Adeno-associated virus

Introduction
The eye is outstandingly well suited for the development of novel therapeutic approaches such as gene therapy. It is easily accessible and allows local application of therapeutic agents with reduced risk of systemic effects. Non-invasive procedures for the determination of ocular structures and functions in both the clinic and the laboratory are better developed than for any other organ. Ease of access to the eye enables manipulation at all levels from subcellular to whole organ and the understanding of the genetics, biochemistry, cell biology, developmental biology, neurobiology and physiology of the eye is highly advanced. Moreover, there are many established small and large animal models of retinal disease. These factors have allowed a more rapid progress in the development of gene therapy for retinal disease compared with that for other neurological disorders.

i) Slitlamp examination of the eye

ii) Vitrectomy

Over the past twelve years we have been at the forefront of investigating the basic principles of gene transfer to the eye with a broad programme of work to develop gene therapy for eye diseases affecting the retina, including inherited retinal degenerations as well as complex diseases such as those associated with retinal and choroidal neovascularisation and posterior uveitis.

We have developed a variety of viral gene therapy vectors. The most promising with respect to clinical use are those based on adeno-associated virus (AAV) (Ali et al Hum Mol Genet 5: 591-4). These are attractive for a number of reasons, not in the least because they are non-pathogenic for humans.

Whilst there are a number of different AAV serotypes, current vectors are almost exclusively based on AAV serotype 2. Following subretinal injection of AAV-2 vectors efficient gene therapy to photoreceptor cells and the retinal pigment epithelium (RPE) has been demonstrated in a variety of animals, including rodents, dogs, and primates with minimal inflammation and toxicity. Recently, AAV-2 vectors, pseudotyped with AAV-4, 5 and 8 capsids have been evaluated in the eye. Subretinal injection of AAV-2/4 vectors results in RPE specific transduction in rodents, dogs and primates. Subretinal injection of AAV-2/5 and especially AAV-2/8, results in faster and more efficient transduction of the photoreceptor cells and RPE than was found with AAV-2/2. These vectors are likely to become the vectors of choice for future use, but currently AAV-2 is still used as it is the best studied serotype.

i) GFP expression on the fundus after subretinal injection of AAV serotype 8

ii) GFP expression in photoreceptor and RPE after subretinal injection of AAV serotype 8

We have use recombinant AAV-2 vectors in a large series of studies to demonstrate the feasibility of gene transfer in rodent models and have demonstrated high efficiency, stability and minimal associated toxicity.

Besides AAV, we are also developing gene therapy vectors based on lentiviruses. Unlike AAV, lentiviruses do not transduce the photoreceptor cell efficiently, but lentiviruses can be used successfully to transduce other tissues in the eye, such as the RPE (Bainbridge et al Gene Ther 8: 1665-8). Lentiviruses have the advantage over AAV that they express more rapidly and to higher levels. In contrast to AAV, lentiviruses integrate their DNA into the host genome, which is am advantage when transducing a dividing tissue, as the transgene is not lost over time. The eye contains many non-dividing though, where it can be a disadvantage, as integration increases the risk of mutagenesis of the host genome. Recently, we have shown that non-integrating (Integrase deficient) lentiviruses can efficiently transduce ocular tissues in the mouse, resulting in long-term transgene expression (Yanez-Munoz et al Nat Med 12: 348-53). This allows us to choose the lentivirus system most suited to our target tissue: an integrating vector for dividing cells or a non-integrating vector for non-dividing cells or transient transgene expression.

i) Schematic of the structure of a lentiviral vector

ii) GFP expression in the RPE after subretinal injection of a lentiviral vector

Gene therapy for photoreceptor defects
Most common forms of retinal dystrophy are caused by mutations in photoreceptor specific genes. Therefore, most of the studies aimed at developing gene therapy for inherited retinal degeneration have focused on the treatment of photoreceptor defects. In 2000 we were the first group to show unequivocal evidence of restoration of photoreceptor structure and function after gene replacement therapy. These experiments were done in a mouse model of retinal dystrophy due to a gene encoding peripherin-2, a photoreceptor-specific protein required for the generation of outer segments (Ali et al Nat Genet 25: 306-10). Subretinal injection of AAV-2 vector carrying the peripherin-2 transgene resulted in the formation of discs and the generation of new structures that, in many cases, were morphologically similar to outer segments. The restoration of photoreceptor structural integrity was reflected in significantly improved ERG responses and a useful improvement of higher visual function.

