1922

Serotonin-Induced Coronary Spasmin a Swine Model

A Minor Role of DefectiveEndothelium-Derived Relaxing Factor

Tohru Fukai, MD; Kensuke Egashira, MD, PhD; Hiroshi Hata, MD;Koutaro Numaguchi, MD; Yuichi Ohara, MD; Teisuke Takahashi, BS;

Hitonobu Tomoike, MD, PhD; Akira Takeshita, MD, PhD

Background. Coronary spasm may be caused by endothelial dysfunction, vascular smooth musclehyperreactivity, or both. We aimed to determine the relative role of endothelial dysfunction and vascularsmooth muscle hyperreactivity in the pathogenesis of coronary artery spasm in the swine model in vivo.Methods and Results. In G6ttingen miniature pigs given a high cholesterol diet, a segment of the left

coronary artery was denuded and irradiated with x-ray (total, 30 Gy). Three months after endothelialdenudation and irradiation, vasomotor responses of the denuded and control sites to agonists wereassessed by quantitative arteriography. Serotonin (10 jag/kg) provoked coronary spasm at the denudedsite (diameter reduction, 79±6%) associated with ST elevation but not at the nondenuded control site(21±6%). Intracoronary infusion of N0-nitro-L-arginine methyl ester (LNNA, an inhibitor of endotheli-um-derived nitric oxide) of 1 and 3 mg/kg potentiated constriction evoked with serotonin (1, 3, 10 pgg/kg)at the control site but did not alter it at the denuded site. However, serotonin-induced constriction afterLNNA was still less at the control site (31±3%) than at the denuded site (80±5%). Endothelium-dependent vasodilation with substance P (0.1, 1, 10 ng/kg), which was inhibited by LNNA, was less(P<.01) at the denuded site than at the control site, whereas vasodilation with the nitrovasodilator SIN-i(0.1, 1, 10 ng/kg) was comparable between the two sites. Histological study revealed regeneratedendothelial cells and intimal thickening at the denuded site.

Conclusions. The results suggest that the denuded segment of the coronary artery with regeneratedendothelium was associated with defective endothelium-dependent vasodilation mediated by nitric oxideand vascular smooth muscle hyperreactivity to serotonin. However, provocation of coronary spasm withserotonin resulted primarily from vascular smooth muscle hyperreactivity but not by defective nitric oxideproduction in this swine model. (Circultion. 1993;88[part 1]:1922-1930.)KEY WORDs * atherosclerosis * endothelium * nitric oxide * substance P

C oronary spasm plays an important role in thepathogenesis of ischemic heart disease.1,2 How-ever, the mechanisms of coronary spasm have

not been well understood. We have created a swinemodel of coronary spasm in which angiographicallydocumentable coronary spasm associated with ST seg-ment elevation could be provoked by histamine orserotonin 3 months after endothelial denudation withhigh cholesterol feeding with or without selective irra-diation.3-5 Using this animal model of coronary spasm,we have demonstrated that coronary spasm occurred atthe site of intimal thickening after denudation in the invivo condition as well as in the excised heart,3-7 suggest-ing that coronary spasm was related to local atheroscle-

Received February 8, 1993; revision accepted June 8, 1993.From the Research Institute of Angiocardiography and Cardio-

vascular Clinic, Kyushu University School of Medicine, Fukuoka,Japan.

Correspondence to Dr Egashira, The Research Institute ofAngiocardiology and Cardiovascular Clinic, Kyushu UniversitySchool of Medicine, 3-1-1, Maidashi, Higashi-ku, Fukuoka, 812Japan.

rotic process but not to neurohormonal mechanisms.We also showed that coronary spasm could be pre-vented by selective blockers of H1- or S2-receptors ofvascular smooth muscle7-9 and that smooth muscle re-activity to histamine or serotonin of excised coronaryartery from the spastic segment was markedly aug-mented compared with that of the artery from thenonspastic segment.8,10 These results suggest that sero-tonin- or histamine-induced coronary spasm in thisswine model results from augmented contraction ofvascular smooth muscle induced by altered receptor-mediated mechanisms.However, the in vitro experiments also revealed that

endothelium-dependent vasorelaxation evoked with se-rotonin and histamine was attenuated in excised coro-nary artery from the spastic segment compared withthat in the artery from the nonspastic segment.10 Suchfindings suggest the possibility that in addition to vas-cular smooth muscle hyperreactivity to agonists, defec-tive endothelium-dependent vasorelaxation may con-tribute to coronary spasm. It has been implicated thatdefective endothelium-dependent vasodilation may play

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

Fukai et al Serotonin-Induced Coronary Spasm in Pigs

an important role in coronary spasm in patients withvariant angina, since acetylcholine provokes coronaryspasm in those patients.1"

This study aimed to determine the role of vascularsmooth muscle hyperreactivity and defective endothelialfunction to serotonin-induced coronary spasm in theswine model. The relative roles of those two mechanismsin coronary spasm must be assessed in the in vivo condi-tion, since coronary spasm is defined angiographically as alocal hyperconstriction of the coronary artery producingmyocardial ischemia.12,12 We examined the decreases inthe coronary artery diameter evoked with serotonin be-fore and after N@-nitro-L-arginine methyl ester (LNNA),which inhibits nitric oxide biosynthesis13-15 at the denudedsite and irradiated segment (the spastic site) and at thenondenuded segment (the nonspastic site). We also exam-ined the increases in the coronary artery diameter evokedwith substance P, an endothelium-dependent vasodilator,before and after LNNA at the two sites.

