INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ATP-SENSITIVE K⁺ (K_ATP) channels are found in many organs, such as the heart (14), brain (20), pancreas (21), and blood vessels (30). They play a prominent role in the protection against oxygen deficiency (1), in insulin secretion (21), and in the regulation of blood flow (6), respectively. At the molecular level, they consist of a tetrameric K⁺ channel pore [K⁺ inward rectifier (Kir)6.1 or Kir6.2] surrounded by a regulatory tetramer called the sulfonylurea receptor (1, 32). K_ATP channels were thought to be localized exclusively in the plasmalemma and, when activated by hypoxia or ischemia, to shorten action potential duration in the cardiac myocytes, reduce calcium influx, and thus reduce cardiac work (1). The reduced calcium load should reduce the possibilities of calcium-induced cell injury (26).

Several lines of evidence suggest an additional hypothesis by implicating mitochondrial K_ATP channels. First, immunoelectron microscopy suggests that Kir6.1- like subunits are also found on the mitochondrial inner membrane (34), although the cDNA coding for mitochondrial Kir is probably neither Kir6.1 nor Kir6.2 (33) and has not yet been cloned. Second, in ischemic preconditioning, the shortening of the action potential duration is not required for amelioration of contractile dysfunction of the canine heart (11), although it is essential in mice (35). Ischemic preconditioning is a condition where a prior ischemic episode protects the heart against subsequent ischemic damage (9). Third, the mitochondrial K_ATP channel blocker 5-hydroxydecanoate (5-HD) can abolish ischemic preconditioning in some studies (27), although not in others (35), whereas mitochondrial K_ATP channel openers such as pinacidil and diazoxide (13) can mimic ischemic preconditioning (2). These arguments for a role of mitochondrial K_ATP channels are challenged by the possibility that diazoxide may not target mitochondrial K_ATP channels (27, 28) but inhibit mitochondrial respiration or oxidant stress at reoxygenation.

Cytoplasmic calcium plays a crucial role during ischemia in both myocyte contraction (10) and cell injury (26), although the link between activation of mitochondrial K_ATP channels and cytoplasmic calcium is still unclear. Could opening of mitochondrial K_ATP channels alter mitochondrial calcium uptake and release (12)? Could other intracellular K_ATP channels perhaps alter calcium handling by the sarcoplasmic reticulum? Either situation might change the cytoplasmic calcium baseline or calcium transients and thus modify the risk of cell injury.

Most work on K_ATP channels and calcium homeostasis has been performed on ventricular myocytes. However, dysfunction of the cardiac atrium is the cause of 50% of all cardiac deaths in Western societies (19). It is not known whether atrial function can be deduced entirely from studies on the ventricle. Therefore, the aim here was to determine whether plasmalemmal or mitochondrial K_ATP channels, or both, regulate calcium baseline and calcium transients in metabolically compromised but otherwise intact atrial myocytes.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Atrial myocyte culture. Myocytes from the atrial appendages of neonate rats (2- to 3-day-old Sprague-Dawley rats) were dispersed and cultured on 22-mm glass slides as previously described (3, 17). Experiments were performed after 1 day and up to 3 days in culture.

Cytoplasmic calcium measurements. The myocytes were exposed for 5-7 min to the cell-permeant calcium indicator fura 2-AM (1 µmol/l) with pluronic acid (1 µl/ml) (both from Molecular Probes) and then placed on the experimental setup in an incubation chamber containing 100 µl HEPES buffer at 30°C. The buffer and setup have been described previously (3). The incubation chamber was perfused at 2 ml/min for drug application (see Protocols) while the preparation was illuminated (40 times/s) with a monochromator (PTI) at wavelengths of 340 and 380 nm. The emitted light was collected through a ×100 fluorescence oil objective (numerical aperture 1.3, Nikon), filtered by a 515 ± 20-nm bandpass filter, directed through an adjustable window overlying the region of interest, and measured with a photomultiplier. The ratio of light emitted during the 340- and 380-nm excitation was calculated on-line by FELIX software (DeltaRam, PTI) and was monotonically related to the free calcium level. Calibrations were performed with fura 2-free acid (Molecular Probes) at free calcium concentrations of 0, 0.1, 0.225, 0.602, 1.35, and 39.8 µmol/l and yielded ratios of 0.488, 0.787, 0.964, 1.345, 1.611, and 2.090, respectively. Calibrations for zero calcium were also performed on the myocytes by perfusion of EGTA-buffered solution containing the calcium ionophore ionomycin (20 µmol/l) (Alexis Biomedicals) and 50 µg/ml palmitoyl lysophospatidyl choline (Sigma) and yielded ratios of 0.543 ± 0.026, thus at most 18 nmol/l. A bipolar fine-tipped stimulation electrode was lowered in the vicinity of the myocyte to trigger calcium transients at 1 Hz. The ratios were displayed on-line (40 samples/s) and stored together with the trigger signal on a personal computer. The myocytes were selected if they contracted in synchrony with the trigger signal. Contractions of the myocytes could not be observed during calcium measurements and were not monitored in this study.

