696

Cocaine-Induced Alterations in Prostaglandin Production inRabbit Aorta

lACC Vol. 19. No.3March I. 1992:696-703

ERIC J. EICHHORN, MD, FACC, SABA E. DEMIAN, MD,* LUIS G. ALVAREZ, MD,JOHN E. WILLARD, MD, SUSAN MOLINA, BS, MT, LORI L. BARTULA, BS,M. DALE PRINCE, BS, LINDSEY R. INMAN, PHD, PAUL A. GRAYBURN, MD, FACC,STUART I. MYERS, MD*Dallas, Texas

To determine if alterations in endothelial prostaglandin produc­tion occur after long-term cocaine use, 26 New Zealand Whiterabbits were randomized to a low fat diet with (n =12) or without(n =14) daily intravenous cocaine (2 mg/kg body weight). Rabbitswere killed at 6 or 12 weeks. Segments of aorta were examined inblinded manner for histologic changes. Additional slices wereincubated in oxygenated Krebs butTer and release of 6-keto­prostaglandin Fla , thromboxane B2 and prostaglandin E2 wasassayed by radioimmunoassay.

Minimal intimal histologic changes were seen in the aorta ofthree cocaine-treated rabbits. At 12 weeks 6-keto-prostaglandinFla was increased in the cocaine group (p =0.063) as comparedwith levels in the control group. When rabbits killed at 6 and 12

Cocaine abuse is a major health hazard in the United States.More than 22 million Americans have tried cocaine at leastonce, and 5 million are current users (1). Recent reports(1-5) document that cocaine abuse can result in myocardialischemia and infarction even in the absence of significantcoronary artery disease. Myocardial ischemia may occurimmediately after the use of cocaine (3,4) or hours to dayslater (6). The pathophysiologic mechanism or mechanismsresponsible for such a diverse temporal relation betweencocaine ingestion and myocardial events are unknown.However, cocaine-induced coronary artery vasoconstrictionhas been reported in patients after intranasal cocaine (7).This vasoconstriction appears to be primarily alpha-

From the Cardiac Catheterization Laboratory of the Dallas VeteransAdministration Medical Center and the *Departments of Internal Medicine(Cardiology Division), Pathology and Surgery at the University of TexasSouthwestern Medical Center, Dallas, Texas. This study was supported inpart by a Veterans Affairs Merit Review Grant, Washington, D.C. andNational Heart, Lung, and Blood Institute Ischemic Specialized Centers ofResearch Grant HL-17669 from the National Institutes of Health, Bethesda,Maryland. It was presented in part at the 40th Annual Scientific Sessions ofthe American College of Cardiology, Atlanta, Georgia, March 1991.

Manuscript received July II, 1991; revised manuscript received August21, 1991, accepted August 29, 1991.

Address for reprints: Eric J. Eichhorn, MD, Cardiac CatheterizationLaboratory (l1lA2), Veterans Affairs and University ofTexas, SouthwesternMedical Center, 4500 South Lancaster Road, Dallas, Texas 75216.

©1992 by the American College of Cardiology

weeks were considered together, increases in thromboxane B2(p =0.044) and a trend to increased prostaglandin E2 (p =0.083)were seen in the cocaine group. The ratio of thromboxane B2 to6-keto-prostaglandin Fla was increased in the cocaine groupcompared with that in the control group (p < 0.02).

These data suggest that an increase in prostaglandin produc­tion occurs in the vascular endothelium of rabbits ingestingcocaine before gross histologic changes are evident. In addition,thromboxane B2 increases disproportionately with respect to6-keto-prostaglandin Fla , suggesting that a milieu for thrombosismay exist in users of cocaine.

(J Am Coli CardioI1992;19:696-703)

adrenergically mediated because it can be blocked with thealpha-adrenergic antagonist, phentolamine (7). Furthermore,cocaine-induced vasoconstriction is more pronounced atsites of angiographically visible atherosclerosis than at sitesof angiographically normal coronary arteries (8). This obser­vation suggests that endothelial modulation of coronaryvasoconstriction may play an important role in the patho­physiologic mechanism of cocaine-induced vasoconstric­tion. It has also been suggested (4,5,9) that cocaine maypotentiate platelet aggregation and thereby contribute to thedevelopment of coronary artery thrombosis.

We (10) have shown that patients who have a history ofchest pain temporally related to cocaine use frequently haveevidence of endothelial dysfunction (that is, acetylcholine­induced coronary vasoconstriction). We (10) have also re­ported histologic evidence of accelerated atherosclerosis inthe coronary arteries of autopsy subjects who had a historyof chronic cocaine abuse. Thus, patients who have chestpain temporally related to cocaine use frequently haveevidence of endothelial dysfunction that may be related tofunctional or structural endothelial changes, or both.

