INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Giardia lamblia is a common cause of both acute and chronic diarrheal disease in humans (reviewed in reference 1). In many regions of the world, giardiasis is endemic and infection is practically universal by 2 years of age (1). In developed countries, infections are more sporadic but nevertheless common whenever fecal contamination occurs, such as with contamination of water supplies or direct person-to-person spread in day care centers. The courses of infections are highly variable among individuals; some infections resolve quickly, whereas others can continue for years. This variability may be due to differences in pathogenicity among parasite isolates as well as to differences in host responses (3).

Several lines of evidence suggested that antibodies and T cells are required to control Giardia infections. Many studies have focused on immunoglobulin A (IgA) since it is found predominantly on mucosal surfaces. Infections of humans and rodents with G. lamblia and Giardia muris lead to the production of parasite-specific antibodies, including antibodies of the IgA isotype (5, 18, 28). Parasites recovered from infected animals were shown to be coated with IgA (19), and IgM and IgG antibodies have been shown to be cytotoxic in vitro by complement-dependent and complement-independent mechanisms (34). Furthermore, hypogammaglobulinemia is often associated with chronic giardiasis in humans (reviewed in references 1 and 40). Together, these data have led to the hypothesis that antibodies, particularly of the IgA isotype, are required to control G. lamblia infections. This idea has been further supported by experimental infections in mice with the related parasite G. muris. Mice depleted of B cells by treatment with anti-IgM antibodies were unable to control G. muris replication (44) as were xid mutant mice, which have reduced numbers of B cells (45). Finally, the discovery of antigenic variation of the surface proteins of G. lamblia was also consistent with a role for antibody in controlling the parasite, since antigenic variation is typically thought to be a mechanism used by microorganisms to evade host antibody responses (2, 9, 35). T cells are also important in controlling Giardia infections. Nude mice and anti-CD4 antibody-injected mice were unable to control G. muris replication (20, 47). Similarly, neonatal nude and SCID mice were unable to control infections with G. lamblia (13). However, it was unclear from these studies whether T cells were directly involved in eliminating the parasites or whether they were needed merely to augment production of antibodies.

Because Giardia replicates only in the lumen of the small intestine, we were interested in directly addressing the role of antibodies in controlling G. lamblia infections as well as defining any other components of the mucosal immune system responsible for controlling parasite infections. We therefore took advantage of a model of acute G. lamblia infection in adult mice (7). In this model parasites are introduced by gavage and replicate in the small intestines of the mice until parasite numbers drop dramatically between 1 and 2 weeks postinfection. However, small numbers of parasites continue to be detectable by culturing the intestinal contents for several months, although they cannot be detected by visual inspection of intestinal contents. We have used both G. lamblia and G. muris to infect B-cell-deficient mice and show that there is little difference in the levels of parasite clearance between these mice and wild-type mice for either parasite. In contrast, we show that -T-cell-receptor (-TCR)-bearing, CD4⁺ T cells are required for control of acute G. lamblia infections in mice and that this requirement persists in the absence of B cells. Finally, infections in several lines of cytokine-deficient mice demonstrate that neither the Th1 nor the Th2 subset is absolutely required for control of acute G. lamblia infections.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. C57BL/6J (wild-type), BALB/cJ (wild-type), C57BL/6J Igh-6 tm1 Cgn (B-cell-deficient, B6 Igh-6 [26]), C57BL/6J Tcrb tm1 Mom (-T-cell-deficient, B6 TCR [32]), C57BL/6J Tcrd tm1 Mom (-T-cell-deficient, B6 TCR [23]), C57BL/6J scid (T- and B-cell-deficient, B6 SCID), BALB/cJ STAT-6 tm1 Gru (signal transducer and activator of transcription 6 [STAT-6]-deficient, BA STAT-6 [25]), C57BL/6J IFN tm1 Ste (gamma interferon [IFN-]-deficient, B6 IFN- [8]), and C57BL/6J IL-4 tm1 Nnt (interleukin-4 [IL-4]-deficient, B6 IL-4 [37]) mice were obtained from the Jackson Laboratories, Bar Harbor, Maine. IL-4 and IL-4 receptor gamma chain (IL-4R)-deficient mice on an inbred BALB/c background (BA IL-4 and BA IL-4R) were kindly provided by Nancy Noben-Trauth (National Institute of Allergy and Infectious Diseases, Rockville, Md.) (37, 38). All mouse experiments were performed in accordance with institutional guidelines.

