Cardiac myosin motor deficits are associated with left ventricular dysfunction in human ischemic heart failure

Abstract

During ischemic heart failure (IHF), cardiac muscle contraction is typically impaired, though the molecular changes within the myocardium are not fully understood. Thus, we aimed to characterize the biophysical properties of cardiac myosin in IHF. Cardiac tissue was harvested from 10 age-matched males, either with a history of IHF or nonfailing (NF) controls that had no history of structural or functional cardiac abnormalities. Clinical measures before cardiac biopsy demonstrated significant differences in measures of ejection fraction and left ventricular dimensions. Myofibrils and myosin were extracted from left ventricular free wall cardiac samples. There were no changes in myofibrillar ATPase activity or calcium sensitivity between groups. Using isolated myosin, we found a 15% reduction in the IHF group in actin sliding velocity in the in vitro motility assay, which was observed in the absence of a myosin isoform shift. Oxidative damage (carbonylation) of isolated myosin was compared, in which there were no significant differences between groups. Synthetic thick filaments were formed from purified myosin and the ATPase activity was similar in both basal and actin-activated conditions (20 µM actin). Correlation analysis and Deming linear regression were performed between all studied parameters, in which we found statistically significant correlations between clinical measures of contractility with molecular measures of sliding velocity and ELC carbonylation. Our data indicate that subtle deficits in myosin mechanochemical properties are associated with reduced contractile function and pathological remodeling of the heart, suggesting that the myosin motor may be an effective pharmacological intervention in ischemia.

NEW & NOTEWORTHY Ischemic heart failure is associated with impairments in contractile performance of the heart. This study revealed that cardiac myosin isolated from patients with ischemic heart failure had reduced mechanical activity, which correlated with the impaired clinical phenotype of the patients. The results suggest that restoring myosin function with pharmacological intervention may be a viable method for therapeutic intervention.

INTRODUCTION

Cardiovascular disease remains the number one cause of mortality worldwide, accounting for 32% of deaths, or roughly 18 million lives per year (World Health Organization). Many types of cardiovascular disease may result in heart failure, in which the heart no longer can adequately pump blood to meet peripheral metabolic demands. Thus, heart failure is not one singular disease process, but instead is a clinical syndrome that reflects diverse structural and/or functional abnormalities of the heart (1). Accordingly, various etiologies of heart failure, such as coronary artery disease, cardiomyopathies, valvular disease, or hypertension, may impact the myocardium differently.

Coronary artery disease (CAD) represents the most common type of cardiovascular disease, comprising roughly one-half of the deaths attributed to cardiovascular disease (2). CAD is associated with the development of atherosclerotic plaques in the coronary vasculature and decreased blood and oxygen supply to the myocardium, often resulting in myocardial infarction. Following infarction, cell death of cardiomyocytes and subsequent fibrosis may decrease contractile function leading to ischemic heart failure (IHF), in which the heart’s pumping mechanism is compromised secondary to ischemia. IHF typically presents as heart failure with reduced ejection fraction (HFrEF), a systolic dysfunction characterized by a reduction in ventricular wall thickness and an increase in left ventricular internal diameter (3). Although organ-level changes have been well documented, molecular changes at the level of myofilaments in end-stage IHF are not fully understood.

Previously, it has been demonstrated that end-stage heart failure of various etiologies is associated with a reduction in myofibrillar ATPase (4–6), which correlates with reduced contractile performance of the myocardium (7). Although myofibrillar ATPase is consistently reduced, there are limited data regarding the function of the myosin motor specifically in end-stage heart failure. Accordingly, using human cardiac biopsies of the left ventricular free wall, we aimed to further characterize the biophysical properties of cardiac myosin in ischemic heart failure, secondary to coronary artery disease. Furthermore, we explored the role of oxidative damage in ischemia through the measurement of protein carbonylation, a common measure of protein oxidation. We hypothesized that myofibrils and myosin extracted from ischemic hearts would have decreased myosin ATPase and velocity in the in vitro motility assay, compared with healthy controls. We also hypothesized that myosin is a target for oxidative damage in ischemic heart failure, which may underlie the mechanochemical defects. To test these hypotheses, we independently extracted myofibrils and myosin from human hearts with and without ischemic heart failure and studied their mechanochemical properties.

METHODS

Procurement of Tissue and Clinical Measures

Myocardial samples were acquired at the University of Kentucky using a published protocol (8). Hearts from five patients with ischemic heart failure (IHF) who underwent cardiac transplantation as well as hearts from five organ donors who had no history of heart failure (NF) were handed to a researcher in the operating room, who then dissected samples from through-wall sections of the apical region of the left ventricle (LV). All the patients who had ischemic heart failure had end-stage disease and had an infarction several months or more before their transplant. The tissue sampling avoided overt scars and thus, the muscle was remote myocardium. However, the precise distance to the infarction was not recorded.

Samples were further dissected transmurally into sub-epicardial, midmyocardial, and subendocardial specimens. Previous data from the laboratory indicate that heart failure has a greater impact on the contractile properties of the middle transmural region of the LV than on samples from the subepicardium or subendocardium (8). The current study compared samples from LV midmyocardium to optimize the probability of detecting important effects.

Specimens were snap-frozen in liquid nitrogen within ∼30 min and stored in the vapor phase of liquid nitrogen until use. Samples were shipped overnight on dry ice to Penn State College of Medicine (Hershey, PA), where they were kept in liquid nitrogen until the day of myofibril or myosin extractions (Fig. 1). The University of Kentucky Institutional Review Board approved all procedures, and subjects or their legally authorized representatives gave written informed consent. Heart transplant recipients (IHF group) had extensive clinical workup including echocardiogram (Table 1) and cardiac catheterization (Supplemental Table S1). Echocardiographic but not catheterization data were available for the NF group.

