It has been reported that exercise after myocardial infarction (MI) attenuates left ventricular (LV) pump dysfunction by normalization of myofilament function. This benefit could be due to an exercise-induced upregulation of endothelial nitric oxide synthase (eNOS) expression and activity. Consequently, we first tested the hypothesis that the effects of exercise after MI can be mimicked by elevated eNOS expression using transgenic mice with overexpression of human eNOS (eNOSTg). Both exercise and eNOSTg attenuated LV remodeling and dysfunction after MI in mice and improved cardiomyocyte maximal force development (Fmax). However, only exercise training restored myofilament Ca2+-sensitivity and sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA)2a protein levels and improved the first derivative of LV pressure at 30 mmHg. Conversely, only eNOSTg improved survival. In view of these partly complementary actions, we subsequently tested the hypothesis that combining exercise and eNOSTg would provide additional protection against LV remodeling and dysfunction after MI. Unexpectedly, the combination of exercise and eNOSTg abolished the beneficial effects on LV remodeling and dysfunction of either treatment alone. The latter was likely due to perturbations in Ca2+ homeostasis, as myofilament Fmax actually increased despite marked reductions in the phosphorylation status of several myofilament proteins, whereas the exercise-induced increases in SERCA2a protein levels were lost in eNOSTg mice. Antioxidant treatment with N-acetylcysteine or supplementation of tetrahydrobiopterin and l-arginine prevented these detrimental effects on LV function while partly restoring the phosphorylation status of myofilament proteins and further enhancing myofilament Fmax. In conclusion, the combination of exercise and elevated eNOS expression abolished the cardioprotective effects of either treatment alone after MI, which appeared to be, at least in part, the result of increased oxidative stress secondary to eNOS “uncoupling.”
- left ventricular remodeling
- endothelial nitric oxide synthase overexpression
left ventricular (LV) remodeling after myocardial infarction (MI) is a compensatory mechanism that serves to restore LV pump function. Despite the apparent appropriateness of LV remodeling to maintain cardiac pump function early after MI, remodeling is an independent risk factor for the development of congestive heart failure (21). The mechanism underlying the progression from LV remodeling to overt heart failure remains incompletely understood. Recently, we (11) reported that exercise training (exercise) after MI attenuates LV pump dysfunction by normalization of myofilament function in mice. The beneficial effects of exercise on cardiac function can be attributed, at least in part, to an exercise-induced upregulation of endothelial nitric oxide (NO) synthase (eNOS) expression and activity (35), which is the major isoform of NOS that is expressed in the coronary vascular endothelium and cardiac myocytes (24). For example, an increase in NO bioavailability has been shown to improve many of the processes perturbed in LV remodeling, including angiogenesis (28), cardiac fibrosis (20), and hypertrophy (33). Indeed, Jones et al. (18) reported that elevated expression of the human eNOS gene in the vascular endothelium in transgenic mice (eNOSTg) not only improved survival but also blunted LV dysfunction and pulmonary congestion after MI. These beneficial effects were ascribed to a reduction in afterload, secondary to a lower peripheral vascular resistance (40), thereby facilitating LV pump function, whereas global LV contractility was unchanged (18). In contrast, exercise improved global LV contractility with no beneficial effect on survival (11). These findings suggest that the beneficial effects of exercise and eNOSTg are only in part overlapping and thus partly complementary.
In view of the complementary mechanisms by which exercise and eNOSTg appear to exert beneficial effects, we tested the hypothesis that combined treatment with exercise and eNOSTg exerts an added benefit and investigated the underlying mechanisms. Contrary to our hypothesis, the results unexpectedly indicated that exercise and eNOSTg abolished the beneficial effects of either treatment alone on global cardiac function after MI. Since a shortage of substrate (l-arginine) and/or cofactors [tetrahydrobiopterin (BH4)] of eNOS can result in “uncoupling” of NOS, thereby generating ROS such as superoxide rather than NO (29), we subsequently tested the hypothesis that the detrimental effects of combined exercise training and eNOSTg are the result of increased oxidative stress due to eNOS uncoupling.
MATERIALS AND METHODS
The experiments complied with National Institutes of Health Guide for Care and Use of Laboratory Animals (NIH Pub. No. 85-23, Revised 1996) and were approved by the Erasmus University Medical Center Animal Care Committee.