MethodsAnimal Preparation

Thirteen male Gottingen miniature pigs (Nihon CreaInc, Tokyo) (4 to 6 months old, 15 to 21 kg in weight)were housed individually under controlled room tem-perature and were fed minipig chow containing 2%cholesterol (Nihon Crea Inc, Tokyo). The denudationof the left coronary artery was undertaken as follows.Pigs were sedated with intramuscular ketamine hydro-chloride (12.5 mg/kg) and were then anesthetized withan intravenous administration of sodium pentobarbital(20 mg/kg). Animals were then intubated and ventilatedvia a positive-pressure respirator (Shinano Inc, Tokyo).Arterial pH, Po2 and Pco2 were kept within normalranges. Using sterile surgery, an 8F green Kifa catheter(Kifa, Stockholm) was inserted from the carotid arteryinto the orifice of the left coronary artery. After anintravenous injection of 3000 units of heparin, themidportion of the left anterior descending coronaryartery was denuded using a balloon catheter as de-scribed.3 Coronary arteriograms were recorded beforeand after the denudation procedure to determine thedenuded site, which revealed mild constriction immedi-ately after denudation. Pigs were then brought back tothe houses and kept for 3 months. To produce advancedatherosclerotic lesions, x-ray was irradiated selectivelyto the denuded area twice at 4 and 6 weeks after theendothelial denudation as described.4,5 X-ray was colli-mated to the area of 5 cm in diameter, and the totaldose of irradiation was 30 Gy. Animals were cared foraccording to the guidelines of American PhysiologicalSociety.

Coronary Arteriography andHemodynamic Measurements

Electrocardiograms in leads I, II, III, V,, and V6 wererecorded. Arterial pressure was measured with a pres-sure transducer (Gould Inc, Cleveland, Ohio) con-nected to the Kifa catheter. Arterial pressure, heartrate, and ECGs were continuously monitored and re-corded on a pen recorder (NEC San-Ei PolygraphSystem, Tokyo).

Selective coronary arteriography was performed in a

alization of the denuded site, using a Toshiba cineangi-ography system (KXO-1250/CAS-CA, Toshiba MedicalInc, Tokyo). Angiograms were recorded on 35-mmcinefilm (Varicath I, Hirata Sangyo Inc, Osaka, Japan)at 48 frames per second. The angle of the projection, theposture of the animal, and the distance from the x-rayfocus to the object and that from the object to the imageintensifier were carefully kept constant during eachexperiment.

Cineangiograms were projected on a screen using acineprojector (ELK-35CB, Nishimoto Sangyo Inc, Osa-ka, Japan), and an end-diastolic frame was selected.The luminal diameters of the denuded and nondenudedsites were measured quantitatively by a caliber as wepreviously described.36 The denuded site was at the leftanterior descending coronary artery as opposed to thecontrol site, which was at the left circumflex artery.Coronary spasm was provoked at the denuded site (thespastic site) by serotonin of 10 ,ug/kg. A segment of thenondenuded site where the baseline diameter was sim-ilar to that at the spastic site was taken as the controlsite.

Experimental ProtocolThree months after endothelial denudation, pigs

were anesthetized with sodium pentobarbital as de-scribed in the Animal Preparation section. Heparin(3000-unit bolus) was administered intravenously every60 minutes. After indomethacin (3 mg/kg) was admin-istered, pigs were allowed to stabilize for 30 minutes,and the following three protocols were done in 13 pigs.

Protocol 1. In 7 pigs, serotonin at the graded doses of1, 3, and 10 ,ug/kg was administered into the leftcoronary artery. Coronary arteriography was performed3 minutes after each dose of serotonin. We waited for atleast 10 minutes before injecting the next dose ofserotonin.

Thirty minutes after the serotonin study, PGF2a at thedose of 10 jig/kg was infused into the left coronaryartery, and coronary arteriography was performed 10minutes after infusion of PGF2,. In the preliminarystudies, we measured the coronary artery diameterevery 5 minutes after intracoronary infusion of PGF2,(10 ,ug/kg) and confirmed that the drug caused asustained decrease in the diameter; the degree ofvasoconstriction was stable from 10 to 40 minutes afterPGF2,. Fifteen minutes after PGF2,,, substance P (anendothelium-dependent vasodilator) at the gradeddoses of 0.1, 1, and 10 ng/kg was infused into the leftcoronary artery. Coronary arteriography was performed3 minutes after each dose of substance P. We waited for10 minutes before the next dose of substance P wasinfused.

Sixty minutes after the completion of experimentswith substance P, LNNA, an inhibitor of nitric oxidebiosynthesis, was infused into the left coronary artery ata dose of 1 mg/kg in 10 minutes. Thereafter, the studieswith serotonin, PGF2a, and substance P were repeated.After LNNA, the order of the drug study was random-ized. It was confirmed in the preliminary study that theeffect of LNNA lasted for at least 1 hour.

Protocol 2. In another 6 pigs, serotonin at the doses of1, 3, and 10 ,ug/kg was infused into the left coronaryartery, and coronary arteriography was performed as

left anterior oblique projection that allowed clear visu-

1923

described in protocol 1. After completion of the exper-

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

1924 Circulation Vol 88, No 4, Part 1 October 1993

iments with serotonin, we waited for 20 minutes, andthen PGF2, (10 pg/kg) was infused. Fifteen minutesafter PGFW2, SIN-i (3-morpholino-syndnonimine), anendothelium-independent nitrovasodilator, was infusedinto the left coronary artery at the graded doses of 0.1,1, and 10 jig/kg. Coronary arteriography was performedbefore and 4 minutes after each dose of SIN-i. Then,LNNA at a dose of 1 mg/kg was infused into the leftcoronary artery, and the experiments with SIN-1 at thegraded doses were repeated.

