The American Journal of Pathology, Vol. 180, No. 6, June 2012

Copyright © 2012 American Society for Investigative Pathology.

Published by Elsevier Inc. All rights reserved.

http://dx.doi.org/10.1016/j.ajpath.2012.02.021

Vascular Biology, Atherosclerosis, and Endothelium Biology

MEF2C Ablation in Endothelial Cells Reduces RetinalVessel Loss and Suppresses Pathologic Retinal

Neovascularization in Oxygen-Induced Retinopathy

Zhenhua Xu,* Junsong Gong,* Debasish Maiti,*Linh Vong,† Lijuan Wu,* John J. Schwarz,† andElia J. Duh*From the Department of Ophthalmology,* Johns Hopkins

University School of Medicine, Baltimore, Maryland; and the

Center for Cardiovascular Sciences,† Albany Medical College,

Albany, New York

Ischemic retinopathies, including retinopathy of pre-maturity and diabetic retinopathy, are major causesof blindness. Both have two phases, vessel loss andconsequent hypoxia-driven pathologic retinal neo-vascularization, yet relatively little is known about thetranscription factors regulating these processes. Myo-cyte enhancer factor 2 (MEF2) C, a member of theMEF2 family of transcription factors that plays animportant role in multiple developmental programs,including the cardiovascular system, seems to have asignificant functional role in the vasculature. We,therefore, generated endothelial cell (EC)–specificMEF2C-deficient mice and explored the role of MEF2Cin retinal vascularization during normal developmentand in a mouse model of oxygen-induced retinopa-thy. Ablation of MEF2C did not cause appreciable de-fects in normal retinal vascular development. How-ever, MEF2C ablation in ECs suppressed vessel loss inoxygen-induced retinopathy and strongly promotedvascular regrowth, consequently reducing retinal avas-cularity. This finding was associated with suppressionof pathologic retinal angiogenesis and blood-retinalbarrier dysfunction. MEF2C knockdown in cultured ret-inal ECs using small-interfering RNAs rescued ECs fromdeath and stimulated tube formation under stress con-ditions, confirming the endothelial-autonomous andantiangiogenic roles of MEF2C. HO-1 was induced byMEF2C knockdown in vitro and may play a role in theproangiogenic effect of MEF2C knockdown on retinalEC tube formation. Thus, MEF2C may play an antiangio-genic role in retinal ECs under stress conditions, and

modulation of MEF2C may prevent pathologic retinal

2548

neovascularization. (Am J Pathol 2012, 180:2548–2560;

http://dx.doi.org/10.1016/j.ajpath.2012.02.021)

Retinal neovascularization (NV) is a major cause of blind-ness in the working-age and pediatric populations owingto ischemic retinopathies, including retinopathy of pre-maturity, diabetic retinopathy, and retinal vein occlu-sions. In these conditions, blood vessel loss [vaso-oblit-eration (VO)] results in hypoxic retina, which secretesgrowth factors that stimulate pathologic retinal angiogen-esis. This destructive angiogenesis could be preventedeither by direct inhibition of the pathologic NV or byreducing retinal vessel loss, thus decreasing the hypoxicstimulus that drives the NV. Current therapeutic strate-gies focus largely on the former approach of inhibiting thelate stage of pathologic NV. There is great potential in thealternative strategy of reducing VO or promoting normalvascular repair, thereby reducing the hypoxic stimulusthat drives the late stage of NV.1–3 It would, therefore, behighly beneficial to learn more about the factors govern-ing the extent of VO and normal vascular regrowth in theischemic retinopathies.

Animal models of oxygen-induced retinopathy (OIR)have been highly useful in providing insights into retinalvascular pathobiology.4 The mouse model of OIR5 hasbeen widely used in part because of the potential forgenetic manipulation6 and has enabled the assessmentof various factors on retinal VO, vascular regrowth afterinjury, and pathologic angiogenesis.2,5–7 This model hasyielded great insights into paracrine factors in the retinalmilieu that regulate the retinal vasculature.2,6,7 Relatively

Supported by research grants from the NIH (EY018138 to E.J.D.), theJuvenile Diabetes Research Foundation (E.J.D.), and a generous gift fromCindy and Jeong H. Kim, Ph.D. E.J.D. is a recipient of an RPB CareerDevelopment Award.

Accepted for publication February 2, 2012.

Supplemental material for this article can be found at http://ajp.amjpathol.org or at http://dx.doi.org/10.1016/j.ajpath.2012.02.021.

Address reprint requests to Elia J. Duh, M.D., Wilmer Eye Institute,Johns Hopkins University School of Medicine, 400 N. Broadway, Smith

Bldg., Room 3011, Baltimore, MD 21287. E-mail: eduh@jhmi.edu.

MEF2C Ablation Reduces Retinal Vessel Loss 2549AJP June 2012, Vol. 180, No. 6

less is known regarding endothelial intrinsic factors andgenes dictating VO, vascular regrowth, and pathologicretinal NV in OIR, including relevant transcription factors.

The transcription factor myocyte enhancer factor 2(MEF2) C is known to play a critical role in cardiovasculardevelopment.8–10 MEF2C is one of four MEF2 proteins,which compose a distinct class of MADS (MCM1-aga-mous-deficiens serum response factor)-box family oftranscription factors.11 MEF2 proteins are central regula-tors of multiple developmental programs and have beendemonstrated to play a pivotal role in the morphogenesisand myogenesis of skeletal, cardiac, and smooth musclecells.12,13 Global deletion of mef2c leads to early embry-onic lethality due to cardiac defects.10 Of particular in-terest to us, MEF2C seems to have a significant functionalrole in the vasculature. MEF2C is expressed in endothe-lial cells (ECs) in vivo9 and in vitro.14,15 Targeted deletionof mef2c in mice leads to severe vascular defects.8,9 Ourlaboratory previously found that MEF2C expression andMEF2-dependent activity in ECs are activated by vascu-lar endothelial growth factor (VEGF).15 To addresswhether MEF2C has a role in retinal vascular regulation,we generated EC-specific MEF2C knockout mice andinvestigated the function of MEF2C in OIR. We found thatinactivation of MEF2C reduces retinal blood vessel lossand promotes retinal revascularization in OIR, therebysuppressing pathologic retinal NV. In addition, knock-down of MEF2C promotes angiogenesis of cultured reti-nal ECs. Together, the results of these studies suggestthat MEF2C has an important inhibitory role in regulatingretinal vascularization under stress conditions.

Materials and Methods

Animals

To generate mice that lack mef2c in ECs, Tie2-Cre trans-genic (Tie2-Cre�) mice16 were first crossed with mice thatcarry the conventional exon 2–deleted allele of mef2c(mef2c�2)10 to generate Tie2-Cre�/mef2c�/�2 mice.These genes are on the same chromosome, making theTie2-Cre and mef2c deletion linked in subsequent breed-ing. These mice were then mated with mef2cloxp/loxp

mice17,18 to generate Tie2-Cre�/mef2cloxp/�2 EC-specificmef2c-deficient mice (referred to as mef2c�EC) and lit-termate mice mef2c�/loxp (referred to as control). Thegenotypes of mef2c�EC mice and control mice wereidentified by PCR analysis of DNA isolated from tail sam-ples. The PCR primers were as follows: Tie2-Cre forward:5=-ACCAGCCAGCTATCAACTCG-3=, Tie2-Cre reverse:5=-TTACATTGGTCCAGCCACC-3=; MEF2C lox forward:5=-TTCAGGTGACCTCATTTGAACC-3=, MEF2C lox reverse:5=-GGAGCCATTGCTCATAAGAAAG-3=; and MEF2C neo-mycin forward: 5=-GGCATGCTGGGGATGCGGTGGGC-3=, MEF2C neomycin reverse: 5=-GTCACCTTAAGACATA-AAGCACCCTCC-3=. ROSA 26 reporter (R26R) mice[B6.129S4-Gt(ROSA)26Sortm1Sor/J; The Jackson Labora-tory, Bar Harbor, ME]19 were used to cross with Tie2-Cretransgenic mice or Tie2-Cre�/mef2c�/�2 mice to monitor

the specific expression of Cre recombinase under the

control of Tie2 promoter. All the animal procedures wereapproved by the Institutional Animal Care and Use Com-mittee of the Johns Hopkins University School of Medi-cine and were conducted in accordance with the Asso-ciation for Research in Vision and OphthalmologyStatement for the Use of Animals in Ophthalmic and Vi-sual Research.

