Modulation of focal and global Ca²⁺ release in calsequestrin-overexpressing mouse cardiomyocytes

Abstract

1. Focal and global Ca²⁺ releases were monitored in voltage-clamped control and hypertrophied calsequestrin (CSQ)-overexpressing mouse cardiomyocytes, dialysed with fluo-3, using rapid (120–240 frames s⁻¹) two-dimensional confocal imaging.

2. Spontaneous focal Ca²⁺ releases (Ca²⁺ sparks) were absent or significantly reduced in frequency in hypertrophied myocytes of CSQ-overexpressing mice compared to their age-matched controls. Sporadic Ca²⁺ sparks seen in CSQ-overexpressing myocytes had intensities and durations similar to those of controls although quantitative analysis showed a trend towards more diffuse focal releases.

3. Activation of Ca²⁺ current (I_Ca) failed to produce the typical sarcomeric Ca²⁺ striping pattern consistently seen in control myocytes. Instead, focal Ca²⁺ releases appeared as a disorganized patchwork of diffuse or ‘woolly’ fluorescence signals, resulting in slowly developing and reduced global Ca²⁺ transients.

4. Although the density of I_Ca in CSQ-overexpressing myocytes was only slightly smaller than that of controls, the inactivation kinetics of the current were greatly reduced, consistent with the much smaller rate of rise of cytosolic Ca²⁺.

5. Enhancement of I_Ca by elevation of [Ca²⁺]_o from 2 to 10 mM or addition of 3 μM isoproterenol (isoprenaline) failed to normalize the frequency of spontaneous Ca²⁺ sparks at rest or the pattern and the magnitude of I_Ca-gated Ca²⁺ transients. Isoproterenol was somewhat more effective than elevation of [Ca²⁺]_o.

6. In sharp contrast, low (0·5 mM) caffeine concentrations that produced no measurable effects on I_Ca or Ca²⁺ transients in control myocytes, re-established spontaneous focal Ca²⁺ releases in CSQ-overexpressing cells, triggered large I_Ca-gated cellular Ca²⁺ transients, and strongly enhanced the kinetics of inactivation of I_Ca.

7. Our data suggest that impaired Ca²⁺ signalling in CSQ-overexpressing myocytes results from reduced co-ordination and decreased frequency of Ca²⁺ sparks. The impaired Ca²⁺ signalling could not be restored by procedures that increased I_Ca, but was mostly restored in the presence of caffeine, which may alter the Ca²⁺ sensitivity of the ryanodine receptor.

In cardiac muscle the influx of Ca²⁺ through the Ca²⁺ channel is critical in triggering Ca²⁺ release from the sarcoplasmic reticulum (SR). The released Ca²⁺ in turn inactivates the L-type Ca²⁺ channel by interacting with the ‘Ca²⁺ sensing’ domains of the C-terminal tail (Soldatov et al. 1998) expressing calmodulin-binding properties (Zühlke et al. 1999; Peterson et al. 1999; Qin et al. 1999). Ca²⁺ mediated cross-signalling between the Ca²⁺ channel and ryanodine receptor (RyR) allows for tight functional coupling between these proteins, such that the influx of Ca²⁺ through the Ca²⁺ channel activates Ca²⁺ release from the RyR and the released Ca²⁺ in turn inactivates the Ca²⁺ channel, thereby inhibiting the release process (Sham et al. 1995). Quantitative analysis of the contribution of the released Ca²⁺ to the inactivation of the Ca²⁺ channel suggests that about 70 % of Ca²⁺ channel inactivation is determined by the Ca²⁺ released via the ryanodine receptor (Adachi-Akahane et al. 1996).

In the cardiac Ca²⁺ signalling cascade, the Ca²⁺ storage capacity of SR and the extent to which Ca²⁺ can be released from the low affinity (K_d≈ 1 mM) Ca²⁺ binding protein calsequestrin (CSQ) (MacLennan & Wong, 1971; Ikemoto et al. 1989) may also play a critical role in the release process (Kawasaki & Kasai, 1994). The structural properties of CSQ and its localization within the lumen of the SR suggest that this protein may interact with the RyR via the junctional proteins junctin and triadin (Jones, et al. 1995; Guo et al. 1996; Zhang et al. 1997). Recently we reported that mouse hearts overexpressing canine CSQ hypertrophied by 50–100 % and developed heart failure (Jones et al. 1998). Isolated myocytes from such hearts were 50–100 % larger and had significant electrophysiological and electromechanical alterations (Knollmann et al. 1998b). The most prominent features of electrical remodelling of these CSQ-overexpressing hearts were prolongation of the action potential and Q-T interval, downregulation of the transient outward K⁺ current (I_to), and greatly slowed kinetics of inactivation of I_Ca (Jones et al. 1998; Knollmann et al. 1998b). Since I_Ca-gated Ca²⁺ release was impaired in CSQ-overexpressing myocytes despite the presence of 10- to 15-fold larger caffeine-releasable Ca²⁺ pools (Jones et al. 1998), we suggested that impaired Ca²⁺ release was mostly responsible for the marked slowing of the inactivation kinetics of I_Ca. In a subsequent report on another strain of transgenic mice overexpressing murine calsequestrin, similar hypertrophy and slowing of the kinetics of I_Ca were described, but the extent of impairment of I_Ca-gated Ca²⁺ release was not quantified (Sato et al. 1998).

In the present report we have quantified the properties of focal and global Ca²⁺ release in this model and have studied the nature of impaired Ca²⁺ signalling in CSQ-overexpressing myocytes by interventions that modify the Ca²⁺ channel or the Ca²⁺ sensitivity of the RyR. We find that agents that enhance Ca²⁺ current are less effective in re-establishing normal Ca²⁺ signalling than those that sensitize the Ca²⁺ release channel. A preliminary report of this work has already appeared (Wang & Morad, 1999).

METHODS

Transgenic mice and cell isolation

Mice expressing 10- to 15-fold higher levels of cardiac calsequestrin, aged 9–21 weeks, and showing significant cardiac hypertrophy were used in this study (Jones et al. 1998). Single ventricular myocytes were enzymatically isolated using a modification of a previously described collagenase-protease technique (Mitra & Morad, 1985). Briefly, mice were deeply anaesthetized with sodium pentobarbital (50 mg kg⁻¹, i.p.), the chest cavity was opened and hearts were excised, resulting in exsanguination. Experiments were conducted in accordance with institutional and NIH established guidelines. The excised hearts were placed in ice-cold incubation solution of the following composition (mM): 30 taurine, 90 NaCl, 5.4 KCl, 10 Hepes, 10 glucose and 1 MgCl₂, titrated to pH 7.2 with NaOH. The aorta was cannulated and perfused for 5 min with an oxygenated incubation solution at 36°C containing 0.5 mM EGTA. The solution was then switched to an isolation solution containing 1 mg ml⁻¹ albumin (Sigma Chemical Co.), 0.12 mg ml⁻¹ protease (Type XIV, Sigma Chemical Co.) and 0.33 mg ml⁻¹ collagenase (Type IV, 204 U mg⁻¹, Worthington Biochemical Corp.) for 8 min. The heart was then removed and the ventricle minced and digested for an additional one to three digestion periods (5-10 min each) in a shaking water bath at 37°C. The resulting cell suspensions were collected after each digestion and stored at room temperature in an incubation solution containing an additional 0.2 mM CaCl₂. This procedure yielded 30–40 % rod-shaped myocytes that were used for up to 10 h.

