Kristina Jörgens,1,2 Sandra J. Stoll,1,2 Jennifer Pohl,1,3 Thomas H. Fleming,4 Carsten Sticht,5

Peter P. Nawroth,4 Hans-Peter Hammes,3 and Jens Kroll1,2

High Tissue Glucose AltersIntersomitic Blood Vessels inZebrafish via MethylglyoxalTargeting the VEGF ReceptorSignaling CascadeDiabetes 2015;64:213–225 | DOI: 10.2337/db14-0352

Hyperglycemia causes micro- and macrovascular com-plications in diabetic patients. Elevated glucose concen-trations lead to increased formation of the highly reactivedicarbonyl methylglyoxal (MG), yet the early consequencesof MG for development of vascular complications in vivoare poorly understood. In this study, zebrafish were usedas a model organism to analyze early vascular effectsand mechanisms of MG in vivo. High tissue glucoseincreased MG concentrations in tg(fli:EGFP) zebrafishembryos and rapidly induced several additional mal-formed and uncoordinated blood vessel structures thatoriginated out of existing intersomitic blood vessels(ISVs). However, larger blood vessels, including the dorsalaorta and common cardinal vein, were not affected. Ex-pression silencing of MG-degrading enzyme glyoxalase(glo) 1 elevated MG concentrations and induced a similarvascular hyperbranching phenotype in zebrafish. MG en-hanced phosphorylation of vascular endothelial growthfactor (VEGF) receptor 2 and its downstream target Akt/protein kinase B (PKB). Pharmacological inhibitors forVEGF receptor 2 and Akt/PKB as well as MG scavengeraminoguanidine and glo1 activation prevented MG-inducedhyperbranching of ISVs. Taken together, MG acts onsmaller blood vessels in zebrafish via the VEGF receptorsignaling cascade, thereby describing a new mechanism

that can explain vascular complications under hyperglyce-mia and elevated MG concentrations.

Diabetes is characterized by chronic hyperglycemia, eitherthrough insulin deficiency or insulin resistance (1). Morethan 30% of diabetic patients suffer from micro- andmacrovascular complications that represent the leadingcauses of death worldwide (2). Although chronic hyper-glycemia is favored as the principle causal factor of com-plications, even good glycemic control cannot reduce therisk of cardiovascular mortality and microvascular compli-cations in type 2 diabetic patients (3). Importantly, 36%of newly diagnosed type 2 diabetic patients already sufferfrom clinically discernible retinopathy (4), and some stud-ies indicate the presence of “diabetes” complications inpatients without detectable hyperglycemia (5). Likewise,intensive glycemic control can only account for a minorproportion of therapeutic effects following myocardial in-farction (6–9).

Formation of glucose metabolites—reactive dicarbonyls—can arise from several sources, including auto-oxidationof glucose, which leads to glyoxal formation, decompositionof amadori products (3-deoxyglucosone) and fragmentation

1Department of Vascular Biology and Tumor Angiogenesis, Centre for Biomedi-cine and Medical Technology Mannheim, Medical Faculty Mannheim, HeidelbergUniversity, Mannheim, Germany2Division of Vascular Oncology and Metastasis, German Cancer Research Center(DKFZ-ZMBH Alliance), Heidelberg, Germany35th Medical Department, Medical Faculty Mannheim, Heidelberg University,Mannheim, Germany4Department of Medicine and Clinical Chemistry, Heidelberg University, Mannheim,Germany5Medical Research Center, Medical Faculty Mannheim, Heidelberg University,Mannheim, Germany

Corresponding author: Jens Kroll, jens.kroll@medma.uni-heidelberg.de.

Received 3 March 2014 and accepted 28 July 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0352/-/DC1.

K.J. and S.J.S. contributed equally to this work.

© 2015 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 64, January 2015 213

COMPLIC

ATIO

NS

of glycerinaldehyde-3-phosphate and dihydroxyacetone phos-phate during glycolysis (methylglyoxal [MG]) (1). The mostreactive dicarbonyl is MG, which can interact in several pro-teins with amino acid side chains of arginine, lysine, andcysteine and thereby induce several posttranslational proteinmodifications and subsequently alter function of those pro-teins (10). MG is detectable in healthy subjects, but it isconstantly degraded by the glyoxalase (glo) system, whichis expressed in the cytosol of all mammalian cells (11). Theglo system is comprised of two enzymes, glo1 and glo2, andits cofactor glutathione. It catalyzes the conversion of MG toD-lactate via the intermediate S-D-lactoylglutathione (11).

MG is elevated in diabetic patients. In plasma of healthypeople, MG concentrations of ;100 nmol/L have beendetected, whereas in diabetic patients, MG levels increaseup to 700 nmol/L (12). A direct link between elevated MGconcentrations and diabetes complications has now beenidentified for the development of metabolic hyperalgesia(13). In this recent study, MG levels above 600 nmol/Lwere identified as the critical threshold for pain in diabeticpatients, and this is based on an increased sensitivitycaused by a direct MG-induced posttranslational modifica-tion of the voltage-gated sodium channel Nav1.8. The path-ological consequences of increased MG concentrations inendothelial cells and for the development of cardiovasculardiseases in diabetic patients are less clear. In vitro studies inimmortalized human microvascular endothelial cell line 1,human umbilical vein endothelial cells, and mouse kidneyendothelial cells showed that elevated concentrations ofglucose increased internal MG levels and that MG directlytargets vascular basement membrane type IV collagen (14),Hsp27 (15), and transcription factor mSin3A (16). In addi-tion to endothelial cells, MG acts on smooth muscle cells,thereby increasing arterial atherogenicity (17), blood pres-sure (18), and salt-sensitive hypertension (19) and delayingwound healing in aged mice (20). Together, these data pro-vide experimental evidence that MG contributes to the de-velopment of vascular cell dysfunction in vitro and tocardiovascular complications in vivo.

Zebrafish have been proven as an excellent modelorganism to study processes of vascular development andfunction, and several mechanisms characterized in zebra-fish share similar functions in mammals. Besides yeast(21), Caenorhabditis elegans (22), and drosophila (23),zebrafish have evolved as a model organism for investiga-tions of metabolic disorders (24), and hyperglycemia-induced pathologies in zebrafish are related to diabeticpatients (25–27).

