Requirement for p38 and p44/p42 Mitogen-Activated Protein Kinases in RAGE-Mediated Nuclear Factor-κB Transcriptional Activation and Cytokine Secretion

All tissue culture reagents were purchased from Life Technologies and were documented to be free of endotoxins. HSA (Fraction V), glyoxylic acid, and sodium cyanoborohydride were obtained from Sigma (St. Louis, MO). N-acetyl-l-cysteine (NAC), genistein, PD98059, and SB203580 were from Calbiochem (San Diego, CA). Anti-phosphotyrosine antibody (PY99) was from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-ERK1/2, anti–phospho-ERK1/2, anti-p38, anti–phospho-p38, anti–c-Jun NH₂-terminal kinase (JNK), anti–phospho-JNK, and anti–phospho-ATF2 antibodies were purchased from New England BioLabs (Beverly, MA). Synthetic peptides were prepared by Dr. Kaijun Ren (Chemistry Department, Texas Biotechnology Corporation) and were >95% pure, as assessed by high-performance liquid chromatography. The expression plasmid for the p38 kinase-dead dominant-negative mutant was provided by Dr. Roger Davis from the Howard Hughes Medical Institute at the University of Massachusetts Medical Center (Worcester, MA). The IκB mutant was a gift from Dr. Chung Y. Hsu at Washington University School of Medicine (St. Louis, MO). Both mutants are in the pcDNA3.1 expression vector. Enzyme-linked immunosorbent assay (ELISA) kits for TNF-α, IL-1β, and MCP-1 were from R&D Systems (Minneapolis, MN).

CML-HSA was prepared as previously described (34). Briefly, 175 mg/ml HSA was incubated at 37°C for 24 h in 0.2 mol/l sodium phosphate buffer (pH 7.8) containing 150 mmol/l glyoxylic acid and 450 mmol/l sodium cyanoborohydride, followed by extensive dialysis against 2 L phospate-buffered saline at 4°C changed twice at 24-h intervals. The resulting CML-HSA preparation contained ∼30% CML-modified lysine residues per molecule albumin based on amino acid analysis. Both the CML-HSA and the control HSA were documented to be endotoxin-free by use of the Limulus amebocyte lysate assay (Sigma E-TOXATE). The lower limits of detection were 0.015 EU/ml (turbid gel), 0.03 EU/ml (soft gel), and 0.06 EU/ml (hard gel).

Human THP-1 monocytic leukemia cells were obtained from the American Type Culture Collection (ATCC TIB-202; Rockville, MD) and maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 50 μmol/l 2-mercaptoethanol, 2 mmol/l glutamine, 100 U penicillin, and 100 μg/ml streptomycin. Cells were grown at 37°C in a 5% CO₂ atmosphere. Log-phase cells were used for all experiments.

Human peripheral monocytes were isolated from healthy donors’ blood (SeraCare, Oceanside, CA) using standard Ficoll-Hypaque (Amersham Pharmacia Biotech, Uppsala, Sweden) density-gradient centrifugation followed by adherence depletion of lymphocytes as previously described (35). Cells were seeded at 4 × 10⁵ cells per well in a 48-well plate and treated with 100 μg/ml HSA or CML-HSA in the absence or presence of SB203580 or PD98059 for 24 h before measurement of TNF-α in the culture media by ELISA.

The human RAGE extracellular domain encompassing residues 23–340 (sRAGE) was expressed and purified from Escherichia coli using the pET thioredoxin system (Novagen, Madison, WI). For antiserum production, BALB/c mice were immunized with 100 μg sRAGE protein in complete Freund’s adjuvant, and an additional 50 μg sRAGE in incomplete Freund’s adjuvant was given every 4 weeks until a high serum titer was achieved. NF-κB reporter neutralizing activity of the antiserum was measured by preincubation with THP-1 cells (1:5 or 1:25 dilution of the antiserum) for 1 h before exposure to CML-HSA.

