Gene therapy of retinal dystrophies: achievements, challenges and prospects

When ¢rst informed of this important discovery, it took this author and his contemporaries back to a meeting co-sponsored by the Retinitis Pigmentosa (RP) Foundation (now the Foundation Fighting Blindness) in the early 1970s. The Canadian and American RP Foundations sponsored a meeting of basic scientists and clinicians at a country resort called The Briars near Toronto. At that period in the history of the ¢eld, there were no credible insights into the molecular aetiology of RP. Although there was evidence for a genetic basis, there were multiple, con£icting theories based on vascular attenuation, in£ammation, lysosomal leakage within the retinal pigment epithelium (RPE) and others. At the conference, I recall that a young, brilliant investigator named Richard Cone (an appropriate name for a vision scientist) o¡ered his opinion that rhodopsin must somehow be involved in RP. Perhaps he was inspired by the burgeoning interest in phototransduction or maybe he was truly clairvoyant. In any event, his prediction turned out to be true.

The path to discovery of rhodopsin as an important player in RP was an arduous one and this success story is worth repeating because it points out the importance of fundamental science in any endeavour that involves the conquest of human disease. Following the development of methods for the puri¢cation of rhodopsin, it required approximately 15 years of hard work before the primary structure of this intrinsic membrane protein was solved (Ovchinnikov et al 1983, Hargrave et al 1983). This crucial information was then utilized by a gifted MD/PhD student named Jeremy Nathans at Stanford University who determined the structure of the human rhodopsin gene (Nathans & Hogness 1984). This information was essential for the screening of human DNA for the ¢rst disease-causing rhodopsin gene allele.

Thus, the prospect for gene-based approaches toward the treatment of inherited retinal diseases became a reality and we are gathered here to assess the past, present and future. Gene therapy to date has enjoyed limited success in non-ocular tissues but is now coming into prominence in treating the retina of the eye. One of the problems encountered in gene therapy of non-ocular tissues has been the premature application of this approach in human patients without su⁄cient investigator awareness of pitfalls. All of the work involving retinal gene therapy to date has been performed in preclinical settings (animal models) and this has been highly bene¢cial to our ¢eld.

Practitioners of the art and science of gene therapy in the eye have had the bene¢t of learning from initial setbacks in other ¢elds, the development of improved virus vectors and the relatively favourable setting in which these vectors are placed into the eye, namely the ‘subretinal space’. This region is not a true space but rather an extracellular matrix-rich compartment that separates the two major components of the retina. These are the neurosensory portion, which includes the photoreceptors, the cells directly involved in initiation of the visual response, and the RPE . a cellular monolayer that nourishes the photoreceptors. Incarceration of the virus

vectors within this space promotes interaction with the target cells, usually the photoreceptors or the RPE, and allows for the use of relatively low doses but high local concentrations around the cells that are to be transduced by the virus. Thus, no sophisticated homing systems are required.

The examples of gene therapy that will be used to support these statements are all taken from preclinical work, as are most of the proposals for future work. Nonetheless, Phase I/II clinical trials may be anticipated within a few years.

Gene replacement and ribozyme strategies

During the past ¢ve years, proof of principle for gene-based therapy of inherited retinal diseases has been established in animal models. Initial studies utilized ¢rstgeneration recombinant adenoviral (rAV) vectors (Bennett et al 1996) and demonstrated moderate, but transient success. The results were compromised in part by the complexity of the vector genome and its products, and signi¢cant stimulation of the immune response to the point that expression of the rescue gene was suppressed (Bennett et al 1996). Even when a minimal (‘gutted’) rAV was used as a vector, transient rescue gene expression was still observed (KumarSingh & Farber 1998). Considerably more promising results have been reported with recombinant adeno-associated (rAAV) virus which, when used as a vector is stripped of all of its genes with the exception of short viral inverted terminal repeats (Hauswirth et al 2000). The parent virus, which is non-pathogenic and stably integrates into an indi¡erent site in the human genome produces circulating antibodies in its host (about 80% of humans have antibodies to this virus) and the rAAV also produces a mild immune response (Bennett et al 1999). However several laboratories have reported long-term stable expression of genes delivered to the retina in rAAV (Bennett et al 1999, Hauswirth et al 2000), even following a second treatment in the presence of circulating antibodies to rAAV capsid proteins (Bennett et al 1999).

