Endothelium-derived relaxing factor release associated with increased endothelial cell inositol trisphosphate and intracellular calcium

Endothelium-Derived Relaxing Factor Release Associated with Increased Endothelial Cell Inositol Trisphosphate and Intracellular Calcium Alex L. Loeb, PhD, Nicholas J. Izzo, Jr., PhD, Randolf M. Johnson, PhD, James C. Garrison, PhD, and Michael J. Peach, PhD

The release of eicosanoids and endothelium-derived relaxing factor (EDRF) from endothelial cells is thought to involve a calcium-dependent step. Using cultured bovine aortic endothelial cells as a model system, we have examined the relation between agonist-induced changes in inositol polyphosphates and calcium levels within the endothelial cells and extracellular calcium on EDRF release. In a superfusion-cascade system, EDRF was detected by the relaxation of a rabbit aortic ring without endotheliurn suspended beneath a column of cultured endothelial cells. Endothelial cell stimulation by bradykinin or melittin induced dose-dependent relaxation of the bioassay ring. In addition, bradykinin and melittin stimulated an increase in intracellular calcium concentration in fura- loaded endothelial cells and an increase in inositol 1,4,Btrisphosphate (lns[l,4,5]Ps) in cells prelabeled with 3H-myoinositol. Bradykinin stimulation produced transient increases in lns(l,4,5)Ps, fura- fluorescence and transient EDRF release. Melittin stimulation induced more prolonged release of EDRF from the endothelial cell column, which was correlated with sustained increases in the fura- signal and the level of lns(l,4,5)Ps. Omission of calcium from the cell superfusate attenuated, but did not eliminate, bradykinin-induced EDRF release and the calcium transient, whereas the melittin-induced responses were only slightly attenuated. Endothelial cells clearly demonstrate receptor-activation of phospholipase C and release of sequestered calcium from subcellular sites in response to Ins( 1,4,5)Ps. These results imply that EDRF release is correlated with increased intracellular calcium levels seen in the absence of extracellular calcium. However, sustained release of EDRF does require influx of extracellular calcium via an undefined mechanism. (Am J Cardiol 1988;62:366-406)

From the Department of Pharmacology, University of Virginia School of Medicine, Charlottesville, Virginia. Address for reprints: Michael J. Peach, PhD, Department of Pharmacology, University of Virginia School of Medicine, Charlottesville, Virginia 22908.

366

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62

any different drugs and hormoneswith vasodilatory activity inducethe releaseof an endotheliurn-derived relaxing factor (EDRF) from endothelial cells. Severalearly studiesdesignedto elucidate how EDRF releasewas regulated indicated that EDRF releasewas dependent on the presenceof extracellular calcium.‘-4 Theseinvestigatorsshowedthat the ability of many agonists to release EDRF was inhibited by pretreatment of intact vesselsegmentswith calcium antagonists’ or by incubation of vascularsegmentsin buffer with calcium omitted.1,2T5 However, it is known that several endothelium-dependentvasodilators do not require the presenceof extracellular calcium to induce EDRF release617 and, in fact, severalagonistswill increaseintracellular free calcium levels in cultured endothelial cellss~g in the absenceof extracellular calcium. Using cultured bovine endothelial cells known to releaseEDRFiO as a model system,we have examined the relation between agonist-induced changesin phospholipase C-mediated formation of inositol polyphosphates and calcium levelswithin the endothelial cells, and extracellular calcium on EDRF release. Inositol 1,4,5-trisphosphate(Ins[ 1,4,5]Ps) acts as an intracellular second messengerto releasecalcium from intracellular stores,” and therefore could be associatedwith EDRF release.

