"Darwinian" tumor-suppression model unsupported in clinical experience.

To the editor: A recent article in Nature, with accompanying News and Views commentary, proposed a “Darwinian” tumor-suppression model of T-cell leukemia (Figure 1) and extrapolated how this may account for leukemogenesis in gene therapy trials for X-linked severe combined immunodeficiency (SCID-X1).1,2 However, we believe that this hypothesis is unsupported by clinical experience in humans in either the gene therapy or allogeneic haematopoietic stem cell transplant (SCT) settings. In the paper, by Martins et al.,1 data presented from murine bone marrow stem cell transplantation studies supported a compelling hypothesis whereby T-cell leukemia developed in the thymus of animals as a result of imbalanced competition between “young” bone marrow–derived stem cells and locally resident “older” T-cell precursors. In the normal situation, bone marrow–derived stem cells continually migrate to the thymus and seed T-cell precursor niches, where self-renewing progenitors can sustain long-term thymopoiesis and support a diverse T-cell repertoire. Figure 1 ”Darwinian” tumor suppression model in mice. (a) T cells develop in the thymus from progenitor cells. The thymus is continuously supplied with progenitor cells from the bone marrow, which replace thymus-resident progenitor cells. Martins ... Martins et al. demonstrated that, in the absence of competition from bone marrow–derived stem cells, resident thymic progenitors in mice are prone to transformation leading to T-cell leukemia. However, the authors proceeded to extrapolate how their murine model might be applicable to gene therapy trials of SCID-X1 in Paris and London, in which 5 of 20 infants developed T-cell leukemia following ex-vivo gamma-retroviral modification of autologous bone marrow stem cells.3,4 If the hypothesis were correct, it would have important and influential consequences on the design and conduct of gene therapy trials for similar conditions. Unfortunately, contradictory evidence derived from more than 40 years of clinical experience in allogeneic SCT for similar disorders was not considered. In humans, SCID-X1 provides a unique host environment that can accommodate donor engraftment even in the absence of preparative chemotherapy. Although host bone marrow stem cells may not be replaced, donor-derived precursors can repopulate vacant thymic niches and support T-cell recovery. We recently reported evidence of long-term thymopoiesis in 53 subjects with “engraftment-permissive” SCID (the majority were SCID-X1) who became high-level T-cell chimeras without significant B-cell or myeloid chimerism after transplantation.5 Sustained, single-lineage donor T-cell recovery in these subjects was indicative of engraftment in thymic, but not bone marrow, niches, a scenario mirrored in specific gene therapy trials in which there was no preconditioning to facilitate engraftment. Yet in these children, and several hundred similar transplant recipients worldwide, there have been no reports of T-cell malignancy. If the “Darwin” suppression hypothesis were applicable, the absence of dominant suppressive effects from bone marrow–derived progenitors in the thymus might have been expected to be associated with an accelerating incidence of T-cell leukemia over time, and this has not been the case. As importantly, T-cell leukemic transformation has arisen not only in unconditioned SCID-X1 patients undergoing retroviral gene therapy but also following chemotherapy-conditioned gene therapy procedures, most notably in seven of nine engrafted subjects treated for Wiskott–Aldrich syndrome.6 According to the proposed hypothesis (if operating as the primary driver), ongoing migration of gene-corrected bone marrow stem cells to the thymus in these children should have suppressed any T cells liable to transformation following retrovirus-mediated mutagenesis. Given the current molecular evidence in humans and in animal models, the most likely explanation is that mutagenesis in retroviral gene therapy studies is primarily attributable to vector-mediated upregulation of proto-oncogenes rather than to evolution of thymic autonomy in the absence of competitive bone marrow–derived populations. Insertion of gamma-retroviral genomes close to known oncogenes has caused transformation of T-cell precursors and has also resulted in acute myeloid leukemia in chronic granulomatous disease in which thymic suppression was of no relevance.7 Thus, although the newly published animal data for competitive tumor suppression appear compelling, a similar link in humans remains elusive and cannot be extrapolated from experiences in either gene therapy or allogeneic transplantation.