A total of 159 eNOSTg mice and 223 wild-type (Wt) littermates in the C57Bl/6J background of either sex (∼12 wk old) entered the study and were randomly assigned to one of eight experimental groups. Data from 147 Wt littermates have been previously reported (11). Sham-operated mice (SH) and mice with MI were housed and were kept sedentary (SHSED and MISED, respectively) or subjected to voluntary exercise training by wheel running (SHEX and MIEX, respectively) for 8 wk. Treadmills were custom built to allow electronic measurement of the distance run by the mice. Voluntary exercise was the training of choice due to its ability to achieve uniform controlled exercise in mice and to minimize stress factors, which are present during forced running and particularly during swimming (5). C57Bl/6J mice were used because they are known to perform excellently in voluntary wheel running (5). The generation of eNOSTg mice has been previously described (43, 44). Briefly, a DNA fragment containing the human eNOS gene (∼6 kb of the 5′-natural flanking sequence, including the native eNOS promoter, and ∼3 kb of the 3′-sequence) was used for microinjection of fertilized oocytes. Transgenic offspring were backcrossed to C57Bl/6J mice for >10 generations. In three additional groups of eNOSTg-MIEX mice, the antioxidant N-acetylcysteine (NAC; n = 11, Sigma) or l-arginine (Sigma) without (n = 10) or with (n = 5) BH4 (Schircks Laboratories) was supplemented either in drinking water (1% NAC or 2.5% l-arginine) or administered by gavage (BH4: 80 mg/kg administered 3 times/wk) over the entire 8-wk followup period.
Mice were weighed, sedated with 4% isoflurane, intubated, and pressure-controlled ventilated with O2-N2O [1:2 (vol/vol)] containing ∼2.5% isoflurane for anesthesia. MI was produced by permanent ligation of the left anterior descending coronary artery (LAD) with a 7-0 silk suture (Aesculap) as previously described (11, 41). SH animals underwent the operation without infarct induction. Eight weeks after mice entered the study, hemodynamic measurements were performed under anesthesia as previously described (11). In brief, short- and long-axis views from M-mode LV echocardiography (ALOKA, Prosound SSD-4000) were performed, LV diameters at end diastole (LVEDD) and end systole were measured, and fractional shortening was calculated. A 1.4-Fr microtipped pressure transducer catheter (SPR-671, Millar Instruments, Houston, TX) was inserted into the LV, and pressure-diameter relations were obtained from M-mode images synchronized with LV pressure by simultaneous ECG recording. At the conclusion of each experiment, right ventricular (RV) weight (RVW) and LV weight (LVW), tibia length (TL), and lung fluid weight were determined. Masson's trichrome staining was used for the analysis of the LV collagen volume fraction in noninfarcted LV samples as previously described (11). Briefly, four fields were randomly selected in two sections of eight mice per group and photographed using an Olympus BH 20 microscope (Olympus) at a magnification of ×400. Within each field, segments representing connective and muscle tissue were identified and manually traced with a digitizing pad and computer image-analysis software (Clemex Vision PE 3.5) to calculate the traced area. The collagen volume fraction was calculated in each field as the sum of all connective tissue areas divided by the sum of all connective tissue and muscle areas and then averaged for each animal per group.
Western blot analysis.
Frozen LV tissue samples (n = 5 mice/group) were homogenized, and protein concentrations were determined as previously described (11). Blots were preincubated in Odyssey blocking buffer (LI-COR Biosciences) and incubated with diluted primary antibodies in blocking buffer containing 0.1% Tween 20. As previously described (11), primary antibodies against phospholamban (PLB), sarco(endo)plasmic reticulum Ca2+-ATPase 2a (SERCA2a), the β1-adrenergic receptor, and phosphorylated Ser16- or phosphorylated Thr17-containing sequences of PLB were used (11). IRDye 800CW- or IRDye 680-conjugated goat anti-rabbit or goat anti-mouse secondary antibodies (LI-COR Biosciences) were used. For SERCA2a, PLB, and phosphorylated PLB detection, a rat cardiac membrane preparation was used as the positive control. For β1-adrenergic receptor detection, a rat brain extract was used as the positive control. Fluorescent signals were detected and quantified using the Odyssey Infrared Imaging System (LI-COR Biosciences).
Force measurements in single permeabilized cardiomyocytes.
