Macular Dystrophy and Cone-Rod Dystrophy Caused by Mutations in the RP1 Gene: Extending the RP1 Disease Spectrum

Abstract

Purpose: To describe the clinical and genetic spectrum of RP1-associated retinal dystrophies.

Methods: In this multicenter case series, we included 22 patients with RP1-associated retinal dystrophies from 19 families from The Netherlands and Japan. Data on clinical characteristics, visual acuity, visual field, ERG, and retinal imaging were extracted from medical records over a mean follow-up of 8.1 years.

Results: Eleven patients were diagnosed with autosomal recessive macular dystrophy (arMD) or autosomal recessive cone-rod dystrophy (arCRD), five with autosomal recessive retinitis pigmentosa (arRP), and six with autosomal dominant RP (adRP). The mean age of onset was 40.3 years (range 14–56) in the patients with arMD/arCRD, 26.2 years (range 18–40) in adRP, and 8.8 years (range 5–12) in arRP patients. All patients with arMD/arCRD carried either the hypomorphic p.Arg1933* variant positioned close to the C-terminus (8 of 11 patients) or a missense variant in exon 2 (3 of 11 patients), compound heterozygous with a likely deleterious frameshift or nonsense mutation, or the p.Gln1916* variant. In contrast, all mutations identified in adRP and arRP patients were frameshift and/or nonsense variants located far from the C-terminus.

Conclusions: Mutations in the RP1 gene are associated with a broad spectrum of progressive retinal dystrophies. In addition to adRP and arRP, our study provides further evidence that arCRD and arMD are RP1-associated phenotypes as well. The macular involvement in patients with the hypomorphic RP1 variant suggests that macular function may remain compromised if expression levels of RP1 do not reach adequate levels after gene augmentation therapy.

Retinitis pigmentosa (RP) encompasses a heterogeneous group of inherited retinal dystrophies characterized by rod photoreceptor degeneration that precedes cone photoreceptor degeneration. The RP1 gene is one of the more than 80 genes associated with RP. Besides RP1, seven other RP genes—BEST1, NR2E3, NRL, RHD12, RHO, RPE65, and SAG—have been associated with both autosomal dominant and autosomal recessive modes of inheritance.¹ In general, RP patients with an autosomal recessive inheritance pattern have a more severe disease course compared with patients with autosomal dominant RP (adRP). This also applies to patients with RP1-associated RP: autosomal recessive RP1 patients generally have a lower age of onset compared with autosomal dominant RP1 patients, as well as a worse long-term prognosis with respect to retaining central vision due to the occurrence of early macular atrophy or cystoid macular edema.^1,2 

The RP1 gene, mapped to chromosome 8q12.1, contains four exons, three of which are coding, and encodes a photoreceptor-specific microtubule-associated protein that plays a vital role in the architecture of both rod and cone photoreceptor outer segments.^3,4 RP1 is located at the photoreceptor axoneme, where it links outer segment discs to the axonemal microtubules, and thereby regulates the length and stability of the axoneme.⁵ The interaction with the microtubules is mediated primarily by two doublecortin (DCX) domains encoded by exons 2 and 3.⁵ In addition, RP1 contains a third putative domain, between amino acid residues 486 and 635, that shares homology with the Drosophila melanogaster bifocal (BIF) protein, which is required for normal photoreceptor morphogenesis.⁶ RP1 mutations that are known to cause adRP are clustered in a relatively small region in exon 4 between amino acid residues 500 and 1053,⁷ and result in the production of a truncated protein with a presumed dominant-negative activity.⁸ In contrast, most mutations located more toward the N- or C-terminus of RP1 result in autosomal recessive RP (arRP).⁷ 

In 2016, Ellingford et al.⁹ identified compound heterozygous mutations in RP1 (i.e., p.Tyr41His and p.Leu172Arg) in a patient diagnosed with macular dystrophy (MD)/presumed Stargardt disease. However, because this new genotype-phenotype correlation was identified in only a single family, they concluded that reevaluation of the clinical phenotype was warranted.⁹ Knowledge about the entire disease spectrum associated with certain genes is important, particularly in view of novel therapeutic options, as the prognosis and disease course between phenotypes may differ markedly. In this study, we report patients diagnosed with MD and cone-rod dystrophy (CRD) in addition to patients with RP, and expand the clinical spectrum associated with mutations in the RP1 gene. 

