Co-altered vaginal Lactobacillus, metabolome and host gene expression associate with the grade of cervical intraepithelial neoplasia in Chinese women

Co-altered vaginal Lactobacillus, metabolome and host gene expression associate with the grade of cervical intraepithelial neoplasia in Chinese women | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Co-altered vaginal Lactobacillus, metabolome and host gene expression associate with the grade of cervical intraepithelial neoplasia in Chinese women Wenkui Dai, Chunlei Guo, Xin Jiang, Yu Liu, Yinan Wang, Qian Zhou, and 9 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4717221/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Vaginal Lactobacillus has been implicated in modulating the risk of cervical intraepithelial neoplasia (CIN) progression. However, there remains a gap in population-based studies elucidating the underlying mechanisms that link Lactobacillus with CIN progression and carcinogenesis. Methods To address this knowledge gap, we conducted an in-depth analysis of vaginal microbiota (VM), metabolome, and host transcriptome profiles in a cohort of 75 Chinese women, stratified into two groups based on their CIN status: low-grade CIN1 (n = 38) and high-grade CIN2+ (n = 37). Results Our findings revealed that samples dominated by Lactobacillus were more prevalent in the CIN1 cohort. Furthermore, the vaginal metabolome displayed a significant interplay with the microbiota, with Lactobacillus emerging as a key influencer. Among the 100 metabolites that distinguished the CIN1 and CIN2 + cohorts, 26 were inversely correlated with Lactobacillus levels, including L-Carnitine and UDP-D-glucose. Conversely, five metabolites, such as Succinic anhydride, exhibited a positive correlation with Lactobacillus abundance. Differential gene expression analysis revealed 176 genes upregulated in the CIN1 cohort compared to the CIN2 + cohort, primarily related to immune responses and negative regulation of cell migration. Notably, COL4A2 and CCBE1, both negatively correlated with L-Carnitine, were among the upregulated genes. Conversely, 82 genes were downregulated in the CIN1 cohort, including TP63 and FOXD1, which positively correlated with UDP-D-glucose. Further mediation analysis suggested that L-Carnitine plays a crucial role in mediating the positive association between Lactobacillus and COL4A2 expression, both of which are enriched in the CIN1 cohort. Similarly, UDP-D-glucose emerged as a mediator in the negative association between Lactobacillus and FOXD1, a gene depleted in the CIN1 cohort. Conclusions These findings provide insights into the complex interplay between vaginal Lactobacillus , the metabolome, and host gene expression patterns associated with CIN progression. The identified Lactobacillus :L-Carnitine:COL4A2 and Lactobacillus :UDP-D-glucose:FOXD1 regulatory axes underscore the potential significance of these pathways in modulating CIN risk. These population-based discoveries hold promise for future research aimed at developing targeted interventions to prevent or delay CIN progression. Human papillomavirus Cervical intraepithelial neoplasia Vaginal microbiota Metabolome Host transcriptome Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Background Persistent human papillomavirus (HPV) infection poses a significant risk for cervical intraepithelial neoplasia (CIN) and cervical cancer [ 1 ]. However, the intricacies of factors modulating persistent HPV infection and CIN progression remain to be understood. Recent studies highlight the potential significance of vaginal microbiota (VM) in influencing these risks [ 2 – 11 ]. VM with dominant Lactobacillus has been demonstrated prevalent in healthy women [ 12 – 13 ]. Cross-sectional studies revealed a decrease in the level of vaginal Lactobacillus and variations in metabolites among HPV-positive women with high-grade CIN, compared to those with lower CIN grades [ 2 – 6 ]. Furthermore, prospective studies suggest causal links between vaginal Lactobacillus and the risk of persistent HPV infection as well as CIN progression [ 7 – 11 ]. Specifically, women with Lactobacillus crispatus -dominant VM exhibited a positive association with natural CIN regression, whereas those with non- Lactobacillus -dominant VM structures faced an increased risk of CIN progression [ 7 – 8 ]. The impact of VM on the risk of CIN progression is partially attributable to its modulation of host immune responses [ 14 – 18 ]. Recent investigations have uncovered associations between VM and various immune mediators, including a negative correlation between Lactobacillus and immune checkpoint inhibitors like LAG-3 and PD-L1 [ 19 – 20 ]. Certain vaginal Lactobacillus strains have been shown to modulate epithelial innate immune responses and activate Langerhans cells [ 15 , 21 ], which are vital in detecting HPV antigens and activating cellular immunity. Moreover, in vitro study has demonstrated the inhibitory effect of vaginal Lactobacillus on the growth of cancer-related Ect1/E6E7 and CaSKi cells, thereby influencing CIN progression and potentially carcinogenesis [ 15 ]. However, there is a paucity of population-based studies exploring the underlying mechanisms that link VM composition to the risk of CIN progression. In this study, we enrolled 75 HPV-positive Chinese women diagnosed with either low-grade or high-grade CIN. Our primary objectives were to elucidate the variations in the VM, metabolome, and host gene expression that are associated with CIN grade, and to assess the potential mediating role of metabolites in VM-host interactions. Our findings are expected to provide valuable population-based evidence for comprehending the influence of VM on the risk of CIN progression, thereby contributing to a deeper understanding of mechanisms underlying this complex relationship. Results Study population This study encompassed 75 HPV-positive women, who provided vaginal swabs and pap-brush samples (Fig. 1 , Table 1). Out of these, 149 vaginal swabs were applied for subsequent analysis, with 75 samples qualifying for VM examination and 74 for vaginal metabolome assessment. Additionally, 43 pap-brush samples were selected for host transcriptome analysis. For included HPV-positive women, there were no significant variations in the average age between the CIN1 and CIN2 + cohorts (Table 1). Most women in both groups were tested positive for HPV16/18-excluded high-risk HPV genotypes, with approximately 21% of women in both groups exhibiting HPV16 positivity. However, the proportion of HPV18-positive women was higher in the CIN1 group compared to the CIN2 + group (Table 1). Among the 75 women included, 4, 15, and 14 women had received the 2-valent, 4-valent, and 9-valent HPV vaccine respectively (Table 1). Furthermore, there were no significant differences observed in other phenotypes such as smoking between CIN1 and CIN2 + cohorts (Table 1). The increased proportion of Lactobacillus -dominant community state type (LD CST) in CIN1 cases compared to those in the CIN2 + cohort The LD CST VM groups comprised Lactobacillus crispatus -dominant (LCD, 16%), Lactobacillus iners -dominant (LID, 45.33%), Lactobacillus gasseri -dominant (LGD, 2.67%), Lactobacillus jensenii-dominant (LJD, 1.33%), and CSTs dominated by more than two Lactobacillus species (MixedLD, 16%) (Fig. S1 A). The VM groups categorized as non- Lactobacillus -dominant (NLD) CST accounted for 18.67% of the 75 microbial samples, exhibiting a heterogeneous community primarily consisting of Gardnerella , Fannyhessea or Streptococcus species (Fig. S1 A). At the phylum level, a major proportion of microbial samples in the CIN1 group (92.11%) were dominated by Firmicutes, whereas this dominance was less prevalent in the CIN2 + group, comprising 78.38% of samples (Fig. S1 B). Conversely, Actinobacteria displayed a lower prevalence in the CIN1 group (5.26%) and a significantly higher proportion in the CIN2 + group (21.62%) (Fig. S1 B). Similarly, we observed a decreased proportion of microbial samples with the NLD CST in the CIN1 group (13.16%) compared to the CIN2 + group (24.32%) (Fig. 2 A-D). Further analysis revealed a slightly higher proportion of LID and a lower proportion of LCD CST in the CIN1 group (LID: 50% vs 40.54% in the CIN2 + group; LCD: 13.16% vs 18.92% in the CIN2 + group) (Fig. 2 C-D). The association of altered vaginal metabolome with Lactobacillus Our Procrustes analysis robustly demonstrated a synergistic interplay between VM and metabolome profiles (Fig. 3 A). Employing Bray-Curtis dissimilarity as a metric, we observed significantly higher similarity within LD samples compared to those within NLD samples as well as those between LD and NLD samples (Fig. 3 B). The two-way orthogonal partial least squares (O2PLS) analysis showed notably higher loading value of Lactobacillus and associated D-lactic acid (Table S1 ). In addition, the dissimilarity among NLD samples mirrored that between LD and NLD groups, underscoring the substantial heterogeneity in metabolic profiles among microbial samples within the NLD CST (Fig. 3 B). To further decipher the metabolic determinants distinguishing CIN1 from CIN2 + cohorts, we implemented a random forest classifier to pinpoint metabolites that notably contributed to the differentiation. Given the high homogeneity of metabolic profiles in LD samples, subsequently we examined associations of those metabolites with vaginal Lactobacillus . Among the top 100 contributing metabolites, 31 exhibited a notable correlation (FDR < 0.1) with vaginal Lactobacillus levels (Fig. 3 C). Specifically, Lactobacillus demonstrated a negative correlation with 26 metabolites, spanning diverse chemical classes such as benzenoids (e.g., Phenylethylamine, Dezocine, Benzoylcholine, 2,4-Dichlorobenzoic acid), lipids and lipid-like molecules ((9,10)-Epoxyoctadecenoic acid, 11-Hydroxy-9-tridecenoic acid, Dodecylbenzenesulfonic acid, LysoPS 14:0), organic acids and derivatives (Thr-Lys, (S)-9-Hydroxy-10-undecenoic acid, N-Undecanoylglycine, 3-Methyl-2-oxovaleric acid), and others including 3-Methyl-1-butylamine, 2-Hydroxypyridine, L-Carnitine, UDP-D-glucose, 3-Dehydroxycarnitine, D-Mannose-6-phosphate, and beta-D-Glucosamine (Fig. 3 D, Table S2 ). Conversely, five metabolites—Indole-5,6-quinone, 4-Hydroxyphenylmaraviroc, Succinic anhydride, L-Kynurenine, belonging to organic oxygen and organoheterocyclic compounds—positively correlated with Lactobacillus levels (Fig. 3 D, Table S2 ). Among the remaining 69 metabolites that did not exhibit notable correlation with Lactobacillus levels, we observed distinct metabolic patterns between the CIN1 and CIN2 + cohorts. Specifically, the CIN1 cohort displayed elevated levels of several lipid and lipid-like molecules, including glycerophospholipids (e.g., PG 34:1, PG(16:0/18:1), PG 36:1; PG(18:0/18:1)), sterol lipids, fatty acyls (adipic acid, 11-Oxohexadecanoic acid, Ethyl hydrogen fumarate), and sphingolipids (Fig. 3 D, Table S2 ). Additionally, increased concentrations of benzene and substituted derivatives (Procaine, Cardanoldiene), organonitrogen compounds, keto acids and derivatives, and hydroxy acids and derivatives were observed in the CIN1 cohort (Fig. 3 D, Table S2 ). In contrast, the CIN1 cohort exhibited decreased levels of other lipid and lipid-like molecules, notably O-Acetyl-L-carnitine and D-(+)-Malic acid (Fig. 3 D, Table S2 ). Furthermore, several organic acids and derivatives, such as beta-Alanine, (S)-3,4-Dihydroxybutyric acid, and Homovanillic acid sulfate, were found to be downregulated in the CIN1 cohort compared to the CIN2 + cohort (Fig. 3 D, Table S2 ). Analysis also revealed decreased levels of organoheterocyclic compounds, particularly indoles and derivatives (e.g., Isorhynchophylline) and dihydrofurans (e.g., Norfuraneol, Furanone A) in the CIN1 cohort (Fig. 3 D, Table S2 ). Cervical transcriptomic differences between CIN1 and CIN2 + cohort Under the stringent threshold of FDR < 0.1, our analysis revealed a differential expression pattern between the CIN1 and CIN2 + cohorts, with 176 host genes upregulated and 82 genes downregulated in the CIN1 cohort (Fig. 4 A). Functional enrichment analysis of the upregulated genes in the CIN1 cohort predominantly associated them with immune responses, cell chemotaxis, negative regulation of cell migration, and B cell differentiation, activation, and proliferation (Fig. 4 B). Notably, several genes exhibiting ≥ 2-fold change in the CIN1 cohort were implicated in innate or adaptive immune responses, including CCR6, CXCR3, CXCR5, CD200, CD79A, CD79B, CD19, CDH6, CADM1, DEFB103A and DEFB103B (Fig. 4 B, Table S3 ). Additionally, CCBE1, IL36RN, SPON1, SERPINA5 and COL4A2 were among the other genes with ≥ 2-fold change enriched in the CIN1 cohort (Fig. 4 B, Table S3 ). In contrast, genes enriched in the CIN2 + cohort were primarily associated with epidermis development, keratinocyte differentiation, keratinization, and epidermal cell differentiation (Fig. 4 B, Table S3 ). Among these, genes with high overall expression levels included FABP5, CDKN2A (encoding p16 protein), TP63, KRT24, UPK3BL1, PHGDH, KCNS1, LGALS7B, and CALML5 (Fig. 4 B, Table S3 ). Random forest analysis pinpointed the importance of differentially-expressed genes(DEGs) to the differences observed between CIN1 and CIN2 + cohorts (Fig. 4 C). Among the top 15 contributors, KCNS1, KRT24 and FOXD1 exhibited significantly higher expression levels in the CIN2 + cohort (Fig. 4 C, Table S3 ). Conversely, among these 15 key contributors, 12 genes, including CCBE1, SERPINA5, SPON1, and COL4A2, were found to have higher expression levels in the CIN1 cohort (Fig. 4 C, Table S3 ). Although the statistical significance was not achieved (FDR > 0.1), we observed an increased expression trend of genes encoding other defensin and defensin-like molecules in the CIN1 + cohort, encompassing DEFB1, DEFB104A, DEFB4A, CAMP, PI3, S100A7, and SLPI (Figure S2 ). In contrast, within the CIN2 + cohorts, an upregulation of genes associated with p16 and Ki67 proteins, which serve as clinical biomarkers for CIN grading, was evident (Figure S2 ). Additionally, genes related to common inflammatory markers, such as interleukin (IL)-1α and macrophage inflammatory protein (MIP)-1α, exhibited elevated expression levels in the CIN2 + cohort (Figure S2 ). The correlation between vaginal Lactobacillus , altered metabolites and cervical gene expression To elucidate the intricate Lactobacillus -metabolite-host gene interactions associated with the CIN grade, we employed linear regression to quantify the correlations between Lactobacillus abundance, the top 100 metabolites contributing to inter-cohort differences, and DEGs. Besides to notable correlation with 31 metabolites (Fig. 3 C, Table S2 ), vaginal Lactobacillus levels correlated with 30 DEGs, including positive correlation with CIN1-enriched CCBE1, COL4A2 and SPON1 as well as negative correlation with CIN1-depleted FOXD1, TLX3 and IL36RN (Fig. 5 A, Table S4 ). Among the 31 metabolites significantly correlated with Lactobacillus (FDR < 0.1) (Table S2 ), all displayed robust associations with the expression of at least one host gene, representing a total of 239 DEGs (Fig. 5 A, Table S4 ). Our analysis revealed 1063 distinct metabolite-DEG correlations (Fig. 5 A, Table S4 ), with key metabolites such as (S)-9-Hydroxy-10-undecenoic acid, N-Undecanoylglycine, (9,10)-Epoxyoctadecenoic acid, UDP-D-glucose, L-Carnitine, Thr-Lys, Phenylethylamine, 3-Methyl-2-oxovaleric acid, Dodecylbenzenesulfonic acid, Dezocine, D-Mannose-6-phosphate, beta-D-Glucosamine, and Succinic anhydride predominating (Fig. 5 A, Table S4 ). Notably, these correlations involved DEGs primarily comprising COL4A2, SPON1, CCBE1, DEFB103A, and DEFB103B, which were upregulated in the CIN1 cohort (Fig. 5 A, Table S4 ). Additionally, genes enriched