August 2011
Volume 52, Issue 9
Free
Retina  |   August 2011
Frizzled 4 Is Required for Retinal Angiogenesis and Maintenance of the Blood-Retina Barrier
Author Affiliations & Notes
  • Kim T. Paes
    From the Departments of Ophthalmology,
  • Ernest Wang
    From the Departments of Ophthalmology,
  • Kathy Henze
    Pathology,
  • Peter Vogel
    Pathology,
  • Robert Read
    Pathology,
  • Adisak Suwanichkul
    Molecular Biology, and
  • Laura L. Kirkpatrick
    Pharmaceutical Discovery, Lexicon Pharmaceuticals, Inc., The Woodlands, Texas.
  • David Potter
    From the Departments of Ophthalmology,
  • Matthew M. Newhouse
    From the Departments of Ophthalmology,
  • Dennis S. Rice
    From the Departments of Ophthalmology,
  • Corresponding author: Dennis S. Rice, Lexicon Pharmaceuticals, Inc., Department of Ophthalmology, 8800 Technology Forest Place, The Woodlands, TX 77381; drice@lexpharma.com
Investigative Ophthalmology & Visual Science August 2011, Vol.52, 6452-6461. doi:10.1167/iovs.10-7146
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Kim T. Paes, Ernest Wang, Kathy Henze, Peter Vogel, Robert Read, Adisak Suwanichkul, Laura L. Kirkpatrick, David Potter, Matthew M. Newhouse, Dennis S. Rice; Frizzled 4 Is Required for Retinal Angiogenesis and Maintenance of the Blood-Retina Barrier. Invest. Ophthalmol. Vis. Sci. 2011;52(9):6452-6461. doi: 10.1167/iovs.10-7146.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: Mice deficient in the secreted protein Norrin or its receptor Frizzled-4 (FZD4) exhibit incomplete vascularization of the neural retina. However, because of early retinal vascular defects in the knockout models, it has not been possible to study FZD4 contribution in ocular neovascular disease. To further understand the role of this signaling pathway in physiological and pathologic angiogenesis, the authors generated a monoclonal antibody that neutralizes FZD4 function in vivo.

Methods.: Antibodies were generated by immunizing Fzd4 knockout mice with the cysteine-rich domain of FZD4. A monoclonal antibody (1.99.25) was discovered that antagonizes Norrin- and WNT3A-induced β-catenin accumulation in vitro. 1.99.25 and an isotype-matched negative control antibody were evaluated in models of developmental retinal angiogenesis, oxygen-induced retinopathy, and retinal angiomatous proliferation. The authors also investigated the role of FZD4 in maintaining the blood-retina barrier in normal adult mice.

Results.: Administration of 1.99.25 inhibited physiological and pathologic sprouting angiogenesis within the retina. Inhibition of FZD4 in developing retinal vascular networks caused the upregulation of PLVAP, a protein normally associated with fenestrated, immature endothelium in the CNS. In the adult neural retina, the administration of 1.99.25 induced PLVAP expression in the deep capillary bed and enabled extravasation of small and large molecules through the blood-retina barrier.

Conclusions.: These results demonstrate that FZD4 is required for physiological and pathologic angiogenesis in the retina and for regulation of retinal endothelial cell differentiation. The authors also show that FZD4 is critical for maintaining the integrity of the mature blood-retina barrier.

