Free Radical Biology and Medicine 53 (2012) 357–365

Contents lists available at SciVerse ScienceDirect

Free Radical Biology and Medicine

0891-58

http://d

n Corr

E-m

journal homepage: www.elsevier.com/locate/freeradbiomed

Original Contribution

Puerarin inhibits the retinal pericyte apoptosis induced by advancedglycation end products in vitro and in vivo by inhibiting NADPHoxidase-related oxidative stress

Junghyun Kim, Ki Mo Kim, Chan-Sik Kim, Eunjin Sohn, Yun Mi Lee, Kyuhyung Jo, Jin Sook Kim n

Traditional Korean Medicine (TKM) Based Herbal Drug Research Group, Herbal Medicine Research Division, Korea Institute of Oriental Medicine,

1672 Yuseongdaero, Yuseong-gu, Daejeon 305–811, South Korea

a r t i c l e i n f o

Article history:

Received 2 September 2011

Received in revised form

25 April 2012

Accepted 27 April 2012Available online 17 May 2012

Keywords:

Advanced glycation end products

Apoptosis

Iso-flavones

NADPH oxidase

Oxidative stress

Retinal pericyte

Puerarin

49/$ - see front matter & 2012 Elsevier Inc. A

x.doi.org/10.1016/j.freeradbiomed.2012.04.03

esponding author. Fax: þ82 42 868 9471.

ail address: jskim@kiom.re.kr (J.S. Kim).

a b s t r a c t

Retinal pericyte loss is one of the histopathological hallmarks of early diabetic retinopathy. Puerarin

(40-7-dihydroxy-8-beta-D-glucosylisoflavone), which is an isoflavone-C-glucoside, causes various phar-

macological effects that include antihyperglycemic and anti-inflammatory activities. In the present

study, we determined the efficacy and possible mechanism of puerarin on the advanced glycation

end product (AGE)-modified bovine serum albumin (BSA)-induced apoptosis of cultured bovine retinal

pericytes and rat retinal pericytes in intravitreally AGE-modified rat serum albumin (RSA)-injected

eyes. Puerarin significantly inhibited pericyte apoptosis, the generation of reactive oxygen species

(ROS), and NADPH oxidase activity by inhibiting the phosphorylation of p47phox and Rac1 which were

induced by the AGE-BSA treatment. The puerarin treatment markedly suppressed the activation of

nuclear factor-kappaB (NF-kB). In addition, the in vivo apoptosis of the retinal pericyte of rats that

was stimulated by the intravitreal injection of AGE-RSA was evidently attenuated by the puerarin

treatment. These results demonstrate that puerarin may exert inhibitory effects on AGE-induced

pericyte apoptosis by interfering with the NADPH oxidase-related ROS pathways and blocking NF-kB

activation, thereby ameliorating retinal microvascular dysfunction.

& 2012 Elsevier Inc. All rights reserved.

Introduction

Diabetic retinopathy is one of the major complications ofdiabetic mellitus and is the primary cause of acquired blindnessin working-age adults. The loss of retinal pericytes is a hallmark ofearly diabetic retinopathy and leads to the development of retinalpathology, including thickening of the basement membrane, theformation of acellular capillaries, retinal hemorrhage, endothelialproliferation, and angiogenesis; these processes ultimately lead toblindness [1]. Several studies have demonstrated that retinalmicrovascular cell apoptosis plays a crucial role in the develop-ment of early diabetic retinopathy [2].

All forms of diabetes are characterized by chronic hyperglycemia,which is responsible for the development and progression of vascularcomplications, such as diabetic nephropathy and retinopathy [3,4]. Itis well known that hyperglycemia induces an increased production ofreactive oxygen species (ROS), particularly in the retinal vessels [5].This hyperglycemia-induced oxidative stress plays a crucial role inthe development of diabetic retinal vascular complications [1,6,7].Therefore, the diabetes-induced retinal vascular dysfunction may berescued by blocking the production of ROS.

ll rights reserved.

0

Puerarin (daidzein-8-C-glucoside) is an isoflavone glycoside thatis isolated from the root of Pueraria lobata and has been used intraditional Korean medicine to treat various diseases for millennia.Some plants contain large quantities of puerarin [8,9]. Puerarin hasbeen studied for its use as an antihyperglycemic agent because itincreases insulin sensitivity and protects the pancreatic islets [10,11].Additionally, puerarin effectively inhibits the formation of advancedglycation end product (AGE) formation; this formation is a character-istic risk factor for diabetic complications [12].

It is widely accepted that puerarin plays a protective role indiabetic retinopathy, but its mechanism remains poorly understood.Therefore, in this study, we investigated the protective effect ofpuerarin on retinal pericyte loss. We studied the AGE-modifiedbovine serum albumin (BSA)-induced apoptotic cell death of culturedbovine retinal pericytes in the presence or absence of puerarin. Inaddition, we examined the preventive effect of puerarin on thealteration of retinal vessels from the intravitreally AGE-modified ratserum albumin (RSA)-injected eyes of rats.

Materials and methods

Primary bovine retinal pericytes culture

Primary retinal pericyte cells were isolated from bovine retinalmicrovessels using the previously published methods [13,14].

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365358

Cells that were derived between the third and fifth passages wereused in this experiment. The cells were plated onto appropriateculture dishes and used in experiments after they reached 80%confluence. The standard culture medium was replaced with freshserum-free medium 16 h before experiments were performed.

Cellular apoptosis detected using flow cytometry

Bovine retinal pericytes were seeded in 12-well plates. Thecells were preincubated with puerarin (1, 5, and 10 mM), NAC(1 mM, Calbiochem, San Diego, CA, USA), or DPI (5 mM, Calbio-chem) for 1 h, and the retinal pericytes were stimulated withvarious concentrations of AGE-BSA (CircuLex, Nagano, Japan) forthe indicated times. Puerarin used in the present study wasisolated as described previously [12]. Nonglycated BSA was usedas a control to ensure that each control well was exposed to analbumin concentration equivalent to that of the treated wells. Thecells were then collected and treated according to the protocolincluded in annexin V-fluorescein isothiocyanate (FITC) ApoptosisDetection Kit (Merck, Darmstadt, Germany) and the percentagesof apoptotic cells were determined using a flow cytometer (FACSCalibur, Becton Dickinson, San Jose, CA, USA). The results wereanalyzed using Cell Quest Pro software (Becton Dickinson).

