Author + information
- Received December 2, 2003
- Revision received January 15, 2004
- Accepted January 27, 2004
- Published online June 2, 2004.
- Gian G Corbucci, MD*,
- Cinzia Perrino, MD†,
- Giuseppe Donato, MD‡,
- Antonio Ricchi, MD§,
- Biagio Lettieri, MD∥,
- Giancarlo Troncone, MD¶,
- Ciro Indolfi, MD#,
- Massimo Chiariello, MD† and
- Enrico V Avvedimento, MD¶,* ()
- ↵*Reprint requests and correspondence:
Dr. Enrico V. Avvedimento, Dipartimento di Biologia e Patologia Cellulare e Molecolare “L. Califano”, University “Federico II,” Naples 80131, Italy.
Objectives We sought to describe the sequence of molecular events during ischemia and reperfusion of the human heart and to determine the activation of stress kinases and deoxyribonucleic acid (DNA) damage response elements on apoptosis in ischemia or reperfusion of the human heart.
Background Brief ischemia is tolerated by cardiac myocytes, but it determines immediate metabolic changes and block of contraction. Prompt restoration of coronary blood flow is inexorably associated with a slow recovery of myocardial contractile function. The prolonged, postischemic contractile dysfunction in the viable tissue is called myocardial stunning. The molecular mechanisms underlying myocardial stunning and ischemia-reperfusion injury are still poorly understood. Their elucidation would be valuable in order to identify novel therapeutic strategies.
Methods We examined human left ventricular samples taken from 20 patients undergoing elective valve surgery before aortic cross-clamping, 20 ± 2 min (brief ischemia), 58 ± 5 min after the cross-clamping period (prolonged ischemia), and 21 ± 4 min after reconstitution of coronary blood flow (reperfusion). Stress kinases and DNA damage sensor proteins (ATM, H2AX, p53) were determined by immunoblotting with specific antibodies. Electron microscopy analysis was carried out on ischemic and reperfused samples. ATP content, reactive oxygen species (ROS) levels, and cytochrome oxidase activity were determined by biochemical assays.
Results Ischemia caused accumulation of ROS, reduction of cytochrome C oxidase and ATP, and activation of stress kinases p38 and Jun terminal kinase. Electron microscopy showed significant mitochondrial swelling in the majority of cells, but no appreciable apoptosis of cardiomyocytes. During ischemia, myocytes were intensely stained by TUNEL, and many cells showed proliferative cell nuclear antigen-positive nuclei. Finally, we found in ischemic tissues increased p53/p21WAFlevels and phosphorylation of histone H2AX, a substrate of ATM kinase, which marks double-strand DNA breaks. Reperfusion caused a robust extracellular signal-regulated kinase-1/2 activation, a marked reduction of TUNEL staining, and persistent activation of ATM checkpoint.
Conclusions These data indicate that ischemia induces extensive DNA damage and activation of ATM checkpoint. Reperfusion allows the repair of the DNA lesions and salvage of ischemic cells.
Brief ischemia is tolerated by cardiac myocytes, but it determines immediate metabolic changes and block of contraction. Prompt restoration of coronary blood flow is inexorably associated with a slow recovery of myocardial contractile function. The prolonged, post-ischemic contractile dysfunction in the viable tissue is called myocardial stunning (1). The molecular mechanisms underlying myocardial stunning and ischemia-reperfusion injury are still poorly understood. Their elucidation would be valuable in order to identify novel therapeutic strategies.
It has been shown in animal models that production of oxygen free radicals (reactive oxygen species [ROS]) soon after restoration of blood flow is responsible for the reperfusion injury (2–4). Nevertheless, the clinical use of antioxidants during the reperfusion therapy of human myocardial infarction did not prevent the stunning phenomenon (5). These data suggest that other mechanisms, possibly acting during the ischemic period, could be involved in the pathogenesis of ischemia and reperfusion injury. Apoptosis has been described in a variety of human cardiovascular diseases, including myocardial infarction (6)and heart failure (7,8). In vivo animal models of myocardial ischemia and reperfusion have shown apoptotic cell death in severely and mildly ischemic regions (9)and after reperfusion (10). However, there is considerable controversy about the extent of cardiac myocytes, which undergo apoptosis in various cardiac diseases, and many investigators now believe that very high numbers may be artifactual as they would not be consistent with survival of the organ.
