Journal of Molecular and Cellular Cardiology 43 (2007) 130–136www.elsevier.com/locate/yjmcc

Original article

Positive effect of darbepoetin on peri-infarction remodeling in a porcinemodel of myocardial ischemia–reperfusion

Catalin Toma ⁎, Dustin P. Letts, Masaki Tanabe, John Gorcsan III, Peter J. Counihan

Cardiovascular Institute, University of Pittsburgh Medical Center, Scaife Hall S559, 200 Lothrop Street, Pittsburgh, PA 15213, USA

Received 1 April 2007; received in revised form 25 April 2007; accepted 14 May 2007Available online 24 May 2007

Abstract

Erythropoietin (Epo) has anti-apoptotic and pro-angiogenic effects in rodent models of myocardial infarction (MI). We tested the hypothesisthat a long-acting Epo derivative (darbepoetin) has a beneficial effect on infarct size and peri-infarct remodeling in a clinically relevant largeanimal model of ischemia–reperfusion. A human acute MI scenario was simulated in 16 domestic pigs by inflating an angioplasty balloon in theproximal left circumflex (LCx) artery for 60 min. The animals were randomized to darbepoetin 30 μg/kg iv or placebo (saline) at the time ofreperfusion. Treatment with darbepoetin did not lead to a reduction in the infarct size at 2 weeks as assessed by histology (30.3±1.8% of thevolume at risk for placebo vs. 33.2±2.5% for darbepoetin). However, significant effects were seen in the peri-infarct region. Histologicalevaluation revealed decreased interstitial fibrosis (6.8±0.7% of myocardial sections area vs. 9.6±0.7%, p=0.02) and increased average capillaryarea (106±3% of the non-infarcted myocardium vs. 89±4%, p=0.003) in the treatment arm in the absence of significant cardiac hypertrophy.This resulted in preserved regional wall motion as assessed by tissue Doppler-derived radial strain (subepicardial radial strain 90.1±21.2% fordarbepoetin vs. 20.3±10.1% for placebo, pb0.05). However, this did not translate to improved wall thickening (126.5±6.0% of diastolicthickness for darbepoetin vs. 119.8±5.4% for placebo, p=NS). Beneficial effects of darbepoetin to peri-infarct remodeling were observed in aclinically relevant model of ischemia–reperfusion. Although the infarct size was not reduced, there was a limited decrease in interstitial fibrosis,increased capillary area and regional functional improvement in darbepoetin-treated animals.© 2007 Elsevier Inc. All rights reserved.

Keywords: Erythropoietin; Myocardial infarction; Remodeling; Fibrosis; Angiogenesis

1. Background

Abrupt occlusion of a coronary artery initiates a cascade ofevents beginning with hypoxic cell death, followed byprolonged adverse remodeling over a period of weeks, withsubsequent systolic dysfunction. This process involves bothapoptosis of the cardiac cells and increased cardiac fibrosis [1]as well as rearrangement of myocardial fibers [2]. Althoughearly restoration of flow remains the best strategy at limitinginfarct size, apoptotic cell death continues secondary toreperfusion injury, which sparked significant interest onstrategies aimed at limiting this phenomenon [3].

Erythropoietin (Epo) is a glycoprotein hormone secreted bythe kidney in response to hypoxia, leading to enhanced

⁎ Corresponding author. Tel.: +1 412 647 3429; fax: +1 412 647 0481.E-mail address: tomac@upmc.edu (C. Toma).

0022-2828/$ - see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.yjmcc.2007.05.014

erythropoiesis. The Epo receptor (EpoR) has been identified inseveral tissues outside the bone marrow, most notably in the heartand brain [4]. Experimental data have demonstrated thatactivation of EpoR in isolated cardiomyocytes decreases therate of apoptosis induced by hypoxic stress [5,6]. In vitro thiseffect is mediated by the PI3K/Akt [6], although ERK-mediatedsignaling may play the predominant role post-myocardialinfarction (MI) [7]. Epomay also have a role in local angiogenesis[8], as well as in the reduction of inflammation following thereperfusion injury [9]. Furthermore, recombinant erythropoietinadministered at the time of infarction was shown to significantlyreduce the infarct size and to improve the LV performance inrodent models of MI [5–7,10–12]. However, these findings werenot consistently reproduced in large animal models ofMI [13,14].In addition, since most of these data were obtained in chroniccoronary occlusionmodels, the potential for clinical application inreperfused MIs remains unclear. Accordingly, our objective was

131C. Toma et al. / Journal of Molecular and Cellular Cardiology 43 (2007) 130–136

to test the hypothesis that the long-acting erythropoietin analogue,darbepoetin, has favorable effects on the infarct size and post-MIremodeling in a clinically relevant porcine model of ischemia–reperfusion.

