The Super-Relaxed State and Length Dependent Activation in Porcine Myocardium

Abstract

Rationale:

Myofilament length dependent activation (LDA) is the key underlying mechanism of cardiac heterometric autoregulation, commonly referred as the Frank-Starling law of the heart. Although alterations in LDA are common in cardiomyopathic states, the precise structural and biochemical mechanisms underlying LDA remain unknown.

Objective:

Here, we examine the role of structural changes in the thick filament during diastole, in particular changes in the availability of myosin heads, in determining both calcium sensitivity and maximum contractile force during systole in permeabilized porcine cardiac fibers.

Methods and Results:

Permeabilized porcine fibers from ventricular myocardium were studied under relaxing conditions at short and long sarcomere length (SL) using muscle mechanics, biochemical measurements, and X-ray diffraction. Upon stretch, porcine myocardium showed the increased calcium sensitivity and maximum calcium activated force characteristic of LDA. Stretch increased diastolic ATP turnover, recruiting reserve myosin heads from the super-relaxed state (SRX) at longer SL. Structurally, X-ray diffraction studies in the relaxed-muscle confirmed a departure from the helical ordering of the thick-filament upon stretch which occurred concomitantly with a displacement of myosin heads towards actin, facilitating cross-bridge formation upon systolic activation. Mavacamten, a selective myosin-motor inhibitor known to weaken the transition to actin-bound power-generating states and to enrich the ordered SRX myosin population, reversed the structural effects of stretch on the thick-filament, blunting the mechanical consequences of stretch; mavacamten did not, however, prevent other structural changes associated with LDA in the sarcomere, such as decreased lattice spacing or troponin-displacement.

Conclusions:

Our findings strongly indicate that in ventricular muscle, LDA and its systolic consequences are dependent on the population of myosin heads competent to form cross-bridges and involves the recruitment of myosin heads from the reserve SRX pool during diastole.

Graphical Abstract

INTRODUCTION

The force of ventricular contraction during a heartbeat is precisely modulated by its preload, established during diastole. This process of heterometric autoregulation, commonly referred as "Frank-Starling Law of the Heart" has as its underlying cellular mechanism events in the sarcomere collectively termed "myofilament length dependent activation" (LDA)¹. At the cellular level, ventricular filling stretches the myofibrils thereby increasing the sarcomere length (SL) as preload increases. Increased SL leads to both increased sensitivity to calcium and maximum contractile force during systole, the hallmarks of LDA. Substantial experimental evidence indicates that LDA is perturbed in cardiomyopathic disease states, such as hypertrophic cardiomyopathy (HCM) or heart failure^2-4. Despite its key role in both normal and pathological cardiac function, however, the molecular mechanisms underlying LDA are still not understood.

Much work has been done (reviewed in^{1, 5}) towards elucidating the complex sarcomeric mechanisms that govern LDA. In these efforts, small-angle X-ray diffraction has been an important tool for investigating the mechanistic role of structural changes in the sarcomere^{1, 6-8}. A key component of the length transducing mechanism is proposed to be the increase in titin-based strain of the thick filament with increasing SL. In diastole, this appears to manifest itself as changes in the ordering of the myosin heads on the surface of the thick filaments. This is accompanied by changes in the troponin complex on the thin filaments that may increase their sensitivity to calcium⁹, implying a communication pathway between thick and thin filaments under diastolic conditions. The increase in systolic force at longer lengths, can also be explained, at least in part, by a strain-dependent model of thick filament activation¹⁰ whereby thick filament activation during systole depends on positive feedback between the stress on the filament and a graded process of switching "ON" of individual myosin motors^{6, 7} from a low energy-consuming "OFF" state. A key concept in this model is that the number of motors available for force production is related to the degree of strain on the thick filament and thus to the loading conditions of the sarcomere.

Under normal physiological conditions, only a small fraction of all available myosin motors engages with actin during a cardiac cycle; the remainder are either unattached but readily available to form cross-bridges, in a state known as the disordered-relaxed (DRX) state, or are sequestered in an ordered, energy-sparing “OFF” state with ultra-low ATP consumption rates (~0.003 s⁻¹), known as the super-relaxed (SRX) state^11-13. Alterations in the balance between these states are thought to modulate myofilament function in disease. For instance, mutations in genes encoding cardiac sarcomeric proteins are known to contribute to hyper-contractility in HCM by releasing excess myosin heads from the SRX state (reviewed in^{14, 15}) thus favoring the formation of excess cross-bridges and increasing overall ATP consumption.

Here we use permeabilized Yucatan mini-pig myocardium to interrogate the diastolic functional, biochemical, and structural basis of LDA using mechanical measurements and small angle X-ray diffraction and compared our findings to published studies on rodent myocardium. While studies of rodent myocardium have provided invaluable insights into our understanding of cardiac physiology, findings from these systems may not be fully extrapolatable to humans. Porcine hearts, however, are closer in size, have similar heart rates as well as similar myosin isoform composition to human hearts; Porcine myocardium delivers excellent x-ray diffraction patterns that have been used to interrogate thick-filament/SRX dysregulation in disease¹⁶. Our mechanical results indicate that permeabilized porcine myocardium shows the increased calcium sensitivity and maximum calcium activated force characteristic of LDA. Structural results from X-ray diffraction showed that unlike findings in intact rat myocardium^{6, 9, 17} myosin heads were shifted towards actin at longer SL accompanied by a loss of helical order of the heads in permeabilized porcine myocardium. This observation was confirmed biochemically, as we show that stretch increases the number of myosin heads in the DRX state favoring cross bridge formation, while depopulating the SRX state at longer SL. In agreement with previous studies from rat myocardium, however, increasing SL results in an increase in the intensity of the third order troponin based meridional reflection (Tn3), and an increase in the M2 forbidden reflection from the thick filaments, in concert with titin-based passive strain in the myofilaments⁹.

