Toxic Vessel Reaction to an Absorbable Polymer-Based Paclitaxel-Eluting Stent in Pig Coronary Arteries

Drug-eluting stents (DES) have significantly reduced restenosis. One of the most intensely studied agents considered and currently in clinical use for DES-based local delivery is paclitaxel. Efficacy of paclitaxel-loaded, biostable polymer-coated stents with polymer-based sustained-release, low- and moderate-dose drug delivery for reducing in-stent neointima has been shown in several animal models, including rat, rabbit and pig.^1–5 These and other seminal animal studies have demonstrated the dose-dependent response of paclitaxel as a locally delivered antiproliferative agent. Toxic effects, including medial necrosis and resultant stent malapposition in pig coronary arteries, were seen at very high doses (12.5 µg/mm stent length) using a nonpolymeric, dip-coated formulation.⁶ Paclitaxel is a potent drug considered to have a narrow toxic-to-therapeutic window when locally delivered.⁷ However, there is a wide range of total drug dosages currently under investigation for clinical use, suggesting a lack of consensus about the optimal dosing and elution profiles for safe but effective inhibition of restenosis. The wide-scale acceptance of DES throughout the world, following the promising results of animal studies and clinical trials, has lead to generally excellent results in the real world outside the trial environment. However, serious adverse events have been reported, including subacute and late thrombosis, aneurysm formation and hypersensitivity reactions following deployment of DES in human beings.^8,9 Hypersensitivity reactions could be caused by the polymer, drug, stent metal or the interaction between these elements.^10,11 Such observations have contributed to ongoing research and development for improved devices. With concerns about delayed healing and late stent thrombosis, there is a trend towards reducing drug dosage and polymer coating thickness. Permanent, biostable polymers may result in chronic heightened inflammation and hypersensitivity reactions; the use of erodible, biocompatible polymers is expected to minimize these risks by ultimately leaving only the metal stent backbone in the vessel wall. Our objective in this study was to investigate a new, second-generation drug-eluting stent comprising an erodible polymer and moderate-dose paclitaxel on a thin-strut cobalt chromium stent platform in a clinically relevant porcine coronary artery model. Methods Sixteen juvenile domestic pigs (37–63 kg) were enrolled in this study. Experimental procedures were performed without notable difficulties according to the recommendations of the consensus advisory panel to the FDA (U.S. Food and Drug Administration).¹² Pigs received stents according to a random assignment constrained by individual coronary anatomy. A total of 45 stents of 3 different types: bare-metal stents (BM), polymer-only coated stents (POLY), and polymer-based paclitaxel-eluting stents (PACL) were implanted in 16 pigs. Thin-strut cobalt chromium of 3.0 mm diameter and 11 mm length was the platform for all stents. Animals received 325 mg aspirin and 150 mg clopidogrel 1 day before implant and daily until termination. Cardiac catheterization was performed with full heparinization (200 units/kg) and stents were implanted using IVUS to obtain a stent-to-artery ratio of ª1.1:1. Post-implant angiography and IVUS were performed to determine stent size, complete apposition and patency. Stent-to-artery ratios by quantitative coronary analysis (QCA) were similar among the groups: 1.09 ± 0.05, 1.09 ± 0.05, and 1.08 ± 0.07 for BM, POLY and PACL, respectively. Three pigs were sacrificed after 1 week, 9 pigs after 1 month, and 3 pigs after 3 months; 1 pig died prematurely (see details in the Results section). At explantation, coronary arteries were perfusion-rinsed with phosphate-buffered saline (PBS) and perfusion-fixed in situ at physiologic pressure with 5% formalin/1.25% glutaraldehyde for 15–20 minutes. Stented vessels were trimmed free from the heart and embedded in methyl methacrylate. Sections from proximal, middle and distal vessel regions were cut using a heavy-duty microtome, collected on glass slides, deplasticized and stained with hematoxylin-eosin and Movat-pentachrome, and analyzed using a computer-assisted image analysis system. Polymer/drug loading. The biodegradable polymer used on the drug-eluting stent system