Discussion

New target validation approaches are aiming for more reliable target identification early in the therapeutic development process. So far, identification of drug targets and initial target validations are usually performed in two-dimensional (2D) cell culture model systems.
Rational design of 3D co-culture models which reflect, at least in parts heterotypic interactions of a real tumor enables investigations of gene function on the phenotypic impact in a model that more closely resembles tumor growth in vivo. In this respect, we have established a high throughput-compatible technique to discover and analyze cancer gene functions in 3D tumor microtissue model systems composed of cancer cells in combination with stromal cells such as fibroblasts. A cancer cell line expressing a green fluorescent protein
constituently was used to distinguish the response of the whole model from the cancer cell population. Fluorescent intensity directly correlated with cancer cell line proliferation whereas the stromal cell populations did not grow over time. Interestingly, siRNA mediated depletion
of one of the targeted reference gene Kif11/Eg5, a member of the mitotic kinesin like motor protein family resulted in significant growth inhibition only in the microtissue model while they were more resistant to Kif11/Eg5 depletion using 2D monolayer cultures. In addition to the differences in the response to Kif11 knock down more robust data generation with higher z-factors were obtained. The results suggest that growth of certain cancer cells in 3D versus
2D can reveal differential dependencies on specific genes for their survival. Moreover, the results highlights that the high throughput-compatible 3D co-culture models will enable new possibilities in the discovery, characterization and validation of gene functions in key biological and pathological processes.