Action potential propagation in transverse-axial tubular system is impaired in heart failure

Abstract

The plasma membrane of cardiac myocytes presents complex invaginations known as the transverse-axial tubular system (TATS). Despite TATS's crucial role in excitation-contraction coupling and morphological alterations found in pathological settings, TATS's electrical activity has never been directly investigated in remodeled tubular networks. Here we develop an ultrafast random access multiphoton microscope that, in combination with a customly synthesized voltage-sensitive dye, is used to simultaneously measure action potentials (APs) at multiple sites within the sarcolemma with submillisecond temporal and submicrometer spatial resolution in real time. We find that the tight electrical coupling between different sarcolemmal domains is guaranteed only within an intact tubular system. In fact, acute detachment by osmotic shock of most tubules from the surface sarcolemma prevents AP propagation not only in the disconnected tubules, but also in some of the tubules that remain connected with the surface. This indicates that a structural disorganization of the tubular system worsens the electrical coupling between the TATS and the surface. The pathological implications of this finding are investigated in failing hearts. We find that AP propagation into the pathologically remodeled TATS frequently fails and may be followed by local spontaneous electrical activity. Our findings provide insight on the relationship between abnormal TATS and asynchronous calcium release, a major determinant of cardiac contractile dysfunction and arrhythmias.

The transverse-axial tubular system (TATS) is a complex network characterized by transverse (t tubules, TT) and longitudinal (axial tubules, AT) components running from one transverse tubule to the next (1–3). The TATS of a myocyte rapidly conducts depolarization of the surface sarcolemma (SS) to the core of the cardiomyocyte (4). The coupling between sarcolemmal Ca²⁺ entry during an action potential (AP) and Ca²⁺ release from sarcoplasmic reticulum (SR) promotes synchronous myofibril activation throughout the myocyte (5). Recent studies highlight that AP-relevant ion channels and transporters are expressed in TATS membrane with different densities and isoforms from those in SS (6, 7). These findings, in combination with diffusional limitations in TATS's lumen (3, 8), raise the possibility that AP may differ among membrane domains. Further interest in the TATS AP stems from the recent finding of loss and disorganization of tubules in several pathological conditions, including heart failure (9–14). Because correlation between morphological TATS alterations and Ca²⁺-release asynchronicity has been observed (14, 15), recording AP propagation in TATS can elucidate the fundamental electrophysiological mechanism linking structural and functional anomalies.

Although the uniformity of the AP across the whole sarcolemma has been mathematically (16) and experimentally (17) proved, a potential alteration of the AP propagation in a disorganized TATS has never been investigated. Here, we develop an optical method to simultaneously measure APs in different sarcolemma domains, allowing optical mapping of AP propagation within intact and pathologically remodeled TATS.

Results

We investigated the electrical properties of TATS in isolated rat ventricular myocytes labeled with di-4-AN(F)EPPTEA (Fig. S1), a new fluorinated voltage-sensitive dye with improved photostability. As shown in Fig. 1 A and B, the diffusion of the dye into the sarcolemma allows specific and homogeneous labeling of the TATS. Conventional electrophysiological measurements performed in labeled myocytes showed that the staining procedure does not influence the electrical properties of the cell (Fig. S2). A custom-made random access multiphoton (RAMP) (18–21) microscope was used to record membrane potential (Vm) simultaneously from multiple sites in real time. The RAMP acousto-optic deflectors rapidly scan linear segments of different membrane domains and perform multiplexed measurements of the two-photon fluorescence (TPF) signal with submillisecond temporal resolution. Fig. 1 A–D is an example of real-time and simultaneous optical recording of 10 elicited APs from three different membrane sites: SS (red), TT (green), and AT (blue). The experiments were performed at room temperature in the presence of an excitation contraction uncoupler (blebbistatin) and 1 mM extracellular [Ca²⁺]. APs were elicited at 2 Hz by field stimulation (black arrowheads in Fig. 1 C and D). Repeated measurements (903 APs from 21 cells from eight rats) showed that AP amplitude (Fig. 1E) and time course (Fig. 1F) measured in TATS are not statistically different from those of SS. Measurements at a high stimulation frequency were also performed to test possible effects of local AP alterations due to cumulative changes in TATS luminal ion concentration (e.g., Ca²⁺ depletion and K⁺ accumulation). The uniformity of AP in all sarcolemma domains is maintained even at high pacing rates (Fig. S3), confirming the expected tight electrical coupling between membrane domains (16, 17). The tight electrical coupling also holds in the subthreshold voltage range during spontaneous depolarization events (Fig. S4).

