Definition

The term gene therapy was previously used to describe replacement of a defective gene with a functional copy by means of gene transfer. The diseases originally targeted for gene therapy were classic, heritable genetic disorders. The term now broadly defines therapy involving the transfer of exogenous genes (complementary DNA [cDNA]) encoding therapeutic proteins into cells to treat disease.¹ The genetically altered cells are made into protein-producing “factories” churning out disease-altering gene products. Specifically, the host cell transcribes the exogenous gene (or transgene) into messenger RNA (mRNA); cytoplasmic ribosomes translate mRNA into the protein product. These products can affect not only the metabolism of cells from which they were made but also the metabolism of adjacent non–genetically altered cells via paracrine mechanisms (Fig. 9–1).

FIGURE 9–1 DNA encoding the gene of interest is constructed into a viral vector that is rendered incapable of replication. The vector is exposed to host cells, attaches to their surface, and is internalized. The released genetic information can either travel to the nucleus, where it may become integrated into the host genome, or remain episomal. It commandeers the normal protein-making machinery of the cell and produces large quantities of transgene.

Exogenous genes are produced and packaged in the laboratory in the following manner. First, cDNA of a gene of interest is constructed by the enzyme reverse transcriptase from mRNA. The cDNA is incorporated into a plasmid, a circular piece of DNA that is self-replicating and capable of delivering exogenous genes into cells, albeit at an inefficient rate. The cDNA plasmid is next integrated into a larger plasmid with a promoter sequence, assembling an expression plasmid. The promoter sequence initiates transcription of the gene of interest by target cells after gene transfer has occurred. The cytomegalovirus promoter is commonly used in gene transfer experiments. This promoter is constitutive, meaning it consistently initiates transcription throughout the life of the gene. The expression plasmid is integrated into either a viral or nonviral molecular vehicle that facilitates transfer of the exogenous gene to cells. These vectors are discussed in the following section.

There are two basic strategies for delivering exogenous genes to target cells. The first is the in vivo method, in which a gene-carrying vector is directly transferred to an intended population of cells within the host. The second approach, known as ex vivo gene therapy, involves removing target cells from the body, genetically altering them in vitro, and reimplanting them in the body (Fig. 9–2). Ex vivo methods are more complex and involve multiple, time-intensive steps. This approach is relatively safer, however, because the genetically altered cells may be observed for abnormal behavior before implantation. The ex vivo strategy allows the opportunity for in vitro selection of cells that express the gene of interest at high levels. Gene transfer via the in vivo method is technically simpler. There are relative advantages and disadvantages to both approaches that depend on the anatomy and physiology of the target organs, the pathophysiology of the disease being treated, the vector of choice, and safety considerations.²

FIGURE 9–2 A, In vivo gene therapy involves direct injection of vector-gene constructs into target tissues within the host. B, In ex vivo method, target cells are harvested from the host and then transduced, expanded, and propagated before reimplantation.

Overview of Vectors

Successful gene therapy generally depends on the efficient transfer of genes to target cells with subsequent expression. Generally, the duration of transgene production to treat disease successfully depends on the disease being targeted. Sustained expression is necessary for chronic conditions such as disc degeneration, whereas brief expression may be sufficient for acute conditions such as bone healing. With few exceptions, naked plasmid DNA is not taken up and expressed by cells effectively. Consequently, vectors are often necessary to package and insert genes into cells in such a way that the genetic information can be expressed. There are two broad categories of vectors, viral and nonviral. Gene delivery involving viral vectors is termed transduction, whereas transfer using nonviral vectors is termed transfection.

Nonviral Vectors

Nonviral vectors include liposomes, DNA-ligand complexes, gene guns, and microbubble-enhanced ultrasound. Liposomes are phospholipid vesicles that deliver genetic material into a cell by fusing with the cell membrane. Liposome vectors are simple, inexpensive, and safe, but drawbacks include transient expression of the transgene, cytotoxicity at higher concentrations, and low efficiency of transfection. DNA-ligand complexes and gene gun are nonpathogenic and relatively inexpensive to construct, but there is concern with lower transfer efficiencies and limited persistence of gene expression. Nishida and colleagues³ showed that ultrasound transfection with microbubbles significantly enhanced the transfection efficiency of plasmid DNA into the nucleus pulposus cells of rats in vivo, observing transgene expression up to 24 weeks. The overall transfection efficiency and level of gene expression of these nonviral vectors are generally inferior, however, to that of viral-mediated gene transfer. Consequently, most current studies involving gene therapy employ viral vectors.⁴

Viral Vectors

Viral vectors take advantage of the natural ability of viruses to infect and deliver genetic information efficiently to specific cell populations. The most commonly used viral vectors are derived from retroviruses, herpes simplex viruses (HSV), adenoviruses, and adeno-associated viruses (AAV). These viruses are often rendered incapable of replication before gene therapy application in an effort to make them less pathogenic. There are inherent merits and drawbacks associated with each viral vector, which are discussed in the following section. The choice of viral vector for gene transfer experiments is based on multiple considerations, including the gene to be delivered, the disease to be treated, and safety considerations.

Retroviruses are small RNA viruses that replicate their genomic RNA into double-stranded DNA (dsDNA) via the action of reverse transcriptase. The dsDNA is integrated into the host genome at a random location where it is able to express transgene for the life of the cell. Exogenous dsDNA is replicated by the transduced cell and passed on to all progeny cells during cell division. Gene delivery with retroviral vectors results in stable, long-term expression because the gene is integrated into the cell’s genome. Because the integration is at a random site, however, the risk of potential mutagenesis of oncogenes exists. Until more recently, this risk was considered only a theoretical possibility, but preliminary reports from a gene therapy trial involving retroviral vectors suggest one of the enrolled subjects developed leukemia as a result of oncogene mutation.⁵ Another disadvantage of the most commonly used retroviral vector, the murine leukemia virus (MLV), is that it infects and transduces only actively dividing cells. For these reasons, MLV is most suited for ex vivo applications. Lentiviruses, another class of retroviruses, are capable of infecting nondividing cells. Their drawbacks lie in their wild-type pathogenicity and complex genomic configuration, which make molecular processing for gene transfer much more intricate.

HSV vectors are dsDNA viruses, and the wild-type virus is a human pathogen that is trophic for sensory neurons. The replication-deficient vectors can infect dividing and nondividing cells of almost all types in vitro and in vivo. In addition, HSV vectors have the capacity to carry large amounts of exogenous DNA. This large carrying capacity allows for the production of HSV vectors that are capable of expressing multiple transgenes, which may be highly desirable for gene therapy applications involving complex disease pathophysiology, as in cancer or arthritis. HSV vectors do not integrate the genes they are carrying into the genome of the target cell. A disadvantage of HSV vectors is that they result in transient expression of the transgene despite efficient transduction. Consequently, these vectors would be insufficient for chronic conditions, such as intervertebral disc degeneration.

Adenoviruses are dsDNA viruses capable of infecting many cell types, including nondividing cells. There are 47 known human serotypes of adenoviruses, with serotypes 2 and 5 most commonly used for gene therapy studies. Wild-type adenoviral infections result in mild respiratory and gastrointestinal illnesses. The ability of adenoviral vectors to transfer genes to target cells is particularly efficient. Consequently, the adenovirus is an appealing option for in vivo gene delivery to quiescent, nondividing cell populations. The adenovirus genome exists as an episome within the nucleus of the infected cell and is not integrated into the genome of the host cell, so the risk of insertional mutagenesis does not exist.

The vectors are relatively easy to engineer in very high titers, in contrast to the HSV vector. A major disadvantage with the adenoviral vector is its short duration of transgene expression in most tissues. The transient expression of gene product is thought to occur because of low-level production of adenoviral antigens by the infected cell, resulting in an immune response directed against these cells. The episomal location of the vector genome is also thought to contribute to the short duration of expression. During cell division, the viral episome is not replicated and instead ultimately is degraded. Research is ongoing to engineer adenoviral vectors to minimize viral protein expression and consequently be less immunogenic.

AAV is a parvovirus with a 4.7-kb single-stranded DNA genome. Wild-type AAV lacks the viral machinery to self-replicate and can reproduce only in association with concomitant viral infection, usually adenovirus. AAV is also capable of infecting many different cell types, dividing and nondividing, but its level of infection efficiency is varied. The wild type is not known to cause disease.

The AAV vector differs from the adenoviral vector in several important ways. First, the AAV vector integrates reliably into a specific site on chromosome 19 in a nonpathogenic manner. Second, AAV does not provoke a significant immune response because the vector fails to express viral gene products after infection of target cells. Wild-type AAV has only two genes, Rep and Cap, which cannot be replicated without the presence of a helper virus. There is no expression of AAV gene products after transduction, theoretically leading to minimal host cell–mediated immune reaction. Although nearly 80% of the population has circulating antibodies against AAV2 (serotype 2) as a result of silent infections, titers of these neutralizing antibodies are usually low. For these reasons, sustained transgene expression can be achieved for 1 year in an immunocompetent host. The main shortcoming of AAV vectors is that they are capable of carrying only small amounts of foreign DNA. In addition, these vectors are difficult to construct and purify in the laboratory without helper virus contamination. There are multiple serotypes of AAV, but AAV2 has been most thoroughly studied for musculoskeletal applications, including degenerative disc disease.

There is a wide range of vector systems with different profiles for delivery efficiency, duration of expression, technical feasibility, and safety. Ongoing investigations are attempting to improve these profiles, and this research is likely to result in enhanced vectors with inducible promoters, tissue-specific promoters, or tissue-specific tropism. As mentioned, the appropriate vector system for a gene therapy application depends on multiple factors, including the method of delivery, pathophysiology of the disease targeted, and the gene selected for transfer.

