Agonist-induced [Ca2+]i waves and Ca2+-induced Ca2+ release in mammalian vascular smooth muscle cells L. A. BLATTER Department

AND W. G. WIER

of Physiology,

University

of Maryland

Blatter, L. A., and W. G. Wier. Agonist-induced [Ca2+]i waves and Ca2+-induced Ca2+ release in mammalian vascular smooth muscle cells. Am. J. Physiol. 263 (Heart Circ. Physiol. 32): H576-H586, 1992.-Focal application of vasopressin to cultured vascular smooth muscle cells (A7r5 cells) elicits first a localized increase of intracellular Ca2+ concentration ( [ Ca2+] J and then a wave of elevated [Ca2+]i that propagates at constant velocity throughout the cell. The cellular mechanisms of such complex spatiotemporal patterns of [Ca”+]i are of interest because they are involved fundamentally in cellular signal transduction in many types of cells. Vasopressin evoked a [Ca”+]i transient even in the absence of extracellular Ca2+, and intracellular perfusion with heparin completely blocked the response to vasopressin stimulation. Therefore the initial response to vasopressin reflects release of Ca2+ from an intracellular myo-inositol-1,4,5trisphosphate (IP,)-sensitive Ca2+ store. We tested four hypotheses on how a localized increase in [Ca2+]i propagates as a [Ca2+]i wave throughout the entire cell: the hypotheses distinguished 1) whether IP, or Ca2+ is the primary intracellular messenger that diffuses, and 2) whether positive feedback on the release of intracellular Ca2+ (Ca2+.) is involved of phospholipase C (further release of Ca‘ 2+ through activation by Ca2+ and increased production of IP, or by Ca2+-induced Ca2+ release). The results of various experimental interventions, which included probing Ca2+i stores (heparin, caffeine, and ryanodine), were compared with predictions from mathematical models for intracellular diffusion, release, and uptake of Ca2+. We conclude that in A7r5 smooth muscle cells, which have been stimulated focally with vasopressin, Ca2+ is released initially by IP3. The localized increase in [Ca2+]i then propagates throughout the cell as a [Ca2+]; wave. Ca2+ activates its own release, through Ca2+-induced release of Ca2+, by diffusing to distant Ca2+-release sites. myo-inositolra-2; digital

1,4,5-trisphosphate;

imaging

A7r5

cells;

vasopressin;

fu-

microscopy

IN MANY CELLS, agonists trigger oscillatory changes of intracellular Ca2+ concentration ([ Ca2+] ;). Sometimes these changes also occur in a propagating spatiotemporal pattern known as [Ca2+]; waves (8, 28, 31). The quantitative study of such events can reveal important information on the mechanisms of Ca2+. regulation as well as on certain mechanisms of signal’ transduction, including long-range intercellular signaling and frequency-encoded signaling (l-3,24,26,33). In this report, the term “[ Ca2+] i wave” will refer specifically to changes in [Ca2+]i that propagate through a cell as a result of positive feedback in the mechanisms that increase [Ca2+]i. Specifically, we investigated the mechanisms by which hormonal signals for contraction are transduced into increased [Ca2+]i in vascular smooth muscle cells. [Ca2+]i was measured with the fluorescent Ca2+-indicator fura-2. We used cultured A7r5 cells, which are a line of smooth muscle cells derived from embryonic rat thoracic aorta (20). These cells have been well characterized biochemically and electrophysiologically and are used extensively for studying Ca2+ homeostasis in vascular H576

0363-6135/92

$2.00 Copyright

School of Medicine,

Baltimore,

Maryland

21201

smooth muscle (see Ref. 27 for references). For imaging [Ca2+];, as required in our study, these cells have the advantage of being quite thin so that spatial resolution is good, and they can be perfused internally with membrane-impermeant Ca2+ indicators and other substances with the use of micropipette electrodes. Stimulation of these cells with extracellular [Args]vasopressin (AVP) activates the phosphoinositide cascade, increasing the intracellular levels of myo-inositol-1,4,5trisphosphate (IP,) and [Ca2+]i in a concentration-dependent manner (9, 34). [Ca2+]i oscillates in these and other smooth muscle cells, but the mechanism remains obscure, and conflicting data exist on the possible involvement of Ca2+-induced Ca2+ release (CICR). Images of [Ca2+]i after application of AVP were made on a much higher temporal resolution than has been done previously (7), affording us the surprising observation of rapid [Ca2+]; waves: the focal stimulation with AVP increased [Ca2+]i in a circumscribed region of the cell from which a wave of elevated [Ca2+]i consequently propagated throughout the cell. We tested four models for the propagation of [Ca2+]; waves in A7r5 cells. These can be distinguished according to whether Ca2+ or IP, is the primary intracellular messenger that undergoes diffusion and whether a positive feedback mechanism for the release of Ca2+ is present. In the absence of positive feedback mechanisms, the simple diffusion of Ca2+ or IP3 would be sufficient to produce an apparent [Ca2+]i wave. The positive feedback mechanisms tested are either a modified cross coupling of the specific actions of IP3 and Ca2+ [IP,-Ca2+ cross coupling (ICC) model; 24-261, in which the rise in [Ca2+]i feeds back to the release of Ca2+ by further amplification of the production of IP3 through stimulation of phospholipase C by Ca2+ (see Refs. 3, 24-26 for references), or CICR (ll), in which a [Ca2+]i wave propagates via diffusion of Ca2+ to distant Ca2+release sites at which it can induce further release. These models were evaluated by comparing spatiotemporal distribution of [ Ca2+]; predicted by a mathematical model with measured spatiotemporal patterns of [Ca2+]i under various experimental conditions. METHODS

Cultured vascular smooth muscle cells. A7r5 cells were obtained from American Type Culture Collection (Rockville, MD) and kept in culture as described elsewhere (35). Once a week the cells were dispersed using culture medium containing 1% trypsin for passage and subculture onto cover slips for later experimentation. Cells from passages 6 to 34 were used, and the experiments were carried out within 1 wk after subculture onto cover slips. In general, the experiments were performed on single isolated cells that were not part of a large confluent cluster of cells. Solutions and drugs. The cells were superfused continuously