Fig. 1.

Electrical properties of TATS. (A) TPF image of a stained rat ventricular myocyte. (Scale bar: 10 μm.) (B) The region in the yellow box of A is shown magnified 10×. The lines mark the probed sarcolemma regions: surface sarcolemma (SS) in red, t-tubule (TT) in green, and axial tubule (AT) in blue. (C) Normalized fluorescence traces from the scanned line indicated in B. APs were elicited at 2 Hz, corresponding to the black arrowheads. (D) Average of 10 sequential APs shown in C. (E) Relative fluorescence change (ΔF/F) at the peak of APs from the three different sarcolemmal regions. (F) AP time course: Peak indicates the time from stimulus end to AP peak; 50 and 90% indicate the time from AP peak to 50 and 90% repolarization, respectively. Each bar represents the mean ± SE.

Multisite optical recordings were also performed in ventricular trabeculae in which cell-to-cell conduction occurs through sarcolemmal gap junctions. Fig. 2 A–D shows examples of APs simultaneously recorded in SS and TATS of two adjacent cells. Trabeculae were locally stimulated in a region ∼2 mm apart from the recording area (Fig. 2A), and an ∼15 ms delay was found between the stimulus and the rapid AP upstroke phase (Fig. 2D), in agreement with the expected myocardium conduction velocity at room temperature (22). According to previous findings (23), the AP profile in the intact tissue is slightly different from that recorded in single myocytes, showing a more pronounced plateau phase. Repeated measurements (880 APs from 13 trabeculae from five rats) showed that AP profiles are identical in SS and TT at both low (0.2 Hz, Fig. 2D) and high (5 Hz, Fig. 2E) pacing rates (see mean data in Fig. S5). These experiments, in addition to confirming the tight electrical coupling between SS and TATS in healthy myocardium, show near-instantaneous AP propagation between the two membrane domains.

Fig. 2.

Action potential propagation in TATS. (A) Bright-field image of a rat ventricular trabecula. The yellow arrowhead marks the stimulation site and the yellow diamond encompasses the recording area. (Scale bar: 1 mm.) (B) TPF image of the area highlighted in yellow in A. (Scale bar: 20 μm.) (C) The region in the yellow box of B shows two adjacent myocytes magnified. (D) Normalized fluorescence traces from the scanned lines indicated in C. APs are elicited at 0.2 Hz, corresponding to the black arrowhead. The traces are the average of 10 sequential episodes. (E) Normalized fluorescence traces (average of 10 episodes) recorded simultaneously from SS and TT during stimulation at 5 Hz.

Disconnecting TTs from SS is expected to prevent AP propagation into TATS. We achieved a physical disconnection of TTs from SS (acute detubulation) with an osmotic shock technique (24). When the two membrane systems were electrically uncoupled, the SS AP was shorter (6), which is consistent with the excess of L-type Ca²⁺ current in TATS (Fig. S6). Although disconnecting TTs from SS would prevent the diffusion of the dye into TATS, if staining is performed before detubulation, even the TTs disconnected from the surface will be labeled, allowing Vm optical measurements in isolated TATS elements. In this class of experiments [stained and then detubulated (S/D)], we find that APs, always present in the SS, are mostly absent in the TATS (Fig. 3 B–D). Fig. 3E shows the distribution of TATS responses, normalized against SS (ΔVm_TATS/SS). In control myocytes [control (Ctrl) in black], the uniformity of AP amplitude in TATS and SS is shown by the Gaussian distribution centered at 1. In detubulated myocytes, the distribution (S/D in light gray) instead predominates near 0. This indicates that APs do not propagate into detached tubules. The amplitude distribution of detubulated myocytes also shows a significant fraction of responses with detectable fluorescence variations, presumably reflecting the electrical activity of tubules still coupled to the SS. In fact, the osmotic shock procedure usually induces incomplete detubulation; most TTs are disconnected whereas a subpopulation maintains connections with SS (Fig. S7).

