Heart and Circulatory Physiology

Caffeine-induced Ca2+ sparks in mouse ventricular myocytes

Michael Ritter, Zhi Su, Kenneth W. Spitzer, Hideyuki Ishida, William H. Barry


Ca2+ sparks are spatially localized intracellular Ca2+ release events that were first described in 1993. Sparks have been ascribed to sarcoplasmic reticulum Ca2+ release channel (ryanodine receptor, RyR) opening induced by Ca2+ influx via L-type Ca2+ channels or by spontaneous RyR openings and have been thought to reflect Ca2+ release from a cluster of RyR. Here we describe a pharmacological approach to study sparks by exposing ventricular myocytes to caffeine with a rapid solution-switcher device. Sparks under these conditions have properties similar to naturally occurring sparks in terms of size and intracellular Ca2+ concentration ([Ca2+]i) amplitude. However, after the diffusion of caffeine, sparks first appear close to the cell surface membrane before coalescing to produce a whole cell transient. Our results support the idea that a whole cell [Ca2+]i transient consists of the summation of sparks and that Ca2+ sparks consist of the opening of a cluster of RyR and confirm that characteristics of the cluster rather than the L-type Ca2+ channel-RyR relation determine spark properties.

  • ryanodine receptors
  • confocal microscopy

caffeine is known to bind to the ryanodine receptor Ca2+ release channel (RyR) in sarcoplasmic reticulum (SR) and to induce opening of the channel (13). This effect of caffeine has been utilized to induce release of Ca2+ from the SR, creating a whole cell Ca2+ transient (11, 12), and the amount of Ca2+ released by caffeine has been used to indicate the extent of loading of the SR with Ca2+ (16,17). The decline of the intracellular Ca2+ concentration ([Ca2+]i) transient in the presence of caffeine has been used to estimate the contribution of other Ca2+ uptake and extrusion systems because persistent opening of the RyR by caffeine eliminates contribution of the SR to Ca2+ sequestration (1). Although these characteristics of the effects of caffeine have been extensively studied, little is known about the pattern by which this release occurs when an isolated cardiac myocyte is abruptly exposed to caffeine. We therefore hypothesized that caffeine-induced whole cell [Ca2+]itransients, especially at the very beginning when [caffeine]i is low, might start with Ca2+ release events that could resemble Ca2+sparks.

In this work, we utilized a novel Nipkow type confocal microscope (7,8) to image Ca2+ in isolated mouse ventricular myocytes during abrupt exposure to caffeine achieved with a rapid solution switcher device (15, 18). Our results indicate that the first manifestation of abrupt exposure to caffeine is the appearance of Ca2+ sparks at the periphery of the cell with subsequent coalescence of sparks toward the interior of the cell to produce a whole cell Ca2+ transient. Analysis of Ca2+sparks, induced by caffeine indicates that these sparks have characteristics similar to those that have been described previously using line scan devices. These results provide further support for the hypothesis that a whole cell [Ca2+]i transient consists of a summation of sparks, suggest that Ca2+ release occurs from a cluster of RyR to produce a spark, and further demonstrate the independence of Ca2+ sparks from L-type Ca2+ channel function.


Adult mouse myocyte preparation.

Adult murine cardiomyocyte isolation was performed as previously described (17). Hearts were obtained from female Balb/c mice 6–8 wk of age. The heart was removed from a deeply anesthetized mouse and attached to a sterile aortic cannula providing continuous retrograde coronary artery perfusion at 37°C at a pressure of 70–90 mmHg and an approximate flow rate of 1.8 ml/min. After perfusion with Ca2+-free modified Tyrode solution for 5 min, hearts were digested with 0.90 mg/ml collagenase D (Boehringer Mannheim) in 25 μM CaCl2 containing modified Tyrode solution for 7–12 min. Modified Tyrode solution consisted of (in mM) 126 NaCl, 4.4 KCl, 1 MgCl2, 18 NaHCO3, 11 glucose, 4 HEPES, and 30 2,3-butanedione monoxime and 0.13 U/ml insulin, gassed with 5% CO2-95% O2, which maintained the pH at 7.4. The left ventricles were cut into small pieces in 100 μM Ca2+-containing modified Tyrode solution. These pieces were gently agitated and then incubated in the same solution containing 2% albumin at 30°C for 20 min. The cell suspension was then centrifuged at 300 rpm for 3 min, and the pellet of the cells was resuspended in 200 μM Ca2+ and 2% albumin Tyrode solution and allowed to settle for 20 min at 30°C. The cells were then resuspended in culture medium composed of 5% heat-inactivated fetal bovine serum (Hyclone), 47.5% MEM (GIBCO), 47.5% modified Tyrode solution, 10 mM pyruvic acid, 4.0 mM HEPES, and an additional 6.1 mM glucose at 30°C in a 5% CO2 atmosphere until use. Isolated cells were all used for experiments within 4 h of isolation.