Protocol 3. This protocol was done in the 6 pigs thatwere used in the study of protocol 2. A week after thestudy of protocol 2, serotonin at the graded doses wasadministered into the left coronary artery before andafter 3 mg/kg ofLNNA as described in protocol 1. Afterthe pretreatment with PGF2a (10 pug/kg), the effects ofsubstance P at the graded doses of 1, 3, and 10 pg/kgwere also studied before and after LNNA at the highdose (3 mg/kg), as described in protocol 1. Since LNNAat the high dose increased arterial pressure by morethan 25 mm Hg, the a-blocker phentolamine (1 mg/kgby bolus followed by infusion of 1 mg/min) was simul-taneously infused intravenously during the study with 3mg/kg of LNNA in order to prevent the systemicpressor effect of 3 mg/kg of LNNA.Each dose of the drug was diluted with 1 mL of

physiological saline, and the same amount of saline wasused to flush the catheter. During the study, the cathe-ter position was fixed, and an equal amount of drugsolution was injected into the left coronary artery.

Histological StudyPigs were killed with a lethal dose of sodium pento-

barbital and exsanguination, and then the heart wasexcised. The left coronary artery was perfused with 2%buffered glutaraldehyde solution at the pressure of 90mm Hg. The spastic and nonspastic segments of the leftcoronary artery were resected and fixed with glutaral-dehyde solution for a few days. Tissue samples used forlight microscopic examination were embedded in paraf-fin, sectioned at 5-1£m thickness, mounted on glassslides, and stained with hematoxylin-eosin, using vanGieson methods. Samples used for scanning electronmicroscopic examination were dried, longitudinally bi-sected, mounted on alminum studs, coated with gold,and examined using a Hitachi H-430 scanning electronmicroscope.

DrugsSerotonin, LNNA, substance P (Sigma Chemical Co,

St. Louis, Mo), SIN-i (Takeda Pharmaceutical Co,Osaka, Japan), and PGF2, (Sigma Chemical Co) wereused. All drugs were diluted with physiological salineimmediately before use.

Statistical AnalysisData are expressed as mean±SE. Student's t tests

were used for comparisons of paired or unpaired data.When serial changes in the coronary artery diameter orhemodynamic variables in response to drugs were com-pared, ANOVA for repeated measures followed byBonferroni's multiple comparison test was used. Com-parisons of hemodynamic parameters and coronaryvasomotion before and after administration of LNNA

value of less than .05 was considered statisticallysignificant.

ResultsBody Weight and the Coronary Artery Diameter inBaseline ConditionBody weight increased from 23±1 kg (ranging from

17 to 25 kg) before endothelial denudation to 36±1 kg(ranging from 28 to 40 kg) (P<.01) 3 months afterendothelial denudation.The baseline diameters of the denuded and nonde-

nuded control sites were 2.1±0.3 and 2.2±0.1 mm

before and 2.6±0.1 and 2.7±0.2 mm (n=13) 3 monthsafter endothelial denudation, respectively. No organicstenosis was seen angiographically at the denuded site.

Effects ofLNNA on Vasoconstriction Evoked WithSerotonin and PGF2GChanges in hemodynamic variables with serotonin

and PGF2, are summarized in Table 1. Mean arterialpressure decreased transiently after each dose of sero-tonin, but it returned to the baseline level by 3 minutesafter drug administration when coronary arteriographywas performed. Heart rate did not significantly changeafter each dose of serotonin. Arterial pressure and heartrate did not change after PGF2,,.

Fig 1 shows angiograms before and after serotonin ina pig. Changes in the luminal diameter evoked withserotonin and PGF2,, before and after LNNA are sum-marized in Table 1 and Fig 2. Before LNNA, thebaseline diameter of the denuded site did not differfrom that in the control site. Serotonin decreased thediameter at the denuded and control sites in a dose-dependent manner. The degrees of serotonin-inducedvasoconstriction at the denuded site were significantlygreater than those at the control site (P<.01 byANOVA). Significant ischemic ST segment elevation(0.2 mV or more) was noted after serotonin at the doseof 10 jig/kg in all pigs. The degree of PGFw,-inducedvasoconstriction at the denuded and control sites (16±2and 15±2%, respectively) did not differ.An infusion of LNNA at the dose of 1 mg/kg in-

creased arterial pressure (P<.05) but did not signifi-cantly change heart rate and the diameter of the de-nuded and control sites (Table 1). The decreases in thediameter evoked with serotonin at the control site wereenhanced after LNNA (P<.01 by ANOVA), whereasthose at the denuded site were not altered by LNNA (1mg/kg) (Fig 2). LNNA (1 mg/kg) did not alter thePGF2,-induced decrease in the diameter at the denudedand control sites (Table 1).The effects of LNNA at the high dose (3 mg/kg) on

vasoconstriction evoked with serotonin and PGF2a aresummarized in Table 2. LNNA (3 mg/kg) increasedmean arterial pressure (P<.01), which was decreased byan intravenous infusion of phentolamine to the levelcomparable to that before LNNA (3 mg/kg). Heart ratedid not significantly change before and after LNNA (3mg/kg). Phentolamine did not change the diameter atthe denuded and control sites. LNNA of 3 mg/kgsignificantly enhanced the serotonin-induced decreasesin the diameter at the control site, as did LNNA of 1mg/kg. However, the decreases in the diameter evoked

were done using two-way ANOVA. The probability with serotonin at the graded doses after LNNA of 3

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

Fukai et al Serotonin-Induced Coronary Spasm in Pigs

TABLE 1. Protocol 1: Vasoconstriction Evoked With Serotonin and PGF2U Before andAfter 1 mg/kg LNNA