Fig. 3.

TATS integrity is crucial for proper AP propagation. (A) Confocal image of a rat ventricular myocyte after formamide-induced osmotic shock (the myocyte is stained before the detubulation: S/D). TTs retain a regular architecture in the cell core whereas they appear fragmented in the subsarcolemmal regions (yellow arrows). (Scale bar: 10 μm.) (B) TPF image of the labeled TATS after the osmotic shock. (Scale bar: 5 μm.) (C) Normalized fluorescence traces from the scanned line indicated in B: no electrical activity is detected in the scanned TATS regions. (D) Average of the 10 sequential episodes shown in C. (E) Distribution of the relative fluorescence change (ΔF/F) measured in TATS and normalized by the ΔF/F measured in SS. The histogram compares the distributions found in two different classes of experiments: control (Ctrl, in black; total number of counts: 903; cells: 21; rats: eight) and stained before formamide-induced osmotic shock (S/D, in light gray; total number of counts: 1,306; cells: 16; rats: three). (F) TPF image from a myocyte stained after formamide-induced osmotic shock (D/S): only a subpopulation of TTs is well labeled. (Scale bar: 5 μm.) (G) Normalized fluorescence traces from the scanned line indicated in F: SS and TT 1 show regular APs, whereas TT 2 and TT 3 display nonregenerative electrical responses. Red asterisks in TT 3 highlight AP failures. (H) Average of 10 episodes for SS, TT 1, and TT 2. Separate averaging of six APs and four subthreshold events in TT 3. (I) The histogram compares the distributions of relative amplitudes found in control and in D/S (gray; total number of counts: 580; cells: 14; rats: three). (J) Fraction of TATS elements showing subthreshold events (STE) in Ctrl, S/D, and D/S myocytes. Each bar represents the expected value ± SE of the experimental binomial distribution.

To investigate the behavior of the tubules that remain connected to the SS, we stained cells after the osmotic shock [detubulated and then stained (D/S)] so that only tubules still continuous with the surface were labeled. As shown in Fig. 3 F–J, even though the tubules are connected to the surface, in some cases AP propagation fails. In Fig. 3G, for example, APs are resolved in the SS but they are absent in some tubules. In particular, in tubule 3, a two-state scenario characterized by regular APs and failure events is seen. By separately averaging APs and failure events (Fig. 3H), we find that the AP profiles in TATS have a similar shape to those in SS whereas the failures are more consistent with subthreshold events. This observation indicates that TATS is capable of self-generating APs. The presence of voltage-gated Na²⁺ channels in the tubular membrane (25) supports this finding. In D/S myocytes, the ΔVm_TATS/SS distribution is similar to that found in controls (Fig. 3I), but ∼12% of the TATS elements display subthreshold ΔVm (Fig. 3J). This indicates that connection with the SS is not sufficient by itself to ensure a full electrical coupling; the integrity of TATS is also required.

