Development of gene therapy for blood disorders: an update

The field of gene therapy as applied to blood disorders has continued to advance rapidly since my review was completed in 2008.¹  That review provided a historical prospective and described the initial successes in treating severe combined immunodeficiency and the first evidence in humans of the genotoxicity of integrating retroviruses. Further success in clinical trials,^2,3  particularly in the immunodeficiencies,^4-7  and also the development of a successful approach for hemophilia B⁸  are among the recent clinical advances.

Much of the effort to develop clinical gene therapy has focused on viral vector systems for gene transfer.^9-11  Viruses have naturally evolved to insert genetic information into target cells, and the field of gene therapy attempts to take advantage of this proclivity. The basic strategy is to express the individual viral proteins on independent expression cassettes within producer cells and introduce the vector genome with the packaging signals and transgene either by transfection or transduction. There are a number of vector systems that are in wide use both experimentally and clinically as summarized in Table 1.

Vectors based on adenovirus have the advantage of having a large transgene capacity allowing large or multiple gene expression cassettes to be included within the vector genome.^10,12  Adenoviral vectors have had limited use in the treatment of blood disorders, although they have been quite effective in facilitating the generation of viral-specific T-cell populations as discussed in detail in a later section.

rAAVs are less immunogenic than adenoviral vectors and are capable of transducing quiescent cells and establishing a stable episome without vector integration in most cases.^11,13  Only the inverted terminal repeats are required as part of the vector genome, and packaging can be achieved with various strategies to develop high-titer vector preparations. Further, the incidence of preexisting immunity is generally low, particularly for certain serotypes such as serotypes 8 and 9. However, rAAV can carry only a transgene up to 4700 nucleotides. The AAV genome is single stranded, and vector preparations are composed of a mixture of vector particles having 1 of the 2 strands of the virus. Upon transduction, the required annealing of the 2 strands delays gene expression. This limitation can be overcome by utilizing a self-complementary design in which the 2 strands of the transgene are on a single hairpin genome in an inverted orientation, which allows quick assembly into a transcription unit following transduction.^14,15 

The γ-retroviral vectors initially used in clinical trials had an intact long terminal repeat (LTR). The LTR enhance-promoter combination was associated with genotoxicity. However, subsequently self-inactivating vectors in which the enhancer-promoter are not present in the integrated vector genome have been shown to be clinically efficacious in the treatment of certain immunodeficiencies.¹⁶  Lentiviral vectors based on HIV-1 were developed largely in response to the challenge of transducing nonmitotic, hematopoietic stem cells.⁹  Although cytokine activation to trigger the cells into the G1 phase of the cell cycle is still required, mitosis is not required as the preintegration complex of HIV vectors can cross the nuclear membrane without mitosis.⁹  Another advantage of the HIV preintegration complex is that it persists in transduced cells for an extended period prior to integration, thereby allowing a longer window of opportunity for integration to occur. Foamy viruses, another type of retrovirus, do not cause disease and have not infected humans.¹⁷  The genome of foamy viruses integrates diffusely in cellular DNA rather than within genes or at regulatory elements (Table 1). A foamy viral vector has been used to achieve stem cell–targeted gene transfer and resolution of leukocyte adhesion deficiency in a canine model.¹⁸ 

The α-retroviral vectors are being considered for therapeutic gene transfer as well.^19-21  Such vectors integrate more uniformly within the genome, particularly within intergenic regions compared with γ-retroviral vectors, which are found near transcriptional start sites and cytosine guanine dinucleotide islands, and lentiviral vectors, which tend to integrate within genes.²²  This favorable integration pattern has been shown to result in lower genotoxicity compared with lentiviral and γ-retroviral vectors.²⁰  A split-packaging design for self-inactivating α-retroviral vectors has been developed and shown to be capable of generating high-titer vector particles.¹⁹  The therapeutic potential for this class of vectors has been shown by functional correction of chronic granulomatous disease (CGD) in a cell line as well as within transduced human cells that engraft in the mouse model.²¹ 

Integrating retroviruses are inherently mutagenic.^16,22  Enhancer-mediated protooncogene activation has resulted in leukemia in a number of participants in the trials for immunodeficiencies.^4-7,16,22,23  Myelodysplasia has also been observed in participants in a trial for CGD secondary to protooncogene activation.²⁴  Various efforts have been made to improve the safety of integrating retroviral vectors. These include elimination of the enhancer element in the context of a self-inactivating vector design, the use of internal cellular promoters with lower activation potential,^25,26  and the addition of insulator elements to the LTRs of the integrated vector.²⁷  Historically, the chicken β-globin locus insulator²⁸  has been used, but more recent studies have identified a number of additional elements that have insulating potential.^29-31 

