Inferring Mutational Chronology In Breast Cancer By Deep-Sequencing Tumor Subpopulations

Heterogeneous breast carcinomas often contain nuclei with multiple distributions of ploidy, each representing a distinct genomic subpopulation. We present an approach called Ploidy-Seq that combines flow-sorting and next-generation sequencing to isolate tumor subpopulations and deep-sequence the nuclei at high coverage (80-100X) to identify somatic mutations. We applied Ploidy-Seq to a triple-negative (ER-, PR-, Her2-) ductal carcinoma that showed three major tumor subpopulations and one stromal subpopulation. In order to distinguish germline from somatic mutations, we first identified all of the SNPs in the 2N stromal cells. We then studied the evolution of different molecular clocks within the breast tumor, including single nucleotide variations (SNVs), indels, structural rearrangements, copy number changes and translocations. By applying set theory operations, we compared the three tumor subpopulations to identify early mutations (present in all subpopulations), intermediate events (shared between two subpopulations) and late mutations (exclusive to each subpopulation). Overall we identified 101 nonsynonymous point mutations of which 31 had occurred early and were present in all tumor subpopulations. Many of the early mutations occurred in cancer genes including THSD7B involved in TGFβ signaling, the CDH26 cadherin-like 26 gene, and the kinesin gene KIF2B. Early mutations were further studied by single-cell sequencing, which showed evidence of clonal expansions within the tumor, suggesting that they provided a major selective advantage. These early mutations are of considerable interest for therapeutic targeting since they are present in all of the tumor subpopulations, and are likely to be driving the growth of the cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 5054. doi:1538-7445.AM2012-5054