Sung Wook Park,1,2 Jang-Hyuk Yun,3,4 Jin Hyoung Kim,1 Kyu-Won Kim,5 Chung-Hyun Cho,3,4 andJeong Hun Kim1,2,6

Angiopoietin 2 Induces PericyteApoptosis via a3b1 IntegrinSignaling in Diabetic RetinopathyDiabetes 2014;63:3057–3068 | DOI: 10.2337/db13-1942

Pericyte loss is an early characteristic change in diabeticretinopathy (DR). Despite accumulating evidence thathyperglycemia-induced angiopoietin 2 (Ang2) has a cen-tral role in pericyte loss, the precise molecular mecha-nism has not been elucidated. This study investigatedthe role of Ang2 in pericyte loss in DR. We demonstratedthat pericyte loss occurred with Ang2 increase in thediabetic mouse retina and that the source of Ang2 couldbe the endothelial cell. Ang2 induced pericyte apoptosisvia the p53 pathway under high glucose, whereas Ang2alone did not induce apoptosis. Integrin, not Tie-2 re-ceptor, was involved for Ang2-induced pericyte apopto-sis under high glucose as an Ang2 receptor. Highglucose changed the integrin expression pattern, whichincreased integrin a3 and b1 in the pericyte. Further-more, Ang2-induced pericyte apoptosis in vitro was ef-fectively attenuated via p53 suppression by blockingintegrin a3 and b1. Although intravitreal injection ofAng2 induced pericyte loss in C57BL/6J mice retina invivo, intravitreal injection of anti-integrin a3 and b1 anti-bodies attenuated Ang2-induced pericyte loss. Taken to-gether, Ang2 induced pericyte apoptosis under highglucose via a3b1 integrin. Glycemic control or blockingAng2/integrin signaling could be a potential therapeutictarget to prevent pericyte loss in early DR.

Diabetic retinopathy (DR) is the leading cause of visualloss in working-aged people and the most common

microvascular complication in diabetic patients despitethe recent improvement in the management of DR viaglycemic control and photocoagulation (1). Macular edema(leakage) and neovascularization (angiogenesis) both causesevere vision loss in DR (2), and pericyte loss is one of theearliest and most characteristic changes of DR (3).

The pericyte plays two major important clinical roles inDR. First, the pericyte enwraps endothelial cells to keepthe integrity of inner blood–retinal barrier (BRB) with therole of microvascular autoregulation (4). Thus, pericyteloss could weaken the inner BRB, even when endothelialcells are intact, and lead to capillary instability and vas-cular leakage in macular edema. Second, microaneurysmand neovascularization occur in proliferating endothelialcells at the site of pericyte loss (5). Because pericyte loss isan early diabetic change, preventing pericyte loss for theprimary prevention of DR would be beneficial.

Although pericyte loss is important in early DR, themechanism by which hyperglycemia leads to pericyte lossremains largely unknown. However, Ang2 plays a critical rolein pericyte loss in DR. Hyperglycemia causes pericyteapoptosis and, ultimately, pericyte loss (6–8). Ang2 increasesin the vitreous of patients with proliferative DR (9). In ad-dition, Ang2 is upregulated by hyperglycemia in the diabeticretina and in endothelial cells (10–12). Ang2 induces pericyteloss in normal mice retina and in mice overexpressing Ang2(10,13). Thus, we postulated that hyperglycemia increasesAng2, which in turn induces pericyte apoptosis in DR.

1Fight against Angiogenesis-Related Blindness Laboratory, Biomedical ResearchInstitute, Seoul National University Hospital, Seoul, Korea2Department of Biomedical Sciences, College of Medicine, Seoul National Uni-versity, Seoul, Korea3Department of Pharmacology and Ischemic/Hypoxic Disease Institute, College ofMedicine, Seoul National University, Seoul, Korea4Cancer Research Institute, College of Medicine, Seoul National University, Seoul,Korea5Department of Pharmacy, Seoul National University, Seoul, Korea6Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea

Corresponding authors: Jeong Hun Kim, steph25@snu.ac.kr; and Chung-HyunCho, iamhyun@snu.ac.kr.

Received 24 December 2013 and accepted 31 March 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1942/-/DC1.

S.W.P. and J.-H.Y. contributed equally to this work.

© 2014 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

Diabetes Volume 63, September 2014 3057

COMPLIC

ATIO

NS

The role of Ang2 in pericyte loss has been studied(10,11,13–15). Apoptosis and migration are both sug-gested mechanisms of pericyte loss by Ang2; however,the precise mechanism by which Ang2 induces pericyteloss has not yet been fully elucidated. Ang2 has beenknown to bind to the endothelial-specific Tie-2 tyrosinereceptor with similar affinity to Ang1 (16). Ang2 acts inan autocrine manner in angiogenesis. This endothelial cell–derived antagonistic ligand of vessel maturation andremodeling controls the Ang1–Tie-2 signaling axis (17). Al-though Ang2 has been postulated to naturally bind to theTie-2 receptor in the pericyte as the Ang-Tie system (11,13),whether Tie-2 indeed serves as a receptor for Ang2-inducedpericyte loss is not clear. Integrin was recently found tomediate platelet-derived growth factor-BB–induced pericyteloss in tumor vessels (18). Also, Ang2 binds to integrinand regulates angiogenesis through integrin signaling(19). Thus, we hypothesized that Ang2 induces pericyteapoptosis via integrin signaling.

In this study, we demonstrated that Ang2 inducedpericyte apoptosis via the p53 pathway under high glucose.Interestingly, integrin, not Tie-2 receptor, was importantfor Ang2-induced pericyte apoptosis under high glucose.High glucose increased integrin a3b1 in pericytes. Further-more, our results showed Ang2-induced pericyte apoptosiswas effectively attenuated by blocking integrin a3b1 invitro and in vivo. Taken together, Ang2 induced pericyteapoptosis via a3b1 integrin signaling in DR.

RESEARCH DESIGN AND METHODS

Cell CulturesHuman umbilical vein endothelial cells (HUVECs; Lonza),human retina microvascular endothelial cells (HRMECs;ACBRI), human brain astrocytes (ACBRI), and humanpericytes (PromoCell) were maintained in EBM-2, M199medium, DMEM with 20% FBS, and pericyte mediacontaining growth factors (PromoCell), respectively. Allcells were cultured at 37°C in an incubator with a humid-ified atmosphere of 95% O2 and 5% CO2.

