Mitochondrial depolarization and electrophysiological changes during ischemia in the rabbit and human heart

Abstract

Instability of the inner mitochondrial membrane potential (ΔΨ_m) has been implicated in electrical dysfunction, including arrhythmogenesis during ischemia-reperfusion. Monitoring ΔΨ_m has led to conflicting results, where depolarization has been reported as sporadic and as a propagating wave. The present study was designed to resolve the aforementioned difference and determine the unknown relationship between ΔΨ_m and electrophysiology. We developed a novel imaging modality for simultaneous optical mapping of ΔΨ_m and transmembrane potential (V_m). Optical mapping was performed using potentiometric dyes on preparations from 4 mouse hearts, 14 rabbit hearts, and 7 human hearts. Our data showed that during ischemia, ΔΨ_m depolarization is sporadic and changes asynchronously with electrophysiological changes. Spatially, ΔΨ_m depolarization was associated with action potential duration shortening but not conduction slowing. Analysis of focal activity indicated that ΔΨ_m is not different within the myocardium where the focus originates compared with normal ventricular tissue. Overall, our data suggest that during ischemia, mitochondria maintain their function at the expense of sarcolemmal electrophysiology, but ΔΨ_m depolarization does not have a direct association to ischemia-induced arrhythmias.

mitochondria are well recognized for their roles in both energy production and cell death and have recently been implicated as mediators of ventricular arrhythmias (1–3, 6–8, 35). A critical parameter of mitochondrial function is the tightly regulated inner mitochondrial membrane potential (ΔΨ_m). Stable ΔΨ_m is necessary for mitochondrial function, whereas hyperpolarization or depolarization through excessive movement of ions (i.e., Ca²⁺, ROS, and H⁺) across the mitochondrial membranes is indicative of a pathophysiological state. Stabilization of ΔΨ_m has been shown to mitigate ventricular arrhythmias (1, 5), reduce apoptosis and necrosis (17, 18, 30), and protect against ischemic heart failure (22).

Ischemia involves complex changes across metabolic and electrophysiological systems. The lack of O₂ during ischemia causes metabolic pathways to switch from oxidative phosphorylation to glycolysis for ATP production. To maintain ΔΨ_m, ATPase synthase reverses direction, consuming dwindling cytosolic ATP and competing with ionic pumps and contractile proteins for energy (32, 40). With increased cytosolic ADP, sarcolemmal ATP-sensitive K⁺ (K_ATP) channels have a greater probability of being open and produce an outward K⁺ current (33). This leads to action potential (AP) duration (APD) shortening and depolarization of the resting membrane potential, which results in a reduction in excitability and conduction block (16).

Measuring ΔΨ_m in the intact myocardium is not trivial and has yielded conflicting results across different methodologies. Several groups have monitored ΔΨ_m using either wide-field optical mapping (21, 26, 38) or microscopy, including two-photon laser-scanning microscopy (13, 23, 28, 37) and confocal microscopy (38). Wide-field optical mapping studies of rat hearts have characterized ΔΨ_m depolarization as prompt (<1 min) and spatially spreading as an actively propagating unidirectional wave (21, 26), whereas two-photon laser-scanning and confocal microscopy has shown delayed [i.e., >15 min (28), >50 min (23), 10–20 min (37), 27–69 min (38)] depolarization that is sporadic within clusters of cells (13, 28, 37). The collapse of ΔΨ_m has been thought to contribute to arrhythmogenic substrates during metabolic stress (1, 42); however, the relationship between ΔΨ_m depolarization and electrophysiological changes in the intact heart is unknown.

In this report, we present a novel method for simultaneous optical imaging of sarcolemmal transmembrane potential (V_m) and ΔΨ_m. We monitored these parameters in mouse, rabbit, and human preparations to determine the relation between ΔΨ_m and electrophysiology during ischemia. Our data indicate that ischemia-associated cardiac contracture induces artifacts in ΔΨ_m, but when contracture is avoided, ΔΨ_m depolarization is sporadic, reversible, and asynchronous with electrophysiological changes. We assessed the spatial association between ΔΨ_m depolarization and electrophysiological changes to determine the role that ΔΨ_m plays in arrhythmia substrate development. Finally, we quantified ΔΨ_m during ischemia-induced focal arrhythmias to establish if mitochondrial function differs in the focal origin compared with the remaining myocardium.

