Introduction: Classification of breast cancer can incorporate immunohistochemical (IHC) detection of ER/PR/HER2/KI67 to stratify the subtypes. High throughput proteomics analysis allows for the expansion of biomarker discovery within the subtypes. We evaluated a cohort of 109 tumors characterized as ER+ (Luminal A and Luminal B1; HER2+ and ER low (1-10%) cases were excluded) compared to ER-/HER2- tumors. Utilizing an integrated bioinformatics approach, we developed a proteomic marker signature to reclassify tumors into ER+(like) and ER-(like) tumors. CPTAC (Proteomic)/TCGA (RNAseq) datasets and larger METBRIC and GSE96058 cohorts were used to validate this marker signature. The selected biomarkers demonstrated significant differences impacting survival outcome.
Methods: Clinical IHC subtyping of core biopsies was used to select a cohort of patients with ER+/HER2- and 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. Proteomic analysis was performed using Thermo Q-Exactive+ LC MS/MS analysis. Differential analysis was applied to select the significantly altered proteins between ER+ and ER- cases, Univariate survival analysis was engaged to filter informative protein/genes using TCGA RNA-Seq data. Nearest centroid analysis was deployed to define the classifier to predict novel molecular subtypes.
Results/Conclusions: We selected 34 proteins/genes from 164 significantly differentially expressed proteins for further analysis. The centroid model constructed with the 34 proteins defined 2 groups: ER+(like) and ER-(like). An additional 4 groups were defined across subtypes: luminal tumors classified both by IHC and marker signature (LL), luminal tumors classified by IHC but marker signature more like triple negative (LT), triple negative tumors classified by IHC but marker signature more like luminal (TL), and triple negative classified by both IHC and marker signature (TT). This marker signature segregated close to 5000 tumors across CPTAC, TCGA, METABRIC and GSE96058 cohorts. Survival analysis in these groups of patients revealed differences in radiation, hormone/radiation, hormone therapy, and hormone/radiation/chemotherapy treatments. In summary using proteomics data we identified a 34 gene/protein marker signature, validated in large external cohorts and exhibited impact on survival and response to therapy. Further, this signature was enriched in metabolism and microenvironmental associated factors that could represent novel targets or development combination strategies based on this signature.
Citation Format: Guisong Wang, Punit Shah, Rick Searfoss, Leigh Fantacone-Campbell, Jeffrey A. Hooke, Brenda Deyarmin, Rebecca N. Zingmark, Stella Somiari, Jianfang Liu, Leonid Kvecher, Bradley Mostoller, Lori A. Sturtz, Praven-Kumar Raj-Kumar, Elder Granger, Linda Vahdat, Mary L. Cutler, Chas Bountra, Rangaprasad Sarangarajan, Hai Hu, Michael A. Kiebish, Albert J. Kovatich, Niven R. Narain, Craig D. Shriver. Reclassification of ER+ (luminal A/luminal B1 minus ER low)-like and ER- like breast tumors based on proteomic/gene and clinical outcome signatures [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1188.