Numerous studies support that Human papillomavirus (HPV) can cause cervical cancer. However, few studies have surveyed the heterogeneity of HPV infected or uninfected (HPV+ and HPV-) cervical cancer (CESC) patients. Integration of scRNA-seq and TCGA data to analyze the heterogeneity of HPV+ and HPV- cervical cancer patients on a single-cell level could improve understanding of the cellular mechanisms during HPV-induced cervical cancer.
CESC scRNA-seq data obtained from the Gene Expression Omnibus (GEO) database and the Seurat, Monocle3 package were used for scRNA-seq data analysis. The ESTIMATE package was used for single-sample gene immune score, CIBERSORT package was used to identify immune scores of cells, and the “WGCNA” package for the weighted correlation network analysis. Univariate Cox and LASSO regression were performed to establish survival and relapse signatures. KEGG and GO analyses were performed for the signature gene. Gene Expression Profiling Interactive Analysis was used for Pan-cancer analysis.
In the HPV+ CESC group, CD8+ T cells and B cells were down-regulated, whereas T reg cells, CD4+ T cells, and epithelial cells were up-regulated according to scRNA-seq data. Survival analysis of TCGA-CESC revealed that increased expression of naive B cells or CD8+ T cells favors the survival probability of CESC patients. WGCNA, univariate Cox, and LASSO Cox regression established a 9-genes survival signature and a 7-gene relapse model. Pan-cancer analysis identified IKZF3, FOXP3, and JAK3 had a similar distribution and effects in HPV-associated HNSC.
Analysis of scRNA-seq and bulk RNA-seq of HPV+ and HPV- CESC samples revealed heterogeneity from transcriptional state to immune infiltration. Survival and relapse models were adjusted according to the heterogeneity of HPV+ and HPV- CESC immune cells to assess the prognostic risk accurately. Hub genes represent similar protection in HPV- associated HNSC while showing irrelevant to other potential HPV-related cancers.