|Home | About | Journals | Submit | Contact Us | Français|
Cardiac adaptation to aerobic exercise training includes improved cardiomyocyte contractility and calcium handling. Our objective was to determine whether cytosolic calcium/calmodulin-dependent kinase II and its downstream targets are modulated by exercise training. A six-week aerobic interval training program by treadmill running increased maximal oxygen uptake by 35% in adult mice, whereupon left ventricular cardiomyocyte function was studied and myocardial tissue samples were used for biochemical analysis. Cardiomyocytes from trained mice had enhanced contractility and faster relaxation rates, which coincided with larger amplitude and faster decay of the calcium transient, but not increased peak systolic calcium levels. These changes were associated with reduced phospholamban expression relative to sarcoplasmic reticulum calcium ATPase, and constitutively increased phosphorylation of phospholamban at the threonine 17, but not at the serine 16 site. Calcium-calmodulin-dependent kinase IIδ phosphorylation was increased at Threonine 287, indicating activation. To investigate the physiological role of calcium/calmodulin-dependent kinase IIδ phosphorylation, this kinase was blocked specifically by autocamtide-2 related inhibitory peptide II. This maneuver completely abolished training-induced improvements of cardiomyocyte contractility and calcium handling, and blunted, but did not completely abolish the training-induced increase in Ca2+ sensitivity. Also, inhibition of calcium/calmodulin-dependent kinase II reduced the greater frequency-dependent acceleration of relaxation that was observed after aerobic interval training. These observations indicate that calcium/calmodulin-dependent kinase IIδ contributes significantly to the functional adaptation of the cardiomyocyte to regular exercise training.
We have previously shown that aerobic capacity is closely related to cardiomyocyte contractility and calcium handling in training [1,2], detraining , post-infarction heart failure  and genetic selection of high versus low exercise capacity . It has also been demonstrated that exercise-induced improvement of contractile function in cardiomyocytes may be linked to increased activity of the cardiac sarcoplasmic reticulum Ca2+ ATPase 2a (SERCA-2a) both in healthy rats [2,6] and in rats with post-infarction heart failure . Uptake of Ca2+ into sarcoplasmic reticulum (SR) by SERCA-2a is an important determinant of the rate of cardiac muscle relaxation. It also determines the Ca2+ loading of the SR and thus the amount of Ca2+ available for release during cardiomyocyte contraction . SERCA-2a is regulated by phospholamban (PLB); unphosphorylated PLB binds to SERCA-2a and inhibits its activity, whereas phosphorylation removes PLB from SERCA-2a and increases Ca2+ uptake rate. Mainly, PLB is phosphorylated by cAMP-dependent protein kinase A (PKA) at serine (Ser)-16 and by Ca2+/calmodulin dependent kinase II (CaMKII) at threonine (Thr)-17 . There are also indications of a PLB-independent action of CaMKII upon SERCA-2a that appears to increase diastolic SR Ca2+ re-uptake .
It is widely accepted that Ser-16 phosphorylation is important for β-adrenergic modulation of cardiac contractility and relaxation, and that CaMKII-mediated Thr-17 PLB phosphorylation increases at high heart rates . Furthermore, CaMKII has been functionally implicated in the mediation of frequency-dependent acceleration of relaxation (FDAR) that contributes to optimal diastolic ventricular filling at increased heart rates , even in PLB-knockout mice .
The purpose of this study was to investigate whether CaMK-mediated changes in Ca2+ signaling contribute to the exercise-mediated effects on inotropy and lusitropy in mice cardiomyocytes.
