Differential response to acute ischemia between isolated contracting hearts and hearts perfused with excitation-contraction uncouplers

Abstract

Acute myocardial ischemia triggers electrophysiological changes, including altered cardiac action potential and conduction slowing. Optical mapping is widely used to study these changes, but most experiments employ excitation-contraction uncouplers to suppress contractile motion and motion artifacts. We hypothesized that contraction suppression with these uncouplers masks ischemic effects, leading to misleading results. We compared Langendorff-perfused, non-contracting hearts treated with blebbistatin to contracting hearts under acute ischemic conditions. Optical mapping with emission ratiometry and motion-tracking post-processing minimized motion artifacts, while ischemia was induced by ligating the obtuse marginal branch of the left circumflex coronary artery. Contracting hearts exhibited faster and more pronounced reductions in action potential duration (APD) and action potential triangulation (2 minutes vs. 14 minutes), along with an increased incidence of spatially discordant alternans (SDA). They also displayed steeper APD restitution slopes, whereas these slopes were flattened in non-contracting hearts. These differences may stem from reduced metabolic demands and the absence of mechano-electric feedback in non-contracting hearts. In contrast, contracting hearts, with higher metabolic activity and mechanical feedback, experienced more severe ischemic changes. These findings highlight the limitations of using blebbistatin-treated, non-contracting hearts in electrophysiological research, as critical ischemic effects may be underestimated. This study underscores the need to integrate mechanical and electrical dynamics in preclinical models to accurately replicate ischemic conditions, enhancing the translational relevance of experimental cardiac research.

Introduction

Acute regional myocardial ischemia rapidly perturbs cardiac electrophysiology, altering ventricular activation and repolarization. Hallmark responses include shortening of action potential duration (APD) and slowing of conduction velocity (CV) (1, 2). These responses have been described across species, form a central feature of electrophysiology remodelling during acute ischemia (3–6).

Optical mapping is a widely utilized experimental tool in preclinical cardiac electrophysiology research, offering high spatiotemporal resolution for investigating electrical activity across the heart (7). However, a significant limitation of this technique is the presence of motion artifacts, which compromise data integrity and accuracy. Historically, motion artifacts have been mitigated through the use of pharmacological agents, such as blebbistatin and 2,3-butanedione monoxime (2,3-BDM), which inhibit cardiac contraction (8, 9). While effective, this approach inherently alters the physiological state of the myocardium, including direct interactions with ion channel activities, metabolic demands, and mechano-electrical feedback (MEF) (8, 9). Recent advancements have shifted focus toward integrating computer vision-based methodologies, such as optical flow motion-tracking algorithms, with ratiometric cardiac optical mapping (10–14). These advancements facilitate the analysis of optical mapping data from contracting hearts, eliminating the need for pharmacological motion suppression and offering a more physiologically relevant model for studying cardiac electrophysiology.

While blebbistatin is widely regarded as a preferred excitation–contraction uncoupler due to its minimal direct effects on cardiac electrophysiology (15, 16), recent evidence suggests that its use may significantly alter baseline electrophysiological parameters (10, 17). These effects are thought to arise not from blebbistatin itself, but from the metabolic changes associated with the absence of mechanical contraction. Contracting hearts exhibit higher oxygen consumption and metabolic demand compared to blebbistatin-arrested, non-contracting hearts (9, 18). Therefore, differences observed at baseline between contracting and non-contracting preparations may reflect the altered energetic state of the myocardium under non-contractile conditions, highlighting the importance of considering metabolic context when interpreting electrophysiological data from blebbistatin-treated hearts. The potential for differential effects of blebbistatin on cardiac electrophysiology during acute ischemia has been hypothesized, particularly in comparison to the behavior of contracting hearts. Contracting hearts, with their inherently higher metabolic demands, may experience more pronounced electrophysiological disturbances under ischemic stress. Despite these implications, no studies to date have systematically compared the electrophysiological responses of non-contracting, blebbistatin-treated hearts with those of contracting hearts under conditions of acute ischemia.

This study aims to address this critical gap by directly comparing the electrophysiological responses of contracting and blebbistatin-treated non-contracting hearts during acute ischemia. Through this comparative approach, we seek to elucidate the potential influence of pharmacological motion suppression on the electrophysiological response of the myocardium to ischemic stress. These findings are expected to provide novel insights into the implications of pharmacological motion suppression in experimental cardiac research and to refine interpretation and design of optical-mapping studies of ischemia.

Materials and Methods

All animal experiments were conducted in accordance with the UK Animals (Scientific Procedures) Act of 1986 and the European Union Commission Directive 2010/63/EU.

