Cellular Biology - Circulation Research

2 downloads 0 Views 344KB Size Report
Daut J, Maier-Rudolph W, von Beckerath N, Mehrke G, Gunther K, ... Paul RJ, Bauer M, Pease W. Vascular smooth muscle: aerobic glycolysis linked to sodium ...
Cellular Biology Influence of Mitochondrial Inhibition on Global and Local [Ca2ⴙ]i in Rat Tail Artery Karl Swärd, Karl Dreja, Anders Lindqvist, Erik Persson, Per Hellstrand Abstract—Inhibition of oxidative metabolism is often found to decrease contractility of systemic vascular smooth muscle, but not to reduce global [Ca2⫹]i. In the present study, we probe the hypothesis that it is associated with an altered pattern of intracellular Ca2⫹ oscillations (waves) influencing force development. In the rat tail artery, mitochondrial inhibitors (rotenone, antimycin A, and cyanide) reduced ␣1-adrenoceptor–stimulated force by 50% to 80%, but did not reduce global [Ca2⫹]i. Less relaxation (about 30%) was observed after inhibition of myosin phosphatase activity with calyculin A, suggesting that part of the metabolic sensitivity involves the regulation of myosin 20-kDa light chain phosphorylation, although no decrease in phosphorylation was found in freeze-clamped tissue. Confocal imaging revealed that the mitochondrial inhibitors increased the frequency but reduced the amplitude of asynchronous cellular Ca2⫹ waves elicited by ␣1 stimulation. The altered wave pattern, in association with increased basal [Ca2⫹]i, accounted for the unchanged global [Ca2⫹]i. Inhibition of glycolytic ATP production by arsenate caused similar effects on Ca2⫹ waves and global [Ca2⫹]i, developing gradually in parallel with decreased contractility. Inhibition of wave activity by the InsP3 receptor antagonist 2-APB correlated closely with relaxation. Furthermore, abolition of waves with thapsigargin in the presence of verapamil reduced force by about 50%, despite unaltered global [Ca2⫹]i, suggesting that contraction may at least partly depend on Ca2⫹ wave activity. This study therefore indicates that mitochondrial inhibition influences Ca2⫹ wave activity, possibly due to a close spatial relationship of mitochondria and the sarcoplasmic reticulum and that this contributes to metabolic vascular relaxation. (Circ Res. 2002;90:792-799.) Key Words: arterial smooth muscle 䡲 metabolic inhibition 䡲 myosin phosphorylation 䡲 calcium waves 䡲 confocal imaging

T

he distribution of blood flow in tissue depends on the sensitivity of arterial tone to the local oxygen tension, but the nature of the mediating signal is yet unclear. Proposed mechanisms include reduced Ca2⫹ inflow across the cell membrane,1– 4 decreased myosin phosphorylation,5 or inhibition of cross-bridge cycling.6 In some cases of vascular hypoxia or mitochondrial inhibition, the intracellular free Ca2⫹ concentration ([Ca2⫹]i), measured as an average from multiple cells, is unchanged or even increased at high stimulus levels while force is reduced.7–10 Thus mechanisms based on decreased global [Ca2⫹]i are insufficient to explain hypoxic relaxation. It is puzzling also that in many studies myosin light chain phosphorylation does not decrease,6,9 even though exceptions are seen.5 Finally, insufficient MgATP for cross-bridge interaction, although clearly a possibility with severe metabolic inhibition, does not seem to be a general explanation for hypoxic relaxation.10 Within the vessel wall, individual cells may exhibit asynchronous Ca2⫹ oscillations during agonist stimulation. These depend on the function of the sarcoplasmic reticulum (SR) and are primarily associated with InsP3-induced Ca2⫹ release.11–14 In addition, highly localized increases in [Ca2⫹]i in

close proximity to the cell membrane due to activation of ryanodine receptors in the SR, termed Ca2⫹ sparks, can influence the activity of neighboring KCa channels, modulating membrane potential and thus voltage-dependent Ca2⫹ inflow.15 [Ca2⫹]i at the cellular level therefore varies both spatially and temporally as demonstrated by confocal microscopy, in contrast to the “global” [Ca2⫹]i measured by conventional imaging methods. This allows for temporal and spatial coding of Ca2⫹ signals, explaining why this single ion can simultaneously regulate a multitude of cellular processes. In addition to the plasma membrane and SR, mitochondria may contribute to the regulation of intracellular Ca2⫹ distribution. Considering that adrenergic stimulation causes increased Ca2⫹ wave activity in vascular tissue,11–14 it is possible that myosin phosphorylation fluctuates in a corresponding manner, whereas measurements after extraction of freezeclamped tissue only reveal global changes. This suggests the possibility that contraction may be modulated via the spatial and temporal pattern of intracellular Ca2⫹ transients as a possible site of action of metabolic inhibition. In addition to mitochondrial inhibitors, the effects of inhibited glycolytic energy supply are of interest, as there is evidence that energy

Original received July 2, 2001; revision received February 28, 2002; accepted March 1, 2002. From the Department of Physiological Sciences, Lund University, Lund, Sweden. Correspondence to Karl Swärd, PhD, Dept of Physiological Sciences, Lund University, BMC F12, S-221 84 Lund, Sweden. E-mail [email protected] © 2002 American Heart Association, Inc. Circulation Research is available at http://www.circresaha.org