Measurement of ROS production

Intracellular ROS production was detected using dihydrodi-chlorofluorescein diacetate (DCF-DA, Invitrogen, Carlsbad, CA,USA). The cells were treated as indicated and then loaded with5 mM DCF-DA for 30 min. Ten thousand cells were analyzed by aflow cytometer (Becton Dickinson), and peroxide-containing cellswere identified as those with increased fluorescence of oxidizedDCF-DA. Nonglycated BSA was used as a control.

Determination of NADPH oxidase activity

The lucigenin-derived enhanced chemiluminescence assaywas used to determine NADPH oxidase activity. The cells weretreated as indicated and detached and resuspended in ice-coldbuffer that contained 1 mM ethylene glycol tetraacetic acid,protease inhibitors, and 150 mM sucrose. The cells were distrib-uted at 5�104/well onto a 96-well microplate luminometer(Synergy HT, Bio-Tek, Winooski, VT, USA). Immediately beforerecording, NADPH (100 mmol/L) and dark-adapted lucigenin(5 mmol/L) were added to cell suspensions. Light emission wasmeasured for 6 min in quadruplicate The data were collected at2 min intervals in order to measure the relative changes inNADPH oxidase activity.

Western blotting

Western blotting was performed as previously described [15].Primary antibodies against phospho-p47phox and phospho-Rac1were obtained from Cell Signaling Technology (Danvers, MA,USA). Antibodies against caspase-3 and poly(ADP-ribose) poly-merase-1 (PARP-1) were obtained from Santa Cruz Biotechnology(Santa Cruz, CA, USA). Anti-b-actin antibody was obtainedfrom Sigma.

TUNEL staining

The pericytes were fixed with 4% paraformaldehyde. ATdT-mediated dUTP nick-end labeling (TUNEL) staining wasperformed with a DeadEnd Fluorometirc TUNEL kit as per themanufacturer’s instructions (Promega, Madison, WI, USA).

NF-kB nuclear translocalization and measuring of NF-kB activity

To evaluate NF-kB nuclear translocalization, immunofluores-cence staining for the NF-kB p65 subunit was performed on thecultured retinal pericytes using a mouse anti-NF-kB p65 antibody(Santa Cruz). NF-kB was visualized using a FITC-conjugated goatanti-mouse IgG (Santa Cruz). For an electrophoretic mobility shiftassay (EMSA), nuclear extracts were prepared using a NE-PER kit(Pierce Biotechnology, Rockfold, IL, USA) according to the manu-facturer’s instructions. The EMSA assay was performed by incu-bating 10 mg of nuclear protein extract with IRDye 700-labeledNF-kB oligonucleotide (LI-COR, Lincoln, NE, USA) or an unlabeledprobe, which was used as a cold competitor. A supershift assaywas performed to identify the shifted bands in the EMSA. Thenuclear extracts were incubated with p50 and p65 antibodies(Santa Cruz) for 20 min before the EMSA was performed. TheEMSA gels were analyzed, and the images were captured andquantified using a LI-COR Odyssey infrared laser imaging system(LI-COR). To confirm and evaluate NF-kB activation, we alsoperformed a luciferase assay. The cells (105 cells/well) wereseeded on 24-well plates the day before transfection. The plas-mids NF-kB-Luc (1 mg/well) (Stratagene, La Jolla, CA, USA) andpRL-SV40 (0.2 mg/well) (Promega, WI, USA) were transfected intocells using Lipofectamine (Invitrogen). The pNF-kB-Luc plasmidconsisted of the NF-kB-binding region followed by the fireflyluciferase reporter gene. The pRL-SV40 plasmid served as aninternal control to normalize the transfection efficiency. Afterexposure to the various treatments, the cells were harvested andlysed in 100 mL of lysis reagent. Twenty microliters of cell lysatewas then mixed with 100 mL of luciferin (the substrate ofluciferase), and the luminescence was detected immediately.The luminescence was measured using a luminometer (SynergyHT, Bio-Tek). All reagents for the luciferase assays were purchasedfrom Promega.

Synthesis of AGE-modified rat serum albumin

AGE-RSA was prepared by incubating RSA (fraction V, lowendotoxin; Sigma) with 500 mM D-glucose under aerobic con-ditions for 10 weeks at 37 1C in the presence of proteaseinhibitors and antibiotics based on published methods [16].Unmodified RSA, which was used as a control, was preparedunder the same conditions but without the addition of sugar.Finally, the preparations were extensively dialyzed againstphosphate buffer to remove free glucose. The extent of advancedglycation was assessed using competitive ELISA with a mono-clonal anti-AGE antibody (TransGenic Inc., Kumamoto, Japan).The content of AGE-RSA was approximately 200-fold higher thanthe nonglycated control. The endotoxin content was determinedusing the Limulus amebocyte lysate assay (E-Toxate kit; Sigma)to be less than 0.015 EU/mL in both solutions.

Intravitreal injection of AGE-modified rat serum albumin

Male Sprague-Dawley rats (8 weeks old) were used inthis study. Each rat was anesthetized using a 1:1 mixture ofxylazine hydrochloride (4 mg/kg) and ketamine hydrochloride(10 mg/kg). Rats were randomly divided into two groups of eightrats as follows: (1) eight rats of the control group, (2) eight ratsof the group treated with puerarin. Rats in the control groupwere injected with a single dose of 6 mg AGE-RSA in a volume of3 mL into the vitreous of the right eye using a microinjector(Hamilton, Reno, NV, USA) under a dissecting microscope. Forthe normal control, 6 mg RSA in a volume of 3 mL was injectedinto the left eye. In the treatment group, puerarin was injected ata concentration of 400 mM in a volume of 1.5 mL which is

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365 359

equivalent to a dose of 10 mM for 1 h, and this was followed bythe injection of 6 mg of AGE-RSA in a volume of 1.5 mL into theright eye. For the negative control, 3 mL of 200 mM puerarin wasinjected into the left eye. Based on the assumption that thevitreous volume of an adult rat eye is approximately 56 mL [17],the final intravitreal concentrations of AGE-RSA and puerarinwere approximately 100 mg/ml and 10 mM, respectively. Theneedle was left in position for 30 to 60 s and then slowlywithdrawn to minimize any fluid loss from the eye. The ratswere monitored regularly so that any infection associated withthe injection site could be detected. Any eye that sufferedinjection-damaged lenses or retinas was excluded from thestudy. There is no infection associated with the injection sitein all eyes. At 1 day after the intravitreal injection, the rats wereanesthetized and killed. All of the experiments were approved bythe Institutional Animal Care and Use Committee of the KoreaInstitute of Oriental Medicine.

