Statistics from Altmetric.com
For several decades, basic research on acquired and inherited retinal degeneration was substantially based on a variety of animal models, most of them originating from spontaneous mutations, others induced by damaging external agents. In the past few years, however, progress in genetic engineering has led to a rapidly growing number of transgenic animals, mostly mice, carrying constructs that lead to disruption or overexpression of candidate genes for retinal degenerations. On the one hand, these new models constitute a powerful and adaptable tool to investigate the role of specific gene mutations and the resulting cellular defects that finally lead to photoreceptor cell death. On the other hand, they extend the spectrum of animal models suitable for the newly arisen field of retinal somatic gene therapy.
To assist researchers and clinicians interested in the field, this article attempts to provide a structured overview on recently developed transgenic animal models as well as on models based on spontaneous mutations and induced degenerations. In this review the authors focus on animal models for photoreceptor degeneration since the rapidly growing field of models for ganglion cell death merits its own review and would be beyond the scope of this article. Even with this restriction, the abundance of information generated especially in the past few years makes the attempt of a complete overview almost illusory. Therefore, we apologise for omissions or shortfalls extant in this review. Furthermore, we want to point out that some of the model systems described have already been used extensively. We will therefore occasionally not cite original publications but rather reviews dealing with the specific model system.
Finally, we did not incorporate strategies using viral vectors and/or pharmacological substances. The specific deletion or overexpression of genes susceptible for the modulation of photoreceptor apoptosis, however, was included. The authors are aware that these two subjects may not be clearly separated in all cases.
Hereditary animal models
In 1923 Clyde E Keeler discovered a mouse strain lacking photoreceptors.1 He speculated that these animals suffered from an inherited defect of retinal development and called these micer (rodless retina) mice. These mice showed an early and rapid degeneration of the outer retina, which resulted in a single row of remaining photoreceptor nuclei in the central retina by postnatal day 21 (p21). After publishing several articles on his findings, the continuous lack of interest from the research community made him abandon the whole mouse strain in the early 1930s. In 1951, Brückner and colleagues reported severe retinal degeneration in a wild mouse strain found in the surroundings of the city of Basle that was soon named therd (retinal degeneration) mouse.2 This mouse model later became the most extensively studied animal model for human autosomal recessive retinitis pigmentosa (RP). However, it took four more decades to elucidate the genetic defect underlying photoreceptor degeneration in the rd mouse: a nonsense mutation in the rod photoreceptor cGMP phosphodiesterase β subunit gene.3 4In 1993, 70 years after Keeler's original observations on ther mouse, Pittler and co-workers demonstrated by polymerase chain reaction (PCR) analysis ofr DNA isolated from old histological sections that the defects in the r mouse and the rd mouse were identical.5
Apart from the rd mouse, a number of hereditary retinal defects have been identified subsequently not only in rodents but also in other animals such as cats,6-10dogs,11-14 and chickens15 16 (Table 1). Some of these represent mutations that are also found in human autosomal recessive RP11 12 14 17 18 or X linked congenital stationary night blindness (CSNB) while others carry mutations in rod structural proteins19 20 or more complex syndromes such as Usher type I.21 Therds (retinal degeneration slow) mouse, for instance, carries an insertional mutation in therds/peripherin gene encoding for a photoreceptor structural protein expressed both in rods and cones.22 Consequently, rod and cone photoreceptor outer segments in homozygous rds mice never develop and photoreceptor nuclei start to die by apoptosis as early as by p21 and at the age of 12 months, the entire outer nuclear layer (ONL) has disappeared.19 Interestingly, the pattern of photoreceptor cell loss is peripheral to central in therds mouse which is in marked contrast to therd mouse where photoreceptor degeneration starts in the central retina progressing towards the periphery.19
Transgenic animal models
In the past two decades our understanding of the molecular events leading to human retinal dystrophy has improved markedly. The history of the identification of genetic loci for inherited retinal degeneration started in the early 1980s when Bhattacharyaet al mapped the gene responsible for X linked retinitis pigmentosa to a subregion of the X chromosome.23 In 1989, Humphries and his co-workers identified the first autosomal dominant RP locus24 and only a year later, Dryja and his group described the first mutations within the rhodopsin gene in patients with autosomal dominant retinitis pigmentosa25 26 followed by numerous other groups (for review see Gal et al 27). Consecutively, mutations in other photoreceptor specific genes such as the β subunit of the cGMP phosphodiesterase,4peripherin,22 and the rod outer segment protein 1 (ROM-1)28 were described.
