Abstract

Familial exudative vitreoretinopathy (FEVR) belongs to a group of genetically and clinically heterogeneous disorders in retinal vascular development. To date, in approximately 50% of patients with FEVR, pathogenic mutations have been detected in FZD4, LRP5, TSPAN12, NDP and ZNF408. In this study, we identified two heterozygous frameshift mutations in RCBTB1 from three Taiwanese cases through exome sequencing. In patient-derived lymphoblastoid cell lines (LCLs), the protein level of RCBTB1 is approximately half that of unaffected control LCLs, which is indicative of a haploinsufficiency mechanism. By employing transient transfection and reporter assays for the transcriptional activity of β-catenin, we demonstrated that RCBTB1 participates in the Norrin/FZD4 signaling pathway and that knockdown of RCBTB1 by shRNA significantly reduced nuclear accumulation of β-catenin under Norrin and Wnt3a treatments. Furthermore, transgenic fli1:EGFP zebrafish with rcbtb1 knockdown exhibited anomalies in intersegmental and intraocular vessels. These results strongly support that reduced RCBTB1 expression may lead to defects in angiogenesis through the Norrin-dependent Wnt pathway, and that RCBTB1 is a putative genetic cause of vitreoretinopathies.

Introduction

First described in 1969, familial exudative vitreoretinopathy (FEVR; [OMIM 133780, 305390, 605750, 601813, 613310, and 616468]) is a rare developmental disorder of the retinal blood vessels characterized by incomplete peripheral vascularization (1,2). FEVR has two prominent clinical features: intrafamilial variability (3–8), in which individuals carrying the same mutation may present with a wide range of severities, from asymptomatic to the most complicated retinal detachments (RDs); and disease asymmetry, with presentations of severity varying between the two eyes (9,10). Coats disease (OMIM 3 00 216), also known as sporadic retinal telangiectasis, is a rare disorder with an estimated incidence of 0.09 in 100 000 in UK (11). In addition to predominantly affecting males and typically affecting only one eye, Coats disease is characterized by telangiectasia of the retinal vessels, subretinal lipid exudation, macular edema and RD resulting from defects in retinal vascular development (12–14). Both FEVR and Coats disease, in addition to other congenital vitreoretinopathies such as persistent hyperplastic primary vitreous (OMIM 611308) and Norrie disease (OMIM 310600), share similar fundus features, including avascularization of the peripheral retina and subretinal exudation. Therefore, they have been proposed to be in a spectrum of vitreoretinopathies (4,6,9,15–18).

Various Mendelian inheritance patterns of FEVR have been documented (15,19–21), mainly autosomal dominant with reduced penetrance estimated to be less than 50% (22) and thus may appear as sporadic cases. To date, mutations in five genes, namely frizzled 4 (FZD4 [OMIM 6 04 579]), low-density lipoprotein receptor-related protein 5 (LRP5 [OMIM 6 03 506]), tetraspanin-12 (TSPAN12 [OMIM 6 13 138]), Norrin (NDP [OMIM 3 00 658]) and zinc finger protein 408 (ZNF408 [OMIM 6 16 454]), have been identified to be highly associated with FEVR (6,15,17,23–26). All of these genes have been found to play a role in normal vascular development in animal models, and the first four genes are crucial components of the Norrin-induced FZD4/β-catenin signaling pathway (23,27). However, mutations in these genes account for only approximately 50% of cases of FEVR (6,17,28) and other vitreoretinopathies such as Coats disease and Norrie disease have also been demonstrated sharing the same genetic factors (NDP and/or FZD4) (29,30). Although Coats disease usually appeared sporadic, there are genetic evidences that Coats disease is linked to the NDP/wnt signaling pathway, one is a somatic mutation in NDP identified in a boy (12) and the other is a germline mutation also in NDP in an affected female who gave birth to a boy diagnosed with Norrie disease (31). These findings have suggested that vitreoretinopathies may result from mutations from the same group of genes and that additional underlying genetic factors for these disorders may also exist and function in early retinal vascular development.

Results

Family recruitment and mutation screening for known FEVR-related genes

Fifteen patients diagnosed with FEVR or Coats disease and their family members were recruited from nine Taiwanese families with informed consent (Fig. 1; Supplementary Material, Fig. S1). We performed direct sequencing of the exons and exon–intron boundaries of the five candidate genes, i.e. FDZ4, LRP5, NDP, TSPAN12 and ZNF408, in the affected participants. We found two previously identified mutations in FZD4 in two families, specifically, c.313A>G in family E1 and c.1282_1285delGACA in family E10; and detected a novel single-nucleotide variant (NM_000266.3:c.-77A>G) at the 5′UTR of NDP in family E9. This was not detected in 320 X chromosomes from ethnically matched population controls. To evaluate whether this 5′UTR mutation affects the NDP expression level, we generated constructs with a wild-type (WT) or mutant 5′UTR upstream of the firefly luciferase coding sequence. According to normalized luciferase activities, we estimated the expression level in the mutant 5′UTR to be half that of the WT 5′UTR (Supplementary Material, Fig. S2). To assess the effect of this mutation on the Norrin-dependent Wnt signaling pathway, we cotransfected the NDP expression constructs with a WT or mutant 5′UTR at two different doses with FZD4, LRP5 expression constructs and a reporter with TCF/LEF-binding sites into ARPE19 cells. The results revealed a dose-dependent reduction of Wnt signaling. A significant reduction of the reporter was evident only in transfections with a higher dose of NDP constructs (150 ng), but not in the group with a lower construct dose (75 ng) (Supplementary Material, Fig. S2). Because of limited information regarding the physiological concentration of Norrin during retinal angiogenesis, we could not reach a definitive conclusion on whether this variant was the major genetic factor and a sufficient cause of FEVR in family E9.

Figure 1.

Pedigrees of two families with heterozygous mutations in RCBTB1 and the clinical features of affected members. (A) Pedigree structure and partial chromatograms of Sanger confirmation sequencing for families E5 and E9. Because of the reduced penetrance and phenotypic variability in FEVR, the mothers in family E9 were assumed to be nonpenetrant carriers. (B) Fundi of three members in family E9. Both eyes were affected in individuals III-2 and III-4. In individual III-2, disc-dragging with macular ectopia in the right eye and total RD in the left eye were evident. The fundi of individual III-4 exhibited a fibrovascular stalk (arrowhead) emanating from the disc to the periphery in the right eye, and disc-dragging with macular ectopia in the left eye. The fundi of carrier II-4 appeared unremarkable. (C) The fundus of individual II-1 of family E5 exhibited marked subretinal lipid exudates (arrowhead), with total exudative detachment in the right eye. The left eye was unremarkable. The diagnosis was Coats disease on the right eye. The retina was reattached after cryopexy, pars plana vitrectomy and removal of the subretinal fibrous cord and encircling buckle. However, a massive lipid exudate remained at the posterior pole. OD, right eye; OS, left eye. *Reference sequences for RCBTB1: NM_018191.3, NC_000013.10.

Figure 1.

Pedigrees of two families with heterozygous mutations in RCBTB1 and the clinical features of affected members. (A) Pedigree structure and partial chromatograms of Sanger confirmation sequencing for families E5 and E9. Because of the reduced penetrance and phenotypic variability in FEVR, the mothers in family E9 were assumed to be nonpenetrant carriers. (B) Fundi of three members in family E9. Both eyes were affected in individuals III-2 and III-4. In individual III-2, disc-dragging with macular ectopia in the right eye and total RD in the left eye were evident. The fundi of individual III-4 exhibited a fibrovascular stalk (arrowhead) emanating from the disc to the periphery in the right eye, and disc-dragging with macular ectopia in the left eye. The fundi of carrier II-4 appeared unremarkable. (C) The fundus of individual II-1 of family E5 exhibited marked subretinal lipid exudates (arrowhead), with total exudative detachment in the right eye. The left eye was unremarkable. The diagnosis was Coats disease on the right eye. The retina was reattached after cryopexy, pars plana vitrectomy and removal of the subretinal fibrous cord and encircling buckle. However, a massive lipid exudate remained at the posterior pole. OD, right eye; OS, left eye. *Reference sequences for RCBTB1: NM_018191.3, NC_000013.10.

