Cytokine Production Associated with Periportal Fibrosis during Chronic Schistosomiasis Mansoni in Humans

Transcription

1 INFECTION AND IMMUNITY, Feb. 2006, p Vol. 74, No /06/$ doi: /iai Copyright 2006, American Society for Microbiology. All Rights Reserved. Cytokine Production Associated with Periportal Fibrosis during Chronic Schistosomiasis Mansoni in Humans L. F. Alves Oliveira, 1 E. C. Moreno, 2 G. Gazzinelli, 3,4 * O. A. Martins-Filho, 3 A. M. S. Silveira, 1 A. Gazzinelli, 5 L. C. C. Malaquias, 1 P. LoVerde, 6 P. Martins Leite, 1 and R. Correa-Oliveira 3 Núcleo de Pesquisa em Imunologia, Faculdade de Ciências da Saúde, Universidade Vale Do Rio Doce, Gov. Valadares, Minas Gerais, Brazil 1 ; Fundação Nacional de Saúde, Minas Gerais, Brazil 2 ; Centro de Pesquisas René Rachou Fiocruz, Belo Horizonte, Minas Gerais, Brazil 3 ; Santa Casa de Misericórdia, Belo Horizonte, Minas Gerais, Brazil 4 ; Escola de Enfermagem, UFMG, Belo Horizonte, Minas Gerais, Brazil 5 ; and State University of New York at Buffalo, Buffalo, New York 6 Received 8 July 2005/Returned for modification 30 August 2005/Accepted 7 November 2005 Volunteers living in an area where schistosomiasis mansoni is endemic were subjected to ultrasound examination and classified into groups according to the levels of fibrosis diagnosed, namely, absence of indications of fibrosis (group 0), incipient fibrosis (group 1), and moderate/severe fibrosis (group 2). Peripheral blood mononuclear cells (PBMC) collected from the volunteers were stimulated with soluble antigens from adult schistosomes or from schistosome eggs, and the production of the cytokines gamma interferon, tumor necrosis factor alpha, transforming growth factor (TGF- ), interleukin-4 (IL-4), IL-10, and IL-13 was determined. Potential associations of the level of fibrosis with age, sex, intensity of infection, and cytokine production were investigated between the three groups. Univariate analysis identified associations of age (>50), gender (male), and absence of eggs/g of feces with moderate/severe fibrosis and an association of intensity of infection (>100 eggs) with incipient fibrosis. When cytokine production in PBMC cultures stimulated by soluble egg antigens was categorized as low or high, significant differences in the distribution of IL-13 levels were established between groups 0 and 2. No significant differences were detected between the groups in the cytokines produced by PBMC cultures stimulated with soluble antigens from adult schistosomes. When all variables were tested in multivariate analyses, only IL-13 was strongly associated with fibrosis (odds ratio 5.8; 95% confidence interval [CI] 1.1 to 30.5). While high levels of TGF- appeared to be associated with protection against fibrosis, the strength of the association was low. * Corresponding author. Mailing address: Centro de Pesquisas Rachou, Fundação Oswaldo Cruz, Avenida Augusto de Lima 1715, Barro Preto, Belo Horizonte, MG , Brazil. Phone: Fax: ggaz@cpqrr.fiocruz.br. Present address: Southwest Foundation for Biomedical Research, San Antonio, Tex. Chronic hepatosplenomegaly in schistosomiasis is encountered more frequently in patients resident in areas of schistosomiasis endemicity showing high transmission rates of the disease, and the level of its occurrence has been associated with the intensity of infection (2, 16). The severity of the disease is a consequence of fibrosis caused by granulomatous reaction around the schistosome eggs that are trapped in the small vessels of the liver. However, only a small percentage of individuals infected in an area of endemicity actually develop a severe form of the disease. Studies with mice which were deficient in the expression of certain cytokines (knockout mice) and had been infected with Schistosoma mansoni showed that the formation of egg granulomas and the ensuing hepatic fibrosis are dependent on the regulation of cytokines (35). However, the factors and sequential processes involved in the development of periportal fibrosis in humans are poorly understood. We have previously demonstrated (34) that in vitro stimulation with soluble egg antigens (SEA) of peripheral blood mononuclear cells (PBMC) derived from schistosome eggnegative subjects produced significantly higher levels of the secreted cytokine gamma interferon (IFN- ) than the PBMC of individuals whose intensity of infection was determined as more than 100 eggs/g of feces. On the other hand, PBMC from the egg-positive group produced significantly higher amounts of interleukin-10 (IL-10) following such stimulation, and a strong inverse correlation between IL-10 and IFN- formation was evident. Moreover, it was shown that the ratio of IFN- / IL-10 influences the levels of tumor necrosis factor alpha (TNF- ). Given that IFN- has been specifically associated with protection against fibrosis (4, 11, 17) and that its production depends on the relative concentration of IL-10, it is conceivable that IL-10 might also play some role in the development of morbidity in schistosomiasis. Furthermore, other cytokines, including IL-13, have already been reported to be involved in the development of periportal fibrosis in patients with chronic schistosomiasis living in areas of endemicity (4, 27, 31, 33). In the present study, we have evaluated the levels of production of the cytokines IFN-, TNF-, transforming growth factor (TGF- ), IL-4, IL-10, and IL-13 following in vitro stimulation with SEA or soluble adult worm antigen preparation (SWAP) of PBMC from individuals of different ages and sexes and with various intensities of infection and degrees of fibrosis. The use of ultrasound enabled the direct evaluation of hepatic fibrosis in large cohorts of subjects, and several criteria are available for grading the severity of the disease based on 1215