0 1992 the American

Physiological

Society

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

[CA*+]~ WAVES

AND CAM+-INDUCED

with a physiological salt solution composed of (in mM) 135 NaCl, 4 KCl, 1 MgC12, 2 CaCl,, 10 dextrose, 10 N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES), and 0.33 NaH2P0, titrated to pH 7.3 with NaOH (-3.9 mM). The addition of CaCl, was omitted from the nominally Ca2+-free superfusion solution. AVP (Sigma Chemical, St. Louis, MO) and ryanodine were made up as concentrated stock solutions. Caffeine was dissolved directly in the superfusion solution. Heparin (mol wt 4,000-6,000; 0.3 or 1 mg/ml) was applied by internal perfusion via a tight-seal patch-clamp pipette. Furaloading and calibration. Single A7r5 cells were loaded with the fluorescent Ca2+ indicator fura- either by exposure to the membrane-permeant form of fura-2, fura- acetoxymethyl ester (fura-2/AM, 5 PM), or by internal perfusion with membrane impermeant fura- (salt) through a tight-seal patch electrode. The composition of the pipette filling solution was (in mM) 130 Cs-glutamate, 10 NaCl, 10 HEPES, 10 CsCl, 1 MgC12, 3 MgATP or 3 Na2ATP, and 0.1 fura-2; pH was adjusted to 7.2. Dye loading and experiments were carried out at room temperature (19OC). [Ca2+], was calculated from the fluorescence signals (F) at excitation wavelengths of 360 and 380 nm according to the formula [Cal, = K&[(R - R,i,)/(R,,, - R)], where & is the dissociation constant and R is F,,, nm/F360 nm (4). Kd was assumed to be 200 nM (4), and Rmin and R,,, were obtained by intracellular calibration (35). /3 was determined to be 0.963 (in solutions), indicating that in our system the excitation wavelength of 360 nm is isosbestic. Experimental setup and digital imaging methods. The experimental setup for fluorescence microscopy and digital imaging of [Ca2+]i has been described in detail previously (5, 35). Briefly, the main components of the system are a Nikon Diaphot inverted microscope, a charge-coupled-device (CCD) camera (Photolux, Photonic Science, Tunbridge Wells, UK) fiber optically coupled to a microchannel plate intensifier (XX 1381,

(:A’+

RELEASE

IN SMOOTH

MUSCLE

H577

Philips Electronic Instruments, Mahwah, NJ), and a real-time image processor (series 151, Imaging Technology, Woburn, MA) under the control of a microcomputer. Images were obtained at video frame rate (30 Hz) and were stored in digital form on a real-time video disk storage system (model 8300 RTD, Applied Memory Technology, Tustin, CA). Computer programs for the data acquisition and analysis were written using the programming language C and the library of subroutines from the series 151, ITEX 151. Fluorescence of fura- was excited at wavelengths of 360 and 380 nm and recorded at the wavelength of 510 nm. Images of [Ca2+li were obtained by ratioing background-subtracted fluorescence images (single frames or average of 2 successive video frames) obtained at an excitation wavelength of 380 nm (Ca2+-sensitive wavelength) with low-noise images of the fluorescence at 360 nm (averages of 256 consecutives video frames) obtained at several times during an experiment. As is typical for smooth muscle cells in culture, A7r5 cells lose their ability to contract. Therefore, ratioing images that were not recorded at exactly the same time are not complicated by motion artifacts. A constant value corresponding to the average fluorescence after removing the cells from the field of view obtained at both excitation wavelengths was subtracted as background. To improve the signal-to-noise ratio, all of the images were filtered spatially by averaging a matrix of 2 X 2 pixels of a single video frame. Some images were averages of two successive video frames, and some were subjected to additional low-pass filtering. Video imaging of intracellular furafluorescence. Imaging rapid changes in [Ca2+Ji in small single cells requires a recording system that should be fast, very sensitive, and free of geometric distortion. To date, intensified CCD video cameras are probably the best compromise in this regard, but they are far from optimal. A major shortcoming is their limited dynamic range. In our system, fluorescence images recorded from single cells are digitized as B-bit images, thus representing 256 intensity levels.

Fig. 1. Imaging of fura- fluorescence and [Ca2+li in A7r5 cells with intensified charge-coupled-device camera. A: background-subtracted fura- fluorescence images (average of 64 consecutive video frames; 380-nm excitation wavelength; fura- salt loaded) recorded at low (left) and high (right) gain of intensifier/camera system. At low-gain setting, fluorescence intensity at any location in cell did not exceed a pixel value of 180 and was thus in linear response range of recording system. At high-gain setting, cell edges could be recorded with a sufficient signal-to-noise ratio, but bright central region of cell saturated intensifier/camera system. B: images of (Ca*+]i calculated from fluorescence images recorded at low- and high-gain settings at excitation wavelengths of 380 (see Fig. 1A) and 360 nm (images not shown). Images are thresholded by setting pixels with a gray level 54 to 0 (threshold value of 4 corresponds to -2% of maximal brightness encountered in image). [Ca*+h image calculated from low-gain fluorescence images (left) represents [Caz+h in central region correctly, but information in cell’s periphery is lost. It also can be seen that nuclear [Caz+] is not significantly different from cytoplasmic [Ca2+], although fluorescence image reveals considerable differences in intensity. [Ca2+li is represented correctly in peripheral regions of cell when measured at high gain (middle), but information on central region cannot be used owing to saturation of recording system. Composite [Ca*+]i image (right), in which missing pixels in low-gain image were substituted from high-gain [Ca2+]i image, shows that in resting A7r5 cell [Ca2+li is distributed homogeneously throughout cell. Calibration bar corresponds to 20 pm horizontally and 20 pm vertically. Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