was PLGA (polylactide coglycolide), a copolymer of polyglycolide (PGA) and polylactide (PLA) which are aliphatic polyesters of the poly (a-hydroxy acids), glycolic and lactic acid. PLGA is hydrophilic and undergoes bulk erosion as water enters the polymer bulk where hydrolytic degradation occurs. Bulk erosion results in polymer surfaces with multiple concavities (Figure 1). Stents were coated by dissolution of PLGA and paclitaxel (final concentration of 10% paclitaxel; drug was GMP grade, Fragron Inc., purity 99.4%). The polymer or polymer/drug mixture was sprayed onto the stents and dried to constant mass. Paclitaxel content of the stent was determined by gravimetric analysis and verified by dissolving the coating off the stents and measuring the paclitaxel content by high-performance liquid chromatography. The in vitro elution of PLGA is known to be slower than in vivo elution attributed to the presence of biological compounds, especially lipids and enzymes.¹³ In vitro release kinetics were determined on expanded stents immersed in 2 ml aliquots of agitated PBS at 37ºC for 50 days. Buffer was changed every 2–3 days and the paclitaxel content determined by ultraviolet spectrophotometry. The biodegradable polymer had a 3–5 µm coating thickness, and the DES contained a 9:1 polymer-to-paclitaxel ratio for a total dose density of 0.3–0.35 µg/mm², or 1.75–2.11 µg/mm of stent length. Histopathology. Sections from each of the stent segments were scored for inflammation, necrosis, intramural hemorrhage and mural thrombus. The nature of the extracellular matrix in the neointima and adventitia was assessed and described. An overall inflammation score was assigned according to the following semiquantitative grading scale: 0 = none; 1 = mild, scattered inflammatory cells; 2 = moderate, inflammatory cells encompassing 14 Necrosis of the tunica media was scored according to a similar grading scale: 0 = none; 1 = mild, focal transmural or nontransmural regions of medial smooth muscle cell (SMC) necrosis involving any portion of the artery; 2 = moderate, transmural medial SMC necrosis involving > 25% of the circumference of the artery; 3 = severe, transmural medial SMC necrosis with involvement of > 50% of the circumference of the artery. A scoring scale was used to grade the extent of vessel injury determined by stent strut penetration into the vessel wall as follows: 0 = strut not in contact with internal elastic lamina (IEL); 1 = strut contacting IEL with profile in neointima; 2 = strut penetrating IEL and profile in media; 3 = strut penetrating media and contacting external elastic lamina (EEL); 4 = strut in adventitia.¹⁵Histomorphometry. Sections were imaged by 20x instrument magnification. Morphometric analysis was performed by computerized planimetry on all levels of all stents. Lumen, IEL and EEL were traced and area measurements obtained; areas of neointima and media were obtained by subtraction. Neointimal thickness at each stent strut was measured. Histologic percent area stenosis [1 - (luminal area/IEL area) x 100] was calculated. We also measured the “gap width”, which was defined as the distance between the strut and IEL, as an index of stent malapposition. Statistical analysis. Descriptive statistics were generated for all quantitative data and expressed as mean ± SEM. Histomorphometric data were evaluated by two-way analysis of variance (ANOVA) in which independent variables were stent type and restudy time. All-pairwise multiple comparisons using the Tukey method were performed if ANOVA probability was Results A single premature death occurred immediately after stent implantation related to anesthesia complications, not to stent reaction per se. Therefore, histopathologic and histomorphometric measurements and analyses were performed for 42 stents from 15 animals; Table 1 illustrates the distribution of stents according to stent type and restudy time. In vitro drug release pharmacokinetics. The release profile indicated a mild initial burst (12.3 ± 0.6% paclitaxel released after 4 days) as the paclitaxel on the coating surface diffused into the surrounding environment. This was followed by a period of slow release (17.3 ± 3.1% paclitaxel released after 15 days), and a subsequent increase in release rate (60.7 ± 4.0% paclitaxel released after 30 days) (Figure 2). In vitro drug release was monitored until no significant change in release rate was observed. Histopathology and