Various assays have been used to evaluate the steps taken to improve vector safety, although, as a noted in a recent review, none of the assays is fully predictive.²²  One factor that may affect safety is the relative distribution of vector integrations (Table 1). For example, lentiviral vectors are less likely to integrate into regulatory elements than γ-retroviral vectors.^32,33  However, recent studies suggest preferred regions of integration for both types of vectors, so the difference in integration pattern is presumably only relative with respect to safety.³²  Also, the methodology for determining vector distribution is still imperfect.³⁴  Alternative mechanisms of oncogenesis include alternative splicing, gene inactivation, truncation of cellular messenger RNA or protein, and micro RNA activation.²²  Each of these mechanisms has been observed in animal models. Clonal dominance developed in 1 participant in a gene therapy trial for thalassemia and was associated with interruption of the HMGA2 gene by the vector genome resulting in the generation of a truncated messenger RNA with missing regulatory elements.³⁵ 

Generally, bone marrow transplantation is used to treat severe immunodeficiencies. The outcome for individuals with a matched related donor is outstanding (>90%) but much less satisfactory for individuals who have a matched unrelated donor or a partially mismatched, related donor.^4-6,36,37  For those patients, gene transfer offers a potentially more satisfactory alternative form of treatment. In early trials, T lymphocytes were transduced, but much of the recent work has focused on transducing bone marrow hematopoietic stem cells expressing the CD34+ phenotype.

X-linked severe combined immunodeficiency (X-SCID) reflects the lack of the common γ-chain that is part of several interleukin (IL) receptors including the IL-7 receptor, which is required for T-cell development. Affected patients have deficient T cells and natural killer cells and poorly functional B cells. As recently reviewed,⁴  between 1999 and 2006, 20 individuals with X-SCID were treated in 2 gene therapy trials, 1 in Paris³⁸  and 1 in London (Table 2).³⁹  All subjects lacked an HLA-identical donor. Seventeen of the 20 treated participants are alive and display nearly full correction of the T-cell deficiency by genetically modified T cells when evaluated between 5 and 12 years after the gene transfer procedure. However, half of the trial participants remain on immunoglobulin replacement. The natural killer cell deficiency also persisted. Older participants with hypomorphic mutations responded less well to the gene therapy procedure, possibly because of loss of thymic function with advancing age. Unfortunately, 5 of the participants developed T-cell leukemia within 3 to 6 years after the gene transfer procedure. Four were successfully treated with standard antileukemic therapy, and 1 died of refractory leukemia.^4,6  Vector integration analysis identified insertions near the LM02 protooncogene in 4 participants. The leukemogenic process was thought to be initiated by vector-mediated protooncogene activation,¹  but other mutations must have occurred over time before evolution to full neoplasia. The initial clinical trials closed after the development of leukemia in the 5 participants. Subsequently, a self-inactivating γ-retroviral vector has been developed for treatment of X-SCID,²⁶  and patients are again being enrolled. A self-inactivating lentiviral vector with an insulator element has also been designed and shown directly not to elevate LM02 expression in T cells.²⁷  Two clinical trials utilizing this vector have also been opened and are enrolling participants.

Deficiency in ADA leads to SCID. Over the years, there have been several attempts to perform gene therapy in such individuals (Table 2). The early efforts were unsuccessful, presumably because of inadequate transduction of target T cells or primitive hematopoietic cells. However, since 2000, 40 patients have been treated in Italy, the United Kingdom, and the United States.^40-43  CD34+ cells were transduced with a γ-retroviral vector encoding the ADA gene. Low-intensity conditioning with either busulfan or melphalan was required to allow engraftment of primitive stem cells with subsequent recovery from the immunodeficiency. Integration site analysis demonstrated vector insertions near protooncogenes, including LM02, but none of these patients developed leukemia. Moreover, only minor changes in transcriptional activity were observed in T-cell clones harboring 1 or 2 copies of the vector genome.⁴⁴ 

