Retinitis pigmentosa and allied conditions today: a paradigm of translational research

Monogenic human retinal dystrophies are a group of disorders characterized by progressive loss of photoreceptor cells leading to visual handicap. Retinitis pigmentosa is a type of retinal dystrophy where degeneration of rod photoreceptors occurs at the early stages. At present, there are no available effective therapies to maintain or improve vision in patients affected with retinitis pigmentosa, but post-genomic studies are allowing the development of potential therapeutic approaches. This review summarizes current knowledge on genes that have been identified to be responsible for retinitis pigmentosa, the involvement of these genes in the different forms of the disorder, the role of the proteins encoded by these genes in retinal function, the utility of genotyping, and current efforts to develop novel therapies.


Introduction
Human retinal dystrophies (RD) are a group of disorders characterized by a primary and progressive loss of photo receptor cells leading to visual handicap. Monogenic RD are rare diseases. The most common form of the disease, retinitis pigmentosa (RP), is characterized by primary degeneration of rod photoreceptors and has an estimated prevalence of around 1 in 4,000 [14], although higher frequencies have been reported in some Asian popu lations (1 in 930 in South India [5], and approximately 1 in 1,000 in China [6]). RP constitutes 85 to 90% of RD cases.
The first symptoms of RP are retinal pigment on fundus examination, and night blindness, followed by progres sive loss in the peripheral visual field, eventually leading to legal blindness after several decades. The clinical aspects of RP are shown in Table 1. The clinical presentation can be macular, cone or conerod dystrophy (CORD), in which the decrease in visual acuity pre dominates over the visual field loss, or it can be the only symptom. Cone dystrophy is an inherited ocular disorder characterized by the loss of cone cells, which are the photoreceptors responsible for central and color vision. Typically, age of onset is early teens, but it can be very variable, ranging from congenital forms of the disease (Leber's congenital amaurosis (LCA)) to lateonset RD.
RP is usually nonsyndromic (70 to 80%), but there are also more than 30 syndromic forms, involving multiple organs and pleiotropic effects, the most frequent being Usher syndrome (USH; approximately 15 to 20% of all RP cases). USH associates RP with sensorineural deafness and sometimes vestibular dysfunction. The second most common syndromic form is BardetBiedl syndrome (BBS), which accounts for 20 to 25% of syndromic forms of RP or approximately 5% of cases of RP. Patients with BBS typically present with RP, obesity, polydactyly, renal abnormalities and mild mental retardation.
It is worth noting that USH and BBS are genetically as heterogeneous as isolated RP. To date, nine genes have been identified for USH and 14 for BBS. The existence of patients lacking mutations in any of the identified genes indicates that at least one more gene remains unidentified for both syndromes.
Other syndromic forms of RP include associations with hearing loss and obesity (Alström syndrome), dysmor phic face and kidney deficiency (SeniorLocken syndrome), and metabolic disorders [7]. Table 2 shows the most common disorders involving nonsyndromic and syn dromic RP.
inheritance have been reported for nonsyndromic cases of RP and their families depending on the geographical origin, the sample size of the study and the methods for clinical ascertainment. A reliable estimate for the percentages of each inheritance pattern could be 15 to 25% for autosomaldominant RP (adRP), 35 to 50% for autosomalrecessive RP (arRP), 7 to 15% for Xlinked RP (xlRP), and 25 to 60% for syndromic RP [9] (Table 3).
However, wellknown genetic phenomena that alter Mendelian inheritance have also been observed in RP. Incomplete penetrance [10] and variable expressivity have been reported in many families with RP. The literature offers many examples of variable degrees of severity of RP among members of the same family carrying the same mutation [11]. In xlRP forms, female carriers sometimes present RP symptoms and can be as affected as male carriers. One explanation for this might be lyonization, that is, the random inactivation of one X chromosome in females to compensate for the double X gene dose during early developmental stages. The inactivation of the X chromosome not carrying the mutation in a cell or cell population that will later develop into the retina could lead to an active mutated RP gene in the female carrier.

