|Home | About | Journals | Submit | Contact Us | Français|
The biochemical differences between the two mammalian cardiac myosin heavy chains (MHC), α-MHC and β-MHC, are well described, but the physiological consequences of basal isoform expression and isoform shifts in response to altered cardiac load are not clearly understood. Mature human ventricle contains primarily the β-MHC isoform. However, the α-MHC isoform can be detected in healthy human ventricle and appears to be significantly down-regulated in failing hearts. The unique biochemical properties of the α-MHC isoform might offer functional advantages in a failing heart that is expressing only the β-MHC isoform. This hypothesis cannot be tested in mice or rats as both species express α-MHC as the predominant isoform.
To test the effects of persistent α-MHC expression on the background of β-MHC, we made transgenic (TG) rabbits that expressed the rabbit α-MHC cDNA in the ventricle such that the endogenous myosin was partially replaced by the transgenically-encoded species. Molecular, histological and functional analyses showed no significant baseline effects in the TG rabbits compared to nontransgenic (NTG) littermates. To determine if α-MHC expression afforded any advantages to stressed myocardium, a cohort of TG and NTG rabbits were subjected to rapid ventricular pacing. While both the TG and NTG rabbits developed dilated cardiomyopathy, the TG rabbits had a higher shortening fraction, less septal thinning, and more normal ±dP/dt than paced NTG rabbits.
Transgenic expression of α-MHC does not have any apparent detrimental effects under basal conditions and is cardioprotective in experimental tachycardia-induced cardiomyopathy.
Two distinct cardiac myosin heavy chains (MHC), α-MHC and β-MHC, are expressed in a species-dependent manner and assemble as an αα-MHC homodimer (called V1) or a ββ-MHC homodimer (V3). Until recently it was thought that the human ventricle contained only β-MHC and reports demonstrating that as much as 30% of the total cardiac MHC RNA pool consisted of α-MHC were met with surprise.1,2 However, it is now clear that significant amounts of α-MHC mRNA are present in human ventricular cardiomyocytes and are translated into protein. Miyata et al detected varying amounts of α-MHC protein in all twelve of the non-failing human ventricles studied. In two samples, approximately 10% of the total myosin consisted of α-MHC and, while one of the samples only had 1.2%, the average was 7.21 ± 3.2%. Strikingly, when 10 failing hearts were examined, no detectable α-MHC protein was found.3 Subsequently, Reiser et al studied samples from 14 non-failing ventricles, 12 individuals with ischemic cardiomyopathy (ICM) and an additional 12 patients with dilated cardiomyopathy (DCM). The non-failing right ventricles had 5 ± 0.8% α-MHC protein while non-failing left ventricle had 2.5 ± 0.7% α-MHC. Both the DCM and ICM ventricles had significantly reduced levels of α-MHC.4 Although the absolute numbers differed between the two studies, the qualitative differences were consistent, indicating that α-MHC protein is down-regulated during the development of cardiac disease. Abraham et al performed serial biopsies on patients with DCM and found that functional improvement correlated well to increased α-MHC expression and concomitant decreased β-MHC mRNA.5 These data support the hypothesis that MHC isoform content is an important determinant of ventricular function and suggest that re-expression or maintenance of the α-MHC isoform might improve the function of failing myocardium.
The direct structure/function relationships between the cardiac MHC isoforms and human heart failure remain obscure with direct evidence of the mechanistic consequences of isoform shifts lacking in larger mammals. The existence of α-MHC in the human ventricle and its down-regulation during the development of heart failure emphasizes the importance of determining these structure-function relationships but directly relevant experiments cannot be done in the mouse as the cardiac isoform composition already consists of α-MHC. The ability to manipulate the genetic complement of the rabbit in a cardiac-specific manner coupled with the rabbit’s biochemical similarities to the human heart make the rabbit the most appropriate model to study cardiac MHC isoform issues as they relate to human disease. As in human hearts, the “slow” β-MHC isoform is predominant and the α-MHC:β-MHC ratio is altered by hormonal and physiological changes. For example, in rabbits and humans, thyrotoxicosis produces a shift toward the V1 isoform, while hypothyroidism, pressure overload or aging result in essentially a complete shift to V3.6
To address the relationship between MHC isoforms and cardiac function, we made TG rabbits that express rabbit α-MHC under the control of the rabbit β-MHC promoter. This strategy does not force expression of a foreign or mutated transgene; rather, we engineered persistent expression of a native isoform that would normally be down-regulated as the animal ages. Under basal conditions, persistent expression of relatively high levels of α-MHC is benign. We then subjected a cohort of TG and nontransgenic (NTG) controls to rapid ventricular pacing to create tachycardia-induced cardiomyopathy (TIC). While both groups of rabbits experienced a decrease in ventricular function, the TG rabbits exhibited significantly better contractile parameters and enhanced preservation of myocardial dimensions.
