JACC Vol. 23, No . 3March 1, 1994:665-71

Dipyridamole Direct'lly

VProliferation In 000Trealmmemit ed'Restenosis After AngioplastyJAI PAL SINGH, PHD, KIMBERLY J . ROTHFUSS, BS, TODD R . WIERNICKI, MS,WILLIAM B. LACEFIELD, PHD, WILLIAM L. KURTZ, MS, RAYMOND F. BROWN, MS,KELLIE A. BRUNE, RVT, DIANNA BAILEY, RVT, GREGORY P . DUNE, PHDIndianapolis, Indiana

Objectives . The effect of dipyridamole on smooth muscle cellproliferation and urevenlion of intional thickening after arterialinjury was investigated .

Background . In addition to andplatelet activity, dipyridamolealso inhibits cell proliferation . We examined whether the antipro-liferative action of dipyridomole on smooth niuscle cells, asdemonstrated here, has a direct effect on intimal thickening aftervascular injury .

Methods. Cell proliferation was determined by measuringdeoxyribonucleic acid (DNA) synthesis and by cell counting . Thein vivo effect of locally delivered dipyridamole was determined ina rabbit model with carotid or femoral artery injury .

Results . Dipyridamole produced a dose-dependent inhibition ofsmooth muscle cell proliferation, producing 50% inhibition at7 jag/mi . Structural analogues SH-869 and mopamidol were 10 to100 times less effective than dipyridamole, suggesting that cellgrowth inhibition may be unrelated to the artiplatelet activity of

Restenosis remains a major long-term complication of per-cutaneous transluminal coronary angioplasty . In 35% to 40%of patients undergoing coronary angioplasty, reocclusionoccurs within the I st 3 to 6 months . At present, no pharma-cologic treatment is available for the prevention of resteno-sis. Studies in humans and in experimental animal modelshave shown that excessive proliferation of vascular smoothmuscle cells may be associated with the pathogenesis ofrestenosis after coronary angioplasty (1-5) . Because plate-lets accumulate at the site of vascular injury after coronaryangioplasty and contribute polypeptide growth factors thatstimulate proliferation of smooth muscle cells, it is thoughtthat platelets play an important role in the development ofrestenosis (3,6) . In addition, a direct role of thrombus inrestenosis has been suggested (3,7,8) . These observationshave led to the testing of antiplatelet agents for the treatment

From Eli Lilly and Company, Indianapolis, Indiana .Manuscript received March 24, 1993 ; revised manuscript received Octo-

ber 18, 1993, accepted October 20, 1993 .AsIdress forcoavawn=w Dr. Jai Pal Singh, Cardiovascular Vesearch,

Building 2812, Lilly Research Laboratories, Indianapolis, Indiana 46285 .

01994 by the American College of Cardiology

665

INTERVENTIGNAL CARDIOLOGY

th Muscle

InAgyWaJolls ill dace

dipyridamole. Inhibition of cell proliferation by dipyridamole wasattenuated by increasing the serum concentration in the culturemedium . Bypassing serum by local delivery of dipyridamole at theperiadventitial site produced 63% inhibition (p < 0 .05) of cellreplication in balloon-injured arteries . Locally delivered dipyri-damole also inhibited intimal thickening (20%, p < 0 .05) afterballoon injury .

Conclusions . Dipyridamole inhibited smooth muscle cell pro-liferation in vitro . This activity was attenuated by serum proteins .Locally delivered dipyridamole inhibited cell replication in arter-ies and intimal thickening after balloon injury. These resultssuggest that although systemic treatment with dipyridamele maynot be efficacious because of inadequate serum levels, its anti-proliferative action on smooth muscle cells may reduce restenosiswhen the drug is delivered locally after coronary angioplasty .

(J Ain Coll Cardiol 1994 ;23:665-71)

of restenosis after coronary angioplasty (9-11) and coronaryartery bypass grafting (12-14) .

Dipyridamole (2,6-bis(diethanolamino)-4,8-dipiperidi-nopyrimido-[5,4-dipyrimidine [Persantine]) has been shownto inhibit both platelet aggregation in humans and thrombusformation in the injured artery of rabbits (15) . It has alsobeen shown to reduce the incidence of thromboembolism inpatients with prosthetic vascular devices (16) and, in somestudies (12), also been shown to reduce reocclusion and thedevelopment of atherosclerosis after coronary artery bypassgrafting . In the rabbit model of restenosis, dipyridamole atdoses higher than those tolerated clinically was found toprevent lumen narrowing after balloon injury (17) . Highdoses of dipyridamole were also found to reduce intimalthickening in the baboon after arterial injury by homocystinefeeding (18) . However, dipyridamole did not reduce theincidence of restenosis after coronary angioplasty in humanclinical trials (9,10) . In a prospective analysis of these data,dipyridamole did not significantly reduce the rate of resten-osis, as defined by a loss of ?50% of the gain in lumendiameter achieved by angioplasty, but the severity of lesionsin the treated group was reduced (11) . That analysis sug-

