Three-dimensional (3D) cultures have the potential to increase the predictive value of pre-clinical drug research and bridge the gap towards anticipating clinical outcome of proposed treatments. However, their implementation in more advanced drug-discovery programs is still in its infancy due to the lack of reproducibility and low time- and cost effectiveness. HCT116, SW620 and DLD1 cells, cell lines with distinct mutations, grade and origin, were co-cultured with fibroblasts and endothelial cells (EC) in 3D spheroids. Clinically relevant drugs, i.e. 5-fluorouracil (5-FU), regorafenib and erlotinib, were administered individually to in CRC cell cultures. In this study, we established a robust, low-cost and reproducible short-term 3D culture system addressing the various complexities of the colorectal carcinoma (CRC) microenvironment. We observed a dose-dependent increase of erlotinib sensitivity in 3D (co-)cultures compared to 2D cultures. Furthermore, we compared the drug combination efficacy and drug-drug interactions administered in 2D, 3D and 3D co-cultures. We observed that synergistic/additive drug-drug interactions for drug combinations administered at low doses shifted towards additive and antagonistic when applied at higher doses in metastatic CRC cells. The addition of fibroblasts at various ratios and EC increased the resistance to some drug combinations in SW620 and DLD1 cells, but not in HCT116. Retreatment of SW620 3D co-cultures with a low-dose 3-drug combination was as active (88% inhibition, relative to control) as 5-FU treatment at high dose (100 μM). Moreover, 3D and 3D co-cultures responded variably to the drug combination treatments, and also signalling pathways were differently regulated, probably due to the influence of fibroblasts and ECs on cancer cells. The short-term 3D co-culture system developed here is a powerful platform for screening (combination) therapies. Understanding of signalling in 3D co-cultures versus 3D cultures and the responses in the 3D models upon drug treatment might be beneficial for designing anti-cancer therapies.