Acute systemic hypoxia induces delayed cardioprotection against ischemia (I)-reperfusion (R) injury via inducible nitric oxide synthase (iNOS)-dependent mechanism. Because CoCl2 is known to elicit hypoxia-like responses, we hypothesized that this chemical would mimic the delayed preconditioning effect in the heart. Adult male mice were pretreated with CoCl2 or saline. The hearts were isolated 24 h later and subjected to 20 min of global I and 30 min of R in Langendorff mode. Myocardial infarct size (% of risk area; mean ± SE, n = 6–8/group) was reduced in mice pretreated with 30 mg/kg CoCl2 (16.1 ± 3.1% vs. 27.6 ± 3.3% with saline control; P < 0.05) without compromising postischemic cardiac function. Higher doses of CoCl2 failed to induce similar protection. Electrophoretic mobility gel shift assay demonstrated significant enhancement in DNA binding activity of hypoxia-inducible factor 1α (HIF-1α) and activator protein 1 (AP-1) in nuclear extracts from CoCl2-treated hearts. Activation of HIF-1α and AP-1 was evident at 30 min and sustained for the next 4 h after CoCl2 injection. In contrast, CoCl2-induced protection was independent of NF-κB activation because no DNA binding or p65 translocation was observed in nuclear extracts. Also, CoCl2-induced cardioprotection was preserved in p50 subunit NF-κB-knockout (KO) mice (11.1 ± 3.0% vs. 25.1 ± 5.0% in saline-treated p50-KO mice; P < 0.05). The infarct-limiting effect of CoCl2 was absent in iNOS-KO mice (20.9 ± 3.0%). We conclude that in vivo administration of CoCl2 preconditions the heart against I/R injury. The delayed protective effect of CoCl2 is achieved through a distinctive signaling mechanism involving HIF-1α, AP-1, and iNOS but independent of NF-κB activation.
- ischemia-reperfusion injury
- myocardial infarction
- nitric oxide synthase
- transcription factors
incessant or intermittent chronic systemic hypoxia can elicit a complex adaptive response that leads to enhanced tolerance to myocardial ischemia-reperfusion (I/R) injury in immature or adult mammals (2, 3, 20, 28). The myocardial endogenous adaptation to hypoxia leading to resistance against subsequent lethal injuries appears to be an inherent part of the so-called “late phase of preconditioning” (5), which may have great promise for eventual clinical applications. Most recently, we reported a similar protective effect by acute exposure to systemic hypoxia in adult mice (36), where the adverse effects of chronic hypoxia, such as pulmonary hypertension and right ventricular hypertrophy, can be effectively avoided. In these studies, inducible nitric oxide synthase (iNOS) played a pivotal role in triggering as well as mediating the hypoxia-induced late cardioprotection (36). Cai et al. (7) recently confirmed these results and reported the involvement of hypoxia-inducible factor 1α (HIF-1α) and erythropoietin (Epo) in the acute systemic hypoxia-induced cardioprotection. Three HIF-1α prolyl-4-hydroxylases (PHDs) (named PHD1, PHD2, and PHD3) effect the proteasome-mediated degradation of HIF by catalyzing the hydroxylation of key proline residues in the HIF-1α subunit under normoxic conditions. When oxygen tension is reduced, PHD-mediated hydroxylation cannot occur and HIF-1α accumulates in the nucleus, resulting in HIF-mediated gene transcription. Because cobalt is the best-known chemical inducer of hypoxia-like responses such as erythropoiesis and angiogenesis in vivo, we further hypothesized that hypoxia-induced protection can be chemically mimicked with CoCl2. Specifically, we were interested in establishing the role of CoCl2 in inducing delayed preconditioning-like effects against I/R injury. The second goal was to identify the underlying signaling mechanisms in CoCl2-induced cardioprotection. In particular, we sought to determine whether CoCl2 activates the key transcription factors including HIF-1α, NF-κB, and activator protein 1 (AP-1) and also utilizes iNOS in inducing delayed preconditioning in the heart. We chose to study these transcription factors because of their potential link with hypoxia-induced cardioprotection (7) and their well-described essential role in cardiac preconditioning (31, 37–39). The preliminary results of this study were presented at the American Heart Association Scientific Sessions 2003 (Orlando, FL) and published in abstract form (35).
