Objective: Tyrphostin AGL-2043 is a potent tricyclic quinoxaline inhibitor of PDGF β-receptor tyrosine kinase (PTK), Kit, and Flt3. We have shown previously that selective inhibition of PDGF β-receptor PTK by tyrphostins markedly reduces SMC proliferation and migration in vitro, reduces neointima formation in balloon-injured porcine femoral arteries, and reduces neointimal stenosis in stented porcine coronary arteries when administered intramurally within biodegradable nanoparticles. The present study was designed to determine the effect of AGL-2043 delivered from a stent-based, biodegradable polymeric coating on neointima formation in the porcine coronary artery model. Methods and Results: Stents coated with biodegradable, polylactic/glycolic acid (PLGA) polymer, with (n=13) or without (n=11) 180 mcg AGL-2043 were implanted into the proximal LAD of 24 Sinclair mini-pigs (34±4 kg) to achieve a 1.1:1 stent/artery diameter ratio. The delivery of drug from stent to tissue was confirmed by high-performance liquid chromatography. After 28 days, histomorphometric analysis showed that in-stent stenosis in animals treated with AGL-2043 was reduced by 50% (51±21% versus 26±10%, p=0.001), the absolute neointimal area was reduced by 44% (2.38±1.04 versus 1.31±0.43 mm2, p=0.004), and the absolute luminal area was increased by 57% (2.19±1.09 versus 3.39±0.59 mm2, p=0.003). There were no significant differences between control and AGL-2043 in injury score (1.24±0.11 vs. 1.15±0.12, p=0.07) or inflammation score (1.19±0.35 vs. 1.07±0.33, p=0.41). Moreover, the difference in % in-stent stenosis between control and treated animals remained highly significant even after normalizing the % stenosis to the degree of injury (p=0.0008) or to the inflammation score (p=0.001). Mortality for this study was zero. Tissue concentration in segments 1 cm proximal and distal to the stents, were negligible or zero at 1 h, 24 h, and 4 weeks after stent implantation. Conclusion: Stent-based delivery of tyrphostin AGL-2043 from a biodegradable polymeric coating reduces in-stent neointimal hyperplasia in porcine coronary arteries by 50% after 28 days and preserves lumen area. Long-term studies should be the next step in testing applicability to the human interventional setting.
Platelet-derived growth factor (PDGF), expressed by platelets, smooth muscle cells (SMCs), and macrophages, has been shown to play an important role in the pathogenesis of injury-induced neointimal formation. PDGF acts as both a mitogen and chemoattractant for SMCs, and mediates the transformation of SMCs from their contractile to proliferative/synthetic phenotype [1–4]. Moreover, the degree of neointimal formation has been shown to depend substantially on both PDGF β-receptor overexpression and its activation by PDGF-BB [5–8].
The tyrphostins are low molecular weight, synthetic compounds whose basic structure can be modified to block specific receptor or intracellular protein tyrosine kinases (PTKs) [9–11]. The profound selective PTK inhibition of these compounds results from competitive interaction with the ATP-binding domain as well as mixed competitive inhibition with substrate-binding subsites [12,13]. Unlike the larger anti-receptor antibodies, the relatively low MW, hydrophobic tyrphostins can more easily gain access to receptor sites deep within tissues such as the arterial media.
Tyrphostin AGL-2043 is a potent, synthetic, tricyclic quinoxaline inhibitor of several PTKs, namely, PDGF β-receptor, Kit, and Flt3. In a series of in vitro experiments, we have shown that AGL-2043 and/or the related tyrphostin, AG-1295, selectively inhibit PDGF-BB-induced β-receptor phosphorylation, selectively inhibit SMC proliferation in culture, and attenuate the outgrowth of SMCs from porcine and human arterial explant tissue . In vivo studies in pigs by our group, using local drug delivery, have shown that AG-1295 reduces neointima formation in isolated balloon-injured femoral arteries by ≈50% when administered intraluminally within polylactic acid-based nanoparticles (90 nm) . We also have shown that AG-1295 reduces neointimal stenosis in the balloon-injured rat carotid artery model when delivered from intraluminal nanoparticles or perivascular polymeric matrices [16,17]. Neointima formation in pig coronary arteries was reduced by ≈25% when AGL-2043-containing polylactide-based biodegradable nanoparticles were delivered intramurally via an oozing, needle balloon prior to stent implantation . However, this model appeared to be associated with marked additional wall damage and hence a less than optimal result.