Preparation of trypsin-digested vessels

The eyes were enucleated from the animals, and the retinaswere isolated. After fixation in 10% formalin for 2 day, the retinaswere incubated in trypsin (3% in sodium phosphate buffer con-taining 0.1 M sodium fluoride to inhibit the DNase activity) forapproximately 60 min. The vessel structures were isolated fromthe retinal cells by gentle rinsing in distilled water. The vascularspecimens were then mounted on slides.

Assessment of apoptosis

Apoptosis was assessed using a TUNEL staining protocolaccording to the manufacturer’s instructions (Promega). Thenumber of TUNEL-positive cells per unit area (mm2) was thendetermined in counted in a total of 5 fields. The numbers ofapoptotic and total cells were counted, and the percentages ofapoptotic cells were calculated.

Immunofluorescence staining

The trypsin digests were immunofluorescently stained aspreviously described [18]. The slides were incubated with amouse anti-8-hydroxygluanine (8-OHdG) antibody (Santa Cruz)and a mouse anti-NG2 antibody (Chemicon, Temecula, CA, USA)for 1 h. To detect 8-OHdG and NG2, the slides were incubatedwith a rodamine-conjugated goat anti-mouse antibody (SantaCruz). The oxidation of guanine to form 8-OHdG acts as a markerof oxidative DNA damage [19]. The antibodies against NG2 labelthe somas of pericytes [20]. The number of pericytes present in5 randomly selected fields was determined by counting thenumber of NG-2-positive cells per square millimeter of capillaryarea. The fluorescence intensity of 8-OHdG was analyzed in5 randomly selected square millimeter areas using ImageJ soft-ware (NIH, Bethesda MD USA).

Statistical analysis

The results were evaluated statistically using one-wayanalysis of variance followed by Tukey’s multiple comparisontest using GraphPad Prism 4.0 (GraphPad Software, San Diego,CA, USA).

Results

Puerarin inhibits pericyte apoptosis induced by AGE-BSA involving

ROS production

The retinal pericyte apoptosis induced by AGE-BSA was exam-ined in the presence of various concentrations of AGE-BSA (0, 10,50, 100, and 200 mg/mL) for 3 h (Fig. 1A) and in the presence of100 mg/mL AGE-BSA for 0, 1, 3, and 6 h (Fig. 1B). It was obviousthat AGE-BSA induced pericyte apoptosis in a time- and dose-dependent manner. The optimal response was achieved at100 mg/mL AGE-BSA. A relatively small increase in apoptosiswas observed when the concentration of AGE-BSA was doubledfrom 100 to 200 mg/mL. The earliest significant increase inpericyte apoptosis induced by AGE-BSA was observed after 1 hof incubation, and the rate of apoptosis reached a plateau at 3 hand remained stable thereafter. Pericyte apoptosis was furtherexamined by Western blot analysis and TUNEL staining after thepericytes were exposed to AGE-BSA. In the pericytes, the cleavageof PARP-1 and caspase-3 was evident when the cells were treatedwith 100 mg/mL AGE-BSA for 3 h (Fig. 1G), and the TUNEL-positivepericytes were highly detectable after they had been treated with100 mg/mL AGE-BSA for 3 h (Fig. 1H). Similarly, ROS production,which was assessed based on DCF fluorescence, significantlyincreased in a concentration-dependent manner after 3 h ofAGE-BSA treatment (Fig. 1D). The optimal response was observedwhen using 100 mg/mL AGE-BSA (which represents 105% of theBSA level). No further increase in ROS production was observedwhen the concentration of AGE-BSA was doubled from 100 to200 mg/mL. An acute stimulation of ROS production by AGE-BSAwas observed, and the earliest significant increase in ROS produc-tion was apparent after 1 h of incubation (Fig. 1E). For allsubsequent experiments, the pericytes were treated with100 mg/mL AGE-BSA for 3 h. To observe the effects of puerarinon AGE-BSA-induced pericyte apoptosis, various concentrationsof puerarin (1, 5, and 10 mM) were added 1 h before the pericyteswere stimulated for 3 h with 100 mg/mL AGE-BSA. As shown inFig. 1C, puerarin inhibited AGE-BSA-induced pericyte apoptosis ina dose-dependent manner, and the maximal inhibitory effect wasobserved at 10 mM. To determine whether the inhibitory roles ofpuerarin in AGE-BSA-induced pericyte apoptosis involved ROS,we observed the effects of an NADPH oxidase inhibitor (DPI,5 mM) and a free radical scavenger (NAC, 1 mM) on AGE-BSA-induced pericyte apoptosis (Fig. 1C). Puerarin (10 mM) markedlyinhibited the production of ROS induced by AGE-BSA, and similarinhibitory effects were observed using DPI and NAC (Fig. 1F).

Blockade of ROS production by puerarin via the inhibition of NADPH

oxidase

To explore the underlying mechanism by which puerarinsuppressed the intracellular oxidative stress, we measured theactivity of NADPH oxidase using lucigenin-enhanced chemilumi-nescence. Stimulation with 100 mg/mL AGE-BSA led to a time-dependent increase in NADPH oxidase activity; the increase at 3 hwas 6.5-fold (Fig. 2A). As demonstrated in Fig. 2B, pretreatment ofthe cells with puerarin reduced the AGE-BSA-dependent NADPHoxidase activation in a dose-dependent manner, and similarinhibitory effects were observed using DPI. NADPH oxidase iscomposed of cytosolic subunits including p47phox and Rac1 [21].Upon activation, the NADPH oxidase subunits become phosphory-lated and translocated from the cytosol to the plasma membrane;this action stimulates enzymatic activity and triggers the produc-tion of ROS [22]. In our experiments, the phosphorylation levelsof the NADPH oxidase subunits (p47phox and Rac1) weresignificantly increased in the pericytes 3 h after treatment with

Fig. 1. The effect of puerarin on AGE-BSA-induced apoptosis and the production of reactive oxygen species (ROS) in bovine retinal pericytes. (A and B) The pericytes were

exposed for 3 h to AGE-BSA (10 to 200 mg/mL) or BSA. (C and D) The pericytes were treated with 100 mg/mL AGE-BSA for 1 to 6 h. (E and F) The pericytes were pretreated

with puerarin, N-acetylcysteine (NAC) or diphenyleneiodonium (DPI) for 1 h, followed by treatment with 100 mg/mL AGE-BSA for 3 h. (A–C) Apoptotic cells were detected

using an FITC-labeled annexin V protein and flow cytometry. (D and E) Intracellular ROS were measured using a DCF-DA fluorescence. (D) Western blot analysis was used

to detect poly(ADP-ribose) polymerase-1 (PARP-1) and caspase-3. (E) TUNEL staining at 100�magnification. Each bar represents the mean7SE from four independent

experiments (nPo0.01 vs control, #Po0.01 vs AGE-BSA).

