Catenin alpha 1 mutations cause familial exudative vitreoretinopathy by overactivating Norrin/beta- signaling

Xianjun Zhu, … , Weiquan Zhu, Zhenglin Yang

J Clin Invest. 2021. https://doi.org/10.1172/JCI139869.

Research In-Press Preview Angiogenesis Genetics

Graphical abstract

Find the latest version: https://jci.me/139869/pdf 1 Catenin alpha 1 mutations cause familial exudative vitreoretinopathy by over-

2 activating Norrin/beta-catenin signaling

3 Xianjun Zhu1,2,6,7,8, Mu Yang1,2,8, Peiquan Zhao3,8, Shujin Li1,2,8, Lin Zhang1, Lulin Huang1, Yi

4 Huang1, Ping Fei3, Yeming Yang1, Shanshan Zhang1, Huijuan Xu1, Ye Yuan1, Xiang Zhang3,

5 Xiong Zhu1, Shi Ma1, Fang Hao1, Periasamy Sundaresan4, Weiquan Zhu5, Zhenglin Yang1,2,6,7,*

6 1Sichuan Provincial Key Laboratory for Human Disease Study, Sichuan Provincial Peo- 7 ple's Hospital, University of Electronic Science and Technology of China, Chengdu, Sichuan, 8 China; 9 2Research Unit for Blindness Prevention of Chinese Academy of medical Sciences 10 (2019RU026), Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital, 11 Chengdu, Sichuan, China; 12 3Department of Ophthalmology, Xinhua Hospital, School of Medicine, Shanghai Jiaotong Uni- 13 versity, Shanghai, China; 14 4Department of Genetics, Aravind Medical Research Foundation, Aravind Eye Hospital, Madu- 15 rai, Tamil Nadu, India; 16 5Department of Molecular Medicine, School of Medicine, University of Utah, Salk Lake City, 17 Utah, USA. 18 6Chinese Academy of Sciences Sichuan Translational Medicine Research Hospital, Chengdu, 19 610072, China. 20 7These authors jointly directed this work. 21 8These authors contributed equally to this work. 22 23 Conflict of interest: The authors have declared that no conflict of interest exists.

24 *Correspondence should be addressed to: Zhenglin Yang, Sichuan Provincial Key Laboratory for 25 Human Disease Gene Study, Sichuan Provincial People's Hospital, University of Electronic Sci- 26 ence and Technology of China, 27 32 The First Ring Road West 2, Chengdu, Sichuan, 610072, China 28 Email: [email protected]. 29 Phone: 86-28-87393375 30 Fax: 86-28-87393596 31

32

1 1 Abstract 2 Familial exudative vitreoretinopathy (FEVR) is a severe retinal vascular disease that causes

3 blindness. FEVR has been linked to mutations in several associated with inactivation of

4 the Norrin/β-catenin signaling pathway, but these account for only approximately 50% of cases.

5 We report that mutations in CTNNA1 (α-catenin) cause FEVR by overactivating the β-catenin

6 pathway and disrupting cell adherens junctions. Three heterozygous mutations in CTNNA1

7 (p.F72S, p.R376Cfs*27 and p.P893L) were identified by exome-sequencing. We further demon-

8 strated that FEVR-associated mutations led to overactivation of Norrin/β-catenin signaling due to

9 impaired interactions within the /catenin complex. The clinical features of FEVR

10 were reproduced in mice lacking Ctnna1 in vascular endothelial cells (ECs) or with overac-

11 tivated β-catenin signaling by an EC-specific gain-of-function allele of Ctnnb1. In isolated

12 mouse lung endothelial cells, both CTNNA1-P893L and F72S mutants failed to rescue either the

13 disrupted F- arrangement or VE-Cadherin and CTNNB1 distribution. Moreover, we dis-

14 covered that compound heterozygous Ctnna1 F72S and a deletion allele could cause similar phe-

15 notype. Furthermore, a LRP5 mutation, which activates Norrin/β-catenin signaling, was identi-

16 fied in a FEVR family and the corresponding knock-in mice exhibited partial FEVR-like pheno-

17 type. Our study demonstrates that precise regulation of β-catenin activation is critical for retinal

18 vascular development and provides new insights into the pathogenesis of FEVR.

19 20

21

22

2 1 Clinical Perspective

2 What Is New?

3 • We identified three heterozygous mutations in CTNNA1 in familial exudative vitreoreti-

4 nopathy (FEVR) patients, and these mutations resulted in overactivation of Norrin/β-

5 catenin signaling and disruption of the cadherin/catenin complex.

6 • Clinical features of FEVR were reproduced in mice lacking Ctnna1 in vascular endothe-

7 lial cells and in mice with an endothelial-cell-specific gain-of-function Ctnnb1 allele.

8 • In a large Indian family with FEVR, we identified an LRP5 mutation (p.P848L) that re-

9 sults in overactivation of Norrin/β-catenin signaling, and we observed clinical features of

10 FEVR in the retina of Lrp5P847L knock-in mice.

11 • The precise regulation of β-catenin activation is critical for normal retinal vascular devel-

12 opment.

13 What Are the Clinical Implications?

14 • Because both in- and overactivation of Norrin/β-catenin signaling can cause defective an-

15 giogenesis, careful monitoring during drug treatment targeting β-catenin is warranted.

16 • The cadherin/catenin complex has the potential to be a therapeutic target for other neo-

17 vascular diseases affecting the blood–brain barrier, which contribute to altered brain

18 function and intellectual disability.

19

20

3 1 Introduction 2 3 Familial exudative vitreoretinopathy (FEVR) is a severe inherited retinal disorder characterized

4 by the incomplete vascularization of the peripheral retina and by the absence or abnormality of

5 the second/tertiary capillary layers in the deep retina (1). These pathological changes are be-

6 lieved to result in neovascularization and exudate, vitreous hemorrhaging, traction from the vitre-

7 ous membranes, displacement (ectopia) of the macula, and the folding and detachment of the ret-

8 ina (2). FEVR is genetically heterogeneous, and its clinical features can be highly variable; fam-

9 ily members with the same mutation can present without symptoms or can exhibit a range of

10 clinical features, including complete blindness (3).

11

12 FEVR can be inherited as an autosomal-dominant, autosomal-recessive, or X-linked disorder.

13 Mutations in 11 genes and one have been identified as causing FEVR: LRP5 (low-density

14 lipoprotein receptor-related protein 5) (4, 5); FZD4 (frizzled 4) (3); NDP (norrin, previously

15 known as Norrie disease pseudoglioma) (6); TSPAN12 (tetraspanin-12) (7-9); ZNF408 (zinc fin-

16 ger protein 408) (2); KIF11 ( family member 11) (10); CTNNB1 (catenin beta 1) (11, 12);

17 ATOH7 (atonal homolog 7) (13); RCBTB1 (RCC1 and BTB domain-containing protein 1) (14);

18 EVR3 (exudative vitreoretinopathy 3) on 11p12-13 (15); ILK (integrin-linked ki-

19 nase) (16) and JAG1 (jagged canonical Notch ligand 1) (17). In addition to FEVR, mutations in

20 LRP5 are associated with osteopenia and osteoporosis (18), whereas mutations in NDP cause

21 Norrie disease, which features intellectual disability (ID) and cognitive impairments (19). Muta-

22 tions in KIF11 are also associated with microcephaly, with or without chorioretinopathy,

23 lymphedema, or ID (20), and mutations in CTNNB1 with syndromic ID (21). Thus, FEVR can

24 feature in syndromes that cause intellectual impairment or disability.

4 1 Previous studies have indicated that the pathogenesis of FEVR may involve disrupted Nor-

2 rin/β-catenin pathway signaling (7, 22-25). Normally, FZD4 and LRP5 form a complex with

3 TSPAN12 in the plasma membrane of endothelial cells (ECs) (7, 22-25). Norrin is an extracellu-

4 lar ligand that binds to FZD4 on ECs with the aid of LRP5 and TSPAN12 to activate down-

5 stream β-catenin/Wnt pathway signaling (7, 22-25). Therefore, mutations in any of the genes en-

6 coding these five (NDP, FZD4, LRP5, TSPAN12, and β-catenin) might lead to FEVR by

7 inactivating the Norrin/β-catenin pathway and altering the expression of its downstream target

8 genes (7, 22-25). However, it remains unclear how mutations in the ZNF408, KIFF11, ATOH7,

9 and RCBTB1 genes and EVR3 on chromosome 11p12-13 contribute to the pathogenesis of

10 FEVR. Mutations in the 11 genes and one locus currently associated with FEVR can explain

11 only approximately 50% of FEVR cases (26, 27), so we performed a whole-exome sequencing

12 (WES) analysis of 49 FEVR families who do not carry mutations in these genes. From this se-

13 quence analysis, we identified three heterozygous mutations in α-catenin (CTNNA1) of the cad-

14 herin/catenin complex associated with the phenotypes of FEVR. CTNNA1 is a core member of

15 the cadherin/catenin complex, which integrates adherens junctions (AJs) with the actin cytoskel-

16 eton and promotes intercellular adhesion (28, 29). CTNNA1 has been reported as a putative tu-

17 mor suppressor in myeloid leukemia (30), glioblastoma (31), skin cancer (32) and gastric cancer

18 (33, 34). Saskens et al. identified mutations in CTNNA1 that cause butterfly-shaped pigment dys-

19 trophy (35). Very recently, a missense CTNNA1 mutation was identified in an age-related macu-

20 lar degeneration (AMD) patient (36). Furthermore, Alexander et al. identified four mutations in

21 CTNNA1 that cause macular pattern dystrophy (37). These findings together introduced a new

22 perspective that CTNNA1 might affect retina development. Using a range of approaches, includ-

23 ing cell biological assays and genetically modified mouse models, we demonstrate in this paper

5 1 that the mutant forms of CTNNA1 identified in our WES analysis disrupt conformation of the

2 cadherin/catenin complex and induce inappropriate overactivation of the Norrin/β-catenin path-

3 way in vascular ECs that causes FEVR.

6 1 Results

2 WES of FEVR families identified mutations in CTNNA1

3 We used WES to identify potentially novel genetic mutations in 49 families (14 unaffected in-

4 dividuals and 86 individuals affected with FEVR) with an autosomal-dominant form of FEVR

5 without mutations in known FEVR genes. Variants identified by WES with a frequency of <

6 0.005 in the dbSNP138, Exome Variant Server, ExAC, and gnomAD databases and absent in

7 1,000 sequenced controls were filtered. These filtered variants were then annotated with ANNO-

8 VAR software (see Methods for detailed information) (Table S1). We selected genes that met all

9 three of the following standards as the top candidate genes for FEVR (38). First, candidate genes

10 harbored at least one “disruptive” variant (nonsense, frameshift, or in splice sites) (39). Second,

11 the variant was present in at least three families of FEVR. Third, the candidate gene was related

12 to Wnt signaling pathways or angiogenesis. CTNNA1 (Cadherin-Associated Protein, Alpha 1,

13 three families) was the best candidate gene that met our stringent criteria. From this analysis,

14 three heterozygous mutations in CTNNA1, a core component of the cadherin/catenin complex,

15 were discovered in three families: Family-3016, Family-3004, and Family-34 with FEVR (Fig-

16 ures 1, Tables S2 and S3).