METHODS

Optical Mapping

An automated dual-camera imaging system was designed for simultaneous optical mapping of ΔΨ_m and V_m. Optical mapping was performed with two orthogonal MiCAM ULTIMA-L CMOS cameras (SciMedia, Costa Mesa, CA), which have high spatial (100 × 100 pixels, 300 μm/pixel) and temporal (500 frame/s) resolution. A dichroic mirror (635-nm cutoff, Omega Optical, Brattleboro, VT) separated fluorescence signals between cameras. RH-237 (500 nmol/l, Life Technologies, Grand Island, NY) provided fluorescence signals for V_m, whereas tetramethylrhodamine methyl ester (TMRM; 150 nmol/l; Life Technologies) provided fluorescence signals for ΔΨ_m (FΔΨ_m). Once administered, both dyes were continuously perfused. FΔΨ_m was band-pass filtered (590 ± 10 or ± 25 nm, Thorlabs, Newtown, NJ), and V_m was long-pass filtered (>700 nm; Thorlabs). A single light emitting diode (LED; 520 ± 5 nm, Prizmatix, Southfield, MI) provided excitation light, and a programmable controller (CompactRio, National Instruments, Austin, TX) was used for precise temporal triggering of the LED and cameras. The excitation-contraction uncoupler blebbistatin (5–10 μmol/l, Caymen Chemical, Ann Arbor, MI) was added to the perfusate to minimize motion-induced artifacts.

Tissue Preparation

The present study was approved by the Institutional Review Board of Washington University School of Medicine. Human subjects gave informed consent. Donated hearts were provided by Barnes-Jewish Hospital and Mid-America Transplant Services (S. Louis, MO). The Institutional Animal Studies Care and Use Committee of Washington University approved the animal protocols of this study, which involved C57BL/6J mice (n = 4) and New Zealand White rabbits (n = 14, 4–5 mo old, 3–4 kg).

Mice were anesthetized with a ketamine-xylazine cocktail (ketamine: 80 mg/kg body wt and xylazine: 10 mg/kg body wt) and heparinized (100 units) by intraperitoneal injection. Rabbits were anesthetized intravenously with a cocktail made of pentobarbital sodium (80 mg/kg body wt) and 1,000 units of heparin. Hearts were removed after a midsternal incision. As previously reported (15, 25), explanted human hearts (n = 7) were arrested with cardioplegic solution [containing (in mmol/l) 10 NaCl, 1.2 CaCl₂, 16 KCl, 16 MgCl₂, and 10 NaHCO₃] in the operating room and were maintained at 4–7°C during a 15-min transportation period to the research laboratory. Donated hearts were then dissected into left ventricular (LV) wedge preparations.

Both heart and wedge preparations were perfused via coronary arteries and superfused with warmed (37°C), oxygenated (95% O₂-5% CO₂) Tyrode solution [containing (in mmol/l) 128.2 NaCl, 1.3 CaCl₂, 4.7 KCl, 1.05 MgCl₂, 1.19 NaH₂PO₄, 20 NaHCO₃, and 11.1 d-glucose]. Tissue was perfused under constant pressure (70–90 mmHg), and a far-field ECG was recorded (PowerLab 26T, AD Instruments, Colorado Springs, CO) throughout experimentation. Mouse, rabbit, and human tissues were continuously paced at cycle lengths of 120, 250, and 1,000 ms, respectively, for the duration of the experiment.

Experimental Protocols

We applied one of following four experimental protocols.

Protocol 1.

The fluorescence stability of TMRM was tested on mouse hearts (n = 4) to ensure that fluorescence changes did not occur from extraneous factors such as photobleaching.

Protocol 2.

Targeted depolarization of ΔΨ_m was accomplished by adding FCCP (2.5 μmol/l, Sigma-Aldrich, St. Louis, MO), a known proton ionophore that dissipates ΔΨ_m, to the perfusate (n = 5 rabbit hearts).

Protocol 3.

