Portal de Periódicos da CAPES

Vascular Endothelial Growth Factor Expression in the Retinal Pigment Epithelium Is Essential for Choriocapillaris Development and Visual Function

2005; Elsevier BV; Volume: 167; Issue: 5 Linguagem: Inglês

10.1016/s0002-9440(10)61231-x

1525-2191

Alexander G. Marneros, Jie Fan, Yoshihito Yokoyama, Hans Peter Gerber, Napoleone Ferrara, Rosalie K. Crouch, Björn Olsén,

Connexins and lens biology

The choroid in the eye provides vascular support for the retinal pigment epithelium (RPE) and the photoreceptors. Vascular endothelial growth factor (VEGF) derived from the RPE has been implicated in the physiological regulation of the choroidal vasculature, and overexpression of VEGF in this epithelium has been considered an important factor in the pathogenesis of choroidal neovascularization in age-related macular degeneration. Here, we demonstrate that RPE-derived VEGF is essential for choriocapillaris development. Conditional inactivation of VEGF expression in the RPE (in VEGFrpe−/− mice) results in the absence of choriocapillaris, occurrence of microphthalmia, and the loss of visual function. Severe abnormalities of RPE cells are already observed when VEGF expression in the RPE is only reduced (in VEGFrpe+/− mice), despite the formation of choroidal vessels at these VEGF levels. Finally, using Hif1arpe−/− mice we demonstrate that these roles of VEGF are not dependent on hypoxia-inducible factor-1α-mediated transcriptional regulation of VEGF expression in the RPE. Thus, hypoxia-inducible factor-1α-independent expression of VEGF is essential for choroid development. The choroid in the eye provides vascular support for the retinal pigment epithelium (RPE) and the photoreceptors. Vascular endothelial growth factor (VEGF) derived from the RPE has been implicated in the physiological regulation of the choroidal vasculature, and overexpression of VEGF in this epithelium has been considered an important factor in the pathogenesis of choroidal neovascularization in age-related macular degeneration. Here, we demonstrate that RPE-derived VEGF is essential for choriocapillaris development. Conditional inactivation of VEGF expression in the RPE (in VEGFrpe−/− mice) results in the absence of choriocapillaris, occurrence of microphthalmia, and the loss of visual function. Severe abnormalities of RPE cells are already observed when VEGF expression in the RPE is only reduced (in VEGFrpe+/− mice), despite the formation of choroidal vessels at these VEGF levels. Finally, using Hif1arpe−/− mice we demonstrate that these roles of VEGF are not dependent on hypoxia-inducible factor-1α-mediated transcriptional regulation of VEGF expression in the RPE. Thus, hypoxia-inducible factor-1α-independent expression of VEGF is essential for choroid development. The choroidal vessels between the retina and the sclera of the eye originate from mesodermal tissue surrounding the newly formed optic cup early in development. The choroidal endothelial cells are derived from the periocular mesenchyme, whereas choroidal melanocytes are derived from cranial neural crest cells.1Torczynski E Normal and abnormal ocular development in man.Prog Clin Biol Res. 1982; 82: 35-51PubMed Google Scholar, 2Etchevers HC Vincent C Le Douarin NM Couly GF The cephalic neural crest provides pericytes and smooth muscle cells to all blood vessels of the face and forebrain.Development. 2001; 128: 1059-1068PubMed Google Scholar It is thought that the choroidal vessels are induced by the retinal pigment epithelium (RPE) during development. In mice in which the RPE is converted to neural retina through transgenic expression of FGF-9 in the RPE, choroidal vessels and melanocytes are absent.3Zhao S Overbeek PA Regulation of choroid development by the retinal pigment epithelium.Mol Vis. 2001; 7: 277-282PubMed Google Scholar RPE cells are likely to maintain choroidal vessels also in the adult. For example, in the advanced form of dry age-related macular degeneration, localized RPE degeneration precedes atrophy of the underlying choroidal vessels.4Sarks JP Sarks SH Killingsworth MC Evolution of geographic atrophy of the retinal pigment epithelium.Eye. 1988; 2: 552-577Crossref PubMed Scopus (645) Google Scholar, 5Sarks SH Ageing and degeneration in the macular region: a clinico-pathological study.Br J Ophthalmol. 1976; 60: 324-341Crossref PubMed Scopus (754) Google Scholar, 6McLeod DS Taomoto M Otsuji T Green WR Sunness JS Lutty GA Quantifying changes in RPE and choroidal vasculature in eyes with age-related macular degeneration.Invest Ophthalmol Vis Sci. 2002; 43: 1986-1993PubMed Google Scholar, 7Young RW Pathophysiology of age-related macular degeneration.Surv Ophthalmol. 