Trichomonas vaginalis Lipophosphoglycan Exploits Binding to Galectin-1 and -3 to Modulate Epithelial Immunity

Abstract

Trichomoniasis is the most common non-viral sexually transmitted infection caused by the vaginotropic extracellular protozoan parasite Trichomonas vaginalis. The infection is recurrent, with no lasting immunity, often asymptomatic, and linked to pregnancy complications and risk of viral infection. The molecular mechanisms of immune evasion by the parasite are poorly understood. We demonstrate that galectin-1 and -3 are expressed by the human cervical and vaginal epithelial cells and act as pathogen-recognition receptors for the ceramide phosphoinositol glycan core (CPI-GC) of the dominant surface protozoan lipophosphoglycan (LPG). We used an in vitro model with siRNA galectin knockdown epithelial clones, recombinant galectins, clinical Trichomonas isolates, and mutant protozoan derivatives to dissect the function of galectin-1 and -3 in the context of Trichomonas infection. Galectin-1 suppressed chemokines that facilitate recruitment of phagocytes, which can eliminate extracellular protozoa (IL-8) or bridge innate to adaptive immunity (MIP-3α and RANTES (regulated on activation normal T cell expressed and secreted)). Silencing galectin-1 increased and adding exogenous galectin-1 suppressed chemokine responses to Trichomonas or CPI-GC/LPG. In contrast, silencing galectin-3 reduced IL-8 response to LPG. Live Trichomonas depleted the extracellular levels of galectin-3. Clinical isolates and mutant Trichomonas CPI-GC that had reduced affinity to galectin-3 but maintained affinity to galectin-1 suppressed chemokine expression. Thus via CPI-GC binding, Trichomonas is capable of regulating galectin bioavailability and function to the benefit of its parasitic survival. These findings suggest novel approaches to control trichomoniasis and warrant further studies of galectin-binding diversity among clinical isolates as a possible source for symptom disparity in parasitic infections.

Keywords: CCL-20 (MIP-3α); CCL5 (RANTES); Interleukin 8 (IL-8); cytokine; galectin; human vagina; inflammation; interleukin; parasite; sexually transmitted infection.

FIGURE 1.

Expression of galectin-1 and -3 by human cervicovaginal epithelial cells. A–D, Texas Red immunofluorescent visualization of galectin-1 (A) and galectin-3 (B) in ectocervical (Ect) cells with isotype IgG control in C and matched bright field in D. E and F, Western blot analysis of galectin-1 (E) and galectin-3 (F) in epithelial cell lysates loaded at equal total protein/lane (25 μg for galectin-1 and 15 μg for galectin-3) as follows: lane 1 = recombinant galectin control; lane 2 = endocervical (End); lane 3 = ectocervical; and lane 4 = vaginal (Vk) cells. G–J, cell-associated (G and H) and soluble (I and J) protein levels of galectin-1 and -3 expressed by endocervical, ectocervical, vaginal cells and in vitro reconstructed three-dimensional ectocervical tissues (VEC), measured by immunoenzyme assays. Bars are mean and S.E. Picograms of galectin/ml of cell culture supernatant or nanograms of galectin/mg of total protein in cell lysates from at least three experiments with each cell line are shown.

FIGURE 2.

rh-galectin-1 and -3 binding to T. vaginalis (TV) and purified T. vaginalis LPG and CPI-GC. A and B, live T. vaginalis bound to galectin-1 and -3 in solution, demonstrated by Western blot (A) and ELISA (B). A, lane 1, T. vaginalis alone; lane 2, T. vaginalis bound to galectin-1; lane 3, T. vaginalis bound to galectin-1 after elution with 100 mm lactose, and lane 4, galectin-1 alone (75 ng). C and D, galectin-1 and -3 bound to T. vaginalis, LPG and CPI-GC fixed on solid phase, detected by ELISA in the presence of 100 mm lactose or no competing sugar. Bars are means and S.E.; p values are from two-way ANOVA and Bonferroni multiple comparison post-test. E, Western blot (WB) analysis demonstrating that all galectin-3 monoclonal antibodies used in the experiments shown in A–E bind the N-terminal (non-lectin) domain of galectin-3.

FIGURE 3.

Kinetics of CPI-GC galectin binding. A–F, kinetic curves of CPI-GC binding to recombinant human galectin-1 (A and C), galectin-3 (B and D), and the plant lectins ricin (E) and tomato lectin (F) evaluated by bio-interferometry. CPI-GCs were prepared from different primary clinical T. vaginalis isolates (UR-1, OC-6, OC-7, OC-8, and OC10), the wild type laboratory strain B7RC2 and its mutant derivatives M-412 and M-2E2.

FIGURE 4.

