Eighteen juvenile farm pigs were studied. Coronary angioplasty was performed in all three coronary arteries of each animal. After PTCA, each coronary artery in each pig was randomized to receive either local delivery of μg BE , vehicle alone or PTCA only. Twelve animals were euthanized at 28 days for morphometric analysis, and four animals were euthanized at seven days for immunohistochemical analysis of vascular smooth muscle cell (SMC) proliferative activity. Two animals died a few days after PTCA and were excluded.

Restenosis is currently the major limitation of percutaneous transluminal coronary angioplasty (PTCA), as seen in up to 30% to 40% of patients. Neointimal proliferation is a fundamental mechanism of restenosis. Although brachy­therapy has been reported to show beneficial effects (1,2), no effective therapy exists yet for neointimal proliferation. Vascular smooth muscle cell (SMC) migration and proliferation occur as early as 36 h after arterial injury (3). In cell culture assays, 17-beta-estradiol (BE) inhibited the migration and proliferation of rat vascular SMCs (4,5). Similar effects have also been shown with human vascular SMCs from the saphenous vein (6). Prolonged systemic administration of estrogen has been shown to inhibit intimal hyperplasia in animal studies (7,8). In the present experiment, we tested the hypothesis that local administration of BE during PTCA could effectively inhibit neointimal proliferation.

From the Department of Medicine, Montreal Heart Institute and University of Montreal, Montreal, Quebec, Canada
Manuscript received October 26,1999; revised manuscript received May 17, 2000, accepted July 12, 2000.

METHODS

Animal preparation. Eighteen juvenile farm pigs (nine sexually immature females with intact ovaries and nine castrated males), approximately three months old and weighing 20 to 25 kg, were studied. The males were castrated at birth. The animals were fed regular pig chow in which the bulk of protein was derived from a nonsoybean source, as soybean contains significant amounts of phytoestrogens, which could potentially influence the effect of BE. The study was approved by and conducted in accordance with the guidelines of the Animal Care and Ethical Research Committee of the Montreal Heart Institute. Before the procedure, animals were given 650 mg of acetylsalicylic acid and 30 mg of nifedipine orally, premedicated with an intramuscular injection of 6 mg/kg body weight of a mixture of tiletamine hydrochloride and zolazepam hydrochloride and given 0.5 mg of atropine. The invasive procedure was performed under general anesthesia with a mixture of isoflurane (1% to 1.5%) and oxygen­enriched air. The right femoral artery was cannulated percutaneously, and an 8F arterial sheath was introduced. After arterial access had been obtained, 100 mg of lidocaine and 250 U/kg of heparin were administered intra-arterially through the sheath. The activated coagulation time was maintained at >300 s throughout the procedure.

Angioplasty and local delivery. Standard PTCA equipment was used. An 8F right Amplatz guiding catheter and right Judkin's guiding catheter were used for coagulation of the left and right coronary arteries, respectively. Heart rate, blood pressure and the electrocardiogram were monitored throughout the experiment. Coronary angioplasty was performed with a balloon size chosen to correspond to a balloon/artery ratio of 1.1:1.2. Three 30-s inflations at 10 arm were performed with a 30-s interval between inflations. All three coronary arteries of each animal were subjected to PTCA. After PTCA, each animal's coronary artery was randomized to receive either 600 ug BE locally, vehicle alone locally or PTCA only. The chemicals BE and its vehicle 2-hydroxypropyl-beta-cydodextrin (HPCD) were purchased from Sigma Chemicals. The InfusaSleeve catheter (LocalMed, Inc., Palo Alto, California) was used for local delivery (9). Five milliliters of the designated substance was delivered at a driving pressure of 10 arm and support balloon pressure of 6 atm. The dose of BE (~30 ug/kg body weight) was chosen as representative of a daily dose of systemic BE therapy used in previous studies, which ranges from 10 to 100 ug/kg.

