Cardioprotective and Antiapoptotic Effects of Heme Oxygenase-1 in the Failing Heart
Background— Heme oxygenase-1 (HO-1) is an inducible stress-response protein that imparts antioxidant and antiapoptotic effects. However, its pathophysiological role in cardiac remodeling and chronic heart failure (HF) is unknown. We hypothesized that induction of HO-1 in HF alleviates pathological remodeling.
Methods and Results— Adult male nontransgenic and myocyte-restricted HO-1 transgenic mice underwent either sham operation or coronary ligation to induce HF. Four weeks after ligation, nontransgenic HF mice exhibited postinfarction left ventricular (LV) remodeling and dysfunction, hypertrophy, fibrosis, oxidative stress, apoptosis, and reduced capillary density, associated with a 2-fold increase in HO-1 expression in noninfarcted myocardium. Compared with nontransgenic mice, HO-1 transgenic HF mice exhibited significantly (P<0.05) improved postinfarction survival (94% versus 57%) and less LV dilatation (end-diastolic volume, 46±8 versus 85±32 μL), mechanical dysfunction (ejection fraction, 65±9% versus 49±16%), hypertrophy (LV/tibia length 4.4±0.4 versus 5.2±0.6 mg/mm), interstitial fibrosis (11.2±3.1% versus 18.5±3.5%), and oxidative stress (3-fold reduction in tissue malondialdehyde). Moreover, myocyte-specific HO-1 overexpression in HF promoted tissue neovascularization and ameliorated myocardial p53 expression (2-fold reduction) and apoptosis. In isolated mitochondria, mitochondrial permeability transition was inhibited by HO-1 in a carbon monoxide (CO)–dependent manner and was recapitulated by the CO donor tricarbonylchloro(glycinato)ruthenium(II) (CORM-3). HO-1–derived CO also prevented H2O2-induced cardiomyocyte apoptosis and cell death. Finally, in vivo treatment with CORM-3 alleviated postinfarction LV remodeling, p53 expression, and apoptosis.
Conclusions— HO-1 induction in the failing heart is an important cardioprotective adaptation that opposes pathological LV remodeling, and this effect is mediated, at least in part, by CO-dependent inhibition of mitochondrial permeability transition and apoptosis. Augmentation of HO-1 or its product, CO, may represent a novel therapeutic strategy for ameliorating HF.
Received August 31, 2009; accepted March 1, 2010.
Heme oxygenase-1 (HO-1) is a rapidly inducible cytoprotective protein that degrades heme to biliverdin, ferrous iron, and carbon monoxide (CO).1 HO-1 mitigates cellular injury by exerting antioxidant, antiapoptotic, and antiinflammatory effects1–4; these benefits of HO-1 have been widely demonstrated in a variety of pathological states including hypoxic lung disease,5 vascular injury,6 and cardiac transplant rejection.7 Conversely, mice with germ-line HO-1 disruption exhibit progressive anemia, inflammation, and oxidative stress with age.8,9 HO-1–mediated cytoprotection likely reflects the effects of its catalytic products, especially CO, because CO recapitulates the cytoprotective profile of HO-1 and can rescue the injurious effects of HO-1 deficiency.1–4,7,10
Clinical Perspective on p 1925
Short-term HO-1–mediated cytoprotection extends to the heart. Heterozygous HO-1+/− mice exhibit exaggerated cardiac injury and dysfunction after ischemia/reperfusion, findings partially rescued by antioxidants.11 In contrast, mice with cardiac-restricted HO-1 overexpression are resistant to ischemia/reperfusion injury, with improved contractile recovery and reduced infarct size, inflammatory cell infiltration, oxidative damage, and apoptosis.12 Similar results were obtained in rat hearts subjected to ischemia/reperfusion 8 weeks after human HO-1 gene transfer13; the improvement in left ventricular (LV) function was maintained for up to 1 year after injury.14 Nevertheless, even though the salutary effects of HO-1 during short-term cardiac ischemic stress have been established, it is not known whether such a paradigm can be extended to chronic heart failure (HF). This issue is particularly important because HF is characterized by prolonged activation of several stress-response systems (eg, inflammatory cytokines, catecholamines) that, although initially compensatory and protective, ultimately produce long-term detrimental effects. Although prior studies have reported that HO-1 is upregulated in human failing hearts15 and in animal models of right ventricular failure,16 it is not clear whether such long-term induction is beneficial or detrimental. Accordingly, in this study we tested the hypothesis that HO-1 upregulation in the failing heart is a cardioprotective adaptation that alleviates pathological LV remodeling.
An expanded Methods section is included in the online-only Data Supplement. All studies were performed in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (Department of Health and Human Services publication NIH 85-23, revised 1996).
Male mice (10 to 16 weeks of age, weighing 25 to 30 g) were used. Cardiac-specific HO-1 transgenic mice were obtained from Dr Shaw-Fang Yet (Harvard Medical School). The HO-1 transgenic mouse line used expresses 4 copies of a human HO-1 transgene under the control of the α-MyHC promoter.12 HO-1 transgenic mice were backcrossed for a minimum of 7 generations into C57BL/6. Nontransgenic littermates were used as controls. For in vivo tricarbonylchloro(glycinato)ruthenium(II) (CORM-3) supplementation studies, C57BL/6 mice were used.
Permanent coronary ligation (or sham operation) in mice was performed as described previously.17 The mice were followed for 4 weeks after operation.
