Skip to main content
NCBI home page
NPJ Biofilms and Microbiomes logoLink to NPJ Biofilms and Microbiomes
. 2026 Jul 31;12:208. doi: 10.1038/s41522-026-01120-6

Role of cervicovaginal microbiome and metabolome in ALA-PDT-mediated HPV clearance in patients with cervical HSIL

Ting Zhang 1,2, Chunyan Zhang 1,2, Shuyu Yuan 1, Yimeng Wang 1, Mengzhuo Zhang 1, Kaili Dai 1, Yiduo Yu 1, Lijuan Zhang 1, Lili Cao 1,2, Mengzhen Zhang 1,2,✉, Yuehui Su 1,2,✉
PMCID: PMC13680506  PMID: 42897152

Abstract

The cervicovaginal microbiome and its associated metabolites play a pivotal role in the persistence of human papillomavirus (HPV) infection. Here, we investigated the contribution of the cervicovaginal microbiome and metabolome to 5-aminolevulinic acid photodynamic therapy (ALA-PDT)-mediated HPV clearance in patients with cervical high-grade squamous intraepithelial lesions (HSIL). 16S ribosomal RNA sequencing analysis identified significant differences in cervicovaginal microbiota before ALA-PDT between the HPV clearance and non-clearance groups. The HPV non-clearance group exhibited higher abundances of Anaerococcus and Gemella. ALA-PDT exerted a more substantial impact on cervicovaginal microbiota diversity in the HPV clearance group, characterized by increased abundances of Finegoldia, Corynebacterium, Lancefieldella, and Winkia, and a decreased abundance of Anaeroglobus following treatment. Global untargeted metabolomics analysis further revealed distinct cervicovaginal metabolic profiles before ALA-PDT between the two groups. The pathways of mineral absorption and valine, leucine, and isoleucine biosynthesis were significantly enriched in the HPV clearance group. Alterations in cervicovaginal metabolic profiles following ALA-PDT were also observed. Moreover, several key differential metabolites were observed to be significantly correlated with Anaeroglobus. Our findings underscore the critical involvement of the cervicovaginal microbiome and metabolome in ALA-PDT-mediated HPV clearance, providing valuable clues for predicting and improving the effect of HPV clearance following ALA-PDT.

Subject terms: Diseases, Microbiology

Introduction

Cervical cancer (CC) is the second most common malignancy among women, following breast cancer. Cervical high-grade squamous intraepithelial lesions (HSIL), as precancerous lesions of CC, carry a significant risk of progression to CC. Studies have demonstrated that approximately one-third of cervical HSIL can progress to invasive cancer in the absence of timely intervention1,2. Therefore, clinical management of cervical HSIL is frequently implemented to reduce the incidence of CC.

Currently, the primary clinical approach for treating cervical HSIL involves surgical resection, including cold knife conization and loop electrosurgical excisional procedure. However, surgical resection compromises cervical anatomy and function, leading to complications such as diminished cervical secretions and adverse reproductive morbidity in a subsequent pregnancy3. With the rising incidence of cervical HSIL among younger women, there is an urgent need for minimally invasive and non-invasive therapies that preserve cervical structure and function while maintaining therapeutic efficacy. Photodynamic therapy (PDT) is a novel, highly selective, and non-invasive therapeutic modality approved for treating some precancerous conditions and solid tumors. PDT has been introduced into clinical practice for treating cervical HSIL in China, demonstrating great therapeutic potential. However, findings from our previous research indicated that although the complete remission rate of PDT in the treatment of cervical HSIL could reach about 90%, approximately 40% of the patients exhibited persistent human papillomavirus (HPV) infection following treatment4–6. This proportion rises by nearly 50% among individuals with concurrent vaginal or vulvar SIL7. Persistent HPV infection after treatment exacerbates the psychological burden of patients and, more importantly, elevates the risk of lesion recurrence.

Microorganisms, metabolites, and immune components constitute key components of the female reproductive tract microenvironment, playing an essential role in maintaining homeostasis and overall reproductive health8. Additionally, disruption of the cervicovaginal microbiome structure contributes to HPV persistence and facilitates the onset and progression of cervical lesions and cancer9. Notably, a profile of microbes characterized by a predominance of Prevotella bivia, Enterococcus durans, and Porphyromonas uenonis, along with a concomitant paucity of Lactobacillus iners and Prevotella disiens, has been associated with HPV persistent infection10. Furthermore, several studies suggest that the microbiota can influence therapeutic efficacy through mechanisms involving microbial translocation, immunomodulation, metabolism, and enzymatic degradation11. For example, the gut microbiome regulates the treatment response to immunotherapy in patients with melanoma, lung cancer, and kidney cancer11. Moreover, the gut and cervical microbiome have been demonstrated to modulate the response to chemoradiation in patients with CC12. Indeed, the interactions between therapy and the commensal microbiome are mutual. The structure of the cervicovaginal microbiome can be affected by various treatment modalities, including surgery, focused ultrasound, and chemoradiation13–16. These findings suggest that the cervicovaginal microbiome influences the treatment outcome of cervical lesions and highlight the potential for manipulating the cervicovaginal microbiome to enhance therapeutic efficacy. Indeed, the significant potential of microbiome manipulation in enhancing gynecological health and improving the therapeutic outcomes of gynecological diseases has been demonstrated11. The strategies employed to manipulate the female reproductive tract microbiomes mainly include the use of vaginal probiotics, prebiotics, novel antimicrobials, biofilm disruptors, and microbiome transplantation11.

Based on the findings of the above studies, it was hypothesized that the cervicovaginal microbiome may play a role in the process of PDT to facilitate HPV clearance in patients with cervical HSIL. Accordingly, 16S ribosomal RNA (rRNA) sequencing was performed to detect and analyze the cervicovaginal microbiome of patients before and after PDT, with the aim of investigating the interaction between the cervicovaginal microbiome and PDT. The underlying mechanism was also preliminarily examined through the metabolomic analysis of cervicovaginal secretions.

Results

Clinical characteristics of participants

A total of 74 patients diagnosed with cervical HSIL who underwent ALA-PDT were enrolled in this study. Of these, 24 patients exhibited persistent HPV infection after treatment (HPV non-clearance group), while 50 patients achieved HPV clearance (HPV clearance group). The mean ages of the HPV non-clearance and HPV clearance groups were 29.4 and 31.1 years old, respectively, with no statistically significant difference in age distribution observed between the two groups (P = 0.297) (Supplementary Table 1). Additionally, the distribution of HPV infections between the two groups did not differ significantly (P = 0.339) (Supplementary Table 1). Utilizing 16S rRNA sequencing data, the community state types (CSTs) of these patients’ pre-treatment samples were classified as follows: 3 CST Ⅰ, 4 CST Ⅱ, 59 CST Ⅲ, and 8 CST Ⅳ (Supplementary Table 2). A comparative analysis indicated no statistically significant difference in CST distributions between the HPV non-clearance and HPV clearance groups (P = 0.480) (Supplementary Table 1). Among the 42 patients with both pre-treatment and post-treatment samples, no statistically significant differences were observed in the distributions of age (P = 0.532), HPV infection (P = 0.166), and baseline CSTs (P = 0.763) between the HPV non-clearance and HPV clearance groups (Supplementary Table 3). Notably, all of the enrolled patients had no smoking history, and they also had no alcohol drinking history during the treatment and follow-up periods.