Since MG elicits alterations on cultured endothelialcells but its effect on blood vessels in vivo is less clear, thisstudy was aimed at identifying early MG-induced patho-logical mechanisms on blood vessels in zebrafish. Highglucose increases MG in tg(fli:EGFP) zebrafish embryosand subsequently alters intersomitic blood vessels(ISVs), which is mediated by an increased activation ofvascular endothelial growth factor (VEGF) receptor 2and Akt/protein kinase B (PKB). This demonstrates

a role of MG in development of hyperglycemia-inducedvascular damages and provides a new in vivo mechanismof early vascular damages induced by MG.

RESEARCH DESIGN AND METHODS

Inhibitors, Antibodies, and ReagentsInhibitors, antibodies, and reagents used included MG solu-tion 40%, D-(+)-glucose, D-mannitol, Akt inhibitor XVIIISC66, p38MAP kinase inhibitor SB203580, PI3 kinase (PI3K)inhibitor LY294002, VEGF receptor 2 inhibitor PTK787/Vatalanib, and aminoguanidine (AMG), rabbit anti-VEGFreceptor 2 (phospho Y1054 + Y1059) antibody, rabbit anti-VEGF receptor 2, goat antiactin (I-19), rabbit anti-Akt,mouse anti-pAkt (S473, 587F11), anti-glo1 (6F10), andhorseradish peroxidase–conjugated antibodies rabbit anti-mouse, rabbit anti-rat, goat anti-rabbit, and rabbit anti-goat.

Zebrafish Husbandry and Zebrafish LinesEmbryos of AB wild-type, tg(fli:EGFP), and tg(nflk:EGFP)were raised and staged as described (28).

Incubation with MG, Glucose, and InhibitorsApproximately 80 fertilized eggs were incubated in a 10 cmpetri dish with 20 mL solutions that were changed daily.Solutions contained egg water, glucose, or MG and 0.2%1-phenyl-2-thiourea and inhibitors were diluted in DMSO.

Injections of Morpholinos, mRNA, and IntracardiacInjectionControl morpholinos and glo1 were diluted to 3 mg/mLwith 1.5 3 p53 morpholino to attenuate possible off-target effects (29) in 0.1 mol/L KCl. hGlo1 or mOrangemRNA were diluted to 100 pg/nL in 0.1 mol/L KCl. Onenanoliter of morpholino or mRNA was injected into theembryos at the one-cell stage. For intracardiac injection ofMG, embryos were injected at 48 hpf with 1 nL control orMG solution into the heart and hyperbranches were ana-lyzed at 96 hpf.

MorpholinosMorpholinos included SB-glo1-MO#1: 59-CATGAAGTCCTACAGGAGACAAATT-39 (targeting intron 1–exon 2 junction);SB-glo1-MO#2: 59-GCCTACAAATAAAACATCCACACAT-39(targeting intron 2–exon 3 junction); SB-pdx1-Mo: 59-GATAGTAATGCTCTTCCCGATTCAT-39 (30); p53-MO: 59-GCGCCATTGCTTTGCAAGAATTG-39; control-MO: 59-CCTCTTACCTCAGTTACAATTTATA-39.

Synthesis of mRNA was performed using the T7mMessage mMachine Kit following the protocol of themanufacturer using a plasmid coding human glo1 (kind giftof Dr. Michael Brownlee, Albert Einstein College of Medi-cine, Bronx, NY). Intracardiac injections at 48 hours postfertilization (hpf) were performed as recently described (31).

MG and Glucose MeasurementsMG was determined by derivatization with 1,2-diamino-4,5-dimethoxybenzene followed by high-performance liquidchromatography of the quinoxaline adduct (32). Intracellu-lar zebrafish glucose concentration was determined indeproteinated total extracts of whole zebrafish embryos

214 Methylglyoxal in Zebrafish Blood Vessels Diabetes Volume 64, January 2015

using the Glucose (HK) Assay Kit according to the man-ufacturer’s instructions (33).

Western Blot AnalysisZebrafish embryos were collected at 48 hpf, and Westernblot was performed as described previously (31).

RT-PCRWhole RNA was isolated as recently described (31). Primersincluded glo1 forward (CCGCGTGTAAAGAGGGAAGCT)and glo1 reverse (GGCAGCATAGACATCCGGTACT). Thefollowing program was used to amplify PCR products:95°C for 2 min (95°C for 30 s, 59°C for 45 s, 72°C for30 s) 3 35 and 72°C for 5 min.

Microscopy and Analysis of Vascular DefectsFor in vivo imaging, tg(fli:EGFP) or tg(nflk:EGFP) em-bryos were anesthetized with 0.003% tricaine and em-bedded in 1% low-melting-point agarose dissolved in eggwater containing 0.003% tricaine and 0.2% 1-phenyl-2-thiourea. Confocal images were performed usinga DMIRE2 confocal microscope with Leica TCS SP2True Confocal Scanner and a DM6000 B confocal micro-scope with Leica TCS SP5 DS scanner. Images were takenwith 400 Hz, 1,024 3 1,024 pixels, and z-stacks wererecorded with 1 mm thickness.

Quantification of Vascular Defects and StatisticalAnalysisFor quantification of altered blood vessels, the first 14 ISVpairs of each zebrafish embryo were analyzed for hyper-branching blood vessels at 96 hpf. A minimum of 100embryos per sample was analyzed. Results are given asmean number of hyperbranches per embryo 6 SE. Forevaluation of endothelial cell quantity, the number ofendothelial cells was counted for six ISV pairs in tg(nflk:EGFP) at 48 and 96 hpf. Results are given as mean num-ber of endothelial cells per 6 ISVs 6 SE. Diameter ofvessels is given as the average of measurements at threedifferent positions in the embryo. Statistical significancebetween different groups was analyzed using Student ttest or Mann-Whitney U test as indicated (SPSS 17.0).P values , 0.05 were considered significant.

