Calcification of Bovine Pericardium: Glutaraldehyde Versus No-React Biomodification

Calcification of Bovine Pericardium: Glutaraldehyde Versus No-React Biomodification Amir Abolhoda, MD, Sumei Yu, MS, J. Rodrigo Oyarzun, MD, John R. McCormick, MD, John D. Bogden, PhD, and Shlomo Gabbay, MD Departments of Cardiothoracic Surgery and Preventive Medicine and Community Health, New Jersey Medical School; and the Department of Applied Chemistry, New Jersey Institute of Technology, Newark, New Jersey Background. Calcific degeneration is the most frequent cause of clinical dysfunction of glutaraldehyde (GA)- pretreated bioprosthetic heart valves. The No-React (NR) process has been shown to be a promising anticalcification treatment. In this comparative study, our objective was to delineate the advantages of the NR treatment over GA. Methods. Bovine pericardial strips pretreated with GA and NR were individually incubated in calcium phosphate solution for 21 days at 37 C. The pretreated bovine pericardium then was implanted subcutaneously in rats and retrieved at 14, 21, and 35 days after-implantation. Mineral and morphologic analyses were performed on each specimen. Results. The NR-treated pericardium revealed significantly reduced in vitro calcification compared with the GA-treated tissue (mean tissue calcium content 1.3-0.2 versus 5.9-0.7 ~g/mg; p < 0.001). Mineral analysis showed progressive calcification of the GA-pretreated pericardium over the period of implantation (calcium content increasing from 49.6-9.6/~g/mg after 2 weeks to 134.3-9.1 /~g/mg at 5 weeks after-implantation). The NR-treated implants had calcified significantly less (p < 0.05) at each corresponding interval. Moreover, morphologic examinations demonstrated a protracted inflammatory response in the form of giant cell and mononuclear cell infiltration associated with intrinsic collagen disruption in the GA-treated tissue; the NR-treated pericardium maintained morphologic integrity with a mild inflammatory response. Conclusions. The NR biochemical process appears not only to attenuate pericardial calcification, but also to abort the host's destructive inflammatory response to the xenograft. (Ann Thorac Surg 1996;62:169-74) labrication of a durable bioprosthetic cardiac valve F from chemically treated xenograft tissues continues to be a challenging prospect. Despite the inherent advantages of biologic valve prostheses, such as pseudoanatomic central flow with superior hemodynamic indices and low thrombogenicity, negating the need for chronic anticoagulation therapy [1], clinical bioprosthetic valve implantation has been plagued by progressive tissue degeneration, resulting in valve dysfunction and ultimate replacement. Reoperation is an eventual outcome in approximately 20% to 30% of bioprosthetic valve recipients by the tenth postoperative year; most failures are attributed to primary tissue failure in the form of calcific degeneration [1]. The pathophysiology of calcific mineralization of bioprosthetic valves has been examined by numerous investigators over the past 2 decades, yet the exact mechanism and contributing factors have not been defined clearly. During the early 1980s, the potential role of the host's immune response in xenograft calcific degeneration was refuted by the description of equivalent chemical and morphologic characteristics in the calcification of porcine Accepted for publication March 7, 1996. Address reprint requests to Dr Gabbay, Department of Cardiothoracic Surgery, UMDNJ--New Jersey Medical School, 185 South Orange Ave, Room G-502, Newark, NJ 07103. bioprosthetic aortic valve cusps in athymic mice [2]. In addition, Levy and associates [3] assigned negligible importance to the host's inflammatory response based on their millipore di~usion chamber subcutaneous implant model simulating the intrinsic calcification that was observed in other experimental models despite the absence of attached host cells. Furthermore, although the in vitro pulse accelerator (fatigue tester) models have demonstrated stress-induced collagen deterioration and its potential role in accelerating the calcification process [4-6], the biopathologic mechanism of bioprosthetic mineralization cannot be explained solely on the basis of dynamic stress. For the past several years, the pendulum of investigations in the area of bioprosthetic calcification has swung toward exploring the seemingly fundamental role of the xenograft biochemical preparation method in inducing tissue calcification. Glutaraldehyde (GA) is currently the standard reagent for preservation and biochemical fixation of fresh bioprosthetic leaflet materials of either bovine pericardium or porcine aortic valve cusp origin. It imparts intrinsic tissue stability (biodegradation resistance) and reduces the antigenicity of the material [7]. However, multiple recent reports have suggested a detrimental role of aldehyde-induced intra- and intermolecular collagen cross-linkages in initiating tissue mineralization [8-11]. In addition, GA has been implicated in 1996 by The Society of Thoracic Surgeons 0003-4975/96/$15.00 Published by Elsevier Science Inc PII S0003-4975(96)00277-9