Figure 1.

Overview of workflow. A diagrammatic representation of the workflow of this project is shown, which includes tissue collection, myofibril preparation, myosin extraction, generation of synthetic thick filaments, and mass spectrometry preparation. Corresponding assays at each level of this study are indicated in green text. All experiments, except select clinical measures, contain n = 5 biological replicates for both NF and IHF. IHF, ischemic heart failure; NF, nonfailing control.

Table 1.

Demographics and echocardiographic data

	NF	IHF	P Value
n	5	5
Demographics
Male, n	5	5
Age, yr	44.53 ± 8.36	52.94 ± 5.38	0.095
BMI, kg/m2	27.9 ± 6.27	30.7 ± 4.44	0.457
Echocardiographic data
Ejection fraction, %	43.6 ± 14.8	20.0 ± 3.5	0.009
LVIDd, cm	4.38 ± 0.30	6.30 ± 1.72	0.042
LVIDs, cm	3.38 ± 0.49	4.83 ± 1.18	0.046
Fractional shortening, %	23.0 ± 8.6	14.0 ± 8.1	0.159

Values are means ± SD; n, number of patients. Note that in the ischemic heart failure (IHF) group, there was an n = 4 for left ventricular internal diameter (LVID) during diastole (LVIDd) and an n = 3 for LVID during systole (LVIDs). Accordingly, fractional shortening, which is calculated from LVID, only contains n = 3 for IHF. These numbers are reflected in P value calculations. BMI, body mass index; NF, nonfailing controls.

Mass Spectrometry

To determine the distribution of myosin isoforms in each sample by liquid chromatography-mass spectrometry (LCMS), a ∼1–4 mg piece of each sample was sent to the University of Vermont on dry ice. The proteins in each sample were reduced, alkylated, and digested into peptides using trypsin (9). In brief, each piece of muscle was placed in a glass bottom dissection chamber containing 150-μL 0.1% Rapigest SF Surfactant (Waters, Corp.) and mechanically triturated with forceps. The solubilized proteins were reduced by addition of 0.75 μL 1 M dithiothreitol (DTT) and heating at 100°C for 10 min. Proteins were alkylated by addition of 22.5 μL of a 100 mM iodoacetamide (Acros Organics) in 50 mM ammonium bicarbonate, followed by a 30-min incubation in the dark at ∼22°C. The proteins were cleaved into tryptic peptides by addition of 25 μL of 0.2 μg/μL trypsin (Promega) in 50 mM ammonium bicarbonate, and incubation for 18 h at 37°C. Following the digestion, the samples were dried down by centrifugal evaporation and reconstituted in 100 μL of a 7% formic acid in 50 mM ammonium bicarbonate solution to inactivate trypsin and degrade Rapigest (1 h, 37°C). Samples were dried down once more and reconstituted in 100-μL 0.1% trifluoroacetic acid (TFA) for further cleavage of Rapigest (1 h, 37°C). Samples were dried down a final time, reconstituted in 100 μL 0.1 TFA, centrifuged for 5 min at 18,800 RCF (Thermo, Sorvall Legend Micro 21 R), and 75 μL of the supernatant were removed for analysis by LCMS (9). Briefly, a 20-μL aliquot of each sample was injected onto a 100 Å, 1.8 μm, 1 × 150 mm, Acquity UPLC HSS T3 column (Waters Corporation) attached to an UltiMate 3000 ultrahigh pressure liquid chromatography (UHPLC) system (Dionex). The UHPLC effluent was directly infused into a Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific). Data were collected in data-dependent MS/MS mode. Peptides were identified by SEQUEST and LC peak areas of peptides of interest, being those unique to, or shared between protein isoforms, were extracted using the Proteome Discoverer 2.2 software package. LC peak areas were imported into excel for further analysis (9).

The relative abundance of individual myosin heavy-chain and light-chain isoforms between groups was determined from the average of the top three LC peak areas for each protein isoform. The LC peak areas were normalized by dividing each value by the average value of the top three LC peak areas of peptides shared between the α- and β-myosin heavy chain isoforms. This normalization accounted for differences in the amount of tissue digested and fraction of each sample injected into the mass spectrometer (9). The average relative abundance ± SD was reported.

Myofibril Preparation

Myofibrils were prepared from human cardiac tissue, modified from previously established protocols (10). Briefly, roughly 200 mg of tissue was placed into 4 mL of rigor buffer, consisting of (in mM) 10 Tris (pH 7.0), 5 EGTA, 132 NaCl, 5 KCl, 1 MgCl₂, 1 NaN₃, 1 DTT, and 0.1 PMSF and 1 µg/mL leupeptin and 1 µM pepstatin A. The tissue was briefly minced with scissors and then homogenized (Brinkmann Homogenizer, Model PT 10/35) four times for 10 s, with 30 s of incubation on ice between homogenizations. Cellular debris was removed via a series of centrifugations (5 min at 2,500 g, 4°C) in a wash buffer (30 mM imidazole, pH 7.0, 60 mM KCl, 2 mM MgCl₂, 1 mM NaN₃, 1 mM DTT, 0.1 mM PMSF, 1 µg/mL leupeptin, and 1 µM pepstatin A) for a total of five spins. The first spin contained 0.5% Triton-X to permeabilize cellular membranes, whereas the final four spins were absent of detergent. To measure the concentration of the myofibrils, a 5% myofibril dilution was prepared for spectrophotometry, in which the formula, (A₂₈₀ − A₃₂₀/ε)·20, where ε = 0.7 mL/(mg·cm) was employed to calculate myofibril protein concentration (10). Integrity of the myofibrils was inspected via light microscopy before experimentation. The target starting myofibril concentration for the ATPase assay was 10 mg/mL. Myofibrils not used immediately for ATPase were combined with equal parts glycerol (1:1) and kept at −20°C.

