Fundus photograph from the right eye of patient 2 shows a small atrophic-appearing foveal lesion and a limited number of yellowish white fundus flecks in the perifoveal region, characteristic of patients with phenotype I.
Fluorescein angiogram from the right eye of patient 2 depicts the absence of a dark choroid, characteristic of patients with phenotype I.
Electroretinogram from the right eye of patient 2 shows normal cone (A) and rod (B) amplitudes, characteristic of patients with phenotype I.
Fundus photograph from the left eye of patient 16 shows the presence of numerous yellowish white fundus flecks and an atrophic-appearing foveal lesion, characteristic of patients with phenotype II.
Fluorescein angiogram from the left eye of patient 15 shows the presence of a silent or dark choroid, characteristic of patients with phenotype II. Hyperfluorescence associated with the fundus flecks and hypopigmentary changes of the retinal pigment epithelium is also apparent.
Electroretinogram from the right eye of patient 18 shows subnormal cone (A) and rod (B) amplitudes.
Fundus photograph from the right eye of patient 23 shows diffuse atrophic-appearing changes of the retinal pigment epithelium throughout the posterior retina that extend anterior to the vascular arcades to the midperipheral retina.
Fluorescein angiogram from the right eye of patient 23 shows diffuse hyperfluorescence and choriocapillaris atrophy within the macular region, characteristic of patients with phenotype III.
Electroretinogram from the left eye of patient 24 shows reduction of cone (A) and rod (B) amplitudes, characteristic of patients with phenotype III.
Fishman GA, Stone EM, Grover S, Derlacki DJ, Haines HL, Hockey RR. Variation of Clinical Expression in Patients With Stargardt Dystrophy and Sequence Variations in the ABCRGene. Arch Ophthalmol. 1999;117(4):504-510. doi:10.1001/archopht.117.4.504
EDWIN M.STONEMD, PHD
To report the spectrum of ophthalmic findings in patients with Stargardt dystrophy or fundus flavimaculatus who have a specific sequence variation in the ABCRgene.
Twenty-nine patients with Stargardt dystrophy or fundus flavimaculatus from different pedigrees were identified with possible disease-causing sequence variations in the ABCRgene from a group of 66 patients who were screened for sequence variations in this gene.
Patients underwent a routine ocular examination, including slitlamp biomicroscopy and a dilated fundus examination. Fluorescein angiography was performed on 22 patients, and electroretinographic measurements were obtained on 24 of 29 patients. Kinetic visual fields were measured with a Goldmann perimeter in 26 patients. Single-strand conformation polymorphism analysis and DNA sequencing were used to identify variations in coding sequences of the ABCRgene.
Three clinical phenotypes were observed among these 29 patients. In phenotype I, 9 of 12 patients had a sequence change in exon 42 of the ABCRgene in which the amino acid glutamic acid was substituted for glycine (Gly1961Glu). In only 4 of these 9 patients was a second possible disease-causing mutation found on the other ABCRallele. In addition to an atrophic-appearing macular lesion, phenotype I was characterized by localized perifoveal yellowish white flecks, the absence of a dark choroid, and normal electroretinographic amplitudes. Phenotype II consisted of 10 patients who showed a dark choroid and more diffuse yellowish white flecks in the fundus. None exhibited the Gly1961Glu change. Phenotype III consisted of 7 patients who showed extensive atrophic-appearing changes of the retinal pigment epithelium. Electroretinographic cone and rod amplitudes were reduced. One patient showed the Gly1961Glu change.
A wide variation in clinical phenotype can occur in patients with sequence changes in the ABCRgene. In individual patients, a certain phenotype seems to be associated with the presence of a Gly1961Glu change in exon 42 of the ABCRgene.
The identification of correlations between specific mutations in the ABCRgene and clinical phenotypes will better facilitate the counseling of patients on their visual prognosis. This information will also likely be important for future therapeutic trials in patients with Stargardt dystrophy.
