The FASEB Journal • FJ Express Full-Length Article

Free radical scavenging inhibits STAT phosphorylationfollowing in vivo ischemia/reperfusion injury

James McCormick,*,1 Sean P. Barry,*,1 Ahila Sivarajah,† Giorgio Stefanutti,§

Paul A. Townsend,‡ Kevin M. Lawrence,† Simon Eaton,§ Richard A. Knight,*Christoph Thiemermann,† David S. Latchman,* and Anastasis Stephanou*,2

*Department of Medical Molecular Biology, §Department of Surgery, The Institute of Child Health,University College London, London, UK; †Centre for Experimental Medicine, Nephrology andCritical Care, The William Harvey Research Institute, St. Bartholomew’s and The Royal LondonSchool of Medicine and Dentistry, Queen Mary, University of London, Charterhouse Square,London, UK; and ‡Human Genetics Division, Southampton General Hospital, University ofSouthampton, Southampton, UK

ABSTRACT The signal transducer and activator oftranscription (STAT) family are latent transcriptionfactors involved in a variety of signal transductionpathways, including cell death cascades. STAT1 hasbeen shown to have a crucial role in regulating cardiaccell apoptosis in the myocardium exposed to ischemia/reperfusion (I/R) injury. The free radical scavenger,tempol, is known to have cardioprotective properties,although little is known about the molecular mecha-nism(s) by which it acts. In the present study, weassessed the levels of phosphorylated STAT1 andSTAT3 and examined whether tempol was able toaffect STAT activation after in vivo cardiac I/R injury.We observed a reperfusion time-dependent increase inthe tyrosine phosphorylation of STAT1 and STAT3 atresidues 701 and 705, respectively. Here we show forthe first time that tempol dramatically reduced STAT1and 3 phosphorylation. The reduction in STAT1 and 3phosphorylation was accompanied by a concomitantdecrease in cellular malondialdehyde (MDA) levels. Toverify the role of STAT1 in modulating the cardiopro-tective effect of tempol, rats were injected with theSTAT1 activator, IFN-�, and tempol during I/R injury.We found that the presence of IFN-� abrogated theprotective effects of tempol, suggesting that the pro-tective effects of tempol may partly operate by decreas-ing the phosphorylation of STAT1. This study demon-strates that careful dissection of the molecularmechanisms that underpin I/R injury may reveal cardio-protective targets for future therapy.—McCormick, J,.Barry, S. P., Sivarajah, A., Stefanutti, G., Townsend, P. A.,Lawrence, K. M., Eaton, S., Knight, R. A., Thiemermann,C., Latchman, D. S., Stephanou, A. Free radical scaveng-ing inhibits stat phosphorylation following in vivo isch-emia/reperfusion injury. FASEB J. 20, E1404–E1410(2006)

Key Words: tempol � heart � rat

Myocardial ischemia results from the occlusion of

blood flow to the left ventricle via the coronary artery,leading to cardiac stunning and cell death, which canlead to loss of cardiac output (1). Although bothnecrotic and apoptotic cell death occur during isch-emia, apoptotic cell death appears to be enhanced afterthe re-establishment of coronary blood flow (reperfu-sion) (2, 3). Inhibition of the apoptotic process incardiac myocytes exposed to simulated ischemia/reper-fusion injury (I/R) or in the ischemic myocardiumusing specific caspase inhibitors results in reducedinfarct size (4, 5).

Reactive oxygen species (ROS) are generated duringischemia and reperfusion injury (6), and these freeradicals promote the onset of cell death, resultingin characteristic cellular changes including lipid per-oxidation, protein denaturation, and double-strandbreaks in DNA (6). The generation of ROS during I/Rhas been demonstrated to induce apoptosis in thecardiac myocyte (7). Tempol (4-hydroxy-2,2,6,6-tetra-methylpiperidine-1-oxyl) is a low molecular weightagent capable of scavenging superoxide anions, andoffers an advantage over many other ROS scavengers asit is able permeate biological membranes in vitro (8).Moreover, tempol was shown to reduce infarct size afterIR injury (8).

The janus-kinase (JAK) family of proteins areactivated in response to various physiological insultsand cytokines, leading to the phosphorylation ofsignal transducer and activator of transcription(STAT) proteins (9, 10). Phosphorylation of STAT1and STAT3 at tyrosine residues 701 and 705, respec-tively, results in their dimerization, leading to nu-clear translocation. STAT1 is thought to promote theonset of apoptosis while STAT3 is believed to abro-

1These authors contributed equally to this work.2Correspondence: Department of Medical Molecular Biol-

ogy, The Institute of Child Health, University College Lon-don, 30 Guilford St., London, WC1N 1EH, UK. E-mail:a.stephanou@ich.ucl.ac.uk

doi: 10.1096/fj.06-6188fje

E1404 0892-6638/06/0020-1404 © FASEB

gate the effects of STAT1 (11). Once in the nucleus,STAT proteins regulate the expression of pro- andantiapoptotic genes, including caspase 1 (9, 10).Both STATs can also be phosphorylated on serineresidue 727, which is thought to be necessary formaximum activation, although the precise mecha-nism is poorly understood.