2. Methods

2.1. Porcine model of ischemia–reperfusion

Male crossbred farm pigs (20–30 kg, n=16) were used in thisstudy. All experiments were done in accordance with theUniversity of Pittsburgh institutional guidelines. The pigs weresedated, intubated and anesthesia was provided using 1.5%inhaled isoflurane. The animals received 300 mg aspirin rectallyand amiodarone iv (30 mg bolus, followed by infusion at 1 mg/min) to prevent the lethal ventricular arrhythmias associated withthis model. Arterial access was obtained by placing 6 Frenchsheath in the right femoral artery. Heparin 10,000 U wasadministered iv. The left main coronary artery was selectivelyengaged using an AL 1 guide catheter positioned across the top ofthe aortic arch. Using a 4-mm compliant angioplasty balloon, theproximal left circumflex artery was occluded for 60 min byinflating the balloon at low pressure (3–4 atmospheres). Theanimals were randomly assigned to either placebo (saline) ordarbepoetin (Amgen Inc., Thousand Oaks, CA) 30 μg/kgadministered iv 5 min prior to reperfusion. The operator wasblinded to the treatment assignment. The animals were thenallowed to recover and returned to their cages. Blood forhematological variables was obtained at baseline and prior tosacrifice.

2.2. Histological analysis

At 2 weeks following the MI, the animals were sedated,intubated, and the heart exposed via a lateral thoracotomy. Thepericardium was opened, and a suture was placed around theproximal/ostial circumflex artery in the same location where theoriginal balloon inflation was performed. Heparin 10,000 U wasadministered iv. The hearts were arrested with KCl andretroperfused with Evans Blue ex vivo at 100 cm H2O pressureto allow for determination of the area at risk. The ventricles weresectioned in 1-cm-thick slices and incubated in triphenyltetra-zolium chloride (TTC) for 10 min to quantify the scar. Thesections were then photographed on both sides with a digitalcamera and area at risk, as well as scar area were measured usingthe NIH Image J software. The myocardial volume at risk (VAR)and the scar volume were derived from these measurements bymultiplying the average areas by the slice thickness.

Samples of myocardium from the infarct area, the peri-infarctregions on both sides, and the interventricular septum (used ascontrol) were fixed in 4% formalin and paraffin-embedded forhistology. A rabbit polyclonal antibody (clone H-194, SantaCruz, CA) was used for EpoR detection. After incubationovernight with the primary antibody at a 1:25 dilution, aperoxidase-based secondary detection system was used (Envi-sion, Dako, Carpinteria, CA) coupled to the 3-amino-9ethylcar-bazole (AEC) chromogen (red color on the slides).

Collagen staining was performed using a modified Masson'sstain from which the nuclear and cytoplasmic stains were omittedfor clear delineation of the fibrotic areas. Morphometric analysisof myocyte size was obtained by measuring the myocyte cross-sectional area at the level of the nucleus (at least 75 individualmeasurements per animal). To identify vascular endothelium,formalin-fixed tissue sections were incubated with biotinylatedlectin (Griffonia Simplicifolia Lectin II at 1:100 dilution)followed by staining with the ABC system (Vector Labs,Burlingame, CA) forming a brown precipitate in the presenceof the primary ligand. Photographs were obtained at 20× forcollagen and 40× for capillary density determinations (at least 5random fields for each data point) and quantified using Image Jsoftware (NIH). Only areas where the myocytes were cross-sectioned were used for capillary morphometry. The density(number of capillaries/40× field) and total capillary area/40× field(0.1mm2)weremeasured and reported relative to the values in theseptum.

