|Home | About | Journals | Submit | Contact Us | Français|
The advent of gene editing has introduced the ability to make changes to the genome of cells, thus allowing for correction of genetic mutations in patients with monogenic diseases. Retinal diseases are particularly suitable for the application of this new technology because many retinal diseases, such as Stargardt disease, retinitis pigmentosa (RP), and Leber congenital amaurosis (LCA), are monogenic. Moreover, gene delivery techniques such as the use of adeno-associated virus (AAV) vectors have been optimized for intraocular use, and phase III trials are well underway to treat LCA, a severe form of inherited retinal degeneration, with gene therapy. This review focuses on the use of gene editing techniques and another relatively recent advent, induced pluripotent stem cells (iPSCs), and their potential for the study and treatment of retinal disease. Investment in these technologies, including overcoming challenges such as off-target mutations and low transplanted cell integration, may allow for future treatment of many debilitating inherited retinal diseases.
The human eye is composed of an anterior segment, which facilitates entry and refraction of light, and a posterior segment, which absorbs light and sends visual signals to the brain. The anterior segment contains the cornea, iris, ciliary body, and lens while the posterior segment consists of the vitreous humor, retina, choroid, and optic nerve (Figure 1). Damage to the posterior segment, such as degeneration of photoreceptors and retinal pigment epithelial (RPE) cells in the retina can lead to poor clinical outcomes.
Research in posterior segment eye diseases is often hindered by a lack of tissue samples. Because the mammalian retina does not regenerate after damage, biopsy of the posterior segment can lead to significant visual consequences and is avoided in living patients. Diseased eye tissues for research are often obtained postmortem at advanced stages of disease. Therefore, the ability to derive induced pluripotent stem cells (iPSCs) from donors with and without ophthalmic disease expands options for research on these diseases. When combined with gene editing, iPSCs hold enormous potential for the study of eye diseases and development of novel therapies .
iPSC lines derived from affected patients can be used with gene editing to study disease through disease modeling, drug screening, and regenerative medicine (Figure 2) [2,3]. iPSC lines can model pathophysiology of genetic disease when derived from affected patients [4,5]. Patient-derived cells can also be used to screen drug libraries and identify therapeutics optimized for specific patients; this is an example of “personalized medicine” . In drug screening, gene editing that targets the mutated gene locus can create controls from patient-derived diseased cells, thus reducing genetic variation and other confounders that may influence phenotypic differences [7,8]. Lastly, in regenerative medicine, gene editing can repair patient-derived iPSCs to generate disease-free autologous transplants for therapeutic use.
Gene editing utilizes nucleases and deoxyribonucleic acid-binding (DNA-binding) domains to specifically target regions of DNA. These strategies include zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs), and clustered, regularly interspaced short palindromic repeats (CRISPR)/CRISPR associated protein 9 (Cas9) systems. ZFNs and TALENs target DNA through protein-DNA interactions with zinc finger proteins  and amino acid repeat variable di-residues , respectively. Both then use endonucleases to cleave DNA. CRISPR technology uses a DNA-targeting domain composed of ribonucleic acid (RNA) and a Cas9 endonuclease effector domain. Because its targeting domain relies on RNA-DNA base pairing rather than protein-DNA interactions, CRISPR/Cas9 systems are more easily customizable than ZFNs and TALENs [11,12].
Many retinal diseases are monogenic; they can be caused by mutations in single genes. Currently, over 250 genes with retinal disease-causing mutations have been identified . These diseases are easier to target with gene editing because only one locus must be repaired. Thus far, research involving gene editing and iPSCs has targeted retinitis pigmentosa (RP) and Leber congenital amaurosis (LCA) because these diseases are monogenic. RP is a genetic disorder of the retina characterized by progressive loss of rod photoreceptors resulting in night blindness, loss of peripheral vision, and blindness, that can be caused by mutations in over 90 genes implicated in disease pathogenesis . LCA is an inherited retinal dystrophy characterized by severe visual impairment, photophobia, nystagmus, and hyperopia caused by mutations in genes involved in the visual cycle . The retina itself is a desirable target for gene-based therapies, as its response to therapy can be imaged over time. It is also inherently immune-privileged, leading to decreased host immune responses against vehicles injected into the eye to deliver gene editing components .
