In early neonates, the fetal circulatory system undergoes dramatic transition to the adult circulatory system. Normally, embryonic connecting vessels, such as the ductus arteriosus and the foramen ovale, close and regress. In the neonatal retina, hyaloid vessels maintaining blood flow in the embryonic retina regress, and retinal vessels take over to form the adult-type circulatory system. This process is regulated by a programmed cell death switch mediated by macrophages via Wnt and angiopoietin 2 pathways. In this study, we seek other mechanisms that regulate this process, and focus on the dramatic change in oxygen environment at the point of birth. The von Hippel-Lindau tumor suppressor protein (pVHL) is a substrate recognition component of an E3-ubiquitin ligase that rapidly destabilizes hypoxia-inducible factor αs (HIF-αs) under normoxic, but not hypoxic, conditions. To examine the role of oxygen-sensing mechanisms in retinal circulatory system transition, we generated retina-specific conditional-knockout mice for VHL (Vhlα-CreKO mice). These mice exhibit arrested transition from the fetal to the adult circulatory system, persistence of hyaloid vessels and poorly formed retinal vessels. These defects are suppressed by intraocular injection of FLT1-Fc protein [a vascular endothelial growth factor (VEGF) receptor-1 (FLT1)/Fc chimeric protein that can bind VEGF and inhibit its activity], or by inactivating the HIF-1α gene. Our results suggest that not only macrophages but also tissue oxygen-sensing mechanisms regulate the transition from the fetal to the adult circulatory system in the retina.

In early neonates, the fetal circulatory system undergoes rapid and dramatic transition to the adult circulatory system to adapt to detachment from the maternal blood supply. Normally, embryonic connecting vessels such as ductus arteriosus, ductus venosus and the foramen ovale close and regress (Merkle and Gilkeson, 2005). In the neonatal retina, it is generally known that the hyaloid vessels that maintain blood flow in the embryonic retina regress, and that retinal vessels assume the role of supplying blood to the retina (Lang, 1997; Lobov et al., 2005). This process is regulated by the programmed cell death switch mediated by macrophages via Wnt and angiopoietin 2 pathways (Lang and Bishop, 1993; Lobov et al., 2005; Rao et al., 2007).

Oxygen concentration is vital to nearly all forms of life on earth via its role in energy homeostasis (Yun et al., 2002). Mammalian embryos develop in a low (~3%) oxygen environment inside the maternal body (Rodesch et al., 1992). However, pups are abruptly exposed to atmospheric oxygen levels (~21%) after birth. This dramatic change in oxygen environment at the point of birth indicates some crucial roles of oxygen concentration in triggering the transition from the fetal to the adult circulatory system.

The von Hippel-Lindau tumor suppressor protein (pVHL) is a substrate recognition component of an E3-ubiquitin ligase that rapidly destabilizes hypoxia-inducible factor αs (HIF-αs) under normoxic, but not hypoxic, conditions (Maxwell et al., 1999). Genetic inactivation of the Vhl gene results in the stabilization and increased activity of HIF-αs, which includes HIF-1α, HIF-2α and HIF-3α. It has been reported that Vhl-null embryos are lethal at embryonic day (E) 11.5-E12.5 owing to abnormal placental vascularization (Gnarra et al., 1997). Endothelial-specific deletion of the Vhl gene using Tie2-Cre mice also results in intrauterine death with hemorrhage at E12.5, and loss of HIF-1α does not rescue the lethality (Tang et al., 2006). Although the vascular endothelial growth factor (VEGF) gene is widely known to be a HIF-1α target (Forsythe et al., 1996), HIF-2α is thought to be the main regulator of the VEGF gene in tissues that express both HIF-1α and HIF-2α (Hu et al., 2003; Tang et al., 2006). These previous studies showed that VHL is required for endothelial cells in a cell autonomous manner. However, surprisingly, it is still largely unknown how VHL contributes to vascular development, particularly how and when the well-known relationship between VHL and oxygen concentration operates in vascular development. In humans, VHL disease is an autosomal dominant syndrome that causes the development of various benign and malignant disorders (Nicholson et al., 1976). The major disorders include retinal, brain and spinal cord angioma, pheochromocytoma, renal cell carcinoma and pancreatic cystadenoma. In the majority of VHL patients, retinal angioma is the first sign of the disease to appear. However, the molecular mechanism of retinal angioma development in VHL disease still remains unknown.

In this study, by using retina-specific conditional-knockout technology, we explore the precise in vivo function of VHL in retinal vascular development, and show that not only macrophages but also tissue oxygen-sensing mechanisms regulate the transition from the fetal to the adult circulatory system in the retina.

Mice

All animal experiments were conducted in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research. In the expression analysis shown in Fig. 1, we used C57BL/6 mice (Clea Japan). Transgenic mice expressing Cre recombinase under control of the Pax6 retina-specific regulatory element α-promoter (α-Cre) (Marquardt et al., 2001) were mated with Vhlfloxed/floxed mice (Haase et al., 2001) (kind gift from Volker H. Haase, University of Pennsylvania), HIF-1αfloxed/floxed mice (Ryan et al., 2000), or HIF-2αfloxed/floxed mice (Gruber et al., 2007). As control littermates for Vhlα-CreKO mice, HIF-1αα-CreKO mice, HIF-1αα-CreKO;Vhlα-CreKO mice and HIF-2αα-CreKO;Vhlα-CreKO mice, we used Vhlfloxed/floxed mice, HIF-1αfloxed/floxed mice, HIF-1αfloxed/floxed;Vhlfloxed/floxed mice and HIF-2αfloxed/floxed;Vhlfloxed/floxed mice without the α-Cre transgene, respectively. In a preliminary study, we found no detectable difference in retinal vascular structures between α-Cre+ and α-Cre− mice, confirming the validity of these control littermates. To monitor Cre expression in α-Cre mice or Vhlα-CreKO mice, we mated these mice with CAG-CAT-EGFP transgenic mice (Kawamoto et al., 2000). All mice used in this study were maintained on a C57BL/6J background.

Preparation of whole-mount samples and cryosections of retinas

Enucleated eyes were fixed for 20 minutes in 4% paraformaldehyde (PFA) in PBS and then dissected as previously described (Fruttiger et al., 1996; Gerhardt et al., 2003; Kubota et al., 2008). The obtained tissues were post-fixed overnight in 4% PFA in PBS and stored in methanol at −20°C. Cryosections of retinas (10 μm) were prepared as previously described (Kurihara et al., 2006), after eyeballs were immersed overnight in 4% PFA.

Immunostaining and in situ hybridization

Immunohistochemistry (IHC) for whole-mount retinas and other tissues was performed as described previously (Fruttiger et al., 1996; Gerhardt et al., 2003; Kubota et al., 2008; Kurihara et al., 2006). The primary antibodies used were: monoclonal anti-PECAM1 (2H8, Chemicon), α-smooth muscle actin (1A4; FITC-conjugated; Sigma-Aldrich), desmin (DAKO), F4/80 (A3-1; Serotec, Oxford, UK), GFAP (G-A-5; Cy3-conjugated; Sigma-Aldrich or Dako), polyclonal type IV collagen (Cosmo Bio), HIF-1α (originally established by immunizing purified fusion proteins encompassing amino acids 416 to 785 of mouse HIF-1α into guinea pigs), HIF-2α (Santa Cruz Biotechnology), glutamine synthetase (Molecular Probes), opsin (Cosmo Bio) and pVHL (Santa Cruz Biotechnology). The secondary antibodies used were Alexa Fluor 488-conjugated IgGs (Molecular Probes) or Cy3/Cy5-conjugated IgGs (Jackson ImmunoResearch). Nuclei were stained with 10 μg/ml Hoechst bisbenzimide 33258 (Sigma-Aldrich) or DAPI (Molecular Probes). For whole-mount in situ hybridization (ISH), retinas were briefly digested with proteinase K and hybridized with digoxigenin (DIG)-labeled antisense RNA probes. When ISH was combined with IHC, IHC was performed after all the ISH procedures were completed. For the BrdU incorporation assay, 100 μg per body weight (g) of BrdU (BD Pharmingen) dissolved in sterile PBS was injected intraperitoneally 2 hours before sacrifice. Isolated retinas were stained using a BrdU immunohistochemistry system (Calbiochem) according to the manufacturer's instructions. When BrdU assays were combined with IHC, the application of the first antibody for each procedure was done simultaneously. FITC-conjugated Dextran (Sigma-Aldrich) was injected into the left cardiac ventricle and allowed to circulate for 2 minutes. After staining, samples were mounted using a Prolong Antifade Kit (Molecular Probes).

