AGER1 regulates endothelial cell NADPH oxidase-dependent oxidant stress via PKC-δ: implications for vascular disease

Abstract

Advanced glycated end-product receptor 1 (AGER1) protects against vascular disease promoted by oxidants, such as advanced glycated end products (AGEs), via inhibition of reactive oxygen species (ROS). However, the specific AGEs, sources, and pathways involved remain undefined. The mechanism of cellular NADPH oxidase (NOX)-dependent ROS generation by defined AGEs, N^ε-carboxymethyl-lysine- and methylglyoxal (MG)-modified BSA, was assessed in AGER1 overexpressing (AGER1⁺ EC) or knockdown (sh-mRNA-AGER1⁺ EC) human aortic endothelial (EC) and ECV304 cells, and aortic segments from old (18 mo) C57BL6-F₂ mice, propagated on low-AGE diet (LAGE), or LAGE supplemented with MG (LAGE+MG). Wild-type EC and sh-mRNA-AGER1⁺ EC, but not AGER1⁺ EC, had high NOX p47^phox and gp91^phox activity, superoxide anions, and NF-κB p65 nuclear translocation in response to MG and N^ε-carboxymethyl-lysine. These events involved epidermal growth factor receptor-dependent PKC-δ redox-sensitive Tyr-311 and Tyr-332 phosphorylation and were suppressed in AGER1⁺ ECs and enhanced in sh-mRNA-AGER1⁺ ECs. Aortic ROS, PKC-δ Tyr-311, and Tyr-332 phosphorylation, NOX expression, and nuclear p65 in older LAGE+MG mice were significantly increased above that in age-matched LAGE mice, which had higher levels of AGER1. In conclusion, circulating AGEs induce NADPH-dependent ROS generation in vascular aging in both in vitro and in vivo models. Furthermore, AGER1 provides protection against AGE-induced ROS generation via NADPH.

increased levels of extracellular advanced glycation end products (AGEs) promote vascular complications related to diabetes and aging (11, 13, 14, 33) by several mechanisms, including the generation of excess intracellular reactive oxygen species (ROS) (15, 22, 44). These can trigger transcriptional factor activity, including NF-κB, activator protein-1, and forkhead box O (7, 16, 68), leading to increased inflammation and/or vascular injury (11, 33, 41). While AGEs have pathophysiological significance, specific in vivo relevant AGE structures, the mechanism(s) by which they trigger ROS, and the dominant oxidant species that are involved remain unknown (1, 45, 73).

NADPH oxidase (NOX), a major cellular system regulating ROS, has been implicated in the induction of increased NF-κB activity after exposure to AGEs (5, 72, 74). NOX is a complex multi-subunit enzyme consisting of cytosolic (i.e., p47^phox, p67^phox, p40^phox, Rac1) and membrane-bound components (i.e., gp91^phox and p22^phox). Homologs of these subunits are found in most cell types (4), and phosphorylation of cytosolic components (i.e., p47^phox) prompts their translocation to the membrane. Assembly of the NOX complex is required for the activated enzyme to catalyze the formation of active superoxide anions (6, 46). In vitro and in vivo studies suggest that many PKC isoforms, including the novel PKC subfamily (including PKC-δ), are upstream regulators of NOX (9, 29, 38). The site-specific activation of PKC-δ, a redox-sensitive kinase, promotes membrane translocation and phosphorylation of NOX p47^phox (2, 55, 62).

AGE receptor 1 (AGER1) is a cell surface-associated receptor that opposes excessive ROS generation by AGEs. AGER1 is linked to the endocytosis and removal of AGEs (15, 49) and suppression of receptor for AGE (RAGE), MAPK, and NF-κB activity, apparently via inhibition of AGE-induced ROS generation (15, 49). One mechanism of AGER1 action is the inhibition of AGE-induced epidermal growth factor receptor (EGFR) transactivation and the Tyr-phosphorylation of ERK, and the serine-36 phosphorylation of p66^shc, which promotes the nuclear localization of FKHRL1 and synthesis of the antioxidant MnSOD (16). This link between AGER1 and innate antioxidative stress (OS) defense was supported by studies in AGER1-transgenic mice fed a high-fat diet (65). The high AGER1 expression in these AGER1 transgenic mice resulted in reduced OS, amelioration of inflammatory vascular injury, and insulin resistance. Thus AGER1 appears to control the activation of distinct cellular pathways initiated by extrinsic AGEs, in which increased ROS generation is a common denominator. However, key events at the interface of specific extracellular AGEs and the cell membrane, especially of vascular endothelium, have not been elucidated. Although AGEs have been suggested to promote NOX activation and ROS, via receptors such as RAGE (72) or EGFR (15), the precise mechanisms involved are not fully understood (1, 45, 73).

AGER1 protein colocalizes with AGE antigens in animal and human endothelial cells (ECs) and vascular tissues (59, 60). Under normal conditions, the levels of AGER1 correlate with ambient AGE levels (35). However, in the presence of sustained, elevated levels of OS (i.e., due to chronic diabetes or kidney disease), AGER1 levels decline (35, 36). This is a potential mechanism for the worsening of oxidant stress and vascular injury with aging and chronic diseases.

Sustained isocaloric restriction of oxidant intake in normal mice, via a low-AGE (LAGE) diet, leads to lower systemic OS, increased resistance to diabetes and cardiorenal disease of aging, and extended lifespan (13). The stable low-OS phenotype in LAGE mice is reversible by a diet supplemented with methylglyoxal (LAGE+MG) (17). Tissue AGER1 expression in LAGE mice is enhanced compared with mice fed standard chow, possibly because AGER1 expression is reduced because of the high content of AGEs found in normal mouse chow (13, 17).