To further elucidate the role of diastolic shifts in myosin head populations in LDA, we utilized mavacamten, a small-molecule cardiac myosin modulator that has been previously shown to prevent the transition of myosin to power-generating, strongly actin-bound states while stabilizing the SRX of myosin. Biochemical results show that mavacamten increases the population of heads in the SRX at both long and short SLs, while reducing the recruitment of DRX heads with stretch. Similarly, mavacamten blunted the outward movement of myosin heads, preserving quasi-helical order at longer SL without affecting stretch-mediated changes in troponin-based reflections. As a result, mavacamten treatment eliminated diastolic tension and calcium sensitivity gains with stretch while blunting, but not eliminating, the changes in maximal force associated with LDA. Taken together, our findings suggest that increased passive strain on the thick filament at longer SL releases myosin heads from the SRX to the DRX state, making them available to interact with actin. Both the increased calcium sensitivity of the thin filament and the maximum force generated in systole (F_max) components of LDA, therefore, are dependent on the pool of myosin heads competent to form cross-bridges.

METHODS

Data Availability.

All reduced data and supporting materials have been provided with the published article. See Supplemental Methods for details. The raw data are available from the corresponding author upon request.

Muscle Mechanics.

Experiments used hearts from young (3 to 5-month-old) male Yucatan mini-pigs. Humane euthanasia and tissue collection procedures were approved by the Institutional Animal Care and Use Committees at Exemplar Genetics. Fiber bundles from left-ventricular papillary muscles from fresh Yucatan mini-pig hearts were prepared as described previously^{16, 18}. Papillary muscle bundles (4 mm x 150-250 μm) were permeabilized in high-relax solution (HR) containing 1% Triton X-100. Preparations were mounted between a force transducer and high-speed motor and adjusted to either short ("Short", 2.0 μm) and long ("Long", 2.3 μm) SLs measured using a video SL system. Muscle preparations were sequentially exposed to eight increasing calcium concentrations first at Short and then at Long SL in the absence or presence of mavacamten (+MAVA, at 1 μM). Upon reaching steady-state force at each pCa, a small stretch was performed (3% increase in fiber length over 250 ms) to evaluate the stiffness of the fibers¹⁹. This maneuver produces a biphasic tension response: an initial, steep, highly calcium-sensitive, increase (Phase I) that then resolves into a second region (Phase 2) with a much shallower slope (figure I) after an abrupt “break point” at the end of the stretch. The first part of the biphasic force response is used to estimate active stiffness¹⁹ and the second component used to estimate Phase 2 stiffness (see Supplemental methods figure I). Phase 2 stiffness was plotted vs. tension and fitted to an exponential function (Y=Y0*exp(k*x). Phase 2 stiffness at zero tension (Y0), was used as an estimate of the stiffness due to passive elements in the fiber. All experiments were done at 22 °C.

LDA/SRX Assay.

Permeabilized myocardial fibers mounted on a on a custom setup modified from¹¹ and incubated in HR solution containing 250 μM (2’/3’-O-N-methylanthraniloyl ATP) (MANT-ATP) for 10 min before washout with relaxing solution containing unlabeled ATP¹¹. Fluorescence images were captured every 5 seconds for 6 minutes for each of Short SL, Long SL, Short SL plus mavacamten and Long SL plus mavacamten. The rate of MANT-ATP turnover as it is hydrolyzed by myosin and replaced by excess unlabeled ATP, was calculated by fitting background-corrected fluorescence to a 2-phase exponential decay curve to estimate the relative proportions of DRX and SRX heads²⁰. The area under the MANT-ATP turnover curve (AUC) was also calculated as an inverse estimate of the ensemble ATPase activity.

X-ray Diffraction.

Small-angle X-ray diffraction experiments were performed at the BioCAT beamline 18ID at the Advanced Photon Source, Argonne National Laboratory²¹. Permeabilized myocardial preparations (described above) were mounted in a trough containing HR solution. SLs were set to either short ("Short", 2.0 μm) and long ("Long", 2.2 μm) SL using light diffraction with a helium-neon laser (633 nm). X-ray patterns were collected from one of four protocols: 1): muscles initially held at Short SL, followed by Long SL and Long SL plus mavacamten; 2): muscles initially held at Short SL, followed by Short SL plus mavacamten; 3): muscles initially held at Short SL, followed by Short at pCa4; 4): muscles initially held at Short SL, followed by Long SL and Long SL at pCa4 from the same preparations in a paired manner. The data were analyzed as described previously^{22, 23}. All experiments were done at 23 °C.

Statistics.

