What is the effect of hyperglycemia on atherogenesis and can it be reversed by aminoguanidine?

ELSEVIER

Diabetes Research and Clinical Practice 30 Suppl. (1996) S 123-S 130

What is the effect of hyperglycemia on atherogenesis and can it be reversed by aminoguanidine?l Richard Bucala* The Picower Institute for Medical Research 350 Community Drive, Manhasset, NY 11030, USA

Abstract

Reducing sugars such as glucose react non-enzymatically with the amino groups of proteins and lipids to initiate a chemical modification pathway known as advanced glycosylation. Recent progress in our understanding of this process has affirmed the hypothesis that advanced glycosylation endproducts (AGEs) play an important role in the evolution of both diabetic and non-diabetic vascular disease. Utilizing newly developed AGE-specific ELISA techniques, AGEs have been identified to be present on a variety of vascular wall, lipoprotein, and lipid constituents. Vascular wall AGEs contribute to vascular pathology by acting to increase vascular permeability, enhance subintimal protein and lipoprotein deposition, and inactivate the endothelium-derived relaxing factor, nitric oxide. Lipid-linked AGEs also have been shown to initiate oxidative modification, thus promoting the formation of oxidized low-density lipoprotein. AGE-specific ELISA analysis has demonstrated a significantly increased level of AGE-modified LDL in the plasma of diabetic patients when compared to normal controls. Furthermore, LDL which has been modified by advanced glycosylation exhibits markedly impaired clearance kinetics in vivo. Thus, AGE-modification impairs LDL-receptor-mediated clearance mechanisms and contributes to elevated LDL levels in patients with diabetes. This concept has been substantiated recently by the clinical observation that administration of the advanced glycosylation inhibitor aminoguanidine to diabetic patients significantly decreases circulating LDL levels. Keywords: Atherosclerosis; Diabetes; Advanced glycosylation; Advanced glycosylation endproduct (AGE); Oxidation

1. I n t r o d u c t i o n

Diabetic patients experience a 3-4-fold increased risk for developing atherosclerosis and vascular insufficiency when compared to the general population. Since diabetes afflicts at least 10 million individuals in the United States alone, the contribution of hyperglycemia and insulin resistance to the overall mortal* Tel.: +1 516 3655090; fax: +1 516 3654200; e-mail: [email protected] I For Discussion, see p. 157.

ity of both heart and cerebrovascular disease is considerable [ 1-3]. Among the pathological processes which appear to be central in the development of atherosclerosis are acquired chemical modifications that affect the function of low-density lipoprotein (LDL) [4]. A dyslipidemia characterized by increased levels of LDL, VLDL, and IDL is common in diabetics and significantly increases the possibility that these patients will suffer from the atherosclerotic complications of heart attack and stroke [5,6]. Lipoproteins isolated from diabetic plasma also exhibit import-

0168-8227/96/$15.00 © 1996 Elsevier Science Ireland Ltd. All rights reserved S S D I 0304-3835(95)01 181-1