Work from other investigators (11) has demonstrated adecrease in in vitro 6-keto-prostaglandin F1a (a stable me­tabolite of prostacyclin [PGI2]) production in umbilical ar­teries from pregnant cocaine abusers. These investigators(11) also found a dose-dependent decrease in phospholipase

0735-1097/92/$5.00

lACC Vol. 19, No.3March I, 199Z:6%-703

EICHHORN ET AL.PROSTAGLANDIN ALTERATIONS WITH COCAINE

697

Az activity in vitro with very large cocaine doses. Becauseprevious work from our group (12,13) has shown that throm­boxane Az is an important mediator of the conversion fromstable to unstable angina, we were interested in examiningthe role of prostaglandins in promoting cocaine-inducedischemia.

Therefore, the purpose of this study was to determine ifcocaine induces biochemical or histologic arterial changes inrabbit aorta. Our hypothesis was: Chronic cocaine adminis­tration alters endothelial prostaglandin production or pro­motes the development of histologic endothelial abnormali­ties, or both.

MethodsProtocol. Twenty-six New Zealand White rabbits were

randomized to a low cholesterol, low fat diet with (n = 12,group A) or without (n = 14, group B) a daily dose ofintravenous cocaine (2 mg/kg body weight). This dose wasbased on previous human studies (7) demonstrating 2 mg/kgto be safe, yet capable of producing a hemodynamicallysignificant effect. At 6 weeks, six group A and six group Brabbits were weighed in blinded manner and killed. At 12weeks of cocaine treatment, six group A and eight group Brabbits were killed in a similar fashion. This protocol wasapproved by the Animal Studies Subcommittee of the DallasVeterans Administration Medical Center, July 1990 andconforms to the "Position of the American Heart Associa­tion on Research Animal Use."

Aorta tissue preparation. On the animal's death, a 3- to5-mm cross section of aorta was removed and fixed in 10%buffered formalin for eventual histologic examination. Thesection was stained with hematoxylin-eosin and examinedhistologically by a pathologist (S.E.D.) who had no knowl­edge of the treatment randomization.

At the time of induced death, a 3- to 5-mm transversesection of aorta adjacent to the histologic slice was removedfor assessment of prostaglandin production. The section wasnot selected on the basis of prior histologic examination. Itwas immediately suspended in Krebs-Henseleit buffer (pH7.4, 37°C) and used to examine the synthesis and release ofprostanoids from group A and group B aortas. At 5, 15, 30and 60 min after suspension in the buffer, 100 JLl of bufferwas removed and immediately frozen at -80°C. Thesesamples were stored at -80°C until radioimmunoassay couldbe performed to analyze for prostanoids. After the lastsample was collected, the slices were analyzed for proteincontent. The tissue slices were weighed and homogenized inO.l-M potassium phosphate buffer with 0.01% triton with aUltraturrex homogenizer (Tekmar) for 150 s. The homoge­nate was spun at 2,000 rpm for 5 min and the supernatantdecanted and analyzed for protein content by the method ofLowry et al. (14).

Radioimmunoassay. Our methodology for radioimmuno­assay of 6-keto-prostaglandin Fla, thromboxane Bz andprostaglandin Ez has been previously described (15). In

brief, these prostanoids were measured from the unextractedeffluent perfusate by radioimmunoassay. The antiserum doesnot distinguish between prostaglandin Eland prostaglandinEz. All samples and standards were measured in duplicate.

Antisera were obtained in lyophilized form from Dr.Lawrence Levine, Department of Biochemistry, BrandeisUniversity (6-keto-prostaglandin Fla and thromboxane Bz)and Advanced Magnetics, Cambridge, Massachusetts (pros­taglandin Ez). Dilutions in assay buffer of 1:5,000, 1:8,000and 1: 12,000 for prostaglandin Ez, 6-keto-prostaglandin Fla

and thromboxane Bz, respectively, were used to obtain 40%to 60% binding of a zero standard "blank." Cross-reactivityof antisera from Dr. Levine has been published previously(16). The cross-reactivity of 6-keto-prostaglandin Fla is < 1%to prostaglandin Fza , prostaglandin Dz, thromboxane Bzand13,14-H2-prostaglandin Ez. The cross-reactivity of throm­boxane Bz antiserum is <1% to prostaglandin Fza, prosta­glandin Dz, 6-keto-prostaglandin Fza, 6-keto-prostaglandinF1a and 13,14-H2-prostaglandin Ez. Cross-reactivity of theprostaglandin Ezantiserum is given by Advanced Magneticsas 50% to prostaglandin EI and <1% to 6-keto-prostaglandinFla'

The standard curve was constructed with use of thesmoothed-spline function (RIA computer software packageSecuRIA, Packard). Unbound, unabsorbed tracer (referredto as "nonspecific binding") was quantified in tubes contain­ing only labeled ligand and dextran-coated charcoal. Interas­say and intra-assay variation was 7% to 9% for 6-keto­prostaglandin F1a and thromboxane Bz and 14% to 15% forprostaglandin Ez. Sensitivity (the amount of unlabeled stan­dard causing 5% reduction from maximal binding of theblank) was approximately 8 pg.

