Journal of NeurochemistryLippincott—Raven Publishers, Philadelphia© 1998 International Society for Neurochemistry

ETA and ETB Specific Ligands Synergistically AntagonizeEndothelin-1 Binding to an Atypical Endothelin Receptor

in Primary Rat Astrocytes

Niels Jensen, Martin Hasselblatt, Anna-Leena Siren, *Lothar Schilling,tMartin Schmidt, and Hannelore Ehrenreich

Departments of Neurology and Psychiatry, Georg-August- University, and Max-Planck-Institute for Experimental Medicine,Gottingen; * Department of Neurosurgery, University Hospital, Mannheim; and t Main Laboratory,

BASF Aktiengesellschaft, Ludwigshafen, Germany

Abstract: Using a whole-cell binding procedure with longincubations at low temperature and subsequent acidstripping, we have characterized an atypical endothelin(ET) receptor in primary rat cortical astrocyte cultures.We found the following: (a) no competition for 1251-ET-1binding by the ETA antagonists BQ-123 and LU 135252or the ETA agonist IRL 1620; (b) weak competition bythe ETB antagonist BQ-788 and by the predominant ETBligand ET-3; (c) potentsynergistic competition of ETA andETB ligands in combination for 1251-ET-1 binding; (d) po-tent competition of FT-i with any of the radioligandsused, 1251-ET-1, 1251-IRL 1620, and [3H]BQ-123;(e) lackof competition of IRL 1620 and BQ-123 with the respec-tive other radioligand; (f) shifting of the amount of acid-strippable 1251-FT-1 binding from 20 to 80% by ETB Ii-gands and to 4% by ETA ligands; and (g) as a control,typical ETA and ETA binding characteristics of the RAT-ifibroblast and the U373MG astrocytoma cell line, respec-tively, under our assay conditions. The unusual bindingproperties of astrocytic ET receptors described in thisstudy appear to be the result of several binding sites inthe receptor for different ET ligands or ligand epitopes.Key Words: Primary astrocytes—Rat—Endothelin re-ceptor—ETA— ET~—Antagonists—Binding sites.J. Neurochem. 70, 473—482 (1998).

Endothelins (ETs), peptides of 21 amino acids, areamong the most potent vasoactive factors known. Theyexert their action via specific binding to at least twodifferent receptor subtypes, ETA and ETB, that belongto the class of heptahelical, G protein-coupled recep-tors (Arai et a!., 1990; Sakurai et a!., 1990; Ohisteinet al., 1996). ET-induced vasoconstriction is mediatedpredominantly by vascular ETA receptors, whereas ac-tivation of vascular ETB receptors leads to vasorelaxa-tion through release of nitric oxide or prostacyclin(Ohlstein et a!., 1996). Since the discovery of ETs in1988 (Yanagisawa et a!., 1988), these peptides have

been shown to be involved in various pathophysiologi-cal conditions. In fact, in cardiovascular medicine,clinical studies using ET antagonists are under way(Warner eta!., 1996). With respect tocerebrovasculardisorders, animal studies point to an efficiency of ETantagonists, e.g., in treating subarachnoid hemorrhage-associated vasospasm or in reducing damage causedby cerebral ischemia (Clozel and Watanabe, 1993). Inthese disorders, predominantly ET receptors located onblood vessels are therapeutically addressed. To reachsmooth muscle cells of cerebral blood vessels, how-ever, the receptor antagonists have to penetrate theblood—brain barrier. Therefore, in principle, the hugepopulation of astrocytes, glia! cells that comprise‘—40% of the cellular brain mass, is reached as well.Astrocytes are known to bind ETs specifically, as wellas to produce these peptides (MacCumber et a!., 1990;Ehrenreich et al., 1991b; Hösli and Hösli, 1991). Thefunction of astrocytic ET receptors, however, and con-sequences of their blockage are still unclear. The tasksattributed to astrocytic ET receptors range from auto-regulatory functions (Ehrenreich et a!., 1991a) to mito-genesis (Supattapone et a!., 1989; MacCumber et a!.,1990), ET eliminatory activity (Wu Wong et a!.,1995), and microcirculatory effects (Koyama andBaba, 1994). In primary cultures of rat astrocytes,mRNA for both the ETA and ETB receptor has beendetected (Ebrenreich et a!., 1993). In contrast, most

Received June 8, 1997; revised manuscript received August 21,1997; accepted September 11, 1997.

Address correspondence and reprint requests to Dr. H. Ehrenreichat Max-Planck-Institute for Experimental Medicine, Hermann-Rein-Sir. 3, D-37075 Gottingen, Germany.

Abbreviations used: BSA, bovine serum albumin; CNPase, 2,3-cyclic nucleotide-3-phosphodiesterase; DMEM, Dulbecco’s modi-fied Eagle’s medium; E-64, trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane; ET, endothelin; GFAP, glial fibrillary acidic pro-tein; PMSF, phenylmethylsulfonyl fluoride.

473

474 N. JENSEN ET AL.

binding studies point to the existence of only one popu-!ation of ET receptors on these cells (Ehrenreich eta!., 1991b; Hama et a!., 1992; Levin et al., 1992; WuWong et a!., 1995, 1996). This obvious discrepancy,as well as the need, in face of the future neurologicalapplications of ET antagonists, to understand astrocyticET binding better, originally prompted us to undertakethe present study. Binding of various different ET ago-fists and antagonists was characterized in cultures ofprimary rat cortical astrocytes. Surprisingly, we ob-tained evidence for the presence of an atypical ETreceptor on these ce!!s.

EXPERIMENTAL PROCEDURES

MaterialsMaterials and animals were obtained from the follow-

ing sources: rats (Wistar) from Charles River Laborato-ries [CRL: (WI) WU BR]; ‘251-Tyr’3-ET-! and ‘251-Tyr’3-IRL 1620 (2,200 Ci/mmol each) from Du Pont—NEN;[3H]Pro3’”°~-BQ-123(30 Ci/mmo!) from Amersham;U373MG human astrocytoma cell line from ATCC; RAT-Irat fibroblast cell line and LU 135252, gifts from Knoll AG,Germany; ET-!, ET-3, BQ-!23, BQ-788, and IRL 1620 fromAlexis; Dulbecco’s modified Eagle’s medium (DMEM),pheny!methylsulfonyl fluoride (PMSF), pepstatin A, leupep-tin, trans-epoxysuccinyl-L-leucylamido-(4-guanidino)bu-tane (E-64), bombesin, and mouse monoclonal antibodyagainst 2,3-cyclic nucleotide-3-phosphodiesterase (CNPase)from Sigma; penicil!in/streptomycin, mouse monoclonal an-tibody against glial fibrillary acidic protein (GFAP), rabbitpolyclona! antiserum against human von Wi!!ebrandt factor8, sheep polyclonal antiserum against mouse IgG (rhoda-mine-labeled), and mouse polyclona! antiserum against rab-bit IgG (fluorescein isothiocyanate-labeled) from Boeh-ringer Mannheim; mouse monoclona! antibody againstCD! lb from Serotec; and rabbit polyclonal antiserumagainst rat fibronectin from Kirkegaard and Perry (Gaithers-burg, MD, U.S.A.).

