Large Deletions of TSPAN12 Cause Familial Exudative Vitreoretinopathy (FEVR)

Abstract

Purpose: Familial exudative vitreoretinopathy (FEVR) is a rare, hereditary visual disorder. The gene TSPAN12 is associated with autosomal dominant inheritance of FEVR. The prevalence and impact of large deletions/duplications of TSPAN12 on FEVR patients is unknown. To glean better insight of TSPAN12 on FEVR pathology, herein, we describe three FEVR patients with TSPAN12 deletions.

Methods: Thirty-three Korean FEVR patients, who previously screened negative for TSPAN12 mutations, mutations in other FEVR-associated genes such as NDP, FZD4, LRP5, and large deletions and duplications of NDP, FZD4, and LRP5, were selected for TSPAN12 large deletion and duplication analyses. Semiquantitative multiplex PCR for TSPAN12 gene dosage analyses were performed, followed by droplet digital PCR (ddPCR) for validation.

Results: Among the 33 patients, three patients were confirmed to carry large TSPAN12 deletions. Two of them had whole-gene deletions of TSPAN12, and the other patient possessed a deletion of TSPAN12 in exon 4. FEVR severity detected in these patients was not more severe than in a patient with TSPAN12 point mutation.

Conclusions: Regarding previously reported proportions of FEVR-associated genes contributing to the disorder's autosomal dominant inheritance pattern in Korea, we determined that patients with TSPAN12 large deletions were more common than patients with single nucleotide variants in TSPAN12. Evaluating TSPAN12 large deletions and duplications should be considered in FEVR screening and diagnosis as well as in routine genetic workups for FEVR patients.

Familial exudative vitreoretinopathy (FEVR) is a rare, genetically heterogeneous disorder that impairs vision and causes retinal detachment. Autosomal dominant inheritance is the most common form of FEVR and is known to be associated with the FZD4, LRP5, and TSPAN12 genes. Recently, the ZNF408 and KIF11 genes were also implicated in FEVR.^1,2 In a previous report by our group, mutational studies regarding FZD4, LRP5, and TSPAN12 were carried out in 51 unrelated FEVR patients lacking NDP mutations. Among them, mutations with high probabilities of being pathogenic were detected in 18 patients.³ FZD4 mutations accounted for the largest proportion of autosomal dominant FEVR cases (13/18 patients, 72.2%), followed by LRP5 (4/18 patients, 22.2%), and TSPAN12 (1/18 patients, 5.6%) mutations. In rest of the 33 patients, no FZD4, LRP5, and TSPAN12 mutations were detected. 

In our previous report, large deletions/duplications in NDP, FZD4, and LRP5 had been screened via multiplex ligation-dependent probe amplification (MLPA) using the SALSA P285-C1 LRP5-NDP-FZD4 (MRC-Holland, Amsterdam, Netherlands). Because TSPAN12 was not included, we looked for large deletions/duplications of TSPAN12 in the patients with no mutation detected to determine the genetic cause of FEVR. Gene dosage analysis by semiquantitative multiplex PCR and droplet digital PCR (ddPCR) were performed. In this study, we discovered and confirmed three cases of FEVR due to TSPAN12 large deletions. 

Methods

Subjects

Among the 51 FEVR patients from our previous study, 33 patients were selected for large deletion/duplication analyses of TSPAN12. These patients were diagnosed with FEVR between January 2008 and December 2012 at the Seoul National University Children's Hospital (Seoul, Korea). The diagnostic criteria for FEVR composed of three of the following: (1) birth at full term or premature birth with no evidence of retinopathy of prematurity, (2) a presence of peripheral retinal avascular area, (3) variable degree of nonperfusion, vascular leakage, or retinal neovascularization in fluorescein angiography. 

These patients were previously screened and determined to be negative for NDP, FZD4, LRP5, and TSPAN12 mutations. Additionally, these patients lacked large deletions and duplications of NDP, FZD4, and LRP5. Whole blood was collected from patients in EDTA blood collection tubes and the DNA was extracted using Gentra PureGene blood kits (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. Experiments were performed according to the Declaration of Helsinki. 

