Introduction

Leber congenital amaurosis (LCA, MIM 204000) is a severe form of inherited retinal dystrophy, characterized by severe visual loss at or near birth, Franceschetti's oculo-digital sign, searching or wandering nystagmus, and pigmentary retinopathy [1], [2]. Visual acuity is rarely better than 20/400 [3]. Fundus changes are extremely variable, ranging from a normal appearance to an obvious pigmentary retinopathy similar to retinitis pigmentosa. Electroretinogram (ERG) recordings are usually extinguished or severely subnormal. In most cases, LCA shows an autosomal recessive pattern of inheritance. However, several families with autosomal dominant LCA have been well documented [2]. The prevalence of LCA is around 1–2 per 80,000 live births, accounting for approximately 20% of cases of inherited blindness among children and more than 5% of all congenital retinopathies [4]. Currently, 16 loci for LCA have been mapped, at which mutations in 15 genes have been identified as being responsible for the disease: GUCY2D [5], CRB1 [6], RPE65 [7], RPGRIP1 [8], AIPL1 [9], LCA5 [10], CRX [11], LRAT [12], TULP1 [13], RDH12 [14], CEP290 [15], RD3 [16], SPATA7 [17], IMPDH1 [18], and MERTK [19].

It has been estimated that mutations in these 15 genes are responsible for about 65% of all LCA cases [2]. The most frequently mutated genes in published studies are CEP290 (15%), followed by GUCY2D (12%), CRB1 (10%), RPE65 (6%), AIPL1 (5.3%), RPGRIP1 (4.2%), LCA5 (1.8%), CRX (1.0%), and MERTK (0.6%). However, the full frequency spectrum of variation in these 15 genes has not been evaluated in East Asia, an area containing one third of the world’s population. In this study, we performed a comprehensive evaluation of variation in these 15 genes in 87 unrelated Han Chinese patients with LCA.

Methods

Mutation screening

The most frequently mutated exons in the 15 genes, plus one intron, were selected based on previous reports for the initial scan for variation in the 87 patients [2]. The primer pairs (Table S1) used to amplify genomic fragments encompassing each of the exons and the adjacent 100 bp or more flanking intronic regions were designed using Primer 3 (http://frodo.wi.mit.edu/primer3/). PCR amplifications were carried out in 10-µl reactions containing 40 ng genomic DNA. Touchdown PCR amplification consisted of a denaturizing step at 95°C for 5 minutes, followed by 35 cycles of amplification (at 95°C for 30 seconds, at 64∼57°C for 30 seconds starting from 64°C with decreasing by 0.5°C every cycle for 14 cycles until remaining at 57°C for 21 cycles, and at 72°C for 40 seconds), and a final extension at 72°C for 10 minutes. The amplicons were sequenced with an ABI BigDye Terminator cycle sequencing kit v3.1 (Applied Biosystems, Foster City, CA), electrophoresed on an ABI3100 or ABI 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA), and analyzed with Seqman software (Lasergene 8.0, DNASTAR, Madison, WI) or Mutation Surveyor (SoftGenetics, State College, PA).

Subsequently, all the remaining exons in 11 of the 15 genes (GUCY2D, CRB1, RPE65, RPGRIP1, AIPL1, LCA5, CRX, SPATA7, CEP290, RDH12 and MERTK) were amplified and sequenced for all patients. For the remaining 4 genes including TULP1, RD3, LRAT, and IMPDH1, because of the rarity of reported variation in these genes and exons, only exons with previously reported variants were analyzed. The exons analyzed in our initial scan and subsequent analyses are listed in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The exons and one intron sequenced in our initial screen and in subsequent analyses.

https://doi.org/10.1371/journal.pone.0019458.t001

A variant was predicted to be pathogenic if it was predicted to be damaging by Blosum 62, predicted to be possibly damaging (a value of 1.5 or greater) by the position specific independent counts algorithm of Polyphen [20], had a SIFT score (which distinguishes tolerated variants from those that are not tolerated) equal to or less than 0.05, and was absent in 192 ethnically matched control chromosomes. Splicing changes were predicted by Automated Splice Site Analyses (https://splice.uwo.ca/).