Mouse Model of OIR

The OIR mouse model was described previously bySmith et al.5 Briefly, mef2c�EC mice and control litter-mates at postnatal day (P)7 along with their nursing moth-ers were put into an oxygen chamber and exposed to75% oxygen for 5 days. At P12, mice were removed fromthe chamber and were maintained in room air for 5 daysuntil P17. Mice were euthanized at different time points,and retinas were harvested for whole-mount staining orwere processed for immunohistochemical staining.

Retinal Whole-Mount Preparations and Staining

Retinal whole-mount staining was performed as previ-ously described.20 Briefly, whole eyes were fixed in 4%paraformaldehyde for 30 minutes at room temperature.After the corneas and lenses were removed, the eyecupswere fixed in 4% paraformaldehyde for 10 more minutes,and the retinas were carefully dissected. After 1 hour ofblocking in 10% normal goat serum in PBST (PBS plus3% Triton X-100; Roche Diagnostics GmbH, Mannheim,Germany), the retinas were incubated with Alexa Fluor594–conjugated isolectin GS-IB4 from Griffonia simplici-folia (Invitrogen, Carlsbad, CA) or rat anti–platelet endo-thelial cell adhesion molecule 1 (PECAM-1) antibody (BDBiosciences, San Jose, CA) at 4°C overnight. For PECAMimmunostaining, retinas were washed in PBST and thenwere incubated with anti-rat IgG conjugated with AlexaFluor 488 or 594 antibodies (Invitrogen) at 4°C overnight.Anti-NG2 chondroitin sulfate proteoglycan (Millipore, Bil-lerica, MA) was used for immunostaining of pericytes.21

After four 1-hour washes in PBST, retinas were flatmounted on slides in Fluoromount-G (Electron Micros-copy Sciences, Hatfield, PA). Retina whole-mount im-ages were taken using a Zeiss AxioVision microscope(Carl Zeiss Microscopy LLC, Thornwood, NY) at �50magnification. For LacZ staining, Tie2-Cre/R26R andR26R mice at P3 and P30 were euthanized, and eyeswere fixed in 0.5% glutaraldehyde for 5 minutes. Afterlens removal, eye cups were fixed in 0.5% glutaralde-hyde for 10 minutes. Retinas were isolated and washed inPBS three times and then were stained in X-Gal stainingsolution (0.1% w/v X-Gal, 5 mmol/L K3Fe(CN)6, 5 mmol/LK4Fe(CN)6.6H2O, and 2 mmol/L MgCl2) at 37°C overnight.After staining, retinas were mounted on slides in Aqua-Poly/Mount solution (Polysciences Inc., Warrington, PA).

Quantitation of Retinal VO and Pathologic NV

Retinal VO and NV were quantitated as previously de-scribed.20 Briefly, images of each of the four quadrants of

retina were imported into Adobe Photoshop CS2 (Adobe

2550 Xu et alAJP June 2012, Vol. 180, No. 6

Systems Inc., San Jose, CA). Retinal segments weremerged to produce one image of each entire retina. VOand pathologic NV areas were quantified by comparingthe number of pixels in the VO or NV area with the totalnumber of pixels in the retina.

Immunohistochemical and ImmunofluorescenceStaining for Retinal Frozen Sections

Mice were sacrificed on P9, and eyes were enucleatedand directly embedded in optimal cutting temperaturemedium (Electron Microscopy Sciences) at �80°C. Sec-tions (10 �m) were cut and fixed in cold 4% paraformal-dehyde. For immunohistochemical staining, sectionswere first blocked with 10% normal swine serum in 0.05mol/L Tris-buffered saline and then were incubated withbiotinylated lectin (1:40; Vector Laboratories, Burlin-game, CA) for 2 hours. After washes in Tris-bufferedsaline, a 1:100 dilution of horseradish peroxidase avidinD (Vector Laboratories) was added. Diaminobenzidinesubstrate solution was applied later to visualize antibodybinding. For immunofluorescence staining, sections werefixed as described previously herein and were blocked in5% normal goat serum for 1 hour. After overnight incu-bation with anti-MEF2C antibody (Cell Signaling Technol-ogy Inc., Beverly, MA) and anti–PECAM-1 antibody (BDBiosciences) at 4°C, sections were washed in Tris-buff-ered saline plus 0.1% Tween 20 and then were stainedwith anti-rabbit IgG conjugated with Alexa Fluor 488 (In-vitrogen) and anti-rat IgG conjugated with Cy3 (JacksonImmunoResearch Laboratories Inc., West Grove, PA).Hoechst 33258 (Invitrogen) was used to stain nuclei.Photographs were taken using a Zeiss LSM 710 confocalmicroscope (Carl Zeiss Microscopy LLC).

Measurement of Apoptosis in Retina Vessels

To detect EC apoptosis in mouse retina under hyperoxicconditions, pups were exposed to 75% oxygen for 16hours, and retinal cryosections were prepared as de-scribed previously herein. Ten-micrometer serial sec-tions were cut and collected from the iris root to theoptic nerve. The apoptotic cells were labeled using theApopTag Plus fluorescein in situ apoptosis detection kit(Millipore). After washing in PBS for 10 minutes, thesections were incubated with anti–PECAM-1 antibody tostain ECs. The retinal sections were also counterstainedwith Hoechst 33258 for nuclei staining. For each section,images were taken using a Zeiss LSM 710 confocal mi-croscope. To quantify the apoptotic ECs, cells with bothPECAM- and fluorescein-positive staining were counted.The data are presented as the mean number of apoptoticECs per section.

Blood-Retinal Barrier Assay

The blood-retinal barrier assay was performed as previ-ously described.22,23 Mice at P17 with OIR were given ani.p. injection of 1 �Ci/g body weight of [3H]-mannitol. One

hour after injection, the mice were sedated and retinas

rapidly removed and dissected to free the lens, vitreous,and any retinal pigment epithelium. The retinas wereplaced into preweighed scintillation vials. The thoraciccavity was opened, and the left superior lobe of the lungwas removed, blotted to remove excess blood, andplaced in another preweighed scintillation vial. A left dor-sal incision was made, and the retroperitoneal space wasentered without entering the peritoneal cavity. The renalvessels were clamped, and the left kidney was removed,cleaned of all fat, blotted, and placed into a preweighedscintillation vial. The remaining droplets on the tissueswere allowed to evaporate for 20 minutes. The vials wereweighed, and the tissue weights were calculated. Onemilliliter of NCSII solubilizing solution (Amersham, Chi-cago, IL) was added to each vial, and the vials wereincubated overnight in a 50°C water bath. Solubilizedtissue was brought to room temperature and decolorizedwith 20% benzoyl peroxide in toluene in a 50°C waterbath. After reequilibration to room temperature, 5 mL ofscintillation fluid CytoScint ES (Fisher Scientific, Pitts-burgh, PA) and 30 �L of glacial acetic acid were added.The vials were stored for several hours in darkness at 4°Cto eliminate chemiluminescence. Radioactivity wascounted using a liquid scintillation counter (LS 6500;Beckman Coulter Inc., Brea, CA). The counts per minuteper milligram tissue measured for lung, kidney, and retinafrom mef2c�EC mice and control mice were used tocalculate the retina/lung and retina/kidney ratios.

Cell Culture

Human retinal ECs (HRECs; Cell Systems, Kirkland, WA)were cultured in EGM-2MV medium (Lonza Inc., Walkers-ville, MD) in a humidified 5% CO2 incubator at 37°C, andthe medium was changed every 2 to 3 days. HRECs weregrown in fibronectin (Invitrogen)-coated dishes and wereused at passages 6 to 10.

siRNA Transfection

HRECs were transfected with 25 nmol/L negative controlsmall-interfering RNA (siRNA) (AM4611; Ambion, Austin,TX) and MEF2C siRNA (#1, 106791; #2, 143535; Ambion)or heme oxygenase-1 (HO-1) siRNA (s6673; Ambion)using siPORT amine transfection reagent (Ambion) ac-cording to the manufacturer’s instructions. Forty-eighthours after transfection, cells were treated with or withoutH2O2 for 15 minutes and then were used for further as-says. Specific knockdown of MEF2C and HO-1 were con-firmed by real-time PCR and Western blot analysis.

