Portal de Periódicos da CAPES

The Molecular Pathogenesis of Giardiasis

1997; Lippincott Williams & Wilkins; Volume: 24; Issue: 1 Linguagem: Inglês

10.1097/00005176-199701000-00018

1536-4801

Michael J.G. Farthing,

Parasitic infections in humans and animals

Giardia intestinalis (formerly known as Giardia lamblia) is the most common human protozoan pathogen. It is perhaps best known for the clinical diversity of the infection that they produce (1-3). In young children and previously unexposed adults, particularly travelers from low- to high-prevalence parts of the world, Giardia can produce serious diarrhoeal illness with intestinal malabsorption, marked weight loss, and in infants and young children, impairment of growth and development. In the vast majority of individuals worldwide, the parasite is apparently carried without significant clinical morbidity. There is no satisfactory explanation for the diverse clinical spectrum seen in giardiasis, and as yet no virulence factors have been identified. The failure to identify the parasite determinants responsible for the production of intestinal injury, diarrhoea, and malabsorption makes it impossible to provide a simple explanation for the molecular pathogenesis of giardiasis (4). However, there have been some important steps forward in understanding the disease mechanisms, particularly the concept of genetic diversity, within the Giardia genus. The recent description of biological variation in association with different genotypes may be the first indication that virulent and relatively avirulent isolates of Giardia intestinalis exist. This observation alone could be of great assistance in the isolation of specific virulence factors in the future. For more than two decades there has been evidence that mucosal immune responses may be involved in the development of the enteropathy associated with giardiasis (5). Repopulation of the villus with relatively immature cells may also be a factor and, according to recent evidence, could occur as a direct response to the parasite without evoking immunological pathways or epithelial cell injury. Finally, Giardia appears to be able to influence a number of important physiological events within the intestinal lumen and to interfere with the digestive process as well as the absorptive function of the small intestinal epithelium (4). GENETIC DIVERSITY OF GIARDIA INTESTINALIS The members of the Giardia genus are characterised by the three pairs of flagella, two nuclei, and ventral disc that distinguishes them from other members of the hexamitidae family (6). There are three major morphological subtypes of Giardia: G. agilis from amphibians, G. muris from mice, and G. intestinalis from humans and some other vertebrates. These different types can be distinguished by the overall shape and dimensions of the trophozoite body and by the distinctive shapes of the median bodies. Two other Giardia isolates have been described: G. psittaci, isolated from a budgerigar, and G. ardeae, from the Great Blue Heron. These can be distinguished from the other subtypes by an absence of the ventrolateral flange in the former and a single (rather than a double) caudal flagellum in the latter. However, only G. intestinalis is known to infect humans and all of these isolates are morphologically identical. Other approaches were required therefore to identify phenotypic and genotypic diversity amongst different human G. intestinalis isolates. Phenotypic Variation Isoenzyme analysis and the classification of isolates according to zymodemes have been used for many years to distinguish different isolates of protozoa. This approach has been used with some success in the identification of virulent and relatively avirulent isolates of Entamoeba histolytica; however, not all investigators accept the absolute reliability of this approach to classification. Application of this technique to the classification of Giardia isolates has resulted in the description of a variety of groupings or subclassifications that demonstrated the geographic diversity of the organism; this technique also provided some of the early evidence indicating that G. intestinalis isolates from humans and animals are similar and in some cases apparently identical (7). The main limitation of this technique, and indeed of all other approaches to classifying Giardia isolates, relates to the lack of any correlation between isoenzyme analysis and the biological behaviour of a particular isolate with respect to virulence. Phenotypic differences have also clearly been demonstrated by surface antigen analysis using immunoblotting or immunoprecipitation techniques (8,9). Although differences have been observed, it has not been possible to develop a reproducible typing system based on immunodominant antigens. This limitation may be related to the organism's ability to display a variety of surface antigens that vary within the same isolates (10). Thus, antigenic variation may be one mechanism by which this organism evades the host immune response. Genotypic Variation Giardia has been estimated to have between 8 and 50 sets of chromosomes with a total genome size of about 1.2 × 107 base pairs (11-13). Giardia nuclei are haploid, and the observed genetic diversity between isolates is thought to be based on clonal divergence. The Giardia genome appears to exhibit considerable genomic plasticity, which extends from an