THP-1 cells were seeded at 5 × 10⁶ cells per 100-mm dish in 10 ml of serum-free medium (SFM) the day before transfection. Transfection was performed by the DEAE-dextran method as described previously (36), with minor modifications. Cells were washed once with SFM and resuspended in 1 ml of the same medium containing 2 μg NF-κB–Luc reporter plasmid (Clontech, Palo Alto, CA) and 200 μg/ml DEAE-dextran (Promega, Madison, WI). The cell-DNA mixture was incubated at room temperature for 20–30 min before washing, centrifugation, and resuspension into fresh SFM. Transfected cells were seeded into 96-well plates at 70,000 cells per well for recovery. After 24 h, cells were treated with the indicated amount of control albumin or CML-modified albumin for 10 min to 24 h before cell harvest. Equivalent amounts of cell lysates, normalized for total protein (Bradford protein assay; Bio-Rad, Palo Alto, CA), were used for measurement of luciferase activity. Luciferase assays were performed using the Steady-Glo luciferase assay system according to the manufacturer’s instructions (Promega), and luminescence was detected in a TopCount microplate scintillation counter, using a single-photon monitor program (Packard Instrument Company, Meriden, CT). For cotransfection experiments, 10 μg (10-fold molar excess) of expression plasmid was used with 1 μg of the NF-κB reporter, with the empty vector serving as a trans-fection control. Transfection efficiency in THP-1 cells was 20–30%, as monitored by cotransfection with a β-galactosidase expression plasmid (Clontech).

Cell extracts were prepared using the protocol of Schreiber et al. (37), with minor modifications. Briefly, 2–4 × 10⁶ cells were collected by centrifugation, washed in Tris-buffered saline, then resuspended by gentle pipetting in 400 μl cold buffer containing 10 mmol/l HEPES, pH 7.9, 10 mmol/l KCl, 0.1 mmol/l EDTA, 0.1 mmol/l EGTA, 1 mmol/l dithiothreitol (DTT), and 0.5 mmol/l phenylmethylsulfonyl fluoride (PMSF). The cell suspension was incubated on ice for 15 min, after which Nonidet-P40 was added to a final concentration of 1%. The homogenate was subjected to centrifugation for 10 s; the pellet was resuspended in 50 μl ice-cold buffer containing 20 mmol/l HEPES, pH 7.9, 0.4 mol/l NaCl, 1 mmol/l EDTA, 1 mmol/l EGTA, 1 mmol/l DTT, and 1 mmol/l PMSF; and the tube was rocked at 4°C for 30 min on a shaking platform then dialyzed against reaction buffer (20 mmol/l HEPES, pH 7.9, 20% glycerol, 50 mmol/l KCl, 0.2 mmol/l EDTA, 1 mmol/l MgCl₂, and 1 mmol/l DTT). The nuclear extract was spun for 5 min in the cold room. Total protein content of the supernatant was determined, with the remainder stored at −70°C in aliquots. Some 5 μg of the nuclear extract was mixed with 2 μg poly dI-dC and ∼200 pg (10,000 cpm) of radiolabeled double-stranded oligonucleotides containing the c-myc NF-κB binding site (CGTTGGGAAAACCCGGGATT) in reaction buffer to a final volume of 30 μl and incubated at room temperature for 25 min. Samples were run on a 5% (30:1) polyacrylamide gel in 0.5× Tris-borate-EDTA buffer.

Radiolabeled probes were prepared by labeling one of the oligonucleotides comprising the DNA binding site and then annealing it to its unlabeled partner. Radiolabeling was performed using T4 polynucleotide kinase and [γ-³²P]ATP. After this reaction, unincorporated nucleotides and the T4 kinase buffer were removed, the unlabeled oligonucleotide was added, and it was heated to 80°C and then allowed to cool slowly to room temperature.

A cDNA-encoding RAGE was generated by reverse transcriptase (RT)–polymerase chain reaction (PCR) using human lung mRNA as a source for the cDNA and primers based on the published full-length sequence (GenBank AF001095). The product was fully sequenced and then subcloned into the eucaryotic expression vector pcDNA3.1(⁺)neo. RAGE lacking its intracellular domain was constructed by PCR-based site-directed mutagenesis using primers to introduce a termination codon immediately after amino acid 364 (RAGE-365t) or amino acid 371 (RAGE-372t). RAGE-365t terminates two amino acids after the hydrophobic region thought to be the transmembrane domain, whereas RAGE-372t retains only a portion of the cytosolic tail that is not well conserved across species.