Western Blot Analysis

After treatment, cells were washed with PBS and lysed inLaemmli sample buffer (Bio-Rad Laboratories, Hercules,CA). The protein samples from total cell lysates weresubjected to 10% SDS-PAGE and were transferred toHybond ECL nitrocellulose membrane (Amersham Bio-sciences, Piscataway, NJ). After incubating with appro-priate primary and secondary antibodies, the blots were

detected using the SuperSignal West Pico or Femto

MEF2C Ablation Reduces Retinal Vessel Loss 2551AJP June 2012, Vol. 180, No. 6

chemiluminescent substrates (Pierce Biotechnology,Rockford, IL). For reprobing, the blots were washed inWestern blot stripping buffer (Thermo Fisher ScientificInc, Rockford, IL) for 10 minutes before proceeding withnew blotting. Rabbit monoclonal MEF2C antibody (1:1000) was purchased from Cell Signaling Technologies.Rabbit polyclonal anti–HO-1 antibody (1:1000) was fromEnzo Life Sciences International Inc. (Plymouth Meeting,PA). Monoclonal glyceraldehyde-3-phosphate dehydro-genase antibody (1:2000; Abcam Inc., Cambridge, MA)was used for loading control.

Apoptosis Assay

For the measurement of in vitro apoptosis, siRNA-trans-fected HRECs were treated with 300 �mol/L H2O2 inendothelial basal medium 2 (EBM2) � 2% fetal bovineserum for 6 hours, and caspase-3/7 activity was mea-sured using a Caspase-Glo 3/7 assay kit (PromegaCorp., Madison, WI) according to the manufacturer’s in-structions.

Tube Formation Assay

EC tube formation was evaluated using a two-layer col-lagen gel mixture assay, as previously described.24–27

Forty-eight hours after siRNA transfection, HRECs weretrypsinized, and 7 � 104 cells were seeded on top of thelower collagen gel layer in each well of the 24-well plate.After overnight culture, cells were then treated with dif-ferent doses of H2O2 in EBM2 � 2% fetal bovine serumfor 5 hours. Medium was then removed, and 150 �L ofcollagen gel mixture was added to each well. After 2hours of gel polymerization at 37°C, 500 �L of mediumcontaining H2O2 was added to the upper layer of colla-gen gel. Eighteen hours later, six random fields in eachwell were chosen and photographed using an Axiovert200M microscope (Carl Zeiss Microscopy LLC), and totaltube length in each field was measured using ImageJversion 1.41o (NIH, Bethesda, MD). Each experimentalcondition was assayed in triplicate, and at least threeindependent experiments were performed.

Cell Proliferation Assay

EC proliferation activity was measured using the CellTiter96 AQueous One Solution Cell proliferation assay (Pro-mega Corp.).28 Twenty-four hours after siRNA transfec-tion, cells were trypsinized and seeded into each well ofa 96-well plate at a cell density of 2500 cells. The cellswere serum starved in EBM2 supplemented with 2% fetalbovine serum for 2 to 3 hours, and the medium wasreplaced with 100 �L of fresh EBM2 with 2% fetal bovineserum containing 10 ng/mL VEGF. After 72 hours of in-cubation, 20 �L of MTS reagent from the kit was added toeach well and was incubated at 37°C in 5% CO2 for 2hours. The plates were shaken thoroughly, and absor-bance was measured using a microplate reader (FLUO-star OPTIMA; BMG Labtech GmbH, Ortenberg, Ger-

many) at 492 and 620 nm.

Cell Migration Assay

Modified Boyden chambers containing polycarbonatemembranes (Transwell, 8-�m pore size; Corning Inc.,Lowell, MA) coated with 10 �g/mL collagen were used formigration assay.28 Forty-eight hours after siRNA transfec-tion, cells were washed twice with PBS and serumstarved in EBM2 � 0.1% bovine serum albumin over-night. The next day, cells were trypsinized and seededonto a collagen I–coated Transwell plate at 6 � 104 perwell, and 10 ng/mL VEGF was added to the lower cham-ber. After incubation at 37°C for 4 hours, the stationarycells on the top of the membrane were removed by acotton swab, and the migrated cells on the bottom of themembrane were stained with DAPI. Ten photographs foreach well were randomly taken under a microscope(Zeiss Axiovert 200) at 20� objective, and cells werecounted using ImageJ.

Recombinant Adenovirus Infection

Adenovirus encoding green fluorescent protein and con-stitutively active MEF2C were gifts from Dr. Jeffery Molk-entin (University of Cincinnati, Cincinnati, OH).29 ThecDNA for constitutively active MEF2C encodes aminoacids 1 to 143 of MEF2C fused to the VP16 transcriptionalactivation domain. Adenoviruses were used to infectHRECs at multiplicities of infection of 10 and 25. Forty-eight hours after infection, cells were washed twice withPBS, and proteins were extracted using Laemmli buffer.

Microarray Analysis

HRECs were transfected with either control or MEF2CsiRNA for 24 hours. For profiling of mRNA, total RNA wasextracted using the RNeasy mini kit (Qiagen Inc., Valencia,CA) according to the manufacturer’s instructions. RNA am-plification and labeling procedures were performed by us-ing the Low RNA Input Fluorescent Linear Amplification kit(Agilent Technologies, Palo Alto, CA). RNA spike-in controls(Agilent Technologies) are added to RNA samples beforeamplification and are labeled according to the manufactur-er’s protocol. Amplified RNA (from control versus MEF2CsiRNA transfection) was labeled with either Cy3 or Cy5, andthe two labeled pools were mixed for microarray analysis.Whole human genome microarrays (G4112F; Agilent Tech-nologies) were used that contain 41,000 probes measuringapproximately 21,000 unique transcripts. Microarrays werescanned using an Agilent G2505B scanner controlled byAgilent Scan Control software version 7.0 (Agilent Technol-ogies). Data were extracted using Agilent Feature Extrac-tion software version 9.1 (Agilent Technologies). Differen-tially expressed targets were identified by using processeddata and PValueLog ratios generated by software.

Statistical Analysis

Data are presented as mean � SD or mean � SEM. Student’st-test was used to perform statistical analysis. Multiple groupswere compared by using one-way analysis of variance. A

value of P � 0.05 was considered statistically significant.

2552 Xu et alAJP June 2012, Vol. 180, No. 6

Results

Expression of MEF2C in Mouse Retina

MEF2C is known to be expressed in ECs systemically.We investigated the localization of MEF2C in the mouseretina at P9 and P35. Immunofluorescence staining anal-ysis demonstrated nuclear localization of MEF2C (Figure 1),consistent with its role as a transcription factor. Co-im-munostaining with endothelial marker PECAM-1 revealedMEF2C expression in ECs from superficial and deep ret-inal vessels at P9 (Figure 1). At P35, retinal vessels con-tinued to exhibit MEF2C staining. Aside from retinal ECs,MEF2C was also strongly expressed in some non-ECtypes distributed in the ganglion cell layer and innernuclear layer (Figure 1).

EC-Specific Ablation of mef2c

Since knockout of the mef2c gene is embryonic lethaldue to severe cardiac defects, we selectively knockedout mef2c in vascular ECs using Tie2-driven Cre ex-pression. A similar approach has previously been used tostudy selective mef2c ablation in multiple other celltypes.18,30–33 EC-specific mef2c-deficient mice weregenerated by crossing Tie2-Cre�/mef2c�/�2 mice withmef2cloxp/loxp mice. Heterozygous mef2c�/loxp mice wereindistinguishable from wild-type mice and served as lit-termate controls. Retinal vascular endothelial targeting of

Figure 1. Expression of MEF2C in retinal ECs. Frozen eye sections from P9mice (top and middle panels) and P35 mice (bottom panel) were labeledwith MEF2C antibody (green), PECAM-1 antibody (red) to visualize vessels,and Hoechst 33258 (blue) for nuclear staining. MEF2C was expressed in ECsfrom superficial and deep vascular layers and was also evident in choroidalvasculature. Besides ECs, there was prominent staining of MEF2C in non-ECsin the ganglion cell layer (GCL), inner nuclear layer (INL), and outer nuclearlayer (ONL). Hoechst counterstaining clearly demonstrated MEF2C localiza-tion primarily in the cell nucleus. Scale bars: 100 �m (top and bottompanels); 20 �m (middle panel).