Previously we demonstrated that STAT1 plays animportant role in promoting apoptotic cell death incardiac myocytes and also in the intact heart after I/Rinjury (11–14). These studies therefore suggest STAT1as a therapeutic target in the ischemic myocardium. Ithas been reported that the polyphenolic agent epigal-locatechin-3-gallate (EGCG), a major constituent ofgreen tea, is a potent inhibitor of STAT-1 phosphory-lation and activation (15). We recently showed thatEGCG reduced STAT-1 phosphorylation and protectedcardiac myocytes against I/R–induced apoptotic celldeath. More interesting is the fact that EGCG reducedthe infarct size and improved hemodynamic recoveryand ventricular function in the ischemic/reperfusedrat heart (16).

Here we demonstrate, using an in vivo model, thatSTAT1 and STAT3 phosphorylation and myocardialinfarct size can be reduced by injection of the freeradical scavenger, tempol. Moreover, IFN-�, a potentactivator of STAT1, not only enhanced infarct sizebut overcame the protective effects of tempol. Theresults presented here demonstrate a potential mech-anism of action for tempol and indicate the thera-peutic use of free radical scavengers in postischemicinsults.

MATERIALS AND METHODS

This study was performed in accordance with the UnitedKingdom, Home Office Animals (Scientific Procedures) Act1986.

In vivo ischemia/reperfusion injury in the rat

Ischemia/reperfusion injury in the rat was performed asdescribed previously (8). Briefly, male Wister rats (255–285 g)were anesthetized with sodium thiopentone (120 mg/kg;Intraval, Essex, UK) by intraperitoneal injection. Rats weresubjected to 25 min of acute myocardial ischemia by occlu-sion of the left anterior descending coronary artery (LAD),followed by reperfusion, or were sham operated (no LADocclusion). After reperfusion, the coronary artery was reoc-cluded; Evans blue dye (VWR, Dorset, UK) was injected intothe left ventricle and the animals sacrificed by an overdose ofanesthetic. The uncolored area (risk area) of the heart wasremoved from the blue area (nonrisk area).

Western blot analysis

Cardiac tissue was snap frozen in liquid nitrogen and lysed inradio-immuno-precipitation assay (RIPA) buffer (0.75 MNaCl, 5% (v/v) Nonidet P-40, 2.5% (w/v) deoxycholate, 0.5%(w/v) SDS, 0.25 M Tris-HCl pH 8.0, 10 mM DTT containingprotease inhibitor cocktail). Western blots were carried out as

described previously (24), using anti-STAT1, anti-STAT3,antiphospho ERK1/2 (Santa Cruz Biotechnology, CA, USA),antiphosphoSTAT1Y701 (Zymed, San Francisco, CA, USA),antiphosphoSTAT1S727 (Upstate, Cambridge, UK) antiphos-pho STAT3Y705, antiphospho STAT3727 (Cell Signaling Tech-nology, Danvers, MA, USA), and anti-GAPDH (Chemicon,Temecula, CA, USA). Horseradish peroxidase-conjugatedsecondary antibodies were from Dako (Glostrup, Denmark).Densitometric analysis was performed using a GS-800 scannerand quantity one software (Bio-Rad, Herts, UK). All statisticalanalyses were performed using Graphpad Prism v4.0 (Graph-Pad Software, San Diego, CA, USA).

Immunohistochemistry

After the indicated treatments, the area at risk was separatedand cut transversely into four slices. The tissue was then fixedin 10% formalin and embedded in paraffin. Immunohisto-chemical analysis was carried out on 4 �m sections withstandard protocols using antibodies specific for phospho-STAT3Y705 (Cell Signaling) and phospho-STAT1Y701 (Zymed,San Francisco, USA). Biotin-conjugated secondary antibodieswere used (Vector Laboratories, Peterborough, UK), fol-lowed by streptavidin-biotin peroxidase complex solution(Dako). Color reactions were developed by incubating sec-tions in 3�3-diaminobenzidine (Sigma-Aldrich, Dorset, UK)and images were captured on a Zeiss Axioscope 2 plusmicroscope.

Determination of tissue malondialdehyde concentration

Levels of malondialdehyde MDA in cardiac tissue were mea-sured by HPLC as described previously (17). All reagents usedwere obtained from Sigma-Aldrich. Protein concentration inthe homogenate was measured by the method of Peterson(25), and MDA was expressed as �mol/g protein.

Drug protocol

For MDA and STAT activation experiments, the followingexperimental groups were used: LAD coronary artery occlu-sion (25 min) and reperfusion (30 min) infused with 1) saline5 min prior to the onset of reperfusion, 2) 100 mg/kg tempol5 min prior to the onset of reperfusion, or 3) sham operation(n�4). For infarct measurements, rats were either shamoperated or subjected to 25 min of ischemia and 2 h ofreperfusion infused with 1) saline, 2) tempol, 3) IFN-�, or 4)tempol � IFN-� (n�5).

Quantification of myocardial tissue injury

At the end of the 2 h reperfusion period, the risk area wasobtained (as above). The tissue from the risk area was cut intosmall pieces and incubated with p-nitroblue tetrazolium(NBT, 0.5 mg/ml) for 30 min at 37°C. The stained tissue wasseparated from the infarcted tissue, weighed, and the infarctsize expressed as a percentage of the risk area.