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365360

100 mg/mL AGE-BSA (Fig. 2C and D; Po0.01), indicating that thecomplex had been activated. In contrast, the AGE-BSA-inducedphosphorylations of p47phox and Rac1 were dramaticallydecreased in pericytes that were treated with puerarin (10 mM,Po0.01).

Blockade of redox-sensitive NF-KB activation by puerarin

To determine the downstream consequences of AGE-BSA-induced ROS production, we investigated whether AGE-BSA couldactivate the redox-sensitive nuclear transcription factor NF-KB. Inits inactive state, NF-KB is maintained in its latent form in thecytoplasm; this form masks the nuclear localization signal. Theexposure of the pericytes to 100 mg/mL AGE-BSA for 3 h resultedin the activation of NF-KB, which was demonstrated by theenhanced translocation of NF-KB to the nucleus (Fig. 3A). EMSAanalysis of the nuclear protein revealed a consistently increasedlevel in the DNA-binding activity of NF-KB after the 3-h AGE-BSAtreatment (a 2.3-fold increase versus the control; Po0.01;Fig. 3B). Puerarin significantly inhibited the nuclear translocationand DNA-binding activity of NF-KB. The supershift assays that

were performed using the anti-p50 and anti-p65 antibodiesconfirmed the specificity of the shifted bands. To further confirmthese data, we analyzed the NF-KB activity using an NF-KB-dependent promoter-based luciferase reporter assay (Fig. 3C).A significant increase in luciferase activity was seen followingAGE-BSA treatment compared with the control (Po0.01), whichwas significantly inhibited by puerarin (Po0.01). In addition,both DPI and puerarin significantly inhibited this increase inluciferase activity, which suggests that the AGE-BSA-stimulatedactivation of NF-KB was redox sensitive and occurred via theactivation of NADPH oxidase.

Puerarin prevents the apoptosis of retinal pericytes in intravitreally

AGE-RSA-injected rat eyes

To determine whether an increase in the levels of AGEs couldcontribute to increased retinal pericyte loss in vivo, we examinedwhether an increased concentration of AGE-RSA in the eye couldinduce pericyte apoptosis in normoglycemic rats. To study this,we intravitreally injected AGE-RSA into a normoglycemic rat eye.Immunohistofluorescence analysis for NG2, which is a marker

Fig. 2. The effect of puerarin on the AGE-BSA-induced activation of NADPH oxidase in bovine retinal pericytes. (A) The pericytes were treated with 100 mg/mL AGE-BSA for

1 to 3 h. (B) The pericytes were pretreated with puerarin or diphenyleneiodonium (DPI) for 1 h, followed by treatment with 100 mg/mL AGE-BSA for 3 h. (C and D) Western

blot analysis was used to detect phospho-p47phox and phospho-Rac1. Each bar represents the mean7SE from four independent experiments (nPo0.01 vs control,#Po0.01 vs AGE-BSA).

Fig. 3. The effect of puerarin on NF-kB activity in AGE-BSA-treated retinal pericytes. (A) Subcellular localization of the NF-kB p65 subunits. (B) Electrophoretic mobility

shift assay for NF-kB. (C) Pericytes were pretreated with puerarin or diphenyleneiodonium (DPI) for 1 h, followed by treatment with 100 mg/mL AGE-BSA for 3 h. The

NF-kB activity was measured using a luciferase assay. Each bar represents the mean7SE from four independent experiments (nPo0.01 vs control, #Po0.01 vs AGE-BSA).

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365 361

for pericytes, revealed AGE-RSA-induced pericyte loss in the ratretinal vessels. In the RSA-injected eyes, the NG2 immunoreac-tivity was concentrated in the cell body of the pericytes (Fig. 4A).

Analysis of the number of NG2-positive cell bodies on thecapillaries revealed that the pericyte density significantlydecreased in the AGE-RSA-injected eyes when compared with

Fig. 4. The effect of puerarin on retinal pericyte loss in intravitreally AGE-RSA-injected rat eyes. Immunofluorescence staining for NG2 (red) in trypsin-digested retinal

vessels from an RSA-injected eye (A), an AGE-RSA-injected eye (B), and a puerarin-treated eye (C). The number of pericytes was determined by counting the number of

NG2-positive cells per mm2 of capillary area. The values in the bar graphs represent the means7SE, n¼4. nPo0.01 vs the control group, #Po0.01 vs the AGE-RSA-

treated group.

Fig. 5. The effect of puerarin on retinal pericyte apoptosis in intravitreally AGE-RSA-injected rat eyes. The trypsin-digested retinal vessels from an RSA-injected eye (A), an

AGE-RSA-injected eye (B), and a puerarin-treated eye (C) were stained with TUNEL (green). (D) Quantitative analysis of the TUNEL-positive nuclei. The values in the bar

graphs represent the means7SE, n¼4. nPo0.01 vs the control group, #Po0.01 vs the AGE-RSA-treated group.

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365362

the control (Fig. 4B, Po0.01). However, the pretreatment withpuerarin significantly inhibited the AGE-RSA-dependent pericyteloss (Fig. 4C, Po0.01). To characterize injury of the retinalpericyte by AGE-RSA, TUNEL staining was performed usingtrypsin-digested retinal vessels. TUNEL analysis can detect cellsin which DNA is fragmenting and is therefore widely used as a

marker for apoptosis [23]. In the retinal trypsin digests of theRSA-injected eyes, a TUNEL-positive nucleus was rarely detected(Fig. 5A). In the AGE-RSA-injected eyes, many TUNEL-positivemicrovascular cells and fragmented nuclei were observed(Fig. 5B). However, treatment of the AGE-RSA-injected eyes withpuerarin prevented the increase in the positive cells that was seen

Fig. 6. Puerarin prevents oxidative DNA damage in retinal vessels derived from intravitreally AGE-RSA-injected rat eyes. The retinal tissues were stained with 8-OHdG,

which is a marker for oxidative DNA damage (red). Representative photomicrographs of the retinal vasculature from an RSA-injected eye (A), an AGE-RSA-injected eye (B),

and a puerarin-treated eye (C). Quantitative analysis of the immunofluorescence intensity for 8-OHdG. The data are expressed as the means7SE (n¼8). *Po0.01 vs the

control group, #Po0.01 vs the AGE-RSA-treated group.

