Immunity, immunoregulation and the ecology of trichuriasis and ascariasis

Authors

  • J. E. Bradley,

    Corresponding author
    1. School of Biology, Nottingham University, Nottingham NG7 2RD, UK
      J. E. Bradley, School of Biology, Nottingham University, Nottingham NG7 2RD, UK (e-mail: Jan.Bradley@Nottingham.ac.uk).
    Search for more papers by this author
  • J. A. Jackson

    1. School of Biology, Nottingham University, Nottingham NG7 2RD, UK
    Search for more papers by this author

J. E. Bradley, School of Biology, Nottingham University, Nottingham NG7 2RD, UK (e-mail: Jan.Bradley@Nottingham.ac.uk).

SUMMARY

Immune responses to human roundworm (Ascaris lumbricoides) and whipworm (Trichuris trichiura) and their role in controlling worm populations are reviewed. Recent immunoepidemiological data implicate Th2-mediated responses in limiting A. lumbricoides and T. trichiura populations. Reinfection studies further suggest that IL-5 cytokine responses are negatively associated with adult recruitment in T. trichiura but not A. lumbricoides and may therefore be involved in negative intraspecific and interspecific interactions mediated through the host immune system. The importance of inducible immunoregulatory networks in the ecology of the host–parasite relationship is considered, with particular regard to possible manipulative strategies by the parasites. This aspect of the worms’ interaction with the host immune system is both poorly known and potentially central to an understanding of parasite population dynamics and the evolutionary pressures that have shaped present-day host–parasite associations. Some possible implications of worm-mediated immunomodulation for the occurrence of bystander infectious diseases in human populations and the management of de-worming programmes are also discussed.

GENERAL INTRODUCTION

Ascaris lumbricoides (the roundworm) and Trichuris trichiura (the whipworm) are two of the most widespread metazoan parasites of human populations. Respectively, they have been estimated to infect 1472 and 1049 million individuals (1) primarily in the tropics and subtropics. Whilst both species can be associated with significant pathology and morbidity, the symptoms attributable to direct pathogenic effects are generally mild or undetected in most hosts. However, the vast numbers of humans affected and the growing perception that worms have strong anti-inflammatory influences on the host immune system (2,3), makes understanding the immunological effects of these infections a potentially vital public health issue. In particular, A. lumbricoides and T. trichiura co-occur geographically with devastating microbial infectious agents such as HIV, Mycobacterium tuberculosis and malaria. This suggests that worm-mediated changes in immune responsiveness (or the lack of such influences when worms are removed by chemotherapy) could have an important impact on infection rates and disease outcomes in these pathogens.

The following review will consider immunological aspects of whipworm and roundworm infections, including the possible role of immune responses in limiting worm populations. It will also discuss, from an evolutionary perspective, the way in which worms (and de-worming) could influence the control of immune responses and the potential effect of worm-mediated immunomodulation on bystander infectious diseases.

BACKGROUND: LIFE-CYCLES AND PATHOLOGY

Both A. lumbricoides and T. trichiura have direct life-cycles, inhabiting the intestinal tract as adults and producing eggs that are voided to the external environment with the faeces. Unlike the hookworms (considered elsewhere in this volume), whose third-stage larvae actively penetrate skin, both A. lumbricoides and T. trichiura are transmitted passively within the egg, being ingested by the host as a result of faecal contamination. Despite the superficial ecological similarities between A. lumbricoides and T. trichiura, the two species are phylogenetically relatively distant, respectively, descending from the two main secernentean and adenophorean lines of parasitic nematodes occurring in vertebrates. Their life histories in the final hosts also differ, with A. lumbricoides larvae undergoing an extensive migration through a series of host tissues and organs, whilst T. trichiura larvae develop entirely in the gut area. Given their phylogenetic divergence and different use of host tissues, the two species might be expected to be physiologically and antigenically relatively distinct. They might also therefore be expected to differ in the immune responses that they elicit and are vulnerable to.

In T. trichiura, larvae mature to the infective L2 stage within the egg. On emergence they penetrate and develop in the mucosal epithelium, with adults only usually establishing in the caecum or colon (whether as the result of an active behaviour or differential mortality is unclear) (4). The anterior of the adult remains within an epithelial tunnel, whilst the expanded posterior section protrudes into the intestinal lumen. Prepatent period is probably 2–3 months and maximum lifespan up to 4 years (4,5), although detailed information is not available and most worms may be lost before 1–2 years.

A more complex series of larval migrations is found in A. lumbricoides (4). Larvae undergo one or possibly two moults in the egg, penetrating the intestinal wall and migrating to the hepatic portal circulation on emergence. Studies of the closely related Ascaris suum in pigs (4), which has a comparable life-history, suggest worms invade the liver 1–2 days later and the lungs by 5–6 days post-infection. From here, L3 larvae migrate up the trachea and are swallowed, reaching the intestinal lumen after 8–9 days where they moult twice and mature at 50–55 days post-infection. Adult A. suum live for up to 55 weeks in pigs but most are lost by 23 weeks.