Reagents and AntibodiesRecombinant mouse and human Ang2, human integrina3b1, and phycoerythrin (PE)-conjugated anti–Tie-2 andmouse IgG antibodies were purchased from R&D Systems.Other reagents and antibodies were anti–Bcl-2 familyantibodies (Epitomics); anti–phospho-p53, anti–polyADP ribose polymerase (PARP), anti-cleaved caspase-3,anti-integrin b1 antibodies (Cell Signaling Technology);anti-p53, anti-Tie1, peroxidase-conjugated secondary anti-bodies (Santa Cruz Biotechnology); anti-Tie2 antibody andH-Gly-Arg-Gly-Asp-Ser-OH (GRGDS) peptide (Millipore);MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide; Sigma-Aldrich); fluorescein isothiocyanate–conjugated annexin V/propidium iodide assay kit (BDBiosciences); anti–neuron-glial antigen 2 (NG2), anti-integrin a3b1, anti-Ang2 antibodies (Abcam); and TUNELfluorescein kit (Roche). Anti-integrin a1 (clone FB12,

MAB1973), anti-integrin a3 (clone P1B5, MAB1952),anti-integrin b1 (clone 6S6, MAB2253), and a-integrinblocking and immunohistochemistry kit (a1-6, v) were pur-chased from Millipore and used for functional blocking.Small interfering (si)RNAs for p53 were purchased fromBioneer (Daejeon, Korea).

AnimalsAll animal experiments in this study were in strictagreement with the Association for Research in Visionand Ophthalmology Statement for the Use of Animalsin Ophthalmic and Vision Research and the guidelinesof the Seoul National University Animal Care and UseCommittee. Eight-week-old, pathogen-free male C57BL/6Jmice were purchased from Central Laboratory AnimalInc. After an 8-h fast, diabetes was induced by oneintraperitoneal injection of freshly prepared streptozotocin(STZ; Sigma-Aldrich) at a concentration of 180 mg/kgbody weight in 10 mmol/L citrate buffer (pH 4.5). Age-matched controls received citrate buffer only. Mice withblood glucose levels .300 mg/dL 4 days after STZ in-jection were deemed diabetic.

Diabetic and nondiabetic mice were killed 6 months afterdiabetes induction. Eyes were collected under deep anesthe-sia and immediately frozen at280°C for ELISA or were fixedin 4% paraformaldehyde for retinal digestion. Glucose levelsand body weight were monitored consecutively, and glycatedhemoglobin was determined before mice were killed.

Retinal Digest PreparationsVascular preparations of whole-mount retinas were per-formed using a trypsin digestion technique. Briefly, retinaswere fixed for at least 24 h in 4% paraformaldehyde,incubated in water for 1 h, and then digested in 2.5%trypsin (Gibco) at 37°C for 1 h. After careful removal of theinner limiting membrane, the retinal vessels were isolatedby careful irrigation with filtered water. The retinal digestsamples were dried and stained with periodic acid Schiffbase for 15 min and hematoxylin.

Morphological Quantification of Pericytes andAcellular CapillariesTo determine numbers of retinal pericytes and acellularcapillary, retinal digest preparations (n = 7–8) were ana-lyzed. Pericytes were identified according to the morphol-ogy and relative location to capillaries. Total pericyteswere counted in 10 randomly selected areas (original mag-nification 3400) in the middle one-third of the retinalcapillary area. The number of pericytes and acellular capil-laries were standardized to the capillary area (numbers ofcells or acellular capillaries/mm2 capillary area). The cap-illary area was calculated using the NIS-Elements AR 3.2program (Nikon, Tokyo, Japan). Samples were evaluatedin a masked fashion.

Intravitreal Injection of Ang2 and Anti-IntegrinAntibodiesNormal 6-week-old male mice were used. An intravitrealinjection was performed under deep anesthesia with 1 mL

3058 Ang2 Induces Pericyte Loss via a3b1 Integrin Diabetes Volume 63, September 2014

containing 100 ng Ang2 and 500 ng anti-integrin a3(clone P1B5) or anti-integrin b1 (clone 6S6). Sterile PBSwas injected for control. After 10 days, retinas underwentdigestion preparation.

Transferase-Mediated TUNEL Assay andImmunofluorescenceRetinal digestion preparations were incubated with anti-rabbit NG2 antibody (1:100) and a TUNEL fluorescein kit.Nuclei were counterstained with DAPI. TUNEL and NG2double-positive cells were evaluated with a fluorescencemicroscope (Nikon).

Cell Viability AssayIn all experiments, 2.5 3 104 cells were seeded into96-well plates. After 24 h, cells were treated with Ang1(300 ng/mL), Ang2 (300 ng/mL) under normal glucose(5 mmol/L glucose), high glucose (25 mmol/L glucose),and high mannitol (5 mmol/L glucose and 20 mmol/Lmannitol) as an osmotic control for 48 h. Cell viabilitywas determined by MTT assay according to the manufac-turer’s instructions. Three independent experiments wereperformed for each experimental condition.

FACS AnalysisTo evaluate apoptosis, 5 3 105 cells were treated withAng2 (300 ng/mL) under normal glucose, high mannitol,and high glucose at 37°C for 48 h. To determine the effectof integrin blocking, cells were treated with anti–integrinblocking antibodies (5 mg/mL) 1 h before the addition ofAng2. The cells were harvested and washed twice in PBS.Cells were stained with fluorescein isothiocyanate annexin-Vand propidium iodide (PI) for 15 min and analyzed by flowcytometry. Annexin V–positive/PI-negative cells were de-termined to be apoptotic.

To evaluate Tie-2 expression, pericytes and HUVECs(1 3 106) were suspended in the complete media (25 mL)for each experiment. PE-conjugated anti-Tie2 antibody wasadded to the each sample at 4°C for 1 h. PE-conjugatedmouse IgG antibody was used as a control.

Quantitative RT-PCRAll RNA was collected and isolated from cells using theRNeasy Plus Mini kit (Qiagen). cDNAs were preparedfrom RNAs (1 mg) using 2.5 mmol/L oligo-dT primers,1 mmol/L deoxyribonucleotide triphosphates, and murineleukemia virus RT. Quantitative (q)PCR assays were per-formed in qPCR Master Mix for SYBR Green PCR MasterMix (Applied Biosystems) using 7900HT real-time PCR(Applied Biosystems). Reaction conditions were 50 cyclesof 95°C for 5 s and 60°C for 20 s for qPCR. Quantitativereal-time PCR was performed using the following primers:Ang2 (forward: 59-ACTGTGTCCTCTTCCACCAC-39andreverse: 59-GGATGTTTAGGGTCTTGCTTT-39); Tie-2 (for-ward: 59-GCTTGCTCCTTTCTGGAACTGT-39 and reverse:59-CGCCACCCAGAGGCAAT-39) (20); Tie-1 (forward: 59-AGAACCTAGCCTCCAAGATT-39 and reverse: 59-ACTGTAGTTCAGGGACTCAA-39); ITGA1 (forward: 59-GGTTCCTACTTTGGCAGTATT-39 and reverse: 59- AACCTTGTCTGA

TTGAGAGCA-39); ITGA3 (forward: 59-AAGGGACCTTCAGGTGCA-39 and reverse: 59-TGTAGCCGGTGATTTACCAT-39);ITGB1 (forward: 59-GAAGGGTTGCCCTCCAGA-39 andreverse: 59-GCTTGAGCTTCTCTGCTGTT-39); and b-actin(forward: 59-GCCGCCAGCTCACCAT-39 and reverse: 59-TCGATGGGGTACTTCAGGGT-39). A mean quantity wascalculated from triplicate qPCR for each sample and nor-malized to the control gene.

