The Concept Of Gene Therapy

Published Date: 02 Nov 2017

Introduction

The concept of "gene therapy" was first postulated in the early 1970s,[1] with the advent of recombinant DNA technology. The basic idea is simple: replace or supplement a damaged gene with a copy of a functional gene in patients suffering from a genetic disorder. In practice, however, this is incredibly difficult. Indeed, twenty years of research was required before the first human patient, a young girl suffering from severe combined immune deficiency due to a defective adenosine deaminase gene (ADA-SCID), was successfully treated using gene therapy in which her T-lymphocytes were isolated from whole blood, grown in culture, transduced with a lentiviral vector, and re-infused into her bloodstream.[2] However, in the subsequent twenty years since that first success, and along with some serious setbacks, much progress has been made.

There are two basic forms of gene transfer used in gene therapy: in vivo, or direct; and ex vivo, or indirect. Direct gene transfer involves the insertion of the therapeutic gene construct directly into somatic cells through local or systemic delivery of a vector (e.g., a virus, liposome, nanoparticle, etc.) containing the construct. Indirect gene transfer requires the removal of cells (e.g., hematopoietic stem cells, immune cells, etc.) from the patient, insertion of the therapeutic gene construct into the cells ex vivo, selection and amplification of modified cells, and reinfusion of the modified cells into the patient.[3] Each method has its pluses and minuses, which will be discussed later.

Selecting Viral Vectors

Viruses are the most well-characterized, and indeed the most obvious, method of gene delivery. In essence, a virus is a gene-delivery machine: viral replication requires the injection of the virus’s genome into a host cell, and the subsequent hijacking of the cell’s macromolecule synthesis machinery to produce more viruses.[4] In terms of gene therapy, the qualities of the therapeutic transgene construct drive the selection of the viral vector. At root, the primary selective criterion is stability of the construct within the cell. For constructs that require long-term stability through integration into the cellular genome, the viral vector must be a retrovirus or lentivirus. However, if the construct only requires transient expression after transduction, vectors such as adenovirus or adeno-associated virus (AAV) can be used.

Under most circumstances, viral vectors used for gene therapy are "replication incompetent," meaning they cannot reproduce once they infect a target cell in the patient. The reason for this is two-fold. First, "replication competent" vectors simply have less genomic real estate for a therapeutic transgene, as they require intact genes for the viral proteins necessary for the production of new viral particles. Second, and more importantly, replication competent viruses tend to cause the deaths of their host cells, either through lysis (as with adenoviruses) or through the induction of apoptosis by cytotoxic T-lymphocytes and/or natural killer cells. In addition, an active infection by replicating viral vectors would provoke a potentially dangerous immune response in the patient. One of the few applications where replication competence would be therapeutically beneficial is in the treatment of solid tumors, in which viruses, modified to specifically target tumor cells, would both kill the cells directly as well as mark them for destruction by the immune system. [5]

An important factor in the choice of vector is the total size of the transgene construct, including viral sequences. Retroviral and lentiviral vectors carry single-stranded (+)RNA genomes of between eight and ten kilobases and, in addition to the therapeutic transgene and its promoter, a functional construct must include: a packaging signal for proper loading of the viral RNA into the capsid; a primer binding site and poly-purine tract for proper reverse transcription; and long terminal repeat sequences at the 3’ and 5’ ends for proper integration into the host genome.[6] Adenoviruses carry significantly larger double-stranded DNA genomes of between 25 and 45kb, and therefore are capable of carrying much larger therapeutic transgene constructs; however, as with retro- and lentiviral vectors, the full construct must include a packaging sequence, as well as 5’ and 3’ inverted terminal repeats for genomic stability.[7] At approximately 4.5kb,[8] the single-stranded DNA genome of adeno-associated viruses (AAV) is the smallest of the viruses covered here; however, the only viral sequences required in the full construct are short 5’ and 3’ inverted terminal repeats.[9]