DOI: 10.1161/01.RES.0000015214.40360.84

792

Swärd et al

Mitochondria and Force in Arteries

793

metabolism is compartmentalized, with preferential support of membrane ion pumps via glycolysis.17

Materials and Methods Preparation and Force Measurements Female Sprague-Dawley rats (Møllegaard, Copenhagen, Denmark) were killed by cervical dislocation as approved by the regional ethics committee. Segments were prepared in nominally Ca2⫹-free modified Krebs solution containing (mmol/L) 135.5 NaCl, 5.9 KCl, 1.2 MgCl2, 11.6 glucose, and 11.6 HEPES, pH 7.35. To remove endothelium, the segment was slid over an insect needle (250 ␮m diameter) 20 to 30 times. Rings (1 mm wide) were mounted for isometric force registration as described.18 After equilibration in Krebs solution containing 2.5 mmol/L Ca2⫹ for 30 minutes at 37°C, preparations were repeatedly contracted with 0.3 ␮mol/L cirazoline (7 minutes) followed by relaxation for 25 minutes, until stable contractions were attained and experimentation begun.

Global Arterial Wall [Ca2ⴙ]i Measurements Arterial segments (2 mm) were turned inside out for optical access to the medial smooth muscle cells. Segments were threaded onto glass capillaries (350 ␮m in diameter) and loaded with 16 ␮mol/L fura-2 AM in Krebs solution (pH 7.4) in the dark for 3 hours. The 340:380-nm ratio of fura-2 fluorescence was determined at room temperature as an indirect measure of intracellular free Ca2⫹ concentration ([Ca2⫹]i).19 These measurements are referred to as “global [Ca2⫹]i,” because they represent the activity of several, although not necessarily all, medial cells.

Confocal Imaging of Local [Ca2ⴙ]i Arteries mounted on glass capillaries were loaded for 80 minutes with 10 ␮mol/L fluo-4 AM in Eppendorf tubes in the dark. They were subsequently mounted in a Zeiss LSM 510 laser scanning confocal microscope as described.19 Fluo-4 was excited at 488 nm, and emission recorded at ⬎505 nm with a ⫻63 oil immersion objective (NA⫽1.25). Pinhole settings were adjusted to give confocal planes of 1 ␮m. Image series were recorded at 0.3 to 0.7 s/image. The fraction of oscillating cells was obtained by dividing the number of cells with spontaneous [Ca2⫹]i oscillations by the total number of cells in the field (maximally 60 cells). Oscillations were defined as sharp deflections from a stable baseline with amplitudes larger than 10% of basal fluorescence, and these were easily distinguished from movement artifacts which only occurred during development and decay of tension, not during maintained contraction (cf, online Movie found in the online data supplement available at http:// www.circresaha.org). Clustered events (⬎10%), separated by incomplete return to baseline, were counted as one oscillation. To reduce clustering, the cirazoline concentration was kept at 0.1 ␮mol/L. Line-scans (20 000 scans at 3 ms each) were obtained longitudinally within the discernible contours of single cells.

Myosin 20-kDa Light Chain Phosphorylation Segments of tail artery mounted on glass capillaries were frozen at 5 minutes after stimulation with cirazoline in 10% (wt/vol) trichloroacetic acid, 20 mmol/L dithiothreitol, in acetone on dry ice. Extraction, urea-glycerol gel electrophoresis, and blotting were as described by Mita and Walsh.20

Chemicals Cirazoline hydrochloride and calyculin A were from ICN; arsenate, rotenone, dinitrophenol, antimycin A, and 2-aminoethyl diphenyl borate (2-APB) were from Sigma; and fura-2 AM and fluo-4 AM from Molecular Probes. Dr Michael P. Walsh (University of Calgary, Alberta, Canada) kindly provided the polyclonal anti-LC20 antibody.

Statistics Summarized data are presented as mean⫾SEM Analysis of variance was used for multiple comparisons. Student’s t test or Wilcoxon

Figure 1. Effect of mitochondrial inhibition on [Ca2⫹]i and force during agonist stimulation of arterial rings. Rings were stimulated with cirazoline (0.3 ␮mol/L, black bars). A through D, Effects of site I through III inhibition (10 ␮mol/L rotenone, 5 ␮mol/L antimycin A, and 3 mmol/L cyanide) and uncoupling (50 ␮mol/L dinitrophenol) on the fura-2 ratio. Breaks represent 20to 25-minute washout periods. F through I, Effects on force. Calibration bars⫽5 mN and 5 minutes. E, Summarized ratio and force (n⫽3 to 6). In E, b indicates ratio under basal conditions; c, integrated ratio and force normalized to the preceding control response. *P⬍0.05. J, Force versus relative ratio for a cumulative concentration-response curve to cirazoline [3, 10, 30 nmol/L (䢇), and 0.3 ␮mol/L (䡩)] and for responses to cirazoline (0.3 ␮mol/L) in the presence of dinitrophenol (ƒ), antimycin A (▫), cyanide (⌬), and rotenone (䡩). Control force at 0.3 ␮mol/L cirazoline (䡩) was 16.6⫾1.2 mN/mm, n⫽9.

matched pairs test was used for single comparisons. A value of P⬍0.05 was considered significant. For n⬍6, determinations were made in rings from separate animals. Modeling was made using MATLAB software (The Math Works Inc).