Cell culturePrimary astrocytes were prepared form cortices of 1-day-

old Wistar rats and grown to confluence under neuron-freeconditions in 24-well plates (2 cm2 per well) in glutamine-free DMEM supplemented with 10% heat-inactivated fetalcalf serum, 50 U/rn! penicillin, and 50 ~tg/ml streptomycinas described previously (Ehrenreich et al., l99lb) - The me-dium was changed twice weekly, and cells were used after2—3 weeks in culture. The RAT-! fibroblast cell line andtheU373MG astrocytoma cell line were grownto confluenceunder the same conditions as the astrocytes, but media weresupplemented with 2 mM glutamine.

Culture purityThe purity of primary astrocyte cultures was determined

by immunofluorescence using a mouse monoclona!antibodyagainst GFAP (1:50) as an astrocyte-specific marker, amouse monoclonal antibody against CD 1 lb (1:1,000) asa microglia-specific marker, a mouse monoclonal antibodyagainst CNPase (1:1,000) as an oligodendrocyte-specificmarker, a rabbit polyclona! antiserum against human vonWi!lebrandt factor 8 (1:50) as an endothelial cell-specificmarker, and a rabbit po!yclonal antiserum against rat fibro-

nectin (1:800) as a fibroblast-specific marker. A rhodamine-labeled sheep polyclonal antiserum against mouse IgG anda fluorescein isothiocyanate-labeled mouse polyclonal anti-serum against rabbit IgG were used as second antibodies.Confluent monolayers of astrocytes showed >97% positivestaining for GFAP. Oligodendrocytes accounted for <2%and microglial cells for <1% of the cells. Fibroblasts orendothelial cells were not detected in these cultures.

Radioligand bindingThe binding medium used for competition and associa-

tion/dissociation experiments wasDMEM without bicarbon-ate, 0.2% bovine serum albumin (BSA; wt/vol), 25 mMHEPES, adjusted to pH 7.4 with NaOH solution. The cellswere washed at room temperature three times with the bind-ing medium and were incubated, still adherent, with 0.5 mlper well of theprepared incubation solution composed of 50pM (competition) or 65—83 pM (association/dissociation)‘251-ET! or ‘251-IRL 1620, or of 2 nM [3H]BQ-!23, andthe unlabeled ligands. Each point was determined at least induplicate. After incubation (usually for 20 h at 4°C),thecells were washed three times with ice-cold binding mediumon ice. The acid stripping was done three times with 0.5 mlof an ice-cold solution of 0.2 M acetic acid and 0.5 M NaC!for 10 mm each on ice, and the portions were pooled forcounting. As alternative approaches to stripping radioligands(which favor their solubility and may permit dissociation ofreceptor-bound ligand without increasing nonspecific bind-ing), the acid-strip solution was replaced by three differentmodifications of a hyperosmotic solution (0.5 M NaC1, 20mMNaHCO

3/Na2CO3, pH 9.4): (a) Triton X-l00: 0.002%;(b) Triton X-lOO: 0.0005% (as nonionic detergent); and(c) triethanolamine: 5 mM (as hydrophilic weak ion-pairingreagent). The remaining cell layer was controlled for com-plete adherence under the microscope and solubilized with0.5 ml per well of a 10mM EDTA and 1% sodium dodecylsulfate (wt/vol) solution. The samples were counted with aWallac 1470 y counteror aWallac 1209 scintillation counterin 5 ml of scintillation fluid. The nonspecific binding in thepresence of I pM unlabeled ligands accounted for <1% ofthe total binding of ‘

251-ET-1 or ‘251-IRL 1620 and for <4%of the total binding of [3H] BQ-123. For incubations at 4°C,the plates were placed on ice 5 mm before the addition ofthe precooled incubation solutions and strictly kept in thecold until the end of the experiment. In the association anddissociation experiments, the dissociation was initiated byadding ET-1 or IRL 1620 in binding medium to give a finalconcentration of 0.2 or 5 ,uM, respectively. For inhibition ofreceptor internalization, cells were pretreated at the incuba-tion temperature for 5 mm with 40 or 80 ,uM phenylarsineoxide or with 0.45 M sucrose in binding medium (Garlandet al., 1994). Then internalization inhibitors were also pres-ent in theincubation solution. For preventing potential degra-dation of ligands, cells were incubated in the presence of apeptidase inhibitor cocktail consisting of 1 mM PMSF, 1~.tMpepstatin A, 100 p~Mleupeptin, and 100 ,uM E-64.

Data analysisThe data were analyzed using the computer programs

LIGAND 4.7 (Munson and Rodbard, 1980) and Excel withits macroprogram Solver (Microsoft). The binding constantsk_

1 and k1 were determined by fitting the data against thebinding model of a second-order association/first-order dis-sociation: d[RL]/dt = k1 [RI [LI + k_1 [RL] (where [RI

J. Neurochem., Vol. 70, No. 2, 1998

ATYPICAL ENDOTHELIN RECEPTOR IN RAT ASTROCYTES 475

is the concentration of receptor, [LI is the concentration ofligand, and [RL] is the concentration of the receptor—ligandcomplex). The binding constants K

0 were determined usingthe equation K0 = k_1/k1 and the dissociation half-timesusing the equation t112 = ln(2)/k_1. The specific bindingwas definedas the difference between total binding (no com-petitor) and nonspecific binding (in the presence of a 1 btMconcentrationof competitor). “Whole binding” was definedas the sum of acid-strippable and acid-resistant binding.Whenever ranges are given, they refer to minimal and maxi-mal data of all experiments.