2.3. Echocardiography

Transthoracic echocardiograms were obtained at baseline,during the circumflex occlusion, and at the time of sacrificeusing a Toshiba Aplio platform (SSA-770A) with tissueDoppler (TD) capabilities and a 3.0-MHz transducer. Short-and long-axis views were obtained from a right parasternalapproach, and 4 and 2 chamber views were obtained from theapex. Wall and cavity dimensions were determined from short-axis views at the level of papillary muscle allowing forcalculation of fractional area shortening. The apical views wereused to calculate the ejection fraction (EF) offline using amodified Simpson approach.

Digitally stored TD short-axis data were analyzed offlineusing custom software (ApliQ, Toshiba Medical Systems Corp).Angle correction was performed as previously described [15]. Acontractile center was manually set in the center of the short-axissection at the mitral valve level. Radial Lagrangian strain wascalculated as time integral of velocity gradient that wasmeasured along the radii of a distance towards the contractilecenter between 2 pixels situated 3 mm apart and visualized as acolor-coded map. The data for the septal wall and postero-lateralwall (the infarcted wall, opposite to the transducer) wererepresented in M-mode using an average of 5 sampling lines1.5 mm apart (total width of the segment investigated 7.5 mm).The color-coded map was used to aid the placement of theM-mode analysis in the postero-lateral wall at the level of theinfarction. The beginning (endocardial) and end (epicardial)points were selected 1–2 mm from the respective surface toensure accurate tracking by the automated software throughoutthe cardiac cycle. Data were divided and averaged into the inner(subendocardial) and outer (subepicardial) myocardial halves.

2.4. Statistics

All data are expressed as mean±SEM. For multiple com-parisons, a two-way ANOVA was performed with a post hoct-test with the Bonferroni correction where appropriate. Two-

Fig. 1. (A) Determination of infarct size and volume at risk (typical exampleshown). The arrows indicate the areas sampled for histology. (B) The volume atrisk (VAR) was similar in both groups, as well as the infarct size (scar) relative tothe VAR or relative to the LV mass (p=NS).

132 C. Toma et al. / Journal of Molecular and Cellular Cardiology 43 (2007) 130–136

tailed unpaired Student's t-test with equal variance was used forcomparisons between continuous variables for data pairs. Chi-squared test was used for comparing categorical variables(mortality). A p-value less than 0.05 was considered significantunless otherwise specified.

3. Results

3.1. Myocardial infarction model

There were 5 deaths in total (31%, 4 in the placebo groupand 1 in the darbepoetin group, p=NS); all the observedmortality occurred within 24 h following the myocardialinfarction and was likely due to ventricular arrhythmias. Thefinal analysis includes 6 animals in the treatment arm and 5 inthe control arm. At 2 weeks, there was no significant increase inthe hemoglobin in the treated animals (10.8±0.7 g/dl at base-line vs. 11.3±0.6 g/dl at 2 weeks); similarly, the white cell andplatelet counts were not significantly affected by darbepoetin(WBC 15.6±1.7 109/l vs. 17.0±1.7 109/l; platelets 299±55109/l vs. 434±89 109/l, baseline vs. 2 weeks, n=6). Similarly,no significant differences were observed between the darbe-poetin-treated animals and placebo at 2 weeks (placebo: Hb9.8±0.5, WBC 14.1±1.3 109/l, platelets 316±50 109/l, n=5).

The volume at risk (VAR) obtained by proximal circumflexocclusion represented approximately 1/3 of the left ventricularvolume (30.3±1.8% for placebo and 33.2±2.5% for darbepoe-tin). Most of the infarcted tissue was limited to thesubendocardial region in both groups (non-transmural infarcts,Fig. 1A). There were no differences between the infarct size (IS)relative to the VAR between the treatment and the placebo arms(32.5±3.6% vs. 28.9±3.4%, p=0.49, Fig. 1B).

3.2. Histology data

EpoR staining indicated strong expression of this protein incardiomyocytes and to a lesser extent in the capillaries (Fig. 2).The protein was similarly detected in the peri-infarct zone, aswell as in the remote myocardium (Figs. 2A and B,respectively). No staining was observed in the fibroblasts inthe scar region. The staining was both cytoplasmic and at thelevel of the cell membrane (Fig. 2).

