Predictive and prognostic value of B-cell gene-expression signatures and B-cell receptor (BCR) repertoire in HER2+ breast cancer: A correlative analysis of the CALGB 40601 clinical trial (Alliance)

Abstract

Background: Although tumor-infiltrating lymphocytes density is associated with increased response and improved outcomes in HER2+breast cancer, BCR clonal diversity repertoire could provide a more informative measure of an individual's immunemediated anti-tumor response. In this study, we focus on the specific role of B-cell gene-expression signatures and BCR repertoire as predictive and prognostic biomarkers in CALGB 40601, a neoadjuvant study of single vs. dual (trastuzumab+lapatinib) HER2 targeting with paclitaxel. Method(s): Gene expression profiling by mRNA sequencing was performed on 265 pretreatment samples and signature scores were calculated by determining the median expression of all genes in a signature. BCR repertoire analysis using V'DJer was assessed on 256 of the samples. The predictive and prognostic value of clinical parameters, signature scores and BCR diversity metrics was tested in a univariate analysis for pathologic complete response (pCR) and event-free survival (EFS). Result(s): Of the>600 expression signatures tested, 10 were significantly associated with both pCR and EFS. Five immune-related signatures were associated with higher pCR and better outcome: a T-helper signature (OR=5.1, HR=0.22, both p<0.05), 2 IgG signatures (OR=2.2 and 1.72, HR=0.61 and 0.70 respectively, all p<0.05) and 2 Bcell signatures (OR=1.63 and 1.27, HR=0.74 and 0.81 respectively, all p<0.05). Patients with a high IgG-signature showed a significant higher pCR rate when treated with a trastuzumab combination regimen (58% vs. 34%, p 0.001). The Ig heavy chain c (IGHG) was the most abundant isotype (median counts =4,800). Patients without assembled IGHG or a high evenness showed a significantly lower pCR rate (27% vs. 54%, p<0.001 and 12% vs. 59%, p<0.001, respectively). Patients with high IGHG counts also showed a significant EFS benefit at 5 years (log rank 0.03). Conclusion(s): B-cell gene expression signatures have a relevant predictive and prognostic value in CALGB 40601. The clinical implementation of these biomarkers could help us to design new neoadjuvant treatment strategies for HER2+ breast cancer.