allelic variation of single genes to polymorphism of entire chromosomes (13). The chromosomes that have been studied most extensively in Giardia are those that encode for rRNA genes. The rRNA genes are present as short arrays close to the telomeres. Extensive and frequent polymorphisms have been noted at the end of these chromosomes, whereas the central chromosome domains are conserved. DNA analysis by restriction fragment polymorphism analysis and DNA fingerprinting has demonstrated marked genetic diversity (14-17); there is, however, close similarity between G. intestinalis isolates obtained from humans and those from domestic and wild animals, supporting the view that giardiasis is a zoonosis. However, more than one DNA fingerprint profile can be obtained from cloned isolates from the same human infection, suggesting that mixed infections with genetically diverse Giardia isolates occur in humans. DNA fingerprint profiles, however, have been shown to change with time and with experimental in vitro pressures, confirming the plasticity of the Giardia genome. Thus, it is unclear whether giardiasis is truly a mixed infection or the process of in vitro culture exploits the plastic genome and artificially leads to clonal divergence (14,18). Recent work from our laboratory suggests that in vitro cultured G. intestinalis isolates with distinct DNA fingerprints have different biological effects in a neonatal rat model of giardiasis with respect to both structural and functional abnormalities in the small intestine (19). This observation may lead the way to the eventual isolation of virulence factors, as it has been possible to link a possible genetic marker with a possible biotype. INTESTINAL COLONISATION: MOLECULAR MECHANISMS Infection is initiated by ingestion of Giardia cysts, generally through contaminated water or food; direct person-to-person transfer does occur. Colonisation of the small intestine is an essential component of Giardia's life cycle and a sine qua non for the production of diarrhoeal disease. Colonisation occurs as a three-step process commencing with excystation, followed by attachment to the small intestine epithelium and multiplication. However, to complete the life cycle, the organism must again encyst, following which cysts are excreted in the faeces and the cycle of faecal-oral transmission can continue. Excystation This process has been achieved in vitro for G. intestinalis and G. muris by exposure of cysts to low pH (1.3-2.7 or 4.0) followed by transfer to neutral pH (20,21). These in vitro conditions closely resemble those in vivo when cysts move from the acid environment of the stomach to the near neutral pH of duodenal and jejunal fluid. However, excystation of G. intestinalis and G. muris can occur at neutral pH (21,22), and G. muris have been shown to exist following exposure to sodium bicarbonate at pH 7.5 (22), consistent with a role for duodenal and pancreatic bicarbonate secretion in the excystation process. Excystation begins within 5-10 min of exposure to a suitable excystation environment; flagella movement is the first sign of activity. The intracellular vacuoles are thought to discharge their contents during excystation (23), suggesting a role for hydrolase activity in the process. Attachment Giardia attaches intimately to the intestinal epithelium and to a variety of other inert substrates such as glass and plastic. The ventral disc is thought to be the primary organelle of attachment. Hydrodynamic forces generated beneath the disc by continuous activity of the ventral flagella have been proposed as a possible attachment mechanism (24). It has also been suggested that flagella activity causes low pressure under the ventral disc due to fluid fluxes around the ventral and marginal grooves. However, attachment occurs with the morphologically distinct G. psittaci, in which the ventrolateral flange is incomplete; thus it is unlikely that hydrodynamic forces can be the sole explanation for disc-mediated attachment (6). The presence of contractile proteins in the peripheral regions of the ventral disc suggests that they participate in attachment (25). Inhibitors of microfilament function, such as cytachalasin-B and low calcium concentrations, inhibit attachment-again supporting a role for cytoskeletal proteins in attachment (26,27). An alternative attachment mechanism has been proposed following the discovery of a mannosebinding surface lectin on Giardia trophozoites. Lectin activity was initially described using the classic approach of a mixed-cell agglutination with mammalian erythrocytes (28). The lectin is distributed throughout the parasite surface and is not specifically localised to the ventral disc (29). The lectin is present within the trophozoite as a prolectin that can be activated by trypsin (30). Giardia lectin (also known as taglin) has been purified and found to have a molecular weight of 28-30 kDa (31). Giardia trophozoites adhere to isolated mammalian intestinal epithelial cells in vitro by mechanisms that are consistent with lectin-mediated attachments (32). Giardia also attaches to Caco-2 cells (the human colon cancer cell line that differentiates to resemble small intestinal epithelial cells) by mechanisms that incorporate both contractile proteins and surface lectin-mediated processes (33). The relative importance of mechanical and lectin-mediated attachment mechanisms has not been