Electrophysiological measurements

Myocytes were whole-cell clamped (Hamill et al. 1981) using borosilicate glass capillaries (resistance 2–3 MΩ) filled with an internal Cs⁺-rich solution (see below) containing 1 mM EGTA and 1 mM fluo-3. The cell capacitance was measured using previously established protocols (Cleemann & Morad, 1991). Currents were neither leak nor capacitance substracted (to obtain high quality data and to control for myocyte viability), but were series resistance compensated. Since imaging of the myocytes with a two-dimensional laser scanner required 10–15 min of dye dialysis, only cells with low leak current and clear striations were included in the final analysis of the results. All myocytes were clamped at a holding potential (V_h) of -70 mV prior to application of test pulses. Ca²⁺ currents were activated with test depolarizations to 0 mV either from -70 mV or more rarely from -50 mV to further suppress the residual TTX-resistant Na⁺ current.

Two-dimensional confocal Ca²⁺ imaging

Ca²⁺ currents and 2-D Ca²⁺ images of a whole-cell-clamped myocyte were simultaneously monitored in cells dialysed with 1 mM fluo-3 and 1 mM EGTA. This concentration of dye in combination with EGTA was chosen to limit the diffusion distance of Ca²⁺ and Ca²⁺-dye complex to about 50 nm (Adachi-Akahane et al. 1996; Cleemann et al. 1998). The combined use of fluorescent and non-fluorescent Ca²⁺ buffers and a confocal apparatus capable of rapidly scanning the cellular image via an accousto-optically steered laser beam (Noran, Odyssey, Madison, WI, USA) made it possible to obtain high resolution images of the rise and fall of cytosolic Ca²⁺ at high spatial and temporal resolution: ∼0.5 μm and 120–240 frames s⁻¹ (Cleemann et al. 1998). Briefly, the confocal apparatus was mounted on an inverted microscope (Zeiss, Axiovert 135 TV) with a water immersion objective lens (Zeiss, C-apochromat, ×40, NA 1.2). The 488 nm excitation beam was generated from an argon ion laser (Omnichrome, Chino, CA, USA). The confocal slit in the rapidly scanned direction was set at 50 μm (0.8 μm in the focal plane), and the fluorescent light emitted was measured with a high efficiency photomultiplier tube (Hamamatsu, Middlesex, NJ, USA). The data were acquired by a Silicon Graphics workstation computer (Indy, Unix operating system, Noran Instruments, Middleton, WI, USA) and stored temporarily in 100 Mbyte random access memory before being transferred to hard disk.

Fluorescence measurements were initiated 6–8 min after rupture of the membrane under the patch pipette. After this period of dialysis, the intracellular fluo-3 concentration was typically close to its equilibrium near the patch electrode, but continued to exhibit slight gradients towards the ends of the cell (Cleemann et al. 1997). Static images of Ca²⁺-dependent fluorescence show the average fluorescence intensity (F₀) calculated from several frames (Figs 1A and B (arrows) and 7A (arrows) and B). Dynamic Ca²⁺ signals were illustrated by sequences of frames showing the change in fluorescence (ΔF) relative to the average fluorescence (F₀) measured either in the presence of scattered focal Ca²⁺ releases (Figs 1 and 7), or immediately before and after Ca²⁺ releases activated by I_Ca (Figs 4A and B and 5A and B). This method of normalization was used to show Ca² sparks as clearly as possible without resorting to the use of contrast enhancement. In another approach (Figs 3A and B, 6A and B and 8A and B) we relied, in part, on a colour scale to present total fluorescence (F) with a large dynamic range. Tracings of cellular Ca²⁺ transients are shown as the average fluorescence of each frame normalized relative to the average resting fluorescence (F/F₀) prior to the ionic or pharmacological interventions (Figs 3F and G, 4E, 5E, 6E and F, and 8E). This type of normalization was used to permit comparison of results from different cells, and allow detection of systematic changes in both cellular transients (ΔF) and resting fluorescence (F_r) during the experiment (Figs 4F, 5F and 8F). Focal Ca²⁺ releases were identified, and followed in time, using a computerized algorithm (Figs 2 and 7). This algorithm (Cleemann et al. 1998) identified local fluorescence maxima by means of a centre-minus-surround detection kernel (inset in Fig. 2), which consists of pixels (0.207 μm spacing) approximating a central positively weighted disc (radius 0.8 μm) and a concentric negatively weighted ring (radius 1–1.5 μm). A new computer algorithm was developed in Visual Basic (Microsoft) to remove constraints of the previous program and allow routine measurements of the amplitude and size of Ca²⁺ sparks by fitting a Gaussian distribution to the local fluorescence intensity (Fig. 1C).

Figure 1. Spontaneous Ca²⁺ sparks in control (panel A) and CSQ-overexpressing (panel B) mouse cardiomyocytes and comparison of their size and magnitude (panel C).

The top images of two cardiomyocytes in panels A and B show the average fluorescence intensity (F₀) obtained from 60 frames recorded at 30 Hz in cells voltage clamped at a holding potential of -70 mV. In recordings obtained from 4 to 6 min after rupture of the patch, the fluorescence intensity is still highest near the dialysing patch electrode (e) suggesting that that the Ca²⁺-sensitive dye (1 mM fluo-3 with 1 mM EGTA as an adjuvant) has not equilibrated fully in the longitudinal direction and/or that the dye in the pipette is partially detected even though it is outside the confocal plane. Nuclear regions (n) and faint longitudinal lines suggestive of fibrils are often seen. The lower images (1, 2 and 3) show three consecutive frames recorded at 33 ms interval in the two cells. The frames show the change in fluorescence (ΔF) so that the outline of the cell and the position of the electrode are seen only as changes in the intensity of the noise. Panel A shows the typical presence in control cells of one or more Ca²⁺ sparks in each frame. Panel B shows one of the rare Ca²⁺ sparks (arrow) in CSQ-overexpressing cells. Notice that the Ca²⁺ sparks are brief so that at 30 Hz they are seen clearly only in single frames and have, at most, a faint afterglow in the following frame (panel B, arrows). Panel C shows a scattergram with regression lines of the normalized intensity (F₁/F₀) and dimension (σ) of Ca²⁺ sparks in control (^) or CSQ-overexpressing (•) cells. The fluorescence intensity of each Ca²⁺ spark was fitted by least-squares approximation by a Gaussian distribution (F₁exp(-[(x–x₀)²+ (y–y₀)²]/2σ²)) characterized by its centre (x₀, y₀), its standard deviation (σ, see inset) and the central increase in fluorescence (F₁) measured relative to the resting fluorescence (F₀).

