Role of Intracellular Alkalosis Induced by Activation of the Protein Kinase C-Dependent Na+-H+ Exchanger

269

Endothelin and Increased Contractilit in Adult Rat Ventricular Myocytes

Role of Intracellular Alkalosis Induced by Activation of the Protein Kinase C-Dependent Na+-H+ Exchanger

Bernhard K. Kramer, Thomas W. Smith, and Ralph A. Kelly

Endothelin,a21-aminoacid vasoactive peptide, isamongthemost potentpositively inotropic agents yet described in mammalian heart. Having demonstrated that endothelin's inotropic effect is due, in part, to an apparent sensitization of cardiac myofilaments to intracellular calcium,wedetermined whether this could be duetoarise inintracellularpH(pHi).Inisolated adult rat ventricular cells loaded with the H+-selective fluorescent probe BCECF, 100 pM endothelin increased contractile amplitude to 190±26% of baseline and pH; by 0.08 0.02 (n=8), whereas 1 nM endothelin increased pHi by 0.13+±0.03 with little further increase in contractility. Amiloride (10-4 M) prevented theincrease inpH; inresponse toendothelin and reduced theinotropic^response by45%, although theinotropic effect could be readily restored by subsequentNH4Cl-induced alkalinization. Similarly, inhibitors ofproteinkinase C (H-7and sphingosine) diminished^orabolished the rise inpH;afterendothelinsuperfusionwhilecausing

a decline in its inotropic effect comparable with that observed with amiloride. Pretreatment withpertussis toxin, which^wehavedemonstrated results incompleteADP-ribosylation ofthe a-subunits of

G.

^andGiGTP-bindingproteins andabolition ofendothelin's positive inotropic effect,onlypartiallyreduced the intracellularalkalinizationinducedbythepeptide, suggesting

acomplexsignal transductionmechanism. Thus, thepositive inotropic action ofendothelin is due in part to stimulation of the sarcolemmal Na'-HW exchanger by ^a protein kinase C-mediated pathway, resulting in a rise in pHi and sensitization of cardiac myofilaments to intracellular Ca2+. (Circulation Research 1991;68:269-279)

T he regulation of the contractile state of the myocardium under normal physiological conditions has long been assumed to be dependent on the extent of mechanical loading of muscle fibers and the degree of activation of the autonomicnervoussystem.Endothelin,apotentvaso- constrictor peptide originally derived from media bathing primary cultures of porcine aortic endothe- lial cells, has recently been reported to act as a

positive inotropic agent in mammalian atrial and ventricular muscle. These reports1-5 suggest that local factors, perhaps released by the microvascular endotheliumand/or endocardium,mightalsodirectly

From the Cardiovascular Division, Departments ofMedicine, Brigham and Women's Hospital and Harvard Medical School, Boston,Mass.

SupportedbygrantsHL-19259 and HL-36141 from the National Institutes ofHealth,byaFacultyDevelopmentAward from the Pharmaceutical Manufacturers Association (R.A.K.), and by a grantfrom thePaul-Martini-Stiftung, Bonn,FRG(B.K.K.).

Address forcorrespondence: RalphA.Kelly, MD, Cardiovas- cularDivision,BrighamandWomen'sHospital,75FrancisStreet, Boston, MA 02115.

Received March8,1990;acceptedSeptember 18, 1990.

regulate myocardial contractile function. Indeed, in isolated ^rat ventricular myocytes, endothelin is the most potent positive inotropic factor we have yet tested,withanEC50 inthis system of 50 pM,3avalue that approximates the reported Kd of endothelin for its receptor in either intact heart or in sarcolemmal membranes.6-8 In addition, the kinetics of endothe- lin'seffectonmyocytecontractility^areunusual, with aslowonsetandsustained duration of action in both single cell and intact tissue preparations, indicating that the peptidemaymodulate theinotropicrespon- siveness of the heartover atime frame of minutesto hours, incontrast totherapidonsetandshorter-lived effects of knownneurohumoral factors.3

The mechanism by which endothelin induces a positive inotropiceffect in cardiac myocytes remains controversial. Although an increase in cytosolic cal- cium has beenobserved in isolated rabbit myocytesat endothelin concentrations above 10 nM,9 we have reported that inisolated, freshlydissociated adultrat ventricular myocytes, endothelin increased contrac- tile amplitude with little or no increase in cytosolic calcium, appearing to enhance myofilament respon-

siveness to calcium via a pertussis toxin-sensitive pathway.3 At concentrations at or below 1 nM that would elicit amaximal increase incontractile ampli- tude, there was little or no increase in the peak systolicordiastoliccalciumlevelduring intracellular calcium transients in myocytes paced at 1.5 Hz, although some increase in intracellular calciumwas variably observed at endothelin concentrations at or above 1 nM.

One potential mechanism by which endothelin could enhance myofilament responsiveness to cal- cium would be the induction of an intracellular alkalosis, since changesinintracellular pH(pHi) are knowntoaffect thecalciumsensitivityof thecontrac- tile apparatus.10-12 Endothelin has been shown to increase

pHi

in glomerular mesangial cells and vas- cular smooth muscle cells, probably by activation of the Na+-H+ antiporter intheplasma membrane.13,14 The purpose ofthepresent studywas toexaminethe effects of endothelin on

pHi

in adult rat cardiac myocytes and to determine whether the increase in contractile amplitude observed with the peptide could be due in part to an increase in sarcoplasmic pH. For the purposes of this report, "endothelin"

refers to the originally described porcine/human endothelin, now termed "endothelin-1."'15

Materials and Methods

Calcium-tolerant isolatedratventricular cellswere preparedusingavariation of the methods described byHaworth et al16and Cheunget al,17as previously described.18 Briefly, hearts from adult (225-250 g) male Sprague-Dawley rats (Charles River Laborato- ries, Wilmington, Mass.)^wereretrogradely perfused for 5 minutes with Krebs-Henseleit (KH) buffer, followed by a 5-minute perfusion with nominally

Ca`4-free

KH buffer and, subsequently, Ca2-free KHbuffer containing 0.05% collagenase and 0.03%

hyaluronidase for 25 minutes. Ventricles were then minced andincubatedinKHbuffercontaining1mM CaCl2, 0.002% trypsin, and 0.002% deoxyribonu- clease,andthecellswerereleasedbytrituration and sedimented in 2% bovine serum albumin. The cells were allowed to attach to 12-mm glass coverslips coated with collagen (Vitrogen, Collagen Corpora- tion, Palo Alto, Calif.) for contractility and fluores- cence measurements.

