Nonclinical evaluation of chronic cardiac contractility modulation on 3D human engineered cardiac tissues

Cardiac contractility modulation (CCM) is a medical device‐based therapy delivering non‐excitatory electrical stimulations to the heart to enhance cardiac function in heart failure (HF) patients. The lack of human in vitro tools to assess CCM hinders our understanding of CCM mechanisms of action. Here, we introduce a novel chronic (i.e., 2‐day) in vitro CCM assay to evaluate the effects of CCM in a human 3D microphysiological system consisting of engineered cardiac tissues (ECTs).


| INTRODUCTION
Cardiac electrophysiology medical devices, including cardiac resynchronization therapy (CRT) and cardiac contractility modulation (CCM), have been developed to treat patients with drug-resistant heart failure (HF).While CRT is the first-line treatment for HF patients displaying low ejection fraction (<35%) and prolonged QRS duration, 1 there remains a significant population of HF patients (e.g., 60-70%) with normal QRS duration who may not be eligible for CRT.
Consequently, there is a significant gap in viable treatment options for these HF populations and CCM is heralded as a potential solution.
The CCM device is implanted in the pectoral region with contact leads placed in the myocardium.3][4] Future devices are expected to be developed to address additional patient populations and device functionalities.
There is a lack of human in vitro tools to evaluate mechanisms of action of novel cardiac electrophysiology medical devices.
Hence, detailed mechanistic elucidation may be useful to inform or make changes to device settings (e.g., treatment schedule, pulse parameters).Likewise, the direct effects of chronic (e.g., continuous and discontinuous) CCM on human cardiomyocyte physiology remain poorly understood.Previous studies have demonstrated a CCM-induced enhancement of myocardial gene expression profile in patients and canine models with HF following chronic discontinuous CCM (e.g., 3-months). 5,6These studies have provided important mechanistic insight into our understanding of chronic CCM.To date, most CCM studies rely heavily on human subjects and costly animal models.In principle, human induced pluripotent stem cell-derived cardiomyocyte (hiPSC-CM) models may be a useful tool to assess the molecular and functional effects of CCM in human cardiac tissue in vitro, 7 complementing in vivo or clinical studies.HiPSC-CMs have been proposed as an in vitro tool to aid the therapeutic development process.Likewise, our previous hiPSC-CMs studies have been included in the US FDA catalog of regulatory science tools to help assess new medical devices. 8ditionally, hiPSC-CMs have been utilized in regulatory submissions. 91][12][13][14] Human 3D cardiac microphysiological systems, including hiPSC-CM engineered cardiac tissues (ECTs), cardiac microtissues or cardiac microbundles, are established models for the acute evaluation of cardiac electrophysiology medical device signals and cardiac drugs. 12,15However, such models may not be appropriate for long-term chronic in vitro studies.In this work, we develop a chronic (i.e., 2-day) in vitro method to stimulate (i.e., discontinuously) hiPSC-CM based ECTs to elucidate the consequences of CCM in human cardiac tissues. 16 demonstrate that ECTs respond to chronic electrical stimulation, mimicking the clinical CCM signal, by increased contraction and intracellular calcium transient amplitude, prolonged contraction duration, and altered gene program relative to paced and time-matched controls.To the best of our knowledge, this is the first in vitro chronic ECT medical device study, elucidating the functional effects of CCM on human cardiac biology.Here, we establish a standardized 3D ECT-based in vitro testbed to quantify the physiological and molecular effects of chronic CCM that can be used to test safety and performance secondary to electrical stimulation.

| Electrical field stimulation
3D ECTs were stimulated with two separate commercial pulse generators, A-M Systems (Model 4100, A-M Systems) and Grass stimulator S88, for the CCM group and control group, respectively.4][25] In this configuration, the tissues are oriented perpendicular between the parallel electrodes.12]26 ECTs were field stimulated at ∼3 times the capture threshold using monophasic square wave pulses at 1 Hz, 5 ms pulse duration, and ∼20 V/cm amplitude (Figure 1D).CCM was delivered at 1 Hz as 2 biphasic pulses, 5.14 ms phase duration (20.56 ms total duration), 28 V/cm pulse amplitude (phaseamplitude), and an interphase interval of zero.The delay between the initial field stimulation pulse and CCM was 30 ms (i.e., time from the beginning of the field stimulation pulse and the beginning of the CCM pulse).CCM pulse parameters were comparable to the standard setting typically used (Figure 1D). 27,283D ECTs were chronically stimulated for 2-days, from Day 6 to 8, in 5 1-h increments per each 24 h.
Following stimulation ECTs were allowed to rest and recover for ∼14-h (i.e., spontaneous beating/unstimulated) before contractility, calcium, and gene expression measurements.