Low-flow ischemia was induced for 90 min by reducing perfusion pressure to 15% of baseline (n = 8 rabbit hearts and n = 7 LV wedges). N₂ was bubbled into the superfusion to prevent O₂ reaching the epicardial surface during ischemia. After 45 min of low-flow ischemia, a bolus of FCCP (100 μmol/l) was administered to collapse any remaining ΔΨ_m. In a subset of rabbit experiments (4 of 8 experiments), the pulmonary artery was cannulated to control perfusate removal from the right ventricle (RV), which helps to maintain the volume of fluid in the RV and the shape of the heart. In human wedge preparations, V_m imaging was performed on all preparations (n = 7), whereas dual ΔΨ_m-V_m imaging was performed on three of seven preparations.

Protocol 4.

No-flow ischemia (30 min) followed by reperfusion was implemented in a single rabbit preparation (n = 1).

Data Analysis

Signal processing.

V_m signals were processed using freely available MATLAB software (24). V_m signals were spatially binned (3 × 3 pixels), filtered (0- to 100-Hz finite impulse response filter), and normalized. Activation times were defined at dV_m/dt_max, and APD values were calculated as the time from activation to 70% of repolarization (APD₇₀). FΔΨ_m signals were spatially binned (3 × 3 pixels) and baseline corrected. FΔΨ_m was normalized from steady-state fluorescence attained just before the administration of FCCP (protocol 2) or low-flow ischemia (protocol 3) to fluorescence after 45 min of FCCP addition. After circulation of FCCP for 45 min, FΔΨ_m was assumed to be at its minimum value (4, 14).

Spatial analysis.

Changes in heart shape were quantified by determining the percent change in pixels over the preparation throughout the experimental protocols. To increase the sample size of protocol 4 and to measure the percent change during no-flow ischemia, the perfusion pump was turned off after the completion of protocol 3 (5 of 8 rabbit preparations). Spatial heterogeneity in FΔΨ_m was calculated on ∼40 × 40-pixel regions by taking the magnitude of the spatial gradient or |d(FΔΨ_m)/dx + d(FΔΨ_m)/dy|. The spatial heterogeneity in FΔΨ_m was compared between ischemia and FCCP administration. The spatial correlation between changes in FΔΨ_m and optical AP (OAP) parameters was also calculated on ∼40 × 40-pixel regions. The percent change in OAP parameters (abscissa) and FΔΨ_m (ordinate) during ischemia were plotted, and the Pearson product-moment correlation coefficient (r) was calculated for 1,600 points (protocol 3, n = 4 rabbit hearts).

Arrhythmia analysis.

Sustained arrhythmias (focal and reentrant) were considered those lasting longer than 30 s and were confirmed both optically and with the ECG. The focal origin was defined as the area of the epicardium activating in the 0 to 0.05 quantile (approximately earliest 5 ms), whereas the epicardium activating in the 0.75 to 1.0 quantile was defined as late activating. During focal activity, FΔΨ_m was calculated in the focal origin, late-activated region, and across the entire epicardial surface.

Statistical Analysis

All data are expressed as means ± SE. P values of <0.05 were considered significant. Statistically significant differences were identified using a Student's t-test or one-way ANOVA. A post hoc Tukey test was performed if ANOVA was significant.

RESULTS

Signal Artifacts: Alterations in Heart Shape

Figure 1 shows a typical example of a rabbit heart during low-flow ischemia. As shown in Fig. 1A, left, at baseline, teal and pink circles emphasize coronary branches, whereas the orange, red, and crimson circles emphasize the edge of the heart. At 1.5 min, low-flow ischemia caused the heart to subtly drift and shrink in size. This led to the coronary branches moving vertically in the optical field of view, which are labeled as blue and magenta circles (Fig. 1A, left). Note that the red and crimson circles are completely off the tissue, whereas the orange circle is located on the curved edge of the heart (Fig. 1A, left). Minimal additional tissue deformation occurred beyond 1.5 min.

Fig. 1.

Signal artifacts in inner mitochondrial membrane potential (ΔΨ_m) fluorescence (FΔΨ_m). A, left: images of an explanted rabbit heart at baseline and at 1.5 and 45 min of low-flow ischemia. The teal, blue, and navy circles as well as pink, magenta, and purple circles highlight the movement of coronary branches. The orange, red, and crimson circles emphasize the change in ventricular shape. Right, FΔΨ_m signals from teal and pink circles (top) and edge circles (bottom). B: binary representation of hearts (left), which was used to quantify imaged surface area (right). Shown is a plot of the percent change in the imaged surface area in rabbit (R) and human (H) preparations during no-flow ischemia (NF_R, n = 6), low-flow ischemia (LF_R, n = 4; LF_H, n = 3), low-flow ischemia with pulmonary artery cannulation (LF-PA_R, n = 4), and administration of FCCP_R (n = 5). Data are expressed as means ± SE.