Clinical Problem

Disorders of the spine often necessitate intervertebral fusion. Although internal fixation devices can successfully achieve temporary stabilization at practically all levels, long-term stability requires osseous consolidation. In contrast to fracture healing, spinal arthrodesis involves deposition of new bone in intersegmental locations that are not biologically structured for bone formation. Consequently, the nonunion rate is 40%^6,7 with single-level fusions and higher when multiple levels are attempted. Although instrumentation has improved the rate of bony union, pseudarthrosis remains a considerable clinical problem. Autogenous bone graft may be scarce in volume in cases such as pediatric fusions and revision surgery. It is associated with substantial donor site morbidity. Owing to these significant obstacles to clinical success, extensive research has been directed at developing molecular therapies to facilitate intersegmental fusion.

Augmentation of Spinal Fusion with Bone Morphogenetic Proteins

Bone morphogenetic proteins (BMPs) are a group of osteoinductive cytokines that play an essential role in the formation and maturation of osseous tissues. Urist and colleagues^8,9 originally recognized these proteins for their ability to form ectopic bone by inducing mesenchymal stem cell differentiation into chondrocytes and osteoblasts. Numerous subsequent animal studies have shown the ability of BMPs to enhance bone deposition at fusion sites.^10–19 This was followed by several clinical trials that validated these preclinical findings. In a study by Patel and colleagues,²⁰ patients undergoing posterolateral lumbar fusion augmented with iliac crest autograft and recombinant human BMP-7 (rhBMP-7) had better outcomes as measured by the Oswestry score and radiographic analysis than patients who received iliac crest autograft alone. In another human trial using rhBMP-2 in interbody fusion cages for single-level lumbar degenerative disc disease, patients who received cages filled with collagen sponge–delivered rhBMP-2 had superior clinical and radiographic results compared with control patients who received cages filled with autogenous bone graft.²¹ Various other clinical trials have similarly shown the efficacy of BMPs to enhance spinal arthrodesis.^22,23

Gene Therapy for Spinal Fusion

Gene therapy represents a potential next step in the evolution of therapies directed at promoting a spinal fusion. Gene therapy techniques to deliver various BMP genes could overcome the barriers of high dosing and complex carrier systems and achieve long-term, controllable BMP expression. The transduced cells would secrete the BMP extracellularly, delivering it to the environment at physiologically appropriate doses for a sustained period, maximizing the osteoinductive potential of these growth factors. In addition, BMP expression could be regulated in a temporal fashion by using vectors with inducible promoters, which would allow the ability to control the activity of the protein tightly to the clinical setting. Another potential advantage is the capacity to deliver gene therapy for spine arthrodesis in a minimally invasive procedure with percutaneous injections to the spine. Lieberman and colleagues²⁴ found an increase in the total volume of new bone with improved histologic quality when the gene for BMP-2 was delivered compared with rhBMP-2 protein alone.

Several studies have shown the feasibility of using gene therapy to enhance spinal fusion. Alden and colleagues²⁵ showed new enchondral bone formation in paraspinal muscles injected with adenoviral BMP-2 constructs (Ad-BMP-2). Important observations made by this study included the absence of bone deposition distant from the injection site and the absence of neural compromise, suggesting that this approach may be safe for the clinical setting. In addition, Helm and colleagues²⁶ documented that the direct injection of Ad-BMP-9 resulted in fusion in a rodent model without the development of nerve root compression or systemic side effects. In a rabbit model, Riew and associates²⁷ showed that mesenchymal cells transduced with BMP-2 can promote spinal arthrodesis.

Many of these studies used first-generation adenoviral vectors for gene delivery to immunocompromised animals, allowing for sustained expression. Gene therapy experiments with immunocompetent animals have led to a relative paucity of bone formation,²⁷ however, owing to the immune response elicited by adenoviral vectors. Further studies using second-generation adenoviral vectors or other vectors such as lentivirus could minimize these responses and maximize gene expression.

Another molecular avenue for bypassing the limitation of adenoviral immunogenicity is to deliver a gene for a factor that is “upstream” from the actions of BMP cytokines. In this way, neither efficient transduction nor sustained duration would be necessary because this factor would start a cascade of BMP activity after only a short period of expression. This intriguing strategy has been developed by Boden and colleagues,^28,29 who showed that a novel intracellular transcription factor, LIM mineralization protein-1 (LMP-1), could be used to upregulate the expression of BMPs and their receptors. LMP-1 initiates a cascade of events intracellularly, which stimulates the secretion of osteoinductive factors, which increases BMP activity. All of the study animals that were implanted with peripheral blood buffy coat cells genetically modified with Ad-LMP-1 showed successful lumbar fusion.²⁸ None of the 10 controlled rabbits had evidence of any bone formation, and the investigators concluded that local gene therapy could reliably induce spinal fusion in an immunocompetent animal.