Results The effect of mitochondrial inhibition with rotenone (site I, 10 ␮mol/L), antimycin A (site II, 5 ␮mol/L), cyanide (site III, 3 mmol/L), and dinitrophenol (uncoupler, 50 ␮mol/L) on arterial wall (global) [Ca2⫹]i and force was tested during stimulation with the ␣1-selective agonist cirazoline (0.3 ␮mol/L). With the exception of uncoupling, short exposures (10 minutes, Figures 1A through 1E) did not reduce stimulated [Ca2⫹]i. All interventions significantly inhibited contraction (Figures 1E through 1I). Longer exposure (2 hours) to

794

Circulation Research

April 19, 2002

Figure 2. Uncoupling but not site I through III inhibition reduces myosin phosphorylation. Myosin regulatory light chain phosphorylation (MLC-P) was measured under basal conditions and after stimulation with cirazoline in the presence and absence of mitochondrial inhibitors. A, Summarized data (n⫽6). B, Original blot with a general MLC antibody. ***P⬍0.001.

antimycin A, but not rotenone or cyanide, reduced cirazoline stimulated [Ca2⫹]i (P⬍0.05, n⫽3, data not shown). In Figure 1J, force is plotted versus global [Ca2⫹]i for a cumulative concentration-response curve with cirazoline (control) and for the highest concentration of cirazoline (0.3 ␮mol/L) in the presence of mitochondrial inhibitors. The inhibitor data fall below and to the right of the control relationship, implying that a considerable fraction (50% to 70%) of force reduction is independent of global [Ca2⫹]i. Myosin light chain phosphorylation (MLC-P) was determined after 25 minutes in relaxing solution and after 5 minutes stimulation with cirazoline following 5 minutes pretreatment with either vehicle or inhibitors. The protocol was thus identical to that used for force determination. As illustrated in Figures 2A and 2B, MLC-P was not significantly reduced by rotenone, antimycin A, or cyanide, whereas dinitrophenol almost totally abolished the phosphorylation response to stimulation (n⫽6 for all). Although we cannot exclude some reduction of MLC-P with antimycin A and cyanide (from 0.38 to ⬇0.30 mol Pi per mol regulatory light chain), phosphorylation, being a global measurement, shows essentially the same response to mitochondrial inhibition as does global [Ca2⫹]i. A straightforward interpretation of these results is that the metabolic sensitivity of vascular tone depends on mechanisms distinct from [Ca2⫹]i elevation and myosin phosphorylation. However, when stable myosin phosphorylation was induced by the membrane-permeable phosphatase inhibitor calyculin A (3 ␮mol/L), the associated force response was reduced by only 26⫾10% on addition of rotenone (Figure 3A; n⫽4). A cirazoline-induced contraction of the same magnitude was reduced by 85⫾4% (n⫽4, P⬍0.001). Contractions induced by high-K⫹, which result from activation of voltage-sensitive Ca2⫹ channels, were also less sensitive to rotenone and antimycin A, although no difference was seen with cyanide (Figures 3B and 3C). The greater effect on global [Ca2⫹]i of uncoupling with dinitrophenol relative to sites I–III inhibition might be due to ATP consumption by mitochondria because reverse-mode activity of the ATP synthase is possible in the presence of

Figure 3. Effect of mitochondrial inhibition on force elicited by phosphatase inhibition or depolarization. A, Effect of rotenone (10 ␮mol/L) on force induced by calyculin A (3 ␮mol/L) or by cirazoline (0.3 ␮mol/L). B and C, Relative residual force after incubation with inhibitors during stimulation with 60 (B) or 140 (C) mmol/L K⫹ (open bars), or 0.3 ␮mol/L cirazoline (solid bars). n⫽4 to 5.

dinitrophenol. When oligomycin (5 ␮g/mL) was used to inhibit the ATP synthase, the stimulated [Ca2⫹]i after uncoupling increased by 27⫾9% (P⬍0.05, n⫽4; cf, online Figure A found in the online data supplement available at http:// www.circresaha.org). Oligomycin had no effect by itself at 10 minutes, but long-term (2-hour) exposure inhibited the [Ca2⫹]i response to cirazoline (not shown). These results suggest that mitochondrial ATP consumption may play some, although limited, role in the [Ca2⫹]i response to uncoupling. Neither glibenclamide (10 ␮mol/L) nor tetraethyl ammonium (TEA, 1 mmol/L), which block KATP and KCa channels, respectively, had any effect on the reduction of [Ca2⫹]i by dinitrophenol (not shown). Even though force reduction by mitochondrial inhibition could not be explained by decreased global [Ca2⫹]i (Figures 1 and 4A), the possibility was considered that temporal fluctuations or spatial compartmentation might explain the results. To investigate this, laser scanning confocal microscopy was used. Frame scanning demonstrated that 9⫾2% (n⫽9) of cells in the vessel media exhibited asynchronous oscillations of [Ca2⫹]i with a broad frequency spectrum. This value increased to 74⫾3% on stimulation with cirazoline (0.1 ␮mol/L, n⫽37; Figure 5A). During stimulation with the

Swärd et al

Mitochondria and Force in Arteries

795

Figure 5. Effect of mitochondrial inhibition on wave activity. A, Percentage of active cells per frame (8 frames per condition, nⱖ3) showing wave activity in unstimulated and stimulated (cirazoline 0.1 ␮mol/L) tail artery, and effects of rotenone (10 ␮mol/L), cyanide (3 mmol/L), antimycin A (5 ␮mol/L), dinitrophenol (50 ␮mol/L), or arsenate (2 mmol/L). B, Wave amplitude (solid columns) and frequency (open columns), with and without (c) mitochondrial inhibitors (ⱖ21 cells). *P⬍0.05.

