|Home | About | Journals | Submit | Contact Us | Français|
Reports of neoplasia related to insertional activation of protooncogenes by retroviral vectors have raised serious safety concerns in the field of gene therapy. Modification of current approaches is urgently required to minimize the deleterious consequences of insertional mutagenesis. In this issue of the JCI, Adjali and colleagues report on their treatment of SCID mice lacking the 70-kDa protein tyrosine kinase, ZAP-70, with direct intrathymic injection of a ZAP-70–expressing T cell–specific lentiviral vector, which resulted in T cell reconstitution. Using lentiviral vectors and in situ gene transfer may represent a safer approach than using retroviral vectors for ex vivo gene transfer into HSCs, avoiding 3 factors potentially linked to leukemogenesis, namely HSC targets, ex vivo transduction and expansion, and standard Moloney leukemia virus–based retroviral vectors.
SCID disorders are a family of genetically determined conditions characterized by a block in T cell development and function, variably associated with defects in other hematopoietic lineages. The genetic and phenotypic heterogeneity among these disorders is considerable. Affected children present with severe opportunistic infections due to a lack of functional mature T cells. The most serious forms of SCID are fatal in the first years of life (1). The only curative standard of therapy is an allogeneic bone marrow transplant from an HLA-matched, related donor. If this is not possible, bone marrow transplantation from an HLA-matched, unrelated or a parental haploidentical donor is performed. The survival rate after HLA-matched transplantation is greater than 75% but falls markedly with the use of alternative donors, including haploidentical family donors and unrelated donors, and chronic disabling complications such as graft-versus-host disease are common. Some children reach full functional immunity following transplantation, but others fail to have their B cell function restored or can experience progressive loss of T cell immunity and even complete graft failure over time. These poor outcomes are particularly common when children are transplanted with non–HLA-matched donor cells, for instance from a parent or an unrelated donor (2, 3).
These unsatisfying outcomes, particularly with alternative donor transplantation, have stimulated intense interest in gene therapy approaches to SCID. The most common SCID subtype is SCID-X1, an X-linked recessive disease characterized by a block in T and natural killer cell differentiation due to mutations in the common γ chain of cytokine receptor complexes. Recently, complete immune reconstitution of young boys with SCID-X1 by transduction of their CD34+ bone marrow cells with an oncoretroviral vector encoding the common γ chain followed by reinfusion of transduced cells without conditioning therapy was reported (Figure (Figure1A)1A) (4, 5). These encouraging results represented the first unequivocal demonstration of the clinical efficacy of gene therapy. However, elation was short-lived, as 3 out of 11 enrolled children in this clinical trial have developed T cell leukemia linked to insertional mutagenesis, specifically, activation of the LIM domain only 2 (known as LMO2) transcription factor locus (6). Thus, at present the risks and potential benefits of gene therapy are being reconsidered. The critical question is whether the high frequency of leukemia in this trial was related to specific features of these children or their target CD34+ cell populations, constitutive expression of this particular signaling transgene, or activity of vector elements such as enhancers or promoters as opposed to the possibility that these serious adverse events represent significant risks from the use of any integrating vector to transduce long-lived stem or progenitor cells.
The integration of replication-incompetent retroviruses was previously thought to be essentially random, and it was estimated that single copies of proviral genomes scattered throughout the genome were unlikely to result in the activation of protooncogenes. Recently, however, the insertion of oncoretroviral vectors has been shown to occur preferentially near transcription start sites of genes, potentially increasing the risk of altering the expression of adjacent endogenous genes, while lentiviral vectors were shown to integrate within genes even more preferentially, but over the entire length of transcription units instead of near transcriptional start sites, in a pattern possibly less likely to affect endogenous gene regulation (7–9). Insertional events related to both retroviral and lentiviral vectors are also enhanced in actively expressed genes. Current lentiviral vectors have deletions of their enhancer and promoter regions and instead use internal transgene-specific regulatory elements. Thus these vectors might be less likely to activate adjoining genomic transcription units than the strong constitutive regulatory elements present in the long terminal repeats of all currently utilized standard retroviral vectors (8). The lack of evidence of insertional leukemogenesis due to HIV integration in AIDS patients suggests that the risk may indeed be much lower for lentiviruses than for murine leukemia virus–derived retroviral vectors, but wild-type HIV can only infect mature T cells, not hematopoietic stem cells. Thus, the comparison may not be fully relevant.
