Acute Biomechanical Effects of Cardiac Contractility Modulation in Living Myocardial Slices From End- Stage Heart Failure Patients

1. Introduction

Cardiac contractility modulation (CCM) is a novel device-based therapy for patients with refractory symptomatic heart failure (HF)[9]. CCM stimulation enhances myocardial contractility by delivering non-excitatory electrical pulses of high voltage and long duration to the failing myocardium during the absolute refractory period[1,25]. As such, CCM does not elicit a new contraction, but provides inotropic enhancement of contractility by increasing systolic levels of intracellular calcium[5,20]. CCM stimulation of the right ventricular septum improved global contractility, ejection fraction, functional outcomes and quality of life, and elicited reverse remodeling similar to cardiac resynchronization therapy in HF patients[9,26]. However, more studies are necessary to determine whether CCM improves long-term outcomes, such as survival and HF hospitalizations[6,13]. Continuous efforts are needed to identify true responders to CCM therapy, understand the exact mechanisms and to optimize the contractile response to CCM stimulation[19]. Recent studies in 3D human engineered cardiomyocytes and neuro-cardiac coculture indicated that CCM stimulation settings may impact cardiomyocyte contractile improvement, but these models still lack mechanical loading and physiological resemblance[7,15]. In the current study, living myocardial slices (LMS) from end-stage HF patients were used to evaluate the acute impact of CCM stimulation parameters on the biomechanical properties of the human failing myocardium.

2. Methods

2.1. Slice Preparation

Left ventricular tissue was obtained from patients with end-stage HF undergoing assist device implantation or cardiac transplantation. All patients consented to the use of their surgical residual material for scientific research (opt-out method of consent), as approved by the Medical Ethical Committee of the Erasmus Medical Center (MEC 2020-0988) and in accordance with local regulations and with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments and guidelines. LMS production was described previously in detail[3,8]. In short, left ventricular biopsies originating from the apex and left ventricular free wall were immediately submerged in 4°C Tyrode slicing buffer (NaCl 136 mM, KCl 5.4 mM, MgCl₂·6H₂O 1 mM, NaH₂PO₄·H₂O 0.33 mM, Glucose 10 mM, CaCl₂·2H₂O 0.9 mM, 2,3-butanedione monoxime 30 mM, HEPES 5 mM, pH 7.4) and trimmed to remove epicardial fat and excessive endocardial trabeculae. The biopsies were then immersed in 37°C 4% low-melting agarose (Agarose II, VWR Chemicals LLC, Solon, OH, USA) and cooled until the gel solidified. Submerged in a 4 °C Tyrode buffer-filled bath, the embedded tissue was sliced by a high-precision cutting vibratome (VT1200S, Leica BioSystems, Nussloch, Germany) with a thickness of 300 µm (vibration amplitude 1.3 mm, blade advance speed 0.07 mm/s). After removal of the agarose, plastic triangles were glued to both ends of the LMS, with longitudinal myocardial fibre orientation in between the triangles.

2.2. Slice Cultivation

The cultivation chambers (InVitroSys GmbH, Munich, Germany) were filled with 2.4 mL of 37 °C culture medium (Gibco Medium-199 (Grand Island, NY, USA) supplemented with 5% penicillin-streptomycin, 5% Insulin-Transferrin-Selenium-X and 50 µM 2-Mercaptoethanol). Culture medium calcium concentration was 1.80 mM. From onset of cultivation, 1.6 mL medium was refreshed after 1 hour and hereafter every 24 hours. The slices were mechanically loaded with a diastolic preload of approximately 1 mN using a custom-made stretcher, of which one end was able to move during contraction, resulting in auxotonic contractions[8]. The preload was readjusted with every medium change. The cultivation chambers were placed in a 37 °C 5% CO₂ incubator and placed on a rocking plate (30 rpm) for continuous oxygen and nutrient supply to the LMS. LMS were continuously cultured with electrical field stimulation using a biphasic square wave main pulse (50 mA), consisting of 3 ms charging and discharging pulses separated by a 1 ms pause interval (total pulse duration: 7 ms). LMS were continuously stimulated at 0.5 Hz during the first hour to optimize oxygen and nutrient supply to the tissue and with 1.0 Hz hereafter to simulate near-physiological conditions.