124

R. Bucala / Diabetes Research and Clinical Practice 30 Suppl. (1996) S123-$130

ant qualitative abnormalities, such as an impaired capacity to be taken up by fibroblast or tissue LDL receptors [7,8]. The precise molecular basis for this uptake abnormality is an area of considerable interest since elevation in the plasma levels of either LDL or its metabolic precursor, VLDL, is an important risk factor for the development of atherosclerosis [13,9]. 2. Oxidative modification One type of chemical modification that has been proposed to render the LDL particle more atherogenic is 'oxidative modification'. LDL that has been oxidized in vitro by exposure to transition metals exhibits diminished recognition by cellular LDL receptors and preferential uptake by macrophage 'scavenger' receptors. As a result of this process, it has been proposed that vascular wall macrophages become transformed into lipid-laden 'foam' cells, leading to the development of fatty streaks and the complex, proliferative lesions that characterize atherosclerotic plaque [4,9]. Over the years it has been established that unsaturated fatty acids (i.e. fatty acids bearing one or more double bonds in their carbon backbone) readily undergo peroxidative degradation. Peroxidation occurs when an unspecified oxidant abstracts the relatively labile allylic hydrogen atom from an unsaturated fatty acid side chain. The addition of molecular oxygen then generates a lipid peroxyl radical. This serves to propagate oxidative degradation by initiating hydrogen atom abstraction at additional unsaturated bonds. Fatty acid peroxidation and breakdown then ensues, rapidly producing shorter chain, ct,flunsaturated aldehydes such as the hydroxyalkenals. These species in turn can react with nucleophilic sites on proteins to form Michael addition products and resonance-stabilized Schiff bases [10-13]. Although transition metal-catalyzed lipid peroxidation occurs readily in vitro and is studied frequently as a model for the 'propagative' phase of oxidative degradation, it should be noted that the identity of the oxidant(s) responsible for initiating lipid peroxidation in vivo remains unsettled. This has led to significant conceptual difficulties in assessing the prevalence and the pathological significance of lipid peroxidation, particularly with respect to atherogenesis [14].

Transition metals catalyze oxidative chemistry by lowering the energy barrier for electron transfer between two substrates. In effect, the unoccupied, outer d orbitals of transition metals act as molecular 'wires' to facilitate the transfer of electrons between lipids and molecular oxygen which, in the absence of metals, is insufficiently reactive toward carboncarbon double bonds. Although oxygen is frequently invoked as the molecule responsible for producing oxidative damage in vivo, it is in fact a poor oxidant. Ground state or triplet oxygen (357g) is incapable of reacting with molecules of singlet multiplicity such as unsaturated fatty acids. Although significant amounts (0.1-1.0 /zM) of the superoxide anion (02-) do form in vivo, this species also is incapable of abstracting bis-allylic hydrogen atoms from unsaturated aliphatic residues. Under acidic conditions, superoxide may initiate oxidative modification by forming perhydroxyl radicals (HOO'), or by reacting with transition metals to form reactive hydroxyl radicals (HO') [10-14]. Nevertheless, low trace metal concentrations and the abundant anti-oxidant capacity of plasma provide strong evidence against the possibility that metal-catalyzed oxidation plays a significant role in mediating lipid peroxidation in vivo [15-17]. 3. Advanced glycosylation In diabetes, persistent hyperglycemia leads to an increase in the level of non-enzymatically glycosylated or 'glycated' protein amino groups [18,19]. This post-translational modification process is initiated by the attachment of glucose residues to primary amino groups and proceeds from freely reversible Schiff bases to more stable, slowly reversible Amadori rearrangement products (Fig. 1). Amadori products were considered several years ago to be a possible source of apolipoprotein modification that might alter LDL metabolism in vivo. Increased amounts of Amadori products in fact have been found to be present in LDL isolated from diabetic patients than in LDL from non-diabetic patients [20]. Early glycosylation products such as the Schiff base or the Amadori product are slowly reversible however, and these adducts do not affect LDL clearance unless the ketoamine linkage has been first stabilized [20]. Although this can be achieved ex vivo by chemical re-

R. Bucala / Diabetes Research and Clinical Practice 30 Suppl. (1996) $123-$130

TIME:

GLUCOSE + NH2-R

O ,'/ C I (CHOH) 4 ~ NH2-R I CH2OH

hours K1 ~,

~-

SCHIFF BASE

H X

H \ ""-

N+H-R // C I (CHOH)4

days K2 ~ ~

~

~-

AMADORI PRODUCT

"~"

CH20 H

CH2-NH-R I C=O I (CHOH)3

weeks, months Kn . ~ ~

~

""

~

"~"