The current study was, therefore, designed to determine the effect of tyrphostin AGL-2043 eluted from a stent-based, biodegradable, polymeric coating on in-stent neointimal growth in porcine coronary arteries.
2.1. Drug-eluting stents
Genius, Megaflex, biodegradable polymer coated stents (EuroCor, Bonn, Germany) were implanted into the LAD of 24 Sinclair mini-pigs (34±4 kg). The biodegradable, polylactic/glycolic acid (PLGA) polymer was impregnated with 180 μg AGL-2043 in 13 pigs. The required dose of tyrphostin was based on previously reported pharmacokinetic studies performed in vitro and in vivo [16–18]. Control animals (n=11) received coated stents without AGL-2043.
All animals were treated with clopidogrel (75 mg/day) and aspirin (100 mg/day) beginning 7 days before stent implantation. Anesthesia was induced with ketamine (20 mg/kg IM) and droperidol (0.2 mg/kg IM). After endotracheal intubation, anesthesia was maintained by a mixture of isofluorane (1–3%), oxygen, and room air. Following systemic heparinization (1 ml, 5000 IU), diagnostic coronary angiography was performed, via the right carotid artery, using a Siemens, Siremobil 245B angiography system (Munich, Germany), and images digitally stored using a computerized image analysis system (Medcon, Israel). The stents (3.0 mm diameter; 15 mm length, 10 atm, for 20 s) were implanted into the proximal LAD with the precise positioning determined by predeployment angiographic measurements to achieve a 1.1:1 stent/artery diameter ratio. After 28 days, under general anesthesia and mechanical ventilation, a left thoracotomy was performed (4th interspace with removal of ribs 4 and 5), the pericardium incised, and the heart and origin of the great vessels exposed. After heparinization (5000 IU), the heart was arrested by rapid IV injection of KCl (40 meq/l). The heart was then excised and the coronary arteries perfused ex vivo (retrograde via the ascending aorta) with normal saline (500 ml with 1 ml heparin [5000 IU], 90 mm Hg) followed by pressure perfusion with 4% buffered formaldehyde (500 ml, 90 mm Hg).
The delivery of AGL-2043 from stent to tissue was confirmed using a modified high-performance liquid chromatography (HPLC) assay  in two additional animals in which stents were implanted in both LAD and RCA (1 h, 2 stents; 24 h, 2 stents). For this analysis, the arrested hearts were excised and the coronary arteries perfused with normal saline to clear the blood. Segments of the arteries (1 cm long) were excised immediately (1 mm) distal and proximal to the stented segment. The stented segment was then excised, cut longitudinally, and the stent wires gently removed with jewelers' forceps. The arterial segments were then washed with saline prior to placement in tubes for analysis.
Blood levels and tissue concentrations (lung, liver, kidney, non-stented RCA, and the segment adjacent to the stented LAD) were determined in 4 of the 24 pigs sacrificed at 4 weeks for histomorphometry (3 treated, 1 control).
2.3. Tissue preparation
After perfusion, the entire heart was post-fixed by immersion in 4% buffered formaldehyde. Segments of the LAD were excised, dehydrated in ethyl alcohol, and embedded in methyl-methacrylate. The site of stent implantation was cut throughout into 2-mm-thick disc segments (6–7 per animal) with the aid of an Isomet 1000 precision rotating diamond saw (Buehler, Lake Bluff, IL). Each 2 mm segment was then screened morphometrically using bright field transillumination microscopy (Fig. 1 top).
All 2-mm segments were then mounted individually on stubs, sectioned (5 μ), deplasticized with 2-butanone, and rehydrated with decreasing concentrations of alcohol and distilled water for staining by the Movat pentachrome technique  and with hematoxylin and eosin (H&E) for correlative histomorphometric and histopathologic analysis (Fig. 1 bottom).
2.4. Histomorphometric analysis
One treated animal was excluded for technical reasons. Computerized morphometric screening was performed first on the 2-mm-thick methyl-methacrylate discs (from 18 of the 23 animals [control: 8 pigs, 49 discs; AGL-2043: 10 pigs, 61 discs]) prior to conventional microtomy. Percent in-stent stenosis was calculated as ([in-stent area−lumen area/in-stent area] × 100).
Morphometric measurements and calculations performed on the 5-μ-thick Movat-stained histologic sections (23 pigs) (control: 11 pigs, 65 sections; AGL 2043: 12 pigs, 68 sections) included vessel size (area bounded by the external elastic lamina [EEL]), lumen area, neointima area (area bounded by the internal elastic lamina [IEL]−area occupied by strut intimal strut profiles−lumen area), and percent in-stent stenosis ([in-stent area−lumen area/in-stent area] × 100).
