Overview of gene therapy, gene editing, and gene silencing

INTRODUCTION — A number of methods are under development for treating genetic disorders, including inherited (eg, monogenic/Mendelian) conditions, and acquired conditions such as cancer and infections. This topic reviews molecular techniques that can be used to alter the sequence or expression of a gene, including gene therapy, gene editing, and gene silencing.

Separate topics discuss related concepts and the clinical or investigational uses of these approaches in specific medical conditions:

●Glossary – (See "Genetics: Glossary of terms".)

●Genetic diseases – (See "Basic genetics concepts: DNA regulation and gene expression".)

●Genetic testing – (See "Genetic testing".)

●Genetic counseling – (See "Genetic counseling: Family history interpretation and risk assessment".)

OVERVIEW

Definitions — The following definitions apply to the methods discussed herein. A more complete glossary of related terms is presented separately. (See "Genetics: Glossary of terms".)

●Gene therapy – Gene therapy refers to the introduction of an exogenous gene or genes into one or more autologous or allogeneic cell types. The new gene may be referred to as a transgene. Gene therapy may be used for a number of reasons (eg, to supply a missing gene, to bypass the role of a missing gene, or to augment therapy for a disease). Gene therapy is in limited clinical use or under investigation in several types of disorders. (See 'Gene therapy' below.)

●Gene editing – Gene editing refers to the creation of sequence-specific alterations in the DNA of a cell using molecular methods that take advantage of site-directed DNA repair after strand breakage. This may be done to correct a pathogenic gene variant or to alter other biologic processes. The target cell may be a multipotent stem cell such as a hematopoietic stem cell, a differentiated somatic cell, or an embryonic stem cell. Gene editing is an investigational method for health care in humans. (See 'Gene editing' below.)

●Gene silencing – Gene silencing refers to the reduction of the expression of a gene (or genes). This may be performed using a number of methods such as short hairpin RNAs (shRNAs) or antisense oligonucleotides (ASOs) that cause RNA interference (RNAi), or by inducing epigenetic alterations that silence transcription of groups of genes. There are a number of potential applications; very few are in clinical use. (See 'Gene silencing' below.)

Epigenetic therapies are discussed separately. (See "Principles of epigenetics", section on 'Therapeutic uses'.)

In vitro versus in vivo transduction — Transduction refers to the process of introducing genetic material into a living cell. Cells can be transduced with a construct (gene therapy, gene editing, or gene silencing) either in vitro (ie, outside of the body) or in vivo (eg, injected into a tumor, inhaled, or administered intravenously). Following in vitro transduction, the modified cells can be administered to the patient. Following in vivo transduction, the targeting construct may home to target cells.

Hematopoietic stem cells can be transduced in vitro or in vivo; studies investigating intraosseous injection or homing to stem cell factor receptor (CD117) on the stem cells are ongoing [1,2].

In vitro manipulations allow greater opportunity for selection and testing of the transduced cells, whereas in vivo administration allows the cells to remain in their physiologic niche, where they may be less likely to be exposed to environmental stresses such as hyperoxia or loss of normal adhesion, and, as a result, they may be less susceptible to genotoxic stress.

In either case, it may be possible to administer repeated doses of the therapy. This is easiest for treatment approaches that do not require cytotoxic therapy to eradicate the bone marrow.

Types of vectors — Constructs for gene therapy or gene editing are delivered to the cells of interest using a vector. The choice of vector can have a major impact on efficacy and safety. A number of vector backbones have been tested since the inception of gene therapy studies. Most vectors are derived from viruses, although other types of vectors such as plasmids can be used.

Characteristics of the vector backbone — Vector backbones differ in a number of characteristics:

●Capacity – Different vectors can accommodate different-sized constructs. Some large genes or groups of genes require a very large-capacity vector.

●Production – Vectors used clinically must be produced in sufficient quantity for administration. Depending on the clinical application and number of patients affected, production demands may influence vector selection.

●Target cells – Some viruses have a tissue-specific tropism. This can be manipulated in some systems. As an example, bispecific antibodies (antibody-based molecules able to recognize two different antigens) that can retarget a virus to another cell type are under investigation in preclinical models [3,4]. Some viruses are only able to infect dividing cells, while others can also infect quiescent cells.

●Integrating versus nonintegrating – Integrating vectors are inserted into the genomic DNA of the host cells and are replicated at every cell division, whereas nonintegrating (episomal) vectors may be lost gradually with progressive cell divisions. Integrating vectors carry risks associated with their insertion into the genome, such as creating new fusion genes, damaging existing genes, and/or leading to increased expression of an adjacent gene from a strong vector promoter/enhancer region. (See 'Integrating versus nonintegrating vectors' below.)

●Expression level – Some genes require a higher expression level, whereas others may be sufficient at lower levels. For diseases with autosomal recessive inheritance, restoration of a small amount of gene function may be sufficient for effective therapy.

●Duration of expression – Some vectors provide long-term expression, which may be necessary for gene deficiencies. Others may have transient expression, which may be ideal for acquired disorders (eg, cancer, infection).

●Immunogenicity – The degree of immunogenicity of the virus used to create the gene therapy vector affects the host immune response, which may in turn lead to side effects as well as loss of potency. As an example, adenovirus infections are common in the general population, and virus-specific memory immune responses may limit the use of certain adenoviral vectors. There are a number of potential approaches to avoiding, reducing, or overriding immunogenicity, as discussed below. (See 'Strategies to reduce immunogenicity' below.)

Integrating versus nonintegrating vectors — Early gene therapy vectors were based on integrating viruses such as gamma retroviruses. These were mostly supplanted by lentiviral vectors in which long-terminal repeat (LTR) enhancers were eliminated, resulting in a lower likelihood that adjacent genes will be activated, a better ability to transduce nondividing cells, and a shorter culture period [5]. (See "Genetics: Glossary of terms", section on 'Enhancer'.)

However, as discussed below, initial reports with certain integrating vector constructs described cases of oncogenesis, including fatal leukemogenesis, and a substantial body of subsequent work has focused on nonintegrating vectors, as discussed below. (See 'Potential concerns with gene therapy' below.)

Nonintegrating vectors include those derived from nonintegrating viruses (eg, adenovirus, adeno-associated virus [AAV], poxviruses such as vaccinia), integration-deficient lentiviruses, and nonviral vectors (eg, plasmids [small, circular, double-stranded DNA molecules capable of replication], artificial chromosomes, nanoparticles) [6,7].

Adenoviral and AAV vectors can infect a variety of cell types including nondividing cells. For instance, different AAV serotypes can be used to infect specific tissues; in particular, AAVs have good efficiency for transducing liver, muscle, and nervous tissue. These viruses potentially avoid oncogenic effects related to insertional mutagenesis or increased expression of unrelated genes that are located near the insertion site. However, there are concerns about very low-frequency integration of AAV into cellular DNA and the potential risk of genotoxicity [8].

Most adenoviral vectors are replication deficient (due to deletion of early region 1A [E1A, a transactivating region required for adenoviral transcription and DNA replication]) and can accommodate large inserts (up to 37 kb in high-capacity "gutless" vectors); they can express larger amounts of recombinant proteins than other vectors [9]. By contrast, the much smaller wildtype AAV is naturally replication-defective and can replicate only in the presence of a helper virus (adenovirus); AAV vectors can only package DNA sequences up to 5 kb.

