Incorporation of the vaginal microbiome into a microphysiological model of the endocervix to study the effects of bacterial vaginosis

Introduction: The cervicovaginal microbiome plays a vital role in women’s health, particularly in protecting the host from contracting and transmitting sexually transmitted infections (STIs). Yet, the mechanism by which microbiota perform this critical function remains poorly understood. Vaginal microbiota dominated by Lactobacillus spp. supports a lower pH ( < 4) through the production of lactic acid, which is thought to be responsible for decreased risk of infection. Indeed, vaginal microbiomes without many Lactobacillus spp. are typically considered non-optimal and have been associated with increased risk of STIs. However, direct mechanistic investigations into the role of the microbiome are challenging as animal models do not recapitulate human physiology and in vitro models lack the needed complexity and often do not integrate the microbiota. As such, we have developed a microfluidic human cervical model that can support the culture of endocervical cells with either an optimal or non-optimal constructed patient-informed microbiota. This model can be infected with STI pathogens to better understand the protective or permissive role of the microbiome and enables mechanistic investigation into the role of the microbiome in cervicovaginal health.

Materials and

Methods: We fabricated a two-component microfluidic system for ease of use: (1) an “insert” which supported cell culture and (2) a “cassette” to establish fluid flow through two insert channels. A2EN endocervical or primary cervical (HCxEC) or vaginal (HVEC) epithelial cells were cultured on the insert for up to 7 days to establish a monolayer and endogenous mucus production before the insert was sealed into a cassette to establish fluid flow. Cells were kept at the air-liquid interface, and media was circulated basally at a flow rate of 1-60 µl/min. Single microbial isolates or patient-informed microbiota were constructed and co-cultured with the endocervical cells. To ensure a functional tissue when incorporating a microbiome into the system, media formulations were optimized and bacterial growth and epithelial inflammation in the device were compared to a standard transwell insert model using qPCR and ELISA. The epithelial cells in the device were inoculated with Chlamydia trachomatis and the role of different types of microbial communities was quantified by measuring infection efficiency.

Results, Conclusions, and Discussions: Microfluidic models were successfully fabricated with a silicone insert and an acrylic cassette (Figure 1A). Epithelial monolayer formation for all cell types and mucus production for cervical cells was confirmed via immunofluorescent staining of claudin, phalloidin, and MUC5B after 6 days of culture. To successfully co-culture microbiota and epithelial cells, media optimization was performed to ensure both cell types could grow in the same culture and did not induce an epithelial inflammatory response. Patient-informed microbiota populations were constructed and a population representing community state types (CST) I and IV were used in the insert. CST I represents a Lactobacillus crispatus rich microbiome and is considered optimal while CST IV is constructed with several strict and facultative anaerobes such as Gardnerella vaginalis and Prevotella bivia to mimic a non-optimal microbiome. MIcrobiota were inoculated onto the apical surface of A2EN cells within the device. HBSS+glucose media was found to be the most effective at maintaining bacterial growth (Figure 1B). The CST IV microbiota increased the concentration of IL-8 in primary cell cultures while the CST I microbiota did not, indicating that the bacteria themselves have an inflammatory effect on the primary cells (Figure 1C). After 24 hours of microbiota culture in the model, epithelial cells were exposed to Chlamydia trachomatis. We found that microbiota of CST I were protective against infection, while microbiota of CST IV were permissive to C. trachomatis infection. D-lactate, produced by Lactobacillus crispatus and Lactobacillus jensenii but not Lactobacillus iners, was shown to be critical to the protection (Figure 1D), indicating a direct mechanism through which the microbiome plays a role in cervicovaginal health and STI pathogenesis. Together, these data validate this model as a tool that can be used to study microbiome-host-pathogen interactions and bacterial vaginosis.
We developed a simple microphysiological model of the human endocervix that supports patient-informed vaginal microbial communities. Our in vitro model will enable mechanistic investigation of the interactions between optimal and non-optimal microbiomes, host cells, and other pathogens. Further, the low cost and ease of use of the design of this microphysiological model enable democratization and translation to the broader scientific community.

Acknowledgements (Optional):