While genetic alterations in several regulators of the cell cycle have a significant impact on the gastric carcinogenesis process, the prognostic role of them remains to be further elucidated. The TCGA-STAD training set were downloaded and the mRNA expression matrix of cell cycle genes was extracted and corrected for further analysis after taking the intersection with GSE84437 dataset. Differentially expressed mRNAs were identified between tumor and normal tissue samples in TCGA-STAD. Univariate Cox regression analysis and lasso Cox regression model established a novel seven-gene cell cycle signature (including GADD45B, TFDP1, CDC6, CDC25A, CDC7, SMC1A and MCM3) for GC prognosis prediction. Patients in the high-risk group shown significantly poorer survival than patients in the low-risk group. The signature was found to be an independent prognostic factor for GC survival. Nomogram including the signature shown some clinical net benefit for overall survival prediction. The signature was further validated in the GSE84437 dataset. In tissue microarray, CDC6 and MCM3 protein expression were significant differences by the immunohistochemistry-based H-score between tumor tissues and adjacent tissues, and CDC6 is an independent prognostic factor for GC. Interestingly, our GSEA revealed that low-risk patients were more related to cell cycle pathways and might benefit more from therapies targeting cell cycle. Our study identified a novel robust seven-gene cell cycle signature for GC prognosis prediction that may serve as a beneficial complement to clinicopathological staging. The signature might provide potential biomarkers for the application of cell cycle regulators to therapies and treatment response prediction.
Keywords: GEO; TCGA; cell cycle; gastric cancer; prognostic model.