Of the 18 animals, two died a few days after PTCA and were excluded; thus, 16 animals were analyzed. Twelve animals were euthanized at 28 days and four at 7 days. After premedication and anesthesia, the right internal jugular vein and common carotid artery were cannulated. After cross­clamping of the descending thoracic aorta through a left lateral thoracotomy, exsanguination was performed, with simultaneous administration of 1 liter of 0.9% sodium chloride solution. The heart was perfusion-fixed in vivo with 2 liter of 10% buffered formalin at 200 mm Hg, removed from the animal and placed in 10% buffered formalin solution. The site of PTCA was identified in relation to adjacent side branches, which served as landmarks. The injured segment was harvested with a 1-cm normal segment proximal and distal to the injured site. Serial sections 3 to 5 mm long were made from the harvested segment, with a minimum of at least three sections (maximum of 5) from each PTCA site. The sections were stored in buffered 10% formalin and dehydrated with increasing concentrations of alcohol, followed by treatment with xylene and paraffin. Each section was cut into 6-um-thick slices with a microtome (Olympus cut 4060 E) and stained with Verhoeff's stain for morphometric analysis.

Morphometric analysis. Measurements were made with a video microscope linked to a 486 personal computer and customized software. A minimum of three sections for each injured segment were analyzed, and the results were averaged. When individual values were widely discrepant for an injured segment (observed in only one injured vessel), all sections demonstrating rupture of the internal elastic lamina (IEL) were included for analysis. Analyses were performed by an examiner who had no knowledge of the treatment assignment of the injured segments. The areas of external elastic lamina (EEL), IEL and lumen were measured by digital planimetry, neointimal area (IEL - lumen area) and medial area (EEL - IEL area) were obtained. The fracture length ( FL) was measured as the length of discontinuity of the injured IEL, and measurement of the extent of injury was controlled by standardization for each artery, expressed in relation to the circumference of the IEL (10,11). The extent of injury was determined by the ratio of  FL to IEL circumference ( FL/IEL). The percent neointima was de­fined as the percent total vessel area occupied by the neointima ([intimal area/EEL] X 100). Morphologic per­cent stenosis was calculated as 100 (1 - lumen/IEL area) (10). The restenotic index was defined as [intimal area/ (intimal area + medial area)]/( FL/IEL circumference) (11). Injury score was determined as previously defined (12).

Immunohistochemistry. After slicing with a microtome and blocking of nonspecific antibodies, the sections were treated with mouse anti-proliferating cell nuclear antigen (PCNA) antibodies and diluted biotinylated goat antimouse antibodies. They were incubated with avidin-biotin (Elite ABC Kit, Vector Laboratories, Burlingame, California), developed with 3,3'-diaminobenzidine (Vector Laboratories) and counterstained with hematoxylin. Porcine liver cells served as the positive control. For each section, a 6-um slice counterstained with hematoxylin without treatment with the primary antibody (mouse anti-PCNA) served as the negative control.

Immunohistochemistry was performed on samples from animals euthanized at seven days. The percentage of proliferating SMCs was obtained by dividing the number of PCNA-positive SMCs by the total number of SMCs in each field; separate measurements were made for neointimal and medial layers. To standardize comparison among treatment groups, measurements were obtained at four fixed locations separated by 90° for each section, and the results were averaged. For each segment, two sections demonstrating a maximal neointimal response were analyzed, and the results averaged.

Statistical analysis. Values are expressed as the mean value ± SEM. Kruskal-Wallis analysis was used for com­parison of data among the three groups. Multiple comparisons between the BE and PTCA-only groups and between the vehicle-only and PTCA-only groups were made using the Student-Newman-Keuls test. Chi-square analysis was used for comparison of proportions. Values were considered significant at p < 0.05.

Characteristics BE PTCA Only Vehicle Alone p Value  Segnents analyzed (n) 12 9 10 NS  Artery size (mm) 2.86 ± 0.1 2.94 ± 0.08 2.94 ± 0.13 NS  Balloon/artery ratio 1.22 ± 0.03 1.2 ± 0.02 1.17 ± 0.03 NS  EELref/EELinj 1.01 ± 0.05 1.16 ± 0.11 1.31 ± 0.15 NS  Fracture length (mm) 1.64 ± 0.27 1.71 ± 0.25 1.81 ± 0.25 NS  Extent of injury 0.25 ± 0.04 0.25 ± 0.04 0.25 ± 0.04 NS  Neointimal area (mm2) 0.40 ± 0.09 0.88 ± 0.20 1.14 ± 0.33 <0.05  Neointimal area/fracture                  length (mm2) 0.25 ± 0.04 0.5 ± 0.07 0.58 ± 0.13 <0.005  Perecnt neointima 12.16 ± 2.57 23.02 ± 3.97 25.46 ± 4.73 <0.025  Neointima/media area 0.59 ± 0.14 1.67 ± 0.43 1.75 ± 0.41 <0.01  Percent stenosis 15.67 ± 3.22 27.51 ± 4.39 30.34 ± 5.40 <0.025  Restenotic index 1.3 ± 0.14 2.4 ± 0.23 2.42 ± 0.22 <0.005  Injury score 1.64 ± 0.1 1.7 ± 0.14 1.77 ± 0.15 NS  EELref = proximal reference segment external elastic limina area; EELinj = injured segment external  elastic lamina area (averaged). Data are presented as the mean value ± SEM.     BE = 17-beta-estradiol; NS = not significant