Myofibrillar ATPase

Myofibrillar ATPase was performed using a malachite green assay, as previously described (11, 12). Briefly, 800-µL myofibril and ATP solutions were prepared separately at increasing CaCl₂ concentrations (pCa₉ to pCa₃), where [Ca] (in µM) = 10^(6-pCa). Myofibril solutions had a final protein concentration of 0.2 mg/mL, whereas ATP solutions had a final concentration of 10 mM ATP. KCl concentration was reduced in each solution in accordance with increasing pCa to maintain similar final ionic strength. ATPase buffer, consisting of (in mM) 30 MOPS (pH 7.0), 167.6 KCl, 7.5 MgCl₂, and 2 EGTA, was used for dilutions for both myofibril and ATP solutions. A color reagent was prepared directly before the ATPase assay, in which 150 mL of 0.045% malachite green was combined with 50 mL of 4.2% ammonium molybdate (MG-AM solution). The color reagent served as an indicator of liberated inorganic phosphate from ATP hydrolysis.

Myofibril and ATP samples at a given pCa were placed in a water bath (37°C) for 5 min before being combined to initiate the reaction. At 1, 2, 3, 4, 6, 8, and 10 min, 100 µL of the reaction was sampled and immediately quenched with 100 µL 2 N HCl and then neutralized with 16 µL of 2 M Tris-3 M NaOH. After 30 s, 800 µL of MG-AM was added and immediately combined with 100 µL of 34% sodium citrate. Color development occurred over 15 min at room temperature, and the absorbance at 620 nm was measured by spectrophotometry to quantify liberated phosphate. Solutions from each of the time points were added in triplicate to a 96-well plate. A linear slope was calculated for the reaction at each pCa, which was converted to ATPase rate using a phosphate standard curve (0–40 µM). ATPase rates from NF versus IHF cardiac tissue were averaged at each given pCa and were fit to a logarithmic-response curve using the Hill equation: Y = y_basal + y_amp·{[Ca²⁺]⁻ⁿ/([Ca²⁺]⁻ⁿ + [pCa₅₀]⁻ⁿ)}, where y_basal is basal ATPase, y_amp is amplitude, maximal ATPase is y_basal + y_amp, n is the Hill coefficient, and pCa₅₀ represents the pCa at one-half maximal ATPase activity.

Cardiac Myosin Extraction from Tissue

Full-length cardiac myosin was extracted directly from human cardiac tissue, adapted from previously described protocols (13, 14). Briefly, left ventricular free wall heart tissue was retrieved from liquid nitrogen, and 150 mg was weighed for extraction. The tissue was placed in extraction buffer, consisting of (in mM) 300 KCl, 150 K₂HPO₄, 10 Na₄P₂O₇, 1 MgCl₂ (pH 6.8), 2 DTT, and 1 PMSF and 10 µg/mL apoprotinin and leupeptin, briefly minced with scissors, and then dounced manually in a glass homogenizer for 10 min. The samples were spun for 5 min at 10,000 g at 4°C and then spun in an ultracentrifuge for 20 min at 53,000 g at 4°C (TLA120.2 rotor) to remove cellular debris. The myosin-containing supernatant was then precipitated by 10-fold dilution with dH₂O for 60 min on ice. The samples were spun once more for 10 min at 10,000 g and 4°C, and the pellets were rinsed with cold dH₂O and resuspended in high-salt resuspension buffer, containing (in mM) 25 imidazole, 600 KCl, 1 EGTA, and 4 MgCl₂ (pH 7.4).

To remove “dead heads” from the samples, myosin was further purified using an actin spindown protocol, in which actin was added at 1.5× the concentration of the myosin sample. The actomyosin was pelleted by ultracentrifugation (95,000 rpm in TLA120.2 rotor, 15 min, 4°C) and then released in MOPS300 buffer, containing (in mM) 10 MOPS (pH 7.0), 300 KCl, 1 EGTA, 1 MgCl₂, and 1 DTT, in the presence of 2 mM ATP (95,000 rpm in TLA120.2 rotor, 10 min, 4°C). Final myosin concentrations were calculated via Bradford assay, and the quality of the extraction was assessed by SDS gel.

Protein Quality Control

All myofibril preparations and myosin extractions were assessed via SDS-PAGE. Myofibril concentration was measured using spectrophotometer absorbance measurements noted above (10), whereas myosin concentration was measured via Bradford assay using BSA as a standard.

In Vitro Motility

The in vitro motility (IVM) assay was performed with full-length human cardiac myosin, adapted from previously established protocols (13, 15). Briefly, microscope coverslips were coated with 1% nitrocellulose in amyl acetate (Ladd Research) and applied to a microscope slide with double-sided tape to create a flow cell. Myosin in MOPS300 buffer at a concentration of 125 µg/mL (0.5 µM) was applied directly to the nitrocellulose surface, and the surface subsequently was blocked with BSA (1 mg/mL). Unlabeled sheared actin (2 µM) followed by ATP (2 mM) was added to ensure blocking of inactive myosin heads (“dead heads”). Fluorescently labeled actin (10 nM final, labeled with phalloidin-Alexa 555) was then added to the flow cell. To initiate motility, an activation buffer containing 0.35% methylcellulose, an ATP regeneration system (2 mM ATP, 5 mg/mL glucose, 46 U/mL pyruvate kinase, and 0.46 mM phosphoenolpyruvate), oxygen scavengers (0.1 mg/mL glucose oxidase, 0.018 mg/mL catalase), and 10 mM dithiothreitol (DTT) was added. The slide was visualized promptly with a NIKON TE2000 microscope (DsRed filter; excitation/emission 555/588 nm) equipped with a 60×/1.4 NA phase objective and a Perfect Focus System. All images were acquired at 1-s intervals for 2 min using a shutter-controlled CoolSNAP HQ2 cooled CCD digital camera (Photometrics) binned 2 × 2. Temperature (25 ± 1°C) was monitored using a thermocouple meter (Stable Systems International). Videos were exported to ImageJ and prepared for automated FAST software motility analysis (16), from which >1,000 actin filaments from one experiment (i.e., slide) per myosin extraction (n = 5) at 0.5 µM myosin were compiled for statistical analysis, NF versus IHF. The sample means of independent experiments were averaged, and the data were presented as a SuperPlot, as previously described (13, 17).

