Advanced Glycation End Product Interventions Reduce Diabetes-Accelerated Atherosclerosis

Six-week-old homozygous apoE-deficient male mice (backcrossed 20 times from C57BL/6 background; Animal Resource Centre, Canning Vale, WA, Australia) were studied in accordance with guidelines provided by the Alfred and Baker Animal Ethics Committee. Mice were rendered diabetic by daily injection of streptozotocin at a dosage of 55 mg/kg for 5 consecutive days. Control animals were given daily injections of vehicle (citrate buffer) alone. Groups of diabetic animals (n = 20/group) were then randomized after 1 week of recovery to receive 1) no treatment; 2) AGE cross-link breaker ALT-711 [4,5-dimethyl-3-(2-oxo2-phenylethyl)-thiazolium chloride; Alteon, Ramsey, NJ], gavaged at 10 mg · kg⁻¹ · day⁻¹ (DALT); or 3) the AGE formation inhibitor AG hydrocarbonate (Fluka Pharmaceuticals) at 200 mg · kg⁻¹ · day⁻¹ (DAG; 1 g/l in drinking water). All groups of animals were followed for 20 weeks. Mice were allowed access to standard mouse diet and water ad libitum. An initial pilot study using two doses of ALT-711 (gavaged at 10 or 20 mg · kg⁻¹ · day⁻¹) revealed no differences between the two doses in terms of the primary end point of plaque accumulation or aortic AGE accumulation. Therefore, the definitive study involving comparison with AG was performed using 10 mg · kg⁻¹ · day⁻¹ ALT-711.

At the conclusion of the study, glycated Hb (GHb) was measured by high-performance liquid chromatography (17). Total cholesterol, HDL, and triglyceride concentrations were measured by autoanalyzer. LDL cholesterol was calculated by the Friedewald formula (18). Systolic blood pressure was assessed by a computerized, noninvasive tail-cuff technique in conscious mice (19).

After 20 weeks, after anesthesia with intraperitoneal pentobarbital sodium (60 mg/kg body wt; Nembutal, Boehringer Ingelheim, Mannheim, Germany), aortas were rapidly dissected and snap-frozen in liquid nitrogen. A subset of aortas from each group (n = 10 mice per group) were removed and fixed in neutral-buffered formalin for subsequent evaluation of lesions.

Two approaches were used to evaluate the atherosclerotic lesions as described previously (16). First, macroscopic assessment of en face whole aorta segments was performed. The entire aorta was cleaned of fat, opened longitudinally, and dissected into three segments: aortic arch, descending thoracic aorta, and abdominal aorta. After staining with Sudan IV-Herxheimer’s solution (Sigma Chemical Co, Carpinteria, CA), aortas were pinned out flat, and images were digitized on a dissecting microscope (Olympus SZX9; Olympus Optics, Tokyo, Japan) equipped with a high-resolution camera (Zeiss Axiocam; Zeiss Instruments, Gottingen, Germany). The digitized images were evaluated, and the lesion area was evaluated by calculation of the proportion of aortic intimal surface occupied by red stain in each of the three aortic segments using Image SXM program software (National Institutes of Health, Bethesda, MD).

The aortic segments studied above were then embedded in paraffin on end and cross-sectioned at 4 μm for subsequent immunohistochemistry, in situ hybridization, and histological evaluation. Secondary histological assessment of lesions was performed on sections stained with Masson’s Trichrome as described previously (16). Total collagen was quantified by calculation of the area of aniline blue staining occupied within the plaques by Optimas 6.2 Software (Optimas 6.2; Video Pro-32, Bedford Park, SA, Australia) associated with a JVC video camera and Olympus microscope.

Fluorescence was assayed in duplicate 20-μl plasma aliquots as described previously (13). Characteristic fluorescence (excitation 370 nm, emission 440 nm) has been previously attributed to the accumulation of AGE cross-links (20,21). Briefly, samples were deproteinated and delipidated by addition of 0.15 mol/l trichloroacetic acid and chloroform (22), followed by centrifugation and removal of the upper aqueous phase. Fluorescence of the supernatant was determined (excitation 370 nm, emission 440 nm) using an on-line high-performance liquid chromatography injector (Waters, Milford, MA) and expressed per arbitrary protein unit at 280 nm.

