Flow-induced changes in Ca2+ signaling of vascular endothelial cells: effect of shear stress and ATP MAKOTO

MO, SUZANNE

G. ESKIN,

AND

WILLIAM

P. SCHILLING

Departments of Surgery and Molecular Physiology and Biophysics, Baylor College of Medicine, Houston, Texas 77030

MO, MAKOTO, SUZANNE C. ESKIN, AND WILLIAM P. SCHILLING. Flow-induced changes in Ca2+ signaling of vascular endothelial cells: effect of shear stress and ATP. Am. J. Physiol. 260 (Heart Circ. Physiol. 29): H1698-H1707,1991.-The effect of hemodynamic flow on apparent cytosolic free Ca2+ concentration ([Ca”‘]i) of cultured bovine aortic endothelial cells was examined in the absence and presence of adenine nucleotides using microfluorimetric analysis of fura- fluorescence. In the absence of adenine nucleotides, flow-induced shear stress produced little change (>> AMP. Adenosine was without effect on [Ca”‘]i. Washout of ATP resulted in the immediate return of [Ca2+]i to basal values, indicating that the effect of ATP was rapidly reversible. Decreasing the flow rate to zero during the sustained phase also resulted in an immediate decrease of [Ca”+]i. Similar results were obtained with ADP and AMP but not with the nonhydrolyzable adenine nucleotide analogues cu,P-methyleneadenosine-5’-diphosphate, ,B,y-imidoadenosineY&phosphate, or &r-methyleneadenosine#-triphosphate. Furthermore, the rate of [Ca”?]i decrease upon cessation of flow during the sustained phase of the response to ATP was inversely proportional to the ATP concentration. These results suggest that hydrolysis of ATP to adenosine by the ectonucleotidase is responsible for the termination of the ATP response under zero-flow conditions. Evaluation of the dose- and flowdependent response of the cells to ATP indicates that convective-diffusive transport of ATP may play an important role in regulation of endothelial cell [Ca2+]i in presence of ectonucleotidase activity and could have important consequences for the regulation of blood flow in the vasculature. purinoceptors; ectonucleotidase; convective-diffusive transport; fura-

CELLS are exposedto avariety of substances in the blood and are subjected to changing levels of hemodynamic shear stress depending on the vascular bed. Flow-induced morphological and functional changes reported in cultured endothelial cells include elongation and alignment of the cells in the direction of flow, rearrangement of the cytoskeleton (11, 25), and increases in 1) prostacyclin production (16), 2) tissue plasminogen activator production (l2), 3) uptake of lowdensity lipoproteins (34), 4) pinocytosis rate (9), and 5) K+ channel current (29). These effects of flow on endoVASCULARENDOTHELIAL

H1698

0363-6135/91

thelial cell function may reflect the transduction of mechanical signals from the cell surface to the cytoplasm as a result of shear stress (8). Recent studies have demonstrated flow-induced, endothelium-dependent dilatation of blood vessels (24, 31, 32) suggesting that shear stress may increase cytosolic free Ca2” ( [Ca2+]i) of the endothelial cell. Consistent with this hypothesis is a recent report by Ando et al. (1). These authors examined the effect of shear stress on [Ca2+]i using endothelial cells loaded with the fluorescent Ca2+ indicator fura-2. Increasing flow rate produced a biphasic [Ca2+]i response similar to that obtained with agonist agents. These changes in [Ca2+]i however, were not shown to be graded with respect to shear stress. It is now well established that changes in [Ca2+]i of vascular endothelial cells produced by vasoactive agents such as ATP (4, 30) cause the release of endothelialderived relaxing (10, 19, 22) and hyperpolarizing factors (5) that influence the tone of the underlying smooth muscle. This effect of extracellular ATP occurs via activation of purinoceptors of the PzY subtype (27) and subsequent stimulation of the phosphoinositol pathway (30). Termination of the ATP response is, at least in part, regulated by the activity of the nucleotidase associated with the extracellular membrane surface of the vascular endothelial cell (21). ATP and ADP are metabolized by ectoadenosinetriphosphatase and/or ectoadenosinediphosphatase (ectoADPase), whereas AMP is metabolized to adenosine via the 5’-nucleotidase (6, 20). Because the potency of ATP and its metabolites for activation of the P2+eceptor follows the sequence ATP > ADP >>> AMP > adenosine (3, 4, 21), changes in blood flow will not only affect the level of shear stress but may also change the balance between ATP delivery to the cell surface and degradation to less active or inactive hydrolysis products. In the present study, we have measured [Ca2+]i of endothelial cells under shear stress in the absence and presence of extracellular ATP using microfluorimetric analysis of fura- fluorescence. Whereas shear stress produced little change in [Ca2+]i, the effect of ATP was found to depend on both the flow rate and the concentration of ATP in the superfusion buffer. Similar flowdependent effects were not seen with the nonhydrolyzable adenine analogues a,@-methyleneadenosine-5’-diphosphate (AMP-CP), p,r-imidoadenosine-5’0triphosphate (AMP-PNP), or ,@,y-methyleneadenosine-5’0triphosphate (AMP-PCP). Thus both blood flow and

$1.50 Copyright 0 1991 the American Physiological

Society

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

SHEAR

STRESS

AND ATP ON ENDOTHELIAL

ectonucleotidase may play an important role in regulation of vascular response through the delivery and degradation of adenine nucleotides. MATERIALS

