Molecular Vision 2020; 26:226-234
Received 19 August 2019 | Accepted 24 March 2020 | Published 26 March 2020
Bing You, Xiaohui Zhang, Ke Xu, Yue Xie, Hanwen Ye, Yang Li
Beijing Institute of Ophthalmology, Beijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical University, Beijing Ophthalmology & Visual Sciences Key Lab. Beijing, China
Correspondence to: Yang Li, Beijing Institute of Ophthalmology, Beijing Tongren Hospital, Hougou Lane 17, Chong Nei Street, Beijing, 100730, China; Phone: 8610-58265915; FAX: 8610-65288561 or 65130796; email: firstname.lastname@example.org
Purpose: Aniridia is a rare congenital panocular disease caused by mutations in PAX6. The purposes of this study were to clarify the mutation features of PAX6 in a cohort of Chinese patients with aniridia and to describe their clinical characteristics.
Methods: We recruited 95 patients from 65 unrelated families clinically diagnosed with aniridia. All patients underwent ophthalmic examinations. Sanger sequencing and multiplex ligation probe amplification of PAX6 were performed to detect intragenic variants and copy number variations (CNVs).
Results: We identified 58 disease-causing mutations in PAX6 in 63 families; the detection rate was 96.9%. The 58 mutations included frameshift indels (27.6%), splice site changes (25.9%), nonsense mutations (20.7%), CNVs (19.0%), missense mutations (3.4%), run-on mutations (1.7%), and a synonymous mutation (1.7%). Clinical examinations revealed that 71 patients had complete or almost complete iris loss, 16 patients showed partial iris loss, and six patients had a full iris but with an abnormal structure.
Conclusions: The results confirmed that mutations in PAX6 are the predominant cause of aniridia, and the majority are loss-of-function mutations that usually result in classical aniridia. In contrast, missense mutations, run-on mutations, and small numbers of splicing mutations mostly lead to atypical aniridia and an intrafamilial phenotypic variability of iris hypoplasia.
Aniridia is a rare congenital panocular disease characterized by a variable degree of hypoplasia of the iris and an incidence rate of between 1:40,000 and 1:96,000 [1-3]. Most patients with aniridia also have nystagmus, foveal hypoplasia, and visual defects. As the disease progresses, 50–90% of patients may develop keratopathy, cataract, and glaucoma in childhood or in the first two decades of life . Aniridia may present as an isolated ocular disorder or as a part of Wilms’ tumor, aniridia, genitourinary abnormalities, mental retardation (WAGR) syndrome, and obesity (WAGRO) syndrome [1-3]. Aniridia is an autosomal dominant disorder with high penetrance and variable expression. Approximately two-thirds of aniridia cases have an affected parent, and the remaining one-third of patients are sporadic, usually carrying a de novo gene mutation. Mutations in the paired box 6 (PAX6; Gene ID: 5080, OMIM: 607108) gene and its associated regulatory regions are major causes of aniridia [1-3].
The PAX6 gene, located on chromosome 11p13, encodes a highly conserved transcriptional regulatory protein that is expressed in the developing eye, brain, spinal cord, and pancreas [4,5]. The protein consists of two DNA binding domains—a paired domain (PD) and a homeodomain (HD) separated by a 79-amino acid linker peptide (LNK)—and a proline, serine, and threonine-rich (PST) transcriptional transactivation domain [4,5]. Thus far, more than 400 distinct PAX6 variants have been reported based on the Human PAX6 Mutation Database) . The majority (about 77%) of the reported intragenic mutations are nonsense, frameshift small insertion or deletion (indel), and splicing mutations that introduce a premature termination codon (PTC) and trigger a nonsense-mediated decay (NMD) process . Missense mutations account for only 11.7% of the intragenic mutations of PAX6 . Large heterozygous genomic deletions at the 11p13 region that include PAX6 or its associated regulatory regions have also been frequently reported [6-9]. These loss-of-function nucleotide variants and haploinsufficiency of PAX6 represent the main mechanism underlying the occurrence of aniridia [6,7].
