Spectrum, frequency, and genotype-phenotype of mutations in SPATA7.

Purpose: To describe the mutation spectrum of SPATA7 and associated ocular phenotypes.
Methods: As part of a continuing examination of the genetic basis of inherited ophthalmic diseases, sequencing variations in SPATA7 were identified in sequencing data from 5,090 probands. Mutations in SPATA7 were identified in 12 Chinese patients from ten families. Family history and clinical data were examined in detail in these patients. To evaluate possible gene-specific fundus changes, the results were combined with data from 66 patients from 50 families previously reported in the literature.
Results: Seven homozygous or compound heterozygous mutations, including two novel mutations (c.367C>T, p.Q123* and c.1083-2A>G) and five known mutations in SPATA7, were identified in ten families, including six families with Leber congenital amaurosis (LCA), three families with juvenile retinitis pigmentosa, and one family with early-onset high myopia. These families accounted for approximately 2.2% (6/269) of LCA and 0.4% (10/2,252) of inherited retinal dystrophies in this case series. A combined analysis of data from the present study and data from 60 families reported in the literature showed that 93.3% (112/120) of mutant alleles were truncation mutations, whereas only about 5.0% were missense mutations, and 1.7% were non-frameshift indels. Common SPATA7-associated fundus changes, including narrow arterioles, a relatively well-preserved macular region, and widespread RPE atrophy resulting in diffuse mottled hypopigmentation in the midperipheral retina, were identified in this cohort and in patients in the literature. Missense mutations were not associated with specific phenotypic features or severity. Conclusions: Narrow arterioles, a relatively well-preserved macular region, and widespread RPE atrophy resulting in diffuse mottling hypopigmentation in the midperipheral retina may be considered early and common fundus changes specific to SPATA7-associated retinopathy. The fact that similar mutations result in varied phenotypes points to the existence of other potential modifiers of the disease. Uncovering the identity of these modifiers might aid the development of novel treatments.

Introduction

Hereditary retinal degeneration (HRD) includes a broad group of hereditary retinal disorders with different patterns of inheritance, varied clinical phenotypes, and a highly polymorphic basis of molecular defects. These disorders are further classified into several subgroups based on the primary involvement of cones or rods, location of the retina, severity of the disease, and accompanying ocular or systemic signs. The disorders can be transmitted as autosomal dominant, autosomal recessive, X-linked, or mitochondrial traits. Thus far, mutations in at least 271 genes have been reported to be responsible for these disorders (RetNet, https://sph.uth.edu/retnet/). A systematic analysis of the coding regions of these genes identified genetic defects in approximately 50–60% of families with HRDs. A lack of mutations in these genes may suggest a genetic defect either in undefined novel genes or in uncovered regions of known genes. However, a proportion of reported mutations may not truly be causative of disease. Clarification of genotype–phenotype correlations, especially gene-specific retinal lesions, will not only help to further identify genetic defects but also clarify the pathogenesis of mutations.

Mutations in the spermatogenesis-associated protein 7 (SPATA7) gene (Gene ID: 55812, OMIM: 609868) have been reported to be responsible for autosomal recessive Leber congenital amaurosis (LCA), retinitis pigmentosa (RP), juvenile RP, and cone or cone-rod dystrophy (CORD) [1-33]. To date, biallelic mutations in SPATA7 have been reported in 54 families. However, the spectrum and frequency of mutations in SPATA7 and their associated characteristic phenotypes have not been systemically analyzed.

In the present study, we systematically analyzed variations in SPATA7, mainly based on a combination of whole exome and targeted exome in-house sequencing data from 5,090 unrelated patients. Seven biallelic mutations in SPATA7, including two novel mutations, c.367C>T (p.Q123*) and c.1083–2A>G, were observed in ten families, including four families briefly reported previously. Systemic analysis of these genotype–phenotype data, as well as related data in the literature, shed light on the spectrum and frequency of mutations in SPATA7 and identified gene-specific changes in the early developmental stage, in addition to novel fundus changes.