| Measurement of contractile properties
A contractility platform and software (CellOPTIQ and Clyde Biosciences), based on video (i.e., pixel) displacement, was used to measure 3D ECT contractility. 12,29,30  script from PDMS micropillar displacement based on the known micropillar spring constant of 2.68 µN/µm.ECT cross-sectional area was calculated assuming an ellipse shape based on the average width (Supporting Information: Figure 1), as previously described. 21Twitch power was calculated by the product of the systolic force and the twitch velocity as previously described. 31For chronic CCM contraction and force measurements (i.e., Day 8) ECTs were used from three independent experiments (i.e., separate seedings, generation, and electrical stimulation).]32 ECTs that were detached from the micropillars, quiescent, or did not achieve capture upon field stimulation were not used for contractile experiments.
All primers were generated and are listed in (Supporting Information: Table S1).Transcripts were normalized to housekeeping gene GAPDH and differential fold changes were calculated as 2 C −ΔΔ T values.

| Mitochondrial activity
For mitochondrial functional assay, ECTs were stained with Mito-

| Chronic CCM increases contractility and contraction duration in ECTs
To assess the chronic effects of CCM on human cardiomyocyte contractility, we evaluated contractile properties of ECTs following stimulation with the standard CCM signal (i.e., two biphasic pulses, 5.14 ms duration, and 30 ms delay) and the currently approved clinical schedule (i.e., 5, 1-h increments in 24 h, on/off) (Figure 1D).
Stimulation was applied for 2-days, and measurements were obtained We observed a mean contraction slope of 558.9 ± 165.6 a.u./s in the control group and 1092 ± 271.1 (p = .0588)in the CCM group.
Similarly, we found a mean relaxation slope of −294.4 ± 82.8 a.u./s in the control group and −534.7 ± 123.9 (p = .0588)in the CCM group, although there were no significant differences in these values (Supporting Information: Table S2).Additionally, chronic CCM resulted in the widening of the contraction trace (i.e., contraction duration 10%−90%) for the entire contraction cycle (Figure 2A−D).
Likewise, chronic CCM resulted in prolonged contraction time in early systole (i.e., time to peak 10%) and prolonged relaxation time for the entire diastolic phase (i.e., time to baseline 10%−90%) relative to that of the control group (Figure 2E).Taken together; these results suggest that chronic CCM increases peak contraction amplitude, widens contraction duration, and prolongs contraction and relaxation time in ECTs relative to paced and time-matched controls 14-h after stimulation was ceased.| 899

| Effect of chronic CCM on ECT mechanical output and energetics
Given that culturing ECTs on PDMS micropillars with a known spring constant permits calculation of peak systolic force, we next quantified ECT mechanical output (i.e., force and power), which are key metrics of cardiac physiology.Field-stimulated ECTs displayed a mean peak systolic force of 10.5 ± 1.6 µN in the control group and 12.8 ± 2.5 µN (p = .4139)in the CCM group.(Figure 3A).Using the peak systolic force and maximum twitch velocity, we calculated a mean maximum contraction twitch power of 823.0 ± 221.0 pW in the control group and 1060 ± 273.0 pW (p = .5030)in the CCM group.In addition, we observed a mean maximum relaxation twitch power of 713.4 ± 197.0 pW in the control group and 754.7 ± 184.0 pW (p = .8818)in the CCM group.We next evaluated how well the video-based pixel-displacement compared with the calculated micropillar-deflection peak systolic force (Figure 3C).The videobased pixel-displacement correlated well with micropillar-deflection, exhibiting a significant positive correlation with values independently calculated from the same ECTs (Figure 3C).To evaluate mitochondrial activity, we quantified mitochondrial membrane potential following chronic CCM.We observed a normalized mean MitoTracker intensity of 1.0 ± 0.08 a.u. in the control group and 1.3 ± 0.2 (p = .0871)in the CCM group (Figure 3D).These results suggest that 2-days of chronic CCM may not significantly alter ECT mechanical output and energetics 14-h after stimulation was ceased.

