Bacterial vaginosis (BV) is a common condition associated with discomfort and negative outcomes such as pre-term birth. BV is a dysbiosis caused by a shift from normal vaginal lactobacilli species to other BV-associated bacteria, including various species of the Mobiluncus genus, broadly classified as “Mobiluncus morphotypes” by their shared curved rod shape on a Gram-stain. Previously, the Fredricks Lab (Vaccine and Infectious Disease Division) has shown that a non-Mobiluncus curved-rod, BV-associated bacterium (BVAB)-1, has high specificity for BV. Unlike most vaginal bacteria, BVAB1 are flagellated, which may help the bacteria to ascend into the upper genital tract and cause further disease.
Innate immune sensors have evolved to specifically detect pathogen components such as flagella. Toll-like receptor 5 (TLR5) is an innate immune receptor that is expressed in the vaginal epithelium and is known to recognize the flagellin monomer. Despite this connection, the role of TLR5-mediated detection of flagellated bacteria in BV has not been studied until now, in a recent Infection and Immunity manuscript led by graduate student Erin dela Cruz of the Fredricks Lab, Vaccine and Infectious Disease Division. The authors hypothesized that TLR5-deficiency would lead to decreased immune activation to Mobiluncus morphotypes, allowing for increased colonization and BV symptoms. To test this, they enrolled both BV-negative and BV-positive women, some of whom were confirmed to lack functional TLR5, and cultured the bacteria present in vaginal fluid. Contrary to their hypothesis, they found that women with TLR5 deficiency had decreased vaginal Mobiluncus spp. or BVAB1 colonization, while Mobiluncus species M. curtisii and non-flagellated bacteria populations were unchanged based on TLR5 status. Additionally, TLR5 functionality had no association with a positive or negative BV diagnosis.
Based on these surprising results, the authors adjusted their hypothesis to posit that TLR5 deficiency somehow protects against colonization by flagellated M. mulieris and BVAB1, but not M. curtisii. To determine the mechanism behind this relationship, they compared the flagellin amino acid sequences (FlaA) of M. mulieris and BVAB1 to identify shared domains. They compared these to sequences from previously characterized TLR5 agonists and flagellated bacteria known to escape TLR5 recognition to identify a shared domain that is likely to be recognized by TLR5. After identifying a 10-amino acid N-terminal sequence conserved across 205 flagellated species, they found that BVAB1 and M. mulieris share a key TLR5 recognition residue in this region while M. curtisii contains a key amino acid substitution, explaining why TLR5 activity influences BVAB1 and M. mulieris populations but not M. curtisii.
To confirm the ability of M. mulieris to stimulate TLR5, the authors cultured heat-killed M. mulieris or M. curtisii with a cell line expressing TLR5 and found that only M. mulieris stimulated a TLR5-specific immune response. Since BVAB1 has never been grown in pure culture and therefore could not be directly cultured with TLR5-expressing cells to test its ability to stimulate the receptor, cervicovaginal lavage (CVL) supernatant from women with high vaginal BVAB1 concentrations was cultured with TLR5-expressing cells, and a strong TLR5 response was induced. These findings suggest that flagellated BVAB may require an inflammatory vaginal environment to thrive, explaining why women lacking TLR5 (and the resulting immune activation pathway) host fewer BVAB1 and M. mulieris bacteria. Although this model initially appears counterintuitive, the authors speculate that a TLR5-mediated proinflammatory response may create a niche for flagellated BVAB in several ways, such as by decreasing the availability of host-derived nutrients needed by competing normal vaginal bacteria or by increasing mucin production, which could be a fuel source for flagellated BVAB.
Interestingly, although M. mulieris and BVAB1 induced TLR5 activation, they did not induce the production of the inflammatory cytokine interleukin-8 (IL-8), a downstream product of TLR5 activation, which initially surprised the authors. However, the sequence alignments compared earlier in this study revealed that FlaA proteins encoded by Mobiluncus spp. are highly variable and often include a “disordered domain” not found on other flagellins. The authors speculate that this additional variable domain could help to evade the immune system, so that the flagellated bacteria can benefit from the early effects of TLR5 stimulation while shutting down immune responses that would be deleterious to themselves.
This work partially illuminates the relationship between TLR5 and flagellated BVAB, revealing a complex interaction in which these pathogenic species may benefit from some level of BVAB-induced inflammatory immune activation. These findings also demonstrate that the “disordered domain” may constitute a novel adaptation for bacteria to influence the host immune response. Going forward, the Fredricks Lab plans to investigate the TLR5-mediated immune response and its role in halting or aiding BVAB; while beneficial for Mobiluncus spp. vaginal colonization, TLR5 responses may prevent these flagellated bacteria from ascending higher in the vaginal tract. Thus, TLR5 may mediate a delicate balance by allowing local BVAB colonization in order to prevent more severe and disseminated pathology.
This work was supported by the National Institute of Health, the American Society for Microbiology Watkins Fellowship, and the University of Washington Achievement Rewards for College Scientists.
UW/Fred Hutch Cancer Consortium members David Fredricks and Thomas Hawn contributed to this work.
Dela Cruz EJ, Fiedler TL, Liu C, Munch MM, Kohler CM, Oot AR, Wallis JM, Wang J, Frishman A, Garcia K, Wiser A, Balkus JE, Srinivasan S, Golob JL, Sycuro LK, Marrazzo JM, Hawn TR, Fredricks DN. Genetic variation in Toll-like receptor-5 and colonization with flagellated bacterial vaginosis-associated bacteria. Infect Immun. 2020 Nov 16;IAI.00060-20. doi: 10.1128/IAI.00060-20. Online ahead of print.