OUP user menu

RAGE mediates oxidized LDL-induced pro-inflammatory effects and atherosclerosis in non-diabetic LDL receptor-deficient mice

Li Sun , Tatsuro Ishida , Tomoyuki Yasuda , Yoko Kojima , Tomoyuki Honjo , Yasuhiko Yamamoto , Hiroshi Yamamoto , Shun Ishibashi , Ken-ichi Hirata , Yoshitake Hayashi
DOI: http://dx.doi.org/10.1093/cvr/cvp036 371-381 First published online: 28 January 2009


Aims Receptor for advanced glycation end products (RAGE) plays a pivotal role in the genesis of diabetic vascular diseases. To further explore the mechanisms underlying atherosclerosis under non-diabetic conditions, we examined the effect of RAGE deficiency on atherosclerosis in hyperlipidaemic mice.

Methods and results RAGE−/− mice were crossed with low-density lipoprotein receptor-deficient (LDLr−/−) mice to generate the double knockout (DKO) mice. After feeding with high-fat diet for 12 weeks, aortic atherosclerotic lesions were analysed histologically in these mice. Although there were no differences in serum levels of glucose and known RAGE ligands between DKO and LDLr−/− mice, DKO mice exhibited a significant decrease in the size and macrophage content in atherosclerotic lesions compared with LDLr−/− mice. Expression of intracellular adhesion molecule-1 and vascular cell adhesion molecule-1 in the aorta was lower in DKO mice than in LDLr−/− mice. Fluorescence-based assays revealed that oxidative stress in the vessel wall was attenuated in DKO mice than in LDLr−/− mice. Cell culture experiments revealed that RAGE mediated oxidative LDL-induced activation of p42/44 mitogen-activated protein kinases and oxidative stress in macrophages.

Conclusion Oxidative LDL may be a ligand of RAGE in the hyperlipidaemic state. RAGE inactivation inhibits the atherosclerosis through reducing oxLDL-induced pro-inflammatory responses and oxidative stress in hyperlipidaemia.

  • Atherosclerosis
  • Adhesion molecule
  • Oxidized LDL
  • Oxidative stress
  • RAGE

1. Introduction

The cardiovascular complications of diabetes are a major cause of patient mortality1 and the underlying mechanisms of the cardiovascular complications of diabetes have not been fully elucidated. Hyperglycaemia drives non-enzymatic glycation and oxidation of proteins and lipids, which enhances irreversible formation of advanced glycation end products (AGEs).2 This process is particularly facile in diabetic and hyperlipidaemic states. Accumulation of pre-formed AGEs in the vessel wall has been shown to promote diabetic vascular diseases.3 The receptor for AGE (RAGE) is a multi-ligand receptor that mediates the action of AGEs. RAGE was initially isolated from the lung, but is also expressed on the surface of vascular endothelial cells, smooth muscle cells, and macrophages.46 The endogenous ligands of RAGE are variable and include AGEs,7 S100/calgranulins family,8 amphoterin, and amyloid-ß peptide.3 Accumulating evidence suggests that RAGE plays a pivotal role in promoting inflammatory processes and endothelial activation, which accelerates atherosclerosis in patients with diabetes.9 Binding of AGEs to RAGE activates multiple intracellular signalling pathways including p21ras, which recruits downstream targets such as mitogen-activated protein kinases (MAPK), and activates NF-κB.10,11 The AGE–RAGE interaction augments inflammatory responses by causing upregulation of cyclooxygenase-25,12 and leads to vascular dysfunction and monocyte activation.13 Diabetes-associated atherosclerotic lesions display increased accumulation of RAGE ligands and enhanced expression of RAGE.9,14

On the other hand, an endogenous secretory isoform of RAGE (soluble RAGE; sRAGE) exerts anti-atherogenic effects by acting as a decoy. Administration of sRAGE attenuates the severity and complexity of atherosclerosis in streptozotocin-induced diabetic apoE-null mice.15,16 Previous studies have indicated that low levels of sRAGE in the plasma correlate with the presence of coronary artery disease in non-diabetic patients.17 These lines of evidence suggest that the RAGE system and its associated ligands have a close relationship not only with diabetic vascular complications but also with non-diabetic vascular diseases. However, little is known about the significance of RAGE expression in atherosclerosis under non-diabetic conditions. In this study, we aimed to explore the role of RAGE in atherosclerosis by seeking to identify novel ligand(s) in the hyperlipidaemic environment.