Different primers were used for Tie-2 RT-PCR (forward:59-TGTTCCTGTGCCACAGGCTG-39 and reverse: 59-CACTGTCCCATCCGGCTTCA-39). PCR products were separatedon 1% agarose gels and visualized using SYBR Safe DNAGel Stain (Invitrogen) under ultraviolet transillumination.

Immunoprecipitation and ImmunoblottingFor immunoprecipitation, Ang2 (500 ng), integrin a3b1(500 ng), and anti-integrin a3b1 antibodies (2 mg) wereincubated with G-Sepharose beads at 4°C overnight (19).Immune complexes were collected by centrifugation andwashed with buffer three times. For immunoblotting, cellswere harvested and lysed in radioimmunoprecipitation as-say buffer with a protease inhibitor cocktail. Protein lysateswere resolved by SDS-PAGE and transferred onto nitrocel-lulose membrane. The membranes were incubated withprimary antibodies (1:1,000) at 4°C overnight and second-ary antibodies (1:5,000) at room temperature for 1 h. Themembranes were incubated with enhanced chemilumines-cent substrate (Pierce) and exposed to film.

Statistical AnalysisStatistical analyses were performed using the standardtwo-tailed Student t test assuming unequal variances, andP , 0.01 was considered statistically significant. Quanti-tative data are given as mean 6 SD. Data in figures aredepicted as mean 6 SE.

RESULTS

Retinal Capillary Pericyte Numbers Are Decreased inDiabetic Mice RetinasThe numbers of pericytes and acellular capillaries in retinaldigestion were compared between mice with 6-month STZ-induced diabetes and age-matched nondiabetic control mice(Fig. 1A). On one hand, the number of retinal capillarypericytes was significantly decreased in the STZ-induceddiabetic mice retinas (932.6 6 70.3) compared with thatof the control group (1,341.8 6 55.9, P , 0.001; Fig. 1B).On the other hand, the number of retinal acellular capil-laries was significantly increased in the STZ-induced dia-betic mice retinas (93.0 6 22.5) compared with that of thecontrol group (28.1 6 11.5, P , 0.001; Fig. 1C). Table 1reports the metabolic and physical parameters of the ex-perimental groups.

Ang2 Is Increased in Diabetic Retinas, and the Sourceof Ang2 Could Be Endothelial CellsThe effect of hyperglycemia on Ang2 expression in STZ-induced diabetic retinas was evaluated by qRT-PCR. Indiabetic retinas at 6 months, Angpt2 mRNA increased

diabetes.diabetesjournals.org Park and Associates 3059

1.87-fold compared with nondiabetic normal retinas (P =0.004; Fig. 2A). Also, Angpt1 mRNA increased 2.54-fold(P = 0.004) and Vegfa mRNA increased 1.6-fold (P =0.027; Supplementary Fig. 1A and B).

Next, to determine the source of Ang2 in the diabeticretinas, we examined the effects of high glucose on invitro ANGPT2 mRNA transcription by qRT-PCR in threemajor components of the BRB: HRMECs, pericytes, andastrocytes. On one hand, high glucose increased ANGPT2mRNA level in HRMECs more than 1.5-fold (1.52 6 0.09,P = 0.003) compared with normal glucose (Fig. 2B). Onthe other hand, high mannitol, an osmotic control, didnot increase ANGPT2 mRNA (1.21 6 0.13, P = 0.094) inHRMECs (Fig. 2B). Also, high glucose did not increaseANGPT2 mRNA in pericytes (1.13 6 0.22, P = 0.418)and astrocytes (1.13 6 0.22, P = 0.418; Fig. 2B and C).These data demonstrate that Ang2 increases in the dia-betic retina and that the source of Ang2 increase is retinalmicrovascular endothelial cells.

Ang2 Plays Synergistic Role in Pericyte ApoptosisUnder High-Glucose ConditionsWe determined the effect of Ang2 on the cell viability andapoptosis of pericytes under high glucose. Ang2 alone didnot affect cell viability in pericytes (97.6 6 3.6%, P =0.221). High glucose reduced cell viability in pericytes(89.4 6 7.9%, P = 0.020; Fig. 3A). Interestingly, Ang2

Figure 1—The number of retinal capillary pericytes is decreased in the diabetic mouse retina. Pericytes were identified in retinal digestpreparations by morphologic criteria (shape, staining intensity, and relative position in the capillary) and quantitated in 6-month STZ-induced diabetic mice (DM) and age-matched controls (Con). A: Representative examples of periodic acid Schiff– and hematoxylin-stainedretinal digest preparations of nondiabetic and diabetic mice after 6 months of diabetes are shown. The arrows indicate pericytes, and thearrowheads indicate acellular capillaries (original magnification 3400; scale bar = 20 mm). The number of retinal pericytes (B) and acellularcapillaries (C) are shown normalized to the area of capillaries (mm2) in which they were counted (n = 8 mice in each group). The bar graphrepresents mean 6 SE. *P < 0.01 by Student t test.

Table 1—Metabolic and physical parameters of theSTZ-induced diabetic and age-matched nondiabeticcontrol mice at 6 months

Nondiabetic Diabetic P value

Body weight (g) 32.46 6 3.22 18.97 6 0.69 ,0.001

Blood glucose(mmol/L) 10.11 6 1.48 32.83 6 1.34 ,0.001

HbA1c (%) 5.33 6 0.88 8.90 6 1.10 ,0.001

HbA1c (mmol/mol) 34.75 6 9.74 73.75 6 11.95

Data are presented as mean 6 SD.