RESULTS

After PTCA and local delivery, the animals were allowed to recover and gained weight steadily. Two animals died 48 and 72 h, respectively, after the procedure and were excluded; thus, 16 animals were studied. Autopsy of the two animals revealed occlusive thrombus at the site of PTCA (in the BE-treated vessel in one pig and in the vessel treated with PTCA only in the other pig).

No changes in heart rate, electrocardiographic variables or blood pressure were observed during local delivery.

Injured segments. The balloon/artery ratio and artery diameter were similar among the three treatment groups (Table 1). Segments with intact IEL in which discernible injury was absent were excluded from analysis (two from the PTCA-only group and one from the vehicle-only group). Two segments were lost during harvesting (one from the vehicle-only group and one from the PTCA-only group).

Morphometric analysis. On morphometric analysis at 28 days, arterial segments treated with local delivery of BE showed significantly less neointimal hyperplasia compared with the other two treatment groups (Fig. 1). This beneficial effect was noted in all of the variables of neointimal response to injury analyzed (Table 1). The extent of injury was similar among the three groups, suggesting that the use of the InfusaSleeve catheter was not associated with enhanced risk of injury. The noninjured segments proximal and distal to the sites of PTCA appeared morphologically normal, with no evidence of inflammation or necrosis.

To exclude the existence of an inhibitory effect on intimal proliferation due to the vehicle, analyses were performed comparing segments treated with vehicle alone and PTCA only, which showed no significant differences. In contrast, significantly less intimal hyperplasia was observed in BE treated segments as compared with segments treated with PTCA only (Fig. 2). Compared with PTCA only or vehicle alone, BE decreased neointimal formation by ~50%.

The existence of a significant protective effect of BE as compared with placebo or vehicle treatment by gender was then assessed (Table 2 and 3). Although the extent of injury in the BE-treated segments appeared slightly greater in males and slightly less in females, compared with the other two treatment groups, they were not statistically significant. A significant protective effect of BE on the neointimal response to injury was demonstrated in both genders.

There were no significant differences in the effect of BE in the different arteries treated (left anterior descending vs. circumflex vs. right coronary artery). Neointimal area (0.28 ± 0.16 vs. 0.47 ± 0.16 vs. 0.46 ± 0.15 mm², p = NS), neointimal area/ FL (0.17 ± 0.04 vs. 0.35 ± 0.11 vs. 0.25 ± 0.06, p = NS) and restenotic index (1.2 ± 0.14 vs. 1.72 ± 0.27 vs. 1.12 ± 0.25, p = NS) were similar in the different arteries treated with BE. Immunohistochemistry. A statistically significant decrease in proliferating SMCs (PCNA-positive SMCs) was seen in the arterial segments treated with BE. Among the different groups, the percentage of PCNA-positive SMCs in the neointima were 0.43 ± 0.26% with BE, 4.26 ± 1.17% with PTCA only and 4.27 ± 1.37% with vehicle alone (p < 0.05 for BE vs. other two groups). There were no significant differences in the percentage of PCNA-positive SMCs in the media between the three groups: 0.4 ± 0.15%, 1.38 ± 0.87% and 1.24 ± 0.79% for BE, PTCA only and vehicle alone, respectively (p = NS).

Vascular remodeling. To determine the effect on vascular remodeling of the agents used, the EEL areas of the injured segment and of the normal vessel proximal to the PTCA site were obtained, and their ratio calculated (10). No significant difference between the groups was noted: 1.01 ± 0.05, 1.16 ± 0.11 and 1.31 ± 0.15, respectively, for BE, PTCA only and vehicle alone (p = NS).