The canonical Wnt signaling pathway is implicated in angiogenesis and differentiation of the blood-brain barrier. 1 4 Canonical Wnt signaling is initiated by ligand binding to the cysteine-rich domain (CRD) of a Frizzled receptor, in complex with LDL-receptor related protein 5 or 6 (LRP5 or LRP6). Ligand binding results in decreased degradation of cytoplasmic β-catenin, which can then translocate to the nucleus. Nuclear β-catenin promotes TCF/LEF-1–mediated transcriptional activation and regulates expression of genes controlling cellular activities, including proliferation, migration, and differentiation. 5  
Several lines of evidence have shown that the Wnt pathway is involved in vascular network growth and differentiation in the central nervous system. 6 Genetic disruption of both Wnt7a and Wnt7b resulted in a lack of proper angiogenesis in the ventral neural tube. A global role for canonical Wnt signaling in angiogenesis was revealed by conditional disruption of β-catenin in endothelial cells, which exhibited a profound block in perineural vascular invasion in the neural tube. 12,4 Recently, the secreted protein Norrin (NDP) signaling, through a receptor complex composed of Frizzled-4 (FZD4), LRP5, and Tetraspanin-12 (TSPAN12), has been shown to direct retinal angiogenesis during development in both mice and humans. 7 Although Norrin is not a prototypical Wnt, it activates the Wnt/β-catenin pathway in endothelial cells. 8,9  
The importance of Norrin and FZD4 in retinal vascular development and disease has spurred the characterization of their cellular expression patterns and the role played by specific cell types in their function. Norrin is expressed exclusively in Müller glia during retinal vascular development, whereas FZD4 is more widely expressed in neurons, endothelial cells, and mural cells of the vasculature. 9,10 Elegant studies by Nathans and colleagues 9 demonstrated that an endothelial cell-specific knockout of Fzd4 reproduces the vascular phenotype observed in the global Fzd4 knockout. Endothelial cell reception of Norrin signaling is influenced by TSPAN12, which binds FZD4, promotes receptor clustering, and enhances transcriptional activation through β-catenin accumulation. 8 Knockout of Fzd4 impacts the earliest stages of superficial retinal vascular network development, 11 confounding experiments aimed at addressing whether the formation of the outer retinal vasculature or the transition of the vasculature to a mature, barrier phenotype are directly dependent on Fzd4. Moreover, Ndp and Fzd4 are expressed in the adult retina, suggesting important ongoing signaling in mature vessels. 9 To address these functions, we developed FZD4 neutralizing antibodies and used one of these to assess the outcome of inhibiting FZD4 in wild-type mice during specific stages of retinal vascular development. We also sought to understand the contribution of FZD4 in ocular vascular disease. 
Methods
Generation of Fzd4 and Vldlr Knockout Mice
The Fzd4 and Vldlr knockout lines were generated by homologous recombination using targeting vectors generated with the Lambda KOS system. 12 Details of genomic clone isolation, target vector construction, Southern blot confirmation of targeted embryonic stem cell clones, and mouse breeding are contained in the Supplementary Methods
Mouse Care and Study
All procedures involving animals were conducted in conformity with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, Institutional Animal Care and Use Committee guidelines, and state and federal laws and standards outlined in the Guide for the Care and Use of Laboratory Animals. Mice were housed at 24°C on a fixed 12-hour/light/12-hour dark cycle and had free access to water and rodent chow (product no. 5021; Purina, St. Louis, MO). 
mAb Generation and Purification
Details of all plasmids used, expression and purification methods for mouse and human FZD4-CRD protein, and expression and purification methods for the human FZD4 cell line are provided in the Supplementary Methods. Antibodies recognizing the FZD4-CRD were generated by immunizing Fzd4 knockout mice. Methods for antibody generation and characterization are provided in the Supplementary Methods. 
Developmental Retinal Angiogenesis Assay
Neonatal hybrid mice (C57BL/6 × 129/SvEvBrd) were generated from matings of wild-type mice. To examine the effect of FZD4-neutralizing mAb 1.99.25 on superficial retinal vascular development, postnatal day (P) 0.5 wild-type neonatal mice received a single subcutaneous dose (30 mg/kg) of either the 1.99.25 or the anti–keyhole limpet hemocyanin (KLH) control mAb, both formulated in Dulbecco's PBS. On P5.5, neonates were euthanatized by carbon dioxide inhalation and decapitation. Eyes were enucleated and processed for wholemount retinal staining with endothelial markers, as described in the histology section. To examine the effects on intraretinal capillary beds, neonates were treated once by intraperitoneal injection of 30 mg/kg 1.99.25 or anti–KLH mAb on P7, when most of the superficial vasculature had developed. Neonates were euthanatized by carbon dioxide inhalation followed by decapitation on P12, P15, and P16 after treatment with the antibodies. For light microscopy, eyes were enucleated and fixed in Davidson's fixative for paraffin sectioning or 4% paraformaldehyde prepared in 0.1 M PBS for immunohistochemical staining of retinal wholemounts, as described in the histology section. Four P15 mice (two anti–KLH mAb-treated controls and two 1.99.25-treated mice) were anesthetized and perfused with 1.25% paraformaldehyde and 1.25% glutaraldehyde in 0.1 M phosphate buffer. Eyes were removed, bisected, and incubated in 2% osmium tetroxide, stained with 2% uranyl acetate, and embedded in Araldite. Ultrathin sections were obtained and collected in formvar-coated slot grids and were subsequently stained with lead citrate. Images were acquired with a transmission electron microscope (2000EX; JEOL, Tokyo, Japan). 
Mouse Model of Oxygen-Induced Retinopathy
The mouse model of oxygen-induced retinopathy (OIR) was performed as previously described. 13 C57BL/6 pups (n = 8 per litter) and nursing mothers were placed in a chamber maintained at 75% oxygen (A84XOV OxyCycler system; BioSpherix, Lacona, NY) for 5 days from P7 to P12. On P12, pups and mothers were placed in clean cages in normal room air conditions for an additional 5 days until P17. Antibody treatment was administered daily by intraperitoneal injection of 30 mg/kg 1.99.25 or anti–KLH mAb control antibody from P12 to P16. To control for inter-litter variability, each litter was divided into treatment and control antibody groups, and a total of four litters were used per experiment. Malnourishment is known to influence the degree of retinal angiogenesis in this model, so only pups with weights greater than 6 g just before necropsy (P17) were used for analysis. On P17, pups were euthanatized by carbon dioxide inhalation, and eyes were enucleated into 4% paraformaldehyde. Wholemount retinas were processed for immunohistochemical staining with isolectin GS-IB4 (see histology section). The retinal avascular area was determined by digital imaging analysis (Photoshop CS; Adobe, San Jose, CA) using the free hand lasso tool. Avascular area results are as percentage total retinal area. Contralateral eyes were immersion fixed in Davidson's fixative (Poly Scientific, Bay Shore, NY) and prepared for paraffin embedding after removal of the cornea, lens, and hyaloid vessels. Serial paraffin sections were mounted on glass slides and stained with hematoxylin. Slides were scanned with an imaging system (ScanScope XT; Aperio, Vista, CA). Sections spaced 16 μm apart were identified and analyzed for neovascularization (Nuclear Algorithm v8.1; Aperio). A minimum of 26 sections from the peripheral retina to the optic nerve head were analyzed to determine the total number of epiretinal neovascular nuclei per animal. 
Mouse Model of Retinal Angiomatous Proliferation
Vldlr knockout mice (C57BL/6 × 129/SvEvBrd strain background) were generated from pairings of homozygous Vldlr knockout parents. Before the onset of subretinal neovascular development, on P12, Vldlr knockout pups were each administered a single intraperitoneal dose of 30 mg/kg 1.99.25 or anti–KLH mAb isotype-matched control. Pups underwent funduscopy and angiography on P20 and P30. Retinas for wholemount analysis were isolated from P16 eyes fixed in 4% paraformaldehyde. Wholemounts were stained with isolectin GS-IB4, and pathologic sprouts were counted from both eyes by masked examiners. The total number of sprouts in both eyes is reported. P20 and P30 eyes were placed in Davidson's fixative and processed for paraffin sections. Sections were stained with anti–PLVAP antibody, as described in the histology section. 
Blood-Retina Barrier Model
C57BL/6 wild-type mice were dosed intraperitoneally with 30 mg/kg 1.99.25 or anti–KLH mAb control once a week for 4 weeks. Funduscopy and angiography were performed at 2 and 4 weeks. Fluorescein angiograms were obtained by intraperitoneal injection of 60 μL fluorescein (10% AK-Fluor, 100 mg/mL) and capturing images at 1 (early phase), 3 (mid phase), and 6 (late phase) minutes after injection using a retinal imaging system (Genesis; Kowa, Nagoya, Japan) and a 78 D condensing lens. At the end of the study, mice were euthanatized and eyes were enucleated into Davidson's fixative for immunohistochemical examination of IgG and PLVAP, as described. 
Histology
For PLVAP and mouse IgG immunohistochemistry, eyes were fixed with Davidson's fixative and embedded in paraffin. Antigen retrieval was performed using citrate buffer, pH 6. Antibodies used were rat anti–PLVAP/MECA-32 (1:1000; BD PharMingen, San Diego, CA) together with biotinylated rabbit anti–rat IgG (1:400, mouse adsorbed; Vector Laboratories, Burlingame, CA) and biotinylated horse anti–mouse IgG (diluted 1:500; Vector Laboratories). Sections were counterstained with hematoxylin. 
Wholemount retinas from P5.5 neonates were labeled with a cocktail of rat anti-CD31 (BD PharMingen), rat anti-CD105 (BD PharMingen), and rat anti-CD102 (BD PharMingen). All antibodies were diluted (1000×) in PBS containing 0.1% Triton-X-100 (PBST). The following day, samples were washed in PBST three times. Retinas were then incubated overnight in a goat anti–rat Alexa 594 (Invitrogen, Carlsbad, CA) secondary antibody solution prepared at 1:500 in PBST. The next day, samples were washed in PBST and mounted on glass slides. Coverslips were applied using mounting medium (Fluoromount G; Electron Microscopy Sciences, Hatfield, PA). Retinal wholemounts for isolectin GS-IB4-Alexa 488 (1:1000; Invitrogen) staining were prepared from eyes fixed overnight in 4% paraformaldehyde. Isolectin GS-IB4 staining was performed overnight at a dilution of 1:1000 in PBST, containing 1 mM CaCl2 and 1 mM MgCl2. The next day, samples were washed three times in PBST, mounted on glass slides with mounting medium (Fluoromount G; Electron Microscopy Sciences), and imaged with a microscope (IX70; Olympus, Tokyo, Japan) equipped with a camera (DP70; Olympus). 
Analysis of retinal vascular beds in the developmental retinal angiogenesis study was performed in regions adjacent to the optic nerve head and at the periphery using confocal microscopy (SP5; Leica). A 20× z-stack of images in each region (775 μm × 775 μm × ∼35–40 μm depth) was acquired at a resolution of 1024 × 1024, 400-Hz scanning speed, 1-μm step size, pinhole of 60.66 μm, gain of 508.4 V, line average of two, and frame accumulation of two. Slices representing the superficial, intermediate, and deep vascular beds were isolated from each stack to generate a maximum projected image representing each layer. In general, the superficial vascular bed represented the first 14 μm, the intermediate layer represented the next 7 to 10 μm, and the deep vascular bed represented the final 16 to 19 μm regardless of treatment or area of retina imaged. 
Results
Antibody-Mediated Inhibition of FZD4 Disrupts Development and Maturation of the Outer Retinal Vasculature
Fzd4 knockout mice (Supplementary Fig. S1) were immunized with recombinant mouse FZD4-CRD. Multiple mice demonstrated robust anti-FZD4 serum titers. Hybridomas were generated, and their supernatants were screened for FZD4-CRD-specific mAbs by ELISA and based on binding to cells expressing human FZD4. After subcloning of candidate hybridomas, several purified mAbs were screened for the ability to block alkaline phosphatase (AP)-Norrin binding to cells expressing FZD4. One mAb, designated 1.99.25, demonstrated high potency in this assay (Supplementary Fig. S2A) and was also shown to antagonize β-catenin accumulation in FZD4/TSPAN12-expressing cells stimulated with either Norrin or WNT3A (Supplementary Fig. S2B). An isotype-matched anti–KLH negative control mAb did not inhibit ligand-induced β-catenin accumulation. 
The retinal vascular deficiency in Fzd4 knockout mice occurs during the first postnatal week. The superficial vascular plexus exhibits decreased complexity, and vascular progression toward the retinal periphery is significantly delayed. 11 To determine whether mAb 1.99.25 antagonized FZD4 signaling in vivo, wild-type pups were dosed on P0.5, and the superficial vasculature was examined on P5.5 (Supplementary Fig. S3). The anti–KLH control mAb did not affect the progression or complexity of the superficial vasculature (Supplementary Fig. S3A). However, a single administration of 1.99.25 shortly after birth delayed the progression of the superficial vascular plexus and decreased the vascular complexity (Supplementary Fig. S3B), recapitulating the phenotype observed in the Fzd4 knockout. 
To assess the direct contribution of FZD4 to outer retinal vasculature formation after the development of a normal superficial vascular plexus, wild-type mice were injected at P7 with 30 mg/kg of either anti–FZD4 mAb 1.99.25 or anti–KLH negative control mAb. Retinal vasculature was examined on P12 and P15. On P12, the superficial vascular plexus and deep capillary bed were well developed in the central (Fig. 1A) and peripheral (Fig. 1C) retina in the anti–KLH control mAb-treated mice. Horizontal sprouts were observed in the intermediate capillary bed in the central retina, whereas only vertical sprouts were observed in the periphery, reflecting the central to peripheral gradient in vascular bed maturation. No vascular abnormalities were observed in the anti–KLH mAb–treated group. In mice treated with mAb 1.99.25, vascular density was decreased in the central retina in both the intermediate and the deep capillary beds (Fig. 1B). Vascular density in the peripheral retina of mice treated with mAb 1.99.25 was even more profoundly decreased (Fig. 1D). The superficial vasculature, which had developed before dosing the FZD4 antibody on P7, was not affected by treatment (compare Figs. 1C and 1D, superficial). By P15, vascular development in anti–KLH mAb-treated mice had progressed significantly in the intermediate capillary bed in both central (Supplementary Fig. S4A) and peripheral (Supplementary Fig. S4C) retina. In contrast, central vascular density in mice treated with mAb 1.99.25 remained noticeably decreased in both the intermediate and the deep capillary beds (Supplementary Fig. S4B). Anti–FZD4 mAb treatment blocked the formation of the deep capillary bed in the peripheral retina and resulted in collections of endothelial cells located posterior to the superficial vascular bed. 
Figure 1.
 