Langendorff Perfusion of the Whole Rabbit Heart

Twelve mixed-sex New Zealand white rabbits (female, n = 10, 2–2.5 kg, 10-12 weeks old; male, n = 2, 2–2.5 kg, 10-12 weeks old) were sourced from Inotiv, UK. Sedation was induced using Domitor (1 mg/mL, 0.125 mL/kg, Orion Pharma, UK) and Ketavet (100 mg/mL, 0.075 mL/kg, Zoetis, UK). After 15 minutes, the rabbits were euthanized by overdose of anesthesia (Pentobarbitone Sodium, 200 mg/mL, 0.8 mL/kg, Animal Care, UK) administered via the marginal ear vein, combined with 2500 IU of Heparin (Panpharma, UK) to prevent coagulation. Following a thoracotomy, hearts were excised, cannulated at the aorta, and retrogradely perfused using a standard Langendorff system with Tyrode’s solution containing (in mmol/L): NaCl 128.2, CaCl₂ 1.3, KCl 4.7, MgCl₂ 1.05, NaH₂PO₄ 1.19, NaHCO₃ 20, and Glucose 11.1. Hearts were maintained at 37°C in an acrylic chamber at a flow rate of 30ml/min. Electrocardiogram (ECG) signals were continuously recorded via needle electrodes connected to a Bio Amp system (ADInstruments, Australia). A bipolar pacing electrode was positioned on the right ventricle of the heart and connected to a Micropace III system (Micropace, MicropaceEP Ltd, California, USA) to perform the stimulation protocol. Monophasic action potential (MAP) recordings were obtained using an MAP electrode connected to a Bio Amp system and a PowerLab 8/30 data acquisition unit (both from ADInstruments, Australia).

Experimental Protocol

An illustration of the experimental protocol is provided in Figure 1. Following excision and cannulation, the hearts were allowed to stabilize for approximately 30 minutes. Six hearts were injected with the excitation-contraction uncoupler blebbistatin (10 μmol/L, Tocris Bio-Sciences, Cambridge, UK) to suppress contraction (non-contracting heart group), while the remaining six hearts were not (contracting heart group). Blebbistatin was prepared according to recommendations from published work (9). To minimize precipitation during dilution from the DMSO stock, the perfusate was heated to 37 °C and vigorously agitated. An inline 1 μm filter (VWR International Ltd, UK) was incorporated to remove any formed crystals and reduce the risk of coronary microvascular obstruction during perfusion. A voltage-sensitive fluorescent dye, Di-4-ANEPPS (2–5 μmol/L, VWR International Ltd, UK), was subsequently introduced via a side port to stain the myocardium. For pre-ligation restitution pacing, ten hearts underwent pacing at pacing cycle lengths (PCL) of 280, 250, 220, and 200 ms, followed by baseline pacing at a PCL of 280 ms. However, two hearts (one from each group) were subjected solely to baseline electrophysiological measurements at a PCL of 280 ms. To induce acute ischemia, a suture (4-0 Vicryl, Ethicon, UK) was placed around the obtuse marginal coronary artery in the left ventricle, as depicted in Figure 1B (19). After ligation, the flow rate is reduced to 20 ml/min to accelerate ischemia. The hearts were paced at intervals of 2, 6, 10, 14, 18, and 22 minutes, using a PCL of 280 ms at each time point after obtuse marginal artery ligation. Following the 22-minute time point, restitution pacing was repeated for the ten hearts.

Figure 1. Ratiometric Optical Mapping to Assess Action Potential Changes During Ligation.

A. Schematic of the experimental setup for emission ratiometric optical mapping The rabbit heart was stained with the voltage-sensitive dye Di-4-ANEPPS and excited using blue-wavelength light (470 ± 20 nm LEDs). Emitted fluorescence passed through a dichroic mirror (610 nm DMLP) and was split into two emission pathways. Two orthogonally oriented cameras, aligned to image the same field of view, captured the signals: Camera 1 recorded long-wavelength fluorescence (>620 nm), and Camera 2 captured short-wavelength fluorescence using a 535 ± 35 nm bandpass filter. A ratiometric analysis of the captured signals was performed to minimize motion artifacts and improve signal quality. Camera image created in BioRender. Hua, Z. (2025) https://BioRender.com/7h5b77o.

B. Schematic showing optical mapping of cardiac action potentials before and after ligation. Ratiometric optical mapping was performed as described above. The heart was either rendered non-contractile using blebbistatin or maintained in a contracting state. The obtuse marginal (OM) branch of the left circumflex artery was ligated, and the heart underwent electrical pacing before and after ligation.

Emission Ratiometry Optical Mapping

Emission ratiometry was performed for both heart groups. A graphic illustration of the emission ratiometry optical mapping setup is shown in Figure 1A. Briefly, the hearts were illuminated by four high-power LEDs (470 ± 20 nm, Cairn Research, UK). Fluorescence was collected and separated using a dichroic mirror (610nm DMLP, Chroma Technology, Vermont, USA) and captured by two CMOS cameras (128 × 80 pixels, Sci-measure, USA), each fitted with an emission filter (> 620nm for camera 1, and 535 ± 35 nm for camera 2, Chroma Technology, Vermont, USA) with a sampling frequency of 250Hz (4ms camera exposure time) (20).

