pp. 149-154 in Intractable Neurological Disorders, Human Genome Research and Society. Proceedings of the Third International Bioethics Seminar in Fukui, 19-21 November, 1993.

Editors: Norio Fujiki, M.D. & Darryl R.J. Macer, Ph.D.

Copyright 1994, Eubios Ethics Institute All commercial rights reserved. This publication may be reproduced for limited educational or academic use, however please enquire with Eubios Ethics Institute.

Development of new strategies for human gene therapy

Mahito Nakanishi
Institute for Molecular and Cellular Biology, Osaka University, Japan

1. Introduction

Since the techniques of molecular biology were introduced into medical research, understanding of the molecular basis of many human diseases has been facilitated dramatically, and it is against this background that the concept of gene therapy was born. As the molecular basis of many diseases have become clear, supplementation of the deficient proteins by delivering and expressing the genes in tissue cells has been recognized as an important therapeutic approach (1-4).

As early as 1980, a scientist in University of California at Los Angeles performed the first gene transfer experiment by introducing a human globin gene into the blood cells of a patient with thalassemia (a disease caused by a deficient globin gene). He undertook the trial, however, without any formal approval by authorized committees in the USA and became the target of criticism. Motivated partly by this incident, the scientific community began to discuss the rules to be followed in gene therapy. In 1985 the National Institute of Health (NIH, USA) made guidelines for gene therapy which all the protocols submitted to NIH for approval since then have had to fulfil this guideline. NIH has revised there guidelines periodically and most other countries planning gene therapy (including Japan) have made their own guidelines based on this one. Of course the first approved full scale gene therapy in 1990 of two child patients with adenosine deaminase (ADA)-deficient severe combined immunodeficiency (SCID) followed this guideline.

The realisation of gene therapy has also depended to a large extent on the improvement of technology to introduce and stably express foreign genetic information in animal cells (5). Namely gene therapy might not have become a realistic proposition without the development of retrovirus vectors. More recently, many other viral and non-viral methods to transfer genetic materials have been proposed for use in gene therapy. Along with this progress, the classical narrow definition of gene therapy as the treatment of genetic disease has been replaced with a broader definition, as the transfer of new genetic material to the cells of an individual, followed by its expression, with resulting therapeutic benefit to the individual (4). Now gene therapy is in the process of becoming one of the important approaches for treatment of cancer and infectious disease.

What kind of cells then are targets for the transfer of a foreign gene? We can in principle divide gene therapy into two types, somatic cell and germ line. The human body consists of these two kinds of cells (somatic cells and germinal cells). Germinal cells (sperm, egg and their precursor cells) can transmit their genetic information to the next generation. All the cells other than germinal cells are somatic cells, including fibroblasts, epithelial cells, muscle cells, neuronal cells and many kinds of tissue cells. These cells are differentiated to have a special function in the body and cannot transmit their genetic information to a descendant. Gene therapy targeted to somatic cells seems not to be harmful to the genetic information of a descendant because of the cell's character. In this view, we can consider somatic gene therapy as a modification of the protein replacement therapy that hitherto has been permitted without any ethical problems. This notion has been widely accepted among experts since the NIH made their guidelines in accordance with it in 1985.

On the other hand, gene therapy targeted at germinal cells has not been accepted yet and is unlikely to be accepted in the future. This is mainly due to ethical and religious reasons we should not create the genetic information to modify the "quality" of human beings permanently. I also wish to point out that the term "therapy" should be used when we treat a patient with some diagnosis (a fertilized egg is not a patient), which presupposes their existence, hence the ban is a matter of course. There are some other difficult problems concerning the gene therapy on an unborn baby. It is probably too early to discuss this possibility, however, because the technology for introduction of genetic information is immature.

To discuss safety and ethical problems of gene therapy, it is important to know the current system for gene therapy well, and be familiar with techniques for introducing genes. In this article I wish to focus on ethical and safety issues and what kind of techniques can be developed to deal with them, after discussing general aspects of gene therapy.

2. Candidate diseases for gene therapy

Firstly, the classical genetic diseases, caused by a single gene defect, are important candidates for gene therapy in its narrow definition (Table 1). The number of genetic diseases for which the responsible genes have been isolated and characterized is continuously growing. We may expect that most of the genes responsible for major genetic diseases (i.e. those with many sufferers and severe symptoms) will be characterized before the end of this century. Among these diseases, those which respond poorly to conventional treatment such as drugs, restrictions and enzyme or hormone therapy replacement are generally considered to be primary candidates. In actual gene therapy of these diseases, however, there is no need to insist on expressing the normal version of the defective gene in all the tissues in which it is originally active. Rather, more flexible alternative approaches, such as expressing an enzyme that bypasses the metabolism of the accumulated harmful material and leads to its excretion, may result in more success.