Tie2-Cre was evaluated by crossing Tie2-Cre transgenic

mice with R26R mice.34 X-Gal staining of whole-mountretina from Tie2-Cre/R26R mice revealed Cre recombina-tion activity exclusively in the retinal vascular network(see Supplemental Figure S1 at http://ajp.amjpathol.org).In contrast, no staining was found in Tie2-Cre–negativeR26R retinas.

Normal Retinal Vascular Development in Micewith EC-Specific Ablation of MEF2C

We first investigated the impact of EC-specific ablation ofmef2c on retinal vascular development in mice. At P5,there was no significant difference in vascularized areabetween mef2c�EC and control mice (Figure 2, A–C). AtP17, the ganglion cell layer, inner nuclear layer, andphotoreceptor layers were histologically similar betweenmef2c�EC and control mice (Figure 2, D and E). Super-ficial and deep retinal vessels were present and similar inboth groups of mice (Figure 2, D and E). Therefore,EC-specific ablation of MEF2C did not cause abnormal-ities in normal retinal vascular development.

Attenuation of Retinal VO and ReducedEC Apoptosis in Endothelial-Specificmef2c-Deficient Mice

We then proceeded to investigate the possible role ofMEF2C on retinal vascular health in a pathologic settingusing the OIR model.5,6 In this widely used model for thestudy of ischemic retinopathies, including retinopathy ofprematurity, hyperoxic exposure is used to induce retinalVO in neonatal mice.5,6 Retinal ischemia, in turn, resultsin the up-regulation of multiple proangiogenic growth fac-tors, stimulating pathologic retinal NV. In this model,physiologic retinal revascularization does occur, albeit ata relatively slow rate. Therefore, therapeutic strategiesthat accelerate retinal revascularization (and thereby re-duce retinal ischemia) in this model significantly inhibitpathologic NV.35,36

We studied the expression of MEF2C in the hyperoxic(P8) and hypoxic (P14) phases of the OIR model. MEF2Cimmunostaining was observed in a subset of ECs at P8.Significantly increased MEF2C expression in ECs wasapparent at P14 (see Supplemental Figure S2 at http://ajp.amjpathol.org). Hoechst counterstaining clearly dem-onstrated MEF2C localization primarily in the cell nu-cleus. We investigated the effect of EC-specific mef2cablation on hyperoxia-induced retinal VO. Hyperoxia-in-duced VO is known to be rapid, reaching a peak level ofVO within 48 hours after the onset of oxygen exposure (ie,P9).6,37 No differences in the retinal vasculature betweenmef2c�EC and control mice were noted at P9 in room air(Figure 3, A and B), consistent with the observations ofnormal retinal vascular development at P5 and P17 (Fig-ure 2). At P9 in OIR, however, there was significantlyreduced retinal VO in mef2c�EC mice (Figure 3, C andD), with approximately 37% less VO in mef2c�EC micecompared with in littermate controls (Figure 3E).

It is known that hyperoxia-induced EC apoptosis con-

tributes to vessel regression.38 We studied EC apoptosis

17. GCL

MEF2C Ablation Reduces Retinal Vessel Loss 2553AJP June 2012, Vol. 180, No. 6

16 hours after initiation of the hyperoxia phase (P7 � 16hours) since the maximal degree of VO (and retinal avas-cularity) is known to occur at P9 in OIR.6 There wassignificant reduction (approximately 28%) in TUNEL-pos-itive ECs in mef2c�EC mice compared with in controlmice (P � 0.007; Figure 3, F–H). We also found prominentTUNEL-positive cells distributed in the inner nuclear layersof control mice and mef2c�EC mice (Figure 3, F and G).

Figure 2. mef2c�EC mice display no significant difference in normal retincontrol mice (A) and mef2c�EC mice (B) at P5 were stained with anti-PECAarea. No significant difference was observed (n � 3 to 7; P � 0.59). Dacounterstaining of hematoxylin in control (D) and mef2c�EC (E) retinas at PScale bars: 500 �m (A and B); 50 �m (D and E).

Marked Improvement in RetinalRevascularization Associated with ReducedPathologic NV in mef2c�EC Mice

After the initial phase of retinal VO, induced by the hy-peroxic stimulus, the retina undergoes a second phasecharacterized by a degree of “physiologic” retinal revas-

lar development compared with control mice. Retina whole mounts fromantibody for retinal vasculature. C: Quantification of the retinal vascularizediven as mean � SD. Histochemical staining of lectin for vasculature and, ganglion cell layer; INL, inner nuclear layer; and ONL, outer nuclear layer.

Figure 3. Decreased VO and EC apoptosis inmef2c�EC mice during the hyperoxia phasein the mouse OIR model. There is no vessel lossin mef2c�EC mice or their control littermates atP9 in room air (RA) (A and B) as demonstratedby staining vasculature with GS lectin–AlexaFluor 549. Retinal VO at P9 (C and D) is shownby staining vasculature with GS lectin–AlexaFluor 549, as described in Materials and Meth-ods. Avascular areas are highlighted with whitedashed lines. Scale bars: 500 �m. E: mef2c�ECmice had significantly less avascular retinal areathan did control mice at P9 (n � 5 to 11 pergroup). Data are given as mean � SD. ***P �0.0001. After 16-hour exposure to 75% oxygen,apoptosis of ECs in control (F) and mef2c�EC(G) retinas was detected using the TUNEL stain-ing (green) method and co-immunostaining withanti-PECAM antibody (red). Nuclei are stainedwith Hoechst 33258 (blue). Arrows show apop-totic ECs. Scale bars: 50 �m. GCL, ganglion celllayer; INL, inner nuclear layer; and ONL, outernuclear layer. H: The number of retinal ECs un-dergoing apoptosis is significantly decreased inmef2c�EC mice (n � 7, 12.0 � 1.8) comparedwith in their control littermates (n � 9, 16.6 �2.3). Data are given as mean � SD. **P � 0.001.

al vascuM (red)ta are g

2554 Xu et alAJP June 2012, Vol. 180, No. 6

cularization (or vascular recovery) and pathologic preret-inal NV.6 The amount of pathologic preretinal NV is in-versely related to the degree of vascular recovery sincethe latter dictates the extent of retinal hypoxia/ischemiathat drives NV.6 As shown in Figure 4, A and B, there wasa dramatic reduction in avascular retinal area inmef2c�EC mice compared with in littermate controls atP12 and P17. Specifically, avascular area was signifi-cantly decreased �50% in endothelial-specific MEF2Cknockout mice compared with in littermate controls atP12. At P17, littermate control mice exhibited an almostfourfold increase in avascular retina compared withmef2c�EC mice (Figure 4B), indicating a markedly higherdegree of retinal revascularization in mef2c�EC micecompared with in controls. Associated with this reductionin avascular retina, mef2c�EC mice exhibited signifi-cantly less (2.3-fold decrease) pathologic retinal NV (Fig-ure 4C). The vessels in the revascularized area inmef2c�EC mice were covered by pericytes, as demon-strated by the expression of NG2 (Figure 4D).

In addition to the structural end point of revasculariza-

tion, we also looked at the functional end point of isch-

emia-induced vascular leakage in OIR. At OIR P17, in-creased blood-retina barrier breakdown was found incontrol mice, an approximately fivefold increase for theleakage ratio of retina to lung and a sixfold increase forthe leakage ratio of retina to kidney (Figure 5). In contrast,endothelial-specific MEF2C knockout mice showed a sig-nificant decrease of the blood-retinal barrier comparedwith control mice at P17 in the OIR model.

