Microphthalmia is a developmental disorder in which the axial length of the globe is >2 standard deviations (SD) below the mean and manifests clinically as an overall small eye . Estimates vary but approximately 19:100,000 children are born with uni- or bilateral microphthalmia, often (∼50%) associated with other anomalies [2, 3]. The extreme form of the phenotype—anophthalmia—is rare. The precise contribution of DNA mutations to the causation of microphthalmia is unknown because mutations in all known disease genes account for <40% of cases and environmental causes are also known to contribute [3, 4]. It is hoped that the identification of additional disease genes will provide deeper insight into the molecular mechanisms that control normal eye development. Additionally, mutation identification is key to informed genetic counseling of the affected families. Our analysis of microphthalmia has revealed several disease genes, all acting recessively as predicted in our highly consanguineous population, including ODZ3, SIX6 and C12orf57 [5-7]. However, it is clear that these loci account for a very small proportion of cases. Here we describe the identification of ALDH1A3 as a disease gene for non-syndromic microphthalmia. Surprisingly, and despite the apparently low contribution of ALDH1A3 and SMOC1 to the allelic spectrum of microphthalmia/anophthalmia, we show that microphthalmia/anophthalmia in one family segregates with mutations in both of these genes.
Family 1 consists of five children with colobomatous microphthalmia or anophthalmia born to two sets of parents who are all first cousins; one sibship is affected with colobomatous microphthalmia and the other anophthalmia (Fig. 1). Clinical features are summarized in Table 1. IV:1 and IV:2 are two male siblings who presented with anophthalmia and syndactyly. On the other hand, their cousins IV:5, IV:7 and IV:8 presented with colobomatous microphthalmia with no involvement of the digits and one of them had cognitive impairment with autistic features. With written informed consent, all members of this family were recruited and blood was drawn for DNA extraction. Genotyping and autozygosity mapping were performed essentially as described before . Consistent with the observation that the phenotypes differed between the two branches, we observed no shared run of homozygosity (ROH) suggesting that microphthalmia/anophthalmia is not caused by a single homozygous mutation in this family. Indeed, autozygome-guided candidate sequencing in the branch segregating anophthalmia with syndactyly revealed a novel homozygous variant in SMOC1 [NM_001034852.2:c.857G>A (p.Arg286His)], encoding a bone morphogenetic protein (BMP) antagonist described recently as a cause of this syndromic form of anophthalmia [8-10]. This variant is absent in the Exome Variant Server and 250 ethnically matched exomes. The affected residue Arg286 is highly conserved in the orthologs of SMOC1 and in its paralog SMOC2 (Fig. S1).
|Patient ID||Gender||Age||Microphthalmia||Anopthalmia||Coloboma||Syndactyly||Developmental delay||Others||Mutationa|
|IV:1||M||18 years||-||Bilateral||-||+||?||?||SMOC1: NM_001034852.2:c.857G>A (p.Arg286His)|
|IV:2||M||13 years||-||Unilateral||-||+||-||-||SMOC1: NM_001034852.2:c.857G>A (p.Arg286His)|
|IV:5||M||9 years||Bilateral||-||OD: small inferior chorioretinal coloboma: OS: inferior iris and chorioretinal coloboma||-||-||Esotropia and hyperopia||ALDH1A3 (NM_000693.2): c.1105A>T (p.Ile369Phe)|
|IV:7||M||2 years||Bilateral||-||Bilateral colobomatous microphthalmia with cyst||-||ALDH1A3 (NM_000693.2): c.1105A>T (p.Ile369Phe)|
|IV:8||M||2 years||Bilateral||-||Bilateral colobomatous microphthalmia with cyst||-||+||ALDH1A3 (NM_000693.2): c.1105A>T (p.Ile369Phe)|
|Proband A||F||20 days||Bilateral||-||OD: severe colobomatous microphthalmos with retinal detachment; OS: inferior iris and chorioretinal coloboma||-||-||-||ALDH1A3 (NM_000693.2): c.434C>T (p.Ala145Val)|
|Sister of Proband A||F||5 years||Bilateral||-||OD: severe colobomatous microphthalmos with retinal detachment; OS: inferior iris and chorioretinal coloboma||-||-||-||ALDH1A3 (NM_000693.2): c.434C>T (p.Ala145Val)|
The three siblings in the other branch (IV:5, IV:7 and IV:8), on the other hand, had five shared ROH, none of which overlapped with a known microphthalmia disease gene. After excluding the possibility of compound heterozygous SMOC1 gene mutation in this branch by Sanger sequencing, we proceeded with exome sequencing followed by autozygome filtration as described previously . Only one variant survived the final filter (homozygous→coding/splicing→absent in dbSNP132→within the shared ROH→absent in Saudi exomes). This variant affected ALDH1A3 [NM_000693.2: c.1105A>T (p.Ile369Phe)] which encodes retinaldehyde dehydrogenase, one of three retinaldehyde dehydrogenases known to be involved in the conversion of retinal to retinoic acid in humans (Fig. 1). As retinoic acid plays a well-established role in eye development and in view of the compelling eye phenotype in Aldh1a3−/− mice, we considered this a strong candidate gene and fully sequenced it in all unsolved microphthalmia cases in our collection that show ROH overlap with it [11, 12]. Of the 10 probands that met this criteria, we were able to identify another homozygous missense change [ALDH1A3: NM_000693.2:c.434C>T (p.Ala145Val); Fig. 1] in a family with two affected (Table 1).
In the course of this study, another group identified two missense and one likely splicing mutation in ALDH1A3 in patients with microphthalmia using a very similar approach to the one we described here . This represents an independent line of evidence linking ALDH1A3 to the causation of microphthalmia in humans. Interestingly, autism and congenital heart disease at a varying frequency were reported in that study and we also show here that both normal and abnormal cognitive development are compatible phenotypes with ALDH1A3 mutations, even within the same family (Table 1). Of note, in addition to the eye developmental defect, Aldh1a3−/− mice die shortly after birth due to birth asphyxia caused by failure of nose development . This suggests that the preponderance of missense variants in the mutation spectrum of ALDH1A3 in humans (five out of six) may reflect ascertainment bias; more severe mutations may result in a syndromic lethal phenotype. Similar to the SMOC1 variant, the two ALDH1A3 variants we report here are not in the Exome Variant Server or our local database of 250 exomes. As shown in Fig. 1, these two variants are conserved not only in ALDH1A3 orthologs but also in the two paralogs ALDH1A1 and ALDH1A2.
In summary, we report the identification of ALDH1A3 as a disease gene for non-syndromic microphthalmia in humans, consistent with the very recent report by Fares-Taie et al. . The finding of mutations in two rare loci in one consanguineous family is a reminder that locus heterogeneity is possible even for such an uncommon phenotype as microphthalmia/anophthalmia and should be considered as a potential pitfall in any homozygosity mapping project.