17

18 For Family-3016 and Family-3004, we sequenced both of the affected family members us-

19 ing WES. After filtering coding variants with frequency MAF > 0.005 in public databases and an

20 in-house database, we ultimately got 51 rare variants in Family-3016 and 40 rare variants in

21 Family-3004, in which the CADD scores of 9 and 1 genes were higher than those of CTNNA1,

22 respectively. However, CTNNA1 was the only gene that met the candidate gene selection stand-

23 ard (Table S1). For Family-34, we sequenced the proband using WES and found 78 rare variants

7 1 after filtering coding variants with frequency MAF > 0.005 in public databases and an in-house

2 database. CTNNA1 p.R376Cfs*27 is a frameshift mutation that truncates the majority of the

3 CTNNA1 protein domains and most likely disrupts protein function (Table S1). Table S1 lists all

4 the rare variants in the three families. Finally, three heterozygous mutations in exon 3

5 (c.215T>C; p.F72S), exon 8 (c.1125_1131delCAGGGAC; p.R376Cfs*27), and exon 18

6 (c.2678C>T; p.P893L) of CTNNA1 were confirmed in patients with FEVR but not in unaffected

7 members of Families 3016, 34, and 3004 using Sanger sequencing analysis (Figure 1A). These

8 mutations affect amino acids that are highly conserved in the genes across different species (Fig-

9 ure 1A). The affected parents of the probands showed lack of peripheral vessels with neovascu-

10 larization and leakage in fundus fluorescein angiography (Figure 1C), however, no visible macu-

11 lar abnormality can be observed in fundus examination of the patients (Figure 1D). Notably, we

12 observed that the p.R376Cfs*27 mutation led to the most severe phenotype due to complete loss

13 of the CTNNA1 protein (Figure 1C and Table S2). However, we could not observe noteworthy

14 discrepancies in clinical phenotype between P893L and F72S mutations.

15

16 Mutant CTNNA1 caused overactivation of the β-catenin pathway

17 α-catenin (CTNNA1) binds to β-catenin (CTNNB1) adjacent to the membrane in AJs and to the

18 actin in the cytoplasma, whereas β-catenin (CTNNB1) binds to the cytoplasmic do-

19 main of VE-Cadherin (CDH5) to form a cadherin/catenin complex (40, 41). α-catenin and VE-

20 Cadherin also regulate Wnt signaling through their interactions with β-catenin by segregating the

21 available pool of β-catenin in the cytosol, thereby inhibiting the expression of downstream genes

22 (42-45). To assess the effect of the CTNNA1 mutations on the biological functions of α-catenin,

23 we used site-directed mutagenesis to introduce the three mutations identified in the FEVR

8 1 families into CTNNA1 expression plasmids. First, we asked whether the mutant α-catenin pro-

2 teins interacted normally with β-catenin. The co-immunoprecipitation results showed that the

3 F72S mutant form of α-catenin completely failed to interact with β-catenin whereas the P893L

4 mutant retained the interaction (Figure 2A). The frameshift mutation (p.R376Cfs*27) of

5 CTNNA1 resulted in nonsense-mediated decay, because no protein could be identified by West-

6 ern blot (Figure 2A). Therefore, two of the three α-catenin mutant proteins diminished their nor-

7 mal interactions with β-catenin. Next, we investigated the potential effects of the identified mu-

8 tant α-catenin protein on the Norrin/β-catenin pathway. To this end, we compared Nor-

9 rin/FZD4/LRP5/β-catenin signaling in the presence or absence of the wildtype (WT) and mutant

10 forms of CTNNA1 using the TopFlash reporter gene system, in which luciferase levels represent

11 signaling pathway activity. The expression of the WT CTNNA1 in HEK293 STF cells led to di-

12 minished luciferase activity compared to that in cells transfected with an empty vector (Figure

13 2B), confirming the inhibitory role of α-catenin on Norrin/β-catenin transcriptional activity (42-

14 45). However, the Norrin/β-catenin signaling pathway activity was significantly increased upon

15 expression of each of the mutant α-catenin proteins (Figure 2B). Compared to their WT protein,

16 the P893L mutant of CTNNA1 showed an approximately 1.3-fold increase in Norrin/β-catenin

17 transcriptional activity. Both F72S and the R376Cfs*27 mutation of CTNNA1 completely failed

18 to inhibit Norrin/β-catenin transcriptional activity, suggesting that the mutant α-catenin proteins

19 might fail to control the available pool of β-catenin in the cytosol.

20

21 To explore whether CTNNA1 mutants could affect endogenous β-catenin transcriptional activ-

22 ity, we performed adenovirus directed overexpression of WT or mutant CTNNA1 in human reti-

23 nal endothelial cells (HRECs). Quantitative real-time PCR (qPCR) analysis was applied to detect

9 1 changes in mRNA levels of β-catenin downstream genes, including CyclinD1 (CCND1), c-Myc

2 (MYC), Occludin (OCLN), and Claudin-5 (CLDN5). As shown in Figure S1, overexpression of

3 WT CTNNA1 significantly reduced the transcript levels of these genes compared to that trans-

4 fected with vector plasmids. Mutant forms of CTNNA1 significantly increased the transcript lev-

5 els of these genes compared to WT CTNNA1, consistent with the notion that CTNNA1 inhibits

6 β-catenin activity (Figure 2B).

7

8 To confirm the involvement of α-catenin in Norrin/β-catenin pathway regulation, we assayed

9 β-catenin activity using the TopFlash reporter gene system following the lentiviral-mediated

10 shRNA knockdown (KD) of CTNNA1 in HEK293 STF cells. The KD efficacy of CTNNA1 in

11 HEK 293STF cells was 87% (Figure 2C). We observed a sixfold increase of Norrin/β-catenin

12 transcriptional activity in CTNNA1 silenced cells compared to the activity measured in cells

13 treated with control shRNA (Figure 2D). Moreover, mRNA levels of endogenous β-catenin regu-

14 lated genes (CCND1, MYC, OCLN, and CLDN5) in CTNNA1 KD or CTNNB1 overexpressed

15 (OE) HRECs were detected using qPCR. Significant elevation in the mRNA levels of these

16 genes was observed in either CTNNA1 KD or CTNNB1 OE HRECs compared to controls (Figure

17 S2). Consistent with previous findings (43-45), these results provide support for the role of

18 CTNNA1 in inhibiting Norrin/β-catenin signaling. Hence, the dysfunction of CTNNA1 resulted in

19 overactivation of the Norrin/β-catenin pathway.

20

21 Dysfunction of CTNNA1 interrupts cell adherens junctions

22 We next transfected adenovirus containing mutant CTNNA1-F72S and CTNNA1-R893L into

23 HRECs to test whether this mutant allele interfered with normal cell assembly and the stability of

10 1 AJs in these retinal ECs. When transiently overexpressed in HRECs, the CTNNA1-P893L mutant

2 and WT proteins were correctly localized in the plasma membrane, whereas the CTNNA1-F72S

3 mutant protein was localized in the cytosol of HRECs. Expression of both CTNNA1-P893L and

4 CTNNA1-F72S led to disruption of F-ACTIN filament organization compared to the WT

5 CTNNA1 protein (Figure 2E). To assess the effect of CTNNA1 deficiency on AJs, we used the

6 lentivirus-mediated shRNA to KD CTNNA1 in HRECs. As shown in Figure 3A and B, CTNNA1

7 expression in KD cells was reduced to approximately 10% of that in control cells. Under normal

8 culture conditions, VE-Cadherin between ECs forms a confluent junction at the cell membrane

9 but is prominently reduced after CTNNA1 KD (Figure 3C). δ-catenin (also known as p120-

10 catenin or CTNND1, CTNND1 hereafter), a protein that binds to and suppresses VE-Cadherin

11 endocytosis, was also discontinuous and reduced on the plasma membrane following CTNNA1

12 KD (Figure 3D) (46). Moreover, the membrane expression of β-catenin was also reduced in

13 CTNNA1 KD cells (Figure 3E). This indicates that normal CTNNA1 levels on the cell membrane

14 are important for maintaining the stability of AJ proteins at cell junctions. In addition, severe dis-

15 organization of F-ACTIN was observed in the CTNNA1 KD cells compared to controls, further

16 supporting the hypothesis that dysfunction of CTNNA1 impairs both AJs and actin cytoskeleton.

17 Furthermore, we observed that KD of CDH5 or CTNND1 in the HRECs resulted in phenotypes

18 similar to those seen in CTNNA1 KD cells, suggesting the essential role of core components in

19 the cadherin/catenin complex (Figure 3).

20

21 Because disruption of the cadherin/catenin complex might release β-catenin from the cell

22 membrane and activate Wnt signaling, which regulates cell proliferation (43-45), we investigated

23 whether KD of CTNNA1 in HRECs could promote proliferation using a 5-ethynyl-2´-

11 1 deoxyuridine (EdU) incorporation assay. As expected, we found increased proliferation upon

2 CTNNA1 KD (Figure S3), which is consistent with studies that have demonstrated that CTNNA1

3 acts as a tumor suppressor and that the depletion of CTNNA1 leads to increased cell proliferation

4 in vitro (32, 34).

5

6 Loss of Ctnna1 in vascular ECs causes angiogenesis defects in mouse retinas

7 We then investigated whether loss of Ctnna1 in mouse vascular ECs could lead to phenotypes

8 of FEVR using genetic modified mouse models. Ctnna1 is essential for early development, and

9 its deletion in vascular ECs using conditional Ctnna1 knockout (KO) mice and Tie2-Cre trans-

10 genic mice resulted in embryonic lethality (Table S4). To overcome this problem, we generated

11 an inducible endothelial KO mouse model by crossing mice carrying the Ctnna1 flox allele with

12 mice carrying Pdgfb-iCreER, in which Cre is fused to ERT2 and expressed under the control of

13 the Pdgfb promoter. Ctnnalflox/flox, Pdgfb-iCre-ER, and Ctnnalflox/flox or Pdgfb-iCre-ER littermate

14 controls were induced by daily intraperitoneal (IP) injections of tamoxifen starting at postnatal

15 day 1 (P1) for four days. The specificity of Cre-mediated recombination was evaluated using the

16 tdTomato reporter (Figure S4A). After induction by tamoxifen, Ctnna1 was deleted in the ECs of

17 Ctnnalflox/flox, Pdgfb-iCre-ER mice (hereafter termed Ctnna1iECKO/iECKO), with the expression of

18 tdTomato restricted to ECs (Figure S4A). Tamoxifen-treated Ctnna1flox/flox or Pdgfb-iCre-ER

19 mice were used as WT control mice.

20

21 Following the administration of tamoxifen from P1 to P4, Ctnna1iECKO/iECKO mice were smaller

22 in body size, with approximately 22% decreased body weight (Figure S4B and C), and failed to

23 thrive. Most of them died at P9 (Table S5), indicating the blunted overall body development

12 1 upon complete depletion of Ctnna1 in mice ECs. The external appearance of heterozygous

2 Ctnna1iECKO/+ mice was similar to that of their WT littermates (Figure S4B). Areas of bleeding

3 were apparent in the eyeballs of Ctnna1iECKO/iECKO mice (Figure S4D), a phenomenon also ob-

4 served in FEVR patients (Figure 1C), indicating that Ctnna1 plays a role in retinal vascular de-

5 velopment. In addition, the eyeballs of Ctnna1iECKO/iECKO mice were smaller than those of their

6 littermate controls, indicating that the development of the retinal vasculature may affect eye size

7 (Figure S4D and E). Bright-field imaging of the retinas and hyaloid vessels of Ctnna1 mutant

8 mice revealed abnormal, blood-red vessel structures (Figure S4D). The hyaloid vasculature of

9 Ctnna1iECKO/iECKO mice showed a phenotype similar to that of Norrin-/- (47), Lrp5-/- (48), Tspan12-

10 /- (7), and Fzd4-/- (49) mice, characterized by its slower regression at P9 compared to littermate

11 controls (Figure 4A). This data suggested that the loss of Ctnna1 might disrupt normal apoptosis-

12 mediated regression of hyaloid vessels.

