Allelic Variation in the VMD2 Gene in Best Disease and Age-Related Macular Degeneration

Abstract

purpose. To assess the allelic variation of the VMD2 gene in patients with Best disease and age-related macular degeneration (AMD).

methods. Three hundred twenty-one AMD patients, 192 ethnically similar control subjects, 39 unrelated probands with familial Best disease, and 57 unrelated probands with the ophthalmoscopic findings of Best disease but no family history were screened for sequence variations in the VMD2 gene by single-strand conformation polymorphism (SSCP) analysis. Amplimers showing a bandshift were reamplified and sequenced bidirectionally. In addition, the coding regions of the VMD2 gene were completely sequenced in six probands with familial Best disease who showed no SSCP shift.

results. Forty different probable or possible disease-causing mutations were found in one or more Best disease or AMD patients. Twenty-nine of these variations are novel. Of the 39 probands with familial Best disease, mutations were detected in all 39 (33 by SSCP and 6 by DNA sequencing). SSCP screening of the 57 probands with a clinical diagnosis of Best disease but no family history revealed 16 with mutations. Mutations were found in 5 of 321 AMD patients (1.5%), a fraction that was not significantly greater than in control individuals (0/192, 0%).

conclusions. Patients with the clinical diagnosis of Best disease are significantly more likely to have a mutation in the VMD2 gene if they also have a positive family history. These findings suggest that a small fraction of patients with the clinical diagnosis of AMD may actually have a late-onset variant of Best disease, whereas at the same time, a considerable fraction of isolated patients with the ophthalmoscopic features of Best disease are probably affected with some other macular disease.

Vitelliform macular dystrophy (Best disease, VMD2, Online Mendelian Inheritance in Man [OMIM] 153700) is an autosomal dominant juvenile-onset macular degeneration characterized by an egg-yolk–like appearance of the macula. ¹ In 1992, the Best disease gene was localized to 11q13. ^{2 3} Subsequent studies refined the chromosomal location of the gene to a 980-kb interval between D11S4076 and uteroglobin (UGB) ⁴ ; and later, the VMD2 gene was identified by positional cloning . ^{5 6} The gene encodes a 585-amino-acid protein known as bestrophin, which is selectively expressed in the retinal pigment epithelium of the eye. Bestrophin has sequence homology with the RFP protein family. These proteins have a conserved sequence of 350 to 400 amino acids and are found in organisms as diverse as the nematode, the fruit fly, and the mouse. The function of this family of proteins is currently unknown. ^{5 6}  

Best disease shares some clinical and histologic features with AMD. Both diseases are characterized by an accumulation of lipofuscin-like material within and beneath the retinal pigment epithelium. ⁷ Geographic atrophy of the retinal pigment epithelium and choroidal neovascularization can occur in both disorders. One feature that distinguishes Best disease from AMD is an abnormality of the electro-oculogram (EOG). Normally, the electrical potential across the retinal pigment epithelium increases almost twofold when an eye moves from darkness to bright light, but in patients with Best disease, this potential changes very little, if at all, under these conditions. That this EOG abnormality is present in affected individuals with ophthalmoscopically normal maculae suggests that it is a primary rather than a secondary phenomenon. However, the mechanism by which bestrophin mutations give rise to an abnormal EOG is not known at this time. 

Age-related macular degeneration is likely to be a genetically heterogeneous disease, and, given the broad phenotypic similarity to Best disease, it is plausible that a fraction of AMD cases could be caused by mutations in the VMD2 gene. ^{2 4 8} This study was undertaken to try to identify disease-causing mutations in the VMD2 gene in individuals affected with either Best disease or AMD. 

Materials and Methods

This study adhered to the tenets of the Declaration of Helsinki and was approved by the Human Subjects Review Committee at the University of Iowa. Informed consent was obtained from all study participants. Thirty-one of the 96 Best probands were ascertained in Iowa, 57 in other centers in the United States, and 8 in Switzerland. The clinical diagnosis of Best disease was based on a subfoveal vitelliform lesion in at least one eye and the relative absence of drusen. Two hundred fifty-seven AMD patients were ascertained in the Retina Clinic of the University of Iowa, and 64 patients were ascertained in the Retina Clinic of the Hôpital Jules Gonin, Lausanne, Switzerland. The diagnosis of AMD was based on the presence of drusen and/or atrophy of the retinal pigment epithelium and the absence of ophthalmoscopic stigmata of other macular diseases such as the presumed ocular histoplasmosis syndrome. Ninety-six normal volunteers more than 40 years of age, and with no family or personal history of AMD were ascertained in Iowa, and 96 normal volunteers with no personal history of eye disease were ascertained in Lausanne. DNA from all study participants was extracted from venous blood using a previously described protocol. ⁹ Ten primer pairs, as previously described, ⁵ were used to amplify the entire coding sequence and exon–intron boundaries of the VMD2 gene in all 609 individuals. For improved sensitivity, exon 4 was amplified in two parts. The primers for exon 4A were 5′-AAAGCTGGAGGAGCCGAG-3′ and 5′-CGTCCAAAGCTGCTCCGTA-3′ and for exon 4B were 5′-GCTGGTGGAACCAGTACGAG-3′ and 5′-CTCCACCCATCTTCCGTTC-3′. 

