Transfection of NK Cells with mRNA or Lentivirus Expressing Chimeric Antigen Receptors Results in Highly Efficient Killing of Lymphoid Malignancies and Compares Favorably with Monoclonal Antibody-Directed ADCC.

Abstract 1696

Poster Board I-722

NK cells frequently do not kill malignant lymphoid target cells except through antibody dependent cellular cytotoxicity (ADCC) when combined with monoclonal antibodies (mAb). Here we compared the ability of the human NK cell line NK-92 to lyse lymphoid cell lines and primary chronic lymphoid leukemia (CLL) cells after transfection with either mRNA or lentivirus coding for Chimeric Antigen Receptors (CAR) against CD19 or CD20. Electroporation (Genepulser, Biorad) of 10 μg mRNA (coding for GFP, CD19-CAR or CD20-CAR) into NK-92 cells was performed at 300V and 150 μF (Leuk. Res 2009;33:1255). For lentivirus transduction, CD19-CAR and CD20-CAR were cloned into a pCL20c IRES-GFP vector and lentivirus stocks for GFP, CD19-CAR and CD20-CAR were obtained from 293T packaging cells. NK-92 cells were transduced by two successive rounds of spinfection in the presence of 8 μg/ml protamine sulfate. Mean transfection efficiencies were 57.2% ± 6.6 for electroporation with mRNA, and 32.6% ± 4.1 for lentivirus transduction. The cytotoxity of the transfected/transduced NK-92 cells against the reference Raji cell line was not affected. The NK-92 resistant cell lines SUP-B15 (CD19+/CD20-) and TMD5 (CD19+/CD20+) became highly sensitive to lysis by transfected NK-92 expressing CD19-CAR, while only TMD5 became sensitive to transfected NK-92 expressing CD20-CAR. Importantly, mRNA-transfected NK-92 showed two-fold higher killing compared to lentivirus-transduced NK-92 despite similar receptor expression. As expected, expression of CAR following mRNA transfection was temporary (48 hours), whereas lentivirus transduced NK-92 maintained CAR expression and cytolytic function for indefinite periods in culture. In order to obtain higher target cell killing by lentivirus transduced cells, NK-92 cells could be highly enriched by cell sorting. Subsequently, we compared CAR-directed killing with ADCC on primary B-CLL patient samples, using NK-92 cells genetically modified by lentivirus expressing anti-CD20 CAR versus high affinity Fc receptor-expressing (FcRγIIIA) NK-92 cells combined with the anti-CD20 monoclonal antibody Rituximab. Initially all CLL cells were resistant to killing by NK-92. Anti-CD20 CAR transfected NK-92 cells showed significantly higher cytotoxicity than ADCC against patients' cells, with less patient-to-patient variability. In conclusions, transfection of NK cells with mRNA coding for CAR against lymphoid surface antigens is a highly efficient method that can easily be scaled up to produce clinical grade material. Lentivirus transduced NK-92 cells show stable expression of CAR, but efficient killing requires sorting or selection for transfected cells. CAR-directed killing by NK-92 compares favorably with anti-CD20 mAb mediated ADCC against primary CLL targets and either approach, alone or in combination, could have clinical relevance.