Figure 4. Mitochondrial inhibition changes the pattern of intracellular Ca2⫹ waves. A, Representative tracing (n⫽4) showing effect of rotenone (10 ␮mol/L, gray bar) on global [Ca2⫹]i during stimulation with cirazoline (0.1 ␮mol/L, black bar). B, Confocal images from an area of arterial wall stimulated with cirazoline before (left) and after (right) addition of rotenone. Circles indicate 3 areas of interest. C, Fluo-4 fluorescence intensity vs time for the areas of interest. Positions of images in B indicated by arrowheads. D, Stacks of 3-ms cellular line-scans in another preparation before (left) and after (right) rotenone treatment.

phosphatase inhibitor calyculin A, waves were present in 42⫾8% of cells (frequency 3.8⫾0.3 waves/min, n⫽10). This agent thus induces an intracellular Ca2⫹ response, but its main effect on contraction is mediated via a potent inhibitory effect on myosin phosphatase, sufficient to elicit a Ca 2⫹ independent contraction.21 Addition of rotenone during stimulation with cirazoline increased the frequency and reduced the amplitude of the oscillations in active cells without altering time-averaged cellular [Ca2⫹]i. One experiment, in which the same cells could be identified before and after addition of rotenone, is shown in Figures 4B and 4C. The majority of the focal increases in [Ca2⫹]i spread along the axis of the cell and proceeded considerable distance as waves. Temporal stacks of line-scans from 2 different cells in one preparation before

and after addition of rotenone clearly illustrate the altered wave pattern provoked by the mitochondrial inhibitor (Figure 4D). In 21 cells from 3 arterial rings, the frequency of oscillations increased by 71%, whereas the amplitude decreased by 46%, on addition of rotenone (Figure 5B). Spatially averaged fluorescence after addition of rotenone was 64⫾10 fluorescence units, versus 53⫾1 before addition (P⬎0.05), supporting the results obtained with fura-2 that rotenone does not lower global [Ca2⫹]i. Cyanide likewise increased the frequency and reduced the amplitude of Ca2⫹ waves (Figure 5B). Both rotenone and cyanide slightly reduced the number of active cells (Figure 5A). Because uncoupling, in contrast to rotenone and cyanide, lowered global [Ca2⫹]i, one might anticipate distinct effects on local [Ca2⫹]i oscillations. At 10 minutes after the addition of dinitrophenol, the oscillatory pattern of [Ca2⫹]i was completely abolished (Figure 5A). Likewise, the site II inhibitor antimycin A caused a pronounced reduction of the number of active cells. In the remaining active cells, the frequency was increased by 20% and the amplitude reduced by 61% (Figure 5B). Consistently, both global [Ca2⫹]i (Figure 1) and Ca2⫹ oscillations (Figure 5) showed parallel tendencies to decrease with time of exposure to antimycin A. Similar effects do not arise with the site I and III inhibitors, because their responses are more stable in time. The effect of inhibited glycolysis on Ca2⫹ waves was investigated using arsenate, which uncouples ATP formation during the conversion of 1,3-diphosphoglycerate to phosphoglycerate. Although it does not inhibit glycolysis, there is no net glycolytic generation of ATP.22 Treatment with arsenate (2 mmol/L) progressively reduced the cirazoline-induced global [Ca2⫹]i response (n⫽4, P⬍0.001; cf, online Figure C). Basal [Ca2⫹]i increased in parallel (n⫽4, P⬍0.01). Similar

796

Circulation Research

April 19, 2002

Figure 6. Global [Ca2⫹]i, force, and wave activity after inhibition of glycolysis. A, Representative trace (n⫽3) of global [Ca2⫹]i during stimulation with cirazoline (0.1 ␮mol/L, black bar) with addition of arsenate (2 mmol/L, gray bar). B, Mean force (black line)⫾SEM (dotted line) after addition of 2 mmol/L arsenate at 0 minutes (n⫽4). Filled symbols indicate mean force at the times when confocal Ca2⫹ images were obtained. Open symbols show the product of percent active cells (C), normalized frequency (D), and normalized amplitude (E) of Ca2⫹ waves.

