Single-cell transcriptomics uncovers HPV-driven immune evasion and establishes a TAB2-based prognostic signature in cervical cancer

Background Persistent human papillomavirus (HPV) infection is a key driver in cervical cancer, yet the molecular mechanisms linking HPV to immune escape remain incompletely understood. Methods We integrated single-cell RNA sequencing (scRNA-seq), bulk transcriptome, and clinical data to analyze cervical cancer heterogeneity. scRNA-seq, inferCNV, and pseudotime analysis characterized HPV-associated cellular features and malignant transformation. hdWGCNA identified core gene modules. Univariate Cox, Lasso, and multivariate Cox regressions were used to construct a prognostic risk model. In vitro functional validation of TAB2 was also performed. Results Eleven major cell populations were identified, with epithelial cells as primary HPV targets. HPV-related gene enrichment increased with disease progression. Malignant cells exhibited significantly higher immune escape scores than non-tumor cells, strongly correlating with HPV infection levels (P < 0.001). hdWGCNA identified three modules linked to immunosuppression, stress response, and proliferation. An eight-gene risk model showed stable prognostic performance across TCGA and GEO cohorts, served as an independent prognostic factor (P < 0.05), and correlated with WNT/TGFβ pathway activation and immune checkpoint genes. TAB2 knockdown significantly suppressed cervical cancer cell proliferation, migration, and invasion in vitro. Conclusion This multi-omics analysis identifies transcriptional programs associated with HPV infection and immune escape in cervical cancer. The eight-gene risk model shows potential for prognosis prediction and patient stratification, offering a theoretical basis for individualized treatment. The mechanistic hypotheses generated warrant further experimental validation.