effect of concurvity in generalized additive models linking mortality to ambient air particulate matter Epidemiol 2003 14 18 23 10.1097/00001648-200301000-00009
Ramsay T Burnett Krewski D Exploring bias in a generalized additive model for spatial air pollution data Environ Health Perspect 2003 111 1283 1288 12896847
Godtliebsen F Marron JS Chaudhuri P Significance in scale space for bivariate density estimation J Comput Graph Stat 2002 11 1 21 10.1198/106186002317375596
|
15955253
|
PMC1183231
|
CC BY
|
2021-01-04 16:36:32
|
no
|
Environ Health. 2005 Jun 14; 4:11
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-11
|
oa_comm
|
==== Front
Int J Equity HealthInternational Journal for Equity in Health1475-9276BioMed Central London 1475-9276-4-81597813010.1186/1475-9276-4-8ResearchHealth information generation and utilization for informed decision-making in equitable health service management: The case of Kenya Partnership for Health program Solomon Nzioka M [email protected] HSV-HIV project, Moi University-Indiana University Partnership Program; Faculty of Health Sciences, Moi University, Eldoret-Kenya2005 24 6 2005 4 8 8 4 8 2004 24 6 2005 Copyright © 2005 Solomon; licensee BioMed Central Ltd.2005Solomon; 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.
Context
The Kenya Partnership for Health (KPH) program began in 1999, and is currently one of the 12 field projects participating in the WHO's 'Towards Unity for Health initiative' implemented to develop partnership synergies in support of the Primary Health Care (PHC) approach [1].
Content
This paper illustrates how Program-linked Information Management by Integrative-participatory Research Approach (PIMIRA) as practised under KPH has been implemented within Trans-Nzoia District, Kenya to enhance community-based health initiatives. It shows how this model is strategically being scaled-up from one community to another in the management of political, social, cultural and economic determinants (barriers and enhancers) of health.
Objective
Target rural communities in the development of a community-based health information management and feedback initiatives that can provide insights on the social, cultural, political and economic determinants of health for utilization in informed health service management.
Key Findings and Achievements
1. Cues for health seeking and health service utilization are determined by the social, cultural, political and economic factors as seen by the individual and as defined by the community but not due to the pathological nature of the illness.
2. Establishment of community-based health surveillance and health action initiatives as the best practices in transferring health as a resource that can be 'owned and guarded' by the community.
3. Establishment of Healthy Villages Initiative (HVI) through which health service delivery and scale-up can be sustained at the community level.
4. Provision of actionable health information necessary for health planning and evaluation of preventive health programs thorough PIMIRA.
Conclusion
It has been realized that for every one person who visits a health facility for medication, there are nine others who had the same condition but sought health care from other sources including self-medication and five others who never sought health care. Innovative means of involving the community in health information management and utilization such as PIMIRA are hence the best ways of guaranteeing equitable delivery of health services that are accessible and sustainable by the community.
built-in health information managementcommunities' health needsservice-linked datasocialculturalpolitical and economic determinants of healthcommunity-based health information managementcommunity-based health surveillancehealth seeking behaviourHealthy Villages Initiative
==== Body
1.0 Introduction
1.1 Setting of KPH program
The focus of this article is to present the success story of closing the information-utilization gap through community-based initiatives by specifically narrowing down to the Program-linked Information Management by Integrative-participatory Research Approach (PIMIRA) and its role in promotion of sustainable health service delivery through Healthy Villages Initiative (HVI) as applied in the Kenya Partnership for Health (KPH) program. KPH is one of the 12 world wide field projects participating in World Health Organization's Towards Unity for Health (TUFH) initiative being implemented in support of the PHC pillars; namely inter-sectoral collaboration, community involvement and use of appropriate technology in order to deliver sustainable health services [2]. The focus of KPH program is capacity building and social structural strengthening at the community level for health information management and utilization in decision making, by use of health facility based routine data on reported morbidity and community based health surveillance. KPH prototype field program is based in Trans-Nzoia District of Rift Valley Province within the Republic of Kenya. It is implemented in the whole district but for effective and efficient monitoring and progress reporting, KPH is comprehensively undertaken in Amuka Sub-location of Kaisagat located in Kwanza Division within Trans-Nzoia District. Trans-Nzoia district is at an attitude of 1800 M above sea level with a high water table coupled with an annual rainfall of 1200 MM and a mean temperature of 18.6°C.
1.1 Background of KPH Program
Moi-University Faculty of Health Sciences offers undergraduate degrees in. Environmental Health, and Nursing whereby a Community-Based Education and Service (COBES) approach is applied in the training programs. COBES sessions are undertaken in each year of study for a period of six continuous weeks whereby students are assigned to a field -based program for hands-on experience. During the second year of Environmental health studies, students may participate in an elective course for six weeks at their place of choice so as to enhance any special skills they may want to develop and utilize during their career. For my elective course, I was assigned to the AMREF Water and Sanitation (WATSAN) project based in Kitui district of Eastern province.
1.3 Problem Statement and Justification
A rapid appraisal of the WATSAN project was one of the key activities I undertook. The two broad areas I considered were the project's performance against pre-set targets and the influence of the project on the beneficiary community's health profile. To the satisfaction of the project implementers, there were definite achievements in the number of springs and shallow wells protected as well as the number of Ventilated Improved Pit (VIP) latrines constructed and replicated by the community. However there was uncertainty in gauging the community's health outcome due to unavailability of service-linked data that can be inferred on the course-effect measurement scale. There was an intrinsic need to share and collaborate data from varied sources; not only from the routine health facility records which were inconsistently maintained but also incomplete. The records did not explain the community's health seeking behaviour and health service utilization. This gap between health information management and utilization formed the bulk of the work that I have been undertaking in search of a solution that can bridge the gap.
2.0 Conceptual framework
2.1 Empirical Argument from Literature Review
From the health information management and decision-making gap identified under the Kitui WATSAN project, I undertook review of approaches that could be applied to delineate the existing gap. From review, it was established that unfortunately this gap also contributes greatly to the utilization or non-utilization of health services due to the social, cultural, political and economic determinants of health that could not be captured from the routine health records. Theo Lipperveld, in his article on Health Information System, suggests that no single data source can provide all of the information required for planning and management of health services [3]. We sought an approach that would promote health information generation and utilization at the community level to solve the intricacies of health disparities due to differences in social, cultural, political and economic determinants of health. Insight was provided by operations research as well as community participatory training approaches that have been developed. However, a critical review of these broad areas of operations research and participatory community training approaches could not provide a solution to the prior identified problem. A search for a model that could not only combine the two broad approaches but also provide a health information-action link from the health services was undertaken.
Such community-based initiatives reviewed and applied over the five years in search of a model that could be applied to close the gap between health information generation, processing, dissemination and utilization of service-linked health information included development and utilization of community-based training toolkit such as in Participatory Hygiene and Transformation Training (PHAST), Child-to-Child and Participatory Rural Appraisal (PRA) approaches as well as a UNDP land planning approach referred to as Mèthode Appliècaution de le Planification et de Resèarchè (MAPIR). Review also covered health promotion models such as the PRECEDE-PROCEED model, the Health Education/Promotion model from Mark et al (1992), and the 21st Century Field model on determinants of Health as adapted in Robert G. et al [4]. The section below gives an outline of the provisions of 21st Century Field model on determinants of Health.
2.1.1 The '21st-Century Field Model'
In the five year review, the 21st field model was found to be the most comprehensive model that considers determinants of health in its broadest sense. It is unlike the modern practice of pathological orientation in which unfortunately the outcomes of preventive programs are also oriented to in the exclusion of social, cultural, political and economic determinants of health and well-being (see figure 1: Determinants of Health: The 21st Century model).
Figure 1 Determinants of Health: The 21st Century model.
2.1.2 'The 21st-Century Field Model' Explanation
According to the 21st century field model on determinants of health, there are three broad levels on which health can be promoted; primary, secondary, and tertiary. The environment variables that should be considered relate to primary prevention ("how do we keep ourselves well?"), secondary prevention ("if we are getting sick, how can we detect these conditions early?"), and finally tertiary prevention ("if we are sick, how do we get the best care?"). Secondary prevention involves activating the health care system to reduce the prevalence of disease and injury which can lead to recovery, disability, or death, or disease and injury can be resolved outside the health care system with appropriate self-care and good decision making.
According to the 'the 21st-century field model', recovery from an illness or injury can return an individual to productive pursuits, which feed back to the global factors at the top of the model. The big question being answered in this model is how the world can attain better health in the 21st century? Then as the model demonstrates, there are many paths of influence, each presenting its own challenges for the future. The components of the model that offer opportunities for improvement through communications and public health intervention are the community/social, Physical and family/individual environments. WHO has estimated that a poor physical environment is responsible for about one-fourth of all preventable diseases. For example, tropical regions are ideal environments for the transmission of many diseases including malaria, schistosomiasis, and diarrhea. Environmental conditions account for an estimated 90 percent of health problems caused by malaria. Environmental threats to human health can be divided into "traditional hazards" associated with a low level of economic development, and "modern hazards" associated with industrialization. Traditional hazards related to poverty and low levels of economic development and include lack of access to safe drinking water, inadequate sanitation in the household. In the community these include indoor air pollution and inadequate solid waste disposal.
With the structural framework outlined in the 21st Century Field Model, PIMIRA takes a further step of addressing the determinants of health at the community level by providing service-linked information for the purpose of enhancing ownership to health as a resource that can be 'cherished and guarded' by the community.
2.1.3 Birth of Program-linked Information Management by Integrative-participatory Research Approach (PIMIRA) as the Information-Utilization Magic Bullet
This article presents the resultant model developed when none of the reviewed models could respond to the initially identified gap under the Kitui WATSAN project; namely that of health information and community participation in health service delivery. The resultant model described hereby is referred to as Program-linked Information Management by Integrative-participatory Research Approach (PIMIRA), which has been developed since 1999. PIMIRA is currently being implemented hand-in-hand with other community advocacy-for-health-action approaches such as the polythene free initiative, Healthy Villages Initiative and the annual Health festivals extravaganza; it is the bond that is transforming the community advocacy-for-health-action approaches to health ownership by the community in the context of 'our health-our good-our responsibility'.
The PIMIRA model has proven to be effective in closing the gap between health information generation, processing, dissemination and utilization by forming the soft bond of providing health information relevant for informed decision-making to the community, health managers, policy makers, health professionals and health training institutions.
2.1.4 PIMIRA Principles
Assumptions in PIMIRA
i. Those who create decisions will be committed to following them through sustainable health development
ii. Communities are capable of accurately describing their present situation/ problems and visualizing possible improvements
Underlying Principles
i. People have beliefs about the causes of diseases which may or may not be consistent with the scientific explanations of the disease
ii. No lasting change in people's behavior may occur without health awareness, understanding and believing in the change
PIMIRA Operational Principles
i. The best way to achieve sustainable health improvements is to take an incremental approach, starting with what the community has and building upon a series of changes
ii. Self-esteem and associative strength is a prerequisite to decision-making and follow-up.
iii. Local people should be the health analysts and presenters of their own health situation while health workers play the role of catalytic facilitation.
iv. Any innovative initiative should apply adaptable participatory learning process (creative) rather than lay down a prescriptive set of tools to be followed.
v. Community Health training by health information visualization reduces marginalization and promotes equitable health access
vi. Multi-disciplinarity and diversification of information sources, tools and techniques avoids uni-sectorial approaches to reality by enabling analysis from varied outcome determinants
2.1.5 Community Based Toolkit (for Operations Research)
i. Pocket Chart (PC) – an investigative and evaluative tool used to collect and tabulate data on where people defecate, where they collect water and their information and communication structure.
ii. Community mapping – showing the available water supply resources, permanent mosquito breeding sites and pillar stones in information and communication within the community as well as disease distribution by lay definitions.
iii. Resource maps – showing the income generating activities the community is involved in.
iv. Flow charts – showing the possible water and food contamination routes as well as the existing information flow structure.
v. Matrix classifications – a set of pictures of common causes and barriers of health and communication. They utilize epidemiological principles to pass health concepts to the community by way of visualization.
vi. Venn diagrams – Collects information about the traditional or modern organizations involved in the management of local water resources and information systems.
vii. Community surveillance tally cards – cards containing visual representations of the main signs of water related diseases and malaria vis-à-vis community routine activities like school attendance and provision of casual labour to the agricultural farms as is usually the case.
viii. Facility morbidity tally sheets – tabulation sheets for recording village specific water related diseases and malaria as they are reported in the community health facilities.
ix. Historical analytic charts and Seasonal calendars amongst others documents how the community have been handling certain diseases and the traditional seasonal calendar scheduling of activities from the community historical perspective.
3.0 KPH Interventions/Preventive Services
Hardware of KPH program includes provision of Insecticide Treated Mosquito nets (ITNs), indoor residual spraying against mosquitoes, safe water supplies and promotion of latrine ownership while the software (means to sustainable social and economic development) includes capacity building and social structural strengthening through community training, enhancement of Income Generating Activities (IGAs) through micro-enterprising, enhanced community involvement in health information surveillance through the operations research (PIMIRA) and finally stakeholder networking through the service-linked information dissemination.
3.1 The Big-4 Questions answered by PIMIRA
i. Can community-based health surveillance on the cultural, social, political and economic factors be utilized in addressing the determinants of health as well as enhancing management of health resources at community level?
ii. Can community involvement in health information management improve access, utilization and provision of equitable health services?
iii. What methods can be applied to promote direct involvement of the community in Health Information Management?
iv. Can community involvement in Health Information Management be utilized for effective scaling-up of health gains within the community?
3.2 The PIMIRA Structural Framework
Program-linked Information Management by Integrative-participatory Research Approach (PIMIRA) answers the over- arching question of 'How do we provide service-linked information to different stakeholders in health care for informed decision-making and for delivery of sustainable health services from the beneficiary community's perspective? Answer to this question is illustrated in Figure 2: Health Systems Management Model: Kenya Partnership for Health Approach
Figure 2 Health Systems Management Model: Kenya Partnership for Health Approach.
Operations Research model: 'PIMIRA' Explanation
The model above represents the framework of a socially accountable health delivery system as projected under the KPH program. In such a system, information management is the power and the strength through which decision and actions regarding the preventive services being offered are monitored, evaluated and improved. The program management team is the conceptual and operational group targeted with developing or delivering socially accountable services answerable to beneficially community health needs. The team is basically a network of a multi-disciplinary oriented working group who oversee all the implementation, monitoring and evaluation of the program with controls or indicators for all social, cultural, political and economic determinants of health. It is an epicenter for centralized service-linked information management and dissemination for timely feedback on the program from all the concerned partners. The input/incoming arrows from the boxed texts to the PMT signify the basic information sources that should be critically sought in the initial design of a socially responsive (cost-effective, high quality, relevant and equitable) health delivery program.
The out-going box from the PMT signifies the concretized input from the other partners while the dotted blue lines represent the levels of action-linked feedback.
3.2.2 A Critique of PIMIRA Model Suitability
Health information is identified as the key to successful partnership and sustainable health service delivery as well as the tool through which community involvement for ownership in health services can be based. PIMIRA was developed from the health education/promotion model by Mark et all (1992). The major deviations (to address weaknesses in Mark's model) are:-
1. Marks' model exclusively focused on health promotion schemes while PIMIRA model focuses on Human Resource for Health (HRH) development by use of community-based tool-kit.
2. Unlike the two feedback levels to project implementation and Project objectives recognized in Marks' model, PIMIRA considers the two and refocuses on the overall health service planners and the community as key decision-making levels for improvement of any health service.
3. PIMIRA model strongly links the health service planners, the community and the training institutions by way of mutually dependent information sources for common/shared health action
4. It puts targeted population into active partnership from the inception to the end of the sponsored health service project through the design, development and use of community-based toolkit.
5. Beneficiary community is the main actor in process evaluation. They continuously gather activity-based health attribute data, develop community health profile and subsequently decide on the cause of action for their own health.
6. Activity specific-health indicators and the community health profile (Objectively Verifiable Indicators) forms the basis for evidence-based health impact and outcome evaluation.
3.3 The Healthy Villages Initiative (Ripples analogy)
This is the default method through which villages have embraced preventive services provided through PIMIRA as 'our health, our responsibility' and hence duplication of the health interventions in the villages through the neighborhood concept of development. This has been realized through public health field days being held within the healthy villages as illustrated in Figure 3: The Healthy Villages: Ripples Analogy.
Figure 3 The Healthy Villages: Ripples Analogy.
A healthy village is defined as an administrative village in which minimum public health requirements to prevent malaria and diarrhea diseases have been undertaken and there has been demonstrable evidence on the latter. In future this will also include HIV/AIDS and Child Health Essential Services (CHES) a package that will include promotion of safe delivery practices, immunization, nutrition and prevention of childhood illnesses as outlined in 2003 UNDP report [5] and by Scott et al [6].
Public health field days are community awareness campaign/show days during which the neighbouring communities/villages are invited for education and awareness on best-integrated preventive practices against malaria (including Intermitted Pregnant Treatment -IPT) and diarrhea diseases. In future this will also include HIV/AIDS and CHES. The following session demonstrates what timely health information can contribute to enhance access and equity of health care services in relation to malaria, diarrhoel diseases and HIV/AIDS. This concept is based on the Ottawa Charter [7] and the 1998 Human Development report [8].
4.0 Achievements through PIMIRA community-based initiatives
Table 1 presents evidence of sustainable health service delivery by closing service-based information-utilization gap. It shows that though malaria had low fatality rate, it is the major contributor to individual financial resource drains in the community directly and indirectly since as collaborated from the community-based health surveys (conducted at base line), many people do not go to look for casual jobs as main contributor to children absenteeism at schools. These factors have direct major impact to the Trans-Nzoia community economic welfare were more than 75% of the natives are in the informal employment sector. Malaria morbidity was seen to directly be leading to reduced per capita earning and hence low purchasing power which narrows down to determination of where, when and how to seek medication. The community based surveillance toolkit implemented under PIMIRA led to the realization of the community's economic potential that can be achieved if malaria morbidity can be reduced and embracement of preventing mosquito breeding and bites within the Health Villages Initiative and hence realization of "Our Health, Our Responsibility".
Table 1 Evidence of sustainable health service delivery by closing service-based information-utilization gap (Malaria as the main economic depressant from the community's perspective)
Top three diseases and their case fatality rate for the year 2001 (cue for community action).
Disease No. Of cases reported Relative % Mobidity No. Dead Case fatality rate
Malaria 18220 77% 180 0.9*
Pneumonia 1836 8% 128 6.9
Diarrhea diseases 3464 15% 80 2.3
Within the Health Villages, School Health clubs, market committees and Village Health Committees are now having shared health objectives in eliminating mosquito-breeding sites. Through public health field days, this has led to the neighboring villages demanding that the same preventive measures be undertaken in their villages. For the first time in many years, a malaria outbreak was not reported in Trans-Nzoia District since 2001 [9]. Through this, the roll back malaria initiative has got roots and is now rolling out from one village to the next and eventually we hope it will spread to the whole district [1].
Related community actions includes social marketing of insecticide treated nets at a subsidized cost from population Service International (PSI), community mapping of mosquito breeding sites and their management, indoor residual spraying and house screening. Achievement of all these measures to 75% coverage has been our mark for a Healthy Village before a public health field day is held. A total of 9 public health field days have since been held and this is an indicator of coverage in the efforts of rolling back malaria in the community and hence increasing their economic productivity. With a district malaria epidemic threshold of 15,000 cases per month, the trend has been as presented in figure 4.
Figure 4 Morbidity Cases for the Epidemic prone period of April to July (1993 – 2002).
4.2 The case of Diarrhoeal Diseases Management as a cue for community action
The reference community uses communal springs as their main sources of drinking water. To enhance community participation, spring committees were selected by the community to support the Village Health Committees in the coordination of spring protection activities such as collection of the community share of contribution. From the program side, this group was utilized for health communication to the community regarding safe water handling even at household level to prevent recontamination. The communal springs were tested for coli-forms before and after protection and test outcome is as presented in figure 5.
Figure 5 Testing outcome on Coli-forms in water springs before and after protection.
Using Moi University, Faculty of Health Sciences mobile paqua lab, the spring protection committee members where utilized in drawing both pre and post protection water samples and were present during incubation to test the levels of contamination. Having obtained the above results, the community Health Workers (CHWs) continued to administer the community-based surveillance tools after which the results was compared to the dispensary records on diarrhoeal diseases. Through subsequent community mapping charts, the community members were amazed to note the distinct changing patterns of diarrhoeal cases reported. Using the protected water sources as reference points to convince the community on the significance of water source protection in the reduction of diarrhoeal cases, more springs are now been protected. The community is now competing to contribute their share and hence demanding that the program contributes its share for their water sources to be protected. Unlike in the past where springs maintenance was the work of the Government or the NGOs who had sponsored their protection, the total number of 9 springs protected under the KPH program are maintained/repaired by the community.
The story of fecal contamination, its subsequent elimination by water source protection and the subsequent reduction of diarrhoeal diseases is taught every time there is a public health field day. This has dramatically increased the demand from the community for protected water sources. They are writing proposals and organizing themselves into spring committees without a significant health professionals' involvement.
All the community-advocacy for action activities are undertaken through self-help groups for which seed funds are given in the form of Insecticide treated mosquito nets as an income generating activity. Scaling up of the Health benefits from one village to another is through public health field days in which neighbouring villages are invited to see, touch and feel public health interventions within the Healthy Villages. This is the concept outlined in the 2003 WHO report [10].
4.3 HIV status as an indicator of testing facilities utilization
Table 2 presents evidence on how closing service-based information-utilization gap can led to sustainable health service delivery. Health facility utilization indicator is inversely proportional to the number of people testing positive out of those who turn up for actual testing. When the percentage indicator is high, most of the people who turn up for testing are actually positive and hence on inference, people who have had strong reasons to believe that they are infected are the only ones utilizing the testing facility. On the other hand, when the percentage indicator is low, it shows that most of the people who turn up for testing are end up being negative and hence on inference people who are utilizing the service have had other reasons or cues to action other than strong reasons to believe that they are infected. These other cues for health action may include and not limited to increased individual/societal health awareness on the need for the service on offer which for the case of HIV/AIDS in Trans-Nzoia District may be directly attributed to the community-based awareness strategies in place such as peer group counseling de-stigmatization campaigns through Village Health Committees, Market committees and School Health Clubs.
Table 2 HIV status as an indicator of utilization of testing facilities (Closing service-based information-utilization gap sustainable health service delivery)
YEAR/ MONTH 1999 2000 2001
No. tested % positive No. tested % positive No. tested % positive
JAN 57 44 15 27 73 29
FEB 94 48 75 52 62 45
MAR 71 37 95 42 119 46
APRIL 55 55 71 45 27 37
MAY 56 45 88 45 139 44
JUNE 47 53 86 29 116 44
JUL 52 35 70 50 115 28
AUG 74 31 75 56 63 37
SEPT 57 54 109 45 68 44
OCT 64 53 41 29 124 29
NOV 67 39 101 54 104 34
DEC 64 58 47 15 92 26
TOTAL 758 552 873 489 1102 443
Service utilization Indicator (%positive/%tested) 73% 56% 40%
Looking at it from another angle, even when the cues for action are taken to be constant, then the utilization indicator in the case of HIV/AIDS shows that there is a progressive annual increase in the number of people seeking testing services at an early stage of infection.
As an evidence of information for decision-making loop completed from the above reporting, the year 2002 witnessed an affirmative action by the District Medical Officer of Health (DMOH) and the District Commissioner (DC) to the effect that data on all HIV/AIDS testing and related services kept by the different NGOs, private and mission organizations within the district should be centralized for analysis on regular intervals of every three months. This was in addition to the establishment and acceptance of the KPH program to operate in the District so as to coordinate data management and utilization for health action at all levels of care within the district.
5.0 Conclusions, recommendations and way-forward
The PIMIRA method has operationalized the 21st Century Field model on determinants of health by addressing the social, cultural, political and economic determinants of ill health as defined in its outline on different environments affecting health and well-being. This is as contained in the PIMIRA operational principles outlined in this document.
Scaling up of this approach to include any new health package or cover a larger area is intrinsically provided for in PIMIRA through its congruent Health Villages initiative. KPH is hence a good case example of promoting health equity and access to health services by closing the information-utilization gap through community-based initiatives.
Recommendation is for logistical and financial resources to be availed for extension of the KPH coverage especially for community-based trainings and seed funds for income generating activities. The success PIMIRA has guaranteed the way forward for sustainable delivery of health services that are based on community's needs (social, political, economic and cultural) and hence establishment of socially responsive health programs.
The way-forward is to scale-up the Healthy villages initiative by including HIV/AIDS support programs for the People Living With AIDs (PLWAs) and People Affected by HIV/AIDs (PAHAs) as well as targeted interventions on Child Health Essential Services (CHES), a package that will include promotion of safe delivery practices, immunization, nutrition and prevention of childhood illnesses.
Finally as noted in the World Health report 2003, with a renewed commitment to the principles of PHC through strengthened health systems, real Millenium Development Goals (MDGs) and other national health priorities can be achieved. Shaping the future looks at the impact of the different threats to health and offers innovative and far-reaching solutions that will shape a better, healthier future for all – but only if the international community act now.
Acknowledgements
I acknowledge the field efforts of Mr. Ulo Benson currently with Merlin International and the logistical support offered by Prof. B.O. Khwa-Otsyula; Dean Faculty of Health Sciences, Moi University during the actual implementation of the PIMIRA project and my dear wife Hellen W. Kamotho for her undying encouragements during those moments of uncertain outcomes. I highly appreciate efforts by the Ministry of Health; Department of Environmental Health (public Health), Trans-Nzoia District and the community of Kaisagat Location for making the PIMIRA project a success. Finally I acknowledge the WHO (Department of Health Service Delivery, OSD) for the seed-fund provided to initiate the PIMIRA project.
==== Refs
Charles Boelen ed Towards Unity For Health (TUFH) Challenges and opportunities for Partnership in health development, A working paper 2000 Geneva: World Health Organization
WHO The Challenge of Implementation: District Health Systems for PHC 1988 Geneva: World Health Organization, 2003
Jack Bryant Charles Boelen Buz Salasky Ed Managing a successful TUFH project (WHO), Mini-Symposium report 2001
Robert EvansG Stoddart GL RG Evans, ML Barer, and TR Marmor "Producing Health, Consuming Health Care," in Why Are Some People Healthy and Others Not? 1994 53
UN Development Programme, Human Development Report Millenium Development Goals 2003 New York: Oxford University Press
Scott RatzanC Gary FilermanL LeSar JohnW Eds Attaining Global Health: Challenges and Opportunities 2000 55 A publication of the Population Reference Bureau
WHO Ottawa Charter For Health Promotion Paper presented at the First International Conference on Health Promotion: The Move Towards a New Public Health. Ottawa, Ottawa, Canada Nov. 17–21, 1986
UN Development Programme Human Development Report 1998 New York: Oxford University Press
Charlote Leighton ed The Health reform and priority Health Services Journal: Health system reforms in improving priority health services 2001 Partnership for Health Reform project, Abt Associates Inc, Bethseda, Meryland
WHO World Health Report 2003, Shaping the future 2003 Geneva: World Health Organization
|
15978130
|
PMC1183232
|
CC BY
|
2021-01-04 16:39:32
|
no
|
Int J Equity Health. 2005 Jun 24; 4:8
|
utf-8
|
Int J Equity Health
| 2,005 |
10.1186/1475-9276-4-8
|
oa_comm
|
==== Front
Int J Equity HealthInternational Journal for Equity in Health1475-9276BioMed Central London 1475-9276-4-91598516510.1186/1475-9276-4-9CommentaryNepal's War on Human Rights: A summit higher than Everest Singh Sonal [email protected] Khagendra [email protected] Edward [email protected] MPH Student, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland, USA2 Department of Medicine, Unity Health System, Affiliate of the University of Rochester, Rochester, NY, USA3 International Student Representative, International Physicians for the Prevention of Nuclear War, Katmandu, Nepal4 Centre for International Health and Human Rights, University of Oxford, UK2005 28 6 2005 4 9 9 17 2 2005 28 6 2005 Copyright © 2005 Singh et al; licensee BioMed Central Ltd.2005Singh 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.
Nepal has witnessed serious human rights violations including arbitrary arrests, detentions, "disappearances", extra judicial executions, abductions and torture carried out by both the Royal Nepalese Army and the Maoist rebels in the 10 years of the "peoples war". Women and children have borne the brunt of the conflict. Massive displacement has led to adverse social and psychological consequences. While the reasons for the conflict are mainly indigenous and rooted in the social and economic in-equities, remedies for health inequities must come not only from the health sector but also from broad social policies and adopting a participatory and conflict-sensitive approach to development. Meanwhile the international community needs to use its leverage to urge both sides to accept a human rights accord and honor international human rights and humanitarian laws, while investigating allegations of abuse and prosecute those responsible.
==== Body
Introduction and recent events
Nepal is a Himalayan kingdom in south Asia, sandwiched between India and China. Serious human rights violations have escalated since the Communist Party of Nepal (Maoists) launched a "people's war," against the government forces in 1996. On February 1st of this year (2005), King Gyanendra of Nepal announced a state of emergency in Nepal, assuming direct rule over the kingdom for a planned 3 years. Political leaders, including the prime minister and opposition leaders, were placed under house arrest. Many student leaders, human right activists and pro-democrats were detained. News media were censored with security personnel patrolling the streets on high alert. According to the king "There is spiraling violence by the rebels which has caused enormous suffering to the people and nation, but the political parties are just fighting among themselves and have been unable to stop it." [1] This dramatic event attempts to strengthen the Royal Nepalese army, an army which has already been accused of serious human rights violations, and has heightened the possibility of further human rights violations in Nepal.
While the reasons for the current conflict are mainly political, we explore some of the social economic and health inequities at the root of the conflict, the current human rights situation in Nepal and its impact on the more vulnerable of the population, mainly women and children. We also explore the role of the international community and developmental and humanitarian Non Governmental Organisations (NGOs) in the current conflict and their successes and failures in promoting the cause of human rights and health equity in Nepal.
Inequities at the root of the "Peoples War"
As one of the poorest countries in the world, Nepal has a gross national income of US$240 per person [2] and a population of more than 23 million, where some 85 % reside in villages and the majority (82%) survive on less than 2 dollars a day. In the last two and a half decades, the country has experienced an average economic growth rate of four per cent. However, the number of people below the poverty line has doubled from 4.7 million in 1976 to 9 million in 2002.[3]
The indicators of health and quality of living are very meager in this part of the world. Life expectancy at birth is 61 years. The maternal mortality ratio is 539 per 100000 livebirths, [4] and the infant mortality rate is 64 per 1000 live births.[5]
There are widespread disparities in the health care indicators between people from different class and geographical areas. This is evident from the comparison of under-five mortality rates (U5MR) among women on the basis of their education level: U5MR for children of uneducated mothers is 121 per 1000 births which is 64 per cent higher than for children of mothers with primary education, and is nearly double that for children of mothers with secondary level education.[5] Similarly, U5MR in urban areas is 93.6 per 1000, whereas in rural and mountainous regions it increases to 147 and 201 per 1000, respectively. [5] There are also differences in immunization coverage, nutritional status and health care delivery. Similarly, the distribution of the total of 3200 physicians in Nepal is imbalanced: the average physician-patient ratio is 4 doctors per 100,000 people in Nepal that rises to only 1 physician per 150, 000 people in remote hilly areas.[6]
The disparities are also reflected in the Human Development Index (HDI), a composite index of education, health (life expectancy at birth (LEB)) and income, which shows a close association with the caste hierarchy. Brahmans, Newars, and Chhetris, are well above the national average while indigenous people, Dalits (untouchables) and Muslims are below the national average.[7] The marginalization of indigenous people and Dalits, and discrimination along caste and gender lines, with the widening rural-urban divide are a key factor in the current conflict.
Originating in the western heartlands region of Nepal, which has some of the worst indicators, the conflict now has spread to all 75 districts. It has led to widespread disruption of infrastructure and affected the delivery of health services throughout the country.[8] The conflict has claimed more than 11,000 lives and human rights violations have escalated since the collapse of a cease-fire between the two sides in August 2003. Nepal has become a country of human generated disasters [9]. Underdeveloped roads and fragile communication links (only 14 phone lines for 1000 people) [10], in a rugged mountainous terrain suited for guerilla warfare, has allowed both sides to perpetuate crimes against civilians with complete disregard for the rule of law.
Human rights abuses - disappearances and torture
US based Human Rights Watch claims that both the Maoist rebels and the Royal Nepalese Army are engaged in regular intimidation and extortion leading to a climate of intense fear in Nepal.[10] The government forces have resorted to large-scale arbitrary arrests, detentions, "disappearances", extra judicial executions and torture including rape.[11,12] Human rights defenders, including lawyers; journalists and members of NGO's have been arrested, tortured, killed or "disappeared" in Nepal. [12] Nepal held the unique distinction for the highest number of "disappearances" of any country in 2003 and 2004.[13] The Maoists have resorted to torture and deliberate and unlawful killings.[11,13] According to INSEC (Informal Sector Service Centre), a human rights organisation, nearly 3000 people were killed and about 26,000 people were abducted in 2004 in Nepal.[14] The Maoists have abducted civilians, including teachers and schoolchildren for the purpose of 'political indoctrination'. [13]
More than 70% of Nepalese prisoners claim to have been tortured while in custody [15] The Centre for Victims of Torture, (CVICT) Nepal, an NGO based in Kathmandu claims that some 16,000 people are subject to torture in Nepal every year, affecting an estimated 100,000 people including family members. [16] According to data compiled by CVICT, at the beginning of the Maoist insurgency 80 percent of the victims were subjected to torture from the state and the remainder by the Maoist rebels. However, a recent study in the mid-western district of Jajarkot showed that the number subjected to torture by Maoists had doubled and reached 40 percent [16]. A recent study by Danish researchers confirmed the presence of torture by both the government forces and Maoists in mid-western Nepal[17] and our survey of Tibetan refugees fleeing to Nepal at the Tibetan Refugee Transit Centre in Kathmandu confirmed the presence of torture by both the security forces and the rebels.[18] The long-term repercussions of torture on health and psychological well-being are considered devastating. [19]
Role of health professionals
Some physicians have contributed to the politics of Nepal in the last decades in the struggle to ensure health as a human right within the broader macroeconomic and political picture.[20] The government has prosecuted physicians for the ethical practice of providing care for those injured, including rebels, violating international ethical standards set by the World Medical Association. [21] The government has issued directives to all health professionals and institutions stating that if health professionals provide treatment without appropriate notification, they will be regarded as supporters of terrorists and be prosecuted according to the Terrorist and Disruptive Activities Ordinance, 2001.[22] The directive was outlined by the Ministry of Health in a public meeting: "doctors working both in government hospitals and private health institutions are liable to government action if they treat terrorists without getting permission from the security wings... if any doctor defies, action will be taken against him or her as per the recently promulgated ordinance against terrorists".[22] This directive puts medical professionals in an impossible situation: during the ongoing conflict, medical professionals are at risk of encountering armed groups demanding treatment for their wounded; however, provision of such treatment might lead to subsequent government prosecution. [21] The Nepalese Medical Council (NMC), the only national body ensuring medical ethics, has remained silent on this issue.
Consequences of conflict
Nepal has witnessed a gradual increase in the incidence of depression, posttraumatic stress disorder and suicide since the beginning of the conflict [23]. Mental health services, which were rudimentary to begin with, have been further fragmented. Health experts estimate the prevalence of mental health problems in Nepal to be as high as 30 %. [23]
Women and children are bearing the major brunt of the war. With literacy among women as low as 36%, [24] the political violence has had a negative impact on women's rights and health.[25] The Maoists have capitalized on the plight of women, who have been marginalized for decades in Nepalese society and enrolled them into the conflict in large numbers [26] Nearly half of all Maoist rebels are women and their sexual exploitation is not uncommon. The conflict has also contributed to an increase in the trafficking of Nepalese women and girls, nearly 5000 to 10,000 a year to Indian brothels. [27] The youth have fled the country in large numbers to Indian cities and the Middle East, leaving women and children behind.
Children have been particularly affected by the insurgency. [28] Some estimate that around 100, 000 children have been affected by the war and the numbers likely to increase to 500,000 as the conflict expands.[29] Conservative estimates in 2003 showed that at least 146 children have died, 2000 have been orphaned, and 3000 have become homeless.[29] While the government vehemently denies the use of child-soldiers, around 10–15% of the recruits are under the age of 18 years (possibly due to birth registration irregularities which are not uncommon in developing nations).[30] The Maoists previously denied the use of child-soldiers. However, according to an estimate in 2000, around 30 % of the Maoist soldiers were children.[31] They have been utilized as informants, porters, and for cultural propaganda. Earlier in 2004 the Maoists announced a plan to create a militia of 50,000 child soldiers.[32] Although the numbers of recruits planned may be ambitiously inflated, they have resorted to mass- abduction of children as young as 12 from schools and classrooms in Western Nepal. The abducted children are indoctrinated and given training in guerilla warfare.[32] This marks a major departure from their previous commitments to avoid recruiting children below the age of 18.
Refugee crisis
Nepal provides support to nearly 100,000 Bhutanese refugees under the aegis of the United Nation High Commission for Refugees (UNHCR). The refugee crisis has been compounded by the fact that many have moved within Nepal as a result of the conflict between Maoists and the government. Most realistic estimates put their number between 100,000 and 200,000. [33] The displaced Nepalese have either flocked to the main cities or fled the conflict to India. Displacement of the Nepalese – population has given rise to social problems commonplace amongst migrant populations. The government, to a large extent, has ignored the plight of internally displaced persons (IDPs). [33]
Human rights conventions being flouted
The civil war in Nepal meets the Geneva Conventions definition of an internal armed conflict. [34,10] The Maoist rebels have an identifiable and organized command structure, both at the national and regional level, are in de-facto control of a significant part of Nepali territory [10]. Both the government of Nepal, which ratified the Geneva Conventions in 1964, and the Maoist rebels have agreed to abide by them. [10] One of the most fundamental protections during internal armed conflicts is contained in Article Three common to the four Geneva Conventions of 1949. This governs the treatment of civilians and captured combatants during internal armed conflicts, and outlaws summary executions, torture and other ill treatment of persons, the taking of hostages, and punishment without fair trial. This has been violated by both parties to the conflict.
