Comprehensive analysis reveals an arachidonic acid metabolism-related gene signature in patients with pancreatic ductal adenocarcinoma

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is highly heterogeneous, making its prognosis prediction difficult. The arachidonic acid (AA) cascade is involved in carcinogenesis. Therefore, the metabolic enzymes of the AA cascade consist of lipoxygenases (LOXs), phospholipase A2s (PLA2s), and cyclooxygenases (COXs) along with their metabolic products, including leukotrienes. Nevertheless, the prognostic potential of AA metabolism-associated PDAC has not been explored. Herein, the mRNA expression patterns and the matching clinical information of individuals with PDAC were abstracted from online data resources. We employed the LASSO Cox regression model to develop a multigene clinical signature in the TCGA queue. The GEO queue and the ICGC queue were employed as the validation queue. There was differential expression of a significant number of AA metabolism-associated genes (56.8%) between PDAC and neighboring nonmalignant tissues in the TCGA queue. Univariate Cox regression demonstrated that 13 of the differentially expressed genes (DEGs) were linked to overall survival (OS) (p < 0.05). A 6-gene clinical signature was developed for stratifying the PDAC patients into two risk groups, with the high-risk group patients exhibiting remarkably lower OS than the low-risk group patients (p < 0.001 in the TCGA data set and the ICGC queue, and p = 0.001 in the GEO data set). The multivariate Cox data revealed the risk score as an independent OS predictor (HR > 1, p < 0.01). The receiver operating characteristic (ROC) curve verified the predictive potential of our signature. The expression and alteration of the six genes in PDAC were also validated using online databases. Functional analyses demonstrated that immune-linked cascades were enriched, and the immune status was remarkably different between the high- and low-risk groups. In summary, an AA metabolism-associated clinical gene signature can be applied for prognostic estimation in PDAC.