1987; 31: 291-306Abstract Full Text PDF PubMed Scopus (529) Google Scholar Furthermore, localized chemical or mechanical destruction of the RPE in animal models results in atrophy of choroidal vessels, loss of endothelial fenestrations, and increased deposition of collagen in the choroid.8Del Priore LV Kaplan HJ Pathogenesis of AMD.Ophthalmology. 1995; 102: 1125-1126Abstract Full Text PDF PubMed Scopus (10) Google Scholar, 9Leonard DS Zhang XG Panozzo G Sugino IK Zarbin MA Clinicopathologic correlation of localized retinal pigment epithelium debridement.Invest Ophthalmol Vis Sci. 1997; 38: 1094-1109PubMed Google Scholar, 10Korte GE Cushin S Delman N Permeability of regenerating and atrophic choriocapillaris in the rabbit.Acta Anat (Basel). 1989; 134: 144-150Crossref PubMed Scopus (13) Google Scholar These observations are consistent with ultrastructural studies that provide evidence for a continuous remodeling of the choroidal vessels in the adult,9Leonard DS Zhang XG Panozzo G Sugino IK Zarbin MA Clinicopathologic correlation of localized retinal pigment epithelium debridement.Invest Ophthalmol Vis Sci. 1997; 38: 1094-1109PubMed Google Scholar, 10Korte GE Cushin S Delman N Permeability of regenerating and atrophic choriocapillaris in the rabbit.Acta Anat (Basel). 1989; 134: 144-150Crossref PubMed Scopus (13) Google Scholar with large vessels showing extensive fenestrations at the side facing the RPE. Thus, RPE cells have an essential role in the development and the maintenance of a fenestrated choroidal vasculature in the eye. We speculate that vascular endothelial growth factor (VEGF), expressed by the RPE, might be a critical mediator of these RPE functions. This hypothesis is based on several observations. First, RPE cells stimulate tube formation of endothelial cells in co-cultures with choroidal endothelial cells, a process that was reduced by adding VEGF-neutralizing antibodies to the cell cultures.11Sakamoto T Sakamoto H Murphy TL Spee C Soriano D Ishibashi T Hinton DR Ryan SJ Vessel formation by choroidal endothelial cells in vitro is modulated by retinal pigment epithelial cells.Arch Ophthalmol. 1995; 113: 512-520Crossref PubMed Scopus (70) Google Scholar Second, VEGF expression by the RPE occurs at the time of choroidal vessel formation (shown in humans), and persists throughout adulthood (shown in humans and mice).3Zhao S Overbeek PA Regulation of choroid development by the retinal pigment epithelium.Mol Vis. 2001; 7: 277-282PubMed Google Scholar, 12Yi X Mai LC Uyama M Yew DT Time-course expression of vascular endothelial growth factor as related to the development of the retinochoroidal vasculature in rats.Exp Brain Res. 1998; 118: 155-160Crossref PubMed Scopus (52) Google Scholar, 13Gogat K Le Gat L Van Den Berghe L Marchant D Kobetz A Gadin S Gasser B Quere I Abitbol M Menasche M VEGF and KDR gene expression during human embryonic and fetal eye development.Invest Ophthalmol Vis Sci. 2004; 45: 7-14Crossref PubMed Scopus (69) Google Scholar Third, cell culture studies demonstrate the presence of VEGF at the basal side of the RPE.14Adamis AP Shima DT Yeo KT Yeo TK Brown LF Berse B D'Amore PA Folkman J Synthesis and secretion of vascular permeability factor/vascular endothelial growth factor by human retinal pigment epithelial cells.Biochem Biophys Res Commun. 1993; 193: 631-638Crossref PubMed Scopus (385) Google Scholar, 15Blaauwgeers HG Holtkamp GM Rutten H Witmer AN Koolwijk P Partanen TA Alitalo K Kroon ME Kijlstra A van Hinsbergh VW Schlingemann RO Polarized vascular endothelial growth factor secretion by human retinal pigment epithelium and localization of vascular endothelial growth factor receptors on the inner choriocapillaris. Evidence for a trophic paracrine relation.Am J Pathol. 1999; 155: 421-428Abstract Full Text Full Text PDF PubMed Scopus (467) Google Scholar This polarized localization is consistent with secretion of VEGF from the basal side of RPE cells. Fourth, VEGF induces endothelial fenestrations in several tissues,16Roberts WG Palade GE Increased microvascular permeability and endothelial fenestration induced by vascular endothelial growth factor.J Cell Sci. 1995; 108: 2369-2379Crossref PubMed Google Scholar and the preferential localization of the choroidal endothelial fenestrations at the side facing the RPE suggests a role for the VEGF-expressing RPE cells in maintaining a fenestrated phenotype of choroidal vessels. To test the hypothesis that RPE-derived VEGF is critical for choroidal vascular development or postnatal function of the choroidal vessels, we used a conditional gene-targeting approach to generate mutant mice that specifically lack VEGF expression in the RPE. A transgenic mouse line expressing Cre recombinase under the control of the tyrosinase-related protein (Trp)-1 promoter was used to generate site-specific somatic mutations in the RPE.17Mori M Metzger D Garnier JM Chambon P Mark