Statistical analysis

All data are expressed as a mean � se of n observations, wheren represents the number of animals in the group. Hemody-namic parameters were analyzed via a 2-way ANOVA, followedby a Bonferroni post test. Risk area and infarct size wereanalyzed via a one-factorial ANOVA, followed by a Bonferroni

E1405FREE RADICAL SCAVENGING IN I/R INJURY

post test for multiple comparisons. A P value �0.05 denotes astatistical significantly difference.

RESULTS

STATs are phosphorylated in a time-dependentmanner after I/R injury in vivo

Using an in vivo coronary occlusion model, we per-formed a time course of I/R injury on the rat myocar-dium to assess the dynamics of STAT phosphorylationover time. Animals were subjected to either shamoperation (no LAD occlusion), 25 min of ischemia, or25 min of ischemia followed by reperfusion for 5 to 120min. As expected, the mean arterial blood pressure fellduring the ischemic insult, but remained constantthroughout the reperfusion time, whereas heart rateremained constant during both ischemia and reperfu-sion (data not shown).

As shown in Fig. 1A, STAT1 tyrosine phosphorylation(Y701) rather than STAT1 serine phosphorylation

(S727) was induced 15 min after reperfusion in the riskarea and reached a maximum at 30 min. Levels ofSTAT1 Y701 phosphorylation decreased dramatically at60 min of reperfusion and remained unchanged up to120 min. Phosphorylation of STAT1 occurred predom-inantly in the risk area, the main area of ROS genera-tion, highlighting the role of activated STAT1 in thedamaged myocardium.

In contrast to STAT1, tyrosine 705 (Y705) phosphor-ylation of STAT3 occurred during ischemia, but duringreperfusion it followed kinetics similar to STAT1, withincreasing levels of Y705 phosphorylation seen duringreperfusion phase, reaching a maximum at 30 min (Fig.1A). Phosphorylation of STAT3 at S727 occurred max-imally during ischemia alone and remained relativelyunchanged during reperfusion, suggesting that STAT3S727 phosphorylation is reperfusion independent (Fig.1B). Similar to STAT1, phosphorylation of STAT3 Y705and S727 was present at basal levels in the nonrisk area,indicating that the injury had only affected the leftventricle, as expected. These data indicated I/R invivo-induced STAT activation may contribute to infarctdevelopment.

Reperfusion injury induces rapid activation of theATM-�H2AX pathway

H2AX has been shown to be activated by hypoxia andreperfusion in solid tumors (18). The ischemic insultfailed to activate H2AX, suggesting that ischemiadoes not lead to large-scale DNA damage (Fig. 2). Incontrast, reperfusion rapidly induced �H2AX in therisk area, showing for the first time that H2AXphosphorylation occurs in a reperfusion-dependentmanner in cardiac tissue. This phosphorylation elic-its time-dependent kinetics similar to STAT1 and 3phosphorylation, highlighting concomitant activa-tion of DNA damage and JAK-STAT pathways. Mini-mal �H2AX activation occurs in the nonrisk area,highlighting the fact that double-strand DNA breaksare confined to areas of tissue suffering the greatestreperfusion-induced injury.

Figure 1. Time course of in vivo reperfusion after ischemiain rat hearts. Rats were sham operated or underwent 25min ischemia, followed by the periods of reperfusionindicated. 20 �g protein from each heart (n�4 per timepoint, risk and nonrisk areas) were separated on SDS-PAGEgels and probed with antibodies for A) total STAT1,phospho STAT1Y701, and phospho STAT1S727, B) TotalSTAT3, phospho STAT3Y705, and phospho STAT3727.Equal protein loading was confirmed by probing with anantibody (Ab) to GAPDH.

γ

Figure 2. 20 �g protein from each heart (n�4 per timepoint, risk and nonrisk areas) were separated on SDS-PAGEgels and levels of DNA damage were assessed by probingfor the phosphorylated form of histone-2AX (H2AX).Equal protein loading was confirmed by immunoblottingfor GAPDH.

E1406 Vol. 20 October 2006 McCORMICK ET AL.The FASEB Journal

STAT1 and STAT3 tyrosine phosphorylation isinduced by ROS

ROS generation is responsible for intracellular dam-age and has been shown to lead to the onset ofapoptosis (7). To verify that tempol had reduced theconcentration of ROS within the intracellular envi-ronment, we measured the cellular concentration ofmalondialdehyde (MDA). MDA is a marker of lipidperoxidation that occurs as a result of the damagingeffects of ROS (17). We found that injection oftempol reduced the concentration of MDA to shamlevels compared with the vehicle control (P�0.002),confirming that free radical concentration had beensuccessfully inhibited during I/R (Fig. 3A). Compar-ing the sham-operated to the vehicle and tempol-treated animals in the nonrisk area, no statisticallysignificant difference in the MDA levels (P��0.2)was observed, indicating that I/R injury does noteffect the nonrisk area (Fig. 3A).