H578

[CA2+]l

WAVES

AND

CA

2+-INDUCED

However, the output of the camera/intensifier system is related linearly to the input intensity only up to a pixel value of 180, which in practice further limits the dynamic range of the system. The differences in furafluorescence intensity recorded ifferent regions of a single vascular smooth muscle cell can be larger than those covered sufficiently by the dynamic response of the camera/intensifier system. This is illustrated in Fig. 1A. Figure 1A (Left) shows a fura- fluorescence image of an A7r5 cell in which all pixel values are within the linear recording range of the camera (i.e., maximum gray level l h. The effect of caffeine was largely reversible. A similar effect was observed (Fig. 3E; n = 4 cells) when A’7r5 cells were stimulated with AVP in the presence of ryanodine (10 PM). In contrast to the effect of caffeine, however, the effect of ryanodine was irreversible (17). The experiments illustrated in Fig. 3 established that 1) the response to AVP is mediated by IP3 because the response is blocked by heparin; 2) IP, releases Ca2+ from internal stores, and Ca2+ entry is not involved directly 3) positive and negative feedCCa2+-free experiments); back mechanisms on Ca2+ release are present (because [Ca2+]; oscillates; 1, 3, 14); and 4) the caffeine- and ryanodine-sensitive store is related functionally to the IP3sensitive store because caffeine and ryanodine alter the IPR-mediated release of Ca 2+. Furthermore, the observa-

RELEASE

IN SMOOTH

MUSCLE

H579

tion that caffeine and ryanodine abolish the release of Ca2+ implies that CICR might exist in these cells. This issue is controversial because evidence has been provided that CICR is not present in vascular smooth muscle (10, 27) and that in smooth muscle two distinct types of Ca2+release molecules (channels), IP,-receptors and ryanodine-receptors, coexist (16, 17). [Ca2+j; maues in A7r5 ceLLs. We observed through the microscope that [ Ca2+]i waves, lasting - 1 s, were elicited by application of AVP in high concentrations. To investigate signal-transducing mechanisms on this time scale, we used a high temporal resolution imaging system for fura- fluorescence. Rapid application of AVP to the bath caused a rise of [Ca2+]i first at the cell’s periphery from where a rapid elevation of [Ca2+]i then moved concentrically toward the center of the cell (Fig. 4A; n = 17 experiments). It seemed possible that this spatiotemporal pattern of [Ca2+]i simply reflected the geometry of the cells and the fact that IP3 is produced only in the surface membrane. (Because the cells are very thin at their edges, the [Ca2+]; tended to rise most rapidly in those regions, where the surface-to-volume ratio is high.) One way to exclude this possibility would be to expose only a small area of the cell to AVP. If the [Ca2+]; spread more rapidly across the cell than it could diffuse, then this possibility, as the sole explanation of the phenomenon, could be rejected. When this experiment was done, by pressured ejection of AVP from a micropipette electrode, [ Ca2+]; rose initially only in the circumscribed area in which the agonist had been ejected (Fig. 4B). The focal increase of [Ca2+]i then gave rise to an elevation of [Ca2+]i that moved at constant velocity along the longitudinal axis of the cell. The average velocity in five cells was 15.9 ,um/s (range 8.9-20.9 pm/s; n = 5; linear regression, r = 0.99). The fact that the data could be fitted well by a linear regression indicates that the propagation velocity is constant (see also Fig. 7). The constant velocity by which the change in [ Ca2+]; moved across cellular regions of different thickness argued against the simple geometrical explanation given above and implied a positive feedback transduction mechanism as in a [Ca2+]i wave. Nevertheless, more detailed experimentation was required to establish the exact nature of the change in [Ca2+];. Models of the spatiotemporal distribution of [Ca2+]i. We tested four explicit hypotheses on the spatiotemporal distribution of [ Ca2+]i that could be elicited by focal application of AVP (Fig. 5). These can be distinguished according to whether Ca2+ or IP3 is the primary intracellular messenger that undergoes diffusion and whether a positive feedback mechanism for the release of Ca2+ is involved. (The previous experiments suggested that positive feedback was present.) In all cases, the spatiotemporal change in [ Ca2+]i is assumed to be initiated locally, in the area in which IP3 is produced and releases Ca2+ from the IP,-sensitive Ca2+ store. ModeZs I and II are not true [Ca2+]; waves in that positive feedback is not involved, but mathematical analyses showed that the spatial pattern of [ Ca2+]i predicted by these two models can be difficult to distinguish from true [Ca2+3; waves, particularly in small cells, and therefore these models had to be considered. In model I, Ca2+ moves only by diffusion

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

0.2

__.

0.4

0.6

0.8

2#8

2

3,6

4,4

5.6

c

&3,4

8.6

0,B

1,B

2.2

;.,

Fig Images of spatiotemporal distribution of [Ca’+], in A7r5 cells. Two successive video frames recorded at video rate (33-ms intervals) were averaged and low-pass filtered. Calibration bar is 20 pm. Nos. beneath images in A-C give time at which image was obtained, in seconds after onset of response. Pseudocolor mapping of [Ca’+li (in nM): dark blue, 0; red, 600 (A and B); and 450 (C), respectively. A: [Ca2+]; transient induced by AVP applied in bath. Ca2+ indicator was introduced by internal perfusion with fura- salt. B: [Ca”‘]; wave triggered by focal application of AVP. Control image (B, top left) shows cell at rest, position of pipette containing AVP (3 wM) to be ejected, and direction of rapid bath flow, which provided that only left end of cell was in contact with AVP. C: spatiotemporal pattern of [Ca2+li in cell in which intracellular caffeine-sensitive Ca2+ store had been depleted by exposure to 10 mM caffeine for >60 min before experiment. Focal increase in [Ca2+li was achieved by wounding cell membrane with tip of microelectrode. Position of microelectrode is indicated in image 0. Cells shown in B and C were fura- acetoxymethyl ester loaded. (As outlined in METHODS, images shown represent changes in [Ca”]; in central, spindle-shaped region of cell. Due to low signal-to-noise ratio of fluorescence signal, [Caz+li could not be imaged reliably in very thin peripheral regions of cell; see also Fig. 1.)

along its concentration gradient from the area of localized initial increase in [Ca2+]i. In model 11, IPs diffuses from the site of its initial production to other areas of the cell, where it releases Ca2+ from IPa sensitive stores. Model III produces a true [Ca2+]i wave and represents a modified ICC model (24-26) in which the rise in [Ca2+]i creates a positive feedback loop by further amplification of the production of IPa through stimulation of phospholipase C by Ca2+. To explain [Ca2+]i waves elicited by focal application of AVP by this model, it must be assumed that

