Comparative Tumor Microenvironment Analysis for HCC and PDAC Using KMplotter

Hepatocellular carcinoma (HCC) and pancreatic ductal adenocarcinoma (PDAC) are highly lethal cancers marked by profound epigenetic and metabolic reprogramming. Among the candidate biomarkers, the DNA methyltransferase DNMT3A and the guanine monophosphate synthetase (GMPS) have emerged as potential prognostic drivers, yet their roles across tumor contexts remain unclear. Here, we demonstrate the application of KMplotter to interrogated pan-cancer transcriptomic and survival datasets encompassing over 7000 patients, complemented by expression profiling of normal, tumor, and metastatic tissues, and integrated tumor microenvironment (TME) analyses. Elevated DNMT3A and GMPS expression correlated with worse overall survival in HCC, particularly in Asian patients, while in PDAC, high DNMT3A but low GMPS expression predicted favorable outcomes. Both genes were consistently upregulated in tumors relative to normal tissues, with further increases in metastatic HCC. Immune deconvolution revealed that DNMT3A was linked to Th2/Treg-enriched niches, whereas GMPS overexpression coincided with high mutational burden or stromal enrichment, fostering immunosuppressive microenvironments. Comparative analysis of toll-like receptor signatures highlighted divergent antigen-sensing pathways, with HCC reflecting viral-driven immune exhaustion and PDAC showing self-antigen-associated signaling. Collectively, these findings position DNMT3A and GMPS as context-dependent biomarkers that integrate metabolic and immune cues to shape prognosis in liver and pancreatic cancer, offering mechanistic insight and translational relevance for patient stratification.