All amplimers were analyzed in a standard fashion using single-strand conformation polymorphism (SSCP) analysis in a single laboratory. All SSCP gels were independently scored by a minimum of two experienced investigators. Amplimers showing a bandshift were reamplified and sequenced bidirectionally using an automated sequencer and dye-terminator chemistry (model 377; Perkin–Elmer, Applied Biosystems, Foster City, CA). Variants were confirmed by restriction endonuclease digestion if a restriction site polymorphism was introduced by the sequence change. The entire VMD2 coding sequence was bidirectionally sequenced in six Best probands with familial disease who failed to exhibit an SSCP shift. This resulted in the discovery of five sequence variations. The amplimers containing three of these variations (Glu300Asp, Glu300Lys, and Asp302Gly) were also sequenced in 43 control individuals. A fourth variation (Lys30Arg) is detectable by restriction digestion, and this assay was also performed in 43 control individuals. The fifth change (Tyr227Cys) was previously found in one Best disease kindred but not in 42 control individuals. ⁶ Statistical analyses were performed with a two-tailed Fisher’s exact test. The study design enabled us to detect a 5% difference between the AMD and control groups with 80% power at the 0.05 significance level. 

Results

Four hundred eighty-two instances of 53 different sequence variations were detected in the VMD2 gene. Thirty-six of these were thought to be probable disease-causing variations because they altered the predicted amino acid sequence of the VMD2 gene product, were absent from the 192 normal volunteers, and were present in individuals with a phenotype (Best disease) previously associated with the VMD2 gene (Table 1) . Four were thought to be possible disease-causing variations, because they altered the predicted amino acid sequence of the VMD2 gene product, were absent from the 192 normal volunteers, and were present in individuals with a phenotype (AMD) that has not been independently associated with the VMD2 gene (Table 1) . Thirty-one of the 40 probable or possible disease-causing mutations had one or more additional features compatible with pathogenicity, including alteration of the size, polarity, or charge of the predicted protein product of the mutant allele; correct segregation of the mutation in seven or more affected individuals; presence in multiple Best pedigrees; and/or alteration of one or more amino acid residues that have been highly conserved in the RFP gene family (Table 1) . 

Thirty-nine of the Best probands had a positive family history of the disease. SSCP analysis detected VMD2 sequence variations in 33 of these individuals. Bidirectional DNA sequencing of the entire VMD2 coding sequence was performed with samples obtained from the remaining six, and a sequence variation was detected in each. Fifty-seven individuals had a subfoveal vitelliform lesion in at least one eye suggestive of Best disease but did not have a positive family history of the disease. SSCP analysis detected a VMD2 sequence variation in only 16 of these patients, a significantly lower (P < 0.0001) fraction than in the group with a positive family history. 

Some clinical information about two of the Best disease families has been published previously. Affected individuals from the family that was the subject of our original linkage study, ² including two individuals who were reported because of dramatic fluctuations in visual acuity, ¹⁰ were found to harbor an Ala243Thr mutation. Another Best patient had been previously reported because of the coexistence of an inflammatory condition known as the multiple evanescent white dot syndrome. ¹¹ This individual (marked with a dot in Fig. 1A ) was found in the present study to have a Tyr227Asn change. Three other members of his family (enclosed by a box in Fig. 1A ) are noteworthy, because they illustrate the incomplete penetrance of the vitelliform phenotype that is known to occur in some individuals that carry a VMD2 mutation. The eldest of the three was first examined at age 66 because her position in the family made her an obligate carrier of a Best disease mutation. She had no visual symptoms, and her visual acuity was 20/15 OD and 20/20 OS. Fundus examination showed healthy discs and vessels and a completely normal macula (Fig. 1B) . EOG testing showed an abnormal Arden ratio. Her 44-year-old daughter was also first examined because of her obligate carrier status. She had no visual symptoms, and her visual acuity was 20/15 OU. Fundus examination showed an entirely normal macula in both eyes (Fig. 1C) . The 9-year-old grandson of the first patient was first examined because of a family history of Best disease in a great uncle, a great aunt, and several cousins (Fig. 1A) . He had no visual symptoms, and his visual acuity was 20/15 OU. However, fundus examination revealed yellow, subfoveal vitelliform lesions in both eyes (Fig. 1D) . 

