Chapter 29: Gene Therapy for Hematologic Diseases

Gene therapy is a promising treatment for several inherited or acquired hematologic disorders. Gene therapy involves the introduction of a functional gene to replace a mutated gene or a therapeutic gene to provide a missing or defective protein to the organism. In some cases, the patient’s blood cells are removed and special, targeted cells such as hematopoietic stem cells (HSCs) are selected for engineering. The therapeutic genes are introduced into a vector and delivered into the targeted cells. These targeted, gene-modified cells are reinfused back into the patient. Because this method modifies cells outside the patient’s body, it is called ex vivo gene therapy (Fig. 29–1A). By contrast in vivo gene therapy describes the therapeutic gene-containing vectors being directly injected into the patient (Fig. 29–1B). In the in vivo case, the gene is expressed, producing a therapeutic protein for treatment. Theoretically, if the gene-modified cells are long-lived and able to expand inside the body, a single gene therapy can be sufficient to provide a lifelong therapeutic effect. Current gene therapy technologies have reached the point that many types of single-gene hematologic deficiency diseases can be permanently corrected, for example, X-linked severe combined immunodeficiency (X-SCID) and adenosine deaminase deficiency severe combined immunodeficiency (ADA-SCID).

Gene therapy has improved significantly in effectiveness since the mid-1980s when the first experiment using stem cell gene transfer was successful. Scientists initially encountered several serious obstacles. First, there was difficulty in delivering a modified gene into HSCs because of their lack of cell-surface receptors and their quiescent state.¹ Second, early X-SCID gene therapy was interrupted because 20 percent of patients developed a T-cell type of leukemia within 3 to 6 years after therapy. This event was caused by gene therapy-related viral vector insertional mutagenesis; the vector contained powerful enhancer elements within its long terminal repeats (LTRs) that inserted close to, and activated the LMO2 protooncogene.² Over 3 decades, these problems have been (for the most part) resolved. With an array of cytokine stimulation cocktails to improve HSC receptivity to engineering, and improved lentiviral vectors, the HSC transduction rate in humans can reach 80 to 100 percent.^3,4 In addition, myeloablative conditioning regimens (e.g., busulfan, melphalan and 1,3-bis-[2-chloroethyl]-1-nitrosourea [BCNU; aka carmustine]) that decrease the number of endogenous stem cells prior to infusion of the engineered HSCs has proven to be an effective method to increase engraftment rates (average: 1 to 2 copies of gene marking/cell).^3,4 High-level stem cell engraftment is a critical factor for the gene therapy of chronic granulomatous disease, X-linked adrenoleukodystrophy (X-ALD) and metachromatic leukodystrophy.^3,4,5,6 Moreover, newer, safer self-inactivating (SIN) viral vectors have been developed in which viral LTR enhancers are completely removed. Using these newer vectors, no patient has developed therapy-related malignancy in several clinical trials, some of which have followed patients for as long as 8 years.^3,4 Gene therapy is no longer a hypothetical form of therapy; some trials have achieved clinical correction for as long as 12 years. Gene therapy has, thus, undergone a renaissance. This chapter discusses two critical technical factors in gene therapy: targeted cells and delivered vectors; briefly reviews the gene therapy of several common hematopoietic genetic deficient diseases; and describes new strategies of in vivo selection and insertion site-targeted gene therapy.

HEMATOPOIETIC STEM CELLS

HSCs are the cell of choice for many gene therapy applications for several reasons. First, many blood cell disorders originate at the stem cell level and, therefore, gene-corrected HSCs are the best candidate for replacement. Second, HSCs are a long-lived and self-renewing population and may reduce or eliminate the need for repeated administration of gene therapy. HSCs are readily obtained in the blood, marrow, or umbilical cord blood. They are also easily selected and manipulated in the laboratory and can be returned to patients relatively easily. Third, engineered HSCs are able to correct defects in all hematopoietic lineages. Fourth, HSCs migrate to several tissues in the body—primarily the marrow, but also the liver, spleen, and lymph nodes. These may be strategic locations for localized delivery of therapeutic agents for disorders unrelated to the hematopoietic system, such as for patients with liver diseases.

