Giardia duodenalis: The Double-edged Sword of Immune Responses in Giardiasis

Overview

Giardiasis is one of the most common intestinal protozoan infections worldwide. The etiological agent, Giardia duodenalis (syn. G. lamblia, G. intestinalis), is a flagellated, binucleated protozoan parasite which infects a wide array of mammalian hosts (Adam, 2001). The symptoms of giardiasis include abdominal cramps, nausea, and acute or chronic diarrhea, with malabsorption and failure of children to thrive occurring in both sub-clinical and symptomatic disease (Thompson et al., 1993). Infections are transmitted by cysts which are excreted in the feces of infected humans and animals. Human giardiasis is distributed worldwide, with rates of detection between 2-5% in the developed world and 20-30% in the developing nations (Farthing, 1994). There is significant variation in the outcome of Giardia infections. Most infections are self-limiting, although re-infection is common in endemic areas and chronic infections also occur. Moreover, some individuals suffer from severe cramps, nausea and diarrhea while others escape these overt symptoms. This review will describe recent advances in parasite genetics and host immunity that are helping to shed light on this variability.

Epidemiology

Recent studies of Giardia have identified eight distinct genotypes within G. duodenalis, only two of which, assemblages A and B, are capable of infecting humans (reviewed in Thompson, 2009). These studies have led to a reevaluation of the zoonotic potential of this organism. Although parasites with both A and B genotypes can infect numerous mammalian species in addition to humans, other genotypes appear to have more restricted host range. Assemblages C and D, for example, are commonly found in dogs, but have yet to be reported in humans. Thus, the idea that human transmission from dogs (and cats and livestock) to humans needs to be reevaluated. Fortunately, new data are becoming available indicating that most cases of giardiasis are due to anthroponotic spread, but zoonotic transmission can and does occur (Snel et al., 2009).

Better understanding of the molecular epidemiology of Giardia will also require reanalysis of studies of a commercially available vaccine for veterinary giardiasis (Olson et al., 2000). This vaccine is essentially a mixture of lyophilized trophozoites of four parasite strains. Since these strains can be grown in culture they are likely from assemblages A and B, but not the assemblages commonly found in and restricted to cats (F), dogs (C and D) or livestock (E). Thus, while some studies have shown some protection against experimental infections (Olson et al., 1996; Olson et al., 1997; Olson et al., 2001), others failed to show such a protection against the parasite (Stein et al., 2003; Uehlinger et al., 2007; Anderson et al., 2004). For example, in one study vaccinated kittens had abnormal stools on fewer days, secreted fewer cysts, and had a significantly higher weight gain in the post-challenge period (Olson et al., 1996). Conversely, Stein and coworkers (2003) did not find any correlation between cats receiving 3 doses of a Giardia vaccine and reduction in cyst shedding compared to unvaccinated kittens. New veterinary vaccines will need to take into account the restricted host ranges of the different genotypes and work around our inability to culture those other than types A and B. Potential human vaccines will need to address the role of immune responses in contributing to pathology and determining which responses are protective, as opposed to those which are merely present.

The factors determining the variability in clinical outcome in giardiasis are still poorly understood (Buret, 2007). However, host factors (such as immune status, nutritional status and age), as well as differences in virulence and pathogenicity of Giardia strains are recognized as important determinants for the severity of infection (Haque et al., 2005). Numerous studies have attempted to correlate the development of symptoms to the presence of either assemblage A or B parasites. While individual studies often find a strong correlation between parasite genotype and virulence, the answer comparing across studies is very unclear. For example, one study in Dutch patients found assemblage A isolates solely in patients with intermittent diarrhea, while assemblage B isolates were present in patients with persistent diarrhea (Homan and Mank, 2001). In contrast, Guerden et al. (2009) found that infections with assemblage B parasites were commonly found in diarrhea patients, but that a high proportion of infections were with mixed assemblages that might have interfered with previous analyses. This may be due to the fact that assigning parasites to specific genotypes usually reflects alleles at loci such as glutamate dehydrogenase, 18S RNA and triose phoshate isomerase (TPI) which are unlikely to be directly associated with virulence. More effort, however, should be directed to understanding mechanisms of virulence and identifying specific parasite virulence factors in order to understand the relative contributions of both the host and the parasite to disease.

Immune responses that control infection

The immune response to microbial pathogens, including Giardia sp., relies on both innate and adaptive components. Although the actual host defense mechanisms responsible for controlling Giardia infections are poorly understood, many studies have demonstrated the development of adaptive immune responses as well as innate mechanisms in humans and other animals (Roxström-Lindquist et al., 2006). Understanding the complex network of immune responses and host-parasite cross-talk should assist us in identifying novel and common targets for the therapeutic intervention of the infection (Solaymani-Mohammadi et al., 2010).

Epidemiological studies suggest that previous infection with Giardia leads to a reduced risk of re-infection and to reduced development of overt symptoms in secondary infections. Analysis of cases in an outbreak at a ski resort in Colorado showed that individuals residing in the community for more than 2 years had a much lower risk of being affected than new residents (Istre et al., 1984). Similarly, a community in British Columbia experienced two outbreaks five years apart and individuals affected in the first outbreak were much less likely to be ill during the second outbreak (Isaac-Renton et al., 1994). Both studies suggest that previous exposure to Giardia produces an immunity to disease. It is unclear in these studies whether prior exposure actually prevented infection, or only if severe symptoms were avoided the second time. Nonetheless, these findings suggest that development of an effective vaccine could be feasible. A recent study in Brazilian children suggests that symptoms are less severe during re-infection, consistent with the idea that previous exposure doesn't prevent infection, but does reduce the pathology which can occur (Kohli, 2008). Additional studies in humans and animal models are, however, needed to determine what types of immune responses mediate this protection.

Studies in animal models require careful interpretation. First, animal immune responses are not always equivalent to that seen in humans. Additionally, many studies have utilized G. muris, a rodent parasite that cannot be cultured, to analyze immune responses in mice. As described below, however, G. muris may be resistant to immune mechanisms capable of killing G. duodenalis. Studies of G. duodenalis are restricted by the inability of many strains of G. duodenalis to colonize adult mice. Experiments with G. duodenalis in gerbils and neonatal mice have also been performed, although immunologic analysis of these animals is difficult. In some cases, understanding the host restriction of G. duodenalis may provide insights into the biology of the host-parasite interaction. For example, certain bacterial flora have been shown to render mice resistant to colonization (Singer and Nash, 2000b). Bacterial flora could inhibit Giardia through direct action against the parasite. Studies in vitro showed that culture supernatants from Lactobacillus johnsonii La1 significantly inhibited the proliferation of G. duodenalis trophozoites in a strain-dependent manner (Pérez et al., 2007). Diverse animal species (and even individuals) differ greatly in the dominant microbiota colonizing their intestinal mucosal surfaces. Thus, the different microbiota could explain why some hosts are resistant to specific parasitic infections. Host immune mechanisms play a key role in regulating the microbiota, e.g. through expression of diverse sets of antimicrobial defensins by Paneth cells (Salzman et al., 2007). Indeed, some Paneth cell defensins have also been show to kill G. duodenalis in vitro (Aley, 1994). However, the effects of defensins on intestinal flora is bidirectional, and changes in flora can also effect production of defensins (Ayabe, 2004). Further analysis of the host-parasite-environment three-way interaction in giardiasis may provide novel means to prevent this infection.

Several lines of evidence suggest that IgA antibodies contribute to protective immunity against giardiasis. While most chronic infections have been reported in patients with no underlying immune abnormality, patients with common variable immunodeficiency (CVID) and Bruton's X-linked aggamaglobulinemia (XLA) are clearly prone to chronic giardiasis (Stark et al., 2009). Patients with these syndromes both lack normal B cell function and reduced production of IgG, suggesting the necessity of antibodies for control of the infection. However, both CVID and XLA have additional defects in immune function and chronic Giardia infection is not common in selective IgA deficiency, suggesting additional layers of complexity in host immunity. As such, the rates of giardiasis in HIV-infected patients are higher than controls in some studies (e.g. Angarano et. al, 1997; Feitosa et. al, 2001; Bachur et al., 2008), although it is still unclear whether epidemiological factors rather than immunosuppression are responsible for these differences (Stark et al., 2009).

