Genetics and Phenotypes of RPE65 Mutations in Inherited Retinal Degeneration

Abstract

purpose. To characterize the spectrum of RPE65 mutations present in 453 patients with retinal dystrophy with an interest in understanding the range of functional deficits attributable to sequence variants in this gene.

methods. The 14 exons of RPE65 were amplified by polymerase chain reaction (PCR) from patients’ DNA and analyzed for sequence changes by single-strand conformation polymorphism (SSCP) and direct sequencing. Haplotype analysis was performed using RPE65 intragenic polymorphisms. Patients were examined clinically and with visual function tests.

results. Twenty-one different disease-associated DNA sequence changes predicting missense or nonsense point mutations, insertions, deletions, and splice site defects in RPE65 were identified in 20 patients in homozygous or compound heterozygous form. In one patient, paternal uniparental isodisomy (UPD) of chromosome 1 resulted in homozygosity for a probable functional null allele. Eight of the disease-associated mutations (Y79H, E95Q, E102X, D167Y, 669delCA, IVS7+4a→g, G436V, and G528V) and one mutation likely to be associated with disease (IVS6+5g→a) have not been reported previously. The most commonly occurring sequence variant identified in the patients studied was the IVS1+5g→a mutation, accounting for 9 of 40 (22.5%) total disease alleles. This splice site mutation, as well as R91W, the most common missense mutation, exists on at least two different genetic backgrounds. The phenotype resulting from RPE65 mutations appears to be relatively uniform and independent of mutation class, suggesting that most missense mutations (15 of 40 disease alleles [37.5%]) result in loss of function. At young ages, this group of patients has somewhat better subjective visual capacity than is typically associated with Leber congenital amaurosis (LCA) type I, with a number of patients retaining some useful visual function beyond the second decade of life.

conclusions. RPE65 mutations account for a significant percentage (11.4%) of disease alleles in patients with early-onset retinal degeneration. The identification and characterization of patients with RPE65 mutations is likely to represent an important resource for future trials of rational therapies for retinal degeneration.

The retinal pigment epithelium (RPE) performs a number of functions critical for visual processing, metabolism, and survival of the photoreceptor cells of the retina. ¹ Mutations in genes expressed in the RPE have recently emerged as an important cause of inherited retinal degeneration. RPE65 was the first identified RPE-specific disease gene as a result of linkage studies that mapped the disease locus in a consanguineous family to the interval containing RPE65 on chromosome 1p31 and identified mutations responsible for childhood-onset severe retinal dystrophy. ² In addition, missense mutations in RPE65 were identified in a patient with LCA type II by using the candidate gene approach. ³ Since these initial reports, a wide range of disease severity has been associated with RPE65 mutations, from congenital blindness to adult-onset retinitis pigmentosa, although the most common phenotype is severe and early-onset retinal degeneration. ^{4 5 6 7 8}  

RPE65 encodes an abundant and evolutionarily conserved 61-kDa protein associated with the smooth endoplasmic reticulum of the RPE. ^{9 10} Although the specific function of RPE65 is not yet known, a body of evidence indicates that it plays an essential role in vitamin A metabolism necessary for the synthesis of the visual pigment chromophore 11-cis retinal. Studies of Rpe65 knockout mice show that disruption of the gene results in a severely depressed electroretinogram (ERG), absence of the rhodopsin photopigment, and accumulation of all-trans retinyl esters in droplets within the RPE. ¹¹ Similar ultrastructural abnormalities are also present in a strain of Swedish Briard dogs that carry a functional null allele of RPE65. ^{12 13} These findings indicate that loss of RPE65 function results in a block in retinoid processing after esterification of vitamin A to membrane lipids, however, the mechanism by which RPE65 participates in retinoid isomerase activity of the RPE remains to be elucidated. 