In addition to the laws of war, the government of Nepal is a party to all the major human rights treaties, including the International Covenant on Civil and Political Rights (ICCPR),[35] which it acceded to in 1991, and the Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment.[36] which prohibits arbitrary arrest and detention, torture and other mistreatment, enforced disappearances, and extrajudicial executions. Efforts initiated by the National Human Rights Commission of Nepal, established in 1999, to have a human rights accord signed have failed. In Jan 2005 the United Nations Human Rights Commissioner; Louis Arbour criticized both the governments and the Maoist leaders in Nepal for not doing enough to tackle human rights violations. [37] The reason for the human rights abuses has been the impunity enjoyed by the security forces under the Terrorist and Disruptive practices Act (TADA). [38]
Political paralysis with dozens of failed governments in the last decade and a crumbling monarchy that witnessed fratricide, patricide and matricide in one chilling day 4 years ago have meant a loss of confidence by the Nepalese people in the political process.[39] And, the recent political turnover has made the situation worse by suspending the fundamental rights to freedom, expression, information, property and free travel. Human rights abuses are increasing as a result of the new regime's deliberate involvement in creating armed militias and other vigilante groups throughout the country encouraging them to conduct offensive attacks against civilians in the name of resisting the Maoists. [40]
Role of the international community
The international community is sharply divided between supporting the Nepalese army, with a dubious record of abuses on one hand, and the brutal Maoist rebel movement on the other. Previously, the US, India and the United Kingdom had supported the Nepalese government with weapons; the US supplying US$29 million US in military aid from 2001 to 2004, [10]largely viewing the Maoist problem as a part of its global "war on terror". On the other hand, the European Union and the United Nations have condemned both sides for human rights abuses. No government has supported the Maoists. Although Nepal's conflict is mainly internal and will require indigenous solutions addressing decades of poverty and inequality, with antagonistic and uncompromising political visions, it is difficult to envisage a solution without the intervention of a third party. It is imperative that the international donor community, which provides for nearly 60 % of Nepal's development budget, particularly the states most active in Nepal – India, the U.S., the U.K. and the European Union – should act decisively and in concert to promote adherence to international human rights and humanitarian law in Nepal. The recent political developments in Nepal have led to a temporary curb in military and developmental aid to Nepal by several countries, although it is unlikely that the government will cut down on military spending, thus developmental projects in the rural areas might be the hardest hit.
Role of humanitarian and development NGOs and conflict sensitive development
While Nepal is flooded with NGOs, and most of them are developmental NGOs paradoxically, development assistance may have unknowingly exacerbated the conflict by perpetuating the same inequalities, which led to the conflict in the first place. Many international agencies have inadvertently contributed to the conflict by raising the expectations of the rural poor. NGO projects have mainly benefited the urban majority while the rural minority still suffers in poverty.[41] Foreign aid, which accounts for nearly 60 % of Nepals developmental budget, may have paradoxically contributed to lopsided development in Nepal. While aid money has favored urban development 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 of centralization of effective power on the one hand, and maintenance of the economic, social and political status quo, on the other hand. The Maoists have forced several international agencies to leave remote western regions, where help is needed the most, while the government has put several administrative roadblocks in the way of international agencies working in Nepal.[42] 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 [43]. Offices of non-governmental organizations (NGOs) have been burnt and volunteers are afraid to work.[43] Even Médecins Sans Frontières (MSF) – not a development NGO but a humanitarian NGO – was forced to curtail its activities last year in Jumla, one of the poorest districts in Midwest, due to the conflict. In May 2005 four international agencies the World Food Programme, Britain's DFID and German GTZ and Dutch SNV aid agencies suspended their program in western Nepal as the rebels attacked aid-workers. [44]
In the context of Nepal – as in many other wars – a more nuanced approach to humanitarian relief and protection and development agendas would be helpful-one that recognizes a clear distinction between humanitarian relief and development. They are not the same and should not be lumped together. The distinction is critical in that it can mean the difference between the relief of the immediate suffering of war, or not. Humanitarian relief is to be given in a manner consistent with universal medical ethics; on the basis of need alone; impartially, with the giver as a neutral agent between the parties to the conflict (with crimes against humanity, war crimes, genocide and egregious violations of international humanitarian law being the outer limit of neutrality), and in a fashion that is independent of government, rebel or third party interests. Meeting all these criteria can conflict with a development agenda that seeks to overcome structural inequities that can be both cause and the conditions of war. It is not that inequities must not nor cannot be addressed, just not in a structural manner by actors who operate under the aegis of humanitarianism in war. To do so is to effectively seek to rewrite public policy and practice, which are in many cases in war, contested and a principle issue of contention in the conflict. To seek to do this is to effectively engage the politics of the particular war, which necessarily means taking sides – sometimes different sides at different times, but taking sides none-the-less, thus endangering the direct provision of humanitarian relief and protection in war. It is not possible – practically or politically – for an agency – NGO or otherwise – to pursue a development agenda in a war while simultaneously providing humanitarian relief.
Development NGOs in Nepal may consider pursueing an approach of "conflict-sensitive development"[44] – development sensitive to the (conflict) environments in which NGOs operate, in order to reduce the negative impacts of their activities – and to increase their positive impacts – on the situation and its dynamics. Development projects can continue in less affected areas with a need for transitional programs in conflict areas that can adapt to the rapidly changing environment. Some agencies have adopted a participatory role in development and have involved neutral local agencies, increasing community participation in their projects with good success. There is also a need for increasing coordination between organizations working in various projects.[45] Multicultural, multilingual and multiethnic representation and participation are essential components in the design of any successful developmental programmes in Nepal. Remedies for health inequities must come not only from the health sector but also from broad social policies that address potential health gaps related to equity e.g. distribution of income. Reducing illiteracy might significantly decrease the vulnerability of women to the effects of other health risks.
Recommendations
There is a need to provide immediate assistance to internally displaced persons; protect the independence and effectiveness of the judiciary; ensure the continued independence, institutional continuity and stability of the NHRC; and ensure the full and unimpeded access of the NHRC, the office of the High Commissioner of Human Rights (OHCHR) and the International Committee of the Red Cross to all detention centers, including Royal Nepal Army barracks. [10] The Nepalese justice system, which lacks independence, training and resources, must step up to the challenge. The OHCHR and other internationally recognized organizations must embark on continuing education programmes and intensive human rights education for Nepalese judges and prosecutors. The Nepalese government and army must allow such persons to carry out their duties without pressure or threats. [46]
It is necessary that both sides comply with international human rights and humanitarian law, in particular prohibitions on attacks on civilians; executing or ill-treating persons in custody; committing "disappearances," abductions and unlawful arrests; and committing acts of extortion or looting.[10] There is a need to investigate all allegations of abuse and appropriately prosecute the perpetrators in accordance with international fair trial standards. There is also an urgent need to implement a human rights accord, which abides by the Geneva conventions and commit both the government and Maoists to abide by clear human rights standards and accept human rights monitoring. This Accord was drawn up by the NHRC and widely promoted by the international and human rights community, including by the High Commissioner for Human Rights during her visit to Nepal in January 2005. The Accord would be a valuable confidence building measure towards future peace negotiations.
The recent establishment of an Office of the High Commissioner for Human Rights in Nepal is an important step towards protecting human rights in Nepal. [47]. Under the April 10 Memorandum of Understanding signed between the government of Nepal and the U.N. High Commissioner for Human Rights the mandate of the mission is to "monitor the observance of human rights and international humanitarian law, bearing in mind the climate of violence and the internal armed conflict in the country. [48] If the mission is to succeed, this mandate must be interpreted broadly to cover all aspects of human rights violations, and not be restricted to reporting on specific cases. If investigations are limited to specific incidents of torture, killing and abduction, the non-functioning of judicial institutions as a primary cause of these abuses will be missed. The U.N. mission will have to engage closely with local people and sympathetic and interested Nepalese, both inside and outside the country. While both sides have welcomed the UN monitoring of human rights in Nepal it will remain to be seen whether this leads to an improvement in the human rights situation. The difficult geographical terrain of Nepal and limited communication links makes the process of human rights monitoring a major challenge. Although the agreement is clear, the international community must remain vigilant to ensure that this agreement is complied with effectively and fully
The international community was effective in reducing the number of disappearances last year and will have to use its leverage in reducing the number of killings. Given the importance that Nepal places on its international image and its dependence on international assistance, the position that the international community adopts will be of critical importance in the coming months. It is therefore important that the international community, when sending a strong message about the importance of restoring democracy, stresses that this must be a democracy with human rights and protection for a pluralist civil society at its core. [49] Failure to address these issues at this critical hour runs the risk of leaving Nepal on a slippery slope of chaos and anarchy. In the meantime, the "People of Nepal", in whose name the war is being fought, will continue to be its main casualty, as they face renewed threats of violence, displacement and hunger with every passing day.
Acknowledgements
The authors wish to acknowledge the help of James Orbinski for his views on the role of NGOs.
==== Refs
Nepal news Proclamation to the nation from his majesty King Gyanendra Bikram Shah Dev Nepal news Kathmandu, Nepal Feb 1 2005. accessed on April 22 2005
Ministry of Finance Economic survey, fiscal year 2001/2002 2002 Kathmandu: His Majesty's Government, Ministry of Finance
UNDP Nepal Human Development Report 2001 – Poverty Reduction and Governance 2002 Pulchowk, Kathmandu, Nepal
Pradhan A Aryal R Regmi G Ban B Govindasamy P Nepal Family Health Survey 1996. Kathmandu: His Majesty's Government, Ministry of Health 1997
Ministry of Health Nepal demographic and health survey 2001 2002 Kathmandu: His Majesty's Government, Ministry of Health, Family Health Division
United Nations Development Programme Human development indicators 2003: Nepal Human Development Report 2003 2003 Oxford: Oxford University Press (accessed 2005 May 4).
Gurung YB Indigenous Peoples Plan for Rural Water Supply and Sanitation (RWSS-II). Report submitted to The World Bank SAMANATA, Institute for Social and Gender Equality Kathmandu 2002
Singh S Impact of long-term political conflict on population health in Nepal CMAJ 171 1499 501 2004 Dec 7 15583204
Jimba M Wakai S Is Gandhi's philosophy of non-violence dead in Nepal? Lancet 2001 358 1017 8 11586993 10.1016/S0140-6736(01)06148-7
Human Rights Watch Between a Rock and a Hard Place: Civilians Struggle to Survive in Nepal's 2004 accessed May 4 2005
Amnesty International Nepal Killing with impunity Amnesty International London UK Jan 20 2004 Accessed Jan 23 2004.
Amnesty International Nepal Human Rights Defenders under threat Amnesty International, London, UK 28 July, 2004 (Accessed Feb 1 2005)
Amnesty International Nepal Fear for safety Fear of disappearance Amnesty International, London, UK 29 July 2004 (Accessed Feb 1 2005)
Informal Sector Service Centre (INSEC) Human rights yearbook 2004 2005 Kathmandu: INSEC Accessed April 22, 2005
Stevenson PC The torturous road to democracy – domestic crisis in Nepal Lancet 2001 358 752 6 11551600 10.1016/S0140-6736(01)05844-5
Centre for Victims of Torture No. Of torture victims increases five-fold due to conflict: CVICT 2004 Nepal news Kathmandu Nepal Accessed Jan 28 2005
Lykke J Timilsena MN Evidence of torture in the mid west region of Nepal Dan Med Bull 2002 49 159 62 12064092
Dolma S Singh S Mills E The dangerous journey of Tibetan refugees fleeing to Nepal (unpublished)
Fazel M Wheeler J Danesh J Prevalence of serious mental disorders in 7000 refugees resettled in western countries: a systematic review Lancet 2005 365 1309 14 15823380 10.1016/S0140-6736(05)61027-6
Maskey M Practicing Politics as Medicine writ large in Nepal Development 2004 47 122 130 10.1057/palgrave.development.1100039
Sharma GK Osti B Sharma B Physicians persecuted for ethical practice in Nepal Lancet 2002 359 1519 11988269 10.1016/S0140-6736(02)08422-2
Anon Government warning to medics The Kathmandu Post 345 2001 Kathmandu, Dec 1
The Kathmandu Post 30 pc of population suffering from psychiatric problems 2002 Dr. Shrestha, Kathmandu, The Kathmandu Post May 20 2002 (accessed Jan19, 2005)
Central Bureau of Statistics Statistical yearbook of Nepal 2001 2001 Kathmandu: His Majesty's Government, National Planning Commission Secretariat
Maskey M Political Disruption, Violence against Women and its Impact on Health in Nepal Development 2003 46 68 71 10.1057/palgrave.development.1110449
Pettigrew J Shneiderman S Women and the Maobadi Ideology and Agency in Nepals Maoist movement 2004 Himal South Asian magazine, Kathmandu accessed April 22 2005
Mishra D Nepal NGOs fear rise in women trafficking Times of India Jan 17 2005
UN Office for the Coordination of Humanitarian affairs Nepal; Focus on Children and the insurgency 18 Jan 2005
Asian Human Rights Commission Children and the People's War in Nepal: 146 Children Killed, 2000 Orphaned, 4000 Recruited and 4000 Homeless ACR Weekly Newsletter 02 22 January 2003. (Accessed Feb 1 2005)
Singh S Post-traumatic stress in former Ugandan child soldiers Lancet 2004 363 1648 15145645 10.1016/S0140-6736(04)16217-X
Ramana PV Nepal: Children in Maoist Ranks. Institute of Peace and Conflict Studies Article No 1400 31st May 2004. accessed May 4 2005
Dhakal S Nepal Communists plan to train 50, 000 child soldiers One World South Asia 27 February 2004 (Accessed Jan 19 2005
Norwegian Refugee Council Nepal A hidden IDP crisis 10 Oct 2003 (Accessed Feb 1 2005)
International Committee of the Red Cross, Commentary, IV Geneva Convention 1958 Geneva: International Committee of the Red Cross
United Nations General Assembly Resolution 2200 A (XXI) of December 16, 1966 (entered into force on March 23, 1976.)
United Nations General Assembly Resolution 39/46 of December (entered into force on June 26, 1987.)
British Broadcasting Corporation Nepal gripped by political crisis BBC corporation Accessed Feb 1 2005.
Nepal An appalling situation Human rights defenders increasingly victims of the internal armed conflict The Observatory 2005 accessed May 4 2005
Marie Lecomte-Tilouine David GellnerN Regicide and Maoist revolutionary warfare in Nepal: Modern incarnations of a warrior kingdom Anthropology Today 2004 20 13 10.1111/j.0268-540X.2004.00247.x
Charles Haviland Nepals rising vigilante violence British Broadcasting Corporation 14 March 2005
Setting priorities for preventive action in Nepal Final report of the Conflict Prevention Initiative Web Conference 2001 Cambridge (MA): Program on Humanitarian Policy and Conflict Research, Harvard School of Public Health (accessed 2005 April 22.
Poudel K Nepal Maoist threat prompts NGOs to retreat One World South Asia 06 April 2004 accessed April 22 2005
Lawday A HIV and Conflict A double emergency 2002 London Save the Children UK
British Broadcasting Corporation Nepal rebels" beat aid workers" 2002 accessed May 19 2005
Gaigals C Leonhardt M Conflict-sensitive approaches to development Safe World, International Alert, and the International Development Research Centre 2001 Accessed April 22 2005
Human Rights Watch Nepal concern for Nepal: Human Rights Concerns for the 61st Session of the U.N. Commission on Human Rights March 10 2005 accessed May 4 2005
United Nations Office of High commissioner for Human rights set to monitor situation in Nepal United Nations Press Release 11 April 2005 accessed April 22 2005
Asian Human Rights Commission Comprehensive monitoring of Human Rights in Nepal Hong Kong May 17 2005
Amnesty International Nepal A long ignored human rights crisis is now on the brink of catastrophe 10–16 Feb 2005 accessed May 4 2005
|
15985165
|
PMC1183233
|
CC BY
|
2021-01-04 16:39:31
|
no
|
Int J Equity Health. 2005 Jun 28; 4:9
|
utf-8
|
Int J Equity Health
| 2,005 |
10.1186/1475-9276-4-9
|
oa_comm
|
==== Front
Front ZoolFrontiers in Zoology1742-9994BioMed Central London 1742-9994-2-101596084810.1186/1742-9994-2-10ReviewThe role of vocal individuality in conservation Terry Andrew MR [email protected] Tom M [email protected] Peter K [email protected] IUCN – The World Conservation Union, Regional Office for Europe, Boulevard Louis Schmidt 64, 1040 Brussels, Belgium2 127 Kennington Avenue, Bishopston, Bristol BS7 9EX, UK3 Duchy College, Stoke Climsland, Callington PL17 8PB, UK2005 16 6 2005 2 10 10 21 3 2005 16 6 2005 Copyright © 2005 Terry et al; licensee BioMed Central Ltd.2005Terry 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.
Identifying the individuals within a population can generate information on life history parameters, generate input data for conservation models, and highlight behavioural traits that may affect management decisions and error or bias within census methods. Individual animals can be discriminated by features of their vocalisations. This vocal individuality can be utilised as an alternative marking technique in situations where the marks are difficult to detect or animals are sensitive to disturbance. Vocal individuality can also be used in cases were the capture and handling of an animal is either logistically or ethically problematic. Many studies have suggested that vocal individuality can be used to count and monitor populations over time; however, few have explicitly tested the method in this role. In this review we discuss methods for extracting individuality information from vocalisations and techniques for using this to count and monitor populations over time. We present case studies in birds where vocal individuality has been applied to conservation and we discuss its role in mammals.
==== Body
Review
Introduction
Signals can contain information useful to conservation [1,2]. Until recently, communication behaviour had a limited role in conservation, being restricted to enhancing captive breeding programs [3] or use in species counts [4]. However, knowledge of how individuals within a population communicate and what they are communicating can generate information ranging from measures of habitat use to genetic fitness [2,5] that can be applied to conservation and that may neither be possible nor desirable to extract using other methods. In this review we shall concentrate on a subsection of communication behaviour that underlies most attempts to gain useful information from signalling, namely individuality. We discuss the different applications and types of information that can be extracted using vocal individuality. Further, we consider the different methods currently applied and the results gained with different taxonomic groups. Finally, we discuss some of the limitations and future directions that this technique can take. As a consequence of our area of research we have concentrated our discussion on the role of acoustic vocal individuality in conservation, although the principles involved equally apply to other signalling modalities, with the possible exception of chemical signals where the techniques may currently be lacking.
Vocal individuality
A pre-requisite for discrimination and individual recognition is that, in the signals being used, there should be low within-individual variation and high between-individual variation [6,7]. Many studies have shown the presence of individually distinctive vocal features in a wide range of animal species and it seems that vocal individuality is most likely a feature of all vocally active species and is caused by a series of genetic, developmental and environmental factors [8,9]. The level of individuality and the difficultly in extracting and using it will differ between species and we discuss these issues later in the review (also see [10-12]).
When discussing the uses of vocal individuality we distinguish between two commonly confused terms; discrimination and identification. Discrimination requires that individuals differ enough at one point in time to be separated. Identification requires that an individual's vocal features remain constant enough to be associated with that individual for periods of time. Identification is harder to demonstrate and has more useful conservation applications than discrimination [2,13]. Discrimination is limited to census tasks, but identification allows individuals to be monitored over time, which can yield useful life-history information [2,13]. It is possible to have discrimination without identification but not vice versa and the extent to which individuals can be identified should be carefully considered in any cost/benefit analysis of using vocal individuality as a monitoring tool.
Vocal individuality and conservation
Although several studies have suggested that vocal individuality could have a role as a useful conservation tool [5,14,15], it has been rarely used or tested in census or monitoring roles (some examples of use and testing are given below). In the following sections we discuss some of the different aspects of conservation biology where the identification of individuals can either yield useful data or be used as a non-invasive marking technique that may offer an alternative to invasive approaches such as mark-recapture or tagging.
Conservation models and identifiable individuals
Several studies by Sutherland, Goss-Custard and colleagues [16-20] have shown the importance of individual differences in predictive models. Assumptions of individual equality have largely come from difficulties associated with collecting and then modelling the baseline demographic data [21]. In most cases individual identification is required to extract this demographic data.
Population viability analysis (PVA) is the most widely used form of conservation model. These models attempt to predict the potential fate of a population given a set of demographic and environmental parameters. PVA models include demographic stochasticity, that is, fluctuations in population size due to random individual differences [22], as a form of 'sampling error' [23]. Mean expected survival and fecundity are estimated from a limited data set and made equal for all individuals; stochasticity is included as the level of variance around this mean when survival is calculated for the model population [23,24]. Increasing the variance increases the risk of extinction faced by the population. However, these assumptions are biologically untenable. The inclusion of non-random individual variation acts to reduce the effects of demographic stochasticity, leading to the result that the predicted extinction risk faced by populations may be overly pessimistic [23,24]. By identifying individuals, it is possible to generate more accurate predictions concerning the nature of individual variation in parameters such as fecundity and survival, which will increase the accuracy of predictions of PVA models [17,23].
Behavioural traits
As well as affecting large-scale population models, identifying individuals can highlight behavioural traits that may affect the conservation value (in terms of effort and resources) of different subsections of a population [5]. In many species there are large asymmetries in lifetime reproductive success, with only a few individuals contributing to the next generation. For example, 17% of a population of common buzzards (Buteo buteo), accounted for 50% of the following year's fledglings [25]. Other bird species have yielded similar values [26]. Kelly [27] showed that a few (8%) cheetah (Acynonyx jubatus) matrilineages produced over 50% of the population over a 20-year period, and studies in other mammals have yielded equally high differences in reproductive success [28,29].
Individuals within populations also differ in whether they are residents or floaters. Floating individuals often occupy lower quality habitat, possibly representing 'sink' populations, and generally do not breed or breed with lower success [30-32]. This difference in breeding status has obvious implications for management strategies. Monitoring fluctuations in the proportion of floating individuals may also be a more revealing indicator of habitat quality [5]. Also, if translocation is considered as a management option, floating individuals may be suitable for translocation [33], as they may show reproductive success equal to other residents when moved to suitable habitat [30].
Sampling bias
In most cases it is not possible or desirable to count an entire population; usually, census techniques take samples from a population and use them to estimate the true population size [34,35]. All sampling techniques make assumptions about the populations they are counting and violation of the assumptions leads to biases in the estimate. Unless accounted for, biases can be pervasive and generate misleading census estimates. Accounting and testing for bias within a census can be difficult without either a complete knowledge of the population or a subset of identified individuals [34]. For example, the use of playback to elicit responses from individuals in acoustic surveys can cause a bias by only attracting certain individuals [36,37]. Using radio-telemetry, Conway et al [36] found both a seasonal and density-dependent bias in response to playback by Yuma clapper rails (Rallus longirostris yumanensisi). Playback elicited responses from 10–40% of males present. This also caused biases in the identification of preferred habitat for the rails [36]. A similar study with black rails (Laterallus jamaicensis) showed differential responses to playback due to factors such as breeding status, with non-nesting males being the most likely to respond [37]. Ogutu & Dublin [38] found playback to be an affective census tool for lion (Panthera leo) and spotted hyena (Crocuta crocuta) populations; however, placement of playback speakers near range borders instigated territorial disputes and prey abundance was also a significant source of bias.
Floating individuals can bias a census as they are usually more mobile than resident individuals. For example, floating male bitterns (Botaurus stellaris) were repeatedly counted in acoustic surveys causing a population of 15 individuals (nine residents and six floaters) to be counted as 25 [39]. This bias was detected by applying vocal individuality information to the population (further details below).
Sampling techniques that catch and mark individuals within a population are likely to also suffer from biases [5]. For example, the key assumptions of most mark-recapture approaches are that individuals are equally likely to be caught and that marks do not affect their subsequent behaviour and survival [34]. However, several studies have shown that capture techniques often only mark a subsection of the population (e.g. [40]). For example, radio-tracked corncrakes (Crex crex) called on 75–80% of nights during one breeding season and this was used to generate census guidelines for that species [41]. However, a subsequent study found a far lower calling rate (~ 40%) in a different population, which led to under-estimates of the population size [40].
Welfare considerations
Although marking techniques can allow the unequivocal identification of individuals, invasive marking of individuals can have both short-term and long-term effects on those individuals. Short-term effects include direct costs of the capture and handling process itself [5]. Longer-term costs include avoidance of the capture area, stress-related susceptibility to disease, increased susceptibility to predation and loss of subsequent reproductive success (e.g. [42]). Both short-and long-term effects of marking are contentious issues, with often equal numbers of articles showing the presence or absence of an effect. An example of this is the debate surrounding the causes of extinction of the Serengeti population of wild dogs (Lycaon pictus). The rapid spread of disease through the population was correlated with capture and handling [43]. However, recent analysis has shown that this relationship was likely to be non-causal and that the administered vaccines failed to take effect during an outbreak of the disease [44]. Another example comes from studies attempting to show effects of neckbands on survival in goose species [45-48].
It can be difficult to study the effects of capture and handling and the subsequent effects of marks on individuals unless there is an individually identifiable (or non-invasively marked) control group for comparison. For example, Nimon et al [49] used artificial eggs to non-invasively measure heart rates in Adélie penguins (Pygoscelis adeliae) and showed that handling individuals pre-disposed them to extreme reactions to human-induced stress [49,50]. However, it is safe to assume that the capture and handling of animals can only have deleterious effects, even if they may not be immediately obvious. Careful consideration has to be given to balancing benefits gained by invasive marking against potential adverse effects on the individuals or population. In our view, the default should be non-invasive marking techniques, such as vocal individuality – a view that has increasingly been adopted by journals in their recommendations to authors (e.g. [51]).
Life-history parameters
Techniques that identify individuals are used to generate life-history data, such as habitat use, survival, recruitment, immigration and emigration. This information can then be used to test hypotheses concerning the agents of decline facing a population or test the effects of management strategies [35,52]. Radio-telemetry is often considered the most productive monitoring technique. However, aside from concerns inherent in catching individuals and biases in data generated, radio-tags have a limited operational life, often of one year or less [53]. If vocalizations remain constant over time, they can provide a long-term monitoring option [53,54].
Summary
We have reasoned that vocal individuality should be widespread (if not ubiquitous) and can generate useful conservation information. Also, as a non-invasive marking method it may provide less biased data than other marking techniques and have fewer adverse welfare implications. We now discuss methods for extracting vocal individuality and some instances in which it has been applied to conservation questions in birds.
Methods used in vocal individuality
The methods used to collect and analyse vocalizations will, in many cases, determine whether vocal individuality is used as monitoring tool. As the amount of equipment, analysis time and specialist knowledge required to discriminate between individuals increases, the likelihood of the technique being used decreases. Thus, in this section we discuss some general points concerning the collecting, processing and analysis of vocalizations for individuality.
Equipment and analysis
Before any statistical analysis takes place, there are a number of practical considerations concerning the process of converting an animal's vocalization into data points. Although some of these considerations may seem trivial, ignoring them can render the approach useless before it has started. This section highlights some of the points that we have encountered in our experience with sound recording and analysis; many texts (e.g. [56]) provide in-depth discussion of these topics.
The best type of microphone to use is determined by features of the vocal signal, predominantly its frequency range [57,58]. Shotgun and parabolic microphones are most commonly used for individuality studies and have different characteristics [57]. Parabolic reflectors amplify the signal entering the microphone and are more directional than shotgun microphones, reducing the reverberations in recordings [57]. However, parabolas cannot have a flat frequency response because the parabola diameter determines the extent to which each frequency is amplified and also determines the lowest frequency that can be recorded (e.g. [57,58]). Species with low frequency vocalizations, e.g. European bitterns (frequency range 100–200 Hz) should be recorded with a shotgun microphone with a flat response in this frequency range.
Similarly, the best type of recorder to use is determined by its ability to capture the sounds of interest with the necessary temporal and frequency resolution. Manufacturers' performance specifications are a starting point, although these may often be optimistic, subject to change with age of the machinery and may differ between individual recorders of the same model (particularly analogue tape recorders). It is therefore advisable to include a calibration sound of known frequency and duration at the beginning of a recording session. Digital recorders (solid-state or DAT) are limited by the sampling rate of their analogue to digital converters, so special care is necessary when recording high frequency sounds. Increasingly, field recordings are being made with devices using data compression to increase storage capacity (e.g. MiniDisc). The data compression algorithms remove parts of the signal that are not perceptible to humans, but they can also remove or distort important biological signal components [57].
When recordings are being made, the record level must be adjusted to prevent the signal becoming overloaded and the consequent introduction of distortion. We recommend that the recordist adjusts levels during recording as automatic gain controls rarely adequately control record levels of animal signals. In many cases it will be advisable to include a high-pass filter between the microphone and the recorder as this has the effect of removing low frequency background noise and signal strength is better represented by the recorder's level meters.
When analogue signals are digitised (either during recording or analysis) the highest frequency of interest must be less than half the sampling frequency (the Nyquist frequency), otherwise the higher frequencies will get mirrored in the new digitised signal [59-61]. Low-pass filtering should be used to remove sound from above the Nyquist frequency and filtering to remove unimportant parts of a recording is a method of minimising the amount of signal processing required [61].
Once a signal has been digitised and filtered, it can be measured. Although complex signal processing, as well as automated measuring algorithms, are becoming increasingly common, most measures are taken from three forms of signal representation; waveform (Fig. 1A), power spectrum (Fig. 1B) and spectrogram (or sonagram, Fig. 1C). The waveform shows changes in amplitude (acoustic pressure) with time and is suitable for taking temporal measures. The power spectrum is generated from the waveform and shows changes in amplitude at different frequencies; this is suitable for measurements of features such as dominant frequency. A spectrogram is a three-dimensional display plotting frequency against time, with amplitude shown as intensity of grey scale or in colour. Spectrograms are very useful for visualising sounds and can be used in qualitative comparisons (e.g. [11,53]), but should not be used for direct measurements. This is because there is a fundamental trade-off when producing spectrograms between the resolution of time and frequency information; high resolution cannot be achieved in both features [62]. We suggest that waveforms should be used for temporal measures and power spectrum for frequency measures. Reported measurements should also always be accompanied by the minimum difference (the cursor increment) that could be measured in addition to resolution limitations imposed by the sampling frequency of A/D conversion.
Figure 1 Three common forms of signal representation. An example of a corncrake call displayed as: (A) a waveform which plots temporal information on the x axis and amplitude on the y axis; (B) a spectrogram which plots temporal information on the x axis, frequency on the y axis and amplitude in the image greyscale; (C) a power spectrum, which plots frequency information on the x axis and sound pressure level on the y axis. The spectrogram was made with a 2 msec time step and a 20 Hz frequency step (Hamming window).
Qualitative assessment of vocal individuality
Qualitative comparisons of sounds can be either visual, through graphic representations such as spectrograms, or aural through recordings or listening in the field. Aural identification is possible and field researchers with extensive experience of their study species can often identify individuals by ear [11]. However, there are few studies that address the use of aural comparisons as a census technique. Gilbert [63] showed that discriminating and identifying individual European bitterns by ear was affected by experience and that only small sample sizes could be discriminated.
More common is the use of visual representations of sounds such as spectrograms to discriminate between and to identify individuals [11,12,14,15,64-67]. In most cases, qualitative comparison is used as the first level of analysis, for example, in selecting sounds to be used in further quantitative tests [67]. In some cases qualitative comparisons have proved more effective than quantitative approaches [68]. Janik [66] compared several techniques used to discriminate between bottlenose dolphin (Tursiops truncates) signature whistles and found that visual comparison was the most effective at separating individuals.
The main advantage of qualitative assessment is that observers are capable of complex pattern recognition, perhaps using the overall 'gestalt' of an image in a comparison [2,65]. Features that may not be adequately represented in quantitative tests are considered in qualitative comparisons. Similarly, changes in small-scale measures can be weighted equally with large-scale features. The use of qualitative assessment contains several potential disadvantages associated with the subjectivity of human decision-making, which can lead to biases or inaccuracies [65,67,69-71]. It is possible to address these issues by measuring the repeatability of an observer's classifications; however, an observer with high repeatability may not make reproducible classifications [71]. It is possible to measure the degree of inter-observer reliability and use this to determine whether different observers can be used to classify the sounds (see [67] for discussion of inter-observer reliability scores).
Quantitative assessment of vocal individuality
Multivariate approaches are commonly applied to vocal individuality studies and are particularly suited to classification tasks [72]. They can be broadly divided into those that discriminate between individuals by finding differences between them and those that find similarities between them. We shall discuss each in more detail.
Discriminant function analysis (DFA) is a multivariate difference statistic commonly used to show which variables best discriminate between two or more groups. It does this by combining the variables with weighting coefficients to create a set of functions that can discriminate the groups. Once these functions are established, they can be used to classify new data into one of the pre-existing groups. This corresponds to the three main uses of DFA in vocal individuality; establishing variation, data reduction and classification. Establishing variation and paring down the parameters used are the first steps in any investigation into vocal individuality. Invariably many measures are taken from the vocalizations (often >20) and not all of them will be effective in discriminating individuals. For example, Gilbert et al [53] measured 23 features of bittern boom trains and used DFA to reduce these down to 7 features which allowed effective discrimination. A specialised form of this analysis is called stepwise discriminant analysis (SDFA) and it enters variables, one by one, into the analysis until there is no further increase in discrimination accuracy [73]. This can be a powerful way of extracting an 'optimal' subset of features and reducing further analysis time [72,73]. Once the discriminant functions have been established, they can be used to classify cases to the pre-existing groups. Measures of classification success (percent cases correctly classified) are often cited and are most often high (>80%). However, this measure does not test the functions' generality [73]. The discriminant functions must be validated with data not used in their creation. Most common statistics packages allow validation by not using some of the cases to produce the discriminant functions, and then classifying these unused data. This validation can either be a jackknife, where half the cases are left ungrouped and then classified, or a leave-one-out test where one case at a time is ungrouped and classified. Classification scores without any validation are almost meaningless in the context of their applicability to vocal individuality. Many studies have used DFA to show vocal individuality in avian [10-13,74,7], canid [78,79] and primate [80] species. Implicit in many of these studies is a potential use of DFA to generate conservation data. However, the question remains whether DFA on its own can generate useful conservation information. In reality the role of DFA classification in conservation is limited [2,81]. This is because in all published examples of the use of DFA for vocal individuality the type of DFA used can only classify vocalizations to particular individuals if all individuals are known; it cannot accommodate vocalizations from new individuals. This limitation is overcome by a non-parametric form of DFA known as adaptive kernel-based DFA, in which the range of values for inclusion into the kernel of existing groups is defined.
Similarity techniques offer a different approach to discrimination that avoids most of problems experienced when using DFA. They do not require complete knowledge of the population being monitored [2,81]. When using similarity techniques two cases are compared and if they are within a pre-defined threshold, they are classified as coming from the same individual. If a new individual joins the population, it should be outside the threshold for all known individuals. The two most commonly used approaches are acoustic space, which compares measures taken from vocalizations and cross-correlation which is used to compare sounds directly [2]. When a series of measurements are taken from a sound they can be used as coordinates that define the location of that vocalization in an acoustic space whose dimensions are determined by the number of variables [13,82]. The Euclidean distance between locations is used as the measure of similarity. This approach has been successfully used to identify individuals within and between seasons in several non-oscine bird species [13,53,83]. Cross-correlation is used to compare representations of whole sounds, most often these are spectrograms [59,84]. Two spectrograms are incrementally passed over each other and at each stage a Pearson correlation coefficient is calculated. The maximum correlation value is used as the similarity measure [84]. Many cross correlation routines move spectrograms only along the time axis and thus two sounds with identical frequency contours but centred on different frequencies would give a low similarity value. This is avoided by routines that move along time and frequency axes. The only conservation application of cross-correlation that we are aware of was to monitor wild turkey (Meleagris gallopavo mexicana) populations [85]. However, it has also been used show individuality [12,86-88], dialect differences [89], vocal learning and development [69,90] in several species of songbird. The advantage of cross-correlation is that it considers the entire sound objectively. However, care has to be taken with the particular sound types being compared, the amount of noise included in the signal and with the settings used to create the representations, as these factors will adversely affect the similarity value generated (for further discussion see [81,84,91]). Once a similarity measure has been used, the distributions of within-individual and between-individual similarity values can be compared (see Fig. 2). With an ideal similarity measure, the two distributions would not overlap, i.e. there would be no chance of making a false identification or discrimination [13]. However, in reality a certain amount of overlap will always occur and the extent of this overlap can be used to show how effective the technique will be (see Fig. 2 and [2,81,92]). Note that it will be impossible to maximise correct identifications and simultaneously minimise false identifications. However, studying the area of overlap can aid the setting of a threshold.
Figure 2 Examples of distributions of within-and between-individual similarity values. Distributions of within-individual (black bars) and between-individual (white bars) pair-wise comparisons in a similarity analysis. The ideal case (A) has no overlap between the two distributions; however, in cases of complete overlap (B) the technique becomes useless. The most common situation is one of partial overlap (C). The extent of this overlap can be used as a measure of confidence in the technique.
A number of complex non-linear models used for speech recognition have also been applied to bioacoustic signal analysis, and may represent the future direction for analysis tools as researchers try to compare increasingly complex signals. Models that have been used in bioacoustic studies include dynamic time warping [93-95], artificial neural networks [96-98], hidden Markov models [95] and linear predictive coding. These models can be divided into those that function by modelling sound production (linear predictive coding and hidden Markov models) or sound perception (neural networks). Currently, the most commonly used of these models are artificial neural networks. Originally designed as models of biological neural networks, they contain a network of inter-connected simple processing units that work in parallel to solve complex classification tasks [99,100]. The connections between the units are weighting coefficients. Neural networks solve classification problems by iteratively adjusting the weighting coefficients and combining them with the parameters until some pre-determined classification error value is achieved. The combination of simple algorithms used to classify highly optimised parameters makes them powerful and versatile tools. Neural networks have several advantages over other techniques. First, they are non-linear and can work with data that cannot be separated with linear classification tools [100,101]. Second, there is a huge range of different neural network types, and they have been applied to many classification and regression tasks. This provides a large source of reference material to draw on [98]. Third, they can be used to find clusters of similar vocalizations, set up groups based on those clusters and then classify new data to one of these groups or create a new group (e.g. Kohonen networks [102]). Neural networks have successfully identified individuals from their vocalizations in tungara frogs (Physalaemus pustulosus) [103,104], fallow deer (Dama dama) [96,105], Gunnison's prairie dog (Cynomys gunnisoni) [106], Stellar sea lions (Eumetopias jubatus) [97] and killer whales (Orcinus orca) [107]. Elsewhere we have shown that different neural network models were able to accurately count and identify individual corncrakes in a series of simulated census and monitoring tasks [98]. These types of model make it possible to create automatic analysis and identification systems [95], which will reduce analysis time, increase the amount of data that can be analysed and make complex classification techniques more available. All of these factors will be important in any conservation application. Note that, as with any statistical tool, they cannot be treated as complete black boxes, and some care has to be taken to adhere to the assumptions or limitations of the models (e.g. neural networks: [101], hidden Markov models: [95]).