A dense fibrotic scar was identified in the postero-lateral wallof all animals with a rim of viable tissue in the subepicardialregion. Morphometric analysis indicated a small non-significantincrease in average myocyte cross-sectional area in the peri-MIregion compared to the remote myocardium in both groups (Fig.3A). There were no significant differences between the twotreatment groups (septum vs. postero-lateral wall, 281±45 μm2

vs. 330±60 μm2 in the placebo arm, 222±27 μm2 vs. 314±52 μm2 in the darbepoetin arm, p=NS between groups).Staining for collagen revealed increased interstitial fibrosis inthe zone bordering the scar in both groups (Fig. 3B) with anincrease in the collagen septation of the muscle fibers. Theamount of interstitial fibrosis was significantly decreased by30% in the animals receiving darbepoetin compared to controls(6.8±0.7% vs. 9.6±0.7% of myocardium, pb0.05, Fig. 3B).

Analysis of capillary density area in the peri-infarct regionrevealed a decreased number of capillaries in the both groups(placebo: 156±11 vs. 200±20 capillaries/field; darbepoetin165±8 vs. 195±15 capillaries/field, n=6, p=0.016) relative tothe non-infarcted myocardium. When normalizing the peri-infarct data to the septal values in each animal, there we nosignificant differences in capillary density between the twogroups (87±7% of septal capillary density vs. 78±5% dar-bepoetin vs. placebo respectively, p=NS). However, a sig-nificantly higher total capillary cross-sectional area in the peri-infarct region was found with darbepoetin vs. placebo (106±3%of septal capillary area vs. 89±4% respectively, p=0.003, Fig.4), consistent with growth of preexisting microvessels.

3.3. Echocardiography data

We have evaluated the LV function by 2D echocardiographyand radial strain analysis at baseline, during the circumflexocclusion, and at 2 weeks. Acutely during circumflex occlusiona significant decrease in the global LV function was observed(EF 49.4±1.5% down to 32.4±2.1%, pb0.01). However, at2 weeks a significant recovery of the global LV function wasobserved in both groups (EF at 2 weeks 48.8±2.6% in the

Fig. 3. (A) Morphometric analysis revealed no differences in the averagemyocyte size in the peri-infarct region relative to the non-infarcted territory(septum). (B) Staining for collagen revealed significantly increased fibrosis peri-infarct compared to the septum in the placebo group. With darbepoetin, peri-infarct fibrosis was significantly reduced by approximately 30% (pb0.05) tovalues comparable to the septal region.

Fig. 2. Expression of erythropoietin receptor (EpoR) in the heart in the peri-infarct region (A) and in the remote non-infarcted myocardium (B). Apolyclonal anti-human EpoR antibody was used, coupled to the Envision/AEC secondary detection system (positive staining is red). There was abundantEpoR expression primarily in the cardiomyocytes, with staining detected incapillaries as well (arrows in panel A).

133C. Toma et al. / Journal of Molecular and Cellular Cardiology 43 (2007) 130–136

placebo arm and 47.3±2.8 in the darbepoetin arm, Table 1,p=NS vs. baseline). Postero-lateral wall thickening decreasedsignificantly during the circumflex occlusion (from 138.0±5.0% of the diastolic wall thickness to 97.0±2.5%). This wasaccompanied by a compensatory increase in the regionalthickening in the opposing septal wall (from 130.0±4.4% to140.9±2.5%). Mixed design ANOVA revealed a significantdecrease in the postero-lateral wall thickening in both groups2 weeks post-MI (pb0.05), with no significant differencebetween placebo and darbepoetin (Table 1).

M-mode radial strain data revealed increased radial strain atbaseline in the subendocardial region compared to thesubepicardial myocardium consistent with prior tissue Dopplerdata [16]. During circumflex occlusion, radial strain in thepostero-lateral wall in both regions decreased to almost zero(Fig. 5). Again, a compensatory increase in the strain valuesrecorded in the opposing septal wall was noted (data notshown). At 2 weeks, the strain in the subendocardial regionremained significantly decreased (scar region) in both groups;in the subepicardial region (peri-infarct area) the strain raterecovered to above-baseline values in the darbepoetin arm but

remained significantly decreased in the placebo arm comparedto baseline (subepicardial radial strain 90.1±21.2% fordarbepoetin vs. 30.3±12.2% for placebo, pb0.05, Fig. 5).

