Orphaned ryanodine receptors in the failing heart

Abstract

Heart muscle is characterized by a regular array of proteins and structures that form a repeating functional unit identified as the sarcomere. This regular structure enables tight coupling between electrical activity and Ca²⁺ signaling. In heart failure, multiple cellular defects develop, including reduced contractility, altered Ca²⁺ signaling, and arrhythmias; however, the underlying causes of these defects are not well understood. Here, in ventricular myocytes from spontaneously hypertensive rats that develop heart failure, we identify fundamental changes in Ca²⁺ signaling that are related to restructuring of the spatial organization of the cells. Myocytes display both a reduced ability to trigger sarcoplasmic reticulum Ca²⁺ release and increased spatial dispersion of the transverse tubules (TTs). Remodeled TTs in cells from failing hearts no longer exist in the regularly organized structures found in normal heart cells, instead moving within the sarcomere away from the Z-line structures and leaving behind the sarcoplasmic reticulum Ca²⁺ release channels, the ryanodine receptors (RyRs). These orphaned RyRs appear to be responsible for the dyssynchronous Ca²⁺ sparks that have been linked to blunted contractility and, probably, Ca²⁺-dependent arrhythmias in diverse models of heart failure. We conclude that the increased spatial dispersion of the TTs and orphaned RyRs lead to the loss of local control and Ca²⁺ instability in heart failure.

Ca²⁺ sparks, the elementary Ca²⁺ release events during the excitation–contraction (EC) coupling process in heart (1–5), are triggered by the opening of L-type Ca²⁺ channels (LCCs) and sum to produce the [Ca²⁺]_i transient (6). Each Ca²⁺ spark arises from a nanometer-sized structural and functional unit or “couplon,” which includes both (i) LCCs in the sarcolemma or transverse tubules (TTs), and (ii) a cluster of ryanodine receptors (RyRs) that reside in the sarcoplasmic reticulum (SR) membrane and face the LCCs across a 15-nm gap (7–9). During the cardiac action potential (AP), the nearly simultaneous opening of LCCs synchronizes the triggering of Ca²⁺ sparks from the RyR clusters (also referred to as Ca²⁺ release units, or CRUs).

Ca²⁺ signaling in heart failure is characterized by contractile dysfunction (10, 11), a reduction in the magnitude of the [Ca²⁺]_i transient (10–13), and depletion of Ca²⁺ from SR Ca²⁺ stores (14–16). Additionally, reduced synchrony in Ca²⁺ sparks triggering (17, 18), and Ca²⁺-dependent arrhythmias (19–21) are common, but current explanations for how these phenomena develop are incomplete. Clues come from recent work on altered Ca²⁺ signaling in healthy cells (under certain conditions) and diverse heart failure (HF) models. Local (subcellular) Ca²⁺ overload can develop rapidly between heart beats if Ca²⁺ release is blunted in normal ventricular myocytes. This alteration was shown to underlie Ca²⁺ transient alternans when RyR open probability or LCC current (I_Ca) was reduced in single cell voltage–clamp experiments (22, 23). In severe but compensated canine model of left ventricular hypertrophy, we recently observed paradoxically wider and more frequent Ca sparks despite reduced global SR load, perhaps due, at least in part, to local regions of SR Ca²⁺ overload (24). In tachycardia-induced HF, Kamp and colleagues (25, 26) have described a regional loss of TTs within ventricular myocytes, but other evidence suggests that TTs remain normal.^†† Additionally, many other factors may contribute to arrhythmogenesis, including reduced K⁺ channel expression, increased Na⁺/Ca²⁺ exchanger activity (21, 27), and altered Ca²⁺ channel gating (28); therefore, possible links between structural changes and dysfunctional Ca²⁺ signaling in this HF model remain unclear. In a cell culture investigation using pig ventricular myocytes, dyssynchronous Ca²⁺ release was observed as the cells in culture dedifferentiated and lost TTs (18). In contrast, in a feline heart failure model, local Ca²⁺ signaling failure was attributed to alterations in early action potential repolarization, and not to structural abnormalities (29). We used this collection of provocative results to inform our planned experiments.