170 ABOLHODA ET AL Ann Thorac Surg CALCIFICATION OF BOVINE PERICARDIUM 1996;62:169-74 C o O E O (11 o O _= I- 7 [3 N0-React pretreated pericardium / GlutaraJdehyde prelreated pericardium J Saline Change No Control of Ca change solution of Ca solution Fig 1. In vitro calcification of bovine pericardium, pretreated with either conventional glutaraldehyde or No-React, after 21 days of incubation at 37 C in a physiologic calcium solution (mean tissue calcium content expressed in ~g/mg +- standard error of the mean; n = 5). Tissue incubation in saline solution over the same time period was used as a control devitalization of the intrinsic connective tissue cells of the bioprosthesis, thus resulting in breakdown of transmembrane calcium regulation and hence contributing to cellassociated calcific deposits [12], A battery of alternative modification regimens have been proposed, with variable success. Among the most notable anticalcification reagents are 2-aminooleic acid, Sodium dodecyl sulfate, and diphosphonates [13-15]. However, GA remains the final sterilization agent in all of these alternative treatment cocktails, with its inherent injurious effects on both cellular and collagenous components of the bioprosthetic tissue. Replacing GA with an alternative preservative would have profound clinical implications if the proposed biochemical modification process retained the advantages of aldehyde fixation while inhibiting or minimizing implant calcification. In a series of pilot studies conducted previously in our laboratory, we have tested bovine pericardial and porcine aortic valve cusp tissue fragments treated with a novel anticalcification process known as No-React (NR), introdticed by Biocor (Belo Horizonte, Brazil). These preliminary Studies have shown great promise of a potentially viable alternative for chemical preparation of cardiac valve bioprostheses, with comparable stress tolerance and superior cytocompatibility [16]. Although the intricacies of the process have not been disclosed fully because of commercial incentives, the treatment involves the following: (1) aldehydes cross-linkage to achieve high resistance to biodegradation, (2) a detox- ification process in solutions of natural endogenous substances with multiple physical variables, and (3) incubation with a surfactant that presumably supports in vivo nonreactivity. The objectiv~es of the present study were to characterize the beneficial biologic properties of the NR treatment and to verify its merit as an anticalcification process. We also examined the role of local foreign body reaction and the host's inflammatory response in implant degeneration. Material and Methods Static In Vitro Model Static in vitro calcification testing was conducted as described previously by Bernacca and colleagues [17]. Bovine pericardial tissue strips of 1 5 cm, pretreated with either conventional GA or NR, were placed in individual vials containing a solution of 135 mmol/l NaCl, 2.2 mmol/l CaCI 2 2H20, and 1.2 mmol/l KH 2 PO 4 in 0.05 mmol/l 3-(N-morpholino) prop'ane-sulfonic acid, buffered to ph 7.40, In a parallel control group, tissue samples were placed in a solution containing 135 mmol/l NaCl in 0.05 mmol/l 3-(N-morpholino) propane-sulfonic acid, with ph adjusted to 7.40. Solutions were filter sterilized under laminar-flow conditions. The vials then were kept in a 37 C incubator for 21 days; the calcifying solution was changed weekly under sterile conditions in half of the vials. The samples and the vials were inspected regularly for evidence of gross extrinsic calcific deposits or bacterial overgrowth. On completion of the incubation period, the tissue samples were retrieved from the vials, washed in several changes of sterile 0.9% NaC1, dried at 90 C overnight, and analyzed 150 A 125 O) E ~ 100 C 0 u 75 E "_~ u 50 0 W ~" 25 o 2 3 5 Time (weeks post-implantation) Fig 2. Comparison of trends of calcification of No-React- versus glutaraldehyde-pretreated bovine pericardium implanted subcutaneously in rats (mean tissue calcium content expressed in I~g/mg 4- standard error of the mean; n = 7).

