Persistent, Asymptomatic Colonization with Candida is Associated with Elevated Frequencies of Highly Activated Cervical Th17-Like Cells and Related Cytokines in the Reproductive Tract of South African Adolescents

Abstract

Cervicovaginal inflammation, nonoptimal microbiota, T-cell activation, and hormonal contraceptives may increase HIV risk, yet associations between these factors and subclinical Candida colonization or hyphae are unknown. We collected cervicovaginal samples from 94 South African adolescents, aged 15 to 19 years, who were randomized to injectable norethisterone enanthate (Net-En), an etonorgesterol/ethinyl estradiol vaginal ring (NuvaRing), or oral contraceptives in the UChoose trial (NCT02404038) at baseline and 16 weeks post-randomization. We assessed cervicovaginal samples for subclinical Candida colonization (by quantitative PCR [qPCR]), hyphae (by Gram stain), microbiota composition (by 16S rRNA gene sequencing), cytokine concentrations (by Luminex), and cervical T-cell phenotypes and activation (by multiparameter flow cytometry). While hormonal contraceptive type did not influence incidence of Candida colonization or hyphae, hyphae presence was associated with significantly elevated concentrations of IL-22, IL-17A and IL-17F, all produced by Th17 cells, but not of other cytokines, such as IL-1β or IL-6, after adjustment for confounders. Subclinical Candida colonization was associated with reduced frequencies of Th17-like cells and elevated frequencies of CCR6-CCR10 T cells. Women with Candida hyphae were less likely to have bacterial vaginosis (BV). Persistent, subclinical colonization with Candida over 16 weeks was associated with significant increases in Th17-related cytokine concentrations and highly activated Th17-like and CCR6-CCR10 T-cell frequencies. These data suggest that vaginal Candida colonization and hyphae increase Th17-related cytokines, but not overall female genital tract inflammation in Sub-Saharan African adolescents. Persistent Candida colonization, even when asymptomatic, may increase Th17 cell frequencies and related cytokines and thereby could subsequently increase HIV risk, although the causal relationship requires confirmation. IMPORTANCE Sub-Saharan African female adolescents are globally at the highest risk of HIV acquisition, and genital inflammation, microbial dysbiosis, and cervical HIV target cell activation are thought to contribute to this risk. Previously, the relationship between these mucosal factors and subclinical vaginal Candida colonization or hyphae has not been described, and the role of HIV-susceptible Th17 cells in mediating anti-Candida immunity in the human female genital tract has not been clearly established. We show that presence of yeast hyphae was associated with increases in Th17 cell-related cytokines and the absence of microbial dysbiosis, and that persistent Candida colonization resulted in significant increases in Th17-related cytokines and highly activated Th17-like cell frequencies. Our results suggest that Th17 cells are important for anti-Candida immunity in the human female genital tract and that prolonged vaginal Candida colonization may contribute to increased HIV risk in Sub-Saharan African adolescents by increasing HIV target cell frequencies and activation.

Keywords: HIV target cells; Sub-Saharan Africa; genital inflammation; mucosal immunity; vaginal candidiasis.

Conflict of interest statement

The authors declare no conflict of interest.

The authors declare no conflict of interests.

Figures

FIG 1

Candida colonization cervicovaginal cytokines at baseline. (A) Heatmap of log10 cytokine concentrations at baseline annotated by cytokine/inflammatory group, bacterial vaginosis (BV) status by Nugent Scoring, presence of bacterial sexually transmitted infections (STIs), yeast hyphae (hyphae) by Gram stain, and Candida colonization by qPCR. (B) Multivariate linear regressions showing the association between genital cytokine concentrations in women with low or high Candida colonization compared to those without any colonization, and in women with hyphae compared to those without. Each association is shown as a β-coefficient and the error bars represent the 95% confidence interval (CI). Associations shown in purple are statistically significant. P values of ≤0.05 were considered statistically significant.

FIG 2

Candida colonization and cervical immune cells at baseline. (A) Cervical Th17 cells (CCR6+ CCR10–) and Th1/Th2-enriched cells (CCR6– CCR10–) were identified by expression of CCR6 and CCR10 on CD4+ T cells. The frequencies of Th17-like and Th1/Th2-enriched cells and expression of HLA-DR and CCR5 on these cells in women with Candida colonization (purple) and in those without (green) was assessed at the baseline, and Wilcoxon rank-sum tests were applied to compare the frequencies between the groups. P values of  ≤0.05 were considered significant. (B) Multivariate linear regressions showing association between T-cell phenotypes in women with low or high Candida colonization compared to that in those without any colonization and in women with hyphae compared to in those without. Each association is shown as a β-coefficient and error bars represent the 95% CI. Associations shown in purple were statistically significant prior to adjusting for multiple comparisons. P values of ≤0.05 were considered statistically significant.