THP-1 monocytes (5 × 10⁶) were solubilized in a lysis buffer containing 50 mmol/l Tris-HCl, pH 7.5, 150 mmol/l NaCl, 1% Nonidet-P40, 0.5% sodium deoxycholate, and protease inhibitor cocktail (Boehringer Mannheim, Indianapolis, IN). Cellular lysates (50 μg) were fractionated on SDS-PAGE and analyzed by immunoblotting. Western blotting was performed using the indicated primary antibodies according to the instructions of each manufacturer. Blots were then stained with the corresponding anti–rabbit or anti–mouse IgG–horseradish-peroxidase conjugate (1:3,000 dilution) and detected by enhanced chemiluminescence reagent (ECL; Amersham, Arlington Heights, IL). Autoradiography was done at room temperature, and the appropriate exposures were quantitated by densitometry.

ELISA was used for quantitative measurement of released TNF-α, IL-1β, and MCP-1. After treatment of the cells, the media was collected and analyzed by ELISA according to each protocol provided by the manufacturer (R&D Systems).

Quantitative data were expressed as means ± SE, based on at least three separate experiments. Unpaired Student’s t test was used to determine the significance of differences among treatments. A P value <0.05 was considered significant.

AGEs have been shown to stimulate NF-κB DNA-binding activity in vitro in a number of cell types (24,26,30,33). To determine whether CML-HSA induces NF-κB transcriptional activity, we used the THP-1 monocytic cell line transiently transfected with an NF-κB reporter gene. Both RT-PCR and Western blot were used to demonstrate the presence of RAGE mRNA and protein, respectively, within THP-1 cells (data not shown). Transfected cells incubated with native HSA or without CML-HSA treatment were used as controls for basal luciferase activity. Treatment with CML-HSA at >100 μg/ml stimulated reporter gene expression by ∼1.7-fold as compared with the basal level (Fig. 1A). CML-HSA exposure initiated a fast and statistically significant NF-κB response, with an initial induction observed as early as 20 min after treatment. Luciferase activity was elevated approximately twofold within the first hour and remained essentially unchanged for 24 h (Fig. 1B). The lack of further luciferase accumulation after 1 h suggests an acute activation and deactivation of NF-κB, whereas the prolonged signal probably reflects the stability of the luciferase protein. The effect of 10, 5, and 1% CML modification of HSA was tested using the NF-κB reporter assay, and a similar 1.5- to 1.7-fold increase in the luciferase signal was observed at modifications of ≥5% (data not shown).

Nuclear extracts from THP cells treated with either TNF-α or CML-HSA showed a large increase in specific binding to the radiolabeled oligonucleotide containing the c-myc NF-κB binding site (Fig. 1C). Specificity was confirmed by demonstrating that inclusion of 10-fold molar excess of unlabeled competitor probe prevented the appearance of specific (double arrows), but not nonspecific, DNA binding (single arrow). Unmodified HSA treatment had no effect on DNA binding over that observed in untreated extracts.

To evaluate the role of RAGE in CML-HSA–mediated NF-κB transcriptional activation, we prepared synthetic peptides based on sequences derived from the RAGE extracellular domain that have high probability of solvent accessibility. A 17–amino acid peptide encompassing residues 102–118 of RAGE neutralized CML-HSA activity in a dose-dependent manner such that a 20-fold molar excess effectively eliminated the NF-κB response (Fig. 2A). Peptides corresponding to amino acids 39–56 and amino acids 67–82 (both predicted to have a high probability of solvent accessibility) had no detectable effect (data not shown). Mouse antiserum to sRAGE also blocked the CML-HSA–induced NF-κB response in a dose-dependent manner (Fig. 2A).

To more clearly define the role of RAGE in CML-HSA signaling to NF-κB, we generated two RAGE cytoplasmic domain truncation mutants: RAGE-365t lacks the whole intracellular domain, whereas RAGE-372t retains a small nonconserved portion of the cytoplasmic tail. Both mutants and wild-type RAGE were individually coexpressed with the NF-κB reporter. In contrast to full-length (wild-type) RAGE, the mutants blocked the NF-κB response to CML-HSA (Fig. 2B). Similar results were seen when changes in NF-κB were measured by DNA binding activity (Fig. 2C). THP-1 cells were transfected with an empty vector (pcDNA3.1), wild-type RAGE, or the RAGE-365t. Both control and wild-type RAGE–transfected cells showed no increase in basal NF-κB DNA binding but exhibited a large increase after incubation with CML-HSA. RAGE-365t decreased the response to CML-HSA. Together, these results demonstrate that the intracellular domain of RAGE is required for CML-HSA–induced signaling to NF-κB and that expression of a RAGE in which this domain has been removed acts to block endogenous RAGE signaling in a trans-dominant manner. RAGE specificity was demonstrated by the inability of the RAGE mutant, the peptide, or RAGE antiserum to block a fourfold increase in luciferase activity induced by TNF-α (data not shown).