Three case studies
Here we detail three cases where vocal individuality has been used to obtain conservation information on bird species or where the potential for generating such information has been investigated. Vocal individuality has been used to correct for census error and bias in bitterns and corncrakes, to follow habitat use in corncrakes and owls, to monitor annual turnover in bitterns and owls and to produce data on which to run survival models in bitterns.
European bittern
The bittern is IUCN red-listed and is strictly protected under the EU Birds and Habitats Directives due to large population declines (>50%) in the last 25 years [108]. Bitterns are found in dense reed-bed habitat and are secretive, making monitoring difficult [11]. The use of leg rings can be a useful method to identify individuals; however, bitterns are rarely seen or caught and the only practical census and monitoring method involves counting calling males [109]. Male bitterns have an individually distinctive long-range vocalization [11,65] – the boom – and booms are used to census and monitor individuals. The bittern is currently the only species we are aware of that is routinely counted and monitored using vocal individuality. Bitterns have been monitored in the United Kingdom in this way for over 10 years, allowing long-term population data to be extracted [53]. Gilbert et al [53] identified males between years by their booms and used these data as the basis for a survival analysis and to monitor movements between populations for the UK populations (although see [12], for possible differences with Italian populations of bitterns).
Terry [110] developed a semi-automatic method that discriminated between relatively large numbers (>40) of bitterns and could also extract the discriminating measures from poor quality recordings. The main discriminating feature was the dominant frequency of the boom and this was found to vary little over very short periods of time (days). However, recordings made between years of four radio-tagged individuals [53] showed within-individual changes over time in the dominant frequency that sometimes exceeded between-individual differences. Bitterns could be identified from year to year by taking several measures of the boom manually [53]. Studies of Italian bitterns also showed that frequency-based features of bittern booms vary over time, and re-identification can be difficult [12].
In areas where vocal individuality is not used, bitterns have been counted using acoustic surveys of calling individuals. In low-density populations, individuals show high site fidelity [53] therefore simply counting calling individuals may be adequate. However, populations at higher densities, where individuals compete more strongly for resources, are likely to contain a number of non-territorial individuals that as a consequence are mobile and vocally active (floaters). Unless these individuals are accounted for they can cause large over-estimates of population size [39].
Corncrake
Corncrakes are a species of land rail that primarily inhabit hay meadows and silage fields [41]. Although population numbers in the UK have increased in recent years due to increased habitat management (currently estimated at 600–700 calling males [111]), they are listed as a species of global conservation concern (IUCN red-listed, [108]) and are on Annex 1 of the EU Wild Birds Directive [109]. As the species is secretive and nocturnal, the only viable method to census a population is by counting calling males in the breeding season. Currently 90% of the UK habitat for corncrakes is monitored in this manner [109].
The census strategy used for corncrakes was developed from radio-tagging studies [41], which found that males rarely move more than 250 m between nights and that on any particular night around 75% of males would call. Based on these findings, censuses were carried out on two nights and if two calling locations were within 250 m of each other on both nights, they were judged to come from the same individual [41]. Corncrakes have individually distinctive vocalizations [13,112], which are consistent over years [13]. The census rules present obvious sources of bias and Peake & McGregor [40] monitored corncrakes using vocal individuality to show that males called far less than anticipated (41% of males per night) and this led to the population being underestimated by up to 30%. They were also able to follow individuals throughout the season and showed that habitat quality affected movement, with males in poor quality habitat moving greater distances. Their study also shows that tracking movements using the standard census approach was in most cases accurate [40]. The main role for vocal individuality in this case would be to provide correction factors to refine standard census rules or as a method for monitoring individual movements in relation to small-scale habitat differences.
Owl species
Owls, like most raptors, occur at low densities and are usually difficult to monitor. Most owl species are territorial and show high site fidelity after dispersal. However, monitoring territory occupancy, survival and recruitment can be difficult for these mostly nocturnal and secretive species. However, most owl species are characterised by their long-range vocalizations and these have been shown to be individually distinctive for several species (e.g. barred owls, Strix varia [113]; Northern saw whet owls, Aegolius arcadicus [114]; eagle owls, Bubo bubo [115]; screech owls, Otus asio [116]; Christmas Island hawk owls, Ninox natalis [117]; tawny owls, Strix aluco [75,118]; pygmy owls, Glaucidium passerinum [76]; African wood owls, Strix woodfordii [77] European scops owls, Otus scops [54] and Seychelles scops owls (Otus insularis [Peake, Currie & McGregor pers comm.]). These vocalizations seem to remain stable during a season and, where studied, were consistent for long periods of time [54,77,115].
Vocal individuality has proved to be a useful tool for mapping territories in these species and monitoring habitat use through the breeding season. Galeotti & Sacchi [54] showed a high annual turnover on breeding territories for European scops owls, with 55–78% of territories occupied by different individuals between years (note that vocal individuality was validated by previous studies in this species). They suggest that this high turnover may be related to their migratory habits and possible high winter mortality, which leaves many vacant territories at the beginning of the breeding season [54]. African wood owls are a sedentary species where both sexes call. They form monogamous pairs that maintain stable territories. Delport et al [77] recorded wood owls at different calling locations over a 12-year period to monitor territory occupancy and turnover for each sex. They found a turnover of 19.3% and 13.7% for males and females respectively. This is one of the few examples where both sexes can be monitored with vocal individuality.
In both the previously mentioned cases [54,77], individuals were not independently marked, and vocal stability is assumed based on previous studies of other owl species [75,76] and the presence of temporal stability in the call features used [77]. However, this assumption of vocal stability has to be treated with care, as if a vocalization changes between years, it will not be possible to determine whether the individual or the vocalization is different.
Potential and limitations of vocal individuality
We have explained how vocal individuality can have a role in conservation and we have presented case studies in which it has been used to generate useful data. For such a potentially useful technique, vocal individuality is surprisingly under-utilised. Our case studies show a heavy bias towards male territorial signalling in avian species. This bias is a natural consequence of an avian bias in bioacoustics and communication generally and the fact that territorial signals are the most obvious. However, we feel that there are other classes of individual and other taxa that could be represented in vocal individuality studies.
In this section we discuss some of the potential limitations of vocal individuality (concentrating particularly on vocal stability and sex differences in vocal activity), we indicate the sort of information required to assess its potential as a conservation tool and we explore the potential of vocal individuality in mammals.
Some limitations of using vocal individuality
All identification techniques contain a number of potential biases and disadvantages that need to be borne in mind in order to minimise the chance of producing misleading results or limiting their explanatory power. Four main problems are encountered when using vocal individuality. First, establishing the extent and stability of individuality requires intensive study with, preferably, independently marked individuals. Second, this intensive study requires knowledge of sound analysis and the equipment used for recording and analysis. Third, vocal individuality is biased towards the most vocally active sections of the population (and there are similar biases in any method based on acoustics). Many factors can affect which portion of a population is vocally active, for example sex (e.g. it is the males of most temperate bird species that are vocally active), age, time of year, breeding status, territorial status (e.g. great horned owls, Bubo virginianus [55]) or dominance. For vocal individuality to provide useful information, these sub-groups have to be known. Fourth, because vocal individuality uses natural variation, there will always be a level of ambiguity in the identification of an individual [2]. We consider the most important current limitations on the value to conservation of vocal individuality to be vocal stability and monitoring females, so we shall discuss these topics in more detail.
Vocal stability over time
Vocal stability over time is difficult to show because ideally it requires independent identification of individuals. The case studies involving corncrakes and bitterns (see above) both used a sample of radio-tagged males to collect recordings over time [12,41,53]. With the exception of Lengagne [115], studies involving owls have looked for temporal stability in vocal features with time as an indication of vocal stability [54,77]. They have also used previous studies showing long-term vocal stability in different owl species to support their assumptions [75,76], although none of those individuals were independently marked. When using vocal individuality to re-identify individuals, for example when monitoring populations, establishing vocal stability needs special consideration and, as with bitterns (see above) may require an increase in effort, equipment and knowledge. Such extra effort may limit the applicability of vocal individuality in monitoring contexts.
Monitoring females
For many species, especially avian, monitoring the presence and movement of females will be difficult without radio-tracking. For example females of most temperate avian species are quiet and do not have long-range vocalizations. Exceptions to this include some non-oscine species such as petrels and some owl species. Thus in many cases population estimates are reported as the number of calling males rather than pairs (e.g. corncrakes and bitterns [108]). Indirect measures of female presence are sometimes possible, for example male corncrakes cease calling for some days when they have attracted a mate [13]. There are also exceptions, for example, female brown-headed cowbirds (Molothrus ater) have a loud long-range individually distinctive call that can be used to identify them [119]. As mentioned above, male and female wood owls have long-range territorial vocalizations [77]. In cases where females do not have obvious long-range vocalisations, it may be possibly to exploit calls used to maintain contact between pairs, with offspring or with other social group members as a way of identifying and tracking females. For example, Campbell et al [97] showed that female Stellar sea lions could be identified from their mother-pup calls.
Colonial breeding species, such as many species of seabird and marine mammal, rely on contact calls (and other cues) to locate their mates and offspring [6,97,120,121]. As such, the vocalisations of these species have been shown to contain high levels of individuality. In most cases, both the males and females vocalise, often with discriminable sex differences. Information from individuality could be combined with other census methods involving calls for these species, for example many species of petrel are burrow-nesting and nocturnal, however, they possess extensive vocal repertoires [122]. These species also readily respond to playback of calls and this method is used to census populations and to determine burrow occupation [123-125]. In some species, primarily tropical and monogamous, duetting is common between pair-bond members, and calls become more similar between the male and female as the bond develops with increasing time (e.g. gibbon species [126]; red-fronted parrots, Poicephalus gulielmi [127]; twites, Acanthus flavirostris [128]. Although this may make the discrimination task more difficult, it may yield information on pair identity when the individuals are separated and the length of the pair bond.
In species with more complex social systems females are also vocally active, and often both sexes can be discriminated by individual and gender [129]. Female chacma baboons (Papio ursinus) have individually distinctive contact and alarm calls [130] and Weiss et al [131] showed that cotton-top tamarin (Saguinus oedipus) calls contain information on individual, sex and group identity. In several social canid species, females are as vocally active as males and take part in long-range vocalizing [132-136]. Vocalizations have been found to be individually distinctive for many of these species [78,132-134]. However, the rate of vocalizing in some canid species is related to social dominance, season, and whether they are transient or resident groups [135,136], and monitoring strategies would have to take account of this.
Vocal individuality as an effective conservation tool
Most of the literature suggesting that vocal individuality has a role in conservation is in fact represented by studies of the presence of vocal individuality and not whether it provides an effective conservation tool. To provide an indication of effectiveness as a conservation tool requires information on several topics and consideration of several issues. We consider the following to be most important.
1) The vocalization used should be easily recordable and provide the best potential for vocal individuality. It is generally considered that long-range advertising vocalizations most readily fulfil these criteria [65]. Species with a repertoire of such vocalizations may present the problem of which vocalization to use for individual identification. The problem can be addressed by either choosing one vocalization (such as the most common and/or distinctive) or by looking at features common to all vocalizations in the repertoire (e.g. [137,138]).
2) The sample size tested should be similar to the number of individuals that will be discriminated when the technique is used. In many cases vocal individuality studies have used smaller sample sizes (10–20 individuals) than if the technique were to be used as a conservation tool.
3) Careful consideration is needed in choosing which measures to take and every effort should be made to pare them down (using principle components analysis or stepwise discriminant analysis) to the most effective discriminators. The best measures will be those that can be extracted from recordings of varying quality. Some techniques standardize recording quality by only accepting those recordings that include specific sections of the signal, usually of lower amplitude, for analysis [53]. Atmospheric conditions affect temporal, amplitude and frequency components of a signal in different ways [139]. Different analysis types are also affected by recording quality, for example the instantaneous frequency of bittern booms can discriminate between individuals even in poor quality recordings [110]. However, if temporal measures are to be taken from a waveform display, recordings have to be of a higher quality because background noise will mask the signal.
4) When multivariate statistical tools are used to discriminate and identify individuals, we recommend the use of similarity techniques (see above) as they do not need the population size to be known and they also allow for the identification of additional individuals.
5) Any application of vocal individuality will not use recordings in isolation to identify individuals. A lot of data can be collected at the time of recording that will reduce the number of individuals that have to be discriminated. The person making the recording can note other simultaneously calling individuals and the locations and times of all recordings. This can be important if a technique seems to lack discrimination power.
6) The most effective test of vocal individuality will be through simulated census and monitoring situations [81,98]. These can be achieved through blind trials and repeated random sampling from known data sets to create population samples of unknown size and composition. We tested the use of neural network models in census and monitoring tasks by using a data set of 30 individuals that was randomly sampled [98]. The same individual could appear several times, and we used neural networks to classify individuals in a series of blind trials. These kinds of test more accurately simulate how the technique will perform as a conservation tool.
7) In most cases the people developing and testing monitoring techniques are not the same as those who will be using them in the field. Thus, one important, if obvious, point is that specific guidelines need to be given to those who will collect and analyse recordings. For many endangered or low-density populations, collecting recordings requires considerable fieldwork (e.g. [53]), and therefore this warrants the most efficient (effort vs. results) analysis possible. It may not be possible to rely on spontaneous vocalizing to collect enough recordings and playback is often used to elicit calls. However, the time of playback or its overuse can also have biasing effects, for example by causing individuals to move off their territories [37,38].
Vocal individuality in mammals
As demonstrated by the examples we have used in this review, applications of vocal individuality have been limited to avian species, usually vocally active species that are difficult to monitor with conventional methods. A notable exception is the study of swift foxes (Vulpes velox) by Darden et al [79]. Many other mammalian species are vocally activity on land and are difficult to monitor conventionally because they occur at low densities in dense habitat. In addition, many have a rich vocal repertoire, with both sexes vocalizing or with vocal communication between parents and offspring. However, the general structure of the mammalian vocal system causes not only complex signal structures (e.g. harmonics) but also modifications (e.g. biphonation and chaotic noise), making it more difficult to analyse (but see [79]). Perhaps more importantly, the signalling context can have large effects on the structure of the sound produced, which can affect its use in identification [140]. This seems to be a common phenomenon in primates. For example, the degree of vocal variability in chimpanzees is related to the amount of social chorusing between males, with the amount of time males spent together increasing the similarity of their calls [140,141]. Several marmoset species show individually distinctive vocalizations when either isolated [142] or in stable social groups, but in novel social conditions the vocal structures change [143]. Despite these potential problems, the number of reports of vocal individuality in mammals indicates that it could be a useful conservation tool (e.g. [144]).
For cetaceans, the underwater acoustic channel is the most important means of communication [145,146]. Studying individual vocal differences in cetaceans and other aquatic mammals has proved to be difficult due to problems in identifying which individual is calling, with the result that most studies have been at the group or population level [145]. More recently, various techniques have been used to locate individual callers either with theodolites, hydrophone arrays, or tags [146]. Also individuals have been identified over longer periods of time using photo-identification, mostly of dorsal fin [147] or tail-fluke features [148]. However, visual identification techniques are less readily applied when large groups of individuals are together [146]. Few studies have been made of individually distinctive features of their vocal behaviour, and none have used this individuality to follow individuals.
The most studied cetaceans are dolphins. Bottle-nose dolphins (Tursiops truncatus) have individually distinctive signature whistles [149-151], and these whistles can remain constant for over a decade [151,152] but they can change depending on the social context [153,154], and individuals are capable of copying the whistles of others [152] which then may be used in matched calling encounters [155]. Signature whistles would seem to be ideal for vocal individuality; however, social effects present several potential problems in their use. Also, the use of signature whistles may decrease when individuals are in groups [154]. It remains to be seen whether acoustic signals can be used to monitor dolphin species.
Large cetaceans are capable of communicating over large distances and are acoustically active, especially during the breeding season [145]. These acoustic signals have the ability to generate information on group identity, body size and interactions [146]. Studying cetaceans in stable social groups has revealed the more complex aspects of communication. Both sperm whales (Physeter macrocephalus) and killer whales (Orcinus orca) produce both individually and group distinctive signals [107,156,157]. The structure of sperm whale clicks also gives information about the size of the caller [158]. Few studies have used individuality to follow individuals (but see Fig. 10 in [152]).
Conclusion
There will be many instances where vocalizations are the only evidence of the presence of members of a population. Often, these populations are of conservation concern and baseline demographic information is difficult to obtain. Identifying individuals using individually distinctive vocalizations offers an alternative to marking that avoids problems associated with handling and sampling biases. As with other monitoring techniques, vocal individuality contains biases that have to be accounted for when it is used. Vocal individuality is not a panacea to all monitoring problems, but where a species is sensitive to disturbance or difficult to monitor (either through its behaviour or because of its environment), utilising its vocal behaviour can provide an effective conservation tool.
Authors' contributions
AMRT carried out the initial literature survey and drafted the first version of the manuscript. All authors expanded and developed the review. PKM produced the final version of the manuscript which was read and approved by all authors.
Acknowledgements
AMRT was funded by the Leverhulme Trust, TMP by the University of Copenhagen and PKM by Statens Naturvidenskablige Forskningsråd grant 21-01-0482. We thank the communication crew, Mark Bee, Vincent Bretagnolle and two anonymous referees for their helpful comments.
==== Refs
Baptista LF Gaunt SLL Clemmens JR, Buchholz R Bioacoustics as a tool in conservation studies Behavioural Approaches to Conservation in the Wild 1997 Cambridge: Cambridge University Press 209 211
McGregor PK Peake TM Gilbert G Sutherland WJ, Gosling M Communication behaviour and conservation Behaviour and Conservation 2000 Cambridge: Cambridge University Press 261 280
Curio E Conservation needs ethology Trends in Ecology and Evolution 1996 11 260 263
Emlen JT Dejong MD Counting birds: the problem of variable hearing abilities Journal of Field Ornithology 1992 63 26 31
McGregor PK Peake TM Caro TM The role of individual identification in conservation biology Behavioral Ecology and Conservation Biology 1998 Oxford; Oxford University Press 31 55
Falls JB Kroodsma DE, Miller EH Individual recognition by sound in birds Acoustic Communication in Birds 1982 2 New York: Academic Press 237 278
Stoddard PK Kroodsma DE, Miller EH Vocal recognition of neighbours by territorial passerines Ecology and Evolution of Acoustic Communication in Birds 1996 Ithaca: Cornell University Press 356 374
McGregor PK Signalling in territorial systems: a context for individual identification, ranging and eavesdropping Philosophical Transactions of the Royal Society of London Series B 1993 340 237 244
Suthers RA Variable asymmetry and resonance in the avian vocal tract: a structural basis for individually distinct vocalizations Journal of Comparative Physiology A 1994 175 457 466
Eakle WL Mannan RW Grubb TG Identification of individual breeding bald eagles by voice analysis Journal of Wildlife Management 1989 53 450 455
Gilbert G McGregor PK Tyler G Vocal individuality as a census tool, practical considerations illustrated by a study of two rare species Journal of Field Ornithology 1994 65 335 348
Puglisi L Adamo C Discrimination of individual voices in male great bitterns (Botaurus stellaris) in Italy The Auk 2004 121 541 527
Peake TM McGregor PK Smith KW Tyler G Gilbert G Green RE Individuality in Corncrake Crex crex vocalisations Ibis 1998 140 120 127
Beightol DR Samuel DE Sonagraphic analysis of the American woodcock's peent call Journal of Wildlife Management 1973 37 470 475
Saunders DA Wooller RD Consistent individuality of voice in birds as a management tool Emu 1988 88 25 32
Goss-Custard JD Caldow RWG Clarke RT LeVDit Durell SEA Urfi J West AD Consequences of habitat loss and change to populations of wintering migratory birds: predicting the local and global effects from studies of individuals Ibis 1995 137 S56 S66
Sutherland WJ From Individual Behaviour to Population Ecology 1996 Oxford: Oxford University Press
Goss-Custard JD Sutherland WJ Krebs JR, Davies NB Individual behaviour, populations and conservation Behavioural Ecology: An Evolutionary Approach 1997 Oxford: Blackwell 373 395
Stillman RA Goss-Custard JD West AD Le V Dit Durell SEA Caldow RWG McGrorty S Clarke RT Predicting mortality in novel environments: tests and sensitivity of a behaviour-based model Journal of Applied Ecology 2000 37 564 588
West AD Goss-Custard JD Stillman RA Caldow RWG Le V Dit Durell SEA McGrorty S Predicting the impacts of disturbance on shorebird mortality using a behaviour-based model Biological Conservation 2002 106 319 328
Reed JM Mills LS Dunning JB Menges ES McKelvey KS Frye R Beissinger SR Anstett MC Miller P Emerging issues in population viability analysis Conservation Biology 2002 16 7 19
Lande R Risks of population extinction from demographic and environmental stochasticity and random catastrophes American Naturalist 1993 142 911 927
Kendall BE Fox GA Variation among individuals and reduced demographic stochasticity Conservation Biology 2002 16 109 116
Fox GA Kendall BE Demographic stochasticity and the variance reduction effect Ecology 2002 83 1928 1934
Kruger O Lindstrom J Lifetime reproductive success in common buzzard, Buteo buteo : from individual variation to population demography Oikos 2001 93 260 273
Newton I The contribution of some recent research on birds to ecological understanding Journal of Animal Ecology 1995 64 675 696
Kelly MJ Lineage loss in Serengeti cheetahs: consequences of high reproductive variance and heritability of fitness on effective population size Conservation Biology 2001 15 137 147
Clutton-Brock TH Clutton-Brock TH Reproductive success Reproductive Success 1988 University of Chicago Press 472 485
Gompper ME Stacey PB Berger J Conservation implications of the natural loss of lineages in wild mammals and birds Conservation Biology 1997 11 857 867
Smith JNM Arcese P How fit are floaters? Consequences of alternative territorial behaviors in a nonmigratory sparrow American Naturalist 1989 133 830 845
Bayne EM Hobson KA Effects of habitat fragmentation on pairing success of ovenbirds: importance of male age and floater behaviour The Auk 2001 118 380 388
Kempenaers B Everding S Bishop C Boag P Robertson RJ Extra-pair paternity and the reproductive role of male floaters in the tree swallow (Tachycineta bicolor) Behavioural Ecology and Sociobiology 2001 49 251 259
Komdeur J Breeding the Seychelles magpie robin Copsychus sechellarum and implications for its conservation Ibis 1996 138 485 498
Bibby CJ Burgess ND Hill DA Mustoe SH Bird Census Techniques 2000 2 London: Academic Press
Greenwood JJD Sutherland WJ Basic Techniques Ecological Census Techniques 2000 Cambridge: Cambridge University Press 11 110
Conway CJ Eddleman WR Anderson SH Hanebury LR Seasonal changes in Yuma clapper rail vocalisation rate and habitat use Journal of Wildlife Management 1993 57 282 290
Legare ML Eddleman WR Buckley PA Kelly C The effectiveness of tape playback in estimating black rail density Journal of Wildlife Management 1999 63 116 125
Ogutu JO Dublin HT The response of lions and spotted hyaenas to sound playbacks as a technique for estimating population size African Journal of Ecology 1998 36 83 95
Eriksen T Booming bitterns Botaurus stellaris : ecology, behaviour and conservation implications MSc thesis 1999 Copenhagen University, Zoological Institute
Peake TM McGregor PK Corncrake Crex crex census estimates: a conservation application of vocal individuality Animal Biodiversity & Conservation 2001 24 81 90
Stowe TJ Hudson AV Corncrake studies in the western isles Royal Society for the Protection of Birds Conservation Review 1988 2 38 42
Moorhouse TP MacDonald DWJ Indirect negative impacts of radio-collaring: sex-ratio variation in water voles Journal of Applied Ecology 2005 42 91 98
de Villiers MS Meltzer DGA van Heerden J Mills MGL Richardson PRK van Jaarsveld AS Handling-induced stress and mortalities in African wild dogs (Lycaon pictus) Proceedings of the Royal Society of London Series B 1995 262 215 220 8524913
Woodruffe R Assessing the risks of intervention: immobilization, radio-collaring and vaccination of African wild dogs Oryx 2001 35 234 244
Castelli PM Trost RE Neck bands reduce survival of Canada geese in New Jersey Journal of Wildlife Management 1996 60 891 898
Menu S Hestbeck JB Gauthier G Reed A Effects of neck bands on survival of greater snow geese Journal of Wildlife Management 2000 64 544 552
Schmutz JA Morse JA Effects of neck collars and radiotransmitters on survival and reproduction of emperor geese Journal of Wildlife Management 2000 64 231 237
Alisauskas RT Lindberg MS Effects of neckbands on survival and fidelity of white-fronted and Canada geese captured as non-breeding adults Journal of Applied Statistics 2002 29 521 537
Nimon AJ Schroter RC Stonehouse B Heart-rate of disturbed penguins Nature 1995 374 415
Culik B Wilson R Penguins crowded out Nature 1991 351 340
Animal Behaviour Guidelines for the treatment of animals in behavioural research and teaching Animal Behaviour 2005 69 i vi
Caughley G Gunn A Conservation Biology in Theory and Practice 1996 Massachusetts: Blackwell Science
Gilbert G Tyler GA Smith KW Local annual survival of booming male great bittern Botaurus stellaris in Britain, in the period 1990–1999 Ibis 2002 144 51 61
Galeotti P Sacchi R Turnover of territorial scops owl Otus scops as estimated by spectrographic analysis of male hoots Journal of Avian Biology 2001 32 256 262
Rohner C The numerical response of great horned owls to the snowshoe hare cycle: consequences of non-territorial 'floaters' on demography Journal of Animal Ecology 1996 65 359 370
Hopp SL Owren MJ Evans CS eds Animal Acoustic Communication: Recent Technical Advances 1998 Berlin: Springer
Kroodsma DE Budney GF Grotke RW Vielliard JME Gaunt SLL Ranft R Veprintseva OD Kroodsma DE, Miller EH Natural sound archives: guidance for recordists and a request for cooperation Ecology and Evolution of Acoustic Communication in Birds 1996 Ithaca: Cornell University Press 474 486
Gerhardt HC Hopp SL, Owren MJ, Evans CS Acoustic signals of animals: recording, field measurements, analysis and description Animal Acoustic Communication: Recent Technical Advances 1998 Berlin: Springer 1 26
Charif RA Mitchell S Clark CW Canary 11 User's Manual 1993 Ithaca: Cornell University Press
Clements M Hopp SL, Owren MJ, Evans CS Digital signal acquisition and representation Animal Acoustic Communication: Recent Technical Advances 1998 Berlin: Springer 27 58
Stoddard PK Hopp SL, Owren MJ, Evans CS Application of filters in bioacoustics Animal Acoustic Communication: Recent Technical Advances 1998 Berlin: Springer 105 128
Beecher M Signalling systems for individual recognition: an information theory approach Animal Behaviour 1989 38 248 261
Gilbert G Vocal individuality as a census and monitoring tool: practical considerations illustrated by a study of the bittern Botaurus stellaris and the black-throated diver Gavia immer PhD thesis 1993 Nottingham University, Department of Life Science
Miller E Dring T Strong PIV, Merideth NH Territorial defense of multiple lakes by common loons: a preliminary report Papers from the 1987 Conference on Loon Research and Management 1988 North American Loon Fund 1 14
McGregor PK Byle P Individually distinct bittern booms: potential as a census tool Bioacoustics 1992 4 93 109
Janik VM Pitfalls in the categorization of behaviour: a comparison of dolphin whistle classification methods Animal Behaviour 1999 57 133 143 10053080
Jones AE ten Cate C Bijleveld CCJH The inter-observer reliability of scoring sonagrams by eye: a study on methods, illustrated on zebra finch songs Animal Behaviour 2001 62 791 801
Clapperton BK Jenkins PF Individuality in contact calls of the Pukeko (Aves: Rallidae) New Zealand Journal of Zoology 1987 14 19 28
Clark CW Marler P Beeman K Quantitative analysis of animal vocal phonology: an application to swamp sparrow song Ethology 1987 76 105 115
Williams JM Objective comparisons of song syllables – a dynamic programming approach Journal of Theoretical Biology 1993 161 317 328
Harper DGC Some comments on the repeatability of measurements Ringing and Migration 1994 15 84 90
Sparling DW Williams JD Multivariate analysis of avian vocalisations Journal of Theoretical Biology 1978 74 83 107 713569
Manly BFJ Multivariate Statistical Methods: A Primer 1986 London: Chapman & Hall
Bauer H-G Nagl W Individual distinctiveness and possible function of song parts of short-toed treecreepers (Certhia brachydactyla). Evidence from multivariate song analysis Ethology 1992 91 108 121
Galeotti P Pavan G Individual recognition of male tawny owls (Strix aluco) using spectrograms of their territorial calls Ethology, Ecology and Evolution 1991 3 113 126
Galeotti P Paladin M Pavan G Individually distinct hooting in male pygmy owls Glaucidium passerinum : a multivariate approach Ornis Scandinavica 1993 24 15 20
Delport W Kemp AC Ferguson JWH Vocal identification of individual African wood owls Strix woodfordii : a technique to monitor long-term adult turnover and residency Ibis 2002 144 30 39
Durbin LS Individuality in the whistle call of the Asiatic wild dog Cuon alpinus Bioacoustics 1998 9 197 206
Darden SK Dablesteen T Pedersen SB A potential tool for swift fox Vulpes velox conservation: individuality of long-range barking sequences Journal of Mammalogy 2003 84 1417 1427
Jorgensen DD French JA Individuality but not stability in marmoset long calls Ethology 1998 104 729 742
Terry AMR McGregor PK Peake TM A comparison of some techniques used to assess vocal individuality Bioacoustics 2001 11 169 188
Nelson DA Song overproduction and selective attrition lead to song sharing in the field sparrow (Spizella puslla) Behavioural Ecology and Sociobiology 1992 30 415 424
Rebbeck M Corrick R Eaglestone B Stainton C Recognition of individual European nightjars Caprimulgus europaeus from their song Ibis 2001 143 468 475
Khanna H Gaunt SLL McCallum DA Digital spectrographic cross-correlation: tests of sensitivity Bioacoustics 1997 7 209 234
Dahlquist FC Schemnitz SD Flachs BK Distinguishing individual male wild turkeys by analyzing vocalisations using a personal computer Bioacoustics 1990 2 303 316
Evans MR Evans JA A computer based technique for the quantitative analysis of animal sounds Bioacoustics 1994 5 281 290
Lessells CM Rowe CL McGregor PK Individual and sex differences in the provisioning calls of European bee-eaters Animal Behaviour 1995 49 244 247
McGregor PK Anderson CM Harris J Seal JR Soul JM Individual differences in songs of fantailed warblers Cisticola juncidis in Portugal Airo 1995 5 17 21
Latruffe C McGregor PK Tavares JP Mota PG Microgeographic variation in corn bunting (Miliaria calandra) song: quantitative and discrimination aspects Behaviour 2000 137 1241 1255
Nelson DA Marler P Palleroni A A comparative approach to vocal learning: intra-specific variation in the learning process Animal Behaviour 1995 50 83 97
Cortopassi KA Bradbury JW The comparison of harmonically rich sounds using spectrographic cross-correlation and principle coordinates analysis Bioacoustics 2000 11 89 127
Peake TM Variation in the vocal behaviour of the corncrake Crex crex : potential for conservation PhD thesis 1997 University of Nottingham, Department of Life Science
Buck JR Tyack PL A quantitative measure of similarity for Tursiops truncatus signature whistles Journal of the Acoustical Society of America 1993 94 2497 2506 8270729
Anderson SE Dave AS Margoliash D Template based automatic recognition of birdsong syllables from continuous recordings Journal of the Acoustical Society of America 1996 100 1209 1219 8759970
Kogan JA Margoliash D Automated recognition of bird song elements from continuous recordings using dynamic time warping and hidden Markov models: a comparative study Journal of the Acoustical Society America 1998 103 2185 2197
Reby D Lek S Dimopoulos I Joachim J Lauga J Aulagnier S Artificial neural networks as a classification method in the behavioural sciences Behavioural Processes 1997 40 35 43
Campbell GS Gisiner RC Helweg DA Milette LL Acoustic identification of female Steller sea lions (Eumetopias jubatus) Journal of the Acoustical Society of America 2002 111 2920 2928 12083225
Terry AMR McGregor PK Census and monitoring based on individually identifiable vocalizations: the role of neural networks Animal Conservation 2002 5 103 111
Hinton GE How neural networks learn from experience Scientific American 1992 105 109
Gurney K An Introduction to Neural Networks 1997 London: UCL Press
Bishop CM Neural Networks for Pattern Recognition 1995 Oxford: Clarendon Press, Oxford
Kohonen T Self-organising Maps 1997 Berlin: Springer
Phelps SM Ryan MJ Neural networks predict response biases of female tungara frogs Proceedings of the Royal Society of London Series B 1998 265 279 285 9523430
Phelps SM Ryan MJ History influences signal recognition: neural network models of tungara frogs Proceedings of the Royal Society of London Series B 2000 267 1633 1639 11467426
Reby D Joachim J Lauga J Lek S Aulagnier S Individuality in the groans of fallow deer (Dama dama) bucks Journal of Zoology London 1998 245 79 84
Placer J Slobodchikoff CN A fuzzy-neural system for identification of species-specific alarm calls of Gunnison's prairie dogs Behavioural Processes 2000 52 1 9 11011104
Deecke VB Ford JKB Spong P Quantifying complex patterns of bioacoustic variation: use of a neural network to compare killer whale (Orcinus orca) dialects Journal of the Acoustical Society of America 1999 105 2499 2507 10212431
Gibbons DW Avery M Baillie S Gregory R Kirby J Porter R Tucker G Williams G Bird species of conservation concern in the United Kingdom, Channel Islands and Isle of Man: revising the Red data list RSPB Conservation Review 1996 10 7 18
Gilbert G Gibbons DW Evans J Bird Monitoring Methods A Manual of Techniques for Key UK Species 1998 Sandy: RSPB
Terry AMR Bioacoustics in conservation: new approaches using vocal individuality PhD thesis 2002 Copenhagen University, Zoological Institute
Gregory RD Noble DG Cranswick PA Campbell LH Rehfisch MM Baillie SR The state of the UK's birds 2000 2001 Sandy: RSPB, BTO and WWT
May L Individually distinctive corncrake Crex crex calls: a pilot study Bioacoustics 1994 6 25 32
Freeman PL Identification of individual barred owls using spectrogram analysis and auditory cues Journal of Raptor Research 2000 34 85 92
Otter K Individual variation in the advertising call of male Northern saw-whet owls Journal of Field Ornithology 1996 67 398 405
Lengagne T Temporal stability in the individual features in the calls of eagle owls (Bubo bubo) Behaviour 2001 138 1407 1419
Cavanagh PM Ritchison G Variation in the bounce and whinny songs of the Eastern screech-owl Wilson Bulletin 1987 99 620 627
Hill FAR Lill A Vocalisations of the Christmas Island hawk-owl Ninox natalis : individual variation in advertisement calls Emu 1998 98 221 226
Appleby BM Redpath SM Variation in the male territorial hoot of the tawny owl Strix aluca in three English populations Ibis 1997 139 152 158
Burnell K Rothstein SJ Variation in the structure of female brown-headed cowbird vocalizations and its relation to vocal function and development Condor 1994 96 703 715
Speirs EAH Davis LS Discrimination by Adelie penguins, Pygoscelis adelie, between the loud mutual calls of mates, neighbours and strangers Animal Behaviour 1991 41 937 944
Aubin T Jouventin P How to vocally identify kin in a crowd: the penguin model Advances in the Study of Behavior 2002 31 243 277
Warham J The Behaviour, Population Biology and Physiology of the Petrels 1996 San Diego: Academic Press
Bretagnolle V Calls of Wilson's storm petrel: functions, individual and sexual recognitions and geographic variation Behaviour 1989 111 98 l12
Bretagnolle V Robinson P Species-specific recognition in birds: an experimental investigation in the Wilson's storm petrel (Procellariiformes, Hydrobatidae) using digitalized signals Canadian Journal of Zoology 1991 69 1669 1673
Berrow SD The use of acoustics to monitor burrow-nesting white-chinned petrels Procellaria aequinoctialis at Bird Island, South Georgia Polar Biology 2000 23 575 579
Geissmann T Duet-splitting and the evolution of gibbon songs Biological Reviews 2002 77 57 76 11911374
Venuto VMRBL African parrot vocalizations and their functional significance Ostrich 2001 224 228
Marler P Mundinger PC Vocalizations, social-organization and breeding biology of twite Acanthus flavirostris Ibis 1975 117 1 6
Benz JJ French JA Leger DW Sex differences in vocal structure in a callitrichid primate Leontopithecus rosalia American Journal of Primatology 1990 21 257 264
Fischer J Hammerschmidt K Cheney DL Seyfarth RM Acoustic features of female chacma baboon barks Ethology 2001 107 33 54
Weiss DJ Garibaldi BT Hauser MD The production and perception of long calls by cotton-top tamarins (Saguinus oedipus): acoustic analyses and playback experiments Journal of Comparative Psychology 2001 115 258 271 11594495
Fromolt KH Kruchenkova EP Russig H Individuality of territorial barking in Arctic foxes, Alopex lagopus Zeitschrift fur Saugetierkunde 1997 62 66 70
Tooze ZJ Harrington FH Fentress JC Individually distinct vocalisations in timber wolves, Canis lupus Animal Behaviour 1990 40 723 730
East ML Hofer H Loud calling in a female-dominated mammalian society: I. structure and composition of whooping bouts of spotted hyenas, Crocuta crocuta Animal Behaviour 1991 42 637 649
East ML Hofer H Loud calling in a female-dominated mammalian society: II. behavioural contexts and functions of whooping of spotted hyenas, Crocuta crocuta Animal Behaviour 1991 42 651 669
Gese EM Ruff RL Howling by coyotes (Canis latrans): variation among social classes, seasons and pack sizes Canadian Journal of Zoology 1998 76 1037 1043
Falls JB Krebs JR McGregor PK Song matching in the great tit (Parus major): the effects of similarity and familiarity Animal Behaviour 1982 30 997 1009
Weary DW Krebs JR Great tits classify songs by individual voice characteristics Animal Behaviour 1992 43 283 287
Wiley RH Richards DG Kroodsma DE, Miller EH Adaptations for acoustic communication in birds: sound transmission and signal detection Acoustic Communication in Birds 1982 2 New York: Academic Press 132 164
Mitani JC Brandt KL Social factors influence the acoustic variability in the long-distance calls of male chimpanzees Ethology 1994 96 233 252
Mitani JC Gros JL Chorusing and call convergence in chimpanzees: tests of three hypothesis Behaviour 1998 135 1041 1064
Jones BS Harris DHR Catchpole CK The stability of the vocal signature in Phee calls of the common marmoset, Callithrix jacchus American Journal of Primatology 1993 31 67 75
Elowson AM Snowdon CT Pygmy marmosets, Cebuella pygmaea, modify vocal structure in response to changed social environment Animal Behaviour 1994 47 1267 1277
Hartig S Individual acoustic identification as a non-invasive conservation tool: an approach to the conservation of the African wild dog Lycaon pictus Bioacoustics 2005 15 35 50
Tyack PL Mann J, Connor R, Tyack P, Whitehead H Functional aspects of cetacean communication Cetacean Societies: Field Studies of Dolphins and Whales 2000 Chicago: University of Chicago Press 270 307
Whitehead H Christal J Tyack PL Mann J, Connor R, Tyack P, Whitehead H Studying cetacean social structure in space and time: innovative techniques Cetacean Societies: Field Studies of Dolphins and Whales 2000 Chicago: University of Chicago Press 65 90
Ford JKB Vocal traditions among resident killer whales (Orcinus orca) in coastal waters of British Columbia (Canada) Canadian Journal of Zoology 1991 69 1454 1483
Dufault S Whitehead H An assessment of changes with time in the marking patterns used for photo-identification of individual sperm whales, Physeter macrocephalus Marine Mammal Science 1995 11 335 343
Caldwell MC Caldwell DK Vocalization of naive captive dolphins in small groups Science 1968 159 1121 1123 5636351
Caldwell MC Caldwell DK Tyack PL Leatherwood S, Reeves RR Review of the signature-whistle hypothesis for the Atlantic bottlenose dolphin The Bottlenose Dolphin 1990 San Diego: Academic Press 199 234
Sayigh LS Tyack PL Wells RS Scott MD Signature whistles of free-ranging bottlenose dolphins Tursiops truncatus – stability and mother offspring comparisons Behavioral Ecology and Sociobiology 1990 26 247 260
Tyack P Hopp SL, Owren MJ, Evans CS Acoustic communication under the sea Animal Acoustic Communication: Recent Technical Advances 1998 Berlin: Springer 163 220
Janik VM Dehnhardt G Todt D Signature whistle variation in a bottlenosed dolphin, Tursiops truncatus Behavioral Ecology and Sociobiology 1994 35 243 248
Janik VM Slater PJB Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls Animal Behaviour 1998 56 829 838 9790693
Janik VM Whistle matching in wild bottlenose dolphins (Tursiops truncatus) Science 2000 289 1355 10958783
Weilgart LS Whitehead H Distinctive vocalizations from mature mature sperm whales (Physeter macrocephalus) Canadian Journal of Zoology 1988 66 1931 1937
Ford JKB Acoustic behavior of resident killer whales (Orcinus orca) off Vancouver Island, British Columbia Canadian Journal of Zoology 1989 67 727 745
Goold JC Signal processing techniques for acoustic measurement of sperm whale body lengths Journal of the Acoustical Society of America 1996 100 3431 3441 8914321
|
15960848
|
PMC1183234
|
CC BY
|
2021-01-04 16:38:34
|
no
|
Front Zool. 2005 Jun 16; 2:10
|
utf-8
|
Front Zool
| 2,005 |
10.1186/1742-9994-2-10
|
oa_comm
|
==== Front
Global HealthGlobalization and Health1744-8603BioMed Central London 1744-8603-1-111604580110.1186/1744-8603-1-11DebateCan context justify an ethical double standard for clinical research in developing countries? Landes Megan [email protected] London School of Hygiene and Tropical Medicine, 1 Keppel Street, London, WC1E 7HT, UK2005 26 7 2005 1 11 11 18 2 2005 26 7 2005 Copyright © 2005 Landes; licensee BioMed Central Ltd.2005Landes; 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 design of clinical research deserves special caution so as to safeguard the rights of participating individuals. While the international community has agreed on ethical standards for the design of research, these frameworks still remain open to interpretation, revision and debate. Recently a breach in the consensus of how to apply these ethical standards to research in developing countries has occurred, notably beginning with the 1994 placebo-controlled trials to reduce maternal to child transmission of HIV-1 in Africa, Asia and the Caribbean. The design of these trials sparked intense debate with the inclusion of a placebo-control group despite the existence of a 'gold standard' and trial supporters grounded their justifications of the trial design on the context of scarcity in resource-poor settings.