Retroviral and Lentiviral Biology

Retroviruses (family Retroviridae) and lentiviruses (Lentiviridae, a genus of Retroviridae) are membrane-enveloped viruses that can only replicate by first inserting their genomes into those of their host cells, through processes called reverse transcription and integration. In brief, a mature retroviral particle contains the viral genome, encoded as a dimerized, linear, single stranded, positive-sense mRNA, and two enzymes: reverse transcriptase (RT) and integrase (INT). Upon successful docking and fusion of the viral envelope with the host cell’s plasma membrane, and subsequent release of the capsid into the cytoplasm, RT translates the viral mRNA into linear, double-stranded cDNA through a coordinated sequence of DNA polymerizations and RNA degradations, using a modified transfer RNA as the initial primer.[10] Upon successful reverse transcription, the viral cDNA combines with INT to form the preintegration complex. INT processes the 3’ termini of the viral cDNA long terminal repeats (LTR) to form 5’ sticky ends, and catalyzes the cleavage of the host DNA and subsequent integration of the viral cDNA into the host genome. The integrated viral cDNA is thereafter referred to as a "provirus," and all subsequent production of viral mRNA and proteins will use the provirus as the template.[10]

For the most part, retroviruses can only integrate their genomes into actively-dividing cells during metaphase, as they lack the ability to enter the intact nucleus through the nuclear pore. Lentiviruses, however, are able to integrate with both dividing and quiescent cells, as they are able to actively transport the preintegration complex through the nuclear pore with the use of viral accessory proteins containing nuclear localization signal motifs. As such, lentiviral proviruses are able to integrate into the genomes of terminally differentiated cells such as neurons, microglia, and macrophages.[11]

It should be noted that provirus integration is inherently mutagenic, and there remains the possibility that host genes and/or gene regulatory sequences can be disrupted; in fact, some of the first retroviruses to be characterized were discovered due to their oncogenicity. Although integration is essentially random, some retroviruses, in particular gamma-retroviruses, show a bias toward gene-dense chromosomal regions, areas of active transcription, and transcriptional start sites.[12, 13] Indeed, there have been cases of integration-induced leukemia in patients who received gene therapy for X-linked SCID using a gamma-retroviral vector.[14, 15] Lentiviruses, on the other hand, appear to have minimal integration bias,[16, 17] although there is in vitro evidence that human immunodeficiency virus (HIV), a lentivirus, does have a bias for integration within active genes.[18]

Adenoviral and AAV Biology

Adenoviruses (family Adenoviridae) are non-enveloped viruses composed of icosahedral capsids containing linear, double-stranded DNA genomes. At each vertex of the capsid, there is a protein structure called a penton that serves as the anchorage for a host-cell-binding fiber.[7] For most serotypes, successful viral entry requires binding of the terminus of the fiber to a "coxsackievirus-adenovirus receptor" (CAR) on the plasma membrane of the target cell, and subsequent interaction between the penton and cellular integrin. Upon penton/integrin interaction, the virus is endocytosed via clathrin-coated pit formation. The viral particle escapes the endosome via an unknown mechanism involving acidification of the endosome, continued penton/integrin interaction, and dissociation of the fibers from the capsid, and is transported to the nucleus along the microtubules. Upon arrival at the nuclear pore, the capsid fully disassembles, and the viral genome is transported into the nucleus where transcription and replication of the viral genome can take place.[19, 20] Although viral proteins are produced in the cytoplasm, they are returned to the nucleus for virion assembly. Release of mature virions requires lysis of the host cell.