Fluorescence Spectroscopy

Fluorescence measurements were performed in a SPEX CM2 dual excitation spectrofluorometer (SPEX Industries, Edison, N.J.) as previously de- scribed.319 Coverslipswith attached cellswereplaced inawater-jacketed chamber(37+0.5°C)on thelight- shielded, heated stage of themicroscope and super- fused with buffer at a rate of 1 ml/min. Coverslips were placed in an aluminum chamber, 25 mm in diameter, with a volume of 300

gl.

The contents of thechambercould beefficiently exchanged within 30 seconds, or more quickly if necessary, using tran- siently higher flow rates. Because of potential ad-

verse effects of constant ultraviolet light exposure, exposures to the excitation light beam (and thus determination of pHi) were limited to brief "snap- shots" of several seconds after each experimental perturbation, with total exposure to the excitation beam limited toseveral minutes for each experiment.

Freshly isolated adult ventricular myocytes were loaded with the pH-sensitive fluorescent dye 2',7'- bis(carboxyethyl)-5,6-carboxyfluorescein (BCECF, Molecular Probes, Eugene, Ore.) by permitting the cells to attachtocoverslips,incubatingthem in super- fusionbuffer(seebelow)with 2,uMBCECF/AM, the acetoxymethyl esterderivative of BCECF, for 20 min- utes at 23°C, and then washing them. A 1 mM BCECF/AM stock solution was prepared in dry di- methylsulfoxide and kept frozen in aliquotsuntiluse.

Aftereachexperiment usinga single cell (onlyone experiment was performed per coverslip), an in situ calibrationof the BCECF fluorescent signal was per- formedwithin eachcell. Myocyteswere calibrated by superfusion with two or three callbration buffers as previously described.'9 Calibration buffers differed from the superfusion medium and containedHEPES/

KOH (4 mM); EGTA (0.5 mM); pyruvate (5 mM);

glucose (5.6 mM); K2ATP (10 mM); the ionophores nigericin (20 ,uM), ionomycin (4 ,M), and carbonyl cyanidem-chlorophenylhydrazone (CCCP)(0.2 ,uM);

and KCl to yield a finalK' concentration of 150. pH wasadjustedat37°C with KOHorHClto6.60 or7.50, respectively. The pHi for each cell was then deter- mined from a linear regression offluorescence ratio versus the pH value of the calibration buffer.

ContractilityMeasurements

Measurements of contractile amplitude of single isolatedratcardiomyocytes without simultaneous ac- quisition of pHidatawere made on the stage of an inverted phase-contrast microscopeaspreviously de- scribed.8"19 Myocyteswerestimulated at25% above threshold witha3-msec square-wave pulsethrough a platinum wire placed in the superfusion liquid con- nected to a stimulator (model S88, Grass Instru- ments, Quincy, Mass.).

Afterequilibration with the superfusion buffer and stabilization of myocyte contractile amplitude at a stimulation rate of 1.5 Hz, the cardiocytes were superfused with buffers containing 100 pMendothe- lin for 7 minutes to obtain a maximal increase in contractile amplitude. Consequently,thesuperfusion time of other drugs was increased proportionately (100 ,mM amiloride or 10 ,M ouabain with or without endothelin),with similartimes of exposure to ultra- violet light, asdescribed in the text, table, and figure legendsbelow. Those cells exposed to varying extra- cellular pH(pH), withorwithout prior exposure to 100 pM endothelin, were exposed to superfusion buffer at each pHO for 3 minutes. Similarly, cells made alkalotic and subsequently acidotic by initial exposure to, and thenwashout of, NH4Cl were first allowed to stabilize for 3 minutes in superfusion buffer at a pHO of 7.4; stabilization was followed by

Krameretal Endothelin-Induced AlkalosisinCardiocytes exposure to 10 mMNH4Cl for 3 minutes and then to

superfusion buffer alone,

pH,

7.4, for another 6 minutes. A similar protocol was followed in which drugs were added before the NH4Cl prepulse tech- nique was begun; however, in this case, the cells were allowed to equilibrate in superfusion buffer plus endothelin and/or amiloride or in superfusion buffer plus endothelin and/or ouabain for 12 minutes. Un- less otherwise indicated, the superfusion buffer con- tained 137 mM NaCl, 4.0 mM KCl, 0.5 mM MgCl2, 0.9 mMCaCl2, 5.6 mM glucose, 4 mM HEPES, and 0.05%bovineserumalbumin at a pH of 7.40 at 37°C.

Forcellssuspended with a bicarbonate-buffered sys- tem, the superfusion bufferwas as described above, except that 20 mM HCO3 was substituted for HEPES and the NaCl concentration was 120 mM.

pH measurements were made using a pH meter (model 140, Corning, Medfield, Mass.) at 37°C, and pHwas adjustedwith NaOHorHCl.

Materials

Hyaluronidase type II, trypsin, bovine serum albu- min (bovine fraction V), amiloride, ouabain, isopro- terenol, D-sphingosine, and CCCP were purchased from Sigma, St. Louis; collagenase and deoxyribonu- clease I were from Worthington Biochemical, Free- hold, N.J.; the protein kinase inhibitor 1-(5-isoquino- linylsulfonyl)-2-methylpiperazine (H-7) was from Seikagaku KogyoCo.,Tokyo; ionomycinandnigericin were obtained from Calbiochem, La Jolla, Calif.; per- tussistoxinwasfromListBiologicals, Campbell, Calif.;

and endothelin (synthetic porcine/human endothe- lin-1)wasfromPeninsula Laboratories,Belmont, Calif.

