RPE65 Variant p.(E519K) Causes a Novel Dominant Adult-Onset Maculopathy in 83 Affected Individuals

Abstract

Purpose: Recessive RPE65-associated retinopathy is a well-known target for gene therapy, whereas dominant RPE65-associated retinopathy, due to the Irish founder variant p.(D477G), has been reported only once until now and is very rare. Here, we present the discovery of a novel, second dominant RPE65-associated retinopathy caused by variant c.1555G>A, p.(E519K).

Methods: Genomic data was investigated in a Belgian discovery cohort (n = 2873) and an international replication cohort (n = 18,796) with inherited retinal disease (IRD). Heterozygous p.(E519K) individuals underwent extensive phenotyping. Haplotype phasing was based on long-read sequencing and microsatellite analysis. Variant p.(E519K) was assessed in vitro using an enzymatic assay, Western blotting, co-immunoprecipitation, cellular thermal shift assay (CETSA), minigene assays, and in silico using protein modeling (AlphaFold).

Results: The monoallelic p.(E519K) variant was found in 83 affected individuals from Belgium, the Netherlands, France, and Canada, all of European ancestry. A shared region of 464 kilobases (kb) confirmed a founder effect. Variant p.(E519K) lowers RPE65 protein expression and enzymatic activity, with altered protein stability predicted and experimentally confirmed. Genotype-phenotype data support dominant inheritance and phenotypic variability, respectively, characterized by late-onset macular dystrophy with two main subtypes.

Conclusions: The discovery of a dominant RPE65-IRD due to founder variant p.(E519K) reduces the diagnostic gap in dominant IRD and highlights a novel target for therapy.

Figure 1.

Workflow through which monoallelic patients for RPE65 c.1555G>A p.(E519K) were identified. Panel 1 to 4 represents the Belgian IRD discovery cohort: WGS analysis (n = 150) and a second step of WES analysis (n = 2175) in IRD cases revealed c.1555G>A, p.(E519K) in 3 and 28 cases, respectively. Targeted testing of IRD patient cohorts that tested negative for mitochondrial retinopathies (MIDD), ABCA4-, BEST1-, and PRPH2-associated maculopathy (n = 518) identified 14 more patients. Segregation analysis (n = 51) showed segregation of p.(E519K) with disease in 30 affected family members. The total number of monoallelic p.(E519K) IRD cases in the Belgian discovery cohort is 75. Panel 5 represents the IRD replication cohort (n = 18,796). The p.(E519K) variant was mined in IRD databases from France, The Netherlands, Germany, the United Kingdom (UK), Scotland, Ireland, and Canada. The total number of monoallelic p.(E519K) IRD cases found in the replication cohort is 8, specifically France (n = 3; of which one was included in this study), the Netherlands (n = 1), and Canada (n = 6). Overall number of monoallelic p.(E519K) IRD cases = 85, of which 83 were included in this study.

Figure 2.

Haplotype reconstruction in all Belgian index patients, one family member (when available) and the French ( patient S34), Dutch (patient S35) and Canadian (patient S31) index cases. (A) The location of the six microsatellite markers is depicted above a schematic representation of the genes in this region (chr1: 67611911-69469522). The red line = minimal shared region (464 kb) and the green line = maximal shared region (1.6 Mb). (B) Results from haplotype analysis. Each grey box represents the two alleles with the patient identifier shown at the top. The c.1555G>A, p.(E519K) variant is indicated in bold and the common allele is shown on the left. The red box highlights the shared markers. NA values = missing data. The conversion key for the microsatellite alleles can be found in Supplementary Table S8.

Figure 3.

Conservation and protein modelling using crystal structures of RPE65. (A) The final segment of the amino acid alignment for RPE65 of 6 vertebrate species and the related carotenoid cleavage dioxygenases from 2 invertebrate species is displayed. The orange box = Asp477 and the red box = Glu519. (B) In the 3FSN crystal structure, RPE65 is observed as a dimer. Both Asp477 and Glu519 are distant from the dimerization site. (C) The monomer crystal structure of RPE65 (4RYX), whereby a positively charged lysine (Lys498) is positioned in the vicinity of Glu519. If the glutamate is replaced by a lysine these residues (Lys498 and Lys519) will repel each other due to their positive charges. His527, important for iron binding in the catalytic center, could be affected by p.(E519K).