The rabbit β-MHC promoter has been described (Sanbe et al, submitted). The full-length rabbit α-MHC cDNA was subcloned into the β-MHC promoter cassette, the promoter/cDNA construct released by NotI digestion and gel purified. Oocyte injections were performed and TG rabbits derived and identified.7 Transcript levels of α-MHC, β-MHC, atrial natriuretic factor (ANF), brain natriuretic protein (BNP), sarcomeric endoplasmic reticulum ATPase (SERCA), phospholamban (PLN), and the myosin light chains were assessed by RNA dot blots and confirmed by quantitative real time PCR.8
Myofibrils were isolated from atrial and ventricular samples.9 Samples were electrophoresed on 6.5% acrylamide gels with a 100:1 acrylamide:bis-acrylamide ratio, stained with Sypro Ruby (Biorad) and digitized with a Typhoon imaging system (Molecular Dynamics). Image J software (NIH) was used for MHC band quantification.
To confirm the presence of functional TG protein, atrial and left ventricular (LV) MHC was purified from individual rabbit hearts and used in actin-activated ATPase activity experiments. F-actin was prepared from acetone powder of chicken pectoralis muscle.10 Actin concentrations ranging from 10–80 μmol were used.11
Tissue was harvested from deeply anesthetized rabbits and processed for light microscopy by an investigator blinded to genotype.12 To determine cardiomyocyte cross-sectional area, 5μm paraffin embedded sections were incubated with tetramethyl rhodamine isothiocyanate (TRITC)-labeled wheat germ agglutinin (Sigma).13 Quantification was accomplished by confocal microscopy and SimplePCI imaging software (Compix Inc). Ten fields per section were captured for analysis and all elliptical myocytes with complete sarcolemmal delineation traced. To ensure that cross-sectional rather than oblique areas were measured, cardiomyocytes were excluded from analysis if roundness determined by the quantification software was <0.66.
Rabbits were anesthetized with either ketamine and acepromazine (basal studies) or isoflurane (TIC studies) for non-invasive transthoracic echocardiograms. Inhaled isoflurane administered via mask was chosen as a lighter anesthetic for the TIC studies because of concern for depressed ventricular function during the pacing protocol. Invasive hemodynamics were made with a 4 French Millar MIKRO-TIP pressure catheter (Millar). The rabbits were anesthetized with ketamine and acepromazine and maintained with 2% isoflurane via mask. The right carotid artery was cannulated with a 4 French sheath (Cook) to introduce the catheter, which was then positioned in the LV under fluoroscopic guidance. Data were recorded with a BIOPAC system. Offline measurements were made by an investigator blinded to genotype.
After sedation with ketamine and xylazine, an echocardiogram was performed. The airway was secured via intubation with a 3-0 endotracheal tube and anesthesia maintained with isoflurane delivered by a positive pressure ventilator. Under sterile conditions, a left thoracotomy was performed and a bipolar steroid-eluting pacing lead sutured to the LV epicardium. The lead was attached to a pacemaker generator (Medtronic) specially modified to permit rapid pacing rates. The generator was implanted in a pocket under the scapula, the incision closed in layers and the rabbits extubated once sufficiently awake. Buprenorphine was administered SQ for post-operative pain management and the rabbits recovered in a pre-warmed isolette until fully awake at which time they were returned to their cages. They were monitored closely in the post-operative period for signs of pain or infection, with no complications noted.
A 5–7 day recovery was allowed between pacemaker implantation and the initiation of ventricular pacing (Figure 2). The rabbit was anesthetized with 2% isoflurane via mask and echocardiography performed to assess function. The generator was activated to pace at 300 beats per minute (bpm) and complete capture confirmed by echocardiography. Isoflurane was discontinued and after recovery the rabbit was returned to its cage. After pacing at 300 bpm for 10 days, the rabbit was anesthetized with isoflurane, function determined and the pacing rate increased to 340 bpm for another 10 days. The same procedure was repeated before increasing the pacing rate to 380 bpm for 10 days, after which time a final echocardiogram was performed and the rabbit euthanized with pentobarbital. A subset of rabbits underwent Millar catheterization of the left ventricle prior to euthanization. Hearts of two NTG and two TG rabbits were prepared for light microscopy and tissues from the remaining rabbits snap frozen in liquid nitrogen and stored at −80 °C for RNA and protein analyses.