Gene Dosage Analysis by Semiquantitative Multiplex PCR

Gene dosage of TSPAN12 was assessed by semiquantitative multiplex PCR to detect deletions or duplications. Eight exons of the TSPAN12 gene were amplified with the HBB and B2M genes as endogenous references. Primers were designed for all coding exons, and all forward primers were labeled with 6-fluorescein amidite (6-FAM) (Table 1). Exons 2, 5, and 8 were amplified in the first tube and exons 1, 3, 4, 6, and 7 in the second tube. After 18 cycles of PCR, the products were analyzed using an ABI Prism 3130xl Genetic Analyzer (Applied Biosystems, Carlsbad, CA, USA) with the GeneMapper ID 3.7 software (Applied Biosystems). For each patient sample, PCR amplification was done in triplicate. The peak height ratio of each exon was calculated by dividing the mean peak heights of the TSPAN12 gene with that of the reference genes, HBB and B2M. Afterward, normalized gene dosage was determined using the following equation:

Table 1

 

 

Primer and Probe Sequences for Semiquantitative Multiplex PCR and ddPCR Assays Targeting TSPAN12 Exons

Table 1

 

Primer and Probe Sequences for Semiquantitative Multiplex PCR and ddPCR Assays Targeting TSPAN12 Exons

 

 

 

FEVR patients with normalized values between 0.8 and 1.2 were determined to have two copies of an exon, while patients with normalized values between 0.3 and 0.7 only possessed one copy of an exon.  

Droplet Digital PCR (ddPCR)

To confirm the TSPAN12 large deletions detected in semiquantitative multiplex PCR, ddPCR was performed. Primers and 5′ 6-FAM dye-labeled probes for TSPAN12 exons 3 and 4 were designed (Table 1), and copy number variation of TSPAN12 was evaluated using the QX200 Droplet Digital PCR system (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Digital PCR master mix solutions (20 μL) were prepared from ddPCR 2X Master mix (Bio-Rad Laboratories, Inc.), and ddPCR reactions also consisted of 0.5 μL 20× primers, TaqMan Probe mix (Applied Biosystems), 7 μL nuclease-free water, and 1 μL DNA template. To normalize the patient data, 20× reference assays for the human eukaryotic initiation factor 2C1 (EIF2C1) gene were used and served as endogenous controls, with the probes labeled at the 5′ end with hexachloro-fluorescein. All probes were labeled with Black Hole Quencher-1 at their 3′ ends. Each reaction mixture was then loaded into an 8-channel DG8 droplet generator cartridge (Bio-Rad Laboratories, Inc.). After loading 70 μL of droplet generation oil, emulsified droplets were created using the QX-200 Droplet Generator (Bio-Rad Laboratories, Inc.). Forty microliters of droplet emulsions were then transferred to a 96-well PCR plate using an Eppendorf Xplorer 8-channel electronic pipette (Eppendorf AG, Hamburg, Germany). The loaded 96-well PCR plate was sealed with foil sheets in the PX1 PCR Plate Sealer (Bio-Rad Laboratories, Inc.). Polymerase chain reaction cycling was performed in a DNA Engine Dyad Peltier Thermal Cycler (Bio-Rad Laboratories, Inc.) using the following conditions: denaturation at 95°C for 10 minutes, and 50 cycles of 94°C for 30 seconds, and 54°C for 1 minute (exon 3), and 50°C for 1 minute (exon 4), with the ramp rate set to 2°C/s. After the PCR reactions were complete, fluorescence of each thermally cycled droplet was measured using the QX200 Droplet Reader (Bio-Rad Laboratories, Inc.). The data were visualized and analyzed using the QuantaSoft software v1.7.4 (Bio-Rad Laboratories, Inc.). Thresholds for positive reactions were manually determined based on results from the no template control (NTC) wells. 