0735-1097/94/$7 .00

6"

SINGH ET AL.DIPYRIDAMOLE AND SMOOTH MUSCLE CELL PROLIFERATION

gested that drug treatment was modestly effective in modi-fying the restenosis process . The biologic basis for thereduction of lesion '.everity by dipyridamole and the con-flicting results with regard to efficacy in animal models andhumans is not clear . In the present study, we report thatdipyridamole inhibits vascular smooth muscle cell prolifera-tion in culture. On the basis of pharmacokinetic studies inhumans and animals (19-21), we estimated that serum levelsof dipyridamole at the dose regimen used in previous coro-nary angioplasty trials (9,10) may not have been sufficient toproduce an antiproliferative effect in vivo . We thereforehypothesized that although circulating serum concentrationsof dipyridamole may not be sufficient to produce an anti-proliferative effect when the drug is delivered systemically, adirect application at the site of injury may inhibit restenosisafter balloon injury . Our results show that local delivery ofdipyridamole produced a significant reduction in intimalthickening in an animal model of vascular smooth musclecell proliferation .

MethodsPrimary cultures of vascular smooth muscle cells. Vas-

cular smooth muscle cells were prepared from New ZealandWhite rabbits as described previously (22) . Aortas wereremoved aseptically ; the adventitia was stripped off; and I-to 2-mm explants were cultured in Dulbecco's modifiedEagle medium containing 10% fetal bovine serum, L-glutamine (2 mmol/liter), penicillin (100 Wad) and strepto-mycin (100 µg/ml) . Smooth muscle cells were observedgrowing from the explants within 5 to 7 days . After 7 to 10days, the explants were removed, and cells were subculturedby trypsinization and propagated in the previously describedmedium .

Determination of DNA synthesis. Deoxyribonucleic acid(DNA) synthesis in confluent cultures of smooth musclecells was determined by 3H-thymidine incorporation, asdescribed in our previous studies (22,23). Smooth musclecells from passages I to 3 were plated into 96-well microtiterplates in Dulbecco's modified Eagle medium containing 10%serum at a starting cell density of 4 x 103 cells/well . After 5days, when the cultures appeared confluent, medium wasreplaced with 0.2 ml of Dulbecco's modified Eagle mediumcontaining 2% platelet-poor plasma, 20 ng/ml of platelet-derived growth factor and I UCi/Ml of 3H-thymidine . Stocksolutions of test compounds were prepared in dimethyl-sulfoxide and then serially diluted to the indicated concen-trations . The final dimethylsulfoxide concentration was:50,1% . The assay plates were incubated at 37°C under 5%carbon dioxide and 95% air. After 24 h, cells were washed inphosphate-buffered saline solution and then fixed in metha-nol. Radioactive thymidine incorporation into DNA wasdetermined by scintillation counting .

Determination of cell growth . Smooth muscle cells frompassages 2 to 3 were plated into 12-well plates (1 x 104cells/well) in Dulbecco's modified Eagle medium containing

JACC Vol. 23, No. 3March 1, 1994 :665-71

10% fetal calf serum . Cells were allowed to attach to platesfor 16 to 18 h at 37°C in 5% carbon dioxide and 95% air . Themedium was then replaced with Dulbecco's modified Eaglemedium containing 2% platelet-poor plasma, 20 ng/ml ofplatelet-derived growth factor and the indicated concentra-tion of compound . After 72 h, cell growth was determined bytrypsinization and counting using a ZM-Coulter counter.