Based on these observations, we assessed the effects of AGER1 expression on AGE-induced ROS generation and signal activation in ECs in vitro and in aortic rings from older mice exposed to MG, a chemically defined AGE. AGER1 overexpression in human EC reduced AGE-stimulated NOX-dependent superoxide anion production and NF-κB p65 activity, via suppression of PKC-δ, a redox-sensitive kinase. Likewise, PKC-δ expression in aortic rings from AGE-restricted (LAGE) mice was reduced, and AGER1 expression was increased. In contrast, PKC-δ expression, as well as NOX-dependent ROS and NF-κB activity, was markedly increased in aortas from older mice fed the diet. In summary, using cell culture and in vivo models, we provide evidence that circulating AGEs induce NADPH-dependent ROS generation in the vasculature. We also demonstrate that AGER1 protects against AGE-induced ROS generation via NADPH.

MATERIALS AND METHODS

Reagents.

Anti-AGER1 (OST48), anti-NF-κB p65, and anti-phosphorylated Tyr-311 and Tyr-332 of PKC-δ were purchased from Santa Cruz Biotechnology; anti-EGFR was from Upstate Biotechnology; anti-Shc and anti-phosphotyrosine PY20 were purchased from BD Biosciences; anti-PKC-δ was from Cell Signaling Technology; and anti-p47^phox from Millipore (Santa Cruz Biotechnology). Rottlerin was obtained from Calbiochem (San Diego, CA). Apocynin was from Sigma-Aldrich (St. Louis, MO), and AG1478 was from Calbiochem (La Jolla, CA). Endotoxin-free bovine serum albumin (BSA), which was used to prepare N^ε-carboxymethyl-lysine (CML)-BSA and MG-BSA, were characterized by HPLC and GC-MS (15, 16). Endotoxin-free AGE-BSA was prepared by incubating endotoxin-free BSA with 0.5 M d-glucose in PBS buffer at 37°C for 6 wk under sterile conditions. An endotoxin-binding affinity column (Pierce, Rockford, IL) was used to remove endotoxin from AGE-BSA and native BSA (tested by a Limulus amoebocyte lysate assay; Bio Whittaker, Walkerville, MD) (15).

Cell culture.

Human ECs (ECV304 cell line) from the European Collection of Cell Cultures (Wiltshire, UK) were cultured in medium 199 supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% FBS at 37°C in a humidified incubator with a mixture of 95% air and 5% CO₂. Human aortic ECs (HAEC) were obtained from ScienCell Research Laboratories (Carlsbad, CA). The cells were cultured in EC medium (Carlsbad, CA) with EC growth supplement and 5% FBS. Before treatment, cells were washed in PBS and incubated with serum-free media for 12 h.

Stable ECV304 cell line with plasmid-based AGER1 shRNA.

AGER1 knockdown was achieved using two SureSilencing short hairpin RNA (shRNA) plasmids provided by SuperArray Bioscience (Frederick, MD): plasmid 1 with insert sequence (5′-3′), CAACCATCGTTGGGAAATCAT; and plasmid 2 with insert sequence (5′-3′), TGACATTCAGCTGGAGTTTGT. ECV304 cells were stably transfected with two shRNA plasmids and selected in G418 containing medium (500 μg/ml) for 6 wk. At least a 70% target protein knockdown was confirmed by Western blot analysis.

Transient transfection.

After overnight incubation in serum-free medium, cells were transiently transfected with wild-type (WT) AGER1 or vector alone (49), using lipofectamine plus reagent (GIBCO-BRL). Forty-eight hours later, cells were treated with test reagents for the desired time period, and cell monolayers and supernatants were harvested for Western blot analysis.

Protein extraction.

Cellular membranes and nuclear proteins were prepared from ECV304, HAEC cells, and mouse aortic segments using a kit purchased from BioChain Institute (Hayward, CA). Cells were centrifuged, and the pellet was washed with PBS 3×, and then ice-cold buffer C (HEPES pH 7.9, MgCl₂, KCl, EDTA, sucrose, glycerol, sodium orthovanadate) was added (1 ml to 1 × 10⁷ cells), following which the mixture was rotated for 20 min at 4°C and then passed through a syringe with a 26- to 30-gauge needle 50–90 times to disrupt cell membranes and release nuclei. The lysate was then centrifuged at 15,000 g for 20 min at 4°C. The supernatant constituted the cytosolic fraction. The pellet was dissolved in buffer N (HEPES pH 7.9, MgCl₂, NaCl, EDTA, glycerol, and sodium orthovanadate) and rotated for 20 min at 4°C. After centrifugation at 15,000 g for 20 min at 4°C, the nuclear proteins were removed with the supernatant. The pellet was resuspended in buffer M (HEPES pH 7.9, MgCl₂, KCl, EDTA, sucrose, glycerol, sodium deoxycholate, NP-40, sodium orthovanadate) and centrifuged at 15, 000 g for 20 min at 4°C to extract the membrane proteins.

Western blot analysis.