2. Methods

2.1 Animal preparations

All animal studies were performed in accordance with the Institutional Guidelines of Kobe University and the Guide for the Care and Use of Laboratory Animals published by the US National Institute of Health (NIH Publication No. 85–23, revised 1996). Homozygous RAGE-deficient mice (RAGE−/−)18 were crossed with low-density lipoprotein receptor-deficient mice (LDLr−/−)19 to yield RAGE−/−LDLr−/− double knockout mice (DKO). Wild-type (WT) C57Bl/6 mice (RAGE+/+) were obtained from Japan Charles River (Yokohama, Japan). Between 6 and 14 mice with the C57Bl/6 background were included in each subsequent experimental subgroup. Littermate LDLr−/− and DKO mice were weaned at 8 weeks of age onto a high-fat diet (0.15% cholesterol, 15.2% protein, 36.2% fat, and 48.7% carbonate) from Oriental Yeast, Inc. (Tokyo, Japan), and maintained on the diet for 12 weeks.

2.2 Characterization of atherosclerotic lesion and blood chemistry

The LDLr−/− and DKO mice were euthanized at the age of 20 weeks (after 12 weeks on the high-fat diet) for the analyses of the atherosclerotic lesions and blood chemistry. Blood was taken by cardiac puncture and centrifuged at 5000 r.p.m. to obtain the plasma. Lipoprotein profiles were determined by LipoSEARCH®, a high-sensitivity lipoprotein profiling system (Skylight Biotech, Akita, Japan). For an intraperitoneal glucose tolerance test (IPGTT), fasted mice received an intraperitoneal injection of glucose (1.2 mg glucose/g body weight), and blood samples were collected at 0, 15, 30, 60, and 120 min after the glucose injection. Blood glucose was measured by the OneTouch Ultra Blood Glucose Monitoring System (LifeScan, Milpitas, CA). HbA1c levels were determined using a DCA2000 analyser (Bayer Medical, Tokyo, Japan).

The aorta was perfused and fixed with 4% paraformaldehyde/PBS, embedded in OCT compound, and sectioned (10-µm thickness). Five consecutive sections, spanning 600 µm of the aortic root, were collected from each mouse and stained with haematoxylin and eosin (H–E), Sudan III, Elastica van Gieson, or Masson’s trichrome. For quantitative analysis of the atherosclerosis, total lesion area of five sections from each mouse was measured with Image J Software by the method reported previously.20 The amount of aortic lesion formation in each animal was measured and calculated as the percentage of lesion area per total area of the aorta.

For determination of macrophage content or oxidized LDL (oxLDL) levels, frozen aortic sections at the same level were incubated with a monoclonal rat anti-mouse macrophage antibody (MOMA-2, 1:100, Biosource International, Camarillo, CA) or a polyclonal antibody against Cu2+-oxidized LDL (1:400, Acris Antibodies GmbH, Germany) overnight at 4°C, then biotinylated anti-rat IgG or anti-rabbit IgG (1:200) at 37°C for 60 min, followed by the LSAB® streptavidin-HRP (Dako Japan, Kyoto, Japan) for 20 min. Peroxidase activity was identified by reaction with the Liquid DAB substrate Chromogen System (Dako). The percentage of macrophages in atherosclerotic lesions was quantitated as described.20

2.3 Immunoblotting

The aortic sample was homogenized in a lysis buffer (50 mM Tris–HCl pH 7.5, 1 mM EGTA, 1 mM DTT, 2 µg/mL leupeptin, 0.001 mM phenylmethylsulfonyl fluoride [PMSF], 10% Glycerol, 20 nM CHAPS). An aliquot (30 µg) of the tissue lysate or serum was subjected to western blotting using primary antibodies: a polyclonal goat anti-mouse CD54 antibody, polyclonal rabbit anti-vascular cell adhesion molecule (VCAM)-1 antibody, anti-intracellular adhesion molecule (ICAM)-1 antibody (1:500, Santa Cruz Biotechnology Santa Cruz, CA), anti-AGE antibody (1:500, TransGenic, Kumamoto, Japan), mouse anti-S100B antibody (1:5000, BD Biosciences, San Jose, CA), or a monoclonal mouse anti-high-mobility group box-1 (HMGB-1) antibody (1:2000, Gene Tex, San Antonio, TX) overnight at 4°C. The blots were exposed to HRP-conjugated anti-IgG antibodies (1:5000 Amersham Biosciences, Little Chalfont, Buckinghamshire, UK), and the protein bands were visualized using the SuperSignal® West Pico Chemiluminescence Substrate (Pierce Biotechnology, Rockford, IL) and quantified. ß-actin was used as an internal control.