Parasites and infections. The clone of G. lamblia GS/M-H7 has been previously described (3). Infections were initiated with 500,000 in vitro-cultivated trophozoites suspended in 0.1 ml of TYI-S-33 medium given by gavage (7). Groups of three mice were euthanized at various time points, and the proximal 10-cm sections of their small intestines were removed and placed in 5 ml of TYI-S-33 medium on ice. The small intestines were minced, and after 15 min on ice, parasites were counted with a hemocytometer. At least four separate grids on the hemocytometer were counted for each mouse, and mice with no detectable parasites were considered to have cleared the infections. Since a single parasite on one grid corresponds to 10⁴ trophozoites/ml, the limits of detection were roughly 10⁴ trophozoites/mouse. G. muris cysts were obtained from Frank Schaefer (U.S. Environmental Protection Agency, Cincinnati, Ohio), and an inoculum of 1,000 cysts/mouse was used to initiate infections by gavage.

VSP analysis. Expression of the H7 variant-specific surface protein (VSP) was analyzed essentially as described previously (3). After quantitation, parasites were cultured in 15 ml of TYI-S-33 medium in glass tubes at 37°C for 15 min to allow attachment. Nonadherent material was decanted, and tubes were filled with phosphate-buffered saline and chilled on ice to detach parasites. After concentration by centrifugation, parasites were then adhered to glass slides at 37°C, fixed in methanol-acetone (1:1), air dried, and stained with monoclonal antibody G10/4, which reacts with the VSP expressed by the H7 clone of the parasite. Staining was visualized with goat-anti-mouse IgG, IgA, IgM, and fluorescein isothiocyanate (Cappel) and viewed on a Zeiss axiophot microscope. Negative controls were incubated without primary antibody and showed no staining.

Anti-CD4 antibody treatment. Mice were depleted of CD4⁺ T cells by administration of 2 mg of the monoclonal antibody GK1.5 given intraperitoneally in phosphate-buffered saline on days 4, 0, +6, and +9 pre- and postinfection (10). Control animals similarly received 2 mg of an isotype-matched monoclonal antibody, GL113. Both antibodies were provided by A. Sher (National Institute of Allergy and Infectious Diseases) as ammonium sulfate-purified ascites fluid. CD4⁺-T-cell depletion was >95% effective as determined by fluorescence-activated cell sorter analysis of splenic lymphocytes with anti-CD3 phycoerythrin and anti-CD8 fluorescein isothiocyanate (Pharmingen, San Diego, Calif.) at the termination of the experiment (data not shown).

Statistics. Comparisons between the numbers of infected animals per group were done by chi-square analyses. Comparisons between the numbers of parasites observed in different groups were done by a z test. Mice with undetectable parasites (<10⁴) were considered to have a statistical value of 0 for statistical comparisons. Both tests were done with Excel software (Microsoft Corp., Redmond, Wash.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

We began by investigating whether antibodies are required for control of acute infections with G. lamblia. Wild-type or B-cell-deficient mice which lack all antibody due to disruption of the Ig heavy-chain (IgH) locus (26) were infected with G. lamblia trophozoites, and parasite loads were determined at various times after infection. Table 1 shows that while all the mice were infected 1 week after infection, B-cell-deficient mice, -T-cell-deficient mice, and SCID mice all had significantly more parasites than did wild-type controls, suggesting that both B-cell and T-cell functions can affect parasite load. However, while both B-cell-deficient and wild-type mice eliminated the majority of parasites by 2 weeks postinfection and had no detectable parasites (<10⁴/mouse) 4 weeks postinfection, -T-cell-deficient mice and scid mice were completely unable to control the parasites, suggesting a crucial role for T cells in control of this parasite (Table 1) (7). There was some indication that SCID mice might harbor more parasites than the B-cell or TCR-deficient mice 7 days postinfection (Table 1, experiment 1), suggesting that the role of T cells and B cells in controlling parasite loads might be additive.