CH20 H

125

,

ADVANCED GLYCOSYLATION ENDPRODUCTS

AFGP PYRRALINE

FFI PENTOSIDINE

CML

CROSSLINES

Fig. I. Pathwayfor the formationof AGEs. duction with borohydride, reactions of this type do not occur in vivo. Instead, Amadori products either dissociate to yield free glucose, or else slowly undergo additional rearrangement reactions to produce late-forming, irreversibly bound moieties termed advanced glycosylation endproducts (AGEs) [ 18,19]. AGEs alter the structural and functional properties of proteins and many of these alterations have been linked to the pathophysiological changes that accompany both long-standing diabetes and normal aging. AGEs progressively crosslink connective tissue collagen for example, and increase connective tissue rigidity [21,22]. Collagen-linked AGEs also act as reactive 'foci' to covalently trap circulating serum proteins such as albumin, lipoproteins, and immunoglobulins [23,241. This accounts in part for the increase in protein deposition and in basement membrane thickening that occurs in the systemic and renal vasculature of diabetic patients. Cell surface receptors that are specific for the recognition, uptake, and degradation of AGE-modified proteins have been identified on circulating monocytes, endothelial cells, and renal mesangial cells 125-27]. AGEs are chemotactic for monocytes, and the receptor-mediated uptake of AGE-modified proteins initiates cytokinemediated processes that promote tissue remodelling [28]. Occupancy of endothelial cell receptors by AGEs leads to increased vascular permeability, down-regulation of the anticoagulant factor thrombomodulin, and increased synthesis of the procoagulant tissue factor [26]. As discussed below, recent studies also have established an important role for AGE-mediated re-dox reactions in vivo. Advanced glycosylation proceeds by a successive series of rearrangement, fragmenta-

tion, and intramolecular oxidation-reduction reactions involving reactive carbonyl compounds and nucleophilic amines. The possibility that AGEs exhibit significant red-ox activity in vivo was first suggested by studies showing that AGEs can inactivate, by direct chemical reaction, the endothelium-derived relaxing factor nitric oxide (NO) [29,30]. In vitro studies showed that the addition of increasing amounts of either lysine-derived or protein-derived AGEs to a solution of authentic nitric oxide produced an apparent first-order inactivation curve for nitric oxide. Invasive blood pressure monitoring showed that animals with experimentally induced diabetes exhibit a defect in nitric oxide-dependent vasodilatory responses that develops over a time course which is consistent with AGE formation and not glycemic fluxes [29]. Furthermore, when the advanced glycosylation inhibitor aminoguanidine was administered either to diabetic animals or to animals who had been given AGEs intravascularly, there was a significant sparing effect on the progression of vasodilatory impairment [29,31]. These data thus were consistent with the concept that nitric oxide activity is modulated in part by acquired AGE modifications which occur on collagen and other vascular wall proteins (Fig. 2).

4. Lipid-advanced glycosylation and oxidative modification Our investigations in the advanced glycosylation of iipids and lipoproteins began several years ago after considering the possibility that phospholipids which contain primary amino groups could react with glucose to form AGEs in much the same way that

126

R. Bucala / Diabetes Research and Clinical Practice 30 Suppl. (1996) $123-$130

cence, and immunoreactive properties as the AGEs that form on proteins. We hypothesized that AGE-mediated oxidationreduction reactions might act within the hydrophobic microenvironment of phospholipids to initiate fatty acid oxidation [32]. Lipid oxidation thus could be initiated independently of transition metals or exogenous, free-radical generating systems and might serve as an important mechanism for initiating lipid oxidation in vivo. This hypothesis was supported by experimental studies which showed that lipid oxidation occurred concomitantly with AGE formation, in the presence of metal chelators, and in the absence of exogenously-added transition metals. Of note, lipid oxidation did not occur when glucose was incubated with PC, indicating that the reaction between glucose and the phospholipid amine was essential for initiating oxidative modification. In further studies, the addition of the advanced glycosylation inhibitor aminoguanidine prevented the formation of both lipidAGEs and lipid oxidation products, verifying the essential role of advanced glycosylation in the initiation of lipid oxidation [32]. Fig. 3 shows an overall scheme for the oxidative modification of fatty acid residues by lipid-linked AGEs. The contribution of advanced glycosylation to LDL oxidation was examined in a similar fashion in