13

14 Flat-mounted retinas obtained from Ctnna1iECKO/iECKO and control mice were evaluated using

15 Isolectin B4 (IB4), which labels blood vessel cells. At P3, the horizontal growth of blood vessels

16 was slower in Ctnna1iECKO/iECKO mice than that in WT controls (Figure S4F and G). At P9, the

17 superficial vessels of Ctnna1iECKO/iECKO mice were significantly enlarged during development, ac-

18 companied by delayed horizontal outgrowth and hyperplasia (Figures 4B–D, G, H, and S4H). In-

19 terestingly, unlike Fzd4+/- (49), Lrp5+/- (48), and Tspan12+/- (7) mice, the vasculature of heterozy-

20 gous Ctnna1iECKO/+ retinas also showed moderately delayed radial outgrowth, indicating that an-

21 giogenesis is more sensitive to Ctnna1 dosage than to the dosages of the above genes (Figure 4B,

22 G). This is consistent with the fact that heterozygous CTNNA1 mutations caused FEVR in hu-

23 mans. To further characterize the above-mentioned phenotypes, we applied an

13 1 immunofluorescence staining assay on frozen retina sections. In control mice, we found that ver-

2 tical branches formed from the primary vasculature at P9 and subsequently established capillar-

3 ies in the outer (OPL) and inner (IPL) plexiform layers (Figures 4C and S4H, left panels). How-

4 ever, the loss of one copy of Ctnna1 delayed vertical growth of the superficial retina vascular

5 plexus (Figures 4C and S4H, middle panels). Both Ctnna1iECKO/iECKO and Ctnna1iECKO/+ retinas

6 exhibited obvious defects in vertical vascular growth into the deeper retinal layers and hyper-

7 plasia of the vascular plexus, and both also lacked vertical secondary and tertiary vessels at P9

8 (Figures 4C and S4H, right panels).

9

10 Given that adult Fzd4-/-, Ndp- and Tspan12-/- retinas completely lack deep vessel layers (7,

11 25), we asked whether adult mice have similar phenotype upon homozygous or heterozygous de-

12 pletion of Ctnna1 in ECs. Because Ctnna1iECKO/iECKO mice died at ~P9 when tamoxifen was in-

13 jected from P1, we performed tamoxifen injection from P6 and found that Ctnna1iECKO/iECKO

14 mice died at ~P13 to P14 (Table S6). It was previously reported that mouse retinal vasculature

15 develops through three stages: superficial radial growth from P1 to P9; deep vessel plexus in the

16 OPL from P7 to P12; and intermediate vessel plexus in the IPL from P11 to P17 (50). Thus, we

17 chose P13 as a checkpoint for deep vessel development. Interestingly, Ctnna1iECKO/iECKO mice

18 (tamoxifen induction from P6) showed abnormality only at the periphery of the superficial vas-

19 culature at P13 (Figure S5), suggesting that CTNNA1 might mostly affect tip cell vessel devel-

20 opment rather than stalk cell vessel stability. Notably, at P13 we found fewer vessels in the OPL

21 and minimal vessels in the IPL of Ctnna1iECKO/iECKO retinas compared to that observed in control

22 retinas (Figure S5B and C). However, at P13, Ctnna1iECKO/+ mice showed only mild delay in ves-

23 sel development in the IPL of the retina compared to control retinas (Figure S6). Therefore, loss

14 1 of Ctnna1 in mouse retinal vascular cells led to angiogenesis defects and partially reproduced

2 clinical features seen in FEVR patients.

3 In addition, VE-Cadherin also appeared diffuse in Ctnna1iECKO/iECKO mice retinas (Figure 4D),

4 further supporting a role of CTNNA1 in the stability of catenin/cadherin complex and normal

5 retinal angiogenesis. In Ctnna1iECKO/iECKO retinas, the staining of a small molecule Evans blue

6 (MW = 961 g/mol) and red blood cells using a Ter119 antibody revealed 50% increase of leak-

7 age along blood vessels (Figure 4E, F and I), providing evidence of exudative bleeding from reti-

8 nal vessels. Furthermore, electron microscope ultrastructural analysis also exhibited discontinu-

9 ous distribution of EC-EC tight junction (Figure S7). Besides, glial fibrillary acidic protein

10 (GFAP) accumulation was observed in Ctnna1iECKO/iECKO retinas (Figure S8A), resembling the

11 phenotype of Fz4-/- mice (24). This accumulation is indicative of activated glial cells and high

12 retinal stress (24). The tip cell marker Esm1 was also observed in the remodeling plexus of

13 Ctnna1iECKO/iECKO retinas (Figure S8B), indicating an abundance of tip over stalk cells in the

14 plexus (51). The tight junction protein Claudin-5, which is also a Wnt downstream target, was

15 increased with an abnormal distribution pattern in Ctnna1iECKO/iECKO retinas (Figure S8C). There-

16 fore, loss of Ctnna1 led to leakage in retinal vessels, similar to the symptoms observed in FEVR

17 patients.

18

19 Loss of Ctnna1 leads to overproliferation of retinal vessels in mouse retinas

20 To further investigate the consequences of loss of Ctnna1 in retinal vascular vessels, we ana-

21 lyzed the mitotic proliferation of vascular ECs using an EdU incorporation assay in WT control

22 versus Ctnna1iECKO/iECKO and Ctnna1iECKO/+ retinas at P5 (Figure 5). In P5 Ctnna1iECKO/iECKO reti-

23 nas, the EdU-labeled proliferating ECs were significantly increased compared to those in WT

15 1 control and Ctnna1iECKO/+retinas (Figure 5A and B), which is similar to the observed overprolif-

2 eration in CTNNA1 KD HRECs (Figure S3). These results indicated that uncontrolled prolifera-

3 tion might contribute to vascular remodeling and vessel hyperplasia. We further compared the

4 retinal vascularization of P5 Ctnna1iECKO/iECKO mice with that of Lrp5 or Fzd4 KO mice (a model

5 of FEVR with inactivation of β-catenin signaling) and observed very similar phenotypes, such as

6 delayed superficial vascular progression, neovascularization at the peripheral region, and exten-

7 sive erythrocyte leakage (Figure 6). However, the vessel density of Ctnna1iECKO/iECKO retinas was

8 evidently increased, whereas it was decreased in Lrp5 or Fzd4 KO mice retinas compared to WT

9 controls (Figure 6C). EC proliferation is induced by vascular endothelial growth factor A

10 (VEGF-A) (52). We examined the abundance of Vegfa (Vegf164) in Ctnna1iECKO/iECKO and con-

11 trol retinas. In the control retinas, Vegfa is expressed ahead of the angiogenic front (Figure 7A,

12 left panel) and very low abundance of Vegfa can be detected in the remodeling plexus. However,

13 in Ctnna1iECKO/iECKO retinas, abnormal expression of Vegfa was observed around the angiogenic

14 front and in the remodeling plexus (Figure 7A, right panel). These data indicated that Vegfa was

15 induced by the remodeling plexus after Ctnna1 deletion. In contrast, no difference in Vegfa

16 abundance was observed around the angiogenic front and in the remodeling plexus in Lrp5 and

17 Fzd4 KO retinas (Figure 7B and C). During sprouting angiogenesis, the behaviors of tip cells at

18 the front and trailing stalk cells are controlled by VEGFR2 and Notch signaling (53). Upregula-

19 tion of DLL4 was observed in Ctnna1iECKO/iECKO retinas (Figure S9). In contrast, diminished

20 DLL4 expression can be observed in the vasculature of Lrp5 and Fzd4 KO retinas (Figure S9).

21 These data suggested that the molecular events underlying angiogenesis defects in

22 Ctnna1iECKO/iECKO and Lrp5 or Fzd4 KO retinas are distinct.

23

16 1 Loss of Ctnna1 disrupts the integrity of the blood–brain barrier in the mouse brain

2 The integrity of the blood–brain barrier (BBB) is important for central nervous system func-

3 tion. We therefore tested whether the loss of Ctnna1 damaged BBB. Obvious bleeding regions

4 were observed in the brains of Ctnna1iECKO/iECKO mice (Figure S10A), most prominently in the

5 cerebellum, implying the occurrence of blood vessel leakage in the brain. To investigate whether

6 BBB integrity was disrupted during postnatal angiogenesis in Ctnna1iECKO/iECKO mice, we in-

7 jected them with Evans blue intraperitoneally at P8, 24 h before sacrifice. As predicted, exten-

8 sive leakage of Evans blue was observed in the whole brains of Ctnna1iECKO/iECKO mice (Figure

9 S10A), indicating a disruption of the entire BBB. Hematoxylin-and-eosin-stained paraffin sec-

10 tions of Ctnna1iECKO/iECKO mice brains were also examined (Figure S10B), and the staining re-

11 vealed that the deletion of Ctnna1 resulted in an altered morphology of the cerebellum compared

12 to that of littermate controls. Regions of hemorrhage were also observed in the Ctnna1iECKO/iECKO

13 cerebellum (Figure S10B). Next, we imaged the cerebellum in P9 animals using contrast-en-

14 hanced X-ray micro-computed tomography (micro-CT) to examine abnormal blood vessels. Sag-

15 ittal projections of the Ctnna1iECKO/iECKO and control cerebellums are shown in Figure S10C.

16 Compared to WT controls, the Ctnna1iECKO/iECKO cerebellum showed additional high-intensity

17 areas, which were reconstituted in 3D and shown to be widely distributed, abnormal tube-like

18 structures (Figures S10D and S11A). Because blood-filled lesions were imaged as high-intensity

19 areas, we concluded that they represented enlarged blood vessels in the Ctnna1iECKO/iECKO cere-

20 bellum, whereas low-intensity holes are considered edematous lesions (54). The presence of pos-

21 itive Ter119 signals outside of the vessels on frozen sections of the Ctnna1iECKO/iECKO cerebellum

22 suggested extensive leakage of erythrocytes from blood vessels (Figures S10E-F and S11B). In

23 addition, enlarged blood vessels and edema-like cavities were observed in these cerebellum

17 1 sections, consistent with the micro-CT observations of the Ctnna1iECKO/iECKO cerebellum (Figure

2 S10D). Brain glial cells were also activated, as reflected by the increased levels of GFAP expres-

3 sion in the Ctnna1iECKO/iECKO cerebellum (Figure S11B). Furthermore, electron microscope ultra-

4 structural analysis confirmed junctional defects in vessels of Ctnna1iECKO/iECKO cerebellum, which

5 included discontinuous distribution of electron-dense tight junctions or leakage of erythrocytes

6 out of vessels (Figure S12). Thus, the loss of Ctnna1 in mice disrupts the integrity of the BBB

7 and impairs the cerebellum’s normal structure and functions, similar to the phenotypes observed

8 in the Fzd4-/- mice (24).

9

10 Loss of Cdh5 in retinal vascular ECs causes angiogenesis defects in mouse retinas

11 Because knockdown of CTNNA1 in ECs reduced the abundance of membrane CDH5 (Figure

12 3), which is a core component of the cadherin/catenin complex and was previously reported to

13 play an important role in retinal vascular development, we analyzed angiogenesis in detail by

14 breeding mice carrying a loxP-flanked Cdh5 allele with tamoxifen-inducible Pdgfb-iCre trans-

15 genic mice to generate Cdh5loxp/loxp, Pdgfb-iCreER (named Cdh5iECKO/iECKO ) (55, 56). We found

16 extensive erythrocyte leakage as well as delayed superficial outgrowth and local hyperdensity of

17 the retinal vasculature of Cdh5iECKO/iECKO mice (Figure S13A), consistent with the phenotype pre-

18 viously reported and closely resembling the phenotype in Ctnna1iECKO/iECKO mouse retinas (55,

19 56). Furthermore, we observed delayed regression of hyaloid vessels in Cdh5iECKO/iECKO mice

20 (Figure S13B), suggesting that disruption of the cadherin/catenin complex might cause FEVR.

21 CDH5 plays an inhibitory role in Norrin/β-catenin signaling

22 We then asked whether CDH5 inhibits Norrin/β-catenin signaling activity as CTNNA1 does.

23 A TopFlash reporter gene assay was applied to demonstrate the effect of CDH5 on Norrin/β-

18 1 catenin transcriptional activity. Overexpression of CDH5 in HEK293 STF cells downregulated

2 the Norrin/β-catenin signaling activity by approximately 70% compared to that in cells trans-

3 fected with an empty vector (Figure S13C). Moreover, after knocking down CDH5 (82% reduc-

4 tion in CDH5 transcriptional level), a 3.5-fold increase in Norrin/β-catenin transcriptional activ-

5 ity was observed in HEK293 STF cells compared to the control cells (Figures S13D and E).

6 These results indicated that CDH5 plays a role similar to CTNNA1 in regulating Norrin/β-catenin

7 signaling activity.