2 1216 ALVES OLIVEIRA ET AL. INFECT. IMMUN. ultrasound examinations of liver and spleen (1). When adjusted for other covariates, the results indicated an important association of IL-13 levels with moderate or severe periportal fibrosis. Although TGF- was found to be negatively associated with the development of fibrosis, the role of this cytokine remains undefined and requires further investigation. MATERIALS AND METHODS The study described in this paper was approved by the ethical committees of the Centro de Pesquisas René Rachou, Fiocruz, Universidade Vale do Rio Doce, and the State University of New York at Buffalo and was conducted in full accordance with accepted guidelines (10). Study area and population. Study subjects were drawn from among residents of Virgem das Graças (Jequitinhonha Valley, Minas Gerais, Brazil). Schistosomiasis is endemic in this area, and other intestinal parasites such as Ascaris lumbricoides, Trichuris trichiura, and species of Ancylostoma and Entamoeba are found in the population. Malaria, leishmaniasis, or Chagas disease, however, are not endemic to the area. Prior to the start of the study, local health care personnel explained the research to the village population, stressing that all individuals would receive appropriate treatment whether or not they eventually agreed to participate in the study. As necessary, individuals were treated with praziquantel at the standard dose (50 mg/kg of body weight) after the blood donation, and the treatment was repeated after 2 months if results from examination of the feces remained positive. Volunteers were enrolled into the study only after informed consent was obtained from each individual or their legal guardian. Clinical histories and quantitative egg counts, as determined from Kato-Katz thick stool smears (21), were collected from 644 residents in the area (95.8% of the total population). From these, 328 volunteers were submitted to physical and ultrasound examinations performed by different physicians, and 91 individuals agreed to donate blood for the cytokine study. The physician performing the ultrasound examinations was not aware of the infection status of the subject nor of the results of the clinical examination. Liver size, portal vein diameter, thickness of the walls of the central and peripheral portal branches, spleen size, and splenic and mesenteric vein diameters were assessed using a conventional portable ultrasound instrument (ultrasonic diagnostic instrument model EUB-200; Hitachi, Tokyo, Japan). Periportal thicknesses were evaluated according to the established criteria for the classification of hepatic fibrosis, with some modifications (1, 25). In this classification, the degree of fibrosis was determined from the central and peripheral echogenic periportal thickness of the liver portal system. Preparation of antigens. Schistosome eggs were isolated from the livers of infected mice exposed 8 weeks previously to cercariae and were homogenized and ground in cold phosphate-buffered saline (PBS). The clear supernatant fluid (SEA) resulting from high-speed centrifugation of the homogenate at 10 5 g for 1 h at 4 C was stored for use (5). Adult S. mansoni worms (male and female) were suspended in ice-cold PBS and sonicated over ice (2 10 min). Insoluble material was removed by ultracentrifugation at 10 5 g for1hat4 C, and the supernatant (SWAP) was stored for use (15). Preparation and stimulation of cultures of PBMC. Human PBMC were purified from heparinized venous blood samples by sedimentation at 400 g for 45 min at 18 C in a density gradient of Ficoll-Hypaque (LMS Litton Biometrics, Kensington, MD). In order to assay the stimulation of cytokines, PBMC were cultured at cells in a total volume of 1 ml of RPMI 1640 (Gibco-BRL, Gaithersburg, MD) supplemented with 5% AB Rh-positive heat-inactivated human serum (Sigma, St. Louis, MO) and 3% antibiotic/antimycotic solution (stock mix contained 10,000 U penicillin, 10 mg streptomycin, and 25 g amphotericin B [Fungizone]/ml; Gibco-BRL) in the presence of SEA or SWAP at a concentration of 25 g/ml. Following incubation for 144 h at 37 C in 5% carbon dioxide, supernatants were collected and immediately frozen ( 70 C) for subsequent determination of cytokine concentration. Determination of cytokines. The concentrations of human cytokines were determined using enzyme-linked immunosorbent assay-based prediction kits from Pharmingen (San Diego, CA), following the instructions issued by the manufacturer. Briefly, for each cytokine, flat-bottomed plates (Immunolon 4; Dynatech, Chantilly, VA) were each coated overnight at 4 C with the appropriate antibody diluted in carbonate-bicarbonate coating buffer (ph 9.5). Following an initial wash with PBS (ph 7.2) containing 0.05% Tween 20 (Sigma) (a rinsing solution that was used for all subsequent washes between incubation steps), the plates were blocked with 10% fetal calf serum in PBS for 1hatroom temperature. Samples or standards were then added to the wells and the plates were incubated overnight at 4 C, washed, and incubated for 1hinthepresence of horseradish peroxidase-conjugated anti-human cytokine antibodies. Finally, 50 l of the substrate o-phenylenediamine (Sigma) with 0.03% hydrogen peroxide was added to each well, and the formation of product determined at 450 nm using an automated enzyme-linked immunosorbent assay reader (Molecular Devices, Sunnyvale, CA). Concentrations of cytokines (pg/ml) in the samples were determined by interpolation from standard calibrations set up on each plate using appropriate amounts of recombinant human cytokines. Data analysis. The data set was organized using EpiInfo (Centers for Disease Control and Prevention) software version 6.04d, and statistical analyses were performed using Stata Corporation (College Station, TX) statistical software version 8.4 (2003). Parametric ( 2 ) and nonparametric (Kruskal-Wallis and Mann-Whitney) tests were employed in the univariate analyses for categorical and continuous variables, respectively. For logistic regression analyses, the extreme values (outliers) of the cytokine responses