We next examined the effect of selective inhibitors of potentially relevant signaling molecules impinging on the reporter response. Cells transfected with the NF-κB reporter were pretreated for 30 min with each compound before stimulation with CML-HSA. As shown in Fig. 3, the antioxidant NAC; the MEK1 (an upstream kinase for ERK1/2) inhibitor, PD98059; and the p38 inhibitor SB203580 completely blocked CML-HSA–induced NF-κB activation. Similarly, genistein, the protein tyrosine kinase (PTK) inhibitor, resulted in a partial but significant reduction (∼35%) in reporter activation. Together, these results suggest a role for reactive oxygen species, PTK, and MAPK in the signaling from RAGE to NF-κB.

Tyrosine phosphorylation is a crucial event that occurs rapidly and at a very early stage to propagate initial signals after activation of many receptors. Since genistein diminished NF-κβ activation by CML-HSA, we evaluated more directly the tyrosine phosphorylation profile after CML-HSA exposure. Stimulation of THP-1 cells with CML-HSA caused rapid phosphorylation (<10 min) on tyrosine residues of several proteins. Prominent phosphotyrosine polypeptides migrated at 75, 85, and 125 kDa (Fig. 4). Interestingly, the induced phosphotyrosine signals on these bands were abolished after 10 min and reappeared 60 min after treatment.

The results shown in Fig. 3 suggest that CML-HSA–mediated signaling to NF-κB involves the MAPK family. To determine which MAPK was involved, we examined MAPK activation by Western blotting using antibodies specific to the active forms of the kinases. As shown in Fig. 5A, stimulation of THP-1 monocytes with CML-HSA enhanced ERK1/2 MAPK activities approximately twofold over basal levels. CML-HSA–mediated ERK activity peaked at 20 min and returned to basal levels by 45 min posttreatment. Moreover, incubation of cells with PD98059 for 30 min before CML-HSA stimulation resulted in complete inhibition of ERK1/2 activation, whereas addition of the p38 inhibitor SB203580 had no detectable effect.

p38 Kinase, another member of the MAPK family, also responded to CML-HSA treatment, as its activity was stimulated 1.3- to 2.1-fold between 5 and 30 min, with maximal kinase activity at 20 min after CML-HSA exposure (Fig. 5B). Like ERKs, p38 activity returned to basal levels 45 min after treatment. SB203580, the selective inhibitor of p38, blocked >80% of the p38 activation. PD98059 had no detectable effect. No changes were observed in the levels of JNK activation after CML-HSA stimulation (Fig. 5C). Concentrations of PD98059 and SB203580 that effectively blocked the NF-κB reporter response did not affect JNK activity.

To establish a direct role for p38 MAPK in RAGE signaling leading to NF-κB transcriptional activation, THP-1 cells were transfected with the NF-κB reporter and an expression plasmid directing the synthesis of a dominant-negative form of p38. Overexpression of the kinase-dead p38 mutant completely blocked CML-HSA–mediated NF-κB activation, as did expression of an IκB mutant resistant to phosphorylation-dependent degradation (Fig. 6A). To further verify that the expressed p38 mutant competes efficiently with the endogenous wild-type kinase, we also measured the p38 kinase activity from empty vector-transfected and p38 mutant–transfected cells. In vitro kinase assays were performed using the transcription factor ATF-2, a downstream target of p38, as a substrate. Although total p38 protein levels were elevated by expressing the mutant (Fig. 6B, left panel), the p38 mutant–containing lysate had much less kinase activity compared with the empty vector–transfected control (Fig. 6B, right panel).

Since it is well established that NF-κB mediates proinflammatory responses by regulating the expression of a number of cytokines (38,39), and since CML-HSA induced p38-dependent NF-κB activation in THP-1 monocytes, we further investigated the role of p38 MAPK and NF-κB in CML-HSA–induced cytokine production. Mock-, p38-, or IκB mutant–transfected THP-1 cells were exposed to CML-HSA for 24 h, and secreted cytokines were measured. Table 1 shows that CML-HSA treatment induced a robust secretion of TNF-α and MCP-1 (∼12- and 6-fold, respectively). IL-1β production was less dramatic but still increased ∼3.5-fold by 24 h. Expression of the p38 dominant-negative mutant or the IκB mutant significantly reduced the CML-HSA–induced secretion of these cytokines, indicating that p38 MAPK and NF-κB are both required in the signaling pathway from RAGE to cytokine production. To strengthen and extend this observation, similar experiments were performed in peripheral blood-derived monocytes. As shown in Table 2, CML-HSA induced an approximately fourfold increase in TNF-α in primary monocytes, a response that was blocked by SB203580 and PD98059.