Two-tailed t-tests were used to compare NTG and TG data gathered under both basal conditions and prior to generator implantation in the pacemaker cohorts. Wilcoxon rank sum tests were used to compare pre- to post-pacing echocardiographic measurements. Statistical significance was set as P≤0.05. All results are reported as mean ± SD.
We derived four stable TG lines, the goal being to create a collection of lines in which varying proportions of the normal β-MHC protein complement was replaced by the transgenically encoded α-MHC isoform. As is the case for the mouse,14,15 total myofilament myosin is controlled such that expression of the transgenically encoded transcript and α-MHC protein results in reduced steady state levels of β-MHC. The degree of replacement is dependent upon the level of TG expression and normal sarcomeric myosin levels are maintained. TG expression varied somewhat between the four lines, with α-MHC replacement ranging from 15%–43% (Figure 1A). We focused on lines 18 and 45, which showed, within the limits of individual variation, essentially identical levels of replacement: approximately 40% ± 3% of the endogenous β-MHC was replaced with α-MHC. At all stages of development the animals appeared to be healthy. Detailed histological analyses were performed to assess for cardiac pathology resulting from persistent expression of α-MHC. Age- and gender-matched TG and NTG myocardium were examined throughout the adult period without striking differences in hypertrophy, interstitial fibrosis and myocyte disarray being found (Figure 1B, C). We also performed RNA dot blots and quantitative real time PCR on samples isolated from line 18 TG and NTG littermates to look for activation of molecular markers of hypertrophy or failure. As expected, the TG rabbits showed significant increases in ventricular α-MHC (P<0.001) and decreases in β-MHC (P=0.01). Transcript levels for ANF, PLN and SERCA2a showed no statistically significant variations (Figure 1D).
To confirm activity of the TG protein, actin-activated ATPase rates were determined with NTG atrial and LV samples as controls. TG LV samples, with a 40/60 mix of α-MHC and β-MHC showed maximal ATPase activity that was approximately 40% that of NTG atrium, which has essentially 100% α-MHC (Figure 1E). These data confirmed that transgenically encoded α-MHC is biochemically active and influences the overall actin-activated ATPase activity as predicted based upon the α-:β-MHC ratio.
We then assessed cardiac structure and function. Serial echocardiograms were performed at 4, 8–9 and 12–15 months on cohorts of TG and NTG littermates (Table 1). At 9 months, the TG animals had slightly smaller LV end-diastolic dimensions and very mild septal hypertrophy. These differences were not apparent in 12–15 month rabbits. At no time did we find differences between the TG and NTG rabbits with respect to ventricular systolic or diastolic function, and no structural heart defects were detected. We also performed cardiac catheterization to directly measure LV pressure. The cohort consisted of 5 TG and 6 NTG mixed gender, 12 month (lines 18 and 45) rabbits. There were no significant differences in systemic blood pressure, LV peak systolic and end diastolic pressures, or maximum and minimum dP/dt (Table 2). Taking the histological, non-invasive and invasive data together, we conclude that persistent TG expression of α-MHC at high levels beyond the juvenile period is benign in the unstressed animal.
As no pathology or alterations at the molecular and cellular levels could be detected, we reasoned that they were a suitable model to test the hypothesis that α-MHC in the ventricle is cardioprotective. Rapid ventricular pacing results in depressed myocardial function in many species with clinical and laboratory findings very similar to that seen in human DCM.16,17 We applied tachycardia induced cardiomyopathy (TIC) to the α-MHC TG rabbits and developed a protocol similar to that of Jeron et al (Figure 2).18 The initial cohort consisted of 6 TG and 7 NTG 15 month line 45 rabbits of mixed gender (2 females and 4 males in the TG group and 2 females and 5 males in the NTG group). All but one animal survived the pacing regimen. One clinically asymptomatic NTG male rabbit died 5 minutes after the pacemaker rate was increased to 380 bpm. A necropsy showed ascites, pleural and pericardial effusions and a large, dilated heart. The remaining TG and NTG animals appeared outwardly normal throughout the protocol with normal cage behavior, grooming, appetite and respiratory effort.