Figure 7. A low concentration of caffeine (0.5 mM) causes spontaneous Ca²⁺ spark activity in CSQ-overexpressing mouse cardiomyocytes.

Panel A shows recordings in a layout similar to Fig. 1. Upper image, average fluorescence intensity (F₀). Lower images (1 and 2), sequential frames showing transient fluorescence changes (ΔF). Panel B shows the distribution and panel C the time course of Ca²⁺ sparks in a single cell exposed to 0.5 mM caffeine. The sample traces and their locations are labelled a-i.

Figure 4. Effect of increased extracellular calcium concentrations on I_Ca-gated Ca²⁺ release in CSQ-overexpressing cells.

Panels A and B show representative sample frames from stacks of fluorescence images (ΔF) recorded at 240 Hz during voltage-clamp depolarization from -70 to 0 mV with normal (2 mM, panel A) and elevated (10 mM, panel B) extracellular Ca²⁺ concentrations. Notice that elevation of [Ca²⁺]_o increased the number of Ca²⁺ sparks but did not change their appearance. The numbered sample frames (7, 8, 9, 11, 13 corresponding to filled symbols in panel E) show the increase in fluorescence (ΔF) after subtraction of the average background fluorescence. Panels C and E show, respectively, I_Ca and the cellular fluorescence signal in CSQ-overexpressing cardiomyocytes with 2 mM (CSQ) and 10 mM (CSQ+10Ca) [Ca²⁺]_o. The fluorescence signals (F) in panel E were normalized relative to the resting cellular fluorescence (F₀) measured with 2 mM [Ca²⁺]_o. Panel D shows average values of the amplitude of the current (I_Ca, right axis,) and its half-time of inactivation (t^1/2, left axis,) recorded with 2 mM (CSQ) and 10 mM (CSQ+10Ca) [Ca²⁺]_o. Similarly, panel F shows the effect of elevated [Ca²⁺]_o on the average values of the resting (F_r,) and transient (ΔF, ▪) cellular fluorescence signals normalized relative to the resting fluorescence (F₀) measured with 2 mM Ca²⁺. The numbers (12) next to the error bars indicate the number of cells examined.

Figure 5. Effect of isoproterenol on I_Ca-gated Ca²⁺ release in transgenic cells.

The layout of this figure is identical to that of Fig. 4. After the initial recording (CSQ), the transgenic cells were exposed to 3 μM isoproterenol for 1 min before testing the drug effect (CSQ+Iso). Panel A, sample frames before isoproterenol. Panel B, sample frames in the presence of isoproterenol. Arrows indicate the location of long lasting releases at a site that is barely visible in panel A, but in clearer view in panel B after some cell shortening. Panel C, membrane currents. Panel D, effect of isoproterenol on ensemble averages of half-time and amplitude of I_Ca. Panel E, Ca²⁺ transients. Panel F, effect of isoproterenol on ensemble averages of resting and transient fluorescence signals.

Figure 3. Comparison of I_Ca-induced Ca²⁺ release in voltage-clamped control (A, C and F) and transgenic (B, D and G) ventricular mouse cardiomyocytes.

Panels A and B show numbered sample frames of total fluorescence (F) recorded at 240 Hz at the times indicated by filled symbols in the tracings of the cellular Ca²⁺ transients (panels F and G). Arrows identify some Ca²⁺ sparks that can be followed from frame to frame. The colour scale provides a calibration of the fluorescence intensity (F) relative to the resting fluorescence (F₀= 1). Panels C and D illustrate the voltage-clamp protocol and the time course of the Ca²⁺ current. The bar charts compare the time course of the Ca²⁺ current (panel E) and the Ca²⁺ signal (panel H) in control (□) and CSQ-overexpressing () cells. The Ca²⁺ current is characterized by its half-time of inactivation (t^1/2, see panel C) while the Ca²⁺ signal is described by both its time-to-peak (TTP, see panel F) and its half-time of decay (t^1/2). The error bars indicate the standard error of the mean and adjacent numbers indicate the number of cells examined.

Figure 6. Caffeine-induced Ca²⁺ release in control (panels A, C and E) and transgenic (panels B, D and F) cardiomyocytes.

The numbered sample frames from a control cell (panel A) and a transgenic cell (panel B) were recorded as total fluorescence (F) at 30 Hz at the times indicated by filled and numbered symbols in the fluorescence tracings (F/F₀) of panels E and F. The timing of the exposure to 10 mM caffeine is indicated by vertical dashed lines and the labelled bars (Caffeine) in panels C and D, which also show the time course of the Na⁺-Ca²⁺ exchange current. Notice that frames 23 and 27 in panel B were recorded at a time when the inward Na⁺-Ca²⁺ exchange current continued to increase, but the Ca^{2 +} signal could no longer be followed because the fluorescence detector was in saturation (horizontal dashed line in panel F) and the Ca²⁺ waves were invading the ends of the cell outside the detection area.

Figure 8. Caffeine in low concentrations strongly enhances the I_Ca-gated Ca²⁺ release in CSQ-overexpressing mouse cardiomyocytes.

The sample frames of total fluorescence (F) were recorded at 120 Hz before (CSQ, panel A) and during (CSQ + 0.5 mM Caff, panel B) exposure to 0.5 mM caffeine during activation of I_Ca (panel C) at the times indicated by filled symbols in the tracings of the cellular fluorescence signals (panel E). Arrows in panels A and B show locations of prominent Ca²⁺ release sites; the dashed lines supposedly indicate alignment of release sites along a z-line; and the colour scale provides the means of calibration. In addition to the caffeine effect (0.5 mM Caff.) the ensemble averages in the bar graphs (panels D and F) also show the effect of washout (Wash).

Figure 2. Time course and distribution of Ca²⁺ sparks in a control cell (panels A and C) and a CSQ-overexpressing cell (panels B and D).