The era of transgenic animals had begun just a few years earlier when in 1981 Wagner et al performed the first successful transgenesis by transplanting a rabbit β globin gene into a mouse embryo.29 Therefore, the retinal research community was among the first to profit from the new possibility of designing specific transgenic animals mimicking genetically caused human disease; disruption or overexpression of the target gene allowed the investigation of the role of a single specific gene product on retinal function in vivo.
Transgenic mice and rats30-32 are among the most commonly used animals to date; nevertheless, other species such as pig33 may also be very useful models since they may show a cone distribution similar to the human eye (Table 2). Furthermore, in the transgenic pig, being a larger mammal, the time course of retinal degeneration may resemble the human disease more closely. Finally, the large size of the pig eye is very well suited to subretinal injections and somatic gene therapy.
Transgenic models for retinal degenerations and dystrophies may be divided into two major subgroups.
TRANSGENIC MODELS MIMICKING HUMAN RP AND/OR MODULATING PHOTORECEPTOR PHYSIOLOGY
Soon after the identification of the first gene mutations leading to RP in human, attempts were made to create transgenic animals carrying analogous mutations. To our knowledge, the earliest transgenic mouse generated for retinal research was the Pro23His rhodopsin mutant mouse created in 1992 by Olsson et al. 34 In the following years, a number of different animal models carrying rhodopsin mutations mimicking autosomal dominant RP were generated in various species.20 30-41
The rds mouse was used for the generation of double mutant mice carrying a peripherin mutation on anrds null background,42 leading to an acceleration of the time course of retinal degeneration observed in the rds mouse.
Several groups investigated the role of gene products playing key parts in photoreceptor physiology. For example, the lack of interphotoreceptor retinoid binding protein (IRBP), responsible for the buffering of retinoids in the extracellular matrix, led to slow photoreceptor degeneration inIRBP −/− mice.43
Among the proteins necessary for phototransduction, the role of the γ subunit of the rod cGMP phosphodiesterase was investigated through the generation of mice lacking a functional PDEγ gene.44These mice showed a rapid retinal degeneration starting as early as p5.
Another very promising recent finding does not primarily implicate photoreceptor cells but rather the retinal pigment epithelium (RPE) where the RPE65 protein is an essential element for vitamin A metabolism. In humans, mutated RPE65 protein not only leads to autosomal recessive, childhood onset severe retinal dystrophy45 but also to autosomal recessive RP46 and Leber's congenital amaurosis.46 47In analogy to the human mutation, a four nucleotide deletion in the RPE65 gene was identified in Swedish briard dogs48 49 as the cause of CSNB50 but also for autosomal recessive, early onset and progressive retinal dystrophy.50 51Transgenic Rpe65−/− mice, because of a lack of rhodopsin in the outer segment discs, show no rod function as demonstrated by electroretinography and seem to be an excellent model for an all cone retina.52 Furthermore, photoreceptors are completely protected from light induced degeneration indicating that rhodopsin is the key trigger in the initiation of the signalling pathway from light to photoreceptor death by apoptosis.53
McCall and co-workers generated a transgenic animal called the rdta mouse.54 This animal expresses the gene for an attenuated diphtheria toxin—under control of a portion of the rhodopsin promoter—which inhibits G protein binding to rhodopsin. Expression of this transgene results in the elimination of rod photoreceptor cell bodies in the ONL and the elimination of any rod mediated ERG responses by p17.54
Recently, Biel et al demonstrated that deletion of the cyclic nucleotide gated channel 3 (CNG3) led to a selective loss of cone function. This all rod retina may serve as a model for human achromatopsia.55
The Rim protein (RmP) is an ABC transporter protein in rod photoreceptor outer segment discs. Recently, Weng and collaborators reported the generation of a mouse carrying a null mutation in the rim protein gene (ABCR).56 abcrknockout mice show increased levels of all-trans-retinaldehyde following light exposure and a dramatic accumulation of the lipofuscin fluorophore A2-E. With increasing age, the loss of RPE cells leads to secondary photoreceptor degeneration. abcr knockout mice may represent a model for human recessive Stargardt's disease and may also provide insights into the pathogenesis of age related macular degeneration (AMD).