Adeno-associated viruses (AAV, family Parvoviridae, genus Dependovirus) are non-enveloped viruses composed of icosahedral capsids containing single-stranded DNA genomes. An AAV virion is roughly one fifth the size of an adenovirus, and inherently replication incompetent. As the name suggests, AAV requires co-infection of the host cell with an adenovirus in order to replicate.[8] The method of infection for AAV is similar to that of adenovirus: receptor binding (in the case of AAV, heparin sulfate proteoglycan and integrin),[21, 22] endocytosis via clathrin-coated pits, escape from acidified endosomes, and translocation to the nucleus.[23]

Immune Response to Viral Infection

One of the greatest hurdles to the use of viruses as vectors for in vivo administration of gene therapy is the patient’s immune response. The immune system does not differentiate between a pathogenic virus and a viral vector carrying a of therapeutic transgene constructs: both are viewed as hazardous invaders bent on subverting a cell’s normal functions for their own purposes. Cells have innate methods to recognize viral DNA (or RNA) and proteins, including the NOD-like receptors and the endosomal Toll-like receptors (TLR), which utilize the NF-B pathway to express interferon, interleukin-1β (IL-1β), and other pro-inflammatory cytokines[24, 25] to impede transduction and to signal natural killer (NK) cells. In addition, the cellular proteasome processes viral proteins for presentation to adaptive immune cells via the class 1 major histocompatibility complex (MHC-I) for presentation to cells of the adaptive immune system, such as cytotoxic CD8+ T-lymphocytes.[26] Upon antigen presentation, the CD8+ T-cell releases cytokines to induce apoptosis in the infected cell.

Free virus can also be endocytosed by professional antigen-presenting cells such as dendritic cells and macrophages, which process the viral proteins for display via the class II major histocompatibility complex (MHC-II). Antigen presentation by these cells recruits and induces the proliferation of CD4+ T-lymphocytes, which are responsible for the activation of antibody-producing B-lymphocytes, stimulation of CD8+ T-lymphocytes and macrophages, and immunological "memory."[26] This form of antigen presentation primes the body to immediately repel any subsequent infection with the same virus, which poses a serious hurdle in gene therapy, where multiple infusions of vector may be needed to deliver sufficient therapeutic constructs to the patient’s cells.

Immune Response in Gene Therapy

As discussed above, the immune system poses a severe challenge to the systemic administration of viral vectors for gene therapy. Adenoviral vectors are especially problematic: an estimated 70% of the population has antibodies to at least one adenoviral serotype, and adenoviruses elicit a very strong innate immune response.[27, 28] Indeed, adenoviral capsid proteins are often administered as vaccine adjuvants. In addition, the lack of transgene integration with the use of adenoviral vectors would require multiple doses of vector as transgene expression attenuates; however, the establishment of immunological memory, and an associated humoral immune response against the vector, would necessitate the use of a different vector serotype for each administration.[27] AAVs, in contrast to adenoviral vectors, elicit a relatively mild innate immune response; however, the single-stranded DNA genome does activate the NF-B pathway via TLR, indicating that an innate response still occurs, although the duration of cytokine expression is much shorter.[28, 29] In addition, some 70% of the population carries neutralizing serum antibodies against the AAV-2 serotype and, since antibody cross-reactivity between serotypes is highly common, a large proportion of that population carries antibodies against all known serotypes. As these antibodies are capable of neutralizing large numbers of virus, even with relatively low serum titers for the antibodies, the use of AAV vectors may be ineffective in patients showing humoral immunity.[30]

Retroviruses and lentiviruses also provoke a robust innate immune response via TLRs, with concomitant reduction in transduction efficiency, as well as an adaptive immune response. Treatment with anti-inflammatory drugs like dexamethasone prior to administration of lentiviral vectors, however, has been shown to reduce the degree of innate immune response and improve transduction efficiency,[31] and the use of vectors stripped of nearly all non-essential viral proteins appears to minimize cytotoxic T-lymphocyte response to transduced cells.[27]

