Bacterial vaginosis (BV) is a polymicrobial imbalance of the vaginal flora that occurs in 1 of 3 women in the U.S.  Little is known about what causes BV or why it predisposes women to sexually transmitted infection, urinary tract infection, placental infection and preterm birth.

Recently, we described the first mouse model of vaginal infection with clinical features of BV, (and thus, the first successful application of the experimental aspects of Koch's postulates to BV) using the "high GC Gram-positive' Gardnerella vaginalis (see Gilbert, 2013). We are now applying this model to define mechanisms of Gardnerella-host interactions.    

Mucus degradation has long been thought to play an important role in BV. One of the biochemical features of BV is the presence of sialidase activity in vaginal specimens. We have shown evidence of mucus sialoglycan depletion in specimens of women with BV and that Gardnerella vaginalis is sufficient to reproduce this phenotype in our animal model (Lewis, 2013).

Learn more about BV and our contributions to this field here.



Urinary tract infection is one of the most common infections in women and is particularly dangerous in pregnancy when it can be associated with higher risks of complications such as preterm birth. We are interested in 1) the role of the vagina as a reservoir of uropathogens, 2) how the composition of the vaginal microbiota influences host susceptibility to different types of  urinary tract infection, and 3) how these and other factors converge during and beyond pregnancy to influence adverse health outcomes.


There is a strong relationship between sexual activities and risk of recurrent UTI (rUTI). This association is commonly thought to be because sex causes mechanical transfer of E. coli from nearby body sites, such as the vagina, into the urinary tract. Like most body sites, the vagina is occupied by a bacterial community (or microbiome). We reasoned that if UTI-causing bacteria are transferred during sex, other vaginal bacteria are likely to be transferred too, and this may account for associations reported in the clinical literature between bacterial vaginosis and UTI. Using this reasoning, we recently developed a mouse model of rUTI by exposing bladder tissue containing dormant E. coli reservoirs to differet types of vaginal bacteria. This model shows that Gardnerella vaginalis, but not Lactobacillus exposures to the bladder can trigger recurrent E. coli UTI and lead to potentially serious, life threatening forms of UTI. Learn more about this work here.