Human retina-in-a-dish: Unlocking the potential to study mechanisms of inherited retinal disease

Inherited retinal diseases affect the structure and function of the retina, leading to moderate to severe vision impairment. In addition to being clinically diverse, they are also genetically heterogeneous and can be attributed to more than 280 genes.¹ Although the genetics causal to these diseases are largely known, there are limitations in the model systems available to study inherited retinal diseases. Over the past decade, the advancements made in stem cell biology have allowed for somatic cells to be reprogrammed back to a pluripotent state, opening up avenues to develop stem cell lines from patient fibroblasts, peripheral blood mononuclear cells, or other patient samples. Along with these scientific advancements, three-dimensional retinal organoids—a human, developing retina-in-a-dish model—have emerged and provided a system to follow the development and function of the retinal neurons in vitro. In the work by Afanasyeva et al., the authors developed retinal organoids from patients with the inherited retinal disease Leber congenital amaurosis (LCA).²

LCA is an autosomal recessive disorder that manifests during infancy and is the most common cause of inherited blindness in children.^3,4 LCA itself is also genetically heterogeneous and has been associated with mutations in more than 25 genes that are important for photoreceptor and retinal pigment epithelium (RPE) function. One of the most severe forms is linked to a mutation in Leber congenital amaurosis 5 (LCA5). LCA5 encodes lebercilin, a protein found in the connecting cilium between the outer and inner segment of photoreceptors that is essential for the trafficking of proteins and vesicles.⁵ Afanasyeva et al. studied the effects of a nonsense variant of LCA5 in vitro and its recapitulation of the clinical phenotype in patient-derived induced pluripotent stem cells (iPSCs) which were then differentiated into retinal organoids.² Using CRISPR-Cas9 technology, they were able to correct the patient iPSCs to create a wild-type isogenic control and confirm the rescue of the lebercilin protein and its localization and function in the gene-corrected line of retinal organoids.

In their paper, Afanasyeva et al. describe the steps taken to obtain the corrected patient lines by first assessing gRNA efficiency and homology-directed repair (HDR) efficiency. They tested two different gRNAs, as well as two single-stranded oligodeoxynucleotide (ssODN) donor templates that contained the intended edit. After predicting editing efficiency in silico, they used TIDER analysis to determine the frequency of total non-homologous end joining and HDR-mediated editing. Due to Cas9:gRNA1 RNP and ssODN1 showing a higher total editing efficiency of 74% and 58.7% HDR efficiency with the addition of the Alt-R HDR Enhancer molecule, they proceeded with those conditions.

After obtaining two iPSC clonal lines that were successfully corrected on both alleles, they assessed any presence of off-target effects using next-generation whole-genome sequencing (WGS). They predicted in silico the off-target editing sites with the highest homology to the gRNA1 target sequence. Out of nine predicted off-target sites, one lay in a gene-coding region in CFAP44. Upon sequencing, they checked at these sites and found no-off target edits. Next, they searched for genomic alterations, such as structural variants or single nucleotide variants, using an automated bioinformatic detection pipeline. The pipeline detected 121 and 119 structural variants in clone no. 1 and clone no. 2, respectively. However, upon manual examination, they were all shown to be false positives, as they were already present in the parental line prior to editing. Through WGS and karyotyping, they were able to validate that the edited iPSCs showed no off-target effects, presented normally, and had pluripotency.

To study the loss of LCA5 in the patient line and its restored form in the gene-edited line, they differentiated patient, isogenic gene corrected, and unrelated control iPSCs into retinal organoids for 120, 150, and 190 days. Their sample size was limited, and additional clones should be tested; however, in all three lines, they saw no visible concerns in retinal organoid differentiation using light microscopy. To assess lebercilin expression and ciliary localization, they performed immunohistochemistry with lebercilin and ciliary marker-specific antibodies. They did not detect any lebercilin protein in the patient retinal organoids, and excitingly they found the gene correction did restore both lebercilin mRNA and protein expression, and rescued protein localization to the same level of the non-isogenic control.

They also looked at rhodopsin and opsin, proteins found in the photoreceptor outer segments that are crucial for the activation of the phototransduction cascade. In the patient, isogenic control, and non-isogenic control organoids, rhodopsin and opsin were found to be in the outer segment of the photoreceptors. Consistent with the murine model of LCA5, in the patient-derived organoids, rhodopsin was also found in the outer nuclear layer (ONL), reasonably from the loss of lebercilin and disrupted protein transport.⁶ This mislocalization of rhodopsin and opsin in the ONL in the patient organoids, but not the isogenic and non-isogenic control, indicates the potential rescue of lebercilin function.

This exciting work shows the emerging power of gene editing technology in investigating disease models that were previously difficult to study. In obtaining patient lines and creating isogenic controls, scientists can control for the same genetic background to tease out what solely is due to the mutation or due to other genomic alterations. Patient-derived iPSCs serve as an extremely valuable resource for disease modeling and studying cellular and molecular mechanisms to gain insights in how diseases develop. An additional benefit for using patient-derived iPSCs is that obtaining cells is minimally invasive, only requiring blood cells or skin fibroblasts. In the field of ophthalmology, retinal organoids have also emerged to be the most ideal in vitro model system for studying retinal development as they can capture a high level of complexity that cannot be recreated in photoreceptor- or RPE-specific cell lines (Figure 1). Retinal organoids can form all the different cell subtypes of the retina and can closely resemble similar proportions of cellular compositions.⁷ Although they are incredibly useful for studying cellular and molecular mechanisms behind the pathophysiology of development and disease, their utility is restricted as they cannot be used to fully evaluate visual function.

Figure 1.

Human stem cell-derived retinal organoids for the study and treatment of inherited retinal disease

The potential applications available for human stem cell-derived retinal organoids are shown in circles. The two green circles highlight key advances made in the work by Afanasyeva et al.² Genetic engineering was utilized to develop isogenic retinal organoids from LCA5 Leber congenital amaurosis patient stem cells. These isogenic retinal organoids were then studied and LCA5 impacts on protein trafficking between inner and outer segments were noted. Created with BioRender.com.

Many inherited retinal diseases are autosomal recessive, making it an ideal candidate for gene therapy. However, LCA5 only accounts for less than 2% of LCA cases and is a particularly severe variant of LCA.^8,9 Since retinal neurons are post-mitotic, once they have been lost, they cannot be regenerated. With LCA5-LCA, the onset of vision loss occurs at birth or within the first few months of life, restricting therapeutic options.¹⁰ Determining the efficacy of gene therapy will be limited by those existing cells and the focus will need to be on protecting those cells and preventing further degeneration. Despite the translational limitations, this study is a step in the right direction by depicting one of the many usages of CRISPR-Cas9 technology and its potential to allow scientists to create models of the same genetic background to study disease.