T LYMPHOCYTES

The use of hematopoietic progenitor or precursor cells for gene therapy theoretically would require repeat application. However, for some inherited and acquired diseases involving T cells (e.g., severe combined immunodeficiency [SCID] and AIDS), the use of T-cell precursors is capable of long-term correction of the T-cell deficiency. The underlying mechanism is still not clear, but some types of T cells, such as memory T cells, are long-lived. The memory T-cell compartment has three subsets: T-memory stem cells, central memory T cells, and effector memory T cells.⁷ Memory T cells and central memory T cells can expand to large numbers of effector T cells when activated. T-cell gene therapy is effective in some forms of cancer therapy. Long-term effects could be related to memory T stem cells. For example, T effector cells can directly eradicate malignant cells. However, cancer cells are often “invisible” to T cells. For this reason, modified T cells have been created by adding a chimeric antigen receptors (CARs) expressing gene, engineering the T cells to recognize cancer-specific antigens such as CD19 on chronic lymphocytic leukemia (CLL) cells.^8,9 Once the CAR-modified T cells are transfused into the patient, they can attack and eradicate the targeted malignant cells. In clinical trials, anti-CD19 CAR T cells profoundly decreased the level of cancer cells in three patients with end-stage CLL.^8,9

An advantage of using gene-modified T cells is that the procedure does not disturb the otherwise normal hematopoietic system. However, this is also a limitation of using T cells in gene therapy. For example, when using T cells to correct X-SCID, although T cells recover to near normal levels, the other affected cell types (e.g., B cells, natural killer [NK] cells) are not corrected. In this approach, the patient would require immunoglobulin replacement and the NK cell deficiency would persist. A study compared the outcomes of several clinical trials of HIV therapy comparing CD4+ T-cell treatment to HSC-based treatment. These trials found that overall, the results from HSCs were better than that of CD4+ T cells.¹⁰ One reason offered for this result is that without stimulation, T cells proliferate at a very low rate, estimated at approximately one division every 3.5 years for naïve T cells, and one division every 22 weeks for memory T cells.¹¹

The majority of gene therapy studies thus far have employed viral vectors, because of their high efficiency of transgene delivery into the human nucleus. For a long-term effect, genome-integrating retroviral/lentiviral vectors have been employed. However, DNA integration causing critical gene mutagenesis has raised concern about long-term safety. This concern has led to the development of persisting but nonintegrating viral vectors.

NON–GENOME-INTEGRATING VIRAL VECTORS

Adenoviral Vector

Adenoviral vectors do not integrate into the genome, but are highly effective in gene therapy because of their ability to efficiently transduce both dividing and nondividing cells, and to persist relatively well in long-lived targeted cells. Adenoviral vectors also have capacity to hold large segments of DNA (e.g., 7.5 kbp); they are easily manipulated with recombinant DNA techniques and have the ability to produce high titers.^12,13 However, adenoviral vector infection enlists a variety of humoral and cellular immune responses.¹⁴ Therefore, adenoviral vector therapy may result in acute toxicity and autoimmunity, and which clears transgene-expressing cells, reducing the efficacy of therapy.¹³ Of interest, this autoimmunity can be targeting to adenovirus-infected cancer cells, and thus used as an oncolytic agent.¹⁵

ADENO-ASSOCIATED VIRAL VECTORS

The adeno-associated viral vector (AAV) has the abilities to infect both nondividing and dividing cells and to persist without vector integration.^16,17 Upon entering host cells, wild-type AAV DNA becomes episomal or integrates into the genome; in contrast, current, modified AAV vectors are designed to lose their integrating ability.¹⁸ AAV can carry a transgene up to 4.7 kbp. The AAV genome is single-stranded DNA and vector preparations are composed of a mixture of vector particles having one of the two strands of the virus. Upon transduction, the required annealing of the two strands delays gene expression. This limitation can be overcome by using a self-complementary design in which the two strands of the transgene are on a single hairpin genome in an inverted orientation, which allows quick assembly into a transcription unit following transduction.¹⁹