Mouse skin collagen was prepared as described previously for rat skin collagen (8). In brief, insoluble collagen was isolated from 1.5-cm pieces of skin after removal of fat and hair with a razor blade and subsequent sequential extractions with 1.0 mol/l NaCl, 0.5 mol/l acetic acid, and delipidation with chloroform:methanol (1:2). The collagen was then lyophilized and stored at −20°C until analyzed for AGE/ALE content. The AGEs/ALEs Nε(carboxymethyl)lysine (CML) and Nε(carboxyethyl)lysine (CEL) were quantified by isotope dilution and selected ion monitoring gas chromatography–mass spectrometry (23) and normalized to their parent amino acid lysine.

Skin collagen (3 mg), prepared as above, was used for cross-linking analysis and added to 7.5 ml of 0.5 mmol/l acetic acid containing 10 μmol/l pepsin (Sigma Chemical Co.) and incubated at 37°C with gentle agitation (60 Hz). Aliquots of 250 μl were removed at 0, 4, 8, 12, 16, and 24 h of digestion. The conditions were chosen on the basis of preliminary studies that showed significant solubilization of collagen from nondiabetic animals at 1 h and maximum solubilization (>98%) after 24 h. The collagen content of these samples was determined after acid hydrolysis (6 mol/l HCl for 24 h at 110°C) using the procedure of Stegemann and Stadler (24). Collagen solubility was expressed as a function of time with the 0 time point providing 0% digestion and the 24-h time point providing 100% digestion. The collagen content at each time point of pepsin digestion was expressed as a percentage of the total recoverable collagen after 24 h. The time taken for digestion of 50% of the collagen was determined for each of the groups using derivatization of area under the curve.

Skin fluorescence (at excitation 370 nm, emission 440 nm) was determined in 24-h acid-hydrolyzed collagen preparations in the presence of 0.1 mol/l NaBH₄, via a flow injection system adapted from Wrobel et al. (22) using a Waters 470 spectrophotometer/fluorometer (Waters). For correcting for differences in sample collagen content, fluorescence intensity was expressed as arbitrary units of fluorescence per milligram of collagen.

A modification of the ABC Ig enzyme bridge technique (25) was used for immunohistochemistry. Formalin-fixed paraffin sections of aortas from previous en face studies were dewaxed and hydrated. Antigen retrieval by microwaving in citrate buffer (1.8 mmol/l citric acid, 8.2 mmol/l trisodium citrate) twice for 5 min was performed for connective tissue growth factor (1:500; rabbit anti-rCTGF; gift of Dr. S. Twigg, Department of Medicine, University of Sydney) (26) and transforming growth factor-β (TGF-β; 1:250; Santa Cruz Biotechnology, Santa Cruz, CA). Slides for collagen immunostaining were porcine pepsin digested (0.4%) in 0.5% nonfat milk powder diluted in Tris-buffered saline containing 0.05% Tween 20. After incubation with 0.3% hydrogen peroxide for 20 min, sections were incubated with protein-blocking agent (Lipshaw, Pittsburgh, PA) for 20 min and then incubated with primary antibody overnight at 4°C. CML is the major epitope recognized by the monoclonal anti-AGE antibody, 4G9, used in this protocol (1:500, Alteon) (27). Other primary antibodies used were polyclonal goat anti-RAGE (1:250; Chemicon, Temecula, CA); goat anti-human collagen I, III, and IV (1:80, 1:75, and 1:800, respectively; Southern Biotechnologies, Birmingham, AL); and mouse anti-human α-smooth muscle actin (α-SMA; 1:50; Dako).

