Novel homozygous large deletion including the 5′ part of the SPATA7 gene in a consanguineous Israeli Muslim Arab family

• Introduction
• Methods
• Results
• Discussion

Abstract

Purpose: To identify the genetic defect in a consanguineous Israeli Muslim Arab family with juvenile retinitis pigmentosa (RP).

Methods: DNA samples were collected from the index patient, her parents, her affected sister, and two non-affected siblings. Genome-wide linkage analysis with 250 K single nucleotide polymorphism (SNP) arrays was performed using DNA from the two affected patients. Owing to consanguinity in the family, we applied homozygosity mapping to identify the disease-causing gene. The candidate gene SPATA7 was screened for mutations with PCR amplifications and direct Sanger sequencing.

Results: Following high-density SNP arrays, we identified several homozygous genomic regions one of which included the SPATA7 gene. Several mutations in SPATA7 have been reported for various forms of retinal dystrophy, including Leber congenital amaurosis (LCA) and juvenile RP. PCR-based sequence content mapping, long-distance PCR amplifications, and subsequent sequencing analysis revealed a homozygous 63.4 kb large deletion that encompasses the 5′ part of the SPATA7 gene including exons 1–5. The mutation showed concordant segregation with the phenotype in the family as expected for autosomal recessive mode of inheritance and is consistent with a diagnosis of juvenile RP.

Conclusions: We report a novel homozygous large deletion in SPATA7 associated with juvenile RP in a consanguineous Israeli Muslim Arab family. This is the first larger deletion mutation reported for SPATA7.

Leber congenital amaurosis (LCA; OMIM 204000) and juvenile retinitis pigmentosa (RP) are closely related inherited retinal dystrophies with largely overlapping clinical characteristics. LCA is the most severe form of inherited retinal degeneration characterized by severe visual loss at or shortly after birth and nystagmus [1]. The incidence of LCA has been variably described as ranging from 1.2 [2] to 3.0 [3] per 100,000 live births. Fundus may be normal at diagnosis or present various abnormalities, including pigmentary retinopathy as seen in patients with retinitis pigmentosa [4]. In most cases, the electroretinogram (ERG) responses are non-detectable, a fact that accentuates the early loss of rod and cone photoreceptor function [5]. Usually, LCA shows an autosomal recessive pattern of inheritance. However, a small number of autosomal dominant LCA cases have been reported [6-8]. In approximately 20% of patients with LCA, mental retardation or behavioral disorders are observed [9]. Juvenile RP has an overall milder phenotype with disease onset occurring in the first years of a child’s life, night blindness, nystagmus, and severely reduced or extinguished ERG responses.

LCA was first described as an intrauterine form of RP [10]. A later classification by Foxman et al. enabled a clear differentiation between LCA and early-onset RP based on the age at onset of symptoms, severity of visual loss, and associated non-ocular abnormalities [11]. The distinction between LCA and juvenile or early-onset RP can become challenging, as the clinical features overlap. It is therefore also expected that the diseases might be caused by mutations in the same gene [12].

Previous work showed that, consistent with its clinical heterogeneity, LCA may be caused by mutations in at least 19 different genes: GUCY2D (Gene ID: 3000; OMIM 600179) [13], RPE65 (Gene ID: 6121; OMIM 180069) [14], SPATA7 (Gene ID: 55812; OMIM 609868) [12], AIPL1 (Gene ID: 23746; OMIM 604392) [15], LCA5 (Gene ID: 167691; OMIM 611408) [16], RPGRIP1 (Gene ID: 57096; OMIM 605446) [17,18], CRX (Gene ID: 1406; OMIM 602225) [7,19], CRB1 (Gene ID: 23418; OMIM 604210) [20], NMNAT1 (Gene ID: 64802; OMIM 608700) [21], CEP290 (Gene ID: 80184; OMIM 610142) [22,23], IMPDH1 (Gene ID: 3614; OMIM 146690) [8], RD3 (Gene ID: 343035; OMIM 180040) [24], RDH12 (Gene ID: 145226; OMIM 608830) [25,26], LRAT (Gene ID: 9227; OMIM 604863) [27,28], TULP1 (Gene ID: 7287; OMIM 602280) [29], KCNJ13 (Gene ID: 3769; OMIM 603208) [30], IQCB1 (Gene ID: 9657; OMIM 609237) (also NPHP5) [31,32], MERTK (Gene ID: 10461; OMIM 604705) [4,33], and OTX2 [34] (Gene ID: 5015; OMIM 600037). Mutations in some LCA genes, such as RPE65 [35], SPATA7 [12], CRB1 [36], RDH12 [26], LRAT [27], and TULP1 [37], are also known to cause autosomal recessively inherited juvenile RP. By contrast, heterozygous mutations in AIPL1 might cause an autosomal dominant form of juvenile RP [38].