RESULTS

High Tissue Glucose Induces Altered Blood Vessels inZebrafish via MGIn order to identify early consequences of elevated MG onblood vessels in vivo, we analyzed blood vessels in tg(fli:EGFP) zebrafish embryos in the presence of increasedglucose or MG concentrations. To this end, tg(fli:EGFP)zebrafish embryos were cultured for 96 h starting at 1 hpfin the presence of 15, 30, or 55 mmol/L glucose solution.Formation and maintenance of blood vessels in zebrafishembryos were analyzed at 24, 48, 72, and 96 hpf. Like-wise, tg(fli:EGFP) zebrafish embryos were incubated in200 or 500 mmol/L MG solution and subsequently ana-lyzed at the indicated time points. Incubation of zebrafishembryos with glucose or MG did not affect formation and

function of major blood vessels, including the dorsal aorta(DA), the posterior cardinal vein (PCV), ISVs, the dorsallongitudinal anastomotic vessel (DLAV), and parachordallymphangioblasts (PL), in up to 96 h of observation time(Fig. 1A and not shown). Likewise, vessel diameter of DA,PCV, ISV, and DLAV and heart beat including blood flowwere unaltered within the different experimental groups(Supplementary Fig. 1 and Supplementary Movies 1–3).However, altered blood vessels in the presence of hightissue glucose or high tissue MG were observed startingat 72 hpf, originating out of ISVs. In addition to appar-ently normal formed ISVs, several additional malformedand uncoordinated blood vessel structures were observeddeveloping out of ISVs but not from larger vessels, in-cluding the DA and PCV (Fig. 1A and B). Hyperbranchingendothelial cells appeared as small vessel structures thatformed from the upper part of ISVs, grew horizontallytoward neighboring ISVs, and partially connected tothese. Quantification of hyperbranches per zebrafish em-bryo at 96 hpf showed a dose-dependent, significant in-crease of hyperbranching endothelial cells in high tissueglucose or high tissue MG zebrafish embryos (Fig. 1B).Similar to high tissue glucose, MG induced the same vas-cular phenotype in tg(fli:EGFP) zebrafish embryos, butnumbers and length of malformed and uncoordinatedendothelial cells were less pronounced in high tissue glu-cose embryos as compared with high tissue MG zebrafishembryos (Fig. 1A and B). Incubation of tg(fli:EGFP) zebra-fish embryos with glucose and MG together further in-creased number of hyperbranching blood vessels (Fig. 1Aand B). Because glucose increases osmotic pressure inaqueous solutions, additional experiments in the presenceof mannitol were performed. Yet mannitol did not affectblood vessels in tg(fli:EGFP) zebrafish embryos (Fig. 1B),indicating that glucose- or MG-induced alterations ofblood vessels in zebrafish are not due to an increasedosmotic pressure.

Based on the hypothesis that hyperglycemia mediatesvascular complications via MG formation, we addressed thequestion of whether hyperglycemia increases MG in zebra-fish embryos. Zebrafish embryos were incubated in thepresence of different glucose or MG concentrations, andlevels of tissue glucose and tissue MG in zebrafish embryoswere analyzed. Glucose incubation with 15, 30, or 55mmol/L glucose solution dose dependently increased in-ternal zebrafish tissue glucose to 21, 23, and 32 nmol/mgprotein, respectively (Supplementary Fig. 2). Likewise, glu-cose dose dependently increased MG in zebrafish embryos,leading to a threefold increase of MG in zebrafish embryosthat were cultured in 30 mmol/L glucose solution reachingan internal tissue MG concentration of ;1.5 nmol/mgprotein (Fig. 1C). Thus high tissue glucose increased tissueMG concentrations in zebrafish. In addition, external MGrapidly increased MG in zebrafish embryos up to 5 nmol/mgprotein, confirming this experimental approach as a usefulstrategy to increase MG in living zebrafish embryos andto analyze early MG-driven effects on blood vessels

diabetes.diabetesjournals.org Jörgens and Associates 215

Figure 1—MG and glucose induce altered ISVs in zebrafish embryos. A: High tissue glucose (23 nmol/mg) and high tissue MG (5.3 nmol/mg)induce formation of hyperbranches (arrows) in the zebrafish trunk vasculature. Light images show overall morphology, and confocalimages show trunk vasculature of 96 hpf tg(fli:EGFP) zebrafish embryos. The black box marks the region displayed in the confocal image.The black scale bar is 500 mm, and the white scale bar is 100 mm. B: Quantification of hyperbranching zebrafish blood vessels at 96 hpf asseen in A in control solution–, mannitol-, MG-, and glucose-treated tg(fli:EGFP) zebrafish embryos at the indicated tissue concentrations;n = 108–158 embryos per group; mean 6 SE. Significance against control is given. C: Increased tissue MG levels in zebrafish embryosfollowing incubation with MG or glucose with indicated concentrations. Whole zebrafish embryo lysates were analyzed at 24 hpf; n = 3;mean 6 SE. Significance against control group is given. D: Intracardiac injection of MG induces hyperbranching of blood vessels at 96hpf. Embryos were injected at 48 hpf with 1 nL control or MG solution into the heart, and hyperbranches were analyzed at 96 hpf; n = 102–110embryos per group. Significance against control is given. Data were analyzed using the Mann-Whitney U test (B and D) or the Student t test(C). *P # 0.05; **P # 0.01; ***P < 0.001. CO, control; Gluc, glucose; MA, mannitol; ns, not significant.

216 Methylglyoxal in Zebrafish Blood Vessels Diabetes Volume 64, January 2015

independently of high glucose. To finally confirm the earlypathological consequences of MG on altered blood vesselsin zebrafish, MG was directly injected into the vasculatureof living zebrafish at 48 hpf, facilitating an intracardiacinjection (31), and formation of blood vessels was ana-lyzed at 96 hpf. Similar to high tissue MG experiments(Fig. 1A), intracardiac injection of MG induced alteredblood vessels originating out of the ISVs (Fig. 1D). Thisindicated a direct and glucose-independent effect of MGon the vasculature in living zebrafish embryos. Next weaddressed the question of whether expression silencing ofglo1, which detoxifies cellular MG to D-lactate, leads to anenhanced formation of MG and eventually to alteredblood vessels in zebrafish as it has been observed underhigh tissue glucose or high tissue MG (Fig. 1). To reduceglo1 expression in zebrafish, an antisense approach wasperformed using two splice-blocking morpholinos target-ing zebrafish glo1 (Fig. 2). Both morpholinos wereinjected into the one-cell stage, and functionality of thesplice-blocking morpholinos was validated by RT-PCR,showing decreased expression of wild-type glo1 but expres-sion of morphant glo1 (Fig. 2C). Subsequently, formationof the vascular system in tg(fli:EGFP) zebrafish embryos upto 96 hpf was analyzed. Similar to high tissue glucose andhigh tissue MG experiments (Fig. 1), reduced expression ofglo1 did not alter formation and function of major bloodvessels (DA, PCV, DLAV, ISV, and PL) in the zebrafishtrunk at 96 hpf. However, both morpholinos induced mal-formed and uncoordinated blood vessel structures thatagain originated out of ISVs (Fig. 2A). Quantification ofhyperbranching blood vessels at 96 hpf revealed approxi-mately three to four altered blood vessels per zebrafishembryo (Fig. 2B), which reflects the same number of al-tered blood vessels as was observed with an internal tissueglucose of 23 nmol/mg protein (Fig. 1B). Remarkably, re-duction of glo1 expression in zebrafish embryos increasedinternal MG concentration by approximately twofold (Fig.2D), which is similar to MG concentrations in high tissueglucose zebrafish (Fig. 1C). These data demonstrate thathigh tissue glucose or genetic reduction of glo1 expressionboth lead to similar elevated MG concentrations and sub-sequently to altered blood vessels in tg(fli:EGFP) zebrafishembryos. A similar vascular phenotype was also identifiedby using a genetic approach to induce hyperglycemia byinterfering with pancreas development in zebrafish. Ex-pression silencing of the transcription factor pdx1 in zebra-fish (30) increased glucose threefold (Fig. 2E) and inducedthe same vascular phenotype (Fig. 2A [bottom] and F) asseen for glo1 loss-of-function experiments (Fig. 2A and B)and for elevated MG and glucose concentrations (Fig. 1Aand B).