Tissue sections were stained consecutively with biotinylated IgG for 10 min and avidin-biotin horseradish peroxidase complex for 15 min (Vectastain ABC ELITE kit; Vector Laboratories, Burlingame, CA) before a substrate solution of 3,3′-diaminobenzidine tetrahydrochloride (Sigma Chemical Co.) was added. Sections were counterstained in Harris’ hematoxylin and mounted in dePex (BDH; Merck, Poole, U.K.). Negative control sections had the omission of the primary antibody. Positive control tissues were also included. Quantification of aortic cross-section immunostaining was completed by computer-aided densitometry (Optimas 6.2; Video Pro-32), whereby a total of 10 slides that contained at least 6 sections (×100) per group were counted. Results were expressed as proportional area of positive staining (28) in cross-sections.

Localization of CTGF mRNA was determined by in situ hybridization as previously described, with the inclusion of a sense riboprobe to determine background hybridization (16). Briefly, antisense CTGF and sense riboprobes labeled with ³⁵S CTP were prepared using the Promega Transcription System (Promega, Madison, WI). Four-micron sections were incubated in 125 μg/ml Pronase E at 37°C, postfixed in 4% paraformaldehyde, and air dried. The slides were then incubated overnight at 60°C in hybridization buffer that contained 2 × 10⁴ cpm/μl ³⁵S-labeled riboprobe, 0.72 mg/ml yeast RNA, 50% deionized formamide, 100 mmol/l dithiothreitol, 10% dextran sulfate, 0.3 mol/l NaCl, 10 mmol/l Na₂HPO₄ (pH 7.5), 5 mmol/l EDTA (pH 8.0), 0.02% BSA, 0.02% Ficoll 400, and 0.02% polyvinylpyrrolidone. Coverslips were removed, and sections were rinsed and then incubated at 37°C in 150 μg/ml RNase A for 1 h. Samples were washed, air dried, and exposed to BioMax MR film for 3–5 days. For microscopic localization, slides were coated in Amersham LM-1 emulsion, dried for 1 h, and incubated at 4°C in a light-proof container with desiccant for a period of 2–4 weeks according to the autoradiography results and developed photographically as previously described.

Results are expressed as means ± SE unless otherwise specified. Analyses were performed by ANOVA followed by post hoc analysis using Fischer’s least significant difference method, correcting for multiple comparisons (Statview V). P < 0.05 was considered to be statistically significant.

Plasma glucose and GHb were increased in all diabetic animals as compared with the control group (Table 1). There was a modest decrease in GHb in AG-treated animals. Diabetic mice also had decreases in body weight, which were not affected by either treatment (Table 1). Systolic blood pressure remained unchanged between control and diabetic mice, although there was a significant reduction in blood pressure noted in the DALT treatment group (Table 1). When compared with untreated diabetic mice, those that were treated with ALT-711 showed significant reductions in plasma total cholesterol, LDL cholesterol, and triglycerides. AG treatment did not affect plasma triglycerides and had a modest effect on HDL cholesterol, but no statistical changes were seen with respect to total or calculated LDL cholesterol (Table 1).

En face analysis of total aortic plaque area revealed an approximately sixfold increase in the untreated diabetic as compared with control mice (Fig. 1A). The lesions throughout the aortas of these diabetic animals appeared complex and fibrous as compared with the control apoE−/− animals, which seemed to have mostly fatty streaks with complex plaques evident only in the aortic arch. Both AG- and DALT-treated animals demonstrated significant abrogation in total plaque area and in the complexity of plaques. There was no attenuation of plaque area seen in any treatment group in the aortic arch region (Fig. 1B), and these lesions appeared fibrous and complex in all groups. By contrast, in the thoracic and abdominal regions, complex plaques were identified primarily in the untreated diabetic group. Each of the treatments administered to diabetic animals afforded a similar degree of attenuation in plaque lesion area in both the thoracic (Fig. 1C) and abdominal (Fig. 1D) regions of the aorta.

Diabetes induced a modest but significant increase in plasma AGE peptides (Fig. 2A). These diabetes-induced increases were completely prevented by each of the treatments.