Mouse echocardiography was performed and analyzed at baseline and 4 weeks postoperatively as described previously.17
LV Pressure Measurement
In a subset of animals, LV mechanical performance was assessed with the use of a Millar pressure-conductance catheter 4 weeks after coronary ligation or sham operation as described previously.18
Isolation of Mouse Cardiomyocytes
Ca2+-tolerant mouse ventricular myocytes were isolated with the use of collagenase digestion as described previously.18 Myocyte cell death was determined with the use of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay after incubation of isolated myocytes with 500 μmol/L H2O2 for 1 hour. Absorbance at 550 nm was used to index cell viability. In some studies, myocytes were also exposed to either 50 μmol/L hemoglobin (CO scavenger) or 100 μmol/L desferoxamine (iron chelator) beginning 30 minutes before addition of H2O2.
H9c2 Cardiomyocytes and HO-1 Transfection
H9c2 cells (rat embryonic cardiomyoblasts) were obtained from the American Type Culture Collection. Ad5/HO-1, the replication-deficient adenoviral vector containing the entire coding region of rat HO-1 complementary DNA, was generated in 293 cells by homologous recombination of pCMV/HO-1 and pJM17, a circularized adenovirus genome lacking its E1 region and a portion of the E3 region. Plaque-isolated viral clones were propagated in 293 cells and then purified over 2 CsCl gradients and titered by plaque assay as described previously.19 Ad5/LacZ containing the β-galactosidase gene driven by the CMV promoter was used as a control vector. H9c2 cells were seeded in 100-mm tissue culture dishes and transiently transfected with Ad5/HO-1 or Ad5/LacZ for 45 minutes at a multiplicity of infection of 10 for 72 hours before treatments.
Masson trichrome staining for collagen was performed as described previously.17 Myocyte cross-sectional area was determined in sections stained with rhodamine-conjugated wheat germ agglutinin (Molecular Probes, Carlsbad, Calif). Oxidative stress was indexed by immunostaining for malondialdehyde-adducted proteins as described previously.20 For HO-1 immunofluorescent staining, we used antimouse HO-1 antibody (Stressgen) and tetramethylrhodamine isothiocyanate–conjugated secondary antibody. Immunostaining with fluorescein isothiocyanate–conjugated isolectin B4 (Vector Laboratories, Burlingame, Calif) was performed to determine tissue capillary density. Apoptosis in tissue and cells was determined with the use of the DeadEnd fluorometric terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL) system (Promega). Optical sections were obtained with a Zeiss LSM510 inverted confocal scanning laser microscope equipped with Enterprise/argon/HeNe lasers and excitation wavelengths appropriate for multichannel scanning. For immunohistological analyses, we typically analyzed 6 fields per heart with an area of 66 mm2 per field (total area, 396 mm2 per heart).
HO-1 Gene Expression
Total RNA was isolated from cardiac tissue as described previously.17,20 HO-1 mRNA was quantified and normalized to mouse β-actin mRNA with the use of real-time polymerase chain reaction with LUX gene-specific primers (Invitrogen, Carlsbad, Calif).
Protein extraction, Western immunoblotting, and densitometry were performed as described previously.17,18,20 Primary antibodies used included antip53, anti-poly-ADP ribose polymerase (PARP), anti-Bax, anti-Bcl-2, anti-β-actin, and anti-α-tubulin from Santa Cruz Biotechnology (Santa Cruz, Calif), and anti-HO-1 from Stressgen.
Measurement of Free Malondialdehyde by Gas Chromatography–Negative Ionization Chemical Ionization–Mass Spectrometry
Tissue malondialdehyde concentration was measured by gas chromatography–negative ionization chemical ionization–mass spectrometry (GC-NICI-MS) as described previously.21 Heart tissue homogenate was derivatized with pentfluorobenzylhydroxylamine, and the pentfluorobenzyl-oxime derivatives of malondialdehyde were measured by GC-NICI-MS with the select ion monitoring mode. Benzaldehyde ring d5 was used as the internal standard. The following ions were monitored for the indicated aldehyde: benzaldehyde d5-m/z 286 (M+-HF) and malondialdehyde m/z 204 (M+-C7H2F5-HFNO-C2H3).
Mitochondrial Membrane Permeability Transition
In adult cardiomyocytes and H9c2 cells, changes in mitochondrial membrane potential (Δψm) were assessed with tetramethylrhodamine methyl ester (TMRM). TMRM fluorescence was measured at 5-minute intervals for 30 minutes in live cells with a 585-nm long-pass filter and laser-scanning confocal microscopy. TMRM is a cationic dye that accumulates in mitochondria in proportion to Δψm. The TMRM concentration chosen (100 nmol/L) does not suppress mitochondrial respiration22 and is nonquenching so that mitochondrial depolarization is accompanied by a decrease in cellular TMRM fluorescence over time followed by cell shortening due to ATP depletion.23The time required for a 2-fold decrease in TMRM fluorescence was taken as an end point for membrane permeability transition (MPT).
Mitochondrial MPT Pore Opening
Mitochondria were isolated from adult mouse hearts, and MPT pore opening in isolated mitochondria was induced by Ca2+ as described previously.24 Isolated mitochondria were resuspended in swelling buffer (containing the following [in mmol/L]: 120 KCl, 10 Tris-HCl [pH 7.4], 20 MOPS, and 5 KH2PO4) to a final protein concentration of 0.25 mg/mL. Mitochondrial swelling induced by pore opening was measured spectrophotometrically as a reduction in absorbance at 520 nm (A520).