2.4 Detection of superoxide by in situ dihydroethidium method

Dihydroethidium (DHE), an oxidative fluorescent dye,21 was used to detect superoxide in segments of aorta and aortic sinus. Fresh unfixed segments of the aortas and hearts were frozen in OCT compound and transverse sections (20 µm) were cut with a cryostat. Sections were then allowed to thaw, half-dried, and incubated in a light-protected chamber at 37°C for 10 min with 2 µmol/L DHE (Molecular Probes, Eugene, OR). After washing twice with ice-cold PBS, the slides were sealed with a mounting medium for fluorescent specimens (Dako Japan). Images were obtained with the use of an LSM5 Pascal laser scanning confocal microscope (Carl-Zeiss Japan, Tokyo, Japan). The excitation wavelength was 488 nm and the emission fluorescence was detected with the use of a 585 nm long-pass filter. Identical laser settings were used for acquisition of images from different groups of mice.

2.5 Generation of RAGE-overexpressing COS7 cells

The cDNAs for mouse full-length RAGE (mRAGE, GenBank NM_007425) and endogenous secretory RAGE (esRAGE, GenBank AB207883) were obtained by RT–PCR using mouse lung cDNA as template. The primer sequences for mRAGE were 5′-ATGAATTCGCCGCCATGCCAGCGGGGACAGCA-3′ and 5′-AAGATATCGAAGTGCCTCAAGGAGGAATTGG-3,′ and for esRAGE they were 5′-ATGAATTCGCCGCCATGCCAGCGGGGACAGCA-3′ and 5′-AAGATATCTTAGTCCGTTCCCTCACCTTCAG-3′. The PCR products were cloned into the pcDNA3.1 expression vector (Invitrogen, Carlsbad, CA). COS7 cells were cultured in Dulbecco’s minimal essential medium (DMEM) containing 10% fetal bovine serum (FBS), and transfected with the mRAGE- or esRAGE-containing plasmid using Lipofectamine™ 2000 (Invitrogen) and transfected cell clones were selected in the presence of 600 µg/mL G418 (Invitrogen). Western blotting was used to confirm the expression of mRAGE (44 kDa) and esRAGE (34 kDa). We successfully obtained stable clones overexpressing mRAGE (COS7mRAGE) or esRAGE (COS7esRAGE). The mock (vector)-transfected cells were used as control (COS7Mock).

2.6 Phosphorylated Erk assay

COS7 cells and mouse intraperitoneal macrophages (MPM) were maintained in DMEM containing 10% FBS. After serum starvation, the cells were stimulated with oxLDL (5 µg/mL for COS7 cells, or 1 µg/mL for MPM) with or without RAGE-neutralizing antibodies (30 µg/mL, R&D Systems, Minneapolis, MN), or esRAGE-rich medium (esRAGE concentration: 100 µl/mL).22 Proteins were extracted from the treated cells at indicated time points by using Na3VO4-lysis buffer (0.5% NP-40, 20 mM Tris–HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.01 mM PMSF, and 1 mM Na3VO4). An aliquot of protein (20 µg) was subjected to western blotting for Erk1/2 (p44/42 MAPK) utilizing the phosphor-p44/42 MAPK (Thr202/Tyr204) (E10) or p44/42 MAPK antibodies (1:1000, Cell Signaling Technology, Danvers, MA).

2.7 LDL–RAGE binding assay

COS7esRAGE cells were cultured in serum-free DMEM for 18 h and the culture medium was collected and used as the esRAGE-rich medium. Culture medium from COS7Mock cells was used as control. Native LDL (nLDL) was isolated from the human plasma by sequential ultracentrifugation.23 For LDL binding studies, nLDL or oxLDL was fluorescently labelled with DiI. The DiI labelling was achieved by adding 3 µg of DiI (Molecular Probe, Eugene, OR) in DMSO (3 mg/mL) into 1 mg of lipoprotein protein and incubation for 8–15 h followed by centrifugation at 13 000 r.p.m. for 20 min. The supernatants were collected as DiI-labelled LDL preparations. The binding of DiI-labelled LDL to COS7mRAGE was monitored by flow cytometry (BD FACSCalibur System, BD Biosciences) in the presence or absence of anti-RAGE antibodies (30 µg/mL) or esRAGE-rich medium. COS7mRAGE or COS7Mock cells (5 × 105 cells/mL, 200 µl) were incubated with 5 µg/mL DiI-labelled LDL for 2 h at 4°C. After washing with PBS, the cells were scraped in 1 mL PBS and collected into a 5 mL-polystyrene round-bottom tube with a cell-strainer cap on ice. The binding data were calculated after subtracting the corresponding non-specific binding which was obtained by adding 250 µg/mL LDL preparation before adding the DiI-labelled LDL preparation to cells.