Coronary Diameter, mmHeart Rate, mAop,

Denuded Site Control Site bpm mm HgBefore 1 mg/kg LNNA

Control 2.5±0.1 2.7±0.2 160+11 145±8Serotonin

1 pg/kg 1.9±0.1t 2.6±0.2 167±11 149±73 ,tglkg 1.5±0.1t 2.5±0.2 167±11 148±610,ug/kg 0.5±0.1t 2.2±0.2 169±7 142±7

PGF2a 101ig/kg 2.1±0.2 2.3±0.2 162±10 148±6After 1 mg/kg LNNAControl 2.4±0.1 2.5±0.1 155±11 154±7*Serotonin

1 gg/kg 1.8±0.1t 2.1±0.2* 156±11 154±63 gg/kg 1.4±0.1 t 1.9±0.2* 156±11 152±610 ,g/kg 0.5±0.1t 1.7±0.2* 156±11 149±5*

PGF2a 10Jg/kg 2.0±0.2 2.2±0.2 150±12 155±6*

mAop indicates mean arterial pressure; bpm, beats per minute.Data are mean±SE; n=7. *P<.05 vs before ANOVA; tP<.01 vs control site; fP<.01 vs before LNNA by ANOVA

of repeated measures followed by Bonferroni's multiple comparison test.

mg/kg did not differ from those after LNNA of 1 mg/kgat the denuded and control sites. The PGF2a-induceddecrease in the diameter at the denuded and controlsites did not differ before and after LNNA (3 mg/kg).

Effects ofLNNA on Vasodilation Evoked WithSubstance P and SIN-I

Changes in hemodynamic variables during infusion ofsubstance P are summarized in Table 3. Arterial pres-sure and heart rate did not significantly change aftereach dose of substance P. The magnitudes of theincrease in the diameter evoked with substance P at thegraded doses were significantly larger in the controlthan the denuded site (P<.05 by ANOVA).LNNA at the doses of 1 and 3 mg/kg attenuated

vasodilation evoked with substance P at the control sitebut not at the denuded site (P<.01) (Fig 3 and Table 3).The high dose of LNNA (3 mg/kg) did not furtherattenuate vasodilation evoked with substance P at thecontrol site as compared with the results with LNNA of1 mg/kg (Table 3).The effects ofLNNA (1 mg/kg) on the SIN-i-induced

increases in the diameter are summarized in Table 4.SIN-1 at the graded doses increased the diameter at thedenuded and control sites in a dose-dependent manner.The increases in the diameter evoked with SIN-i didnot differ between the denuded and control sites andbefore and after LNNA (1 mg/kg) at each site.

Histological StudyExamination by light microscope revealed thickening

of the intima at the denuded site, which was coveredwith the regenerated endothelium. Intimal thickeningwas not seen at the control site (Fig 4).

Scanning electron microscopy disclosed an intactlining of endothelial cells of a fusiform shape at the

control site. The abnormal appearances of the regener-ated multishaped endothelial cells of varying sizes werenoted at the denuded site (Fig 4).

DiscussionThe major findings of this study are fourfold. First, an

intracoronary administration of serotonin provokedcoronary spasm associated with ischemic ST segmentelevation at the denuded site 3 months after endothelialdenudation and x-ray irradiation. Second, endothelium-dependent vasodilation was markedly attenuated at thedenuded site as compared with that at the control site.Third, vasoconstricting response of vascular smoothmuscle to serotonin was markedly augmented at thedenuded site as compared with that at the control site.Fourth, and most important, attenuated endothelium-dependent vasodilation at the denuded site was unlikelyto account for serotonin-induced coronary vasospasm.Thus, coronary vasospasm in this swine model mayresult primarily from augmented vasoconstricting re-sponse of vascular smooth muscle to serotonin.

Defective Endothelium-Dependent Vasodilation at theDenuded Site

In this study, endothelium-dependent vasodilation ofthe coronary artery was examined in two ways. First, weexamined serotonin-induced vasoconstriction beforeand after LNNA at the denuded and control sites. It hasbeen suggested that endothelium-dependent vasodila-tion is largely mediated by nitric oxide, which is synthe-sized in the endothelium from L-arginine,11,12 and thatits synthesis is competitively inhibited by L-arginineanalogues such as LNNA.13-15 It also has been shownthat serotonin has dual effects on vascular smoothmuscle: that is, direct constriction and indirect endothe-lium-dependent relaxation of vascular smooth mus-

1925

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

1926 Circulation Vol 88, No 4, Part 1 October 1993

Control Serotonin 10pg/kg

rL.

FIG 1. Angiograms before (left) and after (right) serotonin in a pig. An intracoronary administration of serotonin (10 pg/kg)

provoked coronary spasm (augmented decrease in the diameter at the denuded site, as indicated by arrows) associated with

ischemic ST changes. Asterisks indicate the segment used as the control site.

cle.16,17 The findings in this study are consistent with thelatter suggestion, since LNNA augmented serotonin-induced coronary vasoconstriction at the control site bypreventing endothelium-dependent (nitric oxide-medi-

b..