To finally prove that high tissue glucose–induced alteredblood vessels in zebrafish are primarily caused by MG, theMG-degrading enzyme glo1 was overexpressed in zebrafishembryos by glo1 mRNA injection (Fig. 3A and B). The abilityof glo1 to degrade MG should lead to decreased MG con-centrations and prevent vessel alterations in zebrafish

embryos. Indeed, glo1 overexpression significantly reducedglucose- or MG-induced altered blood vessel formation inzebrafish (Fig. 3A). In addition to the genetic glo1 over-expression experiments, we have also used AMG thatattenuates MG-induced modification of proteins (34). Tothis end, tg(fli:EGFP) zebrafish embryos with high tissueglucose were analyzed in the presence of 10 or 20 mmol/LAMG that showed a dose-dependent reduction of alteredblood vessels (Fig. 3C). Likewise, the number of alteredblood vessels in tg(fli:EGFP) zebrafish glo1 morphants wassignificantly reduced in the presence of AMG (Fig. 3D). AMGitself prevented high glucose–induced MG formation inzebrafish (Fig. 3E) that confirmed AMG’s ability to reduceMG levels in a living organism. Taken together, the datashow that high tissue glucose causes increased MG concen-trations in zebrafish embryos and MG induces early alter-ations of blood vessels.

MG Targets the VEGF Receptor 2 Signaling Pathway,Leading to Altered Blood Vessels in ZebrafishMG is a highly reactive dicarbonyl molecule that can induceposttranslational protein modifications. This has recentlybeen shown for sodium channel Nav1.8, which is associatedwith enhanced sensory neuron excitability and hyperalgesiain diabetic mice (13). Because of the fact that MG inducesaberrant blood vessels in tg(fli:EGFP) zebrafish embryos(Figs. 1–3), we proved the working hypothesis that MGdirectly targets the VEGF signaling cascade in zebrafishthat acts as the prime angiogenic pathway promoting phys-iological and pathophysiological angiogenesis (35). VEGFbinds to VEGF receptor 2 and activates its intrinsic tyro-sine kinase domain, leading to a rapid and strong auto-phosphorylation of VEGF receptor 2, activation of itsdownstream signaling pathways, and eventually blood ves-sel formation (36). Therefore, increased and permanentactivation of the VEGF receptor signaling cascade couldbe the primary cause for hyperbranching blood vessels inhigh tissue glucose or high tissue MG zebrafish embryos. Inorder to analyze VEGF receptor 2 activation in high tissueMG, tg(fli:EGFP) zebrafish embryos were analyzed forVEGF receptor 2 phosphorylation at 48 hpf. This timepoint was chosen because an MG-induced biochemical acti-vation of an angiogenic marker molecule will take ;2 daysto translate into a vascular phenotype. In control zebrafishembryos, VEGF receptor 2 is weakly phosphorylated. Yet inhigh tissue MG zebrafish, activation of VEGF receptor 2 issignificantly enhanced (Fig. 4A and B). Next we analyzedactivation of protein kinase Akt/PKB that acts downstreamof VEGF receptor 2 and regulates endothelial cell migrationand angiogenesis (36). Similar to VEGF receptor 2, Akt/PKB was weakly activated in control zebrafish embryos at48 hpf (Fig. 4C and D). In contrast, in high tissue MGzebrafish embryos, Akt/PKB phosphorylation was in-creased, but expression of Akt/PKB was unaltered (Fig.4C and D). Lastly we have addressed the question ofwhether increased Akt/PKB phosphorylation in high tissueMG zebrafish is due to an enhanced VEGF receptor 2

diabetes.diabetesjournals.org Jörgens and Associates 217

phosphorylation or whether Akt/PKB could be targeted byMG itself. To this end, high tissue MG zebrafish embryoswere incubated at 46 hpf in the presence of the VEGFreceptor 2 antagonist PTK787/Vatalanib, and phosphor-ylation of Akt/PKB was assessed at 48 hpf (Fig. 4E andF). Interestingly, Akt/PKB phosphorylation was signifi-cantly, but not completely, reduced. This suggests that

MG-driven Akt/PKB phosphorylation in high tissue MGzebrafish is only partially caused by an enhanced up-stream VEGF receptor 2 activation and may operate byother yet unknown mechanisms. Thus the data showthat MG induced hyperactivation of the VEGF receptorsignaling cascade that may lead to nonphysiological an-giogenesis in high tissue MG zebrafish embryos.