in the CIN2 + cohort, like FOXD1 and TLX3, were also implicated (Fig. 5 A, Table S4 ). Further mediation analysis has significantly enriched our understanding of the complex relationships between vaginal Lactobacillus and gene expression patterns, specifically identifying mediation effects of 15 metabolites in the associations between Lactobacillus and 18 DEGs. Importantly, the proportion of mediation effects was statistically significant, as evidenced by FDR < 0.1 and the lowest value of the 95% confidence interval (CI) exceeding zero (Fig. 5 B). Among these mediators, L-Cartinine stands out for its dual role. Negatively correlated with both Lactobacillus abundance and COL4A2 expression (Fig. 5 A, Table S2 ), L-Cartinine acts as a mediator in two opposing associations: it facilitates the positive relationship between Lactobacillus and COL4A2, which is enriched in CIN1-positive samples (Fig. 5 A-B, Table S3 ), while simultaneously mediating the negative association between Lactobacillus and TLX3, which is depleted in CIN1-positive samples (Fig. 5 A-B, Table S4 ). Similarly, UDP-D-glucose exhibits a negative correlation with Lactobacillus levels and is implicated in mediating the negative associations between Lactobacillus and four CIN1-depleted DEGs, notably FOXD1 (Fig. 5 A-B, Tables S2-3). Consistent with this, these DEGs positively correlate with UDP-D-glucose levels (Fig. 5 A, Table S4 ), reinforcing the mediating role of UDP-D-glucose. Additionally, three other metabolites—(S)-9-Hydroxy-10-undecenoic acid, N-Undecanoylglycine, and 9,10-Epoxyoctadecenoic acid—all negatively correlated with vaginal Lactobacillus abundance, and mediate the negative associations between Lactobacillus and their positively correlated DEGs, such as LGALS7B, TLX3, and GLS2 (Fig. 5 A-B, Table S4 ). Furthermore, nine additional metabolites, including 3,5-Dichlorobenzoic acid (neg.M189T37), 3-Dehydroxycarnitine (pos.M146T47), and others, play significant mediating roles in the relationships between vaginal Lactobacillus and 12 DEGs, notably GYS2 and TLX3 (Fig. 5 B). Discussion This study provides compelling evidence of co-alterations in vaginal Lactobacillus , the metabolome, and host gene expression, elucidating the influence of Lactobacillus on host gene expression patterns that correlate with the severity of cervical lesions. Consistent with previous reports, Lactobacillus -dominated VM types were found to be less prevalent among women with high-grade CIN [ 2 – 6 ], compared to those with low-grade CIN. Previous studies have attributed the beneficial effects of vaginal Lactobacillus to their production of D-lactic acid and hydrogen peroxide [ 22 – 25 ], and our study concurs with these findings, demonstrating a notable association between Lactobacillus and D-lactic acid levels. Furthermore, we identified significant correlations between vaginal Lactobacillus and a panel of metabolites that effectively distinguish between CIN1 and CIN2 + cohorts, partly aligning with prior research linking Lactobacillus with vaginal metabolites associated with HPV infection and CIN grade [ 2 , 26 – 27 ]. Extending our investigation to cervical gene expression, this population-based study delves into the transcriptomic profiles of both CIN1 and CIN2 + cohorts. Elevated expression of the p16-encoding CDKN2A gene in the CIN2 + cohort corroborates our pathological findings and the clinical consensus, which highlights positive p16 staining as an indicator of high-grade CIN [ 28 – 29 ]. This validates the feasibility of RNA-Seq analysis using cervical exfoliated cells as a diagnostic tool. Moreover, we discovered that genes depleted in the CIN2 + cohort are primarily associated with immune responses and negative regulation of cell migration, echoing previous reports that implicate altered cell migration and compromised immunity in HPV-infected high-grade CIN and cervical cancer [ 30 – 31 ]. Additional findings, including decreased expression of defensins and defensin-like molecules, alongside increased inflammatory markers in the CIN2 + cohort, are consistent with multiple cohort studies [ 2 , 19 – 20 , 32 – 34 ], reinforcing the robustness of our observations. Further in-depth analysis revealed intriguing correlations among vaginal Lactobacillus populations, metabolites that exhibited high discriminatory power between the two cohorts, and DEGs. Notably, Lactobacillus was inversely correlated with several metabolites, including 9,10-Epoxyoctadecenoic acid, L-Carnitine, N-Undecanoylglycine, (S)-9-Hydroxy-10-undecenoic acid, and UDP-D-glucose, which, in turn, showed positive associations with the oncogene TP63. These findings partially corroborate previous studies emphasizing the protective role of vaginal Lactobacillus in hindering the progression of CIN [ 5 , 7 – 8 ]. In contrast, N-Undecanoylglycine and 9,10-Epoxyoctadecenoic acid exhibited significant negative correlations with the expression of genes (CD19, CD79A, CD79B, CD200, and CCR6) that are enriched in CIN1 and play crucial roles in both innate and adaptive immunity [ 35 – 38 ]. This observation suggests a potential mechanism whereby alterations in these metabolites may impact immune function in the context of CIN. Additionally, L-Carnitine was negatively correlated with the expression of DEFB103A and DEFB103B, which encode defensins, integral components of the innate immune barrier in the female lower genital tract. Notably, decreased expression of these defensins has been reported in high-grade CIN [ 32 , 39 – 41 ], further implicating L-Carnitine in the modulation of innate immune responses during CIN progression. Furthermore, our analysis revealed a notable positive correlation between UDP-D-glucose and the FOXD1 gene. Subsequent analysis demonstrated a significant mediation effect of UDP-D-glucose in the negative association observed between vaginal Lactobacillus and FOXD1 gene expression. Given the emerging evidence indicating the pro-tumorigenic role of FOXD1 in cancer cell proliferation, cancer progression, and poor prognosis [ 42 – 45 ], this finding suggests that vaginal Lactobacillus may impede the progression of CIN by inhibiting FOXD1 expression through downregulation of UDP-D-glucose. While the precise mechanisms underlying this mediation remain elusive, recent studies have highlighted the ability of vaginal Lactobacillus to inhibit the growth of certain vaginal pathogens [ 46 – 47 ]. This, in turn, may lead to a reduction in the production of pathogen-derived UDP-D-glucose and its product, UDP-D-glucuronic acid (UDP-GlcA) [ 48 – 49 ]. Although there is currently no direct evidence linking UDP-GlcA to CIN progression, previous studies have shown that excessive UDP-GlcA can disrupt Golgi function, thereby compromising cellular signaling capabilities [ 50 ]. Moreover, other researchers have demonstrated a positive role for UDP-GlcA in promoting tumor cell migration and cancer metastasis [ 51 – 52 ]. To fully elucidate the role of UDP-D-glucose and UDP-GlcA in the Lactobacillus -mediated inhibition of FOXD1 and CIN progression, future studies are warranted to investigate the underlying molecular mechanisms and validate these findings in larger, well-controlled cohorts. Our study also uncovered a notable mediation effect of L-Carnitine in the positive correlation observed between Lactobacillus and COL4A2 gene expression, suggesting that Lactobacillus positively regulates COL4A2 expression by downregulating the production of L-Carnitine. COL4A2 encodes type IV collagen, which is a major structural component of basement membranes and plays a pivotal role in maintaining the integrity of the basal membrane [ 53 ]. This finding aligns partially with previous research demonstrating elevated levels of type IV collagen in the cervix of individuals with lower grades of cervical lesions [ 54 – 58 ], indicating its potential involvement in the prevention of cervical pathology. Intriguingly, Amiloride has been shown to suppress the migratory and invasive capabilities of HeLa cells by inhibiting the degradation of type IV collagen, achieved through downregulation of type IV collagenase expression [ 59 ]. Additionally, the C-terminal fragment of type IV collagen, known as canstatin, has been identified as a potent inhibitor of tumor growth [ 53 , 60 – 61 ], underscoring the multifaceted role of COL4A2 and its products in modulating cancer progression. Collectively, these findings highlight the significance of COL4A2 expression in cervical epithelial cells, emphasizing its potential role in preventing the progression of CIN. Limitations of the present study warrant acknowledgment. Firstly, the small sample size and the use of cross-sectional multi-omic profiling pose limitations in drawing definitive conclusions. However, these challenges were partially mitigated by the application of rigorous inclusion criteria, specifically excluding participants with common vaginal infections and menopause, which are known to significantly influence VM [ 12 , 62 – 63 ]. Secondly, the utilization of 16S rDNA amplicon sequencing, while informative, constrained our ability to fully comprehend the functional communities within the VM [ 12 , 62 , 64 ]. Consequently, we were unable to delve into intra-group heterogeneity and inter-group variations among vaginal Lactobacillus at the strain level and their functional landscapes. This underscores the need for future studies incorporating more comprehensive sequencing approaches, such as shotgun metagenomics, to address these gaps. Lastly, although our analysis revealed Lactobacillus -metabolite-host linkages that correlated with CIN grade, establishing causality remains an area for further investigation. The observed associations between UDP-D-glucose and L-Carnitine mediating the influence of Lactobacillus on FOXD1 and COL4A2 gene expression, respectively, are intriguing but preliminary. Conclusions Overall, this study presents a compelling case for co-alterations in vaginal Lactobacillus , metabolites, and host gene expression patterns that are associated with CIN grade. Our findings hint at potential roles of vaginal Lactobacillus in modulating the risk of CIN progression, particularly through the mediation of UDP-D-glucose and L-Carnitine on FOXD1 and COL4A2 expression. Nevertheless, further research is imperative to unravel the intricate mechanisms underlying these associations and mediation effects, including longitudinal studies and intervention trials to confirm causality and elucidate the full extent of their clinical implications. Materials and methods Study design A total of seventy-five HPV-positive women with varying grades of CIN were recruited through interviews conducted at the Department of Obstetrics and Gynecology of Peking University Shenzhen Hospital in China. Adhering to the enrollment criteria outlined in our previous study [ 65 ], participants were required to be over 18 years of age, not experiencing menopause, and without a history of cervical ablation, resection surgery, hysterectomy, pelvic radiotherapy, sexual intercourse within three days, douching, or vaginal medication use within seven days, antibiotic exposure within one month, hormone replacement therapy or GnRH-a use within three months, and no autoimmune diseases or HIV infection. Additionally, vaginal smears were tested negative for H 2 O 2 , leukocyte esterase, neuraminidase, b-glucuronidase, and acetylaminoglucosidase to exclude common genital infections. All participants were fully informed and provided written consents prior to the enrollment. Data and sample collection With the aid of a research assistant, the clinician diligently recorded crucial metadata for each participant. This included information on smoking habits, sexual activity history, contraceptive use, gestation and abortion status, HPV genotypes, and vaccination records. To ensure the homogeneity of the samples, vaginal swabs were collected ≥ 3 days following menstruation. During colposcopy, a single physician expertly sampled the posterior fornix for vaginal swabs. All swabs were promptly preserved in a 2 ml sterile tube kept on ice and transferred to -80℃ storage within 30 minutes of collection. For the host transcriptome analysis, cervical exfoliated cells were gently collected using Pap brushes. To maintain RNA integrity, the same skilled technician immediately conducted RNA extraction using a specialized kit (Bacteria RNA Extraction Kit, Vazyme Biotech Co., Ltd China) following sample collection. This ensured the quality and reliability of the data for subsequent analysis. HPV testing and pathology diagnosis For HPV screening, we employed the Roche Cobas®4800-HPV system, encompassing specific detection of HPV16/18 and 12 additional high-risk HPV genotypes (HPV 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68). The diagnosis of CIN was performed by two seasoned pathologists, utilizing p16 immunohistochemical staining as an auxiliary tool to differentiate between CIN1 and CIN2 + lesions. Specifically, cytologically diagnosed CIN2 samples that exhibited p16 positivity were categorized as CIN2+; otherwise, they were classified as CIN1 [ 66 – 67 ]. Microbial DNA extraction, 16S rRNA gene amplicon sequencing and data processing Microbial DNA of vaginal swabs was extracted using the Dneasy PowerSoil Pro Kit (Qiagen, Germany). The concentration and purity of the extracted DNA were assessed via 1% agarose gels on an Agilent5400 platform (Agilent Technologies, Inc., Santa Clara, USA). For library construction, we amplified the 16S rDNA V4-V5 hypervariable regions using specific primers: 515-FR (GTGCCAGCMG CCGCGGTAA) and 926-RR (CCGTCAATTCMTTTRAGTTT). The resultant DNA libraries were sequenced on the Illumina NovaSeq platform (Illumina, San Diego, CA, United States), with a read length of 250 bp. The raw sequencing data underwent rigorous analysis using QIIME2 [ 68 ] for the generation of VM profiles. VM CST analysis We utilized unsupervised hierarchical clustering with average Euclidean linkage to identify distinct VM groups from all 75 women. The nomenclature for each microbiome group was determined based on whether the microbiome profiles exhibited a dominant bacterial genus or species with the abundance of ≥ 50%. For the LD CST that lacked a dominant bacterial genus or species, the microbiome group was specifically designated as MixedLD, indicating its dominance by more than two Lactobacillus species. Untargeted metabolomics analysis To characterize the vaginal metabolome, we employed liquid chromatography-mass spectrometry (LC-MS) utilizing the Thermo Fisher Scientific platform (Ottawa, United States). Raw intensity data were converted to the mzXML format, and ion features were extracted utilizing the Progenesis QI software (v.2.2). These ions underwent rigorous filtering, excluding those missing in over 50% of quality control samples or more than 80% of test samples, as well as those with a relative standard deviation exceeding 30%. To identify the metabolites, we searched the Human Metabolome Database (HMDB, v.5.0) and Kyoto Encyclopedia of Genes and Genomes (KEGG, v.96.0). The resultant metabolite abundance matrix provided the cornerstone for subsequent analysis. RNA-Seq and data processing For RNA sequencing (RNA-Seq), total RNA was extracted from cervical exfoliated cells using the Qiagen RNease Mini Kit. RNA-Seq was performed on the Illumina NovaSeq platform. The raw sequencing reads were first filtered with Cutadapt (v.2.5) [ 69 ] to remove low-quality and adapter sequences. Subsequently, these reads were aligned to the human genome (GRCh38) using Hisat2 (v.2.1.0) [ 70 ]. Gene expression levels were quantified and normalized using RSEM (v.1.3.3) [ 71 ] and DESeq2 (v.1.20) [ 72 ], respectively. Then the DESeq2 (v.1.20) software was applied to analyze the differences of host gene expression between CIN1 and CIN2 + cohort [ 