Methods

Patients

Twenty-two patients (19 families) with a retinal dystrophy and mutations in the RP1 gene were clinically examined at the Radboud university medical center in Nijmegen, The Netherlands (families A, B, M, N , P, and Q); the Tohoku University Graduate School of Medicine in Sendai, Japan (families C–E, J and L); the Kyushu University Hospital, Fukuoka, Japan (families F and K); the Yuko Wada Eye Clinic, Sendai, Japan (family G); and the Amsterdam UMC, The Netherlands (families H, I, O, R and S). The genetic evaluation was performed between June 2013 and May 2018. Informed consent was obtained from all patients before data collection and additional ophthalmic examinations. The study adhered to the tenets of the Declaration of Helsinki, and was approved by the local ethics committees. 

Genetic Analysis

In all families, genomic DNA was extracted from peripheral lymphocytes according to standard procedures. The genetic data of the Japanese patients was obtained in context of another study.¹⁰ In short, whole exome sequencing was performed in patients A-II:2, A-II:4, A-II:5, B-II:8, H-II:2, I-II:2, M-II:2, P-II:13, and Q-III:4. The exome data were analyzed using a vision gene panel consisting of 220 (patients H-II:2 and I-II:2), 342 (patient P-II:13), 366 (patient B-II:7), and 395 genes (patients M-II:2 and Q-III:4) or without the use of a gene filter (family A). See Supplementary Table S1 for an overview of the genes included in these panels. Targeted panel sequencing covering 256 (patient R-IV:2) or 266 vision genes (patients O-II:1 and S-III:4) was performed in a certified DNA diagnostic laboratory, with additional Sanger sequencing for all areas with a coverage below 30 reads (Supplementary Table S1). Mutational screening in patient N-III:9 was performed using an arrayed primer extension microarray for adRP (containing 414 variants in 16 genes), according to a previously described protocol.¹¹ All variants detected by microarray analysis were verified by direct sequencing. Patients C to G and J to L had their RP1 open reading frame screened by means of Sanger sequencing as previously reported.¹⁰ In addition, molecular inversion probes were used to exclude variants in 109 other inherited retinal dystrophy genes in patients C to F, J, and K,¹⁰ and targeted resequencing containing 83 nonsyndromic RP genes was performed in patient L-II:2 (Supplementary Table S1). The pathogenicity of novel missense variants was assessed combining cosegregation analysis and in silico prediction tools, including SIFT and Polyphen-2, and by using the PhyloP, CADD-PHRED, and Grantham scores. For an extensive description of the genetic analysis, see Supplement 1. 

Clinical Evaluation

Clinical data were obtained from the medical records of the patients. In addition, three patients (family A) were reevaluated after the identification of the causative RP1 mutations. We performed a detailed ophthalmic examination, which included visual acuity testing, slit-lamp biomicroscopy, and detailed ophthalmoscopy. Most patients also underwent conventional fundus photography and/or ultra-widefield fundus imaging (Optos P200Tx; Optos, Dunfermline, UK), Goldmann perimetry, as well as spectral-domain optical coherence tomography (SD-OCT) using a confocal scanning laser ophthalmoscope (Spectralis HRA+OCT; Heidelberg Engineering, Heidelberg, Germany; Cirrus; Carl Zeiss Meditec, Inc., Dublin, CA; or Topcon 3D OCT-2000; Topcon, Inc, Tokyo, Japan ). Fundus autofluorescence (FAF) imaging was performed using the Spectralis HRA+OCT 30° × 30° field of view centered on the macula,¹² or the 200° field of view from the ultra-widefield imaging device. All patients underwent full-field ERG, and multifocal electroretinography (mfERG) was performed in patients B-II:8, C-II:2, D-II:1, and E-II:2. Electrophysiological recordings were performed according to the International Society for Clinical Electrophysiology of Vision guidelines and assessed by applying local standard values.^13,14 