but more dramatic effects were obtained using iodoacetic acid, which halts glycolysis (online Figure B). Arsenate treatment for 1 hour (n⫽4) greatly reduced local Ca2⫹ oscillations in the presence of cirazoline (0.1 ␮mol/L, Figure 5A), while spatially averaged fluorescence increased (control: 42⫾3 fluorescence units; arsenate: 54⫾7; arsenate⫹cirazoline: 54⫾5; P⬍0.05, n⫽4), in accordance with global [Ca2⫹]i measurements with fura-2. Global [Ca2⫹]i increased when arsenate was added on top of a stimulation with 0.1 ␮mol/L cirazoline (Figure 6A). In parallel, force increased slightly, peaked at 2 minutes, and then decreased slowly to 20% of the original force at 30 minutes (Figure 6B). The percentage of active cells and the amplitude of Ca2⫹ waves followed essentially the same pattern as force (Figures 6C and 6E). The frequency of Ca2⫹ waves on the other hand increased to almost 2-fold the original frequency at 30

Figure 7. Inhibition of Ca2⫹ waves by thapsigargin or 2-APB reduces adrenergically stimulated force. A, Effects of cirazoline (0.3 ␮mol/L, black bar) and verapamil (1 ␮mol/L, gray bar) on global [Ca2⫹]i. After thapsigargin treatment (10 ␮mol/L, 10 minutes), stimulation was repeated. First and second breaks are 25 and 60 minutes. B, Force recorded in another ring taken through the same protocol. C, Typical traces of Ca2⫹ wave activity under conditions as in A and B, except that 0.1 ␮mol/L cirazoline was used. Da, Concentration-dependent (IC50⫽11 ␮mol/L) inhibition of cirazoline (0.1 ␮mol/L)-induced force by 2-APB. Db through Dd, Percent active cells, normalized frequency, and normalized amplitude of Ca2⫹ waves, together with logistic fit to force data (dotted line). De, Product of data from Db through Dd.

minutes (Figure 6D). The product of the percentage of active cells, wave amplitude, and frequency (open circles in Figure 6B) correlated with force. To test if Ca2⫹ waves are important for force maintenance after stimulation with cirazoline, we used thapsigargin, which is a potent inhibitor of the sarcoplasmic reticulum Ca2⫹ATPase. Agents emptying the Ca2⫹ stores have been reported to cause depolarization of arterial smooth muscle cells15; therefore, the L-type Ca2⫹ channel blocker verapamil was included in the experiment to avoid compensatory Ca2⫹ inflow due to depolarization. Verapamil (1 ␮mol/L) reduced the sustained rise of global [Ca2⫹]i induced by cirazoline (0.3 ␮mol/L) by 22⫾6% (n⫽4, P⬍0.05; Figures 7Aa and 7Ab) and inhibited the associated force by 31⫾3% (P⬍0.05; Figure 7Bb). Waves were still present after addition of verapamil (73⫾8 versus 81⫾3% active cells; Figure 7Cb) with unchanged wave frequency (7.1⫾0.6 versus 8.5⫹0.6

Swärd et al waves/min) and amplitude (66⫾6 versus 67⫾5 fluorescence units, 39 and 35 cells). After thapsigargin treatment, global [Ca2⫹]i in the stimulated preparations was maintained (85⫾7%, Figure 7Ac, versus 78⫾6%, Figure 7Ab; NS), whereas waves were completely abolished (Figure 7Cc) and contraction reduced by an additional 54⫾3% (Figure 7Bc; P⬍0.001). To further establish a role of waves in force maintenance, we used the InsP3 receptor antagonist 2-APB, which dosedependently inhibited cirazoline (0.1 ␮mol/L)-induced force (Figure 7Da) with an IC50 value of 11 ␮mol/L. Force inhibition by 2-APB was associated with dose-dependent reductions of the percentage of active cells (Figure 7Db), Ca2⫹ wave frequency (Figure 7Dc), and amplitude (Figure 7Dd). The product of the number of active cells, wave frequency, and amplitude correlated well with tissueaveraged force (Figure 7De).

Discussion This study demonstrates a novel mechanism for an influence of metabolism on excitation-contraction coupling in vascular smooth muscle. The temporal and spatial pattern of intracellular Ca2⫹ fluctuations (waves) was shown to be affected by inhibition of oxidative or glycolytic metabolism, revealing correlations with force development that are not shown by conventional (global) measurements of either [Ca2⫹]i or MLC-P. Potentially, this explains earlier puzzling observations,7–10 but it should be emphasized that the evidence for a causal relationship between altered Ca2⫹ wave activity and relaxation during metabolic inhibition so far is indirect, as discussed later. A further aspect is that Ca2⫹ waves may influence other cellular processes, such as signaling pathways for gene expression and apoptosis,23 and that the effects shown in this study might reveal mechanisms for long-term effects of hypoxia on cellular function. Further work to explore these possibilities should prove fruitful. In the rat tail artery, normal wave activity was found in preparations where the ryanodine receptors had been functionally inactivated,19 and thus InsP3 receptor activity seems to be the major cause of waves in this vessel. A crucial factor in the generation of waves is the Ca2⫹ sensitivity of the InsP3 receptor,24 implying a positive feedback control to increase the rate of Ca2⫹ release. High Ca2⫹ concentrations exert a negative effect on InsP3 receptor activity,23–25 of possible importance in a negative feedback loop to limit emptying of Ca2⫹ stores. A close spatial contact between mitochondria and the endoplasmic reticulum has been demonstrated in HeLa cells,16 and it is possible that a similar arrangement exists with the sarcoplasmic reticulum in smooth muscle. Indirect evidence for this are the functional results showing that Ca2⫹ released from the SR is rapidly taken up into mitochondria,26 and that mitochondrial inhibition reduces the amount of Ca2⫹ released by InsP3.27 These findings suggest that mitochondria effectively reduce the Ca2⫹ concentration in a restricted space near the SR, as it would otherwise be high enough to inhibit further release. Of direct relevance to the present findings, Jouaville et al28 showed that addition of mitochondrial substrates decreases the frequency and increases the amplitude of