8

9 Loss of Ctnna1 or Cdh5 in isolated mouse lung endothelial cells interrupts cell adherens

10 junctions

11 Because knockdown of CTNNA1 or CDH5 in HRECs disrupts cell AJ proteins (Figure 3), we

12 examined whether the same phenotype could be observed in isolated Ctnna1iECKO/iECKO or

13 Cdh5iECKO/iECKO mouse lung endothelial cells (MLECs). As Figure S14A shows, deletion of en-

14 dothelial Ctnna1 resulted in discontinuous distribution of VE-Cadherin, CTNND1, and

15 CTNNB1. The membrane-bound CTNNA1 protein level was reduced to 11% (Figure S14A and

16 B), indicating efficient Ctnna1 deletion in MLECs. Membrane-bound VE-Cadherin, CTNND1,

17 and CTNNB1 protein levels in Ctnna1iECKO/iECKO MLECs were reduced to 44%, 31%, and 38%,

18 respectively. Similarly, deletion of endothelial Cdh5 resulted in discontinuous distribution of

19 VE-Cadherin, CTNND1, and β-catenin (Figure S14A and C). The membrane-bound protein level

20 of VE-Cadherin was reduced to 19% (Figure S14C), also suggesting efficient Cdh5 excision.

21 Membrane-bound CTNNA1, CTNND1, and CTNNB1 protein levels in Cdh5iECKO/iECKO MLECs

22 were reduced to 48%, 36%, and 35%, respectively. Moreover, as in a previous report (57), we

19 1 found that the F-ACTIN arrangement was also disrupted in Ctnna1- and Cdh5-deficient MLECs

2 (Figure S14A, right panel).

3

4 Rescue effect of CTNNA1 mutants on and ACTIN filament of isolated

5 Ctnna1-deficient MLECs

6 Next, we investigated whether overexpression of WT or CTNNA1 mutants could rescue the

7 disrupted AJ and ACTIN filament in Ctnna1-deficient MLECs. As Figure S15 shows, the ar-

8 rangement of F-ACTIN was well organized in CTNNA1-WT transfected Ctnna1-deficient

9 MLECs, whereas CTNNA1-P893L and F72S mutant proteins failed to rescue the disrupted F-

10 ACTIN arrangement. Moreover, CTNNA1-WT protein was able to restore disrupted VE-Cad-

11 herin and CTNNB1 distribution compared to vector control. CTNNA1-P893L exhibited a com-

12 promised ability to rescue compared to CTNNA1-WT and vector control. Notably, CTNNA1-

13 F72S completely failed to rescue the discontinuous distribution of VE-Cadherin and CTNNB1

14 (Figure S15). To explore whether mutations in CTNNA1 could affect CTNNB1 nuclear translo-

15 cation, we activated Norrin/β-catenin signaling by adding recombinant Norrin. We found that

16 CTNNA1-WT protein significantly inhibited nuclear translocation of CTNNB1 compared to vec-

17 tor, whereas mutant forms of CTNNA1 exhibited a compromised inhibitory role in nuclear trans-

18 location of CTNNB1 compared to WT CTNNA1 (Figure S16). These data provided additional

19 evidence that CTNNA1 mutants compromised AJs.

20

21 F72S mutant allele of Ctnna1 combined with the deletion allele of Ctnna1 phenocopies

22 Ctnna1iECKO/iECKO mice

20 1 To test whether mutation of Ctnna1 in mice could cause a phenotype similar to that of

2 Ctnna1iECKO/iECKO mice, we generated mice carrying the Ctnna1-F72S point mutation. Heterozy-

3 gous mutant mice (Ctnna1F72S/+) were intercrossed to generate homozygous mice, however, we

4 found that the homozygous F72S mutation in the Ctnna1 gene causes embryonic lethality (Table

5 S7). Thus, we examined whether Ctnna1iECKO/F72S compound heterozygous mice exhibited a reti-

6 nal vascular phenotype similar to that of Ctnna1iECKO/iECKO mice. Ctnna1F72S/+ mice were bred

7 with Ctnnalflox/+; Pdgfb-iCre-ER mice (Ctnna1iECKO/+) to generate Ctnna1+/+; Pdgfb-iCre-ER

8 (Ctrl), Ctnna1F72S/+, Ctnnalflox/+; Pdgfb-iCre-ER (Ctnna1iECKO/+) and Ctnna1F72S/flox; Pdgfb-iCre-

9 ER (Ctnna1F72S/iECKO). Confocal images showed that the F72S heterozygous mutation could not

10 cause a significant change in vascular development compared to control littermates, whereas the

11 retinal vasculature of Ctnna1F72S/iECKO compound heterozygous mice showed retarded growth of

12 horizontal vessels and vessel leakage (Figure 8). This phenotype is similar to that seen in

13 Ctnna1iECKO/iECKO mice (Figure 4) and provides another piece of evidence on the pathogenicity of

14 the CTNNA1-F72S mutation.

15

16 Overactivation of β-catenin signaling results in angiogenesis defects in mouse retinas

17 Our in vitro cell biology analysis demonstrated that FEVR-associated mutant α-catenin pro-

18 teins and knockdown of CTNNA1 led to increased Norrin/β-catenin activity. These findings sug-

19 gest that the FEVR phenotype caused by loss of CTNNA1 function might be partially due to ab-

20 normal activation of the Norrin/β-catenin pathway. To investigate whether increased Norrin/β-

21 catenin activity could result in angiogenesis defects, we generated an EC-specific gain-of-func-

22 tion (GOF) allele of Ctnnb1. In this allele, exon 3 of Ctnnb1 was floxed by two loxP sites

23 (Ctnnb1foxedExon3) (Figure S17A). When this allele is crossed on to a Cre-expressing strain, exon

21 1 3 of Ctnnb1 is deleted in Cre-expressing tissue, abolishing the GSK3β binding site of β-catenin

2 (Figure S17A). This mutant β-catenin protein cannot be degraded by the GSK3β complex, and β-

3 catenin remains activated to drive the expression of its downstream genes. As predicted, we ob-

4 served accumulation of β-catenin protein (220% of control) in Ctnnb1foxedExon3/foxedExon3; Pdgfb-

5 iCre-ER (hereafter named Ctnnb1 GOF Homo) mouse retinal ECs (Figure S17B and C). The ex-

6 pression of this Ctnnb1 GOF allele in mouse ECs resulted in retarded superior retinal blood ves-

7 sel growth and defects in vertical vascular growth into the deeper retinal layers, with secondary

8 and tertiary vessels absent at P9 (Figure 9), which is similar to some clinical features of FEVR

9 (3-5, 7, 12, 24, 25). However, the retinal vasculature of P13 Ctnnb1 GOF homo mice only

10 showed mild retarded growth into deep vessel layers (Figure S18), indicating that overactivation

11 of β-catenin signaling could partially affect vessel development. Furthermore, similar to that in

12 Ctnna1iECKO/iECKO, upregulation of Vegfa and DLL4 was also observed in Ctnnb1 GOF homo ret-

13 inas (Figure 7D and S9). These findings demonstrated that β-catenin activity has to be main-

14 tained in a precise range and that increased β-catenin activity can disrupt angiogenesis in the ret-

15 ina (3-5, 7, 12, 24, 25).

16

17 During our WES analysis, we also identified a missense mutation (p.P848L) in the LRP5

18 gene, which leads to a twofold increase of Norrin/β-catenin signaling luciferase activity, in an

19 Indian family with FEVR (Figure S19). We then generated an Lrp5P847L knock-in mouse corre-

20 sponding to the LRP5-848L human mutation to examine its effect on retinal angiogenesis. The

21 Lrp5 knock-in allele led to retarded angiogenesis, as seen in the Ctnnb1 GOF Homo mice, and

22 similar to phenotypes observed in FEVR patients (Figures 10 and S20). This piece of evidence

22 1 provides further support for the role of increased β-catenin activity in FEVR pathogenesis with-

2 out affecting maintenance of barrier function.

3

23 1 Discussion 2 It is established that FEVR in around half of patients is associated with mutations in NDP,

3 FZD4, LRP5, and TSPAN12, which encode components of the Norrin/β-catenin pathway (3-5, 7,

4 12, 24, 25) through inactivation of Norrin/β-catenin signaling. In this study, we demonstrated

5 that mutations of CTNNA1 in the cadherin/catenin complex cause FEVR through overactivation

6 of Norrin/β-catenin signaling and interruption of cell junctions, suggesting that the cad-

7 herin/catenin complex plays an important role in the pathogenesis of FEVR. Mutations of

8 CTNNA1 result in reduced sequestration of β-catenin in the cytosol and increased activity of Nor-

9 rin/β-catenin signaling. Consistent with this finding, mutations in CTNNB1, which encodes a

10 core member of the cadherin/catenin complex and functions as a key effector in the Norrin/β-

11 catenin pathway, were recently reported to be associated with FEVR, further supporting a role

12 for the cadherin/catenin complex in the etiology of these retinal diseases (11, 12). Two types of

13 mutations have recently been identified in CTNNB1, one that decreases β-catenin activity and

14 one that increases β-catenin activity (12), indicating that the transcriptional activity of CTNNB1

15 has to be maintained within a narrow range. These findings, together with our results in this

16 study that transcriptional activity of β-catenin protein has to be precisely regulated within a nar-

17 row window to maintain proper vessel development, suggesting that β-catenin plays a central

18 role in merging the cadherin/β-catenin and Norrin/β-catenin pathways, making it a pivotal pro-

19 tein in the pathogenesis of FEVR.

20

21 Previous studies have indicated that dysfunction of CTNNA1 could result in cancer and devel-

22 opmental diseases (58). However, we did not found cancer in our affected patients. Following

23 ACMG guideline (59, 60), we referred the patient with p.R376Cfs*27 to genetic counseling. Re-

24 cently, Saksens et al. identified three heterozygous missense mutations in CTNNA1 (c.953T>C,

24 1 p.Leu318Ser; c.1293T>G, p.Ile431Met and c.919G>A, p.Glu307Lys) in patients with butterfly

2 macular dystrophy (35). Notably, another CTNNA1 variant (c.160C>T; p.Arg54Cys) was identi-

3 fied in a patient affected with butterfly-shaped pigment dystrophy (35) and in an AMD patient

4 without classical phenotype of butterfly-shaped pigment dystrophy (36). The pathogenicity of

5 this variant remains unclear. Lately, Alexander et al. (37) identified four missense mutations in

6 CTNNA1 (c.965C>T, p.Ser322Leu; c.1316C>T, p.Ser439Phe; c.1294G>A, p.Glu432Lys and

7 c.973A>G, p.Thr325Ala) in patients with macular pattern dystrophy. These identified variants

8 were predicted to be disease-causing by SIFT, affect a residue that is completely conserved

9 among vertebrate species. A chemically induced mouse model Ctnna1tvrm5 exhibited similar pig-

10 mentary abnormalities, focal thickening, and elevated lesions, decreased light-activated ERG re-

11 sponses and loss of photoreceptor cells, and carried a homozygous missense mutation Ctnna1

12 p.Leu436Pro (35). According to an X-ray crystallographic structural model of the human

13 CTNNA1 protein (, PDB: 4IGG) (Figure S22A) (61), residues affected by

14 those four variants associated with butterfly macular dystrophy (p.Glu307, p.Leu318, p.Ile431,

15 and p.Leu436) (35) and four variants with macular pattern dystrophy (p.Ser322, p.Ser439,

16 p.Glu432, and Thr325) (37) are located closely in force-sensing module (domains D3–D4, amino

17 acid residues 260–630) and in protein binding domains (D3a, residues 260–400; D3b, residues

18 400–507), suggesting that mutations in these domains might lead to macular dystrophy. The two

19 missense CTNNA1 mutations associated with FEVR identified in our study are located near dis-

20 tinct helix bundles (p.Phe72 in the N-terminal helix and p.Pro893 in the C-terminal helix, respec-

21 tively) in the crystal structure and these two helix bundles (hereafter termed helix-N and helix-C)

22 are spatially close (Figure S22A). Furthermore, structural analysis of CTNNA1 asymmetric di-

23 mers (chain A and chain B) revealed that the helix-N of chain A was spatially closer to the helix-

25 1 C of chain B than to the helix-N of chain B (Figure S22B), indicating that Phe72 and Pro893 res-

2 idues might participate in similar functions. This was also supported by our data that both vari-

3 ants affected F-actin orientation (Figure 2E and S15) and partially lost the ability to inhibit beta-

4 catenin signaling (Figure 2B, S1 and S16). It is plausible that different CTNNA1 variants have

5 different impacts on the protein functions and lead to distinct disease conditions. Another similar

6 example is the FEVR disease-causing gene LRP5. Mutations in LRP5 resulted in osteopenia and

7 osteoporosis (18), whereas distinct mutations in LRP5 also caused FEVR (4).