Methods

Venous blood was collected in an EDTA coating tube from superficial vein of the arm of each participant using disposable needle after sterilization of skin. The blood samples were stored at 4 °C temporarily for a few hours to a few days before the preparation of genomic DNA. Clinical data were collected from outpatient clinics following the Guidance of Sample Collection of Human Genetic Diseases (863-Plan) by the Ministry of Public Health of China. Written informed consent adherent to the tenets of the Declaration of Helsinki as well as the ARVO statement on human subjects was obtained from all the patients or their guardians before sample and data collection. This study was approved by the institutional review board of the Zhongshan Ophthalmic Center.

Genomic DNA was prepared from leukocytes obtained from the venous blood of the study group. Genetic defects in the probands of each family were screened with either whole exome sequencing (WES) or targeted exome sequencing (TES), and a mutation in one individual was identified with Sanger sequencing (Table 1). The WES or TES data were analyzed through multistep bioinformatics analysis as described previously [34,35]. Variants in SPATA7 were obtained for 5,090 probands with a variety of forms of genetic eye diseases, including 2,252 probands with HRDs (one proband was examined with Sanger sequencing, 2,132 probands were examined with WES described previously, and 119 additional probands were recently examined with TES).

Table 1

Clinical information of the probands and affected siblings with biallelic SPATA7 mutations.

Family ID	Clinic group	Mutation based on NM_018418	Effect	Gender	Age (year) at		Axial length/ refraction	First symptom	Visual acuity		Fundus changes	ERG recording rods and cones
					onset	1st exam			right	left
F01-II:1*	LCA	c.[1183C>T];[1183C>T]	p.[R395*];[R395*]	M	0.3	0.6	NA	PV, ODS, RN	NPL	NPL	AV, YFD	Extinguished
F02-II:1	LCA	c.[367C>T];[1183C>T]	p.[Q123*];[R395*]	F	FMB	4.3	+5.25; +5.75	PV, RV	LP	LP	AV, TD	NA
F03-II:1*	LCA	c.[644_647del];[644_647del]	p.[L215Sfs*];[L215Sfs*]	F	FMB	2.3	NA	PV, ODS, RN	LP	LP	AV, YSB	Extinguished
F03-II:2	LCA	c.[644_647del];[644_647del]	p.[L215Sfs*];[L215Sfs*]	M	FMB	1.3	NA	PV, ODS, RN	LP	LP	AV, YSB	NA
F04-II:1*	LCA	c.[1183C>T];[1183C>T]	p.[R395*];[R395*]	M	0.583333	1.3	NA	PV, ODS, RN	NPL	NPL	AV, TD	Extinguished
F05-II:1	LCA	c.[367C>T];[367C>T]	p.[Q123*];[Q123*]	F	FMB	13	NA	PV, RN	LP	LP	AV, YMD, PPP	Extinguished
F06-II:1	LCA	c.[322C>T];[322C>T]	p.[R108*];[R108*]	M	0.3	2.4	+5.00; +4.75	PV, RV	LP	LP	normal-like	Extinguished
F07-II:1*	jRP	c.[322C>T];[1183C>T]	p.[R108*];[R395*]	M	FMB	3.4	-1.50; -1.25	PV, NB, PA	0.4	0.6	AV, YMD, PPP	Extinguished
F08-II:1	jRP	c.[20_23del];[1183C>T]	p.[V7Efs*19];[R395*]	M	4	7	+3.00;+2.50	PV, NYS	0.02	0.03	AV, YMD, PPP	Extinguished
F09-II:6	jRP	c.[20_23del];[1083-2A>G]	p.[V7Efs*19];[splicing]	M	22	25	NA	NB	HM	LP	WPD, AV, IBP	NA
F10-II:1	myopia	c.[20_23del];[20_23del]	p.[V7Efs*19];[V7Efs*19]	M	NA	5	-8.50; -10.25	PV	NA	NA	NA	NA
F10-II:2	myopia	c.[20_23del];[20_23del]	p.[V7Efs*19];[V7Efs*19]	M	NA	3	24.38; 23.91mm	PV	NA	NA	NA	NA

M=male; F=female; FMB=first few months after birth; NA=not available; PV=poor vision or no pursuit of objects; ODS=ocular digital sign; RN=roving nystagmus; NB=night lindness; LP=light perception; NPL=No pursuit of light; HM=hand motion; AV=attenuated vessels; TD=tapetoretinal degeneration; YFD=Yellowish-white frosted degeneration in midperipheral retina; YSB=Yellowish-white sandy beach in midperipheral retina; YMD=Yellowish-white mottled degeneration in midperipheral retina. PPP=a few tiny pepper-powder like pigmentation; WPD=waxy pale disc; IBP=Irregular black pigmentation; NA=Not available; Ext=Extinguished.