Statistical analysis. Differences in prostaglandin produc­tion between groups A and B were assessed (at 5, 15,30 and60 min of suspension time in buffer) by a repeated measuresanalysis of variance. As the thromboxane Bz data exhibitedsignificant heteroscedasticity, a logarithmic (base 10) trans­formation was used before analysis of data to achievecommon variances between the two groups. Because severalrabbits exhibited no thromboxane release (that is, 0 JLg/mgprotein), their thromboxane Bzvalues were set to 0.01 JLg/mgprotein and then 0.001 JLg/mg protein to avoid losing datawith a logarithmic transformation (loglo 0 does not exist).The data were analyzed with use ofboth values. Because thelower limit of sensitivity of the assay is 0.00008 JLg/ml(approximately 0.00032 JLg/mg protein, assuming 0.63 mgprotein per sample, which was the mean sample proteincontent), these were reasonably low values for thromboxanedetection. The effect of the presence of intimal changes onprostaglandin production was assessed by a 2-factor (intimalchanges, group A vs. B) analysis of variance at 5, 15 and30 min. Differences in heart weight and body weight wereassessed by an unpaired Student t test. Nominal variableswere assessed by contingency table analysis. Statisticalsignificance was established at p < 0.05 and a trend at p <

698 EICHHORN ET AL.PROSTAGLANDIN ALTERATIONS WITH COCAINE

JACC Vol. 19. No.3March I. 199Z:696-703

0.10. All results are expressed as mean values ± 1SD unlessotherwise specified.

ResultsGroup characteristics at induced death. Group A and B

rabbits did not differ in age or gender (chi-square = 0.26). Atthe time ofdeath, there was no difference in body weight (3.5± 0.2 vs. 3.7 ± 0.1 kg, group A vs. group B) or heart weight(6.9 ± 1.0 vs. 6.2 ± 0.8 g). In group A rabbits heart weightwas indexed to body weight and was slightly higher than thatof group B rabbits (2 ± 0.3 vs. 1.7 ± 0.2 g/kg), although thisdifference did not reach statistical significance (p = 0.13).

Prostaglandin results (Table 1). At 6weeks a strong trend(p = 0.082) of increased thromboxane Bz production wasfound in the cocaine-treated rabbits, although there were nodifferences between groups in 6-keto-prostaglandin FI a orprostaglandin Ez release (Fig. 1). At 12 weeks the cocainegroup had increased 6-keto-prostaglandin Fla release (p =0.063) but no significant differences in thromboxane Bz (p =

0.29) or prostaglandin Ez release with respect to the controlgroup (Fig. 2). Although the mean values for thromboxaneBz production were 23 to 34 times higher in the cocainegroup, this difference did not reach statistical significance.No significant differences between groups in prostaglandinEzproduction were seen at 12 weeks (p = 0.12), although thecocaine group had mean values 3 to 10 times larger at both 6and 12 weeks. This finding was due to a subgroup of rabbitswith very large prostaglandin production. When all 26 rab­bits (with 6 and with 12 weeks of treatment) in the twogroups were studied, a significant increase in thromboxaneBzproduction was seen overall (Fig. 3). Overall, no increasein 6-keto-prostaglandin Fla was found when all rabbits inboth groups were studied. For those rabbits with a throm­boxane Hz release of 0 JLg/mg protein, when a value of0.001 JLg/mg protein was used (instead of0.01 JLg/mg protein)to calculate a logarithmic transformation for statistical anal­ysis, no difference (0.01 vs. 0.001 JLg/mg protein) in p valuesoccurred.