Statistics

Datawereanalyzed byStudent'sttestforunpaired observations with a modified Bonferroni correction for multiple comparisons, repeated-measures analy- sis of variance forpaired observations, and Student- Newman-Keuls test for multiple comparisons. Data inthetextandfigures ^areexpressed ^asmean±SEM.

Throughout the article, unless otherwise stated, baseline ^or "basal" values are expressed as 100%

before administration of any drugs ^{or a} change in superfusion buffer pH, and all subsequent data are

expressed ^{as a}percent of baseline ^orbasal values.

Results Effects ofEndothelin onMyocyte pH, and Contractility

As ^we have shown previously,3 endothelin in- creased contractility inisolated ratventricular myo-

cytes; typically, ^no change in contractile amplitude

was apparent for approximately 4 minutes, and a

maximal response was present at 7 minutes. A re- sponsecouldbe observed in all cellsat1pM,andthe

EC50^was approximately 50pM; a maximalresponse was present at 1 nM with an increase of approxi- mately 90-100% over baseline contractility. In the BCECF-loaded cellsreported here, the time course

:cL

Pa

cJ u x

200l ⁰¹⁰

100 n

T

1p l00pA 1nM

[endothelin]

0

0r~

n0 _"

IE '.

2I I- °

C._

c

1u "is

10 nM

FIGURE 1. Bar graph showing changes in intracellular pH

(pHJ) and contractile amplitude after endothelin exposure.

Single cardiacmyocytesfreshly isolated from adultrat ventri- cleswereloaded with theH'-specificfluorescent dye BCECF and superfused with physiological media buffered with HEPES at extracellular pH (pHO) 7.4 over the range of endothelin concentrations indicated. Contractility, expressed asapercentage of the control preendothelin contractile am- plitude, at 8 minutes afteraddition of the peptide is shown in open bars, and the corresponding average change inpHiat 8 minutes isshowninsolid bars. ThepHiin myocytesatpHO7.4, stimulated at 1.5 Hz, was7.09±0.01 (n= 71), and this defines thebaselinepHivalue. TheBCECF intracellularfluorescence signal wascalibratedin situ at the end of each experiment as described in "Materialsand Methods."Averagedata for five to seven cells aregivenfor each point. Changes inpH,and contractility weremaximalat 1 nM. *p<0.05compared with valuesforcontractile amplitude

orpH,

before endothelin. The increase in contractility at 10 pMaveraged33% but did not reachstatisticalsignificance (n=4).

and magnitude of the contractile response to endo- thelin were similar to what we have reported in unloaded orfura 2-loaded cells.3

In adultratventricular myocytes bathed in superfu- sionbuffer atpH 7.4, the baseline

pH,

^was

7.09±0.01

(n=71 cells) incells stimulated at 1.5 Hz. The pHiin resting, nonstimulated cells was approximately 7.14±0.02. Asshown inFigure 1, endothelin increased myocyte pHi. This dose-response relation was shifted somewhat to the right comparedwith the increase in contractility; there was no significant response in

pH,

apparent at 10 pM. The maximal rise in pH corre- sponded to the maximal increase in contractility at 1 nM.Thetimecourseof the change inpHialsodiffered marginally from the increase in contractility; some increase inpHiwasalwaysapparentat2-3 minutes(as shown in Figure 2) although, like the contractility response, the rise in pHi was not maximal for 6-8 minutes(Figure2).Incontrast,noincrease incontrac- tilitywasapparent for about4minutes afterbeginning the endothelin infusion. When cellswere exposedto 1 nM endothelin in a bicarbonate-buffered superfusion medium, the increase inpHiand contractileamplitude were qualitatively similar to that seen with HEPES- buffered media. pHi increased 0.072 pH units, and contractile amplitude increased to 207% of control (n =3).Consequently,becauseof the need forprecision

.~~~~~~~~~~~

271

200 F,n0

0

o

OE'

...I

._ a

0q

_0

C)E

0

x

150+

100

2 ⁴

time(min)

0cO

01 0

C2E^1*

,o

6 8

FIGURE2. Graphshowingtimecourseof changesin intra- cellularpH(pHJ) and contractileamplitude withendothelin.

Freshly isolated ventricular myocytes were superfused with physiological bufferatextracellularpH(pH,)7.4, loaded with theH'-specificfluorescent dyeBCECF, stimulatedat1.5Hz, and exposed to 100pM endothelin at time 0. Changes in contractile amplitude ^are represented by open symbols, and changes in pHi by solid symbols. In neither case was the increasesignificant until4 minutes (n=6).

and^easeofreproducibilityofsuperfusionbufferpH,all subsequent experiments^were done withHEPES-buff- ered media. To exclude thepossibilitythat theincrease in pHi with endothelin ^was secondarily related to an

increase incontractilityper seinthismodel,cellswere

exposed to aconcentration ofisoproterenol (1.25-5.0 nM) that resulted in an increase in contractileampli- tude roughly equivalent to that produced by 100 pM endothelin. Isoproterenolat thisconcentration had^no effectonpHi(-0.01±0.01of baselinepHi,n=3). Also, ouabain (10-4-10-3) routinelytended to decrease pH, inmyocytes superfused atpH0 7.4despite acompara-

ble increase in contractileamplitude.

Effects ofVarying

pH,

andNH4Cl on

Myocyte Contractility

To examine therelation betweenpH andchanges in myocyte contractile amplitudein isolatedratven-

tricularmyocytes paced at 1.5 Hz, pHi^was manipu- lated eitherby varying pH0 orbytheNH4Cl prepulse technique^as described in"Materials and Methods."

Varying pH0 hadapredictable effect^onmyocytepHi;

increasing pHO to 7.9 from 7.4 increased pHi to 7.26+0.03.Thechangesincontractileamplitudewith varying pH0 are shown in Figure 3A. Extracellular alkalinization resulted in amarked increase in ^con- tractile amplitude, with little or no decrease in^con- tractility ata

pH.

of6.9. To determine the effects of priorexposuretoendothelin ^onsubsequentintracel- lularalkalinizationor acidificationduring changes in pH,myocytespretreatedwith100pMendothelin for 7 minutes were subsequently exposed to the ^same protocol asthecontrol cells describedabove(Figure 3A). In this figure, the initial contractile amplitude

was assigned a value of 100%. The increase in contractile amplitude with endothelin at 7 minutes averaged 190+26% of baseline (n=8). As illustrated here, endothelin-pretreated cells hadonly ^a modest further increase incontractile amplitude after^expo-

suretoa

pH,

^{of7.9and littleorno}further declinein contractility whenpHOwas reversed to 6.9.