The ¢rst, striking success with rAAV in an animal model for inherited retinal degeneration was obtained with a ribozyme strategy (Lewin et al 1998). These investigators used transgenic rats carrying the P23H opsin gene mutation, which causes retinitis pigmentosa in humans and is dominant in its action. Coding sequences for hairpin and hammerhead ribozymes were designed to cleave a speci¢c site in the coding sequence of the mutant mRNA, leaving the wild type mRNA intact. These ribozyme coding sequences were then placed under control of the bovine opsin promoter to confer cell-speci¢c expression in rods and packaged into rAAV. Following injection into the subretinal space there was signi¢cant slowing of the rate of photoreceptor degeneration over a period of 3 months. Subsequent studies now indicate signi¢cant, persistent rescue for at least 8 months (LaVail et al 2000). In addition to morphological rescue, physiological

rescue was also observed as measured by comparative ERGs in the injected and contralateral non-injected eye. Hence, there is considerable potential for this strategy in the treatment of dominantly inherited disease where the dominantnegative e¡ect of the mutant allele must ¢rst be removed before replacement gene therapy can be e¡ective. In the case of dominant-negative rhodopsin mutations, replacement of the mutant allele mRNA would probably be necessary as well. It has been shown that disruption of one allele in these mice allows the formation of normal rod outer segments initially but results in slow photoreceptor degeneration over time (Lem et al 1999). Replacement would be even more important for rds/peripherin, whose gene displays haploinsu⁄ciency, as described below.

Others (Ali et al 2000) have recently shown that a replacement minigene vectored by rAAV can achieve partial rescue in rds / mice. This is encouraging, additional evidence for the e⁄cacy of the rAAV strategy. Ali and colleagues were able to elicit the formation of truncated outer segments in rds / , which carries a null, insertion mutation in this gene (Travis et al 1989). Rds/peripherin, the protein product of the rds gene (Connell et al 1991, Travis et al 1991) is a member of a small family of adhesion molecules that are essential for the formation of rod and cone photoreceptor outer segment discs. In its absence, no discs are formed (Nir & Papermaster 1986, Usukura & Bok 1987). Heterozygous (rds+/ ) mice, with a spontaneous null mutation in one allele, have oversized outer segment discs that curl into whorls (Hawkins et al 1985). Thus the gene manifests haploinsu⁄ciency (semidominance). The gene therapy strategy used by Ali et al (2000) was quite successful in rds / mice and would presumably be even more so in rds+/ . However, this is not the genotype for Rds/peripherin-based disease in the human population, which has over 60 di¡erent mutations reported thus far. Most human patients have dominant-negative point mutations (Sohocki et al 2001) rather than null mutations, which would require a combination of targeted ribozyme plus rds gene replacement therapy. Fortunately, all of this could be achieved in the context of an AAV vector and this work is in progress in our laboratory in collaboration with William Hauswirth, Alfred Lewin and Matthew LaVail. To that end, we are using a transgenic mouse model with a P216L, dominant-negative point mutation in Rds/peripherin (Kedzierski et al 1997) as the experimental subject.