72 ]. Finally, the DAVID Knowledgebase (v2023q4) [ 73 ] was utilized to identify enriched functional pathways for DEGs. Bioinformatics and statistics The synergy between the VM and its associated metabolome was meticulously evaluated through the application of Procrustes analysis, implemented in the vegan package of the R software environment. Subsequently, a random forest classifier, consisting of 100 trees and employing leave-one-out cross-validation, was harnessed to assess the pivotal role of individual metabolites in distinguishing the CIN1 from CIN2 + cohort, utilizing the mlr package in R. Prior to conducting the association analysis, all continuous variables were standardized to conform to a standard normal distribution (N~(0, 1)) by employing an empirical normal quantile transformation. Subsequently, we employed linear models to systematically examine the associations between Lactobacillus abundance and the top 100 metabolites distinguishing the CIN1 and CIN2 + cohort, as well as the relationships between these key metabolites and differentially expressed host genes. The linear models were formulated as follows: Metabolite ~ Lactobacillus + Age Gene ~ Metabolite + Age The mediation effects along the Lactobacillus –metabolite–host axis were thoroughly examined using the R mediate package, with the differentially expressed host genes serving as continuous dependent variables, and age included as the covariate. The mediation analysis entailed fitting two linear models: Metabolite i = α 1 + β 1 Lacto + δ 1 T X i + ε i1 Gene i = α 2 + β 2 Lacto + γ 1 Metabolite i + δ 2 T X i + ε i2 where Lacto denotes the relative abundance of vaginal Lactobacillus . Furthermore, Metabolite i signifies each metabolite in the top 100 metabolite differentiating CIN1 from CIN2 + cohort. Additionally, Xi represents a vector of covariates, while Gene i stands for each differentially expressed host gene, also with a significance threshold of FDR < 0.1. After fitting these two modules, the product of two coefficients β 1 γ 1 was interpreted as an estimate of average causal mediated effect and the coefficient β 2 was interpreted as an estimate of the average direct effect. Significant mediation effects were determined based on an FDR < 0.1 for the average mediation effect, with the lowest value of the 95% confidence interval (CI) exceeding zero. All statistical analyses and data visualization were executed using R software (version 4.0.5), adhering to the established significance threshold of FDR < 0.1 for statistical significance. Declarations Ethics approval and consent to participate This study was approved by the Ethics Committee of Peking University Shenzhen Hospital (registration number: 2022-157). All participants were fully informed and then provided signed consent. Consent for publication Not applicable Availability of data and material The raw data of VM, metabolome and host transcriptome analysis were accessible in CNGB Sequence Archive (CNSA) under Project No. CNP0005859. Other data supporting the findings of this study are available from the corresponding authors on reasonable request. Competing interests There is no competing interests to be declared. Funding This study was supported by the National Natural Science Foundation of China (82202826), Shenzhen High-level Hospital Construction Fund (YBH2019-260), Shenzhen Key Medical Discipline Construction Fund (SZXK027), Sanming Project of Medicine in Shenzhen (SZSM202011016), Shenzhen Public Platform for Preservation of Fertility and Reproduction (XMHT20220104049) and Peking University Shenzhen Hospital Scientific Research Fund (KYQD2021075 and KYQD2022132). Authors' contributions W.R., L.S. and D.H. conceived the study. D.H., J.H., J.X.,L.Y. and L.C. recruited and selected attenders. G.C., W.D. and Y.Qing performed the sample collection and associated experimental analysis. D.W. performed data analysis and drafted the manuscript. G.C., J.X., W.Y., Z.Q., X.R. and Y.Qin helped with the data analysis. L.S. polished the manuscript. References Lei J, Ploner A, Elfström KM, Wang J, Roth A, Fang F, et al. HPV Vaccination and the Risk of Invasive Cervical Cancer. N Engl J Med. 2020;383(14):1340-1348. Ilhan ZE, Łaniewski P, Thomas N, Roe DJ, Chase DM, Herbst-Kralovetz MM. Deciphering the complex interplay between microbiota, HPV, inflammation and cancer through cervicovaginal metabolic profiling. EBioMedicine. 2019;44:675-690. Guo C, Dai W, Zhou Q, Gui L, Cai H, Wu D, et al. Cervicovaginal microbiota significantly changed for HPV-positive women with high-grade squamous intraepithelial lesion. Front Cell Infect Microbiol. 2022;12:973875. Chen Y, Qiu X, Wang W, Li D, Wu A, Hong Z, et al. Human papillomavirus infection and cervical intraepithelial neoplasia progression are associated with increased vaginal microbiome diversity in a Chinese cohort. BMC Infect Dis. 2020;20(1):629. Mitra A, MacIntyre DA, Lee YS, Smith A, Marchesi JR, Lehne B, et al. Cervical intraepithelial neoplasia disease progression is associated with increased vaginal microbiome diversity. Sci Rep. 2015;5:16865. Zheng JJ, Song JH, Yu CX, Wang F, Wang PC, Meng JW. Difference in vaginal microecology, local immunity and HPV infection among childbearing-age women with different degrees of cervical lesions in Inner Mongolia. BMC Womens Health. 2019;19(1):109. Mitra A, MacIntyre DA, Ntritsos G, Smith A, Tsilidis KK, Marchesi JR, et al. The vaginal microbiota associates with the regression of untreated cervical intraepithelial neoplasia 2 lesions. Nat Commun. 2020;11(1):1999. Usyk M, Zolnik CP, Castle PE, Porras C, Herrero R, Gradissimo A, et al. Cervicovaginal microbiome and natural history of HPV in a longitudinal study. PLoS Pathog. 2020;16(3):e1008376. Dareng EO, Ma B, Adebamowo SN, Famooto A, Ravel J, Pharoah PP, et al. Vaginal microbiota diversity and paucity of Lactobacillus species are associated with persistent hrHPV infection in HIV negative but not in HIV positive women. Sci Rep. 2020;10(1):19095. Di Paola M, Sani C, Clemente AM, Iossa A, Perissi E, Castronovo G, et al. Characterization of cervico-vaginal microbiota in women developing persistent high-risk Human Papillomavirus infection. Sci Rep. 2017;7(1):10200. Berggrund M, Gustavsson I, Aarnio R, Lindberg JH, Sanner K, Wikström I, et al. Temporal changes in the vaginal microbiota in self-samples and its association with persistent HPV16 infection and CIN2. Virol J. 2020;17(1):147. Holm JB, France MT, Gajer P, Ma B, Brotman RM, Shardell M, et al. Integrating compositional and functional content to describe vaginal microbiomes in health and disease. Microbiome. 2023;11(1):259. Farr Zuend C, Lamont A, Noel-Romas L, Knodel S, Birse K, Kratzer K, et al. Increased genital mucosal cytokines in Canadian women associate with higher antigen-presenting cells, inflammatory metabolites, epithelial barrier disruption, and the depletion of L. crispatus. Microbiome. 2023;11(1):159. Mahajan G, Doherty E, To T, Sutherland A, Grant J, Junaid A, et al. Vaginal microbiome-host interactions modeled in a human vagina-on-a-chip. Microbiome. 2022;10(1):201. Gao Q, Fan T, Luo S, Zheng J, Zhang L, Cao L, et al. Lactobacillus gasseri LGV03 isolated from the cervico-vagina of HPV-cleared women modulates epithelial innate immune responses and suppresses the growth of HPV-positive human cervical cancer cells. Transl Oncol. 2023;35:101714. Breedveld AC, Schuster HJ, van Houdt R, Painter RC, Mebius RE, van der Veer C, et al. Enhanced IgA coating of bacteria in women with Lactobacillus crispatus-dominated vaginal microbiota. Microbiome. 2022;10(1):15. Berard AR, Brubaker DK, Birse K, Lamont A, Mackelprang RD, Noël-Romas L, et al. Vaginal epithelial dysfunction is mediated by the microbiome, metabolome, and mTOR signaling. Cell Rep. 2023;42(5):112474. Delgado-Diaz DJ, Jesaveluk B, Hayward JA, Tyssen D, Alisoltani A, Potgieter M, et al. Lactic acid from vaginal microbiota enhances cervicovaginal epithelial barrier integrity by promoting tight junction protein expression. Microbiome. 2022;10(1):141. Łaniewski P, Cui H, Roe DJ, Chase DM, Herbst-Kralovetz MM. Vaginal microbiota, genital inflammation, and neoplasia impact immune checkpoint protein profiles in the cervicovaginal microenvironment. NPJ Precis Oncol. 2020;4:22. Łaniewski P, Barnes D, Goulder A, Cui H, Roe DJ, Chase DM, et al. Linking cervicovaginal immune signatures, HPV and microbiota composition in cervical carcinogenesis in non-Hispanic and Hispanic women. Sci Rep. 2018;8(1):7593. Song J, Lang F, Zhao N, Guo Y, Zhang H. Vaginal Lactobacilli Induce Differentiation of Monocytic Precursors Toward Langerhans-like Cells: in Vitro Evidence. Front Immunol. 2018;9:2437. Breshears LM, Edwards VL, Ravel J, Peterson ML. Lactobacillus Crispatus Inhibits Growth of Gardnerella Vaginalis and Neisseria Gonorrhoeae on a Porcine Vaginal Mucosa Model. BMC Microbiol. 2015; 15:276. Witkin SS, Mendes-Soares H, Linhares IM, Jayaram A, Ledger WJ, Forney LJ. Influence of Vaginal Bacteria and D- and L-Lactic Acid Isomers on Vaginal Extracellular Matrix Metalloproteinase Inducer: Implications for Protection Against Upper Genital Tract Infections. mBio. 2013; 4(4): e00460–13. Zarate G, Nader-Macias ME. Influence of Probiotic Vaginal Lactobacilli on In Vitro Adhesion of Urogenital Pathogens to Vaginal Epithelial Cells. Lett Appl Microbiol. 2006; 43(2):174–180. Vallor AC, Antonio MA, Hawes SE, Hillier SL. Factors Associated With Acquisition of, or Persistent Colonization by, Vaginal Lactobacilli: Role of Hydrogen Peroxide Production. J Infect Dis. 2001; 184(11):1431–1436. Borgogna JC, Shardell MD, Santori EK, Nelson TM, Rath JM, Glover ED, et al. The vaginal metabolome and microbiota of cervical HPV-positive and HPV-negative women: a cross-sectional analysis. BJOG. 2020;127(2):182-192. Xu H, Liu M, Song Y, Liu L, Xu F, Chen J, et al. Metabolomics variation profiling of vaginal discharge identifies potential targets for cervical cancer early warning. Acta Biochim Biophys Sin (Shanghai). 2022;54(10):1561-1565. Kurman RJ, Carcangiu ML, Herrington CS, Young RH. Tumours of the Cervix: squamous Cell Tumours and Precursors. In: Kurman JR, Carcangiu ML, Herrington CS, Young RH, eds. World Health Organization Classification of Tumours of the Female Reproductive Organs. 4th ed. Lyon: IARC Press. 2014; 172–182. Carozzi F, Gillio-Tos A, Confortini M, Del Mistro A, Sani C, De Marco L, et al. Risk of high-grade cervical intraepithelial neoplasia during follow-up in HPV-positive women according to baseline p16-INK4A results: a prospective analysis of a nested substudy of the NTCC randomised controlled trial. Lancet Oncol. 2013; 14(2):168-176. Wang Q, Song R, Zhao C, Liu H, Yang Y, Gu S, et al. HPV16 E6 promotes cervical cancer cell migration and invasion by downregulation of NHERF1. Int J Cancer. 2019;144(7):1619-1632. Lu Y, Chen Y, Zhang Z, Li M, Chen X, Tu K, et al. HPV16 E6 promotes cell proliferation, migration, and invasion of human cervical cancer cells by elevating both EMT and stemness characteristics. Cell Biol Int. 2022;46(4):599-610. Lebeau A, Bruyere D, Roncarati P, Peixoto P, Hervouet E, Cobraiville G, et al. HPV infection alters vaginal microbiome through down-regulating host mucosal innate peptides used by Lactobacilli as amino acid sources. Nat Commun. 2022;13(1):1076. Łaniewski P, Cui H, Roe DJ, Barnes D, Goulder A, Monk BJ, et al. Features of the cervicovaginal microenvironment drive cancer biomarker signatures in patients across cervical carcinogenesis. Sci Rep. 2019;9(1):7333. Audirac-Chalifour A, Torres-Poveda K, Bahena-Román M, Téllez-Sosa J, Martínez-Barnetche J, Cortina-Ceballos B, et al. Cervical Microbiome and Cytokine Profile at Various Stages of Cervical Cancer: A Pilot Study. PLoS One. 2016;11(4):e0153274. Li X, Ding Y, Zi M, Sun L, Zhang W, Chen S, et al. CD19, from bench to bedside. Immunol Lett. 2017;183:86-95. Harton J, Jin L, Hahn A, Drake J. Immunological Functions of the Membrane Proximal Region of MHC Class II Molecules. F1000Res. 2016;5:F1000 Faculty Rev-368. Shao A, Owens DM. The immunoregulatory protein CD200 as a potentially lucrative yet elusive target for cancer therapy. Oncotarget. 2023;14:96-103. Gómez-Melero S, Caballero-Villarraso J. CCR6 as a Potential Target for Therapeutic Antibodies for the Treatment of Inflammatory Diseases. Antibodies (Basel). 2023;12(2):30. Hubert P, Herman L, Maillard C, Caberg JH, Nikkels A, Pierard G, et al. Defensins induce the recruitment of dendritic cells in cervical human papillomavirus-associated (pre)neoplastic lesions formed in vitro and transplanted in vivo. FASEB J. 2007;21(11):2765-75. Hubert P, Herman L, Roncarati P, Maillard C, Renoux V, Demoulin S, et al. Altered α-defensin 5 expression in cervical squamocolumnar junction: implication in the formation of a viral/tumour-permissive microenvironment. J Pathol. 2014;234(4):464-77. Wolf R, Howard OM, Dong HF, Voscopoulos C, Boeshans K, Winston J, et al. Chemotactic activity of S100A7 (Psoriasin) is mediated by the receptor for advanced glycation end products and potentiates inflammation with highly homologous but functionally distinct S100A15. J Immunol. 2008;181(2):1499-506. Cai K, Chen S, Zhu C, Li L, Yu C, He Z, et al. FOXD1 facilitates pancreatic cancer cell proliferation, invasion, and metastasis by regulating GLUT1-mediated aerobic glycolysis. Cell Death Dis. 2022;13(9):765. Kumegawa K, Yang L, Miyata K, Maruyama R. FOXD1 is associated with poor outcome and maintains tumor-promoting enhancer-gene programs in basal-like breast cancer. Front Oncol. 2023;13:1156111. Prasad P, Kannan B, Sriram G, Jaber M, Khair AMB, Ramasubramanian A, et al. Waterpipe smoke condensate induces epithelial-mesenchymal transformation and promotes metastasis of oral cancer by FOXD1 expression. J Stomatol Oral Maxillofac Surg. 2024;101900. Liu X, Min S, Zhang Q, Liu Y, Zou Z, Wang N, et al. Prognostic and clinicopathological significance of FOXD1 in various cancers: a meta and bioinformation analysis. Future Sci OA. 2024;10(1):FSO901. Liu Z, Zhang Z, Qiu L, Zhang F, Xu X, Wei H, et al. Characterization and bioactivities of the exopolysaccharide from a probiotic strain of Lacto- bacillus plantarum WLPL04. J Dairy Sci. 2017;100:6895–6905. Allonsius CN, Vandenheuvel D, Oerlemans EFM, Petrova MI, Donders GGG, Cos P, et al. Inhibition of Candida albicans morphogenesis by chitinase from Lactobacillus rhamnosus GG. Sci Rep. 2019;9:2900. Petrucca A, Cipriani P, Sessa R, Teggi A, Pustorino R, Santapaola D, et al. Burkholderia cenocepacia vaginal infection in patient with smoldering myeloma and chronic hepatitis C. Emerg Infect Dis. 2004;10(11):1957-9. Amabebe E, Richardson LS, Bento GFC, Radnaa E, Kechichian T, Menon R, et al. Ureaplasma parvum infection induces inflammatory changes in vaginal epithelial cells independent of sialidase. Mol Biol Rep. 2023;50(4):3035-3043. Doshi MB, Lee N, Tseyang T, Ponomarova O, Goel HL, Spears M, et al. Disruption of sugar nucleotide clearance is a therapeutic vulnerability of cancer cells. Nature. 2023;623(7987):625-632. Wang X, Liu R, Zhu W, Chu H, Yu H, Wei P, et al. UDP-glucose accelerates SNAI1 mRNA decay and impairs lung cancer metastasis. Nature. 2019;571(7763):127-131. Gao Q, Cheng B, Chen C, Lei C, Lin X, Nie D, et al. Dysregulated glucuronic acid metabolism exacerbates hepatocellular carcinoma progression and metastasis through the TGFβ signalling pathway. Clin Transl Med. 2022;12(8):e995. Okada M, Yamawaki H. A current perspective of canstatin, a fragment of type IV collagen alpha 2 chain. J Pharmacol Sci. 2019;139(2):59-64. Sorvari T, Sarnesto A, Syrjänen K. Type IV collagen in the basal membrane of human papillomavirus associated premalignant and malignant squamous cell lesions of the uterine cervix. Gynecol Obstet Invest. 1988;26(4):324-31. Stewart CJ, McNicol AM. Distribution of type IV collagen immunoreactivity to assess questionable early stromal invasion. J Clin Pathol. 