Intensity of Intestinal Infection with Multiple Worm Species Is Related to Regulatory Cytokine Output and Immune Hyporesponsiveness

Abstract

Increasing immunological dysfunction (atopy and autoimmunity) in western society may be linked to changes in undetermined environmental agents. We hypothesize that increased exposure to multiple gut worm species promotes stronger immunological regulation. We report here that African children constitutively secrete more immunoregulatory cytokines (interleukin [IL]-10 and transforming growth factor [TGF]- β1) under conditions of hyperendemic exposure to the intestinal nematodes Ascaris lumbricoides and Trichuris trichiura compared with conditions of mesoendemic exposure. Under conditions of hyperendemic exposure, estimators of combined intestinal nematode infection level relate positively to combined constitutive IL-10 and TGF-β1 production and negatively to total immune reactivity (determined as IL-4, interferon-γ, and cellular proliferative responses to Ascaris or Trichuris helminth antigens, Streptococcus pneumoniae bacterial antigen, or the mitogen phytohemaglutinin). Total immune reactivity and anti-inflammatory cytokine production relate inversely. Our data suggest that gut nematodes are important mediators of immunoregulation

The hygiene hypothesis proposes a causal relationship between reduced exposure to microorganisms in the developed world and an increase in allergic disease [1, 2]. Tissue-dwelling helminth parasites (blood flukes and filarial nematodes) have been shown to suppress the symptoms of allergy [3], and experimental helminth infections can ameliorate disease in animal models of autoimmunity [4]. Perhaps more importantly in developing countries, tissue-dwelling helminth infections can modulate the type and magnitude of immune responses not only to parasite-specific antigens but also toward unrelated pathogen groups such as bacteria [5] and viruses [6], an immunological phenomenon that can be reversed after anthelmintic chemotherapy [7]. Mechanistically, there is direct evidence from animal models [8] and from naturally occurring infections [9, 10] that tissue-dwelling helminths initiate an expansion of T-regulatory (T_reg) cells that suppress both Th1 and Th2 effector responses. In the context of the hygiene hypothesis, however, it is debatable as to how much influence these spatially focal infections would have in driving a widespread anti-inflammatory phenotype in developing countries

In contrast, exposure to gastrointestinal worms is almost universal in the developing world, but rare in developed countries [11, 12]. Considering the phylogenetic relatedness of intestinal and tissue-dwelling worms [13] and the broad similarities in the immune responses that they induce [14], these organisms could be key players in the induction of immunoregulatory networks that inhibit the development of immunopathological disorders and modulate the inflammatory response to coinfection with nonhelminth pathogens. Indeed, recent data from an experimental model indicate that the gut worm Heligmosomoides polygyrus can suppress airway hypersensitivity and inflammation via an expansion of regulatory T cells [15]. Here, we ask whether gut worm infection level is directly linked to an elevated anti-inflammatory phenotype in naturally occurring infections. Furthermore, we assess whether immunological hyporesponsiveness is evident and associated with worm-driven immunoregulation

Materials and Methods

Study population and clinical evaluationField sites where gut worms were putatively hyperendemic and mesoendemic were selected on the basis of preexisting epidemiological data (fecal egg counts [FECs], in eggs per gram of feces) gathered by Centre Pasteur du Cameroun. School-aged villagers at Ayéné (3°23′ N, 11°40′ E) and Ngoya (3°57′ N, 11°27′ E) were surveyed for age and anthelmintic treatment history, and they supplied 3 stool samples on consecutive days for analysis of parasite egg output by use of the Kato-Katz thick-smear technique, as detailed elsewhere [16]. Sampling was school-based at both study sites, in each case from a single establishment (which, at Ayéné, represented the only school in a village of <500 inhabitants). Socioeconomic status was uniformly low at Ayéné and higher at Ngoya. No individual who participated in the study had received anthelmintic treatment for more than 1 year. Arithmetic mean FECs for Ascaris lumbricoides and Trichuris trichiura were calculated on the basis of 2–3 samples. To protect against erroneous counts caused by environmental contamination and/or experimental error, individuals with samples that yielded replicate counts that deviated by >30% were asked to provide an additional sample. If a discrepancy of >30% was still evident between all combinations of duplicates, the patient was excluded from the study