A number of studies have used animal models of infection to help clarify the role of antibodies in controlling Giardia infections. Snider et al. (1988) first showed that xid mice (hypogammaglobulinemic mice with a defect in the same kinase that is altered in human XLA) and wild-type mice treated with anti-IgM to deplete B cells developed chronic infections with G. muris. However, when we used gene-targeted B cell deficient mice and G. duodenalis for infections, no defect was apparent, suggesting alternate mechanisms exist to eradicate the infection (Singer and Nash, 2000a). This contradiction has been resolved by studies directly comparing G. muris and G. duodenalis infections in mice lacking the polyIg receptor that cannot transport IgA or IgM into the intestinal lumen (Davids et al., 2006). G. duodenalis infections were controlled in the absence of antibodies, while G. muris infections became chronic. This suggests that mice have additional mechanisms able to kill the human pathogen to which the mouse species is resistant. Identification of these mechanisms in mice and their human counterparts would therefore greatly facilitate development of effective vaccines or immunotherapeutics.

The existence of antibody-independent mechanisms for Giardia elimination does not necessarily mean that antibodies have no role. However, the ability of G. duodenalis to undergo extensive variation of the surface coat antigens, called variant-specific surface proteins (VSPs), likely delays the effectiveness of the antibody response (Nash, 1997). Mice deficient in the cytokine interleukin (IL)-6 develop chronic infections with G. duodenalis (Bienz et al., 2003; Zhou et al., 2003). In the first 2 weeks following infection, these mice make a strong IgA response which reacts with only a small subset of parasites growing in vitro, suggesting a response to a limited subset of parasite VSPs. In contrast, eight weeks after infection IL-6 deficient mice produce IgA reactive with all parasites from in vitro cultures, suggesting that the IgA response now recognizes all possible VSPs, common epitopes on VSPs, or invariant antigens on the parasite. The IL-6 deficient mice cleared their infections at this time point, indicating that such broadly reactive antibodies could indeed confer protection (Zhou et al., 2007).

In contrast to B cells, T cells appear to be required for control of Giardia at all time points post-infection. Nude mice, mice treated with anti-CD4 and mice lacking the T cell receptor β gene all develop chronic infections with G. muris and/or G. duodenalis (Roberts-Thomson and Mitchell, 1978; Stevens et al., 1978; Heyworth et al., 1987; Singer and Nash, 2000a). One role of T cells is to provide help for production of antibodies. These cells also provide help for other aspects of the immune response, although it is unclear which of these are essential for parasite control. Human studies have provided limited information about cytokine levels in sera of patients. In one study, peripheral blood lymphocytes from naïve individuals produced interferon (IFN)-γ in response to trophozoites (Ebert, 1999), and sera from infected adults contained elevated levels of IL-5, IL-6 and IFN-γ (Matowicka-Karna, et al., 2009). Mice lacking either IL-4 or IFN-γ are able to control infections with kinetics similar to wild-type mice (Singer and Nash, 2000a). We have recently shown that mice lacking TNF are delayed in parasite elimination (Zhou et al., 2007). This is consistent with elevated TNF levels in sera of infected children (Bayraktar et al., 2005). Altogether, additional studies are warranted to help us better understand regulation of T cell development and the roles of these immune cells during giardiasis.

Several recent studies have begun to examine how the innate immune system responds to Giardia infection and how this might influence development of T cell responses. An early study from the Kagnoff group examined the response of Caco2 cells to G. duodenalis parasites and saw no change in the expression of numerous cytokines that were induced by other intestinal pathogens (Jung et al., 1995). Roxström-Lindquist et al. (2005) used microarrays to reexamine this interaction and discovered a strong response by a limited number of genes, including several chemokines that might be expected to recruit dendritic cells and lymphocytes to the intestinal mucosa. Unfortunately, studies in vivo have not determined the importance of these factors during infection. Our lab has recently examined the interaction of dendritic cells with Giardia and found that live parasites and parasite extracts activate these cells for antigen presentation, but inhibit secretion of IL-12 while enhancing production of IL-10 (Kamda and Singer, 2009). Inhibition of IL-12 production was partially mediated via IL-10, but could be more completely reversed by inhibiting phosphoinositide 2-kinase with wortmannin. The host receptors and parasite ligands involved in this inhibition remain to be identified. Interestingly, this inhibition of IL-12 production could suggest that the parasite actively restricts the development of severe inflammation and may contribute to the lack of pathology often seen during infections. How dendritic cells and epithelial cells interact to determine the development of T cell responses needs to be determined.