Homozygous or compound heterozygous rare variants in SPATA7 were selected for this study, and their segregation in family members was confirmed with Sanger sequencing. The clinical data and mutation characteristics in the patients were systematically analyzed in conjunction with data reported in the literature [1-33]. The ophthalmic examinations, including visual acuity, axial length, refraction, fundus changes, and electroretinogram, were performed as we described in previous studies [36]. Gene-specific fundus changes were determined by comparing available fundus images with detailed documented descriptions and available clinical data from these patients, as well as from case studies reported in the literature. Factors affecting phenotypic changes were considered, including the type and location of mutations and the patient’s age at the time of the examination.

Results

Mutations in SPATA7 and phenotypes in the cohort

Based on an analysis of sequencing data collected at our clinic on 5,090 probands with genetic eye diseases (including 269 probands with LCA and 1,983 probands with other forms of HRDs), biallelic mutations in SPATA7 were found in ten families, with 12 affected individuals (Figure 1, Table 1). The clinical diagnosis of the probands was LCA in six families, juvenile RP (or early-onset HRD) in three families, and early-onset high myopia in one family (Table 1). In total, seven different mutations were identified in these ten families. Of these mutations, five were known mutations, and two were novel mutations, of which the c.367C>T (p.Q123*) nonsense mutation was predicted to have a 35 score by CADD (https://cadd.gs.washington.edu/), and the c.1083–2A>G mutation was predicted to affect the splicing acceptor site by BDGP and HSF (Table 2). All mutations were confirmed with Sanger sequencing and segregated with disease in the families (Figure 1 and Appendix 1). Therefore, the mutations in SPATA7 were responsible for 2.2% (6/269) of families with LCA and 0.4% (10/2,252, 269+1983) of families with HRD in this Chinese cohort from one institution. All seven mutations were predicted to result in truncated proteins when expressed.

Figure 1

Pedigrees of ten families with biallelic mutations in SPATA7. The arrow indicates the proband in each family. The affected individuals are shown as filled squares (male) or circles (female). Mutations are listed under each family, and their segregation in families is shown in the pedigrees.

Table 2

Mutations in SPATA7 identified in this study and previous literatures.