In this issue of the JCI, Adjali and colleagues (10) took a different approach to gene therapy for SCID. In this case, the authors targeted T cell precursors instead of HSCs by direct intrathymic injection of a ZAP-70–expressing T cell–specific lentiviral vector into ZAP-70–/– SCID mice (Figure (Figure1B).1B). ZAP-70 deficiency is a rare autosomal- recessive inherited subtype of SCID, in which T cell development is blocked at the stage of double-positive cells, resulting in an absence of peripheral CD8+ T cells and the presence of circulating CD4+ T cells that are unresponsive to TCR-mediated stimuli in vitro (11). The differentation block in this disease is due to the inability of double-positive cells to propagate TCR signals in the absence of ZAP-70 kinase activity despite tyrosine phosphorylation of TCR-associated ZAP-70 molecules (12). The protein tyrosine kinase ZAP-70 is normally expressed in thymocytes, mature T cells, and natural killer cells (13).
Using the T cell–specific lentiviral vector injected directly into the thymus of ZAP-70–/– mice, Adjali and colleagues demonstrated partial reconstitution of polyclonal and functional T cells in some mice (10). They chose this approach to avoid transduction of HSCs and precursor cells from other hematopoietic lineages that do not express ZAP-70 and thus do not require correction in this disorder; this potentially reduces the risk of insertional mutagenesis by exposing many fewer precursor cells to insertional events. HSCs may be particularly susceptible to insertional mutagenesis due to their prolonged and increased replicative capacity. Many genes controlling self-renewal and proliferation are expressed in HSCs before being shut off during differentiation, and increased expression of genes has been shown to predispose loci to insertions of retroviral and lentiviral vectors (7, 9). Insertions at these sites could therefore constitutively activate genes that are normally shut off during hematopoietic development, and thus contribute to leukemogenesis. By the targeting of more mature cells via intrathymic injection, this risk might be decreased.
The in situ approach also avoids ex vivo culture of target cells, which is desirable for a number of reasons. Ex vivo culture of HSCs, progenitors, and even mature T cells has been found to result in impaired homing capabilities and functional deficits (14–16). Furthermore, ex vivo culture potentially exposes the cells to infectious agents, serum, and other risks related to complex and prolonged manipulations and removes them from their natural microenvironment. Thus, in situ delivery of a gene transfer vector to T cell precursors without any requirement for ex vivo culture is an inherently attractive approach.
Although the lentiviral vector used in this study (10) was designed to be T cell–specific, it also transduced thymic epithelial cells (17). The interaction of thymocytes with thymic epithelial cells is crucial for the development of T cells. Thus the effects on thymic epithelial cell function by ZAP-70 transduction needs to be further evaluated. Changes in the vector design might be necessary to obtain increased T cell specificity. Intrathymic vector injection is technically very challenging both in mice and humans but probably accounted for successful correction in less than a quarter of the mice treated by Adjali et al. (10). In human subjects suffering from SCID, the need for intrathymic injection will be a real hurdle, since in these patients the thymus is very small, usually less than 1 g, and lacking in thymocytes (18). Thus, it might be difficult to establish this approach for clinical application, even if the success rate could be significantly improved via sophisticated imaging modalities and optimization in large animal models.
The observation that long-term reconstitution of functional T cells is achievable using this approach in mice is encouraging. However, even though T cells can persist for many years, permanent reconstitution and continued production of naive T cells is thought to require correction of HSCs with self-renewing capacity. The durability of transduction of thymic early progenitor cells, which is likely responsible for the positive results in this study, is unclear (10). Longer follow-up studies are necessary for clarification. The ability of the T cell population to massively expand in number suggests the possibility that we could reconstitute the T cell compartment with a few gene-corrected cells, as seen after spontaneous mutation (19) or as demonstrated in this study (10). But is this the right strategy for reaching therapeutic levels of gene-corrected cells? Keeping in mind that expansion of a limited number of precursor cells and an abnormal proliferation advantage potentially contributed to the development of leukemia in the SCID-X1 trial, direct injection of vector into the thymus with subsequent marked expansion of cells now expressing ZAP-70 constitutively may have similar risks. The ZAP-70 transgene is a tyrosine kinase that may become oncogenic if its expression is not at physiologic levels in the appropriate hematopoietic lineages. Further modification of the lentiviral vector so that real lineage restriction and physiologic expression level of the transgene are achieved is probably desirable for improved safety. It will also be important to gain a better understanding of the biology of the thymic components in SCID patients in regard to the characteristics of the cells that are being transduced by the intrathymic injections, in particular if this approach is extended to other forms of SCID. Finally, longer term follow-up studies of many more mice as well as any SCID patients enrolled in trials using this approach will be crucial for understanding whether permanent correction is really possible via a non-HSC–based approach and for assessing leukemogenic risk.
See the related article beginning on page 2287.
Conflict of interest: The authors have declared that no conflict of interest exists.