Participants supplied a finger prick sample of blood for thick-smear microscopic analysis of blood-dwelling filariae (Loa loa and Mansonella perstans) and were palpated for onchocercomas containing adult Onchocerca volvulus No individual in the study was infection positive by these measures. A survey of the wider Ayéné community found O. volvulus to be absent and found L. loa and M. perstans infections to occur at respective prevalences of 11.6% and 1.1% (n=440) across all age groups and at respective prevalences of 6.4% and 0.9% in 10–14-year-olds (n=109). Additionally, there was no evidence of hookworm infection at Ayéné or Ngoya, as determined by the Kato-Katz technique. All participants infected with gut worm were subsequently offered treatment with mebendazole. The study received clearance from the Cameroonian national ethics committee. Informed consent was obtained from the patients and their parents (or guardians), and the guidelines for human experimentation from the Ministry of Health, Cameroon, were followed in the conduct of clinical research

Parasite antigensParasites were obtained as described elsewhere [16, 17]. Trichuris muris was determined to be a suitable and more readily available antigenic substitute for T. trichiura for measurement of anti-Trichuris cellular and humoral immune responses in humans [18]. Ascaris lumbricoides antigen (AlAg) and T. muris antigen (TmAg) were prepared as described elsewhere [16, 17]

Cell cultureIndividuals who supplied stool samples were asked to donate between 5 and 10 mL of venous blood. Whole venous blood was diluted 1:4 in RPMI tissue culture medium containing 50 mmol/L L-glutamine and 80 μg/mL gentamicin. One-milliliter volumes were cultured at 37°C in 5% CO₂ in 48-well tissue culture plates (Corning) for 48 or 120 h in the absence of any stimuli or in the presence of AlAg, TmAg, or phytohemaglutinin (PHA; Sigma) at 5μg/mL and Streptococcus pneumoniae streptolysin-O (SLO; Difco) at 1:50. Eight hundred microliters of culture supernatant was removed and frozen in liquid nitrogen for subsequent cytokine analysis. For the measurement of lymphocyte proliferation, 200μL of diluted whole blood was plated in triplicate into 96-well tissue culture plates (Corning). Blood cells were cultured for 120 h in the absence or presence of antigens, as for the 1-mL cultures, before being pulsed with 1μCi tritiated thymidine (Amersham) for 6 h. Cells were harvested by use of a plate harvester (Tomtec) onto filter paper, and scintillation was measured with a β-scintillation counter, in accordance with the manufacturer’s instructions (Wallac). Mean counts per minute of beta radiation from stimulated cultures were divided by mean cell counts from unstimulated cultures to provide a proliferation index

ImmunoassaySamples of supernatant from the 48-h cultures were assayed for interleukin (IL)-4 and samples of supernatant from the 120-h cultures were assayed for IL-10 and interferon (IFN)-γ by ELISA, as described elsewhere [16, 17]. For the measurement of transforming growth factor (TGF)-β1, samples of supernatant from the 120-h unstimulated cultures were acidified for 10 min to activate latent TGF-β1 prior to ELISA, in accordance with the manufacturer’s instructions (Pharmingen; Becton Dickinson)

Statistical analysisThe strategy of data reduction and hypothesis testing is summarized in figure 1. Accumulations of anti-inflammatory cytokines in cultured blood from Ngoya and Ayéné schoolchildren were compared by use of nonparametric Mann-Whitney U tests. Variations in Ascaris and Trichuris FEC per individual were combined by calculating the mean of the 2 standardized FEC variables to give a measure of overall intestinal worminess. The standardization used was to zero the mean and represent variation in standard deviation units (i.e., [FEC observation − FEC sample mean]/FEC sample standard deviation). Likewise, anti-inflammatory cytokine accumulations in cultured blood were combined per person by averaging the standardized concentrations of IL-10 and TGF-β1 pg/mL to give a measure of combined anti-inflammatory cytokine accumulation. Standardization was carried out to limit bias in the combined variable resulting from greater absolute values or variability in 1 of the 2 constituent variables