The results of our experiments identify a signaling cascade in THP-1 monocytes in which CML-HSA binds RAGE, leading to phosphorylation of p38 MAPK and ERK1/2, causing increased DNA binding and transcriptional activation of NF-κB and secretion of proinflammatory cytokines. Although many AGEs have been identified, including pentosidine, methylglyoxal, pyralline, and imidizalone (1,3,4,9), to date, only CML has been reported to be an adduct on proteins that act as a ligand for RAGE (33). Our findings are consistent with previous observations that 1) RAGE plays a role in chronic cellular activation and tissue injury by serving as a cell-surface receptor for the S100/calgranulin family of polypeptides that are released from activated inflammatory cells and accumulate at sites of chronic inflammation (6); 2) AGE-induced increased endothelial cell tissue factor expression, a membrane-bound glycoprotein that initiates coagulation by its ability to bind factor VII/VIIa, is blocked by RAGE antisense (27); 3) AGE-RAGE interaction induces cell-associated VCAM-1 expression (26); 4) high–molecular weight hyaluronic acid blocks AGE-induced NF-κB activation and proinflammatory cytokine production in J774 mouse macrophages (40); and 5) AGE adducts increase NF-κB DNA binding activity (24,30). Thus, although many studies have shown RAGE-induced NF-κB DNA binding activity, here we further show that this DNA binding leads to increased transcription of NF-κB target genes. These findings, together with the observations that NF-κB is a pleiotropic regulator of many “response-to-injury” genes and that the expression of this cell-surface receptor is upregulated in a number of diverse pathologies, suggest a central role for RAGE in the expression of proinflammatory and prothrombotic genes.

Our results demonstrate that the cytoplasmic domain of RAGE is required for CML adducts to activate NF-κB target gene transcription. These results are consistent with previous studies demonstrating that transient transfection of a cDNA-encoding RAGE lacking the cytoplasmic domain imparted a dominant-negative effect when the ligand was CML-ovalbumin (33) or S100B (a member of the S100/calgranulin family) (6). A dominant-negative effect of the RAGE cytoplasmic tail also was reported on neurite outgrowth and NF-κB activation in N18 neuroblastoma and C6 glioma cells in response to AGE-modified bovine serum albumin and amphoterin (41).

Although the conserved cytoplasmic domain of RAGE is required for CML-HSA to activate NF-κB, only some of the signaling intermediates in the path from RAGE to the nucleus have been demonstrated. Thus, oxidant stress (24), PTK (42), p21Ras, and ERK1/2 MAPK (30) are involved in the AGE-mediated signaling cascade. In addition, it has been suggested that more than one parallel pathway can be activated simultaneously by the cytoplasmic domain of RAGE (41). In that study, it was demonstrated that Rac and Cdc42, the Rho family small GTPases, were responsible for RAGE-mediated neurite outgrowth induced by amphoterin, but not the simultaneous activation of NF-κB–dependent transcription, which was dependent on Ras-MAPK pathway. It is noteworthy that CML-HSA ligation of RAGE induced acute tyrosine phosphorylation followed by either dephosphorylation or degradation. For many transmembrane immunoglobulin-class receptors, signaling is initiated by ligand-induced aggregation that activates receptor-coupled tyrosine kinases (43,44). In most studies of mammalian receptors that signal by coupling to tyrosine kinase activation or that contain intrinsic tyrosine kinase activity, the earliest obligate intracellular event after ligand engagement is tyrosine phosphorylation (45,46). Although our results indicate that tyrosine phosphorylation plays an important role in RAGE signaling, no consensus kinase motif has been identified in the RAGE intracellular domain. Consequently, RAGE probably couples to a tyrosine kinase(s), directly or indirectly, to mediate tyrosine phosphorylation. Currently, intracellular proteins capable of binding to the RAGE cytoplasmic domain remain unknown.