The occurrence of both A. lumbricoides and T. trichiura usually peaks in childhood or early adolescence (5,6) and can be associated with significant direct pathogenic effects (5,7). Heavy T. trichiura infections induce Trichuris dysentery syndrome (TDS) whose symptoms include chronic dysentery, rectal prolapse, anaemia, inhibited growth, and clubbing of the fingers. Ascaris lumbricoides is also known to be associated with nutritional impairment, intestinal pathology, and reduced growth. In acute ascariasis, blockage of the intestine and bile ducts may lead to life-threatening complications. Infections with both species may, furthermore, produce some degree of impaired cognitive function in children (8). Both worms are also likely to have heavy economic impacts: Chan (9) estimated that in 1990, disability-adjusted life years (DALYs), a measure of years of healthy life lost, numbered 10·5 million for A. lumbricoides and 6·4 million for T. trichiura.

IMMUNOLOGY AND HOST RESISTANCE

Studies on A. lumbricoides and T. trichiura suggest that the recolonization of human populations following chemotherapy can be extensive. Elkins et al. (6) found that 11 months after treatment, A. lumbricoides infection intensity in a southern Indian population had recovered to 55% of pre-treatment levels. Bundy et al. (10) reported a 44% recovery in T. trichiura infection intensity 17 months after treatment in a Caribbean community. Such data indicate that acquired resistance generated by past infections is at best partial in many individuals. Age–prevalence and age–intensity relationships for T. trichiura and A. lumbricoides are typically convex, peaking in children or adolescents (5,6). Whilst this is consistent with a slow increase in acquired resistance with age (11), the same patterns could also be explained by age-specific variation in exposure (12). However, in relevant mammalian model systems such as Trichuris suis and A. suum in pigs (13,14) and Trichuris muris in mice (15), strong partial acquired resistance can be generated under some transmission regimens in some genetic types.

Work on murine model systems has emphasized the importance of genetically determined variability in resistance to gastrointestinal (GI) nematodes (16). Studies in congenic mouse strains suggest that both MHC and non-MHC genes may influence susceptibility to T. muris (17) and Heligmosomoides polygyrus (18), whilst Behnke et al. (19) mapped several quantitative trait loci (QTL) associated with resistance to H. polygyrus in mice. Some candidate genes and genetic regions have also been linked with susceptibility to A. lumbricoides and T. trichiura in humans (reviewed by Quinnell (20)). Underlying genetically determined variation in resistance–susceptibility might partly explain the predisposition to infection often observed in field studies of treatment and re-exposure (10). This predisposition (a tendency for pre- and post-treatment infection levels within individuals to be correlated) may also be partially due to variable environmental factors and exposure (12). However, recent quantitative genetic studies (21,22) of pedigrees from Nepalese and Chinese communities have found a substantial genetic component determining variation in susceptibility to T. trichiura (28%) and A. lumbricoides (30–50%).

GI nematode infections induce vigorous immune responses, although the mechanisms involved in putative functional resistance are not fully resolved. Because direct experimental studies are not usually possible in humans, investigators are only able to approach this problem by comparative studies on murine model systems or observational field studies that correlate levels of potentially relevant immune signalling or effector molecules with infection state variables. Both types of study will be considered in the present review. However, it is noted that no direct immunological model has been developed for A. lumbricoides and discussion will concentrate on the most relevant to T. trichiura: T. muris in the mouse. Some supporting data are given for other models when considering phenomena potentially common to GI nematodes in general.

Early field studies on human populations focused on the humoral immune response, examining the relationship between circulating specific immunoglobulins and infection. Whilst it should be noted that the immunoglobulins in peripheral blood may not be representative of those at active infection sites (23), specific responses involving all serotypes (IgM, IgG1-4, IgA, IgE) have been recorded in A. lumbricoides or T. trichiura infections (11,24–26). A strong tendency exists for specific immunoglobulin levels to track worm burden (24,26–28). This could simply be due to the degree of antigenic exposure and involve immunoglobulin molecules that have no role in protection. However, some negative associations have been reported for serotypes of potential functional significance in the expulsion of GI nematodes. Studies on Caribbean populations suggest a negative association between specific salivary IgA (in the dimeric secretory form) and T. trichiura infection intensity (29). This may relate to the activity of IgA in intestinal mucosal secretions. Some studies have also found negative associations between specific IgE and susceptibility to A. lumbricoides in children. Hagel et al. (30) reported a negative association of reinfection with specific IgE in an urban Venezuelan population. McSharry et al. (31) found an inverse association between IgE specific to the A. lumbricoides recombinant allergen ABA-1 and persistent susceptibility in Nigerian individuals producing IgG responses against the same antigen. A negative relationship has also been recorded between A. lumbricoides infection and ABA-1-specific IgE in older members (> 12 years) of a Cameroonian population (Dr J.D. Turner & JEB, manuscript in preparation). Furthermore, T. trichiura faecal egg output in children from this population was negatively related to IgE specific to somatic T. trichiura antigen (32). The importance of antibody responses in immunity to A. lumbricoides is supported by a quantitative genetic analysis of genetic marker data from Nepalese pedigrees (33). This study strongly implicated the TNFSF13B gene on chromosome 13 in resistance/susceptibility. TNFSF13B codes for a member of the TNF superfamily involved in the regulation of B-cell activation and immunoglobulin secretion. The in vivo activity of this cytokine includes promoting the survival of immunoglobulin-secreting cells (34), and might therefore be expected to enhance antibody responses against intestinal nematodes.

