The genetic modification of hematopoietic stem cells (HSCs) holds promise as a therapeutic modality for a variety of disorders, including inherited metabolic disorders, malignancies, and HIV. The recent description of leukemia in children treated with HSC gene therapy1 is a reminder that although this methodology has made significant advances, the ultimate application of a safe and effective stem cell gene therapy requires additional preclinical testing. It needs to be recognized that many HSC gene therapy strategies introduce a foreign gene as part of the therapy, and immune responses to such proteins are not fully understood. It has been demonstrated in studies in nonhuman primates that the delivery of foreign genes using oncoretroviral vectors does in fact elicit an immune response in certain instances, particularly in the setting of nonmyeloablative conditioning. In contrast, in previous reports that have described similar strategies using myeloablative conditioning, such immune responses were not observed, resulting in the conclusion that myeloablation was necessary for tolerance to develop to the transgene product(s). This further emphasizes the importance of suitable large-animal models to evaluate both the safety and efficacy of such approaches.
In this issue, Morris and colleagues (page 492) report the detection of cellular immune responses to foreign gene products in the setting of myeloablation. These observations raise the possibility that previous studies indicating that ablation facilitated the development of tolerance to a transgene product may not have fully appreciated the complexities of gene transfer strategies. In these studies the authors chose to evaluate the use of lentiviral vectors, which can transduce nondividing cells and therefore offer an advantage compared with oncoretroviral vectors. They had previously demonstrated that in baboons that received transplants of oncoretroviral vectors expressing green fluorescent protein (GFP) in a myeloablative transplantation setting, durable GFP expression was observed.2 In contrast, in the studies presented here, using a similar conditioning protocol and lentiviral vectors, labeled cells disappeared relatively rapidly. This unexpected observation prompted the authors to evaluate whether the disappearance of the labeled cells was associated with a cellular immune response. Clear data establishing the evolution of a cellular immune response (cytotoxic T lymphocyte [CTL]) to the transgene product are presented, and development of this immune response was associated with the disappearance of labeled cells. Based on the fact that this was different from what they had observed with lentiviral vectors, the authors evaluated the frequency and phenotype of the transduced cells as well as the behavior of the transduced cells in colony-forming unit (CFU) assays and nonobese diabetic/severe combined immunodeficiency (NOD/SCID) transplants. Neither phenotypic differences nor functional differences in CFU and NOD/SCID assays were observed compared with oncoretroviral vectors. Toxicity of the lentiviral preparations was not observed and the white cell and platelet recovery rates were comparable with those seen in control animals, suggesting that the graft was capable of supporting hematopoietic recovery. The major difference observed was that the lentiviral vectors uniformly produced lower levels of transduction compared with retroviral vectors. This resulted in the animals included in this transplant protocol receiving a lower dose of genetically modified, transgene-expressing cells. Studies by Gur et al have demonstrated that the level of transgene expression may contribute to tolerance and that a high level of expression is required for tolerance of T cells to occur.3 The lower levels of transgene expression in the study by Morris et al using lentiviral vectors may therefore explain the evolution of a cellular immune response, even in the setting of myeloablation. Alternative explanations include the fact that cells transduced with lentiviral vectors may be more immunogenic than those transduced with retroviral vectors, but supportive data are currently lacking. The continuing evolution of improved gene therapy vectors is likely to result in further unexpected observations and reinforces the need for well-executed, large-animal studies to evaluate both safety and efficacy.