1992;45(1):9-15. Pitt MA, Hale RJ, Buckley CH. The distribution of type IV collagen in invasive carcinoma of the uterine cervix. Histopathology. 1992;20(2):139-43. Hirota J, Yoneda K, Osaki T. Basement membrane type IV collagen in oral squamous cell carcinoma. Head Neck. 1990;12(5):400-5. Tong CCL, Koptyra M, Raman P, Rathi KS, Choudhari N, Lin X, et al. Targeted gene expression profiling of inverted papilloma and squamous cell carcinoma. Int Forum Allergy Rhinol. 2022;12(2):200-209. Wang X, Jiang Z, An J, Mao X, Lin F, Sun P. Effect of a synthetic inhibitor of urokinase plasminogen activator on the migration and invasion of human cervical cancer cells in vitro. Mol Med Rep. 2018;17(3):4273-4280. Ma Y, Wu T, Zhou H, He G, Li Y, Wang B, et al. Canstatin represses glioma growth by inhibiting formation of VM-like structures. Transl Neurosci. 2021;12(1):309-319. Zhang F, Duan J, Song H, Yang L, Zhou M, Wang X. Combination of canstatin and arsenic trioxide suppresses the development of hepatocellular carcinoma. Drug Dev Res. 2021;82(3):430-439. France MT, Fu L, Rutt L, Yang H, Humphrys MS, Narina S, et al. Insight into the ecology of vaginal bacteria through integrative analyses of metagenomic and metatranscriptomic data. Genome Biol. 2022;23(1):66. Brotman RM, Shardell MD, Gajer P, Fadrosh D, Chang K, Silver MI, et al. Association between the vaginal microbiota, menopause status, and signs of vulvovaginal atrophy. Menopause. 2018;25(11):1321-1330. Ma B, France MT, Crabtree J, Holm JB, Humphrys MS, Brotman RM, et al. A comprehensive non-redundant gene catalog reveals extensive within-community intraspecies diversity in the human vagina. Nat Commun. 2020;11(1):940. Dai W, Du H, Zhou Q, Li S, Wang Y, Hou J, et al. Metabolic profiles outperform the microbiota in assessing the response of vaginal microenvironments to the changed state of HPV infection. NPJ Biofilms Microbiomes. 2024;10(1):26. Waxman AG, Chelmow D, Darragh TM, Lawson H, Moscicki AB. Revised terminology for cervical histopathology and its implications for management of high-grade squamous intraepithelial lesions of the cervix. Obstet Gynecol. 2012;120(6):1465-1471. Darragh TM, Colgan TJ, Cox JT, Heller DS, Henry MR, Luff RD, et al. The Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions: background and consensus recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. J Low Genit Tract Dis. 2012;16(3):205-42. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37(8):852-857. Marcel M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 2011;17:10–12. Kim D, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with low memory requirements. Nat Methods. 2015;12(4):357-60. Li B, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12:323. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. Sherman BT, Hao M, Qiu J, Jiao X, Baseler MW, Lane HC, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50(W1):W216-W221. Table Table 1 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigures.pdf TableS1.csv TableS2.csv TableS3.csv TableS4.csv Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4717221","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":328259522,"identity":"958cf100-1bab-4fcf-9024-1f374dc8ede1","order_by":0,"name":"Wenkui Dai","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Wenkui","middleName":"","lastName":"Dai","suffix":""},{"id":328259523,"identity":"24dbe6c3-095c-46f1-9697-1f8c08114568","order_by":1,"name":"Chunlei Guo","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chunlei","middleName":"","lastName":"Guo","suffix":""},{"id":328259524,"identity":"918d61b5-6ca5-4120-973b-f9c237cdcfeb","order_by":2,"name":"Xin Jiang","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Jiang","suffix":""},{"id":328259525,"identity":"4831112a-c968-4933-be7f-c54bcc02a2d8","order_by":3,"name":"Yu Liu","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Liu","suffix":""},{"id":328259526,"identity":"662739d0-af5d-4a1a-9319-cd6f96d37571","order_by":4,"name":"Yinan Wang","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yinan","middleName":"","lastName":"Wang","suffix":""},{"id":328259527,"identity":"c7be3e76-e90c-4d2b-b755-49b2fc3d6a67","order_by":5,"name":"Qian Zhou","email":"","orcid":"","institution":"City University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Zhou","suffix":""},{"id":328259528,"identity":"5876099d-4d67-483d-a664-2fcb26d39519","order_by":6,"name":"Ruinan Xu","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ruinan","middleName":"","lastName":"Xu","suffix":""},{"id":328259529,"identity":"c1fca9ae-0395-483a-a317-76b7743cdd6a","order_by":7,"name":"Jun Hou","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Hou","suffix":""},{"id":328259530,"identity":"88967f73-ea64-40cf-a66a-48ba27d975b6","order_by":8,"name":"Di Wu","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Di","middleName":"","lastName":"Wu","suffix":""},{"id":328259531,"identity":"ded8f475-9a94-47c8-b864-2cd92d28d51a","order_by":9,"name":"Qing Yang","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qing","middleName":"","lastName":"Yang","suffix":""},{"id":328259532,"identity":"c03f3d62-6b7f-45ff-b7f2-f1bdaf95578b","order_by":10,"name":"Qin Yang","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qin","middleName":"","lastName":"Yang","suffix":""},{"id":328259533,"identity":"d034045c-0e40-49d9-a526-349262fcdc36","order_by":11,"name":"Changzhong Li","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Changzhong","middleName":"","lastName":"Li","suffix":""},{"id":328259534,"identity":"1e4021fc-e1ae-4d7f-8bce-857bb7723d04","order_by":12,"name":"Shuai Cheng Li","email":"","orcid":"","institution":"City University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Shuai","middleName":"Cheng","lastName":"Li","suffix":""},{"id":328259535,"identity":"a88827c5-b164-4011-bf9d-3d6aed783b01","order_by":13,"name":"Hui Du","email":"","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Du","suffix":""},{"id":328259536,"identity":"d54ac429-e198-4b48-826d-5b4ca348a21a","order_by":14,"name":"Ruifang Wu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuUlEQVRIiWNgGAWjYJCCD0CcwM/MfPABsToYZ4C0SLazJRuQpsXgPI+ZAFHqDW7kPmwubKvLMz7MYMbAUGMTTYSWdMPmmW2Hi80OM6Q9YDiWlttASIvZjTT2x7xtBxK3HWY4bsDYcJgoLYzNvG11iZubGdskSNHCnLiBmZmNOC32Z54xNs84dzhxxmE2ZoMEYvwi2Q60paCsLrG///zHBx9qbAhrAQFmOCuBGOWoWkbBKBgFo2AUYAMAtTxAeTokmy0AAAAASUVORK5CYII=","orcid":"","institution":"Peking University Shenzhen Hospital","correspondingAuthor":true,"prefix":"","firstName":"Ruifang","middleName":"","lastName":"Wu","suffix":""}],"badges":[],"createdAt":"2024-07-10 10:09:28","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4717221/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4717221/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":62271714,"identity":"e2616e14-6ac9-4e12-8a90-1faa1a422b40","added_by":"auto","created_at":"2024-08-12 10:23:35","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":100247,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eWorkflow of this study\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/23d0ec4a653062e27309ce55.png"},{"id":62272368,"identity":"85cb5179-6191-46fb-bad4-9c3cdb73f2e9","added_by":"auto","created_at":"2024-08-12 10:31:35","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":74403,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVM differences between CIN1 and CIN2+ cohort. A.\u003c/strong\u003e Distribution of genus-level Community State Type (CST_G) for each group. This panel depicts the distribution of genus-level community state types (CST_G) in both CIN1 and CIN2+ cohorts. The \u003cem\u003eLactobacillus\u003c/em\u003e-dominant (LD) and non-\u003cem\u003eLactobacillus\u003c/em\u003e-dominant (NLD) CSTs are clearly distinguished, highlighting the differences in microbial composition between the two groups.\u003cstrong\u003e B. \u003c/strong\u003eGenus-level compositions for 75 microbial samples.\u003cstrong\u003eC.\u003c/strong\u003e Distribution of species-level Community State Type (CST_S) for each group. This panel shows the distribution of species-level community state types (CST_S) in the CIN1 and CIN2+ cohorts. Distinct CSTs are identified, including \u003cem\u003eLactobacillus crispatus\u003c/em\u003e-dominant (LCD), \u003cem\u003eLactobacillus gasseri\u003c/em\u003e-dominant (LGD), \u003cem\u003eLactobacillus iners\u003c/em\u003e-dominant (LID), \u003cem\u003eLactobacillus jensenii\u003c/em\u003e-dominant (LJD), and MixedLD (dominated by more than two \u003cem\u003eLactobacillus\u003c/em\u003e species), revealing subtle differences in microbial ecology between the two groups. \u003cstrong\u003eD.\u003c/strong\u003eSpecies-level compositions for 75 microbial samples\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/fd9dc55263cc3144bdfe99d1.png"},{"id":62272370,"identity":"c618a9cd-f680-452f-a283-cd918d1c2566","added_by":"auto","created_at":"2024-08-12 10:31:35","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":154065,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe association of altered vaginal metabolome with \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eLactobacillus \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003elevels. A. \u003c/strong\u003eThis Procrustes analysis visually compares the structure of the VM (represented by blue squares) with that of the vaginal metabolome (represented by orange squares). Black lines connect microbial and metabolic samples from the same individual, illustrating the potential interplay between the microbial community and its metabolic products. \u003cstrong\u003eB. \u003c/strong\u003eBray-Curtis dissimilarities based on metabolic profiles. This panel assesses the Bray-Curtis dissimilarities among individuals grouped by their \u003cem\u003eLactobacillus\u003c/em\u003e-dominant (LD) or non-\u003cem\u003eLactobacillus\u003c/em\u003e-dominant (NLD) CSTs, as well as among individuals with distinct CSTs, based on their metabolic profiles. The results provide insights into how microbial composition influences metabolic variability. \u003cstrong\u003eC. \u003c/strong\u003eRandom forest analysis of metabolites distinguishing CIN1 from CIN2+ cohort. The Random Forest analysis identifies the top 100 metabolites that best distinguish the CIN1 cohort from the CIN2+ cohort. The blue and orange bars represent negative and positive correlations, respectively, between the abundance of these metabolites and \u003cem\u003eLactobacillus\u003c/em\u003elevels. Gray bars indicate no significant association. The Y-axis displays the Gini index, a measure of the variable's importance in the classification. \u003cstrong\u003eD. \u003c/strong\u003eHeatmap of normalized abundance of top 100 metabolites. This heatmap visually represents the normalized abundance of the top 100 metabolites that differentiate between the CIN1 and CIN2+ cohorts. The intensity of color indicates the relative abundance of each metabolite, offering a comprehensive overview of the metabolic differences between the two groups.\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/e2dbae40b4b3e47c6860f6ba.png"},{"id":62271710,"identity":"9b102d47-a393-4711-9e1b-5711c7c8f61b","added_by":"auto","created_at":"2024-08-12 10:23:35","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":97363,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCervical transcriptomic differences between CIN1 and CIN2+ cohort.\u003c/strong\u003e \u003cstrong\u003eA.\u003c/strong\u003e This plot displays 258 genes with significant variation (FDR\u0026lt;0.1) between CIN1 and CIN2+ cohorts, emphasizing the magnitude and statistical significance of expression changes. \u003cstrong\u003eB. \u003c/strong\u003eRandom forest analysis assesses the relative importance of 258 differentially-expressed genes (DEGs, FDR\u0026lt;0.1) in distinguishing CIN1 from CIN2+ cohorts. \u003cstrong\u003eC.\u003c/strong\u003e The heatmap depicts the normalized expression levels of 258 DEGs (FDR\u0026lt;0.1), alongside functional enrichments associated with these genes, offering a comprehensive view of their biological relevance.\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/976b04929910fd15c8a9978e.png"},{"id":62271718,"identity":"a0b9fecb-6304-4b05-86b4-2b86603b8f8d","added_by":"auto","created_at":"2024-08-12 10:23:36","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":678056,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eLactobacillus\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e-metabolite-host gene correlation associated with the CIN grade. A.\u003c/strong\u003e This network illustrates the relationships between vaginal \u003cem\u003eLactobacillus\u003c/em\u003elevels, metabolites that significantly differentiate CIN1 and CIN2+ cohorts, and 239 DEGs (FDR\u0026lt;0.1). Red and blue lines signify positive and negative correlations, respectively, revealing potential regulatory interactions. \u003cstrong\u003eB. \u003c/strong\u003eThis analysis explores the mediation effects of metabolites in the influence of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e abundance on host gene expression patterns, shedding light on the complex interplay between microbial, metabolic, and host genetic factors in CIN progression.\u003c/p\u003e","description":"","filename":"Fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/838077a389a9d135ef9958a8.png"},{"id":75114874,"identity":"630b3964-731c-471b-91ae-f6f7b03d0a1c","added_by":"auto","created_at":"2025-01-30 16:01:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2142010,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/7ea6576f-59a0-42fc-8e89-a1d95cf10d0e.pdf"},{"id":62271715,"identity":"c3e20f8f-c5ce-467b-b5c1-7fc0e6d4bfac","added_by":"auto","created_at":"2024-08-12 10:23:35","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":212685,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/0c279e0886697f03a87c0a91.pdf"},{"id":62271717,"identity":"7b477e95-b030-4ca5-bfa1-840bba046fab","added_by":"auto","created_at":"2024-08-12 10:23:35","extension":"csv","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":63463,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1.csv","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/b4f9f93ef419a764f4d20f29.csv"},{"id":62271712,"identity":"5af82d52-5046-4e39-a04b-1bd304ab819d","added_by":"auto","created_at":"2024-08-12 10:23:35","extension":"csv","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":25659,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.csv","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/e94b38e645ab28a60b304495.csv"},{"id":62273189,"identity":"891bc0fd-1b83-40fc-be9a-78b2a84e70dc","added_by":"auto","created_at":"2024-08-12 10:39:35","extension":"csv","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":58754,"visible":true,"origin":"","legend":"","description":"","filename":"TableS3.csv","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/5ccf757d464d881eb9c4896f.csv"},{"id":62272371,"identity":"b9aecb9b-2b36-4a95-8f92-f73bfdffa223","added_by":"auto","created_at":"2024-08-12 10:31:36","extension":"csv","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":58754,"visible":true,"origin":"","legend":"","description":"","filename":"TableS4.csv","url":"https://assets-eu.researchsquare.com/files/rs-4717221/v1/327ddfbe7c1021dd02fa09d7.csv"}],"financialInterests":"No competing interests reported.","formattedTitle":"Co-altered vaginal Lactobacillus, metabolome and host gene expression associate with the grade of cervical intraepithelial neoplasia in Chinese women","fulltext":[{"header":"Background","content":"\u003cp\u003ePersistent human papillomavirus (HPV) infection poses a significant risk for cervical intraepithelial neoplasia (CIN) and cervical cancer [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. However, the intricacies of factors