Assessment of tissue hypoxia

Detection of hypoxic cells in cryosections was performed using a Hypoxyprobe-1 Plus Kit (Chemicon). In brief, 60 mg/kg of pimonidazole was injected intraperitoneally 30 minutes before sacrifice, and samples were stained with Hypoxyprobe Mab1-FITC.

RT-PCR analysis

Total RNA was prepared from retinal tissues and reverse-transcribed using Superscript II (Invitrogen). A quantitative PCR assay for Vegfa was performed on an ABI 7500 Fast Real-Time PCR System using TaqMan Fast Universal PCR master mixture (Applied Biosystems) and TaqMan Gene Expression Assay Mix of Vegfa (Mm00437304_m1), Csf1r (Mm00432689_m1), and angiopoietin 2 (Mm00545822_m1). Mouse β-actin (Mm00607939_s1) assay mix served as an endogenous control. Data were analyzed with 7500 Fast System SDS Software 1.3.1. All experiments were done with four replicates.

Intravitreous injections

Injections into the vitreous body were performed using 33-gauge needles, as described previously (Gerhardt et al., 2003; Kubota et al., 2008). Sterile PBS (0.5 μl) with or without 1 mg/ml Flt1-Fc chimera proteins (R&D Systems) was injected at postnatal day (P) 4.

Confocal microscopy and quantification

Fluorescence images were obtained using a confocal laser scanning microscope (FV1000; Olympus) at room temperature. Quantification of the cells or substances of interest was usually done in eight random 200 μm × 200 μm fields just behind the sprouting edges of each retina. To construct three-dimensional projections, multiple slices horizontally imaged from the same field of view were integrated by FV10-ASW Viewer (Olympus).

Retinal explant culture

Retinal explant culture was performed using P6 mouse neural retinas based on the protocol we previously described (Ozawa et al., 2007; Ozawa et al., 2004). Briefly, eyes were enucleated and neural retinas were isolated and placed on a Millicell chamber filter (Millipore; pore size, 0.4 Am) with the ganglion cell layer facing up. The chamber was then placed in a 6-well culture plate, containing 50% MEM (GIBCO), 25% HBSS (GIBCO), 25% horse serum (Thermo Trace), 200 mM l-glutamine, and 6.75 mg/ml d-glucose. Explants were incubated at 34°C in 5% CO2. Twenty-four hours after exposure of each concentration of oxygen, the explants were subjected to immunostaining or RT-PCR analysis. The oxygen concentration in the incubator was controlled with nitrogen.

Statistical analysis

Comparison between the average variables of two groups was performed by two-tailed Student's t-test. P-values less than 0.05 were considered to be statistically significant.

HIF-1α expression is rapidly downregulated after birth

As a first step in the investigation of the cellular responses in the retina to the dramatic change in the oxygen environment at birth, we examined the expression patterns of pVHL, HIF-1α and HIF-2α in the developing retina. Expression of pVHL was detected ubiquitously from the inner to the outer side in the retina, and did not change between before and after birth (Fig. 1A-E). By contrast, nuclear staining of HIF-1α was downregulated in the deep retinal layer [the neuroblastic layer (NBL) at this stage] after birth (Fig. 1H-J), despite its ubiquitous expression in embryonic stages (Fig. 1F,G). HIF-2α immunoreactivity was predominantly detected in endothelial cells of hyaloid vessels in the vitreal space (see Fig. S1A-C in the supplementary material) and in retinal vessels in the ganglion cell layer (GCL) (see Fig. S1D,E in the supplementary material). In embryonic retina, hypoxic areas were visualized by a hypoxic probe, pimonidazole, through all layers (Fig. 1K,L), although a relatively hypoxic area after birth was limited to the GCL (Fig. 1M-O), which still expressed HIF-1α (Fig. 1H-J).

Fig. 1.

Expression patterns of pVHL and HIF-1α in the retina before and after birth. (A-E)pVHL staining (green), detected in the soma of cells in the whole retina, is unchanged after birth. (F-J) Nuclear staining of HIF-1α (red), detected ubiquitously in the embryo (F,G), is downregulated in the deep retinal layer (NBL at this stage) after birth (H-J). (K-O)In embryonic retina, hypoxic areas visualized by the hypoxic probe pimonidazole (yellow) spread all through the layers (K,L), although after birth the hypoxic signal was considerably downregulated in the retina, noticeably in the deep retinal layer (M-O) where HIF-1α expression has also declined (H-J). GCL, ganglion cell layer; NBL, neuroblastic layer. Scale bar: 100 μm in A-O.

Fig. 1.

Expression patterns of pVHL and HIF-1α in the retina before and after birth. (A-E)pVHL staining (green), detected in the soma of cells in the whole retina, is unchanged after birth. (F-J) Nuclear staining of HIF-1α (red), detected ubiquitously in the embryo (F,G), is downregulated in the deep retinal layer (NBL at this stage) after birth (H-J). (K-O)In embryonic retina, hypoxic areas visualized by the hypoxic probe pimonidazole (yellow) spread all through the layers (K,L), although after birth the hypoxic signal was considerably downregulated in the retina, noticeably in the deep retinal layer (M-O) where HIF-1α expression has also declined (H-J). GCL, ganglion cell layer; NBL, neuroblastic layer. Scale bar: 100 μm in A-O.

Vhlα-CreKO mice show persistence of the hyaloid vessels independently of macrophages

To elucidate the function of pVHL in postnatal retinal vascularization, we employed a retina-specific mouse Cre line, α-Cre, which shows Cre expression in the deep retinal layer, including in retinal progenitor cell-derived neural cells, under the control of a neural retina-specific regulatory element of murine Pax6 gene, but not in astrocytes and endothelial cells at P10 (Marquardt et al., 2001). We examined the past events of Cre expression in α-Cre mice until early postnatal age by crossing α-Cre mice with a reporter transgenic line, CAG-CAT-EGFP (Kawamoto et al., 2000). At P6, GFP expression was detected in neural cells in the deep retinal layer, but not in astrocytes or endothelial cells of both retinal and hyaloid vasculature (see Fig. S2A-T in the supplementary material). Then, we crossed Vhlfloxed/floxed mice (Haase et al., 2001) with α-Cre mice and generated α-Cre-specific conditional-knockout mice for Vhl (Vhlα-CreKO mice). As control littermates for Vhlα-CreKO mice, we used Vhlfloxed/floxed mice without the α-Cre transgene. At P6, blood flow visualized by intracardiac injection of fluorescein isothiocyanate (FITC)-dextran was detected in both hyaloid and retinal vessels of control mice (Fig. 2A,B,E,G,I). In Vhlα-CreKO mice, however, there was abundant blood flow in hyaloid vessels, poor flow in retinal vessels, and abundant collateral flow from hyaloid vessels to retinal vessels (Fig. 2C,D,F,H,J). Nonetheless, despite the impaired blood flow, the retinal vascular structure could be detected by immunohistochemistry for platelet/endothelial cell adhesion molecule 1 (PECAM1) in Vhlα-CreKO mice (Fig. 2F,H). The persistent blood flow in hyaloid vessels in Vhlα-CreKO mice prompted us to examine the well-known mechanisms of hyaloid vessel regression: reduced perivascular macrophages or angiopoietin 2 deficiency lead to persistence of the hyaloid vessels (Lobov et al., 2005; Rao et al., 2007). Macrophages around the hyaloid vessels were not significantly changed (Fig. 2K-M), and colony stimulating factor 1 receptor (Csf1r) expression, which correlates with the number of macrophages (Kubota et al., 2009), showed no significant reduction in Vhlα-CreKO mice (Fig. 2N). Angiopoietin 2 expression was rather increased in Vhlα-CreKO mice (Fig. 2O). Collectively, Vhlα-CreKO mice show persistence of the hyaloid vessels independently of macrophages and angiopoietin 2.