Discussion
These 'contextual' apologetics are arguably an ethical loophole inherent in current bioethical methodology. However, this convenient appropriation of 'contextual' analysis simply fails to acknowledge the underpinnings of feminist ethical analysis upon which it must stand. A more rigorous analysis of the political, social, and economic structures pertaining to the global context of developing countries reveals that the bioethical principles of beneficence and justice fail to be met in this trial design.
Conclusion
Within this broader, and theoretically necessary, understanding of context, it becomes impossible to justify an ethical double standard for research in developing countries.
==== Body
Introduction
The design of clinical research trials deserves special caution, for such research is always at risk of crossing the fine line between regard for individual rights and potential exploitation of research subjects. Infamous experiments like the Tuskagee Syphilis Study have rendered evident the dangers for individuals when we cross that line. To safeguard human subjects, the international community has agreed on standard ethical principles, particularly the World Medical Association's Declaration of Helsinki; while it is encouraging that these frameworks exist, they remain open to interpretation, revision, and debate. With the 1994 placebo-controlled trials to reduce maternal to child transmission (MTCT) of HIV-1 initiated in Africa, Asia, and the Caribbean, we saw a breach in our consensus concerning the application of these principles, namely how to apply ethical standards to research conducted in the context of resource-poor settings.
In fact, the 'contextual' apologetics for this breach are inherent, I will argue, in current bioethical methodology. As an application of ethical theory, bioethics pays particular attention to context by acknowledging the unique influence of relationships and the immediate environment on an individual's experience. In terms of developing countries, bioethics has grounded its contextual analysis on the discourse of scarcity and sacrifice [1]. In particular, the use of placebo-controlled trials in developing countries has been justified by the contextual considerations of scarcity – trials that would otherwise be deemed unethical in developed countries. This convenient appropriation of 'contextual' analysis simply fails to acknowledge the underpinnings of feminist ethical analysis upon which it must stand. A more rigorous consideration of the political, social, and economic structures pertaining to individual contexts must be sought, and we must ensure that our international ethical standards are scrutinized and applied at the appropriate level. Within this broader, and theoretically necessary, understanding of context, it becomes impossible to justify an ethical double standard for research in developing countries.
Background: The debate over placebo-controlled trials in developing countries
The debate over the application of research ethics in developing countries surfaced with the early prevention of MTCT of HIV trials. While in 1994 there was an existing protocol from the AIDS Clinical Trial Group 076 (ACTG 076) for preventing MTCT, high antiretroviral (ARV) costs and insufficient infrastructure placed the regimen out of reach for the majority of the HIV-infected population in the developing world. To find a more cost-effective and applicable treatment for resource-poor settings, randomized placebo-controlled trials were initiated to investigate a short-course ARV regimen. However, these studies sparked intense debate as they clearly violated the condition of equipoise: that placebo groups are only deemed ethical if there exists sufficient uncertainty regarding the merit of the intervention. In other words, if there exists no gold standard of care then placebos can be justified. The arguments for not providing the 'gold standard' available in developed countries were founded on the existing low 'standard of care' in the context of developing countries. The NIH and CDC, both principal funding organizations of the studies, defended the studies' design: "it is an unfortunate fact that the current standard of perinatal care for the HIV-infected pregnant women in the sites of the studies does not include any HIV prophylactic intervention at all" and that placebo controls "will be the most reliable answer to the question of the value of the study compared to the local standard of care [2]." In opposition, Paul Lurie, wrote to the United States Department of Human Rights and Services: "unless you act now as many as 1002 newborn infants will die of unnecessary HIV infections they will contract from...HIV-infected mothers in nine unethical research experiments funded by your Department [3]." Despite this early opposition, the trials began in 16 countries and included over 12 000 HIV-infected women [4].
Redefining the 'context' of developing countries
While traditional ethical theory seeks fundamental principles to guide our actions, much of the current bioethical literature rejects claims to the effect that morality can be reduced to a set of universal principles [5]. They argue that the agent of moral decision (a patient, family member or physician) is inextricably embedded in a complex web of relationships. By accounting for the uniqueness of each ethical situation, bioethics attempts to apply ethical theory or principles relevantly to the context at hand. The problem then becomes how to define context, for much of our practical application of bioethics hinges on this point. A specifically feminist bioethics also rejects the universal claims of traditional ethical theory, but goes further to place value on the political, economic, and social factors that differentiate individual situations. In doing so, it pays particular attention to power differentials that exist between men and women, rich and poor, developed and developing countries [6]. It is this broader scope of context that must be applied to research standards in the debate over the short-course ARV placebo-controlled trials.
Applying the principles of bioethics at an appropriate level of analysis
Since the Belmont Report defined the four principles of bioethics, namely the principles of non-maleficence, beneficence, justice and autonomy, they have been used to ensure that ethical standards are applied to research. To begin, the principle of non-maleficence states that research must cause no harm to subjects and the principle of beneficence states that due to their participation in research, all possible benefits to subjects should be maximized. This immediately raises some questions for the case at hand. While it can be argued that no outright harm was afforded to the mother-infant pairs in the placebo group since their access to ARVs was no different than it would have been within their country context, it raises the question of to whose standard of care was the trial responsible? Moreover, can we further justify using this low standard of care within a resource-rich, developed world led research trial, thus violating the principle of beneficence? What is missing is an acknowledgement of the interlocking political, social, and economic contextual factors of these trials and an examination of what 'standard of care' ought to mean. Ultimately, the question becomes whether the contextual argument is enough to justify violating the principle of beneficence.
To answer this question, we must first make a distinction between the accessibility of AZT within the developing country as opposed to accessibility in a clinical trial. Supporters of the placebo-control design did not argue that it was financially or logistically unfeasible to provide the gold standard ACTG076 regimen for the control group, but rather that it was unnecessary because of the low standard of care which existed in the developing countries. Lurie and Wolfe argue that this contextualized 'standard of care' justifies withholding a readily available treatment [7]. An individual's right to receive maximum benefits from participating in research is thus determined by their individual socio-economic status, and in turn, their access to health care.
The tenuous nature of this contextualized justification becomes clear when we take a broader view of the factors determining an individual's access to health care. First, we must acknowledge that a person's scope of choice is often determined by forces beyond her own control. As Amartya Sen describes, not only do those in developing countries face economic deprivations, they are in turn subject to substantial 'unfreedoms'. These 'unfreedoms' include lack of employment (or freedom to participate in the market) and lack of access to health care (or freedom to ward off early mortality) [8]. In particular, access to health care is often not determined entirely by individual choice, but rather by the wealth of a country, its commitment to population health, and the distribution of its resources. When one includes this understanding of the economic context that dictates individual access to health care, context itself seems like a grossly unfair justification for violating the principle of beneficence. Here, the 'standard of care' argument only serves to exploit the individual, and take advantage of her circumstance and poverty.
Furthermore, we can expand the scope of context to include both an evaluation of a nation's internal health care priorities and overriding global economic inequities. Recognizing the role of developing nations in the global economy, we see that it is not entirely by choice that developing countries provide a standard level of care that does not include adequate MTCT prophylaxis. Developing countries are in fact given very little option under the continuing reverberations of the 1980s debt crisis. After heavily borrowing from the International Monetary Fund and the World Bank, countries have faced 'forced' economic reform through the structural adjustment policies of these lending institutions such as the devaluation of currency and enforced transition into export-based economies. This economic re-structuring has either required or caused a significant erosion of social service infrastructure, including health care and has particularly impacted many of the most vulnerable populations [9]. By dismissing the international forces that define the range of economic options available to developing countries, we isolate the problems of the developing world and allow ourselves to ignore our relative role and responsibility. Thus, within this broader scope of context, developed nations leading the research in question should recognize their interconnected role in the determinants of global inequity and should caution their support for the 'standard of care' argument which only serves to reinforce such inequity.
When expanding the parameters of context, the use of placebo-controlled trials in developing countries also fails to meet the principle of justice. This principle ensures that those who bear the burden of research risk will ultimately receive the benefits of the research [10]. In other words, specific populations should never be targeted due to their availability or compromised position. In the early debate over the short-course AZT trials, Satcher and Varmus argued that this principle was being satisfied since the trials were specifically investigating cost effective treatments for the HIV epidemic that was disproportionately affecting the world's poorest nations [11]. This argument hinges on a definition of context that isolates developing countries as independently dealing with an overwhelming epidemic and obscures our ability to apply the principle of justice at an appropriate level of analysis.
To assess the principle of justice within a broader understanding of the HIV epidemic, we must determine whether the research subjects would indeed receive adequate benefits for their participation. At the beginning of the debate, Annas and Grodin challenged the notion that an affordable treatment would ever be operational given the exceedingly low health care resources available to developing countries [12]. While the feasibility of implementation may have been questionable in 1998, there has been a global commitment to the prevention of MTCT and some notable successes in reducing the rate of MTCT with short course ARV regimens. Unfortunately, programmes designed for resource poor areas continue to fall short of the overwhelming success seen in developed countries where studies with highly active antiretroviral therapy demonstrate transmission rates of 1% in a non-breastfeeding population [13]. In contrast, the World Health Organization estimates that in developing countries, only 1% to 35% of populations have access to preventative care, with the lowest coverage in countries most significantly affected by the HIV epidemic [14]. Also, in 2002, it was estimated that 800 000 children acquired HIV infections, the vast majority of which were in the developing world [15]. While resource-rich populations are virtually beating vertical transmission, developing nations continue to struggle.
In terms of weighing the research subject's contributions to the trials against their benefits, which ten years later seems partial at best, an expanded context forces us to look at all the potential beneficiaries of this research. It has been argued that advances in more effective and cheaper methods of preventing vertical transmission are as likely to be implemented in developed countries as in developing countries [16]. The fact that these short-course regimens have not become standard in the developed world is due to their sub-standard reduction of transmission in comparison to more complex ARV regimens [17]. However, knowledge generated by these short-course trials has been applied to other research. The promising effects of short-course nevirapine trials in developing countries prompted researchers to study whether adding nevirapine to the more complex gold standard ARV regimen would further reduce transmission rates in the developed world [18]. This demonstrates that knowledge does not stay within segregated developing-developed contexts. Isolating the HIV epidemic as a developing world problem does not adequately acknowledge the global scope of the disease nor the subsequent global benefit of advances in treatment. Given the limited benefits to date of MTCT programs and this larger web of beneficiaries, it seems that while perhaps unintentional, the structure of current research serves to exploit impoverished populations for the benefit of science and more developed nations. In this context, the argument for the placebo-control trials violates the principle of justice.
Where does the debate stand now?
The 'standard of care' debate has continued since the MTCT prevention trials and has prompted the inclusion of paragraph 29 in the Declaration of Helsinki:
the benefits, risks, burdens, and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic method. This does not exclude the use of placebo, or no treatment, in studies where no proven prophylactic diagnostic or therapeutic method exists [19].
This attempt to create more stringent standards for the use of placebo-controlled trials regardless of the contextual standard of care, has been attacked by the international community as "out of touch with contemporary thinking" [20] and overly constitutional [21]. In response to these changes, there has been a flurry of independent international organizations writing their own, opposing ethical standards. Many, such as the Nuffield Council on Bioethics and the Council for International Organizations of Medical Sciences, have come to the conclusion that there should not be an absolute ruling on this matter, but that study design can be qualified when all other standards are satisfied and a sound scientific reason can be given for using a placebo-control arm [22]. This is treading on dangerous territory as the demands of science may outweigh ethical safeguards for individuals. It seems the debate has landed us back at the beginning with the Nuffield Council's new guidelines:
Wherever appropriate, participants in the control group should be offered a universal standard of care for the disease being studied. Where it is not appropriate to offer a universal standard of care, the minimum standard of care that should be offered to the control group is the best intervention available for that disease as part of the national public health system [23].
What determines the appropriateness of offering a universal standard of care may be scientific criteria or, as Schuklenk argues, may be erroneously conflated with economic criteria such as the low standard of care available in resource-poor settings [24]. This dangerous return to justifying a double standard for research in developing countries shows us that instead of building a mature consensus around the application of research ethics to developing countries, the discourse has become even further ensconced in the isolating and narrow context of the developing world.
Conclusion
The arguments put forward to understand the ethical dilemma created by the short-course ARV trials for the prevention of MTCT of HIV should not be interpreted to mean that research should never be done in developing countries. Rather, developed nations need to honestly assess their role in such research, take responsibility for their actions, and abstain from the exploitation of ethical loopholes as provided by the contextual nature of bioethics. It is essential that we consider context in our ethical deliberations, but we must be critical of our definition of context. It would be tragic if we allowed ethical principles to be manipulated for the exploitation of vulnerable populations, the psychological comfort of the true beneficiaries, and the effacement of real differences between individuals and populations. To this end, we must always remember that the inclusion of context is a corrective to traditional ethics, not an invitation to exploitation. As demonstrated above, the framework of the short-course ARV trials is fundamentally challenged when context is taken seriously. However, the current global situation does not engender optimism that this exploitative research constitutes an isolated incident. We must not shirk our own recognized ethical responsibilities – at the heart of research design we must situate the proper articulation of context, towards which I submit the above as a first step.
Abbreviations
AIDS Acquired Immune Deficiency Syndrome
ARV Antiretroviral
AZT Zidovudine
CDC Center for Disease Control
HIV Human Immunodeficiency Virus
NIH National Institutes of Health
MTCT Mother to child transmission
==== Refs
Olweny C Bioethics in developing countries: ethics of scarcity and sacrifice Journal of Medical Ethics 1994 20 169 174 7996563
Angell M The Ethics of Clinical Research in the Third World New England Journal of Medicine 1997 337 847 849 9295243 10.1056/NEJM199709183371209
Bayer R The Debate over Maternal-Fetal HIV Transmission Prevention Trials in Africa, Asia, and the Caribbean: Racist Exploitation or Exploitation of Racism? American Journal of Public Health 1998 88 567 570 9550995
Levine C Placebos and HIV: Lessons Learned Hastings Centre Report 1998 28 43 48
Waluchow W Thomas J Waluchow W, Thomas J Ethical Frameworks for Decision-Making Well and Good: Case Studies in Biomedical Ethics 1998 USA: Broadview Press
Sherwin S No Longer Patient 1992 USA: Temple University Press
Lurie P Wolfe S Unethical Trials of Interventions to Reduce Perinatal Transmission of the Human Immunodeficiency Virus in Developing Countries New England Journal of Medicine 1997 337 853 856 9295246 10.1056/NEJM199709183371212
Sen A Development as Freedom 1999 New York: Anchor Books
Sparr P (ed) Mortgaging Women's Lives 1994 London: Zed Books
National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research The Belmont Report 1978 Washington, DC
Satcher D Varmus H Ethical Complexities of Conducting Research in Developing Countries New England Journal of Medicine 1998 337 1003 1005 9309109
Annas G Grodin M Human Rights and Maternal-Fetal HIV Transmission Prevention Trials in Africa American Journal of Public Health 1998 88 560 563 9550993
Mofenson L Tale of two epidemics: the continuing challenge of preventing mother-to-child transmission of human immunodeficiency virus J Infect Dis 2003 187 721 724 12599044 10.1086/367905
UNAIDS Progress Report on the Global Response to the HIV/AIDS Epidemic 2003 Geneva
UNAIDS Progress Report on the Global Response to the HIV/AIDS Epidemic 2003 Geneva
Annas G Grodin M Human Rights and Maternal-Fetal HIV Transmission Prevention Trials in Africa American Journal of Public Health 1998 88 560 563 9550993
Jackson JB Musoke P Fleming T Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: 18 month followup to the HIVNET012 randomized trial Lancet 2003 362 59 68
Dorenbaum A Cunnigham CK Gelber RD Culnane M Mofenson L Briotto P Rekacewicz C Newell ML Delfraissy JF Cunnigham-Schrader B Mirochnick M Sullivan JL Two-dose intrapartum/newborn nevirapine and standard antiretroviral therapy to reduce perinatal HIV transmission: a randomized trial JAMA 2002 288 189 98 12095383 10.1001/jama.288.2.189
Wendler D Emanuel EJ Lie RK The Standard of Care Debate: Can Research in Developing Countries Be Both Ethical and Responsive to Those Countries's Health Needs? American Journal of Public Health 2004 94 923 928 15249290
Tangwa GB Between universalism and relativism: a conceptual exploration of problems in formulating and applying international biomedical ethical guidelines J Med Ethics 2004 30 63 67 14872078 10.1136/jme.2003.003194
Lie RK Emanuel E Grady C Wendler The standard of care debate: the Declaration of Helsinki versus the international consensus opinion J Med Ethics 2004 30 190 193 15082816 10.1136/jme.2003.006031
McMillan JR Conlon C The ethics of research related to health care in developing countries J Med Ethics 2004 30 204 206 15082819 10.1136/jme.2002.001263
McMillan JR Conlon C The ethics of research related to health care in developing countries J Med Ethics 2004 30 204 206 15082819 10.1136/jme.2002.001263
Schuklenk U The standard of care debate: against the myth of an "international consensus opinion." J Med Ethics 2004 30 194 197 15082817 10.1136/jme.2003.006981
|
16045801
|
PMC1183235
|
CC BY
|
2021-01-04 16:39:02
|
no
|
Global Health. 2005 Jul 26; 1:11
|
utf-8
|
Global Health
| 2,005 |
10.1186/1744-8603-1-11
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-451604865010.1186/1477-7525-3-45ResearchWomen's quality of life is decreased by acute cystitis and antibiotic adverse effects associated with treatment Ernst Erika J [email protected] Michael E [email protected] James D [email protected] George R [email protected] College of Pharmacy, University of Iowa, Iowa City, IA, USA2 Department of Family Medicine, University of Iowa Health Care, Iowa City, IA, USA3 Carver College of Medicine, University of Iowa, Iowa City, IA, USA4 Northeast Iowa Medical Education Foundation, Waterloo, IA, USA2005 27 7 2005 3 45 45 28 4 2005 27 7 2005 Copyright © 2005 Ernst et al; licensee BioMed Central Ltd.2005Ernst 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 acute cystitis is a common infection in women, the impact of this infection and its treatment on women's quality of life (QOL) has not been previously described.
Objectives: To evaluate QOL in women treated for acute cystitis, and describe the relationship between QOL, clinical outcome and adverse events of each of the interventions used in the study.
Methods
Design. Randomized, open-label, multicenter, treatment study.
Setting. Two family medicine outpatient clinics in Iowa.
Patients. One-hundred-fifty-seven women with clinical signs and symptoms of acute uncomplicated cystitis.
Intervention. Fifty-two patients received trimethoprim/sulfamethoxazole 1 double-strength tablet twice daily for 3 days, 54 patients received ciprofloxacin 250 mg twice daily for 3 days and 51 patients received nitrofurantoin 100 mg twice daily for 7 days.
Measurements. QOL was assessed at the time of enrollment and at 3, 7, 14 and 28 days after the initial visit. QOL was measured using a modified Quality of Well-Being scale, a validated, multi-attribute health scale. Clinical outcome was assessed by telephone interview on days 3, 7, 14 and 28 using a standardized questionnaire to assess resolution of symptoms, compliance with the prescribed regimen, and occurrence of adverse events.
Results
Patients experiencing a clinical cure had significantly better QOL at days 3 (p = 0.03), 7 (p < 0.001), and 14 (p = 0.02) compared to patients who failed treatment. While there was no difference in QOL by treatment assignment, patients experiencing an adverse event had lower QOL throughout the study period. Patients treated with ciprofloxacin appeared to experience adverse events at a higher rate (62%) compared to those treated with TMP/SMX (45%) and nitrofurantoin (49%), however the difference was not statistically significant (p = 0.2).
Conclusion
Patients experiencing cystitis have an increase in their QOL with treatment. Those experiencing clinical cure have greater improvement in QOL compared to patients fail therapy. While QOL is improved by treatment, those reporting adverse events have lower overall QOL compared to those who do not experience adverse events. This study is important in that it suggests that both cystitis and antibiotic treatment can affect QOL in a measurable way.
==== Body
Background
Acute uncomplicated cystitis is a common community acquired infection in women, causing an estimated 7 million episodes yearly in the United States [1]. On average, each episode is associated with 1.2 days of missed work or classes [1]. The cost of treating ambulatory patients with symptoms of dysuria is estimated to exceed 1 billion dollars per year in the United States [1].
As with the majority of infectious diseases, empiric therapy is the cornerstone of treatment for acute uncomplicated cystitis. When obtained, urine cultures often take several days for results, making it necessary to begin therapy empirically. Selection of the antibiotic must take into consideration many factors, including results of clinical trials demonstrating efficacy of a chosen regimen, known susceptibility patterns of commonly infecting organisms, the clinical condition of the patient, as well as drug allergies, rate of adverse reactions and cost of treatment. In many settings, particularly health maintenance organizations, treatment algorithms exist for the empiric treatment of acute cystitis [2,3]. In most cases, these guidelines suggest treatment based upon symptoms of disease in otherwise healthy adult women and recommend against obtaining culture and sensitivity tests [2]. At least one study has shown this treatment strategy to be the most cost-effective means of therapy [4].
In order to fully compare potential antimicrobial regimens, factors such as clinical cure rates, complications, including adverse events of treatment, and quality of life (QOL) should be taken into account. While there are several tools available in the literature to evaluate QOL, the majority are applied to patients with chronic diseases such as rheumatoid arthritis or with cancer [5]. One pilot study has been published evaluating QOL in acute cystitis and found that it is significantly affected by the illness [6]. However, the impact of treatment on QOL has not been evaluated in the previously published study.
Additionally, the evaluation of adverse effects related to different treatments and the impact on QOL have not been described previously for cystitis. Most of the published literature comparing the impact of treatment on a patients' QOL is limited to chronic diseases or treatments with severe adverse effects such as cancer chemotherapy [7-10]. We sought to determine the impact of a common infection, cystitis, on QOL and evaluate the impact of three different treatment regimens on QOL in these patients. The addition of QOL information could be useful to primary care physicians when deciding on an antibiotic for empiric treatment of cystitis.
Methods
Patients were recruited from two outpatient family medicine clinics in the state of Iowa, both affiliated with the University of Iowa College of Medicine and College of Pharmacy. Women aged 18 to 65 were eligible for inclusion in the study if they had clinical symptoms of cystitis, including dysuria or frequency, and were considered by their primary physician to have acute uncomplicated cystitis and were not allergic to any study medication. Patients were excluded if they had diabetes, known anatomical abnormalities, urinary symptoms for greater than 7 days, vaginal discharge, or were pregnant. These exclusion criteria were intended to select for acute cystitis patients and exclude patients who may have pyelonephritis or sexually transmitted diseases. All study procedures were approved by the Institutional Review Board at the University of Iowa.
After obtaining written informed consent, patients provided a urine specimen for culture and sensitivity. Patients were then randomized, using a random number generator, to one of three treatments; a) Trimethoprim/Sulfamethoxazole (TMP/SMX) 800 mg/160 mg (Qualitest Pharmaceuticals, Inc, Huntsville, AL) 1 tablet twice daily for 3 days, b) ciprofloxacin 250 mg (Cipro, Bayer Corporation, West Haven CT) 1 tablet twice daily for 3 days or c) nitrofurantoin 100 mg (Macrobid, Proctor & Gamble Pharmaceuticals, Cincinnati, OH) 1 capsule twice daily for 7 days. Nitrofurantoin was given for seven days because clinical studies have shown it to be less effective in a three day regimen compared to TMP/SMX [11]. Medication was provided to the patient by a pharmacist on the research team.
Outcome was assessed by telephone interview using a standardized questionnaire to assess resolution of symptoms, compliance with the prescribed regimen, and occurrence of adverse events. Patients were questioned using a prompted list of symptoms and also asked if they experienced any adverse or side effects from the prescribed regimen. QOL was measured using the Quality of Well-Being, a validated, multi-attribute health scale [12]. This scale was selected because it has been successfully applied to acute illnesses, whereas other quality of life scales, such as the SF-36 Health Survey, are better suited to assess chronic illnesses [4]. Higher scores on the Quality of Well-Being reflect higher QOL. The questionnaire was modified slightly to be administered over the telephone and was re-administered by a trained interviewer at each follow up contact. Four follow-up telephone contacts were necessary to evaluate early and late recurrence of cystitis, and were completed at 3, 7, 14 and 28 days after the initial visit.
Clinical cure was defined as the absence of symptoms and no additional antibiotic treatment for urinary symptoms within the 28 day follow up period. Therefore, if a woman reported continued symptoms at day 7 but did not receive additional antibiotic treatment and had resolution of symptoms on follow up at days 14 and 28, she was considered a clinical cure. These criteria were selected with the purpose of mimicking usual care. Urine culture results were classified as positive (greater that 50,000 colony forming units per milliliter), negative (less than 50,000 colony forming units per milliliter) or contaminated (containing multiple organisms at less than 50,000 colony forming units per milliliter).
The two tailed t-test for independent samples was used to compare QOL by outcome and adverse events. One-way analysis of variance with the Bonferroni correction was used to compare QOL by culture results (positive, negative, contaminated) or drug treatment (TMP/SMZ, Ciprofloxacin, Nitrofurantoin). Differences in cure rates between treatments were compared using the chi square test. Fisher's exact test was used to compare outcome and organism susceptibility. We estimate power of 75% to detect significant differences in QOL, adverse events and outcome at the alpha level of 0.05.
Results
Patient enrolment and demographics
One-hundred-fifty-seven patients met the inclusion criteria and were enrolled in the study. Eleven patients were excluded from the analysis of outcome and adverse events. Of these 11, nine were lost to follow up, one had her symptoms resolve before she took the medication and one was scheduled for surgery and her surgeon wanted her to receive a different medication. One patient experienced severe nausea after one day and was prescribed another antibiotic. This patient was counted as experiencing an adverse event and a clinical failure. See figure 1 for the diagram of patient enrollment.
Figure 1 Diagram of patient enrolment and analysis.
The mean age of women enrolled in the study was 34 years (range 18–69). There were no significant differences in baseline demographic between the randomized groups (Table 1). Of the 157 women enrolled, 57% had positive urine cultures, 26% had negative cultures, and 17% had contaminated specimens. E. coli accounted for 82% of the positive urine cultures (Table 2). Of all organisms isolated, resistance to TMP/SMZ, ciprofloxacin and nitrofurantoin were 7%, 1% and 8%, respectively. For E. coli the resistance rates were 7%, 0% and 4% for TMP/SMX, ciprofloxacin and nitrofurantoin, respectively.
Table 1 Patient Demographics
All Patients TMP/SMZ Ciprofloxacin Nitrofurantoin
Age in years (mean ± SD) 34 ± 12 33 ± 12 32 ± 12 36 ± 12
Weight in kg (mean ± SD) 73 ± 21 73 ± 22 72 ± 22 74 ± 22
Overweight, >20% of ideal body weight (%) 26 29 22 31
Number of concomitant medications (mean ± SD, range) 1 ± 2, 0–12 1 ± 2 1 ± 2 1 ± 2
Received antibiotics in previous 6 months (%) 34 35 30 38
Receiving oral contraceptives or hormone replacement therapy (%) 39 33 48 35
Sexually active (%) 85 89 85 80
Urinary tract infection in previous 6 months (%) 37 40 43 28
Table 2 Organism species and susceptibility results of culture specimens from women with acute uncomplicated cystitis
N % % susceptible TMP/SMX Ciprofloxacin Nitrofurantoin
E. coli 73 82 93 100 95
P. mirabillis 5 5.6 100 100 na
K. pneumonia 4 4.4 100 100 50
S. aureus 3 3.3 100 0 100
C. koseri 2 2.2 100 100 100
B-hemolytic Strep 1 1.1 na na na
S. saphophyticus 1 1.1 0 100 100
Total 89 100
na = not available
Clinical outcomes
Clinical cure rates did not differ by treatment with 88, 82 and 87% of patients treated with TMP/SMX, ciprofloxacin and nitrofurantoin, respectively experiencing a clinical cure (p = 0.7). Patients randomized to TMP/SMX who were infected with a resistant organism had lower cure rates, compared to those infected with a susceptible organism (50% vs. 83%), although this difference was not statistically significant due to the low number of infections with resistant organisms (Table 3). Patients treated with ciprofloxacin appeared to experience adverse events at a higher rate (62%) compared to those treated with TMP/SMX (45%) and nitrofurantoin (49%); however, the difference was also not statistically significant (p = 0.2).
Table 3 Percent (no.) of patients successfully treated by organism susceptibility
Susceptible Resistant p value
TMP/SMZ 83% (43/52) 50% (1/2) 0.9
Ciprofloxacin 76% (41/54) --- --
Nitrofurantoin 80% (41/51) 100% (1/1) 0.6
Quality of life outcomes
Overall, the QOL of patients treated for acute uncomplicated cystitis improved over the study period. Quality of Well Being scores improved for all patients, from a mean (± SD) of 0.68 (± 0.03) at baseline to 0.81 (± 0.11) at day 28 (Figure 2). Patients experiencing a clinical cure had significantly better QOL at days 3 (0.77 vs. 0.72, p = 0.04), 7 (0.82 vs. 0.71, p < 0.001), and 14 (0.83 vs. 0.76, p = 0.01) compared to patients who failed treatment (Figure 3). QOL at baseline was not different between patients successfully treated compared to those who failed therapy (0.68 vs. 0.68, p = 0.6) or day 28 (0.82 vs. 0.79, p = 0.5). There was a statistically significant difference in QOL by culture result (p = 0.004). In the post hoc analysis, patients with positive urine cultures had significantly higher quality of well being scores compared to patients with negative urine cultures (0.69 versus 0.66, p = 0.003) but not compared to patients with contaminated urine cultures (0.68, p = 0.23) at baseline. All three groups had improvement in their QOL over the study period (Figure 4). There was no significant difference in improvement of QOL by treatment assignment (Figure 5).
Figure 2 Quality of Life for all patient at each follow up contact.
Figure 3 Quality of Life in patients by clinical outcome at each follow up contact.
Figure 4 Quality of Life in patients by urine culture results at each follow up contact.
Figure 5 Change in quality of life by drug treatment over the study period.
Adverse effects
Seventy-six of the 146 patients included in the analysis reported 94 adverse events. Gastrointestinal effects and vaginitis were the most commonly occurring adverse events, accounting for 26 (28%) and 25 (27%) of the effects, respectively. Headache (12% of patients) and other central nervous system effects such as dizziness (5%) were the next most common side effects. The remaining 12% of adverse events were a variety of miscellaneous complaints. Patients experiencing an adverse event had significantly lower QOL compared to those not experiencing an adverse event throughout the study period, including differences at baseline (Figure 6).
Figure 6 Patient quality of life by adverse event experience at each follow up contact.
Discussion
We have described the impact of a seemingly minor infection, acute uncomplicated cystitis, on the QOL of women. Our study is the first to examine the effect of successful treatment compared to treatment failure on the QOL. Further, we have also shown that experiencing an adverse effect from treatment also directly impacts QOL. While the three antibiotics used in this study differed in terms of cure rates and adverse effects we found little difference in the impact on QOL between treatments.
Current recommendations suggesting empiric therapy with TMP/SMX without obtaining cultures may need to be modified in the era of increasing resistance [1,3,4,13]. However, altering empiric therapy recommendations to either nitrofurantoin or fluoroquinolone antibiotics, without information on local resistance patterns, may result in unnecessary increases in health care expenditures due to increased drug costs. In primary care offices, where limiting the inappropriate prescribing of antimicrobial therapy is receiving focused attention by the Centers for Disease Control and Prevention, information regarding the emergence of resistance and clinical outcome is essential to providing evidence-based recommendations for empiric antimicrobial therapy for acute uncomplicated cystitis.
In our study, we found the resistance rate of E. coli to TMP/SMX remained low, approximately 7% for women with cystitis. We also found that clinical cure did not differ between the treatment regimens tested. These findings are important because it means TMP/SMX continues to be an effective first line antibiotic in our population.
Our findings support those of a previously published study, indicating that uncomplicated cystitis has a significant, measurable impact on patients' QOL [6]. Patients' QOL was improved throughout the study period (figure 2). QOL was significantly higher in patients experiencing a clinical cure compared to those who failed treatment at days 3, 7 and 14 (figure 3). QOL did not differ between these groups at baseline and 28 days. This indicates all patients were experiencing significant decreases in QOL at baseline. It also indicates we are able to measure the impact of cystitis on QOL. The fact that QOL was not different between patients experiencing a cure or failure of initial therapy, reflects the success of alternate treatment after experiencing the initial failure.
QOL at baseline was not different between patients with positive or contaminated urine cultures. However, QOL did differ significantly at baseline between patients with positive and negative urine cultures. One possible explanation for this difference is that patients with culture negative cystitis are more sensitive to the impact of cystitis symptoms on their QOL compared to those with positive urine cultures. That is to say that while their urine culture was negative, the urinary symptoms they were experiencing had a greater impact on QOL compared with patients that have a higher overall QOL. Alternatively, patients with culture negative cystitis may differ in some other way that impacts their QOL compared to patients with culture positive cystits.
We did not find any significant difference between the drug treatments on patients' QOL throughout the study period (figure 5). While there was a trend toward lower QOL in the group of patients treated with TMP/SMZ at day 3, this difference did not reach statistical significance (p = 0.06). Therefore based upon improvements in QOL, all three treatments may be considered appropriate empirical therapy for cystitis.
Previously, the effect of medication adverse events on QOL has only been demonstrated with more severe adverse events such as those experienced with cancer chemotherapy [14-16]. We have shown that even common side effects of antibiotics such as gastrointestinal effects have a negative impact on patients QOL. While some difference was observed in the rate of adverse events reported between treatment groups, this difference did not reach statistical significance. We also found that patients reporting adverse events had lower QOL throughout the study period, including baseline and at 28 days. This may suggest that patients with lower overall QOL may be more sensitive to the addition of drug therapy to their daily routine. The potential for patients having lower QOL reporting more adverse events has not been previously described. This information along with our findings that patients with negative urine cultures also have lower QOL suggests certain patients experience greater impacts on their QOL from seemingly innocuous occurrences such as urinary symptoms of cystitis or taking antibiotics compared to other patients. It is possible that health care providers may need to approach diagnosis and treatment in these patients differently from other patients. More care may be necessary in evaluating symptoms and laboratory information to decide whether treatment is necessary and more care may be needed in drug therapy selection, for example.