The drifting and shrinking of tissue is reflected in the optical signals (Fig. 1A, right). The interior pixels (i.e., at the coronary branches) continuously collected data from different heart locations, and, in this case, there was an artificial increase in fluorescence. Pixels near the edge of the tissue rapidly decreased fluorescence because they were no longer focused on the tissue. Such shrinking created an appearance of a propagating wave of mitochondrial depolarization near the edge of the heart. The amount of tissue contracture was quantified from binary images during each protocol (Fig. 1B). When the perfusate flow out of the RV was controlled in rabbit hearts during low-flow ischemia (green vs. gray; Fig. 1B), heart shape was maintained and artifacts in the optical signal were minimized. Supplemental Movie SI in the Supplemental Material further shows the extensive shrinking of a rabbit heart during no-flow ischemia, which made it impossible to monitor FΔΨ_m.¹ In the subsequent sections, only preparations that maintained their shape (green, blue, and red; Fig. 1B) are discussed.

Sporadic FΔΨ_m Depolarization

Figure 2 shows that during the administration of FCCP (A) and low-flow ischemia (B), the reduction in FΔΨ_m is sporadic and doesn't propagate as a unidirectional wave. After 3 min of FCCP administration, only the small area near the black circle had reduced FΔΨ_m. This area expanded over the next 2 min, and the region near the pink circle showed a reduction in FΔΨ_m. These areas, as well as others that showed an early reduction in FΔΨ_m, continued to expand until the entire epicardial surface had reduced FΔΨ_m. Low-flow ischemia showed a different spatial pattern, where after the first minute nearly half the 9 × 9-mm region showed reduced FΔΨ_m. These regions did not expand over the next 10 min but instead slightly shrank in area and showed a decrease in FΔΨ_m. After 15 min, FΔΨ_m expanded from these early reduced regions to occupy the entire surface.

Fig. 2.

Sporadic depolarization of FΔΨ_m during FCCP administration and low-flow ischemia. A: spatial changes in FΔΨ_m during FCCP administration. B: spatial changes in FΔΨ_m during low-flow ischemia. In A and B, a 9 × 9-mm region is shown at six time points (left) as well as a 40-min stacked time series (right). The time after FCCP administration and low-flow ischemia is shown in the top right corner (in min). The pink and black circles highlight areas with an early reduction in FΔΨ_m and provide orientation between two- and three-dimensional displays. Areas with increased FΔΨ_m are masked (white) to enhance the visualization of areas with decreased FΔΨ_m.

Dye Stability and Targeted ΔΨ_m Depolarization

The stability of FΔΨ_m (Fig. 3, A–C) and targeted ΔΨ_m depolarization (Fig. 3, D–F) served as positive controls. The stability of FΔΨ_m was monitored in mouse hearts (n = 4) during the time when a metabolic insult is typically applied (gray region). A representative FΔΨ_m signal and spatial FΔΨ_m maps showed similar FΔΨ_m at the beginning and end of stability testing. Overall, FΔΨ_m increased from 2,100 ± 315 to 2,200 ± 290 arbitrary units during stability testing but was not significantly different. Administration of the potent ΔΨ_m uncoupler FCCP (2.5 μmol/l) caused a substantial reduction in FΔΨ_m. The change in FΔΨ_m was monitored in rabbit hearts (n = 5), and representative FΔΨ_m signals and maps are shown in Fig. 3, D and E. FΔΨ_m maps at 30 s, 15 min, and 30 min showed a near-homogenous change in FΔΨ_m, which is shown in Fig. 3F.

Fig. 3.

