Journal of Molecular and Cellular Cardiology 41 (2006) 149–159www.elsevier.com/locate/yjmcc

Original article

PARP inhibition prevents postinfarction myocardial remodeling and heartfailure via the protein kinase C/glycogen synthase kinase-3β pathway☆

Anita Palfi a,b, Ambrus Toth b, Katalin Hanto a,b, Peter Deres a,b, Eszter Szabados a,Zoltan Szereday c, Gyozo Kulcsar d, Tamas Kalai d, Kalman Hideg d, Ferenc Gallyas Jr. b,

Balazs Sumegi b,e, Kalman Toth a,⁎, Robert Halmosi a

a First Department of Medicine, Division of Cardiology, School of Medicine, University of Pecs, 13 Ifjusag str., H-7624 Pecs, Hungaryb Department of Biochemistry and Medical Chemistry, School of Medicine, University of Pecs, Pecs, Hungary

c Department of Dermatology, School of Medicine, University of Pecs, Pecs, Hungaryd Department of Organic and Medicinal Chemistry, School of Medicine, University of Pecs, Pecs, Hungary

e Research Group for Mitochondrial Function and Mitochondrial Diseases, Hungarian Academy of Sciences, Budapest, Hungary

Received 15 September 2005; received in revised form 17 March 2006; accepted 23 March 2006Available online 23 May 2006

Abstract

The inhibition of glycogen synthase kinase-3β (GSK-3β) via phosphorylation by Akt or protein kinase C (PKC), or the activation of mitogen-activated protein kinase (MAPK) cascades can play a pivotal role in left ventricular remodeling following myocardial infarction. Our previous datashowed that MAPK and phosphatidylinositol-3-kinase/Akt pathways could be modulated by poly(ADP-ribose)polymerase (PARP) inhibitionraising the possibility that cardiac hypertrophic signaling responses may be favorably influenced by PARP inhibitors.

A novel PARP inhibitor (L-2286) was tested in a rat model of chronic heart failure following isoproterenol-induced myocardial infarction.Subsequently, cardiac hypertrophy and interstitial collagen deposition were assessed; additionally, mitochondrial enzyme activity and thephosphorylation state of GSK-3β, Akt, PKC and MAPK cascades were monitored.

PARP inhibitor (L-2286) treatment significantly reduced the progression of postinfarction heart failure attenuating cardiac hypertrophy andinterstitial fibrosis, and preserving the integrity of respiratory complexes. More importantly, L-2286 repressed the hypertrophy-associated increasedphosphorylation of panPKC, PKC α/βII, PKC δ and PKC ε, which could be responsible for the activation of the antihypertrophic GSK-3β.

This work provides the first evidence that PARP inhibition beneficially modulates the PKC/GSK-3β intracellular signaling pathway in a ratmodel of chronic heart failure identifying a novel drug target to treat heart failure.© 2006 Elsevier Inc. All rights reserved.

Keywords: PARP inhibition; Intracellular signaling; Glycogen synthase kinase-3β; Heart failure; Protein kinase C; Ventricular remodeling

Abbreviations: BNP, B-type natriuretic peptide; BW, body weight; ERK1/2,extracellular signal-regulated kinase; GSK-3β, glycogen synthase kinase-3β;HE, hematoxylin and eosin; ISO, isoproterenol hydrochloride; JNK, c-jun N-terminal kinase; MAPK, mitogen activated protein kinase; NAD+, nicotinamideadenine dinucleotide; NF-κB, nuclear factor-κB; NIH, National Institutes ofHealth; PARP, poly(ADP-ribose) polymerase; PDC-1α, pyruvate dehydroge-nase complex-1α; PI3K, phosphatidylinositol-3-kinase; PKC, protein kinase C;ROS, reactive oxygen species; SEM, standard error of mean; TBS, Tris-bufferedsaline; TEF-1, transcriptional enhancer factor-1; TL, length of right tibia; WV,weight of ventricles☆ List of special characters: β, μ, ±, <.⁎ Corresponding author. Fax: +36 72 536 146.E-mail address: kalman.toth@aok.pte.hu (K. Toth).

0022-2828/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.yjmcc.2006.03.427

1. Introduction

Increased activation of poly(ADP-ribose) polymerase (PARP)enzyme is a crucial step in the development of oxidative stress-induced cell dysfunction and tissue injury [1,2]. Oxidative stressalso plays a pathogenic role in chronic heart failure that can becharacterized by altered intracellular signaling and geneexpression [3–7]. Several studies demonstrated that inhibitionof PARP enzyme can efficiently reduce oxidative myocardialdamage; nevertheless, little is known about the mechanism ofcardioprotection by PARP inhibitors in chronic heart failure[1,2,8]. This molecular mechanism is worth elucidating because

150 A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

PARP inhibition already represents a potential clinical approachto fight against myocardial remodeling: as we have previouslyreported, carvedilol, an antiadrenergic drug used, among others,to treat the hypertrophic and failing heart, exhibited antioxidantand PARP-inhibitory properties [9,10]. It has been widely knownthat the beneficial effects of PARP enzyme blockade in oxidativeinjuries partially rely on the preservation of the cellular NAD+

and ATP pools [1,2]. Recent studies, however, suggested thatPARP inhibition could beneficially influence various intracellularsignaling routes further contributing to the cardioprotective effectof these agents [2,8,11,12]. Although PARP enzyme-mediateddirect protein interaction and poly(ADP-ribosyl)ation wereshown to influence the function of a variety of transcriptionfactors and to interfere with the expression of severalproinflammatory genes, to date, limited information is availableon how PARP inhibition might modulate intracellular signalingduring the progression of postinfarction heart failure [11].

