Preclinical discovery and clinical validation of predictive markers of response to cetuximab (Erbitux™) in metastatic colorectal cancer

4032 Cetuximab (Erbitux™) is a chimeric monoclonal antibody that binds to the extracellular ligand-binding domain of the epidermal growth factor receptor (EGFR) and inhibits EGFR-dependent signaling. Clinical data has shown that whereas EGFR levels and EGFR mutations do not appear to correlate with response to Cetuximab in colorectal cancer (CRC), an increased number of gene copies for EGFR may be predictive. Our goal is to identify predictive markers of response that will ultimately be developed into molecular diagnostics to enhance the effectiveness of Cetuximab therapy in CRC. Preclinical approaches using transcriptional profiles of 164 primary CRC tumors and 21 sensitive and resistant CRC cell lines have identified a list of genes that may be predictive of response. A Phase II study of Cetuximab monotherapy in patients with metastatic CRC was conducted to examine whether preclinically identified markers could be validated in a clinical setting. Patients with metastatic CRC were enrolled in a study of Cetuximab monotherapy. RNA was isolated from the pre-treatment core biopsies of all patients and was profiled on Affymetrix gene chips. DNA isolated from the pre-treatment core biopsies is being used for EGFR gene copy number analysis as well as mutation analysis. An interim statistical analysis of transcriptional profiling data was performed to examine whether preclinically identified markers are differentially expressed between patients who derive clinical benefit (Partial Responders PR and Stable Disease SD) and those who do not (Progressive Disease PD). Pathway analysis tools were used to identify biological relationships between the predictive markers identified from this clinical study. Several preclinically identified candidates were able to distinguish between the PR/SD and PD patients with high accuracy. The top candidate sensitivity markers identified from this analysis are key players in the EGFR signaling pathway. Their strikingly higher expression levels in responders suggest that tumors addicted to the EGFR pathway respond to Cetuximab. Approximately twenty top response prediction markers are found to be associated with the EGFR network. The identified response prediction markers have the potential to be used either singly or in combination for patient selection to enhance the effectiveness of Cetuximab therapy. This analysis is currently being extended to include a larger set of patients from this study.