Abstract 4566: RNA-seq identifies a TGF-β signature that predicts response to preoperative bevacizumab in breast cancer.

Proceedings: AACR 104th Annual Meeting 2013; Apr 6-10, 2013; Washington, DC Background: Gene expression profiles from tumor biopsies taken during preoperative therapy may identify markers associated with therapy benefit. We have previously shown that changes in TGF-β pathway activity following one dose of bevacizumab (B) are associated with pathologic complete response (pCR) (SABCS, 2012). Here we evaluated the ability of this signature to predict pCR following brief exposure to trastuzumab (T) or nab-paclitaxel (N). Methods: We sequenced transcriptomes of core biopsy RNA from 50 breast tumor pairs obtained from neoadjuvant clinical trials BrUOG 211A/211B. Patients were given one dose of B (n=15) or N (n=17) or T (n=18), followed by combination biologic/chemotherapy (HER2- with B/carboplatin/N; HER2+ with T/carboplatin/N). Paired-end sequencing (74bp) was performed using amplified total RNA from biopsies obtained pre/post 10 day exposure to run-in monotherapy. Transcript abundance and differential expression were estimated assuming Poisson-distributed read-counts. Significance-adjusted log2 fold change (sa-log2FC) of gene expression was calculated by diminishing statistically insignificant (p > 0.05) log2FC values by 100. sa-log2FC values of the TGF-β activity signature genes were used to cluster patients and the clusters were tested for enrichment of pCR cases. Prioritization of the genes in the TGF-β activity signature was performed with a shrunken centroid classifier using 3-fold cross-validation to identify an appropriate shrinkage threshold. Results: PCA and RNA-seq based PAM50 subtyping showed baseline and post-brief exposure samples clustering together across all patients. We identified significantly differentially expressed genes upon brief exposure to therapy associated with pCR in the B arm. Pathway analysis revealed TGF-β signaling, Cell Cycle, DNA Replication and Steroid Biosynthesis pathways to be enriched (p≤0.05) after multiple testing correction. Following the lead on TGF-β signaling, we derived a 107-gene TGF-β activity signature by combining a published TGF-β activity signature (Padua et al, 2008) with the pCR-associated KEGG TGF-β signaling pathway genes in our study. This signature showed a strong separation of pCR cases (p=0.004) by clustering. We further refined this using a shrunken centroid classifier to identify a 61-gene signature of TGF-β activity whose down-regulation is predictive of pCR in the B arm. Evaluation of the TGF-β signature in the T and N arms showed no association with pCR. Conclusions: TGF-β pathway genes whose activity is altered by brief exposure to bevacizumab are associated with pCR; a similar pattern was not observed following exposure to trastuzumab or nab-paclitaxel. This signature may provide an early indication of potential benefit from preoperative treatment with bevacizumab and combination chemotherapy. We are currently in the process of validating these results using a custom RT-PCR panel. Citation Format: Vinay Varadan, Kristy Miskimen, Sitharthan Kamalakaran, Angel Janevski, Nilanjana Banerjee, Nicole Williams, Maysa Abu-Khalaf, William Sikov, Nevenka Dimitrova, Lyndsay Harris. RNA-seq identifies a TGF-β signature that predicts response to preoperative bevacizumab in breast cancer. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 4566. doi:10.1158/1538-7445.AM2013-4566