Position at chr14	Nuleotide change NM_018418	Alleles in probands	Effect	Type	CADD score	REVEL score	BDGP	HSF	Frequency in ExAC	HGMD	Our cohort	First reported
88828457	c.1-23706_372+8679delinsTGG	2	/	Gross deletion	/	/	/		/	DM	/	[12]
88852165	c.3G>A	1	p.?	Initiation	24.7	0.277	/	/	1/79940	DM	/	[6]
88852180	c.18A>G	2	p. =	Splicing	/	/	LDS	BDS	/	DM	/	[27]
88852181	c.19G>A	2	p.V7I	Missense+Splicing	/	/	LDS	BDS	/	DM	/	[23]
88857725	c.20_23delTCAG	5	p.V7Efs*19	Frameshift	/	/	/	/	/	DM	Yes	[32]
88859831	c.189delA	1	p.A64Lfs*3	Frameshift	/	/	/	/	/	DM	/	[31]
88883061	c.245dupA	1	p.D82Efs*16	Frameshift	/	/	/	/	/	DM	/	[31]
88883069	c.253C>T	13	p.R85*	Nonsense	37	/	/	/	2/120686	DM	/	[3]
88883081	c.265_268delCTCA	1	p.L89Kfs*4	Frameshift	/	/	/	/	/	DM	/	[3]
88883104	c.288T>A	12	p.C96*	Nonsense	29	/	/	/	2/120690	DM	/	[15]
88883112	c.296_297delAG	2	p.E99Vfs*5	Frameshift	/	/	/	/	/	DM	/	[22]
88883138	c.322C>T	11	p.R108*	Nonsense	35	/	/	/	4/119988	DM	Yes	[1]
88883156	c.341delA†1	1	p.N114Ifs*23	Frameshift	/	/	/	/	/	DM	/	[13]
88883183	c.367C>T	3	p.Q123*	Nonsense	35	/	/	/	/	/	Yes	Novel
88892575	c.373-1G>A†2	1	/	Splicing	/	/	LAS	BAS	/	DM	/	[13]
88892690	c.487A>T	1	p.K163*	Nonsense	35	/	/	/	/	DM	/	[11]
88892843	c.640delC†3	2	p.Q214Sfs*2	Frameshift	/	/	/	/	/	DM	/	[7]
88892847	c.644_647delTAGT	2	p.L215Sfs*30	Frameshift	/	/	/	/	/	/	Yes	[20]
88892903	c.700dupT	1	p.S234Ffs*2	Frameshift	/	/	/	/	/	DM	/	[29]
88892911	c.708_711delACAA	1	p.K236Nfs*9	Frameshift	/	/	/	/	/	DM	/	[29]
88892966	c.763C>T	3	p.Q255*	Nonsense	39	/	/	/	/	DM	/	[2]
88893049	c.845+1G>A	3	/	Splicing	/	/	LDS	BDS	1/112364	DM	/	[2]
88895690	c.913-2A>G	4	/	Splicing	/	/	LAS	BAS	/	DM	/	[14]
88895739	c.960dupA†4	4	p.P321Tfs*6	Frameshift	/	/	/	/	/	DM	/	[1]
88897545	c.1058dupC	1	p.S354Ffs*4	Frameshift	/	/	/	/	/	DM	/	[19]
88899477	c.1083-2A>G	1	/	Splicing	/	/	LAS	BAS	/	/	Yes	Novel
88899496	c.1100A>G	1	p.Y367C	Missense	26.9	0.522	/	/	2/120094	DM	/	[30]
88899498	c.1102_1103delCT	2	p.L368Efs*4	Frameshift	/	/	/	/	/	DM	/	[11]
88899508	c.1112T>C	2	p.I371T	Missense	25.5	0.234	/	/	39/120050	DM?	Yes	[24]
88903886	c.1161-1G>C	1	/	Splicing	/	/	LAS	BAS	1/120856	DM	/	[2]
88903897	c.1171C>T	3	p.R391*	Nonsense	44	/	/	/	/	DM	/	[23]
88903909	c.1183C>T	13	p.R395*	Nonsense	36	/	/	/	4/120876	DM	Yes	[1]
88903937	c.1211A>G	1	p.E404G	Missense	29.8	0.157	/	/	/	DM	/	[26]
88903941	c.1215G>T	2	p.E405D	Missense	24.4	0.113	/	/	1/120540	DM	/	[13]
88903946	c.1215+5G>A	2	/	Splicing	/	/	none	BDS	/	DM	/	[27]
88904180	c.1216-2A>T	2	/	Splicing	/	/	LAS	BAS	/	DM	/	[13]
88904181	c.1216-1G>A	2	/	Splicing	/	/	LAS	BAS	/	DM	/	[8]
88904195	c.1229_1231delACC†5	1	p.H410del	Inframe deletion	/	/	/	/	/	DM	/	[3]
88904207	c.1241_1252del12	1	p.V414_V417del	Inframe deletion	/	/	/	/	/	DM?	/	[19]
88904339	c.1373delT	4	p.V458Efs*48	Frameshift	/	/	/	/	/	DM	/	[8]
88904361	c.1395delA†6	2	p.Q465Hfs*41	Frameshift	/	/	/	/	/	DM	/	[1]

Notes: LDS, loss of a donor site; LAS, loss of an acceptor site; BDS, Broken Donor Site; BAS, Broken Acceptor Site. †1, previously reported as c.340del; †2, previously reported as c.371-1G>A; †3, previously reported as c.544delC based on NM_018418; †4, some previously reported as c.961dupA; †5, previously named as c.1227_1229delCAC; †6, previously named as c.1546delC. None of the 41 mutations was present in 1000 genome.

