Skip Navigation

Increased Ratio of Tumor Necrosis Factor-α to Interleukin-10 Production Is Associated with Schistosoma haematobium-Induced Urinary-Tract Morbidity

  1. Alex N. Wamachi1,
  2. Jyoti S. Mayadev4,
  3. Peter L. Mungai3,
  4. Phillip L. Magak3,
  5. John H. Ouma1,
  6. Japhet K. Magambo2,
  7. Eric M. Muchiri3,
  8. Davy K. Koech1,
  9. Charles H. King4 and
  10. Chistopher King4,1
  1. 1Kenya Medical Research Institute, Nairobi, Kenya
  2. 1Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya
  3. 1Division of Vector Borne Diseases, Nairobi, Kenya
  4. 1Center for Global Health and Diseases, Case Western Reserve University School of Medicine, Cleveland, Ohio
  1. Reprints or correspondence: Dr. Christopher L. King, Center for Global Health and Disease, Case Western Reserve University School of Medicine, Rm. WRC 4132, 2103 Cornell Rd., Cleveland, OH 44106-7286 (cxk21{at}case.edu).

  1. Presented in part: 12th National Institutes of Allergy and Infectious Diseases International Centers for Tropical Disease Research, Bethesda, MD, 13–15 May 2003.

 
Next Section

Abstract

Bladder and kidney disease, which affect ∼25%–C30% of subjects infected with Schistosoma haematobium, are mediated by T cell-dependent granulomatous responses to schistosome eggs. To determine why only some infected subjects develop disease, we examined the hypothesis that infected Kenyan subjects with ultrasounddetected urinary-tract morbidity (n = 49) had dysregulated cytokine production leading to enhanced granulomatous responses, compared with subjects of similar age and intensity of infection without morbidity (n = 100). Peripheral blood mononuclear cells from subjects with morbidity produced 8-fold greater levels of egg antigen-driven tumor necrosis factor (TNF)-α and had a 99-fold greater mean TNF-α:interleukin (IL)- 10 ratio, compared with subjects without disease. No differences in cytokine response to non-Cegg-derived schistosome antigens were observed between groups. Subjects with morbidity had increased TNF-α production in response to endotoxin, suggesting an innate hyperresponsiveness. These results indicate that increased TNF-α production, relative to that of IL-10, is associated with developing bladder-wall morbidity with S. haematobium infection.

Schistosoma haematobium infects >111 million people worldwide, causing significant urinary-tract morbidity [1]. More than 90% of these infections occur in sub- Saharan Africa. Despite effective chemotherapy, schistosomiasis continues to be endemic in many parts of the developing world because of continuous reexposure. The need to develop complementary control strategies provides an incentive to better understand the host regulation of infection and disease. The spectrum of disease ranges from mild asymptomatic hematuria to advanced disease manifested by hydronephrosis, kidney failure, or squamous-cell carcinoma of the bladder [2]. Most infected individuals, however, do not develop severe disease. Those individuals who do develop disease are more likely to have heavier infections of longer duration [3–5] and an exaggerated host immune response to parasite antigens [6]. The relative contributions of these various risk factors for disease are not known, although such information is of potentially great importance because of the impending expansion of mass treatment and/or eradication programs for schistosomiasis in Africa. Treatment programs will decrease the burden of infection but may not produce consistent reductions in disease. Studies of Schistosoma mansoni infection in mice, baboons, and humans have demonstrated that chronic persistent infection downmodulates host immune responses, thereby limiting immunopathology [7–10]. Treatment may clear infection, but it may also interrupt host immune modulation. As a consequence, intermittent treatment of Schistosoma japonicum or S. mansoni infection has been shown to increase the risk for disease in some individuals [11, 12]. Whether similar mechanisms govern the degree of immune-mediated pathology associated with S. haematobium, the most prevalent human schistosome infection, is not well understood.

S. haematobium infection provides an excellent system to better understand the early stages of granulomatous inflammation and morbidity, because most adult worms reside in the relatively small confines of the vesiculourinary venous plexus, and inflammatory changes are readily detectable by ultrasonography in the bladder wall. Bladder morbidity coincides with early acquisition of infection in children [2]. Adult S. haematobium worms mature in the portal circulation and then migrate to the venous plexus in the bladder via the rectal inferior iliac veins. There, the worms mate and produce ova. Antigens produced by viable ova in tissues stimulate a hostdriven granulomatous inflammation that disrupts the epithelial barrier of the bladder, which permits ova to move from the venule into the bladder lumen [13]. This inflammatory process causes varying degrees of collateral damage to tissues, and the extent of this damage depends on the intensity of the granulomatous response and its anatomic location. For example, inflammation at the site where the ureter enters the bladder may obstruct the ureter, leading to a dilation of the ureter, calyxes of the kidney, and eventual kidney failure [14].

The granuloma is the source of immunologically driven pathology in schistosomiasis [15]. Its initial formation and subsequent modulation is CD4+ T cell dependent and is tightly controlled via a complex series of cytokines and chemokines produced by monocytes and lymphocytes and, eventually, eosinophils [16, 17]. Our central hypothesis is that newly exposed and/or infected children develop a population of schistosome egg antigen-reactive lymphocytes with varying patterns of cytokine responses that affect the magnitude of granulomatous inflammation and risk for disease with subsequent infection. Specifically, tumor necrosis factor (TNF)-α is the decisive proinflammatory mediator essential for granuloma formation [18]. Analogously, interleukin (IL)-10, a potent inhibitor of TNF-α, is critical for the down-modulation of granulomatous inflammation [19]. Recently, we have shown that children and adolescents are more likely to develop bladder-wall pathology if their peripheral blood mononuclear cells (PBMCs) produce more TNF-α relative to IL-10 in response to egg antigens [6]. These observations, however, were limited by small sample size and the restriction of cases only to those with advanced bladder pathology.

The present study extends these findings to a larger population of S. haematobium-infected children, with a greater range of bladder pathology, to evaluate whether the magnitude of the production of egg antigen-driven TNF-α, relative to that of IL-10, is associated with increasing severity of bladder disease. The study was based on a cross-sectional survey of all children and adolescents in a community endemic for S. haematobium in Kenya, to better understand the early immunological events associated with disease. We examined all children identified with bladder-wall pathology who were matched with children of similar age and intensity of infection but without observable bladder pathology. To further understand the basis for increased TNF-α production, relative to that of IL-10, among infected children with disease, we also examined whether these children showed a similar bias in their innate immune response as indicated by increased lipopolysaccharide (LPS)-driven TNF-α production relative to that of IL-10.