In total, 62 female C57BL/6J mice (Møllegaards Breeding Center, Lille Skensved, Denmark), 19-21 g and 8 weeks old at inclusion were randomized to sedentary control or regular aerobic interval training on an inclined (25°) treadmill for 6 weeks, 5 days/week, 1.5 hours/day. After a 20 minute warm-up at 50-60% of maximal oxygen uptake (VO2max), the mice alternated between 8 and 2 min intervals at 85-90% and 50-60% of VO2max, respectively. Thus, since the treadmills were quickly accelerated or decelerated to the preferred speed, accumulated high-intensity running was performed for 56 minutes. VO2max was assessed during exhaustive treadmill running in a custom-made metabolic chamber, before and after the training period, and at the start of every training week, to adjust relative intensity in the training group. Animals were sacrificed 24 hours after the last exercise bout and 1 week after the last VO2max test, to avoid effects of acute exercise. Procedures for handling animals, as well as training and testing VO2max were previously published . The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Experimental protocols were approved by the Norwegian Animal Research Ethics Council.
Single cardiomyocytes from the left ventricle were isolated with a Krebs-Henseleit Ca2+-free buffer, collagenase type 2, and CaCl2 added stepwise to 1.2 mmol/L, as previously described [12,13]. Fura-2/AM-loaded (2 μmol/L, Molecular Probes, Eugene, OR) myocytes were stimulated with bipolar electrical pulses at 0.5, 2, and 5 Hz on an inverted epi-fluorescence microscope (Diaphot-TMD, Nikon, Tokyo, Japan), whereupon cell shortening was recorded by video edge-detection (Model 104, Crescent Electronics, Sandy, UT) and intracellular [Ca2+] ([Ca2+]i) was probed by counting 510 nm emission after exciting with alternating 340 and 380 nm wavelengths (Photon Technology International, Lawrenceville, NJ). The ratio signal of the two excitation wavelengths was subsequently converted to [Ca2+]i by measuring Rmin and Rmax in permeabilized cells and assuming a Kd of 200 nM . During the stimulation protocol, cells were continuously perfused with HEPES buffer (1.2 μmol/L Ca2+, 37° C). In separate experiments, cells were incubated 30 min before and during stimulation protocol with autocamtide-2 related inhibitory peptide II (AIP; 10 μmol/L; Calbiochem, San Diego, CA) to target CaMK. From each condition, 10 stable, consecutive contractions at each stimulation frequency were averaged.
Frozen left ventricles from trained and sedentary mice were homogenized in lysis buffer (30 mmol/L Tris-HCl, pH 7,5, 150 mmol/L KCl, 300 mmol/L Sucrose and 10 mmol/L NaF) and centrifuged for 20 min at 8000 g at 4° C; the supernatant was further ultracentrifuged 30 min at 100,000 g at 4° C. Cytosolic extract was defined as supernatant while the pellet was considered as SR/membrane-enriched fraction .
30 μg of cytosolic and SR extracts from both trained and sedentary mice were loaded on 10% SDS-PAGE for detection of CaMKII, 15% SDS-PAGE was used for PLB detection. Proteins were blotted on nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA) and membranes were blocked with TBS-T/milk (50 mmol/L Tris-HCl pH 7.4, 150 mmol/L NaCl, 0.1% v/v Tween-20, 5% w/v non-fat dried milk) for 1 hour at room temperature, followed by 1 hour room temperature incubation with the following antibodies: total PLB and phospho-Thr-17-PLB antibodies (Badrilla, Leeds, UK), phosphor-Ser-16-PLB antibody (Upstate, Charlottesville, VA), phospho-Thr-287-CaMKII (Affinity Bioreagents, Golden, CO) and total CaMKII, kindly provided by Prof. Harold A. Singer (Albany Medical College, NY). As secondary antibodies, horseradish peroxidase-conjugate goat anti-rabbit and anti-mouse were used (Amersham, Buckinghamshire, UK). Membranes were developed using the enhanced chemiluminescence (ECL) detection system (Amersham).
Data are presented as mean±SD, with significance level p<0.05. Comparisons were made using ANOVA with Scheffe post-hoc tests.