Morphometric parameters (for discs and histological sections) are reported in two ways: (1) the mean across all discs or sections from each animal, and (2) the most narrowed disc or section from each animal.
2.5. Injury and inflammation scores
Degree of injury was estimated in the Movat-stained sections in accordance with the method of Schwartz et al.  where 0=IEL intact; 1=IEL lacerated, media compressed; 2=IEL lacerated, media lacerated, EEL compressed; and 3=IEL lacerated, media lacerated, EEL lacerated, strut in adventitia. The injury scores for each strut were averaged to obtain the mean score for each of the 133 histological sections from the 23 animals. Percent in-stent stenosis also was normalized to injury score to adjust for any differences in degree of injury between control and treated groups.
An inflammation score (value 0 to 3), modified from Schwartz et al.  was assigned to adventitia, media, neointima, and each stent strut site, as well as an overall assessment for each section of all pigs. For this score, 0=none to minimal inflammatory cell infiltrate; 1=mild; 2=moderate; and 3=severe. This analysis was performed on H&E-stained sections.
2.6. Statistical analysis
Data are expressed as mean±S.D. Comparisons of histomorphometric and histopathological parameters between control and treated animals were performed by one-way ANOVA. Correlations were performed by univariate regression analysis with probabilities calculated from ANOVA. The Statview II statistical package (Brain Power) was used for these calculations.
This investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996.
Mortality for this study was zero.
Quantitative coronary angiography performed before and after stent implantation in the LAD established the post-stent-to-pre-stent (stent-to-artery) diameter ratio as 1.093±0.064 (range: 1.027–1.281) for all 23 animals. There was no significant difference in stent-to-artery ratios between control (1.073±0.036 [n=11]) and AGL-2043 treated groups (1.111±0.079 [n=12])(p=0.16). Mean vessel sizes by histomorphometry (as measured by area bounded by the EEL and area bounded by the IEL) were also similar between the groups (EEL: Control, 5.479±0.631 mm2 vs. AGL-2043, 5.832±0.396 mm2, p=0.12; IEL: Control, 4.716±0.57 mm2 vs. AGL-2043, 4.831±0.402 mm2, p=0.58).
3.1. Histomorphometry of 2 mm discs
Initial morphometric screening of the 110, 2 mm thick methyl-methacrylate discs (Figs. 1 top and Fig. 2) showed the % in-stent stenosis in arteries treated with AGL-2043-eluting stents to be significantly lower than in those receiving the coated stents without the drug (16±7% [n=61 discs, 10 pigs] vs. 43±21% [n=49 discs, 8 pigs], p=0.001 by one-way ANOVA). Restricting the analysis to the single, most narrowed segment of each vessel also showed a significant reduction (61%) in stenosis in the treated group versus the controls (22±10% vs. 57±22%, p=0.0005). The reliability of this screening method was supported by univariate regression analysis, which showed a highly significant correlation between % in-stent stenosis determined by morphometry of 2 mm thick discs and % in-stent stenosis from the conventional 5 μ Movat stained sections of these discs (y=0.819x−3.594, r=0.96, p=0.0001).
3.2. Histomorphometry of 5-μm sections
Table 1 shows the morphometric data obtained from the 133 conventional 5-μ-thick histological sections (2 mm intervals) from all 23 animals. As with the initial screening method, the mean percent in-stent stenosis of arteries treated with AGL-2043-eluting stents was significantly less than for arteries receiving coated stents without the drug (50% reduction, by average of all 5 μ sections; 45%, by most-narrowed 5 μ section only) (Fig. 3). Moreover, arteries with AGL-2043-eluting stents had significantly reduced absolute neointima area and significantly increased absolute luminal area, whether by average of all sections or by the most-narrowed section of each artery alone.
|CONTROL n=65 Discs (11 Pigs)||AGL 2043 n=68 Discs (12 Pigs)||P|
|NEOINTIMA AREA (mm2)|
|LUMEN AREA (mm2)|
|% IN-STENT STENOSIS|
|Most narrowed||65±20||36±13 (45%)||0.0006|
|CONTROL n=65 Discs (11 Pigs)||AGL 2043 n=68 Discs (12 Pigs)||P|
|NEOINTIMA AREA (mm2)|
|LUMEN AREA (mm2)|
|% IN-STENT STENOSIS|
|Most narrowed||65±20||36±13 (45%)||0.0006|
Mean (±S.D.) percent in-stent stenosis, absolute area occupied by neointima, and absolute luminal cross-sectional areas in control and AGL-2043-treated arteries presented in two ways: as the average of all 5-μ sections at 2 mm intervals through the entire stented segment of all 23 animals (Average) and as the most-narrowed section of each stented artery (Most Narrowed). P-values were calculated from one-way ANOVA. The percent reduction in in-stent stenosis in the treated animals is shown in parentheses.
