Blockade of receptor for advanced glycation end product (RAGE) attenuates ischemia and reperfusion injury to the liver in mice



Hepatic ischemia/reperfusion (I/R) injury associated with liver transplantation and hepatic resection is characterized by hepatocellular damage and a deleterious inflammatory response. In this study, we examined whether receptor for advanced glycation end product (RAGE) activation is linked to mechanisms accentuating inflammation on I/R in a murine model of total hepatic ischemia. Animals treated with soluble RAGE (sRAGE), the extracellular ligand-binding domain of RAGE, displayed increased survival after total hepatic I/R compared with vehicle treatment. TUNEL assay and histologic analysis revealed that blockade of RAGE was highly protective against hepatocellular death and necrosis on I/R; in parallel, proliferating cell nuclear antigen was enhanced in livers of mice treated with sRAGE. Rapid activation of p38, p44/42, stress-activated protein kinase and c-Jun N-terminal kinase mitogen-activated protein kinases, signal transducer and activator of transcription-3, and nuclear translocation of activator protein-1 was evident at early times on I/R. In the remnants of sRAGE-treated livers, however, activation of each of these signaling and transcription factor pathways was strikingly decreased. sRAGE-treated remnants displayed enhanced activation of nuclear factor κB, in parallel with increased transcripts for the proregenerative cytokine, tumor necrosis factor-α. In conclusion, these data suggest that RAGE modulates hepatic I/R injury, at least in part by activation of key signaling pathways linked to proinflammatory and cell death-promoting responses. We propose that blockade of this pathway may represent a novel strategy to attenuate injury in hepatic I/R and to facilitate regeneration. (HEPATOLOGY 2004;39:422–432.)

Ischemia/reperfusion (I/R) injury to the liver is a common sequela of transplantation or significant hepatic resection. Key roles for endogenous inflammatory cells, such as Kupffer cells, in initiating and sustaining oxidant stress and subsequent potent inflammatory responses have been revealed by extensive investigation in multiple model systems of both cold and warm ischemic injury.1–5 Generation of reactive oxygen species, at least in part by activated liver macrophages, stimulates a second wave of injury, triggered by influx of neutrophils and other inflammatory cells, especially CD4+ T cells and macrophages, into the remnant. Subsequently, release of interleukin, tumor necrosis factor-α (TNF-α), and platelet-activating factor and enhanced expression of cell adhesion molecules, such as intracellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), and chemokines provokes inflammation-mediated liver injury.6–9 Extensive damage to the hepatic vasculature in these settings by accumulation and entrapment of neutrophils, platelets, and mononuclear phagocytes may augment endothelial injury and a “no-reflow” phenomenon, thereby sustaining hepatocyte injury.10, 11 The precise method of death of hepatocytes, endothelial cells, and inflammatory cells in the injured liver remains not fully clarified, but seems to include components of both apoptotic and necrotic pathways.12, 13 Regardless of the method of death, it is established that I/R injury in the liver may overwhelm innate regenerative programs.

The inflammatory response is tightly regulated by an elaborate cascade of proinflammatory mediators. Multiple signaling cascades, including signal transducer and activator of transcription-3, p44/42, p38 mitogen-activated protein kinases (MAPKs) and stress-activated protein kinase (SAPK) or c-Jun N-terminal kinase (JNK) have been implicated in inflammatory and cell death pathways in the perturbed liver activated on I/R.14, 15 Furthermore, the transcription factor nuclear factor κB (NF-κB) likely plays multiple roles in the injured liver, serving as a key regulator of both the hepatic inflammatory as well as regenerative and antiapoptotic response.16–19

The receptor for advanced glycation end products (RAGE) is linked to amplification of the inflammatory response.20 RAGE, a member of the immunoglobulin superfamily, interacts with ligands enriched in inflamed milieu, including (carboxymethyl)lysine-modified adducts, and S100/calgranulins, the latter members of a family of proinflammatory cytokines.21, 22 Engagement of RAGE by proinflammatory ligands sustains inflammatory responses and leads to tissue injury. Blockade of RAGE, employing soluble RAGE (sRAGE), the extracellular ligand-binding domain of RAGE, suppressed delayed-type hypersensitivity and colitis in murine models,22 as well as the induction of collagen-induced arthritis in mice sensitized to and challenged with bovine type II collagen.23 Activation of RAGE triggers a range of signaling cascades linked prominently to I/R injury and regeneration in the liver.20

These considerations led us to hypothesize that activation of RAGE contributes to activation of proinflammatory and tissue-destructive processes on I/R, and that blockade of the receptor may limit immediate deleterious inflammatory mechanisms, and thereby facilitate regenerative potential in the injured liver. In this study, we used a murine model of total hepatic I/R. We show that blockade of RAGE, using sRAGE, improved survival after total hepatic I/R. In parallel, inflammation and cell death were suppressed in the remnant after reperfusion. We propose that blockade of RAGE may be a novel target for restoration of hepatic homeostasis after I/R.