In Figure3B,the changes in contractileamplitude

are plotted as a function of the changes in pHi inducedbyexposuretoand, subsequently, removal of NH4C1. Before exposure to NH4Cl, cells exposed to 100 pM endothelin increased theirpHiby 0.08±0.02 (p<0.05; n=6). The subsequent increases in pHi in both control cells and endothelin-pretreated cells

were similar. Interestingly, the negative inotropic effect ofadecline inpHi^onNH4C1 washoutwasmore

pronounced inendothelin-pretreated cells, fallingto 53% of the contractile amplitude in endothelin- pretreated cells measured just before addition of NH4Cl (open circle with ##). Thiswas in spite ofa

lower

pHi

in control cells. Subsequently, there^{was a} marked rebound of pHi in endothelin-pretreated cells aftertherecoveryfrom NH4CI washout-induced intracellular acidosis, resultingin an increasein

pHi

to7.25±0.05 (n =6)at 9 minutes afterremoval ofthe initial prepulseofNH4C1. This rebound intracellular alkalosis ^was unassociated with any comparable ^re- bound increase incontractility (Figure 3B).

Endothelinand Inhibition ofNa+-H+ Exchange To determine the potential role of the sarcolem- malNa+-H+ exchangerin theintracellular alkaliniza- tion after endothelin infusion,cells^wereexposed to 100pMendothelinwith^orwithoutpretreatmentwith 100 ,M amiloride. Amiloride alone at this ^concen- tration had^noeffectonthecontractilityof thesecells pacedat 1.5Hz andsuperfusedatpH7.4 (103±12%

ofbaseline, n=15, at5 minutes). Amiloride also did notsignificantly affectthepHiofcellspacedat1.5 Hz at a superfusion buffer pH,, of 7.4 (-0.03±0.03,

n=4).Unlike the reportedeffect ofamilorideonthe fluorescence intensity of6-carboxyfluorescein, addi- tion of 10-4 M amiloride did not affect the fluores-

cenceratio ofBCECFinthese cells.20,21

The addition of 100 pM endothelin to cells pre-

treated with

10`

^Mamiloride resultedin^anincrease in contractile amplitude that averaged 149±17% of baseline (n=7), about 45% below that seen with endothelin alone. However, the intracellular alkalin- ization observedwith 100 pM endothelin was com-

pletelyblocked with amiloride (-0.07±0.03, p=NS, n=6). The effects ofvarying pHO orofNH4Cl addi- tion and washout on contractile amplitude of amiloride-pretreated cells, with and without endo- thelin,areshown inFigures4Aand4B,respectively.

In these figures, baseline contractile amplitude (100%) is the contractile response of isolated rat cardiac myocytes pacedat 1.5 Hz after5 minutes of superfusion with buffer containing 100 ,uM amiloride. Myocytes pretreated with amiloride alone (n=7) showed no enhanced contractile response to

superfusion with buffer at pHO 7.9 in contrast to control cells (Figure 3A). However, in cells pre-

treated with amiloride and endothelin at pH 7.9, there ^{was a} marked increase in contractility at this pH0 that was similar to that seen in nonamiloride-

T 9

T 0l

0 T*

9 * 1 1

0 O

0 0

i v

c

Krameretal Endothelin-Induced Alkalosis in Cardiocytes

A.

400- pH7.4 pH 7.9 pH 7.4 pH6.9 pH7.4

300- 1^{, i*}

200- _{0 max.@ 00o0}^{T T}

0 i.⁰

0~~~~~0 ^.0

lo.n ⁰

basal 0 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 time(min)

B.

5uu,

0

*=_0.E

0- a-0

cJ ^a

X0

400.

3004 2004 100~

-0.30 -0.20-0.10 0.00 0.10 0.20 0.30 0.40 0.50 ChangeinpHi

FIGURE3. Graphs showing the effect of changingintracellular pH (pHI) ^onmyocyte contractility. Single isolated ventricular myocyteswereloaded with theH'-specificfluorescent dye BCECF, attachedtocoverslips for simultaneousacquisitionofpHiand contractility data at37°C, and stimulated at 1.5 Hz (see 'Materials and Methods"). Panel A: After equilibration with the superfusionbuffer, thecontractile amplitude^wasmeasured and assignedavalue of 100% ("basal" contractility). The cardiocytes

werethensuperfused with media alone (-; n=8)orwith 100pMendothelin (o;ⁿ⁼7) atpH7.4for7minutes(time0).Endothelin (10OpM)typically increased contractile amplitudeto190-200% ofbasal levels. Subsequently, the^myocyteswereexposed^tomedia with^orwithout 100pMendothelinatthe extracellularpHvalues indicated. *p<0.05 compared with basal contractile amplitude.

Panel B:After equilibrationatextracellularpH 7.4, the^myocytesweremadealkalotic and then acidotic by initial^exposureto10 mMNH4Clfor 3 minutes and by itssubsequent washout. For both control cells (a; n=8) and cells pretreated with 100 pM endothelin (o; n=8), the relation betweenpHiandcontractility isshown duringthe alkalinization and subsequent intracellular acidification. ThebasalpHiforall cells(n=16)^was 7.09, and this point (#)^wasassigned^aApHivalueof 0.00 and^acontractility valueof100%. Eight cellswerethen exposedto100pM endothelin, andpHiandcontractilitywereallowedto equilibrate for 8 minutes(##) before the addition ofNH4Clto thesuperfusion media. The subsequent changesinpHI andcontractile amplitude

areillustrated inendothelin-pretreated cells and controls cells duringNH4Clexposureand washoutasdescribed above.pHiand contractilitywererecordedat90-second intervalstominimizeexposuretoultravioletexcitationlight. Although^errorbars for each pointfor bothparameters^areexcludedfor clarity ofpresentation,anychangeinpHigreaterthan0.05wassignificantatthe p<O.O05 level; thepH, ofagroupof cells^atagiventimepointwascompared with that group's baseline predrug pH by paired analysis with

astandard Newman-Keulstest todetermine the levelofsignificance. Incontrast,thefurtherincreaseincontractileamplitude in the endothelin-pretreated cells above that caused by endothelin alone (100pM;increase inpHI of 0.08+0.02; before NH4Clexposure

