Microbiome and metabolomic changes associated with HPV clearance in women undergoing local excisional treatment for cervical intraepithelial neoplasia

Abstract

Cervical intraepithelial neoplasia (CIN) is a common gynecological condition often associated with persistent human papillomavirus (HPV) infections. Although the loop electrosurgical excision procedure (LEEP) is effective in removing lesions, some patients remain HPV positive post-treatment. In this prospective study, we enrolled reproductive-age women diagnosed with HPV-related CIN and employed a multi-omics analysis of cervicovaginal secretion and cervical tissue microbiomes, alongside non-targeted and targeted metabolomic assessments. We aim to explore the role of the cervicovaginal and intratissue microbiota and associated metabolites on HPV clearance following LEEP. We observed significant shifts in bacterial diversity and composition in both cervicovaginal secretion and cervical tissue samples. Notably, distinct bacterial species, such as Lactobacillus and certain anaerobes (e.g., Prevotella bivia), were correlated with HPV clearance post-LEEP. Metabolomic profiling revealed that the HPV-cleared group exhibited elevated acetic acid levels and significant alterations in glycerophospholipid metabolism, suggesting a potential role in promoting HPV clearance. Correlation analyses demonstrated significant associations between altered bacteria and metabolites with HPV status, with models incorporating both achieving high predictive accuracy. Overall, our findings highlight the importance of the cervicovaginal and intratissue microbiomes and metabolites in facilitating HPV clearance, suggesting potential therapeutic targets for patients with CIN.

Importance: The clearance of human papillomavirus (HPV) after local excisional treatment for cervical intraepithelial neoplasia is crucial for patient health. This study reveals significant alterations in the cervicovaginal secretion and cervical tissue microbiomes, alongside metabolomic changes, which are associated with HPV clearance. Through a comprehensive multi-omics approach, we identified specific bacterial species and metabolic changes that correlate with successful HPV clearance post-loop electrosurgical excision procedure. Notably, the presence of beneficial Lactobacillus species and elevated levels of acetic acid linked to glycerophospholipid metabolism emerged as potential biomarkers for HPV status, suggesting that these factors play a pivotal role in improving treatment outcomes. These findings highlight the potential for microbiome-targeted therapies to enhance HPV clearance and provide insights into the microbial and metabolic mechanisms involved in cervical health.

Keywords: HPV; cervical intraepithelial neoplasia; metabolome; microbiome.

Fig 1

Comparison of the results of the 5R 16S rRNA gene sequencing from cervicovaginal secretions and cervical tissue samples. (A) Study workflow illustrating that participants underwent sequencing of cervicovaginal secretion and cervical tissue 5R 16S rRNA, along with targeted metabolomic analysis of SCFAs and non-targeted metabolomics, followed by bioinformatic integration of microbiome and metabolome data. (B and C) Venn diagrams indicating the distribution of shared and unique microbiota at the phylum level (B) and species level (C). (D) A butterfly diagram showcasing the top 20 bacterial species from cervicovaginal secretion and cervical tissue samples, with shared species highlighted in red.

Fig 2

Analysis of bacterial diversity and composition of cervicovaginal secretions pre- and post-LEEP surgery. (A) Venn diagrams comparing bacterial species before and after LEEP, specifically between pre-NR and post-NR (top) and pre-R and post-R (bottom). Pre, samples collected before LEEP surgery; post, samples collected after LEEP surgery. (B) Alpha diversity measures, using Shannon (top) and Simpson (bottom) indices, comparing pre- and post-LEEP samples for both groups. P values derived from the Mann-Whitney U test. (C) Beta diversity analyzed through PCoA based on Bray-Curtis distances between pre-NR and post-NR (left), and between pre-R and post-R (right), with significance assessed using ANOSIM. (D) Heatmap depicting the top 50 bacterial species across 86 samples. (E) Visualization of bacterial clusters within the analyzed groups. (F) Hierarchical clustering on the average relative proportions of the 30 most abundant species among the four groups based on Bray-Curtis distances. (G) A Venn diagram presenting unique differentially abundant bacterial species (DABs) specific to the R group (Mann-Whitney U test, P < 0.05). (H) A bubble diagram showcasing the four DABs exclusive to the R group.

