Abstract Background: Activating mutations in KRAS, NRAS, and BRAF are known to cause resistance to anti-EGFR therapy. However, only ~40% of colorectal cancer (CRC) patients with RASWT tumors respond to that treatment. We aim to leverage tumor transcriptomes to predict response to anti-EGFR antibody (e.g., cetuximab) treatment and understand intrinsic resistance mechanisms in CRC. Experimental Design: Transcriptomic profiles with RAS mutation status from two clinical (Okita et al., n = 135; Khambata-Ford et al., n = 68), and preclinical cohorts of CRC cell lines (Medico et al., n = 146) and PDX models (Bertotti et al., n = 216) treated with cetuximab were downloaded from Gene Expression Omnibus (GEO). Each cohort was divided into RASWT and RASMut groups, and transcriptomic profiles were used to assign consensus molecular subtypes (CMS) to each sample. Gene Set Enrichment Analysis (GSEA) was performed to identify resistant pathways in each CMS RASWT group. Results: Restricting to RASWT patients (n = 80) in Okita et al. cohort, we observed that CMS2 tumors had significantly higher disease control rates (DCR) [92% (33/36); chi-square p = 0.03] with cetuximab in combination with doublet chemotherapy relative to CMS1 [50% (2/4)], CMS3 [59% (13/22)], and CMS4 [83% (15/18)] tumors. RASWT CMS2 tumors (n = 43) also showed significantly higher DCR [68% (15/22); chi-square p = 0.03] with single-agent cetuximab compared to CMS1 [0% (0/4)], and CMS4 [29% (5/17)] tumors in Khambata-Ford et al. cohort. In multivariate analysis including both CMS and tumor sidedness, only CMS2 (HR = 0.62, p = 0.03) was significant predictor of cetuximab response in Okita et al. cohort. CMS1 tumors showed worst treatment response irrespective of tumor sidedness [DCR, left: 50% (1/2) vs right: 50% (1/2)]. In contrast both left and right-sided CMS2 tumors responded well [DCR, left: 91% (31/34) vs 100% (2/2)] to cetuximab plus chemotherapy [PFS, 7.9 vs 7 months; log-rank test p = 0.7]. In microsatellite stable (MSS) RASWT cell lines (n = 27), CMS2 cells lines were most sensitive [59% (7/17)] relative to CMS4 [0% (0/10)]. In MSS RASWT PDX models (n = 60), CMS2 had highest DCR [75% (18/24)] than CMS1 [36% (4/11)], CMS3 [55% (5/9)], CMS4 [62% (10/16)]. We found Myc, E2F, and mTOR gene sets/pathways were consistently enriched in resistant patients, cell lines, and PDX models (Normalized enrichment score > 1, FDR < 0.25). Small molecule inhibitors of Myc (JQ1), E2F (AZD7762 & MK-8776), and mTOR (Everolimus) pathways in resistant CRC cell line (HT55) exhibited additive, but not synergistic, effects with cetuximab. Conclusions: These data suggest that tumor transcriptional profiles are a better predictor of anti-EGFR response than tumor-sidedness. Use of CMS to predict anti-EGFR response should be validated in a larger, prospective study. Activation of Myc, E2F, and mTOR pathways are associated with cetuximab resistance in colorectal cancer cells. Citation Format: Saikat Chowdhury, Ria Gupta, Valsala Haridas, Mohammad A. Zeineddine, Scott Kopetz, John Paul Shen. Consensus molecular subtypes (CMS) of colorectal cancer predict anti-EGFR response irrespective of tumor sidedness [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1246.
Read full abstract