To investigate the mechanism underlying dyssynchronous Ca²⁺ sparks, we undertook an integrated investigation of single ventricular myocytes from spontaneously hypertensive rats that developed HF. Ca²⁺ sparks and [Ca²⁺]_i transients in these cells were abnormal and dyssynchronous. An investigation of the Ca²⁺ release fluxes revealed that the dyssynchronous sparks were secondary Ca²⁺ release events resulting from local propagation of the Ca²⁺ signal. We also identified a major factor that underlies these signaling abnormalities: spatial dispersion of the TTs. In contrast to the work on cells from tachycardia-induced HF, remodeling and misplacement of TTs rather than TT loss are the primary events. We incorporate these findings into a new mechanistic model of Ca²⁺ signaling dysfunction in HF.

Results

Dyssynchronous Ca²⁺ Release Is Due to Secondary Ca²⁺-Induced Ca²⁺ Release (CICR), Not Dyssynchronous LCC Triggering in Failing Myocytes.

Molecular, structural, and functional changes in cardiac myocytes were investigated in spontaneously hypertensive rats (SHRs) that develop overt heart failure (SHR/HF; see Table 1, which is published as supporting information on the PNAS web site). Fig. 1 A–C shows that SHR/HF cells exhibit reduced amplitude of [Ca²⁺]_i transient. This reduction has been largely attributed to a decrease in the Ca²⁺ content of the SR due to reduced expression and function of SR Ca²⁺ ATPase (SERCA) (14, 15, 30). The smaller [Ca²⁺]_i transients in the SHR/HF myocytes also reveal a change in the spatial character of the Ca²⁺ release, as shown in Fig. 1B. In certain regions (Fig 1B, arrows), Ca²⁺ release does not occur at the time of the initial depolarization, but instead local Ca²⁺ signals in these “defective” regions increase a short time after the initial depolarization. This phenomenon has been seen before in cells isolated from failing rabbit (17) and human (18) hearts and given the name “dyssynchronous Ca²⁺ release.” These delays in the triggering of Ca²⁺ sparks contribute to the prolongation of the [Ca²⁺]_i transient and its slowed relaxation (Fig. 1D). However, how they come about in HF remains elusive. Although such delayed Ca²⁺ releases could arise from delays in the triggering of Ca²⁺ release by late openings of LCCs, they could also result from slowly propagating CICR.

Fig. 1.

Ca²⁺ signaling in normal and failing ventricular myocytes. (A) A 1-Hz train of field stimulated [Ca²⁺]_i transients is shown as line-scan images from a fluo-4-loaded control cell from Wistar-Kyoto (WKY) rats (Upper) and fluorescence records (Lower). (B) A train of stimulated [Ca²⁺]_i transients for age-matched spontaneously hypertensive rats with heart failure (SHR/HF). White arrows mark the positions on line-scan images in which [Ca²⁺]_i releases are missing at fixed locations in consecutive beats within the same cell. Late or dyssynchronous Ca²⁺ release events were revealed after initial missing releases. (Inset) WKY and SHR/HF Ca²⁺ transients on expanded time scale. Red arrowed line denotes prolonged Ca²⁺ release in the failing myocyte. (C) Reduced amplitude of [Ca²⁺]_i transient (F/F₀) in SHR/HF cells. (D) Prolonged [Ca²⁺]_i transient in SHR/HF cells. n = 30–52 cells from four to six hearts. ∗∗, P < 0.01 vs. WKY cells. T₂₀, the plateau duration measured at 20% below the peak level.

Measurement of local Ca²⁺ release flux can be used to distinguish between late triggering and locally propagated CICR. This measurement is possible by using heavy intracellular Ca²⁺ buffering (4 mM EGTA) and a low-affinity Ca²⁺ indicator (Oregon Green 488 BAPTA-5N, 1 mM) (31, 32). The inclusion of EGTA restrains Ca²⁺ diffusion, suppresses global Ca²⁺ elevation, and prevents secondary regenerative CICR. The resulting signals, namely “Ca²⁺ spikes,” therefore result only from SR Ca²⁺ release triggered directly by LCC openings. Thus, the measurement of Ca²⁺ spikes will help dissect secondary CICR from the dyssynchronous LCC triggerings of RyR Ca²⁺ release. Typical results from control and SHR/HF myocytes are shown in Fig. 2A and B, respectively. Marked traces in Fig. 2 A and B show the time courses of Ca²⁺ release events at specific sites in Fig. 2C.

Fig. 2.

