Human Gene Therapy

Vectors derived from the human parvovirus AAV-2 (adeno-associated virus type 2) are among the most promising gene delivery vehicles currently being developed. These vectors are not only capable of transducing a large variety of human cell types in vitro and in vivo, but in immunocompetent animal models can establish longterm gene expression without being pathogenic to the recipient. However, a limitation of this vector system with respect to its clinical application has long been the laborious work needed to prepare high-titer and pure AAV-2 vector stocks. A number of improvements to the basic manufacturing protocol have recently been reported that now allow the production of AAV-2 vectors of significantly higher quality and quantity. This article considers the most relevant approaches taken so far, which include modifications to the conventional transfection/infection protocol as well as the development of helper virus-free packaging methods and the establishment of vector producer cell lines. The various novel protocols are discussed, including their advantages and drawbacks, with a particular focus being put on their prospects for clinical use. Despite these advancements, the development of an ideal AAV-2 vector production method fully suiting clinical requirements obviously remains a challenging issue.

The availability of efficient and nontoxic gene delivery technologies is fundamental to the translation of therapeutic concepts into clinical practice by gene transfer. High-capacity adenoviral (HC-Ad) vectors are characterized by the ability to transduce cells in vitro and in vivo with more than 30 kb of nonviral DNA. This quality allows simultaneous gene transfer of several expression cassettes, large promoters, and some genes in their natural genomic context. Because all viral coding sequences are removed from these vectors, safety is considerably improved compared with previous-generation adenoviral vectors.

DNA degradation is a fundamental problem for any gene therapy or genetic immunization approach, since destruction of incoming genes translates into loss of gene expression. To characterize the biology of DNA degradation after naked DNA injection, the location and levels of tissue nucleases were assessed. Extracts from the serum, kidney, and liver of mice had high levels of calcium-dependent endonuclease activity. High levels of acidic endonuclease activity were identified in the spleen, liver, kidney, and skin with little activity in skeletal or cardiac muscle. Relatively little exonuclease activity was observed in any tissue. The presence of endonucleases in the skin and muscle mediated degradation of 99% of naked DNA within 90 min of injection. This degradation most likely occurred in the extracellular space upstream of other cellular events. Despite this massive destruction, gross tissue nuclease levels did not determine skin-to-muscle transfection efficiency, or site-to-site transfection efficiency in the skin. While gross tissue nuclease levels do not appear to determine differences in transfection efficiency, the presence of robust tissue nuclease activity still necessitates that massive amounts of DNA be used to overcome the loss of 99% of expressible DNA. In addition to destroying genes, the nucleases may play a second role in genetic immunization by converting large plasmids into small oligonucleotides that can be taken up more easily by immune cells to stimulate CpG-dependent Th1 immune responses. For genetic immunization, vaccine outcome may depend on striking the right balance of nuclease effects to allow survival of sufficient DNA to express the antigen, while concomitantly generating sufficient amounts of immunostimulatory DNA fragments to drive Th1 booster effects. For gene therapy, all nuclease effects would appear to be negative, since these enzymes destroy gene expression while also stimulating cellular immune responses against transgene-modified host cells.

To improve gene transfer to CNS neurons, critical elements of herpes simplex virus 1 (HSV-1) amplicons and recombinant adeno-associated virus (AAV) vectors were combined to construct a hybrid amplicon vector, and then packaged via a helper virus-free system. We tested the HSV/AAV hybrid amplicon vectors for transduction efficiency and stability of transgene expression (green fluorescent protein) in primary neuronal cultures from rat fetal ventral mesencephalon, in comparison with traditional HSV amplicon, AAV, or adenovirus (Ad) vectors at the same multiplicity of infection. The HSA/AAV hybrid vectors transduced the highest number of primary neurons in culture 2 days after infection. As compared with all other vectors tested, only hybrid vectors containing the AAV rep gene maintained the 2-day level of transgene expression over 12 days in culture. This rep-containing hybrid vector was then tested for efficiency and safety in the brain. One month after injection into adult rat striatum (1 X 10⁶ transducing units injected), transgene expression was observed within the striatum (ranging from 564 to 8610 cells) and the substantia nigra (via retrograde transport, ranging from 130 to 809 neurons). The HSV/AAV hybrid amplicon vectors transduced predominantly neurons within the striatum, and showed transduction efficacy similar to and in many cases higher than that of HSV amplicon vectors. No immune response was observed in the HSA/AAV hybrid vector-injected brains, as determined by immune markers specific for helper T lymphocytes, cytotoxic T lymphocytes, and microglia. This HSV/AAV hybrid system shows high transduction efficiency and stability in culture. The effective and safe transgene delivery into the nigrostriatal system illustrates its potential for therapeutic application for neurologic disorders, such as Parkinson and Huntington disease.