Neutralizing FZD4 attenuates intraretinal vascular development. P7 mice received either anti–KLH mAb (A, C) or mAb 1.99.25 (B, D), and vasculature was examined on P12 using retinal wholemounts stained with isolectin GS-IB4-Alexa 488. (A, B) Representative images obtained in the central retina. (C, D) Representative images obtained in the peripheral retina. (A, C) Treatment with the anti–KLH mAb had no effect on vascular development. On P12, the superficial vascular plexus and the deep capillary bed are well developed in both the central (A) and peripheral (C) retina, whereas the intermediate capillary bed is still emerging. In the 1.99.25-treated mice, deep capillary bed formation is significantly reduced in the central retina (B) and almost absent in the peripheral retina (D). Vascular sprouts were present at the level of the intermediate capillary bed, but no continuous microvessels were observed in the 1.99.25-treated group. Scale bar, ∼250 μm.
Figure 1.
 
Neutralizing FZD4 attenuates intraretinal vascular development. P7 mice received either anti–KLH mAb (A, C) or mAb 1.99.25 (B, D), and vasculature was examined on P12 using retinal wholemounts stained with isolectin GS-IB4-Alexa 488. (A, B) Representative images obtained in the central retina. (C, D) Representative images obtained in the peripheral retina. (A, C) Treatment with the anti–KLH mAb had no effect on vascular development. On P12, the superficial vascular plexus and the deep capillary bed are well developed in both the central (A) and peripheral (C) retina, whereas the intermediate capillary bed is still emerging. In the 1.99.25-treated mice, deep capillary bed formation is significantly reduced in the central retina (B) and almost absent in the peripheral retina (D). Vascular sprouts were present at the level of the intermediate capillary bed, but no continuous microvessels were observed in the 1.99.25-treated group. Scale bar, ∼250 μm.
Previous studies in the Fzd4 knockout have shown that the vascular endothelium is fenestrated and that endothelial cells lacking FZD4 exhibit increased expression of the plasmalemma vesicle-associated protein Plvap (also called Meca-32), which encodes a structural component of endothelial fenestrae. The fenestration status of the vasculature in the anti–FZD4 antibody-treated mice was evaluated by PLVAP immunohistochemistry. Mice treated with the anti–KLH mAb on P7 and analyzed on P16 exhibited strong PLVAP expression in the normally fenestrated choriocapillaris, whereas the vasculature within the neural retina was negative, as expected (Supplementary Fig. S5A). Mice that received the FZD4 mAb exhibited strong PLVAP expression in both the choriocapillaris and the neural retinal vasculature (Supplementary Fig. S5B). Transmission electron microscopy confirmed the presence of fenestrations within the retina vascular endothelium in mice that received the FZD4 mAb (Supplementary Fig. S5D). Tight junctions were apparent in mice that received either 1.99.25 or the anti–KLH mAb (Supplementary Fig. S5C). 
FZD4 Neutralizing Antibody Inhibits Physiological, but Not Pathologic, Angiogenesis in the OIR Model
To begin to understand the role of FZD4 in retinal vascular disease, we evaluated the effects of mAb 1.99.25 in the mouse OIR model. This model mimics human retinopathy of prematurity and involves both physiological angiogenesis, resulting in normal retinal vascular structures, and pathologic angiogenesis, resulting in the formation of aberrant neovascular tufts at the vitreoretinal boundary. Wild-type mice were subjected to hyperoxia from P7 to P12 and then were returned to room air, at which time they were administered 30 mg/kg of either mAb 1.99.25 or the anti–KLH control mAb daily from P12 to P16. Retinal wholemounts collected from anti–KLH-treated mice on P17 exhibited a central avascular area attributed to hyperoxia-induced vaso-obliteration, a hallmark of this model (Fig. 2A). The avascular area was manifestly increased in mice treated with mAb 1.99.25 (Fig. 2B). Approximately 25.8% ± 2.1% (SD) of the total retinal area lacked vessels in the 1.99.25-treated animals compared with 17.4% ± 2.6% for the anti–KLH-treated group (Fig. 2E). 
Figure 2.
 
Administration of FZD4-neutralizing mAb delays vascular recovery in the OIR model. (A) Retina stained with isolectin GS-IB4 to visualize the vasculature and area of vaso-obliteration (outline) in a P17 animal treated with the anti–KLH control antibody. (B) The area of vaso-obliteration (outline) is larger in the 1.99.25-treated mice. (C) Paraffin section of an eye from an anti–KLH control-treated animal stained with hematoxylin to visualize nuclei associated with neovascular tufts (arrow) located above the ganglion cell layer (gcl). (D) Neovascular tufts (arrow) are present in the 1.99.25-treated animals. Inhibition of FZD4 resulted in clusters (asterisks) of vascular cells in the inner plexiform layer (ipl). (E) Administration of 1.99.25 increased the area of vaso-obliteration on P17 compared with anti–KLH control-treated animals. Each point represents an individual animal. The mean is indicated by the horizontal line (P < 0.0001 using two-tailed Student's unpaired t-test). (F) The total number of nuclei above the inner limiting membrane was not significantly different (P = 0.53) in anti–KLH- versus 1.99.25-treated animals. Scale bars: 600 μm (A, B); 50 μm (C, D).
Figure 2.
 