modulating persistent HPV infection and CIN progression remain to be understood. Recent studies highlight the potential significance of vaginal microbiota (VM) in influencing these risks [\u003cspan additionalcitationids=\"CR3 CR4 CR5 CR6 CR7 CR8 CR9 CR10\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eVM with dominant \u003cem\u003eLactobacillus\u003c/em\u003e has been demonstrated prevalent in healthy women [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Cross-sectional studies revealed a decrease in the level of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e and variations in metabolites among HPV-positive women with high-grade CIN, compared to those with lower CIN grades [\u003cspan additionalcitationids=\"CR3 CR4 CR5\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Furthermore, prospective studies suggest causal links between vaginal \u003cem\u003eLactobacillus\u003c/em\u003e and the risk of persistent HPV infection as well as CIN progression [\u003cspan additionalcitationids=\"CR8 CR9 CR10\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Specifically, women with \u003cem\u003eLactobacillus crispatus\u003c/em\u003e-dominant VM exhibited a positive association with natural CIN regression, whereas those with non-\u003cem\u003eLactobacillus\u003c/em\u003e-dominant VM structures faced an increased risk of CIN progression [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe impact of VM on the risk of CIN progression is partially attributable to its modulation of host immune responses [\u003cspan additionalcitationids=\"CR15 CR16 CR17\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Recent investigations have uncovered associations between VM and various immune mediators, including a negative correlation between \u003cem\u003eLactobacillus\u003c/em\u003e and immune checkpoint inhibitors like LAG-3 and PD-L1 [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Certain vaginal \u003cem\u003eLactobacillus\u003c/em\u003e strains have been shown to modulate epithelial innate immune responses and activate Langerhans cells [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], which are vital in detecting HPV antigens and activating cellular immunity. Moreover, in vitro study has demonstrated the inhibitory effect of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e on the growth of cancer-related Ect1/E6E7 and CaSKi cells, thereby influencing CIN progression and potentially carcinogenesis [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eHowever, there is a paucity of population-based studies exploring the underlying mechanisms that link VM composition to the risk of CIN progression. In this study, we enrolled 75 HPV-positive Chinese women diagnosed with either low-grade or high-grade CIN. Our primary objectives were to elucidate the variations in the VM, metabolome, and host gene expression that are associated with CIN grade, and to assess the potential mediating role of metabolites in VM-host interactions. Our findings are expected to provide valuable population-based evidence for comprehending the influence of VM on the risk of CIN progression, thereby contributing to a deeper understanding of mechanisms underlying this complex relationship.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy population\u003c/h2\u003e \u003cp\u003eThis study encompassed 75 HPV-positive women, who provided vaginal swabs and pap-brush samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Table\u0026nbsp;1). Out of these, 149 vaginal swabs were applied for subsequent analysis, with 75 samples qualifying for VM examination and 74 for vaginal metabolome assessment. Additionally, 43 pap-brush samples were selected for host transcriptome analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFor included HPV-positive women, there were no significant variations in the average age between the CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts (Table\u0026nbsp;1). Most women in both groups were tested positive for HPV16/18-excluded high-risk HPV genotypes, with approximately 21% of women in both groups exhibiting HPV16 positivity. However, the proportion of HPV18-positive women was higher in the CIN1 group compared to the CIN2\u0026thinsp;+\u0026thinsp;group (Table\u0026nbsp;1). Among the 75 women included, 4, 15, and 14 women had received the 2-valent, 4-valent, and 9-valent HPV vaccine respectively (Table\u0026nbsp;1). Furthermore, there were no significant differences observed in other phenotypes such as smoking between CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts (Table\u0026nbsp;1).\u003c/p\u003e \u003cp\u003e \u003cb\u003eThe increased proportion of\u003c/b\u003e \u003cb\u003eLactobacillus\u003c/b\u003e\u003cb\u003e-dominant community state type (LD CST) in CIN1 cases compared to those in the CIN2\u0026thinsp;+\u0026thinsp;cohort\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThe LD CST VM groups comprised \u003cem\u003eLactobacillus crispatus\u003c/em\u003e-dominant (LCD, 16%), \u003cem\u003eLactobacillus iners\u003c/em\u003e-dominant (LID, 45.33%), \u003cem\u003eLactobacillus gasseri\u003c/em\u003e-dominant (LGD, 2.67%), \u003cem\u003eLactobacillus jensenii-dominant\u003c/em\u003e (LJD, 1.33%), and CSTs dominated by more than two \u003cem\u003eLactobacillus\u003c/em\u003e species (MixedLD, 16%) (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eA). The VM groups categorized as non-\u003cem\u003eLactobacillus\u003c/em\u003e-dominant (NLD) CST accounted for 18.67% of the 75 microbial samples, exhibiting a heterogeneous community primarily consisting of \u003cem\u003eGardnerella\u003c/em\u003e, \u003cem\u003eFannyhessea\u003c/em\u003e or \u003cem\u003eStreptococcus\u003c/em\u003e species (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003eAt the phylum level, a major proportion of microbial samples in the CIN1 group (92.11%) were dominated by Firmicutes, whereas this dominance was less prevalent in the CIN2\u0026thinsp;+\u0026thinsp;group, comprising 78.38% of samples (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eB). Conversely, Actinobacteria displayed a lower prevalence in the CIN1 group (5.26%) and a significantly higher proportion in the CIN2\u0026thinsp;+\u0026thinsp;group (21.62%) (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eB). Similarly, we observed a decreased proportion of microbial samples with the NLD CST in the CIN1 group (13.16%) compared to the CIN2\u0026thinsp;+\u0026thinsp;group (24.32%) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-D). Further analysis revealed a slightly higher proportion of LID and a lower proportion of LCD CST in the CIN1 group (LID: 50% vs 40.54% in the CIN2\u0026thinsp;+\u0026thinsp;group; LCD: 13.16% vs 18.92% in the CIN2\u0026thinsp;+\u0026thinsp;group) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC-D).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eThe association of altered vaginal metabolome with\u003c/b\u003e \u003cb\u003eLactobacillus\u003c/b\u003e\u003c/p\u003e \u003cp\u003eOur Procrustes analysis robustly demonstrated a synergistic interplay between VM and metabolome profiles (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Employing Bray-Curtis dissimilarity as a metric, we observed significantly higher similarity within LD samples compared to those within NLD samples as well as those between LD and NLD samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). The two-way orthogonal partial least squares (O2PLS) analysis showed notably higher loading value of \u003cem\u003eLactobacillus\u003c/em\u003e and associated D-lactic acid (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). In addition, the dissimilarity among NLD samples mirrored that between LD and NLD groups, underscoring the substantial heterogeneity in metabolic profiles among microbial samples within the NLD CST (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo further decipher the metabolic determinants distinguishing CIN1 from CIN2\u0026thinsp;+\u0026thinsp;cohorts, we implemented a random forest classifier to pinpoint metabolites that notably contributed to the differentiation. Given the high homogeneity of metabolic profiles in LD samples, subsequently we examined associations of those metabolites with vaginal \u003cem\u003eLactobacillus\u003c/em\u003e. Among the top 100 contributing metabolites, 31 exhibited a notable correlation (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1) with vaginal \u003cem\u003eLactobacillus\u003c/em\u003e levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003eSpecifically, \u003cem\u003eLactobacillus\u003c/em\u003e demonstrated a negative correlation with 26 metabolites, spanning diverse chemical classes such as benzenoids (e.g., Phenylethylamine, Dezocine, Benzoylcholine, 2,4-Dichlorobenzoic acid), lipids and lipid-like molecules ((9,10)-Epoxyoctadecenoic acid, 11-Hydroxy-9-tridecenoic acid, Dodecylbenzenesulfonic acid, LysoPS 14:0), organic acids and derivatives (Thr-Lys, (S)-9-Hydroxy-10-undecenoic acid, N-Undecanoylglycine, 3-Methyl-2-oxovaleric acid), and others including 3-Methyl-1-butylamine, 2-Hydroxypyridine, L-Carnitine, UDP-D-glucose, 3-Dehydroxycarnitine, D-Mannose-6-phosphate, and beta-D-Glucosamine (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Conversely, five metabolites\u0026mdash;Indole-5,6-quinone, 4-Hydroxyphenylmaraviroc, Succinic anhydride, L-Kynurenine, belonging to organic oxygen and organoheterocyclic compounds\u0026mdash;positively correlated with \u003cem\u003eLactobacillus\u003c/em\u003e levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAmong the remaining 69 metabolites that did not exhibit notable correlation with \u003cem\u003eLactobacillus\u003c/em\u003e levels, we observed distinct metabolic patterns between the CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts. Specifically, the CIN1 cohort displayed elevated levels of several lipid and lipid-like molecules, including glycerophospholipids (e.g., PG 34:1, PG(16:0/18:1), PG 36:1; PG(18:0/18:1)), sterol lipids, fatty acyls (adipic acid, 11-Oxohexadecanoic acid, Ethyl hydrogen fumarate), and sphingolipids (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Additionally, increased concentrations of benzene and substituted derivatives (Procaine, Cardanoldiene), organonitrogen compounds, keto acids and derivatives, and hydroxy acids and derivatives were observed in the CIN1 cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn contrast, the CIN1 cohort exhibited decreased levels of other lipid and lipid-like molecules, notably O-Acetyl-L-carnitine and D-(+)-Malic acid (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Furthermore, several organic acids and derivatives, such as beta-Alanine, (S)-3,4-Dihydroxybutyric acid, and Homovanillic acid sulfate, were found to be downregulated in the CIN1 cohort compared to the CIN2\u0026thinsp;+\u0026thinsp;cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Analysis also revealed decreased levels of organoheterocyclic compounds, particularly indoles and derivatives (e.g., Isorhynchophylline) and dihydrofurans (e.g., Norfuraneol, Furanone A) in the CIN1 cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eCervical transcriptomic differences between CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohort\u003c/h2\u003e \u003cp\u003eUnder the stringent threshold of FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1, our analysis revealed a differential expression pattern between the CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts, with 176 host genes upregulated and 82 genes downregulated in the CIN1 cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Functional enrichment analysis of the upregulated genes in the CIN1 cohort predominantly associated them with immune responses, cell chemotaxis, negative regulation of cell migration, and B cell differentiation, activation, and proliferation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). Notably, several genes exhibiting\u0026thinsp;\u0026ge;\u0026thinsp;2-fold change in the CIN1 cohort were implicated in innate or adaptive immune responses, including CCR6, CXCR3, CXCR5, CD200, CD79A, CD79B, CD19, CDH6, CADM1, DEFB103A and DEFB103B (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). Additionally, CCBE1, IL36RN, SPON1, SERPINA5 and COL4A2 were among the other genes with \u0026ge;\u0026thinsp;2-fold change enriched in the CIN1 cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn contrast, genes enriched in the CIN2\u0026thinsp;+\u0026thinsp;cohort were primarily associated with epidermis development, keratinocyte differentiation, keratinization, and epidermal cell differentiation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). Among these, genes with high overall expression levels included FABP5, CDKN2A (encoding p16 protein), TP63, KRT24, UPK3BL1, PHGDH, KCNS1, LGALS7B, and CALML5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRandom forest analysis pinpointed the importance of differentially-expressed genes(DEGs) to the differences observed between CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Among the top 15 contributors, KCNS1, KRT24 and FOXD1 exhibited significantly higher expression levels in the CIN2\u0026thinsp;+\u0026thinsp;cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). Conversely, among these 15 key contributors, 12 genes, including CCBE1, SERPINA5, SPON1, and COL4A2, were found to have higher expression levels in the CIN1 cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAlthough the statistical significance was not achieved (FDR\u0026thinsp;\u0026gt;\u0026thinsp;0.1), we observed an increased expression trend of genes encoding other defensin and defensin-like molecules in the CIN1\u0026thinsp;+\u0026thinsp;cohort, encompassing DEFB1, DEFB104A, DEFB4A, CAMP, PI3, S100A7, and SLPI (Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). In contrast, within the CIN2\u0026thinsp;+\u0026thinsp;cohorts, an upregulation of genes associated with p16 and Ki67 proteins, which serve as clinical biomarkers for CIN grading, was evident (Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Additionally, genes related to common inflammatory markers, such as interleukin (IL)-1α and macrophage inflammatory protein (MIP)-1α, exhibited elevated expression levels in the CIN2\u0026thinsp;+\u0026thinsp;cohort (Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cb\u003eThe correlation between vaginal\u003c/b\u003e \u003cb\u003eLactobacillus\u003c/b\u003e, \u003cb\u003ealtered metabolites and cervical gene expression\u003c/b\u003e\u003c/p\u003e \u003cp\u003eTo elucidate the intricate \u003cem\u003eLactobacillus\u003c/em\u003e-metabolite-host gene interactions associated with the CIN grade, we employed linear regression to quantify the correlations between \u003cem\u003eLactobacillus\u003c/em\u003e abundance, the top 100 metabolites contributing to inter-cohort differences, and DEGs. Besides to notable correlation with 31 metabolites (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e), vaginal \u003cem\u003eLactobacillus\u003c/em\u003e levels correlated with 30 DEGs, including positive correlation with CIN1-enriched CCBE1, COL4A2 and SPON1 as well as negative correlation with CIN1-depleted FOXD1, TLX3 and IL36RN (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAmong the 31 metabolites significantly correlated with \u003cem\u003eLactobacillus\u003c/em\u003e (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1) (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e), all displayed robust associations with the expression of at least one host gene, representing a total of 239 DEGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). Our analysis revealed 1063 distinct metabolite-DEG correlations (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e), with key metabolites such as (S)-9-Hydroxy-10-undecenoic