M Site-specific somatic mutagenesis in the retinal pigment epithelium.Invest Ophthalmol Vis Sci. 2002; 43: 1384-1388PubMed Google Scholar All mice harboring the Trp1-Cre transgene were genotyped for the rd1 (retinal degeneration 1) mutation that was present in the founder mouse of FVB/N background, and mice that did not carry the rd1 mutation were used. In this mouse line, Cre was expressed in the RPE from embryonic day (E) 10.5 (time of RPE differentiation) to postnatal stages resulting in efficient excision of floxed DNA from E10.5 to adulthood.17Mori M Metzger D Garnier JM Chambon P Mark M Site-specific somatic mutagenesis in the retinal pigment epithelium.Invest Ophthalmol Vis Sci. 2002; 43: 1384-1388PubMed Google Scholar The presence of Cre expression in the RPE of this mouse line has been shown by in situ hybridization, which correlated with Cre-mediated excision of floxed DNA in polymerase chain reaction (PCR) experiments using isolated RPE cells from Trp1-Cre mice.17Mori M Metzger D Garnier JM Chambon P Mark M Site-specific somatic mutagenesis in the retinal pigment epithelium.Invest Ophthalmol Vis Sci. 2002; 43: 1384-1388PubMed Google Scholar Trp1-Cre mice (Trp1-Cretg/o) were interbred with floxed-Vegfa mice to generate homozygous mutant mice that lack the expression of VEGF-A (VEGF) in the RPE (Trp1-Cretg/0/VEGFflox/flox mice, referred to as VEGFrpe−/− mice). The generation and genotyping of floxed-Vegfa mice have been described previously.18Gerber HP Hillan KJ Ryan AM Kowalski J Keller GA Rangell L Wright BD Radtke F Aguet M Ferrara N VEGF is required for growth and survival in neonatal mice.Development. 1999; 126: 1149-1159Crossref PubMed Google Scholar Trp1-Cre mice were also interbred with floxed-Hif1a mice to generate homozygous mutant mice that lack the expression of hypoxia-inducible factor-1α (HIF-1α) in the RPE (Trp1-Cretg/0/Hif1aflox/flox mice, referred to as Hif1arpe−/− mice). The generation and genotyping of floxed-Hif1a mice have been described previously.19Schipani E Ryan HE Didrickson S Kobayashi T Knight M Johnson RS Hypoxia in cartilage: HIF-1alpha is essential for chondrocyte growth arrest and survival.Genes Dev. 2001; 15: 2865-2876Crossref PubMed Google Scholar We confirmed that expression of Cre was maintained in the RPE of the mutant mice that we generated with immunostainings using a polyclonal anti-Cre antibody (Novagen, Madison, WI) (Figure 2C). In addition, the presence of Cre-mediated excision of floxed DNA in the eyes of mutant mice was confirmed by PCR analysis. Posterior eyecups (after removal of retinal tissue) of mutant mice were used as a DNA source. PCR primers were designed to flank the loxP sites at which Cre-mediated excision of floxed DNA occurs (VEGFflox/flox PCR primers: UP 5′-gtgtgggcctgtggtccttttc-3′, DW 5′-ctggcctcacctgcattcacatct-3′; Hif1aflox/flox PCR primers: UP 5′-ggaagtaagcacctggaagtagta-3′, DW 5′-aaaaagtattgt-gttggggcagta-3′). To assess expression of VEGF at a single-cell level in embryonic and postnatal RPE cells, we used a mouse line in which Vegf expression is associated with LacZ expression, and performed subsequent staining for β-galactosidase activity.20Miquerol L Gertsenstein M Harpal K Rossant J Nagy A Multiple developmental roles of VEGF suggested by a LacZ-tagged allele.Dev Biol. 1999; 212: 307-322Crossref PubMed Scopus (239) Google Scholar In all timed pregnancies, plug date was defined as E0.5. Genomic DNA, isolated using standard protocols from portions of the embryos, was used for genotyping. For immunohistochemical analysis whole eyes were fixed in 4% paraformaldehyde, followed by infiltration in 30% sucrose. Eyes were embedded in OCT (Tissue-Tek; EMS, Hatfield, PA) and 7-μm cryostat sections were made. Sections were incubated with monoclonal rat anti-mouse CD31 (BD PharMingen, San Jose, CA) and rabbit polyclonal anti-mouse Cre antibodies (Novagen). Sections were incubated with biotinylated secondary IgG (Vector Laboratories, Burlingame, CA), and an ABC kit (Vector Laboratories) was used for detection. For histological analysis eyes from mutant and wild-type littermates were fixed in 2.5% glutaraldehyde and 1.25% paraformaldehyde in 0.1 mol/L cacodylate buffer (pH 7.4) for 1 day. After postfixation in 4% osmium tetroxide, and dehydration steps, the eyes were embedded in TABB epon (Marivac Ltd., Halifax, Canada). Semithin sections were used for methylene blue staining. Ultrathin sections were contrasted with uranyl acetate and lead citrate before evaluation in a 1200EX JEOL electron microscope. For macroscopic examination of choroid tissue, eyes were enucleated, the posterior eyeball was dissected, the retina was removed, and choroidal flat mounts were examined under a dissection microscope. Staining for β-galactosidase activity was performed on embryos at