To date, PAX6-related aniridia in Caucasian patients has been extensively described [10-13]. However, similar studies on the PAX6 mutation spectrum and the clinical features in Chinese patients are limited, especially in large cohorts. In the present study, we screened mutations in PAX6 in 95 patients drawn from 65 unrelated families who showed different extents of hypoplasia of the iris. We described their phenotypes in detail and used minigene analysis to evaluate the influence of three single nucleotide variants (SNVs) on the splicing pattern.
This study was performed according to the tenets of the Declaration of Helsinki for research involving human subjects and was approved by the Beijing Tongren Hospital Joint Committee on Clinical Investigation (Beijing, China). This study also adhered to the ARVO statement on human subjects. Informed consent was obtained from all patients after a full explanation of the procedures. We enlisted 95 patients with aniridia from 65 unrelated families, including five previously described families [14,15]. Of these families, 39 probands had a family history showing an autosomal dominant inheritance pattern, while the remaining 26 probands were sporadic. All patients underwent a standard ophthalmological examination consisting of best-corrected visual acuity (BCVA), slip-lamp biomicroscopy, and fundus examination. About half of the patients also underwent optical coherence tomography (OCT) examinations (Heidelberg OCT Spectralis, Heidelberg, Germany, or Ivue-100, Optovue Inc., Fremont, CA). Aniridia-related keratopathy (ARK) severity was graded into 0 to 4 stages according to the 5-point scale detailed described by Lagali et al.  (Figure 1).
Peripheral blood samples were obtained by venipuncture from all participants and preserved on 4°C before use. The genomic DNA was extracted using a genomic DNA extraction and purification kit (Vigorous, Beijing, China), following the manufacturer’s protocol. Whole exons of the PAX6 gene (including exon 5a) were amplified with 13 pairs of primers, as previously described . PCR assays were carried out using standard reaction mixtures. After an initial denaturation at 95 °C for 5 min, reaction mixes underwent 35 cycles of 95 °C for 30 s, annealing for 30 s, and 72 °C for 30 s followed by 72 °C for 10 min. The sequencing results were compared with a published cDNA sequence for PAX6 (GenBank NM_000280.4). In silico analysis was performed using Mutation Taster, PolyPhen-2, and SIFT when a missense variant was encountered. For variants that had a suspected splicing effect, we performed in silico analysis using the NetGene2 Server, Human Splice Finder (HSF), and Berkeley Drosophila Genome Project (BDGP). The pathogenicity assessment was based on the guidelines presented by the American College of Medical Genetics and Genomics (ACMG) . Cosegregation analysis was performed when a PAX6 variant was detected, and other family members were available.
For 14 patients who had no intragenic PAX6 mutation identified, we performed Multiplex ligation-dependent probe amplification (MLPA) to detect genomic rearrangement in the chromosome 11p13 region. The MLPA analyses were carried out using the SALSA MLPA P219-B2 PAX6 probe mix (MRC Holland, Amsterdam, Netherlands) following the manufacturer’s protocol. The detailed methods had been previously described .
We evaluated an effect on splicing of three SNVs (one synonymous and two missense mutations) using minigene assays (Exontrap, MoBiTec GmbH, Goettingen, Germany) and pET01 Exontrap vectors. Human embryonic kidney 293 (HEK-293) cells and human retinal pigmented epithelium-19 (ARPE-19) cells grown under standard conditions were used for transfection. The cells were verified as authentic with short tandem repeat (STR) analysis (detailed methods and results of STR analysis are described in Appendix 1). Primers containing a 5′-BamHI site and a 3′-NotI site were designed to amplify the exon of interest and about 300–700 bp of flanking introns. Amplified genomic DNA from three patients with an SNV and one patient with a canonical splicing mutation c.357+1G>A (as a positive control) was cloned to pET01 Exontrap vectors, according to the manufacturer’s instructions. Clones with either the mutant or wild-type allele were detected and selected with Sanger sequencing. The selected cloned minigene plasmids were then transfected into HEK-293 or ARPE-19 cells using Lipofectamine 2000 DNA transfection reagent (Invitrogen, Carlsbad, CA), according to the manufacturer’s instructions. Reverse transcription-polymerase chain reaction (RT-PCR) was performed using the FastKing One-Step RT–PCR Kit (Tiangen Biotech). After 48 h, the transfected cells were harvested, and total RNA was extracted with the RNAprep Pure Cell/Bacteria Kit (Tiangen Biotech, Beijing, China). After an initial reverse transcription at 42 °C for 30 min and denaturation at 95 °C for 3 min, reaction mixes underwent 35 cycles of 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s followed by 72 °C for 5 min. The RT-PCR products were analyzed with 2% agarose gel electrophoresis, and each gel-purified DNA band was confirmed with Sanger sequencing.
