Chimeric antigen receptor modified­T (CAR­T) cells rapidly became an essential tool for CD19+ hematologic cancer treatment. Genetic retargeting of a T cell with aCAR enables a new antigen­specificity through the single­chain variable fragment, which is derived from a tumor­specific antibody. CAR also consists oftransmembrane, co­stimulatory, and CD3z signaling domains. This construct allows T cells to recognize cancer­specific antigens in an MHC­independent manner,facilitates activation and release of cytotoxic cytokines when the target is recognized. There are several limitations to effectively exploit CAR­T cells for solid tumortreatment, including the suppressive activity of tumor stroma and poor migration of effectors to tissue. The aim of the project is to improve GD2, and L1CAM targetedCAR­T cells by co­expression with membrane­anchored cytokines. At this point, the lentiviral transduction efficiency reached approximately 60% with the anti­GD2CAR construct, while CAR+ cells were almost equally represented by CD4+ and CD8+ populations. Bicistronic expression of surface cytokine IL15/IL15Radownstream P2A and anti­GD2 CAR led to a significant decrease in transduction efficiency (15% CAR+; 5% IL15+; 5% IL15Ra+) and remains to be optimized.Although the main advantages of IL15/IL15Ra should be observed in vivo, its expression correlates with better survival of T cells in vitro. Both IL15/IL15Ra positiveand negative anti­GD2 CAR­T cells proved to specifically kill GD2+ target cells of neuroblastoma and glioma origin (IMR­32, T98G) at 5:1 – 20:1 E:T ratio. However,degranulation analyzed by CD107a staining was much lower than that of anti­CD19 CAR­T. Thus, GD2 targeted CAR­T cells armored with membrane form of IL15specifically kill GD2+ cancer cells in vitro and may serve as a prospective tool for neuroblastoma, ganglioneuroblastoma and glioma immunotherapy application.