Defective local EC coupling in SHR failing myocytes. (A) Ca²⁺ spikes shown in control cells (WKY). Local EC coupling and SR Ca²⁺ release function was examined by using Ca²⁺ spikes method (31, 32). Briefly, EGTA (4 mM) was added to the pipette-filling solution to limit the diffusion of released Ca²⁺. The local Ca²⁺ release signal was detected with the Ca²⁺ indicator (1 mM Oregon Green 488 BAPTA-5N). Local signals with ΔF/F₀ > 4 SD were considered spike events. (B) Ca²⁺ spikes from SHR/HF cells. White arrows mark the missed events at the potential release sites. (C) Time course of local Ca²⁺ release at sites 1–8 for control and HF cells. (D) The spatially averaged SR Ca²⁺ release flux (Jsr) was decreased in SHR/HF (ANOVA, P < 0.01 vs. WKY group over the voltage range, n = 11–15 cells from approximately four to five hearts). (E) Ca²⁺ spike latency as a function of test voltage. Spike latency was measured from the beginning of the depolarizing pulse to the take-off (2 SD above baseline) of Ca²⁺ spike. (F) Fractional TT-SR activation: percentage of activated TT-SR junctions to total junctions measured. (ANOVA, P < 0.01 vs. WKY group over the voltage range, n = 11–15 cells from four to five hearts).

There are four notable features of the Ca²⁺ release events in SHR/HF cells identified by using the Ca²⁺ spike method. First, immediately upon depolarization, there are many more “missed” events (see e.g., Fig. 2B, sites 6 and 7) in SHR/HF than control (Fig. 2F). This finding is in agreement with the failure of synchronized Ca²⁺ release shown in Fig. 1B. Second, in the presence of a Ca²⁺ chelator (Fig. 2B), each missed release is not followed by a late event even though experiments in the absence of chelator (Fig. 1) and regular RyR distribution in normal cells (Fig. 3) indicate that a potential calcium release site should be present. Third, the synchrony of triggered Ca²⁺ release events is indistinguishable from that in control (Fig. 2 A, B, and E), suggesting that the synchrony of primary voltage-gated LCC-triggered Ca²⁺ releases is unchanged in HF. Simultaneous electrophysiological recording revealed no significant changes in whole-cell I_Ca density except for a moderate slowing of I_Ca inactivation kinetics (see Fig. 5, which is published as supporting information on the PNAS web site), similar to previous reports in other HF models (13, 30). Fourth, the magnitude of the release flux (Jsr) in SHR/HF myocytes is significantly smaller than control (Fig. 2D).

Fig. 3.

Alterations in transverse tubule morphology in failing rat ventricular myocytes. (A) Control myocytes from WKY rats were exposed to 10 μM of the fluorescent lipophilic marker Di-8-ANEPPS for 10 min, and imaged with a confocal microscope. (B) Myocytes from an age-matched SHR/HF rat were imaged as in A. (C) Zoomed-in view (×2.5) of A with high-contrast color look-up table. (D) Zoomed-in view of B. (E) TT line tracing from C. (F) TT line tracing from D. (G) Power versus spatial frequency in the longitudinal (x) dimension, at zero frequency in the y dimension, computed by using Fourier analysis (see Materials and Methods). Failing myocytes display a clear decrease in power at ≈0.5 μm⁻¹, which corresponds to the average TT spacing of 2 μm seen in healthy cells. The 2nd and 3rd harmonic components seen in WKY are almost completely absent in the SHR/HF cells, consistent with the chaotic appearance of TT in the HF cell. (H and I) Density of LEs (H) and TEs (I). Bars represent the percentage of cell pixels positive for Di-8-ANEPPS staining that were part of a continuous line of stained pixels extending for at least 2 μm in either the longitudinal (H) or the transverse (I) direction. See Materials and Methods for a description of the normalization and thresholding steps involved in computing these percentages. ∗∗∗, P < 0.001 vs. WKY controls, n = 9–13 cells from four hearts.

The missed events and the smaller Ca²⁺ releases at the TT–SR junctions (i.e., the reduced Ca²⁺ spike amplitudes) are consistent with local failure of CICR and reduced SR Ca²⁺ content. The absence of delayed release events in spike measurement, however, suggests a possible explanation for the mystery of Ca²⁺ release dyssynchrony in HF: the late sparks seen in SHR/HF (Fig. 1B) seem to arise as regenerative or secondary CICR driven by the primary, LCC-evoked Ca²⁺ sparks that constitute the rising phase of the [Ca²⁺]_i transient. Under conditions for Ca²⁺ spike measurement, intracellular EGTA suppresses the global Ca²⁺ elevation and thereby inhibits the secondary release component, such that only Ca²⁺ spikes directly triggered by LCC openings are observed (33). If the secondary sparks in SHR/HF (Fig. 1B) had been triggered by LCC openings, we would have observed delayed Ca²⁺ spikes during the 300-ms depolarization in Fig. 2B.

