Chelating Activity of Advanced Glycation End-product Inhibitors

Abstract

The advanced glycation end-product (AGE) hypothesis proposes that accelerated chemical modification of proteins by glucose during hyperglycemia contributes to the pathogenesis of diabetic complications. The two most commonly measured AGEs, Nε-(carboxymethyl)lysine and pentosidine, are glycoxidation products, formed from glucose by sequential glycation and autoxidation reactions. Although several compounds have been developed as AGE inhibitors and are being tested in animal models of diabetes and in clinical trials, the mechanism of action of these inhibitors is poorly understood. In general, they are thought to function as nucleophilic traps for reactive carbonyl intermediates in the formation of AGEs; however alternative mechanisms of actions, such as chelation, have not been rigorously examined. To distinguish between the carbonyl trapping and antioxidant activity of AGE inhibitors, we have measured the chelating activity of the inhibitors by determining the concentration required for 50% inhibition of the rate of copper-catalyzed autoxidation of ascorbic acid in phosphate buffer. All AGE inhibitors studied were chelators of copper, as measured by inhibition of metal-catalyzed autoxidation of ascorbate. Apparent binding constants for copper ranged from ∼2 mm for aminoguanidine and pyridoxamine, to 10–100 μm for carnosine, phenazinediamine, OPB-9195 and tenilsetam. The AGE-breakers, phenacylthiazolium and phenacyldimethylthiazolium bromide, and their hydrolysis products, were among the most potent inhibitors of ascorbate oxidation. We conclude that, at millimolar concentrations of AGE inhibitors used in many in vitro studies, inhibition of AGE formation results primarily from the chelating or antioxidant activity of the AGE inhibitors, rather than their carbonyl trapping activity. Further, at therapeutic concentrations, the chelating activity of AGE inhibitors and AGE-breakers may contribute to their inhibition of AGE formation and protection against development of diabetic complications. The advanced glycation end-product (AGE) hypothesis proposes that accelerated chemical modification of proteins by glucose during hyperglycemia contributes to the pathogenesis of diabetic complications. The two most commonly measured AGEs, Nε-(carboxymethyl)lysine and pentosidine, are glycoxidation products, formed from glucose by sequential glycation and autoxidation reactions. Although several compounds have been developed as AGE inhibitors and are being tested in animal models of diabetes and in clinical trials, the mechanism of action of these inhibitors is poorly understood. In general, they are thought to function as nucleophilic traps for reactive carbonyl intermediates in the formation of AGEs; however alternative mechanisms of actions, such as chelation, have not been rigorously examined. To distinguish between the carbonyl trapping and antioxidant activity of AGE inhibitors, we have measured the chelating activity of the inhibitors by determining the concentration required for 50% inhibition of the rate of copper-catalyzed autoxidation of ascorbic acid in phosphate buffer. All AGE inhibitors studied were chelators of copper, as measured by inhibition of metal-catalyzed autoxidation of ascorbate. Apparent binding constants for copper ranged from ∼2 mm for aminoguanidine and pyridoxamine, to 10–100 μm for carnosine, phenazinediamine, OPB-9195 and tenilsetam. The AGE-breakers, phenacylthiazolium and phenacyldimethylthiazolium bromide, and their hydrolysis products, were among the most potent inhibitors of ascorbate oxidation. We conclude that, at millimolar concentrations of AGE inhibitors used in many in vitro studies, inhibition of AGE formation results primarily from the chelating or antioxidant activity of the AGE inhibitors, rather than their carbonyl trapping activity. Further, at therapeutic concentrations, the chelating activity of AGE inhibitors and AGE-breakers may contribute to their inhibition of AGE formation and protection against development of diabetic