Perhaps the most dramatic success with retinal gene therapy to date, at least in the context of its emotional impact, has been the recent work of Acland et al (2001) on the Briard dog. Some of the animals in this breed carry a mutation in the RPE65 gene. Previous studies on mice with a disrupted RPE65 gene by Redmond et al (1998), revealed that the protein product of this gene is essential for the production of 11-cis-retinol, the immediate precursor for 11-cis-retinal. The latter is the essential rod and cone photoreceptor opsin chromophore that initiates vision. It was subsequently discovered that a mutation in this gene is the cause of

the very early onset vision defect in these dogs (Aguirre et al 1998). Also, it was known at that time that this gene is the basis for a form of Leber congenital amaurosis (blindness) in humans (Marlhens et al 1997, Gu et al 1997). Acland et al (2001) using rAAV vector-based gene therapy, reported behavioural and electrophysiological evidence for the acquisition of visual function in young Briard dogs that were sightless prior to therapy. The prospect of treating young children in the same manner and o¡ering them sight is most heart-warming indeed. This is one of two examples of success in the treatment of an inherited retinal disease in which the primary site of expression of the mutant gene is the RPE.

The second example, a very recent one, involving treatment of an RPE-speci¢c disease comes from the venerable RCS (rdy) rat. This was one of the ¢rst putative animal models for human retinitis pigmentosa, but one whose causative gene had been a mystery for over 30 years, until D’Cruz et al (2000) solved this elusive riddle. They found that the defective gene is a receptor tyrosine kinase called Mertk, an RPE membrane receptor apparently essential for the transmembrane signalling event during the daily phagocytosis of photoreceptor outer segment fragments. Hard on the heels of this important discovery was the detection of gene defects in humans as well (Gal et al 2000) ¢nally, establishing the RCS rat as a bona ¢de animal model for retinitis pigmentosa. Vollrath et al (2001) have now partially corrected the retinal dystrophy phenotype in RCS rats by adenovirus-based gene transfer. Transduction of the RCS RPE with wild-type Mertk reversed the phagocytic defect, elicited considerable rescue of photoreceptors from death and improved the cornea-negative scotopic threshold response by two log units in the treated eyes.

The recent work of Acland et al (2001) and Vollrath et al (2001) provides compelling evidence that RPE-based inherited diseases can be treated as readily as those that are photoreceptor based. Interestingly, in both cases, correction of an expressed gene defect in the RPE was able to correct a photoreceptor phenotype. These phenotypes are characterized by absence of rhodopsin regeneration due to lack of 11-cis-retinal chromophore and slow photoreceptor degeneration in the case of the Briard dog and photoreceptor dystrophy and cell death in the RCS rat.

Gene-based trophic factor therapy

The seminal work of LaVail and Steinberg et al (Faktorovich et al 1990, LaVail et al 1992) demonstrated that a variety of growth factors and cytokines, which include leukaemia inhibitory factor (LIF) and ciliary neurotrophic factor (CNTF), rescue rodent photoreceptors from genetic and environmental insult for up to a month when injected as a single bolus into the vitreous cavity of the rat eye. It is not known how these factors exert their rescue e¡ect on photoreceptors. To date,

there is no evidence for CNTF receptors on adult rodent photoreceptors nor are their photoreceptors known to express CNTF. The nearest neighbour in which CNTF protein and mRNA have been detected is the Mˇller glial cell (Kirsch et al 1997). However, Beltran et al (2002) have recently reported expression of the CNTFRa subunit in canine photoreceptors.

Prompted by the observation that CNTF exerts a rescue e¡ect on adult photoreceptors, which do not express CNTF protein or the neuron-speci¢c subunit CNTFRa, several investigators have sought to determine whether the e¡ect of CNTF is direct or indirect by studying CNTF-mediated signalling pathways in retinal neurons and glia. CNTF activates the JAK (Janus tyrosine kinase) and STAT (signal transducers and activators of transcription) signalling pathway (Stahl & Yoncopoulos 1994), hence investigators have looked for upregulation of mRNAs for these proteins as a function of CNTF stimulation. CNTF can also activate ERK1 and ERK2 (Boulton et al 1994) so there is some cross talk for CNTF. Wahlin et al (2000) gave single injections of various factors (BDNF, CNTF and FGF2) into the mouse vitreous and observed a rapid increase in phosphorylated ERK (pERK) in Mˇller cells and an increase in c-fos in Mˇller, amacrine and ganglion cells. These levels returned to baseline by 6^24 h but were followed by increases in Mˇller cell GFAP. Peterson et al (2000), using rats, injected AxokineTM (a recombinant form of CNTF) into the vitreous and observed phosphorylation of STAT3 (pSTAT3) and ras-MAPK. pSTAT3 was localized to nuclei of retinal Mˇller cells, ganglion cells and astrocytes but not photoreceptors. These data, taken together suggest an indirect e¡ect of CNTF on photoreceptor survival in rodents.