Among the 321 patients with AMD, SSCP analysis detected VMD2 coding sequence variations in 5. The clinical characteristics of these individuals are summarized in Table 2 . There was a family history of AMD in two of the AMD patients with coding variations in the VMD2 gene (Table 2) . Three of these five patients had evidence of choroidal neovascularization, and two did not. 

Thirteen changes that did not alter the predicted amino acid sequence of the bestrophin protein were also observed during the study (Table 3) . Twenty of the 53 total sequence variations that were observed have been previously reported ⁵ or submitted to an online database (http://www.uni-wuerzburg.de/humangenetics/vmd2.html). ^{6 12 13}  

Discussion

In this study, 39 unrelated patients were available with a clinical diagnosis of Best disease and a positive family history. VMD2 sequence variations were identified in 33 of these patients (85%) using a screening strategy that consisted of SSCP followed by sequence characterization of observed shifts. By relaxing the diagnostic criteria to include patients with the clinical appearance of Best disease but without a family history, 57 additional patients became available for study. Although 16 sequence variations were detected with SSCP in these 57 individuals, the yield was three times lower (28%). We were able to increase the mutation detection rate to 100% in the familial group by performing automated DNA sequencing of the entire coding region in the patients in whom SSCP showed no bandshift during the initial screening. The limitation to using sequencing alone as a screening assay is that it is 10 times more expensive than SSCP followed by sequencing of observed shifts. The significantly lower frequency of VMD2 mutations in patients with Best-like lesions who have no family history of the disease suggests that a substantial fraction of these patients are in fact phenocopies—that is, they are affected with a molecularly different disease that mimics Best disease ophthalmoscopically. The identification of VMD2 mutations among the group of patients with the clinical diagnosis of AMD suggests that phenocopies also occur in the reverse direction. 

Among the Best disease–associated mutations, we observed one (Tyr227Asn) that was associated with such marked phenotypic variability that molecularly affected individuals in two different families were found to have perfectly normal fundus examinations at ages greater than 50 years. This mutation would be a good choice for use in transgenic animal experiments designed to look for mitigator genes in different murine backgrounds. The possible disease-causing mutations present in the AMD population would also be good candidates for study at the animal model level, because the prevalence of AMD in the general population would make their pathogenicity difficult to demonstrate with statistical arguments alone. The EOG abnormalities associated with Best disease have been very tightly linked to the lipofuscin-accumulating portion of the phenotype since the EOG abnormalities were first recognized in this disease. ^{14 15} However, the recognition of a patient with a normal EOG and a lipofuscin-accumulating phenotype (AMD) suggests that the mechanisms that underlie the lipofuscin accumulation and the EOG abnormality could be distinct. Obviously, an alternative explanation would be that the sequence variation observed in this patient may be a rare non–disease-causing polymorphism that simply meets our empiric criteria for a possible disease-causing mutation. 

Apolipoprotein E, a protein involved in fatty acid metabolism, is deposited in soft drusen. ¹⁶ Similarly, Petrukhin et al. ⁵ have suggested that a defect in fatty acid metabolism results in the lipofuscin deposits seen in Best disease. However RDS and ABCR mutations also result in lipofuscin accumulation within the RPE (causing pattern dystrophy and Stargardt disease, respectively), and neither of these genes is thought to be directly involved in fatty acid metabolism. These diseases demonstrate that genes with quite different functions (a structural protein of the photoreceptor outer segment and an adenosine triphosphate–binding cassette transporter) can result in RPE dysfunction and lipofuscin accumulation. Furthermore, at the present time, no explanation has been advanced for the EOG abnormality seen in most patients with Best disease. 

Homology with other known proteins can sometimes suggest functional domains or residues the mutation of which might reasonably be expected to be more pathogenic than others. The VMD2 gene is an example that this is sometimes easier said than done. The literature regarding the RFP gene family is scant, ^{5 6} (see National Center for Biotechnology Information web site http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?uid=1945425&form=6&db=n&Dopt=g) and the regions thought to be shared by members of the family are open to interpretation. For consistency, we used the homologies suggested by Marquardt et al. ⁶ for the entries in Table 1 . 