2.8 Measurement of O2 production by lucigenin-enhanced chemiluminescence

After overnight serum-starvation, MPM were incubated with oxLDL (5 µg/mL) with or without 100 µM diphenyliodonium (DPI, NADH/NADPH oxidase inhibitor, Sigma-Aldrich, St Louis, MO) for 6 h. The cells were then washed with ice-cold PBS, gently scraped into the lysis buffer and homogenized. The O2 production was measured by lucigenin-enhanced chemiluminescence (CL) assay using a BLR-201 Luminescence Reader (Aloka, Tokyo, Japan). After incubation with the solution (50 mM HEPES, pH 7.4, 1 mM EDTA, 6.5 mM MgCl2, 83 mM sucrose, 250 µM lucigenin, 100 µM NADPH) for 10 min, the reaction was started by adding 100 µl of the homogenates to make the final volume 1 mL. Photon emission was continuously recorded for 15 min. All CL data were reported after subtracting the background CL observed in the absence of homogenates as the average counts per minute for the 15 min recording. In case with the aorta, fresh whole aortas were equilibrated in pre-incubation buffer (HEPES/PSS: DCC = 10 000:1) for 30 min at 37°C, then transferred to the HEPES/PSS with lucigenin (5.1 µg/mL), which was used to check the background. The generation of superoxide anion was expressed as counts per minute per unit mass of cell protein or dry tissue weight.

2.9 Statistical analysis

Data were expressed as mean±SE. Unpaired Student's t-test was used to detect statistically significant differences between two groups. One-way ANOVA was used to compare the differences among three or four groups with Bonferroni’s test for post hoc analysis. Repeated measures ANOVA was used to compare lipid fraction results obtained from HPLC. P < 0.05 was considered statistically significant.

3. Results

3.1 RAGE deficiency attenuated atherosclerotic lesions in LDLr−/− mice

Male and female LDLr−/− and RAGE−/−LDLr−/− (DKO) mice were evaluated for development of atherosclerosis on high-fat dietary challenge. After 12 weeks on high-fat diet, at the age of 20 weeks, atherosclerotic lesion formation was evaluated by Sudan III staining of sections at the level of the aortic valve (Figure 1A). Extent of disease was quantified as total lesion area determined by planimetry. Interestingly, the atherosclerotic lesions were markedly attenuated in DKO mice compared with LDLr−/− mice. The mean lesion size between the two groups was significantly decreased by 54.9% in males and by 32.8% in females comparing DKO mice with LDLr−/− mice (Figure 1B, P < 0.01).

Figure 1

Characterization of atherosclerotic lesions in different genetic models. (A) Representative photomicrographs of atherosclerotic plaque vulnerability. Haematoxylin–eosin stain (H–E), Sudan III-stain, Masson‘s trichrome, Elastica van Gieson, and immunohistochemistry for MOMA-2 were performed in the aortic root sections from LDLr−/− and RAGE−/−LDLr−/− (double knockout) mice fed on high-fat diet for 12 weeks. Total lesion area (B) of five sections in the aortic root from each mouse was quantified morphometrically as described in the Methods. After 12 weeks on high-fat diet, the atherosclerotic lesion areas were significantly attenuated in double knockout compared with LDLr−/− mice. *P < 0.01 vs. corresponding LDLr−/− group (n = 14). Quantitation of lesion area stained with MOMA-2 revealed significantly less accumulation of macrophages in lesions of double knockout compared with LDLr−/− animals (C). *P < 0.01 vs. corresponding LDLr−/− mice (n=12).

3.2 Effect of RAGE deficiency on macrophage content in the atherosclerotic lesions

Immunohistochemical staining of atherosclerotic areas was performed with antibodies chosen to specifically label different cellular components of the lesions. An antibody to MOMA-2 was employed to evaluate the area containing infiltrating macrophages. The MOMA-2-stained area in the aortic sinus lesions was significantly smaller in DKO mice by 48.0% in males and by 61.3% in females, consistent with overall fewer macrophages in these lesions (Figure 1C). There was no significant difference in T-cell and smooth muscle cell composition of the lesions in the LDLr−/− and the DKO mice, as evaluated with immunohistochemistry with antibodies to CD3 and α-SMA (data not shown). Masson’s Trichrome and Elastica van Gieson staining revealed higher elastic collagen fibre contents in the lesion plaque of DKO mice compared with LDLr−/− mice (Figure 1A). Therefore, RAGE deficiency appeared to reduce not only the size and macrophage content but also the complexity of the atherosclerotic lesions.