A GENE ASSOCIATED with Stargardt dystrophy1 and fundus flavimaculatus2 has been mapped to the short arm of chromosome 1. This gene was subsequently identified to be a photoreceptor gene–specific ATP-binding transporter gene (ABCR)3 that is localized to the disc membrane of retinal rod outer segments.4 Sequence changes in this gene were also identified in patients with age-related macular degeneration5 and in a family with an autosomal-recessive form of retinitis pigmentosa.6
The intent of the present study was to identify the spectrum of retinal disease that might be encountered in patients with various sequence changes in the ABCRgene and a diagnosis of Stargardt dystrophy or fundus flavimaculatus. We observed 3 different phenotypes of retinal disease that encompass the wide variation of disease expression that can occur in patients with sequence changes in the ABCRgene.
We screened 66 patients from 56 families diagnosed as having Stargardt dystrophy or fundus flavimaculatus from a group of 330 patients seen by 1 of us (G.A.F.) for variations in coding sequences of the ABCRgene. Ninety-six control patients without Stargardt dystrophy were also screened.
Patients who were screened were selected to represent a spectrum in fundus changes that we encountered in patients classified as having Stargardt dystrophy or fundus flavimaculatus. In this article, the terms Stargardt dystrophy and fundus flavimaculatus are considered synonymous.
An ophthalmic examination was performed by one of us (G.A.F.) on all patients. This included best-corrected visual acuity with a Snellen projection chart or, in individual instances, a Feinbloom Distance Test Chart for the Partially Sighted. Slitlamp examination of the cornea, anterior chamber, lens, and vitreous was also performed. Ocular pressure was determined by applanation tonometry.
A dilated fundus examination with direct and indirect ophthalmoscopy was performed. An electroretinogram (ERG) was obtained by either of 2 procedures previously described.7,8 These recording techniques adhered to the international standard for clinical electrophysiologic measurements.9 All results were compared with the 90% tolerance limits or an appropriate range for a healthy population.7,8 A single-flash waveform was considered nondetectable if the amplitude was less than 10 µV, whereas a 30-Hz flickering stimulus was considered nondetectable if the response was less than 1 µV.
Visual field examination was performed monocularly with a Goldmann perimeter using II-2-e, II-4-e, III-4-e, and V-4-e test targets. The targets were moved from nonseeing to seeing regions. All of the above targets were not necessarily used on each patient.
The 66 patients and 96 controls were screened in the 50 exons of the ABCRgene with single-strand conformation polymorphism analysis followed by automated DNA sequencing of abnormal shifts. Primer sequences used for the assay design were the same as those of Allikmets et al.5 Exon numbers were changed to agree with Gerber et al10(also G. Travis, MD, et al, unpublished data, 1998).
From each patient, 12.5 ng of DNA was used as template in an 8.35-µL polymerase chain reaction containing 1.25 µL of 10X buffer (100 mmol of Tris hydrochloride, pH 8.3, 500 mmol of potassium chloride, and 15 mmol of magnesium chloride); 300 µm each of deoxycytidine triphosphate, deoxyadenosine triphosphate, deoxyguanosine triphosphate, and deoxythymidine triphosphate; 1 pmol of each primer; and 0.25 U of Taq polymerase (Boehringer Mannheim, Indianapolis, Ind). Samples were denatured for 5 minutes at 94°C and incubated for 35 cycles under the following conditions: 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds in a DNA thermocycler (Omnigene, Teddington, Middlesex, UK). After amplification, 5 µL of stop solution (95% formamide, 10 mmol of sodium hydroxide, 0.05% bromophenol blue, and 0.05% xylene cyanol) was added to each sample. Amplification products were denatured for 3 minutes at 94°C and electrophoresed on 6% polyacrylamide and 5% glycerol gels at 25 W for approximately 3 hours. After electrophoresis, the gels were stained with silver nitrate, as described by Bassam et al.11
Polymerase chain reaction products were sequenced using fluorescent dideoxynucleotides on an automated sequencer (model 373; Applied Biosystems, Foster City, Calif). All mutations were recognized by the approximately equal peak intensity of 2 fluorescent dyes at the mutant base. All sequencing was performed bidirectionally.