Heart failure (HF) is characterized by weakened contraction of heart muscle leading to cardiac remodeling, which further weakens the heart and can also cause deadly arrhythmias (26). Loss and disorganization of the TATS are early features of cardiomyocyte remodeling in HF (27). Here we have reproduced a rat model of post-ischemic HF (28) (Fig. S8). We compared the TATS morphology of ventricular myocytes isolated from healthy rat hearts (Fig. 4A, Ctrl) to that of HF myocytes (Fig. 4B, HF). TATS disorganization was quantified calculating the TT component by the fast Fourier transform (FFT) of the myocyte image. Fig. 4C shows that the TT component is statistically reduced (t-test based) in HF in agreement with previous observation (28). We find that HF tubular disorganization alters TATS APs. Similar to that found in acute detubulation, we see a significant fraction of AP failure events in TATS elements connected to SS (Fig. 5 A–D). We find that ∼7% of TATS elements display no active responses, highlighting intracellular AP propagation anomalies. This suggests that TATS morphological remodeling in HF increases the connection resistance between SS and some TATS elements, impeding an adequate conduction of membrane depolarization.

Fig. 4.

Morphological TATS remodeling in HF. TATS confocal images in control (Ctrl) (A) and in post-ischemic HF (B) myocytes stained with di-4-AN(F)EPPTEA. The region in the yellow box is shown magnified 4×. In Ctrl, the TATS pattern is homogeneous and characterized by a FFT peak at 0.5 μm⁻¹. In HF, patchy areas devoid of TATS are present, and the remaining network is characterized by a reduction of the 0.5-μm⁻¹ peak as clearly shown by the power spectrum profile in white. (C) TT component (quantified by FFT peak analysis) measured in Ctrl (black; cell: 17; rats: three) and in HF (orange; cells: 13; rats: two). Each bar represents the mean ± SE.

Fig. 5.

AP propagation is impaired in HF. (A) Representative TPF image of a HF myocyte. (B) Normalized fluorescence traces from the scanned line indicated in A: SS and TT 1 show regular APs, TT 2 and TT 3 display nonregenerative electrical responses, and TT 4 highlights local arrhythmic events (blue asterisks). (C) The histogram compares the distributions of relative amplitudes found in control (Ctrl) and HF (total number of counts: 2,773; cells: 44; rats: three). (D) Fraction of TATS elements showing subthreshold events (STE) in Ctrl and HF. (E) Frequency of TATS spontaneous depolarization (arrhythmic event, or AE) observed in Ctrl, in D/S, and in two categories of HF: HF− is defined as HF TATS elements that do not display regular APs upon stimulation, and HF+ is defined as HF TATS elements showing normal active responses upon stimulation. Each bar represents the mean ± SE.

In the TATS of HF myocytes, a number of local spontaneous depolarizations (arrhythmic events, or AEs) occur (Fig. 5B), which are never observed in control or in detubulated myocytes (Fig. 5E). As shown in Fig. 5E, spontaneous depolarization events are extremely rare in TATS elements of HF myocytes with normal APs, whereas they do occur in ∼55% of TATS elements exhibiting subthreshold events. Spontaneous depolarization events do not propagate to the sarcolemma but remain confined in the TATS element where they arise. This finding supports the hypothesis that the electrical coupling between tubules showing spontaneous depolarization events and the remaining TATS elements is impaired.

Discussion

Combining the advantages of random access microscopy with the improved photostability of a new fluorinated voltage-sensitive dye, we developed an imaging method to simultaneously record APs at multiple sarcolemmal sites with submillisecond temporal and submicrometer spatial resolution. We find that the TATS APs are identical in amplitude and time course to those of the SS. These data confirm that the electrical coupling between SS and TATS is tight and prevails over inhomogeneous membrane current distribution or diffusion restrictions. Thus, any local condition affecting the AP of a single membrane domain influences all domains.

The simultaneous optical recording of APs from different membrane domains in real time not only reveals a sarcolemma voltage space-clamp at unprecedented spatial and temporal resolution but also allows observation of single failure events occurring in structurally altered tubular systems. In fact, sequential recordings from different domains using standard line-scan approaches were unable to unveil AP propagation impairment.