[##]) ^wassignificant onlyatthepoint ofmaximal alkalosis (ApHiof +0.38+0.01; p<O.O05) afterNH4Clexposure. Noneof the changesincontractileamplitudeineithercontrol^orendothelin-pretreated cells^wassignificantly differentfrom baselineafterNH/Cl washout.

pretreated cells (Figure 3A). Subsequent acidifica- tion with buffer at

pH.

6.9 resulted in only an

insignificant fall in contractile amplitude in both amiloride (alone) and amiloride-plus-endothelin- pretreatedcells(Figure 4A).

Withintracellular alkalinization inducedbysuper-

fusion with buffer containing NH4C1 (Figure 4B), there was an increase in myocyte contractility in amiloride-pretreated cells that was approximately half of that observed in control cells not pretreated with amiloride, as shown in Figure 3B. This is con-

sistent with the fact that amiloride with NH4C1 blunted the increase inpHibyabout50% compared with nonamiloride-pretreated cells (note the change in scale ^on the pH axis in Figure 4B). Intracellular alkalinization with NH4C1 ⁱⁿ endothelin- and amiloride-pretreated cells resulted in an increase in contractile amplitude equalto that ^seeninendothe- lin-pretreated cells not exposed to amiloride. This marked increase incontractileamplitude with endo- thelin occurred despite the fact that amiloride with

NH4C1 blunted the risein pHiin endothelin-treated cells, as incontrol cells, byabouthalf, imply

A. B.

400- pH7.4 pH 7.9 pH 7.4 pH6.9 pH7.4

300- 1 X

100 ^A v *

1 .

200-T max. T 4

AA A Aj

1001 A A A A 4I 4'A _{A 4AA}

0bosal 0 1.5 3.0 _4.5 6.0 7.5 9.0 10.5 12.0 time(min)

0

.. W

~CE_ a c cx c 0

oia

on₀

400- 1A

300-

200-- A~~~

10.40 - 0.A

-0.40 -0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 ChongeinpHi

FIGURE 4. Graphsshowingendothelin and inhibitionofNa'-H'exchange. Panel A: Singleisolatedventricular myocyteswere

monitoredforchangesincontractileamplitude whilethe extracellularpH ofthesuperfsion buffer^wasvaried^asshown. Allcells

wereexposedto100MMamiloridefor12minutesbefore beginningtheprotocolattime0;^sevenofthese cellswerealsoexposed

to 100pM endothelin for the final 7minutes (A), and eight of these cells were exposed to amiloride alone (A). Amiloride

pretreatment alone had no significant effect ^on eitherpHi ^or baseline contractile amplitude ofmyocytespaced at 1.5 Hz at extracellularpH 7.4(-0.03+0.03;p=NS). Amiloridepretreatmentreduced thepositiveinotropicresponsetoendothelinbyabout 45% (to 149+17% of the contractile amplitude in amiloride-alone cells). Themeancontractileamplitude of bothgroupsof cells after 12 minutes ofpreincubation with either amiloride aloneoramilorideplusendothelin, pH 7.4, wasassignedavalueof 1(00%

attime0. *p<0.05in the amiloride- andendothelin-pretreatedcellscomparedwith baselinecontractility. PanelB:As described inthelegendtoFigure 3B, the relationbetweenchangesin intracellularpH(pH1)and contractileamplitudeisshownformyocytes pretreated with amiloride alone(A)^oramiloride and endothelin (A) duringexposureto10mMNH4Clfor 3minutesandafterits abrupt withdrawal (see "Materials and Methods'). For 10 amiloride-pretreated cells, the initialvalueforApHiwas0.00andthe valueforcontractility ^was100% (this pointis indicatedby #). Sixofthese cells werethen exposedto 100pMendothelinfor 8 minutes (##)beforeNH4Clwasaddedtothesuperfusionmedia. At thispoint, despitethe initialstatistically insignificantdecline

inpH,with endothelininamiloride-pretreatedcells(-0.07±0.03), contractilityincreasedto144±10%ofcontrol(p<0. 05). Both

groupsofcells (A and A)^werethenexposedto10mMNH4Cl andsubsequent washout,andrecordingsofpHiandcontractilitywere

madeat90-second intervals. As in Figure3B, errorbarsateachpointforbothparametershavebeen excludedfor clarity.Any change in ApHgreaterthan 0.08^wassignificant(p<0.05).Importantly, theincreases in contractileamplitudewithendothelinafter

exposuretoNH4Cl werehighlysignificantandequaledthe absoluteincreasesin contractileamplitudeseenincellseposedto100 pMendothelin withoutamiloride (open circles in Figure 3B). The subsequent decline in pH, and fall in contractile amplitude with NH4C/ washoutwassignificant and didnotrecoverin thecontinuing^presenceof10-4Mamiloride in bothgroupsof cells.

withNH4C1 is dependenton the capacity ofthe cell tobuffersubsequent changes incellularH+ content, because some intracellular accumulation of H+

would have occurred during amiloride pretreatment inthe absence of another H+ extrusion mechanism,

asinthe nominal absence of bicarbonate under these experimentalconditions. Theslower declinetobase- line pHi in amiloride-pretreated cells continuously superfused with NH4C1 is also explained in part by H+ accumulation, since this decline isdependent on

the electrochemical gradient forNH4+ as well as its permeability, which, although less than NH3, is still substantial in thesecells.23324