Panels A and B show the location of all Ca²⁺ sparks (•) recorded in 60 frames at 120 Hz over a period of 500 ms in relation to the edges of the cell (nearly vertical lines) and the probable location of z-lines (panel A, nearly horizontal lines). Panels C and D illustrate the time course of Ca²⁺ sparks plotted as ΔF/F₀, where ΔF is measured with a centre-minus-surround kernel (inset between panels B and D; see Methods), and F₀ is the average fluorescence intensity in the analysed area. In these abnormally active cells, what might appear as very intense, long lasting Ca²⁺ sparks could often be resolved as composites of several smaller Ca²⁺ sparks that were propagating primarily in the direction of the z-lines (→).

Solutions and data analysis

The solution used for cellular equilibration and formation of the gigaseal contained (mM): 137 NaCl, 5.4 KCl, 2.0 CaCl₂, 1 MgCl₂, 10 Hepes, buffered to pH 7.4 with NaOH. In experimental solutions, inward K⁺ currents were suppressed by addition of Ba²⁺ (0.2 mM) and omission of extracellular K⁺. Na⁺ currents were mostly suppressed using 3–6 μM TTX, and holding potentials (V_h) of -70 to -50 mV. We used 200 μM cAMP in all experiments except those where cells were exposed to isoproterenol (3 μM). In some experiments, elevation of [Ca²⁺]_o was achieved by adding the salt without correcting for the increase in osmolarity. Isoproterenol was used at 3 μM and caffeine at 0.5-10 mM. Patch pipettes were filled with solutions containing (mM): 110 CsCl, 20 TEA, 10 Hepes, 5 MgATP, 5 glutathione, 1 EGTA, 1 fluo-3 (Molecular Probes Inc.), with the pH adjusted to 7.3 with CsOH. All experiments were performed at room temperature (22-25°C).

Ensemble values were calculated as means ±s.e.m.

RESULTS

Transgenic mice overexpressing calsequestrin live into adulthood but develop significant cardiac dilatation and hypertrophy 30 days post partum (Knollmann et al. 1998b). Hypertrophy increases progressively with age, leading to heart failure. Isolated CSQ-overexpressing ventricular myocytes are also enlarged (∼2-fold larger membrane capacitance; Jones et al. 1998) and have longer action potentials with significant downregulation of K⁺ channels (I_to and I_K1; Knollmann et al. 1998a,b).

Spontaneous focal Ca²⁺ releases (Ca²⁺ sparks)

Figure 1 shows examples of two-dimensional confocal images of spontaneously occurring Ca²⁺ sparks in myocytes obtained from the hearts of CSQ-overexpressing mice (panel B) and their non-transgenic littermate (panel A). The two top images (F₀) show the average fluorescence intensity measured with the Ca²⁺ indicator dye fluo-3. The outline of each cell and the position of the whole-cell patch-clamp electrode (e) are clearly visible while the fading of the fluorescence signal towards the ends of the cells suggests that, following 6 min of dialysis, longitudinal equilibration of the cell with fluo-3 is still incomplete (see Cleemann et al. 1997). In Fig. 1A and B the lower three images (1, 2 and 3) are sequential frames at 30 Hz measured differentially as the increase in fluorescence (ΔF) relative to the average fluorescence (F₀). Consequently the outline of the cell and the position of the electrode are seen only as variations in the noise. On the other hand, the differential measurement aids the visual detection of Ca²⁺ sparks over the entire cell, including its edges. In control cells, spontaneous focal Ca²⁺ releases were almost always detected in normal Tyrode solution (2 mM [Ca²⁺]_o, V_h=−70 mV), appeared well defined (0.8-2.4 μm in diameter corresponding to 2σ in panel C), and occurred randomly in different regions of the cells, typically at a rate of 1–2 per frame recorded at 30 Hz, covering the major part of the cell. In sharp contrast, spontaneously occurring Ca²⁺ sparks were absent in almost all transgenic myocytes, and in rare exceptions (7 out of 49 cells monitored for 1–2 s) they were generally seen only in a few singular frames (e.g. Fig. 1B).

Recordings at 30 frames s⁻¹ were used to survey the major part of a cell typically for 2 s. In such records, each Ca²⁺ spark was seen clearly only in a single frame, and appeared, at best, as a much fainter diffuse fluorescence in the next frame consistent with a spark duration of less than 33 ms. These images were analysed by approximating individual Ca²⁺ sparks in single frames by a Gaussian distribution (inset of panel C) characterized by its normalized amplitude (F₁/F₀) and standard deviation (σ). Panel C compares the parameters of the sporadic spontaneous Ca²⁺ sparks measured in CSQ-overexpressing cells (•) to a randomly selected subset of the more numerous events in control cells (^). In terms of photons, the overall intensity of Ca²⁺ sparks (evaluated as the average of σ²F₁/F₀) was the same in control (0.91 ± 0.08 μm², n = 60 sparks) and CSQ-overexpressing cells (1.04 ± 0.10 μm², n = 55 sparks). Notice, however, that smaller (σ= 0.4-0.8 μm) but very bright (F₁/F₀= 3-7) Ca²⁺ sparks were only found in control cells (^) while very large sparks (σ > 1.2 μm) of lower intensity (F₁/F₀ < 2) appeared to be a characteristic of CSQ-overexpressing cells (•). This difference is based on a small number of observations, but it raises the possibility that Ca²⁺ sparks in CSQ-overexpressing cells tend to: (a) spread by diffusion during sustained Ca²⁺ release from a single site, (b) spread by Ca²⁺-induced Ca²⁺ releases from a primary release site to some neighbouring sites, or (c) originate from more poorly defined or deformed diadic junctions (for ultrastructural deformity, see Jones et al. 1998).