Vhlα-CreKO mice show poorly formed retinal vessels characterized by excessive vessel regression

As abnormal hyaloid vessel persistence could affect retinal vascularization, we examined the retinal vessels in Vhlα-CreKO mice. In these mice, the entire growth of the retinal vasculature spreading from the optic disc and the branching points were significantly reduced compared with control mice (P=0.013 for spreading distance, P=0.00012 for branching points; Fig. 3A,B,D,E,M,N; see also Fig. S3A,B in the supplementary material). Endothelial tip cells and their filopodia, which are controlled by VEGF and the Delta-like 4/Notch pathway (Gerhardt et al., 2003; Hellstrom et al., 2007; Phng et al., 2009), were also significantly reduced in Vhlα-CreKO mice (P=0.0012 for tip cells; P=0.011 for filopodia; Fig. 3C,F,O,P). A significant decrease in endothelial proliferation (Fig. 3G,J,Q) and excessive vessel regression (Fig. 3H,K,R), characterized by empty vascular sleeves (Phng et al., 2009), were also detected in Vhlα-CreKO mice (P=7.9×10−5 for BrdU+ endothelial cells; P=0.0044 for empty sleeves). Empty sleeves were abundantly detected not only in their stalk cell area but also at their leading edge, suggesting that vessel regression occurs in both tip cells and stalk cells in Vhlα-CreKO mice. Although it is well known that pVHL regulates the expression of various extracellular matrices (Ohh et al., 1998), Col IV expression was not impaired in Vhlα-CreKO mice. Although the number of pericytes associated with endothelial tubes was not changed in Vhlα-CreKO mice (P=0.395; Fig. 3I,L,S), detached pericytes were frequently detected just beyond their sprouting edges. Previously, injection of angiopoietin 2 into the eyes of normal rats was shown to induce a dose-dependent pericyte loss (Hammes et al., 2004), suggesting that upregulated angiopoietin 2 (Fig. 2O) might contribute to pericyte defects in Vhlα-CreKO mice. Collectively, Vhlα-CreKO mice showed poorly formed retinal vessels, presumably owing to persistence of the hyaloid vessels.

Fig. 2.

Vhlα-CreKO mice show persistence of the hyaloid vessels independently of macrophages. (A-D) Representative images of retinas after intracardiac FITC injection (n=8). Vhlα-CreKO mice show collateral flow from hyaloid vessels to retinal vessels (open arrowheads) in the P6 retinas. Blood flow was visualized by intracardiac FITC-conjugated dextran (green) injection. (E-H) Confocal images of FITC- (orange) PECAM1- (blue) labeled P6 retinas; G,H show sections. Poor blood flow is detected in retinal vasculature (asterisk) in Vhlα-CreKO mice compared with control. Relatively abundant dextran was perfused into arteries (A) and veins (V) in control mice, and into hyaloid vessels (arrows) in Vhlα-CreKO mice. (I,J) Schematics show abundant blood flow (red) in retinal vessels in control mice, and in hyaloid vessels in Vhlα-CreKO mice. (K,L) Immunostaining with F4/80 (green) and PECAM1 (blue) on hyaloid vessels in P6 retinas. (M) Quantification of F4/80+ macrophages per 200 μm vessel length. Vessel length was calculated by FV10-ASW Viewer software (n=4). (N,O) Quantitative PCR of Csf1r (N) or angiopoietin 2 (O) for isolated RNA from P6 retinas (n=5). *P<0.05. All panels show P6 retina. Error bars indicate mean ± s.d. Scale bars: 500 μm in A-D; 100 μm in E-H,K,L.

Fig. 2.

Vhlα-CreKO mice show persistence of the hyaloid vessels independently of macrophages. (A-D) Representative images of retinas after intracardiac FITC injection (n=8). Vhlα-CreKO mice show collateral flow from hyaloid vessels to retinal vessels (open arrowheads) in the P6 retinas. Blood flow was visualized by intracardiac FITC-conjugated dextran (green) injection. (E-H) Confocal images of FITC- (orange) PECAM1- (blue) labeled P6 retinas; G,H show sections. Poor blood flow is detected in retinal vasculature (asterisk) in Vhlα-CreKO mice compared with control. Relatively abundant dextran was perfused into arteries (A) and veins (V) in control mice, and into hyaloid vessels (arrows) in Vhlα-CreKO mice. (I,J) Schematics show abundant blood flow (red) in retinal vessels in control mice, and in hyaloid vessels in Vhlα-CreKO mice. (K,L) Immunostaining with F4/80 (green) and PECAM1 (blue) on hyaloid vessels in P6 retinas. (M) Quantification of F4/80+ macrophages per 200 μm vessel length. Vessel length was calculated by FV10-ASW Viewer software (n=4). (N,O) Quantitative PCR of Csf1r (N) or angiopoietin 2 (O) for isolated RNA from P6 retinas (n=5). *P<0.05. All panels show P6 retina. Error bars indicate mean ± s.d. Scale bars: 500 μm in A-D; 100 μm in E-H,K,L.

Vascular defects in Vhlα-CreKO mice are attributable to ectopic VEGF expression

The intact macrophages and preserved angiopoietin 2 expression in Vhlα-CreKO mice caused us to suppose that some other mechanism led to the persistence of the hyaloid vessels. As retinal vascular growth depends on the appropriate spatial distribution of heparin-binding VEGF within the retina (Gerhardt et al., 2003; Kubota et al., 2008; Ruhrberg et al., 2002; Stalmans et al., 2002), we suspected that the vascular defects in Vhlα-CreKO mice were attributable to the abnormal expression pattern of VEGF. In control mice, VEGF expression was predominantly detected in astrocytes beyond the sprouting edge (Fig. 4A,B,E-H). In Vhlα-CreKO mice, marked expression of VEGF was detected in the deep retinal layer (Fig. 4C,D,I-L), and was coincidental with the area of persistent hyaloid vessels. However, typical VEGF expression in astrocytes beyond the sprouting edge was diminished in Vhlα-CreKO mice, despite the fact that their astrocyte plexus was normal (Fig. 4M,N). This VEGF expression pattern in Vhlα-CreKO mice was similar to normal VEGF expression in the embryonic retina (Saint-Geniez et al., 2006). By quantitative RT-PCR analysis, a more than 4-fold significant increase (P=0.00071) in Vegfa expression was detected in the entire retina of Vhlα-CreKO mice (Fig. 4O). Because these expression patterns of VEGF suggested that the vascular defects in Vhlα-CreKO mice could be attributable to ectopic and abundant VEGF expression, we injected a potent VEGF inhibitor, FLT1-Fc chimeric protein (Gerhardt et al., 2003; Kubota et al., 2008), into the eyes of Vhlα-CreKO mice. These FLT1-Fc injections reduced the amount of both collateral blood flow and vascular structures from hyaloid vessels to retinal vessels (Fig. 4P-U). As genetic inactivation of the Vhl gene results in HIF-α stabilization (Kapitsinou and Haase, 2008) and the upregulation of Vegfa expression, which is widely known as a HIF-1α target gene (Forsythe et al., 1996), we examined the expression of pVHL, HIF-1α and HIF-2α in Vhlα-CreKO mice. We found that pVHL immunoreactivity was greatly reduced in the deep layer of the peripheral retina (Fig. 4V,W), where the strong α-Cre-mediated recombination is detected (see Fig. S2 in the supplementary material), except for in astrocytes and endothelial cells (see Fig. S4A-J in the supplementary material). HIF-1α immunoreactivity was increased in this area of Vhlα-CreKO mice (Fig. 4X,Y; see also Fig. S4K-O in the supplementary material). In addition, strong HIF-2α immunoreactivity was detected in invaginating hyaloid vessels, but not in the area of reduced pVHL expression (Fig. 4Z,AA; see also Fig. S4P-T in the supplementary material).