Diseases that respond to the conventional treatments to some degree will also in turn become candidates in the future as technology develops. When we can transfer and express genetic information in tissue cells stably, the treatment will improve the quality of life of patients. Some clinical researchers propose that the these diseases may in fact be the first choice for application until the method of gene therapy is better established, because a patient can be treated with backup therapy should gene therapy not work. The advantages of gene therapy compared to replacement therapy are that the protein produced directly in tissues contains no hazardous contaminants such as pathogenic microorganisms or denatured proteins and that we can also supplement the intracellular proteins such as enzymes, receptors and transcription factors.

Another candidate disease for gene therapy is cancer, an example of diseases due also to acquired genetic mutations. More than half of the protocols approved by NIH have been aimed at treatment of cancer. Recently the molecular basis of cancer has gradually been becoming clearer. Cancer is now known to be caused by genetic alterations such as a mutation or the unregulated expression of "normal" oncogenes and defects in tumour suppressor genes. Currently, a number of researchers are focusing their effort on making extra cytokines or histocompatible antigen in cancer cells or in tumour-infiltrating lymphocytes to enhance natural immunity against cancer. This approach is much more realistic as a cancer therapy than the direct approach to correct a genetic alteration by introduction of a normal gene.

Gene therapy is also proposed for treatment of severe infectious diseases such as acquired immunodeficiency syndrome (AIDS). It is clear, however, that for infectious diseases where the pathology of infection is known the prevention of transmission of pathogens has priority over complex and experimental therapies in infectious diseases.

3. Process for approval of a gene therapy protocol and public access to information

One of the important points in discussing the ethical issues of gene therapy is how and on what criteria a protocol for gene therapy is examined before approval. Gene therapy still has a very experimental hue and its approval involves a multi-layered checking machinery. I want to use cases from the USA as an example. The protocol for any gene therapy project is first examined by the Institutional Review Board (which considers it as a treatment) and by the Biosafety Committee (which considers it as a recombinant DNA experiment) of the organisation to which the principal investigator belongs. After approval at the institutional level, the protocol is submitted to the Recombinant Advisory Committee (RAC) of the NIH and is examined further. Finally, it is submitted to the independent Food and Drug Administration (FDA), where the safety and quality control of the gene transfer vector as a medicine is examined. If it passes all these the therapy becomes officially possible. The details of a protocol submitted to the NIH and discussion that took place in the RAC are published in the journal named Human Gene Therapy, that was first issued in 1990. The technical meetings of the RAC are also open to the public.

The RAC examines the protocol on several points, including its necessity, effectiveness and safety. The first point is whether the gene therapy proposed is necessary to cure the patients. If there is an effective alternative clinical choice, the RAC does not consider the proposed protocol necessary. For example, there was a controversy in the RAC over the first proposal of gene therapy for ADA-deficiency patients, because the condition of these patients might be improved by supplementation with polyethyleneglycol(PEG)-modified bovine ADA protein. In this case, the proposal was finally approved for those patients who had lost response to PEG-ADA therapy, because data was submitted showing extension of the life of cells that had become positive to ADA after gene insertion.

To prove the effectiveness of gene therapy, it is essential to have enough data from animal experiments. Hence, development of an animal model for the target disease by gene knockout is important in assessing the effectiveness of the gene therapy. (Safety issues are discussed below).

One of the current problems is availability of information about the actual results of the approved gene therapy trials. Some investigators refuse to open the clinical results to the public until they have published their results formally in a scientific journal. This makes the approval process incomplete.

4. Technology currently under development for transfer of genetic information and related problems

Under the present state of affairs where no method for targeting tissues has yet been established, there are two ways to introduce genes into target cells. That is the therapeutic genes can be introduced into target tissues either by the "ex vivo" or "in vivo" method. In the ex vivo method, we first remove either a part of the tissues or some cells from a patient, then culture them in an incubator under sterile conditions. While culturing, we can transfer genetic information by any of various methods. After culturing for some more time, we transplant the tissues into the patient. The ex vivo method is especially essential when using retrovirus vectors, because they are extremely inefficient at introducing their genome into tissue cells directly. This approach can be thought of as a technical modification of blood transfusion and tissue transplantation and has been applied successfully to many kinds of cells including lymphocytes, fibroblasts, keratinocytes, hepatocytes, endothelial cells, muscle cells and bone marrow cells.