E12.5 and postnatal day (P)1. Embryos were fixed in a fixative, containing 0.2% glutaraldehyde, 5 mmol/L EGTA, pH 7.3, 2 mmol/L MgCl2, and 2% paraformaldehyde in 0.1 mol/L sodium phosphate. After washing in 0.1 mol/L sodium phosphate (pH 8), tissues were incubated in a solution containing 1 mg/ml X-Gal (Sigma, St. Louis, MO), 5 mmol/L K-ferrocyanide, and 5 mmol/L ferricyanide for 5 to 16 hours at 37°C. Tissues were then infiltrated with 30% sucrose and embedded in Tissue-Tek OCT. Sections were cut on a cryostat, mounted on Superfrost Plus slides, and dried at room temperature. Groups (n = 5) of VEGFrpe−/−, VEGFrpe+/−, and control mice were used for ERG measurements. Scotopic ERG analysis was performed as described.21Fan J Rohrer B Moiseyev G Ma JX Crouch RK Isorhodopsin rather than rhodopsin mediates rod function in RPE65 knock-out mice.Proc Natl Acad Sci USA. 2003; 100: 13662-13667Crossref PubMed Scopus (118) Google Scholar Mice were dark-adapted and anesthetized by using xylazine (20 mg/kg) and ketamine (80 mg/kg). Full-field ERGs were recorded from both eyes by using the electrophysiological system 2000 (UTAS E-2000; LKC Technologies, Gaithersburg, MD). ERGs were recorded in response to 10-μs families of single flashes with a series of light intensities (the unattenuated flash is 2.48 photopic m−2·cd·s) under scotopic conditions. The method for measuring rhodopsin in mouse retinas has been described previously.21Fan J Rohrer B Moiseyev G Ma JX Crouch RK Isorhodopsin rather than rhodopsin mediates rod function in RPE65 knock-out mice.Proc Natl Acad Sci USA. 2003; 100: 13662-13667Crossref PubMed Scopus (118) Google Scholar, 22Marneros AG Keene DR Hansen U Fukai N Moulton K Goletz PL Moiseyev G Pawlyk BS Halfter W Dong S Shibata M Li T Crouch RK Bruckner P Olsen BR Collagen XVIII/endostatin is essential for vision and retinal pigment epithelial function.EMBO J. 2004; 23: 89-99Crossref PubMed Scopus (110) Google Scholar Briefly, retinas were homogenized in 10 mmol/L Tris buffer (pH 7.5) containing 1 mmol/L ethylenediamine tetraacetic acid, 1 mmol/L 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride, protease inhibitor cocktail, and 10 μg DNase I. After centrifugation at 14,200 × g for 15 minutes, pellets were solubilized in 1% dodecylmaltoside (sodium phosphate buffer, pH 7.4) at 4°C for 2 hours, centrifuged (88,000 × g for 10 minutes), and the supernatant analyzed in a spectrophotometer. Difference spectra were determined from measurements before and after bleaching with white light in the presence of freshly prepared 20 mmol/L hydroxylamine, pH 6.0 to 7.0. The concentrations of rhodopsin were calculated using the following extinction coefficient: ε(rhodopsin) = 40,000 mol/L−1cm−1. The ocular vasculature in eyes of 3- to 12-week-old control, VEGFrpe+/−, and VEGFrpe−/− albino mice was analyzed using FITC-coupled dextran and concanavalin A lectin perfusion. After the induction of deep anesthesia, the chest cavity was opened, a needle was inserted into the right ventricle, and a perfusion cannula was inserted into the left ventricle. The mice were first perfused with phosphate-buffered saline (PBS) to wash out erythrocytes. Then fixation was achieved with 1% paraformaldehyde and 0.5% glutaraldehyde. The perfusion was continued with FITC-coupled concanavalin A lectin (40 μg/ml in PBS) (Vector Laboratories). Lectin staining was followed by PBS perfusion to remove excess concanavalin A. The eyes of the mice were enucleated and lens and retina were removed. The remaining posterior eyeballs were flat-mounted in a water-based fluorescence anti-fading medium (Vectashield; Vector Laboratories). In a second set of experiments, mice were perfused with FITC-coupled high-molecular weight dextran (Sigma), before per-fusion with 1% paraformaldehyde and 0.5% glutaraldehyde. Fluorescein angiography was performed on the Topcon Imagenet digital angiography system (TRC 50 VT camera and IMAGEnet 1.53 system; Topcon, Paramus, NJ). The photographs were captured with a 20-D lens in contact with the fundus camera lens, after intraperitoneal injection of 0.1 ml of 1% fluorescein sodium (Akorn, Decatur, IL). Primitive periocular vessels, first seen at E11.5, form the choroidal vascular network at E13.5 in mouse eyes. Choroidal melanocytes migrate between these vessels in the first 3 weeks postnatally.23Rousseau B Dubayle D Sennlaub F Jeanny JC Costet P Bikfalvi A Javerzat S Neural and angiogenic defects in eyes of transgenic mice expressing a dominant-negative FGF receptor in the pigmented cells.Exp Eye Res. 2000; 71: 395-404Crossref PubMed Scopus (43) Google Scholar We analyzed the expression of Vegf in embryonic and postnatal RPE cells using Vegf-LacZ mice. At the time of choroidal vessel formation, at E12.5, Vegf expression was prominent in the RPE (Figure 1, A and B). Expression was also detected in the lens and in the retinal ganglion cell layer (Figure 1, A and B). Immunostaining for blood vessels with anti-CD31 antibodies showed the presence of choroidal vessels at E12.5 (Figure 1C). After development of the choroid, Vegf continued to be expressed in the RPE and was also detected postnatally (Figure 1, D and E). At P1, Vegf expression continued to be expressed in the lens and the retinal ganglion cell layer as well (Figure 1, D and E). To ablate VEGF expression in the RPE, VEGFflox/flox mice were bred with Trp1-Cretg/o transgenic mice to generate Trp1-Cretg/0/VEGFflox/flox mice (referred to as VEGFrpe−/−). Trp1-Cretg/0/VEGF+/+, Trp1-Cre0/0/VEGFflox/flox, and Trp1-Cre0/0/VEGFflox/+ littermates served as controls. None of the control mice displayed morphological eye abnormalities. VEGFrpe−/− mice showed expression of Cre in the RPE (Figure 2C). Excision of floxed DNA was confirmed by PCR analysis of posterior eyecups derived from VEGFrpe−/− mice (Figure 2C). Thus, VEGF is efficiently ablated in the RPE of VEGFrpe−/− mice. The ocular phenotype described here was observed in black, brown, and white mice. VEGFrpe−/− mice displayed severe microphthalmia with complete penetrance. All VEGFrpe−/− mice investigated had a significant reduction in the size of the eyeball (n >100 VEGFrpe−/− mice), whereas all control littermates had normally sized eyes. The extent of microphthalmia showed some variability between the VEGFrpe−/− mice, but in all of these mice the microphthalmia phenotype was very striking and could easily be detected ma-croscopically (Figure 2A). After enucleation, eyes of VEGFrpe−/− mice appeared to be significantly smaller than control eyes (Figure 2B). In addition, an almost translucent appearance of the posterior eyeball was evident in VEGFrpe−/− mice (Figure 2B). Histological analysis of eyes in these mice revealed some melanocytes at the area of the optic nerve region (Figure 2D). The lens occupies the complete interior of the eyeball, and no vitreal space can be seen in VEGFrpe−/− mice (Figure 2D). Eyes of VEGFrpe−/− mice at E13.5, P1, P14, P21, P28, and P56 were examined by light microscopy and electron microscopy. Absence of choroid was observed at all stages examined, whereas normal choroid was evident in control littermates (Figure 3A). Almost no choroidal vasculature was observed in the eyes of VEGFrpe−/− mice (Figure 3B). Instead, underneath the RPE dense collagenous tissue resembling sclera could be seen. Absence of choroidal vessels was further investigated with immunostainings for endothelial cell markers (eg, CD31). Whereas strong immunostaining for CD31 was detected underneath the RPE in control littermates (Figure 3C), almost no immunostaining underneath the RPE was observed in VEGFrpe−/− mice, confirming the absence of a significant choroidal vasculature in these mice (Figure 3D). The retinal vasculature did not appear to be affected in VEGFrpe−/− mice (Figure 3D). Perfusion of VEGFrpe−/− mice with FITC-coupled high-molecular weight dextran or concanavalin A lectin confirmed the absence of choroidal vasculature in most areas of the posterior eyeball (Figure 4G), and only a fine vascular network with thin vessels was seen in some areas of the posterior eyeball (Figure 4, H and I), lacking the typical choriocapillaris as seen in control mice (Figure 4; A to C). Furthermore, no choroidal melanocytes were present, giving the posterior eyeball a translucent appearance when compared with control eyes. Notably, clusters of melanocytes were seen in the area near the optic nerve (Figure 2D).Figure 4Analysis of vasculature in flat mounts of control eyes (A–C), VEGFrpe+/− eyes (D–F), and VEGFrpe−/− eyes (G–I). A and B: A dense vascular network consisting mostly of capillaries (arrows) characterizes the choroid in control mice, supplied by choroidal arteries (arrowhead) (FITC-dextran). C: Concanavalin A lectin staining of the choroidal capillaries (arrow) and arteries (arrowhead) seen in flat mounts of control eyes. VEGFrpe+/− eyes show areas with presence of choriocapillaries (D), while at some areas choriocapillaris was lacking and only thin vessels are seen at these areas (E, F) (FITC-dextran). G: Flat mount of a FITC-dextran-perfused VEGFrpe−/− eye shows iris vessels (arrowhead) at the left. The area of the ora serrata is indicated (*), with the posterior eyeball being located at the right. Absence of choroidal vasculature is seen in most areas of the posterior eyeball in VEGFrpe−/− mice, with the exception of only some small vessels (arrow). H: These vessels form a fine vascular network (arrow) that lacks the typical fenestrated vessels of the choriocapillaris (FITC-dextran). I: Concanavalin A lectin staining of the remaining vessels seen in flat mounts of