Increased VEGF expression in Vhlα-CreKO mice is due to their impaired oxygen-sensing mechanism via HIF-1α

As the VEGF expression pattern in postnatal Vhlα-CreKO retina is similar to normal VEGF expression in the embryonic retina (Saint-Geniez et al., 2006), we expected abnormal VEGF expression in the Vhlα-CreKO retina to be attributable to an impaired oxygen-sensing mechanism and a lack of response to the atmospheric oxygen concentration. As newborn mice do not tolerate oxygen levels below 10% (Claxton and Fruttiger, 2005), we employed an ex vivo retinal culture system, which enabled us to expose the retina to variable oxygen concentrations between 1% and 21% (Fig. 5A). In the control retina culture, strong nuclear HIF-1α expression was induced in 1% oxygen (Fig. 5D), but not in 21% oxygen (Fig. 5B). Consequently, Vegfa expression was significantly higher (P=2.4×10−8) in 1% oxygen than in 21% oxygen (Fig. 5J). In the Vhlα-CreKO retina culture, strong nuclear HIF-1α expression was detected (Fig. 5H), even in 21% oxygen (Fig. 5F), and Vegfa expression in the Vhlα-CreKO retina in 21% oxygen was significantly higher (P=1.1×10−5) than in the control in 21% oxygen; it was equivalent to the level observed in 1% oxygen (Fig. 5J). To determine the significance of the increased nuclear HIF-1α level in hypoxic conditions in this culture system, we generated α-Cre-mediated conditional-knockout mice for HIF-1α (Ryan et al., 2000) (HIF-1αα-CreKO mice). In HIF-1αα-CreKO retina, hypoxia-induced HIF-1α stabilization (Fig. 5K,M) was not detected (Fig. 5O,Q) and, accordingly, increased Vegfa expression in 1% oxygen was significantly suppressed (P=0.00028) compared with in control retina (Fig. 5S). Consistent with the in vivo results (Fig. 1A-E), the expression pattern of pVHL was stable in the deep retinal layer during ex vivo culture, and was not affected by the oxygen concentration (Fig. 5C,E,G,I,L,N,P,R). These data in the ex vivo culture system suggest that the increased VEGF expression in Vhlα-CreKO mice is due to their impaired oxygen-sensing mechanism via HIF-1α.

Fig. 3.

Vhlα-CreKO mice show poorly formed retinal vessels characterized by excessive vessel regression. (A-F) PECAM1 staining of P6 retinas. Note delayed vascular growth (bidirectional arrows in A,D), decreased branching points and tip cells (B,E), and filopodia (C,F) in Vhlα-CreKO mice. Persistent hyaloid vessels in the lower left lobe of Vhlα−CreKO mice were removed mechanically before immunostaining (D). (G-L) Immunostaining with the indicated antibodies. Note the decreased BrdU+ endothelial cells (arrows, G,J), type IV collagen (Col IV)+isolectin empty sleeves (arrowheads, H,K), and detached pericytes beyond sprouting edges (open arrowheads, L) in Vhlα-CreKO mice, although most pericytes in control mice are closely contacted with vasculature (closed arrowheads, I,L). (M-S) Quantification of percentage spreading distance from optic discs (M), the number of branching points in the area behind sprouting edges (N), tip cells per field (O), filopodia per tip cell (P), BrdU+ endothelial cells (Q), Col IV+isolectin vessels (R), and pericytes per 200 μm vessel length (S; n=6, in each genotype). *P<0.05, **P<0.01. All panels show P6 retina. Error bars indicate mean±s.d. Scale bars: 500 μm in A,D; 100 μm in B,E,G-L; 20 μm in C,F.

Fig. 3.

Vhlα-CreKO mice show poorly formed retinal vessels characterized by excessive vessel regression. (A-F) PECAM1 staining of P6 retinas. Note delayed vascular growth (bidirectional arrows in A,D), decreased branching points and tip cells (B,E), and filopodia (C,F) in Vhlα-CreKO mice. Persistent hyaloid vessels in the lower left lobe of Vhlα−CreKO mice were removed mechanically before immunostaining (D). (G-L) Immunostaining with the indicated antibodies. Note the decreased BrdU+ endothelial cells (arrows, G,J), type IV collagen (Col IV)+isolectin empty sleeves (arrowheads, H,K), and detached pericytes beyond sprouting edges (open arrowheads, L) in Vhlα-CreKO mice, although most pericytes in control mice are closely contacted with vasculature (closed arrowheads, I,L). (M-S) Quantification of percentage spreading distance from optic discs (M), the number of branching points in the area behind sprouting edges (N), tip cells per field (O), filopodia per tip cell (P), BrdU+ endothelial cells (Q), Col IV+isolectin vessels (R), and pericytes per 200 μm vessel length (S; n=6, in each genotype). *P<0.05, **P<0.01. All panels show P6 retina. Error bars indicate mean±s.d. Scale bars: 500 μm in A,D; 100 μm in B,E,G-L; 20 μm in C,F.

Deletion of HIF-1α, but not HIF-2α, rescues vascular defects in Vhlα-CreKO mice

The expression patterns of pVHL, HIF-1α, HIF-2α and VEGF in Vhlα-Cre KO mice (Fig. 4) and the data from the ex vivo culture system (Fig. 5) suggested that the increased activity of HIF-1α was responsible for the vascular defects in Vhlα-CreKO mice. Therefore, we generated α-Cre-specific double-knockout mice for VHL and HIF-1α (Ryan et al., 2000) (HIF-1α;Vhlα-CreKO mice). Vascular defects seen in Vhlα-CreKO mice were suppressed in HIF-1α; Vhlα-CreKO mice (Fig. 6A,B,D-G,M-P). The abnormal expression of Vegfa in Vhlα-CreKO was fairly normalized in HIF-1α; Vhlα-CreKO mice (Fig. 6J,K; see also Fig. S5A in the supplementary material; 1.15±0.08-fold compared with control). Conversely, α-Cre-mediated VHL and HIF-2α (Gruber et al., 2007) double-knockout mice (HIF-2α; Vhlα-CreKO mice) showed similar vascular defects and a similar VEGF expression pattern to Vhlα-CreKO mice (Fig. 6C,H,I,L; see also Fig. S5B in the supplementary material). Taken together, HIF-1α, but not HIF-2α, plays essential roles in the appearance of vascular defects in the Vhlα-CreKO retina.

Progressive retinal degeneration in adult Vhlα-CreKO mice

Next, we examined retinal function and circulation in adult Vhlα-CreKO mice. Increased apoptosis of photoreceptors and other neuronal cells was seen in Vhlα-CreKO mice after P14 (Fig. 7A-H) induced decrease of the outer nuclear layer (ONL; Fig. 7I,J), despite the normal development of photoreceptors at P7 (Fig. 7E,J). Gliosis, accompanied by vessels invaginating into the deep retinal layer, destroyed the construction of the ONL (Fig. 7K-T) at P14. As a result, a significant (P=3.3×10−6 for a-wave, P=1.3×10−5 for b-wave) decrease in amplitude and a significant (P=0.046 for b-wave) extension of implicit time in electroretinography (ERG) was observed in Vhlα-CreKO mice at P28 (Fig. 7U-Y). Circulation via hyaloid vessels in the Vhlα-CreKO retina was detected at the age of 8 weeks (see Fig. S6A-D in the supplementary material) and at 18 months (see Fig. S6E-H in the supplementary material). Some of the Vhlα-CreKO mice developed pre-retinal hemorrhage, cataracts and iris neovascularization (Fig. 8A-C), similar to the phenotypes that accompany ischemic retinopathy (Hayreh, 2007).