DISCUSSION

The present study demonstrates, for the first time, to our knowledge, that locally delivered BE decreases neointimal proliferation after PTCA in pigs. The study also shows that
local delivery of BE can be performed safely using the InfusaSleeve catheter, without significant additional injury.

Characteristics BE PTCA Only Vehicle Alone p Value    Extent of injury 0.32 ± 0.05 0.26 ± 0.04 0.26 ± 0.04 NS  Neointimal area (mm2) 0.51 ± 0.13 1.01 ± 0.29 0.99 ± 0.23 NS  Neointimal area/fracture                  length (mm2) 0.25 ± 0.06 0.52 ± 0.09 0.5 ± 0.08 <0.025  Neointima/media area 0.78 ± 0.21 2.05 ± 0.69 1.65 ± 0.4 0.05  Restenotic index 1.24 ± 0.21 1.84 ± 0.27 1.92 ± 0.31 <0.025   Data are presented as the mean value ± SEM.    Abbreviations as in Table 1.

Characteristics BE PTCA Only Vehicle Alone p Value    Extent of injury 0.14 ± 0.01 0.24 ± 0.08 0.21 ± 0.07 NS  Neointimal area (mm2) 0.25 ± 0.07 0.73 ± 0.3 1.51 ± 1.05 NS  Neointimal area/fracture                  length (mm2) 0.25 ± 0.13 0.46 ± 0.11 0.77 ± 0.41 <0.05  Neointima/media area 0.33 ± 0.07 1.23 ± 0.43 1.99 ± 1.16 NS  Restenotic index 1.37 ± 0.2 2.75 ± 0.38 2.93 ± 0.43 <0.05   Data are presented as the mean value ± SEM.    Abbreviations as in Table 1.

Local delivery and injury response. Systemic administration of pharmacologic agents may not achieve a local concentration of the agent to produce a significant effect; in addition, higher doses may result in intolerance or toxicity. Unlike systemic administration, local delivery involves the delivery of a single dose of the agent during PTCA. Therefore, local delivery appears to be a potentially effective method for the prevention of tissue response to injury. Several agents have shown beneficial results in reducing neointimal proliferation after PTCA in animal experiments. Locally delivered ethanol (13), nitric oxide (10), vascular endothelial growth factor (14) and antisense oligonudeotides (15) have resulted in decreased neointimal proliferation after arterial injury. Two clinical trials have demonstrated that local radiation therapy may have the potential to reduce restenosis after PTCA (1,2).

17-Beta-estradiol and injury response. Previous experiments in animals have demonstrated that estrogen administered subcutaneously for up to three weeks inhibited the myointimal response to arterial injury (7,8). Recently, a preliminary report has suggested that short-term subcutaneous estrogen therapy (6 to 17 days) may be effective in reducing the injury response (16). Estrogen administered intramuscularly for at least three weeks has also demonstrated the potential to inhibit vascular SMC proliferation and neointimal hyperplasia (17). However, the efficacy of local delivery of BE to inhibit intimal hyperplasia has not been previously studied.

In the present study, local delivery of BE was associated with a markedly lower proliferatioe response compared with the other two treatment groups. The neointimal area, neointima/media area ratio, restenosic index and morphologic percent stenosis were all significantly lower in the BE-treated arterial segments. The decreased neointimal hyperplasia after treatment with BE is probably due to the inhibition of SMC proliferation. Assessment at different time points could possibly provide additional information on the effect of BE on SMC proliferation. Studies examining SMC proliferation after arterial injury have shown that maximal SMC proliferatioe activity occurs during the first seven days after experimental arterial injury and rapidly decreases beyond seven days (18). We were able to demonstrate significantly lower SMC proliferation at seven days in the arterial segments treated with BE, compared with the PTCA-only and vehicle-only treatment groups. The inhibitory effect of BE on SMC proliferation may be due to a direct effect on SMC proliferation (19) or to enhanced synthesis of nitric oxide by BE (20,21). Nitric oxide is known to inhibit both SMC migration (22) and proliferation (23). The biologic effects of estrogen, like other steroid hormones, involve intracellular receptors (24). Estrogen receptors have been demonstrated in coronary arteries obtained from autopsy specimens in both premenopausal and postmenopausal women (25) and in cell cultures of human saphenous vein and internal mammary artery specimens (26). The beneficial effects of BE-the predominant circulating estrogen in premenopausal women-on vascular injury response may not be replicated by other kinds of estrogens. For example, conjugated equine estrogen was found to have no effect on neointimal proliferation in nonhuman atherosclerotic primate models (27). In this study, it is unlikely that loss of estrogen receptors in the hypercholesterolemic monkeys could be responsible for the lack of response to conjugated equine estrogen on neointimal proliferation, as conjugated equine estrogen treatment did result in a statistically significant decrease in atherosclerotic plaque area and inhibited the progression of atherosclerosis (with no effect on neointimal proliferation) in the same monkeys, thereby demonstrating responsiveness to estrogen. Simultaneous administration of progesterone may attenuate the vascular injury response to BE (28). A sexually dimorphic response to estrogen has been reported after arterial injury, with intact male rats deriving no benefit from estrogen therapy (29). However, this sexually dimorphic effect was not observed in another experiment with gonadectomized rats (8). In the present study, a significant effect of BE on the neointimal proliferatioe response was noted in both genders, although the male animals used in the study had been castrated at birth.