Protocols. Experiments were performed on a total of 89 preparations. In 14 preparations, two control periods were followed by two test periods with either 100 µmol/l diazoxide or 100 µmol/l tolbutamide and by two recovery periods, as shown in Fig. 2. In 44 preparations, the two control periods were followed by six test periods with the uncoupling protonophore CCCP (100, 200, or 400 nmol/l); 5-HD (500 µmol/l) was applied during the third and fourth test period, as shown in Fig. 3. This protocol allowed us to compare, within the same group as well as between groups, the effects of CCCP and 5-HD and was adopted to maximize the chances of observing subtle effects of 5-HD. In 31 preparations, the control periods were followed by six test periods with 400 nmol/l CCCP plus either HMR-1098 (30 µmol/l), a cell-impermeant plasmalemmal K_ATP channel blocker (21), glibenclamide (1 µmol/l), or diazoxide (500 µmol/l) and by two recovery periods, as shown in Fig. 5. With this protocol, the drug effects are tested by comparison with CCCP alone, and drug effects on post-CCCP recovery can also be tested. To minimize bleaching of the preparation, the myocytes were illuminated during the first 20 s of each 2-min period, with experiments lasting 10.3 or 18.3 min. Diazoxide, tolbutamide, glibenclamide, and CCCP were from Sigma, 5-HD was from ICN Biomedicals, and HMR-1098 was from Aventis Pharma Deutschland. Similar concentrations were used previously in our studies (4, 17) or by others (e.g., Ref. 9). A concentration of 100-400 nmol/l CCCP was used to partially mimic a mild metabolic impairment, as verified below.

Electrophysiological controls. Seven atrial myocytes were used in perforated patch-clamp experiments for monitoring the whole cell plasmalemmal K_ATP current. The methods were similar to those described previously (4) except that the membrane was not ruptured while in the cell-attached mode and the pipette solution was supplemented with 5 µg/ml gramicidin (Sigma). This method leaves the cell largely intact (36) and allows for monitoring the K_ATP current-membrane potential curves during drug application.

Data analysis. Calcium transients (ratios) were averaged over 20 s for each interval and each preparation, averaged over all preparations of the same group and same interval, and displayed as a series of consecutive mean signals. The means ± SE were displayed when appropriate, as shown by the example for one myocyte shown in Fig. 2C and for N myocytes in Fig. 2D. For each preparation, the mean ratios were also analyzed for their baseline level, the time constant for the decrease from the peak level (fitted by a first-order exponential), and amplitude (equal to the maximum minus minimum ratio). The amplitudes shown in the Tables 1 and 2 were normalized relative to the mean controls. The parameters were averaged over each group and each interval and were analyzed for statistically significant differences between groups and times by nonparametric ANOVA for repeated measures (SAS Institute; Carey, NC).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Electrophysiological recordings showed that 400 nmol/l CCCP modestly increased the K_ATP current compared with 1,000 nmol/l CCCP. The examples shown in Fig. 1, A and B, show the protocol used. Figure 1C shows, for three cells, over the entire recording period, the relationship between K_ATP current measured at 0 mV and the membrane potential measured at 0 pA, indicating that extremely small increases in membrane current can hyperpolarize the myocytes. Figure 1D shows the mean increase in K_ATP current density ± SE in response to 400 and 1,000 nmol/l CCCP and the strong inhibition by HMR-1098. In contrast, application of 500 µmol/l 5-HD had no effect on basal or CCCP-activated K_ATP current (results not shown).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Electrophysiological controls with gramicidin perforated-patch recordings. A: imposed membrane potential ramp (top) and recorded ATP-sensitive K+ (KATP) current traces (bottom) for baseline (B), 400 nmol/l CCCP (C400), 1,000 nmol/l CCP at peak (C1000), 1,000 nmol/l CCCP at steady state (C1000s), and coadded 30 µmol/l HMR-1098 (HMR). B: corresponding KATP current measured at 0 mV (top) and membrane potential measured at 0 pA (bottom). C: potential-current relationships superimposed for 3 atrial myocytes and measured over all recording periods. D: mean increase in KATP current density (±SE) over baseline for these myocytes, as normalized by the cell capacitance. * P < 0.05 relative to baseline; cP < 0.05 relative to 400 nmol/l CCCP; hP < 0.05 relative to 1,000 CCCP. CCCP (400 nmol/l) increased current density in all 3 cells tested.

