Gene therapy — a novel form of drug delivery

MOLECULAR MEDICINE
cy. Meeting this challenge will require better methodsfor delivering genes, the development of animal models GENE THERAPY — A NOVEL FORM OF DRUG
of disease that adequately mimic diseases in humans, DELIVERY
and well-designed clinical trials to assess the safety andbiologic activity of these products in patients.
The development of novel vectors or vehicles for gene delivery warrants serious attention. Most gene therapyto date has involved the use of viruses as carriers of the GENE therapy is a novel form of drug delivery that gene. The carrier virus may be a retrovirus that lacks enlists the synthetic machinery of the patient’s virtually all viral genes except those required for in- cells to produce a therapeutic agent. Using the body to fecting mammalian cells. Indeed, if properly prepared, treat its own disease overcomes the need to manufac- these viruses are so defective that after they infect the ture highly purified proteins. It also eliminates the need appropriate target cell, they cannot replicate or infect for repeated parenteral administration of proteins (as in other cells. The use of retroviruses is also advanta- hemophilia) or drugs (as in hereditary hypercholester- geous because they are easily generated, the infected olemia) and reduces the difficulties of complying with viruses can be extensively characterized in tissue cul- exogenous-drug regimens. Applications of gene thera- ture before being injected into patients, and the stable py are not limited to rare inherited diseases but extend integration of the virus into the chromosome ensures potentially to common acquired disorders, including its retention by the cell. The risk that retroviral inte- cancer, heart disease, and the acquired immunodefi- gration into the cellular genome could inactivate host ciency syndrome. Gene therapy is thus likely to have tumor-suppressor genes or activate proto-oncogenes broad implications for the future practice of medicine.
appears to be minimized by reducing the efficiency of This introductory article presents an overview of infection so that at most one virus infects a cell. As gene therapy. Succeeding articles will discuss specific vectors, retroviruses have two major limitations. First, advances and the obstacles to various kinds of gene they can harbor genes of only limited size, approxi- therapy for hereditary and acquired diseases. mately 7 kb; although most therapeutic genes of inter- The targets of gene therapy are either intracellular or est can be accommodated, some are too large. Second, extracellular (Fig. 1). In the first example shown in Fig- retroviruses infect only dividing cells, whereas most ure 1, the gene delivered to the cell affects the cell itself tissues in adults consist primarily of nondividing cells.
by replacing a defective or missing gene (Fig. 1A) or by For this reason, retroviruses are generally used for ex providing a product that kills or inhibits the growth of vivo gene delivery — i.e., the injection of cells that cells (Fig. 1B). In the second example, the cell releases were genetically engineered by infection outside the the product of the new gene, which can act locally on body, rather than by in vivo infection with the retrovi- neighboring cells (Fig. 1C) or enter the circulation for delivery to distant cells (Fig. 1D).
Adenoviruses are t he second most commonly used Although easily comprehended in principle, gene viral vector for gene delivery. These viruses have the therapy is not so simple in practice. While gene therapy same size constraints as retroviruses, but they have cer- is likely to become a major part of the medicine of the tain advantages. Adenoviruses readily infect nondivid- future, it is by no means a component of the clinician’s ing cells and can be produced in large numbers, making armamentarium at present. Clinical experience with infection of tissues in an adult more efficient. Moreover, gene therapy over the past five years has shown that in these viruses remain extrachromosomal, thereby reduc- most cases, toxicity is not a serious problem. As with ing the chance of disrupting the cellular genome. A lim- recombinant DNA, the birth of gene therapy appeared itation of adenoviruses is that if delivered in vivo they to open a Pandora’s box of potential complications and infect all tissues, including the germ line, and thus possible toxic effects. However, most of the early wor- could affect subsequent generations. Another problem ries now appear unwarranted. Indeed, it has been sug- is that they are not as defective as retroviruses and gested that clinical tests of gene therapies may no long- could more readily yield infectious virus in the body.