Spermatogenesis associated protein 7 (SPATA7) was identified in 2003 by cloning and characterization of full-length cDNA from rat testes, and therefore designated as RSD-3; accordingly, the homologous human gene was called HSD-3.1 [39]. Human SPATA7 consists of 12 exons spanning a 52.9 kb genomic sequence and maps to chromosome 14q31.3 (Ensembl Genome Browser, transcript ID ENST00000393545, GRCh37 / hg19).SPATA7 encodes a protein of 599 amino acids that is conserved from sea urchins to humans but is absent in lower eukaryotes, such as insects and fungi [12]. In addition to expression in spermatocytes, Spata7 has been reported to be expressed in multiple retinal layers in the adult mouse retina, including the inner segments of photoreceptors as well as the ganglion cell and inner nuclear layers [12]. Expression patterns in the developing and mature mouse retina suggested that Spata7 is important for normal retinal function rather than development [12]. Two isoforms resulting from alternative splicing of exon 3 are predominantly expressed in the testis (isoform lacking exon 3) or in the brain and retina, respectively, proposing specific functions of the two SPATA7 isoforms in spermatozoa and neurons [9]. According to the Human Gene Mutation Database (HGMD), four nonsense mutations, three missense mutations, three splicing mutations, and seven small indel (insertions/deletions) mutations have been identified in SPATA7. Thus, this minor gene for LCA or juvenile RP accounts for only a small percentage of cases.

In an Israeli Muslim Arab family (TR 11) with four children (three female, one male), two children (17- and 4-year-old sisters) were clinically diagnosed with autosomal recessive juvenile RP (Figure 1). Their parents were first-degree cousins and otherwise healthy. The affected siblings were born after normal pregnancies and normal deliveries. When the older affected daughter was 1 year old, the parents noticed nystagmus and vision abnormalities. Upon clinical examination, the diagnosis of juvenile RP was made. The younger affected daughter showed similar symptoms at about the same age, and clinical examination revealed retinal degeneration similar to that in the older sister. The visual deterioration showed considerable progression, resulting in reduced visual acuity (6/24 and 6/60 in the right and left eyes, respectively), severely constricted visual fields, and extinguished full-field ERGs in the older sister (at the age of 17). Funduscopic examination revealed areas of hypopigmentation due to atrophic RPE changes with pigmentary alterations from pigment migration (“salt and pepper fundus”), attenuated arterioles, and optic disc pallor, all characteristic of juvenile RP. OCT imaging showed an overall decrease in retinal thickness with shortening of the photoreceptor outer segments, reduction in the outer nuclear layer, and RPE atrophy. The inner retinal layers were less affected. The morphological findings for the older sister are presented in Figure 2. The two affected siblings were otherwise healthy; neurologic examination demonstrated normal findings without dysmorphic features or skin lesions.

Homozygosity mapping performed with DNA samples of the two affected subjects revealed two larger homozygous regions on chromosome 8 (96.8M–116.1M) and chromosome 14 (84.9M–96.0M) and a few smaller homozygous regions on chromosomes 3, 17, and 22. The only known RD gene that localizes within these homozygous regions was SPATA7 that maps on chromosome 14 (at about 88.8M).