The ratio ofthromboxane B2 to 6-keto-prostaglandin F1arelease was determined to ascertain if the increased prosta­glandin production was disproportional (Fig. 4). Previousinvestigators (17,18) have used a ratio of prostanoids toexamine relative changes in these substances. The cocainegroup had an 8- to 13-fold higher mean value for this ratio at5, 15, 30 and 60 min of prostaglandin production (p < 0.02).These data suggest that aortic thromboxane Bz (whichreflects thromboxane Az production) increases in somerabbits with long-term cocaine use. In addition, 6·keto­prostaglandin Fla (which reflects prostacyclin production)and perhaps prostaglandin Ez may increase with chroniccocaine exposure. In some rabbits chronic cocaine useproduced disproportionate increases in thromboxane Bz ascompared with 6-keto-prostaglandin Fla'

Histologic findings. Three of the rabbits killed at 6 weekshad minimal evidence of histologic changes consistent with

early intimal proliferation (Fig. 5). Although all three ofthese rabbits were in group A, we were unable to establish acorrelation between cocaine use and development of intimalchanges because of the small number of rabbits with changes(chi-square = 3.96, p = 0.047; chi-square with continuitycorrection = 1.89, p =0.17). Because the number of rabbitswith histologic changes was small, there was no relation ofintimal changes to alterations in either 6-keto-prostaglandinFla' thromboxane Bz or prostaglandin Ez.

DiscussionOur study demonstrates that cocaine abuse causes in­

creases in vascular prostaglandin production in rabbits.Furthermore, these data suggest a possible disproportionateincrease in thromboxane Az production as compared withprostacyclin production. These biochemical changes oc­curred despite the absence of significant histologic changes.

Previous studies. Previous studies of cocaine-inducedischemic heart disease (5,7,19) have focused on pharmaco­logic properties of cocaine that promote vasoconstriction.Such studies, although important, fail to explain why myo­cardial ischemic events often occur days or months aftercocaine ingestion (6). We (10) have previously demonstratedthat cocaine abuse is associated with accelerated atheroscle­rosis and a dysfunctional endothelial response to acetylcho­line, suggesting depressed release of one or more endotheli­um-derived relaxation factors. Although these previouslypublished data (10) demonstrated that chronic cocaine abusemay be associated with functional and anatomic abnormali­ties of the coronary artery endothelium, the data from ourcurrent study suggest that biochemical abnormalities mayalso be present. Because previous investigations (20,21)have shown increases in prostacyclin and thromboxaneproduction in the presence of atherosclerosis, an increase inprostaglandin production might not be unexpected in thepresence of intimal proliferation or endothelial injury. How­ever, we found that these biochemical changes were inde­pendent of histologic changes.

Speculation about why cocaine may cause ischemia. Al­though prior studies (3-5,7,19) have suggested that cocainecauses myocardial infarction through both norepinephrine­mediated alpha-adrenergic coronary vasoconstriction anddirect coronary vasoconstriction, it is unclear why coronaryvasospasm and myocardial infarction occur in associationwith acute or chronic use of cocaine in some persons but notin others. Furthermore, the occurrence of myocardial infarc­tion or ischemia well after cocaine ingestion in some patients(6) suggests that these events are not simply a direct conse­quence of cocaine-mediated coronary vasoconstriction inmany cases. The absence of chest pain or myocardialinfarction in some users of "crack" or freebase cocaine(who have high serum levels of cocaine) and its presence insome persons who inhale or "snort" cocaine (and havelower serum levels) also suggests that infarction in users isnot just a dose-related phenomenon of cocaine-induced

lACC Vol. 19. No.3 EICHHORN ET AL. 699March 1. 1992:6%-703 PROSTAGLANDIN ALTERATIONS WITH COCAINE

Table 1. Results at 6 and 12 Weeks for Prostaglandin Production From Group A(cocaine-treated) and B (untreated) Rabbit Aortas

5 Min 15 Min 30 Min 60 Min p Value

6 Weeks

6-keto-prostaglandin FI a

Group A (n = 6) 150 ± 137 166 ± 96 136 ± 75 210 ± 106 0.27Group B (n = 6) 151 ± 98 334 ± 277 327 ± 291 262 ± 163

Thromboxane B2

Group A (n = 6) 63 ± 43 5.0 ± 4.1 3.2 ± 2.5 4.8 ± 3.4 N/AGroup B (n = 6) 4.2 ± 6.6 3.9 ± 6 1.2 ± 2.1 2.1 ± 3.2

Log lO thromboxane BzGroup A (n = 6) 0.54 ± 0.76* 0.24 ± 1.14 0.11 ± 1.06t 0.26 ± 1.13 0.082Group B (n = 6) -1.63 ± 2.12 -0.98 ± 1.58 -1.16 ± 1.31 -1.07 ± 1.44

Prostaglandin EzGroup A (n = 6) 203 ± 13.6 17.4 ± 18 21.4 ± 25.4 32.5 ± 40.7 0.53Group B (n = 6) 21.2 ± 19.8 17.6 ± 8.7 12.8 ± 6.1 12.7 ± 7.6

Thromboxane Bz/6-keto-prostaglandin Fla

Group A (n = 6) 0.12 ± 0.17 0.036 ± 0.035t 0.032 ± 0.029 0.027 ± 0.017 0.26Group B (n = 6) 0.02 ± 0.03 0.007 ± 0.012 0.005 ± 0.01 0.008 ± 0.012

