Extracellular vesicles from vaginal Gardnerella vaginalis and Mobiluncus mulieris contain distinct proteomic cargo and induce inflammatory pathways

Abstract

Colonization of the vaginal space with bacteria such as Gardnerella vaginalis and Mobiluncus mulieris is associated with increased risk for STIs, bacterial vaginosis, and preterm birth, while Lactobacillus crispatus is associated with optimal reproductive health. Although host-microbe interactions are hypothesized to contribute to reproductive health and disease, the bacterial mediators that are critical to this response remain unclear. Bacterial extracellular vesicles (bEVs) are proposed to participate in host-microbe communication by providing protection of bacterial cargo, delivery to intracellular targets, and ultimately induction of immune responses from the host. We evaluated the proteome of bEVs produced in vitro from G. vaginalis, M. mulieris, and L. crispatus, identifying specific proteins of immunologic interest. We found that bEVs from each bacterial species internalize within cervical and vaginal epithelial cells, and that epithelial and immune cells express a multi-cytokine response when exposed to bEVs from G. vaginalis and M. mulieris but not L. crispatus. Further, we demonstrate that the inflammatory response induced by G. vaginalis and M. mulieris bEVs is TLR2-specific. Our results provide evidence that vaginal bacteria communicate with host cells through secreted bEVs, revealing a mechanism by which bacteria lead to adverse reproductive outcomes associated with inflammation. Elucidating host-microbe interactions in the cervicovaginal space will provide further insight into the mechanisms contributing to microbiome-mediated adverse outcomes and may reveal new therapeutic targets.

Fig. 1. Size, morphology, and concentration of extracellular vesicles derived from L. crispatus, G. vaginalis, and M. mulieris grown in NYC culture medium.

a Transmission electron microscopy indicates small, spherical vesicles isolated from bacterial culture samples but not NYC culture medium. b Particle size distributions from Nanoparticle Tracking Analysis indicate the presence of vesicles between 90 and 420 nm in diameter. All scale bars are 500 nm.

Fig. 2. Proteomic analysis of extracellular vesicles derived from L. crispatus, G. vaginalis, and M. mulieris grown in NYC culture medium.

a Out of 1074 total bacterial proteins identified, 90 orthologs were shared between all three species’ isolates, and 247, 491, and 336 were specific to L. crispatus, G. vaginalis, and M. mulieris isolates, respectively. b Select shared and distinct proteins are of functional interest; protein abundance (proteins per million) is shown by heat map. c Proteins were classified according to predicted subcellular localization.

Fig. 3. Uptake of L. crispatus, G. vaginalis, and M. mulieris bEVs by cervical and vaginal epithelial cells over 24 h.

bEV preparations from (a) NYC culture medium (control), (b) L. crispatus, (c) G. vaginalis, and (d) M. mulieris were labeled with rhodamine B isothiocyanate and observed in the cytoplasm of ectocervical (Ect1), endocervical (End1), and vaginal epithelial (VK2) cells after 1, 4, and 24 h of incubation. All scale bars are 20 μm.

Fig. 4. Dose-dependent cytokine responses from cervical and vaginal epithelial cells after 24 h exposure to L. crispatus, G. vaginalis, and M. mulieris bEVs.

IL-8 expression measured by ELISA increases with increasing doses of bEVs from G. vaginalis and M. mulieris, but not L. crispatus and NYC culture medium (control), relative to nontreated (a) ectocervical, (b) endocervical, and (c) vaginal epithelial cells. d Exposure to bacterial EVs does not induce significant changes in lactate dehydrogenase release. 29-plex Luminex array identified a multi-cytokine response to bEVs from G. vaginalis and M. mulieris, but not L. crispatus, expressed as fold-changes relative to NYC culture medium control-treated (e) ectocervical, (f) endocervical, and (g) vaginal epithelial cells (n = 3 per treatment). Error bars indicate standard deviation and asterisks indicate adjusted p-values < 0.05 (*), < 0.01 (**), < 0.001 (***), and <0 .0001 (****) via one-way ANOVA with Tukey’s correction for multiple comparisons.

Fig. 5. Cytokine responses from THP-1 monocytes after 24 h exposure to L. crispatus, G. vaginalis, and M. mulieris bEVs.

29-plex Luminex array identified a multi-cytokine response in THP-1 monocytes exposed to bEVs from G. vaginalis and M. mulieris, and a limited response to exposure to bEVs from L. crispatus, expressed as fold-changes relative to exposure to bEVs prepared from NYC culture medium (n = 3 per treatment). Error bars indicate standard deviation and asterisks indicate adjusted p-values < 0.05 (*), < 0.01 (**), < 0.001 (***), and < 0.0001 (****) via one-way ANOVA with Tukey’s correction for multiple comparisons.

Fig. 6. TLR2-specific pathways activated by L. crispatus, G. vaginalis, and M. mulieris bEVs.

In HEK-TLR2 reporter cells exposed to bEVs for 24 h, G. vaginalis and M. mulieris bEVs induce significantly increased expression of (a) NF-kB and (b) IL-8. c Exposure to any bEV treatment does not induce cell death relative to non-treated controls, as measured by lactate dehydrogenase release. Error bars indicate standard deviation and p-values were calculated via one-way ANOVA with Tukey’s correction for multiple comparisons.