Abstract PS18-11: Molecular characterization of advanced breast cancer according to site of metastasis

Abstract

Abstract Background: There is an increasing need in the clinic to biopsy metastatic disease in patients with advanced breast cancer (ABC). However, the microenvironment and tumor cell biology of breast cancer metastasis is largely unknown. Here, we report a molecular characterization of ABC according to the site of metastasis. Methods: RNA from 184 FFPE metastatic samples were evaluated using the nCounter BC 360 Panel, which includes the expression of 689 BC-related genes and 82 immune-related genes. PAM50 subtypes (Basal-like, HER2-enriched [HER2-E], Luminal A, Luminal B and Normal-like) were also determined. HER2 protein expression was assessed by immunohistochemistry (IHC) in 115 tumor samples, and HER2-low tumors (i.e. 1+ or 2+ and ISH-negative) were identified. Descriptive statistics, significance analysis of microarrays (using False Discovery Rate [FDR]) and logistic regressions were used to identify organ-specific gene expression profiles. Finally, we derived an organ-specific predictor of 209-genes from our cohort, and applied it to the RNAseq-based TCGA PanCancer dataset, which includes 174 glioblastomas multiforme, 424 liver hepatocellular carcinomas and 576 lung adenocarcinomas, among other cancer-types. Results: A single metastatic tumor sample from 184 individual patients with ABC was obtained from bone (18%), liver (17%), skin (15%), brain (12%), breast (13%), lymph nodes (9%), lung (7%), pleura (5%), ovary (2%), muscle (1%) and peritoneum (1%). All PAM50 subtypes were identified across the main organ sites; however, significant differences in subtype distribution were observed (p<0.001). Basal-like subtype was more prevalent in brain, lung and skin metastasis (FDR=0.3%). Unsupervised analysis and principal component analysis revealed brain and liver metastasis as the most distinct. Supervised analysis identified organ-specific genes independently of PAM50 subtype, i.e.: bone-specific genes (WIF1, IBSP, MMP9 and ITGB3); brain-specific genes (CRYAB, SOX10, FGF1 and CHI3L1); liver-specific genes (ALDH1A1, CYP4F3, PCK1, and SFRP2); lung-specific genes (CAV1, WNT5A, PTGS2 and IL6); skin-specific genes (KRT14, KRT5, S100A7 and SERPINB5). Among the organ-specific gene list, 15 (30%) of the 50 PAM50 genes were identified, including up-regulation of FGFR4 in liver, ESR1 in bone, ERBB2 in lung and KRT5 and KRT14 in skin. Regarding ERBB2, a high correlation between ERBB2 mRNA and HER2 IHC expression (0, 1+, 2+ and 3+) was observed (p<0.001). Interestingly, HER2-low disease was identified across all PAM50 subtypes and organ sites, including bone metastasis. Regarding immune-genes, all were found differentially expressed across the main organ sites (FDR<5%) with lung metastasis showing the highest expression (i.e. PDCD1, CD8A, GMZA, IL1B) and liver and brain metastasis the lowest. Finally, the 209-gene organ-specific predictor in PanCancer TCGA identified 96% of glioblastomas multiforme as brain, 98.6% of liver hepatocellular carcinomas as liver and 57.1% of lung adenocarcinoma as lung. Conclusions: The main sites of metastasis in ABC have unique biological features independently of tumor molecular subtype. Our results suggest that treatment strategies based on the site(s) of metastasis should be explored. Citation Format: Fara Brasó-Maristany, Laia Paré, Núria Chic, Olga Martínez-Sáez, Tomás Pascual, Meritxell Mallafré, Blanca González-Farré, Esther Sanfeliu, Débora Martínez, Patricia Galván, Belinda Salinas, Barbara Adamo, Reinaldo Moreno, Maria Vidal, Montserrat Muñoz, Aleix Prat. Molecular characterization of advanced breast cancer according to site of metastasis [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 PS18-11.