MATERIALS AND METHODS
Adult male outbred ICR mice were supplied by Harlan (Indianapolis, IN). Adult male homozygous (−/−) iNOS-knockout (KO) and p50-KO mice as well as B6,129 wild-type (WT) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). All animal experiments were conducted under the guidelines on humane use and care of laboratory animals for biomedical research published by the National Institutes of Health (Pub. No. 85–23, revised 1996).
Chemicals and drugs.
CoCl2, diethyldithiocarbamic acid (DDTC), and LPS as well as all the chemicals for preparing the Krebs-Henseleit perfusion buffer were purchased from Sigma-Aldrich (St. Louis, MO).
Mouse model of global I/R injury.
The methodology of Langendorff-isolated, buffer-perfused mouse heart preparation was previously described in detail (33, 34, 36). This ex vivo model enabled us to measure ventricular contractile function as well as infarct size after 20 min of no-flow global ischemia (I) and 30 min of reperfusion (R).
Experimental groups and protocols.
Mice were randomized into the following 10 experimental groups: saline group (n = 6), ICR mice pretreated with 0.15 ml of 0.9% saline (ip) 24 h before I-R; three CoCl2 groups (n = 6/group), ICR mice pretreated with CoCl2 (dissolved in 0.9% saline; ip) at a dose of 30, 60, or 120 mg/kg, respectively, 24 h before I/R; saline + B6,129 group (n = 8), B6,129-WT mice pretreated with 0.15 ml of saline (ip) 24 h before I/R; CoCl2 + B6,129 group (n = 6), B6,129-WT mice pretreated with CoCl2 (30 mg/kg ip) 24 h before I/R; DDTC + CoCl2 group (n = 6), B6,129-WT mice given DDTC (150 mg/kg ip) 30 min before CoCl2 treatment (30 mg/kg ip) 24 h before I/R; saline + p50-KO group (n = 7), p50-KO mice pretreated with 0.15 ml of saline (ip) 24 h before I/R; CoCl2 + p50-KO group (n = 6), p50-KO mice pretreated with CoCl2 (30 mg/kg ip) 24 h before I/R; CoCl2 + iNOS-KO group (n = 7), iNOS-KO mice pretreated with CoCl2 (30 mg/kg ip) 24 h before I/R. In addition, we used our previously reported data (34), which were collected under similar experimental conditions and protocol in iNOS-KO mice, to serve as the control for the CoCl2 + iNOS-KO group.
DNA binding of transcription factors.
In a parallel series of experiments, ventricular myocardium samples (n = 2 or 3 per group) were collected from ICR mice after 30 min and 1, 2, 4, and 6 h of treatment with saline (0.15 ml), CoCl2 (30 mg/kg), LPS (5 mg/kg), or systemic hypoxia (10% inspired O2 fraction; see Ref. 36 for further methodological details). Tissue samples (∼200 mg each) were ground in liquid nitrogen and homogenized in 1 ml of ice-cold lysis buffer containing (in mM) 20 Tris (pH 7.4), 140 NaCl, 1.5 MgCl2, 1 EGTA, 1 EDTA, 1 DTT, 0.5% NP-40, 0.5 Na3VO4, and a cocktail of protease inhibitors (aprotinin, leupeptin, PMSF). The nuclei were separated by centrifugation, washed once with 1 ml of lysis buffer lacking NP-40, and resuspended in 50 μl of nuclear extraction buffer containing (in mM) 50 Tris·HCl (pH 7.8), 60 KCl, 1 EDTA, 1 EGTA, 2 DTT, 1 PMSF, and 0.5 Na3VO4. After three cycles of freezing and thawing, the nuclear extracts were obtained by centrifugation at 10,000 g for 15 min and then used in the DNA binding assay. The gel shift oligonucleotide of NF-κB (5′-AGTTGAGGGGACTTTCCCAGGC-3′; catalog no. sc-2505, Santa Cruz), HIF-1α (5′-TCTGTACGTCACCACACTCACCTC-3′; catalog no. sc-2625, Santa Cruz), and AP-1 (5′-CGCTTGATGACTCAGCCGGAA-3′; catalog no. sc-2501, Santa Cruz) was labeled with T4 polynucleotide kinase and [γ-32P]ATP. Typically, the binding reaction mixture consisted of 5 μg of protein, 5% glycerol, 1 μg of poly(dI-dC), and 0.1 ng of 32P-labeled NF-κB. The reaction mixture was incubated for 30 min at 30°C. The specific protein-DNA complexes were then separated on 5% PAGE in 0.5× Tris-borate-EDTA buffer. The gels were dried in a glycerol-ethanol mixture and exposed to X-ray film. For quantifying the DNA binding activity, the optical density for each EMSA band was scanned and analyzed with a densitometric system (Bioquant 98).
