The most determinant symptom of Best VMD is the abnormal EOG with a reduced light-peak to dark-trough ratio combined with a normal ERG [19
]. Bestrophin 1, the 585-amino acid protein encoded by the BEST1
], is a member of the RFP-TM family of proteins, so named for their highly conserved arginine, phenylalanine, proline motif [3
], which appears to exhibit properties of Ca2+
]. Bestrophin 1 does not appear to be the channel itself but to act as a modulating subunit; thus channel function would directly correlate to the involved mutation. The apparent role of bestrophin 1 in the regulation of ion transport obviously affects the light peak on EOG; it is unlikely that the light peak defect itself is the cause of vision loss in Best VMD. Any connection between the light-peak deficit in Best VMD and lipofuscin accumulation in the RPE (the most common histopathologic finding in Best VMD) is speculative. However, given that ion transport is a requirement for acidification of phagolysosomal compartments and Ca2+
is a critical regulator of vesicle fusion, either of the proposed functions of bestrophin 1, if impaired, could lead to the accumulation of lipofuscin (and ultimately cause vision loss from lipofuscin toxicity to photoreceptors). To date more than 108 different BEST1 mutations have been reported (see the Human Gene Mutation Database
Here, we report three novel missense changes absent from 192 control chromosomes. All three affected residues are conserved through evolution and were predicted by a structure homology-based method to have an impact on the protein (). Two out of the three mutations occur in exon 2 (p.T4A and p.G15D), which is located in the NH2
cytoplasmic domain of the protein. Interestingly, it has been demonstrated that the p.T6P and p.A10V mutations that affect this domain produce currents with an amplitude >20% that of wild-type bestrophin [25
]. The novel mutations p.T4A and p.G15D are predicted to change the polarity and/or the charge of the NH2
terminus of the protein and therefore may be regarded as disease causing. The third novel mutation, p.I230T, may alter the structure of the protein as it changes a hydrophobic residue located in the transmembrane domain of the protein into a polar residue.
Of note, the p.V9A change had previously been classified as a change of uncertain pathogenicity seemingly conservative by Petrukhin et al. [3
]. In our study this change was regarded as a disease-causing mutation by virtue of its absence from 192 control chromosomes and its Polyphen PSIC score (1.949), which suggests that it may have a functional impact.
Heterozygous mutations in BEST1
, which usually cause typical Best VMD, may also cause adult vitelliform macular degeneration [26
], autosomal dominant bestrophinopathy, and a rare and unique condition called autosomal dominant vitreoretinochoroidopathy [28
]. Burgess et al. [29
] recently reported on compound heterozygous or homozygous mutations in the BEST1
gene as the causative mutations for a distinctive retinopathy, which they named autosomal-recessive bestrophinopathy (ARB). Given that the different diseases caused by BEST1
gene mutations may share common clinical findings, a complete clinical examination of Best VMD patients combined with molecular genetics studies of the BEST1
gene is mandatory for adequate counseling of the families. Interestingly, while in nine of the ten pedigrees reported here the disease segregated as an autosomal dominant trait, in one family the affected patient (FG08) was apparently homozygous for a BEST1
mutation [p.R92C]. The common origin of the patient’s parents’ homozygosity for the mutation is the likely reason, although hemizygosity at the BEST1
locus cannot be excluded. In any case, mutational biallelism raises the question as to whether, instead of Best VMD, the patient may be affected with the autosomal-recessive bestrophinopathy described by Burgess et al. as a null phenotype of bestrophin-1 in humans [29
]. Clinical examination showed that the patient had no ARB-associated, scattered, punctate flecks and retinal edema but presented with bilateral multifocal lesions consistent with the diagnosis of multifocal Best VMD. This phenotype may be considered more severe than that of another patient heterozygous for the p.R92C mutation who is affected with bilateral focal lesions complicated by CNV. This observation differs from that of Bakall et al. who reported on the histopathology of a donor eye from an individual homozygous for the BEST1
p.W93C mutation and concluded that the clinical and pathological effects of homozygosity for the p.W93C mutation are not more severe than those reported for heterozygotes [30
In our series we report a large interfamilial and intrafamilial clinical variability in terms of age of onset, disease progression, stage of the lesions, and visual function. We found no association between BEST1 mutations and expressivity, with respect to age, BCVA, and stage of the disease as evaluated by FAF, FA, and OCT. Mean BCVA impairment showed a statistically significant correlation to a more advanced stage of the disease. This association was independent of the patients’ age. These data suggest that a functional impairment in Best VMD may be related to the progression of the disease rather than to a patient’s age. However, in the current series there was only one family to illustrate the phenotype of each mutation except for one mutation. This probably represents a major limitation of any statistical analysis in proposing a severity scaling.
Interestingly, the p.A243V mutation was found to be associated with late onset in one family of our Best VMD series. This finding is consistent with a previous report of a mild and relatively invariable Best VMD phenotype associated with this mutation [31
]. Even though our study was not designed to investigate disease progression, the absence of phenotype in two siblings of the same family harboring the mutation may be explained by their young age (13 and 17 years). However, it is possible that these two individuals may remain unaffected (normal fundus findings, OCT, and EOG) through their life span as well. Incomplete penetrance is indeed a well known feature in BEST1 disease. Functional and clinical data in our series may support this notion. The heterozygous p.R92G an p.G15D mutations resulted in the earliest disease manifestation (at 2 years of age); however, the same mutation was also responsible for either a later onset (at the age of 11 years for FG06 [p.R92G]) or even no disease manifestation (at the age of 30 years for FG10 [p.G15D]) within the same families (FAMILY FG II and FAMILY FG IV, respectively).
All patients except two had bilateral macular lesions. Two patients presented with unilateral disease, but this could not be related either to their age or to their genotype. Indeed, one of them, aged 27 (CT07), shared the p.T4A mutation with his 23-year-old sibling presenting with bilateral lesions (CT08). Similarly, the second patient, a 70-year-old man (CT11) carried the p.R25W mutation responsible for bilateral lesions in two of his young relatives, aged 10 and 36. Similarly, CNV did not appear to correlate with the mutation, as suggested by the intrafamilial variability of this trait.
Bilateral multifocal Best VMD lesions were diagnosed at the age of 41 in a patient heterozygous for another BEST1 mutation, p.I230T. Two younger relatives (aged 9 and 11) presented with an early-stage lesion; progression is uncertain.
The wide variability of clinical expression of BEST1
mutations within and between families is consistent with previous reports [27
]. Owing to this wide variability of clinical expression, it is difficult to compare our findings with other previously published series. Moreover, we adopted a widely accepted clinical classification, and, based on fundus biomicroscopy, all eyes were graded as showing only one of the progressive stages of Best VMD; thus, for example, in contrast with Boon et al. [30
], we did not consider characteristics attributable to different stages. However, the clinical features reported here for each progressive stage were typical and actually consistent with other Best VMD series.
One limitation of the current study was the absence of real co-segregation analysis for the families with novel reported changes. Another limitation was that we did not perform, systematically, ERG in our patients and thus we were not able to distinguish whether an abnormal light rise on EOG would have been due to either photoreceptor or RPE dysfunction.
Overall in our series, particularities were found in two affected patients showing unilateral Best VMD, in two affected patients showing, in both eyes, multifocal Best VMD, and in four affected patients (six eyes) who were treated for CNV. All these are well known possible features of Best VMD. Three out of 23 patients (13%) with the BEST1 mutation showed normal fundus, OCT, and EOG findings.
In conclusion, variability of clinical expression of BEST1 mutations suggests that cis or trans-acting genetic modifiers may modulate the functional and clinical data.