Ca2+ can stimulate IP, production even in the absence of hormone-receptor interaction. Model IV represents a mechanism in which positive feedback is provided by CICR (11) and a [Ca2+]i wave propagates via diffusion of Ca2+ to distant Ca2+-release sites at which it can induce further release. These hypotheses were evaluated by comparing spatiotemporal distribution of [Ca2+]i predicted by a mathematical model with measured spatiotemporal patterns of [Ca2+]i under various experimental conditions. The first

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

[CA2+lI

Spatio-temporal

pattern

WAVES

AVP-R -+G

G +

-+

pLc

CA 2+-INDUCED

cA2+

RELEASE

IN SMOOTH

Ca2+-release

-b

Ca2+-diffusion

IP,-diffusion

___)

Ca2+-release

__+

Ca2+-diffusion

H581

MUSCLE

of [Ca2+li

2 AVP-R-+

AND

‘J

/, PLC b

DAG IP,

___)

DAG IP

N 3b

Ca2+-release

DAG AVP-R -+G

-+

PLC p LIP,

>DAG --+

Ca2+-release A

_I+

IP,

ICC

Ca2+

CICR

PLC YIP, 1

pDAG

IV

AVP-R+G

-+

PLC %P,

,---b

Ca2+-

ease __+

Fig. 5. Putative cellular mechanisms of spatiotemporal [Ca2+]i transients, A7r5 cells. AVP-R, vasopressin-receptor complex in surface membrane; phospholipase C; DAG, diacylglycerol; IP3, myo-inositol-1,4,5-trisphosphate; Ca2+-induced Ca2+ release .

point to be established was whether simple diffusion of Ca2+, away from an area of high concentration, could, in practice, be reliably distinguished from a [ Ca2+]i wave, which propagates via a positive feedback on Ca2+ release. To observe the spatiotemporal pattern of [Ca2+]; that would arise from diffusion, it was necessary to abolish the activity of cellular processes that could take up or release Ca2+. We incubated A’7r5 cells with caffeine (10 mM) for 1 h and then observed the spatiotemporal pattern of [Ca2+]i in these cells (which had not been stimulated with AVP previously) when [ Ca2+]; was increased focally: Fig. 4C shows images of [Ca2+]i of a cell in which [Ca2+]; had been raised focally by penetrating the cell membrane (wounding) with a micropipette (the electrode was kept in place during the experiment to ensure a continuous source of Ca2+). Figure 6A shows plots of [Ca2+]i along a single line of pixels (line plot or [ Ca2+]i profile) from the same cell. Figure 6B illustrates the theoretical spatiotemporal distribution of Ca2+ produced by diffusion of Ca2+ away from a region of elevated [ Ca2+]i (simulating the micropipette). The experimentally observed profile (Fig. 6A) and diffusional profile predicted by the mathematical model are distinctively different from the [ Ca2+]i profile in a cell stimulated focally with AVP (Fig. 6C). Here, [ Ca2+]i rises rather rapidly from resting levels (Fig. 6C, 1) and forms a steep wave front (Fig. 6C, 2) that propagates at constant velocity across the cell; [Ca2+]i then levels off at an elevated plateau behind the wave front (Fig. 6C, 3). This [Ca2+]i profile compares well with

Ca2+-diffusion

triggered by focal application of AVP, in G, GTP-binding coupling protein; PLC, ICC, IP3-Ca2+ cross coupling; CICR,

that predicted by a mathematical model that simulated a [Ca2+]i wave propagated via CICR (shown in Fig. 6D). Figure 6E shows the control experiment, in which a normal (Ca2+ stores not depleted) A7r5 cell was wounded with a micropipette in the same way as described above (n = 3 cells). The propagating wave front reveals the same characteristics as the wave observed on focal stimulation with AVP (Fig. 6C) but is clearly different from the diffusional pattern (Fig. 6, A and B). Figure 6F represents the prediction from a model that includes a localized source of Ca2+ (wounding pipette) and CICR. If CICR is eliminated from the model, the [Ca2+]; profiles show the typical diffusional pattern, which are represented by the dashed lines. The wounding experiment also was done in cells (Fig. 6G; n = 3 experiments) that had been perfused internally with heparin to block the release of Ca2+ by IP,; the observed [Ca2+]i profiles were very similar to those shown in Fig. 6E. The contour plots shown in Fig. 6H are the result of a simulation of a [ Ca2+]i wave from a two-dimensional mathematical model. This model incorporated the focal stimulation with AVP (localized release of Ca2+ from an IP3-sensitive store), diffusion of Ca2+, and CICR. The [Ca2+]i wave initially spreads in a circular fashion. As soon as the wave front encounters the cell border, the wave front rapidly becomes straight and propagates further at a constant velocity across the cell. This spreading pattern was very similar to that observed in the experiments in which A7r5 cells were stimulated by focal application of AVP (see Fig. 4B).

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

H582

[cA2+lI

WAVES

AND

CA 2+-INDUCED

cA2+

RELEASE

IN SMOOTH

MUSCLE

A 5oor PaI i (n M)

0

C

5OOr

Distance (pm)

100

Distance (pm)

D

3

Oi 0

Distance (pm) F

E SOOr

5oor

Distance (pm)

100

500

[Cali (n w

Distance (pm) G

Distance (pm)

