Abstract P4-06-01: Using multi-omic profiling to unravel the complexity of triple-negative breast cancer (TNBC)

Abstract

Abstract Introduction: TNBCs show marked heterogeneity in their spectrum of genomic alterations, patterns of gene expression, proteomic profiles, and immune microenvironments. Although TNBC heterogeneity has been characterized at each of these levels individually, how variation at one level is associated with differences at other levels is poorly understood. As new targeted and immune-based therapies are developed for TNBC, there is a critical need to better understand this interplay. In this study, we performed “multi-omic” profiling on a cohort of TNBCs in order to determine how various DNA, RNA, protein, and immunologic parameters are correlated. Methods: A cohort of 95 formalin-fixed, paraffin-embedded tumor specimens from primary and metastatic TNBCs was curated. The following parameters in each tumor specimen were quantified: stromal tumor-infiltrating lymphocytes (sTILs) assessed on hematoxylin and eosin stained sections; expression of programmed death-ligand 1 (PD-L1), androgen receptor (AR) and the retinoblastoma tumor suppressor protein (RB) by immunohistochemistry (IHC); and expression of 776 breast cancer-related transcripts (using the nCounter Breast Cancer 360 panel). Targeted tumor DNA sequencing was performed in a subset of cases and tumor mutational burden (TMB) and DNA mismatch repair (MMR) status were determined. Gene expression patterns were analyzed using customized algorithms and we performed descriptive statistics to evaluate associations between DNA, RNA, and protein variables. Results: All 95 TNBC specimens were analyzed for sTILs; IHC for RB, AR, and PD-L1; and transcriptomic profiling. DNA sequencing was performed on 68 cases. The genomic, transcriptomic, and proteomic classification of the cohort is provided in Table 1. AR+ TNBCs (>1% tumor cells on IHC) were more likely to be non-basal (p<0.001), showed higher levels of AR (p<0.001) and FOXA1 (p<0.01) transcripts, were enriched for expression of genes related to mammary differentiation (p<0.001), depleted for expression of genes related to proliferation (p=0.018), “BRCAness” (p<0.001), and homologous recombination deficiency (HRD) (p=0.045), and showed significant upregulation of the immune inhibitory checkpoint molecule TIGIT (p=0.004). RB-positivity (>50% tumor cells on IHC) correlated with higher RB1 transcript levels but not with RB1 deletion/mutation status. RB+ tumors were depleted for expression of genes associated with HRD and BRCAness (p=0.029 and p=0.071). TNBCs showing widespread upregulation of gene expression signatures related to numerous immune processes included those with: (i) higher levels of sTILs, (ii) >1% PD-L1+ stromal mononuclear cells and/or tumor cells, (iii) tumors without RB1 or TP53 alterations. Genomic aberration in the PI3K pathway was associated with increased expression of AR and FOXA1 but not with alterations in immune gene expression. Conclusion: Patterns exist connecting TNBC heterogeneity across the DNA, RNA, and protein levels. Further analyses determining associations between genomic, transcriptomic, and proteomic features in this cohort of TNBC will be presented at the meeting. Genomic, transcriptomic, and proteomic classification of tumors in the entire TNBC cohortTumor site (n=95)Primary64 (67%)Regional nodal4 (4%)Distant nodal or visceral27 (28%)Stromal TILs (n=93)<10%50 (53%)10-49%30 (32%)>50%13 (14%)RB staining (n=95)Positive (>50% tumor cells)52 (55%)AR staining (n=95)Positive (>1% tumor cells)34 (36%)PD-L1 staining (stromal, n=94)Positive (>1% mononuclear cells)34 (37%)PD-L1 staining (tumor, n=94)>1% tumor cells27 (29%)PAM50 subtype (n=90)Basal72 (80%)HER2-enriched15 (17%)Luminal B3 (3%)TNBC subtype (n=91)Basal-like immune activated (BLIA)76 (80%)Basal-like immunosuppressed (BLIS)7 (7%)Mesenchymal (MES)8 (8%)Luminal androgen receptor (LAR)4 (4%)DNA alterations (n=68)TP53 alterations54 (80%)BRCA1/2 mutations14 (21%)RB1 alterations21 (31%)PI3K pathway alterations31 (46%) Citation Format: Shom Goel, Heather Ann Brauer, Yuqi Ren, Kara Gorman, Wafa Osmani, Jessica Pittenger, Chelsea Andrews, Edward T Richardson III, Elizabeth A Mittendorf, MD PhD, Eric P Winer, Stuart Schnitt, Sara M Tolaney. Using multi-omic profiling to unravel the complexity of triple-negative breast cancer (TNBC) [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-06-01.