FIG 3

Yeast hyphae and vaginal microbiota at baseline. (A) The Shannon index, a measure of within-sample species diversity, was compared in women with yeast hyphae and in those without at baseline using a Wilcoxon rank-sum test. Prevalence of community state type (CST)-I (Lactobacillus crispatus-dominated), CST-III (L. iners-dominated) and CST-IV (dominated by diverse, bacterial vaginosis [BV]-associated bacteria) in women with hyphae and in those without. (B) Fold-change differences in the abundances of specific vaginal bacterial taxa in women with hyphae compared to in those without. Dots to the left of 0 represent a fold-change decrease. Red dotted lines represent a 0.5-fold change difference in abundance. Only differential abundances where P ≤ 0.01 after adjusting for multiple comparisons were included.

FIG 4

Distinguishing women with yeast hyphae from those without using cytokine, bacteria, and Th17 cell signatures. Multiomics feature selection model including all cytokines, bacterial taxa merged at the genus level, and Th17-like cell phenotypes which explain the highest variance when comparing women with yeast hyphae to those without. (A) Loading plot showing the contributions of the different selected biomarkers. Color indicates the group in which the variable has the maximum level of expression, using the median. (B) Circos plot showing the expression of the selected biomarkers for women without (outer green line) and those with hyphae (outer orange line), and correlations among the selected biomarkers (inner red and blue lines). Only associations above the threshold correlation cutoff, 0.3, were included. (C) Multiomics feature expression for each participant comparing the integrated cytokine, microbial, and Th17 phenotype profiles of women with hyphae (orange) and of those without (green). Cytokine (purple), bacteria (pink), and Th17-like cell (blue) clusters are shown on the top of the heatmap, and only selected features were included. Color key shows the range of correlation values.

FIG 5

Hormonal contraceptive use and Candida colonization. (A) Prevalence of high, low, or no Candida colonization or hyphae at the baseline and at the 16-week follow-up in women randomized to combined oral contraceptives (COC), injectable Net-En, and NuvaRing. (B) Changes in Candida status of participants who were colonized with Candida (Pos) or Candida-free (Neg) from the baseline to the 16-week follow-up. Each line presents one participant and is color-coded by contraceptive randomization group.

FIG 6

Persistence, clearance, and acquisition of Candida colonization from baseline to 16 weeks, and cervicovaginal cytokines. Paired concentrations of cytokines produced by Th17 cells (A), those important for Th17 cell differentiation (B), and those for Th17 cell regulation (C) from baseline to 16 weeks in women who remained persistently colonized with Candida (Pos-Pos), those who cleared Candida (Pos-Neg), those who remained free of Candida (Neg-Neg), and those who acquired Candida (Neg-Pos). P values prior to multiple-comparison adjustment are displayed. Those remaining statistically significant after multiple-comparison adjustment are indicated by a box. (D) Multivariate linear regressions showing the association between cervicovaginal cytokines and Candida clearance (compared to women remaining colonized) or acquisition (compared to women remaining non-colonized). Each association is shown as a β-coefficient and error bars represent the 95% CI. Associations shown purple were statistically significant prior to adjusting for multiple comparisons. P values of ≤0.05 were considered statistically significant.

FIG 7

Persistence, clearance, and acquisition of Candida colonization from baseline to 16 weeks and cervical immune cell phenotype. Paired frequencies of Th17-like (A) and CCR6– CCR10– CD4+ T cells (B) from baseline to 16 weeks in women who remained persistently colonized with Candida (Pos-Pos), those who cleared Candida (Pos-Neg), those who remained free of Candida (Neg-Neg), and those who acquired Candida (Neg-Pos). P values prior to multiple comparison adjustment are shown. Those remaining statistically significant after multiple-comparison adjustment are indicated by a box. (C) Multivariate linear regressions showing the association between cervical T-cell phenotypes and Candida clearance (compared to women remaining colonized) or acquisition (compared to women remaining non-colonized). Each association is shown as a β-coefficient and error bars represent the 95% CI. P values of ≤0.05 were considered statistically significant.