Nuclear translocation of p65 subunit of NF-κB.
The ventricular samples (n = 2–4/group) were collected from ICR mice 30 min after pretreatment with saline (0.15 ml ip), CoCl2 (30 mg/kg ip), or LPS (5 mg/kg ip). Tissue samples were ground in liquid nitrogen and homogenized in 1 ml of ice-cold buffer containing (in mM) 250 sucrose, 10 Tris·HCl (pH 7.4), 1 EDTA, 1 Na3VO4, and 1 NaF, with protease inhibitor cocktail (Sigma-Aldrich). The homogenate was centrifuged at 1,000 g for 10 min at 4°C, and the supernatant was collected and recentrifuged at 10,000 g for 30 min and recovered as the cytosolic protein. The nuclear pellets were incubated in the lysis buffer containing (in mM) 50 Tris·HCl (pH 7.4), 150 NaCl, 1 EDTA, 1% NP-40, and 1 Na3VO4, with protease inhibitor cocktail, to obtain the soluble protein from nuclear fraction. Fifty micrograms of total protein from either the cytosolic or the nuclear fraction of each sample were separated by SDS-PAGE on 10% acrylamide gels, transferred to a nitrocellulose membrane, and then blocked with 5% non-fat dry milk in Tris-buffered saline + Tween 20. The membrane was subsequently incubated with a rabbit polyclonal antibody (1:200 dilution; catalog no. sc-372, Santa Cruz) specific to the p65 subunit of NF-κB. The secondary antibody was a horseradish peroxidase-conjugated anti-rabbit IgG (1:500 dilution; Amersham). The membranes were developed with enhanced chemiluminescence and exposed to X-ray film.
Data analysis and statistics.
Data are presented as group means ± SE. The difference among experimental groups was compared by unpaired t-test or one-way ANOVA followed by Student-Newman-Keuls post hoc test. P < 0.05 was considered as statistically significant.
Exclusions and baseline cardiac function.
Administration of a bolus dose of CoCl2 up to 120 mg/kg caused no mortality in the mice during the 24-h period. The animals' behavior also appeared normal except for the occurrence of hyperpnea similar to the hypoxia response. A total of 64 mouse hearts were subjected to the I/R protocol for the assessment of ventricular function as well as infarct size. Four hearts (i.e., 5.9% of the 68 perfused hearts) were excluded for one of the following reasons: 1) time delay (>3 min) or aortic damage during aortic cannulation or 2) sustained arrhythmia during the stabilization period. The preischemia baseline levels of ventricular contractile function (measured either by developed force or rate-force product) and coronary flow (normalized against the individual heart wet weight, i.e., ml·min−1·g−1) were not different among all 10 experimental groups (P > 0.05; Table 1). In contrast, the heart rate was lower in 30 mg/kg CoCl2-treated B6,129 and iNOS-KO mice compared with ICR mice treated with higher doses of CoCl2 (60 or 120 mg/kg). Such differences in heart rate may reflect a simple mouse strain difference. Alternatively, tachycardia may represent an early sign of cardiotoxicity at the higher doses of CoCl2.
Effect of CoCl2 on infarct size.