| Chronic CCM enhances intracellular calcium amplitude in ECTs
Next, we investigated the effects of chronic CCM on human cardiomyocyte intracellular calcium handling properties.We found following chronic CCM, spontaneously beating ECTs from the CCM group displayed comparable diastolic intracellular calcium levels as the control group (Figure 4A,B).As expected, based on the increased contraction amplitude, the CCM group displayed increased calcium transient amplitude relative to the control group whereas all other calcium handling parameters investigated were not significantly different (Figure 4A,B) (Supporting Information: Table S3).While the spontaneous beat rate of both groups was reduced relative to that of baseline ECTs (i.e., Day 5), both the control group and the CCM group displayed comparable spontaneous beating activity of 39 ± 2.1 BPM and 37 ± 4.5 BPM (p = .3),respectively.These results suggest that chronic CCM increases ECT intracellular calcium amplitude relative to the control group without impacting spontaneous beat rate.

| Chronic CCM modifies ECT gene profile
To determine the effects of chronic CCM on human cardiac tissue gene expression, we performed transcriptional profiling of a subset of cardiac genes involved in HF (Figure 5).The effects of chronic CCM on myocardial gene expression profile has been studied extensively in patients and canine models but not yet in ECTs.ECT gene expression profile showed differential expression following chronic CCM.Of note, expression of NPPB encoding B-type natriuretic peptide (BNP), a frequently used biomarker of HFe (i.e., upregulated), was significantly reduced in the CCM group (Figure 5).Enzyme-linked immunosorbent assay (ELISA) revealed a mean NTproBNP expression of 466.6 ± 436.8 pg/mL in the control group and 80.24 ± 63.04 pg/mL (p = .6667)in the CCM group (Supporting Information: Figure 2).
Additionally, the gene encoding the sodium-calcium exchanger (SLC8), a component of intracellular calcium handling, which has been reported to be differentially expressed in HF, 35 was significantly increased in the CCM group (Figure 5).These results illustrate the capability of ECTs to evaluate differential gene expression as early as 2-days following chronic CCM in vitro.

| Chronic 3D ECT model
In this study, we establish a robust in vitro tool to quantify the effects of chronic (i.e., 2-days), discontinuous CCM on 3D human ECTs to assist mechanistic elucidation, device development, and regulatory decision-making capabilities.CCM is an intracardiac therapy approved for the treatment of HF with reduced ejection fraction.
However, the chronic effects of the standard clinical CCM pulse parameters on human cardiac tissues have not been completely parameter-dependent manner. 12Likewise, we have demonstrated that a 2D hiPSC-CM model stimulated acutely with CCM displays increased contractility and calcium; however, this system required submaximal extracellular calcium concentration (i.e., [0.5 mM]) and a flexible substrate to elucidate the CCM response. 8,10,11Interestingly, here using 3D ECTs that have not undergone 7-weeks of electrical conditioning we elucidate the chronic CCM effects, with key physiological and molecular features altered in a similar fashion as patients and large animal models.

| Implication for future studies
β-adrenergic signaling has been implicated as a potential mechanism for the enhancement of cardiac function during CCM. 2,26,36,37We have previously shown in 2D hiPSC-CM models that acute CCM effects are likely induced by a mixed mechanism, including sympathetic stimulation through neuronal ganglion and β-adrenergic signaling pathways. 7,10,11armacological inotropes that act via β-adrenergic receptors, such as isoproterenol, exert their effects by increasing the amount of intracellular calcium and cAMP-dependent pathways resulting in positive inotropy, lusitropy, and chronotropy.Following chronic CCM, ECTs displayed positive inotropy including a twofold (p = .0151)increase in the contraction-time interval (AUC) (Supporting Information: Figure 3).Likewise, the CCM group displayed greater than fivefold increase in contraction-calcium loop area, suggesting higher contractile function (Supporting Information: Figure 4).Moreover, while we found chronic CCM ECTs display prolonged contraction duration, we did not investigate if this translates to changes in action potential duration.An important feature of hiPSC-CMs is their utility as an in vitro proarrhythmia model, to detect drug-induced proarrhythmic effects. 19nsequently, following chronic CCM we did not observe an increase in proarrhythmic-like events (data not shown), as assessed by the frequency of ectopic beats during spontaneous beating conditions.