Direct experimental studies in murine model systems have demonstrated that B-cells and immunoglobulins can indeed contribute to protection against GI nematodes. However, the time-lag involved in de novo antibody responses, and the ability of model hosts to expel primary infections in the absence of antibody, suggest immunoglobulins are probably most effective in secondary infections. In the T. muris–mouse model, susceptible SCID mice (B- and T-cell deficient) can be shifted to a resistant phenotype solely by transfers of CD4+ T-cells from resistant BALB/c mice (indicating the B-cell compartment is not essential) (35). Blackwell & Else (36), on the other hand, have demonstrated that susceptible B-cell-deficient µMT mice were restored to a resistant phenotype by reconstitution with naïve wild-type (C57BL/6) splenic B-cells. The effect of B-cells in this system could be related to their antigen-presenting function and signalling activities during T-cell responses, although treatment with specific IgG1 from resistant (NIH) mice was also protective. In Trichinella spiralis, studies of IgE-deficient mice suggest higher levels of this serotype enhance worm clearance (37). In the H. polygyrus–mouse system, negative associations of susceptibility with the production of specific IgG1 and IgA have been found between and within (38) different mouse strains. Passive transfer of purified IgG1 from immune mice also enhances protection in this model (39). In Ascaris and Trichuris spp. infections therefore the MHC restriction of antibody responses against parasite antigens (demonstrated in mice (40,41)) could perhaps contribute to genetic determinism in susceptibility.

The T-cell compartment is known to be of critical importance in resistance to GI nematodes. On the basis of murine models, it has become widely accepted that protection is associated with Th2-induced responses and susceptibility is associated with a Th1 phenotype (see reviews by Finkelman et al. (42) and Gause et al. (43)). In the T. muris–mouse model, for example, resistant inbred strains generate Th2-polarized responses, whilst chronically susceptible strains generate Th1-polarized responses (44). Th2 cytokines associated with functional resistance in different murine models include the interleukins (ILs) IL-4, IL-5, IL-9 and IL-13 (42,44–46), although the involvement of individual cytokines may vary between models (43). In the T. muris model, the pro-inflammatory cytokine tumour necrosis factor alpha (TNF-α) is required for resistance (47). However, the role of TNF-α in resistance–susceptibility is context-dependent, with high levels also occurring during Th1 responses in susceptible hosts (48). IL-10 is also important for resistance to T. muris, due to its regulatory effect on Th1 responses (48). Finkelman et al. (42) have suggested that different GI nematode species may stimulate the same, essentially stereotypical, Th2 cascade and that functional resistance against particular species depends on different subsets of mediators and effectors within this cascade.

In humans, Cooper et al. (26) showed that PBMCs stimulated with A. lumbricoides antigens in ex vivo assays showed stronger Th2 responses in infected individuals from rural Ecuadorian communities than in uninfected subjects from an adjacent urban environment. Greater frequencies of IL-4 and IL-5-secreting cells relative to IFN-γ secreting cells, and a greater absolute production of IL-5, occurred in the infection-exposed subjects. A protective role for Th2 cytokine responses in resistance to GI nematodes in humans was corroborated within a Cameroonian population hyperendemic for A. lumbricoides (49). Here, an age-stratified analysis indicated that a negative association occurred between ex vivo whole blood IL-9 and IL-13 responses to parasite antigen and A. lumbricoides faecal egg counts (FECs) in the older (> 12 years) members of the study group. Age-specific negative associations between A. lumbricoides and T. trichiura infection and ex vivo whole blood IL-13 and IL-5 responses to parasite antigen have also now been reported in a different Cameroonian population (50,51). Furthermore, Cooper et al. (52) found that atopic responses (usually inversely associated with helminth infection) in Ecuadorian schoolchildren were significantly positively associated with the frequency of PBMCs secreting IL-4 and IL-5 when stimulated by Ascaris antigens in ex vivo assays. Although there was no direct association between IL-4 or IL-5 secretion and A. lumbricoides infection status, the highest responses occurred in the group of subjects who were both atopic and uninfected.