Cardiopulmonary bypass during open-heart surgery is a controlled human model of myocardial ischemia and reperfusion caused by aortic cross-clamping. Patients undergoing cardiopulmonary bypass often postoperatively develop heart failure due to myocardial stunning, and they require inotropic support.
We analyzed the activation of stress kinases signaling cascade and of markers of DNA damage in left ventricles during cardiopulmonary bypass. We found that acute ischemia robustly activated stress kinases. The majority of myocytes became TUNEL-positive and accumulated deoxyribonucleic acid (DNA) lesions, as shown by positive proliferative cell nuclear antigen (PCNA) staining and activation of ATM, the enzyme that senses DNA damage. Reperfusion was associated with reduced TUNEL staining, although ROS accumulation was stimulated.
Patient selection and surgical protocol
The study was performed on left ventricular human samples obtained from 20 consecutive patients during elective valvular surgery (aortic and/or mitralic) at the Department of Cardiothoracic Surgery at “Brotzu” Hospital in Cagliari (Italy). Biopsies were obtained from 12 men and eight women (mean age, 56 ± 6 years) (Table 1). The protocol was approved by the Institutional Ethics Committee (University of Cagliari), and informed consent was obtained from all patients. Patients with left ventricular ejection fraction <40%, those requiring preoperative inotropic or intraaortic balloon pump support, those with uncontrolled systemic diseases (diabetes, hypertension, or renal failure), those receiving nitrate medications or statins were excluded. Routine anesthetic and surgical protocols were used in all patients enrolled. The same surgeon performed all operations. The surgical outcome was uneventful in all patients, and there were no major complications (myocardial infarction, death). The hearts were maintained in pO2= 140 mm Hg for the entire period of ischemia. All patients received blood cardioplegic solution plus KCl. Briefly, blood cardioplegia was administered to all patients immediately after cross-clamping of the aorta. Cardioplegia was given first, antegrade, through the aortic root, and then retrogradethrough the coronary sinus with boluses of cold cardioplegia every 20 min. Patients were maintained in mild hypothermia (32°C rectal temperature); before removal of the cross-clamp, an additional warm cardioplegia was administered, followed by warm blood.
Left ventricular biopsies
Biopsies (10 to 20 mg) were taken at four time points during the operation: 1) before aortic cross-clamping (control); 2) 20 ± 2 min after the cross-clamping period (20'); 3) 58 ± 5 min after the cross-clamping period (58'); and 4) after 21 ± 4 min from reconstitution of coronary blood flow (reperfusion). Samples were snap-frozen in liquid nitrogen and stored at −80°C before use.
Immunoblotting and antibodies
Fourteen different groups of ventricular specimens (Table 1) were lyzed in ice-cold RIPA buffer (20 mM Tris-HCl pH 7.4, 0.15 M NaCl, 10 mM EDTA, 0.25% Triton X-100, 0.05% Tween-20, 0.1% SDS) containing aprotinin (5 μg/ml), leupeptin (10 μg/ml), pepstatin (2 μg/ml), and 0.5 mM phenylmethylsulfonyl fluoride (PMSF). Proteins (40 to 60 μg) were separated on SDS-polyacrilamide gels and transferred to nitrocellulose membranes. Commercially available primary antibodies anti-p38 (rabbit, Santa Cruz, sc-535, Santa Cruz, California) and phospho-p38 (rabbit, Cell Signaling, Beverly, Massachusetts 9211), anti-extracellular signal-regulated kinase (ERK)-1/2 (rabbit, Santa Cruz, sc-154) and phospho-ERK1/2 (mouse, Santa Cruz, sc-7383), anti-Jun terminal kinase (JNK) (rabbit, Upstate Biotechnology, Waltham, Massachusetts) and p-JNK (rabbit, Upstate Biotechnology), anti-p53 (mouse, Santa Cruz), anti-PCNA (mouse, Dako, clone PC 10) and anti-p21WAF(goat, Santa Cruz), anti H2Ax (rabbit, Upstate Biotechnology) were used. Immunoblotting was performed as previously described (11), and bands were quantified by densitometry. Each experiment and its densitometric analysis were independently repeated at least 3 ×.