established. The surface lectin may operate in the initial attachment between Giardia and the epithelium, since, in contrast to ventral disc-mediated attachment, the parasite does not need to be in a specific orientation to the mucosa. Subsequently, the parasite may reorientate in a ventral surface-down position where mechanical forces may be more effective. Multiplication Giardia trophozoites divide by binary fission. Although it has been suggested that "sexual reproduction" may occur, there is no firm evidence of exchange of genetic material between organisms. Giardia is a luminal parasite and is presumed to be dependent on luminal nutrients. Bile promotes growth of Giardia in vivo and in vitro (34-37). Other related protozoa and intestinal bacteria do not exhibit a similar degree of dependence (38). It has been proposed that bile salts promote the uptake of preformed biliary phospholipid, which the organism is unable to synthesise de novo (39). Carbohydrate and a relatively low oxygen tension are also required for growth (40). In vitro culture requires additional nutrients, including iron, ascorbic acid, and cysteine (37). Long-term in vitro cultivation generally requires the presence of mammalian serum. Encystation Encystation is essential for the parasite to complete its life cycle and survive in the external environment or be carried by other hosts such as domestic and wild animals. Encystation has been achieved in vitro by exposing trophozoites to high concentrations of conjugated bile salts and myristic acid at approximately neutral pH (41-43). Encystation is associated with the expression of specific encystation antigens (21-39 kDa) during the first 6-18 h. Higher molecular weight antigens (66-103 kDa) appear later. These antigens are transported to the cyst wall and are probably involved in wall structure. SMALL INTESTINAL PATHOLOGY AND PATHOPHYSIOLOGY Giardia is usually found in close proximity to the apical surface of the enterocyte, often penetrating deep into the intestinal crypts. Giardia is not an invasive organism, although there have been reports of trophozoites being found within the mucosa (2); overall, this should be regarded as a rare event. One of the difficulties in explaining the pathology and pathophysiology of giardiasis relates to the extensive spectrum of disease expression, which ranges from symptom-free carriage of the organism to chronic diarrhoea and malabsorption. This clinical variation may relate to host factors, which may be genetically or environmentally determined, or to the severity of the infection (as judged by parasite load within the small intestine). As discussed previously, there is preliminary evidence to suggest that Giardia isolates may vary in virulence, which could be a major determinant of disease expression. Pathology A variety of structural abnormalities of the small intestinal mucosa have been reported in humans and in animal models of giardiasis (44-53). In humans, giardiasis is associated with the complete repertoire of abnormalities of villous architecture, ranging from entirely normal light microscopic appearances to partial or subtotal villous atrophy. The majority of individuals have either normal or relatively mild villous shortening, usually associated with an increase in crypt depth. However, studies in human small intestine are relatively limited and inevitably biased by cases that are investigated in a hospital setting, which almost certainly represent the more severe end of the giardiasis clinical spectrum. In human giardiasis, there may be no light microscopic abnormalities in the proximal small intestine; in hospital-based series, probably 20-25% of patients fall into this category (48,49). Less than 10% of adults with giardiasis will have subtotal villous atrophy. There is often an increase in chronic inflammatory cells in the lamina propria and in the epithelium in human giardiasis. Experimental infections in gerbils, mice, and rats can produce similar abnormalities of mucosal architecture, although the villous and crypt abnormalities are usually relatively mild (54). The gerbil provides a particularly good model with which to study small intestinal structure and function, as weanling gerbils develop diarrhoea and have significant morphological abnormalities by day 6 of infection, with reduction in villous height in the duodenum and increase in crypt depth in the duodenum, jejunum, and ileum (55,56). In the ileum, there is a small but significant increase in villous height. These early changes in villous and crypt morphology occur in the absence of any inflammatory infiltrate in the lamina propria and with no increase in the numbers of intraepithelial lymphocytes. We have recently reported similar findings in a neonatal rat model of giardiasis (19). Two of three strains tested produced reduction in villous height but no significant increase in crypt death. In addition, there was no decrease in epithelial cell height nor an increase in the number of intraepithelial or lamina propria lymphocytes. In human giardiasis, even when villous architecture appears normal by light microscopy, ultrastructural changes such as shortening and disruption of microvilli are present (45,57,58). The gerbil model has confirmed the significance of these observations, demonstrating a marked reduction in microvillus membrane surface area in both jejunum and ileum-although these