Purpose
To assess the prevalence of PRPH2 in autosomal dominant retinitis pigmentosa (adRP), to report 6 novel mutations, to characterize the biochemical features of a recurrent novel mutation, and to study the clinical features of adRP patients.
Design
Retrospective clinical and molecular genetic study.
Methods
Clinical investigations included visual field testing, fundus examination, high-resolution spectral-domain optical coherence tomography (OCT), fundus autofluorescence imaging, and electroretinogram (ERG) recording. PRPH2 was screened by Sanger sequencing in a cohort of 310 French families with adRP. Peripherin-2 protein was produced in yeast and analyzed by Western blot.
Results
We identified 15 mutations, including 6 novel and 9 previously reported changes in 32 families, accounting for a prevalence of 10.3% in this adRP population. We showed that a new recurrent p.Leu254Gln mutation leads to protein aggregation, suggesting abnormal folding. The clinical severity of the disease in examined patients was moderate with 78% of the eyes having 1-0.5 of visual acuity and 52% of the eyes retaining more than 50% of the visual field. Some patients characteristically showed vitelliform deposits or macular involvement. In some families, pericentral RP or macular dystrophy were found in family members while widespread RP was present in other members of the same families.
Conclusions
The mutations in PRPH2 account for 10.3% of adRP in the French population, which is higher than previously reported (0%-8%) This makes PRPH2 the second most frequent adRP gene after RHO in our series. PRPH2 mutations cause highly variable phenotypes and moderate forms of adRP, including mild cases, which could be underdiagnosed.
In the retina, the human peripherin-2 gene ( PRPH2 ; MIM #179605), also known as RDS (retinal degeneration slow), encodes peripherin-2, a transmembrane glycoprotein localized in the rim regions of photoreceptor outer segment discs. Peripherin-2 forms homo- and heterotetramers with its paralog protein ROM1 (retinal outer segment membrane protein 1; MIM #180721). These oligomers are essential for the stabilization of the disc rims and are required to pile up the discs as compact, elongated structures. Mutations in PRPH2 cause a wide range of autosomal dominant retinal dystrophies, either with involvement of the peripheral retina such as retinitis pigmentosa, cone-rod dystrophy, and even 1 case of retinitis punctata albescens, or with predominant involvement of the macula such as adult vitelliform macular dystrophy, cone dystrophy, pattern dystrophy, and central areolar choroidal atrophy. In addition, the PRPH2 p.Leu185Pro substitution has also been associated with ROM1 mutations in a digenic form of retinitis pigmentosa.
Among the variety of retinal degenerations caused by PRPH2 mutations, autosomal dominant retinitis pigmentosa (adRP) is the most frequent condition. Typical symptoms of RP include night blindness and progressive visual field constriction, eventually progressing toward total blindness after several decades. The prevalence of RP is approximately 1/3500 to 1/4000 and the mode of inheritance can be autosomal dominant (30%-40%), autosomal recessive (50%-60%), or X-linked (5%-15%). RP is the most genetically heterogeneous clinical entity of inherited retinal disorders, with 69 disease-causing genes currently known in this condition ( www.sph.uth.tmc.edu/retnet ), including 24 genes causing adRP. The prevalence of the known genes in adRP ranges from 26.5% to 16.6% for the most frequently found mutations in RHO (MIM #180380), to many genes accounting for less than 1% of the adRP families. Among those genes, the prevalence of PRPH2 mutations varies widely from 0% to 8% of the cases of adRP in cohorts of different origins, but no accurate prevalence data are available for the French population. Also, as usually found in adRP, the severity of the PRPH2 genetic form is considered as moderate, but it is not known whether there are important variations of severity inside the PRPH2 genetic category. Therefore, we sought PRPH2 mutations in a large cohort of 310 adRP families originating mainly from France. We found novel mutations, characterized the biochemical features of 1 novel mutation, and analyzed the clinical features of the affected patients.
Methods
Patients
Three hundred and ten index patients were included in the study. Informed and written consent was obtained for all patients participating in the study. Patients of European origins were recruited from 10 different clinical centers in France. The study (# 2008-A01238-47) received the authorization from the Sud méditerranée IV ethical board committee (# 08 10 05 from 04/11/2008), was approved by the French regulation agency for medication (AFSSAPS # B81319-70), and is registered at http://clinicaltrials.gov (# NCT01235624). The investigators followed the tenets of the Declaration of Helsinki.
Clinical Investigations
Patients had standard ophthalmologic examination (refractometry, visual acuity, slit-lamp examination, applanation tonometry, and funduscopy). Kinetic visual fields were determined with a Goldmann perimeter with targets V4e, III4e, and I4e. Optical coherence tomography (OCT) measurement of the macula was performed using an OCT-3 system (Stratus model 3000; Carl Zeiss Meditec, Dublin, California, USA) or with a spectral-domain OCT (Spectralis, Heidelberg, Germany) with software version 3.0. Autofluorescence measurements were obtained with the HRA2 Heidelberg retinal confocal angiograph (Heidelberg Engineering, Dossenheim, Germany) and fundus pictures were taken. Full-field electroretinograms (ERGs) were recorded using a Ganzfeld apparatus (Metrovision, Pérenchies, France) with a bipolar contact lens electrode on maximally dilated pupils according to the ISCEV protocol.
For numerical values, visual acuity was measured with Snellen charts in decimal numbers. Goldmann visual field was quantified by counting the number of subdivisions of the Goldmann grid within the areas of the V4e isopter and expressed as a percentage of the normal visual field. Correlations between visual parameters (visual acuity, visual field, and ERG amplitudes) and age were investigated with the coefficient correlation of ranks of Spearman with a confidence interval at 95%, calculated by a Fisher transformation.
Mutation Screening
Genomic DNA was isolated from 10 mL peripheral blood leukocytes using standard salting-out procedure. Coding exons and adjacent intronic sequences of the PRPH2 gene (NM_000322.4; primer pairs and polymerase chain reaction [PCR] conditions are available on request) were sequenced with an Applied Biosystems 3130xL genetic analyzer (Applied Biosystems, Foster City, California, USA) using a BigDye Terminator cycle sequencing ready reaction kit V3.1 (Applied Biosystems) following the manufacturer’s instructions. Sequence analysis and mutation identification were performed using Collection and Sequence Analysis software package (Applied Biosystems). SIFT, PolyPhen2, and Align GVGD were used to predict possible impacts of missense variants. The genomic sequence environment of putative splice-site mutations was analyzed using Human Splicing Finder and MaxEnt.
Genotyping of Microsatellite Markers and Linkage Analysis
PCR was carried out in 25 μL final volume containing 50 ng genomic DNA, 5 pmol of each primer, 0.2 mM dNTPs (MP Biochemicals, Asse-Relegen, Belgium), 2 mM MgCl 2 , PCR buffer, and 1 unit of DNA polymerase (AmpliTaq Gold; Applied Biosystems). Initial denaturation at 95 C for 10 minutes was followed by 35 cycles of denaturation at 94 C for 30 seconds, specific annealing temperature for 30 seconds, and extension at 72 C for 1 minute. A final extension step was performed at 72 C for 10 minutes. The PCR products were diluted and mixed with Genescan 400HD ROX size standard and subsequently analyzed on an Applied Biosystems 3130xL genetic analyzer (Applied Biosystems). Results were analyzed with GeneMapper software (version 4.0; Applied Biosystems).
Two-point LOD scores were calculated with Superlink-online ( http://bioinfo.cs.technion.ac.il/superlink-online/ ). The phenotype was analyzed as an autosomal dominant and fully penetrant trait with an affected allele frequency of 0.001.