2.3. CCM Stimulation

On day one of LMS culture, CCM field stimulation was administered as a second biphasic pulse (Figure 1). CCM stimulation was performed with 40 ms delay after the excitatory stimulus, 80 mA amplitude and total pulse duration 20 ms to simulate clinical settings. Next, different CCM stimulation parameters (delay, amplitude and duration) were tested, as specified in Figure 1. The stimulation threshold was determined by increasing baseline pulse amplitude in steps of 5 mA until contractile capture. The effect of CCM pulse amplitude (1 to 4 times the individual LMS stimulation threshold), pulse duration (11, 21, 31 and 41 ms) and pulse delay (10, 20, 30, 40, 50, 70, 100, 150 and 200 ms) were evaluated. Each experiment consisted of 5 minutes continuous CCM stimulation, followed by a 15 minutes recovery period before starting the next experiment.

2.4. Functional Refractory Period

For each LMS, the functional refractory period (FRP) was determined using decremental pacing with a fixed S1-rate (1000ms) and decreasing the programmed extra S2-stimuli (950 to 100 ms). The first S1-S2 interval that did not show contractile capture on the S2-stimulus was defined as the FRP.

2.5. Force-Frequency Relationship

The force-frequency relationship (FFR) was determined by increasing stimulation frequencies from 30 to 180 bpm with increments of 30 bpm and 1 minute in between intervals. A comparison was made of the FFR between baseline and during CCM stimulation to assess the effect of CCM at higher stimulation frequencies.

2.6. Contractile Measurements

Contractile measurements were extracted using the peak analysis module of LabChart 8 software (ADInstruments). Start and end of the peak were chosen at 10% away from the baseline to compensate for baseline noise [2]. Contractile force was continuously measured by a magnetic force transducer in the cultivation chamber. Contractility was assessed for 60 seconds directly before each protocol and during the fifth minute of CCM stimulation. For each contraction, force amplitude (F_max), peak area (AUC), contraction duration (CD), contraction duration at 50% of the maximum amplitude (CD₅₀), time to peak (TTP), time to relaxation (TTR), steepest positive slope (+dF/dt), and steepest negative slope (−dF/dt) were evaluated (Figure 1). A positive inotropic response was defined as a ≥ 5% increase in F_max compared to baseline, and a negative response as ≥ 5% decrease. A neutral response was defined as a response in Fmax between -5% and 5%. LMS with a positive inotropic response were selected for biomechanical effect analyses. A separate biomechanical analysis was performed in LMS with a positive inotropic response. LMS were excluded from analysis for the respective protocols if baseline noise or sensor errors interfered with the contractile signal.

2.7. Statistical Analysis

Statistical analyses were performed using R software (version 4.4.0; R foundation for statistical Computing, Vienna, Austria) with the “lmerTest” statistical package. Continuous baseline data were presented as mean ± standard deviation (SD) if normally distributed and as median and interquartile range (IQR) otherwise. Categorical data were presented as number (%). All contractile parameters were presented as median (IQR), as the majority of this data has a non-normal distribution. Differences in contractile parameters were assessed using a clustered Wilcoxon Signed Rank test. Relative values of biomechanical parameters were calculated as ${\displaystyle \frac{(\mathrm{C}\mathrm{C}\mathrm{M}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}–\mathrm{b}\mathrm{a}\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e})}{\mathrm{b}\mathrm{a}\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}}}·100\mathrm{\%}$. A linear mixed-effect model was developed with random slopes for stimulation frequency and nested random intercepts for LMS in individual patients to capture higher correlations within patients and the repeated measures within LMS. Fixed effects used were the application of CCM, stimulation frequency and their interaction term. The model was fit using under restricted maximum likelihood with the Nelder-Mead optimizer[4]. T-tests using Satterthwaite’s method were employed to obtain p-values of the fixed effects[12]. P-values were considered statistically significant if P ≤ 0.05.

3. Results

3.1. LMS Characteristics

Baseline characteristics of the study population are summarized in Table 1. No patients were previously treated with CCM therapy. LMS were produced from left ventricular biopsies of 7 patients with end-stage HF (age: 39 ± 18.5 years, 3 male), resulting in 64 LMS. Median FRP of HF-LMS was 275 (260 – 330) ms. No HF-LMS had an FRP below 200 ms and no extra induced contractions were observed during CCM stimulation.