3060 Ang2 Induces Pericyte Loss via a3b1 Integrin Diabetes Volume 63, September 2014

aggravated cell death under high glucose in pericytes(72.4 6 2.9%, P = 0.002; Fig. 3A). Next, pericyte apoptosiswas assessed by annexin-V/PI flow cytometric analysis.Of importance, the number of apoptotic pericytes was in-creased under high glucose (7.7 6 0.3%, P , 0.001) com-pared that of the control group (2.4 6 0.4%). Furthermore,Ang2 significantly aggravated high glucose–induced pericyteapoptosis (25.9 6 0.4%, P , 0.001), whereas Ang2 alonedid not induce apoptosis (2.7 6 0.1%, P = 0.298) undernormal glucose (Fig. 3B and Supplementary Fig. 2A).

Next, we determined the effect of Ang2 on the cellviability and apoptosis of HRMECs under high-glucoseconditions. Ang1 and Ang2 increased cell viability inHRMECs under high glucose (114.4 6 8.8% [P = 0.014]and 109.5 6 2.3% [P = 0.040], respectively; Fig. 3C). Asexpected, Ang2 significantly decreased the apoptotic cellpopulation in HRMECs (Fig. 3D, and Supplementary Fig.2B). These data suggested that Ang2 plays a synergisticrole in pericyte apoptosis under a high-glucose conditionand has a protective effect on the endothelial cell.

Ang2 Induces Pericyte Apoptosis via the p53 PathwayUnder High-Glucose ConditionsWestern blot studies confirmed that Ang2 induced theapoptosis pathway with an increase of Bax, cleaved PARP,and cleaved caspase-3 under high glucose but not under

normal glucose (Fig. 4A). Next, we aimed to identify themechanism that mediates Ang2-induced pericyte apopto-sis under high glucose. We found that Ang2 induced p53phosphorylation (Fig. 4B) and, subsequently, p53 accumu-lation (Fig. 4C) under high glucose. Then, to determinethe role of the p53 pathway for the observed Ang2-mediated pericyte apoptosis, we treated pericytes withcontrol siRNA or two different p53 siRNAs. The p53siRNAs effectively downregulated p53 expression (Fig. 4D)and also attenuated Ang2-induced pericyte apoptosis underhigh glucose (Fig. 4E). Interestingly, Ang2 phosphorylatedextracellular signal–related kinase (ERK), but not Akt(Fig. 4F). This ERK phosphorylation by Ang2 wasinhibited by PD98059 (ERK inhibitor), and the ERKinhibitor attenuated Ang2-induced p53 phosphoryla-tion (Fig. 4G). These data suggest that Ang2 inducespericyte apoptosis via the p53 pathway under high-glucose conditions.

Integrin, Not Tie-2, Is Important for Ang2-InducedPericyte Apoptosis Under High Glucose As an Ang2ReceptorTo determine whether the Tie-2 receptor is related withAng2-induced pericyte apoptosis, Western blot analysis(Fig. 5A) and RT-PCR (Fig. 5B) for Tie-2 and Tie-1 wereperformed on lysates obtained from HUVECs and pericytes.

Figure 2—Ang2 is increased in diabetic retinas, and the source of Ang2 could be endothelial cells. A: Angpt2mRNA levels were determinedin 6-month STZ-induced diabetic mice (DM) retinas compared with age-matched control (Con) mice by qRT-PCR and normalized to Rn18smRNA. Angpt2 expression increased in 6-month STZ-induced DM retinas (n = 6 mice in each group). ANGPT2 mRNA transcription isinduced by high glucose (HG; 25 mmol/L) in HRMECs (B) but not in pericytes (C) or astrocytes (D). HRMEC, pericytes, and astrocytes wereincubated for 48 h under 25 mmol/L HG, 20 mmol/L mannitol plus 5 mmol/L glucose (HM), or 5 mmol/L glucose (NG) as an osmotic control.ANGPT2 mRNA transcription was assessed by qRT-PCR, with actin as an internal control. ANGPT2 mRNA levels were normalized toACTIN mRNA and reported as fold-induction compared with cells exposed to 5 mmol/L glucose. Bar graph represents mean 6 SE. *P <0.01, #P > 0.05 by Student t test.

diabetes.diabetesjournals.org Park and Associates 3061

Interestingly, pericytes did not express Tie-2 or Tie-1,whereas HUVECs expressed Tie-2 and Tie-1. These datawere confirmed by qRT-PCR. TIE2 mRNA and TIE1 mRNAlevels were significantly lower in pericytes than in HUVECs(Fig. 5C). Compared with HUVECs with Tie-2 expression,pericytes did not express Tie-2 in FACS analysis (Fig. 5D).

To test whether integrins may serve as receptors forAng2 in pericyte apoptosis under high glucose, weincubated pericytes under high glucose for 15 min withAng2 and Gly-Arg-Gly-Asp-Ser (GRGDS) peptides, whichcan inhibit integrins that bind the Arg-Gly-Asp (RGD)sequence (21). GRGDS (0.5 mg/mL) attenuated Ang2-induced p53 phosphorylation (Fig. 5E). This result showedthat integrin signaling is involved in Ang2-inducedp53 phosphorylation. Of importance, these data sug-gested that integrin, not Tie-2, is important for Ang2-induced pericyte apoptosis under high glucose as an Ang2receptor.

High Glucose Increases Integrin a1, a3, and b1 inPericytesAs shown in Fig. 3C and Fig. 5E, Ang2 induced pericyteapoptosis via the integrin receptor under high glucose butnot under normal glucose. We hypothesized that highglucose preconditioned pericytes susceptible to Ang2 by

changing the integrin pattern. We screened the integrin asubunits (a1–6 and av) to determine which integrin sub-unit is responsible for Ang2-induced pericyte apoptosis.Integrin a1 and a3–blocking antibodies attenuated Ang2-induced p53 phosphorylation (Fig. 6A).

Next, to determine whether high glucose changesthe integrin expression pattern, we performed qRT-PCRand Western blot studies for a1, a3, and b1. The choicewas made because integrin a1 and a3 can form the het-erodimer with only integrin b1 (22). High glucose in-creased mRNA levels at 24 and 48 h for ITGa1 (1.86-fold and 2.20-fold, P , 0.01; Fig. 6B), ITGa3 (1.50-foldand 1.83-fold, P , 0.01; Fig. 6C), and ITGb1 (1.50-foldand 1.43-fold, P , 0.01; Fig. 6D), respectively. In addi-tion, high glucose increased integrin a1, a3, and b1 ex-pression (Fig. 6E). However, the integrin a1 was rarelyexpressed compared with integrin a3. In this regard, weperformed a coimmunoprecipitation assay to show directbinding of Ang2 to integrin a3b1. Indeed, Ang2 directlybound to integrin a3b1 (Fig. 6F).