Several effector mechanisms have been proposed to eradicate Giardia infections (reviewed in Eckmann, 2003). These include phagocytosis by macrophages, secretion of defensins, nitric oxide (NO) and mucins by epithelial cells as well as the recruitment and activation of mucosal mast cells. Macrophages have been shown to ingest and kill Giardia trophozoites in vitro, but a role in vivo has not been demonstrated (reviewed by Smith, 1985). Likewise, while some α-defensins can lyse trophozoites in vitro (Aley, et al., 1994), Eckmann (2003) reported that mice lacking MMP-7, the protease which converts pro-defensin peptides to active defensins, had no deficit in controlling G. muris infections. NO is cytostatic for G. duodenalis in vitro, but mice lacking inducible nitric oxide synthase (iNOS) eradicate G. duodenalis infections normally (Eckmann et al., 2000; Li et al., 2006; Andersen et al., 2006). Interestingly, the neuronal NOS isoform did have a defect in parasite control, but this was more likely due to effects on intestinal motility and not a direct effect on the parasite. The activity of mucins against the parasite has also only been shown in vitro (Roskens and Erlandsen, 2002).

In contrast to these other effector mechanisms, a significant role for mast cells in controlling Giardia has been demonstrated in vivo in several animal models. In an early study, mast cell deficient (wf/wf) mice, unlike their wild-type littermates, developed chronic giardiasis after oral administration of G. muris. Furthermore, the susceptible BALB/c mice injected with the anti-histamine and anti-serotonin drug, cyproheptadine, also showed prolonged infections with G. muris, implying some roles for mast cells in host's ability to eliminate the parasite (Erlich et al., 1983). A potential role for mast cells was suggested by Hardin et al. (1997), who examined jejunal macromolecular transport, epithelial permeability, and mucosal and connective tissue mast cell counts in Mongolian gerbils infected with G. duodenalis. In this study, mast cell counts were correlated with epithelial permeability and antigen uptake, suggesting a link between development of other immune responses and mast cell recruitment. This was further supported by data from our laboratory showing that c-kit^w/wv mice infected with G. duodenalis failed to make parasite-specific IgA, mount a mast cell response, or to eliminate the infection (Li et al., 2004). Anti-c-kit-treated C57BL/6 mice had normal IgA responses, lacked mast cell responses, and failed to control the infection within 10 days. Hence, while the requirement for mast cells is clear, their role in development of antibody responses is not. A recent study, however, found that mast cells can help drive B cell switching to production of IgA (Merluzzi et al., 2010). We also recently described an additional role for mast cells in mediating smooth muscle contractions and contributing to changes in intestinal motility following infection (see below). It is likely that this contributes to both protection and pathology in giardiasis.

If mast cells are needed for elimination of Giardia infection, then prior or concurrent infection with intestinal helminthes that induce mastocytosis might be expected to prevent infection. However, pre-infection with Trichinella spiralis led to an increase in the number of trophozoites in the intestine and a delay in Giardia elimination, despite a large increase in mast cell numbers and normal production of anti-Giardia IgA (von Allmen et al., 2006). Thus, while mast cells are necessary for Giardia elimination, they are clearly not sufficient on their own. Indeed, the strong Th2 environment created by T. spiralis infection may make the intestinal tract more conducive to Giardia survival. Given that in most endemic areas where giardiasis is endemic, the overwhelming majority of inhabitants are infected with other parasitic protozoa and helminths as well, additional study of the interactions between G. duodenalis and other infections is clearly needed.

Murine models of infection have also been used to study possible new vaccine strategies for giardiasis. A recombinant live attenuated vaccine, based on a gene segment of the VSP (VSPH7) expressed in a Salmonella typhimurium vaccine strain, was evaluated for its potential to induce local and systemic immune responses (Stager et al., 1997). The recombinant vaccine elicited synthesis of serum IgG as well secretory IgA (sIgA) against both bacterial antigens and the VSP. Vaccinated mice were never challenged with live parasites; however, several experiments have examined the possibility of transmission-blocking vaccines that target cyst antigens. Larocque et al. (2003) showed that oral immunization with recombinant cyst wall protein (CWP2) could elicit IgA and mixed Th1/Th2 responses and reduce cyst shedding after live challenge with G. muris cysts. Similarly, delivery of CWP2 expressed in Streptococcus gordonii or Lactococcus lactis elicited high levels of antibodies in the vaccinated mice and also reduced cyst shedding after challenge (Lee and Faubert, 2006 a,b). Most recently, a DNA vaccine encoding CWP2 delivered to the intestinal mucosa via S. typhimurium led to production of antibody, a mixed Th1/Th2 response and reduced cyst shedding compared to control mice (Abdul-Wahid and Faubert, 2007). Analysis of vaccine efficacy in mice lacking elements of the immune system, e.g. IL-4 deficiency, would be useful in determining which of the observed responses actually contribute to the observed protection.