~~Vasodllatlon Fig. 2. Modcl relating the inactivation of NO by sub-endothelial bound AGEs. Nitric oxide is produced enzymaticallyby endothelial cells after stimulation by acetylcholine (ACh), and nonenzymatically by nitrosovasodilatorspro-drugs such as nitroglycerin (NTG). Inset shows an arteriole in cross-section.

polypeptide amines form AGEs [32]. In model studies, buffered suspensions of phosphatidylethanolamine (PE) or phosphatidyicholine (PC) were incubated with glucose and the metal chelator EDTA at 37°C. PE but not PC (which contains a blocked, quaternary amine) was observed to react with glucose to form products with the same absorbance, fluores-

0

NH 2

glucose

O / R

O

II

II

II

0

0

o

0

0

•~

/ R

0

AGE

I R

AGE y

0

O

CH PE

CH 3

3

CH 3 PE-AG E

Fig. 3. Scheme for the formation of AGEs on phospholipid head groups, followed by AGE-initiated oxidative damage to fatty acid side chains. PE, phosphatidylethanolamine;PE-AGE,AGE-modifiedphosphatidylethanolamine;R, unspecified fatty acid side chain.

R. Bucala / Diabetes Research and Clinical Practice 30 Suppl. (1996) $123-S130

vitro. Purified human LDL was incubated with glucose (in the presence of metal chelators) and analyzed for both advanced glycosylation and oxidative modification. Incubation of LDL with 200 mM glucose for 3 days resulted in the formation of readily measurable levels of AGEs on both lipid and apoprotein. These studies also indicated that lipid-linked AGEs formed more rapidly than ApoB-AGEs, reaching a specific activity 100-fold greater than the ApoB-linked AGEs. These data were consistent with a previously described enhanced rate of advanced glycosylation in the lipid versus the polar, apoprotein phases. Measurements of oxidative modification further showed that LDL was oxidized concomitantly with the formation of AGEs [32]. Although thiobarbituric acid is widely used to test for lipid oxidation, the thiobarbituric acid reaction is relatively non-specific and can produce chromophores after reaction with a number of possible aldehydes. In more recent studies, gas chromatographyelectron impact mass spectrometry analyses have identified specific oxidation products which form during either phospholipid- or LDL-advanced glycosylation. One product formed mass ions which were consistent with the structure of 4-hydroxyhexenal, an unsaturated fatty acid oxidation product which forms by the oxidation of docosahexaenoic acid, the most prevalent fatty acid of the n-3 series [33]. These data provide further structural support for the concept that lipid-advanced glycosylation reactions can act directly to oxidize unsaturated fatty acids. The recent identification of an AGE-reactive intermediate that is trapped by reaction with the advanced glycosylation inhibitor aminoguanidine has provided an important conceptual framework for understanding how lipid-linked AGEs might act to initiate the oxidation of fatty acid side chains. Two years ago, an aminoguanidine-derived triazine was isolated from a solution of aminoguanidine and a propylamino-substituted Amadori product which were incubated under physiological conditions [34]. The formation of the triazine was shown to proceed by the reaction of aminoguanidine with the reactive intermediate: 1-propylamino-i,4-dideoxyosone. The amine-substituted dideoxyosone is an interesting compound from the viewpoint that it may lead to pro-oxidant effects during advanced glycosylation reactions. Once attached to phospholipid head

127

groups, dideoxyosones and related AGEs may interact with fatty acid side chains to initiate red-ox cycling directly within the hydrophobic, lipid microenvironment.