RESULTS

Assay conditions/internalizationWhole-ce!! procedures using intact adherent cells

were preferred to membrane preparations to retainmore physiological conditions during the incubation.To reach equilibrium conditions of association and dis-sociation in the binding experiments, long incubationtimes had to be used. However, the probability of ii-gands being internalized into the living cell or of liganddegradation increases with longer incubations. There-fore, the acid-strip method has been applied toestimatethe amount of internalized radioactive ligand. After therespective incubation time, the intact cell layer waswashed with a hyperosmolar solution of pH 2.5 andhighionic strength. Underthese conditions, a!! radioac-tively labeled ligand that is extracellularly bound tothe receptor should be washed off (acid-strippable),and the remaining radioactivity (acid strip-resistant)is assumed to represent the internalized part (Resinket a!., 1990; Stojilkovic et a!., 1992; Takasuka et a!.,1992; Hildebrand et al., 1993; Marsault et a!., 1993;Garland et a!., 1994; Wu Wong et a!., 1995, 1996).As initial experiments resulted inhigh amounts of acidstrip-resistant radioactivity (“internalized”), we de-cided to incubate at 4°C,a temperature known to pre-vent vesicle-mediated internalization (Resink et al.,1990; Stojilkovic et a!., 1992; Takasuka et a!., 1992;

FIG. 1. Association and dissociation of1251ET1 and 1251-IRL 1620 on primary ratcortical astrocytes at 4°C.Adherent astro-cytes in 24-well plates were incubated fordifferent times with 83 pM 125l..~71(A) or65 pM 1251-IRL 1620 (B) at 4°C.To initiatedissociation, 0.2 ~sMET-1 after 3 h (A) or5 1sM IRL 1620 after 8 h (B)wasadded tosome wells. After washing, the cells weresubjected to an acid-strip procedure using0.5 M NaCl/0.2 M acetic acid solutionthree times for 10 mm each. The strippingsolutions were pooled (acid-strippablebinding), and the remaining intact celllayer was solubilized (acid strip-resistantbinding) and counted. Shown is the sum ofacid-strippable and acid-resistant binding(whole binding) from one representativeexperiment of four (1251-ET-1) or two (l25l~

IRL 1620) independent experiments ondifferent cell batches.

Hi!debrand et al., 1993; Marsau!t et a!., 1993; Garlandet a!., 1994; Wu Wong et a!., 1995).

Association and dissociationFigure 1 shows the association and dissociation of

‘25I-ET-l and 1251-IRL 1620, an ETB-specific agonist,obtained with living adherent astrocytes. For each timepoint, who!e binding was determined as the sum ofacid-strippable and acid-resistant binding. Equilibriawere reached after ‘—‘10 h for 1251-ET-1 and 16 h for125I-IRL 1620. The incubation period of the followingcompetition assays was chosen based on these timeestimates. k

1 (association rate constant) and k_1 (disso-ciation rate constant) were determined, and K0 wascalculated using the equation K0 = k~1/k1and resultedin KDETI = 1.5 x 10_2 h~/4.0 X iO~M’ h’ 4pM and KDIRL 1620 = 7.2 X 102 h’/1.7 X i0~M’

40 pM. The dissociation half times were t112 ETI

= 46 h and t112 ~ 1620 = 10 h. Despite an incubationat 4°Cto prevent internalization, the acid-strippablebinding accounted for only 21—25% (n = 4) of thewho!e ‘

251-ET-1 binding and 80—81% (n = 2) of thewhole ‘25I..IRI~1620 binding at all time points. Theacid-resistant part did not increase over time at theexpense of the strippable part as would have to beexpected for an internalization process, suggesting thatthe acid-resistant part may not represent internalizedradioligand. To control for a possible acid-induced in-crease in nonspecific 125I-ET-1 binding, additional ex-periments using 0.1

1uM ET- 1 in the acid-strip solutiondid not yield different results. In addition, using alter-native alkaline stripping solutions, the amount of strip-resistant ‘

251-ET- 1 binding remained high (83—93%)under all conditions tested.

Internalization inhibitorsFurther assays (20 h, 4°C) were performed using

hyperosmolar sucrose (0.45 M) in the incubation me-dium, which is known to inhibit endocytotic internal-

J. Neurochem., Vol. 70, No. 2, 1998

476 N. JENSEN ET AL.

ization (Garland et al., 1994). Addition of sucrose didnot remarkably alter strippable binding (range: 76—96% of controls without sucrose; n = 2) or wholebinding (range: 80—98% of controls without sucrose;n = 2) of different ET receptor ligands and !igandconcentrations. Phenylarsine oxide, another commoninternalization inhibitor acting through covalent modi-fication of proteins (Garland et al., 1994), at concen-trations of 40 and 80

1uM also failed to reduce the acid-resistant part.

DegradationTo test for a possible degradation of ligands during

the longincubation periods, a peptidase inhibitor cock-tail, made of PMSF, pepstatin A, leupeptin, and E-64,was added to the incubation mediumtogether with 1251..

ET- 1 and different concentrations of the competingligands. Neither whole binding (range: 93—111% ofcontrols; n = 2) nor percentage of acid-strippab!e bind-ing (range: 87—106% of controls) was altered in thepresence of this cocktail in incubations for 20 h at 4°Cor for 2 h at 37°C.

Final incubationOmitting BSA from the incubation medium lowers

the total binding of1251-ET- 1 to 7% of its binding in

the presenceof BSA. Therefore, all further incubationswere performed for 20 h at 4°Cin the presence of 0.2%BSA (wt/vol), without inhibitors of internalization ordegradation unless otherwise indicated.

‘251-ET-1 and 125I-IRL 1620 competitionFigure 2 shows competition curves for ‘25I-ET- 1 and

‘251..IRL 1620, a selective ETB !igand (Takai et a!.,

1992). Competing for 1251-ET- 1 binding, the K, forET-1 was 113 pM, and the curve was best fitted with asingle binding site (p < 0.0001, compared with a two-site model). In contrast, IRL 1620 and the specific

ETA antagonists BQ-l23 (Ihara et al., 1995) and LU135252 [compound (S)-6a in Riechers et a!., !996]did not appreciably compete (<11% inhibition at 1

1.tM), and ET-3, a predominant ETB ligand, displayedonly weak competition (47% inhibition at 1 ~.tM) (notall curves are shown). Competing for ‘

251-IRL 1620binding, the K, values of IRL 1620 and ET-! were1.84 nM and 400 pM, respectively. BQ-123 did notcompete with 1251-IRL 1620. For the homologous com-petition curves, the amount of acid-strippable bindingremained constant in a!! experiments over the wholeconcentration range [20—30% 125I-ET-1 (n = 5) and81—84% 1251-IRL 1620 (n = 3); curves not showni.Again, this finding makes it unlikely that the resistantpart represents internalization.

