Colorectal cancer (CRC) is a heterogeneous disease, with variable molecular pathogenesis and responses to treatments. About 40–50% of patients with locally advanced CRC exhibit resistance to therapy and develop recurrent disease. Current CRC staging based on histopathology and imaging has a limited ability to predict prognosis, indicating that the development of molecular classifications of CRC is needed to enable effective disease management. Of note, several independent studies have recently proposed new CRC classification systems based on distinct gene-expression profiles. Although these studies identified different numbers of tumour subtypes, they all concurred on the identification of one highly aggressive subtype of CRC, which is characterized by the expression of stem-cell and mesenchymal genes (SSM subtype), and is associated with a poor prognosis. Two studies, published in Nature Genetics, now highlight that the predictive power of the SSM subtype relies on the identification of a gene programme expressed by stromal components rather than epithelial tumour cells.
In the first study, Eduard Batlle and colleagues interrogated patient transcriptomic datasets to assess the relative expression of stromal and epithelial genes in each CRC subtype. Following this approach, the investigators discovered that all poor prognosis subtypes identified by the recently published molecular classifications of CRC rely on the expression of a gene signature by cancer-associated fibroblasts (CAFs) rather than on the expression of genes in the cancer cells. Importantly, Batlle highlights that “the CAF programme is upregulated in patients who, despite being classified within a good prognosis subtype, displayed an elevated risk of relapse after therapy.” Furthermore, Batlle and colleagues showed that this gene programme is induced by TGF-β signalling. “These analyses gave rise to several hypotheses that we tested through a novel technology that enables the in vitro propagation of patient-derived CRC samples in the format of 3D tumour organoids,” explains Batlle. These organoids are close surrogates of the disease and, therefore, a useful tool to study tumour behaviour. In this model, the use of a TGF-β receptor 1 inhibitor to block the crosstalk between cancer cells and the microenvironment in patient-derived organoids and xenografts halted disease progression, preventing the formation of metastases. “We are now completing the development of a prognostic test that improves the identification of patients with a poor prognosis,” says Batlle. Emphasizing the importance of these results, he concludes, “they may pave the way to test anti-TGF-β therapies in poor prognosis patients with CRC in the future.”
This is a preview of subscription content, access via your institution