acid, N-Undecanoylglycine, (9,10)-Epoxyoctadecenoic acid, UDP-D-glucose, L-Carnitine, Thr-Lys, Phenylethylamine, 3-Methyl-2-oxovaleric acid, Dodecylbenzenesulfonic acid, Dezocine, D-Mannose-6-phosphate, beta-D-Glucosamine, and Succinic anhydride predominating (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). Notably, these correlations involved DEGs primarily comprising COL4A2, SPON1, CCBE1, DEFB103A, and DEFB103B, which were upregulated in the CIN1 cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). Additionally, genes enriched in the CIN2\u0026thinsp;+\u0026thinsp;cohort, like FOXD1 and TLX3, were also implicated (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFurther mediation analysis has significantly enriched our understanding of the complex relationships between vaginal \u003cem\u003eLactobacillus\u003c/em\u003e and gene expression patterns, specifically identifying mediation effects of 15 metabolites in the associations between \u003cem\u003eLactobacillus\u003c/em\u003e and 18 DEGs. Importantly, the proportion of mediation effects was statistically significant, as evidenced by FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1 and the lowest value of the 95% confidence interval (CI) exceeding zero (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Among these mediators, L-Cartinine stands out for its dual role. Negatively correlated with both \u003cem\u003eLactobacillus\u003c/em\u003e abundance and COL4A2 expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e), L-Cartinine acts as a mediator in two opposing associations: it facilitates the positive relationship between \u003cem\u003eLactobacillus\u003c/em\u003e and COL4A2, which is enriched in CIN1-positive samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e), while simultaneously mediating the negative association between \u003cem\u003eLactobacillus\u003c/em\u003e and TLX3, which is depleted in CIN1-positive samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSimilarly, UDP-D-glucose exhibits a negative correlation with \u003cem\u003eLactobacillus\u003c/em\u003e levels and is implicated in mediating the negative associations between \u003cem\u003eLactobacillus\u003c/em\u003e and four CIN1-depleted DEGs, notably FOXD1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B, Tables S2-3). Consistent with this, these DEGs positively correlate with UDP-D-glucose levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e), reinforcing the mediating role of UDP-D-glucose. Additionally, three other metabolites\u0026mdash;(S)-9-Hydroxy-10-undecenoic acid, N-Undecanoylglycine, and 9,10-Epoxyoctadecenoic acid\u0026mdash;all negatively correlated with vaginal \u003cem\u003eLactobacillus\u003c/em\u003e abundance, and mediate the negative associations between \u003cem\u003eLactobacillus\u003c/em\u003e and their positively correlated DEGs, such as LGALS7B, TLX3, and GLS2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B, Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFurthermore, nine additional metabolites, including 3,5-Dichlorobenzoic acid (neg.M189T37), 3-Dehydroxycarnitine (pos.M146T47), and others, play significant mediating roles in the relationships between vaginal \u003cem\u003eLactobacillus\u003c/em\u003e and 12 DEGs, notably GYS2 and TLX3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study provides compelling evidence of co-alterations in vaginal \u003cem\u003eLactobacillus\u003c/em\u003e, the metabolome, and host gene expression, elucidating the influence of \u003cem\u003eLactobacillus\u003c/em\u003e on host gene expression patterns that correlate with the severity of cervical lesions. Consistent with previous reports, \u003cem\u003eLactobacillus\u003c/em\u003e-dominated VM types were found to be less prevalent among women with high-grade CIN [\u003cspan additionalcitationids=\"CR3 CR4 CR5\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], compared to those with low-grade CIN. Previous studies have attributed the beneficial effects of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e to their production of D-lactic acid and hydrogen peroxide [\u003cspan additionalcitationids=\"CR23 CR24\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], and our study concurs with these findings, demonstrating a notable association between \u003cem\u003eLactobacillus\u003c/em\u003e and D-lactic acid levels.\u003c/p\u003e \u003cp\u003eFurthermore, we identified significant correlations between vaginal \u003cem\u003eLactobacillus\u003c/em\u003e and a panel of metabolites that effectively distinguish between CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts, partly aligning with prior research linking \u003cem\u003eLactobacillus\u003c/em\u003e with vaginal metabolites associated with HPV infection and CIN grade [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eExtending our investigation to cervical gene expression, this population-based study delves into the transcriptomic profiles of both CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts. Elevated expression of the p16-encoding CDKN2A gene in the CIN2\u0026thinsp;+\u0026thinsp;cohort corroborates our pathological findings and the clinical consensus, which highlights positive p16 staining as an indicator of high-grade CIN [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. This validates the feasibility of RNA-Seq analysis using cervical exfoliated cells as a diagnostic tool. Moreover, we discovered that genes depleted in the CIN2\u0026thinsp;+\u0026thinsp;cohort are primarily associated with immune responses and negative regulation of cell migration, echoing previous reports that implicate altered cell migration and compromised immunity in HPV-infected high-grade CIN and cervical cancer [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Additional findings, including decreased expression of defensins and defensin-like molecules, alongside increased inflammatory markers in the CIN2\u0026thinsp;+\u0026thinsp;cohort, are consistent with multiple cohort studies [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e], reinforcing the robustness of our observations.\u003c/p\u003e \u003cp\u003eFurther in-depth analysis revealed intriguing correlations among vaginal \u003cem\u003eLactobacillus\u003c/em\u003e populations, metabolites that exhibited high discriminatory power between the two cohorts, and DEGs. Notably, \u003cem\u003eLactobacillus\u003c/em\u003e was inversely correlated with several metabolites, including 9,10-Epoxyoctadecenoic acid, L-Carnitine, N-Undecanoylglycine, (S)-9-Hydroxy-10-undecenoic acid, and UDP-D-glucose, which, in turn, showed positive associations with the oncogene TP63. These findings partially corroborate previous studies emphasizing the protective role of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e in hindering the progression of CIN [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn contrast, N-Undecanoylglycine and 9,10-Epoxyoctadecenoic acid exhibited significant negative correlations with the expression of genes (CD19, CD79A, CD79B, CD200, and CCR6) that are enriched in CIN1 and play crucial roles in both innate and adaptive immunity [\u003cspan additionalcitationids=\"CR36 CR37\" citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. This observation suggests a potential mechanism whereby alterations in these metabolites may impact immune function in the context of CIN. Additionally, L-Carnitine was negatively correlated with the expression of DEFB103A and DEFB103B, which encode defensins, integral components of the innate immune barrier in the female lower genital tract. Notably, decreased expression of these defensins has been reported in high-grade CIN [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan additionalcitationids=\"CR40\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e], further implicating L-Carnitine in the modulation of innate immune responses during CIN progression.\u003c/p\u003e \u003cp\u003eFurthermore, our analysis revealed a notable positive correlation between UDP-D-glucose and the FOXD1 gene. Subsequent analysis demonstrated a significant mediation effect of UDP-D-glucose in the negative association observed between vaginal \u003cem\u003eLactobacillus\u003c/em\u003e and FOXD1 gene expression. Given the emerging evidence indicating the pro-tumorigenic role of FOXD1 in cancer cell proliferation, cancer progression, and poor prognosis [\u003cspan additionalcitationids=\"CR43 CR44\" citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e], this finding suggests that vaginal \u003cem\u003eLactobacillus\u003c/em\u003e may impede the progression of CIN by inhibiting FOXD1 expression through downregulation of UDP-D-glucose. While the precise mechanisms underlying this mediation remain elusive, recent studies have highlighted the ability of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e to inhibit the growth of certain vaginal pathogens [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. This, in turn, may lead to a reduction in the production of pathogen-derived UDP-D-glucose and its product, UDP-D-glucuronic acid (UDP-GlcA) [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. Although there is currently no direct evidence linking UDP-GlcA to CIN progression, previous studies have shown that excessive UDP-GlcA can disrupt Golgi function, thereby compromising cellular signaling capabilities [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. Moreover, other researchers have demonstrated a positive role for UDP-GlcA in promoting tumor cell migration and cancer metastasis [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. To fully elucidate the role of UDP-D-glucose and UDP-GlcA in the \u003cem\u003eLactobacillus\u003c/em\u003e-mediated inhibition of FOXD1 and CIN progression, future studies are warranted to investigate the underlying molecular mechanisms and validate these findings in larger, well-controlled cohorts.\u003c/p\u003e \u003cp\u003eOur study also uncovered a notable mediation effect of L-Carnitine in the positive correlation observed between \u003cem\u003eLactobacillus\u003c/em\u003e and COL4A2 gene expression, suggesting that \u003cem\u003eLactobacillus\u003c/em\u003e positively regulates COL4A2 expression by downregulating the production of L-Carnitine. COL4A2 encodes type IV collagen, which is a major structural component of basement membranes and plays a pivotal role in maintaining the integrity of the basal membrane [\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e]. This finding aligns partially with previous research demonstrating elevated levels of type IV collagen in the cervix of individuals with lower grades of cervical lesions [\u003cspan additionalcitationids=\"CR55 CR56 CR57\" citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e], indicating its potential involvement in the prevention of cervical pathology. Intriguingly, Amiloride has been shown to suppress the migratory and invasive capabilities of HeLa cells by inhibiting the degradation of type IV collagen, achieved through downregulation of type IV collagenase expression [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. Additionally, the C-terminal fragment of type IV collagen, known as canstatin, has been identified as a potent inhibitor of tumor growth [\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e], underscoring the multifaceted role of COL4A2 and its products in modulating cancer progression. Collectively, these findings highlight the significance of COL4A2 expression in cervical epithelial cells, emphasizing its potential role in preventing the progression of CIN.\u003c/p\u003e \u003cp\u003eLimitations of the present study warrant acknowledgment. Firstly, the small sample size and the use of cross-sectional multi-omic profiling pose limitations in drawing definitive conclusions. However, these challenges were partially mitigated by the application of rigorous inclusion criteria, specifically excluding participants with common vaginal infections and menopause, which are known to significantly influence VM [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. Secondly, the utilization of 16S rDNA amplicon sequencing, while informative, constrained our ability to fully comprehend the functional communities within the VM [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]. Consequently, we were unable to delve into intra-group heterogeneity and inter-group variations among vaginal \u003cem\u003eLactobacillus\u003c/em\u003e at the strain level and their functional landscapes. This underscores the need for future studies incorporating more comprehensive sequencing approaches, such as shotgun metagenomics, to address these gaps. Lastly, although our analysis revealed \u003cem\u003eLactobacillus\u003c/em\u003e-metabolite-host linkages that correlated with CIN grade, establishing causality remains an area for further investigation. The observed associations between UDP-D-glucose and L-Carnitine mediating the influence of \u003cem\u003eLactobacillus\u003c/em\u003e on FOXD1 and COL4A2 gene expression, respectively, are intriguing but preliminary.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eOverall, this study presents a compelling case for co-alterations in vaginal \u003cem\u003eLactobacillus\u003c/em\u003e, metabolites, and host gene expression patterns that are associated with CIN grade. Our findings hint at potential roles of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e in modulating the risk of CIN progression, particularly through the mediation of UDP-D-glucose and L-Carnitine on FOXD1 and COL4A2 expression. Nevertheless, further research is imperative to unravel the intricate mechanisms underlying these associations and mediation effects, including longitudinal studies and intervention trials to confirm causality and elucidate the full extent of their clinical implications.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStudy design\u003c/h2\u003e \u003cp\u003eA total of seventy-five HPV-positive women with varying grades of CIN were recruited through interviews conducted at the Department of Obstetrics and Gynecology of Peking University Shenzhen Hospital in China. Adhering to the enrollment criteria outlined in our previous study [\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e], participants were required to be over 18 years of age, not experiencing menopause, and without a history of cervical ablation, resection surgery, hysterectomy, pelvic radiotherapy, sexual intercourse within three days, douching, or vaginal medication use within seven days, antibiotic exposure within one month, hormone replacement therapy or GnRH-a use within three months, and no autoimmune diseases or HIV infection. Additionally, vaginal smears were tested negative for H\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e, leukocyte esterase, neuraminidase, b-glucuronidase, and acetylaminoglucosidase to exclude common genital infections. All participants were fully informed and provided written consents prior to the enrollment.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eData and sample collection\u003c/h2\u003e \u003cp\u003eWith the aid of a research assistant, the clinician diligently recorded crucial metadata for each participant. This included information on smoking habits, sexual activity history, contraceptive use, gestation and abortion status, HPV genotypes, and vaccination records. To ensure the homogeneity of the samples, vaginal swabs were collected\u0026thinsp;\u0026ge;\u0026thinsp;3 days following menstruation. During colposcopy, a single physician expertly sampled the posterior fornix for vaginal swabs. All swabs were promptly preserved in a 2 ml sterile tube kept on ice and transferred to -80℃ storage within 30 minutes of collection.