WAS is an X-linked disorder characterized by microthrombocytopenia, immunodeficiency, eczema, and a proclivity to develop an autoimmune disorder and/or lymphoma secondary to a mutation in the WAS protein (WASp) gene.^5,45  The disease is variably severe depending on the exact nature of the mutation in the WASp gene, but individuals with early onset severe forms have been identified and are the appropriate candidates for attempted gene therapy.⁴⁶  An initial clinical trial was conducted in Hanover, Germany, using a γ-retroviral vector in which the LTR enhancer/promoter combination drove expression of the WASp gene (Table 2).²³  Gene transfer was targeted to bone marrow and/or peripheral blood CD34+ cells ex vivo, and participants received myelosuppression prior to reinfusion of the transduced cells. Although 9 of the 10 participants that were enrolled did well clinically for several years, ultimately, 4 developed leukemia secondary to insertional mutagenesis.^4,7  Subsequently, self-inactivating lentiviral vectors were developed by a number of groups.^25,47-50  Vectors have been tested in which the WASp coding sequences are under the control of various promoters.⁴⁹  Full correction of the WAS phenotype was achieved in the murine model with the vector having a retroviral promoter, whereas only partial correction was achieved with the vector having the 1.6-kb WAS promoter fragment.⁴⁹  We have also found that the γ-retroviral promoter is much stronger than either the EF1α or the WAS 1.6-kb promoter in various model systems.⁵⁰  Several other studies suggested that a 1.6-kb promoter fragment from the WASp gene achieves adequate levels for correction in the preclinical model as well as normal levels of expression in human CD34+ cells.^25,51  Methodology for large-scale manufacture of the lentiviral vector has permitted the initiation of a number of clinical trials.^7,52  A recent report describes correction of the immune deficiency in 3 participants in a clinical trial who have now been followed up to 18 months after the gene therapy procedure.⁵³ 

CGD reflects a deficiency in neutrophil function due to mutations in nicotinamide adenine dinucleotide phosphate oxidase that result in chronic bacterial and fungal infections. The clinical characteristics of the disease have recently been reviewed in detail.⁵⁴  Nicotinamide adenine dinucleotide phosphate is a complex of 5 separate proteins, each of which may be deficient and cause the CGD phenotype, although the most common defect is in gp91^phox.⁴⁴  Early trials done without myeloablation or limited myeloablation resulted in only transient production of genetically modified neutrophils, although a clinical benefit with clearance of a chronic infection was noted in several of the early patients.^1,5,54-56  The initial clinical trials used γ-retroviral vectors with an intact LTR, and not surprisingly, protooncogene activation occurred resulting in myelodysplasia in all of the successfully treated participants (Table 2).^24,57  In contrast to X-SCID, ADA-SCID, and WAS, the gene-corrected cells in CGD do not have an engraftment advantage, and thus, as treatment efforts evolve, regimens are likely to be more fully myeloablative.⁵⁷  In an effort to achieve greater safety, lineage- and stage-restricted lentiviral vectors are being developed for treatment of CGD.^58,59  Other vector systems are also being considered, although they are in much earlier stages of development.^21,60 

Many leukodystrophies are potential candidates for treatment by gene therapy.^61,62  To date, 2 have been studied in detail, X-linked ALD (X-ALD) and metachromatic leukodystrophy (MLD). X-ALD is caused by mutations in the ABCD1 gene that encodes a transporter protein. This deficiency leads to the accumulation of very long chain fatty acids in plasma and tissue and progressive demyelination in the central nervous system (CNS). X-ALD has been successfully treated with hematopoietic stem cell transplantation, although cessation of progression of the disorder occurs 12 to 18 months after treatment. To date, 4 patients have received autologous hematopoietic stem cells transduced with the lentiviral vector encoding the transporter (Table 2).⁶³  Arrest of progression of the disorder occurred in the first 2 patients that were treated.⁶¹  Over time, their percentage of genetically modified myeloid and lymphoid cells stabilized at ∼10%. The favorable outcome following stem cell transplantation, whether allogeneic or autologous and gene corrected, is thought to rely on the migration of monocytes/macrophages into the CNS and subsequent conversion to microglia that are long-lived and provide the therapeutic benefit. Integration site analysis in these 2 participants of the trial demonstrated the typical lentiviral pattern with no evidence of protooncogene activation.⁶⁴ 

MLD is an inherited autosomal recessive disorder secondary to a deficiency of the lysosomal enzyme arylsulfatase A.⁶¹  Massive accumulation of nonmetabolite sulfatides damages both the central and peripheral nervous systems. A mouse knockout model of MLD has been developed and used for the exploration of gene therapy approaches.⁶⁵  Stem cell–targeted gene transfer followed by autologous transplantation is one approach.⁶⁶  An alternative that is also being explored in the mouse model is the direct injection of AAV vectors encoding ARSA into the CNS.⁶⁷  Early correction is essential because the most common form of MLD develops in the second year of life with rapid, progressive CNS dysfunction. Thus, the introduction directly of AAV vectors into the brain seems more preferable than stem cell–targeted gene transfer in that correction of the phenotype with the latter approach requires several months, during which the patients continue to deteriorate.