To assess the effects of TIC at the molecular level, we compared RNA isolated from paced TG and NTG rabbits to unpaced NTG rabbits and found that both TG and NTG paced rabbits had significant increases in atrial ANF and BNP expression and increased ventricular BNP transcript (Figure 3). The differential levels of ANF and BNP between the cardiac compartments are similar to those noted by investigators using TIC in rabbits.19 Paced TG and NTG rabbits had significant decreases in atrial SERCA and PLN RNA compared to non-paced NTG, but transcript levels were not significantly down-regulated in the ventricles.
We predicted a significant difference between paced TG and NTG in the relative levels of α- and β-MHC at both the RNA and protein levels. As expected, RNA dot blots showed significant increases in ventricular β-MHC message in the paced NTG rabbits and ventricular α-MHC RNA for the TG paced rabbits. At the protein level, this correlated with a MHC complement of 100% β-MHC in myofibril preparations from paced NTG animals and persistent α-MHC accumulation in paced TG samples (Figure 4A), with the ratio of α-:β-MHC in the TG animals shifting from 40:60 under basal conditions to 60:40 after pacing.
Light microscopy revealed that rapid ventricular pacing produced histological changes consistent with dilated cardiomyopathy in both TG and NTG rabbits (Figure 4B). We found that while both groups of paced animals exhibited considerable variation in cardiomyocyte size, compared to a non-paced NTG control, both had a significantly smaller mean cardiomyocyte cross sectional area (NTG-paced 90.0 ± 40.9 μm3, TG-paced 88.8 ± 40.9 μm3, NTG-control 108.6 ± 46.8 μm3, P < 0.0001 for both paced groups vs. control, Figure 4C) and evidence of more extensive interstitial fibrosis.
At the whole organ level, there were no differences between TG and NTG discerned by the pre-operative echocardiogram (Table 3). At the protocol’s termination, the change in IVS thickness was significantly different between NTG and TG, with a decrease in septal thickness seen in NTG rabbits (ΔNTG −0.06 ± .04 cm vs. ΔTG 0.03 ± .05 cm, P=.02). At the final study, the change in shortening fraction in paced NTG rabbits was significantly different than that of TG rabbits (ΔNTG −19 ± 5% vs. ΔTG −14 ± 4%, P=.02). The pre- to post-pacing change in heart rate-corrected velocity of circumferential fiber shortening (VCFc), a pre-load independent measure of contractility, was not statistically different (ΔNTG −0.74 ± .3 circ/sec vs. ΔTG −0.52 ± .41, P = .36), which was likely due to greater variability in the TG cohort. Figure 4D–F shows representative M-mode tracings from the final echocardiograms.
To further investigate cardiac function, 10 rabbits (5 NTG and 5 TG of mixed gender) underwent the pacing protocol and Millar catheterization at the terminal study. Compared to NTG paced rabbits, TG paced animals had higher systemic blood pressure, higher peak left ventricular systolic pressure, a higher +dP/dtmax and lower −dP/dtmin(Table 4). There was no significant difference in heart rate or left ventricular end diastolic pressure. These invasive measures of ventricular function are consistent with the echocardiographic data
To begin to explore the functional significance of ventricular MHC diversity, we created TG rabbits expressing the α-MHC cDNA under control of the rabbit β-MHC promoter. This promoter was chosen as it provides significant levels of transgene expression in both unstressed and stressed ventricles. We hypothesized that under basal conditions, sufficient cardiac reserve exists to support the increased ATP demand inherent to the α-MHC isoform, and confirmed the lack of pathology in the α-MHC TG rabbits despite significant accumulation of enzymatically active α-MHC. Even at 36 months (the oldest animals in our colony) there are no apparent detrimental effects from transgenic α-MHC expression in the ventricles (data not shown).