5. Lipoprotein (LDL) advanced glycosylation To further define the relationship between advanced glycosylation and LDL oxidation in vivo, we analyzed LDL from both non-diabetic and diabetic individuals by AGE-specific ELISA for the presence of phospholipid-AGEs and ApoB-AGEs [35]. LDLadvanced glycosylation also was examined in patients with diabetes and renal insufficiency. Reactive AGEs circulate in high concentrations during renal insufficiency and these moieties can continue to react with plasma components such as LDL [36,37]. There was a significant elevation in the levels of ApoBAGE and lipid-AGE in LDL when each of the patient groups was compared to the control (non-diabetic/ normal renal function) group. The highest levels of AGE-modification were observed in patients with renal insufficiency, pointing to the important role of circulating, reactive AGE-peptides in producing AGE-modified LDL. LDL from diabetic or renal insufficient individuals also showed significantly greater oxidative modification than the LDL from control, non-diabetic individuals. Although consistent with previous studies suggesting an increase in lipid oxidation in diabetes [38,39], it remains uncertain whether these measurements of LDL oxidation reflect an actual increase in lipid oxidation in vivo, or simply an increase in the susceptibility of LDL to oxidation once LDL is removed from the plasma for biochemical analysis. Nevertheless, linear regression analysis of these data revealed that there was a significant correlation between the level of AGE modification and LDL oxidation [32]. Chemical modification of basic residues within the LDL-receptor binding domain of ApoB has been shown to interfere with the ability of LDL to undergo receptor-mediated uptake and degradation [40,41]. Advanced glycosylation similarly modifies the lysine and arginine residues of proteins, and it was reasoned that LDL modified by AGEs might exhibit markedly delayed clearance kinetics in vivo. When the plasma clearance of AGE-LDL (modified in vitro) was examined in transgenic mice expressing the human

128

R. Bucala / Diabetes Research and Clinical Practice 30 Suppl. (1996) $123-$130

LDL-receptor, markedly delayed clearance kinetics were observed when AGE-LDL was compared to control, native LDL (Fig. 4) [35]. It is important to note that the level of AGE-modification in these experiments was comparable to that observed in diabetic/ESRD patients in vivo (-80 U AGE/mg ApoB). Overall, there was a significant relationship between the extent of ApoB-advanced glycosylation and impaired plasma clearance of LDL in vivo [35].

with a 19% decrease in total cholesterol, a 19% decrease in triglycerides, and a 28% decrease in LDLcholesterol. Taken together, these data provide strong support for the concept that lipoprotein advanced glycosylation plays an important role in inhibiting normal pathways of LDL clearance in vivo.

7. Conclusions The elevated circulating level of AGE-modified

6. Lowering of plasma LDL in human subjects by aminoguanidine To begin to assess the contribution of advanced glycosylation to altered LDL clearance kinetics in human subjects, we examined the lipoprotein profiles of diabetic patients who were enrolled in a 28-day, double-bind placebo-controlled trial of aminoguanidine [35]. Eighteen patients received aminoguanidine at an average daily dose of 1200 mg and 8 patients received placebo. Blood samples were obtained at the initiation and at the termination of treatment and analyzed for total cholesterol, triglycerides, VLDL-cholesterol, LDL-cholesteroi, HDLcholesterol, Hb-AGE, and HbAlc (Table 1). The efficacy of aminoguanidine as an inhibitor of advanced glycosylation in this study was verified by the observation that circulating Hb-AGE levels decreased by almost 28% in the aminoguanidine-treated group [42]. Aminoguanidine therapy was also associated

A.

14. Non-Transgenlc

0

Mice

;o

1

200 '

300

Time (minutes)

B. o 3

iz 14.

Fig. 4. Plasma clearance of native and AGE-LDL in control, nontransgenic mice and transgenic mice expressing the human LDLreceptor. (A) Control (non-transgenic) mice injected with native LDL (O) or AGE-LDL (e). (B) Mice transgenic for the human LDL receptor injected with native LDL (O) or AGE-LDL (e). The ratio of AGE-LDL to native LDL was calculated and averaged for all mice at each time point. The mean clearance ratio of AGE-LDL to native-LDL in the transgenic mice was determined to be 1.35 ± 0.03 (P < 0.001 by one-way ANOVA). (C) Relationship between the extent of ApoB-AGEs and the AGE-LDL/ native-LDL clearance ratio. ApoB-AGEs were measured by ELISA in 4 different preparations of LDL that were subjected to plasma clearance studies. (O) Control (native) LDL (I.2U AGE/mg ApoB); (&) LDL modified with synthetic AGE-peptides in the presence of 800 mM aminoguanidine, yielding 2.0 U AGE/mg ApoB. (V) LDL modified with synthetic AGE-peptide in the presence of 400 mM aminoguanidine, yielding 8.0 U AGE/mg ApoB. (,~) LDL modified with AGE-peptide alone (80 U AGE/rag APOB) (from Ref. [35]).