GENOME-INTEGRATING VIRAL VECTORS

Retroviral Vectors

A feature of retroviral vectors is that they have the ability to stably incorporate their viral DNA into the host genome, which can result in long-term expression of the transgene. Most retroviral vectors are γ-retroviral or lentiviral vectors. Lentiviral vectors are usually HIV-1 based, or may also contain elements of simian immunodeficiency virus, and have at least three advantages over γ-retroviral vectors; first, the lentiviral vector preintegration complex is able to cross the nuclear membrane in host cells even in the absence of mitosis and, therefore, is able to transduce nondividing cells, such as HSCs.²⁰ Second, the lentiviral vector preintegration complex is more stable and persists longer, which improves the likelihood of integration.²⁰ Third, γ-retroviral vectors prefer to integrate near gene transcription start sites, such as the CpG islands and conserved noncoding sequences and conserved transcriptional factor binding sites, whereas lentiviral vectors are more evenly distributed, reducing the likelihood of driving the expression of a deleterious gene(s).²² Lentiviral vectors have thus emerged as an important advance for gene therapy in the hematologic disorders.

X-LINKED SEVERE COMBINED IMMUNODEFICIENCY

X-SCID is a single-gene-deficient disease caused by mutations in the gene of the interleukin-2 receptor, γIL2RG (Chap. 80).²³ Lack of functional γIL2RG results in a lack of T and NK cells and poorly functional B cells. X-SCID is a fatal disease, and without medical interventions, patients often die within the first 2 years of life.²³ During 1999 to 2006, 20 X-SCID children were treated in two gene therapy trials.^24,25 Because patients lacked an human leukocyte antigen (HLA)-identical donor, precluding stem cell transplantation as a curative therapy, all 20 patients were given a single γ-retroviral vector-mediated gene therapy in which a wild-type γIL2RG gene was delivered into patients’ T cells. From 5 to 12 years of observation after the gene transfer procedure, 17 of the 20 treated participants are alive and display nearly full correction of the T-cell deficiency by genetically modified T cells.^24,25

ADENOSINE DEAMINASE DEFICIENCY SEVERE COMBINED IMMUNODEFICIENCY

Adenosine deaminase (ADA) deficiency is an autosomal-recessive inherited disorder caused by mutation of ADA gene on chromosome 20 (Chap. 80). ADA deficiency leads to inhibition of DNA synthesis, which is particularly toxic to lymphocytes because they are some of the most mitotically active cells. This condition ultimately causes SCID. ADA-SCID is almost always fatal by 2 years of age. Gene therapy with a normal ADA gene expressed in autologous HSC is a potentially curative treatment. However, early attempts at gene therapy did not shown any clinical benefit. In the late 1990s, when improved retroviral vector and preinfusion chemotherapy conditioning were instituted, success occurred. Since 2000, 40 patients have been treated in Italy, Great Britain, and the United States.^26,27,28 In these studies HSCs were transduced with a γ-retroviral vector encoding the ADA enzyme. Low-intensity conditioning with either busulfan or melphalan was used to increase engraftment of stem cells. In some patients the corrected ADA gene can be detected in blood mononuclear cells for more than 9 years, and the ADA enzyme remains at a normal level.²⁷

WISKOTT-ALDRICH SYNDROME

Wiskott-Aldrich syndrome (WAS) is the result of a mutation in the gene encoding the WAS protein (WASp).²⁸ The gene is located on the short arm of the X chromosome and WAS is an X-linked recessive genetic disorder. WASP activates actin polymerization in almost all blood cells. Hereditary deficiency in WASP is associated with microthrombocytopenia, recurrent infections, eczema, high incidence of autoimmunity and hematopoietic malignancy (lymphoma or leukemia).²⁹ In 2010, three patients in Italy were administered a lentiviral vector engineered to express WASp following busulfan conditioning. All three WAS patients showed excellent multilineage engraftment with an average of 0.4 to 0.9 correct gene copy per genome persisting to 30 months. Symptoms of WAS improved substantially. Pretreatment eczema resolved between 6 and 12 months after therapy. The frequency and severity of infections progressively decreased, and cytomegalovirus replication was controlled, allowing withdrawal of antiinfectious prophylaxis in two patients. Improvement of platelet count resulted in discontinuation of platelet transfusions.³