Remodeling of the left ventricle following myocardialinfarction ultimately progresses into heart failure, which is aleading cause of mortality worldwide [5]. Cardiac remodeling isassociated with alterations in intracellular signaling includingthe inhibition of glycogen synthase kinase-3β (GSK-3β) viaphosphorylation by Akt, protein kinase C (PKC), p70-S6kinase, p90-RSK and protein kinase A [13]. These pathwaysand the mitogen-activated protein kinases (MAPK) have allbeen demonstrated to alter their activation state in response tohypertrophic stimuli and may therefore contribute to myocardialhypertrophy [5–7]. We have recently found that inhibition ofPARP protected cells from oxidative stress by modulating thephosphorylation and activation of Akt and/or MAP kinases in acell culture system, ex vivo and in vivo animal models ofmyocardial injury and also in a rodent septic shock model.These findings also suggest that PARP inhibition couldmodulate these signaling pathways which are also participantsin the development of cardiac hypertrophy [17,18]. Notably,ventricular remodeling is also accompanied by changes inenergy metabolism, mitochondrial dysfunction and by in-creased expression of fetal genes and extracellular matrixcomponents [14–16].

To elucidate the role of protein kinase signaling in themechanism of cardioprotection afforded by PARP inhibition, L-2286 (Fig. 1), a novel quinazoline-type PARP inhibitor wastested in a rat model of chronic heart failure after isoproterenol-induced myocardial infarction [8]. L-2286 attenuated themyocardial hypertrophy and interstitial deposition of type IIIcollagen, and beneficially influenced the mitochondrial enzymeactivity in the failing heart. In addition, our study firstdemonstrated that PARP inhibition could reduce the progression

Fig. 1. Chemical structure of L-2286 (2-[(2-Piperidin-1-yletil)thio]quinazolin-4(3H)-one).

of chronic myocardial remodeling possibly via its ability tointerfere with PKC-mediated GSK-3β inhibition.

2. Materials and methods

2.1. Chronic heart failure model

Male CFY-strain Sprague–Dawley rats (350–380 g) receivedtwo subcutaneous injections (separated by a 24-h interval) of80 mg/kg isoproterenol (ISO, Sigma-Aldrich Co, Budapest,Hungary) [19]. Twenty-four hours after the second injection, thesurviving animals were randomly assigned to receive either5 mg/kg/die L-2286 (n = 15), an effective and selective water-soluble PARP inhibitor, or water (n = 12). PARP inhibitortreatment was delayed 24 h to avoid suppression of infarct sizeby the PARP inhibition [20,21]. L-2286 was given for 8 weeks.The drug-only control group (C + L-2286) received 5 mg/kg/dieL-2286 for 8 weeks. At the end of the 8-week period, bodyweights were measured and standard electrocardiogram (ECG)was recorded to determine R wave amplitude and J pointdeviation (lead II) (Schiller AG Electrocardiograph, Switzer-land). Animals were subsequently sacrificed, their hearts wereremoved, the atria and great vessels were trimmed from theventricles and the weight of the ventricles was measured. It wasthen normalized to the body mass and the length of right tibia.Hearts were freeze-clamped or fixed in 10% formalin. Theinvestigation conformed to the Guide for the Care and Use ofLaboratory Animals published by the U.S. National Institutes ofHealth (NIH Publication No. 85–23, revised 1996), and wasapproved by the Animal Research Review Committee of theUniversity of Pecs Medical School.

2.2. Determination of plasma B-type natriuretic peptide

Blood samples were collected into the Lavender Vacutanertubes containing EDTA and aprotinin (0.6 TIU/ml of blood),centrifuged at 1600×g for 15 min at 4 °C. Supernatants (plasma)were collected and kept at −70 °C. Plasma B-type natriureticpeptide-45 (BNP-45) levels were determined by enzymeimmunoassay method (BNP-45, Rat EIA Kit, Phoenix Phar-maceuticals Inc., CA, USA).

2.3. Measurement of mitochondrial enzyme activity

NADH:cytochrome c oxidoreductase was measured asdescribed previously [22]. Enzyme activity was determinedby measuring the rate of cytochrome c reduction at 550 nmin a medium containing 50 mmol/l sodium-phosphate,1 mmol/l sodium-azide, 1.5 mM NADH and 50–75 μgmitochondrial protein/ml, pH 7.5. The reaction was startedby addition of 40 μl cytochrome c.

2.4. Histology

Ventricles fixed in formalin were sliced and embedded inparaffin. Sections (5-μm thick) were cut serially from base toapex. Ten to twelve slices at 1-mm intervals were stained with

151A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

hematoxylin and eosin (HE). The sections were mounted onslides and projected at a magnification of 400× and photo-micrographs were taken. Mean myocyte diameters on HEstained sections were calculated by measuring 100 cells perspecimen in the region of the cell nucleus using the two-pointdistance function of the TelPath analyzer system (bollman.com,2000). Type III collagen was stained as a marker of interstitialfibrosis on frozen sections, 5-μm thick by the Vector M.O.M.™Kit (Vector Laboratories Inc., Burlingame, CA, USA) stainingprocedure. After fixing and dehydrating, sections were washedin Tris-buffered saline (TBS) containing 0.5% Tween 20, pH7.6. To block the endogenous peroxidase activity, sections wereincubated in 3%H2O2 then washed in TBS. After 1 h incubationwith M.O.M.™ mouse IgG blocking reagent (containing 20%normal rat serum), sections were washed in TBS, and incubatedin the working solution of M.O.M.™ diluent for 5 min. Primarymouse antisera against type III collagen (1:1000, MonoclonalAnti-Collagen, Type III, Sigma-Aldrich Co, Budapest, Hun-gary) diluted in M.O.M.™ diluent reacted at room temperaturefor 30 min, followed by two 2-min rinses in TBS. BiotinylatedAnti-Mouse IgG Reagent was then applied for 10 min andsections were washed twice for 2 min in TBS. VECTASTATINABC Reagent was applied for 5 min that was followed by two2-min rinses in TBS. Sections were then stained with VectorNovaRED Substrate (Vector Laboratories Inc., Burlingame,CA, USA) for 5 min, washed in distilled water, dehydrated,mounted on slides and projected at a magnification of 100×.Sections were quantified with the NIH ImageJ analyzer system.All histological and immunohistological samples were exam-ined by an investigator in a blinded fashion.