\u003c/p\u003e \u003cp\u003eFor the host transcriptome analysis, cervical exfoliated cells were gently collected using Pap brushes. To maintain RNA integrity, the same skilled technician immediately conducted RNA extraction using a specialized kit (Bacteria RNA Extraction Kit, Vazyme Biotech Co., Ltd China) following sample collection. This ensured the quality and reliability of the data for subsequent analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eHPV testing and pathology diagnosis\u003c/h2\u003e \u003cp\u003eFor HPV screening, we employed the Roche Cobas\u0026reg;4800-HPV system, encompassing specific detection of HPV16/18 and 12 additional high-risk HPV genotypes (HPV 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68). The diagnosis of CIN was performed by two seasoned pathologists, utilizing p16 immunohistochemical staining as an auxiliary tool to differentiate between CIN1 and CIN2\u0026thinsp;+\u0026thinsp;lesions. Specifically, cytologically diagnosed CIN2 samples that exhibited p16 positivity were categorized as CIN2+; otherwise, they were classified as CIN1 [\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMicrobial DNA extraction, 16S rRNA gene amplicon sequencing and data processing\u003c/h2\u003e \u003cp\u003eMicrobial DNA of vaginal swabs was extracted using the Dneasy PowerSoil Pro Kit (Qiagen, Germany). The concentration and purity of the extracted DNA were assessed via 1% agarose gels on an Agilent5400 platform (Agilent Technologies, Inc., Santa Clara, USA). For library construction, we amplified the 16S rDNA V4-V5 hypervariable regions using specific primers: 515-FR (GTGCCAGCMG CCGCGGTAA) and 926-RR (CCGTCAATTCMTTTRAGTTT). The resultant DNA libraries were sequenced on the Illumina NovaSeq platform (Illumina, San Diego, CA, United States), with a read length of 250 bp. The raw sequencing data underwent rigorous analysis using QIIME2 [\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e] for the generation of VM profiles.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eVM CST analysis\u003c/h2\u003e \u003cp\u003eWe utilized unsupervised hierarchical clustering with average Euclidean linkage to identify distinct VM groups from all 75 women. The nomenclature for each microbiome group was determined based on whether the microbiome profiles exhibited a dominant bacterial genus or species with the abundance of \u0026ge;\u0026thinsp;50%. For the LD CST that lacked a dominant bacterial genus or species, the microbiome group was specifically designated as MixedLD, indicating its dominance by more than two \u003cem\u003eLactobacillus\u003c/em\u003e species.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eUntargeted metabolomics analysis\u003c/h2\u003e \u003cp\u003eTo characterize the vaginal metabolome, we employed liquid chromatography-mass spectrometry (LC-MS) utilizing the Thermo Fisher Scientific platform (Ottawa, United States). Raw intensity data were converted to the mzXML format, and ion features were extracted utilizing the Progenesis QI software (v.2.2). These ions underwent rigorous filtering, excluding those missing in over 50% of quality control samples or more than 80% of test samples, as well as those with a relative standard deviation exceeding 30%. To identify the metabolites, we searched the Human Metabolome Database (HMDB, v.5.0) and Kyoto Encyclopedia of Genes and Genomes (KEGG, v.96.0). The resultant metabolite abundance matrix provided the cornerstone for subsequent analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eRNA-Seq and data processing\u003c/h2\u003e \u003cp\u003eFor RNA sequencing (RNA-Seq), total RNA was extracted from cervical exfoliated cells using the Qiagen RNease Mini Kit. RNA-Seq was performed on the Illumina NovaSeq platform. The raw sequencing reads were first filtered with Cutadapt (v.2.5) [\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e] to remove low-quality and adapter sequences. Subsequently, these reads were aligned to the human genome (GRCh38) using Hisat2 (v.2.1.0) [\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e]. Gene expression levels were quantified and normalized using RSEM (v.1.3.3) [\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e] and DESeq2 (v.1.20) [\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e], respectively. Then the DESeq2 (v.1.20) software was applied to analyze the differences of host gene expression between CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohort [\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e]. Finally, the DAVID Knowledgebase (v2023q4) [\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e] was utilized to identify enriched functional pathways for DEGs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eBioinformatics and statistics\u003c/h2\u003e \u003cp\u003eThe synergy between the VM and its associated metabolome was meticulously evaluated through the application of Procrustes analysis, implemented in the vegan package of the R software environment. Subsequently, a random forest classifier, consisting of 100 trees and employing leave-one-out cross-validation, was harnessed to assess the pivotal role of individual metabolites in distinguishing the CIN1 from CIN2\u0026thinsp;+\u0026thinsp;cohort, utilizing the mlr package in R. Prior to conducting the association analysis, all continuous variables were standardized to conform to a standard normal distribution (N~(0, 1)) by employing an empirical normal quantile transformation.\u003c/p\u003e \u003cp\u003eSubsequently, we employed linear models to systematically examine the associations between \u003cem\u003eLactobacillus\u003c/em\u003e abundance and the top 100 metabolites distinguishing the CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohort, as well as the relationships between these key metabolites and differentially expressed host genes. The linear models were formulated as follows:\u003c/p\u003e \u003cp\u003eMetabolite\u0026thinsp;~\u0026thinsp;\u003cem\u003eLactobacillus\u003c/em\u003e\u0026thinsp;+\u0026thinsp;Age\u003c/p\u003e \u003cp\u003eGene\u0026thinsp;~\u0026thinsp;Metabolite\u0026thinsp;+\u0026thinsp;Age\u003c/p\u003e \u003cp\u003eThe mediation effects along the \u003cem\u003eLactobacillus\u003c/em\u003e\u0026ndash;metabolite\u0026ndash;host axis were thoroughly examined using the R mediate package, with the differentially expressed host genes serving as continuous dependent variables, and age included as the covariate. The mediation analysis entailed fitting two linear models:\u003c/p\u003e \u003cp\u003eMetabolite\u003csub\u003ei\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;α\u003csub\u003e1\u003c/sub\u003e\u0026thinsp;+\u0026thinsp;β\u003csub\u003e1\u003c/sub\u003eLacto\u0026thinsp;+\u0026thinsp;δ\u003csub\u003e1\u003c/sub\u003e\u003csup\u003eT\u003c/sup\u003eX\u003csub\u003ei\u003c/sub\u003e\u0026thinsp;+\u0026thinsp;ε\u003csub\u003ei1\u003c/sub\u003e\u003c/p\u003e \u003cp\u003eGene\u003csub\u003ei\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;α\u003csub\u003e2\u003c/sub\u003e\u0026thinsp;+\u0026thinsp;β\u003csub\u003e2\u003c/sub\u003eLacto\u0026thinsp;+\u0026thinsp;γ\u003csub\u003e1\u003c/sub\u003eMetabolite\u003csub\u003ei\u003c/sub\u003e\u0026thinsp;+\u0026thinsp;δ\u003csub\u003e2\u003c/sub\u003e\u003csup\u003eT\u003c/sup\u003eX\u003csub\u003ei\u003c/sub\u003e\u0026thinsp;+\u0026thinsp;ε\u003csub\u003ei2\u003c/sub\u003e\u003c/p\u003e \u003cp\u003ewhere Lacto denotes the relative abundance of vaginal \u003cem\u003eLactobacillus\u003c/em\u003e. Furthermore, Metabolite\u003csub\u003ei\u003c/sub\u003e signifies each metabolite in the top 100 metabolite differentiating CIN1 from CIN2\u0026thinsp;+\u0026thinsp;cohort. Additionally, Xi represents a vector of covariates, while Gene\u003csub\u003ei\u003c/sub\u003e stands for each differentially expressed host gene, also with a significance threshold of FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1. After fitting these two modules, the product of two coefficients β\u003csub\u003e1\u003c/sub\u003eγ\u003csub\u003e1\u003c/sub\u003e was interpreted as an estimate of average causal mediated effect and the coefficient β\u003csub\u003e2\u003c/sub\u003e was interpreted as an estimate of the average direct effect. Significant mediation effects were determined based on an FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1 for the average mediation effect, with the lowest value of the 95% confidence interval (CI) exceeding zero.\u003c/p\u003e \u003cp\u003eAll statistical analyses and data visualization were executed using R software (version 4.0.5), adhering to the established significance threshold of FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1 for statistical significance.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Ethics Committee of Peking University Shenzhen Hospital (registration number: 2022-157). All participants were fully informed and then provided signed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe raw data of VM, metabolome and host transcriptome analysis were accessible in CNGB Sequence Archive (CNSA) under Project No. CNP0005859. Other data supporting the findings of this study are available from the corresponding authors on reasonable request.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere is no competing interests to be declared.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the National Natural Science Foundation of China (82202826), Shenzhen High-level Hospital Construction Fund (YBH2019-260), Shenzhen Key Medical Discipline Construction Fund (SZXK027), Sanming Project of Medicine in Shenzhen (SZSM202011016), Shenzhen Public Platform for Preservation of Fertility and Reproduction (XMHT20220104049) and Peking University Shenzhen Hospital Scientific Research Fund (KYQD2021075 and KYQD2022132).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eW.R., L.S. and D.H. conceived the study. D.H., J.H., J.X.,L.Y. and L.C. recruited and selected attenders. G.C., W.D. and Y.Qing performed the sample collection and associated experimental analysis. D.W. performed data analysis and drafted the manuscript. G.C., J.X., W.Y., Z.Q., X.R. and Y.Qin helped with the data analysis. L.S. polished the manuscript.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eLei J, Ploner A, Elfstr\u0026ouml;m KM, Wang J, Roth A, Fang F, et al. HPV Vaccination and the Risk of Invasive Cervical Cancer. N Engl J Med. 2020;383(14):1340-1348.\u003c/li\u003e\n \u003cli\u003eIlhan ZE, Łaniewski P, Thomas N, Roe DJ, Chase DM, Herbst-Kralovetz MM. Deciphering the complex interplay between microbiota, HPV, inflammation and cancer through cervicovaginal metabolic profiling. EBioMedicine. 2019;44:675-690.\u003c/li\u003e\n \u003cli\u003eGuo C, Dai W, Zhou Q, Gui L, Cai H, Wu D, et al. Cervicovaginal microbiota significantly changed for HPV-positive women with high-grade squamous intraepithelial lesion. Front Cell Infect Microbiol. 2022;12:973875.\u003c/li\u003e\n \u003cli\u003eChen Y, Qiu X, Wang W, Li D, Wu A, Hong Z, et al. Human papillomavirus infection and cervical intraepithelial neoplasia progression are associated with increased vaginal microbiome diversity in a Chinese cohort. BMC Infect Dis. 2020;20(1):629.\u003c/li\u003e\n \u003cli\u003eMitra A, MacIntyre DA, Lee YS, Smith A, Marchesi JR, Lehne B, et al. Cervical intraepithelial neoplasia disease progression is associated with increased vaginal microbiome diversity. Sci Rep. 2015;5:16865.\u003c/li\u003e\n \u003cli\u003eZheng JJ, Song JH, Yu CX, Wang F, Wang PC, Meng JW. Difference in vaginal microecology, local immunity and HPV infection among childbearing-age women with different degrees of cervical lesions in Inner Mongolia. BMC Womens Health. 2019;19(1):109.\u003c/li\u003e\n \u003cli\u003eMitra A, MacIntyre DA, Ntritsos G, Smith A, Tsilidis KK, Marchesi JR, et al. The vaginal microbiota associates with the regression of untreated cervical intraepithelial neoplasia 2 lesions. Nat Commun. 2020;11(1):1999.\u003c/li\u003e\n \u003cli\u003eUsyk M, Zolnik CP, Castle PE, Porras C, Herrero R, Gradissimo A, et al. Cervicovaginal microbiome and natural history of HPV in a longitudinal study. PLoS Pathog. 2020;16(3):e1008376.\u003c/li\u003e\n \u003cli\u003eDareng EO, Ma B, Adebamowo SN, Famooto A, Ravel J, Pharoah PP, et al. Vaginal microbiota diversity and paucity of Lactobacillus species are associated with persistent hrHPV infection in HIV negative but not in HIV positive women. Sci Rep. 2020;10(1):19095.\u003c/li\u003e\n \u003cli\u003eDi Paola M, Sani C, Clemente AM, Iossa A, Perissi E, Castronovo G, et al. Characterization of cervico-vaginal microbiota in women developing persistent high-risk Human Papillomavirus infection. Sci Rep. 2017;7(1):10200.\u003c/li\u003e\n \u003cli\u003eBerggrund M, Gustavsson I, Aarnio R, Lindberg JH, Sanner K, Wikstr\u0026ouml;m I, et al. Temporal changes in the vaginal microbiota in self-samples and its association with persistent HPV16 infection and CIN2. Virol J. 2020;17(1):147.\u003c/li\u003e\n \u003cli\u003eHolm JB, France MT, Gajer P, Ma B, Brotman RM, Shardell M, et al. Integrating compositional and functional content to describe vaginal microbiomes in health and disease. Microbiome. 2023;11(1):259.\u003c/li\u003e\n \u003cli\u003eFarr Zuend C, Lamont A, Noel-Romas L, Knodel S, Birse K, Kratzer K, et al. Increased genital mucosal cytokines in Canadian women associate with higher antigen-presenting cells, inflammatory metabolites, epithelial barrier disruption, and the depletion of L. crispatus. Microbiome. 2023;11(1):159.\u003c/li\u003e\n \u003cli\u003eMahajan G, Doherty E, To T, Sutherland A, Grant J, Junaid A, et al. Vaginal microbiome-host interactions modeled in a human vagina-on-a-chip. Microbiome. 2022;10(1):201.\u003c/li\u003e\n \u003cli\u003eGao Q, Fan T, Luo S, Zheng J, Zhang L, Cao L, et al. Lactobacillus gasseri LGV03 isolated from the cervico-vagina of HPV-cleared women modulates epithelial innate immune responses and suppresses the growth of HPV-positive human cervical cancer cells. Transl Oncol. 2023;35:101714.\u003c/li\u003e\n \u003cli\u003eBreedveld AC, Schuster HJ, van Houdt R, Painter RC, Mebius RE, van der Veer C, et al. Enhanced IgA coating of bacteria in women with Lactobacillus crispatus-dominated vaginal microbiota. Microbiome. 2022;10(1):15.\u003c/li\u003e\n \u003cli\u003eBerard AR, Brubaker DK, Birse K, Lamont A, Mackelprang RD, No\u0026euml;l-Romas L, et al. Vaginal epithelial dysfunction is mediated by the microbiome, metabolome, and mTOR signaling. Cell Rep. 2023;42(5):112474.\u003c/li\u003e\n \u003cli\u003eDelgado-Diaz DJ, Jesaveluk B, Hayward JA, Tyssen D, Alisoltani A, Potgieter M, et al. Lactic acid from vaginal microbiota enhances cervicovaginal epithelial barrier integrity by promoting tight junction protein expression. Microbiome. 2022;10(1):141.\u003c/li\u003e\n \u003cli\u003eŁaniewski P, Cui H, Roe DJ, Chase DM, Herbst-Kralovetz MM. Vaginal microbiota, genital inflammation, and neoplasia impact immune checkpoint protein profiles in the cervicovaginal microenvironment. NPJ Precis Oncol. 2020;4:22.\u003c/li\u003e\n \u003cli\u003eŁaniewski P, Barnes D, Goulder A, Cui H, Roe DJ, Chase DM, et al. Linking cervicovaginal immune signatures, HPV and microbiota composition in cervical carcinogenesis in non-Hispanic and Hispanic women. Sci Rep. 2018;8(1):7593.\u003c/li\u003e\n \u003cli\u003eSong J, Lang F, Zhao N, Guo Y, Zhang H. Vaginal Lactobacilli Induce Differentiation of Monocytic Precursors Toward Langerhans-like Cells: in Vitro Evidence. Front Immunol. 2018;9:2437.\u003c/li\u003e\n \u003cli\u003eBreshears LM, Edwards VL, Ravel J, Peterson ML. Lactobacillus Crispatus Inhibits Growth of Gardnerella Vaginalis and Neisseria Gonorrhoeae on a Porcine Vaginal Mucosa Model. BMC Microbiol. 2015; 15:276.