LEUKODYSTROPHIES, X-LINKED ADRENOLEUKODYSTROPHY, AND METACHROMATIC LEUKODYSTROPHY

There are two major leukodystrophies, X-ALD and metachromatic leukodystrophy (MLD); both have been successfully treated with gene therapy.^4,6 X-ALD is a severe genetic demyelinating disease caused by mutations in the ABCD1 gene on the X chromosome that encodes the adrenoleukodystrophy (ALD) protein, an adenosine triphosphate binding cassette transporter. ALD deficiency leads to the accumulation of very-long-chain fatty acids and progressive demyelination in the central nervous system. In 2009, two French children were reported to have autologous HSC gene therapy with a lentiviral vector encoding wild-type ABCD1.⁶ The patients were given myeloablative regimen conditioning with cyclophosphamide and busulfan. ALD protein was expressed in 23 and 25 percent of blood mononuclear cells in the two patients. Over 3 years of followup, corrected ABCD1 was found in 7 to 14 percent of granulocytes, monocytes, and T and B lymphocytes. Cerebral demyelination was arrested.⁶

MLD is an autosomal recessive disorder caused by mutations in ARSA gene in chromosome 22 that encodes arylsulfatase A (ARSA). Deficiency in ARSA leads to sulfatide accumulation, eventually destroying the myelin sheath of the nervous system. In 2013, three children with the disease received autologous HSC gene therapy with a lentiviral vector encoding wild-type ARSA.⁴ After myeloablative conditioning with busulfan, engraftment was excellent with 45 to 80 percent gene marking levels. ARSA activity was restored to above normal values in the hematopoietic lineages and the cerebrospinal fluid.⁴ There was a clear therapeutic benefit. In X-ALD and MLD gene therapies, corrected cells did not have an obvious selective advantage. However, in both cases the gene marking levels were high (>10 percent) for as long as 2 years.^4,6 These results indicate that successful gene therapy can be achieved with very high viral transduction rates, even absent a transduced cell selection mechanism.

Avoiding Graft-versus-Host Disease

Gene therapy can lead to graft-versus-host disease (GVHD), especially after an allogeneic stem cell transplantation. To prevent GVHD, a suicide gene is introduced in the transferred cells; in the event of GVHD, a prodrug is given to activate the suicide gene to kill the transduced cell. This “safety mechanism” can also guard against uncontrolled proliferation of targeted cells. Traditionally, the suicide gene is herpes simplex virus thymidine kinase (HSV-TK), which can be made “suicidal” in the presence of ganciclovir (GCV). Once GCV is phosphorylated by HSV-TK, it turns to a nucleotide analogue that inhibits DNA synthesis and kills the cell.³⁰ Another new suicide gene system was developed by an inducible caspase-9 protein (iCasp9), which activates the mitochondrial apoptotic pathway. Its use is based on caspase-9 dimerization to an active form. The catalytic domain of caspase-9 is fused to a modified FK506-binding domain that can be homodimerized by a chemical inducer, such as AP1903 (Fig. 29–2). In a clinical trial, four patients who had undergone stem cell transplantation for relapsed acute leukemia were treated with the modified T cells. A single dose of AP1903 was given when GVHD developed, which eliminated more than 90 percent of the modified T cells within 30 minutes and eliminated the clinical signs of GVHD.³¹ In comparison to HSV-TK, iCasp9 offers several advantages. It induces cell death faster, does not rely on cell proliferation, which makes it ideal for cancer stem cell killing; and, because caspase-9 is downstream of the mitochondria, the sensitivity of cells to its activation should be independent of the antiapoptotic BCL-2 family of proteins, which is often upregulated in hematologic malignancies.