2.5. Cell culture model

H9c2 (2-1) cardiomyoblasts were from American TypeCulture Collection (Wesel, Germany). The cells were main-tained as monolayer adherent culture in Minimum EssentialEagle's Medium containing 1% antibiotic–antimycotic solu-tion, 1 mM sodium pyruvate and 10% fetal calf serum in humid5% CO2 atmosphere at 37 °C. Rat neonatal primary cardio-myocytes were prepared exactly as described byWang et al. andcultured overnight in humid 5% CO2 atmosphere at 37 °C [23].The cells were treated for 1 or 24 h with 1 μM ISO together withor without 5 μM L-2286, then the medium was discarded, andthe cells were processed for Western blotting. All experimentswere repeated three times.

2.6. Western blot analysis

Fifty milligrams of heart samples or a confluent 100 mmplate of cultured cells was homogenized in ice-cold Tris buffer(50 mmol/l, pH 8.0) containing 50 mM sodium vanadate andprotease inhibitor cocktail (Sigma-Aldrich Co., Budapest,Hungary) and harvested in 2× concentrated SDS-polyacryl-amide gel electrophoresis sample buffer. Proteins wereseparated on 12% SDS-polyacrylamide gel and transferred tonitrocellulose membranes. After blocking (2 h with 3% nonfatmilk in Tris-buffered saline), membranes were probed

overnight at 4 °C with antibodies recognizing the followingantigenes: phospho-specific Akt-1/protein kinase B-α Ser473

(1:1000), nonphosphorylated Akt/PKB (1:1000), phospho-specific glycogen synthase kinase (GSK)-3β Ser9 (1:1000),phospho-specific extracellular signal-regulated kinase (ERK1/2) Thr183–Tyr185 (1:1000), phospho-specific p38 mitogen-activated protein kinase (p38-MAPK) Thr180–Gly–Tyr182

(1:1000), phospho-specific c-Jun N-terminal kinase (JNK)(1:1000), phospho-specific protein kinase C (PKC) (pan) βIISer660 (1:1000), phospho-specific protein kinase C α/βII(PKCα/βII) Thr638/641 (1:1000), phospho-specific protein kinaseC δ Thr505 (1:1000), phospho-specific protein kinase C δ Ser643

(1:1000; Cell Signaling Technology, Beverly, MA, USA),phospho-specific protein kinase C ε Ser729 (1:1000; Upstate,UK), nonphosphorylated PKC (1:1000), N-terminal domain ofactin (1:10000; Sigma-Aldrich Co, Budapest, Hungary).Membranes were washed six times for 5 min in Tris-bufferedsaline (pH 7.5) containing 0.2% Tween (TBST) before additionof goat anti-rabbit horseradish peroxidase-conjugated secondaryantibody (1:3000 dilution; Bio-Rad, Budapest, Hungary).Membranes were washed six times for 5 min in TBST andthe antibody–antigen complexes were visualized by means ofenhanced chemiluminescence. The results of Western blots werequantified by NIH ImageJ program.

2.7. Statistical analysis

Statistical analysis was performed by analysis of varianceand all of the data were expressed as the mean ± SEM.Significant differences were evaluated by use of unpairedStudent's t test and P values below 0.05 were considered to besignificant.

3. Results

3.1. PARP inhibition improved the gravimetric parameters inISO-induced chronic heart failure

After myocardial infarction, the initial loss of myocardiumcan initiate progressive ventricular enlargement and remodel-ing. The surviving myocytes undergo hypertrophy character-ized by increases in both cell length and diameter. In ourstudy, gravimetry performed 8 weeks after ISO-inducedmyocardial infarction showed significantly elevated ventricu-lar weight (P < 0.05) as well as ventricular weight normalizedto body weight (P < 0.01) and to tibia length (P < 0.001)(Table 1). L-2286 treatment prevented the unfavorablechanges in these parameters (Table 1). Additionally, L-2286administration also attenuated the ISO-induced weight gain(Table 1). When added alone, L-2286 did not change thegravimetric parameters (Table 1).

3.2. L-2286 moderated the electrocardiographic changesassociated with chronic heart failure

Decreased R wave amplitude as well as J point depression(P < 0.01) were observed on the electrocardiogram of the ISO

Fig. 2. L-2286 inhibited the heart failure-induced elevation in plasma BNP level.Plasma BNP levels were determined by an EIA method as described in Materialsand methods. ISO: animals 8 weeks after ISO administration; ISO + L-2286:animals treated with L-2286, 8 weeks after ISO administration; C + L-2286:animals treated with L-2286 for 8 weeks. Values are means ± SEM, *P < 0.01(vs. Control group), †P < 0.05 (vs. ISO group).

Table 1PARP inhibition improves gravimetric parameters in ISO-induced chronic heartfailure

Control C + L-2286 ISO ISO + L-2286

BW (g) 403 ± 53.7 401.2 ± 36.3 431.3 ± 28.6 a 397.2 ± 22.3 b

WV (g) 1.02 ± 0.078 0.981 ± 0.07 1.132 ± 0.123 a 0.977 ± 0.074 b

WV/BW(g)

2.59 ± 0.14 2.47 ± 0.15 2.62 ± 0.19 a 2.48 ± 0.15 b

WV/TL(g/mm)

0.023 ± 0.001 0.022 ± 0.001 0.025 ± 0.0025 a 0.022 ± 0.0013 b

Conditions for performing gravimetry are detailed in Materials and methods.BW: body weight, WV: weight of ventricles, TL: length of right tibia, WV/BW:weight of ventricles/body weight ratio, WV/TL: weight of ventricles/length oftibia, ISO: animals 8 weeks after ISO administration; ISO + L-2286: animalstreated with L-2286, 8 weeks after ISO administration; C + L-2286: animalstreated with L-2286 for 8 weeks. All values are means ± SEM.a P < 0.05 (vs. Control group).b P < 0.05 (vs. ISO group).