3.3 Effect of RAGE deficiency on metabolic parameters

Plasma lipid profiles were obtained from LDLr−/− and DKO mice on high-fat diet. The HPLC analysis revealed that cholesterol levels in each lipoprotein fraction and LDL-cholesterol/HDL-cholesterol ratio were similar for DKO and LDLr−/− mice (Figure 2A and B). Similarly, there was no significant difference in triglyceride levels between DKO and LDLr−/− mice (Figure 2C). There was no significant difference in body weight and fasting blood glucose before and after receiving high-fat diet among the mouse groups (Table 1). IPGTT and HbA1C levels were similar among the mouse groups before and after receiving the high-fat diet (Figure 2D and E).

Figure 2

Plasma lipid profile and other ligand levels in different genetic models. Lipoprotein fractions were separated by HPLC, and the levels of cholesterol (A) and triglycerides (C) in the fraction were measured. High density lipoprotein-cholesterol/low-density lipoprotein-cholesterol ratio (B) was calculated. There were no significant differences in the lipoprotein profile between LDLr−/− and double knockout groups (n = 9–12). After 12 weeks of high-fat diet, the intraperitoneal glucose tolerance test and HbA1C were evaluated; no significant difference was found among the groups (D and E, n = 9).

View this table:
Table 1

Body weight and fasting blood glucose levels in different genetic models

Body weight (g)Blood glucose (g/L)
Age (weeks)6820620
RAGE+/+18.79 ± 2.7822.95 ± 3.9733.51 ± 2.94133.55 ± 54.81135.54 ± 54.11
RAGE−/−18.00 ± 0.6320.67 ± 1.0332.00 ± 6.4199.00 ± 19.18133.33 ± 55.87
LDLr−/−17.13 ± 2.0221.98 ± 1.7631.03 ± 5.51113.38 ± 46.54138.22 ± 63.10
 DKO15.89 ± 1.6519.14 ± 2.0630.39 ± 4.3282.00 ± 28.70125.58 ± 52.53
RAGE+/+15.25 ± 2.1217.13 ± 0.8323.85 ± 3.62124.17 ± 32.17130.29 ± 44.51
RAGE−/−16.46 ± 1.8118.58 ± 1.0823.82 ± 3.15100.00 ± 28.66130.38 ± 53.05
LDLr−/−15.27 ± 1.9817.50 ± 2.3124.08 ± 3.80111.59 ± 48.77137.18 ± 40.69
 DKO14.46 ± 1.0617.47 ± 1.6222.91 ± 3.5070.43 ± 30.16127.88 ± 39.00
  • There were no significant differences in body weight and fasting blood glucose levels among RAGE+/+, RAGE−/−, LDLr−/−, and DKO mice (n = 14).

Next we examined the levels of known-ligands of RAGE in these mice. Western blotting revealed that the plasma levels of AGEs and S100B were similar among WT RAGE+/+ mice, RAGE−/−, LDLr−/−, and DKO mice both in males and females (Figure 3A). Also, there were no significant differences in HMGB-1 levels in the aorta among these mouse groups (Figure 3B).

Figure 3

Receptor for advanced glycation end products deficiency reduced ICAM-1 and VCAM-1 expression in the aorta. Serum AGEs and S100B (A) levels and HMGB-1 levels (B) in the aortic tissue were analysed by western blotting, which show no significant difference among WT, RAGE−/−, LDLr−/−, and double knockout mice (n = 9). The ICAM-1 and VCAM-1 expression was similar in chow diet groups (C), but significantly decreased in double knockout compared with LDLr−/− after receiving high-fat diet (D); *P < 0.05, n = 7. Lucigenin-enhanced chemiluminescence assay was performed in the fresh mouse aorta (E). Chow diet mice and WT, RAGE−/− mice exhibited low oxidative stress, and high-fat diet induced oxidative stress in LDLr−/− mice, which was significantly decreased (45%) in double knockout compared with LDLr−/− mice (*P < 0.05, n = 7).

3.4 Effect of RAGE deficiency on expression of endothelial adhesion molecules

Endothelial adhesion molecules are known to be downstream molecules of the RAGE signalling pathway and the earliest markers of vascular inflammation. Therefore, we evaluated the expression of ICAM-1 and VCAM-1 in the aortic sinus. Western blotting revealed that ICAM-1 and VCAM-1 were all low and did not differ among the mouse groups with chow diet (Figure 3C). However, in the hyperlipidaemic condition, ICAM-1 levels in LDLr−/− mice were 2.1 times higher than those in WT mice, and DKO mice exhibited a statistically significant 68% (P < 0.01) decrease in ICAM-1 expression compared with LDLr−/− mice (Figure 3D). Similarly, VCAM-1 levels in the aorta of LDLr−/− mice were two times higher compared with those in WT mice. A statistically significant 79% decrease (P < 0.01) in VCAM-1 expression was found in DKO mice compared with LDLr−/− mice. These findings indicate that RAGE deficiency inhibited the expression of pro-inflammatory adhesion molecules in the hyperlipidaemic states.