Results

Fifty-three different variants were identified in 44 of the 87 patients (50.6%), involving 78 of the 88 alleles (11 homozygous and 56 heterozygous) (Tables S2 and S3). These include 8 homozygous or 12 compound heterozygous cases, 5 cases in whom two mutations in one gene and a single mutation in a second gene were identified (triallelic), 3 patients in whom two mutations in different genes were identified (digenic), and 16 patients in whom only a single heterozygous variant was identified. Of the 53 variants identified, 35 (66%) were novel and predicted to be pathogenic, 9 (17%) were known to be pathogenic, and 9 (17%) were novel neutral or unknown effects. An initial scan of 55 amplicons (out of 215 in total, 25.6%) detected 83.3% (65/78) of the mutant alleles detected by the full scan. Sequencing the remaining exons in 11 of the 15 genes detected only 13 additional mutant alleles (16.9%).

The percentage of patients who had variants in the 15 genes were, in decreasing order: GUCY2D 16.1% (14/87), CRB1 11.5% (10/87), RPGRIP1 8% (7/87), RPE65 5.7% (5/87), SPATA7 4.6% (4/87), CEP290 4.6% (4/87), CRX 3.4% (3/87), LCA5 2.3% (2/87), MERTK 2.3% (2/87), AIPL1 1.1% (1/87), and RDH12 1.1% (1/87) (all variants were taking into account for those patients who had variants in more than one gene). For individual genes, the majority of variants were found in one exon (RPE65, CRX, LCA5, AIPL1, and RDH12), two exons (CRB1, RPGRIP1, SPATA7, CEP290 and MERTK), or three exons (GUCY2D) (Table 2). Further analysis revealed that 9 exons (4.19% fragments sequenced, 9/215) contained 50.8% (39/78) of the variant alleles detected in this study (Table S4). Polymorphisms detected in these genes are shown in Table S5. Clinical data for the 44 patients in which variants were found were listed in Table S6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Frequencies of variant alleles detected in the individual exons of each gene.

https://doi.org/10.1371/journal.pone.0019458.t002

Discussion

Leber congenital amaurosis is the earliest occurring and most severe inherited retinal dystrophy. Since the initial identification of mutations to GUCY2D as a cause of LCA in 1996 [5], mutations in a total of 15 genes have been identified as being responsible for LCA, accounting for approximately 65%–70% of LCA cases [2], [27], [28]. However, the variant frequencies for these genes vary between different ethnic groups. In northern Pakistan, the genes most commonly mutated in LCA are RPGRIP1 (29% of families), AIPL1 (21% of families), and LCA5 (21% of families); whereas in Caucasian populations, mutations in RPGRIP1, AIPL1, and LCA5 account for only 4.2%, 5.3%, and 1.8% of LCA cases, respectively [2]. Although CEP290 is the most commonly identified mutant gene in Caucasian LCA patients, accounting for 15% of LCA cases [29], mutations in CEP290 have not been detected in LCA patients from Korea [30], Saudi Arabia [31], northern Pakistan [32], or southern India [33]. An understanding of the frequency spectrum of variants in these 15 genes in different populations will not only facilitate genetic diagnosis and genetic counseling for this disastrous disorder, but will also identify patients who may potentially benefit from gene therapy or other possible interventions [31]. Because the NPHP5 gene was only recently reported as related to LCA by Stone EM, et al [34], this gene was not included into our study.

Here, we performed comprehensive mutational analysis of 15 genes in 87 unrelated Chinese patients with LCA. Of the 15 genes, all coding exons in 11 genes were analyzed by cycle sequencing. For the remaining 4 genes which have only infrequently been identified as causal for LCA, only exons in which variants had previously been identified were sequenced. The percentage of patients who had variants in the 15 genes were, in decreasing order: GUCY2D 16.1% (14/87), CRB1 11.5% (10/87), RPGRIP1 8% (7/87), RPE65 5.7% (5/87), SPATA7 4.6% (4/87), CEP290 4.6% (4/87), CRX 3.4% (3/87), LCA5 2.3% (2/87), MERTK 2.3% (2/87), AIPL1 1.1% (1/87), and RDH12 1.1% (1/87). Thus, the frequencies of variants in different genes vary remarkably in different populations. GUCY2D was the gene most frequently mutated in Chinese patients while the genes found to be most frequently mutated in other populations were RPGRIP1 in people from northern Pakistan [32], CRB1 in the Spanish [28], and CEP290 in Caucasians [15]. The overall rate variant detection was 50.6% (44/87) in this study, which is comparable to the worldwide variations detection rate of 65% if the difference between the frequency of variants in CEP290 is taken into consideration (CEP290 mutations occur in 15% of Caucasian patients but in only in 6.4% of Chinese patients) [2].