AND

METHODS

Materials. The sodium salts of ATP, ADP, AMP, and AMP-CP, AMP-PCP, the lithium salt of AMP-PNP, adenosine, Triton X-100, ethylene glycol-bis(@-aminoethyl ether)-N,N,N’,N’-tetraacetic acid (EGTA), and bovine serum albumin were purchased from Sigma Chemical (St. Louis, MO). Fura- pentapotassium salt and the acetoxymethylester of fura- (fura-2/AM) were obtained from Molecular Probes (Eugene, OR). Bradykinin was obtained from Calbiochem (San Diego, CA). Solutions. N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES)-buffered saline (HBS) contained salts in the following concentrations: 140 mM NaCl, 5 mM KCl, 1 mM MgC12,lO mM D-glucose, 1.8mM CaClg, and 15 mM HEPES, pH adjusted to 7.40 for 37°C with NaOH. The solution osmolarity was -300 mosM. Isolation and culture of vascular endothelial cells. Bovine aortic endothelial cells (BAECs) were harvested as previously described (15) and cultured using Dulbecco’s modified Eagle’s medium (GIBCO) supplemented with 10% heat-inactivated fetal bovine serum (Hyclone, Logan, UT), 50 pg/ml streptomycin, 50 pg/ml penicillin, 100 pg/ml neomycin, and 2 mM glutamine. Calf pulmonary artery endothelial cells (CPAEs) were a generous gift of Dr. Peter Del Vecchio and were cultured as previously described (13). Primary BAECs or BAECs and CPAEs in the sixth to eighteenth passage were seeded on glass slides (75 X 38 mm; Fisher Scientific) that had been pretreated for 2 h with 0.5 M NaOH. Seeding density was ~6 X lo4 cells/cm2. Seeded slides were placed in 100 mm culture dishes with 10 ml complete medium and incubated at 37°C in a humidified air atmosphere with 5% COZ. Human umbilical vein endothelial cells (HUVECs) were isolated and cultured on glass slides as previously described (12). All experiments were performed 48-96 h after seeding. Flozu system. Confluent BAEC monolayers were exposed to- controlled steady laminar flow in a parallelplate flow chamber (17) using two infusion/withdrawal syringe pumps (Harvard Medical). Shear stress (7; dyn/ cm2) was calculated by the following equation: 7 = 6Qp/ bh2, where Q equals the flow rate (ml/s), p equals the viscosity (0.01 poise), b is the width of the flow channel (1.46 cm) and h is the height of the flow channel (0.0218 cm). The volume of the flow channel was 0.127 ml. Flow rate was adjusted to 0.07, 0.35, 0.7, 2.8, 7.0, or 17.5 ml/ min to yield shear stress values of 0.1, 0.5, 1.0, 4.0, 10, and 25 dyn/cm2, respectively. In some experiments, cells were subjected to either 4.6 or 11 dyn/cm” shear stress for 24 h before measurement of [ Ca2+]i (see below) as described previously (12, 16). Measurement of [Ca2+]i. Cells on the glass slides were incubated in complete Dulbecco’s modified Eagle’s medium containing fura=2/AM (6 PM; added fro; a stock solution of fura-B/AM in dimethyl sulfoxide) for 30 min at 37OC. Cells were washed twice with HBS and incubated in HBS for at least 30 min at room temperature

CELL

H1699

[CA2+]i

before measurement of fluorescence. Cells were washed with HBS immediately before measurement to ensure removal of extracellular fura- that may have diffused out of cells after the initi .a1 loading. S lides were mounted on the flow chamber and placed on the stage of a Nikon Diaphot inverted microscope (Nikon, Garden City, NY) that was optically connected to an SLM 8000 spectrophotofluorimeter(SLM Instruments, Urbana, IL). Diffuse, uniform distribution of fura- in the cytosol was observed under high magnification (~1,000). For all experimental traces shown, measurements were performed using a Nikon x20 Fluorescence objective. The number of cells in the field of view was ~100. Excitation wavelength alternated between 355 and 385 nm (slit width of 8 nm) every 0.5 s, and emission fluorescence was recorded at 500 t 10 nm using a band-pass filter. An IBM-XT computer (IBM, Boca Raton, FL) with SLM software calculated the fluorescence ratio, R, where R equals F& FsG5,the ratio of fluorescence intensities at 355 and 385 nm excitation, respectively. Autofluorescence was determined on glass slides without cells or on slides with unloaded cells and was subtracted before calculation of R. The magnitude of the calcium-insensitive component A

200.

1pMATp

so -250 2

0 -50 .

C

1 pM ATP 1

zeroFlow

200. 150.

100 b Flow(7.oml/nh) 503 . . . 2

4

Time

.

.

6

0

(minutes)

~- .-.-

-

--

FIG. 1. Effect of extracellular ATP on [Ca’+]i under flow. A: furaloaded bovine aortic endothelial cells grown on a glass slide were mounted on a parallel plate flow chamber, and fluorescence was monitored as described in MATERIALS AND METHODS. At the arrow indicating flow, superfusion of the cells with HEPES-buffered saline (HBS) at a rate of 7 ml/min (10 dyn/cm2) was initiated. The medium was changed at the time indicated to HBS containing 1 PM ATP with no discontinuity of flow. B: same conditions as in A with return of the medium to HBS without ATP at the arrow indicating washout. C: same conditions as in A, until flow was terminated in the presence of ATP at the arrow indicating zero flow.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

H1700

SHEAR

STRESS

AND

ATP

ON

of fura- fluorescence (33) was less than 2% of the total signal as determined by manganese quenching. For calibration, fura- loaded cells were harvested from glass slides and resuspended in 30 ~1 of HBS containing 0.1% bovine serum albumin. Triton X-100 (0.1% final concentration) was added to half of the cell suspension and Triton X-100 and EGTA (9 mM final concentration, pH = 8.5) were added to the remaining cells. A small aliquot

ENDOTHELIAL

CELL

[CA’+],

(2 ~1) of each cell suspension was mounted in an empty flow chamber for measurement of maximum and minimum fluorescence ratio, R,,, and R,,,, respectively. [Ca”‘], was calculated by the following equation