Figure 1

Schematic plan of the statistical approach used to address the study hypotheses H1–H3. FEC, fecal egg count; IL, interleukin; GLM, general linear modeling; PCA, principal components analysis; TGF, transforming growth factor

Variation within the 11 responses to parasite antigens, bacterial antigen and mitogen was reduced by use of a standard data reduction technique, principal components analysis (PCA), as described elsewhere [18]. The first principal component (PC1) possessed moderate to high positive coefficients for all 11 immune response variables (see table 3 in the Results section), and thus reflected total immune reactivity (PC1 = total immune reactivity). A total of 3 hypotheses were tested (figure 1): H1, variation in anti-inflammatory cytokine accumulation is related to variation in intestinal worminess; H2, variation in immune reactivity is related to variation in intestinal worminess; H3, variation in immune reactivity is related to variation in anti-inflammatory cytokine accumulation. To determine the goodness-of-fit of response variables to a normal distribution pattern, raw or log₁₀ transformed variables were analyzed by use of a Kolmogorov-Smirnov test for normality. General linear modeling was used to test H1–H3. Combined anti-inflammatory cytokine accumulation (H1) or total immune reactivity (H2 and H3) were assigned as response variables. For all models, age (5 levels) and sex were added as fixed factors, and platelet count and white blood cell count were added as continuous covariates. Combined intestinal worminess was added as a continuous covariate in the model used to test H1, and graded intestinal worminess was added as a fixed factor in the model used to test H2. In the model used to test H3, combined anti-inflammatory cytokine accumulation was added as a continuous covariate. Only the significance of main effects was tested. To test H2, intestinal worminess was graded into levels of zero, mild, moderate, or severe by use of cutoffs (see figure 3A in the Results section). A 2-level “graded intestinal worminess” factor was generated that compared the zero and mild worminess categories with moderate and severe worminess

Table 3

Principal components analysis of immunological reactivity

Figure 3

A Classification of intestinal worm infection in Ayéné schoolchildren. Roundworm and whipworm infections were classified as zero (absence of parasite eggs in stool), light (fecal egg count [FEC] per gram of feces, 1–4999 for roundworm and 1–2999 for whipworm), moderate (FEC per gram of feces, 5000–19,999 for roundworm and 3000–4999 for whipworm) or heavy (FEC per gram of feces, >20,000 for roundworm and >5000 for whipworm). B Effect of level of intestinal worm infection on immune reactivity. Bars mean total immune reactivity (first principal component [PC1] arbitrary units [AU]) ± standard error of measurement. Schoolchildren who had no infection or were lightly infected with roundworm and/or whipworm (n=10) were compared with the remainder of the study group, who were moderately to heavily infected with either or both parasites (n=39). There was a significant reduction in total immune reactivity in the latter group, compared with the former group (*, variance ratio [F]_1,40 = 6.248; P=.017). The significance of this relationship was preserved when immune reactivity in individuals who were not infected and individuals with light infection with roundworm (n=22) were compared with moderately and heavily infected individuals (n=27) (*, F_1,40=4.643; P=.037) but not when immune reactivity was compared between whipworm grades. C Significant inverse relationship between anti-inflammatory cytokine accumulation and total immune reactivity (F_1,40=9.99; P=.003). Line linear regression best fit (added as a guide)

After hypothesis testing with combined variables and in cases where significant effects were identified, post-hoc testing was carried out using individual immune response and parasitological variables, under the same conditions. All analyses were carried out with SPSS for Windows (version 11; SPSS)

Results

Our field sites were in Centre Province, Cameroon. By cross-sectional sampling we identified 2 sites, the rural village of Ayéné 80 km southwest of the capital, Yaoundé, and the semiurban locale of Ngoya on the outskirts of Yaoundé. Human roundworm (A. lumbricoides) and whipworm (T. trichiura) were endemic in both villages, but the villages differed markedly in their infection levels (table 1). Using World Health Organization definitions of gut worm endemicity (>50,000 per gram of feces for A. lumbricoides and >10,000 per gram of feces for T. trichiura), Ayéné is a category I community for roundworm and whipworm (heavy infections affect >10% of the population), whilst Ngoya is a category II community (heavy infections affect <10% of the population) [19]