The functional effectors at the base of the Th2 cascade are likely to include eosinophilia, IgA, IgE, mastocytosis and increased mucus secretion. There are also physiological changes in the intestine, including increased mucosal permeability and smooth muscle contractivity, driven by IL-4 and IL-13. Together these may produce an environment in which worms tend to become trapped in mucus and be swept from the gut (46). As noted above, different subsets of effectors may operate against different nematode species (42). Mast cells, for example, may not be important in the expulsion of Trichuris spp. but appear to be a central element of protective responses against some other GI nematodes (53). In human populations, IL-5 (an important mediator of eosinophil responses) has been found to have a strong negative association with reinfection in some GI nematodes (50,51). This is consistent with the effectiveness of IL-5-driven Th2 cellular responses being greatest against larvae, rather than established adult infections (54). Jackson et al. (51) also found differential effects of IL-5 on the recruitment of T. trichiura and A. lumbricoides following treatment, with only T. trichiura showing a significant negative association between reinfection and ex vivo whole blood IL-5 responses to parasite antigen. This may be explained by the different life-histories of the two species. Whilst invading T. trichiura larvae remain localized in the gut area throughout their development, A. lumbricoides preadults undergo an extensive migration through the circulatory system, liver and lungs, during which the developing stages would be sequestered from localized Th2 responses in the gut. The induction of IL-5 responses could potentially therefore be important both in intraspecific competition in T. trichiura and in interspecific interactions between A. lumbricoides and T. trichiura mediated through the host's immune system.

WORMS AND IMMUNOREGULATION

Th1 and Th2 responses represent the two main arms of the adaptive immune system driving effector mechanisms. Apart from T-helper cells that promote effector activity, a subset with regulatory properties (Treg) also exists that can dampen immune responses in both effector arms. These typically signal other cells through the immunosuppressive cytokines IL-10 and TGF-β. Different functional types amongst Treg cells have not been fully characterized, but a distinction can be made between natural and adaptive cells (55). The former are specific to self-antigens and influence T-cells and APCs by direct cell–cell contact, independent of cytokines. Adaptive Treg cells are induced in the periphery, are specific to tissue-specific and foreign antigens and influence other cells by direct contact and cytokine signals. Natural Treg cells and at least some adaptive Treg subsets are characterized by persistent CD25 expression. Whilst worms might be expected to benefit directly if they could induce adaptive Treg cells (through a reduction in specific effector activity), it is also possible that effects on natural Treg cells might be evolutionarily adaptive (perhaps in terms of maintaining the host niche by limiting anti-self immune responses during infection).

Although the variety of effector mechanisms marshalled by signals from T-helper cells potentially allow flexible and appropriate responses against a range of different pathogens, the control systems involved in T-cell function may also be vulnerable to manipulation by parasites. It has long been known that some GI nematode infections can suppress host immune responses (56). Studies in human populations (primarily on tissue-dwelling helminths such as schistosomes and filarial nematodes) suggest that whilst worms generally shift the host phenotype towards Th2, chronic infections are also often associated with immune hyporesponsiveness and T-cell anergy (3). This is assumed to be through induction of the Treg T-helper cell phenotype. In some cases hyporesponsiveness is antigen-specific (57), but it can also extend to polyclonal responses (58), perhaps dependent on the intensity of infection (59). Worm-mediated hyporesponsiveness cannot be explained by clonal deletion of antigen-specific cells, as specific responses can be restored after chemotherapy (60).

Apart from the suppression of immune responses, worms may also subvert host immunity through the induction of inappropriate effectors. Thus, A. lumbricoides is known to induce vigorous total and specific IgE responses (11). It has been suggested that the induction of high levels of polyclonal IgE may modulate immediate hypersensitivity reactions by competitive binding of FcɛR1 receptors (30). Competitive blocking of epitopes by IgG4 may also be important in compromising specific IgE responses (11,61). This is supported by a recent study (Dr J.D. Turner & JEB, manuscript in preparation) in a Cameroonian population exposed to A. lumbricoides, that found a positive association between infection intensity and IgG4:IgE ratio for immunoglobulin specific to the parasite-derived allergen ABA-1.

It is well established that various excretory–secretory (ES) products from parasitic nematodes can modulate mammalian immune responses. These may act by directly interfering with immune effectors, or by influences on upstream mediators involved in induction and control. Most data are available for filarial nematodes, whilst GI nematodes are still relatively poorly studied. Many substances with immunomodulatory effects have been described from filarial species, including serine and cysteine protease inhibitors (62), cytokine-like proteins (63) and phospholipid components bound to proteins (64). Amongst GI nematodes, Nippostrongylus brasiliensis in mice secretes a cysteine protease inhibitor, nippocystatin (65), that interferes with antigen processing and specific responses. Unknown proteins in adult N. brasiliensis ES products have also been found to influence host responses towards Th2 (66), whilst L3 ES material inhibits the bronchoalveolar recruitment of neutrophils in LPS-induced inflammation (67). Based on an expressed sequence tag (EST) study Harcus et al. (68) found that secreted signalling proteins in this species may be undergoing relatively accelerated diversification, possibly attributable to strong selective pressure from the host immune system. Amongst studies particularly relevant to the worms considered in this review, Paterson et al. (69) found that inoculation of BALB/c mice with A. suum body fluid during the induction phase of immune responses to heterologous antigens suppressed delayed-type hypersensitivity reactions and promoted a Th2-biased response (with increased total serum IgE and reduced IFN-γ production by lymphocytes). There was also a tendency for lymphocytes from treated animals to produce increased regulatory (IL-10) cytokine responses. Based on cells from BALB/c mice, Deehan et al. (70) found that glycosphingolipids from A. suum suppressed T-cell proliferation induced by F(ab′)2 fragments of anti-murine Ig and by LPS. These products also suppressed IL-12p40 production by LPS/IFN-γ-stimulated peritoneal macrophages.

