Colorectal cancer (CRC) is a frequently lethal disease with heterogeneous outcomes and drug responses. To resolve inconsistencies among the reported gene expression–based CRC classifications and facilitate clinical translation, we formed an international consortium dedicated to large-scale data sharing and analytics across expert groups. We show marked interconnectivity between six independent classification systems coalescing into four consensus molecular subtypes (CMS) with distinguishing features: CMS1 (MSI Immune, 14%), hypermutated, microsatellite unstable, strong immune activation; CMS2 (Canonical, 37%), epithelial, chromosomally unstable, marked WNT and MYC signaling activation; CMS3 (Metabolic, 13%), epithelial, evident metabolic dysregulation; and CMS4 (Mesenchymal, 23%), prominent transforming growth factor β activation, stromal invasion, and angiogenesis. Samples with mixed features (13%) possibly represent a transition phenotype or intra-tumoral heterogeneity. We consider the CMS groups the most robust classification system currently available for CRC – with clear biological interpretability – and the basis for future clinical stratification and subtype–based targeted interventions.

Cancer cells consume large quantities of glucose and primarily use glycolysis for ATP production, even in the presence of adequate oxygen 1,2. This metabolic signature (aerobic glycolysis or the Warburg effect) enables cancer cells to direct glucose to biosynthesis, supporting their rapid growth and proliferation 3,4. However, both causes of the Warburg effect and its connection to biosynthesis are not well understood. Here we show that the tumor suppressor p53, the most frequently mutated gene in human tumors, inhibits the pentose phosphate pathway (PPP) 5. Via the PPP, p53 suppresses glucose consumption, NADPH production, and biosynthesis. The p53 protein binds to glucose-6-phosphate dehydrogenase (G6PD), the first and rate-limiting enzyme of the PPP, and prevents the formation of the active dimer. Tumor-associated p53 mutants lack the G6PD-inhibitory activity. Therefore, enhanced PPP glucose flux due to p53 inactivation may increase glucose consumption and direct glucose toward biosynthesis in tumor cells.

Colon cancer is a clinically diverse disease. This heterogeneity makes it difficult to determine which patients will benefit most from adjuvant therapy and impedes the development of new targeted agents. More insight into the biological diversity of colon cancers, especially in relation to clinical features, is therefore needed. We demonstrate, using an unsupervised classification strategy involving over 1,100 individuals with colon cancer, that three main molecularly distinct subtypes can be recognized. Two subtypes have been previously identified and are well characterized (chromosomal-instable and microsatellite-instable cancers). The third subtype is largely microsatellite stable and contains relatively more CpG island methylator phenotype-positive carcinomas but cannot be identified on the basis of characteristic mutations. We provide evidence that this subtype relates to sessile-serrated adenomas, which show highly similar gene expression profiles, including upregulation of genes involved in matrix remodeling and epithelial-mesenchymal transition. The identification of this subtype is crucial, as it has a very unfavorable prognosis and, moreover, is refractory to epidermal growth factor receptor-targeted therapy.