0

0E'5

-a

5E._

0:5

Ur

0

100 r

75 F

- D site (LNNA-)0 D site (LNNA+)

C site (LNNA-)C site (LNNA+)

] NS

25 r

o0 10

Serotonin (pg/kg)

FIG 2. Plot shows serotonin-induced vasoconstriction beforeand afterLNNA (1 mg/kg). Thepercentage ofdecreases in thediameter induced by serotonin at the graded doses was

significantly greater after than before LNNA at the control (C)site but did not differ before and after LNNA at the denuded(D) site. **Indicates P<.01 by ANOVA.

ated) vasodilation (Tables 1 and 2 and Fig 2). However,in contrast to the findings at the control site, LNNA didnot alter serotonin-induced vasoconstriction at the de-nuded site. Second, we examined coronary vasodilationevoked with substance P after pretreatment with PGF2,before and after LNNA. PGF2, caused mild and com-parable vasoconstriction at the denuded and controlsites before and after LNNA (Tables 1 and 2). It isshown that substance P induces endothelium-depen-dent vasodilation and has no direct effect on vascularsmooth muscle.18-21 Before LNNA, substance P causedsignificantly greater vasodilation at the control site thanat the denuded site (Fig 3). Furthermore, LNNA signif-icantly attenuated substance P-induced vasodilation atthe control site but did not alter it at the denuded site.Thus, coronary vasodilation mediated by nitric oxideevoked with two agonists, serotonin and substance P,was more markedly attenuated at the denuded site thanat the control site. It appears that vascular smoothmuscle response to nitric oxide did not differ betweenthe denuded and control sites, since SIN-i-inducedvasodilation was comparable between the two sites(Table 4). These findings suggest that endothelium-dependent vasodilation via nitric oxide in response toserotonin and substance P was defective at the previ-ously denuded and spastic sites.We confirmed histologically that the endothelium was

regenerated 3 months after denudation, although the

v ..... 7: ......... . :.:, i..B. : :s 2 ., ..... (. ivii>sLi i S L.ti tSf Ad.....< . . ....... :: 7 ' iv J - - ^ * . ##a 7 - Ad9 i''4" '^"'' 'wg ; 'any

-o f >A .; # ::: '.:.__.+. Fo ___. : s I,<:. A;v . i.| ' 3 st ' . . 44 .' . i.'{t . .^

#,+ Si. -S i i: :4x, ''t y . , i* X fi' MP.ie' i i,S ::: :. .:a -.50 t

1

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

Fukai et al Serotonin-Induced Coronary Spasm in Pigs 1927

TABLE 2. Protocol 3: Vasoconstriction Evoked With Serotonin and PGF2, Before andAfter 3 mg/kg LNNA

Coronary Diameter, mmHeart Rate, mAop,

Denuded Site Control Site bpm mm HgBefore 3 mg/kg LNNAControl 2.7±0.1 2.7±0.1 157±10 142±8Serotonin

1 jig/kg 2.1±0.1* 2.6±0.1 157±11 140±53/hg/kg 1.5±0.1 * 2.4±0.1 157±8 140±510 1£g/kg 0.7±0.1 * 2.3±0.1 153±7 135±5

PGF2oa 10/hg/kg 2.2±0.2 2.1±0.2 152±10 145±8After 3 mg/kg LNNA

Control

Before phentolamine 2.6±0.1 2.6±0.1 141±8 160±10tAfter phentolamine 2.6±0.2 2.6±0.2 152±8 142±10

Serotonin1 /g/kg 2.1 ±0.1* 2.4±0.1t 157±5 140±63/hg/kg 1.6±0.2* 2.1 ±0.1t 158±8 137±610 jig/kg 0.8±0.2* 2.0±0.1t 162±7 135±5

PGF2. 10/hg/kg 2.2±0.1 2.3±0.1 150±8 147±6

mAop indicates mean arterial pressure; bpm, beats per rData are mean±SE; n=6. *P<.01 vs control site; tP<.

followed by Bonferroni's multiple comparison test.

regenerated endothelial cells varied in shape and size incontrast to the uniform fusiformed cells at the controlsite. Thus, our findings indicate that endothelium-de-pendent vasodilation of the denuded site was defectiveeven though the endothelium was regenerated. Thesefindings in vivo are consistent with those of the in vitrostudy.'0,2223 Yamamoto et al10 reported that endotheli-um-dependent relaxation to serotonin in the denudedsegment with regenerated endothelial cells was im-paired in the swine. Shimokawa and Vanhoutte22,23reported that even in the swine without cholesterol diet,the endothelium-dependent relaxations to serotoninand aggregating platelets were impaired from 4 to 24weeks after endothelium removal in the previouslydenuded arteries with regenerated endothelium.

It is interesting to note that substance P causedcoronary vasodilation after LNNA at the denuded andcontrol site (Table 3 and Fig 3). Since the increase inthe dose of LNNA did not further attenuate substanceP-induced coronary vasodilation (Table 3), it was un-likely that inhibition of nitric oxide synthesis by LNNAwas inadequate. We consider the possibility that sub-stance P-induced coronary vasodilation after LNNAwas mediated by endothelium-dependent mechanismsother than release of nitric oxide, since substance P hasno direct vasorelaxing effect on vascular smooth mus-cle.18-21 It has been suggested that substance P mayevoke vasorelaxation by facilitating release of endothe-lium-dependent hyperpolarizing factor in pig coronaryartery.24 Our findings indicate that substance P-inducedcoronary vasodilation after LNNA was comparablebetween the denuded and control sites (Fig 3), suggest-ing the possibility that the release of endothelium-

LNNA by ANOVA of repeated measures

dependent hyperpolarizing factor by substance P maynot be attenuated at the denuded site, although synthe-sis of nitric oxide is defective. However, this possibilityneeds to be examined in future studies.