Comparison of cervicovaginal microbiome before ALA-PDT between the HPV clearance and non-clearance groups

To investigate the role of cervicovaginal microbiota in ALA-PDT-mediated HPV clearance in patients with cervical HSIL and to identify potential microbial predictive biomarkers for ALA-PDT outcome, 16S rRNA sequencing was performed on the cervicovaginal secretion samples obtained from all participants prior to ALA-PDT. The results revealed that the alpha diversity of the cervicovaginal microbiota before ALA-PDT was comparable between the HPV clearance and non-clearance groups (Fig. 1a). In contrast, the beta-diversity analysis using principal coordinate analysis (PCoA) and Bray–Curtis distances demonstrated a significant difference between the two groups (P = 0.043) (Fig. 1b). To further identify the microbes contributing to the observed microbial compositional differences, a Venn diagram was constructed, indicating 978, 473, and 262 amplicon sequence variants (ASVs) unique to the HPV clearance group, HPV non-clearance group, and shared between both groups, respectively (Fig. 1c). Compared to the HPV clearance group, the HPV non-clearance group demonstrated a reduced abundance of Firmicutes_D, Bacteroidota, Fusobacteriota, and Firmicutes_A (In the SILVA Release 132 database, Firmicutes_A and Firmicutes_D are classified as distinct phylum-level clades due to phylogenetic differences, although both are part of the traditional phylum Firmicutes.) and an increased abundance of Actinobacteriota at the phylum level (Fig. 1d). At the genus level, the HPV non-clearance group exhibited lower abundances of Lactobacillus, Prevotella, Alloscardovia, Sneathia, and Enterococcus_H, and higher abundances of Bifidobacterium, Streptococcus, and Staphylococcus (Fig. 1e). Furthermore, LEfSe analysis identified three taxa, Anaerococcus, Gemella, and Gemellaceae, with significant differences in relative abundances between the two groups, which were increased in the HPV non-clearance group (Fig. 1f, g). Detailed LEfSe results are provided in Supplementary Table 4.

Fig. 1. Comparison of cervicovaginal microbiome before ALA-PDT between the HPV clearance and non-clearance groups.

Fig. 1

Open in a new tab

a, b Analysis of the diversity of cervicovaginal microbiota before ALA-PDT in the HPV clearance (N) and non-clearance (P) groups. a Alpha diversity boxplot (Chao1, Observed species, Simpson, Shannon, Faith’s PD, Pielou’s evenness, and Good’s coverage indexes) in the HPV clearance and non-clearance groups. b PCoA using Bray–Curtis distances of beta diversity. c–e Analysis of the composition of cervicovaginal microbiota before ALA-PDT in the HPV clearance and non-clearance groups. c Venn diagram showing numbers of exclusive and shared ASVs between the HPV clearance and non-clearance groups. d Taxonomic column diagram at the phylum level showing the top 20 most abundant identified phyla. e Taxonomic column diagram at the genus level showing the top 20 most abundant identified genera. f, g Distinctive bacterial species in the HPV clearance and non-clearance groups. f Histogram of LDA score among taxa with significant differences between the two groups. g Taxonomic cladistics of LDA Effect Size (LEfSe) analysis.

Comparison of cervicovaginal metabolome before ALA-PDT between the HPV clearance and non-clearance groups

Given the involvement of metabolic alterations in the interaction between the host and microbiota, global untargeted metabolomics was performed on the cervicovaginal secretion samples obtained before ALA-PDT. Partial Least Squares Discriminant Analysis (PLS-DA) was employed to visualize the distribution trends of metabolites, revealing differences in metabolite distribution between the HPV clearance and HPV non-clearance groups (Fig. 2a). The reliability of the PLS-DA model was validated through cross-validation and permutation testing (Fig. 2b). As illustrated in Fig. 2c, 187 metabolites exhibited significant differential expression between the two groups, including 81 upregulated metabolites and 106 down-regulated metabolites in the HPV non-clearance group compared to the HPV clearance group. Z-scores, derived from the relative content of metabolites, were used to illustrate the overall trend of changes and the degree of difference in metabolites between the two groups. The top 100 differential metabolites are presented in Fig. 2d. Subsequent KEGG pathway enrichment analysis of the differential metabolites revealed that the pathways of mineral absorption and valine, leucine, and isoleucine biosynthesis were significantly enriched in the HPV clearance group compared to the HPV non-clearance group (Fig. 2e, Supplementary Table 6). Detailed differential metabolites and KEGG pathway enrichment analysis results are provided in Supplementary Tables 5 and 6.

Fig. 2. Comparison of cervicovaginal metabolome before ALA-PDT between the HPV clearance and non-clearance groups.

Fig. 2

Open in a new tab

a Overall distribution trends of samples before ALA-PDT in the HPV clearance (N) and non-clearance (P) groups by the PLS-DA. b PLS-DA model evaluation by cross-validation and permutation test. c, d Differently expressed metabolites between the two groups. c Statistical histogram of differently expressed metabolites between the two groups (P vs N). d Z-score showing the overall change trend and difference degree of the top 100 differential metabolites. e KEGG pathway enrichment analysis of differential metabolites. The count of differential metabolites in the most enriched pathways and the impact of differential metabolites on the most enriched pathways were illustrated by a bar chart of the influence factors of the metabolic pathway. The horizontal axis shows impact values for various metabolic pathways, while the vertical axis lists these pathways. The numbers indicate the corresponding number of metabolites involved in the pathways. The color gradient reflects the P values, with red hues indicating smaller P values and blue hues indicating larger P values.

Impact of ALA-PDT on the composition and diversity of cervicovaginal microbiota

To investigate the impact of ALA-PDT on the cervicovaginal microbiome, 16S rRNA sequencing was performed on post-treatment cervicovaginal secretion samples, as well as self-matched baseline samples, obtained from 42 patients. The results indicated that ALA-PDT affected the alpha diversity of the cervicovaginal microbiota, particularly in terms of richness (Chao1 and Observed species indices) and coverage (Good’s coverage index) (Fig. 3a). Additionally, beta diversity analysis revealed a significant difference in the cervicovaginal microbiome before and after ALA-PDT (P = 0.035) (Fig. 3b). Subsequently, the differences in the composition of the cervicovaginal microbiota before and after ALA-PDT were further investigated. The Venn diagram revealed that 1237, 621, and 313 ASVs were unique to the post-treatment and pre-treatment groups and shared between both groups, respectively (Fig. 3c). At the phylum level, the abundance of Bacteroidota, Fusobacteriota, and Firmicutes_A was lower after ALA-PDT compared to before ALA-PDT. In contrast, the abundance of Actinobacteriota was higher than before ALA-PDT (Fig. 3d). At the genus level, the abundance of Prevotella, Streptococcus, and Sneathia decreased, while the abundance of Fannyhessea and Enterococcus_H increased after ALA-PDT (Fig. 3e). Furthermore, LEfSe analysis identified the taxa of Ezakiella, Actinomycetaceae, Leptotrichia_A, Lancefieldella, and Enterobacterales_A as having significant differences in relative abundances between the post-treatment and pre-treatment groups, all of which were significantly increased in the post-treatment group (Fig. 3f, g). Detailed LEfSe results are presented in Supplementary Table 7. These findings collectively suggest that ALA-PDT induces changes in the composition and diversity of the cervicovaginal microbiota.

Fig. 3. Impact of ALA-PDT on the composition and diversity of cervicovaginal microbiota.

Fig. 3

Open in a new tab

a, b Analysis of the diversity of cervicovaginal microbiota before and after ALA-PDT. a Alpha diversity boxplot (Chao1, Observed species, Simpson, Shannon, Faith’s PD, Pielou’s evenness, and Good’s coverage indexes) in the pre-treatment (A) and post-treatment (B) groups. b PCoA using Jaccard distances of beta diversity. c–e Analysis of the composition of cervicovaginal microbiota before and after ALA-PDT. c Venn diagram showing the numbers of exclusive and shared ASVs between the pre-treatment and post-treatment groups. d Taxonomic column diagram at the phylum level showing the top 20 most abundant identified phyla. e Taxonomic column diagram at the genus level showing the top 20 most abundant identified genera. f, g Distinctive bacterial species in the pre-treatment and post-treatment groups. f Histogram of LDA score among taxa with significant differences between the two groups. g Taxonomic cladistics of LDA Effect Size (LEfSe) analysis.

Impact of ALA-PDT on the metabolites in the cervicovaginal microenvironment

Global untargeted metabolomics was performed on the cervicovaginal secretion samples obtained from patients before and after ALA-PDT to examine the effect of ALA-PDT on the cervicovaginal metabolome. The results of Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) revealed a significant difference in the distribution of metabolites between cervicovaginal microenvironments before and after ALA-PDT (Fig. 4a). The reliability of the OPLS-DA model was validated through cross-validation and permutation testing (Fig. 4b). As illustrated in Fig. 4c, following ALA-PDT, the levels of 296 metabolites were significantly increased, while the expression of 502 metabolites was significantly decreased in the cervicovaginal microenvironment. Additionally, the overall trend of changes and degree of difference in metabolites within the cervicovaginal microenvironment before and after ALA-PDT is represented by Z-scores, with the top 100 differential metabolites displayed (Fig. 4d). KEGG enrichment analysis revealed that the pathways of transforming growth factor-β (TGF-β) signaling and arginine biosynthesis, among others, appeared to be altered following ALA-PDT (Fig. 4e). However, these alterations did not achieve statistical significance. The detailed findings of differential metabolites and KEGG pathway enrichment analysis are provided in Supplementary Tables 8 and 9, respectively. In conclusion, ALA-PDT notably influenced the distribution of metabolites within the cervicovaginal microenvironment.