complications. advanced glycation end-product ascorbic acid aminoguanidine Nε-(carboxymethyl)lysine diaminophenazine diethylenetriaminepentaacetic acid heptafluorobutyric acid pyridoxamine phenacyldimethylthiazolium bromide phenacylthiazolium bromide reverse phase-high performance liquid chromatography The Maillard or browning reaction between sugars and proteins leads to the formation of chemical modifications and cross-links in proteins, known as advanced glycation end-products (AGEs)1 (1Baynes J.W. Thorpe S.R. Turtle J.R. Kaneko T. Osato S. Diabetes in the New Millennium. Endocrinology and Diabetes Research Foundation, Sydney1999: 337-350Google Scholar). These products contribute to the age-dependent chemical modification of long lived tissue proteins (2Thorpe S.R. Baynes J.W. Drugs & Aging. 1996; 9: 69-77Crossref PubMed Scopus (269) Google Scholar), and accelerated formation of AGEs during hyperglycemia is implicated in the development of diabetic complications (1Baynes J.W. Thorpe S.R. Turtle J.R. Kaneko T. Osato S. Diabetes in the New Millennium. Endocrinology and Diabetes Research Foundation, Sydney1999: 337-350Google Scholar, 2Thorpe S.R. Baynes J.W. Drugs & Aging. 1996; 9: 69-77Crossref PubMed Scopus (269) Google Scholar). Consistent with the AGE hypothesis, AGE inhibitors, such as aminoguanidine (AG) (3Nilsson B.O. Inflamm. Res. 1999; 48: 509-515Crossref PubMed Scopus (198) Google Scholar) and pyridoxamine (PM) (4Khalifah R.G. Baynes J.W. Hudson B.G. Biochem. Biophys. Res. Commun. 1999; 257: 251-258Crossref PubMed Scopus (252) Google Scholar), inhibit the formation of AGEs in various proteins in vitroand in collagen in vivo (5Degenhardt T.P. Alderson N.L. Arrington D.D. Beattie R.J. Basgen J.M. Steffes M.W. Thorpe S.R. Baynes J.W. Kidney Internat. 2001; (in press)Google Scholar, 6Alderson N.L. Metz T.O. Chachich M.E. Baynes J.W. Thorpe S.R. Diabetes. 2001; 50 (Abstr. 696P): A172Crossref Google Scholar) and retard the development of diabetic complications in animal models. The two most commonly measured AGEs, Nε-(carboxymethyl)lysine (CML) and pentosidine, are glycoxidation products formed by sequential glycation and oxidation reactions (7Baynes J.W. Thorpe S.R. Diabetes. 1999; 48: 1-9Crossref PubMed Scopus (2143) Google Scholar), and traces of transition metal ions in physiological buffers are potent catalysts of the formation of these AGEs in vitro (8Chace K.V. Carubelli R. Nordquist R.E. Arch. Biochem. Biophys. 1991; 288: 473-480Crossref PubMed Scopus (82) Google Scholar). AGE inhibitors are nucleophilic compounds designed to trap reactive carbonyl or dicarbonyl intermediates in AGE formation. However, as shown here, some of these compounds also have potent chelating activity, making it difficult to dissect the antioxidative, chelating activity from the carbonyl trapping activity of the inhibitors in in vitro studies. In the present work, we measured the chelating activity of AGE inhibitors by determining the concentration required for 50% inhibition of copper-catalyzed autoxidation of ascorbic acid in phosphate buffer. Using this assay, we show that AGE inhibitors have weak-to-potent metal chelating activity, and that, in fact, many experiments demonstrating the activity of AGE inhibitors in vitro are likely to be measuring the chelating activity rather than carbonyl trapping activity of these compounds. We also show that AGE-breakers and their hydrolysis products are potent chelators of copper and conclude that both the carbonyl trapping and chelation activity of AGE inhibitors and AGE-breakers may be involved in their therapeutic mechanism of action. Unless otherwise noted, all chemicals were purchased from Sigma/Aldrich Chemical Co. Monobasic and dibasic sodium phosphate salts were purchased from EM Science (Gibbstown, NJ); heptafluorobutyric acid (HFBA) from Acros (Pittsburgh, PA), and nitric acid and HPLC grade acetonitrile from Fisher Scientific (Pittsburgh, PA). Tenilsetam ((±)-3-(2-thienyl)-2-piperazinone) (9Munch G. Taneli Y. Schraven E. Schindler U. Schinzel R. Palm D. Riederer P. J. Neural Transm. 