Background/aims Microphthalmia-anophthalmia-coloboma (MAC) are congenital eye malformations causing a significant percentage of visually impairments in children. Although these anomalies can arise from prenatal exposure to teratogens, mutations in well-defined genes originate potentially heritable forms of MAC. Mutations in genes such as CHX10, GDF6, RAX, SOX2 and OTX2, among others, have been recognised in dominant or recessive MAC. SOX2 and OTX2 are the two most commonly mutated genes in monogenic MAC. However, as more numerous samples of MAC subjects would be analysed, a better estimation of the actual involvement of specific MAC-genes could be made. Here, a comprehensive mutational analysis of the CHX10, GDF6, RAX, SOX2 and OTX2 genes was performed in 50 MAC subjects.
Methods PCR amplification and direct automated DNA sequencing of all five genes in 50 unrelated subjects.
Results Eight mutations (16% prevalence) were recognised, including four GDF6 mutations (one novel), two novel RAX mutations, one novel OTX2 mutation and one SOX2 mutation. Anophthalmia and nanophthalmia, not previously associated with GDF6 mutations, were observed in two subjects carrying defects in this gene, expanding the spectrum of GDF6-linked ocular anomalies.
Conclusion Our study underscores the importance of genotyping large groups of patients from distinct ethnic origins for improving the estimation of the global involvement of particular MAC-causing genes.
- eye coloboma
- Eye (Globe)
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Eye development is a multistep, highly complex process requiring the coordinated interaction of numerous genes during embryonic life.1 Prenatal environmental or genetic insults disturbing this process, acting in isolation or in combination, can result in a variety of congenital eye malformations that are responsible of approximately 25% of severe visual impairments in children.2 3 Among these anomalies, the spectrum of microphthalmia-anophthalmia-coloboma (MAC) occurs with an approximate frequency of two in 10 000 in the general population.3 4
Anophthalmia, the most severe eye structural anomaly, is defined by the complete absence of ocular tissue in the orbit, while microphthalmia refers to eyes with axial lengths of two standard deviations below the age-adjusted mean.5 Coloboma is a cleft caused by absent eye tissue, typically occurring in the inferonasal quadrant of the eye due to defective closure of the optic cup embryonic fissure.6
The aetiology of MAC is diverse and includes multifactorial inheritance as well as chemical or biological teratogenic agents disrupting normal intrauterine eye development.7 8 MAC malformations can coexist within a single family, indicating that they are not separate disorders, but rather represent alternative expressions of a same genetic defect.9 10 Accordingly, in recent years it has been recognised that a subset of familial or isolated MAC cases arises from mutations in particular genes.11 These monogenic cases exhibit a high risk of familial transmission, most commonly in an autosomal dominant or in an autosomalrecessive fashion. Thus, molecular defects in genes such as SOX2,12 13 RAX14 15 OTX216 17, CHX1018–20 and GDF621 22, among others, are responsible for a highly variable spectrum of congenital eye anomalies, including MAC. Considerable intra- and interfamilial phenotypic heterogeneity is frequently associated with mutations in these genes, suggesting that additional genetic and/or environmental factors plays a role in modulating the final ocular phenotype associated to a particular mutation. Mutations in genes such as BCOR,23 MFRP,24 STRA625 and GDF326 have also been shown to cause rare syndromic or isolated forms of MAC.
Currently, the two most frequently mutated genes causing monogenic MAC are SOX2 and OTX2, accounting for 10% and 3% of cases, respectively.13 17 27 However, this percentage can vary because the number of genes implicated in Mendelian forms of MAC is expected to grow during the next few years. Importantly, the molecular analysis of large groups of MAC patients from different ethnic backgrounds would lead to a better estimation of the actual contribution of each gene and also to detect novel pathogenic mutations.
To address this issue, we performed a comprehensive screening of the MAC-causing genes SOX2, RAX, OTX2, CHX10 and GDF6 in a group of 50 unrelated Mexican subjects suffering from bilateral congenital MAC spectrum of malformations.
Diagnosis of MAC spectrum in each participating individual was based on specialised ophthalmologic clinical examination and/or imaging (orbit USG/CT scan) data. Exhaustive genealogical information was recorded and the possible occurrence of extraocular somatic defects and/or developmental abnormalities was investigated by a clinical geneticist in all subjects.