Table 1

Study group characteristics

In prior studies examining the cellular immune response to gut worms, we recorded notable levels of the anti-inflammatory cytokine IL-10, a key cytokine mediator of T_reg cells [20], accruing in unstimulated cultures of whole blood obtained from individuals infected with gut worms. We therefore compared 5-day blood culture accumulations of IL-10 and TGF-β1, another regulatory cytokine linked to T_reg activity [20], for children from Ayéné with the results for children from Ngoya. We examined an age range between 8 and 15 years because it is during this period of life that roundworm and whipworm attain maximal intensities of infection in communities where they are endemic [21]. All resting cultures produced measurable IL-10, but accumulations in cultures of samples from Ayéné schoolchildren were significantly higher than those in cultures of samples from Ngoya schoolchildren (P<.001) (figure 2A). TGF-β1 could be detected in a proportion of all resting cultures after 5 days (59% of cultures from Ayéné and 53% of cultures from Ngoya) (figure 3A). TGF-β1 accumulations were also more elevated in Ayéné, compared with Ngoya, although this difference was not statistically significant. Thus Cameroonian children with high-level exposure to intestinal worms clearly demonstrated enhanced counter-inflammatory cytokine expression, compared with more lightly exposed children of similar age and genetic background

Figure 2

(A) Accumulations of interleukin (IL)-10 and transforming growth factor (TGF)-β1 after 120 hour culture of peripheral blood derived from schoolchildren from Ngoya (mesoendemic for roundworm and whipworm) and Ayéné (hyperendemic for roundworm and whipworm). **, P<.001). (B) The significant positive relationship between variation in combined intestinal worminess and combined anti-inflammatory cytokine accumulation (F = 6.01, P=.019). Line is linear regression best fit added as a guide. (C) The positive but non-significant relationships between roundworm or whipworm FEC and IL-10 or TGF-β1 accumulations in peripheral blood. Lines are linear regression best fits added as a guide

We therefore tested the hypothesis that anti-inflammatory phenotype was directly proportional to increasing level of intestinal worminess in our sample from Ayéné. Using FEC as a measurement of worm burden, we examined the effect of total intestinal worminess (the average of a schoolchild’s roundworm and whipworm FEC) on grouped IL-10 and TGF-β1 accumulations. Because fluctuations in white blood cell counts could potentially influence net cytokine production in cultures of whole blood and because platelets store TGF-β1 [22] that could be released during culture, we included white blood cell counts and platelet counts as covariates in our statistical model along with terms for other potentially confounding variables (host age and sex) (table 2). After considering the effects of these parameters, there was a strong positive correlation between total intestinal worminess and constitutive anti-inflammatory cytokine secretions among Ayéné schoolchildren (P=.019) (figure 2B). If constitutive secretions in peripheral blood of either IL-10 or TGF-β1 were compared with individual worminess (roundworm or whipworm infection level), we observed that both intestinal worm species demonstrated positive trends in relation to IL-10 or TGF-β secretion (figure 2C). None of these relationships attained statistical significance, however, suggesting that overlapping effects of both intestinal parasites promote the IL-10 + TGF-β1 anti-inflammatory phenotype in a density-dependent manner

Table 2

General linear model of the relationship between combined intestinal worminess and combined anti-inflammatory cytokine accumulation in cultured blood

We then asked whether intestinal worminess and the associated anti-inflammatory phenotype could affect immunological responsiveness among Ayéné schoolchildren. We assayed recall responses in whole blood to roundworm antigen, whipworm antigen, and S. pneumoniae bacterial antigen, as well as responsiveness to mitogen (11 separate variables). Measurements comprised IFN-γ responses representative of type 1 T helper lymphocytes (Th1), IL-4 responses representative of Th2 lymphocytes, and lymphocyte proliferation. We used a standard data reduction technique, PCA, to condense this comprehensive immunological survey into groupings of responses that had a high degree of intercorrelation. From this analysis we identified a predominant trend covering 35.4% of total variation within the 11 response measurements, with positive contributions from IFN-γ, IL-4, and proliferative responses (i.e., total immune reactivity) (table 3). Thus the PCA indicated that immune responses that belonged to the Th1 or Th2 phenotype, invoked against worm and bacterial antigens, as well as nonspecific responses to mitogen, were mixed rather than polarized in most schoolchildren in our cohort and shared a similar pattern (increased Th1 = increased Th2-like response)