The basic strategy with this approach is to genetically modify donor T cells ex vivo before infusion for enhancement of engraftment and prevention of viral infections. GVHD may develop in this context so that methodologies have been evolved to introduce a suicide gene into the T lymphocytes before infusion into patients. By the time the T cells are ablated, engraftment has been established and immunity is generally sufficient to protect from infections. A number of genetic systems have been used or explored experimentally.⁶⁸  Two have been tested in clinical trials, 1 with a vector having the thymidine kinase gene, and posttreatment is with ganciclovir.⁶⁸  The other system involves caspase-9.⁶⁹  T cells containing this gene can be abrogated using a dimerizing drug (AP1903). Overall, this approach shows promise for control of GVHD.

Adoptive transfer of T-lymphocyte populations has been used as a strategy for preventing and treating viral infections in immunocompromised individuals.⁷⁰  Adenoviral-transduced, antigen-presenting cell lines were used to derive the T cell–specific populations to treat or prevent viral infections in stem cell transplant recipients.^70,71  More recent studies rely on the generation of cytotoxic T lymphocytes using immunostimulatory cells that have been transduced with peptide mixtures of the viral antigens.^72,73  Current efforts to use gene transfer for prevention or treatment of viral infections are focused on AIDS.^74-76  The goal in these efforts is to introduce 1 or more genetic elements into autologous hematopoietic stem cells that are then used to reconstitute the hematopoietic system of the AIDS patient to create a population of HIV-resistant T cells. Several studies have focused on the knockdown or knockout of the CCR5 gene as it encodes a potent coreceptor for HIV infection. Inhibition of critical processes such as viral entry or replication reflects an alternative strategy of stem cell modification for HIV therapy.⁷⁴ 

A number of creative approaches have been developed to use gene transfer in the context of treating cancer.^1,77  These include generating T-cell populations in vitro with specificity for antigens expressed on tumor cells for reinfusion, the development of tumor vaccines by expressing genes that enhance the immune response to injected tumor cells, and the creation of CARs on T cells in which a single-chain antibody with specificity for an antigen expressed on human tumor cells linked to internal domains that participate in cell activation.^77,78  Remission has been achieved in patients with B-lymphoid malignancies who received CAR-modified T cells (Table 2).^79-81  Fortunately, no evidence of T-cell transformation has been observed in these patients, although the period of observation is still rather short. Oncolytic adenoviruses with relative tumor cell specificity are being developed for a variety of cancers.⁸²  Species D adenovirus oncolytics have been shown to infect and kill B-cancer cells, but development of genetic variants to enhance activity has not yet been reported, nor have clinical trials begun to date.⁸³  Another strategy to use gene transfer to enhance cancer treatment involves the transfer of a drug-resistance gene into autologous hematopoietic stem cells that are given to patients in the context of myelosuppression. Treatment is then given to amplify the genetically modified cell populations.⁸⁴  This strategy has been tested in a small clinical trial that demonstrated that intensification of chemotherapy was feasible, and there was an apparent improvement in outcome.⁸⁵  Overall, gene transfer technology continues to offer many opportunities to improve the treatment of individuals with malignant hematopoietic disorders.

Several different approaches have been considered for using viral vectors to achieve therapy for hemophilia.^1,86  Over the years, Amit Nathwani at University College London and Andrew Davidoff at St. Jude Children’s Research Hospital have shown that intravenous injection of rAAV8 particles resulted in long-term hepatic production of human factor IX in the rhesus macaque model. These observations have now been translated into a clinical trial (Table 2).⁸  To date, 10 participants have been enrolled, and all show evidence of vector-derived FIX production. The length of follow-up ranged from 6.7 months to 3.3 years. The FIX levels varied from 1% to as much as 8% or 9% in the individual patients at different times, with participants that received the highest dose having the higher levels. Seven have stopped routine prophylaxis with recombinant FIX, and the remaining 3 individuals have reduced the frequency of recombinant protein infusion. The only complication in the trial was the development of transient elevation in transaminases in 1 and perhaps in 2 other participants, which was abrogated with a short course of prednisolone.⁸  Overall, these results are highly encouraging with respect to application of gene therapy for FIX deficiency. Efforts are underway to modify the AAV capsid to reduce its immunogenicity.⁸⁷  Despite the larger size of the FVIII gene, an AAV vector has been constructed with a modified coding sequence capable of being packaged and transmitting an expressible FXIII gene.⁸⁸  A clinical trial for hemophilia A is in the early stages of planning.