Histomorphometry Qualitative observations. At 1 week there was minimal neointimal formation for all stent types. The tunica media was compressed and thinned at the site of stent strut contact and there was a thin mural thrombus frequently present for all three groups. There was, however, a marked transmural medial necrosis and medial hemorrhage present in all three sections (proximal, middle and distal) of two PACL samples (Figure 3); there were no samples showing medial necrosis and hemorrhage in the BM or POLY groups. Inflammatory cell infiltrates at the luminal surface and in the thin mural thrombus, as well as in the adventitia and perivascular space consisting primarily of polymorphonuclear leukocytes (neutrophils) and monocyte-macrophages, were seen in all groups. For 1-month specimens, divergent microscopic morphologies were seen between the BM and POLY on the one hand, and the PACL on the other. Bare-metal and polymer-only showed the appearance typical of stented porcine coronaries, including: concentric or slightly eccentric fibrocellular neointima formation with proteoglycan-rich matrix in the deep neointima, and collagenous matrix in the more adluminal region; confluent squamous periluminal cell layer; and leukocyte involvement in the form of both moderate numbers of foreign-body giant cells adjacent to the stent struts, as well as occasional diffuse or focal histiolymphocytic infiltrates with eosinophils in the adventitia and perivascular space (Figure 4). In these stent types, stent struts compressed the tunica media without evidence of necrosis or hemorrhage. In marked contrast, arteries implanted with the paclitaxel-coated stents showed medial necrosis, separation of the tunica media from the stents, poorly developed neointima that consisted of inspissated thrombus and amorphous proteinaceous material with some fibrocellular organization and abundant inflammatory infiltrates, incomplete endothelialization, and inflammatory cells in the media, adventitia and perivascular space (Figure 4C). For 3-month samples, all groups showed a well-healed appearance; concentric fibrocellular neointima with well-organized layers of circumferentially-oriented spindle-shaped cells in a collagen-rich extracellular matrix was present (Figure 5). For PACL, this morphology was present on the adluminal aspect of the intima, whereas in the regions close to the stent, a loose-appearing, proteoglycan-rich matrix was observed. The periluminal cell layer consisted of confluent, flattened endothelial or endothelial-like cells, and there was little inflammation except for occasional foreign-body giant cells adjacent to stent strut profiles, and rare focal or diffuse histiolymphocytic infiltrates containing occasional eosinophils. Inflammation was notably more prevalent in the PACL group. Neointima of arteries implanted with PACL was notably thicker than the other groups. In some vessels implanted with BM, particularly in the proximal sections of more markedly tapering vessels, there was separation of the stent from the media, with the struts located eccentrically in the lumen on a “pedestal” of mature fibrous neointimal tissue connecting the struts with the vessel wall. Measured and scored variables. Histopathologic and histomorphometric descriptors of vessel responses were found to vary according to stent type and restudy time, substantiating and quantifying the morphological qualitative observations. These descriptors included gap width, intimal thickness, histological % stenosis, necrosis score, inflammation score and strut injury score. Illustrative micrographs of vessel responses are shown in Figure 6. Gap width (distance between strut and IEL). At 1 week, there was no effect of stent type on the width of the gap between the stent struts and IEL; the mean gap width in the paclitaxel group was 0.02 mm ± 0.01 vs. bare-metal 0.01 mm ± 0.01 (p = 0.964; Figure 7). The polymer-only coated stent had a gap width of 0.01 mm. At 1 month, gap width varied according to stent type; PACL showed greater gap width compared to the BM and POLY groups (0.22 mm ± 0.02 vs. 0.03 mm ± 0.02 and 0.02 mm 0.01, respectively; p Intimal thickness. At 1 week, there was no effect of stent type on intimal thickness. The mean intimal thickness in the BM group was 0.04 mm ± 0.01 vs. 0.02 mm ± 0.01 in the PACL group (p = 0.987; Figure 8). The polymer-only coated stent had an intimal thickness of 0.03 mm. At 1 month, intimal thickness