152 A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

group 8 weeks after myocardial infarction (Table 2). Theseelectrocardiographic changes were significantly improved(P < 0.05) by L-2286 treatment (Table 2). When addedalone, L-2286 did not cause any electrocardiographic changes(Table 2).

3.3. L-2286 inhibited the heart failure-induced elevation inplasma BNP level

Elevation of plasma BNP level is associated with the severityof heart failure, while low plasma BNP concentrations providestrong evidence against heart failure. Plasma BNP level in ourstudy showed a significant increase in the ISO group 8 weeksafter myocardial infarction (P < 0.01) (Fig. 2). On the otherhand, L-2286 treatment significantly reduced the plasma BNPlevel (P < 0.05) suggesting that the PARP inhibition was able todecrease the severity of postinfarction heart failure. Whenadded alone, L-2286 did not change plasma BNP levels (Fig. 2).

3.4. L-2286 prevented the oxidative damage of respiratorycomplexes in chronic heart failure

Ventricular remodeling is accompanied by altered mito-chondrial energy metabolism. Under our experimental condi-tions, chronic heart failure induced a partial inactivation of

Table 2L-2286 mitigated electrocardiographic changes associated with chronic heartfailure

Control C + L-2286 ISO ISO + L-2286

R waveamplitude(mV)

0.592 ± 0.025 0.51 ± 0.03 0.175 ± 0.02 a 0.4 ± 0.031 b

J point (mV) 0.12 ± 0.012 0.1 ± 0.023 −0.038 ± 0.04 a 0.05 ± 0.02 b

Conditions for performing electrocardiography are detailed in Materials andmethods. ISO: animals 8 weeks after ISO administration; ISO + L-2286: animalstreated with L-2286, 8 weeks after ISO administration; C + L-2286: animalstreated with L-2286 for 8 weeks. All values are means ± SEM.a P < 0.01 (vs. Control group).b P < 0.05 (vs. ISO group).

respiratory complexes I–III assessed by the measurement ofNADH:cytochrome c oxidoreductase activity (P < 0.01) (Fig.3A). L-2286 could partially protect (P < 0.05) NADH:cytochrome c oxidoreductase activity in the failing heartmitochondria. Importantly, the expression of mitochondrialenzymes (including NADH dehydrogenase (complex I) 43,53, 70 subunits and pyruvate dehydrogenase complex-1α)showed no changes in either the untreated or L-2286-treatedheart samples, therefore the reduced activity of mitochondrialenzymes can be completely attributed to posttranslationaloxidative injury (Fig. 3B). When added alone, L-2286 did notcause any changes in the activity of the respiratory complexes(Fig. 3).

3.5. L-2286 attenuated the myocardial hypertrophy andinterstitial collagen deposition

Histological analysis revealed marked myocyte hypertrophyin failing rat hearts (Fig. 4B) compared to the control group(P < 0.001) (Fig. 4A). In addition, interstitial type III collagendeposition increased from 12.2 ± 1.2% in control hearts to18.4 ± 1.4% in ISO-treated animals (P < 0.001) (Figs. 5A–B).L-2286 treatment significantly decreased the mean myocytediameter (P < 0.005) (Figs. 4C–D) and prevented the interstitialfibrosis (type III collagen deposition 15.4 ± 1.4%) (P < 0.01)(Figs. 5C–D). When added alone, L-2286 did not affectmyocardial hypertrophy or interstitial collagen deposition (datanot shown).

3.6. L-2286 reduced GSK-3β and PKC phosphorylation infailing myocardium

GSK-3β was slightly phosphorylated in control animals, butbecame strongly phosphorylated 8 weeks after myocardialinfarction in the ISO group (Fig. 6A). However, this strongphosphorylation was diminished by L-2286 administration (Fig.6A). The moderate Akt phosphorylation present in the controlsamples increased slightly in the ISO group, but the L-2286treatment did not influence the phosphorylation state of Akt

Fig. 3. L-2286 prevented the oxidative damage of respiratory complexes in chronic heart failure. (A) Activity of NADH:cytochrome c oxidoreductase in isolatedmitochondria from failing rat hearts. Mitochondrial enzyme activities were measured as described in Materials and methods. ISO: animals 8 weeks after ISOadministration; ISO + L-2286: animals treated with L-2286, 8 weeks after ISO administration; C + L-2286: animals treated with L-2286 for 8 weeks. Data areexpressed as the percent of control values (means ± SEM), *P < 0.01 (vs. Control group), †P < 0.05 (vs. ISO group). (B) Expression of mitochondrial enzymes (NADHdehydrogenase complex (complex I)-43, 53, 70 subunit, pyruvate dehydrogenase complex-1α) in heart failure rats. Conditions for Western blot analysis were detailedin Materials and methods. Representative immunoblots from five experiments with similar results and densitometric evaluation are shown. PDC-1α: pyruvatedehydrogenase complex-1α; ISO: animals 8 weeks after ISO administration; ISO + L-2286: animals treated with L-2286, 8 weeks after ISO administration; C + L-2286: animals treated with L-2286 for 8 weeks.

153A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

(Fig. 6A). The level of GSK-3β phosphorylation was not inaccordance with that of Akt, one of the upstream kinases ofGSK-3β, therefore the L-2286-induced decrease in GSK-3βphosphorylation was probably independent of Akt.

Fig. 4. L-2286 attenuated the ISO-inducedmyocardial hypertrophy. Sections obtainedwere stained with hematoxylin and eosin and mean myocyte diameters were calculatMaterials and methods. ISO: animals 8 weeks after ISO administration; ISO + L-2286as means ± SEM, *P < 0.001 (vs. Control group), †P < 0.005 (vs. ISO group). When

The overall (pan) phosphorylation state of protein kinase Cwas moderate in control animals, but became strongly enhanced8 weeks after myocardial infarction without any change in theexpression of total PKC (Fig. 6B). As far as the PKC isoforms

from hearts of control (A), ISO-treated (B) and ISO + L-2286-treated animals (C)ed (D). Conditions for measurement of mean myocyte diameter were detailed in: animals treated with L-2286, 8 weeks after ISO administration. Values are givenadded alone, L-2286 did not affect myocardial hypertrophy (data not shown).