Equal amounts of proteins extracted from cells or mouse aortas were separated on 8 or 10% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes. The membranes were blocked in TTBS (Tris-buffered saline with 0.1% Tween 20) containing 5% dry milk for 1 h and then probed with primary antibodies in TTBS with 5% dry milk overnight at 4°C. After washing, primary antibodies associated with the membranes were detected by appropriate peroxidase-conjugated secondary antibodies and the ECL chemiluminescence system (Roche, Indianapolis, IN). For immunoprecipitation, cells incubated overnight in serum-free medium were treated with the desired agonists at 37°C for the indicated times. Cells were rinsed in ice-cold PBS buffer and lysed in RIPA lysis buffer. After cells were disrupted by brief sonication, cellular protein concentration was determined. Anti-EGFR antibody (10 μg) or anti-PKC-δ antibody (5 μg) were added to cell lysates (800 μg), and the mixture was incubated for 1 h at room temperature, followed by further incubation at 4°C with 50 μl of protein G/A plus agarose beads (Santa Cruz Biotechnology) and constant rotation overnight. The immunoprecipitates were collected and washed by lysis buffer 3×. Proteins were then loaded onto 10 or 8% SDS-PAGE for Western blot analysis.

Lucigenin-enhanced chemiluminescence.

NOX activity in cells or mouse aortic rings was measured by lucigenin-enhanced chemiluminescence (39, 40, 52). In brief, confluent ECV304 cells or mouse aortas, cleared of adherent fat and cut into 2-mm rings, were preincubated for 45 min at 37°C in Krebs-HEPES buffer (99 mM NaCl, 4.7 mM KCl, 1 mM KH₂PO₄, 1.2 mM MgSO₄, 2.5 mM CaCl₂, 25 mM NaHCO₃, 20 mM Na-HEPES, and 11 mM glucose, pH 7.4) containing 100 μM of NADPH, with or without 1 mM apocynin (Sigma-Aldrich, St. Louis, MO), or EGFR inhibitor AG1478 (Calbiochem, La Jolla, CA). Cells or aortic rings were then exposed to CML-BSA (150 μg/ml), MG-BSA (30 μg/ml), or BSA (150 μg/ml) for 15 min at 37°C, with 5 μM lucigenin (Sigma-Aldrich, St. Louis, MO). After incubation, the plate was placed into a TopCount Microplate scintillation/luminescence count (Packard, Boston, MA), and tissue-dependent photon emission per second per well was monitored over a 20-min period. The NOX activity was expressed as relative light units per milligram of cellular protein in cells or aortic rings. Superoxide production was normalized to the weight of dried aortic tissue samples (39, 74).

Animals and AGE-supplemented dietary study.

C57BL/6 mice (Jackson Laboratory, Bar Harbor, MA) were individually caged and provided free access to water. At 2 mo of age, mice (F0) were placed on a LAGE diet, which had been prepared under reduced heat, but was otherwise identical to standard mouse chow (13, 17). The LAGE diet contained 3.4 × 10⁴ units of protein-associated CML and 0.14 × 10⁴ nmol of protein-associated MG in each gram of chow (17). LAGE diet-fed mice were maintained on the same diet during gestation, and pups from F2, (n = 6, males/females 3/3) were pair fed either a LAGE diet or a LAGE diet to which MG-BSA was added (LAGE+MG) (1 mg/g food) (17). The LAGE+MG diet contained 6.4 × 10⁴ units of protein-associated CML and 0.38 × 10⁴ nmol of protein-associated MG in each gram of chow (17). All animal studies were approved by the Mt. Sinai School of Medicine Animal Center Institutional Review Board.

Measurement of AGEs.

The concentration of AGEs in sera and protein extracts from aortic rings was determined by ELISA, using monoclonal antibodies for CML epitopes (4G9; Alteon, Northvale, NJ) and MG derivatives, i.e., MG-H1 (3D11) prepared in the authors′ laboratory (13, 17).

Determination of F2-isoprostanes (8-isoprostane).

Blood samples from each of the diet groups was collected at 3 and 18 mo for measurement of the endogenous lipid peroxidative product, 8-isoprostane, in fresh plasma using an enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI) (13).

RNA extraction and real-time RT-PCR.

Total RNA was extracted from freshly isolated mouse aortic segments using Trizol (Molecular Probes). Total RNA was reverse transcribed using Superscript III RT (Invitrogen). Quantitative SYBR Green real-time PCR assay was performed to measure the expression of mRNA for NOX isoform p47, gp91, and Rac1. The thermal cycle condition includes 40 cycles of denaturation at 95°C for 15 s, annealing at 55°C for 20 s, and elongation at 72°C for 30 s. Sequences of the primers used for real-time PCR were as follows: for p47, forward primer 5′-GGCACAAAGGACAATCCATC-3′, reverse primer 5′-AGGGATAGGAGCCGTCTAGG-3′; for gp91, forward primer 5′-TCAAGTGTCCCCAGGTATCC-3′, reverse primer 5′-CTTCACTGGCTGTACCAAAGG-3′(36); for Rac1, forward primer 5′-TGCGTTCCCTGGAGAGTACATCC-3′, reverse primer 5′-TTGAGTCCTCGCTGTGTGAGTGC-3′(52). β-actin and GAPDH gene products were used for internal normalization. The transcript copy number of target genes was determined based on their threshold cycle values.

Statistical analysis.

Experiments were repeated at least three times each. Results were expressed as means ± SD. Statistical significance between groups was determined by Student's t-test. A probability value of P < 0.05 was considered significant.

RESULTS

AGER1 suppresses AGE-induced NF-κB/p65 nuclear translocation in ECs.