tended to vary by stent type (0.15 mm ± 0.02 vs. 0.16 mm ± 0.03 and 0.06 mm ± 0.01 for BM, POLY and PACL groups, respectively; p = 0.08). There was no difference in intimal thickness between the BM and POLY groups at 1 month (p = 0.980). At 3 months the PACL group showed greater intimal thickness compared to the BM group (0.48 mm ± 0.14 vs. 0.07 mm ± 0.02; p Histological % area stenosis. At 1 week and 1 month post-stent implantation, there was no effect of stent type on the area of neointimal stenosis between the BM and PACL groups, 10% ± 3 vs. 9% ± 1 (p = 0.983; Figure 9). The polymer-only coated stent had a histological stenosis of 12%. There was also no effect at 1 month (25% ± 3 vs. 30 ± 8% and 23% ± 2, for BM, POLY and PACL, respectively; p = 0.908). However, at 3 months, histological % stenosis varied significantly according to stent type, with the highest level of 59% ± 11 for the PACL group vs. 17% ± 2 for the BM group; p Discussion Animal models play an instrumental role in developing and improving stent technology, a role that continues to evolve in the era of localized drug delivery. Safety is a principal concern for any stent technology, and the porcine coronary stent model appears useful in its assessment. The critical failure mode for stents includes acute, subacute or late closure because stent thrombosis nearly always has catastrophic implications. Long-term neointimal stimulation is a secondary concern for stent safety, especially with DES. Toxicity induced by high local drug concentration remains an ongoing concern, and significant arterial pathologies have been demonstrated in animals with high-dose local delivery of paclitaxel. The combination of stent design, polymer type and coating thickness, total drug dose and the time-course of elution are all important factors in determining the tissue dosing over the active or therapeutic period of the DES. Although the clinical trial data for the two flagship drug-eluting stents, Taxus™ (Boston Scientific Corp., Natick, Massachusetts) and Cypher^® (Cordis Corp., Miami, Florida) show favorable results for limiting restenosis,^16–21 adverse vascular pathologies including delayed healing, non-reendothelialization and rebound neointimal stimulation, as well as vessel toxicities such as medial necrosis, stent malapposition and chronic severe inflammation, may still be of major concern. Excessive inflammation related to polymer composition and the consequent adverse effect of neointimal stimulation may not manifest early because proliferation is suppressed by the therapeutic drug compound. Furthermore, toxic effects of the eluted drug itself may not appear within standard clinical trial endpoints. In this study, we evaluated a new DES comprising a bioabsorbable polymer eluting a moderate paclitaxel dose (1.75–2.11 µg/mm) on a thin-strut, cobalt chromium stent platform. In vitro elution studies of this formulation over 8 weeks indicated that only 60% of the paclitaxel load was eluted by that time; total polymer degradation was found to take 10–12 weeks. However, in vivo studies (results not shown) on polymer-only coated stents showed complete polymer degradation in 4–6 weeks, suggesting a faster drug elution time in this setting. The PACL group showed a consistent and profound vascular toxicity compared to tissue responses of the bare-metal and polymer-coated controls. The significant reduction in neointimal thickness in the PACL group compared to the BM group at 1 month is consistent with the inhibitory effect of paclitaxel on stent-induced smooth muscle cell (SMC) proliferation seen in other studies. However, there was a concomitant severe necrosis in the tunica media, leading to dilation of the vessel wall away from the stent and thrombus deposition in the resultant gap subjacent to the stent. These observations point to a troublesome toxic response that was not observed with polymer-only and bare-metal stents, therefore, it appears attributable to the presence of paclitaxel. Our 3-month results illustrate that PACL arteries underwent significant rebound neointimal thickening compared to bare-metal controls. The amorphous parastrut material that filled the space between the stent circumference and dilated IEL at 1 month disappeared and was replaced by an aggressive proliferation of proteoglycan-rich in-stent neointima. Although neointimal formation at the interim and standard evaluative time point of 1 month was successfully suppressed by local drug delivery, this temporary suppression appeared unsustainable. Since the slow-release modulation of low-dose paclitaxel from Taxus stents has not been reported to elicit toxicities such as those observed in this study, it seems likely that the rapid release of drug from the stent is responsible for these phenomena (despite the moderate dose of drug compared to other stents currently in use or under clinical investigation). Similar medial necrosis, stent malapposition and chronic neointimal stimulation have also been documented previously for a burst-release paclitaxel dip-coated stent at very high doses.