Advances in understanding Giardia: determinants and mechanisms of chronic sequelae

Unscrambling strain virulence and pathogenicity

Of at least six recognized species of Giardia, only G. lamblia causes infections in humans. G. lamblia isolates are further divided into eight assemblages, designated A to H. Among the assemblages, only assemblages A and B and their respective subtypes cause human infection. The relative proportion of assemblage A to assemblage B infection varies both temporally and spatially, with a predilection for more assemblage B infections in endemic settings [87–90].

Whole-genome sequencing has revealed that assemblage A and assemblage B laboratory isolates are quite dissimilar and may be better categorized as separate species [91]. Despite this genetic divergence, attempts to ascribe clinical variability and pathogenicity based on assemblage designation have been inconclusive. In separate studies, either assemblage A [92,93] or assemblage B [94,95] more strongly associates with diarrhea. A limitation in assessing assemblage-specific pathogenicity in naturally acquired infection is that targets used to differentiate between assemblages A and B—β-giardin (bg), glutamate dehydrogenase (gdh), triose phosphate isomerase (tpi), and the small subunit 18S rRNA genes—are not known virulence factors. Furthermore, studies that incorporate multilocus genotyping demonstrate a substantial proportion of specimens with mixed results [96,97], raising the possibility of heterologous infections or even inter-assemblage recombination events [98,99]. Expanded comparative genomics studies investigating several assemblage A and assemblage B isolates may further clarify strain-specific pathogenicity.

In experimental conditions, Giardia strain is the strongest predictor of disease outcome. In initial volunteer studies, the axenized assemblage B isolate GS/M caused more symptoms than other isolates, a finding that was reproduced in murine models [30,100]. Laboratory assemblage A and assemblage B isolates have different growth characteristics in standard media conditions, and the former are more readily recoverable; however, assemblage B isolates are more infectious [84,85]. Historically, there has been significant experimental heterogeneity in laboratory studies [101], and this limits comparisons between investigators. When tested under identical conditions, however, there are apparent assemblage-dependent differences in induction or disruption of interleukin-8 (IL-8) from epithelial cells in vitro [102,103] and ex vivo biopsies from patients with inflammatory bowel disease [104] and immune-mediated disaccharidase deficiency in murine models [84]. Furthermore, in vitro models demonstrate that strain-dependent changes occur in epithelial cells [77] and that mixed assemblage co-cultivation, or cultivation with a clone expressing mixed assemblage traits, induces more epithelial cell apoptosis and tight-junction disruption than mono-infection with either assemblage A or B [105]. Finally, proximity to a natural source may also be relevant given that recently human-passaged strains were more infectious in an animal model than laboratory strains were [106] and that gerbil-passaged cysts achieve prolonged infections [85].

Advances in Giardia biology and gene regulation using transcriptomic and proteomic approaches aim to identify new virulence traits that could be confirmed in human infections. For example, as a possible means of evading host humoral and innate defenses, Giardia has a repertoire of 20 to 200 kDa cysteine-rich proteins that densely coat the surface of trophozoites, termed variant surface proteins (VSPs). Although only a single VSP is expressed on an individual trophozoite, the breadth of potential VSPs (73 to 270 or more) [91,107] in a given strain is associated with enhanced virulence [108]. How VSP switching occurs every 6-13 generations and what events trigger post-transcriptional VSP regulation [107,109] and trafficking may help unravel parasite determinants of transient colonization opposed to chronic infection. The role of Giardia cathepsins as disease determinants also warrants further consideration [102,103].