Acid-strippable/resistant ‘251-ET-1 bindingFigure 3 shows heterologous competition curves for

the ETA specific antagonists BQ-123 and LU 135252,the ETB specific agonists ET-3 and IRL 1620, and theETB specific antagonist BQ-788 (Ishikawa et al., 1994)with 1251-ET-1 as the radioligand. Although a substan-tial competition was not demonstrated under the condi-tions used, surprisingly, these ligands altered theamount of acid-strippable binding in a characteristicway: the ETA specific ligands BQ-123 and LU 135252decreased the acid-strippable binding from 20—30%(n = 14) to 4—8% (n = 7), rendering the bindingalmost completely acid-resistant (Fig. 3A). A compa-rable effect can be observed witha medium containingthe antibiotic sulfamethoxazol. An isomer of this sub-stance has been reported to be an ETA ligand (Steinet a!., 1994). In contrast, the ETB specific ligands ET-3, IRL 1620, and BQ-788 raised the percentage ofacid-strippable binding from 20—30% (n = 14) to 75—84% (n = 11), yielding the butterfly patterns of Fig.3B. Using alternative alkaline stripping solutions, no

FIG. 2. Competition of ET-1, IRL 1620,and BQ-123 for 1251-ET-1, 1251-IRL1620, or [3H]BQ-123binding to pri-mary rat cortical astrocytes. Adherentastrocytes in 24-well plates were incu-bated for 20 h at 4°Cwith 50 pM 125l

ET-1 (A), 50 pM ‘25l-IRL 1620(B), or 2nM [3H]BQ-123 (C) and the respectiveunlabeled compounds. The K, valuescalculated from the whole binding(sum of acid-strippable and acid-resis-tant binding) are as follows: 1251-ET-1:ET-1, 113 pM (n = 5); IRL 1620, NC*(n = 3); BQ-123, NC* (n = 5). 125l-IRL1620: ET-1, 400 pM (n = 2); IRL 1620,1.84 nM(n = 3); BQ-123, NC°(n= 3).[3H]BQ-123:ET-1, 415 pM (n = 3);IRL 1620, NC*(n = 2); BQ-123, 6.7 nM(n = 3). Shown are the mean valuesfor whole binding of one representativeexperiment of several independent ex-periments on different cell batches.°NC,no competition up to 1 ~M.

J. Neurochem., Vol. 70, No. 2, 1998

ATYPICAL ENDOTHELIN RECEPTOR IN RAT ASTROCYTES 477

FIG. 3. Competition of specific ETAand ETB ligands and of combinationsof ligands for 1251-ET-1 binding to pri-mary rat cortical astrocytes. Adherentastrocytes were incubated for 20 h at4°Cwith 50 pM l25l~~1and the re-spective ligands. Using the acid-stripmethod, the whole binding was dividedinto an acid-strippable and an acid-re-sistant part. A: Competition for 1251-ET-1 binding of the ETA specific ligands LU135252 (n = 3) and BQ-123 (n = 4).B: Competition for 1251-ET-1 binding ofthe ET

5 specific ligands ET-3 (n = 5),IRL 1620 (n = 3), and BQ-788(n = 3).C: Competition for i25l..~1 binding ofETB specific ligands in the presence ofETA specific ligands: ET-3 + 100 nMLU 135252 (n = 2), ET-3 + 100 nMBQ-123 (n = 2), IRL 1620 + 100 tiMBQ-123 (n = 2), and BQ-788 + 100nM BQ-123 (n = 2). Shown is one rep-resentative experiment of several inde-pendent experiments on different cellbatches. Continuous line, whole bind-ing; 0, acid-strippable binding; •,acid-resistant binding.

alteration of the amount of strippable ‘251-ET- 1 binding

was observed with any of the ET ligands.To get further insight into these effects, we tested

combinations of ETA and ETB ligands for their compe-tition with 1251-ET-1 binding to primary rat astrocytes.Figure 3C shows such competition experiments usingETB ligands in the presence of a 100 nMconcentrationof the ETA !igand BQ-123 or LU 135252. Althougheach class of substances alone is not able to competemeasurably with 1251-ET-1 binding, a combination ofthem turned out to be verypotent. Under these circum-stances, the acid-strippable binding almost disap-peared. This unexpectedly strong synergism betweenETA and ETB ligands was also seen when ETB ligandswere applied as the constant part in the presence of

varying amounts of ETA ligands (LU 135252 + 100nM IRL 1620 and BQ-123 + 100 nM BQ-788; datanot shown).

[3H] BQ-123 bindingIf ETA ligands display such a potent synergism with

ETA ligands, there should be [3H]BQ- 123 binding de-tectable. Indeed, specific [3H]BQ- 123 binding can bedemonstrated in primary rat astrocytes (Fig. 2C). TheK, values derived from competition experiments are asfollows: BQ-123, 6.7 nM; ET-l, 415 pM; LU 135252,2.1 nM; BQ-788, ~=100nM; and ET-3, ‘—‘1

1.tM (notall curves are shown). IRL 1620 up to 1 ~t/vfdid notshow competition. Under all conditions tested, theII

3HIBQ-l23 binding was 97—100% acid-strippable.

J. Neurochem., Vol. 70, No. 2, 1998

478 N. JENSEN ET AL

The Bm~i),for 3H J BQ- 123 in two independent experi-ments resulted in 80% and 83% of Bmax ET-i~ This,again, points to a receptor population to which all thetested radioligands, i.e., 125I-ET-1, 125I-IRL 1620, and[3H]BQ-123, bind.

ETA receptor?To exclude the possibility that ETA receptors would

display an abnormal binding behavior caused by theunusual assay conditions (4 h, 20°C), we tested theET binding characteristics of the rat fibroblast cell lineRAT-i, which is known to express high amounts ofETA receptors (Ambar and Sokolovsky, 1993). Thefibroblasts, however, show a typical ETA binding pat-tern under our incubation conditions (Fig. 4). ET-iand LU 135252, but not ET-3, are potent competitorsof both 125I-ET-1 and tI3HIBQ-123. In all conditionstested, the percentage of acid-strippable binding re-mained nearly constant at 65—74% (n = 2) of thewhole 1251-ET-1 binding. These data suggest that thepredominant ET binding site in astrocytes is not a typi-cal ETA receptor.