In the noninfarcted remote area of the myocardium, isometric force was measured in single permeabilized cardiomyocytes of 5 mice/group at different Ca2+ concentrations and a sarcomere length of 2.2 μm as previously described (11, 42). In brief, mechanically isolated myocytes were permeabilized in relaxing solution containing 0.5% Triton X-100 (5 min), a treatment that also removes soluble and membrane-bound kinases and phosphatases, which may alter the phosphorylation status of myofibrillar proteins. To remove Triton X-100, cells were washed twice in relaxing solution. Subsequently, a single myocyte was attached between a force transducer and a piezoelectric motor. After a complete force-pCa series had been obtained, myocytes were incubated in relaxing solution containing the exogenous catalytic subunit of PKA, and a second force-pCa series was obtained.
Myofilament protein phosphorylation status.
To determine the phosphorylation status of myofilament proteins [myosin-binding protein C (MyBP-C), troponin T (TnT), troponin I (TnI), myosin light chain 2 (MLC-2), desmin, and tropomyosin], LV samples (n = 6 mice/group) were separated on gradient gels and stained with Pro-Q Diamond phosphoprotein gel stain in conjunction with SYPRO ruby staining of the gels as previously described (11, 51). Phosphorylation signals for myofilament proteins were normalized to the intensities of SYPRO ruby-stained MyBP-C bands and analyzed using the luminescent image analyzer las-3000 and Aida image analyzer.
Superoxide anion generation in the noninfarcted remote myocardium was determined using dihydroethidium (DHE) fluorescence (30). Hearts from 4 mice/group were excised, and LV tissue samples were washed in ice-cold saline, embedded in Tissue-Tek, frozen in liquid nitrogen-cooled isopentane, and stored at −80°C. Tissue sections of 5 μm were cut using a cryostat and stained with 10 μM DHE at 37°C for 30 min. Fluorescent images were obtained using an inverted fluorescence microscope system (Axiovert S100, Zeiss) with a digital camera (Axiocam, Zeiss) and acquisition software (Axiovision, Zeiss) through a ×20 objective (LD ACHROPLAN, 0.40 korr, Ph2, Zeiss) and filterset 09 (no.487909-0000, exitation: 450- to 490-nm band-pass filter and emission: >515-nm long-pass filter). The generation of superoxide was demonstrated by red fluorescent labeling. Images were analyzed on an offline microscopy image-analysis software system (Clemex Technologies) to quantify the amount of fluorescence.
Data were analyzed using three-way ANOVA (genotype × MI × exercise) followed by post hoc testing with Student-Newman-Keuls test when appropriate. Survival was analyzed by the Kaplan-Meier method and log-rank (Mantel-Cox) test. Significance was accepted when P < 0.05. Data are means ± SE.
Survival and exercise.
Wt-MISED mice demonstrated 60% survival compared with Wt-SHSED mice (Fig. 1A), which significantly improved to 78% in eNOSTg-MISED mice. In contrast, exercise had no significant effect on survival in Wt-MI mice and even tended to increase mortality in eNOSTg-MI mice (P = 0.063). Both Wt-MIEX and eNOSTg-MIEX mice initially ran shorter distances per day compared with their corresponding SH groups (Fig. 1B), but neither MI nor eNOSTg significantly influenced total running distance over the 8-wk period (Fig. 1C).
LV remodeling and global LV function.
eNOSTg-SHSED mice exhibited slightly lower mean aortic pressure and LV systolic pressure levels but had normal global LV function compared with Wt-SHSED mice (Table 1). eNOSTg-SH mice exhibited a slightly but not significant larger LVEDD, as reflected by the small rightward shift of the LV pressure-diameter relation, compared with the corresponding Wt-SH groups (Fig. 2 and Table 1), but LVW, RVW, and lung fluid weights were virtually identical (Table 1 and Fig. 3). Exercise did not affect LVW, RVW, or lung fluid weight in either Wt-SH or eNOSTg-SH mice, although it reduced LV interstitial collagen content and LVEDD only in Wt-SH mice. Similarly, exercise had minimal effects on LV systolic and diastolic function in either Wt-SH or eNOSTg-SH mice, although levels of LV fractional shortening and LV dP/dtmin were slightly lower in eNOSTg-SHEX mice compared with Wt-SHEX mice (Fig. 3 and Table 1).
