Immune remodeling in the cervicovaginal microenvironment associated with HPV status, vaginal community state type, and cytological grade
HPV-positive samples show reduced IL-4 and IL-17E/IL-25 compared with HPV-negative samples, but the immune profile varies non-monotonically across microbial community state types and cytological categories, without linear progression with lesion severity.
| Population | Women with cervicovaginal samples stratified by HPV status, vaginal community state type (CST I–V), and cytological category (NILM, LSIL, ASC-H) |
|---|---|
| Exposure | Characterization of the cervicovaginal immune microenvironment (cytokine/chemokine panel) combined with multi-kingdom metagenomic analysis (bacteria, viruses, fungi) using CLR |
| Comparator | Comparisons between groups: HPV+ vs. HPV−; CSTs against each other; LSIL and ASC-H vs. NILM |
| Outcome | Cervicovaginal IL-4 (HPV+ vs HPV−); Cervicovaginal IL-17E/IL-25 (HPV+ vs HPV−); IFN-α2 by vaginal community state type (CST II vs others); Cytokine profile in LSIL vs NILM; Immune profile in ASC-H vs LSIL/NILM; Phocaeicola–IL-10 association (after global BH-FDR correction); Rountreeviridae–MIP-1β association (after global BH-FDR correction) |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| Cervicovaginal IL-4 (HPV+ vs HPV−) | reduced in HPV+ vs HPV−; exact effect size and 95% CI not reported | — | Low | — | 1 studies |
| Cervicovaginal IL-17E/IL-25 (HPV+ vs HPV−) | reduced in HPV+ vs HPV−; exact effect size and 95% CI not reported | — | Low | — | 1 studies |
| IFN-α2 by vaginal community state type (CST II vs others) | relatively elevated in CST II (L. gasseri-dominant); exact effect size and 95% CI not reported | — | Low | — | 1 studies |
| Cytokine profile in LSIL vs NILM | lower IL-4, IFN-γ, IL-17F, PDGF-AB/BB, IL-17E/IL-25, PDGF-AA in LSIL vs NILM; exact effect sizes and 95% CI not reported | — | Low | — | 1 studies |
| Immune profile in ASC-H vs LSIL/NILM | distinct non-progressive pattern in ASC-H; exact effect size and 95% CI not reported | — | Low | — | 1 studies |
| Phocaeicola–IL-10 association (after global BH-FDR correction) | significant after global BH-FDR correction; direction and effect size not reported | — | Low | — | 1 studies |
| Rountreeviridae–MIP-1β association (after global BH-FDR correction) | significant after global BH-FDR correction; direction and effect size not reported | — | Low | — | 1 studies |
Context
Persistent high-risk HPV infection is the principal cause of cervical cancer, yet not all infections progress to severe dysplasia. The local immune and microbial microenvironment may modulate this risk, and systematic characterization is needed to identify progression biomarkers. Prior studies assessed cytokines or microbiome separately; this analysis integrates both in the same sample set.
What the study showed
HPV-positive samples showed reduced IL-4 and IL-17E/IL-25 compared with HPV-negative samples (exact magnitudes not reported in the available text). IFN-α2 was relatively elevated in CST II (L. gasseri-dominant) and IL-3 was specifically reduced in CST V (L. jensenii-dominant), in a non-monotonic pattern across CSTs. LSIL showed lower IL-4, IFN-γ, IL-17F, PDGF-AB/BB, IL-17E/IL-25, and PDGF-AA relative to NILM, while ASC-H exhibited a distinct rather than uniformly progressive immune profile. After global BH-FDR correction, only two taxon–cytokine associations remained significant: Phocaeicola–IL-10 and Rountreeviridae–MIP-1β.
How it was done
Cross-sectional observational study published in 2026 in Frontiers in Cellular and Infection Microbiology. Cervicovaginal samples were analyzed for cytokine/chemokine profiles and multi-kingdom microbiome (bacteria, viruses, fungi) by sequencing with CLR analysis. Taxon–cytokine associations were tested with prespecified prevalence and abundance filtering and BH-FDR correction (local and global). Exact sample size was not reported in the available text.
Effect magnitude
Effect sizes (RR, OR, SMD, MD) and confidence intervals were not reported in the available text; findings are described as statistically significant associations after BH-FDR correction, without absolute numerical values for most markers.
Risk of bias
Cross-sectional design precludes causal inference about the direction of immune–microbiome associations. Sequencing depth for viruses and fungi was substantially lower than for bacteria, making viral and fungal findings explicitly exploratory per the authors. Sample size was not reported in the abstract, limiting power assessment. No risk-of-bias tool was mentioned; confounders such as sexual behavior, antibiotic use, and hormonal status were not discussed in the available text.
What this study does NOT prove
This study does not prove causality between immune profile, microbial composition, and cervical lesion progression. Findings cannot be generalized to populations with different HPV prevalence, microbial profiles, or sociodemographic characteristics distinct from the studied sample.
In clinical practice
Findings do not support immediate changes in clinical management. Cervicovaginal immune profiling is not monitored in standard practice, and no clinical thresholds have been established for the studied markers. Clinicians should await longitudinal studies confirming whether identified associations predict lesion progression or regression.
Limitations
Cross-sectional design precludes causal inference about the direction of immune–microbiome associations. Sequencing depth for viruses and fungi was substantially lower than for bacteria, making viral and fungal findings explicitly exploratory per the authors. Sample size was not reported in the abstract, limiting power assessment. No risk-of-bias tool was mentioned; confounders such as sexual behavior, antibiotic use, and hormonal status were not discussed in the available text.
What is still missing
Longitudinal studies are needed to determine whether immune differences identified here precede cytological progression or are a consequence of it. Replication in independent cohorts with adequate sample size and confounder analysis is essential before any translational application.
Technical appendix
Version history
- 1.0 · 2026-11-21 — Auto-generated under Evidence Standard v1.0