Calcium transients were measured, at first, in the absence of CCCP. As an example, Fig. 2A shows the raw signals (ratio) recorded in one atrial myocyte over 20-s periods during baseline control, diazoxide exposure, and recovery. Figure 2, B-D, illustrates on an expanded time scale the first four calcium transients of the same cell (B), the time-averaged means over 20 s from that myocyte (C), and finally the group-averaged calcium transients for the diazoxide (D, left) and tolbutamide (D, right) experiments. Neither diazoxide nor tolbutamide caused significant changes in the calcium transients.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Procedure for recording and averaging the fura 2 emission ratios and experiments with diazoxide (D) and tolbutamide (T) (both 100 µmol/l). A: raw signals (ratios) from one myocyte over three 20-s periods. B: corresponding superimposed first 4 calcium transients. C: signal averages over 20 s ± SE. D: group-averaged signals ± SE for N myocytes. C, baseline control; R, recovery. Note that each averaged ratio transient spans a 0.9-s interval, starting at the time indicated (in min) below. There were no significant group differences in the normalized amplitude, baseline ratio, or recovery time constant.

CCCP (100-400 nmol/l) caused a concentration-dependent decrease in amplitude during the first and second test period, as shown by the group-averaged traces shown in Fig. 3, A, D, E, G, and H (SE not shown for clarity). The amplitude then slowly recovered despite the continuous application of CCCP (Fig. 3, E and H) and further recovered, albeit partially, after withdrawal of CCCP. Application of the mitochondrial K_ATP channel blocker 5-HD (500 µmol/l) during the third and fourth test period to 200-400 nmol/l CCCP had no significant effects relative to CCCP alone, as shown by the superposition of the traces for the 5-HD test and the vehicle time control (Fig. 3, F and I). Comparison of the traces for 100 nmol/l CCCP plus 5-HD and vehicle (Fig. 3, A-C) showed a slow increase in baseline for CCCP. This increase was most marked for 400 nmol/l CCCP (Fig. 3, G and H). Note also that the vehicle alone (ethanol) and/or time caused a small and only partially reversible decrease in the amplitude of the calcium transient. The data were further analyzed by calculating the change in ratio over baseline, group averaged, and displayed as a superposition of tests and vehicle time controls with the means ± SE (Fig. 4). All test traces for 200-400 nmol/l CCCP perfectly superposed on their respective controls. A slight difference only was observed in the comparison of 100 nmol/l CCCP plus 5-HD relative to the vehicle minus 5-HD (Fig. 4, B and C). Table 1 shows a detailed statistical analysis for the normalized amplitude, baseline, and time constant. For the 200-400 nmol/l CCCP experiments, only one comparison showed a small significant difference in normalized amplitude after 5-HD (53.4% vs. 64.2%); one comparison showed a small significant difference in the time constant (162.3 vs. 123.8 ms), whereas there were no significant differences in baseline. For the 100 nmol/l CCCP + 5-HD vs. vehicle comparison, the significance of the larger time constants with 5-HD was due to the larger time constants in the controls before the application of 5-HD.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent effect of the protonophore CCCP on calcium transients in atrial myocytes and the lack of effect of the mitochondrial KATP channel blocker 5-hydroxydecanoate (5-HD; 500 µmol/l). Tests are shown in A, D, and G; vehicle (V) time controls are shown in B, E, and H; and both are superimposed in C, F, and I. The averaging procedure and symbols are the same as in Fig. 2. Note that each test has its own vehicle time control. CCCP concentrations are in nanomoles per liter. Further analysis is shown in Fig. 4, and statistics with the number of experiments are shown in Table 1.

View larger version (42K): [in this window] [in a new window]   Fig. 4.   Superposition of group-averaged -ratios ± SE for the CCCP-5-HD experiments shown in Fig. 3 at the 4-min (A, D, and G), 10-min (B, E, and H), and 14-min (C, F, and I) time points. Note the match of signals for tests vs. vehicle time controls during the application of 500 µmol/l 5-HD plus 200 or 400 nmol/l CCCP.