The decrease in MDA levels is coupled with de-creased tyrosine phosphorylation of both STAT1and STAT3 (Fig. 3B). Densitometry revealed that ty-rosine phosphorylation of STAT1 was reduced by anaverage of 70 � 5% compared with the vehicle control,while STAT3 was decreased 45 � 7% (Fig. 3C). Immuno-histochemical staining of tissue sections from the risk areaconfirms that I/R-induced STAT phosphorylation, whichwas predominantly nuclear, was indeed reduced by freeradical scavenging (Fig. 4). These data imply that STAT1and STAT3 phosphorylation in I/R can be attenuated by

a single bolus injection of tempol, which may allowtempol to protect against cell death on the onset ofreperfusion.

Activated STAT1 contributes to infarct development,in vivo

Although activated STAT1 is induced after I/R injuryand has been implicated in promoting apoptosis (11–14), it was unclear whether this activation contributedto infarct development, in vivo. To test this, rats weresubjected to I/R injury with injection of the potentSTAT1 activator, IFN-�. IFN-� was introduced to allowcontinual activation of STAT1 for up to 120 min(postischemia), at which point the infarct area can bemeasured. Baseline values of mean arterial blood pres-sure (MAP), heart rate, and pressure rate index (PRI)were not significantly different between groups (datanot shown) compared with sham-operated animals; 25min of LAD occlusion followed by 2 h of reperfusiondid not cause significant alterations in MAP, heart rate,or pressure rate index compared with their respectivecontrols. Neither IFN-� or IFN-� plus tempol had anysignificant effect on MAP, HR or PRI (data not shown).

The mean values for the risk area were similar in allanimal groups studied and ranged from 48 � 2% to 51 �4% (data not shown) compared with sham-operated ani-mals; 25 min of LAD occlusion, followed by 2 h of reperfu-sion, caused a significant increase in infarct size. Administra-tion of IFN-� (25 �g/kg) 5 min before reperfusion caused asignificant increase (by 30%) in myocardial infarct size

Figure 3. Rats were sham operated or underwent 25 min of ischemia and 30min of reperfusion with infusion of saline or 100 mg/kg tempol, after whichrisk and nonrisk areas were dissected. A) The MDA lipid peroxidation assay wascarried out using equal amounts of tissue (n�5); results are expressed in �molMDA/g protein (*P�0.001) compared with I/R. B) 20 �g protein from eachgroup (n�5) were immunoblotted for pSTAT1Y701, pSTAT3Y705, pSTAT3S727,total STAT1, and total STAT3. C) Densitometric analysis of pSTAT1Y701andpSTAT3Y705 normalized to total STAT and expressed as % change from vehicle(*P�0.05).

E1407FREE RADICAL SCAVENGING IN I/R INJURY

compared with vehicle (Fig. 5A); IFN-�-induced activation ofSTAT1 in the risk area was confirmed by Western blot (Fig.5B). In the presence of IFN-�, tempol no longer had anyaffect on STAT1 phosphorylation; strikingly, IFN-� com-pletely abrogated the cardioprotective affect of tempol. Thisstrongly implicates STAT1 activation as a main target intempol-mediated cardioprotection

DISCUSSION

In the present study we show for the first time signifi-cant changes in levels of phosphorylated STAT1 andSTAT3 in the infracted area in the in vivo myocardiumexposed to I/R injury. The free radical scavenger,tempol, was shown to reduce STAT1 and STAT3 phos-phorylation. Previous studies have demonstrated thatadministration of tempol in the same manner decreasesinfarct size (8). Taken together, these data imply thatthe activation of STAT1 and 3 are important functionswithin the infarct area and may contribute to cardiacremodeling. Furthermore, the decrease in infarct sizemay be partly due to the inhibition of the JAK-STATpathway by tempol. The molecular pathway resulting incell death after free radical generation is complex andnot fully understood. Our studies also suggest that freeradical production may trigger distinct kinases that areresponsible for STAT1 or SATAT3 phosphorylationand that by reducing STAT phosphorylation by usingtempol treatment, the extent of myocardial damage isreduced, which could in part explain the reduction ininfarct size previously reported (8).

STAT1 has been shown to induce apoptosis after I/Rin vitro, whereas STAT3 was found to elicit opposite andthus protective effects (11, 12). It is paradoxical, there-fore, that both STAT1 and STAT3 phosphorylation wasreduced by tempol treatment in vivo. Although there isless overall activated STAT3, it may still demonstrate itsprotective effect. High levels of serine-phosphorylated

STAT3 were seen during ischemia that were not furtherenhanced upon reperfusion, but were reduced uponintroduction of tempol. In contrast, however, low-levelserine phosphorylation of STAT1 was observed duringischemia but diminished upon the onset of reperfu-sion. Phosphorylation of STAT on S727 has been shownto enhance transcriptional activity of both STAT1 andSTAT3 (9–11). Thus, the serine phosphorylation ofSTAT3 rather than STAT1 further suggests that STAT3is still functioning to protect the myocardium from I/Rinjury. The generation of STAT3 S727 mutant miceindicated that these animals were highly susceptible todoxorubicin-induced heart failure, signifying the im-portance of this residue (19). The reduction in phos-phorylation at S727 after reperfusion suggests rapidactivation of a distinct phosphatase as a possible feed-back mechanism to inhibit full STAT transcriptionalactivity.