were eliminated and the remaining values categorized as low or high with respect to the median value. The outliers were identified after an exploratory data analysis and construction of boxplot graphs of all cytokines grouped by fibrosis. The outliers are extreme values that are unlikely to occur and probably result from laboratory artifacts. Five outliers were eliminated (SEA data): IFN-, one; IL-10, one; IL-4, one; and TGF-, two. In order to evaluate the associations of the various covariates that might influence the development of periportal fibrosis, multivariate analyses were conducted using a logistic regression model. Initially, univariate analyses were performed in order to select the variables to be included in the multivariate analyses. Any variable whose univariate test produced a significance level P of 0.20 was considered a candidate for the multivariate model, together with all variables of known biologic importance for fibrosis in schistosomiasis, even if such variables did not achieve the stated significance. In the second step, logistic regressions were conducted separately and included (i) demographic and parasitological variables and (ii) the levels of cytokines produced after antigen stimulation. All of the variables that achieved a P value of 0.15 in the intermediate logistic regression were retained for the final multivariate analysis. Those variables that attained a P value of 0.05 were considered significantly associated with fibrosis after adjustment for the other investigated covariates. In the logistic regression, a stepwise procedure was used, commencing with a model containing all of the selected variables and followed by successive backward elimination. In each case, statistical significance was determined by a likelihood ratio test (20). In order to establish the risk factors associated with fibrosis, three comparative models were considered, in which the subjects were grouped according to different degrees of fibrosis, i.e., model A (group 0 versus group 1 plus group 2), model B (group 0 plus group 1 versus group 2), and model C (group 0 versus group 1 versus group 2). RESULTS Study population. Among the 91 subjects who took part in the cytokine study, 48 (52.8%) were females and 43 (47.2%) were males, and their ages ranged from 14 to 85 years (mean, years; median, 53 years; 25th-75th percentile range, 36 to 62 years). A total of 55 individuals (60.4%) presented parasitic infections caused variously by Ascaris lumbricoides, Ancylostoma duodenale, Trichuris trichiura, Enterobius vermicularis, and species of Taenia. Fifty-one (56.0%) participants were, however, egg positive for schistosomiasis, and their infection levels ranged from 12 to 1,272 eggs/g of feces (median, 72 eggs/g of feces; 25th to 75th percentile range, 24 to 160 eggs/g). Of these egg-positive subjects, 30 (58.8%) exhibited infection levels lower than 100 eggs/g of feces and 21 (41.2%) presented more than 100 eggs/g of feces. According to the ultrasound examinations conducted on all participants, 47 individuals (51.6%) showed no signs of periportal fibrosis (absence of clinical and ultrasound evidence of hepatic fibrosis or other alterations in the liver or spleen; periportal thickness, 2.9 mm) and were assigned to group 0. Incipient periportal fibrosis (periportal thickness, 3 to 5 mm) was exhibited by 22 subjects (24.2%) and these made up group 1, while group 2 was composed of the 15 individuals (16.5%) who showed moderate fibrosis (periportal thickness, 5 to 7

3 VOL. 74, 2006 CYTOKINE PRODUCTION AND FIBROSIS IN SCHISTOSOMIASIS 1217 Variable TABLE 1. Characterization of the study group Result for: Group 0 (n 47) Group 1 (n 22) Group 2 (n 22) Mean age SD b (yr) Egg counts (eggs/g of feces) Range , Mean SD Median c (25th 75th percentile range) (0 72) (0 248) (0 4) Age groups d yr 22 (46.8) e 14 (63.6) 4 (18.2) 50 yr 25 (53.2) 8 (36.4) 18 (81.8) 0.01 Gender f Female 28 (58.3) 14 (29.2) 6 (12.5) Male 19 (44.2) 8 (18.6) 16 (37.2) 0.02 Egg distribution g No eggs 18 (38.3) 6 (27.3) 16 (72.7) eggs/g of feces 20 (42.6) 6 (27.3) 4 (18.2) 100 eggs/g of feces 9 (19.1) 10 (45.4) 2 (9.1) 0.01 Other parasitic diseases Negative 21 (44.7) 9 (40.9) 6 (27.3) Positive 26 (55.3) 13 (59.1) 16 (74.7) 0.38 P value a a Significance levels based on 2 test (categorical variables) or Kruskal-Wallis and Mann-Whitney tests (continuous variables): minimum accepted level of significance was a P value of b Significant differences determined between groups 0 and 2 (P 0.036), and between 1 and 2 (P 0.001). c Significant differences detected between groups 0 and 1 (P 0.075), between 0 and 2 (P 0.013), and between 1 and 2 (P 0.002). d Significant differences determined between groups 0 and 1 (P 0.192), between 0 and 2 (P 0.022), and between 1 and 2 (P 0.002). e Values in parentheses are percentages. f Significant differences determined between groups 0 and 2 (P 0.012), and between 1 and 2 (P 0.015). g Significant differences determined between groups 0 and 2 (P 0.013), and between 1 and 2 (P 0.002). mm) and the 7 individuals (7.7%) who presented severe fibrosis (periportal thickness, 7 mm). Subjects with moderate and severe hepatic fibrosis were included in the same group because of the low frequency of occurrence of severe fibrosis. Information concerning age, gender, infection level, and the presence of other parasitic diseases with respect to the different degrees of hepatic fibrosis defined for the three groups is provided in Table 1. All measured variables, except for that associated with the presence of other parasitic diseases, showed significant differences (P 0.20) between at least some of the groups. With respect to mean age, differences were observed between groups 0 and 2 (P 0.036) and groups 1 and 2(P 0.001). When this variable was allocated into the 14- to 50-year and 50-year age group categories, differences were also observed between groups 0 and 1 (P 0.192). Differences in gender were observed