Because adenoviruses and adeno-associated viruses do not normally integrate into the host genome, vector DNA will remain episomal and will be eliminated when the cell divides or dies. Furthermore, because of their tropism to liver following systemic administration, retargeting is essential for their use in other tissues.

Another disadvantage of adenoviral vectors is their immunogenicity. Potential consequences are twofold:

●Inability to receive gene therapy due to prior immunity – Many individuals have developed immunity to adenoviruses from common respiratory infections (see "Pathogenesis, epidemiology, and clinical manifestations of adenovirus infection"). These infections or prior exposures from a previous dose of gene therapy can elicit serotype-specific neutralizing antibodies, and cells that have been transduced with adenoviral constructs can be killed by CD8⁺ cytotoxic T lymphocytes that can recognize different serotypes [10,11]. It may be possible to design vectors that bypass or override immunogenicity; this is an active area of research. (See 'Strategies to reduce immunogenicity' below.)

●Immune/inflammatory complications of treatment – In certain cases, immune/inflammatory reactions to an adenoviral gene therapy construct can be severe. These reactions may preferentially affect the liver as hepatic macrophages take up viral particles and secrete inflammatory cytokines. Antiviral antibodies may also contribute. Certain adenoviral serotypes have produced acute inflammatory reactions, one of which was fatal [12-14].

A smaller percentage of viral vectors under study are integrating vectors derived from retroviral backbones [15]. These include lentiviral vectors such as those derived from human immunodeficiency virus (HIV; with pathogenic genes removed), foamy virus (FV), murine leukemia virus (MLV), and avian sarcoma and leukosis virus (ASLV). Lentiviruses are appealing because they can infect nondividing cells. However, clinical trials using integrating vectors have been associated with serious adverse events including leukemogenesis, highlighting the significance of the safety concerns related to the site of insertion into the host cell genome [16-18].

To reduce the potential for insertional mutagenesis and related adverse effects from retroviral vectors, the genes responsible for replication competence are provided in a separate "helper" plasmid; this prevents the gene therapy vector from replicating independently and inserting into other sites in the genome. Other modifications include a self-inactivating (SIN) design in which promoter/enhancer regions, which could cause transcription from genes near the integration site, are removed.

Another type of integrating vector takes advantage of transposons (mobile DNA elements that can translocate between different chromosomes and plasmids), which use a transposase enzyme to cut and paste the gene of interest into the host genome [19]. Similar safety measures to those used with retroviruses are included (eg, providing the transposase enzyme in a separate plasmid).

Strategies to reduce immunogenicity — Strategies to reduce or override immunogenicity of adenoviral vectors include one or more of the following:

●Use of adenovirus serotypes to which most people have not been exposed (eg, group B adenovirus serotype 35, which has a special affinity for hematopoietic cells) [20-23].

●Use of adenovirus serotypes that infect other species (eg, chimpanzee or dog) [24,25].

●Insertion of genes that inhibit immune or inflammatory responses (eg, genes from the adenoviral early region 3 [E3] that is responsible for downregulating host responses) [26].

●Deletion of viral genes that may create immunogenic epitopes [27-29]. However, deletion of E1 and E4 did not eliminate immunogenicity [12-14].

●Induction of immune tolerance [30-32].

Another strategy is to direct the use of adenoviral vectors to clinical settings in which transient high-level expression is optimal (eg, limited treatment of cancer or an infection).

These approaches remain an active area of research.

AAV vectors are less immunogenic than adenoviral vectors, but immunogenicity remains an important issue that interferes with sustained AAV expression [8]. Interestingly, immunogenicity was not predicted in animal models but occurred in clinical trials, likely due to the fact that most individuals have memory immune responses to AAVs due to past infections. Reduced vector doses and use of immunosuppression have been adopted in many subsequent clinical trials. Details of safety, efficacy, and immunosuppressive strategies are presented separately. (See "Gene therapy and other investigational approaches for hemophilia".)

AAV vector can be found in body fluids for several weeks after systemic administration but have not been associated with any identifiable issues in animal and human trials.

There are a number of ongoing clinical trials using AAV for multiple applications including hemophilia, muscular dystrophy, and severe heart failure, and several AAV-based gene therapies have been approved by the US Food and Drug Administration (FDA). (See "Gene therapy and other investigational approaches for hemophilia" and "Investigational therapies for management of heart failure", section on 'Gene therapy' and "Duchenne and Becker muscular dystrophy: Glucocorticoid and disease-modifying treatment", section on 'Gene transfer via viral vectors'.)

GENE THERAPY

Clinical applications of gene therapy — Gene therapy is a fundamentally different approach from most other therapeutics, in that it aims to treat the underlying cause of a disease rather than the symptoms [9]. Gene therapy is in limited clinical use and is under investigation in various disorders [33].

Inherited single gene disorders — Gene therapy is an appealing way to provide a normally functioning gene to an individual who has inherited a pathogenic gene variant or variants that interfere with normal gene function. In many cases, the level of expression does not need to be fully restored to normal; in some cases, very low-level expression is sufficient to greatly improve a disease.

Often, a variant with special properties is used (anti-sickling variant in beta globin for sickle cell disease, more hemostatically active form of factor IX in hemophilia B). In addition to providing more favorable clinical effects, use of a transgene that differs in sequence from an endogenous gene facilitates monitoring of the transgene expression level.

For hematologic disorders and immunodeficiency syndromes that might otherwise be treated with allogeneic hematopoietic cell transplantation (HSCT), gene therapy using modified autologous hematopoietic stem cells (HSCs) and autologous HSCT offers the ability to provide hematopoietic reconstitution with normally functioning autologous blood cells while avoiding the risk of graft-versus-host disease (GVHD). Restoration of normal cell counts may be more rapid than with allogeneic HSCT, potentially reducing the risks associated with cytopenias post-HSCT (anemia, bleeding, and infection).

For ophthalmologic and dermatologic conditions, it may be possible to apply the gene therapy construct locally (directly into the eye or onto the skin).

Examples of genetic disorders that may be amenable to gene therapy include the following:

●Ophthalmologic – Retinal defects are good candidates for gene therapy because the eye is suitable for direct intraocular injection [34]. The eye is also immune privileged (ie, able to accept foreign antigens without mounting an immune/inflammatory response), potentially reducing the likelihood of adverse inflammatory reactions and/or rejection of the vector. In addition, it is possible to treat the eyes independently, allowing assessment of efficacy in one eye before treating the other and reducing the potential for visual impairment.

•A gene therapy for vision loss (voretigene neparvovec-rzyl [Luxturna]) was approved by the US Food and Drug Administration (FDA) in December 2017 for treating the congenital retinal degenerative disorder Leber's congenital amaurosis 2 (LCA2) in individuals with confirmed biallelic pathogenic variants in RPE65 [35]. The construct is based on an AAV2 vector that delivers a wildtype human RPE65 complementary DNA (cDNA) [36]. In preclinical studies, the vector was mainly detected in intraocular fluids, with some transient shedding in tears and serum; there were no adverse events related to the vector. Patients treated with this agent on clinical studies who have been followed for up to nine years have experienced improvements in vision without therapy-associated serious adverse effects, although there were cases of increased intraocular pressure and macular events that mostly resolved. RPE65-associated retinal dystrophy affects approximately 1000 to 3000 patients in the United States. The cost of this gene therapy at the time of its licensing was approximately USD $850,000 for a one-time treatment.