Conclusion
In summary, we have found that patients experiencing cystitis have significant decreases in QOL and that QOL is improved by effective treatment. Furthermore we have found that patients reporting adverse events have lower overall QOL and may be impacted to a greater extent by simple infections and adverse events associated with treatment.
Authors' contributions
EJE conceived the study, participated in its design and coordination, performed the statistical analysis and drafted the manuscript. MEE participated in the study design, coordination and patient recruitment and helped revise the manuscript. JDH participated in patient recruitment and helped revise the manuscript. GRB participated in study design, data analysis and manuscript revision. All authors read and approved the final manuscript.
Acknowledgements
We would like to acknowledge the Society of Infectious Diseases Pharmacists Pfizer Research Award for the financial support of this study.
==== Refs
Hooton TM Stamm WE Diagnosis and treatment of uncomplicated urinary tract infection Infect Dis Clin North Am 1997 11 551 581 9378923 10.1016/S0891-5520(05)70373-1
Stamm WE Hooton TM Management of urinary tract infections in adults [see comments] N Engl J Med 1993 329 1328 1334 8413414 10.1056/NEJM199310283291808
Barry HC Cost-effectiveness analysis [letter; comment] J Fam Pract 1997 44 524; discussion 524 5 9191613
Barry HC Ebell MH Hickner J Evaluation of suspected urinary tract infection in ambulatory women: a cost-utility analysis of office-based strategies [see comments] J Fam Pract 1997 44 49 60 9010371
Anderson JP Kaplan RM Coons SJ Schneiderman LJ Comparison of the Quality of Well-being Scale and the SF-36 results among two samples of ill adults: AIDS and other illnesses J Clin Epidemiol 1998 51 755 762 9731924 10.1016/S0895-4356(98)00046-8
Ellis AK Verma S Quality of life in women with urinary tract infections: is benign disease a misnomer? J Am Board Fam Pract 2000 13 392 397 11117334
Patients' Perceptions of Quality of Life After Treatment for Early Prostate Cancer J Clin Oncol 2003 21 3777 3784 14551296 10.1200/JCO.2003.02.115
Downs TM Sadetsky N Pasta DJ Grossfeld GD Kane CJ Mehta SS Carroll PR Lubeck DP Health Related Quality of Life Patterns in Patients Treated With Interstitial Prostate Brachytherapy for Localized Prostate Cancer-Data From CaPSURE J Urol 2003 170 1822 1827 14532784 10.1097/01.ju.0000091426.55735.f0
Fukuhara S Lopes AA Bragg-Gresham JL Kurokawa K Mapes DL Akizawa T Bommer J Canaud BJ Port FK Held PJ The F Dopps W Health-related quality of life among dialysis patients on three continents: The Dialysis Outcomes and Practice Patterns Study Kidney Int 2003 64 1903 1910 14531826 10.1046/j.1523-1755.2003.00289.x
Zubarik R Benway D Moses P Ferrentino N Health-related quality of life (HRQOL) subsequent to esophageal adenocarcinoma staging with endoscopic ultrasound (EUS) Am J Gastroenterol 2003 98 S218 10.1016/S0002-9270(03)01430-8
Hooton TM Winter C Tiu F Stamm WE Randomized comparative trial and cost analysis of 3-day antimicrobial regimens for treatment of acute cystitis in women [see comments] Jama 1995 273 41 45 7654268 10.1001/jama.273.1.41
Kaplan RM Bush JW Berry CC Health status: types of validity and the index of well-being Health Serv Res 1976 11 478 507 1030700
Carlson KJ Mulley AG Management of acute dysuria. A decision-analysis model of alternative strategies Ann Intern Med 1985 102 244 249 3871318
Ballatori E Roila F Impact of Nausea and Vomiting on Quality of Life in Cancer Patients During Chemotherapy Health Qual Life Outcomes 2003 1 46 14521717 10.1186/1477-7525-1-46
Di Maio M Perrone F Quality of Life in elderly patients with cancer Health Qual Life Outcomes 2003 1 44 14525617 10.1186/1477-7525-1-44
Gravis G Bladou F Salem N Macquart-Moulin G Serment G Camerlo J Genre D Bardou VJ Maraninchi D Viens P Weekly administration of docetaxel for symptomatic metastatic hormone-refractory prostate carcinoma Cancer 2003 98 1627 1634 14534878 10.1002/cncr.11687
|
16048650
|
PMC1183236
|
CC BY
|
2021-01-04 16:38:13
|
no
|
Health Qual Life Outcomes. 2005 Jul 27; 3:45
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-45
|
oa_comm
|
==== Front
Int J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 1479-5868-2-71595816810.1186/1479-5868-2-7ResearchTest-retest reliability of a questionnaire to assess physical environmental factors pertaining to physical activity Evenson Kelly R [email protected] Aileen P [email protected] Department of Epidemiology, School of Public Health, University of North Carolina-Chapel Hill, Chapel Hill, North Carolina, USA2005 15 6 2005 2 7 7 4 7 2004 15 6 2005 Copyright © 2005 Evenson and McGinn; licensee BioMed Central Ltd.2005Evenson and McGinn; 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
Despite the documented benefits of physical activity, many adults do not obtain the recommended amounts. Barriers to physical activity occur at multiple levels, including at the individual, interpersonal, and environmental levels. Only until more recently has there been a concerted focus on how the physical environment might affect physical activity behavior. With this new area of study, self-report measures should be psychometrically tested before use in research studies. Therefore the objective of this study was to document the test-retest reliability of a questionnaire designed to assess physical environmental factors that might be associated with physical activity in a diverse adult population.
Methods
Test and retest surveys were conducted over the telephone with 106 African American and White women and men living in either Forsyth County, North Carolina or Jackson, Mississippi. Reliability of self-reported environmental factors across four domains (e.g., access to facilities and destinations, functionality and safety, aesthetics, natural environment) was determined using intraclass correlation coefficients (ICC) overall and separately by gender and race.
Results
Generally items displayed moderate and sometimes substantial reliability (ICC between 0.4 to 0.8), with a few differences by gender or race, across each of the domains.
Conclusion
This study provides some psychometric evidence for the use of many of these questions in studies examining the effect of self-reported physical environmental measures on physical activity behaviors, among African American and White women and men.
environmentexerciseleisure activitiesquestionnairesreproducibility of results
==== Body
Background
Physical activity improves health and quality of life and reduces the risk for several leading causes of death [1]. Yet despite these documented benefits, many adults do not obtain the recommended amounts of physical activity [1]. Barriers to physical activity occur at multiple levels: individual, interpersonal, organizational, community, and public policy or society factors. These factors fit within the framework of the socioecologic model [2,3]. Several studies have reviewed the literature on correlates of physical activity among adults and each has shown that until more recently the focus has been on individual and interpersonal levels of this framework and not on the broader contextual measures [4-7].
In a 1998 review, Sallis et al [8] recommended pursuing a range of strategies to improve the conceptualization of the environment for physical activity, identifying behavior settings in which people are most likely to be physically active, and identifying characteristics of settings that appear to decrease or increase the likelihood of physical activity in that setting. Since that time the research in this field has proliferated. With a new area of study, self-reported measures are needed that have been tested psychometrically on diverse populations, including assessment of reliability and validity if appropriate.
Pikora et al [9] developed a framework for assessing potential environmental influences of walking and cycling based on a review of the literature, interviews, and a Delphi study. The framework included the following physical environmental domains: destination, functionality, safety, and aesthetic. The destination feature relates to the availability of public and private facilities. The functionality feature reflects the physical attributes of the street and path that make up the fundamental structural aspects of the local environment, such as the type and width of the street and the volume, speed, and type of traffic. The safety feature represents both personal safety and traffic safety. The aesthetic feature included both streetscape (e.g., trees, garden and street maintenance, cleanliness, pollution) and views (e.g., sights, architecture).
Using the framework of Pikora et al [9], we developed survey questions or used other published questions to assess these features, since we hypothesized they might be associated with physical activity of adults. In addition, we also developed questions on the natural environment and the use of physical activity facilities. The purpose of this study was to evaluate the psychometric properties of the survey by evaluating the test-retest reliability of the questionnaire in a diverse adult population. In addition, we explored whether reliability differed by race or gender, since we wanted to ensure that the questionnaire was reliable among African American and White women and men, the target for a survey we were conducting.
Methods
Sample
A telephone survey was conducted using a computer assisted telephone interview system (CATI) between January and July 2003 on a random sample of non-institutionalized adults 18 years or older residing in two regions: Forsyth County, North Carolina (NC) and the metropolitan statistical area (MSA) of Jackson, Mississippi (MS). Disproportionate sampling was used for Forsyth County in order to ensure representation for less urban areas outside of the Winston-Salem metropolitan area within the county. Respondents were randomly chosen in two stages: the first stage at the household level and the second stage at the individual level. Surveys were only conducted in English.
A sampling company (Genesys Marketing Systems Group) provided a listing of residential household phone numbers and Clearwater Research, Inc (Boise, Idaho) conducted the telephone surveys. They used Behavioral Risk Factor Surveillance System (BRFSS) telephone survey protocols [10] of up to 15 call attempts for each sampled phone number distributed across weekday, weeknight, and weekends. The average length of the telephone interview was 27 minutes.
Reliability Interviews
Overall 1662 men and women completed the baseline survey. At the end of the interview, 1448 adults were asked if they would be willing to participate in a retest interview. The remaining 214 adults were not asked to participate in a retest interview, because the interview quota was complete. Among these 1448 adults, 76% (n = 1104) agreed to be called back for the retest survey. Reliability information was collected from a 6% (n = 106) purposeful sample of women and men, to ensure approximately equal numbers of participants from both sites, by gender, and by race. The mean time between interviews was 16.8 days (standard deviation 4.2, range 9–30 days). On average, the reliability survey was 5 minutes shorter than the original survey, mainly due to the exclusion of questions on the random selection process and their familiarity with the survey procedures. Each participant provided consent and the study was approved by the Institutional Review Board at the University of North Carolina. Participants were paid $5 for their participation for each survey they completed.
Questionnaire
Most questions on perceived environmental factors were unique to this study, but developed based on work and existing questionnaires by others [11-15]. The questions we examined for test-retest reliability are listed in Appendix 1. While we were guided by the framework of Pikora et al [9], it should be noted that we did not develop questions to ascertain all elements of their framework. Furthermore, some of the questions may fall in more than one feature of their framework. For example, hills could be considered as part of the functional feature, but we chose to group it under the natural environment. It should also be noted that while the majority of the 51 questions under evaluation focused on perceived environmental factors including availability of facilities, there were five questions that focused on frequency of use (how often do you walk to destinations and how often do you use private facilities, public facilities, public schools, and places of worship for physical activity).
Access to Facilities and Destinations
We expanded the destination definition of Pikora et al [9] which related to the availability of facilities within one's neighborhood, to also include access to facilities within the home. In part, this was due to the fact that we were interested in all types of leisure activity as well as transportation activity, and not just walking and cycling. In our survey, destinations were assessed by asking whether participants had places within walking distance. Participants were also asked whether their neighborhoods had sidewalks, trails, or parks and playgrounds and whether lack of these facilities was a barrier to their activity. Other questions pertained to having places to exercise, having equipment or facilities at home, and the availability, use, and quality of private facilities, public facilities, and public schools. Participants were also asked about the use and quality of activity facilities at places of worship.
In addition to exploring item-by-item responses, we calculated an "availability of physical activity facilities index" by adding the responses to the availability of private recreational facilities, public recreational facilities, and public school facilities together, with assignments of excellent = 4, good = 3, fair = 2, poor = 1, or none = 0. The higher the score the more facilities were available to participants.
Functionality and Safety
In our questionnaire, we combined the functionality and safety features of the framework of Pikora et al [9]. Functionality included questions on heavy traffic, speeding cars, noise, and crosswalks and traffic signals. For safety, we asked questions relating to dogs, personal safety, and street lighting. A "crime safety index" was also collected, as developed by Saelens et al [15], using the 6 items found in Appendix 1 and assigning values 1 through 4 to the response options, ranging from strongly disagree to strongly agree, with the last 3 questions (items 4, 5, 6) reverse coded. The score was calculated by adding the 6 items together (with the lower number indicating more crime in the neighborhood) and taking the mean, such that the score ranged from 1 to 4.
Aesthetics and Natural Environment
For the aesthetic feature, we included items on trees, pollution, and trash, litter, or graffiti. Questions pertaining to the natural environment included items on hills and weather.
Physical Activity
Physical activity was assessed by asking if the adults had participated in any moderate or vigorous activity for at least 10 minutes at a time, using questions from the year 2001 BRFSS core module on physical activity [16]. If they responded "yes" to either question, then they were asked how many days per week did they engage in the activity for at least 10 minutes at a time and how much total time per day they spent doing these activities. We grouped participants into three levels based on current physical activity recommendations [17]: those who met recommendations (defined as being moderately active for at least 30 minutes for 5–7 days a week or vigorously active at least 20 minutes for 3–7 days a week), those who were insufficiently active (defined as some physical activity, but not enough to meet recommendations), and those who were inactive (not participating in any moderate or vigorous physical activities for at least 10 minutes at a time in a usual week).
Socio-demographics and Health
All respondents were asked questions regarding age, race, education, and employment. Employment was grouped into two categories: employed or not employed (out of work, homemaker, student, retired, or unable to work). General health was assessed by asking, "Would you say your general health is: excellent, good, average, fair, or poor?" Respondents were also asked, "Are you limited in doing any physical activity or exercise because of a disability or health problem?" If they answered "yes", the respondent was asked if this disability of health problem was mild, moderate, or severe.
Statistical Analyses
Analyses were conducted overall and by gender and race. Intraclass correlation coefficients (ICC) were calculated to assess reliability, based on a one-way analysis of variance [18], along with 95% confidence intervals. The ICC was the proportion of total variance in the measure (subject variability and measurement error) that was due to the true differences between participants (subject variability). We also calculated overall kappa (2 level) and weighted kappa (> = 3 level) coefficients for categorical variables, which were similar to the ICC and thus are not reported. Although the ICC ranges from 0 to 1, in a few cases when the sample size was small, the lower confidence interval fell below 0 and is reported as such. As a rough guide, we followed the ratings suggested by Landis and Koch [19] an agreement level: 0–0.2 poor, 0.2–0.4 fair, 0.4–0.6 moderate, 0.6–0.8 substantial, and 0.8-<1.0 almost perfect. For the index measures, Cronbach alpha coefficients were calculated to assess internal consistency. SAS version 8.01 was used for all analyses.
Results
Among the sample of 106 adults, approximately one-quarter were from each race-gender group: n = 27 African American women, n = 25 African American men, n = 30 White women, and n = 24 White men. Approximately half were from Forsyth County, NC and half were from Jackson, MS. The mean age of participants was 48 years (range 18 to 82 years). The mean and median length of time living at the residence was 13.4 and 7.4 years, respectively (interquartile range 1.8 to 23.0 years). Based on the self-reported data, 42.5% met recommendations for physical activity, 44.3% were insufficiently active, and 13.2% were inactive. Other descriptive characteristics are listed in Table 1 (Additional file: 1).
The test-retest reliability of all measures is reported overall (Table 2, Additional file: 1) and by gender and race (Table 3, Additional file: 1). When exploring differences in reliability by gender or race, we discuss here those measures where the ICC differed by at least two categories, according to Landis and Koch [19].
Reliability of Items on Access to Facilities and Destinations
Items on general access and availability for places to exercise, home equipment and facilities, and neighborhood attributes of sidewalks, trails, and parks/playground (including whether these were a barrier to physical activity) showed moderate to substantial reliability, except for the lower reliability found on the item asking if lack of parks was a barrier to physical activity. Only two items meaningfully differed by gender (i.e., how often do you walk to those places and quality of worship facilities were both lower among women). Several items differed by race, with higher reliability found for African Americans on one question (i.e., having places to exercise) and lower reliability on three questions (i.e., how often using equipment at home, lack of sidewalks, lack of parks or playgrounds). The availability of physical activity facilities index had substantial test-retest agreement and the Cronbach alpha coefficient was 0.81. When examining component questions, the item on availability of public recreational facilities showed higher reliability among men. In general the questions on quality had lower reliability, which may be due in part to the difficult in assessing quality and due to the small sample size because of the skip patterns for those questions.
Reliability of Items on Functionality and Safety
For functionality and safety, the items assessing whether characteristics were a problem in their neighborhood and whether those items were barriers to their physical activity showed moderate to substantial reliability. Only the questions on noise showed differences by gender and race. Reliability was higher among (1) Whites compared to African Americans when determining whether noise was a problem in their neighborhood and (2) among men compared to women when determining whether excessive noise was a barrier to physical activity. The crime safety index had substantial reliability overall, with one component item (i.e., if walkers and bikers on the streets can be seen) performing poorly. For Whites, the component question on talking with people when walking also performed poorly.
Reliability of Items on Aesthetics
The six items pertaining to trash, trees, and pollution represented the aesthetics domain. These items generally were moderately reliable. Reliability was higher among men compared to women and among Whites compared to African Americans when assessing if lack of trees were a problem in their neighborhood. African Americans had higher reliability than White participants on the question pertaining to whether exhaust fumes was a barrier to their physical activity.
Reliability of Items on Natural Environment
The four items pertaining to hills and weather represented the natural environment and generally showed moderate reliability. No meaningful differences were identified by gender and only one difference was found by race. Reliability was somewhat higher among African Americans when determining whether bad weather was a problem.
Discussion
This study documents the test-retest reliability of a questionnaire designed to assess physical environmental factors that might be associated with physical activity in a diverse adult population. Many of the items and scales had moderate and sometimes substantial reliability (ICC between 0.4 to 0.8), with a few differences by gender or race, across each of the domains (e.g., awareness and access to facilities, functionality and safety, aesthetics, and natural environment).
Awareness and Access to Facilities and Destinations
Within the domain of awareness and access to facilities and destinations, we evaluated 21 survey items. In a 2002 review, Humpel et al [20] concluded that most studies examining the relationship between self-reported accessibility to physical activity facilities and physical activity have shown a positive association. We developed a 3-item index to assess availability of physical activity facilities index, which demonstrated substantial reliability. These indices asked about private, public, and public school recreational facilities. Although we asked about use of physical activity facilities at places of worship, we did not inquire about availability of those facilities, and therefore did not include it in our index measures. Future studies may want to consider adding and testing the addition of this question to the index. In general, the questions on quality of these activity facilities had lower reliability, which may be due in part to the difficult in assessing quality and due to the small sample size because of the skip patterns for those questions.
Several studies have examined items on neighborhood features hypothesized to be correlated with physical activity, as first used by Sallis et al [11], to determine presence or absence of characteristics such as sidewalks, hills, heavy traffic, street lights, and unattended dogs. We chose to expand this concept, to ask whether the respondent agreed or disagreed that lack of these items were a problem in their neighborhood, and whether those items were a barrier to their physical activity. The items on availability and barriers to physical activity pertaining to sidewalks, trails, and parks generally displayed acceptable reliability. The question on places to go within walking distance came from another survey [14]. In another study of 344 multi-ethnic women, this question showed substantial reliability (ICC 0.75) [21], which was similar to our findings (ICC 0.63 overall, ICC 0.57 for women).
Functionality and Safety
Functionality refers to design features of the built environment, such as noise, traffic, speeding cars, and crosswalks [9]. Increased functionality and safety are hypothesized to be associated with participation in physical activity [20,22]. Most of the items we examined in this domain were newly derived and displayed acceptable reliability. The crime safety index, which came from Saelens et al [15], had substantial reliability in this study (ICC 0.68). In Saelens et al reliability study of a similar sample size, the ICC for the crime safety index was a bit higher, with an ICC of 0.80; however, these estimates both fall within the same broad category as being "substantial" in reliability.
Aesthetics and Natural Environment
Aesthetic features, such as trees, pollution, and trash, can be difficult to measure but are likely associated with whether or not an individual chooses to be active outdoors [20]. The natural environment, such as weather and hills, are also hypothesized to be associated with physical activity. The survey items for the natural environment and most items for aesthetics were newly derived and displayed moderate reliability.
Implications
These self-reported measures can be used to document individual level perceptions of the neighborhood environment and to explore their association with physical activity. There is also interest in studying the level of agreement between these neighborhood perceptions as compared to objective measures of the neighborhood environment, and the strength of their relationships with physical activity [23,24], which have not been adequately explored in the literature. In some cases, if studies find that perceived measures do not adequately represent a similar objectively measured construct, then this could point to interventions. For example, if an individual perceives that there are no trails in their neighborhood, but the objective measure indicates that there is one nearby, this could indicate the need for better promotion of existing resources. These self-reported measures can also be used to aggregate up responses to create a neighborhood level measure.
Future studies can also utilize these questions to further explore the mediators of change in physical activity. For example, it may be that the perception of an environmental factor as a barrier to physical activity could be a mediating variable (to be targeted in an intervention) between perceived environmental factors and physical activity behavior. This could not be tested within this cross-sectional study but could be explored in future studies.
Study Limitations
Our study had several limitations. First, despite the short time between administrations, true changes, while unlikely, could have occurred between surveys, which would weaken the reported reliability estimates. Second, in the analysis we could not account for whether the same interviewer administered the test and retest interview, but all retests were conducted using similar standardized methodology. Third, the generalizability of this study is somewhat limited in that it was conducted in only two geographic areas and only among African American and White participants. Some of our reliability estimates, especially in the stratified analysis by gender and race, were not very precise due to the sample size. However, we were able to explore whether reliability of these questions differed by race or gender, providing estimates that might be more useful to studies focusing on certain race or gender groups. Future studies should consider their population under study, to determine whether these findings might be generalizable to their sample.
One challenge of this research is deciding if and how to define neighborhood. In this survey, we decided to define "neighborhood" for the participant as a 20 minute walk or one mile from their home. However, for the availability of private facilities, public facilities, public schools, and places of worship, we defined the "community" as the area within a 20-minute drive from their home. In doing this, it may have presented challenges to the participant about recalling geographic areas that may not be familiar to them. There may also have been other factors that affected recall, and thus reliability results, such as disability and the length of time living in the neighborhood. In this study, only 19% of participants reported a moderate or severe disability that might limit their physical activity and 75% of participants reported living at their address 1.8 years or longer. Thus, these were stratification factors that could not be adequately explored with this data.
Conclusion
In conclusion, this study provides some psychometric evidence for the use of many of these environmental measures in future studies examining the effect of self-reported environmental measures on physical activity behaviors for African American and White women and men. The evidence for test-retest reliability of the questionnaire is especially important, as work in this field is expanding rapidly. To date, one study has compared objective to self-reported estimates of several physical environmental measures we used here (e.g., traffic, sidewalks, lights, and crime) [24]. Further work in this area will help to better understand which measures to use and how to interpret findings from these measures.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KRE and APM conducted the study, KRE drafted the manuscript, and APM assisted with the analysis and interpretation. Both authors read and approved the final manuscript.
Supplementary Material
Additional file 1
Click here for file
Acknowledgements
This work was supported by a grant from the American Heart Association. We thank Fang Wen for assisting with these analyses. We also acknowledge Clearwater Research Inc. for the survey data collection and would especially like to thank John Hetherington and Patty Burke. We would also like to thank the two anonymous reviewers for their helpful comments.
==== Refs
U.S. Department of Health and Human Services Physical Activity and Health: A Report of the Surgeon General 1996 Atlanta, GA, U.S. DHHS, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion
McLeroy KR Bibeau D Steckler A Glanz K An ecological perspective on health promotion programs Health Educa Q 1988 15 351 377
Sallis JF Owen N Glanz K, Lewis FM and Rimer BK Ecological models Health behavior and health education Theory, research, and practice 1997 2nd San Francisco, CA, Jossey-Bass 403 424
Dishman RK Sallis JF Bouchard C, Shephard RJ and Stephens T Determinants and interventions for physical activity and health Physical Activity, Fitness, and Health 1994 Champaign, Human Kinetics Publishers
Sallis JF Owen N Physical Activity and Behavioral Medicine 1998 Thousand Oaks, CA, Sage
Trost SG Owen N Bauman AE Sallis JF Brown W Correlates of adults' participation in physical activity: review and update Med Sci Sports Exerc 2002 34 1996 2001 12471307 10.1097/00005768-200212000-00020
Eyler AE Wilcox S Matson-Koffman D Evenson KR Sanderson BK Thompson J Wilbur JE Young DR Correlates of physical activity among women from diverse racial/ethnic groups: A review J Womens Health Gen Based Med 2002 11 239 253 10.1089/152460902753668448
Sallis JF Bauman A Pratt M Environmental and policy interventions to promote physical activity Am J Prev Med 1998 15 379 397 9838979 10.1016/S0749-3797(98)00076-2
Pikora T Giles-Corti B Bull F Jamrozik K Donovan R Developing a framework for assessment of the environmental determinants of walking and cycling Soc Sci Med 2003 56 1693 1703 12639586 10.1016/S0277-9536(02)00163-6
Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System User’s Guide 1998 Atlanta, GA, U.S. Department of Health and Human Services, Centers for Disease Control and Prevention 1 585
Sallis JF Johnson MF Calfas KJ Caparosa S Nichols JF Assessing perceived physical environmental variables that may influence physical activity Research Quarterly Exercise Sport 1997 68 345 351
Brownson RC Eyler AA King AC Shyu YL Brown DR Homan SM Reliability of information on physical activity and other chronic disease risk factors among US women aged 40 years or older Am J Epidemiol 1999 149 379 391 10025482
King AC Castro C Wilcox S Eyler AA Sallis JF Brownson RC Personal and environmental factors associated with physical inactivity among different racial/ethnic groups of US middle- and older-aged women Health Psychology 2000 19 354 364 10907654 10.1037//0278-6133.19.4.354
Eyler AA Matson-Koffman D Rohm Young D Wilcox S Wilbur J Thompson J Sanderson B Evenson KR A quantitative study of correlates of physical activity among women from diverse racial/ethnic groups: the Women's Cardiovascular Health Network Project - introduction and methodology Am J Prev Med 2003 25 5 14 14499804 10.1016/S0749-3797(03)00159-4
Saelens BE Sallis JF Black JB Chen D Neighborhood-based differences in physical activity: An environment scale evaluation Am J Public Health 2003 93 1552 1558 12948979
Centers for Disease Control and Prevention Prevalence of physical activity, including lifestyle activities among adults - United States, 2000-2001 Morb Mort Week Rep 2003 52 764 769
Pate RR Pratt M Blair SN Haskell WL Macera CA Bouchard C Buchner D Ettinger W Heath GW King AC et al Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine JAMA 1995 273 402 407 7823386 10.1001/jama.273.5.402
Streiner DL Norman GR Health Measurement Scales: A Practical Guide to their Development and Use 1995 Second Oxford, Oxford Medical Publications
Landis JR Koch GG The measurement of observer agreement for categorical data Biometrics 1977 33 159 174 843571
Humpel N Owen N Leslie E Environmental factors associated with adults' participation in physical activity: A review Am J Prev Med 2002 22 188 199 11897464 10.1016/S0749-3797(01)00426-3
Evenson KR Eyler AA Wilcox S Thompson JL Burke JE Test-retest reliability of a questionnaire on physical activity and its correlates among women from diverse racial and ethnic groups Am J Prev Med 2003 25(3Si) 15 22 14499805 10.1016/S0749-3797(03)00160-0
Saelens BE Sallis JF Frank LD Environmental correlates of walking and cycling: findings from the transportation, urban design, and planning literatures Ann Behavior Med 2003 25 80 91 10.1207/S15324796ABM2502_03
Brownson RC Chang JJ Eyler AA Ainsworth BE Kirtland KA Saelens BE Sallis JF Measuring the environment for physical activity: A comparison of the reliability of 3 questionnaires for physical activity across the United States Am J Public Health 2004 94 473 483 14998817
Kirtland KA Porter DE Addy CL Neet MJ Williams JE Sharpe PA Neff LJ Kimsey Jr. CD Ainsworth BE Environmental measures of physical activity supports: perception versus reality Am J Prev Med 2003 24 323 331 12726870 10.1016/S0749-3797(03)00021-7
|
15958168
|
PMC1183237
|
CC BY
|
2021-01-04 16:38:04
|
no
|
Int J Behav Nutr Phys Act. 2005 Jun 15; 2:7
|
utf-8
|
Int J Behav Nutr Phys Act
| 2,005 |
10.1186/1479-5868-2-7
|
oa_comm
|
==== Front
Int J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 1479-5868-2-81602950710.1186/1479-5868-2-8ResearchIncreasing weight-bearing physical activity and calcium-rich foods to promote bone mass gains among 9–11 year old girls: outcomes of the Cal-Girls study French Simone A [email protected] Mary [email protected] Jayne A [email protected] John H [email protected] Peter [email protected] Dianne [email protected] Kristine [email protected] Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota, USA2 School of Nursing, University of Minnesota, Minneapolis, Minnesota, USA2005 19 7 2005 2 8 8 24 11 2004 19 7 2005 Copyright © 2005 French et al; licensee BioMed Central Ltd.2005French et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A two-year, community-based, group-randomized trial to promote bone mass gains among 9–11 year-old girls through increased intake of calcium-rich foods and weight-bearing physical activity was evaluated.
Methods
Following baseline data collection, 30 5th-grade Girl Scout troops were randomized to a two-year behavioral intervention program or to a no-treatment control group. Evaluations were conducted at baseline, one year, and two years. Measures included bone mineral content, density, and area (measured by DXA), dietary calcium intake (24-hour recall), and weight-bearing physical activity (physical activity checklist interview). Mixed-model regression was used to evaluate treatment-related changes in bone mineral content (g) for the total body, lumbar spine (L1-L4), proximal femur, one-third distal radius, and femoral neck. Changes in eating and physical activity behavioral outcomes were examined.
Results
Although the intervention was implemented with high fidelity, no significant intervention effects were observed for total bone mineral content or any specific bone sites. Significant intervention effects were observed for increases in dietary calcium. No significant intervention effects were observed for increases in weight-bearing physical activity.
Conclusion
Future research needs to identify the optimal dosage of weight-bearing physical activity and calcium-rich dietary behavior change required to maximize bone mass gains in pre-adolescent and adolescent girls.
==== Body
Background
Osteoporosis currently affects over 25 million people in the United States [1,2] and low bone mineral density (BMD) is a major factor involved in bone fractures in postmenopausal women and the elderly. [3,4] Variation in bone mass accumulation during childhood and adolescence may be an important determinant of the risk of sustaining osteoporotic fractures during later adulthood, since 45% of the adult bone mass is built and enlarged during adolescence. [4,5] Thus, maximizing bone mass in youth may be one of the most effective preventive strategies available. [5]
Although an estimated 60–80% of the variance in peak bone mass (PBM) is attributable to genetic variables, [6-9] eating and physical activity behaviors represent modifiable variables that make small but significant contributions to the attainment of (PBM). [10] From a population perspective, small increases in PBM achieved through eating and physical activity behavior changes could significantly decrease the population-based fracture risk later in life. [10]
To date, the majority of primary prevention interventions for increasing bone mineral mass among children and adolescents have focused on increasing calcium intake through calcium supplementation from pills or fortified foods. [11-15] The study design of these primary prevention trials has been individual-randomized trials with no behavioral change intervention component. In general, results of these studies have shown positive associations between calcium (CA) supplementation and increases in bone mass. [11-16] However, bone mass gains are not sustained once supplementation ceases. [11,12,16]
Several physical activity primary prevention trials have been implemented in school-based settings (for example in physical education classes) and have randomized either individuals or classrooms. [17,18] Similar to the CA supplement trials, however, the studies have not included a behavioral change intervention component. Results of the available studies suggest a positive impact of weight-bearing physical activity (WBPA) on bone mass gains in children, particularly in the femoral neck, spine and total body. [16,19,20] However, little is known about the amount and type of physical activity (PA) needed to increase BMD gains or whether there is a threshold or dose-response relationship.
It has been hypothesized that WBPA and CA intake may produce additive or interactive effects on bone mass gains. We are aware of only one randomized controlled trial that simultaneously targeted increased intake of CA-rich foods and WBPA to evaluate bone mass gains among children. [20] Sixty-six eight-year old girls were randomized to either a low-impact or high-impact PA class during school (three times per week for 20 minutes), and consumed either CA-fortified foods or the same unfortified foods for 8.5 months. No behavioral change intervention component was included. Results showed that the effects of increases in CA intake and WBPA on rate of bone gain varied at different skeletal sites. Clearly, additional research is needed to clarify the independent and interactive effects of each bone-related behavior, WBPA and CA intake, on site-specific and generalized bone mass gains in girls. Furthermore, community-based studies will provide information about the effectiveness of behaviorally-based, primary prevention interventions that are implemented in settings that are widely generalizable.
The present paper describes the primary outcomes of a community-based, group-randomized trial to increase bone mass gains among 9–11 year-old girls by targeting increases in intake of CA-rich foods and WBPA. It was hypothesized that the behavioral intervention would result in increases in dietary CA intake and WBPA, and in the rate of bone mass accrual. Innovative aspects of the present study included 1) its focus on changing eating and PA behaviors to increase bone mass growth; 2) its focus on pre-adolescent girls, a group at risk for declines in CA intake and PA levels as they develop into adolescence; and 3) its unique collaboration with a community-based organization (the Girl Scouts of America) as a channel to implement a health behavior intervention.
Methods
Study Overview
Thirty 5th-grade Girl Scout troops were recruited and randomized to a two-year behavioral intervention program (n = 15 troops) or to a no-treatment control group (n = 15 troops). The behavioral program was implemented during 5th and 6th grades by trained troop leaders as part of the regular troop meetings. The intervention program was based on Social Cognitive Theory [21] and consisted of ten 90-minute activity-based sessions during each of the two years. It focused on the development of behavioral skills to choose CA-rich foods and to engage in WBPA (see Table 1). Behavioral goals for the intervention were to increase daily dietary CA intake to 1300 mg/day (increase of 4 daily servings of CA rich foods; about 800 CA mg/day) and to increase WBPA to 120 minutes per week. A continuously available, interactive web-based program, and a one-week summer camp between 5th and 6th grade years, were implemented as components of the intervention program. Parents also were targeted through the web-based program. Control troops did not receive any program and conducted their usual troop meeting activities during the two-year intervention period. Evaluation was conducted with individual girls and a parent at clinic visits at baseline prior to randomization, at one-year follow-up, and at the end of the study (two year follow-up).
Table 1 CAL-GIRLS INTERVENTION PROGRAM (Selective Version)
INTERVENTION COMPONENT 5th GRADE 6th GRADE
TROOP MEETINGS (behavioral curriculum) Calcium Matching Game
Bone Bead Bracelets
CAL-Girls Club Radio Show
Bone Builder Breaks
CAL-Girls Web Site
Grocery Store Scavenger Hunt
Calcium-rich Food Sculptures
Jump Rope w/U of M Staff
Bone-opoly Game
Creating Calci"yum" Cereal Treats
CAL-Girls Time Capsules
"Got Milk?" Mustaches & Photos
CAL-Girls Parent Party Digiwalkers
CAL-Girls Web Site
CAL-Girls Smoothies
CAL-Girls Makin' Movies & Filming at MTN Studio
Decorate CAL-Girls T-shirts
Fruit & yogurt parfaits/sundaes
CAL-Girls on the Grow! Game
CAL-Girls Are Great! Booklets
Aerobic Step-building Workshop
YWCA/YMCA Step-Aerobics Class
Calcium-rich Party dips & snacks
Step Right Up! Parent Party
CAL-Girls HOME ACTIVITIES Under Our Roof Activities
• Find 5 calcium-rich foods in your house. Place a CAL-Girls sticker on each item.
• Explore the CAL-Girls Web Site and post a message to Calcy the Cow.
• Spend at least 20 minutes each week for two weeks doing Bone Builder Activities with your family.
• Create three calcium-rich meals with your family. CAL-Girls Challenge Cards Examples: • Add your favorite kind of cheese to a snack, soup or main dish.
• Ask an adult what calcium-rich foods they ate today
• Visit a fitness trail. Spend 20 minutes doing Bone Builder physical activities.
• Ask adults in your family to come up with FIVE reasons why it's important to eat Bone Builder Foods.
CAL-Girls WEB SITE Calcium News
Bone Building Activities
Web Site Fun
Virtual Grocery Store
All About Cheese!
The Cow Connection
In Your Community
Sing, Shake, Celebrate
Mostly Milk
Is It Over? Welcome Back!
Big News on Bones
Spread the Word
Time for Bones!
The Miracle of MILK!
Bigger, Better & Stronger Bones!
Step Up for Strong Bones!
"Y" Strong Bones
What's Your Calcium Sign?
Bones Last a Lifetime!
CAL-Girls SUMMER CAMP* COHORT ONE (2001) COHORT TWO (2002)
Jump Rope
Recycling & The Environment
Ice Cream in a Bag
Farmer's Market
Creating Cereal Necklaces
Making Pizza
Homemade Fruit Leather or Yogurt Pancakes
Swimming
Camp Olympics Jump Rope
Germs & Hand-washing
Food Group Bingo
Food Jeopardy
Milk, and Sugar in Soda Pop
Milk Crossword Puzzle
Ice cream in a Bag
Making Sun Pizzas
Swimming
Autograph Books
*NOTE: The CAL-Girls Summer Camp was offered in the summer between 5th and 6th grade.
Troop Recruitment
The study was conducted in collaboration with two local councils of the Girl Scouts of America. Girl Scout Council staff were active collaborators in the program development and implementation. Troops from the Minneapolis and St Paul metropolitan area were recruited to take part in the study through mailed fliers to troop leaders and troop leader meeting announcements. Troop eligibility criteria were: 1) troop size = 8 girls; 2) parental consent and girl assent from each troop member to participate in troop meetings with intervention program activities; and 3) troop plans to remain together at minimum two more years. Troops that met eligibility criteria were scheduled for a recruitment presentation by the study Project Director or the Principal Investigator at a regularly scheduled troop meeting. Recruitment visits included a brief video showing several Girl Scouts from our pilot study completing clinic measures and the pilot Girl Scout troops engaging in intervention activities. A written study description and consent forms were distributed and discussed at the recruitment meeting. Parents were invited to ask questions and were asked to sign the consent form at the meeting. Girls were asked to sign assent forms. The study was reviewed and approved by the University of Minnesota Research Participants Internal Review Board.
Intervention Program Implementation
The intervention program was implemented by trained troop leaders with support from research staff. Troop leaders assigned to the intervention group were trained by research staff two times per year during each of the two intervention years. Most training sessions were held at the Division of Epidemiology and Community Health. The first training session focused on intervention activities for sessions one through five, and the second training session focused on intervention activities for sessions six through ten. Research staff reviewed the intervention session content, demonstrated activities with troop leaders, and answered questions. All materials and supplies needed for implementing the intervention activities were provided to troop leaders. Grocery store gift certificates were provided to troop leaders to cover the costs of the CA-rich snacks that were purchased for each intervention session.