between groups 0 and 2 (P 0.012) and1and2(p 0.015), while differences in intensity of infection were observed between groups 0 and 1 (P 0.075), 0 and2(p 0.013), and 1 and 2 (P 0.002). Univariate analysis thus permitted the identification of an association of age ( 50 years), gender (male), and absence of eggs/g of feces with moderate/severe fibrosis (group 2). Furthermore, an association of age ( 50 years) and intensity of infection ( 100 eggs) with incipient fibrosis (group 1) was observed. Levels of cytokines produced on stimulation of PBMC by SEA and SWAP. Table 2 shows the median levels of cytokines measured in supernatants collected from SEA-stimulated PBMC cultures that had been prepared from samples from subjects within the three defined study groups. As these data were not in a normal distribution, univariate analyses were performed using the Kruskall-Wallis nonparametric test: significant associations were shown only for cytokines IL-10 and IL-13 (P 0.20). Application of the Mann-Whitney test demonstrated that the median value of IL-13 in group 2 was significantly higher than that in group 0 (P 0.065), which was in marked contrast to results for IL-10, where the median value in group 2 was significantly lower than the corresponding values TABLE 2. Cytokines produced by SEA-stimulated cultures of PBMC derived from subjects without fibrosis (group 0) and those with incipient (group 1) and moderate/severe (group 2) fibrosis Cytokine Median value a for: Group 0 Group 1 Group 2 P value b IL IL IL TGF TNF IFN a Concentrations in pg/ml. b Significance levels based on Kruskal-Wallis test (median values); minimum accepted level of significance was a P value of 0.20.

4 1218 ALVES OLIVEIRA ET AL. INFECT. IMMUN. FIG. 1. The production of IL-10 and IL-13 in SEA-stimulated cultures of PBMC derived from subjects without fibrosis (group 0) and those with incipient (group 1) and moderate/severe (group 2) fibrosis. The data shown are mean values standard errors. Within each panel, bars bearing the same lower-case letter are significantly different according to the Mann-Whitney test (a, P 0.009; b, P 0.056; and c, P 0.065). in groups 0 and 1 (P and P 0.056, respectively) (Fig. 1). The levels of cytokines IFN-, IL-10, IL-4, TGF-, TNF-, and IL-13 in SEA-stimulated PBMC cultures derived from the three groups were categorized as low or high according to the median value of each cytokine titer, and the results are presented in Table 3. The numbers of patients evaluated in the cytokine assay varied because of the limited volume of culture supernatant available. Only for IL-10 and IL-13 were significant between-group differences observed in the percentages of patients exhibiting low and high levels of cytokines: for IL-10, there were significant differences between values for groups 0 and2(p 0.012) and 1 and 2 (P 0.092), while for IL-13, the significant difference was between groups 0 and 2 (P 0.024). No significant differences were observed in the percentages of individuals with low and high cytokine levels produced by SWAP stimulation of PBMC cultures derived from the three different groups of subjects (data not shown). Risk factors associated with fibrosis. The variables selected for inclusion in the intermediate logistic model from the demographic and parasitological data obtained in this study were age, gender, and egg distribution. Furthermore, even though only IL-13; and IL-10 showed significant associations in the univariate analyses, all of the cytokines produced by SEAstimulated PBMC were included in the logistic regression method, since they have been variously reported by others to be associated with periportal fibrosis. In model A, the variables that attained a significance level of 0.15 were egg distribution, IL-13, TGF-, while in model B, the significant variables were age, gender, egg distribution, IL-4, TNF-, TGF-, and IL-13; in model C, they were gender, egg distribution, IL-4, TNF-, TGF-, IL-13, and IFN-. After final multivariate adjustment, the variables that remained in the models were IL-13 and TGF- (P 0.05). Table 4 shows the degree of association of IL-13 and TGF- with fibrosis, as evaluated by odds ratio (OR) after adjusting for the other investigated covariates, for the three full models. In model A, in which all subjects exhibiting some level of fibrosis (i.e., members of groups 1 and 2) were taken together and compared with individuals without fibrosis (group 0), the variable positively associated with fibrosis was IL-13 (OR 3.1). In contrast, TGF- was negatively associated with fibrosis (OR 0.3); however, this association was not apparent in the univariate analysis (P 0.280) even though an association was apparent in logistic regression analysis. In model B, subjects with incipient fibrosis (group 1) were placed together with individuals without fibrosis (group 0) and compared with those exhibiting moderate/severe fibrosis (group 2): only IL-13 was positively associated with moderate/severe fibrosis (OR 4.0). TABLE 3. Percentages of individuals with low and high cytokine levels produced by SEA-stimulated cultures of PBMC derived from subjects without fibrosis (group 0) and those with incipient (group 1) and moderate/severe (group 2) fibrosis Group 0 Group 1 Group 2 Cytokine a No. % No. % No. % P value IFN- Low High IL-10 c Low High IL-4 Low High TGF- Low High TNF- Low High IL-13 d Low High a Cytokine levels were categorized as low or high according to the median value of the cytokine titer and compared among the different groups. b Significance levels based on 2 test (categorical variables); minimum accepted level of significance was a P value of c Significant differences determined in the percentages of individuals with low and high levels of IL-10 between groups 0 and 2 (P 0.012), and between groups 1and2(P 0.092). d Significant differences determined in the percentages of individuals with low and high levels of IL-13 between groups 0 and 2 (P 0.024).