The low frequency of mutations in the AMD patient population (1.5%) was not significantly different from that present in the control population (0/192, 0%). Two unrelated AMD patients were found to harbor an identical nonsense mutation in codon 149. EOGs were not obtained from these patients, and thus they may represent misdiagnosed Best patients. However, nonsense mutations appear to be very rare in Best disease (http://www.uni-wuerzburg.de/humangenetics/vmd2.html).Thus, another possibility is that truncating mutations produce haploinsufficiency and a late-onset AMD phenotype. Two other AMD patients with missense changes underwent EOG testing. In one patient, who harbored an Arg105Cys sequence change, the Arden ratio was found to be normal in both eyes (2.16 OD, 1.85 OS) suggesting that this patient did not have Best disease. In another, who harbored a Glu119Gln variation and whose right fundus is depicted in Figure 2 , the Arden ratio was abnormal (1.0 OU). This patient may have Best disease resembling AMD. A further possibility is that certain bestrophin mutations produce both an AMD phenotype and EOG abnormalities. Of interest, the Glu119Gln variation has been reported to produce another non–Best disease phenotype (bull’s eye maculopathy) in a 57-year-old white female. ⁸ The latter study similarly suggested that some other rare VMD2 variants may predispose to a small fraction of AMD. ⁸  

In summary, we have identified a number of novel VMD2 sequence variations in patients with macular disease. The majority of these were associated with the phenotype of Best disease, whereas a few were observed in patients with AMD. Although overall the sequence variations in AMD patients were not statistically significant, we cannot exclude that some rare VMD2 variants may result in AMD. Patients with the clinical diagnosis of Best disease are significantly more likely to have a mutation in the VMD2 gene if they also have a positive family history. 

 Supported by National Institutes of Health Grant EY-10539, the Carver Charitable Trust, The Ruth and Milton Steinbach Foundation, The Grousbeck Family Foundation, The Foundation Fighting Blindness, Swiss National Science Foundation Grant 32-053750.98, the Fondation Telethon Action Suisse, The Kyle Curran Memorial Fund, and an unrestricted grant from Research to Prevent Blindness.

 Submitted for publication June 28, 1999; revised September 29 and November 11, 1999; accepted December 2, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Probable and Possible Disease-Causing Mutations in the VMD2 Gene

Exon	Sequence Change	Nucleotide Change	Maximal LOD Score*	Effect	Familial Probands (n = 39)	Unrelated Best Probands (n = 57)	AMD Probands (n = 321)	Normals (n = 192)	Conserved in RFP Proteins, †
2	Thr6Arg	ACA-AGA	0.90	Charge	1	0	0	0	–
2	Phe17Cys	TTC-TGC	0.80	Polarity	—	1	0	0	–
2	Trp24Cys, ‡	TGG-TGT	—	Polarity	1	0	0	0	+
2	Arg25Trp	CGG-TGG	—	Charge	—	1	0	0	+
2	Lys30Arg, §	AAG-AGG	0.20	—	3	0	0	0	–
3	Phe80Leu	TTC-TTA	—	—	—	1	0	0	–
4	Thr91Ile	ACC-ATC	—	Polarity	1	0	0	0	–
4	Pro101Thr	CCG-ACG	—	Polarity	1	0	0	0	–
4	Arg105Cys	CGC-TGC	—	Charge	—	0	1	0	–
4	Glu119Gln, ‡	GAG-CAG	—	Charge	—	0	1	0	–
4	Asn133Lys	AAC-AAG	—	Charge	—	1	0	0	–
4	Gly135Ser, ‡	GGC-AGC	—	—	—	1	0	0	–
4	Leu140Arg	CTG-CGG	—	Charge	—	1	0	0	–
4	Arg141His	CGC-CAC	—	—	—	1	0	0	+
4	Lys149Stop	AAG-TAG	—	Termination	—	0	2	0	–
5	Ala195Val	GCG-GTG	0.17	—	1	1	0	0	–
5	Ile201Thr	ATC-ACC	—	Polarity	1	0	0	0	+
5	Leu207Ile	GTC-ATC	—	—	1	2	0	0	–
5	Ile211Thr	ATC-ACC	—	Polarity	—	1	0	0	–
6	Arg218Cys, ‡	CGT-TGT	0.30	Charge	2	2	0	0	+
6	Arg218His	CGT-CAT	—	—	1	1	0	0	+
6	Gly222Val	GGA-GTA	0.07	—	1	0	0	0	–
6	Leu224Pro	CTG-CCG	—	—	1	0	0	0	–
6	Tyr227Asn, ‡	TAC-AAC	4.73	—	2	0	0	0	+
6	Tyr227Cys, ‡ , §	TAC-TGC	—	—	1	0	0	0	+
7	Ala243Thr	GCG-ACG	11.17	Polarity	2	0	0	0	–
7	Val275Ile	GTC-ATC	—	—	—	0	1	0	–
7	Phe276Leu	TTC-TTG	0.94	—	—	1	0	0	–
7	Del281Phe	Del TTC	—	Deletion	—	1	0	0	+
8	Asn296His	AAC-CAC	—	Charge	1	0	0	0	–
8	Pro297Ala, ‡	CCC-GCC	1.60	—	1	0	0	0	+
8	Glu300Asp, ‡ , §	GAG-GAC	1	—	6	0	0	0	–
8	Glu300Lys, §	GAG-AAG	0.30	Charge	2	0	0	0	–
8	Del Asp301	Del GAT	—	Deletion	2	0	0	0	+
8	Asp302Gly, §	GAT-GGT	—	Charge	1	0	0	0	–
8	Asp302Val	GAT-GTT	—	Charge	1	0	0	0	–
8	Glu306Gly	GAG-GGG	—	Charge	1	0	0	0	+
8	Glu306Asp	GAG-GAC	—	—	1	0	0	0	+
8	Thr307Ala	ACC-GCC	—	Polarity	1	0	0	0	–
8	Thr307Ile, ‡	ACC-ATC	1.94	Polarity	2	0	0	0	–

