Brief Communication. Alkaline Phosphatase Activity of Glutaraldehyde-Treated Bovine Pericardium Used in Bioprosthetic Cardiac Valves

Transcription

1 844 Brief Communication Alkaline Phosphatase Activity of Glutaraldehyde-Treated Bovine Pericardium Used in Bioprosthetic Cardiac Valves Anthony R. Maranto and Frederick J. Schoen Bioprosthetic valves fail frequently because of pathological mineralization, a process that begins in cell remnants of the glutaraldehyde (GLUT) fixed tissue. Other pathological cardiovascular calcification and physiological mineralization in skeletal/dental tissues are both largely initiated in cell-derived membranous structures (often called "matrix vesicles"), and the enzyme alkaline phosphatase (AP) likely has an important function in the pathogenesis of mineral nucleation. This study tested the hypothesis that AP might also be present in and contribute to calcification of bioprosthetic valves. AP activity of fresh and GLUT-treated bovine pericardium was measured by the conversion of p-nitrophenyl phosphate to p-nitrophenol. Following 24 hours in 0.6% HEPES-buffered GLUT and storage for 2 weeks in 0.2% GLUT, considerable AP hydrolytk activity remained in GLUT-treated tissue relative to that of fresh tissue (\ mmi, 24 vs. 45 fimol reaction product/min/mg tissue protein, respectively), although binding was somewhat reduced (K m, 1.9 x 10 3 vs. 1.4 x 10 3 fim substrate, respectively). Enzyme reaction product was demonstrated in both fixed and fresh tissue by light microscopic histochemical studies, confirming the biochemical results. Reaction product was noted along membranes of vascular endothelial cells and interstitial fibroblasts, the sites of early calcine deposits hi bioprosthetic valves, by ultrastructural examination of GLUT-treated tissue. We conclude that GLUTtreated bovine pericardium retains much of the hydrolytlc activity of AP, an enzyme associated with normal skeletal and pathological cardiovascular and noncardio vascular mineralization, and suggest that further examination of the mechanistic role of this enzyme may stimulate new approaches for slowing or preventing calcification of bioprosthetic tissue. (Circulation Research 1988;63: ) Mineralization is the primary cause of failure of bioprosthetic cardiac valves. 1 Cellular elements in the xenograft tissue devitalized by glutaraldehyde (GLUT) cross-linking are the predominant sites of mineral deposition in both experimental and clinical implants. 23 In skeletal and dental tissues, normal (i.e., physiological) mineralization largely occurs initially in extracellular membranous particles derived from cellular remnants, often called "matrix vesicles." 4 Mineralization in matrix vesicle like membranous fragments is also noted in cardiovascular diseases characterized by dystrophic calcification, including atherosclerosis 5 and calcific valvular degeneration. 6 There is substantial evidence that the enzyme alkaline From the Department of Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts. Supported in part by National Institutes of Health Grant HL Address for reprints: Frederick J. Schoen, MD, PhD, Department of Pathology, Brigham and Women's Hospital, 75 Francis Street, Boston, MA Received January 12, 1988; accepted May 2, phosphatase (AP) is important in the pathogenesis of physiological mineralization. Matrix vesicle membranes are enriched in AP. 7 In vitro calcification of chick embryonic bone requires matrix vesicle pyrophosphatase activity 8 and inhibitors of AP activity reduce mineralization. 9 Moreover, AP has been noted at mineralization loci This study tested the hypothesis that AP is present in and might contribute to mineralization of bioprosthetic valve tissue. Despite the widely held view that the lengthy GLUT treatments employed in the fabrication and storage of bioprostheses should be expected to destroy all metabolic activities characteristic of viable tissue, it is well documented that the aldehyde cross-linking obtained in tissue fixation for pathological studies incompletely quenches most enzyme activities. Recovery of enzyme activity ranges from 10% to 75%. 12 The results of the present study demonstrate that considerable AP activity persists in bovine pericardium after GLUT treatments of the type used to manufacture bioprosthetic valves. Moreover, this activity is localized to membranous cell remnants known to be initial sites of mineralization. 2-3

