Prospects for gene therapy

strategies are being investigated. These include administration of recombinant neurotrophic factors, photoreceptor, RPE or stem cell transplantation, and gene therapy. Of these di¡erent approaches, the most signi¢cant progress has probably been made with regard to gene therapy. E⁄cient in vivo gene transfer to photoreceptor cells and the RPE following subretinal injection of adenoassociated viral (AAV) vectors resulting in stable transgene expression has been demonstrated in a variety of animals, including rodents, dogs and primates with minimal in£ammation and toxicity (Bennett et al 1999, Dudus et al 1999, Sarra et al 2002, Bainbridge et al 2003). The use of cell-speci¢c promoters enables transgene expression to be restricted to either photoreceptors or RPE, providing an additional level of control. A number of recent studies have clearly demonstrated either retardation of photoreceptor cell loss or functional improvement in animal models of retinal degeneration following AAV-mediated gene transfer (McGee Sanftner et al 2001, Liang et al 2001, Acland et al 2001, Ali et al 2000). Here I will outline some of our work on the development of gene therapy for inherited retinal degeneration and compare the e⁄cacy of treating of photoreceptor and RPEspeci¢c defects.

We have focused most of our e¡orts on developing gene replacement therapy for inherited retinal degeneration using Prph2Rd2/Rd2 mice. These animals, formerly known as rds (retinal degeneration slow) mice are a well-characterized model of retinal degeneration. They are homozygous for a null mutation in the Prph2 gene, encoding the structural glycoprotein, peripherin 2, which is essential for outer-segment formation (Travis et al 1991). Due to the failure to develop photoreceptor outer segments, Prph2Rd2/Rd2 mice have extremely limited electroretinogram (ERG) responses. This facilitates reliable quanti¢cation of functional improvement following treatment making the Prph2Rd2/Rd2 mouse a very useful tool for assessing the e⁄cacy of gene therapy protocols. We have demonstrated improvement of photoreceptor ultrastructure at a number of time points following injection of an AAV vector carrying a peripherin 2 cDNA driven by a rhodopsin promoter (AAV.rho.rds) (Ali et al 2000, Sarra et al 2001). We have now analysed treated animals at weekly intervals using ERG (Schlichtenbrede et al 2003a,b). We have analysed ERG changes in a series of increasing stimulus intensities, using trace pattern and b-wave amplitude as indicators of functional improvement at several time points following treatment. Although there is some inter-animal and test^retest variability, the ERG has proven to be very reliable tool in assessing therapeutic bene¢t. In the absence of treatment, we found the concordance between the b-wave amplitudes of the two eyes of a single animal was very high, both in normal and Prph2Rd2/Rd2 mice. Another phenomenon demonstrated in the recordings was the variability in rates of degeneration in di¡erent mice. This was in contrast to non-degenerating mice (wild-type CBA), where considerably less variation was seen over large a number of animals. The

di¡erence in variability may be explained by a di¡erence in genetic heterogeneity. The Prph2Rd2/Rd2 mice are on a CBA background but are not congenic. This is consistent with the observation in humans that within families manifesting the same degeneration and the same gene defect there can be a very large variation in the rate of degeneration and vision loss (Apfelstedt-Sylla et al 1995).

The b-wave pattern in the animals treated with AAV.rho.rds resemble that of a normal animal. Furthermore the presence of oscillatory potentials on rise of the b-wave indicate that the transverse connections of the outer retina involving amacrine cells are also functioning. The average b-wave amplitude in the rescued retina following single time point injection at 4 weeks post injection is 97 mV at the 100 mcds/m2 stimulus. For a wild-type mouse (CBA) the b-wave amplitude at the same intensity is around 400 mV. Thus the b-wave amplitude in treated animals is approximately 25% of that in normal wild-type mice. Given the subretinal injection technique reaches approximately 50% of retina and about 70% of photoreceptors are transduced (Sarra et al 2002), the degree of the physiological rescue is consistent with the anticipated number of transduced photoreceptors.

