Altered DNA methylation in ion transport and immune signalling genes is associated with severity in Pancreatic Ductal Adenocarcinoma

Background: Pancreatic ductal adenocarcinoma (PDAC) is one of the leading cancers worldwide and has a poor survival, with a relative five-year survival rate of only 8.5%. In this study we investigated epigenetic marks associated with PDAC severity and prognosis, through studying alterations in DNA methylation patterns. Methods: DNA methylome for tumor and adjacent normal tissue samples from PDAC patients (n=7) were generated using Illumina 450K bead chips. Differentially methylated positions (DMPs) were identified with |delta beta| > 0.2 and p-value<0.01 by comparing tumors with the adjacent normal tissues. Validation of differential methylation and associated gene expression at selected genes was carried out in an independent cohort PDAC patient. Results: We identified 76 DMPs in PDAC patients that mapped to 43 genes. Among them, 44.7% (n=34) were hypo-methylated and 55.3% (n=42) were hyper-methylated DMPs in cancer samples. The trends of change in methylation at these 76 DMPs from well to moderate were like that from moderate to poorly differentiated cancer samples. The gradual trend in differential methylation was observed both in our cohort and the TCGA-PAAD cohort, suggesting methylation marks can serve as early indicators of disease pathology. Altered promoter methylation, which may affect gene expression, was observed for transcription regulators (BHLHE23, GSC2, FOXE1 and TWIST1), gated ion channels (KCNA6, and CACNB2), tumor suppressors (RASSF1, SPRED2, and NPY) and genes functioning in interferon signalling (SIGIRR, MX2, and OAS2). We also have compared the TCGA-PAAD dataset with normal pancreatic tissue data from GTEx V8 dataset leading to a confluent observation. Conclusions: We reported the first study on methylome in PDAC tumors from patients in India. We identified altered DNA methylation associated with increasing severity in PDAC among some genes like SIGIRR, MX2 along with other previously reported loci. We also concluded a confluence in our observation when comparing the TCGA-PAAD dataset with GTEx V8 dataset.