Ann Thorac Surg ABOLHODA ET AL 171 1996;62:169-74 CALCIFICATION OF BOVINE PERICARDIUM Fig 3. Light photomicrographs of No-React (NR)- and glutaraldehyde (GA)-pretreated bovine pericardial explants after 3 weeks. (A) NR-pretreated implant showing well-preserved collagen fascicular bundles with no foreign body reaction (hematoxylin and eosin). (B) NR-pretreated implant stained with the yon Kossa technique; no calci~c deposits are seen. (C) GA-preserved explant, illustrating intense inflammatory cellular infiltration (InF-R) of the collagen structure and giant cell formation (Gc) (hematoxylin and eosin). (D) van Kossa stain of GA-pretreated tissue demonstrating destructive calcium (CA++)deposits. (A, 100; B, 100; C, 200; D, 100.) for calcium content by atomic absorption spectrophotometry. Subcutaneous Implantation in Rats Weanling Sprague-Dawley rats (SD Strain; Taconic Laboratories, German Town, NY) were used at age 4 weeks (60 to 80 g). The animals were injected intraperitoneally with two distinct anesthetic cocktails of equal efficacy; either pentobarbital (40 to 60 mg/kg) or ketamine! xylazine (ketamine 40 to 87 mg/kg; xylazine 5 to 13 mg/kg). Each animal received two strips of tissue, one each of GA- and NR-pretreated bovine pericardium, in separate subcutaneous pouches in the anterior abdominal wall. The wounds were closed with 5.0 Vicryl suture material. The rats were fed Lab Rodent Diet (Purina Meals Inc). They received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Re- sources and published by the National Institutes of Health (NIH publication 86-23, revised 1985), At 14, 21, and 35 days after-implantation, the animals were sacririced with a lethal intraperitoneal dose of thiopental (300 mg/kg), and the pericardial specimens were retrieved. A small portion of each specimen was fixed immediately in 10% neutral buffered formalin for light microscopy examination. The remainder of each sample was used for mineral analysis. Mineral Analyses Pericardial tissue was washed in sterile saline and dried to constant weight in a 90 C desiccator oven. The tissue concentration of calcium was determined using previously described techniques [18] by flame atomic absorption spectrophotometry (Perkin-Elmer Model 603; Perkin-Elmer, Norwalk, CT) after digestion with a 3:1 mixture of double-distilled 70% nitric and perchloric acids (GFS Chemicals, Columbus, OH). National Institutes of Standards and Technology bovine liver (SRM

172 ABOLHODA ET AL Ann Thorac Surg CALCIFICATION OF BOVINE PERICARDIUM 1996;62:169-74 Fig 4. Light photomicrographs of No-React (NR)- and glutaraldehyde (GA)-pretreated bovine pericardium harvested 5 weeks after subcutaneous implantation in rats. (A) NR-treated pericardium maintained intrinsic collagen architecture (hematoxylin and eosin). (B) GA-pretreated explant demonstrating persistent cellular permeation of the pericardial fibrosa accompanied by intrinsic disruption (FBR) (hematoxylin and eosin). (C) Higher magnification of slide in 4B illustrating multinuclear giant cells with associated destruction of collagen bundles (CoL-B) (hematoxylin and eosin). (D) GA-pretreated explant showing calcific deposits (CA+ +), detectable even with the hematoxylin and eosin technique. (InF-R = inflammatory cellular infiltration.) (A, x100; B, x100; C, x400; D, 100.) 1577a, Gaithersburg, MD) was used as a quality control sample for all analyses. Concentrations were expressed as/~g/mg dry tissue weight (mean + standard error of the mean). Statistical significance was determined by a twotailed independent t test. Morphologic Analyses Sample fragments removed for histologic evaluation were fixed immediately in 10% neutral buffered formalin, dehydrated in graded concentrations of ethanol, cleared in xylene, and embedded in paraffin according to standard methods. Sections 5 p~m thick were stained with hematoxylin and eosin and Von-Kossa stain for demonstration of calcium deposits. Results Mineral Analyses STATIC IN VITRO TESTINC. Figure 1 summarizes the differential in vitro calcification of GA- versus NR-treated pericardial tissues with either weekly changes or no changes of the calcifying solution. With weekly changes of solution, the mean calcium content of GA-treated pericardium after 21 days of incubation (5.9-0.7 ~g/mg) was significantly higher than the calcium content of NR-treated pericardium (1.3 --- 0.2 /~g/mg) (p < 0.001). With no changes of solution, the absolute calcium content was less in either tissue, but was still significantly higher in GA-treated pericardium (p < 0.001). Interval gross inspection of GA-treated pericardium during the 21-day incubation period revealed progressively larger numbers of surface calcific deposits; this trend was not noted on the surface of the NR-pretreated tissue. SUBCUTANEOUS IMPLANTATION. Figure 2 summarizes the results for in vivo calcification of the GA- versus NRtreated bovine pericardium at 2, 3, and 5 weeks afterimplantation. The mean calcium content of CA-treated pericardium after 2, 3, and 5 weeks of subcutaneous