Exome sequencing and identification of potentially pathogenic variants

To detect disease-associated novel variants, we performed whole-exome sequencing in affected individuals from families E5 (II-1), E6 (II-1) and E9 (III-2 and III-4). For each case, the average coverage was at least 90-fold, with more than 90% of the targeted exons covered at least 20 times. Because of the availability of a pair of third-degree relatives (i.e. III-2 and III-4 in E9), we first analyzed their exome data. After conducting variant filtering to fulfill an autosomal dominant or X-linked inheritance and to exclude known variants in dbSNP 138, we identified 130 candidate variants in both III-2 and III-4 in family E9 (Supplementary Material, Table S1), 27 of which passed confirmation based on information provided by the Ensembl Genome Browser. To further narrow down the candidate gene list, we included the exome data of II-1 in E5 and II-1 in E6, and applied the modified overlap strategy (32) by focusing on novel variants in the same gene in at least two of the three families. We identified two novel variants with predicted damaging effects in RCBTB1 in families E5 and E9. In family E5, we detected a single-nucleotide deletion (NM_018191.3:c.707delA) in RCBTB1 for individual II-1, who had been diagnosed with Coats disease, and in his unaffected father (individual I-1) (Fig. 1A). In family E9, a canonical splice-site variant (NM_018191.3:c.1172+1G>A, NC_000013.10:g.50118872C>T) was found in two cousins diagnosed with FEVR and their asymptomatic mothers (Fig. 1A). Neither of these two variants was detected in 186 unrelated, ethnic-matched population controls and was not present as rare variants in the Phase 3 1000 Genomes variants (33). The clinical characteristics of these two families are summarized in Table 1 and Figure 1. We also performed direct sequencing of all the exons and exon–intron boundaries of RCBTB1 in the affected participants of the other recruited families. The results revealed only common single-nucleotide polymorphisms (SNPs) rather than the novel damaging SNVs (Supplementary Material, Table S2).

Table 1.

Clinical characterization of subjects

Family ID E9
 
E5
 
Individual ID III-2 III-4 II-2 II-4 I-1 I-2 II-2 
Sex 
Mutation NDP: c.-77A>G
RCBTB1: c.1172+1G>A (p.Glu349Glyfs*17) 
RCBTB1: c.707delA (p.Asn236Thrfs*11) – 
Onset age Congenital Congenital – – 3 y/o – – 
Clinical features 
 Anterior segment 
  OD Unremarkable Status post intraocular lens implantation Unremarkable Unremarkable Aphakia ND ND 
  OS Cornea opacity Unremarkable Unremarkable 
 Fundoscopic exam 
  OD Macular ectopia Fibrovascular stalk Unremarkable Unremarkable Exudative RD ND ND 
  OS Traction RD and retinal break Macular ectopia Unremarkable 
Family ID E9
 
E5
 
Individual ID III-2 III-4 II-2 II-4 I-1 I-2 II-2 
Sex 
Mutation NDP: c.-77A>G
RCBTB1: c.1172+1G>A (p.Glu349Glyfs*17) 
RCBTB1: c.707delA (p.Asn236Thrfs*11) – 
Onset age Congenital Congenital – – 3 y/o – – 
Clinical features 
 Anterior segment 
  OD Unremarkable Status post intraocular lens implantation Unremarkable Unremarkable Aphakia ND ND 
  OS Cornea opacity Unremarkable Unremarkable 
 Fundoscopic exam 
  OD Macular ectopia Fibrovascular stalk Unremarkable Unremarkable Exudative RD ND ND 
  OS Traction RD and retinal break Macular ectopia Unremarkable 

Reference sequences for NDP and RTBCB1 are NM_000266.3 and NM_018191.3, respectively.

+, Present; −, absent; F, female; ID, intellectual disability; M, male; ND, not documented; OD, right eye; OS, left eye.

Effect of novel variants on gene products

The RCBTB1 gene contains 13 exons encoding a 531-amino-acid protein with two predicted domains, namely RCC1 at the N-terminus and the BTB domain at the C-terminus (Fig. 2A), and it is ubiquitously expressed in various tissues (Supplementary Material, Fig. S3). The single-nucleotide deletion (c.707delA) in family E5 is located in exon 7, whereas the novel variant (c.1172+1G>A) in family E9 was located at the splicing donor site of intron 10. To verify the effect of the splice-site variant on the RCBTB1 transcript, we performed RT-PCR followed by amplicon sequencing, where we used the leukocyte RNA from the participants of family E9. An aberrant transcript with a skipped exon 10 was found, and a frameshift with a truncated protein was predicted (NP_060661.3:p.Glu349Glyfs*17) (Fig. 2A and B). Immunoblot analysis of RCBTB1 using total proteins from LCLs from carriers of family E9 revealed an expression level that was approximately half that of control LCLs without predicted truncated proteins (Fig. 2C), suggesting the haploinsufficiency of this variant on RCBTB1.

Figure 2.

Effect of splice-site mutation on mRNA and protein expression. (A) RCBTB1 gene structure and predicted domains. The UTRs and coding sequences are indicated in the clear and solid boxes, respectively, and are labeled with exon numbers. Domains in RCBTB1 (UniProt Q8NDN9) include RCC1 (blue) and BTB (green). The nucleotide positions and predicted changes in amino acids of the two heterozygous mutations identified in families E5 and E9 are indicated below the gene structure. The predicted additional amino acid residues located downstream of the frameshift mutations are in red. RCC1, the regulator of chromosome condensation 1; BTB, BR-C, ttk and bab. (B) Aberrant RTBCB1 transcripts with skipped exon 10 were identified through RT-PCR by using RNA extracted from peripheral blood leukocytes from the members of family E9. Direct sequencing of the RT-PCR products revealed the direct joining of exons 9 and 11 of RCBTB1, and a frameshift mutation with premature termination (p.Glu349Glyfs*17) was predicted. Messenger RNA without reverse transcription was used as the template for the negative control, RT(–). (C) Immunoblot analysis of RCBTB1 (apparent mobility at ∼55 kDa) in LCLs derived from the members of family E9 or baseline controls. The predicted truncated form (∼38 kDa) was not detected, and approximately half the expression level of WT RCBTB1 was noted in all heterozygotes from family E9.

Figure 2.

Effect of splice-site mutation on mRNA and protein expression. (A) RCBTB1 gene structure and predicted domains. The UTRs and coding sequences are indicated in the clear and solid boxes, respectively, and are labeled with exon numbers. Domains in RCBTB1 (UniProt Q8NDN9) include RCC1 (blue) and BTB (green). The nucleotide positions and predicted changes in amino acids of the two heterozygous mutations identified in families E5 and E9 are indicated below the gene structure. The predicted additional amino acid residues located downstream of the frameshift mutations are in red. RCC1, the regulator of chromosome condensation 1; BTB, BR-C, ttk and bab. (B) Aberrant RTBCB1 transcripts with skipped exon 10 were identified through RT-PCR by using RNA extracted from peripheral blood leukocytes from the members of family E9. Direct sequencing of the RT-PCR products revealed the direct joining of exons 9 and 11 of RCBTB1, and a frameshift mutation with premature termination (p.Glu349Glyfs*17) was predicted. Messenger RNA without reverse transcription was used as the template for the negative control, RT(–). (C) Immunoblot analysis of RCBTB1 (apparent mobility at ∼55 kDa) in LCLs derived from the members of family E9 or baseline controls. The predicted truncated form (∼38 kDa) was not detected, and approximately half the expression level of WT RCBTB1 was noted in all heterozygotes from family E9.

Functional characterization of rcbtb1 in zebrafish

To examine the role of RCBTB1 in angiogenesis in vivo, we conducted morpholino (MO)-mediated knockdown of rcbtb1 in the Tg(fli1:EGFP) zebrafish line (34) to investigate its effects on embryonic blood vessel development. Compared with those injected with control MOs containing five mismatched nucleotides, morphants with rcbtb1 translation-blocking MOs revealed comparable gross development except some anomalies in the patterns of the vasculatures in the intersegmental vessels (ISVs) and the intraocular vessels (IOVs). Some ISVs exhibited a smaller average area and even appeared as a streak of cells without lumen. These abnormalities resembled truncated ISVs at 3 days post fertilization (dpf) and could be partially rescued through coinjection with human RCBTB1 mRNA (Fig. 3A and B). In a similar manner, injections with rcbtb1 splicing MOs resulted in dose-dependent aberrant splicing of rcbtb1 (Supplementary Material, Fig. S4) as well as thinner ISVs, suggesting that rcbtb1 plays a role in early embryonic angiogenesis in zebrafish. To accurately model clinical phenotypes of FEVR, we further investigated the effect of rcbtb1 knockdown on IOVs in zebrafish using the ndp translation-blocking MOs as a positive control. As expected, injections with ndp MOs resulted in narrow IOVs and an increased area of avascularization at 4 dpf, without inducing obvious effects on the ISVs (Fig. 3C–E). We categorized the IOV phenotypes of the ndp morphants into three classes according to the proportion of the avascular area: class I was the baseline; class II indicated thinner IOVs and an avascular area range between 25 and 50% and class III represented thinner IOVs and an avascular area larger than 50% (Fig. 3C). Compared with the IOV phenotypes in the ndp morphants, a milder effect was found in the rcbtb1 morphants, specifically 79 and 87% in classes II to III for rcbtb1 and ndp morphant groups, respectively. To determine the genetic interactions between Norrin signaling and rcbtb1 during IOV development in vivo, we investigated the effect of combined MOs targeting ndp and rcbtb1. Coinjections of half-dose ndp MOs (3 ng) with rcbtb1 mismatch-control MOs (3 ng) resulted in mild IOV anomalies (25% of class II, but 0% of class III), whereas coinjections of half-dose ndp MOs (3 ng) with rcbtb1 MOs (3 ng) severely impaired the IOV structure, and resembled those observed in the working-dose ndp morphants (Fig. 3C and D, respectively). These findings suggest that rcbtb1 is involved in the Norrin-dependent IOV development in zebrafish.