Initiation of double-strand DNA breaks during apo-ptosis causes accumulation of the phosphorylated formof the H2AX (�H2AX) through activation of the ataxiatelangiectasia mutated (ATM) and ATM-related (ATR)kinases, leading to recruitment of DNA repair factors(20). Here we show for the first time that I/R injuryrapidly induces the activation of �H2AX in the risk areawhereas minimal �H2AX activation occurs in the non-risk area. We recently demonstrated a role of STAT1 inthe ATM-mediated pathway (21), and our present datamay suggest that STAT1 may also modulate the ATMpathway that is known to modulate the apoptoticprocess via the ATM-p53 pathway.

The present study suggests that during I/R injury, abalance exists between proapoptotic STAT1 and anti-apoptotic STAT3 that determines the response of thecell to I/R injury. Indeed, STAT3 has been shown toprotect cells from IFN-�-induced apoptosis by abrogat-ing the proapoptotic effects of STAT1 (23). Moreover,STAT3-deficient mice are more sensitive to I/R-in-

Figure 4. Tissue was fixed in 10% formaldehyde,mounted on wax blocks, cut into 4 �m sections,followed by immunohistochemical staining us-ing pSTAT1Y701 and pSTAT3Y705 antibodies. A)Tyrosine phosphorylation of STAT3 is signifi-cantly decreased on tempol treatment. B) Thesame result for tyrosine phosphorylated STAT1.

E1408 Vol. 20 October 2006 McCORMICK ET AL.The FASEB Journal

duced apoptosis and have larger infarct sizes. It wasshown recently that granulocyte colony-stimulating fac-tor (G-CSF) activates both STAT1 and STAT3 in car-diac myocytes (24). The protective effect of G-CSF inmyocardial infarction is mediated, however, throughSTAT3, which was activated to a greater extent thanSTAT1. Similarly, we demonstrated that increasing theratio of STAT3 to STAT1 in transfection experimentsprotected cardiac myocytes from hypoxia/reoxygen-ation-induced cell death (12). In the present study,injection of IFN-�, a potent activator of STAT1, exac-erbated cardiac damage as evidenced by increasedinfarct sizes. Furthermore, IFN-�-induced STAT1 acti-vation during reperfusion significantly reduced thecardioprotective effect of tempol, indicating that thisantioxidant’s antiapoptotic effect is likely to be medi-ated partly through reduction in activated STAT1.Since we previously showed that the cardioprotective

affects of a natural antioxidant green tea are mediatedthrough STAT1 (16), this highlights the role of STAT1as a general molecular target for antioxidants in theischemic myocardium.

J.M. and S.B. are Ph.D. students funded by the BritishHeart Foundation, London, UK.

REFERENCES

1. Maxwell, S. R, and Lip, G. Y. (1997) Reperfusion injury: a reviewof the pathophysiology, clinical manifestations and therapeuticoptions. Int. J. Cardiol. 58, 95–117

2. Eefting, F., Rensing, B., Wigman, J., Pannekoek, W. J., Liu,W. M., Cramer, M. J., Lips, D. J., and Doevendans P. A. (2004)Role of apoptosis in reperfusion injury. Cardiovasc. Res. 61,414–426

3. Olivevetti, G., Quani, F., Sala, R., Lagrasta C., Corradi, D.,Bonacina, E., Gambert, S. R, Cigoa, E., and Anvera, P. (1994)Acute myocardial infarction in humans is associated with theactivation of programmed myocyte cell death in the survivingportion of the heart. J. Mol. Cell. Cardiol. 28, 2005–2016

4. Stephanou, A., Brar, B. K., Scarabelli, T., Knight, R. A., andLatchman, D. S. (2001) Distinct initiator caspases are requiredfor the induction of apoptosis in cardiac myocytes in ischaemiaversus reperfusion injury, Cell Death Differentiation 2002 8, 434–435

5. Scarabelli, T. M, Stephanou, A., Pasini, E., Comini, L., Raddino,R., Knight, R A., and Latchman, D. S. (2002) Different signallingpathways induce apoptosis in endothelial cells and cardiacmyocytes during ischaemia/reperfusion injury. Circ. Res. 90,745–748

6. Kukreja, R. C., and Hess, M. L. (1992) The oxygen free radicalsystem: from equations through membrane-protein interactionsto cardiovascular injury and protection. Cardiovasc. Res. 26,641–655

7. Flaherty, J. T. (1991) Myocardial injury mediated by oxygen freeradicals. Am. J. Med. 91, 79S–85S

8. McDonald, M. C., Zacharowski, K., Bowes, J., Cuzzocrea, S., andThiemermann, C. (1999) Tempol reduces infarct size in rodentmodels of regional myocardial ischemia and reperfusion. FreeRadic. Biol. Med. 27, 493–503

9. Decker, T., Stockinger, S., Karaghiosoff, M., Muller, M., andKovarik, P. (2002) IFNs and STATs in innate immunity tomicroorganisms. J. Clin. Invest. 109, 1271–1277

10. O’Shea, J. J., Gadina, M., and Schreiber, R. D. (2002) Cytokinesignaling in 2002: new surprises in the Jak/Stat pathway. Cell 109Suppl., S121–S131