[Ca”‘l, = [(R - Rm,,)/(Rmax - R)~O’XJ~~SWI% where F386T and FZssEare the fluorescence values at 385 nm excitation in the presence of saturating and zero Cazf, respectively, and Kd equals 224 nM at 37°C (23). R,,, values ranged from 6.2 to 8.3, R,,, from 0.40 to 0.47, and F385E/F385Tfrom 6.4 to 10.6. These calibration values are similar to those obtained for authentic fura- pentapotassium salt using the same calibration procedure. Changes in the fluorescence ratio, R, were considered to reflect true changes in [Ca”‘], only if reciprocal changes in Fss5 and FZ8, were observed. In order to adequately control the temperature of the superfusion buffer at all flow rates, the entire flow system and microscope interface was surrounded by a plastic housing and the temperature was maintained at 37°C using 2 air curtain incubators (Nicholson Precision Instruments, Bethesda, MD). The traces shown are representative of experiments replicated at least three times. Where indicated, [Ca”‘], values are reported as means + SE and n equals the number of determinations. After each fura- experiment, the flow system (plastic tubing and flow chamber) was washed with 100% ethanol and rinsed extensively with deionized water to prevent contamination of subsequent experiments by the presence of agonist agents. This procedure was particularly important for the Ca2+ ionophore ionomycin. This compound was found to adhere to the plastic surfaces of the flow system. Residual ionomycin remained associated with the plastic for several weeks if washed only with water or saline solution, giving rise to what appeared to be flow-dependent changes in [Ca”‘], (data not shown). These responses were eliminated by extensive washing with ethanol or by complete replacement of the tubing after each ionomycin experiment. Although less pronounced, similar results were obtained after use of high concentrations of the agonists, bradykinin, histamine, and the adenine nucleotide analogues. RESULTS

Fura- loaded endothelial cells cultured on glass slides were employed for the measurement of [Ca”‘], in the absence and presence of extracellular ATP (Fig. 1). Increasing flow from zero to 7.0 ml/min (shear stress = 10 dyn/cm2) produced a small decrease in [Ca”‘], followed by a small increase to an apparent steady level. The decrease observed after the onset of flow was associated with a simultaneous decrease in both Fas5 and FZs6 and reflects washout of extracellular fura- that leaked from cells before the initiation of flow. However, the small subsequent increase in [Ca”‘], may represent a shear FIG. 2. Effect of shear stress on endothelial cell morphology. Cells were subjected to shear stress for 24 h at 37°C as previously described (17). Photomicrographs show a representative field of stationary control monolayers (A) and cells subjected to either 4.6 dyn/cm’ (B) or 11 dyn/cm’ shear stress (C). Direction of fluid flow was from top to bottom.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

SHEAR

STRESS

AND

ATP

ON

stress-dependent phenomenon. This small ( 300), primary HUVECs (n = 26), or subcultured CPAEs (n = 25). Additionally, we examined [ Ca2+]i in BAECs subjected to either 4.6 or 11 dyn/cm2 shear stress for 24 h and subsequently loaded with fura- under stationary conditions. The cells exposed to 11 dyn/cm2 shear stress were elongated and aligned in the direction of fluid flow, whereas the cells exposed to 4.6 dyn/cm2 shear stress for 24 h were not aligned (Fig. 2). These preconditioned cells also exhibited little change in [Ca2+]i upon exposure to different levels of shear stress but responded to application of ATP (Fig. 3). Addition of ATP (1 PM) under continuous flow conditions increased [Ca2+]i within 10 s from a basal value of 56.0 t 2.5 nM to a peak value of 319 t 54.4 nM (n = 12; Fig. 1). After the peak response, [ Ca2+]i declined slowly with time. As expected, the sustained component of the [Ca2+]i response was sensitive to removal of extracellular Ca2+ (data not shown), indicating that this phase reflects Ca2+ influx from the extracellular space. Removal of ATP from the superfusion buffer without change in the flow rate resulted in the immediate return of [Ca2+]i to basal levels (Fig. 1B) demonstrating that the effect of ATP is rapidly reversible and dependent on continuous receptor stimulation. A similar decrease in [Ca”‘]i was li

A

ll0pMAl.P

II

400

Flow Rote (ml/min.)

-

2.8 1

.

c\l

17.5

-

0

2.8

11

1

. 1 lOpMAT? ’

B

ENDOTHELIAL

CELL

H1701

[CA*+]i

observed when the flow rate was decreased to zero during the sustained phase of the [Ca”‘]i response to ATP (Fig. 1C). The decrease in [Ca2+]i observed in response to the reduction in flow rate to zero was examined as a function of extracellular ATP concentration (Fig. 4). The peak change in [Ca”‘]i increased in a dose-dependent fashion with an ED5* of -10 PM. After termination of flow during the sustained phase of the response, [Ca2+]; decreased at a rate that was inversely proportional to the ATP concentration. These results suggest that a potent degradative or removal mechanism for ATP exists at the endothelial cell membrane and that the response of the cells to ATP under static conditions may be limited by ATP diffusion. Previous studies using vascular endothelial cells have shown that ATP stimulates the release of Ca2+ from internal stores and the influx of Ca2+ from the extracellular space as a result of interaction with PsYpurinoceptors (4, 30). Likewise it has been demonstrated that the ectonucleotidase of vascular endothelial cells will hydrolyze ATP, ADP and AMP to adenosine (6, 20). To test the hypothesis that the termination of the sustained phase of the [Ca2+]i response to ATP after reduction of flow rate to zero reflects hydrolysis of ATP to less active metabolites, the relative potency for elevation of [Ca2+]i by ATP, ADP, AMP and adenosine was examined (Fig. 5). ADP (1 PM) and AMP (100 PM) both produced increases in [Ca2+]; under flow conditions. However, adenosine was without effect at the highest concentration examined (100 PM). The rank order of potency determined from the peak of the [Ca”‘]i response was ADP > ATP >>> AMP > adenosine consistent with P2,-receptor binding by these agonists (3, 4, 21). Upon cessation of flow, the sustained [Ca”‘]; response seen during application of both ADP and AMP returned rapidly to basal levels in a fashion similar to that obtained with ATP. This result is consistent with the ability of the ectonucleotidase to metabolize both of these nucleotides to adenosine. Adenosine (100 PM) had no effect on the [Ca2+]; response of the cells to ATP (1 PM). At the concentration of ATP used in the above exper-

g600

400

t

1000 t

200

Flow Rat. (ml/min.)