3.2. Biomechanical Effects of CCM on Contractility

Biomechanical effects of CCM are shown in Table 2. CCM stimulation (80 mA, 21 ms duration, 40 ms delay) significantly enhanced F_max in HF-LMS (n = 60) compared to baseline (CCM: 1229 (587 – 2658) vs. baseline: 1066 (529 – 2128) µN, p = 0.05). Maximum +dF/dt increased significantly with CCM stimulation (CCM: 10145 (5086 – 22260) vs. baseline: 8148 (4109 – 16488) µN/s, p = 0.05) and –dF/dt (CCM: -6968 (-16660 - -3692) vs. baseline: 6461 (-13355 - -3020) µN/s, p = 0.04). In addition, an increased AUC was observed (CCM: 297 (151 – 562) vs. baseline: 243 (129 – 464) µN.s, p = 0.05). CD, TTP and TTR were not affected by CCM.

Variation in F_max was observed between LMS in response to CCM, even within different LMS produced from the same patient. 42 LMS (70.0%) had a positive inotropic response to CCM, 11 LMS (18.3%) showed a negative inotropic response to CCM and 7 LMS (11.7%) had a neutral response. In LMS with a positive inotropic response (n = 42), CCM increased F_max by 19.5% (CCM: 1324 (801 – 2738) vs. baseline: 1066 (626 – 2113) µN, p = 0.03), while enhancing AUC (CCM: 306 (192 – 581) vs. baseline: 243 (156 – 461) µN.s, p = 0.04) and reducing CD₅₀ (CCM: 217 (202 – 242) vs. baseline: 224 (216 – 252) ms, p = 0.04). These results are presented in Table 3. In LMS with a positive inotropic response, F_max increased gradually for up to 3 minutes and then stabilized (Figure 2) which disappeared after cessation of CCM stimulation.

Variation in F_max was also observed between patients in response to CCM stimulation (Table 4). Patients with ischemic cardiomyopathy (n = 3) demonstrated an increase of F_max (68.4, 35.2 and 5.5%), while both patients with chemo-induced dilated cardiomyopathy responded with a decrease (-0.3 and -4.2%). In both patients, greater CCM delay, duration and to lesser extent amplitude resulted in maximal increase of F_max to 34.5 and 14.5%, respectively.

3.3. Effect of CCM Stimulation Delay

LMS displayed enhanced F_max in a delay dependent manner, as illustrated in Figure 3 and Suppl. Table A. The AUC, dF/dt and -dF/dt enhanced with increasing delay settings (Figure 3). A reduction of TTP, was observed at delay ≤ 30 ms (delay 30ms; TTP; CCM: 158 (138 - 176) vs. baseline: 170 (159 - 185), p = 0.05).

The fraction of LMS exerting a positive inotropic response to CCM increased with increasing delay settings (Figure 4). At the shortest delay of 10 ms, 34 LMS (57 %) demonstrated a positive inotropic response, which increased to ≥ 70 % of LMS with delays ≥ 40 ms. Of note, 33% of LMS demonstrated a negative inotropic response at delay of 10 ms. This fraction subsided with increasing delay settings (Figure 4).

3.4. Effect of CCM Stimulation Duration

Longer stimulation duration showed increased F_max (Suppl. table B). Stimulation duration of 11 ms did not significantly increase F_max, and only 20 LMS (41%) demonstrated a positive inotropic response at this duration. The fraction of LMS with a positive inotropic response increased to 78% with 41 ms stimulation duration (Figure 4).

3.5. Effect of CCM Stimulation Amplitude

The fraction of LMS with a positive inotropic response increased from 4 LMS (7%) at 1 times threshold to 40 LMS (71%) at 4 times threshold (Figure 4). The effect of CCM stimulation amplitude on median F_max is demonstrated in Suppl. Table C.