Optical Mapping and MAP Data Processing

Optical mapping data were processed and analyzed using custom-made Python scripts. A flowchart illustrating the motion correction process for contracting hearts is presented in Figure 2A. Initially, the original video from camera 1 underwent contrast enhancement using a kernel size of 5 to mitigate light intensity variations caused by the voltage-sensitive fluorescent dye, thereby ensuring constant intensity over time - a fundamental assumption of the optical flow algorithm. The motion of the heart in the contrast-enhanced camera 1 video was subsequently tracked using the Farneback optical flow algorithm, implemented via the optimap(https://github.com/cardiacvision/optimap/) library (12, 21). A reference frame was selected prior to depolarization for accurate motion tracking. The displacement vectors derived from the tracking algorithm were applied to both original camera 1 and camera 2 videos to ensure consistent spatial warping across both datasets. The warped camera 1 and camera 2 videos were ratioed to generate motion-corrected videos for subsequent analysis. In this context, motion correction involved motion tracking followed by image warping and ratiometry. Examples of these motion-corrected signals are shown in Figure 2B. For non-contracting hearts, motion tracking and warping were not required, as there was no contractile movement. Instead, only ratiometry was performed to enhance fluorescent signal intensity and eliminate common artifacts, such as inhomogeneous dye loading and uneven illumination. Following motion correction or direct ratiometry, all videos underwent the same post-processing steps: spatial filtering using a 5×5 box kernel, temporal filtering via a moving average filter with a 5-frame window, and baseline drift correction using a moving average filter with a window length corresponding to the pacing cycle length.

Figure 2. Motion Correction Workflow for Contracting Heart Optical Mapping and Comparison of Optical Signals Pre- and Post-Correction.

A. Schematic representation of the motion correction process. Images from Camera 1 undergo contrast enhancement to reduce light intensity variations. Farneback, an optical flow-based motion tracking algorithm, is applied to generate displacement vectors that map heart motion across frames. These vectors are used to warp both Camera 1 and Camera 2 images, resulting in motion-corrected frames. Emission ratiometry is performed between the motion-tracked images from both cameras, enhancing signal accuracy by minimizing other artifacts.

B. Comparison of optical action potentials before and after motion correction in contracting hearts. A representative image of the heart with selected regions of interest is shown, along with the corresponding optical signals. The data from the contracting heart underwent both motion tracking and ratiometric processing.

C. Comparison of optical action potentials before and after ratiometry in non-contracting hearts. As the heart is non-contractile, only ratiometry was applied to remove residual artifacts. A representative image shows the selected regions of interest along with the corresponding optical signals.

MAP data (originally sampled at 1 kHz) were downsampled to 250 Hz to match the sampling rate of the optical recordings for direct comparison. Subsequently, they underwent similar temporal filtering and baseline correction as applied to the optical data.

To evaluate the accuracy of motion correction, simultaneous MAP and optical action potential (OAP) signals were compared. Because the MAP electrode occupied a portion of the field of view, optical recordings could not be obtained from the exact same site; therefore, OAP signals from a neighboring region were used for comparison. The MAP and OAP signals showed good agreement in waveform morphology and timing both pre- and post-ligation (Figure S1).

Electrophysiology Parameters Analysis

All analyses were conducted below the level of ligation (i.e. the ischemic area and the corresponding area of pre-ischemia). APD, action potential morphology, and CV were evaluated at baseline and at each post-ligation time point as described above (n=6 in each group). APD₅₀ and APD₈₀, representing the time from activation (defined as the maximum rate of change of voltage, dv/dt) to 50% and 80% repolarization, respectively, were calculated for comparative analysis. Action potential morphology was assessed using the triangularity index, defined as the ratio of the area outside the triangle formed by the action potential peak and the repolarization time at 80% repolarization to the area of the triangle itself (22). CV was calculated using the single-vector method, by computing the ratio of spatial distance between two points along the dominant direction of wave propagation to the time delay obtained from the activation map. The measurement was constrained to the region below the ligation point to specifically assess conduction within the ischemic area. CV was not calculated from the pacing electrode location to the point of maximum activation time because the pacing electrode was placed at the back of the heart to assist with motion tracking, as changes in brightness due to its presence within the imaging field of view could interfere with accurate tracking. Additionally, the analysis was intentionally focused on wave propagation within the ischemic region, rather than across the entire heart. In cases where multiple wavefronts were present, CV was computed in a region where wavefronts remained distinct and unidirectional, avoiding areas where wavefronts merged or changed direction. Given the variability in baseline APD, triangularity index, and CV among individual hearts, percentage changes between each post-ligation time point and the pre-ligation baseline were calculated to quantify reductions in APD, triangularity, and CV between contracting and non-contracting hearts.