Immunohistochemical analysis of the nonfluorescent AGE CML in aortic cross-sections revealed significant increases in the untreated diabetic group (Fig. 2B and3A). There was extensive CML staining localized within the vascular wall, particularly in smooth muscle cells as well as in the endothelial and adventitial layers (Fig. 3B). There was also CML staining within the cellular portion of diabetic plaques. The elastic laminae, however, remained unstained. The most effective attenuation of this vascular wall staining of CML was seen in the DALT groups (Fig. 3C), although the AG-treated group also showed significant reductions in staining for CML (Fig. 3D). It is interesting that there was still some CML immunostaining identified within plaques of the DALT group.

Quantification of RAGE expression by immunohistochemistry identified a significant increase in this receptor in untreated diabetic mice as compared with control apoE−/− mice (Fig. 2C). RAGE was localized throughout the aortic wall, particularly within the endothelium and adventitia, although there was also staining evident within the plaques of the untreated diabetic group (Fig. 3E and F). Each of the treatments prevented the diabetes-induced expression of RAGE within the aorta (Figs. 2C and3G and H). No immunostaining was evident in the negative control sections in which the primary antibody was omitted.

Skin collagen from diabetic animals was less soluble than collagen taken from control animals. The time taken to reach 50% digestion of the diabetic collagen with pepsin was 8.4 h as compared with 3.8 h in control animals (Fig. 4A). Each of the apoE−/− diabetic groups, which received treatment, demonstrated increased solubility of their skin collagen. The DALT and DAG groups showed similar findings with 50% digestion achieved at ∼5 h in both groups.

Skin collagen–associated CML was increased by diabetes and was attenuated by both treatments, although AG seemed to be more effective (Fig. 4B). There was no increase in the AGE/ALE CEL, although AG did reduce collagen CEL to below control levels (Fig. 4C). Finally, untreated diabetic skin collagen showed increased AGE fluorescence (excitation 370 nm, emission 440 nm) as compared with control mouse skin collagen (Fig. 4D). Treatment of diabetic groups with ALT-711 or AG prevented the increase in skin collagen fluorescence.

The total collagen content in lesions was >10 times greater in the untreated diabetic aortic lesions as compared with nondiabetic controls (Table 2). This was related to significant increases in each of the collagens studied, specifically collagens I, III, and IV (Table 2). Treatment with ALT-711 attenuated the overall increase in plaque collagen content by ∼50% but displayed differential effects on the three types of collagen studied. The reduction with ALT-711 treatment was most profound for collagen IV, which was normalized to nondiabetic levels, with collagen III attenuated by ∼50%, whereas there was no effect on the increase in type I collagen. AG completely abrogated the diabetes-induced increase in total collagen, which was related in part to reductions in each of the collagen subtypes (Table 2).

Collagen I was identified throughout the aortic wall and within plaques of the diabetic apoE−/− mice. ALT-711 seemed to reduce the collagen content of the wall only, in contrast to AG, which provided no reduction in either plaque or aortic wall content of collagen I. Collagen III was localized to the outer layer of the elastic laminae throughout the vessel wall of diabetic animals. There was also staining perpendicular to the elastic laminae transversing the aortic media. In addition, collagen III was evident in the fibrous cap of complex plaques. Neither ALT-711 nor AG reduced the collagen III content immediately over plaques, although there was significant attenuation within the vessel wall in each of the treatment groups. Collagen IV was identified throughout the internal elastic laminae, particularly toward the outer surface. These layers were thickened in the diabetic apoE−/− mice with collagen IV deposition also evident within plaques. Both treatment groups showed a reduction in the content of collagen IV within the vessel wall and plaques, although the superficial surface staining of plaques remained unchanged.

Aortic CTGF protein expression was induced by diabetes (Fig. 5A). CTGF mRNA and protein were localized within aortic plaques and in the media of diabetic mice (Fig. 6). Both treatments attenuated these increases in CTGF gene and protein expression to levels seen in control apoE−/− mice. A similar pattern was seen for TGF-β1 expression, which was increased in diabetic apoE−/− aortas as compared with control mice and was reduced in the treated groups (Fig. 5B).