The immunogenicity of the therapeutic transgenic product must also be taken into account. The products of transgenes of non-human origin may be recognized as "non-self" and induce an MHC-I mediated immune response from CD8+ T-lymphocytes and, if excreted by the transduced cell and processed by antigen presenting cells, induce a neutralizing humoral immune response.[31] As such, whenever possible, therapeutic transgenes should be of human origin. The type of promoter used to drive transgene expression can also affect the intracellular inflammatory response and attenuation of expression, with viral promoters being the most immunogenic, and the most likely to be silenced through methylation.[28] The use of cell and/or tissue-specific promoters of human origin may reduce the likelihood of attenuation, especially when incorporated into a transgenic construct intended for genomic integration. This would also have the added effect of preventing transgene expression in off-target cells.[32]

Transient Infection, Permanent Integration: Homologous Recombination

The development of endonucleases that are capable of targeting specific DNA sequences opens up a new possibility for gene repair: the excision of a defective gene, or sequence of a gene, and subsequent replacement with functional sequence, leaving intact the endogenous expression controls.[33, 34] While retroviral and lentiviral vectors are capable of integrating therapeutic gene constructs into host-cell genomes, they do not remove the defective gene itself. In addition, there remains the possibility of insertional oncogenesis at the integration site,[12] as seen in patients who developed leukemia after European trials of a lentiviral vector to treat X-SCID.[14, 15] By harnessing the ability of these new targeted endonucleases to cleave specific locations in the host genome, and taking advantage of the cell’s endogenous homologous recombination machinery, it is possible to avoid the possible negative outcomes inherent in genomic integration.[33]

Zinc-finger endonucleases (ZFN) are fusion proteins consisting of a defined number of C2H2 zinc-finger DNA-binding domains, each of which can recognize a series of 3-4 nucleotides, and a non-specific restriction enzyme, FokI.[35] The series of Zn-finger domains can be engineered for high specificity to DNA sequences, allowing the restriction enzyme to cleave only at the specified location. Transcription-activator-like effector nucleases (TALEN) function in a similar fashion; however, the DNA binding domain is derived from transcription-activator-like effector proteins produced by the pathogenic bacterium Xanthomonas.[36] In each case, the FokI domain requires dimerization in order to cleave DNA and generate a double-strand break, accomplished by providing two distinct ZFN or TALEN proteins, each targeted to a region immediately upstream or downstream of the intended cleavage site. Upon cleavage, one of two repair mechanisms can be employed. If an exogenous therapeutic DNA construct is introduced to the nucleus along with the targeted nucleases, the cell’s homology-directed repair machinery is activated, using the exogenous DNA as the template for repair and thereby "fixing" the target gene. Alternatively, the break could be induced in a "safe harbor" region (i.e., a non-coding, non-regulatory area of the genome) for integration of a full gene construct without the possibility of insertional oncogenesis.[34] In the absence of exogenous template DNA, the cell’s non-homologous end joining machinery is activated to close the break, introducing frame-shift mutations and, therefore, "knocking out" the target gene.[34, 37]

Because gene editing using ZFNs and TALENs does not require integrase, retroviral and lentiviral vectors are not necessary for this method. Indeed, the administration of a single adenoviral or AAV vector (or simply transfection with a plasmid) carrying a construct encoding the two nuclease genes and the template DNA has the potential to permanently modify the target cell while the construct itself is only transiently expressed.[34] The immunological concerns regarding adenoviral vectors would remain; however, as the transgene would be integrated in this model, the need for additional administrations of vector is reduced, thereby obviating the effect of humoral immunity against the vector itself. Alternatively, a non-integrating lentiviral vector (NILV) may be used to reduce the severity of immune response. NILVs are lentiviruses with mutations in their integrase, their LTRs, or both, which prevent their cDNA from becoming integrated into the host genome. Instead, the cDNA exists as a transient episome in the host nucleus.[38]

It should be noted that both ZFNs and TALENs have the potential for off-target DNA binding and genotoxicity due to improper DNA cleavage. Although increasing the number of DNA-binding domains increases the specificity of the target sequence, the use of too many domains may allow non-specific targeting.[39] In addition, the nuclease proteins are xenogenous, and therefore can be recognized as "non-self" by the immune system;[40] however, insofar as their expression is transient, their effect on long-term success of the therapy should be minimal.