Administration of FZD4-neutralizing mAb delays vascular recovery in the OIR model. (A) Retina stained with isolectin GS-IB4 to visualize the vasculature and area of vaso-obliteration (outline) in a P17 animal treated with the anti–KLH control antibody. (B) The area of vaso-obliteration (outline) is larger in the 1.99.25-treated mice. (C) Paraffin section of an eye from an anti–KLH control-treated animal stained with hematoxylin to visualize nuclei associated with neovascular tufts (arrow) located above the ganglion cell layer (gcl). (D) Neovascular tufts (arrow) are present in the 1.99.25-treated animals. Inhibition of FZD4 resulted in clusters (asterisks) of vascular cells in the inner plexiform layer (ipl). (E) Administration of 1.99.25 increased the area of vaso-obliteration on P17 compared with anti–KLH control-treated animals. Each point represents an individual animal. The mean is indicated by the horizontal line (P < 0.0001 using two-tailed Student's unpaired t-test). (F) The total number of nuclei above the inner limiting membrane was not significantly different (P = 0.53) in anti–KLH- versus 1.99.25-treated animals. Scale bars: 600 μm (A, B); 50 μm (C, D).
Epiretinal neovascular tufts anterior to the inner limiting membrane were observed in mice treated with either anti–KLH control mAb (Fig. 2C) or mAb 1.99.25 (Fig. 2D); quantification of cell nuclei associated with these tufts did not reveal a significant difference between treatment groups (Fig. 2F). The total number of nuclei in the 1.99.25-treated group averaged 2364 ± 907 (SD), and in the KLH mAb-treated control group it averaged 2070 ± 916. Interestingly, collections of vascular cells were frequently observed in the inner plexiform layer (IPL) of the 1.99.25-treated mice (Fig. 2D, asterisks) but not of the anti–KLH-treated mice. The effects of mAb 1.99.25 in the mouse OIR model show that inhibition of FZD4 prevents revascularization of the retina after hyperoxia without affecting the formation of the pathologic neovascular tufts. Importantly, elevated ocular VEGF-A protein levels, induced by the relative hypoxia in the OIR model, were not affected by treatment with either anti–KLH mAb or FZD4-neutralizing mAb 1.99.25 (Supplementary Fig. S6). 
FZD4 Neutralizing Antibody Inhibits Pathologic Angiogenesis in the Vldlr Knockout Model of Retinal Angiomatous Proliferation
Fzd4 knockout mice do not exhibit vascular abnormalities in the choriocapillaris, and our results with mAb 1.99.25 in the OIR model show no effect of inhibiting FZD4 on pathologic angiogenesis originating from the superficial retinal vasculature. To assess the role played by FZD4 in pathologic angiogenesis derived from the outer retinal vasculature, we turned to the very low density lipoprotein receptor (VLDLR) knockout model of retinal angiomatous proliferation (RAP). Vldlr knockout mice were generated by homologous recombination (Supplementary Fig. S7). Intraretinal neovascular sprouts were observed as early as P14, consistent with previous reports in Vldlr knockout mice. 14 Subsequently, these sprouts extended through the outer nuclear layer, toward the retinal pigment epithelium, and finally anastomosed with each other and the choriocapillaris. 
Vldlr knockout mice each received a single intraperitoneal injection of 30 mg/kg anti–KLH control or mAb 1.99.25 on P12. Four days later, on P16, retinal wholemounts were prepared and stained with isolectin B4 Alexa 488 to visualize the vasculature. Inhibition of FZD4 resulted in a dramatic 98% decrease in the incidence of pathologic, intraretinal neovascular sprouts compared with anti–KLH-treated control mice (compare Figs. 3A and 3B). The average number of neovascular sprouts was 131 ± 70 (SD) in the anti–KLH group compared with 2.6 ± 1.9 in the FZD4 mAb-treated group (P < 0.0001; Fig. 3C). To explore the duration of this effect, a single injection of either 1.99.25 or anti–KLH control mAb was given to homozygous Vldlr knockout mice on P12, followed by angiography and histology on either P20 or P30. Angiograms obtained on either P20 or P30 from mice treated with the anti–KLH control mAb exhibited progressively worsening areas of focal fluorescein leakage emanating from the deep capillary beds (Figs. 4A, P20; 4C, P30). In contrast, focal areas of fluorescein leakage were not observed in angiograms obtained from P20 or P30 Vldlr knockout mice treated with mAb 1.99.25. Instead, 1.99.25-treated mice exhibited diffuse fluorescein leakage at P20 (Fig. 4B) that was still present at P30 but was reduced in intensity (compare Figs. 4B and 4D). 
Figure 3.
 
FZD4-neutralizing mAb inhibits pathologic angiogenesis in the Vldlr model. Vldlr knockout pups received either a single injection of 30 mpk 1.99.25 or anti–KLH control on P12. On P16, retinal wholemounts were prepared from both eyes of each animal, stained with isolectin GS-IB4-Alexa 488, and imaged using confocal microscopy. (A, top) Confocal projection (z plane) from a Vldlr knockout that received the anti–KLH mAb control. Numerous pathologic sprouts were observed in the outer nuclear layer (onl), beneath the vascular beds located in the outer plexiform layer (opl) and the inner plexiform layer (ipl). A confocal line scan in the x-y plane located within the photoreceptors revealed many pathologic sprouts in the anti–KLH mAb-treated Vldlr knockout mice. The optic nerve head is circled. (B) Confocal projection (z plane) illustrates the dramatic reduction in pathologic sprouts in Vldlr knockout mice treated with 1.99.25. Image (x-y plane) taken at the level of the photoreceptors in a retina isolated from a Vldlr knockout mouse that received mAb 1.99.25. A few sprouts were present near the optic nerve head (circle). Scale bar, ∼100 μm. (C) Quantification of vascular sprouts in either anti–KLH control- or 1.99.25-treated animals. Each point represents the average number of neovascular sprouts in both retinas of a single animal. Inhibition of FZD4 significantly blocks the formation of these pathologic sprouts (P < 0.0001).
Figure 3.
 
FZD4-neutralizing mAb inhibits pathologic angiogenesis in the Vldlr model. Vldlr knockout pups received either a single injection of 30 mpk 1.99.25 or anti–KLH control on P12. On P16, retinal wholemounts were prepared from both eyes of each animal, stained with isolectin GS-IB4-Alexa 488, and imaged using confocal microscopy. (A, top) Confocal projection (z plane) from a Vldlr knockout that received the anti–KLH mAb control. Numerous pathologic sprouts were observed in the outer nuclear layer (onl), beneath the vascular beds located in the outer plexiform layer (opl) and the inner plexiform layer (ipl). A confocal line scan in the x-y plane located within the photoreceptors revealed many pathologic sprouts in the anti–KLH mAb-treated Vldlr knockout mice. The optic nerve head is circled. (B) Confocal projection (z plane) illustrates the dramatic reduction in pathologic sprouts in Vldlr knockout mice treated with 1.99.25. Image (x-y plane) taken at the level of the photoreceptors in a retina isolated from a Vldlr knockout mouse that received mAb 1.99.25. A few sprouts were present near the optic nerve head (circle). Scale bar, ∼100 μm. (C) Quantification of vascular sprouts in either anti–KLH control- or 1.99.25-treated animals. Each point represents the average number of neovascular sprouts in both retinas of a single animal. Inhibition of FZD4 significantly blocks the formation of these pathologic sprouts (P < 0.0001).
Figure 4.
 
FZD4-neutralizing mAb induces diffuse vascular leakage in Vldlr knockout mice. Vldlr knockout mice were treated on P12 with either anti–KLH negative control or FZD4 mAb 1.99.25. (A) Focal vascular leakage is present on P20 in mice treated with the anti–KLH antibody. (B) Diffuse vascular leakage is present on P20 in mice treated with mAb 1.99.25. (C) Focal vascular leakage is present at P30 in mice treated with the anti–KLH antibody. (D) Faint, diffuse vascular leakage is present at P30 in mice treated with mAb 1.99.25. The angiogram shown in (B) was obtained approximately 1 minute after fluorescein injection. The remaining angiograms were captured approximately 6 minutes after injection.
Figure 4.
 