5 VOL. 74, 2006 CYTOKINE PRODUCTION AND FIBROSIS IN SCHISTOSOMIASIS 1219 TABLE 4. Cytokines produced by SEA-stimulated cultures of PBMC associated with the degree of fibrosis according to the models employed in the multivariate analysis d Cytokine Adjusted odds ratio a (95% CI) P value b Model A (n 91) (group 0 versus groups 1 2) TGF- 0.3 ( ) IL ( ) Model B (n 56) (groups 0 1 versus group 2) IL ( ) Model C (n 56) (group 0 versus group 1 versus group 2) TGF- 0.2 ( ) c IL ( ) c a The odds ratio was adjusted for all variables included in the study. CI, confidence interval. b Significance value in the final model: minimum accepted level of significance was a P value of c Significant differences detected between groups 0 and 2. d Group 0, absence of fibrosis; group 1, incipient fibrosis; group 2, moderate/ severe fibrosis. Finally, in model C, in which subjects without fibrosis (group 0) were compared with individuals presenting incipient fibrosis (group 1) and with those exhibiting moderate/severe fibrosis (group 2), it was apparent that IL-13 was strongly associated with fibrosis (OR 5.8) and that TGF- was inversely associated with fibrosis (OR 0.2), as also shown in model A. It is noteworthy that no other cytokines showed significant associations with the degree of fibrosis in any of the models tested by multivariate analyses. DISCUSSION The aim of this work was to determine the immunological factors involved in the development of hepatosplenomegaly in schistosomiasis patients living in an area where S. mansoni is endemic. The production of the cytokines IFN-, IL-10, IL-4, IL-13, TGF-, and TNF- by subjects who were indicated by ultrasound examination as showing no evidence of fibrosis (group 0), exhibiting incipient fibrosis (group 1), or presenting moderate/severe fibrosis (group 2) was studied. The diagnosis of moderate/severe fibrosis by ultrasound is consistent with the pathology of biopsy fragments (1). However, there is currently a lack of pathology confirming the ultrasound interpretation of incipient fibrosis, and such indications may be considered to represent either initial fibrosis or an inflammatory cellular infiltration focus induced by schistosome egg antigens. For this reason, potential associations between cytokine levels and the degree of fibrosis were compared by considering three different models involving the three defined groups. If data were not grouped by fibrosis, IFN- and IL-4 were found to be higher in the age group of 50 years; IFN-, IL-4, TNF-, and IL-13 were significantly higher in individuals without other parasitic infections; and IFN-, TNF-, and IL-13 showed higher median values in egg-negative individuals than in egg-positive individuals, but IL-10 gave the opposite result. However, in multivariate analysis where all variables were corrected for, only IL-13 production correlated with fibrosis. In the present study, the group exhibiting incipient fibrosis constituted a younger population than that presenting moderate to severe fibrosis (Table 1), thus suggesting a progression of the disease from group 1 to group 2. The results of univariate analysis indicated that moderate/severe fibrosis was more frequent in subjects over 50 years old (P 0.01) and especially in males rather than females (P 0.02), which confirms previous findings (3, 4, 17, 29). However, the progression of pathology in schistosomiasis patients is not linear and may also be influenced by the genetic backgrounds of the individuals (6, 29). It is interesting to note from the univariate analysis that the intensity of infection (eggs/g of feces) was highest in the group with incipient fibrosis (group 1). This suggests that while intensity of infection may be sufficient to trigger the process of fibrosis, it is apparently not necessary for progression to severe fibrosis, a process that clearly depends on a number of other important factors. The intensities of infection and degrees of fibrosis apparently fluctuate considerably between different studies, and this may explain why a positive association between them has been found in some investigations (29) but not in others (4, 33). In the present work, it is shown that parameters including age, gender, and intensity of infection are confounders, and all must be considered in order to establish potential associations between the production of cytokines and periportal fibrosis. It has previously been considered that risk of fibrosis was associated with different cytokine profiles, and depended upon both age and gender. Booth et al. (4) concluded that children presented the lowest overall risk of fibrosis, but that this was probably because they had not been exposed long enough for lesions to produce this effect. However, these authors reported that low production of IL-10 was a common factor in all cases where fibrosis was detected in a child. The results of the present study demonstrate that 72.7% of patients with moderate/severe fibrosis (group 2) showed low levels of IL-10 (Table 3) and that the median and mean values of IL-10 