abnormalities were transient (56). This decrease in the height of microvilli was uniformly found but was not specifically related to sites of trophozoite attachment. There is evidence from human studies to suggest that the extent of the mucosal abnormality relates to the severity of the diarrhoea (48,49). Disaccharidase Activity In human giardiasis, morphological abnormalities have been associated with a reduction in lactase, sucrase, and maltase activities in the microvillus membrane (49,59). Similar observations have been reported in experimental infections in mice, gerbils, and rats (19,55,60-62). Reduction in disaccharidase activities is maximal when diarrhoea and villous morphological abnormalities are most profound. In the neonatal rat model of giardiasis, precocious expression of sucrase was demonstrated at 10 days in the three genotypically distinct isolates that were tested (19). Although precocious sucrase expression in the small intestine has not previously been reported in a parasitic infection, it is well-described in other situations where there is mucosal injury or other systemic disturbances inducing a stress response. Lymphoid Nodular Hyperplasia Lymphoid nodular hyperplasia has been associated with both chronic giardiasis and immune deficiency (63-66). Several studies have examined the prevalence of giardiasis in patients with hypogammaglobulinaemia and found it to occur in 29-71% of cases. However, in one study from India, 25 patients were described with giardiasis and lymphoid nodular hyperplasia but no immunoglobulin deficiency (64). Thus, the relationship among lymphoid nodular hyperplasia, hypogammaglobulinaemia, and giardiasis remains unclear although it would appear that any two can occur in combination without any direct indication of pathogenesis. There are also no clear indications as to the pathogenesis of lymphoid nodular hyperplasia, although several studies have shown a predominance of IgM-producing B cells within the mucosa and lymphoid nodules, suggesting there might be immune "overreactivity" to a luminal antigen, possibly with failure to switch from IgM to IgA production within the intestine. Intestinal Transport Abnormalities Several studies in animal models suggest that there are functional consequences of the structural abnormalities and the reduction in disaccharidase activity. In gerbils, basal transport of sodium and chloride ions was not different in noninfected controls, but glucose-stimulated sodium absorption was significantly reduced in the jejunum but not in the ileum of infected gerbils in experiments using stripped small intestinal mucosa mounted in Ussing chambers (56). Perfusion studies in vivo in animals have also shown impaired water, sodium, and chloride absorption in response to glucose, although basal transport was similar to that in controls (19,56). In a neonatal rat model of infection, basal transport of water, sodium, and chloride ions was impaired with some animals actually in a net secretory state for sodium and chloride ions (19). Perfusion of a lactose-containing solution enhanced these transport abnormalities. Studies with brush border membrane vesicles from infected mice provided further evidence that there is impairment of glucose and amino acid transport (67,68). Although phenotypic and genotypic differences have been evident among different Giardia isolates for some years, the absence of any clearly defined virulence factors has made it difficult to ascertain whether any variation in phenotype or genotype has a biologically relevant correlate in the form of diverse virulence biotypes. Previous experimental infection studies in humans, however, have shown that different isolates can produce variable immune responses in the host and that these differences can result in differing susceptibilities to rechallenge and differing durations of infection before natural clearance. More recently, we studied the biological consequences of experimental infection in 8-day-old neonatal rats with three distinct G. intestinalis genotypes as judged by DNA fingerprinting (19). The isolates differed in their ability to impair fluid and electrolyte absorption and in the proabsorptive response to perfusing a lactose-electrolyte solution. Isolates differed in virulence potential with respect to reduced lactase activity and impaired glucose absorption. These differences in pathophysiology could not be attributed to quantitative differences in parasite load or duration of infection, since these variables were identical for all three isolates. In addition, there were no major differences in the local intestinal immune response to Giardia antigens, as judged by sIgA production, indicating that these apparent differences in virulence could not be attributed to differences in immune-mediated clearance (69). These findings in neonatal rats cannot be extrapolated directly to human infection although they suggest that Giardia isolates vary in virulence, possibly accounting for at least some of the clinical diversity seen in this infection. MOLECULAR MECHANISMS OF DIARRHOEA AND MALABSORPTION The mechanisms by which Giardia causes diarrhoea and intestinal malabsorption remain controversial. Early ideas on pathogenesis suggested that Giardia trophozoites act as a mechanical barrier to absorption or compete for host nutrients (4), but the large functional reserve of the small intestine and the relatively