Figure 3.

Functional characterization of RCBTB1 through morpholino (MO)-mediated gene knockdown in zebrafish embryos. (A) Morphant rcbtb1 zebrafish showed thinner and irregular ISVs. Lateral views of trunk vasculatures (dorsal side facing up) in fli1:EGFP transgenic larvae at 3 dpf injected with the control MO, rcbtb1-ATG MO, or combinations of rcbtb1-ATG MO and mRNA encoding human WT RCBTB1. ISVs project from the dorsal aorta toward the dorsal longitudinal vessel at the top. (B) Quantitative analysis of the average areas of eight ISVs counted posteriorly from the urogenital pole in different morphant groups normalized to the average of no-injection controls. The values are represented as the mean ± standard error of the mean (SEM). Asterisks indicate a significant difference, with a P of <0.05. (C) Morphant rcbtb1 zebrafish exhibited moderate defects in IOV development. The images depict the IOVs of fli1:EGFP transgenic zebrafish larvae at 4 dpf. Phenotypes of the IOVs were categorized into three classes: baseline (class I), moderately affected (class II) and severely affected (class III). The IOVs in class I had a normal width and a regular radial configuration, with an avascular area occupied <25% of the total space. However, narrow and less regular radial vasculatures with an avascular area ranging between 25 and 50%, or >50%, were classified into classes II and III, respectively. (D) Semiquantitative analyses of IOV phenotypes in morphants injected with the control MO (6 ng), rcbtb1-ATG MO (6 ng), ndp ATG MO (6 ng), or combinations of 3 ng of the rcbtb1 control or ATG MO with 3 ng of ndp ATG MO. At least 40 larvae per condition from three independent experiments were evaluated. (E) Comparison of IOV and ISV phenotypes in fli1:EGFP transgenic larvae at 4 dpf, with the injection of 6 ng ndp ATG MO or rcbtb1-ATG MO. Both morphant groups had class II IOV phenotypes, but the thinner and irregular ISV phenotypes (white arrowhead) were noted only in rcbtb1-ATG morphants.

Figure 3.

Functional characterization of RCBTB1 through morpholino (MO)-mediated gene knockdown in zebrafish embryos. (A) Morphant rcbtb1 zebrafish showed thinner and irregular ISVs. Lateral views of trunk vasculatures (dorsal side facing up) in fli1:EGFP transgenic larvae at 3 dpf injected with the control MO, rcbtb1-ATG MO, or combinations of rcbtb1-ATG MO and mRNA encoding human WT RCBTB1. ISVs project from the dorsal aorta toward the dorsal longitudinal vessel at the top. (B) Quantitative analysis of the average areas of eight ISVs counted posteriorly from the urogenital pole in different morphant groups normalized to the average of no-injection controls. The values are represented as the mean ± standard error of the mean (SEM). Asterisks indicate a significant difference, with a P of <0.05. (C) Morphant rcbtb1 zebrafish exhibited moderate defects in IOV development. The images depict the IOVs of fli1:EGFP transgenic zebrafish larvae at 4 dpf. Phenotypes of the IOVs were categorized into three classes: baseline (class I), moderately affected (class II) and severely affected (class III). The IOVs in class I had a normal width and a regular radial configuration, with an avascular area occupied <25% of the total space. However, narrow and less regular radial vasculatures with an avascular area ranging between 25 and 50%, or >50%, were classified into classes II and III, respectively. (D) Semiquantitative analyses of IOV phenotypes in morphants injected with the control MO (6 ng), rcbtb1-ATG MO (6 ng), ndp ATG MO (6 ng), or combinations of 3 ng of the rcbtb1 control or ATG MO with 3 ng of ndp ATG MO. At least 40 larvae per condition from three independent experiments were evaluated. (E) Comparison of IOV and ISV phenotypes in fli1:EGFP transgenic larvae at 4 dpf, with the injection of 6 ng ndp ATG MO or rcbtb1-ATG MO. Both morphant groups had class II IOV phenotypes, but the thinner and irregular ISV phenotypes (white arrowhead) were noted only in rcbtb1-ATG morphants.

RCBTB1 participates in the Norrin/β-catenin signaling pathway by regulating the nuclear accumulation of β-catenin

We determined the role of RCBTB1 in the Norrin/β-catenin signaling by performing TCF/LEF reporter assays in ARPE19 cells. Cells were cotransfected with plasmids expressing FZD4 and LRP5, reporter constructs and one of the two RCBTB1-targeting shRNA plasmids (shRCBTB1-1 or shRCBTB1-2) with different knockdown efficiency, or the pLKO.1 vector as a sham control. At 24 h after transfection, the cells were cultured in the presence of 0, 62.5 or 125 ng/mL recombinant human Norrin (rhNorrin) for another 24 h. Compared with no-ligand and mutant reporter controls (M51), both concentrations of rhNorrin can effectively activate TCF/LEF-driven expression of luciferase in a dose-dependent manner (Fig. 4A). In the RCBTB1 knockdown groups, we estimated the Norrin-induced activation to be approximately 50 and 33% of that of the vector control under high and low doses of Norrin, respectively, and the reduction of Norrin-induced signaling was proportional to the knockdown efficiency of the shRNA plasmids (Fig. 4A; Supplementary Material, Fig. S5). A more pronounced reduction of Norrin-induced signaling upon RCBTB1 knockdown at a lower dose of Norrin may reflect a synergistic effect between NDP and RCBTB1, indicating that RCBTB1 may be involved in the Norrin/β-catenin signaling pathway. To further investigate the ligand-specific response of RCBTB1 in Norrin-dependent signaling, cells cotransfected with FZD4, LRP5, reporter constructs and RCBTB1-targeting shRNA plasmids or the pLKO.1 vector were incubated in a medium with 100 ng/mL Wnt3a, and we also found a reduction in reporter activity for the RCBTB1 knockdown group (Fig. 4A). These results suggest that RCBTB1 may be a common downstream component of the FZD4/LRP5 signaling pathway. Because β-catenin is tightly regulated in both cytoplasm and nucleus, the effect of RCBTB1 on β-catenin in different cellular compartments in ARPE19 cells with or without Norrin/Wnt3a activation was analyzed and quantified by subcellular fractionation and immunoblot analyses. The quality of the fractionation was assessed using Lamin B1 and GAPDH as markers for nuclear and cytosolic fractions, respectively (Fig. 4B). In APRE19 cells transfected with FZD4/LRP5, the nonphosphorylated β-catenin was upregulated significantly in the nuclear fraction under Norrin/Wnt3a activation for 4 h. However, the increase of nuclear β-catenin under Norrin/Wnt3a activation was abolished in RCBTB1 knockdown groups without affecting the level of β-catenin in the cytosolic fractions (Fig. 4B and C). This effect was similarly revealed at the single-cell level in immunofluorescence analysis using an anti-β-catenin antibody in ARPE19 cells transfected with FZD4, LRP5 and either RCBTB1-targeting shRNA or the pLKO.1 vector under serum-free media with 250 ng/mL rhNorrin or Wnt3a for 4 h. The cells in RCBTB1 knockdown groups revealed significantly reduced nuclear accumulation of β-catenin under treatments with either Norrin or Wnt3a (Fig. 4D), suggesting that RCBTB1 is located upstream of β-catenin and regulates its nuclear accumulation in the FZD/β-catenin signaling pathway.

Figure 4.