Fig. 5. L-2286 reduced the ISO-induced interstitial collagen deposition. Sections obtained from hearts of control (A), ISO-treated (B) and ISO + L-2286-treatedanimals (C) were immunostained with an antibody against type III collagen; additionally, densitometric evaluation is shown (D). Conditions for immunohistologywere detailed in Materials and methods. ISO: animals 8 weeks after ISO administration; ISO + L-2286: animals treated with L-2286, 8 weeks after ISO administration.Values are given as means ± SEM, *P < 0.001 (vs. Control group), †P < 0.01 (vs. ISO group). When added alone, L-2286 did not affect interstitial collagen deposition(data not shown).

154 A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

are concerned, PKC α/βII and PKC δ (Ser643) seemed to bepartially responsible for the ISO treatment-associated increasesin pan PKC phosphorylation. However, the phosphorylation ofPKC ε (Ser729) increased considerably and seemed to beaccounting for the majority of the changes observed in pan PKCphosphorylation in the ISO-treated animals (Fig. 6B). Mostimportantly and in concert with the GSK-3β phosphorylationstate, the enhanced phosphorylation of pan PKC, PKC α/βIIand PKC δ (Ser643) and PKC ε (Ser729), seen in thepostinfarcted samples was significantly attenuated by thePARP inhibitor treatment. Since GSK-3β can be phosphory-lated by PKC leading to its inactivation, we propose that thePARP inhibition-caused reduced GSK-3β phosphorylation wasprobably due to the decreased activation of the PKC pathway.The phosphorylation state of PKC δ (Thr505), however,remained unchanged during the different treatments used.When added alone, L-2286 did not cause significant changesin GSK-3β and PKC phosphorylation (Figs. 6A and B).

3.7. L-2286 had no influence on the phosphorylation of MAPkinases

Extracellular signal-regulated kinase (ERK) was moderatelyphosphorylated in control samples and remained so in ISO-treated as well as ISO + L-2286-treated hearts (Fig. 6C). Incontrast, the moderate phosphorylation of both p38-MAPK andc-jun N-terminal kinase (JNK) seen in control hearts waselevated in heart failure animals (ISO group), but L-2286 could

not mitigate the phosphorylation of these kinases (Fig. 6C).When added alone, L-2286 did not change significantly thephosphorylation of MAP kinases (Fig. 6C).

4. Discussion

Experimental models of chronic heart failure can be inducedby several methods, including coronary ligation, electricalshock, cardiac pacing, pressure or volume overload and toxicdepression of cardiac function. All of these interventions resultin progressive myocardial and neurohumoral changes, inde-pendently of the primary stimulus. Subcutaneous administrationof beta-adrenoceptor agonist isoproterenol produces patchymyocardial necrosis in a dose-related way with an intactcoronary vasculature [19]. Furthermore, isoproterenol is rapidlymetabolized so acute adverse effects of the drug can be avoided.Administration of the drug produces graded myocardial celldeath and rapidly impairs left ventricular function, at leastpartially, through free radical generation [19,24]. The initial cellloss leads to consequent myocardial hypertrophy and ventric-ular enlargement. Pathologic hypertrophy following myocardialinfarction often progresses into heart failure, which is one of theleading causes of morbidity and mortality worldwide andcharacterized by a progressive deterioration in heart function[5,24,25]. Intriguingly, PARP inhibition was reported toalleviate cardiac dysfunction and improve vascular relaxationin postinfarcted rats; however, the effects of PARP inhibition onthe (sub)cellular and metabolic parameters in the failing heart

155A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

are still elusive [22]. Therefore, we examined how a novelPARP inhibitor agent (L-2286) interfered with the progressionof postinfarction heart failure, including changes in cellularmass, interstitial matrix composition, metabolic parameters,gene expression and intracellular signaling.

Cardiac remodeling is characterized by inadequate myocytehypertrophy and accumulation of extracellular matrix structuralproteins (fibrillar collagen, type I and III collagen), which causetissue stiffness and adversely affect myocardial viscoelasticity,ultimately leading to diastolic and systolic dysfunctions[5,15,16]. In addition, perivascular fibrosis and disorganizedhypertrophy can exhaust the coronary blood flow reserve andthe consequent hypoperfusion may provoke myocardial ische-mia [5,21]. In our study, L-2286 exerted a beneficial effect onthe progression of postinfarction remodeling by attenuatingmyocardial hypertrophy and interstitial deposition of type IIIcollagen in the failing hearts.

Cardiac hypertrophy and failure also involve a severedysfunction of mitochondrial energy and substrate metabolism.Impaired mitochondrial function is associated with a decline inhigh-energy phosphates, reduced oxygen consumption anddecreased tissue content and activity of complex I through IVofthe respiratory chain [14]. Accordingly, the activity of complexI to III (NADH:cytochrome c oxidoreductase) has dropped inour heart failure model accompanied with unaltered expressionof either complex I (43, 53, 70 subunit) or pyruvatedehydrogenase complex (1α subunit). Since this enzymedysfunction is mainly attributed to posttranslational inactivationby reactive oxygen species, PARP inhibition by L-2286efficiently helped preventing this event.

In response to hypertrophic stimuli, cardiac fetal-like geneprogram is initiated entailing, among others, the expression andsecretion of natriuretic factors (BNP, ANP) [5,15,26,27]. Leftventricular wall stretch and volume overload trigger the releaseof plasma B-type natriuretic peptide from the ventricles,therefore elevated BNP concentration specifically signalsimpaired left ventricular function and heart failure [13,28]. Inconcert with this, plasma BNP level increased in thepostinfarcted animals, which was attenuated by L-2286treatment.