Sequence changes were considered to be "possibly disease causing" if they were expected to change the amino acid structure of the ABCR protein and if they were also present in less than 1% of the 192 control alleles.
Table 1shows the age, sex, vision at the most recent visit, fluorescein angiographic findings, and clinical phenotype for 29 patients with a possible disease-causing mutation in the ABCRgene. Each of the 29 patients were from different pedigrees. However, the mother of patient 12 and the father of patient 29 were also found to have compound heterozygous sequence variations in the ABCRgene. Table 2shows ERG amplitudes from the worse eye in patients with different clinical phenotypes. Three major clinical phenotypes were observed. Twelve patients (phenotype I) showed a generally small, atrophic-appearing foveal lesion surrounded by parafoveal or perifoveal yellowish white lesions (Figure 1). These lesions were essentially restricted to a region surrounding the fovea, although in patient 2 (Table 1), a few yellowish white deposits were seen anterior to the perifoveal region and nasal to the optic disc. In isolated patients, the foveolar region appeared grossly spared. In these patients, visual acuity tended to remain normal or only slightly reduced. In 10 of 12 patients who underwent fluorescein angiography, none showed a dark or silent choroid, which has been observed in patients with Stargardt dystrophy (Figure 2).12,13 Central scotomas were observed in all 12 patients with phenotype I.
Electroretinograms were obtained on 10 of these 12 patients; 9 of whom had normal cone and rod amplitudes (Figure 3). These patients would have been classified as having stage 1 disease as described by Fishman.14 Of interest, 9 of the 12 patients showed a Gly1961Glu sequence variation on 1 allele of the ABCRgene (Table 1), none of whom showed a dark choroid or abnormal ERG findings.
A second group of 10 patients (phenotype II) had numerous yellowish white fundus lesions that were not restricted to the parafoveal or perifoveal region but rather, occurred throughout the posterior pole, often extending anterior to the vascular arcades and nasal to the optic disc (Figure 4). Seven of 10 patients on whom a fluorescein angiogram was obtained showed a dark choroid (Figure 5). Eight patients showed a central scotoma, 1 patient showed a paracentral scotoma, and another patient did not undergo visual field testing. Eight of 10 patients underwent ERG testing. In 2 patients, the amplitudes were normal; in 6 patients, subnormal responses were observed (Figure 6). Cone amplitudes were reduced more extensively than were rod amplitudes (Table 2). None of these patients showed the Gly1961Glu sequence variation in the ABCRgene, and would have been classified as having stage 2 disease in the classification by Fishman.14
A third group of 7 patients (phenotype III) showed extensive atrophic-appearing changes of the retinal pigment epithelium throughout the posterior pole extending beyond the vascular arcades (Figure 7). Presumably, most fundus flecks had resorbed in these patients. Atrophic changes of the choriocapillaris were also observed either diffusely or regionally, with the greatest extent observed within the posterior pole between the vascular arcades (Figure 8). An ERG was obtained on 6 of these 7 patients and showed a notable reduction in cone and rod function (Figure 9). One 65-year-old patient had nondetectable responses. In a previous publication,14 other patients with this phenotype had moderately to markedly subnormal ERG cone and rod responses. Visual field testing done on 5 of the 7 patients showed central scotomas with (3 patients) or without (1 patient) peripheral restriction and only residual temporal islands (1 patient) (Table 1). Such patients were previously categorized as having stage 3 or stage 4 disease.14 One of the 7 patients in this subgroup showed a compound heterozygous sequence variation in which 1 allele had a Gly1961Glu change. A variation in the second allele involved codons 1620-1622 and replaced 6 base pairs (ATAACA) with 4 base pairs (TCCT), for a net loss of 2 base pairs. This resulted in a premature stop signal 24 codons later.