Ouabain, Endothelin, and Myocyte

pHi

To compare the response of endothelin-treated isolated rat ventricular myocytes to changes in

pHi

with theirresponse afterpretreatment with another positive inotropic agent, myocytes were superfused witheither 10O Mouabainfor 12minutesor10- M ouabain for 5 minutes; superfusion was followed by the addition of100pM endothelin for an additional 7minutes(datanotshown). Higherconcentrations of ouabain alone (10-_10--3 M) resulted in initial in-

creases in myocyte contractile amplitude to 160-

200% of baseline; these increases ^were not associ- ated with ^any significant fall in pHibut often ^were followed by the development ofcontracture either beforecompletionofthe protocolsorduring calibra- tion of the BCECFfluorescence signal.After^astable increase in myocyte contractile amplitude had been achieved, myocytes were exposed to superfusion buffer containing either

10`

M ouabain alone ^or ouabain plus endothelin; the sequential changes in buffer pHO ^{were as} described in Figure 3A, and NH4C1 prepulse and washout were as described in Figure 3B. Ouabain alone at 10-5M increased^con- tractile amplitude to 112±7% of baseline (n=11), but there ^{was no} significant change in

pHi

(-0.02+0.02,n=4). Addition of 100 pMendothelin resulted in ^anadditive increaseincontractile ampli- tudeto 191+22%ofthat^seenwithouabain

(10`

M) alone.

The increase in^myocyte contractilityonalkaliniz- ing the cell interior by eitherraisingpHi, ^or adding NH4C1 in cells pretreated with ouabain alone was

reduced comparedwith theincreaseincontrolcells,

even when the slightly higher baseline contractility

with ouabain is considered.This^mayhavebeendue

to a tendency toward lower cytoplasmic pH during

CU)

00

0 L .6

0Cc0)_eijJ

k

Krameret al Endothelin-InducedAlkalosis inCardiocytes alkalinization in ouabain-pretreated cells compared

with control cells. For example, the maximum in- crease in pHi that was achieved after exposure to NH4C1 was +0.42±0.05 in control cells, compared with +0.35±0.02 inouabain-treated cells.

Ofinterest was a sustained increase in contractile amplitude in ouabain- andendothelin-pretreated cells that subsequently underwent intracellular acidifica- tionby either acidification of the externalmedium or NH4washout(data not shown). This is incontrastto the response of control cells or cells treated with endothelin alone, which demonstrated a fall in con- tractile amplitude toward or below baseline during intracellular acidification. This suggests that the com- bination of ouabain and endothelin results inagrad- ual increase in contractile amplitude with time that is relativelyunresponsive to further swings in pHi.

Endothelin, p111 and Protein Kinase C

Endothelin increases cytoplasmic pH in isolated adultrat cardiac myocytes via anamiloride-sensitive mechanism,presumably theNa+-H+exchangeratthe sarcolemmalmembrane,asdocumented above. Since activation of the Na+-H+ antiporter in a variety of cell types is knowntobe mediated by proteinkinase C, twoinhibitors ofthis enzyme, H-7 and the struc- turally unrelated compound sphingosine, were used to determine the involvement ofprotein kinase C in the sequence ofevents leading to intracellular alka- linization with endothelin. The increase inpHiafter exposure to 1 nM endothelin was completely inhib- ited by a 45-minute preincubation of myocytes in superfusionbuffer with 50 ,uM H-7 (n=4), as shown inFigure 5. Pretreatment of myocytes with 200 ,uM sphingosine for 10 minutes followed by 1 nM endo- thelin reduced the subsequent alkalinization of the cell interior to 0.053±0.006 pHunits (40.2% ofthe response ofcells treated with endothelin alone, Fig- ure 5). Incubation with this concentration ofsphin- gosine was restricted to 10 minutes; this time limit was setbecause our own experience has shown that toxic effects of this drug occur during incubations lasting longer than 20 minutes. The increase in contractile amplitude in H-7-pretreated and sphin- gosine-pretreated cells subsequently exposed to en- dothelin was 156+21% and 136±7% of control, respectively, approximately half of that seen with endothelin alone, as shown inFigure 5.

We3 have previously demonstrated that pertussis toxin completelyinhibits the increase incontractility observed with endothelin. To determine its effecton the intracellular alkalinization induced by endothe- lin, quiescent cellswere pretreated with 100 ng/ml pertussis toxin for 3 hours. This protocol had no effect on the contractile amplitude of cells paced subsequentlyat1.5 Hzandresulted in^an augmenta- tion of the contractile response ^to 83-agonists, as

expected. Pertussis toxin pretreatment also has been shown to reduce subsequent

[32P]ADP-ribosylation

of an endogenous G protein a-subunit by 98% in these cells.3 Pertussis toxin incubation blunted the

'..C_

E!

02

Co1- 0

control H-7 aphingosine ^PT

Sc a 0 0 .0- I

E .

CLX

£a.C.__

FIGURE5. Bargraph showing role ofprotein kinase Cand resistance topertussistoxin (PT) in the intracellularpH(pH,)

responseto endothelin.pHo,extracellularpH.Freshlyisolated adult ratventricularmyocytes were exposed to 1 nMendothe- lin, and contractility (open bars) andpH, (filled bars) were recorded at 8 minutes. Eight cells were preincubated with either50pMH-7 or 200pMsphingosinetodeterminewhether these inhibitorsofprotein kinase C would affect the alkalin- izationresponse toendothelin.H-7completelyblockedthe rise inpH,with 1 nMendothelin (*p<O.OS),althoughneither the declinein contractility with H-7 nor thechangesinpHiand contractility with sphingosine reached statistical significance aftercorrectionsformultiplecomparisons. Pertussistoxin (100 ng/ml for3 hours) had a nonsignificant effect onpHibut

reproducibly inhibited thecontractile response to endothelin (p<o.0l).

rise in pHi in response to 1 nM endothelin to 0.070±0.032 pH units (n=5), although this was not significantly different fromthe increase noted with- out pertussis toxin. However, as documented previ- Ously,3 theamplitude of contraction in response to 1 nMendothelin afterpertussis toxin pretreatment did not increase significantly overbaseline at 7 minutes

(119±20%, p=NS).