H

500

Distance (pm) Fig. 6. [ Ca2+]i profiles recorded along longitudinal axis of cells (left column) and [Ca”+] profiles predicted by mathematical model (right column) incorporating diffusion and intracellular release of Ca2+. A, E, and G, [Ca2+]i profiles recorded from cells in which rapid and focal increase of [Ca2+]i was achieved by wounding cell membrane with tip of microelectrode in presence of 2 mM extracellular [Ca2+] ([Ca”+]J. Profiles show [Ca2+]i at rest, [Ca2+]i immediately after wounding cell membrane, and series recorded at 333-ms intervals after wounding. Arrowhead, position of micropipette. A: wounding experiment in cell pretreated with 10 mM caffeine for >60 min (same cell as shown in Fig. 4C). B: [Ca2+]i profiles predicted by mathematical model in which it was assumed that [Ca2+]i changed exclusively by diffusion from small region of cell in which it had been released initially by IP,. C: [Ca2+]i profiles recorded from cell shown in Fig. 4B, which had been stimulated focally with AVP. 1, rising phase; 2, wavefront; 3, plateau of elevated [Ca2+]i. D: [Ca2+]i profiles predicted by mathematical model in which it was assumed that Ca2+ was released from internal stores by CICR when [Ca2+]i rose above defined threshold value. In this model, [Ca2+]i wave was initiated by raising [ Ca2+]i in small peripheral region of -3% of total cell volume. This simulated localized release of Ca2+ by IP3. Propagation of elevated [Ca2+]i results from combination of diffusion of Ca2+ and release of Ca2+ from sarcoplasmic reticulum. Model reproduces diffusional foot, which is maintained throughout cell. E: [Ca2+]i profiles from wounding experiment in cell bathed in standard Tyrode solution (2 mM [Ca”+]J. F: [Ca2+]i profiles (solid curues) predicted by model that incorporates constant source of Cai+ (wounding pipette) and CICR. Dashed curues, predicted [Ca2+]i profiles for same time intervals when CICR was eliminated from simulation. In this case profiles show typical diffusional pattern. G: same type of experiment as shown in E but carried out in cell in which binding of IP3 to its receptor was blocked by internal perfusion with heparin. H: contour plots of changes in [Ca2+]i predicted by Sdimensional model that incorporates local IP3-sensitive release of Ca2+ (in region of 8 x 8 volume elements at left end of model cell corresponding to - 1% of total cell volume), CICR, and diffusional Ca 2+ fluxes. Plots l-18 separated by 333 ms. [Ca2+]i between 2 iso-[Ca2+]i lines is 25 nM (cell length, 120 pm; parameters for CICR and resting [Ca2+]i same as in D).

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

[cA2+lI

WAVES

AND

2+-INDUCED

CA

To further distinguish between mechanisms that involve positive feedback for the release of Ca2+ or solely diffusion of Ca2+, we analyzed the propagation velocities of [Ca2+]i waves observed under the above experimental conditions. The propagation velocity was measured by recording time and distance in which [Ca2+]i rose to half the final plateau value of a [ Ca2+]; wave. Because no plateau of elevated [Ca2+]; could be observed in the experiments shown in Fig. 6A, the velocity was measured when [Ca2+]; rose to 200 nM. The level of 200 nM corresponds to the average half-plateau value of a [Ca2+]i wave. Figure 7A shows a plot of the propagated distances as a function of time (= velocity) observed in cells that had been stimulated focally with AVP (n = 5 experiments) and had been wounded with a micropipette under control conditions (n = 3 experiments), during internal perfusion with heparin (n = 3 experiments), and after exposure to caffeine (n = 3 experiments). Figure 7B shows the average propagation velocities (in pm/s) ob-

A

0 w, control + w, hepafin l w, caffeine o AVP (focal)

100

0 +

Distance wo

6

Time (set)

B

30

Velocity (km/set)

0

AVP (focal)

W w control heparin

W

caffeine

Fig. 7. A: propagation velocity. Distance traveled across A7r5 cell by front of elevated [Ca2+]i plotted as function of time. 0, expts in which [Ca2+]i wave was triggered by focal stimulation with AVP (n = 5). +, solid symbols, expts in which localized increase of [Ca2+]i was produced by wounding cell with micropipette. This was done under control conditions (+, n = 3), during internal perfusion with heparin (+, n = 3), and after preincubation with caffeine (m, n = 3). B: bar graph of average propagation velocity (error bars, SD) of [Ca2+]i waves induced by focal stimulation with vasopressin (AVP, focal), and by wounding (w) cells with micropipette under control conditions, during internal perfusion with heparin, and in presence of caffeine. Statistics: no assumptions regarding distribution of data were made, and due to small sample sizes we used a nonparametric procedure to compare differences of means for statistical significance. We used a l-tailed Wilcoxon rank sum test (Mann-Whitney U-test) for independent samples. * P 5 0.02, statistically significant difference.

cA2+

RELEASE

IN SMOO’

H583

H MUSCLE

400

300

[c 1 ai

(nM)

2oo 100

0

5

Time (set) Fig. 8. Time course of rise of [Cali triggered by focal application of AVP. [Ca2+]i was measured at video frame rate (33 ms) in small (15 pm2) cytoplasmic region of cell. During passage of [Ca2+]i wave, [Ca2+]i rose at maximal rate of -1 PM/S until it reached a new maintained (plateau) level. After passage of wave front, [Ca2+]i decreased at much slower rate. Rapid rise of [Ca2+]i was consistently preceded by smaller and slower increase of [Ca”+]i, the diffusional foot (arrowhead).

served under these different experimental conditions. The propagation velocities measured in the experiments with focal stimulation with AVP, after wounding under control conditions, and after internal perfusion with heparin are 15.9, 18.9, and 15.3 pm/s, respectively, and are not significantly different. The propagation velocities in the above experiments were constant because in each experiment the distance/time plots can be fitted well by linear regression (r varies between 0.98 and 1.00). In contrast, the propagation velocity of elevated [ Ca2+]; measured from cells that had been wounded after incubation with caffeine was on average only 4.8 pm/s and statistically significantly lower (P = 0.02, one-sided MannWhitney U-test for independent samples). The fit of the data by linear regression was not as good but was still considered good (average r = 0.95). Further support for the involvement of CICR in the propagation of AVP-induced [Ca2+]i waves came from the detailed analysis of the onset of the wave. [Ca2+]i was measured in a small cytoplasmic region (15 hm2) of the cell and was plotted as a function of time. The sharp rise in [ Ca2+]i is preceded consistently by a small gradual increase of [ Ca2+]; that we refer to as the “diffusional foot” (Fig. 8, arrowhead). This is consistent with the idea that a small amount of Ca2+ diffuses in front of the wave due to the large concentration gradient and facilitates release of Ca2+ from distant Ca2+-sensitive release sites. The average maximal rise in [Ca2+]i measured in the steep linear area of the wave front was 722 PM/S (range 0.4-1.0 PM/S). DISCUSSION