The missed events and associated secondary releases at fixed sites in consecutive beats (Fig. 1B) suggest that release dyssynchrony results from a structural rather than a functional defect. This defect causes excitation-Ca²⁺ release (EC) uncoupling but allows secondary CICR.

TT System Remodeling in SHR Failing Hearts.

In an analysis of the mechanisms of defective Ca²⁺ signaling and decreased EC coupling “gain” in heart failure, Gomez et al. (34) suggested that spatial remodeling of SR or TTs may contribute to this process. This speculation has been controversial. At one extreme is the report that very large TT loss occurs in volume overload HF (25), but there is also evidence against TT loss in HF.^†† The complexity of the process is now well appreciated (35, 36). If spatial remodeling did occur as myocytes responded to prolonged stress and if this remodeling involved SR-TT organization, then it may provide a mechanistic link between the disease process and dyssynchronous Ca²⁺ release.

In mammalian ventricular myocytes, the system of TTs is a network of tubular invaginations roughly 250 nm in diameter with many transverse elements (TEs) and a few longitudinal elements (LEs) that normally form an imperfect rectilinear grid (37, 38). This cell-wide, regularly distributed TT network ensures nearly instantaneous electrical excitation and synchronous triggering of Ca²⁺ release through the entire cytoplasm during membrane depolarization (4). Fig. 3 A and B shows images of the TTs in ventricular myocytes from control and SHR/HF hearts. The spatial dispersion of TTs in SHR/HF is readily apparent by inspection and by quantitative comparison. The close-up views (Fig. 3 C and D) and the diagram panels (Fig. 3 E and F) highlight a dramatic TT reorganization characterized by a loss of TEs, a gain in LEs, and an overall chaotic appearance of TT in HF. Both a 2D Fourier analysis (Fig. 3G) and a quantitative computation of the TE and LE densities in the TT images (Fig. 3 H and I; see Materials and Methods) confirmed the dramatic changes in the TT organization in SHR/HF cells. Such profound TT remodeling would alter EC coupling and the triggering of Ca²⁺ sparks. The nature of the change, however, depends on the relative positions of the LCCs and the Ca²⁺ release units.

Reorganization of ECC Proteins in SHR Heart Failure: “Orphaned” RyRs.

LCCs or dihydropyridine receptors (DHPRs) in TTs, whose Ca²⁺ currents serve as the primary physiological trigger of Ca²⁺ sparks that sum to produce the [Ca²⁺]_i transient (2, 3, 39), are known to be located near RyR clusters (7–9). As shown in Fig. 4B and E, the spatial distribution of DHPRs in HF becomes disorganized and irregular, similar to the changes in the TT network. On the other hand, RyR distribution retains its regular striated appearance (Fig. 4 A and D), indicating that RyRs do not readily follow the TTs and DHPRs to reorganize in SHR/HF. By overlaying the two signals, we found that the clear colocalization of these proteins, as described by previous investigators (8), has been changed by the development of HF (Fig. 4 C and F). Quantitative measurements corroborated that the degree of colocalization between RyRs and DHPRs is significantly decreased in SHR/HF, compared with control cells (Fig. 4 G and H). Thus, as the TTs bearing DHPRs move away from the Z-line structures, they seem to leave behind the RyR clusters as functionally orphaned elements. By this means, the voltage dependence and magnitude of the cellular I_Ca resulting from LCC openings remain unchanged, but with slower Ca²⁺-dependent inactivation. This change results from weaker or loss of local retrograde signaling, due to an increased average distance, between RyRs and LCCs.

Fig. 4.

Reorganization of EC coupling proteins in failing ventricular myocytes: Orphaned RyRs. (A) Immunofluorescence image of a WKY cell for RyRs (red). (B) Coimmunofluorescence image (with A above) for DHPRs (green). (C) Colocalization image for A and B. (D–F) As in A–C for SHR/HF cell. (G) 2D protein colocalization analysis was achieved by using unbiased automatic thresholding methods (see Materials and Methods). The percentage of DHPRs colocalized with RyRs was reduced in HF. (H) Measurement of colocalization of RyRs with DHPRs. ∗∗, P < 0.01; ∗∗∗, P < 0.001 vs. WKY, n = 15 and 10 cells for WKY and SHR/HF, respectively.