Nutritional determinants of disease

It is increasingly recognized that, within the luminal ecology of the human intestine, resident microbiota modify ingested nutrients and their metabolites, influencing nutrient availability and uptake. Giardia lacks fundamental enzymes necessary for generating critical biomolecules such as cholesterol and therefore must rely on acquisition of these materials from the luminal environment [110–112]. The parasite acquires nutrients primarily via bulk-phase uptake in endocytic vacuoles localized to the exposed dorsal (non-adherent) side [113], adheres to epithelial cells using the ventral disc [114,115], and may direct fluid flow toward adherent parasites via coordinated flagellar motion [116]. Arginine utilization represents an example of how shunting of key nutrients may provide an advantage to the parasite at a detriment to the host. Stadelmann and colleagues [117] demonstrated that Giardia isolates expressing arginine deiminase rapidly deplete arginine, resulting in epithelial cell cell-cycle arrest. Since both Giardia and epithelial cells use arginine for growth, an arginine-depleting mechanism could explain why, under conditions of arginine scarcity such as severe protein malnutrition, there is increased susceptibility to villus shortening during Giardia infection [85,118]. For the host, decreased availability of arginine also impairs inducible nitric oxide production by epithelial cells and may alter cellular immune responses [119,120], whereas Giardia uses the sequestered arginine primarily for energy production via the arginine dehydrolase pathway. Host genetically determined pathways that facilitate arginine acquisition or utilization despite Giardia infection may be important for maintaining host resiliency during infection. For example, cellular arginine uptake occurs through the cationic transporter (CAT), and CAT expression is increased in some cell types in the presence of APO-E 4/4 allelotype [121]. The APO-E 4/4 allelotype confers protection against diarrhea in children [122,123] and Cryptosporidium infection in mice [121], and may also associate with better cognitive outcomes in Giardia-infected children [124].

There is mounting recognition of an intricate relationship between host nutrition and mucosal immune responses. Malnourished children, for example, have diminished seroconversion to some components of the oral polio vaccine [125]. In experimental models, malnutrition confers enhanced susceptibility to Cryptosporidium [126–128] and enteroaggregative Escherichia coli (EAEC) [129]. Zinc deficiency enhances susceptibility to Giardia [130] and diminishes inflammatory responses to EAEC [131]. In a chronic giardiasis model, protein malnutrition led to decreased expression of IL-4 and IL-5 and diminished B-cell populations in intestinal tissues, villus blunting, and growth failure despite a parasite burden similar to that of well-nourished infected controls [85]. These findings suggest that, in children with chronic giardiasis, nutritional status could similarly determine disease severity.

Microbiota and co-pathogen interactions

There is increasing recognition of the role of the nearly 10¹⁴ resident bacteria in regulating host metabolism during states of both obesity and undernutrition [36,132] as well as immunity. In experimental giardiasis, the microbiota composition was shown to have more influence over susceptibility to infection than the absence of CD4⁺ T cells [133]. Thus, several experimental models rely on manipulating host microbiota through the use of antibiotic-containing water to promote infection [84,85,103]. Though sometimes necessary, the depletion of microbiota in these models may exclude important Giardia effects on microbiota that indirectly influence pathogenesis. Although attempts to identify small-bowel bacterial overgrowth in children with giardiasis have failed to show an association [134], microbiota from duodenal contents of Giardia-infected patients caused more inflammation than axenized Giardia trophozoites in germ-free mice [135]. Chen and colleagues [86] also observed that diverse populations of mucosally associated bacteria, including Lactobacillus, Streptococcus, and Staphylococcus species, are increased during experimental Giardia infection and after clearance of the parasite. Associated increases in the pro-inflammatory cytokines interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α), and myeloperoxidase were also seen [86]. The persistent translocation of mucosally associated bacterial microbiota has been proposed to explain post-infection irritable bowel syndrome and chronic fatigue [31], sequelae present in up to 5% of individuals with prior giardiasis [136], particularly in those whose initial course was more protracted [137]. Mechanisms by which Giardia could influence microbiota composition include alterations in the luminal nutritional environment as previously stated, as well as dampening the epithelial cell inflammatory responses through post-translational reductions in IL-8 and decreased nitric oxide synthesis [102,120,138–140].