*  The frequency of the mutant allele was assumed to be 1% for this calculation.

†  The homologies suggested by Marquardt et al. ⁶ were used for the entries in this column (see text).

‡  Previously reported.

§  Detected by DNA sequencing (see text).

Figure 1.

 

(A) Pedigree of a family affected with Best disease caused by a Tyr227Asn change in the VMD2 gene. This sequence variation was confirmed in all affected individuals by SSCP. The individual marked with a dot was found to have the multiple evanescent white dot syndrome as well as Best disease. ¹¹ Fundus photographs of the individuals whose symbols are outlined by the box are shown in (B) through (D). (B) Fundus of the asymptomatic 66-year-old grandmother of the individual depicted in (D). (C) Fundus of the asymptomatic 44-year-old mother of the individual depicted in (D). (D) Fundus of a 9-year-old boy with classic vitelliform lesions characteristic of Best disease.

Table 2.

 

Clinical Features of AMD Patients with VMD2 Sequence Variations

VMD2 Variation	Clinical Details	Visual Acuity	Age
* >Arg105Cys	Cuticular drusen both eyes (EOG 2.16 OD, 1.85 OS)	20/30 OD, 20/20 OS	69
Lys149stop*	Drusen and choroidal neovascular membranes both eyes	20/400 OU	83
Lys149stop	Geographic atrophy both eyes	20/400 OD, 20/70 OS	86
Glu119Gln*	Drusen right eye, drusen and an occult choroidal neovascular membrane left eye (EOG 1.0 OU)	20/25 OU	78
Val275Ile	Choroidal neovascular membrane right eye, disciform scar right eye	20/200 OD, 20/100 OS	87

*  Positive family history of AMD.

Table 3.

 

Polymorphisms in the VMD2 Gene

Sequence Change	Best Probands (n = 96)	AMD Probands (n = 321)	Normal Subjects (n = 192)	Exon
Leu37Leu*	6, †	NA	7, †	2
Thr55Thr	0	1	0	3
Ile73Ile*	17	40	24	3
IVS4-24C-T*	22	128	70	Intron 4
Leu206Leu	2	0	0	5
IVS5-6C-T	0	1	0	Intron 5
IVS6 del (TCC)3*	15	50	29	Intron 6
Ile232Ile*	0	0	1	6
IVS8 del T	0	0	2	Intron 8
Pro341Pro*	1	5	2	9
Thr470Thr*	NA	NA	NA	10
Ser519Ser*	NA	NA	NA	10
Thr536Thr*	NA	NA	NA	10

*  Previously reported.

†  Detected by DNA sequencing (see text).

Figure 2.

 

Photograph of the right fundus of a 78-year-old patient with the clinical diagnosis of AMD and a Glu119Gln variation in the VMD2 gene.

The authors thank Luan Streb, Kim Vandenburgh, Robin Hockey, Heidi Haines, Gretel Beck, and Chris Taylor for their excellent technical assistance.