•Another gene therapy construct has been developed for individuals with X-linked retinitis pigmentosa due to biallelic pathogenic variants in RPGR. A study from 2020 reported the results of injection of a codon-optimized RPGR in an AAV8 vector in a bleb under the retina in 18 men with RPGR variants, many of whom had advanced retinal degeneration [37]. There was a dose-dependent improvement in visual function and subjective improvement in visual clarity and field of vision.

The role of gene therapy in retinal disorders is discussed in more detail separately. (See "Retinitis pigmentosa: Treatment", section on 'Gene therapy'.)

●Dermatologic – Gene therapy applied topically to the skin has been investigated as a therapy for dystrophic epidermolysis bullosa, a disorder caused by pathogenic variants in the COL7A1 gene, which encodes the alpha 1 chain of type VII collagen. Type VII collagen is the main constituent of anchoring fibrils. (See "Epidermolysis bullosa: Epidemiology, pathogenesis, classification, and clinical features", section on 'Dystrophic epidermolysis bullosa' and "Overview of the management of epidermolysis bullosa", section on 'Gene therapy'.)

●Clotting factor deficiencies – Hemophilia is highly amenable to gene therapy because small increases in the level of the deficient clotting factor are sufficient to transform the clinical phenotype from one of severe, life-threatening bleeding to a mild bleeding disorder. Because the coagulation proteins circulate in the bloodstream, they can in principle be produced from one of several cell types with access to the intravascular space. An AAV vector that targets the liver and expresses Factor IX, etranacogene dezaparvovec-drlb (Hemgenix), was approved by the US Food and Drug Administration (FDA) in November 2022 for treatment of hemophilia B. Patients treated with this construct have demonstrated stable factor IX activity for up to two years [38]. Gene therapy for hemophilia A has proven more challenging because a greater increase in factor VIII is needed for a good therapeutic effect. Gene therapy for hemophilia is discussed in more detail separately. (See "Gene therapy and other investigational approaches for hemophilia".)

●Respiratory disorders – Cystic fibrosis is a single gene disorder that could be corrected or improved with gene therapy. Because the major clinical complications arise in the respiratory epithelia, delivery of the gene therapy construct could potentially be done by inhalation. This raises the possibility of repeat dosing using a nonintegrating vector.

●Hemoglobin disorders – Hemoglobinopathies (sickle cell disease, thalassemia) are an attractive target for gene therapy because production of a small amount of normal globin chains (or reactivation of fetal hemoglobin expression) can dramatically improve red blood cell (RBC) function and raise the hemoglobin level sufficiently to reduce disease complications. However, the normal globin must be produced in developing RBC precursors, and therefore, to be clinically useful, the transgene must be introduced into HSCs that can expand sufficiently to produce enough mature RBCs. In August 2022, the FDA approved a gene therapy for beta thalassemia, betibeglogene autotemcel (Zynteglo); this therapy is delivered to autologous HSCs in a lentiviral vector and requires HSCT [39]. (See "Investigational therapies for sickle cell disease", section on 'Gene therapy and gene editing' and "Management of thalassemia", section on 'Gene therapy and other stem cell modifications' and "Hematopoietic stem cell transplantation in sickle cell disease".)

●Immunodeficiencies – Immunodeficiency syndromes due to single-gene defects such as adenosine deaminase deficiency (ADA), X-linked severe combined immunodeficiency (SCID), and Wiskott-Aldrich syndrome are good candidates for gene therapy because the missing gene can be provided in HSCs [40]. ADA deficiency was the first application of gene therapy to treat immunodeficiency. Similar to hemoglobinopathies, this approach requires autologous HSCT using an integrating vector inserted into autologous HSCs. Therapy with a gamma-retrovirus-based vector expressing ADA was approved in 2016 in Europe but has not been approved in the United States; one case of leukemia was reported in 2020 [41]. Subsequently, alternative viral vector approaches have been developed to reduce the risk of leukemogenesis; a trial in 50 individuals with ADA-SCID used ex vivo transduction of autologous HSCs with a self-inactivating lentiviral vector encoding human ADA reported robust immune reconstitution without major adverse events [42]. Several other conditions have been investigated, as discussed separately. (See "Overview of gene therapy for inborn errors of immunity".)

●Neurologic disorders – An intravenous AAV gene therapy, onasemnogene abeparvovec-xioi (Zolgensma), was approved for spinal muscular atrophy (SMA) by the FDA in May 2019 [43,44]. It replaces the function of the missing or nonworking SMN1(survival motor neuron 1) gene. As of August 2022, this treatment has been used in over 2300 patients. Steroids are given postinjection to help reduce the risk of hepatotoxicity. However, two deaths have been reported due to acute liver toxicity. (See "Spinal muscular atrophy", section on 'Disease-modifying therapy'.)

A gene therapy construct used to treat metachromatic leukodystrophy caused by deficient arylsulfatase A (ARSA) activity, atidarsagene autotemcel (arsa-cel), appears promising and has been approved in some European countries. This therapy is delivered to autologous HSCs in a lentiviral vector and requires HSCT. (See "Metachromatic leukodystrophy", section on 'Treatment'.)

A lentivirus-based gene therapy, elivaldogene autotemcel (Skysona), was FDA-approved in September 2022 for cerebral adrenoleukodystrophy, caused by abnormalities in the ABCD1 gene, which is required for production of the adrenoleukodystrophy protein [45]. (See "Clinical features, evaluation, and diagnosis of X-linked adrenoleukodystrophy".)

Delandistrogene moxeparvovec (Elevidys) is an AAV-based gene therapy that was FDA-approved in June 2023 for Duchenne muscular dystrophy; this construct produces micro-dystrophin, a shortened version of the dystrophin protein [46]. (See "Duchenne and Becker muscular dystrophy: Clinical features and diagnosis".)

●Other single gene disorders – Several other genetic disorders are potential candidates for gene therapy, including certain inherited metabolic defects, inherited bone marrow failure syndromes, and other single gene disorders. These are discussed in separate topic reviews. (See "Dyskeratosis congenita and other telomere biology disorders", section on 'Therapies to enhance telomere function' and "Overview of the management of epidermolysis bullosa", section on 'Gene therapy' and "Retinitis pigmentosa: Treatment", section on 'Gene therapy' and "Crigler-Najjar syndrome", section on 'Investigational therapies'.)

Both lentiviral-modified HSCs and AAV vectors are under clinical investigation to treat a number of other genetic disorders, including hemophilia A, sickle cell disease, mucopolysaccharidosis, Gaucher disease, Wiskott-Aldrich syndrome, chronic granulomatous disease, SCID, and others [33,47].

Cancer therapy — There are several approaches to using gene therapy as a component of cancer treatment. Some of these may be appropriate for hematologic malignancies as well as solid tumors. Unlike correction of single gene disorders, gene therapy for cancer that attempts to kill malignant cells presents additional challenges related to additional mutations or immunologic changes used by the cancer cells to survive.