VEGFrpe−/− eyes. Original magnifications: ×20 (A, C, D–F, I); ×10 (B, H); ×2 (G).View Large Image Figure ViewerDownload Hi-res image Download (PPT) RPE cells in adult VEGFrpe−/− mice showed dystrophic changes with a reduction in the content of melanin granules. The granules that were present appeared irregularly shaped and melanin granules with the typical fusiform shape were almost lacking (Figure 5; B to E). The basal infoldings of the RPE were disorganized and reduced. Furthermore, the apical villi of the RPE did not interdigitate with photoreceptors regularly (Figure 5B). The thickness of the RPE was reduced as well (Figure 5; B to E). Ectopic RPE cells were occasionally observed in the photoreceptor outer segment layer. Notably, a single basal lamina layer was seen underlying the RPE in VEGFrpe−/− mice, adjacent to the dense collagenous scleral tissue (Figure 5F). Thus, only the innermost layer of the five-layered Bruch's membrane was present in these mice, suggesting that the formation of a proper Bruch's membrane depends on the presence of choroidal vessels (Figure 5F). Notably, RPE differentiation seemed to occur normally in VEGFrpe−/− embryos (Figure 3F). Photoreceptor outer segments did not display the regular alignment seen in control eyes, and were significantly shorter and disorganized. Focal areas of neural retina had only very few photoreceptor outer segments and some areas displayed even complete absence of outer segments (Figure 5B). Outer segments contained also irregular disk membranes (Figure 5E). Photoreceptor inner segments appeared disorganized as well (Figure 5B). There was a significant reduction in outer segment lengths in retinas of VEGFrpe−/− mice at all ages examined. Furthermore, photoreceptor nuclei were occasionally dislocated within the retina, some nuclei being localized right next to RPE cells (Figure 5, C and D). The morphological abnormalities of the RPE cells and the photoreceptors suggest an impaired visual function in VEGFrpe−/− mice. To assess photoreceptor function, scotopic ERG responses were recorded and rhodopsin content in eye homogenates was measured. Both a- and b-wave amplitudes were significantly reduced in VEGFrpe−/− mice when compared to ERGs of control mice (Figure 6). Consistent with the abnormal ERGs, retinal rhodopsin content in VEGFrpe−/− eyes was also significantly reduced (9.45 ± 13.04 pmol/retina; mean ± SD) when compared to control littermates (486.6 ± 43.27 pmol/retina; mean ± SD) (P < 0.005) (Figure 6). These results indicate an abnormal retinal function in VEGFrpe−/− mice. A key transcriptional regulator of VEGF expression in many cells is hypoxia-inducible factor-1α (HIF-1α),24Semenza GL HIF-1 and tumor progression: pathophysiology and therapeutics.Trends Mol Med. 2002; 8: S62-S67Abstract Full Text Full Text PDF PubMed Scopus (948) Google Scholar which is also expressed in RPE cells.25Ding K Scortegagna M Seaman R Birch DG Garcia JA Retinal disease in mice lacking hypoxia-inducible transcription factor-2alpha.Invest Ophthalmol Vis Sci. 2005; 46: 1010-1016Crossref PubMed Scopus (31) Google Scholar, 26Slomiany MG Rosenzweig SA IGF-1-induced VEGF and IGFBP-3 secretion correlates with increased HIF-1 alpha expression and activity in retinal pigment epithelial cell line D407.Invest Ophthalmol Vis Sci. 2004; 45: 2838-2847Crossref PubMed Scopus (92) Google Scholar, 27Martin G Schlunck G Hansen LL Agostini HT Differential expression of angioregulatory factors in normal and CNV-derived human retinal pigment epithelium.Graefes Arch Clin Exp Ophthalmol. 2004; 242: 321-326Crossref PubMed Scopus (83) Google Scholar VEGF expression can be induced by exposure to low oxygen tension, and HIF-1α is a key mediator of hypoxia-induced VEGF expression.24Semenza GL HIF-1 and tumor progression: pathophysiology and therapeutics.Trends Mol Med. 2002; 8: S62-S67Abstract Full Text Full Text PDF PubMed Scopus (948) Google Scholar To assess if VEGF expression in the RPE is dependent on HIF-1α, we generated mutant mice that specifically lack HIF-1α expression in the RPE (Hif1arpe−/− mice) and compared the eyes of these mice with VEGFrpe−/− eyes. Eyes of Hif1arpe−/− mice appeared normal in size and shape, and histological examination showed normal choroid tissue with choroidal vasculature and melanocytes (Figure 7B). Electron microscopic examination of the choroidal vasculature in Hif1arpe−/− mice revealed the presence of endothelial fenestrations (Figure 7C), as seen in control mice. Thus, Hif1arpe−/− mice do not have the abnormal choroidal phenotype seen in VEGFrpe−/− mice. This demonstrates that HIF-1α is not a critical transcriptional regulator of VEGF expression in the RPE during ocular development. VEGFrpe+/− mice show no significant microphthalmia. However, a focal loss of choroidal pigmentation