Response of A7r5 cells to stimulation with A VP, Stimulation of A7r5 cultured smooth muscle cells with vasopressin increased [ Ca2+] i in a concentration-dependent manner. Furthermore, the time interval between exposure to the agonist and the peak of the [Ca2+]; transient is also concentration dependent. We could exclude the possibility that this delay is related to diffusion of the AVP in the bath by placing the cell into a rapid flow that provides exchange of the extracellular solution in (1 s. It

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

H584

[CA2+lr

WAVES

AND

CA 2+-INDUCED

is much more likely that this delay reflects the physiological processes that are involved in relating receptor binding of the agonist and the release of Ca2+ from internal stores. Berridge et al. (2) have pointed out that this delay could depend on a rate-limiting step associated with the receptor transduction process that generates the appropriate second messenger or, alternatively, that the resting level of the messenger is somewhat below the threshold necessary to trigger the release; the delay (or latency) then dep#endson how lo w it takes to increase the concentration of the messenger to this threshold level. The response to AVP is mediated by IP3 because it can be blocked completely by intracellular application of heparin. Removal of Ca”,+ showed that AVP acts through Ca2+ release from intracellular stores, and the occurrence of [ Ca2+] i oscillations indicates the existence of positive and negative feedback : mechanisms on Ca2+ release. Furthermore, the fact that caffeine and ryanodine are able to abolish the release of Ca2+ triggered by AVP 1) indicates that in A7r5 cells the IP,-sensitive store is not independent from a caffeine- and ryanodine -sensitive store, and 2) implies the presence of CICR. Cellular mechanism of [Ca’+]; waves in A7r5 cells. Using digital imaging of [Ca2+]i, we observed that a focal increase of [ Ca2+] i gave rise to [ Ca2+]i waves that propagated at constant veloc ity throughout the entire cell.1 A localized increase of [Ca2+]i was achieved either by focal application of AVP, by wounding the cell membrane with the tip of a microelectrode, or (in 1 experiment) by breaking the seal of a patch electrode containing millimolar [Ca2+]. We tested four models for intracellular mechanisms underlying the spatiotemporal distribution of [Ca2+]; observed after focal application of AVP. These models are summarized in Fig. 5. Our experiments provide clear evidence against model I (simple diffusion of Ca2+). A localized increase in [ Ca2+]i due to either focal stimulation with AVP or wounding of the cell membrane with a micropipette typically triggers a propagated [ Ca2+.]i wave with the typical characteristics (see Fig. 6C) of a rapid rise from resting levels, the steep wave front, and the sustained plateau of elevated [ Ca2+]i that eventually recovered at a much slower rate after passage of the wave front. Under conditions in which the release and uptake of Ca2+ by a caffeine-sensitive Ca2+ store was disabled by exposure to 10 mM caffeine (Figs. 4C and 6A), the spatiotemporal pattern of [Ca2+]; was distinctly different: the change in [Ca”+]; showed a typical diffusional profile that lacked the elevated plateau of increased [Ca2+]i and that spread at a significantly lower velocity. Models II and III are unlikely to represent the primary mechanism by which AVP-induced [Ca2+]i waves propagate because [Ca2+]i waves can be triggered even when the IP, receptor is blocked (Fig. 6G), and the mathematical model shows that simple diffusion of IP3 1 As shown in Fig. 4B, the [Ca2+]i wave tends to slow down slightly during passage through the nuclear region (see also Ref. 31). This finding is not inconsistent with CICR because in this region the density of Ca2+, stores is lower. This is also well reproduced by the threedimensional mathematical model (data not shown). Although it is unlikely, the possibility that the nucleus forms a diffusional barrier for Ca2+ cannot be excluded .

cA2+

RELEASE

IN SMOOTH

MUSCLE

cannot produce a [Ca2+]; wave that propagates at constant velocity. (The [Ca2+]i wave would slow down as a consequence of the pattern of diffusion of IP3.) The observation that in the presence of caffeine a localized increase in [Ca2+]i spreads throughout the cell at a significantly slower rate argues in favor of the hypothesis that by exposure to caffeine a positive-feedback mechanism for the release and/or uptake of Ca2+ had been removed. However, AVP still is able to trigger [Ca2+]i transients in the presence of caffeine or ryanodine (at least initially), even in cells that had been incubated with caffeine or ryanodine for >l h. Repetitive stimulation with AVP in the presence of caffeine or ryanodine progressively abolished the [ Ca2+]i transients. In summary, [Ca”+]; waves can be triggered by a localized increase in [Ca2+]i in these cells in the total absence of AVP. That these spatiotemporal patterns of [Ca’+]; are actually waves and involve positive feedback is shown clearly by the fact that they are distinctly different from patterns produced by diffusion of Ca2+ alone. Such [Ca2+]i waves could be explained by the hypothesis that Ca2+ diffuses to sites at which it activates phospholipase C and that the wave therefore propagates by Ca2+-activated IP3 production and IP,-mediated release (model III), but similar patterns are observed even when IP3mediated release is blocked by heparin. Thus we favor model IV, in which Ca2+, released by IP3, diffuses to sites at which it can release Ca2+ by CICR. CICR in vascular smooth muscle. The typical profile of a [Ca2+]i wave elicited by focal AVP is constant throughout the entire cell and has a feature that is consistent with CICR. The presence of the diffusional foot that precedes the sharp rise of [Ca”+]; (Fig. 8) is consistent with a mechanism by which Ca2+ diffuses along its concentration gradient and, as soon as a critical threshold concentration is reached, triggers its own release. Release is presumed to be through a caffeine- and ryanodine-sensitive Ca2+-release channel. In this regard, the maximum rate at which [ Ca2+]i can rise from resting levels is of particular interest. The maximum rate of rise of [ Ca2+]i that we observed in cells that had been stimulated focally with AVP was - 1 PM/S. (This is a conservative estimate because the measurements were taken from [Ca2+]i images in which changes of [Ca2+]i are averaged over 33 ms). This rate of change in [Ca2+]i is comparable to the lowest rates of change that caused a measurable development of tension in cardiac cells due to CICR (12). The presence of CICR in smooth muscle is controversial. Ehrlich and Watras (10) reported that IP3 but not caffeine was able to release Ca2+ from vesicles of aortic SR and to induce channel opening when the vesicles were incorporated into a planar bilayer. In contrast, Kobayashi et al. (21) have reported a caffeine-sensitive Ca2+ store in permeabilized rat arterial smooth muscle cells. Furthermore, caffeine induces contraction in arterial ring preparations and attenuates agonist-induced contraction and elevation of [Ca2+]i in vascular smooth muscle preparations (7). Nevertheless, Missiean et al. (27) concluded, from the lack of effect of ryanodine and caffeine on 45Ca2+ fluxes induced by AVP, that CICR is not present.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