We find that acute detachment by osmotic shock of most tubules from SS prevents AP propagation not only in the disconnected tubules, but also in some of the tubules that remain connected with SS. This indicates that the structural integrity of the TATS is crucial to maintaining a functional electrical coupling between the TATS and SS. In agreement with a validated mathematical model (16), the electrical coupling between SS and TATS can be quantified in terms of a mean resistance (R) of the tubular system. Taking into account that the TATS represents a parallel combination of TTs, , where n_TT is the number of attached TTs. Consequentially, a reduction of n_TT increases the resistance and worsens the electrical coupling between the two compartments. This can increase the probability that the depolarization wavefront propagating from the surface to TATS fails in reaching the AP threshold in the tubular membrane. To confirm this scenario, we used a computational model of rat ventricular myocyte (16) to predict how TATS structural alterations affect AP propagation. We find that when n_TT is markedly reduced, even small variations in tubular geometry (e.g., tubular stenosis and/or an increment of the effective tubular length related to the reduction in t-tubular mouths) can cause AP propagation failure (Fig. S9). This result suggests that the redundancy of TATS interconnections (axial tubules) represents a safety mechanism to guarantee a proper sarcolemmal space-clamp in the whole myocyte.

The pathological implications of this finding were investigated in failing hearts. We find that AP propagation into the remodeled TATS frequently fails, leading to subthreshold voltage variations. Following the same consideration as above, the HF-related tubular remodeling may increase the connection resistance between the TATS element and the surrounding sarcolemma, leading to local electrical uncoupling even in the presence of the membrane continuity suggested by the presence of the dye molecule.

Tubular subthreshold responses cannot activate local Ca²⁺ channels and trigger Ca²⁺ release from the SR. Recent observations in HF models (26) have shown that both t-tubular loss and L-type Ca²⁺ channel redistribution to the surface contribute to increase the number of uncoupled orphan ryanodine receptors (RyRs). Here we show that, in addition to orphan RyRs, Ca²⁺ release units coupled to dysfunctional TATS elements also fail to be recruited. This indicates that the morphological analysis of TT reduction and RyR-Ca²⁺ channel colocalization may underestimate the number of nonrecruited Ca²⁺ release units in HF.

In TATS elements exhibiting AP failure, we also find local spontaneous electrical activity (i.e., AEs). Electrically uncoupled TATS elements are presumably characterized by smaller capacitance than those tightly coupled to the SS. Thus, in accord with our experimental observations, even a small localized inward current (e.g., a Na⁺-Ca²⁺ exchanger current due to local Ca²⁺ sparks) can potentially induce a large local ΔVm sufficient to reach the AP threshold. We observed that AE do not propagate to the whole sarcolemma and thus cannot directly promote generalized cellular activity. On the other hand, spontaneous tubular activities can trigger local asynchronous SR Ca²⁺ release, which contributes to nonuniform myofilament activation (29) and SR Ca²⁺ content depletion (30), promoting contractile dysfunction in HF.

Furthermore, with the same mechanism, electrically uncoupled TATS elements can amplify Ca²⁺ sparks, leading to generalized proarrhythmogenic Ca²⁺ waves (23). As reported by Song et al. (26), detubulation can generate “orphan,” unregulated RyR clusters having a higher propensity toward diastolic spontaneous Ca²⁺ release. Accordingly, our results suggest that, in HF cells, spontaneous Ca²⁺ release from unregulated RyRs can elicit AEs in uncoupled TATS. In our view, two complementary mechanisms may cooperate in determining HF-related arrhythmogeicity: unregulated RyRs (which increase the rate of Ca²⁺ sparks) and uncoupled TATS elements, generating a positive feedback that turns local sparks into generalized Ca²⁺ waves, which are triggers for arrhythmias.

Materials and Methods

Myocardial Preparations.

Male Wistar Han rats (300–350 g) are used for the experiments in accord with local regulations for laboratory animal use. Ventricular myocytes and right ventricular trabeculae are isolated as previously described (31, 32).

Acute Detubulation.

Detubulation was induced by osmotic shock as previously described (24). Briefly, 1.5 M formamide was added to the cell suspension for 15–20 min; the cells were then resuspended in standard, formamide-free solution.