ETB receptor?To investigate whether typical ETA receptors would

account for the unexpected binding behavior of ETligands in astrocytes under the assay conditions used,we tested the human astrocytoma cell line U373MG,which has been shown to express a homogeneous ETBreceptor population (Wu Wong et a!., 1995). Indeed,a typical ETB binding profile using our assay conditions(4 h, 20°C) was observed in this cell line (Fig. 4).ET-i and ET-3 were potent competitors of ‘251-ET-iwitha K, of 30 pMand 240 pM, respectively. The acid-strippable part of the binding remained nearly constant(range: 37—44%; n = 2). LU 135252 showed weakcompetition, and BQ-123 at 1 btM did not competewith ‘25I-ET-i. [3HIBQ-l23 failed to exert specificbinding. These observations, in turn, question the pre-

dominant presence of typical ETA receptors in astro-cytes.

Bombesin/ET receptor?These unusual binding properties raise the question

of whether a new receptor type outside the classicalETA/ETB scheme is present in primary rat astrocytes.Bombesin in concentrations up to 1 p~Mfailed to com-pete for ‘251-ET-I or [3H]BQ-i23 binding and did notaffect the acid-strippable binding. This makes aninvolvement of the recently discovered bombesin/ETreceptor (Li et al., 1996) in astrocytes unlikely. Inaddition, bombesin did not compete with 1251-ET-i or[3HIBQ-123 binding to RAT-i or U373MG cells.Effects of time and temperature

The unexpected binding behavior of the astrocyticET receptors could be a result of the long incubationat low temperatures. To exclude this possibility, theincubation conditions were changed from 20 h at 4°Cto 2 h at 37°C(Fig. 5). The ‘251-ET-i binding becamemore acid-resistant [the strippable binding decreasedfrom 20—21% to 7—12% (n = 2; not shown)I, andthe K, for ET-i doubled from 113 pM to 208 pM withan unaltered Bmax. The competition by ET-3 improvedbut was still not complete at 1 ~tM(the IC

50 decreasedfrom ‘--‘10 1.tM to ‘--‘ iOO nM). Despite these alterations,the effect of ET-3 on the amount of acid-strippablebinding was still present: 100 nM ET-3 increased thestrippable binding from 7—12% to 34—40% (n = 2;not shown). Also, the synergism between ETA andETB ligands persisted: 100 nM LU 135252 decreasedthe IC50 of ET-3 from ‘--‘100 nM to ‘--‘3 nM, and 100nM BQ-123 decreased the IC50 of BQ-788 from ‘=2.5~.tMto ~-~0.4~.tM(not shown).

DISCUSSIONAcid strip-resistant binding is not internalized

The acid-strip method has commonly been appliedto ET receptors for the determination of ligand inter-

FIG. 4. Comparison of the E~bindingprofile of different cell types, expressingeither the classical ETB receptor(U373MG, human astrocytoma) or theclassical ETA receptor (RAT-i fibro-blasts) with the primary rat corticalastrocyte ET receptor. Shown is the per-centage of specific binding of 50 pM 125l

Er-i or of 2 nM [3H]BQ-1 23 in the pres-

ence of different unlabeled competitors.The lower parts of the columns in the1251ET1 experiments represent theamount of acid-strippable binding. °NB,no specific binding detectable.

J. Neurochem., Vol. 70, No. 2, 1998

ATYPICAL ENDOTHELIN RECEPTOR IN RAT ASTROCYTES 479

FIG. 5. Competition of representative ET receptor ligands for1251ET1 binding to primary rat cortical astrocytes. Adherent as-trocytes were incubated with 50 pM 1251-ET-i and increasingamounts of competitors using previous standard conditions(37°C,2 h). Shown is the sum of acid-strippable and acid-resis-tant binding. The K

1 values, as calculated from whole binding,are 208 pM for Er-i, =50 nM for ET-3, and =400 nM for BQ-788. BQ-i 23 up to i 1iM did not compete. Shown are representa-tive curves from two or three independent experiments.

nalization during incubation of whole cells (Re sink etal., 1990; Stojilkovic et al., 1992; Hildebrand et al.,1993; Marsault et al., 1993; Wu Wong et a!., 1995).A high percentage of acid strip-resistant

1251-ET-lbinding to human ETA and ETB transfected COS cellsafter incubation for 3 h at 4°Chas been reported earlier(Takasuka et a!., 1992). The present results suggestthat the acid strip-resistant binding to ET receptorsunder the incubation conditions used is not the resultof ligand internalization into primary rat cortical astro-cytes. This assumption is based on the following obser-vations: (a) Vesicle-mediated internalization can beeffectively blocked by incubations at 4°Cin previouslytested systems (Garland et a!., 1994). (b) In the pres-ent study, commonly accepted internalization inhibi-tors, hyperosmolar sucrose and phenylarsine oxide(Garland et al., i994), failed to alter the amount ofacid-strippable binding. (c) In association experimentsperformed at 4°C,the ratio of acid-strippable to wholebinding remains constant from the beginning (timepoint 1 h) to the equilibrium at 10—16 h. Comparableasymptotic acid-resistant 125I-ET-l binding kinetics onintact cells have been described by others (Marsaultet al., 1993; Desmarets et al., 1996). An internalizationprocess should result in higher amounts of internalizedradioligand if the incubation time is increased. (d) Inhomologous binding experiments at 4°C,the amountof acid-strippable binding is constant over the wholerange of unlabeled ligand concentrations used. Highdoses of competitor would be expected to compete for

an internalization mechanism, as well as for binding.(e) Under the assay conditions used, IRL i620, ET-3, LU 135252, or BQ-123, coincubated with ‘251-ET-1, profoundly alters the amount of acid-strippable bind-ing. However, there is no effective competition as ex-pected for an internalization process. (f) The bindingof the ETA antagonist [3H 1 BQ- 123 is entirely acid-strippable, and that of the ETA agonist 1251-IRL 1620is 81—84% strippable under all conditions tested. Incontrast, the nonselective agonist 125I-ET-l displayshighly variable (range: 4—82%) strippable binding,depending on the unlabeled ligands added. (g) Usingastrocytic membrane preparations, a binding patterncomparable to that described for intact cells can beobserved with respect to both acid stripping and thesynergistic competition of ET ligands (data notshown). Taken together, the acid-resistant part of 125J..