[CA~+]~ WAVES

AND

CAM+-INDUCED

cA2+

RELEASE

IN SMOOTH

MUSCLE

H585

Shin et al. (30) have even shown that, in the same pop- bath application of AVP, IP3- and Ca2+-sensitive release ulation of rat vascular smooth muscle cells, one sub- of Ca2+ act in concert, thus leading to a higher level population of cells (in the presence of Ca2+ ) responded to Of[Ca2+]i. a stimulation with caffeine or angiotensinP1 with a tranWe thank Drs. Joseph Kao and Don Gill for valuable comments on sient increase of [ Ca2+]i, whereas another subpopulation the manuscript. responded only to the hormonal stimuli and was unreFinancial support was provided by National Heart, Lung, and Blood Institute Grant HL-29473 (W. G. Wier) and by the American Heart sponsive to the caffeine stimulation. Association, Maryland Affiliate (L. A. Blatter). It has been proposed that IP,-sensitive and -insensitive Part of this work has been published in abstract form (6). Ca2+ stores can coexist in the same cell (1, 3). It is genAddrc-s for reprint requests: L. A. Blatter, Dept. of Physiology, Univ. erally assumed that these pools are, at least functionally, of Maryland School of Medicine, 660 W. Redwood St., Baltimore, MD separate entities. In this two-pool model, IP3 releases 21201. 3040% of the total releasable Ca2+ from nonmitochonReceived 31 May 1991; accepted in final form 7 April 1992. drial Ca2+ pools. The anatomic location and identity of the IP,-sensitive and -insensitive pools remains uncer- REFERENCES tain (1, 3). In contrast, it has been reported that smooth 1. Berridge, M. J. Calcium oscillations. J. Biol. Chem. 265: 95839586, 1990. muscle has two Ca2+ pools, one of which is sensitive only 2. Berridge, M. J., P. H. Cobbold, and K. S. R. Cuthbertson. to IP3 and the other of which is sensitive to both IP3 and Spatial and temporal aspects of cell signalling. PhiZos. Trans. R. CICR (16,17). The relative amounts of the two pools vary Sot. Lond. B Biol. Sci. 320: 325-343, 1988. considerably among different types of smooth muscle 3. Berridge, M. J., and R. F. Irvine. Inositol phosphates and cell signalling. Nature Lond. 341: 197-205, 1989. (17). Our results support a two-pool model in which the D. J., and W. G. Wier. Mechanism of release of two pools are interdependent. Caffeine is able to deplete a 4. Beuckelmann, calcium from sarcoplasmic reticulum of guinea-pig cardiac cells. J. caffeine- (and Ca2+-) sensitive pool but affects the IP3Physiol. Lond. 405: 233-255, 1988. sensitive pool only indirectly. Nevertheless, the IP3-sen5. Blatter, L. A., and W. G. Wier. Intracellular diffusion, binding, sitive pool becomes depleted in the presence of caffeine and compartmentalization of the fluorescent calcium indicators indo-l and fura-2. Biophys. J. 58: 1491-1499, 1990. (or ryanodine) when the release from this pool is triggered 6. Blatter, L. A., and W. G. Wier. Focal application of vasorepeatedly by stimulation with AVP. The possible interpressin to vascular smooth muscle cells triggers calcium-waves as action between the two pools may occur on the level of revealed by digital imaging microscopy. Biophys. J. 59: 235a, 1991. restoring resting [ Ca2+]i after stimulation with AVP: 7. Bova, S., W. F. Goldman, X.-J. Yuan, and M. P. Blaustein. Influence of Na+ gradient on Ca 2+ transients and contraction in Ca2+ released from the IP,-sensitive store is taken up by vascular smooth muscle. Am. J. Physiol. 259 (Heart Circ. Physiol. both pools and also is transported across the cell mem28): H409-H423, 1990. brane by the Ca2+ pump and Na-Ca exchange. The pres8. Cuthbertson, K. S. R., and P. H. Cobbold (Editors). Oscillaence of caffeine or ryanodine, however, keeps the Ca2+tions in cell calcium (collected papers and reviews). Cell Calcium 12: 61-268, 1991. sensitive store in a “leaky” state that, as a consequence of 9. Doyle, V. M., and U. T. Ruegg. Vasopressin induced producrepetitive stimulation, leads to a net loss of Ca2+ and tion of inositol trisphosphate and calcium efflux in a smooth consequently also to a depletion of the IP,-sensitive store. muscle cell line. Biochem. Biophys. Res. Commun. 131: 469-476, An alternative explanation for the effect of caffeine is 1985. that caffeine alters the threshold amount of IP, required 10. Ehrlich, B. E., and J. Watras. Inositol 1,4,5-trisphosphate activates a channel from smooth muscle sarcoplasmic reticulum. to release Ca2+ from internal stores, as has been observed by Parker and Ivorra (29) in Xenopus oocytes. In A7r5 11. Nature Lond. 336: 583-586, 1988. Fabiato, A. Calcium-induced release of calcium from the cardiac vascular smooth muscle cells, this mechanism seems unsarcoplasmic reticulum. Am. J. Physiol. 245 (Cell Physiol. 14): likely because the amplitude of the [Ca2+]i transient Cl-c14, 1983. evoked by the first application of AVP in the presence of 12. Fabiato, A. Time and calcium dependence of activation and inactivation of calcium-induced release of calcium from the sarcaffeine was unchanged (see Fig. 30). This was indecoplasmic reticulum of a skinned canine cardiac Purkinje cell. J. pendent of the fact that the cells were exposed to caffeine Gen. Physiol. 85: 247-289, 1985. for only a few minutes or were incubated with caffeine 13. Fleischer, S., and M. Inui. Biochemistry and biophysics of for >l h. excitation-contraction coupling. Annu. Rev. Biophys. Biophys. Chem. 18: 333-364, 1989. In conclusion, we have found that AM cells possess 14. Harootunian, A. T., J. P. Y. Kao, S. Paranjape, and R. Y. both IP3- and Ca2+-sensitive Ca2+ release. Stimulation Tsien. Generation of calcium oscillations in fibroblasts by posiwith AVP first triggers IP,-sensitive release of Ca2+ from tive feedback between calcium and IP3. Science Wash. DC 251: internal store. The increased [Ca2+]i then activates 75-78, 1991. A. L., and R. D. Keynes. Movements of labelled CICR. CICR is the positive feedback mechanism by 15. Hodgkin, calcium in squid giant axons. J. Physiol. Lond. 138: 253-281, 1957. which a localized increase of [Ca2+]; can propagate as a Iino, M. Calcium dependent inositol trisphosphate-induced cal[Ca2+]; wave throughout the entire cell. Further evidence 16. cium release in the guinea-pig taenia caeci. Biochem. Biophys. Res. for this hypothesis comes from the observation that the Commun. 142: 47-52, 1987. plateau [ Ca2+]i of a [ Ca2+]i wave triggered by focal appli17. Iino, M., T. Kobayashi, and M. Endo. Use of ryanodine for functional removal of the calcium store in smooth muscle cells of cation of AVP (Fig. 4B) is consistently lower than the the guinea-pig. Biochem. Biophys. Res. Commun. 152: 417-422, maximal increase in [Ca2+]i that was observed when 1988. the whole cell was exposed to the agonist (Fig. 4A). A 18. Irvine, R. F., and R. M. Moor. Micro-injection of inositol possible explanation for this observation is that the pla1,3,4,5-tetrakisphosphate activates sea urchin eggs by a mechateau TCa2+li is solely the result of CICR, whereas, during nism dependent on external Ca 2+. Biochem. J. 240: 917-920,1986.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