3.5 Effect of RAGE deficiency on oxidative stress in hyperlipidaemic states

Lucigenin-enhanced CL was performed to quantitate the production of superoxide in the whole vessels (Figure 3E). In normolipidaemic states, oxidative stress in the mouse aorta was very low. When the mice were fed with high-fat diet, aortic O2 production in WT and RAGE−/− mice was not changed. In contrast, feeding LDLr−/− mice with high-fat diet markedly increased oxidative stress in the aorta, which was significantly lower in DKO mice. These results indicated that RAGE deficiency inhibited oxidative stress in the vessel wall under the hyperlipidaemic condition.

The localization of O2 production in the aortic lesions was determined by in situ DHE staining. While elastic lamina in the vessel wall was visualized as green signals under confocal microscopy, O2 production was seen as red signals. In mice fed with normal chow, O2 production in the aortic sinus was barely detectable (Figure 4A). In the hyperlipidaemic environment, however, LDLr−/− mice exhibited a marked O2 production in the aortic lesion compared with WT or RAGE−/− mice (Figure 4B), which was markedly decreased in DKO mice (Figure 4B). A closer observation revealed that the O2 production throughout the vessel wall was decreased in the DKO mice compared with that of LDLr−/− mice in hyperlipidaemic states (Figure 4C).

Figure 4

Superoxide production in mouse aortic system. Dihydroethidium assay was performed in the frozen mouse aortic sinus (A and B) and descending aorta (C) to measure the generation of O2. After high-fat diet (B and C), increased fluorescence associated with dihydroethidium staining (red dots) was detected in the aortic sinus and descending aorta isolated from LDLr−/− and double knockout mice compared with corresponding chow diet-fed mice (A). Arrows indicate O2 producing cells. The corresponding haematoxylin and eosin staining (H–E) and elastic lamina (EL) are shown. Each experiment was performed in a group of seven animals.

3.6 OxLDL activates Erk pathway through RAGE

Because oxLDL contains AGE epitopes,24 we tested if oxLDL can directly activate RAGE. We have generated COS7 cells which stably overexpress mouse RAGE (COS7mRAGE) and endogenous secretory RAGE (COS7esRAGE). RAGE is not expressed in the control COS7 cells (COS7Mock) (Figure 5A). We stimulated these cells with oxLDL, and evaluated the phosphorylation of Erk1/2 by western blotting. As shown in Figure 5B, oxLDL increased the phosphorylation of Erk1/2 in COS7mRAGE cells, but not in COS7Mock cells.

Figure 5

OxLDL activate Erk1/2 through receptor for advanced glycation end products. (A) Western blotting confirmed the expression of mRAGE (44 kDa) and esRAGE (34 kDa) in the cell lysate and medium, from COS7 mRAGE and COS7 esRAGE cells, respectively. After 24 h serum-starvation, the COS7Mock and/or COS7mRAGE cells were stimulated with oxLDL, RAGE neutralizing antibody, or esRAGE-rich medium for the indicated time points. Levels of phosphorylated- and total-Erk1/2 (P-Erk and T-Erk, respectively) were evaluated by western blotting in the graph (n = 5). OxLDL activated Erk in COS7mRAGE cells (solid bars), but not in COS7Mock cells (open bars) (B). Anti-receptor for advanced glycation end products antibody or esRAGE did not activate Erk in COS7 mRAGE cells (C). COS7mRAGE or COS7Mock cells were incubated with 5 µg/mL DiI-labelled nLDL (D) or oxLDL (E) for 2 h at 4°C, in the presence of esRAGE-rich medium (esRAGE+) or control medium (esRAGE−). (F) COS7mRAGE or COS7Mock cells were incubated with 5 µg/mL Di-I-labelled oxLDL with or without anti-receptor for advanced glycation end products antibodies. DiI-oxLDL-bound cells were assessed using flow cytometry. While esRAGE did not affect the nLDL-binding to COS7mRAGE or COS7Mock cells (D), esRAGE (E), and anti-receptor for advanced glycation end products antibodies (F) significantly inhibited the binding of oxLDL to COS7mRAGE cells but not to COS7Mock cells. Data are express as mean±SE (*P < 0.05, **P < 0.01, n = 5–6).