Figure 2—Endogenous expression silencing of glo1 and pdx1 increases MG or glucose levels and induces vessel hyperbranching inzebrafish embryos. A: Morpholino-mediated reduction of glo1 or pdx1 expression induces formation of hyperbranches (arrows) in tg(fli:EGFP) zebrafish embryos. Morpholinos SB-glo1-MO#1, SB-glo1-MO#2, and SB-pdx1-MO were injected into the one-cell stage, and tg(fli:EGFP) zebrafish embryos were analyzed at 96 hpf. The black box marks the region displayed in the confocal image. The black scale bar is500 mm, and the white scale bar is 100 mm. B: Quantification of hyperbranching blood vessels in glo1 morpholino-injected tg(fli:EGFP)zebrafish embryos at 96 hpf; n = 146–175 embryos per group; mean 6 SE. Significance against control (Control-MO) is given. C:Functionality of splice-blocking morpholinos SB-glo1-MO#1 and SB-glo1-MO#2. RT-PCR of Control-MO-, SB-glo1-MO#1-, and SB-glo1-MO#2-injected (3 ng each) zebrafish embryos at 24 hpf show wild type and generation of morphant glo1 signals. D: MG levelswere elevated following reduction of glo1 expression. Zebrafish lysates were analyzed at 48 hpf; n = 3; mean 6 SE. Significance againstcontrol group (Control-MO) is given. E: Inhibition of pancreas development in zebrafish by pdx1 expression silencing induced hypergly-cemia. Zebrafish lysates were analyzed at 48 hpf; n = 3; mean 6 SE. Significance against control group (Control-MO) is given. F:Quantification of hyperbranching blood vessels in pdx1-MO–injected tg(fli:EGFP) zebrafish embryos at 96 hpf; n = 62–86 embryos pergroup; mean 6 SE. Significance against control (Control-MO) is given. Data were analyzed using the Mann-Whitney U test (B and F ) or theStudent t test (D and E). **P # 0.01; ***P < 0.001. MO, morpholino; Morph, morphant; ns, not significant; WT, wild type.

218 Methylglyoxal in Zebrafish Blood Vessels Diabetes Volume 64, January 2015

Therefore, MG-induced increase of the VEGF signalingcascade could be the primary cause for early morpholog-ical vascular abnormalities leading to several additionalmalformed and uncoordinated blood vessel structures in

tg(fli:EGFP) zebrafish embryos. Interestingly, altered bloodvessels were not formed because of an increase in endo-thelial cell proliferation in zebrafish embryos, as neitherhigh tissue glucose nor high tissue MG increased the

Figure 3—Blood vessel hyperbranching in 96 hpf zebrafish embryos is induced by MG. A: Overexpression of glo1 reduced high tissue glucose–and high tissue MG–mediated (23 and 5.3 nmol/mg, respectively) hyperbranching of blood vessels in 96 hpf zebrafish. tg(fli:EGFP) zebrafishembryos were injected with 100 pg mOrange mRNA or 100 pg glo1 mRNA into the one-cell stage, and quantification of blood vessel hyper-branching is shown at 96 hpf. B: Representative Western blot for glo1 protein levels in 24 hpf zebrafish lysates after injection of 100 pg mOrangeor 100 pg glo1 mRNA. C and D: Quantification of blood vessel hyperbranching in 96 hpf tg(fli:EGFP) zebrafish embryos in the presence of 10 or20 mmol/L AMG. AMG treatment reduced high tissue glucose–induced (23 nmol/mg) (C) and glo1 knockdown–induced (D) hyperbranching at 96hpf; n = 104–131 (C) or 107–124 (D) embryos per group; mean 6 SE. Significance against control (Control or Control-MO) is given. E: AMGreduces MG levels in zebrafish embryos. High tissue glucose zebrafish embryos were incubated in the presence of AMG. Zebrafish lysates wereanalyzed at 24 hpf for MG formation; n = 3; mean6 SE. Significance against control group is given. Data were analyzed using the Mann-WhitneyU test (A, C, and D) or Student t test (E). **P # 0.01; ***P < 0.001. Gluc, glucose; MO, morpholino; ns, not significant.

diabetes.diabetesjournals.org Jörgens and Associates 219

number of endothelial cells in 48 and 96 hpf tg(nflk:EGFP)zebrafish embryos (Supplementary Fig. 3A and B).

To finally prove the working hypothesis that MG altersactivation of the VEGF signaling cascade leading toaltered blood vessels in zebrafish, pharmacological inhib-itors against VEGF receptor 2 and important downstreamsignaling molecules were used. At 48 hpf, when the majortrunk vessel in zebrafish have already been formed (37)and where VEGF receptor 2 is strongly activated (Fig. 4Aand B), a low dose (0.2 mmol/L) of VEGF receptor 2 an-tagonist PTK787/Vatalanib was added to tg(fli:EGFP) hightissue MG and high tissue glucose zebrafish embryos and,in this concentration, did not affect physiological bloodvessel development in zebrafish (38). Subsequently, bloodvessels in tg(fli:EGFP) zebrafish embryos were analyzed at96 hpf. High tissue glucose or high tissue MG zebrafishembryos again showed hyperbranching blood vessels; yetthe vascular phenotype was less pronounced, as seen inFigs. 1 and 3, which is very likely due to the cytotoxicityof the solvent DMSO (Figs. 5B and 6C and D). Sim-ilar experiments in the presence of PTK787 completely

reverted altered blood vessel formation, and inhibitor-treated embryos were indistinguishable from control-treatedanimals (Fig. 5A and B). The same data were observedusing 5 mmol/L LY294002 that blocks activation of PI3Kacting as an essential downstream regulator of VEGFreceptor 2 (36) (Fig. 6A and C). In addition, Akt/PKBthat is strongly phosphorylated in zebrafish embryos inthe presence of high tissue MG (Fig. 4C and D) and thatis specifically blocked by 1 mmol/L SC66 prevented vas-cular alterations in high tissue MG as well as high tissueglucose tg(fli:EGFP) zebrafish embryos (Fig. 6B and D).Finally, pharmacological inhibition of p38MAP kinase,which is considered as a negative regulator of VEGF sig-naling (39,40), was performed in high tissue glucose orhigh tissue MG tg(fli:EGFP) zebrafish embryos. Incubationof zebrafish embryos with 50 mmol/L p38MAP kinase in-hibitor SB203580 did not prevent vessel hyperbranch-ing, but interestingly, SB203580 alone already inducedadditional, malformed, and uncoordinated blood vesselstructures in control zebrafish embryos at 96 hpf (Sup-plementary Fig. 4).