1994; 8: 193-208Crossref PubMed Scopus (77) Google Scholar, 10Shoda H. Miyata S. Liu B.F. Yamada H. Ohara T. Suzuki K. Oimomi M. Kasuga M. Endocrinology. 1997; 138: 1886-1892Crossref PubMed Google Scholar) was a gift from Aventis Pharma, Germany, and OPB-9195 ((±)-2-isopropylidenehydrazono-4-oxo-thiazolidin-5-ylacetanilide) (11Nakamura S. Makita Z. Ishikawa S. Yasumura K. Fujii W. Yanagisawa K. Kawata T. Koike T. Diabetes. 1997; 46: 895-899Crossref PubMed Google Scholar,12Miyata T. Ueda Y. Asahi K. Izuhara Y. Inagi R. Saito A. Van Ypersele De Strihou C. Kurokawa K. J. Am. Soc. Nephrol. 2000; 11: 1719-1725PubMed Google Scholar) from Otsuka Pharmaceuticals, Japan. Phenacylthiazolium bromide (PTB) and phenacyldimethylthiazolium bromide (PMTB) were synthesized according to the methods of Vasan et al. (13Vasan S. Zhang X. Zhang X. Kapurniotu A. Bernhagen J. Teichberg S. Basgen J. Wagle D. Shih D. Terlecky I. Bucala R. Cerami A. Egan J. Ulrich P. Nature. 1996; 382: 211-212Crossref PubMed Scopus (426) Google Scholar) and Wolfenbüttel et al. (14Wolffenbüttel B.H. Boulanger C.M. Crijns F.R. Huijberts M.S. Poitevin P. Swennen G.N Vasan S. Egan J.J. Ulrich P. Cerami A. Levy B.I. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4630-4634Crossref PubMed Scopus (379) Google Scholar), respectively, and purified to homogeneity by reverse-phase HPLC. 2,3-Diaminophenzine (DAP) was synthesized according to Soulis et al. (15Soulis T. Sastra S. Thallas V. Mortensen S.B. Wilken M. Clausen J.T. Bjerrum O.J. Petersen H. Lau J. Jerums G. Boel E. Cooper M.E. Diabetologia. 1999; 42: 472-479Crossref PubMed Scopus (47) Google Scholar). PTB, PMTB, and DAP were homogeneous by reverse phase-HPLC analysis and by electrospray ionization liquid chromatography-mass spectrometry. All water used in these experiments was distilled and then deionized (15–18 MΩ) using a mixed bed ion-exchanger with activated charcoal polisher. Chelex-100 (∼80% water content) was suspended and allowed to settle several times in water to remove soluble chelator and fine particles (16Van Reyk D.M. Brown A.J. Jessup W. Dean R.T. Free Radic. Res. 1995; 23: 533-535Crossref PubMed Scopus (29) Google Scholar). To remove metal ions from phosphate buffer, chelex-100 resin (7 g/liter, wet weight) was added to 50 mm phosphate buffer, pH 7.4, in a 2-liter plastic container and rocked gently overnight at room temperature. The buffer was stored at room temperature in the same container, allowing the resin to settle to the bottom; aliquots were removed for experiments, as needed. All incubations (1.5-ml total volume) were conducted in chelex-treated 50 mm phosphate buffer, pH 7.4, in a water bath at 30 °C; reagents were pre-equilibrated at this temperature for at least 5 min. In preliminary experiments, the kinetics of oxidation of AA was determined in quartz spectrophotometer cuvettes by measuring the rate of decrease in absorbance at 265 nm, as described by Buettner (17Buettner G.A. Methods Enzymol. 1990; 186: 125-127Crossref PubMed Scopus (77) Google Scholar); the cuvettes were precleaned by soaking overnight in 1% nitric acid. AA (500 μm) was stable (>99% recovery) in chelex-treated 50 mm phosphate buffer, pH 7.4, for at least 1 h. Addition of 500 nm CuCl2 yielded a half-life of approximately 1 h. For studies with AGE inhibitors, the copper and inhibitor were preincubated in phosphate buffer for 5 min, and then the ascorbate was added from a concentrated solution (75 μl of 10 mm AA in water) to initiate the reaction. Aliquots (150 μl) were removed at 0, 15, 30, 45, and 60 min and transferred to autoinjector vials containing 15 μl of 10 mm diethylenetriaminepentaacetic acid (DTPA) to quench the metal-catalyzed oxidation reaction. Samples were analyzed by RP-HPLC, using a Supelcosil LC-18 column (25 cm × 4.6 mm, 5 μm) (Supelco; Bellefonte, PA) on a Shimadzu (Columbia, MD) LC-10Ai liquid chromatograph equipped with a SIL-10Ai auto-injector and an SPD-10AV UV-Vis detector. Solvents for HPLC were: A, 0.1% HFBA in water and B, 50% acetonitrile. The gradient was 100% A at 0 min, increase to 80% B from 0 to 4 