8

9 The knockdown of CTNNA1 in HREC resulted in dissociated AJ proteins of the cad-

10 herin/catenin complex on the cell membrane (Figure 3), such as VE-Cadherin (CDH5), β-catenin

11 (CTNNB1), and δ-catenin (CTNND1). F-ACTIN was disorganized in HRECs expressing the mu-

12 tant CTNNA1 proteins and in CTNNA1 KD cells, indicating that functional disruption of

13 CTNNA1 damages the actin cytoskeleton network. Although CTNNA1-P893L might not affect

14 its binding with CTNNB1 and cell localization as CTNNA1-F72S does (Figure 2A), this mutant

15 protein indeed disrupts the F-ACTIN network (Figure 2E) and activates β-catenin signaling com-

16 pared to WT CTNNA1 (Figure 2B), suggesting that the F-ACTIN network might play a role in

17 retinal vascular development. In addition, in our rescue experiment using adenovirus, both

18 CTNNA1-P893L and F72S mutant proteins failed to rescue the disrupted F-ACTIN arrangement,

19 VE-Cadherin and CTNNB1 distribution in isolated MLECs (Figure S15 and S16). However, the

20 exact mechanism by which P893L mutation causes FEVR remains elusive and needs further in-

21 vestigation.

22

23 The crucial role of CTNNA1 in the pathogenesis of FEVR is strongly supported by the retinal

24 blood vessel phenotypes observed in the Ctnna1 EC-specific KO mouse model. Both

26 1 Ctnna1iECKO/+ and Ctnna1iECKO/iECKO mice showed incomplete vascularization of the peripheral

2 retina, absence of the second and tertiary capillary layers in the deep retina, delayed hyaloid vas-

3 culature regression, and vascular leakage in the retina (Figure 4), reproducing the main clinical

4 features of FEVR patients. It is true that the Ctnnal iECKO/+ model exhibited only a milder pheno-

5 type than what was observed in FEVR patients. This might be due to a species difference. An-

6 other similar example is the Lrp5 or Fzd4 KO model for FEVR. Lrp5 or Fzd4 heterozygous KO

7 mice did not show any visible defect in angiogenesis, but homozygous KO mice exhibited

8 FEVR-like phenotypes (25, 49). In contrast to that of Lrp5 or Fzd4 KO models with inactivation

9 of β-catenin signaling, loss of Ctnna1 in Ctnna1iECKO/iECKO resulted in overactivation of β-catenin

10 signaling, disruption of Vegfa gradient (Figure 7), and drastic upregulation of notch ligand DLL4

11 (Figure S9). Abnormal distribution of Vegfa in the remodeling plexus might result in high den-

12 sity of blood vessels in retina. Consequently, both CTNNA1 KD HRECs and Ctnna1iECKO/iECKO

13 exhibited increased EC proliferation, possibly induced by overactivation of β-catenin signaling

14 and vascular-leakage-associated upregulation of Vegfa (Figure 5, 7 and S3). DLL4 plays an in-

15 hibitory role of tip cell formation and promotes stalk cell differentiation (62). A high concentra-

16 tion of DLL4 might inhibit radial growth of the retinal vasculature (62, 63). Constitutive activa-

17 tion of β-catenin in Ctnnb1 GOF also led to upregulation of notch ligand DLL4 (Figure S9).

18 These results demonstrate that the transcriptional activity of β-catenin protein has to be precisely

19 regulated within a narrow window to maintain proper vessel development.

20

21 Ctnna1iECKO retinas exhibited increased proliferation and delayed vascular expansion. This

22 paradox can be explained by the following two factors. First, tip cell migration depends on nor-

23 mal Vegfa gradient (52). Loss of Ctnna1 leads to activation of β-catenin signaling and disruption

27 1 of normal Vegfa gradient, because increased expression of Vegfa was observed in both the re-

2 modeling plexus and angiogenic front of the retina (Figure 7A). This abnormal Vegfa gradient

3 resulted in excessive angiogenesis in the remodeling plexus rather than vascular expansion. Sec-

4 ond, binding of CTNNA1 with was previously reported to be crucial to maintaining mi-

5 gration polarity in ECs (64), and loss of Ctnna1 in mouse ECs might impair collective polarity,

6 leading to deficient outward angiogenesis.

7

8 The cadherin/catenin complex plays a critical role in cellular AJs and maintains the integrity

9 of cell–cell connections (65). In the cadherin/catenin complex, VE-Cadherin interacts with sev-

10 eral binding partners, such as α-catenin, β-catenin, and p120-catenin, to maintain normal cell–

11 cell adheren and dynamics (66). Dysfunction of AJs might result in vascular diseases (67), and

12 indeed, VE-Cadherin/β-catenin signaling controls vascular EC survival (42). A recent study by

13 Yamamoto et al., combined with our work (Figure S13), demonstrated that hemorrhaging in the

14 peripheral retina, incomplete retina vascular development, defective sprouting into the deeper

15 retina, and local hyperdensity of the retinal vasculature near the angiogenic front were exhibited

16 in Cdh5iECKO/iECKO mutant mice (55, 56). These retinal vascular development defects reproduce

17 the main clinical phenotypes of FEVR, supporting the hypothesis that disruption of the cad-

18 herin/catenin complex causes FEVR.

19

20 In contrast to the previously reported role of NDP, FZD4, LRP5, and TSPAN12 mutations in

21 decreasing Norrin/β-catenin signaling (3-5, 7, 12, 24, 25), FEVR-causing CTNNA1 mutations

22 overactivate Norrin/β-catenin signaling. In addition, it is interesting that we identified a FEVR-

23 causing LRP5 mutation with increased Norrin/β-catenin transcriptional activity, in contrast to

28 1 those loss-of-function LRP5 mutations identified in patients with FEVR (4, 5). Consistent with

2 this finding, a recent study reported that a FEVR-associated CTNNB1 mutation (p.R710C) in-

3 creased the transcriptional activity of Norrin/β-catenin as well as a loss of function mutation

4 (p.H720*) (12), suggesting that that overactivation of Norrin/β-catenin is associated with FEVR.

5 In addition, overactivation of Norrin/β-catenin signaling by the overexpression of Norrin in

6 transgenic mice disrupted embryonic angiogenesis (25), further supporting our hypothesis that

7 overactivation of Norrin/β-catenin signaling could cause defective angiogenesis. Importantly, we

8 demonstrated that mice expressing an EC-specific GOF of the Ctnnb1 allele exhibited main

9 FEVR phenotypes, including an incomplete vascularization of the peripheral retina and lack of

10 the second and tertiary capillary layers in the deep retina (Figure 9). It thus appears that precisely

11 controlled β-catenin activation is critical for normal retinal vascular development. Both the loss

12 of β-catenin activity, caused by mutations in the NDP/FZD4/LRP5-/TSPAN12/CTNNB1 path-

13 way, and the overactivation of β-catenin activity via mutations in CTNNA1/CTNNB1 complex

14 are associated with FEVR. Notably, the developmental of vascular defects in the GOF allele of

15 Ctnnb1 and Lrp5P847L/P847L mice was much milder than that observed in Ctnna1iECKO/iECKO mice

16 (Figures 4, 9, 10, S5, S18, and S 21), due to the possibility that overactivation of beta-catenin has

17 a limited effect on vascular development.

18

19 Based on our observation of AJ defects among CTNNA1-depleted HRECs, we proposed that

20 dysfunction of the cadherin/catenin complex and excessive β-catenin transcriptional activity

21 could cause FEVR pathogenesis. In fact, similar to the phenotypes of Fzd4 and Lrp5 homozy-

22 gous KO mutants, loss of Ctnna1 in ECs resulted in abnormal proliferation of EC (Figure 6), ex-

23 tensive leakage in retinal vessels (Figure 4F and I), retarded radical growth, lack of capillary

29 1 layers in the deep retina, and delayed hyaloid vasculature regression, which were also observed

2 in the retinas of mice with the GOF allele of Ctnnb1 and Lrp5P847L (Figures 4, 9, and 10). Thus,

3 this study not only provides new insight into the pathogenesis and therapeutic targets for FEVR,

4 but also sheds light on pathogenesis clues for other blindness diseases accompanied by neovas-

5 cularization. A future challenge will be to determine precisely how altered cadherin/catenin com-

6 plex function and overactivated Norrin/β-catenin independently contributes to the pathogenesis

7 of FEVR and other neovascularization blindness diseases.

8

9 Taken together, these findings demonstrate that dysfunction of the cadherin/catenin complex

10 and dysregulation of Norrin/β-catenin transcriptional activity is of key importance for the devel-

11 opment of FEVR and potentially other neovascular diseases. Of course, our finding is tentative

12 and would benefit from replication in other cohorts/FEVR families.

13 14

30 1 Methods

2 A full description of the methods is presented in the online-only Data Supplement.

3 Whole-exome sequencing

4 Exome sequencing was performed on DNA samples of the index patients. The raw sequence

5 data reported in this paper have been deposited in the Genome Sequence Archive in National Ge-

6 nomics Data Center (68), Beijing Institute of Genomics (China National Center for Bioinfor-

7 mation), Chinese Academy of Sciences, under accession number HRA000554 that are publicly

8 accessible at https://bigd.big.ac.cn/gsa.

9 Experimental animals

10 All animal protocols were approved by the Animal Care and Use Committee of Sichuan Pro-

11 vincial People’s Hospital. All experimental procedures and methods were performed in accord-

12 ance with the approved study protocols and relevant regulations. All mice were on the C57BL/6

13 background. Mice were housed under a 12:12-hour light: dark cycle at 25°C and had unrestricted

14 access to food and water.

15 Floxed Ctnnal mice were obtained from the Jackson Laboratory (stock number 004604;

16 https://www.jax.org/strain/004604) (69). Ctnna1flox/flox mice were mated to Pdgfb-iCre transgenic

17 mice (70) to generate Ctnna1iECKO/iECKO mice with inducible alleles to inactivate Ctnna1 in ECs.

18 To monitor the efficiency of the Cre-mediated deletion of the floxed Ctnna1 exon, a tdTomato

19 reporter was used (stock number 007914; strain name: B6.Cg-Gt [ROSA] 26Sortm14 [CAG-

20 tdTomato]Hze/J, also named Ai14D, http://jaxmice.jax.org/strain/007914.html) (71). The re-

21 porter contains a loxP-flanked STOP cassette that prevents the transcription of the downstream

22 CAG promoter-driven red fluorescent protein variant tdTomato. The STOP cassette is removed

23 in the Cre-expressing tissue(s), and tdTomato is expressed. Because this CAG promoter-driven

31 1 reporter construct was inserted into the Gt (ROSA) 26Sor locus, tdTomato is expressed only in

2 tissues that express Cre.

3 Floxed Cdh5 mice were obtained from Cyagen (https://www.cyagen.com/cn/zh-cn/sperm-

4 bank-cko/12562).

5 Lrp5 KO mice were obtained from the Jackson Laboratory (stock number 005823,

6 https://www.jax.org/strain/005823).

7 Fzd4 KO mice were obtained from the Jackson Laboratory (stock number 012823,

8 https://www.jax.org/strain/012823).

9 Ctnnb1floxedExon3/+ mice were generated by Viewsolid Biotech (Beijing, China) using the

10 CRISPR/Cas9 nickase technique via a design described by Harada et al. (72). Two loxP sites

11 with the same orientation were placed upstream and downstream of exon 3 of the Ctnnb1 gene.