Peripherin-2 Expression and Western Blots
Wild-type (WT) and p.Leu254Gln (L254Q) mutant were cloned into the pPICZ expression vector containing the c-myc epitope and the polyhistidine ( His)6 -tag as described before ; the nucleotide sequence was confirmed by Eurofins MWG (Ebersberg, Germany) using automated DNA sequencing. Pichia pastoris cells (strain KM71H) were transformed with the Pme I linearized expression vector, stably transformed cells were spread on YPD plates (1% yeast extract, 2% peptone, 2% glucose, 2% agar) with media containing 100 μg/mL zeocin. Cells were cultured, harvested, and stored at −80 C as described before. Cells were lysed upon further processing and membranes containing the WT or L254Q proteins were isolated using differential centrifugation as described previously. The membranes were dissolved in 1% n -dodecyl- β -D-maltoside (DDM) using sequentially an 18G, 19G, and 25G needle. His-tagged WT or L254Q proteins were purified using Ni-NTA agarose (final buffer 10 mM NaPO4, 150 mM NaCl, 200 mM imidazole, and 0.1% n -dodecyl- β -D-maltoside). Reducing sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was performed by mixing 1:1 (v:v) with 2× loading buffer containing 1% β-mercaptoethanol and incubated for 5 minutes at room temperature prior to loading of the gel. Nonreducing SDS-PAGE was performed by mixing 1:1 (v:v) with 2× loading buffer without β-mercaptoethanol and immediate loading after mixing. Transfer to the polyvinylidene fluoride membrane and probing—using cmyc-tagged murine monoclonal (Cell Signaling Technology, Danvers, Massachusetts, USA) as primary and anti-mouse horseradish peroxidase–conjugated (Promega, Fitchburg, Wisconsin, USA) as secondary antibody—was done as described before.
Results
Identification of Recurrent and Novel PRPH2 Mutations
A cohort of 310 French families with autosomal dominant retinitis pigmentosa was screened for the 3 exons of the PRPH2 gene ( NM_000322.4 ). We found that 32 probands (10.3%) carried a mutation. A total of 15 different mutations were identified ( Table 1 ). Nine of them were previously described, including 1 nonsense (p.Arg46*) and 8 missense mutations (p.Leu126Pro, p.Cys165Tyr, p.Trp179Arg, p.Ser198Arg, p.Gly208Asp, p.Phe211Leu, p.Pro216Ser, and p.Cys222Ser). Six others were novel, including 4 missense (p.Asp194Glu, p.Trp246Cys, p.Ala253Glu, and p.Leu254Gln), 1 frameshift (p.Val69Cysfs*30), and 1 splice site (c.829-4C>G) mutation. All mutations co-segregated with the disease phenotype in available family members ( Figures 1 and 2 ). The novel mutations were not identified in 96 ethnically matched control individuals and were not present in the public human SNP databases (including dbSNP, Ensembl, HapMap, the 1000 Genomes project and Exome Variant Server).
Nucleotide Change | Exon | Protein Change | Region | PolyPhen2 | SIFT | a-GVGD a | EVS | Reference |
---|---|---|---|---|---|---|---|---|
c.136C>T | 1 | p.Arg46* | D1 | N.A. | N.A. | N.A. | 0/13 006 | Meins et al |
c.205delG | 1 | p.Val69Cysfs*30 | 2 nd TMD | N.A. | N.A. | N.A. | 0/13 006 | Present study |
c.377T>C | 1 | p.Leu126Pro | D2 | Prob. | APF | C65 | 0/13 006 | Renner et al |
c.494G>A | 1 | p.Cys165Tyr | D2 | Prob. | APF | C65 | 0/13 006 | Souied et al |
c.535T>C | 1 | p.Trp179Arg | D2 | Prob. | APF | C65 | 0/13 006 | Bareil et al |
c.582T>A | 2 | p.Asp194Glu | D2 | Pos. | TOL | C35 | 0/13 006 | Present study |
c.594C>G | 2 | p.Ser198Arg | D2 | Prob. | APF | C65 | 0/13 006 | Sullivan et al |
c.623G>A | 2 | p.Gly208Asp | D2 | Pos. | APF | C65 | 1/13 006 | Kohl et al |
c.631T>C | 2 | p.Phe211Leu | D2 | Prob. | APF | C15 | 0/13 006 | Ekström et al |
c.646C>T | 2 | p.Pro216Ser | D2 | Pos. | TOL | C65 | 0/13 006 | Fishman et al |
c.664T>A | 2 | p.Cys222Ser | D2 | Prob. | APF | C65 | 0/13 006 | Downs et al |
c.738G>C | 2 | p.Trp246Cys | D2 | Prob. | APF | C65 | 0/13 006 | Present study |
c.758C>A | 2 | p.Ala253Glu | D2 | Prob. | TOL | C65 | 0/13 006 | Present study |
c.761T>A | 2 | p.Leu254Gln | D2 | Prob. | APF | C65 | 0/13 006 | Present study |
c.829-4C>G | Int. 2-3 | Splice site defect (p.Glu276_Val277insGln) | 4 th TMD | N.A. | N.A. | N.A. | 0/13 006 | Present study |
a a-GVGD scores: amino acid substitutions on a 7-scale scoring system, from C0 (neutral) to C65 (the most likely pathogenic); C35 is considered intermediate.