The occurrence of alternans and the restitution relationship between APD and diastolic interval (DI) were analyzed from both pre- and post-ligation restitution pacing recordings (n=5 in each group). Action potential duration (APD) alternans was systematically characterized based on beat-to-beat variations in APD₇₀. Alternans was defined as a difference in APD₇₀ (ΔAPD₇₀) exceeding 10 milliseconds, persisting for at least five consecutive beats, and affecting a minimum of 5% of the mapped area (23). To further classify the spatial organization of alternans, regions meeting these criteria were categorized as spatially concordant alternans (SCA) when APD fluctuations occurred uniformly across the affected area, or as spatially discordant alternans (SDA) when out-of-phase regions were separated by a nodal line. This approach provided a consistent and quantitative framework for evaluating alternans patterns under varying experimental conditions. Restitution analysis was performed using a 5×5 pixel grid located immediately below the ligation point. A mono-exponential function APD_n+1 = a − b × exp(−DI_n/τ) was fitted to restitution curves using MATLAB’s curve-fitting toolbox (MATLAB 2023a) (24). The maximum gradient from each restitution curve was calculated for comparison. CV restitution in the ischemic region was analyzed in the same manner as APD, using mono-exponential curve fitting and calculating the maximum restitution gradient. Moreover, ECG parameters such as QRS duration and absolute ST height were analyzed under pre- and post-ligation conditions in contracting and non-contracting hearts using LabChart 8 (ADInstruments, Dunedin, New Zealand).

Statistical Tests

Repeated-measures one-way ANOVA with Dunnett’s multiple comparisons test was used to compare APD₅₀, APD₈₀, triangularity index, CV, QRS duration, and ST height at baseline (pre-ligation) and at each post-ligation time point. For percentage changes in APD, triangularity, and CV from pre-ligation data, a two-way ANOVA was conducted. Fisher’s exact test was used to assess the incidence of alternans. A paired t-test was applied to compare the maximum gradients of the electrical restitution curves between pre- and post-ligation data. APD and CV from the non-ischemic region were compared across baseline (pre-ligation) and multiple post-ligation time points using the Friedman test (non-parametric), followed by Dunn’s multiple comparisons test. Two-way ANOVA was used to compare baseline APD values (APD₅₀ and APD₈₀) between contracting and non-contracting hearts. An unpaired t-test was applied to compare baseline triangularity index, APD restitution gradient, CV, and CV restitution gradient between contracting and non-contracting hearts. All statistical analyses were performed using Prism 10, with statistical significance defined as a p-value < 0.05.

Results

Action Potential Changes in Acute Ischemia

Figure 3A presents representative optical action potential traces and APD₈₀ maps from the ischemic area of the heart in pre-ligation and post-ligation (2 mins, 10 mins, 18 mins and 22 mins) conditions in both non-contracting (top row) and contracting hearts (bottom row). Action potential signals and APD maps from ischemic and non-ischemic regions are also shown in Figure S2. APD shortening was observed following ligation in both non-contracting and contracting hearts. However, the temporal dynamics of this phenomenon differed significantly between the two groups. In non-contracting hearts, APD shortening occurred gradually, with statistically significant reductions in both APD₅₀ (Figure S3A) and APD₈₀ observed from 14 minutes post-ligation, persisting thereafter (Figure 3B). In contrast, contracting hearts exhibited a more pronounced and rapid reduction in APD, with significant shortening of both APD₅₀ and APD₈₀ detected as early as 2 minutes post-ligation.

Figure 3. Changes in Action Potential Duration (APD₈₀) Following Ligation in Contracting and Non-Contracting Hearts.

A. Representative optical action potential traces and APD₈₀ maps for non-contracting (top row) and contracting (bottom row) hearts at pre-ligation and at 2, 6, 10, 14, 18, and 22 minutes post-ligation. Non-contracting hearts show a gradual progressive reduction in APD across the ischemic region, while contracting hearts demonstrate more pronounced APD shortening.

B. Temporal profiles of APD₈₀ in non-contracting (left) and contracting (right) hearts from the ischemic region. Data are presented as mean ± SD.C. Percentage change in APD₈₀ relative to pre-ligation baseline at each time point for non-contracting (gray bars) and contracting hearts (blue bars).

Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001; ns = not significant.

At baseline (pre-ligation), APD and triangularity values were significantly longer in non-contracting hearts compared to contracting hearts, reflecting intrinsic differences in repolarization under mechanical load (Figure S4A–S4B). To account for this disparity, both baseline and post-ligation values were assessed, and percentage changes were calculated to more directly compare ischemia-induced APD shortening between groups. To quantify the extent of APD shortening between the two groups, percentage changes in APD₅₀ (Figure S3B) and APD₈₀ were analysed (Figure 3C). A divergent trend was subsequently observed: from 2 minutes post-ligation until the end of the recording, APD in contracting hearts remained relatively stable, whereas it gradually decreased in non-contracting hearts. At 2 and 6 minutes post-ligation, contracting hearts demonstrated significantly greater reductions in both APD₅₀ and APD₈₀ compared to non-contracting hearts. However, beyond 6 minutes post-ligation, the differences in APD shortening between the groups were not statistically significant.

It is important to note that there were no significant differences in the temporal profiles of APD between pre- and post-ligation conditions in either non-contracting (Figure S5A) or contracting (Figure S5C) hearts within the nonischemic region.

Action potential morphology was quantified using a triangularity index (Figure 4A), defined as the ratio of the area outside the triangle formed by the action-potential peak and the 80% repolarization time to the area of that triangle; lower values indicate a more triangular morphology.

Figure 4. Changes in Action Potential Morphology (Triangularity Index) Following Ligation in Contracting and Non-Contracting Hearts.