M

METHODS Endothelial cell culture: Endothelial cells from bovine

aorta were obtained as previously described.lOCells were grown to confluence in Waymouth’s medium (Gibco) supplementedwith 10% fetal bovine serum (Hyclone) plus penicillin (100 U/ml) and streptomycin (100 pug/ ml). At confluence, the cells were passagedor used for calcium flux experiments. Microcarrier culture: As describedpreviously,1°dispersed endothelial cells were seededonto Cytodex@3 microcarriers (Pharmacia) and placed into siliconized roller bottles and rolled at 1 rpm. When confluent on the microcarriers (about 1 week), aliquots were removedand washedwith physiologic salt solution in preparation for study. Column superfusion: An endothelial cell column was preparedby placing confluent endothelial cells grown on microcarriers (5 X lo7 cells) into a water-jacketed columnt” maintained at 37°C. The cell column was superfused at 3 ml/min with warmed, oxygenatedphysiologic salt solution of the following composition (in mM): 111 sodium chloride, 5 potassium chloride, 1 sodium phos-

phinate, 0.5 magnesiumchloride, 25 sodium bicarbonate, 2.5 calcium chloride (CaClz), 11.1dextrose.“0” calcium buffer was identical in composition except that CaC12 was omitted and dextrose increased to 18.6 mM. EDRF releasefrom the cell column was bioassayed using endothelium-denudedrabbit thoracic aortic rings. Each ring was mounted directly below the cell column and was superfusedas just described. The ring was allowed to equilibrate at 2.0 g resting tension for 90 minutes. Phenylephrine (0.4 PM) was added to the buffer reservoir to constrict the ring. Drugs used to stimulate EDRF releasewere infused for 1 minute into the superfusion systemjust upstreamof the cell column to the desired final concentration in the superfusate. To determine the effect of extracellular calcium on EDRF releasefrom the cell column, the normal calcium superfusatewas exchanged for “0” calcium buffer. To maintain normal calcium concentrations (2.5 mM) on the bioassayring, 250 mM CaC12was introduced below the column at a flow rate of 0.03 ml/min. After 5 minutes exposuresto “0” calcium buffer, the column was stimulated with an agonist and the tone of the bioassay ring monitored. After each“0” calcium experiment, the preparation was switched back to normal calcium and was retested to determine whether endothelial cell function had been altered irreversible. Calcium fluorescence measurements: Endothelial cells were harvested at confluence and resuspendedin Hank’s balancedsalt solution (Gibco) at 2 X lo6 cell/ml with 1.26mM calcium. After 30 minutes incubation with 10 PM fura-2/acetoxymethyl ester at 37’C, the cells were washed twice and resuspendedin a balanced salt solution containing 1 mg/ml bovine serum albumin.8 Changesin endothelial cell intracellular free calcium concentrationsin responseto agonists could be detected by monitoring the fura- fluorescencesignal. Fura- fluorescencewas measuredusing a SPEX Flurolog spectrofluorometer with excitation at 340 nm and emission set at 505 nm. To determine whether extracellular calcium contributed to the intracellular calcium transients, experimentswere performed in the presenceof 4 mM ethyl-

FIGURE 1. Bioassay of relaxing activity released from cultured bovine aortic endothelium. Representative tracings are shown, depicting grams of phenylephrineinduced (2 X IO-’ Rl) active tension in endothelium-denuded rings of rabbit thoracic aorta, which were super-fused with the effluent from the bovine aortic endothelial cell column, as described in Methods section. At the times marked, either 1 X lo-’ M bradykinin (BK) (A and B) or 3 &ml melittin, (C and 0) were infused into the endotheiial cell column. In A and C, the super-fusing buffer contained 2.5 mEq Cat+. In B and D, a Ca++-deficient buffer super-fused the bovine aortic endothelial cell, and 2.5 mEq of Cat+ was added to the effluent.

eneglycol-bis ,&aminoethyl ether)-N,N’-tetraacetic acid (EGTA) to chelate extracellular calcium. Intracellular calcium concentrationswere calculated as previously described.8 lnositol trisphosphate: Confluent endothelial monolayers were labeled with 3H-myoinositol (10 pCi/ml) for 24 hours in normal growth medium. Before assay, the cells werewashedtwice with physiologic salt solution and incubated for 20 minutes in lithium chloride (10 mM). Agonists were then added for specified times and the reaction stoppedby addition of ice-cold perchloric acid (500 mM). The 2 isomers of inositol trisphosphate(Ins[ 1,4,5]Psand (Ins[l,3,4]Ps)-were isolated by highperformance liquid chromatography as described-previously.l2 RESULTS