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365 363

in normal eyes (Fig. 5C). To measure the oxidative damage thatoccurred in the AGE-RSA-injected eyes, the retinal vasculature wasimmunostained for 8-OHdG. A representative pattern of the immu-nohistochemical localization of 8-OHdG in the retinal vessel isshown in Fig. 6. The 8-OHdG marker exhibited nuclear and/orperinuclear localization in the retinal pericytes. Increased immunor-eactivity of 8-OHdG was observed in the AGE-RSA-injected eye(Fig. 6B), and treatment with puerarin suppressed the expression of8-OHdG compared to the AGE-RSA-injected eye (Fig. 6C).

Discussion

Growing evidence suggests that the overproduction of ROSmay be the key initiating event that leads to the long-termdevelopment of diabetic complications [24]. However, the specificmechanisms that link hyperglycemia with oxidative stress anddiabetic retinopathy are poorly understood. One of the majorconsequences of hyperglycemia is the formation of AGEs. AGEshave been implicated in the pathogenesis of diabetic retinopathy,and inhibiting the formation of AGEs can improve diabeticretinopathy [25–27]. In addition, the interaction between theAGE and the receptors for AGE (RAGE) elicits the generation ofROS and the expression of proinflammatory cytokines [25,26].AGEs have also been directly linked with the apoptotic cell deathof retinal pericytes [25,28,29]. AGE-induced apoptosis is mediatedby increasing oxidative stress or via proapoptotic cytokinesinduced by the AGE/RAGE interaction [30–32].

The present study provides strong evidence suggesting that aglycated protein stimulates ROS production in retinal pericyteslargely via the activation of NADPH oxidase. AGE-induced oxida-tive stress is believed to play a pivotal role in the retinal vascularresponse, the major pathogenesis of which is the loss of retinalpericytes [32,33]. Therefore, blocking the production of excessROS will significantly ameliorate vascular injuries in diabetics. Inthis study, we have demonstrated for the first time that puerarininhibits AGE-induced pericyte apoptosis in vitro and attenuates

pericyte damage in the intravitreally AGE-RSA-injected rat eyein vivo. In addition, long-term administration of puerarin (10 and50 mg/kg/day) also prevented several vascular alterations, such aspericyte ghost, acellular formation, and blood–retinal barrierbreakage, in streptozotocin-induced diabetic rats (electronic sup-plementary material [ESM]. This antiapoptotic effect of puerarinon retinal pericytes is mediated through the blocking of p45phox/Rac1-dependent NADPH oxidase activation and ROS production.These findings suggest that puerarin may possess protectivebeneficial effects against diabetic retinopathy.

In a previous study, puerarin was shown to protect againsthigh glucose-induced vascular dysfunction in the rat thoracicaorta [34]. Pharmacological research has demonstrated thatpuerarin exerts a protective effect against myocardial reperfusioninjury [35] and diabetic retinopathy [36]. In addition, puerarinpossesses antioxidant properties, such as its ability to scavengefor ROS [37]. Despite its known antidiabetic activity, the effects ofpuerarin on diabetic retinal vascular complications have not yetbeen clarified.

In our in vitro system, the effective dose of puerarin to treatretinal pericyte damage was as low as 1 mM, and the maximaldose was 10 mM. Puerarin was also effective in vivo at a dose of10 mM. Although some reports have described the pharmacoki-netics of puerarin, the results have been contradictory [38,39].Zhu et al. reported that the peak plasma concentrations ofpuerarin following the oral consumption of 100 mg/kg puerarinin male Zucker rats was 13.83 mM [40], and Prasain et al. reportedthat the peak plasma concentration of puerarin following the oralconsumption of 50 mg/kg puerarin in male spontaneously hyper-tensive rats was 8.51 mM [41]. Based on previous reports and ourresults, we chose a dose of 10 mM puerarin to mimic the in vivoconcentrations of puerarin. Our finding regarding the effectivedose of puerarin in vitro is consistent with the in vivo results.

Puerarin is currently available in the market as oral prepara-tions, such as pellets, granules, and capsules. However, thesepreparations have low bioavailability in human [42]. The peakplasma concentration of puerarin after oral administration of

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365364

400 mg puerarin to a healthy volunteer was 0.13 mM [43]. Thepeak plasma concentration of puerarin following the oral con-sumption of 2 g kudzu extract containing 95 mg puerarin in adultmales was 0.11 mM [44]. This peak plasma concentration ofpuerarin after oral administration in human was 100-fold lowerthan that observed in rats. Due to poor oral bioavailability, theintravenous administration of puerarin is widely used in a clinicalsetting [45]. But the clinical efficacy of puerarin intravenousinjection is limited by severe and acute toxic side effects, suchas intravenous hemolysis. It is very necessary to seek a noveldelivery system for puerarin. To improve the oral bioavailabilityof puerarin, a puerarin microemulsion system and puerarinliposome were successfully developed [45–47]. When microemul-sion formulation of puerarin was dosed in mice, the oral bioavail-ability of puerarin microemulsion was 15.82-fold higher than thatof puerarin [46].

It is still unclear how puerarin could work well at a relativelylower dose. Because puerarin is an isoflavone C-glycoside, it ismore resistant to metabolic deactivation and enzymatic hydro-lysis than the O-glycosides daidzein and genistein. Recent studieshave demonstrated that daidzein and genistein are subjected toextensive metabolism in the gut and liver, and this may affecttheir biological properties. These effects include decreases in theirantioxidant activities and reduced effects on inflammation andcell adhesion [48,49]. Therefore, the presence of low levels ofpuerarin metabolites in tissues may have particular relevance toits strong beneficial function.

The mechanisms by which puerarin exerts its antiapoptoticeffect remain unclear. Considerable evidence suggests that anincreased level of ROS is responsible for hyperglycemia- andAGEs-induced retinal pericyte apoptosis [32,33]. Therefore, weinvestigated whether the antiapoptotic effect of puerarin wasmediated through the inhibition of ROS generation. Our datashowed that puerarin significantly inhibited ROS overproductionin vitro and inhibited oxidative damage in vivo. NADPH oxidasemay be the most important source of ROS production in vascularcells [50]. In the present study, DPI (which is a specific inhibitor ofthe NADPH-dependent oxidase) markedly suppressed ROS over-production and NADPH oxidase activation in the AGE-BSA-treatedpericytes. This result suggests that the main enzyme source forROS production in the AGE-BSA-treated pericytes is derived fromNADPH oxidase activity. Importantly, DPI and the classicalantioxidant NAC simulated the inhibitory effect of puerarinon AGE-BSA-induced pericyte apoptosis, suggesting that theantiapoptotic effect of puerarin is related to the inhibition of theNADPH oxidase-dependent overproduction of ROS.