Mitochondria and Force in Arteries

797

Figure 8. Modeling predicts regulation of force by Ca2⫹ wave frequency and amplitude independently of global [Ca2⫹]i. A, Fluorescence traces from the same cell before (black) and after (gray) addition of rotenone, linearly transformed (Equation 1) to [Ca2⫹]i. Mean (global) [Ca2⫹]i values are shown by bars to the right. B, Further transformation with a logistic function (Equation 2) gives cellular force transients. C, Mean force and [Ca2⫹]i for the 3 cells in Figure 4 were obtained for a range of physiologically reasonable calibrations by varying the calibration factor (k) between fluorescence and [Ca2⫹]i. Equation 1: [Ca2⫹]i⫽Fluorescence (AU)⫻k⫹m; Equation 2: Force (%)⫽(100⫺a)/(1⫹([Ca2⫹]i / EC50)␣⫹a; where k is the calibration factor (AU), chosen to give peak [Ca2⫹]i⫽750 nmol/L in A, varied in C; m⫽30 nmol/L (not critical ⬍100 nmol/L); EC50⫽700 nmol/L (unpublished data from permeabilized tail artery); ␣⫽⫺4.7 (from Gomez and Swärd32); and a⫽0.98 (not critical).

Ca2⫹ waves in Xenopus laevis oocytes, effects that were blocked by rotenone. A central consideration in interpreting the present results is whether an altered wave pattern toward increased frequency and decreased amplitude is able to explain a decrease in force production in spite of maintained (or at least nearly so) levels of global [Ca2⫹]i and MLC-P. Relevant factors are that the relationships between [Ca2⫹]i and MLC-P, and between MLC-P and force, are both nonlinear and time-dependent. A simple model may be used to illustrate the impact of frequency and amplitude variations of Ca2⫹ waves on cell force to allow comparison with the observed effects. A sigmoidal [Ca2⫹]i-force relationship with an EC50 of 7 · 10⫺7 mol/L, obtained from experiments on skinned preparations (unpublished data) was used for conversion of fluorescence traces to force before and after treatment with rotenone (Figures 8A to 8B). Calculation of mean values from the individual traces shows that the inhibitor has a relatively greater effect on force than on [Ca2⫹]i. Because the fluo-4 recording does not allow an absolute measure of [Ca2⫹]i, the data in Figures 8A and 8B were generated using an assumed calibration factor giving physiologically reasonable [Ca2⫹]i values. To investigate the effect of calibration, a range of [Ca2⫹]i values was assigned to the means calculated for the 3 cells in Figures 4B and 4C. This also allows for any Ca2⫹ sensitization caused by the adrenergic stimulation. Averaging over the cells yielded the curves shown in Figure 8C. At low and intermediate [Ca2⫹]i, representing expected values during

798

Circulation Research

April 19, 2002

adrenergic stimulation in the tail artery,18 force was considerably reduced for similar Ca2⫹ levels during mitochondrial inhibition. A further prediction of the model is that force can be reduced as a consequence of reduced interwave [Ca2⫹]i because this would affect the position of the wave peaks vis à vis the threshold for contraction. The partial relaxation seen during L-type channel blockade (cf, Figure 7), which did not abolish wave activity, could thus be due to a decrease in Ca2⫹ influx across the cell membrane, affecting interwave [Ca2⫹] and apparent in the measurements of global [Ca2⫹]. Different contributions to activator Ca2⫹ by influx via L-type channels may also explain discordant results in the literature, as in pressurized mesenteric resistance arteries addition of nifedipine was found to totally relax force, but not to eliminate wave activity.12 The model allows for both [Ca2⫹] versus MLC-P and for MLC-P versus force being nonlinear.29 –31 On the other hand, it does not incorporate temporal latencies inherent in activation and force development, which will cause phase shifts between [Ca2⫹]i and force. Effects of such phase shifts, if significant, would be expected to decrease the amplitude of force when oscillation frequency increases, as the system will function as a low-pass filter. Ca2⫹ waves engaging only a section of a cell at a time might be expected to cause a segmental force development influenced by cell and tissue elasticity. However, we never observed length oscillations in individual cells during agonist-stimulated asynchronous wave activity. An interesting comparison is the study by Peng et al13 in rat mesenteric resistance arteries. During the initial phase of low-level adrenergic stimulation, asynchronous Ca2⫹ waves and slowly developing force were observed, whereas after prolonged stimulation, intracellular Ca2⫹ oscillations became synchronous between cells and force started to oscillate. This suggests that although cells are able to contract and relax in phase with the Ca2⫹ transients, segmental movements are only observed when tissue stress varies synchronously over sufficiently large areas to allow measurable changes in elastic extension. Some level of metabolic sensitivity is seen with all modes of activation tested in this study, ie, in addition to adrenergic stimulation, depolarization by high-K⫹ solution, and inhibition of myosin phosphatase activity using calyculin A. It is not likely that inhibited wave activity accounts for the relaxation with these latter types of stimulation. In addition, uncoupling of oxidative phosphorylation using 2,3dinitrophenol caused a relaxation that was more pronounced than that produced by site I through site III inhibitors and was accompanied by decreased MLC-P. Thus, both stimulus conditions and modes of metabolic inhibition can be found that do no fit the pattern considered here with respect to the role of wave activity. This is to be expected because excitation-contraction coupling, and contraction itself, are energy-dependent at several steps. In a recent comprehensive investigation of the energetics of hypoxic relaxation in porcine coronary artery, Shimizu et al10 were able to rule out changes in either intracellular pH, ATP, or inorganic phosphate concentrations as causes of the response. Yet, given sufficiently potent metabolic inhibition, any of these factors may change sufficiently to cause loss of contractile force. In