Ann Thorac Surg ABOLHODA ET AL 173 1996;62:169-74 CALCIFICATION OF BOVINE PERICARDIUM implantation was 49.6 + 9.6, 82.5 + 10.4, and 134.3-9.1 ~g/mg, respectively. Comparatively, the mean calcium content of the NR-treated pericardium was significantly lower (p < 0.05) at each corresponding interval (19.6 -+ 6.0, 32.3 + 12.2, and 21.4 -+ 5.2 /~g/mg, respectively). Furthermore, there was no evidence of a trend of progression of calcification in the NR-treated pericardium over the 5 weeks of observation. Morphologic Analyses The earliest histologic evidence of inflammatory mononuclear infiltration in the GA-pretreated bovine pericardium was noted in specimens harvested 2 weeks after implantation, with remarkable protracted progression of this phenomenon extending to 3 and 5 weeks after implantation (Figs 3C, 4B, 4C). Giant cell formation and associated intrinsic disruption of collagen architecture were hallmarks of the destructive changes observed in the GA-treated pericardium. On the other hand, the NRpretreated pericardium harvested after 3 and 5 weeks of implantation revealed a markedly attenuated inflammatory cellular response, limited to the serosal surface of the pericardium, and preservation of the fascicular organization of the collagen bundles (Figs 3A, 4A). Calcification of the GA-pretreated pericardium was observed histologically as early as 3 weeks after implantation (Fig 3D). Mineral deposits increased with time, and a confluence of calcific sites often permeated the tissue planes (Fig 5B). The NR-pretreated explants, however, showed no histologic evidence of calcific deposits after 3 weeks (Fig 3B). After 5 weeks of implantation, early calcific deposits in the form of small punctuated lesions were observed in one sample only; yet the morphologic integrity of the NR-treated pericardium was intact (Fig 5A). Comment The results of this study authenticate the ability of the NR treatment to delay significantly the degenerative mineralization of bovine pericardial tissue. In both our in vitro and in vivo static calcification models, the NR-treated pericardium calcified at a significantly diminished rate compared with conventional GA-treated tissue. Furthermore, our experiments show that although initiation of tissue mineralization does not require a host-mediated response, the in vivo milieu potentially has an important role in accelerating the calcification process. This phenomenon is further illustrated by the intense inflammatory reaction and the resulting tissue disruption noted in our morphologic examinations of the GA-treated pericardium and the absence of such destructive features in the NR-treated tissue. Even more remarkable is the protracted nature of this inflammatory response, which is concurrent with the progressive calcification of GApreserved tissue. Establishing a link between inflammatory response and accelerated tissue mineralization demands further proof, yet we believe it is a concept overlooked in Fig 5. Light photomicrographs of 5-week No-React- and glutaraldehyde-pretreated subcutaneous implants stained with the von Kossa technique. (A) No-React-treated pericardial tissue showing a few small punctuated calcium crystals with remarkably preseraed couagenous morphology. (B) Severe intrinsic disruption and surface ulcerations of glutaraldehyde-pretreated implant, with confluent sheets of calciflc deposits (CA+ +) replacing normal architecture. (xlo0.) previous investigations. The millipore chamber models merely address the cellular components of the inflammatory response and underestimate the role of chemical mediators of inflammation. With our growing knowledge of cytokines and molecular mediators of inflammation, such as tumor necrosis factor-a and transforming growth factor-/3, and their potentially injurious local effects mediated through paracrine cascades, the promoting role of the inflammatory response in bioprosthetic tissue degeneration can be clarified. We are currently conducting experiments in our laboratory to quantify the tissue levels of some of these mediators in explanted bovine pericardium and in porcine aortic valve cusps. In conclusion, the NR anticalcification treatment is a superior alternative to conventional GA treatment as a biologic modifier of xenograft tissue; it appears not only to arrest the degenerative mineralization of tissue implants, but also to abolish the host's inflammatory response to the xenograft. Long-term clinical in vivo circulatory trials are needed before recommending this