Non-parametric statistical analyses were performed using GraphPad Prism 7 (Graphpad Software). The results are given as mean ± SEM. Mechanical data were analyzed by Wilcoxon matched-pairs test (figure 1, and figure 3) and unpaired Mann-Whitney test (figure 2). One-way Brown-Forsythe and Welch ANOVA with Dunnett’s T3 multiple comparisons test with individual variances computed for each comparison was performed on SRX assays (figure 4). All the X-ray data was analyzed by Wilcoxon matched-pairs test except for the data in figure 5D which were analyzed by Mann-Whitney test. Symbols on figures: ns: p>=0.05, *: p<0.05, **: p<0.01, ***: p<0.001 and ****: p<0.0001.

Figure 1. Stretch effects on tension/pCa parameters.

A: Normalized tension/pCa curves at Short SL (closed, solid line) and Long SL (open, dotted line) for porcine myocardium. B. Normalized maximal tension (Tmax) response to stretch; Short SL (black bars) and Long SL (grey bars). C. Ca²⁺ sensitivity response (pCa50) to stretch; Short SL (black bars) and Long SL (grey bars). D. Ca²⁺ cooperativity response (Hill slope; n) to stretch; Short SL (black bars) and Long SL (grey bars). (n =8) *P<0.05, **P<0.01, ***P<0.001.

Figure 3. Stiffness effects of stretch and mavacamten on porcine muscle fibers.

A: Active stiffness (passive normalized) vs. pCa at Short SL (black circles, solid line) and Long SL (grey circles, dotted line). B: Phase 2 Stiffness vs. Tension at Short SL (black circles, solid line) and Long SL (grey circles, dotted line). C: Active Stiffness (normalized by passive stiffness) at pCa5.8 for Short SL (black bars) and Long SL (grey bars, n = 8). D. Active Stiffness (normalized by passive stiffness) at pCa5.8 for Short SL (black red hatched bars) and Long SL (grey red hatched bars) in the presence and absence of 1 μM mavacamten (n = 7). E: Passive stiffness (Y0; stiffness at T0) at Short SL (black bars) and Long SL (grey bars) in the presence (n = 7) and absence (n = 8) of 1 μM mavacamten (red hashes). *P<0.05, **P<0.01, ***P<0.001.

Figure 2. Stretch effects of mavacamten on parameters of LDA.

A: Normalized tension/pCa curves showing the effects of 1 μM mavacamten at Short SL (A; solid line) and Long SL (B; dotted line) for porcine myocardium in the absence (black) and presence of 1 μM mavacamten (red). Mava reduced maximal tension at Short and Long SL by 32 and 33%, respectively. C. Percent change with stretch (vs. Short SL) of Tmax for untreated (Long SL; black bars) and fibers in the presence of 1 μM mavacamten (Long SL + Mava; red bars). D. Percent change with stretch (vs. Short SL) of Ca²⁺ sensitivity (pCa₅₀) for untreated (Long SL; black bars) and fibers in the presence of 1 μM mavacamten (Long SL + Mava; red bars). E. Change in Hill slope with stretch (vs. Short SL) of Hill slope for untreated (Long SL; black bars) and in the presence of 1 μM mavacamten (Long SL + Mava; red bars). F. Percent change with stretch (vs. Short SL) of submaximal tension (tension at pCa6.4) for untreated (Long SL; black bars) and fibers in the presence of 1 μM mavacamten (Long SL + Mava; red bars). (n = 8). *P<0.05, **P<0.01, ***P<0.001.

Figure 4. Effects of stretch and mavacamten on MANT-ATP dissociation on pig myocardium.

A: Timecourse of MANT-ATP dissociation with stretch for native pig myocardium (short SL (black); long SL (grey). B: Effects of 1 uM mavacamten on MANT-ATP dissociation at short SL (short, untreated (black); short, 1 uM mavacamten (red)). C. Effects of stretch and mavacamten on MANT-ATP dissociation at short and long SL's in the presence and absence of 1 uM mavacamten (short SL, untreated (black); long SL, untreated (grey); short SL, 1 uM mavacamten (red); long SL, 1 uM mavacamten (pink). For figures A, B, and C, dashed lines represent relative Half-life of the Fast and Slow phases (black, untreated; red, 1 uM mavacamten). D: Area under the curve (AUC) of the MANT-ATP dissociation timecourse (through 600s) with stretch in the absence and presence of 1 uM mavacamten (short SL, untreated (black bars); long SL untreated (grey bars), 1 uM mavacamten (red hashes). DRX population (%) (E) and DRX ATPase rate (s⁻¹) (F) of MANT-ATP dissociation with stretch in the absence and presence of 1 uM mavacamten (short SL, untreated (black bars); long SL untreated (grey bars), 1 uM mavacamten (red hashes). SRX population SRX (%) (G) and SRX ATPase rate (s⁻¹) (H) of MANT-ATP dissociation with stretch in the absence and presence of 1 uM mavacamten (short SL, untreated (black bars); long SL untreated (grey bars), 1 uM mavacamten (red hashes). (n = 8 for Short SL, n = 6 for Long SL, n = 4 for Short SL + Mava and Long SL + Mava. **: p<0.01, ***: p<0.001 and ****: p<0.0001.

Figure 5. Effects of stretch and mavacamten on equatorial reflections on permeabilized pig myocardium.

