Cervicovaginal microbiota in female sex workers with bacterial vaginosis: A metagenomic perspective

Dear Editor,

Bacterial vaginosis (BV) is a common vaginal dysbiosis, characterized by depletion of protective lactobacilli and an overgrowth of opportunistic, facultative anaerobic bacteria.¹ BV has been implicated in adverse reproductive and gynaecological outcomes, including preterm birth, miscarriage, and an increased susceptibility to sexually transmitted infections (STIs) such as HIV and non-gonococcal/non-chlamydial pelvic inflammatory disease.² Globally, BV affects approximately 23% to 29% of women of reproductive age, with prevalence varying based on region and ethnicity.³ However, it disproportionately affects high-risk populations, particularly female sex workers (FSWs) who remain underrepresented in microbiome research.⁴ Clinically, BV often presents with abnormal vaginal discharge, a characteristic fishy odour, and vaginal itching or irritation. Diagnosis is typically made using Amsel’s criteria or Nugent scoring. However, these conventional methods are limited by subjectivity and lack the resolution needed to detect new morphotypes.

In this cross-sectional study, we aimed to characterise the cervicovaginal microbial communities of FSWs diagnosed with BV in India using high-throughput 16S rRNA gene-based metagenomic sequencing. Cervicovaginal swabs were collected from non-menstruating FSWs who had complaints of abnormal vaginal discharge between October 2021 and December 2022. Participants were recruited through a network of non-governmental organizations affiliated with a regional STI reference, research and training centre. Informed written consent was obtained from all participants. BV diagnosis was confirmed based on Nugent scoring,⁵ and only consecutive samples scoring between 7 and 10 were included for this study. In parallel, all samples were screened for Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis using standard diagnostic assays. Genomic DNA was extracted using the Chelex® method⁶ (Sigma Aldrich, MO, USA), purified with AMPure XP beads (Beckman Coulter, USA), and subjected to the amplification of the V3-V4 hypervariable region of the 16S rRNA gene using the Quick-16S^TM NGS Library Prep Kit (Zymo Research, USA). Amplicon libraries were sequenced on the Illumina MiSeq platform (2 × 300 bp paired-end reads). Bioinformatic and statistical analyses were performed using the Qiime2 pipeline (version 2024.5) and R (2024.09.1). Reads were demultiplexed, quality filtered to remove low-quality sequences, and denoised using the DADA2 algorithm to generate amplicon sequence variants. Taxonomic assignment was performed using a pre-trained Naïve Bayes classifier trained on the Silva 138-99 reference database.

All samples tested negative for N. gonorrhoeae and C. trachomatis, while T. vaginalis was detected in 4% of cases. The majority of participants were aged between 29 and 39 years and reported symptoms consistent with BV, including thin homogenous discharge (83.3%), malodour (50.5%), and itching (33.3%), persisting for approximately 1 month. All participants tested negative for HIV. High-throughput sequencing of the 30 BV-positive samples yielded an average of 0.1 million reads per sample, revealing a microbial landscape dominated by four major bacterial phyla: Firmicutes (41.8%), Actinobacteriota (26.9%), Bacteroidota (17.1%), and Fusobacteriota (10.8%), cumulatively accounting for >96% of the total community. Within Firmicutes, the most dominant phylum, the predominant genera included Lactobacillus (19.5%), Shuttleworthia (6.8%), Megasphaera (4.3%), Aerococcus (2.6%), Dialister (1%), and Mycoplasma (0.7%). Actinobacteriota were primarily represented by Gardnerella (26.6%) and Atopobium (5.2%), while Bacteroidota included Prevotella (9.2%), Prevotella_7 (2.6%), and Porphyromonas (1%). Sneathia (10.9%) constituted the major representative of Fusobacteriota. Minor contributions were noted from Proteobacteria (2.8%), primarily Escherichia-Shigella (3.1%); Patescibacteria (0.3%), including Saccharimonadales (0.2%); and Campylobacterota (0.1%) represented by Campylobacter (0.1%) [Figure 1]. Hierarchical clustering of genus-level profiles demonstrated distinct microbial community patterns across individuals with variable dominance of Lactobacillus, Gardnerella, and Sneathia [Figure 2]. Spearman’s rank correlation analysis showed a positive correlation between Sneathia and Prevotella (ρ = 0.724, p = 0.00000193) and a negative correlation between Sneathia and Lactobacillus (ρ = -0.588, p = 0.000322), suggesting potential synergistic and antagonistic microbial interactions. Clinically, participants exhibiting higher Prevotella and Sneathia levels more frequently reported persistent discharge and malodour, whereas those enriched in Lactobacillus had milder symptom profiles. Notably, the sample positive for T. vaginalis displayed a similar dysbiotic profile.

Close

Figure 1:

(a) Stacked bar plot showing the relative abundance of bacterial genera from phylum to genus level across cervicovaginal samples (left panel). (b) Relative abundance of the top 20 dominant genera in the cervicovaginal samples. Relative abundances were computed by first normalising feature counts within each sample to account for differences in sequencing depth, followed by averaging genus-level abundances across the entire cohort. The y-axis represents the cumulative relative abundance (%), and each coloured segment corresponds to one of the top 20 genera (right panel).

Export to PPT

Close

Figure 2:

Heatmap of the top 20 most abundant bacterial genera across all samples. The heatmap displays the relative abundance (row-scaled) of the top 30 genera across all samples with hierarchical clustering applied to both genera (rows) and samples (columns). Color intensity represents the scaled abundance (Z-score), ranging from low (blue) to high (red).

Export to PPT

These findings underscore the microbial heterogeneity within the high-risk cohort and confirm the predominance of BV-associated anaerobes. Our observations align with findings from Afro-Caribbean and Chinese cohorts, where FSWs also exhibited elevated levels of Prevotella and Sneathia and increased microbial diversity. The depletion of lactobacilli and enrichment of these pro-inflammatory genera may contribute to mucosal inflammation, epithelial disruption, and immune dysregulation, potentially exacerbating susceptibility to STIs and reproductive outcomes.^7-8 These shifts in microbial communities highlight the clinical importance of dysbiosis and the need for microbiome-informed diagnostics and the development of targeted therapeutic strategies, particularly in high-risk populations. Limitations of this study include the absence of a healthy control group, a relatively small sample size, and the inherent genus-level resolution of 16S rRNA gene sequencing. Despite these constraints, our findings provide insight into the cervicovaginal microbial ecology of FSWs diagnosed with BV in India and support the integration of microbiome-based surveillance into STI prevention and reproductive health strategies targeting vulnerable populations.