Since both MG and CML AGEs have been shown to induce NF-κB activity in endothelial and other cell types (5, 7, 14), the effect of AGER1 on the nuclear translocation of NF-κB/p65 was assessed. Serum-starved ECV304 cells were treated with CML-BSA (CML), MG-BSA (MG), or BSA for 24 h. Immunoblot analysis of nuclear proteins revealed a significant increase in nuclear NF-κB/p65 in CML-treated and MG-treated cells, but not in BSA-treated mock cells (Fig. 1). Despite high levels of nuclear NF-κB/p65 protein, there was no loss of NF-κB/p65 protein from the cytoplasmic fraction (7). The AGE-induced NF-κB/p65 nuclear translocation was completely prevented in AGER1⁺ ECV304 cells (Fig. 1). However, in mock cells and in cells transduced with shRNA-AGER1, nuclear translocation of NF-κB/p65 was enhanced (Fig. 1). These data indicate that the mechanism by which AGER1 prevents AGE-induced NF-κB activation could be via blocking the nuclear translocation of NF-κB/p65.

Fig. 1.

Advanced glycation end-product receptor 1 (AGER1) expression decreases advanced glycation end-product (AGE)-induced NF-κB/p65 nuclear translocation. AGER1-ovexpressing (AGER1⁺), AGER1 knockdown [short hairpin RNA (shRNA)-AGER1⁺], or mock ECV304 cells were stimulated with N^ε-carboxymethyl-lysine (CML)-BSA (150 μg/ml), methylglyoxal (MG)-BSA (30 μg/ml), or BSA (150 μg/ml) for the indicated time intervals. Immunoblots for NF-κB/p65 protein in nuclear and cytoplasmic extracts are shown. β-Tubulin is the loading control. Data are representative of three identical experiments.

NOX activation is required for AGE-induced NF-κB/p65 nuclear translocation.

Since overproduction of ROS and NF-κB activation has been linked to NOX (8, 12, 26, 31), ECV304 cells were exposed to CML, MG, or BSA for 24 h, in the presence or absence of NOX inhibitors (diphenyleneiodonium or apocynin) to determine if NOX activity is required for AGE-induced NF-κB/p65 nuclear translocation. AGE-treated cells coincubated with apocynin showed a significantly lower level of NF-κB p65 nuclear translocation compared with AGEs alone (CML, by 80%; and MG, by 77%), and the levels were also lower in the presence of diphenyleneiodonium (CML, 68% and MG, 51%) (Fig. 2, A and B).

Fig. 2.

Endothelial NADPH oxidase (NOX) activation is required for AGE-induced NF-κB/p65 nuclear translocation. A: ECV304 cells were stimulated with CML-BSA, MG-BSA, or BSA, as per Fig. 1, in the presence or absence of NOX inhibitors diphenyleneiodonium (DPI; 10 μM) or apocynin (APO; 300 μM) for 24 h. The nuclear and cytoplasmic fractions were probed with anti-NF-κB/p65 or anti-β-tubulin. B: densitometric analysis of 3 independent experiments. Shown are means ± SD. *P < 0.01 vs. unstimulated controls (C). #P < 0.05 vs. stimulated cells.

AGER1 inhibits AGE-induced NOX activation.

AGE-induced NOX activity was assessed in ECV304 cells by lucigenin-enhanced chemiluminescence, to assess superoxide anion production (39). After stimulation with CML or MG for 20 min, NOX-derived superoxide production was significantly reduced by apocynin (Fig. 3A). Comparison of AGER1⁺ cells to mock cells revealed that superoxide production was significantly reduced in AGER1⁺ cells (Fig. 3B).

Fig. 3.

CML-BSA- and MG-BSA-stimulated superoxide anion production is largely NOX dependent in endothelial cells (ECs) and is inhibited by AGER1. A: ECV304 cells were stimulated with CML-BSA, MG-BSA, or BSA, as per Fig. 1, for 20 min, with or without apocynin (1 mM). Lucigenin-enhanced chemiluminescence (LEC) was used to assess NOX-dependent superoxide production. Untreated cells served as controls (C). Data are shown as percent relative LEC units (%RLU) above control. Inset: total RLU × 10³/1 × 10⁶ cells. *P < 0.05, AGE-treated vs. none. #P < 0.05, AGE⁺ apocynin vs. AGE⁺ without an inhibitor. B: AGER1⁺ or mock ECV304 cells treated as in A. Data are means ± SD from 3 independent experiments. Inset: total RLU × 10³ units/1 × 10⁶ cells. *P < 0.05 vs. mock controls. #P < 0.05, AGE-treated AGER1⁺ vs. AGE-treated mock cells.

Since phosphorylation and association of the NOX cytosolic subunit, p47^phox, with membrane-bound components is a critical step for enzyme activity (21), AGER1⁺ and shRNA-AGER1⁺ were compared with mock ECV304 cells. Stimulation of mock cells with CML-BSA and MG-BSA led to a time-dependent increase in p47^phox translocation to membranes, accompanied by a decrease in cytosolic p47^phox protein (Fig. 4A). Note that the AGE-mediated translocation of p47^phox protein was greater in shRNA-AGER1⁺ cells than in mock cells at 15 and 30 min in cells treated with CML-BSA and MG-BSA, but not in BSA-treated cells (Fig. 4B). This translocation was significantly attenuated in AGER1⁺ cells at all time points studied. Therefore, the reduction of p47^phox protein translocation to the membrane by AGER1 may be indirectly responsible for the inhibition of NF-κB/p65 nuclear translocation.

Fig. 4.