12 Ctnnb1floxedExon3/+mice were mated to Pdgfb-iCre transgenic mice (70) to generate

13 Ctnnb1floxedExon3/+ Pdgfb-iCre-ER mice.

14 Ctnna1 F72S knock-in mice carrying the F72S mutation corresponding to the human F72S

15 mutation (named Ctnna1em1XJZ, hereafter Ctnna1F72S) were generated using the CRISPR/Cas9

16 nickase technique. gRNA sequence: AAGCAACTGAGAATTTCTTGG. Donor oligo with the

17 sequence of CCATGTTTTGGCTGCATCTGTTGAACAAGCAACTGAGAATTCCTT-

18 GGAAAAGGGGGATAAAATTGCAAAAGAGAGCCAGT were co-injected into the

19 C57BL/6J zygotes to introduce F72S point mutation into the mouse genome. Ctnna1F72S target-

20 ing allele was screened using a sequencing PCR product.

21 Lrp5 knock-in mice carrying the P847L mutation corresponding to the human P848L mutation

22 (named Lrp5em1XJZ, hereafter Lrp5P847L) were generated using the CRISPR/Cas9 nickase tech-

23 nique. gRNA sequence: GACGATCTGCCCTACCCGTTTGG. Donor oligo with the sequence

32 1 of TATGTGCTATGTCCCCGCACAGGTCAGGAGCGCATGGTGATAGCTGAC-

2 GATCTGCCCTACTGTTTGGCCTGACTCAATATAGCGATTACATCTACTGGACTGACT

3 GGAACCTGCATAGCATT were co-injected into the C57BL/6J zygotes to introduce P847L

4 point mutation into the mouse genome. Lrp5P847L targeting allele was screened using a sequenc-

5 ing PCR product.

6 Western blot analysis

7 The primary and secondary antibodies used for Western blotting are listed in Table S12 in the

8 online-only Data Supplement. Uncropped immunoblots are shown in the online-only Data Sup-

9 plement.

10 Statistical analysis

11 All data are presented as mean ± standard deviation. Animals were assigned randomly to ex-

12 perimental groups. Western blotting signals were detected using Image Quant LAS 500 (GE Life

13 Sciences), and ImageJ software was used to quantify the detected signals. Statistical analysis was

14 performed via GraphPad Prism 6.0. The data sets were tested for normal distribution using

15 Shapiro-Wilk test. If the data set is not normally distributed, non-parametric statistic is used. p-

16 values were calculated by two-tailed Student’s t-test or by one- or two-way ANOVA followed by

17 a, Tukey, Dunnett or Sidak’s multiple comparisons test as appropriate. A p < 0.05 was consid-

18 ered statistically significant.

19 Study Approval

20 This research was carried out in accordance with the tenets of the Declaration of Helsinki and

21 was approved by the ethical oversight committee of Sichuan Provincial People's Hospital,

22 Xinhua Hospital, Shanghai Jiaotong University and Aravind Eye Hospital. Written informed

33 1 consent was obtained from subjects who participated in this study or from the legal guardians of

2 minors.

3 Author contributions

4 Z.Y. and XJ.Zhu designed and supervised the study. P.Z., X. Z., P. F. and P. S recruited the

5 participants. L.H. and Z.Y performed the sequencing analysis. XJ.Zhu, S.L., M.Y., L.Z., W.Z.,

6 and Z.Y. performed the animal analysis, cell biology, immunohistochemistry and gene expression

7 studies. L.Z. Y.H., Y.Y., YM.Y., XZ, S.Z. and H.X. performed the construction and mutation of

8 plasmids. S.M. and F.H. were responsible for sample preservation and DNA extraction. S.Z., X.Z.

9 and H.X. were responsible for animal breeding. XJ.Zhu, M.Y., S.L. and Z.Y. wrote the manuscript.

10 All authors critically revised and gave final approval to this manuscript.

11 Acknowledgments

12 The authors want to thank all patients and their family members for participating in this study.

13 This research project was supported by: the National Precision Medicine Project

14 (2016YFC0905200), the National Natural Science Foundation of China (81790643 to Z.Y.),

15 (81970841, 81770950 to XJ.Zhu), (82000913 to S.L.), (82071009 to L.Z.), (81770964, 81470642

16 to P.Z.), (81770963 to P.F.), (81670895 to L.H.); CAMS Innovation Fund for Medical Sciences

17 (2019-12M-5-032), the Department of Science and Technology of Sichuan Province of China

18 (2020ZYD037 to Z.Y., 20YZHY0011 to XJ.Zhu), (2017JQ0024, 2016HH0072 to L.H.),

19 (2018YSZH0020 to L.Z.), (2019M653382 to S.L.).

20 Data and materials availability

21 All data are available in the main text or the supplementary materials. 22

23

34 1 References

2 3 1. Criswick VG, and Schepens CL. Familial exudative vitreoretinopathy. American journal 4 of ophthalmology. 1969;68(4):578-94. 5 2. Collin RW, Nikopoulos K, Dona M, Gilissen C, Hoischen A, Boonstra FN, et al. ZNF408 6 is mutated in familial exudative vitreoretinopathy and is crucial for the development of 7 zebrafish retinal vasculature. Proceedings of the National Academy of Sciences of the 8 United States of America. 2013;110(24):9856-61. 9 3. Robitaille J, MacDonald ML, Kaykas A, Sheldahl LC, Zeisler J, Dube MP, et al. Mutant 10 frizzled-4 disrupts retinal angiogenesis in familial exudative vitreoretinopathy. Nature 11 genetics. 2002;32(2):326-30. 12 4. Toomes C, Bottomley HM, Jackson RM, Towns KV, Scott S, Mackey DA, et al. 13 Mutations in LRP5 or FZD4 underlie the common familial exudative vitreoretinopathy 14 locus on chromosome 11q. American journal of human genetics. 2004;74(4):721-30. 15 5. Jiao X, Ventruto V, Trese MT, Shastry BS, and Hejtmancik JF. Autosomal recessive 16 familial exudative vitreoretinopathy is associated with mutations in LRP5. American 17 journal of human genetics. 2004;75(5):878-84. 18 6. Chen ZY, Battinelli EM, Fielder A, Bundey S, Sims K, Breakefield XO, et al. A mutation 19 in the Norrie disease gene (NDP) associated with X-linked familial exudative 20 vitreoretinopathy. Nature genetics. 1993;5(2):180-3. 21 7. Junge HJ, Yang S, Burton JB, Paes K, Shu X, French DM, et al. TSPAN12 regulates 22 retinal vascular development by promoting Norrin- but not Wnt-induced FZD4/beta- 23 catenin signaling. Cell. 2009;139(2):299-311. 24 8. Nikopoulos K, Gilissen C, Hoischen A, van Nouhuys CE, Boonstra FN, Blokland EA, et 25 al. Next-generation sequencing of a 40 Mb linkage interval reveals TSPAN12 mutations 26 in patients with familial exudative vitreoretinopathy. American journal of human genetics. 27 2010;86(2):240-7. 28 9. Poulter JA, Ali M, Gilmour DF, Rice A, Kondo H, Hayashi K, et al. Mutations in TSPAN12 29 cause autosomal-dominant familial exudative vitreoretinopathy. American journal of 30 human genetics. 2010;86(2):248-53. 31 10. Robitaille JM, Gillett RM, LeBlanc MA, Gaston D, Nightingale M, Mackley MP, et al. 32 Phenotypic overlap between familial exudative vitreoretinopathy and microcephaly, 33 lymphedema, and chorioretinal dysplasia caused by KIF11 mutations. JAMA 34 ophthalmology. 2014;132(12):1393-9. 35 11. Dixon MW, Stem MS, Schuette JL, Keegan CE, and Besirli CG. CTNNB1 mutation 36 associated with familial exudative vitreoretinopathy (FEVR) phenotype. Ophthalmic 37 genetics. 2016;37(4):468-70. 38 12. Panagiotou ES, Sanjurjo Soriano C, Poulter JA, Lord EC, Dzulova D, Kondo H, et al. 39 Defects in the Cell Signaling Mediator beta-Catenin Cause the Retinal Vascular 40 Condition FEVR. American journal of human genetics. 2017;100(6):960-8. 41 13. Khan K, Logan CV, McKibbin M, Sheridan E, Elcioglu NH, Yenice O, et al. Next 42 generation sequencing identifies mutations in Atonal homolog 7 (ATOH7) in families with 43 global eye developmental defects. Hum Mol Genet. 2012;21(4):776-83. 44 14. Wu JH, Liu JH, Ko YC, Wang CT, Chung YC, Chu KC, et al. Haploinsufficiency of 45 RCBTB1 is associated with Coats disease and familial exudative vitreoretinopathy. Hum 46 Mol Genet. 2016;25(8):1637-47. 47 15. Downey LM, Keen TJ, Roberts E, Mansfield DC, Bamashmus M, and Inglehearn CF. A 48 new locus for autosomal dominant familial exudative vitreoretinopathy maps to 49 chromosome 11p12-13. American journal of human genetics. 2001;68(3):778-81.

35 1 16. Park H, Yamamoto H, Mohn L, Ambuhl L, Kanai K, Schmidt I, et al. Integrin-linked 2 kinase controls retinal angiogenesis and is linked to Wnt signaling and exudative 3 vitreoretinopathy. Nat Commun. 2019;10(1):5243. 4 17. Zhang L, Zhang X, Xu H, Huang L, Zhang S, Liu W, et al. Exome sequencing revealed 5 Notch ligand JAG1 as a novel candidate gene for familial exudative vitreoretinopathy. 6 Genetics in medicine : official journal of the American College of Medical Genetics. 7 2019. 8 18. Gong Y, Slee RB, Fukai N, Rawadi G, Roman-Roman S, Reginato AM, et al. LDL 9 receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell. 10 2001;107(4):513-23. 11 19. Berger W, Meindl A, van de Pol TJ, Cremers FP, Ropers HH, Doerner C, et al. Isolation 12 of a candidate gene for Norrie disease by positional cloning. Nature genetics. 13 1992;1(3):199-203. 14 20. Ostergaard P, Simpson MA, Mendola A, Vasudevan P, Connell FC, van Impel A, et al. 15 Mutations in KIF11 cause autosomal-dominant microcephaly variably associated with 16 congenital lymphedema and chorioretinopathy. American journal of human genetics. 17 2012;90(2):356-62. 18 21. Tucci V, Kleefstra T, Hardy A, Heise I, Maggi S, Willemsen MH, et al. Dominant beta- 19 catenin mutations cause intellectual disability with recognizable syndromic features. The 20 Journal of clinical investigation. 2014;124(4):1468-82. 21 22. Li VS, Ng SS, Boersema PJ, Low TY, Karthaus WR, Gerlach JP, et al. Wnt signaling 22 through inhibition of beta-catenin degradation in an intact Axin1 complex. Cell. 23 2012;149(6):1245-56. 24 23. Smallwood PM, Williams J, Xu Q, Leahy DJ, and Nathans J. Mutational analysis of 25 Norrin-Frizzled4 recognition. The Journal of biological chemistry. 2007;282(6):4057-68. 26 24. Xu Q, Wang Y, Dabdoub A, Smallwood PM, Williams J, Woods C, et al. Vascular 27 development in the retina and inner ear: control by Norrin and Frizzled-4, a high-affinity 28 ligand-receptor pair. Cell. 2004;116(6):883-95. 29 25. Ye X, Wang Y, Cahill H, Yu M, Badea TC, Smallwood PM, et al. Norrin, frizzled-4, and 30 Lrp5 signaling in endothelial cells controls a genetic program for retinal vascularization. 31 Cell. 2009;139(2):285-98. 32 26. Gilmour DF. Familial exudative vitreoretinopathy and related retinopathies. Eye. 33 2015;29(1):1-14. 34 27. Salvo J, Lyubasyuk V, Xu M, Wang H, Wang F, Nguyen D, et al. Next-generation 35 sequencing and novel variant determination in a cohort of 92 familial exudative 36 vitreoretinopathy patients. Investigative ophthalmology & visual science. 37 2015;56(3):1937-46. 38 28. Kobielak A, and Fuchs E. Alpha-catenin: at the junction of intercellular adhesion and 39 actin dynamics. Nat Rev Mol Cell Biol. 2004;5(8):614-25. 40 29. Piao HL, Yuan Y, Wang M, Sun Y, Liang H, and Ma L. alpha-catenin acts as a tumour 41 suppressor in E-cadherin-negative basal-like breast cancer by inhibiting NF-kappaB 42 signalling. Nature cell biology. 2014;16(3):245-54. 43 30. Liu TX, Becker MW, Jelinek J, Wu WS, Deng M, Mikhalkevich N, et al. Chromosome 5q 44 deletion and epigenetic suppression of the gene encoding alpha-catenin (CTNNA1) in 45 myeloid cell transformation. Nat Med. 2007;13(1):78-83. 46 31. Ji H, Wang J, Fang B, Fang X, and Lu Z. alpha-Catenin inhibits glioma cell migration, 47 invasion, and proliferation by suppression of beta-catenin transactivation. J Neurooncol. 48 2011;103(3):445-51. 49 32. Silvis MR, Kreger BT, Lien WH, Klezovitch O, Rudakova GM, Camargo FD, et al. alpha- 50 catenin is a tumor suppressor that controls cell accumulation by regulating the

