Obesity promotes pancreatic cancer through the ECM-PI3K-AKT signaling axis

Abstract Objective To investigate obesity signature genes as potential biomarkers of pancreatic cancer and their possible molecular regulatory mechanisms. Methods Based on the mRNA-Seq data of obesity and pancreatic cancer from GEO and TCGA-GTEx databases, we screened four obesity signature genes using a random forest tree algorithm and analyzed their expression in pancreatic cancer, and constructed a nomogram risk prediction model and performed preliminary validation of the predictive efficiency of the model. Then the tumor immune profile of pancreatic cancer was assessed based on the CIBERSORT algorithm, and the correlation between the four obesity genes and the abundance of tumor immune cell infiltration, and immune checkpoints in pancreatic cancer was discussed. Results The diagnostic ROC curves suggested that the four obesity genes showed good diagnostic efficacy for pancreatic cancer, hinting that they may be potential predictive targets for pancreatic cancer, and the pancreatic cancer risk prediction model based on the four obesity-related genes showed good risk prediction efficacy for pancreatic cancer in the obese population, which has good potential for clinical application. The study also found that the upregulation of COL1A2 expression may activate the PI3K/ART signaling pathway through the mediation of integrin IGTA to promote the development of pancreatic cancer, closely associated with the developm- ent and progression of pancreatic cancer. Conclusion The obesity signature genes CCDC80, COL1A2, DPYSL3 and BCAT1 are potential predictive targets for new pancreatic cancer in obese people, and the upregulation of COL1A2 expression may promote pancreatic cancer development by activating the PI3K/ART signaling pathway.