Early Success: Ashanti DeSilva and ADA-SCID

As mentioned in the introduction, the first successful clinical application of therapeutic gene transfer was performed in 1990. The patient, a four-year-old girl named Ashanti DeSilva, suffered from a form of severe combined immune deficiency due to a homozygous mutation in her adenosine deaminase (ADA) gene. ADA is involved in purine metabolism, and gene deficiency leads to a buildup of 2’-deoxyadenosine within the cytoplasm, which is toxic to T-lymphocytes. In addition, this overabundance of 2’-deoxyadenosine inhibits ribonucleotide reductase, preventing the production of other deoxynucleotides and, therefore, DNA synthesis during S-phase. ADA deficiency also leads to a buildup of S-adenosyl-L-homocysteine, an intermediate product of the purine salvage pathway that is toxic to immature lymphocytes.[41] The monogenic nature of the disease makes it an attractive target for gene therapy.

The gold standard of care for patients with ADA-SCID is bone marrow transplantation; however, enzyme replacement therapy using PEGylated bovine ADA (PEG-ADA) is also used as an interim therapy, or if a histocompatible donor cannot be found.[2, 41, 42] In Ashanti DeSilva’s case, her responsiveness to PEG-ADA was rapidly declining after two years of enzyme replacement therapy, and no suitable bone marrow donor could be found. As her health deteriorated, her parents enrolled her in a small proof-of-concept clinical trial, in which her T-lymphocytes were isolated from her blood by apheresis and treated with a retroviral vector, derived from Moloney murine leukemia virus, carrying the functional human ADA gene. The cells were then amplified in culture for approximately ten days and reinfused into her bloodstream; however, there was no selection for enrichment of transduced cells.[2]

Within six months of the therapy, DeSilva’s peripheral T-lymphocyte levels increased to normal levels, and remained at those levels for the next four years. Her ADA level, previously undetectable, increased to a level consistent with a patient hemizygous for the gene and, as with her T-lymphocyte counts, remained at that level for the next four years. Although she did continue with enzyme supplementation therapy at the recommendation of her doctors, she no longer demonstrated a loss of responsiveness to the treatment, and remained at the PEG-ADA levels used prior to the gene therapy.[2] Information on her current health status is extremely difficult to find, but the most recent mention of her online, dating from 2012, indicates that she is healthy.[43]

Anti-Vector Response Gone Mad: The Gelsinger Case

The immune response to viral vectors does not simply pose a threat to the success of the therapy; it can also pose a threat to the life of the patient. An extremely strong immune reaction to an adenoviral vector led to the death of Jesse Gelsinger during a Phase I safety trial at the University of Pennsylvania in 1999, and to the most significant setback in the nascent gene therapy field.

Jesse Gelsinger was an eighteen-year-old from Arizona who suffered from an ornithine transcarbamylase (OTC) deficiency. The OTC enzyme is involved in the processing of the nitrogenous waste products (particularly ammonia) formed by protein catabolism into urea, which is subsequently excreted by the kidneys. OTC deficiency is a rare X-linked, single-gene defect, occurring in between 1:40,000 and 1:80,000 live births. Complete deficiency is typically fatal in early infancy, as excessive serum ammonium levels (hyperammonemia) will lead to encephalopathy and death. Roughly half of patients with partial deficiency will die before the age of five without a liver transplant; this makes the gene a good candidate for transient transgene supplementation, to buy time before a liver transplant or until a drug and dietary regimen can be implemented.[27, 44] Mr. Gelsinger’s deficiency was not inherited, and comparatively mild; in his case, the deficiency was due to a spontaneous mutation during embryogenesis, leading to mosaicism. His condition was controlled through drug therapy and strict dietary management, although he had suffered several bouts of hyperammonemia over his lifetime, the most recent occurring nine months prior to his participation in the gene therapy study.[44-46]