Integrin a3b1 Suppression Inhibits Ang2-InducedPericyte ApoptosisFrom the result of integrin expression in pericytes underhigh glucose, integrin a1b1 or a3b1 were supposed to be

Figure 3—Ang2 plays synergistic role in pericyte apoptosis under high glucose (HG; 25 mmol/L glucose). The effects of Ang2 on the cellviability and apoptosis of pericytes and HRMECs under HG conditions were determined. Pericytes and HRMECs were incubated for 48 hwith and without Ang1 (300 ng/mL) or Ang2 (300 ng/mL) under HG and compared with control (Con). A and C: Cell viability was assessed byMTT assay. Ang2 induced cell death under HG in pericytes (A) but not in HRMECs (C). Pericytes (B) and HRMECs (D) were stained withannexin-V fluorescein isothiocyanate and PI and analyzed by flow cytometry. Cell apoptosis was expressed as the percentage of apoptoticcells in total cell populations. B: HG induced pericyte apoptosis, and Ang2 aggravated that apoptosis. D: Ang2 showed a protective effecton HRMEC apoptosis. The bar graph represents the mean 6 SE of three independent experiments. HM, high mannitol (5 mmol/L glucoseand 20 mmol/L mannitol); NG, normal glucose (5 mmol/L glucose). *P < 0.01 by Student t test.

3062 Ang2 Induces Pericyte Loss via a3b1 Integrin Diabetes Volume 63, September 2014

the possible receptor of Ang2. To determine whetherAng2 induced pericyte apoptosis through integrin a1b1or a3b1, pericytes were incubated under 25 mmol/L highglucose with Ang2 and anti-integrin a1, a3, and b1 anti-bodies for 48 h. Interestingly, Ang2-induced pericyteapoptosis under high glucose was attenuated by anti–integrin a3 and b1 antibodies but not by the anti–

integrin a1 antibody (Fig. 7A, and Supplementary Fig.3A). These results were confirmed for p53 expression onWestern blot analysis (Fig. 7B). Ang2 significantly in-creased p53 expression under high glucose (2.7 6 0.2,P , 0.01) but was significantly attenuated by anti–integrin a1, a3, and b1 antibodies (1.5 6 0.2, 1.0 60.1, and 0.7 6 0.1, respectively; P , 0.01).

Figure 4—Ang2 induces pericyte apoptosis via the p53 pathway under high glucose (HG; 25 mmol/L glucose). A: Western blot analysis forBax, Bcl-2, Bcl-xL, cleaved caspase-3, and cleaved PARP were performed on lysates obtained from pericytes treated with Ang2 (300 ng/mL)under normal glucose (NG; 5 mmol/L glucose), high mannitol (HM; 5 mmol/L glucose and 20 mmol/L mannitol), and HG for 48 h comparedwith control (Con). B: Western blot analysis for phospho (p)-p53 (Ser15) was performed on lysates obtained from pericytes treated with Ang2for 15, 30, and 60 min under HG (25 mmol/L). C: Western blot analysis for p53 was performed on lysates obtained from pericytes treated withAng2 for 24 and 48 h under HG (25 mmol/L). D: After the pericyte transfection with control siRNA or p53 siRNA, Western blot analysis for p53was performed on cell lysates with Ang2 for 48 h under HG (25 mmol/L), with b-tubulin used as a loading control. Data represent threeindependent experiments. E: Apoptotic cell counts were assessed by FACS analysis 48 h after Ang2 treatment in siRNA-transfectedpericytes. The bar graph represents the mean 6 SE of three independent experiments. *P < 0.01 by Student t test. F: Ang2 was treatedfor 5, 15, 30, and 60 min under HG. Determination of p-ERK, ERK, p-Akt, and Akt was by Western blot. G: Pericytes were preincubated withWortmannin (1 mmol/L) or PD98059 (20 mmol/L) for 1 h and treated with Ang2 for 15 min under HG. Determination of p-p53, p53, p-ERK,ERK, p-Akt, and Akt was by Western blot, with b-tubulin used as a loading control.

diabetes.diabetesjournals.org Park and Associates 3063

On the basis of in vitro experiments, we next intra-vitreously injected normal mice with 100 ng Ang2 withand without 500 ng anti-a3 or 500 ng anti-b1 antibodies.Ten days after the intravitreal injection, the isolated ret-inas were digested with trypsin for pericyte evaluation(Fig. 7C). The in vivo intravitreal injection of Ang2 in-duced pericyte loss in the retina compared with that inPBS injected control mice (1,047 6 52 and 1,401 6 109cells/mm2, P , 0.001). Ang2-induced pericyte loss wassignificantly attenuated by anti-a3 or anti-b1 antibodies(1,354 6 148 and 1,380 6 66 cells/mm2, P , 0.001;Fig. 7D). In addition, Ang2-induced TUNEL and NG2double-positive pericytes were decreased in mice injectedwith anti-a3 or anti-b1 antibodies (Fig. 7E). These in vivodata suggest that Ang2 induced pericyte loss via anapoptotic mechanism in the retina. Overall, these datademonstrate that integrin a3 and b1 are important forAng2-induced pericyte apoptosis via the p53 pathway.

DISCUSSION

In this study, we demonstrated that pericyte loss occurredwith Ang2 increase in diabetic mice retinas (10,11,13).Previously, Tie-2 was known to be related with Ang-

induced pericyte survival and recruitment (11); however,there was no direct evidence that Ang2 induced pericyteapoptosis via Tie-2. Unlike the previous study (11), peri-cytes did not express Tie-2 mRNA or protein in our study.Although Ang2 alone did not induce apoptosis in vitro(11), Ang2 induced pericyte apoptosis under high glucosevia the p53 pathway (Fig. 4). Thus, we postulated thatAng2 would induce Tie-2–independent apoptosis in peri-cytes. Accumulating evidence supports the role of integrinin Ang2 activity in endothelial cells (19); however, little isknown about the Ang2/integrin system in the pericyte.We thus focused on the integrin receptors on pericytes.