In the present study, we showed that retina-specific conditional-knockout mice for the Vhl gene exhibit persistent hyaloid vessels independently of macrophage function, which are sustained until adulthood. These vascular defects in Vhlα-CreKO mice are rescued by either local VEGF inhibition or genetic deletion of HIF-1α, but not of HIF-2α. These results suggest not only macrophages but also tissue oxygen-sensing mechanisms regulate the transition from the fetal to the adult circulatory system in retina.

Fig. 4.

Vascular defects in Vhlα-CreKO mice are attributable to ectopic VEGF expression. (A-L) Whole-mount (A-D) or section (E-L) in situ hybridization for VEGF combined with immunostaining with the indicated antibodies. Although VEGF expression is detected in astrocytes located in the avascular area (arrows) of control mice, abundant VEGF expression is detected in the deep retinal layer (open arrowheads) where persistent hyaloid vessels invaginate in Vhlα-CreKO mice. (M,N) Immunostaining with PECAM1 (green) and PDGFRα (red) on P6 retinas. Despite a normal astrocyte plexus (asterisks), vessel regression (open arrowheads) occurs in Vhlα-CreKO mice. (O) Quantitative PCR of Vegfa RNA isolated from P6 retinas (n=6). (P-U) Fluorescence microscopic images in P6 retinas perfused with FITC-dextran (P-R), and confocal images labeled with PECAM1 (S-U). FLT1-Fc injection into the eyes of Vhlα-CreKO mice reduces collateral flow (arrows, R) and vascular structures (arrows, U) that exist abundantly in Vhlα-CreKO mice injected with vehicle (open arrowheads in Q,T). (V-AA) Immunostaining of P6 retinal sections with the indicated antibodies. Although pVHL expression is greatly reduced (open arrowheads, W), HIF-1α immunoreactivity is increased (closed arrowheads, Y) in the deep retinal layer of Vhlα−CreKO mice. HIF-2α staining is detected in invaginating hyaloid vessels in Vhlα-CreKO mice (arrows, AA). **P<0.01. Error bars indicate mean ± s.d. Scale bars: 500 μm in A-L; 200 μm in P-AA; 100 μm in M,N.

Fig. 4.

Vascular defects in Vhlα-CreKO mice are attributable to ectopic VEGF expression. (A-L) Whole-mount (A-D) or section (E-L) in situ hybridization for VEGF combined with immunostaining with the indicated antibodies. Although VEGF expression is detected in astrocytes located in the avascular area (arrows) of control mice, abundant VEGF expression is detected in the deep retinal layer (open arrowheads) where persistent hyaloid vessels invaginate in Vhlα-CreKO mice. (M,N) Immunostaining with PECAM1 (green) and PDGFRα (red) on P6 retinas. Despite a normal astrocyte plexus (asterisks), vessel regression (open arrowheads) occurs in Vhlα-CreKO mice. (O) Quantitative PCR of Vegfa RNA isolated from P6 retinas (n=6). (P-U) Fluorescence microscopic images in P6 retinas perfused with FITC-dextran (P-R), and confocal images labeled with PECAM1 (S-U). FLT1-Fc injection into the eyes of Vhlα-CreKO mice reduces collateral flow (arrows, R) and vascular structures (arrows, U) that exist abundantly in Vhlα-CreKO mice injected with vehicle (open arrowheads in Q,T). (V-AA) Immunostaining of P6 retinal sections with the indicated antibodies. Although pVHL expression is greatly reduced (open arrowheads, W), HIF-1α immunoreactivity is increased (closed arrowheads, Y) in the deep retinal layer of Vhlα−CreKO mice. HIF-2α staining is detected in invaginating hyaloid vessels in Vhlα-CreKO mice (arrows, AA). **P<0.01. Error bars indicate mean ± s.d. Scale bars: 500 μm in A-L; 200 μm in P-AA; 100 μm in M,N.

Considering that HIF-1α in the outer side of the retina is dramatically downregulated after birth but that pVHL expression in the retina does not differ between embryos and neonates (Fig. 1), the change of environmental oxygen concentration must be important for the role of pVHL in this transition (Rodesch et al., 1992). VEGF expression in the deep retinal layer during embryonic but not postnatal retinal stages (Saint-Geniez et al., 2006) supports the idea that the abnormal VEGF distribution in postnatal Vhlα-CreKO retina represents a defective VEGF expression switch from the embryonic pattern to the postnatal one. Mice selectively expressing single isoforms of VEGF (VEGF120 or VEGF188) or overexpressing VEGF under the control of a lens-specific promoter show a persistence of hyaloid vessels as well as defective retinal vascularization, supporting the concept that spatial distribution of VEGF regulates the transition from the embryonic to the adult circulatory system in the retina (Mitchell et al., 2006; Stalmans et al., 2002).

Fig. 5.

Increased VEGF expression in Vhlα-CreKO mice is due to their impaired oxygen-sensing mechanism via HIF-1α. (A) Schematics show the experimental procedure: isolated retinas from control or conditional-knockout mice were unfolded, placed on a chamber filter, and exposed to each concentration of oxygen. Retinal explants of control mice (B-E,K-N), Vhlα-CreKO mice (F-I), or HIF-1αα-CreKO mice (O-R) were exposed to 21% oxygen (B,C,F,G,K,L,O,P) or 1% oxygen (D,E,H,I,M,N,Q,R). (J,S) Quantitative PCR of Vegfa RNA isolated from retinal explants (n=6, respectively). Note, strong HIF-1α staining (red) is induced by 1% oxygen (D,H,M) except in HIF-1αα-CreKO retina (Q), and by genetic inactivation of VHL (F,H). pVHL expression was not detected in Vhlα-CreKO retina (G,I). (J) Correlated with the expression of HIF-1α, Vegfa expression was upregulated even under normoxia conditions in Vhlα-CreKO retina. (S) In HIF-1αα-CreKO retina, hypoxia-induced Vegfa expression was significantly suppressed. **P<0.01. Error bars indicate mean±s.d. Scale bars: 50 μm in B-I,K-R.

Fig. 5.

Increased VEGF expression in Vhlα-CreKO mice is due to their impaired oxygen-sensing mechanism via HIF-1α. (A) Schematics show the experimental procedure: isolated retinas from control or conditional-knockout mice were unfolded, placed on a chamber filter, and exposed to each concentration of oxygen. Retinal explants of control mice (B-E,K-N), Vhlα-CreKO mice (F-I), or HIF-1αα-CreKO mice (O-R) were exposed to 21% oxygen (B,C,F,G,K,L,O,P) or 1% oxygen (D,E,H,I,M,N,Q,R). (J,S) Quantitative PCR of Vegfa RNA isolated from retinal explants (n=6, respectively). Note, strong HIF-1α staining (red) is induced by 1% oxygen (D,H,M) except in HIF-1αα-CreKO retina (Q), and by genetic inactivation of VHL (F,H). pVHL expression was not detected in Vhlα-CreKO retina (G,I). (J) Correlated with the expression of HIF-1α, Vegfa expression was upregulated even under normoxia conditions in Vhlα-CreKO retina. (S) In HIF-1αα-CreKO retina, hypoxia-induced Vegfa expression was significantly suppressed. **P<0.01. Error bars indicate mean±s.d. Scale bars: 50 μm in B-I,K-R.

Fig. 6.