Genes involved in retinitis pigmentosa
The overwhelming pool of genetic data that has become available since the identification of the first mutation associated with RP in humans (a proline to histidine change at amino acid position 23 in rhodopsin, reported by Dryja et al. in 1990 [12]) has revealed the genetics of RD to be extremely complex. Research into the molecular causes of RD has revealed the underlying disease genes for about 50% of cases, with more than 200 genetic loci described [13]. These genes are responsible not only for RP, but also for many other different clinical entities such as LCA, macular degeneration and CORD. To date, 26 genes have been identified for arRP and 20 for adRP, and two genes on the X chromosome (xlRP). For a number of these genes, some mutations in the same gene lead to autosomaldominant forms, while some other mutations lead to autosomalrecessive forms. Different mutations in several genes lead to syndromic forms such as USH or isolated RP (USH2A gene) or non syndromic deafness (MYO7A, CH23, PCDH15, USH1C and USH1G), and mutations in the same gene can cause different clinical entities, as has been observed for ABCA4, which is implicated in arRP, autosomalrecessive macular dystrophies (arMD) and autosomalrecessive CORD (arCORD). Furthermore, most of the mutations causing RP are exclusive to one or a few individuals or families. Common mutations and hot spots are rare; therefore, there is a need for large and timeconsuming mutation screenings to achieve a molecular diagnosis of RP in patients. In addition, there is no clear genotype phenotype correlation and, in many cases, relatives  bearing the same mutation display very different forms of RP in terms of age of onset and severity. Many genes and proteins are associated with RD. These proteins are involved in retinal functions, but they can also play other roles such as degradation of proteins in the retinal pigment epithelium, and ionic interchange or trafficking of molecules in the ribbon synapse of photo receptors. Tables 4, 5, 6 and 7 summarize the genes involved in RD, their chromosomal locations and func tions, and the proteins they encode. The major pathways involved in pathogenesis of RP are discussed below.

Phototransduction
Phototransduction is the process through which photons are converted into electrical signals. It begins with the lightinduced isomerization of the ligand of rhodopsin, which is 11cis retinal, and the activation of rhodopsin. Rhodopsin undergoes a change in conformation upon photoexcitation and activates the G protein transducin. GDPbound inactive transducin exchanges GDP for GTP, and GTPbound active transducin increases the activity of cGMP phosphodiesterase. The result is decreased levels of cGMP in the cytoplasm, and this causes the closing of cGMPgated ion channels and leads to mem brane hyperpolarization. The recovery of the photo trans duction process is carried out by the phosphorylation of rhodopsin by a receptorspecific kinase, rhodopsin kinase. The phosphorylated photoactivated rhodopsin is bound by arrestin, thereby terminating activity of the receptor in the signal transduction process. Mutations in the gene encoding rhodopsin (RHO) are responsible for adRP, arRP and dominant congenital stationary night blindness. Mutations in the genes for cGMP phospho diesterase alpha and beta subunits (PDE6A and PDE6B, respectively) are responsible for arRP and dominant congenital stationary night blindness. Mutations in the genes encoding the rod cGMPgated channel alpha and beta subunits (GUCA1A and GUCA1B, respectively) are responsible for arRP, while arrestin (SAG) is involved in Oguchi disease. The genes encoding guanylate cyclase activating protein 1B (GUCA1B) and cone alpha subunit of cGMP phosphodiesterase (PDE6C) are responsible for dominant MD and arCORD, respectively.

Visual cycle
After isomerization and release from the opsin protein, all trans retinal is reduced to alltrans retinol, and it travels back to the retinal pigment epithelium to be 'recharged' . It is first esterified by lecithin retinol acyl transferase and then converted to 11cis retinol by RPE65. Finally, it is oxidized to 11cis retinal before traveling back to the rod outer segment, where it can again be conjugated to an opsin to form a new functional rhodopsin. Many proteins involved in the chemical transformation and transport for retinoids are causative agents of RD. Mutations in the gene that encodes the retinal pigment epitheliumspecific 65kDa protein (RPE65) can cause arRP or autosomal recessive LCA (arLCA); ABCA encodes a retinal ATP binding cassette transpor ter, and mutations lead to a wide variety of clinical symptoms, including arRP, autosomal recessive Stargardt disease and arCORD; the gene IRBP1 encodes the inter photoreceptor retinoid binding protein and mutations cause arRP; LRAT encodes lecithin retinol acyltransferase and mutations cause arRP and arLCA. Mutations in up to 13 different genes involved in the visual cycle lead to different retinal degenerations, highlighting the importance of this biochemical pathway in the physiology of vision.

Phagocytosis of photoreceptor discs
The stacks of discs containing visual pigment molecules in the outer segments of the photoreceptors are con stantly renewed. New discs are added at the base of the outer segment at the cilium, and old discs are displaced up the outer segment and engulfed by the apical processes of the pigment epithelium. They are then broken down by lysis. Photoreceptor outersegment discs are phagocytosed by the pigment epithelium in a diurnal cycle. Among the different proteins involved in this process, only MERTK, the gene encoding cmer protooncogene receptor tyro sine kinase, has been identified as causing arRP.