ET- 1 binding seems to represent different binding sitesor modi rather than an internalization process.

One-receptor hypothesisCan these binding sites or modi be located in one

receptor, or do they result from interactions of differentreceptors or receptor subtypes? The followingobserva-tions support a one-receptor concept: (a) The bindingdata for ‘251-ET-l are best fitted with the assumptionof one binding site. Moreover, earlier work from ourlaboratory showed a single site of ‘251-ET-3 binding inprimary rat astrocyte cultures upon incubation at roomtemperature for 2 h (Ehrenreich et al., 1991b). Similardata have been reported by others (Hama et a!., i992;Levin et al., 1992). (b) The percentage of acid strippa-ble 125I-ET-i binding varied over a large range from4% (with ETA specific ligands) over ‘--.25% (homolo-gous competition) to 84% (with ETB specific ligands).Additional receptor populations with a constantamount of acid-strippable binding, like the fibroblastETA or the astrocytoma ETB receptor, should restrictthis range if expressed in significant amounts. Thus,the predominant cortical astrocytic ET receptors seemto con ist of a homologous population and to havebinding sites or modi that differ in their acid sensitivity.

Which receptor subtype?The question remains which ET receptor subtype is

functionally expressed inour primary cortical astrocytecultures where mRNA of both ETA and ETB receptorshas been demonstrated by reverse transcription—poly-merase chain reaction (Ehrenreich et a!., i993). In thepresent study, typical ETB binding behavior, with ET-I and ET-3 being nearly equipotent, could be demon-strated using the human astrocytoma cell lineU373MG, which has been described earlier to expressa pure ETB receptor population (Wu Wong et al.,1995). In contrast, in our astrocyte preparations, ET-3has a K at least 240 timeshigher than ET-i. Moreover,[3HJBQ-i23 shows specific binding to astrocytes, butnot to U373MG cells. A failure of [3H]BQ-i23 tobind to ETB receptors is known from Girardi heart cells

J. Neurochem., Vol. 70, No. 2, 1998

480 N. JENSEN ET AL.

(Ihara et a!., 1995). This would point against typicalETB receptors being involved in cortical astrocyte cu!-tures.

Using a fibroblast cell line known to express highamounts of ETA receptors, typical ETA binding behav-ior (Ihara et a!., 1995; Wu Wong et a!., 1996) couldbe reproduced under the assay conditions used in thisstudy. A completely missing competition of specificETA antagonists for 125I-ET-1 binding, as found onour astrocytes, has never before been observed in anymammalian ETA receptor preparation. Therefore, it isunlikely that a “classical” ETA receptor is the predom-inant ET receptor in our astrocytes.

An amazing binding similarity is seen with ET re-ceptors found in Xenopus laevis (Kumar et al., 1994)and in the rainbow trout (Lodhi et a!., 1995), desig-nated ETAX and ETAF, respectively. In both prepara-tions, and in agreement with our results, ET-3 has a‘=300-fold lower affinity than ET-1 when competingfor 1251-ET-1 binding, and the highly specific and po-tent ETA and ETB ligands, BQ-123 and BQ-3020, re-spectively, and sarafotoxin S6c have very low affinitiesfor these receptors. The effects of other radioligands,ligand combinations, or acid stripping have not beentested on these receptors. Our present data could indi-cate the existence of a rat homologue of these receptorsthat is expressed in high concentrations in primarycortical astrocytes.

The binding of specific ETA and ETB (radio) ligandsor their synergism in competing for bound 125I-ET-ion the same type of ET receptor has not been describedpreviously. The remaining question of this study is thatof the identity of the predominant rat cortical astrocyticET receptor. If it is not a new ET receptor subtypecomparable to nonmammalian ET receptors, it couldbe either a posttranslationally or posttranscriptiona!!ymodified ETA or ETB receptor, possibly in the form ofan ETA/ETB hybrid receptor.

Different binding sites of the receptorThe predominant cortical astrocytic receptor shows

the following atypical behavior: (a) no competition ofETA or ETB specific ligands alone for 125I-ET-1 bind-ing, but in combination; (b) homologous competitionof each ligand, ‘25I-IRL 1620 or [3H]BQ-123, withitself, but not with the respective other ligand; and (c)potent ET- 1 competition, independent of the ETA orETB specific radioligand used. This can be explainedwith the simplified mode! shown in Fig. 6. SpecificETA and ETB ligands bind to either one of two differ-ent, nonoverlapping binding sites of the receptor. ET-1, however, is able to bind to both binding sites withsimilar affinity (the K~values described here for ET-1 at ‘25I-IRL 1620 and [3HIBQ-123 labeled receptorsare 400 pM and 415 pM, respectively). Even if onesite is blocked, ET-1 is still able to bind to the otherone. Therefore, only blockade of both sites can hinderET- 1 from binding to the receptor.

FIG. 6. Model ofthe astrocytic ET receptor with different bindingsites, explaining the atypical ET binding behavior of primary ratcortical astrocytes. We propose that 1251-ET-i is able to bind totwo different binding sites of the predominant cortical astrocyticEr receptor. The site on the left side of the figures is specificfor ETB ligands and acid strip-resistant with regard to 1251-ET-1binding. The other site on the right side of the figures is specificfor ErA ligands and sensitive to acid stripping. Only a combina-tion of specific ligands for both sites prevents the binding of125l~~ET~~iA mechanism fixing the bound 125l-ET-1 into a low-dissociation complex is indicated by a third binding site on thetop. A Acid-strippable 125l-ET-1 binding to the astrocytic Elreceptor with an ETB specific ligand bound simultaneously. B:Acid strip-resistant 1251-ET-1 binding to the astrocytic El recep-tor with an ETA specific ligand bound simultaneously.