Deletion of HIF-1α but not HIF-2α, rescues vascular defects in Vhlα-CreKO mice. (A-I) Blood flow visualized by intracardiac FITC-dextran injection (A-C), or confocal images from mice of indicated genotypes labeled with PECAM1 at P6 (D-I). Vascular defects [collateral flow from hyaloid vessels to retinal vessels (open arrowheads), delayed vascular growth (bidirectional arrows), decreased branching points, tip cells and filopodia] seen in Vhlα-CreKO mice are abolished by deletion of HIF-1α, but not HIF-2α, in Vhlα-CreKO mice. (J-L) Whole-mount in situ hybridization for VEGF (blue) combined with immunostaining with Col IV (green). Ectopic VEGF expression seen in Vhlα-CreKO mice is abolished by deletion of HIF-1α, but not HIF-2α, in Vhlα-CreKO mice. (M-T) Quantification of percentage spreading distance from optic discs (M,Q), the number of branching points in the area behind sprouting edges (N,R), tip cells per field (O,S), filopodia per tip cell (P,T; n=6 for each genotype). *P<0.05, **P<0.01. Error bars indicate mean ± s.d. Scale bars: 500 μm in A-C,D,F,H; 100 μm in E,G,I,J-L.

Fig. 6.

Deletion of HIF-1α but not HIF-2α, rescues vascular defects in Vhlα-CreKO mice. (A-I) Blood flow visualized by intracardiac FITC-dextran injection (A-C), or confocal images from mice of indicated genotypes labeled with PECAM1 at P6 (D-I). Vascular defects [collateral flow from hyaloid vessels to retinal vessels (open arrowheads), delayed vascular growth (bidirectional arrows), decreased branching points, tip cells and filopodia] seen in Vhlα-CreKO mice are abolished by deletion of HIF-1α, but not HIF-2α, in Vhlα-CreKO mice. (J-L) Whole-mount in situ hybridization for VEGF (blue) combined with immunostaining with Col IV (green). Ectopic VEGF expression seen in Vhlα-CreKO mice is abolished by deletion of HIF-1α, but not HIF-2α, in Vhlα-CreKO mice. (M-T) Quantification of percentage spreading distance from optic discs (M,Q), the number of branching points in the area behind sprouting edges (N,R), tip cells per field (O,S), filopodia per tip cell (P,T; n=6 for each genotype). *P<0.05, **P<0.01. Error bars indicate mean ± s.d. Scale bars: 500 μm in A-C,D,F,H; 100 μm in E,G,I,J-L.

Previously, Lang and colleagues clearly demonstrated that macrophages mediate hyaloid vessel regression through the paracrine action of WNT7B (Lobov et al., 2005). They showed that a lack of macrophages in PU.1−/− mice caused a significant delay in hyaloid vessel regression, and that intraocularly injected wild-type, but not Wnt7b-mutant, macrophages ameliorated this delay. Moreover, angiopoietin 2 has also been shown to be involved in hyaloid vessel regression via the dual effect of suppressing survival signaling in the endothelial cells of hyaloid vessels and stimulating Wnt ligand production by macrophages (Rao et al., 2007). Interestingly, the macrophage/microglia number in Vhlα-CreKO mice was not affected (Fig. 2K-M), but angiopoietin 2 expression was increased (Fig. 2O), which suggests that an oxygen-sensing mechanism mediated by the VHL/HIF-1α/VEGF system operates hyaloid vessel regression independently of macrophages, Wnt and angiopoietin 2.

Fig. 7.

Vhlα-CreKO mice develop retinal degeneration in adults. (A-H) TUNEL-positive cells (red) counterstained with Hoechst (blue) are increased in the ONL in Vhlα-CreKO mice (E-H) compared with in control mice (A-D) after P14. (I,J) ONL thickness is decreased in a time-dependent manner in Vhlα-CreKO mice (n=3, respectively). (K-T) Merged images for glial fibrillary acidic protein (GFAP), Hoechst and BrdU (K,P), and IHC for PECAM1 (yellow, L,Q), glutamine synthetase (GS, green, M,R), opsin (red, N,S), and merged images for GS and opsin (O,T) in control (K-O) and Vhlα-CreKO mice (P-T). Proliferating reactive glia are detected by BrdU (arrows, P), and gliosis destroys the construction of the ONL (open arrowheads, R-T) in Vhlα-CreKO mice at P14. (U-Y)ERG analysis at P28. (U) Representative wave responses from individual control or Vhlα-CreKO mice. Decreased amplitude (V,W) and prolonged implicit time (Y) are detected in Vhlα-CreKO mice (n=6). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. *P<0.05, **P<0.01. Error bars indicate mean ± s.d. Scale bar: 100 μm in A-H,K-T.

Fig. 7.

Vhlα-CreKO mice develop retinal degeneration in adults. (A-H) TUNEL-positive cells (red) counterstained with Hoechst (blue) are increased in the ONL in Vhlα-CreKO mice (E-H) compared with in control mice (A-D) after P14. (I,J) ONL thickness is decreased in a time-dependent manner in Vhlα-CreKO mice (n=3, respectively). (K-T) Merged images for glial fibrillary acidic protein (GFAP), Hoechst and BrdU (K,P), and IHC for PECAM1 (yellow, L,Q), glutamine synthetase (GS, green, M,R), opsin (red, N,S), and merged images for GS and opsin (O,T) in control (K-O) and Vhlα-CreKO mice (P-T). Proliferating reactive glia are detected by BrdU (arrows, P), and gliosis destroys the construction of the ONL (open arrowheads, R-T) in Vhlα-CreKO mice at P14. (U-Y)ERG analysis at P28. (U) Representative wave responses from individual control or Vhlα-CreKO mice. Decreased amplitude (V,W) and prolonged implicit time (Y) are detected in Vhlα-CreKO mice (n=6). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. *P<0.05, **P<0.01. Error bars indicate mean ± s.d. Scale bar: 100 μm in A-H,K-T.

A possible concern with our data might be the slightly, but significantly, increased branching points observed in HIF-1α; Vhlα-CreKO mice (Fig. 6G,N). Independently of HIF-α proteolysis, pVHL is known to be involved in extracellular matrix assembly and turnover (Ohh et al., 1998). Moreover, erythropoietin has been reported to be regulated by VHL and HIFs (Chen et al., 2008; Rankin et al., 2007). These previous findings suggest the involvement of multiple candidates in addition to HIF-1α and VEGF, and might explain the minor vascular changes observed in HIF-1α;Vhlα-CreKO mice. However, the dramatic rescue effects obtained by FLT1-Fc injection (Fig. 4P-U), or by the genetic inactivation of HIF-1α, but not HIF-2α (Fig. 6), show that the VHL/HIF-1α/VEGF cascade plays crucial roles in the transition from the fetal to the adult circulatory system in the retina.

Fig. 8.

Vhlα-Cre KO mice show phenotypes of hypoxic retinopathies. (A) Pre-retinal hemorrhage observed by fundus photography, (B) iris neovascularization (open arrowheads), and (C) cataract formation (arrow) in Vhlα-CreKO mice.

Fig. 8.

Vhlα-Cre KO mice show phenotypes of hypoxic retinopathies. (A) Pre-retinal hemorrhage observed by fundus photography, (B) iris neovascularization (open arrowheads), and (C) cataract formation (arrow) in Vhlα-CreKO mice.

Our current study has possible clinical implications. Adult Vhlα-CreKO mice show similar characteristics (Fig. 8) to human ischemic retinopathies, such as diabetic retinopathy, retinal vessel occlusion, and retinopathy of prematurity (Hayreh, 2007), and could provide clues for exploring the mechanisms of these human diseases. It has been recently shown that systemic administration of a prolyl hydroxylase inhibitor protects from retinal vaso-obliteration in response to hyperoxia (Sears et al., 2008), suggesting that pVHL functions in retinopathy of prematurity.

Retinal hemangioma in VHL patients and ischemic retinopathies are usually treated by laser photocoagulation that disrupts photoreceptors and increases the oxygen supply for the other neural cells from the choroidal vasculature bed, although it may cause night blindness and visual field defects (Yu and Cringle, 2001). Recently, various kinds of VEGF inhibitors (Brown and Regillo, 2007) have become possible candidates for use against ophthalmic diseases, although destruction of the physiological vasculature (Fischer et al., 2007) and neuronal dysfunction (Saint-Geniez et al., 2008) caused by long-term VEGF inhibition caution against the continuous administration of these reagents. Suppression of HIF-1α might be an alternative strategy against these diseases in the future.