The coapplication of glibenclamide or HMR-1098 with 400 nmol/l CCCP caused a striking increase in the amplitude of the calcium transients and baseline (Fig. 5, D and G) compared with CCCP alone (superpositions in Fig. 5, F and I). The coapplication of diazoxide had little effect (Fig. 5C). Analysis of the ratio changes (Fig. 6) clearly indicated that the traces for diazoxide and vehicle were superposed (A-C) at the 4-, 10-, and 14-min time points, whereas there was a sharp increase in baseline and amplitude upon the application of glibenclamide and HMR-1098 (D and G). Upon withdrawal of CCCP, there was no further recovery in the amplitude in the glibenclamide- or HMR-1098-treated myocytes, whereas a significant recovery from CCCP was observed for the controls and in diazoxide-treated myocytes (Fig. 7). Table 2 details the statistical analysis for the normalized amplitude, baseline ratio, and time constant. The most striking features are the significantly larger transient amplitudes and recovery time constants with glibenclamide and HMR-1098 relative to the control group.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Effect of glibenclamide (G; 1 µmol/l) and HMR-1098 (H; 30 µmol/l) and the lack of effect of diazoxide (D; 500 µmol/l) in metabolically compromised atrial myocytes. Tests are shown in A, D, and G; vehicle time controls are shown in B, E, and H; and both are superimposed in C, F, and I. The averaging procedure and symbols are the same as in Fig. 2. Note that each test has its own vehicle time control. The CCCP concentration was 400 nmol/l. Further analysis is shown in Figs. 6 and 7 and Table 2.

View larger version (45K): [in this window] [in a new window]   Fig. 6.   Superposition of group-averaged -ratios ± SE for the experiments shown in Fig. 5 at the 4-min, 10-, and 14-min time points. Note the match of signals for diazoxide vs. vehicle time control plus 400 nmol/l CCCP (A-C) and the large differences for glibenclamide (D and E) and HMR-1098 (G-I).

View larger version (34K): [in this window] [in a new window]   Fig. 7.   Minimal and maximal ratio transients ± SE for the data shown in Fig. 5. The vehicle group (A) includes the 3 time-control groups shown in Fig. 5. The column display emphasizes the upward jump in baseline ratio and peak amplitude with glibenclamide (C) and HMR-1098 (D) and the minor effects of diazoxide (B). See Fig. 5 for an explanation of the symbols; see Table 2 for statistics.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

There are two new findings in this study on atrial myocytes: 1) blockade of plasmalemmal K_ATP channels potently increases the amplitude, recovery time constant, and baseline of the calcium transients in metabolically impaired atrial myocytes; and 2) modulation of mitochondrial K_ATP channel activity has little effect on electrically triggered cytoplasmic calcium transients in either metabolically intact or impaired cells.

Role of plasmalemmal K_ATP channels in shaping cytoplasmic calcium transients. Both glibenclamide and HMR-1098 caused a dramatic increase in the amplitude, baseline, and recovery time constant of the calcium transients in CCCP-treated myocytes relative to the controls (Figs. 5-7 and Table 2). Only a minor part of this increase may be due to a potential reversal of an effect of ethanol. This vehicle for CCCP caused a minor decrease in the transient amplitude and had no effect on the recovery time constant or baseline. Interestingly, the recovery time constant was significantly reduced relative to the baseline control after withdrawal of CCCP and HMR-1098. This could be explained if HMR-1098, by partly maintaining the calcium transients, also maintained the magnitude of the contractions during CCCP and thus increased the metabolic stress. As a consequence, the ADP-to-ATP ratio may have increased and caused the opening of sarcolemmal K_ATP channels as the HMR-1098 was withdrawn. These results imply that blockade of sarcolemmal K_ATP channels of a metabolically stressed heart could have detrimental consequences on cellular metabolism and survival. It is speculated that this situation may arise during treatment of diabetic heart patients with sulfonylurea-derived drugs.

No significant role for mitochondrial K_ATP channels in shaping cytoplasmic calcium transients. The role of mitochondrial calcium uptake and release in shaping cytoplasmic calcium transients is controversial. In ventricular myocytes challenged by caffeine, the mitochondria contribute <3% to the total removal of cytoplasmic free calcium, whereas calcium uptake by the reticulum accounts for >60% (24). However, one could argue that the caffeine-induced, massive release of calcium from the sarcoplasmic reticulum overwhelms potential calcium uptake/release systems in the mitochondria. A more physiological approach consists in recording calcium transients during the natural contraction/relaxation cycle (15), where mitochondrial calcium uptake and release is estimated to be "substantial." For these reasons, we monitored cytoplasmic calcium transients in response to triggered depolarizations of the atrial myocytes while applying mitochondrial K_ATP channel modulators.