Dye stability test and targeted ΔΨ_m depolarization. A: representative FΔΨ_m signals during the dye stability test. B: spatial FΔΨ_m maps early in dye loading (T₁) as well as at the beginning (T₂) and end (T₃) of stability testing. C: FΔΨ_m intensity (FL) at T₂ and T₃ (n = 4 mice). D: representative rabbit heart showing the spatial location of FΔΨ_m signals (right) during targeted depolarization with FCCP administration. E: FΔΨ_m maps at 30 s, 15 min, and 30 min after FCCP administration. F: summary (n = 5 rabbits) of epicardial FΔΨ_m. All data are expressed as means ± SE. AU, arbitrary units; RV, right ventricle; LV, left ventricle; LA, left atrium; LAD, left anterior descending coronary artery.

Low-Flow Ischemia in Rabbit Hearts

To investigate the relation between ΔΨ_m and electrophysiology, we applied low-flow ischemia in rabbit hearts (n = 4) and subsequently injected a bolus of FCCP at 45 min to collapse any remaining ΔΨ_m. Three representative FΔΨ_m signals (black, teal, and pink) and spatial maps are shown in Fig. 4, A and B. During ischemia, FΔΨ_m had three phases: 1) FΔΨ_m decreased for 2 min, 2) FΔΨ_m increased and stabilized for ∼10 min, 3) FΔΨ_m decreased for the remainder of ischemia. At 2, 12, and 45 min, FΔΨ_m was 93.0 ± 2.5%, 98.1 ± 2.4%, and 69.3 ± 2.4%, respectively (Fig. 4C). Plots of APD₇₀ (gray) and conduction velocity (CV; teal) over the same time course are shown in Fig. 4D and demonstrate the asynchrony between electrophysiological and FΔΨ_m changes. Spatial FΔΨ_m heterogeneity was found to be significantly greater (P < 0.05) in ischemia (Fig. 4E) compared with targeted depolarization with FCCP (Fig. 3), which emphasizes the large spatial variation in FΔΨ_m during ischemia.

Fig. 4.

Metabolic and electrophysiological changes during low-flow ischemia in explanted rabbit hearts (n = 4). A: three FΔΨ_m signals. B: sequential FΔΨ_m maps throughout the protocol. The black, teal, and pink circles illustrate the spatial location of FΔΨ_m signals in A. C: summary (n = 4) of the changes in FΔΨ_m during low-flow ischemia. D: summary (n = 4) of the changes in action potential duration at 70% repolarization (APD₇₀; gray) and conduction velocity (CV; teal). E: spatial FΔΨ_m heterogeneity during low-flow ischemia versus targeted ΔΨ_m depolarization (Fig. 3). All data are expressed as means ± SE. *P < 0.05.

The spatial correlation between FΔΨ_m and changes in electrophysiological parameters was determined. Figure 5, A and B, shows a representative correlation calculation at 30 min of low-flow ischemia. Electrophysiological maps are shown as changes from baseline (ΔAPD₇₀ and ΔCV), where large APD₇₀ shortening and conduction slowing are shown as red. Blue FΔΨ_m values identify areas with depolarized mitochondria. When FΔΨ_m and ΔAPD₇₀ were plotted together (Fig. 5B, top), there was a strong inverse relation; however, when FΔΨ_m and ΔCV were plotted together (Fig. 5B, bottom), no relation was present. The summary of r values (n = 4; Fig. 5C) showed that FΔΨ_m-ΔAPD₇₀ had a progressively stronger inverse relation, reaching an r value of −0.58 ± 0.15 at 45 min. Spatial changes in CV were not associated with changes in FΔΨ_m, and, at 45 min of ischemia, FΔΨ_m-ΔCV had an r value of 0.04 ± 0.09.

Fig. 5.

Spatial correlation between changes in FΔΨ_m and optical action potential parameters during low-flow ischemia. A: representative ADP₇₀, FΔΨ_m, and CV maps at 30 min, where electrophysiological parameters are shown as changes from baseline (ΔAPD₇₀ and ΔCV). B: pixel-by-pixel plot of FΔΨ_m against ΔAPD₇₀ (top) and FΔΨ_m against ΔCV (bottom). C: summary (n = 4 rabbits) of r values for FΔΨ_m-ΔAPD₇₀ (top) and FΔΨ_m-ΔCV (bottom) at 2, 5, 10, 20, 30, and 45 min. All data are expressed as means ± SE.