Protein Carbonylation Western Blots

To assess differences in protein oxidation, we used a Protein Carbonyl Assay Western Blot Kit (Abcam, ab178020) (18). Residual myosin samples from extractions were flash-frozen in liquid nitrogen and stored at −80°C. Samples were thawed in parallel and diluted down to 2.0 µM in preparation for a derivation reaction with 2,4 dinitrophenylhydrazine (DNPH), which reacts specifically with protein carbonyl groups, a common marker of protein oxidation. Each protein sample was subjected to a DNPH reaction (+) and a derivation control reaction (−) for 15 min at room temperature. The samples then were neutralized and run on a 15% polyacrylamide gel. The gels were transferred to a PVDF membrane, which was incubated with DNPH primary antibody overnight, washed, and then incubated with horseradish peroxidase (HRP) goat anti-rabbit secondary antibody before development with enhanced chemiluminescence (ECL). Both antibodies were supplied with the Abcam kit. SDS-PAGE was performed with roughly 800 ng of protein per well to get an adequate signal for the essential light chain (ELC), whereas 100 ng of protein per well was appropriate to minimize the signal for the myosin heavy chain (MHC). Coomassie gels were run in tandem to confirm protein concentration of the protein samples.

Densitometry of myosin heavy chains in both Western blots as well as Coomassie SDS gels were calculated using ImageJ software (National Institutes of Health, NIH). Oxidation densitometry was normalized to a carbonylated BSA standard on each blot, and protein concentration was calculated using a myosin standard. The degree of protein oxidation was expressed as the ratio between Western blot densitometry:Coomassie densitometry for each protein sample. Ratios were averaged by group (nonfailing vs. failing) and then compared statistically by a Student’s t test.

Reconstitution of Synthetic Thick Filaments

Synthetic thick filaments (STFs) were reconstituted from purified myosin, adapted from previously established protocols (19). Briefly, myosin in high salt (MOPS300) was diluted 10-fold with MOPS0 (0 mM KCl) to reduce the ionic strength (final concentration of 30 mM KCl). At lower ionic strength, myosin molecules in solution can reconstitute into thick filaments, previously coined synthetic thick filaments (STFs). The diluted sample was placed on a rocker at 4°C overnight, and a precipitate was visible in the solution the following day. Samples were gently homogenized, and 50 µL was used to assess filament formation by ultracentrifugation. The samples were spun in an ultracentrifuge for 10 min at 95,000 rpm and 4°C (TLA120.2 rotor). The supernatant was removed, and the pellet was resuspended in 50 µL of 6 M urea. Supernatant and pellet samples were run in tandem on an SDS gel to assess the degree of filament formation.

STF ATPase

An Applied Photophysics stopped-flow apparatus was used to perform NADH-coupled ATPase assays to measure steady-state actin-activated ATPase of STFs. Briefly, STFs were assayed in the absence and presence of F-actin (20 μM), in the NADH-coupled assay (13), in MOPS30 (30 mM KCl) at 25°C. Following the addition of 1 mM ATP, changes in absorption at 340 nm were measured for 200 s, and the transients were fit to a linear function. A standard curve generated from known ADP concentrations was used to determine the ATPase rate in the absence and presence of actin. STF concentration was based on Bradford assay of full-length myosin before dilution.

Actin Preparation

F-actin used in myosin extraction and STF ATPase assay was purified from rabbit skeletal acetone powder, as previously described (20).

Statistics

All measures were compared using parametric Student’s t tests (NF vs. IHF) with statistical significance set at P ≤ 0.05. All clinical data, the myofibrillar ATPase, the myosin in vitro motility, the STF ATPase, and protein carbonylation measures included an n = 5 separate donor hearts, unless otherwise indicated. Correlation analysis was performed to examine relationships between clinical and molecular measures. Strong correlations were then fit to Deming linear regression to account for variability in X and Y variables and to assess for statistical significance, where P ≤ 0.05.

RESULTS

Donor Characteristics

The two groups (NF and IHF) each consisted of five age-matched males, with similar body mass index (BMI; Table 1). Before transplant, all patients with IHF were classified as class IV heart failure (most severe) in the New York Heart Association (NYHA) Functional Classification, indicating that they were symptomatic at rest. Moreover, each patient had documented history of ischemia (e.g., chronic total occlusion of LAD, stent of LAD/circumflex arteries, or coronary artery bypass graft). From the available clinical data, the IHF group had a significantly depressed ejection fraction (20.0 ± 3.5%), indicative of severe systolic dysfunction. This was significantly different (P = 0.009) compared with the NF group (43.6 ± 14.8%), which was within the borderline range of ejection fraction (40–50%). The high standard deviation and borderline classification of the NF group are considered in discussion. In addition, hearts from the IHF group demonstrated a dilated phenotype characteristic of heart failure, with significantly increased left ventricular internal diameter during both diastole (LVIDd; IHF = 6.30 ± 1.72 cm; NF = 4.38 ± 0.30 cm; P = 0.042) and systole (LVIDs; IHF = 4.83 ± 1.18 cm; NF = 3.38 ± 0.49 cm; P = 0.046) relative to NF controls (Table 1). Fractional shortening (FS) also was calculated from available echocardiogram data, where FS = LVIDd – LVIDs/LVIDd × 100%. The NF group demonstrated an FS of 23.0 ± 8.6%, whereas the IHF group demonstrated 14.0 ± 8.1 falling within the range of severe LV dysfunction (P = 0.159) (21). Other clinical data including prescription medications (NF and IHF) and cardiac catheterization data (IHF only) such as cardiac index (CI), cardiac output (CO), estimated O₂ consumption, stroke volume (SV), and pulmonary vascular resistance (PVR) are reported in Supplemental Table S1.