SPATA7 mutations have been reported for various forms of retinal dystrophy, including LCA and juvenile RP. Therefore, we thought to perform a mutation screening of SPATA7 by applying PCR amplification and direct Sanger sequencing. During this screening, we consistently failed to amplify exons 1–5 in the affected index patient, while further downstream amplicons were readily amplified. We thus hypothesized the presence of a homozygous deletion in the affected patients with a telomeric breakpoint in intron 5 of SPATA7. Taking advantage of the homozygous state of the affected patients, we further refined the localization of the telomeric and centromeric deletion breakpoint with PCR-based STS content mapping and LD PCR-based breakpoint junction amplifications (for details, see Methods), using this strategy, we mapped the centromeric deletion breakpoint to a 6 kb interval within the intergenic sequence between SPATA7 and its upstream gene potassium channel, subfamily K, member 10 (KCNK10; Gene ID 54207; OMIM 605873).

With refined breakpoint mapping, an approximately 6.5 kb deletion junction fragment was successfully amplified. We then used restriction site content mapping to further narrow the localization of the deletion breakpoint within the junction fragment and to design a novel primer pair directly flanking the deletion breakpoints. The reduced size of this PCR product facilitated subsequent Sanger sequencing to cover and define the exact breakpoints (Figure 3). This approach revealed a deletion at the SPATA7 locus that encompasses 63,411 bp and is accompanied by a small insertion of three extra nucleotides between the centromeric and telomeric breakpoints (g.88.828.457_88.891.867delinsTGG; c.1–23706_372+8679delinsTGG, ex. 1–5; GRCh37 / hg19). Since the centromeric breakpoint is located 23,417 bp upstream of SPATA7 5′ untranslated region (UTR) and the telomeric breakpoint is located 708 bp upstream of SPATA7 exon 6, the deletion included exons 1, 2, 3, 4, and 5 of the SPATA7 gene.

To test segregation of the deletion allele, all family members were analyzed with diagnostic PCR amplifications with (1) primers directly flanking the deletion breakpoints (deletion specific assay) and (2) with primers amplifying exon 2 of SPATA7 (wild-type specific assay). In doing so, we observed concordant segregation of the deletion in the family as expected for autosomal recessive mode of inheritance (Figure 4). The index patient and her affected sister were homozygous deletion carriers, whereas the unaffected parents and the two unaffected siblings were heterozygous mutation carriers.

To obtain clues about the mechanism underlying the occurrence of this deletion, bioinformatics analyses of the sequences flanking the deletion breakpoints were performed. However, except for a few unspecific short motifs distant to the actual breakpoints, we did not detect repetitive sequences or other segments of sequence homology that may indicate a homology-based mechanism underlying the deletion. We therefore propose non-homologous end joining (NHEJ), a mechanism that accurately rejoins double-strand breaks, as the most likely mechanism underlying the occurrence of this deletion. In line with this assumption, NHEJ is frequently accompanied by the insertion of one to four nucleotide(s) at the breakpoint junction site [41] as also seen in our family.

Since recessive diseases in consanguineous families are usually caused by homozygous mutations due to identity-by-descent, homozygosity mapping serves as an efficient tool for defining candidate chromosomal loci for diseases. After homozygosity mapping, we picked the correct candidate gene in a consanguineous Israeli Muslim Arab family. Using PCR and subsequent Sanger sequencing, we identified a homozygous large deletion in SPATA7 that caused juvenile RP in this family. We determined the centromeric and telomeric breakpoints of the deletion that is accompanied by a small insertion of three extra nucleotides at the site of deletion (g.88.828.457_88.891.867delinsTGG; c.1–23706_372+8679delinsTGG). The deletion includes exons 1–5 of SPATA7 as well as about 23.4 kb of upstream sequences that likely contain important regulatory sequence elements such as the proximal promoter and proximal transcription factor binding sites. We therefore argue that the mutation (g.88.828.457_88.891.867delinsTGG; c.1–23706_372+8679delinsTGG, ex. 1–5) results in a complete lack of SPATA7 gene or protein expression. Analysis of the breakpoint junctions did not provide evidence of a homology-based mechanism or non-B DNA structures underlying the occurrence of this deletion. Therefore, NHEJ seems to be the most likely mechanism.

SPATA7 is a known LCA gene also reported to cause juvenile RP. Nonetheless,SPATA7 is a minor gene for LCA and juvenile RP that accounts for only a small percentage of cases. We observed concordant segregation of the deletion in the family as expected for the autosomal recessive mode of inheritance. The two affected sisters showed similar clinical manifestations, and their functional and morphological findings support the diagnosis of juvenile RP. Both affected siblings were otherwise healthy with no non-ocular abnormalities. The observed clinical features were in line with previously reported cases [12,42]; the SPATA7 mutation caused an infantile-onset severe rod-cone dystrophy with early degeneration of the peripheral retina but comparatively preserved foveal structure, as seen on the OCT images.