Low-Flow Ischemia in Human Wedge Preparations

To determine the clinical relevance of our observations in mice and rabbits, we investigated the relation between ΔΨ_m and electrophysiology in human ventricular preparations. Figure 6A shows a representative LV wedge preparation with the optical field of view (black square) and the location of optical maps (white outline) highlighted. A transmural gradient in APD₇₀ was seen in five of seven preparations at baseline and was maintained in four of five preparations throughout low-flow ischemia (Fig. 6B). Subepicardial, midmyocardial, and subendocardial APD₇₀ changes during ischemia are shown for all preparations (7 of 7 prepartations) in Fig. 6C.

Fig. 6.

Metabolic and electrophysiological changes during low-flow ischemia in human LV wedge preparations. A: representative LV wedge showing the spatial location of the pacing electrode (*), optical maps (white outline), and the field of view (black square) during the optical mapping experiment. B: representative ADP₇₀ maps at baseline and 45 min of ischemia. C: summary (n = 7) of APD₇₀ changes during ischemia. D: FΔΨ_m maps at baseline and at 5 and 45 min into ischemia. E: summary (n = 3) of subendocardial (Endo), midmyocardial (Mid), and subepicardial (Epi) FΔΨ_m after the initiation of ischemia. All data are expressed as means ± SE. *P < 0.05, Endo vs. Epi; **P < 0.05, Endo vs. Mid.

The change in FΔΨ_m was also quantified for the subepicardium, midmyocardium, and subendocardium during low-flow ischemia (n = 3; Fig. 6, D and E). Upon the initiation of low-flow ischemia, all three regions decreased FΔΨ_m, with the subepicardium, midmyocardium, and subendocardium reaching minimums of 80.5% (at 4 min), 81.3% (at 4 min), and 92.7% (at 3 min), respectively. The midmyocardium and subendocardium exhibited an increase in FΔΨ_m after reaching their minimum values. Subepicardium FΔΨ_m was statistically different (P < 0.05) compared with the subendocardium at 10–35 and 45 min, whereas subendocardium FΔΨ_m was statistically different (P < 0.05) compared with the midmyocardium at 5 min. Thus, as in rabbits, the initial ΔΨ_m depolarization is reversible, but the extent varies transmurally.

Sustained Ventricular Arrhythmias

After 45 min of FCCP perfusion (protocol 2), 0% of rabbit hearts (0 of 5 hearts) had a sustained arrhythmia, whereas after 45 min of low-flow ischemia (protocol 3), 25% of rabbit hearts (2 of 8 hearts) and 0% of wedge preparations (0 of 7 preparations) had a sustained arrhythmia (Fig. 7A). The combination of low-flow ischemia and bolus injection of FCCP resulted in multiple episodes of sustained arrhythmias in 100% rabbit hearts (6 of 6 hearts) and 100% human preparations (3 of 3 preparations).

Fig. 7.

Arrhythmias during metabolic insults. A: arrhythmia inducibility during FCCP administration, low-flow ischemia (Isc), and low-flow ischemia plus FCCP bolus. In B–D, symbols are as follows: white/black asterisks, location of the pacing electrode; black and pink circles, location of optical signals; blue asterisk, stimulus artifact. B: activation map, optical action potentials (black and pink), and far-field ECG (gray) at baseline (top) and 69.5 min into ischemia (bottom). FCCP was added at 45 min. C: activation map during focal activity, where the focal origin (pink) and late-activated region (black) are shown. D: FΔΨ_m map at 70 min. E: summary (n = 5) of FΔΨ_m calculated in the focal origin, late-activated region, and across the entire epicardial surface (total). All data are expressed as means ± SE.

Sustained arrhythmias in rabbit hearts included focal activity (n = 11 episodes, Supplement Movie SII), reentry (n = 2 episodes, Supplement Movie SIII), and multiwave reentry (n = 18 episodes, Supplement Movie SIV). Focal activity and reentry are uncommon for human wedge preparations. Reentry is rare because the size of the wedge in relation to the wavelength (APD × CV) usually prevents the occurrence of reentry. Supplementary Movie SV shows an example of a stable (likely reentrant) arrhythmia in a human LV wedge preparation. However, arrhythmias in other wedge preparations could not be categorized because of the limited optical field of view.

