Abstract
The relatively rapid progress of gene therapy for retinal disorders, compared with gene therapy for many other conditions, testifies to the suitability of the retina as a target for gene therapy. The eye is small, compartmentalized, and isolated from the rest of the body by the blood–retinal barrier. Retinal cells can be targeted effectively either by intravitreal injection or subretinal delivery of modest doses of vector. The presence of the blood–retinal barrier not only reduces the extent of vector dissemination outside the eye, but it also limits the severity of immune responses to both the vector and the transgene product, avoiding what has proven to be a major barrier to effective gene therapy for other disorders.
Perhaps the most important reason to be optimistic about the future prospects for retinal gene therapy is the availability of viral vectors that can effectively, and with minimal toxicity, transduce photoreceptor cells, the most important target cell type for the treatment of inherited retinal dystrophies. Subretinal administration of AAV vectors, especially those derived from AAV serotypes 5 and 8, has been shown to lead to highly efficient photoreceptor cell transduction (Allocca et al., 2007) and prolonged, potentially life-long, expression of the transgene. As most cells in the retina do not normally divide, the lack of integration of vector DNA into the host cell genome is an advantage rather than a limitation of AAV, as it diminishes the risk of malignant transformation that can be associated with viral integration.
In principle, it may be an advantage to use nonviral vectors for retinal gene therapy, particularly if they can be manufactured more cost effectively than viral vectors. As an organ, the eye is one of the best organs with which to test the usefulness of new vectors: there are a wide range of cell types, from endothelium to neurons; many different models of disease; and many other vectors that have already been tested in the eye with which to compare efficacy. To date, the efficiency and duration of nonviral vector-mediated gene transfer to the retina are rather poor compared with the best viral vectors. Whereas viral vectors have evolved to enter the cell efficiently and deliver their genome to the host nucleus, naked DNA–liposome complexes lack this ability. In this issue, Delgado and colleagues report how they have attempted to overcome these limitations, using rational nonviral vector design (Delgado et al., 2012). They have used dextran in the liposome preparation to enhance adhesion of the vector to clathrin, shifting vector entry from caveola-mediated to clathrin-mediated endocytosis. They have also added protamine, which contains a nuclear localization signal, in order to promote translocation of the vector to the host cell nucleus. In vitro assessment using retinal pigment epithelium (RPE) cells showed a significant increase in transgene expression compared with a vector lacking these two proteins. Although there was some evidence of transduction after intraocular administration, further studies are necessary to show that this type of vector can provide long-lasting transgene expression in vivo and lead to effective rescue of an animal model. Nevertheless, these experiments suggest that inclusion of proteins or perhaps protein domains to “piggyback” onto specific cellular pathways may substantially improve the usefulness of nonviral gene transfer.
Unsurprisingly, all three recent clinical trials of gene therapy for retinal dystrophy, and the majority of successful proof-of-concept studies to date, involve gene supplementation therapy for the treatment of autosomal recessive forms of retinal degeneration. However, substantial progress has also been made in the development of treatment strategies for autosomal dominant disease. Disease gene-independent approaches, such as the expression of neurotrophic or antiapoptotic factors, have been shown to prolong slightly the survival of photoreceptor cells in various animal studies. However, a more promising approach currently is the development of disease gene-specific approaches, such as RNA interference (RNAi)-based gene silencing. Although specific silencing of the mutated alleles is conceptually elegant, this approach is hindered by the large number of dominant mutations that may require targeting, particularly in the gene encoding rhodopsin, and the difficulty in achieving specific recognition of single-nucleotide variants. Therefore, most commonly used strategies target both mutated and normal endogenous mRNA species, while supplementing the cell with a modified or “hardened” mRNA that is not recognized by the silencing construct. The past 5 years have seen the contemporaneous development of siRNA tools to silence endogenous rhodopsin mRNA species and transgene expression cassettes to drive sufficiently high levels of the hardened rhodopsin transgene. An important study that combined silencing and supplementation in a mouse model of autosomal dominant retinitis pigmentosa (ADRP), due to a P347S rhodopsin mutation, showed structural and functional preservation of photoreceptor cells for at least 5 months, strengthening the validity of this approach (Millington-Ward et al., 2011). However, in that study the silencing construct and the supplementation cassette were provided in two separate vectors. In this issue, Mao and colleagues show, for the first time, long-term rescue of retinal structure and function, using a single AAV vector to silence and supplement rhodopsin in another model of ADRP, the P23H Rho mouse (Mao et al., 2012). Retinal activity in this study was sustained at a constant level from 2 months posttreatment until termination at 9 months. These impressive results bode well for the translation of rhodopsin gene silencing and supplementation therapy through to clinical application.
Irrespective of the approach used, a detailed description of the disease phenotype in humans is vital for the development of effective gene therapy protocols. This is required in order to determine the potential for therapy including evaluation of the optimal time frame and area for treatment, and to identify suitable markers/phenotypic changes that will allow a timely readout of efficacy. In this issue, Dinculescu and colleagues provide a detailed clinical assessment of the retinal phenotype of a 19-year-old patient with a homozygous null mutation in the RPE-specific MFRP gene (Dinculescu et al., 2012). Although the presence of developmental foveal malformation might be regarded as a contraindication for gene therapy in the fovea, the authors have found extrafoveal preservation of rod and cone function, suggesting that this area may be a suitable target for gene supplementation therapy. The study goes on to show that a naturally occurring mouse model of Mfrp deficiency can be treated successfully with a modified AAV8 viral vector carrying a murine Mfrp transgene, leading to improved retinal function and prolonged survival of the photoreceptors. Even though the authors identify a requirement for further animal work, this study, with the combination of proof of concept in the mouse and detailed description of the patient phenotype, provides a solid basis for the development of MFRP gene supplementation therapy.
Over the course of 20 years, gene therapy for retinal degeneration has moved from a hypothetical possibility to experimental clinical reality, but until now only for AAV-mediated gene supplementation for recessive disease, perhaps the least complicated type of gene therapy. However, these recent successes have thrust the retina into the limelight as one of the most promising target tissues for gene therapy, and as a result, the research effort into retinal gene therapy has expanded, both in academic and commercial settings. Consequently, progress seems more rapid than ever. Although it is impossible to say with certainty how the field will evolve in the coming years, it seems likely that, with the rapidly evolving technology and increasing understanding of the pathophysiology of retinal disorders, gene therapy will soon become a realistic option for many more retinal degenerations. The main challenge over the next 20 years is not just to rescue more mouse models (there are more than 100 single-gene defects that lead to retinal dystrophy), but to translate more of our findings into clinical studies and to optimize treatments in patients.