⁶ The data obtained from the present study provide additional evidence for a systematic set of coronary biological reactions to those DES which display analogous vessel wall toxicities. In a previous investigation, actinomycin-D was tested as an antirestenotic agent in the same animal model. A strikingly similar pattern of medial necrosis, stent malapposition, excessive inflammation and long-term rebound neointimal stimulation was observed.^22,23 It is of interest to note that the preclinical evaluations of sirolimus-eluting stents did not show early vessel toxicities such as medial necrosis and stent malapposition, but did demonstrate rebound neointimal formation.^24,25 In that case neointima was not excessively stimulated, but was equivalent to bare-metal at 3 months, whereas neointimal suppression was observed at 1 month. These observations point to a common and fundamental drive towards post-stenting vessel wall homeostasis, with an underlying pattern of chronic fibroproliferation of the neointima when the rate of healing has been suppressed early with potent antiproliferative drugs. Vascular toxicity compounds this effect and exacerbates fibrocellular proliferation. Stent design^26–28 and type and thickness of polymer are likely important cofactors affecting the tissue response to the drug-eluting device. Rapid drug release in vivo seems to be a likely factor in the vessel toxicities we observed, but direct assessment of in vivo pharmacokinetics is required for verification. Further necessary development work with this stent system may include optimizing the control of the drug-release rate. This study highlights and underscores the importance of rigorous and thorough preclinical evaluations, especially in vivo evaluations of new DES. It emphasizes the necessity of careful microscopic evaluation of histopathologic samples from a range of restudy times, along with objective critical interpretation of the data. Study limitations. The small numbers of animal subjects should be considered and the statistical analysis of data viewed accordingly when interpreting these results. Animal models do not precisely simulate responses to DES in humans.²⁹ We did not perform in vivo pharmacokinetics which may have provided correlative information to our histologic observations and in vitro drug-release measurements. Acknowledgement. We would like to acknowledge the expert technical assistance of Cindy Baranowski for histologic section preparation and the contribution of Jeremy Ollernshaw, PhD for assistance with project management. Conclusions Despite in vitro data showing slow, sustained release within the range of dosages and elution rates reported for other commercially available or investigational paclitaxel-eluting stents, porcine coronary implants demonstrated a time-dependent phenomenon of vessel wall toxicity culminating in excessive neointimal formation for paclitaxel elution from a bioabsorbable polymer. This study suggests that the therapeutic window for paclitaxel may not be as broad as currently inferred, and that efforts to improve DES technology should consider early vessel wall toxicity as well as chronic neointimal suppression. Hence, thorough preclinical testing to define this window for each new device is important.