Inherited defects in vitamin A metabolism and other RPE-specific functions are likely to have a unique significance for research into the causes and treatment of retinal degeneration, since disorders affecting the RPE are, in principle, more accessible to therapeutic intervention than disorders directly affecting the proteins of the photoreceptor cells. Patients with RPE-specific defects may therefore be candidates for targeted therapies likely to become available in the near future. A necessary prerequisite for relevant clinical trials is the large-scale ascertainment and characterization of such individuals. However, few RPE65 patients have been well characterized, and the range of associated phenotypes has not been fully defined. In addition, a number of questions about the disease caused by RPE65 mutations remain unanswered, due in part to the limited amount of data on the mutation spectrum of RPE65. 

We now report the results of RPE65 mutation analysis in a large collection of patients with retinal dystrophy from the United States and Europe. Our findings include the identification of a number of novel mutations. Our data are presented in the context of other known RPE65 mutations, including analysis of mutation class, prevalence, and associated phenotype. 

Methods

Subjects

RPE65 mutation screening was performed in patients with retinal dystrophy representing diverse forms of the disease. Informed consent was obtained from the patients, according to study protocols approved by university internal review boards for human subject studies. Our research followed the tenets of the Declaration of Helsinki. 

Patient Evaluation

Ophthalmic examinations of patients treated in our clinics included slit lamp biomicroscopy; assessment of visual acuity, color perception, and visual fields; electroretinograms (ERGs); and ascertainment of family history. For other patients, clinical descriptions were provided by local ophthalmologists. 

Mutation Screening

DNA was prepared from whole blood using standard methods. After polymerase chain reaction (PCR) amplification of individual or groups of exons, single-strand conformation polymorphism (SSCP) analysis was used to screen for DNA sequence changes in and near gene coding regions using oligonucleotide primers and conditions published previously. ^{2 14} Direct DNA sequence analysis using chain terminator cycle sequencing technology (Amersham Pharmacia Biotech, Piscataway, NJ) with the same primer pairs used for PCR amplification was used to confirm suspected DNA sequence changes, as well as for primary screening in approximately one third of cases. 

Sequence Analysis

Secondary structure predictions were made using the predictive algorithms of Chou and Fasman ¹⁵ and Garnier et al. ^{16 17} Potential posttranslational modification sites were identified using Prosearch to scan the Prosite database. ¹⁸ Coefficients of splice site efficiency were calculated according to Shapiro and Senapathy. ¹⁹  

Genotype Analysis

For analysis of individual mutation-associated haplotypes the microsatellite-type polymorphism located 2.8 kb upstream of the RPE65 transcription start site (locus D1S2803) was used. ²⁰  

Results

Mutation Screening

To determine the relative contribution of RPE65 mutations to the causes of inherited retinal degenerations, we screened a group of 453 unrelated patients with various forms of retinal dystrophy using a combination of SSCP and direct DNA sequencing. Our analysis resulted in the identification of 21 different disease-associated DNA sequence variants in 20 patients that predict missense or nonsense point mutations, insertion, deletion, and splice site defects in RPE65 (Table 1 , upper). These sequence variants were present in homozygous (homo) form in 7 patients and in compound (cmpd) heterozygous form in 13 patients. Disease relevance was suggested by segregation analysis (data not shown), with two notable observations. In one case (patient arRP850), two different sequence variants (1114delA and T457N) were detected on the same allele. Because 1114delA predicts a functional null mutation, the pathogenic potential of the T457N missense mutation cannot be determined. In a second case (patient 1723), segregation data were not compatible with Mendelian inheritance, in that the homozygous IVS1+5g→a mutation present in the patient was found to be carried only by his father. Analysis of three DNA polymorphisms in RPE65 and 29 genetic markers spread out along both arms of chromosome 1 showed the patient to be homozygous for the paternal allele in all cases, with no inconsistencies seen for any other chromosome tested (data not shown). These findings indicate that the patient is homozygous for the IVS1+5g→a mutation due to uniparental isodisomy (UPD) of chromosome 1 (Thompson D, Gal A, unpublished observations, June 2000). 

