Disturbed Repolarization-Relaxation Coupling During Acute Myocardial Ischemia Permits Systolic Mechano-Arrhythmogenesis

Abstract

Background: The heart's mechanical state feeds back to its electrical activity, potentially contributing to arrhythmias. Mechano-arrhythmogenesis has been mechanistically explained during electrical diastole, when cardiomyocytes are at their resting membrane potential. During electrical systole, cardiomyocytes are refractory right after the onset of depolarization, while during repolarization in physiological conditions, they seem to be protected from systolic mechano-arrhythmogenesis by near-simultaneous restoration of resting membrane potential and cytosolic calcium concentration ([Ca²⁺]_i): repolarization-relaxation coupling (RRC). Yet, late-systolic mechano-arrhythmogenesis has been reported in ischemic myocardium, with unclear underlying mechanisms. We hypothesize that ischemia-induced alteration of RRC gives rise to a vulnerable period for mechano-arrhythmogenesis.

Methods: Acute left ventricular regional ischemia was induced by coronary artery ligation in Langendorff-perfused rabbit hearts, with mechanical load controlled by an intraventricular balloon. Mechanical activity was assessed by echocardiography and arrhythmia incidence by ECG. Single left ventricular cardiomyocytes were exposed to simulated ischemia or pinacidil (ATP-sensitive potassium channel opener). Stretch was applied in diastole or late systole using carbon fibers. Stretch characteristics and arrhythmia incidence were assessed by sarcomere length measurement. In both models, RRC was assessed by simultaneous voltage-[Ca²⁺]_i fluorescence imaging and mechano-arrhythmogenesis mechanisms were pharmacologically tested.

Results: In whole hearts, acute regional ischemia leads to systolic stretch and disturbed RRC at the ischemic border. These electro-mechanical changes were associated with waves of arrhythmias, which could be reduced by mechanical unloading, electro-mechanical uncoupling, or buffering of [Ca²⁺]_i. In left ventricular cardiomyocytes, physiological RRC is associated with a low incidence of systolic mechano-arrhythmogenesis, while a vulnerable period emerged by prolonged RRC during ischemia. The increase in systolic mechano-arrhythmogenesis was reduced by restoring RRC, chelating [Ca²⁺]_i, blocking mechano-sensitive TRPA1 (transient receptor potential ankyrin 1) channels, or buffering reactive oxygen species levels.

Conclusions: Prolonged RRC allows for late-systolic mechano-arrhythmogenesis in acute ischemia, involving contributions of elevated [Ca²⁺]_i, TRPA1 activity, and reactive oxygen species, which represent potential antiarrhythmic targets.

Keywords: arrhythmia; calcium; feedback; ion channels; rabbits; reactive oxygen species; systole.

Figure 1.

Altered left ventricular (LV) mechanical activity during acute regional ischemia in rabbit whole hearts. A, Schematic illustration of the experimental setup for the whole-heart regional ischemia model. B, Representative 2-dimensional echocardiographic image of the LV, showing the normally perfused remote tissue, the ischemic border, and the central ischemic tissue regions. C, Example of measured radial strain over one heartbeat in each of the 3 regions before (left) and 30 minutes after (right) ligation of the anterior branch of the left circumflex coronary artery positioned at one third of the distance between the LV base and apex. D, Summary data on maximum (wall thickening, left) and minimum (wall thinning, right) radial strains before (0 minutes) and after (30 minutes) coronary artery ligation, averaged over 3 beats per heart, demonstrating an ischemia-induced decrease in contractile function in ischemic and border tissue, and stretch of ischemic border tissue. Differences between pre- and 30 minutes post-ligation were assessed by paired, 2-tailed, Student t tests.

Figure 2.

Arrhythmia incidence during acute regional ischemia in rabbit whole hearts. A, Representative ECG and left ventricular (LV) pressure recordings 30 minutes after coronary artery ligation, showing examples of ectopic excitation (left) and self-terminating arrhythmic activity (right). Incidence of ectopic excitation (blue, left y axis) and self-terminating arrhythmic activity (red, right y axis) over 60 minutes (binned into 5-minute periods) after artery ligation. Experimental groups include hearts with: (B) contracting, physiologically loaded LV; (C) contracting, mechanically unloaded LV; and (D) noncontracting, physiologically loaded LV. In addition, results are shown for hearts with a contracting, physiologically loaded LV treated with: (E) BAPTA-AM (1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid acetoxymethyl ester, to buffer cytosolic calcium); or (F) dantrolene to stabilize ryanodine receptors in their closed state. G, Average arrhythmia incidence in the 20 to 45 minute window after artery ligation (phase 1b). Whether the LV of each group was physiologically loaded and/or contracting is indicated by the tick (✓) marks below the x axis labels. Differences between interventions and loaded hearts were assessed by 1-way ANOVA with Dunnett post hoc tests.