small metabolic mass of the parasite make these hypotheses untenable. There is evidence, however, that Giardia can produce variable degrees of mucosal injury and at the same time influence conditions in the intestinal lumen that could impair nutrient digestion and absorption. Enteropathy Ultrastructural Damage Giardia trophozoites attach to the epithelium and have been shown by electromicroscopy to disrupt and distort mcrovilli at the site where the ventral disc interfaces with the microvillus membrane (70). It is possible therefore that physical factors involved in this cell-cell interaction might account for microvillus damage at sites of adherence. There is some evidence to suggest that Giardia itself produces, and possibly release, cytopathic substances into the intestinal lumen (68). As yet no parasite product has been identified to account for the morphological damage observed in the small intestine. Giardia does, however, contain a number of thiol proteinases which might attack surface glycoproteins and disrupt microvillus membrane integrity (71-73). In addition, the surface mannose-binding lectin of Giardia may contribute to epithelial injury (29-32). Dietary plant lectins can directly damage intestinal epithelial cells and produce microvillus membrane abnormalities very similar to those seen in giardiasis (74,75). It remains to be established whether any of these parasite products are injurious to the host enterocytes. Villus-Crypt Architecture Abnormalities Whatever the mechanism by which Giardia damages villus epithelial cells and presumably produces increased epithelial cell loss, there would appear to be a predictable crypt cell response with an increase in crypt depth and crypt cell proliferation. In other conditions in which this response occurs, such as coeliac disease, there is repopulation of the villus by relatively immature enterocytes with reduced absorptive capacity. Increased intestinal proliferation has been confirmed in the gerbil model of giardiasis (56), but using thymidine kinase activity as a marker of maturity, there is no evidence in the jejunum or ileum that the cells repopulating the villous are less mature than those in controls. From these data, it seems likely that the structural and functional abnormalities observed in the microvillus membrane relate to direct injury rather than to a secondary mechanism increasing crypt cell production. Work in our laboratory has shown, however, that Giardia can directly stimulate the enzyme ornithine decarboxylase (ODC), a rate-limiting enzyme in polyamine synthesis and cell proliferation, in Caco-2 cells in vitro (76) and in neonatal rat small intestine in vivo (77). Stimulation of ODC requires cell-cell contact, presumably during the attachment process. Dead trophozoites or subcellular fractions of Giardia have no effect. Whether this mechanism could have a role in producing a hyperproliferative state in the small intestinal crypt requires further investigation. Failure to discover a simple explanation for epithelial damage has promoted speculation that immunological mechanisms may be involved akin to those thought to be responsible for the villus atrophy of coeliac disease. In human giardiasis there is a variable immune response within the mucosa, although infection is often associated with an increase in the number of lamina propria and intraepithelial lymphocytes (49,52,78,79). These inflammatory changes have been more difficult to reproduce in experimental animals (19,80). However, there is compelling evidence that T cell activation alone can produce villus atrophy (81). Enteropathy occurring in intestinal graft-versushost disease and rejection of transplanted intestinal allografts is characterised by villous atrophy, crypt cell hyperplasia, and a lymphocytic infiltrate. Using human foetal small intestinal explants, it has been possible partly to characterise the mechanisms involved by activating T cells with pokeweed mitrogen or anti-CD3 antibody; both approaches produced villus atrophy, crypt hyperplasia, and increased interleukin-2 production, confirming T cell activation (81). Further support for this hypothesis has been obtained from studies in experimental G. muris infection in athymic nu/nu mice (80). Despite prolonged infection, the alteration of the villus-crypt cell ratio is less severe in nu/nu mice than in immunocompetent controls. When lymphocytes from the spleens of immunologically intact mice were injected into athymic infected mice, histological abnormalities in the small intestine became more pronounced. However, reduction in the villus-crypt cell ratio did not occur in the immunocompromised mice before reconstitution, and it seems likely that T cell mechanisms are also involved (80). In addition, immunosuppression in mice results in more profound effects on disaccharidase activities in conventional animals, indicating that epithelial damage is not solely dependent on immune function (82). Furthermore, although intraepithelial lymphocytes are frequently increased in number in giardiasis and have been implicated in immune-mediated damage in experimental G. muris infection, intraepithelial lymphocyte numbers increased after villus shortening when the decrease in brush-border disaccharidase had already occurred (60). Further evidence to support immunopathogenic mechanisms was, however, provided in studies in which G. muris-infected mice