Releaseof EDRF from the endothelial cell column was detectedby relaxation of a preconstrictedendotheliurn-denudedrabbit aortic ring suspendedbeneaththe cell column. Stimulation of EDRF releasewith either bradykinin (0.1 PM) or melittin (3 pg/ml) produced transient relaxations of the bioassayring (Fig. 1, A and C). EDRF releasefrom the cell column was unaffected by pretreatment with indomethacin (28 PM), and no relaxations were seenwhen bradykinin or melittin was infused over the bioassay tissue in the absenceof endothelial cells (data not shown). When extracellular calcium was removed from the buffer superfusing the cell column (“0” calcium buffer) for 5 minutes, bradykinin and melittin were still able to induce EDRF release.The relaxations were more transient and in the caseof bradykinin, reducedby 56 f 17% (averageof 3 experiments)(Fig. 1, B and D). The magnitude of the melittin relaxation wasnot altered by calcium deletion. The changes in relaxation profile were most likely due to alterations in EDRF releasefrom the cell column since the calcium concentration in the buffer bathing the ring was kept constant. To determine whether changesin intracellular calcium might be associatedwith EDRF release,dispersed

60.B BK 5.r 43. 21. OJ. 0

Melittin -D

I

-Calcium ..

. . . . ..I. 5 10

. . . . . . . ..!............... 0 5 15

10

, 15

I 20

TIME (min) THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5,1988

376

A SYMPOSIUM: CALCIUM, THE CARDIOVASCULAR SYSTEM AND THE KIDNEY

endothelial cellswere loadedwith the fluorescentcalcium indicator dye, fura-2. Changes in intracellular calcium within the endothelial cell in responseto melittin and bradykinin could be detected by monitoring the furafluoresencefrom cell suspension.Typical fura- fluorescenceresponsesto bradykinin and melittin are shown in Figure 2. Note that the bradykinin-induced calcium signal wasbiphasic, with a rapid transient peak falling off to a sustainedplateau. In contrast, melittin induced a rapid and sustained increase in fluorescence that was maintained for at least 10 minutes. Pretreatment with EGTA to remove extracellular calcium for 1 minute before the addition of bradykinin or melittin to the cell suspension decreasedthe basal intracellular free calcium. EGTA eliminated the sustainedplateau in the bradykinin transient, with almost no effect on the initial rapid increasein free calcium, suggestingthat the source of calcium was from an internal storagesite and that extracellular calciurn was responsiblefor the sustainedplateau (Fig. 3). In

contrast, there was no change in the melittin-induced transient after EGTA pretreatment, suggestingthat melittin was able to sustain a large increasein intracellular calcium entirely from an internal source. Ins( 1,4,5)Psis known to act as a secondmessengerto releasecalcium from internal storeswithin a wide variety of cell types.” It was therefore of interest to determine whether bradykinin and melittin could stimulate endothelial cell phospholipaseC to produce this compound and, if so, whether it could be correlated to the calcium transients and EDRF release. In confluent endothelial cells prelabeled with 3Hmyoinositol, both bradykinin and melittin stimulated rapid increase in Ins( 1,4,5)Ps (Table I). The Ins( 1,4,5)Ps transients could be correlated closely with the calcium transients (Fig. 2 and 3). Bradykinin produced a 2.75fold increase in Ins(1,4,5)P3 at 30 seconds, which had returned to basal levels by 10 minutes. Melittin, on the other hand, stimulated a 1.87-fold increase in