Clinical data on the 12 patients in the ten families are summarized in Table 1, including nine male patients and three female patients. The age of onset varied from 3 months after birth to 22 years. Poor vision or no pursuit of objects was the most common initial sign and was present in nine of ten probands, while night blindness was the initial sign in the other patients. Visual acuity ranged from no pursuit of light to 0.6. Nonrecordable cone and rod responses were recorded in seven probands using electroretinography. Fundus changes were age-dependent, with some common features possibly specific to mutations in SPATA7.

Fundus images were available for six patients from five families, including one patient with fundus images taken at three different times. The youngest patient (F01-II:1) was a 9-month-old boy with no pursuit of light for 6 months. Electroretinography recordings in this case showed no recordable responses of rods and cones. A fundus examination revealed attenuation of retinal arterioles and a relatively normal-looking posterior retina, except the absence of foveal reflex. Tiny spot-like RPE atrophy formed a frosted appearance in the midperipheral retina (Figure 2A,B). A peculiar, novel fundus change was observed in two affected siblings when they were aged 2 years and 4 months (F03-II:1) and 1 year and 4 months (F03-II:2), with yellowish-white sand-like deposits in the midperipheral retina (Figure 2C–F). Interestingly, similar retinal degeneration was absent in part of the inferior portion, close to the area of the optic fissure in the left and right eyes of the two siblings. The posterior fundus and optic disc were relatively normal, except slightly attenuated retinal arterioles (Figure 2C–F). In addition, diffuse and mottling hypopigmentation in the midperipheral retina was observed in probands from two families (F05-II:1, Figure 2G,H; F07-II:1, Figure 2I).

Figure 2

Fundus images from patients with biallelic mutations in SPATA7. Representative fundus images show yellowish-white frosted degeneration in the midperiphery in patient F01-II:1 (A and B) and yellowish-white sand-like deposits in the midperiphery sparing the inferior area close to the optic fissure (C–F), in addition to yellowish-white mottled degeneration in the midperiphery (G–I). The homozygous novel mutation c.367C>T was identified in the F05-II:1, which is highlighted in red in G and H.

Age-related fundus changes were observed in the proband from F07. Fundus images for F07-II:1 were obtained when the patient was aged 3 years and 5 months (Figure 3 A), 7 years and 5 months (Figure 3B,C), and 8 years and 7 months (Figure 3D–I). The proband attended our clinic at the ages of 3 years and 5 months due to a complaint of “bumping into obstacles.” Ocular examination revealed a relatively normal posterior retina (Figure 3A), with a few minor grayish-white spots in the midperiphery. Electroretinography demonstrated nonrecordable responses for rods and cones. By the time the child was 7 years and 5 months, the grayish-white spots on the midperipheral retina were clearly obvious (Figure 3B,C). At the age of 8 years and 7 months, diffuse and mottling hypopigmentation with a few small intraretinal pigments of midperipheral retina was visible, displaying a salt-and-pepper appearance in the focal area (Figure 3D–I). In addition, a salt-and-pepper-like fundus change was recorded in F08-II:1 when the child was 7 years old. Bone spicule-like pigmentation in the midperiphery was recorded in one patient (F09-II:6) at the age of 25 years, indicating that fundus changes may worsen with age. Based on a comparison of the fundus images with documented descriptions of fundus changes in the case records, in some patients, a fundus with a relatively normal-looking posterior retina (macular region) would likely be described as normal by some ophthalmologists in general outpatient clinics, despite the presence of significant degeneration in the midperiphery. In summary, the age-related change in the fundus observed in patient F07-II:1 with time may be common in patients with mutations in SPATA7, as discussed in the phenotypic characterization section.