Discussion

Dyssynchronous Ca²⁺ Release in Heart Failure: A Mechanistic Model.

By integrating these findings, we propose a mechanistic model that can explain functional disruption of EC coupling in SHR/HF and may provide an explanation of the origin of dyssynchronous Ca²⁺ signaling. The reorganization of TTs produces an array of repositioned DHPRs, and this restructuring strands RyRs at the Z lines, leaving some RyRs orphaned. The orphaned RyR clusters that become physically separated from their LCC partners do not respond to LCC openings in the normal manner, and the local failure of Ca²⁺ release seen immediately upon depolarization is a result of this restructuring. The orphaned clusters can, however, be activated later to release Ca²⁺ but with variable latencies as local Ca²⁺ is elevated by nearby Ca²⁺ sparks that have been triggered normally. It is also noteworthy that the structurally more normal regions, which do respond to depolarization, release smaller amounts of Ca²⁺, as shown in Fig. 2, probably due to the decreased SR Ca²⁺ content.

According to this model, dyssynchronous Ca²⁺ release should occur in normal cells when conditions are manipulated appropriately. This dyssynchronous release has been demonstrated in atrial cells (40), in model systems by small depolarizations with high background Ca²⁺ influxes to produce Ca²⁺ alternans (23), in dedifferentiated myocytes in culture that, among other things, become partially detubulated (18, 41), and in massively detubulated myocytes subjected to osmotic shock (42).

TT Remodeling of Fetal Phenotype: A Unique Finding from SHR Model?

The morphology of TTs in HF is a topic of continuing investigation and may well be different in different models. The SHR model mimics human hypertensive HF, which develops over a long period. Failing SHR myocytes undergo strikingly structural remodeling, characterized by reappearance of prevalent longitudinal tubules, a fetal phenotype that normally disappears during early development (43). This observation is in general agreement that a failing heart is trying to rejuvenate by switching on early development programs. A model of rapid pacing (in dog) produces a rapidly progressive HF and may be a model for volume overload HF. In this pacing-induced HF model, Kamp and colleagues (25, 26) reported loss of TTs. Brief preliminary reports on human heart failure found no change in TT density and spatial organization^†† or, like us, found TT spatial reorganization.^‡‡ The differences in the models, the severity of the disease, and the time course of the development of the disease probably account for the differences in observations.

Finally, it should be noted that structurally and/or functionally orphaned RyR clusters, such as those we have identified here, are in fact fairly common features of diverse myocardial cell types. In comparatively small cells with an underdeveloped TT system, such as mammalian atrial and avian ventricular myocytes, RyRs are coupled to LCCs at the external cell membrane (peripheral couplings) but not in the cell interior (44, 45). Propagating Ca²⁺ waves that activate the orphaned central RyRs seem to be part of the normal EC coupling process in these cells. In ventricular myocytes, naturally orphaned RyRs appear on corbular or extended junctional SR that is not associated with TTs (46, 47), but they are many fewer in number than the coupled RyRs. The function of these naturally orphaned RyRs at these regions is not well understood. They may be functionally distinct from those created by long-term pathology. Because tight coupling between most RyR clusters and LCCs is essential to maintain local control of CICR (48, 49), a significant increase in the number of orphaned RyR clusters in pathology should lead to a decrease in coupling efficacy and increased instability in Ca²⁺ release, as we have shown here.

In summary, we have investigated the paradox of Ca²⁺ instability: dyssynchronous Ca²⁺ release in HF where SR Ca²⁺ content is low. We conclude that the HF-associated Ca²⁺ instability arises from structural changes in the relationship between the TTs and the SR that disrupt the local control of RyRs by LCCs. These structural alterations and mechanistic remodeling may also contribute to other cardiomyopathies. Our findings thus complement other cellular and molecular changes that may contribute to defective and/or dyssynchronous Ca²⁺ signaling in HF. These changes include action potential alterations (29, 50) and altered RyR2 sensitivity (51, 52) (see also ref. 53 for a review).

Materials and Methods

Confocal Ca²⁺ Imaging.

Field-stimulated Ca²⁺ transients.

Ventricular myocytes were isolated from SHRs with overt HF or age-matched Wistar-Kyoto rats by using standard enzymatic methods (54). Myocytes were loaded with Ca²⁺ indicator fluo-4 AM (10 μM for 15 min). In a recording chamber, cells were field-stimulated at 1 Hz until steady state. Then, confocal line-scan imaging was performed by a Zeiss LSM 510 confocal microscope equipped with an argon laser (488 nm) and a ×63, 1.3 NA oil immersion objective. Line-scan images were acquired at sampling rate of 1.92 ms per line, along the longitudinal axis of the cell. Digital image processing was performed by using custom-devised routines with idl programming language (Research Systems, Boulder, CO).