FIGURE 2. Fura- measurements of cytosollc Ca++ concentration in suspensions of (bovine aortii endothelial cells). Bovine aortic endothelial cells (1 X 106/ml) were suspended in buffer containing 1 mM CaCla. Where indicated, 3 X lo-* M bradykinin (BK) and 3 &ml melittin (MEL) were added to the cuvette. At the end of the run, the cells were lysed with 50 mM digitonin, to saturate the fura- with Cat+, and then 1 mM manganous chloride was added to quench the fluorescence. Cytosolic calcium concentrations were caleulated as previously described.* ----==t

0

100

200

300

4DO

TIME (set> ,

I:-

320

22

180

L-J

EK

MnClz I W

100

200

300

TIME (set> 386

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62

400

FIGURE 3. Effect of EGTA on cytosolii cakium transients in fura-2-loaded bovine aortic endothelial ceils. Cells were suspended in the same buffer as in Figure 2. At the time indicated, 2 mM EGTA was added to the cuvette, followed by 3 X 10e8 M bradykinin (BK) and 3 &ml melittin (MEL). Before lysing the cells with 50 mM digitonin, an excess of Ca++ (5 mM) was added back to the cuvette.

TABLE I Effect of Bradykinin and Melittin on lnositol Trisphosphate

in Endothelial Cells

Percent Increase After Stimulation (100% = Basal) 30 Seconds

600 Seconds

Drug

lns(1,4,5)P3

lns(1.3,4)P3

lns(1,4,5)P3

lrlS(1,3,4)P3

Bradykinin (1 PM) Melittin (3 pg/ml)

275 rL-65 185&24

216 f 72 188rt35

98fr24 233 f 43

199*46 230 f 171

Endothelial stimulation.

cell monolayers were labeled with 3H-myolnositol for 24 hours (IO &i/ml). washed with physiologic buffer, and preincubated with lithium chloride for 20 minutes lnositol phosphates were extracted into perchloric acid and neutralized before analysis by high-performance liquid chromatography.

Ins(1,4,5)Ps, which was maintained for at least 10 minutes. The maintenance of elevated Ins(1,4,5)P3 was unexpectedbut correlated well with the sustainedcalcium transient stimulated by melittin. These data suggest that an increase in intracellular calcium in response to bradykinin and melittin could be a result of Ins(1,4,5)Ps-induced calcium release from internal stores. Bradykinin and melittin also induced an increase in Ins(1,3,4)Px at 30 seconds,which was sustained for at least 10 minutes with both agents. DISCUSSION

The regulation and control of EDRF releasehas been a difficult problem to study. However, most investigators concludethat calcium most likely doesplay an important role in this process.1J~4~5J Theseinvestigatorsshowedthat EDRF responseswere significantly attenuated or abolished when extracellular calcium was deleted from the tissue-bathingfluid. In contrast, Loeb et al6 and Dusting and Macdonald7 reported that endothelium-dependent relaxation in responseto melittin and arachidonic acid wasnot inhibited by elimination of extracellular calcium. In addition, we8 and others9 have shown that calcium transients within endothelial cells were not abolished after stimulation with agonists that release EDRF under “0” calcium conditions, or in the presenceof EGTA.* In the present study, we have demonstratedthat EDRF release was correlated with increased levels of Ins(1,4,5)Psand intracellular calcium in cultured endothelial cells. A recent study by Pirotton et ali3 has shown a correlation betweenIns( 1,4,5)IPsand intracellular calcium in endothelium in responseto adenosinediphosphate and adenosinetriphosphate. As shown in Figure 1, bradykinin and melittin were able to releaseEDRF from cultured endothelium evenin the absenceof extracellular calcium. In the caseof bradykinin, neither the relaxation or the calcium transient was as pronouncedin the absenceas in the presenceof calcium. In addition, EGTA pretreatment abolished the sustained plateau in fura- fluorescence after bradykinin stimulation. Although the magnitude of relaxation and the fura- signal after melittin stimulation were not affectedby calcium depletion, the time courseof the relaxation was shortened.These data suggestthat the magnitude of the total responseand perhaps the sustainedrelease of EDRF does require an influx of extracellular calcium. More important, the initial signal for EDRF production/release from the endothelium appearsto be an in-