In eight additional patients, six DNA sequence variants were detected on only one RPE65 allele (Table 1 , lower). The disease relevance of three of these sequence variants (R85H, K294T, and N321K) is uncertain, because in two cases (R85H and N321K), the patients were heterozygous, although they were the children of consanguineous marriages. In addition, the case of N321K, this sequence change was detected (on one allele) in screens of 50 control individuals from the general population. This variant, however, has been detected in an unrelated patient in apparently compound heterozygous form. ²¹ In the third case (K294T), in one of three families, only one of two affected siblings carried this sequence change. In contrast, disease relevance seems likely for the other three sequence variants identified in heterozygous form (IVS1+5g→a, A132T, and IVS6+5g→a), because these are either found in unrelated patients in homozygous form (this work and Reference 4) and/or predict functional null alleles. In the latter three heterozygous patients, mutation screening of an additional 650 bp of sequence from the RPE65 proximal promoter region, as well as the sequence from− 2361 to −2599 containing potential Oct-1 and E47/Th1 sites, ²⁰ did not result in the identification of a second disease-associated sequence variant. 

Missense Mutations

The disease-associated missense mutations detected in our studies predict amino acid substitutions resulting in changes in charge (R91W, R91Q, E95Q, and D167Y), polarity (Y79H and Y368H), translation initiation (M1T), and (potentially) structure (P363T, G436V, and G528V). All missense mutations affect amino acid residues that are conserved among human, bovine, canine, rat, chicken, and salamander sequences, ^{9 10 12 22 23 24} and were not present in 50 control individuals. This finding reflects the high overall conservation of the protein, ranging from 98.7% identity between humans, cattle, and dogs, to 85.7% identity between humans and salamanders. The positions of the amino acid substitutions were distributed relatively evenly throughout the length of the protein. The predicted protein structure contained a high proportion of β-pleated sheet (in a ratio of two to one with α-helical regions), with 8 of 10 missense mutations located within β-pleated sheet domains (not shown). The amino acid substitutions did not predict major changes in overall protein folding or structure (using the Chou and Fasman or Garnier et al. algorithms ^{15 16 17} ), and only substitutions at arginine-91 were predicted to perturb the localβ -pleated sheet structure (changing it to α-helix). The sequence changes also did not disrupt predicted posttranslational modification sites, including consensus sequences for N-linked glycosylation, protein kinase C phosphorylation, casein kinase II phosphorylation, tyrosine kinase phosphorylation, and N-linked myristoylation. It should be noted, however, that the functional significance of these sites has not been established, and there is evidence that the mature protein contains neither O- nor N-linked glycans. ²⁵ The sequence changes also did not create new donor or acceptor splice site consensus sequences with calculated splicing coefficients ¹⁹ greater than or equal to the values for nearby native sites (data not shown). Thus, the effect of missense mutations is unlikely to be at the level of the transcript. Disease-associated missense mutations in this population (15 of 40 disease alleles [37.5%] occurred with slightly lower frequency than mutations predicting functional null alleles. 

Null Mutations

In our studies, more than half of the disease-associated DNA sequence variants identified (11/20) predicted functional null alleles resulting from nonsense mutations (E102X, R124X, and R234X), a 1-bp insertion (144insT), small deletions (344del20, 669delCA, 831del8, and 1114 delA), and splice site mutations (IVS1+5g→a, IVS7+4a→g, and IVS8 + 1g→t). The mutations are not clustered and are likely to result in the production of truncated protein, in some cases containing unrelated amino acid residues, or more likely, in complete absence of the protein due to greatly reduced mRNA/protein stability. 