Fig. 4. Impact of ALA-PDT on the metabolites in the cervicovaginal microenvironment.

Fig. 4

Open in a new tab

a Overall distribution trends of samples in the pre-treatment (A) and post-treatment (B) groups by the OPLS-DA. b OPLS-DA model evaluation by cross-validation and permutation test. c, d Differently expressed metabolites between the two groups. c Statistical histogram of differently expressed metabolites between the two groups (B vs A). d Z-score showing the overall change trend and difference degree of the top 100 differential metabolites. e KEGG pathway enrichment analysis of differential metabolites. The count of differential metabolites in the most enriched pathways and the impact of differential metabolites on the most enriched pathways were illustrated by a bar chart of the influencing factors of the metabolic pathway. The horizontal axis shows impact values for various metabolic pathways, while the vertical axis lists these pathways. The numbers indicate the corresponding number of metabolites involved in the pathways. The color gradient reflects the P values, with red hues indicating smaller P values and blue hues indicating larger P values.

Comparison of ALA-PDT’s impact on the composition and diversity of the cervicovaginal microbiota between the HPV clearance and non-clearance groups

The alterations in cervicovaginal microbiome during ALA-PDT were compared between the HPV clearance and non-clearance groups to investigate whether the different effects of ALA-PDT on the composition and diversity of cervicovaginal microbiota contribute to distinct HPV clearance outcomes. The results indicated that ALA-PDT influenced the alpha diversity of the cervicovaginal microbiota, particularly microbial coverage, in the HPV clearance group, while no such effect was observed in the HPV non-clearance group (Fig. 5a, Supplementary Fig. 1a). Furthermore, no differences were observed in the beta diversity of the cervicovaginal microbiota before and after ALA-PDT in either the HPV clearance or non-clearance group (P = 0.235 and P = 0.174) (Fig. 5b, Supplementary Fig. 1b). Moreover, analysis of the difference in the composition of cervicovaginal microbiota before and after ALA-PDT in the HPV clearance group revealed that 794, 424, and 228 ASVs were unique to the post-treatment group, pre-treatment group, and shared between both groups, respectively (Fig. 5c). At the phylum level, a reduction in the abundance of Bacteroidota, Fusobacteriota, and Firmicutes_A was observed following ALA-PDT, whereas the abundance of Actinobacteriota was increased (Fig. 5d). At the genus level, the abundance of Prevotella, Sneathia, Streptococcus, Dialister, and Parvimonas was reduced after ALA-PDT, while the abundance of Fannyhessea and Enterococcus_H was increased (Fig. 5e). LEfSe analysis identified significant post-treatment increases in the abundance of Finegoldia, Corynebacterium, Enterobacterales_A, Lancefieldella, and Winkia, whereas the abundance of Anaeroglobus was significantly decreased (Fig. 5f, g). Detailed LEfSe results are presented in Supplementary Table 10. These findings suggest that the HPV clearance outcomes in patients with cervical HSIL following ALA-PDT are associated with the regulation of cervicovaginal microbiota composition and diversity by ALA-PDT.

Fig. 5. ALA-PDT’s impact on the composition and diversity of the cervicovaginal microbiota in the HPV clearance group.

Fig. 5

Open in a new tab

a, b Analysis of the diversity of cervicovaginal microbiota before and after ALA-PDT in the HPV clearance group. a Alpha diversity boxplot (Chao1, Observed species, Simpson, Shannon, Faith’s PD, Pielou’s evenness, and Good’s coverage indexes) before (N_1) and after (N_2) ALA-PDT in the HPV clearance group. b PCoA using Jaccard distances of beta diversity. c–e Analysis of the composition of cervicovaginal microbiota before and after ALA-PDT in the HPV clearance group. c Venn diagram showing the numbers of exclusive and shared ASVs between the pre-treatment (N_1) and post-treatment (N_2) groups. d Taxonomic column diagram at the phylum level showing the top 20 most abundant identified phyla. e Taxonomic column diagram at the genus level showing the top 20 most abundant identified genera. f, g Distinctive bacterial species before and after ALA-PDT in the HPV clearance group. f Histogram of LDA score among taxa with significant differences between the pre-treatment and post-treatment groups. g Taxonomic cladistics of LDA Effect Size (LEfSe) analysis.

Comparison of ALA-PDT’s impact on the metabolites in the cervicovaginal microenvironment between the HPV clearance and non-clearance groups

ALA-PDT-associated alterations in cervicovaginal microenvironmental metabolite distributions were compared between the HPV clearance and non-clearance groups. OPLS-DA was performed, indicating significant differences in metabolite distributions in the cervicovaginal microenvironment pre- and post-ALA-PDT within HPV clearance and non-clearance groups (Fig. 6a, Supplementary Fig. 2a). The reliability of the OPLS-DA models for HPV clearance and non-clearance groups was validated through cross-validation and permutation testing (Fig. 6b, Supplementary Fig. 2b). As illustrated in Fig. 6c, following ALA-PDT, the levels of 247 metabolites were significantly increased in the HPV clearance group, while the expression of 407 metabolites was significantly decreased. In the HPV non-clearance group, 147 metabolites exhibited significantly increased expression following ALA-PDT, while the expression of 317 metabolites was significantly decreased. (Supplementary Fig. 2c). Comparatively, the HPV clearance group exhibited more metabolite changes before and after ALA-PDT than the HPV non-clearance group. Additionally, Z-values were employed to quantify the overall change trends and degrees of differences in differential metabolites in the two groups, with the top 100 differential metabolites presented in Fig. 6d and Supplementary Fig. 2d. KEGG pathway enrichment analyses indicated that the pathways of arginine biosynthesis, TGF-β signaling, and mammalian target of rapamycin (mTOR) signaling seemed to be more altered in the HPV clearance group compared to the HPV non-clearance group, with exhibiting a down-regulated trend (Fig. 6e, Supplementary Fig. 2e). Detailed results of differential metabolites and KEGG pathway enrichment analysis are provided in Supplementary Tables 11–14. These results collectively indicated that the regulation of ALA-PDT on metabolite distribution within the cervicovaginal microenvironment in patients with cervical HSIL differs between the HPV clearance and non-clearance groups. Moreover, the impact seems to be more pronounced in those who achieved HPV clearance following treatment.

Fig. 6. ALA-PDT’s impact on the metabolites in the cervicovaginal microenvironment in the HPV clearance group.

Fig. 6

Open in a new tab

a Overall distribution trends of samples in the HPV clearance group before (N_1) and after (N_2) ALA-PDT by the OPLS-DA. b OPLS-DA models evaluation by cross-validation and permutation test. c, d Differently expressed metabolites before and after ALA-PDT in the HPV clearance group. c Statistical histogram of differently expressed metabolites before and after ALA-PDT in the HPV clearance (N_2 vs N_1). d Z-score showing the overall change trend and difference degree of the top 100 differential metabolites in the HPV clearance group. e KEGG pathway enrichment analysis of differential metabolites in the HPV clearance group. The count of differential metabolites in the most enriched pathways and the impact of differential metabolites on the most enriched pathways were illustrated by a bar chart of the influencing factors of the metabolic pathway. The horizontal axis shows impact values for various metabolic pathways, while the vertical axis lists these pathways. The numbers indicate the corresponding number of metabolites involved in the pathways. The color gradient reflects the P values, with red hues indicating smaller P values and blue hues indicating larger P values.

Correlation analysis between altered cervicovaginal microbes and metabolites during ALA-PDT

A correlation analysis between the differential microbes and metabolites before and after ALA-PDT in the HPV clearance group was conducted. The findings indicated that several associations between microbes and metabolites achieved statistical significance (|r| > 0.3, FDR < 0.05). The overall correlation exhibited a moderate intensity and presented a complex pattern of both positive and negative correlations. Notably, a significant correlation was observed between the abundance of Anaeroglobus and the expression of several key differential metabolites, including phosphoribosyl formamidocarboxamide, 12-HETE, L-Leucine, Cytidine monophosphate, Xanthurenic acid, 8-Methoxykynurenate, 3,4-Dihydroxymandelaldehyde, 4-hydroxyphenylacetate, L-Glutamine, and D-Alanyl-D-alanine, among others (Fig. 7). This suggests that the alterations in metabolite expression and distribution within the cervicovaginal microenvironment following ALA-PDT in the HPV clearance group are closely associated with the abundance of Anaeroglobus. Detailed results of the correlation analysis are presented in Supplementary Table 15.

Fig. 7. Correlation analysis between altered cervicovaginal microbes and metabolites during ALA-PDT.

Fig. 7

Open in a new tab

Correlation analysis of the differential microbes and metabolites following ALA-PDT in the HPV clearance group. Red represents a positive correlation, and blue represents a negative correlation. (*FDR < 0.05; **FDR < 0.01).

Discussion

As a minimally invasive treatment, ALA-PDT has great potential in the clinical management of cervical HSIL. However, approximately 40% of the patients experience persistent HPV infection following ALA-PDT, which exacerbates psychological distress and elevates the risk of lesion recurrence4–6. Cervicovaginal microbiota has been indicated to be closely associated with the acquisition, persistence, and clearance of HPV, as well as the onset, progression, and treatment of CC and precancerous lesions11,17. Moreover, the metabolites produced by cervicovaginal microbiota play a crucial role in these processes. Accordingly, this study integrates metabolomics to investigate the impact of cervicovaginal microbiota on the efficacy of ALA-PDT in clearing HPV in patients with cervical HSIL.

A significant association was identified between elevated pre-treatment abundances of Anaerococcus and Gemella in cervicovaginal microbiota and persistent HPV infection following ALA-PDT in this study. Anaerococcus is recognized as one of the predominant anaerobic microorganisms implicated in bacterial vaginosis. Studies have reported that the bacteria associated with bacterial vaginitis can produce nitrosamines, which compromise the immune system by inducing cytokine release, facilitating the persistent HPV infection18,19. Furthermore, Gemella has been identified as negatively correlated with vaginal concentrations of immune mediators, including C-C motif chemokine ligand 2 (CCL2), CCL3, secretory leukocyte protease inhibitor, and C-X-C motif chemokine 10 (CXCL10)20. Our findings, supported by these studies, suggest that Anaerococcus and Gemella may serve as potential biomarkers for predicting the efficacy of ALA-PDT in clearing HPV in patients with cervical HSIL.

In this study, ALA-PDT exerted a significant influence on the diversity and composition of the cervicovaginal microbiota. Notably, the HPV clearance group exhibited more pronounced microbial diversity alterations during ALA-PDT than the HPV non-clearance group. These findings suggest a potential functional involvement of the cervicovaginal microbiota in mediating the HPV-clearing effects of ALA-PDT. Finegoldia, mainly Finegoldia magna, was identified as the most pronounced differential abundance in the cervicovaginal microenvironment before and after ALA-PDT within the HPV clearance group. Moreover, a significant post-ALA-PDT increase in the abundance of Finegoldia was observed. Previous research has indicated that Finegoldia is significantly more prevalent in the vaginal microbiota of HPV-negative women compared to HPV-positive women, which is consistent with our findings21. However, Finegoldia magna has been identified in patients with vaginosis and has been associated with high-risk HPV infection, as well as an increase in the odds of cervical intraepithelial neoplasia grade 2 or higher (CIN2+)22–24. These findings appear to contradict our results, indicating that Finegoldia is associated with HPV infection and clearance, but the specific relationship between them is rather complex and requires further investigation. A significant increase in the post-treatment abundance of Corynebacterium was also observed in the HPV clearance group. Prior research has demonstrated that Corynebacterium stimulates macrophages to secrete interferon-β (IFN-β) by activating the toll-like receptor 3 (TLR3) signaling pathway, thereby contributing to the defense against viral infections25. Moreover, Corynebacterium has been shown to enhance antibody production by promoting dendritic cell differentiation and upregulating a Th2-skewed immune response26. These observations are consistent with the findings of the present study. Lancefieldella, a member of the Atopobiaceae family, has been implicated in the pathogenesis of HPV infection, cervical dysplasia, and cancer27. Conversely, it has been demonstrated that Lancefieldella can stimulate a pro-inflammatory innate immune response, leading to the upregulation of interleukin (IL)-1β, IL-6, and IL-8 secretion, which can exert an antiviral effect28. In the present study, a positive correlation was observed between the abundance of cervicovaginal Lancefieldella and HPV clearance following ALA-PDT, providing further evidence for its beneficial role in HPV clearance. The abundance of Winkia was also significantly elevated following ALA-PDT in the HPV clearance group, while the abundance of Anaeroglobus, particularly Anaeroglobus vaginalis, was markedly decreased. Nevertheless, the specific roles of Winkia and Anaeroglobus in HPV infection and clearance remain unclear, and more systematic and in-depth studies are necessary.

The distribution of cervicovaginal metabolites have been shown to be associated with HPV infection and persistence, as well as with cervical lesions. This includes metabolites associated with amino acids and lipids, biogenic amines, organic acids and their derivatives, and TCA cycle-related metabolites, among others29–33. Additionally, studies have indicated that amino acid metabolism, carbohydrate metabolism, and lipid metabolism are all related to HPV clearance after treatment34,35. In this study, a large number of metabolites were also identified to be associated with HPV clearance mediated by ALA-PDT. In comparison to the HPV non-clearance group, the pathways of mineral absorption and valine, leucine, and isoleucine biosynthesis were significantly enriched in the HPV clearance group before ALA-PDT. Previous research has indicated that the pathway significantly enriched for mineral absorption is associated with HPV infection36. Furthermore, HPV-positive cell lines have been observed to exhibit upregulated biosynthesis of valine, leucine, and isoleucine compared to HPV-negative cell lines37. These observations seem to contradict our findings, suggesting that the pathways of mineral absorption and the biosynthesis of valine, leucine, and isoleucine may exhibit inverse relationships with HPV infection in the general population and with persistent HPV infection following treatment. Additionally, alterations in cervicovaginal metabolic profiles following ALA-PDT were also observed. Comparatively, the pathways of arginine biosynthesis, TGF-β signaling, and mTOR signaling exhibited a down-regulated trend following ALA-PDT in the HPV clearance group. Previous research has established a correlation between arginine metabolism and synthesis and the clearance of HPV after treatment35. The signaling pathways of TGF-β and mTOR have been implicated in enhancing viral replication and facilitating the progression of viral infections38–40. Nevertheless, the role of these pathways in ALA-PDT-mediated HPV clearance remains to be explored.

Additionally, the correlation analysis between the differential microbes and metabolites before and after ALA-PDT in the HPV clearance group revealed a significant positive correlation between the abundance of Anaeroglobus and the expression of phosphoribosyl formamidocarboxamide, 12-HETE, L-leucine, cytidine monophosphate, xanthurenic acid, 8-methoxykynurenate, 3,4-dihydroxymandelaldehyde, 4-hydroxyphenylacetate, L-glutamine, and D-alanyl-D-alanine. These metabolites are primarily involved in amino acid metabolism, purine metabolism, central carbon metabolism, pyrimidine metabolism, and arachidonic acid metabolism, all of which have been implicated in HPV infection and persistence35,41–43. However, it is unclear whether Anaeroglobus participates in ALA-PDT-mediated HPV clearance through regulating these metabolic pathways. Further investigation is necessary.