To compare the time course of spontaneous Ca²⁺ sparks in control and CSQ-overexpressing myocytes the imaging speed was increased to 120 or 240 Hz, which reduced the scanned region (Fig. 2). In such experiments we were forced to choose the rare CSQ-overexpressing myocytes that produced a larger number of spontaneous diadic Ca²⁺ releases. Considering the variability in ultrastructural and electrophysiological properties of myocytes from CSQ-overexpressing hearts, we accepted the possibility that the more active CSQ-overexpressing myocytes were somewhat atypical of the fully developed stages of cardiac impairment. To obtain a valid comparison we therefore used for comparison control cells that also had a relatively high incidence of Ca²⁺ sparks and thus ascertained that the analysed CSQ-overexpressing myocytes were obtained from hearts with established hypertrophy and exhibited the characteristic slow kinetics of inactivation of I_Ca and Ca²⁺ transients (see Fig. 3). A computerized algorithm was used to locate the Ca²⁺ sparks and follow their development (Cleemann et al. 1998). The spatial resolution of these measurements was determined by a detection kernel, which measured the average fluorescence intensity over a disc with 0.8 μm radius relative to a surrounding ring of 1–1.5 μm radius. The selected traces of Fig. 2 suggest that Ca²⁺ sparks were of similar amplitude and duration in control (panels A and C, traces a–m) and in CSQ-overexpressing myocytes (panels B and D, traces a–j). The duration of focal Ca²⁺ release (15-20 ms) was somewhat briefer than that observed in most other studies (Cannell et al. 1994, 1995), probably because we have used both a centre-minus-surround kernel, and 1 mM EGTA as a secondary non-fluorescent Ca²⁺ buffer (for detailed analysis and rationale of this approach, see Cleemann et al. 1998). The 2-D confocal measurements made it possible to distinguish isolated focal Ca²⁺ releases from compound releases that spread slowly in the direction of the z-lines in control cells (Fig. 2A and C, arrows) and collectively lasted longer than the spatially confined releases. Such propagated releases were also seen occasionally in CSQ-overexpressing cells (Fig. 2B and D, traces e and f and h and i), but we found no indication that this mechanism significantly enhances the overall level of spontaneously released Ca²⁺. The cells illustrated in Fig. 2 had unusually high levels of spontaneous Ca²⁺ release, but the properties of each Ca²⁺ spark were similar to those typically observed in less active cells. When analysis was limited to Ca²⁺ sparks that showed no sign of propagation, and the analysis was performed with identical settings (detection threshold, dimensions of detection kernel, Cleemann et al. 1998), the average durations of sparks in control (17.3 ± 2.8 ms, n = 4 cells) and CSQ-overexpressing myocytes (17.6 ± 2.0 ms, n = 8 cells) were comparable.

Confocal images of I_Ca-gated Ca²⁺ release

Figure 3 compares the Ca²⁺ current and the profiles of rise of intracellular Ca²⁺ in a control and CSQ-overexpressing myocyte. Two-dimensional confocal images of Ca²⁺ obtained at 4.17 ms intervals (240 frames s⁻¹) shows that activation of I_Ca in the control myocyte leads to the development of sparks (arrows in panel A, frame 5; first filled circle in panel F) that appear as bright spots which tend to fuse into a sarcomeric Ca²⁺ striping pattern in later frames (Fig. 3A, frames 6 and 7; cf. Cleemann et al. 1998). This pattern of focal Ca²⁺ release generates rapidly activating and coordinated Ca²⁺ release (Fig. 3F). In CSQ-overexpressing myocytes, on the other hand, Ca²⁺ sparks do not develop into the characteristic sarcomeric Ca²⁺ stripes (Fig. 3B). Instead, large diffuse areas of Ca²⁺ fluorescence with embedded Ca²⁺ sparks (Fig. 3B, arrows) can be consistently observed during the slowly rising cytosolic Ca²⁺ transients (Fig. 3G). Note that the slow rise in cytosolic Ca²⁺ transients is accompanied by much slower inactivation of I_Ca (Fig. 3D), suggesting impaired Ca²⁺ signalling in CSQ-overexpressing myocytes (Adachi-Akahane et al. 1996; Jones et al. 1998).

The bar charts in Fig. 3 document the significance of these findings based on pooled data from a number of control and transgenic myocytes. Panel H compares the kinetics of Ca²⁺ transients and shows that the time-to-peak (TTP, see panel F) and the half-time (t^1/2) of its decay are much longer in CSQ-overexpressing myocytes, consistent with slower rise and fall of global cytosolic Ca²⁺. Similarly, panel E shows that the half-time of inactivation of I_Ca (t^1/2, see Fig. 3C) in transgenic myocytes was prolonged more than 2-fold compared to the control myocytes. It is likely that the absence of coordinated and rapid Ca²⁺ release in CSQ-overexpressing myocytes is mostly responsible for the slower inactivation of I_Ca in transgenic mice, as has been described for Ca²⁺ currents in rat ventricular myocytes (Adachi-Akahane et al. 1996). In this analysis the half-time, rather than exponential analysis, was used to approximate the inactivation kinetics of the channel, as relatively brief 50–60 ms pulses were used in confocal imaging experiments to limit photo-damage caused by intense laser light.

Enhancement of I_Ca is not sufficient to restore Ca²⁺ signalling

Since the degree of activation of RyRs in part depends on a rapid rise of Ca²⁺ in the microenvironment of the Ca²⁺ channel-RyR complexes, we examined whether increasing the influx of Ca²⁺ through the channel would restore the impaired Ca²⁺ signalling in CSQ-overexpressing myocytes. To ensure that the Ca²⁺ content of the SR remained constant when 2 or 10 mM Ca²⁺ was used as charge carrier through the Ca²⁺ channel, a sequence of five conditioning depolarizing pulses in 2 mM Ca²⁺ were applied prior to the rapid and short (0.5 s) application of the 10 mM Ca²⁺ solution. Figure 4 shows the effect of elevation of extracellular Ca²⁺ from 2 to 10 mM on I_Ca (panels C and D), global cellular Ca²⁺ transients (panel E and F) and focal Ca²⁺ releases (panels A and B). Elevation of Ca²⁺ enhanced I_Ca (panel D, 32 ± 9 %, mean ±s.e.m., n = 12), increased the cellular Ca²⁺ transients (ΔF/F₀ in panel F, by 100 ± 9 %, mean ±s.e.m., n = 12), and decreased the half-time of inactivation of I_Ca (t^1/2, panel D, 28 ± 7 %, mean ±s.e.m., n = 12); but the sporadic pattern of focal Ca²⁺ releases and their size, duration and uncoordinated nature remained unchanged.

Figure 5 shows that isoproterenol was somewhat more effective than elevation of Ca²⁺ in restoring the coordinated Ca²⁺ release, whether monitored as confocal images (frames of panels A and B) or as global intracellular Ca²⁺ transients (panels E and F). In the presence of isoproterenol, cytosolic Ca²⁺ transients were increased more than 2-fold (133 ± 35 %, n = 9) as the resting Ca²⁺ signal increased by 65 ± 5 % (mean ±s.e.m., n = 9, F_r/F₀ in panel F). Isoproterenol did not reduce the time-to-peak of global Ca²⁺ transients (panel E) or accelerate the inactivation kinetics of I_Ca significantly (panel D), consistent with the idea that coordinated Ca²⁺ release was not fully re-established. The increase in the number of Ca²⁺ sparks detected shortly after activation of I_Ca (frames 7–9 of panels A and B) appeared to be roughly proportional to the increase in the cellular Ca²⁺ transients (panel E), and occurred without significant changes in the size of the individual Ca²⁺ sparks. In this context, note that strong, long lasting focal Ca²⁺ releases, of the type seen near the edge of the cell in the lower right corner in each frame of both panels A and B (arrows), were detected both in the absence and presence of isoproterenol, but were too rare to contribute significantly to the global Ca²⁺ transient.