Pretreatment of ICR mice with a relatively low dose of CoCl2 (30 mg/kg) 24 h before I/R resulted in reduction of infarct size (16.1 ± 3.1%) compared with saline-treated mice (27.6 ± 3.6%, P < 0.05; Fig. 1). This infarct-limiting effect was not observed in mice treated with 60 or 120 mg/kg CoCl2 (P > 0.05 vs. saline treated). CoCl2 (30 mg/kg) also protected B6,129-WT mice (12.0 ± 2.8%) compared with saline-treated B6,129-WT mice (22.8 ± 3.4%, P < 0.05; Fig. 2). The delayed cardioprotection induced with CoCl2 was abolished by DDTC (150 mg/kg), an antioxidant, given 30 min before CoCl2 treatment (infarct size 21.5 ± 6.1% vs. 22.8 ± 3.4% in saline + B6,129 group, P > 0.05; see Fig. 2). The infarct-limiting effect of CoCl2 was preserved in mice lacking the p50 subunit of NF-κB (11.1 ± 3.0% vs. 25.1 ± 5.0% in saline + p50-KO group; P < 0.05). In contrast, CoCl2 failed to reduce infarct size in iNOS-KO mice, i.e., 20.9 ± 3.0% in the CoCl2 + iNOS-KO group (n = 7) vs. 23.5 ± 3.8% in nontreated iNOS-KO mice (n = 10) as previously reported (Ref. 34; P > 0.05).
Postischemic cardiac contractile function and coronary flow.
There was no significant difference in the preischemic values of developed force and force-rate product (Fig. 3) among all experimental groups. After I/R, the rate-force product was depressed in all groups (Fig. 3), primarily because of the decrease in developed force, because the heart rate remained reasonably constant at the preischemic level as shown in Table 1. CoCl2 treatment resulted in mild and statistically insignificant improvement in postischemic cardiac contractile function (Fig. 3). In addition, administration of the cardioprotective dose of CoCl2 had no effect on postischemic coronary flow compared with the saline-treated group (P > 0.05; data not shown).
DNA binding activity of transcription factors.
A basal level of DNA binding activity for HIF-1α (Fig. 4) and AP-1 (Fig. 5) was detected with EMSA from the nontreated control hearts. Treatment with CoCl2 (30 mg/kg) resulted in rapid increase in the DNA binding activity of HIF-1α and AP-1, which was maintained for the next 4 h before returning to the baseline level 6 h after treatment (Figs. 4 and 5, bottom, time course of averaged densitometry results). In contrast, no DNA binding activity of NF-κB was observed in the CoCl2-treated hearts throughout the 6-h postdrug treatment period (Fig. 6). Interestingly, exposure of mice to systemic hypoxia (10% inspired O2) for 2–4 h did not activate NF-κB (Fig. 6). On the other hand, LPS (5 mg/kg; served as positive control) caused a robust increase in the DNA binding activity of NF-κB after 30 min, with further increases by 1 and 2 h after treatment (Fig. 6). These results were corroborated by the lack of nuclear translocation of p65 in all four individual heart samples after treatment with the cardioprotective dose of CoCl2 i.e., 30 mg/kg (Fig. 7, top), similar to the saline-treated mice (Fig. 7, bottom left). In contrast, LPS caused remarkable translocation of p65 from the cytosolic to the nuclear fraction (Fig. 7, bottom right) compared with saline-treated mice (Fig. 7, bottom left).
Although acute systemic hypoxia induces delayed preconditioning in the heart, it is not known whether such protection can be realized by chemical induction of hypoxic response with CoCl2. In the present study, we have demonstrated, for the first time, that a single bolus dose of CoCl2 (30 mg/kg) induced a delayed preconditioning-like cardioprotective effect, as evidenced by reduction of infarct size after global I/R in the mouse heart (Fig. 1). In fact, we initially used the drug dose of 60 mg/kg (ip), which was previously reported by Bergeron et al. (4) to protect neonatal rat brain against I/R injury. However, the 60 mg/kg dose and another higher dose (120 mg/kg) were not able to reduce infarct size in the adult mouse heart (Fig. 1).