addition, as pointed out by Shimizu et al,10 the Ca2⫹ sensitivity of contraction is by definition decreased when force is reduced in the presence of elevated [Ca2⫹]i, and in addition to effects mediated by altered myosin phosphorylation, this may involve a putative thin filament–mediated regulatory system. Whatever the nature of such energy-limited steps, they do not produce additive effects, ie, even if inhibition of a later step in the sequence produces (nearly) as much relaxation as that of an earlier step, this would be revealed only when the early step is bypassed. The fact that InsP3 receptor blockade by 2-APB caused relaxation that correlated with decreased wave activity indicates that waves in fact support force development. Furthermore, several observations suggest that wave activity is a better correlate to force than global [Ca2⫹]i, as inhibition of glycolytic ATP production by arsenate caused a timedependent relaxation correlating with decreased wave activity (active cells⫻frequency⫻amplitude), even though global Ca2⫹, as well as wave frequency, increased during the process. Similarly, force development in the presence of verapamil demonstrates that inflow of Ca2⫹ via L-type membrane channels is not necessary to support force; however, the subsequent elimination of waves by exposure to thapsigargin caused relaxation despite maintained elevation of global [Ca2⫹]i. Both thapsigargin and arsenate are expected to decrease extrusion of Ca2⫹ from the cell because thapsigargin inhibits Ca2⫹ uptake into the SR and arsenate inhibits glycolytic energy production,22 and thus, eg, Na⫹-K⫹ pump activity,17 which is of importance for Na⫹-Ca2⫹ exchange over the plasma membrane. Even though the different inhibitors of oxidative phosphorylation all produced similar results, there was a weaker relaxing effect of cyanide than of rotenone and antimycin A. This may have been due to the fact that cyanide has a low solubility at neutral pH and thus is difficult to administer in defined concentration. Metabolic inhibition may in some cases produce effects similar to hypoxia,7 so in addition to elucidating mechanisms due to metabolic inhibition or dysfunction, the present results may be relevant to hypoxic vasorelaxation. In summary, this study shows that inhibition of oxidative or glycolytic metabolism leads to relaxation of the vascular wall that may occur through an altered pattern of Ca2⫹ wave activity, which affects contractile activity of individual cells in the tissue but is not reflected in measurements of global [Ca2⫹]i or MLC-P.

Acknowledgments This study was supported by the Swedish Medical Research Council (project 04X-28). K.S. was supported by a postdoctoral fellowship from Svenska Sällskapet för Medicinsk Forskning.

References 1. Daut J, Maier-Rudolph W, von Beckerath N, Mehrke G, Gunther K, Goedel-Meinen L. Hypoxic dilation of coronary arteries is mediated by ATP-sensitive potassium channels. Science. 1990;247:1341–1344. 2. Franco-Obregon A, Lopez-Barneo J. Low PO2 inhibits calcium channel activity in arterial smooth muscle cells. Am J Physiol. 1996;271: H2290 –H2299. 3. McHugh D, Beech DJ. Modulation of Ca2⫹ channel activity by ATP metabolism and internal Mg2⫹ in guinea-pig basilar artery smooth muscle cells. J Physiol. 1996;492:359 –376.