EC AGER1⁺ inhibits CML-BSA- or MG-BSA-induced NOX p47^phox translocation to cell membranes. A: membrane and cytosolic fractions from AGER1⁺, shRNA-AGER1⁺, or mock ECV304 cells, stimulated with CML-BSA, MG-BSA, or BSA, as per Fig. 1, for the indicated times were subjected to Western blot analysis by anti-p47^phox antibody. B: densitometric analyses showing membrane-to-cytosolic ratio from 3 independent experiments are shown as means ± SD. *P < 0.05 vs. 0 min. #P < 0.05 vs. mock ECV304 cells at the same time points.

AGER1 blocks AGE-mediated NOX activity through negative regulation of EGFR.

Since activation of EGFR by oxidants induces NOX (19, 27, 28, 42 ) and AGEs promote EGFR activation (15), EGFR could serve as an upstream link in AGE-induced NOX activity (56, 61). AGE-BSA, which contains both CML and MG modifications, stimulated p47^phox protein translocation from the cytoplasm to the membrane (Fig. 5A). The AGE-induced translocation was prevented in the presence of the EGFR inhibitor AG1478 in HAEC cells (Fig. 5B). Similar results were found in ECV304 cells (data not shown). An EGFR inhibitor (AG1478) led to a significant reduction in chemiluminescence in ECV304 cells, consistent with reduced superoxide production (Fig. 6A). These results indicated that transactivation of EGFR was involved in AGE-promoted NOX activation and ROS generation.

Fig. 5.

AGE-stimulated NOX p47^phox membrane translocation requires epidermal growth factor receptor (EGFR) transactivation. Human aortic ECs (HAEC) pretreated with AG1478 (10 μM) for 15 min were exposed to AGE-BSA (150 μg/ml) for 5–30 min. Membrane and cytosolic fractions were subjected to Western blot analysis and probed with anti-p47^phox antibody. β-Tubulin was used as a control. HAEC was stimulated by AGE-BSA alone (A) or after treatment with AG1478 (B) and AGE-BSA for 5–30 min. Data are representative of 3 experiments. Similar data were also obtained from ECV304 cells (data not shown).

Fig. 6.

EC NOX-dependent superoxide anion production by AGEs requires EGFR transactivation, but is inhibited by AGER1. A: ECV304 cells were stimulated with CML-BSA, MG-BSA, or BSA, as per Fig. 1, for 20 min. ECV304 cells were stimulated with CML-BSA, MG-BSA, or BSA in the presence of AG1478 (10 μM). B: wild-type (WT), EGFR⁺, and EGFR⁺-AGER1⁺ ECV304 cells were AGE-stimulated as in A. LEC was used to measure NOX-dependent superoxide production. Data are means ± SD from 3 experiments. Untreated cells were used as controls (C). A and B: percent RLU increase above control. Insets: total RLU (units × 10³/1 × 10⁶ cells). *P < 0.05 vs. no-inhibitor control or untreated WT cells. #P < 0.05 vs. no-inhibitor or WT cells treated with CML-BSA or MG-BSA. §P < 0.05 vs. EGFR⁺-treated cells.

Superoxide production triggered by CML-BSA and MG-BSA was also assessed in EGFR overexpressing (EGFR⁺) cells and in EGFR⁺ AGER1 coexpressing (EGFR⁺-AGER1⁺) ECV304 cells. Superoxide production promoted by either CML-BSA or MG-BSA was significantly increased in EGFR⁺ cells compared with WT cells. However, in treated EGFR⁺-AGER1⁺ cells superoxide production remained at the levels present in treated WT cells (Fig. 6B). These data suggest that EGFR potentiates AGE-mediated superoxide production, and that AGER1 suppresses AGE-mediated NOX activity, in part through the EGFR pathway.

AGER1⁺ interferes with the interaction of EGFR with PKC-δ promoted by AGEs.

Oxidants, e.g., H₂O₂, have been implicated in the activation of NOX by PKC-δ, via EGFR transactivation (10, 24, 28). Since AGEs promote, and AGER1 blocks, EGFR activation and adaptor recruitment, i.e., p66^Shc (16, 49), PKC-δ was assessed in AGE-stimulated mock AGER1⁺ or shRNA-AGER1⁺ ECV304 cells. After 15 min of AGE stimulation, precipitation with anti-PY20 revealed that EGFR Tyr-phosphorylation was increased in mock cells (Fig. 7A). Immunoprecipitation with anti-PKC-δ and anti-p66^Shc revealed that PKC-δ and p66^Shc coprecipitated with EGFR, possibly as a complex. Furthermore, reduction of AGER1 levels (shRNA-AGER1⁺ cells) resulted in increased levels of activated EGFR and PKC-δ in the precipitate. Examination of precipitates using anti-PKC-δ and anti-p66^Shc antibodies revealed that PKC-δ and p66^Shc coprecipitated with EGFR, possibly as a multicomponent complex. Furthermore, reduction of AGER1 levels (shRNA-AGER1⁺ cells) resulted in increased levels of activated EGFR and PKC-δ in this complex. In contrast, AGE-mediated complex formation was not found in AGER1⁺ cells, and PKC-δ was not associated with EGFR (Fig. 7A). The net effect of AGER1 in blocking EGFR/PKC-δ interactions and subsequent NOX activation could provide a mechanism to explain the negative effects of AGER1 on ROS generation by AGEs (56, 61, 64).

Fig. 7.