36 1 localization and activity of the transcriptional coactivator Yap1. Sci Signal. 2 2011;4(174):ra33. 3 33. Buckley CD, Tan J, Anderson KL, Hanein D, Volkmann N, Weis WI, et al. Cell adhesion. 4 The minimal cadherin-catenin complex binds to actin filaments under force. Science. 5 2014;346(6209):1254211. 6 34. Majewski IJ, Kluijt I, Cats A, Scerri TS, de Jong D, Kluin RJ, et al. An alpha-E-catenin 7 (CTNNA1) mutation in hereditary diffuse gastric cancer. The Journal of pathology. 8 2013;229(4):621-9. 9 35. Saksens NT, Krebs MP, Schoenmaker-Koller FE, Hicks W, Yu M, Shi L, et al. Mutations 10 in CTNNA1 cause butterfly-shaped pigment dystrophy and perturbed retinal pigment 11 epithelium integrity. Nature genetics. 2016;48(2):144-51. 12 36. de Breuk A, Acar IE, Kersten E, Schijvenaars M, Colijn JM, Haer-Wigman L, et al. 13 Development of a Genotype Assay for Age-Related Macular Degeneration: The EYE- 14 RISK Consortium. Ophthalmology. 2020. 15 37. Tanner A, Chan HW, Pulido JS, Arno G, Ba-Abbad R, Jurkute N, et al. Clinical and 16 genetic findings in CTNNA1-associated macular pattern dystrophy. Ophthalmology. 17 2020. 18 38. Gong B, Zhang H, Huang L, Chen Y, Shi Y, Tam PO, et al. Mutant RAMP2 causes 19 primary open-angle glaucoma via the CRLR-cAMP axis. Genetics in medicine : official 20 journal of the American College of Medical Genetics. 2019;21(10):2345-54. 21 39. Novarino G, Fenstermaker AG, Zaki MS, Hofree M, Silhavy JL, Heiberg AD, et al. 22 Exome sequencing links corticospinal motor neuron disease to common 23 neurodegenerative disorders. Science. 2014;343(6170):506-11. 24 40. Drees F, Pokutta S, Yamada S, Nelson WJ, and Weis WI. Alpha-catenin is a molecular 25 switch that binds E-cadherin-beta-catenin and regulates actin-filament assembly. Cell. 26 2005;123(5):903-15. 27 41. Yamada S, Pokutta S, Drees F, Weis WI, and Nelson WJ. Deconstructing the cadherin- 28 catenin-actin complex. Cell. 2005;123(5):889-901. 29 42. Carmeliet P, Lampugnani MG, Moons L, Breviario F, Compernolle V, Bono F, et al. 30 Targeted deficiency or cytosolic truncation of the VE-cadherin gene in mice impairs 31 VEGF-mediated endothelial survival and angiogenesis. Cell. 1999;98(2):147-57. 32 43. Hwang SG, Yu SS, Ryu JH, Jeon HB, Yoo YJ, Eom SH, et al. Regulation of beta-catenin 33 signaling and maintenance of chondrocyte differentiation by ubiquitin-independent 34 proteasomal degradation of alpha-catenin. The Journal of biological chemistry. 35 2005;280(13):12758-65. 36 44. Simcha I, Shtutman M, Salomon D, Zhurinsky J, Sadot E, Geiger B, et al. Differential 37 nuclear translocation and transactivation potential of beta-catenin and . The 38 Journal of cell biology. 1998;141(6):1433-48. 39 45. Takahashi N, Ishihara S, Takada S, Tsukita S, and Nagafuchi A. Posttranscriptional 40 regulation of alpha-catenin expression is required for Wnt signaling in L cells. 41 Biochemical and biophysical research communications. 2000;277(3):691-8. 42 46. Su W, and Kowalczyk AP. The VE-cadherin cytoplasmic domain undergoes proteolytic 43 processing during endocytosis. Mol Biol Cell. 2017;28(1):76-84. 44 47. Luhmann UF, Lin J, Acar N, Lammel S, Feil S, Grimm C, et al. Role of the Norrie 45 disease pseudoglioma gene in sprouting angiogenesis during development of the retinal 46 vasculature. Investigative ophthalmology & visual science. 2005;46(9):3372-82. 47 48. Lobov IB, Rao S, Carroll TJ, Vallance JE, Ito M, Ondr JK, et al. WNT7b mediates 48 macrophage-induced programmed cell death in patterning of the vasculature. Nature. 49 2005;437(7057):417-21.

37 1 49. Xu Q, Wang Y, Dabdoub A, Smallwood PM, Williams J, Woods C, et al. Vascular 2 Development in the Retina and Inner Ear : Control by Norrin and Frizzled-4, a High- 3 Affinity Ligand-Receptor Pair. Cell. 2004;116(6):883-95. 4 50. Wang Y, Rattner A, Zhou Y, Williams J, Smallwood PM, and Nathans J. Norrin/Frizzled4 5 signaling in retinal vascular development and blood brain barrier plasticity. Cell. 6 2012;151(6):1332-44. 7 51. del Toro R, Prahst C, Mathivet T, Siegfried G, Kaminker JS, Larrivee B, et al. 8 Identification and functional analysis of endothelial tip cell-enriched genes. Blood. 9 2010;116(19):4025-33. 10 52. Gerhardt H, Golding M, Fruttiger M, Ruhrberg C, Lundkvist A, Abramsson A, et al. VEGF 11 guides angiogenic sprouting utilizing endothelial tip cell filopodia. The Journal of cell 12 biology. 2003;161(6):1163-77. 13 53. Jakobsson L, Franco CA, Bentley K, Collins RT, Ponsioen B, Aspalter IM, et al. 14 Endothelial cells dynamically compete for the tip cell position during angiogenic 15 sprouting. Nature cell biology. 2010;12(10):943-53. 16 54. Zhou Z, Tang AT, Wong WY, Bamezai S, Goddard LM, Shenkar R, et al. Cerebral 17 cavernous malformations arise from endothelial gain of MEKK3-KLF2/4 signalling. 18 Nature. 2016;532(7597):122-6. 19 55. Gaengel K, Niaudet C, Hagikura K, Lavina B, Muhl L, Hofmann JJ, et al. The 20 sphingosine-1-phosphate receptor S1PR1 restricts sprouting angiogenesis by regulating 21 the interplay between VE-cadherin and VEGFR2. Dev Cell. 2012;23(3):587-99. 22 56. Yamamoto H, Ehling M, Kato K, Kanai K, van Lessen M, Frye M, et al. Integrin beta1 23 controls VE-cadherin localization and blood vessel stability. Nat Commun. 2015;6:6429. 24 57. Lampugnani MG, Zanetti A, Breviario F, Balconi G, Orsenigo F, Corada M, et al. VE- 25 cadherin regulates endothelial actin activating Rac and increasing membrane 26 association of Tiam. Mol Biol Cell. 2002;13(4):1175-89. 27 58. Ding L, Ellis MJ, Li S, Larson DE, Chen K, Wallis JW, et al. Genome remodelling in a 28 basal-like breast cancer metastasis and xenograft. Nature. 2010;464(7291):999-1005. 29 59. Green RC, Berg JS, Grody WW, Kalia SS, Korf BR, Martin CL, et al. ACMG 30 recommendations for reporting of incidental findings in clinical exome and genome 31 sequencing. Genetics in medicine : official journal of the American College of Medical 32 Genetics. 2013;15(7):565-74. 33 60. Kalia SS, Adelman K, Bale SJ, Chung WK, Eng C, Evans JP, et al. Recommendations 34 for reporting of secondary findings in clinical exome and genome sequencing, 2016 35 update (ACMG SF v2.0): a policy statement of the American College of Medical 36 Genetics and Genomics. Genetics in medicine : official journal of the American College 37 of Medical Genetics. 2017;19(2):249-55. 38 61. Rangarajan ES, and Izard T. Dimer asymmetry defines alpha-catenin interactions. Nat 39 Struct Mol Biol. 2013;20(2):188-93. 40 62. Benedito R, Roca C, Sorensen I, Adams S, Gossler A, Fruttiger M, et al. The notch 41 ligands Dll4 and Jagged1 have opposing effects on angiogenesis. Cell. 42 2009;137(6):1124-35. 43 63. Yan X, Yang Z, Chen Y, Li N, Wang L, Dou G, et al. Endothelial cells-targeted soluble 44 human Delta-like 4 suppresses both physiological and pathological ocular angiogenesis. 45 Sci China Life Sci. 2015;58(5):425-31. 46 64. Carvalho JR, Fortunato IC, Fonseca CG, Pezzarossa A, Barbacena P, Dominguez- 47 Cejudo MA, et al. Non-canonical Wnt signaling regulates junctional mechanocoupling 48 during angiogenic collective cell migration. eLife. 2019;8. 49 65. Ratheesh A, and Yap AS. A bigger picture: classical and the dynamic actin 50 cytoskeleton. Nat Rev Mol Cell Biol. 2012;13(10):673-9.

38 1 66. Dejana E, Orsenigo F, and Lampugnani MG. The role of adherens junctions and VE- 2 cadherin in the control of vascular permeability. Journal of cell science. 2008;121(Pt 3 13):2115-22. 4 67. Dejana E, Tournier-Lasserve E, and Weinstein BM. The control of vascular integrity by 5 endothelial cell junctions: molecular basis and pathological implications. Dev Cell. 6 2009;16(2):209-21. 7 68. National Genomics Data Center M, and Partners. Database Resources of the National 8 Genomics Data Center in 2020. Nucleic Acids Res. 2020;48(D1):D24-D33. 9 69. Vasioukhin V, Bauer C, Degenstein L, Wise B, and Fuchs E. Hyperproliferation and 10 defects in epithelial polarity upon conditional ablation of alpha-catenin in skin. Cell. 11 2001;104(4):605-17. 12 70. Claxton S, Kostourou V, Jadeja S, Chambon P, Hodivala-Dilke K, and Fruttiger M. 13 Efficient, inducible Cre-recombinase activation in vascular endothelium. Genesis. 14 2008;46(2):74-80. 15 71. Madisen L, Zwingman TA, Sunkin SM, Oh SW, Zariwala HA, Gu H, et al. A robust and 16 high-throughput Cre reporting and characterization system for the whole mouse brain. 17 Nat Neurosci. 2010;13(1):133-40. 18 72. Harada N, Tamai Y, Ishikawa T, Sauer B, Takaku K, Oshima M, et al. Intestinal 19 polyposis in mice with a dominant stable mutation of the beta-catenin gene. EMBO J. 20 1999;18(21):5931-42. 21

22

39 1

2 Figure 1. CTNNA1 mutations in families with familial exudative vitreoretinopathy (FEVR).

3 (A) FEVR pedigrees and Sanger sequencing analysis showing inheritance of FEVR (patients are

4 denoted in black). Three heterozygous mutations in CTNNA1 were identified. Black arrows indi-

5 cate the proband of each family. Red arrows indicate the changed nucleotides. Affected amino

40 1 acids are denoted in red and are conserved among different species. (B) Schematic representation

2 of the CTNNA1 protein domains showing the location of variants identified in this study. (C and

3 D) Fundus fluorescein angiography (C) and fundus photography (D) of 31-year-old patient (Ⅰ:1)

4 in family-3016, 25-year-old patient (Ⅰ:2) in family-34, 27-year-old patient (Ⅰ:2) in family-3004

5 and normal person.