G. lamblia undergoes antigenic variation during infections of immunocompetent hosts (2, 35). The VSPs expressed by parasites in the inoculum are replaced by antigenically distinct VSPs, presumably due to elimination of the parasites expressing the original VSP and outgrowth of parasites which have spontaneously switched to expression of new VSPs. Gottstein and Nash showed that neonatal nude mice were capable of selecting for antigenic variants but that neonatal SCID mice were not (13), and Stäger and Müller recently showed that selection of antigenic variants failed to occur in B-cell-deficient mice, showing that B cells are required for selecting antigenic variants of G. lamblia (46). Our analysis of the VSPs expressed by the parasites recovered from all of these strains confirmed that B cells were required for the selection of antigenic variants (data not shown). Interestingly, antigenic variants were selected for in the TCR mutant mice even though the parasites could not be eliminated. It is possible that IgM is able to select for antigenic variants or that the T cells are able to provide help for isotype switching in the TCR-deficient mice. Thus, while antibodies are essential for selection of VSPs, they are not required for control of parasite numbers in the gut.

G. muris naturally infects mice and other rodents. Studies of G. muris infections in anti-IgM-treated mice and xid mutant mice have suggested that antibodies are required for parasite clearance (44, 45). We wanted to see if our results with G. lamblia in B-cell-deficient mice differed from the results of these earlier studies due to differences between G. lamblia and G. muris. We therefore infected wild-type and B-cell-deficient mice with G. muris cysts. While the clearance of G. muris in mice is slower than for G. lamblia, both wild-type and B-cell-deficient mice eliminated the majority of parasites by 3 weeks postinfection (Table 2). Thus, the data indicate that antibodies are not required for control of either G. lamblia or G. muris infections. Importantly, our results are not unique to the adult-mouse model of G. lamblia infections, as G. muris is also eliminated in B-cell-deficient mice. Furthermore, our results can be directly compared to those of earlier experiments. The differences between our results and those previously published for G. muris are likely due to conditions in B-cell-deficient mice resulting from disruption of the IgH locus (26) that are different from conditions in mice with the xid mutation or following treatment with anti-IgM antibodies (44, 45).

Since antibodies were not required for control of acute Giardia infections in mice, we examined the role of T cells in these infections. CD4⁺ T cells were shown to be required for control of G. muris infections (20). In order to determine whether these T cells were required to augment antibody production or if they acted through a distinct pathway, 4 days prior to infection with G. lamblia we treated both wild-type and B-cell-deficient mice with anti-CD4 antibody. As had been seen for G. muris, depletion of the CD4⁺ T cells prevented G. lamblia clearance in wild-type mice (Table 3). In addition, depletion of CD4⁺ T cells in B-cell-deficient mice prevented the elimination of the parasites, suggesting that the role of these T cells is not merely to provide help for antibody production and that there is a distinct antibody-independent, but T-cell-dependent, pathway leading to elimination of the parasite.

The T-cell compartment in the small intestine contains both -TCR- and -TCR-bearing T cells. While the former are known to recognize peptide antigens bound to molecules encoded in the major histocompatibility complex, it is still unclear which ligands are recognized by the latter, although recent data suggest that they recognize invariant molecules expressed by intestinal epithelial cells under certain conditions (15). We therefore infected mice deficient in either T-cell subset due to disruption of either the TCR or - locus (23, 32). As seen in Table 4, the response to G. lamblia is clearly dependent on the -TCR-bearing T cells and not the -TCR-bearing T cells. Most of the mice were infected with Giardia at day 7 postinfection. The low numbers of parasites seen in the wild-type and TCR-deficient mice in this experiment at day 7 suggest that immunity-mediated clearance may have already begun by day 7. While wild-type and TCR-deficient mice eliminated the majority of parasites by 2 weeks postinfection, the TCR-deficient mice could not control this acute phase of the infection. The reduction in the number of parasites from week 1 to week 4 postinfection seen in the TCR-deficient mice (Table 1, experiment 2) was not reproducible (Table 1, experiment 1).

In many other parasitic infections the immune response becomes biased towards either the Th1 or the Th2 subset (42). To determine if either subset was required for protection in this model of G. lamblia infection, we infected mice deficient in the Th1 subset due to disruption of the IFN- gene (8) or in the Th2 subset due to disruptions of the gene for IL-4, IL-4R, or STAT-6 (25, 37, 38). Table 5 shows that all of these mice were able to control the acute phase of the infection in a manner similar to the way wild-type mice could, although mice lacking IFN- showed a slight delay in eliminating the parasites in one experiment. Interestingly, the STAT-6 knockouts lacked parasites even on day 7, suggesting that they might in fact be hyperimmune to this parasite or that they were refractory to infection initially.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we clearly demonstrate that a T-cell-dependent mechanism is required for control of acute G. lamblia infections in adult mice. However, since our assay for parasites cannot discriminate between low parasite burdens and complete elimination of infections, no role for T cells in controlling chronic G. lamblia infections can be assigned at this time. TCR-deficient mice maintain high parasite burdens in the small intestine for over 4 weeks postinfection. In addition, depletion of CD4⁺ T cells from either wild-type or B-cell-deficient mice significantly enhances infections, suggesting that the role of these T cells is distinct from their ability to augment antibody production. We also show that T cells expressing TCRs but not those expressing TCRs are able to mediate clearance of G. lamblia. Finally, we show that neither IFN--producing Th1 cells nor IL-4- or IL-13-dependent Th2 cells are absolutely required for control of this parasite.