could still be observed in the eyes of these mice (Figure 8, A and B). In areas at which choroidal vessels and melanocytes were present, RPE cells showed abnormalities similarly as seen in VEGFrpe−/− eyes (Figure 8C). In such areas, RPE cells next to choroidal vessels were flattened and contained abnormal pigment granules that were reduced in number (Figure 8D). Apical interdigitations of RPE cells with photoreceptors were also reduced. Fluorescein angiography demonstrates the focal loss of RPE cell pigmentation in VEGFrpe+/− eyes as distinct areas of hyperfluorescence, due to the presence of fluorescein in the underlying choroidal vasculature (Figure 8, E and F). The presence of choroidal vessels in albino VEGFrpe+/− mice was further shown in choroidal flat mounts of FITC-dextran perfused eyes. Notably, focal loss of choriocapillaris was observed in eyes of VEGFrpe+/− mice, which showed only some thin vessels at these areas (Figure 4; D to F). Our data show that deficiency of VEGF expression in the RPE results in the absence of choroid. This lack of choroid tissue in VEGFrpe−/− mice was also reflected in the absence of a proper Bruch's membrane, and only a single basal lamina under the RPE was seen adjacent to scleral collagen fibers. Thus, VEGF is essential for choroid development. Although vascular development is a very complex process that is believed to involve a large number of growth factors (such as fibroblast growth factors, platelet-derived growth factors, VEGFs, angiopoetins, and so forth), we show here that VEGF is necessary for choroid development, and that its loss cannot be compensated for by the expression of other angiogenic growth factors. Such a critical role of a single growth factor for choroid development has not been demonstrated for any other growth factor involved in angiogenic processes. For example, lack of the angiogenic FGF-2 in Fgf2−/− mice had no apparent effect on choroid development.28Ortega S Ittmann M Tsang SH Ehrlich M Basilico C Neuronal defects and delayed wound healing in mice lacking fibroblast growth factor 2.Proc Natl Acad Sci USA. 1998; 95: 5672-5677Crossref PubMed Scopus (494) Google Scholar, 29Ozaki H Okamoto N Ortega S Chang M Ozaki K Sadda S Vinores MA Derevjanik N Zack DJ Basilico C Campochiaro PA Basic fibroblast growth factor is neither necessary nor sufficient for the development of retinal neovascularization.Am J Pathol. 1998; 153: 757-765Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar However, it is possible that in addition to VEGF other growth factors influence choroidal vessel formation and maturation. A possible role for fibroblast growth factors in choroidal vascular maturation has been suggested based on the observation of immature choroidal vessels in mutant mice overexpressing a dominant-negative FGFR-1 in the RPE.23Rousseau B Dubayle D Sennlaub F Jeanny JC Costet P Bikfalvi A Javerzat S Neural and angiogenic defects in eyes of transgenic mice expressing a dominant-negative FGF receptor in the pigmented cells.Exp Eye Res. 2000; 71: 395-404Crossref PubMed Scopus (43) Google Scholar, 30Rousseau B Larrieu-Lahargue F Bikfalvi A Javerzat S Involvement of fibroblast growth factors in choroidal angiogenesis and retinal vascularization.Exp Eye Res. 2003; 77: 147-156Crossref PubMed Scopus (56) Google Scholar In addition to the absence of choroid, a dramatic microphthalmia was observed in VEGFrpe−/− mice. This finding defines a potential molecular mechanism for microphthalmia seen after selective RPE cell ablation during ocular development, such as in transgenic mice expressing the diphtheria toxin-A gene under the Trp1-gene promoter.31Raymond SM Jackson IJ The retinal pigmented epithelium is required for development and maintenance of the mouse neural retina.Curr Biol. 1995; 5: 1286-1295Abstract Full Text Full Text PDF PubMed Scopus (189) Google Scholar The observed microphthalmia in VEGFrpe−/− mice is most likely a consequence of the absence of choroidal vessels and indicates an important role of the choroidal vascular network for the growth of the eye. The Cre strain used, Trp1-Cre, expresses Cre recombinase in the RPE starting at E10.5, the stage of RPE differentiation in the mouse. Thus, excision of the Vegfa gene in VEGFrpe−/− mice results in RPE-specific deficiency of VEGF expression from the onset of RPE differentiation. Notably, Trp1-Cre mice do not express Cre recombinase in the choroidal melanocytes and vessels, or periocular mesenchyme. Therefore, the observed absence of choroid can be attributed to the lack of VEGF expression in the RPE. It needs to be pointed out that in Trp1-Cre mice, Cre recombinase expression was detected also in cells of the ciliary pigment epithelium and of the ciliary margin of the retina at E10.5, in addition to its expression in the RPE.17Mori M Metzger D Garnier JM Chambon P Mark M Site-specific somatic mutagenesis in the retinal pigment epithelium.Invest Ophthalmol Vis Sci. 2002; 43: 1384-1388PubMed Google Scholar The ciliary margin of the retina has been reported to contain retinal progenitor cells (in fish and amphibians).32Reh TA Levine EM Multipotential stem cells and progenitors in the vertebrate retina.J Neurobiol. 