We identified 47 distinct intragenic disease-causing mutations in PAX6 and 11 larger genomic deletions in the chromosome 11p13 region in 63 families, for a detection rate of 96.9%. In two sporadic cases, no disease-causing mutation was detected. The 47 intragenic mutations included 16 frame-shift indel, 15 splicing effect, 12 nonsense, two missense, one run-on, and one synonymous mutation (Appendix 2). The distribution of each intragenic mutation in PAX6 is shown in Figure 2A. Of the 11 different-length large genomic deletions (from at least 11 kb to 6.2 Mb), seven included a part or all of the PAX6 gene, three encompassed PAX6 and WT1 (Gene ID: 7490, OMIM: 607102), and one involved only the PAX6 downstream-associated regulatory region (Appendix 2, Figure 2B). The most common mutation was p.X423Lfsext*36, which was identified in four probands with an allele frequency of 6% (4/63). Mutation p.R203X and a gross deletion encompassing ELP4 (Gene ID: 26610, OMIM: 606985), PAX6, and DKFZp686K1684 were each detected in two probands with an allele frequency of 3% (2/63), while the remaining 55 mutations were each found only in one proband. Of the 58 mutations, 29 mutations were identified for the first time in the present study. These novel mutations were not found in any public databases, such as the Exome Variant Server and 1000 Genomes, and were classified as pathogenic variants based on the ACMG guidelines (Appendix 2).
Two novel missense mutations and one synonymous mutation were predicted by NetGene2 or HSF to affect splicing; therefore, we performed minigene assays to reveal the influences of these mutations on PAX6 pre-mRNA splicing. The synonymous mutation c.174C>T and the positive control c.357+1G>A introduced a new donor splicing site within exon 6, resulting in a 185 bp and 108 bp shortening of exon 6, respectively (Figure 3A). In contrast, the two missense mutations c.359T>A and c.780T>G did not show any abnormal splicing bands in either HEK-293 or ARPE-19 cells (Figure 3B,C).
In total, 93 patients from 63 unrelated families (24 sporadic and 39 with family history) were found to carry disease-causing mutations in PAX6. Cosegregation analyses performed in 56 (89%) unrelated families revealed 21 sporadic probands, each carrying a de novo mutation. (See Appendix 3 for a summary of the clinical findings for the 93 patients.) This cohort of patients included 43 women and girls and 50 men and boys. Their age at last examination ranged from 2 months to 71 years, and their BCVA varied from 0.8 to no light perception (NLP). The patients showed different extents of iris hypoplasia: 67 patients had complete loss of the iris (Figure 4A), four patients had almost complete iris loss (Figure 4B), 16 patients had partial iris loss (Figure 4C), and six patients had a full iris but with abnormal structures (Figure 4D). Overall, 93% of the patients (70/75) whose fundi were examinable with ophthalmoscopy or OCT presented foveal hypoplasia (Figure 4E,F). In addition, 87% of the patients (59/68) whose ARK was evaluated and graded with color anterior segment photography showed stage 1–4 keratopathy (Supplementary Table S2-Appendix 3). We also observed nystagmus (87%), cataract (87%), glaucoma (15%), microcornea (13%), tunica vasculosa lentis (TVL) or persistent pupillary membrane (12%), ectopia lentis (3%), and optic nerve hypoplasia (ONH; 1%) in the patients. Three sporadic patients who presented with non-ocular complications were diagnosed with WAGR (026130 and 026155) and WAGRO (026046) syndromes. We observed intrafamilial phenotype heterogeneity in five families (026027, 026058, 026059, 026123, and 026151; Figure 5).