To demonstrate that there was a direct interaction between oxLDL and RAGE, we examined the effect of esRAGE or neutralizing antibodies against RAGE on the binding between LDL and COS7mRAGE cells. The anti-RAGE antibodies or esRAGE did not activate Erk1/2 (Figure 5C). Treatment of COS7Mock cells or COS7mRAGE cells with esRAGE did not affect the nLDL binding to the cells (Figure 5D). However, the addition of esRAGE significantly inhibited the binding of oxLDL to COS7mRAGE cells, but not the binding to COS7Mock cells (Figure 5E). Similarly, anti-RAGE antibodies markedly inhibited the binding of oxLDL to COS7mRAGE cells, but not to COS7Mock cells (Figure 5F). These results suggest that oxLDL by itself may serve as a ligand for RAGE and activate the RAGE signalling pathway under hyperlipidaemic conditions.

3.7 RAGE deficiency reduced proliferation and oxidative stress in macrophages

Because oxLDL is one of the most potent stimulator for initiation of atherosclerosis, we examined the existence of oxLDL in our mouse models by immunohistochemistry. OxLDL was hardly detectable in normolipidaemic WT and RAGE−/− mice (Figure 6A, top), and the expression was increased in hyperlipidaemic mice (Figure 6A, bottom). In contrast, oxLDL accumulated in the aortas of LDLr-/− and DKO mice fed with chow diet, and the accumulation was particularly augmented by feeding of high-fat diet. To clarify the mechanisms for the reduced macrophage content in the atheroma in DKO mice, we isolated peritoneal macrophages from RAGE−/− and RAGE+/+ mice and evaluated oxidative stress and Erk1/2 activation, since Erk1/2 is a key mediator of cell proliferation. OxLDL induced the expression of phosphorylated Erk1/2 both in RAGE+/+ and RAGE−/− macrophages, but phosphorylation of Erk1/2 in RAGE−/− macrophages was significantly lower than that in RAGE+/+ macrophages (Figure 6B).

Figure 6

OxLDL activates Erk1/2 through receptor for advanced glycation end products in macrophages. OxLDL immunoreactivity was detected in mouse aortic sinus (A). In chow diet mice and high-fat diet WT and RAGE−/− mice oxLDL was very low, and in LDLr−/− mice with high-fat diet oxLDL was high and receptor for advanced glycation end products deficiency decreased it. In negative control, the primary antibody was replaced with non-specific rabbit immunoglobulins. After 36 h serum-starvation, the RAGE−/− or wild-type (RAGE+/+) mouse intraperitoneal macrophages were stimulated with 1 µg/mL oxLDL for the indicated time points. Levels of total and phosphorylated Erk1/2 (T-Erk, and P-Erk, respectively) were evaluated by western blotting. The graph data represent the average of three independent experiments. Although Erk1/2 is activated by oxLDL both in RAGE+/+ and in RAGE−/− macrophages, the levels of phosphorylated Erk1/2 were significantly lower in RAGE−/− macrophages (B). *P < 0.05 vs. the corresponding RAGE+/+ value (n = 5). The RAGE−/− or RAGE+/+ mouse intraperitoneal macrophages were incubated with 5 µg/mL oxLDL for 6 h. Lucigenin-enhanced chemiluminescence assay revealed that oxLDL-induced oxidative stress in RAGE+/+ macrophages was six times higher than that in RAGE−/− macrophages, which was inhibited by diphenyliodonium, an inhibitor of NADH/NADPH oxidase (C). (*P < 0.05, **P < 0.01, n = 5).

Finally we evaluated the oxidative stress in RAGE+/+ and RAGE−/− macrophages by the lucigenin assay. OxLDL markedly increased oxidative stress in RAGE+/+ macrophages, which is partly but significantly inhibited by addition of diphenyliodonium (DPI), an inhibitor of NADH/NADPH oxidase. In contrast, the oxLDL-induced oxidative stress was abolished in RAGE−/− macrophages (Figure 6C). These findings suggest that RAGE mediates the oxLDL-induced proliferation and oxidative stress in macrophages.