Research staff assisted with the intervention program implementation in several ways. Research staff conducted all sessions in which the girls received training on accessing and completing the web-based activities, and were responsible for creating and managing the website, and responding to girls' and parents' queries regarding web program access. Research staff also were responsible for implementing the home-based intervention activities, community-based intervention activities and the summer camp.
Clinic Visit
All evaluation data were collected by trained research staff during clinic visits held at the University of Minnesota. Clinic visits were conducted at baseline, prior to randomization, and after the first and second years of the study (a total of three clinic visits). Girls were scheduled individually with a parent. Clinic visits required about two hours to complete. Ninety-five percent of the girls who were enrolled in the troops (n = 322 / 340) completed the baseline clinic visit. Reasons for non-completion included repeated scheduling problems (n= 6) and refusal due to concerns about the bone scan (n= 12). All data collectors and bone densitometry technicians received standardized training and met certification criteria prior to collecting participant data. Bone densitometry technicians were state-certified to conduct radiology scans, and also received study-specific bone densitometry training.
Measures
Bone Mass
Bone mass was measured using dual-energy x-ray absorptiometry (Hologic, Inc; QDR-4500; Waltham, MA). Change in bone mineral content (BMC) was the primary outcome variable for the study because it best reflects bone mass change in developing children. [22] Bone mineral density (BMD) and bone area (BA) were also measured and analyzed, and are presented for descriptive comparison with the published literature. Standard positioning protocol and software was used to complete and analyze the scans. [23] BMC (gm) was measured for the total body and at the lumbar spine (L1-L4), total hip, including one sub-region (femoral neck), and one-third (distal) radius. Separate scans were conducted for total body and for each sub-region. At all data collection periods, for hip scans, the femoral neck box was kept constant. However, at follow-up assessments, if necessary, the hip region of interest (ROI) was enlarged to allow for growth. Unossified vertebral transverse processes were outlined manually. A high correlation (r = .98) was observed between the two repeated baseline scans for participants in cohort one (n = 193). Therefore, each site was scanned only once for participants in cohort two (n = 129) and for subsequent cohort scans. Body composition calculations for fat mass (g), lean mass (g), and percentage of body weight that is fat were derived from DXA measures.
Ten inter-operator reliability scans at each bone site were collected during each data collection period. Averaged across assessment periods, BMC reliability was high (whole body r = .995; spine (L1-L4) r = .991; total hip r = .988; femoral neck r = .975; 1/3 distal radius r = .983) and percent bias was approximately 1% at each site. For additional quality control, each technician reviewed the scans measured by the other technician. Scan interference such as gross movement (total body) or casts (radius) eliminated data from at least one of the five scans for 40 girls at baseline and 10 girls at either follow-up period. The lumbar spine system phantom was scanned daily to maintain quality control, and the coefficients of variation (CV) for our machine was as follows: BMC CV = 0.61%; BMD CV = 0.39%; BA CV = 0.47%.
Calcium Intake
The Nutritional Data System for Research [24] from the University of Minnesota' s Nutrition Coordinating Center was used to collect information regarding dietary intake. Dietary intake was measured by trained and certified dietary interviewers using a single 24-hour recall with a multiple-pass approach. A single recall was used instead of multiple recalls because the purpose of the recalls was to estimate group means, not individual intake. [25,26] Upon completion of the interview, the girls' parent was asked specific questions regarding the type and dose of vitamin/supplement intake, particularly CA supplements and CA-fortified products. A certified nutrition coordinator reviewed all recall data for quality assurance. Data on dietary energy and dietary CA intake (without supplements) were examined.
Weight Bearing Physical Activity
Weight bearing physical activity was measured using the Physical Activity Checklist Interview (PACI). [27] The PACI is an interview-administered instrument to assess the previous day recall of 24 physical activities. Among children, moderate validity has been reported for the PACI when compared to objective measures of PA with a correlation of .51 between the PACI and a heart rate index, and a correlation of .33 between the PACI and an accelerometer. [27] Girls reported minutes spent in each activity during three day-parts (morning, afternoon, evening). Intensity of activity was estimated by girls' self-report of how out of breath they were during each bout of the activity reported (none of the time, some of the time, or most of the time). MET values were assigned to each of the 24 activities [28] and were weighted based on self-reported intensity. [29]
Physical activities were defined by a consensus among the study investigators as either weight-bearing or non-weight-bearing. WBPA behaviors were judged to exert significant gravitational force and inertial impact on the body (e.g., jumping rope, jogging, aerobic dance), and classification was consistent with other research that measured the vertical ground reaction forces of physical activities among youth. [30] The only activities that were not included in the WBPA calculation were: swimming, bicycle riding, and household chores. A measure of osteogenic (bone-forming) PA was calculated to quantify the dose of bone-forming PA reported. [31,32] Osteogenic scores were calculated as: Ground Reaction Force × [loge(minutes +1) × (sessions per week -1)]. [32]
Weight and height
Weight and height were measured by trained and certified technicians following a standardized protocol. [33] Weight was measured to the nearest .1 kg on a calibrated balance-beam scale. Height was measured to the nearest millimeter using a wall-mounted stadiometer; two height measurements were taken and averaged for data analyses. Girls were measured in cloth gowns without shoes. Body mass index was calculated as weight (kg)/height (m)2.
Body Composition
Lean body mass (g) and body fat mass (g) were assessed by the DXA scan. Lean mass (without bone) and fat mass values were translated to kilograms (kg) and divided by body weight (kg) to compute percentage body lean and percentage body fat.
Sexual maturation
Sexual maturation was measured using direct observation by trained female staff. Breast and pubic hair development were rated separately as stages 1–5 according to standardized procedures. [34] Girls also completed two self-report measures of pubertal development: one using a pictorial format based on Tanner staging [35,36] and one using a series of questions about the girl's physical development from two published studies. [37,38] Approximately one-third (35%–46%) of girls declined to complete the observed pubertal staging protocol at one of the three data collection points. Therefore, a multiple imputation procedure was used to estimate the missing observed values (breast and pubic hair stages separately) from their available self-reported physical development data (i.e., rating of maturity, overall rating of physical development, breast growth, rating of body build), measured height and weight, and age (specific items and procedures can be obtained from the authors). For analyses, imputed data were only used when observed data were not available. Imputation was used in 33% of the 338 girls for both breast development and pubic hair.
Medical History Information
At each data collection point, parents reported their child's history of diseases and conditions that might affect bone development. Girls' data were flagged for the following diseases/conditions: bone-related diseases, daily use of steroids known to affect bone metabolism, multiple broken bones, treated growth problem, or other medical conditions (e.g., Turner's syndrome, cerebral palsy).
Demographic Information
Demographic Information was collected via questionnaire from the parent who accompanied the girl to the clinic. Ninety-two percent of the accompanying parents were the girls' mothers; 7% were fathers; and 1% were other female guardians. Parental education, annual household income, parent and child race/ethnicity, and marital status were self-reported.
Process Evaluation Data: Troop Leader Implementation Fidelity and Troop-Level Intervention
Participation
Data on troop participation and troop leader implementation were collected from several sources.
Fidelity of Implementation
Fidelity of Implementation was measured based on direct observation of troop meetings by trained research staff, who observed 60% of all sessions (six of ten sessions in grade 5 and six of ten sessions in grade 6). At each session, staff observed 1) the troop leader's program material coverage; and 2) whether troop leaders verbally encouraged girls during the troop meetings to do WBPA and eat CA-rich foods outside the troop meetings.
Troop Participation
Troop Participation was measured based on four troop-level indicators: 1) research staff observations of the proportion of girls who attended selected troop meetings (six of ten sessions in grade 5; six of ten sessions in grade 6); 2) proportion of home activities completed by the girls in both years of the intervention; 3) web program participation (proportion of web program activities completed by the girls);and 4) proportion of girls in each troop attending the summer camp. These data were summarized by troop to create a troop-level participation score.
Statistical Analysis
All statistical analyses were conducted using SAS (SAS Release 8.2). [39] The primary outcome analysis was change in BMC and was conducted using mixed-model regression. One common model was developed for all analyses. However, each bone site (total body, spine (L1-L4), total hip, femoral neck, and one-third distal radius) and bone mass outcome [BMC (gm); BMD (gm/cm2) and BA (cm2)] was analyzed separately. For the mixed-model regression, individual girls were nested within troops and troops were nested within treatment condition. Treatment condition was the primary independent variable. Because the study is a randomized controlled trial, the outcome may legitimately be analyzed without adjustment for covariates. However, covariate adjustment is likely to be necessary in a group-randomized trial because the small number of troops to be randomized increases the probability that the realized randomization will not balance confounders fully. This is especially true when the area is a growth outcome as the girls enter puberty. Moreover, adjusting for measures of maturation increases the precision of the estimate of the dependent variable.
The following covariates were included in the bone mass outcome models: baseline menarcheal status (no/yes), and change in menarcheal status over the two years ("no change" or "achieved menarchy"); baseline height and change in height at interim and follow-up; baseline Tanner stage breast development and change in Tanner stage of breast development (categorical: no change, +1, +2 or +3 stages advanced from baseline), baseline weight, baseline age, and race (reference group: Caucasian). The change in weight was not included as it was thought that the intervention might have an impact on the rate of change in weight. The interim and final measures (adjusted for baseline) were treated as correlated outcomes within girl (ie., repeated measures). The troops were treated as two random effects, one for level and the other for change in troop means as correlated outcomes. Because the intervention was applied to the intact social groups (the troops), the proper error for assessing the effect of intervention is variation at the level of the changes in troop means (i.e., based on the component of variance identified as the troop-by-time interaction). [40]
A similar mixed-model regression was developed to examine secondary behavioral outcome variables: changes in dietary CA intake and WBPA. Models in which change in dietary CA intake was examined were adjusted for total energy intake in addition to the covariates listed above. Height and change in height were not included as covariates in the analysis of change in WBPA or CA intake.
All thirty troops remained in the study for the entire study duration (i.e., troop-level retention was 100%). Ninety-five percent of the girls enrolled in the 30 troops at baseline completed the baseline clinic evaluation (322/340). Of this group, 300 completed the second clinic visit, and 296 completed the third clinic visit. Overall, the individual retention rate was 92%. In addition, 16 girls joined a troop within the two years after baseline; seven girls completed the second and third clinic visits, two girls completed only the second clinic visit, and seven girls completed only the third clinic visit. Data from these girls were included in the calculation of troop means used in the analysis.
Girls were flagged for reporting a lifetime history of bone-related diseases or medical conditions that may affect bone metabolism (e.g., daily steroid use, treated growth problem), or a condition that may inhibit participation in the intervention (e.g., Turner's syndrome, cerebral palsy). Twenty-three, fourteen and twelve girls were flagged accordingly at baseline, one-year and two-year follow-up visits, respectively. Analyses conducted without including data from these girls showed materially identical results and therefore are not presented.
Results
Descriptive Characteristics
Table 2 shows the baseline demographic, physical and behavioral variables of the troops by treatment group condition. Parents of participating girls were mostly white, married, college-educated and upper-middle income.
Table 2 Baseline and follow-up demographic and physiological variables among 30 Girl Scout troops (troop-level means) by treatment group (n = 15 control, n = 15 intervention)
Label Baseline Follow-up 1 Follow-up 2 Two-yr Intervention Effect
C I C I C I EffectA SE
Demographics
Girls' race (% white) 90 90 -- -- -- -- -- --
Family income (% > $90 k) 51 46 -- -- -- -- -- --
Parent's education (% = college degree) 50 59 -- -- -- -- -- --
Girls' age (years) 10.5 10.4 11.5 11.4 12.4 12.3 0.04 0.02
Physical Development
Breast Stage 1 (%) 37 47 30 40 1 1 -- --
Breast Stage 2 (%) 55 51 53 55 24 23 -- --
Breast Stage 3 (%) 7 2 17 5 44 53 -- --
Breast Stage 4 (%) 0 0 0 0 21 16 -- --
Breast Stage 5 (%) 0 0 0 0 10 7 -- --
Pubic Hair Stage 1 (%) 62 73 23 34 5 7 -- --
Pubic Hair Stage 2 (%) 24 21 33 33 25 23 -- --
Pubic Hair Stage 3 (%) 13 6 28 25 29 38 -- --
Pubic Hair Stage 4 (%) 2 0 15 9 38 31 -- --
Pubic Hair Stage 5 (%) 0 0 1 0 5 1 -- --
Menarcheal (% yes) 2 0 14 6 34 35 -- --
Body Size and Composition
Height (cm) 143.5 142.7 149.7 149.2 155.6 155.5 0.77 0.46
Weight (kg) 40.5 39.5 45.1 44.2 49.9 49.8 0.90 0.75
BMI (kg/m2) 19.5 19.3 20.0 19.8 20.5 20.5 0.28 0.26
Fat Mass (%) 26.1 25.4 25.6 25.3 24.7 24.7 0.62 0.57
Lean Mass (%) 71.2 71.8 71.5 71.9 72.2 72.3 -0.58 0.55
A Net difference between change in intervention and control troop-level means, I = intervention troops; C = Control troops
Girls' average age was 10.5 years, 43% were categorized as Tanner stage 1 for breast development, and 1.4% were menarcheal. Calcium intake at baseline averaged 1265 mg/day. Average total energy expenditure in WBPA was 340 kcal/day and WBPA weighted MET score averaged 488, similar to the PACI scores for moderate-to-vigorous PA among 5th grade girls in a school-based nutrition and PA intervention. [24,41] At baseline, control troop mean age was significantly higher than the intervention group mean (p < .05). However, the age differences were not clinically or practically meaningful.
Process Evaluation
Table 3 shows data on the intervention troop leader implementation (intervention fidelity) and troop participation in intervention activities (n = 15 troops). Troop leader implementation was consistently high across all troops. Troop participation levels showed some variability. However, no troops had poor participation when considering participation and exposure to all of the intervention components. The intervention component with the highest participation was troop meeting attendance (ranging from 65% to 94% in 5th grade, and 61% to 97% in 6th grade). The intervention component with the lowest participation was use of the program website (repeated use among 50% of girls in the 5th grade, and 24% of the girls in 6th grade).
Table 3 Process evaluation for CAL-Girls intervention troops (n = 15)
Intervention participation 5th grade 6th grade
Website Mean Range Mean Range
Logged in at least once (%) 81.6 42.9–100.0 46.5 9.1–75.0
Logged in more than once (%)* 50.1 10.0–100.0 24.9 0.0–60.0
Posted at least one message on bulletin board (%) 46.5 7.1–100.0 7.6 0.0–62.5
Number of website hits 34.6 12.3–89.9 13.5 0.8–117.5
Number of quizzes completed 0.87 7.7–2.4 -- --
Number of Puzzled Patty activities completed -- -- 0.64 0.0–3.1
Camp attendance (%) 45.4 22.2–77.8 -- --
Proportion of troop attending meetings (%) 84.9 65.5–94.0 81.7 61.1–97.6
Proportion of troop completing home activities (%) 51.8 32.2–91.7 44.3 18.5–73.8
Intervention fidelity
Troop activities completed (%) 86.1 74.2–96.9 80.2 61.2–89.3
Leaders followed guide all or most of time (%) 95.9 75.0–100.0 95.3 66.7–100.0
Leaders encouraged girls to continue behaviors outside of troop (%) 85.6 60.0–100.0 72.4 33.3–100.0
*Variable not included in website factor score.
Effects of the Intervention on Bone Mass Gains
Table 4 shows treatment-related changes in bone mass for each bone site. Treatment group differences in change in BMC were not significant for total body, total hip, femoral neck or 1/3 distal radius. Changes in total body BMC slope over time by treatment group (Figure 1) and in femoral neck BMC slope over time by treatment group (Figure 2) show similar increases among both the intervention and control groups.
Table 4 Baseline and follow-up dietary, physical activity, and bone variables among 30 Girl Scout troops (troop-level means) by treatment group (n = 15 control, n = 15 intervention)
Label Baseline Follow-up 1 Follow-up 2 Two-year Intervention Effect
C I C I C I EffectA SE
Dietary Intake
Calcium (mg)* 1274 1199 1245 1401 1310 1394 92 82
Energy (kcal) 2139 2136 2108 2138 2127 2047 -- --
Physical Activity
Weight-bearing PA score@ 507 448 539 472 555 531 49 91
Osteogenic score$ 29 24 25 23 25 22 -- --
Bone Mineral Area (BA)
Total body (cm2) 1287 1274 1443 1421 1600 1606 8.1 7.2
Total hip (cm2) 24.1 24.0 26.3 26.3 28.3 28.2 -0.26 0.11
Spine (L1-4) (cm2) 38.6 38.2 42.7 42.1 46.7 46.7 .25 .33
Femoral neck (cm2) 4.0 4.0 4.2 4.2 4.4 4.5 .08 .03
1/3 distal radius (cm2) 2.1 2.1 2.2 2.2 2.3 2.3 .01 .01
Bone Mineral Content (BMC)
Total body (g) 1106 1084 1292 1241 1495 1485 -4.9 14.5
Total hip (g) 17.9 17.3 20.9 20.4 24.2 23.7 -.63 .29
Spine (L1-4) (g) 26.0 25.3 31.6 30.4 38.1 37.8 -.06 .58
Femoral neck (g) 2.8 2.8 3.1 3.1 3.5 3.5 .03 .04
1/3 distal radius (g) 1.1 1.1 1.2 1.2 1.3 1.3 .01 .01
Bone Mineral Density (BMD)
Total body (g/cm2) 0.85 0.85 0.89 0.87 0.93 0.92 -.003 .01
Total hip (g/cm2) 0.74 0.72 0.79 0.77 0.85 0.83 -.010 .01
Spine (L1-4) (g/cm2) 0.67 0.66 0.73 0.71 0.81 0.80 -.011 .02
Femoral neck (g/cm2) 0.71 0.69 0.75 0.72 0.80 0.78 -.008 .007
1/3 distal radius (g/cm2) 0.52 0.51 0.54 0.54 0.57 0.57 .001 .004
A Adjusted net difference between change in intervention and control troop-level means (adjusted for baseline age, race, weight; baseline and change in height, menarcheal status, Tanner breast stage).
* Adjusted for energy intake; and above variables, not including height or change in height)
@ Calculated variable of total weight-bearing physical activity minutes, MET value, and self-reported intensity.
Adjusted for above variables (in A), not including energy intake, height or change in height.
$ Calculated variable of ground reaction force value × (Ln (minutes in activity +1)
Figure 1 Whole Body: BMC (2x SE bars). CALGIRLS Study.
Figure 2 Femoral Neck of R. Hip: BMC (2xSE bars). CALGIRLS Study
Effects of the Intervention on WBPA and Dietary CA Intake
Table 4 shows adjusted changes in WBPA and dietary CA intake by treatment group. Overall, girls increased their CA intake and their WBPA over the two-year intervention. No significant treatment-related differences in change in WBPA were observed. Increases in dietary CA were significantly greater among intervention troops compared to control troops over the two-year period. However, both intervention and control troops remained at or near recommended CA levels throughout the study period.
Discussion
The results of the present study showed that a community-based behavioral intervention to increase dietary CA intake and WBPA among 9–11 year old girls enrolled in the Girl Scouts program was not effective in increasing bone mass gains, or frequency of WBPA, over a two-year period. Significant increases in dietary CA intake were observed as a result of the intervention. However, all troops were close to recommended dietary CA levels throughout the study period. These null results were observed despite the documented high levels of troop intervention implementation and participation. Nevertheless, these results provide data that may be useful in designing future dietary and PA behavioral interventions that target youth in a community setting.
Reasons for the lack of significant effects of the intervention on behavior or bone changes are not clear. It is not surprising that no significant increases in bone mass changes were observed as a result of the intervention, since no significant changes in WBPA were produced, and CA intake was at recommended levels at baseline. Perhaps most surprising is the lack of behavior change observed among intervention troops, despite the high level of implementation and participation in the intervention activities. The troop leaders assigned to the intervention were well trained, motivated and thorough in their implementation of the intervention program activities during the troop meetings. Participation among girls was consistently high. The intervention activities comprised a significant portion of the troop meeting time for a two-year period.
These results suggest that a more structured intervention may be needed to impact WBPA changes when implemented in a community-based setting. Previous PA interventions that targeted bone mass changes among children were conducted using a more controlled intervention design. PA increases were achieved through supervised PA programs with a standardized frequency, duration and type of PA. [16,18-20,22,42,43] By contrast, the behavioral intervention program implemented in the present study focused on educational activities, and behavioral skill-building such as goal-setting, self-monitoring and incentives. A more structured intervention, one that increases the frequency, intensity and duration of WBPA, may be needed to increase WBPA enough to impact bone mass gains. This type of structured intervention may be more easily implemented in a school-based setting in which daily physical education classes are part of the normal schedule. [18,19,22,42] In addition, greater involvement of parents in structuring the home environment to increase WBPA and CA-rich food opportunities for their child may be needed to increase the magnitude of the intervention-related behavior changes and thereby increase bone mass gains. These more structured environmental intervention components may be coupled with a behavioral change intervention component for maximum effects on behavior change.
Baseline levels of CA intake and WBPA may also be important factors that influence the magnitude of the intervention effects on behavior changes. The girls in the present study had high baseline levels of CA intake, thus making further increases more challenging compared to a population with lower initial CA levels. Moreover, the high mean CA intake at baseline approached recommended intake for this age group (NIH, 1994). Additional CA may not have been associated with additional gains in bone mass even if the intervention was successful in producing further increases in CA intake. For the WBPA intervention component, a more frequent, structured, and supervised PA program and or a program that focuses on specific types of PA that are the most osteogenic, such as jumping, may prove more effective in increasing BMC gains.
Strengths of the present study include its strong study design, high level of implementation fidelity and participation, high-quality measurement of outcomes, and high levels of participant retention in the evaluation cohort. An additional strength was its implementation in a community-based youth group by trained community-based volunteers. Potential weaknesses include the use of a population with initially high levels of one of the behaviors targeted for increase (CA intake); lack of structure in the WBPA behavioral component; the use of an intervention without a strong environmental change component; and the reliance on self-report measures of PA.
The present study found that despite high levels of program implementation and participation, a community-based behavioral intervention to increase bone mass gains in 9–11 year old girls was not successful in changing bone mass outcomes or WBPA. Future research is needed to further explore the effectiveness of community-based behavioral interventions for dietary CA and WBPA increases that target children. However, a useful prelude to such research would be a series of additional controlled studies that examine physical activity dosages and the types of activities that produce the strongest effects on bone mass gains in targeted age groups. [20] A similar set of systematic dosage studies for change in CA intake in populations that have an initially low or high CA level are needed. Together, the results of such studies could inform a more focused and potentially more effective behavioral intervention for bone mass gains targeting children that could be implemented in a free-living setting in the community.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SAF developed and directed the overall study and wrote the manuscript. MS contributed to the study design, direction and implementation and manuscript writing. JAF was the project director responsible for daily oversight of intervention evaluation activities and contributed to the manuscript writing. JHH contributed to the project design and oversight, data analysis and manuscript writing. PH contributed to the study design, data analysis and manuscript writing. DNS contributed to the study design, development and manuscript writing. KE contributed to the manuscript writing.
Acknowledgements
This research was supported by NIH RO1 HD37743.
The authors wish to thank the Girl Scouts Councils of Greater Minneapolis and St Croix Valley for their collaboration on the study.
Dr Pedro Teixeria served as Guest Editor for this paper.
==== Refs
NIH Consensus Development Panel on Optimal Calcium Intake Optimal calcium intake JAMA 1994 272 1942 1948 7990248 10.1001/jama.272.24.1942
Riggs BL Melton LJ Involutional osteoporosis N Engl J Med 1986 314 1676 1686 3520321
Matkovic V Kosital K Simonovic I Buzina R Brodarec A Nordin BE Bone status and fracture rates in two regions of Yugoslavia Am J Clin Nutr 1979 32 540 549 420146
Matkovic V Fontana D Tominac C Goel P Chestnut CH Factors that influence peak bone mass formation: a study of calcium balance and the inheritance of bone mass in adolescent females Am J Clin Nutr 1990 52 878 888 2239765
Carrie-Fassler AL Bonjour JP Osteoporosis as a pediatric problem Pediatr Clin N Am 1995 42 811 824
Smith DM Nance WE Kange KW Christian JC Johnston CC Genetic factors in determining bone mass J Clin Investig 1973 52 2800 2808 4795916
Slemenda CW Christian JC Williams CJ Norton JA Johnston CC Genetic determinants of bone mass in adult women: a re-evaluation of he twin model and the potential importance of gene interaction on heritability estimates J Bone Miner Res 1991 6 561 567 1887818
Pocock NA Eisman JA Hopper JL Yeates MG Sambrook PN Eben S Genetic determinants of bone mass in adults: a twin study J Clin Investig 1987 80 706 710 3624485
Kelly PJ Eisman JA Sambrook PN Interaction of genetic and environmental influences on peak bone density Osteoporos Int 1990 1 56 60 2133642 10.1007/BF01880417
Heaney RP Abrams S Dawson-Hughes B Looker A Marcus R Matkovic V Weaver C Peak bone mass Osteoporos Int 2000 11 985 1009 11256898 10.1007/s001980070020
Lloyd T Andon MB Rollings N Martel JK Landis R Demers LM Eggli DF Kieselhorst K Kulin HE Calcium supplementation and bone mineral density in adolescent girls JAMA 1993 270 841 844 8340983 10.1001/jama.270.7.841
Johnston CC Miller JZ Slemenda CW Reister TK Hui S Christian JC Peacock M Calcium supplementation and increases in bone mineral density in children N Engl J Med 1992 327 82 87 1603140
Chan GM Hoffman K McMurray M Effects of dairy products on bone and body composition in pubertal girls J Pediatr 1995 126 551 556 7699532
Bonjour J Carrie A Ferrari S Clavien H Slosman D Theintz G Rizzoli R Calcium-enriched foods and bone mass growth in prepubertal girls: A randomized, double-blind, placebo-controlled trial J Clin Investig 1997 99 1287 1294 9077538
Cadogan J Eastell R Jones N Barker ME Milk intake and bone mineral acquisition in adolescent girls: randomized, controlled intervention trial BMJ 1997 315 1255 1260 9390050
French SA Fulkerson JA Story M Increasing weight-bearing physical activity and calcium intake for bone mass growth in children and adolescents: A review of intervention trials Prev Med 2000 31 722 731 11133340 10.1006/pmed.2000.0758
Snow-Harter C Bouxsein ML Lewis BT Carter DR Marcus R Effects of resistance and endurance exercise on bone mineral status of young women: A randomized exercise intervention trial J Bone Miner Res 1992 7 761 769 1642145
Morris FL Naughton GA Gibbs JL Carlson JS Wark JD Prospective ten-month exercise intervention in premenarchal girls: Positive effects on bone and lean mass J Bone Miner Res 1997 12 1453 1462 9286762
MacKelvie LK Khan KM McKay HA Is there a critical period for bone response to weight-bearing exercise in children and adolescents? A systematic review Br J Sports Med 2002 36 250 257 12145113 10.1136/bjsm.36.4.250
Iuliano-Burns S Saxon L Naughton G Gibbons K Bass SL Regional specificity of exercise and calcium during skeletal growth in girls: A randomized controlled trial Bone Miner Res 2003 18 156 162
Bandura A Social foundations of thought and action 1986 Englewood Cliffs, NJ: Prentice-Hall
Bass S Pearce G Bradney M Hendrich E Delmas PD Harding A Seeman E Exercise before puberty may confer residual benefits in bone density in adulthood: Studies in active prepubertal and retired female gymnasts J Bone Miner Res 1998 13 500 507 9525351
Bradney M Pearce G Naughton G Sullivan C Bass S Beck T Carlson J Seeman E Moderate exercise during growth in prepubertal boys: Changes in bone mass, size, volumetric density, and bone strength: A controlled prospective study J Bone Miner Res 1998 13 1814 21 9844098
Nutrition Coordinating Center, University of Minnesota, Minneapolis MN, USA (November, 1999). Nutrition Data System for Research, software version 4.02. Food and Nutrient Database 30. NDS-R Software, 1998–1999. Regents of the University of Minnesota. Nutrition Coordinating Center. All rights reserved
Block GA A review of validations of dietary assessment methods Am J Epidemiol 1982 115 492 505 7041631
Crawford PB Obarzanek E Morrison J Sabry ZI Comparative advantage of 3-day food records over 24-hour recall and 5 – day food frequency validated by observation of 9- and 10-year old girls J Amer Diet Assoc 1994 94 626 630 8195550 10.1016/0002-8223(94)90158-9
Sallis JF Strikmiller PK Harsha DW Feldman HA Ehlinger S Stone EJ Williston J Woods S Validation of interviewer- and self-administered physical activity checklists for fifth grade students Med Sci Sports Exerc 1996 28 840 851 8832538
Ainsworth BE Haskell WL Leon AS Jacobs DR Montoye HJ Sallis JF Paffenbarger RS Classification of energy costs of human physical activities Med Sci Sports Exerc 1993 25 71 80 8292105
Sallis JF Prochaska JJ Taylor WC A review of correlates of physical activity of children and adolescents Med Sci Sports Exerc 2000 32 963 975 10795788 10.1097/00005768-200005000-00014
DeRidder CM Kemper HCG Bertens MJM van Gameren ACL Ras E Voogd J Delemarre van der Waal HA Concurrent validity of a weight-bearing activity questionnaire in prepubertal and pubertal girls and boys Ann Hum Biol 2002 29 237 246 12031134 10.1080/03014460110075701
Turner CH Robling AG Designing exercise regimens to increase bone strength Exer and Sport Sci Rev 2003 31 45 50 10.1097/00003677-200301000-00009
Austen S Saxon L Daly R Bass SL Background loading history influences the osteogenic response to a exercise program in young boys: a randomized controlled trial unpublished data
Lohman TG Roche AF Martorell R Anthropometric Standardization Reference Manual 1988 Human Kinetics Books: Champaign, IL
Tanner JM Growth at adolescence 1962 2 Oxford: Blackwell Scientific Publications
Van Wieringen JC Wafelbakker F Verbrugge HP Haas JH Growth Diagrams, 1965 Netherlands 1971 Wolters-Noordhoff Publishing: Groningen, The Netherlands
Roche AF Wellens R Attie KM Siervogel RM The timing of sexual maturation in a group of US white youths J Pediat Endocrinol & Metab 1995 8 11 18
Petersen AC Crockett L Richards M Boxer A A self-report measure of pubertal status: Reliability, validity, and initial norms J Youth Adoles 1988 17 177 133
Morris NM Udry JR Validation of a self-administered instrument to assess stage of pubertal development J Youth Adoles 1980 9 271 280
SAS/STAT Release 82 2001 Cary, NC: SAS Institute, Inc
Murray D 1998 New York: Oxford University Press 9 15
Luepker RV Perry CL McKinlay SM Nader PR Parcel GS Stone EJ Webber LS Outcomes of a field trial to improve children's dietary patterns and physical activity: The Child and Adolescent Trial for Cardiovascular Health (CATCH) JAMA 1995 275 768 776 8598593 10.1001/jama.275.10.768
Fuchs RK Bauer JJ Snow CM Jumping improves hip and lumbar spine bone mass in prepubescent children: A randomized controlled trial J Bone Miner Res 2001 16 148 156 11149479
Blimkie CJR Rice S Webber CE Martin J Levy D Gordon CL Effects of resistance training on bone mineral content and density in adolescent females Can J Physiol Pharmacol 1996 74 1025 1033 8960394 10.1139/cjpp-74-9-1025
|
16029507
|
PMC1183238
|
CC BY
|
2021-01-04 16:38:04
|
no
|
Int J Behav Nutr Phys Act. 2005 Jul 19; 2:8
|
utf-8
|
Int J Behav Nutr Phys Act
| 2,005 |
10.1186/1479-5868-2-8
|
oa_comm
|
==== Front
J Inflamm (Lond)Journal of Inflammation (London, England)1476-9255BioMed Central London 1476-9255-2-71599626910.1186/1476-9255-2-7ResearchColocalization of endogenous TNF with a functional intracellular splice form of human TNF receptor type 2 Scherübl Christoph [email protected] Wulf [email protected]ütze Stephan [email protected] Thomas [email protected]ännel Daniela N [email protected] University of Regensburg, Institute of Immunology2 Institute of Medical Microbiology and Hygiene, D-93042 Regensburg3 University Hospital of Schleswig-Holstein Campus Kiel, Institute of Immunology, D-24105 Kiel, Germany2005 4 7 2005 2 7 7 31 3 2005 4 7 2005 Copyright © 2005 Scherübl et al; licensee BioMed Central Ltd.2005Scherübl 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
Tumor necrosis factor (TNF) is a pleiotropic cytokine involved in a broad spectrum of inflammatory and immune responses including proliferation, differentiation, and cell death. The biological effects of TNF are mediated via two cell surface TNF receptors: p55TNFR (TNFR1; CD120a) and p75TNFR (TNFR2; CD120b). Soluble forms of these two receptors consisting of the extracellular domains are proteolytically cleaved from the membrane and act as inhibitors. A novel p75TNFR isoform generated by the use of an additional transcriptional start site has been described and was termed hicp75TNFR. We focused on the characterization of this new isoform as this protein may be involved in chronic inflammatory processes.
Methods
Cell lines were retroviraly transduced with hp75TNFR isoforms. Subcellular localization and colocalization studies with TNF were performed using fluorescence microscopy including exhaustive photon reassignment software, flow cytometry, and receptosome isolation by magnetic means. Biochemical properties of the hicp75TNFR were determined by affinity chromatography, ELISA, and western blot techniques.
Results
We describe the localization and activation of a differentially spliced and mainly intracellularly expressed isoform of human p75TNFR, termed hicp75TNFR. Expression studies with hicp75TNFR cDNA in different cell types showed the resulting protein mostly retained in the trans-Golgi network and in endosomes and colocalizes with endogenous TNF. Surface expressed hicp75TNFR behaves like hp75TNFR demonstrating susceptibility for TACE-induced shedding and NFκB activation after TNF binding.
Conclusion
Our data demonstrate that intracellular hicp75TNFR is not accessible for exogenously provided TNF but colocalizes with endogenously produced TNF. These findings suggest a possible intracellular activation mechanism of hicp75TNFR by endogenous TNF. Subsequent NFκB activation might induce anti-apoptotic mechanisms to protect TNF-producing cells from cytotoxic effects of TNF. In addition, the intracellular and not TACE-accessible splice form of the hp75TNFR could serve as a pool of preformed, functional hp75TNFR.
==== Body
Background
TNF is a pleiotropic cytokine involved in a broad spectrum of inflammatory and immune responses including proliferation and cytotoxicity in a variety of different cell types [1]. Two distinct receptor molecules with an apparent molecular mass of 55 kDa (p55TNFR, TNFR type 1) and 75 kDa (p75TNFR, TNFR type 2) have been identified and their corresponding cDNAs cloned [2-5]. The p55TNFR is expressed rather constitutively on a broad spectrum of different cell types and has been shown to mediate most of the commonly known biological effects of TNF [6,7]. In contrast, expression of the p75TNFR seems to be modulated by various stimuli. However, there are only a few cellular responses that can be attributed exclusively to signalling via the p75TNFR, e.g. proliferation of NK cells [8], B cells [9], thymocytes [10], and mature T cells [11], and GM-CSF secretion of T lymphocytes [12]. Moreover, the p75TNFR has been shown to be preferentially activated by membrane-bound TNF [13]. Although the intracellular domains of the two TNFR show only little similarity they share activities like NFκB activation. While p55TNFR is capable of mediating these effects when expressed at physiologically relevant levels, induction of NFκB via the p75TNFR alone was observed only in cells overexpressing this receptor subtype [14,15].
The extracellular domain of both TNFR is suceptible to proteolytic cleavage. Agents like the natural ligand TNF, LPS, anti CD3- antibodies, and other stimuli induce a rapid receptor shedding in several cell types including macrophages, T- cells, and granulocytes [16-19].
High levels of soluble p75TNFR are found in sera of patients suffering from cancer [20], HIV [21], sepsis [22], and several autoimmune diseases like rheumatoid arthritis [23] and systemic lupus erythematodes [24]. Expression of a secreted soluble p75TNFR isoform, generated by differential splicing, was recently described to be elevated in rheumatoid arthritis [25].
A novel p75TNFR isoform generated by the use of an additional transcriptional start site has been described and was termed hicp75TNFR [26]. Exon 1 that contributes to the signal peptide in human p75TNFR is replaced by Exon1a consisting of an Alu element which was exonized during evolution in both mouse and human[27]. Several cell lines e.g. activated macrophages express hicp75TNFR in parallel to hp75TNFR [26] and hicp75TNFR mRNA upregulation was observed in mouse livers after injection of LPS in mice sensitized with D-GalN (unpublished observation). While the relevance of soluble TNFR as inhibitory molecules is generally accepted the function of an intracellular TNFR in inflammatory processes remains elusive. In this study we determined the localization of hicp75TNFR and tested possible ways of activation by exogenous and endogenous TNF.
Methods
Cell culture and reagents
HEK 293 cells and NIH 3T3 cells were maintained in Dulbeccos's Mod Eagle Medium (Invitrogen, Karlsruhe, Germany) supplemented with 10% heat- inactivated fetal calf serum (PAN Biotech GmbH, Aidenbach, Germany) and 50 μg/ml gentamycin (PAA Laboratories, Linz, Austria). p55TNFR and p75TNFR double-deficient fibroblasts (TNFR1/2) were generated in our lab by simian virus 40 large T-immortalization of murine fibroblasts from TNFR1 and TNFR2 double knock-out mice [28]. L929 cells and TNFR1/2 knock-out fibroblasts were grown in RPMI 1640 medium (Sigma-Aldrich Chemie GmbH, Deisenhofen, Germany) supplemented with 10% heat- inactivated fetal calf serum and 50 μg/ml gentamycin. The human p75TNFR-specific monoclonal mouse antibody 80M2 and rabbit serum 80M [29] were kindly provided by P. Scheurich (University of Stuttgart, Germany). The mouse monoclonal anti-myc antibody (9E10) was purchased from Invitrogen (Karlsruhe, Germany). Polyclonal rabbit IgG antibodies anti-human TNF were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Secondary antibodies for immunostaining: rabbit anti-mouse FITC was from DakoCytomation GmbH (Hamburg, Germany), and goat anti-rabbit TRITC from Sigma-Aldrich.