12 Weeks

6-keto-prostaglandin Fla

Group A (n = 6) 427 ± 1m 769 ± 270t 590 ± 182 571 ± 140 0.063Group B (n = 8) 204 ± 107 365 ± 158 579 ± 471 590 ± 275

Thromboxane BzGroup A (n = 6) 51.7 ± 88.8 44.4 ± 75.6 36.4 ± 62 323 ± 51.5 N/AGroup B (n = 8) 1.7±1.7 1.3 ± 1.4 1.3±1.3 1.4 ± 1.2

Log lO thromboxane BzGroup A (n = 6) 0.51 ± 1.89 0.65 ± 1.51 03 ± 1.61 0.48 ± 1.48 0.29Group B (n = 8) -0.59 ± 1.54 -0.63 ± 1.19 -0.61 ± 1.19 -0.07 ± 0.87

Prostaglandin EzGroup A (n = 6) 26.2 ± 33.lt 31.3 ± 46.5 28.5 ± 43.2 33.1 ± 46.4 0.12Group B (n = 8) 33 ± 4.6 4.8 ± 4 2.9 ± 3.6 9.8 ± 13

Thromboxane Bz/6-keto-prostaglandin FI a

Group A (n = 6) 0.133 ± 0.244 0.041 ± O.06t 0.048 ± 0.08 0.05 ± 0.079 0.16Group B (n = 8) 0.008 ± 0.008 0.003 ± 0.003 0.002 ± 0.002 0.003 ± 0.003

6 and 12 Weeks (all rabbits)

6-keto-prostaglandin FI a

Group A (n = 12) 289 ± 188t 467 ± 369 363 ± 272 390 ± 222 0.88Group B (n = 14) 181 ± 103 352 ± 208 471 ± 411 450 ± 282

Thromboxane BzGroup A (n = 12) 29 ± 64.5 24.7 ± 55.1 19.8 ± 45.3 18.6 ± 37.7 N/AGroup B (n = 14) 2.8 ± 4.5 2.4 ± 4.1 1.3 ± 1.6 1.7 ± 2.2

Log lO thromboxane BzGroup A (n = 12) 0.53 ± 1.37* 0.44 ± 1.29* 0.21 ± 1.3* 037 ± 1.26t 0.044Group B (n = 14) -1.04 ± 1.82 -0.78 ± 1.33 -0.85 ± 1.22 -0.53 ± 1.23

Prostaglandin EzGroup A (n = 12) 233 ± 243 243 ± 34.4 25 ± 34t 32.8 ± 41.6t 0.083Group B (n = 14) 11 ± 15.7 103 ± 9 7.2 ± 6.9 11 ± 10.8

Thromboxane Bz/6-keto-prostaglandin F1a

Group A (n = 12) 0.126 ± 0.2* 0.038 ± 0.047* 0.04 ± 0.058* 0.Q38 ± 0.056* 0.019Group B (n = 14) 0.014 ± 0.022 0.005 ± 0.008 0.003 ± 0.007 0.005 ± 0.008

*p < 0.05, tp < 0.1 tp < 0.005 versus group B. Values are expressed as ng prostaglandin/mg protein (mean ± I SO). N/A = not applicable because ofheteroscedasticity.

vasospasm. In addition, recent studies (22,23) have chal­lenged the hypothesis that cocaine causes vasoconstrictionor vasospasm in humans. Thus, other predisposing or per­missive factors may create a milieu facilitating cocaine­induced vasoconstriction or thrombosis, or both. Such a

milieu may permit vasoconstriction or thrombus formationto occur, even without the presence of cocaine.

We propose that coronary tone is maintained through thebalance of various endothelial and humoral factors (13,24­28). Several of these factors are responsible for coronary

100

lACC Vol. 19, No.3March I, 1992:696-703

900p-o.o&3

900 froo

I6-keto-PGF1 a. 900 f(ngImg 900

protein) 400 f ?000

900 2roo

0 ro :Il 00 «l Ii) Iil roTime

2 (Minutes)

450

400

350

000

700 EICHHORN ET AL.PROSTAGLANDIN ALTERATIONS WITH COCAINE

Ii)'---'---'"-"'---'-----'_...L-......o ro ID 00 ~ Ii) Iil ro

Time(MinUtes)

Ii) p-o.12

«l

I I00

PGE 2:Il(ngImg

protein)ro

0 000 10 :Il 00 «l Ii) Iil ro

Time(Minutes)