eNOSTg had no effect on infarct size measured 24 h after ligation (39 ± 2% of the total LV area in 6 eNOSTg-MISED mice vs. 43 ± 3% of the total LV area in 8 Wt-MISED mice). Also, there was no effect of elevated eNOS expression or exercise on infarct size 8 wk after LAD ligation as the infarct area was 18 ± 1%, 20 ± 1%, 19 ± 1%, and 18 ± 1% of the LV total area in Wt-MISED, Wt-MIEX, eNOSTg-MISED, and eNOSTg-MIEX mice, respectively. MI resulted in marked LV dilation (Fig. 2 and Table 1), LV hypertrophy, pulmonary congestion (increase in lung fluid weight), RV hypertrophy, and increased collagen content of the remote surviving LV myocardium in Wt-MISED mice (Fig. 3), which was associated with depressed systolic and diastolic LV function (Fig. 3 and Table 1). LV remodeling and dysfunction were also observed in eNOSTg-MISED mice. However, compared with Wt-MISED mice, LV diameter and collagen content remained significantly lower, whereas LV fractional shortening remained significantly higher, and pulmonary congestion was prevented in eNOSTg-MISED mice. Interestingly, the first derivative of LV pressure at 30 mmHg (LV dP/dtP30), an index of global LV contractility, was slightly lower than in Wt-MISED mice, suggesting that the improved LV fractional shortening was the result of the lower aortic blood pressure in eNOSTg-MISED mice rather than the result of improved LV contractility. Similar to eNOSTg, exercise had no effect on MI-induced LV hypertrophy but blunted LV dilation, interstitial fibrosis, pulmonary congestion, and RV hypertrophy and improved fractional shortening. Unlike eNOSTg, exercise had no effect on aortic blood pressure but improved LV dP/dtP30. Despite these complementary actions, we unexpectedly observed that exercise of eNOSTg-MI mice aggravated LV remodeling, interstitial fibrosis, LV dysfunction, pulmonary congestion, and RV hypertrophy (Table 1 and Figs. 2 and 3).
Western blot analysis.
Protein levels of β1-adrenoceptors and SERCA2a were significantly lower in eNOSTg-SH mice compared with Wt-SH mice, whereas PLB and phosphorylation at either its Ser16 or Thr17 site were not significantly different (Fig. 4). Exercise training had minimal effects on β1-adrenoceptors, SERCA2a, PLB, the SERCA2a-to-PLB ratio, and PLB phosphorylation at the Thr17 site in both Wt-SH and eNOSTg-SH mice but produced a small increase in PLB phosphorylation at the Ser16 site in Wt-SH mice.
MI resulted in a significant decrease in protein levels of β1-adrenoceptors and SERCA2a, with no effect on PLB, so the SERCA2a-to-PLB ratio was lower in the remote surviving myocardium of Wt-MISED mice compared with Wt-SHSED mice (Fig. 4). Furthermore, PLB phosphorylation at the Ser16 site increased, whereas phosphorylation at the Thr17 site decreased. In eNOSTgSED mice, MI did not produce further alterations in β1-adrenoceptors and SERCA2a levels and had no effect on PLB or its phosphorylation status. Exercise in Wt-MI mice normalized levels of β1-adrenoceptors and SERCA2a, with no effect on PLB or its phosphorylation. In contrast, exercise in eNOSTg-MI mice had no beneficial effect on Ca2+-regulatory protein levels.
Force development in single permeabilized cardiomyocytes.
Passive force was significantly lower in eNOSTg-SHSED (2.0 ± 0.3 kN/m2) compared with Wt-SHSED (3.0 ± 0.4 kN/m2) mice, but Fmax and myofilament Ca2+ sensitivity (pCa50) were similar in both groups (Fig. 5). Exercise had no effect on Fmax or Ca2+ sensitivity in Wt-SH mice but increased Fmax in eNOSTg-SH mice. Treatment with the catalytic subunit of PKA decreased Ca2+ sensitivity to a similar extent in all four SH groups (average change in pCa50: 0.060 ± 0.003), indicating preserved PKA signaling in eNOSTg-SHSED, Wt-SHEX, and eNOSTg-SHEX mice.
In WtSED mice, MI produced a reduction in Fmax and an increase in Ca2+ sensitivity (Fig. 5). The latter was corrected by the addition of the catalytic subunit of PKA, indicating that the increase in Ca2+ sensitivity resulted from reduced PKA signaling. eNOSTg in MISED was associated with a marked increase in Fmax but did not significantly affect Ca2+ sensitivity (P = 0.27 vs. Wt-MISED mice). Conversely, exercise in Wt-MI mice restored both Fmax and Ca2+ sensitivity to SH levels. The combination of exercise and eNOSTg in MI mice increased Fmax even further, whereas Ca2+ sensitivity decreased to levels observed in Wt-MIEX mice. PKA decreased Ca2+ sensitivity in eNOSTg-MISED and eNOSTg-MIEX mice. However, even after treatment with PKA, Ca2+ sensitivity remained higher in eNOSTg-MISED compared with Wt-MISED mice. These findings suggest that, in contrast to Wt mice, loss of PKA-mediated protein phosphorylation does not contribute to increased Ca2+ sensitivity in eNOSTg mice after MI.