4. Discussion

A variety of inflammation-driven processes and oxidative stress reactions are involved in the disease evolution of diabetic macroangiopathy, from the formation of the fatty streak to the development of advanced plaques.25 In particular, RAGE and its ligands are known to play a crucial role in atherosclerosis. Expression of AGE and RAGE was increased in the atherosclerotic area in apoE-null mice and human diabetic atherosclerotic plaques.9 RAGE ligands activate a range of vascular cells, such as endothelial cells, macrophages, and vascular smooth muscle cells, through activation of multiple intracellular signalling pathways. The activation of RAGE can stimulate p21ras, Erk1/2, p38 and SAPK/JNK, rho GTPases, phophoinositol-3 kinase, and the JAK/STAT pathway, and has downstream consequences such as activation of NF-κB.7,2628 It has been reported that the blockade of RAGE pathway by administration of soluble RAGE suppressed and stabilized the atherosclerosis in diabetic apoE null mice.16, 29

In the present study, we showed that RAGE deficiency resulted in a marked reduction of the atherosclerotic lesions in LDLr−/− mice. The finding is similar to a recent study by Harja et al.22 They generated RAGE−/−apoE−/− DKO mice and showed that RAGE deficiency attenuates atherosclerosis. They also demonstrated that the RAGE-ligand S100b/calgranulins and oxLDL-containing AGE epitopes contribute to a great extent to vascular inflammation, endothelial dysfunction, and atherosclerosis by activation of RAGE and its downstream JNK signalling. The present study has confirmed the inhibitory effect of RAGE deficiency on atherosclerosis utilizing another murine model of atherosclerosis, LDLr−/− mice. Moreover, we have extended the role of RAGE in atherogenesis to macrophages under hyperlipidaemic conditions. When we examined the composition of the plaques, we found a significant reduction of macrophage content in the plaques of DKO groups compared with LDLr−/− groups. Our cell culture experiments revealed overexpression of RAGE in RAGE-non-expressing COS7 cells resulted in the oxLDL-induced activation of Erk1/2, which is blocked by the treatment of esRAGE or RAGE-neutralizing antibodies. In contrast, oxLDL-induced activation of Erk1/2 was attenuated in RAGE−/− macrophages compared with RAGE+/+ macrophages. In the light of the fact that oxLDL contains AGE epitopes,24 our findings strongly imply the oxLDL by itself can induce the phosphorylation of Erk1/2 by activation of RAGE in macrophages. We speculate that the RAGE-mediated proliferation of macrophages may at least in part account for the reduced macrophage content and atherosclerosis in DKO mice.

The present study has characterized another target in the downstream signalling pathway of RAGE. While LDLr−/− mice displayed high levels of oxidative stress, the DKO mice exhibited significantly weaker oxidative stress in the aorta. Oxidative stress plays a critical role in the progression of atherosclerosis and vascular inflammation. In hyperlipidaemia, excess LDL in the plasma can undergo oxidation to yield oxLDL, and oxLDL in turn exhibits a variety of pro-atherogenic effects in the vascular wall. Our data indicate that RAGE can mediate oxLDL-induced production of reactive oxygen species as well as proliferation of macrophages. Oxidative stress has been reported to increase the expression of endothelial adhesion molecules. In fact, VCAM-1/ICAM-1 expression in the aorta was increased in hyperlipidaemic condition in the LDLr−/− mice and was decreased by RAGE inactivation, which is consistent with the previous report.22 These findings suggest that the reduction of oxidative stress by RAGE deficiency may partly account for the reduction in atherosclerosis progression and monocyte recruitment in the plaque. On the other hand, in the present study, RAGE deficiency did not affect plasma lipid or glycaemic profiles. Taken in combination with our in vitro data, it is likely that the atherogenic dyslipidaemia, specifically oxLDL, activated RAGE and its downstream signalling to promote atherosclerosis in hyperlipidaemia.

In summary, targeted inactivation of RAGE resulted in a significant decrease in the size and complexity of atherosclerotic lesions in LDLr−/− mice. The reduced atherosclerosis in the DKO mice was associated with reduced expression of ICAM-1 and VCAM-1 and oxidative stress in the vessel wall, as well as reduced proliferation of macrophages. OxLDL appears to be a new ligand of RAGE in the hyperlipidaemic condition, and its binding to RAGE enhances macrophage proliferation and oxidative stress. Here, we have further characterized the function of RAGE in the pathogenesis of atherosclerosis in hyperlipidaemia. In as much as this pathway appears to play a key role in atherosclerotic plaque formation, RAGE blockade may be useful as a therapeutic approach to inhibit atherosclerosis not only in diabetic but also in hyperlipidaemic states.


This study was supported by Grants-In-Aid for Scientific Research, Global- and 21st Century COE Program, from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and Japan Foundation of Cardiovascular Research.


We express sincere thanks to Dr Masamitsu Kuriyama for his help and advise in experiments on molecular biology and cell culture.

Conflict of interest: none declared.


View Abstract