Our results also show that CoCl2 selectively activated HIF-1α (Fig. 4) and AP-1 (Fig. 5), suggesting the role of these transcription factors in CoCl2-induced delayed cardioprotection. These data are in agreement with the previous study by Cai et al. (7), which suggested a causal relationship between HIF-1α activation and delayed cardioprotection. In contrast, we failed to demonstrate the effect of CoCl2 on activation of NF-κB, as shown by lack of augmentation in DNA-binding activity after CoCl2 or systemic hypoxia treatment (for up to 6 h; Fig. 6). Also, CoCl2 did not cause protein translocation of the p65 subunit of NF-κB from the cytosolic to the nuclear fraction (Fig. 7). Furthermore, genetic deletion of the p50 subunit of NF-κB did not abolish the cardioprotective effect in mice (Fig. 2). p50-KO mice are deficient in the p105 precursor to p50 and thus are unable to assemble the transcriptionally active p50-p65 heterodimer (27). These results were surprising, because NF-κB has been shown to be an essential transcription factor in the delayed preconditioning induced with brief episodes of I/R (37) or pharmacological agents such as anisomycin (39), diazoxide (31), and the adenosine A3 receptor agonist N6-(3-iodobenzyl) adenosine-5′-N-methyluronamide (38). Our data suggest that delayed pharmacological cardioprotection can also be achieved without the activation of NF-κB.
Another interesting finding in this study is that the putative antioxidant DDTC blocked CoCl2-induced cardioprotection, suggesting that reactive oxygen species (ROS) may serve as the triggers/messengers for the downstream effects of CoCl2. It has been demonstrated that CoCl2 increases ROS generation or oxidative stress in divergent cell types (8, 29). Chandel et al. (8) showed that CoCl2 stabilizes HIF-1α through ROS generation by a nonenzymatic, nonmitochondrial mechanism in the hepatocytes. Although our study did not address the direct role of ROS in stabilization of HIF-1α (Fig. 4) in the heart, such a possibility exists because CoCl2 activated HIF-1α and DDTC blocked the CoCl2-induced protection (Fig. 2). In addition, CoCl2 may also stabilize HIF-1α by antagonizing Fe2+, which is an essential cofactor along with the oxygen for prolyl hydroxylation that causes protein degradation of HIF-1α. It is also known that the AP-1 transcription factor is made up of a family of regulatory proteins that can be activated by oxidative stress. Moreover, previous studies have proposed an important role of ROS in the signaling cascade leading to delayed cardioprotection (5). Additionally, the observation that DDTC blocked CoCl2-induced delayed cardioprotection, which is independent of NF-κB (Figs. 6 and 7), indirectly suggests that DDTC is not a selective inhibitor of NF-κB. Recent cellular evidence also suggested that ROS do not mediate NF-κB activation (14). This may explain why the signaling mechanism of CoCl2-induced protection involves ROS but not NF-κB.
Our data also suggest that CoCl2-induced delayed protection is dependent on iNOS because the infarct-limiting effect was absent in the iNOS gene-deficient mice (Fig. 2). Many studies have demonstrated that iNOS is an essential mediator of the delayed phase of myocardial preconditioning induced by brief episodes of I/R (5), acute systemic hypoxia (36), and a number of pharmacological agents (31, 33, 38, 39). It could be postulated that iNOS is also required in mediation of CoCl2-induced delayed protection, which is primarily regulated by HIF-1α and AP-1 but not NF-κB. This assumption is based on a previous study that demonstrated the upregulation of iNOS by systemic hypoxia via HIF-1α in adult rat cardiac myocytes (16). As demonstrated previously (31, 33, 38, 39), the upregulation of iNOS may increase myocardial NO generation that in turn opens mitochondrial ATP-sensitive K+ channels (25), which have been considered as the potential end-effector for myocardial cardioprotection induced by various stimuli including systemic hypoxia (2, 3) and drugs (31, 38).