Mechanisms of Coronary Spasm in the Swine ModelThe major aim of this study was to determine the

relative roles of endothelial dysfunction and vascularsmooth muscle hyperreactivity in serotonin-induced cor-onary spasm in the swine model. Vascular smooth musclereactivity to serotonin was assessed by examining coro-nary vasoconstriction evoked with serotonin afterLNNA. Serotonin-induced coronary vasoconstriction af-ter LNNA was significantly greater (P<.01) at the de-nuded site than at the control site (Fig 2), indicating thatvascular smooth muscle at the denuded site was hyper-reactive to serotonin 3 months after denudation. Theresults shown in Fig 2 suggest that the contribution ofdefective nitric oxide production to serotonin-inducedhypercontraction at the denuded site may be a functionof the serotonin dose. More important, after LNNA,serotonin at the highest dose (10 jig/kg) caused thecoronary diameter reduction of only 31+±1% at thecontrol site and did not cause myocardial ischemia,whereas serotonin-induced vasoconstriction at the de-nuded site was approximately 80% before and afterLNNA and was associated with myocardial ischemia (Fig2). These findings strongly suggest that inhibition of nitricoxide synthesis was not enough to cause coronary vaso-spasm and that serotonin-induced coronary vasospasm inthis swine model resulted primarily from hypercontrac-tion of vascular smooth muscle to serotonin.

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

1928 Circulation Vol 88, No 4, Part 1 October 1993

TABLE 3. Protocols 1 and 3: Vasodllation Evoked With Substance P Before and AfterLNNA at Doses of 1 mg/kg and 3 mg/kg

Coronary Diameter, mmHeart Rate, mAop,

Denuded Site Control Site bpm mm HgBefore 1 mg/kg LNNA

Control§ 2.1 ±0.2 2.3±0.2 162±10 148±6

Substance P

0.1 ng/kg intracoronary 2.1 ±0.1t 2.7±0.1 161±10 148±5

1 ng/kg intracoronary 2.3±0.1 t 2.8±0.1 166±10 148±6

10 ng/kg intracoronary 2.4±0.1* 3.0±0.1 172±7 146±6

After 1 mg/kg LNNA

Control§ 2.0±0.2 2.2±0.2 150±12 155±6*Substance P

0.1 ng/kg intracoronary 2.0±0.2 2.3±0.1* 153±10 153±51 ng/kg intracoronary 2.2±0.2 2.4±0.1* 153±11 152±510 ng/kg intracoronary 2.3±0.1 2.5±0.1* 153±10 149±4

Before 3 mg/kg LNNAControl§ 2.2±0.2 2.1 ±0.2 152±10 145±8

Substance P

0.1 ng/kg intracoronary 2.2±0.2t 2.6±0.1 153±6 142±111 ng/kg intracoronary 2.3±0.1 t 2.8±0.1 153±6 142±1110 ng/kg intracoronary 2.4±0.1* 2.9±0.1 155±7 140±8

After 3 mg/kg LNNA

Control§ 2.2±0.1 2.3±0.1 150±8 147±6Substance P

0.1 ng/kg intracoronary 2.2±0.1 2.4±0.1* 157±8 142±61 ng/kg intracoronary 2.3±0.1 2.5±0.1* 157±5 142±810 ng/kg intracoronary 2.4±0.2 2.6±0.2* 162±9 137±3

mAop indicates mean arterial pressure; bpm, beats per minute.Data are mean±SE; n=7. *P<.05, tP<.01 vs control site by ANOVA of repeated measurements followed by

Bonferroni's multiple comparison test; *P<.01 vs before LNNA by ANOVA of repeated measurements followed byBonferroni's comparison test. §Diameters, heart rate, and mAop after PGF2a.

These findings in pigs are consistent with those inpatients with vasospastic angina.2526 Recent clinical stud-ies by Egashira et aP25 and Yamamoto et al6 havedemonstrated that endothelium-dependent vasodilationevoked with substance P was preserved at the spastic sitewhere coronary spasm was provoked by acetylcholine orergonovine, suggesting preservation of endothelial func-tion at the spastic coronary artery segments. They sug-gested that acetylcholine- or ergonovine-induced coro-nary vasospasm in patients with variant angina resultedprimarily from hyperreactivity of vascular smooth muscleto acetylcholine or ergonovine but not from local endo-thelial dysfunction.2526 It should be noted that endothe-lium-dependent vasodilation evoked with substance Pwas markedly attenuated at the denuded and spastic sitein our swine model, whereas it was preserved at thespastic site in patients with variant angina.2526 Thedifference in the results between patients and the swinemodel might have resulted from the fact that the pigs hadundergone endothelial denudation.We cannot exclude the possibility that other endothe-

lium-related mechanisms such as endothelium-depen-

I-

ES

E

-0

03aSSc,c

40

30

20

10

o

* D site (LNNA-)0 D site (LNNA+)- C site (LNNA-)C- C site (LNNA+)

**

1NS

0.1 10

Substance P (nglkg)FIG 3. Plot shows substance P-induced vasodilation beforeand afterLNNA (1 mg/kg). Thepercentage ofincreases in thediameter induced by substance P at the graded doses wassignificantly greater after than beforeLNNA at the control (C)site but did not differ before and after LNNA at the denuded(D) site. **Indicates P<.01 byANOVA.