Aerobic interval training induced a marked adaptive response in aerobic capacity and cardiomyocyte contractile function and Ca2+ handling. VO2max increased by 35% (Figure 1A), and isolated cardiomyocytes were ~20% longer and ~30% wider in trained mice than in sedentary controls (117±12 vs. 98±10 and 27±6 vs. 21±4 μm cell length and width for trained and sedentary mice, respectively; p<0.01 for both comparisons). Typical examples of cardiomyocyte shortening and the Ca2+ transient are shown in Figure 1B and C, respectively. Fractional shortening and Ca2+ transient amplitude were not significantly different between trained and sedentary groups at 0.5 Hz stimulation, but at physiological contraction rates (2-5 Hz), fractional shortening and Ca2+ transient amplitudes were larger in trained mice; at 5 Hz stimulation, fractional shortening and Ca2+ transient amplitude were 60% and 50% larger, respectively, after aerobic interval training (Figure 1D-G). The larger Ca2+ transient amplitude occurred despite reduced systolic and diastolic Ca2+ levels; the reduction in diastolic Ca2+ level was greater than in systolic Ca2+ level (example traces on Figure 1C, and Figure 1F). However, diastolic shortening during the stimulation protocol was not different between cardiomyocytes from sedentary and trained mice (Figure 1G). Since the amplitude of the fractional shortening increased with training, but peak systolic [Ca2+]i did not, it indicates that myofilament sensitivity to Ca2+ has shifted with aerobic interval training. Therefore, we plotted fractional shortening during systole (corrected for diastolic shortening during increased stimulation frequencies) to peak systolic Ca2+ in each cardiomyocyte, and found that this index of Ca2+ sensitivity increased by ~20% after training (Figure 3C). Moreover, increased Ca2+ sensitivity was also indicated during diastole (Figure 3D).
SERCA-2a protein levels in SR extracts increased by ~25% after aerobic interval training (Figure 2A), whereas PLB remained unchanged (Figure 2B). This alone reduces PLB-to-SERCA-2a protein ratio, such that SERCA-2a is likely to be less inhibited in hearts from exercise trained mice. In hearts from trained mice, phosphorylation was selectively elevated at Thr-17 (Figure 2C), whereas phosphorylation at Ser-16 remained unchanged (Figure 2D). Hearts from trained mice also had elevated phosphorylation at Thr-287 of CaMKIIδ (the main cardiac CaMK isoform), while total protein amount was unchanged (Figure 2E). Since CaMKII is known to phosphorylate Thr-17 PLB , these observations suggest that adaptation to training includes elevated SERCA-2a activity and improved inotropy and lusitropy because of CaMKII activation in cardiomyocytes from trained mice.
To further investigate the role of CaMKII in mediating the exercise training-induced adaptive improvement in contractility and Ca2+ handling, we also studied the response of cardiomyocytes to CaMKII blockade. In isolated cardiomyocytes, the training-induced increase in fractional shortening was completely blunted after pre-incubation with AIP (CaMKII inhibitor), whereas no change occurred in cells from sedentary mice (Figure 3A). Parallel changes occurred in the Ca2+ transient; in trained cardiomyocytes, acute treatment with the same kinase inhibitor reduced amplitude to a level similar to controls (Figure 3B). When plotting systolic shortening to peak systolic Ca2+ (Figure 3C) and maximal diastolic shortening to diastolic Ca2+ (Figure 3D), it was revealed that inhibition of CaMKII blunted, but did not completely remove the training-induced improvement on Ca2+ sensitivity during both systole and diastole, respectively. No effects were observed in cardiomyocytes from sedentary mice.
As expected, time to 50% cell re-lengthening decreased with increasing stimulation frequencies in both trained and control cardiomyocytes. No difference between groups was observed at 0.5 Hz whereas time to re-lengthening was shorter in trained cardiomyocytes at 2 and 5 Hz (p<0.01). The difference in re-lengthening between sedentary and exercised groups was larger at 5 Hz than at 2 Hz (p<0.03). Interestingly, the frequency-dependent acceleration of relaxation (FDAR) with increasing stimulation frequencies, as well as the differences between trained and untrained hearts, were absent after pre-incubating cells with AIP (Figure 3E).