Examples include the following:

●Antitumor immune response – Gene therapy can be used to augment the immune response and immune killing of tumor cells. As examples:

•Ex vivo modification of autologous T lymphocytes to make them express a chimeric antigen receptor (CAR) can be used to direct the patient's own lymphocytes to their own tumor cells; this is referred to as CAR-T therapy or T-cell immunotherapy. The use of T cells for CAR therapy has shown remarkable efficacy in certain lymphoid malignancies [48,49]. Commercial CAR-T cell products (eg, tisagenlecleucel, axicabtagene ciloleucel) became available for clinical use in 2017. The use of HSCs for CAR therapy is under investigation [50]. Clinical use of CAR-T therapies is presented separately. (See "Treatment of relapsed or refractory acute lymphoblastic leukemia in adults", section on 'CAR-T' and "Multiple myeloma: Treatment of third or later relapse", section on 'Chimeric antigen receptor T cells' and "Diffuse large B cell lymphoma (DLBCL): Suspected first relapse or refractory disease in patients who are medically fit", section on 'CD19-directed chimeric antigen receptor-T cell therapy' and "Diffuse large B cell lymphoma (DLBCL): Second or later relapse or patients who are medically unfit", section on 'Chimeric antigen receptor T cell therapy'.)

•Talimogene laherparepvec (T-VEC, Imlygic) is a modified herpes simplex virus (HSV) engineered to express granulocyte-macrophage colony-stimulating factor (GM-CSF); it was approved in 2015 in the United States and Europe for treating unresectable melanoma [6,51]. The virus is injected directly into cutaneous, subcutaneous, or nodal lesions; the dose is based on the size of the lesion. Its mechanism is thought to involve recruitment and activation of immune cells to the site of the tumor. In a trial in which 436 patients were randomly assigned to receive T-VEC or subcutaneous GM-CSF, T-VEC was associated with higher rates of overall response and durable response and a trend towards better overall survival (hazard ratio [HR] 0.79, 95% CI 0.62-1.00) [52]. This therapy and its use in melanoma are discussed in more detail separately. (See "Cutaneous melanoma: Management of local recurrence".)

•Intravesical instillation of nadofaragene firadenovec (Adstiladrin) received FDA approval for treatment of BCG-unresponsive invasive bladder cancer in December 2022. This therapy uses a nonreplicating recombinant adenovirus to deliver a copy of interferon alpha-2b into urothelial cells. In a multicenter, open-label study of intravesical administration once every three months for up to four doses, 55 of 103 patients (53 percent) with carcinoma in situ had a complete response, and this response was maintained in 25 of these individuals (46 percent) at 12 months [53].

Other gene therapies targeting the immune response are under investigation.

•Modification of other immune cells such as natural killer (NK) cells or macrophages to express a CAR is also possible [54]. A proposed advantage of this approach is that autologous cells are not required to generate the CAR cells, which means the therapy could be used as an allogeneic "off the shelf" product; clinical trials with such products suggest that these approaches are effective [55]. Even if the host mounted an immune response against the CAR-NK or CAR-macrophages, it might take several days, during which the CAR cells could continue killing the tumor. (See "Treatment of relapsed or refractory chronic lymphocytic leukemia", section on 'Chimeric antigen receptor T cells'.)

●Direct tumor lysis – Oncolytic viruses are an approach to selectively or preferentially killing tumor cells due to viral replication and tumor cell lysis. Nonintegrating viruses have been engineered with mutations that allow viral replication and cell lysis in tumor cells but not normal cells (ie, conditionally replicating adenoviruses).

•ONYX-015 was the first oncolytic or conditionally replicating adenovirus created. It was based on the concept that adenoviruses can kill cells by inactivating the tumor suppressor p53 via the early region 1B (E1B)-55 kD protein, but a large number of tumors lack p53; thus, the viral gene E1B could be removed, leading to a virus that is only able to replicate in and kill p53-negative cancer cells. Normal cells with wildtype p53 would be unharmed. This idea was considered to be a brilliant advance, but apparently the proposed mechanism was incorrect because replication of the E1B-deleted virus was found to be p53 independent. ONYX-015 may be able to kill tumor cells by an unrelated mechanism, possibly involving the protein Y-box-binding factor, which is selectively expressed in cancer cells and can substitute for the function of E1B in viral replication [56,57]. Initial studies of ONYX-015 appeared promising, but the technology was subsequently sold to a company that did not perform trials that included a survival endpoint, so its true efficacy remains unclear [56,58].

A slightly altered version of ONYX-015, Oncorine, was subsequently approved and licensed in China in 2006 for treatment of refractory nasopharyngeal cancer in combination with chemotherapy, based on a surrogate endpoint [59]. In a randomized trial of 82 patients with nasopharyngeal carcinoma, administration of a p53-altered adenovirus in combination with radiotherapy resulted in significantly longer disease-free survival compared with radiotherapy alone [60]. In a randomized trial involving 107 patients with oral cancer, those treated with surgical resection followed by both radiotherapy and a p53 recombinant adenovirus had a lower relapse rate than those treated with surgery and radiotherapy alone [61].

ONYX-015 and related oncolytic viruses are not approved for clinical use in the United States. However, over 10 other oncolytic adenovirus vectors, most of which are being given in combination with immunotherapy, are being investigated in clinical trials in the United States.

•Transduction of tumor cells with a "suicide" gene (eg, herpes simplex virus thymidine kinase [HSV-tk]) that would render them susceptible to the antiviral drug ganciclovir has been proposed as a strategy for certain localized malignancies such as retinoblastoma and mesothelioma [62,63].

•Direct injection of glioblastomas with a gamma-retroviral vector encoding the enzyme cytidine deamidase (vocimagene amiretrorepvec; Toca 511) was considered a promising therapy initially [64]. When a patient receiving this therapy is given systemic 5-fluorocytosine (5-FC), the tumor cells expressing the cytidine deaminase enzyme convert the 5-FC to 5-FU, which has antineoplastic activity. Thus, this strategy results in very high local concentrations of the cytotoxic agent with reduced systemic toxicity. In a series of 56 patients with glioblastoma who were treated using this approach, 16 (30 percent) had some clinical benefit, and 6 (11 percent) had a durable complete or partial response for six months or more. However, a randomized trial showed no benefit of injecting Toca 511 into the tumor resection cavity and treatment with 5-FC compared with standard care [65].

•A genetically modified, live-attenuated poliovirus, which has tropism for glioblastoma cells, was developed to cause cell lysis without damaging normal neuronal tissues; 61 patients with recurrent glioblastoma who were treated with this construct had better two-year and three-year survival than historical controls [66]. The construct was delivered directly to the tumor. The patients were immunized against poliovirus prior to receiving the construct, to prevent possible development of polio.

Some trials using genetically engineered viruses to treat cancer have used direct intratumoral injection, while other therapies such as CAR-T cells are given systemically. In principle, a genetically engineered virus or cell therapy could be administered by either route (injected into tumor cells or administered intravenously). In either case, dissemination of the virus would be critical to killing cancer cells that were not injected and/or that have metastasized elsewhere in the body.

Intralesional and systemic approaches have both been well tolerated, with the most common side effects including influenza-like symptoms (eg, fever, chills, malaise) [52,60,61]. An exception is the cytokine release syndrome (CRS) that results from massive immune activation in individuals with a large tumor burden. These subjects and other immunotherapy strategies for cancer are discussed in more detail separately. (See "Principles of cancer immunotherapy".)

Protection of cells from drug toxicity — Another approach to gene therapy for cancer is to protect normal (noncancerous) cells from chemotherapy drugs, allowing higher doses of chemotherapy to be administered. As an example, some investigations have focused on introducing a drug-resistance gene into normal HSCs as a way to protect them from chemotherapy that is toxic to HSCs in the bone marrow, causing cytopenias and associated risks (eg, bleeding risk with thrombocytopenia, infectious risk with neutropenia).