Swärd et al 4. Miller AL, Morales E, Leblanc NR, Cole WC. Metabolic inhibition enhances Ca2⫹-activated K⫹ current in smooth muscle cells of rabbit portal vein. Am J Physiol. 1993;265:H2184 –H2195. 5. Coburn RF, Moreland S, Moreland RS, Baron CB. Rate-limiting energydependent steps controlling oxidative metabolism-contraction coupling in rabbit aorta. J Physiol. 1992;448:473– 492. 6. Obara K, Bowman PS, Ishida Y, Paul RJ. Effects of hypoxia on [Ca2⫹]i, pHi and myosin light chain phosphorylation in guinea-pig taenia caeci. J Physiol. 1997;503:427– 433. 7. Taggart MJ, Wray S. Hypoxia and smooth muscle function: key regulatory events during metabolic stress. J Physiol. 1998;509:315–325. 8. Aalkjaer C, Lombard JH. Effect of hypoxia on force, intracellular pH and Ca2⫹ concentration in rat cerebral and mesenteric small arteries. J Physiol. 1995;482:409 – 419. 9. Taggart MJ, Menice CB, Morgan KG, Wray S. Effect of metabolic inhibition on intracellular Ca2⫹, phosphorylation of myosin regulatory light chain and force in rat smooth muscle. J Physiol. 1997;499:485– 496. 10. Shimizu S, Bowman PS, Thorne G III, Paul RJ. Effects of hypoxia on isometric force, intracellular Ca2⫹, pH, and energetics in porcine coronary artery. Circ Res. 2000;86:862– 870. 11. Iino M, Kasai H, Yamazawa T. Visualization of neural control of intracellular Ca2⫹ concentration in single vascular smooth muscle cells in situ. EMBO J. 1994;13:5026 –5031. 12. Miriel VA, Mauban JR, Blaustein MP, Wier WG. Local and cellular Ca2⫹ transients in smooth muscle of pressurized rat resistance arteries during myogenic and agonist stimulation. J Physiol. 1999;518:815– 824. 13. Peng H, Matchkov V, Ivarsen A, Aalkjaer C, Nilsson H. Hypothesis for the initiation of vasomotion. Circ Res. 2001;88:810 – 815. 14. Ruehlmann DO, Lee CH, Poburko D, van Breemen C. Asynchronous Ca2⫹ waves in intact venous smooth muscle. Circ Res. 2000;86:e72– e79. 15. Jaggar JH, Porter VA, Lederer WJ, Nelson MT. Calcium sparks in smooth muscle. Am J Physiol. 2000;278:C235–C256. 16. Rizzuto R, Pinton P, Carrington W, Fay FS, Fogarty KE, Lifshitz LM, Tuft RA, Pozzan T. Close contacts with the endoplasmic reticulum as determinants of mitochondrial Ca2⫹ responses. Science. 1998;280: 1763–1766. 17. Paul RJ, Bauer M, Pease W. Vascular smooth muscle: aerobic glycolysis linked to sodium and potassium transport processes. Science. 1979;206: 1414 –1416. 18. Lindqvist A, Nordström I, Malmqvist U, Nordenfelt P, Hellstrand P. Long-term effects of Ca2⫹ on structure and contractility of vascular smooth muscle. Am J Physiol. 1999;277:C64 –C73.

Mitochondria and Force in Arteries

799

19. Dreja K, Nordström I, Hellstrand P. Rat arterial smooth muscle devoid of ryanodine receptor function: effects on cellular Ca2⫹ handling. Br J Pharmacol. 2001;132:1957–1966. 20. Mita M, Walsh MP. ␣1-Adrenoceptor–mediated phosphorylation of myosin in rat-tail arterial smooth muscle. Biochem J. 1997;327:669 – 674. 21. Ishihara H, Ozaki H, Sato K, Hori M, Karaki H, Watabe S, Kato Y, Fusetani N, Hashimoto K, Uemura D, Hartshorne DJ. Calciumindependent activation of contractile apparatus in smooth muscle by calyculin-A. J Pharmacol Exp Ther. 1989;250:388 –396. 22. Harris RA. In: Devlin TM, ed. Textbook of Biochemistry. New York, NY: Wiley Liss; 1992:308 –309. 23. Berridge MJ, Lipp P, Bootman MD. The versatility and universality of calcium signalling. Nat Rev Mol Cell Biol. 2000;1:11–21. 24. Miyakawa T, Mizushima A, Hirose K, Yamazawa T, Bezprozvanny I, Kurosaki T, Iino M. Ca2⫹-sensor region of IP3 receptor controls intracellular Ca2⫹ signaling. EMBO J. 2001;20:1674 –1680. 25. Cardy TJ, Traynor D, Taylor CW. Differential regulation of types-1 and -3 inositol trisphosphate receptors by cytosolic Ca2⫹. Biochem J. 1997; 328:785–793. 26. Gurney AM, Drummond RM, Fay FS. Calcium signalling in sarcoplasmic reticulum, cytoplasm and mitochondria during activation of rabbit aorta myocytes. Cell Calcium. 2000;27:339 –351. 27. McCarron JG, Muir TC. Mitochondrial regulation of the cytosolic Ca2⫹ concentration and the InsP3-sensitive Ca2⫹ store in guinea-pig colonic smooth muscle. J Physiol. 1999;516:149 –161. 28. Jouaville LS, Ichas F, Holmuhamedov EL, Camacho P, Lechleiter JD. Synchronization of calcium waves by mitochondrial substrates in Xenopus laevis oocytes. Nature. 1995;377:438 – 441. 29. Fitzsimons DP, Herring BP, Stull JT, Gallagher PJ. Identification of basic residues involved in activation and calmodulin binding of rabbit smooth muscle myosin light chain kinase. J Biol Chem. 1992;267:23903–23909. 30. Itoh T, Suzuki A, Watanabe Y, Mino T, Naka M, Tanaka T. A calponin peptide enhances Ca2⫹ sensitivity of smooth muscle contraction without affecting myosin light chain phosphorylation. J Biol Chem. 1995;270: 20400 –20403. 31. Malmqvist U, Arner A, Makuch R, Dabrowska R. The effects of caldesmon extraction on mechanical properties of skinned smooth muscle fibre preparations. Pflügers Arch. 1996;432:241–247. 32. Gomez M, Swärd K. Long-term regulation of contractility and calcium current in smooth muscle. Am J Physiol. 1997;273:C1714 –C1720.