Transgenlc Mice

0

i

L

100

200

300

Time (minutes)

C.

1.5 --I

<

1.0 1()

ApoB-AGE (U AGE/mg protein)

100

129

R. Bucala / Diabetes Research and Clinical Practice 30 Suppl. (1996) $123-$130

Table I Biochemical analysis of blood specimens obtained from diabetic patients who received aminoguanidine (n = 18) or placebo control (n = 8) for 28 days Treatment

Cholesterol

Triglyceride

VLDL

LDL

HDL

HbAGE

HbAlc

Aminoguanidine P value

81.3 ± 7.2 <0.04

81.0 ± 6.2 <0.02

68.5 +_28.7 NS

71.9 ± 9.9 <0.05

104.7 ± 10.9 NS

72.7 ± 7.5 <0.025

89.7 +_4.2 NS

Placebo P value

97.4±5.4 NS

89.8__.5.8 NS

96.4±7.1 NS

100.7±11.2 NS

96.7+_16.4 NS

90.8__.6.7 NS

100.0±4.5 NS

Values are expressed as percent (mean ± SE) of baseline value for each patient group (day 28 value/day 0 value × 100). P values are by paired Student's t-test. NS, not significant (from Ref. [35]). L D L in diabetic patients together with the impaired plasma c l e a r a n c e kinetics o f A G E - L D L point to the important, contributory role o f L D L - a d v a n c e d glycosylation in diabetic atherogenesis. A decrease in rec e p t o r - m e d i a t e d c l e a r a n c e o f A G E - L D L is likely to act in c o n c e r t with abnormalities in lipoprotein production, increases in lipoprotein oxidation and vascular wall lipoprotein trapping, and alterations in endothelial cell function to p r o d u c e the rapidly progressive vasculopathy o f diabetes or renal insufficiency. L o w e r levels o f A G E - m o d i f i e d L D L occur in nondiabetic/non-renally impaired individuals as well and o v e r a t i m e period o f m a n y years, a d v a n c e d g l y c o s y lation also may contribute to the age-related develo p m e n t o f atherosclerosis in the general population.

References [1] Kannel, W.B. and McGee, D.L. (1979) Diabetes and cardiovascular disease. J. Am. Med. Assoc. 241, 2035-2038. [2] Ruderman, N.B. and Haudenschild, C. (1984) Diabetes as an atherogenic factor. Prog. Cardiovasc. Dis. 26, 373-412. [3] WHO Study Group (1985) Diabetes Mellitus. WIIO Tech. Rep. Ser. 727, 1-113. [4] Steinberg, D., Parthasarathy, S., Carew, T.E., Khoo, J.C. and Witztum, J.L. (1989) Modifications of low-density lipoprotein that increases its atherogenicity. Mech. Dis. 320, 915924. [5] Jensen, T., Stender, S. and Deckert, T. (1988) Abnormalities in plasma concentrations of lipoproteins and fibrinogen in type I diabetic patients with increased urinary albumin excretion. Diabetologia 31,142-145. [6] Brown, W.V. (1994) Lipoprotein disorders in diabetes mellitus. Med. Clin. N. Am. 78, 143-161. [7] Lopes-Virella, M.F., Sherer, G.K., lees, A.M., Wohitmann, H., Mayfield, R., Sagel, J., LeRoy, E.C. and Colwell, J.A. (1982) Diabetologia 22, 430-436. [8] Hiramatsu, K., Bierman, E.L. and Chait, A. (1985) Metabo-

[9]

[10]

[11]