Recent years have seen a burst in our understanding of theintracellular signaling mechanisms which mediate extrinsic andintrinsic growth stimuli into coordinated gene expressionchanges. Hypertrophy is ignited and maintained by vasoactivepeptides, peptide growth factors, hormones and neurotransmit-ters, which all act upon MAP kinases, JAK/STAT, CaMK/calcineurin routes as well as glycogen synthase kinase-3β[5,6,29–33]. Differential activation of MAPKs results inspecific cardiac morphologic and functional phenotypes:while ERK1/2 activation leads to a concentric form ofhypertrophy with enhanced cardiac function, p38-MAPK andJNK activation gives rise to maladaptive hypertrophy [5].Interestingly, JNK and p38-MAPK are activated in failinghuman hearts, while ERK activation remains at a physiologiclevel [5,7]. Furthermore, several protein kinases including Akt,p70-S6 kinase, p90-RSK, protein kinase C and protein kinase Ahave been implicated in cardiac hypertrophy, commonly

phosphorylating, thereby suppressing the antihypertrophicactivity of GSK-3β [7,13,30,34]. The ability of GSK-3β toinhibit hypertrophy in response to β-adrenergic stimulation,pressure overload and calcineurin activation suggests that GSK-3β represents a point of cross-talk and convergence of varioushypertrophic signaling pathways [7,13,30,34]. GSK-3β phos-phorylation via a phosphatidylinositol-3-kinase dependentpathway can be catalyzed by at least two protein kinases: Aktand ILK [31,34]. Growth factors, however, can regulate GSK-3β not only through the PI3K/Akt pathway, but also via analternative PKC pathway [35]. Most importantly, several PKCisoforms, i.e. PKCα, PKCβ, PKCγ, PKCη and PKCδ, canphosphorylate GSK-3β at regulatory serine residues [36].Furthermore, altered gene expression of differential cardiacPKC isozymes including PKCδ, PKCε, PKCα and PKCβI havebeen recently reported in myocardial hypertrophy [37–40].

In a previous study, we reported similar patterns ofelevation in phosphorylation of GSK-3β, ERK1/2 and p38during the first 24 h in the presence and absence of L-2286following ischemia and ISO administration. However, therewere marked differences in phosphorylation of Akt and JNKbetween the two treatments characterized by a transientincrease in the first 4 h followed by a decline in case ofISO treatment but a phosphorylation pattern similar to that ofphospho-ERK1/2 and phospho-p38 MAP kinases in case ofischemia–reperfusion [8]. These changes in the activation ofkinase signaling pathways were most probably due to theinduced oxidative stress rather than the ISO per se, since wecould not detect any significant changes in the phosphoryla-tion of the studied kinases when we treated H9c2(2-1)cardiomyoblast—or rat neonatal primary cardiomyocytecultures with ISO in the presence and absence of L-2286 inthe present study (data not shown). In contrast to the acuteeffects of ISO, 8 weeks after myocardial infarction, we foundno increase in ERK1/2 activity, but experienced a two-foldelevation in Akt, p38-MAPK and JNK phosphorylation.Notably, upon L-2286 administration, these kinase activitiesremained unaltered. By sharp contrast, heart failure followingISO injections upregulated GSK-3β phosphorylation, whichcorrelated with the activation of several PKC isoforms ratherthan that of Akt. Most importantly, PARP inhibition in theinfarcted hearts simultaneously mitigated both GSK-3β andPKC phosphorylation. Since GSK-3β can integrate a varietyof antihypertrophic signals and transmit them to NFAT and c-Jun, the L-2286-induced PKC-mediated GSK-3β activationmay have contributed to the rescue of failing hearts in ourexperimental model [14].

The mechanism by which PARP inhibition diminished theactivity of PKC still needs to be elucidated. Nevertheless,oxidative stress in chronic heart failure can increase PKCactivation via various pathways. Firstly, oxidative stress candirectly or indirectly stimulate PKC via growth factor receptorphosphorylation or lipid secondary messengers (derived fromthe reactions of other receptor-regulated enzymes such asphospholipases A2, D and C) [41]. Secondly, poly(ADP-ribose) polymerase activation could also contribute to PKCactivation, because ROS-induced poly(ADP-ribosyl)ation

156 A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

depletes nicotinamide adenine dinucleotide (NAD+) pools ordirect poly(ADP-ribosyl)ation blocks glycerinaldehyde 3-phosphate dehydrogenase (GAPDH) enzyme [42,43]. Theinhibition of GAPDH can divert the glycolytic flux towards the

formation of α-glycerophosphate and diacylglycerol (DAG),the latter being the activator of PKC [43]. On the other hand,reactive oxygen species can also directly oxidize and hinderGAPDH [44].

Fig. 6. L-2286 treatment inhibits hypertrophic stimuli-induced GSK-3β phosphorylation via protein kinase C signaling in chronic heart failure. (A) Western blotanalysis of Akt and GSK-3β phosphorylation and densitometric evaluation is shown. (B) Western blot analysis of pan PKC, PKC α/β, PKC δ (Thr505), PKC δ(Ser643) and PKC ε (Ser729) phosphorylation and densitometric evaluation is shown. (C) Western blot analysis of ERK, p38-MAPK and JNK phosphorylation anddensitometric evaluation is shown. Actin is shown as loading control. ISO: animals 8 weeks after ISO administration; ISO + L-2286: animals treated with L-2286,8 weeks after ISO administration; C + L-2286: animals treated with L-2286 for 8 weeks. Representative immunoblots from five experiments with similar results areshown, *P < 0.001 (vs. Control group), †P < 0.001 (vs. ISO group).

157A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

Notably, PARP inhibition primarily interferes with theactivity of nuclear PARP and, in turn, maintains NAD+ andATP pools avoiding NAD+ depletion-induced inactivation ofGAPDH. Furthermore, we have previously reported that PARPinhibitors could prevent the ROS-induced inactivation ofmitochondrial respiratory complexes and the consequentmitochondrial ROS formation [22]. We therefore suggestthat PARP inhibition can blunt PKC activation eventually bypreserving NAD+ pools and abrogating ROS production,thereby diminishing GAPDH inactivation and DAGformation.