HUMAN IMMUNODEFICIENCY VIRAL INFECTION

The HIV infects helper T cells, such as CD4+ T cells, macrophages, and dendritic cells. The infection kills CD4+ T cells and cripples cell-mediated immunity (Chap. 81). Without treatment, average survival time after infection with HIV is estimated to be 9 to 11 years. In 2007, a gene therapy trial demonstrated that AIDS can be cured. A patient with an HIV infection and acute myeloid leukemia was given allogenic stem cell transplantation from a selected donor, whose two copies of the chemokine (C-C motif) receptor 5 gene (CCR5) were lost because of mutations.³² CCR5 is the major cellular coreceptor used by HIV to infect CD4+ T cells. The patient has since remained off anti-HIV drugs for 7 years. This case generated enormous interest in gene therapy approaches to cure HIV by blocking CCR5 expression. One of the most promising methods is to create loss of function mutations of CCR5 with zinc-finger nucleases (ZFNs). ZFNs are artificial restriction enzymes generated by fusing zinc fingers to a nonspecific double-strand DNA cleavage protein, a truncated Fox1.³³ A zinc finger can be engineered to a target 18 to 24 bp sequences in CCR5 DNA. ZFNs can repeatedly cut the DNA at a targeted site and eventually mutate CCR5 when repair errors occur. In a 12-patient clinical trial, the patient’s CD4+ T cells were infected ex vivo with a ZFN-expressing adenoviral vector to disrupt CCR5.³⁴ The modified cells were then reintroduced. CD4 T cells lacking CCR5 are resistant to HIV infection. Ultimately, the CCR5-mutated cells replace those vulnerable CCR5 wild-type cells. The infusion immediately increased the circulating CD4+ T cell count from a median of 488 × 10⁹/L to 1517 × 10⁹/L in the first week. The modified cells were also found in T-cell–rich gut-associated lymphoid tissues. Patients were off antiviral drugs for 84 days during which time the circulating CD4+ T cells dropped while viral load spiked. Antiviral treatment had to be reinstituted.³⁴ The CCR5-mutated CD4+ T-cell count remained stable even during the drug-off period, consistent with their resistance to HIV killing. However, CCR5-mutated CD4+ T cells apparently did not expand quickly, which may explain why in this trial the modified cell infusion alone was insufficient to control the HIV infection. Another trial used a lentiviral anti-HIV small interfering RNA (siRNA) expression vector to modify HSCs.³⁵ Although long-term (18 months) expression of siRNA in multiple blood cell lineages was observed, modified cell levels were less than 0.4 percent. The fact that both modified HSCs and T cells failed to repopulate indicated that the in vivo selection of modified HIV-resistant cells may be weak or the modified T cells do not have a proliferative reaction to HIV infection. In both cases, the engraftment rate was very poor (T cell <10 percent and HSC <0.2 percent), which could indicate that the modified cells do not have a sufficient starting number for repopulation. New trials are underway with an improved lentiviral vector, engraftment protocols and a methylguanine DNA methyltransferase (MGMT) in vivo selection mechanism.³⁶ The cure for HIV by gene therapy has shown promise, but some clinical obstacles have to be overcome to achieve success.

DISORDERS OF HEMOGLOBIN

Thalassemia and sickle cell disease represent the most common single-gene defect diseases worldwide (Chaps. 48 and 49). β-Thalassemia is caused by a mutation in the β-globin gene, resulting in reduced adult hemoglobin A (HgbA) and severe anemia.³⁷ Therefore, gene therapy is used to express a normal β-globin gene. Many efforts have been made to use stem cell gene therapy. However, success has been very limited.³⁷ Although a weak survival advantage for corrected red cells at an early mature stage was observed, the in vivo selection alone appears insufficient to achieve a sustained correction.³⁸ It is predicated that it would require 20 percent of the primitive hematopoietic cells to be genetically modified, and the gene expressed at near normal levels in those cells, to achieve a definitive therapeutic benefit. Even higher levels of corrected cells (approximately 100 percent) might be required to cure the disease.³⁸ The first successful clinical trial was reported in 2007.³⁹ An 18-year-old patient with severe β-thalassemia dependent on monthly red cell transfusion since age 3 years, received HSC lentiviral β-globin gene therapy. The viral transduction rate was approximately 30 percent. The patient continued receiving transfusions for 16 months after the transplantation, at that point the therapeutic HgbA was sufficient and maintained at a sufficient level until 33 months (Fig. 29–3). During the final 21 months, 100 percent of HgbA was from modified cells and the patient was transfusion-free. However, the therapy effect was later found to be from a dominant clone (>60 percent of all viral insertion sites in nucleated blood cells at 24 months), in which the viral insertion causes transcriptional activation of HMGA2 in erythroid cells.³⁹ Nevertheless, the patient remained with a good quality life with a stable Hgb of 9 to 10 g/dL, transfusion-free, and cancer-free for up to 7 years.⁴⁰ The dominant clone might indicate a strong in vivo selection, however, the level of gene modified cells in this patient has been never greater than 21 percent, and the highest level noted in blood was 10.9 percent and in erythroblasts was 3.3 percent, which is below that predicted. The dominant clone remained stable over time. However, HMGA2 overexpression was detected in erythroid cells, which could enhance in vivo selection and proliferation of the corrected cells. Nevertheless, this is but a single case; whether the in vivo selection of the dominate clone had a significant role is not clear. More trials are needed.

