Conserved metabolic enzymes as vaccine antigens for giardiasis

Abstract

Giardia lamblia is a leading protozoal cause of diarrheal disease worldwide. Infection is associated with abdominal pain, malabsorption and weight loss, and protracted post-infectious syndromes. A human vaccine is not available against G. lamblia. Prior studies with human and murine immune sera have identified several parasite antigens, including surface proteins and metabolic enzymes with intracellular functions. While surface proteins have demonstrated vaccine potential, they can exhibit significant variation between G. lamblia strains. By comparison, metabolic enzymes show greater conservation but their vaccine potential has not been established. To determine whether such proteins can serve as vaccine candidates, we focused on two enzymes, α-enolase (ENO) and ornithine carbamoyl transferase (OCT), which are involved in glycolysis and arginine metabolism, respectively. We show in a cohort of patients with confirmed giardiasis that both enzymes are immunogenic. Intranasal immunization with either enzyme antigen in mice induced strong systemic IgG1 and IgG2b responses and modest mucosal IgA responses, and a marked 100- to 1,000-fold reduction in peak trophozoite load upon oral G. lamblia challenge. ENO immunization also reduced the extent and duration of cyst excretion. Examination of 44 cytokines showed only minimal intestinal changes in immunized mice, although a modest increase of CCL22 was observed in ENO-immunized mice. Spectral flow cytometry revealed increased numbers and activation state of CD4 T cells in the small intestine and an increase in α4β7-expressing CD4 T cells in mesenteric lymph nodes of ENO-immunized mice. Consistent with a key role of CD4 T cells, immunization of CD4-deficient and Rag-2 deficient mice failed to induce protection, whereas mice lacking IgA were fully protected by immunization, indicating that immunity was CD4 T cell-dependent but IgA-independent. These results demonstrate that conserved metabolic enzymes can be effective vaccine antigens for protection against G. lamblia infection, thereby expanding the repertoire of candidate antigens beyond primary surface proteins.

Fig 1. Production and purification of recombinant ENO and OCT.

Recombinant forms of α-enolase (ENO) and ornithine carbamoyl transferase (OCT) were produced in E. coli after induction with isopropyl-β-d-thiogalactopyranoside (IPTG), purified by nickel affinity chromatography, and visualized by SDS-PAGE and Coomassie Blue staining. Arrow heads depict the recombinant proteins.

Fig 2. Immunogenicity of ENO and OCT in humans.

A,B. Sera from G. lamblia-infected patients (circled dots) and age-matched, presumed-unexposed controls (black dots) were tested for levels of IgG (A) and IgA (B) against the indicated antigens (Ag) by multiplex bead immunoassays. Each data point represents one individual, red bars show the geometric means; *P<0.05 vs controls as determined by Mann-Whitney U test. C. Correlation of ENO-specific IgG and IgA levels. The red dashed line represents the linear correlation curve. The black oval encloses all control values. D. Correlation of levels of IgG specific for ENO and OCT. The red lines outline the different groups with low or high antibody levels against the two antigens, with the lower left quadrant including all controls.

Fig 3. Immunogenicity and protective capacity of ENO and OCT in mice.

A-C. C57BL/6 mice were immunized by three intranasal administrations two weeks apart with ENO or OCT, and cholera toxin (CTX) as an adjuvant. CTX toxin alone was used as a control. B. Mice were bled, and plasma IgG levels against the respective antigens (Ag) were determined by ELISA (means ± SE, n = 4 mice/group; open symbols, immunized mice; closed symbols, CTX-treated controls). C. Two weeks after the last immunization, mice were orally inoculated with G. lamblia GS/M trophozoites, and live trophozoites were enumerated in the small intestine (SI) after 5 days. Data from three separate experiments (i.e., separate inoculations on different days) were combined in the graph. Each dot represents an individual animal. The red lines represent geometric means, the dashed black line the detection limit of the assay. *P<0.05 vs CTX control as determined by Kruskal-Wallis test with Dunn’s post hoc test. D. Mice immunized with ENO (n = 5 mice) or OCT (n = 5 mice) and CTX-treated controls (n = 5 mice) were orally infected with G. lamblia GS/M trophozoites and fecal cyst output was determined at the indicated intervals. Data are shown as means and SE. *P<0.05 vs CTX control at the same time as determined by t-test.

Fig 4. Effect of antigen combinations on protection against G. lamblia.

A,B. C57BL/6 mice were immunized by alternating intranasal administrations of two antigens (a and b) on a weekly basis for a total of three doses for each, using cholera toxin as the adjuvant (A). Cholera toxin (CTX) alone was used as a control. One week after the last immunization, the C57BL/6 mice were orally inoculated with G. lamblia GS/M trophozoites, and live trophozoites were enumerated in the small intestine after 5 days. The graph in panel B shows the decrease in log₁₀ of the trophozoite numbers in the small intestine relative to those of CTX controls for the indicated antigen combinations (ENO, α-enolase; OCT, ornithine carbamoyl transferase; α1G, α1-giardin; UPL, uridine phosphorylase-like protein-1). Each dot represents an individual animal. Data are compiled from five separate experiments, with a mean and SE of the log₁₀ trophozoite load in the CTX controls of 5.78 ± 0.71 (range 4.83–6.41) on day 5. Red lines represent geometric means, the dashed black line shows the mean of the controls. All six immunized groups, which had received either one of three antigens (ENO, UPL, α1G) or two antigens (ENO+OCT, ENO+UPL, ENO+ α1G), were significantly protected (*P<0.05 by Kruskal-Wallis test with Dunn’s post hoc test), while none of these groups were significantly different from any other immunized group. Significances were omitted from the graph for clarity.