However, in human populations, evidence for an association between GI nematodes and immunosuppressive T-cell subsets or cytokines is currently limited. In a small field study, Geiger et al. (71) found that ex vivo PBMC IL-10 responses to A. lumbricoides-derived antigen were unrelated to individual infection status with T. trichiura and A. lumbricoides. However, Turner et al. (49) reported a negative relationship between ex vivo whole blood IL-10 responsiveness to parasite-derived antigen and A. lumbricoides FEC in an older (12–36 years) host age class at one locality in Cameroon. In another Cameroonian population, Jackson et al. (51) found a significant association between a measure of differential susceptibility to A. lumbricoides and T. trichiura and ex vivo whole-blood IL-10 responsiveness to parasite antigen. Here, IL-10 secretion in older individuals (14–57 years age class) tended to increase in hosts with more fecund A. lumbricoides infections and less fecund T. trichiura infections. This indicates IL-10 could be important in chronic A. lumbricoides infections. A further study of ex vivo whole-blood TGF-β responses to parasite antigen in a Cameroonian community suggests these may be positively related to the presence of A. lumbricoides infection, but only in very heavy infections (Dr J.D. Turner & JEB, unpublished data). Taken together, the existing field data indicate there may be no simple, all-encompassing relationship between GI nematode faecal egg counts and regulatory cytokines. Levels of these mediators could potentially be elevated both in Th2-biased resistant phenotypes (where successful anti-worm responses are accompanied by a significant regulatory component, as in some resistant murine models discussed above) and in chronically susceptible phenotypes (where worms persist through immunodulatory activities). Further field studies, especially comparative analyses of age and infection-delineated strata within populations, may give a greater insight into the role of Treg subsets and anti-inflammatory cytokines in human GI nematode infection.

IMMUNOLOGICAL EFFECTS ON BYSTANDER INFECTIONS

It has been demonstrated in murine models that some GI nematodes can modulate immune responses to heterologous antigens (72) and affect heterologous resistance to bystander infectious agents (73). As discussed above, worms in humans may often tip the Th1/Th2 balance towards the Th2 phenotype. They also putatively induce the development of regulatory T-cell subsets (associated with IL-10 and TGF-β production) that effect a generalized reduction in responsiveness of both the Th2 and Th1 arms. Overall, therefore, helminthiases would be expected to down-regulate the Th1 responses that are critical to the immune control of many viral, bacterial and protozoan infections. This is consistent with field studies in human populations showing that worm infections may inhibit immune responsiveness to (58), and reduce the effectiveness of vaccination against some microbial pathogens (74). In the case of filarial nematode (Onchocerca volvulus) infections in a Cameroonian population, Stewart et al. (58) found that the ex vivo proliferation and IL-4 responsiveness of PBMCs in response to adult worm and mycobacterial antigen was negatively associated with microfilaraemia. However, there was also an age-dependent (and perhaps exposure-related) shift towards Th2-polarized antigen-specific responses.

Worms could potentially therefore have significant effects on some microbial infections, particularly in the case of pathogens like HIV or Mycobacterium spp., where the Th1/Th2 balance influences infection outcome (2). GI nematode infection has been positively associated with pulmonary tuberculosis in a Brazilian population (75). Intestinal helminthiasis has also been linked to accelerated progession of HIV (2) because of the chronic immune activation and increased expression of HIV co-receptors observed in worm-infected Ethiopian migrants (76–78). This is supported by studies indicating that PBMCs from worm-infected individuals are more susceptible to HIV-1 in vitro (79) and that positive association between worms and HIV-1 viral load occurs in asymptomatic HIV-positive Ethiopians (80). Furthermore, several studies have proposed epidemiological links between A. lumbricoides infection in humans and the occurrence of malaria. Surveys in Thailand suggest a negative relationship between cerebral malaria and A. lumbricoides infection (81) and also between cerebral malaria and malaria-associated renal failure and jaundice and intestinal helminth infection in general (82). A hypothesis put forward to explain this is that chronic activation of the CD23/NO pathway by helminths (82) might reduce the expression of adherence receptors and reduce severe malaria symptoms due to cell adherence phenomena. Prevalence of P. falciparum is, on the other hand, associated positively with intestinal helminth infection in Thailand (83) (perhaps a consequence of the general immunosuppressive activities of worms?). Helminth infections may also affect the population dynamics of different malarial species and genotypes in this region. Nacher et al. (84) found the frequency of mixed P. falciparum and P. vivax infections was positively associated with A. lumbricoides infection and Chaorattanakawee et al. (85) found that T. trichiura infection was associated with increased numbers of P. falciparum genotypes within infections. In contrast to the Thai studies, Le Hesran et al. (86) found that severe malaria was positively associated with the occurrence of A. lumbricoides in Senegalese populations. This variability of effect suggests that the causal relationships underlying associations between helminths and bystander pathogens in field situations may be complex and context-dependent.