To address whether the observed histological variablescorrelate with contractile function, differential strain (defined assubepicardial strain–subendocardial strain) was used to char-acterize the relative functional improvement in the peri-MI zonein each animal. In the darbepoetin-treated animals, we found astrong inverse linear correlation between the amount of collagenand differential strain (r2 =0.86, p=0.008). On the other hand,there was only a weak correlation between capillary size anddifferential strain (r2 =0.40, p=0.17).

4. Discussion

Our results indicate that darbepoetin administration at thetime of reperfusion did not lead to a chronic reduction in infarctsize, but had a positive remodeling effect in the survivingmyocardium bordering the scar.

Previous rodent data demonstrate a significant reduction ofthe scar with erythropoietin, even when administered after theacute event [5,6,9–12]. In large animals, as well as in our study,the effects are less robust. Hirata and colleagues showed that

Fig. 5. Radial strain distribution in the lateral wall at baseline, during circumflexocclusion and at 2 weeks after MI in placebo (A) vs. darbepoetin (B). The tissueDoppler-derived strain was higher at baseline in the inner (circles) vs. outer halfof the myocardium (squares). During circumflex occlusion, the strain valuesdecreased significantly in both regions close to zero (pb0.01). At 2 weeks, thestrain in the subendocardial regions remained significantly decreased in bothgroups (pb0.05) consistent with the presence of a scar in that region. In the outerlayer of the postero-lateral wall myocardium (peri-infarct zone) the strain valueremained low in the placebo group, but recovered significantly to values aboveaverage in the darbepoetin group (*pb0.05 vs. baseline and vs. placebo).

Fig. 4. (A) Analysis of the capillary size and density revealed increased averagecapillary size in the peri-infarct area compared to the remote myocardium in thedarbepoetin-treated animals (typical example shown). (B) When normalizing forthe values measured in the remote non-infarcted territory (septum), the totalcapillary area was significantly increased in the darbepoetin-treated animalscompared to placebo (p=0.003).

134 C. Toma et al. / Journal of Molecular and Cellular Cardiology 43 (2007) 130–136

Epo administered to dogs at the time of permanent coronaryligation led to a reduction of infarct size, but not when the drugwas given 6 h after the acute event [17]. The current standard ofcare for MI is prompt reperfusion, and the above observationsmay not be equally relevant in this setting. Indeed, recent data inpigs [13] and sheep [14] involving a temporary coronaryocclusion failed to demonstrate a reduction in infarct size withEpo at the time of reperfusion. Our results are in agreement withthese findings. This is an important observation with respect totranslating the protective effect of Epo to clinical applications as

Table 1Echocardiographic parameters at baseline, during the circumflex occlusion andafter 2 weeks for placebo and darbepoetin

Baseline LCx occlusion Placebo Darbepoetin

EF (%) 49.4±1.5 32.4±2.1 48.8±4.5 47.3±2.8LVEDD (mm) 38.4±0.9 36.2±1.1 43.3±2.7 43.9±1.2Septal thickening (%) 130.0±4.4 141.0±7.3 138.3±6.2 131.1±4.5PLW thickening (%) 138.0±5.0 97.0±2.5 119.8±5.4 126.5±6.0

The data at baseline and during circumflex occlusion represent the average forboth groups. The ejection fraction (EF) decreased during circumflex occlusion,and recovered to baseline values in both groups. There was small but significantincrease in LV end-diastolic diameter (LVEDD) in both groups (pb0.05). Theregional thickening (systolic thickness expressed as percent of diastolicthickness) in the postero-lateral wall (PLW) was significantly decreased inboth groups post-MI (p=0.02 by ANOVA) compared to baseline with nosignificant improvement with darbepoetin.

our model reproduces the current management of acute MI inhumans, with mechanical reperfusion and concomitant use ofantiplatelet and anticoagulation therapies. The observed dis-crepancy between large animal and rodent data may be relatedto differences in myocardial perfusion between the models, as isconceivable that the thinner rodent myocardium is moreresistant to ischemia with greater potential for recovery thanthe thick porcine LV wall. Of note is a recent small phase I trialinvolving administration of darbepoetin in patients with acuteMIs demonstrating a lack of change in the infarct size but asuggestion of beneficial remodeling in the surviving myocar-dium [18].