Myofilament protein phosphorylation status.
There were no differences between Wt-SHSED, Wt-SHEX, eNOSTg-SHSED, and eNOSTg-SHEX mice with respect to the phosphorylation status of several myofilament proteins (Fig. 5). Also, MI did not produce alterations in the phosphorylation status of myofilament proteins in either WtSED or eNOSTgSED mice. Exercise increased MLC-2 phosphorylation in Wt-MI mice. However, the combination of exercise and eNOSTg caused marked decrements in the phosphorylation status of MyBP-C, TnI, desmin, MLC-2, and tropomyosin after MI.
DHE reacts with superoxide anions to form ethidium bromide, which, in turn, intercalates with DNA to provide nuclear fluorescence as a marker for superoxide anion generation. As shown in Fig. 6, DHE fluorescence was enhanced in the noninfarcted remote myocardium after MI compared with SH mice. Exercise normalized the MI-induced increase in DHE fluorescence. The combination of exercise and eNOSTg after MI even further increased the DHE fluorescence, which was abolished by NAC treatment.
Effect of antioxidant treatment.
Adjunctive treatment of eNOSTg-MIEX mice with the antioxidant NAC improved survival from 60% in eNOSTg-MIEX mice without NAC to 85% (data not shown; P < 0.05) and prevented LV remodeling, interstitial fibrosis, global LV dysfunction, pulmonary congestion, and RV hypertrophy (Table 2). In contrast, NAC treatment had no significant effect on LVW and LV function in Wt-SHSED, eNOSTg-SHSED, and eNOSTg-SHEX mice (data not shown). At the myofilament level, treatment of eNOSTg-MIEX mice with NAC produced further increases in Fmax, with no effect on myofilament Ca2+ sensitivity, and partially reversed the phosphorylation of myofilament protein MLC-2, tropomyosin, and desmin. NAC had no effect on MyBP-C, TnI, and TnT phosphorylation in eNOSTg-MIEX mice.
Effect of l-arginine and BH4 treatment.
To investigate whether the increased oxidative stress resulted from a shortage of eNOS substrate and/or cofactor BH4, we investigated whether l-arginine supplementation alone or together with BH4 ameliorated the detrimental effects of combining exercise and eNOS overexpression after MI. Treatment of eNOSTg-MIEX mice with l-arginine had no significant effects on LV remodeling, dysfunction, and backward failure (data not shown), suggesting that substrate deficiency was not the cause of the increased oxidative stress. Treatment of eNOSTg-MIEX mice with l-arginine/BH4 did not affect heart rate (530 ± 20 beats/min) or mean aortic pressure (65 ± 2 mmHg) but normalized LV collagen content, lung fluid weight, RVW, fractional shortening, and LV dP/dtP30 (Fig. 7) and increased survival from 59% in eNOSTg-MIEX mice to 80% in eNOSTg-MIEX mice with l-arginine/BH4 (P < 0.05). l-Arginine/BH4 had no effect on protein levels of β1-adrenoceptors but increased protein levels of SERCA2a (P = 0.056) with no effect on PLB, so that the SERCA2a-to-PLB ratio increased (P < 0.05). At the myofilament level, l-arginine/BH4 increased Fmax, Ca2+ sensitivity, and phosphorylation of myofilament proteins MyBP-C, TnT, and MLC-2, but not of TnI, desmin, or tropomyosin. These observations are consistent with the concept that deficiency of BH4 in these mice resulted in eNOS “uncoupling” and thereby increased superoxide production.
The present study investigated the impact of elevated eNOS expression and exercise training on LV remodeling and dysfunction after MI in mice at the global LV and cardiomyocyte level, either as a single or as combined therapy. The main findings were as follows: 1) in mice with MI, single therapy of either exercise or elevated eNOS expression improved global LV fractional shortening and cardiomyocyte force development and reduced collagen content and attenuated pulmonary congestion, whereas only eNOSTg improved survival; 2) the combination of elevated eNOS expression and exercise abolished the beneficial effects of either treatment alone after MI; and 3) these beneficial effects were restored in part by cotreatment with either the antioxidant NAC or the combination of l-arginine and BH4 supplementation. The implications of these findings will be discussed.