Our study does have some limitations. Firstly, the modest sample size, particularly the limited number of post-treatment samples and the imbalance between the HPV clearance and HPV non-clearance groups, may reduce statistical power and increase the risk of false positives in high-dimensional analyses. To address this, we employed permutation tests, cross-validation, FDR correction, and longitudinal analysis based on paired samples. Nevertheless, the findings should be considered exploratory and require validation in larger independent cohorts. Despite our efforts to control for confounders that could influence the research outcomes, certain potential confounders, such as dietary habits, were not fully controlled for. Additionally, this study demonstrated correlations between cervicovaginal microbiome and metabolome and ALA-PDT-mediated HPV clearance, but it does not establish direct causality. Consequently, to clarify the roles of the cervicovaginal microbiome and metabolome in ALA-PDT-mediated HPV clearance, a more comprehensive, systematic, and in-depth investigation is required.

This study comprehensively characterized the cervicovaginal microbiome and metabolome involved in ALA-PDT-mediated HPV clearance in patients with cervical HSIL, utilizing 16S rRNA sequencing and global untargeted metabolomics. These results identified multiple cervicovaginal microbes and metabolites associated with HPV clearance following ALA-PDT and preliminarily described the potential relationships between these cervicovaginal microbes and metabolites, offering valuable clues for predicting and improving the effect of HPV clearance following ALA-PDT. Furthermore, the findings contribute to a deeper understanding of ALA-PDT-mediated HPV clearance.

Methods

Study population and sample collection

A total of 74 patients diagnosed with cervical HSIL (CIN2, cervical intraepithelial neoplasia grade 2) who underwent topical 5-aminolevulinic acid PDT (ALA-PDT) at the First Affiliated Hospital of Zhengzhou University between April 2023 and October 2024 were included in this study. Among these patients, 24 exhibited persistent HPV infection following treatment. Cervicovaginal secretions were collected from all participants prior to the first treatment. Post-treatment samples of cervicovaginal secretions were collected 3 months after completion of therapy. The timing for post-treatment samples collection was established based on the following criteria: (1) The initial follow-up in the clinical implementation of ALA-PDT is scheduled between 3 and 6 months post-treatment; (2) Treatment outcomes are assessed using data from this initial follow-up; (3) Delaying the follow-up increases the risk of introducing confounding factors, such as the administration of interferons and antibiotics; (4) Previous studies have indicated no statistically significant differences in HPV clearance rates within the 3–6 month period post-treatment. Post-treatment samples were obtained from 42 of the 74 patients. Among them, 16 patients exhibited persistent HPV infection following treatment, while 26 patients achieved HPV clearance. Sample collection was conducted based on the following criteria: (1) Women with cervical HSIL (cervical intraepithelial neoplasia II) confirmed by cervical biopsy who were receiving treatment for the first time; (2) no systemic or vaginal topical antibiotic or antifungal therapy within 1 month before sampling; (3) no vaginal douching and administration within 48 h prior to sampling; (4) no sexual activity within 48 h before sample collection; (5) no treatment other than ALA-PDT during the study period. A detailed clinical trial flowchart is provided in Fig. 8. From the initiation of treatment until the first follow-up visit, the patients were advised to refrain from sexual activity as much as possible. In the event that sexual activity occurred, the use of a condom was necessary. All participants voluntarily agreed to participate in this study, and informed consent was obtained from each individual. This study was approved by the Ethics Committee of Zhengzhou University (approval number 2022-KY-1232).

Fig. 8.

Fig. 8

Open in a new tab

Clinical research flowchart.

Samples were collected from the upper third of the vaginal lateral wall, the cervical surface, and the cervical canal using three sterile swabs. These samples were pooled, and the procedure was repeated subsequently. All swab specimens were immediately stored at −80 °C for subsequent analysis.

HPV testing

DNA extracted from the swabs were tested for 21 types of HPV DNA (HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 53, 56, 58, 59, 66, 68, 26, 82, 73, 6, 11 or 81) using the fluorescence-based Bioperfectus Multiplex Real-Time (BMRT) Human Papillomavirus Genotyping Real-Time PCR Kit (Jiangsu Bioperfectus Technologies Co., Ltd., Taizhou, Jiangsu Province, China). The BMRT HPV kit was used according to the manufacturer’s instructions. Briefly, eight reactions per sample were performed simultaneously. Among them, the reactions A, B, C, D, E, F, and G were prepared to detect and differentiate in FAM™/ VIC® (HEX)/ROX™ fluorescent channels, HPV16/18/31, HPV59/66/53, HPV33/58/45, HPV56/52/35, HPV68/51/39, HPV82/26/73, and HPV6/11/81, respectively. In addition, an internal control (IC) with a reference gene, in reaction tube H (FAM™ channel), was set to identify possible PCR inhibition and to confirm the reliability of the reagents in this kit. PCR amplification was conducted in a total reaction volume of 20 µL, which comprised 2 µL DNA samples, 10 µL of PCR mix, and 8 µL of reaction mix. Positive and blank controls were also included in each specimen PCR detection. All reactions with UDG were pre-incubated at 50 °C for 5 min, followed by an initial denaturation at 95 °C for 10 min, and 45 cycles at 95 °C for 10 s, 58 °C for 40 s. PCR was performed on the Linegene 9600 real-time PCR instrument (Hangzhou Bioer Technology Co., Ltd., Hangzhou, Zhejiang Province, China). To validate the PCR reaction, the cycle threshold (Ct) values of positive controls had to be less than or equal to the cut-off value of 30.0 in FAM™, VIC® (HEX), and ROX™, and the blank controls had to be undetectable. For each of the 21 detected HPV genotypes, the qualitative reference values of positive cut-off were assessed by the manufacturer using ROC curves based on clinical trial results. Specimens with Ct values less than or equal to the cut-off value of one given HPV type were considered as positive for this HPV genotype. Conversely, specimens with Ct values above the cut-off value of one given HPV type were considered negative. Specimens with Ct values of the internal reference gene channel above the cut-off value of 36.7 were considered invalid.

Photodynamic therapy

The patients were placed in the lithotomy position. Prior to treatment, the vaginas and cervix were cleaned with sterile 0.9% sodium chloride. The lesions were completely covered with a thin cotton saturated with 20% ALA thermosensitive gel (Shanghai Fudan-Zhangjiang BioPharmaceutical Co., Ltd., Shanghai, China). After an incubation period of 3–4 h, irradiation of the cervical surface was performed for 30 min using light at a wavelength of 635 nm and a density of 80–120 J/cm2 (LED-IBS; Wuhan Yage PhotoElectronic Co. Ltd, Wuhan, China). The treatment was performed every 10–14 days for 6 times, and colposcopy was performed before each treatment to assess the size and scope of the lesions.

Follow-up and assessment of clinical efficacy

At 3 months post-treatment, HPV testing, liquid-based cytology tests (TCT), and colposcopy were conducted, during which any lesions identified as suspicious under colposcopy were subjected to biopsy for pathological evaluation. (1) Lesion: Complete remission of lesions was defined as the absence of abnormal pathology, as well as no suspicious lesions found by colposcopy. Partial remission was defined as degradation of the lesion level. Persistence was defined as a pathologic diagnosis of CIN2, whereas progression was defined as an upgrade of the lesion level. (2) HPV: Patients who tested positive for any of the baseline HPV types during HPV testing were classified as the HPV non-clearance group, whereas those who tested negative for all 21 HPV types were classified as the HPV clearance group.

DNA extraction and 16S rRNA gene sequencing

Total genomic DNA was extracted using the OMEGA Soil DNA kit (M5635-02; Omega Bio-Tek, Norcross, GA, USA), following the manufacturer’s instructions. The quantity of extracted DNA was assessed using a NanoDrop NC2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), while extracted DNA quality was evaluated via 1.2% agarose gel electrophoresis.

The V3–V4 hypervariable region of the bacterial 16S rRNA gene was amplified by polymerase chain reaction (PCR) using the forward primer 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and the reverse primer 806R (5′-GGACTACHVGGGTWTCTAAT-3′). Thermal cycling conditions included an initial denaturation at 98 °C for 5 min, followed by 25 cycles of denaturation at 98 °C for 30 s, annealing at 53 °C for 30 s, and extension at 72 °C for 45 s, with a final extension step at 72 °C for 5 min. PCR amplicons were purified using Vazyme VAHTSTM DNA Clean Beads (Vazyme, Nanjing, China) and quantified with the Quant-iT PicoGreen dsDNA assay kit (Invitrogen, Carlsbad, CA, USA). Following the individual quantification, amplicons were pooled in equal concentration, and pair-end 2 × 250 bp sequencing was conducted on the Illumina NovaSeq platform using NovaSeq 6000 SP reagent kit (500 cycles) at Suzhou PANOMIX Biomedical Tech Co., LTD.