2 Maranto and Schoen Alkaline Phosphatase in Bioprosthetic Tissue 845 Materials and Methods Pericardium Parietal pericardium was harvested from a freshly killed 2-3-week-old calf, trimmed of adherent fat, and rinsed in 0.9% saline buffered with 50 mm HEPES (ph 7.4). Tissue samples were treated with 0.6% GLUT in HEPES-buffered saline for 24 hours at 4 C, stored in 0.2% GLUT in the same buffer for 2 weeks at 4 C, and washed extensively in buffered saline prior to analysis. These treatments are similar to those used in the manufacture of clinical bioprosthetic valves and in previous experiments. 13 Untreated tissue was transported on gauze moistened with buffered saline at 4 C and analyzed within 2 hours of harvest. Enzyme Assay AP extracts of pericardium were prepared by a modification of the procedure of VanBelle. 13 Ten 3 x 4-mm pieces of thoroughly washed tissue were macerated in a mortar and pestle with five changes each of 1 ml distilled water and 1 ml 2-butanol. The resulting homogenates were pooled and centrifuged at 600g for 10 minutes, and the aqueous phase was assayed without further purification. Protein content was determined by the method of Lowry et al. 14 Hydrolysis of the substrate p-nitrophenyl phosphate (PNPP, Sigma Chemical Company, St. Louis, Missouri) was performed in 0.15 M 2-amino-2- methyl-1-propanol (Sigma) buffer at ph 10.4 and 37 C by a modification of the procedure of Cyboron and Wuthier. 15 Assays were initiated by the addition of a greater than 10-fold excess of substrate and terminated after 30 minutes by the addition of 0.1N NaOH. The appearance of p- nitrophenol (PNP) was measured at 405 nm using a molar extinction coefficient of 4,545 cm" 1 M" 1 as determined from a standard curve of PNP (Sigma). The rate of PNP formation (V o ) was determined for several concentrations of substrate and expressed as micromoles PNP per minute per milligram protein. At each substrate concentration, three aliquots of homogenate were used. Kinetic parameters (K m and VmjJ were estimated by a linear least-squares fit on a double reciprocal plot of 1/V O versus 1/[PNPP]. Cytochemistry For light microscopy, 10-^tm thick sections of untreated and GLUT pericardium were cut with a cryomicrotome, mounted on glass slides, and air dried at room temperature. Sections were incubated at 37 C for 30 minutes in a medium for AP modified from Gomori 16 containing 50 ml of 5 mm MgCl 2 in distilled water, 1 ml Napthol AS-MX phosphate in ph 10.5 buffer (Sigma), and 25 mg Fast Red TR or Fast Blue RR (Sigma). Following incubation, sections were rinsed and mounted without counterstaining. For ultrastructural localization of AP, pericardium treated with GLUT as above was minced into 1 x 1-mm pieces and incubated at 37 C for 15 TABLE 1. Kinetic Parameters of Alkaline Phosphatase From Fresh and GlutaraMehyde-Treated Bovine Pericardium Fresh 1.4 Glutaraldehyde-treated 1.9 K m Gtmol PNP/min/mg (10 3 /im PNPP) protein) minutes in a glycerophosphate and lead citrate medium, ph 9.5." Specimens were rinsed, fixed with 1% OsO 4, dehydrated with ethanol, and embedded in Epon 812 (Polysciences Inc, Warrington, Pennsylvania). Ultrathin sections without additional staining were examined with a JEOL JSM- 100 transmission electron microscope. Controls for both light and electron microscopy included tissue incubated without substrate or with the addition of 0.1 mm L-p-bromotetramisole (Sigma), a specific inhibitor of AP. 18 Results The K m and V^, of AP extracted from fresh and glutaraldehyde-treated pericardium are compared in Table 1. Considerable hydrolytic activity was demonstrable in both untreated and fixed tissues with the n-butanol extraction employed. Over half of the original AP activity was measured after GLUT treatment, although binding was somewhat reduced. Some of the lost