Metronidazole treatment rapidly reduces genital inflammation through effects on bacterial vaginosis-associated bacteria rather than lactobacilli

Abstract

BackgroundBacterial vaginosis (BV) causes genital inflammation and increases HIV risk, whereas a vaginal microbiota dominated by Lactobacillus species is associated with immune quiescence and relative HIV protection. BV treatment reduces genital inflammation, but it is unclear whether this reduction is driven by a decrease in BV-associated bacteria or an increase in Lactobacillus species.METHODSTo evaluate the short-term effect of standard BV treatment on genital immunology and the vaginal microbiota, vaginal swabs were collected immediately before and after metronidazole treatment for BV and analyzed with multiplex ELISA, metagenomic sequencing, and quantitative PCR.RESULTSTopical metronidazole treatment rapidly reduced vaginal levels of proinflammatory cytokines, chemokines, and soluble immune markers of epithelial barrier disruption. Although the vaginal microbiota shifted to dominance by L. iners or L. jensenii, this proportional shift was primarily driven by a 2 to 4 log10-fold reduction in BV-associated bacteria absolute abundance. BV treatment induced no change in the absolute abundance of L. crispatus or L. iners and only minor (<1 log10-fold) increases in L. gasseri and L. jensenii that were not independently associated with reduced inflammation in multivariable models.CONCLUSIONThe genital immune benefits that are associated with Lactobacillus dominance after BV treatment were not directly attributable to an absolute increase in lactobacilli, but rather to the loss of BV-associated bacteria.Trial REGISTRATIONParticipants were recruited as part of a randomized controlled trial (ClinicalTrials.gov NCT02766023) from 2016 to 2019.FUNDINGCanadian Institutes of Health Research (PJT-156123) and the National Institute of Allergy and Infectious Diseases (HHSN2722013000141 and HHSN27200007).

Keywords: Bacterial infections; Cytokines; Immunology; Microbiology.

Conflict of interest statement

Conflict of interest: CRC is a paid consultant for Osel Inc.

Figures

Figure 1. Flow chart of study design.

This flow diagram was adapted from the original clinical trial by Cohen et al. (17) and expanded to include the selection criteria and sample size for this substudy.

Figure 2. Rapid impact of topical metronidazole on vaginal immunology and markers of epithelial barrier disruption.

Log10-transformed vaginal levels of (A) IL-1α, (B) IL-6, (C) IL-8, (D) IP-10, (E) MIP-1β, (F) MIP-3α, (G) soluble E-cadherin, (H) MIG, and (I) MMP-9 were compared before and immediately after metronidazole treatment, using Wilcoxon’s matched-pairs signed-rank test (n = 48). P values are reported above graph.

Figure 3. Rapid impact of topical metronidazole on detectability of vaginal soluble immune factors.

Detectability of (A) IFN-α2a and (B) IL-17A was compared before and immediately after metronidazole treatment using McNemar’s test (n = 48). P values are reported above each graph.

Figure 4. Shift of the vaginal microbiome to Lactobacillus spp. predominance after metronidazole treatment of BV.

(A) The stacked bar plot shows the relative abundances of the most common vaginal bacterial taxa among participants with clinical BV prior to treatment, organized by community state type (CST; n = 45). (B) The stacked bar plot shows the relative abundances of the most common bacterial taxa immediately after metronidazole treatment, organized by CST (n = 32).

Figure 5. Impact of metronidazole treatment on the relative abundance of BV-associated bacteria and vaginal Lactobacillus species.

Forest plot showing the impact of metronidazole treatment on the relative abundance of L. crispatus, L. iners, L. gasseri, and L. jensenii and the BV-associated bacteria Prevotella, G. vaginalis, A. vaginae, and Megasphaera (n = 32). Boxes represent the mean percentage change and whiskers represent the 95% CI. P values were determined with Wilcoxon’s matched-pairs signed-rank test.

Figure 6. Impact of metronidazole treatment on the absolute abundance of BV-associated bacteria and vaginal Lactobacillus species.

Forest plot shows change in absolute abundance of L. crispatus, L. iners, L. gasseri, and L. jensenii and the BV-associated bacteria Prevotella, G. vaginalis, A. vaginae, and Megasphaera, represented by log10-transformed fold change (n = 48). Boxes represent the mean log10-transformed fold change and whiskers represent the 95% CI. P values were determined with Wilcoxon’s matched-pairs signed-rank test.