Vehicle. 17-Beta-estradiol is a lipophilic compound with poor solubility in aqueous solutions; therefore, it needs a vehicle for parenteral administration. 2-Hydroxypropylbeta-cyclodextrin, a starch derivative, has been successfully tested as an effective excipient for protein drugs (30). The pharmacokinetics of HPCD are similar to that of insulin, and the toxic dose (nephrotoxicity) has been estimated to be 200 mg/kg in rats (31). The dose of HPCD used to dissolve BE in the present study was 0.63 mg/kg, far below the toxic dose. Furthermore, HPCD has been used for administration of ophthalmic preparations and intravenous anesthetic agents in humans (32,33). In the present study, the neointimal response to injury was similar in HPCD-treated arteries and arteries undergoing PTCA only. Therefore, HPCD itself did not exert any significant influence on neointimal proliferation.

Clinical evidence. Retrospective studies in humans have shown no benefit of hormonal replacement therapy on angiographic restenosis after PTCA (34), although one study did show a beneficial effect after directional atherectomy (35). However, conjugated estrogen (and not BE) was the predominant form of estrogen used in many of these patients, and no information on concomitant use of progesterone is available. In addition, it is not known whether hormone replacement therapy could achieve a sufficient concentration of estrogen locally at the PTCA site to exert a significant response.

Study limitations. It is possible that spillover of BE into the systemic circulation could occur during local delivery. Also, tissue uptake and retention of BE after local delivery were not measured. However, the results of the morphometric and immunohistochemical analyses dearly demonstrate a significant difference in the arteries treated with BE compared with untreated vessels, which is highly suggestive of an effect due to the locally delivered BE. The efficacy of the InfusaSleeve catheter in delivering various agents intramurally has been reported in previous studies. At a support balloon pressure similar to that used in the present study, a sixfold higher concentration in the arterial wall was achieved by local delivery, as compared with the arterial wall concentration achieved by systemic administration (36). In another study, persistent retention in the arterial wall of locally delivered tissue factor pathway inhibitor was demonstrated for at least 48 h after delivery (37). A strong argument against a systemic effect due to spillover is that, in the present model, where all three treatment arms were applied in each animal, the systemic effects of BE (if any) due to spillover would be expected to affect all three coronary arteries in each animal equally, providing equal protection to all vessels. The differences in the results between the arteries treated with BE compared with PTCA only and vehicle alone cannot therefore be attributed to such systemic effects. Furthermore, BE that enters the circulation is rapidly eliminated, mostly by the liver (38).

Conclusions. We have shown that a single dose of BE delivered locally during PTCA in a porcine model has the potential to decrease neointimal hyperplasia significantly, probably by inhibition of SMC proliferation. Previous studies have demonstrated that treatment with BE is associated with inhibition of proliferation of human vascular SMCs in cell culture assays (6). The local delivery of BE can be performed safely with the InfusaSleeve catheter, without risk of additional injury. The local administration of BE is therefore a promising new approach that may be useful in preventing the proliferative response after PTCA. Its usefulness in preventing restenosis after PTCA merits further investigation.

Acknowledgments
The authors are indebted to Pascale Geoffroy, MSc, and Julie Lebel for assistance in the laboratory; to Martin G. Sirois, PhD, Dominique Lauzier and Veronique Philibert for assistance with immunohistochemistry and morphometry; and to Stanley Nattel, MD, FACC, for his valuable suggestions in the manuscript preparation.