6

41 1

2 Figure 2. CTNNA1 mutations result in Norrin/β-catenin signaling overactivation, α-catenin

3 mislocalization, and F-actin disorganization. (A) Western blot analysis of CTNNA1 (WT and

4 mutants) co-immunoprecipitated with CTNNB1 and VE-Cadherin. An empty vector was used as

5 negative control. (B) Results of luciferase reporter assay in HEK293 STF cells. Cells were trans-

6 fected with plasmids containing CTNNA1 (WT, p.F72S, p.R376C*fs27, or p.P893L) or an empty

7 vector (pCDNA3.1). Plasmids were co-transfected with LRP5, FZD4, NDP, and Renillar-

8 eniformis (PGL4.1). The activity of WT protein was normalized as 1. Error bars, SD (standard

9 deviation). p-values were calculated from multiple comparisons in one-way ANOVA with Dun-

10 nett’s multiple comparisons test (n = 4); * p < 0.05, **** p < 0.0001. (C) qPCR analysis

42 1 demonstrated efficient shRNA-mediated knockdown of CTNNA1 in HEK 293STF cells. Error

2 bars, SD. p-values, Student’s t-test (n = 3); ** p < 0.01. (D) ShRNA-mediated knockdown of

3 CTNNA1 in the 293STF cell line led to elevated luciferase activity. Error bars, SD. p-values, Stu-

4 dent’s t-test (n = 4); *** p < 0.001. (E) Adenovirus (Ad)-mediated overexpression of wildtype

5 (WT) and mutant CTNNA1 (F72S, P893L) in HRECs. Anti-FLAG (CTNNA1) or F-ACTIN an-

6 tibody was co-stained with DAPI, respectively. Scale bars, 25 µm. Experiments were performed

7 at least three times independently.

8

9

43 1

2 Figure 3. CTNNA1, CDH5, and CTNND1 are essential for the integrity of the cad-

3 herin/catenin complex in HRECs. (A) Immunofluorescence images of HRECs transfected with

4 shRNA targeting CTNNA1, CDH5, CTNND1, or control shRNA. Anti-CTNNA1, VE-Cadherin,

5 CTNND1, CTNNB1, or F-ACTIN antibody (red) was co-stained with DAPI (blue), respectively.

6 Scale bars, 25 µm. (B–E) Quantification of membrane signal intensity of CTNNA1, VE-Cad-

7 herin, CTNND1, and CTNNB1 in CTNNA1, CDH5, CTNND1 KD, and control HRECs. Error

8 bars, SD. p-values were calculated from multiple comparisons in one-way ANOVA with

44 1 Dunnett’s multiple comparisons test (n = 8), *** p < 0.001, **** p < 0.0001. Experiments were

2 performed at least three times independently.

3

45 1

46 1 Figure 4. Conditional knockout of Ctnna1 in mice causes severe retinal vascularization de-

2 fects. (A) DAPI staining of hyaloid vessels in the eyes of control and Ctnna1iECKO mice, showing

3 that hyaloid vessel regression was significantly delayed in the eye. Scale bar, 250 µm. (B) Com-

4 pared to those of littermate controls (Ctrl), P9 flat-mounted retinas of Ctnna1iECKO/+ showed de-

5 layed radial growth of the superficial vascular plexus, with moderate neovascularizations at the

6 angiogenic front. Ctnna1iECKO retinas showed retarded vascular growth and hyperplasia of the

7 primary vascular plexus. Vessels were stained with isolectin-B4 (IB4, red). Scale bar, 500 µm.

8 (C) Retinal frozen sections of P9 Ctrl, Ctnna1iECKO/+, and Ctnna1iECKO mice were co-stained with

9 IB4 (red) and DAPI (blue). Scale bars, 25 µm. The vertical growth of the superficial retina vas-

10 cular plexus was delayed in Ctnna1iECKO/+ (middle panel). In Ctnna1iECKO retina, profound de-

11 fects in vertical vascular growth into the deeper retinal layers was observed, and the vascular

12 plexus became hyperplastic. No secondary or tertiary vessels were observed in these retinas. (D)

13 VE-Cadherin (green) and IB4 (red) staining of P9 Ctnna1iECKO and Ctrl retinas. VE-Cadherin

14 was disorganized in Ctnna1iECKO retinas. White dotted boxes indicate enlarged regions, detailed

15 on right. Scale bars, 25 µm and 10 µm. (E) P9 flat-mounted Ctnna1iECKO retinas showed exten-

16 sive leakage of Evans blue (white arrow) and visible, enlarged blood vessels (yellow arrow),

17 compared to the retinas of Ctrl retinas. (F) Ctnna1iECKO, Ctnna1iECKO/+, and control retinas were

18 co-stained with IB4 (green) and Ter119 (red). Extensive leakage of erythrocytes was observed

19 (white arrows) in Ctnna1iECKO/+ and Ctnna1iECKO retinas. Scale bars, 25 µm. (G–I) Quantification

20 of vascular progression, vascular density, and vessel leakage. Error bars, SD. p-values were cal-

21 culated from multiple comparisons in one-way ANOVA with Tukey’s multiple comparisons test

22 (n ≥ 6), ns, no significance, * p<0.05, **** p < 0.0001. Experiments were performed at least

23 three times independently.

47 1

2 Figure 5. Ctnna1 deletion in mouse endothelial cells increased their proliferation at P5. (A)

3 Vascular cell proliferation of Ctrl, Ctnna1iECKO/+, and Ctnna1iECKO/iECKO mice at the vitreal sur-

4 face was measured with EdU labeling at P5. Images captured at higher magnification are shown

5 in the right panel. Scale bar, 250 µm (left), 25 µm (right). (B) Quantification of EdU positive

6 cells per vascular area, error bars, SD. p-values, p-values were calculated from multiple compari-

7 sons in one-way ANOVA with Tukey’s multiple comparisons test (n = 5 mice for each group).

8 ns, no significance, *** p < 0.001, **** p < 0.0001. Experiments were performed at least three

9 times independently.

10

48 1

2 Figure 6. Comparison of retina vascular phenotypes of Ctnna1 endothelial conditional

3 knockout mice and Lrp5 or Fzd4 KO mice. (A) P5 Ctnna1iECKO, Lrp5-/-, and Fzd4-/- mice

49 1 showed similar phenotypes, such as delayed superficial vascular progression, neovascularization

2 at the periphery region, and extensive erythrocyte leakage. However, compared to that of Ctrl

3 retinas, the vascular density of Ctnna1iECKO retinas was increased, whereas it was decreased in

4 Lrp5-/- and Fzd4-/- retinas. White arrows denote leakage areas. Scale bar, 250 μm. Quantification

5 of vascular progression and density are shown in (B) and (C), respectively. Error bars, SD. p-

6 values were calculated from multiple comparisons in one-way ANOVA with Dunnett’s multiple

7 comparisons test (n = 4). ** p < 0.01, **** p < 0.0001. Experiments were performed at least

8 three times independently.

9

50 1

51 1 Figure 7. Vegfa distribution in Ctnna1 endothelial conditional knockout, Lrp5 KO, Fzd4

2 KO, and Ctnnb1 GOF Homo mice. (A) Vegf164 (green) and IB4 (red) staining of P6 Ctrl and

3 Ctnna1iECKO retinas. Abnormal distribution and elevated expression of Vegf164 expressed by

4 both astrocytes and endothelial cells were observed in the angiogenic front and remodeling

5 plexus in Ctnna1iECKO retinal vessels. (B and C) In P6 Lrp5+/-, Lrp5-/-, Fzd4+/-, and Fzd4-/- retinas,

6 Vegf164 was normally localized in the angiogenic front and absent in the remodeling plexus.

7 Scale bars, 25 µm. (D) In P6 Ctrl and Ctnnb1 GOF Homo retinas, abnormal distribution and ele-

8 vated expression of Vegf164 expressed by both astrocytes and endothelial cells were observed in

9 the angiogenic front, whereas in the remodeling plexus, only endothelial-derived Vegf164 was

10 elevated. Experiments were performed at least three times independently.

11

52 1

2 Figure 8. Ctnna1F72S/iECKO mice showed retinal vasculature similar to that of

3 Ctnna1iECKO/iECKO mice. (A) Anti-Ter119 (green) and IB4 (red) immunofluorescence of P7 con-

4 trol, Ctnna1F72S/+, Ctnna1iECKO/+, and Ctnna1F72S/iECKO mice retinas. Scale bars, 200 μm. (B-D)

53 1 Quantification of vascular progression, relative vascular density, and vessel leakage. Error bars,

2 SD. p-values were calculated from multiple comparisons in one-way ANOVA with Tukey’s mul-

3 tiple comparisons test(n≥6), ns, no significance, * p<0.05, *** p < 0.001, **** p < 0.0001. Ex-

4 periments were performed at least three times independently.

5

54 1

2 Figure 9. Gain of function of Ctnnb1 allele in mice caused retinal vascularization defects.

3 (A) P6 flat-mounted retinas from Ctrl, Ctnnb1 foxedExon3/+; Pdgfb-iCre-ER heterozygous (Ctnnb1

4 gain of function [GOF], Het) and Ctnnb1 foxedExon3/foxedExon3; Pdgfb-iCre-ER homozygous (Ctnnb1

5 GOF, Homo) mice, stained with IB4. Compared to those of littermate controls, retinas of the het-

6 erozygous Ctnnb1 GOF and homozygous Ctnnb1 GOF mice showed incomplete retinal vascular-

7 ization. Scale bar, 500 μm. (B) Quantification of vascular progression at P6. Error bars, SD. p-

8 values were calculated from multiple comparisons in one-way ANOVA with Tukey’s multiple

55 1 comparisons test (n = 12). * p<0.05, *** p < 0.001, **** p < 0.0001. (C) DAPI staining of hya-

2 loid vessels in the eyes of control and Ctnnb1 GOF Homo mice, showing relatively delayed hya-

3 loid vessel regression in the eyes of Ctnnb1 GOF Homo mice. Scale bar, 250 μm. (D) Retinal

4 frozen sections of P9 Ctrl and Ctnnb1 GOF Homo mice were co-stained with IB4 (red) and

5 DAPI (blue). Scale bars, 25 μm. Experiments were performed at least three times independently.

6

7

8

56 1

2 Figure 10. Mutation of Lrp5 P847L in mice caused retinal vascularization defects. (A) P6

3 flat-mounted retinas from Lrp5 P847L/+ and Lrp5 P847L/P847L mice, stained with IB4. Compared to

4 those of Lrp5 P847L/+ mice, retinas of the Lrp5 P847L/P847L mice showed incomplete retinal vascular-

5 ization. Scale bar, 500 μm. (B) Quantification of vascular progression of Lrp5 P847L/+ and Lrp5

6 P847L/P847L mice retinas at P6. Error bars, SD. Student’s t-test (n = 8). ****: p < 0.0001. (C) DAPI

7 staining of hyaloid vessels in the eyes of Lrp5 P847L/+ and Lrp5 P847L/P847L mice, showing delayed

8 hyaloid vessel regression in the eyes of Lrp5 P847L/P847L mice. Scale bar, 250 µm . (D) Retinal

57 1 frozen sections of P9 Lrp5 P847L/+ and Lrp5 P847L/P847L mice were co-stained with IB4 (red) and

2 DAPI (blue). Scale bars, 25 μm. Experiments were performed at least three times independently.

3

58