We have previously proposed the CD20 molecule as a novel suicide gene for T lymphocytes in the context of allogeneic bone marrow transplantation, because CD20 can be used both as a selection marker and as a killer gene after exposure to the anti-CD20 therapeutic antibody rituximab. We now report on preclinical studies using this novel system, in which the best transduction protocol, reproducibility, yield, feasibility, and functionality of the transduced T lymphocytes have been investigated with a large donor series. Wild-type human CD20 cDNA was transduced into human T lymphocytes, using a Moloney-derived retroviral vector. Alternative protocols were tested by employing either one or four spinoculations (in which cells are centrifuged in the presence of retroviral vector supernatant) and stimulating T cells with phytohemagglutinin (PHA) or anti-CD3/CD28. One spinoculation alone was sufficient to obtain approximately 30% CD20-positive cells within four experimental days. Four spinoculations significantly increased transduction to 60%. A small difference in transduction efficiency was observed between the two stimulation methods, with PHA being superior to anti-CD3/CD28. Transduced cells could be purified on immunoaffinity columns, with purity reaching 98% and yield being on average 50%. Finally, 86-97% of immunoselected T lymphocytes could be killed in vitro with rituximab and complement. More importantly, the CD20 transgene did not alter the functionality of T lymphocytes with respect to allogeneic recognition and cytotoxic response, anti-Epstein-Barr virus cytotoxic response, antigenic response to tetanus toxoid antigen, interleukin 2 (IL-2), IL-4, and interferon gamma production; chemotaxis in the presence of stromal cell-derived factor 1, phenotype for several activation markers including HLA-DR, CD25, CD69, and CD95, and T cell repertoire.