Expression of various neurotrophic factors such as CNTF and GDNF in the eye can significantly prolong the survival of photoreceptor cells in models of retinitis pigmentosa, even though these methods do not correct the underlying defect.

Photoreceptor outer segment structures are formed after gene therapy with AAV expressing peripherin

By combining the AAV-2-peripherin treatment with an AAV-2 or lentivirus construct containing a GDNF gene an improved rescue of the defect could be shown compared to eyes treated with either construct separately, indicating that neurotrophic support may form a useful tool in the development of gene therapy for photoreceptor defects (Buch et al Mol Ther 14:700-9).

Since, we have shown successful gene replacement therapy for another photoreceptor defect caused by mutations in the Rpgrip gene. In mice with this defect, we could not only improve the photoreceptor function but also prolong their survival. Further attempts are ongoing to improve on these results with neurotrophic factors and more efficient constructs and using a large animal model of the disease (Pawlyk et al IOVS 46:3039-45).

Restoration of RPGR localisation at the connecting cilium after gene therapy with AAV expressing RPGRIP

Currently we are also working on the development of gene therapy for several other, more common, forms of severe retinal degeneration caused by mutations in photoreceptor cell specific genes, such as RPGR (XRP3) and AIPL1.

Gene therapy for RPE defects
Although RP caused by defects in the RPE are relatively rare, they are attractive targets for therapy, as there are a number of reasons why RPE defects may generally be more amenable to treatment. Firstly, in the case of an RPE defect, the photoreceptor cells are inherently healthy and therefore more likely to survive if the function of the RPE can be restored. Secondly, the transduction of the RPE by AAV and lentiviruses is more efficient than the transduction of photoreceptors. Finally, RPE cells are in contact with a number of photoreceptors and successful treatment of one RPE cell can therefore be expected to result in improved function in more than one photoreceptor cell.

To date we have investigated the treatment of two forms of RP caused by RPE defects. The RCS rat has a naturally occurring mutation in a gene encoding a member of the Axl subfamily of tyrosine kinases, Mertk, which in the eye is expressed only in the RPE. In the absence of functional Mertk, the RPE is unable to phagocytose the shed outer segment material, which accumulates in the subretinal space. As a result RPE and photoreceptors cells are lost and the retina degenerates. We performed gene replacement therapies in these animals using AAV-2 and lentiviral vectors and observed correction of the phagocytic defect, slowing of photoreceptor loss and preservation of the retina (Smith et al Mol Ther 8: 188-95; Tschernutter et al Gene Ther 12: 694-701). Retinitis pigmentosa caused by mutations in the MERTK gene is rare and although a number of patients have been identified, it is unlikely that a clinical trial for this disease will be started soon.

Preservation of retinal integrity after gene therapy with lentivirus expressing Mertk

The second form of retinal degeneration caused by an RPE defect is a more promising prospect for developing a clinical trial. The RPE65 gene is involved in the recycling of all-trans¬-retinal to 11-cis-retinal, which is required for the phototransduction pathway. In its absence healthy photoreceptors lack photopigment and cannot function. In 2001 successful gene therapy of dogs with a naturally occurring mutation in the RPE65 gene was shown by Acland et al. Since then we have shown that we can repeat the treatment of dogs and mice with our AAV vectors containing the human RPE65 gene driven by a human RPE65 promoter fragment. In mice restoration of photoreceptor cell function can be shown by ERG analysis. In dogs improved photoreceptor function was shown not only by electrophysiology, but also by showing improved vision in behavioural analyses. Based on these results and on the results of safety studies, we have obtained permission to start a clinical trial for the treatment of this form of retinal degeneration. More information on the trial, which started with the treatment of the first patient in February 2007, is available here.