WORMS AS MODULATORS OF INNATE RESPONSES

The development of adaptive anti-pathogen responses is critically influenced by early signalling from antigen-presenting cells (APCs). Given the position of APCs at the apex of the control sequences leading to anti-parasite effector responses, growing research interest is focusing on the possibility that worm infections could modulate downstream host immune responses through influences on these cells (3). Innate responses in APCs are triggered when pathogen-associated molecular patterns (PAMPs) bind to germ-line-encoded pattern recognition receptor molecules, such as the Toll-like receptors (TLRs). The responses activated by a PAMP tend to be stereotypical for a generic class of pathogen with which that PAMP is associated. They include signals that induce T-helper cells to develop towards a Th1, Th2 or Treg phenotype (Figure 1). T-helper cell polarization occurs primarily through the activities of dendritic cells (DCs), although other APCs such as macrophages and monocytes (which include DC progenitors) and B-cells may also be important. The signalling pathways between APCs and T-cells that lead to Th1, Th2 or Treg polarization have not yet been fully resolved (87–89). However, it is well known that ligation of TLRs by bacterial PAMPs is associated with the secretion of cytokines of the IL-12 family and Th1-polarizing activity by DCs (87). Apart from the identity of a PAMP, functional dendritic cell responses may also be influenced by tissue factors and the general cytokine environment (87). Thus, different tissues may produce signals favouring a T-cell phenotype suitable for the local environment. Also, feedback of cytokine signals from polarized T-cells may influence DC maturation and reinforce their control activities towards a particular phenotype (90). For example, DCs maturing in the presence of IL-10 or TGF-β induce regulatory T-cell phenotypes in vitro (91,92).

Figure 1.

Schematic representation of phenotype induction in T-helper cells and the consequences for immune responses against gastrointestinal (GI) nematodes. Pathogen-associated molecular patterns (PAMPs) from worms trigger pattern-response receptors (PRRs) on antigen-presenting cells of the innate immune system, including dendritic cells (DCs). Activated DCs then polarize T-helper cell phenotype. Different PAMPs induce individual DCs to influence T-helper cells towards a Th1, Th2 or Treg phenotype. Polarized T-helper cells then differentially recruit or suppress immune effectors with distinct consequences for anti-worm responses. Bacterial PAMPs are known to trigger Th1 responses (87), and GI nematode PAMPs are known to trigger Th2 responses (99). Other actions shown for GI nematode PAMPs are hypothetical. Blue panel: innate immune compartment; yellow panel: adaptive immune compartment. Cytokine signals associated with innate and adaptive phases of the immune response are shown in italics.

Studies in schistosomes have shown that Schistosoma mansoni soluble egg antigen inhibits IL-12 production by murine DCs and induces Th2-polarized adaptive responses (93). Furthermore, evidence is accumulating that schistosome-derived molecules influence signalling from human innate immune cells. Human monocytes are activated to produce anti-inflammatory (IL-10) and pro-inflammatory (IL-6, TNF-α) mediators by schistosome glycolipids (94). Van der Kleij et al. (95) identified a phosphatidylserine fraction from schistosome eggs that activated TLR2 and induced IL-10 expression. Furthermore, in field studies on schistosome-exposed populations (96) it was found that pro-inflammatory IL-8 and TNF-α PBMC responses to the bacterial TLR4 ligand LPS and to the schistosome-derived TLR2 ligand were suppressed in infected compared to uninfected children. However, IL-6 and IL-8 PBMC responses to two other schistosome glycolipids (without a known TLR affinity) were elevated in the infected individuals. A phosphorylcholine-containing glycoprotein ES product from filarial nematodes is also known to influence maturing murine DCs towards a Th2-driving phenotype (97). In the case of GI nematodes, Shinkai et al. (98) demonstrated that SCID mice infected with Nippostrongylus brasiliensis were able to produce Th2-polarized cellular responses in the absence of B- and T-cells. Also, Balic et al. (99) found that in N. brasiliensis an ES glycoprotein fraction induced the maturation of DCs with Th2-polarizing activity. Very little is known of the potential influences on APCs by A. lumbricoides and T. trichiura. However, Lochnit et al. (100) found a neutral glycosphingolipid fraction derived from A. suum (a close relative of A. lumbricoides) induced pro-inflammatory (IL-1, IL-6, TNF-α) ex vivo cytokine responses in human PBMCs. Also, in a recent field study, Jackson et al. (manuscript submitted) examined ex vivo monocyte cytokine responses to bacterial TLR2 and TLR4 ligands in a Tanzanian population exposed to A. lumbricoides, T. trichiura and hookworm. Total FECs (and species-specific FECs for T. trichiura and hookworm) were positively associated with pro-inflammatory TNF-α responses to TLR4 stimulation, but unrelated to regulatory cytokine expression (TGF-β, IL-10). These results are potentially at odds with a ‘modified Th2’ paradigm (3) that worms induce immunosuppressive effects. They are consistent, however, with a necessary role for TLR4 in the establishment of chronic infection in the murine T. muris model (101).