Integrins are cell adhesion molecules that are expressedon the surface of endothelial cells and pericytes. Thecontribution of endothelial and mural cell integrins toangiogenesis has been studied. Pericytes express integ-rins, including the collagen receptors a1b1 and a2b1;the laminin receptors a3b1, a6b1, a6b4, and a7b1; thefibronectin receptors a4b1 and a5b1; and the osteoponinreceptors a8b1 and a9b1 (22). Cytokines and extracellularmatrix change integrin subtypes in pericytes (23). On thebasis of the differential response to Ang2 in the pericytesdepending on the glucose concentration (Fig. 3B and 4A),

Figure 5—Integrin, not Tie-2, is important for Ang2-induced pericyte apoptosis under high glucose as an Ang2 receptor. Western blot (A)and RT-PCR (B) for Tie-2 and Tie-1 expression were performed on lysates obtained from HUVECs and pericytes. b-Tubulin was used as aninternal control. C: TIE2 and TIE1 mRNA transcriptions were assessed by qRT-PCR. Actin was used as an internal control. TIE2 and TIE1mRNA levels were normalized to ACTIN mRNA and reported as fold induction compared with HUVECs. *P < 0.01 by Student t test. D:HUVECs and pericytes were analyzed by flow cytometry for Tie-2 expression. E: Western blot analysis for phospho (p)-p53 (ser15) and p53were performed on lysates obtained from pericytes treated with Ang2 (300 ng/mL) or GRGDS (0.5 mg/mL) under 25 mmol/L high glucose(HG) for 15 min. b-Tubulin was used as a loading control. Data represent three independent experiments. Con, control.

3064 Ang2 Induces Pericyte Loss via a3b1 Integrin Diabetes Volume 63, September 2014

we hypothesized that high glucose could change the integ-rin subtype more susceptible to Ang2. During qRT-PCRarray for screening the integrin subtype change responseto high glucose, only ITGa1, ITGa2, ITGa3, and ITGb1mRNA increased more than 1.5-fold among ITGa1,ITGa2, ITGa3, ITGa4, ITGa5, ITGa6, ITGa7, ITGav, andITGb1 mRNA (data not shown). In addition, basal ITGa3

mRNA was more than six times higher than ITGa1 orITGa2 mRNA in pericytes even under normal glucose con-ditions. Thus, we concluded that integrin a3 and b1 areboth predominant and high-glucose inducible integrins inpericytes. The proportion of integrin a1 in pericytes is toominor to be responsible for Ang2-induced pericyte apopto-sis (Fig. 6E and 7B). In addition, many integrins, including

Figure 6—High glucose (HG; 25 mmol/L) increases integrin a1, a3, and b1 in pericytes. A: Western blot analysis for phospho (p)-p53(Ser15) and p53 were performed on lysates obtained from pericytes treated with Ang2 (300 ng/mL) or various integrin-blocking antibodies(5 mg/mL, a1, a2, a3, a4, a5, a6, and av) under HG for 15 min. b-Tubulin was used as a loading control. Data represent three independentexperiments. B–D: Pericyte were incubated under HG for 24 and 48 h. ITGa1 (B), ITGa3 (C), and ITGb1 (D) mRNA transcriptions wereassessed by qRT-PCR. Actin was used as an internal control. ITGa1, ITGa3, and ITGb1 mRNA levels were normalized to actin mRNA andreported as fold induction compared with control. *P < 0.01 by Student t test. E: Western blot analyses for integrin a1, a3, and b1 wereperformed on lysates obtained from pericytes incubated under HG and normal glucose (NG; 5 mmol/L) for 48 h. b-Tubulin was used asa loading control. Data represent three independent experiments. F: After incubation of Ang2 (500 ng), integrin a3b1 (500 ng), and integrina3b1 antibody, immune complexes were coimmunoprecipitated to show direct binding of Ang2 to integrin a3b1 and then were analyzed forAng2 and integrin b1. The same amounts of recombinant Ang2 and integrin a3b1 were used as an input.

diabetes.diabetesjournals.org Park and Associates 3065

Figure 7—Ang2-induced pericyte apoptosis is inhibited by suppression of integrin a3b1. A and B: Pericytes were incubated under 25mmol/L high glucose (HG) with Ang2 (300 ng/mL) and anti-integrin a1, a3, and b1 antibodies (Ab; 5 mg/mL) for 48 h. A: Pericyte apoptosiswas analyzed by FACS and expressed as the percentage of apoptotic cells in total cell populations. B: Western blot analysis for p53 wasperformed on lysates obtained from pericyte. b-Tubulin was used as a loading control. Data represent three independent experiments.(ImageJ quantitation; n = 3, mean 6 SD). *P < 0.01 compared with Ang2 treatment by two-tailed Student t test. C: Ang2 (100 ng), with andwithout anti-a3 or anti-b1 antibodies (500 ng), was intravitreously injected to normal mice (n = 8 mice in each group). Eyes were enucleated10 days after the intravitreal injection, and the isolated retinas were digested with trypsin for pericyte evaluation. Pericytes were identified inperiodic acid Schiff– and hematoxylin-stained retinal digest preparations by morphologic criteria. Representative examples of retinal digestpreparations are shown. Arrows indicate representative pericytes. D: The number of pericytes was normalized to the area of capillaries(mm2) in which they were counted. *P < 0.01 by Student t test. E: Retinal digest preparations were immunostained with NG2 (red), TUNEL(green), and DAPI (blue). White arrows indicate TUNEL-positive pericytes. C and E: Original magnification 3400. Scale bar = 20 mm. Bargraph represents mean 6 SE.

3066 Ang2 Induces Pericyte Loss via a3b1 Integrin Diabetes Volume 63, September 2014

avb3, a5b1, aIIbb3, avb6, and a3b1, recognize the tri-peptide RGD in their ligands (24). However, integrin a1b1recognizes a configuration of residues formed by arginineand aspartic acid residues, not RGD. Attenuation of p53 byGRGDS supported that Ang2-induced pericyte apoptosiswas effectively attenuated by blocking integrin a3b1 (Fig.5E and 7B). Integrin receptors induced by hyperglycemia inDR could make the pericyte more susceptible to the Ang2,and in turn, lead to the pericyte apoptosis by the Ang2/integrin signaling pathway.

The laminin receptor a3b1 integrin is expressed invascular endothelial cells and acts to suppress pathologicalangiogenesis in the retina and other organs (25,26). How-ever, a role for integrin a3b1 in the pericyte is yet to bedetermined (22). Inhibition of b1 integrin induceda rounded morphology of the pericytes, suggesting peri-cyte adhesive properties were affected or that these cellswere undergoing apoptosis (27). In this study, we sug-gested that integrin a3b1 was important for Ang2-induced apoptosis. This is notably different from theclassical concept that Tie-2 mediates the effect of Ang2as an Ang2 receptor. Absence of Tie-2 in the human peri-cytes used in this study was confirmed by various experi-ments, including RT-PCR with two different primers,Western blot, and FACS analysis (Fig. 5A–D). In addition,Ang1 and Ang2 can both directly bind to some integrinsand thereby signal in the absence of Tie-2 (28–31). Thus,we suggested that integrin mediated Ang2-induced apo-ptosis at least in Tie-2–negative pericytes.