Gene therapy for ocular neovascularisation
Ocular neovascularisation is a central feature of the major causes of blindness; pathological angiogenesis occurs in retinopathy of prematurity (ROP), proliferative diabetic retinopathy and age-related macular degeneration; leading causes of blindness in infants, individuals of working age and the elderly, respectively. Current treatments are of limited efficacy and associated with significant adverse effects. Neovascularisation causes loss of vision through increased vascular permeability leading to retinal oedema, vascular fragility resulting in haemorrhage or fibro-vascular proliferation with tractional and rhegmatogenous retinal detachment.

Choroidal neovascularisation at the site of damage to the Bruch's membrane

Angiogenesis is an attractive target for therapeutic intervention since it represents a final common pathway in processes that are multifactorial in aetiology and is the event that typically leads directly to visual loss. Angiogenesis is a multi-step process that is controlled by complex interactions between growth factors, including vascular endothelial growth factor (VEGF), extracellular matrix and cellular components, the net outcome being determined by the balance of angiogenic and angiostatic elements. The therapeutic manipulation of one of a combination of these elements offers the potential means to control neovascularisation in the eye. We have established a program of research to develop gene delivery of angiostatic proteins for the treatment of ocular neovascular disorders. This work is based on the premise that the local delivery of genes encoding angiostatic proteins offers the option of their targeted, sustained and regulatable expression after a single surgical procedure to introduce a vector to an intraocular site. We have investigated the hypothesis that angiogenesis in the retina and choroid can be inhibited by the intra-ocular delivery of recombinant viruses carrying genes encoding angiostatic proteins. We have demonstrated that local expression of angiostatic molecules reduced the extent of retinal neovascularisation in appropriate experimental models and have also shown that transgene expression can be specifically targeted to sites of active angiogenesis in these models and regulated by the incorporation of the hypoxia-responsive element in the expression cassette. (Bainbridge et al Gene Ther 9 320-6; Bainbridge et al Gene Ther 10:1049-54; Balaggan et al Gene Ther 13: 1153-65).

Staining of hypoxic areas (green) in a retinal flat mount

Although the angiostatic effect of single growth factors can be substantial, this approach fails to completely abolish angiogenesis in vivo. Angiogenesis is highly complex process in which the effects of cytokines can be contextual; uncontrolled neovascularisation may be the result of uninhibited parallel pathways or compensatory mechanisms. Ongoing clinical trials suggest that therapy with a single angiostatic growth factor may be insufficient to counteract the effect of the multiple angiogenic factors expressed endogenously. Moreover, the complete inhibition of angiogenic cytokines such as VEGF is likely to be undesirable since these typically mediate critical physiological functions when expressed at basal levels in normoxia. We are now investigating the hypothesis that hypoxia-driven angiogenesis may be specifically, powerfully and more appropriately controlled by targeting each of two critical upstream steps; activation of hypoxia-inducible transcription factor-1 (HIF–1) and overproduction of nitric oxide.

Lentivirus expressing IL1ra reduces infiltrating cells in the posterior segment

AAV expressing IL10 reducaes signs of uveitis in the retina

Gene therapy for uveitis
The immune privileged environment of the eye results, in part, from active control by immunoregulatory factors such as IL-10 and TGFß. In spite of these (and other) specialised mechanisms, autoimmune uveitis is a major cause of blindness. Current treatments are non-specific and cause systemic effects. Disease progression coincides with leukocyte infiltration and production of pro-inflammatory cytokines such as IFN?, TNF? and IL-1, and cytotoxic mediators such as nitric oxide, resulting in the perpetuation of infiltration, local tissue damage and photoreceptor cell death. Experimental auto-immune uveitis (EAU) and endotoxin induced uveitis (EIU) are induced, murine models of posterior and anterior uveitis respectively. Our aim is to decrease photoreceptor loss and prevent disease onset by modulating the immune response seen in EAU with AAV and lentivirus delivered immunoregulatory signals.

We have shown that AAV carrying the IL10 gene can substantially reduce severity of inflammatory disease in animals after induction of EAU (Broderick et al Mol Ther 12: 369-73). We have also demonstrated a reduction in disease severity in mice with EIU after treatment of the eyes with lentiviruses carrying either IL10 or IL1ra. For an optimal treatment effect it is likely that we will need to treat with more than one gene simultaneously, as several inflammatory pathways are often active.

This page last modified 18 December, 2012 by xxx