A: 2-D diffraction pattern from permeabilized resting pig myocardium. B: equatorial intensity ratio from Short SL to Long SL permeabilized pig myocardium in the absence and presence of 50 μM mavacamten (n = 31 for Short SL, n = 18 for Long SL, n = 13 for Short SL+ Mava and n = 10 Long SL+ Mava). C: lattice spacing from Short SL to Long SL permeabilized pig myocardium in the absence and presence of mavacamten (n = 38 for Short SL, n = 20 for Long SL, n = 18 for Short SL+ Mava and n = 14 Long SL+ Mava). D: equatorial intensity ratio from Short SL (n = 9) to Long SL (n = 6) permeabilized pig myocardium at pCa 4. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

RESULTS

Muscle Fiber Mechanics.

Permeabilized porcine muscle preparations showed a left-upward shift in the tension/pCa relationship with stretch (figure 1A). A stretch of 14.5 ± 0.6% was required to increase SL from Short to Long (inset, figure 1A). This stretch was associated with increases in maximal force (Normalized tension (% Max): 99.3 ± 0.6 to 111.8 ± 4.2%, p<0.05, figure 1B), calcium sensitivity (pCa₅₀: 5.802 ± 0.015 to 5.932 ± 0.026, p<0.01, figure 1C), and Hill slope (n: 2.86 ± 0.23 vs. 3.65 ± 0.30, p<0.01, figure 1D), all hallmarks of LDA. Stretch also increased submaximal (at pCa6.4) tension, a surrogate for end-diastolic tension in vivo (figure 2F).

As expected, mavacamten (at Short SL) decreased the tension produced at each pCa (figure 2A) reducing maximal tension (32%) without affecting pCa₅₀ or Hill slope (data not shown). Mavacamten blunted the gain in tension associated with stretch with no change in the ability of mavacamten to reduce maximal tension (32% at Short SL vs. 33% at Long SL, figure 2B). Quantitatively, mavacamten abolished the effects of stretch on both pCa₅₀ (LDA pCa50: −8.7 ± 5.2% vs. 35.6 ± 5.6% change from Short SL; p<0.001, figure 2D) and submaximal tension (tension @ pCa6.4: −3.1 ± 9.3 vs. 131.4 ± 40.2 % change from Short SL; p<0.01, figure 2F). There was no significant effect of mavacamten on LDA parameters of relative changes in maximum tension (T_max) (+12.64 ± 4.24 vs. +6.24 ± 5.19 %Change from Short SL, p=0.356) or Hill slope (0.79 ± 0.20 vs. 0.63 ± 0.18 % change from short SL, p=0.574).

Stiffness.

Active and passive stiffness was estimated using a stretch-hold protocol ¹⁹ as described in the methods. Stretched muscle fibers have higher active stiffness (see Supplemental Methods) from pCa6.2-5.4 (all p<0.05, figure 3A), suggesting a stretch-mediated increase in the number of active (systolic) cross-bridges and which was abolished by myosin inhibition with mavacamten. For example, at the pCa5.8 (the pCa₅₀ of the porcine muscle fiber), stretch increases stiffness (491.5 ± 55.5 vs. 847.4 ± 108.0 kN/m², p<0.05, figure 3C) while in the presence of mavacamten, stiffness is unchanged (350.0 ± 67.8 vs. 278.7 ± 79.6 kN/m², p=0.62, figure 3D).

Stretch markedly increased passive stiffness (50.5 ± 6.5 to 244.1 ± 22.8 kN/m², p<0.01, figure 3E), as estimated by extrapolation of the phase 2 stiffness to zero tension (figure 3B), suggesting that there are structural changes in the fiber, possibly due to titin or to increased inter-filament interactions, leading to increased passive stiffness at Long SLs. While calcium independent acto-myosin interactions have been proposed as significant contributors to passive stiffness at physiological temperature (37°C), these interactions are minimal at the temperature used for these studies (22 °C) in the absence of dextran ^{24, 25}, so are unlikely contributors to this effect. Mavacamten further increased passive stiffness at Long SL (244.1 ± 22.8 vs. 359.3 ± 29.5 kN/m², p<0.05, figure 3E), suggesting that the enhanced interactions of the myosin heads with the thick filament backbone induced by mavacamten stiffens the thick filament.

Loaded MANT- ATP Turnover Rates.

Upon stretching porcine myocardium, there was a leftward shift in the MANT-ATP turnover curve (figure 4A), resulting in a significant decrease in the area under the curve (AUC) (66.3 ± 1.2 vs. 53.62 ± 0.9 fluorescence*s, Short vs. Long SL respectively, figure 4D, p < 0.0001), indicating increased ATPase activity under relaxing conditions at Long SL. With stretch, the fast phase of MANT-ATP turnover shows a significant increase, indicating an increase in the population of myosin heads in the disordered relaxed state (80.5 ± 0.6 vs. 84.6 ± 0.4 %, at Long SL, figure 4E, p = 0.0005). A modest increase in the ATPase rate of these heads was also observed (0.044 ± 0.0004 vs. 0.059 ± 0.001 s⁻¹, at Long SL, figure 4F, p < 0.0001), possibly due to the reduction in lattice spacing at long SL. Stretch decreased the population of myosin heads in the SRX state (15.4 ± 0.5 vs. 12.0 ± 0.4 % at Long SL, figure 4G, p = 0.0005), suggesting the recruitment of myosin with stretch. No changes in the intrinsic SRX ATPase rate were observed (0.0064 ± 0.0003 vs. 0.0064 ± 0.0003 s⁻¹, Short vs. Long SL, figure 4H, p > 0.99).