Rabbit model of vascular injury. 3H-Thynddine incorpo-ration studies . New Zealand White male rabbits (2 .3 to 3 .0 kg)were anesthetized with ketamine hydrochloride (50 mg/kgbody weight intramuscularly) and xylazine (50 mg/kg intra-muscularly) . Aseptic surgical fields were prepared over thescapulae and the ventral surface of the neck by shaving andwashing the skin with 70% isopropanol, Alzet osmoticpumps (model 2ML2, Alza Corp .) conta, ping dipyridamolesolution or vehicle (10% dimethylsulfoxide, 90% polyethyl-ene glycol-300) were implanted in a subcutaneous dorsalpocket . The dose of dipyridamole delivered adventitiallywas varied by altering the concentration in the pump .Renathane catheters (Braintree Scientific) were fitted to theflow moderator on the osmotic pump, and the free end of thecatheter was tunneled subcutaneously to the ventral surfaceof the neck . The free end of the renathane catheter wasanchored to tissues adjacent to a central region of the rightcommon carotid artery, thereby providing long-term deliv-ery of vehicle or dipyridamole to the adventitial surface ofthe carotid artery . Immediately before and 4 days after pumpimplantation, arterial blood pressure was measured in eachrabbit through an indwelling catheter placed in the femoralartery for this purpose . Four days after pump implantation,a 3F embolectomy balloon catheter (model CV1040, Amer-ican V . Mueller) with the balloon deflated was introducedinto the right internal carotid artery and advanced in thecommon carotid artery to the aortic arch . The balloon wasinflated, and the catheter was withdrawn to a point near itsinsertion. The inflation volume produced balloon diametersthat routinely stretched the artery to approximately 1 .8 timesthe normal artery diameter, as measured while the inflatedballoon was in the artery . After two additional balloonpasses, the catheter was withdrawn . The internal carotidartery was ligated, and the rabbits were allowed to recover .This procedure allowed maintenance of blood flow in thecommon carotid artery. Seventy-two hours after ballooninjury, the rabbits were anesthetized as previously de-scribed. A 1-cm segment of the right common carotid arteryand a similar segment from the non-injured left commoncarotid artery were excised, and cell proliferation activitywas determined by modifications of methods used previ-ously (24). In short, arteries were immersed in 1 .5 ml ofDulbecco's modified Eagle medium containing I mg/ml ofbovine serum and I ALCi of 3 H-thymidine/ml and wereincubated in 5% carbon dioxide and 95% air at 37°C . After4 h of incubation, the arterial segments were rinsed inphosphate-buffered saline solution followed by 10% trichlo-roacetic acid (to remove free 3H-thymidine), placed in scin-tillation cocktail and counted for beta emission . The disin-

?ACC Vo' . 23, No . 3March I L 1994:665-71

Figure 1 . Inhibition of vascular smooth muscle cell pro-liferatior, by dipyridarnole . Left, Confluent cultures ofrabbit aorta smooth muscle cells in 96-well microtiterplates were incubated in Duibecco's modified Eaglemedium containing 2 17c platelet-poor plasma . 20 ngiml ofplatelet derived growth factor, I yCi/ml of 'H-1hyrnidineand the indicated concentrations of dipyridaniole . 'H-Thymidine incorporation into deoxyribonucleic acid(DNA) was determined after 24 h . Data shown here aremean value ± SD from four different experiments (eachexperiment was performed in triplicate) . Right, Smoothmuscle cells were seeded in 12-well plates in Dulbecco'smodified Eagle medium containing 10% serum . Afterovernight incubation, medium was changed to Dulbec-co's modified Eagle medium containing 2% platelet-poorplasma, 40 nglml of platelet-derived growth factor andthe indicated concentrations of dipyridamole . Cellgrowth was determined after 4 days by counting using aZM Coulter counter . Data are plotted as mean value ±SD from triplicate observations . 'p < 0 .05 comparedwith control .

tegrations per minute . reflecting thymWine incorporationinto DNA, were normalized to tissue weight, and the result-ing values in the balloon-injured segment were divided bythe weight-normalized values in the noninjured segment ofthe contralateral artery .

Infitnal thickening studies . Rabbits were .prepared forsterile surgery as previously described, except that the skinoverlying the right groin rather than the neck was sterilized .Alzet osmotic pumps were implanted as previously de-scribed, and the free end of the drug delivery catheter wastunneled subcutaneously to the adventitial surface of the rightfemoral artery at a point 1 .5 cm anterior to the bifurcation ofthe saphenous and femoral arteries. The catheter tip wassutured to tissues adjacent to the femoral artery . The rabbitsrecovered and were maintained for 4 days before theyunderwent reanesthetization and balloon injury to the rightfemoral artery . A 2F embolectomy balloon catheter withballoon deflated was introduced through an arieriotomy intothe saphenous artery and advanced retrogradely into theabdominal aorta . The balloon was inflated and the catheterwithdrawn until the balloon was just anterior to the insertionsite. Two additional passes of the inflated balloon weremade, and the catheter was removed . The saphenous arterywas ligated, and the skin was closed with a suture . Bloodflow was maintained in the aortoiliofemoral circuit by way ofthe popliteal and caudal femoral arteries distal to the saphe-nous-carotid artery bifurcation . Fourteen days after ballooninjury, the rabbits were anesthetized as previously describedand perfusion fixed with Ringer's solution followed by 10%buffered zinc formalin (Z-fix, Anatech Ltd .). The femoralartery was excised and transected into three 5-mm segments,and the segments were embedded in paraffin . Each segmentwas sectioned, stained with aldehyde-fuchsin elastic tissuestain and analyzed morphometrically for cross-sectionalintimal area . The mean area of the three segments served torepresent the intimal response for each animal .

SINGH ET AL .

667DIPYRiDAN-IOLF. AND SMOOTH MUSCLE CELL PROLIFERATION

60

10

20

30

40Dipyridamote (pg/ml)

Statistical analysis. In vivo data for group comparisonwere analyzed by the Fisher protected least significantdifference test . Data from in vitro DNA synthesis and cellproliferation were analyzed using analysis of variance/Dunnett method on actual values .

