Overcoming adaptive feedback resistance to KRAS inhibition in colorectal cancer - Project Summary/Abstract Although KRAS is mutated in 20% of all cancers and 40% of colorectal cancer (CRC), it has long been considered an “undruggable” target 1. Recently, novel covalent inhibitors selective for KRASG12C have entered the clinic, offering the unprecedented opportunity to target KRAS directly, and other mutation-specific KRAS inhibitors (i.e. G12D) are under development 2,3. However, prior efforts to target the RAS-MAPK pathway have been hampered by adaptive feedback, which drives pathway reactivation and resistance, particularly in CRC. For example, BRAF inhibition in BRAFV600 CRC leads to loss of ERK-dependent negative feedback and RTK- mediated pathway reactivation, leading to response rates of only ~5%, compared to ~35% in lung cancer and >50% in melanoma 4,5. Similarly, while early clinical data with KRASG12C inhibitors show promising response rates of >35% in lung cancer, response rates in CRC appear much lower (~10%) with limited durability, suggesting a similar mode of adaptive resistance may be operant in KRASG12C CRC 2,3. In support of this hypothesis, our preliminary studies have suggested that robust adaptive feedback signals lead to rapid pathway reactivation and lack of response in KRASG12C CRC models 6. However, prior studies in BRAFV600 CRC—including preclinical and clinical collaborations between Drs. Corcoran and Kopetz—have demonstrated that combination therapies targeting adaptive feedback signaling (e.g. EGFR) can improve clinical outcome, with the first such combination FDA-approved this year (Corcoran et al, Cancer Discovery 2018; Kopetz et al, NEJM, 2019)7-10. Similarly, our preliminary data support the importance of targeting adaptive feedback in KRASG12C CRC, but suggest complex feedback signaling that will require strategies beyond targeting EGFR to optimize outcome. Here, we propose to define the key mechanisms of resistance to KRAS inhibition in CRC and devise therapeutic strategies to overcome resistance. To accomplish this goal, we propose to leverage a unique collection of ~100 patient-derived CRC organoids and a bank of ~300 CRC PDXs, generated through the MGH/MIT/Broad U54 DRSC and the MDACC U54 PDXNet teams, respectively. We will deploy these novel tools to comprehensively map the adaptive feedback response to KRASG12C inhibition in vivo using clinically-relevant PDX and patient- derived organoid xenografts (PDOX) CRC models. In parallel, we will model the evolution of resistance in vivo to evaluate the potential role of RTK plasticity in driving resistance to specific KRAS inhibitor combinations and will identify candidate mechanisms of acquired resistance through genomic analysis of serial tumor biopsies and cfDNA from CRC patients on KRAS inhibitor combination trials. Utilizing this enhanced mechanistic understanding, we will devise and test novel therapeutic strategies in vivo in our patient-derived models.