The effects of AGER1 on AGE-promoted EGFR, PKC-δ, and p66^Shc association. A: AGER1⁺ or shRNA-AGER1⁺ ECV304 cells were stimulated with MG-BSA (30 μg/ml) for 15 min. Cell lysates were immunoprecipitated (IP) using anti-EGFR antibody and immunoblotted with anti-p66^Shc, anti-PKC-δ, anti-phosphotyrosine PY20, anti-AGER1, and anti-EGFR antibodies. Representative images of 3 independent experiments are shown. B: ECV304 cells and HAEC cells were exposed to MG-BSA (30 μg/ml), H₂O₂ (5 mM), or EGF (10 ng/ml) for 15 min. Endogenous PKC-δ was immunoprecipitated with anti-PKC-δ antibody, and immunoblotting was performed using anti-pY311, anti-pY332, and anti-PKC-δ antibodies. Images are representative of 3 experiments. C: ECV304 and HAEC cells were exposed to MG-BSA (30 μg/ml) for 15 min, with or without AG1478 (10 μM) (for 15 min). Endogenous PKC-δ was immunoprecipitated using anti-PKC-δ antibody, and immunoblotting was performed using anti-pY311, anti-pY332, anti-p66^Shc, and anti-PKC-δ antibodies. Images are representative of 3 experiments. D: HAEC stimulated by AGE-BSA (150 μg/ml) and rottlerin (2 μM) for up to 30 min. Membrane and cytosolic fractions were subjected to Western blot analysis and probed with anti-p47^phox antibody. β-Tubulin was used as control. Data are representative of 3 experiments. Similar data were also obtained from ECV304 cells (data not shown).

AGEs promote site-specific phosphorylation of PKC-δ in AGE-stimulated cells.

Since PKC-δ exhibits catalytic activity in response to H₂O₂ (43, 51, 66), this isoform was immunoprecipitated from ECV304 and HAEC cells stimulated with AGEs, H₂O₂, or EGF for 15 min. Unstimulated cells served as controls. Phosphorylation of specific PKC-δ tyrosine residues was assessed by immunoblotting with antibodies to phospho-Tyr-332 (pY332) and phospho-Tyr-311 (pY331). Both Tyr-332 and Tyr-311 were phosphorylated in ECV304 and HAEC cells treated with AGEs or H₂O₂ (Fig. 7B), but little or no effect was noted in EGF-treated cells (Fig. 7B). These data were consistent with redox-dependent PKC-δ phosphorylation at Tyr-332 and Tyr-311 (43, 51, 66). Since PKC-δ phosphorylation appeared to be oxidant sensitive but is not modified by EGF, cells were exposed to AGEs, with or without the EGFR kinase inhibitor AG1478 for 15 min and immunoprecipitated with anti-PKC-δ. PKC-δ Tyr-phosphorylation (at Tyr-332 and Tyr-311) was markedly attenuated in cells pretreated with AG1478 (Fig. 7C), suggesting that EGFR activation is necessary for AGE-stimulated Tyr-phosphorylation of PKC-δ at redox-sensitive sites. To determine whether the activation of PKC-δ is directly involved in NOX activation by AGEs, a selective PKC-δ inhibitor (rottlerin) (34) was used to assess time-dependent p47^phox translocation in AGE-stimulated HAEC cells. The addition of rottlerin completely blocked AGE-induced membrane translocation of p47^phox protein (Fig. 7D), compared with AGEs alone (Fig. 5A). These data suggested that PKC-δ is a key part of the pathway involved in AGE-induced NOX activity in HAECs.

In vivo studies: increased levels of ambient AGEs in mice are associated with increased PKC-δ activation, NOX expression, ROS, and NF-κB activity.

One of two groups of pups from dams fed a LAGE diet was maintained on the LAGE diet, and the second group was fed the same LAGE diet that had been supplemented with MG-modified BSA (LAGE+MG) (17). These diets were maintained throughout life. The LAGE+MG group at 18 mo had significantly higher levels of both CML and MG in serum (Fig. 8A) and in aortic tissue (Fig. 8B), and higher levels of plasma oxidized lipids (8-isoprostanes) (Fig. 8C), compared with the LAGE group. There were no differences between the two groups at baseline with respect to these three measures. Aortic rings from 18-mo-old LAGE+MG mice produced significantly higher amounts of superoxide anion, compared with age-matched LAGE animals and compared with both groups at baseline (Fig. 8D). This increase was completely abolished by the addition of apocynin to extracts of aortic rings in both groups at 18 mo of age. These data suggest that vascular NOX activation and ROS production in vivo resulted from the addition of a specific AGE (MG) to the food.

Fig. 8.

Effect of oral methyl-glyoxal on NOX activation and NF-κB activity in vivo. CML and MG serum levels (A), tissue CML and MG levels (B), plasma 8-isoprostane (C), and superoxide anion production (D), in the presence or absence of apocynin (1 mM), were assessed in aortic rings from 3- or 18-mo C57BL6-F₂ mice, fed low AGE (LAGE) or MG-supplemented LAGE (LAGE+MG) (n = 6/group). LEC was used to measure NOX-dependent superoxide production. Data are means ± SD of 3 experiments. *P < 0.05 vs. 3 mo. #P < 0.05 vs. 18-mo LAGE+MG. §P < 0.05 vs. 18 mo without apocynin treatment. E: aortic PKC-δ Tyr-311 and Tyr-332 phosphorylation. Aortic tissue homogenates from 18-mo LAGE- or LAGE+MG-fed mice were separated by SDS-PAGE and immunoblotted with the relevant antibodies. Densitometric analysis is of 3 independent experiments. Shown are means ± SD. *P < 0.05 vs. 18-mo LAGE+MG. F: transcript gene copy number of NOX isoforms p47^phox, gp91^phox, and Rac1 in 18-mo mouse aortic rings by real-time RT-PCR (n = 5/group). *P < 0.05 vs. LAGE+MG. G: aortic NF-κB/p65 nuclear translocation. Nuclear and cytoplasmic fractions were separated from 18-mo LAGE or LAGE+MG mouse aortas and probed with anti NF-κB/p65 or β-tubulin antibody. Densitometric analysis of 3 independent experiments are shown. Shown are means ± SD. *P < 0.05 vs. LAGE+MG. H: AGER1 protein expression in aortas from 18-mo LAGE- and LAGE+MG-fed mice. Densitometric data are shown. *P < 0.05 vs. LAGE+MG.