Figure 3.

Disturbed repolarization-relaxation coupling (RRC) during acute regional ischemia in rabbit left ventricle. A, Schematic illustration of the dual-excitation (using camera frame-synchronized light-emitting diodes, LEDs), dual-emission fluorescence imaging approach projecting the 2 emission wavelengths (via a 585/50 nm and 800/200 nm band-pass filter) onto a single scientific complementary metal-oxide-semiconductor (sCMOS) camera for near-simultaneous measurement of voltage and cytosolic calcium ([Ca²⁺]_i) in rabbit whole hearts (left), and representative action potentials (AP, blue) and Ca²⁺ transients (CaT, red) along a line across the ischemic border 30 minutes after coronary artery ligation (right). B, Representative AP and CaT traces (4 mm and −4 mm in A), showing AP duration at 50% repolarization (APD₅₀, blue dashed line) and CaT duration at 80% return to diastolic levels (CaTD₈₀, red dashed line), as well as the resulting RRC duration in the perfused (left, gray) and ischemic tissue (right, green), demonstrating an ischemia-induced prolongation of RRC duration. C, Maps of APD₅₀ (left), CaTD₈₀ (middle), and RRC duration (right) across the left ventricle before artery ligation and 30 minutes after the onset of ischemia. D, Measurements of APD₅₀ (left), CaTD₈₀ (middle), and RRC duration (right) over 45 minutes from artery ligation at points across the ischemic border, with the dashed box indicating values that differed significantly from preligation values. Differences compared with preligation were assessed by 1-way ANOVA, with Dunnett post hoc tests.

Figure 4.

Arrhythmias elicited by transient stretch of rabbit left ventricular (LV) cardiomyocytes. A, Brightfield image of a rabbit LV cardiomyocyte before (top) and during (bottom) various levels of axial stretch, applied using a carbon fiber-based system. Scale bar, 10 µm. B, Schematic of the stretch protocol, including timing information for stretch during RRC. C, Representative measurement of sarcomere length in a cardiomyocyte during simulated ischemia with 1 Hz pacing (orange dots) and stretch (green segments) applied in diastole (first and third stretch) or in late systole during repolarization-relaxation coupling (RRC, second and fourth stretch), none of which resulted in an arrhythmia (normal paced rhythm, NPR). D, Stretch-induced contraction (blue curve segment) on diastolic stretch during simulated ischemia. E, Arrhythmic activity (red segment) that spontaneously resolved, initiated by a stretch during RRC in simulated ischemia conditions. F, Sustained arrhythmic activity (red segment) after a diastolic stretch during pinacidil exposure, which was terminated by application of an additional stretch (rescue stretch; same cells as in Video S4).

Figure 5.

Disturbed repolarization-relaxation coupling (RRC) in rabbit left ventricular (LV) cardiomyocytes. A, Schematic of the single-excitation, dual-emission fluorescence imaging approach, utilizing a single camera with an image splitter for simultaneous measurement of voltage and intracellular calcium ([Ca²⁺]_i) in rabbit LV cardiomyocytes. B, Representative action potential (AP, blue) and Ca²⁺ transients (CaT, red) with indication of the duration of RRC (=excitation-contraction coupling [ECC] duration+CaT duration at 80% return to diastolic levels [CaTD₈₀]-AP duration at 50% repolarization [APD₅₀]) in control (CTRL, top left), during simulated ischemia (SI; top right) or pinacidil (PIN; bottom left) application—both of which lead to an increase in RRC duration (green shaded area)—and on mitigation of the SI-induced RRC lengthening by glibenclamide (GLIB, bottom right). C, Measurements of APD₅₀ (blue) and CaTD₈₀ (red) in CTRL, SI, or SI following pretreatment with GLIB (GLIB+SI). D, Measurements of APD₅₀ and CaTD₈₀ in CTRL conditions, or after exposure to PIN or GLIB+PIN. E and F, RRC duration in each experimental condition. Differences between conditions were assessed by 1-way ANOVA, with Tukey post hoc tests, and between APD₅₀ and CaTD₈₀ or SI and pinacidil by paired, 2-tailed, Student t tests. ^####P=1.3×10⁻⁴ for SI vs PIN.

Figure 6.