We have established a positive correlation between the area of retina transduced and the magnitude of the ERG. This is re£ected in the di¡erence between the b- wave amplitudes following single time point and repeated injections. The average b-wave amplitude was 90 mV and 140 mV for the single time point and repeatedly injected animals respectively. This corresponds to a functional rescue over the untreated fellow eye of 80% and 250% respectively.

We have also evaluated the impact of gene replacement therapy on higher visual function. Demonstrating an improvement of central visual responses and correlating this to changes in the retina is a critical step for validating gene therapy approaches for inherited retinal disease. The e¡ect of treatment on higher visual function was assessed by recording from central visually responsive neurons in the superior colliculus and improvements were correlated in individual animals with retinal function and histological and biochemical changes. Although gene replacement therapy only partially restores photoreceptor morphology, it results in a 300% increase of the visual cycle protein rhodopsin, leading to retinal function improvement, re£ected by a 250% higher b-wave amplitude. This corresponds to 15% of the rhodopsin levels in normal animals and 25% of the normal b-wave amplitude. Prph2Rd2/Rd2 mice with improved ERGs also had signi¢cantly higher central visual responses (166% increase at 24 cd/m2). These ¢ndings suggest that gene replacement therapy leading to even relatively modest structural improvement may result in improved visual function.

Despite the clear functional improvement, bene¢cial e¡ects of treatment appear to be transient. The onset of the rescue is consistent with the time course of AAV expression with a lag of approximately two weeks before e¡ective transgene expression levels can be recorded. The improvement in ERG is clear and

consistent between 3^12 weeks for all animals with repeated injections. The longest period of signi¢cant functional bene¢t is around 14 weeks. At later time points the di¡erences between treated and untreated eyes are no longer signi¢cant. We have also observed a decline in ultrastructural quality of the outer segments over time even when we treat young animals. This might be explained either by increased disturbance of photoreceptor cell physiology over time, irrespective of outer segment induction, or by inappropriate transgene expression levels. We have also observed that the number of induced outer segments decrease over time and that there is no reduction in the loss of photoreceptor cells following therapy (Sarra et al 2001). Our investigations have excluded procedure-related damage, vector toxicity and immune responses as major factors which might counteract the bene¢ts of ultrastructural improvements (Sarra et al 2001). There are a number of possible alternative explanations. We consider the major factors to be delayed onset or inappropriate levels of transgene expression or an insu⁄cient transduction rate. A combination of all three and/or other factors may be important.

We wanted to determine whether we could improve photoreceptor survival as well as improving function by combining neuroprotection with gene replacement. Intraocular delivery of a variety of neurotrophic factors has been widely investigated as a potential treatment for retinal dystrophy. A number of studies have demonstrated the e¡ect of vector-mediated ciliary neurotrophic factor (CNTF) gene expression on photoreceptor cell loss, suggesting that this may be an e¡ective treatment for human retinal dystrophies. These studies have focused on animals in which there is very little function and concentrated on the morphology of the retina. A recent report, however, has evaluated long-term AAV-mediated CNTF gene expression in a transgenic mouse model of retinal disease caused by a dominant mutation in the Prph2 gene, in which retinal function is relatively normal and only declines slowly with photoreceptor cell loss. Bok et al (2002) found a 23% preservation of photoreceptor cells as compared to the untreated side at 5 months after treatment, but observed reduced ERG recordings following treatment. In order to evaluate CNTF gene delivery as a potential treatment, we used the Prph2Rd2/Rd2 mouse. CNTF was expressed intraocularly using AAV-mediated gene delivery either by itself or, in a second treatment group, combined with AAV-mediated gene replacement therapy of peripherin 2 (Schlichtenbrede et al 2003a,b). We con¢rmed in both groups of animals that CNTF reduces the loss of photoreceptor cells. Visual function, however, as assessed over a time course by ERG, was signi¢cantly reduced compared with untreated controls. Furthermore CNTF gene expression negated the e¡ects on function of gene replacement therapy. In order to test whether this deleterious e¡ect is only seen when degenerating retina is treated, we recorded ERGs from wild-type mice following intraocular injection of AAV

expressing CNTF. Here a marked deleterious e¡ect was noted, in which the b- wave amplitude was reduced by at least 50%.