α-SMA protein expression, as assessed by immunohistochemistry, was increased in the aortas from untreated diabetic as compared with control mice (Fig. 5C). This parameter was reduced by each of the treatments, although not to control levels (Fig. 5C).

The present study has clearly demonstrated that two disparate approaches to prevent aortic AGE accumulation—with either an inhibitor of AGE formation, AG, or an AGE cross-link breaker, ALT-711—are associated with attenuation of plaque area in a model of diabetes-associated atherosclerosis. These effects occurred in the context of reduced accumulation of AGEs within the aortas; reduced expression of RAGE, which has been implicated in the mediation of many of the biological effects of AGEs; alterations in both AGEs and ALEs present on skin collagen; and a reduction in the expression of prosclerotic growth factors and extracellular collagenous proteins.

AGE accumulation in aortic atheroma has been previously identified in nondiabetic patients (29). The present study confirms that AGEs are increased in diabetic apoE−/− mice as reported by Park et al. (10). Increases in aortic CML were found in association with increased CML in skin collagen. Analysis of skin collagen AGEs provides information about historical glycemic control as evidenced in type 1 diabetic patients in the Diabetes Control and Complications Trial (30) and significantly correlates with the long-term macrovascular outcomes of these patients recently published for the Epidemiology of Diabetes Interventions and Complications study (31), a follow-up trial of Diabetes Control and Complications Trial subjects. Recently, CML has been recognized as an ALE as well as an AGE because it can be formed during either lipid or carbohydrate oxidation (32). Thus, it may be that the increased CML in the atherosclerotic plaques and walls of the aortas observed in the present study is the result of both glycoxidation and peroxidation of lipoproteins and lipids. It is interesting that in the context of aortic plaque reduction, the ALT-711 treatment group had a reduction in plasma LDL cholesterol but lipid levels remained elevated. Our data provide the first report of the effects of ALT-711 on serum lipids, particularly in the apoE−/− mouse. Although in human trials, AG has demonstrated lipid-lowering effects, previous data from other groups have suggested that in rodents, AG has no or minimal effect on plasma cholesterol and a modest effect on triglycerides (8). This is consistent with the present study in which no major effect of AG was demonstrated on the lipid profile in the diabetic apoE−/− mice. Importantly, despite different effects on lipids, both ALT-711 and AG were equally effective in reducing aortic CML content. The relative role of these agents on glycoxidation and lipoxidation in mediating their antiatherosclerotic effects remains to be determined. These pathways need to be considered further, however, with the recent demonstration that other agents that inhibit AGE accumulation may also influence lipid- and glucose-dependent pathways (8).

An important finding in the present study was the regional efficacy of the AGE inhibitors in attenuating plaque formation. Whereas both ALT-711 and AG had no significant effects on atherosclerosis within the aortic arch, previous studies by our own group using the same model have shown that angiotensin-converting enzyme (ACE) inhibition was effective in the aortic arch as well as in the thoracic and abdominal regions of the aorta (16). This suggests that the renin-angiotensin system (RAS), specifically the local RAS rather than systemic blood pressure, may be particularly important in plaque accumulation within this region of the aorta, especially in the context of a reduction in blood pressure by ALT-711, without effects on plaque formation at this site. The effects of ALT-711 and AG on plaque accumulation in the thoracic and abdominal aorta were similar to those seen with perindopril. This suggests that metabolic pathways such as AGE accumulation predominate at these sites. Furthermore, with the recent demonstration both in vitro and in vivo that ACE inhibitors can act as potent inhibitors of AGE formation (33,34), it is possible that the effects of perindopril at these sites may be partly related to its ability to reduce AGE levels.

