647P - Patient-derived xenograft approach to explore differential responses to diet dependent on genotypic or phenotypic characteristics of colorectal cancer

Background

Colorectal cancer (CRC) results from the combination of genetic alterations and environmental risk factors. Several evidences indicate that the diet may have a role in prevention and progression of CRC, but more research is needed to clarify the heterogeneity of dietary associations with CRC.

Methods

Patient-derived xenografts (PDXs) were generated implanting subcutaneously human CRCs at different staging into immunodeficient mice, and feeding with control normal diet (ND) and classical western diet (WD). Tumor phenotype, transcriptomic, genetic and metabolomic were analyzed.

Results

The response to WD was heterogeneous among all generated PDXs. Only one out of three showed a greater growth in mice fed with WD. Particularly, WD induced transcriptomic perturbations of several genes related to epigenetic regulations and tumor progression, remodeling tumor microenvironment, promoting mucin secretion, mitochondria dysfunction and metabolic reprogramming. Interestingly, the three PDXs responded to the same WD with a distinct metabolomic profile that was validated through the machine learning. With whole exome sequencing, PDXs did not share the same somatic cancer mutations that could play a key role in conditioning response to a specific diet. Consistently, the same heterogeneous response to WD in tumor growth was observed in xenografts generated with CRC cell lines. To strictly mimics human diet, we tested in comparison to classic WD, humanized WD diet that displayed a different tumor behavior. The analysis of patient derived organoids developed from PDX fed with humanized diets revealed that the altered metabolic reprogramming acquired in vivo is also maintained in vitro in absence of any stimulus, suggesting a metabolic memory of cancer cells.

Conclusions

Specific CRC somatic mutations, modifying several signaling pathway and tumor metabolism promote tumor growth and progression, and could be predictive of a response to a specific diet. In this scenario, PDX could be considered as a suitable approach to study the response to the diet and to provide a screening of different diets in order to create a dietary guidance for CRC patients.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Humanitas University.

Funding

PRIN.

Disclosure

All authors have declared no conflicts of interest.