NADPH oxidase comprises a membrane-integrated cyto-chrome b558 that is composed of gp91phox, p22phox, and severalcytosolic regulatory subunits (p47phox, p40phox, p67phox, andthe small GTP-binding protein Rac1 or Rac2) [51]. The activationof these NADPH oxidase subunits is required for the activationof AGE-BSA-induced NADPH oxidase and ROS production. Inour study, treatment with puerarin significantly inhibited theAGE-BSA-induced phosphorylation of p47phox and Rac1; thisresult suggested a possible mechanism for the inhibitory effectof puerarin on NADPH oxidase activation. A crucial role of Rac1has been demonstrated in diabetes-induced vascular injury [52].It has also been reported that the activiaton of p47phox andp67phox activation are dependent on Rac1 [40]. A failure tophosphorylate p47phox results in the failure to produce super-oxide in stimulated cells [53]. Therefore, the inhibition of Rac1 bypuerarin may be responsible for its inhibitory effect on theactivation of p47phox and subsequent NADPH oxidase activationand ROS production.

The activation of the transcription factor NF-kB is importantfor the regulation of genes in response to cellular stress [54,55].

NF-kB is recognized as an important redox-sensitive transcriptionfactor and has been implicated in the development of diabeticretinopathy [56]. The hyperglycemic activation of NF-kB inducesaccelerated pericyte loss via the induction of NF-kB-controlledproapoptotic molecules [56]. We have demonstrated thatAGE-BSA stimulates the translocation of NF-KB to the nucleusand that puerarin can almost completely inhibit this transloca-tion. Furthermore, puerarin and DPI inhibited AGE-BSA-inducedNF-KB activity. Collectively, these data strongly suggest that ROSplay an important role in AGE-BSA-induced NF-KB activation.

In conclusion, this study has clearly demonstrated that puer-arin attenuates the AGE-BSA-induced apoptosis of retinal peri-cytes in vitro through blocking ROS production and the p47phox-and Rac1-dependent NADPH oxidase activation. Using the ratintravitreal AGE-RSA injection model to assess the efficacy ofpuerarin in vivo, we also found that puerarin attenuated theconsequent increase in oxidative damage in the retinal vesselsand inhibited pericyte apoptosis. These results suggest thatpuerarin may be an effective therapeutic candidate againstdiabetes-associated retinal pericyte loss.

Acknowledgment

This research was supported by a grant [K10040, K11040] fromthe Korea Institute of Oriental Medicine (KIOM).

Appendix A. Supplementary materials

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.freeradbiomed.2012.04.030.

References

[1] Chen, B. H.; Jiang, D. Y.; Tang, L. S. Advanced glycation end-products induceapoptosis involving the signaling pathways of oxidative stress in bovineretinal pericytes. Life Sci 79:1040–1048; 2006.

[2] Yatoh, S.; Mizutani, M.; Yokoo, T.; Kozawa, T.; Sone, H.; Toyoshima, H.;Suzuki, S.; Shimano, H.; Kawakami, Y.; Okuda, Y.; Yamada, N. Antioxidantsand an inhibitor of advanced glycation ameliorate death of retinal micro-vascular cells in diabetic retinopathy. Diabetes Metab. Res. Rev 22:38–45;2006.

[3] Wisse, B. E.; Kim, F.; Schwartz, M. W. Physiology. An integrative view ofobesity. Science 318:928–929; 2007.

[4] Brem, H.; Tomic-Canic, M. Cellular and molecular basis of wound healing indiabetes. J. Clin. Invest. 117:1219–1222; 2007.

[5] Rees, M. D.; Kennett, E. C.; Whitelock, J. M.; Davies, M. J. Oxidative damage toextracellular matrix and its role in human pathologies. Free Radic. Biol. Med.44:1973–2001; 2008.

[6] Kowluru, R. A. Diabetic retinopathy: mitochondrial dysfunction and retinalcapillary cell death. Antioxid. Redox Signal. 7:1581–1587; 2005.

[7] Manea, A.; Constantinescu, E.; Popov, D.; Raicu, M. Changes in oxidativebalance in rat pericytes exposed to diabetic conditions. J. Cell. Mol. Med.8:117–126; 2004.

[8] Cherdshewasart, W.; Subtang, S.; Dahlan, W. Major isoflavonoid contents ofthe phytoestrogen rich-herb Pueraria mirifica in comparison with Puerarialobata. J. Pharm. Biomed. Anal. 43:428–434; 2007.

[9] Kirakosyan, A.; Kaufman, P. B.; Warber, S.; Bolling, S.; Chang, S.; Duke, J. A.Quantification fo major isoflavonoids and L-canavanine in several organs ofKudzu vine (Pueraria lobata) and in starch samples derived from Kudzu roots.Plant Sci 164:883–888; 2003.

[10] Hsu, F. L.; Liu, I. M.; Kuo, D. H.; Chen, W. C.; Su, H. C.; Cheng, J. T.Antihyperglycemic effect of puerarin in streptozotocin-induced diabetic rats.J. Nat. Prod. 66:788–792; 2003.

[11] Xiong, F. L.; Sun, X. H.; Gan, L.; Yang, X. L.; Xu, H. B. Puerarin protects ratpancreatic islets from damage by hydrogen peroxide. Eur. J. Pharmacol.529:1–7; 2006.

[12] Kim, J. M.; Lee, Y. M.; Lee, G. Y.; Jang, D. S.; Bae, K. H.; Kim, J. S. Constituents ofthe roots of Pueraria lobata inhibit formation of advanced glycation endproducts (AGEs). Arch. Pharm. Res. 29:821–825; 2006.

J. Kim et al. / Free Radical Biology and Medicine 53 (2012) 357–365 365

[13] Capetandes, A.; Gerritsen, M. E. Simplified methods for consistent andselective culture of bovine retinal endothelial cells and pericytes. Invest.Ophthalmol. Vis. Sci. 31:1738–1744; 1990.

[14] Kondo, T.; Hosoya, K.; Hori, S.; Tomi, M.; Ohtsuki, S.; Takanaga, H.;Nakashima, E.; Iizasa, H.; Asashima, T.; Ueda, M.; Obinata, M.; Terasaki, T.Establishment of conditionally immortalized rat retinal pericyte celllines (TR-rPCT) and their application in a co-culture system using retinalcapillary endothelial cell line (TR-iBRB2). Cell Struct. Funct. 28:145–153;2003.

[15] Kim, K. M.; Jung, D. H.; Jang, D. S.; Kim, Y. S.; Kim, J. M.; Kim, H. N.; Surh, Y. J.;Kim, J. S. Puerarin suppresses AGEs-induced inflammation in mouse mesan-gial cells: a possible pathway through the induction of heme oxygenase-1expression. Toxicol. Appl. Pharmacol. 244:106–113; 2010.

