Mechanosensitive Ca2+ transients in endothelial cells from human umbilical vein

Proc. Natd. Acad. Sci. USA

Vol. 91, pp. 2940-2944, April 1994 Cell Biology

Mechanosensitive Ca 2+ transients in endothelial cells from human umbilical vein MASAHIRO OIKE, GuY DROOGMANS, AND BERND NILIUS* KU Leuven, Laboratorium voor Fysiologie, Campus Gasthuisberg, B-3000 Leuven, Belgium

Communicated by Erwin Neher, December 27, 1993 (received for review October 14, 1993)

Probes) dissolved in normal Krebs' solution for 20 min at room temperature and thereafter for another 20 min at 37TC. The extracellular solution was a modified Krebs' solution, containing 132.2 mM NaCl, 5.9 mM KCl, 1.2 mM MgC2, 1.5 mM CaCl2, 11.5 mM glucose, 11.5 mM Hepes-NaOH (pH 7.3). The pipette solutions in all experiments contained 100 mM potassium aspartate, 40 mM KCl, 5 mM NaCl, 1 mM MgCl2, 0.5 mM Na2ATP, 10 mM Hepes, 0.1 mM EGTA (pH 7.2). Cell swelling was induced by a hypotonic solution (HTS), which was obtained by diluting the normal Krebs' solution with an osmolarity of 290 mosM to an osmolarity of 185 mosM. The ionic concentrations in this solution were 94.6 mM NaCl, 4.2 mM KC1, 0.9 mM MgCl2, 1.1 mM CaCl2, 8.2 mM glucose, 8.2 mM Hepes. Ca2+-free HTS contained 0.7 mM EGTA. To exclude an effect of changes in ionic concentrations, a series of experiments were performed in which the HTS described above was compared to an isotonic solution with the same ionic composition to which 81.7 mM D-mannitol was added. The system for Ca2+ measurements and its calibration are based on a previously described method (6). Single cells were excited alternately with light of 360 and 390 nm via a rotating filter wheel (speed between 2 and 3 per s) and fluorescence was measured at 510 nm. Apparent concentration of free calcium was calculated from the fluorescence ratio of the background corrected fluorescence signals. The patch-clamp technique was applied in the standard whole-cell mode or in the perforated patch configuration using a List EPC-9 patchclamp amplifier. Membrane currents were filtered at 1 kHz with an eight-pole Bessel filter and digitized on line at sample intervals of 250 As. In the experiments with heparin (100 MM; 5 kDa; Sigma H-5640), neomycin (1 mM; Sigma), protein kinase A (PKA) inhibitory peptide [3 p;M; amino acid residues 5-24 of the heat stable PKA inhibitor (7); kindly provided by M. Bollen, Department of Biochemistry, KU Leuven, Belgium], and for some patches with arachidonic acid (AA) (10 p;M; Sigma), which were all added to the pipette solution, we have used ruptured patches. In all other experiments we used nystatinperforated patches (8) (a stock solution of nystatin at 50 mg/ml dissolved in dimethyl sulfoxide was diluted 1000 times in the pipette solution immediately before use). The following compounds were added to the bath: AA (Sigma), nordihydroguaiaretic acid {4,4'-(2,3-dimethyl-1,4-butanediyl)bis[1,2benzenediol]; NDGA} (4 ,uM; Sigma), 4-bromophenacyl bromide (pBPB; Sigma), cyclosporin A (10 ,uM; Sandoz), indomethacin (4 ,uM; Sigma). All experiments were performed at room temperature (20°C-22°C).

We have investigated the changes in intracelABSTRACT lular calcium concentration ([Ca2+]) in human endothelal cells induced by mechanial stretch due to osmotic cell swelling. Hypotonic solutions also activate a Cl- conductance that has been described elsewhere and mainly serves to clamp the membrane potential at negative values to provide a driving force for Ca2+ influx. The Increase in [Ca2W+ caused by hypotonic solutive t tions is due to release from inouitol-1,4,5Ca2+ pools and a subsequent Ca2+ influx, apparently activated by store depletion. These changes in [Ca2e+ are completely abolished if the phospholipase A2 (PLA2) activity is inhibited by either 4-bromophenacyl bromide or cyClosporin A. Arachidonic acid, applied either extracellularly or intracellularly via the patch pipette, mimics the mechanosensitive response even in cells with blocked PLA2. Metabolites of the ipo- and cyclooxygenase pathways can be excluded. Phospholipase C activation and the protein kinase A pathway are not involved in this cal mechanical response. Although no specific ph tools for probing the role of PLA2 are available, our evidence suggests that mecansensitivity in endothelial cells may be modulated by ar onic acid.

