Abstract PD8-05: Comparative analysis of the genomic landscape of breast cancers from women of African and European ancestry

Objectives: Paucity of data on populations of African Ancestry in clinical trials continues to limit our ability to design and implement innovative solutions to narrow the breast cancer survival gap amongst Africans, African Americans, and European Americans. We have developed a cross-continent research infrastructure to examine the spectrum of genomic alterations in breast tumors from West Africa and subsequently, to compare them to tumors from African American women and women of European Ancestry in The Cancer Genome Atlas (TCGA) database.

Methods: Consecutive women with breast cancer presenting for treatment at the University College Hospital, Ibadan and at Lagos State University Teaching Hospital, Lagos, Nigeria gave informed consent and were recruited to the West African Breast Cancer Study (WABCS) between 2013-2016. Tumor-normal pairs were subjected to exome and/or high-depth (90x) genome sequencing. High confidence somatic mutations (substitutions, insertions/deletions and structural variants) were obtained by using multiple variant callers. Furthermore, 1,089 exomic and 80 genomic breast tumor-normal pairs from TCGA were harmonized with WABCS samples, resulting in a cohort of 147 West Africans (147 exome; 40 genome), 154 African Americans (154 exome; 31 genome), and 776 Caucasians (776 exome; 43 genome).

Results: Across the exomes, genes commonly altered in breast cancer in TCGA are also altered in women of African ancestry, but the mutational spectrum is quite different, demonstrating overrepresentation of tumors with aggressive phenotypes. Overall, TP53 (65%), ERBB2 (27%), and GATA3 (17%) showed statistically significant higher alteration frequencies in West Africans and African Americans. In contrast, PIK3CA (24%) was less frequently mutated. Of note, GATA3 mutation was statistically significantly more frequent in Nigerians (39%) and African Americans (16.7%) compared to Caucasians (10.5%), in ER-positive cancers. Analysis on Structural Variants (SV), on the other hand, has shown that the genome-wide SV counts among three populations are comparable in ER-negative cancers, while Nigerians have significantly more SV counts compared to African Americans (P=0.0013) or European Americans (P=2.9x10-5) in ER-positive cancers. Similarly, genome-wide substitution patterns in ER+ and ER- cancers varied widely by race/ethnicity. In ER- cases, West Africans carried the highest proportion of canonical APOBEC-associated substitutions, particularly C>T transitions. Conversely, European Americans with ER+ disease showed a higher proportion of C>T than both West Africans (Welch t-test P = 0.044) and African Americans (Welch t-test P = 0.011). Mutation signature analyses highlighted multiple APOBEC signatures, with notable contribution differences across ancestry and ER status. A signature likely corresponding to DNA damage repair correlated with the proportion of genetic ancestry, being most prevalent in European Americans and least common in Nigerians, particularly in ER-negative cancers, with African Americans showing a degree of this signature's contribution in between the two populations (linear model adjusted for age, P=1.0x10-10).

Conclusions: Overall, our data suggests mutation spectra differences in across race/ethnicity and geography. Identification of molecular characteristics such as higher rates of HER2 enriched tumors and higher rates of GATA3 mutations in ER positive tumors are beginning to reveal the genomic basis of race-associated phenotypes and outcomes in breast cancer. Population differences in frequency and spectrum of mutations should now inform the design of innovative clinical trials that improve health equity and accelerate Precision Oncology care in diverse populations.

Citation Format: Olopade OI, Pitt JJ, Riester M, Odetunde A, Yoshimatsu T, Labrot E, Ademola A, Sanni A, Okedere B, Mahan S, Nwosu I, Leary R, Ajani M, Johnson RS, Sveen E, Zheng Y, Wang S, Fitzgerald DJ, Grundstad J, Tuteja J, Clayton W, Khramtsova G, Oludara M, Omodele F, Benson O, Adeoye A, Morhason-Bello O, Ogundiran T, Babalola C, Popoola A, Morrissey M, Chen L, Huo D, Falusi A, Winckler W, Obafunwa J, Papoutsakis D, Ojengbede O, White KP, Ibrahim N, Oluwasola O, Barretina J. Comparative analysis of the genomic landscape of breast cancers from women of African and European ancestry [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD8-05.

Abstract P6-03-17: Genomic landscape of breast cancers from women of African ancestry across the diaspora

Objectives: Of all ethnic/racial groups, age-standardized mortality rate from breast cancer is highest for African American women in the US for reasons that remain understudied. The paucity of genomic studies of breast tumors across the African Diaspora further restricts our understanding of the biology of breast cancer in underserved populations. To gain a better understanding of the genomic landscape of breast cancer in women of African Ancestry, we have developed a cross continent translational research infrastructure to examine the spectrum of genetic alterations in breast tumors from West Africa compared to the spectrum of alterations observed in tumors from African-American and other women who are predominantly white in The Cancer Genome Atlas (TCGA) dataset.