Mavacamten significantly slowed MANT-ATP turnover (figure 4B and 4C for Short and Long SL respectively) resulting in significant increases in the AUC at both Short (66.3 ± 1.2 vs. 93.6 ± 1.8 fluorescence*s, p < 0.0001) and Long SL (53.6 ± 0.9 vs. 76.2 ± 1.3 fluorescence*s, p < 0.0001) (see figure 4D), consistent with a decrease in diastolic ATPase activity. Mavacamten decreased the DRX population and the ATPase rates at both SLs (DRX population: 80.5 ± 0.6 vs. 75.9 ± 1.2 % at Short SL (p = 0.014) and 84.6 ± 0.4 vs 77.5 ± 0.7 % at Long SL (p = 0.014), figure 4E; DRX ATPase rate: 0.0441 ± 0.0004 vs. 0.0332 ± 0.001 s⁻¹ at Short SL (p = 0.0004) and 0.0587 ± 0.001 vs. 0.0658 ± 0.001 s⁻¹ at Long SL (p = 0.03), figure 4F). Conversely, mavacamten increased the SRX population (15.4 ± 0.5 vs. 23.4 ± 0.7 % at Short SL (p = 0.0005) and 12.0 ± 0.3 vs. 17.1 ± 0.5 % at Long SL (p = 0.0008), figure 4G) and reduced the SRX ATPase rate (0.0064 ± 0.0003 vs. 0.0029 ± 0.0005 s⁻¹ at Short SL (p = 0.004) and 0.0064 ± 0.0003 vs. 0.005 ± 0.0004 s⁻¹ at Long SL (p = 0.19), figure 4H), consistent with a stabilization of low-energy consuming myosin states.

Despite the increased population of myosin heads in the SRX, and a reduction in the DRX population with mavacamten, stretch decreased the AUC of MANT-ATP turnover both in the presence and absence of mavacamten (both p<0.01) suggesting that stretch is still able to recruit reserve SRX myosin heads after mavacamten treatment.

Changes in Equatorial X-ray Reflections with Stretch and Mavacamten.

Permeabilized porcine myocardium produced high quality X-ray diffraction patterns (figure 5A) showing strong equatorial and meridional reflections as well as well-developed myosin-based layer lines in resting muscle. The equatorial intensity ratio (I_1,1/I_1,0) is widely used as an indicator of the proximity of myosin heads to actin^{22, 26}. The I_1,1/I_1,0 intensity ratio increased (figure 5B) when muscles were stretched at pCa8 from Short SL (0.26 ± 0.01) to Long SL (0.31 ± 0.02) (p<0.01). This change in I_1,1/I_1,0 upon stretch indicates a shift of mass from the thick-filament backbone, presumably myosin heads, towards actin at longer SL. These effects, due to increases in SL under diastolic conditions, were blunted by mavacamten (50 μM) treatment (I_1,1/I_1,0 = 0.28 ± 0.02 at Long SL, p<0.05), a compound known to bring myosin heads to adopt a configuration more tightly associated with the thick filament backbone in a highly ordered state, presumably the SRX state. Lattice spacing (d_1,0) decreased from 41.5 ± 0.3 nm at Short SL to 39.8 ± 0.3 nm at Long SL (p<0.0001) and was further decreased to 38.6 ± 0.3nm with mavacamten treatment (p<0.001) (figure 5C). While a reduction in lattice spacing with increasing SL is expected from previous studies⁹ the further reduction in lattice spacing with mavacamten could be for multiple reasons including reduced electrostatic repulsion between the myofilaments when myosin heads are held close to the thick filament backbone^{27, 28}, and/or structural inter-filament interactions potentially mediated by myosin-binding protein C (MyBP-C)²⁹.

The I_1,1/I_1,0 intensity ratio in permeabilized and intact contracting mammalian skeletal^{22, 30} and cardiac muscle^{31, 32} is closely correlated to the number of attached, force producing cross-bridges. Figure 5D shows that I_1,1/I_1,0 is larger in contracting muscle preparations at Long SL (0.86 ± 0.09) than in preparations at Short SL (0.62 ± 0.04, p < 0,05) indicating that more cross-bridges are associated with actin at longer SL during contraction, consistent with the increased force.

Changes in Meridional X-ray Reflections with Stretch and Mavacamten.

The meridional reflections, as shown in figure 6A, report on axially repeating structures in the myofilaments. The intensity of the third order myosin meridional reflections (I_M3), primarily arising from axially adjacent myosin heads, decreased ~12% when muscle length increased from Short to Long SL (p<0.05). In contrast, I_M3 increased relative to either Short or Long SL, after mavacamten treatment (figure 6B, p<0.05). There were no significant changes in the axial spacings of the M3 reflection with changes in SL or mavacamten treatment.

Figure 6. Effects of stretch and mavacamten on meridional reflections on permeabilized pig myocardium.