were given trophozoite extract after clearing the primary infection (83). This approach resulted in a rapid decrease in mucosal disaccharidase activity, which was thought to be immune-mediated. The antigen challenge effect was most marked in the C3H/HeN mouse, which is genetically highly susceptible to G. muris infection. Whether Giardia lectin can act as a mitogen and directly activate T cells remains to be established. However, the mannose-binding lectin concanavalin A is a potent mitogen, which might support a similar role for the Giardia lectin. Evidence for Variable Parasite Virulence In addition to the strain-dependent variation in the impairment of fluid, electrolyte, and monosaccharide absorption in neonatal rat small intestine, we have shown similar effects on small intestinal morphology and disaccharidase activity (19). Villus height in animals infected with the Giardia strains PO1 and WB was reduced by 11 and 22%, respectively, although the third isolate VNB3 failed to produce a significant reduction in villus height. Despite a lack of morphological damage, VNB3 had the most profound effect on fluid, electrolyte, and glucose absorption. PO1 and WB produced significant reductions in lactase activity, and all three isolates promoted precocious expression of sucrase activity, in common with other stressful or injurious activities in the small intestine. Again, evidence of strain variation was apparent, with PO1 producing significantly more sucrase activity than either VNB3 or WB. Thus, it would appear that there are parasite determinants that may be variably expressed by different isolates and that appear to influence the degree of morphological and functional disturbance in the small intestine. Maldigestion In addition to the morphological abnormalities in the small intestine that may contribute to the pathophysiology of diarrhoea and malabsorption, a variety of luminal factors have been identified that would operate primarily on digestive processes but that would ultimately contribute to nutrient, fluid, and electrolyte absorption. Bacterial Overgrowth There is some evidence to suggest that symptomatic giardiasis is associated with increased numbers of aerobic and/or anaerobic bacteria in the proximal small intestine. In one study from India, 8 of 17 patients (47%) with steatorrhoea had more than 104 aerobic bacteria cultured from duodenal fluid, whereas no patient from a giardiasis control group without steatorrhoea had bacterial overgrowth (84). Three of the patients with steatorrhoea also had anaerobes present. Tomkins et al. found increased numbers of aerobic bacteria in 9 of 14 symptomatic overland travellers (64%) with giardiasis returning to the U.K.; two also had anaerobes present (85). Bacterial overgrowth can produce architectural abnormalities in the small intestine similar to those seen in giardiasis (86), and it may thus have a role in producing mucosal injury. However, in both of these studies one might expect to find increased numbers of bacteria in the proximal small intestine regardless of the presence of giardiasis; thus, the specificity of these observations requires confirmation. There have been no studies in patients with giardiasis in the industrialised world to examine whether associated bacterial overgrowth occurs in this group of patients and whether it is an relevant mechanism of diarrhoea in giaridiasis. Bile Salt Deconjugation The removal of the glycine or taurine conjugates from bile salts reduces their solubility in aqueous solution and thus reduces their efficacy in micelle formation within the intestinal lumen. In addition, free bile salts are membranotoxic and can cause intestinal secretion, thus potentially contributing to the pathogenesis of diarrhoea (86). Tandon et al. found evidence of bile salt deconjugation in all of their Indian subjects with bacterial overgrowth and in 40% of giardiasis controls without malabsorption (84), but other studies have not confirmed these findings (38). Giardia itself does not have the ability to deconjugate bile salts (38,87). Bile Salt Uptake by Giardia Bile and bile salts appear to make important contributions to the life cycle of Giardia. The presence of bile at low concentrations has been shown to stimulate parasite growth and reduce generation time and thus may be an important colonisation factor for this parasite (36,37,39). The effect can be reproduced partly by the addition of conjugated bile salts alone, which appear to increase the uptake of cholesterol and membrane phospholipid that the parasite is unable to synthesise de novo (39). More recently it has been shown that parasites grown in bile are larger than those grown in a bile-free medium and that the presence of bile alters antigenic expression by the parasite (88). High concentrations of bile salts trigger parasite encystation. During the course of bile stimulation experiments it became apparent to us that the parasite was consuming conjugated bile salts (39). Further studies showed that this was relatively specific for Giardia and that uptake appeared to be occurring by a carrier-mediated active-transport process (89). Bile salt transport occurred against a concentration gradient and was dependent on intact energy metabolic pathways in the parasite. Bile salts were internalised and found to be in the cytosol fraction following differential centrifugation. The metabolic advantage of