min, return to 100% A from 4 to 4.1 min, then re-equilibrate in 100% A from 4.1 to 9 min. The absorbance of ascorbic acid was measured at 244 nm (Fig. 1), and the peak area was integrated to estimate the percent of AA remaining versustime. The equations for the least-squares fits to these plots (triplicate analyses) were determined with SigmaPlot 2000 (Jandel Scientific) and used to determine the concentration of each compound that inhibited the rate of AA oxidation by 50% (IC50). All assays with AGE inhibitors were performed in triplicate by HPLC, measuring residual ascorbate by absorbance at 244 nm (Fig.1). None of the compounds tested co-eluted with or otherwise interfered with the measurement of ascorbate by this RP-HPLC method. In preliminary studies, we adjusted experimental conditions to obtain a concentration of cupric ion that would catalyze the oxidation of AA (500 μm) with an approximate half-life of 1 h. As shown in Fig.2 (lower lines; without AG or PM), this was achieved by addition of 500 nmCuCl2 to 50 mm chelex-treated phosphate buffer, pH 7.4, at 30 °C. This half-life was comparable with the half-life of ascorbate in 0.1 m phosphate buffers prepared in our laboratory; however, buffers prepared with different batches of phosphate salts varied in their catalytic activity because of batch-to-batch variation in metal ion content. The AGE inhibitors, AG and PM, inhibited the oxidation of AA in a concentration-dependent manner (Fig. 2). The linearity of the semilogarithmic plots illustrates that the AGE inhibitors did not alter the pseudo-first order kinetics of the autoxidation reaction. Graphical analysis of the rate of ascorbate oxidation versusinhibitor concentration indicated that percent inhibition was not a linear function of inhibitor concentration. The linear relationships between percent inhibition and the square of the inhibitor concentration, shown in Fig. 3, were consistent with formation of 2:1 or higher valency complexes between AG or PM and Cu(II) ions. These data yielded IC50 values of ∼2.5 mm for AG and ∼1 mm for PM. Amine nitrogens are common ligands for transition metal ions and pyridoxine (IC50 ∼3.6 mm) and pyridoxal (IC50 ≫ 5 mm), with a hydroxymethyl or carboxaldehyde group, in lieu of the aminomethyl group of PM, were less inhibitory than PM itself (Fig. 3). and the acid were potent inhibitors of AA with IC50 values and The of and was an function of inhibitor concentration, that the may have some residual catalytic activity. compounds and on metal-catalyzed oxidation of As shown in Fig. the AGE inhibitors, DAP and IC50 values of and respectively, OPB-9195 (Fig. yielded a with a at of the AGE-breakers, and PMTB, were also potent inhibitors of ascorbate with IC50 values of and (Fig. the half-life of and were and respectively, in phosphate buffer at 30 R. J. E. and J. W. in we also measured the activity of the hydrolysis products in the ascorbate oxidation For these experiments, the AGE-breakers were for in buffer to addition of the ascorbate. As also shown in Fig. the hydrolysis products IC50 values of μm and were among the most potent inhibitors of metal-catalyzed oxidation of ascorbate. The with (Fig. that during the of the may have to the inhibitory activity of this compound at higher To the of AGE inhibitors and AGE-breakers as inhibitors of metal-catalyzed oxidation in we also measured the of on the kinetics of ascorbate oxidation. As shown in Fig. among all the compounds was the most potent inhibitor of ascorbate oxidation on a The IC50 for nm was of the concentration of Cu(II) in the reaction at least binding for The was a in on the metal chelating activity of designed for Free Radic. 1990; 9: PubMed Scopus Google Scholar, J.W. Biochem. 1991; 42: PubMed Scopus Google Scholar). inhibitors and with activity were shown to have chelating activity, that their chelating activity proteins from chemical by metal-catalyzed autoxidation of sugars or ascorbate Free Radic. 