Ca2+ imaging.

Myocytes were incubated in HEPES solution containing 10 μM fluo 3-acetoxymethyl ester (fluo 3-AM, Molecular Probes) at 30°C in the dark for 30 min, washed twice in a dye-free HEPES solution, and then attached to laminin-coated glass coverslips. The loading solution was prepared by diluting a 100 μM fluo 3 stock solution, which contained 0.45% pluronic F-127 (Molecular Probes), 10% DMSO, and 90% heat-inactivated FCS (Gibson). HEPES solution consisted of (in mM) 126 NaCl, 4.4 KCl, 1.0 MgCl2, 1.08 CaCl2, 24 HEPES, 13 NaOH, 11 glucose, and probenecid (pH 7.4). For fluo 3 fluorescence measurements, coverslips were placed on an inverted epifluorescence microscope (Nikon) and myocytes were superfused with HEPES. Excitation light was generated with a 488-nm argon laser (model 5500A Laser, Ion Laser Technology). Excitation time was reduced by a computer-controlled shutter in front of the laser leaving the shutter open for only the imaging time (<1.5 s/measurement) to minimize photobleaching and phototoxicity. [Ca2+]i transients were evoked by exposing cells to 10 mM caffeine for 500 ms with a rapid solution-switcher device, which changes the solution bathing a cell in 3–4 ms (15). Localized Ca2+ release events in field-stimulated myocytes pretreated with nifedipine (20 μM) were also studied.

The image generated was detected with a Nipkow-type confocal system (Yokogawa, Tokyo, Japan), as previously described (7, 8). A Nikon Diaphot inverted microscope and a ×40 fluor oil objective lens (Nikon) were used. Fluorescence was detected via a band-pass filter (530 nm) by a charge-coupled device camera (XR GEN3, Solamere Technology). Images were recorded on a Macintosh 9600/233 computer at video rate and analyzed with NIH Image 1.62a software. For whole cell Ca2+ transient measurements, the entire cell was included in a region of interest. [Ca2+]imeasurements were expressed as F/F0, where F is fluorescence at a given time point and F0 is the resting fluorescence of the cell. In some experiments, a photomultiplier tube was used to measure the time course of the whole cell [Ca2+]i transient, as previously described (17).

Measurement of [Ca2+]ispark diameter and location.

Before analysis a rank filter was applied to reduce noise, which replaced each pixel by the median value in its 3 × 3 neighborhood. Ca2+ spark diameters were then measured from the points where the fluorescence intensities were 25% above average resting cell level. The location of a spark was defined as the distance between the cell edge and the fluorescence peak within a spark. All values are given in micrometers. Results are expressed as means ± SE.


In these myocytes under these experimental conditions, spontaneous Ca2+ release events were observed to occur at a rate of less than one per whole cell confocal layer per second. Ca2+ release events induced by caffeine are illustrated in Fig. 1, which shows consecutive images of an isolated mouse ventricular myocyte taken 33 ms apart after exposure to 10 mM caffeine for 500 ms with a rapid solution-switcher device. A clearly visible [Ca2+]i increase (shown in pink) is first seen along the cell surface membrane and in this case is more pronounced at the upper end of the cell because the caffeine-containing solution arrived at this end first. Figure 1,image 2, shows four local Ca2+ release sites (arrowheads) and one confluent site. The [Ca2+]i elevation then proceeded down the length of the cell and from the subsarcolemmal region to the interior. The time course of changes in whole cell [Ca2+]i triggered by a caffeine pulse is shown in Fig. 2. With our experimental setup it took >33 ms after exposure to caffeine until the first rise in [Ca2+]i could be detected. The peak [Ca2+]i was reached at 233 ms after switcher activation. With the same rapid solution switcher and a photomultiplier tube to measure whole cell [Ca2+]i transients, peak [Ca2+]i was reached in 277 ± 16 ms (n = 14 where n is no. of cells, means ± SE). This time course of the whole cell [Ca2+]i transient induced by abrupt exposure to caffeine is consistent with previous findings reported by O'Neill et al. (11).

Fig. 1.

Example of a caffeine-induced Ca2+ transient in an isolated mouse ventricular myocyte. Images 1–6 were taken consecutively 33 ms apart. First image was taken 33 ms after switching. Bar in image 1, 10 μm. Arrowheads in image 2 show local Ca2+ release sites.

Fig. 2.

Time course of whole cell intracellular calcium concentration ([Ca2+]i) transient triggered by abrupt exposure to 10 mM caffeine (n = 15 cells, means ± SE). F, fluorescence at a given time point; F0, resting fluorescence of cell.