Homozygous mutations were identified in a number of cases, often related to the family and population structure. For example, family LH22 was having homozygous T55M mutations in GUCY2D, families LH32 and QT585 showing homozygous S611P mutations in CRB1, and family RP208 showing a single base deletion in RPGRIP1 are from consanguineous matings in an isolated Chaoshan population. Families QT521 and QT608 showing splicing and M784R mutations in GUCY2D respectively, family LH24 with a R395X mutation in SPATA7 and family QT338 showing an 18 base pair insertion in AIPL1 are not know to be consanguineous, but are from isolated populations.

Identifying the most frequently mutated exons in these 15 genes will greatly facilitate detection of variation in LCA patients in a clinical testing. Our initial scan of 25.6% (55 of 215) of exons, based on the frequencies of variation in other populations, detected 83.3% (65/78) of all the mutant alleles we detected. This indicates that reducing the amount of labor and cost by 75% would still result in detection of over three quarters of the variation in our samples. Based on our study of 87 Chinese patients, sequencing only the 9 most polymorphic (4.18% of the 215 sequenced regions) exons (exons 2, 11, and 12 of GUCY2D; exons 6, 11, and 9 of CRB1; exon 4 of RPE65; exon 3 of RPGRIP1; and exon 6 of CEP290) would result in detection of 50.8% (39/78) of all variants found in this study. This would use less than 5% of the labor and cost of a comprehensive sequencing strategy and detect over 50% of the variation present in other groups of Chinese patients. Screening these exons in order of amount of expected variation would be predicted to cut expenses significantly while still being accurate and comprehensive. Our results provide a key bridge between bench and bed side and should make genetic diagnose of LCA in Chinese patients more accessible and practical. This should greatly enhance the clinical genetic counseling, diagnosis, and early intervention of LCA in the Chinese population. These results also highlight the importance of analyzing the causative genes and their exons in different ethnic groups in a systematic and population-specific fashion.

Another important point from our study is that some sequence variations might mistakenly be thought to be causative mutations for LCA if only a single individual gene were analyzed. For example, the heterozygous c.164C>T (p.T55M) variant in GUCY2D might be incorrectly labeled a common mutation for LCA because it was found in 3 unrelated patients with LCA but not in 96 controls, was predicted to be a pathogenic mutation by Polyphen, and was identified as a causative mutation in SIFT. This might be the case for patient LH22, who had homozygous c.164C>T mutations, but it may not cause LCA in patient QT453 since QT453 had homozygous nonsense mutations (G1226X) in CRB1. Even though both mutations themselves could be pathogenic, digenic mutations may not necessarily cause disease, since one of the phenotypically normal parents in families LH16, QT479, QT453, QT509, and RP208 presumably carry digenic mutations. In this study, digenic variants were detected in 3 patients, including LH15 (GUCY2D and SPATA7), QT659 (CRB1 and LCA5), and RP143 (CEP290 and MERTK). Whether these mutations are indeed causative and what types of digenic variations may be responsible for LCA needs additional study. For the 5 patients (LH16, QT479, QT509, QT453, and RP208) with 3 or 4 variants in two or 3 genes, their clinical phenotypes are comparable to patients with homozygous or compound heterozygous mutations in a single gene (Table S6), suggesting that the third or fourth mutant allele may not have an additive effect and may be more likely to act as a benign variant. In such cases, the 6 additional variant alleles in these 5 patients should not be counted toward the total number of mutant alleles and, therefore, the adjusted number of mutant alleles detected should be 72, not 78 (from 44 patients). In an earlier genotyping microarray study, Zernant, et al. determined 7.3% of LCA patients carry a third mutant allele [35], a greater fraction than would be expected by chance. On the other hand, single heterozygous variants were detected in genes known to cause autosomal recessive LCA in 16 patients (15 predicted to be pathogenic mutations and 1 predicted to be neutral). This might simply represent a failure to detect a second variant. Conversely, it is possible that neutral or silent variants or polymorphisms in the 15 genes sequenced here or in other genes might serve as disease modifier alleles with another major gene defect occurring simultaneously [36]. Additional comprehensive studies, or perhaps further work with animal models, are necessary to answer this question.

In conclusion, systematic analysis of the full frequency spectrum of variation in the 15 selected genes not only gives us an overview of the molecular etiology of LCA in Chinese but also provides useful biomarkers for genetic counseling. In the near future, patients in whom pathogenic mutations are identified could become potential participants for gene therapy, and thus, identification of efficient and effective diagnostic approaches based on the population genetics specific to the patient will become increasingly important.

Supporting Information

Acknowledgments

We thank all patients and controls for their participation.