Previous SectionNext Section

Subjects, Materials, and Methods

Study area and selection of study population. Study subjects were selected from a cross-sectional survey conducted 3–6 months before the current study, as described elsewhere [20], in which all subjects were examined by ultrasound for S. haematobium infection and associated morbidity in the Msambweni Division of the Kwale district, southern Coast Province, Kenya, an area where S. haematobium is endemic. No community- based treatment for schistosomiasis had been conducted during the preceding 8 years in this population. From this initial survey, we selected all children 5–18 years old residing in 2 villages, Vidungeni and Marigiza, who had detectable bladder pathology and S. haematobium infection. These 2 villages were selected because of the willingness of the communities to participate. Infected children with morbidity were matched to 2 (sometimes 3) children without bladder-wall abnormalities who were of similar age (±1 year) and intensity of infection (±20% of egg counts), on the basis of data collected in the initial survey. Three children selected as control subjects did not participate. One child had moved out of the community, and 2 others refused to participate. Ultrasonography for morbidity and intensity of infection was repeated for study subjects (table 1). Exclusion criteria were pregnancy, unwillingness to participate in the study, or plan to move from the area within the next 2 years. All subjects identified with schistosomiasis were treated with praziquantel (40 mg/kg) immediately after we collected peripheral blood for immunological studies. The study protocol was approved by the institutional review board at Case Western Reserve University and the scientific review board and ethical review committee of the Kenya Medical Research Institute. Informed consent was obtained from all participants. Permission for publishing the present study was provided by the Director of the Kenya Medical Research Institute, Kenya.

Measurement of urinary-tract morbidity. Images of the bladder and urinary tract were obtained by use of a portable ultrasound machine (SonoSite 180; SonoSite) and were recorded on a thermal printer (figure 1). These were subsequently scored, by 2 experienced radiologists at separate readings, for renal size, irregularities of the kidney lumen, hydronephrosis, hydroureter, and the number and severity of areas in the bladder wall that were irregular or thickened (in millimeters). Both readings were blinded to the previous readings and the patient's current and previous infection status. The scoring of the severity of bladder-wall morbidity was slightly modified on the basis of a classification described elsewhere [21, 22] and similar to that of King et al. [6] and was scored on the basis of the extent of bladder-wall pathology: grade 0, bladder-wall thickness <5 mm; grade 1, bladder-wall thickness 5–7 mmor any areas of bladderwall irregularity; grade 2, maximum bladder-wall thickening of 8–9 mm with ⩾2 areas of involvement; grade 3, maximum bladder-wall thickening of ⩾10 mm with ⩾1 areas of involvement or any polyp, mass, or tumor. Grade 1 hydroureter was defined on the basis of World Health Organization (WHO) criteria, in which dilation of the ureter can be visualized (e.g., <1 cm) either in the proximal portion or in the distal third. Grade 2 hydroureter was defined by gross dilation, with the maximal point of dilation >1 cm. Grade 1 hydronephrosis was defined as moderate dilation of the renal pelvis (>1 cm) but conservation of renal parenchyma (e.g., >1 cm, measured as the distance between the renal pelvis and capsula). Grade 2 hydronephrosis was defined as dilation of the renal pelvis accompanied by diminished renal parenchyma (e.g., <1 cm, the distance between the renal pelvis and capsula).

Figure 1.
Figure 1.

Different ultrasound images of Schistosoma haematobium- infected children (7–11 years old) with a bladder mass (top), hydronephrosis with dilated calyxes (arrow, middle panel), and hydroureter showing dilated ureters in the posterior bladder wall (bottom). The distance between each white dot on the right of ultrasound images represents 1 cm.

The presence and intensity of infection was determined by filtering 2 10-mL aliquots of clean-caught urine collected through polycarbonate filters (Nucleopore) between 10:00 am and 2:00 pm on 2 separate days. The number of ova was counted on each filter, and the average number of eggs was calculated per milliliter of urine, as described elsewhere [23]. The case and control subjects were selected and transported from their village to the nearby Msambweni District Hospital. There, the subjects underwent a physical exam and donated 10 mL of venous blood. The blood was used for the immunological assays, and a blood smear was prepared to determine the presence of Plasmodium falciparum infection. An aliquot of plasma was examined for the presence of Wuchereria bancrofti as determined by the Og4C3 antigen-detection assay [24]

Antigens and mitogens. S. haematobium worm antigen (SWAP) and S. haematobium egg antigen (SEA) were prepared as saline extracts of adult-stage parasites or from freshly obtained ova [25]. The soluble antigen preparations were passed through Detoxigel affinity columns (Pierce Chemical) to remove endotoxin. Endotoxin in these preparations was <0.5 ng/ mL, 5–50-fold less than that required for LPS stimulation of cytokine production by human lymphocytes [26]. Preparations of crude parasite antigens did not stimulate lymphocyte proliferation or cytokine production by PBMCs or by enriched CD14+ cells prepared by immunomagnetic negative selection from healthy, North American donors. Purified protein derivative (PPD) was prepared by culture supernatants of Mycobacterium tuberculosis purchased from Evans Medical. The mitogens phorbol 12-myristate-13-acetate (PMA) and ionomycin were purchased from Calbiochem.

Isolation of PBMCs and culture conditions for cytokine analysis. PBMCs were isolated from heparinized venous blood and separated by density-gradient centrifugation by use of Ficoll- Paque Plus (Amersham Pharmacia Biotech). The PBMCs were then resuspended in RPMI 1640 supplemented with 10% fetal calf serum, 4 mmol/L L-γ-glutamine, 25 mmol/L HEPES, and 40 εg/mL gentamicin (C-RPMI; BioWhittaker). PBMCs were then cultured in duplicate at 2×106 cells/mL in C-RPMI to a total volume of 1 mL in the presence or absence of antigens or mitogens. The following antigens or mitogens were added into the cell cultures: SWAP (50 εg/mL), SEA (10 εg/mL), PPD (1:200), PMA (50 ng/mL) plus ionomycin (1 εg/mL), phytohemagglutinin (PHA; 2 εg/mL), and LPS (10 ng/mL). Cells were incubated at 37°C in 5% CO2, and supernatants were collected at 48 and 96 h. Supernatants from cultures stimulated with LPS were harvested at 6 and 24 h. Culture supernatants were immediately frozen at −70°C for the later determination of cytokine production. The collection times were chosen on the basis of previous results of kinetic studies for optimal antigen- driven cytokine production.