The study in which Gelsinger enrolled was designed to test the safety of the vector and transgene in otherwise healthy patients with controlled partial OTC deficiency. The eighteen study participants were divided into four three-person cohorts, one four-person cohort, and one two-person cohort, with each cohort receiving a single administration of vector at a different dosage. Gelsinger was a member of the two-person cohort that received the highest dose of vector (6.00x1011 particles/kg).[45, 47] The bioethics panel that approved the study protocol argued that, because the study was looking specifically at the safety of the therapy and not its effectiveness as a treatment, and since no clinical benefit was anticipated, proper informed consent could not be received from patients, or the parents of patients, with life-threatening OTC deficiency. The informed consent paperwork cited three major risks: the possibility of liver inflammation, potentially life-threatening, up to and including hepatic failure; the possibility that immune response to the adenoviral vector could damage the liver; and the possibility that receiving the vector in the trial would induce immune memory, and thereby prevent the participants from receiving future therapeutic treatments. Participants were advised that liver damage or failure would likely require a liver transplant, that a liver biopsy would be required prior to participation in the study, and that the biopsy itself had a small risk of complications up to and including death.[44]

Gelsinger enrolled in the trial in June of 1999, soon after his 18th birthday, with the hope that his participation in the study would later help children suffering from OTC deficiency.[46] Testing upon his enrollment in the study and immediately prior to vector administration showed that his urea synthesis capacity was a mere 6-10% of normal (compared to the 26-100% capacity of the other study participants), was considered symptomatic for hyperammonemia, and he had a neutralizing antibody titer of 80, within the range of other study participants (range 20-1280, mean 151.11, median 20).[45, 47] On September 13, 1999, Gelsinger was administered 3.8x1013 particles (as per his weight and the dose per kg) of transgene-containing, replication-incompetent adenoviral vector via catheter into the right hepatic artery. He was the last participant to receive the virus, and the second to receive this dosage. Side effects of treatment among the previous study participants were headache, chills, fever, nausea and vomiting, back and muscle pain, thrombocytopenia, hypophosphatemia, anemia, and elevated transaminase levels; however, none of the side effects were life-threatening.[45, 47]

Eighteen hours after vector administration, Gelsinger developed jaundice and altered mental status, which had not been seen in the other participants. He became comatose as his serum ammonia levels progressively increased, peaking at 393μmol/dL, approximately ten times above normal. Gelsinger subsequently developed systemic inflammatory response syndrome (SIRS), leading to acute respiratory distress (ARDS, requiring extracorporeal membrane oxygenation, i.e. an artificial lung), disseminated intravascular coagulation, multiple organ system failure, edema, and irreversible neurological damage. The family decided to terminate life support, and Jesse Gelsinger died on September 17, 1999, 80 hours after onset of jaundice and 98 hours after vector infusion.[44-46]

With the family’s permission, an autopsy was performed, and it was determined that Gelsinger’s death was due to a fulminant innate immune reaction to the adenoviral vector, with acute inflammation and neutrophil infiltration in the lungs, tissue necrosis in the liver indicative of toxic insult, widespread necrosis of the spleen, and an increase in myeloid progenitors in the bone marrow with concomitant decrease in erythroid progenitors. Tubular necrosis was seen in the kidneys, although no deposition of immunoglobulin complexes was detected. Serum levels of the pro-inflammatory interleukins 10 and 16 were seen, and were the highest of the eighteen participants. Quantitative real-time PCR of tissue samples from multiple organs for both the therapeutic OTC construct and for the adenoviral E2a gene indicate that the vector had disseminated throughout the body, most likely after fully saturating the adenoviral receptors in the liver, which showed the highest levels of OTC and E2a DNA. The spleen, lymph nodes, and bone marrow showed the next highest levels of OTC and E2a DNA, indicating that the vectors selectively targeted organs rich in antigen presenting cells, and leading to a strong cytokine response to capsid proteins. The rapid onset of the immune response would indicate that there was minimal activation of the adaptive immune system, which generally takes days, rather than hours, to target an invading pathogen without previous exposure.[44, 45] It was noted that the other member of the high-dose cohort did not suffer a similar immunological reaction; however, there were several differences between the two participants. Although they were roughly the same age, Gelsinger was male, and the other participant was female; Gelsinger had a four-fold higher titer of neutralizing antibodies and activated T-cells against adenovirus serotype 5, upon which the vector was based; and Gelsinger’s OTC deficiency was much more severe.[45, 47]