Patients were diagnosed based on their (initial) symptoms, fundus abnormalities, ERG findings, and overall course of the disease. They received the diagnosis MD if they experienced central vision loss without symptoms of night blindness, ophthalmoscopy and multimodal imaging revealed no signs of peripheral involvement, in the presence of an intact peripheral visual field, normal to marginally abnormal scotopic ERG responses, and normal to moderately reduced photopic ERG responses. Patients with a panretinal phenotype received the diagnosis CRD or RP, depending on which photoreceptor function was affected first. CRD was diagnosed when the onset of loss of central vision preceded that of night blindness, presence of a central scotoma, no or mild constriction of the visual field, and reduced ERG responses in a cone-rod pattern. Patients were diagnosed with RP when they presented with night blindness, constriction of the visual field, and reduced ERG responses in a rod-cone pattern. The clinical distinction between MD and CRD, however, can be difficult and is sometimes arbitrary in view of the significant clinical overlap. In addition, in individual MD patients the MD may progress in a more generalized disorder that fits the criteria of CRD. In the present study, we therefore grouped the MD and CRD spectrum into a single category for further analysis. 

Results

Genetic Findings

Genetic analysis detected 16 unique variants in the RP1 gene, including six newly identified variants, in the 19 families (Fig. 1).¹⁰ The novel variants included five nonsense and frameshift mutations, as well as the p.Val190Gly missense variant that is likely pathogenic according to the guideline proposed by the American College of Medical Genetics and Genomics.¹⁵ The variant segregates with the disease (M8, family H), affects a highly conserved amino acid that is located in the DCX domain, is extremely rare in the gnomAD database (1/249046), and is predicted to be pathogenic with a high Grantham score (109/215) and CADD-PRHED score (22.2) including a high SIFT (pathogenic) and PolyPhen-2 score (probably damaging). Heterozygous variants in other inherited retinal dystrophy genes are listed in Supplementary Table S2. 

Figure 1

 

Pedigrees of the families included in this study. Squared boxes indicate men, circles indicate women, filled symbols represent affected persons, and unfilled symbols represent unaffected persons. The plus sign denotes the wild-type allele, and the arrow indicates the proband of the family. Double lines point out consanguineous marriages, the number above the lines indicates the degree of consanguinity. Where relatives were available (families A, B, C, D, H, I, and N), the mutations segregate with the disease. Families C to F, J, and K have previously been described by Nikopoulos et al.^{10 §}Novel variants.

Notably, compound heterozygous mutations were identified in all patients with autosomal recessive MD (arMD)/autosomal recessive CRD (arCRD). In particular, the p.Arg1933* mutation, which was identified in a compound heterozygous state with another nonsense or frameshift mutation in six of eight families with arMD/arCRD, from both East Asian and Caucasian origins (Fig. 1). The remaining two families with arMD/arCRD (families B and H) carried a heterozygous missense mutation (p.Phe180Cys or p.Val190Gly) located in exon 2, in the region that encodes the DCX domain, together with a frameshift mutation on the second allele. Families with an identical combination of mutations were identified; p.Arg1933* in combination with p.Tyr1352Alafs*9 was previously identified in four families of East Asian origin (families C– F).¹⁰ 

In contrast, in five of six families with arRP, homozygous nonsense or frameshift variants were identified. Remarkably, the p.Tyr1352Alafs*9 mutation that was found in patients with arMD/arCRD in combination with p.Arg1933* in families C to F, was identified in a homozygous state in two families with arRP (families J and K), as described.¹⁰ In addition, the p.Pro124Alafs*20 mutation that causes arMD/arCRD in combination with the p.Val190Gly missense mutation in patient H, was also identified in a homozygous state in arRP family N. 

Clinical Findings

An overview of the individual clinical characteristics of the patients with an RP1 mutation is provided in the Table. The clinical characteristics stratified by phenotype are provided in Supplementary Table S3. Of the 22 patients, 11 patients (7 arMD and 4 arCRD patients from eight families) were diagnosed with arMD/arCRD, 5 patients (five families) with arRP, and 6 patients (six families) with adRP. Patient B-II:8 was initially diagnosed with a hydroxychloroquine-associated maculopathy as she received treatment with hydroxychloroquine for rheumatoid arthritis. The correct diagnosis of RP1-associated MD was made based on her clinical presentation that was atypical for hydroxychloroquine maculopathy, the low cumulative dose of 280 g of hydroxychloroquine, and an affected sister free of the medication. Subsequent molecular genetic testing revealed compound heterozygous mutations in the RP1 gene in both of them. 

Table

 

Clinical Features of Patients Carrying Pathogenic Mutations in RP1

 

Patients With arMD/arCRD

The patients with arMD/arCRD presented the latest of the three phenotypes (mean age of onset 40.3 years; SD 13.1 years, range 14–56 years), although they showed overlap in age at onset with the adRP patients (mean 26.2 years, SD 8.2 years, range 18–40 years, P = 0.025) (Supplementary Table S3). Their initial symptom was a decrease in visual acuity or metamorphopsia, sometimes accompanied by photophobia (patient C-II:2) or night blindness (patient F-II:3 and G-II:2). With progression of the disease, 8 of 11 patients developed photophobia. All patients were myopic (range of spherical equivalents: −9.50 diopters [D] to −0.25 D). Biomicroscopy revealed several types of lens opacities in 6 of 11 patients of whom patient F-II:3 underwent cataract extraction at the age of 61 years (Table, Supplementary Table S3). The course of the visual acuity for each patient is represented in Figure 2. Ten patients revealed mild to moderate visual acuity impairment during working life, which eventually led to acuity levels of 20/400 in the eighth decade (patient A-II:2). However, patient C-II:2 already had a visual acuity of 20/400 at the age of 41 years. Visual field testing revealed an absolute central scotoma, except for a paracentral scotoma in the left eye of patient B-II:8 (age 55 years), and a central, relative scotoma in patient B-II:3 (age 64 years). In addition, the visual field was mildly constricted in patients B-II:3, F-II:3, and G-II:2. 

Figure 2

 

Graph showing the course of visual acuity over time in patients with mutations in the RP1 gene. The three phenotypes are indicated with different colors: arMD/arCRD in blue, arRP in orange, and adRP in red. The visual acuity of the best eye is displayed. Snellen visual acuity was converted into logMAR. A logMAR value of 1.9 was assigned to counting fingers (CF), 2.3 to hand movements (HM), and 2.7 to light perception (LP).

Ophthalmoscopy showed RPE alterations or atrophy in the macula (Table, Fig. 3). Attenuation of the retinal vessels was present in patients A-II:4, B-II:3, F-II:3, and G-II:2, focal bone spicule pigmentations were observed in patient F-II:3, and a single nummular pigmentation in patient G-II:2. A bull's eye maculopathy was noticed in the left eye of patient B-II:8 (age 55) and both eyes of patient E-II:2 at the most recent examination. In patient A-II:5, the bull's eye maculopathy had progressed to macular atrophy with foveal involvement (Fig. 4). A posterior staphyloma was visible in patient A-II:2 with spherical equivalents of −8.38/−8.00 D. SD-OCT images revealed loss of the outer retinal layers in the macula, with sparing of the foveal photoreceptors in the patients with a bull's eye maculopathy and the left eye of patient B-II:3 (age 64), in which the photoreceptors and RPE at the fovea were preserved in a foveal sparing-like pattern (Fig. 3F). The right eye of patient B-II:3 initially showed foveal sparing at the age of 56 years, although shortly afterward the foveal cells degenerated as well, with a corresponding loss in visual acuity. FAF imaging showed a round to oval zone of reduced FAF or a speckled pattern of alternating normal and decreased FAF, bordered by a band of increased FAF (Fig. 3); except for patient E:II-2, in whom an oval zone of increased FAF was visible without a reduced autofluorescence signal. FAF images of patients F-II:3 and G-II:2 were not acquired. Full-field ERG ranged from normal rod and cone responses to patients in whom both cone and rod responses were moderately or severely reduced (Table). The mfERG, performed in patients B-II:8, C-II:2, D-II:1, and E-II:2, was severely reduced in all patients except for patient B-II:8 in whom a decreased response in the parafoveal ring was visible in the right eye and moderately reduced responses in the left eye. 

Figure 3

 

Multimodal images of patients with mutations in RP1. (A–C) Multimodal images of the left eye of patient A-II:4. (A) Composite fundus photograph showing a well-demarcated area of chorioretinal atrophy involving the fovea. (B) This area corresponds with an oval zone of absent FAF, bordered by a small residual band of increased FAF. (C) SD-OCT scan revealing loss of the outer retinal layers in the macula. (D–F) Multimodal imaging of the left eye of patient B-II:3. (D) Composite fundus photograph showing macular atrophy, attenuated retinal vessels, and no hyperpigmentation. (E) FAF image revealing decreased autofluorescence surrounded by a hyperautofluorescent ring. (F) SD-OCT scan showing preservation of the RPE, ellipsoid zone, and external limiting membrane layer in the fovea, and an epiretinal membrane. (G–I) Multimodal imaging of the left eye of patient B-II:8, at the age of 55 years. (G) Fundus photograph showing a bull's eye maculopathy, a normal aspect of the optic disc and vasculature, and no intraretinal hyperpigmentation. (H) FAF shows a corresponding oval area of speckled hyper- and hypoautofluorescence, surrounded by a ring of hyperautofluorescence. (I) SD-OCT image revealing the parafoveal loss of outer retinal layers with preservation of the ellipsoid zone and external limiting membrane layer in the fovea. (J–L) Multimodal imaging of the right eye of patient D-II:1. (J) Ultra-widefield fundus photograph showing RPE alterations in the macula, peripapillary atrophy, no hyperpigmentations, and a normal aspect of the optic disc and retinal vessels. (K) Ultra-widefield FAF image showing a speckled pattern of hyper- and hypoautofluorescence, surrounded by a band of increased FAF. (L) OCT scan showing generalized loss of photoreceptor inner and outer segments, and an epiretinal membrane. (M–O) Multimodal imaging of the left eye of patient J-II:2. (M) Ultra-widefield fundus photograph showing RPE atrophy, bone spicule pigmentation, and attenuation of the retinal vessels. (N) Autofluorescence image revealing decreased autofluorescence in the macula and along and surrounding to the vascular arcades. (O) SD-OCT scan showing generalized loss of the outer retinal layers in the macula. (P–R) Multimodal imaging of the left eye of patient N-III:9. (P) Composite fundus photograph showing profound peripheral atrophy, severe attenuation of retinal vessels, extensive bone spicule and nummular pigmentation, and waxy pallor of the optic disc. (Q) FAF image showing a hyperautofluorescent ring in the macula, surrounded by a speckled pattern of hyper- and hypoautofluorescence, and nummular areas of decreased macular and peripheral autofluorescence. (R) SD-OCT scan revealing the preservation of the photoreceptor layer at the fovea.

Figure 4

 

Multimodal images of the left eye of patient A-II:5 over an interval of 6 years. (A) Fundus photograph showing a bull's eye maculopathy, some peripapillary atrophy, and a normal aspect of the optic disc and retinal vessels. (B) The FAF image reveals an oval area of hyperautofluorescence containing a ring of hyper- and hypoautofluorescence. (C) Horizontal SD-OCT showing perifoveal loss of the outer retinal layers. (D) Fundus photograph 6 years later showing macular atrophy, and (E) FAF showing the corresponding area of hypoautofluorescence surrounded by a zone of increased autofluorescence. (F) The OCT scan shows the loss of the outer retinal layers that now also involves the fovea.