[16] Vlassara, H.; Striker, L. J.; Teichberg, S.; Fuh, H.; Li, Y. M.; Steffes, M. Advancedglycation end products induce glomerular sclerosis and albuminuria innormal rats. Proc. Natl. Acad. Sci. USA 91:11704–11708; 1994.

[17] Berkowitz, B. A.; Lukaszew, R. A.; Mullins, C. M.; Penn, J. S. Impaired hyaloidalcirculation function and uncoordinated ocular growth patterns in experi-mental retinopathy of prematurity. Invest. Ophthalmol. Vis. Sci. 39:391–396;1998.

[18] Kim, O. S.; Kim, J.; Kim, C. S.; Kim, N. H.; Kim, J. S. KIOM-79 preventsmethyglyoxal-induced retinal pericyte apoptosis in vitro and in vivo.J. Ethnopharmacol. 129:285–292; 2010.

[19] Beckman, K. B.; Ames, B. N. Oxidative decay of DNA. J. Biol. Chem.272:19633–19636; 1997.

[20] Hughes, S.; Chan-Ling, T. Characterization of smooth muscle cell and pericytedifferentiation in the rat retina in vivo. Invest. Ophthalmol. Vis. Sci45:2795–2806; 2004.

[21] Cross, A. R.; Segal, A. W. The NADPH oxidase of professional phagocytes—

prototype of the NOX electron transport chain systems. Biochim. Biophys. Acta1657:1–22; 2004.

[22] Wu, D. C.; Teismann, P.; Tieu, K.; Vila, M.; Jackson-Lewis, V.; Ischiropoulos,H.; Przedborski, S. NADPH oxidase mediates oxidative stress in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine model of Parkinson’s disease. Proc. Natl.Acad. Sci. USA 100:6145–6150; 2003.

[23] Charriaut-Marlangue, C.; Ben-Ari, Y. A cautionary note on the use of theTUNEL stain to determine apoptosis. Neuroreport 7:61–64; 1995.

[24] Rosen, P.; Nawroth, P. P.; King, G.; Moller, W.; Tritschler, H. J.; Packer, L. Therole of oxidative stress in the onset and progression of diabetes and itscomplications: a summary of a Congress Series sponsored by UNESCO-MCBN,the American Diabetes Association and the German Diabetes Society.Diabetes Metab. Res. Rev 17:189–212; 2001.

[25] Yamagishi, S.; Amano, S.; Inagaki, Y.; Okamoto, T.; Koga, K.; Sasaki, N.;Yamamoto, H.; Takeuchi, M.; Makita, Z. Advanced glycation end products-induced apoptosis and overexpression of vascular endothelial growthfactor in bovine retinal pericytes. Biochem. Biophys. Res. Commun. 290:973–978; 2002.

[26] Yamagishi, S.; Hsu, C. C.; Taniguchi, M.; Harada, S.; Yamamoto, Y.; Ohsawa,K.; Kobayashi, K.; Yamamoto, H. Receptor-mediated toxicity to pericytes ofadvanced glycosylation end products: a possible mechanism of pericyte lossin diabetic microangiopathy. Biochem. Biophys. Res. Commun. 213:681–687;1995.

[27] Yamagishi, S.; Nakamura, K.; Matsui, T. Advanced glycation end products(AGEs) and their receptor (RAGE) system in diabetic retinopathy. Curr. DrugDiscov. Technol 3:83–88; 2006.

[28] Denis, U.; Lecomte, M.; Paget, C.; Ruggiero, D.; Wiernsperger, N.; Lagarde, M.Advanced glycation end-products induce apoptosis of bovine retinal peri-cytes in culture: involvement of diacylglycerol/ceramide production andoxidative stress induction. Free Radic Biol. Med. 33:236–247; 2002.

[29] Mizutani, M.; Kern, T. S.; Lorenzi, M. Accelerated death of retinal micro-vascular cells in human and experimental diabetic retinopathy. J. Clin. Invest.97:2883–2890; 1996.

[30] Kaji, Y.; Amano, S.; Usui, T.; Oshika, T.; Yamashiro, K.; Ishida, S.; Suzuki, K.;Tanaka, S.; Adamis, A. P.; Nagai, R.; Horiuchi, S. Expression and function ofreceptors for advanced glycation end products in bovine corneal endothelialcells. Invest. Ophthalmol. Vis. Sci. 44:521–528; 2003.

[31] Kasper, M.; Roehlecke, C.; Witt, M.; Fehrenbach, H.; Hofer, A.; Miyata, T.;Weigert, C.; Funk, R. H.; Schleicher, E. D. Induction of apoptosis by glyoxal inhuman embryonic lung epithelial cell line L132. Am. J. Respir. Cell Mol. Biol.23:485–491; 2000.

[32] Yamagishi, S.; Nakamura, K.; Matsui, T.; Inagaki, Y.; Takenaka, K.; Jinnouchi,Y.; Yoshida, Y.; Matsuura, T.; Narama, I.; Motomiya, Y.; Takeuchi, M.; Inoue,H.; Yoshimura, A.; Bucala, R.; Imaizumi, T. Pigment epithelium-derived factorinhibits advanced glycation end product-induced retinal vascular hyperper-meability by blocking reactive oxygen species-mediated vascular endothelialgrowth factor expression. J. Biol. Chem. 281:20213–20220; 2006.

[33] Yamagishi, S.; Inagaki, Y.; Amano, S.; Okamoto, T.; Takeuchi, M.; Makita, Z.Pigment epithelium-derived factor protects cultured retinal pericytes fromadvanced glycation end product-induced injury through its antioxidativeproperties. Biochem. Biophys. Res. Commun. 296:877–882; 2002.

[34] Meng, X. H.; Ni, C.; Zhu, L.; Shen, Y. L.; Wang, L. L.; Chen, Y. Y. Puerarinprotects against high glucose-induced acute vascular dysfunction: role ofheme oxygenase-1 in rat thoracic aorta. Vascul. Pharmacol 50:110–115; 2009.

[35] Fan, L. L.; Sun, L. H.; Li, J.; Yue, X. H.; Yu, H. X.; Wang, S. Y. The protectiveeffect of puerarin against myocardial reperfusion injury. Study on cardiacfunction. Chin. Med. J. 105:11–17; 1992.

[36] Teng, Y.; Cui, H.; Yang, M.; Song, H.; Zhang, Q.; Su, Y.; Zheng, J. Protective effect ofpuerarin on diabetic retinopathy in rats. Mol. Biol. Rep. 36:1129–1133; 2009.