Genomic DNA from patients and controls was isolated from peripheral blood leucocytes using the semiautomated QuickGene system (Fujifilm, Tokyo, Japan). The entire coding sequences of CHX10 (located at chromosome 14q24.3 and composed of five exons), GDF6 (8q22.1, two exons), OTX2 (14q21–22, three coding exons), RAX (18q21.3, three exons) and SOX2 (3q26.3-q27, one exon) genes and their exon/intron boundaries were amplified by PCR using pairs of primers derived from the normal published sequences (see supplementary table). Direct sequencing of PCR-amplified products from all five genes was performed using the Big Dye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, California, USA). Samples were run in an ABI Prism 310 Genetic Analyser (Applied Biosystems) and gene sequences were compared using the Ensembl Database (http://www.ensembl.org) transcript IDs ENST00000325404 (SOX2), ENST 00000261980 (CHX10), ENST00000287020 (GDF6), ENST00000408990 (OTX2) and ENST00000334889 (RAX).
Nucleotidic variations were confirmed in each case sequencing the antisense strand from a new PCR amplicon. For excluding novel mutations being polymorphisms, DNAs from 200 ethnically matched healthy subjects (400 alleles) of Mexican Mestizo origin were tested by direct nucleotide sequencing, PCR-restriction length fragment polymorphism (RFLP) and/or high-resolution melting typing for each particular mutation. When available, parental DNAs of cases with proved mutation were analysed to confirm the de novo origin of the probands' mutations. All five genes were entirely sequenced in all subjects except in patients 2, 5, 31 and 41.
Computational assessment of mutations
Two sequence homology-based programs were used to predict the functional impact of the novel mutations found in this study: PolyPhen (polymorphism phenotyping; http://genetics.bwh.harvard.edu/pph) and PANTHER (Protein ANalysis THrough Evolutionary Relationships, http://www.pantherdb.org). PolyPhen predicts how damaging a particular variant may be by using a set of empirical rules based on sequence, phylogenetic and structural information characterising a particular variant. PANTHER is a statistical method for scoring the ‘functional likelihood’ of different amino acid substitutions on a wide variety of proteins and uses evolutionarily related sequences to estimate the probability of a given amino acid occurs at a particular position in a protein.
Fifty Mexican unrelated probands (27 female and 23 male) were ascertained in this study. All of them were of Mexican Mestizo origin and originated from distinct regions of the country. Table 1 summarises the ocular findings in the whole group. Nine familial cases were identified, including three with a likely autosomal dominant inheritance and six with apparent autosomal recessive transmission. Only the index cases from these families were enrolled for the molecular study. A total of 25/50 (50%) cases had bilateral microphthalmia, 11/50 (22%) cases had unilateral microphthalmia with contralateral eye malformations (most commonly coloboma), 6/50 (12%) cases exhibited bilateral anophthalmia, 6/50 (12%) cases had microphthalmia/anophthalmia and 2/50 (4%) cases had unilateral anophthalmia with contralateral coloboma. Twelve patients (24%) exhibited major extraocular malformations. Cranio-facial anomalies were the most common major defects found, followed by musculoskeletal, genitourinary and brain/neurological anomalies.
Eight mutations, four of them novel mutations, were identified in our 50 index patients MAC cohort, yielding a prevalence of 16% (90% CI 7.47 to 24.53) (table 2). Specifically, four heterozygous missense mutations in GDF6, two heterozygous mutations in RAX, one nonsense mutation in OTX2, and one 20-base pair deletion in SOX2, were recognised. Table 2 summarises the molecular and clinical findings in the eight patients carrying mutations. The clinical and genetic data of the patient carrying the SOX2 intragenic deletion have been published previously.28 Two novel RAX mutations, T50P and R110G, were demonstrated in this study (figure 1D,E). These mutations were not found in 400 ethnically matched alleles from normal controls using either direct sequencing or high-resolution melting analysis on a Rotor Gene 6000 real time thermocycler and using the Type-it HRM PCR kit (Qiagen, Mexico City, Mexico). In addition, a novel GDF6 M154T mutation was detected in a single individual (figure 1A). This change was absent from a set of 400 ethnically matched control chromosomes tested by direct DNA sequencing (100 alleles) or PCR-RFLP (300 alleles) using the restriction enzyme BsmFI. In addition, a novel R89X nonsense mutation in OTX2 was demonstrated in a single bilaterally anophthalmic patient (figure 1C). This mutation was absent in DNA from parents of the patient indicating a de novo origin (data not shown).
Previously described mutations also identified in our study included the A249E substitution in GDF622 in three unrelated patients (figure 1B) and c.70del20 in SOX2.28 A GDF6 Q119R mutation was also detected in two unrelated MAC subjects from our group, but as this variant was also observed in several control alleles it was classified as a benign polymorphism.