Figure 3

Fundus images showing age-dependent changes in a patient (F07-II:1). The fundus appeared nearly normal in the posterior area, with a few minor grayish-white spots in the midperiphery at the age of 3 years and 5 months (A), and progressed to significant yellowish-white spots in the midperiphery at the age of 7 years and 5 months (B, C). By the age of 8 years and 7 months, diffuse mottling hypopigmentation, with a few small intraretinal pigments, was visible (D–I). Mild salt-and-pepper-like changes were observed (E, H).

Discussion

In this study, the clinical phenotypes of 12 patients from ten families with biallelic mutations in SPATA7 were described, including six families with newly identified c.367C>T (p.Q123*) and c.1083–2A>G mutations. Common age-dependent SPATA7-associated fundus changes were identified based on an analysis of these patients. The data suggested that mutations in SPATA7 may contribute to approximately 2.2% of LCA and 0.4% of HRD cases in Southern Chinese.

Biallelic mutations in SPATA7 have been detected in a total of 78 patients in 60 families including those in this study and described in previous reports [1-33] involving 41 mutations (Table 2 and Appendix 2) based on transcript NM_018418.5. Fifty of the 60 families were described by other groups, and the other ten were from this study cohort. There are no reports of an association of heterozygous mutations in SPATA7 with eye-related diseases. The 41 mutations could be classified as follows: nonsense (eight mutations in 59 alleles), frameshift (15 mutations in 30 alleles), loss of splicing site (ten mutations in 20 alleles), missense (four mutations in six alleles), in-frame deletion (two mutations in two alleles), gross deletion encompassing exon 1 to exon 5 (one mutation in two alleles), and loss of initiation (one mutation in one allele; Table 2). Most of these mutant alleles (93.3%, 112/120) would result in a truncated protein if expressed (Appendix 3), and most of the transcripts would be degraded by nonsense-mediated decay. These mutations were spread into all 12 coding exons, with enrichment in exons 5 and 11 (Figure 4). Four mutations, c.253C>T (p.R85*), c.1183C>T (p.R395*), c.288T>A (p.C96*), and c.322C>T (p.R108*), were relatively common and affected 13, 13, 12, and 11 of the 120 mutant alleles, respectively (Table 2 and Figure 4). All of these mutations are rare in the general population based on existing databases (Table 2).

Figure 4

A schematic diagram of the mutation spectrum frequency in SPATA7. The two horizontal bars represent the coding regions based on two alternative splicing isoforms. Most loss-of-function mutations are listed on the top of the upper bar. A gross deletion involving the first five exons, four missense mutations, and two in-frame deletions are listed under the upper bar. The line height indicates the number of mutated alleles in 60 families.

Of the 78 patients from 60 families with biallelic mutations in SPATA7, the clinical diagnosis was LCA in 41 families (52 patients), RP in ten families (11 patients), juvenile RP in six families (eight patients), CORD in three families (five patients), and myopia in one family (two patients; Appendix 2) [1-33]. Among siblings in one family with a homozygous c.1112T>C (p.I371T) mutation, a 52-year-old brother had RP, and his 48-year-old sister had CORD [24]. A male patient who had been diagnosed with CORD at 9 years old was subsequently diagnosed with RP at 21 years of age [18]. Clinical information on the other three patients diagnosed with CORD was not available. Eight patients diagnosed with juvenile RP had symptoms in the first few years after birth, but six of them were examined they were 7 years old (Appendix 2). Among the 11 patients diagnosed with RP, there were no clinical data available for six patients from six families. For the other five patients with RP, all underwent examination at the age of 25 years old or older: Two had symptoms at the age of 22 or 25 years, one had no symptoms at the age of 26 years, and the symptoms of the other two patients were not documented. Overall, in 28 of the 60 families, the initial symptom recorded was poor vision or nystagmus or both in 24 probands and night blindness in four probands. Age at onset varied from 3 months to no symptoms at age 26 years [5]. Visual acuity was recorded in 33 of 78 patients, and it ranged from no light perception to nearly normal. Four affected children had no light perception, and five patients had best visual acuity better than 0.5 (20/20; Appendix 2).

