Late Ca2+ Sparks and Ripples During the Systolic Ca2+ Transient in Heart Muscle Cells.

RATIONALE: The development of a refractory period for Ca2+ spark initiation after Ca2+ release in cardiac myocytes should inhibit further Ca2+ release during the action potential plateau. However, Ca2+ release sites that did not initially activate or which have prematurely recovered from refractoriness might release Ca2+ later during the action potential and alter the cell-wide Ca2+ transient.
OBJECTIVE: To investigate the possibility of late Ca2+ spark (LCS) activity in intact isolated cardiac myocytes using fast confocal line scanning with improved confocality and signal to noise. METHODS AND
RESULTS: We recorded Ca2+ transients from cardiac ventricular myocytes isolated from rabbit hearts. Action potentials were produced by electric stimulation, and rapid solution changes were used to modify the L-type Ca2+ current. After the upstroke of the Ca2+ transient, LCSs were detected which had increased amplitude compared with diastolic Ca2+ sparks. LCS are triggered by both L-type Ca2+ channel activity during the action potential plateau, as well as by the increase of cytosolic Ca2+ associated with the Ca2+ transient itself. Importantly, a mismatch between sarcoplasmic reticulum load and L-type Ca2+ trigger can increase the number of LCS. The likelihood of triggering an LCS also depends on recovery from refractoriness that appears after prior activation. Consequences of LCS include a reduced rate of decline of the Ca2+ transient and, if frequent, formation of microscopic propagating Ca2+ release events (Ca2+ ripples). Ca2+ ripples resemble Ca2+ waves in terms of local propagation velocity but spread for only a short distance because of limited regeneration.
CONCLUSIONS: These new types of Ca2+ signaling behavior extend our understanding of Ca2+-mediated signaling. LCS may provide an arrhythmogenic substrate by slowing the Ca2+ transient decline, as well as by amplifying maintained Ca2+ current effects on intracellular Ca2+ and consequently Na+/Ca2+ exchange current.

Cardiac excitation–contraction coupling is mediated at the cellular level by the near-synchronous activation of ≈104 microscopic Ca2+ release events called Ca2+ sparks.[1,2] This occurs because the cardiac action potential (AP) opens L-type Ca2+ channels (LTCC) in the surface membrane to produce a local increase in Ca2+, which in turn opens Ca2+-sensitive channels (ryanodine receptors) in the adjacent junctional sarcoplasmic reticulum membrane (jSR).[3] The spatial restrictions associated with this “local control mechanism” provide this signal transduction pathway both high gain and stability[1,4] and forms the cornerstone of our current understanding of excitation–contraction coupling, explaining the time- and voltage-dependence of the regenerative Ca2+ release process.[5-7] Ca2+ sparks normally occur with high probability at the start of the Ca2+ transient,[8,9] and Ca2+ release during the Ca2+ spark is terminated in ≈10 ms, probably via SR depletion–dependent processes.[10] The cytoplasmic Ca2+ concentration then returns toward resting levels in a few hundred milliseconds as Ca2+ is pumped back into the SR (via SERCA2a [the sarco/endoplasmic reticulum Ca ATPase]) and across the surface membrane (mainly via NCX [sodium/calcium exchanger]).[4,11,12] This currently accepted view of excitation–contraction coupling has led to changes in the time course of Ca2+ decline being generally attributed to changes in SERCA2a and NCX activities with smaller contributions from a sarcolemmal Ca2+-ATPase and mitochondria.[13,14] However, continued SR release (or leak) should oppose SR reuptake and slow the time course of the Ca2+ transient, as seen in a phospholamban knockout mice with CamKIIδc overexpression.[15]

Meet the First Author, see p

Methods

The data that support the findings of this study are available from the corresponding author on reasonable request.

Ventricular cardiomyocytes from New Zealand White rabbits were field-stimulated at 0.5 Hz at 1.2× threshold at 22°C. The enzymatic method for isolating rabbit epicardial myocytes have been described previously.[16] To avoid possible problems associated with cell dialysis, whole-cell patch clamp techniques were not used. Cell movement artifacts were prevented by adding 10 mmol/L 2,3-butanedione monoxime to normal Tyrode superfusion solution. It should be noted that late Ca2+ sparks (LCSs) were also seen in the absence of 2,3-butanedione monoxime and at 37°C (see Online Figure I). 2,3-Butanedione monoxime may modify the relative importance of triggers for Ca2+ sparks in the experiments shown here (see Online Supplement). Fast block of LTCCs was achieved by local superfusion with Cd2+ (10 or 100 μmol/L) plus 1 μmol/L sulforhodamine-B in normal Tyrode superfusion solution from a pressurized micropipette, the fluorescence of which allowed determination of the local concentration of Cd2+ (and hence degree of LTCC block).