These hematologic toxicities are typically dose limiting for these drugs, and a source of HSCs that is resistant to cytotoxic chemotherapy could allow higher doses of therapy to be administered more safely. As an example, HSCs have been transduced with a variant form of methylguanine methyltransferase (MGMT) that allows them to survive high doses of chemotherapy drugs metabolized by this enzyme system (eg, temozolomide or carmustine [BCNU]) [67]. In a small study, seven patients with high-grade glioblastoma treated with modified HSCs were given very high doses of chemotherapy and had improved survival compared with historical controls [50].

Potential concerns with gene therapy — Concerns have been raised about the efficacy and safety of gene therapy, both theoretical and based on adverse patient outcomes:

●Immune and inflammatory reactions/immunogenic response – As mentioned above, gene therapy vectors may be ineffective in some individuals due to prior immunity to the virus used to generate the vector construct, or they may induce complications due to an inflammatory reaction to viral proteins, which may be intense. (See 'Integrating versus nonintegrating vectors' above.)

●Insertional mutagenesis/genotoxicity – Genotoxicity, including insertional mutagenesis, aberrant expression of genes near the insertion site, and creation of oncogenic fusion proteins, is a serious concern based on outcomes from early gene therapy trials. Acute lymphocytic leukemia (ALL) and other T-cell lymphoproliferative disorders developed in 5 of 20 children who participated in early gene therapy trials for immunodeficiency syndromes [5,16,17]. Several of these malignancies involved viral insertions near the LIM domain only 2 (LMO2) or other known proto-oncogenes. Some were effectively treated with chemotherapy, but there were some deaths. Myelodysplasia and myeloid leukemias occurred in other trial participants [18,68]. Details of the outcomes for specific immunodeficiency syndromes are presented in detail separately. (See "Overview of gene therapy for inborn errors of immunity".)

A large amount of research into vector design followed, with the goal of altering vector sequences such as promoter and enhancer sequences in the vector long-terminal repeats (LTRs) that were responsible for these genotoxic effects. Improvements have included the development of self-inactivating (SIN) vectors that lack strong enhancer sequences, the introduction of "insulator sequences" that separate promoters from enhancers and/or promote an open chromatin configuration, and the inclusion of microRNA sequences that could block expression of the transgene in certain cell types (eg, stem cells) [69-71]. Subsequently, over 100 patients have been treated on clinical gene therapy trials (not including CAR-T cells) using various combinations of these enhancements, and safety appears to be greatly improved, with no adverse events related to lentiviral vectors [72]. For the most part, efficacy has been quite good. However, the risk of genotoxicity must be considered in all therapies that seek to manipulate the genome (eg, gene therapy and gene editing). (See 'Potential adverse effects of gene editing' below.)

●Cost – Cost may be an insurmountable barrier to gene therapy for many individuals [73]. Cost is challenging to assess as these potentially curative treatments are given once rather than continuously [33].

•An AAV-based gene therapy for a rare genetic disorder was priced at over USD $1 million, which was cited as a reason for low clinical uptake and the eventual decision not to pursue further approval for this therapy [72].

•On the other hand, gene therapy for hemophilia B has been estimated to save over USD $200,000 annually for those who no longer need routine factor prophylaxis [74]. The reported list price for the AAV-based therapy for hemophilia B, etranacogene dezaparvovec-drlb (Hemgenix), was USD $3.5 million at the time of FDA approval [73].

GENE EDITING

Methods and development — Unlike gene therapy, which provides a new gene in a separate vector, gene editing involves directly altering the endogenous genomic sequence of a cell. As its name implies, gene editing is sequence-specific and seeks to make highly specific genomic sequence alterations. The target cells could in principle include any number of cell types including a pluripotent stem cell from an early-stage embryo, a multipotent tissue-specific stem cell, or a mature somatic cell such as a T lymphocyte. After the genome has been edited in a dividing cell, all progeny would be expected to contain the edited version of the gene. Preclinical studies have been performed in various cell types.

Editing a genetic sequence requires that the DNA be cut (eg, a single strand cut or double-strand break [DSB] introduced by an endonuclease enzyme) at a unique position and the cut ends brought together or repaired (figure 1). During the repair process, the original DNA sequence in the vicinity of the cut is replaced by a new, altered sequence. The new sequence is introduced using a template that is delivered into the target cell together with the endonuclease enzyme. The template contains the desired edit.

For the therapy to be effective and safe, the precision of DNA recognition, cleavage, and repair must be extremely high, with exact in-frame changes and an absence of off-target effects (changes to genes or sequence that are not the intended target). This approach has been used effectively in cellular and animal model systems. Several biologic systems and types of nucleases provide these capabilities for editing endogenous double-stranded DNA:

●CRISPR-Cas9 – CRISPR (clustered regularly interspersed short palindromic repeats) and the CRISPR-associated protein 9 (Cas9) endonuclease are part of an evolutionarily conserved, ancient bacterial system for responding to viruses that infect bacteria (bacteriophages) [75]. CRISPR arrays are regions of DNA in bacterial genomes that can be used to store genetic information from infecting viruses and mobile genetic elements. When the bacteria are reinfected with a previously encountered virus, the CRISPR sequence can be transcribed into RNA that can guide effector proteins to the homologous region in the viral genome, where it can be cleaved by endonucleases. Cas9 is one such endonuclease; there are multiple others that can locate and cleave DNA with a specific sequence (eg, Cpf1, Cas12).

In the laboratory, Cas9 can be guided to virtually any DNA sequence by changing the sequence of the guide RNA (gRNA) to match the DNA sequence of interest. Optimized gRNAs can be used to make the gene targeting more specific [76,77]. Cas9 and gRNA can be delivered into cells using the viral and nonviral vectors described above. (See 'Types of vectors' above.)

●Prime editing and base editing – These methods both introduce single strand DNA breaks instead of double strand breaks.

•Prime editing – In prime editing, only one of the strands in the DNA is cut and a modified guide RNA provides the new (edited) sequence [78]. The nonedited strand is then nicked by a Cas9 enzyme and repaired using the new sequence as a template. This method appears to have a very high fidelity for editing small DNA changes (point mutations or small deletions) such as those responsible for sickle cell disease or a common variant in cystic fibrosis.

•Base editing – Base editing also only cuts one of the DNA strands, rather than both. It is used to change a single base (create a single point mutation) in the DNA.

●Zinc finger nucleases (ZFNs) – A family of enzymes referred to as zinc finger nucleases (ZFNs) is able to cleave specific sites in DNA and can be directed to the gene of interest [79]. ZFNs were generated as hybrid proteins by fusion of a DNA-binding module composed of several zinc finger arrays with a DNA-cleaving module from the restriction endonuclease FokI [80]. FokI cleavage is based on dimer formation between its catalytic domains, requiring two zinc finger FokI hybrid proteins to be generated. One monomer binds to the forward DNA strand and the other monomer binds to the reverse DNA strand [80].

Each "finger" recognizes approximately three base pairs of DNA, and the endonuclease can only cut DNA when two ZFNs bind to both strands of the target DNA in the correct orientation. The two FokI monomers must be in close proximity to allow dimer formation, catalytic activity, and generation of double strand breaks (DSBs) [80].