Until now, no larger deletion as observed in this study has been reported for SPATA7. The Human Gene Mutation Database (HGMD) lists only seven small indel mutations, predominantly resulting in a frameshift and a premature stop codon (for instance the c.265_268delCTCA / p.L89KfsX3 mutation in Mackay et al.’s study [42]). In addition, three missense mutations, three splicing mutations, and four nonsense mutations (with c.253C>T / p.R85* [42] causing the earliest premature transcription termination codon in exon 5 and c.1183C>T / p.R395* causing the latest premature transcription termination codon in exon 11 [12]), are assumed to cause disease. Wang et al. [12] suggested a correlation between the severity of mutant alleles and the resulting clinical phenotype (i.e., the mutant alleles associated with LCA are likely to be more severe loss-of-function alleles than those associated with juvenile RP). However, this was not confirmed in the family reported here, since the large deletion detected at the SPATA7 gene locus was the cause of the overall milder clinical phenotype of juvenile RP.

The Database of Genomic Variants (DGV), a catalog of human genomic structural variations of clinically unaffected, healthy individuals, lists two deletions of 1442 bp and 1486 bp, respectively, in SPATA7 in a single Yoruban individual (NA18507) as reported by McKernan et al. [43]. The smaller of these two deletions maps to intron 5 and coincides with several smaller deletions in this region observed in several population studies. The telometric breakpoint of the large deletion reported in our family is also located in intron 5, however >1,000 bp downstream of the cluster of deletions annotated in DGV. The larger 1,486 bp deletion encompasses the alternatively spliced exon 3 of the SPATA7 gene. The consequence of such a loss of exon 3 is unknown; at least the reading frame of the coding sequence is maintained and results in the shortened SPATA7 isoform. Interestingly, DGV also lists a series of small sequence duplications in intron 2 that overlap with the deduced 5′ breakpoint of the 1,486 bp deletion in NA18507. These findings indicate the existence of two clusters of small copy number variations (CNVs) in SPATA7 that are with the exception of the 1,486 bp deletion in NA18507 restricted to intronic sequences and most likely without functional consequence.

SPATA7 was first identified in the rat testis. Then, researchers showed that Spata7 is also expressed in multiple layers of the mature mouse retina [12]. Since two different isoforms were shown to be expressed in the testis (transcript lacking exon 3) or in the brain and retina (transcript containing exon 3), specific functions of the two SPATA7 isoforms in spermatozoa and neurons were proposed [9]. Analyzing the mouse retina, Wang et al. [12] showed that Spata7 is present in the ganglion cell layer, the inner nuclear layer, and the inner segment of the photoreceptors at P21. Concerning the uniform distribution of the protein in the cytoplasm of the inner segment, it seems not to be a typical cilium protein and might be involved in a novel pathway [12]. The relatively late onset of Spata7 expression in the retina further suggested that Spata7 is important for normal retinal function rather than development [12]. Abulimiti et al. [44] showed that homozygous knockout mice exhibit impaired rhodopsin transportation to the outer segments due to the absence of Spata7 protein function. Nonetheless, additional studies are needed for a detailed understanding of the SPATA7 function and how different mutant alleles cause either LCA or the overall milder phenotype of juvenile RP, especially since the mutation reported here most likely results in a complete lack of SPATA7 protein function but causes the overall milder phenotype of juvenile RP. Thus, our findings are contrary to the suggestion of Wang et al. [12] implying a correlation between the severity of mutant alleles and the resulting clinical phenotype.

In conclusion, a genome-wide screen for homozygosity combined with a candidate gene evaluation within one homozygous region on chromosome 14 resulted in the identification of a pathogenic mutation in SPATA7. The homozygous large deletion encompasses exons 1–5 and is suggested to result in a complete lack of SPATA7 gene or protein expression. Nonetheless, the mutation was found to cause the overall milder clinical phenotype of juvenile RP, in preference to the more aggressive progression of LCA.