11. Stephanou, A. (2004) Role of STAT-1 and STAT-3 in ischaemia/reperfusion injury. J. Cell Mol. Med. 8, 519–525

12. Stephanou, A., Brar, B. K., Scarabelli, T. M., Jonassen, A. K.,Yellon, D. M., Marber, M. S., Knight, R. A., and Latchman, D. S.(2000) Ischemia-induced STAT-1 expression and activation playa critical role in cardiomyocyte apoptosis. J. Biol. Chem. 275,10002–10008

13. Stephanou, A., Scarabelli, T., Brar, B. K., Nakanishi, Y., Mat-sumura, M., Knight, R. A., and Latchman, D. S. (2001) Induc-tion of apoptosis and Fas/FasL expression by ischaemia/reper-fusion in cardiac myocytes requires serine 727 of the STAT1 butnot tyrosine 701. J. Biol. Chem. 276, 28340–28347

14. Stephanou, A., Scarabelli, T., Townsend, P. A., Bell, R., Yellon,D. M., Knight, R. A., and Latchman, D. S. (2002) The carboxyl-terminal activation domain of the STAT-1 transcription factorenhances ischaemia/reperfusion-induced apoptosis in cardiacmyocytes. FASEB J. 16, 1841–1843

15. Menegazzi, M., Tedeschi, E., Dussin, D., de Prati, A. C., Cava-lieri, E., Mariotto, S., and Suzuki, H. (2001) Anti-interferongamma action of epigallocatechin-3-gallate mediated by specificinhibition of STAT1 activation. FASEB J. 15, 1309–1311

16. Townsend, P. A., Scarabelli, T. M., Pasini, E., Gitti, G., Me-negazzi, M., Suzuki, H., Knight, R. A., Latchman, D. S., andStephanou, A. (2004) Epigallocatechin-3-gallate inhibits STAT-1

Figure 5. Prolonged STAT1 activation with IFN-� overcomesthe cardioprotective effects of tempol. A) Infarct size wasmeasured in rats subjected to the surgical procedure alone(sham n�5) or LAD coronary artery occlusion (25 min) andreperfusion (2 h) with infusion (5 min prior to reperfusion)of either saline (I/R n�7), tempol (n�7), IFN-� (n�6), ortempol and IFN-� together (n�6). *P � 0.05 compared withI/R vehicle. B) Western blot analysis of STAT1 phosphoryla-tion from heart tissues extracted from the experiment inpanel A Similar blots were observed from 3 different heartsform each group.

E1409FREE RADICAL SCAVENGING IN I/R INJURY

activation and protects cardiac myocytes from ischemia/reper-fusion-induced apoptosis. FASEB J. 18, 1621–1623

17. Stefanutti, G., Pierro, A., Vinardi, S., Spitz, L., and Eaton, S.(2005) Moderate hypothermia protects against systemic oxida-tive stress in a rat model of intestinal ischemia and reperfusioninjury. Shock 24, 159–164

18. Hammond, E. M., Dorie, M. J., and Giaccia, A. J. (2003)ATR/ATM targets are phosphorylated by ATR in response tohypoxia and ATM in response to reoxygenation. J. Biol. Chem.278, 12207–12213

19. Shen, Y., La Perle, K. M., Levy, D. E., and Darnell, J. E., Jr.(2005) Reduced STAT3 activity in mice mimics clinical diseasesyndromes. Biochem. Biophys. Res. Commun. 330, 305–309

20. Thiriet, C., and Hayes, J. J. (2005) Chromatin in need of a fix:phosphorylation of H2AX connects chromatin to DNA repair.Mol. Cell. 18, 617–622

21. Townsend, P. A., Cragg, M. S., Davidson, S. M., McCormick,J., Barry, S., Lawrence, K. M., Knight, R. A., Hubank, M.,Chen, P. L., Latchman, D. S., and Stephanou, A. (2005)STAT-1 facilitates the ATM activated checkpoint pathwayfollowing DNA damage. J. Cell Sci. 118, 1629 –1639 Epub 2005Mar 22.

22. Hilfiker-Kleiner, D., Hilfiker, A., Fuchs, M., Kaminski, K.,Schaefer, A., Schieffer, B., Hillmer, A., Schmiedl, A., Ding, Z.,and Podewski, E., et al. (2004) Signal transducer and activator oftranscription 3 is required for myocardial capillary growth,control of interstitial matrix deposition, and heart protectionfrom ischemic injury. Circ. Res. 95, 187–195 Epub 2004 Jun 10

23. Shen, Y., Devgan, G., Darnell, J. E., Jr., and Bromberg, J. F.(2001) Constitutively activated Stat3 protects fibroblasts fromserum withdrawal and UV-induced apoptosis and antagonizesthe proapoptotic effects of activated Stat1. Proc. Natl. Acad. Sci.U. S. A. 98, 1543–1548

24. Harada, M., Qin, Y., Takano, H., Minamino, T., Zou, Y., Toko,H., Ohtsuka, M., Matsumura, K., and Sano, M., et al. (2005)G-CSF prevents cardiac remodeling after myocardial infarctionby activating the Jak-Stat pathway in cardiomyocytes. Nat. Med.11, 305–311 Epub 2005 Feb 20

25. Peterson, G. L. (1977) A simplification of the protein assaymethod of Lowry et al. which is more generally applicable. Anal.Biochem. 83, 346–356

Received for publication March 30, 2006.Accepted for publication May 8, 2006.