2.8

t

17.5

1

0

I 3

800

r

600

g

400

2.8

11

I

3

1

1

I

L

I

I

6

9

12

15

18

I

. Flow cl.0 mew

1

Time (minutes) FIG. 3. Effect of shear stress [Ca2+]i of preconditioned cells. Cell monolayers shown in Fig. 2, B and C, were loaded with fura- and subsequently used for measurement of [Ca2+]i. At the times indicated by the arrows, the cells were sequentially subjected to 2.8, 17.5, 0, and 2.8 ml/min superfusion with HBS corresponding to shear stress values of 4, 25, 0, and 4 dyn/cm2, respectively. The medium was changed to HBS containing 10 ,uM ATP at the time shown without changing the flow rate. Flow was subsequently terminated at -16 min. A shows [Ca2+]i values for low shear stress cells (4.6 dyne cmB2 24 h) and B gives values for the high shear stress cells (11 dyne crnm2.24 h). l

2

4

6

6

Time (minutes) FIG. 4. Dose-dependent effect of ATP under flow. Superfusion of the cells with HBS at 7 ml/min was initiated as in Fig. 1. The medium was changed to HBS containing either 0, 0.1, 1, 10, 100, or 1,000 PM ATP. Flow was terminated in the continued presence of each concentration of ATP at the arrow indicating zero flow. For comparison, the 6 traces are superimposed. The lowest trace (unlabeled) was obtained in the absence of ATP.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

H1702

SHEAR

2

4

STRESS

6

AND

ATP

ON

ENDOTHELIAL

CELL

[CA2+]i

8

TIME (minutes) FIG. 5. Effect of adenine nucleotides under flow. Superfusion of the cells with HBS at 7 ml/min (10 dyn/cm2) was initiated as in Fig. 1. The medium was changed to HBS containing either 1 PM ATP, 1 pM ADP, 100 PM AMP, or 100 PM adenosine. For comparison, the 4 traces are superimposed. Flow was terminated in the continued presence of each agent at the arrow indicating zero flow.

C

lc#~AMPcP 1

.

iments, hydrolysis of ATP to AMP would be sufficient to terminate the response since the potency of AMP at the P+eceptor is much less than that for ATP. This result suggests that it is the ectoADPase activity that may contribute most to the regulation of Ca2+ signaling via the hydrolysis of ADP to AMP. To test this hypothesis, experiments were performed using the ADP analogue AMP-CP, which is resistant to hydrolysis by both the ectonucleotidase and the ectopyrophosphatase (7, 21). AMP-CP produced dose-dependent changes in [Ca2+]i over the concentration range of 0.01-l mM (Fig. 6). The response was qualitatively identical to that obtained with ATP. [Ca2+]i peaked and subsequently returned slowly toward basal values in the continued presence of the agonist. Likewise, the effect of AMP-CP on [Ca”‘]i was rapidly reversed when the cells were superfused with HBS in the absence of the agonist (Fig. 6B). However, unlike ATP or ADP, the sustained component of the [Ca”‘]i response to AMP-CP was unaffected by reduction of the flow rate to zero (Fig. 6C). This result was obtained at each AMP-CP concentration examined and suggests that hydrolysis of ADP by the ectonucleotidase is responsible for the rapid decrease in the ATPinduced [ Ca2+]i response observed upon termination of flow. Identical results were obtained using maximal and submaximal concentrations of the nonhydrolyzable ATP analogue AMP-PCP (0.1-l mM; Fig. 7) or maximal concentrations of AMP-PNP (0.1-l mM; Fig. 8). However, experiments with lower concentrations of AMPPNP (10400 PM) revealed a slow decrease in [Ca2+]i upon cessation of flow (Fig. 9). Endothelial cells also express angiotensin converting enzyme, which is responsible for the inactivation of extracellular bradykinin. Thus a similar reduction in the sustained component of the bradykinin response may be expected upon cessation of flow. To test this hypothesis, the cells were stimulated with 0.5, 1.0, and 10 nM bradykinin under flow conditions. The profiles obtained upon addition and washout of bradykinin during the sustained phase of the [Ca”‘]i response were essentially identical to those produced by the adenine nucleotides

2

4

8

8

Time (minutes) FIG. 6. Effect of AMP-CP under flow. A: at the arrow indicating flow, superfusion of the cells with HBS was initiated at a rate of 7 ml/ min (10 dyn/cm2). The medium was changed to HBS containing 100 pM AMP-CP at the time indicated. B: same conditions as in A with return of the medium to HBS without AMP-CP at the arrow indicating washout. C: same conditions as in A, until flow was terminated in the presence of AMP-CP at the arrow indicating zero flow.

and their nonhydrolyzable analogues (Fig. 10). The sustained component of the [Ca2+]i response was unaffected by cessation of flow in the presence of bradykinin at concentrations (X0 nM) yielding maximum change in [ Ca2+]i. However, at low concentrations (0.5 and 1 nM), reduction of flow to zero resulted in a small but reproducible decline in [Ca2+]i that was never observed in the presence of captopril (100 nM), a specific inhibitor of angiotensin converting enzyme (Fig. 11). Captopril alone had no effect on [Ca2+]i. These results suggest that the response of the endothelial cells to bradykinin, like the adenine nucleotides, may be affected by the activity of degradative ectoenzymes. The results presented thus far suggest that the delivery of agonist agents, as indicated by the change in [Ca2+]i, may be regulated by flow. To further address this hypothesis, the effect of changing flow at different concentrations of ATP was examined (Fig. 12). When the cells were superfused with HBS or HBS containing 100 nM ATP, there was no detectable increase in [Ca2+];. The small decline in [Ca2+]i observed with increasing flow rate reflects washout of extracellular fura- that diffused from the cell monolayer before the elevation of the flow rate. In contrast, flow-dependent increases in [Ca2+]i were observed with concentrations of ATP ranging from