Fundus color images were available for 15 patients, including six patients in this cohort and ten patients in published literature [1,3,7,12,24,30]. A documentary record of the fundus changes was available for four cases in this cohort and 19 cases in published literature [1-3,5,7,12,14,18,26]. Based on these records, the fundus may appear normal or nearly normal at an early stage (i.e., a few months after birth), with mild narrow arterioles and a relatively well-preserved macular region, with possible mild hypopigmentation in the midperipheral region. However, within a few years, retinal degeneration gradually progresses to widespread atrophy of the RPE, leading to diffuse hypopigmented grayish-white spots or mottled degeneration in the midperipheral region. Minimal intraretinal pigmentation may appear by school age, with a gradual change to salt-and-pepper appearance. Bone spicule-like pigmentation was not observed in patients until at least the age of 25 years but was common in patients aged approximately 40–50 years old. Narrow arterioles, a relatively well-preserved macular region, and widespread RPE atrophy resulting in diffuse mottling hypopigmentation in the midperipheral retinal region may be considered early and common fundus changes specific to SPATA7-associated retinopathy (Figure 2).

Except age-dependent variations in the disease phenotypes, retinal degeneration and accompanied visual function varied markedly among the individuals with mutations in SPATA7, from no symptoms at the age of 26 years old [5] or nearly normal visual acuity [1,3] to early blindness. Significant phenotypic variation was also present in siblings with the same mutations [5,24]. Previous research suggested that a truncation mutation in the C-terminal portion might be associated with milder phenotypes, such as RP [1]. However, several patients with RP did have truncation mutations in the N-terminal region [5,12,15,17]. In addition, two siblings (F10-II:1 and F10-II:2) with a truncation mutation in the N-terminal region showed myopia without obvious fundus changes. It might be that a truncated protein with partial function is produced by a new start codon downstream of the mutational position, or myopia is an earlier feature than significant fundus changes, which would be similar to a case with mutations in RBP3 as described in a previous study [37]. As 93.3% of mutations in SPATA7 are truncation mutations, the location of the truncation mutation does not appear to correlate with the severity of the manifestations in individuals with mutations in SPATA7, and there is no firm evidence for other genotype–phenotype correlations. Thus, by exclusion, this suggests that other unknown factors may act as modifiers to affect the severity of the disease phenotype [3,5]. The latter has been well demonstrated in recent studies on other diseases [38]. Sparing of the inferior portion of the retina close to the optic fissure (Figure 2C–F) also suggests that potential intrinsic factors might modify the expression of the diseases. Understanding the role of such factors in ocular diseases and the underlying mechanisms may have therapeutic potential.

In addition, four missense mutations and two in-frame deletions were detected in nine patients from seven families in previous reports [3,13,19,24-26,30]. Except c.1112T>C (p.I371T), all others were compound heterozygous mutations with another truncation mutation. Two members of one family with a homozygous c.1112T>C mutation had different types of HRD: RP in the brother and CORD in the sister [24]. Among the other seven patients, six patients had LCA, and one had RP. The phenotypes of patients with a missense or in-frame deletion appear to be comparable to those of patients with biallelic truncation mutations. The pathogenicity of missense or in-frame deletions requires additional studies because of two unusual facts: 1) If loss-of-function mutations are the main cause of the disease, the phenotypes associated with missense mutations would be expected to be mild; 2) if both missense and loss-of-function mutations are causative, there should be more missense mutations detected, especially in this era of next-generation sequencing.

In summary, mutations in SPATA7 were identified in ten families with phenotypes varying from LCA (most common) to arRP (less common) to rare early-onset myopia. Two novel truncation mutations were identified, including c.367C>T (p.Q123*) and c.1083–2A>G. The clinical and genetic data in our series, together with a systematic analysis of previously published data, showed that 93.3% (112/120) of mutant alleles in these families were truncation mutations. Narrow arterioles, a relatively well-preserved macular region, and widespread RPE atrophy resulting in diffuse mottling hypopigmentation in the midperipheral retina may be considered early and common fundus changes suggestive of SPATA7-associated retinopathy. A genotype–phenotype analysis revealed the characteristics of pathogenic mutations and the potential problem of missense variants, as well as findings suggestive of potential modifiers. The findings of the present study can provide a valuable reference for clinical gene tests as well as for future studies.