RPE65 Mutation Prevalence

In our total population of 453 patients screened, 339 were from an unselected collection of patients with retinal dystrophy, and 114 were included on the basis of a clinical diagnosis of LCA or early-onset retinal dystrophy. Of the latter 114 patients, 13 were found to have mutations in both RPE65 alleles (11.4%). Our data further show that RPE65 mutations accounted for 2.1% (7/339) of patients with autosomal recessive retinal dystrophy. The IVS1+5g→a mutation was the most common of all sequence variants identified, accounting for 9 of 40 total disease alleles (22.5%). Haplotype analysis indicated that the IVS1+5g→a allele occurred on at least two genetic backgrounds (Fig. 1 , lanes 1, 2, and 3). Substitutions at arginine-91 were also common, with mutations at this position occurring in four families. Analysis of these families indicated that the R91W allele also arose independently on at least two genetic backgrounds (Fig. 1 , lanes 4, 5, and 6). Among single nucleotide changes, transitions (13/23) occurred at a higher frequency than transversions (10/23). Eight of the disease-associated mutations identified in the present study (Y79H, E95Q, E102X, D167Y, 669delCA, IVS7+4a→g, G436V, and G528V) and one mutation likely to be associated with disease (IVS6+5g→a) have not been reported previously. 

Patient Phenotypes

The patients included for screening in our study were affected by retinal degeneration associated with a range of disease presentations, with many instances of severe, early-onset forms of disease represented. The phenotype of patients carrying RPE65 mutations, however, appeared to be more uniform than that of the screening population as a whole. A summary of relevant data for patients with both disease alleles identified is shown in Table 2 . In most patients with RPE65 mutations, disease was diagnosed in infancy, with visual impairment frequently associated with nystagmus, night blindness, and a tendency to fixate on light. Photophobia was not observed in this group. Constricted visual fields were documented at young ages when measured. In most cases, the retina appeared pale without significant pigment accumulation and with RPE atrophy in the periphery. The data available for very young patients indicate that rod ERGs are undetectable, and cone ERGs is severely diminished at the earliest ages measured (approximately 1 year old), with cone ERGs becoming unrecordable by 5 to 7 years of age. ⁷ Despite such poor ERG indicators, the visual performance of several patients in bright light was sufficient to permit attendance at regular school during the elementary years. At older ages, often during the secondary school years, visual acuity was greatly reduced. However, a number of patients retained residual islands of central or peripheral vision into their third decade, with only one report of no light perception in a patient 25 years old. In the group of 20 patients in whom disease-associated mutations in both RPE65 alleles were identified, 10 (50%) patients had two apparent null alleles, 5 (25%) patients had two missense mutations, and 5 (25%) patients had both an apparent null allele and a missense mutation. However, no obvious difference in visual performance appeared to exist among patients affected with different categories of mutations. For example, patient arRP341 carried two apparent null alleles (IVS1+5g→a, 144insT), patient LCA820 carried a homozygous missense mutation (R91W), and patient 2711 carried a missense mutation and an apparent null allele in compound heterozygous form (Y368H, IVS1+5g→a). At young ages, the clinical descriptions and best visual acuities of all three patients (20/100–20/200) were virtually indistinguishable. 

Discussion

As a result of screening a large and diverse collection of individuals with various presentations of retinal dystrophy, we found that RPE65 mutations are a common cause of congenital or early-onset retinal degeneration, responsible for the disease in 11.4% of patients with mutations present in homozygous or compound heterozygous form. This prevalence is approximately equal to that reported for RPE65 mutations in patients with LCA in whom both mutations were identified in two earlier studies (7/45, 15.6% ⁴ ; 5/54, 9.3% ⁶ ) and is greater than that reported in a recent study (7/176, 4% ⁸ ). UPD, the situation in which an individual inherits two copies of a specific chromosome from one parent and no copy from the other, ²⁶ was also identified in our studies as the cause of retinal dystrophy in one patient as a result of reduction to homoallelism for RPE65. The disease-associated mutations identified by us, along with other published RPE65 mutations, are shown on the schematic of the gene structure in Figure 2 . The mutations identified in our studies account for approximately half of all mutations currently known, and are representative of the group as a whole in terms of mutation type, location, and associated phenotype. 

The autosomal recessive nature of the disease resulting from RPE65 mutations predicts that missense mutations result in loss or considerable reduction of protein function. For two of the missense mutations identified in our studies, R91W and P363T, the connection to disease pathogenesis is very likely, because these sequence variants segregate with the disease phenotype in patients in homozygous form. This was also the case for the mutations M1T and A132T identified in homozygous form by others. ⁴ Most of the other missense mutations identified in our studies (Y79H, R91Q, E95Q, D167Y, P363T, Y368H, G436V, and G528V) are likely to be pathogenic based on genetic evidence. In contrast, there was no strong evidence linking four of the identified missense mutations (R85H, K294T, N321K, and T457N) to the disease state. These sequence changes may therefore represent rare variants of the RPE65 gene which are not, in themselves, causal for the observed phenotype, but which could play a role, for example, in multifactorial retinal diseases with late onset. ²⁷ Indirect evidence that certain amino acid substitutions produce subtle changes in RPE65 activity has been obtained in recent studies showing association between an Rpe65 polymorphism and susceptibility to light damage in strains of inbred mice. ²⁸ As the characterization of the vitamin A cycle continues to develop on a molecular level, it should be possible to devise strategies to explore the possibility that defects in multiple genes may have an interactive effect in causing disease. 

The mechanisms by which RPE65 mutations, in general, contribute to pathogenesis are not yet known and, in part, await elucidation of the specific role of the protein in RPE physiology. Because all known missense mutations affect highly conserved residues but do not appear to disrupt protein folding or posttranslational processing and do not cluster within the linear protein sequence, it may be that these mutations inactivate a functional domain or domains present in the folded protein structure. Such mutations may be predicted to interfere with protein–protein interactions, subcellular localization, ligand binding, or intrinsic enzymatic activity necessary for the synthesis of 11-cis retinal. Each of these hypotheses will be testable when assays of the specific function(s) of the RPE65 protein become available. 

Our inability to detect a second RPE65 mutant allele in three patients identified as having one probable disease-associated mutation, a general finding also reported by other groups, ^{4 6 8 21} may be due to the presence of large deletions or other rearrangements undetected by current screening methods. Alternatively, mutations may occur in other regions of the RPE65 gene not analyzed in our study, including promoter and intronic regions. Mutation screening of these sequences is not practical at this time, because the critical elements that regulate RPE65 promoter activity have not yet been identified, ²⁰ and no highly conserved intronic sequences beyond the splice site consensus sequences are known. ²⁹ Other possibilities include dominant effect of the mutation, digenic inheritance involving the mutation of a second interacting gene, ³⁰ or causal mutations in an unrelated gene. Another important issue to address in future studies is whether heterozygous individuals are at increased risk for vision loss in later life, especially in association with aging. 

The severity of the disease resulting from mutations in RPE65 appears to be largely independent of the mutation types present in these patients. Previous studies have suggested that severe vision loss is the result of null mutations affecting both RPE65 alleles, whereas milder forms of disease result in cases when at least one of the two mutations is a missense allele. ^{5 31} Such a correlation has also been proposed to exist for disease severity and mutations in the ATP-binding cassette transporter of rods (ABCR) gene in autosomal recessive Stargardt disease, fundus flavimaculatus, cone–rod dystrophy, and retinitis pigmentosa. ³² However, our studies now show that a severe phenotype can result from a number of different combinations of RPE65 null and missense mutations. Together, these findings suggest the possibility that some missense mutations may result in true null alleles, whereas others may simply reduce the effectiveness of the protein product. Alternatively, or in addition, variability in disease severity may be determined by modifier genes that impact RPE65-related cell biology. Resolution of this issue also awaits the development of functional tests of mutant RPE65 protein. 