When LaVail and associates attempted to extend their studies on injected CNTF rescue to various mouse models for RP, they achieved only limited success and had none with rds / mice (LaVail et al 1998). They suggested that better methods of CNTF delivery might be indicated. Supporting this hypothesis, Cayouette et al (1998), using intravitreal injections of an adenovirus-vectored, secreted form of CNTF, observed transient photoreceptor rescue in rds / mice. They observed an improvement in the scotopic and photopic ERG and, surprisingly, enhanced opsin expression, in spite of the fact that these null mice cannot express Rds/peripherin. We have recently had success in rescuing the photoreceptors of transgenic mice that carry a Rds/peripherin P216L point mutation with AAV-vectored, secreted CNTF (Bok et al 2002). However, our result, which was not transient in terms of morphological rescue, was also very di¡erent with respect to the ERG. An unexpected result was suppression of the scotopic a and b-wave and photopic b-wave in the injected eye compared to the non-injected, contralateral eye, even though there was signi¢cant morphological rescue of rods in the injected eye as judged by the thickness of the outer nuclear layer. Additionally, there was a striking change in photoreceptor nuclear morphology from the condensed

chromatin pattern typical of rods to a more di¡use pattern reminiscent of cones of bipolar cells. Liang et al (2001) reported similar aberrations when using rAAVvectored CNTF in transgenic rats carrying P23H and S334ter mutations whereas Lau et al (2000) reported photoreceptor rescue in tandem with ERG rescue when S334ter rats were treated with rAAV-mediated ¢broblast growth factor 2.

Conclusions

Gene therapy of retinal dystrophies has enjoyed extraordinary progress since the ¢rst encouraging preclinical study was reported just seven years ago. It is likely that the ¢rst human safety trials will be conducted within the next ¢ve years and it is essential that the most favourable gene or genes be chosen for these initial trials. Failures with gene therapy are well known in other ¢elds and ours can ill a¡ord such a setback. This meeting should include a discussion of the genes that are most likely to be amenable to successful transfer into human patients.

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

Aguirre GD, Baldwin V, Pearce-Kelling S, Narfstr m K, Ray K, Acland GM 1998 Congenital stationary night blindness in the dog: common mutation in the RPE65 gene indicates founder e¡ect. Mol Vis 4:23

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

Bhattacharya SS, Wright AF, Clayton JF et al 1984 Close genetic linkage between X-linked retinitis pigmentosa and a restriction fragment length polymorphism identi¢ed by recombinant DNA probe L1.28. Nature 309:253^255

Beltran WA, Kijas JW, Zhang Q, Aguirre GD 2002 Cloning and expression of ciliary neurotrophic factor receptor (CNTFRa) in the canine retina. ARVO (abstr 2679)

Bennett J, Tanabe T, Sun D et al 1996 Photoreceptor cell rescue in retinal degeneration (rd) mice by in vivo gene therapy. Nat Med 2:649^654

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

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

Boulton TG, Stahl N, Yancopoulos GD 1994 Ciliary neurotrophic factor/leukemia inhibitory factor, interleukin6/oncostatin M family of cytokines induces tyrosine phosphorylation of a common set of proteins overlapping those induced by other cytokines and growth factors. J Biol Chem 269:11648^11655

Cayouette M, Behn D, Sendtner M, Lachapelle P, Gravel C 1998 Intraocular gene transfer of ciliary neurotrophic factor prevents death and increases responsiveness of rod photoreceptors in the retinal degeneration slow mouse. J Neurosci 18:9282^9293