A range of studies from other groups have clearly defined that agents that interrupt AGE pathways influence the vascular tree. Administration of ALT-711 to aging patients abrogated vascular dysfunction attributed to vessel wall stiffening, in the absence of blood pressure reduction (15). AG has also been shown to have beneficial effects on vessel elasticity and endothelial dysfunction (3). Because AG is also an inhibitor of inducible nitric oxide synthase (35), it is possible that some of the beneficial effects of this drug may be via this pathway. Indeed, genetic deficiencies in inducible nitric oxide synthase have been shown to be protective against the development of atheroma (36). Nevertheless, the ability of ALT-711 to attentuate atherosclerotic plaque formation in the present study and previous experiments documenting the efficacy of sRAGE in this model (37) emphasize that the predominant antiatherosclerotic effect of AG is likely to be mediated via its actions as an inhibitor of AGE formation.

ALT-711 and the related compound N-phenacylthiazolium bromide (38) have been reported by investigators to be AGE cross-link breakers capable of the promoting the cleavage and removal of preformed AGEs (39). Indeed, although previously demonstrated to be an AGE cross-link breaker (12), ALT-711 has also been shown to reduce the accumulation of CML (13), and this effect has been postulated to be due to chelation of copper ions, thereby reducing copper catalyzed glycoxidation (40,41). One must be cautious about the interpretation of these in vitro data as these authors also did not find an effect of ALT-711 in contrast to studies defining its ability to cleave preformed AGEs (12). Skin collagen from the diabetic apoE−/− mice was heavily cross-linked as assessed by a reduction in its solubility in pepsin and an increase in fluorescent AGE content. Treatment of diabetic apoE−/− mice with ALT-711 or AG reversed this defect in solubility and reduced the accumulation of AGEs. It is likely that the mechanism of action of these two compounds in conferring these effects differs. Although ALT-711 restores collagen solubility and reduces aortic and skin AGE accumulation, there was only partial attenuation of the diabetes-induced increase in aortic collagen content in this group. This was in contrast to the DAG treatment group, which showed a similar level of aortic collagen content to that seen in control apoE−/− mice. This suggests that AG may be providing its beneficial effects on aortic lesions by preventing both the deposition of vascular collagen and, subsequently, AGE cross-link formation. This is also supported by the results of assessment of individual collagen subtypes. It is interesting that neither drug removed collagen localized at the luminal surface of the plaques. It is tempting to speculate that this may be beneficial to plaque stability in light of recent evidence suggesting that collagen ratios (42) and fibromuscular cap composition (43) are important contributors to plaque stability. sRAGE administration in this model has also been reported to be associated with less collagen accumulation (37).

The present study demonstrated a blood pressure–lowering effect of ALT-711. This is consistent with preliminary clinical data from the as-yet-unpublished SILVER/SAPPHIRE trials in isolated systolic hypertension in which ALT-711 has conferred positive effects on blood pressure, albeit it in a nondiabetic context (www.alteon.com). One cannot exclude a role for blood pressure in the acceleration of atherosclerosis because our group has previously shown beneficial effects in the diabetic apoE−/− model with the ACE inhibitor perindopril (16). Although this effect of the ACE inhibitor has, in general, been interpreted as suggesting a role for blockade of the RAS as the mediator of reduced aortic plaque formation, a direct effect of ALT-711 on blood pressure cannot be excluded. Nevertheless, AG also attenuated atherosclerosis without affecting systemic blood pressure. Furthermore, other studies have shown that hypertension does not account for accelerated atherosclerosis in apoE−/− mice in association with vascular dysfunction (44). Indeed, our own studies in diabetic apoE−/− mice that were treated with the calcium channel blocker amlodipine have identified no improvement in aortic lesions despite normalization of blood pressure (45). In support of this observation, studies by Miyata et al. (33) have also previously demonstrated that nifedipine, another dihydropyridine calcium channel blocker, has no effect on AGE formation.

