Molecular Subtype Identification and Prognostic Prediction of Pancreatic Cancer Based on m6A / m5C / m1A

ABSTRACT

Pancreatic cancer, characterized by an unfavourable prognosis, necessitates early diagnosis and prompt therapeutic intervention. This study aimed to delineate the functional implications of m6A/m1A/m5C‐related genes in pancreatic carcinogenesis and establish a prognostic model. Utilizing data from the TCGA‐PAAD cohort (n=178) combined with GTEx normal pancreatic tissues (n=332), we identified 43 genes associated with m ⁶ A/m ¹ A/m ⁵ C methylation pathways. Significant expression differences in these genes were observed between neoplastic and non‐neoplastic tissues, correlating with multiple biological pathways. Consensus clustering divided pancreatic cancer patients into two distinct molecular subtypes exhibiting pronounced survival differences. A prognostic risk‐scoring model was developed based on nine m ⁶ A/m ¹ A/m ⁵ C‐related genes, demonstrating efficacy in predicting patient outcomes. We constructed a nomogram integrating clinical variables with the risk score to enhance prognostic precision. Among these genes, a comprehensive investigation was conducted using immunohistochemistry and multicolor immunofluorescence to elucidate that higher TRMT61B expression in tumour tissues is significantly associated with higher pancreatic cancer stage, and might relate to an immunosuppressive and dysfunctional tumour immune microenvironment. This study reveals the role of m ⁶ A/m ¹ A/m ⁵ C associated genes, especially TRMT61B, in the epigenetic regulation of pancreatic cancer, providing evidence supporting their potential as novel prognostic biomarkers.