| Study limitations
Our study has several limitations.For example, the standard CCM parameters and schedule were selected to best mimic those of the clinical CCM device.8][39] Likewise, to enable an accelerated experimental workflow, only the effects following 2-days of chronic CCM were assessed in this study.Prolonged chronic CCM (e.g., weeks) is needed but outside the scope of this work.
The observed CCM-induced effects remained 14-h after the stimulation was stopped, suggesting long-term functional and molecular changes.However, we do not know how long following CCM these effects remain or if they were greater immediately after CCM was ceased.
We did not investigate shorter or longer time points following CCM to determine if the response will be augmented, reversed, or plateaued.
Furthermore, while we evaluated the long-term effects of CCM, we did not investigate the effects while the CCM signal was actively turned on. 40ther the effects were only evaluated during spontaneous beating or standard 1 Hz field stimulation.This is of critical importance because in patients, CCM is turned on and off (i.e., discontinuous) thus there may be a modified response while the signal is actively being delivered.
The hiPSC-CMs used in this study represent an apparently "healthy" background, whereas CCM is indicated for HF patients.Utilizing a 2D hiPSC-CM LMNA-related dilated cardiomyopathy (DCM) disease model, we have previously demonstrated an acute CCM-induced increase of contraction amplitude. 11As such, the 3D ECT model will be extended to various diseased backgrounds including HF, DCM and HCM.Depending on the specific disease phenotype, personalized diseased hiPSC-CM and hiPSC-cardiac fibroblast ECTs may display enhanced CCM response relative to 'healthy' models. 41Ts used in this study did not undergo a 7-week electrical conditioning protocol before experiments to enhance functionality as before. 12More functionally enhanced ECTs are expected to have an augmented functional response and gene profile.ECTs that have not undergone 7-weeks electrical conditioning display several features of immature cardiac tissue including spontaneous beating, negative forcefrequency, and lack post-rest potentiation. 42Of note, while we observed altered gene expression profile, it was not identical to the previous reports in canines and patients.This is likely a result of using more complex in vivo models and an experimental duration of 3-months compared to 2-days investigated here.
Utilization of complex culture conditions inherently reduces experimental throughput.As such, we were not able to produce a truly high-throughput assay.The nature of the ECT model precluded sufficient biological material for detailed protein and phosphorylation analysis (e.g., Western blot).Hence, we leveraged ECT conditioned media to assess NTproBNP secretion to support gene expression findings, thus multiplexing the assay.Furthermore, the limited number of tissues may have resulted in certain endpoint assays being underpowered to reach statistical non-significance.
We recognize there are inherent differences in cardiac contractility and force outputs.While both measurements typically have a synergistic relationship, direct force measurements (i.e., force transducer) are a preferred method but are limited by throughput and technical expertize. 12Here, we used standard video-based pixel displacement and a calculated indirect force measurement based on micropillar-deflection to enable usability of the tool.

| CONCLUSION
This work provides a chronic (i.e. calculated as the maximum and minimum time derivative of the contractility amplitude, respectively.To quantify peak systolic force, blinded contraction videos were analyzed by a custom MATLAB F I G U R E 1 Chronic 3D ECT CCM model.(A) Schematic of ECT generation depicting bright field images of preplated hiPSC-CMs (20X, scale bar 0.1 mm) and vCFs (X10, scale bar 1.0 mm).Dissociated cells were seeded in PDMS molds for 5 days to form ECT (4X, scale bar 0.5 mm).(B) ECT compaction and electrical stimulation schedule.(C) Image of commercial pulse generator (left) and custom 10 cm stimulation dish (right).(D) Electrical stimulation waveforms for control group (i.e., Paced only) (Top) and CCM group (Bottom) depicting standard biphasic CCM waveform in red (i.e., two biphasic pulses, 5.14 ms duration, 30 ms delay).3D, three-dimensional; CCM, cardiac contractility modulation; ECT, engineered cardiac tissue; hiPSC-CMs, human induced pluripotent stem cell-derived cardiomyocytes; PDMS, polydimethylsiloxane; vCFs, ventricular cardiac fibroblasts.
ECTs were washed twice for 15 min with FluoroBrite DMEM.A minimum of 30 min were allowed for de-esterification before imaging ECTs.PDMS molds with ECTs attached were inverted in 48-well glass bottom plates (MatTek) containing 300 µL per well of conditioned iCell Cardiomyocytes Maintenance Medium (#M1003, Fujifilm Cellular Dynamic, Inc.).ECTs were imaged by an inverted fluorescence microscope (Zeiss).The camera attached to the side port of the microscope was used to position the ECT for fluorescence-based calcium transient measurement using a X40 objective.
14-h later (Figure1B).Contractile properties were compared to that of control ECTs exposed to an identical field stimulation and recovery schedule (Figure1).When quantified with video-based pixel displacement, field-stimulated ECTs from the CCM group exhibited increased inotropic response including significantly increased peak contraction amplitude relative to the control group (Figure2A−C).
Chronic CCM enhances ECT contraction amplitude.(A) Representative contraction recordings for Day 8 ECTs following 2-day stimulation treatment for the control group (black line) and CCM group (red line) under field stimulation at 1 Hz.Chronic discontinuous CCM induces increased contraction amplitude and prolonged contraction duration 14-h after stimulation was ceased.(B) Average contraction trace for control group (black line) and CCM group (red line) under field stimulation at 1 Hz.(C−E) Summary data graphs.Data are mean ± SEM. n = 10-17 per group.*p < .05(CCM group vs. control group).Contraction time in the CCM group is significantly prolonged at 10% (Time to peak 10%).Relaxation time is significantly prolonged for the entire diastolic phase (Time to baseline 10%−90%).a.u., arbitrary unit; CCM, cardiac contractility modulation; C.D., contraction duration; T.T.B.L., time to baseline; T.T.P., time to peak.