before

creasein intracellular calcium. Bradykinin, a receptormediated agonist, and melittin, which may not be receptor mediated,both stimulate phospholipaseC to produce an increasein Ins( 1,4,5)P3,the well-characterizedsecond messengerknown to releasecalcium from internal storage sites.” It is well known that bradykinin-induced stimulation of phosphoinositidemetabolism is calcium-independent in endothelial cells.i4J5 Ryan et ali6 recently reported that several hormones that release EDRF stimulate an increasein intracellular calcium.16Therefore, it was no surprise that bradykinin was able to activate phospholipase C and mobilize intracellular calcium, even in the absenceof extracellular calcium. However, it was a novel finding that in the cells tested, the entire physiologic responsepathway wasintact after depletion of extracellular calcium. This finding held true in the caseof the bee venom peptide, melittin, as well. Melittin was interesting in that after endothelial cell stimulation, phospholipaseC activation was either continuous, or Ins( 1,4,5)Psbreakdown was inhibited, for at least 10minutes, assuggestedin Table I. This finding was consistentwith both the fura- data, showing a sustained increase in intracellular calcium, and with EDRF production, as shown by a more prolonged relaxation of the bioassayring. The mechanism behind this sustainedincreasein Ins( 1,4,5)P3levels and intracellular calcium is being investigated. Although extracellular calcium is not required for EDRF production or release,it doesappear to be necessary for sustained production of the factor. This was demonstratedby the attenuation of both the magnitude and duration of bradykinin-induced EDRF release,and by the shortenedduration of the melittin-induced relaxation under “0” calcium conditions. In the caseof bradykinin, the sustainedplateau of the calcium transient in the presenceof extracellular calcium (Fig. 2) may be associated with the more pronouncedmagnitude and duration of EDRF release. The influx of extracellular calcium would be attenuated by calcium entry blockers, which have also been shown to decreaseEDRF responsesto certain agonists.1,8 Rubanyi et all7 reported that the calcium channel agonist, BAY K 8644,stimulated the release of EDRF from canine femoral artery. Others have failed to demonstratethe presenceof calcium channelsin endotheliumi8 or blockadeof EDRF releaseby calcium entry blockers.3JyThesediscrepanciesmay be explained by the findings of Toyo and Bevan who showedthat the dependenceof EDRF on extracellular calcium varied along the

THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5,1988

396

A SYMPOSIUM: CALCIUM,

THE

CARDIOVASCULAR SYSTEM AND THE KIDNEY

length of the aorta and its branches.Theseresults suggest heterogeneity among endothelial cells with regard to mechanismsthat regulate calcium influx. There is evidencesuggestinga role for sodium/calcium exchangein the releaseof EDRF. l9The increasein intracellular calcium induced in endothelium by these various agonists results in an activation of phospholipaseAz21 and the releaseof arachidonic acid, which is metabolizedto prostacyclin. In conclusion,it appearsthat agoniststhat are known to releaseEDRF from cultured endothelial cells interact with membranecomponentsthat are coupledto phospholipase C. PhospholipaseC activation stimulates the formation of inositol trisphosphate, which triggers the release of calcium from intracellular storage sites, as has been demonstrated in other systems.” The elevation of intracellular free calcium initially through Ins( 1,4,5)Ps and then through calcium influx from outside the cell stimulates releaseof the relaxing factor. REFERENCES 1. Singer HA, PeachMJ. Calcium- and endothelial-mediated vascular smooth muscle relaxation in rabbit aorta. Hypertemion 1982:4(suppl II):1 9-25. 2. Long CJ, Stone TW. The release of endothelial-derived relaxing factor is calcium-dependent. Blood Vessels 1985;22:205-208. 3. Miller RC, Schoeffter P, Stoclet JC. Insensitivity of calcium-dependent endothelial stimulation in rat isolated aorta to the calcium entry blockerjlunarizine. Br J Pharmacol 1985:85:4%1-487. 4. Rapoprt RM, Draznin MB, Murad F. Mechanisms of adenosine triphosphate-. thrombin- and trypsin-induced relaxation of rat thoracic aorta. Circ Res I984;55:468-479. 5. Griffith TM, Edwards DH, Newby AC, Lewis MJ, HendersonAH. Production of endothelium derived relaxantfactor is dependent on oxidativephosphorylation and extracellular calcium. Cardiouasc Res 1986:20:7-12. 6. Loeb AL, JohnsRA, PeachMJ. Phospholipase activation by melittin releases