Retinal development
Retinal cells are specialized neurons structured in layers. Their patterns of connectivity are crucial, and the correct development of these cells is essential for retinal function. This development is regulated by the precise expression of genes in the right cell type and at the right time, and this regulation is mediated by the synergistic/antagonistic action of a limited number of transcription factors. Muta tions in the conerod otxlike photoreceptor homeo box transcription factor (encoded by the gene CRX) are responsible for adRP, adLCA, arLCA and adCORD; mutations in the neural retina leucine zipper (encoded by the gene NRL) can lead to adRP and arRP. Mutations in the gene encoding the tubbylike protein 1 (TULP1) can cause recessive RP or LCA. RAX2 encodes the retina and anterior neural fold homeobox 2 transcription factor, and mutations are responsible for CORD. Mutations in NR2E3 encoding the nuclear receptor subfamily 2 group E3 cause arRP or adRP. Mutations in the genes encoding prominin (PROM1) and semaphorin 4A (SEMA4A) lead to adCORD. OTX2 (encoding orthodenticle homeobox 2 protein) mutations are associated with adLCA. Finally, defects in TSPAN12 (tetraspanin 12) are associated with familial exudative vitreoretinopathy.

Photoreceptor structure
Although the majority of RD phenotypes appear to result from defects at a single genetic locus, at least one form of RP appears to require coinheritance of defects in the unlinked genes RDS, which encodes peripherin/RDS, and ROM1, which encodes retinal outersegment membrane protein 1. These proteins are components of the poly peptide subunits of an oligomeric transmembrane protein complex, which is present at photoreceptor outerseg ment disc rims and is essential for the correct incidence of light into the discs. Another protein, fascin 2, encoded by FSC2, appears to play a role in the assembly or stabilization of inner segment and calycal process actin filament bundles in photoreceptors and probably regulates the inner segment actin cytoskeleton.

Ciliary structure and function
Photoreceptors have an inner segment that contains the cell organelles and an outer segment composed almost exclusively of optic discs. The connecting cilium connects the inner and outer segments. These discs are constantly renewed and a high number of molecules must travel from the inner segment to the outer segment through the connecting cilium. The development and architecture of the connecting cilium, the correct folding of the involved proteins and the links between the cilium and its surrounding region (calycal process, extracellular matrix) have been shown to be essential for retinal function. Furthermore, as cilia are specialized structures present in many other tissues, defects in the protein components of the cilia and chaperones involved in their development can cause not only isolated RD but also conditions that include RD among their symptoms, such as USH or BBS.
Recently, it has been demonstrated that some ciliary proteins act as positive or negative phenotypic modifiers on defects in other proteins. Some examples are: USH2A, which encodes the large extracellular protein usherin, and defects are responsible for arRP and USH; the genes USH1C and DFNB31, which encode the scaffolding proteins harmonin and whirlin; and CDH23 and PCDH15, which encode the cellcell adhesion proteins cadherin 23 and protocadherin 15, respectively.

mRNA splicing
PremRNA splicing is a critical step in mammalian gene expression. Mutations in genes involved in the splicing processes or spliceosome are associated with a wide range of human diseases, including those involving the retina. Among the genes involved in mRNA splicing, mutations in PRPC8 (human homolog of yeast pre mRNA splicing factor C8), PRP31 (human homolog of yeast premRNA splicing factor 31), HPRP3 (human homolog of yeast premRNA splicing factor 3), PAP-1 (PIM1 kinase), TOPORS (topoisomeraseIbinding arginine/serinerich protein) and SNRNP200 (small nuclear ribonucleoprotein, 200 kDa) are associated with adRP [1418], although the mechanisms behind the process remain unclear.

Other functions
Many other genes and proteins are associated with RD. These proteins have a wide spectrum of functions, such as degradation of proteins in the retinal pigmented epithelium (RPE), ionic interchange, trafficking of molecules in the ribbon synapse of photoreceptors and many others. In addition, the functions of some proteins that have been associated with RD are still unknown.