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

Fukai et al Serotonin-Induced Coronary Spasm in Pigs

TABLE 4. Protocol 2: Vasodliation Evoked With SIN-1 Before and After LNNACoronary Diameter, mm

Heart Rate, mAop,Denuded Site Control Site bpm mm Hg

Before 1 mg/kg LNNA

Control* 2.0+0.3 2.1 ±0.3 163+7 150+7

SIN-10.1 ng/kg 2.3±0.2 2.4±0.1 164±8 148±8

1 ng/kg 2.4±0.1 2.5±0.2 168±8 147±8

10 ng/kg 2.6±0.2 2.7±0.2 168±7 145±7

After 1 mg/kg LNNA

Control* 2.0±0.5 2.1±0.4 160±8 153±5

SIN-i0.1 ng/kg 2.2±0.2 2.2±0.1 168±8 155±7

1 ng/kg 2.3±0.1 2.4±0.2 167±8 154±7

10 ng/kg 2.6±0.2 2.6±0.2 169±7 151±7

mAop indicates mean arterial pressure; bpm, beats per minute.Data are mean±SE; n=7. There were no significant differences between control and denuded sites as assessed

by ANOVA. *Diameters, heart rate, and mAop after PGF2a.

dent hyperpolarizing factors,27-29 vasoconstricting fac-tors such as endothelin,30 and radical species31 mighthave contributed to the provocation of coronary spasm

Light Microgral

by serotonin. Future studies are needed to clarifycontributions of such mechanisms. We excluded theeffects of prostacyclin or other cyclooxygenase products

ph Scanning ElectronMicrograph

Control Site

Denuded Site

FIG 4. Histological images at the denuded and control sites. Examination by light microscope revealed thickening of the intimaat the denuded site (bar represents 600 gm), which was covered with the regenerated endothelium. Intimal thickening was not seenat the control site (bar represents 400 gtm). Scanning electron microscopy disclosed an intact lining of endothelial cells of afusiform shape at the control site (bar represents 50 gum). The abnormal appearances of the regenerated multishaped endothelialcells of varying sizes were noted at the denuded site (bar represents 50 gm), as shown by Nagasawa et al.4

1929

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

1930 Circulation Vol 88, No 4, Part 1 October 1993

on coronary vasospasm because experiments were per-formed in the presence of indomethacin.

ConclusionsThe results demonstrated that endothelium-depen-

dent vasodilation mediated by nitric oxide was attenu-ated at the denuded and irradiated segments of coro-nary artery where coronary spasm was induced byserotonin. However, attenuated endothelium-depen-dent vasodilation via nitric oxide was unlikely to accountfor coronary spasm, since serotonin-induced vasocon-striction after LNNA at the control site was markedlyless than that at the denuded and spastic sites. It isreasonable to consider that serotonin-induced coronaryspasm in this swine model resulted primarily fromhyperreactivity of vascular smooth muscle to serotoninat the denuded site. We do not exclude the possibility,however, that several endothelium-derived factors dif-ferent from nitric oxide may be involved in the patho-genesis of coronary spasm.

AcknowledgmentsThis study was supported by scientific research grant No.

02404045; a grant-in-aid for scientific research on priorityareas No. 03268226 from the Ministry of Education, Scienceand Culture, Tokyo; a grant for scientific research from theNaito Foundation, Tokyo; Japan Heart Foundation researchgrant for 1992, Tokyo; and a grant from the Japanese Foun-dation of Cardiovascular Research, Tokyo. The authors aregrateful to Hiromitsu Kasuya, BS; Makoto Sakata. BS; andChikako Towatari for their technical assistance.

References1. Maseri A, Severi S, L'abbate A, Chierchia S, Marzilli M, BallestraAM, Parodi 0, Biagini A, Distante A. "Variant" angina: oneaspect of a continuous spectrum of vasospastic myocardialischemia: pathogenetic mechanisms, estimated incidence andclinical and coronary arteriographic findings in 138 patients. Am JCardiol. 1978;42:1019-1035.

2. Maseri A, Davies G, Hackett D, Kaski JC. Coronary artery spasmand vasoconstriction: the case for a distinction. Circulation. 1990;81:1983-1991.

3. Shimokawa H, Tomoike H, Nabeyama S, Yamamoto H, Araki H,Nakamura M, Ishii Y, Tanaka K. Coronary artery spasm inducedin atherosclerotic miniature swine. Science. 1983;221:560-562.

4. Nagasawa K, Tomoike H, Hayashi Y, Yamada A, Yamamoto T,Nakamura M. Intramural hemorrhage and endothelial changes inatherosclerotic coronary artery after repetitive episodes of spasmin x-ray-irradiated hypercholesterolemic pigs. Circ Res. 1989;65:272-282.

5. Kuga T, Tagawa H, Tomoike H, Mitsuoka W, Egashira S, Ohara Y,Takeshita A, Nakamura M. Role of coronary artery spasm inprogression of organic coronary stenosis and acute myocardialinfarction in a swine model: importance of mode of onset andduration of coronary artery spasm. Circulation. 1993;87:573-582.

6. Egashira K, Tomoike H, Yamamoto Y, Yamada A, Hayashi Y,Nakamura M. Histamine-induced coronary spasm in regions ofintimal thickening in miniature pigs: roles of serum cholesterol andspontaneous or induced intimal thickening. Circulation. 1986;74:826-837.

7. Yamamoto Y, Tomoike H, Egashira K, Kobayashi r, Kawasaki T,Nakamura M. Pathogenesis of coronary artery spasm in miniatureswine with regional intimal thickening after balloon denudation.Circ Res. 1987;60:113-121.

8. Ohara Y. Kuga T, Hirakawa Y. Hata H. Tomoike H, Nakamura M.Alkalosis augments serotonin-induced luminal constriction at thex-ray irradiated vessel in hypercholesterolemic miniature swine.Circulation. 1990;82(suppl III):III-767. Abstract.

9. Satoh S, Tomoike H, Mitsuoka W, Egashira S, Tagawa H, Kuga T,Nakamura M. Smooth muscles from spastic artery segments showhypercontractility to histamine. Am J Physiol. 1990;259:H9-H13.