Changes in time to 50% Ca2+ decay paralleled those in time to 50% re-lengthening. Time to 50% decay of the Ca2+ transient decreased with increasing stimulation frequencies in both trained and control cardiomyocytes. No difference between groups was observed at 0.5 Hz, but time to Ca2+ decay was shorter in trained cardiomyocytes at 2 and 5 Hz (p<0.01, Figure 3F). Additionally, the difference in Ca2+ decay between trained and untrained hearts was larger at 5 than at 2 Hz (p<0.01). The changes in Ca2+ handling disappeared after CaMK inhibition (Figure 3D & E). These findings indicate that a major component of the training-induced increase in cardiomyocyte contractility is related to changes in Ca2+ handling, specifically in CaMKII-mediated regulation of SERCA-2a activity.
This is the first study to characterize the effects of regular aerobic interval training on Ca2+ signaling in mouse hearts. The main finding was that training reduces the PLB-to-SERCA-2a protein expression ratio and constitutively increases phosphorylation of cytosolic Thr-287 CaMKIIδ and its downstream target Thr-17 PLB in cardiomyocytes. The physiological importance of CaMK for exercise training-induced improvements in inotropy and lusitropy was confirmed in electrically stimulated cardiomyocytes.
Cardiac output is a major determinant of oxygen transport and aerobic metabolism; hence, increasing cardiac output is an essential adaptation to aerobic endurance training . The cellular basis for this involves improved contractile function of the cardiomyocyte, which we have demonstrated occurs with exercise training in rats [1-3]. The present study confirms that these changes occur in cardiomyocytes isolated from exercise trained mice, and identify sub-cellular mechanisms that contribute to improved inotropy and lusitropy.
Aerobic interval training increased protein expression of SERCA-2a without changing total PLB protein levels, and thereby reduced the PLB/SERCA-2a ratio. Since PLB is the primary inhibitor of SERCA-2a [8,16], the reduced ratio will increase the Ca2+ sensitivity and the enzyme activity (Vmax) of SERCA-2a. Increased SERCA-2a mRNA  and protein  expressions have previously been reported in exercise trained rats. Further increases in the Ca2+ sensitivity (reduced affinity for Ca2+, Km) of SERCA-2a would result from the increased phosphorylation of the Thr-17 of PLB [16,18], although it has also been suggested that Thr-17 phosphorylation of PLB also may increase the Vmax of SERCA-2a [19,20].
The cause of the increase in phosphorylation status is not clear, but the current study observed an increase in the phosphorylation of the Thr-287 residue of cytosolic CaMKIIδ, while total CaMKIIδ protein expression was unchanged. Since CaMKII is known to specifically phosphorylate Thr-17 PLB, these data suggest a causal link. This hypothesis was supported by the effect of the specific CaMK-inhibitor AIP, which was able to substantially blunt the exercise training-induced increase in fractional shortening and Ca2+ transient amplitude. Previous work has shown that CaMKII has a series of targets in the cell, including the L-type Ca2+ channel, the ryanodine receptor 2 (RyR2), and SERCA-2a [8,9]. Hence, inhibiting CaMK with AIP may have limited the inotropic response of exercise by acting on a series of sites in the excitation-contraction coupling process that collectively abolished the training-induced effects. In contrast to Thr-17 PLB, the Ser-16 residue (PKA site) does not appear to be modulated by aerobic interval training.
Improved myofilament sensitivity to Ca2+ is an important adaptation of the rat heart to regular exercise training [1-3,21], and is now also indicated in the mouse heart. The CaMK-blockade maneuver in trained cardiomyocytes caused Ca2+ sensitivity to revert toward levels closer to sedentary controls; hence, suggesting that CaMK partly modulates the training-induced improvement in Ca2+ sensitivity. A proposed mechanism for improved Ca2+ sensitivity has been that troponins T  and I  undergo a shift in the isoforms expressed and that atrial myosin light chain 1 increases expression levels . It has also been shown that CaMKII can act to mediate troponin I function  and the promoter-region of the atrial myosin light chain-1 gene . It now appears that CaMKII may be able to acutely induce a shift in the myofilament sensitivity to Ca2+ to a more energetically favorable state after exercise training, although other studies have suggested that PKA mediates Ca2+ sensitivity . This is conceivable also in the current study, since CaMKII-blockade only partially blunted the training-induced increase in Ca2+ sensitivity.