Microbiome bioinformatics analysis

Microbiome bioinformatics analysis was performed using QIIME2 (version 2019.4) with slight modifications based on the official tutorials (https://docs.qiime2.org/2019.4/tutorials/). Briefly, raw sequence data were demultiplexed using the demux plugin, followed by primer cutting with the cutadapt plugin. Subsequently, sequences were subjected to quality filtering, denoising, merging, and chimera removal using the DADA2 plugin. Non-singleton ASVs were aligned using MAFFT software and utilized to construct a phylogenetic tree with FastTree2. Alpha-diversity metrics and beta-diversity metrics were estimated using the diversity plugin, with samples rarefied to 2851 sequences per sample. Taxonomic classification of ASVs was performed using the classify-sklearn naïve Bayes taxonomy classifier within the feature-classifier plugin, with reference to the SILVA Release 132 Database. Sequence data analyses were conducted using QIIME2 and R packages (version 3.2.0). Alpha diversity metrics at the ASV level, including Chao1 richness estimator, Observed species, Shannon diversity index, Simpson index, Faith’s phylogenetic diversity, Pielou’s evenness, and Good’s coverage, were calculated from the ASV table in QIIME2 and presented as box plots. The Kruskal-Wallis rank-sum test and Dunn’s test were employed to confirm the significance of the differences. Beta diversity analysis assessed the structural variation in microbial communities among samples using Jaccard, Bray–Curtis, and UniFrac distance metrics. These variations were visualized through PCoA, nonmetric multidimensional scaling, and unweighted pair-group method with arithmetic means hierarchical clustering. Principal component analysis (PCA) was also conducted based on the genus-level taxonomic composition. The significance of differentiation in microbiota structure among groups was evaluated using permutational multivariate analysis of variance, analysis of similarities, and permutation multivariate dispersion. Taxonomic compositions and abundances were visualized using MEGAN and GraPhlAn software. A Venn diagram was constructed using the R package “VennDiagram” to illustrate the shared and unique ASVs among samples or groups based on the ASV presence irrespective of relative abundance. Taxonomic abundances at the ASV levels were statistically compared among samples or groups using MetagenomeSeq, and the results were visualized in the form of Manhattan plots. Linear discriminant analysis effect size (LEfSe), an analytical approach that integrates the non-parametric Kruskal-Wallis test and the Wilcoxon rank-sum test with the effect size estimation from linear discriminant analysis, effectively identifies taxa with differential abundance across groups. Furthermore, OPLS-DA was applied as a supervised model to elucidate microbiota variation among groups, utilizing the R package “Muma.” Random forest analysis was conducted to differentiate the samples from distinct groups using QIIME2 with default settings. Nested stratified k-fold cross-validation was implemented for automated hyperparameter optimization and sample prediction. Microbial functional profiling was performed using phylogenetic investigation of communities by reconstruction of unobserved states based on the MetaCyc (https://metacyc.org/)and Kyoto encyclopedia of genes and genomes (KEGG, https://www.kegg.jp/) databases. The BioDeep online analysis platform (https://www.biodeep.cn) was also used to perform additional bioinformatic analyses.

Cervicovaginal community state types classification

Using 16S rRNA sequencing data, cervicovaginal CSTs classification was performed employing a custom script based on the following criterion: A sample was designated as CST I, II, III, or V if it exhibited a dominant species (>50% abundance) of L. crispatus, L. gasseri, L. iners, or L. jensenii, respectively. Conversely, samples lacking a dominant species or exhibiting a dominant anaerobic strain other than Lactobacillus were classified as CST IV.

Metabolite extraction and global untargeted metabolomic analysis

Samples were transferred into a centrifuge tube, and 2000 µL methanol containing 4 ppm of 2-amino-3-(2-chloro-phenyl)-propionic acid was added. The samples were vortexed for 30 s, followed by sonication at room temperature for 15 min. Subsequently, the samples were centrifuged for 10 min at 12,000 rpm and 4 °C. The supernatant was filtered using a 0.22 μm membrane and transferred into detection bottles for liquid chromatography (LC)-mass spectrum (MS) analysis.

The LC analysis was performed using a Vanquish Ultra High-Pressure Liquid Chromatography System (Thermo Fisher Scientific, USA). Chromatography was conducted using an ACQUITY UPLC ® HSS T3 column (2.1 × 100 mm, 1.8 µm; Waters, Milford, MA, USA), maintained at 40 °C. The flow rate and injection volume were set at 0.3 mL/min and 2 μL, respectively. For LC-electrospray ionization (ESI) positive-mode MS analysis, the mobile phases consisted of (B2) 0.1% formic acid in acetonitrile (v/v) and (A2) 0.1% formic acid in water (v/v). The separation was conducted using the following gradient: 0–1 min, 10% B2; 1–5 min, 10%–98% B2; 5–6.5 min, 98% B2; 6.5–6.6 min, 98%–10% B2; 6.6–8 min, 10% B2. For LC-ESI negative-mode MS analysis, the mobile phase consisted of B3 (acetonitrile) and A3 (5 mM ammonium formate). Separation was performed using the following gradient: 0–1 min, 10% B3; 1–5 min, 10%–98% B3; 5–6.5 min, 98% B3; 6.5–6.6 min, 98%–10% B3; 6.6–8 min, 10% B3.

Metabolite detection by MS was performed using an Orbitrap Exploris 120 (Thermo Fisher Scientific, USA) equipped with an ESI source. Simultaneous MS1 and MS/MS data acquisition was conducted in Full MS-ddMS2 mode (data-dependent MS/MS). Instrument parameters were as follows: sheath gas pressure, 40 arb; aux gas flow, 10 arb; spray voltage, 3.50 and −2.50 kV for positive and negative ESI modes, respectively; capillary temperature, 325 °C; MS1 scan range, m/z 100–1000; MS1 resolving power, 60,000 FWHM; number of data-dependent scans per cycle, 4; MS/MS resolving power, 15,000 FWHM; normalized collision energy, 30%; dynamic exclusion time, automatic.

Metabolomics data analysis

The metabolomics data analysis was performed by first converting the raw data to mzXML format using MSConvert in the ProteoWizard software package (version 3.0.8789). The data were processed using R XCMS software (version 3.12.0) for feature detection, retention time correction, and alignment. Key parameter settings were as follows: parts per million (ppm) = 15; peak width = c(5, 30); mzdiff = 0.01; and method = centWave. Batch effects were subsequently removed by correcting the data based on quality control (QC) samples. Metabolites with a relative standard deviation greater than 30% in QC samples were excluded and were not used in subsequent data analysis. The metabolites were identified by accurate mass and MS/MS data, which were compared against several databases, including Human Metabolome Database (http://www.hmdb.ca), MassBank (http://www.massbank.jp/), Kyoto Encyclopedia of Genes and Genomes (KEGG) (https://www.genome.jp/kegg/), LipidMaps (http://www.lipidmaps.org), mzCloud (https://www.mzcloud.org), and a metabolite database developed by Panomix Biomedical Tech Co., Ltd. (Suzhou, China). The molecular weight of metabolites was determined based on the m/z (mass-to-charge ratio) of the parent ions in MS data. The molecular formula was predicted using ppm and adduct ions and subsequently matched with the database. Simultaneously, MS/MS data from the quantitative table were compared with the fragment ions and other relevant information about each metabolite in the database to enable the MS/MS-based identification of metabolites. Two multivariate statistical analysis models, unsupervised and supervised, were applied to distinguish between groups (PCA, PLS-DA, and OPLS-DA) using the R ropls package (version 1.22.0). The reliability of the PLS-DA and OPLS-DA models was validated through cross-validation and permutation testing. The statistical significance of the P value was determined through statistical tests between groups. The Benjamini–Hochberg method was used for false positive correction. Finally, biomarker metabolites were identified by combining the P value and OPLS-DA variable importance in projection (VIP) and fold change (the ratio of differences between groups). By default, metabolites with a P value < 0.05 and VIP value > 1 were considered to exhibit significant differential expression. Differential metabolites were subsequently subjected to pathway analysis using MetaboAnalyst, which integrates results from powerful pathway enrichment and pathway topology analyses. The identified metabolites were mapped to the KEGG pathway for biological interpretation of higher-level systemic functions. The metabolites and their corresponding pathways were visualized using the KEGG Mapper tool.