A: low orders of meridional reflections from permeabilized resting pig myocardium. Intensity of the third order myosin meridional reflections (B) and the second order myosin meridional reflections (C), from Short SL (n = 18 for I_M3, and n = 21 for I_M2), Long SL (n = 18 for I_M3, and n = 21 for I_M2), and Long SL + Mava (n = 13 for I_M3, and n = 14 for I_M2). D: Spacing of the sixth order myosin meridional from Short SL (n = 20) Long SL (n = 20) and Long SL + Mava (n = 15). E. The intensity of the second (right) and third (left) order troponin meridional reflections from Short SL (n = 19 for I_Tn3, and n = 20 for I_Tn2), Long SL (n = 19 for I_Tn3, and n = 20 for I_Tn2), and Long SL + Mava (n = 14 for I_Tn3, and n = 9 for I_Tn2). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

The second order myosin meridional reflection (M2) is one of the so-called "forbidden" reflections (figure 3A) that arise from a regionally distorted surface lattice, most likely corresponding to the D-zone of the thick filament⁸, of myosin heads on the thick filament backbone^{33, 34}. The M2 intensity (I_M2) increased by ~21% when muscle length increased from Short to Long SL (figure 6C, p<0.01). There were no significant changes in the axial spacings of the M2 reflection with SL or mavacamten treatment. These results indicate an increased departure from strict helical ordering of the myosin heads around the thick filament with increased SL. The increase in I_M2 was reversed by mavacamten treatment (I_M2 ~ 40% less than at Long SL, p<0.01).

The M6 spacing (S_M6), whose intensity primarily comes from the backbone of the thick filament³⁵ increased by 0.4% when SL increases from short to long (figure 6D, p<0.01) indicating an increase in thick filament strain at longer SL. S_M6 decreased about 0.6% (relative to Long SL, p<0.001) after treatment with mavacamten, indicating either a reduction in the strain experienced by the thick filament or an inhibition of its extension in response to stretch i.e. increased backbone stiffness. The intensities of the M6 reflections (I_M6) did not change significantly with stretch.

The intensity of the third order troponin reflection (I_Tn3) increased by about 7% when SL increased from Short to Long SL (p<0.0001). There were no significant changes of I_Tn3 after mavacamten treatment at Long SL. The intensity of the second order troponin reflection (I_Tn2) decreases ~10% when SL increased from Short to Long SL (p<0.05) with no change with mavacamten treatment (figure 6E). These intensity changes indicate that there are alterations in the structure of the troponin complex with increasing SL and that the increased ordering of myosin heads by mavacamten had no effect on this process at Long SL. These results are consistent with those reported in intact twitching rat myocardium previously reported ⁹ that suggest that titin-based passive strain dependent structural changes in the thick filament are transmitted to the thin filaments as part of the mechanism of LDA.

Myosin Layer Lines Changes with Stretch and Mavacamten.

The quasi-helical arrangement of myosin heads around the thick filament backbone gives rise to the myosin-based layer lines parallel to the equator (figure 7A). The distance from the first maximum in the intensity distribution on a layer line to the meridian is inversely related to the radius to the center of mass of the cross-bridges (R_m)²².The first maxima on the first myosin layer line moved inwards towards the meridian when muscle was stretched from a short SL (black line in figure 7B) to a long SL (red line in figure 7B). This implies that R_m increased when SL increased from Short (14.6 ± 0.12 nm) to Long (15.0 ± 0.12 nm) (p<0.05) SL. This change in R_m caused by stretch appeared to be reversed by mavacamten treatment (figure 7C, R_m = 14.83 ± 0.15 nm) but this change was not significant, possibly due to an insufficient number of measurements (n=5, P=0.31)

Figure 7. Effects of stretch and mavacamten on myosin layer lines on permeabilized pig myocardium.

A. First order myosin-based layer lines from permeabilized resting pig myocardium. (B), intensity profile of first order myosin-based layer lines (inside white box in panel A) from Short SL (black) to Long SL (red). D: The center of mass of the cross-bridges (R_m) from Short SL, Long SL permeabilized pig myocardium in the absence and presence of 50 μM mavacamten. E. Intensity of first order myosin-based layer lines from Short SL, Long SL permeabilized pig myocardium in the absence and presence of 50 μM mavacamten. (n = 11 for Short SL, n = 6 for Long SL, n = 5 for Short SL + Mava and n = 5 for Long SL). *P<0.05, **P<0.01.

The MLL1 intensity (I_MLL1) tracks the number of ordered myosin heads and a decrease in I_MLL1 has been proposed as an indicator of the muscle thick filament transitioning to an ON state¹⁰. The normalized I_MLL1 decreased from 2.65 ± 0.11 at a Short SL to 1.94 ± 0.17 at Long SL (p<0.05) indicating a loss of ordering of myosin heads upon stretch and a regain of ordering after mavacamten treatment (3.34 ± 0.25, p<0.01) (figure 7D).

DISCUSSION

Here we investigated the role of the structural changes in the thick filament, in particular, recruitment of reserve myosin heads during diastole in myofilament length dependent activation (LDA) in porcine cardiac muscle. We show that healthy porcine left-ventricular permeabilized myocardium presents not only robust LDA, with increased maximal tension (T_max), calcium sensitivity (pCa₅₀), and cooperativity (increased Hill coefficient, n_H) with stretch, but also yields high quality X-ray diffraction patterns that can be used to elucidate the structural alterations that occur with stretch. We demonstrate, both structurally and biochemically, as well as by leveraging a clinical-stage myosin-inhibitor, that LDA and its systolic consequences in ventricular muscle are dependent on the population of myosin heads competent to form cross-bridges and involves the diastolic recruitment of myosin heads from the reserve SRX pool.

