Abstract
Perineural invasion (PNI) is a prominent characteristic of pancreatic ductal adenocarcinoma (PDAC). PNI is associated with tumor progression, local recurrence and neuropathic pain; therefore, the identification of biomarkers associated with PNI may be beneficial in assessing the prognosis for patients with PDAC. Using an in vivo model of PNI, five pancreatic cancer cell lines (PANC-1, CFPAC-1, CAPAN-2, SW1990 and ASPC-1) were divided into two groups: High-(comprising PANC-1, CFPAC-1 and CAPAN-2) and low PNI (comprising SW1990 and ASPC-1). Differentially expressed genes (DEGs) between the two groups were identified using the GSE26088 dataset, and were regarded as PNI-associated genes. A total of 445 DEGs associated with PNI (fold change >1.5 or <0.66; P<0.05) were identified, which included 176 up- and 269 downregulated genes. Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis and function annotation were performed, and the NetworkAnalyst database was used for protein-protein interaction network analysis to identify hub genes. A total of 20 hub genes (gene degree, ≥6) were identified. PNI was associated with the function 'chemokine signaling pathway'. The DEGs and hub genes were validated using the GSE102238 dataset and clinical tissue microarrays. Fibroblast growth factor 2 (FGF2) and catenin α 2 were demonstrated to be associated with PNI using the GSE102238 dataset. Furthermore, clinical tissue microarray analysis demonstrated that FGF2 was associated with PNI and poor prognosis. The present study provided a potential method for the reliable identification of PNI-associated genes, although further investigation is required to validate these results.