A. Triangularity index maps (top) for non-contracting and contracting hearts, and a representative action potential signal (bottom) showing labeled regions inside and outside the triangle, used for triangularity index calculation.Traces illustrate changes in action potential shape over time (pre-ligation and at 2, 6,10,14 and 18 minutes post-ligation).

B. Time course of triangularity index values in non-contracting (left) and contracting (right) hearts. Data are presented as mean ± SD.

C. Percentage change in triangularity index relative to pre-ligation baseline at various time points for non-contracting (gray bars) and contracting hearts (blue bars). Contracting hearts show a greater reduction in triangularity index compared to non-contracting hearts at early time points post-ligation.

Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001; ns = not significant.

Under ischemic conditions, a more triangular action potential morphology was observed in both contracting and non-contracting hearts. However, the temporal dynamics of this change differed significantly between the groups. In contracting hearts, the triangularity index decreased rapidly, with significant reductions observed as early as 2 minutes post-ligation. In non-contracting hearts, by contrast, the response was more gradual, with significant reductions in the triangularity index detected only after 14 minutes post-ligation (Figure 4B).

As shown in Figure 4C, contracting hearts exhibited a greater initial reduction in the triangularity index compared to non-contracting hearts. However, at later time points, the magnitude of changes in the triangularity index was comparable between the two groups.

In addition to optical APD measurements, ECG parameters were assessed to provide complementary insight into electrophysiological changes. QRS duration showed no statistically significant change in either contracting or noncontracting hearts. However, absolute ST height showed a statistically significant increase in contracting hearts at 18 minutes post-ligation compared to noncontracting hearts, indicating a difference in ischemic behavior between the two conditions (Figure S6).

Conduction Changes in Acute Ischemia

Figure 5A illustrates the representative activation maps demonstrating the impact of ischemia induced by coronary ligation (indicated by the black cross in post-ligation maps) on CV in both contracting and non-contracting hearts. Following ligation, the reduction in the spacing between isochronal lines and the increased density of isochronal contours indicate a significant slowing of CV in both groups.

Figure 5. Changes in Conduction Velocity (CV) Following Ligation in Contracting and Non-Contracting Hearts.

A. Representative activation maps before and at 22 minutes after ligation in non-contracting (left) and contracting (right) hearts. The pink cross indicates the ligation point and visible CV slowing is observed below the level of ligation, as evidenced by reduced spacing and increased density of isochronal contours. CV was measured at a pacing cycle length of 280 ms.

B. Time course of CV changes in non-contracting (left) and contracting (right) hearts. Data are presented as mean ± SD, showing a significant reduction in CV following ligation.

C. Percentage change in CV relative to pre-ligation baseline for non-contracting (gray bars) and contracting (blue bars) hearts. Ligation-induced CV changes are not statistically significant between contracting and non-contracting hearts.

Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001; ns = not significant.

Unlike the temporal dynamics observed in action potential changes, CV reduction was significant in both contracting and non-contracting hearts as early as 2 minutes post-ligation (Figure 5B). Although not statistically significant at any individual time point, contracting hearts exhibited a larger initial drop in conduction velocity (CV) following coronary ligation, which then remained relatively stable over time. In contrast, non-contracting hearts showed a more gradual and progressive decline in CV throughout the ischemic period. As a result, the magnitude of CV reduction in non-contracting hearts eventually surpassed that of contracting hearts, despite the absence of statistically significant differences between groups at any measured time point. (Figure 5C). Notably, there were no significant differences in the temporal profiles of CV between pre- and post-ligation conditions in either noncontracting (Figure S5B) or contracting (Figure S5D) hearts within the nonischemic region. Moreover, comparison of CV restitution slope in noncontracting hearts showed a decreasing trend post-ligation, although it did not reach statistical significance. In contrast, the CV restitution slope significantly decreased post-ligation in contracting hearts (Figure S7).

Action Potential Alternans Incidence in Acute Ischemia

Representative examples of SCA and SDA in contracting and non-contracting hearts post-ligation are illustrated in Figure 6A and Figure 6B.

Figure 6. Incidence of Concordant and Discordant Alternans in Contracting and Non-Contracting Hearts Before and After Ligation.

A. Representative action potential duration (APD) maps and optical action potential traces from non-contracting hearts, illustrating spatially concordant (top row) and spatially discordant alternans (bottom row). In concordant alternans, APD maps and traces from two consecutive beats show consistent spatial changes. In discordant alternans, they exhibit alternating spatial patterns. Concordant and discordant alternans occurred at 200 ms and 220 ms, respectively. Representative examples were obtained from different hearts.

B. Representative APD maps and optical traces from contracting hearts, also demonstrating spatially concordant (top row) and discordant alternans (bottom row). Concordant and discordant alternans occurred at pacing cycle lengths of 220 ms and 200 ms, respectively. Representative examples are from the same heart.

C. Incidence of alternans in non-contracting and contracting hearts before (pre-ligation) and after ligation (post-ligation). Bars indicate the number of hearts with alternans (gray) and without alternans (orange). Alternans incidence increases post-ligation in both contracting and non-contracting hearts, but only contracting hearts show statistical significance (p = 0.0476).