Mitochondria were resuspended in respiration buffer (pH 7.2) containing the following (in mmol/L): 225 mannitol, 70 sucrose, 10 KH2PO4, 1 EGTA. A Clark electrode was used to measure the oxygen content of the mitochondrial suspension with 10 mmol/L pyruvate and 5 mmol/L malate. State 4 respiration was measured at baseline; 330 μmol/L ADP was added to stimulate state 3 respiration. The respiratory control ratio was defined as the state 3 to state 4 respiration ratio. The ADP to O ratio was calculated with the use of total oxygen consumption during state 3 respiration.
Preparation of the CO Donor CORM-3
CORM-3, a water-soluble transition metal carbonyl drug that stably releases CO, was synthesized following a published protocol,25 and its structure was confirmed by infrared spectroscopy and nuclear magnetic resonance. The synthesized drug was stored at −20°C, and fresh solution was made just before use. For control experiments, CORM-3 was inactivated by dissolving it in Krebs-Henseleit buffer and allowing CO liberation overnight at room temperature.
Several statistical techniques were employed. For 2-group comparisons, we used the unpaired 2-sample t test. For comparisons of >2 groups, we used 1-way ANOVA if there was 1 independent variable, 2-way ANOVA if there were 2 independent variables (eg, genotype and ligation status), and 2-way repeated-measures ANOVA for matched observations over time with 2 independent variables. To adjust for multiple comparisons, we performed Student-Newman-Keuls posttest for 1-way ANOVA and Bonferroni posttest for 2-way ANOVA. Pairwise comparisons were made between sham groups across genotypes, sham versus HF within each genotype, and HF groups across genotypes. Animal survival was evaluated by the Kaplan–Meier method, and the log-rank test was used to compare survival curves between nontransgenic sham and HF, HO-1 transgenic sham and HF, and nontransgenic and HO-1 transgenic HF. A value of P<0.05 was considered significant. Continuous data are summarized as mean±SD.
HO-1 Is Upregulated in Murine Postinfarction HF
Ten-week-old C57BL/6 mice were evaluated 4 weeks after coronary ligation or sham operation. Figure IA in the online-only Data Supplement shows short-axis LV sections from a sham and failing heart and M-mode echocardiograms from a mouse at baseline and 4 weeks after myocardial infarction. In this example, a large infarct is present along the anterior and lateral aspects of the LV. Echocardiography revealed marked chamber remodeling with increased LV dimensions and reduced fractional shortening. As shown in Figure IB in the online-only Data Supplement, there was a 2-fold increase in cardiac HO-1 mRNA and protein levels in HF over sham groups, without changes in HO-1. HO-1 immunofluorescent staining from wild-type (WT) sham and HF hearts also indicated increased myocyte HO-1 expression in HF compared with sham groups (Figure IC in the online-only Data Supplement). For comparison, staining from nontransgenic and HO-1 transgenic hearts is also shown, indicating robust HO-1 expression in myocytes but not in the large vessels in HO-1 transgenic hearts.
HO-1 Overexpression Improves Postinfarction Survival and LV Remodeling
Baseline echocardiography confirmed the absence of a cardiac phenotype in HO-1 transgenic mice (Table I in the online-only Data Supplement). Kaplan–Meier survival curves after coronary ligation or sham operation in nontransgenic and HO-1 transgenic mice (Figure 1A) revealed that whereas mortality in nontransgenic HF was markedly higher than that in nontransgenic sham, myocyte-specific HO-1 overexpression imparted a survival benefit at 28 days after infarction. Figure 1B and 1C show short-axis LV sections, M-mode echocardiograms, and group data from nontransgenic and HO-1 transgenic sham and failing hearts. There was significant chamber dilatation (increased LV end-diastolic volume and LV end-systolic volume) and systolic dysfunction (reduced LV ejection fraction) in both HF groups over sham. However, in comparison with nontransgenic HF, LV dilatation and dysfunction were attenuated in HO-1 transgenic HF. Indeed, in the example in Figure 1B, this occurred despite larger overall infarct size in the HO-1 transgenic mouse. The LV/tibia length ratio showed a similar response pattern (Figure 1D), indicating less hypertrophy in HO-1 transgenic HF. Hemodynamic recordings (Figure 1F) and pressure-volume loops (Figure 1G) indicated improved contractility, LV dilatation, LV filling pressure, and diastolic function in HO-1 transgenic HF compared with nontransgenic HF, without appreciable differences between sham groups (Table). Notably, overall group infarct size was similar in nontransgenic and HO-1 transgenic HF (Figure 1E), suggesting that differences in remodeling and survival were independent of the degree of initial injury. Collectively, the data show that HO-1 overexpression alleviates postinfarction LV remodeling and improves mechanical performance.