Autosomal Recessive RP

The arRP patients were affected at the earliest age, with a mean age of onset of 8.8 years and an onset at the age of only 5 years in the youngest patient (SD 2.8; range 5–12 years; Supplementary Table S3). The initial symptom in the arRP patients was night blindness in all cases and all patients were myopic (range of spherical equivalents: −9.00 D to −1.63 D; Table, Supplementary Table S3). Despite their young age, all four patients for whom data were available already showed lens opacities at a mean age of 33 years. The visual acuity deteriorated from adolescence or early adulthood to levels of light perception in the sixth decade of life and constriction of the visual field with a residue of less than 10 degrees at the most recent examination was present in three of four patients (Table, Fig. 2). 

Ophthalmoscopy showed the three hallmark RP signs—bone spicule pigmentations, attenuation of the retinal vessels, and pallor of the optic disc—often accompanied by profound peripheral RPE atrophy. In addition, macular atrophic lesions were present in four of five patients, with sparing of the fovea in two of them (Table). This was confirmed by an intact ellipsoid zone layer in the fovea on SD-OCT (Fig. 3R, Supplementary Table S3). Finally, ERG responses were severely reduced or nonrecordable under scotopic and photopic conditions in all RP patients. 

Autosomal Dominant RP

Patients with adRP presented with night blindness, at a mean age of 26.2 years (SD 8.2; range 18–40 years). Five of six patients were myopic (range of spherical equivalents: −10.00 to +4.00), except for patient O-II:13, who was hyperopic with spherical equivalents of +1.50 D and +4.00 D (Table). Biomicroscopy revealed several types of lens opacities in four of six patients of whom three patients underwent cataract extraction (Table, Supplementary Table S3). The visual acuity was relatively preserved, as all patients retained a visual acuity of 20/80 or better, even at the age of 72 (patient O-II:13) (Fig. 2, Supplementary Table S3). However, the visual field showed severe constriction with a central residue of less than 10 degrees in two of six patients at a mean age of 48 years (Table, Supplementary Table S3). 

The patients with adRP also showed bone spicule pigmentation and attenuation of the retinal vessel, often accompanied by profound peripheral RPE atrophy and macular atrophy in four of six patients (Table). However, the fovea was spared from the atrophic lesions in all patients (except for the left eye of patient O-II:1). Patient O-II:1 previously underwent a pars plana vitrectomy with peeling of the internal limiting membrane in both eyes for macular holes. Cystoid macular edema was present in patient Q-III:4, with a moderate response to oral carbonic anhydrase inhibitors. Finally, ERG responses were nonrecordable under scotopic conditions and severely reduced or nonrecordable under photopic conditions. 

Discussion

Mutations in the RP1 gene have previously been described in patients with adRP and arRP. In the present study, we provide a detailed clinical description of these phenotypes and report two additional RP1-associated diagnoses: arMD and arCRD, which may represent a single spectrum of retinal degeneration. 

The patients with RP1-associated arMD/arCRD presented with a decrease in visual acuity or metamorphopsia, generally first noticed in the fourth decade, which eventually progressed to legal blindness. Considerable macular abnormalities were observed at the most recent examination including a bull's eye maculopathy in two patients. This bull's eye maculopathy may eventually progress to macular atrophy with foveal involvement, as was observed in patient A-II:5. Unfortunately, longitudinal data of the other patients to confirm this hypothesis were not available. 