Discussion MeasurementofpH,in Cardiomyocytes

Since the introduction ofsecond-generation fluo- rescent dyes such as fura 2 and BCECF for the measurementofintracellular ionactivities, theaccu- ratecalibration of theirfluorescence signalhas been hamperedby severalmethodologicalproblems. With the in situ calibrationtechnique described above and in "Materials and Methods," the intracellular BCECFsignalcanbeconfidentlycalibrated overthe rangeofpHivalues achieved in this report, asdocu- mentedpreviously.19The valueobtained for baseline pHiinthisstudyfor adultratventricular cells super- fusedatpHO7.4(37°C) andstimulated at1.5Hzwas

7.09+±0.01, which is in accord with values recently reportedby Wallert and Frohlich23 and byEisneret a125 in these cells and in other cardiac tissue prepa- rations. TheslightlylowerbaselinepHireportedhere comparedwith values obtained by others in resting (quiescent) cells is likely due to the decline in

pHi

thataccompaniespacing, asdescribedbyBountraet al.26,27

275

The observations reported here were made on unloaded cells; that is, the cells underwentshorten- ing against an internal load that is presumed to be constant (isotonic). However, recentobservationsby LeeandAllen28 and Kenteta129 support theutilityof theisolated myocyte modelasrelevantin thestudyof cardiaccontractility. Finally,ourlaboratoryhasdem- onstratedthatinisolated ratmyocytes subjectedtoa staircasepacing protocol from 0.5-3.0Hzthe ampli- tude of contraction correlated closely with the in- crease in thevelocity of contraction (S.Borzak, J.D.

Marsh, personal communication, September 1990).

Myocyte Contractility andpH,

Reports from several laboratories,30-33 including ourown,3 have now confirmed the original observa- tions of Ishikawa et al,2 who determined that endo- thelin is a potent inotropic agent in cardiac muscle that appears to sensitize cardiac myofilaments to intracellular Ca'+. One potentialmechanism for this effect is an increase in intracellular pH. At a given level of intracellular Ca2+, in muscle not maximally contracted and in the presence of metabolic sub- strates, contractile amplitudeand developed tension increased upon alkalinization of the milieu surround- ing the myofilaments and fell in response to acido- sis.10 With the advent in the last decade of ion- selectivemicroelectrodes and fluorescentdyes, these observations have been confirmed and expand- ed.11,24-26,34-37

In the isolated rat ventricular myocyte model re- ported here, control cells underwent a significant increase in contractileamplitude with induction ofan intracellular alkalosis, although the subsequent intra- cellular acidosis didnotresult inafallbelow baseline in contractile amplitude in the absence of amiloride pretreatment (see Figures 2 and 3). However, when theintracellular acidosiswas notpreceded byasignif- icant alkalosis, as in thedata shown in Figure3Bfor amiloride-pretreated cells,alargeandsustained fall in contractileamplitudeoccurred.Thus, prioralkaliniza- tion of the cell blunts thenegativeinotropiceffect ofa subsequentintracellularacidosis, resultinginahyster- esisphenomenon intherelation of contractile ampli- tudeto pHi,as hasbeen notedpreviously in Purkinje fibers34 and singleisolatedratventricular myocytes.25 Interestingly,thehysteresisincontractile functionwas also apparent in myocytes pretreated with a concen- tration of ouabain(10-5 M) sufficientto cause onlya 10-15% rise in baseline contractile amplitude (data not shown). Presumably, this can be explained by a rise in intracellular Na+ activity that decreased the sarcolemmal Na+ gradient and resulted in both a slowed recovery from the intracellular acid load after NH4C1 washout and a reduced Ca2+ efflux (or in- creased Ca21 influx) via Na4-Ca2' exchange, thereby maintaining contractile amplitude despite the intra- cellular acidosis. Indeed, several reports have noted that thenegative inotropiceffect ofalow

pHi

canbe modifiedconsiderablyorreversedbyarise in cell Na+

and, consequently, Ca2+.242635 39

Endothelin, Myocyte Contractility, and Na+-H+Exchange

Endothelin has been shown to increase pHi in severaltissuesincluding mesangial cells13 andvascu- lar smooth muscle14by activating the plasmalemmal Na+-H+ exchanger. We didnotobserve the transient early acidification reported in mesangial cells by Simonsonetal,'3 whichwasonly observedat concen- trations of endothelin above 10 nM. Endothelin has also been shown to activate phospholipase C and phosphoinositide hydrolysis in a number oftissues, including cardiac cells, with subsequent activation of protein kinase C30,40-42; this activation pattern pre- sumably leads to phosphorylation of the Na+-H+

antiporter and results in the observed intracellular alkalinization.

Asshown inFigure 1, endothelin does increasepHi incardiacmyocytes,withahalf-maximalresponseof 0.08pHunitsat approximately 100 pM in this model.

Endothelin increased the rate of recovery of pHi from NH4C1 washout-induced acidosis and also caused a significant overshoot of pHi after NH4C1 washout,beyond the baseline increase in

pHi

induced by the peptide alone. The timecourseof theincrease in

pHi

afteraddition of endothelin to the superfusion media overlaps the time course of the contractile response to the peptide; althoughasmallincrease in pH could be detected within 1-2 minutes, no in- crease in contractility was apparent for about 4 minutes. Amiloride pretreatment with subsequent exposuretoendothelin prevented the riseinpHiand produced a concomitant 45% decrease in the maxi- mal contractile response compared with endothelin alone.

The concentration of amiloride used in these ex- periments (100 ,M) was chosen because it was approximately10-foldgreaterthanthereportedIC50 of the Na+-H+ exchanger in the sarcolemmal mem- braneof these cells at 140mMNa+23 and because it was sufficient to largely prevent the recovery ofpH, uponwashout ofNH4C1incontrol cells.Although the concentration of amiloride used here is somewhat lowerthanthatreportedto affect other sarcolemmal Na+ carriers, including Na4-Ca2' exchange, we can- not exclude other nonspecific effects of amiloride in these cells that might also have affected myocyte contractile amplitude. Nevertheless, it is likely that both the decline in pHi and the reduced inotropic response to endothelin in amiloride-pretreated cells wererelated, sincetwoinhibitors of protein kinase C, H-7 and sphingosine, that produced complete or partial inhibition of the endothelin-induced rise in

pHi

also caused a proportionate reduction in the inotropic effect of the peptide.