HO-1 Overexpression Attenuates Hypertrophy, Fibrosis, and Oxidant Stress in the Failing Heart
Cell membrane staining with rhodamine wheat germ agglutinin revealed myofiber hypertrophy in both nontransgenic and HO-1 transgenic HF compared with sham-operated hearts (Figure 2A). The degree of hypertrophy, however, was attenuated in HO-1 transgenic HF hearts, confirming the gravimetric data (Figure 1D). To evaluate myocardial oxidative stress, we evaluated both protein-bound malondialdehyde by immunostaining and free malondialdehyde by GC-NICI-MS. As seen by the immunostains in Figure 2B, malondialdehyde-modified proteins were increased in both HF groups over sham but significantly less so in HO-1 transgenic HF. Figure 2C shows representative GC chromatograms (left), the m/z values of ions 286 (benzaldehyde d5) and 204 (malondialdehyde) (center), and the GC-NICI-MS quantitative group data (right). HF in nontransgenic mice increased tissue malondialdehyde levels 2-fold over those of nontransgenic sham mice. In contrast, myocyte-specific HO-1 overexpression completely prevented the HF-associated increase in tissue malondialdehyde. These data indicate a potent antioxidant effect of HO-1 in the remodeling heart. Additionally, Masson’s trichrome staining (Figure 2D) revealed less collagen deposition in the remote myocardium of HO-1 transgenic HF hearts, suggesting an antifibrotic effect of HO-1.
HO-1 Is Antiapoptotic and Promotes Neovascularization in the Failing Heart
Figure 3A shows representative TUNEL stains (with troponin I costaining) from the remote and border zone of a nontransgenic HF heart, illustrating TUNEL-positive nuclei. Quantitative data demonstrated that nontransgenic and HO-1 transgenic HF groups exhibited significantly greater apoptosis of both myocytes and nonmyocytes (differentiated on the basis of attendant troponin I costaining) than their respective sham. The apoptotic rate we observed in the failing heart (≈0.50%) was both consistent with that reported for human HF (≈0.12% to 0.70%)26 and physiologically significant because sustained apoptotic rates as low as 0.23% induce dilated cardiomyopathy in mice.27 However, HO-1 transgenic HF hearts exhibited a significant reduction in myocyte and nonmyocyte apoptotic rate versus nontransgenic HF. As a second measure of apoptosis, we determined myocardial levels of cleaved and uncleaved PARP protein. As seen in Figure 3B, the ratio of cleaved to uncleaved PARP increased in nontransgenic HF over sham (indicating enhanced apoptosis) but did not change appreciably between HO-1 transgenic sham and HO-1 transgenic HF, consistent with the TUNEL stains. Moreover, the protein level of p53, a central transcriptional activator of multiple proapoptotic genes, was increased 2-fold in nontransgenic HF. This increase was suppressed in HO-1 transgenic HF (with p53 levels ≈50% lower than in nontransgenic mice) as well as in HO-1 transgenic sham hearts (Figure 3B). Because p53 also has prominent antiangiogenic effects,28 we also determined the effects of cardiac HO-1 overexpression on tissue neovascularization after infarction. As shown in Figure 3C, isolectin staining revealed that LV capillary density (excluding LV scar) was markedly decreased in nontransgenic HF in comparison with sham-operated hearts but was maintained in HO-1 transgenic HF. Moreover, among the HF groups, border zone capillary density was also significantly higher in HO-1 transgenic hearts. These results establish that HO-1 exerts important antiapoptotic effects and promotes neovascularization in the failing heart, potentially related in part to p53.
HO-1 Suppresses Mitochondrial Respiration, MPT, and Cardiomyocyte Apoptosis in a CO-Dependent Manner
We next evaluated the effects of HO-1 on mitochondrial respiration, MPT, and cardiomyocyte apoptosis. In HO-1 transgenic hearts, robust levels of HO-1 were associated with the mitochondrial fraction (≈40% of total), indicating physical proximity of HO-1 to the mitochondrial pore in these hearts but not in WT hearts (Figure 4A). Respiration studies were performed in freshly isolated mitochondria from nontransgenic and HO-1 transgenic hearts in the presence or absence of the CO donor CORM-3 (iCORM-3 control) or the CO scavenger hemoglobin. ADP to O ratios were similar among all experimental groups, ranging from 2.1±0.05 to 2.8±0.22, indicating a relatively constant relationship between ATP synthesis and oxygen consumption (data not shown). In contrast, the ratio of state 3 to state 4 respiration (respiratory control ratio) was suppressed in both HO-1 transgenic mitochondria as well as nontransgenic mitochondria incubated with CORM-3 but not inactive iCORM-3 (Figure 4B). Both HO-1 transgenic mitochondria and nontransgenic mitochondria treated with CORM-3 showed a general increase in absolute state 3 and state 4 respiration (Figure II in the online-only Data Supplement), although, overall, respiration was suppressed. The increase in CO did not change coupling because the ADP to O ratio was not affected (data not shown). Moreover, the suppressed state 3 to state 4 ratio in HO-1 transgenic mitochondria was normalized on incubation with the CO scavenger hemoglobin, indicating CO dependence of this effect. Ca2+-induced mitochondrial swelling assays were then performed to assess MPT (Figure 4C). Robust swelling was induced by Ca2+ in nontransgenic cardiac mitochondria; this was prevented by the MPT pore inhibitor cyclosporine A as well as by CORM-3 (but not iCORM-3). HO-1 transgenic cardiac mitochondria were resistant to swelling, a response that was reversed by coincubation with hemoglobin. Nontransgenic mitochondria pretreated with recombinant HO-1 protein were also less susceptible to Ca2+-induced MPT (Figure III in the online-only Data Supplement).