The initial reports describing RPE65 mutations defined the associated phenotype as a childhood-onset, severe retinal dystrophy ² and as LCA. ³ Subsequently, it has been proposed that patients with LCA who have RPE65 mutations can be distinguished from patients who have mutations in the photoreceptor-specific guanylate cyclase gene, RetGC1, on clinical grounds. ^{6 33} We find that many RPE65 patients share a common phenotype characterized by poor but useful visual function in early life (measurable cone ERGs) that declines dramatically throughout the school age years. In addition, a number of these patients retain residual islands of peripheral vision, although considerably compromised, into the third decade of life. Thus, the phenotype resulting from RPE65 mutations appears relatively uniform, possibly because each mutation exerts its effect by producing similar deficits in RPE function. It seems likely that the use of various diagnostic designations for these patients, including LCA II, early-onset severe retinal dystrophy, autosomal recessive retinal dystrophy, and early severe retinitis pigmentosa, merely reflects usage preferences by individual ophthalmologists rather than actual phenotypic differences that define patient subtypes. 

The phenotype and functional defects resulting from RPE65 mutations, as well as the existence of both mouse ¹¹ and canine ¹² models of the disease, makes this patient population attractive candidates for future therapeutic trials focused on manipulation of the vitamin A cycle. Patients with the RPE65 mutation who have onset of disease in infancy and who retain reasonable visual function that is lost only over the course of many years would seem to be ideal subjects for therapeutic intervention. Identifying these individuals at young ages will enhance therapeutic opportunities. Research from many laboratories over the next few years will determine which of the many approaches currently under study, including gene therapy, RPE transplantation, and retinoid and survival factor therapy, may have the greatest potential for success in this group of patients. Meanwhile, in anticipation of such trials, continued characterization of this patient population, as well as the corresponding animal models, remain important goals for the immediate future. 

 Presented in part at the annual meetings of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 1998 and 1999.

 Supported by National Institutes of Health (Grants EY05627, EY06094, EY07003, EY09193), The Foundation Fighting Blindness, Research to Prevent Blindness, Daniel Matzkin Research Fund, the Deutsche Forschungsgemeinschaft Grants GA 210/12-1, SFB 430/C2, and the German and British Pro Retina Associations.

 Submitted for publication April 24, 2000; revised August 2, 2000; accepted August 30, 2000.

 Commercial relationships policy: N.

 

Table 1.

 

Summary of RPE65 Sequence Variants Detected in Patients with Retinal Dystrophy

Exon/Intron	Mutation	Predicted Effect	Restriction Site	Patient	Status
Exon 1	56T→C	M1T	+Mae II	arRP192	Cmpd
Intron 1	IVS1+5g→a	Inactive splice site	+Ssp I	arRP181	Homo
				arRP341	Cmpd
				arRP697	Cmpd
				arRP713	Cmpd
				arRP850	Cmpd
				LCA821	Cmpd
				1723	Homo/UPD
				2711	Cmpd
Exon 2	144ins T	Frame shift	+Mse I	arRP341	Cmpd
Exon 3	289T→C	Y79H	No enzyme	arRP76	Cmpd
Exon 4	325C→T	R91W	−Rsa I	arRPL	Cmpd
				arRP192	Cmpd
				LCA820	Homo
Exon 4	326G→A	R91Q	+Sun I	1024	Cmpd
Exon 4	337G→C	E95Q	+Hinf I	arRP76	Cmpd
Exon 4	344del 20	Frame shift		1348	Homo
Exon 4	358G→T	E102X	−Apo I	arRP849	Homo
				arRP697	Cmpd
Exon 5	424C→T	R124X	−Taq I	arRP114	Cmpd
Exon 6	553G→T	D167Y	−Mbo I	arRP188	Cmpd
Exon 6	669del CA	Frame shift	+Mun I	arRP713	Cmpd
				LCA816	Cmpd
Exon 7	754C→T	R234X	+AlwN I	arRP114	Cmpd
Intron 7	IVS7+4a→g	Inactive splice site	+Alu I	1024	Cmpd
Exon 8	831del 8	Frame shift	+Apo I	arRP188	Cmpd
Intron 8	IVS8+1g→t	Inactive splice site	+Mse I	PMK18/1	Homo
Exon 10	1114delA	Frameshift	No enzyme	arRP850	Cmpd
Exon 10	1141C→A	P363T	−BspM I	PMK30/265	Homo
Exon 10	1156T→C	Y368H	+Nla III	arRPL	Cmpd
				2711	Cmpd
Exon 12	1361G→T	G436V	+Rsa I	LCA821	Cmpd
Exon 14	1637G→T	G528V	+Rsa I	LCA816	Cmpd
Intron 1	IVS1+5g→a	Inactive splice site	+Ssp I	LCA826	Hetero
Exon 4	308G→A	R85H	+Msl I	arRP870	Hetero
Exon 5	448G→A	A132T	No enzyme	arRP95	Hetero
Intron 6	IVS6+5g→a	Inactive splice site	No enzyme	arRP476	Hetero
Exon 9	935A→C	K294T	No enzyme	208	Hetero
				858	Hetero
				1369	Hetero
Exon 9	1017T→G	N321K	No enzyme	PMK29/1	Hetero
Exon 13	1424C→A	T457N	+Tsp509 I	arRP850	Cmpd