A novel fusion gene consisting of the open reading frame of a double-mutant (Phe22-Ser31) dihydrofolate reductase (dmDHFR) cDNA fused to the open reading frame of cytidine deaminase (CD) was constructed and characterized for the purpose of conferring simultaneous resistance to methotrexate (MTX) and cytosine arabinoside (ara-C). The kinetic properties of purified recombinant dmDHFR-CD fusion protein were compared with those of purified CD and dmDHFR. The fusion protein was found to retain enzymatic properties of both dmDHFR and CD, in that the K_m and K_cat values of purified dmDHFR-CD protein were found to be virtually identical to those of CD and dmDHFR alone. Retrovirus-mediated expression of dmDHFR-CD in NIH 3T3 cells conferred significant resistance (10- to 12-fold) against MTX and ara-C, compared with mock- and single gene-infected cells and the level of resistance obtained was similar to that of cells expressing both CD and dmDHFR from a retroviral bicistronic vector. Infection of mouse bone marrow cells with the dmDHFR-CD construct also showed high levels of resistance to MTX and ara-C in a CFU-GM assay. This fusion protein confers resistance to two antineoplastic agents that differ in their mechanism of action, and may be useful in the design of gene transfer strategies for protection of target cells against multiple drugs. Since high-dose ara-C and MTX are used in the treatment of lymphomas, this vector may be of value in protecting human hematopoietic progenitor cells from the toxicity of these antimetabolites.

Current clinical gene therapy protocols for the treatment of human immunodeficiency virus type 1 (HIV-1) infection involve the ex vivo transduction and expansion of CD4⁺ T cells derived from HIV-positive patients at a late stage in their disease (CD4⁺ cell count <400 cells/mm³). We examined the efficiency of transduction and transgene expression in adult bone marrow(BM)- and umbilical cord blood (UCB)-derived CD34⁺ cells induced to differentiate into T cells and monocytes in vitro with an MuLV-based vector encoding the neomycin resistance gene and an intracellular antibody directed against the Tat protein of HIV-1 (sFvtat1-C_κ). The expression of the marker gene and the effects of antiviral construct on subsequent challenge with monocytotropic and T cell-tropic HIV-1 isolates were monitored in vitro in purified T cells and monocytes generated in culture from the transduced CD34⁺ cells. Transduction efficiencies of CD34⁺ cells ranged between 22 and 27%. Differentiation of CD34⁺ cells into T cells or monocytes was not significantly altered by the transduction process. HIV-1 replication in monocytes and CD4⁺ T cells derived from CD34⁺ cells transduced with the intracellular antibody gene was significantly reduced in comparison with the degree of HIV replication seen in monocytes and CD4⁺ T cells derived from CD34⁺ cells transduced with the neomycin resistance gene alone. Further, T cells and monocytes derived from CD34⁺ cells transduced with the intracellular antibody gene were demonstrated to express the sFvtat1-C_κ transgene by RT-PCR and had a selective growth advantage in cultures that had been challenged with HIV-1. These data demonstrate that sFvtat1-C_κ inhibits HIV-1 replication in T cells and monocytes developing from CD34⁺ cells and supports the continuing development of a stem cell gene therapy for the treatment of HIV-1 infection.

Preclinical studies were designed to investigate the safety of recombinant adenoviruses infused into the portal vein of adult rhesus monkeys, as well as the safety and efficacy of readministration of these agents. The vectors used were recombinant adenoviruses, the E1 region of which was replaced with a marker gene expression cassette. Four 3- to 5-kg rhesus monkeys underwent portal vein cannulation, and infusion of escalating doses of recombinant first-generation vector. Serial sequential liver biopsies were performed, and necropsies were performed out to 14 months. X-Gal histochemical analysis of the liver showed evidence of dose-dependent increased gene transfer throughout the liver. Quantitative analysis of histopathology showed that portal inflammation was also present in transduced livers, and occurred in a dose-dependent manner. Severe toxicity, including mortality, was noted at the highest dose of vector. Readministration of a second vector was associated with the same degree of toxicity as the first vector, but prompted a much more vigorous neutralizing antibody response. The data suggest that intraportal administration and readministration of recombinant adenoviral E1-deleted vectors are feasible and safe. Vector administration at the highest dose (1 X 10¹³ particles/kg) was associated with severe clinical and biochemical toxicity, and significant gene expression was associated with transaminitis. Readministration of vector is safe, but gene transfer is limited by the presence of neutralizing antibody.