Figure 2. Changes in prostaglandin production at 12 weeks. At 12weeks release of 6-keto-prostaglandin Fla (PGFla), thromboxane Bzand prostaglandin Ez(PGEz)were assayed at 5, 15,30 and 60 min forgroups A (+) and B (0). No differences were seen in prostaglandinEz•Although a 23- to 34-fold increase in mean values for thrombox­ane Bzwas seen in the cocaine group, this difference did not reachstatistical difference (p = 0.29). A strong trend to increased 6-keto­prostaglandin F1a release was found in the cocaine group (p =0.063). Values are expressed as mean values ± SEM.

f

f

o'---'---'"_"'---'-----'_...L-......o ro :Il 00 ~ Ii) Iil ro

Time(Minutes)

PGE2

00 I t I(ng!mg IDprotein)

10

ro ID 00 ~ Ii)Time

(minUtes)

Figure 1. Changes in prostaglandin production at 6 weeks. At 6weeks release of 6-keto-prostaglandin Fla (PGFla), thromboxane Bzand prostaglandin Ez(PGEz)were assayed at 5,15,30 and 60 min forgroups A (+) and B (0). No differences were seen in 6-keto­prostaglandin F1a and prostaglandin Ez. A trend to increasedthromboxane Bz (p = 0.082) was found at 6 weeks. Values areexpressed as mean values ± SEM.

constriction, including thromboxane A2, norepinephrine, sero­tonin, neuropeptide Y and endothelin. Among the vasodila­tors are endothelium-derived relaxation factors, substanceP, vasoactive intestinal peptide, prostacyclin, nucleosidesand nucleotides (adenosine triphosphate and diphosphateand adenosine), histamine, serotonin (with intact endothe­lium) and bradykinin. Arelative lack of endothelium-derivedrelaxation factors and an increase in the ratio of thrombox­ane A2 to prostacyclin may predispose a person to vasocon­striction, vasospasm or thrombosis, especially in the pres­ence of a vasoconstricting agent such as cocaine.

Cocaine-induced vasospasm and thrombosis. Thus, co­caine abuse may induce an altered endothelial milieu thatpermits vasoconstriction or thrombosis, or both. Such ab­normalities include reduced release of endothelium-derivedrelaxation factors and increased production of prostaglan­dins, with disproportionate increases in thromboxane (com­pared with prostacyclin). As we (10) and others (29) have

previously shown atherosclerosis in postmortem studies insome abusers, flow may be further altered by these lesions.Development of these lesions may also predispose to throm­bosis and infarction. Our group (8) has previously demon­strated accentuated vasoconstriction in response to cocaineadministration at sites ofatherosclerosis in noncocaine abus­ers undergoing cardiac catheterization. Thus, cocaine ad­ministration, in the presence of this altered milieu, maypredispose to vasospasm and thrombosis. Whereas thesepatients often have angiographically normal coronary arter­ies (5,10), they may have intimal changes that are notdetected by angiography. Although the ability of cocaine toinduce vasoconstriction may be endothelium independent(5), dysfunctional endothelium may playa role in modulatingcoronary vasoconstriction in the presence of cocaine or maypromote in situ thrombosis and vasoconstriction even in theabsence of cocaine.

JACC Vol. 19, No.3March I, 1992:696-703

EICHHORN ET AL.PROSTAGLANDIN ALTERATIONS WITH COCAINE

701

0.20 p-o.16

0.<1'

0.16

0.20

0.06

0.16

TxB2 /6-keto PGFl~1212 weeks 0.06

! I0.<1' I0

0 10 :aJ :Il 40 50 81 'lOTime

0.20 (Minutes)

0.16

Ip.o.o19

TxB /6-ketoPGF 0.122 la

ADRabbits 0.08

0.<1' I I f0

0.000 10 :aJ :Il~ 50 81 'lO

(Minutes)

TxB2 /6-keto PGF1~·12

6 weeks

000

1000 10 :aJ :Il 40 50 00 '10

lime(minutes)

f fp-ll.044

I flog'Ihromboxane

f t tB2 fOOfrlo ngImg -1protein)

-20 10 :aJ :Il 40 50 00 '10

'lime(minutes)

50

p_o.oIl3

140

PGE 2 :Il

I I I(ng/mgprotEin) :aJ

10 ~ 2 ~Q

p-0.88

: !I6-keto-PGFla

~ ~ f~ 2

10:aJ:Il 4050 00'10lime

(minutes)

Figure 3. Overall changes in prostaglandin production at 6 and 12weeks. Release of 6-keto-prostaglandin FIll (PGF1a), thromboxaneB2 and prostaglandin E2 (PGE2) were assayed for all the rabbits(those killed at 6weeks and those killed at 12 weeks) at 5, 15,30 and60 min for groups A (+) and B (0). No differences were seen in6-keto-prostaglandin Fill release. However, increases in thrombox­ane B2 (p = 0.044) and a trend toward increased prostaglandin E2(p = 0.083) were detected. Values are expressed as mean values ±SEM.