3.6. Force-Frequency Relationship

HF-LMS demonstrated predominantly a neutral or negative FFR before CCM stimulation, as demonstrated by a decrease of median F_max with increasing stimulation frequencies (Figure 5). Positive, neutral and negative FFR were all observed when applying simultaneous CCM stimulation (Figure 5). A decrease in sample size was observed at higher stimulation frequencies, due to intolerance of HF-LMS to high frequency stimulation. Panel B of Figure 5 illustrates a similar FFR in 24 LMS with capture at all stimulus frequencies. The results of the linear mixed-effect model are presented in supplementary figure D. F_max decrease with increasing stimulation frequencies was significantly less steep in LMS with CCM stimulation compared to without CCM stimulation (β CCM: -0.04 vs. β no CCM: -0.12, p-interaction < 0.001). Residuals of the linear mixed-effect model were normally distributed based on visual inspection.

4. Discussion

4.1. Key Findings

This is a proof-of-concept study using CCM stimulation in LMS with heart failure phenotype, showing an acute increase of maximum contractile force in the majority of LMS. The strength of response was dependent on stimulation parameters and the improved biomechanical profile seemed to be determined for a specific subset of patients. Increasing CCM pulse delay, duration and amplitude resulted in a higher fraction of LMS with a positive inotropic response. Significantly higher force generation was observed with increasing delay and duration, but not amplitude. Furthermore, CCM stimulation attenuated the negative force-frequency relationship in the failing myocardium. All biomechanical effects were abolished upon cessation of CCM therapy.

4.2. Effect of CCM on the Biomechanical Profile of HF-LMS

In our study, CCM stimulation with near-clinical settings (delay 40 ms, amplitude 80 mA, duration 21 ms) enhanced maximum and total contractile force, and contraction and relaxation slopes in LMS obtained from patients with HF, while CD, TTP and TTR were not affected. In other words, CCM stimulation increases cardiac contractility without altering systolic and diastolic intervals. Similar increased contractile force and slopes were observed in pre-clinical studies[5,7,14,15], and their clinical counterparts demonstrated no significant alterations of echocardiographic ejection times or diastolic filling times[16,25]. In contrast, pre-clinical studies found a small increase in TTP in 3D cardiomyocytes and endocardial trabeculae obtained from explanted failing hearts[5,7]. Model specificity and loading conditions could explain these differences. In our study, biomechanical properties of HF-LMS were assessed using auxotonic loading, displaying greater clinical relevance in comparison to unloaded cardiomyocytes or isometric loaded trabeculae from failing hearts[8].

4.3. Individual Response of LMS to CCM Stimulation

A larger fraction of LMS showed a positive inotropic response when CCM pulse amplitude, duration and delay increased. These variations were not reported in other pre-clinical studies, and might be unique for our model, as different layers of the human myocardium were used. Intrinsic transmural heterogeneity has been demonstrated previously in LMS by Pitoulis et al.[18]. It could also be specific to the failing myocardium. Other pre-clinical studies generally use cardiomyocytes with little cell-to-cell coupling and lacking micro-architectural structures as present in the patient[7,15]. However, Burkhoff et al. demonstrated no individual differences when CCM was applied to intact human trabeculae of patients with end-stage HF (n = 6)[5]. Nevertheless, a subset of positive responders was seen in our study and raises the question whether this effect could be attributed to the underlying disease. A correlation and stratification according to aetiology could not be performed due to the patient sample size and variance in the number of LMS per patient (2 – 20 LMS), however, patients with ischemic cardiomyopathy were amongst the strongest inotropic responders independent of CCM settings contrary to the two patients with chemo-induced dilated cardiomyopathy which seemed to benefit more from a greater delay, amplitude and duration. Further studies should investigate this potential effect of underlying aetiology on CCM response. All things considered, our observations suggest that CCM stimulation should optimally strive to be longer in pulse duration and amplitude in order to target a larger fraction of the myocardium.

Although clinical CCM stimulation has not been associated with ventricular arrhythmias, delay settings must be chosen with care, as high-voltage signals have the potential to induce ventricular depolarization or even arrhythmias when administered outside the absolute refractory period. Clinically, CCM is applied 20 – 45 ms after local sensing. Our results demonstrate that increasing delay past 45 ms is beneficial to enhance F_max and AUC in LMS, corroborating previous reports using cardiomyocytes[7,23]. As such, using the maximal delay settings within current safety regulations may already contribute to reaching a larger proportion of tissue and better contractile performance. Duration of the absolute refractory period of ventricular cardiomyocytes is approximately 200 – 300 ms[10]. In our study, all HF-LMS had an FRP above 200 ms and no extra induced contractions were observed during CCM stimulation, suggesting that delay settings could potentially be extended safely beyond 45 ms. Inherently, implications for safety should be explored further before implementation and should be correlated with the individual FRP.