Production of high-titer rAAV is essential for in vivo clinical application. One limiting factor may be the failure of existing systems to replicate the packaging genome in such a way that expression of Rep and Cap proteins is coordinately amplified. DISC-HSV (disabled single-cycle virus) is a genetically modified herpes simplex virus (HSV) that by deletion of glycoprotein H (gH) is infectious only if propagated in a complementing cell line. In this study, we have used DISC-HSV as a helper for rAAV replication, and have simulated to some extent the amplication of the rep and cap genomes seen in wtAAV infection by incorporating both these and vector sequences in HSV amplicons. Facilitated production of AAV Rep and Cap proteins translates into a considerably improved recovery of rAAV, which transduces cells of the neuroretina in vivo with high efficiency. The potential for contamination with infectious herpes particles is eliminated by the use of noncomplementing (gH^-) cell lines to propagate the virus, and by standard purification methods. The use of DISC-HSV and herpes-derived amplicons for production of rAAV may be a useful strategy for future in vivo studies and for clinical application.

Bone marrow stromal cells can be used as an alternative source of cells for neural transplantation and repair. Here, the efficacy of genetically modified marrow stromal cells was examined in a rat model of Parkinson disease. Rat marrow stromal cells (rMSCs) and human marrow stromal cells (hMSCs) were genetically engineered by transduction with retroviruses encoding tyrosine hydroxylase (TH) and GTP cyclohydrolase I, the enzyme necessary for production of the tetrahydrobiopterin cofactor for TH (BH₄). Transduced cells synthesized 3,4-dihydroxyphenylalanine (_L-DOPA) in vitro and maintained their multipotentiality after retroviral transduction. To examine the cells in vivo, transduced rMSCs were injected into the striatum of 6-hydroxydopamine-lesioned rats. _L-DOPA and metabolites were detected by microdialysis in the denervated striatum of rats that received doubly transduced rMSCs. Also, there was a significant reduction in apomorphine-induced rotation when compared with controls. The cells engrafted and survived for at least 87 days. However, expression of the transgenes ceased at about 9 days, an observation consistent with reports from other laboratories in which similar retroviruses were used to express transgenes in the brain.

Intracameral fibrin formation, a complication of ocular inflammation and intraocular operations, sometimes results in glaucoma and/or corneal damage leading to permanent visual loss. We transferred a therapeutic gene to the corneal endothelium in order to use it as a therapeutic organ. A plasmid encoding tissue plasminogen activator (tPA) was injected into the anterior chamber of rats and electric pulses (EPs) were given subsequently, which transferred a plasmid gene to a highly selected area of corneal endothelium with no inflammation. The biologically active tPA was clearly present for 4 days after treatment. Fibrin formation induced by YAG laser-generated bleeding in the anterior chamber decreased significantly more in treated eyes than in control eyes. Corneal opacity was significantly lower in treated eyes than in control eyes and histological damage was not apparent in the treated eyes. This genetic modification allows us to use the corneal endothelium to treat various ocular diseases and could be a new and effective type of pharmacologic gene therapy.

Accurate quantification of gene transfer (or gene correction) is a universal challenge in the field of gene therapy. In developing a clinical trial of lymphocyte gene therapy for Hunter syndrome (mucopolysaccharidosis type II), methods using Southern blot or automated DNA sequencing technology were employed, but found to be laborious and subject to considerable variation. As an alternative approach, we explored a real-time kinetic PCR assay appropriate to new instrumentation (PE Biosystems model 7700). A TaqMan probe was designed to hybridize directly across the exon 2-exon 3 junction of the iduronate-2-sulfatase transgene cDNA. In this assay system, cDNA from the retroviral vector L2SN generates a PCR product that is 84 nucleotides long and readily quantified by TaqMan probe binding and subsequent cleavage. Evaluation of this method demonstrated sensitivity over at least 5 logs with respect to the standard (vector plasmid pL2SN). There was no detectable signal from genomic DNA from nontransduced cells, thus indicating the specificity of this assay. The sample preparation method used to prepare specimens was a relatively simple cell lysis procedure, without DNA extraction, and represents a significant advancement over the more complex methods of DNA extraction that are typically used for such assays. This specific assay, and comparison to previous methods, illustrates the utility of a new method that is readily generalized to many gene therapy studies, and that has the potential to be extended to measure gene expression by means of quantitative RT-PCR.