Regulation of soluble galectin-1 and -3 in the human cervical epithelial space by trichomonas. Cervical epithelial cells were infected with Trichomonas (wild type strain B7RC2) in the presence of escalating doses of rh-galectin-1 (A and C) or galectin-3 (B and D). The x axis on each plot shows concentration of rh-galectin-1 or galectin-3 added to the infection model. The y axis shows extracellular levels of galectins quantified in the cell culture supernatants by ELISA after 24 h of incubation with Trichomonas and the particular recombinant galectin. A and B show the levels of endogenous galectin-3 and -1 found in the medium after the treatment with exogenously added galectin-1 and -3, respectively. A, only endogenous galectin 3 was available for measure because no exogenous rh-galectin-3 was added. The endogenous galectin-3 in this case went up as a result of stimulating the cultures with escalating doses of recombinant galectin-1. Similarly, in B we show endogenous galectin-1 measured when no exogenous galectin-1 was added. In contrast, C and D show both the endogenous and the exogenous galectin-1 or -3 levels, respectively. p values are from ANOVA and Bonferroni multiple comparison test. **, p < 0.01; ***, p <0.001; two-way ANOVA, Bonferroni multiple comparisons test. Bars represent mean and S.E. from duplicate cultures in one of three experiments.

FIGURE 5.

Effect of siRNA knockdown on extracellular and cell-associated levels of galectin-1 and -3 and effect of Trichomonas on cell-free levels of both galectins. A, cell-associated levels per total cellular protein content in endocervical epithelial clones stably transfected with shRNA control (ctrl) or siRNA for galectin-1 knockdown (gal-1KD) or galectin-3 knockdown (gal-3KD); B, secreted extracellular levels of galectin-1 and -3 in the shRNA control, gal-1KD, and gal-3KD supernatants; C and D, cell-free levels of secreted galectins following infection for 24 h with T. vaginalis wild type (WT TV) laboratory strain B7RC2 and its mutant derivative m4.12. p values are from ANOVA and Bonferroni multiple comparisons test. Bars represent mean and S.E. from five experiments in A and three experiments in B–D. *, p < 0.05; **, p < 0.01; ***, p <0.001; two-way ANOVA, Bonferroni multiple comparisons test.

FIGURE 6.

LPG regulation of cell-associated galectin-1 and -3 in the human cervical epithelial space. Control (shRNA ctrl) and galectin knockdown epithelial cell clones were treated for 24 h with rh-galectin-1 (A) or galectin-3 (B) in the presence of T. vaginalis (wild type) lipophosphoglycan (LPG) or vehicle control (medium or PBS), and the levels of cell-associated galectins (both endogeneous and exogenous) were determined. p values are from ANOVA and Bonferroni multiple comparisons test. Bars represent means and S.E. from duplicate cultures in one of three experiments. *, p < 0.05; **, p < 0.01; ***, p <0.001; two-way ANOVA, Bonferroni multiple comparisons test.

FIGURE 7.

Immune regulation by endogenous and exogenous galectin-1 and galectin-3. Control (shRNA ctrl) and galectin knockdown endocervical epithelial cell clones were treated for 24 h with escalating doses of rh-galectin-1 (A, C, and E) or galectin-3 (B, D, and F) in the presence of T. vaginalis (wild strain B7RC2) lipophosphoglycan (LPG) or vehicle control (medium or PBS). The chemokine production was assessed in cell culture supernatants. Bars are means and S.E. from duplicate cultures representing three experiments. *, p < 0.05; **, p < 0.01; ***, p <0.001; two-way ANOVA, Bonferroni multiple comparisons test.

FIGURE 8.

Chemokine responses to T. vaginalis (TV) parasites, LPG, and CPI-GC from laboratory strains and clinical isolates that showed different affinity to galectin-3. IL-8 and MIP-3α were quantified by Meso Scale Discovery multiplex in 24-h supernatants from endocervical (End), ectocervical (Ect), or vaginal (Vk) epithelial cells after stimulation with live parasites (A and B), purified LPG (C and D), or CPI-GC (E–G) from wild type (wt) B7RC2, mutants (M-4.12 and M-2E2) and clinical isolates UR1 and OC6. Bars represent means and S.E. from duplicate or triplicate cultures in three experiments performed with each cell line. *, p < 0.05; **, p < 0.01; ***, p <0.001; two-way ANOVA, Bonferroni multiple comparisons test.

FIGURE 9.

Schematic summary of the different scenarios of T. vaginalis signaling to the human epithelial cells via LPG-galectin binding, emerging from our experimental results. Top panel, modest inflammation caused by the wild type protozoa that results from their good affinity to both galectin-3 and galectin-1, where galectin-3 mediates proinflammatory chemokine production, e.g. IL-8, and galectin-1 plays a dual role by suppressing chemokines and stimulating galectin-3 expression. 2nd panel, higher proinflammatory response caused by the wild type protozoa in the absence of endogenous galectin-1 (achieved in our model by galectin-1 knockdown) and preserved galectin-3 signaling. 3rd panel, suppressed inflammatory response to the wild type protozoa in the presence of endogenous galectin-1 and absence of endogenous galectin-3 (achieved by galectin-3 knockdown). 4th panel, suppressed inflammation in response to protozoan mutants that lack affinity to galectin-3 and have reduced but relatively preserved affinity to galectin-1.