Figure 4—MG induces increased phosphorylation of VEGF receptor 2 and Akt/PKB in zebrafish embryos at 48 hpf. A: RepresentativeWestern blot for increased VEGF receptor 2 phosphorylation in high tissue MG (5.3 nmol/mg) zebrafish embryos. pVEGF receptor 2antibody was used to detect pVEGF receptor 2 in whole tg(fli:EGFP) zebrafish lysates at 48 hpf. Total (t)VEGF receptor 2 serves as loadingcontrol. B: Quantification of VEGF receptor 2 phosphorylation after MG treatment (n = 3). C: Representative Western blot for increased Akt/PKB phosphorylation in high tissue MG (5.3 nmol/mg) zebrafish embryos. pAkt antibody was used to detect phosphorylated Akt/PKB inwhole tg(fli:EGFP) zebrafish lysates at 48 hpf. Total (t)Akt serves as loading control. D: Quantification of Akt/PKB phosphorylation (n = 3). Eand F: Inhibition of VEGF receptor 2 activation in zebrafish by PTK787/Vatalanib (PTK787) reduced high tissue MG–driven Akt/PKBphosphorylation. PTK787 was added at 46 hpf, and MG-induced activation of Akt/PKB was analyzed at 48 hpf (n = 3). Data were analyzedusing the Student t test. *P # 0.05. CO, control; VEGFR2, VEGF receptor 2.

220 Methylglyoxal in Zebrafish Blood Vessels Diabetes Volume 64, January 2015

In order to identify additional pathways that maytranscriptionally regulate aberrant blood vessel forma-tion, we have performed microarray analyses using96-hpf-old zebrafish embryos that were kept in thepresence of MG. Interestingly, we have not observed

major transcriptional regulations of important angio-genic molecules (Supplementary Table 1). However,MG treatment interfered with metabolic pathwaysthat are related to glucose metabolism (SupplementaryTable 2).

Figure 5—Inhibition of VEGF receptor 2 phosphorylation by PTK787/Vatalanib prevents blood vessel hyperbranching in zebrafishembryos. A: Altered blood vessels in high tissue glucose (23 nmol/mg) and high tissue MG (5.3 nmol/mg) tg(fli:EGFP) zebrafish embryoswere reduced after incubation with 0.2 mmol/L PTK787 for 2 days, starting at 48 hpf. Light images show overall morphology, andconfocal images show trunk vasculature of 96 hpf tg(fli:EGFP) zebrafish embryos. The black box marks the region displayed in theconfocal image. The black scale bar is 500 mm, and the white scale bar 100 mm. B: Quantification of blood vessel hyperbranching intg(fli:EGFP) zebrafish embryos at 96 hpf; n = 101–103 embryos per group; mean 6 SE. Significance against control is given. Data wereanalyzed using the Mann-Whitney U test. ***P < 0.001. Gluc, glucose.

diabetes.diabetesjournals.org Jörgens and Associates 221

Together the data show that high tissue glucose inducedMG formation in tg(fli:EGFP) zebrafish embryos and thathigh tissue MG alone induced early vascular abnormalitiesin small blood vessels in the zebrafish trunk, which arecaused by MG-driven activation of the VEGF receptor sig-naling cascade.

DISCUSSIONIn this study, MG was identified as cause for early vascularalterations in zebrafish embryos based on the followingobservations: 1) MG accumulates in zebrafish under hightissue glucose as well as in glo1 loss-of-function experi-ments, 2) MG acts on endothelial cells in zebrafish and alters

Figure 6—Inhibition of PI3K and Akt/PKB phosphorylation prevents blood vessel hyperbranching in zebrafish embryos. Altered blood vesselsin high tissue glucose (23 nmol/mg) and high tissue MG (5.3 nmol/mg) tg(fli:EGFP) zebrafish embryos were reduced after incubation with 5mmol/L PI3K inhibitor LY294002 (A) or 1 mmol/L Akt/PKB inhibitor SC66 (B) for 2 days, starting at 48 hpf. Light images show overallmorphology, and confocal images show trunk vasculature of 96 hpf tg(fli:EGFP) zebrafish embryos. The black box marks the region displayedin the confocal image. The black scale bar is 500 mm, and the white scale bar is 100 mm. Quantification of blood vessel hyperbranching intg(fli:EGFP) zebrafish embryos at 96 hpf after incubation with 5 mmol/L LY294002 (C) or 1 mmol/L SC66 (D); n = 101–105 (C) or 106–123 (D)embryos per group; mean6 SE. Significance against control is given. Data were analyzed using the Mann-Whitney U test. ***P< 0.001. Gluc,glucose; ns, not significant.

222 Methylglyoxal in Zebrafish Blood Vessels Diabetes Volume 64, January 2015

ISVs via increased phosphorylation of VEGF receptor 2 andAkt/PKB, and 3) MG-mediated pathological blood vessel for-mation in zebrafish is prevented by pharmacological inhibi-tion of VEGF receptor 2 and Akt/PKB activation as well asby lowering internal MG concentrations (Fig. 7).

The data further advance the recently developed conceptthat hyperglycemia-induced formation of MG acts as animportant cause for development of late-stage complicationsin diabetic patients (13). This study identified in zebrafisha new mechanistic link to how high tissue glucose–inducedMG formation can initiate vascular complications in verte-brates. VEGF receptor 2 and protein kinase Akt/PKB wereidentified in vivo as MG targets that mediate altered anduncoordinated blood vessel function and nonphysiologicalangiogenesis in smaller but not in larger blood vessels. Im-portantly, MG induces vascular alterations in zebrafish in-dependently of high tissue glucose, indicating that MG actsas a primary trigger for high glucose–induced vascular com-plications in this model. In contrast to diabetic rats, whereVEGF receptor 2 protein is elevated (41), in this study, MGdid not significantly regulate VEGF receptor 2 expressionand expression of other angiogenic proteins (Fig. 4 and

Supplementary Table 1), which further supports the conceptthat modifications of VEGF receptor 2 or Akt/PKB proteinsby MG are the cause for altered blood vessels in the zebrafishembryos. One possible mechanism could be that MG directlymodifies amino acids in the VEGF receptor 2 and Akt/PKBproteins. Such a direct, MG-induced posttranslational modi-fication has recently been shown for Akt/PKB in smoothmuscle cells, where MG directly modifies amino acid Cys77of Akt/PKB, thereby leading to enhanced phosphorylation atSer473 and Thr308 (42). Alternatively, enhanced activationof VEGF receptor 2 and Akt/PKB could be an indirect effectinduced by altered function of pathways that regulate VEGFreceptor 2 and Akt/PKB activation level.