Transfection and transduction
NIH 3T3 cells were transiently transfected with SuperFect Transfection Reagent (Quiagen, Hilden, Germany) according to the manufacturers instruction. In order to produce retrovirus containing supernatant HEK 293 cells were transiently cotransfected with the packaging construct pCl-10 A1 [30] and the retroviral vector pQCXIP (BD Biosciences Clontech, CA, USA) containing the gene of interest using the HEPES-buffered saline calcium phosphate method [31]. Medium was replaced 5 h post transfection. Supernatants were collected after 2 days and filtered through a low-protein binding 0.45 μm pore-size filter (Acrodisc Syringe filter, PALL Corporation, MI, USA) in the presence of 8 μg/ml polybrene (Sigma-Aldrich). Target cells were infected during 2 days by changing the virus containing medium in 12 h steps and afterwards selected for positivity either by FACS-sorting in the case of fluorescent cells or by puromycin (Sigma- Aldrich) selection.
Fluorescence microscopy
Transduced NIH 3T3 cells were seeded in Lab- Tek II Chamber Slide systems (Nunc GmbH & Co. KG, Wiesbaden, Germany). The following day they were fixed 15 min with 4% paraformaldehyde in phosphate-buffered saline and permeabilized with 0.1% Triton X-100 for 5 min. After blocking with 1% bovine serum albumin in phosphate-buffered saline primary antibody (2 μg/ml) was added followed by a FITC or TRITC labelled secondary antibody. Before mounting the coverslide with MoBiGLOW Mounting Medium (MoBiTec GmbH, Göttingen, Germany) nuclei were stained with Dapi (Sigma-Aldrich). Staining of the different compartments and nuclei was performed on living cells by using Hoechst 33342, ER-Tracker Blue-White DPX, GolgiTracker BODIPY TR C5-ceramide, MitoTracker Red CMXRos, LysoTracker Red DND-99 (all from Molecular Probes, Eugene, OR, USA) according to the manufacturers instruction. The vector ENDO-eCFP was purchased from BD Biosciences Clontech (CA, USA). After washing with phosphate-buffered saline cells were fixed with 4% paraformaldehyde in phosphate-buffered saline at 4°C for 5 min followed by 10 min at room temperature. Coverslides were mounted with MobiGLOW Mounting. Fluorescent optical sections for each color were obtained with a conventional Zeiss Axiovert microscope equipped with a piezoelectric z-axis focus device (Physik Instrumente, Waldbronn, Germany). Images were taken with a charge-coupled device (CCD) camera (4096 levels of gray, -15°C peltier cooled, Princeton Instruments, Trenton, USA) and processed by MetaMorph software (Universal Imaging Corp., West Chester, USA). The light haze contributed by fluorescent-labelled structures located above and below the plane of optimal focus was mathematically reassigned to its proper place of origin (EPR, Exhaustive Photon Reassignment software Scanalytics, Massachusetts, USA) after accurate characterization of the blurring function of optical system (point spread function, PSF). During restoration, EPR used the PSF image data to refocus light and haze in the raw specimen image.
ELISA
To detect soluble extracellular domains of either hicp75TNFR or hp75TNFR the human p75TNFR (80 kDa) Module Set (Bender MedSystems, Vienna, Austria), and human sTNFRII/TNFRSF1B Duo Set ELISA development (R&D Systems, MN, USA) were used. Cells (4 × 104) were seeded in triplicates into 48-wells culture plates, and incubated either with hTNF, medium alone, or the TACE-inhibitor TAPI-0 (Biomol, Hamburg, Germany) in concentrations as indicated. After 24 h supernatants were collected and tested according to the manufacturers instructions. The OD of ABTS was determined at 405 nm.
Flow cytometry
Expression of transduced hp75TNFR isoforms on the cell surface was detected by flow cytometry on a FACStar Plus (Becton Dickinson, San Jose, CA, USA) using the PE-coupled specific rat anti-human p75TNFR monoclonal antibody and PE-labelled rat IgG2b (both BD, Heidelberg, Germany) as isotype control. For FACS analysis cells (1 × 106/tube) were blocked with PBS containing 10% heat inactivated FCS for 30 min on ice and then incubated with 20 μl of antibody solution on ice for 30 min. The YFP-tagged fusion proteins were directly detectable in FL-1 without any additional staining step. YFP-positive cells were sorted using a FACStar Plus. For intracellular staining FIX&PERM (Caltag, Hamburg, Germany) was used.
Western Blot and Immunoprecipitation
Cells (4 × 106) were seeded in 10 cm dishes and grown overnight. After 24 h cells were washed with ice-cold PBS and lysed in 1 ml of buffer (150 mM NaCl, 50 mM Tris· HCl, pH 7.4,1 mM EDTA, 1% Triton X-100, NP-40 1%, Na-deoxycholate 0,25%) containing a mixture of protease inhibitors (Complete™ EDTA-free tablets, Roche, Mannheim, Germany). After centrifugation lysates were precleared for 4 h in 20 μl of protein G-Sepharose (Amersham Biosciences, Uppsala, Sweden), and immunoprecipitated with 80M2 (10 μg/ml) a mouse monoclonal antibody against the extracellular domain of human p75TNFR [29] and 20 μl of protein G-Sepharose for 16 h at 4 C. Pellets were washed three times in PBS, resolved on 8% SDS-PAGE under reducing conditions, and transferred to PVDF membranes. These membranes were blocked in 1% non-fat milk powder and detected with a rabbit polyclonal antibody (0.1 μg/ml) H-202 (Santa-Cruz Biotechnology, Inc., CA, USA) binding to the intracellular domain of hp75TNFR isoforms, to specifically detect the full-length protein. As secondary antibody a goat anti-rabbit IgG-HRP (Dilution 1:2000, Sigma- Aldrich) was used. The blots were then incubated with HRP substrate (enhanced chemiluminescence substrate NOWA solution A and B; MoBiTec GmbH, Germany) and developed by exposure to film (Hyperfilm; Amersham Biosciences).
Cloning
PCR amplification was performed using sense primers for hicp75TNFR-myc (5'-AAAGGATCCCCCATGGCGAAACCCCTC-3') and for hp75TNFR-myc (5'-AAAGGATCCCCCATGGCGCCCGTCGCCGTC-3') and antisense primers for both (5'GGGCTCGAGTCACAGATCCTCTTCTGAGAT-3'). The template in each case was the myc-tagged cDNA in pcDNA 3.1 hygro (Invitrogen, Karlsruhe, Germany). The resulting PCR products were cloned using BamHI/xhoI into a modified proviral vector, pQCXIP (BD Biosciences Clontech, CA, USA) containing an additional xhoI site on the 3' end of the multiple cloning site. To create eYFP-tagged fusion proteins the whole MCS of pQCXIP (BD Biosciences Clontech, CA, USA) was substituted by the MCS and eYFP coding fragment from pEYFP- N1 (BD Biosciences Clontech, CA, USA). In this modified vector both myc-tagged cDNA's were cloned using EcoRI/BamHI in frame with eYFP using sense primers for hicp75TNFR-myc (5'-CCGAATTCCCAGCCATGGCGAAACCCCTC-3') and for hp75TNFR-myc (5'-CCCAAGCTTGAATTCCCAGCCATGGCGCCCGTCGCCGTC-3') and antisense primers for both (5'CGGGATCCCGCAGATCCTCTTCTGAGATG-3'). All expression constructs have been verified by sequencing.
Isolation of magnetically labelled TNF receptosomes
To isolate TNF plasma membrane receptors after ligand binding and internalization biotinylated TNF (Fluorokine-Kit, R&D Systems, Wiesbaden, Germany) and magnetically labelled 50 nm MACS Streptavidin Microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) were used and the receptosomes isolated by magnetic means after different time points as described recently [32]. Cells were incubated in a total volume of 250 μl cold Dulbecco's Modified Eagles Medium (DMEM) supplemented with 25 Mm HEPES (Invitrogen, Karlsruhe, Germany) with 100 μl (400 ng) of biotinylated TNF for 1 h on ice. Thereafter, 200 μl of the MACS Streptavidin Microbeads solution was added and cells were incubated for 1 h on ice. TNF receptor clustering and formation of magnetized TNF receptosomes was achieved by incubation at 37°C for different times. The labeled cells were mechanically homogenized using steel beads in a 0.25M sucrose buffer, supplemented with 0.015M HEPES, 100 mg/L MgCl2, pH 7.4 and the Protease Inhibitors Set from Roche Diagnostics (Mannheim, Germany) at 4°C. A postnuclear supernatant was submitted to magnetic separation of TNF receptosomes in a high-gradient magnetic field generated in a custom-built free-flow magnetic chamber (German patent held by S. Schütze).
Characterization of TNF receptosome-associated proteins
Receptosome proteins were separated by SDS-PAGE and analyzed by Western blotting for signature proteins of endocytosis and vesicular trafficking by antibodies against clathrin, (Transduction Lab., Lexington, UK), Rab5, Vti1b (StressGene Biotec. Corp. Victoria, Canada), and cathepsin D (Calbiochem-Novabiochem GmbH, Germany) as described [32].
Affinity chromatography
Purified rhTNF (BASF, Ludwigshafen, Germany) was coupled to CNBr-activated Sepharose 4B according to the protocol provided by Amersham Biosciences (capacity 16–32 mg/g) with the following modifications: The gel was swollen with 1 mM of HCl and washed three times for 5 min each, using a total of 200 ml of 1 mM HCl/g of gel. rhTNF was dissolved at 7.5 mg/3 ml of phosphate-buffered saline and dialyzed three times against 500 ml of 0.1M NaHCO%3, pH 8.3, 0.5M NaCl. The volume of the ligand was made up to 5 ml with the same buffer and mixed with the washed gel (2.5 mg of rhTNF/ml of swollen gel slurry) by rotating gently overnight at 4°C. Excess ligand was washed away with the same buffer, and the remaining active groups were blocked with 0.1M Tris-HCl, pH 8.0 for 2 h at room temperature. The gel was then washed with 3 cycles of alternating pH, using 0.1M sodium acetate buffer, pH 4.0, and 0.1M Tris-HCl, pH 8.0, each containing 0.5M NaCl. The conjugate was stored in PBS at 4°C. Clarified supernatant from transduced TNFR1/2 knock-out fibroblasts was precipitated with ammonium sulfate. The pellet was resuspendet in PBS, dialysed against PBS and loaded onto a rhTNF affinity column by using FPLC (Bio-Rad Laboratories, Inc., CA, USA). After washing with PBS, the bound material was eluted with 100 mM glycine, 100 mM NaCl, pH 2.6, into microcentrifuge tubes containing 1M Tris-HCl, pH 7.5. Fractions were analyzed by 8% SDS-PAGE and following Coomassie Blue staining or Western blotting.
Determination of cytotoxic activity of TNF on L929 cells
Cells (L929) were seeded in 96-well microtiter plates at 2.5 × 104 cells/well. After 24 h serial dilutions of rhTNF were added in the presence of actinomycin D (2 μg/ml). After 24 h, cell viability was assessed by adding MTT (Sigma- Aldrich) for 4 h. Cells were lysed with SDS and OD were determined at 450 nm.
Results
Localization of tagged human p75TNFR isoforms
To obtain a detailed picture of the subcellular localization of the hicp75TNFR we used YFP-tagged receptor constructs and compared the staining with compartment-specific fluorescent trackers. As shown in Fig. 1 the big barrel structure of the YFP protein is not altering the staining pattern of the hp75TNFR protein when compared to the relatively small myc-tag. While the hp75TNFR stained in the manner of a typical membrane-localized protein, the hicp75TNFR exhibited a punctuated, vesicle-like pattern localized perinuclearly and throughout the cytoplasm. In addition, we also transfected cells with the corresponding cDNAs both transiently and stably to rule out potential epiphenomena created by viral transduction. The data obtained by transfection were in line with those obtained by transduction (data not shown).
Figure 1 Localization of tagged hp75TNFR and hicp75TNFR molecules in transduced NIH 3T3 cells. Cells transduced either with myc-tagged (A;C) or YFP-tagged (B;D) hp75TNFR (A;B) or hicp75TNFR (C;D) were analyzed microscopically. For detection of YFP the cells were fixed and analyzed. For detection of the myc-tag cells were fixed and permeabilized followed by incubation with a monoclonal mouse anti-human c-myc (9E10) and a secondary anti-mouse FITC antibody. While hp75TNFR staining is mainly found on the plasma membrane, hicp75TNFR staining is perinuclear, punctuated, and distributed throughout the cytoplasm. Nuclei: Dapi (A;C), Hoechst 33342 (B;D).
Subcellular localization of human icp75TNFR
When TNFR localization was compared to different cellular compartments none of the two hp75TNFR variants co-localized with the ER (Fig. 2A; 2B), mitochondria (Fig. 2E; 2F), or lysosomes (Fig. 2G; 2H). Both receptor forms colocalized with the Golgi apparatus (Fig. 2C; 2D). A clear discrimination of colocalization with perinuclear budded endosomes resembling the trans-golgi network (TGN) was observed when hp75TNFR staining was compared to hicp75TNFR staining. Whereas the hp75TNFR was quickly guided through this compartment (Fig. 2I) it seemed that most of the hicp75TNFR-YFP molecules were stored in endosomal structures as documented by the strong colocalization of hicp75TNFR with ENDO-eCFP (Fig. 2J).
Figure 2 Localization of hp75TNFR isoforms in transduced NIH 3T3 cells. Cells transduced either with YFP-tagged hp75TNFR (A; C; E; G; I) or YFP-tagged hicp75TNFR (B; D; F; H; J) were costained with ER-Tracker (A; B), Golgi-Tracker (C; D), Mito-Tracker (E; F), Lyso-Tracker (G; H) and ENDO-eCFP (I; J). While hp75TNFR staining is mainly found on the plasma membrane and in the Golgi apparatus, hicp75TNFR shows no plasma membrane staining and colocalizes with the Golgi apparatus and endosomal compartments of the trans-Golgi network.
Cell surface expression of human p75TNFR isoforms
Although transfected hicp75TNFR was not detectable microscopically, we used flow cytometry to determine whether hicp75TNFR becomes detectable on the cell membrane after transduction.
Fig. 3A shows that both L929 cell lines either transduced with hp75TNFR or hicp75TNFR expressed the same number of the respective hp75TNFR isoform. After fixation and permeabilization a specific PE-labelled antibody against an epitope present on both hp75TNFR and hicp75TNFR stained the transduced L929 cell lines with equal intensities (Fig. 3A) demonstrating that the used antibody has the same affinity to both isoforms and that the cells express a comparable amount of this epitope. Without cell permeabilization hp75TNFR isoforms exposed on the outer cell membrane became detectable which is in line with the microscopic observations (Fig. 2). Surprisingly, hicp75TNFR molecules were also stained on the cell membrane, however to a lower extent than hp75TNFR molecules (Fig. 3B).
Figure 3 Cell surface expression of hp75TNFR isoforms on transduced L929 cells. Expression of hp75TNFR isoforms on L929 cells either transduced with control vector (shaded peak), hp75TNFR (black line) or hicp75TNFR (gray line) was analyzed by flow cytometry with (A) or without (B) permeabilization. After permeabilization the same number of epitopes were accessible in both cell lines. Without permeabilization cells expressing hp75TNFR present more epitopes on the cell surface.
Soluble ectodomain of human p75TNFR isoforms
Soluble forms of both TNF receptors are described, consisting of the shed extracellular domains of the p55TNFR or p75TNFR, respectively. The release can be enhanced by stimulation with TNF, LPS, or phorbol ester and can be inhibited by TAPI (TNF-α processing inhibitor). This event mainly takes place on the cell surface and is caused by metalloproteases [33].
We tested transduced L929 cells sorted for comperable hp75TNFR-YFP and hicp75TNFR-YFP expression concerning the release of soluble receptor by ELISA (Fig. 4A). L929 cells transduced with hp75TNFR-YFP as well as hicp75TNFR-YFP-transduced L929 cells released soluble hp75TNFR into the supernatant (Fig. 4A). More soluble hp75TNFR was measured than soluble hicp75TNFR. In the supernatant from control L929 cells transduced with YFP no soluble TNFR was found. This data indicate that hicp75TNFR is less affected by shedding. In both cell lines shedding was TACE-dependent as indicated by its inducibility with rhTNF and its attenuation with the specific TACE inhibitor TAPI. TNF was increasing the shedding by 25% in the hicp75TNFR tranductants and by 30% in the hp75TNFR transductants. TAPI reduced rhTNF induced shedding to 58% in the hicp75TNFR transductants and to 60% in the hp75TNFR transductants, showing that activitiy of TACE was equal in both cell lines and therefore not the reason for the lower soluble hicp75TNFR in the supernatant.
Figure 4 Shedded ectodomain of hp75TNFR isoforms. (A)The extracellular domain of both hp75TNFR isoforms is released constitutively (black bars) into the supernatant of L929 cell either transduced with hp75TNFR or hicp75TNFR, respectively. Shedding is increased by rhTNF (6 ng/ml; light gray bars) and attenuated by TAPI (100 nM, dark gray bars). (B)The extracellular domain of hicp75TNFR is bioactive as shown by neutralization of rhTNF in a TNF cytotoxicity assay on L929 cells. Supernatant from L929 cells either transduced with icp75TNFR-YFP (●), or YFP alone (○) were tested. DMEM (▼) served as control medium. All values are given as mean ± S.D. of triplicate cultures. Four independent experiments gave similar results.
When the soluble hp75TNFR from the supernatant of hicp75TNFR-YFP transduced L929 cells was tested in a TNF cytotoxicity assay on L929 cells (Fig. 4B) the inhibitory activity of the shedded extracellular domain of the hicp75TNFR was demonstrated. This indicated that the ectodomain of hicp75TNFR is biologically active in binding and neutralizing TNF.
Biochemical characterization of human icp75TNFR
The Exon 1a of hicp75TNFR does not encode a typical signal peptide and is, therefore, not cleaved off in the biosynthetic process by signal peptidases in contrast to the sequence encoded by Exon 1 of hp75TNFR. To test whether this results in a different apparent molecular weight, we isolated full length protein from total lysates of NIH 3T3 cells either transduced with hp75TNFR or hicp75TNFR, respectively, by immunoprecipitation with a monoclonal mouse antibody raised against the extracellular domain of hp75TNFR (80M2).
Immunodetection was done with a polyclonal rabbit IgG fraction (sc-7862) detecting the intracellular domain of hp75TNFR. The full-length mature protein resulting from hicp75TNFR cDNA has 85 kDa, which is about 10 kDa more than the apparent molecular mass of hp75TNFR (Fig. 5A). In addition, the hp75TNFR also appears as a faint band with about 50 kDa as already described earlier as a possibly differently glycosylated hp75TNFR species [34,35]. A prominent double band was also stained in the case of hicp75TNFR at about 50 kDa possibly indicating different molecular masses for icp75TNFR molecules resulting from potential O-glycosylation sites at threonin 7 and serin 11 of the hicp75TNFR-specific exon 1a (unpublished observation).
Figure 5 Biochemical characterization of hicp75TNFR.(A) The hp75TNFR from NIH3T3 cells either transduced with hp75TNFR or hicp75TNFR, respectively, were immunoprecipitated with monoclonal antibodies against the extracellular domain of hp75TNFR (80M2) and stained after blotting with antibodies to the intracellular domain on hp75TNFR (sc-7862). (B) Affinity purified soluble hicp75TNFR from supernatant of transduced TNFR1/2 knock-out fibroblasts was stained after blotting with a rabbit serum against the extracellular domain of hp75TNFR (80M).
To determine whether the difference in the apparent molecular weight is due to the presence of the additional presequence in the extracellular domain, the soluble extracellular domain of hicp75TNFR was purified from cell culture supernatant by affinity chromatography. To avoid contamination with shed mouse TNFR we used TNFR1/2 knock-out fibroblasts transduced with hicp75TNFR. The purified soluble hicp75TNFR was stained with anti-hp75TNFR antiserum (80M) and showed a single band at about 50 kDa (Fig. 5B). Since the soluble hp75TNFR has been published with an apparent molecular weight of 40kDa [36,37] these data indicate that the higher molecular mass of hicp75TNFR might result from the additional 24 amino acids at the N-terminal end encoded by Exon1a and/or different glycosylation.
Activation of human icp75TNFR
Internalization of p55TNFR after binding of exogenous TNF has been published [38]. To test whether exogenous TNF interacts with hicp75TNFR a new method recently described by Schneider-Brachert et al. was used [32]. We isolated receptosomes at different time points in a magnetic field and followed maturation of the resulting endosomes by testing for different marker proteins recruited to the complex (Fig. 6A).
Figure 6 Colocalization of hicp75TNFR with endogenous or exogenous TNF. (A) L929 cells were transduced with hicp75TNFR-YFP and exposed to biotinylated TNF. Receptosomes were isolated after different times, subjected to SDS-PAGE and analyzed by immunostaining of blotted proteins. (B) L929 cells expressing hTNF were transduced with hicp75TNFR and stained with rabbit anti-human TNF antibodies and mouse monoclonal antibody anti-human p75TNFR (80M2).
Over time after binding, clathrin as a marker for endocytosis is decreasing as well as the early endosomal marker Rab4. After 30 minutes the late endosomes fuse with vesicles from the TGN, as documented by an increasing amount of Vti1. Even though this could be the point in time when exogenous TNF could be passed from the internalized TNF-p55TNFR complex to the hicp75TNFR, no hicp75TNFR was isolated in receptosomes after 30 minutes indicating that no intracellular ligand passing is occurring. After 60 min the receptosomes end up in lysosomes indicated by Cathepsin D staining.
To evaluate a possible intracellular activation of hicp75TNFR with endogenously produced TNF L929 stably transfected with full-length, transmembrane TNF were transduced with hicp75TNFR (Fig. 6B). After double staining colocalization of human endogenous TNF and hicp75TNFR was observed in intracellular perinuclear compartments indicating that endogenous TNF could possibly activate the hicp75TNFR.
Discussion
TNF exerts pleiotropic biological activities affecting proliferation, differentiation, or functions in a wide variety of cell types by interacting with its two distinct receptors p55TNFR and p75TNFR [17,18]. Soluble TNF receptors, generated either by proteolytic cleavage [16-19] or differential splicing [25] contribute to the balance of TNF-mediated effects by neutralization of the ligand. The biological function of a mainly intracellularly expressed isoform of hp75TNFR was not clear. Since hicp75TNFR is lacking a typical leader sequence at the N-terminus we used YFP-tagged receptor constructs to investigate whether the protein is directed to a specific intracellular compartment. Localization of the YFP-tagged hicp75TNFR was not different from the corresponding myc-tagged hicp75TNFR. In this way it was not necessary to perform permeabilization and incubation steps, that could probably alter the 3D structure of the cells in colocalization studies with compartment-specific dyes. In case of the hp75TNFR-YFP the expected pattern of a typical plasma membrane-bound receptor was seen, comparable to native p75TNFR expression in human umbilical vein endothelial cells [39].
The hicp75TNFR-YFP protein was not retained in the ER. Also, this receptor is not a mitochondrial protein as has been described by crossreaction with an anti-hp75TNFR antibody [40]. Exogenously added TNF internalizes after binding to the p55TNFR and is transported to the lysosomes [38,41]. The hicp75TNFR did not colocalize with lysosomes and was not found to interact with endocytosed exogenous TNF.
Besides being enriched in the final compartment, overexpressed cellular proteins are usually observed in the Golgi apparatus. The hicp75TNFR strongly colocalized with the Golgi apparatus and with budding endosomal vesicles of the trans-Golgi network (TGN) as indicated by the coexpression with a labelled endosomal marker. No hicp75TNFR staining was observed on the plasma membrane or any other intracellular compartment.
The limitations of sensitivity of fluorescent microscopy become obvious considering the intracellular but not membrane staining of p55TNFR by confocal microscopy. This TNFR is also predominantly seen in the TGN [39,42], even though it is a well characterized plasma membrane protein. Flow cytometrical analysis confirmed indeed that both isoforms of the hp75TNFR can be found on the cell surface with stronger expression of the hp75TNFR compared to the hicp75TNFR isoform. The hicp75TNFR expression reminds of the human transferrin receptor which is also synthesized without a typical leader sequence and, therefore, localized predominantly within the cell. Only a small amount is localized in the plasma membrane [43] where the protein acts as a receptor. It has been shown that a transmembrane domain is sufficient to translocate proteins into the membrane of the ER from where they travel to the plasma membrane via the secretory pathway [43,44].
As a direct consequence of the cell surface expression the hicp75TNFR becomes accessible to the TNF-α convertase (TACE/ADAM-17), a transmembrane disintegrin metalloproteinase of the ADAM family of proteases. The ADAM-17-dependency on shedding of the extracellular domain of hicp75TNFR is demonstrated by the TNF inducibility and reduced shedding in the presence of the specific TACE inhibitor TAPI. In general, lower amounts of soluble hicp75TNFR were detected in supernatants of transduced cells compared to soluble hp75TNFR. These data indicate that the hicp75TNFR molecules emerging on the cell membrane do not behave differently than the hp75TNFR molecules. The intracellular pool of hicp75TNFR does not seem to be affected by shedding that takes place on the cell membrane. Since after stimulation cells are depleted of TNFR [33] restoration of the cell surface with preformed functional TNFR is a very quick mechanism without the need of protein neo-biosynthesis. Such a theory of stored TNFR has also been discussed for intracellular p55TNFR molecules [45].
The soluble hicp75TNFR is biologically active and able to competitively bind TNF. Full-length p55TNFR is released in exosome-like vesicles as published recently [46]. We could exclude this possibility for the hicp75TNFR because of the clear TACE-dependent shedding and the apparent molecular weight of full length and affinity purified soluble hicp75TNFR. The soluble hicp75TNFR showed a clear band at ~50 kDa while full length hicp75TNFR appears at ~85 kDa, which is in both cases about 10 kDa more than the respective form of the hp75TNFR [37]. This increased molecular weight can be explained by the additional N-terminal sequence of 24 amino acids encoded by Exon 1a.
Addressing the question whether exogenous TNF could possibly activate the hicp75TNFR we used the elegant method of receptosome isolation [32]. The results show that internalized TNF was not passed over from the internalized p55TNFR to hicp75TNFR. Due to higher affinity of soluble TNF to the p55TNFR than to the p75TNFR [47] such passing over was rather unlikely. On the other hand, the affinity of hp75TNFR to membrane-bound TNF is relatively high [10]. Therefore, the second possibility of interaction of endogenous TNF with hicp75TNFR was tested using L929 cells stably expressing 26 kDa pre-TNF and transduced with hicp75TNFR. Overexpression of both ligand and receptor became necessary because TNF-producing cells lines such as macrophages could not be transduced efficiently with hicp75TNFR expression constructs. The distribution pattern of overexpressed endogenous TNF is similar to the TNF produced by macrophages upon LPS stimulation [48]. Clear colocalization of endogenous TNF and intracellularly localized hicp75TNFR was detected. Activation of hicp75TNFR by endogenous TNF in the TGN is difficult to unequivocally demonstrate due to the presence of soluble TNF and cell membrane-localized hicp75TNFR in this system. For intracellular p55TNFR it has already been shown that activation did not occur by exogenously added TNF [42]. Parallel expression of ligand and receptor in the same cell could lead to intracellular complexes that are rapidly degraded [49]. Alternatively, intracellular activation of the hicp75TNFR by endogenous TNF cannot be excluded.
Expression of TNF as well as hicp75TNFR by cells such as activated macrophages could lead to the activation of intracellular NFκB pathways and to the expression of NFκB-dependent anti-apoptotic proteins. The importance of NFκB in preventing apoptosis has clearly been demonstrated as fibroblasts of p65/RelA-deficient mice are highly susceptible for TNF induced apoptosis [50] and cells expressing a dominant negative mutant of IκB are also very sensitive to TNF-induced apoptosis [51,52]. Therefore, such an intracellular activation mechanism of hicp75TNFR by endogenous TNF with subsequent NFκB-dependent activation of anti-apoptotic mechanisms might protect TNF-producing cells from cytotoxic effects of TNF. As a result, survival of activated immune cells could thus become sustained.
Intracellular activation of the p75TNFR by endogenous TNF seems particularly interesting in the light of the recent publication by Kim and The [53] demonstrating costimulatory function of the p75TNFR activation for T cell activation. Thus, TNF not only serves as a central mediator of innate immunity by activating the p55TNFR but also by stimulation of the p75TNFR contributes to the induction of adaptive immune responses.
In conclusion, the results of this study show that the alternative splice variant of human p75TNFR is strongly retained in the TGN where it could function as a storage pool of preformed p75TNFR that is not affected by shedding. Upon emerging on the cell surface hicp75TNFR is functionally not longer distinguishable from hp75TNFR. The consequences of hicp75TNFR colocalization with endogenous TNF for hicp75TNFR signalling in TNF producing cells remains to be analyzed.
Conclusion
The results of this study show that the alternative splice variant of the human p75TNFR is strongly retained in the TGN where it colocalizes with endogenous TNF. The consequences of this colocalization for hicp75TNFR signalling in TNF producing cells remains to be analyzed. Upon emerging on the cell surface hicp75TNFR is functionally not longer distinguishable from hp75TNFR. The hicp75TNFR could function as a storage pool of preformed p75TNFR that is not affected by shedding. Potential intracellular activation of hicp75TNFR by endogenous TNF could lead to elevated NFκB levels and eventually protect TNF-producing cells from TNF-induced cell death.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CS participated in experimental design, carried out most experimental procedures and put together the manuscript. WS-B established the protocol for retroviral transduction and provided essential vectors. SS carried out the receptosome isoloation and the corresponding western blots. TH participated in experimental design, cloning and discussions. DNM participated in study design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Eva Pocsik for providing the L929 cells stably transfected with human TNF, and Peter Scheurich for the anti-human p75TNFR antibodies 80M and 80M2. This research was supported by the Deutsche Forschungsgemeinschaft (MA760/13-1). CS was an associate member of the Graduiertenkolleg GRK 760
==== Refs
Wajant H Pfizenmaier K Scheurich P Tumor necrosis factor signaling Cell Death Differ 2003 10 45 65 12655295 10.1038/sj.cdd.4401189
Lewis M Tartaglia LA Lee A Bennett GL Rice GC Wong GH Chen EY Goeddel DV Cloning and expression of cDNAs for two distinct murine tumor necrosis factor receptors demonstrate one receptor is species specific Proc Natl Acad Sci U S A 1991 88 2830 2834 1849278
Loetscher H Pan YC Lahm HW Gentz R Brockhaus M Tabuchi H Lesslauer W Molecular cloning and expression of the human 55 kd tumor necrosis factor receptor Cell 1990 61 351 359 2158862 10.1016/0092-8674(90)90815-V
Schall TJ Lewis M Koller KJ Lee A Rice GC Wong GH Gatanaga T Granger GA Lentz R Raab H . Molecular cloning and expression of a receptor for human tumor necrosis factor Cell 1990 61 361 370 2158863 10.1016/0092-8674(90)90816-W
Smith CA Davis T Anderson D Solam L Beckmann MP Jerzy R Dower SK Cosman D Goodwin RG A receptor for tumor necrosis factor defines an unusual family of cellular and viral proteins Science 1990 248 1019 1023 2160731
Smith CA Farrah T Goodwin RG The TNF receptor superfamily of cellular and viral proteins: activation, costimulation, and death Cell 1994 76 959 962 8137429 10.1016/0092-8674(94)90372-7
Vercammen D Vandenabeele P Declercq W Van de CM Grooten J Fiers W Cytotoxicity in L929 murine fibrosarcoma cells after triggering of transfected human p75 tumour necrosis factor (TNF) receptor is mediated by endogenous murine TNF Cytokine 1995 7 463 470 7578985 10.1006/cyto.1995.0063
Mason AT McVicar DW Smith CA Young HA Ware CF Ortaldo JR Regulation of NK cells through the 80-kDa TNFR (CD120b) J Leukoc Biol 1995 58 249 255 7543923
Erikstein BK Smeland EB Blomhoff HK Funderud S Prydz K Lesslauer W Espevik T Independent regulation of 55-kDa and 75-kDa tumor necrosis factor receptors during activation of human peripheral blood B lymphocytes Eur J Immunol 1991 21 1033 1037 1850359
Grell M Becke FM Wajant H Mannel DN Scheurich P TNF receptor type 2 mediates thymocyte proliferation independently of TNF receptor type 1 Eur J Immunol 1998 28 257 263 9485205 10.1002/(SICI)1521-4141(199801)28:01<257::AID-IMMU257>3.0.CO;2-G
Tartaglia LA Goeddel DV Reynolds C Figari IS Weber RF Fendly BM Palladino MAJ Stimulation of human T-cell proliferation by specific activation of the 75-kDa tumor necrosis factor receptor J Immunol 1993 151 4637 4641 8409424
Vandenabeele P Declercq W Vercammen D Van de CM Grooten J Loetscher H Brockhaus M Lesslauer W Fiers W Functional characterization of the human tumor necrosis factor receptor p75 in a transfected rat/mouse T cell hybridoma J Exp Med 1992 176 1015 1024 1328463 10.1084/jem.176.4.1015
Grell M Douni E Wajant H Lohden M Clauss M Maxeiner B Georgopoulos S Lesslauer W Kollias G Pfizenmaier K . The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor Cell 1995 83 793 802 8521496 10.1016/0092-8674(95)90192-2
Haridas V Darnay BG Natarajan K Heller R Aggarwal BB Overexpression of the p80 TNF receptor leads to TNF-dependent apoptosis, nuclear factor-kappa B activation, and c-Jun kinase activation J Immunol 1998 160 3152 3162 9531270
Heller RA Song K Fan N Chang DJ The p70 tumor necrosis factor receptor mediates cytotoxicity Cell 1992 70 47 56 1320461 10.1016/0092-8674(92)90532-H
Porteu F Nathan C Shedding of tumor necrosis factor receptors by activated human neutrophils J Exp Med 1990 172 599 607 2165128 10.1084/jem.172.2.599
Lantz M Bjornberg F Olsson I Richter J Adherence of neutrophils induces release of soluble tumor necrosis factor receptor forms J Immunol 1994 152 1362 1369 7905502
Crowe PD VanArsdale TL Goodwin RG Ware CF Specific induction of 80-kDa tumor necrosis factor receptor shedding in T lymphocytes involves the cytoplasmic domain and phosphorylation J Immunol 1993 151 6882 6890 8258697
Aderka D Sorkine P Abu-Abid S Lev D Setton A Cope AP Wallach D Klausner J Shedding kinetics of soluble tumor necrosis factor (TNF) receptors after systemic TNF leaking during isolated limb perfusion. Relevance to the pathophysiology of septic shock J Clin Invest 1998 101 650 659 9449699
Aderka D Englemann H Hornik V Skornick Y Levo Y Wallach D Kushtai G Increased serum levels of soluble receptors for tumor necrosis factor in cancer patients Cancer Res 1991 51 5602 5607 1655258
Hober D Benyoucef S Delannoy AS De Groote D Ajana F Mouton Y Wattre P High plasma level of soluble tumor necrosis factor receptor type II (sTNFRII) in asymptomatic HIV-1-infected patients Infection 1996 24 213 217 8811358 10.1007/BF01781095
Schroder J Stuber F Gallati H Schade FU Kremer B Pattern of soluble TNF receptors I and II in sepsis Infection 1995 23 143 148 7499002 10.1007/BF01793854
Cope AP Aderka D Doherty M Engelmann H Gibbons D Jones AC Brennan FM Maini RN Wallach D Feldmann M Increased levels of soluble tumor necrosis factor receptors in the sera and synovial fluid of patients with rheumatic diseases Arthritis Rheum 1992 35 1160 1169 1329774
Aderka D Wysenbeek A Engelmann H Cope AP Brennan F Molad Y Hornik V Levo Y Maini RN Feldmann M . Correlation between serum levels of soluble tumor necrosis factor receptor and disease activity in systemic lupus erythematosus Arthritis Rheum 1993 36 1111 1120 8393677
Lainez B Fernandez-Real JM Romero X Esplugues E Canete JD Ricart W Engel P Identification and characterization of a novel spliced variant that encodes human soluble tumor necrosis factor receptor 2 Int Immunol 2004 16 169 177 14688072 10.1093/intimm/dxh014
Seitz C Muller P Krieg RC Mannel DN Hehlgans T A Novel p75TNF Receptor Isoform Mediating NFk B Activation J Biol Chem 2001 276 19390 19395 11279196 10.1074/jbc.M101336200
Singer SS Mannel DN Hehlgans T Brosius J Schmitz J From "Junk" to Gene: Curriculum vitae of a Primate Receptor Isoform Gene Journal of Molecular Biology 2004 342 883 886 15328599 10.1016/j.jmb.2004.06.070
Krippner-Heidenreich A Tubing F Bryde S Willi S Zimmermann G Scheurich P Control of receptor-induced signaling complex formation by the kinetics of ligand/receptor interaction J Biol Chem 2002 277 44155 44163 12215450 10.1074/jbc.M207399200
Grell M Scheurich P Meager A Pfizenmaier K TR60 and TR80 tumor necrosis factor (TNF)-receptors can independently mediate cytolysis Lymphokine Cytokine Res 1993 12 143 148 8394147
Naviaux RK Costanzi E Haas M Verma IM The pCL vector system: rapid production of helper-free, high-titer, recombinant retroviruses J Virol 1996 70 5701 5705 8764092
Sambrook J Russel DW Molecular cloning: A laboratory manual Molecular cloning: A laboratory manual 2001 3nd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY
Schneider-Brachert W Tchikov V Neumeyer J Jakob M Winoto-Morbach S Held-Feindt J Heinrich M Merkel O Ehrenschwender M Adam D Mentlein R Kabelitz D Schutze S Compartmentalization of TNF Receptor 1 Signaling; Internalized TNF Receptosomes as Death Signaling Vesicles Immunity 2004 21 415 428 15357952 10.1016/j.immuni.2004.08.017
Pedron T Girard R Chaby R TLR4-dependent lipopolysaccharide-induced shedding of tumor necrosis factor receptors in mouse bone marrow granulocytes J Biol Chem 2003 278 20555 20564 12663667 10.1074/jbc.M203551200
Pimentel-Muinos FX Seed B Regulated commitment of TNF receptor signaling: a molecular switch for death or activation Immunity 1999 11 783 793 10626900 10.1016/S1074-7613(00)80152-1
Corti A D'Ambrosio F Marino M Merli S Cassani G Identification of differentially glycosylated forms of the soluble p75 tumor necrosis factor (TNF) receptor in human urine Eur Cytokine Netw 1995 6 29 35 7795172
Kohno T Brewer MT Baker SL Schwartz PE King MW Hale KK Squires CH Thompson RC Vannice JL A second tumor necrosis factor receptor gene product can shed a naturally occurring tumor necrosis factor inhibitor Proc Natl Acad Sci U S A 1990 87 8331 8335 2172983
Pennica D Lam VT Mize NK Weber RF Lewis M Fendly BM Lipari MT Goeddel DV Biochemical properties of the 75-kDa tumor necrosis factor receptor. Characterization of ligand binding, internalization, and receptor phosphorylation J Biol Chem 1992 267 21172 21178 1328224
Schutze S Machleidt T Adam D Schwandner R Wiegmann K Kruse ML Heinrich M Wickel M Kronke M Inhibition of receptor internalization by monodansylcadaverine selectively blocks p55 tumor necrosis factor receptor death domain signaling J Biol Chem 1999 274 10203 10212 10187805 10.1074/jbc.274.15.10203
Bradley JR Thiru S Pober JS Disparate localization of 55-kd and 75-kd tumor necrosis factor receptors in human endothelial cells Am J Pathol 1995 146 27 32 7856733
Ledgerwood EC Prins JB Bright NA Johnson DR Wolfreys K Pober JS O'Rahilly S Bradley JR Tumor necrosis factor is delivered to mitochondria where a tumor necrosis factor-binding protein is localized Lab Invest 1998 78 1583 1589 9881958
Mosselmans R Hepburn A Dumont JE Fiers W Galand P Endocytic pathway of recombinant murine tumor necrosis factor in L-929 cells J Immunol 1988 141 3096 3100 2902145
Jones SJ Ledgerwood EC Prins JB Galbraith J Johnson DR Pober JS Bradley JR TNF recruits TRADD to the plasma membrane but not the trans-Golgi network, the principal subcellular location of TNF-R1 J Immunol 1999 162 1042 1048 9916731
Prinz B Stahl U Lang C Intracellular transport of a heterologous membrane protein, the human transferrin receptor, in Saccharomyces cerevisiae Int Microbiol 2003 6 49 55 12730712
Zerial M Melancon P Schneider C Garoff H The transmembrane segment of the human transferrin receptor functions as a signal peptide EMBO J 1986 5 1543 1550 3017701
Wang J Al Lamki RS Zhang H Kirkiles-Smith N Gaeta ML Thiru S Pober JS Bradley JR Histamine antagonizes tumor necrosis factor (TNF) signaling by stimulating TNF receptor shedding from the cell surface and Golgi storage pool J Biol Chem 2003 278 21751 21760 12646554 10.1074/jbc.M212662200
Hawari FI Rouhani FN Cui X Yu ZX Buckley C Kaler M Levine SJ Release of full-length 55-kDa TNF receptor 1 in exosome-like vesicles: a mechanism for generation of soluble cytokine receptors Proc Natl Acad Sci U S A 2004 101 1297 1302 14745008 10.1073/pnas.0307981100
Grell M Wajant H Zimmermann G Scheurich P The type 1 receptor (CD120a) is the high-affinity receptor for soluble tumor necrosis factor Proc Natl Acad Sci U S A 1998 95 570 575 9435233 10.1073/pnas.95.2.570
Shurety W Merino-Trigo A Brown D Hume DA Stow JL Localization and post-Golgi trafficking of tumor necrosis factor-alpha in macrophages J Interferon Cytokine Res 2000 20 427 438 10805378 10.1089/107999000312379
Decoster E Vanhaesebroeck B Boone E Plaisance S De Vos K Haegeman G Grooten J Fiers W Induction of unresponsiveness to tumor necrosis factor (TNF) after autocrine TNF expression requires TNF membrane retention J Biol Chem 1998 273 3271 3277 9452442 10.1074/jbc.273.6.3271
Beg AA Baldwin ASJ Activation of multiple NF-kappa B/Rel DNA-binding complexes by tumor necrosis factor Oncogene 1994 9 1487 1492 8152812
Brown K Gerstberger S Carlson L Franzoso G Siebenlist U Control of I kappa B-alpha proteolysis by site-specific, signal-induced phosphorylation Science 1995 267 1485 1488 7878466
Van Antwerp DJ Martin SJ Kafri T Green DR Verma IM Suppression of TNF-alpha-induced apoptosis by NF-kappaB Science 1996 274 787 789 8864120 10.1126/science.274.5288.787
Kim EY Teh HS Critical role of TNF receptor type-2 (p75) as a costimulator for IL-2 induction and T cell survival: a functional link to CD28 J Immunol 2004 173 4500 4509 15383581
|
15996269
|
PMC1183239
|
CC BY
|
2021-01-04 16:36:23
|
no
|
J Inflamm (Lond). 2005 Jul 4; 2:7
|
utf-8
|
J Inflamm (Lond)
| 2,005 |
10.1186/1476-9255-2-7
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-131592705810.1186/1743-0003-2-13ResearchInfluence of a portable audio-biofeedback device on structural properties of postural sway Dozza Marco [email protected] Lorenzo [email protected] Becky [email protected] Laura [email protected] Fay B [email protected] Angelo [email protected] Department of Electronics, Computer Science, and Systems, University of Bologna, Bologna, Italy2 Neurological Science Institute, Oregon Health & Science University, Portland (OR), USA2005 31 5 2005 2 13 13 24 1 2005 31 5 2005 Copyright © 2005 Dozza et al; licensee BioMed Central Ltd.2005Dozza 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
Good balance depends on accurate and adequate information from the senses. One way to substitute missing sensory information for balance is with biofeedback technology. We previously reported that audio-biofeedback (ABF) has beneficial effects in subjects with profound vestibular loss, since it significantly reduces body sway in quiet standing tasks.