The biological responses of vascular endothelial cells to mechanical forces, such as shear stress and mechanical stretch, are very diverse. Some of them develop fast, while others occur within the range of several hours (1, 2). One of the earliest responses to mechanical activation is an increase in intracellular [Ca2W] ([Ca2W+, which, for instance, activates the production of nitric oxide (3). Until now, there is no clear indication by which mechanism endothelial cells act as mechanosensors. We have shown recently (4) that mechanical stimulation of endothelial cells activates a Cl- conductance that is not related to the changes in [Ca2~+i. Activation of this conductance may clamp the resting potential of the endothelial cells at negative values to provide a sufficiently high driving force for Ca2+ influx (4). In the present report, we have studied the changes in [Ca2eli that occur upon mechanical stimulation and the various signal transduction pathways that might be involved in the mechanosensitivity of endothelial cells.

MATERIALS AND METHODS Endothelial cells were isolated, as described in more detail previously (5), from human umbilical cord veins by a collagenase digestion method and grown in medium 199 containing 10%6 human serum, 2 mM L-glutamine, 100 units of penicillin per ml, and 100 pg of streptomycin per ml. We have used only nonconfluent voltage-clamped cells in our experiments that were responsive to stimulation with histamine (100 ,pM). For intracellular Ca2+ measurements, cells were incubated with 2 ,LM fura-2/AM (acetoxymethylester) (Molecular

Abbreviations: AA, arachidonic acid; [Ca2+]i, intracellular free calcium concentration; PLA2, phospholipase A2; PLC, phospholipase C; PKA, protein kinase A; Ins(1,4,5)P3, inositol 1,4,5trisphosphate; NDGA, nordihydroguaiaretic acid; pBPB, 4-bromophenacyl bromide. *To whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

2940

Cell Biology: Oike et al.

Proc. Nadl. Acad. Sci. USA 91 (1994)

RESULTS Superfusion of voltage-clamped endothelial cells with Ca2+free HTS induces cell swelling and membrane stretch, which are accompanied by a transient increase in [Ca2+]i (Fig. 1A). Internal application of hypertonic solution via the patch pipette induces a similar Ca2+ transient in cells superfused with normotonic Ca2+-free solutions, suggesting that these Ca2+ signals are due to changes in osmolarity rather than to changes in extracellular ionic strength. These Ca2+ transients developed within 20-100 s and reached their maximum within 4-6 min, after which [Ca2+]i decayed to its basal value. To quantify our results, we have determined the amplitude ofthe maximum increase of [Ca2+]i above its resting level. These values showed a large variability, especially between various batches ofcells. Values ranged between 0.1 and 0.4 juM, with a mean value of 0.23 ± 0.03 juM (n = 50 cells). At the same time, an inward current was activated at a holding potential of -40 mV. This current has been described in detail elsewhere (4) and has been identified as a current through a small conductance, Ca2+-independent Cl- channel. A

HTS

2941

This current was completely blocked by 1,9-dideoxyforskolin, but this compound did not affect the increase in [Ca2+ ] (n = 9 cells). Fig. 1B shows a similar Ca2+ transient obtained by omitting mannitol from a normotonic Ca2+-free extracellular solution, keeping all other ionic concentrations constant. A subsequent stimulation with a supramaximal concentration of histamine (100 pAM), which has been shown previously (9, 10) to induce a Ca2+ transient by depleting intracellular inositol 1,4,5-trisphosphate [Ins(1,4,5)P3d-sensitive Ca2+ pools, did not induce any significant increase in [Ca2e]i (0.05 + 0.04 pM in five cells). It can therefore be concluded that cell swelling induces Ca2+ release and depletion of intracellular agonistsensitive Ca2+ stores. Depletion of these Ins(1,4,5)P3sensitive stores by pretreatment with thapsigargin (Fig. 1C) in Ca2+-free solution completely abolished the Ca2+ response induced by HTS. A similar result was obtained if the stores were depleted by exposure to histamine in Ca2+-free solution before application of HTS. Other mechanical stimuli, such as shear stress induced by directing a stream ofnormotonic solution along the cell surface

15aHTS

A

1 0Ca

L.

a

C14

=L

.