The alteration in ERG trace may re£ect the morphological changes caused by CNTF gene expression. The changes we have observed . a less well-ordered outer nuclear layer (ONL) and, in animals also treated with a gene replacement vector, less outer segment material . are consistent with recent reports that suggest that exposure to high levels of CNTF may have an impact on photoreceptor di¡erentiation (Bok et al 2002, Ca¡e et al 2001). However, CNTF has been shown to exert an indirect e¡ect on photoreceptor cell survival through cells of the inner retina and Mu«ller cells. Wahlin et al (2001) have demonstrated that following intraocular injection of recombinant CNTF, signalling pathways are activated in Mu«ller, ganglion and amacrine cells, but not photoreceptors. This e¡ect was found both in normal and degenerating rodent retinae. It is the inner retinal neurons and Mu«ller cells that are responsible for generating the b- wave. Thus the e¡ect of CNTF on these cells may also be counteracting the potential bene¢t of photoreceptor preservation. Furthermore, CNTF overexpression might stimulate remodelling of the inner retina, leading to a change in cell function, a decrease in b-wave amplitude and subsequently to dedi¡erentiation of the photoreceptor cell, as re£ected by the changes in chromatin staining (Bok et al 2002) It has been suggested that in the course of retinal degeneration, loss of input from photoreceptors leads to input-dependent secondary neurons of the inner retina ‘seeking replacement’ for the input of lost photoreceptors. It would appear that the remodelling of the inner retina that normally occurs in the course of degeneration is exacerbated by the expression of CNTF.

Our studies have shown that intraocular treatment with CNTF may result in unwanted side e¡ects. These e¡ects might be tolerable provided CNTF administration is only temporary. This might be achieved through the use of an inducible-promoter or a pharmacological slow release device with later removal of the ocular insert. To achieve a sustained morphological rescue without causing retinal damage a ¢ne balance would have to be reached, requiring detailed dose^ response studies, further complicated by the relatively high individual variability in the degenerating retina. Alternatively, other neurotrophic molecules should be evaluated for safety and e⁄cacy. In retinal dystrophies, the degenerating retina undergoes a highly complex remodelling process, where any treatment-induced negative stimuli must be avoided.

To date most of the studies aimed at developing gene therapy for inherited retinal degenerations have focused on treatment of photoreceptor cell defects. Although we have shown that we can improve photoreceptor structure and function following gene replacement therapy in the Prph2Rd2/Rd2 mouse, so far we have been unable to slow the loss of cells. It is perhaps not surprising that we

have been unable to a¡ect the degeneration in this animal model . it has a major cellular defect. The Prph2Rd2/+ mouse on the other hand has a relatively mild phenotype of abnormal outer segments, and we are able to slow the degeneration by AAV-mediated delivery of peripherin 2.

Various forms of retinal degeneration are caused by RPE defects and there are a number of animal models of disease, including the RPE657/7 dog and the Royal College of Surgeons (RCS) rat. There are a number of reasons why RPE-speci¢c defects may generally prove to be more amenable to treatment. The transduction e⁄ciency of RPE is often much higher than that of photoreceptor cells and partial correction of RPE function may have signi¢cant impact on photoreceptor function and survival. Recently, Acland et al (2001) described highly e¡ective treatment of RPE657/7 dogs with rAAV containing the RPE65 gene.