 Homo, homozygous; Hetero, heterozygous; Cmpd, compound heterozygous. Key to fonts: Bold, novel sequence variants reported in this study; Bold italics, novel sequence variants reported in our previous studies ^{2 7 21} ; Regular, sequence variants reported in this study also identified in patients in other studies ^{3 4 6 8} ; Italics, sequence variants reported in our previous study also identified in patients in other studies. ^{3 4 6 8} Nucleotide positions are numbered from the transcription initiation site located 54 bp upstream of the initiation codon.

Figure 1.

 

Genotype analysis of six patients with retinal dystrophy with RPE65 mutations IVS1+5g→a or R91W using D1S2803. Alleles of the RPE65 intragenic microsatellite polymorphism D1S2803 were determined for patients with IVS1+5g→a mutations: arRP713 (cmpd), arRP181 (homo), and arRP341 (cmpd) (lanes 1, 2, and 3); and for patients with R91W mutations: arRP192 (cmpd), LCA820 (homo), and arRPL (cmpd; lanes 4, 5, and 6). The data indicate that both IVS1+5g→a and R91W arose independently on at least two different alleles.

Table 2.

 

Clinical Characteristics of Patients with Disease-Associated RPE65 Mutations

Patient	Mutation	Age at Onset and Diagnosis	Visual Acuity, Visual Fields, Refraction	ERG Data	Comments and Recently Ascertained Visual Function
arRPL	R91W, Y368H	7 mo, nystagmus, fixation on light	7 y, VA 20/100; VF V4+; III4, II/4−	7 mo, rod and cone ERGs residual; 7 y, ERGs nonrecordable	7 y, current age*
arRP76	Y79H, E95Q	3 y, nightblind; 20 y, diagnosis RP	No data	No data	40 y, able to read; 58 y, small islands of residual vision
arRP114	R124X, R234X	Birth, nightblind; 5 y, diagnosis RP	No data	No data	17 y, able to read; 42 y, LPO
arRP181	IVS1+5g→a, IVS1+5g→a	Young child, nightblind	No data	No data	Attended regular school; 10 y, RP diagnosis; 14 y, able to read; 18 y, legally blind
arRP188	D167Y, 831del8	Birth, nightblind	No data	No data	19 y, unable to walk alone; 25 y, no LP, †
arRP192	M1T, R91W	2 y, nightblind	6 y, central VA acceptable in bright light	No data	18 y, poor VFs; 38 y, able to read; 48 y, LPO
arRP341	IVS1+5g→a, 144ins T	4 mo, nystagmus fixation on light	6 y, VA 20/200, RE and 20/100 LE; VF III4+, II4− hyperopic	1 y, rod absent, cone abnormal ERG; 6 y, ERGs nonrecordable	8 y, VF and monocular VAs unchanged*
arRP697	IVS1+5g→a, E102X	2 y, nystagmus, nightblind; 6 y, onset of VA loss	Best VA 20/200, myopic	18 y, ERG nonrecordable	18 y, legally blind; 31 y, unable to read; 37 y, CF, VFs unrecordable