Caffeine improves the efficacy of Ca²⁺ release in CSQ-overexpressing myocytes

It has been previously reported that caffeine-releasable Ca²⁺ stores in CSQ-overexpressing myocytes were markedly enhanced even though I_Ca-gated Ca²⁺ release was impaired (Jones et al. 1998). In the next series of experiments, we attempted to examine the effects of high and low concentrations of caffeine on direct or I_Ca-gated Ca²⁺ release in CSQ-overexpressing transgenic myocytes. Figure 6 compares the effect of a rapid and short application (‘puff’) of 10 mM caffeine in control and transgenic myocytes. In control cells dialysed with 1 mM fluo-3 and 1 mM EGTA, rapid application of 5–10 mM caffeine triggered a rapid release of Ca²⁺, the time course of which decayed slowly in the presence of caffeine (panel A, frames 13-70; panel E). Note that the rise in global Ca²⁺ concentrations activated only a brief and small Na⁺-Ca²⁺ exchange current (I_Na-Ca; Callewaert et al. 1989), in part because cytosolic Ca²⁺ is well buffered by 1 mM EGTA and 1 mM fluo-3 (panel C; see also Adachi-Akahane et al. 1996). In sharp contrast, in similarly Ca²⁺-buffered CSQ-overexpressing myocytes, the caffeine-induced Ca²⁺ release was so large as not only to activate a large I_Na-Ca (panel D), but also to fully saturate the fluorescent dye signal (panel F, dashed line). These findings are consistent with the global measurements of Ca²⁺ in CSQ-overexpressing myocytes previously reported using high concentrations of fura-2 as Ca²⁺ buffer (Jones et al. 1998). Confocal images recorded at 30 Hz show that the pattern of caffeine-induced Ca²⁺ release was noticeably different in control (panel A) and transgenic (panel B) myocytes. In control cells, local Ca²⁺ releases were initiated at many sites, first close to the surface of the cell (panel A, frames 13 and 15) then at its centre (frames 17-21), and were followed by a more homogeneous rise in cytosolic Ca²⁺ that faded rapidly (frames 23-27), except for the delayed response from the nuclear region (frames 27 and 70). In contrast, the caffeine-induced Ca²⁺ signal in transgenic cells was typically initiated at only a few sites near the cell surface (panel B, frames 15 and 17), and then appeared to spread as a wave with some focal releases ahead of the front to the entire cell (frames 19 and 21), saturating the detector (frames 23 and 27) and causing noticeable cell shortening (frame 70), often irreversibly. Thus, it appears that the large Ca²⁺ release triggered by caffeine in CSQ-overexpressing cells may activate propagating Ca²⁺ waves (cf. Cheng et al. 1996).

To explore the effect of caffeine on spontaneous or Ca²⁺ channel-gated Ca²⁺ release in transgenic myocytes with intact Ca²⁺ stores, we tested the effects of lower concentrations of caffeine (0.5 mM) on focal and global Ca²⁺ releases. Figure 7A (frames 1 and 2) shows the spontaneous occurrence of a number of Ca²⁺ sparks in a CSQ-overexpressing myocyte clamped to -70 mV following 1 min exposure to 0.5 mM caffeine (cf. Fig. 1). This effect was observed in 20 out of 24 quiescent CSQ-overexpressing myocytes. In comparison, incubation of myocytes in 1 μM isoproterenol induced spontaneous Ca²⁺ sparks in 9 out of 13 quiescent CSQ-overexpressing myocytes. Figure 7 also shows the distribution (panel B) and time course of development of Ca²⁺ sparks in solution containing 0.5 mM caffeine (panel C, traces a–i). These focal Ca²⁺ releases occurred for the duration of caffeine exposure (seconds to minutes), were markedly suppressed or absent upon washout of caffeine, and had intensities and durations similar to those seen in non-transgenic myocytes in the absence of caffeine (cf. Fig. 2).

Figure 8 compares I_Ca-gated Ca²⁺ transients and focal Ca²⁺ releases in the same transgenic myocyte in the absence and presence of 0.5 mM caffeine. Consistent with the findings of Fig. 3, depolarization of CSQ-overexpressing myocytes to 0 mV activated a slowly inactivating Ca²⁺ current (panel C, CSQ) which failed to release significant coordinated focal Ca²⁺ releases (panel A, frames 4-6) or global Ca²⁺ transients (panel E, CSQ). In the presence of 0.5 mM caffeine, however, global Ca²⁺ releases (panel E, CSQ + 0.5 mM Caff) recovered in response to activation of I_Ca. However, although some alignment in Ca²⁺ release sites was occasionally observed (dashed lines in panel A, frame 5 and panel B, frame 4), the characteristic Ca²⁺ striping pattern seen in control cells did not fully develop (panel B). The cellular response of CSQ-overexpressing myocytes in the presence of caffeine appeared to be composed of a large number of ‘woolly’ sparks, which were resolved most clearly shortly after depolarization or at the edges of the cells (panel B, arrows in frames 4 and 5), as they tended to fuse rapidly in the cell interior. It should be noted that the slow inactivation kinetics of I_Ca in CSQ-overexpressing myocytes (t^1/2, panel D) were also significantly enhanced in the presence of 0.5 mM caffeine (panel C), consistent with the restoration of more effective Ca²⁺ release. Caffeine, however, had little or no effect on the peak magnitude of the Ca²⁺ current (Fig. 8D). The 3- to 4-fold enhancement of the cellular Ca²⁺ release induced by 0.5 mM caffeine was completely reversible (ΔF/F₀, panel F) after washout of caffeine. The smaller increase in resting Ca²⁺ signal observed in the presence of caffeine, however, was not reversible (F_r/F₀, panel F), suggesting that photoinactivation, in part, contributes to the rise in the background signal.

In a few cells, where we used 1 and 2 mM concentrations of caffeine, we found that 1.0 mM caffeine consistently increased I_Ca and similarly enhanced the global Ca²⁺ transients. Caffeine at a concentration of 2 mM or higher, on the other hand, often induced partial Ca²⁺ release with transient suppression and then steady-state potentiation of I_Ca (data not shown).

DISCUSSION

The present data indicate that the impairment of Ca²⁺ release in myocytes overexpressing cardiac CSQ is associated with a reduced frequency of spontaneous or Ca²⁺ channel-gated Ca²⁺ sparks and the disappearance of the coordinated Ca²⁺ release which generates the sarcomeric Ca²⁺ striping pattern in non-transgenic cardiomyocytes. Effective Ca²⁺ signalling could be restored by low (0.5 mM) concentrations of caffeine, but not by agents that enhance I_Ca (isoproterenol and high [Ca²⁺]_o), suggesting that impairment of I_Ca-gated Ca²⁺ release in CSQ-overexpressing myocytes may result from decreased Ca²⁺ sensitivity of the ryanodine receptors or their number in the diadic junctions.