MI-induced LV dysfunction in mice.
In agreement with previous observations, we observed that permanent LAD ligation in mice resulted in significant LV remodeling 8 wk later, as characterized by LV dilation, hypertrophy, and increased collagen deposition in the remote surviving myocardium. LAD ligation also produced marked LV dysfunction, as characterized by decrements in LV pump function (fractional shortening) and indexes of global LV contractility (LV dP/dtP30) and relaxation (dP/dtmin and time constant), resulting in LV backward failure, as reflected by pulmonary edema and RV hypertrophy (11, 41, 45). In agreement with observations in swine (42) and mice (11), we found that remodeling of the noninfarcted myocardium was associated with altered myofilament function, as characterized by decreased Fmax and increased Ca2+ sensitivity of tension development in single permeabilized cardiomyocytes.
Exercise training after MI.
In agreement with a recent study from our laboratory (11), we observed that in mice with a large MI (comprising ∼43% of the LV mass), exercise does not aggravate LV remodeling, as relative LV mass and infarct geometry were unchanged, whereas exercise reduced collagen content and tended to decrease LVEDD. Furthermore, exercise attenuated LV dysfunction and ameliorated LV backward failure. At the cellular and molecular level, exercise improved myofilament function, and particularly increased Fmax, whereas protein levels of β1-adrenergic receptor (28%, P = not significant) and SERCA2a (47%, P < 0.05) increased. These Western immunoblot data contrast somewhat with a recent study from our laboratory (11) in which we used the enhanced chemiluminescence method instead of the accurate infrared-based Odyssey method and observed different degrees of increases in protein levels of the β1-adrenergic receptor (48%, P <0.05) and SERCA2a (12%, P = not significant) compared with the present study.
eNOS overexpression after MI.
NO can modulate many of the processes involved in cardiac remodeling. For example, long-term administration of nitrates (NO donor compounds) limited LV remodeling in patients with a recent MI (12a). Conversely, eNOS-deficient mice demonstrated exaggerated LV dilation and hypertrophy and lower fractional shortening and dP/dtmax 4 wk after MI, suggesting that eNOS-derived NO attenuates LV dysfunction and remodeling (34). In the present study, we investigated the effects of elevated eNOS expression on global LV function and isolated cardiomyocyte function after MI. In agreement with a study by Jones et al. (18), we found that mice overexpressing eNOS by 10- to 12-fold had improved survival, global LV pump function, LV dilation, and pulmonary edema after MI and additionally observed that eNOSTg blunted interstitial fibrosis. Conversely, eNOSTg did not attenuate cardiac hypertrophy or improve global cardiac contractility and relaxation parameters. Interestingly, all of these improvements by elevated eNOS expression occurred without significant differences in baseline LV geometry, morphology, or function in SH mice. The only difference that was noted under basal conditions was that protein levels of the β1-adrenergic receptor and SERCA2a were markedly lower. The latter findings mirror the increase in β1-adrenoceptor mRNA levels in eNOS−/− mice (1) and the increase in SERCA2a protein expression in inducible NOS−/− mice (19). Interestingly, S-nitrosylation of SERCA2a has been shown to increase its activity (37). Thus, it could be speculated that an increased excitation-contraction coupling, arising from the elevated cardiomyocyte NO levels and subsequent protein S-nitrosylation, acted to maintain normal cardiac contractile function, in the face of lower protein expression levels of β1-adrenergic receptors and SERCA2a (16).
The mechanism underlying the improved survival and cardiac function in eNOSTg mice after MI remains unclear. Based on the lack of increase in indexes of contractility or relaxation, it has been suggested that the beneficial effects of eNOSTg on the post-MI heart are principally due to an afterload reduction (18), which occurs secondary to the lower systemic vascular resistance and hence lower aortic and LV systolic pressure (40). However, cardiomyocyte-specific overexpression of eNOS has been shown to exert beneficial effects on LV structure and function after MI (17), suggesting that eNOS overexpression can improve LV function by systemic vascular as well as local myocardial effects.