FZD4-neutralizing mAb induces diffuse vascular leakage in Vldlr knockout mice. Vldlr knockout mice were treated on P12 with either anti–KLH negative control or FZD4 mAb 1.99.25. (A) Focal vascular leakage is present on P20 in mice treated with the anti–KLH antibody. (B) Diffuse vascular leakage is present on P20 in mice treated with mAb 1.99.25. (C) Focal vascular leakage is present at P30 in mice treated with the anti–KLH antibody. (D) Faint, diffuse vascular leakage is present at P30 in mice treated with mAb 1.99.25. The angiogram shown in (B) was obtained approximately 1 minute after fluorescein injection. The remaining angiograms were captured approximately 6 minutes after injection.
Histologic sections from P20 and P30 homozygous Vldlr knockout mice were studied to further understand vascular responses to mAb 1.99.25. Collections of endothelial cells were observed adjacent to the RPE in Vldlr knockout mice treated with the anti–KLH control mAb. These pathologic vascular structures were PLVAP immunoreactive (Figs. 5A, 5C, arrows), which is consistent with their leaky nature, as demonstrated by angiography. Mice treated with FZD4 neutralizing antibody lacked these lesions (Figs. 5B, 5D). Strikingly, though normal outer retinal vasculature in anti–KLH mAb-treated mice on P20 and P30 was negative for PLVAP immunoreactivity, as would be expected in mice of these ages, the capillary beds of Vldlr knockout mice treated with the FZD4 neutralizing antibody exhibited intense immunoreactivity for PLVAP on P20 (Fig. 5B). PLVAP immunoreactivity was detectable in retinal capillaries in P30 mice that were treated 18 days earlier with the FZD4 mAb, but it was less intense than that observed on P20 (Figs. 5B, 5D, compare arrows). These data show that inhibition of FZD4 blocks formation of the RAP-like lesions that arise from the outer retinal vasculature in Vldlr knockout mice. Angiogram and PLVAP expression results further suggest that FZD4 plays an important role in regulating the permeability of the developing outer retinal vasculature in these mice. 
Figure 5.
 
FZD4-neutralizing mAb induces PLVAP expression. PLVAP immunohistochemistry was performed on retinas from P20 (A, B) and P30 (C, D) Vldlr knockout mice treated with either anti–KLH control (A, C) or 1.99.25 (B, D) on P12. (A) Focal, pathologic sprouts are present in anti–KLH-treated P20 mice. These sprouts are adjacent to the retinal pigment epithelium (RPE) and are intensely immunoreactive for PLVAP (brown stain). (B) Pathologic sprouts were sparse in 1.99.25-treated cases. However, much of the vasculature (arrows) in these P20 mice exhibited intense PLVAP immunoreactivity. Compare capillaries in the inner plexiform layer (ipl) and the outer plexiform layer (opl) in A and B. (C) Pathologic sprouts (arrow) remain intensely PLVAP immunoreactive in P30 mice treated with the anti–KLH antibody. PLVAP immunoreactivity is not observed elsewhere in the retinal vasculature. (D) Pathologic sprouts were extremely rare in 1.99.25-treated mice. The capillary beds still exhibited PLVAP immunoreactivity on P30, though less intensely than on P20. Scale bar, ∼50 μm.
Figure 5.
 
FZD4-neutralizing mAb induces PLVAP expression. PLVAP immunohistochemistry was performed on retinas from P20 (A, B) and P30 (C, D) Vldlr knockout mice treated with either anti–KLH control (A, C) or 1.99.25 (B, D) on P12. (A) Focal, pathologic sprouts are present in anti–KLH-treated P20 mice. These sprouts are adjacent to the retinal pigment epithelium (RPE) and are intensely immunoreactive for PLVAP (brown stain). (B) Pathologic sprouts were sparse in 1.99.25-treated cases. However, much of the vasculature (arrows) in these P20 mice exhibited intense PLVAP immunoreactivity. Compare capillaries in the inner plexiform layer (ipl) and the outer plexiform layer (opl) in A and B. (C) Pathologic sprouts (arrow) remain intensely PLVAP immunoreactive in P30 mice treated with the anti–KLH antibody. PLVAP immunoreactivity is not observed elsewhere in the retinal vasculature. (D) Pathologic sprouts were extremely rare in 1.99.25-treated mice. The capillary beds still exhibited PLVAP immunoreactivity on P30, though less intensely than on P20. Scale bar, ∼50 μm.
Inhibition of FZD4 in Normal, Mature Retinal Vasculature Results in a Loss of Barrier Function
To understand whether FZD4 continues to regulate barrier function in mature retinal vasculature, adult naive C57BL/6J mice were treated with either the anti–KLH control mAb or 1.99.25 at 30 mg/kg weekly for up to 4 weeks. After baseline angiography, administration of anti–KLH mAb did not appreciably affect retinal vascular permeability after 2 weeks (two doses total) or 4 weeks (four doses total) of treatment (Fig. 6A). The FZD4 mAb, however, caused diffuse fluorescein leakage in the deep capillary bed after 2 weeks and 4 weeks (Fig. 6B) of treatment. Mice treated for 4 weeks with the anti–KLH negative control did not exhibit PLVAP immunoreactivity in the neural retinal vasculature, whereas staining was present in the choroidal capillaries, consistent with their nonbarrier phenotype (Fig. 6C). In contrast, PLVAP immunoreactivity was observed in retinal deep capillaries in mice treated with mAb 1.99.25 (Fig. 6D). 
Figure 6.
 
FZD4-neutralizing mAb induces fluorescein leakage and PLVAP expression in adult mice. Angiograms of adult wild-type C57BL/6J mice obtained after four weekly doses of either anti–KLH (A) or 1.99.25 (B). Treatment with the FZD4 mAb resulted in diffuse fluorescein leakage. Retinal capillaries (C, arrow) lack PLVAP immunoreactivity in anti–KLH-treated mice. Treatment with mAb 1.99.25 results in increased PLVAP expression (brown stain) in retinal capillaries (D, arrows). (E) Immunohistochemistry with anti–mouse IgG intensely stains retinal vasculature in anti–KLH-treated animals. The vitreous (asterisk) is weakly stained. (F) Treatment with 1.99.25 resulted in intense IgG immunoreactivity in the vitreous (asterisk) and outer plexiform layer (opl). IgG-like immunoreactivity also appears at the level of cone outer segments. Scale bar, ∼50 μm (CF).
Figure 6.
 