Two patients (026071 and 026136) had no mutation in PAX6, and both showed partial loss of the iris (Appendix 4).
All 71 patients with complete or almost complete iris loss carried either gross genomic deletions or frameshift indel, nonsense, or splicing effect mutations. All the splicing effect mutations identified in the 71 patients were predicted to introduce a PTC, except mutation c.357+1G>A (which resulted in an in-frame deletion of PAX6) and mutation c.-128–1G>C. In contrast, six patients with the run-on mutation p.X423Lfsext*36 all presented with a partial loss of the iris, and four patients with two missense mutations displayed full or partial iris loss. Of the five families showing intrafamilial phenotypic variability, two families were detected with missense mutations (p.V120E and p.N260K), two pedigrees had splicing effect mutations (c.-128–1G>C and c.1183+5G>A), and one family had a small frameshift deletion (p.P286Lfs*79). A novel mutation c.-128–1G>C in intron 2 was presumed to result in a skipping of exon 3 (77 bp); this mutation was located in the upstream of the coding region of PAX6. Mutation c.1183+5G>A, located in intron 12 and presumed to escape the NMD process (Figure 2A), was predicted to introduce a frameshift at amino acid 396 and add 15 residues to the C-terminal region of the PAX6 protein. Of the 68 patients whose ARK was graded, their ARK severity increased with aging; however, the patients carrying non-NMD mutations seemed to have milder ARK than the patients carrying gross deletions and NMD mutations (Appendix 3).
Five patients from family 026016 carried a gross deletion that encompassed only ELP4 and DCDC1 (Gene ID: 341019, OMIM: 608062). All five of these patients had relatively better visual acuity than that observed in either the patients with gross deletions involving PAX6 or in the patients with NMD mutations. These five patients were also the only ones in this cohort who did not present with foveal hypoplasia.
In this study, we screened the PAX6 gene in 95 patients from 65 unrelated families, making this the largest cohort of patients with congenital aniridia examined thus far in the Chinese population. We identified mutations in the PAX6 gene in 96.9% of the probands with aniridia, which was similar to the previous reports on two recent large cohort studies [11,12]. However, the present detection rate was much higher than the one (55%) previously reported in a Chinese cohort that included 38 probands .
The 58 distinct mutations identified in the present study included almost all types of mutation in PAX6. Of the 47 intragenic mutations, 19 were detected in the PD (40%), followed by ten mutations in the LNK regions (21%), ten mutations in the PST transactivation domain (21%), five mutations in the HD (11%), and three mutations in the 5′ untranslated region (6%). Analysis of the different types of mutations in PAX6 and their locations can help distinguish potential functional effects of the nature of the mutations. The 16 frameshift indels and 12 nonsense mutations were located between exons 5 and 11, where a PTC can activate the NMD process . The minigene assay results demonstrated that the synonymous mutation c.174C>T also introduced a PTC in exon 6, as previously described . Among the 15 splice mutations, nine were predicted to introduce a PTC that could subsequently activate the NMD process, while the remaining six mutations were relatively more complicated. The minigene assays indicted that the mutation c.357+1G>A resulted in an in-frame 36-amino-acid deletion in exon 6, which was consistent with a previous observation . As the deletion covered the C-terminal subdomain of the paired domain, this mutation might significantly disrupt DNA binding . Mutations c.-129+2T>A, c.-128–1G>C, and c.-52+1G>T in intron 2 and 3 were presumed to result in skipping of exon2 (188 bp) or exon 3 (77 bp), while exon 2 and exon 3 were located in the upstream of the coding region of PAX6. A previously reported mutation c.142–139T>C in intron 5–1 resulted in a significant increase in transcription of the 436 amino acid isoform (including exon 5a) and changed the ratio of isoforms 436 and 422 . Mutation c.1183+5G>A in intron 12, a location that was predicted to escape the NMD process, was presumed to result in 411 amino acids of PAX6. As this mutation is located on the non-NMD region, it might cause aniridia by dominant negative mechanism. Additionally, the two missense mutations with the full-length protein, as well as a C-terminal extension mutation p.X423Lfsext*36 (the most common mutation in the current cohort), also caused aniridia by a dominant negative mechanism.