Cytokine measurement. Antigen-driven cytokine production by PBMCs was determined by ELISA, as described elsewhere [6].

Data analysis. Cytokine production was log-normally distributed; therefore, the data were log transformed, including the ratios of IL-10 to TNF-α, before statistical analysis. Differences between means were determined by the paired Student's t test (2-tailed), comparing log-transformed cytokine levels in subjects with disease and their matched control subjects. Differences between proportions were analyzed by x2 test. P < .05 was considered to be significant.

Previous SectionNext Section

Results

Study population. The overall prevalence of S. haematobium infections in subjects of all ages in a sublocation of Msambweni, an administrative unit that included the 2 study villages, was 46% (1462/3181). Among those infected, 27% (392/1462) had evidence of bladder-wall disease, and 7.5% had evidence of hydroureter and/or hydronephrosis. The overall prevalence of bladder-wall disease among infected children and adolescents (5–18 years old; 180/825 [22%]) was lower than that of infected adults (>18 years old; 212/637 [33%]; x2 = 24.1 P < .001). In the 2 study villages within the Msambweni area, 49 (15%) of 322 infected children and adolescents (5–18 years old) had morbidity, and all agreed to participate in the study (table 1). Bladder disease was defined as ultrasound-detectable bladderwall thickness >5.0 mm in ⩾1 area of the bladder on the basis of WHO criteria [21]. The degree of bladder disease ranged from mild, grade 1 (n = 17); to moderate, grade 2 (n = 21); to severe, grade 3 (n = 9). Two additional subjects were also classified as case subjects on the basis of the presence of irregularities in the posterior bladder wall (where the ureter enters the bladder) leading to advanced hydronephrosis or hydroureter, even though the bladder-wall thickness was <5 mm.

The presence of mild hydronephrosis or hydroureter (grade 1) was also observed in 12 subjects—8 who also had bladderwall pathology and 4 who did not. The 4 subjects with mild hydronephrosis and no detectable bladder-wall disease were included in the group without disease because they did not have any evidence of hydronephrosis during the initial ultrasound screening 3 months earlier. Grade 1 hydronephrosis and hydroureter in the absence of detectable bladder-wall disease may not be a consequence of urinary schistosomiasis [21, 27, 28] and can be transient (P.L.M., unpublished data). Additional characteristics of subjects with and without bladder-wall morbidity are shown in table 1, and they show similar age, intensity of infection, and rates of coinfection with malaria and W. bancrofti. In summary, the proportion of infected children and adolescents with urinary-tract morbidity was lower than that of infected adults, with many heavily infected subjects showing no evidence of urinary-tract disease.

Relationship of bladder-wall pathology with parasite-antigen- driven TNF-α and IL-10 production. TNF-α is essential for hepatic granuloma formation caused by S. mansoni infection [18], and its increased production by SEA-specific lymphocytes may be associated with exaggerated granulomatous responses to ova trapped in the bladder wall, producing abnormalities that are detectable by ultrasound. Net SEA-driven TNF-α production was 8-fold higher in subjects with bladder pathology, compared with those without detectable disease (figure 2A). This difference was specific to egg antigens—SWAP and PPD stimulated equivalent amounts of TNF-α between the 2 groups (figure 2A). Spontaneous TNF-α production was similar between subjects with disease (geometric mean, 8 pg/mL [95% confidence interval {CI|, 3–20]) and those without (geometric mean, 13 pg/mL [95% CI, 5–33]).

Figure 2.
Figure 2.

Net parasite antigen (antigen minus spontaneous)-driven tumor necrosis factor (TNF)-α production by peripheral blood mononuclear cells from children with (n = 49) and without (n = 100) bladder- wall morbidity, collected 96 h after antigen stimulation for all 149 subjects studied (A). B, net parasite antigen-driven cytokine production in children with moderate to mild bladder-wall pathology (grade 1–2; n = 38) and severe disease (grade 3; n = 9). Bars, geometric means with 95% confidence intervals. ***P < .001, subjects with vs. subjects without bladder-wall disease (A) or severity of disease (B) (paired Student's t test of log-transformed data). PHA, phytohemagglutinin; PPD, purified protein derivative; SEA, S. haematobium egg antigen; SWAP, Schistosoma haematobium worm antigen.

The mitogens, PHA or PMA plus ionomycin, stimulated a robust response in all subjects, indicating viable lymphocytes. There was no difference between PHA-driven TNF-α production for subjects with disease, compared with those without disease (figure 2A) or in response to stimulation with PMA plus ionomycin (data not shown).

To examine the relationship between the severity of bladderwall disease and SEA-induced TNF-α production, subjects with disease were divided into those with mild to moderate pathology (grades 1–2) and those with severe pathology (grade 3) (figure 2B). Subjects with severe pathology generated 9-fold greater levels of TNF-α than those with less morbidity. There was no significant difference in SWAP, PPD, or mitogen-driven TNF-α production between the 2 groups (data not shown).

It is possible that diminished TNF-α production among control subjects may have occurred as a consequence of crossregulation by increased IL-10 levels [29]. As shown in figure 3, children without bladder-wall pathology produced ∼3-fold more SEA-driven IL-10, compared with children with bladderwall pathology (P = .01). No difference in net IL-10 production was observed between the 2 groups in response to SWAP, PPD, or PHA (figure 3). Constitutive (background) IL-10 production by PBMCs was also similar between subjects with disease (geometric mean, 253 pg/mL [95% CI, 128–501]) and those without disease (geometric mean, 255 pg/mL [95% CI, 162–402]). There was a trend toward lower SEA-driven IL-10 production by PBMCs from subjects with severe bladder morbidity (grade 3; geometric mean, 321 pg/mL [95% CI, 50–2116]), compared with those with mild disease (grade 1–2; geometric mean, 449 pg/mL [95% CI, 225–914]), although this difference was not significant. This failure to reach statistical significance may have been due to the small sample size of subjects with severe disease.

Figure 3.
Figure 3.

