Abstract
Glioblastoma (GBM) is the most common and lethal malignant brain cancer in adults. Immunotherapy has emerged as a potentially powerful approach to achieve long-term survival in patients with GBM. Antigen presenting cells (APCs) play a central role in priming cancer-specific immune responses due to their ability to sample and present tumor neoantigens to the immune system. We hypothesis that antigen presenting cells could be induced by transdifferentiated from GBM in-situ so that these locally created APC will find themselves pre-planted in the tumor microenvironment (TME) with full access to tumor neoantigens. Top ten ranked candidate fate determinants for the GBM-DC transdifferentiation were predicted by NETZEN, an integrated deep-learning and gene network-based ranking artificial intelligence (AI) platform for precision medicine. We successfully transdifferentiated the murine GBM cell line GL261 into CD45+ immune cells accompanying morphological changes to less adhering cells by a combination of four factors (PU.1, IRF8, BATF3, ID2). Of these induced CD45 positive cells, a significant fraction also expresses high levels of the myeloid marker CD11b and antigen presenting molecules MHCII and MHCI by flow cytometry, suggesting these induced CD45+ cells are myeloid lineage APCs (iAPC).These iAPC exhibit phagocytic property, tested by incubating pHrodo Red bioparticles conjugated with Zymosan, an antigen found on the surface of fungi. More importantly, these induced iAPC appear to have lost their proliferative capacity characteristic of the parental GBM cells. Total live cell numbers were significantly reduced in 4F-induced culture compared to the EV control. In conclusion, we successfully transdifferentiated mouse GBM cells into APC-like cells based on NETZEN prediction and our work can potentially provide a novel therapeutic approach for developing an in-situ APC vaccine immunotherapy for GBM, and for other cancers.