In the present cohort, we detected 11 gross deletions in 12 probands (19%, 12/63). The proportion of patients carrying large deletions (19%) was similar to that observed previously in 66 families with aniridia (15%) , but it was much lower than the proportion determined previously in 91 Russian families (31%) . Of the 12 probands in the present study who presented with large deletions, six (50%) had a family history; this rate was much higher than the previously reported rate of about 20% [9,11,12]. All the patients with gross deletions presented typical aniridia (complete or almost complete absence of the iris), and their visual defects were no more severe than the defects found in patients with PTC mutations. Consistent with previous observations [23,24], the patients with gross deletions sparing PAX6 had relatively mild visual defects. In this cohort, we found only three probands harboring gross deletions involving PAX6 and WT1, and these patients all presented with clinical manifestations of WAGR or WAGRO syndrome. Two developed glaucoma at an early age (at 1 and 8 years old); this percentage seemed to be higher than the percentage observed in patients with gross deletions not involving WT1.
In this large cohort of patients with aniridia, we observed genotype–phenotype correlations to some extent. We found that about 96% (68/71) of the patients with typical aniridia carried either an intragenic mutation in PAX6, which would induce the NMD process, or they carried a large deletion involving PAX6. In contrast, the majority of patients (68%, 15/22) with partial hypoplasia of the iris or with a full iris with abnormal structure had either a run-on or missense mutation, or they had splicing mutations that did not involve the NMD process. None of the four patients with missense mutations from two unrelated families displayed typical aniridia, which was consistent with previous observations [7,10]. Six patients with mutation p.X423Lfsext*36 from four unrelated families all presented with partial hypoplasia of the iris; however, none of them had the exudative vascular retinopathy described previously in two patients with the same mutation . In the present cohort, patients from five unrelated families showed intrafamilial phenotypic variability. The mutations identified in these five families included two missense, two splicing, and one frameshift deletion. As expected, the two splicing mutations did not appear to involve the NMD process. The BCVA of the patients was usually related to patient age, ARK severity, cataract, glaucoma, foveal hypoplasia, and the choice of treatment or surgery. The five patients with gross deletion involving only DCDC1 and ELP4 had relative better visual acuity; however, we did not observe any other solid correlation between BCVA and the genotypes. This might be because the majority of the patients (83%) carried a gross deletion or NMD mutation, while the proportion of patients with missense mutations was small.
We did not identify any mutations in PAX6 in two probands (patient 026071 and 026136) who presented with partial hypoplasia of the iris. These probands’ disease-causing mutations might be located in the deep intronic regions of PAX6 or in the regulatory elements downstream of PAX6 [25,26], which were not screened, or they may reside in other genes (possibly FOXC1-Gene ID: 2296, OMIM: 601090, PITX2- Gene ID: 5308, OMIM: 601542, or TRIM44- Gene ID: 54765, OMIM: 612298) [9,27].
In conclusion, the present results further demonstrate that mutations in PAX6 are the predominant cause of aniridia, and that the majority of these mutations are loss-of-function mutations that usually result in a classical aniridia phenotype. In contrast, the missense mutations, run-on mutations, and a small number of splicing mutations, which do not involve the NMD process, mostly lead to atypical aniridia and intrafamilial phenotypic variability.
APPENDIX 1. The methods and results of STR analysis for HEK-293 and ARPE-19 cell line authentication.
APPENDIX 2. Presumed pathogenic variants identified in this study and analysis of the variants by predictive programs.
APPENDIX 3. SThe clinical features and the PAX6 mutation screening results of the patients with aniridia in this study.
APPENDIX 4. Colored anterior segment (CAS) and fundus photographs of patients with negative PAX6 findings.
This work was supported by the National Key R&D Program of China, 2016YFC0905200. The funding organization had no role in the design or conducting of this research. Financial support: This work was supported by the National Key R&D Program of China, 2016YFC0905200. The funding organization had no role in the design or conduct of this research.