Although the specificity of ZFN binding is high due to the two hybrid proteins and intervening linker sequence, considerable engineering is required to produce zinc fingers capable of binding to any desired DNA sequence, making the process expensive, laborious, and difficult to reproduce [80].

●TALENs – Transcription activator-like (TAL) proteins are a type of protein produced by a plant pathogen (Xanthomonas); they are capable of altering plant gene expression. TALs can be modified to have endonuclease activity by fusing a TAL effector DNA-binding domain (composed of 33 to 35 amino acid repeats) to a DNA cleavage domain [81]. These complexes are termed TALENs (transcription activator-like effector nucleases). Similar to ZFNs, the sequence-independent FokI enzyme functions as a site-specific nuclease, and TALENs can efficiently induce DSBs that require repair by nonhomologous end-joining (NHEJ) and homology-directed repair (HDR) in human pluripotent stem cells and somatic cells [81].

TALENs have the ability to recognize specific DNA sequences through two hypervariable amino acid residues at positions 12 and 13, termed repeat-variable di-residues (RVDs). Although TALENs can be engineered for virtually any sequence, their use is limited by the necessity of engineering novel proteins for each target site, similar to ZFNs [82]. Similar to CRISPR-Cas9 and ZFNs, TALENs are delivered into cells using the delivery vectors described above (see 'Types of vectors' above). However, in vivo delivery of TALENs is hampered by their large size (approximately 3 kb for a single TALEN) and the repetitive nature of TALEN arrays, leading to packaging difficulties into certain viral delivery systems [83,84]. (See 'Characteristics of the vector backbone' above.)

Once the endogenous DNA sequence has been cut, the edits are made by the cell as the sequence is repaired (figure 1). One of two major cellular DNA repair mechanisms (homology-directed repair or nonhomologous end-joining) is triggered for editing processes that create DSBs. For prime editing there is only a single strand break and thus DSB repair is not required:

●Homology-directed repair – HDR uses a new sequence template to repair DSBs, thereby introducing a new sequence that can correct deleterious variants [85]. This process allows for precise editing at the DNA level. The template for the repair can be provided as a donor DNA (eg, a single-strand oligonucleotide DNA with the desired sequence) together with the endonuclease; the cellular DNA repair machinery uses the template during repair of the DSB. This method is favored in most applications because it provides the greatest precision. Research continues to determine the best methods for ensuring the fidelity and specificity of the edits.

●Nonhomologous end-joining – NHEJ repairs the two cut ends of the DNA so that they join again, but it often creates small insertions or deletions (indels) at the repair site. The indels can potentially disrupt target genes by altering the reading frame, in turn causing RNA degradation or production of a nonfunctional protein [86]. NHEJ can be used to disrupt a disease-causing gene that is overexpressed or that acts by a dominant negative mechanism [86]. Alternatively, NHEJ can be used to restore the reading frame of a mutated gene. As an example, in a mouse model of Duchenne muscular dystrophy (mdx mouse), NHEJ has been used to create in-frame removal of exons 20 to 23, thus preventing the transcription of a missense variant in exon 23 that results in a premature stop codon and disruption of dystrophin expression [87].

In some cases, the related phenomenon of micro-homologous end-joining (MHEJ) may be used to fuse two DNA strands together [85]. Unlike NHEJ, MHEJ uses short homology sequences near the DNA break to align the broken ends. This is an error-prone mechanism that is frequently associated with alterations of the original DNA sequence [88].

Clinical use of gene editing — Gene editing is not in routine clinical use, but one gene editing construct was approved in the United Kingdom in 2023, and a small number of patients have been treated with other constructs, mostly as part of research studies [89].

Examples of approved and investigational clinical applications:

●Sickle cell disease and beta thalassemia – The first gene editing-based treatment for a monogenic disorder was approved in the United Kingdom (UK) in late 2023, for a construct that targets BCL11A in hematopoietic stem cells. The product of the BCL11A gene regulates the switch from production of gamma globin, which does not carry the sickle cell or thalassemia disease variants, to beta globin, which does. The targeting construct (exa-cel, previously known as CTX-001) reverts the fetal-to-adult switch so that unaffected fetal hemoglobin is produced. (See "Fetal hemoglobin (Hb F) in health and disease", section on 'Hemoglobin switching and downregulation of Hb F expression'.)

A study from 2021 treated one patient with sickle cell disease and one with beta thalassemia using autologous hematopoietic cells modified by exa-cel; both patients did well for more than a year [90].

Unpublished data used in the UK application followed nearly 100 patients treated with the construct, including 29 with sickle cell disease and 42 with beta thalassemia who were followed for long enough to determine safety and efficacy [91]. The construct is under review in the United States and Europe.

●CAR-T-based cancer therapies – Clinical trials using ex vivo CRISPR-edited chimeric antigen receptor (CAR)-T cells for cancer therapy began in 2018 [92]. Several CAR-T cell cancer therapies have been approved by the FDA. (See "Principles of cancer immunotherapy", section on 'CAR-T cells'.)

●HIV – Hematopoietic stem and progenitor cells that lack the CCR5 receptor for HIV could be used to provide HIV-resistant hematopoiesis, including production of HIV-resistant T cells, in an HIV-infected individual [79]. A case report described an individual who was infected with HIV and underwent allogeneic hematopoietic stem cell transplant (HSCT) as part of therapy for acute lymphoblastic leukemia, using cells that were gene edited to ablate CCR5 prior to infusion [93]. The level of CCR5 disruption in peripheral CD4-positive cells was low (3 percent) and insufficient to allow discontinuation of antiretroviral drugs for HIV. However, the CCR5-negative stem cells persisted for over a year and were able to differentiate, and there was no evidence of gene editing-related adverse effects. Further research regarding the safety and efficacy of this approach is needed (eg, to enable engraftment of higher frequencies of gene-edited stem and progenitor and to monitor for tumorigenesis over time).

●Other single gene disorders – Studies are ongoing in various monogenic disorders in which correction of a pathogenic gene variant in hematopoietic or other tissue-specific stem cells could be curative. Examples include the construct approved for sickle cell disease above and others under investigation for hemoglobinopathies, immunodeficiency syndromes, hemophilia, and cystic fibrosis; the rationales are discussed above in the section on gene therapy. (See 'Inherited single gene disorders' above.)

As with other molecular treatments, targeting to the appropriate tissue(s) and efficient delivery are important considerations.

●Targeting – Studies that take advantage of targeting a gene editing construct directly to hematopoietic stem cells in vivo are ongoing [2]; this approach would bypass the need to administer myeloablative chemotherapy and perform hematopoietic stem cell transplantation. (See 'Inherited single gene disorders' above and 'In vitro versus in vivo transduction' above.)

Correction of a pathogenic gene variant in other affected tissues includes:

•Respiratory epithelium (eg, in cystic fibrosis [CF]) [7].

•Retinal epithelium in an individual with Leber congenital amaurosis [94]. (See "Retinitis pigmentosa: Clinical presentation and diagnosis".)

●Delivery – Delivery efficiency has been a significant limitation for other applications. However, lipid nanoparticles (LNPs) have been found to efficiently deliver CRISPR/Cas9 components to hepatocytes in preclinical trials [95]. For example, intravenous administration of LNPs containing components to disrupt the transthyretin (TTR) alleles is under investigation for hereditary TTR amyloidosis in a phase I trial [96]. Early clinical trials using LNPs to deliver CRISPR/cas9 components are pending for other disorders, including hereditary angioedema (targeting disruption of kallikrein expression) and familial hypercholesterolemia (targeting PCSK9 expression).