MD and CRD may show large overlap in clinical and genetic findings. This overlap also occurs in time: an MD phenotype can eventually progress to a more generalized disorder and converge into a CRD phenotype. This may explain the intrafamilial differences in family B, as patient B-II:8 shows abnormalities limited to the macular region, whereas her older sister displays generalized disease, which may represent a later disease stage. It might also explain the more severe phenotype in patient F-II:3, who carries the same mutations as the younger patients C to E with less severe disease. In addition, the mean age of onset of both phenotypes was the same (P = 0.817). Therefore, arMD and arCRD might represent the longitudinal progression of macular/cone predominant RP1-associated disease. Nevertheless, some patients may never show progression to a CRD phenotype and other (genetic) modifiers may also exert their effect on the final phenotype. Because the focus of our manuscript lies on the description of patients with a phenotype that does not match the existing phenotypes with predominant rod involvement, we have combined them into the all-embracing term “arMD/arCRD.” 

The arMD/arCRD shows similarities with other MDs (e.g., central areolar choroidal dystrophy and Stargardt disease), multifocal pattern dystrophy simulating Stargardt disease, and cone dystrophies, and, because of its later age of onset, can mimic AMD.^16–18 It is, however, important to distinguish these disorders to provide the patient with valuable and correct prognostic information and accurate treatment. An adequate family history, the absence of drusen, and the symmetrical presentation of the macular atrophy can help the clinician in distinguishing hereditary forms of MDs from AMD.¹⁷ The absence of irregular yellowish (pisciform) and/or hyperautofluorescent flecks can help to differentiate RP1-associated disease from Stargardt disease, pseudo-Stargardt, and certain cases of central areolar choroidal dystrophy, and ERG responses can help to differentiate from cone dystrophies. 

The phenotype of the RP patients was in accordance with earlier reports on RP caused by RP1 mutations in literature. The five patients with arRP experienced night blindness in the first or second decade of life and showed early involvement of the macular region, whereas the age of onset in the six patients with adRP was between the second and fourth decades, and the fovea remained relatively intact during the course of the disease.^1,2 Myopia, in varying degrees, is a common feature, and was found in 96% of patients. This association of myopia and RP1-related disease has previously been described, particularly in arRP patients.^19,20 

Together, these four retinal dystrophies now form the spectrum of RP1-associated disease. It is, however, important to realize that the RP phenotypes are not simply a more extensive form of the more centrally located form of RP1-associated disease. This is illustrated by the visual acuity in the adRP patients that can remain relatively preserved even in advanced RP cases as well as the different predominantly affected photoreceptors 

The phenotype is, to a large extent, determined by the location and severity of the mutations in the RP1 gene. Several classifications of mutations have been proposed.^7,21,22 In Supplementary Figure S1, we provide the latest overview of mutations and their location. The mutations responsible for adRP reside in a hotspot region in exon 4 between amino acid residues 500 and 1053, and are expected to result in a truncated protein with dominant-negative activity.⁸ In contrast, arRP is caused by the presence of two nonsense or frameshift mutations in RP1. Patients with arMD/arCRD carry a heterozygous variant that is expected to have a mild effect on protein function (p.Phe180Cys, p.Val190Gly, or p.Arg1933*) in combination with a more severe nonsense or frameshift mutation, or a combination of two predicted mild variants such as p.Gln1916* and p.Arg1933* in family 1 and the compound heterozygous missense variants (p.Tyr41His and p.Leu172Arg) reported by Ellingford et al.⁹ in a single MD patient (Fig. 1). The p.Arg1933* variant, a recurrent variant in the Japanese population (allele frequency: 0.6%), does not cause a retinal dystrophy in homozygous carriers, at least not before the age of 80 years.¹⁰ Although, in combination with a pathogenic variant, such as p.Glu334fs*22, p.Tyr1352Alafs*9, or another likely hypomorphic variant p.Gln1916*, the effect of this hypomorphic allele seems to be sufficient to cause retinal disease (Fig. 1). 