Previous studies have shown that vascular endothelialcells are the primary source of Ang2 (32–35). High glucoseincreases Ang2 mRNA in HRMECs (36). The endothelialcell is regarded as potential source of Ang2 in the retinaby transient rapid release from the Weibel-Palade bodyand chronic upregulation of Ang2 (32,37,38). Chronic hy-perglycemia upregulates Ang2 in the diabetic retina(36,39), and Ang2 upregulation is causally involved inthe pathogenesis of pericyte loss in DR (10). Hyperglyce-mia induces pericyte loss not only by increasing Ang2 inendothelial cells but also by changing the integrin subtypeprone to Ang2 in pericytes. Our data support that glyce-mic control is important to prevent Ang2-induced peri-cyte loss in early DR. Glycemic control is the effectivetreatment to reduce the progression of DR (40–42).

We showed that Ang2 induced pericyte apoptosisunder high glucose via the p53 pathway. Furthermore,Ang2-induced pericyte apoptosis was effectively attenu-ated in vitro and in vivo by blocking integrin a3b1. Integ-rin a3 and b1 blockers reduced p53 signaling by Ang2.Ser-15 of p53 is phosphorylated by a mitogen-activatedprotein kinase (ERK1/2)-dependent pathway, and thisstep is required for apoptosis to occur (43).

Others have suggested that hyperglycemia causespericyte loss by apoptosis (6–8). In this study, we postu-lated that hyperglycemia-induced Ang2 overexpression inendothelial cells, in turn, caused apoptotic cell death ofpericytes. In an animal model of DR, high glucose was

induced by STZ injection as early as 3 days. Then, Ang2had increased at 3 months with pericyte loss, and finally,acellular capillary at 6 months (10,44). The mechanism bywhich capillaries become acellular is largely unknown.Apoptosis preceding the formation of acellular capillar-ies in retinas from diabetic rats and humans is onemechanism by which endothelial cells could be elimi-nated from the diabetic capillary (8). Retinal capillarycoverage with pericytes is crucial for the survival of en-dothelial cells. Loss of pericytes is related to the increaseof acellular capillaries from hyperglycemic injury (5). Ourdata indicate Ang2 did not directly induce endothelialcell death (Fig. 3C and D); thus, we postulated that anacellular capillary could cause endothelial cell death sec-ondary to pericyte loss by Ang2.

In conclusion, we demonstrated that pericyte lossoccurred with Ang2 increase in the diabetic mice retinas.Ang2 induced pericyte apoptosis via the p53 pathwayunder high glucose. High glucose increased integrin a3b1in pericytes. Interestingly, integrin was involved in Ang2-induced pericyte apoptosis. Furthermore, Ang2-inducedpericyte apoptosis was effectively attenuated by blockingintegrin a3b1 both in vitro and in vivo. Taken together,Ang2 induced p53-dependent pericyte apoptosis via a3b1integrin signaling in DR. We suggest that glycemic controlor blocking Ang2/integrin signaling could be a potentialtherapeutic target to prevent pericyte loss in early DR.

Funding. This study was supported by the Seoul National University ResearchGrant (800-20130338 to Je.H.K.), the Seoul National University Hospital ResearchFund (03-2013-0070 to Je.H.K.), the Pioneer Research Center Program of theNational Research Foundation for the Ministry of Education, Science and Tech-nology (NRF/MEST; 2012-0009544 to Je.H.K.), the Bio-Signal Analysis Technol-ogy Innovation Program of NRF/MEST (2009-0090895 to Je.H.K.), the GlobalResearch Laboratory Program of NRF/MEST (2011-0021874 to K.-W.K.) and theNRF of Korea grant funded by the Korea government (2012R1A2A2A01012400,2011-0030739 to C.-H.C. and 2012-0006019 to Je.H.K.).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. S.W.P. and J.-H.Y. performed the experimentsand wrote the manuscript. Ji.H.K. and K.-W.K. analyzed the data and reviewedthe manuscript. C.-H.C. and Je.H.K. designed the study and critically revised themanuscript. Je.H.K. is the guarantor of this work, and, as such, had full access toall the data in the study and takes responsibility for the integrity of the data andthe accuracy of the data analysis.

References1. Antonetti DA, Klein R, Gardner TW. Diabetic retinopathy. N Engl J Med 2012;366:1227–12392. Moss SE, Klein R, Klein BE. The incidence of vision loss in a diabeticpopulation. Ophthalmology 1988;95:1340–13483. Cogan DG, Toussaint D, Kuwabara T. Retinal vascular patterns. IV. Diabeticretinopathy. Arch Ophthalmol 1961;66:366–3784. Shepro D, Morel NM. Pericyte physiology. FASEB J 1993;7:1031–10385. Hammes HP, Lin J, Renner O, et al. Pericytes and the pathogenesis ofdiabetic retinopathy. Diabetes 2002;51:3107–31126. Romeo G, Liu WH, Asnaghi V, Kern TS, Lorenzi M. Activation of nuclearfactor-kappaB induced by diabetes and high glucose regulates a proapoptoticprogram in retinal pericytes. Diabetes 2002;51:2241–2248