04

C5

B

HTS

C

4 min

Thapsigargin 2pM

B 2 | | \ Nc|

I

6

(8)

HTS

1.5 -

=

6

(3) 4 min ...............

D

Shear

E

Stretch

1.0 2

0.5 8) (4)

6

(4)

(4)

(4)

10

5

4 min

4 min

FIG. 1. Mechanical activation of a membrane current and Ca2+ release in endothelial cells. (A) Exposure of endothelial cells to hypotonic Ca2+-free bath solutions activates an inward current (upper trace) and induces a transient increase in [Ca2+]i (lower trace). The holding potential was -40 mV. (B) Activation of a Ca2+ signal by hypotonic stress at constant extracellular ion concentrations by omitting mannitol from the superfusion solution (HTS) in Ca2+-free solutions. A subsequent application of a supramaximal concentration of histamine did not cause any substantial increase in [Ca2e]j, indicating that the Ins(1,4,5)P3-sensitive Ca2+ stores are depleted. (C) HTS, applied after depletion of intracellular Ca2+ stores with 2 ,uM thapsigargin, does not evoke a Ca2+ transient. (D and E) Shear stress and direct mechanical stretch also induce a transient Ca2+ increase in Ca2+-free solutions. All these mechanical stimuli were applied to cells voltage clamped at -40 mV.

[Ca21e

[mM]

FIG. 2. Changes in [Ca2+]i after a short period of cell swelling are modulated by extracellular [Ca2W] ([Ca2W+. (A) Swelling of endothelial cells by HTS evoked an increase in [Ca2+]i together with a C1current. The recovery of [Ca2+]1 in normotonic solution is accelerated in Ca2+-free -solution, but [Ca2+h] increases again after resubmission of extracellular Ca2+. Apparently a Ca2+-entry pathway is activated by the exposure to HTS, but it is not accompanied by a significant change in transmembrane current. (B) Dependence of [Ca2+]i on the extracellular [Ca2+] in cells that had been exposed previously to HTS (solid symbols) or not (open symbols). It is obvious that [Ca2+]j strongly depends on extraceliular [Ca2+J after exposing them to HTS but not under control conditions. This dependence on extracellular Ca2+ may also explain the increases in [Ca2+]i after switching back to the normotonic solution (A) that has a higher extracellular [Ca2+1 than the hypotonic solution (1.1 M in HTS compared to 1.5 M in the normal physiological solution). Number of cells is indicated in parentheses.

2942

Cell Biology: Oike et al.

Proc. Natl. Acad. Sci. USA 91 (1994)

or direct mechanical stretch induced by moving a patch pipette in close contact with the cell surface, also induced transient changes in intracellular Ca2+ (Fig. 1 D and E). In the presence of extracellular Ca2+, the HTS-induced increase in [Ca2+]i was sustained, suggesting a contribution of Ca2+ influx to the mechanically evoked Ca2+ signal. After reperfusion with normotonic solution, [Ca2e]i slowly decayed to its resting level with t I2 = 328 ± 122 s (n = 5 cells). It is well established that store depletion by agonists activates a Ca2+-entry pathway in many nonexcitable cells (11, 12). Fig. 2A provides some evidence for activation of a similar mechanism by store depletion induced by cell swelling. In the presence of extracellular Ca2+, cell swelling induced a sustained increase in [Ca2e]i and activated a Cl- current. A subsequent superfusion with normotonic solution was accompanied by a slow decline of [Ca2+]i, which was accelerated after removal of extracellular Ca2+. A subsequent readmission of extracellular Ca2+ caused a pronounced increase in [Ca2+]i. In contrast, similar changes in extracellular Ca2t, if applied to the same cell but before induction of cell swelling, did not affect the level of intracellular calcium. The dependence of the level of intracellular calcium on the extracellular Ca2+ concentration is shown in Fig. 2B and strongly suggests that store depletion induced by cell swelling may activate a Ca2+-entry mechanism. We were unable, however, to detect any changes in transmembrane current during these changes in [Ca2+]j. This current is either too small and beyond the resolution of our technique, or these changes might occur through some electroneutral mechanism. Activation of phospholipase C (PLC) was not involved in the mechanically induced release of intracellular Ca2+. The elevation of [Ca2+]i due to HTS was not affected if the endothelial cells were loaded via the patch pipette with either 1 mM neomycin, a blocker of PLC, or with the low molecular weight heparin that blocks Ins(1,4,5)P3 receptors. In contrast, both substances completely abolished the histamine (100 uM)-evoked Ca2+ signals in the same cell. A modulatory role of PKA is also unlikely, because the [Ca2+1j responses to mechanical stimulation were not affected if endothelial cells were loaded via the patch pipette with a peptide consisting of