In vivo gene editing in the eye has been conducted successfully in mice. In one study, adeno-associated virus serotype 8 (AAV8) vectors delivered CRISPR/Cas9 systems causing knockdown of neural retina leucine zipper (Nrl) in mouse models of RP, including rhodopsin knockout mice, RHO-P347S rhodopsin mutant, and phosphodiesterase 6β hypomorphic mutation . In these RP models, Nrl-ablated rods were resistant to degeneration and cone survival increased.
In vivo gene editing may also provide treatment for retinal neovascular diseases, which have diverse causes including proliferative diabetic retinopathy, wet age-related macular degeneration (AMD), and retinopathy of prematurity. Existing treatment for these diseases involve long-term therapy with biologics targeting vascular endothelial growth factor (VEGF) [18-20]. In a recent study, intravitreal injection of AAV with a CRISPR/Cas9 system targeting VEGF receptor 2 (VEGFR2) led to decreased neovascularization and VEGFR2 expression in murine models of oxygen-induced retinopathy and laser-induced choroidal neovascularization .
Use of gene editing in vivo in humans is also on the horizon. Editas Medicine and Allergan are hoping to target a variant of LCA, LCA10, with gene editing in clinical trials . LCA10 is caused by mutations in centrosomal protein 290 (CEP290), which encodes for a centrosomal protein localized in the connecting cilium of photoreceptors . The IVS26 c.2991+1655 A>G mutation in CEP290, which causes the insertion of a cryptic exon with a premature stop codon in intron 26, is found in about 21 percent of patients with LCA . While clinical trials are still in development, past work by this group usedCRISPR/Cas9 to excise the disease-causing CEP290 mutation in primary fibroblasts derived from patients homozygous for the IVS26 mutation [25,26]. This correction led to increased amounts of wildtype CEP290 protein and correction of ciliogenesis. Treatment of visual diseases in vivo is a promising area to begin research on gene editing therapies.
The eye has already served as a target of gene-based approaches due to the monogenic nature of many retinal diseases and accessibility of the eye. AAV vectors  and lentiviral vectors  have delivered gene therapy to the eye in humans. AAV vectors have a size limit of 5.2 kilobases (kb), which precludes delivery of complementary DNA of larger proteins . However, dual AAV therapies have been developed that capitalize on concatemerization of AAV genomes through splicing or homologous recombination to deliver longer coding sequences [30,31]. Lentiviral vectors accommodate cassettes of 8 to 10 kb, allowing for transduction of larger genes . However, lentiviral vectors integrate into the genome of transduced cells, presenting the possibility of insertional mutagenesis.
Stargardt disease, an inherited macular dystrophy caused by mutations in the ATP-binding cassette, sub-family A, member 4 (ABCA4) gene, has been targeted by gene therapy [33,34]. Phase I/II clinical trials to treat Stargardt disease with gene therapy are ongoing (ClinicalTrials.gov Identifier: NCT01367444, NCT01736592). These studies involve subretinal injection of SAR422459, a lentiviral vector, to deliver a functioning ABCA4 gene to photoreceptors of patients homozygous for ABCA4 mutations and with significant visual impairment. These studies will assess safety and delay in retinal degeneration in the short- and long-term.
Gene editing offers several advantages over gene therapy because the corrected gene is inserted at the endogenous gene locus allowing for endogenous transcriptional control. In contrast, AAV vectors remain episomal in the nucleus. Lentiviral integration sites, while not random, are very diverse . In addition, gene editing causes permanent modification of the genome, leading to stable protein expression. Gene editing can build on knowledge derived from gene therapy experiments. For example, concerns over the AAV vector size limit led to the discovery of Staphylococcus aureus Cas9, which is smaller in size than Streptococcus pyogenes Cas9 . The smaller size of S. aureus Cas9 enables larger genes to be co-packaged with Cas9 within one AAV, leading to the decreased need for dual AAV approaches.
While gene editing shows promise for retinal disease treatment in vivo, gene editing can also be used with iPSCs to study disease pathophysiology in vitro. Patient-derived iPSCs can model diseases and serve as drug screening tools if they display a disease phenotype. Protocols for the generation of retinal pigment epithelial (RPE) cells from iPSCs have been optimized to generate well-differentiated cell lines, making patient-derived iPSCs well-suited for study of diseases affecting the RPE [37-40]. Differentiation of iPSCs into pure populations of photoreceptor-like cells has also recently been refined . iPSC-derived cell lines from patients affected by variants of RP , LCA [5,42], Usher syndrome , Best disease , and gyrate atrophy  have been generated and used to understand the pathophysiology of these disorders. Because RP and LCA10 are currently targets of gene editing-based therapies, iPSC modeling of these diseases will be covered here.
iPSC-derived photoreceptor cells have helped elucidate the pathophysiology of X-linked RP caused by retinitis pigmentosa GTPase regulator (RPGR) mutations. In a recent study, iPSC-derived photoreceptor cells were generated from patients with RPGR mutations and X-linked RP . Through co-immunoprecipitation studies, it was found that RPGR mutations led to disrupted binding of RPGR with gelsolin, a protein involved in filamentous actin turnover. Decreased gelsolin activity led to increased actin polymerization, mislocalization of rhodopsin, and photoreceptor degeneration. In this study, control iPSCs were derived from unaffected family members.
iPSC-derived models contribute to the study of LCA10 pathophysiology. iPSCs generated from fibroblasts of a patient homozygous for the IVS26 c.2991+1655 A>G mutation in CEP290 were differentiated into optic cups . These optic cups demonstrated decreased CEP290 production and mislocalization of RPGR leading to disrupted ciliogenesis. Splicing defects caused by the mutation were preferentially exacerbated in photoreceptor cells as compared to fibroblasts and RPE, possibly due to photoreceptor-specific splicing mechanisms. In this study, control iPSCs were generated from commercial fibroblasts from unaffected individuals.
In the above experiments seeking to describe disease phenotypes, control cell lines were derived from different patients without disease. However, this introduces confounders such as genetic background, differences in the retinal cell differentiation process, or genetic alteration introduced using iPSC reprogramming . As the use of gene editing expands, future experiments could modify diseased cells to generate control cells with the same genetic background as affected cells, thus negating many of the concerns stated above.
In theory, iPSC systems allow for generation of large numbers of diseased cells for drug screening. However, large-scale production of iPSC-derived RPE cells is still challenging due to the complex protocols involved and growth factors required for differentiation [47,48]. As such, large drug screens thus far are not widely used in the study of eye disease. Research is ongoing to develop more efficient methods to generate iPSC-derived RPE cells. For example, chetomin and nicotinamide have been identified as small-molecule inducers that allow for generation of RPE-like cells from iPSCs in a one-exposure step . This protocol enables production of larger quantities of pure iPSC-derived RPE cells and monolayers for screening experiments.
In one study of iPSC-derived RPE cells from patients with AMD, screening of several antioxidant drugs revealed curcumin to be protective of RPE cells from oxidative stress-induced cell death . In another study, iPSC-derived rod photoreceptors from RP patients with mutations in retinitis pigmentosa 1 (RP1), retinitis pigmentosa 9 (RP9), peripherin 2 (PRPH2), or rhodopsin (RHO) genes, were used to screen the vitamins alpha-tocopherol, ascorbic acid, and beta-carotene for protective effects. Alpha-tocopherol was found to be protective of rod photoreceptor cells from patients with RP9 mutation, but not in those from patients with other mutations .
In these studies, control cells were not matched using cells from the same patient, nor were the controls age-matched. However, by repairing mutations in cells from affected patients, gene editing can generate control cells that differ only in one gene. This technology provides the most rigorous comparison of control and diseased cells.
iPSCs can potentially create specific cells for autologous transplantation. Other than one clinical trial using iPSC-derived RPE cells, clinical trials have used only human embryonic stem cell (hESC)-derived cells [52,53]. hESC lines are derived from the inner cell mass of unused preimplantation embryos from in vitro fertilization . hESC-derived RPE cells have been utilized in several clinical studies [55,56]. hESC use is more prevalent than iPSC use because deleterious mutations can occur during the extensive culture process in which somatic cells are reprogramed into iPSCs and differentiated into RPE cells. This was recently highlighted in a clinical trial where autologous iPSC-derived RPE cells were transplanted into human retinas. While one transplanted subject had stable vision a year after surgery, the second subject’s iPSC-derived RPE cells were not transplanted because they developed copy-number variants and single nucleotide variations, the most concerning of which was found in an oncogene [57,58].
In experimental systems seeking to generate autologous transplants, gene editing could be used to repair gene mutations prior to transplant. Therapeutically, autologous transplants are advantageous due to the lower risk of immune rejection. Gene editing with CRISPR/Cas9 has repaired RPGR gene mutations, a cause of X-linked RP, in iPSCs derived from affected patients . As of the writing of this review, these repaired iPSCs have not yet been differentiated into RPE cells to assess function.
Transplantation of disease-free cells into human retina is well-tolerated. In phase I/II clinical trials, hESC-derived RPE cells were transplanted into one eye of nine patients with Stargardt disease and nine patients with AMD [55,56]. Transplanted cells were well-tolerated for up to 37 months after transplantation.
While gene editing holds promise for the treatment and study of genetic disease, challenges remain in the implementation of this technology. Off-target mutations are possible, creating mutations at unintended regions of the genome . In a recent study, two mice underwent CRISPR/Cas9 repair of a phosphodiesterase 6b mutation and developed off-target mutations . Whole-genome sequencing revealed 117 insertion-deletion mutations and 1,397 single-nucleotide variants in both corrected mice that were not found in the uncorrected mouse. Furthermore, mutations were unpredictable because DNA surrounding these regions had poor homology with the targeting RNA.
Gene editing in regenerative medicine may also be limited by efficiency of transplanted cell integration. While early studies showed integration of transplanted photoreceptor-like cells into the retina , more recent studies conclude that transfer of intracellular contents between donor and host cells, rather than cell integration, is more prevalent after subretinal photoreceptor transplantation [63-65]. These studies suggest that rather than direct photoreceptor replacement, future approaches to transplantation must capitalize on the transfer of intracellular contents to deliver protein or nucleic acid to host photoreceptors.
Gene editing is contributing to the understanding of and therapeutic development for retinal diseases. Gene editing has treated retinitis pigmentosa and retinal neovascularization in mice, offering optimistic outlooks on gene editing in vivo. Clinical trials for gene editing of LCA10 are in development. In addition to in vivo applications, gene editing may be applied to personalized medicine when used with iPSCs. These approaches include generating controls for disease modeling and drug screening, and ex vivo repair of gene mutations in regenerative medicine. While gene editing and iPSCs show promise, further work is required to ensure safety regarding off-target mutations from gene editing and mutagenesis that may occur during the derivation and differentiation of iPSCs. Despite these challenges, gene editing technology has made rapid advances and is a valuable tool in understanding and treating retinal diseases.
This work has been supported in part by an unrestricted/challenge award to Yale Eye Center from the Research to Prevent Blindness (RPB), Inc., New York, NY, USA; and the Foundation Fighting Blindness, Columbia, MD, USA.
Katherine Chuang, BS, prepared a draft of the manuscript; Mark Fields, MPH, PhD, and Lucian Del Priore, MD, PhD, offered guidance on the topic, outlined manuscript construction, provided mentorship during the writing process, and critical revisions of the manuscript.