A representative example of substrate development for focal activity in a rabbit heart is shown in Fig. 7, B–D. OAPs (black and pink) and far-field ECGs (gray) are shown at baseline and 69.5 min after the initiation of ischemia. The activation maps, OAPs, and far-field ECG emphasize the predictable response to ischemia, i.e., conduction slowing and repolarization shortening. Figure 7, C and D, shows the activation sequence initiated by 5.86-Hz focal activity at 70 min and the corresponding FΔΨ_m map. The optical signals have reversed order, and the pacing stimulus no longer captures the tissue. The location of a focal origin was in the optical field of view in four of four rabbit hearts with pulmonary artery cannulation. In a single experiment, there were two focal origin sites for different sustained arrhythmias. FΔΨ_m was quantified (n = 5; Fig. 7E) across the entire epicardium (54.9 ± 8.7%), in the focal origin (49.4 ± 10.6%), and in the late-activating region (54.9 ± 4.6%) and was not significantly different (P = 0.87).

DISCUSSION

Here, we present, for the first time, a dual-imaging methodology for optical mapping of sarcolemmal and mitochondrial membrane potentials. Our data show that 1) ΔΨ_m depolarization does not propagate as a unidirectional wave, 2) ΔΨ_m depolarization is sporadic and reversible during ischemia, 3) ΔΨ_m depolarization is spatially associated with APD shortening late in ischemia but not associated to conduction slowing, and 4) the origin of focal arrhythmias is not related to ΔΨ_m depolarization. Overall, mitochondrial potential and sarcolemmal electrophysiological properties are asynchronous during ischemia, and no connection between mitochondrial depolarization and arrhythmogenic substrates was found.

Comparison With Previous Reports

Sporadic depolarization during ischemia contradicts previous wide-field optical mapping reports (21, 26) but is in good agreement with previous microscopy studies that imaged multiple myocytes (∼10 myocytes) per experiment. Microscopy studies (23, 28, 37, 38) on intact hearts showed individual myocytes having isolated FΔΨ_m loss within 10–60 min, whereas neighboring myocytes were unaffected and maintained FΔΨ_m. In addition to microscopy studies, Berkich and colleagues (4) used a lipophilic isotope to measure ΔΨ_m in rat hearts and demonstrated that significant depolarization started after 35 min. We observed FΔΨ_m loss on the majority of the epicardial surface after 12 min of low-flow ischemia and a 30.7% decrease at 45 min (Fig. 4).

Effect of V_m and ΔΨ_m on FΔΨ_m

Measurement of fluorescence relies on the Nernstian distribution of lipophilic cations to accumulate inside negatively charged compartments, especially the cytosol and mitochondria. During the progression of ischemia, changes in FΔΨ_m are not only dependent on ΔΨ_m depolarization but also depolarization of V_m. V_m depolarizes ∼25 mV during ischemia, which will lead to FΔΨ_m loss that is not associated with ΔΨ_m depolarization. Mathematical models (31, 39, 40) have been developed to quantify FΔΨ_m changes within the mitochondrial matrix, cytosol, and extracellular space and can determine the influence of V_m depolarization on FΔΨ_m. A 25-mV depolarization of V_m will cause 4.3% FΔΨ_m loss when ΔΨ_m is held at −180 mV. Thus, V_m depolarization will have a minimal effect on FΔΨ_m, especially early in the protocol, when V_m is just beginning to depolarize.

Transmural FΔΨ_m

Here, we report the first transmural recordings of FΔΨ_m. The distinct transmural FΔΨ_m response to low-flow ischemia could be due to functionally different mitochondria or substrate availability (9, 29). Overall, both morphometric and functional studies (27, 36) on mitochondria have shown few differences transmurally; however, the quantity of glycogen and activity of glycogen phosphorylase have been found to be significantly greater in the subendocardium compared with the subepicardium (19, 20). Ichihara and Abiko (19) showed that ligation of a coronary artery leads to subepicardial and subendocardial decreases in glycogen content, but the subendocardium does so more rapidly and markedly. Excess substrate consumption may allow for the subendocardium to maintain FΔΨ_m and have a greater ability to reverse FΔΨ_m loss.