Aerobic interval training-induced improvement of diastolic function was demonstrated by the faster rates of cardiomyocyte re-lengthening and decay of the Ca2+ transient, in line with exercise trained rats [1-3]. These effects were partly dependent upon CaMKII being available in the cytosol, as CaMK-inhibition by AIP partly removed the exercise training-induced effects when cells were paced at frequencies as close to physiological mouse heart rates as experimentally possible. Given the proposed effect on Ca2+ re-sequestration (see above), CaMKIIδ and its downstream targets appear to enhance cardiomyocyte SR Ca2+ uptake after exercise training.
FDAR ensures appropriate ventricular filling at high heart rates, and results from accelerated SR Ca2+ uptake independent of Ca2+ removal out of the cell . However, the contributions of PLB and CaMKII to FDAR remain controversial . No differences were evident between sedentary and exercise trained for re-lengthening and Ca2+ decay rates at low stimulation frequencies, but a clear training-induced acceleration of relaxation was clear at high stimulation frequencies. The fact that CaMK-inhibition abolished this training-induced effect, suggests that CaMK-dependent Thr-17 PLB phosphorylation regulates FDAR and that this is mediated in response to regular exercise training. Our hypothesis is strengthened by reports suggesting that CaMKII-mediated phosphorylation during low stimulation frequencies is small, but increases substantially when stimulation frequencies increase , and that Thr-17, but not Ser-16, is the phosphorylation residue on PLB that mediates FDAR , in line with our results. Since Thr-287 CaMKII and Thr-17 PLB constitutively increased phosphorylation status, as observed by the immunoblots 24 hours after the last exercise bout, it suggests that physiological heart rates or equivalent stimulation frequencies are necessary to elicit the CaMKII-mediated effect. However, FDAR is robustly present in PLB-knockout mice, where it remains sensitive to CaMKII inhibitors [11,29], suggesting that other regulatory targets may be mediated by CaMK; perhaps compensatory to loss of PLB function, whereas other studies did not detect any effects of CaMK inhibitors on FDAR in rat cardiomyocytes . Hence, the mechanistic understanding of FDAR is incomplete.
We have extended previous studies of exercise training-induced changes of cardiomyocyte contractility and Ca2+ handling in rats to mice. Improved inotropy and lusitropy appear as general adaptations to exercise training in cardiomyocytes across species. These changes in cardiomyocyte contractility in response to exercise training appear to be partly caused by improved SR Ca2+ uptake via a combination of reduced PLB/SERCA-2a protein ratios and by phosphorylation of Thr-17 PLB. Increased myofilament sensitivity also contributes to the final positive inotropic response. Since the Thr-287 residue of cytosolic CaMKIIδ increased phosphorylation upstream of Thr-17 PLB, it suggests a regulatory pathway modulated by exercise training. Acute inhibition of CaMK abolished the training-induced increase in fractional shortening by reversing the effects on intracellular Ca2+ handling and myofilament Ca2+ sensitivity.
This study was supported by grants from the Norwegian University of Science and Technology (OJK, OE, UW), the Norwegian Council on Cardiovascular Diseases (OE, UW), the Norwegian Research Council (funding for sabbatical research at UCSD for OE), the St Olavs Hospital Fund for Heart Research (OJK, OE, UW), the British Heart Foundation (GLS), the European Union FP6 grant LSHM -CT-2005-018833, EUGeneHeart (GC), National Institutes of Health (HL078797-01A1, GC), the Italian Ministry of Health and Education (GC) and the following charity foundations: Torstein Erbo, Blix Family, EWS, Hoel, Agnes Sars, Anders Nordheim, and Sigrid Wolmar (OJK, OE, UW).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.