Myofibril ATPase

In the myofibrillar ATPase assay, both NF and IHF groups demonstrated a similar response to increasing calcium concentrations (Fig. 2). For standard parameters of the Hill equation, the myofibrillar ATPase in the IHF samples was not statistically different compared with NF. The basal ATPase was 49.91 ± 8.66 nmol P_i/min/mg in NF compared with 28.52 ± 6.77 nmol P_i/min/mg in the IHF group (P = 0.088), whereas the maximal ATPase was 168.71 ± 15.52 nmol P_i/min/mg in the NF group compared with 125.33 ± 12.33 nmol P_i/min/mg in the IHF group (P = 0.059). There was no shift in calcium sensitivity with a pCa50 of 6.20 ± 0.12 in the NF group compared with 6.16 ± 0.11 in the IHF group (P = 0.800). Other parameters of the Hill equation including amplitude (Maximum ATPase – Basal ATPase) and the Hill coefficient were not significantly different and have been reported in Table 2.

Figure 2.

Myofibrillar ATPase. The myofibrillar ATPase activity was measured at increasing calcium concentrations from pCa9 to pCa3, where [Ca] [in µM = 10^(6-pCa)]. Both nonfailing (NF, blue; n = 5) and ischemic heart failure (IHF, red; n = 5) samples demonstrated a similar calcium sensitivity (pCa₅₀). Each data point is the average of 5 ATPase experiments, each done on a separate myofibril preparation from a separate heart (1 technical replicate per preparation), expressed as means ± SD, and these data have been fit to the Hill equation. Numerical data for the Hill equation parameters are summarized in Table 2, expressed as means ± SE.

Table 2.

Myofibrillar ATPase data

Parameter	NF	IHF	P Value
Patients, n	5	5
Basal ATPase	49.91 ± 8.66	28.52 ± 6.77	0.088
Amplitude	118.8 ± 12.88	96.81 ± 10.18	0.216
Maximal ATPase	168.71 ± 15.52	125.33 ± 12.23	0.059
Hill coefficient, n	1.521 ± 0.722	1.662 ± 0.810	0.899
pCa50	6.20 ± 0.12	6.16 ± 0.11	0.800

Values are means ± SE; n, number of patients. IHF, ischemic heart failure; NF, nonfailing control; pCa₅₀, pCa at one-half maximal ATPase activity.

In Vitro Motility

The in vitro motility assay demonstrated that myosin extracted from IHF samples (727 ± 40 nm/s) had a decreased ability to translocate actin compared with NF counterparts (853 ± 86 nm/s; P = 0.018), a 15% decline in sliding velocity that is indicative of impaired motor function. The data graphically are represented as a SuperPlot, and the mean of the sample means ± SD is reported in Fig. 3.

Figure 3.

Myosin extraction and in vitro motility. A: representative SDS gel, demonstrating steps of the myosin extraction protocol. Lanes 2–3 contain the NF and IHF samples following the initial high-salt extraction, lanes 4–5 contain the same samples following ultracentrifugation, and lanes 6–7 contain the final myosin samples following an actin spin down and release, where MHC is myosin heavy chain, ELC is essential light chain, and RLC is regulatory light chain. B: in vitro motility. Actin sliding velocities from the unloaded in vitro motility assay represented as a SuperPlot, in which each triangle represents the mean of each individual experiment from separate myosin preparations (n = 5), the color-coded transparent dots represent the range of velocities corresponding with each triangle, and the black line represents the mean of the sample means. Numerical data for motility are displayed. IHF, ischemic heart failure; NF, nonfailing control.

Mass Spectrometry

Tissue samples were analyzed by LCMS to determine whether there were differences in the expression of myosin heavy chain (MHC), myosin regulatory light chain (RLC), and myosin essential light chain (ELC) isoforms. Peptides specific to the α (MYH6)- and β (MYH7)-MHC, ventricular RLC (MYL2), ventricular ELC (MYL3), and atrial ELC (MYL4) isoforms were observed. There were no apparent differences in the relative expression in either of these isoforms between sample groups (Table 3). The β-MHC and ventricular ELC isoforms, representing 98.3 ± 1.5 and 98.7 ± 1.2% of the total MHC and ELC pools, appeared to be the predominate isoforms expressed in both sample groups, based on the absolute abundances of their top three LC peak areas from each of these proteins. The public MS database is available at https://massive.ucsd.edu. Data set MSV000090858.

Table 3.

Relative abundance of myosin by mass spectrometry

	IHF/NF
n	5
β/Total MHC	1.00 ± 0.02
MYL2 (ventricular RLC)	0.97 ± 0.13
MYL3 (ventricular ELC)	0.95 ± 0.03
MYL4 (atrial ELC)	0.99 ± 1.32

Values are means ± SD; n, number of patients. ELC, essential light chain; IHF, ischemic heart failure; MHC, myosin heavy chain; NF, nonfailing control; RLC, regulatory light chain.