We next measured Δψm and MPT in living adult cardiomyocytes or H9c2 cells. WT or HO-1 transgenic cardiomyocytes were loaded with TMRM, and phenylarsine oxide (PAO) (20 μmol/L) was used to induce MPT. The time required for a 2-fold decrease in TMRM fluorescence was taken as an end point for MPT. TMRM mitochondrial labeling revealed the typical cardiomyocyte pattern of fluorescent bands oriented along the longitudinal axis of adult cardiomyocytes or a striking punctuate pattern in H9c2 cells (Figure IV in the online-only Data Supplement). In WT cardiomyocytes, PAO induced MPT within 15 minutes, as indicated by the decline in TMRM fluorescence (Figure 5A). The fluorescence decline was largely prevented by cyclosporine A, showing that the change was specifically a consequence of MPT. In comparison to WT myocytes, MPT induction was significantly delayed in HO-1 transgenic cardiomyocytes, analogous to the results of the mitochondrial swelling studies. Moreover, pretreatment of H9c2 cardiomyocytes with CORM-3, but not iCORM-3, attenuated PAO-induced MPT (Figure V in the online-only Data Supplement), and H9c2 cells transfected with HO-1 adenovirus, which augmented HO-1 colocalization with mitochondria (Figure VI in the online-only Data Supplement), were more resistant to H2O2-induced MPT, as indexed by TMRM labeling (Figure VII in the online-only Data Supplement). Taken together, these results suggest that HO-1, in large part via CO, inhibits both mitochondrial respiration and MPT in cardiomyocytes.
We further evaluated whether the products of HO-1 catalysis had analogous protective effects on cardiomyocyte apoptosis and cell death. H2O2-induced apoptosis was evaluated by TUNEL staining, and cell death was evaluated by MTT assay. H2O2 (100 μmol/L) induced robust apoptosis (Figure 5B) and cell death (Figure 5C) in WT myocytes but not in HO-1 transgenic myocytes. The resistance of HO-1 transgenic myocytes to H2O2-induced cell death was abrogated on cotreatment with hemoglobin and, to a lesser extent, with the iron chelator desferoxamine. However, the interpretation of the latter was confounded by the finding that desferoxamine alone also induced cell death (albeit to a lesser degree) in both WT and HO-1 transgenic myocytes. Taken together, analogous to its effect on MPT, HO-1 is antiapoptotic in cardiomyocytes, an effect in large part a consequence of its product CO (and perhaps ferrous iron). Collectively, these data suggest that CO-mediated prevention of MPT may be an important mechanism underlying the antiapoptotic effects of HO-1 in HF.
CORM-3 Treatment In Vivo Alleviates Postinfarction LV Remodeling
Having found that CO is a key cytoprotective mediator of the antiapoptotic effects of HO-1, we tested whether CORM-3 treatment would attenuate LV remodeling after myocardial infarction in vivo. C57BL/6 mice were subjected to coronary ligation and, starting 4 days later, were administered 40 mg/kg IP CORM-3 daily for 24 days, after which LV remodeling and function were assessed with echocardiography and tissue morphometry. The biological activity of this dose of CORM-3 was tested by measuring blood carboxyhemoglobin. Before drug administration, carboxyhemoglobin in the blood was ≈1%, whereas 2 to 6 hours after injection, carboxyhemoglobin increased to 5.9±0.6% and remained at this level for 24 hours, indicating long-lasting systemic delivery of CO. Figure 6A shows representative LV sections from a sham animal, a HF animal without CORM-3, and a HF animal with CORM-3 treatment. In comparison with untreated HF, the CORM-3–treated heart had less LV dilatation and remodeling. Group echocardiographic and gravimetric data 4 weeks after infarction revealed that sustained CORM-3 administration attenuated LV dilatation, improved systolic function, and reduced LV hypertrophy despite equivalent infarct size between untreated and treated groups (Figure 6B and 6C). Hemodynamic data also indicated better LV mechanical performance with CORM-3, with improvements in dP/dtmax normalized for instantaneous LV pressure, time constant of LV relaxation, and LV end-diastolic pressure over untreated HF. The CORM-3–mediated benefits occurred despite equivalent afterload in the treated and untreated groups (LV peak systolic pressure, 84±4 versus 84±10 mm Hg, untreated versus CORM-3–treated HF; P=NS) (Figure 6D). Moreover, as was the case with HO-1 transgenic mice, infarcted mice treated with CORM-3 also exhibited markedly reduced p53 expression (Figure 6E) and less myocardial apoptosis by TUNEL staining (Figure 6F) than untreated HF mice. Figure 6G depicts expression of Bax, a proapoptotic protein that can be induced by p53, and Bcl-2, an antiapoptotic protein that can be repressed by p53.29 Compared with untreated HF, CORM-3–treated HF hearts exhibited significantly augmented Bcl-2 levels and a trend (although not statistically significant) toward reduced Bax expression. These results are consistent with a p53-mediated effect in CORM-3–treated HF hearts and are also in agreement with the observed reduction in apoptotic rate. These data demonstrate that exogenous CO administration by CORM-3 alleviates postinfarction LV remodeling and apoptosis and that, in the failing heart, exogenous CO augments the cardioprotective effects of HO-1 upregulation.
