Recurrent Patterns of DNA Copy Number Alterations in Tumors Reflect Metabolic Selection Pressures

Human tumors exhibit recurrent patterns of DNA amplifications and deletions across diverse cancer types. These patterns are suggestive of conserved selection pressures during tumor evolution but cannot be fully explained by known oncogenes and tumor suppressor genes. Applying principal component analysis (PCA) to genome-wide DNA copy number alteration (CNA) from 17 tumor types, we identified previously unreported CNA genomic sub-signatures. Combining these human signatures in a cross-species comparison with signatures from mouse models of cancer allowed for synteny-based elimination of passenger genes. Examination of the cross-species refined CNA signatures revealed coordinate enrichment of copy number alterations in core glycolysis enzymes. Notably, these PCA-defined CNA signatures are predictive of glycolytic phenotypes, including FDG-avidity of patient tumors. Using an experimental immortalization system, we show that exogenous expression of metabolic enzymes alters the copy number status of the corresponding endogenous loci, supporting the hypothesis that these metabolic genes act as drivers within the conserved CNA amplification regions. Finally, using mass spectrometry-based metabolomics, we demonstrate that cells with glycolytic CNA signatures channel an increased proportion of glucose-derived carbon into pentose phosphate-associated biosynthetic pathways including nucleotide synthesis. Taken together, our experimental and computational data support a model in which glycolysis-linked selective pressures encountered during tumorigenesis (eg, redox stress and senescence) shape the highly recurrent DNA copy number alterations found in aneuploid human tumors.