bile salt uptake for the parasite has not been defined. Theoretically, consumption of host bile salts during chronic diarrhoea could deplete the bile salt pool and might contribute to malabsorption by impairing micellar solublisation of ingested fat and decreasing the effectiveness of pancreatic lipase, the action of which is bile salt-dependent. Inhibition of Host Digestive Enzymes Intraluminal concentrations of trypsin, chymotrypsin, and lipase have been shown to be reduced in symptomatic patients with giardiasis (90-92). There is no evidence that this reduction is due to a failure of pancreatic exocrine secretion, but it could be related to the more recent observation that live Giardia trophozoites and trophozoite sonicates inhibit trypsin activity and lypolysis in vitro (93,94). The mechanisms by which the parasite inhibits these enzyme activities has not been established but could be related to the direct interaction between a parasite product, such as its own proteinases, and the host enzyme. However, the pancreas has a large functional reserve and the magnitude of the reduction observed in clinical studies is itself unlikely to account for malabsorption. However, the pancreas could contribute to the cascade of abnormalities that together impair nutrient absorption and contribute to diarrhoea and malabsorption. A MODEL FOR PATHOGENESIS From the foregoing discussion it is clear that no single virulence factor or unifying mechanism explains the pathogenesis of giardiasis. Work to date would certainly suggest that infection can be associated with small bowel enteropathy with an associated reduction in disaccharidase activities and impairment of fluid, electrolyte, and monosaccharide absorption. However, epithelial dysfunction and diarrhoea in humans can occur in the absence of overt abnormalities of villus-crypt architecture, as exemplified by the observation that ≈25% of symptomatic individuals with giardiasis presenting to hospital clinics have no small bowel abnormality by light microscopy (49,78) and the finding in our own laboratory that one Giardia isolate (VNB3) did not change villous height but had a profound effect on small intestinal absorptive function (19). Although the T cell activation hypothesis is attractive for explaining the enteropathy, there is no evidence as yet that this is important in human infection and the evidence in animal models is conflicting. However, preliminary factors that might impair digestive processes could well be part of a malabsorptive consortium, but it seems unlikely that they alone can explain the diarrhoea and malabsorption associated with this infection. Thus, we are left with the somewhat unsatisfactory situation of having to accept, in our present state of ignorance, a multifactorial and possibly a multistep process of pathogenesis. One approach to the development of a model for pathogenesis (Fig. 1) is to consider the three common clinical presentations of the infection, namely (a) asymptomatic carrier state; (b) acute self-limiting diarrhoea; and (c) chronic diarrhoea with intestinal malabsorption. For the asymptomatic carrier state, one must presume that either the host is infected with a relatively avirulent strain that lacks key virulence factors or that host immune and possibly nonimmune defence mechanisms are operating sufficiently to control parasite numbers and render infection subclinical. It is unknown whether the carrier state inevitably begins with acute diarrhoeal illness or whether the entire period of carriage remains clinically silent. The prepatent period for acute giardiasis is about 7-10 days. This would be consistent with a relatively long period for colonisation and establishment of infection and the relatively slow process of parasite multiplication. This time span is of course quite different from that of acute bacterial diarrhoea such as that due to enterotoxigenic Escherichia coli (ETEC), which generally begins within 1-2 days of ingesting the organism. There is a much shorter generation time for ETEC and also a rapid release of protein secretatory enterotoxins. For Giardia the time to symptoms would be entirely consistent with the failure to detect the enterotoxin activity in the organism. The prepatent period would certainly be compatible with the involvement of immunopathogenetic mechanisms. In chronic infection with malabsorption, the multifactorial process involving a cascade of mucosal and luminal events is probably involved. Failure to irradicate the organism might lead to progressive disruption of epithelial structure and function, and there is some evidence to suggest that the degree of villous shortening and crypt hyperplasia is quite closely related to the severity of the intestinal malabsorption. Intestinal fat malabsorption activates the ileocolonic "brake," which results in retardation of intestinal transit; in time this would favour bacterial over-growth. In uncontrolled chronic infection, one might anticipate that the parasite load would gradually increase, so that trophozoite numbers would be sufficient to make a significant impact on the activity of luminal hydrolytic enzymes and possibly to make biologically relevant reductions in intraluminal bile salt concentrations. Thus, there would be a destructive malabsorptive cycle that would not be broken until treatment was initiated with an antigiardial agent.FIG. 1: . A model of pathogenesis.