In this study, we show for the first time that HO-1 induction in the failing heart is a cardioprotective adaptation that alleviates postinfarction pathological LV remodeling, an effect mediated in part by CO-dependent antiapoptotic actions that are associated with, and likely due to, prevention of MPT. Several lines of evidence support these conclusions. First, HO-1 expression was upregulated 2-fold in chronically remodeled failing myocardium remote from the infarct and well after the formation of a stable scar. Second, HO-1 overexpression and gain of function in the heart ameliorated LV dilatation and dysfunction, hypertrophy, interstitial fibrosis, and oxidative stress and improved tissue neovascularization. These results indicate both a cardioprotective (antihypertrophic, antioxidant, antifibrotic, and proangiogenic) role of HO-1 in HF and a potential therapeutic effect of enhancing HO-1 function over and above the 2-fold upregulation seen in HF. Third, our observations in HO-1 transgenic mice establish an in vivo antiapoptotic effect of HO-1 in the failing heart that mitigates cell loss, p53 expression, and pathological remodeling. Fourth, in isolated mitochondria and cardiomyocytes, we found that HO-1, primarily via CO, suppresses the mitochondrial respiratory control ratio and MPT and inhibits apoptosis. Finally, sustained CO donor treatment in vivo alleviated postinfarction remodeling and attenuated apoptosis in a manner similar to HO-1 overexpression, further underscoring the potential utility of enhancing the HO-1 axis in chronic HF. Taken together, these findings indicate that augmentation of HO-1 and/or its product, CO, may represent a novel and beneficial therapeutic approach in HF.
HO-1 Upregulation Is a Cardioprotective Adaptation in the Failing Heart
The cytoprotective effects of HO-1 are mediated by its product CO (which is generated exclusively by the HO system), the production of the antioxidants biliverdin and bilirubin, and the degradation of excess amounts of the pro-oxidant heme.1,2 Previous studies have shown cardioprotective actions of HO-1 in models of acute stress (eg, ischemia/reperfusion injury).11–14 However, the role of HO-1 in chronic cardiac remodeling and HF is virtually unknown. Although prior studies have reported HO-1 upregulation in remodeled myocardium resulting from mechanical overload and neurohormonal stimulation,15,16 as well as in experimental and human HF,15,30 its significance is poorly understood. Specifically, no information is available on whether HO-1 is an important modulator of pathological remodeling and, if so, which mechanisms underlie such an effect. These questions are particularly important because chronic upregulation of a compensatory system may not necessarily induce the same effects seen during the acute stress response (for example, proinflammatory cytokines are acutely protective but detrimental over the long term); hence, the short-term actions of HO-1 cannot necessarily be extrapolated to a chronic setting. Moreover, the same catalytic products that are responsible for HO-1–mediated short-term cardioprotection (bilirubin, free iron, or CO) can potentially induce toxicity at inordinately high (or inappropriately sustained) levels.1
Our results establish that the upregulation of HO-1 in the failing heart ameliorates detrimental remodeling. Using genetically engineered mice, we found that sustained HO-1 expression in failing myocardium promotes neovascularization and limits oxidative stress, myofiber hypertrophy, interstitial fibrosis, apoptosis, and p53 expression. These responses were associated with improved chamber remodeling and systolic and diastolic function. The central importance of HO-1 in HF is highlighted by its ability to favorably modulate oxidative stress, neovascularization, and apoptosis, 3 key events that are known to independently influence remodeling. Our studies in HO-1 transgenic mice suggest that further enhancement of HO-1 activity beyond the level that is normally achieved in the failing heart may be a novel therapeutic strategy to prevent progression of disease. One important, although not exclusive, mechanism of these beneficial effects appears to be the moderation of apoptosis.
HO-1 Modulation of Apoptosis and Role of CO
Although HO-1 is known to inhibit apoptosis,1–3 the precise mechanisms by which this occurs remain obscure. Apoptosis in the failing heart involves multiple mechanisms that produce heightened activity of extrinsic death-receptor pathways or intrinsic mitochondrial and/or endoplasmic reticulum–linked pathways.31 Emerging evidence supports an important role for mitochondria as arbiters of cell fate in the failing heart. Intrinsic pathways converge on the mitochondria to induce mitochondrial remodeling and dysfunction, which in turn leads to the release of apoptogenic proteins such as cytochrome c into the cytosol and the activation of terminal caspase cascades.31 A central event that is thought to trigger mitochondrial dysfunction and cytochrome c release is MPT, with subsequent opening of the mitochondrial pore and mitochondrial swelling.31,32 Although in early stages, brief MPT could be protective, prolonged opening of the pore is usually considered detrimental because it triggers both necrosis and apoptosis.32 Indeed, MPT contributes to the loss of mitochondrial function observed in the failing heart,33 and LV remodeling is associated with an increase in mitochondrial pore opening in hearts both 12 and 18 weeks after coronary ligation.34 In addition, cyclosporine A, a potent blocker of MPT, improves mitochondrial function in failing cardiomyocytes.33
A key finding of our study is that HO-1 reduces the respiratory control ratio in cardiac mitochondria and confers increased resistance to prolonged MPT triggered by a variety of inducers including PAO, calcium overload, and oxidative stress (H2O2), suggesting that reduced respiration/phosphorylation coupling and improved mitochondrial membrane stability in the face of cellular stress may underlie the cardioprotective and antiapoptotic effects of HO-1 in HF. HO-1 also produced greater resistance to apoptosis and cell death in cardiomyocytes, akin to the in vivo findings in HO-1 transgenic HF. Importantly, the beneficial effects of HO-1 on respiration, MPT, and apoptosis were associated with physical approximation of HO-1 with the mitochondria and were prevented on scavenging of CO but recapitulated by CORM-3, indicating that CO plays a primary role in the modulation of mitochondrial function and apoptosis. The in vivo observations that cardiac myocyte–restricted overexpression of HO-1 also reduced nonmyocyte apoptosis in HO-1 transgenic HF further support the notion that a small, freely diffusible molecule such as CO mediates these effects. Moreover, our results in cardiomyocytes are consistent with previous studies demonstrating that CO inhibits p53 expression and mitochondrial cytochrome c release in vascular smooth muscle cells exposed to proinflammatory cytokines,35 suppresses free radical production in activated macrophages,36 and protects PC12 cells from peroxynitrite-induced apoptosis by preventing the depolarization of mitochondrial transmembrane potential.37 It is possible that HO-1–mediated inhibition of apoptosis may be related to other mechanisms in addition to CO because iron chelation with desferoxamine also mitigated the antiapoptotic effect of HO-1 to some degree. Regardless, our observations indicate that CO-dependent inhibition of MPT, attendant mitochondrial remodeling, and apoptosis can contribute to the potent antiapoptotic effect of HO-1 in the failing heart.