er require special consideration by the Recombinant A third problem is that the current generation of ad- DNA Advisory Committee to be performed, but should enoviral vectors is immunogenic, which reduces the require the same rigorous evaluation as other novel length of time available for the expression of the gene medical treatments. Exceptional review would thus be and makes repeated administration of the vector im- limited to unprecedented applications or issues — for example, the use of new viral vectors, gene therapy for There are also nonviral methods of gene delivery.
fetuses, or approaches that involve altering the germ Liposomes, or lipid vesicles, which combine readily line. Currently, the main challenge is achieving effica- with cell membranes, are being used to deliver genes,sometimes as aerosols. Liposomes are also being testedin conjunction with viruses to enhance localized deliv- From the Department of Molecular Pharmacology, Stanford University School ery of genes. Direct injection of naked DNA plasmids of Medicine, Stanford, CA 94305-5332, where reprint requests should be ad-dressed to Dr. Blau.
is possible, but for unknown reasons this method ap- Downloaded from nejm.org on August 24, 2012. For personal use only. No other uses without permission. Copyright 1995 Massachusetts Medical Society. All rights reserved. Example 1:
Gene affects
Product kills or inhibits the growth of cells Sample application: CFTR gene is delivered Sample application: Tumor cells receive tumor necrosis factor or TK gene followed by Example 2:
Gene affects
neighboring
or distal tissue
Sample application: The implantation of cells or injection of DNA leads to the production of genetically altered myoblasts or injection of growth factors that stimulate angiogenesis.
plasmid DNA results in muscle that secretes Figure 1. Four Types of Gene Therapy.
CFTR denotes cystic fibrosis transmembrane conductance regulator, and TK thymidine kinase.
pears to function primarily in heart and skeletal mus- therapy, cells express the genes continuously. As a re- cle. The disadvantage of this method is that relatively sult, the production of the therapeutic protein cannot few cells take up the DNA (1 to 3 percent), and thus be modulated. In diseases such as hemophilia regula- only small amounts of the encoded protein are pro- tion of the new gene is not critical, but such diseases duced. Therefore, the primary and most important cur- are the exception. In most other diseases, gene regula- rent use of DNA plasmids as gene vectors is in vaccine tion is desirable; indeed, constitutive expression of the development, since the small amount of protein pro- introduced gene may be detrimental or even life-threat- duced by the few cells that take up the plasmid can elic- ening. To overcome this problem, yeast-gene or bacteri- al-gene regulatory systems have been adapted for use in An important aspect of gene-delivery systems is the mammalian cells. These inducible systems appear ad- ability to regulate the expression of the introduced vantageous because they affect the expression of intro- gene. With the vectors that are now approved for gene duced but not resident genes. There is no toxicity, and Downloaded from nejm.org on August 24, 2012. For personal use only. No other uses without permission. Copyright 1995 Massachusetts Medical Society. All rights reserved. the gene-inducing agent can be administered orally. For a number of existing viral and plasmid vectors. Some example, it is possible to regulate the expression of a desirable properties might include the ability to incor- gene of interest, such as one encoding a hormone, by porate large genes, the absence of immunogenicity, the means of hybrid microbial and mammalian proteins potential to direct the vector to certain cell populations and microbial DNA regulatory elements that respond — possibly by incorporating elements that recognize to either the antibiotic tetracycline or the progesterone specific membrane components and enhance uptake — antagonist mifepristone (RU 486). In principle, these the ability to incorporate elements to limit the expres- new regulatory systems allow genes to be turned on and sion of the delivered genes to particular types of cells, off and the level of the therapeutic protein varied over and the ability to turn on or off and modulate the levels time. The complex goal of regulating genes through the of gene expression in response to either nonmammali- use of endogenous biologic signals, such as glucose lev- an exogenous regulators or endogenous signals such as els in the case of diabetes, is also being pursued but will Animal models of human genetic diseases have been Perhaps the ideal vector will be completely synthetic, invaluable for testing and comparing different vectors a composite of DNA-sequence components derived from for gene delivery. They provide a convenient means of Table 1. Gene-Therapy Protocols That Have Been Approved for Clinical Trials by the Recombinant DNA Advisory Inherited diseases*
Cystic fibrosis transmembrane conductance regulator ficiency due to adenosine deaminase deficiency Acquired diseases
HIV-cleaving ribozyme; antisense (complementary RNA sequence) to HIV TAR (transactivation-response element) along with transdominant nega-tive Rev (regulator of HIV gene expression)† Herpesvirus thymidine kinase followed by ganci- clovir treatment (toxic only to cells expressing herpesvirus thymidine kinase) Antisense RNA to c-fos and c-myc RNA Antisense RNA to insulin-like growth factor-1 RNA Multidrug resistance (MDR-1) for chemoprotection B7 cofactor to induce T-cell costimulation Cytokines to induce immune response to tumor (in- terleukin-2, 4, and 7; granulocyte–macrophage Interferon-g to induce immune response to tumor *The approach used is shown in Figure 1A.
†HIV denotes the human immunodeficiency virus.
‡The approach used is shown in Figure 1C.
§The approach used is shown in Figure 1B.
Downloaded from nejm.org on August 24, 2012. For personal use only. No other uses without permission. Copyright 1995 Massachusetts Medical Society. All rights reserved. evaluating in vivo the plasma and tissue concentrations, problem is to test gene therapies in well-defined ani- time course, and distribution of the expressed proteins.
mal models of cancer in which specific mechanisms However, mouse models created by disrupting the re- can be established — for example, those in which ei- sponsible gene often do not manifest the disease in the ther an oncogene is introduced or a tumor-suppressor same way as humans with the equivalent genetic defect.
For example, the mdx mouse model of Duchenne’s mus- A number of clinical trials of gene therapy have been cular dystrophy harbors the same defective gene as the completed or are under way. They can provide informa- patient but has only transient muscle weakness. Similar- tion that cannot be gleaned from preclinical tests in an- ly, the mouse model of cystic fibrosis does not have life- imals. The clinical trials listed in Table 1 were ap- threatening pulmonary disease, although there is some proved by the Recombinant DNA Advisory Committee accumulation of mucus in the lung. The mouse model of and are designed to meet the major goals outlined in the Lesch–Nyhan syndrome does not have the charac- Figure 1. In the ex vivo approach, cells are genetically teristic profound neurologic symptoms, including self- engineered and characterized in tissue culture before mutilating behavior, or gout, but has reduced dopamine being implanted in the body. In the in vivo approach, levels. However, these models manifest at least some of the gene is delivered and expressed in situ within the the properties of the human diseases, and an under- body. A current limitation of clinical trials is that very standing of how they circumvent the problems found in few viral vectors have been approved for use in hu- their human counterparts may be informative. Nonethe- mans. Thus far the trials have not shown convincingly less, animal models that more closely mimic human dis- that gene therapy is effective in treating disease in hu- ease would be invaluable for preclinical tests of efficacy.
mans. For this reason, Dr. Harold Varmus, Director of Mouse models of acquired human diseases such as the National Institutes of Health, has initiated a major cancer also have their limitations. The cancers in ani- evaluation of the field to ensure that future efforts will mals that are used to test novel therapies are frequent- ly transplantable tumors that grow rapidly in the host, Despite the tremendous attraction of and interest in with nearly 100 percent of the cells undergoing division gene therapy, the field is still in its infancy. Quick cures — a situation quite atypical of most human tumors.
for genetic or acquired diseases are not to be expected. These animal tumors grow readily in culture and quickly develop into cancers in vivo with a highly re- ECOMMENDED
producible phenotype. By contrast, genetic models of Friedmann T. A brief history of gene therapy. Nat Genet 1992;2:93-8.
cancer in mice that resemble the disease in humans Gossen M, Freundlieb S, Bender G, Müller G, Hillen W, Bujard H.
Transcriptional activation by tetracyclines in mammalian cells. Sci- more closely pose difficulties, because to assess the ef- ficacy of treatment in them often requires a very large Mulligan RC. The basic science of gene therapy. Science 1993;260: sample and long-term studies. The solution to this Downloaded from nejm.org on August 24, 2012. For personal use only. No other uses without permission. Copyright 1995 Massachusetts Medical Society. All rights reserved.

Source: http://www.bioscientist.us/news/files/public/1345795743_74_FT2335_nejm199511023331808.pdf