We initially compared total immune reactivity in schoolchildren who had light intestinal worminess (i.e., zero or light roundworm infection and light whipworm infection) with schoolchildren who had moderate to heavy intestinal worminess (i.e., moderate or heavy roundworm infection and/or moderate or heavy whipworm infection) (figure 3A). After accounting for potential confounders, total immune reactivity was significantly reduced in the group that had moderate to heavy intestinal worminess (P=.017) (figure 3B). We then considered the effects of species-specific worminess on total immune reactivity (figure 3B). Both moderate to heavy roundworm and whipworm infections were associated with reduced total immune reactivity. However, this association was only significant for roundworm infection (P=.037), indicating that high-level roundworm infection was relatively more important than high-level whipworm infection in the association between intestinal worminess and reduced immune reactivity in Ayéné schoolchildren. When these analyses were repeated with continuous rather than categorized parasite variables, there was still a significant (negative) association between total worminess and total immune reactivity (P=.045). FEC variables for both species were nonsignificantly negatively associated with total immune reactivity, although the relationship for Ascaris FEC (P=.108) approached the significance cutoff (P=.05)

In addition, we considered whether there was evidence linking the intestinal worminess–driven anti-inflammatory phenotype and reduced immunological reactivity toward helminth-specific, bacteria-specific, and nonspecific stimuli. There was a highly significant inverse relationship between total IL-10 and TGF-β1 accumulation in 5-day cultures and total immune reactivity (P=.003) (figure 3C). When this was dissected into relationships between individual immune reactivity variables for worm-specific, bacteria-specific, or nonspecific stimuli and either constitutive IL-10 or TGF-β1 accumulations, an overarching inverse relationship was apparent between accumulation levels of TGF-β1 and recall responsiveness, regardless of whether it was a Th1-type, Th2-type, or proliferative response and whether the antigen stimulus was worm-specific, bacteria-specific, or even nonspecific (figure 4). Accumulations of IL-10, on the other hand, demonstrated a more directed negative relationship with Th2-type responses, most notably to worm antigen stimulation (figure 4). These associations link intestinal worminess–related IL-10 and TGF-β1 with reduced immune reactivity, both to worm antigens and to unrelated stimuli

Figure 4

Relationship between individual measurements of immune reactivity toward parasite antigens, bacterial antigen and mitogen and accumulations of interleukin (IL)-10 or transforming growth factor (TGF)-β1 in cultured blood. F values are with 1 and 40 degrees of freedom. Lines are linear regression best fit added as a guide

Discussion

We have demonstrated that constitutive levels of the regulatory cytokines IL-10 and TGF-β1 are enhanced in direct relation to the intestinal worm burden and provide evidence that this elevation in anti-inflammatory cytokine secretion in peripheral blood induces immunological hyporesponsiveness. This observation is striking when considering the plethora of other microorganisms encountered and the many other environmental, dietary, and lifestyle factors that are typically varied in the setting of a natural human infection, all of which could potentially induce or suppress IL-10 and/or TGF-β1 output. We infer from our data, therefore, that chronic gut-worm infection plays an important role in driving human immunoregulatory networks. Thus, our findings provide a mechanism for the observed lower incidence of autoimmune disease, allergies, and asthma in communities where gut worms are endemic [23, 24]; the way in which gut worm infections negatively influence the expression of allergy [25] or the effectiveness of vaccination against bacterial infectious agents [26, 27]; and the association between successful deworming and an increase in Th2-mediated allergic reactions [28] and restoration of vaccine efficacy [26, 27]