In contrast to these uses in somatic cells (including somatic stem cells), genome editing that would create a change in the germline has not been studied clinically. Ethical considerations were brought to the attention of the general public following a report that genome editing had been performed without medical oversight in late 2018, as discussed below. (See 'Ethical concerns with germline genome editing' below.)

Potential adverse effects of gene editing — The potential for off-target effects, such as cleavage or genetic modification of DNA regions other than the intended target site, is a serious concern with gene editing; disruption of oncogenes, tumor suppressor genes, and/or DNA repair genes can result in significant cellular toxicity and/or development of cancer [86]. Efforts to reduce off-target effects and to increase the specificity and fidelity of genome editing systems are in progress [86,97,98]. The specificity of gene editing is also affected by the dose of gene therapy construct administered as well as the expression pattern of the gene editing system and the number, type, and differentiation stage of the edited cells [86].

It was shown in a 2018 study that cells with a wild-type p53 gene have reduced efficiency of gene editing [99]. It follows that cells that have mutant p53 are more efficiently edited. However, it is well known that the mutations affecting the p53 gene enhance the risk of cancer transformation [100]. Thus, these cells are potentially at an even higher risk for transformation due to gene editing than cells with wild-type p53.

Another potential concern for clinical translation of genome editing is the possibility that the recipient's immune system will mount an adaptive immune response to the corrected gene that was previously absent [86]. As an example, an individual with severe hemophilia A whose immune system had not previously been exposed to factor VIII might develop inhibitory antibodies to factor VIII, similar to those seen with factor VIII infusions (protein therapy) [101]. (See "Inhibitors in hemophilia: Mechanisms, prevalence, diagnosis, and eradication", section on 'Mechanisms of formation and action'.)

These concerns are mostly similar to those of gene therapy. (See 'Potential concerns with gene therapy' above.)

More specific to gene editing is the possibility that genomic variations in some populations may be more likely to lead to off-target effects (eg, if some populations have a region of sequence homology in an unrelated gene that is identical to the gene of interest, the unrelated gene may also undergo editing). One study estimated that the likelihood of single-nucleotide variants (SNVs) creating off-target sites in a human genome would be approximately 1.5 to 8.5 percent, depending on the genome and site-selection method; however, mutations may be rare at targeted sites or may not have functional consequences [102].

Despite these potential adverse events, there is widespread consensus in the medical and scientific communities that genome editing of somatic cells has the potential to cure inherited disorders and possibly some cancers, and as such its pursuit is reasonable, as long as appropriate attention is paid to these concerns.

Ethical concerns with germline genome editing — In contrast to gene editing in somatic cells, the concept of gene editing in human gametes (egg or sperm) or embryos to permanently modify the germline raises significant ethical concerns. These concerns were addressed in an opinion article from 2015 and several consensus documents published in 2017 to 2018 and reaffirmed in 2023 [103-108]. Details include:

●Lack of ability to consent before birth

●Lack of differentiation between research and clinical applications

●Equitable access and allocation of resources

●Exploitation for nontherapeutic modifications (eg, "enhancement" of a feature rather than treatment/prevention of a disease)

●Potential unanticipated adverse effects

●Potential effects on future generations, including need for monitoring and lack of consent

●Potential for discrimination due to health risks and/or enhanced performance

Some of these concerns apply to gene therapy and gene editing of adult stem cells as well.

An announcement in late 2018 that two girls were born after germline genome editing to prevent expression of the HIV receptor CCR5 was met with strong statements condemning the practice in general and highlighting the lack of sound medical basis or safety considerations for these specific individuals [109-112]. The editing had ostensibly been performed to prevent transmission of HIV, but several experts have commented that this is not a reasonable approach to controlling HIV transmission and that CCR5 is also implicated in neuronal plasticity, learning, and memory [110,113]. Other concerns have focused on the lack of appropriate safeguards to protect the children and lack of public discussion and input regarding the personal and societal harms of such an approach.

Despite these concerns, it is possible that genome editing in the germline may someday be used to "correct" a pathogenic mutation or to overcome the adverse effects of a pathogenic mutation (eg, by altering a downstream component of a pathway) and thus cure or ameliorate disease [114]. This contrasts with the use of germline genome editing to enhance or alter nondisease traits, which is not considered medically ethical.

GENE SILENCING

Overview of mechanisms — Unlike gene therapy and gene editing, gene silencing does not add to or alter the primary genetic information in the patient's cells; rather, it aims to use molecular methods to reduce the expression level of one or more genes, shutting down expression of a defective dominant gene with or without provision of a wildtype allele [81]. Gene silencing can be extremely specific (eg, targeting a single gene using RNA interference [RNAi]), or it can be more global (eg, targeting a group of genes via epigenetic changes).

RNA interference — RNAi is an evolutionarily conserved mechanism in bacteria and eukaryotes for reducing gene expression in response to small (approximately 20 to 25 nucleotides) exogenously or endogenously derived double-stranded RNAs (dsRNAs) [115-117]. These dsRNAs, also called small interfering RNAs (siRNAs), can bind tightly by antisense base pairing to target messenger RNAs (mRNAs) to which they have homologous sequence. This causes the mRNAs to be degraded, which in turn reduces their translation into proteins (ie, reduces their expression). RNAi is thought to be a type of immune defense against invading genetic material as well as an alternative form of gene regulation that can be turned on or off during development.

The sequence-specific silencing of genes by RNAi can be triggered by different types of RNAs including siRNAs, short hairpin RNAs (shRNAs), microRNAs (miRNAs), and other noncoding RNAs (ncRNAs) such as long ncRNAs and pyknons (repeated elements frequently found in the 3' untranslated regions of genes) [118].

The process of RNAi commences with the processing of long dsRNA and complex hairpin precursors into shorter duplexes, or siRNAs, by the enzyme Dicer. The siRNAs are then loaded into a large complex of proteins (the RNA-induced silencing complex [RISC]) that eventually degrades the target mRNA, resulting in inhibition of the target gene expression [118].

Clinically, silencing can be achieved with RNA constructs provided in several forms with different properties, including siRNAs, antisense oligonucleotides (ASOs; AS-ODN), shRNAs encoded from a viral vector, or small RNAs encapsulated in particles that improve their stability.

Two major concerns with RNAi therapies are their short half-life (due to rapid degradation) and their potential for immune stimulation.

●A number of modifications are under investigation for increasing the stability of siRNA constructs, including chemical modifications of the RNA bases or backbone, conjugation to nanoparticles or lipids, or use of a viral vector to deliver the sequence for an shRNA [115,119].

●High doses of siRNAs are known to switch on the innate immune response and the production of cytokines [120,121].

Epigenetic modifications — Epigenetic modifications involve addition of methyl groups or other side chains to DNA or histone proteins. Typically, these changes are distributed widely across the genome. These changes are a major feature of normal gene regulation. They have been adapted for use in several disorders such as cancer and hemoglobinopathies, as discussed separately. (See "Principles of epigenetics".)