Unlike Epo, darbepoetin is longer acting and may confersome benefit on post-MI remodeling despite of apparent lack ofeffect on acute cell death. A recent study in rats undergoingischemia–reperfusion revealed that the darbepoetin cardiopro-tection was achieved at similar doses to the ones used in thisstudy [12]. Our data demonstrate a consistent structural andfunctional improvement in the peri-infarct myocardium in theabsence of a significant change in the hemoglobin. The amountof interstitial fibrosis was significantly decreased compared toplacebo in agreement with data from other groups [11,19,20].

135C. Toma et al. / Journal of Molecular and Cellular Cardiology 43 (2007) 130–136

Data in isolated cardiac fibroblasts indicate that the Epointracellular signaling is present with the activation of the Jak2-STAT3 pathway [21]. However, we did not detect significantEpoR expression in fibroblasts in our study. Another possibilityis that Epo has an anti-inflammatory effect, with decreasedaccumulation of macrophages at the infarct site and decreasedsubsequent fibrosis [9]. To put these observations in perspective,prior histology data indicate that interstitial and replacementfibrosis accounts for significantly more collagen accumulationthan the post-MI scar with the development of ischemic cardio-myopathy [1].

The angiogenic effect of erythropoietin has been welldocumented, with mobilization of bone marrow endothelialprogenitor cells documented in response to darbepoetin [8,17].Epo also stimulates growth of preexisting cardiac capillaries, tothe same extent as VEGF [22], and has been shown to protectthe microvascular endothelium to ischemic injury [23]. Inrodent models of MI, erythropoietin administration leads to anincreased capillary density [24]. In our model, a loss ofcapillaries in the peri-infarct area was apparent, likely due toischemic endothelial cell death. With darbepoetin, we foundonly a trend towards increased capillary density, however asignificant increase in total capillary area was seen (Fig. 4).These findings are consistent with stimulation of growth ofpreexisting microvessels, as previously reported [22]. Our datawith darbepoetin are remarkably similar to the angiogeniceffects of VEGF-D in skeletal muscle, where increased vasculararea correlated with increased flow in the region [25].

Staining for the EpoR revealed abundant expression of thisprotein primarily in the cardiomyocytes (Fig. 2). Although thespecificity of the antibody used in this study was recentlyquestioned [26], this is unlikely to represent non-specific bindingof the primary antibody, as the staining pattern was similar to theone obtained with several different primary antibodies, with mostof the receptor found internalized as previously reported [27].

The above histological observations, in particular theantifibrotic effect, correlated well with regional functionalimprovement in the peri-infarct region as detected by radialstrain measurements. This method allows for measurement ofmyocardial contractile function in discrete regions of themyocardium with less influence from translational motioncompared to direct visual assessment [15,28]. Radial strainanalysis is most accurate when the ultrasound beam is at anarrow angle with the radii of the measurement; such is the casein this study [29]. One pitfall, however, is that the discrete strainin the subepicardium may be overestimated due to referencepoints falling 1–2 mm outside the myocardium given thelimitation in the minimal distance between the points used tomeasure radial strain (3 mm in our case). However, we believethat this will induce little error in our measurements as thepericardium represents a fixed point with regard to radialmotion. Furthermore, we found a significantly decreasedsubepicardial strain during circumflex occlusion, indicatingthat we are able to measure relative changes in radial strain inthis region. Using measurement of radial strain we found thatthe contractile function is preserved in the subepicardial region(peri-infarct zone) in the distribution of the infarct-related artery

with darbepoetin, but not with placebo (Fig. 5). It should benoted that the observed overall beneficial changes noted withdarbepoetin were modest and the improved regional function asdetected by radial strain did not reflect in a global LV functionimprovement.