1991; PubMed Scopus Google P. S. J. A. Biophys. PubMed Scopus Google Scholar). In the present we have the chelating activity of various AGE inhibitors to determine the of metal chelation in the mechanism of action of AGE A reaction for the formation of AGEs from AA in vitro is shown in Fig. autoxidation the browning and of by and is for the formation of the glycoxidation products, and such as are potent inhibitors of formation of these AGEs from ascorbate J.M. Thorpe S.R. Baynes J.W. 1990; PubMed Scopus Google Scholar) and glucose Thorpe S.R. Baynes J.W. Diabetes. 1994; PubMed Google Scholar) (Fig. In AGE inhibitors, such as AG and PM, are thought to function as carbonyl traps (3Nilsson B.O. Inflamm. Res. 1999; 48: 509-515Crossref PubMed Scopus (198) Google Scholar, R.G. Baynes J.W. Hudson B.G. Biochem. Biophys. Res. Commun. 1999; 257: 251-258Crossref PubMed Scopus (252) Google Scholar), with carbonyl intermediates formed on autoxidation of against chemical modification of nucleophilic on (Fig. experiments however, that AGE inhibitors are also transition metal chelators and inhibit the autoxidation of to formation of reactive carbonyl compounds. AGE inhibitors IC50 values for copper from the millimolar for AG and PM to values for carnosine, and These results that AGE inhibitors have two mechanisms of action in inhibition of formation of reactive carbonyl and nucleophilic reaction with these concentrations of AG and inhibitors have been in many studies, our Thorpe S.R. Baynes J.W. Diabetes. 1994; PubMed Google Scholar, M. H. A. Ulrich P. Cerami A. PubMed Scopus Google Scholar, J. 1995; PubMed Scopus Google Scholar, R.G. P. Hudson B.G. J. 1997; PubMed Scopus Google Scholar), on the mechanism of action of AGE These inhibitor concentrations have been because of the concentrations of glucose used to AGE formation in in vitro reactions. on the present work, however, it is likely that at concentrations in the mm higher than by autoxidation this inhibition is not with the chelex-treated phosphate buffers used in our experiments, also with phosphate buffers containing traces of mixed metal ions. results were for AG and PM in the of was less as a of ascorbate autoxidation reactions at AGE inhibitors, such as carnosine, and have metal binding activity are likely to primarily as chelators at the millimolar concentrations commonly for these reagents in the of AGE in two studies in a of AGE inhibitors were described S. S. I. Res. Commun. 2000; PubMed Scopus Google Scholar, S. S. I. Biochem. Biophys. Res. Commun. 1999; PubMed Scopus Google Scholar) are likely to inhibit AGE formation their chelation, rather than carbonyl trapping activity. For AG and PM, however, A. Biochem. 2000; PubMed Scopus Google Scholar) and J.M. A.J. Thorpe S.R. Baynes J.W. J. 2000; PubMed Scopus Google Scholar) have been as to carbonyl compounds during glycoxidation and that, in addition to their chelating activity, they also as of carbonyl compounds. experiments have to be performed with AGE inhibitors to their mechanism of action as carbonyl IC50 for inhibition of copper-catalyzed oxidation of ascorbic acid by various and AGE inhibitory concentration. OPB-9195 was less at higher and a in phosphate a in phosphate inhibitory concentration. OPB-9195 was less at higher and a in phosphate buffer. in a The estimate of the IC50 for inhibition of ascorbate oxidation is a of the chelation activity of AGE is the of constants for metal binding by the of ascorbate K. Y. J. Sci. 23: PubMed Scopus Google Scholar), the AGE inhibitor and the phosphate buffer ions in rather than an for the AGE In most for pyridoxamine Arch. Biochem. Biophys. PubMed Scopus Google Scholar) and 2000; Google Scholar), metal binding constants have not been of the of the the IC50 for an AGE inhibitor with buffer, metal and ascorbate concentration, and with the to a in the would transition metal ions than ascorbate. in the and concentrations of metal ion in buffers also the inhibitors, on their binding constants with AGE the and of chelation in studies on AGE solution is to these studies in the of a such as that the of the AGE inhibitor to chelation would be However, itself is a potent inhibitor (>99% at 1 Thorpe S.R. Baynes J.W. Diabetes. 