Knockdown of RCBTB1 significantly reduced the activation of the Norrin- or Wnt3a-mediated Wnt/β-catenin pathway by reducing the nuclear accumulation of β-catenin. (A) Luciferase assays were conducted using ARPE19 cells. The cells were cotransfected with shRNA targeting RCBTB1 (shRCBTB1-1 and shRCBTB1-2) or the vector control (pLKO.1), with expression vectors encoded for human FZD4, LRP5 and pmirGlo-derived dual luciferase constructs with seven human TCF/LEF-binding sequences (M50) or mutant TCF/LEF-binding sequences (M51) at the firefly luciferase promoter. The transfected cells were treated with 0, 62.5, 125 ng/mL rhNorrin or 100 ng/mL Wnt3a ligands. Compared with no ligand control and mutant reporter controls (M51), both rhNorrin and Wnt3a can activate TCF/LEF-driven expression of luciferase and the activation was significantly reduced by the expression of RCBTB1-targeting shRNA. The RLA is the ratio of Firefly and Renilla luciferase activities and is presented as the mean ± SEM (*P < 0.05, ns, not significant). (B) Subcellular fractionation and immunoblot analysis of ARPE19 cells cotransfected with expression vectors encoding human FZD4 and LRP5 with or without RCBTB1-targeting shRNA (shRCBTB1-1), in the presence of 250 ng/mL rhNorrin or Wnt3a for 4 h. The nonphosphorylated β-catenin was upregulated significantly in the nuclear fraction in control groups, which was abolished in RCBTB1 knockdown groups without affecting the level of β-catenin in the cytosolic fractions. (C) Quantification of (B) in at least triplicate. The non-P-β-catenin levels in each groups were normalized to control groups without Norrin or Wnt3a treatment and presented as the mean ± SEM (*P < 0.05, ns, not significant). (D) Immunofluorescence with anti-β-catenin (green) in ARPE19 cells cotransfected with expression vectors encoding human FZD4 and LRP5, shRNA targeting RCBTB1 (shRCBTB1-1) or the vector-only control (pLKO.1) in the presence of 250 ng/mL rhNorrin or Wnt3a for 4 h. Significant nuclear accumulation of β-catenin was noted in the vector control group rather than in the RCBTB1 knockdown group. The arrowheads indicate the locations of nuclei in cells.

Figure 4.

Knockdown of RCBTB1 significantly reduced the activation of the Norrin- or Wnt3a-mediated Wnt/β-catenin pathway by reducing the nuclear accumulation of β-catenin. (A) Luciferase assays were conducted using ARPE19 cells. The cells were cotransfected with shRNA targeting RCBTB1 (shRCBTB1-1 and shRCBTB1-2) or the vector control (pLKO.1), with expression vectors encoded for human FZD4, LRP5 and pmirGlo-derived dual luciferase constructs with seven human TCF/LEF-binding sequences (M50) or mutant TCF/LEF-binding sequences (M51) at the firefly luciferase promoter. The transfected cells were treated with 0, 62.5, 125 ng/mL rhNorrin or 100 ng/mL Wnt3a ligands. Compared with no ligand control and mutant reporter controls (M51), both rhNorrin and Wnt3a can activate TCF/LEF-driven expression of luciferase and the activation was significantly reduced by the expression of RCBTB1-targeting shRNA. The RLA is the ratio of Firefly and Renilla luciferase activities and is presented as the mean ± SEM (*P < 0.05, ns, not significant). (B) Subcellular fractionation and immunoblot analysis of ARPE19 cells cotransfected with expression vectors encoding human FZD4 and LRP5 with or without RCBTB1-targeting shRNA (shRCBTB1-1), in the presence of 250 ng/mL rhNorrin or Wnt3a for 4 h. The nonphosphorylated β-catenin was upregulated significantly in the nuclear fraction in control groups, which was abolished in RCBTB1 knockdown groups without affecting the level of β-catenin in the cytosolic fractions. (C) Quantification of (B) in at least triplicate. The non-P-β-catenin levels in each groups were normalized to control groups without Norrin or Wnt3a treatment and presented as the mean ± SEM (*P < 0.05, ns, not significant). (D) Immunofluorescence with anti-β-catenin (green) in ARPE19 cells cotransfected with expression vectors encoding human FZD4 and LRP5, shRNA targeting RCBTB1 (shRCBTB1-1) or the vector-only control (pLKO.1) in the presence of 250 ng/mL rhNorrin or Wnt3a for 4 h. Significant nuclear accumulation of β-catenin was noted in the vector control group rather than in the RCBTB1 knockdown group. The arrowheads indicate the locations of nuclei in cells.

Discussion

By employing exome sequencing, bioinformatic analysis and the modified overlap strategy, we identified two frameshift mutations in RCBTB1 (c.707delA and c.1172+1G>A) in two unrelated Taiwanese families affected by Coats disease and FEVR. To date, in addition to the five FEVR-associated genes, RCBTB1 was identified as the sixth putative causative gene in two out of nine families, reflecting the genetic heterogeneity of FEVR and suggesting that other genetic factors may also contribute to the development of vitreoretinopathies.

RCBTB1, also known as CLLD7, is located at chromosome 13q14.3, which is frequently deleted in cases of B-cell chronic lymphocytic leukemia and other neoplasms (35). Although RCBTB1 was thought to be a tumor suppressor gene (35), its physiologic and pathologic roles remain unclear because of limited experimental data (36). RCBTB1 is composed of two domains: the N-terminal RCC1 domain may act as a guanine exchange factor for Ran (37), a critical protein for nuclear transport and the BTB domain is implicated in protein–protein interactions (38). Moreover, a previous study found that the BTB domain of RCBTB1 can act as a substrate adaptor for the CUL3-based E3 ligase and can interact with UbcM2, a highly conserved E2 ubiquitin-conjugating enzyme (39). The abundant expression of UbcM2 in the retina (40) and another E3 ubiquitin ligase, Fbxw7, has been demonstrated to act as a potent positive regulator in angiogenesis by inhibiting Notch signaling in zebrafish (41). These findings raise the possibility that RCBTB1-mediated ubiquitination may be involved in the regulation of angiogenic pathways such as Norrin-induced β-catenin signaling. How these multiple roles and interactions of RCBTB1 are conducted at the cellular level has yet to be investigated.

To link RCBTB1 to vascular development in vivo, knockdown of rcbtb1 by morpholinos in zebrafish was performed to investigate the role of rcbtb1 in angiogenesis. Although some poor correlation between phenotypes in germline mutants and those in morpholino-mediated gene knockdown were reported (42), injections with two different rcbtb1 MOs, a mismatch control, the rescue experiment using human WT and mutant RCBTB1 mRNA as well as CRISPR-mediated transcriptional knockdown were carried out to minimize the possibilities for false-positive phenotypes and off-target effects of MOs in this study (Fig. 3; Supplementary Material, Figs S6 and S7). In this way, delayed sprouting of trunk vessels and defects in ISVs and IOVs were revealed in rcbtb1 morphants, which mimic the avascularization in FEVR and Coats disease under haploinsufficient RCBTB1 genotypes and thus support the correlation between the genotypes and phenotypes. Moreover, compared with the phenotypes of rcbtb1 knockdown, the effect of reduced ndp is restricted to the IOVs but not observed in the ISVs. Further, we showed genetic interaction between NDP and RCBTB1 in vitro and in vivo. These findings suggest that ubiquitously expressed RCBTB1 may act as a common downstream component of angiogenesis in general, while the tissue- or organ-specific angiogenesis is guided by other specific factors, e.g. Norrin and TSPAN12.

In the patient cohort, we identified two heterozygous frameshift mutations in RCBTB1, one of which is a splice-site variant (c.1172+1G>A) resulting in the skip of exon 10 in RCBTB1 transcripts, thereby leading to half the amount of normal RCBTB1 in LCLs. Furthermore, partially reducing RCBTB1 by using shRNA in the transiently transfected cell model and MOs in zebrafish significantly lowered Norrin-induced β-catenin signaling by reduced nuclear accumulation of β-catenin and led to poorly developed vasculatures, respectively. These results suggest that RCBTB1 haploinsufficiency may lead to the observed disease phenotypes. In addition, according to the reporter assay, the novel variant at the 5′UTR of NDP (c.-77A>G) may have resulted in a halved Norrin level in family E9, in which both children affected were hemizygous for the mutant allele, whereas their unaffected mothers were heterozygous. On the basis of the Norrin and RCBTB1 levels estimated according to the genotypes of the members of family E9 (Figs 1B and 2C) and additional evidence obtained from cell transfection experiments (Fig. 4A; Supplementary Material, Fig. S2B) and zebrafish morphant analyses (Fig. 3D), we propose a possible disease mechanism, in which normal angiogenesis can be achieved as long as the Norrin-induced signals exceed a threshold. Thus, combined genetic factors, such as mutations in RCBTB1 and NDP, both influencing Norrin-induced signaling, enhance disease susceptibility in a dynamic microenvironment, leading to defects characterized by a wide spectrum of severity in retinal angiogenesis.

