Across race/ethnicities, breast cancer incidence and mortality rates markedly differ. Numerous studies have demonstrated that individuals of African ancestry acquire aggressive, early-onset breast cancers more frequently than other populations for reasons that remain unexplained. The sources of these disparities are complex, and a comprehensive characterization of mutation landscapes amongst indigenous Africans, African Americans (AA), and Caucasian breast tumors has not been performed.

We generated high-depth whole genome sequencing on 31 Nigerian breast cancers (90x) along with matched blood (30x). Breast cancer whole genomes (tumor-normal pairs) from The Cancer Genome Atlas were harmonized with our samples, resulting in a cohort of 31 Nigerians (17 estrogen receptor negative, ER-), 31 AA (22 ER-), and 43 Caucasians (19 ER-). High confidence somatic mutations (substitutions and insertions/deletions) were obtained by using multiple variant callers. Regardless of race, ER- tumors carried similar numbers of mutations than their estrogen receptor positive (ER+) counterparts (Welch t-test p = 0.57 - 0.82). TP53 (64%) was the most frequently mutated gene in ER- disease, while canonical PIK3CA activating mutations were prevalent in ER+ cases (33.3%). Additionally, tumor suppressor genes RB1, NF1, and PTEN were disrupted via structural rearrangements in ∼6 to 15% of samples. Rearrangements in the H3K27 methylation regulator EZH1 were identified in six Caucasians but only one individual with African _ancest. Notably, within coding regions, no striking mutation rate differences amongst races were identified. However, global substitution patterns in ER+ and ER- cancers varied widely by race/ethnicity. In ER- cases, Nigerians carried the highest proportion of canonical APOBEC-associated substitutions, particularly C>T transitions. Conversely, Caucasians with ER+ disease showed a higher proportion of C>T than both Nigerians (Welch t-test p = 0.044) and AA (Welch t-test p = 0.011). Kataegis, or clustered mutations, was most prevalent in Nigerian samples, regardless of ER status. Evidence for kataegis was often corroborated by structural variant breakpoints and aberrant copy number states at the hypermutated locus. Mutation signature analyses highlighted multiple APOBEC signatures, with moderate contribution differences across race and ER status.

Overall, our data suggests potential mutation spectra differences in Caucasian, African American and indigenous African breast tumors. Identification of these molecular characteristics by ancestry and geography may help understand race-associated phenotypes and exposures that drive outcomes in breast cancer.

Citation Format: Jason J. Pitt, Toshio F. Yoshimatsu, Yonglan Zheng, Jason Grundstad, Jigyasa Tuteja, Jiebiao Wang, Abayomi Odetunde, Galina Khramtsova, Wendy Clayton, Adeyinka Ademola, Temidayo O. Ogundiran, Adenike T. Adeniji-Sofoluwe, Millicent Obajimi, Adewunmi Adeoye, Chinedum Babalola, Oladosu A. Ojengbede, Christopher O. Olopade, Oluwasola A. Olayiwola, Lin Chen, Dezheng Huo, Kevin P. White, Olufunmilayo I. Olopade. Whole genome sequencing reveals different patterns of mutational mechanisms in breast tumors between women of African and European descent. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4494.