Ca²⁺ spikes and SR release function measurements.

Ca²⁺ spikes and SR release function measurements have been described (31, 32).

TT Imaging and Analysis.

Myocytes were loaded with the lipophilic membrane marker Di-8-ANEPPS (10 μM in Tyrode solution for 10 min). After washout, TTs were examined by using the confocal microscope. Stacks of images were acquired at 0.25 or 0.5 μm intervals in z axis. The center images from each cell were selected for analysis.

Analysis.

Two complementary approaches were used to analyze quantitatively the observed changes in TT structure. The power spectrum of each image was computed as the magnitude of the 2D spatial Fourier transform by using the idl programming language. The average intensity of each image was subtracted before computing transforms to correct for differences in staining between images. Results shown are of power as a function of spatial frequency in the longitudinal (x) dimension at zero frequency in the transverse (y) dimension. Frequency components along other directions did not exhibit any obvious differences between control and SHF/HF groups. Because images of TTs from different cells were acquired with slightly different pixel spacing, the power spectrum from each cell was interpolated onto a regular grid so that spectra could be summed to produce plots of “average” power for each group.

The second method of analysis was to quantify the pixels within each cell that were part of a stained line extending for at least 2 μm in either the longitudinal or transverse directions. Several steps were taken in this analysis to correct for differences in image intensity between cells and ensure a bias-free result. First, the overall intensity histogram of each image was plotted, and the nadir between the histogram’s two peaks was defined as the “cell threshold.” This empirical threshold could differentiate between cell pixels and background pixels with >98% sensitivity and specificity. Pixels positive for Di-8-ANEPPS staining were then distinguished from negative pixels by using the “tubule threshold,” defined as the value that selected 30% of the total pixels within the cell mask. We then computed whether each pixel was part of a continuous line extending at least 2 μm, oriented at up to ±20° relative to the direction of interest. The percentages of cell pixels that were part of these LEs or TEs are displayed in Fig. 3 H and I, respectively. Before performing this computation, 35 pixels (roughly 3 μm) were removed from the edges of each cell image to ensure that staining of the outer cell membrane did not artificially inflate the result.

Double Immunostaining of DHPRs and RyRs.

Double labeling was performed on isolated rat myocytes with anti-RyR and anti-DHPR primary antibodies. Rabbit polyclonal anti-DHPR (β2a subunit) antibody was a generous gift from M. Marlene Hosey (55). Mouse monoclonal anti-RyR antibody was purchased from Affinity BioReagents (Neshanic Station, NJ).

Isolated myocytes were initially fixed with 100% ethanol (>24 h). The fixed cells were washed by BSA (1%) and normal goat serum (2%) containing PBS solution (blocking buffer) for four times at 15-min intervals. Primary antibodies were then diluted (1:100) in blocking buffer simultaneously and incubated with myocytes overnight at 4°C. After washing out of uncombined antibodies (four times × 15 min, with blocking buffer), secondary antibodies (Alexa Fluor 488-conjugated goat anti-rabbit IgG and Alexa Fluor 633-conjugated goat anti-mouse IgG) were added (1:100) to blocking buffer and incubated with the cells for 2 h, followed by washout of four times × 15 min with blocking buffer. Finally, the cells were mounted on slides and examined with a confocal microscope (Zeiss LSM 510, 100 × 1.3 NA oil immersion objective). Quantitative protein–protein colocalization analysis was performed offline by using imaris (kindly provided by Bitplane Inc., Saint Paul, MN), with unbiased automatic thresholding methods (56).

Statistics.

Data were expressed as mean ± SE. Student’s t test and ANOVA analysis were applied when appropriate. P < 0.05 was considered statistically significant.

Abbreviations

TT

transverse tubules

RyR

ryanodine receptors

LCC

L-type Ca²⁺ channel

DHPR

dihydropyridine receptor

SR

sarcoplasmic reticulum

EC

excitation-contraction

CICR

Ca²⁺-induced Ca²⁺release

SHR

spontaneously hypertensive rat

HF

heart failure

WKY

Wistar-Kyoto rat

TE

transverse element

LE

longitudinal element

I_Ca

LCC current.