FZD4-neutralizing mAb induces fluorescein leakage and PLVAP expression in adult mice. Angiograms of adult wild-type C57BL/6J mice obtained after four weekly doses of either anti–KLH (A) or 1.99.25 (B). Treatment with the FZD4 mAb resulted in diffuse fluorescein leakage. Retinal capillaries (C, arrow) lack PLVAP immunoreactivity in anti–KLH-treated mice. Treatment with mAb 1.99.25 results in increased PLVAP expression (brown stain) in retinal capillaries (D, arrows). (E) Immunohistochemistry with anti–mouse IgG intensely stains retinal vasculature in anti–KLH-treated animals. The vitreous (asterisk) is weakly stained. (F) Treatment with 1.99.25 resulted in intense IgG immunoreactivity in the vitreous (asterisk) and outer plexiform layer (opl). IgG-like immunoreactivity also appears at the level of cone outer segments. Scale bar, ∼50 μm (CF).
To assess whether the vascular leakage induced by inhibiting FZD4 enabled the extravasation of molecules larger than fluorescein (376 Da), sections were stained for the presence of mouse IgG (approximately 150,000 Da). In anti–KLH control-treated retinas, only the vasculature itself was immunoreactive with mouse IgG (Fig. 6E), whereas in animals receiving the 1.99.25 mAb, intense IgG staining was also observed in the vitreous (Fig. 6F, asterisk). These results show that the inhibition of FZD4 in mature retinal vasculature leads to a loss of blood-retina barrier integrity that permits both low and high molecular weight molecules to pass out of circulation into a normally privileged tissue. 
Discussion
We have generated a monoclonal antibody (1.99.25) that binds the FZD4-CRD and antagonizes Norrin and WNT3A-induced β-catenin accumulation in vitro. Administration of this antibody to wild-type mice on P0 recapitulated the delay in the progression of the superficial vascular plexus observed during the first postnatal week in the Fzd4 knockout, 9,11 validating mAb 1.99.25 as a valuable tool to further characterize FZD4 function in vascular physiology. Administration of 1.99.25 on P7 (subsequent to the normal development of the superficial vascular plexus) inhibited the formation of the intermediate and deep capillary beds in a central to peripheral gradient. This conclusively shows that FZD4 is directly involved in the formation of these capillary beds and that the defective outer retinal vascular development seen in the Fzd4 knockout is not simply a consequence of delayed formation of the superficial vasculature. 
To determine the potential contribution of the FZD4 signaling pathway to VEGF-driven pathologic neovascularization, mAb 1.99.25 was studied in the OIR model. In the hyperoxic phase of the OIR model (P7-P12), the central superficial vasculature undergoes vascular regression and apoptosis, resulting in a large area of vaso-obliteration. This occurs, in part, because of decreased levels of VEGF-A resulting from increased oxygen saturation. Formation of the deep capillary beds is also suspended during the hyperoxic phase. The relative hypoxia in the retina induced on return to room air on P12 leads to a marked increase in VEGF-A levels, 15,16 and the superficial vasculature in the central retina repopulates the area of vaso-obliteration and deep capillary bed formation resumes. 
In OIR mice treated with mAb 1.99.25, the area of vaso-obliteration remained large compared with the negative control group, and regrowth of capillary beds was prevented despite high levels of VEGF-A. Similarly, mice lacking Ndp fail to vascularize the inner retina, leading to increased levels of hypoxia-inducible factor-1α and VEGF-A. 17 The knockout phenotype and our results with mAb 1.99.25 in the OIR model suggest that FZD4 is required for the retinal vasculature to undergo angiogenesis when VEGF is at levels that would be expected to elicit such a response. 
To further investigate the relationship between FZD4 and VEGF-induced retinal angiogenesis, we turned to the Vldlr knockout model of human RAP. VLDLR is a member of the LDL receptor family and plays a critical role in CNS development in mice. 18 In addition to affecting neuronal migration, knockout of Vldlr causes abnormal neovascularization, characterized by the proliferation of retinal vessels through the outer retina into the subretinal space and, in some instances, merging with the choriocapillaris. 14,19 These vascular lesions are associated with locally elevated levels of VEGF, and their formation is significantly attenuated by pharmacologic inhibition of VEGF signaling. 20,21 The dramatic reduction in pathologic sprouting after treatment with mAb 1.99.25 demonstrates that development of these lesions is also critically dependent on FZD4. 
FZD4 is the only known Norrin receptor. The highly similar retinal vascular phenotypes of Ndp and Fzd4 knockout mice suggest that Norrin's effects during retinal blood vessel development are likely to be mediated exclusively through FZD4 and, conversely, that Norrin is the predominant ligand for FZD4 in this compartment. Patterning defects in the retinal vasculature of Ndp knockout mice can be rescued by ectopic, lens-specific overexpression of Norrin, 22 showing that local Norrin gradients do not act to guide vascular sprouting. Our results with mAb 1.99.22 in both the OIR and RAP models, which involve strongly proangiogenic environments in which VEGF plays a dominant role in directing vascular expansion, further support the idea that FZD4 signaling is necessary for endothelial cells to be angiogenically competent. 9  
Vldlr knockout mice exhibit focal vascular leakage emanating from the pathologic sprouts. Vascular permeability of these lesions was also evidenced in our experiments by their strong expression of the endothelial pore component, PLVAP, which was absent in nonlesional capillaries. During development, endothelial cell expression of PLVAP is downregulated as the vasculature matures and assumes a barrier phenotype. 23 Treatment of Vldlr knockout pups with mAb 1.99.25 induced diffuse vascular leakage within the retina and upregulation of PLVAP expression throughout the deep capillary bed. Pan-retinal expression of PLVAP was also observed after treatment with FZD4 mAb in naive wild-type mice on P0 and P7 and in the OIR model. To study this effect in the mature, healthy retina, we treated naive, adult, wild-type mice, causing the same vascular leakage and re-expression of PLVAP, both of which were reversible after the cessation of antibody treatment. Beyond the induction of vascular permeability, we did not observe any obvious loss of normal capillary architecture after 4 weeks of antibody treatment. 
During angiogenesis in the CNS, endothelial cells interact with resident cells in the neural parenchyma and lose expression of PLVAP while upregulating proteins involved in barrier structure and function (e.g., Claudin-3 and GLUT1). Seminal work by Stewart and Wiley 24 showed that vascular acquisition of blood-brain barrier characteristics is induced by factors residing within the brain microenvironment. In vitro and in vivo data demonstrate that this change in endothelial cell phenotype is driven by several Wnts and increased activity of nuclear β-catenin. 2 4 In the developing and adult retina, Müller cells express Norrin, whereas FZD4 is more widely expressed. 11,25 Among their many critical functions, Müller cells construct the glia limitans of retinal capillary beds and, together with astrocytes, induce the formation of the blood-retina barrier. 26,27 Fzd4,Ndp, Lrp5, and Tspan12 knockout mice all exhibit aberrant fenestration of vascular endothelium, and endothelial cells from these knockouts express increased levels of PLVAP. 8,9 Antibody-mediated inhibition of FZD4 produced a fenestrated phenotype in mature capillaries, perhaps through neutralization of Norrin-FZD4 interactions at the level of Müller glia and vascular cells. Over a 2-week mAb washout period, PLVAP expression induced by FZD4 inhibition decreased and fluorescein leakage was no longer observed. Reversion of the vasculature to a mature barrier phenotype indicates a requirement for sustained FZD4 activity under normal conditions and suggests that this pathway is important for physiological regulation of vascular permeability. 
Supplementary Materials
Text s1, PDF - Text s1, PDF 
Footnotes
 Disclosure: K.T. Paes, Lexicon Pharmaceuticals, Inc. (E); E. Wang, Lexicon Pharmaceuticals, Inc. (E); K. Henze, Lexicon Pharmaceuticals, Inc. (E); P. Vogel, Lexicon Pharmaceuticals, Inc. (E); R. Read, Lexicon Pharmaceuticals, Inc. (E); A. Suwanichkul, Lexicon Pharmaceuticals, Inc. (E); L.L. Kirkpatrick, Lexicon Pharmaceuticals, Inc. (E); D. Potter, Lexicon Pharmaceuticals, Inc. (E); M.M. Newhouse, Lexicon Pharmaceuticals, Inc. (E); D.S. Rice, Lexicon Pharmaceuticals, Inc. (E)
The authors thank Ken Coker, Billie Key, Xiao Feng, Colin Lin, and Joe Roethele for expert technical assistance. 
References
Cattelino A Liebner S Gallini R . The conditional inactivation of the beta-catenin gene in endothelial cells causes a defective vascular pattern and increased vascular fragility. J Cell Biol. 2003;162:1111–1122. [CrossRef] [PubMed]
Daneman R Agalliu D Zhou L Kuhnert F Kuo CJ Barres BA . Wnt/beta-catenin signaling is required for CNS, but not non-CNS, angiogenesis. Proc Natl Acad Sci U S A. 2009;106:641–646. [CrossRef] [PubMed]
Liebner S Corada M Bangsow T . Wnt/beta-catenin signaling controls development of the blood-brain barrier. J Cell Biol. 2008;183:409–417. [CrossRef] [PubMed]
Stenman JM Rajagopal J Carroll TJ Ishibashi M McMahon J McMahon AP . Canonical Wnt signaling regulates organ-specific assembly and differentiation of CNS vasculature. Science. 2008;322:1247–1250. [CrossRef] [PubMed]
van Amerongen R Nusse R . Towards an integrated view of Wnt signaling in development. Development. 2009;136:3205–3214. [CrossRef] [PubMed]
Tam SJ Watts RJ . Connecting vascular and nervous system development: angiogenesis and the blood-brain barrier. Annu Rev Neurosci. 2010;33:379–408. [CrossRef] [PubMed]
Ye X Wang Y Nathans J . The Norrin/Frizzled4 signaling pathway in retinal vascular development and disease. Trends Mol Med. 2010;16:417–425. [CrossRef] [PubMed]
Junge HJ Yang S Burton JB . TSPAN12 regulates retinal vascular development by promoting Norrin- but not Wnt-induced FZD4/beta-catenin signaling. Cell. 2009;139:299–311. [CrossRef] [PubMed]
Ye X Wang Y Cahill H . Norrin, frizzled-4, and Lrp5 signaling in endothelial cells controls a genetic program for retinal vascularization. Cell. 2009;139:285–298. [CrossRef] [PubMed]
Wang Y Huso D Cahill H Ryugo D Nathans J . Progressive cerebellar, auditory, and esophageal dysfunction caused by targeted disruption of the frizzled-4 gene. J Neurosci. 2001;21:4761–4771. [PubMed]
Xu Q Wang Y Dabdoub A . Vascular development in the retina and inner ear: control by Norrin and Frizzled-4, a high-affinity ligand-receptor pair. Cell. 2004;116:883–895. [CrossRef] [PubMed]
Wattler S Kelly M Nehls M . Construction of gene targeting vectors from lambda KOS genomic libraries. BioTechniques. 1999;26:1150–1156, 1158, 1160. [PubMed]
Smith LE Wesolowski E McLellan A . Oxygen-induced retinopathy in the mouse. Invest Ophthalmol Vis Sci. 1994;35:101–111. [PubMed]
Heckenlively JR Hawes NL Friedlander M . Mouse model of subretinal neovascularization with choroidal anastomosis. Retina. 2003;23:518–522. [CrossRef] [PubMed]
Pierce EA Foley ED Smith LE . Regulation of vascular endothelial growth factor by oxygen in a model of retinopathy of prematurity. Arch Ophthalmol. 1996;114:1219–1228. [CrossRef] [PubMed]
Pierce EA Avery RL Foley ED Aiello LP Smith LE . Vascular endothelial growth factor/vascular permeability factor expression in a mouse model of retinal neovascularization. Proc Natl Acad Sci U S A. 1995;92:905–909. [CrossRef] [PubMed]
Luhmann UF Lin J Acar N . Role of the Norrie disease pseudoglioma gene in sprouting angiogenesis during development of the retinal vasculature. Invest Ophthalmol Vis Sci. 2005;46:3372–3382. [CrossRef] [PubMed]
Rice DS Curran T . Role of the reelin signaling pathway in central nervous system development. Annu Rev Neurosci. 2001;24:1005–1039. [CrossRef] [PubMed]
Yannuzzi LA Negrao S Iida T . Retinal angiomatous proliferation in age-related macular degeneration. Retina. 2001;21:416–434. [CrossRef] [PubMed]
Chen Y Hu Y Lu K Flannery JG Ma JX . Very low density lipoprotein receptor, a negative regulator of the wnt signaling pathway and choroidal neovascularization. J Biol Chem. 2007;282:34420–34428. [CrossRef] [PubMed]
Dorrell MI Aguilar E Jacobson R . Antioxidant or neurotrophic factor treatment preserves function in a mouse model of neovascularization-associated oxidative stress. J Clin Invest. 2009;119:611–623. [CrossRef] [PubMed]
Ohlmann A Seitz R Braunger B Seitz D Bosl MR Tamm ER . Norrin promotes vascular regrowth after oxygen-induced retinal vessel loss and suppresses retinopathy in mice. J Neurosci. 2010;30:183–193. [CrossRef] [PubMed]
Hallmann R Mayer DN Berg EL Broermann R Butcher EC . Novel mouse endothelial cell surface marker is suppressed during differentiation of the blood brain barrier. Dev Dyn. 1995;202:325–332. [CrossRef] [PubMed]
Stewart PA Wiley MJ . Developing nervous tissue induces formation of blood-brain barrier characteristics in invading endothelial cells: a study using quail-chick transplantation chimeras. Dev Biol. 1981;84:183–192. [CrossRef] [PubMed]
Ye X Smallwood P Nathans J . Expression of the Norrie disease gene (Ndp) in developing and adult mouse eye, ear, and brain. Gene Expr Patterns. 2010;11:151–155. [CrossRef] [PubMed]
Janzer RC Raff MC . Astrocytes induce blood-brain barrier properties in endothelial cells. Nature. 1987;325:253–257. [CrossRef] [PubMed]
Tout S Chan-Ling T Hollander H Stone J . The role of Muller cells in the formation of the blood-retinal barrier. Neuroscience. 1993;55:291–301. [CrossRef] [PubMed]
Figure 1.
 