In conclusion, our data demonstrate that administration of alarge single dose of intravenous darbepoetin at the time ofreperfusion in a clinically relevant model of MI favorablyaffects the peri-infarct remodeling despite a lack of reduction inthe infarct size. The observed effects, both at the level ofmorphology and functional improvement are modest comparedto previous data in rodents, an important observation whenaiming at translating this therapy to the bedside.

Acknowledgments

We would like to thank David Fischer for his excellentassistance with the animal procedures and to Kimberly Fuhrerand Lisa Chedwick for their expert assistance with histology.This work was supported by a grant from Amgen Inc., Thou-sand Oaks, CA.

References

[1] Beltrami CA, Finato N, Rocco M, Feruglio GA, Puricelli C, Cigola E, et al.Structural basis of end-stage failure in ischemic cardiomyopathy inhumans. Circulation 1994;89(1):151–63.

[2] Zimmerman SD, Criscione J, Covell JW. Remodelling in the myocardiumadjacent to an infarction in a pig left ventricle. Am J Physiol 2004;287:H2697–704.

[3] Kloner RA, Rezkalla SH. Cardiac protection during acute myocardialinfarction: where do we stand in 2004? J Am Coll Cardiol 2004;44(2):276–86.

[4] Maiese K, Li F, Chong ZZ. New avenues of exploration for erythropoietin.JAMA 2005;293(1):90–5.

[5] Calvillo L, Latini R, Kajstura J, Leri A, Anversa P, Ghezzi P, et al.Recombinant human erythropoietin protects the myocardium fromischemia–reperfusion injury and promotes beneficial remodeling. ProcNatl Acad Sci U S A 2003;100(8):4802–6.

[6] Parsa CJ, Matsumoto A, Kim J, Riel RU, Pascal LS, Walton GB, et al. Anovel protective effect of erythropoietin in the infarcted heart. J Clin Invest2003;112(7):999–1007.

[7] Miki T, Miura T, Tanno M, Nishihara M, Naitoh K, Sato T, et al.Impairment of cardioprotective PI3K-Akt signaling by post-infarctventricular remodeling is compensated by an ERK-mediated pathway.Basic Res Cardiol 2007;120:163–70.

[8] Bahlmann FH, DeGroot K, Duckert T, Niemczyk E, Bahlmann E, BoehmSM, et al. Endothelial progenitor cell proliferation and differentiation isregulated by erythropoietin. Kidney Int 2003;64(5):1648–52.

[9] Rui T, Feng Q, Lei M, Peng T, Zhang J, Xu M, et al. Erythropoietin preventsthe acute myocardial inflammatory response induced by ischemia/reperfu-sion via induction of AP-1. Cardiovasc Res 2005;65:719–27.

[10] Moon C, Krawczyk M, Ahn D, Ahmet I, Paik D, Lakatta EG, et al.Erythropoietin reduces myocardial infarction and left ventricular func-tional decline after coronary artery ligation in rats. PNAS 2003;100(20):11612–7.

[11] Nishiya D, Omura T, Shimada K, Matsumoto R, Kusuyama T, Enomoto S,et al. Effects of erythropoietin on cardiac remodeling after myocardialinfarction. J Pharmacol Sci 2006;101(1):31–9.

[12] Gao E, Boucher M, Chuprun JK, Zhou R, Eckhart A, Koch WJ.Darbepoetin alpha, a long-acting erythropoietin analogue, offers novel anddelayed cardioprotection for the ischemic heart. Am J Physiol Heart CircPhysiol (in press).

136 C. Toma et al. / Journal of Molecular and Cellular Cardiology 43 (2007) 130–136

[13] Kritensen J, Maeng M, Rehling M, Berg JS, Mortensen UM, Nielsen SS,et al. Lack of acute cardioprotective effect from preischemic erythropoietinadministration in a porcine coronary occlusion model. Clin Physiol FunctImaging 2005;25:305–10.

[14] Olea FD, Vera Janavel G, De Lorenzi A, Cuniberti L, Yannarelli G, CabezaMeckert P, et al. High-dose erythropoietin has no long-term protectiveeffects in sheep with reperfused myocardial infarction. J CardiovascPharmacol 2006;47:736–41.