Clinical Manifestation According to Genetic Mutations/Anomalies

Clinical characteristics used to determine FEVR staging,⁴ mainly angiographic findings and visual acuity at last follow-up visit, were analyzed. 

Statistical Analysis

For the statistical analysis, visual acuity was converted from Snellen value to logMAR value. Statistical analyses were performed using SPSS software version 21.0 (SPSS, Inc., Chicago, IL, USA). P values less than 0.05 were considered to be statistically significant. 

Results

Semiquantitative Multiplex PCR Detected Large TSPAN12 Deletions

Semiquantitative multiplex PCR for gene dosage analyses of TSPAN12 were performed on 33 patients, and among them, three patients were detected to carry large deletions of TSPAN12. Two patients (Patients #1 and #2) were heterozygous for the TSPAN12 deletion, and another patient was found to have a deletion in exon 4 (Patient #3) (Fig. 1). Additionally, a symptomatic brother of Patient #2 (Patient #2s) was also tested and found to be positive for this deletion. 

Figure 1

 

TSPAN12 deletions in FEVR patients. Semiquantitative multiplex PCR was used to normalize the ratio of TSPAN12 exons 4 and 8 (targets) to HBB genes (endogenous references). Dots denote normalized dosages of each patient's samples. All assays were done in triplicate. Patients #1, #2, and #2s possessed half the ratio in both exons. Patient #3 had a single deletion in exon 4.

ddPCR Validated TSPAN12 Deletions in FEVR Patients

To validate the large deletions detected in semiquantitative multiplex PCR, we designed primer and probe sets for TSPAN12 exons 3 and 4 and performed a ddPCR assay. In the three patients with whole-gene deletions, Patients #1, #2, and #2s, exons 3 and 4 positive droplets were approximately half the number of the reference gene positive droplets (Fig. 2), demonstrating that these patients possessed heterozygous deletions as per measured TSPAN12 copy numbers (Fig. 3). For Patient #3, only exon 4 positive droplet counts were approximately half of the reference gene positive droplets. 

Figure 2

 

TSPAN12 ddPCR analysis of FEVR patients. Data is from seven patient samples, including three normal controls and four patients with TSPAN12 deletions. The y-axis represents the raw number of positive droplets from the target exon and the reference gene. Asterisks denote heterozygous deletion of the target, TSPAN12, as compared with the EIF2C1 reference gene.

Figure 3

 

Determination of TSPAN12 large deletions by ddPCR. Deletion of exon 3 was detected in Patients #1, #2, and #2s. Deletion of exon 4 was detected in Patients #1, #2, #2s, and #3.

Clinical Manifestations of TSPAN12 Deletions in FEVR Patients

Patient #1 was born with a cleft lip and low set, dysmorphic ears. He showed developmental delays with FEVR-features associated in both eyes (Fig. 4). He also suffered from systolic congestive heart failure. Karyotype analysis showed a deletion of chromosome 7, del(7)(q31.3q33). Considering that the location of TSPAN12 is 7q31.31, we suspected that whole-gene deletion was the pathogenic cause for FEVR in this patient. The patient died at 20 months of age. Patient #2, who also had a whole-gene deletion, had a brother, Patient #2s, who shared the same FEVR diagnosis. Patient #2 was originally found to carry a rare TSPAN12 variant, c.484G>A, which seemed to be homozygous via sequencing analysis. This variant was not detected in her brother, which implied that it was not the cause of their shared symptoms and highlighted the possibility of a large deletion in this exon. Gene dosage analysis confirmed such a deletion, demonstrating that the c.484G>A variant detected was actually in a hemizygous state for Patient #2. Both siblings did not show any other systemic symptoms. Lastly, Patient #3, with an exon 4 deletion, presented with esotropia of his left eye when he was 3-months old. Fundoscopy revealed dragging of the optic disc with macular ectopia in the left eye, confirming that the esotropia was secondary to retinopathy. 

