The three patients in this study showed an intrafamilial phenotypic variability that led to different clinical diagnoses. Intrafamilial variability has been reported before in a family with four patients carrying homozygous c.2991+1655A>G mutations.
4 Since the amount of wild-type
CEP290 product, quantified in lymphoblast RNA, was comparable in all patients of that family (data not shown), another variant in other gene(s) may influence the phenotype and the intrafamilial variability. In the family described in this study, the intrafamilial variability between the proband and her two cousins may be explained by the heterozygous p.Leu394SerfsX3 mutation in the
MERTK gene, which was present in the LCA patient (VI-6) and absent in the two EOSRD patients (VI-3 and VI-4).
MERTK is associated with autosomal recessive RP
30–32 and is strongly expressed in the RPE and macrophages, where it plays a role in the RPE phagocytosis process.
33 CEP290 is localized in the connecting cilia of photoreceptors and is suggested to play a role in microtubule-associated protein transport in the cilia.
2,3,34 In view of the difference in function and localization, a physical interaction between these proteins seems unlikely. Moreover, one
MERTK p.Leu394SerfsX3 mutation and one
CEP290 c.2991+1655A>G mutation together do not lead to symptoms of retinal dystrophy (in individual V-4). In addition, human and murine carriers of heterozygous pathogenic
MERTK mutations do not show signs of retinal dystrophy as well. However, cultured RPE cells from the RCS rats that carry a homozygous truncating deletion in
MERTK in vitro phagocytose only a few photoreceptor outer segments, in contrast to wild-type cells, in which many phagosomes appear.
35 Therefore, RPE cells with only 50% of the amount of functional MER tyrosine kinase receptor may have a lower turnover of photoreceptor outer segment phagocytosis, which does not have a measurable pathogenic effect in otherwise healthy individuals, but could have a cumulative (or modifying) effect in patients in whom the production of photoreceptor outer segments is disturbed, in this case by two pathogenic mutations in
CEP290. The phenomenon of modifier alleles has been described—for example, tri-allelic mutations resulting in a more severe phenotype in Bardet-Biedl patients and a modifier
RPGRIP1L allele that is suggested to result in a more severe retinal phenotype in ciliopathies.
36,37 Nevertheless, it remains speculative whether the
MERTK mutation functions as a modifier allele. The a priori chance for an individual to carry a pathogenic mutation in a retinal dystrophy gene is ∼10%, assuming that there are 67 autosomal recessive retinal dystrophy genes that all account for a similar proportion of cases, as calculated by Rivolta et al.
38 in 2002. Currently, 44 nonsyndromic autosomal recessive retinal dystrophy genes have been identified (
http://www.sph.uth.tmc.edu/retnet/ University of Texas Houston Health Science Center, Houston, TX) that account for ∼50% of cases, indicating that the calculation of Rivolta et al. could be realistic. Therefore, we estimate that 10% of retinal dystrophy patients are expected to carry a heterozygous mutation in a second gene. Eventually, the identification of combinations of mutations in more than one retinal dystrophy gene by next-generation sequencing, together with a precise description of phenotypes, may uncover the unknown influences of additional heterozygous mutations.