HEMOPHILIA

Hemophilia is an X-linked single-gene defect, of which 70 percent of affected patients display inheritance and 30 percent develop from de novo somatic mutations (Chap. 124). There are two major forms of hemophilia: hemophilia A, caused by loss-of-function mutations of the gene encoding clotting factor VIII (FVIII), and hemophilia B, the result of mutations in the gene encoding clotting factor IX (FIX). Hemophilia A accounts for 80 percent of patients and hemophilia B for 20 percent.⁴¹ The absence of either FVIII or FIX severely impairs the ability to generate thrombin and, subsequently, fibrin, leading to spontaneous bleeding when the factor levels fall below approximately 5 percent of normal. Theoretically, gene therapy using a lentiviral vector that permanently expresses a normal FVIII or FIX gene in the patient could cure either disease. However, after 2 decades of intense research, gene therapy has been very difficult. FVIII and FIX are produced in hepatocytes, not in derivatives of HSCs.⁴² Therefore, hemophilia is not a circumstance for HSC-based gene therapy. The emerging approach to gene therapy for hemophilia is by using in vivo gene therapy (see Fig. 29–1). In this approach, viral particles are injected into a patient’s vein, muscle, hepatic artery, or omentum.⁴³ Initially, five clinical trials with retroviral, adenoviral, or AAV vectors failed to achieve long-term expression of the coagulation factor and no measurable clinical benefit was observed.⁴³ However, a trial by a British-American team reported in 2011 showed exceptional results.¹⁵ This group focused on hemophilia B. The FIX gene, unlike the FVIII gene, is small and easy to insert into an AAV vector, and 1 to 2 percent of the normal blood levels of FIX is sufficient to markedly reduce the bleeding risk.⁴³ A new improved AAV vector (AAV8) was used. This vector has a self-complementary genome to improve transduction efficiency, and was designed to produce fivefold higher levels of capsid protein to reduce a potential cytotoxic T-cell response and increase liver tropism. The AAV vector is not genome-integrating and maintains itself as an intracellular episome. Its gene expression in growing cells is transient because episomes may be lost with each cell division. But in quiescent tissues the AAV vector is capable of mediating long-term gene expression as episomal chromatin.⁴⁴ A single intravenous infusion of the vector was given to six adult male hemophilia B patients who had been treated with recombinant FIX for many years. No notable acute or chronic toxicities were observed. All six patients displayed stable FIX expression at 2 to 11 percent of normal blood levels for 3 years. Four of these patients discontinued recombinant FIX treatment and remained free of spontaneous hemorrhage.^15,43,45 The same research team is attempting a similar approach for hemophilia A gene therapy; however, FVIII gene expression has been inefficient. One reason is gene size. The FVIII coding sequence is 7 kb, which far exceeds the normal packaging capacity of AAV vectors. By modifying the B domain, the FVIII size was reduced to a 5.2-kb AAV expression cassette, which is more efficiently packaged. Also, a hybrid liver-specific promoter was introduced into the vector. The resulting new AAV vector has shown high (15 percent of normal) FVIII expression for 20 to 45 weeks in four macaques.⁴⁶ This AAV vector will be used in a trial of hemophilia A gene therapy in the near future.