3.3. Injury score
The injury score (0–3), assessed strut-by-strut across all 5 μ Movat-stained histological sections of all 23 animals, did not differ significantly between control animals and those receiving AGL-2043-eluting stents (1.239±0.106 vs. 1.146±0.12, p=0.07). When % in-stent stenosis was normalized to the injury score of each animal (mean % in-stent stenosis/mean injury score) the highly significant reduction in % in-stent stenosis in animals receiving the AGL-2043-eluting stent was preserved (control, 41±15% vs. AGL-2043, 22±7%, p=0.0008), pointing to a true effect of this treatment.
3.4. Inflammation score
Inflammatory cell infiltrates were primarily mononuclear cells consisting of lymphocytes, plasma cells, and monocytes. Table 2 summarizes the section-by-section (H&E), strut-by-strut analysis of inflammatory cells for each artery. Although occasional foci of moderate-to-severe extent were found in these vessels, the overall pattern was one of mild inflammatory cell infiltration with mild infiltration in the adventitia, minimal in the media, diffusely mild in the neointima favoring the abluminal portions, and minimal-to-mild strut-associated inflammation. No significant differences were detected between control animals and those receiving AGL-2043-eluting stents. Additional studies using specific immunohistochemical markers will be required to determine whether any differences exist in specific cell types between the control and treated groups. Nonetheless, as with the injury score, normalization of % in-stent stenosis to the overall inflammation score of each animal (mean % in-stenosis/mean inflammation score) showed preservation of the marked statistical difference between control and treated arteries (p=0.001).
Each number represents the mean (±S.D.) inflammation score (0–3 [see Methods]) of all 5-μ histological sections in control and AGL-2043-treated arteries. Scores estimate the extent of inflammatory cells in the adventitia, media, neointima, and around the struts (average score of all struts of each section), as well as an overall score for the entire section. P-values were calculated from one-way ANOVA.
3.5. Drug levels
Tissue concentrations of AGL-2043 immediately adjacent to the stents were 48 and 212 ng/mg at 1 h (RCA and LAD, respectively) and 196 and 89 ng/mg at 24 h; whereas the tissue concentrations of all segments excised 1 cm proximal and 1 cm distal to the stents were negligible (0.8–4.1 ng/mg). Blood levels were zero in all samples withdrawn before, and 4 weeks after stent implantation. Tissue concentrations of AGL-2043 at 4 weeks were zero in the kidney, lung, liver, non-stented RCA, and LAD adjacent to the sites of stent implantation, further emphasizing the effectiveness of this local delivery device.
This study demonstrates that selective inhibition of PDGF β-receptor PTK by tyrphostin AGL-2043, eluted from a stent-based biodegradable polymeric coating, inhibits in-stent neointima formation and preserves lumen area in the porcine coronary model.
Initial histomorphometric screening of the 2-mm-thick methyl-methacrylate discs showed a reduction of 63% (average of all sections) or 61% (most narrowed section) in in-stent stenosis in arteries treated with AGL-2043-eluting stents versus control arteries, whereas subsequent analysis of the 5-μ-thick histological sections (2 mm intervals) showed 50% and 45% reductions, respectively. Nonetheless, regression analysis showed a highly significant correlation between % in-stent stenosis derived from morphometry of the 2-mm-thick discs and that obtained from the conventional 5 μ Movat-stained histological sections discs (r=0.96, p=0.0001), and the effect of the AGL-2043 on % in-stent stenosis was highly significant as assessed by both techniques. The advantages of the discs cut with a rotary diamond saw are speed and excellent preservation of stent-to-tissue architecture, permitting rapid evaluation of % in-stent stenosis with minimal distortion. However, resolution is poor and reliable evaluation of the extent of injury and of tissue and cell pathology could be achieved only with the 5-μ-thick histological sections despite the distortion from the microtome knife and processing for staining. Thus, comprehensive evaluation of these stented vessels appears to require both techniques.