STF ATPase

Synthetic thick filaments (STFs) were formed efficiently by decreasing the ionic strength of the myosin-containing solution, which was confirmed by ultracentrifugation and SDS PAGE (Fig. 4). We found no difference in the basal-level ATPase (0 μM F-actin) activity (NF = 0.11 ± 0.02 s⁻¹; IHF = 0.10 ± 0.03 s⁻¹; P = 0.552). Basal ATPase values were considerably higher than previously reported in STFs (19), likely due to differences in buffer conditions and methodology. We also found no differences in actin-activated (20 μM F-actin) conditions (NF = 0.23 ± 0.08 s⁻¹; IHF = 0.23 ± 0.04 s⁻¹; P = 1.00). Both NF and IHF STFs demonstrated similar fold-change with actin activation, reported in Table 4 (Fig. 4).

Figure. 4.

Synthetic thick filament formation and ATPase activity. A: representative SDS gel demonstrating similar propensity for filament formation in synthetic thick filaments (STFs) for nonfailing (NF) and failing (F) samples. The filamentous myosin is found in the pellet after ultracentrifugation. B: STF ATPase in the absence of actin (basal, 0 µM actin) and actin-activated conditions (20 µM actin). Data are from 5 separate myosin preparations from 5 hearts/group. Numerical data and P values for STF ATPase have been provided in Table 4 (means ± SD).

Table 4.

Synthetic thick filaments ATPase data

	NF	IHF	P Value
n	5	5
Basal ATPase	0.11 ± 0.02	0.10 ± 0.03	0.552
Actin-activated ATPase	0.23 ± 0.08	0.23 ± 0.04	1.000
Fold actin-activation	2.1 ± 0.6	2.5 ± 0.5	0.285

Values are means ± SD; n, number of patients. IHF, ischemic heart failure; NF, nonfailing control.

Protein Carbonylation Western Blot Analysis

Protein oxidative damage was assessed via a protein carbonylation Western blot kit (Abcam) in NF and IHF samples. The intensity of the myosin heavy chain (MHC) and essential light chain (ELC) bands was evaluated by densitometry of the Western blots and normalized by the protein concentration as assessed by Coomassie gels. There was no significant difference in the MHC measures in NF (0.51 ± 0.14) compared with IHF (0.58 ± 0.15; P = 0.50), and the ratio for ELC carbonylation was also not significantly different between NF (0.39 ± 0.15) and IHF (0.62 ± 0.26; P = 0.14; Table 5). Interestingly, there was little to no signal for the regulatory light chain (RLC) in the Western blots for oxidative damage (Fig. 5).

Table 5.

Myosin carbonylation data

	NF	IHF	P Value
n	5	5
MHC oxidation/protein concentration	0.51 ± 0.14	0.58 ± 0.15	0.50
ELC oxidation/protein concentration	0.39 ± 0.15	0.62 ± 0.26	0.14

Values are means ± SD; n, number of patients. ELC, essential light chain; IHF, ischemic heart failure; MHC, myosin heavy chain; NF, nonfailing control.

Figure 5.

Carbonylation Western blot analysis. A: representative Western blot that contains nonfailing (NF) and failing (F) myosin protein samples with or without DNPH derivation of carbonyl compounds. The PVDF membrane was probed with α-DNPH 1° antibody, α-goat anti-rabbit 2° antibody, and enhanced chemiluminescence before visualization in a dark room. Blank lanes between samples contain DNPH-derivation control reactions, in which DNPH is not detected. B: representative SDS gel used to calculate protein concentration of myosin samples, relative to myosin standard of known concentration. Degree of carbonylation was expressed as densitometry of Western blot:densitometry of protein via Coomassie staining, which has been included in Table 5 (means ± SD). Data include 5 separate myosin preparations from 5 hearts/group. Actin was not examined because it included a mixture of endogenous actin and skeletal actin used for myosin purification. DNPH, 2,4 dinitrophenylhydrazine.

Correlation Analysis and Deming Linear Regression

To investigate relationships between clinical and molecular measures, correlation analysis and Deming linear regression were performed on all studied variables. Many variables shared weak or moderate correlations. However, statistically significant strong correlations were found between various combinations of ejection fraction, fractional shortening, sliding velocity, and ELC oxidation. Sliding velocity shared a strong direct relationship with ejection fraction (r = 0.79, P = 0.006) as well as fractional shortening (r = 0.77, P = 0.025), demonstrating a link between clinical and molecular measures. Sliding velocity also shared a strong inverse relationship with ELC carbonylation (r = −0.70, P = 0.025), whereas ejection fraction and ELC carbonylation shared a negative moderate correlation (r = −0.64, P = 0.048). A correlation matrix between all measured variables is included in Fig. 6, where statistically significant relationships are highlighted.

Figure 6.

Correlation and deming regression analysis. A: correlation matrix analysis was performed between clinical and molecular measures to assess for relationships between variables. Positive (direct) correlations are indicated in blue, and negative (indirect) correlations are indicated in red, along with corresponding Pearson’s correlation coefficients. B–E: related variables indicated by correlation analysis were fit to Deming linear regression. Blue data points reflect measures from NF group, whereas red data points reflect measures from IHF group. B and C: strong direct relationships between ejection fraction and sliding velocity (n = 10), as well as fractional shortening and sliding velocity (n = 8), respectively, linking clinical and molecular measures. D and E: strong indirect relationships between ELC carbonylation and sliding velocity (n = 10), as well as ELC carbonylation and ejection fraction (n = 10), respectively. Pearson’s correlation coefficients (r) were calculated via the correlation matrix, whereas P values were calculated via Deming regression analyses. ELC, essential light chain; IHF, ischemic heart failure; NF, nonfailing control.