FANCONI ANEMIA

Fanconi anemia (FA) is caused by mutations in Fanconi genes, which encode the DNA repair proteins that form a function complex (Chap. 35). FA cells are hypersensitive to DNA crosslinking agents.⁴⁷ Sixteen Fanconi genes have been described. A defect in one of them will lead to FA. The disease is characterized by a high risk of developing marrow failure and later myelodysplasia, acute leukemia, or cancers of other tissues.⁴⁷ More than half of patients with FA are the result of FANCA gene mutations; therefore current gene therapy has focused on FANCA insufficiency. Gene therapy for FA is particularly challenging because of the low numbers of HSCs in the stage of marrow failure, and FA cells are extremely sensitive to DNA damage when exposed to myelosuppressive drugs used to condition the patient for stem cell transplantation. In a rare case, two identical twins had inherited FANCA mutations but with normal DNA repair in their blood stem cells. Functional FANCA in blood cells was found to be restored by a spontaneous intrauterine self-correcting somatic mutation in a single HSC. The fact that a single HSC was sufficient to restore a fully normal blood system indicates that FANCA gene therapy may require transduction of only a few HSCs.⁴⁸

In 2011, an international working group was established to facilitate the development of gene therapy for FA.⁴⁹ The initial protocol included delivery of a normal FANCA gene by a third-generation lentiviral vector into HSC and increasing HSC number by in vitro HSC expansion using a combination of HOXB4 and DELTA-1 proteins.⁵⁰

GENE THERAPY FOR CANCER

Gene therapy for cancer has been widely exploited. A review was published detailing the new developments in this field.⁵¹ This chapter has described a few significant new approaches.

One of the most creative new approaches to cancer-targeted gene therapy is the use of CAR for CLL, in which patient’s T cells were modified to target their own cancers (Chap. 92).^7,8 Another strategy is to enhance conventional chemotherapy by protecting the marrow cells through gene transfer. Chemotherapy has a limited therapeutic window because of its severe toxic effect on marrow cells, and its leukemogenic potential. Because the lethal effect of chemotherapy is mainly DNA damage, especially methylating O⁶-guanine, to overcome the side effects, a strong chemoresistant DNA repair gene, a mutant (P140K) of MGMT was introduced into brain tumor patients’ autologous HSC by a γ-retroviral vector, thus the patient’s transduced marrow progenitor cells could be protected by the modified MGMT, permitting them to tolerate more cycles of chemotherapy. A phase I clinical trial demonstrated that intensification of chemotherapy was feasible, and there was an improvement in therapy outcome and patient survival in these small studies.^52,53 A similar result was also observed in a recent clinical trial with a lentiviral vector (reported at the ASH meeting 2014).

IN VIVO SELECTION AND O⁶-METHYLGUANINE-DNA METHYLTRANSFERASE SELECTIVE METHOD

Whether stem cell gene therapy can cure a disease is dependent on the functionality of the gene-corrected cell. In the early successful gene therapy trials for X-SCID and ADA-SCID, a low (0.1 to 1.0 percent) engraftment rate was sufficient for long-term gene correction, and the single most important factor was that the transgene conferred a selective survival advantage on the transgene-bearing cell.³⁸ In contrast, a main reason for poor responses in other gene therapy trials has been that the gene-corrected cells have no, or only a weak selective advantage, such that the gene-transduced cell levels are insufficient to provide a clinical meaningful benefit. Therefore, in vivo selection is a key element in the success of clinical gene therapy.