4.4. CCM and the Force Frequency Relationship in Failing Myocardium

Increasing stimulation frequencies resulted in a reduction of F_max in LMS from HF patients. A flat or negative FFR is a characteristic feature of the failing myocardium and is often associated with deterioration of intracellular calcium handling[17]. In the failing myocardium, downregulated protein expression of the sarcoplasmic reticulum (SR) Ca²⁺ ATPase 2a (SERCA2a) and increased calcium extrusion by the sodium-calcium exchanger results in lower SR storage capacity[17]. In HF-LMS, a neutral or positive FFR was observed more often when CCM stimulation was applied simultaneously. Although there was no significant difference in FFR with and without CCM stimulation, visualization of the FFR curve is suggestive of preservation of contractility at higher stimulation frequencies. CCM is able to improve SERCA2a calcium affinity by rapid phosphorylation of phospholamban, thus increasing calcium sequestration in the SR[20]. Potentially, improved calcium handling by CCM contributes to a partial reversal of the negative FFR in the failing myocardium. Hashimoto et al. demonstrated that SERCA2a gene transfer in transgenic mice improved calcium handling and twitch force especially at higher stimulation frequencies, which increase the intracellular calcium load[11]. Detailed calcium imaging is required to elaborate these findings. Potential improvement of the FFR by CCM might partly explain the improvements in functional parameters, such as peak VO2, 6-minute walking distance and quality of life, that were demonstrated in various clinical trials[9].

4.5. Clinical Perspective

CCM stimulation is a promising therapeutic strategy that has been applied in hundreds of patients clinically, yet despite its potential benefits, remains to be universally established in heart failure management guidelines.[13,24] Heart failure is a heterogeneous condition and response to CCM therapy can vary greatly depending on factors such as stimulation settings, heart failure etiology and degree of cardiac dysfunction. CCM aims to enhance myocardial contractility by providing electrical stimulation during the refractory period, which may be more beneficical for certain subgroups of patients. Therefore, studies are warranted to identify which patients would benefit most from CCM. In this context, HF-LMS offer valuable insights as they contain disease phenotype where the tissue is mechanically loaded, electrically stimulated whilst maintaining (micro)architectural form and function..[21,22] Compared to other experimental models, HF-LMS better resemble heart failure etiology, structure and physiology. Moreover, LMS can be cultured for weeks, or even months, offering a sustained model for long-term studies.

In our study, we demonstrated an individualized response of HF-LMS to different stimulation settings, an effect that has not been demonstrated in other models. This response variation may be attributed to differences in underlying etiology of heart failure. Based on these observations, we believe that future research using HF-LMS should focus on: 1) the response to therapy in patients with various underlying etiologies, 2) the determintation of optimal stimulation settings, and 3) the effects of chronic stimulation over weeks and months on contractility, therapeutic response and the force-frequency relationship.

5. Limitations

This is a proof-of-concept study focussing on the direct effects of CCM in a novel biomimetic model using human LMS derived from patients undergoing cardiac surgery. The use of human tissue is scarce, resulting in a small sample size which should be increased in future studies to allow for examination of the effect of the underlying disease. Furthermore, a large variability in LMS biomechanical profiling is present, which is inherent to the use of a dynamic system. Nevertheless, our study clearly shows specific trends when using CCM. Healthy human cardiac tissue is even more scarce and in our expertise has a different biomechanical profile compared to HF-LMS. Therefore, each LMS was used as their own control. Next, this study investigated the acute effects of CCM stimulation parameters on contractile performance of HF-LMS, but did not investigate optimal combined settings for maximal contractile improvement nor did we extend to chronic CCM stimulation.

6. Conclusion

We present evidence that CCM inotropic response is stimulation parameter dependent (delay, duration, amplitude) using a novel biomimetic setup with human HF-LMS. Furthermore, CCM can partially mitigate the negative FFR of HF-LMS.