Reduced penetrance and asymmetric retinal phenotypes have long been noted in vitreoretinopathies (22). In all affected cases with RCBTB1 mutations, asymmetry in retinal manifestation was obvious as documented in Table 1. In addition, there are heterozygous nonpenetrating parents with RCBTB1 mutations in both families E5 and E9. These characteristics were also found in family E10 with a reported FZD4 mutation (Supplementary material, Fig. S1A; clinical feature not shown). These phenomena could also result from the effect of genetic background in heterozygous individuals and microenvironmental factors in the eyes. Recently, a rare variant in RCBTB1 was detected in a consanguineous family with retinal ciliopathy (43), a disease of photoreceptors. A follow-up electroretinogram (ERG) was performed in the proband, individual III-2, of family E9 (Supplementary Material, Fig. S1B). It was revealed that the rod and cone responses were preserved in his relatively normal eye with only mild disc dragging, while the ERG was flat in the other eye with severe RD. Based on his follow-up histories during past 20 years, including the preserved ERG, lack of night blindness and stable clinical features of FEVR without bony spicules or sheathing vessels, all suggest that he is a patient with FEVR without clinical overlap of retinitis pigmentosa (RP). Besides, the pathophysiologies of vitreoretinopathies and RP are quite different, where angiogenesis is the primary defect in vitreoretinopathies and usually asymmetric while RP affects the photoreceptors and almost always in both eyes. Thus, homozygous RCBTB1 variants in retinal ciliopathy could be an example of clinical heterogeneity.

In addition to known FEVR-associated genes, many other angiogenesis-related genes have been identified in zebrafish (44), indicating the presence of numerous complex signaling pathways involved in angiogenesis. On the basis of the findings, we can speculate the existence of additional unidentified genetic factors associated with FEVR. The frequency of gene mutations may reflect their roles in angiogenesis, i.e. mutations in crucial rate-limiting factors such as FZD4 and LRP5 tend to be more frequently identified, whereas fewer mutations occur in the regulatory components that modulate the major signaling pathways. A recent study involving a cohort of 92 FEVR families identified mutations in LRP5 and FZD4 in 19% and 15% of cases, respectively. Of the families recruited, 48.5% had a confirmed molecular diagnosis (28), indicating that the Norrin-induced β-catenin signaling pathway plays an essential role in retinal angiogenesis. To date, mutations have been identified in only approximately 50% of cases, providing an opportunity to investigate (retinal) angiogenesis by identifying additional genetic factors associated with FEVR and related vitreoretinopathy.

In conclusion, we identified RCBTB1 as a gene associated with vitreoretinopathy and found that it plays a role in retinal angiogenesis through Norrin-induced β-catenin signaling. Further investigation of its physiological role may elucidate its detailed functions in retinal angiogenesis.

Subjects, Materials and Methods

Ethics statement for human subject research

This study was approved by the Institutional Review Boards of Taipei Veterans General Hospital (for FEVR) and Cheng-Hsin General Hospital (for Coats disease). All research studies involving humans were conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from all participants.

Patient recruitment

Fifteen patients diagnosed with FEVR or Coats disease and their family members were recruited from nine Taiwanese families and provided informed consent (Fig. 1; Supplementary Material, Fig. S1). The patients provided a detailed history by answering questions in order to rule out retinopathy of prematurity, and ophthalmic examinations included assessments of corrected visual acuity, slit-lamp examinations, intraocular pressure application and dilated fundus examination with photographs with or without fluorescein angiography. The fundus features were grouped into the following categories for diagnosis: avascular zone only; extraretinal neovascularization; exudative RD; tractional RD; rhegmatogenous RD; retinal breaks without RD; retinoschisis and macular ectopia. Approximately 20 ml of peripheral venous blood was drawn, with 10 ml being used for genomic DNA and total RNA isolation conducted according to our established protocol and the other 10 ml being used for the establishment of LCLs by the Bioresource Collection and Research Center (BCRC), Hsinchu, Taiwan, for additional RNA and protein analyses.

Direct sequencing of candidate genes and mutation analysis

Genomic DNA was extracted from the peripheral blood leukocytes of all participants, according to standard lab protocols. All exons and exon–intron boundaries of the candidate genes (NDP, FZD4, LRP5, TSPAN12 and ZNF408) were amplified under optimized PCR conditions. Subsequently, direct sequence analysis was conducted using the ABI PRISM Big Dye Terminator Cycle Sequencing V2.0 Ready Reaction Kit and ABI PRISM 3730 DNA analyzer (Applied Biosystems). The allele frequencies of all three novel variants (c.-77A>G of NDP; c.707delA and c.1172+1G>A of RCBTB1) in at least 150 Taiwanese controls were assessed by conducting direct Sanger sequencing. Reference sequences for NDP and RTBCB1 are NM_000266.3 and NM_018191.3, respectively. These variants have been submitted to the ClinVar database (http://www.ncbi.nlm.nih.gov/clinvar/).

Exome sequencing analysis and variant filtering

Exome sequencing was performed by the Genome Research Center of National Yang-Ming University for four affected individuals (i.e. II-1 of family E5, II-1 of family E6, III-2 and III-4 of family E9) after excluding possible pathogenic mutations in the five candidate genes. Exon sequences were enriched from genomic DNA by using the Agilent SureSelect All Exon 50 Mb or ROCHE NimbleGen SeqCap EZ Exome V3+UTR (96 Mb) kits and were paired-end sequenced on an Illumina HiSeq 2000 or 2500 platform. The CLC Genomics Workbench v7.0.4 was used for basic quality assessment and variant calling. For each patient sample, at least 90% of the targeted region was covered by at least 20 folds, as displayed in Supplementary Material, Table S1. Variants that were not in the dbSNP138 common-variant list were filtered for being heterozygous and were present in both III-2 and III-4 of family E9. Filtering was followed by removing variants listed in the in-house exome database for unrelated phenotypes, and the remaining variants were analyzed using the Variant Effect Predictor of the Ensembl Genome Browser. The 130 variants with predicted consequences of in-frame deletion, missense, stop-gained or splice-site variants were individually verified using Ensembl (release 75, Feb. 2014). The remaining 27 variants were then prioritized through segregation according to the results obtained from Sanger sequencing, pathway analyses with known candidate genes and the presence of other variants in the same gene in individuals II-1 of family E5 and/or II-1 of family E6.

Plasmids and constructs

The WT and mutant TCF/LEF-binding sequences were amplified using M50 Super 8× TOPFlash and M51 Super 8× FOPFlash plasmids (Addgene) as the template, respectively, and were cloned upstream of the firefly luciferase gene in pmirGlo vectors (Promega). The expression constructs for human NDP with the 5′UTR region, FZD4, and LRP5 were generated using pcDNA3.1myc-His vectors (Invitrogen). Site-directed mutagenesis was performed through PCR by using Phusion DNA polymerase (New England Biolabs) as well as the corresponding primers shown in Supplementary Material, Table S3, to generate mutant constructs. All DNA constructs after site-directed mutagenesis were validated through Sanger sequencing. Two RCBTB1-targeting shRNA plasmids and the pLKO.1 vector were obtained from the National RNAi Core Facility at the Institute of Molecular Biology/Genomic Research Center, Academia Sinica (Taipei, Taiwan), with the following oligo sequences: shRCBTB1-1 (TRCN0000160987): CCGGGCCAAATTACAAGTGGGTGAACTCGAGTTCACCCACTTGTAATTTGGCTTTTTTG, shRCBTB1-2 (TRCN0000273669): CCGGCCTTACTGTGAAGGAATAATTCTCGAGAATTATTCCTTCACAGTAAGGTTTTTG, pLKO.1 empty vector: CCGGACACTCGAGCACTTTTTG.

Cell culture and transient transfection

ARPE19 cells (ATCC number: CRL-2302) were cultured in DMEM/F12 (GIBCO®) supplemented with 10% fetal bovine serum (FBS) in a 5% CO2 incubator. LCLs derived from Epstein-Barr-virus-transformed lymphocytes (BCRC, Hsinchu, Taiwan) from the members of family E9 were cultured in RPMI 1640 supplemented with 20% FBS, l-glutamine and penicillin–streptomycin solution (GIBCO®) in a 5% CO2 incubator standing in the upright position. Transient transfections of ARPE19 with expression vectors or shRNA were performed using Lipofectamine® 3000, according to manufacturer instructions.

RNA extraction and RT-PCR

Total RNA was extracted from peripheral blood leukocytes or LCLs by using the TRIzol® Reagent (Ambion) according to the manufacturer protocol. Reverse transcriptase reactions were performed on 500 ng of total RNA by using M-MLV Reverse Transcriptase (Life Technologies), and 1 μl of cDNA was used to conduct PCR. The PCR products were electrophoresed in a 2% TBE gel, and the products of the aberrant transcripts were excised from the gel and extracted using the Zymoclean™ Gel DNA Recovery Kit (Zymo Research) prior to sequencing. Details of the primer sequences are shown in Supplementary Material, Table S3.