Because the [Ca2+]i transient develops relatively slowly compared with a transient induced by an electrical signal, we were interested in the characteristics of the early [Ca2+]i increases. Figure3 shows fluorescent intensities along a line drawn through the center of an early Ca2+ release event. We found that these first [Ca2+]i rises were consistently seen at the end of the cell where the caffeine first arrived and in very close proximity to the cell surface membrane. [Ca2+]i peaks of these early events showed an amplitude (F/Fo) of 1.94 ± 0.06 (n = 16) and were on average 2.8 ± 0.2 μm away from the cell edge (n = 14). The diameter of these “sparks” was 3.3 ± 0.3 μm (n = 14). In field-stimulated, nifedipine (20 μM)-pretreated myocytes, localized [Ca2+]i release events occurred at the same time throughout the cell and had a diameter of 2.7 ± 0.1 μm (n = 17), slightly but significantly smaller than sparks induced by caffeine (P < 0.05). The amplitude of electrically induced sparks (F/Fo) was not different (1.98 ± 0.06).

Fig. 3.

Fluorescent intensity plot along a line drawn through center of a caffeine-evoked spark. Point zero on horizontal axis corresponds to cell edge.


Recording spatially localized intracellular Ca2+ release events, so-called Ca2+ sparks, is complicated by the fact that during an electrical stimulus, sparks disappear within the whole cell Ca2+ transient. In the past, various groups have applied different techniques to avoid this problem. For example, spontaneous sparks in resting cells can be observed (4). Other techniques include the use of high concentrations of intracellular Ca2+ buffers to limit spatial spread of local Ca2+ release, application of Ca2+ channel blockers such as verapamil to limit frequency of spark occurrence by partially blocking the L-type Ca2+ channel, induction of a small Ca2+ current via depolarization to −30 mV to limit transmembrane trigger Ca2+ entry, and photolysis of caged Ca2+ (6, 9, 10).

In our experiments, rapid exposure of mouse ventricular myocytes to caffeine also appears to cause localized Ca2+ release events. Caffeine-induced sparks appear to be independent of activation of the L-type Ca2+ channel, confirming the observation of Cheng et al. (4) that initiation of sparks does not require Ca2+ influx via the Ca2+ channel. The initiation of caffeine-induced sparks is dependent on the time it takes to expose the cell to caffeine-containing solution, which is 3–4 ms in our case (15), and on the diffusion time of caffeine to the RyRs. Therefore, the sparks we observed originated in a diffusion-dependent pattern close to the cell surface membrane, whereas sparks triggered by an electrical pulse are seen along the T tubules throughout the cell (6, 14). Cannell et al. (3) described spatial nonuniformities in triggered whole cell [Ca2+]i transients, which they hypothesized could be secondary to a temporal and spatial summation of a large number of Ca2+ sparks. In these experiments we show clearly that a caffeine-induced [Ca2+]i transient begins at independent release sites, each of which shows the properties of a Ca2+ spark. These independent release sites appear to culminate in a whole cell [Ca2+]i transient, further indicating that [Ca2+]i transients consist of a large number of sparks.

Ca2+ entry via the L-type Ca2+channel seems to trigger a cluster containing a few RyRs that form the spark. Bridge et al. (2) estimated by noise analysis that >18 RyR may be involved in the production of a Ca2+ spark and that their opening is an “all-or-none” event. Our findings suggest that caffeine-induced Ca2+ sparks also originate from a cluster of RyR openings. Caffeine-induced Ca2+ sparks might be expected to be different from electrically triggered sparks. First, caffeine affects the gating behavior of RyRs by increasing frequency and duration of open time without changing their conductance in lipid bilayer measurements (11). Also, caffeine-activated RyRs do not inactivate, demonstrating a steady-state behavior on a minute time scale (11). The fact that the caffeine-induced sparks we observed had a slightly larger diameter than electrically stimulated ones could reflect this difference. The morphology of the electrically induced sparks we detected is consistent with previous reports of electrically induced (2) and spontaneous (5) mouse ventricular myocyte spark characteristics (diameter 2.5 μm, amplitude 1.7–2.5) measured by line scan. Further studies using higher time resolution will be needed to better define caffeine-induced spark properties. However, the use of caffeine pulses could provide a valuable new approach to study early Ca2+ release events independent of Ca2+ influx through the L-type Ca2+channel, which by itself follows complex kinetics. Also the caffeine pulse technique does not require voltage clamp and resulting dialysis of the cell interior.


We are indebted to Pamela Larson for preparation of the manuscript.


  • Address for reprint requests and other correspondence: W. H. Barry, Div. of Cardiology, Univ. of Utah Health Sciences Center, 50 North Medical Dr., Salt Lake City, UT 84132 (E-mail:whbarry{at}med.utah.edu).

  • This work was partially supported by the Deutsche Forschungsgemeinschaft and by a Specialized Center of Research on Heart Failure Grant HL-53733.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract