The problem: which mouse model would be best representative of each human CRC subtypes?
Colorectal cancer (CRC) primary tumours have been molecularly classified into four consensus molecular subtype (CMS1-4)1. While CMS classification provides valuable prognostic information, its ability to identify subtype-specific responses to therapies remains an area of active research. Among all pre-clinical models, genetically engineered mouse models (GEMMs) in particular provide an ideal system to improve pre-clinical drug testing within a native immunocompetent host 23. They represent a valuable tool to test novel treatments that may benefit specific subtypes of tumours, making it essential to ensure the chosen models accurately recapitulate biological signalling and phenotypes underpinning human cancer subtypes 4. Therefore, accurate and robust classification of mouse CRCs according to human subtypes is a critical step to improve disease-positioning of models and translation of findings from the pre-clinical setting.
In this study for the first time, we developed an R package, namely MmCMS, which provides a publicly-available tool to classify mouse tumour tissue using three options (A-C) of increasing complexity, from gene-level to biological pathways. This package provides an important standardised approach for researchers to enable more reproducible and comparable classification of CRC mouse models, aligned to the biology underpinning human CRC subtypes.
Take home message!
- Our MmCMS-B and C options, which are based on broad biological knowledge-based approaches, could overcome the limitations of individual gene-based cancer subtyping method (MmCMS-A).
- We found our biological knowledge-based approaches are less influenced by nonbiological factors such as normalization methods.
- Our results suggest the presence of intra-genotype CMS subtype heterogeneity within mouse tumours, indicating that the same mutational driver events can result in variable downstream transcriptional signalling, emphasising that faithful mouse model alignment with human tumour signalling should not be based on mutation alone.
- Guinney, J. et al. The consensus molecular subtypes of colorectal cancer. Nat. Med. 21, 1350–1356 (2015).
- Roper, J. & Hung, K. E. Priceless GEMMs: genetically engineered mouse models for colorectal cancer drug development. Trends Pharmacol. Sci. 33, 449–455 (2012).
- Belmont, P. J. et al. Cross-species analysis of genetically engineered mouse models of MAPK-driven colorectal cancer identifies hallmarks of the human disease. Dis. Model. Mech. 7, 613–623 (2014).
- Cancer models for reverse and forward translation. Nature cancer vol. 3 135 at https://doi.org/10.1038/s43018-022-00346-5 (2022).