Whole cell lysate preparation

The cultured cells were lysed with RIPA buffer (50 mm Tris–HCl, pH 8.0, 150 mm NaCl, 0.1% SDS, 1% NP-40 and 0.5% sodium deoxycholate) supplemented with 1× protease inhibitor cocktail (cOmplete Protease Inhibitor Cocktail Tablets, Roche). The cell lysates were cleared through centrifugation at 14 000 × g for 20 min, and the protein concentration was quantified using the Bradford protein assay (BioRad).

Nuclear and cytosolic fractionation

The method for subcellular fractionation was based on Dr David Ron's Lab's protocol (http://ron.cimr.cam.ac.uk/protocols.html). Briefly, cells were washed twice with ice-cold calcium-tris-buffered saline (CTMS) (10 mm Tris–HCl, pH 7.4, 140 mm NaCl, 2 mm CaCl2). For one 10-cm dish, cells were scraped from the dish in 1 ml of CTBS buffer supplemented with 2 mm DTT, 5 mm EDTA and 1× protease inhibitor cocktail (Roche). The harvested cells were pelleted, resuspended in 0.4 ml serum-free media containing with 7.5 mmN-ethylmaleimide (Sigma-Aldrich) (45) and incubated at room temperature for 10 min. After washed in ice-cold CTBS, cells were lysed in Harvest buffer (10 mm HEPES, pH 7.9, 50 mm NaCl, 0.5 m sucrose, 0.1 mm EDTA, 0.5% Triton × 100) containing 1× protease inhibitor cocktail (ROCHE) and 1× phosphatase inhibitor cocktail (Sigma). The cleared supernatant is the cytoplasmic/membrane fraction. The washed nuclear pellets were lysed in buffer 2 × C (20 mm HEPES, pH 7.9, 1 m NaCl, 0.2 mm EDTA, 0.2 mm EGTA, 0.2% NP40). The cleared supernatant contains the nuclear fraction.

Immunoblot analysis

For immunoblot analysis, protein lysates were electrophoresed on 10% SDS-PAGE gels and transferred to PVDF membranes (Millipore). Antibodies to RCBTB1 (ab154649, Abcam®), active (nonphospho) β-catenin (#8814, Cell Signaling Technology®), Lamin B1 (ab133741, Abcam®), GAPDH (NB300-322, Novus Biologicals) and β-actin (8H10D10, Cell Signaling Technology®) were used to detect and quantify the proteins. Chemiluminescent images were captured using the BioSpectrum Imaging System (UVP), and signal intensities were analyzed using ImageJ (Image Processing and Analysis in Java, http://imagej.nih.gov/ij/).

Dual luciferase assay

In 24-well plates, 105 ARPE19 cells/well were transfected with 550 ng of DNA by using 1 μl of the Lipofectamine® 3000 Reagent (Invitrogen). The DNA mixture contained 150 ng each of the expression constructs for FZD and LRP5, 150 ng of the RCBTB1-targeting shRNA plasmid or vector control and 100 ng of reporter constructs (M50) or mutant reporter constructs (M51). The cells were stimulated 24 h post-transfection with 0, 62.5 or 125 ng/mL human recombinant Norrin (rhNorrin) or 100 ng/mL Wnt3a (R&D Systems), or by cotransfecting the indicated amount of NDP-expressing constructs with WT or mutant alleles for 16–18 h. Firefly and Renilla luciferase activity was measured using the Dual-Luciferase® Reporter Assay System (Promega) according to the manufacturer instructions. Relative luciferase activity (RLA) is the ratio between the Firefly and Renilla luciferase activities and is used to normalize transfection efficiency and changes in cell survival and growth.

Immunofluorescence

ARPE19 cells transfected with DNA mixture contained 150 ng each of the expression constructs for FZD and LRP5, 200 ng of the RCBTB1-targeting shRNA plasmids or vector controls were cultured in four-well chamber slides (Nunc® Lab-Tek®). Twenty-four hours after transfection, culture media were replaced with serum-free only media or media containing 250 ng/ml rhNorrin or rhWnt3a for another 4 h and then washed with PBS for three times. The cells were fixed using 4% paraformaldehyde at 37°C for 15 min and followed by three PBS washes. Then 0.25% Triton X-100 in PBS was added to permeabilize cells at room temperature for 5 min and followed by PBST wash for three times. The cells were incubated in 1% BSA in PBST for 1 h and then incubated in 1:100 diluted anti-β-catenin antibody (NBP1-54467SS, Novus) at 4°C overnight. After washed for three times with PBST, the slides were incubated in 1:200 diluted AlexaFluor488-conjugated Donkey anti-Mouse IgG H&L (ab150105, Abcam) for 1 h and afterwards washed with PBST for three times. After mounted with SlowFade Gold Antifade Mountant with DAPI (Life Technologies, S36942), the slides were examined by confocal microscope FV10i and analyzed by the FV10-ASW 4.0 Viewer (Olympus).

Morpholinos and zebrafish embryo manipulations

The Tg(fli1:EGFP) zebrafish were maintained and mated as described (34), and all experiments were conducted with the approval of the Institutional Animal Care and Use Committee, National Health Research Institutes (NHRI-IACUC-103003). Translation-blocking (GAGGCCACTTACTCACATCCACCAT), mismatched-control (GAcGCCAaTTACTaACATaCACaAT) and splice blocking (GTTAAAAAGCATTCCCTCACCTCAC) MOs for the zebrafish rcbtb1 ortholog and translation-blocking (GAGCGACCGAGTTCCTCATAGTGTC) MOs for ndp were designed and synthesized using LLC (Gene Tools) and diluted in nuclease-free, deionized, sterile water supplemented with 0.8% phenol red. To determine the most effective dose of the MOs, 2.3 nL of diluted MO (containing 2, 4 and 6 ng) was injected into cells at one- or two-cell-stage embryos by using an oil-PicoPump pv280 (World Precision Instruments). For in vivo rescue experiments, human WT RCBTB1 mRNAs were prepared using the mMESSAGE mMACHINE Kit (Ambion) by following the manufacturer instructions. The mRNAs (200 pg) were coinjected with MOs as described. After injection, the embryos were cultured at 28.5°C in the E3 embryo medium and were subsequently phenotyped at 3 dpf or 4 dpf through confocal microscopy by using the Leica AF6000 LX. The average areas of eight ISVs counted posteriorly from the urogenital pole were analyzed using ImageJ, and thresholds were optimized to enable the imaging signals to cover the ISV area. The pixels in the covered area were then counted and normalized to those of the control group. For characterizing the intraocular phenotype at 4 dpf, injected embryos were classified into three classes of phenotypes according to the relative severity and the ratio of the avascularization area compared with the MO-injected (6 ng) control embryos. To determine the efficiency of splice blocking, the RNA was isolated from 60 control MOs and different doses of rcbtb1 splice MO-injected embryos at 1 dpf or 2 dpf by using RNAzol®RT (Molecular Research Center) according to the manufacturer protocol. Reverse-transcriptase reaction was performed using 1 μg of RNA and Maxima Reverse Transcriptase (Thermo Scientific) according to the manufacturer protocol. PCR was then conducted, as follows: (forward: CGGACCTCATGTTTTGCTGG and reverse: AACCCCAGCTGGCCATTAC). Fragments were resolved through electrophoresis, extracted from the gel by using the Zymoclean™ Gel DNA Recovery Kit (Zymo Research), and Sanger-sequenced as described.

Statistical analysis

The means and standard error of the mean for all experimental data were calculated. Comparisons between the experimental conditions were conducted using a two-tailed Student t-test for unpaired samples, with α < 0.05 considered to indicate significance.

Authors’ contributions

Project design: M.-Y.C., J.-H.W., J.-H.L., S.-J.C., Y.-J.J.; manuscript writing: J.-H.W., M.-Y.C., S.-J.C., Y.-J.J.; project coordination and senior leaders of the groups: M.-Y.C., S.-J.C., Y.-J.J.; patient recruitment and phenotyping: J.-H.L., H.-M.C., Y.-C.K., Y.-C.C., S.-J.C.; genetic sequencing and interpretation: J.-H.W., M.-Y.C., C.-T.W., T.-T.L.; cell biology experiments: J.-H.W., M.-Y.C., C.T.-W.; and zebrafish experiments: J.-H.W., K.-C.C., Y.-J.J.

Supplementary Material

Supplementary Material is available at HMG online.

