Multi-Omics Profiling Reveals Glycerolipid Metabolism-Associated Molecular Subtypes and Identifies ALDH2 as a Prognostic Biomarker in Pancreatic Cancer

Abstract

Background: The reprogramming of lipid metabolism, especially glycerolipid metabolism (GLM), plays a key role in cancer progression and response to therapy. However, the role and molecular characterization of GLM in pancreatic cancer (PC) remain unclear. Methods: A pan-cancer analysis of glycerolipid metabolism-related genes (GMRGs) was first conducted to assess copy-number variants, single-nucleotide variations, methylation, and mRNA expression. Subsequently, GLM in PC was characterized using lipidomics, single-cell RNA sequencing (scRNA-seq), and spatial transcriptomic analysis. A cluster analysis based on bulk RNA sequencing data from 930 PC samples identified GLM-associated subtypes, which were then analyzed for differences in prognosis, biological function, immune microenvironment, and drug sensitivity. To prioritize prognostically relevant GMRGs in PC, we employed a random forest (RF) algorithm to rank their importance across 930 PC samples. Finally, the key biomarker of PC was validated using PCR and immunohistochemistry. Results: Pan-cancer analysis identified molecular features of GMRGs in cancers, while scRNA-seq, spatial transcriptomics, and lipidomics highlighted GLM heterogeneity in PC. Two GLM-associated subtypes with significant prognostic, biofunctional, immune microenvironmental, and drug sensitivity differences were identified in 930 PC samples. Finally, ALDH2 was identified as a novel prognostic biomarker in PC and validated in a large number of datasets and clinical samples. Conclusions: This study highlights the crucial role of GLM in PC and defines a new PC subtype and prognostic biomarker. These findings establish a novel avenue for studying prognostic prediction and precision medicine in PC patients.

Keywords

• pancreatic cancer
• glycerolipid metabolism
• molecular classification
• multi-omics technologies
• ALDH2