production for this group were significantly lower than those for the group composed of subjects without fibrosis (P 0.027) (Table 2; Fig. 1). However, this has been observed only in univariate analysis. Hoffmann et al. (19) established that mice deficient in IL-10 were associated specifically with an increased risk of hepatic fibrosis, while others have shown that IL-10 is important in reducing the pathology of acute schistosomiasis and that the level of IL-10 decreases in patients with hepatosplenomegaly (12, 26, 30). However, in the present study, when the cytokine profiles were adjusted for other variables, the association of fibrosis and IL-10 was no longer observed. In contrast, the present data demonstrate a strong association between high levels of production of IL-13 and the development of severe fibrosis (P 0.029; OR 5.8) and serve to reinforce the importance of this cytokine in human fibrosis. Moreover, increased levels of IL-13 were observed in PBMC that had been stimulated by SEA but not in those stimulated by SWAP, indicating that the response was primarily directed towards egg antigens. An important role for IL-13 in the development of liver fibrosis in mice and in humans has been pointed out by a number of investigators (4, 7, 8, 9, 13, 33). It is known that IL-13 promotes collagen production through at least three distinct, but overlapping, mechanisms (for a review

6 1220 ALVES OLIVEIRA ET AL. INFECT. IMMUN. see reference 33), and the expression in mice of the decoy receptor IL-13Ra2, which blocks IL-13 function, has been shown to also curtail hepatic fibrosis (9, 23). Interestingly, increased levels of IL-13Ra2 have been detected in the sera of individuals heavily infected with S. mansoni and living in a region of endemicity (28). The present study provides some evidence that TGF- may be associated with protection against fibrosis, and the covariate was maintained in models A and C because it gave a better fit in both cases. However, for TGF-, the confidence interval of the OR was close to unity, and hence the strength of its association with fibrosis requires further investigation. TGF- is a pleiotropic cytokine that is known to be involved in fibrosis by virtue of its ability to induce collagen deposition (36). In the present study, high levels of TGF- were apparently associated with a low level of fibrosis in model A (comparison of group 0 [no fibrosis] with groups 1 plus 2 [incipient and moderate/ severe fibrosis]; OR 0.3) and in model C (comparison of group 0 [no fibrosis] with group 2 [moderate/severe fibrosis]; OR 0.2). Since OR provides a good estimate of relative risk, it does appear that TGF- plays a protective role in the development of fibrosis. In a recent study (33), however, no differences in TGF- levels in SEA-stimulated PBMC supernatants were detected between groups of patients with different degrees of hepatic fibrosis. Indeed, these authors suggested the occurrence of an intermittent induction of fibrosis by TGF- : the intermittent production of this cytokine has been demonstrated in baboons infected with S. mansoni, which suggested an association with fibrosis (14). Nevertheless, several reports indicate that TGF- is a regulatory cytokine that is produced by regulatory T cells and provides an effective mechanism of control of the progression of fibrosis in association with IL-10 (18, 24, 32). The results of the present work highlight the high production of IL-13 associated with moderate/severe fibrosis and the association of intensity of infection with the incidence of incipient fibrosis in subjects living in an area where schistosomiasis is endemic. Furthermore, our data show that TGF- is not positively associated with fibrosis; rather, they support the hypothesis of a fibrogenesis pathway that is dependent on IL-13 but not dependent on TGF- (22). The major limitation of the study, however, was that cause and effect could not be separated, because exposure and disease were measured at the same time. Thus, the experimental design could not determine, for example, whether high levels of IL-13 favored the development of fibrosis or whether the PBMC of individuals with severe fibrosis produced larger amounts of IL-13. However, a follow-up of this population will be conducted in order to evaluate the incidence of periportal fibrosis, the factors involving the development of fibrosis, and the cytokine profile associated with regression of fibrosis after treatment, and also to analyze other potential confounders including frequency of water contact, time of residence in areas of endemicity, and infection with schistosomes alone compared to coinfection with other helminths. ACKNOWLEDGMENTS This work was supported by grants from Programa de Apoio a Grupos de Excelência-PRONEX/CNPq/FINEP (Brazil), National Institutes of Health NIH-ICIDR/grant AI , Fogarty Global Infectious Diseases Training grant D43TW6580, Fundação de Amparo a Pesquisa do Estado de Minas Gerais-FAPEMIG (Brazil), Conselho Nacional de Desenvolvimento Científico e