arRP713	IVS1+5g→a, 669delCA	3 y, nystagmus, nightblind, poor VA	16 y, VA 20/200, VF V4 20°, III4−, hyperopic	29 y, ERG nonrecordable	30 y, VA 20/600; 35 y, LPO, VFs nonrecordable
arRP849	E102X, E102X	Infant, nystagmus, nightblind, poor VA	6 y, VF ring scotomas	6 y, rod ERG absent; 12 y, cone ERG absent	16 y, VA 20/400
arRP850	IVS1+5g→a, T457N, 1114delA	3 y, nystagmus	4 y, VA 20/100; VF III4+, II4−, hyperopic, astigmatism	6 y, rod and cone ERGs nonrecordable	6 y, current age*
LCA816	669del CA, G528V	6 mo, nystagmus, nightblind, poor VA	Best VA 20/200, myopic, astigmatism, 15 y, VF III4−, V4 10°	15 y, ERG nonrecordable	Attended school for the blind, 30 y, no LP and 20/400, VF nonrecordable
LCA820	R91W, R91W	3 y, nystagmus, nightblind, poor VA	Best VA 20/200, plano 23 y, VF V4 8°	23 y, ERG nonrecordable	30 y, VA 20/100 and 1/40, VF V/4 3°
LCA821	IVS1+5g→a, G436V	Birth, nystagmus, nightblind, poor VA	Best VA 20/100, hyperopic, astigmatism	19 y, ERG nonrecordable	19 y, rapid deterioration; 21 y, VF V4 5°; 25 y, VA 20/200 and 1/40
PMK18/1	IVS8+1g→t, IVS8+1g→t	Young child	9 y, VA 20/60 RE, CF LE	No data	11 y, current age, †
PMK30/265	P363T, P363T	3–7 y, nystagmus, nightblind	CF and hand movements	No data	Severe visual handicap between 5 and 12 y, †
1024	R91Q, IVS7+4a→g	Toddler, nystagmus, nightblind	20 y, VA 20/400; refraction−200	20 y, rod ERG absent, cone ERG abnormal	Residual central and peripheral field islands
1348	344del20, 344del20	Before 5 y, nystagmus	7 y, VA 20/200; refraction+100	7 y, abnormal cone ERG	16 y, entered school for blind using low vision aids
1723	IVS1+5g→a, IVS1+5g→a	Before 5 y, nystagmus	8 y, VA 20/200; refraction +400	35 y, rod and cone ERGs not detectable	50 y, small residual islands of vision
2711	IVS1+5g→a, Y368H	5 y, nystagmus, nightblind	5 y, VA 20/60–20/100 10 y, VF 90° to V4e, cannot see V2e	5 y, rod and cone ERGs not detectable	30 y, VF peripheral residual islands only to IV4e, V4e

 RP, retinitis pigmentosa; VA, visual acuity; VF, visual fields; CF, count fingers; LPO, light perception only.

*  For clinical description, see Lorenz et al. ⁷

†  For clinical description, see Gu et al. ²

Figure 2.

 

Schematic of the gene structure of RPE65 showing the positions of the 14 exons and patient mutations published to date (from References 2–8, 21, and 34), with mutations reported in the present study shown in bold.

The authors thank David Hanna, Angela Cassar, and Jingtong Gao for excellent technical assistance.