A wide variety of extracellular signals regulate phosphorylation-dependent activation of MAPK family members that subsequently modulate transcriptional responses altering cellular function in normal and pathological settings. Although there is considerable cross-talk within these signaling cascades, some generalities have emerged regarding the specific types of information that are processed through the three MAPK family members. Thus, ERK typically mediates cell proliferation in response to growth factors, whereas p38 and JNK respond to cytokines and cellular stress inducers. Previously, RAGE ligands have been reported to activate ERK in arterial smooth muscle cells (30,47), rat pheochromocytoma PC12 cells (30), and renal tubulus cells in culture (48), indicating that this pathway is common across a variety of cell types. Our results implicate both ERK and p38 as key intermediates in CML-HSA activation of RAGE signaling leading to NF-κB–mediated target gene expression in a human monocytic cell line. The specificity of these responses is indicated by the lack of activation of JNK, the third MAPK family member. Additional specificity is evident since p38δ is not inhibited by SB203580, suggesting that other p38 isoforms must mediate the response to CML-HSA (49). These observations on p38 appear not to be unique to the AGE ligand used, CML-HSA, since amphoterin-induced RAGE signaling in C6 glioma cells activated p38 as well as p44/p42 (ERK) and stress-activated protein kinase/JNK MAPKs (50). Indeed, the extent of induction for ERK1/2 and p38 was comparable in both studies. Our inability to detect JNK activation in response to CML-HSA may be due to differences in cell types (THP-1 monocytes versus C6 glioma), exposure times (10–60 vs. 90 min), and ligand used (CML-HSA versus amphoterin), which may have different affinities for RAGE. Other recent reports have shown closely related results; CML-HSA increased NF-κB DNA binding through RAGE (33), and glycated serum albumin stimulated ERK activity and NF-κB DNA binding (30), but unfortunately, p38 was not evaluated in these studies.

Mechanisms that couple activation of MAPK to NF-κB target gene induction are complex. It is particularly important to note that NF-κB DNA binding activity is necessary, but not sufficient, for transcription, as shown in several reports and reviews (51). For example, IκB degradation and NF-κB DNA binding were insufficient for TNF- and IL-1β–induced NF-κB–dependent transcription (52). In addition, inhibition of TNF-induced p38 activation blocked NF-κB–dependent gene expression without decreasing DNA binding activity (53), leading to speculation that these intermediates converge at a more distal point. A subsequent report further delineated a mechanism wherein p38 phosphorylated the TATA-binding protein TFIID, increasing its binding to the p65 subunit of NF-κB (54), stabilizing the transcription complex and mediating gene expression. This mechanism was further supported by observations that inhibitors of p38 and ERKs cooperatively regulated the transactivation potential of the p65 subunit (55). Thus, increased DNA binding activity shown here and by others is only an intermediate readout of RAGE signaling and illustrates the importance of analyzing target gene expression, as we report here, for secretion of the cytokines IL-1β, TNF-α, and MCP-1. Finally, our observations that both p38 and ERK mediate RAGE-dependent NF-κB target gene induction suggest that RAGE may be a pleiotropic integrator of AGEs that signal both growth factor–like proliferative responses as well as cytokine-like proinflammatory responses.

The present studies demonstrate that p38 activation is necessary for CML-HSA–mediated RAGE-dependent NF-κB activation, induction of target gene expression, and secretion of proinflammatory cytokines from monocytes. The identification of a RAGE signaling axis involving p38 extends our understanding of AGE-induced cellular dysfunction and provides a possible mechanism for p38 involvement in inflammation, arthritis, septic shock, and myocardial dysfunction after ischemia (56,57,58,59). Furthermore, since AGEs—particularly CML adducts—are prevalent in diseases such as diabetes, renal failure, chronic inflammation, and Alzheimer’s disease (6,32,33,60), our findings may provide another pathophysiologically relevant mechanism linked to these disorders. Understanding the integrative aspects of RAGE in the diverse pathologies of AGEs provides a key challenge for future studies.

We thank Dr. Kaijun Ren (Chemistry Department at Texas Biotechnology Corporation) for the synthesis of peptides used in the RAGE inhibition studies, and Drs. Roger Davis (Howard Hughes Medical Institute at the University of Massachusetts Medical Center, Worcester, MA) and Chung Y. Hsu (Washington University School of Medicine, St. Louis, MO) for the p38 kinase–dead and IκB dominant-negative mutants, respectively.