Furthermore, PKC-δ Tyr-311 and Tyr-332 tyrosine phosphorylation was higher in aortic tissue from LAGE+MG mice (Fig. 8E). Moreover, aortic tissue from LAGE+MG mice had increased levels of p47^phox, p91^phox, and Rac1 mRNA, compared with LAGE mice (Fig. 8F). Nuclear levels of NF-κB/p65 were also increased in the LAGE+MG mice (Fig. 8G). The LAGE mice had higher levels of tissue AGER1 expression at 18 mo of age (Fig. 8H). The levels of AGER1 were inversely correlated with reduced levels of PKC-δ, NOX gene expression, and NF-κB/p65 nuclear translocation, compared with the LAGE+ MG group. Taken together, these data indicated that the maintenance of normal levels of AGER1 protein expression in vivo was associated with protection against vascular oxidant stress due to food-derived AGEs. Thus the suppression of AGER1 that occurs as a result of chronic exposure to high levels of oral AGEs could lead to increased vascular injury.

DISCUSSION

The origin of excessive ROS contributing to the increased risk of vascular disease has been unclear. The present study provides in vitro and in vivo evidence that the addition to food of a specific AGE moiety, which is normally present in vivo and in many foods, increases superoxide anion production in cultured vascular ECs and aortic tissues in vivo. ROS production in EC by specific AGEs was mediated by enhanced site-specific Tyr-phosphorylation of PKC-δ, a redox-sensitive kinase that modulated vascular NOX activity, and NF-κB/p65 nuclear translocation. Both ROS production and site-specific Tyr-phosphorylation of PKC-δ were suppressed in the presence of high levels of AGER1. Thus the inhibitory action of AGER1 appears to be mediated by suppression of EC membrane NOX assembly and ROS generation. These events induce NF-κB transcriptional activity, which was also found to be suppressed by AGER1 in the present study. The in vitro data were corroborated in aortic rings isolated from aging mice exposed to an AGE-supplemented (LAGE+MG) diet. Whereas aortic rings from mice kept on a LAGE diet had normal AGER1 levels, the AGER1 levels were reduced in LAGE+MG mice. The findings highlight the fact that “extrinsic” AGE structures can promote ROS and modulate AGER1 expression, and that these AGEs are a common part of the environment. Collectively, the findings reinforce the importance of EC AGER1 as a critical element of the defense system against vascular disease.

The levels of ROS, which serve as second messengers for redox-sensitive transcriptional factors, are carefully titrated, but high levels can be cytotoxic (3). Ambient AGE-modified molecules are among the physiologically active substances that can disturb this balance (5, 70). In the present study, using nuclear translocation of NF-κB/p65 in ECs as a key biological end point, we found that NF-κB/p65 translocation was markedly increased by the addition of two chemically defined AGE moieties, CML and MG. The fact that similar results were obtained by structurally unrelated AGEs that are present in large amounts in standard western food emphasizes the importance of AGEs as prooxidants and the fact that both AGEs are AGER1 ligands (71). Moreover, the prooxidant properties of AGEs presented to the surface of ECs appear to be largely due to NOX-derived superoxide anions, since NF-κB/p65 nuclear translocation was largely inhibited by a NOX-specific inhibitor (26, 31, 58).

ROS induction by either AGEs or H₂O₂ has been previously shown to be blocked by AGER1 overexpression in some cell types (15, 16, 49) and is now extended to human EC. Specifically, NOX-derived superoxide production, as well as NF-κB activation in response to AGEs, was significantly decreased in ECs overexpressing AGER1. The specificity of this response was demonstrated by the fact that silencing of AGER1 promoted further NOX-derived superoxide production. These events, coupled with the significant decrease in p47^phox membrane translocation in AGE-treated AGER1-overexpressing EC, show that AGER1 acts as a potent regulator of NOX in these cells.

The AGE-induced translocation of p47^phox to the membrane was blocked by a selective inhibitor of PKC-δ, a kinase sensitive to extrinsic oxidants (37, 53), suggesting that induction of PKC-δ kinase activity is required for NOX activation and superoxide anion generation by AGEs in ECs. Consistent with this view, AGEs induced marked phosphorylation of PKC-δ at residues Tyr-311 and Tyr-332, as effectively as H₂O₂. Therefore, this isoform may constitute an important EC membrane-associated component of AGE signaling, although further studies are required to confirm the dependence of these events on site-specific phosphorylation of PKC-δ.

