The inferred modulation of correlated vaginal microbiota and metabolome by cervical differentially expressed genes across distinct CIN grades

Abstract

Background

In vitro studies have demonstrated the modulation of vaginal microbiota (VM) by cervical peptides which levels varied with the status of HPV infection and cervical intraepithelial neoplasia (CIN) grades. However, there is a deficiency in population-based studies investigating the modulation of VM compositions and metabolome by cervical differentially expressed genes (DEGs) across different grades of CIN.

Methods

This study included 43 HPV-positive women, classified into low-grade (CIN1, n = 23) and high-grade (CIN2 + , n = 20) groups. Vaginal swabs were collected for both microbiota and metabolome analysis. Cervical exfoliated cells were collected for RNA-Seq analysis.

Results

We identified 258 differentially expressed genes (DEGs), among which 176 CIN1-enriched genes were linked to immune responses, cell chemotaxis, negative regulation of cell migration, and B cell differentiation, activation, and proliferation. Eighty-two genes upregulated in CIN2 + cohorts were associated with epidermis development and keratinization. Then, we identified 5,686 paired correlations between DEGs, VM, and metabolome, with 2,320 involving Lactobacillus. Further analysis revealed Lactobacillus as the primary determinant of metabolic profiles, followed by Gardnerella, Faecalibacterium, Aerococcus and Streptococcus, such as the notable positive correlation between Lactobacillus with D-lactic acid and DL-indole-3-lactic acid. Applying mediation analysis, we found that Lactobacillus mediated the association of 14 CIN1-enriched DEGs, such as COL4A2, CCBE1 and SPON1, with the production of 57 metabolites, including D-lactic acid, oleic acid and various amino acids. Additional analysis indicated significant mediation effects of 79 metabolites on the association of DEGs with the growth of Lactobacillus, Gardnerella, Fannyhessea and Aerococcus.

Conclusions

Our findings provide valuable population-based evidence for the inferred modulation of correlated VM and metabolome by cervical DEGs across different CIN stages.