Funding

This study was supported by grants from Ministry of Science and Technology (MOST) (104-2314-B-010-049) and Cheng-Yang Collaborative Research Foundation (98F117CY02, 99F167CY05) to M.-Y.C.; the High-throughput Genome Analysis and RNAi Core Facilities supported by the National Core Facility Program for Biotechnology (NSC-103-2319-B-010-001 and NSC100-2319-B-001-002) for Illumina sequencing and RNAi reagents; EBV-transformation service of the Resource Center supported by the National Research Program for Biopharmaceuticals (SB3, NSC100-2325-B-080-001). The work was also supported by grants from the National Health Research Institutes, Taiwan (MG-104-PP-12 and MG-104-PP-13) and the Ministry of Science and Technology, Taiwan (MOST103-2311-B-400-001 and MOST104-2319-B-400-001) to Y.-J.J.

Acknowledgements

We thank all the patients and their family members for participation; Professors Chen-Kung Chou, Ming-Yuan Cheng and Yann-Jang Chen for their valuable discussions and comments. We also thank Dr Didier Y. R. Stainier for dCas9 and gRNA plasmids and the staff in the Zebrafish Facility of NHRI for their efforts in maintaining fish stocks. Also, we are grateful to Chien-Ming Wang and Chih-Hao Tang for their technical help.

Conflict of Interest statement. None declared.