Finally, all of the results presented suggest that oxygen-sensing mechanisms mediated by VHL/HIF-1α/VEGF regulate the transition from the fetal to the adult circulatory system in the retina, and could represent a theoretical basis for the retinal vascular abnormalities in human VHL disease and the development of ischemic retinopathies. It will be interesting to determine whether this oxygen-sensing system, the VHL/HIF-1α pathway, is involved in the transition of the circulatory system in other parts, such as ductus arteriosus, which would confirm its generality for future studies.

We thank Ichie Kawamori, Rie Takeshita, Haruna Koizumi and Taiga Shioda (Keio University, Tokyo, Japan) for technical assistance. This study was supported by the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT) [Grant-in-Aid for the 21st Century COE Program and Global COE Program at Keio University, and Scientific Research (No. 21791710) to T.K.]. Y.K. was supported by Keio Kanrinmaru Project. Experiments using the anti-HIF-1α antibody were supported by the Japan Society for the Promotion of Science Grant-in-Aid for Creative Scientific Research 17GS0419. M.S. is the Leader of the Global COE Program for Human Metabolomic Systems Biology and H.O. is the Leader of the Global COE Program for Stem Cell Medicine supported by MEXT.

Brown
D. M.
,
Regillo
C. D.
(
2007
).
Anti-VEGF agents in the treatment of neovascular age-related macular degeneration: applying clinical trial results to the treatment of everyday patients
.
Am. J. Ophthalmol.
144
,
627
-
637
.
Chen
J.
,
Connor
K. M.
,
Aderman
C. M.
,
Smith
L. E.
(
2008
).
Erythropoietin deficiency decreases vascular stability in mice
.
J. Clin. Invest.
118
,
526
-
533
.
Claxton
S.
,
Fruttiger
M.
(
2005
).
Oxygen modifies artery differentiation and network morphogenesis in the retinal vasculature
.
Dev. Dyn.
233
,
822
-
828
.
Fischer
C.
,
Jonckx
B.
,
Mazzone
M.
,
Zacchigna
S.
,
Loges
S.
,
Pattarini
L.
,
Chorianopoulos
E.
,
Liesenborghs
L.
,
Koch
M.
,
De Mol
M.
, et al.
. (
2007
).
Anti-PlGF inhibits growth of VEGF(R)-inhibitor-resistant tumors without affecting healthy vessels
.
Cell
131
,
463
-
475
.
Forsythe
J. A.
,
Jiang
B. H.
,
Iyer
N. V.
,
Agani
F.
,
Leung
S. W.
,
Koos
R. D.
,
Semenza
G. L.
(
1996
).
Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1
.
Mol. Cell. Biol.
16
,
4604
-
4613
.
Fruttiger
M.
,
Calver
A. R.
,
Kruger
W. H.
,
Mudhar
H. S.
,
Michalovich
D.
,
Takakura
N.
,
Nishikawa
S.
,
Richardson
W. D.
(
1996
).
PDGF mediates a neuron-astrocyte interaction in the developing retina
.
Neuron
17
,
1117
-
1131
.
Gerhardt
H.
,
Golding
M.
,
Fruttiger
M.
,
Ruhrberg
C.
,
Lundkvist
A.
,
Abramsson
A.
,
Jeltsch
M.
,
Mitchell
C.
,
Alitalo
K.
,
Shima
D.
, et al.
. (
2003
).
VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia
.
J. Cell Biol.
161
,
1163
-
1177
.
Gnarra
J. R.
,
Ward
J. M.
,
Porter
F. D.
,
Wagner
J. R.
,
Devor
D. E.
,
Grinberg
A.
,
Emmert-Buck
M. R.
,
Westphal
H.
,
Klausner
R. D.
,
Linehan
W. M.
(
1997
).
Defective placental vasculogenesis causes embryonic lethality in VHL-deficient mice
.
Proc. Natl. Acad. Sci. USA
94
,
9102
-
9107
.
Gruber
M.
,
Hu
C. J.
,
Johnson
R. S.
,
Brown
E. J.
,
Keith
B.
,
Simon
M. C.
(
2007
).
Acute postnatal ablation of Hif-2alpha results in anemia
.
Proc. Natl. Acad. Sci. USA
104
,
2301
-
2306
.
Haase
V. H.
,
Glickman
J. N.
,
Socolovsky
M.
,
Jaenisch
R.
(
2001
).
Vascular tumors in livers with targeted inactivation of the von Hippel-Lindau tumor suppressor
.
Proc. Natl. Acad. Sci. USA
98
,
1583
-
1588
.
Hammes
H. P.
,
Lin
J.
,
Wagner
P.
,
Feng
Y.
,
Vom Hagen
F.
,
Krzizok
T.
,
Renner
O.
,
Breier
G.
,
Brownlee
M.
,
Deutsch
U.
(
2004
).
Angiopoietin-2 causes pericyte dropout in the normal retina: evidence for involvement in diabetic retinopathy
.
Diabetes
53
,
1104
-
1110
.
Hayreh
S. S.
(
2007
).
Neovascular glaucoma
.
Prog. Retin. Eye Res.
26
,
470
-
485
.
Hellstrom
M.
,
Phng
L. K.
,
Hofmann
J. J.
,
Wallgard
E.
,
Coultas
L.
,
Lindblom
P.
,
Alva
J.
,
Nilsson
A. K.
,
Karlsson
L.
,
Gaiano
N.
, et al.
. (
2007
).
Dll4 signalling through Notch1 regulates formation of tip cells during angiogenesis
.
Nature
445
,
776
-
780
.
Hu
C. J.
,
Wang
L. Y.
,
Chodosh
L. A.
,
Keith
B.
,
Simon
M. C.
(
2003
).
Differential roles of hypoxia-inducible factor 1alpha (HIF-1alpha) and HIF-2alpha in hypoxic gene regulation
.
Mol. Cell. Biol.
23
,
9361
-
9374
.
Kapitsinou
P. P.
,
Haase
V. H.
(
2008
).
The VHL tumor suppressor and HIF: insights from genetic studies in mice
.
Cell Death Differ.
15
,
650
-
659
.
Kawamoto
S.
,
Niwa
H.
,
Tashiro
F.
,
Sano
S.
,
Kondoh
G.
,
Takeda
J.
,
Tabayashi
K.
,
Miyazaki
J.
(
2000
).
A novel reporter mouse strain that expresses enhanced green fluorescent protein upon Cre-mediated recombination
.
FEBS Lett.
470
,
263
-
268
.
Kubota
Y.
,
Hirashima
M.
,
Kishi
K.
,
Stewart
C. L.
,
Suda
T.
(
2008
).
Leukemia inhibitory factor regulates microvessel density by modulating oxygen-dependent VEGF expression in mice
.
J. Clin. Invest.
118
,
2393
-
2403
.
Kubota
Y.
,
Takubo
K.
,
Shimizu
T.
,
Ohno
H.
,
Kishi
K.
,
Shibuya
M.
,
Saya
H.
,
Suda
T.
(
2009
).
M-CSF inhibition selectively targets pathological angiogenesis and lymphangiogenesis
.
J. Exp. Med.
206
,
1089
-
1102
.
Kurihara
T.
,
Ozawa
Y.
,
Shinoda
K.
,
Nagai
N.
,
Inoue
M.
,
Oike
Y.
,
Tsubota
K.
,
Ishida
S.
,
Okano
H.
(
2006
).
Neuroprotective effects of angiotensin II type 1 receptor (AT1R) blocker, telmisartan, via modulating AT1R and AT2R signaling in retinal inflammation
.
Invest. Ophthalmol. Vis. Sci.
47
,
5545
-