In addition to the mechanistic data, the study alsoshows how MG acts on blood vessels in vivo. Several highglucose– and MG-driven effects on the vasculature werereported from cultured endothelial cells with controver-sial results (14,43,44). In this study, vascular alterationswere only observed in a subset of blood vessels, namely, inISVs, but not in larger blood vessels, including the DA andthe PCV. This also shows in zebrafish a differential sus-ceptibility and regulation of endothelial cell function and

Figure 7—Model of MG action in zebrafish embryos. Hyperglycemia leads to elevated glucose influx in endothelial cells. Due to highintracellular glucose concentrations, glycolysis is constantly activated, which results in increased MG formation. High intracellular MGconcentrations activate phosphorylation of VEGF receptor 2 and its downstream target Akt/PKB. Consequently, MG induces altered bloodvessels in zebrafish, which could be prevented in the presence of MG scavenger AMG and by overexpression of glo1. A similar vascularphenotype is induced by expression silencing of glo1 that increases intracellular concentrations of MG in zebrafish. Altered MG-inducedblood vessels could be blocked by inhibition of VEGF receptor 2 phosphorylation (PTK787), PI3K activation (LY294002), or Akt/PKBphosphorylation (SC66).

diabetes.diabetesjournals.org Jörgens and Associates 223

angiogenesis by high tissue MG. In this study, MG rapidlyincreases activation of VEGF receptor 2 and protein ki-nase Akt/PKB in zebrafish, leading to nonphysiologicalangiogenesis. In contrast to other rodent models thatare usually used in experimental diabetic research andthat need long periods of hyperglycemia to develop late-stage diabetes complications, zebrafish embryos alreadydeveloped vascular complications within 4 days of hightissue glucose or high tissue MG. Thus the data highlightzebrafish as a model organism to analyze hyperglycemiaand MG-induced pathological vascular alterations and itsmechanisms.

In conclusion, the data identified a mechanism inzebrafish for how elevated MG concentrations alter smallblood vessels via modifications of the VEGF receptor 2signaling cascade. Because MG is increased under hightissue glucose but also induces vascular alterations alone,this study identified MG as an important contributor tohigh glucose–induced vascular alterations.

Acknowledgments. The authors thank Katrin Bennewitz (Centre for Bio-medicine and Medical Technology Mannheim, Medical Faculty Mannheim, Hei-delberg University and German Cancer Research Center [DKFZ-ZMBH Alliance])and Maria Muciek (Medical Research Center, Medical Faculty Mannheim, Heidel-berg University) for technical assistance.Funding. This study was supported by Deutsche Forschungsgemeinschaft(SFB1118; projects A4, B1, C03, S01, SFB/TR23; project Z5, IRTG1874/1;projects SP6, SP9; IS67/5-1) and the Hopp Foundation. The authors acknowl-edge the support of the Core Facility Live Cell Imaging Mannheim at the Centrefor Biomedicine and Medical Technology Mannheim (DFG INST 91027/10-1FUGG).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. K.J. and S.J.S. performed experiments, analyzeddata, and wrote the manuscript. J.P., T.H.F., and C.S. performed experiments andanalyzed data. P.P.N. and H.-P.H. gave conceptual and technological advice. J.K.conceived and designed the study and wrote the manuscript. J.K. is the guarantorof this work and, as such, had full access to all the data in the study and takesresponsibility for the integrity of the data and the accuracy of the data analysis.Prior Presentation. Parts of this study were presented at Keystone Sym-posia on Molecular and Cellular Biology “Complications of Diabetes (X7),” WhistlerConference Centre, Whistler, British Columbia, Canada, 23–28 March 2014.

References1. Brownlee M. Biochemistry and molecular cell biology of diabetic compli-cations. Nature 2001;414:813–8202. Donnelly R, Emslie-Smith AM, Gardner ID, Morris AD. ABC of arterial andvenous disease: vascular complications of diabetes. BMJ 2000;320:1062–10663. Hemmingsen B, Lund SS, Gluud C, et al. Intensive glycaemic control forpatients with type 2 diabetes: systematic review with meta-analysis and trialsequential analysis of randomised clinical trials. BMJ 2011;343:d68984. 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;352:837–8535. Skriver MV, Borch-Johnsen K, Lauritzen T, Sandbaek A. HbA1c as predictorof all-cause mortality in individuals at high risk of diabetes with normal glucosetolerance, identified by screening: a follow-up study of the Anglo-Danish-DutchStudy of Intensive Treatment in People with Screen-Detected Diabetes in PrimaryCare (ADDITION), Denmark. Diabetologia 2010;53:2328–2333

6. Cheung NW, Wong VW, McLean M. The Hyperglycemia: Intensive InsulinInfusion in Infarction (HI-5) study: a randomized controlled trial of insulin infusiontherapy for myocardial infarction. Diabetes Care 2006;29:765–7707. Malmberg K. Prospective randomised study of intensive insulin treatment onlong term survival after acute myocardial infarction in patients with diabetesmellitus. DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute MyocardialInfarction) Study Group. BMJ 1997;314:1512–15158. Malmberg K, Rydén L, Efendic S, et al. Randomized trial of insulin-glucoseinfusion followed by subcutaneous insulin treatment in diabetic patients withacute myocardial infarction (DIGAMI study): effects on mortality at 1 year. J AmColl Cardiol 1995;26:57–659. Malmberg K, Rydén L, Wedel H, et al.; DIGAMI 2 Investigators. Intensemetabolic control by means of insulin in patients with diabetes mellitus and acutemyocardial infarction (DIGAMI 2): effects on mortality and morbidity. Eur Heart J2005;26:650–66110. Lo TW, Westwood ME, McLellan AC, Selwood T, Thornalley PJ. Binding andmodification of proteins by methylglyoxal under physiological conditions. A kineticand mechanistic study with N alpha-acetylarginine, N alpha-acetylcysteine, andN alpha-acetyllysine, and bovine serum albumin. J Biol Chem 1994;269:32299–3230511. Thornalley PJ, Rabbani N. Glyoxalase in tumourigenesis and multidrug re-sistance. Semin Cell Dev Biol 2011;22:318–32512. Fleming T, Cuny J, Nawroth G, et al. Is diabetes an acquired disorder of reactiveglucose metabolites and their intermediates? Diabetologia 2012;55:1151–115513. Bierhaus A, Fleming T, Stoyanov S, et al. Methylglyoxal modification ofNav1.8 facilitates nociceptive neuron firing and causes hyperalgesia in diabeticneuropathy [published correction appears in Nat Med 2012;18:1445]. Nat Med2012;18:926–93314. Dobler D, Ahmed N, Song L, Eboigbodin KE, Thornalley PJ. Increased di-carbonyl metabolism in endothelial cells in hyperglycemia induces anoikis andimpairs angiogenesis by RGD and GFOGER motif modification. Diabetes 2006;55:1961–196915. Schalkwijk CG, van Bezu J, van der Schors RC, Uchida K, Stehouwer CD,van Hinsbergh VW. Heat-shock protein 27 is a major methylglyoxal-modifiedprotein in endothelial cells. FEBS Lett 2006;580:1565–157016. Yao D, Taguchi T, Matsumura T, et al. High glucose increases angiopoietin-2 transcription in microvascular endothelial cells through methylglyoxal modifi-cation of mSin3A. J Biol Chem 2007;282:31038–3104517. Rabbani N, Godfrey L, Xue M, et al. Glycation of LDL by methylglyoxal in-creases arterial atherogenicity: a possible contributor to increased risk of car-diovascular disease in diabetes. Diabetes 2011;60:1973–198018. Wang X, Desai K, Chang T, Wu L. Vascular methylglyoxal metabolism andthe development of hypertension. J Hypertens 2005;23:1565–157319. Guo Q, Mori T, Jiang Y, et al. Methylglyoxal contributes to the developmentof insulin resistance and salt sensitivity in Sprague-Dawley rats. J Hypertens2009;27:1664–167120. Fleming TH, Theilen TM, Masania J, et al. Aging-dependent reduction inglyoxalase 1 delays wound healing. Gerontology 2013;59:427–43721. Tyedmers J. Yeast as a model to understand mechanisms and con-sequences of protein modifications by metabolism derived toxic products. ExpClin Endocrinol Diabetes 2012;120:179–18122. Mendler M, Schlotterer A, Morcos M, Nawroth PP. Understanding diabeticpolyneuropathy and longevity: what can we learn from the nematode Caenor-habditis elegans? Exp Clin Endocrinol Diabetes 2012;120:182–18323. Teleman AA, Ratzenböck I, Oldham S. Drosophila: a model for un-derstanding obesity and diabetic complications. Exp Clin Endocrinol Diabetes2012;120:184–18524. Jörgens K, Hillebrands JL, Hammes HP, Kroll J. Zebrafish: a model for un-derstanding diabetic complications. Exp Clin Endocrinol Diabetes 2012;120:186–18725. Alvarez Y, Chen K, Reynolds AL, Waghorne N, O’Connor JJ, Kennedy BN.Predominant cone photoreceptor dysfunction in a hyperglycaemic model of non-proliferative diabetic retinopathy. Dis Model Mech 2010;3:236–245