activity may simply reflect reduced extractability of enzyme from treated tissue and thereby underestimate the actual activity in fixed tissue. Thus, a substantial fraction of the original AP activity survived initial treatment and 2 weeks of storage in GLUT. AP activity demonstrable by light microscopic histochemistry was similar before and after GLUT treatment (Figure 1). However, overall morphology was better preserved in treated tissue, which allowed better identification of reaction sites. Enzyme activity was most prominent in the cells lining small blood vessels near the serosal surface and in connective tissue cells in the fibrosa. In agreement with light microscopic findings (Figure 2), ultrastructural examination of GLUTtreated tissue revealed that AP reaction product was deposited along membranes of vascular endothelial cells and interstitial fibroblasts. Mitochondrial and plasma membranes reacted in all cells. In fibroblasts, reaction product was also found along large vacuoles, which probably arose from the unavoidable tissue autolysis in bioprosthetic tissue preparation. Neither collagen bundles nor cell nuclei had reaction product. Control tissues incubated without substrate or with the addition of L-pbromotetramisole were unreactive in either the light or electron microscopic incubation media. Discussion Retained Alkaline Phosphatase Activity Following Glutaraldehyde Treatment The nonviability of GLUT-treated bioprosthetic tissue is usually equated with the loss of all normal

3 846 Circulation Research Vol 63, No 4, October *- % v$4*-"-rs v ^ \ FIGURE 1. Cryomicrotome sections of pericardium stained for alkaline phosphatase. A: Untreated (fresh) tissue with staining of cells in thefibrosa (arrows). B and C: Glutaraldehyde-treated tissue, demonstrating staining of connective tissue cells (arrows) and walls of blood vessels (arrowheads). x375. metabolic activities. This is clearly an inaccurate concept in view of the results of the present study, which demonstrate that bovine pericardium treated with GLUT by methods used in making biopros- thetic valves retained the hydrolytic activity of AP, an enzyme associated with normal and pathological mineralization in other tissues The crude nbutanol extract from GLUT-treated pericardium had approximately half the affinity (Km, 1.9 x 103 vs. 0.7 x 103 fim) and one tenth the activity ( V ^ 24 vs. 220 ^.mol/min/mg) as highly purified AP from epiphyseal cartilage assayed with the same substrate under similar conditions.15 Cytochemical examination revealed that enzyme activity was associated with membranes, as in other tissues,9 and was completely inhibited by 0.1 mm L-p-bromotetramisole, a stereospecific uncompetitive inhibitor of AP in liver, kidney, and bone.18 There was strong correlation between the distribution of AP activity and the sites of initial and predominant mineral accumulation observed in clinical explants and tissue derived from circulatory and subcutaneous models of bovine pericardium and porcine aortic valve.1-3 Interstitial connective tissue cells and endothelial cells lining small blood vessels, frequently among the earliest cells to mineralize, were found to react strongly for AP. Ultrastructurally, AP was found along the membranes of vesicles, mitochondria, and other cell structures in which early mineral deposits are accumulated. Collagen fibers, which appear to mineralize later, were not reactive. It is unlikely that the results can be explained by inadequate tissue treatment with GLUT since aldehyde uptake is complete within hours3 and the conditions of treatment used incorporate a high degree of GLUT cross-links.19 Implications of Residual Alkaline Phosphatase Activity The discovery of AP activity in GLUT-treated pericardium strengthens the comparison between mechanisms of mineralization in bioprosthetic tissue and those in other cardiovascular and noncardiovascular tissues. 