Whether resident microbiota in the small intestine are protective against Giardia or promote chronic gastrointestinal disease may be conditional on their composition. When used as probiotics in experimental models, Lactobacillus species (L. johnsonii, L. casei, and L. rahmnosus GG) have been shown to promote Giardia clearance [141]. Concurrently, these Lactobacillus species enhance pro-inflammatory responses and improve histopathological parameters [142–144]. Furthermore, bacteriocins produced by L. acidophilus (P106) and L. plantarum (P164) reduced parasite adhesion [145,146]. The fermented milk product kefir, which contains a combination of lactic acid bacteria (lactobacillus, lactococci, and Leuconostoc), acetic acid bacteria, and yeast, protected mice against G. lamblia trophozoite infection and enhanced pro-inflammatory responses, including IFN-γ and TNF-α, with a decrease in CD4⁺ T-cell expansion [147]. Interestingly, activation of Toll-like receptor 2 (TLR2) signaling in murine dendritic cells, as occurs in response to some Lactobacillus species, in combination with TLR1 or TLR6, leads to increased IL-12/23p40, IL-23, and IL-10 secretion in the presence of Giardia lysate [148], whereas Giardia lysates decrease IL-12p70 and IL-23 but increase IL-10 responses to TLR4 agonists [148]. The down-regulatory effect of Giardia lysate on dendritic cell responses to TLR4 agonists has been shown to be dependent on phosphoinositide 3-kinase inhibition and may depend on Giardia arginine deiminase [78,149]. In contrast, Giardia lysates have no regulatory effect on bovine monocyte-derived dendritic cells [150].

The influence of Giardia-mediated changes in the luminal ecology may also alter susceptibility to other co-infecting enteropathogens, which are now recognized to be more common than mono-infections in children in resource-limited settings [34]. Furthermore, only a narrow margin separates the mean number of co-enteropathogens found in non-diarrheal specimens (4.3 ± 0.1) from the number in diarrheal specimens (5.6 ± 0.1) [34]. In some populations, up to 75% of Giardia infections include co-enteropathogens, most commonly with Vibrio cholerae and rotavirus [151] and elsewhere with norovirus and enteropathogenic E. coli [46]. Cases of co-infections with Tropheryma whipplei have also been reported [152]. In rural Ecuador, the OR for diarrhea with rotavirus infection was 6-14.8, but G. lamblia with rotavirus increased the OR to 11-24.13. However, Giardia co-infection with enterotoxigenic Escherichia coli, Campylobacter, Cryptosporidium, and enteroinvasive Escherichia coli did not demonstrate an increased risk of diarrhea, suggesting specificity to Giardia co-pathogen interactions [153]. On the other hand, associations between decreased risk of acute diarrhea in Giardia-infected cohorts raise questions that require additional investigation of whether Giardia decreases susceptibility to some pathogens or attenuates their severity. Cotton and colleagues [103], for example, demonstrated that Giardia infection diminished neutrophil chemotaxis and inflammation in a model of Clostridium difficile toxin-induced colitis.

Host responses and disease pathogenesis

The duality of host immune response in Giardia pathogenesis is complex and has been reviewed extensively [83,154]. Increasingly, redundant mechanisms of immunologic control are being identified, including chemokines (CCL2, CCL20, CXCL1, CXCL2, and CXCL3) from epithelial cells [155], IL-6 derived from dendritic cells and possibly mast cells [156–158], TNF-α [159], α-defensins activated by matrix metalloprotease 7 (Mmp7) [159], and nitric oxide generated from either nitric oxide synthase 1 (NOS1) or NOS2 [159–161]. Data from isolated outbreaks and experimental models demonstrate a mixed immune response consisting of both antibody production [162,163] and mucosal and systemic Th1-type CD4⁺ T-cell responses [84,164,165]. Studies in populations with serial outbreaks demonstrate that exposure to Giardia can promote protective immunity [166], which is characterized by a predominance of circulating T cells with a Th1-memory phenotype (CD25⁺CD26^bright cells that produce IFN-γ) [165]. Also, the absence of CD4⁺ T cells leads to more prolonged infection in experimental models [84].