Methods
In this paper, we present the effects of a portable prototype of an ABF system on healthy subjects' upright stance postural stability, in conditions of limited and unreliable sensory information. Stabilogram diffusion analysis, combined with traditional center of pressure analysis and surface electromyography, were applied to the analysis of quiet standing tasks on a Temper foam surface with eyes closed.
Results
These analyses provided new evidence that ABF may be used to treat postural instability. In fact, the results of the stabilogram diffusion analysis suggest that ABF increased the amount of feedback control exerted by the brain for maintaining balance. The resulting increase in postural stability was not at the expense of leg muscular activity, which remained almost unchanged.
Conclusion
Examination of the SDA and the EMG activity supported the hypothesis that ABF does not induce an increased stiffness (and hence more co-activation) in leg muscles, but rather helps the brain to actively change to a more feedback-based control activity over standing posture.
==== Body
Background
Maintaining balance is a complex task accomplished by the brain through the fusion and interpretation of sensory information. When sensory information from vestibular, somatosensory, and visual systems [1-3] are not accurate and/or adequate, balance will be compromised. Although, in many cases, the loss of peripheral sensory information is not curable or reversible, the brain can compensate for the loss of sensory information by relying more on the other sensory channels [4,5].
The purpose of biofeedback (BF) systems for postural control is to provide additional sensory information about body equilibrium to the brain [6]. In the last few years, different sensors, encoding algorithms, and information restitution devices have been combined to develop promising BF systems for postural control [7-9]. The major design goals were focused on portability, usability, economy, and effectiveness in balance improvements [8,10-12].
The development of these BF systems has been facilitated by the availability of lightweight, miniaturized, and economical sensors such as accelerometers, inclinometers, and gyroscopes [13]. The use of these sensors makes BF devices inexpensive, unsusceptible to shadowing effect, and not limited in the measurement field, in contrast to dynamometric platforms and motion analysis systems, which are commonly used in laboratory settings [14,15]. In addition, due to their size and weight, these sensors can measure body segment movement without hindering natural motor execution.
More detail is needed in understanding how biofeedback information interacts with the brain or, from a neuroscience perspective, how the brain uses artificial BF information and combines it with natural sensory information. We believe that understanding this interaction is fundamental for further developing effective BF systems.
An interesting analysis in the understanding of how the brain may use BF information for postural control was proposed by Collins and De Luca [16]. These authors developed a statistical-biomechanics method for analyzing force platform data recorded during quiet standing, called stabilogram diffusion analysis (SDA). SDA was applied to center of pressure (COP) data and it disclosed that COP tends to drift away from a relative equilibrium point over short-term observation intervals (less than 1-second long), whereas COP tends to return to a relative equilibrium point over long-term observation intervals. These results took Collins and De Luca to suggest that the motion of the COP is not purely random, and that SDA may be able to give insight on the amount of open-loop and closed-loop postural control applied by the central nervous system for maintaining balance [17]. SDA was used in several contexts, e.g. to evaluate the effect of spaceflight [18], visual input [19,20], and age-related changes [21,22] on postural stability. Chiari el al [20] developed and validated a new nonlinear model for extracting parameters from SDA diagrams, reducing from 6 to 2 the number of the parameters used to characterize the structural properties of COP. Rocchi et al. [23] found that these new parameters may be useful adjuncts to evaluate postural control strategies in patients with Parkinson's disease and may allow the comparison of different deep brain stimulation electrode sites based on their effect on structural properties of the COP.
In this paper, we investigate the effect on postural stability of a portable, accelerometry-based, audio biofeedback (ABF) system recently developed by the authors [9]. Standing with eyes closed on Temper™ foam will be used to evaluate the effects of artificial auditory cues to enhance the limited (from the eyes) and unreliable (from the feet) natural sensory information. Measurements include COP recorded by a force platform under the feet, trunk acceleration measured by the ABF sensors, and EMG signals from the leg muscles. SDA according to [20], traditional COP analysis [24], and muscle activation analysis according to [25] were performed in order to evaluate the effect of ABF on healthy young subject's upright posture.
These analyses were aimed to answer two questions: (1) do structural properties of postural sway change with ABF? And, if so, (2) in which way will this help in understanding the mechanisms underlying ABF efficacy and in improving the design of a rehabilitation strategy for balance disorders?
In this paper, we present evidence that supports the hypothesis that ABF does not induce a purely biomechanical increase in stiffness (and hence more co-activation) in the leg muscles, but rather ABF helps the brain actively adapt its control activity over standing posture.
Methods
Participants
Eight healthy subjects participated in this study (5 males and 3 females, aged 23.5 ± 3.0 yrs, range 21–28 yrs). All participants were free from any neurological, orthopaedic, hearing, or vestibular disorder. Informed consent form was obtained from each subject. The form was prepared in accordance with the Oregon Health and Science University Ethical Committee and respected the declaration of Helsinky, 1964.
Apparatus and procedure
Subjects performed 10, 60-second trials standing with eyes closed on Temper™, 4"-thick foam. COP displacement was recorded with an AMTI OR6-6 force plate. An ABF system [9] was used to provide subjects with additional balance information related to trunk acceleration. The ABF system used a sensor, based on 2-D accelerometers (Analog Device ADXL203) mounted on the subject's back (L5), to create an audio stereo sound representing the acceleration sensed along the anterior-posterior (AP) and the medial-lateral (ML) direction. A laptop, Toshiba Celeron 2.3 GHz, was dedicated to convert the accelerations into stereo sounds. Commercial headphones were used by the subjects to listen to the ABF sound. The ABF system is described in detail in [9] and illustrated in Figure 1. In short, the stereo sound provided by the ABF system consisted of two sine waves, one for the left ear channel and one for the right ear channel. Pitch, volume and left/right balance of the stereo sound were modulated to represent the 2-D acceleration information. Specifically, when the subject swayed forward, and consequently the acceleration increased in the anterior direction, the sound got louder in volume and higher in pitch. When the subject swayed backward, and consequently the acceleration increased in the posterior direction, the sound got louder in volume and lower in pitch. When the subject moved right and, consequently, the acceleration increased in the right direction, the sound got louder in the right ear channel and lower in the left one. When the subject moved left, and consequently the acceleration increased in the left direction, the sound got louder in the left ear channel and lower in the right one. The sound dynamics was optimized for each trial by taking as a reference the first 10-second recordings of each trial. The equations used for the pitch, volume, and left/right balance modulation can be found in [9]. Each subject was instructed to maintain balance during the trials by taking advantage of the ABF information, when available. Five trials with ABF and 5 trials without ABF were performed in randomized order by each subject. Before the experimental session, the subjects were instructed on how ABF codes trunk acceleration into sound, and performed free-movement trials until they felt confident in performing the full experiment.
Figure 1 ABF system device and protocol. The ABF consisted of (1) a sensor mounted on the trunk that measured accelerations along AP and ML axes, (2) a laptop acquiring acceleration from the sensor and processing the ABF sound, (3) a pair of headphones the subject wore for listening to the sound. In this figure is also shown the protocol in which a healthy subject is standing with eyes closed on a temper foam pad placed on a force plate. At the bottom right of the figure are statokinesigrams in condition with and without ABF from a representative subject.
Data recording
For each standing trial, ground reaction forces and torques were recorded from the force plate with a 100-Hz sampling frequency. COP displacement was computed offline from the force plate data after applying a 10-Hz cut-off, zero phase, low-pass Butterworth filter. Accelerations from the trunk along AP and ML direction were collected with a 100 Hz sampling frequency. EMG was recorded from right leg muscles, Tibialis (TI), Soleus (SO), and Gastrocnemius (GA) with two surface electrodes fixed about 6–8 cm apart along the length of each muscle belly; the ground electrode was fixed on a bony area of the right Hallux. The EMG signals were acquired with a 100-Hz sampling frequency, amplified 20000 times, band-pass filtered (71-2652 Hz), integrated with a 6th order Butterworth low-pass filter with a cut-off of 100 Hz (National Semiconductor MF6-100), and full-wave rectified.
Data analysis
From AP COP data, the root mean square distance (COP-RMS) and the frequency comprising the 95% of the power (F95%) were extracted according to Prieto et al. [24].
From the acceleration sensed at trunk level along AP direction we computed the root mean square value (Acc-RMS).
In addition, two stochastic parameters were included in the analyses. These parameters characterize a previously developed model that describes with continuity the transition among the different scaling regimes found in the COP time series [20] The model is described by the following equation:
V(Δt) = K Δt2H(Δt)
where V(Δt) is the variance of COP displacement, computed at time-lag Δt, and H is the scaling exponent, also called Hurst exponent. This is assumed to follow a sigmoid law in the time interval (Δt):
In this way, the features extracted from COP data are the following (see [20] for more details):
- K is an estimate of the diffusion coefficient of the random process obtained by sampling the COP time series at the sampling frequency 1/ΔTc.
- ΔTc represents the time-lag at which the real process corresponds to a purely random behavior, and where it switches from a persistent (positively correlated, and hence interpreted in terms of feed-forward control) to an anti-persistent (negatively correlated, and hence interpreted in terms of feedback control) behavior [16].
Mean muscular activity was calculated from the full wave rectified EMG of each muscle. Muscle activity was expressed as percentage of the maximal recorded activity for each muscle in each subject. This procedure allowed a reliable comparison of muscle activity between-subjects. The EMG signals were further processed applying a zero phase, low pass-filter with a 2 Hz cut-off in order to obtain tension curves according to Olney and Winter [25]. These tension curves were cross-correlated to determine the amount of co-activation between the muscles recorded.
Statistical analysis
Paired T-tests were performed to determine the effect of ABF on the different parameters extracted from COP, acceleration and EMG data collected. The threshold for statistical significance was set to p = 0.05.
Results
Subjects' confidence and comfort
All participants reported ABF sound was comfortable and its way of representing the information was intuitive. In fact, none of the subjects needed more than two, free-movement trials before feeling ready to start the experiment.
Subjects' sway
ABF significantly influenced subjects' balance on the foam. The percentage change induced by ABF on all sway parameters, either measured at the trunk level with the accelerometer or at the feet level with the force platform, is shown in Figure 2. Figure 2 also reports significance levels of the parameter changes occurred while using the ABF. The general results shown in Figure 2 are specified in detail in the following.
Figure 2 Effect of ABF on sway. The percent change of using ABF on the sway parameters is shown. COP-RMS and F95% were extracted from the AP COP displacement according to [24]. Acc-RMS was extracted from AP acceleration recorded at trunk level (L5). K and ΔTc were derived by applying the method proposed by Chiari et al. [20] on the SDA diagrams [16]. Asterisks indicate statistical significance: * p < 0.05 and ** p < 0.01. The reductions of K, COP-RMS and Acc-RMS are a consistent evidence of the reduction of sway amplitude shown by the subject using ABF. The increasing of F95% suggests that the postural control applied by the CNS when ABF is available was increased. The reduction of ΔTc suggests a major active closed-loop postural control exercised by the CNS.
Center of Pressure analysis
Center of pressure displacement in the AP direction was significantly influenced by ABF. T-test results revealed significant effects of ABF on COP-RMS (p = 0.015). This effect is shown by a consistent reduction of COP-RMS for 7 out of 8 subjects as shown in Table 1 (column 7). Average reduction of COP-RMS was 10.7%. Columns 1 and 4 of Table 1 also show the subject-by-subject values of COP-RMS without and with ABF, respectively. The last three subjects (#6, #7, #8) were females and showed smaller COP-RMS, as expected considering their smaller heights [26].
F95% increased with ABF for 7 out of 8 subjects (Table 1, column 8) but this result was not significant (p = 0.42). The values of F95% are also reported for each subject in both conditions (Table 1, columns 2 and 5). Average increase of F95% due to ABF was 6.2% as shown in Figure 2.
It is worth noting that subject #8 behaved as an outlier (Figure 3), compared to the other subjects since she was the only one who showed opposite changes in COP-RMS and F95% while using ABF. Performing the T-Tests, after eliminating this outlier, increased the significance of using ABF on COP-RMS and on F95% (p = 0.002 and p = 0.02, respectively). These results better match the results already published in [9]. The outlying behavior of subject #8 will be investigated further in the discussion.
Table 1 ABF effect on sway parameters Parameters. COP-RMS, F95%, and Acc-RMS are reported, subject-by-subject, for trials with and without ABF. Percentage differences between these two conditions are also reported. Standard deviations are indicated in parenthesis.
COP-RMS
(NO – ABF) [mm] F95 %
(NO – ABF) [Hz] Acc-RMS
(NO – ABF) [mm/s2] COP-RMS
(ABF) [mm] F95 %
(ABF) [Hz] Acc-RMS
(ABF) [mm/s2] COP-RMS
%difference F95%
%difference Acc-RMS
%difference
Subj. #1 10.79 (2.84) 0.99 (0.05) 137 (48) 9.57 (1.86) 1.18 (0.16) 118 (13) -11.2 19.1 -14.1
Subj. #2 9.91 (2.77) 1.20 (0.29) 142 (27) 9.50 (2.26) 1.30 (0.20) 120 (23) -4.1 8.7 -15.6
Subj. #3 9.21 (2.94) 1.16 (0.14) 121 (23) 8.61 (1.42) 1.37 (0.07) 113 (21) -6.5 18.0 -7.0
Subj. #4 10.23 (1.50) 1.43 (0.08) 117 (30) 8.80 (1.74) 1.49 (0.12) 100 (12) -13.9 4.1 -14.6
Subj. #5 8.50 (0.93) 1.49 (0.22) 143 (46) 6.90 (1.35) 1.53 (0.28) 115 (19) -18.8 2.6 -19.3
Subj. #6 9.62 (1.55) 1.34 (0.30) 126 (43) 7.35 (0.88) 1.34 (0.09) 89 (20) -23.6 0.0 -29.2
Subj. #7 6.37 (1.48) 1.60 (0.07) 64 (8.3) 5.19 (0.59) 1.94 (0.12) 51 (4.7) -18.5 20.8 -20.1
Subj. #8 6.08 (1.19) 1.78 (0.25) 48 (6.3) 6.75 (1.41) 1.37 (0.16) 39 (3.8) 10.9 -23.1 -17.3
Average 8.84 (1.75) 1.37 (0.26) 112 (36) 7.83 (1.54) 1.44 (0.15) 93 (31) -10.7 (10.92) 6.2 (14.4) -17.2 (6.3)
Figure 3 Antithetic behaviour of subject #8. COP-RMS percentage change using ABF is reported on the horizontal axis and F95% percentage change using ABF is reported on the vertical axis. The values of each subject from Table 1 are plotted. Subject #8 clearly behaves antithetically to the other subjects.
Acceleration analysis
Acceleration sensed at trunk level (L5) in AP direction was significantly reduced by ABF. T-test results also revealed significant effects of ABF on Acc-RMS (p = 0.0009). Acc-RMS was reduced by ABF across all subjects, as shown in Table 1 (last column).
Average reduction of Acc-RMS was 17.2% (Figure 2). Columns 3 and 7 of Table 1 also show the subject-by-subject values of Acc-RMS without and with ABF, respectively. The last three subjects were females and showed smaller Acc-RMS, as expected considering their smaller heights [26].
Stabilogram diffusion analysis
SDA diagrams plotted from AP COP data, were also significantly influenced by ABF (Figure 4). As a consequence, the parameters K and ΔTc characterizing the SDA diagram were both significantly decreased by ABF (Figure 2). Average K reduction was 9.3% (p = 0.02), whereas average ΔTc reduction was 33.9% (p = 0.018). Table 2 reports the subject-by-subject values of K and ΔTc in both conditions tested. Subject #8 and subject #7 are the only ones who showed a slight increase in K.
Figure 4 Effect of ABF on postural control strategy. SDA diagrams for one representative subject. Two conditions are reported: without ABF (black) and with ABF (gray). The behaviour of K and ΔTc used to parameterize the SDA diagrams is also shown. This figure suggests that, using ABF, subjects decrease the amount of sway by increasing the closed-loop (feedback) posture control.
Table 2 ABF effect on SDA parameters Parameters. K and ΔTc are reported, subject-by-subject, for trials with and without ABF. Percentage differences between these two conditions are also reported. Standard deviations are indicated in parenthesis.
K (NO-ABF) [mm2] Δtc (NO-ABF) [s] K (ABF) [mm2] Δtc (ABF) [s] K % difference Δtc % difference
Subj. #1 100 (57) 0.42 (0.21) 86 (15) 0.38 (0.17) -14.6 -9.9
Subj. #2 70 (29) 0.51 (0.31) 66 (20) 0.41 (0.34) -7.4 -20.5
Subj. #3 75 (41) 0.52 (0.29) 65 (20) 0.29 (0.12) -13.3 -45.3
Subj. #4 80 (21) 0.81 (0.46) 70 (14) 0.39 (0.14) -11.1 -52.0
Subj. #5 47 (13) 0.32 (0.08) 39 (10) 0.26 (0.16) -18.1 -19.7
Subj. #6 64 (12) 0.27 (0.08) 61 (9) 0.20 (0.09) -5.7 -26.1
Subj. #7 32 (7) 0.17 (0.06) 34 (9) 0.09 (0.01) 6.6 -47.4
Subj. #8 35 (14) 0.29 (0.09) 38 (13) 0.19 (0.06) 5.8 -34.3
Average 63 (23) 0.41 (0.20) 57 (18.5) 0.27 (0.11) -9.3 (9.2) -33.9 (15.3)
Muscle activity analysis
Muscle activity of TI, GA, and SO was not influenced by ABF. Overall, the mean activity, expressed as a percentage of the maximal activity recorded from each single muscle across all the trials of a subject, did not change significantly due to ABF (see Figure 5A). TI activity showed a trend toward increasing in trials with ABF (p = 0.17) but this change was particularly clear only for subjects #4 and #7.
Figure 5 Effect of ABF on muscle activity. Estimates of muscle activity levels (Fig. 5A) and muscular co-activation (Fig. 5B) for different pairs of muscles (TI-GA, TI-SO, GA-SO) are shown. Average values are reported for trials with (light gray) and without (dark gray) ABF. Error bars represent standard deviations. As shown in Figure 5A, using ABF does not change significantly the activity of the muscles analyzed (p values from T-Test are reported). This suggests that the major amount of postural corrections induced by ABF does not involve a major average activity of the muscles TI, GA, and SO in the leg. As shown in Figure 5B, using ABF does not change significantly the co-activation between the muscles analyzed (p values from T-Test are reported). This suggests that the major amount of postural corrections induced by ABF does not involve a major co-activation of the muscles TI, GA, and SO in the leg.
Muscle co-activation of ankle agonists-antagonists did not change significantly due to the ABF (see Figure 5B). Co-activation between TI and GA was small both with (r2 = 0.11) and without (r2 = 0.08) ABF. Similarly small was the co-activation between TI and SO with (r2 = 0.14) and without (r2 = 0.09) ABF. As expected, co-activation between the agonists muscles, GA and SO, was instead large (r2 = 0.39 in trials with ABF and r2 = 0.46 in trials without ABF). Figure 5B reports the coefficient of determination r2, which indicates the amount of muscular co-activation, for all pairs of muscles analyzed in trials with and without ABF.
Discussion
Using the proposed ABF device, all healthy subjects included in this study could sway less when standing in a particularly challenging condition, with vision unavailable and somatosensation partly unreliable. All subjects, in fact, reduced their AP Acc-RMS (see Table 1). In this way, subjects were further from their stability limits and, consequently, more stable. Trunk stabilization resulted in smaller corrective torques at the ankles, and hence smaller COP displacements. All but one subjects (Subj. #8) showed a significant decrease in AP COP-RMS (Fig. 2). During ABF, postural corrections in leg muscles were smaller but more frequent in number, as suggested by the increase in F95% of the COP. Future studies involving more sophisticated techniques for the acquisition and analysis of the EMG signals will be needed to validate this hypothesis. This result suggests that ABF may partially substitute for the lack of visual and somatosensory information for postural control by taking the postural control system towards a new steady state associated with a different control strategy.
Examination of the SDA and the EMG activity supported the hypothesis that ABF does not induce an increased stiffness (and hence more co-activation) in leg muscles, but rather helps the brain to actively change to a more feedback-based control over standing posture. Representative SDA diagrams reported in Figure 4 suggest that ABF contributes to a general reduction of both the diffusion coefficient K and the transition time ΔTc. Downward shifts of the SDA diagrams, described by smaller diffusion coefficients, reflect a reduced stochastic activity of the COP, and hence a more tightly regulated control system [16]. Shorter transition times reflect an earlier switching between persistent and anti-persistent behaviors, and hence more prompt reactions to perturbations of the postural control system [27]. In summary, these results support the hypotheses that ABF: 1) increases postural stability in stance, and 2) results in a more prominent role for feedback control over feed-forward control. Hence, the solution proposed by the brain with ABF seems to involve more feedback control for a more stable sway.
Interestingly, our results differ from the results observed by Rougier in quiet stance experiments with visual BF [28]. With visual BF, SDA diagrams only changed some local properties (local slopes) over short or long observation intervals but did not shift significantly, consistent with little, if any, change in K. Furthermore, with visual BF, closed-loop control operated over longer observation-times, suggesting that feed-forward control expanded over feedback control. Such a different behavior between auditory and visual BF may be due to the peculiar, non-redundant role of different senses in multi-sensory integration for the control of posture [29]. Whereas vision provides information about the external environment, it allows predictions of forthcoming events in the scene (feed-forward control) [30]. In contrast, hearing, compared to vision, may be more important for postural reactions to disturbing stimuli (feedback control). This result can also be related to the different processing times required by the central nervous system for visual and auditory stimuli with auditory reaction times significantly faster than visual reaction times. Finally, another factor which may explain the different outcomes of the two BF-studies is the selection of two, different, input variables (COP for visual BF and Acceleration from the trunk for ABF). It is widely accepted that upper- and lower- body segments are controlled separately [31].
Both predictive (feed-forward) and reactive (feedback) control need to be used in order to have an adequate interaction with the environment for postural stability. For this reason, it's hard to determine the relative validity of audio and visual BF. Rather, it may be important, in a rehabilitation setting, to identify which one of the two components of postural control (feed-forward or feedback) needs more reinforcement or substitution in a particular patient, and consequently design an optimised BF treatment.
The outlying results observed for Subj. #8 need to be discussed individually. This woman in fact did not decrease COP-RMS and K, and did not increase F95%, although, similarly to the other subjects, she decreased Acc-RMS and ΔTc (these changes were consistent across the whole population). Hence, with ABF she actually swayed less and she showed the same increase of feedback control. Nonetheless, either due to her small body size or to a slightly different control scheme, she obtained these goals with a different strategy. Figure 6 reports her muscle activities and co-activations. It can be seen how she generally increases muscle activity with ABF (Figure 6A), in particular with a large increase in the activity of posterior muscles, GA and SO. It should be noted, however, that also the estimated co-activations (Figure 6B) look pretty dissimilar compared with the ones of the other subjects, shown in Figure 5B. Particularly low is the co-activation of agonists muscles GA-SO without ABF, which ABF partly contributes to enlarge. For all these reasons, her postural behavior in the proposed task should be looked as an outlying behavior and more analyses are needed, on a larger population, to assess the real influence of body size or usual control strategies on the responsiveness to ABF.
Figure 6 Muscle activity and co-activation in subject #8. The antithetic behaviour of subject #8 for muscles activity (Fig. 6B), and for muscles co-activation (Fig. 6A) is shown. Figure 6A reports the estimates of muscular activity for TI, GA, and SO muscle. Average values expressed in percentage are reported for trials with (light gray) and without (dark gray) ABF. Error bars represent standard deviations. The percent activity was calculated taking as one-hundred-percent reference the trial with the highest muscle activation recorded. Even if muscle activity looks higher in trials with ABF for all muscles, only SO activity changed significantly while using ABF (p values from T-Test are reported; since the number of samples is five, it is convenient to report also the powers which were respectively: 0.09, 0.41, 0.53). This suggests that a major amount of activity of the muscles TI, GA, and SO was exercised by this subject while using ABF. Figure 6B reports the estimates of muscular co-activation for different pairs of muscles: TI-GA, TI-SO, and GA-SO. Average values are reported for trials with (light gray) and without (dark gray) ABF. Error bars represent standard deviations. Even if co-activation looks higher in trials with ABF for all couples of muscles while using ABF, muscles co-activation does not change significantly (p values from T-Test are reported; since the number of samples is five it is convenient to report also the powers which were respectively: 0.20, 0.14, 0.23). This suggests that a major amount of co-activation of the muscles TI, GA, and SO was exercised by this subject while using ABF.
Many earlier biofeedback systems used audio alarms to notify the user of abnormal values of monitored parameters (e.g. [32]). The present ABF system is novel in the use of nonlinear coding functions and in the customization of these functions for each subject and task [9]. Although the current ABF system may interfere with use of hearing for communication, it may be quite useful during the rehabilitation and training process. Plans are underway to improve the current ABF system by making it wireless for increased portability and equipping it with a communication module for remote control, recording, and monitoring. Different sonification procedures will also be tested and compared in a near future. Specifically, 3-D generated sound with a HRTF (Head Related Transfer Function) or immersive sound may be even more effective signals for improving stance balance.
Conclusion
We have investigated the attributes of a portable instrument that feeds back trunk acceleration in order to help subjects to reduce their postural sway during stance. The instrument meets requirements for an adequate biofeedback system that may find interesting applications not only as a rehabilitation device in the clinic, but also in the home care setting, and when doing community mobility training outside the traditional clinical setting. In fact, it has appropriate bandwidth and sensitivity, smoothness and delay of the acoustic signal generator, as well as portability. Acoustic information related to trunk movement allowed subjects in the present experiment to increase postural stability when sensory information from both vision and the surface were compromised by eye closure and stance on foam. We provided evidence that the balance improvement was not a stiffening at the ankle, but rather the brain actively adapted its control strategy over standing posture with more feedback-based control.
List of abbreviations
ABF = audio biofeedback
Acc-RMS = root mean square of the acceleration
AP = anterior-posterior
BF = biofeedback
COP = center of pressure
COP-RMS = root mean square of the COP
EMG = electromyography
F95% = frequency comprising the 95% of the power
GA = gastrocnemius
ML = medial-lateral
SDA = stabilogram diffusion analysis
SO = soleus
TI = tibialis
Competing interests
The authors of this paper declare that they applied for a patent (patent application PCT/IB2004/001679) concerning the audio-biofeedback device used for the study described in this paper. The patent is property of the University of Bologna and of the Oregon Health & Science University.
Acknowledgements
We thank Jim Frank, Maria Grazia Benedetti, and Marco Knaflitz for the advice about electromyography analysis. We also thank Charles Russell and Andrew Owings for technical assistance.
==== Refs
Horak FB Shupert CL Mirka A Components of postural dyscontrol in the elderly: a review Neurobiol Aging 1989 10 727 738 2697808 10.1016/0197-4580(89)90010-9
Lord SR Ward JA Age-associated differences in sensori-motor function and balance in community dwelling women Age Ageing 1994 23 452 460 9231937
Corriveau H Prince F Hebert R Raiche M Tessier D Maheux P Ardilouze JL Evaluation of postural stability in elderly with diabetic neuropathy Diabetes Care 2000 23 1187 1191 10937520
Horak FB Macpherson JM Rowell RB, Shepherd JT Postural equilibrium and orientation Published for the American Physiology Society by Oxford University Press 1996 New York: Published for the American Physiology Society by Oxford University Press 255 292
Horak FB Hlavacka F Somatosensory loss increases vestibulospinal sensitivity J Neurophysiol 2001 86 575 585 11495933
Moore S Woollacott M The use of biofeedback to improve postural stability Phys Ther Practice 1993 2 1 19
Wall C IIIWeinberg MS Schmidt PB Krebs DE Balance prosthesis based on micromechanical sensors using vibrotactile feedback of tilt IEEE Trans Biomed Eng 2001 48 1153 1161 11585039 10.1109/10.951518
Tyler M Danilov Y Bach YR Closing an open-loop control system: vestibular substitution through the tongue J Integr Neurosci 2003 2 159 164 15011268 10.1142/S0219635203000263
Chiari L Dozza M Cappello A Horak FB Macellari V Giansanti D An accelerometry-based system for balance improvement using audio-biofeedback IEEE Trans Biomed Eng 2005
Bazzarelli M Durdle N Lou E Raso J Hill D A wearable networked embedded system for the treatment of scoliosis Stud Health Technol Inform 2002 91 383 386 15457761
Kadkade PP Benda BJ Schmidt PB Wall C III Vibrotactile display coding for a balance prosthesis IEEE Trans Neural Syst Rehabil Eng 2003 11 392 399 14960115 10.1109/TNSRE.2003.819937
Dozza M Chiari L Horak FB Audio- biofeedback improves balance in patients with bilateral vestibular loss Arch Phys Med Rehabil 2005 86 1401 1403 16003671 10.1016/j.apmr.2004.12.036
Luinge HJ Veltink PH Inclination measurement of human movement using a 3-D accelerometer with autocalibration IEEE Trans Neural Syst Rehabil Eng 2004 12 112 121 15068194 10.1109/TNSRE.2003.822759
Mayagoitia RE Lotters JC Veltink PH Hermens H Standing balance evaluation using a triaxial accelerometer Gait Posture 2002 16 55 59 12127187 10.1016/S0966-6362(01)00199-0
Mayagoitia RE Nene AV Veltink PH Accelerometer and rate gyroscope measurement of kinematics: an inexpensive alternative to optical motion analysis systems J Biomech 2002 35 537 542 11934425 10.1016/S0021-9290(01)00231-7
Collins JJ De Luca CJ Open-loop and closed-loop control of posture: a random-walk analysis of center-of-pressure trajectories Exp Brain Res 1993 95 308 318 8224055 10.1007/BF00229788
Collins JJ De Luca CJ Upright, correlated random walks: A statistical-biomechanics approach to the human postural control system Chaos 1995 5 57 63 12780156 10.1063/1.166086
Collins JJ De Luca CJ Pavlik AE Roy SH Emley MS The effects of spaceflight on open-loop and closed-loop postural control mechanisms: human neurovestibular studies on SLS-2 Exp Brain Res 1995 107 145 150 8751072 10.1007/BF00228026
Collins JJ De Luca CJ The effects of visual input on open-loop and closed-loop postural control mechanisms Exp Brain Res 1995 103 151 163 7615030 10.1007/BF00241972
Chiari L Bertani A Cappello A Classification of visual strategies in human postural control by stochastic parameters Hum Mov Sci 2000 19 817 842 10.1016/S0167-9457(01)00024-0
Collins JJ De Luca CJ Burrows A Lipsitz LA Age-related changes in open-loop and closed-loop postural control mechanisms Exp Brain Res 1995 104 480 492 7589299 10.1007/BF00231982
Laughton CA Slavin M Katdare K Nolan L Bean JF Kerrigan DC Philips E Lipsitz LA Collins JJ Aging, muscle activity, and balance control: physiologic changes associated with balance impairment Gait Posture 2003 18 101 108 14654213 10.1016/S0966-6362(02)00200-X
Rocchi L Chiari L Cappello A Gross A Horak FB Comparison between subthalamic nucleus and globus pallidus internus stimulation for postural performance in Parkinson's disease Gait Posture 2004 19 172 183 15013506 10.1016/S0966-6362(03)00059-6
Prieto TE Myklebust JB Hoffmann RG Lovett EG Myklebust BM Measures of postural steadiness: differences between healthy young and elderly adults IEEE Trans Biomed Eng 1996 43 956 966 9214811 10.1109/10.532130
Olney SJ Winter DA Predictions of knee and ankle moments of force in walking from EMG and kinematic data J Biomech 1985 18 9 20 3980492 10.1016/0021-9290(85)90041-7
Chiari L Rocchi L Cappello A Stabilometric parameters are affected by anthropometry and foot placement Clin Biomech (Bristol, Avon) 2002 17 666 677 12446163 10.1016/S0268-0033(02)00107-9
Chiari L Cappello A Lenzi D Della CU An improved technique for the extraction of stochastic parameters from stabilograms Gait Posture 2000 12 225 234 11154933 10.1016/S0966-6362(00)00086-2
Rougier P Influence of visual feedback on successive control mechanisms in upright quiet stance in humans assessed by fractional Brownian motion modelling Neurosci Lett 1999 266 157 160 10465697 10.1016/S0304-3940(99)00272-4
Maurer C Mergner T Peterka RJ A multisensory posture control model of human upright stance Prog Brain Res 2003 142 189 201 12693262
Cerminara NL Edge AL Marple-Horvat DE Apps R The lateral cerebellum and visuomotor control Prog Brain Res 2004 148 213 226 15661193
Buchanan JJ Horak FB. Voluntary control of postural equilibrium patterns Behav Brain Res 2003 143 121 40 12900039 10.1016/S0166-4328(03)00038-X
Wong AM Lee MY Kuo JK Tang FT The development and clinical evaluation of a standing biofeedback trainer J Rehabil Res Dev 1997 34 322 327 9239626
|
15927058
|
PMC1183240
|
CC BY
|
2021-01-04 16:37:40
|
no
|
J Neuroengineering Rehabil. 2005 May 31; 2:13
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-13
|
oa_comm
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.