Recent successes in early clinical trials with drug-eluting stents (DES) using the antiproliferative agents Sirolimus (rapamycin) and paclitaxel (Taxol) have been quite promising [21–23]. Indeed, this has contributed directly to the rapid expansion in the world market for effective drug-eluting stents. Rapamycin inhibits smooth muscle cell proliferation by blocking the cell cycle in the G1 (growth) phase when RNA is produced and proteins are synthesized. In so doing, it prevents the cell from entering the S (DNA synthesis) phase, thus preventing replication. The specific target of rapamycin is target of rapamycin (TOR), considered to be the central protein kinase element of the downstream signaling pathway controlling mRNA translation and cell growth [24–26]. Paclitaxel is the potent, natural prototype of the taxane family of anticancer agents. It works by binding to polymerized tubulin, thereby stabilizing it against disassembly, thus inhibiting cell mitosis (M-phase of cell cycle).
Long-term follow-up concerning the local effects of these drugs on the various cellular and matrix elements of the arterial wall is still incomplete, as is information concerning the fate of the non-biodegradable polymer, e.g., with respect to its integrity, likelihood of fragmentation, and interaction with surrounding tissue and blood components . Accumulating evidence suggests the possibility of development of drug resistance to rapamycin through genetic mutations or compensatory changes in mTOR-regulated proteins . Most recently, there have been reports, still being evaluated, of sub-acute thrombosis occurring within the first 30 days after implantation of the Cypher (Rapamycin) stent and of hypersensitivity reactions including pain, rash, respiratory alterations, hives, itching, fever, and blood pressure changes . Likewise, overexpression of the multidrug resistance gene (MDR-1), molecular changes in the target protein β-tubulin, changes in mitosis checkpoint proteins, changes in apoptotic regulatory proteins, and overexpression of cytokines such as IL-6, have been suggested as possible mediators of resistance to Taxol [29,30]. However, the relevance of such resistances to single, or multiple, localized doses of such compounds at the site of stent implantation remains uncertain.
Thus, as with antimicrobial therapy, it is likely that effective DES treatment will require a varied armamentarium in order to cope with developing changes in devices and materials, untoward tissue responses and drug resistances, as well as a variety of specific vascular requirements, such as stenosis of vein grafts and coronary aneurysm.
Tyrphostin AGL-2043 is a potent, synthetic, tricyclic quinoxaline inhibitor of PDGF β-receptor protein tyrosine kinase, as well as of Kit and Flt3 . The targeted inhibition of PDGF-BB-induced β-receptor phosphorylation causes selective and potent inhibition of SMC migration and proliferation as well as mitogenesis, with only a minor inhibitory effect on endothelial cell growth [14,15], thereby minimizing interference with vessel wall healing. Drugs that interfere early in the cell cycle (early G1 or pre-G1) are considered to be cytostatic and elicit less cellular necrosis and inflammation than agents that affect the cell cycle at a later stage . Thus, blocking SMC proliferation by upstream inhibition of PDGF receptor-mediated signal transduction in the manner shown by AGL-2043 may provide an alternative with low potential toxicity. Although further studies are necessary to determine whether products of degradation of the biodegradable polymer might trigger cellular responses with long-term effects, effective uneventful dissolution of such polymers might represent an important advantage over the potential long-term consequences of disruption of the integrity of non-biodegradable stent coatings.
To date, we have shown that selective inhibition of PDGF β-receptor protein tyrosine kinase (PTK) by the tyrphostin family of PTK blockers markedly reduces PDGF-BB-induced β-receptor phosphorylation, selectively inhibits SMC proliferation in culture with little effect on endothelial cells, attenuates outgrowth of SMCs from porcine and human arterial explant tissue, reduces neointima formation after intraluminal injection in balloon-injured porcine femoral arteries, and reduces neointimal stenosis in stented porcine coronary arteries when administered intramurally within biodegradable nanoparticles. We now report that tyrphostin AGL-2043, eluted from a biodegradable polymeric coating, markedly reduces in-stent stenosis in porcine coronary arteries after 28 days. Although further, longer term studies are necessary to rule out possible “catch-up” in the treated animals, and to determine the full pharmacokinetic profile including dose–response relationships as well as efficacy in the presence of various pre- and/or post-treatment regimens, these studies, taken together with our previous in vitro and in vivo local delivery studies with this class of compounds [9–11,14–18], appear to provide strong evidence in support of further testing for application to the human interventional setting.
SDG is the Chutick Professor of Cardiac Studies, GG is a member of David R. Bloom center of Pharmacy, and AL is the Wolfson Family Professor of Biochemistry, The Hebrew University.
US Food and Drug Administration, FDA Talk Paper, 2003; T03-71.