Molecular Characterization of HPV Integrations in High-risk HPV- positive Women and Constructing an Integration Gene-Based Prognostic Risk Model for Cervical Cancer

Abstract

Abstract Background Cervical cancer (CC) ranks fourth in global cancer incidence and represents the second leading cause of cancer deaths. HPV integration into the human genome is a key molecular event in the progression of cancer carcinogenesis. Methods This study utilized the high-risk HPV-positive women samples to investigate the HPV integration through the high-throughput Viral Integration Detection (HIVID). Subsequently, the expression of recurrent integration genes at both the mRNA and protein level was determined in GEIPA, and the Human Protein Atlas (HPA). Furthermore, the prognostic risk model of HPV integration genes was constructed using the univariate Cox, least absolute shrinkage and selection operator (LASSO) regression and multivariate Cox analysis from the Cancer Genome Atlas (TCGA) datasets. Then, the prognostic risk model was evaluated with Kaplan-Meier (KM) survival curves and receiver operating characteristic (ROC) curves. The CC patients in TCGA were stratified into high and low risk groups and the differences in PI3K-AKT pathway gene enrichment and immunotherapeutic response differences were analyzed between the two subgroups. Finally, the deferentially expressed immune-related genes (immune DEGs) between the high-risk and low-risk groups were subjected to functional enrichment analysis. Results The overall HPV integration rate in high-risk HPV-positive women was 8.75%. We identified 467 integration sites in 82 out of 937 samples. We detected 28 recurrent genes for virus integration in 318 integration genes (reads > = 6), the most frequent were KLF5 (n = 5), LINC00392 (n = 5), BCL11B (n = 3) and TP63 (n = 3).