In contrast, in the in vivo method the vectors carrying the desired gene are administered directly to tissues in the patient's body mainly by physical targeting. In vivo methods will be preferred in the future because the ex vivo method requires advanced technical skills and has high costs. The former is also less risky to patients. This approach has been applied to hepatocytes, and to airway epithelial cells using aerosol inhalation.

Introduction of genetic information into target tissues can be mediated using recombinant viruses by viral vectors or by nonviral, e.g. chemical or physical, methods (the term "vector" is used to refer to a vehicle carrying a gene). The viral vectors currently used for gene therapy are retrovirus vectors and adenovirus vectors. The most practical nonviral method is lipofection. These vectors each have their own advantages and disadvantages.

Retrovirus vector can integrate their genome into a host cell genome after infecting a cell, and the integrated vector DNA behaves just like a part of the host genome. The genetic information introduced is therefore very stable, and the viruses themselves are modified so that they cannot replicate and produce more viral particles. Retrovirus vectors are the vectors used most widely in gene therapy research and have been improved in efficiency and safety. As retrovirus vectors cannot transfer genetic information into nondividing cells, and cannot be concentrated, they are used in the ex vivo method. Their effectiveness in the in vivo method remains unclear.

Adenovirus vectors are constructed by deleting the early gene of the virus thus preventing replication and replacing it with the foreign gene to be introduced. Because the early gene is essential for virus replication, adenovirus vectors are thought not to be replicable. The expression of a gene transferred by an adenovirus vector is more efficient than that of retrovirus vectors, but a major drawback is that expression is transient. Adenovirus vectors are also highly immunogenic, so they cannot be administered repeatedly.

Lipofection uses an electrostatically bonded complex of positively charged lipids and negatively charged DNA as a vector. The complex fuses with the cell membrane and delivers DNA into the cytoplasm. Generally the efficiency of transfer of genetic information by lipofection is lower than by viral vectors. Lipofection has another disadvantage in that the expression of the gene is transient. However the complex of lipid and DNA is more stable and hence easier to handle and also is safer and more easily subject to quality control when compared to viral vectors.

For every therapy the balance of effectiveness and side effects is a problem and we assess the effectiveness of the therapy by examining whether the merit may overcome the risk. When using virus vectors for gene therapy, we need to examine the possible risks even more cautiously than for non-viral methods, because they can be considered as a kind of live vaccine. In early studies, contamination by a wild type retrovirus of a vector preparation caused an extremely high incidence of malignant tumours (leukaemias) in monkeys. Since then, the packaging cells (the cells producing recombinant retrovirus) have been improved so as not to produce wild-type virus, resulting in safer vector preparations so contamination has ceased to be such a problem. Adenovirus vectors also have the limitation of toxicity at high multiplicity of infection. Adenovirus vectors have been reported to cause severe pneumonia when administered directly into the air way.

Another potential risk of gene therapy using retroviruses is the possibility that DNA integrated into the host genome at random may effect the host's gene expression. For example, the strong promotor element in a retrovirus vector (the part that governs transcription) may activate a "silent oncogene" abnormally if integrated into a site near this gene. This is not merely an imagined problem as examples of abnormal activation of oncogenes by recombination have been found in several malignant lymphomas. There is a greater chance of "hitting" the potential oncogene as the number of target cells increases. The size of the DNA of a human cell is approximately 6 billion base pairs. So if we transfer DNA into 6 billion human cells (about 60 grams) and if DNA integrates into the genome of these cells completely at random, we can expect the transferred DNA to integrate at any point in the human genome, and the risk increases. If all the cells in the liver (0.3 trillion cells) received DNA, the risk would increase a further 50 fold. Although the retrovirus vectors are generally considered to be safe in practice at this point in time, there is still no clear evidence over this question.

5. Development of new strategies for human gene therapy

As described above, current methods for the transfer of genetic information, both viral or nonviral, although they all have their own strengths and weaknesses, still have a number of problem points before we can have ideal gene therapy. In my opinion, such an ideal method for gene transfer should fulfil the following criteria:
a) Efficiency of introduction of genetic information is sufficiently high with every kind of cell and furthermore, without toxicity.
b) Gene expression is stable and lifelong if possible.
c) Expression of the introduced gene does not affect that of the host genome and is conversely not affected by the gene expression of host cells.
d) Genetic materials should be present as an independent replicon separate from the host cell genome to avoid affecting the host genome.
To develop techniques satisfying these criteria, we need to investigate new strategies by first examining the principles of gene transfer.