Pathology and the Immune Response

Pathology in giardiasis is understood to arise in several ways. These include breakdown of the epithelial barrier, defects in the epithelial brush border, secretion of chloride ions and hypermotility of the intestinal smooth muscles. The role of parasite virulence factors and host immune responses in contributing to these mechanisms is beginning to be clarified, although much more remains to be understood.

Breakdown of the epithelial barrier has been shown to occur in human disease and animal models of giardiasis (Hardin et al., 1997; Troeger et al., 2007; Zhou et al., 2007). This allows for macromolecules and electrolytes to pass into the sub-mucosa, bypassing normal uptake by epithelial cells. The paracellular flow of nutrients and electrolytes can contribute to nutrient malabosrption by reducing electrochemical gradients needed for proper uptake and can cause inflammation in some individuals, probably through activation of innate immune effectors like macrophages. However, inflammation is often not observed in human giardiasis and does not correlate well to symptomatic disease (Oberhuber et al., 1997). In humans, epithelial barrier defects correlated with alteration in the level of expression of claudin-1, a component of the tight junctions (Troeger et al., 2007). In tissue culture models, changes in another tight junction protein, zona occludens (ZO)-1, as well as the cytoskeletal proteins actin and α-actinin have also been demonstrated (Teoh et al., 2000; Buret et al., 2002). Giardia has also been shown to increase epithelial permeability by inducing apoptosis in epithelial cells both in vivo and in vitro (Chin et al., 2002; Troeger et al., 2007). Interestingly, the magnitude of the apoptotic response in vitro is highly strain-dependent, although no specific parasite markers have been identified yet which can induce apoptosis (Chin et al., 2002). Finally, excess ion secretion was recently demonstrated ex vivo in biopsy material from giardiasis patients, suggesting that this too may contribute to pathology during this infection (Troeger et al., 2007).

The excretion and/or secretion of virulence factors by the parasite has been speculated to be responsible in part for the pathology induced during human giardiasis. Jiménez et al. (2004) demonstrated that oral administration of excretory/secretory antigens from G. intestinalis elicited a robust Th2 response (i.e. IgG1, IgG2, IgE) as well as moderate to profound histological changes. They tentatively attributed these changes to be potentially due to the parasite's proteinase activity. Ringqvist et al. (2008) identified three major secreted proteins (ornithine carbamoyl transferase, arginine deiminase, and enolase) released by Giardia. The release of these proteins was trigged upon contact with host epithelial cells, and these secreted proteins were able to impair host innate immune mechanisms, including the production of nitric oxide.

Another group has reported the presence of an excretory/secretory molecule in the parasite that can induce fluid accumulation in ligated loops of rabbit ileum (Shant et al., 2002). A 58 kDa enterotoxin purified from the excretory-secretory product of the parasite was able to decrease host GTPase activity (Shant et al., 2004). In this study, activation of protein kinase A (PKA) as well an increase in intracellular calcium ions were also observed after enterotoxin treatment of mouse enterocytes. Subsequently, these authors demonstrated that purified 58 kDa enterotoxin induced increased levels of phospholipase C gamma1, inositol triphosphate, and an upregulation in the protein kinase C (PKC; Shant et al., 2005). An N-terminal peptide sequence for the putative toxin has been reported. However, this peptide sequence is not represented in the virtual proteome of the prototype assemblage A strain WB (Morrison et al., 2007). It is possible that this peptide is encoded in a portion of the genome which has not been assembled, or that the WB strain lacks this molecule. Further molecular characterization of this toxin would provide a significant advance in developing tools to combat this disease.

The protective roles of intraepithelial lymphocytes (IEL) against parasitic infections have been characterized in a number of protozoan and helminth infections (Guk et al., 2003; Moretto et al., 2004; Little et al., 2005). In giardiasis, however, it was recently proposed that increased numbers of IELs in patients with chronic disease in fact were associated with a malabsorption syndrome including the impairment of the glucose/sodium uptake (Troeger et al., 2007).