174 ABOLHODA ET AL Ann Thorac Surg CALCIFICATION OF BOVINE PERICARDIUM 1996;62:169-74 modality as the preferred chemical modification treatment for cardiac valve bioprostheses. We express our gratitude for the assistance of Mr Francis W. Kemp and Dr Shenggao Han with the atomic absorption spectroscopy analyses. We also thank Mr James A. Jetko for his superb technical assistance with histologic preparations. References 1. Schoen FJ, Kujovich JL, Levy RJ, Sutton MSJ. Bioprosthetic valve failure. Cardiovasc Clin 1987;18:289-97. 2. Levy RJ, Schoen FJ, Howard SL. Mechanism of calcification of porcine bioprosthetic aortic valve cusps: role of T lymphocytes. Am J Cardiol 1983;52:629--31. 3. Levy RJ, Schoen FJ, Levy JT, Nelson AC, Howard SL, Oshry LJ. Biologic determinants of dystrophic calcification and osteocalcin deposition in glutaraldehyde-preserved porcine aortic valve leaflets implanted subcutaneously in rats. Am J Pathol 1983;113:143-54. 4. Gabbay S, Kresh J. Bioengineering of mechanical and biologic heart valve substitutes. In: Morse D, Steiner RM, Fernandez J, eds. Guide to prosthetic cardiac valves. New York: Springer-Verlag, 1985:239. 5. Gabbay S, Bortolotti U, Iosif M. Mechanical factors influencing the durability O f heart valve pericardial bioprostheses. Trans Am Soc Artif Intern Organs 1985;32:282-7. 6. Thurbikar MJ, Deck JD, Aovad J, et al. Role of mechanical stress in calcification of aortic bioprosthetic valves. J Thorac Cardiovasc Surg 1983;86:115--25. 7. Carpentier A, Lemaigre G, Robert L, et al. Biological factors affecting long-term results of valvular heterografts. J Thorac Cardiovasc Surg 1969;58:467-83. 8. Angell WW, Angell JD, Kosek JC. Twelve-year experience with glutaraldehyde-preserved porcine xenografts. J Thorac Cardiovasc Surg 1982;83:493-502. 9. Schoen FJ, Collins JJ, Cohn LH. Long term failure rate and morphologic correlations in porcine bioprosthetic heart valve. Am J Cardiol 1983;51:957-64. 10. Inamura E, Sawotani O, Koyanagi H, et al. Epoxy compounds as new cross linking agent for porcine aortic leaflets. Subcutaneous implant studies in rats. J Cardiovasc Surg 1989;4:50-7. 11. Bernacca GM, Dimitri WR, Fisher AC, Mackay TG, Wheatley DJ. Chemical modification of bovine pericardium and its effects on calcification in the rat subdermal model. Biomaterials 1992;13:345-52. 12. Schoen FJ, Tsao JW, Levy RJ. Calcification of bovine pericardium used in cardiac valve bioprostheses: implications for the mechanism of bioprosthetic tissue mineralization. Am J Pathol 1986;123:134-45. 13. Chen W, Schoen FJ, Levy RJ. Mechanism of efficacy of 2-amino oleic acid for inhibition of calcification of giutaraldehyde-pretreated porcine bioprosthetic heart valves. Circulation 1994;90:323-9. 14. Levy RJ, Levy JT, Schoen FJ, et al. Prevention of bioprosthetic heart valve calcification. Circulation 1983;68(Suppl 3):395. 15. Levy RJ, Hawley MA, Schoen FJ, et al. Inhibition by diphosphonate compounds of calcification of porcine bioprosthetic heart valve cusps implanted subcutaneously in rats. Circulation 1985;71:349-56. 16. Gabbay S, Chuback JA, Khavarian C, Donahoo J, Oyarzun JR, Scoma R. In vitro and animal evaluation with the Biocor No-React TM anticalcification treatment for bioprostheses (new concept in anticalcification treatment). In: Gabbay S, Frater RW, eds. New horizons and the future of heart valve bioprostheses. Austin, Texas: Silent Partners, Inc, 1994:73-91. 17. Bemacca GM, Mackay TG, Wheatley DJ. In vitro calcification of bioprosthetic heart valves: report of a novel method and review of the biochemical factors involved. J Heart Valve Dis 1992;1:115-30. 18. Bogden JD, Kemp FW, Han S, et al. Dietary calcium and lead interact to modify maternal blood pressure, erythropoiesis, and fetal and neonatal growth in rats during pregnancy and lactation. J Nutr 1995;125:990-1002.