Molecular diagnosis in retinal dystrophies
The first step toward the diagnosis of RD at the molecular level is genotyping; this allows a more precise prognosis of the possible future clinical evolution of RD, and it can be followed by genetic counseling. Moreover, genetic testing is crucial for the inclusion in human genespecific clinical trials aimed at photoreceptor rescue. However, genetic and phenotypic heterogeneity limit mutation detection, rendering molecular diagnosis very complex. While sequencing remains the gold standard, this is costly and time consuming, and so alternative diagnostic approaches have been recently implemented.
One such alternative diagnostic approach is the use of microarray platforms to detect RP mutations. The most widely used are the specificdisease chips for different types of RD. They contain the previously identified muta tions on the responsible known genes Identification rates (identifying at least one diseaseassociated mutation) depend on the geographical origin and ethnicity of the patient, and they currently stand at 47 to 78% for Stargardt disease [1923], 28 to 40% for CORD [21,23,24,25], 28 to 46% for LCA [26,30], 45% for USH 1, and 26% for USH 2 [31,32]. These represent inexpensive and rapid firststep genetic testing tools for patients with a specific RD diagnosis.
In addition, other highthroughput DNA sequencing platforms targeted to hundreds of genes are being developed. They have been designed to contain either genes limited to exons [33] or fulllength retinal disease genes, including introns, promoter regions or both [34]. Other chipbased cosegregation analyses for autosomal recessive forms and LCA have also been designed, but these analyses requires the inclusion of samples from all the members of the family, both healthy and affected [35,36].
Indirect genetic tools for linkage analysis and/or homozygosity mapping are also being used for RD genotyping, mainly for research purposes. However, increasing availability and low costs have made homo zygosity mapping a particularly appealing approach for the molecular diagnosis of RD [37].
The analytical validity of these procedures has been proved. However, their clinical validity remains to be established for every ethnic group, specific array and type of retinal disease. Clinical applications are also somewhat limited due to the fact that many RP genes are still unknown, and mutations may lie outside of commonly tested regions.

Perspectives for future therapeutics
Currently, optical and electronic devices are the only tools available to improve vision in some patients with RD. In the majority of cases, there are no effective therapies available to prevent, stabilize or reverse monogenic RD.
A key goal in developing an effective therapy for RD is the understanding of its pathophysiology, and the identi fication of the molecular events and disease mechanisms occurring in the degenerative retina. Based on advances in knowledge about these processes, several novel therapeutic strategies are currently being evaluated, includ ing pharmacological treatments, gene therapy and cell therapy.
RD disorders are initiated by mutations that affect rod and/or cone photoreceptors and cause subsequent degeneration and cellular death. Consequently, thera peutic strategies are focused on targeting the specific genetic disorder (gene therapy), slowing or stopping photoreceptor degeneration or apoptosis (growth factors or calciumblocker applications, vitamin supplements, endogenous cone viability factors), or even the replace ment of lost cells (transplantation, use of stem or precursor cells).
Before these strategies can be applied to humans, animal models, preclinical studies and appropriately designed human clinical trials are needed to test different treatments and provide information on their safety and efficacy. According to the ClinicalTrials.gov database, 44 interventional clinical trials for RP have been or are being carried out [38].

Pharmacological therapies
Developing an effective pharmacological therapy for RD must be based on the knowledge of the molecular events and major disease mechanisms and the extent to which they overlap. Current therapies target these pathogenic mechanisms.

Vitamin supplementation and chaperone treatments
Results from experimental effects on animal models [39] and a randomized controlled doublemasked clinical trial [40] have suggested possible clinical benefits of vitamin A supplementation in RP. However, the use of these supplements in other genetic forms of RD, such as ABCA4related diseases (arRP, arCORD, and autosomal recessive Stargardt MD), may accelerate the accumulation of toxic lipofuscin pigments in the retinal pigment epithelium, and thus worsen photoreceptor degeneration. As a result, avoidance of vitamin A supplementation is recommended for people with Stargardt disease.
Another viable approach to RP therapy is the use of pharmacological chaperones [41]. Pharmacological chaperones target protein structure, while chaperone inducers (for example, geldanamycin, radicicol and 17AAG) and autophagy inducers (for example, rapa mycin) stimulate degradation, manipulating the cellular quality control machinery. Some studies have suggested that the rod opsin chromophore (11cis retinal) and retinal analogues (for example, 9cis retinal) can act as pharmacological chaperones, whereas rapamycin is effective against the toxic gain of function, but not the dominantnegative effects of mutant rod opsin [41].