Our findings provide examples of the extent of retinal changes that can be observed in patients with possible disease-causing mutations in the ABCRgene who would be classified as having Stargardt dystrophy or fundus flavimaculatus. The spectrum of disease severity in our cohort varied from a localized atrophic-appearing lesion within the fovea surrounded by perifoveal yellowish white flecks to severe and diffuse atrophy of the retinal pigment epithelium and choriocapillaris. Several patients showed a phenotype consistent with the original description of Stargardt dystrophy,15 and in 1 patient, the presence of moderately diffuse yellowish white fundus flecks and the absence of an atrophic-appearing lesion within the fovea was similar to patients traditionally considered to have fundus flavimaculatus.16 Therefore, these 2 disorders likely represent allelic variations of a retinal disease that results from different mutations within the same ABCRgene. In this regard, it would not seem prudent to consider them as 2 distinctly different retinal diseases. However, the distinction might be useful for monitoring the natural history and visual prognosis for such patients.17
The relative number of patients observed in each of our 3 phenotype groups is likely to vary depending on the means of ascertainment. Because we intentionally selected patients with disparate severities of disease, our percentage of patients within each phenotype may not be representative of a randomly selected cohort of such patients. Furthermore, consideration should be given to the likely possibility that patients with phenotype II may have appeared similar to those with phenotype I at an earlier stage of disease regarding the number and distribution of fundus yellowish white lesions. However, it is our impression that those classified as phenotype II likely manifest a dark choroid even at earlier stages of disease, unlike those with phenotype I in whom a dark choroid was not observed in our group of such patients. With 1 exception, the Gly1961Glu sequence variation was observed in patients with a phenotype I expression. Whether this observation represents an inadvertent bias of ascertainment must await the assessment of a larger number of patients. We suspect that at least some patients with phenotype III possibly, or even likely, showed a fundus appearance similar to phenotype II during earlier stages of disease. Thus, a classification of patients into these 3 phenotypes is not intended to imply a comprehensive or rigid segregation of disease phenotypes but rather is suggested as a practical means of describing the extent of phenotypic variation that occurs in patients with various ABCRgene sequence changes.
Only 41 of 132 ABCRalleles of the 66 study patients exhibited a coding change believed to be compatible with involvement in the disease process. Moreover, 17 of 29 patients who exhibited a possible disease-causing mutation had such sequence changes in only 1 of the 2 ABCRalleles. Taken together, these observations suggest that our present mutation assay does not detect most disease-causing mutations in the ABCRgene, including promoter mutations, intronic mutations (inactivating an enhancer, activating a suppressor, or activating a cryptic splice site), and heterozygous deletions involving polymerase chain reaction primer binding sites. More comprehensive correlations of genotype with various phenotypes can be made when additional patients with sequence variations on both alleles of the ABCRgene are identified.
Whenever many true disease-causing mutations are undetectable by an assay, it is possible that an apparently disease-associated sequence variation is actually a non–disease-causing "marker polymorphism" in the same allele as the true mutation. For example, suppose that the Gly1961Glu change is in fact a non–disease-causing polymorphism present in a relatively small portion of the general population. If a true disease-causing mutation occurred by chance in the promoter region of such a "1961-marked" allele, then one would expect an enrichment of the Gly1961Glu sequence change among a group of patients with Stargardt dystrophy (compared with the healthy population) simply because this detectable polymorphism is physically linked to the undetectable disease-causing mutation.
However, even if such an event were to occur, clinical associations with the marker polymorphism (such as the association between the Gly1961Glu sequence change and phenotype I in this article) would be meaningful only as long as the frequency of the polymorphism in the general population was relatively low.
Accepted for publication December 1998.
This study was supported by grants from the Foundation Fighting Blindness, Hunt Valley, Md, and grant EY 10539 from the National Eye Institute, National Institutes of Health, Bethesda, Md.
We thank Kimberlie Vandenburgh, Tumara Clark, and Gretl Beck, MS, University of Iowa, Iowa City, for their valuable technical assistance.
Reprints: Gerald A. Fishman, MD, UIC Eye Center, University of Illinois at Chicago, M/C 648, 1855 W Taylor St, Chicago, IL 60612 (e-mail: firstname.lastname@example.org).