Although it is therefore becoming clear that worms can influence signalling by innate cells, more immunoepidemiological and model studies are required to interpret the effects of GI nematode infection on the human innate immune system.

POTENTIAL MANIPULATIVE STRATEGIES BY PARASITES

Immunological phenotype during infection results from the potentially conflicting influences of host immunoregulatory control and parasite immunomodulatory activities on the developing immune response. It is possible that this conflict could produce outcomes non-optimal for, and not fully under the control of, either the host or parasite. It is therefore difficult to rigorously assess whether the putative Th2/Treg host response phenotype in chronic helminth infections results from a manipulative strategy by parasites, a protective response by the host, a mutually adaptive compromise between these, or an uncontrolled and suboptimal outcome for both parties. In Figure 2 the three different T-cell phenotypes (Th1, Th2, Treg) are considered in terms of the potential selective pressures their induction might exert on host immunoregulatory strategy and parasite manipulative strategy in worm infections. Consideration of these pressures suggests that, as a strategy, from the point of view of either the host or the parasite, the Th2/Treg scenario presents a number of apparent paradoxes.

Figure 2.

Schematic figure showing potential selection pressures (on host immunoregulatory strategy and on parasite manipulative strategy) resulting from bias towards different T-helper phenotypes (Th1, Th2, Treg) in gastrointestinal (GI) nematode infections. Possible positive and negative pressures are indicated by arrows.

In the case of the parasite, murine models indicate that Th1 responses are clearly associated with the lack of ability to expel worms from the gut. It might therefore be expected that at least some intestinal worms should attempt to trigger such responses. Furthermore, an emphasis on Th2 might be expected to be deleterious to worms in the gut. One explanation for the deliberate induction of Th2 pathways by worms might be if this increased the relative fitness of individuals able to avoid killing or expulsion. In particular, a Th2 response phenotype could result in a concomitant immunity effect, where larval infections are killed more effectively than adults (54). Brown & Grenfell (102) have suggested, on the basis of theoretical analyses, that parasite-induced concomitant immunity is a plausible scenario, even when the costs of host manipulation lead to a conflict between group and individual interests (i.e. where some individuals within an infection can exploit others by ceasing to actively manipulate the host). In contrast to Th2-mediated processes, the Treg aspect of the host response has direct benefits for worms, tending to reduce immune attack and maintain niche integrity by decreasing the danger of immune-mediated self-damage by the host. However, some indirect negative consequences of a Treg/Th2 immune phenotype could arise from a failure of the host to resist microbial infections (usually controlled by Th1 responses), thereby compromising the long-term stability of the parasite's niche.

For the host, the Th2/Treg phenotype might appear to be a maladaptive outcome. Individuals become permissive to worm infections and the bias away from Th1 responses may compromise their defences against bystander microbial infections. Essentially, this is the basis for linking de-worming treatment and potential improvements in infectious disease resistance and vaccination (2), discussed in more detail below. However, evolutionary pressures favouring a ‘deliberate’ Th2/Treg strategy by the host could include reduction in the energetic costs of mounting effector responses and the avoidance of immunopathological damage. The Th2 character of these responses may also limit recolonization of the host by juveniles and prevent the accumulation of very large adult worm populations.

These considerations, and the apparent ubiquity of the Th2/Treg chronic infection phenotype suggest that it may be an adaptive compromise for both the host and the worm. However, given the complexity of the selective pressures acting (Figure 2) it is also possible that more than one evolutionary trajectory is possible – and that anti-inflammatory strategies by both the host and worms may not always be the rule. It may also be that different pressures act on different worm life-history stages. The short-term individual interests of an invading L4 larva (in terms of surviving to maturity) may lie in generating ineffective Th1 inflammatory responses. Established worms, on the other hand, would increase their relative fitness by preventing the establishment of new competitors – perhaps through the induction of Th2 responses in the gut. Given the greater body size of established adults compared to invading larvae and juveniles, their relative ability to undertake immunomodulatory activities may be much greater (per individual). However, the degree of larval infection pressure might also have a bearing on any adult vs. juvenile conflict. Interestingly, a recent study of a Cameroonian population hyperendemic for A. lumbricoides found greater Th2 polarization and an association of Th2 cytokines with resistance to A. lumbricoides in older (12–36 years) but not younger (4–11 years) age classes. A second study in a relatively lightly infected Cameroonian population (50) found more polarized Th2 phenotypes and an association between Th2 cytokines and resistance to A. lumbricoides and T. trichiura in younger (4–13 years) but not older (14–57 years) individuals. Possibly this difference could be explained by lower larval infection pressure at the second site and a greater immunomodulatory influence of adult worms? If this were true, then the lack of association between Th1 responses and faecal egg counts in previous field studies might occur because pro-inflammatory mediators are more likely to be promoted by pre-reproductive parasites.