1. Virmani R, Kolodgie FD, Farb A. Drug-eluting stents: Are they really safe? Am Heart Hosp J 2004;2:85–88. 2. Schwartz RS, Chronos N, Virmani R. Preclinical restenosis models and drug-eluting stents: Still important, still much to learn. J Am Coll Cardiol 2004;44:1373–1385. 3. Drachman DE, Edelman ER, Seifert P, et al. Neointimal thickening after stent delivery of paclitaxel: Change in composition and arrest of growth over six months. J Am Coll Cardiol 2000;36:2325–2332. 4. Hong MK, Kornowski R, Bramwell O, et al. Paclitaxel-coated Gianturco-Roubin II (GR II) stents reduce neointimal hyperplasia in a porcine coronary in-stent restenosis model. Coron Artery Dis 2001;12:513–515. 5. Finkelstein A, McClean D, Kar S, et al. Local drug delivery via a coronary stent with programmable release pharmacokinetics. Circulation 2003;107:777–784. 6. Heldman AW, Cheng L, Jenkins GM, et al. Paclitaxel stent coating inhibits neointimal hyperplasia at 4 weeks in a porcine model of coronary restenosis. Circulation 2001;103:2289–2295. 7. Rogers CD. Drug-eluting stents: Clinical perspectives on drug and design differences. Rev Cardiovasc Med 2005;6(Suppl 1):S3–S12. 8. Iakovou I, Schmidt T, Bonizzoni E, et al. Incidence, predictors, and outcome of thrombosis after successful implantation of drug-eluting stents. JAMA 2005;293:2126–2130. 9. Virmani R, Guagliumi G, Farb A, et al. Localized hypersensitivity and late coronary thrombosis secondary to a sirolimus-eluting stent: Should we be cautious? Circulation 2004;109:701–705. 10. Vermani R, Farb A, Guagliumi G, Kolodgie FD. Drug-eluting stents: Caution and concerns for long term outcome. Coron Artery Dis 2004;15:313–318. 11. Farb A, Heller PF, Shroff S, et al. Pathological analysis of local delivery of paclitaxel via a polymer-coated stent. Circulation 2001;104:473–479. 12. Schwartz RS, Edelman ER, Carter A, et al. Drug-eluting stents in preclinical studies: Recommended evaluation from a consensus group. Circulation 2002;106:1867–1873. 13. Baek N, Lee J, Park K. Aqueous N, N-diethylnicotinamide solution as a medium for accelerated release study of paclitaxel. J Biomater Sci Polymer Edn 2004;15:527–542. 14. Kornowski R, Hong MK, Tio FO, et al. In-stent restenosis: Contributions of inflammatory responses and arterial injury to neointimal hyperplasia. J Am Coll Cardiol 1998;31:224–230. 15. Schwartz RS, Huber KC, Murphy JG, et al. Restenosis and the proportional neointimal response to coronary artery injury: Results in a porcine model. J Am Coll Cardiol 1992;19:267–274. 16. Sousa JE, Costa MA, Abizaid A, et al. Lack of neointimal proliferation after implantation of sirolimus-coated stents in human coronary arteries: A quantitative coronary angiography and three-dimensional intravascular ultrasound study. Circulation 2001;103:192–195. 17. Morice MC, Serruys PW, Sousa JE, et al. A randomized comparison of a sirolimus-eluting stent with a standard stent for coronary revascularization. N Engl J Med 2002;346:1773–1780. 18. Moses JW, Leon MB, Popma JJ, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315–1323. 19. Stone GW, Ellis SG, Cox DA, et al. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221–230. 20. Moses JW, Mehran R, Nikolsky E, et al. Outcomes with the paclitaxel-eluting stent in patients with acute coronary syndromes: Analysis from the TAXUS-IV trial. J Am Coll Cardiol 2005;45:1165–1171. 21. Stone GW, Ellis SG, Cox DA, et al. TAXUS-IV Investigators. One-year clinical results with the slow-release, polymer-based, paclitaxel-eluting TAXUS stent: The TAXUS-IV trial. Circulation 2004;109:1942. 22. Robinson K, Chronos N, Royal J, et al. Actinomycin-D drug-eluting stents preserve lumen size and inhibit fibrocellular neointima in pig coronary arteries. Circulation 2001;104:2396 II–506. 23. Robinson KA, Serruys P, Stack RS, Virmani R. Results and lessons learned form unsuccessful drug-eluting stent programs: Failure of the actinomycin D coating to inhibit in-stent restenosis despite potent biologic effects in pig coronary arteries. In: Stone GW, Serruys PW, Baim DS, et al (eds.) Textbook of Coronary Stenting (1st Editon) (in press). 24. Carter AJ, Aggarwal M, Kopia GA, et al. Long-term effects of polymer-based, slow-release, sirolimus-eluting stents in a porcine coronary model. Cardiovasc Res 2004;63:617–624. 25. Suzuki T, Kopia G, Hayashi S, et al. Stent-based delivery of sirolimus reduces neointimal formation in a porcine coronary model. Circulation 2001;104:1188–1193. 26. Hwang CW, David AB, Edelman ER. Physiological transport forces govern drug distribution for stent-based delivery. Circulation 2001;104:600–605. 27. Rogers CD. Optimal stent design for drug delivery. Rev Cardiovasc Med 2004;(5 Suppl )2:S9–S15. 28. Hwang CW, Wu D, Edelman ER. Impact of transport and drug properties on the local pharmacology of drug-eluting stents. Int J Cardiovasc Interv 2003;5:7–12. 29. Virmani R, Kolodgie FD, Farb A, Lafant A. Drug eluting stents: Are human and animal studies comparable? Heart 2003;89:133–138.