Three groups of samples were fixed with 10% buffered formalin for conventional histology, TUNEL examination, and PCNA staining (Table 1). The DNA fragments were determined by the in situ apoptosis detection kit POD (Roche, Indianapolis, Indiana) according to manufacturer's instructions.
Ultrastructural analysis of human left ventricular specimens was performed after fixing three groups of samples in 3% glutaraldehyde/0.1 M phosphate buffer (Table 1). Samples were processed (post-fixation and dehydration) for embedding in Epon-Araldite (West Chester, Pennsylvania). Ultra-thin sections, stained with uranyl acetate and lead citrate, were studied by using a Zeiss (Aresc, Milan, Italy) EM 900 electron microscope.
The intracellular ATP of cardiac tissues from 10 different patients (Table 1) was measured with the ATP bioluminescent assay kit (Sigma, St. Louis, Missouri). ATP content was calculated by comparison with a standard curve derived from known concentrations of ATP, ranging from 0.01 to 10 pmol and was expressed in μmol/g protein.
Activity of mitochondrial enzymes
Ten different groups of tissue samples (Table 1) were homogenized at 4°C in medium A (0.3 M sucrose, 10 mM KH2PO4, 1 mg/ml bovine serum albumin, pH 6.50). Mitochondria were immediately isolated, and mitochondrial proteins were measured. Mitochondrial enzyme assays were performed in a DWS (Beckman DU 640, Beckman, Fullerton, California). All assays were performed at 37°C in 1 ml of medium A. We calculated the cytochrome c oxidase (COX) activity from the initial pseudolinear rate of cytochrome c oxidation at 550 nm – 580 nm. All experiments were performed using different amounts of mitochondria (10, 5, 2, and 1 mg of mitochondrial protein). For each experiment, a series of five assays from the same sample was performed. The enzymatic activity was expressed in μmol/min/g protein.
Data are expressed as mean ± SE. For all analyses, a minimum value of p < 0.05 was considered statistically significant; when present, a p value <0.01 was specified. When possible, for the comparison of results over time in the same patient, we used repeated measures analysis of variance (ANOVA). When, for technical issues, results were not available at all time points in all patients, multigroup comparisons were calculated using ANOVA with Bonferroni correction.
Differential activation of MAPKs (ERK1/2, p38, and JNK) during acute ischemia and reperfusion of human left ventricles
Ischemia and reperfusion of the heart activate apoptosis in animal models (6,9,10). Upstream of cell death promoters and executioners, there are several protein kinases that, once activated by phosphorylation, trigger the phosphorylation of transcription factors and other signaling proteins involved in cell survival (13–18). We examined the activation of ERK, JNK, and p38 at early stages of ischemia and reperfusion of the human heart. Total proteins derived from the ventricular samples of patients (Table 1) undergoing aortic clamping were immunoblotted with specific antibodies that recognize the phosphorylated forms of ERK1/2, JNK, and p38. Figure 1shows that ischemia differentially activates the three stress protein kinases indicated above: 1) ERK1/2 was not induced by 20' and 58' ischemia; 2) JNK was activated by ischemia in a time-dependent fashion (3.12 ± 0.22-fold at 58' vs. CON); 3) p38 was transiently induced at 20' of ischemia and at 58' returned to the baseline (2.14 ± 0.21-fold in 20' vs. CON). Reperfusion, on the other hand, induced a robust and significant stimulation of all three kinases, including ERK1/2, which was not induced by ischemia. The activation, shown by the accumulation of phosphorylated bands, was specific and selective because the total mass of protein of each kinase analyzed was not modified by ischemia or reperfusion (Fig. 1, lower panels).