In conclusion, our study is the first demonstration thatPARP inhibition can halt the progression of cardiachypertrophy into failure partially by promoting GSK-3βactivity via direct or indirect interruption of upstream proteinkinase C signaling. Here we also report that PARPinhibition protects against cellular changes in the post-

infarcted myocardium such as fibrosis, cardiomyocytehypertrophy, mitochondrial dysfunction and fetal geneexpression. The PARP inhibition-induced alterations inintracellular signaling upon myocardial remodeling furtherchallenge the classical view that protection by PARPinhibitors exclusively relies on the preservation of NAD+

and consequently of ATP stores.

Acknowledgments

We thank Gaborne Horvath, Imrene Szilagyi, Laszlo Giranand Bertalan Horvath for excellent technical help. This workwas supported by grants from the Hungarian National ResearchFoundations (OTKA T48334, T47152, F46594, T034320),from the Hungarian Ministry of Health, Social and FamilyAffairs (ETT 512/2003), from the South-TransdanubianCooperative Research Centre, University of Pecs (PTE DD-

158 A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

KKK) and by a grant from the National Research andDevelopment Program (NKFP 2001 1/026).

References

[1] Virag L, Szabo C. The therapeutic potential of poly(ADP-ribose)polymerase inhibitors. Pharmacol Rev 2002;54:375–429.

[2] Szabo G, Liaudet L, Hagl S, Szabo C. Poly(ADP-ribose) polymeraseactivation in the reperfused myocardium. Cardiovasc Res 2004;61:471–80.

[3] Lopez Farre A, Casado S. Heart failure, redox alterations, and endothelialdysfunction. Hypertension 2001;38:1400–5.

[4] Dhalla AK, Singal PK. Antioxidant changes in hypertrophied and failingguinea pig hearts. Am J Physiol 1994;266:H1280–5.

[5] Lips DJ, de Windt LJ, van Kraaij DJ, Doevendans PA. Moleculardeterminants of myocardial hypertrophy and failure: alternative path-ways for beneficial and maladaptive hypertrophy. Eur Heart J2003;24:883–96.

[6] Yamazaki T, Yazaki Y. Molecular basis of cardiac hypertrophy. Z Kardiol2000;89:1–6.

[7] Haq S, Choukroun G, Lim H, Tymitz KM, del Monte F, Gwathmey JL,et al. Differential activation of signal transduction pathways in humanhearts with hypertrophy versus advanced heart failure. Circulation2001;103:670–7.

[8] Palfi A, Toth A, Kulcsar G, Hanto K, Deres P, Bartha E, et al. The role ofAkt and MAP kinase systems in the protective effect of PARP inhibition inLangendorff perfused and in isoproterenol damaged rat hearts. J PharmacolExp Ther 2005;315:273–82.

[9] Habon T, Szabados E, Kesmarky G, Halmosi R, Past T, Sumegi B,et al. The effect of carvedilol on enhanced ADP-ribosylation and redblood cell membrane damage caused by free radicals. Cardiovasc Res2001;52:153–60.

[10] Fowler MB. Carvedilol prospective randomized cumulative survival(COPERNICUS) trial: carvedilol in severe heart failure. Am J Cardiol2004;93:35B–9B.

[11] Zingarelli B, Hake PW, O'ConnorM, Denenberg A,Wong HR, Kong S, et al.Differential regulation of activator protein-1 and heat shock factor-1 inmyocardial ischemia and reperfusion injury: role of poly(ADP-ribose)polymerase-1. Am J Physiol, Heart Circ Physiol 2004;286:H1408–15.

[12] Zingarelli B, Hake PW, O'Connor M, Denenberg A, Kong S, Aronow BJ.Absence of poly(ADP-ribose) polymerase-1 alters nuclear factor-kappa Bactivation and gene expression of apoptosis regulators after reperfusioninjury. Mol Med 2003;9:143–53.

[13] Antos CL, McKinsey TA, Frey N, Kutschke W, McAnally J, Shelton JM,et al. Activated glycogen synthase-3 beta suppresses cardiac hypertrophyin vivo. Proc Natl Acad Sci USA 2002;99:907–12.

[14] van Bilsen M, Smeets PJ, Gilde AJ, van der Vusse GJ. Metabolicremodelling of the failing heart: the cardiac burn-out syndrome?Cardiovasc Res 2004;61:218–26.

[15] Tan FL, Moravec CS, Li J, Apperson-Hansen C, McCarthy PM, Young JB,et al. The gene expression fingerprint of human heart failure. Proc NatlAcad Sci USA 2002;99:11387–92.

[16] Sun Y, Zhang JQ, Zhang J, Lamparter S. Cardiac remodeling by fibroustissue after infarction in rats. J Lab Clin Med 2000;135:316–23.

[17] Veres B, Radnai B, Gallyas Jr F, Varbiro G, Berente Z, Osz E, et al.Regulation of kinase cascades and transcription factors by a poly(ADP-ribose) polymerase-1 inhibitor, 4-hydroxyquinazoline, in LLP-inducedinflammation in mice. J Pharmacol Exp Ther 2004;310:247–55.

[18] Tapodi A, Debreceni B, Hanto K, Bognar Z, Wittmann I, Gallyas Jr F, et al.Pivotal role of Akt activation in mitochondrial protection and cell survivalby poly(ADP-ribose) polymerase-1 inhibition in oxidative stress. J BiolChem 2005;280:35767–75.

[19] Teerlink JR, Pfeffer JM, Pfeffer MA. Progressive ventricular remodeling inresponse to diffuse isoproterenol-induced myocardial necrosis in rats. CircRes 1994;75:105–13.

[20] Kulcsar G, Kalai T, Osz E, Sar PC, Jeko J, Sumegi B, et al. Synthesis andstudy of new 4-quinazolinone inhibitors of the DNA repair enzyme poly(ADP-ribose) polymerase (PARP). ARKIVOC 2003:121–31.

[21] Pacher P, Liaudet L, Mabley J, Komjati K, Szabo C. Pharmacologicinhibition of poly(adenosine diphosphate-ribose) polymerase may repre-sent a novel therapeutic approach in chronic heart failure. J Am CollCardiol 2002;40:1006–16.