Rat Model of Post-Ischemic HF.

Male Wistar Han rats (190–230 g, Harlan Laboratories SRL) were properly anesthetized with fentanyl and droperidole. Myocardial infarction is induced by ligation of the left anterior coronary artery as previously described (28). Relatively large infarct areas are developed due to this procedure, comprising a large portion of the anterior free wall of the left ventricle and the apex. Cardiac function is monitored with echocardiography before surgery and is periodically checked after the intervention. Four weeks after the infarction, a left ventricular dilatation occurs together with a loss of contractile function (Fig. S8). Rats were killed 6–8 wk after surgery and used for cell isolation.

Di-4-AN(F)EPPTEA Dye Synthesis.

6-Dibutylamino-5-fluoro-naphthalene-2-carboxaldehyde (33) (30.1 mg, 100 μmol) and 1-(3-triethylammoniopropyl)-4-methylpyridinium dibromide (34) (39.6 mg, 100 μmol) are mixed in 4 mL of ethanol, and then two drops of pyrrolidine are added (Fig. S1). The solution was stirred at room temperature for 17 h and turned red after reaction. Solvent was removed by rotary evaporation, and the residue was purified by chromatography (SiO2-amino, 7.5:92.5 MeOH/CH2Cl2) to give a red solid (36.0 mg, 53%).

RAMP.

The excitation was provided by a high-power, passively mode-locked fiber laser operating in the 1,064-nm spectral range (200 fs width pulses at 80 MHz repetition rate). The scanning head was developed by using an acousto-optical deflection (AOD) system composed of two AODs crossed at 90°. To compensate for the larger dispersion due to two crossed AODs, we used an acousto-optic modulator placed at 45° with respect to the two axes of the AODs (35). An oil immersion objective was used in experiments with isolated cells, and a water immersion objective was used in experiments with trabeculae. Emission filters of 655 ± 20 nm were used. The RAMP microscope is capable of commutating between two positions of the focal plane in ∼4 μs. In a typical measurement, we probed 5–10 different sarcolemmal sites. The length of the scanned lines ranged from 2 to 10 μm with integration time per membrane pass ∼100 μs, leading to a temporal resolution on the order of 0.5–1 ms. The signal-to-noise ratio for real-time optical AP detection was ∼10.

Confocal Image Analysis.

The TATS morphology was analyzed in terms of TT density using a method based on FFT. We quantified the TT density by measuring the ∼0.5 μm⁻¹ peak component (related to TT periodicity) of the power spectrum of the image. In detail, the fraction of the 0.5 μm⁻¹ component was obtained by normalizing the area under the peak to the total area subtending the power spectrum curve. An open source imaging processing software (ImageJ 1.43u) was used to perform the analysis.

Optical Trace Analysis.

Optical data were analyzed with a custom-made software written in LabVIEW 7.1 (National Instruments). A Vm-independent bleaching effect during each individual line scan was observed; this distortion to the two-photon fluorescence (TPF) time course was normalized from the recordings by fitting to a monoexponential decay. The probability distribution of ΔVm_TATS/SS was displayed by a relative frequency histogram. The histogram (binning size = 0.15) was calculated by measuring the peak ΔF/F in TATS and normalized by the peak ΔF/F in SS for each episode of each element. The TATS elements showing subthreshold events were scored using the AP threshold ΔF/F = 0.037 in agreement with the experiments described in Fig. S4. Approximating the binomial distribution with a normal distribution (central limit theorem), the confidence interval was calculated as:

where is the fraction of successes in a Bernoulli trial process estimated from the statistical sample, and N is the sample size. The central limit theorem was not applied to a binomial distribution where the fraction . In these cases, the Wilson score interval was applied:

Spontaneous depolarizations (i.e., AEs) were scored when, in the recording trace, we found that ΔF/F > 0.074 (twofold AP threshold) not correlated to the stimulus. The binomial proportion confidence interval was calculated as described above.

Additional Methods.

Detailed methodology is described in SI Materials and Methods.