The hemoglobin disorders, severe β-thalassemia and sickle cell anemia, were identified early as potential targets for therapeutic gene transfer.¹  However, despite considerable effort over the years to develop vector systems and to test them in animal models, progress into the clinic has been limited and not fully successful.^1,89  We have estimated that it would require 20% of the primitive hematopoietic cells to be genetically modified to achieve a definitive therapeutic benefit.⁸⁹  Although levels of gene transfer of 7% to 14% were observed in 2 participants in a clinical trial for X-linked ALD,⁶²  comparable levels have not been achieved in individuals with thalassemia. In 1 small trial, 2 patients failed to engraft with the transduced stem cells and had to receive backup bone marrow for hematopoietic rescue (Phillip Leboulch, personal communication). The other participant developed significant amounts of the transgene globin product, but this apparently primarily reflected dominance by a large clone having a retroviral integration into the HMG2A gene.³⁵  This participant remains transfusion independent for a period now up to 7 years.

There has been a long-standing interest in reactivating the fetal γ-globin gene in an effort to compensate for deficient β-globin synthesis in thalassemia or to inhibit sickling in patients with sickle cell anemia.^90,91  Efforts in this area were advanced with the discovery that BCL11A is a major regulator of developmentally stage-specific repression of the γ-globin gene.⁹²  Genetic deletion of BCL11A has been used to correct sickle cell disease in adult mice by interference with fetal hemoglobin silencing.⁹³  In addition, therapeutic levels of fetal hemoglobin have been achieved in erythroid progeny of β-thalassemic cells after lentiviral vector–mediated gene transfer of an inhibitory BCL11A short hairpin RNA.⁹⁴  Efforts to advance these preclinical observations to small-scale clinical trials are of great interest.

Fanconi anemia is an inherited bone marrow failure disorder characterized by aplastic anemia and an enhanced risk for the development of leukemia. The syndrome may occur as a consequence of a defect in 1 of at least 15 genes. The development of gene therapy approaches for this disorder have focused on the protein encoded by the FANCA gene,^95,96  which is most commonly mutated. An international working group has been established to chart the path and facilitate the development of gene therapy for Fanconi anemia.^95,96  Gene therapy for patients with Fanconi anemia is particularly challenging because of low numbers of hematopoietic stem cells and sensitivity to myelosuppressive regimens.

Over the past decade, the use of chimeric nucleases to create specific double-stranded breaks to alter specific sites in the genome has progressed significantly.⁹⁷  Several classes of enzymes have been developed that enable a broad range of genetic modification by inducing DNA double-strand breaks that stimulate error-prone nonhomologous end joining or homology-directed repair at specific genomic locations.⁹⁸  Various viral vectors including AAV⁹⁹  lentiviral¹⁰⁰  and adenoviral vectors¹⁰¹  have been adapted for various purposes to achieve gene editing in target cells. The methodology has been used to establish HIV-1 resistance in T cells by disrupting the HIV coreceptor CC chemokine receptor 5.¹⁰²  The mutation that results in synthesis of a sickle β-globin has been corrected in human induced pluripotent stem (iPS) cells providing proof of principal that this is a potentially viable approach.^103-105  The nuclease methodology has also been used to achieve correction of α-thalassemia in iPS cells¹⁰⁶  as well as genetic correction of β-thalassemia mutations in such cells.¹⁰⁷  Another potential application is the creation of genomic safe harbors that facilitate high β-globin transgene expression. Genomic safe harbors are identified by screening of iPS cells from thalassemic individuals to identify clones with favorable integration sites.¹⁰⁸  Basic research continues to increase the repertoire of target sites amenable to genetic modification.¹⁰⁹  This technology could also be applicable to other blood disorders such as immunodeficiencies or CGD.

The author thanks Derek A. Persons and Brian P. Sorrentino for their review of the manuscript and for providing useful suggestions for improvement, and Pat Streich for her assistance in the preparation of this manuscript.

This work was supported by grants from the National Institutes of Health National Heart, Lung, and Blood Institute (P01HL 53749 and R01DK095169), the Assisi Foundation (94-000 R12), and the American Lebanese Syrian Associated Charities.

Contribution: A.W.N. wrote and approved the manuscript.

Conflict-of-interest disclosure: The author declares no competing financial interests.

Correspondence: Arthur W. Nienhuis, St. Jude Children’s Research Hospital, 262 Danny Thomas Pl – MS#341, Memphis, TN 38105; e-mail: arthur.nienhuis@stjude.org.