[37] Guerra, M. C.; Speroni, E.; Broccoli, M.; Cangini, M.; Pasini, P.; Minghett, A.;Crespi-Perellino, N.; Mirasoli, M.; Cantelli-Forti, G.; Paolini, M. Comparisonbetween chinese medical herb Pueraria lobata crude extract and itsmain isoflavone puerarin antioxidant properties and effects on rat liverCYP-catalysed drug metabolism. Life Sci. 67:2997–3006; 2000.

[38] Qian, X. W.; Chen, X. Z.; Li, D. B. Low-dose ropivacaine-sufentanil spinalanaesthesia for caesarean delivery: a randomised trial. Int. J. Obstet. Anesth.17:309–314; 2008.

[39] Prasain, J. K.; Peng, N.; Moore, R.; Arabshahi, A.; Barnes, S.; Wyss, J. M. Tissuedistribution of puerarin and its conjugated metabolites in rats assessed by liquidchromatography-tandem mass spectrometry. Phytomedicine 16:65–71; 2009.

[40] Zhu, L. H.; Wang, L.; Wang, D.; Jiang, H.; Tang, Q. Z.; Yan, L.; Bian, Z. Y.;Wang, X. A.; Li, H. Puerarin attenuates high-glucose-and diabetes-inducedvascular smooth muscle cell proliferation by blocking PKCbeta2/Rac1-depen-dent signaling. Free Radic. Biol. Med. 48:471–482; 2010.

[41] Prasain, J. K.; Peng, N.; Acosta, E.; Moore, R.; Arabshahi, A.; Meezan, E.;Barnes, S.; Wyss, J. M. Pharmacokinetic study of puerarin in rat serum byliquid chromatography tandem mass spectrometry. Biomed. Chromatogr.21:410–414; 2007.

[42] Wu, H.; Zhou, A.; Lu, C.; Wang, L. Examination of lymphatic transport ofpuerarin in unconscious lymph duct-cannulated rats after administration inmicroemulsion drug delivery systems. Eur. J. Pharm. Sci. 42:348–353; 2011.

[43] Ma, Z.; Wu, Q.; Lee, D. Y.; Tracy, M.; Lukas, S. E. Determination of puerarin inhuman plasma by high performance liquid chromatography. J. Chromatogr. BAnalyt. Technol. Biomed. Life Sci. 823:108–114; 2005.

[44] Qin, F.; Huang, X.; Zhang, H. M.; Ren, P. Pharmacokinetic comparisonof puerarin after oral administration of Jiawei-Xiaoyao-San to healthyvolunteers and patients with functional dyspepsia: influence of disease state.J. Pharm. Pharmacol 61:125–129; 2009.

[45] Yue, P. F.; Yuan, H. L.; Li, X. Y.; Yang, M.; Zhu, W. F.; Xiao, X. H. Developmentand optimization of intravenous puerarin emulsions formation by a novelcomplex-phase inversion-homogenization technology. Chem. Pharm. Bull.55:1563–1568; 2007.

[46] Wu, H.; Lu, C.; Zhou, A.; Min, Z.; Zhang, Y. Enhanced oral bioavailability ofpuerarin using microemulsion vehicle. Drug Dev. Ind. Pharm 35:138–144;2009.

[47] Gu, Y. Z.; Zhou, W.; Zhai, G. X. Preparation of puerarin liposome and its oralabsorption in rat. Zhong Yao Cai 30:970–973; 2007.

[48] Rimbach, G.; De Pascual-Teresa, S.; Ewins, B. A.; Matsugo, S.; Uchida, Y.;Minihane, A. M.; Turner, R.; VafeiAdou, K.; Weinberg, P. D. Antioxidant andfree radical scavenging activity of isoflavone metabolites. Xenobiotica33:913–925; 2003.

[49] Turner, R.; Baron, T.; Wolffram, S.; Minihane, A. M.; Cassidy, A.; Rimbach, G.;Weinberg, P. D. Effect of circulating forms of soy isoflavones on the oxidationof low density lipoprotein. Free Radic. Res. 38:209–216; 2004.

[50] Inoguchi, T.; Sonta, T.; Tsubouchi, H.; Etoh, T.; Kakimoto, M.; Sonoda, N.; Sato,N.; Sekiguchi, N.; Kobayashi, K.; Sumimoto, H.; Utsumi, H.; Nawata, H.Protein kinase C-dependent increase in reactive oxygen species (ROS)production in vascular tissues of diabetes: role of vascular NAD(P)H oxidase.J. Am. Soc. Nephrol. 14:S227–232; 2003.

[51] Groemping, Y.; Lapouge, K.; Smerdon, S. J.; Rittinger, K. Molecular basis ofphosphorylation-induced activation of the NADPH oxidase. Cell 113:343–355; 2003.

[52] Vecchione, C.; Aretini, A.; Marino, G.; Bettarini, U.; Poulet, R.; Maffei, A.;Sbroggio, M.; Pastore, L.; Gentile, M. T.; Notte, A.; Iorio, L.; Hirsch, E.; Tarone,G.; Lembo, G. Selective Rac-1 inhibition protects from diabetes-inducedvascular injury. Circ. Res. 98:218–225; 2006.

[53] El Benna, J.; Faust, R. P.; Johnson, J. L.; Babior, B. M. Phosphorylation of therespiratory burst oxidase subunit p47phox as determined by two-dimen-sional phosphopeptide mapping. Phosphorylation by protein kinase C,protein kinase A, and a mitogen-activated protein kinase. J. Biol. Chem271:6374–6378; 1996.

[54] Purcell, N. H.; Tang, G.; Yu, C.; Mercurio, F.; DiDonato, J. A.; Lin, A. Activationof NF-kappa B is required for hypertrophic growth of primary rat neonatalventricular cardiomyocytes. Proc. Natl. Acad. Sci. USA 98:6668–6673; 2001.

[55] Higuchi, Y.; Otsu, K.; Nishida, K.; Hirotani, S.; Nakayama, H.; Yamaguchi, O.;Matsumura, Y.; Ueno, H.; Tada, M.; Hori, M. Involvement of reactive oxygenspecies-mediated NF-kappa B activation in TNF-alpha-induced cardiomyo-cyte hypertrophy. J. Mol. Cell. Cardiol. 34:233–240; 2002.

[56] Romeo, G.; Liu, W. H.; Asnaghi, V.; Kern, T. S.; Lorenzi, M. Activation ofnuclear factor-kappaB induced by diabetes and high glucose regulates aproapoptotic program in retinal pericytes. Diabetes 51:2241–2248; 2002.