The alteration of acid-strippable binding upon addi-tion of subtype-specific ligands can be explained ifone assumes the ETA ligand binding site to be acid-strippable and the ETB site to be acid strip-resistantwith regard to ‘25I-ET- 1 binding. As long as only ET-1 binds to the receptor, the .—~75%acid-resistant partis due to ET-1 binding to the ETB site. The ~25%acid-strippable part is caused by ET-1 binding to theETA site. A site-specific ligand, when added, is “forc-ing” the ET-1 molecule to bind to the respective othersite, thereby altering the strippable part of 125I-ET-1binding. Whereas IRL 1620, BQ-123, and LU 135252are very site-specific,ET-3 and, especially, BQ-788 are~tbleto bind to the other site at higher concentrationsas well, resulting in their weak 1251-ET- 1 competition.During a complex formation process, the bound ET- Iseems to be fixed by a second binding step into astable complex with low dissociation (Fig. 6). Thecompetition process apparently takes place before thiscomplex is formed, as indicated by a more efficientinhibition at shorter incubations, and by a 30—40 timeshigher K, in competition experiments than K

0 fromassociation and dissociation experiments (113 pM ver-sus 4 pM for ET-1 and 1.84 nM versus 40 pM forIRL 1620).

The model presented in Fig. 6 shows some similari-ties to models developed by others using chimericETA/ETB receptors (Sakamoto et a!., 1993; Becker etal., 1994). They also postulated separate binding siteson the receptor for different parts of the ET molecule.Similar models have been developed for many hepta-

J. Neurochem., Vol. 70, No. 2, 1998

ATYPICAL ENDOTHELIN RECEPTOR IN RAT ASTROCYTES 481

helical peptide receptors, e.g., for opioid receptors (Ta-kemori and Portoghese, 1992). The simultaneous bind-ing of different, and even natural, ligands to one recep-tor is well established on ligand-gated ion channels andhas also been reported for many heptahelical receptors(Kostenis and Mohr, 1996).

In our mode!, a combination of ETA and ETB antago-nists would be expected to synergistically antagonizethe functional response of astrocytic ET receptors toET- 1 stimulation. Indeed, we found that a combinationof ETA and ETB antagonists synergistically increasesthe amount of immunoreactive ET- 1 in the supernatantof primary rat cortical astrocytes, cells known to pro-duce ET-l (Ehrenreich et a!., 1991a), whereas eachof the antagonists by itself is ineffective (data to bepublished elsewhere). The atypical binding behaviorof the astrocytic ET receptor described in the presentarticle may explain this effect. Taken together, thesefindings in cultured rat astrocytes may contribute toour understanding of the cerebra! ET system. As withall rat studies, their relevance for human astrocytesremains to be confirmed.

Acknowledgment: We thank Knoll AG (Ludwigshafen,Germany) for providing LU 135252. Heike Kamrowski-Kruck’s help with cell culture preparations is gratefully ac-knowledged. This work has been supported by the GermanResearch Foundation (Eh 133/1-2,3). A.-L.S. is supportedby the Humboldt Foundation.

REFERENCES

Ambar I. and Sokolovsky M. (1993)Endothelin receptors stimulateboth phospholipase C and phospholipase D activities in differentcell lines. Eur. .1. Pharmacol. 245, 31—41.

Arai H., Hod S., Aramori I., Ohkubo H., and Nakanishi 5. (1990)Cloning and expression of a eDNA encoding an endothelinreceptor [see comments]. Nature 348, 730—732.

Becker A., Haendler B., Hechler U., and Schleuning W. D. (1994)Mutational analysis of human endothelin receptors ETA andET5. Identification of regions involved in the selectivity forendothelin 3 or cyclo-(D-Trp-D-Asp-Pro-o-VaI-Leu). Eur. .1.Biochem. 221, 951—958.

Clozel M. and Watanabe H. (1993) BQ-123, a peptidic endothelinETA receptor antagonist, prevents the early cerebral vasospasmfollowing subarachnoid hemorrhage after intracisternal but notintravenous injection. Life Sci. 52, 825—834.

Desmarets J., Gresser 0., Guedin D., and Frelin C. (1996) Interac-tion of endothelin-1 with cloned bovine ETA receptors: bio-chemical parameters and functional consequences. Biochemistry35, 14868—14875.

Ehrenreich H., Anderson R. W., Ogino Y., Rieckmann P., Costa T.,Wood G. P., Coligan J. E., Kehrl J. H., andFauci A. S. (199Ia)Selective autoregulation ofendothelins in primary astrocyte cul-tures: endothelin receptor-mediated potentiation of endothelin-I secretion. New Biol. 3, 135—141.

Ehrenreich H., Kehrl J. H., Anderson R. W., Rieckmann P., VitkovicL., Coligan J. E., andFauci A. S. (199lb) A vasoactive peptide,endothelin-3, is produced by and specifically binds to primaryastrocytes. Brain Res. 538, 54—58.

Ehrenreich H., Costa T., Clouse K. A., Pluta R. M., Ogino Y., Coli-gan J. E., and Burd P. R. (1993) Thrombin is a regulator ofastrocytic endothelin-1. Brain Res. 600, 201—207.

Garland A. M., Grady E. F., Payan D. G., Vigna S. R., and Bunnett

N. W. (1994) Agonist-induced internalization of the substanceP (NK1) receptor expressed in epithelial cells. Biochem. J. 303,177—186.

Hama H., Sakurai T., Kasuya Y., Fujiki M., Masaki T., and GotoK. (1992) Action of endothelin-l on rat astrocytes through theETB receptor. Biochem. Biophys. Res. Commun. 186, 355—362.

Hildebrand P., Mrozinski J. E. Jr., Mantey S. A., Patto R. J., andJensen R. T. (1993) Pancreatic acini possess endothelin recep-tors whose internalization is regulated by PLC-activatingagents.Am. J. Physiol. 264, G984—G993.

Hösli E. and Hösli L. (1991) Autoradiographic evidence for endo-thelin receptors on astrocytes in cultures of rat cerebellum,brainstem and spinal cord. Neurosci. Lett. 129, 55—58.

Ihara M., Yamanaka R., Ohwaki K., Ozaki S., Fukami T., IshikawaK., Towers P., and Yano M. (1995) [

3H]BQ-123,a highlyspecific andreversible radioligand for the endothelinETA recep-tor subtype. Eur. J. Pharmacol. 274, 1—6.

Ishikawa K., Ihara M., Noguchi K., Mase T., Mino N., Saeki T.,Fukuroda T., Fukami T., Ozaki S., Nagase T., Nishikibe M.,and Mitsuo Y. (1994)Biochemical and pharmacologicalprofileof a potent and selective endothelin B-receptor antagonist, BQ-788. Proc. Nati. Acad. Sci. USA 91, 4892—4896.