Figure 4. The ratios of thromboxane B2 to 6-keto-prostaglandin FIll(TxBi6-keto PGF la) are shown for rabbits killed at 6 weeks or 12weeks and the total group ofrabbits killed at each time period at 5,15, 30 and 60 min for groups A (+) and B (0). Although this ratiowas increased at 6 and 12 weeks, this increase did not reachstatistical significance until all rabbits were analyzed (p < 0.019).

This milieu may change somewhat in the initial stages ofcocaine abuse as intimal proliferation or accelerated athero­sclerosis develops. In our study the increase in thromboxanewas greater at 12 than at 6 weeks in a subgroup of rabbits. Inaddition, in some rabbits 6-keto-prostaglandin Fla was un­changed or slightly decreased at 6 weeks but was increasedat 12 weeks. It is unclear what the effect of cocaine is after12 weeks.

Study limitations. The major limitation of this study wasthe short time period in which the rabbits were treated. Mostdrug abusers have a much longer history of abuse and manymay use cocaine more than once each day. We might haveseen more intimal changes and more pronounced changes inprostaglandin production had we studied more frequent andlonger periods of cocaine administration. In addition, be­cause the control rabbits (group B) received no injections

such as saline solution, we cannot rule out an effect of theinjecting process on the endothelium. We used pure indus­trial grade cocaine that contained none of the many impuri­ties found in street-purchased cocaine. Such impurities mayhave additional effects on the development of intimal his­tologic changes and on prostaglandin production.

Because our preparations were devoid of circulatingplatelets and white blood cells, the contribution of theseelements to prostaglandin production was not assessed. Invivo, at sites of endothelial injury and constriction, plateletsmay aggregate and release mediators that promote sponta­neous reductions in coronary blood flow and thrombosis(12,13,30). Among these mediators are serotonin and throm­boxane Az. Thus, our assessment of endothelial thrombox­ane release may have underestimated the total amount ofthromboxane Az present in the coronary artery in vivo.However, the relative increase in vascular thromboxaneproduction, especially in the presence of coronary vasocon-

702 EICHHORN ET AL.PROSTAGLANDIN ALTERATIONS WITH COCAINE

lACC Vol. 19. No.3March I, 1992:696-703

Figure 5. Histologic changes in the rabbit aorta can be seen in thesephotomicrographs. Above, a normal rabbit aorta (group B). Below, across section of aorta from a cocaine-treated rabbit with earlyintimal hyperplasia and vacuolation. The arrows point to the abnor­mal changes. Hematoxylin-eosin x50, reduced by 25%.

striction due to cocaine use, may initiate or perpetuateplatelet mediator release, promoting a cascade of aggrega­tion and further mediator release.

Although the buffer, which was changed at 5, 15, 30 and60 min, could be a weak stimulus for prostaglandin release,it should have effected equal changes in the two prospec­tively randomized groups. Our reproducibility data havebeen previously published (15).

ReferencesI. Isner JM, Estes NAM III. Thompson PO, et al. Acute cardiac events

temporally related to cocaine abuse. N Engl J Med 1986;315:1438-43.2. Simpson RW, Edwards WD. Pathogenesis of cocaine-induced ischemic

heart disease. Arch Pathol Lab Med 1986;110:479-84.3. Zimmerman FH. Gustafson GM, Kemp HG Jr. Recurrent myocardial

infarction associated with cocaine abuse in a young man with normalcoronary arteries: evidence for coronary artery spasm culminating inthrombosis. J Am Coli Cardiol 1987;9:964-8.

4. Smith HWB, Liberman HA, Brody SL, Battey LL, Donohue BC, MorrisDC. Acute myocardial infarction temporally related to cocaine use.Clinical, angiographic, and pathophysiologic observations. Ann InternMed 1987;107:13-8.

5. Isner JM, Chokshi SK. Cardiovascular complications of cocaine. CurrProbl Cardioll99I;16:95-123.

6. Nademanee K, Gorelick DA, Josephson MA, et al. Myocardial ischemiaduring cocaine withdrawal. Ann Intern Med 1989;111:876-80.

7. Lange RA, Cigarroa RG, Yancy CWo et al. Cocaine-induced coronaryartery vasoconstriction. N Engl J Med 1989;321:1557-62.

8. Flores ED, Lange RA. Cigarroa RG, Hillis LD. Effect of cocaine oncoronary artery dimensions in atherosclerotic coronary artery disease:enhanced vasoconstriction at sites of significant stenoses. J Am ColiCardiol 1990;16:74-9.