CoCl2 is a water-soluble compound that was traditionally used to treat anemia in pregnant women and infants (15) and used to treat refractory anemia in patients undergoing long-term hemodialysis (6) since the 1940s. Despite an earlier report showing the direct action of cobalt in inhibiting hypoxic contracture and preserving ATP in adult rat myocardium (10), the ability of CoCl2 to induce cytoprotective effects was not recognized until recently. For instance, long-term oral intake of low-dose CoCl2 (5 mg/kg for 4 wk) was found to enhance postischemic ventricular function in adult rats (11). Chronic treatment with CoCl2 (75 mg/kg, 3 times/wk for 5 wk) caused significant angiogenesis in rat myocardium (24), and such an increase in arteriolar and capillary supply was not seen in the Epo-treated animals. Most recently, the remarkable cytoprotection induced by CoCl2 or cobalt protoporphyrin against lethal I/R injury has been observed not only in the heart (18) but also in other vital organs including the brain (4), liver (17), and kidneys (19). CoCl2 has been shown to reduce apoptotic death induced by tert-butyl hydroperoxide and serum deprivation, as measured by DNA fragmentation in hepatoma HepG2 cells (23). Pretreatment with cobalt protoporphyrin markedly reduced the number of apoptotic myocytes/endothelial cells in the rat heart 24 h after cold I/R injury (18). This antiapoptotic effect resulted in a significant improvement in cardiac graft survival (18). Therefore, CoCl2, the low-cost and water-soluble chemical compound, may have potential use as an effective inducer of ischemia-resistant phenotype in multiple organs under the settings of organ transplantation or bypass surgeries.
Despite its beneficial effects, CoCl2 was identified as a contributing etiologic factor for the outbreak of so-called “Quebec beer-drinkers' cardiomyopathy” in the late 1960s (1, 22). Several investigators attempted to simulate this human pathological condition in the animal models of cobalt-induced cardiomyopathy (12, 13, 21, 30). The remarkable individual variance of cobalt-induced cardiotoxicity was well recognized by several investigators (1, 12, 13, 32). It was suggested (1, 32) that the cardiotoxic potential of cobalt could be affected by a number of factors, including route of administration, preexisting heart damage, age, length of exposure, nutritional status, and dietary composition (e.g., low protein and high alcohol concentration). For example, the canine model of cobalt-induced cardiomyopathy required a daily intravenous dose of 5 mg/kg cobalt sulfate in conjunction with a low-protein, low-thiamine diet (30). Chronic cobalt exposure (6 mo) resulted in a marked decrease in the activity of manganese superoxide dismutase (a major antioxidant enzyme) as well as in mitochondrial ATP production in rat myocardium (9). Cobalt-induced heart failure occurred only with a 110 mg/kg cumulative dose of cobalt sulfate (i.e., the daily intravenous injection continued for >3 wk). On the other hand, several studies in 1955 (15, 26) reported no mortality and no appreciable toxic effect on thyroid or liver function in human patients under CoCl2 therapy (75–100 mg daily for 6–9 mo) or in laboratory rodents receiving high oral doses of CoCl2 (40 mg/kg per day) for up to 4 mo. Similarly, recent rat studies (11, 24) observed no severe systemic toxic effects and/or mortality with CoCl2.
In summary, we have demonstrated for the first time that pretreatment with a low dose of CoCl2 induces delayed protection against myocardial infarction in adult mice. Furthermore, our data show the involvement of ROS, HIF-1α, AP-1, and iNOS, but not NF-κB, in the cellular signaling mechanisms of CoCl2-induced late cardioprotection. These results suggest that CoCl2 initiates a distinctive NF-κB-independent signaling paradigm that leads to late cardioprotection. Further studies are needed to identify the cause-and-effect role of other protective protein(s) such as heme oxygenase 1, which is known to be induced by cobalt (17–19, 29).
This study was supported by National Heart, Lung, and Blood Institute (NHLBI) grants HL-51045 and HL-59469 to R. C. Kukreja and by American Heart Association, Mid-Atlantic Affiliate Grant 0060289U to L. Xi. C. Yin was supported by NHLBI Training Grant T32-HL-07537.
Present address for M. Taher: Dept. of Pharmacology, Rush University Medical Center, 1735 W. Harrison St., Chicago, IL 60612.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by the American Physiological Society