E1410 Vol. 20 October 2006 McCORMICK ET AL.The FASEB Journal

The FASEB Journal • FJ Express Summary

Free radical scavenging inhibits STAT phosphorylationafter in vivo ischemia/reperfusion injury

James McCormick,*,1 Sean P. Barry,*,1 Ahila Sivarajah,† Giorgio Stefanutti,§

Paul A. Townsend,‡ Kevin M. Lawrence,† Simon Eaton,§ Richard A. Knight,*Christoph Thiemermann,† David S. Latchman,* and Anastasis Stephanou*,2

*Department of Medical Molecular Biology, §Department of Surgery, Institute of Child Health,University College London, London, UK; †Centre for Experimental Medicine, Nephrology andCritical Care, The William Harvey Research Institute, St. Bartholomew’s and The Royal LondonSchool of Medicine and Dentistry, Queen Mary, University of London, Charterhouse Square,London, UK; and ‡Human Genetics Division, Southampton General Hospital, University ofSouthampton, Southampton, UK

To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.05-6188fje

SPECIFIC AIMS

We previously demonstrated that STAT1 plays an im-portant role in promoting apoptotic cell death incardiac myocytes after ischemia/reperfusion (I/R) in-jury. The mechanism and pathways of STAT activationafter I/R are still unclear. Using an in vivo model, wedemonstrate that STAT1 and STAT3 phosphorylationand myocardial infarction can be reduced by the freeradical scavenger, tempol. Moreover, IFN�, a potentactivator of STAT1, not only enhances infarct size butcan overcome the protective effects of tempol. Theresults presented here demonstrate a potential mecha-nism of action for tempol and indicate the therapeuticuse of free radical scavengers for targeting STAT1activation in postischemic insults.

PRINCIPAL FINDINGS

1. STAT1 and STAT3 are phosphorylated in a time-dependent manner after I/R injury in vivo

Using an in vivo coronary occlusion model, we per-formed a time course of I/R injury on the rat myocar-dium in order to assess the dynamics of STAT1 andSTAT3 tyrosine or serine phosphorylation over time. Asshown in Fig. 1A, STAT1 tyrosine phosphorylation(Y701), rather than STAT1 serine phosphorylation(S727), was induced 15 min after reperfusion in the riskarea (infarct area) compared with the nonrisk area andreached a maximum at 30 min. Levels of STAT1 Y701phosphorylation decreased dramatically at 60 min ofreperfusion and remained unchanged for up to 120min. Thus, STAT1 activation occurred predominantlyin the risk area, the main area of reactive oxygenspecies (ROS) generation, highlighting the role ofactivated STAT1 in the damaged myocardium.

In contrast to STAT1, STAT3 tyrosine 705 (Y705)phosphorylation was seen during ischemia and afterthe reperfusion phase, reaching a maximum at 30 min(Fig. 1A). Similarly, STAT3 S727 phosphorylation wasobserved during ischemia and after the reperfusion(Fig. 1B).

2. STAT1 and STAT3 phosphorylation is inducedby ROS

ROS generation is responsible for intracellular damageand has been shown to lead to the onset of apoptosis.To verify that tempol had reduced the concentration ofROS within the intracellular environment, we mea-sured the cellular concentration of malondialdehyde(MDA). MDA is a marker of lipid peroxidation thatoccurs as a result of the damaging effects of ROS. Wefound that injection of tempol reduced the concentra-tion of MDA to sham levels compared with the vehiclecontrol, confirming that free radical concentration hadbeen successfully inhibited during I/R (see Fig. 3A,full-length article). Comparing the sham-operated tothe vehicle and tempol-treated animals in the nonriskarea, no statistically significant difference in the MDAlevels (P��0.2) was observed, indicating that I/Rinjury does not effect the nonrisk area (see Fig. 3A,full-length article).

The decrease in MDA levels was associated withdecreased tyrosine phosphorylation of both STAT1 andSTAT3 (see Fig. 3B, full-length article). Immunohisto-chemical staining of tissue sections from the risk area

1These authors contributed equally to this work.2Correspondence: Department of Medical Molecular Biol-

ogy, Institute of Child Health, University College Lon-don, 30 Guilford St., London, WC1N 1EH, UK. E-mail:a.stephanou@ich.ucl.ac.uk

doi: 10.1096/fj.06-6188fje

21150892-6638/06/0020-2115 © FASEB

confirms that I/R-induced STAT phosphorylation,which was predominantly nuclear, was indeed reducedby free radical scavenging (see Fig. 4, full-length arti-cle). These data imply that STAT1 and STAT3 phos-phorylation in I/R can be attenuated by the freescavenging/antioxidant property of tempol, and thismay allow tempol to protect against cell death on theonset of reperfusion.