In the RCS rat, the phagocytosis of photoreceptor cell debris (mainly composed of shed outer segments) by the RPE is defective due to a 409 bp deletion in the receptor tyrosine kinase gene Mertk (D’Cruz et al 2000). MERTK is involved in the recognition and binding of outer segment debris, possibly due to the appearance of phosphatidylserine in the outer lea£et of the plasma membrane of the shed disks. In the absence of functional MERTK the accumulation of outer segment material in the subretinal space results in the loss of photoreceptor cells and the degeneration of the retina. Photoreceptor cell loss is detectable histologically and electroretinographically from postnatal day 18 onward and subretinal accumulation of debris is apparent at that time. Depending on pigmentation of the animal the degeneration is virtually complete after 2 to 3 months; little electroretinographical activity is detectable and few photoreceptor cells remain. Recently, retinitis pigmentosa (RP) patients have been identi¢ed that have mutations in the MERTK gene (Gal et al 2000). In these patients the disease is characterized by night blindness at a young age, progressing rapidly to a loss of peripheral vision until a small island of central vision remains.

Recently, Vollrath et al (2001) reported the e¡ect on RCS rats of subretinal injection of recombinant adenovirus (rAd) containing the Mertk gene under control of the CMV promoter. As is the case with Ad, adeno-associated virus (AAV) is able to transduce RPE cells with a high e⁄ciency. In contrast with rAd vectors, transgene expression mediated by rAAV vectors is maintained over long periods; after injection into rodents, expression typically lasts throughout the lifetime of the animal. As there is also no notable decrease in rAAV transgene expression a year after treatment of non-human primates (Bennett et al 1999), it is likely transgene expression in humans will also persist for long periods of time. In order to develop a treatment for this form of RP that might be suitable for clinical application, we have evaluated gene therapy approaches in the RCS rat using rAAV vectors.

Our results show that following transduction of RPE cells in the RCS rat with rAAV expressing Mertk the phagocytotic function of these cells are restored and the accumulation of photoreceptor cell debris in the subretinal space is reduced (Smith et al 2003). The rate of photoreceptor cell degeneration that results from the deposition of debris is slowed after treatment, even though it is not completely prevented. Photoreceptor function is still detectable in treated eyes 9 weeks after injection, whereas untreated eyes at the same time point provide no recordable activity. ERG recordings from one-month old RCS rats are already lower than in wild-type animals, suggesting that damage to photoreceptors has occurred by the time treatment with AAV vectors may take e¡ect. Although the late onset of AAVmediated expression may have a major impact on the e⁄cacy of treating RCS rats, this would not be a concern for the treatment of patients. Whereas retinal degeneration in the rats is complete by 3 months, in RP patients with mutations in the MERTK gene the disease process occurs over several decades. The incomplete rescue in this model might also be explained by the fact that not all the debris between the photoreceptors and RPE is cleared. Not only has debris accumulated by the time AAV-mediated Mertk expression occurs, the level of transgene expression may not be su⁄cient to clear the backlog. Furthermore, only part of the retina is transduced. We estimate that subretinal injections in the rat cover roughly 50% of the retina. Even if the photoreceptor cells in these areas of bene¢t from a reduction in debris, degenerating photoreceptors in other parts of the retina may have a negative impact on photoreceptor survival in treated areas of the retina. Despite the partial nature of the rescue, loss of function in the treated rats was slowed by several months. A similar approach in patients might be expected to slow the degeneration by years, rather than months, given the di¡erence in the rates of degeneration in rats and humans. These results therefore provide strong support for developing AAV-based gene therapy approaches for RP patients with mutations in MERTK. We are now screening families for mutations in this gene in anticipation of a clinical trial.

Acknowledgements

I am very grateful to the Foundation Fighting Blindness, The Sir Jules Thorn Charitable Trust, The British Retinitis Pigmentosa Society and The Wellcome Trust for supporting our work on gene therapy for inherited retinal degeneration.

References

Acland GM, Aguirre GD, Ray J et al 2001 Gene therapy restores vision in a canine model of childhood blindness. Nat Genet 28:92^95

Ali RR, Sarra GM, Stephens C et al 2000 Restoration of photoreceptor ultrastructure and function in retinal degeneration slow mice by gene therapy. Nat Genet 25:306^310

172 ALI

Apfelstedt-Sylla E, Theischen M, Ruther K et al 1995 Extensive intrafamilial and interfamilial phenotypic variation among patients with autosomal dominant retinal dystrophy and mutations in the human RDS/peripherin gene. Br J Ophthalmol 79:28^34