Results

Inhibition of vascular smooth muscle cell proliferation bydipyridamole. FigLft- I shows the effect of dipyridamole onproliferation of vascular smooth muscle cells . Early passagesmooth muscle cells were grown to confluence in 96-wellrnicrotiter plates and then transferred to medium containing2% platelet-poor plasma . Our previous studies have shownthat growth-arrested quiescent cultures of smooth musclecells exhibit a low level of DNA synthesis . Treatment withplatelet-derived growth factor produces approximately six-fold to eightfold stimulation of DNA synthesis, as mea-sured by 3H-thymidine incorporation over a 24-h period(22,23) . Figure 1 (left) shows that addition of dipyridamole toplatelet-derived growth factor-stimulated cells resulted in adose-dependent inhibition of 3H-thymidine incorporation inDNA. Fifty percent inhibition of DNA synthesis wasachieved at 0.035 Aglml of dipyridamole . Similar resultswere obtained when smooth muscle cell growth was stimu-lated with other mitogens (e.g ., fibroblast growth factor,epidermal growth factor) or in combination with platelet-derived growth factor. Light microscopic examination ofdipyridamole-treated culures showed normal cell morphol-ogy. Biochemical analysis of cell viability using release o fintracellular lactate dehydrogenase as an indicator of celldeath also showed that inhibition of DNA synthesis bydipyridamole was not due to cellular toxicity (data notshown). These results suggested that dipyridamole inhibitedgrowth factor-induced DNA synthesis in vascular smoothmuscle cells .

668

SINGH ET AL.DIPYRIDAMOLE AND SMOOTH MUSCLE CELL PROLIFERATION

To establish that inhibition of DNA synthesis representedan inhibition of cell proliferation, the effect of dipyridamoleon exponentially growing smooth muscle cells was deter-mined. Smooth muscle cells were seeded in 12-well tissueculture plates in Dulbecco's modified Eagle medium contain-ing l0% serum. After 20 h, when the cells had attached toplates, medium was replaced with Dulbecco's modifiedEagle medium containing 2% platelet-poor plasma and vary-

60ing concentrations of dipyridamole . Cells were then stimu-lated by addition of platelet-derived growth factor . Cultureswere incubated for 4 days, and cell growth was then deter- 40 -1

mined by a change in cell number using a ZM Coultercounter. In control cultures, without dipyridamole, smoothmuscle cells grew approximately fourfold from the initialseeding density . Addition of dipyridamole produced a dose-dependent inhibition of cell proliferation . Fifty percent inhi-bition of cell growth was achieved at 7 p /ml (Fig . I , right) .These results confirm that dipyridamole acted as an inhibitorof vascular smooth muscle cell proliferation .

Effect of serum on the tiprolliferative activity of dipyri -.On the basis of the antiplatelet activity of dipyri-

damole, two randomized, double-blind, placebo-controlledclinical trials have been performed to test the efficacy ofdipyridamole in preventing restenosis after coronary angio-plasty (9,10) . In these studies dipyridamole was found to beineffective in reducing the rate of restenosis (9,10) . Thesedata could be interpreted to indicate that inhibition ofplatelet aggregation may not be sufficient to prevent resten-osis. Our results demonstrating inhibition of vascularsmooth muscle cell proliferation by dipyridamole led us toreinvestigate the potential reasons for the lack of efficacy ofdipyridamole . We reasoned that although the lack of efficacyof dipyridamote could be due to incomplete prevention ofplatelet adhesion/aggregation or to production of growthfactors by cells other than platelets, or both, an adequateblood level of dipyridamole could be anticipated to affectrestenosis through a direct antiproliferative mechanism . Inthe coronary angioplasty trials, the dose regimen used was acombination of aspirin and dipyridamole (300 and 75 mg,respectively, three times a day) . This dose may produce anestimated serum concentration of 0.5 to 1 .0 pg/mi. Becausedipyridamole has been shown to bind to serum proteins, wetested whether such binding reduces the free concentrationof dipyridamole, rendering it ineffective for inhibition ofsmooth muscle cell proliferation. The effect of dipyridamoleon DNA synthesis in the presence of varying concentrationsof serum was determined as shown in Figure 1 . The resultspresented in Figure 2 show that increasing the concentrationof serum in the culture medium resulted in a pronouncedattenuation of the antiproliferative activity of dipyridamole .For example, the concentration of dipyridamole required for50% inhibition of DNA synthesis in cells incubated inmedium containing 1% serum was 0 .03 pg/ml. Increasing theserum concentration to 20% resulted in a 15-fold increase inthe concentration of dipyridamole (0 .45 pg/ml required toachieve 50% inhibition). The effect of plasma on the growth

ec

0.01

JACC Vol . 23, No . 3March 1, 1994 :665-71

9

2

20

.1

1

10

13tpyrfdamote (tigaml)