1998; 36: 206-220Crossref PubMed Scopus (206) Google Scholar, 33Harris WA Perron M Molecular recapitulation: the growth of the vertebrate retina.Int J Dev Biol. 1998; 42: 299-304PubMed Google Scholar Thus, it cannot be entirely excluded that the retinal abnormalities observed in VEGFrpe−/− mice may, at least in part, be a consequence of the lack of VEGF expression in some cells in the neural retina. In development, VEGF expression can be induced through various stimuli. For example, in many cells hypoxia leads to stabilization of the transcriptional regulator HIF-1α and induced expression of VEGF. In turn, this stimulates angiogenesis. However, VEGF expression can also be induced independently of hypoxia-mediated HIF-1α stabilization, for example by glucose deprivation. In fact, Hif1a−/− embryos have increased VEGF expression compared to wild-type embryos.34Kotch LE Iyer NV Laughner E Semenza GL Defective vascularization of HIF-1alpha-null embryos is not associated with VEGF deficiency but with mesenchymal cell death.Dev Biol. 1999; 209: 254-267Crossref PubMed Scopus (351) Google Scholar To elucidate if hypoxia is the main stimulus for VEGF expression in the RPE preceding the development of the choroidal vascular network, we ablated HIF-1α expression in the RPE. Our data show that VEGF expression in the RPE during development can occur independently of HIF-1α, because Hif1arpe−/− eyes are not microphthalmic and develop a choroidal vasculature in contrast to VEGFrpe−/− mice. In conclusion, RPE-derived VEGF is a key factor in choroid development, and its expression is not dependent on HIF-1α. VEGF may induce the formation of the choroidal vasculature through a direct mitogenic effect on endothelial precursor cells in the periocular mesenchyme. In addition to its mitogenic effects on endothelial cells,35Leung DW Cachianes G Kuang WJ Goeddel DV Ferrara N Vascular endothelial growth factor is a secreted angiogenic mitogen.Science. 1989; 246: 1306-1309Crossref PubMed Scopus (4581) Google Scholar VEGF is also a potent inducer of vascular tube formation.11Sakamoto T Sakamoto H Murphy TL Spee C Soriano D Ishibashi T Hinton DR Ryan SJ Vessel formation by choroidal endothelial cells in vitro is modulated by retinal pigment epithelial cells.Arch Ophthalmol. 1995; 113: 512-520Crossref PubMed Scopus (70) Google Scholar The lack of these vessel-inducing properties of VEGF is consistent with a diminished development of the choroidal vasculature. The absence of choroidal melanocytes, which are derived from cranial neural crest cells and which migrate into the developed choroidal vascular network postnatally, could be the result of the lack of a direct promigratory effect of VEGF on these melanocytes. Alternatively, the absence of choroidal melanocytes could be a result of the deficiency of a promigratory factor that is secreted by choroidal endothelial cells and thus be a secondary consequence of the lack of the development of the choroidal vasculature. It is possible that the functional and morphological defects observed in the RPE and the photoreceptors of VEGFrpe−/− mice may in part be secondary consequences of the lack of choroidal vasculature and therefore impaired nutritional support. However, there is also the possibility that VEGF itself has a role in maintaining RPE function. VEGF receptors on RPE cells have been hypothesized to mediate such an autocrine effect of VEGF on RPE cells.36Hoffmann S Masood R Zhang Y He S Ryan SJ Gill P Hinton DR Selective killing of RPE with a vascular endothelial growth factor chimeric toxin.Invest Ophthalmol Vis Sci. 2000; 41: 2389-2393PubMed Google Scholar The observation of abnormal RPE cells in VEGFrpe+/− mice even adjacent to choroidal vessels, suggests an autocrine role for VEGF in maintaining RPE function. However, it cannot be excluded that the morphological abnormalities of the choroidal vessels that were observed in VEGFrpe+/− mice may contribute to the defects in the RPE and photoreceptor cells in these mice. Although our findings demonstrate a critical role for VEGF in eye development, the role of VEGF in the aging eye remains to be determined. We thank Drs. P. Chambon, M. Mark, and M. Mori for providing Trp1-Cre mice; Dr. R. Johnson for providing Hif1a-floxed mice; Dr. A. Nagy for Vegf-LacZ mice; and Dr. Naomi Fukai, Sofiya Plotkina, Elizabeth Benecchi, and Maria Ericsson for excellent technical support.