Connell G, Bascom R, Molday L, Reid D, McInnes RR Molday RS 1991 Photoreceptor peripherin is the normal product of the gene responsible for retinal degeneration in the rds mouse. Proc Natl Acad Sci USA 88:723^726

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

Dryja TP, McGee TL, Hahn LB et al 1990 Mutations within the rhodopsin gene in patients with autosomal dominant retinitis pigmentosa. N Engl J Med 323:1302^1307

Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVail MM 1990 Photoreceptor degeneration in inherited retinal dystrophy delayed by basic ¢broblast growth factor. Nature 347:83^86

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

Gu SM, Thompson DA, Srikumari CR et al 1997 Mutations in RPE65 cause autosomal recessive childhood-onset severe retinal dystrophy. Nat Genet 17:194^197

Hargrave PA, McDowell JH, Curtis DR et al 1983 The structure of bovine rhodopsin. Biophys Struct Mech 9:235^244

Hauswirth WW, LaVail MM, Flannery JG, Lewin AS 2000 Ribozyme gene therapy for autosomal dominant retinal disease. Clin Chem Lab Med 38:147^153

Hawkins RK, Jansen HG, Sanyal S 1985 Development and degeneration of retinal in rds mutant mice: photoreceptor abnormalities in the heterozygotes. Exp Eye Res 41:701^720

Kedzierski W, Lloyd M, Birch DG, Bok D, Travis GH 1997 Generation and analysis of transgenic mice expressing P216L-substituted rds/peripherin in rod photoreceptors. Invest Ophthalmol Vis Sci 38:498^509

Kirsch M, Lee M-Y, Meyer V, Wiese A, Hofmann H-D 1997 Evidence for multiple, local functions of ciliary neurotrophic factor (CNTF) in retinal development: expression of CNTF and its receptor and in vitro e¡ects on target cells. J Neurochem 68:979^990

Kumar-Singh R, Farber DB 1998 Encapsidated adenovirus mini-chromosome-mediated delivery of genes to the retina: application to the rescue of photoreceptor degeneration. Hum Mol Genet 7:1893^1900

Lau D, McGee LH, Zhou S et al 2000 Retinal degeneration is slowed in transgenic rats by AAVmediated delivery of FGF-2. Invest Ophthalmol Vis Sci 41:3622^3633

Lem J, Krasnoperova NV, Calvert PD 1999 Morphological, physiological and biochemical changes in rhodopsin knockout mice. Proc Natl Acad Sci USA 96:736^741

LaVail MM, Unoki K, Yasumura D, Matthes MT, Yancopoulos GD, Steinberg RH 1992 Multiple growth factors, cytokines and neurotrophins rescue photoreceptors from the damaging e¡ects of constant light. Proc Natl Acad Sci USA 89:11249^11253

LaVail MM, Yasumura D, Matthes MT et al 1998 Protection of mouse photoreceptors by survival factors in retinal degenerations. Invest Ophthalmol Vis Sci 39:592^602

LaVail MM, Yasumura D, Matthes MT et al 2000 Ribozyme rescue of photoreceptor cells in P23H transgenic rats: long-term survival and late-stage therapy. Proc Natl Acad Sci USA 97:11488^11493

Lewin AS, Drenser KA, Hauswirth WW 1998 Ribozyme rescue of photoreceptor cells in a transgenic rat model of autosomal dominant retinitis pigmentosa. Nat Med 4:967^971

Liang F-Q, 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 Marlhens F, Bareil C, Gri¡oin JM et al 1997 Mutations in RPE65 cause Leber’s congenital

amaurosis. Nat Genet 17:139^141

McWilliam P, Farrar GJ, Kenna P et al 1989 Autosomal dominant retinitis pigmentosa (ADRP): localization of an ADRP gene to the long arm of chromosome 3. Genomics 5:619^622

Nathans J, Hogness DS 1984 Isolation and nucleotide sequence of the gene encoding human rhodopsin. Proc Natl Acad Sci USA 81:4851^4855