[12] [ 13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

lism of low-density lipoprotein from patients with diabetic hypertriglyceridemia by cultured human skin fibroblasts. Diabetes 34, 8-14. Goldstein, J.L. and Brown, M.S. (1977) The low-density lipoprotein pathway and its relation to atherosclerosis. Annu. Rev. Biochem. 46, 897-930. Kanner, J., German, J.B. and Kinsella, J.E. (1987) Initiation of Jipid peroxidation in biological systems. Crit. Rev. Food Sci. Nutr. 25, 317-365. E,sterbauer, H., Sehaur, R.J. and Zollner, H. (1991) Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Rad. Biol. Med. I 1, 81-128. Halliwell, B.M. and Guneridge, J.M.C. (1989) Free Radicals in Biology and Medicine. Clarendon Press, Oxford. Bielski, B.H.J and Cabelli, D.E. (1991) Highlights of current research involving superoxide and perhydroxyl radicals in aqueous solutions. Int. J. Radiat. Biol. 59, 291-319. Dix, T.A. and Aikens, J. (1992) Mechanisms and biological relevance of lipid peroxidation initiation. Chem. Res. Toxicol. 6, 2-18. Klaasen, C.D. (1985) In: A.G. Gillman, L.S. Goodman, T.W. Rail and F. Murad (Eds.), The Pharmacological Basis of Therapeutics. Macmillan, New York, pp. 1605-1627. Frei, B., Yamamoto, Y., Niclas, D. and Ames, B.N. (1988) Evaluation of an isoluminol chemiluminescence assay for the detection of hydroperoxide in human blood plasma. Anal. Biochem 175, 120-130. Frei, B., Stocker, R. and Ames, B.N. (1988) Antioxidant defenses and lipid peroxidation in human blood plasma. Proc. Natl. Acad. Sci. USA 85, 9748-9752. Njoroge, F.G. and Monnier, V.M. (1989) The chemistry of the Maillard reaction under physiological conditions: a review. Prog. Clin. Biol. Res. 304, 85-107. Bucala, R. and Cerami, A. (1992) Advanced glycosylation: chemistry, biology, and implications for diabetes and aging. Adv. Pharmacol. 23, 1-34. Witztum, J.L., Mahoney, E.M., Branks, M.J., Fisher, M., Elam, R. and Steinberg, D. (1982) Nonenzymatic glycosylation of low-density lipoprotein alters its biologic activity. Diabetes 31,283-291. Schnider, S.L. and Kohn, R.R. (1982) Effects of age and diabetes mellitus on the solubility of collagen from human

130

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

R. Bucala /Diabetes Research and Clinical Practice 30 Suppl. (1996) $123-$130

skin, tracheal cartilage and dura mater. Exp. Gerontol. 17, 185-194. Monnier, V.M., Kohn, R.R. and Cerami, A. (1984) Accelerated age-related browning of human collagen in diabetes mellitus. Proc. Natl. Acad. Sci. USA 81,583-587. Brownlee, M., Pongor, S. and Cerami, A. (1983) Covalent attachment of soluble proteins by nonenzymatically glycosylated collagen; role in the in situ formation of immune complexes. J. Exp. Med. 158,1739-1744. Brownlee, M., Vlassara, H. and Cerami, A. (1985) Nonenzymatic glycosylation products on collagen covalently trap low-density lipoprotein. Diabetes 34, 938-941. Vlassara, H., Brownlee, M. and Cerami, A. (1985) Highaffinity receptor-mediated uptake and degradation of glucose-modified proteins: a potential mechanism for the removal of senescent macromolecules. Proc. Natl. Acad. Sci. USA 82, 5588-5592. Esposito, C., Gerlach, H., Brett, J., Stem, D. and Vlassara, H. (I 989) Endothelial receptor-mediated binding of glucosemodified albumin is associated with increased monolayer permeability and modulation of cell surface coagulant properties. J. Exp. Med. 170, 1387-1407. Skolnik, E.Y., Yang, Z., Makita, Z., Radoff, S., Kirstein, M. and Vlassara, H. (1991) Human and rat mesangial cell receptors for glucose-modified proteins: potential role in kidney tissue remodelling and diabetic nephropathy. J. Exp. Med. 174, 931-938. Vlassara, H., Brownlee, M., Manogue, K, Dinarello, C. and Cerami, A. (1988) Cachectin/TNF and IL-I induced by glucose-modified proteins: role in normal tissue remodeling. Science 240, 1546-1548. Bucala, R., Tracey, K.J. and Cerami, A. (1991) Advanced glycosylation products quench nitric oxide and mediate defective endothelium-dependent vasodilation in experimental diabetes. J. Clin. Invest. 87,432-438. Hogan, M., Cerami, A. and Bucala, R. (1992) Advanced glycosylation endproducts block the antiproliferative effect of nitric oxide. J. Clin. Invest. 90, 1110-1115. Vlassara, H., Fuh, H., Makita, Z. Krungkrai, S., Cerami, A. and Bucala, R. (1992) Exogenous advanced glycosylation endproducts induce complex vascular dysfunction in normal animals: a model for diabetic and aging complications. Proc. Natl. Acad. Sci. USA 89, 12043-12047. Bucala, R., Makita, Z., Koschinsky, T., Cerami, A. and Vlassara, H. (1993) Lipid advanced glycosylation: Pathway