Among the novel mutations, the truncating p.Val69Cysfs*30 mutation led to a premature termination located within the second transmembrane α-helix of peripherin-2. No affected family members were available to test the familial segregation for the p.Asp194Glu mutation ( Figure 2 , Bottom right), but Asp194 is conserved in 16 peripherin-2 orthologs ( Figure 3 ) and is surrounded by residues Lys193 and Arg195, which have been found mutated previously. Moreover, the substitution p.Asp194Glu was predicted to be damaging by PolyPhen2 and align-GVGD programs but not by SIFT ( Table 1 ). For the mutations p.Trp246Cys and p.Ala253Glu, both residues at positions 246 and 253 are also evolutionary conserved ( Figure 3 ), and Trp246 has been previously found mutated in p.Trp246Arg. These 2 mutations were predicted to be damaging by PolyPhen2, align-GVGD, and SIFT but tolerated by SIFT for p.Ala253Glu ( Table 1 ).
We identified 4 families (PHRC057, PHRC069, PHRC161, and PHRC162) with the novel missense mutation, c.761T>A (p.Leu254Gln), with all affected subjects heterozygous for the mutation except 2 homozygous brothers (II:2 and II:3) in Family PHRC161. These 2 subjects had presumed consanguineous parents, while unaffected individuals did not carry the mutation ( Figure 2 , Left). The evolutionary conserved Leu254 is located in the D2 loop ( Figures 3 and 8 ) and the substitution p.Leu254Gln is predicted to be damaging by PolyPhen2, SIFT, and align-GVGD programs ( Table 1 ). In order to investigate whether p.Leu254Gln was a founder mutation, we genotyped the microsatellite markers D6S1575, D6S1549, D6S1552, D6S282, and D6S1650 that spanned the 2.98 Mb surrounding PRPH2 in the available DNA samples in the 4 families. We found that all affected members of the 4 families shared an identical allele for the 5 markers, except Patient II:2 of Family PHRC161, who had a cross-over between D6S1552 and D6S1549 ( Figure 2 , Left). Since the 4 families originated from the same area in the south of France, this indicates a founder effect. We confirmed the linkage at this locus with microsatellite markers reaching a maximum cumulated LOD (logarithm of odds) score of 4.484 for D6S1575 ( Figure 2 , Left). Since many patients carried the p.Leu254Gln, we performed biochemical investigations of the mutated peripherin-2. The WT and the mutated L254Q peripherin-2 proteins were expressed in yeast. We found that both purified WT and L254Q mutant showed monomers and formed dimers ( Figure 4 ). However, aggregates, which were present in both WT and mutated protein extracts, were much more abundant with the L254Q mutant. In addition, in the absence of the reducing agent β-mercaptoethanol in the sample buffer, the amounts of monomeric and dimeric L254Q were dramatically decreased compared to the WT. Thus, the L254Q mutant exhibited a strong tendency to form large aggregates, which might suggests abnormal folding for L254Q mutant.
Five independent families (PHRC011, PHRC084, PHRC197, PHRC276, and Fam716) had the c.829-4C>G mutation ( Figure 2 , Top right). Two algorithms (Human Splicing Finder and MaxEnt) predicted that the c.829-4C>G mutation would create an acceptor splice site located 3 base pairs upstream of the natural splice site and lead to the in-frame insertion of 1 glutamine between amino acids 276 and 277 (p.Glu276_Val277insGln) in the fourth transmembrane α-helix of peripherin-2 ( Figure 8 ). In 4 of the 5 families where several family members were available, the mutation was found to co-segregate with the disease. Only individual IV:2 of Family PHRC197 harbored the mutation and was presumed to be unaffected, but he was never examined. No common haplotype for 5 microsatellite markers (D6S1575, D6S1549, D6S1552, D6S282, and D6S1650) surrounding PRPH2 was found (data not shown) and the families were not originating from the same area, suggesting that c.829-4C>G could be a mutation hot spot.