Computational assessment of mutations
The PolyPhen and PANTHER algorithms were used to predict whether the identified amino acid changes were likely to disrupt proteins function. According to results from this analysis, mutations were classified in two groups: (1) certainly disease-causing changes, as R89X in OTX2 (a nonsense mutation predicting a prematurely truncated protein), as well as GDF6 A249E22 and SOX2 c.70del20, previously demonstrated to be undoubtedly deleterious; and (2) probably disease-causing mutations, including the novel M154T mutation in GDF6 that was clearly deleterious by PANTHER but not by Polyphen, the RAX R110G mutation that was predicted to be deleterious by Polyphen but not by PANTHER, and the T50P mutation in RAX that was predicted to be non-deleterious by Polyphen.
We have identified eight MAC cases caused by mutations in single eye-related genes. Three of these cases carried a recurrent point mutation in GDF6, one carried a novel missense mutation in GDF6, two cases exhibit novel heterozygous mutations in RAX, one carried a novel nonsense mutation in OTX2, and one carried an intragenic SOX2 deletion. The rate of mutations in our cohort was 16% (8/50), a percentage greater than that observed in previous studies analysing monogenic causes of MAC.3 13 27
In our study, four out of 50 unrelated MAC subjects carried GDF6 mutations, yielding a frequency of 8% (90% CI 1.69 to 4.31), which is considerably greater than the 1.2% (6/489) GDF6 mutation rate recently found in a large MAC cohort.22 The data presented here suggest that, at least in our population, GDF6 is a major MAC gene and should be considered as the initial candidate in mutational screening programmes. Two different GDF6 mutations were identified in our study, including A249E (three subjects) and the novel M154T (one subject). Mutation A249E, located at the putative propeptide GDF6 domain, was recently demonstrated by Asai-Coakwell et al22 in two unrelated MAC patients. Here, three unrelated individuals carried the A249E change, indicating that residue A249 is a hotspot for GDF6 mutations. Previous functional analysis of the GDF6 A249E mutation using a SOX9-reporter luciferase assay showed that this variant has a reduced activity compared with wild-type GDF6.22
A novel GDF6 M154T mutation, also situated at the propeptide domain, was identified in our study. Although methionine at position 154 is replaced by threonine in GDF6 protein in a wide number of species (data not shown), its absence in 400 ethnically matched control alleles supports it as a disease-associated variant. Moreover, computational assessment of the M154T mutation by the PANTHER algorithm indicates that this variant is pathogenetic. Interestingly, subject 24 harbouring this novel M154T mutation exhibits bilateral anophthalmia, which, to our knowledge, is the most severe ocular phenotype associated to a GDF6 mutation reported to date.21 22 Although more studies are necessary, it is possible that M154T could be a severely eye-damaging GDF6 allele. Interestingly, patient 48 carrying the recurrent A249E GDF6 mutation exhibited nanophthalmia, a subtype of microphthalmia in which eyes are extremely short but without concomitant malformations; to the best of our knowledge, this anomaly has not been previously associated to a GDF6 mutation.21 22 Thus, the finding of anophthalmia and nanophthalmia in patients carrying GDF6 mutations from this study allows the expansion of the phenotypic spectrum of eye malformations associated to defects in this gene.
Extraocular phenotypes were also observed in subjects with GDF6 mutations from our cohort (table 2). For example, case 27 carrying the A249E mutation exhibited cleft lip and palate. To our knowledge, this is the first example of a GDF6 mutation associated with cleft lip and palate and it is a remarkable finding since GDF6 has demonstrated to be highly expressed in palate and teeth.29 30 Thus, although it is tempting to postulate that facial clefts in our microphthalmic patient 27 are related to the GDF6 A249E mutation, it is also possible that this anomaly could be a coincidental finding. It must be noted that the A249E change was recently shown to underlie sporadic and familial cases of Klippel–Feil syndrome.31 Hence, the GDF6 A249E mutation can originate a broad phenotypic spectrum including anophthalmia, nanophthalmia and microphthalmia with or without facial clefts (present study), eye coloboma with postaxial polydactyly,22 as well as Klippel–Feil syndrome without ocular involvement.22 31 Consequently, our data support the variation A249E as being a recurrent mutation in MAC subjects with or without extraocular malformations.