34 Matrix vesicles from epiphyseal cartilage contain high levels of AP,8 and specific inhibitors of AP activity block mineralization.9 Matrix vesicle-like membranes in mineralizing atherosclerotic plaques also have AP activity.3 The most likely role of AP in promoting mineralization is that it participates in transporting phosphate across membranes. This proposal is supported by the enzyme's high affinity for phosphate and the ability of inhibitors to block phosphate uptake.9*20 If AP were involved in bioprosthetic valve mineralization, then new possibilities might be considered for slowing or preventing this process. One strategy might be to eliminate enzyme activity from the implant by extraction or inhibition.20 AP can be extracted by detergents, n-butanol, or phosphatidylinositol-specific phospholipase.20 Inhibitors fall into two classes: 1) competitive, such as phosphate and its analogues: arsenate, vanadate, phenylene-l,3-diphosphonate, and phosphonoacetaldehyde; and 2) uncompetitive, such as certain

4 Maranto and Schoen Alkaline Phosphatase in Bioprosthetic Tissue 847 m * or-* FIGURE 2. Ultrastructural demonstration of alkaline phosphatase activity in glutaraldehyde-treated pericardium. A (with inset): Capillary endothelial cell with reaction product associated with membranes of plasmalemmal vesicles (arrowheads), mitochondria (m), the plasmaiemma (arrow), and intraluminal debris (d). Nucleus (n) is unstained. B: Fibroblast with activity localized to mitochondria (m), irregular vacuoles (v), and the plasmaiemma (arrow). Nucleus (n) and collagen fibers (c) are unreactive. Bar, 0.5 fim in each. L-amino acids and their analogues, such as theophylline, L-tetramisole, and L-p-bromotetramisole.20 Host and implant modifications that reduce mineralization of bioprostheses in experimental models may reduce tissue AP activity. Systemic or local administration, or cuspal pretreatment with diphosphonate compounds profoundly inhibit bioprosthetic valve mineralization.1 Since some diphosphonates are competitive inhibitors of AP, this treatment could potentially block the enzyme's active site. Pretreatment with the detergent sodium dodecyl sulfate removes phospholipids, believed to be promoters of mineralization and also inhibits calcification of bioprosthetic valve tissue implanted subcutaneously in rats.1 It is likely that this treatment also extracts AP. Successful elimination of AP activity from implants may be precluded by reversal of inhibition or reaccumulation of enzyme from the host. Although sodium dodecyl sulfate reduces mineralization of subcutaneous implants, sodium dodecyl sulfate appears to have little consistent benefit in inhibiting mineralization in long-term circulatory implants, perhaps due to enzyme reaccumulation from the blood.1 Although a more effective strategy might involve chronic suppression of AP in the host, systemic toxicity may limit the ability to take this approach. Theophylline and L-tetramisole are currently administered to humans for other purposes, but at therapeutic plasma concentrations (80 /im and 2 /xm, respectively), their inhibitory effect on AP is negligible,13 and higher doses are associated with significant toxicity. Limitations to This Study Although this study demonstrated that alkaline phosphatase activity is retained following the chemical treatment used in the fabrication of bioprosthetic heart valves, it remains to be demonstrated whether this enzyme is contributory to the calcification of bioprosthetic tissue. Moreover, this study used calf pericardium, while widely used bioprosthetic valves have been constructed mostly from glutaraldehyde-treated porcine aortic valve and, to a lesser extent, from adult bovine pericardium. Although the alkaline phosphatase activity of adult bovine pericardium and porcine aortic valve certainly need to be determined in future investigations, the morphological, kinetic, and biochemical features of mineralization of pericardium, both adult and calf, and porcine valve are almost identical.13 It would also be interesting to know whether calcified and noncalcified bioprosthetic valves removed following in vivo function have this enzyme activity. Finally, the propensity for bioprosthetic valves to calcify has been shown to have host, implant, and

