Limiting numbers of G156A O(6)-methylguanine-DNA methyltransferase-transduced marrow progenitors repopulate nonmyeloablated mice after drug selection.

The limited efficacy of hematopoietic gene therapy can be improved by in vivo selection for transduced long-term repopulating cells (LTRC). We selected for G156A MGMT (triangle upMGMT) transduced LTRC present in 5 x 10(4) to 100 x 10(4) marrow cells infused into nonmyeloablated mice by the administration of O(6)-benzylguanine (BG) and BCNU every 3 to 4 weeks. To facilitate engraftment, mice were given a nonablative dose of BG and BCNU before infusion. Without selection, triangle upMGMT was not detected in any hematopoietic colony-forming units (CFU) 24 to 30 weeks after infusion. After BG and BCNU, triangle upMGMT(+) CFU were frequently detected, and their proportions increased with each treatment cycle. After 2 to 3 cycles of BG and BCNU, many mice were stably reconstituted with 75% to 100% triangle upMGMT(+) CFU for at least 6 months, representing up to 940-fold enrichment. Thus, BG and BCNU stem cell toxicity allows triangle upMGMT-transduced LTRC to repopulate the bone marrow. This degree of selection pressure in nonmyeloablated mice is far greater than that observed in previous drug-resistance gene transfer studies. These data support our approved clinical trial to select for drug-resistant, transduced hematopoietic cells, potentially decreasing cumulative drug-induced myelosuppression in patients with cancer. These data also suggest that triangle upMGMT may be a potent, dominant, selectable marker for use in dual gene therapy.