Order/disorder transitions in the myosin heads and LDA in porcine myocardium.

Here we showed that permeabilized porcine myocardium shows many of the X-ray diffraction signatures of an OFF to ON transition of the thick filament¹⁰, when muscle is stretched under diastolic conditions (pCa8). The increase in S_M6 at longer SL indicates an increase in thick filament strain, resulting from increases in titin-based passive tension with SL⁹ but not the full OFF to ON structural transition characterized by a much larger spacing change seen when calcium activated. However, changes in the relative distributions of myosin heads between SRX and the DRX states with stretch are indicated by the decrease in I_MLL1 at Long SL. An increase in I_1,1/I_1,0 at Long SL is also consistent with an order-disorder transition that allows myosin heads to move closer to the thin filaments, thus facilitating cross-bridge formation. Similarly, the intensity of the M3 reflection, I_M3, decreases at Long SL, also consistent with an order-disorder transition. These observations are supported by the MANT-ATP turnover experiments (figure 4) that indicated increased ATP turnover at Long SL with a reduction of myosin heads in the SRX state and, hence, an enrichment of the pool of disordered heads. All these results indicate that increases in SL in relaxed permeabilized porcine myocardium trigger a reduction in the population of heads closely associated with the thick filament backbone and a concomitant gain in the pool of heads readily available to form cross-bridges.

Effects of Mavacamten on Length Dependent Activation.

Here we show that at physiologically relevant concentrations, mavacamten blunts the gains in systolic tension and calcium sensitivity that occurs with increasing SL. Since mavacamten is known to stabilize the SRX state, this observation further supports the notion that transitions in and out of the DRX/SRX states contribute to increased calcium sensitivity at longer lengths characteristic of LDA.

The MANT-ATP turnover curves shown in figure 4 indicate that stretch induces a transition from the SRX to a disordered relaxed state. Figure 4 also shows that mavacamten enriches the population of heads in the SRX state, consistent with previous data supporting stabilization of the SRX state¹⁶. In the X-ray diffraction experiments shown in figures 5 - 7, mavacamten (50 μm) restores almost all the thick filament based structural alterations with stretch back to the non-stretched state, further confirming that stretch mechanically releases myosin heads from the ordered SRX state to a disordered relaxed state with a reversal of this transition with mavacamten.

Interestingly, although mavacamten treated myocardium shows increased numbers of heads in the SRX at both Short and Long SLs, the relative transitions from the SRX to the DRX states with stretch were comparable in the presence and absence of mavacamten. Our results from permeabilized porcine myocardium are similar to recent results from permeabilized human myocardium³⁶ in that both studies show some degree of systolic LDA with increasing SL after mavacamten treatment. While both studies observed an increase in T_max with increasing SL, Awinda et al. observed an increase in calcium sensitivity, our data indicate a decrease. Together, these data indicate that either mavacamten-stabilized SRX heads remain sensitive to titin-based mechanical strain in the thick filaments at longer SL⁹ but cannot contribute to force or that the mavacamten-stabilized heads at resting lengths are only a subset of the population recruited by stretch. The latter possibility appears more likely since at the mavacamten concentration used, maximal force is merely blunted, not abolished, indicating a residual population of heads not stabilized in the SRX by mavacamten under our experimental conditions.

A recent study from Nelson et al³⁷ showed that, at least in soleus muscle myofibrils, mavacamten primarily affected myosin heads in the P and D zones of thick filaments, where MyBP-C is absent, with little effect on heads in the C-zone where MyBP-C is present. This implies that while heads in the D zone in mavacamten treated muscle are held closely associated with the backbone and presumably insensitive to stretch, the quasi-helically arranged heads in the C zone are free to be released by stretch. If this behavior were to be confirmed in cardiac muscle, it would provide an explanation for the coexistence of both stretch sensitive and stretch insensitive populations of myosin heads on the same thick filaments.

Why are there differences between intact rodent and permeabilized porcine myocardium?

A remarkable finding of the current study is that stretch under relaxed conditions (at pCa8) translocated the mass of myosin heads in permeabilized porcine myocardium away from the thick filament (increased I_1,1/I_1,0 and R_m), likely reflecting the observed shifts in the populations of myosin heads between the SRX and the DRX states as indicated by the stretch-induced acceleration in MANT-ATP turnover kinetics. One might expect a priori, consistent with in silico studies³⁸, that an outward radial shift of cross-bridge mass along with the increased availability of myosin heads with stretch during diastole could serve to explain the increased calcium sensitivity and F_max of LDA. These observations are, however, inconsistent with those reported for intact rat myocardium, either ex vivo (decrease)^{6, 7, 9, 17} or in vivo (minimal change)³⁹. While the temperature at which the experiments were done could potentially be a confounding variable, the small differences in temperature between our experiments (at 22-23°C) and those of Ait-Mou et al (25°C)⁹ or Caremani et al (27°C)¹⁷ are unlikely to significantly affect our interpretations.