Clinical Characterization of Patients With PRPH2 Mutations
From 27 to 67 patients were available for clinical analysis, depending on the type of examination. On average, the age at presentation was 45.2 ± 17.5 years (n = 44, range 13-78). The initial symptom was night blindness with an apparent age of onset at 30.8 ± 13.8 years (n = 29, range 10-57). Almost half the patients (31/67, 46%) were emmetropic (spherical equivalent −1 to +1), 36% were myopic (SE < −1), and 18% were hypermetropic (SE > +1), showing a skew toward moderate myopia ( Figure 5 , Top row, left).
We found that cataract, typically present in adult patients with retinitis pigmentosa, was encountered mostly in patients older than 40 ( Figure 5 , Top row, right). Visual acuity (VA) (decimal fraction) was variable with age ( Figure 5 , Middle row, left), 29 of 81 eyes (35.8%) having a normal visual acuity (VA = 1) in patients aged 32.3 ± 15.2 years (range 13-61), 34 of 81 eyes (42.0%) having a moderately decreased VA (0.9-0.5) in patients aged 47.9 ± 15.4 years (range 29-78), and 18 of 81 eyes (22.2%) having a severely decreased VA (≤0.4) in patients aged 61.2 ± 6.4 years (range 43–72). The decrease in VA was significantly correlated with age (r = -0.64; P < .001). The visual field also decreased progressively with age ( Figure 5 , Middle row, right). We found that 32 of 62 patients (51.6%) kept more than 50% of their visual field, being aged 37.7 ± 13.9 years (range 16–59), while 30 of 62 (48.4%) had lost more than 50%, being aged 54.8 ± 17.6 years (range 16–78). The decrease in visual field was significantly correlated with age (r = -0.56; P < .001). The rod ERG (dim blue) was recordable (b wave ≥10 μV) in 26 of 60 eyes (43.3%) from patients aged 34.2 ± 16.8 years (range 16–61) and was undetectable in 34 of 60 eyes (56.7%) from patients aged 54.2 ± 11.5 years (range 35–78) ( Figure 5 , Bottom row, left). The cone ERG (30 Hz flicker) was recordable (b wave amplitude ≥5 μV) in 49 of 54 eyes (90.7%) from patients aged 43.9 ± 18.5 years (range 16–78) and was undetectable in 5 of 54 eyes (10.3%) from patients aged 52.4 ± 6.8 years (range 45–58) ( Figure 5 , Bottom row, right). Both the rod and the cone ERG decrease was correlated with age, r = −0.62 and −0.44; P < .001 for rod and cone function, respectively.
Fundus examination revealed the presence of pigment deposits in 73% of the patients with a mean age of 45 ± 18 years. Fundus autofluorescence imaging revealed abnormalities in 62.9% (age 47 ± 18 years), including macular autofluorescence ring and atrophic spots in periphery ( Table 2 ). On OCT examination, the majority of patients retained their ellipsoid zone at the fovea (70.4%), whereas a minority had a cystoid macular edema (14.3%). We noticed that some patients had macular involvement with either normal, moderately reduced ( Figure 6 , Row 1, outer left), or severely decreased ( Figure 6 , Row 1, inner left) visual acuity. Some patients had a mild RP with a few spots of atrophy in the retinal periphery and macular sparing ( Figure 6 , Row 1, inner right). In other cases, typical pigment deposits and widespread atrophy in the midperipheral retina was present ( Figure 6 , Row 1, outer right). Some patients showed a pericentral localization of the retinal lesions even if other members of the family had a widespread form ( Figure 6 , Row 2, outer left). This was particularly evident in Family PHRC281 carrying the p.Pro216Ser ( Figure 6 , Row 2, inner left to outer right), in which a family member (III:2) had a pericentral localization of the retinal lesions sharply delimited from the unaffected peripheral retina while her sister (III:1) had a typical widespread retinitis pigmentosa. In a few circumstances, the presence of yellow deposits was noticed, as in Family PHRC305 carrying the p.Pro216Ser, in which the mother had typical retinitis pigmentosa ( Figure 6 , Row 3, outer left) and the son a vitelliform foveal deposit but no signs of retinitis pigmentosa ( Figure 6 , Row 3, inner left).