Impaired Ca²⁺ release in CSQ-overexpressing transgenic myocytes

Although the caffeine-releasable Ca²⁺ pools in CSQ-overexpressing myocytes were enhanced (Fig. 6) 10- to 15-fold (Jones et al. 1998), the Ca²⁺ channel-gated Ca²⁺ release was small and uncoordinated (Fig. 3). We considered the possibilities that this reduction might occur because the unitary events associated with diadic Ca²⁺ release were either: (a) smaller in amplitude, (b) smaller in dimensions, or (c) fewer in number. To distinguish between these possibilities we studied spontaneous Ca²⁺ releases in control solutions and found that they occurred rarely in transgenic myocytes (Fig. 1A and B), but when they did sporadically occur, had the same magnitude as those of control cells (Fig. 1C). We cannot exclude the possibility that the release of Ca²⁺ in CSQ-overexpressing myocytes is distributed over a somewhat larger area (Fig. 1C), but found little evidence that this might result from the prolongation of the release process and the subsequent dispersion by diffusion. The measured duration (15-20 ms) of focal Ca²⁺ release in control and transgenic myocytes was somewhat longer than the value obtained by similar 2-D confocal measurements in rat ventricular myocytes (Cleemann et al. 1998), but was considerably shorter than line-scan estimates of the duration of Ca²⁺ sparks (Cheng et al. 1993; Cannell et al. 1994, 1995; Lopez-Lopez et al. 1995). It remains to be determined if this is due to differences in technique (target size (inset of Fig. 2), use of Ca²⁺ buffers, detection thresholds, scan mode, etc.) or reflects species-dependent variations.

Since large (5-10 mM) concentrations of caffeine produced propagated Ca²⁺ waves in CSQ-overexpressing cells, but not in control myocytes (Fig. 6A and B), we considered whether a similar mechanism might also contribute to the modulation of focal Ca²⁺ releases. We found that spontaneous Ca²⁺ releases sometimes showed saltatory propagation along z-lines, but this mechanism was not more pronounced in CSQ-overexpressing than in control myocytes (Fig. 2). It is more likely that propagated waves represent cells that are unusually active or Ca²⁺ overloaded.

The characteristic Ca²⁺ striping pattern of the I_Ca-activated Ca²⁺ release pattern in control cells (Fig. 3A) was absent (Figs 3B, 4 and 5) or only partially apparent (Fig. 8A and B, dashed lines) in transgenic myocytes. The analysis of spontaneous Ca²⁺ sparks and the patterns of Ca²⁺ release triggered by I_Ca indicate that the CSQ-overexpressing cells are capable of producing focal Ca²⁺ releases similar to those seen in control cells but with much lower frequency. This impairment may be, in part, related to downregulation of the expression of ryanodine receptors (Jones et al. 1998). In addition, ultrastructural distortion previously reported (Jones et al. 1998) may also alter the strategic location of RyRs, thereby diminishing their ability to sense influx of Ca²⁺ through the Ca²⁺ channel and trigger regenerative diadic Ca²⁺ releases. While transgenic cells show a spectrum of properties depending on the progression of hypertrophy with age (Knollmann et al. 1998a), b it should be noted that the cells selected for detailed spark analysis came from hearts with established hypertrophy and characteristic slow Ca²⁺ kinetics of I_Ca and Ca²⁺ transients (Fig. 3).

Ca²⁺ channel activity in transgenic mice

The most prominent change in the biophysical properties of the Ca²⁺ channel in CSQ-overexpressing transgenic myocytes is its slow inactivation compared to control mice (Figs 3D, 4C, 5C and 8C) such that t^1/2 is increased from 8 to 24 ms (Fig. 3E). Such slow kinetics were first described by Jones et al. (1998) and were confirmed by Sato et al. (1998) in another murine CSQ-overexpressing transgenic strain. This effect appears to be unrelated to structural properties of the channel, as the kinetics of Ba²⁺ current through the Ca²⁺ channel of control and transgenic myocytes remained similar (data not shown). Furthermore, since the peak value of I_Ca shows little (< 25 %) or no change, the marked decrease in the rate of inactivation could not have resulted from a decrease in the influx of Ca²⁺ through the channel. We conclude therefore that the marked slowing of the kinetics of inactivation of I_Ca in CSQ-overexpressing myocytes of transgenic mice occurs secondary to the impaired Ca²⁺ release from the ryanodine receptor (Figs 3, 4, 5 and 8; see also Jones et al. 1998). Such Ca²⁺ cross-talk has been quantified in rat ventricular myocytes where Ca²⁺ release appears to determine 60–70 % of the Ca²⁺-induced inactivation of the Ca²⁺ channel (Lee et al. 1985; Adachi-Akahane et al. 1996).

Recovery of I_Ca-gated Ca²⁺ release

Low doses of caffeine (0.5 mM) largely restored the ability of CSQ-overexpressing cells to produce both spontaneous Ca²⁺ sparks at rest (cf. Figs 1 and 7), and I_Ca-gated cellular Ca²⁺ transients (cf. Figs 3 and 8). In experiments with sufficient signal-to-noise ratio to resolve individual Ca²⁺ sparks, it was consistently observed that augmentation of cellular Ca²⁺ transients was accompanied, primarily, by an increased number of Ca²⁺ sparks (Figs 4 and 5) which, in the presence of 0.5 mM caffeine, tended to fuse together (Fig. 8). Individual Ca²⁺ sparks, under these conditions, were seen clearly only near the edges of the cell (Fig. 8). Considering also that the more clearly resolved spontaneous Ca²⁺ releases in transgenic cells are similar to those in control cells with respect to their amplitude, duration and confinement, it is likely that the major effect of caffeine is to increase the number of focal Ca²⁺ releases.