Figure 4.

Variant p.(E519K) affects expression, enzymatic activity and thermal stability of RPE65. (A) RPE65 protein expression following single or co-transfection of vectors encoding WT-RPE65, p.(D477G), or p.(E519K) proteins in HEK293-T cells. A shift was observed for the p.(D477G) variant. CRALBP and ß-tubulin protein expression were used as transfection and loading controls. NTC, no template control. (B) Relative RPE65 protein expression, normalized using CRALBP protein expression, is shown. A significant decrease in protein expression is observed after overexpression of p.(D477G) and p.(E519K), compared to WT-RPE65. Data are represented as median ± IQR. n = 18 for WT, p.(D477G) and p.(E519K), n = 9 for WT + p.(D477G) and WT + p.(E519K), * is P < 0.05 according to a Kruskal-Wallis test with Dunn's multiple comparisons test. (C) p.(E519K) RPE65 showed reduced retinol isomerase activity (56%) compared to WTRPE65 when measured in an in cellulo minimal visual cycle assay. (D) In co-IP experiments, ARPE-19 cells were either single or co-transfected with constructs encoding HA-tagged WT-RPE65 and/or MYC-tagged p.(D477G) or p.(E519K) protein. Co-IP was performed, using the anti-HA and anti-MYC antibodies and western blotting with either anti-HA or anti-MYC antibody was used to determine coprecipitation. The results demonstrate an interaction between the p.(D477G) and WT-RPE65 proteins, but not between p.(E519K) and WT-RPE65 proteins. Input represents total protein extract, FT (flow-through) the leftover unbound fraction, and B (bound) proteins specifically bound by the antibodies and precipitated by the beads. (E, F) Results from a CETSA assay to assess thermal stability after transfection of WT-RPE65 (E) and co-transfection of WT and p.(E519K) (F) in HEK293-T cells. The IC₅₀, the temperature at which 50% of the protein is denatured and aggregated, was lowered in the WT + p.(E519K) condition. Immunoblots are depicted above the graphs, labels 1 to 12 represent the temperatures at which the cells were incubated (30, 37, 39.8, 42.8, 46.3, 48.8, 50.8, 54.1, 58.7, 62.1, 65.0, and 66.9°C, respectively). A nonlinear regression was performed on the data and resulting means with standard deviation are shown in the graphs. R² quantifies goodness of fit.

Figure 5.

SW-AF imaging (top) and color fundus imaging (bottom) illustrating the representative distinct p.(E519K)-associated phenotypes: mottled phenotype, showing small diffusely demarcated hypo-fluorescent spots, which can be accompanied by hyper-fluorescent flecks (B–E) and pattern dystrophy, which has more organized hyper-fluorescent changes and is often accompanied by chorioretinal atrophic lesions (F–J). (A) Right eye of a healthy control. (B) Left eye of a 31-year-old female patient (F12-II:1). (C) Right eye of a 43-year-old male patient (F11-II:2). (D) Left eye of a 29-year-old female patient (F2-III:1). (E) Right eye of a 47-year-old female patient (F11-II:1). (F) Right eye of a 59-year-old female patient (S10). (G) Right eye of a 78-year-old male patient (F4-I:1). (H) Right eye of a 72-year-old female patient (F3-II:2). (I) Left eye of a 62-year-old female patient (F5-III:7). (J) Left eye of a 58-year-old male patient (F8-II:5). SW-AF, short-wavelength autofluorescence.

Figure 6.

Vitelliform lesions. (A) Color fundus (top), horizontal trans foveolar OCT B-scan (middle) and SW-AF fundus (bottom) image of the left eye of patient F6-II:1 at age 33 years showing foveal vitelliform lesion on “Mottled Phenotype” background. (B) Color fundus (top) and SW-AF fundus (bottom) image of the left eye of patient S08 at age 59 years showing foveal vitelliform lesion on mild “Pattern Dystrophy” background. OCT, optical coherence tomography; SW-AF, short wavelength autofluorescence.