Role of disturbed repolarization-relaxation coupling (RRC) in mechano-arrhythmogenesis. Incidence of stretch-induced contractions (blue) and self-sustained arrhythmias (red) on stretch of rabbit left ventricular (LV) cardiomyocytes, applied during diastole or RRC, in control (CTRL) conditions, or during exposure to simulated ischemia (SI), glibenclamide (GLIB)+SI, pinacidil (PIN), or GLIB+PIN. Arrhythmia incidence was calculated as the number of stretch-induced contractions or self-sustained arrhythmias, divided by the number of stretch stimuli applied (m) per condition. Differences in arrhythmia incidence were assessed using χ² contingency tables and Fisher exact test.

Figure 7.

Contribution of cytosolic calcium concentration ([Ca²⁺]_i) to mechano-arrhythmogenesis. A, Incidence of stretch-induced contractions (blue) and self-sustained arrhythmias (red) upon stretch of rabbit left ventricular (LV) cardiomyocytes, applied either during diastole or during repolarization-relaxation coupling (RRC), during exposure to simulated ischemia (SI) or pinacidil (PIN), as well as in combination with BAPTA-AM (1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid acetoxymethyl ester, to buffer [Ca²⁺]_i) or dantrolene (to stabilize ryanodine receptors in their closed state). Arrhythmia incidence was calculated as the number of stretch-induced contractions or self-sustained arrhythmias, divided by the number of stretch stimuli applied (m) per condition. Dashed gray lines show arrhythmia incidence for stretch during diastole or RRC in control (CTRL) conditions (see first pair of bars in Figure 6). B, Representative voltage (blue) and Ca²⁺ (red) signals simultaneously recorded by fluorescence imaging in a cardiomyocyte exposed to PIN, showing a sustained arrhythmia, induced by stretch (purple) during the period of prolonged RRC, leading initially to depolarization and an associated increase in [Ca²⁺]_i above normal systolic levels, followed by spontaneous oscillations in [Ca²⁺]_i (which subsequently resolved, followed by normal paced beats; not shown). Blue inset, Scaled voltage and Ca²⁺ signals (top) and phase plot of the first beat in the inset (bottom), showing that before stretch, changes in voltage precede changes in [Ca²⁺]_i. Red inset, Scaled voltage and Ca²⁺ signals (top) and phase plot of the first beat in the inset (bottom), showing [Ca²⁺]_i oscillations preceding changes in voltage during the stretch-induced sustained arrhythmia, suggesting aberrant Ca²⁺ handling as the driving mechanism. C, Fraction of stretch stimuli applied for which stretch and/or release occurred during RRC in a subset of cells exposed to SI. Stimuli that resulted in arrhythmias are shown in red (all of which were self-sustained arrhythmias). Differences in arrhythmia incidence were assessed using χ² contingency tables and Fisher exact test.

Figure 8.

Contribution of TRPA1 (transient receptor potential ankyrin 1) channels to mechano-arrhythmogenesis. A, Representative current recording by cell-attached patch clamp (holding potential, E=+40 mV) of rabbit left ventricular (LV) cardiomyocytes in control (CTRL) conditions and after switching to AITC (allyl isothiocyanate)-containing solution (50 μmol/L; to activate TRPA1 channels). Inset, Detail of channel activation with visible single-channel events. B, Quantification of AITC-induced current changes (left; 3 minutes total application including a 15 second lag time for AITC to reach the cardiomyocyte) and time-matched, same-batch cardiomyocytes in CTRL conditions (right). C, Incidence of stretch-induced contractions (blue) and self-sustained arrhythmias (red) after stretch of cardiomyocytes during diastole in CTRL and during exposure to AITC, AITC+BAPTA-AM (1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid acetoxymethyl ester, to buffer cytosolic calcium concentration, [Ca²⁺]_i), or AITC+HC-030031 (to block TRPA1 channels)_. Arrhythmia incidence calculated as the number of stretch-induced contractions or self-sustained arrhythmias, divided by the number of stretch stimuli applied (m) per condition. D, Incidence of mechano-arrhythmogenesis upon stretch of cardiomyocytes during diastole or repolarization-relaxation coupling (RRC) during exposure to pinacidil (PIN), PIN+streptomycin (a nonspecific blocker of mechano-sensitive ion channels), or PIN+HC-030031. E, Incidence of mechano-arrhythmogenesis during exposure to simulated ischemia (SI), or to SI with additional exposure to HC-030031, NAC (N-acetyl-l-cysteine, to chelate reactive oxygen species, ROS), or DPI (diphenyleneiodonium chloride, to block ROS production). Dashed gray lines in D and E show arrhythmia incidence for stretch during diastole or RRC in CTRL conditions (see first pair of bars in Figure 6). Current differences were assessed with 2-tailed, paired, Student t test. Differences in arrhythmia incidence were assessed using χ² contingency tables and Fisher exact test.