These findings are consistent with earlier studies that revealed that AGEs promote EGFR transactivation via EGFR/Shc/Grb2 complex formation (15). Since PKC-δ operates in the context of such a complex with EGFR (51), this kinase could represent a key target for AGEs. Indeed, the marked induction of PKC-δ by AGEs was suppressed in the presence of an EGFR kinase-specific inhibitor. Since EGF had no effect on PKC-δ activity (51), the response to AGEs was consistent with a mechanism of EGFR activation by oxidants, which is distinct from that directly caused by EGF interactions with EGFR (15, 25, 50, 51). Furthermore, AGE-induced PKC-δ recruitment and activation in EC were completely blocked in the presence of AGER1 overexpression, whereas AGER1 silencing had the opposite effect. These findings provide further evidence to support the postulate that AGER1 plays an important role in the suppression of excess ROS generated in EC after exposure to extrinsic AGEs (Fig. 9). Alternatively, AGEs could promote, and AGER1 could suppress, EGFR/PKCδ activity via RAGE-promoted Src kinases (18), since oxidant-induced EGFR activation by the Src family of kinases has been described (20, 75). While AGER1 negatively interferes with RAGE-dependent NF-κB activation in mesangial cells (49), this must be examined in EC. While AGER1 physically interacts with surface EGFR (15), RAGE has not been found to contribute to similar interactions.

Fig. 9.

Reactive oxygen species (ROS) generation is increased by AGEs and suppressed by AGER1 in vascular EC. Putative mechanism: 1) extrinsic AGEs promote EC membrane redox phosphorylation of PKC-δ via EGFR transactivation; 2) PKC-δ prompts membrane translocation of NOX p47^phox and NOX-dependent superoxide anion generation, which facilitates p65 movement to the nucleus; 3) EGFR/PKC-δ also induces Ser-36 phosphorylation of p66^Shc, inhibiting Forkhead-Box class O (FOXO) antioxidant function. AGER1 is physically associated with EGFR and its adaptors. Through its long extracellular domain, AGER1 binds AGEs, preventing overt ROS generation and adverse events via this route. A LAGE diet is associated with protected levels of AGER1 and amelioration of vascular changes in aging mice. AGER1 is reduced by prolonged AGE excess.

The in vivo relevance of these in vitro findings was further supported by in vivo studies. While a role for NOX has been suggested in many disease conditions associated with high-systemic OS and AGEs, including diabetes, obesity, and cardiovascular disease (23, 32, 57, 64, 67), the direct effects of external sources of oxidants, such as AGE-modified nutrients on vascular cells, have been addressed in only a few studies (14, 48). The LAGE mice utilized in the present study were propagated from a line raised on a LAGE diet. As a result, they had stably decreased levels of systemic OS and an extended life span (13, 17). The evidence supporting oral AGEs as prooxidants in vivo is at least twofold. First, when the LAGE diet was supplemented with MG (LAGE+MG), there was an increase in AGEs and OS in vivo (17). Second, long-term feeding with a high-AGE diet resulted in a shortened life span (13, 17). In the present report, aortic rings from mice fed the MG-supplemented diet generated higher superoxide anions ex vivo, which was completely blocked by a NOX-specific inhibitor. High levels of ROS in aortic rings from these mice were also associated with higher levels of PKC-δ Tyr-311 and Tyr-332 phosphorylation and NOX gene expression, relative to age-matched LAGE mice. These observations suggested that PKC-δ is an upstream regulator of vascular NOX-induced ROS in vivo, in the presence of a high AGE burden.

Signal pathways regulated by AGE-derived ROS are cell type specific and may be associated with an increase of proinflammatory or proapoptotic events, and/or with suppression of antioxidant systems (37, 53, 54, 63). The significant increase in nuclear translocation of NF-κB/p65 in aortic rings from LAGE+MG-fed mice, but not LAGE-fed mice, confirms the role of NF-κB as one of many transcriptional elements likely to negatively impact on vascular endothelium in animals and humans who are chronically exposed to high levels of external prooxidant AGEs (7).

PKC-δ Tyr-phosphorylation was attenuated, and AGER1 expression was higher in aortic rings of LAGE mice (13). However, mice on the LAGE+MG diet had low AGER1 levels and higher levels of PKC-δ Tyr-phosphorylation. The reciprocal relationship between vascular PKC-δ and AGER1 levels suggests that AGER1 suppresses oxidant stress by suppressing NOX activation both in vivo and in vitro. The inverse correlation between AGER1 and NF-κB/p65 activity in aortic tissue suggests that the anti-inflammatory effects of AGER1 are partly exerted via this mechanism. These findings implicate AGER1 in the negative regulation of NOX via PKC-δ in vivo following the ingestion of common extrinsic oxidants, including AGEs.

In conclusion, the present studies show that external AGEs can increase oxidant stress in EC via targeted site-specific phosphorylation of redox-sensitive PKC-δ, proceeding via EGFR transactivation and driving NOX-dependent superoxide generation and NF-κB transcriptional activity. These events are inhibited by AGER1 in vitro and in vivo by a mechanism that may partly involve AGE binding and internalization by AGER1. The end result is a reduction of the access of AGEs to EGFR and PKC-δ and suppression of downstream events (Fig. 9). The data also raise the possibility that EGFR may contribute to a broader spectrum of conditions than previously thought, i.e., vascular disease due to diabetes or aging. On the other hand, if AGER1 levels are attenuated, the vascular endothelium may be deprived of an important protective mechanism, and external AGEs could stimulate abnormal cell growth via EGFR. Laboratory animal food, as well as human food, may serve as common sources of oxidants in amounts that could have pathological effects. In view of these data and previous in vivo findings (30, 47, 69, 71), reduced exposure to AGEs from external sources may confer sustained protection against vascular oxidant injury that has heretofore been attributed to aging.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant R01 HL073417 (to H. Vlassara).

DISCLOSURES

I am not aware of financial conflict(s) with the subject matter or materials discussed in this manuscript with any of the authors, or any of the authors' academic institutions or employers.