Figure 4

 

Color fundus photographs of patients with large TSPAN12 deletion. (A, B) A 4-month-old boy (patient #1 in table 2) with whole-gene deletion of TSPAN12. Peripheral avascular area with incomplete vascular development in the right eye (A) and the left eye (B). (C, D) A 4-month-old girl (patient #2 in Table 2) with whole-gene deletion of TSPAN12. (C) Peripheral avascular area with incomplete vascular development in the right eye (D). Fibrovascular stalk to inferotemporal direction is concomitant with macular ectopia in the left eye. (E, F) A 20-month-old boy (patient #3 in Table 2) with exon 4 deletion in TSPAN12 gene. (E) Extraretinal fibrovascular membrane is present along with temporal avascular border in the right eye. (F) Fibrovascular stalk to temporal direction is concomitant with extramacular retinal detachment in peripheral retina of the left eye.

Table 2

 

Clinical Characteristics of Patients With TSPAN12 Pathogenic Variants or Large Deletions

 

Upon comparing a patient with a TSPAN12 point mutation (Patient #4) and our patients with large TSPAN12 deletions, patients with the large deletions displayed relatively milder phenotypes (Table 2). However, it is important to note that data were limited as only one patient carrying a point mutation was available in this study. 

Genotypes and Phenotypes of FEVR

Patients #1, #2, and #3, along with Patient #2s, were added to the total pool of genetically confirmed FEVR patients described in our previous study³ to assess the scope of FEVR phenotypes. Visual acuity and FEVR stage at last follow-up were analyzed to see whether there were significant differences according to the genes involved (FZD4, LRP5, or TSPAN12). Indeed, there were significant differences for either FEVR stage (P = 0.021, 1-way ANOVA) or visual acuity (P = 0.019, 1-way ANOVA) at last follow-up according to the particular gene involved. However, post-hoc analyses using the Tukey test showed that, respectively, FEVR stage was significantly lower (P = 0.015) and visual acuity significantly higher in patients with the FZD4 mutation than in patients with the LRP5 mutation (P = 0.004). There was no significant difference in FEVR stage or visual acuity between FZD4- and TSPAN12-mutation patient groups nor between LRP5- and TSPAN12-mutation patient groups. 

Discussion

To our knowledge, there has not been a report on FEVR patients with large deletions of TSPAN12. There are several published reports highlighting patients with chromosome 7 deletions in the 7q31 region, but this is the first study to report FEVR patients genotypically confirmed for such pathogenic large deletions. In a previous report of a patient with a 7q31.2-q32.2⁵ deletion, the patient had speech and language disorders due to the deletion of the FOXP2 gene, which maps to 7q31.1 and is implicated in speech and language impairment. Because TSPAN12 also resides in this deletion range, we expected the patient to have FEVR-related features. However, this study contained a limited description of “mottling of retinal pigmentation” for this patient. To note, a mild FEVR phenotype may have gone undetected because the patient had no visual impairment yet mottled retinal pigmentation. Another report of a patient confirmed with a 5.4-Mb deletion in the 7q31.31 region had congenital nystagmus due to bilateral persistent hyperplastic primary vitreous (PHPV), which phenotypically overlaps with FEVR.⁶ However, this patient was determined to not have FEVR. He had dysmorphic facial features such as low set ears and developmental delays, which led to an autism spectrum disorder (ASD) diagnosis. Haploinsufficiency of CADPS2 and TSPAN12 genes in the 7q31.31 region has been shown to contribute to ASD and PHPV, respectively. The patient in our study with the 7q31.3-q33 deletion shared similar phenotypes with this previously reported case,⁶ such as dysmorphic facial features and developmental delays. However, ASD-related features were not significant in our patient because he suffered from more complicated symptoms such as heart failure and did not survive long enough to fully dissect the scope of his symptoms. 

