The Frank-Starling mechanism involves deceleration of cross-bridge kinetics and is preserved in failing human right ventricular myocardium

Human tissue collection.

Explanted hearts were obtained in the operating room, flushed immediately after removal from living donors, and thereafter transferred to the laboratory in cold cardioplegic solution containing (in mM) 110 NaCl, 16 KCl, 16 MgCl₂, 10 NaHCO₃, and 0.5 CaCl₂. Hearts were procured and treated with identical protocols and timing regardless of their source.

All human tissues were experimented on with approval from the Institutional Review Board (IRB) of The Ohio State University and in accordance with the Declaration of Helsinki. Informed consent was acquired from cardiac transplant patients. Nontransplantable donor hearts (n = 9) were acquired from Lifeline of Ohio Organ Procurement; ejection fraction of these hearts, measured prior to procurement, was in the normal range, and none of the nonfailing hearts had signs of coronary bypass surgery or prior myocardial infarctions, while wall thicknesses, heart weights, and heart weight-to-body weight ratios were all in the normal range. End-stage failing hearts (n = 10) were acquired from patients undergoing cardiac transplantation at The Ohio State University Wexner Medical Center. All hearts, regardless of source, were processed with the same protocol. The characteristics of these hearts are provided in Table 1 and Table 2. An additional four nontransplantable donor hearts (not included in the tables) were used only for assessing the effects of temperature on the rate of tension redevelopment (k_tr).

Trabecula isolation.

Within 30 min, the right ventricle of each heart was transferred from the cardioplegic solution to a cold modified Krebs-Henseleit solution (K-H) previously bubbled with 95% O₂-5% CO₂ containing (in mM) 137 NaCl, 5 KCl, 0.25 CaCl₂, 20 NaHCO₃, 1.2 NaH₂PO₄, 1.2 MgSO₄, 10 dextrose, and 20 2,3-butanedione monoxime (BDM) with a pH of 7.4. Small and linear right ventricular trabeculae were isolated and kept in this solution at 0–4°C until the time of the experiment (0–8 h). Muscles were placed in custom-made setups as previously described for animal models (37), and the perfusion solution was changed to another modified K-H without BDM and containing 0.25 mM CaCl₂. This solution was maintained at 37°C and continuously bubbled with 95% O₂-5% CO₂, resulting in a pH of 7.4. Stimulation was initiated at baseline frequency of either 0.5 Hz (muscles used to assess baseline cardiac function) or 1 Hz (muscles used for k_tr experiments), and the CaCl₂ concentration of the solution was incrementally and slowly raised to 2 mM over the time course of ∼15 min. Muscles were gradually stretched until an increase in developed force was not matched by an increase in resting tension. This length, designated as optimal length (L_opt), corresponds to sarcomere length of ∼2.2 μm, which is near or at the in vivo sarcomere length at end-diastole (41).

Baseline trabecula function.

Right ventricular trabeculae were stabilized at L_opt and 0.5 Hz for ∼15 min. In the endocardial wall of human hearts, the smallest trabeculae are virtually always highly fibrogenous. Good contracting trabeculae have a size that is typically just over 150 μm. This size, in combination with ultrastructure (more fibrotic than small rodents), does not allow for unambiguous detection of sarcomere length with laser diffraction (30). In very rare cases (outliers, not representative of the average endocardial trabecula) a measurable sarcomere length pattern can be seen at rest, but this inevitable disappears during contraction. Thus length variations are set and denoted in percentage of L_opt rather than absolute sarcomere length. After contractile and kinetic data were collected at this optimal length, the length-tension relationship was determined at constant frequency of 0.5 Hz from L₉₀ (90% of L_opt, corresponding to in vivo sarcomere length similar to the end-systolic phase) to L_opt. The force-frequency relationship (FFR) was determined at L_opt from 0.5 Hz to 3 Hz. The β-adrenergic response was determined by adding incremental amounts of isoproterenol for a final concentration of 1 μM at L_opt and 0.5 Hz. Data from multiple muscles of the same heart were averaged and used for the final analysis. Nonfailing muscles (n = 10 from 8 hearts) were 288 ± 32 μm in width, 191 ± 22 μm in thickness, and 3.7 ± 0.5 mm in length. Failing muscles (n = 13 from 9 hearts) were 538 ± 44 μm in width, 358 ± 29 μm in thickness, and 3.2 ± 0.4 mm in length. At the baseline stimulation rate of 0.5 Hz, muscle thickness did not adversely impact contractile strength; the correlation coefficient between muscle thickness and force (R²) was only 0.08. At higher frequency, this correlation was still very weak (R² = 0.28) and thus does not impact data significantly.

k_tr experiments in intact cardiac trabeculae.

