Skip to content
PhageCocktails
PreclinicalGardnerella vaginalisTarget #47 of 50

Recurrent bacterial vaginosis

Bacterial vaginosis (BV) is the most common vaginal dysbiosis in reproductive-age women and is driven by the replacement of protective Lactobacillus species with a polymicrobial, Gardnerella vaginalis-anchored biofilm adherent to the vaginal epithelium. Standard antibiotics (metronidazole, clindamycin) clear symptoms but penetrate this biofilm poorly and indiscriminately suppress commensal lactobacilli, so recurrence within 3-12 months exceeds 50-60%. Phages and especially phage-derived endolysins are well suited here because they can be made exquisitely genus-specific for Gardnerella, leaving the protective lactobacilli intact, and because endolysins enzymatically degrade the peptidoglycan scaffold that holds the biofilm together. This targeted, biofilm-dissolving, resistance-resilient mode of action directly addresses the two failure points of antibiotics: biofilm recalcitrance and collateral damage to the healthy microbiome.

How phages act here

Mechanism

Two related strategies are in play. (1) Lytic Gardnerella phages (e.g., vB_Gva_AB1) infect and lyse specific clinical strains, but Gardnerella's deep strain/clade diversity makes whole-phage host range narrow, which motivates cocktail approaches. (2) The more advanced strategy uses recombinant/engineered phage endolysins (peptidoglycan hydrolases) applied directly as proteins. The lead molecule, PM-477/BNT331, is a genetically engineered, domain-optimized endolysin whose cell-wall-binding domain confers genus-level specificity: it kills all tested Gardnerella species (MIC90 ~1 ug/mL) yet does not lyse Lactobacillus crispatus, L. iners, or other commensals. Critically, endolysins attack peptidoglycan from the outside, so they physically dissolve established Gardnerella biofilms at low microgram concentrations (minimum biofilm eradication concentrations of 1-4 ug/mL), outperforming metronidazole and clindamycin in biofilm assays. Because they act on a conserved structural target rather than a metabolic pathway, resistance develops far more slowly than to antibiotics (only a marginal MIC rise after 25 passages, versus rapid high-level metronidazole resistance). This selectivity also preserves the option of phage-antibiotic or phage-probiotic synergy (debulking biofilm to let lactobacilli re-colonize).

Where it stands

Current evidence

Evidence has progressed from preclinical to first-in-human. PhagoMed Biopharma (acquired by BioNTech in 2021) developed the engineered Gardnerella endolysin PM-477, later clinical candidate BNT331. Robust preclinical data (Landlinger et al., 2022) showed Gardnerella-selective killing, superior biofilm eradication, and resistance avoidance. A 2025 ex vivo pharmacodynamics study (Tisakova et al., npj Biofilms and Microbiomes) treated vaginal samples from 49 women with BV and achieved >=94% reduction of viable Gardnerella at 20-50 ug/mL over 19 hours while sparing/allowing proliferation of L. crispatus, and these data defined the human dose. BioNTech then ran a registered Phase I first-in-human trial of BNT331 (NCT06469164 / BNT331-01): a randomized, double-blind, placebo-controlled study of single ascending doses in healthy women (Part A) and multiple ascending doses in women diagnosed with BV (Part B), using a vaginal insert formulation. The trial started July 2024 and is listed as completed (primary completion July 2025). Separately, academic groups (Arroyo-Moreno et al., 2022) have identified dozens of additional anti-Gardnerella endolysins, broadening the pipeline. No phage/endolysin product is yet approved for BV; the field is at the Phase I / proof-of-concept stage.

Evidence confidence: medium

The data

Key studies & trials

Who is working on it

Programs & centers

BioNTech SE (BNT331 endolysin; Phase I trial NCT06469164)PhagoMed Biopharma GmbH (developer of PM-477; acquired by BioNTech 2021)Brunel University London / McCarthy lab (anti-Gardnerella endolysin discovery, Arroyo-Moreno et al.)Medical University of Vienna (Farr group; ex vivo BV pharmacodynamics)Ghent University (Vaneechoutte lab; Gardnerella biofilm and endolysin testing)

The possibility

If the Phase I results hold, BV could become the first common condition routinely treated with a phage-derived enzyme rather than an antibiotic, with a vaginal insert that dissolves the Gardnerella biofilm in hours while leaving protective lactobacilli to repopulate the niche. Pairing a genus-specific endolysin with a Lactobacillus crispatus probiotic could turn today's revolving door of recurrence into a one-and-done reset of the vaginal microbiome. Because endolysins target a conserved structural wall rather than a mutable metabolic pathway, this approach may stay effective for decades where antibiotics are already failing.

Scientific & educational content. As of 2026 no bacteriophage therapeutic is approved as a marketed drug in the United States or European Union. Phage therapy is available only through clinical trials, compassionate-use / expanded-access pathways, and national magistral frameworks. Nothing here is medical advice or an offer to sell a therapeutic.