The large, almost 2- to 2.5-fold, increase in the myocyte surface area noted earlier (Jones et al. 1998) and confirmed here in CSQ-overexpressing myocytes may, in part, distort the ultrastructural organization of the DHP-ryanodine receptor complex making the Ca²⁺ release mechanism, which requires an organized dyadic microdomain, less efficient. This would especially be true if the level of expression of ryanodine receptor were depressed, as reported by Jones et al. (1998). In transgenic mice overexpressing murine cardiac calsequestrin, Sato et al. (1998) report no decrease in the level of cardiac ryanodine receptors or the associated junctional proteins junctin or triadin, but do report strongly modified kinetics of I_Ca, consistent with the compromised I_Ca-gated Ca²⁺ release. These authors did, however, find that the ryanodine-associated FKB12 protein was suppressed. Removal of FKB12 protein in vitro has been reported to be associated with development of multiple substate openings and ‘chaotic’ behaviour of the ryanodine receptor (Brillantes et al. 1994; Marks, 1996). Our 2-D confocal imaging, however, failed to confirm significant focal Ca²⁺ sparks at rest in CSQ-overexpressing myocytes (Fig. 1). Discrete sparks associated with the activation of Ca²⁺ channels were also rare and appeared to be replaced by fairly disorganized sparks which may be responsible for the slow and uncoordinated global intracellular Ca²⁺ rise in CSQ-overexpressing myocytes. Although ultrastructural studies might suggest distortions in dyadic junctions with SR (Jones et al. 1998), the finding that low concentrations of caffeine re-establish Ca²⁺ signalling is not quite consistent with major ultrastructural changes in the DHP-RyR complex or its immediate microenvironment; rather it may reflect increased sensitivity of ryanodine receptors to Ca²⁺ (Rousseau & Meissner, 1989; Sitsapesan & Williams, 1990). An alternative possibility for the impairment of Ca²⁺ signalling is that high levels of luminal Ca²⁺ may directly, or via the junctional protein, reduce the sensitivity of ryanodine receptor to Ca²⁺ or Ca²⁺ signalling via the Ca²⁺ channel. Thus, the impairment in Ca²⁺ signalling may be related primarily to functional interaction of Ca²⁺ channels with ryanodine receptors, rather than distortions in the ultrastructural architecture of the DHP-RyR complex.

Enhancement of I_Ca by elevation of [Ca²⁺]_o and isoproterenol was always less effective than low doses of caffeine in restoring I_Ca-gated Ca²⁺ release and producing only modest enhancements of the cellular Ca²⁺ transients (Figs 4F and 5F), which generally were related, but not strictly proportional, to changes in the magnitude and rate of inactivation of I_Ca (Figs 4D and 5D). Details of the observed responses may therefore reflect the fact that Ca²⁺ release is sensitive to changes in unitary Ca²⁺ current (Santana et al. 1996) and that inactivation of I_Ca is modulated not only by Ca²⁺ release, but also by screening charge effects of divalent cations, and modal shifts in the gating of I_Ca (McDonald et al. 1994). The variable effect of isoproterenol on recovery of Ca²⁺ signalling (data not shown) may reflect progression of the disease with age (Knollmann et al. 1998a) and the unexpected finding that enhancement of I_Ca by isoproterenol is pharmacologically different from that in control cells suggests a higher activity of phosphodiestrases in CSQ-overexpressing myocytes (Zhang et al. 1997; Knollmann et al. 1998b). Irrespective of these variations with age and experimental conditions, it seems unlikely that the β-adrenergic pathway can restore effective Ca²⁺ signalling in myocytes from CSQ-overexpressing mice with established cardiac hypertrophy.

The enhancements of the I_Ca-triggered Ca²⁺ releases by 10 mM [Ca²⁺]_o, 3 μM isoproterenol, and, especially, 0.5 mM caffeine were paralleled at rest by changes in the basal Ca²⁺ signal (F_r/F₀, Figs 4F, 5F and 8F). This increase in resting signal failed to return fully to baseline after washout of these inotropic agents (e.g. caffeine, Fig. 8F), suggesting that: (1) the effect of caffeine on [Ca]_i is long lasting, (2) resting cytosolic Ca²⁺ concentrations are gradually increasing (Cleemann & Morad, 1991), (3) dialysis of fluo-3 is still in progress at the time of measurement, or (4) photo-damaged fluo-3, though producing an increase in cellular fluorescence, is insensitive to Ca²⁺ (Lipp et al. 1996). It is likely that the apparent effect of caffeine on basal [Ca²⁺]_i (Fig. 8E and F) is real since it is much larger than the effects observed with elevation of [Ca²⁺]_o (Fig. 4E and F) and isoproterenol (Fig. 5E and F). To distinguish clearly between the different possibilities and examine the physiological consequences of this finding it would be useful to perform accurately calibrated ratiometric measurements using, for example, indo-1 or fura-2 as Ca²⁺ probes.

Animal models of cardiac hypertrophy and impaired Ca²⁺ signalling

The pathophysiology of impaired Ca²⁺ signalling, cardiac hypertrophy, and heart failure seen in CSQ-overexpressing mice is distinctly different from the phenotype associated with other genetically modified animal models, including mice overexpressing the Na⁺-Ca²⁺ exchanger, Ca²⁺-ATPase, phospholamban, or the α_1c subunit of the Ca²⁺ channel, as well as spontaneously hypertensive rats, and salt-sensitive mice. Overexpression of the Na⁺-Ca²⁺ exchanger (Adachi-Akahane et al. 1997; Terracciano et al. 1998), Ca²⁺-ATPase (He et al. 1997; Yao et al. 1998), and phospholamban (Chu et al. 1997; Masaki et al. 1998), or the knockout of phospholamban (Masaki et al. 1997) produced no change in myocyte or heart size and did not impair Ca²⁺ signalling. The overexpression of the α_1c subunit of the Ca²⁺ channel produced a small (10-20 %) increase in cell size, but did not alter Ca²⁺ release (Muth et al. 1999). The overexpression of Ca²⁺-ATPase or phospholamban and knockout of phospholamban appeared to regulate Ca²⁺ signalling only by altering the Ca²⁺ content of the SR. For instance, the increased SR Ca²⁺ content of phospholamban-deficient cardiomyocytes was reported to be directly related to an increased frequency of nearly unchanged Ca²⁺ sparks (Santana et al. 1997) and could be normalized by decreasing the Ca²⁺ content of the SR. Thus up- or downregulation of the Ca²⁺ content is not sufficient by itself to produce impairment of Ca²⁺ signalling or cardiac hypertrophy. In comparison, cells from hypertrophied hearts of spontaneously hypertensive rats produced larger Ca²⁺ sparks without a change in the SR Ca²⁺ content (Shorofsky et al. 1998).

Overexpression of CSQ and the α_1c subunit of the Ca²⁺ channel was unexpectedly accompanied by impaired β-adrenergic signalling (Zhang et al. 1997; Muth et al. 1999). Whether the defect in β-adrenergic signalling is a prerequisite for cardiac hypertrophy and failure, as previously suggested (Cho et al. 1999), remains to be determined. The impaired Ca²⁺ signalling of CSQ-overexpressing myocytes did not recover significantly in the presence of β-adrenergic agonists, as was reported for those of salt-sensitive mice in heart failure (Gomez et al. 1997). It might be of some interest to see if a modulation of Ca²⁺ signalling by caffeine is as effective in other models of heart failure as it is in murine CSQ-overexpressing myocytes. Although CSQ levels do not appear to change with cardiac disease in man, it is of interest that CSQ overexpression in mice exhibits the syndrome of hypertrophy and electrical remodelling almost identical to that found in human heart failure.