Net parasite antigen (antigen minus spontaneous)-driven interleukin (IL)-10 production by peripheral blood mononuclear cells from children with (n = 49) and without (n = 100) bladder-wall morbidity, collected 96 h after antigen stimulation. Bars, geometric means with 95% confidence intervals. *P < .01, subjects with vs. subjects without bladderwall disease (paired Student's t test of log-transformed data). PHA, phytohemagglutinin; PPD, purified protein derivative; SEA, Schistosoma haematobium egg antigen; SWAP, S. haematobium worm antigen.

Increased IL-10 production relative to TNF-α in response to egg antigens may protect an individual from developing bladder pathology. The mean ratio of net SEA-driven production by PBMCs of IL-10 to TNF-α for each subject was, on the average, 99-fold greater among subjects without bladder pathology, compared with subjects with pathology (figure 4; P< .0001). This increased IL-10:TNF-α ratio among controls fell into 2 groups-one that was higher relative to subjects with pathology (e.g., ratio >100) and one that was similar to subjects with pathology. These 2 control groups were different in other respects that may contribute to the observed differences in SEAdriven IL-10 and TNF-α production. Among subjects without disease, those with an IL-10:TNF-α ratio >100 were older (mean, 12.3 vs. 10.7 years; P = .03), had higher ova counts (geometric mean, 46.7 vs. 21.6 ova/mL; P = .02), and were more likely to be infected with lymphatic filariasis (11/37 [30%] vs. 3/31 [13%]; P = .07). Together, these results indicate that, among subjects with heavier S. haematobium infection or who were coinfected with other helminths, there was greater IL-10, relative to TNF-α, production.

Figure 4.
Figure 4.

Ratio of net egg antigen-driven interleukin (IL)-10/tumor necrosis factor (TNF)-α production in culture supernatants in subjects with and without bladder morbidity (P < .0001, Student's t test). Left, all subjects. Middle and right panels, IL-10:TNF-α ratios stratified by intensity of infection. Each point indicates a separate subject. Lines, geometric means.

Dependence of SEA-driven TNF-α and IL-10 production on CD4+ cells. In a subset of subjects reactive to egg antigens (n = 11; figure 5), the depletion of CD4+ T cells from PBMCs resulted in an 83% loss of SEA-driven TNF-α and a 78% loss of IL-10 production. Thus, the majority of SEA-driven TNFa and IL-10 production was supported by CD4+ T cells.

Figure 5.
Figure 5.

Effect of CD4+ T cell depletion on levels of egg antigen-driven tumor necrosis factor (TNF)-α and interleukin (IL)-10 production in 11 subjects. Bars, geometric means with 95% confidence intervals. **P< .01 and ***P < .001, cytokine production after CD4+ cell depletion from peripheral blood mononuclear cells (PBMCs) vs. without depletion cultured at 2⩾106 cells/mL. SEA, Schistosoma haematobium egg antigen.

Relationship of SWAP-driven interferon (IFN)-γ and IL- 13 with bladder-wall pathology. SEA-driven IFN-γ has been implicated in granuloma formation [16, 30, 31]. We observed no significant difference in the amount of spontaneous IFN-γ production by PBMCs from subjects with disease, compared with those without disease, 96 h after stimulation (table 2; P = .2). Net SEA- and SWAP-induced IFN-γ tended to be lower among subjects without bladder-wall pathology, compared with those with disease, although these differences were not statistically significant (P = .061 and P = .21, respectively). This is consistent with the suppressive effects of IL-10 on SWAP-driven IFN-γ production [10].

Figure 6.
Figure 6.

Lipopolysaccharide-induced tumor necrosis factor (TNF)-α and interleukin (IL)-10 production in 24-h peripheral blood mononuclear cell cultures from subjects with (solid bars; n = 49) and without (hatched bars; n = 100) morbidity. Bars, geometric means with 95% confidence intervals. *P < .05 and **P < .01.

Table 1.
Table 1.

Study population.

Table 2.
Table 2.

Parasite antigen-driven interferon (IFN)-γ and interleukin (IL)-13 production in subjects with and without bladder morbidity.

IL-13 is produced by CD4+ T cells in granulomas that induce fibrosis [32], and its enhanced production may also be associated with an increased risk for bladder-wall disease. No difference in SWAP-driven IL-13 production at 48 (data not shown) and 96 h was observed between subjects with and without bladder disease (table 2).

Increased TNF-α and IL-10 production with LPS stimulation by PBMCs from case subjects, compared with control subjects. LPS is a potent activator of CD14+ cells through engagement of Toll-like receptor-4 [33]. Because the innate immune responses help shape the adaptive immune response, enhanced LPS-driven TNF-α production may favor the differentiation of antigen-reactive T cells to a phenotype that generates more TNF-α and less IL-10. TNF-α production 6 h after LPS activation was similar between subjects with disease (geometric mean, 3844 pg/mL [95% CI, 2836–5210]) versus those without (geometric mean, 4975 pg/mL [95% CI, 4189–5908]; P = .24). The amount of net LPS-driven IL-10 detected was <60 pg/mL in all subjects examined at this time point. By contrast, LPS-driven TNF-α and IL-10 at 24 h was significantly higher in subjects with disease versus those without (figure 6; P = .04 and P = .009, respectively), which indicates an enhanced innate responsiveness in subjects with bladder-wall disease, compared with those without.

It is possible that SEA itself may have mitogenic properties that might also directly stimulate TNF-α production by antigenpresenting cells (APCs) in short-term (6 h) cultures. SEA stimulated similar amounts of TNF-α production by PBMCs in subjects with and without disease (geometric mean, 44 pg/mL [95% CI, 21–73] and 45 pg/mL [95% CI, 24–61], respectively), which indicates little mitogenic activity of SEA in study subjects.

Previous SectionNext Section

Discussion

The results of this population-based cross-sectional study, performed in a highly endemic area of Kenya, show that the magnitude of the host inflammatory responses in S. haematobium- infected children and adolescents correlates with bladder and urinary-tract morbidity as detected by ultrasound. Subjects with morbidity have increased production of TNF-α (relative to IL- 10) by PBMCs in response to SEA, compared with subjects without disease who were of similar age and intensity of infection. Children with the most severe disease had the highest levels of SEA-induced TNF-α production—51-fold greater, compared with control subjects, and 8-fold greater than those with less-severe disease. The observations confirm an earlier, smaller study of S. haematobium-infected children with severe disease who had a similar 54-fold greater production of TNF-α by PBMCs in response to SEA, compared with control subjects [6]. The current findings also show that subjects with lesssevere disease have correspondingly less SEA-induced TNF-α production and suggest that the amount of Ag-induced TNF-α contributes to the extent of bladder disease observed.

The best predictor of disease was the ratio of IL-10 to TNF-α production, which is consistent with other studies that have shown that these 2 cytokines have opposing effects on host inflammatory responses [29]. Subjects who had a ratio of >100: 1 of SEA-driven IL-10:TNF-α had a diminished risk for disease. This high ratio of IL-10:TNF-α occurred in approximately onehalf of subjects with no disease (figure 4). The remaining subjects had an IL-10:TNF-α ratio similar to that observed among subjects with disease, which indicates that other risk factors must also contribute to the development of disease. These additional risk factors likely include the duration and intensity of infection. Although the present study attempted to control for these 2 risk factors by matching on the duration and intensity of infection between case and control subjects, this matching is only approximate, given that age may be a poor surrogate for duration of infection and that the quantification of ova in urine crudely estimates infection intensity.

The high ratio of IL-10:TNF-α observed in some subjects without disease may be linked to coinfection with other helminths, such as lymphatic filariasis, which has been associated with increased IL-10 production [34, 35]. Conversely, other coinfections may participate in an exaggerated host response to SEA. Salmonella, for example can colonize the gut and tegument of adult worms [36, 37], which may contaminate ova with endotoxin and enhance the local inflammatory responses in the bladder.

We also observed an overall increased response by LPS-reactive cells in PBMC cultures (primarily monocytes and dendritic cells) among subjects with disease, compared with those without disease. This increased LPS-driven TNF-α and IL-10 production by these APCs may favor the differentiation of SEA-reactive CD4+ T cells to a phenotype with greater production of TNF-α relative to that of IL-10. The reason why some subjects develop this phenotype in APCs is unknown, but it may be related to genetic differences. A strong heritable component for infection-associated disease is unlikely in this population, however. A recent population-based pedigree study in the same and surrounding communities estimated only a 14% contribution of heritable effects for bladder disease [20].

TNF-α has been previously shown to play a critical role in granuloma formation and fibrosis that are, in turn, associated with greater morbidity. This has been shown both in murine models of S. mansoni [18, 38–40] and in S. mansoni-infected humans [41–43]. The present study, along with our previous article [6], are the first, to our knowledge, to link the pattern of host immune response to urinary-tract morbidity associated with S. haematobium infection. TNF-α is a highly pleiotrophic cytokine, with a broad spectrum of biological effects that can be linked to a robust granulomatous response. It is expressed by monocytes, as well as by both Th1- and Th2-type cells, and it can exist as a membrane or soluble form [44]. TNF-α is a key proinflammatory molecule that increases local blood flow, up-regulates adhesion molecules on the vascular endothelium, stimulates chemokine production, enhances the function of APCs, and activates monocytes. All of these factors promote the recruitment of macrophages, T and B cells, and eosinophils to viable ova in the bladder wall, to form granulomas [45, 46]. TNF-α also stimulates T cells to retain or up-regulate the expression of CCR7 [47], a chemokine receptor whose engagement by chemokines produced by macrophages in the granuloma [48] increases lymphocyte affinity of integrin binding, thereby inhibiting SEA-specific lymphocytes from leaving the granuloma. This will have the effect of further amplifying the granulomatous response.

By contrast, IL-10 inhibits many of the proinflammatory effects of TNF-α. IL-10 suppresses APC function, inhibits the expression of various adhesion molecules [29], limits the activation and proliferation of SEA-specific T cells [10], and suppresses the expression of CCR7 on T cells [49]. Because IL-10 can reduce integrin expression and binding, this allows many of the above cells to emigrate from the granuloma and limit the build-up of lymphocytes in the granuloma. Stimulation of IL-10 and TNF-α by SEA are dependent on CD4+ T cells, and our previous studies have shown that CD4+ cells are the major source of SEA-driven TNF-α and IL-10 production [6, 10]. Moreover, the stimulation of PBMCs from naive subjects or enriched CD14+ cells with our crude soluble preparations of SEAs generated <50 pg/mL of TNF-α or IL-10, which indicates little stimulation of non-T cells.

No study has directly compared ultrasound-detected abnormalities in the bladder wall with the extent of granulomatous inflammation, because of the unacceptable morbidity associated with cytoscopy and biopsy of the bladder wall in children and adolescents. However, indirect evidence supports such a relationship. First, the extent of bladder-wall thickening and irregularity correlate with the degree of granulomatous inflammation, according to autopsy studies of S. haematobium-infected subjects who died from other causes [14]. Second, urinary cytoscopy results in infected children and adults have demonstrated lesions in the bladder that are similar to those observed in the autopsy specimens [14, 50, 51]. Finally, the presence of these lesions are correlated with ultrasound-detected changes in the bladder wall [51].

An inherent difficulty in studying immune responses to human schistosomiasis is the reliance on PBMCs, which are assumed to reflect local inflammatory responses in the bladder or other tissues. Several recent studies in murine models of schistosomiasis have supported this assumption, showing that SEA-specific lymphocytes transit through egg-induced granuloma and, in the process, enter the peripheral circulation and would be present in PBMCs [45, 52].

A limitation of the present study is the cross-sectional study design that matches for age and intensity of infection between subjects with and without disease. Such matching limits the ability to independently evaluate the role of these 2 variables in contributing to morbidity and to extrapolate these findings to the overall population, because matched pairs are not randomly selected [53]. We selected this study design because previous studies had shown that the duration and intensity of infection have a strong influence on morbidity [5, 28] and because immunological studies of all 322 infected children without disease in the study villages were not feasible. Random selection of these children produced too broad a range of age and intensity of infection to allow us to retrospectively stratify on these variables.

Unlike the study of host immune responses to other human schistosome infections, the study of urinary schistosomiasis has a potential advantage for the study of local inflammatory responses, in that it involves measuring inflammatory mediators in urine. For example, urine levels of eosinophilic cationic protein, a major protein released by eosinophils, has been shown to correlate with presence of ultrasound-detected morbidity [54–56]. Granulocyte colony-stimulating factor has also been identified in urine of patients with urinary schistosomiasis [57]. We did not examine whether differences in the amounts or types of cytokines or chemokines in urine correlate with disease.

In conclusion, our studies reinforce recent reports that human T cell clones that express high levels of TNF-α relative to IL-10 are more likely to form granulomas in vitro [58]. These T cells appear to also exist in vivo, and their preferential expansion likely increases the risk for bladder-wall morbidity during S. haematobium infection. This enhanced immune responsiveness may also increase the risk for persisting morbidity following treatment and subsequent reinfection—our study of this is currently in progress. Ultimately, mass treatment programs and, possibly, vaccination, combined with changes in local ecology to reduce transmission, are the key elements for the elimination of urinary-tract morbidity associated with S. haematobium. However, a better understanding of risk factors for disease may allow more aggressive targeting of these subjects or populations with the above control measures.

Previous SectionNext Section

Acknowledgments

We appreciate the cooperation of the children and their parents in Vidungeni and Marigiza for participating in the study; the technical help of A. Omollo, E. Mzungu, S. Kesuka, R. Ngando, E. Ireri, and H. Kadzo; the help of S. Esha and V. Saidi in collecting blood samples and monitoring children in the villages; G. Watetu for preparing and maintaining the databases; and Ron Blanton and Abe Stavitsky for providing valuable comments about the manuscript.

Previous SectionNext Section

Footnotes

  • Financial support: National Institute of Allergy and Infections Disease (grant AI45473); National Institutes of Health (Fogarty Training Grant).

  • Received July 2, 2003.
  • Accepted June 18, 2004.
Previous Section
 

References

  1. 1.↵
    1. van der Werf M,
    2. de Vlas S
    . Morbidity and infection with schistosomes or soil-transmitted helminths. Rotterdam: World Health Organization and Erasmus University; 2001. p. 1-103.
  2. 2.↵
    1. Mahmoud AAF,
    2. James S
    1. King CH
    . Disease in schistosomiasis haematobia. In: Mahmoud AAF, James S, editors. Schistosomiasis. Vol. 3. London: Imperial College Press; 2001. p. 265-95.
  3. 3.↵
    1. Smith DH,
    2. Warren KS,
    3. Mahmoud AA
    . Morbidity in schistosomiasis mansoni in relation to intensity of infection: study of a community in Kisumu Kenya. Am J Trop Med Hyg 1979;28:220-9.
    Abstract/FREE Full Text
  4. 4.
    1. Christie J,
    2. Crouse D,
    3. Pineda J,
    4. Anis-Ishak E,
    5. Smith J,
    6. Kamel I
    . Patterns of Schistosoma haematobium egg distribution in the human lower urinary tract. I. Noncancerous lower urinary tracts. Am J Trop Med Hyg 1986;35:743-51.
    Abstract/FREE Full Text
  5. 5.↵
    1. King CH,
    2. Keating C,
    3. Muruka J,
    4. et al
    . Urinary tract morbidity in schistosomiasis haematobia: associations with age and intensity of infection in an endemic area of Coast Province, Kenya. AmJ Trop Med Hyg 1988;39:361-8.
    Abstract/FREE Full Text
  6. 6.↵
    1. King CH,
    2. Malhotra I,
    3. Mungai P,
    4. et al
    . Schistosoma haematobium-induced urinary tract morbidity correlates with increased tumor necrosis factor-α and diminished interleukin-10 production. J Infect Dis 2001;184:1176-82.
    Abstract/FREE Full Text
  7. 7.↵
    1. Boros D
    . The role of cytokines in the formation of the schistosome egg granuloma. Immunobiology 1994;191:441-50.
    MedlineWeb of Science
  8. 8.
    1. Fallon PG
    . Immunopathology of schistosomiasis: a cautionary tale of mice and men. Immunol Today 2000;21:29-35.
    CrossRefMedlineWeb of Science
  9. 9.
    1. Grogan JL,
    2. Kremsner PG,
    3. Deelder AM,
    4. Yazdanbakhsh M
    . Antigenspecific proliferation and interferon-γ and interleukin-5 production are down-regulated during Schistosoma haematobium infection. J Infect Dis 1998;177:1433-7.
    Abstract/FREE Full Text
  10. 10.↵
    1. King CL,
    2. Medhat A,
    3. Malhotra I,
    4. et al
    . Cytokine control of parasitespecific anergy in human urinary schistosomiasis. IL-10 modulates lymphocyte reactivity. J Immunol 1996;156:4715-21.
    Abstract
  11. 11.↵
    1. Olds G,
    2. Olveda R,
    3. Wu G,
    4. et al
    . Immunity and morbidity in schistosomiasis japonicum infection. Am J Trop Med Hyg 1996;55:121-6.
    MedlineWeb of Science
  12. 12.↵
    1. Richter J
    . Evolution of schistosomiasis-induced pathology after therapy and interruption of exposure to schistosomes: a review of ultrasonographic studies. Acta Trop 2000;77:111-31.
    CrossRefMedlineWeb of Science
  13. 13.↵
    1. Doenhoff M,
    2. Musallam R,
    3. Bain J,
    4. McGregor A
    . Studies on the hostparasite relationship in Schistosoma mansoni-infected mice: the immunological dependence of parasite egg excretion. Immunology 1978;35:771-8.
    MedlineWeb of Science
  14. 14.↵
    1. Smith JH,
    2. Christie JD
    . The pathobiology of Schistosoma haematobium infection in humans. Hum Pathol 1986;17:333-45.
    CrossRefMedlineWeb of Science
  15. 15.↵
    1. Cheever AW,
    2. Yap GS
    . Immunologic basis of disease and disease regulation in schistosomiasis. Chem Immunol 1997;66:159-76.
    MedlineWeb of Science
  16. 16.↵
    1. Mahmoud AAF,
    2. James S
    1. King CL
    . Initiation and regulation of disease in schistosomiasis. In: Mahmoud AAF, James S, editors. Schistosomiasis. Vol. 3. London: Imperial College Press; 2001. p. 213-64.
  17. 17.↵
    1. Rumbley CA,
    2. Sugaya H,
    3. Zekavat SA,
    4. El Refaei M,
    5. Perrin PJ,
    6. Phillips SM
    . Activated eosinophils are the major source of Th2-associated cytokines in the schistosome granuloma. J Immunol 1999;162:1003-9.
    Abstract/FREE Full Text
  18. 18.↵
    1. Amiri P,
    2. Locksley RM,
    3. Parslow TG,
    4. et al
    . Tumour necrosis factor alpha restores granulomas and induces parasite egg-laying in schistosomeinfected SCID mice. Nature 1992;356:604-7.
    CrossRefMedline
  19. 19.↵
    1. Sadler CH,
    2. Rutitzky LI,
    3. Stadecker MJ,
    4. Wilson RA
    . IL-10 is crucial for the transition from acute to chronic disease state during infection of mice with Schistosoma mansoni. Eur J Immunol 2003;33:880-8.
    CrossRefMedlineWeb of Science
  20. 20.↵
    1. King CH,
    2. Blanton RE,
    3. Muchiri EM,
    4. et al
    . Low heritable component of risk for infection intensity and infection-associated disease in urinary schistosomiasis among Wadigo village populations in Coast Province, Kenya. Am J Trop Med Hyg 2004;70:57-62.
    Abstract/FREE Full Text
  21. 21.↵
    1. World Health Organization
    . Ultrasound in schistosomiasis: a practical guide to the standardized use of ultrasonography for assessment of schistosomiasis-related morbidity. Based on Second InternationalWorkshop in Naimey, Niger, 1996. Geneva: World Health Organization, 2000. [Accessed March 2002]. http://www.who.int/tdr/publications/publications/pdf/ultrasound.pdf.
  22. 22.↵
    1. Medhat A,
    2. Zarzour A,
    3. Nafeh M,
    4. et al
    . Evaluation of an ultrasonographic score for urinary bladder morbidity in Schistosoma haematobium infection. Am J Trop Med Hyg 1997;57:16-9.
    Abstract/FREE Full Text
  23. 23.↵
    1. Peters PA,
    2. Warren KS,
    3. Mahmoud AA
    . Rapid, accurate quantification of schistosome eggs via nuclepore filters. J Parasitol 1976;62:154-5.
    CrossRefMedline
  24. 24.↵
    1. More SJ,
    2. Copeman DB
    . A highly specific and sensitive monoclonal antibody-based ELISA for the detection of circulating antigen in bancroftian filariasis. Trop Med Parasitol 1990;41:403-6.
    MedlineWeb of Science
  25. 25.↵
    1. Boros DL,
    2. Warren KS
    . Delayed hypersensitivity-type granuloma formation and dermal reaction induced and elicited by a soluble factor isolated from Schistosoma mansoni eggs. J Exp Med 1970;132:488-507.
    Abstract
  26. 26.↵
    1. de Waal Malefyt R,
    2. Haanen J,
    3. Spits H,
    4. et al
    . Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigen-specific human T cell proliferation by diminishing the antigen-resenting capacity of monocytes via downregulation of class II major histocompatibility complex expression. J Exp Med 1991;174:915-24.
    Abstract/FREE Full Text
  27. 27.↵
    1. Richter J,
    2. Wagatsuma Y,
    3. Aryeetey M,
    4. Feldmeier H
    . Sonographic screening for urinary tract abnormalities in patients with Schistosoma haematobium infection: pitfalls in examining pregnant women. BullWorld Health Organ 1996;74:217-21.
  28. 28.↵
    1. Subramanian AK,
    2. Mungai P,
    3. Ouma JH,
    4. et al
    . Long-term suppression of adult bladder morbidity and severe hydronephrosis following selective population chemotherapy for Schistosoma haematobium. Am J Trop Med Hyg 1999;61:476-81.
    Abstract
  29. 29.↵
    1. Moore KW,
    2. de Waal Malefyt R,
    3. Coffman RL,
    4. O'Garra A
    . Interleukin-10 and the interleukin-10 receptor. Annu Rev Immunol 2001;19:683-765.
    CrossRefMedlineWeb of Science
  30. 30.↵
    1. Chensue SW,
    2. Terebuh PD,
    3. Warmington KS,
    4. et al
    . Role of IL-4 and IFN-γamma in Schistosoma mansoni egg-induced hypersensitivity granuloma formation: orchestration, relative contribution, and relationship to macrophage function. J Immunol 1992;148:900-6.