5 848 Circulation Research Vol 63, No 4, October 1988 mechanical determinants. 1 The present study addressed only considerations related to the inherent susceptibility to calcification of the substrate tissue, and neither supports nor refutes potential additional contributions of mechanical factors ("wear and tear") or host factors (the effects of recipient age or immunological processes) to the pathophysiology of mineralization. In conclusion, GLUT-treated bovine pericardium retains the major fraction of original AP activity, an enzyme associated with normal and pathological cardiovascular and noncardiovascular calcification. This suggests that the generally held view that bioprosthetic tissue has lost all normal metabolic activities is incorrect, and that further examination of the role of this and related enzymes in the calcification of bioprosthetic tissue may stimulate new approaches for inhibiting or preventing mineralization. Acknowledgments The authors are grateful to Drs. H. Clarke Anderson and Robert J. Levy for helpful comments and review of the data, to Sara Murray and Helen Shing for technical assistance, and to Roberta Baxter for typing this manuscript. References 1. Schocn FJ, Kujovich JL, Levy RJ, St. John Sutton M: Bioprosthetic valve failure. Cardiovasc Clin 1988;18/2: Valente M, Bortolotti U, Thiene G: Ultrastructural substrates of dystrophic calcification in porcine bioprosthetic valve failure. Am J Pathol 1985;119: Schoen FJ, Tsao JW, Levy RJ: Calcification of bovine pericardium used in cardiac valve bioprostheses: Implications for the mechanisms of bioprosthetic tissue mineralization. Am J Pathol 1986;123: Anderson HC: Mineralization by matrix vesicles. Scan Electron Microsc 1984,2: Tanimura A, McGregor DH, Anderson HC: Matrix vesicles in atherosclerotic calcification. Proc Soc Exp Biol Med 1983; 172: Kim KM: Calcification of matrix vesicle in human aortic valve and aortic media. Fed Proc 1976;35: Anderson HC, Matsuzawa J, Sajdera SW, Ali SY: Membranous particles in calcifying cartilage matrix. Trans NYAcad Sci 1979;32: Anderson HC, Reynolds JJ: Pyrophosphate stimulation of calcium uptake into cultured embryonic bones. Fine structure of matrix vesicles and their role in calcification. Dev fl/o/1973;34: Register TC, Warner GP, Wuthier RE: Effect of L- and D-tetramisole on 32Pi and 45Ca uptake and mineralization by matrix vesicle-enriched fractions from chicken epiphyseal cartilage. J Biol Chem 1984;259: Ali SY, Sajdera SW, Anderson HC: Isolation and characterization of calcifying matrix vesicles from epiphyseal cartilage. Proc Natl Acad Sci USA 1970;67: Kanabe S, Hsu HHT, Cecil RNA, Anderson HC: Electron microscopic localization of adenosine triphosphate (ATP>hydrolyzjng activity in isolated matrix vesicles and reconstituted vesicles from calf cartilage. J Histochem Cytochem 1983 ;31: Hayat MA: Fixation for Electron Microscopy. New York, Academic Press, 1981, pp VanBelle H: Kinetics and inhibition of alkaline phosphatase from canine tissues. Biochim Biophys Ada 1972 ;289: Lowry OH, Rosebrough NO, Farr AL, Randall RJ: Protein measurement with the Folin phenol reagent. J Biol Chem 1951;193: Cyboron GW, Wuthier RE: Purification and initial characterization of intrinsic membrane-bound alkaline phosphatase from chicken epiphyseal cartilage. J Biol Chem 1981 ;256: Gamori G: Alkaline phosphatase of cell nuclei. J Lab Clin 17. Lewis PR: Metal precipitation methods for hydrolytic enzymes, in Lewis PR, Knight DP (eds): Staining Methods for Sectioned Material. New York, Elsevier/North Holland, 1977, pp Borgers M: The cytochemical application of new potent inhibitors of alkaline phosphatase. J Histochem Cytochem 1973;21: Golomb G, Schoen FJ, Smith MS, Linden J, Dixon M, Levy RJ: The role of glutaraldehyde-induced cross-links in calcification of bovine pericardium used in cardiac valve bioprostheses. Am J Pathol 1987;127: Wuthier RE, Register TC: Role of alkaline phosphatase, a polyfunctional enzyme, in mineralizing tissues, in Veis A (ed): The Chemistry and Biology of Mineralized Tissues. New York, Elsevier/North Holland, 1984, pp KEY WORDS calcification pericardium alkaline phosphatase bioprosthetic valve