MEF2C Knockdown Rescues Retinal ECs fromOxidative Stress–Induced Cell Death

The present results in the mouse model of OIR indicatethat MEF2C plays an antiangiogenic role in ECs in thissetting since ablation of the mef2c gene results in re-duced VO and EC apoptosis and in a dramatic increasein revascularization. To gain further insights into MEF2Cfunction, we investigated the effect of MEF2C knockdownin cultured HRECs. In particular, we were interested in theresponse to oxidative stimuli, as oxidative stress is known

Figure 4. Ablation of MEF2C in ECs stimulatesvascular recovery and inhibits pathologic NV. A:Retinal VO and NV of mef2c�EC mice and con-trol mice at P12 and P17 are shown by stainingretinal vasculature with GS lectin–Alexa Flour549. Avascular areas are highlighted with whitedashed lines. Scale bars: 500 �m. B: Quantifi-cation of VO area at P12 and P17. mef2c�ECretinas show significantly enhanced vascular re-covery compared with control retinas (n � 3 to4 for P12; n � 5 to 10 for P17). *P � 0.05, ***P �0.001. C: mef2c�EC mice have reduced patho-logic NV areas at P17 (n � 5, 4.4% � 1.9%)compared with littermate control mice (n � 10,10.3% � 2.7%). ***P � 0.001. D: The staining ofpericytes in mef2c�EC and control mice at P17 inthe OIR model. Whole-mount retinas were stainedwith NG2 antibody (green) for pericytes and withPECAM-1 antibody (red) to visualize vessels. Scalebars: 200 �m. Data are given as mean � SD.

to be a major regulator of the retinal vascular response in

MEF2C Ablation Reduces Retinal Vessel Loss 2555AJP June 2012, Vol. 180, No. 6

the OIR model.2 Since EC-specific ablation of mef2c re-duced EC apoptosis (Figure 3H), we first investigated theeffect of MEF2C knockdown on retinal EC survival.MEF2C protein levels in HRECs were significantly re-duced by two different MEF2C siRNA species comparedwith control siRNA (Figure 6A). Exposure to H2O2 signif-icantly increased EC apoptosis, as expected (Figure 6B).However, EC apoptosis was dramatically reduced byMEF2C knockdown compared with control siRNA (Figure6B). These data indicate that inhibiting MEF2C expres-sion might help retinal ECs survive under oxidative stressconditions.

MEF2C Knockdown in Retinal ECs StimulatesTube Formation under Basal Conditions andOxidative Stress

In the OIR model, mice with EC-specific ablation dis-played significantly increased retinal revascularization(Figure 4B). We investigated the effect of MEF2C knock-down on in vitro angiogenesis. We used a tube formationassay that has previously been used for retinal ECs.24–27

Under basal conditions and oxidative stress, knockdownof MEF2C using two different siRNAs resulted in signifi-cantly increased retinal EC tube formation compared withcontrol siRNA (Figure 7 and see Supplemental Figure S3at http://ajp.amjpathol.org). The degree of stimulation of

Figure 5. EC-specific MEF2C knockout mice displayed decreased retinalvascular leakage at P17 in the OIR model. A: The leakage ratio of retina tolung. B: The leakage ratio of retina to kidney. Data are given as mean � SEM(n � 4 to 10). *P � 0.01. RA, room air.

tube formation by MEF2C knockdown was comparable

with the stimulation of tube formation by VEGF (10 ng/mL)in control siRNA-transfected HRECs (see SupplementalFigure S4A at http://ajp.amjpathol.org). Although MEF2Cknockdown resulted in significant stimulation of tube for-mation, it had no significant effect on VEGF-induced cellproliferation or migration (see Supplemental Figure S4, Band C, at http://ajp.amjpathol.org). These data suggest aninhibitory role of MEF2C in angiogenesis under basal andstress conditions.

HO-1 Induction Is Involved in the ProangiogenicEffect of MEF2C Knockdown on EC TubeFormation

Given the role of MEF2C as a transcription factor, welooked at possible changes in gene expression resultingfrom the inhibition of MEF2C. We found that knockdownof MEF2C resulted in significant up-regulation of multipleangiogenesis-related genes through cDNA array analy-sis, including HO-1, angiopoietin 2, angiopoietin-like 4,fibroblast growth factor 2, and transforming growth factor� 2 (Table 1). There was no change in VEGF-A expres-sion in this microarray study (data not shown). Of thesegenes, induction of HO-1 was especially significant andinteresting. HO-1 is an essential enzyme in heme catab-olism and plays a major protective role in cells againststress condition. We confirmed by Western blot analysisthat MEF2C knockdown resulted in significant up-regula-

Figure 6. MEF2C regulates H2O2-induced apoptosis in HRECs. A: HRECswere transfected with negative control siRNA (siCTL) or one of two differentMEF2C siRNAs (siMEF2C #1 and #2) for 48 hours, and knocking down ofMEF2C was confirmed by Western blot analysis to detect MEF2C proteinlevel. B: After siRNA transfection, cells were treated with 300 �mol/L H2O2

for 6 hours. Knocking down of MEF2C significantly decreases H2O2-inducedapoptosis in HRECs. Data are from the average values of five independent

experiments. Data are given as mean � SD. *P � 0.05, **P � 0.01. GAPDH,glyceraldehyde-3-phosphate dehydrogenase.

2556 Xu et alAJP June 2012, Vol. 180, No. 6

tion of HO-1 (Figure 8A). Consistent with this finding,overexpression of constitutively active MEF2C resulted insignificant down-regulation of HO-1 (see SupplementalFigure S5 at http://ajp.amjpathol.org).

To determine the functional importance of HO-1 inMEF2C action, we performed experiments with siRNA-mediated knockdown of HO-1 and MEF2C. HO-1 siRNAwas effective at reducing HO-1 protein expression (Fig-ure 8B). The beneficial effect of MEF2C knockdown onHREC tube formation observed in Figure 7 was abro-

Figure 7. MEF2C regulates the tube formation ability of HRECs with orwithout H2O2 treatment. A: Representative images of tube formation assaywith HRECs after transfection with control siRNA (siCTL) and MEF2C siRNA(siMEF2C). Scale bar � 100 �m. B: After treatment with 300 �mol/L H2O2, cellswith decreased MEF2c expression exhibit stronger tube formation ability than docontrol cells. Data are from the average values of four independent experiments.Data are given as mean � SD. *P � 0.01 compared with siCTL.

Table 1. Multiple Angiogenesis-Related Genes Up-Regulated in M

Gene symbolAccession

number

HMOX1 NM_002133 Heme oxygenase (decycPLAU NM_002658 Plasminogen activator, uTGFB2 NM_003238 Transforming growth facANGPTL4 NM_139314 Angiopoietin-like 4FGF18 NM_003862 Fibroblast growth factorFGF2 NM_002006 Fibroblast growth factorANGPT2 NM_001147 Angiopoietin 2S100A7 NM_002963 S100 calcium binding prHS6ST1 NM_004807 Heparan sulfate 6-O-sulfJUN NM_002228 Jun oncogeneCEACAM1 NM_001024912 Carcinoembryonic antigeLAMA5 NM_005560 Laminin, alpha 5

ARHGAP22 NM_021226 Rho GTPase activating protein

gated when there was concomitant HO-1 knockdown. Weobserved this under basal conditions and in the presenceof oxidative stress stimuli (Figure 8, C and D). This findingindicates that induction of HO-1 plays an important role inthe beneficial effects of MEF2C knockdown on HREC,although there are likely to be other genes that are alsoinvolved in this beneficial effect.

Discussion

Although current therapies for ischemic retinopathies fo-cus largely on the late stage of pathologic retinal NV,there is great rationale for the strategy of reducing theretinal avascularity that provides the driving force forretinal NV. Such a strategy could result in reducing VO,increasing retinal revascularization, or a combination ofthe two. It is, therefore, of great benefit to learn moreabout the mechanisms that govern these processes. Agreat deal of insight has been gained using the OIRmodel, especially regarding the paracrine factors in theretinal milieu that regulate the retinal vasculature in thismodel. In contrast, relatively less is known about the ECintrinsic factors that play a role, including transcriptionfactors that regulate ECs in OIR.