To further investigate the effects of eNOSTg on cardiomyocyte function, we investigated myofilament function and myofilament protein phosphorylation status. Interestingly, we noted that similar to exercise training, elevated eNOS expression increased the Fmax of isolated permeabilized cardiomyocytes but that, in contrast to exercise, elevated eNOS expression did not correct the high myofilament Ca2+-sensitivity in MI and had no effect on MLC-2 protein phosphorylation and β1-adrenergic receptor and SERCA2a protein levels. These findings suggest that in addition to the afterload reduction, an improvement in myofilament Fmax could have contributed to the eNOSTg-induced improvement in LV pump function after MI. The mechanism by which eNOSTg increased Fmax remains to be determined, but it is of interest to note that Sun et al. (37, 38) have recently shown that S-nitrosylation of myofilament proteins, including the myofilament protein α-myosin heavy chain, MLC kinase 1, and myomesin, may be involved in the cardioprotection by ischemic preconditioning against ischemia-reperfusion-induced sarcomeric damage in mice. It remains to be investigated whether S-nitrosylation of myofilaments is altered and contributes to the enhanced myofilament contractility in our model.
The beneficial effects of exercise on cardiac function can be ascribed, at least in part, by an exercise-induced upregulation of eNOS expression and activity (35). The effects of either exercise or eNOSTg in MISED mice on LV geometry, collagen content, and function were indeed similar. However, survival in MI mice was improved by eNOSTg but not by exercise. The differing effects of exercise versus eNOS overexpression on survival are not easily explained. Figure 4 shows that protein levels of β1-adrenergic receptors and SERCA2a were lower in eNOSTg compared with WT mice, but this provides an unlikely explanation as lower levels of these proteins would be expected to result in a worse outcome rather than an improved outcome. It should be noted that mortality in C57Bl/6J mice after permanent coronary artery ligation is principally caused by cardiac rupture of the infarct area within the first 2 wk after MI (7). It is therefore tempting to speculate that the lower systemic vascular resistance and mean aortic pressure in eNOSTg compared with Wt mice (40) may have prevented cardiac rupture, thereby enhancing post-MI survival in eNOSTg mice.
Combination of exercise and eNOS overexpression: role of ROS.
Exercise training in both eNOSTg-SH and Wt-SH mice had no effect on relative LVW, RVW, or lung fluid weight, although eNOSTg-SHEX mice displayed slightly lower levels of LV systolic and diastolic function compared with Wt-SHEX mice. Furthermore, LVEDD and interstitial collagen levels were slightly elevated in eNOSTg-SHEX mice compared with Wt-SHEX mice. Exercise had no effect on the protein expression of SERCA2a and β1-adrenergic receptors, myofilament phosphorylation, or Ca2+ sensitivity of myofilaments in either eNOSTg-SH or Wt-SH mice but resulted in increased Fmax in only eNOSTg-SH mice. These findings suggest minimal alterations in LV responses to exercise training in eNOSTg-SH mice.
The complementary actions of exercise (11) and eNOSTg (18) in MI mice initially led us to hypothesize that the combination of exercise and eNOSTg could have an added benefit on LV remodeling and function after MI. However, we unexpectedly observed that combining the two treatments abolished the beneficial effects of either modality alone. LV remodeling, pump dysfunction, backward failure, and interstitial fibrosis were all aggravated compared with either treatment alone.
It could be forwarded that the slightly lower heart rate in the eNOSTg-MIEX group could have contributed to the aggravated LV dysfunction in these mice. However, regression analysis of heart rate and either LV dP/dtP30 or fractional shortening showed that the lower heart rate could not explain the aggravated dysfunction.
Despite the marked decrease in the level of phosphorylation of various myofilament proteins in eNOS-MIEX mice, myofilament force measurements revealed that pCa50 was similar and Fmax was even higher in eNOSTg-MIEX mice compared with Wt-MIEX mice. This was a surprising finding, because decreased phosphorylation levels of MyBP-C (25), MLC-2 (14), and tropomyosin (39) would all have been expected to decrease rather than increase Fmax. Similarly, the decreased phosphorylation levels of TnI and TnT (3) and MyBP-C (8) would have been expected to increase Ca2+ sensitivity, although it cannot be excluded that such an increase may have been counterbalanced by the decrease in MLC-2 phosphorylation (27, 31). Notwithstanding these surprising observations, the increase in Fmax will likely result in an increased myofilament force development at higher (systolic) Ca2+ concentrations (14, 15). Indeed, we have previously observed a decrease in Fmax in the remodeled myocardium in pigs (42) and mice (11) with a recent infarction that was associated with systolic dysfunction. Conversely, exercise training restored Fmax in mice with an infarction, which was associated with an improvement in global LV systolic function (11). These observations suggest that worsening of global LV systolic dysfunction in eNOSTg-MIEX compared with Wt-MIEX mice is not readily explained by the increase in myofilament Fmax.