Neutralizing FZD4 attenuates intraretinal vascular development. P7 mice received either anti–KLH mAb (A, C) or mAb 1.99.25 (B, D), and vasculature was examined on P12 using retinal wholemounts stained with isolectin GS-IB4-Alexa 488. (A, B) Representative images obtained in the central retina. (C, D) Representative images obtained in the peripheral retina. (A, C) Treatment with the anti–KLH mAb had no effect on vascular development. On P12, the superficial vascular plexus and the deep capillary bed are well developed in both the central (A) and peripheral (C) retina, whereas the intermediate capillary bed is still emerging. In the 1.99.25-treated mice, deep capillary bed formation is significantly reduced in the central retina (B) and almost absent in the peripheral retina (D). Vascular sprouts were present at the level of the intermediate capillary bed, but no continuous microvessels were observed in the 1.99.25-treated group. Scale bar, ∼250 μm.
Figure 1.
 
Neutralizing FZD4 attenuates intraretinal vascular development. P7 mice received either anti–KLH mAb (A, C) or mAb 1.99.25 (B, D), and vasculature was examined on P12 using retinal wholemounts stained with isolectin GS-IB4-Alexa 488. (A, B) Representative images obtained in the central retina. (C, D) Representative images obtained in the peripheral retina. (A, C) Treatment with the anti–KLH mAb had no effect on vascular development. On P12, the superficial vascular plexus and the deep capillary bed are well developed in both the central (A) and peripheral (C) retina, whereas the intermediate capillary bed is still emerging. In the 1.99.25-treated mice, deep capillary bed formation is significantly reduced in the central retina (B) and almost absent in the peripheral retina (D). Vascular sprouts were present at the level of the intermediate capillary bed, but no continuous microvessels were observed in the 1.99.25-treated group. Scale bar, ∼250 μm.
Figure 2.
 
Administration of FZD4-neutralizing mAb delays vascular recovery in the OIR model. (A) Retina stained with isolectin GS-IB4 to visualize the vasculature and area of vaso-obliteration (outline) in a P17 animal treated with the anti–KLH control antibody. (B) The area of vaso-obliteration (outline) is larger in the 1.99.25-treated mice. (C) Paraffin section of an eye from an anti–KLH control-treated animal stained with hematoxylin to visualize nuclei associated with neovascular tufts (arrow) located above the ganglion cell layer (gcl). (D) Neovascular tufts (arrow) are present in the 1.99.25-treated animals. Inhibition of FZD4 resulted in clusters (asterisks) of vascular cells in the inner plexiform layer (ipl). (E) Administration of 1.99.25 increased the area of vaso-obliteration on P17 compared with anti–KLH control-treated animals. Each point represents an individual animal. The mean is indicated by the horizontal line (P < 0.0001 using two-tailed Student's unpaired t-test). (F) The total number of nuclei above the inner limiting membrane was not significantly different (P = 0.53) in anti–KLH- versus 1.99.25-treated animals. Scale bars: 600 μm (A, B); 50 μm (C, D).
Figure 2.
 
Administration of FZD4-neutralizing mAb delays vascular recovery in the OIR model. (A) Retina stained with isolectin GS-IB4 to visualize the vasculature and area of vaso-obliteration (outline) in a P17 animal treated with the anti–KLH control antibody. (B) The area of vaso-obliteration (outline) is larger in the 1.99.25-treated mice. (C) Paraffin section of an eye from an anti–KLH control-treated animal stained with hematoxylin to visualize nuclei associated with neovascular tufts (arrow) located above the ganglion cell layer (gcl). (D) Neovascular tufts (arrow) are present in the 1.99.25-treated animals. Inhibition of FZD4 resulted in clusters (asterisks) of vascular cells in the inner plexiform layer (ipl). (E) Administration of 1.99.25 increased the area of vaso-obliteration on P17 compared with anti–KLH control-treated animals. Each point represents an individual animal. The mean is indicated by the horizontal line (P < 0.0001 using two-tailed Student's unpaired t-test). (F) The total number of nuclei above the inner limiting membrane was not significantly different (P = 0.53) in anti–KLH- versus 1.99.25-treated animals. Scale bars: 600 μm (A, B); 50 μm (C, D).
Figure 3.
 