The accumulation of collagen I was identified throughout the vessel wall and on the luminal side of plaques from diabetic apoE−/− mice. It is interesting that although the increase in vascular collagen I accumulation was partially abrogated by treatment with ALT-711, the plaque content of collagen I remained unchanged. By contrast, AG attenuated the increases in both the aortic wall and the plaque for collagen I. Furthermore, diabetes-induced increases in collagen III and IV, although attenuated in the vascular wall, were not reduced within plaques from treated animals. Because the presence of collagen within the plaques may provide plaque stability (42), the region-specific depletion of collagen I within the vessel wall but not within the plaque itself by ALT-711, in contrast to depletion at both sites by AG, may ultimately represent a potential therapeutic advantage of ALT-711 over agents such as AG. However, interpretation of these findings remains speculative because this difference between agents was not noted for all collagen subtypes. In addition, AGEs have been shown to induce the proliferation (46) and activation of vascular smooth muscle cells (47) to produce fibrillar collagen. This is supported in the present study by the amelioration of diabetes-induced α-SMA expression by each of the AGE reducing treatments.

The profibrotic cytokine CTGF was increased in the atherosclerotic vessels and may be contributing to the deposition of matrix proteins in the vasculature of these diabetic apoE−/− mice. Recent studies by our own group have demonstrated increases in aortic CTGF in this experimental model of atherosclerosis and diabetes (16). Furthermore, we have previously shown that AG (26) and ALT-711 (13) have the ability to reduce the expression of CTGF in diabetic nephropathy. Therefore, the reduction in CTGF expression in response to the inhibitors of AGE accumulation as demonstrated in the present study was not surprising. These findings provide further evidence for the view that this growth factor plays a pivotal role in mediating collagen deposition in various contexts, including diabetes.

RAGE has been identified as a modulator of inflammation (48) and as a major contributor to atherogenesis (10); therefore, not surprisingly, diabetes increased the expression of RAGE in the vascular wall in apoE−/− mice. The reduction in aortic RAGE expression that was seen after ALT-711 or AG treatment has also been reported after administration of sRAGE in this same model (10). In all of these studies, the various treatments were associated with a reduction in atherosclerosis. It remains to be determined whether the reduction in RAGE is just a manifestation of less vascular disease or represents a central mechanism for mediating the antiatherosclerotic effects of different treatments that modulate the advanced glycation pathway. The benefits of sRAGE in this model have been interpreted as indicative of the direct participation of RAGE in the development of diabetes-associated atherosclerosis. However, an alternative interpretation of the results is that sRAGE is acting primarily as a ligand to bind AGEs, thereby promoting their removal and preventing their interaction with other receptors to promote vascular injury. RAGE also acts as a receptor to other inflammatory molecules such as S-100, which may also promote atherosclerosis (48). However, the findings of the present study emphasize the benefits of blocking AGE accumulation per se and are consistent with a mechanism whereby sRAGE confers antiatherosclerotic effects via suppression of AGE-mediated pathways. The recent advent of a viable RAGE knockout mouse will greatly facilitate further elucidation of the specific role of RAGE in this field (49).

The present study has identified that AGEs play an important role in diabetes-associated atherosclerosis, because ALT-711 and AG both were effective in reducing both atherosclerotic plaque formation and complexity. These findings extend our understanding of the central role of AGEs in atherosclerosis, particularly in the diabetic context, and provide clinical investigators with an increasing number of therapeutic options to explore as part of the approach to reduce the burden of cardiovascular disease in diabetes.

This work was completed with support from the Juvenile Diabetes Research Foundation (JDRF), the National Heart Foundation of Australia, and U.S. Public Health Service Grant DK-19971. J.M.F. is a JDRF Post-Doctoral Research Fellow. M.C.T. is a recipient of a National Health and Medical Research Council (NHMRC) Postgraduate Medical Scholarship. M.L. is supported by grants from the Finnish Academy, Einar and Karin Stroem’s Foundation, The Helsingin Sanomat Centennial Foundation. and Paavo Nurmi Foundation. W.C.B. is an Australian Kidney Foundation postgraduate scholar. T.J.A. is a recipient of a Career Development Award jointly funded by the NHMRC and Diabetes Australia.

We thank Paula Aldersea and Gavin Langmaid for expert care of the animals throughout the study and Maryann Arnstein for technical expertise.