Clinical applications of gene silencing — A number of gene-specific RNAi therapies are commercially available, and many others are under investigation. Examples include:

●Neuromuscular and neurologic disorders

•Four ASO products have been approved by the FDA for Duchenne muscular dystrophy, including eteplirsen (Exondys 51), which binds to dystrophin and causes alteration in exon splicing, restoring the function of dystrophin. Over three years of follow-up, patients treated with eteplirsen had a slower decline in disease progression compared with historical controls [122].

•An ASO product, nusinersen (Spinraza), administered intrathecally, is approved by the FDA for the spinal muscular atrophy; a meta-analysis found that treatment was effective for infants [123].

An ASO directed against huntingtin (HTT), which produces an abnormal form of the huntingtin protein in Huntington disease [124]. (See "Huntington disease: Management", section on 'Investigational therapies'.)

•An ASO directed against superoxide dismutase (SOD1) that may be injected intrathecally to treat amyotrophic lateral sclerosis (ALS) [125]. In a phase III study, however, there was no significant difference in disease progression compared with placebo and serious neurologic adverse events occurred in 7 percent of treated recipients [126]. (See "Disease-modifying treatment of amyotrophic lateral sclerosis".)

•ASOs or epigenetic regulators may increase frataxin gene expression in Friedreich ataxia [127].

●Hemophilia – An ASO directed against antithrombin that may be used to shift the hemostatic balance and reduce the risk of bleeding in individuals with hemophilia. (See "Hemophilia A and B: Routine management including prophylaxis", section on 'Prophylactic therapies under development'.)

●Hemoglobinopathies – A shRNA directed at BCL11A (a repressor of fetal hemoglobin synthesis) that may be used to treat beta hemoglobinopathies such as sickle cell disease and beta thalassemia [128]. A study from 2021 treated six patients with sickle cell disease with autologous HSCs transduced with a lentiviral vector encoding this shRNA [129]. This produced sustained responses; however, three patients developed myeloid neoplasms, and the study was placed on hold while awaiting analysis of the cause. (See "Investigational therapies for sickle cell disease", section on 'Gene therapy and gene editing' and "Management of thalassemia", section on 'Gene therapy and other stem cell modifications' and "Fetal hemoglobin (Hb F) in health and disease", section on 'Hemoglobin switching and downregulation of Hb F expression' and "Hematopoietic stem cell transplantation in sickle cell disease".)

●Amyloidosis – Antisense-based therapies that reduce expression of the protein transthyretin (TTR), which forms pathogenic amyloid deposits in TTR-amyloidosis:

•An ASO directed against TTR (inotersen [Tegsedi]), which is administered subcutaneously once weekly, was approved by the FDA in 2018 [130]. A related product (vutrisiran [Amvuttra]), approved in June 2022, can be dosed less frequently (once every three months).

•A lipid-encapsulated siRNA directed against TTR (patisiran [Onpattro]), which is administered intravenously every three weeks, was approved by the FDA in 2018 [131,132].

Evidence for the efficacy of these therapies and their clinical use are presented separately. (See "Overview of amyloidosis", section on 'Treatment' and "Cardiac amyloidosis: Treatment and prognosis", section on 'Disease-specific therapy for ATTR amyloidosis'.)

●Lipid disorders – An ASO product for treatment of familial hypercholesterolemia, mipomersen (Kynamro), which is injected subcutaneously once weekly and lowers apolipoprotein B levels, was FDA approved in 2013 [133]. However, Kynamro was associated with adverse effects including hepatotoxicity and injection site reactions, which led to discontinuation in a majority of patients, and FDA approval was withdrawn in 2019.

●Infections – There are several ongoing clinical trials, using both siRNA and ASO approaches, for treatment of hepatitis B. In most trials, treatment is combined with a nucleoside analogue. Dose-dependent reductions in hepatitis B surface antigen (HBsAg) have been observed. In a trial of the ASO bepirovirsen, 10 percent of patients treated with ASO alone had undetectable HBsAg 24 weeks after discontinuing treatment [134].

Additional examples continue to accrue [135]. As more patients have been treated with ASO products, a number of acute drug toxicities, but only rare severe reactions, have been observed, especially hepatotoxicity, nephrotoxicity, and hypersensitivity reactions [136].

Notably, RNAi products are significantly less costly to develop compared with gene therapy approaches.

Approaches that alter epigenetic regulation are discussed separately. (See "Principles of epigenetics", section on 'Therapeutic uses'.)

SUMMARY

●Overview – Several approaches are available for manipulating gene sequences and gene expression. Gene therapy involves the introduction of a new gene; gene editing involves altering the primary DNA sequence of an existing gene; and gene silencing uses genetic techniques to reduce the expression level of a gene. These therapies can be targeted to a variety of cell types, and they can be introduced into cells outside the body or administered directly to the patient. There are a number of types of vectors available with different properties (eg, integrating versus nonintegrating, more or less immunogenic). Advantages and disadvantages of some of the commonly used vectors are described above. (See 'Overview' above.)

●Gene therapy

•Single gene disorders – For monogenic (single gene) disorders in which an individual carries a pathogenic gene variant, gene therapy may be used to provide a normal copy of the gene or a modified version of a gene with special properties such as increased activity. (See 'Inherited single gene disorders' above.)

•Cancer – In cancer immunotherapy, gene therapy can provide the patient's T cells with a modified chimeric antigen receptor (CAR) that brings T cells and tumor cells in close proximity to enhance tumor cell killing. Oncolytic adenoviral vectors can selectively replicate in tumor cells, leading to cell lysis or increased sensitivity to chemotherapy and/or radiotherapy. (See 'Cancer therapy' above.)

•Adverse effects – A major concern of gene therapy is genotoxicity (introduction of deleterious abnormalities in other genes); this resulted in leukemia in early gene therapy trials. A second concern is the potential for immunogenicity, which can result in acute inflammatory reactions, sometimes severe, and which can possibly lead to rejection of the gene therapy vector. These concerns have been substantively addressed by modifying gene therapy vectors, leading to a small number of therapies being approved for clinical use in the United States in 2017. (See 'Potential concerns with gene therapy' above.)

●Gene editing – Gene editing involves the site-specific alteration of an existing gene. Editing the genetic sequence requires that a region of DNA be cut by an endonuclease (eg, CRISPR-Cas9 system) and the two cut ends brought together, often with a new or corrected sequence inserted between them (figure 1).

There are many clinical settings in which gene editing of somatic cells (including hematopoietic stem cells) may be useful, and some preclinical and early clinical studies have been performed. The precision of DNA recognition, cleavage, and repair must be extremely high; concerns about the fidelity of editing need to be addressed. Ethical issues apply to germline editing using eggs, sperm, or embryos. (See 'Gene editing' above.)

●Gene silencing

•RNA interference – RNA interference (RNAi) uses molecular methods (eg, introduction of an antisense oligonucleotide [ASO] or short hairpin RNA [shRNA]) that take advantage of the evolutionarily conserved machinery to degrade messenger RNA and in turn reduce the expression level of a gene or group of genes. Several gene silencing therapies are clinically available or under development. (See 'Gene silencing' above.)

•Epigenetic therapies – Gene silencing using epigenetic therapies is discussed separately. (See "Principles of epigenetics".)

●Clinical examples – Clinical applications of these methods are discussed in more detail in topic reviews on the specific conditions for which they are used. Selected hyperlinks are provided above. (See 'Clinical applications of gene therapy' above and 'Clinical applications of gene silencing' above.)