Bulk and single-cell transcriptome profiling reveal the metabolic heterogeneity in human breast cancers

Abstract

An emerging view regarding cancer metabolism is that it is heterogeneous and context-specific, but it remains to be elucidated in breast cancers. In this study, we characterized the energy-related metabolic features of breast cancers through integrative analyses of multiple datasets with genomics, transcriptomics, metabolomics, and single-cell transcriptome profiling. Energy-related metabolic signatures were used to stratify breast tumors into two prognostic clusters: cluster 1 exhibits high glycolytic activity and decreased survival rate, and the signatures of cluster 2 are enriched in fatty acid oxidation and glutaminolysis. The intertumoral metabolic heterogeneity was reflected by the clustering among three independent large cohorts, and the complexity was further verified at the metabolite level. In addition, we found that the metabolic status of malignant cells rather than that of nonmalignant cells is the major contributor at the single-cell resolution, and its interactions with factors derived from the tumor microenvironment are unanticipated. Notably, among various immune cells and their clusters with distinguishable metabolic features, those with immunosuppressive function presented higher metabolic activities. Collectively, we uncovered the heterogeneity in energy metabolism using a classifier with prognostic and therapeutic value. Single-cell transcriptome profiling provided novel metabolic insights that could ultimately tailor therapeutic strategies based on patient- or cell type-specific cancer metabolism.

Keywords: breast cancer; metabolism; single-cell RNA sequencing.