DISCUSSION

In this study, we used cardiac tissue harvested from heart transplant recipients secondary to ischemic heart failure (IHF) and age-matched organ donors of similar BMI with a nonfailing (NF) phenotype (Fig. 1). With this tissue, we demonstrate functional deficits in full-length cardiac myosin in IHF compared with NF that correlate with clinical echocardiographic measures of contractility, supporting that myosin mechanochemical deficits are associated with left ventricular dysfunction in ischemic heart failure.

Interpretation of Myofibrillar ATPase

The study of myofibrils is a powerful tool, as these subcellular organelles bridge the gap between whole muscle or tissue-level measurements and those of purified protein (10). Because myofibrils contain endogenous thin and thick filaments with associated regulatory proteins, myofibrillar ATPase is thought to represent physiological ATPase more closely than purified myosin ATPase and is well correlated with muscle fiber contractile performance (7). In the current study, our myofibrillar ATPase data in IHF compared with NF were not significantly different at basal or maximal ATPase, and there were no changes in calcium sensitivity (Fig. 2, Table 2). Contrary to our data, there is a long history of decreased myofibrillar ATPase in human failing hearts reported in the literature, regardless of etiology. Over the past 60 years, reduced myofibrillar ATPase in heart failure has been reported in hypertensive heart disease (22), congestive heart failure (23), coronary artery disease, idiopathic cardiomyopathy, and immunologically rejected hearts (4, 5, 24). Notably, we failed to reject the null in our sample population, with P values of 0.088 at basal conditions and 0.059 at maximal ATPase. Given these conflicting findings, we further investigated potential thick filament deficits in IHF through biochemical and mechanical changes of cardiac myosin between our two groups.

Mechanochemical Properties of Myosin

Although numerous studies have examined myofibrillar ATPase in heart failure, far fewer have investigated mechanical or chemical properties of myosin specifically. The in vitro motility (IVM) assay provides a window into a myosin’s motor function by measuring the speed at which it can translocate fluorescent actin on a glass coverslip (15), whereas our myosin ATPase experiment is a measure of myosin’s enzymatic activity in a thick filament environment (19). Existing reports in the IVM and/or myosin ATPase assays in human heart failure are variable, pending the etiology, duration, and severity of the heart failure, as well as the region of myocardium sampled (4, 6, 24, 25).

Our data demonstrate a 15% reduction in unloaded actin sliding velocity when propelled by myosin extracted from IHF cardiac tissue compared with NF counterparts (Fig. 3). Importantly, this difference was observed in the absence of myosin heavy chain or light chain isoform shifts (Table 3). Thus, our results support that myosin isoform shifts cannot account for deficits in myosin mechanochemistry in failing hearts, in agreement with previous work (4, 25). Excitingly, we demonstrate strong positive correlations between sliding velocity with clinical measures of contractility (ejection fraction and fractional shortening) in our sample, suggesting that the ability of isolated myosin to propel actin in unloaded conditions may serve as an indicator for global ventricular function.

The same isolated myosin reconstituted as STFs, however, showed no changes in basal or actin-activated ATPase (Fig. 4, Table 4). These observations imply either a mechanical or biochemical deficit in myosin from IHF hearts. Velocity in the in vitro motility assay can be expressed as v = d/t_on, where d is the step size of the myosin motor and t_on is the time spent attached to actin in each ATPase cycle (26). Thus, we predict that in the IHF myosin, there is either 1) a decrease in the step size of the myosin motor, secondary to biochemical changes induced by ischemic insults or 2) slower detachment kinetics in unloaded conditions that may be further exacerbated in loaded conditions. Although we observed no changes in steady-state ATPase, it is still plausible that ADP release is slower in IHF, which would account for reduced unloaded velocity without impacting overall rate of ATPase. Optical trapping or stopped flow transient kinetic assays of detachment (ADP release and ATP-induced dissociation experiments) would be suitable follow-up experiments to assess these possibilities. To further investigate potential biochemical alterations of myosin secondary to ischemia, we explored the degree of oxidative damage in our extracted myosin.

Potential Role of Oxidative Damage

Given an established role of reactive oxygen species in heart disease (27, 28), we set out to measure protein oxidation of the extracted myosin from our samples. Through DNPH derivation reactions and Western blot analysis, we probed for carbonylation, a common and irreversible type of protein oxidation previously shown to be associated with myosin deficits (29). We found that there were no differences in the intensity of the signal associated with the ELC or MHC between NF and IHF myosin samples (Fig. 5, Table 5). In both groups, the MHC and ELC consistently demonstrated some degree of carbonylation, whereas the regulatory light chain (RLC) showed little to no signal in all NF and IHF samples, suggesting minimal protein oxidation of this component of myosin (Fig. 5). This is supported by previous mass spectrometry results, which favors the hypothesis that RLC is structurally protected in the hinge region at the transition between subfragments 1 and 2, whereas ELC is more exposed in its proximal location at the lever arm (30).

Though we demonstrated no changes between groups, ELC carbonylation did share statistically significant negative correlations with actin sliding velocity and ejection fraction, imploring future studies for potential mechanisms. Previously, Prochniewicz et al. (30) demonstrated that oxidation of methionine residues in MHC and ELC, but not RLC, were associated with decreases in shortening velocity and force in skeletal muscle fibers, as well as decreases in K⁺-ATPase in myofibrils, myosin, and S1. Other studies have demonstrated that carbonylation of cardiac or skeletal myosin is associated with decrements in actin sliding velocity, decreased force, and changes in myosin’s duty ratio assessed via optical trapping (29, 31). These studies support that oxidative changes in myofilaments are implicated in myosin mechanochemical dysfunction. Though we chose carbonylation as a proxy for global oxidation, this does not discount other forms of potential oxidative damage and/or posttranslational modifications in myofilaments that previously have been reported to be altered in heart disease including nitration (32, 33), acetylation (34), glutathionylation (35), methylglyoxalation (36), and citrullination (37). Recent work demonstrated a reduction in acetylation of the K-951 residue of subfragment-2 in human IHF samples, which was proposed to impact tail flexibility and myosin’s ability to adopt the autoinhibitory state (38). Importantly, our mass spectrometry measurements were performed on tissue samples, whereas our isolated myosin was designated for the carbonylation assay. There were no data indicating specific posttranslational modifications in IHF in whole tissue samples via MS. Based on our carbonylation results, however, future analyses may consider performing mass spectrometry on isolated myosin specifically to assess oxidative changes to ELC.