At present there is no obvious explanation for the discrepancy between intact rat and permeabilized porcine myocardium. It is clear, however, that the two preparations are not equivalent. In Caremani et al.’s experiments on intact twitching rodent myocardium¹⁷, X-ray signatures of an OFF to ON transition in the thick filaments were only evident under systolic and not in diastolic conditions. Here many of the signatures of the OFF to ON transition, not seen in resting intact muscle, were observed in permeabilized, relaxed porcine myocardium. One difference could be that the diastolic calcium concentration in intact twitch rodent myocardium at a low pacing rate (~1 Hz) is 10-fold higher (pCa7)⁴⁰ than with our experiments that were performed at pCa8. It has been argued⁵ that the interactions maintaining myosin heads associated with the thick filament backbone in intact muscle in diastole may become weaker at longer SL but not enough to be detected as changes in the X-ray patterns. Such weakened interactions may only be evident as large-scale structural changes in permeabilized muscle because the system is perturbed away from the complex intracellular milieu and reduced lattice spacing in vivo⁵. Weakened head-backbone interactions in intact myocardium, although not detectable in X-ray patterns under diastolic conditions, could result in increased strain dependent release of myosin heads and subsequent increases in F_max in systole¹⁷. Future studies of intact porcine myocardium could help resolve these species-specific differences in behavior.

Conclusions.

Permeabilized porcine myocardium shows robust LDA that is blunted by the cardiac myosin inhibitor mavacamten. In our studies of relaxed permeabilized porcine myocardium, both the equatorial intensity ratio (I_1,1/I_1,0), the radial position of myosin heads increased at longer SL. Similarly, the intensities of the third order myosin meridional reflection (I_M3) and the first myosin layer line (I_MLL1) decreased, indicating a loss of ordering of myosin heads around the thick filament backbone. These observations, when coupled with a stretch-mediated increase in MANT-ATP disassociation, strongly indicate that during stretch at diastole myosin heads are released from the SRX state and move away from thick filament backbone and become disordered, readily available to form cross-bridges.

Most of the thick filament structural and mechanical changes induced by stretch were reversed by mavacamten, a compound that known to enrich the population of myosin heads in an ordered SRX state and prevent the transition of myosin heads to actin-bound, force-generating states. Mavacamten treatment in porcine myocardium not only increases the population of myosin heads in the SRX state but also stiffens the thick filaments. Mavacamten treatment blunted both changes in calcium sensitivity and in F_max along with the thick filament based structural changes without blunting the stretch-induced changes in troponin reflections and lattice spacing. Both the increased calcium sensitivity of the thin filament and the maximum force generated in systole (F_max) in LDA are dependent on the pool of myosin heads competent to form cross-bridges and involves the diastolic recruitment of myosin heads from the reserve SRX pool.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• The Frank Starling law of the heart has its cellular basis in myofilament length-dependent activation (LDA).

• LDA is the observation that when sarcomeres are passively stretched both the amount of force at a given calcium concentration and the maximum force increases.

• Myosin heads can adopt an ordered, low-energy consuming state known as the super-relaxed (SRX) state in diastole

What New Information Does This Article Contribute?

• Both the calcium sensitivity and the increased maximum force characteristic of LDA depend on the available pool of myosin heads competent to form crossbridges.

• Myosin heads are recruited from a low-energy consumption state known as the super-relaxed (SRX) state when muscle is stretched during diastole.

In the healthy heart the force of contraction is intrinsically controlled by the level of preload exerted on the muscle (i.e., greater force at greater stretch), in a phenomenon commonly referred as the "Frank-Starling Law of the Heart" which is based on a process of sarcomeric heterometric autoregulation termed "myofilament length dependent activation" (LDA). LDA is traditionally associated with increases in calcium-sensitivity and thin-filament activation at greater muscle/sarcomere lengths. This work demonstrates, both structurally and biochemically, as well as by leveraging a novel clinical-stage myosin-inhibitor (mavacamten), that LDA in ventricular mammalian muscle (from pigs) is also dependent on the population of myosin motors competent to form cross-bridges and involves the diastolic recruitment of reserve myosin heads from a low-energy consumption state known as the super-relaxed (SRX) state. As LDA is altered in disease states such as hypertrophic cardiomyopathy and heart failure, these findings not only broaden the physiological understanding of intrinsic force production and energy utilization in the heart but suggest thick-filament targeted interventions as potential new therapeutic avenues for these pathological states.

Nonstandard Abbreviations and Acronyms:

LDA

Length dependent activation

SL

Sarcomere length

HCM

Hypertrophic cardiomyopathy

DRX

Disordered-relaxed

SRX

Super-relaxed

Tn3

Third order troponin reflection

F_max

Maximum force

HR

High-relax solution

AUC

Area under the MANT-ATP turnover curve

T_max

Maximum tension

I_1,1/I_1,0

The equatorial intensity ratio

d_1,0

Lattice spacing

I_M3

The intensity of the third order myosin meridional reflection

M2

The second order myosin meridional reflection

S_M6

The spacing of the sixth order myosin meridional reflection

I_Tn2

The intensity of the second order troponin reflection

R_m

Radius to the center of mass of the cross-bridges

I_MLL1

The intensity of first myosin based layer line