Figure 2. Effect of serum concentration on inhibition of cell prolif-eration by dipyridamole . Confluent cultures of smooth muscle cellsin 96-well microliter plates were incubated in Dulbecco's modifiedEagle medium containing 40 ng/ml of platelet-derived growth factorand 1% (solid circles), 10Jo (open circles) or 20% (triangles) serum inthe presence of various concentrations of dipyridamole . Deoxyribo-nucleic acid (DNA) synthesis during 24-h incubation was deter-mined as described in Figure 1 . The Inset shows inhibition of cellproliferation, as determined by change in cell number at the end of4 days in the presence of a fixed concentration (5 Ag/ml) of dipyri-damole when the cells were incubated in growth medium containingeither 2% or 20% platelet-poor plasma . Data shown here are meanvalue ± SD of triplicate observations. *p < 0.05 w';,h respect to 1%serum. **p < 0.05 with respect to 1% and 10% serum . With theexception of the highest dose of dipyridamole, each curve differs(p < 0.05) from each of the other two curves .

inhibitory activity of dipyridamole was also determined bymeasuring the change in cell number. Figure 2 (inset) showsthe inhibition of cell proliferatiot by 5 ttg/mi of dipyridamolewhen the cells were incubated in medium containing 2% or20% plasma . It is important to note that in both of theseexperiments, the growth medium was supplemented withplatelet-derived growth factor to achieve a similar level ofgrowth stimulation at each of the indicated serum or plasmaconcentrations used . These results suggested that the pres-ence of increasing concentration of serum attenuated thegrowth inhibitory activity of dipyridamole . Therefore, it maybe that the oral dose of dipyridamole used in the clinicaltrials may not have been adequate to affect smooth musclecell proliferation after coronary angioplasty in humans .

Antiproliferative activity of dipyridamole may be distinctfrom its antiplatelet activity. Several analogues of dipyri-damole have been described that exhibit reduced binding toserum proteins . Dipyridamole analogues SH-869 and mop-idamol have been shown to inhibit platelet aggregation andexhibit lesser binding to serum proteins (24) . We examinedthe effect of mopidamol and SH-869 on vascular smoothmuscle cell proliferation in vitro . Our results show that

JACC Vol . 23 . No . 3March 1 . 1994 :665-71

0.01

.1

1

10

100

Dipyrldamale (pglml)

Figure 3. Effect of dipyridamole analogues with antiplatelet activityon smooth muscle cell proliferation . Smooth muscle cells weregrown to confluence in 96-well microtiter plates . Cells were thentransferred to medium containing 2% platelet-poor plasma, 20 nglmlof platelet-derived growth factor, I pCi/ml of 'H-thymidine andvarious concentrations of dipyridamole (solid circles), SH-869 (opencircles) or mopidamol (triangles) . Deoxyribonucleic acid (DNA)synthesis was determined as in Figure 1 . Data shown here are meanvalue ± SD of triplicate observations . *p < 0.05 with respect todipyridamole. dpm = disintegrations per minute .

although both agents have been shown to be as effective asdipyridamole against platelet aggregation, their antiprolifer-ative activity on smooth muscle cells was greatly differentfrom that of dipyridamole . The concentrations of SH-869and mopidamol required for inhibition of DNA synthesis invascular smooth muscle cells were 10 and 100 times higher,respectively, than that of dipyridamole (Fig . 3). These re-sults suggest that the antiproliferative activity of dipyri-damole may be unrelated to its antiplatelet activity .

Effect of dipyridamole on cell proliferation in rabbit arter-ies after balloon injury. To test our hypothesis that bypass-ing serum binding by local delivery may allow dipyridamoleto inhibit smooth muscle cell proliferation in arteries, wedetermined the effect of dipyridamole in a vascular injurymodel of restenosis . Carotid or femoral artery injury wasproduced with an embolectomy balloon catheter. As de-scribed under Methods, this procedure involving endothelialdenudation and stretch produced intimal thickening that washighly uniform from rabbit to rabbit . Initially, the effect ofdipyridamole on cell proliferation after injury to rabbitcarotid arteries was determined by measuring DNA synthe-sis ex vivo. The method for determining DNA synthesis inwhole arteries is similar to that described previously (24) .The ratio of DNA synthesis, as determined by 3H-thymidineincorporation, in the injured and uninjured control arterieswas used as an indicator of stimulation of cell proliferation inresponse to injury. In uninjured arteries where cell prolifer-

StNGH ET AL .

669MPYMDAMOLE AND SMOOTH MUSCLE CELL PROLIFERATION

180

cc 160

140

19 120PM

X o80-Go .