MEF2C is expressed in ECs,9,14,15 and global deletionof mef2c leads to severe vascular defects in mice, sug-gesting an important functional role of MEF2C in ECs.8,9

We were, therefore, interested in the possible role thatMEF2C might play in ECs in the retina in developmentaland pathologic settings. We first confirmed the expres-sion pattern of MEF2C in the retina and found specificnuclear-localized MEF2C in vascular ECs and other celltypes in the retinal ganglion cell layer and inner nuclearlayer at P9 and P35 (Figure 1). MEF2C expression wasnot detected in photoreceptors at this early time point,although a recent report found MEF2C to be expressed ininterneurons in the inner nuclear layer and rod photore-ceptors in adult retinas.39

The mouse retina has been useful as a model forphysiologic and developmental angiogenesis, in largepart because development of the retinal vasculature oc-curs postnatally in mice.6 Although global deletion ofMEF2C results in extensive vascular malformations in the

siRNA-Treated HRECs

ameFold

enrichment P value

2.75 3.55 � 10�09

e 2.18 8.91 � 10�07

a 2 2.31 3.97 � 10�07

2.3 2.55 � 10�07

2.85 2.97 � 10�05

ic) 1.95 2.61 � 10�05

1.64 8.94 � 10�04

7 2.12 7.43 � 10�04

erase 1 1.65 9.62 � 10�03

1.5 5.88 � 10�03

ted cell adhesion molecule 1 1.57 5.77 � 10�03

1.53 4.31 � 10�03

EF2C

Gene n

ling) 1rokinastor bet

182 (bas

otein Aotransf

n-rela

22 1.61 1.22 � 10�03

MEF2C Ablation Reduces Retinal Vessel Loss 2557AJP June 2012, Vol. 180, No. 6

embryo and yolk sac,8,9 we found no effect of EC-specificablation of mef2c on normal retinal vascular development(Figure 2). Because mice with global deletion of MEF2Calso exhibit severe cardiac defects,10 the vascular de-fects observed in these mice may be secondary to thecirculatory problems resulting from the severely mal-formed heart.

We next investigated the possible role of MEF2C inregulating retinal ECs under pathologic circumstancesusing the OIR model. We found that EC-specific ablationof mef2c reduced retinal VO and dramatically promotedretinal revascularization in OIR. We found a significantdecrease in retinal vessel loss in mice with EC-specificablation of mef2c during the hyperoxia phase at P9 (Fig-ure 3, C–E), which is known to be the point at which peakVO (and retinal avascularity) is reached in OIR.6 This wasassociated with a significant decrease in retinal EC apop-tosis in mutant mice.

After P9 in the OIR model, mice undergo a relativelyslow process of retinal revascularization, first during thehyperoxia phase from P9 to P12 and subsequently duringthe phase of relative hypoxia that commences when miceare returned to a room air environment.6 In mice withEC-specific ablation of mef2c, we found substantial en-hancement of revascularization during the hyperoxia andhypoxia phases (Figure 4, A and B). Consequently, thedegree of retinal avascularity was dramatically reducedin mutant mice by P17. This was associated with a sig-nificant reduction in pathologic retinal NV in these mice atP17, as expected (Figure 4C), since the degree of retinalavascularity is a major driver of pathologic NV.6 The phys-iologic revascularization in the knockout was associatedwith pericytes (Figure 4D), which are implicated in vesselstabilization.40 In addition, the knockout mice exhibited asignificant reduction in blood-retinal barrier dysfunction(Figure 5), supporting the conclusion that endothelial abla-tion of MEF2C promotes physiologic revascularization andnormalization of the retinal vasculature.

We complemented the studies in mice with in vitrostudies in cultured HRECs, using an siRNA approach tosuppress MEF2C expression. We exposed HRECs toconditions of oxidative stress since this is thought to be amajor factor in vessel loss in OIR.2 Consistent with thefindings in mouse retina during OIR, MEF2C deficiencypromoted angiogenesis in retinal ECs in vitro. Specifi-cally, MEF2C knockdown reduced apoptosis in retinalECs exposed to oxidative stress (Figure 6) and enhancedEC angiogenesis in a tube formation assay (Figure 7).

Figure 8. HO-1 mediates the inhibitory effect of MEF2C in regulating ECtube formation under oxidative stress conditions. A: Down-regulation ofMEF2C increases HO-1 protein expression in HRECs. HO-1 protein wasdetected by Western blot analysis 48 hours after HRECs were transfected withMEF2C siRNA (siMEF2C). B: HO-1 siRNA (siHO-1) is effective in knockingdown HO-1 expression in HRECs. MEF2C and HO-1 protein were detectedby Western blot analysis 48 hours after siRNA transfection. C: Knockdown ofHO-1 significantly abrogates increased tube formation in siMEF2C-trans-fected HRECs with or without H2O2 treatment. Representative images fortube formation assay are shown. Scale bar � 100 �m. D: Quantitation of tubeformation in siRNA-transfected HRECs. Data are from the average values of

four independent experiments. Data are given as mean � SD. *P � 0.01. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; siCTL, control siRNA.

2558 Xu et alAJP June 2012, Vol. 180, No. 6

These in vitro findings are in line with the results in OIR,which demonstrate a reduction in retinal EC apoptosisand a dramatic enhancement in retinal revascularizationin mice with EC ablation of mef2c. These results confirman EC-autonomous role for MEF2C in regulating angio-genesis.

To gain further insights into the antiangiogenic actionsof MEF2C in vitro, we used microarray analysis to inves-tigate gene expression changes after siRNA-mediatedMEF2C knockdown in retinal ECs. HO-1 expression wassignificantly increased and was of particular interest as apossible mechanism for inhibition of EC apoptosis andenhancement of angiogenesis, especially under oxida-tive stress conditions. HO-1 is well-known to play animportant protective role for blood vessels, in large partby mitigating the effects of oxidative stress.41–43 We,therefore, directed additional studies toward the possiblerole of HO-1 in the HREC culture system. We confirmedthat knockdown of MEF2C significantly increased HO-1expression at the protein level in HRECs (Figure 8A). Inaddition, siRNA-mediated knockdown of HO-1 sup-pressed retinal EC tube formation in vitro under normal orstress conditions (Figure 8, C and D). In addition, con-comitant knockdown of HO-1 significantly abrogated theeffect of MEF2C knockdown on enhancing retinal ECtube formation (Figure 8, C and D). This finding indicatesthat the induction of HO-1 is an important mechanism ofthe proangiogenic effects of knocking down MEF2C, atleast in vitro, although there are likely to be other genesinvolved in this beneficial effect. Whether MEF2C has adirect or indirect effect on HO-1 expression requires fur-ther investigation. It will be of interest to investigate therole of HO-1 in the regulation of retinal ECs in OIR.

Overall, these findings in vivo and in vitro indicate thatMEF2C plays an important role in retinal ECs in regulatingcellular response to stress conditions. These findings inmice were performed using a Tie2-Cre system, which iswidely used to ablate gene expression in ECs. Of note,this system has also been found to be active in somehematopoietic stem cells.44 Therefore, since macro-phages, myeloid progenitor cells, and endothelial pro-genitor cells, which are derived from hematopoietic stemcells, could also affect revascularization in OIR, we can-not exclude a contribution from these other cell types tothe effects we observed.

It is becoming increasingly appreciated that reducingretinal avascularity may be an important therapeutic strat-egy for ischemic retinopathies, therefore making it criticalto gain a better understanding of retinal VO and revas-cularization in OIR. Much has been discovered about theparacrine factors in the retinal milieu that regulate theretinal vascular response to OIR. Research into retinalavascularity in animal models of OIR has identified mol-ecules that can either reduce VO or enhance revascular-ization in OIR, including insulin-like growth factor bindingprotein-3,45,46 erythropoietin,47 and Norrin.48 In addition,pathogenic molecules have been identified in OIR, in-cluding thrombospondin-1,49 tumor necrosis factor-�,50

NADPH oxidase,51–53 peroxynitrite/nitrosative stress,54,55

trans-arachidonic acids,56 and semaphorin 3A.36,57 How-

ever, less is known about the molecular determinants of

the retinal EC response to OIR, which affects vasculardegeneration and revascularization. The present studyindicates that MEF2C is an important transcription factorthat modulates the response of retinal ECs to stress con-ditions. Inhibiting MEF2C in retinal ECs protects the reti-nal vasculature against degeneration and enhances nor-mal revascularization in OIR. Modulation of MEF2C inretinal ECs could, therefore, be a therapeutic strategy toreduce retinal avascularity in ischemic retinopathies.