Correlational field studies are potentially subject to confounding from significant unmeasured variables. Most important among these is likely to be socioeconomic status, which might correlate both with exposure to worms, due to poor sanitation and hygiene, and with variation in nutrition and its possible effects on immune responsiveness [29]. Comparison between the Ngoya and Ayéné sites, which had, respectively, higher and lower socioeconomic status, could have been partly confounded by this difference. However, socioeconomic variation within the Ayéné sample, upon which most of our analysis rests, would have been far more limited because status at this site was uniformly low

Our experimental design allowed us to consider the effects of polyparasitism of the gut by multiple worm species on regulatory cytokine output. The results of this analysis indicate that helminth polyparasitism and cosusceptibility to multiple, heavy worm infections are relatively more important than the effects of individual species per se in driving the immunoregulatory phenotype. This is important when considering that multiple gut-worm infections are the norm in communities where these parasites are endemic and cosusceptibility is a feature of gut-worm infection [11, 30, 31]. Because there is also evidence that cosusceptibility occurs for gut-dwelling and tissue-dwelling helminths [30, 31], an additional layer of immunoregulatory potential may be exerted in susceptible individuals in areas where such infections are coendemic

It has been hypothesized that worm-driven immunoregulatory networks, exemplified by the induction of IL-10 and or TGF-β secreting T_reg cells, represent a parasite survival strategy that enables suppression of an effective immune response [14]. Gut helminths are generally considered to benefit from the modulation of Th2-like responses because data generated in model systems clearly demonstrate that Th2 cytokines drive effector mechanisms at the site of infection that lead to parasite expulsion [32, 33]. Furthermore, we and others have observed that Th2 cytokines and antibody responses linked to Th2 activity are inversely associated with human intestinal helminth infection [16, 17] and with reinfection following chemotherapeutic intervention [34, 35]. Our data are consistent with this hypothesis; accumulations of secreted IL-10 and TGF-β1 are inversely associated with Th2 (IL-4) recall responses to parasite antigens. However, in the case of TGF-β1, we have concluded that this suppressive cytokine is also associated with diminished Th1 responsiveness to bacterial antigen and a nonspecific stimulus. This may be a consequence of gut worm–driven regulatory activity spilling over onto unrelated adaptive cellular responses in heavily polyparasitized individuals. Alternatively, given that TGF-β1 expression increases during wound healing and has thus been implicated in the regulation of such responses [36], the suppression of Th1 responses by TGF-β may be a consequence of tissue repair responses to damage of the gut mucosa, liver, or lungs by heavy worm infection and continuous larval reinvasion. A further possibility is that enteric bacterial infections triggered by worm-mediated disruption of the gut mucosa might also up-regulate counterinflammatory cytokine expression

Dissection of the mechanism by which gut worms promote secretion of IL-10 and TGF-β1 is valuable in understanding why allergies and autoimmunity are increasing in the developed world and may present new avenues for treating or preventing such disorders. In fact, gut worms have already been demonstrated to have therapeutic value in the treatment of inflammatory bowel disease [37]. Furthermore, considering the frequency with which intestinal worm infections occur in humans (approximately 1 billion people are infected with at least 1 species) [11, 12] and their overlapping geographical distribution with devastating diseases such as HIV infection, malaria, and tuberculosis [38], our findings have important implications for the rationale of programs for the control of gut worms, the evaluation of trial vaccine efficacy, and community-directed vaccination strategies, especially those that target school-age children

Acknowledgments

We wish to thank the people of Ayéné and Ngoya for participating in the study. We acknowledge the support and assistance of the technical staff at Centre Pasteur du Cameroun, Dr. Leo Basco, Dr. Marie-Claire Rowlinson, and Dr. Matthew Taylor. We are grateful for the additional laboratory facilities supplied to us by the Organisation de Coordination pour la lutte contre les Endémies en Afrique Centrale (OCEAC), Yaounde, Cameroon

Presented in part: 11th International Congress for Parasitology, Glasgow, United Kingdom, August, 2006

Potential conflicts of interest: none reported

Financial support: Wellcome Trust (grant 055939/Z/99/Z)

^a Present affiliation: Department of Biology, University of York, Heslington, York, United Kingdom (J.D.T.); Institut de Recherche pour le Développement, UR24, 4 Montpellier, France (M.B.)