Vaginal microbiome-host interactions modeled in a human vagina-on-a-chip

Abstract

Background: A dominance of non-iners Lactobacillus species in the vaginal microbiome is optimal and strongly associated with gynecological and obstetric health, while the presence of diverse obligate or facultative anaerobic bacteria and a paucity in Lactobacillus species, similar to communities found in bacterial vaginosis (BV), is considered non-optimal and associated with adverse health outcomes. Various therapeutic strategies are being explored to modulate the composition of the vaginal microbiome; however, there is no human model that faithfully reproduces the vaginal epithelial microenvironment for preclinical validation of potential therapeutics or testing hypotheses about vaginal epithelium-microbiome interactions.

Results: Here, we describe an organ-on-a-chip (organ chip) microfluidic culture model of the human vaginal mucosa (vagina chip) that is lined by hormone-sensitive, primary vaginal epithelium interfaced with underlying stromal fibroblasts, which sustains a low physiological oxygen concentration in the epithelial lumen. We show that the Vagina Chip can be used to assess colonization by optimal L. crispatus consortia as well as non-optimal Gardnerella vaginalis-containing consortia, and to measure associated host innate immune responses. Co-culture and growth of the L. crispatus consortia on-chip was accompanied by maintenance of epithelial cell viability, accumulation of D- and L-lactic acid, maintenance of a physiologically relevant low pH, and down regulation of proinflammatory cytokines. In contrast, co-culture of G. vaginalis-containing consortia in the vagina chip resulted in epithelial cell injury, a rise in pH, and upregulation of proinflammatory cytokines.

Conclusion: This study demonstrates the potential of applying human organ chip technology to create a preclinical model of the human vaginal mucosa that can be used to better understand interactions between the vaginal microbiome and host tissues, as well as to evaluate the safety and efficacy of live biotherapeutics products. Video Abstract.

Fig. 1

Characterization of a microfluidic human vagina-on-a-chip model. A A diagram of a two-channel organ chip viewed from above (left) and a higher magnification micrograph through a cross section of a vagina chip showing a stratified human vaginal epithelial cells cultured in the top channel atop a 50-μm thick porous membrane (dashed lines indicate top and bottom surface) with human uterine fibroblasts cultured on the opposed side of the membrane in the lower channel (right). B Representative cross-sectional immunofluorescence micrographs of human vagina chips showing squamous stratified vaginal epithelium immunostained for CK14, CK15, CK5, CK13, Involucrin, ZO-1, E-cadherin, DSG-3, DSG-1, and F-actin (to show all cells). Dashed line, upper boundary of the porous membrane; yellow, different markers; magenta, DAPI-stained nuclei. C A graph showing the changes in apparent permeability (P _app) of the vaginal tissue barrier measured on-chip for two human donors 05328 (Hispanic) and 04033 (Caucasian) measured by quantifying Cascade Blue transport. Data are presented as mean ± s.d.; n = 4. D RT-qPCR results showing relative mRNA expression of ESR1, PGR, PCK1, GCGR, KRT15, CLDN17, and ZO-1 in the vagina chip before (D0) and after differentiation on day 10 in the presence or absence of 0.4 nM (blue) or 4 nM (green) β-estradiol (D10)

Fig. 2

Culture of L. crispatus consortia in the vagina chip. A Total CFU/chip determined by quantifying non-adherent bacteria in effluents from the apical epithelial channel combined with counts of adherent bacteria measured within epithelial digests at 72 h compared to the original inoculum (T0). B Percent of OC1, OC2, OC3, and C0006A1 consortia bacteria adherent to the epithelium in vagina chips calculated by quantifying the bacteria recovered from the chip (digest) 72 h post-inoculation compared to the T0 inoculum. Significance was calculated by one-way ANOVA; *, P < 0.05 vs. C0006A1, ^#, P < 0.05 vs. OC1. C The pH values measured in the medium within the apical channel of the vagina chip cultured in the absence or presence of the OC1, OC2, OC3, and C0006A1 consortia at 72 h post-inoculation compared to pH measured in the basal channel. D Percent viability of vaginal epithelial cells assessed by calculating the number of live cells relative to control using trypan blue exclusion assay. In A–D, each data point indicates one chip; different colored points indicate chips from different donors; data are presented as mean ± s.d. In the lower graphs, results of metagenomics-based strain ratio analyses of the adherent bacteria from the L. crispatus OC1 (E), OC2 (F), and OC3 (G) consortia present within epithelial digests relative to the original inoculum are shown after 72 h of direct culture with the vaginal epithelium on-chip. 1–4 indicate results from 4 different chips

Fig. 3

D-lactate production in vagina chip. A D-lactate and L-lactate levels measured in the effluent of the epithelial channel of vagina chip cultured in the absence (control) or presence single strain (C0006A1) or multi-strain (OC1, OC2, OC3) L. crispatus consortia at 72 h post-inoculation. Each data point indicates one chip; data are presented as median with 95% CI. Significance was calculated by one-way ANOVA; *, P < 0.05 vs. control, ^#, P < 0.05 vs. OC1, ^¥, P < 0.05 vs. C0006A1. B Table showing median D-lactate concentrations (mM) measured in the vagina of women with Lactobacillus-dominated communities compared to concentrations measured in vagina chips cultured with different L. crispatus consortia (OC1, OC2, OC3) and the C0006A1 strain

Fig. 4

Suppression of the innate immune response by L. crispatus containing consortia in the vagina chip. The levels of cytokines (IL-6, IL-8, IL-1α, IL-1β, and IP-10) measured in effluents of vagina chips cultured with OC1, OC2, OC3, and C0006A1 consortia are show relative to control chips without bacteria. Each data point indicates one chip; different colored points indicate chips from different donors. Data are presented as mean ± sem; significance was calculated by one-way ANOVA; ***, P < 0.0001; **, P < 0.001

Fig. 5

Culture of non-optimal G. vaginalis containing consortia in the vagina chip. A Total CFU/Chip of BVC1 and BVC2 consortia bacteria measured in effluents from the vaginal epithelium-lined channel at 24, 48, and 72 h and epithelial tissue digest at 72 h relative to the starting inoculum (T0). Total, total CFUs measured in effluents + digest; Digest, CFU in the tissue digest. B pH measured in medium in the epithelial channel of vagina chips cultured in the presence or absence of G. vaginalis containing and BVC1 and BVC2 consortia for 72 h. C Percent viability of vaginal epithelial cells cultured on-chip in the presence of BVC1 or BVC2 consortia assessed by calculating the number of live cells relative to control using a trypan blue exclusion assay. D The levels of cytokines (IL-6, IL-8, IL-1α, IL-1β, and IP-10) measured in effluents of vagina chips cultured with BVC1 and BVC2 consortia are shown relative to control chips. Each data point indicates one chip; data are presented as mean ± sem; significance was calculated by one-way ANOVA; *, P < 0.01,**, P < 0.001