In contrast to T cells, B cells and antibodies are not required for control of acute G. lamblia or G. muris infections in mice (Tables 1 to 3). Antibodies can be cytotoxic to Giardia in vitro (33), however, and may be able to reduce the parasite load during the acute phase of infection (Tables 1 and 3 but not Table 2). It is also possible that antibodies may be essential for controlling parasites during the chronic phase of infection. Indeed, Stäger and Müller found that G. lamblia infections persisted longer in B-cell-deficient mice than in wild-type mice, consistent with a role for antibody during the chronic phase of infection (46). Antibodies are also required for selection of VSPs during infections. Importantly, however, our results demonstrate that an antibody-independent mechanism exists and is essential for control of the acute phase of Giardia infections.

How can our experiments demonstrating that T cells are required to control G. lamblia infections be reconciled with the results of previous studies indicating an essential role for B cells in controlling G. muris infections (44, 45)? Since in our experiments both G. muris and G. lamblia infections are controlled in B-cell-deficient mice, it seems likely that the differences observed are due to differences between the mice we have used and those used in earlier studies. While Snider et al. used either wild-type mice treated with anti-IgM or xid mutant mice, we have used mice with a disruption of the IgH locus (44). Disruption of the IgH locus produces mice completely devoid of both antibodies and B cells (26). It is also now known that anti-IgM treatment can have effects on the immune system in addition to the depletion of B cells and antibody. For example, Fc receptor ligation on macrophages can inhibit the production of IL-12 in response to lipopolysaccharide (48, 49). Thus, anti-IgM treatment might have affected cells other than B cells in ways which prevented control of the parasite. Similarly, while xid mutant mice, which have a defect in the btk gene, exhibit abnormal B-cell development, they also have abnormal mast cells that are deficient in cytokine production following Fc receptor cross-linking (17). Thus, the interpretation of earlier experiments should be recast in light of recent findings. The earlier results seen with G. muris do not necessarily imply a requirement for B cells in controlling infections. The ability of IgH-targeted mice to control Giardia infections is therefore most consistent with antibody not being essential during the acute phase of infection. Importantly, the fact that depleting CD4⁺ T cells in these B-cell-deficient mice prevented the infection from being controlled argues that this antibody-independent pathway for control of the parasite is indeed T cell dependent (Table 3).

In addition to the results of G. muris infections in mice, the association between chronic Giardia infection and hypogammaglobulinemia in humans has also been used to argue for a requirement for B cells in the immune response to Giardia. Hypogammaglobulinemia, however, can have multiple etiologies (51). These include various T-cell defects (43) as well as mutations of the btk gene, which, as noted above for xid in mice, is expressed by cells in addition to B cells. Thus, the association of chronic giardiasis with hypogammaglobulinemia cannot exclude the presence of a T-cell-dependent, but antibody-independent, mechanism for controlling this parasite during the acute phase of infections. Alternatively, antibody may play an important role in controlling parasites during the chronic phase of infection, when low numbers of parasites persist in the host for several months. This persistence may be particularly important in human disease, since infections can last a long time and recur. Our study did not address the role of antibody in chronic infections.

Antigenic variation is a common way for pathogenic microorganisms to evade the host antibody response. The experiments with B- and T-cell-deficient mice reported here show that while B cells are necessary to select for antigenic variants, T cells are not. This finding is in contrast to the requirement for T cells, but not B cells, to eliminate the majority of parasites during the acute phase of the infection. The discovery of the antigenic variation of the major surface protein of G. lamblia augured poorly for the development of a vaccine against G. lamblia, since the common belief was that antibodies reactive with this protein are the major protective mechanism against this parasite (34, 46a). This suggests that antigenic variation may be important in the chronic phase of infection or perhaps in the transmission of the parasite. The discovery of antibody-independent pathways for eliminating the parasite, however, once again makes the idea of a protective vaccine feasible. Indeed, it has previously been shown that infection of gerbils with G. lamblia isolates with different surface coats can induce some cross-protective immunity (3). A study of gerbils has also shown that administration of steroids to suppress cellular immunity can lead to reactivation of chronic infections (27), consistent with an important role for T cells in controlling the parasite. The existence of these T-cell-dependent pathways may also allow development of novel immunotherapies for giardiasis.