5552
.
Lang
R. A.
(
1997
).
Apoptosis in mammalian eye development: lens morphogenesis, vascular regression and immune privilege
.
Cell Death Differ.
4
,
12
-
20
.
Lang
R. A.
,
Bishop
J. M.
(
1993
).
Macrophages are required for cell death and tissue remodeling in the developing mouse eye
.
Cell
74
,
453
-
462
.
Lobov
I. B.
,
Rao
S.
,
Carroll
T. J.
,
Vallance
J. E.
,
Ito
M.
,
Ondr
J. K.
,
Kurup
S.
,
Glass
D. A.
,
Patel
M. S.
,
Shu
W.
, et al.
. (
2005
).
WNT7b mediatesmacrophage-induced programmed cell death in patterning of the vasculature
.
Nature
437
,
417
-
421
.
Marquardt
T.
,
Ashery-Padan
R.
,
Andrejewski
N.
,
Scardigli
R.
,
Guillemot
F.
,
Gruss
P.
(
2001
).
Pax6 is required for the multipotent state of retinal progenitor cells
.
Cell
105
,
43
-
55
.
Maxwell
P. H.
,
Wiesener
M. S.
,
Chang
G. W.
,
Clifford
S. C.
,
Vaux
E. C.
,
Cockman
M. E.
,
Wykoff
C. C.
,
Pugh
C. W.
,
Maher
E. R.
,
Ratcliffe
P. J.
(
1999
).
The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis
.
Nature
399
,
271
-
275
.
Merkle
E. M.
,
Gilkeson
R. C.
(
2005
).
Remnants of fetal circulation: appearance on MDCT in adults
.
Am. J. Roentgenol.
185
,
541
-
549
.
Mitchell
C. A.
,
Rutland
C. S.
,
Walker
M.
,
Nasir
M.
,
Foss
A. J.
,
Stewart
C.
,
Gerhardt
H.
,
Konerding
M. A.
,
Risau
W.
,
Drexler
H. C.
(
2006
).
Unique vascular phenotypes following over-expression of individual VEGFA isoforms from the developing lens
.
Angiogenesis
9
,
209
-
224
.
Nicholson
D. H.
,
Green
W. R.
,
Kenyon
K. R.
(
1976
).
Light and electron microscopic study of early lesions in angiomatosis retinae
.
Am. J. Ophthalmol.
82
,
193
-
204
.
Ohh
M.
,
Yauch
R. L.
,
Lonergan
K. M.
,
Whaley
J. M.
,
Stemmer-Rachamimov
A. O.
,
Louis
D. N.
,
Gavin
B. J.
,
Kley
N.
,
Kaelin
W. G.
Jr
,
Iliopoulos
O.
(
1998
).
The von Hippel-Lindau tumor suppressor protein is required for proper assembly of an extracellular fibronectin matrix
.
Mol. Cell
1
,
959
-
968
.
Ozawa
Y.
,
Nakao
K.
,
Shimazaki
T.
,
Takeda
J.
,
Akira
S.
,
Ishihara
K.
,
Hirano
T.
,
Oguchi
Y.
,
Okano
H.
(
2004
).
Downregulation of STAT3 activation is required for presumptive rod photoreceptor cells to differentiate in the postnatal retina
.
Mol. Cell. Neurosci.
26
,
258
-
270
.
Ozawa
Y.
,
Nakao
K.
,
Shimazaki
T.
,
Shimmura
S.
,
Kurihara
T.
,
Ishida
S.
,
Yoshimura
A.
,
Tsubota
K.
,
Okano
H.
(
2007
).
SOCS3 is required to temporally fine-tune photoreceptor cell differentiation
.
Dev. Biol.
303
,
591
-
600
.
Phng
L. K.
,
Potente
M.
,
Leslie
J. D.
,
Babbage
J.
,
Nyqvist
D.
,
Lobov
I.
,
Ondr
J. K.
,
Rao
S.
,
Lang
R. A.
,
Thurston
G.
, et al.
. (
2009
).
Nrarp coordinates endothelial Notch and Wnt signaling to control vessel density in angiogenesis
.
Dev. Cell
16
,
70
-
82
.
Rankin
E. B.
,
Biju
M. P.
,
Liu
Q.
,
Unger
T. L.
,
Rha
J.
,
Johnson
R. S.
,
Simon
M. C.
,
Keith
B.
,
Haase
V. H.
(
2007
).
Hypoxia-inducible factor-2 (HIF-2) regulates hepatic erythropoietin in vivo
.
J. Clin. Invest.
117
,
1068
-
1077
.
Rao
S.
,
Lobov
I. B.
,
Vallance
J. E.
,
Tsujikawa
K.
,
Shiojima
I.
,
Akunuru
S.
,
Walsh
K.
,
Benjamin
L. E.
,
Lang
R. A.
(
2007
).
Obligatory participation of macrophages in an angiopoietin 2-mediated cell death switch
.
Development
134
,
4449
-
4458
.
Rodesch
F.
,
Simon
P.
,
Donner
C.
,
Jauniaux
E.
(
1992
).
Oxygen measurements in endometrial and trophoblastic tissues during early pregnancy
.
Obstet. Gynecol.
80
,
283
-
285
.
Ruhrberg
C.
,
Gerhardt
H.
,
Golding
M.
,
Watson
R.
,
Ioannidou
S.
,
Fujisawa
H.
,
Betsholtz
C.
,
Shima
D. T.
(
2002
).
Spatially restricted patterning cues provided by heparin-binding VEGF-A control blood vessel branching morphogenesis
.
Genes Dev.
16
,
2684
-
2698
.
Ryan
H. E.
,
Poloni
M.
,
McNulty
W.
,
Elson
D.
,
Gassmann
M.
,
Arbeit
J. M.
,
Johnson
R. S.
(
2000
).
Hypoxia-inducible factor-1alpha is a positive factor in solid tumor growth
.
Cancer Res.
60
,
4010
-
4015
.
Saint-Geniez
M.
,
Maldonado
A. E.
,
D'Amore
P. A.
(
2006
).
VEGF expression and receptor activation in the choroid during development and in the adult
.
Invest. Ophthalmol. Vis. Sci.
47
,
3135
-
3142
.
Saint-Geniez
M.
,
Maharaj
A. S.
,
Walshe
T. E.
,
Tucker
B. A.
,
Sekiyama
E.
,
Kurihara
T.
,
Darland
D. C.
,
Young
M. J.
,
D'Amore
P. A.
(
2008
).
Endogenous VEGF is required for visual function: evidence for a survival role on muller cells and photoreceptors
.
PLoS ONE
3
,
e3554
.
Sears
J. E.
,
Hoppe
G.
,
Ebrahem
Q.
,
Anand-Apte
B.
(
2008
).
Prolyl hydroxylase inhibition during hyperoxia prevents oxygen-induced retinopathy
.
Proc. Natl. Acad. Sci. USA
105
,
19898
-
19903
.
Stalmans
I.
,
Ng
Y. S.
,
Rohan
R.
,
Fruttiger
M.
,
Bouche
A.
,
Yuce
A.
,
Fujisawa
H.
,
Hermans
B.
,
Shani
M.
,
Jansen
S.
, et al.
. (
2002
).
Arteriolar and venular patterning in retinas of mice selectively expressing VEGF isoforms
.
J. Clin. Invest.
109
,
327
-
336
.
Tang
N.
,
Mack
F.
,
Haase
V. H.
,
Simon
M. C.
,
Johnson
R. S.
(
2006
).
pVHL function is essential for endothelial extracellular matrix deposition
.
Mol. Cell. Biol.
26
,
2519
-
2530
.
Yu
D. Y.
,
Cringle
S. J.
(
2001
).
Oxygen distribution and consumption within the retina in vascularised and avascular retinas and in animal models of retinal disease
.
Prog. Retin. Eye Res.
20
,
175
-
208
.
Yun
Z.
,
Maecker
H. L.
,
Johnson
R. S.
,
Giaccia
A. J.
(
2002
).
Inhibition of PPAR gamma 2 gene expression by the HIF-1-regulated gene DEC1/Stra13: a mechanism for regulation of adipogenesis by hypoxia
.
Dev. Cell
2
,
331
-
341
.

Competing interests statement

The authors declare no competing financial interests.

Supplementary information