Uniform action potential repolarization within the sarcolemma of in situ ventricular cardiomyocytes

Abstract

Previous studies have speculated, based on indirect evidence, that the action potential at the transverse (t)-tubules is longer than at the surface membrane in mammalian ventricular cardiomyocytes. To date, no technique has enabled recording of electrical activity selectively at the t-tubules to directly examine this hypothesis. We used confocal line-scan imaging in conjunction with the fast response voltage-sensitive dyes ANNINE-6 and ANNINE-6plus to resolve action potential-related changes in fractional dye fluorescence (DeltaF/F) at the t-tubule and surface membranes of in situ mouse ventricular cardiomyocytes. Peak DeltaF/F during action potential phase 0 depolarization averaged -21% for both dyes. The shape and time course of optical action potentials measured with the water-soluble ANNINE-6plus were indistinguishable from those of action potentials recorded with intracellular microelectrodes in the absence of the dye. In contrast, optical action potentials measured with the water-insoluble ANNINE-6 were significantly prolonged compared to the electrical recordings obtained from dye-free hearts, suggesting electrophysiological effects of ANNINE-6 and/or its solvents. With either dye, the kinetics of action potential-dependent changes in DeltaF/F during repolarization were found to be similar at the t-tubular and surface membranes. This study provides what to our knowledge are the first direct measurements of t-tubule electrical activity in ventricular cardiomyocytes, which support the concept that action potential duration is uniform throughout the sarcolemma of individual cells.

Figure 1

Confocal imaging of the cardiomyocyte action potential in Langendorff-perfused mouse hearts in the presence of 50 μM cytochalasin D and 1 μM ryanodine. (A) Full-frame mode (XY) image obtained from a heart loaded with ANNINE-6. The green bar demarks the position of line-scan mode data acquisition along the end-to-end membrane junction between two neighboring cardiomyocytes. The heart was electrically paced at 3 Hz via right epicardial point stimulation. The image was acquired 150 min after dye loading. The middle panel shows a magnified view of the blue boxed region in the full frame-mode image. Sharp lines (black arrows in the middle panel) correspond to action potential-related decreases in ANNINE-6 fluorescence. The right panel shows magnified views of ANNINE-6 fluorescence distribution around cardiomyocyte nuclei. (B) Line-scan (X-t) image of the region in A demarked by the green line. The image is composed of 9000 horizontally stacked line-scans. Each scan line encompassed 128 pixels along a 16.84-μm line integrating for 3.2 μs at each pixel. Line-scan repetition rate was 1024 Hz. Lower trace shows the volume conducted electrocardiogram that was recorded during image acquisition. Asterisks indicate electrical stimulation artifacts. (C) Expanded record of changes in ANNINE-6 fluorescence during the cardiomyocyte action potential. The left panel shows the unprocessed line-scan image of fluorescence changes during the action potential boxed in B. Normalization of the fluorescence signal (F/F₀) removed the nonuniformity in the signal (right panel). Scale bar = 50 ms vertically, 5 μm horizontally. (D) Time course of spatial averaged ANNINE-6 fluorescence from the normalized line-scan image in C.

Figure 2

Scheme for off-line signal processing. Graphic representation of the off-line signal processing steps that were used to correct for dye bleaching-induced baseline drift, to calculate relative fluorescence changes [ΔF/F] and to improve the signal/noise ratio. (A) Line-scan image of action potential-related changes in ANNINE-6 fluorescence that was obtained from an end-to-end junctional membrane of two juxtaposed cardiomyocytes in situ during pacing at 3 Hz. (B) By averaging 128 pixels in each line of the line-scan image in the vertical direction, the time course of the spatial averaged fluorescence [F(t)] was determined. To obtain the time course of normalized fractional changes in dye fluorescence [ΔF/F₀(t)] the baseline [F₀(t)] was estimated (denoted by the red solid line) using a combination of a 250-sample moving averaging filter and a 4th-order Butterworth low-pass filter with a cut-off frequency of 3 Hz. (C) Rendition of ΔF = F(t) − F₀(t) normalized to F₀(t). (D) To increase the SNR of single-trial recordings, plots of ΔF/F₀(t) were processed with a custom-designed optimal filtering algorithm. A filtered 2-s segment from the trace in C is shown (red line). See Appendix A for more detailed description of the optimal filtering algorithm. (E) Recordings of normalized changes in fractional fluorescence occurring during each consecutive action potential in a line-scan image were aligned using both cross-correlation analysis and pacing cycle, and the ensemble average was calculated (red symbols). A single action potential that was obtained by processing the trace in C with the optimal filtering algorithm was superimposed (blue symbols).

Figure 3

Confocal line-scan imaging of voltage changes across the outer sarcolemma and t-tubule membrane of in situ cardiomyocytes. (A) Frame-mode and line-scan confocal images obtained from Langendorff-perfused mouse hearts loaded with ANNINE-6 (left) or ANNINE-6plus (right). Images were acquired during electrical point stimulation at 3 Hz in the presence of 50 μM cytochalasin D and 1 μM ryanodine to dissociate contraction from excitation. White lines denote positions of the line-scan mode data acquisitions along end-to-end (a and d) and side-to-side (b and e) junctions between juxtaposed cardiomyocytes, and across multiple t-tubules (c and f) within single cardiomyocytes. Line-scan images (A, a–f) show 1-s segments of each of the 8.64-s long line-scan image acquisitions. Small letters next to the line-scan images refer to the scan regions delineated in the frame-mode images. Green scale bars = 8 μm. (B) Superimposed time plots of normalized fractional changes in dye fluorescence [ΔF/F₀(t)] for the three recording sites per dye as indicated. Tracings were digitally filtered as outlined in the Appendix. (C) Superimposed tracings of the ensemble average of all 25 action potential-induced fluorescence transients per line-scan image for each of the three recording sites. (D) Bar graphs comparing properties of action potential-induced fluorescence changes in t-tubule and surface membranes in the presence of 50 μM cytochalasin D and 1 μM ryanodine. (Left panel) Peak ΔF/F₀ during action potential phase 0 depolarization. (Right panel) Action potential durations at various time points of repolarization. Values are mean ± 1 SD from 84 and 45 line-scans obtained in eight ANNINE-6- and four ANNINE-6plus-loaded mouse hearts, respectively. E, end-to-end membrane junction; S, side-to-side membrane junction; T, t-tubular membranes. ^∗p < 0.025 versus peak ΔF/F₀ at side-to-side membrane junctions and t-tubular membranes of the same dye (paired t-test). †p < 0.05 versus ANNINE-6 (t-test). Peak ΔF/F₀ and action potential durations were measured from the ensemble average traces.