FΔΨ_m and Arrhythmias

Monitoring ΔΨ_m in conjunction with V_m is advantageous for understanding the role of metabolism in the genesis of arrhythmias. During the ischemia protocol, we observed both focal activity and reentrant arrhythmias. It has been theorized that spatiotemporal heterogeneities in mitochondrial function may be the underlying mechanism for the genesis of ischemia-induced arrhythmias in that they promote arrhythmogenic substrates, including heterogeneous AP repolarization and depressed excitability (1).

During ischemia, we observed large spatial heterogeneities in mitochondrial function within hearts (Fig. 4E) and determined the spatial association between ΔΨ_m and electrophysiological parameters. Our data indicate that early in ischemia, ΔΨ_m polarization is not associated with APD shortening but has an increasing inverse association throughout the protocol. The early dissociation reflects mitochondria reversing and maintaining their potential while APs begin to shorten. Mitochondrial ATP synthase can compete for cytosolic ATP to maintain ΔΨ_m and delay depolarization. The K_ATP channel is very sensitive to the metabolic state of the cytosol, and opening of <1% of channels can result in significant AP shortening (34). It is important to note that the end of repolarization is not equivalent to the refractory period in ischemic tissue, which has postrepolarization refractoriness (11). Thus, heterogeneous AP repolarization during ischemia is an unreliable predictor of reentrant arrhythmias. Slowing of conduction greatly increases the likelihood of the occurrence of arrhythmias, but our data show that ΔΨ_m depolarization is not spatially associated with conduction slowing.

Simultaneous monitoring of ΔΨ_m and V_m allowed for detailed analysis of focal arrhythmias and showed that the focal origin does not have different mitochondrial function compared with the rest of the myocardium. ΔΨ_m depolarization does not seem to have a direct association with ischemia-induced arrhythmias; however, the metabolic state of the tissue likely plays a role in arrhythmogenesis. Injury current has also been proposed as the source of ectopic activity during ischemia and is the electrotonic flow of current between regions of tissue with large variation in membrane potential (10, 12). Accumulation of regional extracellular K⁺ is a major cause for differences in membrane potential, and thus K_ATP channel sensitivity to the cytosolic nucleotide concentration is an important element in producing an injury current.

Limitations

There are several limitations to this study. First, tissue preparations were perfused with the excitation-contraction uncoupler blebbistatin, consequently leaving hearts in an energetically favorable condition. The metabolic demand of different isolated heart models has been previously reported. Using NADH imaging, Wengrowski and colleagues (41) elegantly showed the increase delay in metabolic changes when using biventricular working, Langendorff-perfused, and mechanically arrested models during ischemia. In addition, we used a model of low-flow ischemia instead of no-flow ischemia or regional ischemia by coronary ligation. Venable and colleagues (38) have provided a well-founded argument against the use of wide-field optical mapping of ΔΨ_m during no-flow ischemia. The pool of fluorophores remains constant with no-flow ischemia and thus requires the imaging technique to resolve dye redistribution between cellular compartments. Moreover, explanted hearts were denervated, leaving sympathetic/parasympathetic tone absent, and differences in cardiac cell populations were not considered. Finally, analysis of focal activity was performed on the epicardial surface, whereas the source could be deeper within the myocardium.

Conclusions

Ischemia involves a complex interplay between metabolic and electrophysiological systems. Mitochondrial depolarization is sporadic, reversible, and asynchronous with sarcolemmal electrophysiological changes. Mitochondrial depolarization is not directly associated with ischemia-induced arrhythmias.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants R21-HL-112278, R01-HL-114395, and R01-HL-115415 (to I. R. Efimov). M. S. Sulkin was supported by an American Heart Association predoctoral fellowship. F. S. Ng was funded by British Heart Foundation Travel Fellowship Grant FS/11/69/29017 and a Clinical Lectureship from the UK National Institute for Health Research.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: M.S.S., B.J.B., F.S.N., and I.R.E. conception and design of research; M.S.S., B.J.B., M.T., and S.R.G. performed experiments; M.S.S. and S.R.G. analyzed data; M.S.S., F.S.N., and I.R.E. interpreted results of experiments; M.S.S. and I.R.E. prepared figures; M.S.S. and I.R.E. drafted manuscript; M.S.S., B.J.B., F.S.N., and I.R.E. edited and revised manuscript; M.S.S., B.J.B., M.T., S.R.G., F.S.N., and I.R.E. approved final version of manuscript.