224 Methylglyoxal in Zebrafish Blood Vessels Diabetes Volume 64, January 2015

26. Gleeson M, Connaughton V, Arneson LS. Induction of hyperglycaemia inzebrafish (Danio rerio) leads to morphological changes in the retina. Acta Diabetol2007;44:157–16327. Olsen AS, Sarras MP Jr, Intine RV. Limb regeneration is impaired in an adultzebrafish model of diabetes mellitus. Wound Repair Regen 2010;18:532–54228. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF. Stages ofembryonic development of the zebrafish. Dev Dyn 1995;203:253–31029. Robu ME, Larson JD, Nasevicius A, et al. p53 activation by knockdowntechnologies. PLoS Genet 2007;3:e7830. Jurczyk A, Roy N, Bajwa R, et al. Dynamic glucoregulation and mammalian-like responses to metabolic and developmental disruption in zebrafish. Gen CompEndocrinol 2011;170:334–34531. Stoll SJ, Bartsch S, Augustin HG, Kroll J. The transcription factor HOXC9regulates endothelial cell quiescence and vascular morphogenesis in zebrafishvia inhibition of interleukin 8. Circ Res 2011;108:1367–137732. McLellan AC, Phillips SA, Thornalley PJ. The assay of methylglyoxal in bi-ological systems by derivatization with 1,2-diamino-4,5-dimethoxybenzene. AnalBiochem 1992;206:17–2333. Kunsst A. Methods of Enzymatic Analysis. 3rd ed. Bergmeyer HU, Ed. NewYork, Academic Press, 1984, p. 163–17234. Lo TW, Selwood T, Thornalley PJ. The reaction of methylglyoxal withaminoguanidine under physiological conditions and prevention of methylglyoxalbinding to plasma proteins. Biochem Pharmacol 1994;48:1865–187035. Adams RH, Alitalo K. Molecular regulation of angiogenesis and lym-phangiogenesis. Nat Rev Mol Cell Biol 2007;8:464–47836. Olsson AK, Dimberg A, Kreuger J, Claesson-Welsh L. VEGF receptor sig-nalling - in control of vascular function. Nat Rev Mol Cell Biol 2006;7:359–371

37. Epting D, Wendik B, Bennewitz K, Dietz CT, Driever W, Kroll J. The Rac1regulator ELMO1 controls vascular morphogenesis in zebrafish. Circ Res 2010;107:45–5538. Hollenbach M, Stoll SJ, Jörgens K, Seufferlein T, Kroll J. Different regulationof physiological and tumor angiogenesis in zebrafish by protein kinase D1(PKD1). PLoS ONE 2013;8:e6803339. Issbrücker K, Marti HH, Hippenstiel S, et al. p38 MAP kinase—a molecularswitch between VEGF-induced angiogenesis and vascular hyperpermeability.FASEB J 2003;17:262–26440. Matsumoto T, Turesson I, Book M, Gerwins P, Claesson-Welsh L. p38MAP kinase negatively regulates endothelial cell survival, proliferation, anddifferentiation in FGF-2-stimulated angiogenesis. J Cell Biol 2002;156:149–16041. Huang J, Li X, Li M, et al. Effects of intravitreal injection of KH902,a vascular endothelial growth factor receptor decoy, on the retinas ofstreptozotocin-induced diabetic rats. Diabetes Obes Metab 2012;14:644–65342. Chang T, Wang R, Olson DJ, Mousseau DD, Ross AR, Wu L. Modification ofAkt1 by methylglyoxal promotes the proliferation of vascular smooth muscle cells.FASEB J 2011;25:1746–175743. Chou E, Suzuma I, Way KJ, et al. Decreased cardiac expression of vascularendothelial growth factor and its receptors in insulin-resistant and diabeticStates: a possible explanation for impaired collateral formation in cardiac tissue.Circulation 2002;105:373–37944. Liu H, Yu S, Zhang H, Xu J. Angiogenesis impairment in diabetes: role ofmethylglyoxal-induced receptor for advanced glycation endproducts, autophagyand vascular endothelial growth factor receptor 2. PLoS ONE 2012;7:e46720

diabetes.diabetesjournals.org Jörgens and Associates 225