H586

[CAM+]* WAVES

AND

CAM+-INDUCED

19. Irvine, R. F., R. M. Moor, W. K. Pollock, P. M. Smith, and K. A. Wreggett. Inositol phosphates: proliferation, metabolism and function. Philos. Trans. R. Sot. Lond. B Biol. Sci. 320: 281298, 1988. 20. Kimes, B. W., and B. L. Brandt. Characterization of two putative smooth muscle cell lines from rat thoracic aorta. Exp. CeZZ Res. 98: 349-366, 1976. and M. Nakamura. Complete 21. Kobayashi, S., H. Kanaide, overlap of caffeine- and K+ depolarization-sensitive intracellular calcium storage site in cultured rat arterial smooth muscle cells. J. BioZ. Chem. 261: 15709-15713, 1986. 22. Kobayashi, S., A. V. Somlyo, and A. P. Somlyo. Heparin inhibits the inositol 1,4,5trisphosphate-dependent, but not the independent, calcium release induced by guanine nucleotide in vascular smooth muscle. Biochem. Biophys. Res. Commun. 153: 625-631, 1988. M. J., and R. J. Podolsky. Ionic mobility in 23. Kushmerick, muscle cells. Science Wash. DC 166: 1297-1298, 1969. 24. Meyer, T. Cell signaling by second messenger waves. CeZZ 64: 675-678, 1991. T., and L. Stryer. Molecular model for receptor-stim25. Meyer, ulated calcium spiking. Proc. NatZ. Acad. Sci. USA 85: 5051-5055, 1988. T., and L. Stryer. Calcium spiking. Annu. Rev. Bio26. Meyer, phys. Biophys. Chem. 20: 153-174, 1991. L., I. Declerck, G. Droogmans, L. Plessers, 27. Missiaen, H. De Smedt, L. Raeymaekers, and R. Casteels. Agonistdependent Ca2+ and Mn2+ entry dependent on state of filling of Ca2+ stores in aortic smooth muscle cells of the rat. J. Physiol.

CAM+ RELEASE

IN SMOOTH

MUSCLE

Lond. 427: 171-186, 1990. 28. Neylon, C. B., J. Hoyland, W. T. Mason, and R. F. Irvine. Spatial dynamics of intracellular calcium in agonist-stimulated vascular smooth muscle cells. Am. J. Physiol. 259 (Cell Physiol. 28): C675-C686, 1990. 29. Parker, I., and I. Ivorra. Caffeine inhibits inositol trisphosphate-mediated liberation of intracellular calcium in Xenopus oocytes. J. Physiol. Lond. 433: 229-240, 1991. 3. Shin, W. S., T. Toyo-oka, M. Masuo M., Y. Okai, H. Fujita, and T. Sugimoto. Subpopulations of rat vascular smooth muscle cells as discriminated by calcium release mechanisms from internal stores. Circ. Res. 69: 551-556, 1991. 31 Takamatsu, T., and W. G. Wier. Calcium waves in mammalian ’ heart: quantification of origin, magnitude, waveform, and velocity. FASEB J. 4: 1519-1525, 1990. C. W., M. J. Berridge, A. M. Cooke, and B. V. L. 32 Taylor, Potter. Inositol 1,4,5-trisphosphorothioate, a stable analogue of inositol trisphosphate which mobilizes intracellular calcium. Biothem. J. 259: 645-650, 1989. 33 Tsien, R. W., and R. Y. Tsien. Calcium channels, stores, and oscillations. Annu. Rev. CeZZ BioZ. 6: 715-760, 1990. M. Lazdunski, and C. Frelin. 34. Vigne, P., J.-P. Breittmayer, The regulation of cytoplasmic free Ca2+ concentration in aortic smooth muscle cells (A7r5 line) after stimulation by vasopressin and bombesin. Eur. J. Biochem. 176: 47-52, 1988. 35. Wier, W. G., and L. A. Blatter. Ca2+-oscillations and Ca2+waves in mammalian cardiac and vascular smooth muscle cells. CeZZ CaZcium 12: 241-254, 1991. l

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (155.247.166.234) on August 1, 2018. Copyright © 1992 American Physiological Society. All rights reserved.