diabetes.diabetesjournals.org Park and Associates 3067

7. Miller AG, Smith DG, Bhat M, Nagaraj RH. Glyoxalase I is critical for humanretinal capillary pericyte survival under hyperglycemic conditions. J Biol Chem2006;281:11864–118718. Mizutani M, Kern TS, Lorenzi M. Accelerated death of retinal microvascularcells in human and experimental diabetic retinopathy. J Clin Invest 1996;97:2883–28909. Watanabe D, Suzuma K, Suzuma I, et al. Vitreous levels of angiopoietin 2and vascular endothelial growth factor in patients with proliferative diabeticretinopathy. Am J Ophthalmol 2005;139:476–48110. Hammes HP, Lin J, Wagner P, et al. Angiopoietin-2 causes pericyte dropoutin the normal retina: evidence for involvement in diabetic retinopathy. Diabetes2004;53:1104–111011. Cai J, Kehoe O, Smith GM, Hykin P, Boulton ME. The angiopoietin/Tie-2system regulates pericyte survival and recruitment in diabetic retinopathy. InvestOphthalmol Vis Sci 2008;49:2163–217112. Yao D, Taguchi T, Matsumura T, et al. High glucose increases angiopoietin-2transcription in microvascular endothelial cells through methylglyoxal modificationof mSin3A. J Biol Chem 2007;282:31038–3104513. Pfister F, Feng Y, vom Hagen F, et al. Pericyte migration: a novel mecha-nism of pericyte loss in experimental diabetic retinopathy. Diabetes 2008;57:2495–250214. Pfister F, Wang Y, Schreiter K, et al. Retinal overexpression of angiopoietin-2 mimics diabetic retinopathy and enhances vascular damages in hyperglycemia.Acta Diabetol 2010;47:59–6415. Feng Y, vom Hagen F, Pfister F, et al. Impaired pericyte recruitment andabnormal retinal angiogenesis as a result of angiopoietin-2 overexpression.Thromb Haemost 2007;97:99–10816. Maisonpierre PC, Suri C, Jones PF, et al. Angiopoietin-2, a natural antag-onist for Tie2 that disrupts in vivo angiogenesis. Science 1997;277:55–6017. Augustin HG, Koh GY, Thurston G, Alitalo K. Control of vascular morpho-genesis and homeostasis through the angiopoietin-Tie system. Nat Rev Mol CellBiol 2009;10:165–17718. Hosaka K, Yang Y, Seki T, et al. Tumour PDGF-BB expression levels de-termine dual effects of anti-PDGF drugs on vascular remodelling and metastasis.Nat Commun 2013;4:212919. Felcht M, Luck R, Schering A, et al. Angiopoietin-2 differentially regulatesangiogenesis through TIE2 and integrin signaling. J Clin Invest 2012;122:1991–200520. Rabascio C, Muratori E, Mancuso P, et al. Assessing tumor angiogenesis:increased circulating VE-cadherin RNA in patients with cancer indicates viabilityof circulating endothelial cells. Cancer Res 2004;64:4373–437721. Ruoslahti E. RGD and other recognition sequences for integrins. Annu RevCell Dev Biol 1996;12:697–71522. Silva R, D’Amico G, Hodivala-Dilke KM, Reynolds LE. Integrins: the keys tounlocking angiogenesis. Arterioscler Thromb Vasc Biol 2008;28:1703–171323. Tigges U, Boroujerdi A, Welser-Alves JV, Milner R. TNF-a promotes cerebralpericyte remodeling in vitro, via a switch from a1 to a2 integrins. J Neuro-inflammation 2013;10:3324. Plow EF, Haas TA, Zhang L, Loftus J, Smith JW. Ligand binding to integrins.J Biol Chem 2000;275:21785–2178825. Watson AR, Pitchford SC, Reynolds LE, et al. Deficiency of bone marrowbeta3-integrin enhances non-functional neovascularization. J Pathol 2010;220:435–44526. Lim Y, Jo DH, Kim JH, et al. Human apolipoprotein(a) kringle V inhibits ischemia-induced retinal neovascularization via suppression of fibronectin-mediated an-giogenesis. Diabetes 2012;61:1599–1608

27. Carnevale E, Fogel E, Aplin AC, et al. Regulation of postangiogenic neo-vessel survival by beta1 and beta3 integrins in collagen and fibrin matrices.J Vasc Res 2007;44:40–5028. Carlson TR, Feng Y, Maisonpierre PC, Mrksich M, Morla AO. Direct celladhesion to the angiopoietins mediated by integrins. J Biol Chem 2001;276:26516–2652529. Hu B, Jarzynka MJ, Guo P, Imanishi Y, Schlaepfer DD, Cheng SY. Angio-poietin 2 induces glioma cell invasion by stimulating matrix metalloprotease 2expression through the alphavbeta1 integrin and focal adhesion kinase signalingpathway. Cancer Res 2006;66:775–78330. Bezuidenhout L, Zilla P, Davies N. Association of Ang-2 with integrin beta 2controls Ang-2/PDGF-BB-dependent upregulation of human peripheral bloodmonocyte fibrinolysis. Inflammation 2009;32:393–40131. Thomas M, Felcht M, Kruse K, et al. Angiopoietin-2 stimulation of endo-thelial cells induces alphavbeta3 integrin internalization and degradation. J BiolChem 2010;285:23842–2384932. Gale NW, Thurston G, Hackett SF, et al. Angiopoietin-2 is required forpostnatal angiogenesis and lymphatic patterning, and only the latter role isrescued by Angiopoietin-1. Dev Cell 2002;3:411–42333. Huang YQ, Li JJ, Hu L, Lee M, Karpatkin S. Thrombin induces increasedexpression and secretion of angiopoietin-2 from human umbilical vein endothelialcells. Blood 2002;99:1646–165034. Hackett SF, Wiegand S, Yancopoulos G, Campochiaro PA. Angiopoietin-2plays an important role in retinal angiogenesis. J Cell Physiol 2002;192:182–18735. Oh H, Takagi H, Suzuma K, Otani A, Matsumura M, Honda Y. Hypoxia andvascular endothelial growth factor selectively up-regulate angiopoietin-2 in bo-vine microvascular endothelial cells. J Biol Chem 1999;274:15732–1573936. Rangasamy S, Srinivasan R, Maestas J, McGuire PG, Das A. A potential rolefor angiopoietin 2 in the regulation of the blood-retinal barrier in diabetic reti-nopathy. Invest Ophthalmol Vis Sci 2011;52:3784–379137. Fiedler U, Scharpfenecker M, Koidl S, et al. The Tie-2 ligand angiopoietin-2is stored in and rapidly released upon stimulation from endothelial cell Weibel-Palade bodies. Blood 2004;103:4150–415638. Fiedler U, Reiss Y, Scharpfenecker M, et al. Angiopoietin-2 sensitizes en-dothelial cells to TNF-alpha and has a crucial role in the induction of in-flammation. Nat Med 2006;12:235–23939. Ohashi H, Takagi H, Koyama S, et al. Alterations in expression of angio-poietins and the Tie-2 receptor in the retina of streptozotocin induced diabeticrats. Mol Vis 2004;10:608–61740. Chew EY, Ambrosius WT, Davis MD, et al.; ACCORD Study Group; ACCORDEye Study Group. Effects of medical therapies on retinopathy progression in type2 diabetes. N Engl J Med 2010;363:233–24441. The Diabetes Control and Complications Trial Research Group. The effect ofintensive treatment of diabetes on the development and progression of long-termcomplications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977–98642. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucosecontrol with sulphonylureas or insulin compared with conventional treatment andrisk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352:837–85343. Shih A, Davis FB, Lin HY, Davis PJ. Resveratrol induces apoptosis in thyroidcancer cell lines via a MAPK- and p53-dependent mechanism. J Clin EndocrinolMetab 2002;87:1223–123244. Feit-Leichman RA, Kinouchi R, Takeda M, et al. Vascular damage ina mouse model of diabetic retinopathy: relation to neuronal and glial changes.Invest Ophthalmol Vis Sci 2005;46:4281–4287

3068 Ang2 Induces Pericyte Loss via a3b1 Integrin Diabetes Volume 63, September 2014