Regarding the proportion of each gene's contribution to FEVR in the Korean population,³ pathogenic genetic anomalies were detected in 21 patients among our original 51 patient cohort, with three more patients added to this current study due to detecting large TSPAN12 deletions. This resulted in 41.2% of patients being genetically confirmed for FEVR. FZD4 mutations still accounted for the largest proportion (13/51 patients, 25.5%), followed by LRP5 (4/51 patients, 7.8%), and TSPAN12 mutations (4/51 patients, 7.8%). In our cohort, the number of FEVR patients with TSPAN12 mutations was the same as those patients with LRP5 mutations. This was probably due to less LRP5 variants detected than expected. In another study, which considered the disease-associated variants of FEVR-related genes in a single cohort,⁷ 92 FEVR patients were screened for five genes associated with FEVR using next-generation sequencing (NGS). LRP5 was the most frequently mutated gene (19.6% of patients) in this cohort, followed by FZD4 (15.2%), TSPAN12 (8.7%), NDP (6.5%), and ZNF408 (1.1%). Though this study was different from our study in that minor portion of the patients with NDP mutation were included and large deletion/duplication of the related genes were not analyzed, the proportion of TSPAN12 variants being around 8% was almost the same as our cohort. However, the proportions of FZD4 and LRP5 variants were respectively higher and lower in our cohort than in the cohort of the abovementioned study. 

Interestingly, in our study, the proportion of large deletions in TSPAN12 was higher than that of point mutations, which had not been observed in other FEVR-related genes. Among FZD4-mutation positive patients reported by our group, only one patient was confirmed for a whole-gene deletion while 12 had single nucleotide variants or small deletions. Even in cases with positive NDP mutations tested by our group, large deletion cases were less frequently detected than point mutations (data not shown). It would be interesting to find out whether this high contribution of large deletions in TSPAN12 is also observed in other populations as well. Additionally, the phenotypes of the patients with large deletions were no more severe than those of a patient with a point mutation, although large deletions are generally considered more harmful than point mutations. Though not statistically significant, the patient with a TSPAN12 point mutation (Patient #4) showed more severe symptoms in both eyes as compared with our patients with large TSPAN12 deletions in this study. Upon reviewing other previously reported cases,^8–10 there were patients with TSPAN12 missense mutations who possessed high stage FEVR with retinal detachments. Nevertheless, it is hard to conclude whether these TSPAN12 missense mutations showed dominant negative effects due to limited description of the fundus examination results. The difference in phenotypes according to the mutation type should be further considered in more large-scale studies. 

To detect these large deletions/duplications of TSPAN12, we performed a semiquantitative multiplex PCR along with a ddPCR assay, which is known for its improved sensitivity and resolution. Traditional methods for large deletion detection, such as real-time quantitative PCR (qPCR) or MLPA, have several limitations. Quantitative PCR is known for its imprecise quantitation and needs of standardized materials, while designing and optimizing a new panel for MLPA can be labor intensive and time consuming. The use and application of ddPCR in large deletion detection and copy number variation analyses have been reported in several studies.^11–13 The use of ddPCR in our study efficiently enhanced our resolution of TSPAN12 dynamics in FEVR patients, and the use of such an assay may be a viable choice in better guiding FEVR diagnosis and estimating prevalence. 

In conclusion, we detected TSPAN12 large deletions more frequently than TSPAN12 point mutations among our cohort of FEVR patients. Large deletion/duplication analyses of TSPAN12 should be taken into consideration and evaluated in routine genetic workups for FEVR patients as gene dosage abnormalities and copy number variation of FEVR-related genes may be more prevalent than originally suspected. 

Acknowledgments

Disclosure: S.H. Seo, None; M.J. Kim, None; S.W. Park, None; J.H. Kim, None; Y.S. Yu, None; J.Y. Song, None; S.I. Cho, None; J.H. Ahn, None; Y.H. Oh, None; J.-S. Lee, None; S. Lee, None; M.-W. Seong, None; S.S. Park, None; J.Y. Kim, None