Statistical analysis

Basic characteristics of the participants were analyzed using SPSS 27.0 software and were described using means, frequencies, and proportions. Comparisons of characteristics between different groups were conducted by Mann-Whitney U, Pearson χ2, or Fisher’s exact tests, as appropriate. All results were considered statistically significant at a P value less than 0.05. Correlation analysis was performed using Circlize V0.4.15. The microbial abundances were center log ratio (CLR) transformed to remove compositionality bias. Zero values were addressed using a multiplication substitution strategy, wherein each zero was replaced with half of the smallest non-zero value within each sample. This was followed by normalization and the calculation of CLR values. For metabolomics data, a log₂ (x + 1) transformation was employed to stabilize variance. Using the transformed data, Spearman rank correlation analysis was conducted to assess the associations between microbial species at the genus level and corresponding metabolites, with multiple testing corrections implemented via the Benjamini–Hochberg method. Results that achieved statistical significance (|r| > 0.3, FDR < 0.05) were subsequently visualized using heatmaps.

Supplementary information

Supplementary Information (811KB, pdf)
Supplementary Data Table 2 (14.5KB, xlsx)
Supplementary Data Table 4 (30.5KB, xls)
Supplementary Data Table 5 (36.8KB, xlsx)
Supplementary Data Table 6 (21.5KB, xlsx)
Supplementary Data Table 7 (32KB, xls)
Supplementary Data Table 8 (145.7KB, xlsx)
Supplementary Data Table 9 (21.8KB, xlsx)
Supplementary Data Table 10 (24.7KB, xls)
Supplementary Data Table 11 (141.2KB, xlsx)
Supplementary Data Table 12 (85.6KB, xlsx)
Supplementary Data Table 13 (29.6KB, xlsx)
Supplementary Data Table 14 (18KB, xlsx)
Supplementary Data Table 15 (218.7KB, xlsx)

Acknowledgements

This work was supported by Henan Provincial Science and Technology Research Project (Grant Number: 222102310196; 242102310283), Henan Provincial Key Scientific Research Project Plan of Higher Education Institutions (Grant Number: 25A320031), Henan Provincial Medical Science and Technology Joint Construction Project (Grant Number: LHGJ20220381), and Postdoctoral Research Foundation of the First Affiliated Hospital of Zhengzhou University. Additionally, the authors thank the patients who provided cervicovaginal secretion samples for this study.

Author contributions

This study was designed by T.Z., M.Z.Z. and Y.H.S. Samples and data were collected by C.Y.Z., S.Y.Y., Y.M.W., M.Z.Z., K.L.D., Y.D.Y. and L.J.Z. Funding was acquired by M.Z.Z., Y.H.S., and T.Z. Data were analyzed and interpreted by T.Z. The original draft of the manuscript was written by T.Z. Reviewing and editing of the manuscript was done by T.Z., M.Z.Z. and Y.H.S. All authors discussed the results and commented on the manuscript.

Data availability

Key raw data of 16S rRNA sequencing are available in the National Center for Biotechnology Information Sequence Read Archive (BioProject ID: PRJNA1267277; https://www.ncbi.nlm.nih.gov/sra/PRJNA1267277). Metabolomics data are available in the Open Archive for Miscellaneous Data of National Genomics Data Center (OMIX ID: OMIX010438; https://ngdc.cncb.ac.cn/omix/preview/C8hSqTig).

Code availability

All key analytical parameters, software versions, and methodological details related to 16S rRNA sequencing and global untargeted metabolomics have been comprehensively documented in the “Methods” section. No custom codes were used. All analyses relied on publicly available software with parameters as specified.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Mengzhen Zhang, Email: 13683806877@163.com.

Yuehui Su, Email: fccsuyh@zzu.edu.cn.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41522-026-01120-6.