Kostenis E. and Mohr K. (1996) Two-pointkinetic experiments toquantify allosteric effects on radioligand dissociation. TrendsPharmacol. Sci. 17, 280—283.

Koyama Y. andBaba A. (1994)Endothelins are extracellular signalsmodulating cytoskeletal actin organization in rat cultured astro-cytes. Neuroscience 61, 1007—1016.

Kumar C., Mwangi V., Nuthulaganti P., Wu H. L., Pullen M., BrunK., Aiyar H., Morris R. A., Naughton R., and Nambi P. (1994)Cloning and characterization of a novel endothelin receptorfrom Xenopus heart. J. Biol. Chem. 269, 13414—13420.

LevinE. R., Frank H. J. L., andPedram A. (1992)Endothelin recep-tors on cultured fetal rat diencephalic glia. J. Neurochem. 58,659—666.

Li Y., Kumar C., and Rosen C. A. (1996) Human endothelin-bombesin receptor. PCI Int. AppI. WO 96/11946, PCT/US94/11843.

Lodhi KM., Sakaguchi H., Hirose S., and Hagiwara H. (1995)Localization and characterization of a novel receptor for endo-thelin in the gills of the rainbow trout. J. Biochem. (Tokyo)118, 376—379.

MacCumber M. W., Ross C. A., and Snyder S. II. (1990)Endothelinin brain: receptors, mitogenesis, and biosynthesis in glial cells.Proc. Nati. Acad. Sci. USA 87, 2359—2363.

Marsault R., Feolde E., and Frelin C. (1993) Receptor externaliza-tion determines sustained contractile responses to endothelin- 1in the rat aorta. Am. J. Physiol. 264, C687—C693.

Munsoti P. J. and Rodbard D. (1980) LIGAND: a versatile compu-terized approachfor characterization of ligand-binding systems.Anal. Biochem. 107, 220—239.

Ohlstein E. H., Elliott J. D., Feuerstein G. Z., and Ruffolo R. R.Jr. (1996) Endothelin receptors: receptor classification, novelreceptor antagonists, and potential therapeutic targets. Med. Res.Rev. 16, 365—390.

Resink T. J., Scott Burden T., Boulanger C., WeberE., and BuhlerF. R. (1990) Internalization of endothelin by cultured humanvascular smooth muscle cells: characterization and physiologi-cal significance. Mol. Pharmacol. 38, 244—252.

Riechers H., Albrecht H. P., Amberg W., Baumann E., Bernard H.,Bohm H. J., Klinge D., Kling A., Muller S., Raschack M., UngerL., Walker N., and Wernet W. (1996) Discovery and optimiza-tion of a novel class of orally active nonpeptidic endothelin-Areceptor antagonists. J. Med. Chem. 39, 2123—2128.

Sakamoto A., Yanagisawa M., Sawamura T., Enoki T., Ohtani I.,Sakurai T., Nakao K., Toyo-oka I., and Masaki 1. (1993)Distinct subdomains of human endothelin receptors determinetheir selectivity to endothelinA-selective antagonist and endo-thelinB-selective agonists. J. Biol. Chem. 268, 8547—8553.

Sakurai I., Yanagisawa M., Takuwa Y., Miyazaki H., Kimura S.,

J. Neurochem., Vol. 70, No. 2, 1998

482 N. JENSEN ET AL.

Goto K., and Masaki T. (1990) Cloning of a eDNA encodinga non—isopeptide-selective subtype of the endothelin receptor[see comments]. Nature 348, 732—735.

Stein P. D., Hunt J. T., Floyd D. M., Moreland S., Dickinson K. E.,Mitchell C., Liu E. C., Webb M. L., Murugesan N., Dickey J.,McMullen D., Zhang R., Lee V. G., Serafino R., Delaney C.,Schaeffer T. R., and Kozlowski M. (1994) The discovery ofsulfonamide endothelin antagonists and the development of theorally active ETA antagonist 5-(dimethylamino)-N-(3,4-di-methyl-5-isoxazolyl)- l-naphthalenesulfonaniide. J. Med. Chem.37, 329—331.

Stojilkovic S. S., Balla T., Fukuda S., Cesnjaj M., Merelli F., Krsma-novic L. Z., and Catt K. J. (1992) Endothelin ETA receptorsmediate the signaling and secretory actions of endothelins inpituitary gonadotrophs. Endocrinology 130, 465—474.

Supattapone S., Simpson A. W., and Ashley C. C. (1989) Free cal-cium rise and mitogenesis in glial cells caused by endothelin.Biochem. Biophys. Res. Commun. 165, 1115—1122.

Takai M., Umemura I., Yamasaki K., Watakabe T., Fujitani Y., OdaK., Urade Y., Inui T., Yamamura T., and Okada T. (1992)A potent and specific agonist, Sue- [G1u9,Ala”1 5] endothelin-1(8—21), IRL 1620, for the ETB receptor. Biochem. Biophys.Res. Commun. 184, 953—959.

Takasuka T., Adachi M., Miyamoto C., Furuichi Y., and WatanabeT. (1992) Characterization of endothelin receptors ETA andETB expressed in COS cells. J. Biochem. (Tokyo) 112, 396—400.

Takemori A. E. and Portoghese P. 5. (1992) Selective naltrexone-derived opioid receptor antagonists. Annu. Rev. Pharmacol.Toxicol. 32, 239—269.

Warner T. D., Elliott J. D., and Ohlstein E. H. (1996) Californiadreamin’ ‘bout endothelin: emerging new therapeutics. TrendsPharmacol. Sci. 17, 177—181.

Wu Wong J. R., Chiou W. J., Magnuson S. R., and Opgenorth T. J.(1995) Endothelin receptor in human astrocytoma U373MGcells: binding, dissociation, receptor internalization. J. Pharma-col. Exp. Ther. 274, 499—507.

Wu Wong J. R., Chiou W., Magnuson S. R., Bianchi B. R., andLin C. W. (1996) Human astrocytoma U138MG cells expresspredominantly type-A endothelin receptor. Biochim. Biophys.Acta 1311, 155—163.

Yanagisawa M., Kurihara H., Kimura S., Tomobe Y., KobayashiM., Mitsui Y., Yazaki Y., Gob K., and Masaki T. (1988)A novel potent vasoconstrictor peptide produced by vascularendothelial cells. Nature 332, 411—415.

J. Neurochein., Vol. 70, No. 2, 1998