Cervicovaginal bacteria are a major modulator of host inflammatory responses in the female genital tract

Abstract

Colonization by Lactobacillus in the female genital tract is thought to be critical for maintaining genital health. However, little is known about how genital microbiota influence host immune function and modulate disease susceptibility. We studied a cohort of asymptomatic young South African women and found that the majority of participants had genital communities with low Lactobacillus abundance and high ecological diversity. High-diversity communities strongly correlated with genital pro-inflammatory cytokine concentrations in both cross-sectional and longitudinal analyses. Transcriptional profiling suggested that genital antigen-presenting cells sense gram-negative bacterial products in situ via Toll-like receptor 4 signaling, contributing to genital inflammation through activation of the NF-κB signaling pathway and recruitment of lymphocytes by chemokine production. Our study proposes a mechanism by which cervicovaginal microbiota impact genital inflammation and thereby might affect a woman's reproductive health, including her risk of acquiring HIV.

Figure 1

16S rRNA sequencing analysis of cervical swabs reveals low Lactobacillus abundance and four distinct bacterial community structures. (A) Heatmap of bacterial taxa identified by 16S V4 sequencing of cervical swabs collected from 94 women. Cervicotypes (CTs) were determined based on the dominant species: Lactobacillus crispatus (CT1), L. iners (CT2), Gardnerella (CT3), and mixed microflora containing Prevotella (CT4). Nugent scores and bacterial alpha diversity are also shown. (B) Principal coordinates analysis using the weighted UniFrac distance metric on the same 94 samples, colored by CT. See also Figure S1 and Table S1.

Figure 2

16S V4 sequencing and whole-genome shotgun sequencing (WGS) identify very similar bacterial abundances for all CTs, with higher resolution taxonomic identification provided by WGS. Participant #6 represents CT1 (A), #1 represents CT2 (B), #7 represents CT3 (C), #11 represents CT4 (D). See also Figure S2.

Figure 3

Asymptomatic women display a broad range of baseline genital inflammation that is not explained by STIs. (A) Heatmap of cervicovaginal lavage cytokines from 146 women, each represented by a column. Nugent scores and active STIs (blue: C. trachomatis, N. gonorrhoea, T. vaginalis, and/or HSV-2 positive; gray: negative for that STI) are also displayed. Principal component (PC) analysis was performed on the normalized cytokine concentrations and the first PC (PC1) explained 41% of variation. (B) Pie chart of the STI prevalence in women in the highest quartile of inflammation (n=35). See also Figure S3 and Table S2.

Figure 4

Genital pro-inflammatory cytokine levels vary significantly with microbial community structure and most strongly correlate with CT4 bacterial communities. (A) Cytokine PC1 values from women with bacterial communities CT1–4. (B) Odds ratios and 95% confidence intervals representing the likelihood that a woman with cervicotype 4, an active chlamydia infection, bacterial vaginosis (Nugent score >7), an active Trichomonas infection, or cervicotype 3 is within the top quartile of proinflammatory cytokine levels (as determined by cytokine PC1) versus below the 75^th percentile. (C–H) Cervicovaginal lavage IL-1α, IL-1β, TNF-α, IFN-γ, IL-10, and IL-8 concentrations in women with bacterial communities CT1–4. Median and IQR shown. Significance levels were determined by a Kruskal-Wallis test and asterisks denote post-test significance level (* p < 0.05; ** p < 0.01; *** p < 0.001; n=94). See also Figure S4.

Figure 5

Intraindividual longitudinal genital microbiome changes correlate with pro-inflammatory cytokine levels. (A) Cervicovaginal IL-1α, IL-1β, and TNF-α concentrations with bacterial taxa identified by 16S sequencing from matched longitudinal cervical swabs and CVLs collected from two representative participants, #13 and #14. (B–C) Cervicovaginal IL-1α, IL-1β, and TNF-α concentrations from serial time points with either no change in CT (B) or a change (C) (one-tailed Wilcoxon matched pairs test). The order of time points was standardized such that the lower cervicotype was entered as the first time point and the higher cervicotype was entered as the second time point, regardless of the actual chronological order. Any transition, e.g. from CT1 to CT2 and from CT3 to CT4, was considered to be a CT change. See also Figure S5 and Table S3.

Figure 6

Bacterial species highly correlated with pro-inflammatory cytokines in vivo also stimulate vaginal epithelial cells to produce the same cytokines in vitro. (A) Ridge regression analysis with cytokine PC1 as the response variable and bacterial abundances as the predictor variables. The bacterial prevalence amongst the 94 subjects is also shown; taxa must have an abundance of at least 1% within a sample to be considered present. (B–E) IL-1α, IL-1β, TNF-α, and IL-8 secretion by vaginal epithelial cells after in vitro application of 15 log₁₀ CFU/cm² of each bacterial species. MALP-2 (a TLR 2/4 agonist; 25nM) treatment was used as a positive control (value denoted by dashed line). P values were determined by a Dunnett test, compared to Lactobacillus crispatus. Data shown as mean and SEM, with two biological replicates. See also Figure S6 and Table S4.

Figure 7

Transcriptional profiling suggests that CT4 bacterial products are sensed by cervical antigen presenting cells (APCs), which contribute to genital inflammation by producing a myriad of pro-inflammatory cytokines and critical T cell chemoattractants. (A, B) CD14 (A) and CD11c (B) staining of endocervical tissue. Scale bar: 50 µm. (C) Heatmap of genes differentially expressed in cervical APCs from women with Lactobacillus dominant communities (>95% abundance) (from left to right: 2 in CT1, 3 in CT2) versus those with high Prevotella abundance (>25%) (n=5 in CT4). Women with STIs were excluded. Only genes with a false discovery rate < 0.05 are shown. Gene set enrichment analysis was used to determine statistically significant similarities with annotated gene sets. (D) Number of live endocervical CD38+ HLADR+ CCR5+ CD4 T cells from women in the highest and lowest quintile of cytokine PC1 (Mann-Whitney, n=16 per group). See also Figure S7 and Table S5.