2 40'

coXW 200 -_

Control 0 .6 6 60 600Dipyridamole (gg/day)

Figure 4 . Dose-dependent inhibition of "H-thymidine incorporationby adventitially delivered dipyridamole in balloon-injured rabbitcarotid arteries . Arteries from control and dipyridamole-treatedrabbits were excised 72 h after balloon injury . Cell proliferationactivity in injured and noninjured arteries from each rabbit wasdetermined by - H-thymidine incorporation in deoxyribonucleic acid(DNA). The ratio of 'H-thymidine incorporation (weight-normalizeddisintegrations per minute) in injured and uninjured arteries, ameasure of stimulation of DNA synthesis in response to ballooniniv, y, was determined for control and dipyridamole-treated rabbitslees Methods) . The results shown on the vertical axis are expressedas 'H-thymidine incorporation in dipyridamole-treated comparedwith control rabbits . The control bar represents 21 rabbits, and eachdipyridamole bar represents the mean value ± SEM for three toeight rabbits . *p < 0 .05 versus control .

ation is known to be low, the DNA synthesis activity wasalso very low. DNA synthesis was stimulated 6- to 10-foldwhen measured at 72 h after balloon injury . To determine theeffect of dipyridamole on cell proliferation in the vessel wall,dipyridamole or vehicle was delivered locally at the site ofinjury by placing the tip of the osmotic pump catheter at theadventitial site . Dipyridamole was delivered at the indicatedrates . The level of cell proliferation was clearly lower inarteries treated with dipyridamole than in arteries treatedwith vehicle (Fig . 4) . Inhibition of cell proliferation bydipyridamole appeared to be dose -dependent, producing asmuch as 63% inhibition at 600 µg/day. Although the ex vivo3H-thymidine uptake method measures total cell replicationactivity in the arterial segment, immunostaining of injuredarteries with smooth musc% ; cell actin-specific antibodyHHF-35 and macrophage-specific antibody RAM-1l showedthat in this low 6o.,4estero! diet model the media and intimawere composed primarily of smooth muscle cells . Therefore,a majority of cell replication activity shown here is due tosmooth muscle cells and perhaps a minor component frommesenchymal cells in the adventitia . These results suggestedthat treatment with dipyridamole reduced the number ofreplicating cells in the vessel wall, resulting in a reducedlevel of DNA synthesis .

A dose of 300 ;kg/day of dipyridamole was selected todetermine the effect on intimal thickening in rabbit arteriesafter balloon injury . Adventitial dipyridamole delivery wasinitiated 4 days before injury and continued until 14 daysafter injury, at which time the rabbits were perfusion fixed

SINCH ET AL670 •DIPYRIDAMOLE AND SMOOTH MUSCLE CELL PROLIFERATION

March 1, 1994:665-71

4-

0.18

0.15

•

0.12

0.03

0.00Vehicle Dipyridamole

(300 pgiday)

F

S. Inhibition of intimal thickening in rabbit carotid artery byadventitial dipyridamole 14 days after balloon injury . Intimal cross-sectional area was determined for three sections per artery collectedat 5-mm intervals (see Methods for details) . The three sections wereaveraged to provide a single value for each rabbit . Each barrepresents the mean value t SENT for 17 rabbits . °p < 0 .05 varsusvehicle .

and processed for histologic and morphometric studies andplanimetry to determine intimal area. Figure 5 shows thattreatment of injured arteries with dipyridamole produced a19.6% inhibition of intimal thickening . This effect on intimalresponse to injury was small but statistically significant (p0.037). In this model, long-term intravenous heparin admin-istration at a maximal tolerated dose (23 mg/day) inhibitedintimal thickening by 29% (Dubs GP, unpublished observa-tions). Therefore, a 20% inhibition by low levels of dipyri-damole (300 pzg/day) constitutes a substantial effect bycomparison. These results demonstrating inhibition of cellproliferation and intimal thickening after balloon injurysupport the hypothesis that systemic administration of di-pyridamole may be ineffective in reducing restenosis, per-haps because of inadequate plasma free drug concentrationand binding to serum proteins . However, local application ofdipyridamole at the site of injury may prove efficacious .

Discussion11 of vascular snWh muscle cell proliferation by

dip . Our studies have shown that dipyridamoleinhibited proliferation of vascular smooth muscle cells, asevidenced by dose-dependent inhibition of DNA synthesisand proliferation of exponentially growing cells . In long-termcultures, the concentration of dipyridamole required toproduce 50% inhibition of cell growth was -7 frg/ml. Wehave also shown that inhibition of smooth muscle cell growthwas not due to cytotoxicity of dipyridamole . Dipyridamole-treated cells exhibited normal morphology in cultures . Bio-chemical measurements of cell viability using release ofintracellular lactate dehydrogenase as an indicator of celldeath also showed absence of toxicity during the experimen-tal period . Furthermore, our studies showed that the effectof dipyridamole was reversed on washing the cultures andrestimulating with platelet-derived growth factor (Singh JP,

JACC Vol . 23 . No. 3

0•

.09

unpublished observations) . Dipyridamole also inhibitedDNA synthesis in smooth muscle cells stimulated withseveral of the known smooth muscle cell mitogens, includingplatelet-derived growth fibroblast factor and epidermalgrowth factor (data not shown). Because multiple growthfactors may be involved in smooth muscle proliferation inresponse to vascular injury in vivo, these properties ofdipyridamole may have important advantages in relation toits therapeutic potential .