Alternatively, it could be speculated that perturbations in β1-adrenergic signaling and Ca2+ homeostasis were involved in the loss of beneficial effect of exercise on LV dysfunction in eNOSTg-MI mice. This notion is supported by the different effects of exercise on β1-adrenoceptors and SERCA2a expression in eNOSTg-MIEX versus Wt-MIEX mice (Fig. 4). The exact role of alterations in Ca2+ homeostasis, including measurements of Ca2+ transients in single intact cardiomyocytes, should be the subject of future studies.
The adverse effects of combined exercise and eNOSTg on LV remodeling and function were prevented by the antioxidant NAC, suggesting a role for ROS in the adverse effects. Interestingly, phosphorylation of myofilaments was partly restored and Fmax further increased, suggesting that ROS-induced myofilament protein dephosphorylation had blunted the increase in Fmax produced by combined exercise and eNOSTg. Although the mechanism by which ROS induced dephosphorylation of myofilament proteins remains to be determined, it is of interest to note that S-nitrosylation is an important inhibitor of phosphatase activity, including protein phosphatase (PP)-1 (4, 50), so that increased oxidative stress could have resulted in increased phosphatase activity. Accordingly, heart failure is associated with increased PKC protein content and activity (47) and activated PP-1, as a result of which MLC-2 phosphorylation is decreased (6). This effect is enhanced by increased ROS via the activation of PKC (12). Only recently, Vahebi et al. (39) showed that chronic activation of p38α-MAPK in a transgenic mouse model reduced myofilament Fmax in association with decreased tropomyosin phosphorylation. As ROS have been shown to enhance p38α-MAPK signaling (36), blockade of ROS with the antioxidant NAC and the concomitant enhancement of Fmax might be explained, at least in part, by the restored tropomyosin phosphorylation. Besides these effects on myofilament function, ROS as well as NO can directly influence the function of the sarcoplasmic reticulum ryanodine receptor, i.e., the Ca2+-release channel (49), and SERCA2a (37, 46) and may thereby influence Ca2+ homeostasis.
We hypothesized that increased ROS production was due to uncoupling of eNOS, which may have resulted from a shortage of substrate and cofactors (10, 23, 26, 48), as a result of elevated eNOS expression in conjunction with exercise training (13). Indeed, the combination of BH4 and l-arginine supplementation improved survival and attenuated LV remodeling, dysfunction, and backward failure in eNOSTg-MIEX mice. At the myofilament level, l-arginine/BH4 supplementation increased Fmax, myofilament Ca2+ sensitivity and phosphorylation of myofilament proteins MyBP-C, TnT, and MLC-2. Furthermore, l-arginine/BH4 treatment resulted in a significant increase in protein levels of SERCA2a and the SERCA2a-to-PLB ratio. Taken together, our observations are consistent with the concept that reduced levels of BH4 resulted in eNOS uncoupling and thereby in elevated oxidative stress. Direct measurements of eNOS homodimers (26) could be part of future studies to confirm that eNOS uncoupling indeed occurs in eNOSTg-MIEX mice and is reversed by l-arginine/BH4 treatment (32).
The present study indicates that exercise and elevated eNOS expression as single therapies attenuated LV remodeling, interstitial fibrosis, and pulmonary congestion and improved global LV fractional shortening and cardiomyocyte force development, whereas only eNOS overexpression improved survival in mice with a large MI. The combination of exercise and eNOS overexpression abolished the beneficial effects of either treatment alone. This detrimental interaction appears to be principally the result of increased oxidative stress secondary to uncoupling of eNOS. Several investigators (9, 22) have proposed a role for eNOS gene therapy in patients with cardiovascular disease. Provided that these findings can be extended to humans, the results of the present study could be interpreted to suggest that eNOS gene therapy of LV dysfunction in combination with physical exercise (as part of lifestyle changes) may benefit from careful food supplementation with l-arginine and BH4, as this may avoid potential eNOS uncoupling and increased oxidative stress.
This work was supported by The Netherlands Heart Foundation Grant 2000T038 (to D. J. Duncker).
- Copyright © 2009 the American Physiological Society