F I G U R E 3
Chronic CCM effect on ECT mechanical output and energetics.(A) Peak systolic force analysis.Effect of chronic discontinuous CCM on ECT systolic force generation 14-h after stimulation was ceased.(B) Maximal contraction power analysis.Effect of chronic discontinuous CCM on ECT contraction and relaxation power 14-h after stimulation was ceased.Data are mean ± SEM. n = 10-17 per group for Day 8 ECTs following 2-day stimulation treatment and 14-h rest period for the control group (black) and CCM group (red) under field stimulation at 1 Hz.(C) Contraction force vs pixel displacement correlation.Correlation between peak pixel displacement (amplitude) and peak force measurements calculated from displacement of PDMS pillars (n = 27), *p = < .00001.(D) Normalized Mitochondrial activity per nucleus.Summary bar graphs data are mean ± SEM. n = 8-9 per group.*p < .05. a.u., arbitrary unit; CCM, cardiac contractility modulation; r, pearson's correlation coefficient; P, power.defined.Here, we quantified the chronic functional effects of the standard clinical CCM parameters on ECTs in vitro using video-based and fluorescence imaging.We discovered that chronic discontinuous CCM, 2-days (i.e., 10 total CCM hours) promoted markedly increased contraction and calcium transient amplitude and altered gene expression profile in ECTs.We previously demonstrated that functionally enhanced commercial 3D ECTs, following 7-weeks of electrical conditioning, displayed increased contraction force when stimulated acutely (i.e., seconds) with various CCM signals in a

F
I G U R E 4 Chronic CCM enhances ECT calcium transient amplitude.(A) Representative calcium transient for Day 8 ECTs following 2-day stimulation treatment and 14-h rest period for the control group (black line) and CCM group (red line) under spontaneous beating conditions.Chronic discontinuous CCM induces increased calcium transient amplitude and no significant difference in diastolic calcium or calcium decay rate.(B) Summary data graphs.Data are mean ± SEM. n = 5 per group.*p < .05(CCM group vs. control group) Ca, calcium; CaT, calcium transient; CCM, cardiac contractility modulation; Vmax, maximum velocity.

, 2 -
day) in vitro CCM evaluation tool and can be used to support safety or performance studies for future CCM device signals as well as other cardiac electrophysiological medical device signals in general.Our report is the first chronic in vitro CCM study using human ECTs.We examined whether chronic CCM affected key aspects of excitation-contraction coupling such as contraction, intracellular calcium handling, and cardiac-specific molecular profile.Here we demonstrate several important findings.(1) The 3D ECT model responds to chronic clinical CCM parameters in vitro as early as 2-days.(2) Chronic CCM results in prolonged functional changes 14-h after CCM was stopped.(3) The CCM group displayed enhanced contraction and intracellular calcium transient amplitude.(4) Likewise, the CCM group displayed altered myocardial gene profile.Furthermore, it will be interesting to study the effects of additional novel cardiac electrophysiology medical device signals.Toward that goal, we are currently evaluating cardiac electrophysiology medical device signals in a variety of innovative in vitro hiPSC models to assist mechanistic elucidation and regulatory decision making.