amino acid residues 5-24 of the heat stable PKA inhibitor (7) that specifically blocks the catalytic subunit of PKA. The quantified effects of these various putative mediators are summarized in Fig. 3, which shows the HTS-induced increase in [Ca2eJi for different experimental conditions; neither of these conditions significantly altered the response to HTS. The variability of the HTS-induced responses between different batches ofcells can be judged from the scatter of the control values in this figure. It has been reported that AA, but not its metabolites, is able to release Ca2+ from internal stores (13-15). Fig. 4A shows that AA (10 ,uM) induces a transient increase in intracellular Ca2+ in endothelial cells exposed to Ca2+-free solutions. The dependence of these changes in [Ca2 I on the concentration of AA is shown in Fig. 4B. The increase in [Ca2+]i by extracellularly applied AA still occurred in cells loaded with heparin (Fig. 4C). Histamine, however, was unable to evoke a release of intracellular Ca2+ in heparin-loaded cells (data not shown), indicating that costimulation of Ins(1,4,5)P3 receptors by AA is unlikely. This similarity between the AA and the mechanically induced changes in [Ca2J+] may point to a stimulation of phospholipase A2 (PLA2) by HTS. Inhibition of PLA2 by pretreating the cells with pBPB (10 liM) or cyclosporin A (16) (10 ,uM) completely abolished the Ca2+ response induced by HTS (Fig. 5 A and B). In contrast, the release of intracellular Ca2+ induced by AA was not significantly affected in pBPBpretreated cells (Fig. 4C), indicating that the inhibitory effects on the HTS responses cannot be entirely explained by nonspecific effects of these compounds. The concentration dependence of the inhibitory effect of pBPB is shown in Fig. 5C. Also, the Ca2+ transients induced by the other mechanical stimuli-i.e., shear stress and mechanical stretch of the cell-were completely abolished in the presence of 10 uM pBPB. The mechanically induced increase of [Ca2+]i still occurred in the presence of blockers of cyclooxygenase (indomethacin) or lipoxygenase (NDGA). It is therefore likely that the increase in [Ca2i]i is caused by AA rather than by some of its metabolites. The quantified effects of these substances are included in Fig. 3. HTS

0.4 .4

ly

1 C-

0.2-

.x;

/I

r_---.

:

:: ::.::...:.

. 1i

... . ....

1)

1 (4

(7):

.

+

1-T7?

.

v

:. L..

....:...:...

0.0-

c _

>

*0

E Co a o

o

Z

0

0 C 0

FiG. 3. Evaluation of various possible pathways of signal transduction in mechanically stimulated endothelial cells. Effects of inhibitors of the PLC and PKA pathways on changes in [Ca2+]i induced by cell swelling. Pooled data from endothelial cells clamped at a potential of -40 mV. Neither of these procedures significantly affected the swelling-induced increase in [Ca2+hi. On the right are shown the effects of lipoxygenase and cyclooxygenase inhibitors. These substances also did not significantly alter the Ca2+ responses, suggesting that metabolites of AA are not involved in the sigial transduction. Neomycin (1 mM), heparin (500 NM), and PKA inhibitor (PKI; 3 M) were loaded into the cell via the patch pipette, and measurements were done after 15-20 min. Indomethacin (Indo; cyclooxygenase blocker; 4 ,uM) and NDGA (lipoxygenase blocker; 4 PM) were applied to the bath throughout the experiment. (Inset) How A[Ca2]-ii.e., the increase in [Ca2e]i above the resting level-has been determined is shown. Large differences between amplitudes of the control HTS responses reflect the large variability between various batches of cells. Control and test values were always obtained from the same batch of cells. Number of cells is indicated in parentheses.

Cell Biology: Oike et A

Proc. Natl. Acad. Sci. USA 91 (1994) A

A

B

Arachidonic Acid

*=i

2943

HTS

I

4m

L

L;

*hk

4 min

4 mn control +Heparin +pBPB pBPB cyclosporin A

control

(10)

B

0.3-

C

0.4 -

2 0.2

i (6)

_;q cli

I

C 0.2-

CD