Abstract B14: Inherent resistance to BRAF inhibitors in BRAFV600E colorectal tumors is driven by ligand-dependent activation of the EGFR pathway

Abstract

Abstract Constitutive activation of the RAS, RAF, mitogen-activated protein kinase (MEK), and extracellular signal regulated kinase (ERK) pathway drives tumor cell proliferation and survival in many cancer types, and activating mutations in KRAS or BRAF occur in over 30% of human tumors. It has been demonstrated that BRAF inhibitors are effective selectively against BRAFV600E tumors and not RAS mutant tumors, despite BRAF functioning as a key effector downstream of RAS and upstream of MEK. Recent Phase I clinical data with RAF inhibitors has demonstrated a significant response rate (81%) in metastatic melanoma patients with BRAFV600E-positive tumors (N Engl J Med. 2010;363:809–19). Interestingly, in contrast to this remarkable response in melanoma, the response rates in BRAFV600E-positive colon tumors have been low in comparison (5.2%) (Kopetz et al., ASCO 2010). It has recently been reported that elevated EGFR levels in a subset of BRAFV600E lines render them resistant to BRAFV600E selective inhibitors (Prahallad et al., Nature. 2012; Corcoran et al., Cancer Discovery. 2012). However, the kinetics of EGFR pathway activation in response to inhibitors and underlying mechanisms of resistance are not entirely clear. In this study, we demonstrate that BRAFV600E/EGFR high tumors show a biphasic response to BRAF inhibitors, whereby the MAPK pathway is effectively blocked upon acute treatment, but subsequently rebounds and becomes resistant to inhibition. Components of the negative feedback loop downstream of the MAPK pathway including Sprouty, MKP3, and CDC25C are downregulated during acute treatment but rebound upon prolonged treatment, coincident with phospho-ERK suppression and subsequent reactivation. We show that this effect is not only dependent on high EGFR levels but also requires EGFR ligand stimulation that can be mediated through autocrine signaling. Resistance to this pathway is associated with increased RAS-GTP levels driving wild-type RAF activation upon extended treatment with B-RAF inhibitors. Importantly, immunohistochemical staining of tissue microarrays representative of both BRAFWT and BRAFV600E mutant CRC tumors with an EGFR antibody illustrates that the BRAFV600E mutation in colorectal cancer is associated with a significantly higher percentage of tumors carrying membrane EGFR compared to BRAFWT CRC tumors, with the median of membrane EGFR positive cells in BRAFV600E tumors 56.5% vs. 15% of cells in BRAFWT tumors, suggesting that activation of both pathways may be required for tumorigenesis in this tissue type. Finally, using both in vitro and in vivo models, we demonstrate that, in BRAFV600E-positive colon tumors with high EGFR pathway activity, inherent resistance to BRAF-selective inhibitors can be overcome by combination treatment with clinically relevant inhibitors of EGFR signaling, providing a potential treatment strategy for BRAFV600E/ EGFR high tumors. In summary, these results provide mechanistic insight into the crosstalk between BRAFV600E and EGFR signaling in CRC, shed light into the low response rate of BRAFV600E mutant colorectal cancer to BRAF inhibitors as single agents, and provide a rationale for the use of BRAF and EGFR inhibitors as combination therapy in this indication.