More extensive Methods are given in the Online Supplement.

Results

By using high sensitivity detectors and improved confocality, we have been able to detect additional Ca2+ release events (LCSs) during the entire time course of the cellular Ca2+ transient. The left panel in Figure 1A shows a typical Ca2+ transient recorded from a ventricular myocyte using confocal line scanning,[9] and the right panel shows a high-pass filtered and contrast-enhanced version of this image in which these LCS can be seen more clearly. Individual LCS (Figure 1B) had an increased amplitude compared with diastolic Ca2+ sparks (Figure 1C) but did not exhibit any changes in duration (≈40 ms) or spatial full width at half maximum (≈1.8 μm; data not shown). The increase in amplitude of the LCS can be explained by a reduction in cytoplasmic Ca2+ buffering power because of cytoplasmic Ca2+ binding sites (such as troponin) becoming occupied during the Ca2+ transient.[14] A simple, obvious, explanation for the genesis of the LCSs would be that they arise from jSR that was either not activated during the upstroke of the Ca2+ transient or was uncoupled or orphaned[17] from t-tubules (which carry the AP to the cell interior). To examine this idea, we labeled t-tubules (Figure 1D) and measured the Euclidian distance to the LCS centroid (Figure 1E) and compared the latencies for the upstroke of the Ca2+ transient at LCS sites to overall Ca2+ transient latency (Figure 1F). Perhaps unexpectedly, LCS occurred close to t-tubules and at positions where the Ca2+ transient developed with a shorter than average latency (P<0.05). This suggests that LCS usually arise from sites that are not orphaned but are, in fact, well coupled, a view also supported by the presence of an extensive t-tubule network (top and middle panels, Figure 1D) that did not show any signs of the disruption associated with orphaning.[17]

Figure 1.

Late Ca A, High-resolution recordings (confocal pinhole set to 1 Airy unit) of Ca2+ release during normal Ca2+ transients (left) at 0.5 Hz. Image enhancement by subtraction of the low-pass filtered transient shows LCS more clearly (white boxes; right). B, Comparison of Ca2+ sparks at rest (top left) and LCS (top right) showed high spatiotemporal similarity. The temporal profiles of averaged events (23 Ca2+ sparks and 18 LCS) show that LCSs have a similar time course to resting Ca2+ sparks in the same cell (bottom). C, Average LCS amplitude was greater than resting Ca2+ sparks (*P<0.05, **P<0.01, 196 LCS, 42 diastolic n/N=12/6). D, Di-8-ANEPPS (4-(2-[6-(dioctylamino)-2-naphthalenyl]ethenyl)-1-(3-sulfopropyl)pyridinium) labeling of t-tubules (top) was used to calculate a distance map to nearest t-tubule (middle). The line scan position is indicated by the white line and the location of LCS by white circles. LCS occurred near t-tubules and at regions with early evoked Ca2+ release (bottom). E, The median distance to nearest t-tubule was shorter for LCS sites than the cell average distance (n/N=11/5). Kolmogorov–Smirnov test. F, Ca2+ transient latency at sites with LCS was shorter than the cell average latency (n/N=28/8). *P<0.05.

After activation, Ca2+ spark sites enter a refractory period,[18,19] which should oppose any secondary activation. The spontaneous Ca2+ spark refractory period has been estimated by analyzing the inter-Ca2+ spark interval from ryanodine-modified rat[18] or CamKIIδc-overexpressing mouse[15] myocytes. We performed a similar analysis on LCS sites: Figure 2A shows 2 LCS sites with different inter-LCS intervals. Initial LCS (red arrows) followed the AP-evoked Ca2+ transient with a latency distribution shown as in Figure 2B. Only a few LCS are seen shortly after the onset of the Ca2+ transient, but this increases and reaches a peak before declining with a half time of ≈500 ms, which is similar to the half time of the whole-cell Ca2+ transient (also shown in Figure 2B behind the histogram bars) in this species at room temperature.[20] The time between the first and second LCS (blue arrows Figure 2A) is shown in Figure 2C, and the amplitude of the second event compared with the first is shown in Figure 2D. Because LCS amplitude restitution matches the reported rabbit jSR refilling time course,[19] jSR refilling probably determines LCS amplitude recovery. In this regard, LCS behave similarly to ryanodine-stimulated diastolic Ca2+ sparks.[18] By dividing Ca2+ spark probability by an exponential fit, the first and second LCS activation probability can be derived.[18] It is apparent that the LCS activation probability is (essentially) the same for both the first and second LCS (Figure 2E), suggesting that recovery from refractoriness is a dominant factor in their genesis—but what is their trigger?

Figure 2.

Restitution and refractory behavior of late Ca Time and amplitude restitution of LCS reveal their sarcoplasmic reticulum (SR) load dependence. A, The interval of successive LCS arising from the same location (top), and their relative amplitudes were measured in background-subtracted recordings (bottom). B, The probability density function of LCS after evoked Ca2+ release was similar to the time course of the normalized average Ca2+ transient (solid color behind histogram bars) and decreased exponentially from the maxima (red line). C, The probability of a second LCS at the same site had a similar time dependence to the first LCS (B). D, After the first LCS, the amplitude of the second LCS increased with a time constant, τ=77 ms. E, The probability of LCS activation (PLCS) was essentially the same after evoked Ca2+ release (red) or after a prior LCS (blue; P=0.86, extra sum-of-squares F test). B, 350 LCS from n/N=19/9; C and D, 79 LCS pairs from n/N=10/8.

During the long plateau phase of the AP, LTCCs continue to open stochastically[21,22] and might trigger LCS as individual release sites recover from their refractory period. If this is the case, one might predict that LTCC inhibition should reduce the number of LCS. As illustrated in Figure 3A, blocking ≈75% of LTCCs with extracellular Cd2+ (a fast LTCC open-state blocker[23]) paradoxically increased the number of LCS (Figure 3C). However, straightforward interpretation of this experiment is complicated by the block of LTCCs during the upstroke of the AP, reducing jSR site activation (Figure 3D). This effect would increase release site availability because fewer sites would then be refractory at later times. Importantly, these data also show that the increased number of LCS produced in this condition make a significant contribution to the time course of the Ca2+ transient (Figure 3B and 3E). To remove the complication arising from changes in the number of jSR release sites activated during the AP, we rapidly applied a higher concentration of Cd2+ just after the upstroke of the Ca2+ transient to selectively block later LTCC openings. To show the effect of such early LTCC blockade clearly, we show an exemplar (Figure 3F) that had a higher rate of LCS production shortly after the upstroke of the Ca2+ transient, which contributed to an extended rising phase which can be seen under condition of lower SR load (eg, Grantham and Cannell[21]). The time of arrival and local concentration of Cd2+ was measured by fluorescently labeling the solution with sulforhodamine-B which, with the Kd for Ca2+ block, allowed us to estimate that ≈90% LTCCs should be blocked after the upstroke of the Ca2+ transient (Figure 3G). This reduced the number of LCS, but only by 40% (Figure 3H), supporting the idea that LCS can also be triggered by the rise in cytosolic Ca2+ during the Ca2+ transient. As expected, the application of Cd2+ shortly after the stimulus and upstroke of the AP had no effect on the amplitude of the Ca2+ transient (Figure 3I) but slightly shortened its duration (Figure 3G and 3J), which is likely to be as result of AP shortening[24] (because of LTCC blockade) and the reduced number of LCS.

Figure 3.

Effects of L-type Ca Blocking LTCC before or after evoked Ca2+ release shows LCS can be triggered by late LTCC activity. A, Line scan recordings of consecutive Ca2+ transients in normal Tyrode’s (NT) and in NT+10 μmol/L Cd2+. The extracellular solution was rapidly changed between contractions to preserve sarcoplasmic reticulum (SR) load. Cd2+ application produced more spatial nonuniformities in early Ca2+ release and increased the number of LCS. B, The Ca2+ transient decay was clearly delayed in Cd2+. C, Cd2+ increased mean LCS frequency by ≈40%. D, Cd2+ slightly decreased the Ca2+ transient amplitude and (E) increased its duration compared with NT. F, Rapid application of NT+100 μmol/L Cd2+ (after the Ca2+ transient upstroke) preserved early Ca2+ release but decreased the number of LCS. G, The Ca2+ transient upstroke was preserved in Cd2+ (green line) compared with NT (blue). H, Mean LCS frequency was reduced by ≈40% by Cd2+. I, Average Ca2+ transient amplitude was not reduced by rapid Cd2+ application while the Ca2+ transient duration was decreased slightly (J; P=0.05). C–E, n/N=15/5; H–J, n/N=12/4. **P<0.01, ***P<0.001 paired t test.

LCS production is a part of a continuum of behavior that spans the low spontaneous Ca2+ spark rate during diastole (≈1 per 100 μm/s scanned) to high rates (≈104 per 100 μm/s) during the upstroke of the Ca2+ transient.[8,9] Normally, Ca2+ sparks interact weakly,[25] but when SR and cytoplasmic Ca2+ levels increase, Ca2+ waves can develop from the sequential recruitment of Ca2+ spark sites.[26] The spatiotemporal relationship between LCS site activation (Figure 4A) was analyzed by calculating autocorrelograms (Figure 4B). In most of our experiments, such autocorrelation analysis showed only a time-dependent relationship between spark sites, reflecting the refractory period (Figure 4C) described earlier. However, in a subset of cells that were more highly Ca2+ loaded (Figure 4D), the autocorrelogram showed multiple peaks, indicating that some LCS were both spatially and temporally correlated (Figure 4E). The right panel of Figure 4D illustrates the chevron patterns in LCS production that can be seen by eye, and analysis of the 2D autocorrelogram showed an apparent propagation velocity between LCS sites ≈114 μm/s (Figure 4F), similar to typical macroscopic Ca2+ wave propagation velocities.[26,27] We call these novel propagating LCS events Ca2+ ripples as (1) they are smaller in amplitude, (2) do not propagate over the entire cell, and (3) occur during the declining phase of the Ca2+ transient, although they are clearly related to the well-known phenomenon of Ca2+ waves which can occur during the diastolic period in cardiac myocytes.[3,26,27]

Figure 4.

Late Ca A, Unprocessed high resolution recording showing multiple LCS. B, The autocorrelogram of (A) reveals the time dependence between repeating events (indicated by yellow arrow), with a mean delay of 145 ms, as highlighted in (C). D, left, Further pacing resulted in increased Ca2+ load and many more LCS. Right, Emphasizes how LCS appear to trigger additional sites forming multiple propagating Ca2+ ripples (marked by yellow chevron overlays). E, The 2D autocorrelogram shows that some LCS have both temporal and spatial relationships to other LCS. F, Calculation of propagation velocities in the 2D autocorrelation showed a dominant peak at ≈114 μm/s. Points indicated in (C) and (F) were highly significant (P<10–5) (blue lines show mean±5 SD of scrambled data).

Discussion

The presence of LCS during the normal Ca2+ transient has important implications for understanding the complex interplay of SR Ca2+ release site activation, refractoriness, SR Ca2+ reuptake, and triggers for CICR (Ca-induced Ca release; see Online Figure II). Reduced LTCC activation cannot only disrupt the initial phase of Ca2+ release, leading to dyssynchrony,[28,29] but also increase the number of LCS that can slow the decline of the Ca2+ transient. It is notable that reduced LTCC activation can also occur with pathological changes in t-tubules,[30] APs,[13] or signal transduction cascades.[3] Increased SR leak in a CamKIIδc-overexpressing mouse model has been shown to slow the decline of the Ca2+ transient, and some LCS activity during the Ca2+ transient can be seen in Figure 7 of that paper.[15] These results are also consistent with the ability of release sites to recover from refractoriness sufficiently quickly for some fraction to become reactivated either by cytosolic Ca2+ or LTCCs. While some uncertainty exists in the relative roles of cytoplasmic Ca2+, LTCC, and NCX in triggering LCS and their effect on the Ca2+ transient time course (an uncertainty compounded by 2,3-butanedione monoxime used to inhibit movement artifacts—see Online Supplement), it is clear that LCS production will be sensitive to all of these triggers. In the case of heart failure, any increase in LCS production could exacerbate the existing problem of slowed Ca2+ reuptake because of decreased SERCA2a activity.[31] Further complexity is added by the changes in Ca2+ transient time course also affecting LTCC gating via prolongation of the AP because of NCX-generated current during the declining phase of the Ca2+ transient,[13] as well as the differential responses of coupled and uncoupled release sites.[32]

Under normal conditions, LCS production is initially inhibited by the refractory period after Ca2+ spark activation,[18] but the time course of recovery is shorter than the duration of the plateau of the AP during which a sizeable LTCC current flows. Thus, LCS are more likely to occur late in the AP, and slowing the decline of the Ca2+ transient may contribute to the antagonism between inward NCX current and repolarization reserve.[13] As illustrated in Online Figure II, some pathological changes in the excitation–contraction coupling cycle could increase the probability of LCS which, in turn, may prolong the duration of the Ca2+ transient[15] and AP duration. This forms a new positive feedback pathway that will promote AP prolongation and further Ca2+ influx via LTCC, further destabilizing Ca2+ cycling and increasing all forms of Ca2+ leak.[33]

While more work is needed to fully explore the implications of the novel results presented here, it is now apparent that SR Ca2+ release in the form of LCSs can continue at lower rates throughout the cardiac Ca2+ transient rather than solely during the upstroke of the Ca2+ transient as usually modeled.

Acknowledgments

This work was supported by the British Heart Foundation (grant RG/12/10/29802) and Medical Research Council (MR/N002903/1).

Disclosures

None.