Despite evidence for the role of T cells in protective immunity, clinical observations suggest that humoral responses are critical for immunity. These studies demonstrate greater prevalence of Giardia cysts and more symptomatology in patients with hypogammaglobulinemia (that is, x-linked agammaglobulinemia [Bruton's disease] and common variable immunodeficiency) [167–169] than those with T-cell deficits (advanced AIDS or thymic aplasia). Decreased secretory IgA variably associates with increased risk [170] and impaired IgA responses are seen in children with persistent Giardia infection [171]. In animal models, the majority of genes expressed in experimental murine infections are related to antibody production [159], and IgA-deficient knockout mice have difficulty clearing Giardia [172]. The polymeric Ig receptor, which transports IgA and IgM across the epithelium, is essential for eliminating G. muris from murine hosts, further supporting possible dependence on mucosal secretory antibody responses for eradication of infection [173].

In addition to facilitating parasite clearance, the secondary host immune response to Giardia may promote pathologic changes [84,174]. CD8⁺ T cells may contribute to villus [85] and microvillus [175] shortening in G. lamblia infection as demonstrated in models of murine giardiasis (G. muris) [80,81,84]. In addition, brush border actin cytoskeletal changes in ezrin and villin and disaccharidase deficiency after G. lamblia infection are dependent on T-cell responses [174]. Clinical findings also demonstrate that subsets of infected symptomatic adults [74] and some children show increased lymphocytic infiltrates as well as decreased villus-to-crypt ratios [72] on histological biopsy. Decreased villus surface area in patients with chronic giardiasis coincides with increased T-cell populations and epithelial cell apoptosis [74], and may similarly suggest immune-mediated diffuse microvillus shortening seen in some patients with giardiasis [82]. A murine model using freshly acquired gerbil-passaged G. lamblia cysts in combination with antibiotic-containing water to reduce resident microbiota similarly demonstrated persistent parasitism, with increased mucosal T cells and epithelial cell apoptosis through 9 weeks in immunocompetent hosts, suggesting that some Giardia strains that evade host responses can promote chronic mucosal inflammation [85].

Such chronic and repeated Giardia exposures are common in children in low-income settings who also have a high prevalence of malnutrition and an increasingly recognized gut dysfunction with chronic intestinal inflammation, increased gut permeability, and reduced villus length, termed environmental enteropathy [11]. Few studies have characterized immune responses to Giardia in these children. Kohli and colleagues [87] identified inflammatory diarrhea by using fecal lactoferrin during the first case (but not subsequent cases) of Giardia, and Long and colleagues [176] demonstrated increased IL-4, IL-5, monocyte chemoattractant protein-1 (MCP-1), and IFN-γ but decreased IL-8 in children with more prolonged episodes of Giardia. Thus, relationships between intestinal inflammatory response and Giardia infection may influence disease outcomes in these children.

Elucidation of immune-protective responses in contrast to those that might directly promote enteropathy in these children could also help guide vaccine development. Currently, the only licensed vaccine, GiardiaVax, is for use only in canines and has modest efficacy [177]. Attempts to identify correlates of protection in children in low-resource endemic settings are inconclusive. For example, data from Bangladesh demonstrated that, although human breast milk has direct anti-Giardia properties and may contain high concentrations of anti-Giardia antibodies [18], there is no correlation between the presence of Giardia-specific antibodies in breast milk and incidence, severity, or recurrence of Giardia in these children [178]. Potential approaches to vaccine development include targeting the repertoire of G. lamblia VSPs, since purified VSPs can protect against primary infection in experimental models [179]. Another vaccine candidate, α1-giardin expressed in an attenuated Salmonella vector, showed efficacy in mice, but α-enolase and ornithine carbamoyl transferase did not [180]. Challenging field vaccine application is the recognition that some mucosal vaccines are less effective in children in low-resource settings, with an as-of-yet unidentified but hypothesized immunologic hurdle in the setting of malnutrition [125], adding another layer of complexity to this preventive strategy.

Disclosures

The authors declare that they have no disclosures.