The data shown in Figure 1, although descriptive, discriminated ischemia from reperfusion because: 1) ERK1 was specifically induced by reperfusion; 2) JNK was induced in a time-dependent fashion; 3) p38 was activated by reperfusion, but only transiently by ischemia.
Mitochondrial stress induced by ischemia-reperfusion
Mitochondria are the seat of important cellular functions including essential pathways of intermediate metabolism, amino acid biosynthesis, fatty acid oxidation, steroid metabolism, apoptosis, and, most importantly, oxidative energy metabolism. To determine more precisely the site and the degree of the stress induced by ischemia in cardiomyocytes, we analyzed the structure and the biochemical properties of mitochondria derived from ischemic and reperfused samples. Figure 2Ashows EM pictures of cells derived from ischemic and reperfused samples. The mitochondria appeared markedly enlarged in all the samples, and their structural organization (cristae) was significantly altered. We noticed, however, that swelling of internal membranes was more severe at 20 min of ischemia relative to control and 58 min or reperfused samples. Also, mitochondria from reperfused sample, although enlarged, showed a certain degree of structural organization (Fig. 2A, reperfusion and data not shown). To obtain a biochemical and functional analysis of mitochondrial metabolism under ischemia and reperfusion, we determined the activity of COX, ATP, and ROS content in our samples.
Figure 2Bshows the enzymatic activity of COX, ATP, and ROS content of 10 ischemic and reperfused samples. Cytochrome C oxidase activity and ATP levels were depressed in ischemic samples and remained lower than control levels after the 20-min reperfusion. Reperfusion induced a significant, albeit partial, ATP recovery relative to ischemic samples (Fig. 2B). Oxygen free radical content, analyzed by malondialdehyde, was increased during ischemia and further increased during reperfusion (Fig. 2B).
Transient TUNEL staining of ischemic and reperfused ventricular myocytes
Analysis by EM of the samples indicated that only a minority of cells underwent apoptosis. There was a consistent fraction of the cells (approximately 50% to 60%) that contained swollen mitochondria, glycogen deposits with intact nuclei, and membranes, but we were unable to find cardiac cells with signatures of apoptosis (data not shown). Staining of the sections of ventricular tissue with TUNEL, a widely used assay that labels with terminal transferase nicks and breaks in the DNA molecules, showed a significant fraction of positive cells during ischemia and reperfusion stained by TUNEL (Fig. 3). In several control samples analyzed (time 0), we never detected more than 1% to 2% of positive cells. The fraction of the stained cells significantly increased with the time (30 ± 5% at 20'; 46 ± 4% after 58' of ischemia). Reperfusion was associated with a significant reduction of TUNEL-positive cells (20 ± 3%, p < 0.01 vs. 20' and 58' of ischemia). Comparative analysis of the samples with hematoxylin and eosin (HE) indicated that the majority of stained nuclei were of cardiomyocytes, and they were not apoptotic or necrotic (Fig. 3and insets).
Trigger of DNA damage checkpoints in ischemic and reperfused myocytes
The massive TUNEL staining of the ventricular myocytes induced by ischemia contrasted with the absence of morphologic alterations of the cells. Massive apoptosis would not be consistent with the survival of the organ and would result in a significant organ failure not found in the patients analyzed (Methods section, Patient selection). To solve this paradox, we first determined caspase 3 in our sample by immunoblot. Figure 4Ashows that there is no sign of caspase activation in ischemic or reperfused samples. We also analyzed, in the same ventricular samples, several molecular markers triggered by damaged DNA. DNA damage is sensed and signaled by p53 and ATM kinase. p53 induces the cdk inhibitor p21WAF; ATM, on the other hand, phosphorylates a residue in the histone H2AX, contiguous to the DNA double-strand breaks (19). Figure 4Bshows a significant activation of p53 and p21WAFduring ischemia and reperfusion. Phosphorylation of histone H2Ax was also stimulated, indicating a significant activation of ATM kinase, in ischemia and reperfusion (Fig. 4B). Under both conditions, the major DNA damage checkpoints (p53 and ATM) were activated.
