Abstract PS5-34: Identification of proteomics-based biomarkers for ER+/HER2- breast cancer stratification: Implications on clinical outcome

Abstract

Abstract Introduction: Improving stratification of breast cancer (BC) patients based on molecular signatures for treatment responses and clinical outcomes is a critical unmet need. Currently, endocrine therapy is first-line for hormone receptor positive (HR+) BC. Chemotherapy is added to patients with high-risk luminal BC. In practice, levels of Ki-67 are used to distinguish luminal tumors as low-risk luminal A (LA) and high-risk luminal B (LB) for adjuvant therapy decisions. Herein, proteomic tumor assessment from ER-positive HER2-negative (ER+/HER2-) BC patients was utilized to define molecular subtyping, estimate congruency between proteomic subtyping and traditionally used Ki67 marker, and define a new set of potential predictive and prognostic therapeutic biomarkers for ER+/HER2- BC patients. Method/Result: Clinical immunohistochemistry (IHC) subtyping of core biopsies was used to select a cohort of 86 BC patients with ER+/HER2- primary tumors from flash-frozen surgical samples. The positive/negative status of ER/PR/HER2 was defined using updated ASCO 2020 guidelines. Ki-67 status was determined using the 2011 St. Gallen’s International Expert Consensus recommendations. The cohort includes 28 LA (Ki67 < 14%) cases and 58 LB1 (Ki67 >= 14%) cases. Integrated consensus clustering algorithms with the most varying proteins in our cohort were applied to identify proteomic subtypes. Two distinct separations were observed from the analysis, resulting in one cluster enriched with LA (40 cases) and the other enriched with LB1 (46 cases) called by Fisher’s exact test. These clusters matched 100% with the clusters generated using 900+ proteins common to the 1500+ proteins used in the CPTAC-BC proteomics-based subtyping analysis (Mertins et al. Nature 2016). The differential analyses demonstrated that there is no significant difference between Ki67-defined subtypes and proteomics-defined subtypes (LA-enriched vs. LA cases, LB1-enriched vs. LB1 cases),indicating they are consistent in the molecular profile. Differential analysis was performed to compare LB1-enriched versus LA-enriched cases, resulting in 672 significantly differentially expressed proteins defined at false discovery rate (FDR) < 0.05 and |log2(fold change)|>1. 353 of the 672 proteins were correlated with mRNA at Pearson correlation > 0.39 as reported in the CPTAC-BC study or cBioPortal for Cancer Genome, and their coding genes were used for progression free interval (PFI) analysis based on TCGA RNA-seq data in the TCGA ER+/HER2- cases (662 cases, c.f. Huo et al. JAMA Oncology 2017). 90 of the 353 coding genes significantly associated with PFI were detected at p-value<0.05. Unsupervised hierarchical clustering method and principal component analysis (PCA) of the 90 genes were applied to our cohort to investigate the clustering performance and 94.2% of the cases were clustered correctly using support vector machine (SVM) method after PCA analysis. Biological process and molecular function GO term over-representation analyses of the 90 coding genes were performed separately. Some significant and biologically meaningful GO terms were identified at FDR<0.05. Conclusions: We identified a set of biomarkers that can be potentially employed as proteomic or gene signatures to stratify ER+/HER2- BC into low risk and high-risk groups. Disclaimers The contents of this publication are the sole responsibility of the author(s) and do not necessarily reflect the views, opinions, or policies of Uniformed Services University of the Health Sciences, The Henry M Jackson Foundation for the Advancement of Military Medicine Inc., the Department of Defense or the Departments of the Army, Navy or Air Force. Mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government. Citation Format: Guisong Wang, Punit Shah, Rick Searfoss, J. Leigh Fantacone Campbell, Jeffrey A. Hooke, Brenda Deyarmin, Rebecca N. Zingmark, Stella Somiari, Jianfang Liu, Leonid Kvecher, Lori A. Sturtz, Praveen-Kumar Raj-Kumar, Elder Granger, Linda Vahdat, Mary L. Cutler, Rangaprasad Sarangarajan, Hai Hu, Michael A. Kiebish, Albert J. Kovatich, Niven R. Narain, Craig D. Shriver. Identification of proteomics-based biomarkers for ER+/HER2- breast cancer stratification: Implications on clinical outcome [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS5-34.