10. Yamamoto Y, Tomoike H, Egashira K, Nakamura M. Attenuationof endothelium-related relaxation and enhanced responsiveness ofvascular smooth muscle to histamine in spastic coronary arterialsegments from miniature pigs. Circ Res. 1987;61:772-778.

11. Yasue H, Horio Y, Nakamura N, Fujii H, Imoto N, Sonoda R,Kugiyama K, Obata K, Morikami Y, Kimura T. Induction ofcoronary artery spasm by acetylcholine in patients with variantangina: possible role of the parasympathetic nervous system in thepathogenesis of coronary artery spasm. Circulation. 1986;74:955-963.

12. Oliva P, Potts D, Pluss R. Coronary artery spasm in Prinzmetalangina: documentation by coronary arteriography. N Engi J Med.1973;288:745-751.

13. Rees DD, Palmer RMJ, Hodson HF, Moncada S. A specificinhibitor of nitric oxide formation from L-arginine attenuates en-dothelium-dependent relaxation. BrJPharmacol. 1989;96:418-424.

14. Ishii K, Chang B, Kerwin JF, Huang ZJ, Murad F. N@-nitro-L-arginine: a potent inhibitor of endothelium-derived relaxing factorformation. Ear J Pharmacol. 1990;176:219-223.

15. Moore PK, AL-Swayeh OA, Chong NWS, Evans RA, Gibson A.L-NG-nitro-arginine (L-NOARG), a novel, L-arginine-reversibleinhibitor of endothelium-dependent vasodilation in vitro. Br JPharmacol. 1990;99:408-412.

16. Woodman OL, Dusting GJ. N-nitro-L-arginine causes coronaryvasoconstriction and inhibits endothelium-dependent vasodila-tation in anesthetized greyhounds. Br J Pharmacol. 1991;103:1407-1410.

17. Cappelli-Bigazzi M, Nuno DW, Lamping KG. Evidence of a rolefor compounds derived from arginine in coioniary response toserotonin in vivo. Am J Physiol. 1991;261:H404-H409.

18. Bossaller C, Habib GB, Yamamoto H, Williams C, Wells S, HenryPD. Impaired muscarinic endothelium-dependent relaxation andcyclic guanosine 5'-monophosphate formation in atherosclerotichuman coronary artery and rabbit aorta. J Clin Invest. 1987;79:170-174.

19. Forstermann U, Mugge A, Alheid U, Haverich A, Frohlich JC.Selective attenuation of endothelium-mediated vasodilation in ath-erosclerotic human coronaay arteries. Circ Res. 1988;62:185-190.

20. Toda N, Okamura T. Endothelium-dependent and -independentresponses to vasoactive substances of isolated human coronaryarteries. Am J Physiol. 1989;257:988-995.

21. Chester AH, Oneil GS, Moncada S. radjkarimi S. Yacoub MH.Low basal and stimulated release of nitric oxide in atheroscleroticepicardial coronary arteries. Lancet. 1990;336:897-900.

22. Vanhoutte PM, Shimokawa H. Endothelium-derived relaxingfactor and coronary vasospasm. Circulation. 1989;80:1-9.

23. Shimokawa H, Vanhoutte PM. Natural course of the impairmentof endothelium-dependent relaxations after balloon endothelium-removal in porcine coronary arteries: possible dysfunction of apertussis toxin-sensitive G protein. Circ Res. 1989;65:740-753.

24. Beny JL, Brunet PC. Neither nitric oxide nor nitroglycerinaccounts for all the characteristics of endothelially mediated vaso-constriction of pig coronary arteries. Blood Vessels. 1988;25:308-311.

25. Egashira K, Inou T, Yamada A, Hirooka Y, Takeshita A. Pre-served endothelium-dependent vasodilation at the vasospastic sitein patients with variant angina. J Clin Invest. 1992;89:1047-1052.

26. Yamamoto H, Yoshimura H, Noma M, Kai H, Suzuki S, Tajimi T,Sugihara M, Kikuchi Y. Preservation of endothelium-dependentvasodilation in the spastic segment of the human epicardialcoronary artery by substance P. Am Heart J. 1992;123:298-303.

27. Vanhoutte PM. The end of quest? Nature. 1987;327:459-460.28. Feletou M, Vanhoutte PM. Endothelium-dependent hyperpolar-

ization of canine smooth muscle. BrJPharmacol. 1988;93:515-524.29. Chen G, Suzuki H, Weston AH. Acetylcholine releases endothe-

lium-derived hyperpolarizing factor and EDRF from rat bloodvessels. Br J Pharmacol. 1988;95:1165 -1174.

30. Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M,Mitsui Y, Yazaki Y, Goto K, Masaki T. A novel potent vasocon-strictor peptide produced by vascular endothelial cells. Nature.1988;332:411-415.

31. Kontos HA, Wei EP, Kukreja RC, Ellis EF, Hess ML. Differencesin endothelium-dependent cerebral dilatation by bradykinin andacetylcholine. Am J Physiol. 1990;258:H1261-H1266.

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from

T Fukai, K Egashira, H Hata, K Numaguchi, Y Ohara, T Takahashi, H Tomoike and A Takeshitaendothelium-derived relaxing factor.

Serotonin-induced coronary spasm in a swine model. A minor role of defective

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1993 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.88.4.1922

1993;88:1922-1930Circulation. 

http://circ.ahajournals.org/content/88/4/1922World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer available in the

Permissions in the middle column of the Web page under Services. Further information about this process isOnce the online version of the published article for which permission is being requested is located, click Request

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Circulation Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on May 16, 2018

http://circ.ahajournals.org/D

ownloaded from