To obtain an independent marker of DNA repair, we stained ischemic and reperfused samples with PCNA antibodies that recognize the proliferating nuclear antigen involved in DNA repair. Figure 4Cshows PCNA-positive nuclei in cells at 58 min of ischemia (ca. 10%). The signal was not present in control samples (less than 1%) and was not significantly reduced in the reperfused heart (ca. 8%).
These data strongly suggest that stressed cardiomyocytes accumulate DNA lesions and activate the sensor apparatus, which signals the damage to the repairing machinery. TUNEL-positive cells represent myocytes with damaged DNA, not apoptotic cells.
We established a close correlation between acute ischemia in the heart and activation of DNA damage checkpoint. To our knowledge, there have been no reports to date of DNA damage and activation of ATM in acute ischemia. We report here that ischemia results in a massive and transient induction of DNA damage in ventricular cardiomyocytes. Electron microscopy indicates that the primary targets of ischemia-reperfusion injury are the mitochondria, because they are altered (swelling) in the vast majority of the cells at earlier times during ischemia (Fig. 2). We believe that the activation of stress kinases is secondary to the mitochondrial unbalance of ROS (11). This idea is supported by the increase of ROS in ischemic and reperfused samples and by the marked reduction of COX and ATP content in ischemic samples (Fig. 2B).
Stress response in ischemic cardiomyocytes
During brief and prolonged ischemia, ERK1/2 was not activated, while p38 was rapidly activated compared with JNK. Under very similar conditions, ERK1/2 was found active in ischemic atrial appendages (18). These differences might reflect a different response between the ventricular versus atrial myocytes or a different degree of stress imposed on the hearts before the analysis. It has been reported that these kinases are differentially responsive to different types of stress: JNK is sensitive to misfolded proteins accumulation (20), and p38 is activated by osmotic stress (21). Extracellular signal-regulated kinase-1/2 on the other hand seems to be activated by a variety of stimuli. In isolated cardiomyocytes and endothelial cells, sustained or higher H2O2concentrations are able to activate all three types of stress kinases, while lower H2O2concentrations transiently activate ERK1/2 and JNK. Moreover, we found that ERK1/2 activated by Ki Ras stimulates mitochondrial MnSOD (11). The robust accumulation of ROS in reperfused samples might be responsible for ERK1/2 activation (Figs. 1 and 2). Extracellular signal-regulated kinase-1/2 activation seems to have protective effects in reperfusion injury (22). On the other hand, p38 mediates cellular damage in ischemia-reperfusion (23). Extracellular signal-regulated kinase-1/2 activation under these conditions might be an important factor to reduce mitochondrial stress (24).
Apoptosis or DNA damage and activation of repair machinery?
Ischemia and reperfusion induce a considerable stress on mitochondria, as demonstrated by reduction of COX, ATP content, and mitochondrial swelling (Fig. 2B). Traditionally, apoptosis has been viewed as a necessary outcome of these events. Apoptosis of cardiomyocytes has been reported in various pathologic settings (6–10). Most reports are mainly based on DNA fragmentation, a biochemical hallmark of apoptosis (25), detected from the DNA ladder pattern on gel electrophoresis and/or DNA in situ nick end-labeling (TUNEL). However, evidence of the characteristic ultrastructure, a morphologic hallmark of apoptosis (26), is rare in cardiomyocytes.