406

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62

prostacy& and an endothelium-derived relaxing factor from rabbit aorta (abstr). Blood Vessels 1986;23:86. 7. Dusting GJ, Macdonald PS. Endothelium-dependent vasodilation: role of beta-adrenoceptors, calcium, and cytochrome P-450 (abstr). Blood Vessels 1986;23:66. 8. PeachMJ, Singer HA, Izzo NJ Jr, Loeb AL. Role of calcium in endotheliumdependent relaxation of arterial smooth muscle. Am J Cardiol 1987;59:3SA43A. 9. Luckhoff A, BusseR. Increasedfree calcium in endothelial cells under stimulation with adenine and nucleotides. J Cell Physiol 1986;126:414-420. 10. Loeb AL, Johns RA, Milner P, Peach MJ. Endothelium-derived relaxing factor from cultured eelIs. Hypertension 1987;9:suppl IH:l86-192. 11. Berridge MJ, Irvine RF. Inositol trisphosphate, a novel second messenger in cellular signal transduction. Nature 1984:312:315-321. 12. Johnson RM, Garrison JC. Epidermal growth factor and angiotensin II stimulate formation of inositol 1,4,5- and inositol 1,3,4-trisphosphte in hepatocytes: differential inhibition by pertussis toxin and phorbol 12-myristate 13acetate. J Biol Chem 1987:262:17285-l 7293. 13. Pirotton S, Raspe E, Demolle D, Erneux C, BoeynaemsJ. Involvement of inositol 1,4,5-trisphosphate and calcium in the action of adenine nucleotide on aortic endothelial cells. J Biol Chem 1987;262:17461-17466. 14. Derian CK, Moskowitz MA. Polyphosphoinositide hydrolysis in endothelial cells and carotid artery segments. J Biol Chem 1986;261:3831-3837. 19. Lambert TL. Kent RS. Whorton AR. Bradvkinin stimulation of inositol polyphosphate production in porcine aortir eniothelial cells. J St%1 Chem 1986:261:15288-1529.7. 16. Ryan US, JohnsA, Van BreemanC. Role of calcium in receptor-mediated endothelial cell responses. Chest 1988;93;105S-1 O&Y. 17. Rubanyi GM, Schwartz A, VanHoutte PM. The calcium agonist Bay K8644 and (+)202,791 stimulate the release of EDRFfrom caninefemoral arteries. Eur J Pharmacol 1985;107:143-144. 18. Colden-Stanfield M, Schilling WP, Ritchie AK, Eskin SG, Navarro LT, Kunze DL. Bradykim’n-induced increases in cytosolic calcium and ionic currents in cultured bovine aortric endothelial cells. Circ Res 1987,61:632-640. 19. Winquist RJ, Bunting PB, Schofield TL. Blockade of endothelium-dependent relaxation by the amiloride analog dichlorobenzamil: possible role of Na+/ Caz+ exchange in the release of endothelium-derived relaxant factor. J Pharmacol Exp Ther 1985:235:644-650. 20. Tayo FM, Bevan JA. Extracellular calcium dependence of contraction and endothelium-dependent relaxation varies along the length of the aorta and its branches. J Pharmacol Exp Ther 1987;240:594-601. 21. Hong SL, Deykin D. Activation ofphospholipase A2 andphospholipase C in pig aortic endothelial cells synthesizing prostacyclin. J Biol Chem 1982; 257:7151-7154.