Tecnológico-CNPq (Brazil), and the UNDP World Bank/WHO Special Program for Research and Training in Tropical Diseases. We thank Maria de Fátima da Silva, Marlucy Rodrigues Lima, Lilia Cardoso Moreira, and Ivanete dos Santos Nascimento (UNIVALE, Governador Valadares, Minas Gerais, Brazil) for their technical assistance in various aspects of the field and laboratory work. REFERENCES 1. Abdel-Wahab, M. F., G. Esmat, A. Farrag, Y. A. el-boraey, and G. T. Strickland Grading of hepatic schistosomiasis by the use of ultrasonography. Am. J. Trop. Med. Hyg. 46: Arap Siongok, T. K., A. A. Mahmoud, J. H. Ouma, K. S. Warren, A. S. Muller, A. K. Handa, and H. B. Houser Morbidity in schistosomiasis mansoni in relation to intensity of infection: study of a community in Machakos, Kenya. Am. J. Trop. Med. Hyg. 25: Booth, M., B. J. Vennervald, N. B. Kabatereine, F. Kazibwe, J. H. Ouma, C. H. Kariuki, E. Muchiri, H. Kadzo, E. Ireri, G. Kimani, J. K. Mwatha, and D. W. Dunne Hepatosplenic morbidity in two neighboring communities in Uganda with high levels of Schistosoma mansoni infection but very different durations of residence. Trans. R. Soc. Trop. Med. Hyg. 98: Booth, M., J. K. Mwatha, S. Joseph, F. M. Jones, H. Kadzo, E. Ireri, F. Kazibwe, J. Kemijumbi, C. Kariuki, G. Kimani, J. H. Ouma, N. B. Kabatereine, B. J. Vennervald, and D. W. Dunne Periportal fibrosis in human Schistosoma mansoni infection is associated with low IL-10, low IFN-, high TNF-, or low RANTES, depending on age and gender. J. Immunol. 172: Boros, D. L., and K. S. Warren Delayed hypersensitivity-type granuloma formation and dermal reaction induced and elicited by a soluble factor isolated from Schistosoma mansoni eggs. J. Exp. Med. 132: Chevillard, C., C. E. Moukoko, N. E. Elwali, J. H. Bream, B. Kouriba, L. Argiro, S. Rahoud, A. Mergani, S. Henri, J. Gaudart, Q. Mohamed-Ali, H. A. Young, and A. J. Dessein IFN-gamma polymorphisms (IFN-gamma 2109 and IFN-gamma 3810) are associated with severe hepatic fibrosis in human hepatic schistosomiasis (Schistosoma mansoni). J. Immunol. 171: Chiaramonte, M. G., A. W. Cheever, J. D. Malley, D. D. Donaldson, and T. A. Wynn Studies of murine schistosomiasis reveal interleukin-13 blockade as a treatment for established and progressive liver fibrosis. Hepatology 34: Chiaramonte, M. G., L. R. Schopf, T. Y. Neben, A. W. Cheever, D. D. Donaldson, and T. A. Wynn IL-13 is a key regulatory cytokine for Th2 cell-mediated pulmonary granuloma formation and IgE responses induced by Schistosoma mansoni eggs. J. Immunol. 162: Chiaramonte, M. G., M. Mentink-Kane, B. A. Jacobson, A. W. Cheever, M. J. Whitters, M. E. Goad, A. Wong, M. Collins, D. D. Donaldson, M. J. Grusby, and T. A. Wynn Regulation and function of the interleukin 13 receptor alpha 2 during a T helper cell type 2-dominant immune response. J. Exp. Med. 197: Conselho Nacional de Saúde Comissão Nacional de Ética em Pesquisa, Resolução 196/96. [Online.] Czaja, M. J., F. R. Weiner, S. Takahashi, M. A. Giambrone, P. H. van der Meide, H. Schellekens, L. Biempica, and M. A. Zern Gammainterferon treatment inhibits collagen deposition in murine schistosomiasis. Hepatology 10: Falcão, P. L., L. C. C. Malaquias, O. A. Martins-Filho, A. M. S. Silveira, M. Passos, A. Prata, G. Gazzinelli, R. L. Coffman, and R. Correa-Oliveira Human schistosomiasis mansoni: IL-10 modulates the in vitro granuloma formation. Parasite Immunol. 20: Fallon, P. G., E. J. Richardson, G. J. McKenzie, and A. N. McKenzie Schistosome infection of transgenic mice defines distinct and contrasting pathogenic roles for IL-4 and IL-13: IL-13 is a profibrotic agent. J. Immunol. 164: Farah, I. O., P. W. Mola, T. M. Kariuki, M. Nyindo, R. E. Blanton, and C. L. King Repeated exposure induces periportal fibrosis in Schistosoma mansoni-infected baboons: role of TGF-beta and IL-4. J. Immunol. 164: Gazzinelli, G., N. Katz, R. S. Rocha, and D. G. Colley Immune response during human schistosomiasis mansoni. X. Production and standardization of an antigen-induced mitogenic activity by peripheral blood mononuclear cells from treated, but not active cases of schistosomiasis. J. Immunol. 130: Gryseels, B The epidemiology of schistosomiasis in Burundi and its consequences for control. Trans. R. Soc. Trop. Med. Hyg. 85: Henri, S., C. Chevillard, A. Mergani, P. Paris, J. Gaudart, C. Camilla, H. Dessein, F. Montero, N. E. Elwali, O. K. Saeed, M. Magzoub, and A. J. Dessein Cytokine regulation of periportal fibrosis in humans infected with Schistosoma mansoni: IFN-gamma is associated with protection against