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

for lipid oxidation in vivo. Proc. Natl. Acad. Sci. USA 90, 6434-6438. van Kuijk, F.J., Thomas, D.W., Stephens, R.J. and Dratz, E.A. (1986) Occurrence of 4-hydroxyalkenals in rat tissues determined as pentafluorobenzyl oxime derivatives by gas chromatography-mass spectrometry. Biochem. Biophys. Res. Commun. 139, 144-149. Chert, H.J.C. and Cerami, A. (1993) Mechanism of inhibition of advanced glycosylation by aminoguanidine in vitro. J. Carbohydr. Chem. 12, 731-742. Bucala, R., Makita, Z., Vega, G., Grundy, S., Koschinsky, T., Cerami, A. and Vlassara, H. (1994) Modification of LDL by advanced glycosylation endproducts contributes to the dyslipidemia of diabetes and renal insufficiency. Proc. Natl. Acad. Sci. USA 91,9441-9445. Makita, Z., Radoff, S., Rayfield, E., Yang, Z., Skolnik, E., Delaney, V., Friedman, E.A. and Vlassara, H. (1991) Advanced glycosylation endproducts in patients with diabetic nephropathy. N. Engl. J. Med. 325, 836-842. Makita, Z., Bucala, R., Rayfield, E.J., Friedman, E.A., Kaufman, A.M., Korbet, S.M., Barth, R.H., Winston, J.A., Fuh, H., Manogue, K., Cemmi, A. and Vlassara, H. (1994) l)iabetic-uremic serum advanced glycosylation end products are chemically reactive and resistant to dialysis therapy. Lancet 343, 1519-1522. Nishigaki, I., Hagihara, M., Tsunekawa, H., Maseki, M. and Yagi, K. (1981) Lipid peroxide levels of serum lipoprotein fractions of diabetic patients. Biochem. Med. 25, 373-378. Morel, D.W. and Chisolm, G. (1989) Antioxidant treatment of diabetic rats inhibits lipoprotein oxidation and cytotoxicity. J. Lipid Res. 30, 1827-1834. Mahley, R.W., lnnerarity, T.L., Pitas, R.E., Weisgraber, KH., Brown, J.H. and Gross, E. (1977) Inhibition of lipoprotein binding to cell surface receptors of fibroblasts following selective modification of arginyl residues in arginine-rich and B apoproteins. J. Biol. Chem. 252, 7279-7287. Mahley, R.W., lnnerarity, T.L, Weisgraber, K.H. and Oh, S.Y. (1979) Altered metabolism (in vivo and in vitro) of plasma lipoproteins after selective chemical modification. J. Clin. Invest. 64, 743-750. Makita, Z., Vlassara, H., Rayfield, E., Cartwright, K, Friedman, E., Rodby, R., Cerami, A. and Bucala, R. (1992) Hemoglobin-AGE: a circulating marker of advanced glycosylation. Science 258, 651-653.