D. Incidence of discordant alternans in non-contracting and contracting hearts before (pre-ligation) and after ligation (post-ligation). Bars indicate the number of hearts exhibiting discordant alternans (gray) and no discordant alternans (orange). Discordant alternans are more prevalent in contracting hearts post-ligation, with statistical significance (p = 0.0476).

A statistically significant increase in APD alternans was observed only in contracting hearts (p = 0.0476), whereas the increase in non-contracting hearts did not reach statistical significance (p = 0.1667), (Figure 6C). Further analysis revealed differential patterns in the occurrence of SDA. In non-contracting hearts, the incidence of SDA did not show a significant increase following ligation. In contrast, contracting hearts demonstrated a significant increase in SDA incidence post-ligation (p = 0.0476), as shown in Figure 6D.

Restitution Curve Changes Following Ligation

Figure 7. shows representative electrical restitution curves in non-contracting (A) and contracting (C) hearts, recorded pre- and post-ligation. Panels B and D display the corresponding maximum slopes of the restitution curves for non-contracting and contracting hearts, respectively. In non-contracting hearts, the maximum slope was reduced post-ligation, resulting in a flattened restitution curve (from 1.43 ± 0.39 to 0.80 ± 0.51, p = 0.0225) (Figure 7B). Conversely, contracting hearts exhibited an increase in the slope of the restitution curve, leading to a steeper restitution curve following ligation (from 1.08 ± 0.60 to 2.87 ± 0.81, p = 0.0442) (Figure 7D).

Figure 7. APD Restitution Curves and Maximum Restitution Gradients in Contracting and Non-Contracting Hearts Before and After Ligation.

A. Representative APD restitution curves for non-contracting hearts before (magma) and after (blue) ligation. The curves illustrate the relationship between action potential duration (APD) and the preceding diastolic interval (DI). Black dots represent mean data points, with black error bars indicating standard deviation. Fitted curves are shown in the corresponding colors. A flattening of the restitution curve was observed after 22 minutes of ligation.

B. Comparison of the maximum restitution gradients before and after ligation in non-contracting hearts. Bars represent mean ± SD. A significant decrease in the maximum restitution gradient was observed after ligation.

C. APD restitution curves for contracting hearts before (magma) and after (blue) ligation. Plot structure and labeling are as described in Panel A. Following ligation, a steepening of the restitution curve was observed.

D. Maximum restitution gradients before and after ligation in contracting hearts. Bars represent mean ± SD. A significant increase was observed post-ligation.

Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001; ns = not significant.

Discussion

Traditional optical mapping experiments frequently utilize pharmacological agents, such as blebbistatin, to suppress cardiac contraction and mitigate motion artifacts. While effective in reducing artifacts, this approach abolishes mechanical activity and lowers metabolic demands, both of which influence baseline cardiac electrophysiology. This study highlights the impact of contraction suppression on electrophysiological response to acute ischemia. Key findings include: 1) APD, action potential morphology, exhibited more rapid and pronounced changes in contracting hearts early post-ligation; 2) acute ischemia significantly slowed CV in both contracting and non-contracting heart; 3) spatially discordant alternans occurred more frequently in contracting hearts after ligation; 4) electrical restitution slopes show opposing trends in acute ischemia, with steepening in contracting hearts and flattening in non-contracting hearts.

Action Potential and Conduction Changes during Acute Ischemia

The differences in APD and triangulation between contracting and non-contracting rabbit hearts were most pronounced during the initial phase of acute ischemia, specifically within the first 10 minutes following coronary ligation. This early phase is a critical window during which rapid electrophysiological remodeling unfolds in response to abrupt loss of perfusion. The timing and extent of these changes may be influenced by the heart’s metabolic status, which is substantially influenced by mechanical contraction. Contracting hearts, which continue to generate mechanical force, exhibit higher energetic demands. Although ATP and other energetic profiles were not measured in our study, previous Langendorff perfusion studies have shown that contracting hearts consume more oxygen and accumulate NADH more rapidly than blebbistatin-arrested, non-contracting hearts - indirect evidence of more rapid ATP depletion in the presence of contraction, likely leading to accelerated metabolic stress and possibily early activation of ATP-sensitive potassium (KATP) channels (18).

As a result, contracting hearts in our study exhibited APD shortening as early as 2 minutes post-ligation, followed by relative stabilization for the remainder of the recording. This temporal profile is similar to an in vivo rabbit study in which APD₉₀, measured using monophasic action potentials, shortened rapidly within the first 5 minutes of coronary artery occlusion and then stabilized—suggesting a plateau in repolarization despite ongoing ischemia (25). In contrast, the lower metabolic burden in non-contracting hearts appeared to delay these effects, resulting in a more gradual decline in APD and triangulation over time, although the difference in terminal APD values between the two groups was not statistically significant.