Despitetheevidencethatarise inpHicontributed to theinotropic action ofendothelin, there is aclear dissociation between changes in pHi and the pep- tide's inotropic effect. This dissociation is indicated by the 40-50% increase in contractile amplitude and the absence of an increase in pHi in amiloride-

Krameretal Endothelin-Induced Alkalosis inCardiocytes pretreated cells after exposure to 100 pMendothelin.

In addition, no change in pHi could be detected at concentrations ofendothelin below 10 pM, aconcen- tration that reproducibly causes a 10-25% increase in contractile amplitude in these cells at 8 minutes.

Also, a large increase in contractile amplitude in endothelin-treated cells was achieved with NH4C1 superfusion in the presence of amiloride(Figure 4B);

this increase in contractility was comparable with that seen in the absence of amiloride (Figure 3B) despite the lower maximal pHi with NH4C1 in amiloride-pretreatedcells. Thisimplies that although anintracellularalkalinization contributes to the ino- tropiceffect of endothelin, this effect is maximal after an increase in pH of approximately 0.20. Finally, alkalinizationcanoccurinthe absence of an increase incontractility, as occurred in pertussis toxin-treated cells (Figure 5). Thus, although it is neither the essential nor the sole modulator of the increase in contractile state, the rise in

pHi

with endothelin contributes to the maximal increase in contractile amplitude with the peptide and appears to be initi- atedbyapertussis toxin-resistant second-messenger pathway, while the increase in contractile amplitude ismediated viaa Goor GiGTP-binding protein.

Increased activity of the Na+-H+ exchanger could lead to conditions that would favor reduced

Ca2'

efflux orincreased influx by Na+-Ca2+ exchange, al- thoughagainwewereunable todetect anincrease in eithersystolicordiastolicCa21 concentration in myo- cytesexposedtoless than 1nMendothelin.3 Interest- ingly,afall inintracellular calcium transients has been described during NH4Cl-induced intracellular alkalo- sis, perhaps due to increased Ca2+ uptake by the sarcoplasmic reticulum and/or reduced Na+-H+ ex- change after introduction of the weak base.43 Thus, the failure of our previous work3 to document a

changeincalcium transientsaccompanyingthe inotro- pic effect of endothelin could have been due to the offsettingeffects ofasmallrise in intracellular calcium with endothelin andafall inCa2+ duetothe intracel- lularalkalinization inducedbythepeptide. However, thisexplanation isunlikely; unlike theintroduction of a weak base, the mechanism by which endothelin increases pHi involves stimulation of Na+-H+ ex- change, which, as noted above, should lead to a modest increase incytosolic Ca2+.

Endothelin'sRole in CardiacMyocyte CellBiology and Function

Intracellular alkalinization could also have other roles in normal myocyte physiology aside from the effectsoncontractilitydescribed here. Arise inpHiis clearlyassociated with amitogenicresponse inmany cells and with hypertrophy invascular smooth mus-

cle.42"4445 In addition, endothelin has been shown to have mitogenic effects in vascular smooth muscle cells46 and other cell types.47-50

Importantly,

the mitogenic effectsof endothelin and,^wheretheyhave beenmeasured, theeffects of endothelinonpH and ontheexpressionof certainproto-oncogenes, suchas

c-fos and c-junassociated with someforms ofhyper- trophic growth in cardiac myocytes,51 occur at rela- tively low concentrations (i.e., <1 nM), comparable withthe concentrations used here that induce maxi- mal increases in contractile amplitude and pHi. Ac- tivation ofphospholipase Cby endothelin has been well documented in both established cell lines and primary cultures of vascular smooth muscle cells,52 endothelial cells,49 andglial cells,53 although onlyat

higher concentrations, typically with an EC50 of 10 nMor higher.

Theresponse ofavariety of cell lines andprimary cultures pretreated with pertussis toxin and then exposed toendothelin alsoimplicates acomplexand perhaps tissue-specific signal transduction mecha- nism for this peptide, comprising several indepen- dently regulated pathways. In freshly isolated rat ventricular cells, asdocumented here, pertussis toxin completely inhibited theinotropic effect after expo- sure to endothelin, while variably reducing the con- comitantintracellular alkalinizationbynomorethan 50% (Figure 5). In contrast, pertussis toxin had no effect on the endothelin-induced rise in inositol trisphosphate

(EC5O,

10nM) infreshlydissociatedrat atrialcells.54 Adissociation between phospholipase C and phospholipase A2 activation in vascular smooth muscle cells hasbeen notedbased ontheirresponse to pertussis toxin,41 and a similar large increase in phospholipase A2 activiation with endothelin was seen in mesangial cells, but only at concentrations near 0.1 ,uM.50 This is well above the dose range necessaryforinositoltrisphosphate generation (1-10 nM) orfor the induction ofmitogenesis, intracellular alkalinization, and the stimulation of c-fos (<1 nM).

Thus, it is possible that other actionsofendothelin on ventricular myocytes might be demonstrable at concentrations greater than 1 nM. It isunknown what local tissue concentrationsof endothelin arerelevant.

Although the reported Kd for receptor binding of endothelin-1 to cardiac tissue is in the subnanomolar range,6-8 much higher concentrations could occur under pathological conditions. Nevertheless,the data reported here indicate that endothelin, in amounts well below 1 nM, will result in increased contractile function and an intracellularalkalinizationthat could facilitate the development ofmyocyte hypertrophy.

Acknowledgments

We acknowledgethe assistance and advice of Drs.

Steven Borzak and Martin Reers in thedevelopment of thecalibration method for BCECF in contractile cells. We also acknowledge the superb technical assistance of Mary O'Neill and Stephanie Murphy and the advice and support of Dr. HodaEid.

References

1. Yanagisawa M, Kurihara H, KimuraS,TomobeY,Kobayashi M, MitsuiY, YazakiY, Goto K, Masaki T: A novelpotent vasoconstrictorpeptide produced byvascular endothelial cells.

Nature1988;332:411-415

277