1994; PubMed Google of AGE formation from glucose and that reaction times would be at glucose concentration. We that it may be to the of AGE inhibitors by the browning of proteins by are 10–100 times reactive than glucose with that concentrations of be of by also leads to formation of AGEs, cross-links and of the modification of by glucose G. Biophys. 1999; PubMed Scopus Google Scholar). Maillard reactions of and in the of chelators Thorpe S.R. Baynes J.W. J. Biochem. 1999; PubMed Scopus Google Scholar). on the browning of proteins by be conducted in the of chelator such as that the chelating activity of the inhibitor would not the kinetics of AGE formation in The of the chelating activity of AGE inhibitors in vivo is difficult to the inhibition of ascorbate oxidation with (Fig. is to metal-catalyzed oxidation and formation of Maillard products from glucose in and AGEs are present on in vivo T. Kurokawa K. Ypersele Strihou C. Thorpe S.R. Baynes J.W. Kidney 1998; PubMed Scopus Google Scholar). is likely that to are in with to proteins, and metal to proteins such as J. A. Free Radic. 1997; PubMed Scopus Google Scholar) or S. M. D. R. C. C. G. 2000; PubMed Scopus Google Scholar), may and of the proteins in the of Biophys. 1997; PubMed Scopus Google Scholar). binding with and chelators such as AGE inhibitors may in the by the of metal ion complexes in a such as AG or PM, not inhibit autoxidation reactions at therapeutic concentrations the of metal ions from and for in metal-catalyzed to proteins in and and 1999; Google Scholar, M. J. Free Radic. 2000; PubMed Scopus Google Scholar) have that AGEs on proteins may to transition metal to AGE inhibitors remove metal ions from these they may both inhibit AGE formation chelation carbonyl and also metal-catalyzed to from the of of and copper in AGE the therapeutic of the chelating activity of AGE chelators are not used for the of diabetic chelators have in of A. Free Radic. 1999; PubMed Scopus Google Scholar) and Diabetologia. 2001; PubMed Scopus Google Scholar) in the AGE inhibitors in their IC50 values in the ascorbate from the millimolar for AG and PM, to the for carnosine, and concentrations of of their both AG and PM were in the formation of AGEs and development of complications in animal models of diabetes (5Degenhardt T.P. Alderson N.L. Arrington D.D. Beattie R.J. Basgen J.M. Steffes M.W. Thorpe S.R. Baynes J.W. Kidney Internat. 2001; (in press)Google Scholar). the the concentration of in to mm) PubMed Scopus Google Scholar) the therapeutic concentration of μm J. 1990; PubMed Scopus Google Scholar), is comparable with IC50 in the ascorbate (25 Fig. The therapeutic concentrations of DAP and OPB-9195 have not been are also in the that chelation would contribute to their AGE inhibitory activity in The AGE-breakers and were among the most potent inhibitors of copper-catalyzed oxidation of and the hydrolysis products of these compounds chelating activity. of leads to of the a compound Biochem. 1999; PubMed Scopus Google Scholar). are ligands for copper and transition the chelating activity of the hydrolysis products of and also against as as AG S. Bucala R. Y. Res. 1999; PubMed Scopus Google Scholar), it is also that the hydrolysis products may have carbonyl trapping activity. The chelating of AGE-breakers is that they may also have some activity, the of to and of synthesized We have shown that AGE inhibitors have copper chelating activity. we have not studied in inhibitory were with phosphate buffers containing a of metal ions. The metal chelating activity of containing and most contributes to the chelating activity of AGE inhibitors in vitro and be in their mechanism of action in assays for AGE inhibition for metal chelation by with inhibitor concentrations their IC50 or by using as and of products by the AGE inhibitors is also for their mechanism of action. the chelating carbonyl trapping activity of AGE inhibitors is for the mechanism of action of these and the of chelation in and for inhibitors of diabetic complications. We of for the of phenacylthiazolium bromide and phenacyldimethylthiazolium