Serum attenuates the antip iferative activity of dipyri-damole. Previous studies have shown that alpha-acid glyco-protein present in serum effectively binds dipyridamole .After intravenous administration of dipyridamole, as muchas 99% of the dipyridamole in serum remains bound to serumproteins (25) . We have shown that the presence of serum inculture medium reduced the antiproliferative activity ofdipyridamole. From the pha acokinetic studies in humansand animals, the estimated serum levels of dipyridamoleachieved at the dose regimen used in the coronary angio-plasty trials are expected to be 0 .5 to 1 .0 ,ug1ml. Because99% of dipyridamole is bound to serum proteins, the freedipyridamole concentration could therefore be as low as0.005 to 0.01 fcglml . These concentrations are 5 to 1,times lower than that required for inhibition of smoothmuscle cell proliferation in vitro . On the basis of thesecalculations, we hypothesized that a dipyridamole concen-tration sufficient to inhibit smooth muscle cell proliferationafter vascular injury may not be achieved by systemicdelivery. However, administration of dipyridamole at thesite of balloon injury may produce local concentrations highenough to inhibit smooth muscle cell proliferation .

Inhibition of intimal thickening by locally deliver dipyr-idamole. The results presented here show that local deliveryof dipyridamole at the periadventitial site significantly inhib-ited smooth muscle cell proliferation and intimal thickeningin response to balloon injury . At the maximal dose ofdipyridamole used in these studies, the total cell replicationactivity in the vessel wall was inhibited by 63%®. Intimalthickening determined 2 weeks after balloon injury wasreduced by 20%. The dipyridamole dose, delivery vehicleand delivery system used in this series of experiments wereprobably suboptimal and may have resulted in less efficacythan would have been manifested with a more favorabledelivery system . Because dipyridamole is extremely insolu-ble in biocompatible vehicles, attempts to obtain solutionswith higher drug concentrations have not been successful .Further studies with more suitable vehicles or deliverysystems may provide greater inhibition of intimal thickening .The delivery system (osmotic pump) and the vehicle used inthese studies may not be suitable in the clinical setting .However, these experiments have provided evidence for anin vivo antiproliferative activity of dipyridamole . In theclinical setting, sustained delivery of dipyridamole may beachieved by intraluminal infusion of drug microparticles ordrug incorporated into an implantable biodegradable stent .Recent advances in the development of new catheter-based

JACC V& 23, No . 3March 1, I%T&351

DIPYRIDAMOLE AND SMOOTH MUSCLE CELL PROLIFERATION

delivery systems and vascular stents have shown promisefor their potential use in local delivery of therapeutic agents .In the case of coronary artery bypass grafting, a biodegrad-able sheet carrying the drug may be wrapped around the graftto release the drug at the local site over a desired period .

Dipyridamole has several interesting characteristics thatmake it an attractive drug for local pharmacotherapy aftercoronary angioplasty. Its pharmacologic activities (e.g .,vasodilator, antiplatelet and antiproliferative) are all favor-able for vessel wall biology in preventing resteqosis aftercoronary angioplasty . Dipyridamole is a Food and DrugAdministration-approved drug and is generally consideredsafe for use in humans, even at large doses . It is fluorescent,which allows detection and quantification in tissues forpharmacologic local delivery studies and determination ofrelease kinetics from sustained-release systems without theuse of radioisotopic labeling . Its physicochemical properties,especially solubility characteristics in aqueous versus or-ganic phases, are suitable for preparation of microparticlesuspensions or incorporation into stents . In recent studies,we successfully prepared a preparation of dipyridamoleparticles with a uniform size of 5 to 7 um . A stable micro-particle preparation in biocompatible surfactants was ob-tained for direct intraluminal infusion with a perfusion bal-loon catheter . These and other future studies will allowexamination of potential applications of dipyridamole as alocal therapy for the treatment of restenosis .

keferences1 . Clowes AW, Reidy MA, Clowes MM. Kinetics of cellular proliferation

after arterial injury . Lab Invest 1983,83 :327-33 .2 . Hanke H, Strohschneider T, Oberhoff M, Betz E . Karsch KR. Time

course of smooth muscle cell proliferation in the intima and media ofarteries following experimental angioplasty . Circ Res 1990 :67 :651-9 .