Interestingly, the minimum APD in contracting hearts was slightly higher than in non-contracting hearts, despite the more rapid onset of shortening. This observation may be explained by mechanoelectric feedback mechanisms. Stretch-activated channels, which are engaged by mechanical deformation during contraction, have been shown to modulate membrane potential and repolarization. While mechanoelectric feedback has been linked to APD shortening during acute ventricular stretch, other studies have shown that sustained pressure or compression can prolong APD, suggesting a context-dependent role (e.g., Valsalva or aortic compression models) (26, 27). Although these mechanical perturbations are more extreme than those in a working Langendorff heart, it is plausible that beat-to-beat mechanical activity exerts subtle modulating effects on repolarization, limiting the extent of APD abbreviation in contracting hearts.

Additional support for these dynamics comes from prior studies in isolated rabbit and human ventricular myocytes and Langendorff-perfused rabbit hearts. In these models, transient APD prolongation was observed during the first few minutes of metabolic inhibition, attributed to suppression of the transient outward current (Ito), followed by APD shortening driven by KATP channel activation (28, 29). In our study, a similar transient prolongation trend was seen at 2 minutes in non-contracting hearts, though it did not reach statistical significance. No such prolongation was observed in the contracting hearts, possibly because the APD dynamics had already advanced past this transient phase by the first measurement point.

Taken together, these observations highlight the importance of capturing early ischemic responses when evaluating repolarization and conduction. Although terminal APD values between contracting and non-contracting hearts may converge later in ischemia, they arise from distinctly different temporal trajectories. This distinction emphasizes the physiological relevance of mechanical contraction in shaping ischemic electrophysiology and underscores the value of time-resolved assessments in future studies.

Although the differences in CV between contracting and non-contracting hearts were not statistically significant at any time point, the temporal patterns diverged in a manner that closely mirrored those observed for APD. Both groups exhibited a clear reduction in CV as early as 2 minutes after coronary ligation, suggesting that early ischemic conduction slowing is a shared response, likely driven by mechanisms such as membrane depolarization and subsequent sodium channel inactivation (30), extracellular potassium accumulation (31, 32), and early gap junction uncoupling (33). These ionic changes occur rapidly after coronary occlusion and are generally considered to be largely independent of mechanical contraction, which may explain the similarity in early CV slowing between the two preparations.

However, while CV in contracting hearts stabilized after the initial drop and remained relatively constant over the course of the study, non-contracting hearts showed a more gradual and progressive decline in CV, echoing the temporal profile observed in APD. This pattern suggests that although mechanical activity may not exert a large direct influence on CV under acute ischemic conditions, it may accelerate the onset of metabolic stress, thereby indirectly affecting conduction dynamics over time.

Although many mechanistic pathways (e.g. ATP/NADH dynamics, ionic channels such as Ito and KATP, connexin remodeling, or tissue motion/load) could contribute to the observed differences, we did not measure any of these directly in the present study. Therefore, all mechanistic interpretations offered here are hypothesis-generating and should be tested in future work.

Spatially Discordant Alternans in Contracting Hearts After Acute Ischemic Events

Numerous studies have reported the occurrence of alternans under acute ischemic conditions (34–37). The formation of alternans is complex, involving cellular mechanisms like membrane ion channel dynamics, including APD restitution and cellular calcium dynamics described by the 3R theory, as well as spatial conduction dynamics such as CV restitution (38, 39). The multifactorial nature of alternans formation makes it challenging to elucidate the precise mechanisms underlying the increased prevalence of SDA in contracting hearts. Nonetheless, insights from simulation studies may provide some explanations. Radszuweit et al. demonstrated in their simulation that small stretch-activated currents, associated with mechanical contraction, can exert significant effects, potentially inducing a transition from in-phase to off-phase alternations (i.e., from SCA to SDA) (40). Similarly, Hazim et al. showed that the bifurcation of alternans can be modulated by the strength of stretch-activated currents in their simulations. Considering that acute ischemia distorts cardiac mechanics, it is plausible that mechanical factors, beyond the traditionally considered electrical factors, may contribute to the observed increase in SDA incidence (41). However, further studies are required to investigate this possibility, highlighting the importance of incorporating mechanical contraction into future experimental and pathological models.

Electrical Restitution in Contracting Hearts After Acute Ischemia

APD–DI restitution quantifies how APD varies with the preceding DI (42). We report the maximum slope as a descriptive electrophysiological metric. Although a steep slope (gradient >1) has been hypothesized to favour cardiac arrhythmia in some settings, its predictive value is context-dependent and influenced by other factors such as short-term electrical “memory” and calcium handling; therefore, we do not infer arrhythmic risk from these measurements in this study (42).

Acute ischemia is typically associated with flattening of the restitution curve, as observed in non-contracting hearts (34, 43, 44). Several factors can contribute to such flattening, including short-term electrical “memory” effects and spatial heterogeneity in extracellular potassium (45, 46). In contrast, the contracting hearts in this study demonstrated a steepening of the restitution curve. We interpret these opposing trends as differences in electrophysiological dynamics between contracting and non-contracting preparations; the mechanistic determinants warrant dedicated study.