Reduction in levels of some intestinal disaccharidases, sucrase, lactase, maltase, and trehalase, have been shown to occur in human giardiasis (Levinson and Nastro, 1978) and in experimental infections in mice (Gillon et al., 1982), rats (Cevallos et al., 1995), and gerbils (Buret et al., 1991). In one study of G. muris infections in mice carried out by Daniels and Belosevic (1995), disaccharidase deficiency during primary infection was strongly associated with parasite number. In another study with the same model, the decrease in jejunal disaccharidase activities correlated with a diffuse shortening of brush border microvilli (Buret et al., 1990). Scott et al. (2000) showed that brush border damage and diassacharidase deficiency did not develop in nude mice infected with G. muris. Adoptive transfer of CD8⁺, but not CD4⁺ T cells, from infected wild-type mice into naïve recipients induced both symptoms, indicating that immune responses are key to the brush border defects commonly observed following infection. It still remains unclear how Giardia infection induces CD8⁺ T cell responses and how these cells induce the damage. Interestingly, since CD8⁺ T cells are not required to control the infection (Singer and Nash, 2000a), it is theoretically possible to induce protective immunity without inducing nutrient malabsorption.

The final mechanism contributing to pathology is intestinal smooth muscle hypercontactility. These muscular contractions contribute to the severe cramps associated with symptomatic disease, but also increase the rate of transit of gut contents through the intestines reducing the time available for nutrient uptake and water reabsorption. Recent work in our lab and others has demonstrated that increased transit rates occur in wild-type mice following G. duodenalis and G. muris infections, but not in SCID mice even though these mice have higher parasite loads (Li et al., 2006; Andersen et al., 2006). Increased transit requires coordinated contractions and relaxation of smooth muscle. Interestingly, nitric oxide (NO) production from the neuronal isoform of nitric oxide synthase (NOS1) is required for enhanced transit. NO is normally an inhibitor of muscle contractions in the enteric system. In contrast, we have shown that enhanced muscular contractions occur following infection due to release of mast cell granule contents (Li et al., 2007). Activation of muscular contractions could be triggered in organ baths by addition of the hormone cholecystokinin (CCK), and this response was blocked by agents that interfered with mast cell function. Given that Giardia parasites consume bile and that a normal physiological function of CCK is to induce contraction of the gall bladder and release of bile into the small intestine, it appears as though the parasite has evolved to co-opt the host's normal physiology in order to provide an essential nutrient for its growth. Moreover, variation in the ability of different parasites to consume bile or in the host to produce CCK or recruit mast cells following infection could help explain the variation in pathology seen in giardiasis.

Several studies have shown differences in pathology among Giardia isolates. There is contradicting evidence showing that a specific genotype of G. duodenalis (i.e. assemblage A or B) is associated with symptomatic disease. Singh et al. (2009), for example, showed that most isolates examined from symptomatic patients in Nepal were assemblage B. In contrast, other studies identified assemblage A to be associated with symptomatic disease (Haque et al., 2005). Furthermore, Kohli et al. (2008) found no differences in the ability of assemblage A and B parasites to cause symptomatic infections in children in Brazil. One possibility is that the differences observed (if any) reflect variation within these genotypic classifications, especially the presence or absence of specific virulence factors. Alternately, given that host immune responses are, in part, responsible for pathological effects induced by G. duodenalis, it is possible that different isolates may elicit host immune responses in different ways. For example, Williamson et al. (2000) found distinct differences in the infection dynamics, histopathological responses and serum antibody responses in a neonatal mouse model of infection with two different strains of G. duodenalis (although the genotypes were not reported). Collection of immune response data in parallel with patient symptoms and parasite genotypes may provide answers to this difficult issue.

Summary

Immune responses to Giardia can be an effective mechanism to control this infection. Development of vaccines that reduce the severity of infection could be possible. Infections with Giardia commonly lead to strong immune responses characterized primarily by the production of anti-parasite IgA. Repeated infections appear to produce an immune status in humans where symptoms of disease are reduced, although parasites continue to colonize the host and produce cysts. Recent insights into the molecular epidemiology of this infection, are allowing better assessment of the role of parasite variation in the range of clinical presentations which are commonly observed during giardiasis. These advances are occurring simultaneously with progress in understanding the role of immune response in contributing to pathology as well as protection. Development of vaccines that reduce the severity of infection should therefore be possible, although vaccines which can generate sterile immunity may be more difficult to achieve. Recent insights into the mechanisms involved in initiating immune responses and how these responses both eliminate parasites and contribute to pathology have provided a much better framework for this work to be carried out. Numerous areas of research, however, require more attention. These include how results in animal models relate to human disease, how the different elements of the innate and adaptive immune responses are regulated, and which mechanisms are responsible for parasite prevention of secondary infections as opposed to elimination of the primary infection.