An investigation into Gelsinger’s death by the United States Food & Drug Administration and the National Institutes of Health Recombinant DNA Advisory Committee discovered several deficiencies in the informed consent information, the study protocol, the ultimate application of the study protocol, reportage of side effects and protocol modifications to the NIH, FDA, and University of Pennsylvania institutional review board (IRB), and disclosure of conflicts of interest by the investigators.[48-50] As a result of their findings, the Committee recommended: the development of standards for vector potency, strength, quality, and toxicity; the establishment of a central database of vector safety and toxicity data; much more stringent requirements for pre-clinical data regarding toxicity, kinetics, receptor distribution and concentration; the inclusion of empty-vector controls in future studies, when possible; and improved informed consent procedures, to better clarify risks and benefits to participants.[44, 50] In addition, the National Bioethics Advisory Commission concluded that the study’s use of otherwise healthy patients, who would receive no therapeutic benefit from administration of the transgene, was ethically questionable, and noted that the principal investigator and the University of Pennsylvania improperly disclosed financial interests in the company that produced the vectors, indicating a potential conflict of interest.[51]

Pros and Cons of Direct and Indirect Gene Therapy

Direct administration of viral vectors to a patient, whether by systemic or by organ-specific introduction, is the most versatile method to provide a therapeutic transgene, as it can be used to deliver transgenes to both solid organs (e.g., the liver, heart, and CNS) and "liquid" organs (e.g., blood and, to an extent, the immune system). However, direct administration of viral vectors bears the risk of both innate and adaptive immune response to the vectors and their transgenic payloads; such a response, as seen with Jesse Gelsinger, is potentially fatal.[45] Even if the immune response to the vector is not life-threatening, it is capable of reducing or eliminating the therapeutic effect of the transgene, whether through the production of neutralizing antibodies against the vector, or the induction of NK and/or CD8+ T-lymphocytes to kill transduced cells.[31] However, a recent clinical trial for a gene therapy to treat lipoprotein lipase deficiency found success in mitigating immune response to transduced cells by pre-dosing with methylprednisolone, and administering the immunosuppressive drugs cyclosporine and mycophenolate mofetil for twelve weeks after vector injection.[52] Other gene therapy protocols may also benefit from similar short-term immunosuppressive therapy.

The use of indirect administration of therapeutic transgenes minimizes the chance of systemic immune response, as the process involves removal of target cells from the body and the introduction of vector in vitro. Transgene introduction is often followed by in vitro cell culture for a period of time prior to reinfusion, during which the intracellular pro-inflammatory response to the vector should fade.[28] This method is most useful for treatment of immune and blood disorders, as the affected cells are readily isolated from patients’ bone marrow and/or peripheral blood.[34] Indeed, this method was used to treat the first patient to successfully receive gene therapy, Ashanti DeSilva,[2] and has proved successful in other clinical trials. However, the use of indirect gene therapy to treat genetic disorders in solid organs, or diseases that affect multiple organ systems, is significantly more difficult. Researchers have considered the use of autologous mesenchymal stem cells (MSC),[53] autologous induced pluripotent stem cells (iPSC),[54] and allogeneic (or, possibly, "cloned") embryonic stem cells (ESC)[55] for ex vivo genetic modification and reimplantation. However, engraftment into the target organs, especially with the use of iPSCs and ESCs, remains a major hurdle,[34] although "pre-differentiating" the cells prior to implantation may improve the outcome.[56, 57]