Abstract 3750: Genomic alterations underlie a pan-cancer metabolic transcriptome shift

Abstract

Abstract Background Hypoxia is a key physiological and microenvironmental difference between tumour and normal tissues. Hypoxia induces DNA damage, genomic aberrations and reprogramming of DNA repair mechanisms. A crucial mechanism in a highly dynamic hypoxic tumour microenvironment is metabolic reprogramming of tumour cells. Genome-wide multi-modal molecular profiles have been generated, however it remains unclear whether upregulation of metabolic genes is substantially enriched with somatically acquired alterations. To answer this, we performed a pan-cancer integrative analysis of genomic and transcriptomic profiles of 10 TCGA tumour types (n = 5,500). Methods 2,750 metabolic genes were extracted from KEGG pathways, and differentially upregulated genes were identified for each tumour type. High-ranking (10%) upregulated genes were assessed for correlation between mRNA abundance and DNA copy-number changes, and subsequently tested for metabolic-enrichment using genome-wide permutation analysis. For each tumour type, highly correlated genesets were subsequently examined for their potential as drug targets and biomarkers of disease relapse. Results Integrative analysis of mRNA and copy-number data revealed clusters of tumours that exhibit similar landscape of metabolic gene profiles, and those having unique metabolic copy number aberrations. Within these profiles, we identify a core set of genes that exhibit strong correlation (Spearman's ρ>0.3; q<0.001) between DNA and RNA profiles (candidate drivers) across a range of tumour types. Compared to genome-wide correlation patterns, the correlated genes were significantly over-represented in breast, glioblastoma multiforme and ovarian cancers (p<0.05). Of note, the correlation analysis uncovered breast cancer subtype-specific heterogeneity in altered metabolic profiles with aggressive breast cancers (Basal and HER2-enriched) demonstrating higher number of candidate drivers. mRNA correlation analysis performed on the top candidate drivers revealed an overall trend of positive correlation between hypoxia signature and metabolic genes. This highlights hypoxia as a potential cause of metabolic dysregulation. Further in-vitro analysis of one of the top candidate drivers, SQLE amplification, showed a marked sensitivity to inhibition in hypoxia compared to normoxia, highlighting hypoxia dependence of genetic and metabolic reprograming in aggressive breast cancers (Basal and HER2-enriched) and colon cancer. Conclusions We isolate tumour-type specific and core pan-cancer metabolic gene signatures revealing within and cross disease heterogeneity in metabolic profiles. We highlight a core set of pan-cancer metabolic candidate drivers which are recurrently over-expressed due to genomic amplifications induced by hypoxia. Citation Format: Syed Haider, Alan McIntyre, Ruud GPM van Stiphout, Laura M. Winchester, Simon Wigfield, Adrian L. Harris, Francesca M. Buffa. Genomic alterations underlie a pan-cancer metabolic transcriptome shift. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3750. doi:10.1158/1538-7445.AM2015-3750