Nonfailing right ventricular trabeculae (n = 9 from 9 hearts) used for k_tr experiments had an average width of 406 ± 50 μm, thickness of 267 ± 34 μm, and length of 2.7 ± 0.4 mm. The failing right ventricular muscles (n = 10 from 10 hearts) had an average width of 352 ± 26 μm, thickness of 234 ± 17 μm, and length of 3.3 ± 0.3 mm. Two of the muscles were previously subjected to the length-tension (at 0.5 Hz) and FFR (from 0.5 Hz to 3 Hz) experiments as described above. Muscles were stabilized at L_opt and 1 Hz for 10–15 min. In some of the muscles (nonfailing: n = 7 of 9, failing: n = 7 of 10), a brief FFR (1 Hz, 1.5 Hz, 2 Hz) was initially determined before any k_tr experiments. Thereafter, muscles were restabilized at 1 Hz for 10–15 min, and this frequency was used for the remainder of the experiments. Muscles were exposed to a high-K⁺/Ca²⁺, low-Na⁺ solution containing (in mM) 20.6 NaCl, 121.4 KCl, 6 CaCl₂, 20 NaHCO₃, 1.2 MgSO₄, 1.2 NaH₂PO₄, and 10 dextrose in order to induce a contracture (Fig. 1A). During this contracture, k_tr maneuver was performed at submaximal (a subset of muscles) and maximal (all muscles) tension levels as previously described for rats and rabbits (37). Figure 1B shows a typical k_tr tracing. The solution was switched back to K-H, muscles were stabilized at new lengths, L₉₅ (95% of L_opt) and L₉₀ (90% of L_opt), for 10–15 min, and k_tr experiments were performed at each of these lengths as described above. Thereafter, muscles were restabilized at L_opt, and k_tr experiments were performed at this length for a second time (designated L_opt,repeat) in order to compare to the first L_opt measurement and assess reproducibility.

The temperature dependence of k_tr was assessed at maximal K⁺ contracture tension in another set of nonfailing donor hearts, not used for previous experiments, at temperatures of 37°C, 32°C, and 27°C. Muscles were stabilized at L_opt, 1 Hz, and each of these temperatures for 15 min prior to k_tr experiments.

Statistical and data analysis.

Tension and kinetic measurements were collected and analyzed with custom-made programs in LabVIEW (National Instruments). Muscle tensions were normalized to the cross-sectional area of the muscles and expressed as millinewtons per square millimeter. The k_tr tracings were analyzed from the time of the residual tension, remaining after the slack-restretch maneuver (not different between failing and nonfailing at L_opt), to when the muscle reaches maximum tension during the overshoot. Origin 7 software (OriginLab) was used to fit k_tr tracings to the equation F = F_max[1 − e^−k_tr(t)] + F_res and calculate EC₅₀ of isoproterenol with the equation F = F_min + F_max[Fⁿ/(Fⁿ + EC₅₀)]. Statistical analysis (KaleidaGraph) was performed with unpaired t-test, two-way ANOVA, or one-way ANOVA for repeated measures followed by Bonferroni post hoc test when appropriate. Statistical significance was set at P < 0.05. All data are shown as means ± SE.

Length-tension relationship, force-frequency relationship, and β-adrenergic response in right ventricular myocardium.

In agreement with previous studies on length-dependent force development in human myocardium (27, 47), increasing muscle length stepwise from L₉₀ to L₉₅ to L_opt increased developed tensions in both nonfailing and failing right ventricular muscles to the same degree at a baseline frequency of 0.5 Hz (Fig. 2A). The FFR of nonfailing right ventricular myocardium was positive (up to 2 Hz), while failing right ventricular myocardium had a negative response (Fig. 2B). Both nonfailing and failing right ventricular myocardium responded positively to increasing isoproterenol concentrations (Fig. 2C); however, the EC₅₀ of nonfailing muscles was significantly lower than that of failing muscles, indicating β-adrenergic desensitization and/or downregulation in the latter group (Fig. 2D).

Kinetics of contraction are unaffected in right ventricle of failing myocardium and slow down with increasing muscle length.

We analyzed kinetic parameters of the same right ventricular muscles in Fig. 2A, which also showed blunted FFR and β-adrenergic desensitization. Schematic definitions of these kinetic parameters are provided in Fig. 1D. The maximal rate of tension rise (+dF/dt) and maximal rate of tension decline (−dF/dt) were normalized to the developed twitch tensions in order to obtain a pure kinetic parameter in units of second⁻¹ (28, 29) (Fig. 3). Within each group, +dF/dt/F_max accelerated as muscles were shortened from L_opt to L₉₀ (P < 0.05); however, there were no differences between nonfailing and failing myocardium at each respective muscle length at stimulation frequency of 0.5 Hz (P > 0.05). The relaxation parameter −dF/dt/F_dev, with exception of L₉₅ (P < 0.05), was also similar between the two groups. No major alterations in other kinetic parameters were observed between nonfailing and failing myocardium (P > 0.05). Additional kinetic parameters of these muscles are provided in Table 3.

k_tr measurements are temperature dependent.