The gene-corrected cells in the majority of hematologic genetic diseases have no in vivo selection advantage.³⁸ To overcome this problem, a second, selectable gene can be used to be coexpressed with the gene that corrected the defect. The second gene turns the cells into selectable cells.³⁸ The selectable gene can be cloned into a vector along with correcting gene driven by a single or separate promoter. Experiments have identified several selectable genes including multidrug resistance protein, dihydrofolate reductase, and MGMT.³⁸ MGMT shows the most promising results in large animal and humans trials. MGMT encodes a DNA repair enzyme O⁶-alkyguanine-DNA-alky-transferase, which confers resistance to the cytotoxicity of chemotherapy, such as BCNU and temozolomide (TMZ).⁵⁴ A MGMT mutant, P140K-MGMT, has at least a 50-fold stronger effect on drug resistance resistance.⁵⁵ P140K-MGMT–expressed cells can be exposed to BCNU and TMZ selection pressure (Fig. 29-4). Clinical trials have demonstrated that P140K-MGMT protects the gene-modified cells from TMZ-induced toxicity.⁵² This strategy has also been used in a mouse model of HIV gene therapy.³⁶

Successful gene therapy with satisfactory gene marking levels (approximately 30 percent) are dependent on three factors: a high HSC transduction rate, a high engraftment rate, and in vivo selection.³⁸ Other trials indicate that if there are excellent transduction rates (80 to 90 percent) and excellent engraftment rates (gene copy marking/genome >0.5), these two factors can be sufficient to achieve a sustained gene correction, even without an in vivo selection mechanism.⁴ Nevertheless, if there is a low transduction rate and low engraftment rate, MGMT-mediated in vivo selection will be a very valuable tool to improve the likelihood of successful gene therapy.

OVERCOMING GENOTOXICITY BY TARGETED-INSERTION GENE THERAPY

To achieve sustained gene correction, some gene therapy approaches have used integrated vectors such as γ-retroviral or lentiviral vectors. Gene therapy for X-SCID,⁵⁶ WAS,⁵⁷ and chronic granulomatous disease (CGD)⁵⁸ has found that γ-retroviral insertion in the vicinity of protooncogenes is associated with the development of lymphoproliferative and myeloproliferative neoplasms. Improved lentiviral vectors have added safety features, such as no preference for integration near promoters, removal of viral promoter-enhancers, and self-inactivation. However, even with these new features, lentiviral vector-induced clonal dominance in human⁶ and murine leukemia have been reported.⁵⁹ DNA insertion is the most important factor to determine whether a therapy-related cancer will occur, especially as 80 to 90 percent of lentiviral vector insertions are within gene regions.²¹ Gene therapy with viral vector insertion, in which the insertion site is uncontrollable, raises the risk of a secondary clonal disease. The FDA has not approved a single-gene therapy in the United States, largely because of the risk of uncontrollable insertional mutagenesis.

One way to reduce this risk is to select preferred sites of DNA insertion. Gene-targeting and gene-editing technologies could make this possible. Gene editing describes insertion of DNA at a desired location.³³ It starts with artificially engineered nucleases, such as ZFNs, which can create a double-stranded break (DSB) at a targeted DNA sequence anywhere in the human genome. The DSBs is repaired by homologous recombination (HR) repair or nonhomologous end-joining repair. The ZFNs can be cotransfected with another plasmid (donor DNA plasmid), in which a desired DNA sequence, such as a transgene has been inserted within a sequence that is homologous to the flanking sequences of the DSB. When this DNA sequence is used as a template by HR, it would result in insertion at a targeted location (Fig. 29–5). Using ZFN, a DNA (up to 9.6 kb)⁶⁰ can be inserted precisely into a selected location of the human genome. There are several known “safe harbors” in the human genome. For example, the PPP1R12C gene on chromosome 19 (AAVS1 locus) is frequently integrated by an AAV and such an AAV-based integration is not associated with any subsequent pathologic events.⁶¹ Several genes have been inserted into the AAVS1 site in human cells, including stem cells, using ZFN gene-targeting methods.^60,62 The gene correction rate of gene editing is still very low (approximately 1 percent).⁶³ Fortunately, gene targeting is a fast-growing field. Many new technologies have developed in recent years. The transcription activator-like effector nuclease (TALEN)⁶⁴ and the clustered, regularly interspaced, short palindromic repeats (CRISPR)⁶⁵ technology show promise. Both TALEN and CRISPR are much easier and faster to use than the ZFN approach. These new gene-targeted technologies together may solve the retroviral insertion mutagenesis safety problems.