Deconstructing the metabolic phenome of molecularly defined subsets of sporadic colorectal cancer to identify novel metabolic vulnerabilities.

Our understanding of the complexity of the cancer genome, as well as the clonal evolutionary processes from which cancer emerges has changed our perception from “one gene, one drug” approach to a “multi-gene, multi-drug” model to guide therapeutic strategies (Dienstmann et al., 2017). Metabolic reprogramming is recognised as an emerging hallmark of cancer development. Hence, just as gene expression-based classifications for CRC (Guinney et al., 2015) provide the basis for subtype-specific targeted therapies, extracting a comprehensive overview of the metabolic properties of molecularly-defined subtypes of CRC can provide a fundamentally new dimension to cancer diagnosis and treatment.

Our preliminary mass spectrometry-based imaging data of human primary tumours indicate that there is evidence for distinct metabolic subtypes even within the same tumour reflecting metabolic intratumour heterogeneity. We hypothesise that to some extend this heterogeneity is driven by the inherent cellular genetic diversity but is also influenced by changes in the microenvironment. Characterisation of metabolically distinct subtypes of CRC and the genetic and microenvironment determinants that underlie them will allow the identification of novel metabolic vulnerabilities and ultimately more effective therapeutic interventions.

This project is part of the CRUK Grand Challenge Award funded program that aims to map the entire molecular make-up of tumours using the latest technologies in high-throughput mass spectrometry imaging.

The first objective of this study will be to unravel the metabolic phenome of CRC by performing an extensive mass-spec metabolomics profiling of well-characterised cell lines and primary tumours. This will be followed by extensive genomic profiling of the metabolically distinct sub-regions using next generation sequencing (NGS) and expression microarrays with the focus to unveil novel gene alterations-metabolic phenotype interactions that are associated with molecularly-defined subtypes of CRC. In parallel, we will aim to generate a collection of colorectal cancer patient-derived xenografts (PDXs) and derive matched organoid models, aiming at recovering distinct metabolic phenotypes in pre-clinical models for genetic and pharmacological inhibition screening. The comparison of GEMMS with organoids/PDXs and clinical samples will allow us to make an assessment on model stability and heterogeneity and establish appropriate models for understanding the therapy sensitivity pattern of representative colorectal subytpes.

Probabilistic graphical models can assist with understanding metabolic heterogeneity and building a bridge between biology and computational modelling. The third objective of our first aim will be to study the evolutionary aspects of key metabolic changes e.g. hypoxia, acidosis using computational modelling based on parameters obtained from mass-spec imaging data. This will give us an opportunity to gain understanding of how these metabolic changes might affect tumour growth and notably also the response to therapy. Finally, Aim 2 will uncover and biochemically dissect the signalling nodes that contribute in the metabolic restructurings of a representative subtype of CRC e.g. tumours with high microsatellite instability with the goal to identify novel metabolic vulnerabilities. Novel therapeutic interventions will be tested across all available PDO and representative PDX models with or without CRC tumours with mismatch repair deficiency (Poulogiannis et al., 2010) alone and/or in combination with inhibition of co-dependenet e.g. DNA repair pathways. By combining this in-depth metabolic/genetic analysis with thorough investigation of the molecular mechanisms underlying gene function, we hope to be able to exploit weaknesses in cancer's armour and ultimately provide an entirely novel repertoire for CRC diagnosis and treatment.