References

  • 1.McCredie, M. R. et al. Natural history of cervical neoplasia and risk of invasive cancer in women with cervical intraepithelial neoplasia 3: a retrospective cohort study. Lancet Oncol.9, 425–434 (2008). [DOI] [PubMed] [Google Scholar]
  • 2.Castellsagué, X. Natural history and epidemiology of HPV infection and cervical cancer. Gynecol. Oncol.110, S4–S7 (2008). [DOI] [PubMed] [Google Scholar]
  • 3.Kyrgiou, M. et al. Obstetric outcomes after conservative treatment for cervical intraepithelial lesions and early invasive disease. Cochrane Database Syst. Rev.11, Cd012847 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Wang, B. et al. Comparative study of topical 5-aminolevulinic acid photodynamic therapy and surgery in treating vaginal high-grade squamous intraepithelial lesions following hysterectomy. Photodiagnosis Photodyn. Ther.45, 103921 (2024). [DOI] [PubMed] [Google Scholar]
  • 5.Wang, B. et al. The effect of local photodynamic therapy with 5-aminolevulinic acid in treating different grades of cervical intraepithelial neoplasia. Photodiagnosis Photodyn. Ther.40, 103196 (2022). [DOI] [PubMed] [Google Scholar]
  • 6.Tang, Y. et al. Therapeutic effects of topical photodynamic therapy with 5-aminolevulinic acid on cervical high-grade squamous intraepithelial lesions. Photodiagnosis Photodyn Ther.39, 102884 (2022). [DOI] [PubMed]
  • 7.Su, Y. et al. Effect and rational application of topical photodynamic therapy (PDT) with 5-aminolevulinic acid (5-ALA) for treatment of cervical intraepithelial neoplasia with vaginal intraepithelial neoplasia. Photodiagnosis Photodyn Ther.37, 102634 (2022). [DOI] [PubMed]
  • 8.Li, H. et al. The interaction between microorganisms, metabolites, and immune system in the female genital tract microenvironment. Front. Cell. Infect. Microbiol.10, 609488 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Kyrgiou, M. & Moscicki, A. B. Vaginal microbiome and cervical cancer. Semin. Cancer Biol.86, 189–198 (2022). [DOI] [PubMed] [Google Scholar]
  • 10.Chao, X. et al. Research of the potential biomarkers in vaginal microbiome for persistent high-risk human papillomavirus infection. Ann. Transl. Med.8, 100 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Łaniewski, P., Ilhan, Z. E. & Herbst-Kralovetz, M. M. The microbiome and gynaecological cancer development, prevention and therapy. Nat. Rev. Urol.17, 232–250 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Colbert, L. E. et al. The gut and cervical microbiome promote immune activation and response to chemoradiation in cervical cancer. Cancer Cell 10.2139/ssrn.3199993 (2018). [DOI]
  • 13.Zhang, H. et al. Cervical microbiome is altered in cervical intraepithelial neoplasia after loop electrosurgical excision procedure in China. Sci. Rep.8, 4923 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Janicka-Kośnik, M., Sarecka-Hujar, B., Jakiel, G. & Słabuszewska-Jóźwiak, A. Cervicovaginal microbiome after cervical intraepithelial neoplasia treatment. a protocol for systematic review and meta-analysis. Int. J. Environ. Res. Public Health18, 10.3390/ijerph18179050 (2021). [DOI] [PMC free article] [PubMed]
  • 15.Wang, W., Liu, Y., Yang, Y., Ren, J. & Zhou, H. Changes in vaginal microbiome after focused ultrasound treatment of high-risk human papillomavirus infection-related low-grade cervical lesions. BMC Infect. Dis.23, 3 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Jiang, L. et al. Influence of pelvic intensity-modulated radiation therapy with concurrent cisplatin-based chemotherapy of cervical cancer on the vaginal microbiome. Front. Oncol.11, 615439 (2021). [DOI] [PMC free article] [PubMed]
  • 17.Moscicki, A. B., Shi, B., Huang, H., Barnard, E. & Li, H. Cervical-vaginal microbiome and associated cytokine profiles in a prospective study of HPV 16 acquisition, persistence, and clearance. Front. Cell. Infect. Microbiol.10, 569022 (2020). [DOI] [PMC free article] [PubMed]
  • 18.Muzny, C. A., Łaniewski, P., Schwebke, J. R. & Herbst-Kralovetz, M. M. Host-vaginal microbiota interactions in the pathogenesis of bacterial vaginosis. Curr. Opin. Infect. Dis.33, 59–65 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Tidbury, F. D., Langhart, A., Weidlinger, S. & Stute, P. Non-antibiotic treatment of bacterial vaginosis—a systematic review. Arch. Gynecol. Obstet.303, 37–45 (2021). [DOI] [PubMed] [Google Scholar]
  • 20.Florova, V. et al. Vaginal host immune-microbiome interactions in a cohort of primarily African-American women who ultimately underwent spontaneous preterm birth or delivered at term. Cytokine137, 155316 (2021). [DOI] [PMC free article] [PubMed]
  • 21.Chao, X. P. et al. Correlation between the diversity of vaginal microbiota and the risk of high-risk human papillomavirus infection. Int. J. Gynecol. Cancer29, 28–34 (2019). [DOI] [PubMed] [Google Scholar]
  • 22.Murphy, E. C. & Frick, I. M. Gram-positive anaerobic cocci-commensals and opportunistic pathogens. FEMS Microbiol. Rev.37, 520–553 (2013). [DOI] [PubMed] [Google Scholar]
  • 23.Logel, M., El-Zein, M., Franco, E. L. & Gonzalez, E. Species-level characterization of the cervicovaginal microbiota and its role in human papillomavirus-associated cervical carcinogenesis. J. Med. Virol.96, e29764 (2024). [DOI] [PubMed]
  • 24.Logel, M., Tope, P., El-Zein, M., Gonzalez, E. & Franco, E. L. A narrative review of the putative etiologic role and diagnostic utility of the cervicovaginal microbiome in human papillomavirus-associated cervical carcinogenesis. J. Med. Virol.96, e70027 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Kanmani, P. et al. Respiratory commensal bacteria Corynebacterium pseudodiphtheriticum improves resistance of infant mice to respiratory syncytial virus and Streptococcus pneumoniae superinfection. Front. Microbiol.8, 1613 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Qingzhen, H. et al. Corynebacterium pyruviciproducens promotes the production of ovalbumin specific antibody via stimulating dendritic cell differentiation and up-regulating Th2 biased immune response. Vaccine30, 1115–1123 (2012). [DOI] [PubMed] [Google Scholar]
  • 27.Jimenez, N. R., Mancilla, V., Łaniewski, P. & Herbst-Kralovetz, M. M. Immunometabolic contributions of atopobiaceae family members in human papillomavirus infection, cervical dysplasia and cancer. J. Infect. Dis. 10.1093/infdis/jiae533 (2024). [DOI] [PMC free article] [PubMed]
  • 28.Maarsingh, J. D., Łaniewski, P. & Herbst-Kralovetz, M. M. Immunometabolic and potential tumor-promoting changes in 3D cervical cell models infected with bacterial vaginosis-associated bacteria. Commun. Biol.5, 725 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Ilhan, Z. E. et al. Deciphering the complex interplay between microbiota, HPV, inflammation and cancer through cervicovaginal metabolic profiling. EBioMedicine44, 675–690 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Borgogna, J. C. et al. The vaginal metabolome and microbiota of cervical HPV-positive and HPV-negative women: a cross-sectional analysis. BJOG127, 182–192 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Shen, S., Zhao, S., Shan, J. & Ren, Q. Metabolomic and microbiota profiles in cervicovaginal lavage fluid of women with high-risk human papillomavirus infection. Sci. Rep.15, 796 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Pu, X. et al. Microbial and metabolic profiles associated with HPV infection and cervical intraepithelial neoplasia: a multi-omics study. Microbiol. Spectr.13, e0019225 (2025). [DOI] [PMC free article] [PubMed]
  • 33.Zhao, S. et al. Vaginal microbiome and metabolome profiles among HPV positive and HPV negative women based on stratification of vaginitis. J. Med. Virol.97, e70385 (2025). [DOI] [PubMed]
  • 34.Li, X. et al. Targeted and untargeted metabolomics uncovering the effects of Jiawei Ermiao Granule on patients with persistent HPV infection. Front. Cell. Infect. Microbiol.15, 1550908 (2025). [DOI] [PMC free article] [PubMed]
  • 35.You, H. et al. Modulation of metabolic pathways and EndMT inhibition by a traditional Chinese herbal formula in the treatment of high-risk infections. Front. Cell. Infect. Microbiol.14, 1497965 (2024). [DOI] [PMC free article] [PubMed]
  • 36.Shen, Y., Zhu, H., Zhao, R. & Wang, H. Vaginal microbiota metabolome alteration in HPV-positive women. Eur. J. Obstet. Gynecol. Reprod. Biol.314, 114727 (2025). [DOI] [PubMed] [Google Scholar]
  • 37.Arizmendi-Izazaga, A. et al. Exometabolome and molecular signatures associated with HPV 16 in cervical cancer: integrative metabolomic and transcriptomic analysis for biomarker discovery. Molecules30, 10.3390/molecules30193909 (2025). [DOI] [PMC free article] [PubMed]
  • 38.Park, S. O., Kumar, M. & Gupta, S. TGF-β and iron differently alter HBV replication in human hepatocytes through TGF-β/BMP signaling and cellular microRNA expression. PLoS ONE7, e39276 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Rochette, L., Zeller, M., Cottin, Y. & Vergely, C. GDF15: an emerging modulator of immunity and a strategy in COVID-19 in association with iron metabolism. Trends Endocrinol. Metab.32, 875–889 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Tomé, D. Amino acid metabolism and signalling pathways: potential targets in the control of infection and immunity. Eur. J. Clin. Nutr.75, 1319–1327 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Jaiswal, S., Mishra, V., Majumder, S., Wangikar, P. P. & Sengupta, S. Metabolomic profiling reveals grade-specific niacinamide accumulation and its therapeutic potential via SIRT1-CD38-EMT axis modulation in cervical cancer progression. Biochim. Biophys. Acta Mol. Cell Res.1872, 119994 (2025). [DOI] [PMC free article] [PubMed]
  • 42.Liu, X. et al. Multiomics analysis of metabolic heterogeneity in cervical cancer cell lines with or without HPV. Front. Oncol.13, 1194462 (2023). [DOI] [PMC free article] [PubMed]
  • 43.Wang, Y. et al. Age-specific vaginal microecological dysbiosis associated with HPV infection: a large-scale cross-sectional study with targeted functional sequencing validation. Front. Cell. Infect. Microbiol.15, 1722367 (2025). [DOI] [PMC free article] [PubMed]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information (811KB, pdf)
Supplementary Data Table 2 (14.5KB, xlsx)
Supplementary Data Table 4 (30.5KB, xls)
Supplementary Data Table 5 (36.8KB, xlsx)
Supplementary Data Table 6 (21.5KB, xlsx)
Supplementary Data Table 7 (32KB, xls)
Supplementary Data Table 8 (145.7KB, xlsx)
Supplementary Data Table 9 (21.8KB, xlsx)
Supplementary Data Table 10 (24.7KB, xls)
Supplementary Data Table 11 (141.2KB, xlsx)
Supplementary Data Table 12 (85.6KB, xlsx)
Supplementary Data Table 13 (29.6KB, xlsx)
Supplementary Data Table 14 (18KB, xlsx)
Supplementary Data Table 15 (218.7KB, xlsx)

Data Availability Statement

Key raw data of 16S rRNA sequencing are available in the National Center for Biotechnology Information Sequence Read Archive (BioProject ID: PRJNA1267277; https://www.ncbi.nlm.nih.gov/sra/PRJNA1267277). Metabolomics data are available in the Open Archive for Miscellaneous Data of National Genomics Data Center (OMIX ID: OMIX010438; https://ngdc.cncb.ac.cn/omix/preview/C8hSqTig).

All key analytical parameters, software versions, and methodological details related to 16S rRNA sequencing and global untargeted metabolomics have been comprehensively documented in the “Methods” section. No custom codes were used. All analyses relied on publicly available software with parameters as specified.

Articles from NPJ Biofilms and Microbiomes are provided here courtesy of Nature Publishing Group

RESOURCES