Our study demonstrates that, instead of B cells, CD4⁺, -TCR-bearing T cells are essential and that they function in an antibody-independent pathway to promote parasite clearance (Tables 1, 3, and 4). The requirement for CD4⁺ T cells was not unexpected since anti-CD4 treatment was previously shown to prolong G. muris infections (20). Similarly, because the parasite is noninvasive, it would be unlikely for CD8⁺, class I major histocompatibility complex-restricted T cells to be important in their control. Indeed, class I-deficient mice (₂ microglobulin knockouts) controlled parasites as well as wild-type controls (data not shown). Finally, data from human studies may also suggest a role for T cells in controlling Giardia infections. Giardia infections are associated with nodular lymphoid hyperplasia in humans, a disease associated with decreased B- and T-cell functions (29, 36, 54).

A prominent feature of the mucosal immune system is the presence of T cells bearing TCRs (24). These, however, are not required for protection against Giardia (Table 4). Instead, the more conventional CD4⁺ -TCR-bearing T cells are required for control of parasite replication. This situation is similar to that in intracellular intestinal infections by other parasitic protozoa, e.g., Eimeria vermiformis and Cryptosporidium parvum, in which -TCR-bearing T cells are also thought not to be important in immunity, although they can regulate pathology (39, 41, 55). Thus, no essential role for these -TCR-bearing T cells in protection against intestinal infections can yet be assigned.

Infections with many protozoan and helminth parasites result in T-cell responses strongly biased toward either the Th1 or the Th2 subset (42). In particular, nematodes in the small intestine often lead to very strong Th2 responses, and these responses are, in some cases, absolutely required for elimination of these parasites (14, 50). On the other hand, mice deficient in IFN-, STAT-6, IL-4, and IL-4R all control Giardia infections (Table 5), indicating that Th1 and Th2 cells are not required together. This finding may reflect the fact that in the absence of IFN--producing Th1 cells, Th2 cells are sufficient for control of the parasite or that in the absence of Th2 cells, Th1 cells are sufficient. Alternately, the important cytokines for controlling Giardia may be produced by both the Th1 and the Th2 subset or neither subset. For example IL-3, which can augment mast cell activation in vitro (31), is produced by both Th1 and Th2 cells (33), and transforming growth factor  is typically produced by mucosal T cells sometimes labeled Th3 cells (22, 56). There was some delay in parasite clearance observed in the IFN--deficient mice (Table 5), suggesting that the Th1 response might have a role in controlling Giardia infections. However, these mice were able to completely eliminate the parasites by 4 weeks postinfection, indicating that this response is not essential. This result is consistent with earlier reports showing IFN- production by human intraepithelial and blood lymphocytes in response to G. lamblia parasites (11) and enhanced G. muris infections in mice treated with anti-IFN- antibody (52). We are currently examining the production of cytokines in the intestine during infection of wild-type mice with G. lamblia.

The ability of antibody-deficient mice to eliminate most of the parasites in an acute infection, however, suggests an alternate (and T-cell-dependent) effector mechanism for controlling Giardia. Several groups have previously implicated mast cells in the control of G. muris infections (12, 16, 53). And while c-kit^w/wv mutant mice that are deficient in mast cells are also unable to control infections with G. lamblia (our unpublished observations), it remains to be shown that reconstitution of these mice with mast cells restores immunity to Giardia and that other aspects of the c-kit mutant phenotype are not responsible. And while several IgE-independent pathways have been shown to exist for the activation of mast cells in vitro and in vivo (reviewed in reference 31 and see references 6, 21, and 30), it is difficult to reconcile a role for mast cells with the kinetics of the anti-Giardia response and the lack of effect of eliminating IL-4 from the system since it is a mast cell growth factor (31). G. lamblia has also been shown to be susceptible to defensins, antimicrobial peptides whose production might be influenced by T-cell products (4). The T-cell-dependent pathways responsible for controlling Giardia replication remain to be elucidated.