The best interpretation of the data shown here is that the majority of TUNEL-positive cells are myocytes with damaged DNA, which activates stress kinases and checkpoint proteins, such as p53 and ATM. Note the rapid accumulation of p53 in Figure 4A, is most likely caused by impaired degradation of the protein. Although the TUNEL assay is not very specific, we believe that the repair machine is at work in these samples, because we find cells, during ischemia and reperfusion, positive to PCNA (Fig. 4C). The type of DNA lesions are, most likely, double-strand breaks, because ATM and H2AX phosphorylation are specifically activated by these lesions (19). The fraction of TUNEL-positive cells (insets in Fig. 3) was significantly lower in reperfused samples compared with 20 and 58 ischemia samples. Taken together, these data indicate a novel aspect of the stress induced by reperfusion. We suggest that DNA damage is initiated during ischemia (Figs. 2 and 3) and activates the DNA repair machinery (Fig. 4). Reperfusion, on the other hand, by restoring cellular respiration, determines partial recovery of ATP content and high ROS levels. However, in the same samples, reperfusion was associated with a reduction of TUNEL. We propose that, under these conditions, the repair apparatus is already completely activated and more capable of coping with oxidative stress. Taken together, these observations might explain why restoration of coronary blood flow does not result in complete restoration of cardiac function and also why repetitive ischemic episodes are better tolerated by cardiac myocytes (preconditioning). It is possible that the lesions, initiated at high levels during ischemia, are rescued by repair pathways. For example, ERK1/2 and PKC epsilon form mitochondrial signaling modules that mediate protection of cardiac cells by phosphorylation and inactivation of BAD (24).
The nature of the samples shown in our study precludes a long-term study. We cannot determine the fate of stressed ventricular cells, days and months after the ischemia-reperfusion. We presume that the majority of the cells repaired the DNA lesions, because the patients did not show any functional alterations after surgery. It would be of interest to study the long-term consequences of repaired DNA lesions on chromatin organization and transcription.
In conclusion, the data shown here indicate that ventricular myocytes accumulate DNA damage during ischemia and reperfusion. Under these conditions, only a minor fraction of the stressed cells succumbs to apoptosis, and the vast majority of the cells repair DNA lesions. It remains to be determined if the cells that survive the acute stress undergo gross DNA lesions that compromise chromatin organization and transcriptional efficiency.
☆ This work was partly supported by AIRC (Associazione Italiana di Ricerca contro il Cancro) and MURST (Ministero dell'Università e della Ricerca Scientifica e Tecnologica) grants to Dr. Avvedimento. Drs. Corbucci and Perrino contributed equally to this work.
- cytochrome c oxidase
- deoxyribonucleic acid
- electron microscopy
- extracellular signal-regulated kinase
- Jun terminal kinase
- proliferative cell nuclear antigen
- reactive oxygen species
- Received December 2, 2003.
- Revision received January 15, 2004.
- Accepted January 27, 2004.
- American College of Cardiology Foundation
- Braunwald E,
- Kloner R.A
- Bolli R,
- Jeroudi M.O,
- Patel B.S,
- et al.
- Bolli R,
- Jeroudi M.O,
- Patel B.S,
- et al.
- Flaherty J.T,
- Pitt B,
- Gruber J.W,
- et al.
- Saraste A,
- Pulkki K,
- Kallajoki M,
- Henriksen K,
- Parvinen M,
- Voipio-Pulkki L.M
- Wong S.H,
- Knight J.A,
- Hopfer S.M,
- Zaharia O,
- Leach C.N Jr..,
- Sunderman F.W Jr.
- Lazou A,
- Bogoyevitch M.A,
- Clerk A,
- Fuller S.J,
- Marshall C.J,
- Sugden P.H
- Laderoute K.R,
- Webster K.A
- Talmor D,
- Applebaum A,
- Rudich A,
- Shapira Y,
- Tirosh A
- Burma S,
- Chen B.P,
- Murphy M,
- Kurimasa A,
- Chen D.J
- Liu Y.F
- Aggeli I.K,
- Gaitanaki C,
- Lazou A,
- Beis I
- Schulman D,
- Latchman D.S,
- Yellon D.M
- Baines C.P,
- Zhang J,
- Wang G.W,
- et al.
- Ohno M,
- Takemura G,
- Ohno A,
- et al.
- Kanoh M,
- Takemura G,
- Misao J,
- et al.