Anti-apoptotic therapy and neuroprotection: endogenous cone viability factors and growth factors
The key goals in pharmacological therapy for RD are neuroprotection and the inhibition of proapoptotic pathways, or the activation of endogenous antiapoptotic signaling systems [42]. Neuroprotection of photoreceptor cells is primarily targeted at structural preservation, and also preventing loss of function. The neuroprotective factors include one 'survival' factor (rodderived cone viability factor (RdCVF)) and four different neurotrophic factors (ciliary neurotrophic factor, basic fibroblast growth factor, brainderived neurotrophic factor and nerve growth factor) that delay rod degeneration in some animal models of RP [43].
RdCVF is a protein that increases cone survival. Injections of this protein in p.P23H rats induced an increase in cone cell number and a further increase in the electroretinogram, indicating that RdCVF can not only rescue cones but can also significantly preserve their function [44].
Ciliary neurotrophic factor has shown efficacy in different animal models, and has progressed to phase II/ III clinical trials in earlystage and latestage RP [45]. It has been administered by encapsulated cell technology, which allows the controlled, continuous and longterm administration of protein drugs in the eye, where the therapeutic agents are needed, and does not subject the host to systemic exposure [46].

Gene-based therapy
Many RDassociated genes have been identified and their functions elucidated. Over the past decade, there has been a substantial effort to develop gene therapy for inherited retinal degeneration, culminating in the recent initiation of clinical trials.
A variety of monogenic recessive disorders could be amenable to treatment by gene replacement therapy through the delivery of healthy copies of the defective gene via replicationdeficient viral vectors [47,48]. Pre limi nary results from three clinical trials indicate that the treatment of a form of LCA by gene therapy can be safe and effective. Phase I clinical trials of gene therapy targeting the gene RPE65 [4951] are being conducted in three different medical centers: Moorfields Eye Hospital, UK [49], the Children's Hospital of Philadelphia, USA, [51], and the Universities of Pennsylvania and Florida, USA. [50], For some autosomaldominant forms of RP or LCA, in which expression of a mutant allele has a gainof function effect on photoreceptor cells, or a dominant negative mechanism or a combination of both, gene therapy is likely to depend on efficient silencing of the mutated allele [52]. Genesilencing strategies for these conditions include RNA interference by microRNA based hairpins (Prph2 animal model), short hairpin RNAs (IMPDH1 gene murine model), RNA interference by microRNA combined with gene replacement (transgenic mouse simulating human RHOadRP), and antisense oligonucleotide technologies.

Cell therapy
Adult stem cells isolated from the retinal pigment epithelium at the ciliary body margin can differentiate into all retinal cell types, including photoreceptors, bipolar cells and Müller glia. Animal experiments have shown that, in response to environmental cues, they can repopulate damaged retinas, regrow neuronal axons, repair higher cortical pathways, and restore pupil reflexes, light responses and basic pattern recognition. When transplanted into a damaged retina, the progenitor cells integrate with the retina, forming a protective layer that preserves existing cells and increases photoreceptor density that is, neurogenesis can be fostered by recruit ment of endogenous stem cells into damaged areas or by transplanted stem cells [53].
Clinical trials using human fetal neural retinal tissue and retinal pigment epithelium cells and adult stem cells are in progress. A phase I clinical trial to repair damaged retinas in 50 patients with RP and agerelated macular degeneration has been conducted in India. Phase I clinical trials to repair damaged retinas in patients with RP degeneration have been conducted using autologous stem cells derived from bone marrow, injected either near the cornea or intravitreally (ClinicalTrials.gov NCT01068561). Preliminary results have shown visual improvement.
Additionally, a noninvasive cellbased therapy consisting of systemic administration of pluripotent bonemarrowderived mesenchymal stem cells to rescue vision and associated vascular pathology has been tested in an animal model for RP, resulting in preservation of both rod and cone photoreceptors and visual function [54]. These results underscore the potential application of mesenchymal stem cells in treating retinal degeneration.

Concluding remarks
To date, more than 200 genes associated with RD have been identified; they are involved in many different clinical entities such as RCA, LCA, USH, CORD and MD. The most surprising outcome of these findings is the exceptional heterogeneity involved: a high number of diseasecausing mutations have been detected in most RD genes, mutations in many different genes can cause the same disease, and different mutations in the same gene may cause different diseases. This genetic hetero geneity underlies a high clinical variability, even among family members with the same mutation. The RD genes involve many different pathways, and expression ranges from very limited (for example, expressed in rod photo receptors only) to ubiquitous.
Gaining knowledge of the genetic causes and pathways involved in the photoreceptor degeneration underlying these disorders is the first step in implementing the correct clinical management and a possible prevention or cure for the disease.
An increasing number of clinical trials are exploring different therapeutic approaches with the aim of treating inherited retinal dystrophies. Phenotypic characteriza tion and genotyping are crucial in order to provide patients with potential personalized treatment. Further research into the mechanisms underlying photoreceptor degeneration and retinal cell apoptosis should also bring us closer to the goal of developing efficient and safe therapies.