Fig 5. Cross-protection against divergent G. lamblia strains.

BALB/c mice were immunized three times with ENO, using CTX as the adjuvant (A). CTX alone and naïve untreated mice were used as controls. Two weeks after the last immunization, the mice were orally inoculated with either G. lamblia WB or GS/M trophozoites, and live trophozoites were enumerated in the small intestine (SI) after 5 days (B). Each dot represents an individual animal. The red lines represent geometric means, the dashed black line the detection limit of the assay. Significance was evaluated by ANOVA, followed by a post-hoc Tukey test. *P<0.05 vs CTX controls, ^#P<0.05 vs naive controls for the respective strain. CTX-treated and naïve mice were not significantly different for either G. lamblia strain.

Fig 6. Role of IgA in ENO- and OCT-induced immune protection.

A,B. C57BL/6 mice were immunized with ENO or OCT, and cholera toxin (CTX) as adjuvant, as detailed in Fig 3A. CTX alone was used as a control. Plasma and mucosal washes from the small intestine were obtained and tested by ELISA for IgA and IgM antibodies against the immunizing antigens. Data are means ± SE (n = 4 mice/group). C. IgA-deficient mice (KO) were immunized with ENO (n = 8 mice) or OCT (n = 10 mice), or treated with CTX alone as a control (n = 8 mice). Two weeks after the last immunization, mice were orally inoculated with G. lamblia GS/M trophozoites, and live trophozoites were enumerated in the small intestine (SI) after 5 days. Each dot represents an individual animal. Data are compiled from two separate experiments (inoculations). The red lines represent geometric means, the dashed black line the detection limit of the assay. *P<0.05 vs. CTX control as determined by Kruskal-Wallis test with Dunn’s post hoc test.

Fig 7. IgG isotypes after ENO and OCT immunization.

C57BL/6 mice were immunized with ENO or OCT, and cholera toxin (CTX) as adjuvant, as detailed in Fig 3A. CTX alone was used as a control. Plasma and mucosal washes from the small intestine were obtained and tested by ELISA for the indicated IgG isotypes against the immunizing antigens. Data are means ± SE (n  = 4 mice/group).

Fig 8. Intestinal cytokine expression after G. lamblia infection of ENO-immunized mice.

C57BL/6 mice were immunized with ENO and cholera toxin (CTX) as adjuvant, as detailed in Fig 3. Treatment with CTX alone and naïve mice were as controls. ENO-immunized and CTX-only mice were orally infected with G. lamblia GS/M, while naïve mice were left uninfected. After 5 days, extracts were prepared from the mid-small intestine and levels of the indicated cytokines were assayed by multiplex immunoassay. Data are mean ± SE (n = 6 mice/group). Significances were determined by t-test.

Fig 9. Spectral flow cytometric analysis of intestinal immune cell populations.

C57BL/6 mice were immunized with ENO and CTX (green symbols), as detailed in Fig 3. Treatment with CTX alone (red symbols) and untreated naive mice (open black symbols) were used as controls. ENO-immunized and CTX-only mice were orally infected with G. lamblia GS/M, while naïve mice were left uninfected. After 5 days, single-cell suspensions were prepared from the small intestine. Cells were stained with a viability dye and a cocktail of antibodies against the indicated markers and analyzed by spectral flow cytometry. Dot plots depict representative examples for the gating strategies, while the graphs show the results for individual animals, with means indicated by horizontal bars. Data are compiled from two separate experiments. Significances were calculated by t-test (*p<0.05 for infected CTX controls vs naïve mice; #p<0.05 for infected ENO-immunized mice vs infected CTX controls). The graph in the gray-shaded area shows trophozoite counts obtained from the same cohorts.

Fig 10. Spectral flow cytometric analysis of immune cell populations in the mesenteric lymph nodes.

C57BL/6 mice were immunized with ENO and CTX (green symbols), as detailed in Fig 3. Treatment with CTX alone (red symbols) and untreated naive mice (open black symbols) were used as controls. ENO-immunized and CTX-only mice were orally infected with G. lamblia GS/M, while naïve mice were left uninfected. After 5 days, single-cell suspensions were prepared from the mesenteric lymph nodes. Cells were stained with a viability dye and a cocktail of antibodies against the indicated markers and analyzed by spectral flow cytometry. Dot plots depict representative examples for the gating strategies, while the graphs provide the results for individual animals, with means indicated by horizontal bars. Significances were calculated by t-test (*p<0.05 for infected CXTX controls vs naïve mice; #p<0.05 for infected ENO-immunized mice vs infected CTX controls).

Fig 11. CD4 T cell dependence of vaccine-induced protection against G. lamblia.

CD4-deficient (A, KO) and C57BL/6 mice (A, WT) mice, and Rag-2-deficient (B, KO) mice, were immunized intranasally with the indicated antigens and cholera toxin (CTX) as an adjuvant. CTX alone was used as a control. Two weeks after the last immunization, mice were orally inoculated with G. lamblia GS/M trophozoites, and live trophozoites were enumerated in the small intestine after 5 days. The graphs show the decrease in log₁₀ of the trophozoite numbers in the small intestine relative to those of controls for individual mice. Data are compiled from two separate experiments, with mean day 5 log₁₀ trophozoite loads in the CTX controls of 6.87 and 7.31 (A, WT mice), 6.53 and 5.06 (A, CD4 KO), and 7.20 and 6.99 (B, Rag-2 KO). The red lines represent geometric means and the dotted lines represent the trophozoite load in controls. *P<0.05 vs. CTX controls as determined by Kruskal-Wallis test with Dunn’s post hoc test (A); N.S., not significant.