Alternative Interpretations, Limitations, and Future Directions

Importantly, our motility assay only was performed under unloaded conditions. The introduction of load to the system with α-actinin or utrophin would allow generation of force-velocity curves, and thus a measure of molecular power output (39). We predict that our 15% decrement observed in unloaded conditions would be exacerbated in loaded conditions and that myosin from IHF hearts would demonstrate decreased power relative to NF samples. The current study also addresses myosin motor-specific insult in the absence of collagen deposition. We observed a reduction in actin sliding velocity in purified myosin from IHF samples, which removes any potential impact of fibrosis from the system. Although fibrosis unequivocally contributes to reduced contractility in ischemic hearts (40, 41), our data suggest that the contractile machinery is also damaged. Deficits to myosin specifically provide support for myosin activators in the pharmacological treatment of systolic heart failure (42–44). Treatment of myofibrillar or myosin samples from heart failure with a cardiac activator such as OM or danicamtiv in an attempt to recover depressed function would be an appropriate follow-up study.

Furthermore, we recognize that deficits in other aspects of the contractile apparatus may contribute to ventricular dysfunction in heart failure. For example, we are unable to exclude the possibility of the changes in myosin autoinhibition (e.g., fraction of myosin heads in the superrelaxed state), as we did not perform single turnover assays nor measure common targets of phosphorylation associated with myosin regulation such as RLC or myosin binding protein C (45). Given the purported role of myosin regulation and sequestration in hypertrophic cardiomyopathy (46) and dilated cardiomyopathy (47), it is plausible that myosin autoinhibition may be impacted in heart failure. Posttranslational modifications could impact critical residues on the myosin molecule and subsequently affect the rate at which myosin can enter and exit the interacting heads motif (IHM), as previously suggested (38).

Although our work focused on the thick filament, changes to the thin filament previously have been documented in heart failure. Canton et al. (48) reported increased carbonylation of actin and tropomyosin and increased disulfide cross-bridge formation in tropomyosin in human failing hearts, which correlated with contractile impairment and reduced myocardial viability. Anderson et al. found a correlation between a shift in troponin-T isoforms and decreased myofibrillar ATPase (49), a result that led several groups to attribute heart failure to changes in the thin filament. Although we recognize that the thin filament is impacted during heart failure, our data provide evidence that myosin-containing thick filaments also are targets of adverse changes in IHF. We emphasized the role of contractile function in ischemic heart failure, though we also recognize that numerous reports have indicated changes in Ca²⁺-handling (50, 51) and mitochondrial/metabolic function (52–54) in various etiologies of heart failure. We predict that functional deficits of the contractile mechanism in heart failure are further compounded by such changes in calcium homeostasis and/or metabolic regulation.

The benefits and challenges of experimentation with human tissue also should be considered. First, procurement of such specimens is very difficult for most research groups and is only available when clinicians have decided not to transplant the heart from the organ donor into another patient. Potential reasons for this decision include but are not restricted to organ size, potential infection, and immunological status. Once the tissue is obtained, it is difficult to control for all the potential confounders (i.e., sex, age, genetic background, comorbidities, and clinical treatments) that might influence the data. Collectively, these factors contribute to variability, and in our case, the borderline ejection fraction of the nonfailing group. Despite these challenges, tissues from organ donors remain the gold standard control group for basic science investigations of human myocardium. In this work, we were able to compare clinical and molecular measures from tissue between two distinct clinical groups matched for age, sex, and body composition.

Conclusion Statement

In this study, we used human cardiac tissue harvested from diseased hearts (ischemic heart failure, IHF) and compared it with cardiac tissue from organ donors of similar age and BMI with a nonfailing (NF) cardiac phenotype. With this tissue, we demonstrate functional motor deficits in cardiac myosin in IHF compared with NF, and these motor properties were associated with clinical measures of ejection fraction and fractional shortening, as well as a molecular measure of ELC carbonylation. Collectively, these data provide evidence that impairments to the contractile apparatus, specifically the myosin motor, are associated with left ventricular dysfunction in human ischemic heart failure. Future studies should aim to measure changes in molecular power output through implementation of load to the system. Other studies may aim to characterize the effect of pharmacological agents (e.g., myosin activators) in recovering depressed contractile function across various heart failure etiologies.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.21717797.v1.

GRANTS

This work was supported by National Institute on Alcohol Abuse and Alcoholism Grant R37AA011290 (to C.H.L.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.S.C. and C.M.Y. conceived and designed research; D.V.R., J.G., N.B.W., A.M.P., and C.M.Y. performed experiments; D.V.R., J.G., G.N.M., N.B.W., M.J.P., K.S.C., and C.M.Y. analyzed data; D.V.R., G.N.M., N.B.W., C.H.L., M.J.P., K.S.C., and C.M.Y. interpreted results of experiments; D.V.R., G.N.M., and C.M.Y. prepared figures; D.V.R. and C.M.Y. drafted manuscript; D.V.R., G.N.M., M.J.P., K.S.C., and C.M.Y. edited and revised manuscript; D.V.R., J.G., G.N.M., N.B.W., A.M.P., C.H.L., M.J.P., K.S.C., and C.M.Y. approved final version of manuscript.