As indicated at the start of this section, the conflictive nature of the host–parasite relationship means it is difficult to assign an observed infection outcome to ‘strategy’ by either the host or parasite. How, then, might the possibility of ‘deliberate’ parasite immunomodulatory effects be identified? Perhaps the best evidence for an adaptive manipulation strategy by worms would be the demonstration of parasite molecules with an unambiguously specific action against, or relationship to, host immune mediators or effectors. For example, this could include secretory products that directly mimic host signalling molecules (63,103) as a result of convergent evolution. This would require that the similarity between functional areas of the host and parasite molecules be greater than expected in two independently evolving organisms at the same level of evolutionary divergence. Comparative functional studies on the secretory products of intraspecific strains with different immunomodulatory abilities may also be revealing. In the case of the host, the existence of dedicated pattern response receptors recognizing worm PAMPs and producing Th2/Treg mediators on stimulation (95) might be evidence for a ‘deliberate’ strategy.

IMPLICATIONS FOR DE-WORMING HUMAN POPULATIONS

Vertebrates and the main groups of worms infecting them have co-evolved for hundreds of millions of years – probably since the appearance of primitive tetrapods in the case of nematodes and even earlier for platyhelminth parasites. Worm infections are also near ubiquitous in present-day vertebrate wildlife populations and there is no reason to assume they have been significantly less common in the recent past. Given these facts, it seems reasonable to assert that mammals probably evolved in an environment where individuals were under constant infection pressure from worms. The mammalian immune system may thus have been selected to work optimally whilst expressing the immunoregulatory phenotype typically induced by chronic worm infections. In relation to this idea, it has been proposed (104) that failure to express the worm-induced immunological phenotype in helminth-free populations could potentially explain the frequency of diseases involving immune dysfunction/misregulation in western countries. This is supported by a number of strands of evidence. Firstly, endemic helminth infection and the elevated incidence of atopy have mutually exclusive geographical distributions, respectively, in the developed and developing world. Secondly, atopy involves misregulated responses by the Th2 arm of the immune system that seems to have evolved to provide protection against worms and other macroparasites. Thirdly, negative associations between atopy and helminth infections have also been observed in a number of field studies in developing countries (reviewed elsewhere in this volume).

Amongst human populations, the distribution of major pathogenic helminths coincides geographically with many serious microbial diseases, such as tuberculosis, HIV and malaria. In terms of immunoregulatory influences, GI nematodes (including A. lumbricoides and T. trichiura) could be particularly important due to their very wide distribution (1). Although protective against atopy, one of the major effects of the worm-induced Treg/Th2 phenotype is to suppress the Th1 inflammatory responses that are of critical importance in effective immune responses against most microbial diseases. It has been anticipated that the removal of worms might increase resistance to bystander infections (105,106) by affecting the Th1/Th2 balance and restoring ‘normal’ levels of T-cell responsiveness to antigens. However, it should be noted that mechanisms employed by hosts to control severe microbial infections, and to limit immunopathology, may also have evolved to perform optimally within the worm-induced Treg/Th2 phenotype. It is therefore possible that the absence of worms might lead to stronger, misregulated Th1 inflammatory reactions – and to immunopathological damage. This could be important where infectious diseases with severe immunopathological effects, such as malaria, are endemic. Field records of decreased severe malaria in helminth-infected individuals (82) support a role for worms in the control of immunopathology. Furthermore, TGF-β secretion is negatively associated with infection immunopathology in a murine model of malaria (107). Also, Schopf et al. (48) found that mortality and morbidity in IL-10-deficient mice infected with T. muris could largely be attributed to opportunistic bacterial infections and the unregulated action of the pro-inflammatory mediator IL-12. These experimental studies support a role for the suppressive cytokines associated with worm infections (IL-10, TGF-β) in ameliorating immunopathology due to bystander infections.

The limited information currently available therefore poses important questions about the effect de-worming human populations infected with GI nematodes could have on host susceptibility to different pathogens and on pathogen-induced immunopathology. In order to plan de-worming strategies with more confidence, further study is urgently required to assess the effect of de-worming on disease caused by important immunopathogenic bystander infections.

Ancillary