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

SHEAR

IA 200

STRESS

AND ATP ON ENDOTHELIAL

CELL

Hl703

[CA2+]i

lOOp&iAMPPNP 0.5 mM AMP-PCP

-

1

1

t 150 =

KM3 100 - Flow (7 ml/min,)

1OOpMAMPPNP

200

IB

0.5 mM AMP-PCP

IC

0.5 mM AMP-PCP

150 100

.

a

2

4

Time

50

2

4 Time (minutes)

6

FIG. 7. Effect of AMP-PCP under flow. A: at the arrow indicating flow, superfusion of the cells with HBS was initiated at a rate of 7 ml/ min (10 dyn/cm2). The medium was changed to HBS containing 0.5 mM AMP-PCP at the time indicated. B: same conditions as in A with return of the medium to HBS without AMP-PCP at the arrow indicating washout. C: same conditions as in A, until flow was terminated in the presence of AMP-PCP at the arrow indicating zero flow.

0.5 to 10 PM in the superfusion buffer. As ATP concentration increased, the threshold flow rate for the [Ca”‘]; response decreased and the magnitude of [Ca2+]i elevation at any one flow rate increased. These results demonstrate that the effect of ATP is dependent on flow rate and ATP concentration. DISCUSSION

Does shear stress increase [Ca2+]i of vascular endotheZial cells? The suggestion that shear stress may affect

[Ca2+]i derives from studies in which changes in vascular reactivity, normally ascribed to elevation of [Ca2+]; of the endothelial cell with concomitant release of endothelial-derived paracrine factors, are observed with fluid flow (24, 31, 32). Shear stress, the mechanical force tangential to the endothelium, is considered to be the most injurious factor associated with blood flow (18). When cultured endothelial cells are subjected to steady shear stress in vitro, the cells elongate and align in the direction of fluid flow (14, 26). Since morphological changes undoubtedly reflect a second messenger-mediated event, we examined the role of [Ca2+]i using fura2 loaded endothelial cells under shear stress. However,

.

.

0

(minutes)

FIG. 8. Effect of AMP-PNP under flow. A: at the arrow indicating flow, superfusion of the cells with HBS was initiated at a rate of 7 ml/ min (10 dyn/cm’). The medium was changed to HBS containing 100 PM AMP-PNP at the time indicated. B: same conditions as in A with return of the medium to HBS without AMP-PNP at the arrow indicating washout. C: same conditions as in A, until flow was terminated in the presence of AMP-PNP at the arrow indicating zero flow.

we found that shear stress produced little change in [Ca2+]i. In contrast, the cells responded with sustained changes in [Ca2+]i when challenged under flow conditions with either BK or ATP, two known endothelium-dependent vasodilators. Similar experiments performed using either primary or subcultured BAECs, primary cultures of HUVECs, or subcultured CPAEs revealed that the relative lack of shear stress affect on [Ca2+]i is a general phenomenon and not related to the cell type. We also evaluated the effect of shear stress on [Ca2+]i in cells that had been prealigned. These cells also exhibited little change in [Ca”‘]i upon subsequent challenge with shear stress. Thus, under different shear stresses and in a number of different endothelial cell types, shear stress had little effect on [Ca2+]i. The lack of shear stress effect on [Ca2+]; suggests that some signal messenger other than [ Ca2+]i may be involved in the morphological changes of vascular endothelial cells. Ando et al. (1) reported flow-induced changes in [ Ca2+]i of cultured vascular endothelial cells. [Ca2+]i increased upon initiation of fluid flow in a biphasic manner similar to the profile observed upon application of agonist agents; [Ca2+]i peaked within seconds and subsequently declined to a sustained level that was sensitive to removal of extracellular Ca2+. The reason for the discrepancy between the results of the present study and those of Ando et al. (1) is unknown. It appears however,

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

H1704

SHEAR

STRESS

AND ATP ON ENDOTHELIAL

CELL

[CA2+]i

A 10 nM Brdyidnin 120

1

10 pM AMP-PNP

t

1 B

120

10 pM AMP-PNP

C

10 nM Bmdykinin

C 10 /A AMP-PNP

120

2

4

Time 2

4 Time

6 (minutes)

FIG. 9. Effect of a submaximal concentration of AMP-PNP under flow. A: at the arrow indicating flow, superfusion of the cells with HBS was initiated at a rate of 7 ml/min (10 dyn/cm2). The medium was changed to HBS containing 10 PM AMP-PNP at the time indicated. B: same conditions as in A with return of the medium to HBS without AMP-PNP at the arrow indicating washout. C: same conditions as in A, until flow was terminated in the presence of AMP-PNP at the arrow indicating zero flow.

that the majority of their data was obtained using medium Ml99. As we have recently shown, flow-dependent changes in [Ca2+]i can be obtained using this medium since it contains 1.8 PM ATP (28). Likewise, the results of the present study clearly show that flow-dependent changes in [Ca2+]i can be obtained with HBS supplemented with low concentrations of ATP (Fig. 12). Although Ando et al. (1) stated that the changes in [Ca2+]i with shear stress were also obtained using minimum essential medium and phosphate-buffered saline, the responses were not shown to be graded with respect to shear stress, and the peak changes in [ Ca2+]i were stated to be variable from experiment to experiment. Flow-dependent effect of ATP on [Ca2fli of vascular endothelial cells. Although there was little effect of shear

stress per se on [Ca2+]i, the results of the present study demonstrate that the effect of ATP on Ca2+ signaling depends on flow rate. The initial response of the cells to each of the different adenine nucleotide analogues was essentially identical when examined under flow conditions. The agonists rapidly increased [ Ca2+]i, which peaked within lo-20 s and slowly returned toward basal

6

8

(minutes)

FIG. 10. Effect of bradykinin under flow. A: at the arrow indicating flow, superfusion of the cells with HBS was initiated at a rate of 7 ml/ min (10 dyn/cm2). The medium was changed to HBS containing 10 nM bradykinin at the time indicated. B: same conditions as in A, until flow was terminated in the presence of bradykinin at the arrow indicating zero flow. C: same conditions as in A with return of the medium to HBS without bradykinin at the arrow indicating washout.

values. The rank order of potency for the peak response was ADP = ATP > AMP-PNP > AMP-CP > AMPPCP > AMP. Upon washout of each agonist during the sustained phase of the response, [Ca2+]i rapidly returned to prestimulus levels suggesting that [Ca2+]i is dependent on continuous receptor stimulation and that the dissociation of the agonist from the receptor is rapid. The fate of the adenine nucleotides after dissociation from the receptor however, was revealed by reduction of the flow rate to zero during the sustained phase of the response. Those adenine nucleotides that are thought to be metabolized by the ectonucleotidase, specifically, ATP, ADP, and AMP all showed a rapid decrease in [Ca2+]i upon cessation of flow. In contrast, the [Ca2+]i response of those analogues thought to be resistant to hydrolysis by ectoenzymes, specifically, AMP-CP and AMP-PCP, was unchanged by reduction of flow to zero during the sustained phase. This result was obtained for both maximum and submaximal concentrations of these analogues and for maximum concentrations of AMP-PNP. However, cessation of flow during the sustained phase of the response to low concentrations of AMP-PNP produced a reduction in [Ca2+]i. Since a similar response was not obtained with low concentrations of AMP-CP, which is resistant to hydrolysis by both the ectonucleotidase and

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

SHEAR

STRESS

AND ATP ON ENDOTHELIAL

CELL

H1705

[CA2+]i

Flow Rate (ml/mm.)

0.35 400

100

0.70

7.0

2.8

17.5

1~

ATP 0 nM

5oJ

300 200

‘3

]

100 nM 6-l

I

n

100

B

52

1 nM BK

50I

500 nM

rkL n

n

n

1

300 200

n

- Flow (7.0 ml/min.)

100

l

1

1

I

I

I

2

4

6

8

Time

5oJ

n

n

n

(minutes)

FIG, 11.. Effect of a submaximal concentration of bradykinin under flow. A: at the arrow indicating flow, superfusion of the cells with HBS was initiated at a rate of 7 ml/min (10 dyn/cm2). The medium was changed to HBS containing 1 nM bradykinin at the time indicated. Flow was terminated in the presence of bradykinin at the arrow indicating zero flow. B: same conditions as in A with captopril (100 nM) added along with the bradykinin.

the ectopyrophosphatase (7, 21), the response observed with AMP-PNP probably reflects hydrolysis by the ectopyrophosphatase. The metabolite of this reaction, AMP, would then be degraded by 5’-nucleotidase. There are two additional factors that could contribute to the decrease seen upon cessation of flow during the sustained phase of the [Ca2+]i response. First, the sensitivity of the purinoceptor may be affected by shear stress. This change in sensitivity may reflect a shear stressdependent change in either 1) the affinity of the receptor for ATP or 2) the coupling of receptor occupation to subsequent steps in the signal transduction pathways leading to an apparent increase in agonist efficacy. This possibility seems unlikely for the following reasons. Reduction of flow to zero during maximal stimulation with all of the adenine analogues (including ATP) had no effect on the [Ca2+]i response, suggesting that shear stress does not affect agonist efficacy by alteration of one of the postreceptor binding steps in the signal transduction pathway. Likewise, reduction of flow rate during stimulation by submaximal concentrations of the nonhydrolyzable adenine analogues AMP-CP and AMPPCP did not change [Ca”‘]i, suggesting that a shear stress-dependent change in receptor affinity does not occur. Thus shear stress appears to have little direct effect on the responsiveness of the P2,-receptor. Similar results were obtained with maximal concentrations of bradykinin or submaximal concentrations in the presence of the angiotensin converting enzyme inhibitor, captopril, suggesting that shear stress also has little effect on bradykinin receptor sensitivity. Second, shear stress may have a direct effect on the metabolism of the

5OJ

0

I

I

I

5

10

15

Time (minutes)

FIG. 12. Effect of flow rate and ATP concentration on [Ca”+]i. Cells were superfused with HBS containing the indicated concentrations of ATP. Cells were equilibrated at a flow rate of 0.07 ml/min for 20 min before time 0. At the times shown in each trace, flow rate was increased for 1 min to the value indicated and then returned to 0.07 ml/min for 2 min before each subsequent elevation of flow rate. Flow rates of 0.07, 0.35,0.7, 2.8, 7.0, and 17.5 ml/min correspond to shear stress values of 0.1, 0.5, 1, 4, 10, and 25 dyn/cm2.

adenine nucleotides i.e., shear stress may induce inactivation of the ectonucleotidase. However, experiments using dextran to increase viscosity and hence, shear stress, more than sevenfold (see shear stress equation in MATERIALS AND METHODS) at a constant flow rate and constant ATP concentration showed no increase in [Ca2’]i, suggesting that shear stress has little or no effect on nucleotidase activity (data not shown). Although the results of the present study do not eliminate the possibility that shear stress may affect some aspect of signal transduction, it is clear that basal [Ca2+]i and the response of the endothelial cells to agonist agents is remarkably insensitive to shear stress. Role of convective-diffusive transport in the agonist response. Besides the mechanical forces accompanying blood flow, there are two transport phenomena that contribute to the effective concentration of ATP at the cell surface, convection and diffusion (Fig. 13). The concentration profile, and thus the delivery of ATP to the cell surface, depends on convection as a result of flow rate, and diffusion of ATP in the direction normal to the cell surface. Additionally, however, the effect of ATP on [Ca2+]i of vascular endothelial cells appears to reflect a

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

HI706

SHEAR

STRESS

AND ATP ON ENDOTHELIAL

CELL

[CA2+]i

tive agents on vascular endothelial tion.

ine

tea2+i 1 1 FIG. 13. Model for flow-dependent effects of ATP on endothelial cells depicting steady laminar flow over the surface of the vascular endothelial cells that contain the ectonucleotidase enzymes. As shown by the fluid flow arrows, velocity of fluid flow is zero at the cell surface and increases parabolically in the y-direction. Since flow at the cell surface is negligible, the delivery of ATP to the purinoceptor is controlled primarily by diffusion. Degradation of ATP establishes a concentration gradient for ATP in the solution bathing the cells as indicated by the decreasing size of the ATP concentration [ATP] as the cell membrane is approached. However, the concentration gradient for ATP near the cell surface changes as a function of flow rate. At low flow rate, degradation of ATP by the ectonucleotidase exceeds the rate of diffusion and the steady-state concentration of ATP at the cell surface remains low. However, at high flow rate, convection enhances the delivery of ATP from upstream. Thus the diffusion of ATP will exceed the rate of degradation by the ectonucleotidase.

balance between delivery of ATP to the receptors at the cell surface and the rate of degradation to less active or inactive hydrolysis products (Fig 13). Thus the steadystate concentration of ATP at the cell surface will be a function of flow rate, diffusion, and ectonucleotidase activity. At high flow rates, sufficient ATP is delivered from upstream to elicit a maximum response at each concentration. However, at lower flow rates, the steadystate balance is shifted toward degradation and diminished response. While the concentration of ATP in the blood is still controversial, best estimates are in the submicromolar range under normal conditions with increases into the tens of micromolar range under conditions such as circulatory shock (6, 21) or vascular injury and platelet aggregation (2). Since the threshold concentration of ATP or ADP for stimulation of the Pa,-receptor is -0.1 PM, it is possible that the flow-dependent changes in ATP response observed in the present study could play an important role in the regulation of vascular tone under physiological and pathophysiological conditions. Likewise, changes in the ectonucleotidase activity could modify the flow-dependent profile observed. Although little is known concerning the long-term regulation of ectonuc leotidase activity by the cell, the intracellular molecular events associated with si.gnal transduction (e.g. elevated inositol-trisphosphate, diacylglycerol, cyclic nucleotides, and/or [ Ca2+]J may modulate ectonucleotidase activity and thus the response of the cells to ATP. Although the presence of specific receptors and degradative enzymes for the adenine nucleotides and bradykinin at the endothelial cell surface is well established, the results of the presen t study demonstrate for the first time the combined role of fluid flow and inactivation mechanisms in the steady-state effect of vasoac-

cell signal transduc-

The excellent technical assistance of Lydia T. Sturgis, Lekha Rajan, and J. Gary Meszaros is gratefully acknowledged. We thank Drs. Larry V. McIntire, Matthias U. Nollert, and Scott L. Diamond for helpful discussions. This study was supported by Grant 871317 from the American Heart Association, and by National Heart, Lung, and Blood Institute Grants HL-37044, HL-44119, and HL-23016. This work was done during the tenure of an Established Investigatorship of the American Heart Association awarded to W. P. Schilling. Address for reprint requests: W. P. Schilling, Dept. of Molecular Physiology and Biophysics, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. Received 5 July 1990; accepted in final form 10 January 1991. REFERENCES 1. ANDO, J., T. KOMATSUDA, AND A, KAMIYA. Cytoplasmic calcium response to fluid shear stress in cultured vascular endothelial cells. In Vitro Cell. Dev. Biol. 24: 871-877, 1988. 2. BORN, G. V. R., AND M. A. A. KRATZER. Source and concentration of extracellular adenosine triphosphate during haemostasis in rats, rabbits and man. J. Physiol. Land. 354: 419-429,1984. 3. BURNSTOCK, G., AND C. KENNEDY. Is there a basis for distinguishing two types of P2-purinoceptor? Gen. Phurmacol. 16: 433-440, 1985. 4. CARTER, T. D., T. J. HALLAM, N. J. CUSACK, AND J. D. PEARSON. Regulation of P2,-purinoceptor-mediated prostacyclin release from human endothelial cells by cytoplasmic calcium concentration. Br. J. Pharmucol. 95: 1181-1190,1988. 5. CHEN, G., AND H. SUZUKI. Calcium dependency of the endothelium-dependent hyperpolarization in smooth muscle cells of the rabbit carotid artery. J. Physiol. Lond. 421: 521-534, 1990. 6. COADE, S. B., AND J. D. PEARSON. Metabolism of adenine nucleotides in human blood. Circ. Res. 65: 531-537, 1989. 7. CUSACK, N. J., J. D. PEARSON, AND J. L. GORDON. Stereoselectivity of ectonucleotidases on vascular endothelial cells. Biochem. J. 214: 975-981,1983. 8. DAVIES, P. F. How do vascular endothelial cells respond to flow? News Physiol. Sci. 4: 22-25, 1989. 9. DAVIES, P. F., C. F. DEWEY, S. R. BUSSOLARI, E. J. GORDON, AND M. A. GIMBRONE. Influence of henodynamic forces on vascular endothelial function: in vitro study of shear stress and pinocytosis in bovine aortic cells. J. Clin. Invest. 73: 1121-1129, 1984. Role of the intima in 10. DE MEY, J. G., AND P. M. VANHOUTTE. cholinergic and purinergic relaxation of isolated canine femoral arteries. J. Physiol. Land. 316: 347-355, 1981. 11. DEWEY, C. F., JR., S. R. BUSSOLARI, M. A. GIMBRONE, JR., AND P. F. DAVIES. The dynamic responses of vascular endothelial cells to fluid shear stress. J. Biomed. Eng. 103: 177-185,198l. 12. DIAMOND, S. L., S. G. ESKIN, AND L. V. MCINTIRE. Fluid flow stimulates tissue plasminogen activator secretion by cultured human endothelial cells. Science Wash. DC 243: 1483-1485, 1989. 13. ELLIOTT, S. J., AND W. P. SCHILLING. Carmustine augments the effects of tert-butyl-hydroperoxide on calcium signaling in cultured pulmonary artery endothelial cells. J. Biol. Chem. 265: 103-107, 1990. 14. ESKIN, S. G., AND L. V. MCINTIRE. Hemodynamic effects on atherosclerosis and thrombosis. Sem. Thromb. Hemostusis 14: l70174,1988. 15. ESKIN, S. G., H. D. SYBERS, L. TREVINO, J. T. LIE, AND J. E. CHIMOSKEY. Comparison of tissue-cultured bovine endothelial cells from aorta and saphenous vein. In Vitro 14: 903-910,1978. 16. FRANGOS, J. A., S. G. ESKIN, L. V. MCINTIRE, AND C. L. IVES. Flow effects on prostacyclin production by cultured human endothelial cells. Science Wash. DC 227: 1477-1479, 1985. 17. FRANGOS, J. A., L. V. MCINTIRE, AND S. G. ESKIN. Shear stress induced stimulation of mammalian cell metabolism. Biotech. Bioeng. 32: 1053-1060,1988. 18. FRY, D. L. Acute vascular endothelial changes associated with increased blood velocity gradients. Circ. Res. 22: 165-167, 1968. 19. FURCHGOTT, R. Role of endothelium in response of vascular smooth muscle. Circ. Res. 53: 557-573,1983.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