FZD4-neutralizing mAb inhibits pathologic angiogenesis in the Vldlr model. Vldlr knockout pups received either a single injection of 30 mpk 1.99.25 or anti–KLH control on P12. On P16, retinal wholemounts were prepared from both eyes of each animal, stained with isolectin GS-IB4-Alexa 488, and imaged using confocal microscopy. (A, top) Confocal projection (z plane) from a Vldlr knockout that received the anti–KLH mAb control. Numerous pathologic sprouts were observed in the outer nuclear layer (onl), beneath the vascular beds located in the outer plexiform layer (opl) and the inner plexiform layer (ipl). A confocal line scan in the x-y plane located within the photoreceptors revealed many pathologic sprouts in the anti–KLH mAb-treated Vldlr knockout mice. The optic nerve head is circled. (B) Confocal projection (z plane) illustrates the dramatic reduction in pathologic sprouts in Vldlr knockout mice treated with 1.99.25. Image (x-y plane) taken at the level of the photoreceptors in a retina isolated from a Vldlr knockout mouse that received mAb 1.99.25. A few sprouts were present near the optic nerve head (circle). Scale bar, ∼100 μm. (C) Quantification of vascular sprouts in either anti–KLH control- or 1.99.25-treated animals. Each point represents the average number of neovascular sprouts in both retinas of a single animal. Inhibition of FZD4 significantly blocks the formation of these pathologic sprouts (P < 0.0001).
Figure 3.
 
FZD4-neutralizing mAb inhibits pathologic angiogenesis in the Vldlr model. Vldlr knockout pups received either a single injection of 30 mpk 1.99.25 or anti–KLH control on P12. On P16, retinal wholemounts were prepared from both eyes of each animal, stained with isolectin GS-IB4-Alexa 488, and imaged using confocal microscopy. (A, top) Confocal projection (z plane) from a Vldlr knockout that received the anti–KLH mAb control. Numerous pathologic sprouts were observed in the outer nuclear layer (onl), beneath the vascular beds located in the outer plexiform layer (opl) and the inner plexiform layer (ipl). A confocal line scan in the x-y plane located within the photoreceptors revealed many pathologic sprouts in the anti–KLH mAb-treated Vldlr knockout mice. The optic nerve head is circled. (B) Confocal projection (z plane) illustrates the dramatic reduction in pathologic sprouts in Vldlr knockout mice treated with 1.99.25. Image (x-y plane) taken at the level of the photoreceptors in a retina isolated from a Vldlr knockout mouse that received mAb 1.99.25. A few sprouts were present near the optic nerve head (circle). Scale bar, ∼100 μm. (C) Quantification of vascular sprouts in either anti–KLH control- or 1.99.25-treated animals. Each point represents the average number of neovascular sprouts in both retinas of a single animal. Inhibition of FZD4 significantly blocks the formation of these pathologic sprouts (P < 0.0001).
Figure 4.
 
FZD4-neutralizing mAb induces diffuse vascular leakage in Vldlr knockout mice. Vldlr knockout mice were treated on P12 with either anti–KLH negative control or FZD4 mAb 1.99.25. (A) Focal vascular leakage is present on P20 in mice treated with the anti–KLH antibody. (B) Diffuse vascular leakage is present on P20 in mice treated with mAb 1.99.25. (C) Focal vascular leakage is present at P30 in mice treated with the anti–KLH antibody. (D) Faint, diffuse vascular leakage is present at P30 in mice treated with mAb 1.99.25. The angiogram shown in (B) was obtained approximately 1 minute after fluorescein injection. The remaining angiograms were captured approximately 6 minutes after injection.
Figure 4.
 
FZD4-neutralizing mAb induces diffuse vascular leakage in Vldlr knockout mice. Vldlr knockout mice were treated on P12 with either anti–KLH negative control or FZD4 mAb 1.99.25. (A) Focal vascular leakage is present on P20 in mice treated with the anti–KLH antibody. (B) Diffuse vascular leakage is present on P20 in mice treated with mAb 1.99.25. (C) Focal vascular leakage is present at P30 in mice treated with the anti–KLH antibody. (D) Faint, diffuse vascular leakage is present at P30 in mice treated with mAb 1.99.25. The angiogram shown in (B) was obtained approximately 1 minute after fluorescein injection. The remaining angiograms were captured approximately 6 minutes after injection.
Figure 5.
 
FZD4-neutralizing mAb induces PLVAP expression. PLVAP immunohistochemistry was performed on retinas from P20 (A, B) and P30 (C, D) Vldlr knockout mice treated with either anti–KLH control (A, C) or 1.99.25 (B, D) on P12. (A) Focal, pathologic sprouts are present in anti–KLH-treated P20 mice. These sprouts are adjacent to the retinal pigment epithelium (RPE) and are intensely immunoreactive for PLVAP (brown stain). (B) Pathologic sprouts were sparse in 1.99.25-treated cases. However, much of the vasculature (arrows) in these P20 mice exhibited intense PLVAP immunoreactivity. Compare capillaries in the inner plexiform layer (ipl) and the outer plexiform layer (opl) in A and B. (C) Pathologic sprouts (arrow) remain intensely PLVAP immunoreactive in P30 mice treated with the anti–KLH antibody. PLVAP immunoreactivity is not observed elsewhere in the retinal vasculature. (D) Pathologic sprouts were extremely rare in 1.99.25-treated mice. The capillary beds still exhibited PLVAP immunoreactivity on P30, though less intensely than on P20. Scale bar, ∼50 μm.
Figure 5.
 
FZD4-neutralizing mAb induces PLVAP expression. PLVAP immunohistochemistry was performed on retinas from P20 (A, B) and P30 (C, D) Vldlr knockout mice treated with either anti–KLH control (A, C) or 1.99.25 (B, D) on P12. (A) Focal, pathologic sprouts are present in anti–KLH-treated P20 mice. These sprouts are adjacent to the retinal pigment epithelium (RPE) and are intensely immunoreactive for PLVAP (brown stain). (B) Pathologic sprouts were sparse in 1.99.25-treated cases. However, much of the vasculature (arrows) in these P20 mice exhibited intense PLVAP immunoreactivity. Compare capillaries in the inner plexiform layer (ipl) and the outer plexiform layer (opl) in A and B. (C) Pathologic sprouts (arrow) remain intensely PLVAP immunoreactive in P30 mice treated with the anti–KLH antibody. PLVAP immunoreactivity is not observed elsewhere in the retinal vasculature. (D) Pathologic sprouts were extremely rare in 1.99.25-treated mice. The capillary beds still exhibited PLVAP immunoreactivity on P30, though less intensely than on P20. Scale bar, ∼50 μm.
Figure 6.
 
FZD4-neutralizing mAb induces fluorescein leakage and PLVAP expression in adult mice. Angiograms of adult wild-type C57BL/6J mice obtained after four weekly doses of either anti–KLH (A) or 1.99.25 (B). Treatment with the FZD4 mAb resulted in diffuse fluorescein leakage. Retinal capillaries (C, arrow) lack PLVAP immunoreactivity in anti–KLH-treated mice. Treatment with mAb 1.99.25 results in increased PLVAP expression (brown stain) in retinal capillaries (D, arrows). (E) Immunohistochemistry with anti–mouse IgG intensely stains retinal vasculature in anti–KLH-treated animals. The vitreous (asterisk) is weakly stained. (F) Treatment with 1.99.25 resulted in intense IgG immunoreactivity in the vitreous (asterisk) and outer plexiform layer (opl). IgG-like immunoreactivity also appears at the level of cone outer segments. Scale bar, ∼50 μm (CF).
Figure 6.
 
FZD4-neutralizing mAb induces fluorescein leakage and PLVAP expression in adult mice. Angiograms of adult wild-type C57BL/6J mice obtained after four weekly doses of either anti–KLH (A) or 1.99.25 (B). Treatment with the FZD4 mAb resulted in diffuse fluorescein leakage. Retinal capillaries (C, arrow) lack PLVAP immunoreactivity in anti–KLH-treated mice. Treatment with mAb 1.99.25 results in increased PLVAP expression (brown stain) in retinal capillaries (D, arrows). (E) Immunohistochemistry with anti–mouse IgG intensely stains retinal vasculature in anti–KLH-treated animals. The vitreous (asterisk) is weakly stained. (F) Treatment with 1.99.25 resulted in intense IgG immunoreactivity in the vitreous (asterisk) and outer plexiform layer (opl). IgG-like immunoreactivity also appears at the level of cone outer segments. Scale bar, ∼50 μm (CF).
Text s1, PDF
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×