The novel GDF6 M154T mutation was identified in subject 24 with from our cohort. Two aspects should be discussed in the context of this new mutation: first, this patient show the most severe eye phenotype (bilateral anophthalmia) associated with a GDF6 mutation as all previously reported GDF6 mutations were found in microphthalmia or colobomata cases;21 22 second, although threonine instead of methionine at position 154 is present in GDF proteins in a wide number of species, including mouse, rat, frog, zebra fish and cow, some primates such as macaque and orangutan exhibit methionine at this position. These data suggest that GDF6 methionine 154 is species-specific and that its mutation could account for the severe eye phenotype observed in our patient. Supporting this notion, the GDF6 protein of the presimian primitive primate Tarsius syrichta possesses threonine at position 154.
The frequency of GDF6 mutations found in our study, 8%, is similar to that reported for SOX2 mutations (10%) in subjects with microphthalmia/anophthalmia in distinct series.13 27 32 Thus, our results emphasise the importance of analysing distinct populations for improving the estimation of the involvement frequency of specific MAC genes.
Two novel heterozygous RAX mutations, R110G and T50P, were recognised in patients from the present work. The R110G change is located 26 residues before the homeodomain and was predicted to be deleterious by the Polyphen algorithm. Moreover, analysis of 400 control alleles excluded this variant as a common polymorphism. Arginine 110 is highly conserved among RAX proteins of diverse species. The novel RAX T50P mutation is located 10 residues after the end of the octapeptide domain, and although it was predicted to be non-deleterious by Polyphen, its absence from a set of 400 control chromosomes and from at least 98 additional alleles from MAC individuals from this study argues against this change being a benign polymorphism. Previously, two subjects with MAC spectrum malformations were demonstrated to carry the RAX compound heterozygous mutations Q147X/R192Q and p.Ser222ArgfsX62 and p.Tyr303X, respectively, indicating a recessive mode of inheritance.14 15 More recently however, a heterozygous R66T mutation in RAX was identified in a patient with eye colobomata, suggesting that inactivation of only one RAX allele is sufficient to cause eye malformations.33 Our finding of two unrelated MAC patients with heterozygous mutations in RAX supports that a single allelic mutation in this gene is sufficient to alter ocular morphogenesis.
OTX2 mutation was identified only in one patient in our cohort of 50 MAC individuals. This subject was a 6-month-old baby with bilateral anophthalmia who carried the novel R89X nonsense mutation. This change is located within the OTX2 homeodomain and predicts a prematurely truncated protein. Previously, OTX2 mutations were found in up to 2% of subjects with anophthalmia/microphthalmia.17 Fifteen distinct OTX2 mutations have been documented to date in subjects with severe eye malformations frequently accompanied of brain anomalies, developmental delay and combined pituitary hormone deficiency.16 17 34–36 As observed in our case, most OTX2 mutations reported to date originate premature truncation of the protein product.
Several studies have estimated that SOX2 mutations are responsible of approximately 10% of cases of anophthalmia/microphthalmia.13 27 32 In our study, however, a single mutation of SOX2 was identified in a group of 50 MAC subjects (2%, 90% CI −0.56 to 8.56%). This mutation was the common 20 bp deletion c.70del20 occurring at the 5′ end of the gene and predicting a truncated protein.13 37 The very low incidence of SOX2 mutations in this cohort is interesting and emphasises the importance of genotyping distinct MAC populations for establishing more accurate gene mutation frequencies. It must be noted that a limitation of our study is that it was based entirely on nucleotide sequencing, a technique that misses large gene deletions/duplications. So, several of our SOX2-negative patients could potentially harbour gene rearrangements. Finally, mutations in CHX10, which have been previously identified in autosomal recessive forms of MAC,18–20 were not recognised in any subject from our sample.
In conclusion, our study showed a 16% frequency of monogenic aetiology for MAC spectrum in this particular sample. It should be noted however that mutations in non-analysed MAC genes as GDF3, STRA6 or BCOR could account for several of our cases. The two most commonly mutated genes in our sample were GDF6 and RAX, a finding that stresses the importance of genotyping distinct populations for assigning more accurate mutational rates to specific genes involved in eye developmental anomalies. Provided these results are representative of Mexican population, our findings could help to establish a strategy for prioritise the analysis of MAC genes in large groups of affected subjects.
This work was partially supported by a Conacyt grant (71110).
Competing interest None.
Ethics approval The study was approved by the Internal Review Board of the Institute of Ophthalmology ‘Conde de Valenciana’, at Mexico City.
Provenance and peer review Not commissioned; externally peer reviewed.