9. Togna G. Tempesta E. Togna AR, Oolci N, Cebo B. Caprino L. Plateletresponsiveness and biosynthesis of thromboxane and prostacyclin inresponse to in vitro cocaine treatment. Haemostasis 1985;15:100-7.

10. Eichhorn EJ, Grayburn PA, Bedotto JB, Willard JE, Demian SE,Willerson JT. Abnormal endothelium-dependent coronary vasoreactivityand intimal proliferation in patients with chronic cocaine abuse (abstr).Clin Res 1990J8:293A.

II. Cejtin HE, Parsons MT, Wilson L. Cocaine use and its effect on umbilicalartery prostacyclin production. Prostaglandins 1990;40:249-57.

12. Hirsh PO, Hillis LD, Campbell WB, Firth BG, Willerson JT. Release ofprostaglandins and thromboxane into the coronary circulation in patientswith ischemic heart disease. N Engl J Med 1981 ;304:685-91.

13. Willerson JT, Hillis LD, Winniford M, Buja LM. Speculation regardingmechanisms responsible for acute ischemic heart disease syndromes.J Am Coli Cardiol 1986;8:245-50.

14. Lowry O. Rosenbrough N, Farr A, Randall R. Protein measurement withthe folin phenol reagent. J Bioi Chem 1951 ;193:265-75.

15. Reed MK. Taylor B, Myers SI. The effect of hypoxia on rat splanchnicprostanoid output. Prostaglandins 1989;38:599-608.

16. Levine L. Measurement of arachidonic acid metabolites by radioimmu­noassay. In: Fahey JL, Friedman H, Rose NR, eds. Manual of ClinicalLaboratory Immunology. American Society for Microbiology, Washing­ton. D.C.. 1976:685.

17. Serneri GGN, Abbate R, Gensini GF, Panetta A, Casolo GC, Carini M.TxA2 production by human arteries and veins. Prostaglandins 1983;25:753-66.

18. Kent RS, Diedrich SL, Whorton AR. Regulation of vascular prostaglan­din synthesis by metabolites of arachidonic acid in perfused rabbit aorta.J Clin Invest 1983;72:455-65.

19. Rongione AJ. Isner JM. Cocaine-induced contraction of vascular smoothmuscle is inhibited by calcium channel blockade (abstr). J Am ColiCardioI1989;13(suppl A):78A.

20. FitzGerald GA, Smith B, Pedersen AK, Brash AR. Increased prostacy­c1in biosynthesis in patients with severe atherosclerosis and plateletactivation. N Engl J Med 1984;310:1065-8.

21. Mehta JL. Lawson 0, Mehta P, Saldeen T. Increased prostacyclin andthromboxane A2 biosynthesis in atherosclerosis. Proc NaIl Acad Sci USA1988;85:4511-5.

22. Majid PA. Cheirif J, Abukhalil J. Cocaine-induced vasoconstriction: factor fancy? (abstr). Circulation 1990;82(suppl III):III-278.

23. Bedotto JB, Lee RW, Lancaster LD. Olajos M, Goldman S. Cocaine andcardiovascular function in dogs: effects on heart and peripheral circula­tion. J Am Coli CardiolI988;11:1337-42.

24. Vatner SF. Higgins CB, Braunwald E. Effects of norepinephrine oncoronary circulation and left ventricular dynamics in the conscious dog.Circ Res 1974;34:812-23.

IACC Vol. 19, No.3March 1, 1992:696-703

EICHHORN ET AL.PROSTAGLANDIN ALTERATIONS WITH COCAINE

703

25. Yanagisawa M, Kurihara H, Kimura S, et al. A novel potent vasocon·strictor peptide produced by vascular endothelial cells. Nature 1988;332:411-5.

26. Furchgott RF, Zawadzki IV, Jothianandan D. Endothelial cells as medi­ators of vasodilation ofarteries. J Cardiovasc PharmacoI1984;6:S336-43.

27. Harrison DG. From isolated vessels to the catheterization laboratory:studies of endothelial function in the coronary circulation of humans.Circulation 1989;80:703-6.

28. Ganz P, Alexander RW. New insights into the cellular mechanisms ofvasospasm. Am J CardioI1985;56:1IE-5E.

29. Dressler FA, Malekzadeh S, Roberts WC. Quantitative analysis ofamounts of coronary arterial narrowing in cocaine addicts. Am J Cardiol1990;65:303-8.

30. Eichhorn EJ, Grayburn PA, Willard JE, et al. Spontaneous alterations incoronary blood flow velocity before and after coronary angioplasty inpatients with severe angina. J Am Coli Cardioll991;17:43-52.