Establishment of a Prognostic Stemness Signature for Cervical Squamous Cell Carcinoma

Abstract

Background: The aim of this study was to construct a robust stemness-related gene signature for predicting prognosis of cervical squamous cell carcinoma (CSCC).Methods: Expression data for the PCBC database-derived pluripotent stem cell (PSC) samples were collected using the one-class logistic regression (OCLR) method to calculate stemness indexes (mRNAsi) of samples derived from the TCGA dataset. Functions of possible mRNAsi-related stemness genes extracted through WGCNA were then examined by enrichment analysis. Most representative stemness genes for prognosis prediction were screened to construct a stemness-related gene signature by shrinkage estimate and univariate analysis. Next, the TCGA dataset and the GSE44001 external dataset were incorporated into that model and classified to evaluate the model efficiency and stability in patient prognosis prediction and classification according to the Riskscore. The associations between the Riskscore and clinical characteristics as well as relevant signaling pathways were also explored. Moreover, the prognosis predicting efficiency of the stemness-related gene signature was compared with those of CSCC prognostic signatures reported in other studies.Results: According to the findings, mRNAsi showed significant correlation with key oncogene mutation degrees (including DMD, KMT2C, EP300 and MUC4), infiltrating stroma cells and the CIMP classification for CSCC cases. The 8-stemness gene signature in this study achieved high stability and accuracy in prognosis prediction for CSCC cases. In the meanwhile, the model provided diverse therapeutic targets to precisely treat CSCC in clinical practice based on various subtype-specific stemness genes.Conclusion: Our present study suggested that the 8 stemness gene signature can help to screen out novel stem-related diagnostic indicators, therapeutic targets and prognostic predictors in CSCC.