Although these findings are consistent within our study, there are reports, albeit with limited clinical information, that conflict with our findings. For example, p.Arg1933* was found in trans with the p.Tyr834* nonsense variant in a patient with arRP.²³ In addition, other previously reported variants, such as p.D202E, p.I1988Nfs*3, and p.I2061Sfs*12, could also be predicted to be mild or hypomorphic variants considering their location; however, they have been associated with arRP, although the clinical phenotype has not been described in detail.^23–25 The explanation for this clinical heterogeneity remains to be elucidated, but might be explained by additional (genetic-) modifying factors, the presence of a structural variant or a variant in the non-coding regions of RP1, pathogenic variants in another RP gene, or inaccurate phenotyping, the latter of which should be considered particularly in advanced RP and CRD, in which the distinction can be difficult and depends on patients' self-reported disease course. A potential genetic modifier for RP1-associated disease may be RP1L1, because these proteins have synergistic roles in the photoreceptor axoneme.²⁶ However, we did not find any rare variants in the RP1L1, although intronic or structural variants cannot be excluded (Supplementary Table S2). 

RP1 is an interesting candidate for gene therapy, because of its relatively high prevalence in RP.¹ However, important challenges must be overcome before gene therapy for RP1 can reach the clinic. For example, knowledge about the natural course of the disease is required to be able to evaluate treatment efficacy, particularly in view of the different phenotypes. In addition, in case of gene augmentation therapy, the maximum cargo capacity of the AAVs (∼4.7 kb) is too limited to fit the RP1 gene (6.5 kb),^27,28 and the optimal expression levels of the RP1 protein need to be determined.⁸ This could be more important than previously anticipated if our hypothesis that a hypomorphic variant in combination with a deleterious variant causes arMD/arCRD is correct. Insufficient dosage of the RP1 gene by gene transfer may mimic this condition and could in that case result in the development of an iatrogenic MD. Obviously, this would be a serious concern when treating arRP patients with preserved macular function with the goal to halt the progression of central visual field loss. Although patients with adRP might also benefit from an elevation of wild-type RP1 levels,⁸ the mutant protein that possesses dominant-negative activity remains present in the cell and competes with the wild type. Therefore, other genetic therapies, such as treatment with antisense oligonucleotides, of which proof-of-concept has been shown in an animal model with RHO-associated RP,²⁹ or genome editing may be alternative approaches. 

In conclusion, mutations in the RP1 gene can lead to different clinical phenotypes, varying from RP to arMD/arCRD, depending on the residual RP1 function. Together, these dystrophies form a spectrum of RP1-associated phenotypes that can be clinically distinguished from each other based on the clinical findings, inheritance pattern, age of onset, and disease course. However, additional longitudinal studies are essential to improve the diagnostic process and to study the role of potential modifiers. With the advent of novel therapeutic options such as gene therapy, recognition of the entire clinical spectrum associated with RP1 mutations is essential to aid the selection of patients eligible for treatment, and to evaluate the effect of the treatment provided. 

Acknowledgments

The authors thank Annemiek Struijk and Ralph Florijn for kindly providing clinical and genetic data, and Bjorn Bakker for excellent technical assistance. 

Supported by the Japan Agency for Medical Research and Development (TN, JP17lk1403004;KMN 17ek0109213h0002). The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement n. 310644 (MACULA), the European Union's Horizon 2020 research and innovation program under Grant Agreement no. 634479 (EYE-RISK), the Maculafonds No. MF-2018–58_FZ, and The Swiss National Science Foundation (Grant No. 176097). The funding organizations had no role in the design or conduct of this research. 

Disclosure: S.K. Verbakel, None; R.A.C. van Huet, None; A.I. den Hollander, None; M.J. Geerlings, None; E. Kersten, None; B.J. Klevering, None; C.C.W. Klaver, None; A.S. Plomp, None; N.L. Wesseling, None; A.A.B. Bergen, None; K. Nikopoulos, None; C. Rivolta, None; Y. Ikeda, None; K.-H. Sonoda, None; Y. Wada, None; C.J.F. Boon, None; T. Nakazawa, None; C.B. Hoyng, None; K.M. Nishiguchi, None