Methods: Peripheral blood and breast cancer biopsy tissues were collected from 214 patients enrolled in the West Africa Breast Cancer Study (WABCS) at the University of Ibadan/University College Hospital (UI/UCH) and at Lagos State University Teaching Hospital (LASUTH). Blood DNA as well as breast cancer tissue DNA and RNA were extracted at the Novartis Institutes for Biomedical Research (NIBR), UI/UCH, and LASUTH using a modified protocol of PAXgene Tissue DNA and RNA extraction method. Whole-exome (WES) and transcriptome (RNA-seq) sequencing were performed on the Illumina HiSeq2000 platform at NIBR. Single Nucleotide Variants (SNVs) and insertions/deletions (indels) were called using MuTect and Pindel, while Copy Number Alterations (CNAs) were called using an in-house implementation of the ABSOLUTE method. Observed mutations were compared against those reported in the TCGA dataset. ER, PR and HER2 status were determined by immunohistochemistry (IHC) at UI/UCH, LASUTH and UChicago.

Results: WES data for 95 tumors have been analyzed thus far. Genes commonly mutated in breast cancer in TCGA are also mutated in WABCS but the mutational spectrum is vastly different. TP53 (64%), MYC (31%), and GATA3 (26%), showed significantly higher alteration frequencies in WABCS and African Americans. In contrast, PIK3CA (20%), CDH1 (2%), and MAP3K1 (2%) were less frequently mutated in women of African ancestry. In addition to the high proportion with TP53 mutations, the proportion with HER2 positive subtype of 42.1% and triple-negative subtype of 37.9% suggest that tumors with the most aggressive features are overrepresented in breast cancer patients in West Africa.

Conclusions: In the first study of its kind, high throughput genomic analysis of the largest cohort of women of African ancestry has uncovered alterations in cancer genes, some of which may be amenable to treatment with targeted therapies. Furthermore, we provide evidence that population differences in frequency and spectrum of mutations should drive the design and deployment of precision medicine initiatives. Only then can we develop innovative interventions to reduce the unacceptably high rates of mortality from breast cancer in underserved and under resourced populations.

Citation Format: Olopade OI, Odetunde A, Riester M, Yoshimatsu T, Labrot E, Ademola A, Sanni A, Okedere B, Mahan S, Nwosu I, Leary R, Ajani M, Johnson RS, Sveen E, Zheng Y, Clayton W, Khramtsova G, Oludara M, Omodele F, Benson O, Adeoye A, Morhason-Bello O, Ogundiran T, Babalola C, Popoola A, Morrissey M, Huo D, Falusi A, Winckler W, Obafunwa J, Papoutsakis D, Ojengbede O, Ibrahim N, Oluwasola O, Barretina J. Genomic landscape of breast cancers from women of African ancestry across the diaspora. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P6-03-17.

Association of Proliferating Macrophages with High Grade, Hormone Receptor Negative Breast Cancer

Macrophages, a key cell in the inflammatory cascade, have been associated with poor prognosis in cancers, including breast cancer. Tumor associated macrophages (TAM) have also been shown to play a role in invasion and metastases. In this study, we investigated the role of a subset of macrophages known as proliferating macrophages (promacs) in breast cancer. We examined the relationship between promacs and clinico-pathologic characteristics such as tumor size, grade, lymph node metastasis, hormone receptor status, molecular subtype, and survival. This study was conducted at two independent institutions (University of California, San Francisco and University of Chicago) using two independent cohorts of patients with breast cancer. Formalin-fixed, paraffin-embedded sections and/or tissue microarrays were double-stained with anti-CD68 (a macrophage marker) and anti-PCNA (a proliferation marker) antibodies. The number of total macrophages as well as promacs per high power field were counted. Molecular subtypes were determined from gene expression array analyses. Outcomes were available for a subset of the cases and Kaplan Meier survival curves were generated.Tumor tissue from two different cohorts of patients with breast cancer were stained and analyzed at two independent institutions. In both studies, promac density was significantly correlated with higher grade, hormone-receptor negative tumors, and a basal-like subtype. In contrast there was no correlation between numbers of promacs and tumor size, stage, or number of involved lymph nodes. The presence of increased promacs was a significant predictor of survival in these treated patients. Proliferating macrophages are more abundant in high grade tumors, and predict a worse outcome, independent of stage and grade. These findings, corroborated at two independent institutions, suggest that the presence of promacs is associated with tumor progression and increased promac density may serve as a prognostic indicator for poor outcomes. Novel therapies that are able to target and halt the activity of promacs may be a promising strategy for the treatment of promac rich tumors.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3042.