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWitkin SS, Mendes-Soares H, Linhares IM, Jayaram A, Ledger WJ, Forney LJ. Influence of Vaginal Bacteria and D- and L-Lactic Acid Isomers on Vaginal Extracellular Matrix Metalloproteinase Inducer: Implications for Protection Against Upper Genital Tract Infections. mBio. 2013; 4(4): e00460\u0026ndash;13.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eZarate G, Nader-Macias ME. Influence of Probiotic Vaginal Lactobacilli on In Vitro Adhesion of Urogenital Pathogens to Vaginal Epithelial Cells. Lett Appl Microbiol. 2006; 43(2):174\u0026ndash;180.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eVallor AC, Antonio MA, Hawes SE, Hillier SL. Factors Associated With Acquisition of, or Persistent Colonization by, Vaginal Lactobacilli: Role of Hydrogen Peroxide Production. J Infect Dis. 2001; 184(11):1431\u0026ndash;1436.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eBorgogna JC, Shardell MD, Santori EK, Nelson TM, Rath JM, Glover ED, et al. The vaginal metabolome and microbiota of cervical HPV-positive and HPV-negative women: a cross-sectional analysis. BJOG. 2020;127(2):182-192.\u003c/li\u003e\n \u003cli\u003eXu H, Liu M, Song Y, Liu L, Xu F, Chen J, et al. Metabolomics variation profiling of vaginal discharge identifies potential targets for cervical cancer early warning. Acta Biochim Biophys Sin (Shanghai). 2022;54(10):1561-1565.\u003c/li\u003e\n \u003cli\u003eKurman RJ, Carcangiu ML, Herrington CS, Young RH. Tumours of the Cervix: squamous Cell Tumours and Precursors. In: Kurman JR, Carcangiu ML, Herrington CS, Young RH, eds. World Health Organization Classification of Tumours of the Female Reproductive Organs. 4th ed. Lyon: IARC Press. 2014; 172\u0026ndash;182.\u003c/li\u003e\n \u003cli\u003eCarozzi F, Gillio-Tos A, Confortini M, Del Mistro A, Sani C, De Marco L, et al. Risk of high-grade cervical intraepithelial neoplasia during follow-up in HPV-positive women according to baseline p16-INK4A results: a prospective analysis of a nested substudy of the NTCC randomised controlled trial. Lancet Oncol. 2013; 14(2):168-176.\u003c/li\u003e\n \u003cli\u003eWang Q, Song R, Zhao C, Liu H, Yang Y, Gu S, et al. HPV16 E6 promotes cervical cancer cell migration and invasion by downregulation of NHERF1. Int J Cancer. 2019;144(7):1619-1632.\u003c/li\u003e\n \u003cli\u003eLu Y, Chen Y, Zhang Z, Li M, Chen X, Tu K, et al. HPV16 E6 promotes cell proliferation, migration, and invasion of human cervical cancer cells by elevating both EMT and stemness characteristics. Cell Biol Int. 2022;46(4):599-610.\u003c/li\u003e\n \u003cli\u003eLebeau A, Bruyere D, Roncarati P, Peixoto P, Hervouet E, Cobraiville G, et al. HPV infection alters vaginal microbiome through down-regulating host mucosal innate peptides used by Lactobacilli as amino acid sources. Nat Commun. 2022;13(1):1076.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eŁaniewski P, Cui H, Roe DJ, Barnes D, Goulder A, Monk BJ, et al. Features of the cervicovaginal microenvironment drive cancer biomarker signatures in patients across cervical carcinogenesis. Sci Rep. 2019;9(1):7333.\u003c/li\u003e\n \u003cli\u003eAudirac-Chalifour A, Torres-Poveda K, Bahena-Rom\u0026aacute;n M, T\u0026eacute;llez-Sosa J, Mart\u0026iacute;nez-Barnetche J, Cortina-Ceballos B, et al. Cervical Microbiome and Cytokine Profile at Various Stages of Cervical Cancer: A Pilot Study. PLoS One. 2016;11(4):e0153274.\u003c/li\u003e\n \u003cli\u003eLi X, Ding Y, Zi M, Sun L, Zhang W, Chen S, et al. CD19, from bench to bedside. Immunol Lett. 2017;183:86-95.\u003c/li\u003e\n \u003cli\u003eHarton J, Jin L, Hahn A, Drake J. Immunological Functions of the Membrane Proximal Region of MHC Class II Molecules. F1000Res. 2016;5:F1000 Faculty Rev-368.\u003c/li\u003e\n \u003cli\u003eShao A, Owens DM. The immunoregulatory protein CD200 as a potentially lucrative yet elusive target for cancer therapy. Oncotarget. 2023;14:96-103.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eG\u0026oacute;mez-Melero S, Caballero-Villarraso J. CCR6 as a Potential Target for Therapeutic Antibodies for the Treatment of Inflammatory Diseases. Antibodies (Basel). 2023;12(2):30.\u003c/li\u003e\n \u003cli\u003eHubert P, Herman L, Maillard C, Caberg JH, Nikkels A, Pierard G, et al. Defensins induce the recruitment of dendritic cells in cervical human papillomavirus-associated (pre)neoplastic lesions formed in vitro and transplanted in vivo. FASEB J. 2007;21(11):2765-75.\u003c/li\u003e\n \u003cli\u003eHubert P, Herman L, Roncarati P, Maillard C, Renoux V, Demoulin S, et al. Altered \u0026alpha;-defensin 5 expression in cervical squamocolumnar junction: implication in the formation of a viral/tumour-permissive microenvironment. J Pathol. 2014;234(4):464-77.\u003c/li\u003e\n \u003cli\u003eWolf R, Howard OM, Dong HF, Voscopoulos C, Boeshans K, Winston J, et al. Chemotactic activity of S100A7 (Psoriasin) is mediated by the receptor for advanced glycation end products and potentiates inflammation with highly homologous but functionally distinct S100A15. J Immunol. 2008;181(2):1499-506.\u003c/li\u003e\n \u003cli\u003eCai K, Chen S, Zhu C, Li L, Yu C, He Z, et al. FOXD1 facilitates pancreatic cancer cell proliferation, invasion, and metastasis by regulating GLUT1-mediated aerobic glycolysis. Cell Death Dis. 2022;13(9):765.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eKumegawa K, Yang L, Miyata K, Maruyama R. FOXD1 is associated with poor outcome and maintains tumor-promoting enhancer-gene programs in basal-like breast cancer. Front Oncol. 2023;13:1156111.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePrasad P, Kannan B, Sriram G, Jaber M, Khair AMB, Ramasubramanian A, et al. Waterpipe smoke condensate induces epithelial-mesenchymal transformation and promotes metastasis of oral cancer by FOXD1 expression. J Stomatol Oral Maxillofac Surg. 2024;101900.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eLiu X, Min S, Zhang Q, Liu Y, Zou Z, Wang N, et al. Prognostic and clinicopathological significance of FOXD1 in various cancers: a meta and bioinformation analysis. Future Sci OA. 2024;10(1):FSO901.\u003c/li\u003e\n \u003cli\u003eLiu Z, Zhang Z, Qiu L, Zhang F, Xu X, Wei H, et al. Characterization and bioactivities of the exopolysaccharide from a probiotic strain of Lacto- bacillus plantarum WLPL04. J Dairy Sci. 2017;100:6895\u0026ndash;6905.\u003c/li\u003e\n \u003cli\u003eAllonsius CN, Vandenheuvel D, Oerlemans EFM, Petrova MI, Donders GGG, Cos P, et al. Inhibition of Candida albicans morphogenesis by chitinase from Lactobacillus rhamnosus GG. Sci Rep. 2019;9:2900.\u003c/li\u003e\n \u003cli\u003ePetrucca A, Cipriani P, Sessa R, Teggi A, Pustorino R, Santapaola D, et al. Burkholderia cenocepacia vaginal infection in patient with smoldering myeloma and chronic hepatitis C. Emerg Infect Dis. 2004;10(11):1957-9.\u003c/li\u003e\n \u003cli\u003eAmabebe E, Richardson LS, Bento GFC, Radnaa E, Kechichian T, Menon R, et al. Ureaplasma parvum infection induces inflammatory changes in vaginal epithelial cells independent of sialidase. Mol Biol Rep. 2023;50(4):3035-3043.\u003c/li\u003e\n \u003cli\u003eDoshi MB, Lee N, Tseyang T, Ponomarova O, Goel HL, Spears M, et al. Disruption of sugar nucleotide clearance is a therapeutic vulnerability of cancer cells. Nature. 2023;623(7987):625-632.\u003c/li\u003e\n \u003cli\u003eWang X, Liu R, Zhu W, Chu H, Yu H, Wei P, et al. UDP-glucose accelerates SNAI1 mRNA decay and impairs lung cancer metastasis. Nature. 2019;571(7763):127-131.\u003c/li\u003e\n \u003cli\u003eGao Q, Cheng B, Chen C, Lei C, Lin X, Nie D, et al. Dysregulated glucuronic acid metabolism exacerbates hepatocellular carcinoma progression and metastasis through the TGF\u0026beta; signalling pathway. Clin Transl Med. 2022;12(8):e995.\u003c/li\u003e\n \u003cli\u003eOkada M, Yamawaki H. A current perspective of canstatin, a fragment of type IV collagen alpha 2 chain. J Pharmacol Sci. 2019;139(2):59-64.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSorvari T, Sarnesto A, Syrj\u0026auml;nen K. Type IV collagen in the basal membrane of human papillomavirus associated premalignant and malignant squamous cell lesions of the uterine cervix. Gynecol Obstet Invest. 1988;26(4):324-31.\u003c/li\u003e\n \u003cli\u003eStewart CJ, McNicol AM. Distribution of type IV collagen immunoreactivity to assess questionable early stromal invasion. J Clin Pathol. 1992;45(1):9-15.\u003c/li\u003e\n \u003cli\u003ePitt MA, Hale RJ, Buckley CH. The distribution of type IV collagen in invasive carcinoma of the uterine cervix. Histopathology. 1992;20(2):139-43.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eHirota J, Yoneda K, Osaki T. Basement membrane type IV collagen in oral squamous cell carcinoma. Head Neck. 1990;12(5):400-5.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eTong CCL, Koptyra M, Raman P, Rathi KS, Choudhari N, Lin X, et al. Targeted gene expression profiling of inverted papilloma and squamous cell carcinoma. Int Forum Allergy Rhinol. 2022;12(2):200-209.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWang X, Jiang Z, An J, Mao X, Lin F, Sun P. Effect of a synthetic inhibitor of urokinase plasminogen activator on the migration and invasion of human cervical cancer cells in vitro. Mol Med Rep. 2018;17(3):4273-4280.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMa Y, Wu T, Zhou H, He G, Li Y, Wang B, et al. Canstatin represses glioma growth by inhibiting formation of VM-like structures. Transl Neurosci. 2021;12(1):309-319.\u003c/li\u003e\n \u003cli\u003eZhang F, Duan J, Song H, Yang L, Zhou M, Wang X. Combination of canstatin and arsenic trioxide suppresses the development of hepatocellular carcinoma. Drug Dev Res. 2021;82(3):430-439.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eFrance MT, Fu L, Rutt L, Yang H, Humphrys MS, Narina S, et al. Insight into the ecology of vaginal bacteria through integrative analyses of metagenomic and metatranscriptomic data. Genome Biol. 2022;23(1):66.\u003c/li\u003e\n \u003cli\u003eBrotman RM, Shardell MD, Gajer P, Fadrosh D, Chang K, Silver MI, et al. Association between the vaginal microbiota, menopause status, and signs of vulvovaginal atrophy. Menopause. 2018;25(11):1321-1330.\u003c/li\u003e\n \u003cli\u003eMa B, France MT, Crabtree J, Holm JB, Humphrys MS, Brotman RM, et al. A comprehensive non-redundant gene catalog reveals extensive within-community intraspecies diversity in the human vagina. Nat Commun. 2020;11(1):940.\u003c/li\u003e\n \u003cli\u003eDai W, Du H, Zhou Q, Li S, Wang Y, Hou J, et al. Metabolic profiles outperform the microbiota in assessing the response of vaginal microenvironments to the changed state of HPV infection. NPJ Biofilms Microbiomes. 2024;10(1):26.\u003c/li\u003e\n \u003cli\u003eWaxman AG, Chelmow D, Darragh TM, Lawson H, Moscicki AB. Revised terminology for cervical histopathology and its implications for management of high-grade squamous intraepithelial lesions of the cervix. Obstet Gynecol. 2012;120(6):1465-1471.\u003c/li\u003e\n \u003cli\u003eDarragh TM, Colgan TJ, Cox JT, Heller DS, Henry MR, Luff RD, et al. The Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions: background and consensus recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. J Low Genit Tract Dis. 2012;16(3):205-42.\u003c/li\u003e\n \u003cli\u003eBolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37(8):852-857.\u003c/li\u003e\n \u003cli\u003eMarcel M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 2011;17:10\u0026ndash;12.\u003c/li\u003e\n \u003cli\u003eKim D, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with low memory requirements. Nat Methods. 2015;12(4):357-60.\u003c/li\u003e\n \u003cli\u003eLi B, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12:323.\u003c/li\u003e\n \u003cli\u003eLove MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550.\u003c/li\u003e\n \u003cli\u003eSherman BT, Hao M, Qiu J, Jiao X, Baseler MW, Lane HC, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50(W1):W216-W221.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Human papillomavirus, Cervical intraepithelial neoplasia, Vaginal microbiota, Metabolome, Host transcriptome","lastPublishedDoi":"10.21203/rs.3.rs-4717221/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4717221/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eVaginal \u003cem\u003eLactobacillus\u003c/em\u003e has been implicated in modulating the risk of cervical intraepithelial neoplasia (CIN) progression. However, there remains a gap in population-based studies elucidating the underlying mechanisms that link \u003cem\u003eLactobacillus\u003c/em\u003e with CIN progression and carcinogenesis.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eTo address this knowledge gap, we conducted an in-depth analysis of vaginal microbiota (VM), metabolome, and host transcriptome profiles in a cohort of 75 Chinese women, stratified into two groups based on their CIN status: low-grade CIN1 (n\u0026thinsp;=\u0026thinsp;38) and high-grade CIN2+ (n\u0026thinsp;=\u0026thinsp;37).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eOur findings revealed that samples dominated by \u003cem\u003eLactobacillus\u003c/em\u003e were more prevalent in the CIN1 cohort. Furthermore, the vaginal metabolome displayed a significant interplay with the microbiota, with \u003cem\u003eLactobacillus\u003c/em\u003e emerging as a key influencer. Among the 100 metabolites that distinguished the CIN1 and CIN2\u0026thinsp;+\u0026thinsp;cohorts, 26 were inversely correlated with \u003cem\u003eLactobacillus\u003c/em\u003e levels, including L-Carnitine and UDP-D-glucose. Conversely, five metabolites, such as Succinic anhydride, exhibited a positive correlation with \u003cem\u003eLactobacillus\u003c/em\u003e abundance. Differential gene expression analysis revealed 176 genes upregulated in the CIN1 cohort compared to the CIN2\u0026thinsp;+\u0026thinsp;cohort, primarily related to immune responses and negative regulation of cell migration. Notably, COL4A2 and CCBE1, both negatively correlated with L-Carnitine, were among the upregulated genes. Conversely, 82 genes were downregulated in the CIN1 cohort, including TP63 and FOXD1, which positively correlated with UDP-D-glucose. Further mediation analysis suggested that L-Carnitine plays a crucial role in mediating the positive association between \u003cem\u003eLactobacillus\u003c/em\u003e and COL4A2 expression, both of which are enriched in the CIN1 cohort. Similarly, UDP-D-glucose emerged as a mediator in the negative association between \u003cem\u003eLactobacillus\u003c/em\u003e and FOXD1, a gene depleted in the CIN1 cohort.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThese findings provide insights into the complex interplay between vaginal \u003cem\u003eLactobacillus\u003c/em\u003e, the metabolome, and host gene expression patterns associated with CIN progression. The identified \u003cem\u003eLactobacillus\u003c/em\u003e:L-Carnitine:COL4A2 and \u003cem\u003eLactobacillus\u003c/em\u003e:UDP-D-glucose:FOXD1 regulatory axes underscore the potential significance of these pathways in modulating CIN risk. These population-based discoveries hold promise for future research aimed at developing targeted interventions to prevent or delay CIN progression.\u003c/p\u003e","manuscriptTitle":"Co-altered vaginal Lactobacillus, metabolome and host gene expression associate with the grade of cervical intraepithelial neoplasia in Chinese women","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-12 10:23:30","doi":"10.21203/rs.3.rs-4717221/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9bb087a9-c787-4609-9ccb-529421db58df","owner":[],"postedDate":"August 12th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-01-30T15:53:49+00:00","versionOfRecord":[],"versionCreatedAt":"2024-08-12 10:23:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4717221","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4717221","identity":"rs-4717221","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}