    Abstract
  31. 31.↵
    1. Hirata M,
    2. Kage M,
    3. Hara T,
    4. Yoneda Y,
    5. Zhang M,
    6. Fukuma T
    . Schistosoma japonicum egg granuloma formation in the interleukin-4 or interferongamma deficient host. Parasite Immunol 2001;23:271-80.
    CrossRefMedlineWeb of Science
  32. 32.↵
    1. Chiaramonte MG,
    2. Donaldson DD,
    3. Cheever AW,
    4. Wynn TA
    . An IL-13 inhibitor blocks the development of hepatic fibrosis during a T-helper type 2-dominated inflammatory response. J Clin Invest 1999;104:777-85.
    MedlineWeb of Science
  33. 33.↵
    1. Takeda K,
    2. Kaisho T,
    3. Akira S
    . Toll-like receptors. Annu Rev Immunol 2003;21:335-76.
    CrossRefMedlineWeb of Science
  34. 34.↵
    1. King CL,
    2. Mahanty S,
    3. Kumaraswami V,
    4. et al
    . Cytokine control of parasite-specific anergy in human lymphatic filariasis: preferential induction of a regulatory T helper type 2 lymphocyte subset. J Clin Invest 1993;92:1667-73.
    MedlineWeb of Science
  35. 35.↵
    1. Mahanty S,
    2. Mollis SN,
    3. Ravichandran M,
    4. et al
    . High levels of spontaneous and parasite antigen-driven interleukin-10 production are associated with antigen-specific hyporesponsiveness in human lymphatic filariasis. J Infect Dis 1996;173:769-73.
    Abstract/FREE Full Text
  36. 36.↵
    1. Young SW,
    2. Higashi G,
    3. Kamel R,
    4. el-Abdin AZ,
    5. Mikhail IA
    . Interaction of salmonellae and schistosomes in host-parasite relations. Trans R Soc Trop Med Hyg 1973;67:797-802.
    CrossRefMedlineWeb of Science
  37. 37.↵
    1. LoVerde PT,
    2. Amento C,
    3. Higashi GI
    . Parasite-parasite interaction of Salmonella typhimurium and schistosoma. J Infect Dis 1980;141:177-85.
    Abstract/FREE Full Text
  38. 38.↵
    1. Adewusi OI,
    2. Nix NA,
    3. Lu X,
    4. Colley DG,
    5. Secor WE
    . Schistosoma mansoni: relationship of tumor necrosis factor-alpha to morbidity and collagen deposition in chronic experimental infection. Exp Parasitol 1996;84:115-23.
    CrossRefMedlineWeb of Science
  39. 39.
    1. Joseph AL,
    2. Boros DL
    . Tumor necrosis factor plays a role in Schistosoma mansoni egg-induced granulomatous inflammation. J Immunol 1993;151:5461-71.
    Abstract
  40. 40.↵
    1. Leptak CL,
    2. McKerrow JH
    . Schistosome egg granulomas and hepatic expression of TNF-α are dependent on immune priming during parasite maturation. J Immunol 1997;158:301-7.
    Abstract
  41. 41.↵
    1. Zwingenberger K,
    2. Irschick E,
    3. Vergetti Siqueira JG,
    4. Correia Dacal AR,
    5. Feldmeier H
    . Tumour necrosis factor in hepatosplenic schistosomiasis. Scand J Immunol 1990;31:205-11.
    CrossRefMedlineWeb of Science
  42. 42.
    1. Mwatha JK,
    2. Kimani G,
    3. Kamau T,
    4. et al
    . High levels of TNF, soluble TNF receptors, soluble ICAM-1, and IFN-γamma, but low levels of IL-5, are associated with hepatosplenic disease in human schistosomiasis mansoni. J Immunol 1998;160:1992-9.
    Abstract/FREE Full Text
  43. 43.↵
    1. Henri S,
    2. Chevillard C,
    3. Mergani A,
    4. et al
    . Cytokine regulation of periportal fibrosis in humans infected with Schistosoma mansoni: IFN-γamma is associated with protection against fibrosis and TNF-αlpha with aggravation of disease. J Immunol 2002;169:929-36.
    Abstract/FREE Full Text
  44. 44.↵
    1. Beutler BA
    . The role of tumor necrosis factor in health and disease. J Rheumatol 1999;26 (Suppl 57):16-21.
  45. 45.↵
    1. Rumbley CA,
    2. Zekavat SA,
    3. Sugaya H,
    4. Perrin PJ,
    5. Ramadan MA,
    6. Phillips SM
    . The schistosome granuloma: characterization of lymphocyte migration, activation, and cytokine production. J Immunol 1998;161:4129-37.
    Abstract/FREE Full Text
  46. 46.↵
    1. Hogan LH,
    2. Wang M,
    3. Suresh M,
    4. Co DO,
    5. Weinstock JV,
    6. Sandor M
    . CD4+ TCR repertoire heterogeneity in Schistosoma mansoni-induced granulomas. J Immunol 2002;169:6386-93.
    Abstract/FREE Full Text
  47. 47.↵
    1. Zhang W,
    2. Chen Z,
    3. Li F,
    4. et al
    . Tumour necrosis factor-alpha (TNFalpha) transgene-expressing dendritic cells (DCs) undergo augmented cellular maturation and induce more robust T-cell activation and antitumour immunity than DCs generated in recombinant TNF-αlpha. Immunology 2003;108:177-88.
    CrossRefMedlineWeb of Science
  48. 48.↵
    1. Qiu B,
    2. Frait KA,
    3. Reich F,
    4. Komuniecki E,
    5. Chensue SW
    . Chemokine expression dynamics in mycobacterial (type-1) and schistosomal (type-2) antigen-elicited pulmonary granuloma formation. Am J Pathol 2001;158:1503-15.
    MedlineWeb of Science
  49. 49.↵
    1. Takayama T,
    2. Morelli AE,
    3. Onai N,
    4. et al
    . Mammalian and viral IL-10 enhance C-C chemokine receptor 5 but down-regulate C-C chemokine receptor 7 expression by myeloid dendritic cells: impact on chemotactic responses and in vivo homing ability. J Immunol 2001;166:7136-43.
    Abstract/FREE Full Text
  50. 50.↵
    1. Abdel-Salam E,
    2. Ehsan A
    . Cytoscopic picture of Schistosoma haematobium in Egyptian children correlated to intensity of infection and morbidity. Am J Trop Med Hyg 1978;27:774-8.
    Abstract/FREE Full Text
  51. 51.↵
    1. Abdel-Wahab M,
    2. Ramzy I,
    3. Esmat G,
    4. el Kafass H,
    5. Strickland G
    . Ultrasound for detecting Schistosoma haematobium urinary tract complications: comparison with radiographic procedures. J Urol 1992;148:346-50.
    MedlineWeb of Science
  52. 52.↵
    1. Rumbley CA,
    2. Phillips SM
    . The schistosome granuloma: an immunoregulatory organelle. Microbes Infect 1999;1:499-504.
    CrossRefMedlineWeb of Science
  53. 53.↵
    1. Kelsey J,
    2. Marmot M,
    3. Stolley P,
    4. Vessey M
    1. Selvin S
    . Statistical analysis of epidemiologic data. In: Kelsey J, Marmot M, Stolley P, Vessey M, editors. Monographs in epidemiology and biostatistics. Vol. 17. New York: Oxford University Press; 1991.
  54. 54.↵
    1. Reimert CM,
    2. Mshinda HM,
    3. Hatz CF,
    4. et al
    . Quantitative assessment of eosinophiluria in Schistosoma haematobium infections: a new marker of infection and bladder morbidity. Am J Trop Med Hyg 2000;62:19-28.
    Abstract
  55. 55.
    1. Leutscher P,
    2. Ramarokoto CE,
    3. Reimert C,
    4. Feldmeier H,
    5. Esterre P,
    6. Vennervald BJ
    . Community-based study of genital schistosomiasis in men from Madagascar. Lancet 2000;355:117-8.
    MedlineWeb of Science
  56. 56.↵
    1. Leutscher PD,
    2. Reimert CM,
    3. Vennervald BJ,
    4. et al
    . Morbidity assessment in urinary schistosomiasis infection through ultrasonography and measurement of eosinophil cationic protein (ECP) in urine. Trop Med Int Health 2000;5:88-93.
    CrossRefMedlineWeb of Science
  57. 57.↵
    1. Mahmoud LA,
    2. Robinson WA,
    3. Entringer MA
    . Urinary granulocyte colony- stimulating factor in bilharziasis. Am J Trop Med Hyg 1982;31:518-21.
    Abstract/FREE Full Text
  58. 58.↵
    1. Silva-Teixeira DN,
    2. Contigli C,
    3. Goes AM
    . Cytokine profile associated to effector functions of human T cell clones specific for Schistosoma mansoni antigens. Hum Immunol 1998;59:219-24.
    CrossRefMedlineWeb of Science
| Table of Contents

This Article

  1. J Infect Dis. (2004) 190 (11): 2020-2030. doi: 10.1086/425579
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most