Colorectal cancer (CRC) is a heterogeneous disease that can currently be subdivided into four distinct consensus molecular subtypes (CMS) based on gene expression profiling. The CMS4 subtype is marked by high expression of mesenchymal genes and is associated with a worse overall prognosis compared to other CMSs. Importantly, this subtype responds poorly to the standard therapies currently used to treat CRC. We set out to explore what regulatory signalling networks underlie the CMS4 phenotype of cancer cells, specifically, by analysing which kinases were more highly expressed in this subtype compared to others. We found AKT3 to be expressed in the cancer cell epithelium of CRC specimens, patient derived xenograft (PDX) models and in (primary) cell cultures representing CMS4. Importantly, chemical inhibition or knockout of this gene hampers outgrowth of this subtype, as AKT3 controls expression of the cell cycle regulator p27^KIP1. Furthermore, high AKT3 expression was associated with high expression of epithelial-mesenchymal transition (EMT) genes, and this observation could be expanded to cell lines representing other carcinoma types. More importantly, this association allowed for the identification of CRC patients with a high propensity to metastasise and an associated poor prognosis. High AKT3 expression in the tumour epithelial compartment may thus be used as a surrogate marker for EMT and may allow for a selection of CRC patients that could benefit from AKT3-targeted therapy.

Previously, colorectal cancer (CRC) has been classified into four distinct molecular subtypes based on transcriptome data. These consensus molecular subtypes (CMSs) have implications for our understanding of tumor heterogeneity and the prognosis of patients. So far, this classification has been based on the use of messenger RNAs (mRNAs), although microRNAs (miRNAs) have also been shown to play a role in tumor heterogeneity and biological differences between CMSs. In contrast to mRNAs, miRNAs have a smaller size and increased stability, facilitating their detection. Therefore, we built a miRNA‐based CMS classifier by converting the existing mRNA‐based CMS classification using machine learning (training dataset of n = 271). The performance of this miRNA‐assigned CMS classifier (CMS‐miRaCl) was evaluated in several datasets, achieving an overall accuracy of ~ 0.72 (0.6329–0.7987) in the largest dataset (n = 158). To gain insight into the biological relevance of CMS‐miRaCl, we evaluated the most important features in the classifier. We found that miRNAs previously reported to be relevant in microsatellite‐instable CRCs or Wnt signaling were important features for CMS‐miRaCl. Following further studies to validate its robustness, this miRNA‐based alternative might simplify the implementation of CMS classification in clinical workflows.