Mutational patterns in the breast cancer mitochondrial genome, with clinical correlates.

The presence of mitochondrial DNA (mtDNA) mutations in human cancer has long been recognized, but their functional significance has remained obscure. Debate persists as to whether the mutations help drive the tumor, or are bystander events. Here, we analyze next-generation mtDNA sequence data from 99 breast cancer patients. High depth coverage enables detection of even low-level heteroplasmic variants, and data from matched normal tissue allow us to distinguish between shifts in heteroplasmy and acquired mutations. Somatic mtDNA mutations are found in 73 (73.7%) of patient tumors, and dramatic shifts from the initial germline allele proportions are observed for many heteroplasmies. Clustering of somatic mutations in promoter and replication regions, and also in genes coding for electron transport chain complex I, suggest selection for mutations affecting critical mitochondrial processes. Furthermore, statistical tests for Darwinian selection reveal evidence for positive and relaxed negative selection for somatic missense mutations. We also observe a dramatic decrease in per-cell mtDNA content in tumor tissues, as well as a surprising positive correlation between somatic mtDNA mutational burden and patient survival. Taken together, our results support the view that somatic mtDNA mutations are not solely bystander events, but have significance in cancer from both biological and clinical perspectives. We also anticipate that the catalog of heteroplasmies and somatic mutations presented here will serve as a reference for future studies of cancer mitochondrial genomes.