Comprehensive Registration of DNA Sequence Variants Associated with Inherited Retinal Diseases in Leiden Open Variation Databases

Inherited retinal diseases (IRDs) display an enormous degree of allelic and genetic heterogeneity. For example, 60 genes are currently implicated in retinitis pigmentosa (RP) (https://sph.uth.edu/RetNet/), the most prevalent of these diseases, with an incidence of 1/3,700 in the US [Boughman et al., 1980]. Moreover, more than 10,000 variants have been identified in 202 IRD-associated genes that explain 30%–90% of the cases, depending on the genetic and clinical subtype. Low- and medium-throughput mutation analysis techniques such as Sanger sequencing and allele-specific primer extension analysis (http://www.asperbio.com/asperophthalmics/), respectively, are widely available for many IRDs. High-throughput sequence analysis is becoming available for routine diagnostics [Audo et al., 2012; Booij et al., 2011; Coppieters et al., 2011; Glöckle et al., 2013; Neveling et al., 2012; Shanks et al., 2012], which will lead to a vast increase in the number of IRD-associated sequence variants.

We believe that a comprehensive IRD mutation registration has become increasingly important for several reasons. First, approximately one-third of DNA sequence variants are predicted to result in the absence of protein function, whereas the functional implications of the remaining variants are more difficult to assess. The proper registration of these variants will enable statistical analysis of their frequencies in patients versus “normal” persons. Second, the role of modifier variants in IRDs is still poorly understood, and a proper registration of all putative causal variants will facilitate these kinds of studies. Very few examples have thus far been described in which variants in more than one gene explain the IRD phenotype. Digenic inheritance has been found for variants in PRPH2 and ROM1 that cause RP [Kajiwara et al., 1994], and there is evidence for digenic inheritance of PDZD7 and USH2A variants in Usher syndrome [Ebermann et al., 2010], and FOXC1 and PITX2 variants in Axenfeld-Rieger disease [Kelberman et al., 2011]. Nonpenetrance of IRD-associated variants is a poorly understood phenomenon. The penetrance of protein-truncating PRPF31 variants in families with autosomal dominant RP is determined by the mRNA expression level of the “normal” PRPF31 allele [Vithana et al., 2001], that is controlled by several trans-acting factors [Venturini et al., 2012]. Modifiers have also been proposed to explain the variable phenotypic expression of a homozygous BBS1 missense variant, p.Met390Arg, that ranges from no IRD phenotype, nonsyndromic RP, to full-blown Bardet–Biedl syndrome (BBS) [Badano et al., 2006; Beales et al., 2003; Estrada-Cuzcano et al., 2012]. Furthermore, modifier alleles in RPGRIP1L and TTC21B have been identified in persons with ciliopathies [Davis et al., 2011; Khanna et al., 2009]. It has been estimated that up to 50% of individuals in the general population carries a causal autosomal recessive RD variant [Nishiguchi and Rivolta, 2012], the relevance of which as modifiers on retinal phenotypes caused by variants in other genes remains to be seen. The use of high-throughput sequencing techniques will increasingly yield variant information from multiple genes. Third, phase 1/2 clinical trials based on gene augmentation are underway for choroideremia (CHM), Leber congenital amaurosis (MERTK, RPE65), Stargardt disease (ABCA4), and Usher syndrome type 1 (MYO7A) (see: http://www.clinicaltrials.gov/ct2/home), and many preclinical studies have been performed [den Hollander et al., 2010]. As it can be foreseen that trials will be designed for a growing number of rarely mutated IRD genes, it will be very important to have complete knowledge of all genetically diagnosed persons. Fourth, comprehensive IRD mutation databases would also connect families with rare genetic defects through contact information of genetic centers, which will enable them to share experiences and difficulties in disease management, and foster the development of new treatments. Examples of this kind of patient empowerment are the RDH12-associated foundations “RDH12 Fund For Sight” (http://www.rdh12.org) and “Candle in the Dark” (http://www.candleinthedark.eu/), as well as the CRB1- and LCA5-associated foundations (http://www.crb1.org/; http://www.tfrr.org/). Fifth, the far majority of IRD-associated variants identified in diagnostic facilities has not been published or submitted to public mutation repositories, which is an enormous waste of knowledge. Finally, immediate reporting of all unclear cases, the variants of unknown significance, as well as all variants ruled out to be pathogenic, would greatly help the community.