If CO-mediated MPT prevention is primarily responsible for the antiapoptotic and cardioprotective effect of HO-1 in HF, then bypassing the HO-1 system with pharmacological delivery of CO should achieve similar benefits. Our studies with CORM-3 support this concept because sustained delivery of exogenous CO (CORM-3) recapitulated the effects of HO-1 transgenesis on pathological LV remodeling and dysfunction, apoptosis, and p53 expression. Importantly, the CORM-3–induced remodeling responses occurred without appreciable changes in systolic pressure, indicating a direct effect of CO on the myocardium rather than an indirect effect due to changes in afterload. Many prior studies of CO-mediated effects in vivo have used CO inhalation as a means of delivery, an approach that is quantitatively difficult to standardize with regard to dosing and therefore prone to potential toxicity. Moreover, CO inhalation exerts many nonspecific and potentially untoward effects. For these reasons, we used CORM-3, a carrier of CO that releases CO in a predictable and stable manner.25 The dose that we used increased carboxyhemoglobin to ≈6% long term and was identical to that used by Motterlini and colleagues25 in a murine cardiac transplant rejection model. Further studies will be required to optimize the dose of CORM-3 for sustained CO delivery to the heart and to comprehensively determine the mechanisms underlying CORM-3–mediated improvements in LV remodeling. Nevertheless, our present findings suggest that pharmacological CO delivery may be of potential therapeutic value in HF and that feasibility of this approach should be explored in the clinical setting.
In summary, the studies reported herein reveal a novel pathophysiological role of HO-1 that was heretofore unrecognized. Our results establish that sustained HO-1 upregulation in the failing heart is an important beneficial adaptation that serves to counteract detrimental LV remodeling via antioxidant, antihypertrophic, antifibrotic, and proangiogenic effects. We have also identified an in vivo antiapoptotic action of HO-1 in the failing heart that mitigates progressive cell loss. Our in vitro analyses support the concept that this effect is related, at least in part, to CO-mediated stabilization of mitochondrial pore opening. Consistent with a central role for CO in HO-1–dependent long-term cardioprotection, exogenous CO delivery in vivo with the use of CORM-3 also improved postinfarction LV remodeling and dysfunction and reduced myocardial apoptosis. These findings have both conceptual and practical implications. From a conceptual standpoint, they reveal a new facet of the pathophysiology of HF, that is, the cardioprotective role of HO-1. From a practical standpoint, augmentation of the HO-1 axis and/or ambient CO levels should be explored as a therapeutic approach to limit pathological LV remodeling in HF.
Sources of Funding
This work was supported by National Institutes of Health grants HL-78825, ES-11860, HL-55757, HL-70897, HL-76794, and HL-55477 and by a Veterans Affairs Merit Award.
Otterbein LE, Choi AM. Heme oxygenase: colors of defense against cellular stress. Am J Physiol. 2000; 279: L1029–L1037.
Brouard S, Otterbein LE, Anrather J, Tobiasch E, Bach FH, Choi AM, Soares MP. Carbon monoxide generated by heme oxygenase 1 suppresses endothelial cell apoptosis. J Exp Med. 2000; 192: 1015–1026.
Christou H, Morita T, Hsieh CM, Koike H, Arkonac B, Perrella MA, Kourembanas S. Prevention of hypoxia-induced pulmonary hypertension by enhancement of endogenous heme oxygenase-1 in the rat. Circ Res. 2000; 86: 1224–1229.
Sato K, Balla J, Otterbein L, Smith RN, Brouard S, Lin Y, Csizmadia E, Sevigny J, Robson SC, Vercellotti G, Choi AM, Bach FH, Soares MP. Carbon monoxide generated by heme oxygenase-1 suppresses the rejection of mouse-to-rat cardiac transplants. J Immunol. 2001; 166: 4185–4194.
Poss KD, Tonegawa S. Heme oxygenase 1 is required for mammalian iron reutilization. Proc Natl Acad Sci U S A. 1997; 94: 10919–10924.
Poss KD, Tonegawa S. Reduced stress defense in heme oxygenase 1-deficient cells. Proc Natl Acad Sci U S A. 1997; 94: 10925–10930.