SHEAR

STRESS

AND ATP ON ENDOTHELIAL

20. GORDON, E. L., J. D. PEARSON, AND L. L. SLAKEY. The hydrolysis of extracellular adenine nucleotides by cultured endothelial cells from pig aorta: feed-forward inhibition of adenosine production at the cell surface. J. Biol. Chem. 261: 15496-15504, 1986. 21. GORDON, J. L. Extracellular ATP: effects, sources and fate. Biochem. 22. GORDON,

J. 223: 309-319,1986. J. L., AND W. MARTIN.

Endothelium-dependent relaxation of the pig aorta: relationship to stimulation of “Rb+ efflux from isolated endothelial cells. Br. J. Pharmacol. 79: 531-541,1983. 23. GRYNKIEWICZ, G., M. POENIE, AND R. Y. TSIEN. A new generation of Ca2’ indicators with improved fluorescence properties. J. BioZ. Chem. 260: 3440-3450,1985. HOLTZ, J., U. F~RSTERMANN, BASSENGE. Flow-dependent,

U. POHL, M. GEISLER, AND E. endothelium-mediated dilatation of epicardial coronary arteries in conscious dogs: effects of cyclooxygenase inhibition. J. Cardiouasc. Pharmacol. 6: 1161-1169, 1984. 25. IVES, C. L., S. G. ESKIN, AND L. V. MCINTIRE. Mechanical effects on endothelial cell morphology: in vitro assessment. In Vitro Cell. 24.

Deu. Biol. 26. IVES, C.

22: 500-507,

1986.

L., S. G. ESKIN, L. V. MCINTIRE, AND M. E. DEBAKEY. The importance of cell origin and substratum in the kinetics of endothelial cell alignment in response to steady flow. Trans. Am. Sot. Artif. Intern. Organs 29: 269-274, 1983. 27. MARTIN, W., N. J. CUSACK, J. S. CARLETON, AND J. L. GORDON. Specificity of Pz,-purinoceptor that mediates endothelium-dependent relaxation of the pig aorta. Eur. J. Pharmacol. 108: 295-299, 1985.

CELL

H1707

[CA*+]i

28. MO, M., S. G. ESKIN, M. U. NOLLERT, L. V. MCINTIRE, AND W. P. SCHILLING. Flow-induced changes in cytosolic free calcium of cultured vascular endothelial cells (Abstract). FASEB J. 4: A835, 1990.

2g OLESEN, S. P., D. E. CLAPHAM, AND P. F. DAVIES. Haemodynamic shear stress activates a K+ current in vascular endothelial cells. l

Nature Land. 331: 168-170,1988. S., E. RASPE, D. DEMOLLE, C. ERNEUX, 3o* PIROTTON, NAEMS. Involvement of inositol 1,4,5-trisphosphate

AND J. BOEYand calcium in the action of adenine nucleotides on aortic endothelial cells. J. BioZ. Chem. 262: 17461-17466,1987. 31. POHL, U., J. HOLTZ, R. BUSSE, AND E. BASSENGE. Crucial role of endothelium in the vasodilator response to increased flow in vivo.

Hypertension 32. RUBANYI,

Dallas

8: 37-44,

1986.

G. M., J. C. ROMERO, AND P. M. VANHOUTTE. Flowinduced release of endothelium-derived relaxing factor. Am. J. 19): H1145-H1149, 1986. PhysioZ. 250 (Heart Circ. Physiol. 33. SCANLON, M., D. A. WILLAMS, AND F. S. FAY. A Ca*+-insensitive form of fura- associated with polymorphonuclear leukocytes. J. BioZ. Chem. 262: 6308-6312,1987. E. A., B. L. STEINBACH, R. M. NEREM, AND C. J. 34. SPRAGUE, SCHWARTZ. Influence of a laminar steady-state fluid imposed wall shear stress on binding, internalization, and degradation of lowdensity lipoproteins by cultured arterial endothelium. Circulation 76: 648-656,

1987.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on October 4, 2018. Copyright © 1991 American Physiological Society. All rights reserved.