Fig 3

Analysis of bacterial diversity and composition of cervical tissue pre- and post-LEEP surgery. (A) Venn diagrams comparing bacterial species before and following LEEP, specifically for pre-NR vs post-NR (top) and pre-R vs post-R (bottom). (B) Assessment of alpha diversity via Shannon (top) and Simpson (bottom) indices between the sample pairs, using the Mann-Whitney U test for statistical analysis. (C) Beta diversity determined through PCoA based on Bray-Curtis distances between pre-NR and post-NR (left), as well as between pre-R and post-R (right), with significance evaluated by ANOSIM. (D) Heatmap illustrating the top 50 bacterial species across 68 samples. (E) Visualization of bacterial clusters within the assessed groups. (F) Hierarchical clustering on the average proportions of the 30 most abundant species among the four groups based on Bray-Curtis distances. (G) A Venn diagram representing unique DABs for the R group (Mann-Whitney U test, P < 0.05). (H) A bubble diagram illustrating 11 DABs identified within the R group.

Fig 4

Metabolic profiling of the cervicovaginal metabolome pre- and post-LEEP surgery. (A) PCA plots based on Bray-Curtis distances for SCFAs comparing pre-NR vs post-NR (top) and pre-R vs post-R (bottom), with significance derived from ANOSIM. (B) Box plots depicting SCFA levels for comparing pre-NR and post-NR (top), as well as pre-R and post-R (bottom), before and after LEEP. (C and E) PCA plots for non-targeted metabolomics relating to pre-NR vs post-NR (C) and pre-R vs post-R (E), assessed for significance via ANOSIM. (D and F) Partial least squares discriminant analysis was utilized for comparisons between pre-NR and post-NR (D), alongside the pre-R and post-R (F), with model validity confirmed through 200 permutations. (G) Volcano plots display metabolic alterations between pre-NR vs post-NR (left) and pre-R vs post-R (right), with notable metabolites indicated by red (upregulated) and blue (downregulated) dots based on VIP score and Mann-Whitney U test thresholds. (H) A Venn diagram showing unique differential metabolites specific to the R group. (I) A bubble chart detailing the 20 enriched KEGG pathways associated with the different metabolites shown in panel H. (J) Heatmap illustrating significantly altered metabolites related to glycerophospholipid metabolism.

Fig 5

Correlation analysis among altered metabolites, cervicovaginal bacterial species, and clinical indices in the R group. (A) Correlation of clinical indices (HPV status and pH) with the nine altered metabolites related to glycerophospholipid metabolism and SCFAs, alongside four cervicovaginal DABs. (B) Receiver operating characteristic (ROC) curves displaying discriminatory signatures from the four bacterial species and nine metabolites for the cohort of 86 samples. (C) Pearson’s correlation analysis between DABs and clinical indices. ^*P < 0.05, **P < 0.01, or ***P < 0.001. (D) Linear regression of A. lactolyticus against pH change. The solid black line indicates a statistically significant linear relationship (P < 0.05), and the shaded regions represent the 95% confidence intervals. Each black dot represents an individual sample. (E) Correlation analysis of altered metabolites with clinical indices. *P < 0.05, **P < 0.01, or ***P < 0.001. (F) Linear regression for the five metabolites and HPV status.

Fig 6

Correlation analysis among altered metabolites, cervical tissue bacterial species, and clinical indices in the R group. (A) Correlation of clinical indices (HPV status and pH) with the 9 altered metabolites related to glycerophospholipid metabolism and SCFAs, along with 12 cervical tissue DABs. (B) Receiver operating characteristic (ROC) curves constructed to display discriminatory signatures from the 12 bacterial species and 9 metabolites for the 68 sample cohort. (C) Pearson’s correlation analysis between DABs and clinical indices. *P < 0.05, **P < 0.01, or ***P < 0.001. (D) Linear regression analyses for P. bivia and HPV status, and Streptococcus infantis and pH change. The solid black line represents a statistically significant linear relationship (P < 0.05), while the shaded areas illustrate the 95% confidence intervals. Each black dot corresponds to an individual sample. (E) Correlation analysis of altered metabolites with clinical indices. *P < 0.05, **P < 0.01, or ***P < 0.001. (F) Linear regression for the seven indicated metabolites and HPV status.