Yoshida T, Maulik N, Ho YS, Alam J, Das DK. H(mox-1) constitutes an adaptive response to effect antioxidant cardioprotection: a study with transgenic mice heterozygous for targeted disruption of the heme oxygenase-1 gene. Circulation. 2001; 103: 1695–1701.
Yet SF, Tian R, Layne MD, Wang ZY, Maemura K, Solovyeva M, Ith B, Melo LG, Zhang L, Ingwall JS, Dzau VJ, Lee ME, Perrella MA. Cardiac-specific expression of heme oxygenase-1 protects against ischemia and reperfusion injury in transgenic mice. Circ Res. 2001; 89: 168–173.
Melo LG, Agrawal R, Zhang L, Rezvani M, Mangi AA, Ehsan A, Griese DP, Dell'Acqua G, Mann MJ, Oyama J, Yet SF, Layne MD, Perrella MA, Dzau VJ. Gene therapy strategy for long-term myocardial protection using adeno-associated virus-mediated delivery of heme oxygenase gene. Circulation. 2002; 105: 602–607.
Liu X, Simpson JA, Brunt KR, Ward CA, Hall SR, Kinobe RT, Barrette V, Tse MY, Pang SC, Pachori AS, Dzau VJ, Ogunyankin KO, Melo LG. Preemptive heme oxygenase-1 gene delivery reveals reduced mortality and preservation of left ventricular function 1 yr after acute myocardial infarction. Am J Physiol. 2007; 293: H48–H59.
Hamid T, Gu Y, Ortines R, Bhattacharya C, Wang GW, Xuan YT, Prabhu SD. Divergent TNF receptor-related remodeling responses in heart failure: role of NF-κB and inflammatory activation. Circulation. 2009; 119: 1386–1397.
Luo J, Hill BG, Gu Y, Cai J, Srivastava S, Bhatnagar A, Prabhu SD. Mechanisms of acrolein-induced myocardial dysfunction: implications for environmental and endogenous aldehyde exposure. Am J Physiol. 2007; 293: H3673–H3684.
Li Q, Guo Y, Tan W, Ou Q, Wu WJ, Sturza D, Dawn B, Hunt G, Cui C, Bolli R. Cardioprotection afforded by inducible nitric oxide synthase gene therapy is mediated by cyclooxygenase-2 via a nuclear factor-κB dependent pathway. Circulation. 2007; 116: 1577–1584.
Srivastava S, Chandrasekar B, Gu Y, Luo J, Hamid T, Hill BG, Prabhu SD. Downregulation of CuZn-superoxide dismutase contributes to β-adrenergic receptor-mediated oxidative stress in the heart. Cardiovasc Res. 2007; 74: 445–455.
Srivastava S, Vladykovskaya E, Barski OA, Spite M, Kaiserova K, Petrash JM, Chung SS, Hunt G, Dawn B, Bhatnagar A. Aldose reductase protects against early atherosclerotic lesion formation in apolipoprotein E–null mice. Circ Res. 2009; 105: 793–802.
Levraut J, Iwase H, Shao ZH, Vanden Hoek TL, Schumacker PT. Cell death during ischemia: relationship to mitochondrial depolarization and ROS generation. Am J Physiol. 2003; 284: H549–H558.
van Empel VP, Bertrand AT, Hofstra L, Crijns HJ, Doevendans PA, De Windt LJ. Myocyte apoptosis in heart failure. Cardiovasc Res. 2005; 67: 21–29.
Liu XM, Chapman GB, Peyton KJ, Schafer AI, Durante W. Carbon monoxide inhibits apoptosis in vascular smooth muscle cells. Cardiovasc Res. 2002; 55: 396–405.
Heme oxygenase-1 (HO-1), which degrades heme to biliverdin, ferrous iron, and carbon monoxide (CO), is rapidly inducible and cardioprotective during acute stress. However, its pathophysiological role in chronic heart failure (HF) is unknown. Using myocyte-restricted HO-1 transgenic mice, we evaluated whether HO-1 upregulation in the failing heart is a beneficial adaptation that alleviates pathological remodeling. Nontransgenic and HO-1 transgenic mice underwent either sham operation or permanent left coronary ligation to induce HF. After 4 weeks, compared with nontransgenic mice with HF, HO-1 transgenic mice with HF exhibited improved postinfarction survival and significantly less left ventricular dilatation and dysfunction, cardiac hypertrophy, fibrosis, and oxidative stress, together with improved tissue neovascularization and reduced myocardial p53 expression and apoptosis. In isolated mitochondria, mitochondrial permeability transition was inhibited by HO-1 in a CO-dependent manner and was recapitulated by the CO donor tricarbonylchloro(glycinato)ruthenium(II) (CORM-3). HO-1–derived CO also prevented H2O2-induced cardiomyocyte apoptosis and cell death. Finally, sustained in vivo treatment with CORM-3 alleviated postinfarction left ventricular remodeling, p53 expression, and apoptosis in wild-type mice. Our results establish that sustained HO-1 upregulation in HF is an important beneficial adaptation that serves to counteract detrimental left ventricular remodeling via antioxidant, antihypertrophic, antifibrotic, and proangiogenic effects. We have also identified an in vivo antiapoptotic action of HO-1 in the failing heart that is related, at least in part, to CO-mediated stabilization of mitochondrial pore opening. Therefore, augmentation of HO-1 or its product, CO, should be explored as a therapeutic approach to limit pathological left ventricular remodeling and myocyte loss in HF.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109.905471/DC1.