Table 1: Examples of genetic diseases that are potential gene therapy candidates

Figure 1: Scheme of gene introduction mediated by a fusogenic liposome

For efficient introduction of genetic materials (DNA and RNA) into living animal cells, first we need to deliver these molecules across the cell membrane. Our approach is based on our finding that liposomes (small bilayer lipid bubbles) deliver their contents to the cytoplasm directly and efficiently after incubation with Sendai virus at 37íC (6). A scheme depicting this phenomenon is shown in Figure 1. At first, simple liposomes consisting of phospholipid and cholesterol, and containing the macromolecule (proteins or DNA) inside, fuse with Sendai virus in a receptor-independent manner. (The single-strand virus genomic RNA is inactivated in advance by ultra-violet light (UV) to prevent virus gene expression.) Next the resultant complex (fusogenic liposome), with virus envelope proteins outside, fuses with the cell membrane in a receptor-dependent manner in a similar way to infection with native virus particles.

Using these fusogenic liposomes, we were able to introduce directly and express E.coli beta-galactosidase gene and human hepatitis B virus surface antigen gene in a living rat liver (7). More recently, we have purified and characterized unilamella fusogenic liposomes in detail (Nakanishi et al. manuscript in preparation). Examination of purified fusogenic liposomes by electron microscopy and by dynamic laser light scattering showed that fusogenic liposomes had an unilamella structure with virus spike outside and an average diameter of 379 nm (Figure 2) and the cavity originated from the liposome inside. Suspensions of fusogenic liposomes contained 4 x 109 fusogenic particles per ml and could transform 4 x 105 human cells with a drug resistant gene. The suspension could be stably frozen in liquid nitrogen indefinitely without loss of function. This data clearly showed that unilamella fusogenic liposomes were extremely effective and well characterized tools for gene transfer into tissue cells of living animals.

If we can devise a suitable form for the encapsulated genes, the fusogenic liposomes have great potential to become ideal vehicles for gene therapy because they can encapsulate any type of nucleic acid and deliver it directly into cells. As an ideal vector to be encapsulated in fusogenic liposomes, and to function as an independent replicon in human cells, we have started a project to develop human artificial chromosomes.

The findings described above also encourages us to develop a new type of virus vector using the gene expression system of Sendai virus. Sendai virus has a number of interesting characteristics making it suitable as a gene expression vector. These are: 1) transcribed mRNAs are monocistronic (only one protein is encoded); 2) the virus replicates and transcribes mRNA in the cytoplasm without affecting the gene expression of the host cell; 3) the virus is not pathogenic to humans; 4) the complete sequence of the genomic RNA has already been determined; 5) the negative strand genomic RNA does not cause interviral recombination; 6) there is a mutant virus strain that infects persistently and can coexist stably without interrupting host cell function. Recently we have established the mechanism of persistent infection, a characteristic most important for development of virus vectors as tools for stable gene expression (8). On the basis of these findings, we are investigating the possibility of developing new RNA vectors as independent replicons.

We believe we can develop a new ideal vector system through basic research on molecular and cellular biology and the collaboration of scientists in various research areas, including pharmacologists, chemists and physicians.

Figure 2: Structure of an unilamella liposome (A), an unilamella fusogenic liposome (B) and a Sendai virus particle (C).

1. Friedmann, T. (1989) Progress toward human gene therapy. Science 244: 1275-1281.
2. Anderson, W.F. (1992) Human gene therapy. Science 256: 808-813.
3. Miller, A.D. (1992) Human gene therapy comes of age. Nature 357: 455-460.
4. Morgan, R.A. & Anderson, W.F.(1993) Human gene therapy. Ann. Rev. Biochem. 62: 191-217.
5. Mulligan, R.C. (1993) The basic science of gene therapy. Science 260: 926-932.
6. Nakanishi, M. et al. (1993) Efficient introduction of contents of liposomes into cells using HVJ (Sendai virus).Exp. Cell Res. 159: 926-932.
7. Kato, K. et al. (1991) Expression of Hepatitis B Virus Surface Antigen in Adult Rat Liver: Co-introduction of DNA and Nuclear Protein by a Simplified Liposome Method. J. Biol. Chem. 263: 8929-8937.
8. Kondo, T. et al. (1993) Temperature-sensitive phenotype of a mutant Sendai virus strain is caused by its insufficient accumulation of the M protein. J. Biol. Chem. 268: 21924-21930.

To next chapter
To contents list
To book list
To Eubios Ethics Institute home page