References

1
Criswick
V.G.
,
Schepens
C.L.
(
1969
)
Familial exudative vitreoretinopathy
.
Am. J. Ophthalmol.
 ,
68
,
578
594
.
2
Canny
C.L.
,
Oliver
G.L.
(
1976
)
Fluorescein angiographic findings in familial exudative vitreoretinopathy
.
Arch. Ophthalmol.
 ,
94
,
1114
1120
.
3
Zaremba
J.
,
Feil
S.
,
Juszko
J.
,
Myga
W.
,
van Duijnhoven
G.
,
Berger
W.
(
1998
)
Intrafamilial variability of the ocular phenotype in a Polish family with a missense mutation (A63D) in the Norrie disease gene
.
Ophthalmic Genet.
 ,
19
,
157
164
.
4
Robitaille
J.M.
,
Wallace
K.
,
Zheng
B.
,
Beis
M.J.
,
Samuels
M.
,
Hoskin-Mott
A.
,
Guernsey
D.L.
(
2009
)
Phenotypic overlap of familial exudative vitreoretinopathy (FEVR) with persistent fetal vasculature (PFV) caused by FZD4 mutations in two distinct pedigrees
.
Ophthalmic Genet.
 ,
30
,
23
30
.
5
Berger
W.
,
Kloeckener-Gruissem
B.
,
Neidhardt
J.
(
2010
)
The molecular basis of human retinal and vitreoretinal diseases
.
Prog. Retin. Eye Res.
 ,
29
,
335
375
.
6
Nikopoulos
K.
,
Venselaar
H.
,
Collin
R.W.
,
Riveiro-Alvarez
R.
,
Boonstra
F.N.
,
Hooymans
J.M.
,
Mukhopadhyay
A.
,
Shears
D.
,
van Bers
M.
,
de Wijs
I.J.
et al
. (
2010
)
Overview of the mutation spectrum in familial exudative vitreoretinopathy and Norrie disease with identification of 21 novel variants in FZD4, LRP5, and NDP
.
Hum. Mutat.
 ,
31
,
656
666
.
7
Robitaille
J.M.
,
Zheng
B.
,
Wallace
K.
,
Beis
M.J.
,
Tatlidil
C.
,
Yang
J.
,
Sheidow
T.G.
,
Siebert
L.
,
Levin
A.V.
,
Lam
W.C.
et al
. (
2011
)
The role of Frizzled-4 mutations in familial exudative vitreoretinopathy and Coats disease
.
Br. J. Ophthalmol.
 ,
95
,
574
579
.
8
Schafer
N.F.
,
Luhmann
U.F.
,
Feil
S.
,
Berger
W.
(
2009
)
Differential gene expression in Ndph-knockout mice in retinal development
.
Invest. Ophthalmol. Vis. Sci.
 ,
50
,
906
916
.
9
Gilmour
D.F.
(
2015
)
Familial exudative vitreoretinopathy and related retinopathies
.
Eye (Lond.)
 ,
29
,
1
14
.
10
Gal
M.
,
Levanon
E.Y.
,
Hujeirat
Y.
,
Khayat
M.
,
Pe'er
J.
,
Shalev
S.
(
2014
)
Novel mutation in TSPAN12 leads to autosomal recessive inheritance of congenital vitreoretinal disease with intra-familial phenotypic variability
.
Am. J. Med. Genet. A.
 ,
164A
,
2996
3002
. .
Epub 32014 Sep 36722
.
11
Morris
B.
,
Foot
B.
,
Mulvihill
A.
(
2010
)
A population-based study of Coats disease in the United Kingdom I: epidemiology and clinical features at diagnosis
.
Eye (Lond)
 ,
24
,
1797
1801
.
12
Black
G.C.
,
Perveen
R.
,
Bonshek
R.
,
Cahill
M.
,
Clayton-Smith
J.
,
Lloyd
I.C.
,
McLeod
D.
(
1999
)
Coats’ disease of the retina (unilateral retinal telangiectasis) caused by somatic mutation in the NDP gene: a role for norrin in retinal angiogenesis
.
Hum. Mol. Genet.
 ,
8
,
2031
2035
.
13
Shields
J.A.
,
Shields
C.L.
(
2002
)
Review: Coats disease: the 2001 LuEsther T. Mertz lecture
.
Retina
 ,
22
,
80
91
.
14
Smithen
L.M.
,
Brown
G.C.
,
Brucker
A.J.
,
Yannuzzi
L.A.
,
Klais
C.M.
,
Spaide
R.F.
(
2005
)
Coats’ disease diagnosed in adulthood
.
Ophthalmology
 ,
112
,
1072
1078
.
15
Chen
Z.Y.
,
Battinelli
E.M.
,
Fielder
A.
,
Bundey
S.
,
Sims
K.
,
Breakefield
X.O.
,
Craig
I.W.
(
1993
)
A mutation in the Norrie disease gene (NDP) associated with X-linked familial exudative vitreoretinopathy
.
Nat. Genet.
 ,
5
,
180
183
.
16
Kashani
A.H.
,
Brown
K.T.
,
Chang
E.
,
Drenser
K.A.
,
Capone
A.
,
Trese
M.T.
(
2014
)
Diversity of retinal vascular anomalies in patients with familial exudative vitreoretinopathy
.
Ophthalmology
 ,
121
,
2220
2227
.
17
Poulter
J.A.
,
Ali
M.
,
Gilmour
D.F.
,
Rice
A.
,
Kondo
H.
,
Hayashi
K.
,
Mackey
D.A.
,
Kearns
L.S.
,
Ruddle
J.B.
,
Craig
J.E.
et al
. (
2010
)
Mutations in TSPAN12 cause autosomal-dominant familial exudative vitreoretinopathy
.
Am. J. Hum. Genet.
 ,
86
,
248
253
.
18
Simunovic
M.P.
,
Maberley
D.A.
(
2014
)
Familial exudative vitreoretinopathy mimicking macular telangiectasia type 1
.
Can. J. Ophthalmol.
 ,
49
,
e28
e30
.
19
Gow
J.
,
Oliver
G.L.
(
1971
)
Familial exudative vitreoretinopathy. An expanded view
.
Arch. Ophthalmol.
 ,
86
,
150
155
.
20
Jiao
X.
,
Ventruto
V.
,
Trese
M.T.
,
Shastry
B.S.
,
Hejtmancik
J.F.
(
2004
)
Autosomal recessive familial exudative vitreoretinopathy is associated with mutations in LRP5
.
Am. J. Hum. Genet.
 ,
75
,
878
884
.
21
van Nouhuys
C.E.
(
1982
)
Dominant exudative vitreoretinopathy and other vascular developmental disorders of the peripheral retina
.
Doc. Ophthalmol.
 ,
54
,
1
414
.
22
Kashani
A.H.
,
Learned
D.
,
Nudleman
E.
,
Drenser
K.A.
,
Capone
A.
,
Trese
M.T.
(
2014
)
High prevalence of peripheral retinal vascular anomalies in family members of patients with familial exudative vitreoretinopathy
.
Ophthalmology
 ,
121
,
262
268
.
23
Collin
R.W.
,
Nikopoulos
K.
,
Dona
M.
,
Gilissen
C.
,
Hoischen
A.
,
Boonstra
F.N.
,
Poulter
J.A.
,
Kondo
H.
,
Berger
W.
,
Toomes
C.
et al
. (
2013
)
ZNF408 is mutated in familial exudative vitreoretinopathy and is crucial for the development of zebrafish retinal vasculature
.
Proc. Natl. Acad. Sci. USA
 ,
110
,
9856
9861
.
24
Nikopoulos
K.
,
Gilissen
C.
,
Hoischen
A.
,
van Nouhuys
C.E.
,
Boonstra
F.N.
,
Blokland
E.A.
,
Arts
P.
,
Wieskamp
N.
,
Strom
T.M.
,
Ayuso
C.
et al
. (
2010
)
Next-generation sequencing of a 40 Mb linkage interval reveals TSPAN12 mutations in patients with familial exudative vitreoretinopathy
.
Am. J. Hum. Genet.
 ,
86
,
240
247
.
25
Robitaille
J.
,
MacDonald
M.L.
,
Kaykas
A.
,
Sheldahl
L.C.
,
Zeisler
J.
,
Dube
M.P.
,
Zhang
L.H.
,
Singaraja
R.R.
,
Guernsey
D.L.
,
Zheng
B.
et al
. (
2002
)
Mutant frizzled-4 disrupts retinal angiogenesis in familial exudative vitreoretinopathy
.
Nat. Genet.
 ,
32
,
326
330
.
26
Toomes
C.
,
Bottomley
H.M.
,
Jackson
R.M.
,
Towns
K.V.
,
Scott
S.
,
Mackey
D.A.
,
Craig
J.E.
,
Jiang
L.
,
Yang
Z.
,
Trembath
R.
et al
. (
2004
)
Mutations in LRP5 or FZD4 underlie the common familial exudative vitreoretinopathy locus on chromosome 11q
.
Am. J. Hum. Genet.
 ,
74
,
721
730
.
27
Junge
H.J.
,
Yang
S.
,
Burton
J.B.
,
Paes
K.
,
Shu
X.
,
French
D.M.
,
Costa
M.
,
Rice
D.S.
,
Ye
W.
(
2009
)
TSPAN12 regulates retinal vascular development by promoting Norrin- but not Wnt-induced FZD4/beta-catenin signaling
.
Cell
 ,
139
,
299
311
.
28
Salvo
J.
,
Lyubasyuk
V.
,
Xu
M.
,
Wang
H.
,
Wang
F.
,
Nguyen
D.
,
Wang
K.
,
Luo
H.
,
Wen
C.
,
Shi
C.
et al
. (
2015
)
Next-generation sequencing and novel variant determination in a cohort of 92 familial exudative vitreoretinopathy patients
.
Invest. Ophthalmol. Vis. Sci.
 ,
56
,
1937
1946
.
29
Aponte
E.P.
,
Pulido
J.S.
,
Ellison
J.W.
,
Quiram
P.A.
,
Mohney
B.G.
(
2009
)
A novel NDP mutation in an infant with unilateral persistent fetal vasculature and retinal vasculopathy
.
Ophthalmic Genet.
 ,
30
,
99
102
.
30
Meindl
A.
,
Berger
W.
,
Meitinger
T.
,
van de Pol
D.
,
Achatz
H.
,
Dorner
C.
,
Haasemann
M.
,
Hellebrand
H.
,
Gal
A.
,
Cremers
F.
et al
. (
1992
)
Norrie disease is caused by mutations in an extracellular protein resembling C-terminal globular domain of mucins
.
Nat. Genet.
 ,
2
,
139
143
.
31
Lin
P.
,
Shankar
S.P.
,
Duncan
J.
,
Slavotinek
A.
,
Stone
E.M.
,
Rutar
T.
(
2010
)
Retinal vascular abnormalities and dragged maculae in a carrier with a new NDP mutation (c.268delC) that caused severe Norrie disease in the proband
.
J. AAPOS
 ,
14
,
93
96
.
32
Ng
S.B.
,
Bigham
A.W.
,
Buckingham
K.J.
,
Hannibal
M.C.
,
McMillin
M.J.
,
Gildersleeve
H.I.
,
Beck
A.E.
,
Tabor
H.K.
,
Cooper
G.M.
,
Mefford
H.C.
et al
. (
2010
)
Exome sequencing identifies MLL2 mutations as a cause of Kabuki syndrome
.
Nat. Genet.
 ,
42
,
790
793
.
33
Auton
A.
,
Brooks
L.D.
,
Durbin
R.M.
,
Garrison
E.P.
,
Kang
H.M.
,
Korbel
J.O.
,
Marchini
J.L.
,
McCarthy
S.
,
McVean
G.A.
,
Abecasis
G.R.
(
2015
)
A global reference for human genetic variation
.
Nature
 ,
526
,
68
74
. .
34
Ellertsdottir
E.
,
Lenard
A.
,
Blum
Y.
,
Krudewig
A.
,
Herwig
L.
,
Affolter
M.
,
Belting
H.G.
(
2010
)
Vascular morphogenesis in the zebrafish embryo
.
Dev. Biol.
 ,
341
,
56
65
.
35
Mabuchi
H.
,
Fujii
H.
,
Calin
G.
,
Alder
H.
,
Negrini
M.
,
Rassenti
L.
,
Kipps
T.J.
,
Bullrich
F.
,
Croce
C.M.
(
2001
)
Cloning and characterization of CLLD6, CLLD7, and CLLD8, novel candidate genes for leukemogenesis at chromosome 13q14, a region commonly deleted in B-cell chronic lymphocytic leukemia
.
Cancer Res.
 ,
61
,
2870
2877
.
36
Zhou
X.
,
Munger
K.
(
2010
)
Clld7, a candidate tumor suppressor on chromosome 13q14, regulates pathways of DNA damage/repair and apoptosis
.
Cancer Res.
 ,
70
,
9434
9443
.
37
Bischoff
F.R.
,
Ponstingl
H.
(
1991
)
Catalysis of guanine nucleotide exchange on Ran by the mitotic regulator RCC1
.
Nature
 ,
354
,
80
82
.
38
Hadjebi
O.
,
Casas-Terradellas
E.
,
Garcia-Gonzalo
F.R.
,
Rosa
J.L.
(
2008
)
The RCC1 superfamily: from genes, to function, to disease
.
Biochim. Biophys. Acta
 ,
1783
,
1467
1479
.
39
Plafker
K.S.
,
Singer
J.D.
,
Plafker
S.M.
(
2009
)
The ubiquitin conjugating enzyme, UbcM2, engages in novel interactions with components of cullin-3 based E3 ligases
.
Biochemistry
 ,
48
,
3527
3537
.
40
Mirza
S.
,
Plafker
K.S.
,
Aston
C.
,
Plafker
S.M.
(
2010
)
Expression and distribution of the class III ubiquitin-conjugating enzymes in the retina
.
Mol. Vis.
 ,
16
,
2425
2437
.
41
Izumi
N.
,
Helker
C.
,
Ehling
M.
,
Behrens
A.
,
Herzog
W.
,
Adams
R.H.
(
2012
)
Fbxw7 controls angiogenesis by regulating endothelial Notch activity
.
PLoS One
 ,
7
,
e41116
.
42
Kok
F.O.
,
Shin
M.
,
Ni
C.W.
,
Gupta
A.
,
Grosse
A.S.
,
van Impel
A.
,
Kirchmaier
B.C.
,
Peterson-Maduro
J.
,
Kourkoulis
G.
,
Male
I.
et al
. (
2015
)
Reverse genetic screening reveals poor correlation between morpholino-induced and mutant phenotypes in zebrafish
.
Dev. Cell
 ,
32
,
97
108
.
43
Coppieters
F.
,
Ascari
G.
,
Karlstetter
M.
,
Bauwens
M.
,
De Rocker
N.
,
Boel
A.
,
Vleminckx
K.
,
Van der Eecken
M.
,
Leroy
B.P.
,
Meire
F.
et al
. (
2015
)
Identification of RCBTB1 as novel disease gene for retinal ciliopathy (Program number 283). Presented at the 65th Annual Meeting of the American Society of Human Genetics, October 6–10, 2015, Baltimore, MD
.
44
Alvarez
Y.
,
Cederlund
M.L.
,
Cottell
D.C.
,
Bill
B.R.
,
Ekker
S.C.
,
Torres-Vazquez
J.
,
Weinstein
B.M.
,
Hyde
D.R.
,
Vihtelic
T.S.
,
Kennedy
B.N.
(
2007
)
Genetic determinants of hyaloid and retinal vasculature in zebrafish
.
BMC Dev. Biol.
 ,
7
,
114
.
45
Tinnikov
A.A.
,
Samuels
H.H.
(
2013
)
A novel cell lysis approach reveals that caspase-2 rapidly translocates from the nucleus to the cytoplasm in response to apoptotic stimuli
.
PLoS One
 ,
8
,
e61085
.

Author notes

These authors contributed equally to this work.