3 . Ip JH, Fuster V, Badimon J, Taubman MB, Chesebro JH . Syndromes ofaccelerated atherosclerosis : role of vascular injury and smooth musclecell proliferation . J Am Coll Cardiol 1990 ;15 :1667-87.

4. Austin GE, Ratcliff NB, Hollman J, Tabei S, Phillips DF . Intimalproliferation of smooth muscle cells as an explanation for recurrentcoronary artery stenosis after percutaneous transluminal angioplasty . JAm Coll Cardiol 1989 :6 :369-75 .

5. Nobuyoshi M, Kimura T, Ohishi H, et al . Restenosis after percutaneoustransluminal coronary angioplasty : pathologic observation in 20 patients .J Am Coll Cardiol 1991 ;17 :433-9 .

6 . Bernardo S, Fuster V, Israel DH, et al. Platelet inhibitor agents incardiovascular disease: an update . J Am Coll Cardiol 1989-,14:813-36.

SINGH ET AL

671

7. Steel PM, Chesebro JH . Stanson AW, et al . Balloon angioplasty : naturalhistory of pathophysiological response to injury in a pig model . Circ Res1985:57 :105-12 .

8 . Essed CE . Van denBrand M, Becker A . Transluminal coronary angio-lplasty and early restenosis . Br Heart J 1983 :49 :393-6 .

9 . Schwartz L . Bourassa MA, Lesperance J, et al . Aspirin and dipyridamolein the prevention of restenosis after PTCA . N Engl J hued 1988 ;318 :1714-9 .

10 . White CW, Knudson M, Schmidt D, et al . Neither ficlopidifte noraspirin-dipyridamole prevent restenosis post PICA : results from a ran-domized placebo-controlled multicenter trial [abstract] . Circulation1987 ;76 Suppl IVAV-213,

11 . Schwartz L, Lesperance J, Bourassa MG, et al . The role of antiplateletagents in modifying the extent of restenosis following percutaneoustransluminal coronary angioplasty . Am Heart J 1990 ;119:232-6 .

12 . Chesebro JH, Fuster V . Elveback LIZ, et al . Effect of dipyridamole andaspirin on late vein-graft patency artery coronary bypass operations . NEngl J Med 1984 :310 :209-14 .

13 . Goldman S . Copeiand J, Moritz T, et al . Saphenous vein graft patencyyear after coronary artery bypass surgery and effects of antiplatelettherapy . Circulation 1989 ;80 :1190-7 .

14 . Harker LA . Bernstein E, Dilley R, et al . Failure of aspirin plus dipyri-dam0e to prevent reslenosis after carotid endarterectomy . Ann InternMed 1992 :116:731-6 .

15 . Emmons PR, Harrison MJG, Honour A J . e t al . Effect of dipyridamolederivative on thrombus formation in the rabbit . Nature 1965 ;108 :255-7,

16 . Sullivan JM, Harken DE, Gorlin R . Pharmacologic control of thrombo-embolic complications of cardiac valve replacement . N Engl J Med

17 . Faxon DP, Sanborn TA, Haudenschild CC . Ryan TJ . Effect of antiplatelettherapy on restenosis after experimental angioplasty . Am J Cardiol1984,53 :72C-6C .

18 . Harker L, Ross R, Slichter J, Scott CR . Homocystein-induced arterio-sclerosis . J Clin Invest 1976 ;58 :731-41 .

19 . Rajah SM . Crow MJ, Penny AF, Ahnad R, Watson DA . The effect ofdipyridamole on platelet function: correlation with blood levels in man . BrJ Clin Pharmacol 1977 ;4129-33 .

20. Mahoiny C . Wolfram KM, Cocchetto DM, Bjornsson TD. Dipyridai . -iolckinetics . Clin Pharmacol Titer 1982 :31 :330-8 .

21 . Rosenfild J . Buchanan MR. Reilly PA, Turpie AGG . Dipyridamoledisposition after chronic administration : effect of aspirin . Thromb Res1983-,Suppi IV :137-42 .

22 . Bonin PD, Fici GJ, Singh JP. Interleukin-1 promotes proliferation ofvascular smooth muscle cells in coordination with PDGF or a monocytederived growth factor (MDGF) . Exp Cell Res 1989 ;181 :474-82,

23 . Singh JP, Bonin PD . Inhibition of proliferation of fibroblasts by lazaroids(21-aminosteroids) . Life Sci 1991 -,49 :2653-6 .

24. Asotra S, Foegh M, Conte JV, Cai BR, Ramwell PW. Inhibition of3H-thymidine incorporation after balloon angioplasty. Transplant Proc1989 ;21 :3695-6 .

25 . Niewiarowski S, Lukasiewicz H, Nath N, Sha AT . Inhibition of humanplatelet aggregation by dipyridamole and two related compounds and itsmodification by acid glycoproteins of human plasma . J Lab Clin Med1975 :86 :64-76.