A 3D microtumour system that faithfully represents ovarian cancer minimal residual disease

BackgroundBulk cancer and minimal residual disease (MRD) are characterised by different molecular drivers and therefore necessitate different therapeutic strategies. However, there are currently no 3D models that can faithfully recapitulate MRD ex vivo for therapy development. MethodsA microfluidic technique was implemented to construct 3D microtumours, in which tumour cells, either by themselves or with fibroblasts, were encapsulated in viscous hydrogels. The 3D microtumours were analysed for their response to first-line chemotherapeutics and characterised through RNA-Seq, by comparing them to both 2D cultures and clinical samples. ResultsOur microfluidic platform guarantees the fabrication of 3D microtumours of tailorable size and cell content, which recreate key features of tumours such as hypoxia, characteristic organization of the cytoskeleton and a dose-response to chemotherapeutics close to the physiological range. The 3D microtumours were also used to examine non-genetic heterogeneity in ovarian cancer and could fully reflect the recently described "Oxford Classic" five molecular signatures. The gene expression profile of 3D microtumours following chemotherapy treatment closely resembled that of MRD in ovarian cancer patients, showing the upregulation of genes involved in fatty acid metabolism. We demonstrate that these 3D microtumours are ideal for drug development by showing how they support the identification of a promising inhibitor of fatty acid oxidation, perhexiline, which specifically targets chemotherapy-resistant MRD ovarian cancer cells and not bulk cancer cells. ConclusionWe have obtained the first 3D model of ovarian cancer MRD by using microtumours generated through microfluidics. This system is ideal for high-throughput drug screening and, given its versatility, it can be readily extended to additional types of cancer, as well as accommodate multiple cell types to generate complex tumour microenvironments.