Limitations and Future Directions

This study has several limitations that merit consideration. First, while the importance of incorporating mechanical information is well-recognized, the optical flow motion tracking algorithms employed in this study detect the apparent motion of brightness patterns in a sequence of images rather than the true physical motion field. Moreover, these algorithms are unable to capture the three-dimensional structural and mechanical dynamics of the heart. Using a panoramic optical mapping system may address this limitation by enabling more comprehensive assessments of cardiac mechanical and structural behavior (47).

Second, the restitution analysis conducted in this study was restricted to the region below the level of ligation (i.e. ischemic zone), as this area provides a consistent geometric reference point. This restriction precluded the evaluation of spatial variations in restitution properties, which are known to vary across different regions of the heart (48). Such variations are critical for a comprehensive understanding of cardiac electrophysiology.

Third, the CV analysis utilized a single-vector approach, despite ischemia being associated with highly heterogeneous conduction patterns. While a more localized, vector-based assessment might provide additional mechanistic insight, such an approach was limited in this study by two main factors: the relatively low sampling rate of 250 frames per second (fps) and the complex bidirectional propagation of activation waves across the imaged heart surface. These limitations hinder the accurate resolution of local conduction dynamics. We acknowledge that a frame rate of 250 fps is not ideal for CV analysis under baseline (pre-ligation) conditions, where rapid conduction allows only a few isochronal lines to be captured. However, following coronary ligation, conduction slows substantially, and the same temporal resolution becomes more appropriate for estimating CV. Thus, while our single-vector approach has limitations, it remains sufficient for capturing meaningful trends in conduction slowing during acute ischemia.

Finally, during ischemia, several studies have demonstrated the presence of post-repolarization refractoriness, indicating that APD is not always a reliable surrogate for tissue refractoriness (49, 50). Accurate assessment of refractoriness requires an S1–S2 pacing protocol. While such protocols are straightforward to implement in non-contracting, blebbistatin-treated hearts, applying them in contracting hearts poses a significant challenge: the extra stimulus (S2) often induces abrupt, non-periodic mechanical deformation that disrupts the underlying motion field. This sudden distortion compromises the performance of our current Farnebäck-based optical flow algorithm. As a result, we were unable to reliably measure true refractoriness in the contracting preparation. Developing motion-tracking approaches that are robust to such perturbations is a clear priority for future work.

Conclusion

This study provides key insights into the electrophysiological differences between contracting hearts and blebbistatin-treated non-contracting hearts under acute ischemic conditions. Using advanced optical mapping and computational analysis, we demonstrate that suppression of mechanical contraction significantly alters the myocardial response to ischemic stress. Notable findings include more rapid and pronounced APD shortening, a higher incidence of spatially discordant alternans (SDA), and a steeper APD restitution slope in contracting hearts, along with a trend toward greater conduction velocity (CV) impairment in non-contracting hearts during the later stages of ischemia.

The accelerated APD shortening, increased SDA prevalence, and steeper restitution observed in contracting hearts highlight the physiological relevance of preserving mechanical contraction when studying ischemia-induced electrophysiological remodeling. Importantly, differences in APD and triangulation between contracting and non-contracting hearts were most pronounced during the early phase of ischemia (within the first 10 minutes), emphasizing this window as a critical period for mechanistic investigation. These observations suggest that mechanical contraction plays a key role in shaping the temporal dynamics of repolarization changes, and that omitting this factor may limit our understanding of mechano-electric interactions. Future research should aim to incorporate three-dimensional mechanical measurements and enhance the spatial and temporal resolution of optical mapping to better capture the complex coupling between mechanical and electrical behavior during ischemia. Such advancements may ultimately refine experimental design and mechanistic interpretation in optical mapping studies of acute ischemia.

While pharmacological uncouplers such as blebbistatin are highly effective at minimizing motion artifacts during optical mapping, they also eliminate mechanical contraction and its associated energetic demands and feedback mechanisms. This can influence the trajectory of ischemic remodeling, particularly in the early stages when mechanical and metabolic responses are tightly linked. Nevertheless, non-contracting heart models remain valuable for isolating metabolic contributions and facilitating high-resolution imaging. When used in conjunction with contracting models, they offer complementary perspectives that help disentangle the complex interplay between mechanical and metabolic influences on cardiac electrophysiology. Collectively, our findings advocate for the use of physiologically integrated models to capture the full spectrum of mechano-electric interactions during acute ischemia, while also recognizing the mechanistic insights gained from simplified, non-contracting systems.

New & Noteworthy.

This study highlights key differences in acute ischemic responses between contracting and blebbistatin-treated non-contracting rabbit hearts. Contracting hearts showed faster, more severe action potential duration reductions, increased spatially discordant alternans, and steeper restitution slopes, emphasizing the role of mechano-electric feedback and higher metabolic demands. These findings challenge reliance on non-contracting models in electrophysiological research, underscoring the need for models integrating mechanical and electrical dynamics to improve the translational relevance of ischemic studies.

Data Availability

The data underlying this article will be shared upon reasonable request to the corresponding author.

Detailed code and tutorials for the optimap library can be found at the following link: https://cardiacvision.github.io/optimap/main/tutorials/