Humphries and collaborators designed the rhodopsin knockout mouse as a model serving two different purposes—firstly, to provide a genetic background on which other mutant opsin transgenes may be expressed and, secondly, to study the introduction of functional rhodopsin genes through somatic gene therapy.rho−/− mice do not develop rod outer segments (ROS) and photoreceptor degeneration starts as early as p24. At 8 weeks of age, no scotopic electroretinogram (ERG) response can be detected.57
Sieving and collaborators recently generated a mouse model for autosomal dominant CSNB by crossing rhodopsin G90D transgenic mice with rhodopsin knockout (rho−/− ) mice.58 In addition, mice lacking rhodopsin kinase (RK)59 as well as arrestin knockout mice60may both serve as models for Oguchi's disease, a human condition that causes autosomal recessive CSNB.RK−/− mice reared in cyclic light show a 50% shortening of ROS by 6 weeks of age whereas arrestin knockout mice show a progressive thinning of the ONL starting at p100.
TRANSGENIC MODELS MODULATING REGULATORY GENES OF APOPTOSIS
Apoptosis is a tightly regulated form of cell death that occurs physiologically during organ development in the retina61and other tissues but also in a variety of pathological conditions such as cancer and degenerative disorders. In the retina, apoptosis is the final common pathway of photoreceptor cell death not only in transgenic animal models for RP20 41 and the model of light induced photoreceptor degeneration62 63 but also in human RP.64 Inhibition of apoptosis by modulating potential regulatory genes may therefore be a means to retard or even stop the time course of retinal degeneration.
THE BCL-2 FAMILY
The mammalian cell death suppressor gene Bcl-265 is a member of a group of proteins involved in the regulation of apoptosis.66 Both death antagonists (for example, Bcl-2, Bcl-XL, Bcl-w, Bfl-1, Bag-1, Mcl-1, A1) and agonists (for example Bax, Bak, Bcl-XS, Bad, Bid, Bik, Hrk) belong to the Bcl-2 family. Data on the role of Bcl-2 in regulating photoreceptor apoptosis are ambiguous. Although overexpression of Bcl-2 delayed apoptotic photoreceptor death in several RP animal models, introduction of the transgene did not influence the final outcome of photoreceptor degeneration67-69 and specific overexpression of human Bcl-2 in Müller cells led to early postnatal Müller cell apoptosis followed by photoreceptor degeneration in a transgenic mouse line.70 Similarly, Bcl-XL, a potent inhibitor of apoptosis in many cell types, was unable to efficiently inhibit photoreceptor apoptosis when overexpressed in therd mouse.67 68
Little is known about the role of other Bcl-2 family members in the retina: the role of the pro-apoptotic Bcl-2 family member Bax on photoreceptor apoptosis was studied in Bax deficient mice. These studies indicated that Bax is involved in the control of developmental photoreceptor apoptosis in wild type mice but not of photoreceptor degeneration in the rd mouse.71
The tumour suppressor gene p53, a sequence specific DNA binding transcription factor, is an important regulator of apoptosis in a variety of systems and tissues. However, both p53 dependent and p53 independent apoptosis has been described. In the retina, the lack of p53 did not protect from photoreceptor apoptosis in the model of light induced retinal degeneration72 whereas it delayed photoreceptor death by 3 days in the rdmouse.73
The proto-oncoprotein c-Fos is a member of the AP-1 transcription factor complex involved in the regulation of apoptosis in many systems. In contrast with p53, lack of c-Fos completely protected photoreceptors from light induced apoptosis.63 74 but had no effect on the degeneration process in the rdmouse.75
AP-1 may be constituted of members of the Fos and Jun families of proteins. In the retina, AP-1 primarily consists of c-Fos and junD.76 We therefore also investigated the role of JunD in light induced photoreceptor degeneration. However, no difference in the extent of light induced photoreceptor apoptosis was found betweenjunD−/− andjunD+/+ mice.77
Inducible animal models
In the early 1960s Werner Noëll started to investigate the effect of light on inherited retinal degeneration and in 1965 he showed that photoreceptor degeneration in the Royal College of Surgeon (RCS) rat, a model for inherited retinal dystrophy, was dramatically enhanced when animals were exposed to light: “On the basis of this reasoning, I cannot help but be impressed by the fact that excessive light destroys the visual cells and the pigment epithelium of albino as well as pigmented animals”.78 Since then, it has been postulated that light exposure may enhance inherited photoreceptor degeneration. In the past few years, a number of experimental studies performed in new transgenic models has further supported this hypothesis56 59 60 79-81 and it was also hypothesised that light exposure may enhance the progression of RP in humans.82 83
It was conceivably based on his earlier findings in the RCS rat (D Organisciak, personal communication) that Noëll developed the first inducible animal model for retinal degeneration by damaging photoreceptors in wild type rats using bright light.84 The advantages of an external damaging stimulus are evident. Animals can mature and develop a normal retina until the stimulus is provided; the latter being very flexible and adjustable for timing and intensity so that more or less severe photoreceptor damage can be obtained. Such a tightly tunable model therefore is an excellent tool to investigate the molecular stages of apoptotic photoreceptor death. Light induced photoreceptor degeneration is nowadays used as a model in its own right by a number of groups in a variety of different light conditions in wild type and transgenic animals62 63 85-95 and it has been shown repetitively that photoreceptor death in these models occurs by apoptosis (see Table3).62 63 85-87 94
Another well known method of inducing selective photoreceptor apoptosis is the intraperitoneal administration of N-methyl-N-nitrosourea (MNU). Depending on its dose, MNU leads to a rapid degeneration of photoreceptors leaving other cell layers unaffected.96-100 However, the molecular mechanisms underlying MNU induced photoreceptor apoptosis are not known yet. Other systems of induced photoreceptor damage include the implantation of iron particles into the vitreous101 102 or the intravitreal injection of l-ornithine hydrochloride.103
Currently, many of the animal models described in this review play a more important part than ever, serving a variety of different purposes: firstly, they are used for the study of the molecular mechanisms leading to photoreceptor degeneration. Secondly, they are the basis for therapeutic attempts to retard or even stop photoreceptor apoptosis using a multitude of different approaches. Finally, they serve as a platform for the establishment of experimental somatic gene therapy, probably the most promising approach to photoreceptor rescue (reviewed by Sharma and Ehinger,104 Aliet al,105 and Chong and Bird106).
Although still restricted and fragmentary, our knowledge about molecular mechanisms leading to photoreceptor degeneration increased markedly in the past two decades. It is therefore conceivable to speculate that ophthalmology may be among the first clinical specialties to benefit from “molecular medicine” once new therapeutic strategies evolve.
The authors thank T P Williams for critical reading of the manuscript. Supported by the Swiss National Science Foundation, No 31-40791.94, Sandoz Foundation, Basle, Switzerland, Bruppacher Foundation, Zurich, Switzerland, and Ian and Caroline Leaf and family, Gland, Switzerland.