Acknowledgments

We thank Jeremy Nathans, Gerard Lutty, and ImranBhutto for helpful suggestions, Yanshu Wang for techni-cal assistance regarding retinal whole-mount preparationand staining, Rhonda Grebe for confocal microscopetraining, and Wayne Yu (Sidney Kimmel ComprehensiveCancer Center at Johns Hopkins University) for help withthe microarray study.

References

1. Hartnett ME: The effects of oxygen stresses on the development offeatures of severe retinopathy of prematurity: knowledge from the50/10 OIR model. Doc Ophthalmol 2010, 120:25–39

2. Sapieha P, Joyal JS, Rivera JC, Kermorvant-Duchemin E, Sennlaub F,Hardy P, Lachapelle P, Chemtob S: Retinopathy of prematurity: un-derstanding ischemic retinal vasculopathies at an extreme of life.J Clin Invest 2010, 120:3022–3032

3. Smith LE: Through the eyes of a child: understanding retinopathythrough ROP the Friedenwald lecture. Invest Ophthalmol Vis Sci2008, 49:5177–5182

4. Madan A, Penn JS: Animal models of oxygen-induced retinopathy.Front Biosci 2003, 8:d1030–d1043

5. Smith LE, Wesolowski E, McLellan A, Kostyk SK, D’Amato R, SullivanR, D’Amore PA: Oxygen-induced retinopathy in the mouse. InvestOphthalmol Vis Sci 1994, 35:101–111

6. Stahl A, Connor KM, Sapieha P, Chen J, Dennison RJ, Krah NM,Seaward MR, Willett KL, Aderman CM, Guerin KI, Hua J, Lofqvist C,Hellstrom A, Smith LE: The mouse retina as an angiogenesis model.Invest Ophthalmol Vis Sci 2010, 51:2813–2826

7. Kermorvant-Duchemin E, Sapieha P, Sirinyan M, Beauchamp M,Checchin D, Hardy P, Sennlaub F, Lachapelle P, Chemtob S: Under-standing ischemic retinopathies: emerging concepts from oxygen-induced retinopathy. Doc Ophthalmol 2009, 120:51–60

8. Bi W, Drake CJ, Schwarz JJ: The transcription factor MEF2C-nullmouse exhibits complex vascular malformations and reduced car-diac expression of angiopoietin 1 and VEGF. Dev Biol 1999, 211:255–267

9. Lin Q, Lu J, Yanagisawa H, Webb R, Lyons GE, Richardson JA, OlsonEN: Requirement of the MADS-box transcription factor MEF2C forvascular development. Development 1998, 125:4565–4574

10. Lin Q, Schwarz J, Bucana C, Olson EN: Control of mouse cardiacmorphogenesis and myogenesis by transcription factor MEF2C. Sci-ence 1997, 276:1404–1407

11. Shore P, Sharrocks AD: The MADS-box family of transcription factors.Eur J Biochem 1995, 229:1–13

12. Olson EN, Perry M, Schulz RA: Regulation of muscle differentiation bythe MEF2 family of MADS box transcription factors. Dev Biol 1995,172:2–14

13. Potthoff MJ, Olson EN: MEF2: a central regulator of diverse develop-mental programs. Development 2007, 134:4131–4140

14. Hosking BM, Wang SC, Chen SL, Penning S, Koopman P, Muscat GE:SOX18 directly interacts with MEF2C in endothelial cells. BiochemBiophys Res Commun 2001, 287:493–500

15. Maiti D, Xu Z, Duh EJ: Vascular endothelial growth factor induces

MEF2C and MEF2-dependent activity in endothelial cells. Invest Oph-thalmol Vis Sci 2008, 49:3640–3648

MEF2C Ablation Reduces Retinal Vessel Loss 2559AJP June 2012, Vol. 180, No. 6

16. Koni PA, Joshi SK, Temann UA, Olson D, Burkly L, Flavell RA: Con-ditional vascular cell adhesion molecule 1 deletion in mice: impairedlymphocyte migration to bone marrow. J Exp Med 2001, 193:741–754

17. Vong LH, Ragusa MJ, Schwarz JJ: Generation of conditionalMef2cloxP/loxP mice for temporal- and tissue-specific analyses. Gen-esis 2005, 43:43–48

18. Vong L, Bi W, O’Connor-Halligan KE, Li C, Cserjesi P, Schwarz JJ:MEF2C is required for the normal allocation of cells between theventricular and sinoatrial precursors of the primary heart field. DevDyn 2006, 235:1809–1821

19. Soriano P: Generalized lacZ expression with the ROSA26 Cre re-porter strain. Nat Genet 1999, 21:70–71

20. Connor KM, Krah NM, Dennison RJ, Aderman CM, Chen J, GuerinKI, Sapieha P, Stahl A, Willett KL, Smith LE: Quantification ofoxygen-induced retinopathy in the mouse: a model of vessel loss,vessel regrowth and pathological angiogenesis. Nat Protoc 2009,4:1565–1573

21. Ozerdem U, Grako KA, Dahlin-Huppe K, Monosov E, Stallcup WB:NG2 proteoglycan is expressed exclusively by mural cells duringvascular morphogenesis. Dev Dyn 2001, 222:218–227

22. Derevjanik NL, Vinores SA, Xiao WH, Mori K, Turon T, Hudish T,Dong S, Campochiaro PA: Quantitative assessment of the integrityof the blood-retinal barrier in mice. Invest Ophthalmol Vis Sci 2002,43:2462–2467

23. Huang H, Van de Veire S, Dalal M, Parlier R, Semba RD, Carmeliet P,Vinores SA: Reduced retinal neovascularization, vascular permeabil-ity, and apoptosis in ischemic retinopathy in the absence of prolylhydroxylase-1 due to the prevention of hyperoxia-induced vascularobliteration. Invest Ophthalmol Vis Sci 2011, 52:7565–7573

24. Im E, Venkatakrishnan A, Kazlauskas A: Cathepsin B regulates theintrinsic angiogenic threshold of endothelial cells. Mol Biol Cell 2005,16:3488–3500

25. Im E, Kazlauskas A: Regulating angiogenesis at the level of PtdIns-4,5-P2. EMBO J 2006, 25:2075–2082

26. Im E, Motiejunaite R, Aranda J, Park EY, Federico L, Kim TI, Clair T,Stracke ML, Smyth S, Kazlauskas A: Phospholipase C� activationdrives increased production of autotaxin in endothelial cells andlysophosphatidic acid-dependent regression. Mol Cell Biol 2010, 30:2401–2410

27. Watanabe D, Takagi H, Suzuma K, Suzuma I, Oh H, Ohashi H,Kemmochi S, Uemura A, Ojima T, Suganami E, Miyamoto N, Sato Y,Honda Y: Transcription factor Ets-1 mediates ischemia- and vascularendothelial growth factor-dependent retinal neovascularization. Am JPathol 2004, 164:1827–1835

28. Xu Z, Maiti D, Kisiel W, Duh EJ: Tissue factor pathway inhibitor-2 isupregulated by vascular endothelial growth factor and suppressesgrowth factor-induced proliferation of endothelial cells. ArteriosclerThromb Vasc Biol 2006, 26:2819–2825

29. Xu J, Gong NL, Bodi I, Aronow BJ, Backx PH, Molkentin JD: Myocyteenhancer factors 2A and 2C induce dilated cardiomyopathy in trans-genic mice. J Biol Chem 2006, 281:9152–9162

30. Khiem D, Cyster JG, Schwarz JJ, Black BL: A p38 MAPK-MEF2Cpathway regulates B-cell proliferation. Proc Natl Acad Sci U S A:2008, 105:17067–17072

