Metabolic subtyping and targeting of pancreatic ductal adenocarcinoma

Abstract

Pancreatic Ductal Adenocarcinoma (PDAC) is one of the deadliest malignancies in the world due to late diagnosis and lack of effective treatment. Thus, novel therapies are desperately needed. PDAC is presented in two molecular subtypes, quasi-mesenchymal (QM, basal-like) and classical (pancreatic progenitor). Recent studies showed that these subtypes clinically differ as well with classical subtype being more therapy responsive and with somewhat better prognosis. This study aimed to identify which metabolic pathways are activated in one and the other subtype and whether these metabolic differences are functionally preserved in relevant preclinical PDAC models that may be further used for the development of subtype-specific metabolic therapies. To this aim, established PDAC cell lines, patient-derived xenografts, primary cells and patient samples were used as models. Gene expression analyzes (RNA-seq, microarray, qPCR and GSEA) revealed that QM subtype is dominated by the metabolic transcripts involved in glycolysis and hypoxia whereas classical subtype is branded with gene sets involved in lipid and fatty acid metabolism. Functional seahorse metabolic flux assays also supported the transcriptome data that glycolysis and fatty acid oxidation are active metabolic processes in QM and classical subtype, respectively. Furthermore, immunohistological approaches identified lactate transporter MCT4 as a putative marker of glycolytic / QM PDACs in patient-related materials. However, functional heterogeneity and different levels of glycolysis and fatty acid oxidation usage were also present among members of the same subtype. Accordingly, analyzed chemical inhibitors of glycolysis and lipid metabolism did not present strong subtype dependent activity. Even though glycolysis and fatty acid oxidation are dominant in QM and classical subtype respectively, identified functional heterogeneity indicates fine-tuning of metabolic dependencies and mixed usage of different metabolic pathways even in one transcriptional subtype. In a personalized approach, a patient with a transcriptional signature of QM and functionally active glycolysis would be an obvious candidate for metabolic targeting of glycolysis. However, due to the mixed usage of different metabolic pathways as well, further preclinical and clinical efforts are needed for evaluating the potential of combinatorial usage of metabolic inhibitors.