26P - Pre-clinical modelling and treatment of BRAF mutated colorectal cancer

Background

10% of colorectal cancers (CRC) have a BRAF mutation (BM). This is a poor prognostic group with distinct clinical characteristics. Robust preclinical models that reflect this CRC subtype are currently lacking. We have developed a novel, aggressive, immune competent, genetically engineered mouse model (GEMM) that aligns well with human BM CRC.

Methods

4 genes were targeted with the villinCre^ER: activating Braf ^V600E/+ (B), deletion of Trp53 ^fl/fl (P), deletion of ALK5^fl/fl and NOTCH1 intracellular domain expression Rosa26 ^N1icd/+ (N). This generated BPNA mice. Mice were induced and aged until clinical endpoint. Clinically relevant treatments were benchmarked, including BRAFi (Dabrafenib) combined with EGFRi (Erlotinib). Mouse tumour organoids were derived and high throughput drug screening was performed.

Results

BPNA mice developed endpoint tumours in the proximal colon in a short time period (37 DPI). Mice had high rates of T4 disease (57%) and, despite short latency, developed peritoneal metastasis in 5/27 mice (19%). Primary tumours had proficient mismatch repair (pMMR). Tumours classified as CMS4, aligned with the BM1 transcriptional subtype and displayed regenerative cancer stem cell markers. Therapeutic interventions based on the BEACON CRC (BRAFi+EGFRi) combinations were tested. Both BRAFi and EGFRi as monotherapies had no significant survival advantage (OS) in vivo. However in combination, mice lived twice as long with median OS of the combination group 66 DPI vs 32 DPI in the vehicle control (log-rank p<0.001). DUSP6 was only significantly suppressed in the BRAF/EGFRi arm suggesting effective MAPK targeting requires combination therapy. Through organoid drug screening, 28/166 compounds were identified as hits of which 3/5 top hits were proteasome inhibitors.

Conclusions

The BPNA GEMM aligns with human disease in terms of primary tumour characteristics and metastatic pattern. Mice also respond to standard of care therapeutics similarly to humans creating a good benchmark for further investigation. Organoid drug screens have highlighted a potentially novel drug class for BM CRC treatment. Future work will utilize these mice to test therapeutics enabling robust pre-clinical evidence to de-risk future novel early phase trials for BM CRC.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Cancer Research UK.

Disclosure

O.J. Sansom: Financial Interests, Institutional, Funding: Novartis, Cancer Research Technology, Redx. All other authors have declared no conflicts of interest.