31. Li H, Radford JC, Ragusa MJ, Shea KL, McKercher SR, Zaremba JD,Soussou W, Nie Z, Kang YJ, Nakanishi N, Okamoto S, Roberts AJ,Schwarz JJ, Lipton SA: Transcription factor MEF2C influences neuralstem/progenitor cell differentiation and maturation in vivo. Proc NatlAcad Sci U S A 2008, 105:9397–9402

32. Verzi MP, Agarwal P, Brown C, McCulley DJ, Schwarz JJ, Black BL:The transcription factor MEF2C is required for craniofacial develop-ment. Dev Cell 2007, 12:645–652

33. Wilker PR, Kohyama M, Sandau MM, Albring JC, Nakagawa O,Schwarz JJ, Murphy KM: Transcription factor Mef2c is required for Bcell proliferation and survival after antigen receptor stimulation. NatImmunol 2008, 9:603–612

34. Claxton S, Kostourou V, Jadeja S, Chambon P, Hodivala-Dilke K,Fruttiger M: Efficient, inducible Cre-recombinase activation in vascu-lar endothelium. Genesis 2008, 46:74–80

35. Connor KM, SanGiovanni JP, Lofqvist C, Aderman CM, Chen J,Higuchi A, Hong S, Pravda EA, Majchrzak S, Carper D, Hellstrom A,Kang JX, Chew EY, Salem N Jr, Serhan CN, Smith LE: Increaseddietary intake of omega-3-polyunsaturated fatty acids reduces path-

ological retinal angiogenesis. Nat Med 2007, 13:868–873

36. Joyal JS, Sitaras N, Binet F, Rivera JC, Stahl A, Zaniolo K, Shao Z,Polosa A, Zhu T, Hamel D, Djavari M, Kunik D, Honore JC, PicardE, Zabeida A, Varma DR, Hickson G, Mancini J, Klagsbrun M,Costantino S, Beausejour C, Lachapelle P, Smith LE, Chemtob S,Sapieha P: Ischemic neurons prevent vascular regeneration ofneural tissue by secreting semaphorin 3A. Blood 2011, 117:6024 –6035

37. Lange C, Ehlken C, Stahl A, Martin G, Hansen L, Agostini HT: Kineticsof retinal vaso-obliteration and neovascularisation in the oxygen-induced retinopathy (OIR) mouse model. Graefes Arch Clin ExpOphthalmol 2009, 247:1205–1211

38. Alon T, Hemo I, Itin A, Pe’er J, Stone J, Keshet E: Vascular endothelialgrowth factor acts as a survival factor for newly formed retinal vesselsand has implications for retinopathy of prematurity. Nat Med 1995,1:1024–1028

39. Escher P, Schorderet DF, Cottet S: Altered expression of the tran-scription factor Mef2c during retinal degeneration in Rpe65�/� mice.Invest Ophthalmol Vis Sci 2011, 52:5933–5940

40. Kielczewski JL, Hu P, Shaw LC, Li Calzi S, Mames RN, GardinerTA, McFarland E, Chan-Ling T, Grant MB: Novel protective prop-erties of IGFBP-3 result in enhanced pericyte ensheathment, re-duced microglial activation, increased microglial apoptosis, andneuronal protection after ischemic retinal injury. Am J Pathol 2011,178:1517–1528

41. Dulak J, Loboda A, Jozkowicz A: Effect of heme oxygenase-1 onvascular function and disease. Curr Opin Lipidol 2008, 19:505–512

42. Dulak J, Deshane J, Jozkowicz A, Agarwal A: Heme oxygenase-1 andcarbon monoxide in vascular pathobiology: focus on angiogenesis.Circulation 2008, 117:231–241

43. Yachie A, Niida Y, Wada T, Igarashi N, Kaneda H, Toma T, Ohta K,Kasahara Y, Koizumi S: Oxidative stress causes enhanced endothe-lial cell injury in human heme oxygenase-1 deficiency. J Clin Invest1999, 103:129–135

44. Constien R, Forde A, Liliensiek B, Grone HJ, Nawroth P, HammerlingG, Arnold B: Characterization of a novel EGFP reporter mouse tomonitor Cre recombination as demonstrated by a Tie2 Cre mouseline. Genesis 2001, 30:36–44

45. Kielczewski JL, Jarajapu YP, McFarland EL, Cai J, Afzal A, Li CalziS, Chang KH, Lydic T, Shaw LC, Busik J, Hughes J, Cardounel AJ,Wilson K, Lyons TJ, Boulton ME, Mames RN, Chan-Ling T, GrantMB: Insulin-like growth factor binding protein-3 mediates vascularrepair by enhancing nitric oxide generation. Circ Res 2009, 105:897–905

46. Lofqvist C, Chen J, Connor KM, Smith AC, Aderman CM, Liu N,Pintar JE, Ludwig T, Hellstrom A, Smith LE: IGFBP3 suppressesretinopathy through suppression of oxygen-induced vessel lossand promotion of vascular regrowth. Proc Natl Acad Sci U S A2007, 104:10589 –10594

47. Chen J, Connor KM, Aderman CM, Smith LE: Erythropoietin defi-ciency decreases vascular stability in mice. J Clin Invest 2008, 118:526–533

48. Ohlmann A, Seitz R, Braunger B, Seitz D, Bosl MR, Tamm ER: Norrinpromotes vascular regrowth after oxygen-induced retinal vessel lossand suppresses retinopathy in mice. J Neurosci 2010, 30:183–193

49. Wang S, Wu Z, Sorenson CM, Lawler J, Sheibani N: Thrombospondin-1-deficient mice exhibit increased vascular density during retinalvascular development and are less sensitive to hyperoxia-mediatedvessel obliteration. Dev Dyn 2003, 228:630–642

50. Gardiner TA, Gibson DS, de Gooyer TE, de la Cruz VF, McDonaldDM, Stitt AW: Inhibition of tumor necrosis factor-� improves physio-logical angiogenesis and reduces pathological neovascularization inischemic retinopathy. Am J Pathol 2005, 166:637–644

51. Saito Y, Geisen P, Uppal A, Hartnett ME: Inhibition of NAD(P)Hoxidase reduces apoptosis and avascular retina in an animal modelof retinopathy of prematurity. Mol Vis 2007, 13:840–853

52. Saito Y, Uppal A, Byfield G, Budd S, Hartnett ME: Activated NAD(P)Hoxidase from supplemental oxygen induces neovascularization inde-pendent of VEGF in retinopathy of prematurity model. Invest Ophthal-mol Vis Sci 2008, 49:1591–1598

53. Al-Shabrawey M, Bartoli M, El-Remessy AB, Platt DH, Matragoon S,Behzadian MA, Caldwell RW, Caldwell RB: Inhibition of NAD(P)Hoxidase activity blocks vascular endothelial growth factor overex-pression and neovascularization during ischemic retinopathy. Am J

Pathol 2005, 167:599–607

2560 Xu et alAJP June 2012, Vol. 180, No. 6

54. Beauchamp MH, Sennlaub F, Speranza G, Gobeil F Jr, Checchin D,Kermorvant-Duchemin E, Abran D, Hardy P, Lachapelle P, Varma DR,Chemtob S: Redox-dependent effects of nitric oxide on microvascu-lar integrity in oxygen-induced retinopathy. Free Radic Biol Med2004, 37:1885–1894

55. Brooks SE, Gu X, Samuel S, Marcus DM, Bartoli M, Huang PL,

Caldwell RB: Reduced severity of oxygen-induced retinopathy ineNOS-deficient mice. Invest Ophthalmol Vis Sci 2001, 42:222–228

56. Kermorvant-Duchemin E, Sennlaub F, Sirinyan M, Brault S, Andelfin-ger G, Kooli A, Germain S, Ong H, d’Orleans-Juste P, Gobeil F Jr, ZhuT, Boisvert C, Hardy P, Jain K, Falck JR, Balazy M, Chemtob S:Trans-arachidonic acids generated during nitrative stress induce athrombospondin-1-dependent microvascular degeneration. Nat Med2005, 11:1339–1345

57. Duh EJ: Sema 3A resists retinal revascularization. Blood 2011, 117:5785–5786