[22] Halmosi R, Berente Z, Osz E, Toth K, Literati-Nagy P, Sumegi B. Effect ofpoly(ADP-ribose) polymerase inhibitors on the ischemia–reperfusion-induced oxidative cell damage and mitochondrial metabolism inLangendorff heart perfusion system. Mol Pharmacol 2001;59:1497–505.

[23] Wang Y, Huang S, Sah VP, Ross Jr J, Brown JH, Han J, et al. Cardiacmuscle cell hypertrophy and apoptosis induced by distinct members of thep38 mitogen-activated protein kinase family. J Biol Chem 1998;273:2161–8.

[24] Grimm D, Elsner D, Schunkert H, Pfeifer M, Griese D, BruckschlegelG, et al. Development of heart failure following isoproterenoladministration in the rat: role of the renin–angiotensin system.Cardiovasc Res 1998;37:91–100.

[25] Cleland JG, Swedberg K, Follath F, Komajda M, Cohen-Solal A, AguilarJC, et al, Study Group on Diagnosis of theWorking Group on Heart Failureof the European Society of Cardiology. The EuroHeart Failure surveyprogramme—A survey on the quality of care among patients with heartfailure in Europe: Part 1. Patient characteristics and diagnosis. Eur Heart J2003;24:442–63.

[26] Ogawa Y, Itoh H, Nakao K. Molecular biology and biochemistry ofnatriuretic peptide family. Clin Exp Pharmacol Physiol 1995;22:49–53.

[27] Hystad ME, Klinge R, Spurkland A, Attramadal H, Hall C.Contrasting cardiac regional responses of A-type and B-typenatriuretic peptide to experimental chronic heart failure. Scand JClin Lab Invest 2000;60:299–309.

[28] Clerico A, Emdin M. Diagnostic accuracy and prognostic relevance of themeasurement of cardiac natriuretic peptides: a review. Clin Chem2004;50:33–50.

[29] Kim S, Iwao H. Activation of mitogen-activated protein kinases incardiovascular hypertrophy and remodeling. Jpn J Pharmacol 1999;80:97–102.

[30] Hardt SE, Tomita H, Katus HA, Sadoshima J. Phosphorylation ofeukaryotic translation initiation factor 2Bepsilon by glycogen synthasekinase-3beta regulates beta-adrenergic cardiac myocyte hypertrophy. CircRes 2004;94:926–35.

[31] Haq S, Choukroun G, Kang ZB, Ranu H, Matsui T, Rosenzweig A, et al.Glycogen synthase kinase-3beta is a negative regulator of cardiomyocytehypertrophy. J Cell Biol 2000;151:117–30.

[32] Liao P, Georgakopoulos D, Kovacs A, ZhengM, Lerner D, Pu H, et al. Thein vivo role of p38 MAP kinases in cardiac remodeling and restrictivecardiomyopathy. Proc Natl Acad Sci USA 2001;98:12283–8.

[33] Molkentin JD. Calcineurin and beyond: cardiac hypertrophic signaling.Circ Res 2000;87:731–8.

[34] Condorelli G, Drusco A, Stassi G, Bellacosa A, Roncarati R, Iaccarino G,et al. Akt induces enhanced myocardial contractility and cell size in vivo intransgenic mice. Proc Natl Acad Sci USA 2002;99:12333–8.

[35] Ballou LM, Tian PY, Lin HY, Jiang YP, Lin RZ. Dual regulation ofglycogen synthase kinase-3beta by the alpha1A-adrenergic receptor. J BiolChem 2001;276:40910–6.

[36] Fang X, Yu S, Tanyi JL, Lu Y, Woodgett JR, Mills GB. Convergence ofmultiple signaling cascades at glycogen synthase kinase 3: Edg receptor-mediated phosphorylation and inactivation by lysophosphatidic acidthrough a protein kinase C-dependent intracellular pathway. Mol CellBiol 2002;22:2099–110.

[37] Braun MU, Szalai P, Strasser RH, Borst MM. Right ventricularhypertrophy and apoptosis after pulmonary artery banding: regulation ofPKC isozymes. Cardiovasc Res 2003;59:658–67.

[38] Koide Y, Tamura K, Suzuki A, Kitamura K, Yokoyama K, Hashimoto T,et al. Differential induction of protein kinase C isoforms at the cardiachypertrophy stage and congestive heart failure stage in Dahl salt-sensitive rats. Hypertens Res 2003;26:421–6.

[39] Braun M, Simonis G, Birkner K, Pauke B, Strasser RH. Regulationof protein kinase C isozyme and calcineurin expression inisoproterenol induced cardiac hypertrophy. J Cardiovasc Pharmacol2003;41:946–54.

159A. Palfi et al. / Journal of Molecular and Cellular Cardiology 41 (2006) 149–159

[40] Bayer AL, Heidkamp MC, Patel N, Porter M, Engman S, SamarelAM. Alterations in protein kinase C isoenzyme expression andautophosphorylation during the progression of pressure overload-induced left ventricular hypertrophy. Mol Cell Biochem 2003;242:145–52.

[41] Gopalakrishna R, Jaken S. Protein kinase C signaling and oxidative stress.Free Radic Biol Med 2000;28:1349–61.

[42] Du X, Matsumura T, Edelstein D, Rossetti L, Zsengeller Z, Szabo C, et al.Inhibition of GAPDH activity by poly(ADP-ribose) polymerase activates

three major pathways of hyperglycemic damage in endothelial cells. J ClinInvest 2003;112:1049–57.

[43] Minchenko AG, Stevens MJ, White L, Abatan OI, Komjati K, Pacher P,et al. Diabetes-induced overexpression of endothelin-1 and endothelinreceptors in the rat renal cortex is mediated via poly(ADP-ribose)polymerase activation. FASEB J 2003;17:1514–6.

[44] Janero DR, Hreniuk D, Sharif HM. Hydroperoxide-induced oxidativestress impairs heart muscle cell carbohydrate metabolism. Am J Physiol1994;266:C179–88.