7 VOL. 74, 2006 CYTOKINE PRODUCTION AND FIBROSIS IN SCHISTOSOMIASIS 1221 fibrosis and TNF-alpha with aggravation of disease. J. Immunol. 169: Hesse, M., C. A. Piccirillo, Y. Belkaid, J. Prufer, M. Mentink-Kane, M. Leusink, A. W. Cheever, E. M. Shevach, and T. A. Wynn The pathogenesis of schistosomiasis is controlled by cooperating IL-10-producing innate effector and regulatory T cells. J. Immunol. 172: Hoffmann, K. F., A. W. Cheever, and T. A. Wynn IL-10 and the dangers of immune polarization: excessive type 1 and type 2 cytokine responses induce distinct forms of lethal immunopathology in murine schistosomiasis. J. Immunol. 164: Hosmer, D. W., Jr., and S. Lemeshow Applied logistic regression. John Wiley & Sons, New York, N.Y. 21. Katz, N., A. Chaves, and J. Pellegrino A simple device for quantitative stool thick smear technique in Schistosoma mansoni. Rev. Inst. Med. Trop. São Paulo 14: Kaviratne, M., M. Hesse, M. Leusink, A. W. Cheever, S. J. Davies, J. H. McKerrow, L. M. Wakefield, J. J. Letterio, and T. A. Wynn IL-13 activates a mechanism of tissue fibrosis that is completely TGF-beta independent. J. Immunol. 173: Kawakami, K., J. Taguchi, T. Murata, and R. K. Puri The interleukin-13 receptor alpha2 chain: an essential component for binding and internalization but not for interleukin-13-induced signal transduction through the STAT6 pathway. Blood 97: Kitani, A., I. Fuss, K. Nakamura, F. Kumaki, T. Usui, and W. Strober Transforming growth factor (TGF)- 1-producing regulatory T cells induce Smad-mediated interleukin 10 secretion that facilitates coordinated immunoregulatory activity and amelioration of TGF- 1-mediated fibrosis. J. Exp. Med. 198: Magalhães, T. V., G. Gazzinelli, M. C. Alvarez, F. C. Lima e Silva, L. A. Fraga, A. M. Silveira, A. Gazzinelli, J. Bethony, P. LoVerde, I. R. Caldas, R. Correa-Oliveira, and A. Prata Comparative clinical and ultrasound study of egg-negative and egg-positive individuals from Schistosoma mansoni low morbidity endemic areas, and hospitalized patients with hepatosplenic disease. Rev. Soc. Bras. Med. Trop. 38: Malaquias, L. C. C., P. L. Falcão, A. M. S. Silveira, G. Gazzinelli, A. Prata, R. L. Coffmann, V. R. Pizziolo, and R. Corrêa-Oliveira Cytokine regulation of human immune response to Schistosoma mansoni: analysis of the role of IL-4, IL-5 and IL-10 on peripheral blood mononuclear cell responses. Scand. J. Immunol. 46: Marguerite, M., M. C. Gallissot, M. Diagne, C. Moreau, M. M. Diakkhate, M. Roberts, F. Remoue, A. Thiam, C. H. Decam, F. Rogerie, F. Cottrez, J. L. Neyrink, A. E. Butterworth, R. F. Sturrock, J. P. Piau, B. Daff, M. Niang, I. Wolowczuk, G. Riveau, C. Auriaut, and A. Capron Cellular immune responses of a Senegalese community recently exposed to Schistosoma mansoni: correlations of infection level with age and inflammatory cytokine production by soluble egg antigen-specific cells. Trop. Med. Int. Health 4: Mentink-Kane, M. M., A. W. Cheever, R. W. Thompson, D. M. Hari, N. B. Kabatereine, B. J. Vennervald, J. H. Ouma, J. K. Mwatha, F. M. Jones, D. D. Donaldson, M. J. Grusby, D. W. Dunne, and T. A. Wynn IL-13 receptor alpha 2 down-modulates granulomatous inflammation and prolongs host survival in schistosomiasis. Proc. Natl. Acad. Sci. USA 101: Mohamed-Ali, Q., N. E. Elwali, A. A. Abdelhameed, A. Mergani, S. Rahoud, K. E. Elagib, O. K. Saeed, L. Abel, M. M. Magzoub, and A. J. Dessein Susceptibility to periportal (Symmers) fibrosis in human Schistosoma mansoni infections: evidence that intensity and duration of infection, gender, and inherited factors are critical in disease progression. J. Infect. Dis. 180: Montenegro, S. M., P. Miranda, S. Mahanty, F. G. C. Abath, K. M. Teixeira, E. M. Coutinho, J. Brinkman, I. Gonçalves, L. A. W. Domingues, A. L. Domingues, A. Sher, and T. A. Wynn Cytokine production in acute versus chronic human schistosomiasis mansoni: the cross-regulatory role of interferon- and interleukin-10 in the responses of peripheral blood mononuclear cells and splenocytes to parasite antigens. J. Infect. Dis. 179: Mwatha, J. K., G. Kimani, T. Kamau, G. G. Mbugua, J. H. Ouma, J. Mumo, A. J. C. Fulford, F. M. Jones, A. E. Butterworth, M. B. Roberts, and D. W. Dunne High levels of TNF, soluble TNF receptors, soluble ICAM-I, and IFN-gamma, but low levels of IL-5, are associated with hepatosplenic disease in human schistosomiasis mansoni. J. Immunol. 160: Remoue, F., D. To Van, A. M. Schacht, M. Picquet, O. Garraud, J. Vercruysse, A. Ly, A. Capron, and G. Riveau Gender-dependent specific immune response during chronic human schistosomiasis haematobia. Clin. Exp. Immunol. 124: Ribeiro de Jesus, A., A. Magalhães, D. G. Miranda, R. G. Miranda, M. I. Araujo, A. A. de Jesus, A. Silva, L. B. Santana, E. Pearce, and E. M. Carvalho Association of type 2 cytokines with hepatic fibrosis in human Schistosoma mansoni infection. Infect. Immun. 72: Silveira, A. M., G. Gazzinelli, L. F. Alves-Oliveira, J. Bethony, A. Gazzinelli, C. Carvalho-Queiroz, M. C. Alvarez, F. C. Lima-Silva, A. Prata, P. T. LoVerde, and R. Correa-Oliveira Human schistosomiasis mansoni: intensity of infection differentially affects the production of interleukin-10, interferon-gamma and interleukin-13 by soluble egg antigen or adult worm antigen stimulated cultures. Trans. R. Soc. Trop. Med. Hyg. 98: Wynn, T. A Fibrotic disease and the T(H)1/T(H)2 paradigm. Nat. Rev. Immunol. 4: Zhu, H., L. Zeng, D. Zhu, and Y. Yuan The role of TGF-beta 1 in mice hepatic fibrosis by schistosomiasis japonica. J. Tongji Med. Univ. 20: , 329. Editor: J. F. Urban, Jr.