The sequence variants reported in peer-reviewed journals suffer from poor annotation and non standard description. Guidelines for the proper nomenclature of sequence variants are available [http://www.hgvs.org/mutnomen/recs.html; den Dunnen and Antonarakis, 2000], and their proper use can be tested using computational tools [https://mutalyzer.nl/; Wildeman et al., 2008]. These guidelines however are not rigorously used nor required by most journals. The time-consuming process of data retrieval and assessment has hindered a systematic collection of sequence variants as funding for this important activity is scarce. Large-scale registration of sequence variants associated with IRDs thus far has been performed by retina international (http://www.retina-international.org/sci-news/databases/mutation-database/) and the Human Gene Mutation Database (HGMD). The retina international database has been updated inconsistently, and contains sequence variants of 78 IRD genes. HGMD consists of an open access version (http://www.hgmd.org) and a “professional” version (http://www.biobase-international.com/product/hgmd), requiring a subscription and access fee. Approximately 9,000 variants have been registered for 177 IRD genes. Comprehensive gene-specific databases were set up for the Norrie disease pseudoglioma (NDP) gene, associated with X-linked Norrie disease and vitreoretinopaty [http://www.medmolgen.uzh.ch/research/eyediseases/norriedisease/Norrinmutations.html], and CEP290, associated with Leber congenital amaurosis, early-onset RP, and several syndromic retinopathies (http://www.lovd.nl/CEP290; http://medgen.ugent.be/cep290base/). A large proportion of variants in BBS-associated genes can also be accessed online (https://lovd.euro-wabb.org/status.php). Comprehensive Leiden Open Variation Databases (LOVDs) have been developed for all nine genes implicated in Usher syndrome, (https://grenada.lumc.nl/LOVD2/Usher_montpellier/USHbases.html). Only for six other IRD-associated genes (AIPL1, http://www.LOVD.nl/AIPL1; CHM, http://www.lovd.nl/CHM; LCA5, http://www.LOVD.nl/LCA5; RDH5, http://www.LOVD.nl/RDH5; SEMA4A, http://www.LOVD.nl/SEM4A4; TULP1, http://www.LOVD.nl/TULP1), comprehensive LOVD databases have thus far been created. LOVD databases for any gene can be searched for using the URL format [Gene Symbol].lovd.nl.

The following criteria need to be met for a proper registration of sequence variants and tested individuals: (1) Web-based open access database; (2) Complete registration of all published sequence variants and tested individuals with verified causative variants; (3) Easy upload of new variants; (4) Accurate assessment of mutation data; (5) Regular updating of databases. The proper maintenance of any gene variant database relies on dedicated efforts and long-term commitments.

We propose the following plan to reach our final goal, that is, to create and maintain comprehensive and accurate mutation databases for all IRD-associated sequence variants: (1) Create gene-specific variant databases for all IRD-associated genes; (2) Invite and assign curators for at least 3-year terms; (3) Use the existing LOVD databases to update the remaining (~160) LOVDs in the next 5 years using all published peer-reviewed data; (4) Contact labs world-wide performing diagnostic services for these genes and invite them to share all (unpublished) data; (5) Set up an organization that safeguards yearly updates for all IRD genes; (6) Promote LOVD registration of sequence variants and affected persons with peer-reviewed journals.

In conclusion, we strongly advocate the development and use of gene-specific databases for IRDs, which will be a great asset for the scientific community and will prove invaluable for persons with rare IRDs.