Figure 4

Comparison of the optical action potential at a single t-tubule and at the surface membrane. (A) Frame-mode image taken from a heart loaded with ANNINE-6plus. The blue and white line denote positions of the line-scan mode data acquisitions at a side-to-side junction and along a single t-tubule, respectively. Middle and right panel show zoom-in views of the regions around the labeled side-to-side and t-tubule membrane positions. The thick red line in the middle panel represents ANNINE-6plus fluorescence originating from outer sarcolemmal membranes of two adjacent cardiomyocytes, whereas the red lines perpendicular to the thick line are t-tubules of either myocyte. (B) Line-scan images that were obtained by repetitive scanning at a rate of 1,024 lines/s along the white lines in A. (C) Overlay of filtered ΔF/F₀(t) traces for the two scan positions demarked in A. (D) Expanded record of changes in ANNINE-6plus fluorescence during the action potential. The left panel shows the normalized (F/F₀) line-scan image during the action potential boxed in B. Image in the right panel was low-pass filtered to give a spatial resolution of 0.2 μm, and the contrast enhanced. Scale bars = 50 ms vertically, 2 μm horizontally. (E) Kinetic comparison of the normalized fluorescence signal from the three regions (colored bars) indicated in C. Each region is 2 μm wide. Note the similarity in the time course of the action potential-related fluorescence changes from the three regions.

Figure 5

Kinetic comparison of the optical and electrical signals. (A) Transmembrane action potentials recorded with an intracellular microelectrode from the anterior left ventricular epicardium of a Langendorff-perfused mouse heart during electrical point stimulation at 3 Hz. Arrows indicate stimulation artifacts. The recording was obtained in the presence of 50 μM cytochalasin D and 1 μM ryanodine. The signal was low-pass filtered at 10 kHz and sampled at 20 kHz. (B) Plot of action potential-induced fractional fluorescence changes [ΔF/F₀(t)]. The fluorescence signal was resolved from a confocal line-scan image that was obtained by repeatedly scanning along an end-to-end junction at a rate of 1042 Hz. The fluorescence signal was processed as described in the Appendix A. (C) Comparison of action potential duration in optical and electrical recordings. Intracellular microelectrode recordings were carried out in hearts loaded with ANNINE-6plus. Values are mean ± 1 SD from 45 optical and 124 electrical measurements that were obtained in four and five hearts, respectively, in the presence of 50 μM cytochalasin D and 1 μM ryanodine. Optical recordings from all three outer membrane compartments were included. No statistically significant differences were found (t-test; p > 0.05). (D) Overlay of the initial portion of an optical and electrical action potential at expanded timescale. (E) Power spectral density distributions of the electrical and optical signal in A and B, respectively. Spectra were generated over frequencies ranging from 0 to the Nyquist frequency of the respective recording method (521 Hz for the optical and 10 kHz for the microelectrode technique). The spike at 2015 Hz represents the phase 0 depolarization of the transmembrane action potentials. Inset shows the power spectrum over the frequency range from 2010 to 2020 Hz.

Figure 6

Properties of ANNINE-6plus. (A) Voltage dependence of ANNINE-6plus fluorescence. ANNINE-6plus fluorescence changes (expressed in [− ΔF/F₀]) were plotted as a function of the magnitude of the transmembrane voltage change [ΔV] relative to the resting membrane potential. Data points were derived from the ensemble averages of the tracings shown in A and B of Fig. 5. Only data points during the action potential repolarization were included in the analysis. The line is a linear regression of the data (slope: 0.0217; y axis intercept: 0.0092; R = 0.994). (B) Effect of changes in coupling interval on optical action potential duration. Plot of action potential-induced fluorescence changes [F(t)] that were recorded during spontaneous, irregular electrical activity of a Langendorff-perfused mouse heart loaded with ANNINE-6plus. The fluorescence signal was resolved from a confocal line-scan image that was obtained by repeatedly scanning along an end-to-end junction at a rate of 1042 Hz. The recording was obtained in the presence of 50 μM cytochalasin D and 1 μM ryanodine.

Figure 7

Comparison between the time course of dye bleaching and the time course of the peak amplitude of the fractional fluorescence change [ΔF/F] during confocal line-scan imaging. (A) Time course of total fluorescence [F(t)] and fractional fluorescence change [ΔF/F₀(t)] for the first and last three action potentials of 8.64-s long line-scan recordings obtained with ANNINE-6 (left panels) and ANNINE-6plus (right panels). (B) Decay of F(t) normalized to the zero time point and decrease in peak ΔF/F₀(t) for ANNINE-6 and ANNINE-6plus. Values are mean ± 1 SD from 20 independent line-scans for each dye. All line-scan images were acquired in the presence of 50 μM cytochalasin D and 1 μM ryanodine.