3. Activated STAT1 contributes to infarctdevelopment, in vivo

Although activated STAT1 is induced after I/R injuryand has been implicated in promoting apoptosis, it wasunclear whether this activation contributed to infarctdevelopment, in vivo. To test this, rats were subjected toI/R injury with injection of the potent STAT1 activator,IFN-�. IFN-� was introduced to allow continual activa-tion of STAT1 for up to 120 min (postischemia), atwhich point the infarct area can be measured. Baselinevalues of mean arterial blood pressure (MAP), heartrate, and pressure rate index (PRI) were not signifi-cantly different between groups (data not shown).Compared with sham-operated animals, 25 min of LAD

occlusion followed by 2 h of reperfusion did notcause significant alterations in MAP, heart rate, orpressure rate index compared with their respectivecontrols. Neither IFN-� alone or IFN-� plus tempolhad any significant effect on MAP, HR, or PRI (datanot shown).

Administration of IFN-� (25 �g/kg) 5 min beforereperfusion caused an increase of 30% in myocardialinfarct size compared with vehicle (Fig. 2A); IFN-�induced activation of STAT1 in the risk area wasconfirmed by Western blot (Fig 2B, lower panel). In thepresence of IFN-�, tempol no longer affected STAT1phosphorylation, and IFN-� completely abrogated thecardioprotective affect of tempol. This strongly impli-cates STAT1 activation as a main target in tempol-mediated cardioprotection.

Figure 1. Time course of STAT1 and STAT3 phosphorylationafter ischemia/reperfusion injury. Protein lysates from heartwere separated on SDS-PAGE gels and probed with antibodiesfor A) total STAT1, phospho STAT1Y701, and phosphoSTAT1S727 and B) total STAT3, phospho STAT3Y705, and phos-pho STAT3727. Equal protein loading was confirmed by probingwith an antibody (Ab) to GAPDH.

Figure 2. Prolonged STAT1 activation with IFN-� overcomes thecardioprotective effects of tempol. A) Infarct size was measuredin rats subjected to the surgical procedure alone (sham n�5) orto LAD coronary artery occlusion (25 min) and reperfusion (2h) with infusion (5 min before reperfusion) of either saline (I/Rn�7), tempol (n�7), IFN-� (n�6), or tempol and IFN-�together (n�6). *P � 0.05 compared with I/R vehicle. B)Western blot analysis of STAT1 phosphorylation fromhearts tissues extracted from the experiment of panel A.Similar blots were observed from 3 different hearts formeach group.

2116 Vol. 20 October 2006 McCORMICK ET AL.The FASEB Journal

CONCLUSIONS AND SIGNIFICANCE

In the present study we show for the first time signifi-cant changes in the levels of phosphorylated STAT1and STAT3 in the infracted area in the in vivo myocar-dium exposed to I/R injury. The free radical scavenger,tempol, was shown to reduce STAT1 and STAT3 phos-phorylation. Previous studies have demonstrated thatadministration of tempol in the same manner decreasesinfarct size. Taken together, these data imply that theactivation of STAT1 and 3 are important functionswithin the infarct area and may contribute to cardiacdamage. Furthermore, the decrease in infarct size maybe partly due to the inhibition of the janus-kinase(JAK)-STAT pathway by tempol. The molecular path-way resulting in cell death following free radical gener-ation is complex and not fully understood. Our studiessuggest that free radical production may trigger distinctkinases, which are responsible for STAT1 or SATAT3phosphorylation, and that by reducing STAT phos-phorylation with tempol treatment, the extent of myo-cardial damage is reduced, which could explain in partthe reduction in infarct size as reported.

The present study suggests that during I/R injury abalance exists between proapoptotic STAT1 and anti-apoptotic STAT3 that determines the response of thecell to I/R injury (Fig. 3). Indeed, STAT3 has beenshown to protect cells from IFN-�-induced apoptosis byabrogating the proapoptotic effects of STAT1. More-over, STAT3-deficient mice are more sensitive to I/R-induced apoptosis and have larger infarct sizes. It wasrecently shown that granulocyte colony-stimulating fac-tor (G-CSF) activates both STAT1 and STAT3 in car-diac myocytes. The protective effect of G-CSF in myo-cardial infarction, however, is mediated throughSTAT3, which was activated to a greater extent thanSTAT1. Similarly, we previously demonstrated that in-creasing the ratio of STAT3 to STAT1 in transfectionexperiments protected cardiac myocytes from hypoxia/reoxygenation-induced cell death. In the present study,injection of IFN-�, a potent activator of STAT1, exac-

erbated cardiac damage, as evidenced by increasedinfarct sizes. Furthermore, IFN-�-induced STAT1 acti-vation during reperfusion significantly reduced thecardioprotective effect of tempol, indicating that thisantioxidant’s antiapoptotic effect is likely to be medi-ated partly through a reduction in activated STAT1.Since we previously showed that the cardioprotectiveaffects of a natural antioxidant green tea are mediatedthrough STAT1, this highlights the role of STAT1 as ageneral molecular target for antioxidants in the ischemicmyocardium.

Figure 3. Overview of the JAK/STAT pathway activated afterischemic preconditioning (IP) or ischemia/reperfusion(I/R) injury. Free radical production (reactive oxygen spe-cies) is known to mediate the damaging effects of IR. Theextent of cardiac myocyte damage after I/R as a result of celldeath will depend on the balance and levels of activatedSTAT1 (pST1) or STAT3 (pST3) modulating pro- and anti-apoptotic genes. Agents that act as ROS scavengers such astempol may also modulate the functional activity of STATs.

2117FREE RADICAL SCAVENGING IN I/R INJURY