We utilized k_tr experiments during K⁺ contracture to assess cross-bridge cycling kinetics. For further characterization, we first determined the relationship between temperature and this kinetic parameter. Reducing temperature from 37°C to 32°C and then further to 27°C decreased k_tr in our human nonfailing muscles from 18.1 ± 2.1 s⁻¹ to 14.1 ± 1.4 s⁻¹ and to 6.6 ± 0.9 s⁻¹, respectively (n = 4; ANOVA, P < 0.05). This reflects a Q₁₀ of ∼2.8 (Fig. 1C). This is close to the Q₁₀ values (range of ∼2–3) that have been reported previously for k_tr in intact and permeabilized rat cardiac preparations (11, 19, 37). Furthermore, a previous report indicated that the minimum frequency of dynamic stiffness has a minimum Q₁₀ of 2.7 in human intact preparations (43). The effect of temperature on our k_tr recordings in humans suggests that this parameter is measuring a cellular process, most likely cross-bridge cycling kinetics.

Twitch contractile kinetics and tensions in muscles used for k_tr experiments.

In parallel, a separate set of muscles was used for assessing the effects of muscle length on cross-bridge cycling kinetics by utilizing the k_tr experiments. FFR measurements in a subset of these muscles also showed that nonfailing right ventricular myocardium had a positive FFR (from 1 Hz to 2 Hz) while failing right ventricular muscles had a flat FFR within the same frequency range (data not shown). Contractile parameters were measured in these muscles after they were stabilized at each muscle length at 1 Hz and before K⁺ contracture and k_tr experiments. The +dF/dt/F_dev parameter was sped up as muscle lengths were decreased at a stimulation frequency of 1 Hz (Fig. 4) similar to 0.5 Hz (Fig. 3). Additionally, −dF/dt/F_dev also sped up at 1 Hz as muscle lengths were decreased. Furthermore, both +dF/dt/F_dev and −dF/dt/F_dev were similar between nonfailing and failing myocardium at each respective muscle length. Muscle length regulated the tension developed by these muscles, similar to the previous muscles shown in Fig. 2 (Table 4). Active developed (F_dev) and resting (F_rest) twitch tensions were not significantly different between nonfailing and failing hearts at each respective muscle length (P > 0.05). Within each group, F_dev and F_rest decreased as muscles were shortened from L_opt to L₉₀ (Table 4). We did observe some rundown in developed tension over the long time course of the k_tr experiments. This was quantified by repeating the experiments at L_opt for a second time (L_opt,repeat). Although this rundown seemed more pronounced and statistically significant in the nonfailing group (P < 0.05), the failing group did not reach statistical significance (P = 0.1083). This can partially be explained by our recent study (35) that shows that twitch-dependent rundown is an excitation-contraction (EC) coupling phenomenon and may likely be caused by deteriorating sarcoplasmic reticulum (SR) function. The nonfailing myocardium has a potential larger possibility for a loss of SR function, as this is already reduced in failing myocardium, as reflected by the negative FFR (24, 44).

k_tr is unchanged in failing myocardium; muscle length modulates k_tr.

After stabilization and measurement of twitch parameters at a particular muscle length (L_opt, L₉₅, L_90, L_opt,repeat), k_tr experiments were performed in order to assess cross-bridge cycling kinetics while the trabeculae were under a brief K⁺ contracture. These measurements were made after the muscles were stabilized at stimulation frequency of 1 Hz (equivalent to 60 beats/min). The maximal developed tension during K⁺ contracture decreased as muscles were shortened from L_opt to L₉₅ to L₉₀ (Fig. 5A). This tension was increased when muscles were restretched to L_opt,repeat. However, the tension of L_opt vs. L_opt,repeat was reduced in nonfailing samples (P < 0.05) but unchanged in failing samples (P = 1). There was no statistical difference (P > 0.05) in the K⁺ contracture tension between nonfailing and failing groups at each muscle length.

In all of the muscles used for length studies, k_tr was measured twice at L_opt during two separate K⁺ contractures, one toward the beginning and the other at the end of the experimental series. The k_tr values of the first L_opt were not statistically different compared with L_opt,repeat in both the nonfailing and failing groups (P = 1 for both groups) (Fig. 5C). In both nonfailing and failing samples, k_tr (during maximal K⁺ tension) was accelerated as muscles were shortened from L_opt to L₉₀ (P < 0.05) and decreased as muscles were restretched back to L_opt,repeat (P < 0.05) (Fig. 5C). There were no significant differences (P > 0.05) in k_tr values between nonfailing and failing groups at each muscle length (Fig. 3, B and C). The ratios of F_res (residual force following k_tr) to F_K⁺ (K⁺ contracture force before k_tr) for the nonfailing group were 0.32 ± 0.03 (L_opt), 0.27 ± 0.06 (L₉₅), 0.23 ± 0.07 (L₉₀), and 0.30 ± 0.06 (L_opt,repeat) (P = 0.43). The ratios for the failing group were 0.18 ± 0.04 (L_opt), 0.2 ± 0.04 (L₉₅), 0.2 ± 0.05 (L₉₀), and 0.18 ± 0.06 (L_opt,repeat) (P = 0.97). We also measured k_tr during submaximal K⁺ contracture tension levels in a subset of the muscles. During submaximal activation levels, k_tr was not different between nonfailing and failing samples (data not shown).