Bainbridge JWB, Mistry A, Schlichtenbrede F et al 2003 Stable rAAV-mediated transduction of rod and cone photoreceptors in the canine retina. Gene Ther 10:1336^1344

Bennett J, Maguire AM, Cideciyan AV et al 1999 Stable transgene expression in rod photoreceptors after recombinant adeno-associated virus-mediated gene transfer to monkey retina. Proc Natl Acad Sci USA 96:9920^9925

Bessant DA, Ali RR, Bhattacharya SS 2001 Molecular genetics and prospects for therapy of the inherited retinal dystrophies. Curr Opin Genet Dev 11:307^316

Bok D, Yasumura D, Matthes MT et al 2002 E¡ects of adeno-associated virus-vectored ciliary neurotrophic factor on retinal structure and function in mice with a p216l rds/peripherin mutation. Exp Eye Res 74:719^735

Ca¡e AR, Soderpalm AK, Holmqvist I, van Veen T 2001 A combination of CNTF and BDNF rescues rd photoreceptors but changes rod di¡erentiation in the presence of RPE in retinal explants. Invest Ophthalmol Vis Sci 42:275^282

D’Cruz PM, Yasumura D, Weir J et al 2000 Mutation of the receptor tyrosine kinase gene Mertk in the retinal dystrophic RCS rat. Hum Mol Genet 9:645^651

Dudus L, Anand V, Acland GM et al 1999 Persistent transgene product in retina, optic nerve and brain after intraocular injection of rAAV. Vision Res 39:2545^2553

Gal A, Li Y, Thompson D et al 2000 Mutations in MERTK, the human orthologue of the RCS rat retinal dystrophy gene, cause retinitis pigmentosa. Nat Genet 26:270^271

Liang FQ, Aleman TS, Dejneka NS et al 2001 Long-term protection of retinal structure but not function using RAAV.CNTF in animal models of retinitis pigmentosa. Mol Ther 4: 461^472

McGee Sanftner LH, Abel H, Hauswirth WW, Flannery JG 2001 Glial cell line derived neurotrophic factor delays photoreceptor degeneration in a transgenic rat model of retinitis pigmentosa. Mol Ther 4:622^629

Sarra GM, Stephens C, de Alwis M et al 2001 Gene replacement therapy in the retinal degeneration slow (rds) mouse: the e¡ect on retinal degeneration following partial transduction of the retina. Hum Mol Genet 10:2353^2361

Sarra GM, Stephens C, Schlichtenbrede FC et al 2002 Kinetics of transgene expression in mouse retina following sub-retinal injection of recombinant adeno-associated virus. Vision Res 42:541^549

Schlichtenbrede FC, MacNeil A, Bainbridge JWB et al 2003a Intraocular gene delivery of ciliary neurotrophic factor results in signi¢cant loss of retinal function in normal mice and in the Prph2Rd2/Rd2 model of retinal degeneration. Gene Ther 10:523^527

Schlichtenbrede FC, da Cruz L, Stephens C et al 2003b Long-term evaluation of retinal function in Prph2Rd2/Rd2 mice following AAV-mediated gene replacement therapy. J Gene Med 5:757^764

Smith AJ, Schlichtenbrede FC, Tschernutter M, Bainbridge JW, Thrasher AJ, Ali RR 2003 AAV-mediated gene transfer slows photoreceptor loss in the RCS rat model of retinitis pigmentosa. Mol Ther 8:188^195

Travis GH, Sutcli¡e JG, Bok D 1991 The retinal degeneration slow (rds) gene product is a photoreceptor disc membrane-associated glycoprotein. Neuron 6:61^70

Vollrath D, Feng W, Duncan JL et al 2001 Correction of the retinal dystrophy phenotype of the RCS rat by viral gene transfer of Mertk. Proc Natl Acad Sci USA 98:12584^12589

Wahlin KJ, Adler R, Zack DJ, Campochiaro PA 2001 Neurotrophic signaling in normal and degenerating rodent retinas. Exp Eye Res 73:693^701