[15] Dohi K, PinskyMR, Kanzaki H, Severyn D, Gorcsan III J. Effects of radialleft ventricular dyssynchrony on cardiac performance using quantitativetissue Doppler radial strain imaging. J Am Soc Echocardiogr 2006;19(5):475–82.

[16] Matre K, Fannelop T, Dahle GO, Heimdal A, Grong K. Radial strain gradientacross the normal myocardial wall in open-chest pigs measured with Dopplerstrain rate imaging. J Am Soc Echocardiogr 2005;18(10):1066–73.

[17] Hirata A,Minamino T, AsanumaH, FujitaM,WakenoM,MyoishiM, et al.Erythropoietin enhances neovascularization of ischemic myocardium andimprovesleftventriculardysfunctionaftermyocardialinfarctionindogs.JAmCollCardiol 2006;48(1):176–84.

[18] Lipsic E, van der Meer P, Voors AA, Westenbrink BD, van den Heuvel AF,de Boer HC, et al. A single bolus of a long-acting erythropoietin analoguedarbepoetin alfa in patients with acute myocardial infarction: a randomizedfeasibility and safety study. Cardiovasc Drugs Ther 2006;20(2):135–41.

[19] Krause KT, Jaquet K, Geidel S, Schneider C, Mandel C, Stoll HP, et al.Percutaneous endocardial injection of erythropoietin: assessment ofcardioprotection by electromechanical mapping. Eur J Heart Fail 2006;8(5):443–50.

[20] Li L, Takemura G, Li Y, Miyata S, Esaki M, Okada H, et al. Preventiveeffect of erythropoietin on cardiac dysfunction in doxorubicin-inducedcardiomyopathy. Circulation 2006;113(4):535–43.

[21] Parsa CJ, Kim J, Riel RU, Pascal LS, Thompson RB, Petrofski JA, et al.Cardioprotective effects of erythropoietin in the reperfused ischemic heart: apotential role for cardiac fibroblasts. J Biol Chem 2004;279(20):20655–62.

[22] Jaquet K, Krause K, Tawakol-Khodai M, Geidel S, Kuck KH.Erythropoietin and VEGF exhibit equal angiogenic potential. MicrovascRes 2002;64(2):326–33.

[23] Chong ZZ, Kang JQ, Maiese K. Erythropoietin is a novel vascularprotectant through activation of Akt1 and mitochondrial modulation ofcysteine proteases. Circulation 2002;106(23):2973–9.

[24] Van der Meer P, Lipsic E, Henning RH, Boddeus K, van der Velden J,Voors AA, et al. Erythropoietin induces neovascularization and improvescardiac function in rats with heart failure after myocardial infarction. J AmColl Cardiol 2005 (Jul 5);46(1):125–33.

[25] Rissanen TT, Markkanen JE, Gruchala M, Heikura T, Puranen A, KettunenMI, et al. VEGF-D is the strongest angiogenic and lymphangiogeniceffector among VEGFs delivered into skeletal muscle via adenoviruses.Circ Res 2003 (May 30);92(10):1098–106.

[26] Elliott S, Busse L, Bass MB, Lu H, Sarosi I, Sinclair AM, et al. Anti-Eporeceptor antibodies do not predict Epo receptor expression. Blood 2006(Mar 1);107(5):1892–5.

[27] Depping R, Kawakami K, Ocker H, Wagner JM, Heringlake M, NoetzoldA, et al. Expression of the erythropoietin receptor in the human heart.J Thorac Cardiovasc Surg 2005;130(3):877–8.

[28] Weidemann F, Jamal F, Kowalski M, Kukulski T, D'Hooge J, Bijnens B,et al. Can strain rate and strain quantify changes in regional systolicfunction during dobutamine infusion, B-blockade, and atrial pacing-implications for quantitative stress echocardiography. J Am Soc Echocar-diogr 2002;15(5):416–24.

[29] Marwick TH. Measurement of strain and strain rate by echocardiography:ready for prime time? J Am Coll Cardiol 2006;47(7):1313–27.