We then tested whether the α-MHC TG rabbits respond differently than NTG rabbits when subjected to cardiac stress. Although all models of heart failure have their own set(s) of limitations, we chose to begin with TIC because the model shares many features of human DCM and chronic tachycardia is one of the differential diagnoses of DCM. Additionally, this model has been used extensively in many experimental species, including the rabbit. The model demonstrates decreased LV contractility and cardiac output with increased LVEDP and LV wall stress. The myocytes are elongated with a decreased diameter, resulting in a net decrease in cellular volume. Extracellular matrix function and myocyte adhesion capacity are decreased and plasma cateholamines, endothelin and renin activity are all increased.16,17
Our paced NTG rabbits lack any detectable α-MHC protein in the ventricles, a feature consistent with human DCM.3,4 In the TG animals, the β-MHC promoter continued to express α-MHC. Both paced TG and NTG animals demonstrated activation of the neuro-hormonal axis with increased expression of ANF and BNP. We did not find down-regulation of SERCA or PLN in paced ventricles, suggesting the rabbits were still in a compensated period, consistent with their outwardly normal appearance. In agreement with the lack of significant changes between the TG and NTG RNA levels (excluding the engineered shift in MHC isoform ratios), we found no significant differences in cardiomyocyte size. We did note marked variability in cross-sectional area ranging from hypertrophied to atrophied myocytes, similar to the histological features of human dilated cardiomyopathy.20
While cardiac function decreased in both groups as a result of rapid ventricular pacing, the TG rabbits clearly had better function than their NTG counterparts by both invasive and non-invasive measures of cardiac function. The precise mechanism of the TG advantage is not yet known. A non-specific hypertrophic response to persistent α-MHC expression (and thus more functional sarcomeres per cardiomyocyte) can be ruled out based on our light microscopy findings. Another possibility is a load change leading to altered contractility. We assessed function using both preload sensitive (shortening fraction and ±dP/dt) and afterload sensitive (shortening fraction and VCFc) indices since a truly load-independent measure of contractility has remained elusive.21 The superiority of the TG rabbits could not be due solely to favorable changes in load as both TG and NTG rabbits experienced a similar increase in LV dimension (preload) and both had similar end-systolic wall stress (afterload) at the completion of the pacing protocol (data not shown).
Taken together, our data show that moderate levels of α-MHC protein appear to be cardioprotective in TIC. It has been suggested that the disappearance of α-MHC in failing human ventricle is a compensatory mechanism aimed at preserving function by increasing myocardial efficiency. If this were true, the α-MHC TG rabbits should have fared worse with TIC than NTG rabbits. This was clearly not the case. The data thus support an alternative hypothesis: loss of α-MHC in failing myocardium is actually detrimental and re-expression of this isoform might improve cardiac function. The unique kinetic, mechanical and contractile properties of the two cardiac MHC isoforms may explain how low levels of α-MHC and its modulation during the development of cardiac disease could be functionally significant. When comparing V1 to V3, V1 has a higher ATPase activity22 and a faster maximum velocity of shortening, but a lower tension-time integral.23,24 These mechanical properties are inherent to the myosin molecule, since V1 myosin, in the in vitro motility assay, generates faster actin filament velocities but lower average force per myosin molecule when compared to V3.22,25 Single molecule studies in the laser trap suggest that differences in the kinetics of the cross-bridge cycle for V1 and V3 myosins are the major determinant of cardiac myosin’s mechanical performance.26,27 Clearly there are relationships between MHC isoform content and whole organ function as well. In rats, low level expression (12% replacement) of β-MHC on the α-MHC background resulted in altered systolic and diastolic function.28 Isolated unpaced hearts had a 17% decrease in LV pressure, a 31% decrease in dP/dtmax and a 45% increase in dP/dtmin. These changes were greater than expected for the percent replacement, indicating that the relationship between contractility and relative isoform content may be non-linear, which is consistent with the known cooperative nature of thick-thin filament interactions.
The differences in ATP requirement between the two isoforms also warrant consideration as failing myocardium is “energy starved.” As both ATP and creatine are depleted in hypertrophied and failing hearts, one might expect α-MHC TG rabbits to be less tolerant of cardiac stress than NTG rabbits. A possible explanation is that the paced rabbits were still within a compensated phase such that normal [ATP] was maintained. Another possibility is that the functional effects of α-MHC expression at the cardiomyocyte level precluded the decreases in fatty acid oxidation, glucose uptake and/or extra-mitochondrial production of ATP that eventually lead to ATP depletion. Additional studies to explore the energetic consequences of persistent α-MHC expression, under both basal and stress conditions, will be needed to resolve these issues.
In the future, it will be of interest to extend these studies to the lines that exhibit lower levels of replacement (line 69) to determine if the effects are dose dependent or depend upon some threshold of α-MHC being reached and maintained. Similarly, we can extend our studies to other models of cardiac stress such as pressure overload and myocardial ischemia to determine if α-MHC is advantageous in other cardiomyopathies. Because in many of its salient characteristics the TIC rabbit model reflects human disease, the data may be relevant to clinical cardiology given the functional and biochemical similarities between the rabbit and human heart. The data suggest that altering the MHC isoform ratios may eventually prove beneficial in the treatment of human heart failure.
Sources of Support: National Institutes of Health Grants HL69799, HL60546, HL52318, HL60546, HL56370 (J.R.)