RAGE, receptor for advanced glycation end products; sRAGE, soluble RAGE; I/R, ischemia and reperfusion; NF-κB, nuclear factor κB; MAPK, mitogen-activated protein kinase; SAPK/JNK, stress-activated protein kinase/c-Jun N-terminal kinase; EMSA, electrophoretic mobility shift assay; TNF-α, tumor necrosis factor-α; MPO, myeloperoxidase; TUNEL, transferase-mediated dUTP nick end labeling; ALT, alanine aminotransferase; PCNA, proliferating cell nuclear antigen; IL-6, interleukin-6.

Materials and Methods

Induction of Total Ischemia/Reperfusion.

Male C57BL/6J mice (Charles River, Boston, MA) at 8 to 9 weeks of age were used in all experiments and were maintained in a temperature-controlled room with alternating 12-hour light–dark cycles. All experiments were approved by the Institutional Animal Care and Use Committee of Columbia University. Briefly, mice were anesthetized and a midline incision was performed; all ligaments surrounding the liver were freed. A vascular microclamp was used to interrupt the blood supply to the median and lateral left lobes. After 75 minutes of hepatic ischemia, the clamp was removed, initiating hepatic reperfusion. The nonischemic caudate and right lobes were resected, leaving only ischemic tissue in place. The amount of liver tissue removed did not exceed 30% of the total hepatic mass. No bleeding was observed after resection. Sham control mice underwent the same protocol without ischemia and resection. Mean murine sRAGE was prepared and purified from a baculovirus expression system; any detectable lipopolysaccharide was removed by Detoxi-gel columns (Pierce Chemical Co., Rockford, IL).22 sRAGE was administered at a dosage of 100 μg daily (based on dose-response experiments performed previously).22 Control mice received equal volumes of vehicle (phosphate-buffered saline [PBS]). Treatment was begun 12 hours before surgery and continued until sacrifice, death, or 7 days after I/R. Mice were killed after 1, 6, or 18 hours of reperfusion, and blood and liver tissues were retrieved for analyses.

Terminal Deoxynucleotidyl Transferase-Mediated dUTP Nick End Labeling (TUNEL) Assay.

Formalin-fixed sections of liver tissues (5 μm) were deparaffinized in xylene and were hydrated in graded ethanol. The TUNEL-positive cells were identified using the ApopTaq Red In Situ Apoptosis Kit (Intergen Company, Purchase, NY). For each liver section, 5 high-powered fields (×400) were examined and the numbers of positive cells were counted.

Blood Analysis and Myeloperoxidase (MPO) Assay.

Blood was obtained by cardiac puncture at the time of sacrifice for analysis of serum alanine aminotransferase (ALT) and plasma prothrombin time by Ani Lytics Incorporate (Gaithersburg, MD). Serum samples were also analyzed for TNF-α using ELISA Kits from R&D systems (Minneapolis, MN). Liver MPO activity was measured as an index of neutrophil accumulation as described before.24 MPO activity was measured spectrophotometrically and data were expressed as change in absorbance at 460 nm per minute.

Immunohistochemistry and Histologic Analysis.

Consecutive sections (5 μm) from paraffin-embedded liver were prepared for hematoxylin-eosin staining and proliferating cell nuclear antigen (PCNA) immunohistochemistry. PCNA immunohistochemistry was performed with DAKO animal research Kit (DAKO Corporation, Carpinteria, CA). PCNA-positive staining nuclei were counted in five high-power fields, and the results were expressed as a percentage of total number of hepatocyte nuclei. For activated NF-κB immunohistochemistry, formalin-fixed sections were used. After heat antigen retrieval, nonspecific binding was blocked and sections were incubated with a rabbit polyclonal antibody against p65 NF-κB subunits (Santa Cruz Biotechnology Inc., Santa Cruz, CA) at 1:50 dilution. A biotinylated goat anti-rabbit immunoglobulin G (IgG; Vectastain ABC kit; Vector Laboratories, Burlingame, CA) was used as a secondary antibody at 1:200 dilution and was visualized with enhanced horseradish peroxidase conjugated streptavidin and a substrate chromogen. For RAGE staining or for localizing the RAGE antigen to distinct cell types, single and double immunofluorescence staining was performed in frozen sections using the following antibodies: rabbit anti-mouse RAGE IgG,22 rat anti-mouse CD68 (Serotec, Raleigh, NC), hamster anti-mouse CD11c and rat anti-mouse CD31 (Pharmingen, Franklin Lakes, NJ), and mouse anti-human alpha smooth muscle actin IgG (α-SMA; Sigma). Briefly, after fixation and blocking with normal serum, the sections were incubated with the primary anti-RAGE antibody at a 1:50 dilution, were washed with PBS, and were visualized using a goat anti-rabbit biotinylated secondary antibody and streptavidin conjugated to Alexa Fluor 555 Green (Molecular Probes, Inc., Eugene, OR). In other double-staining procedures, the slides were incubated with another primary antibody (anti-CD11c or the indicated antibodies) and were visualized using a biotinylated secondary antibody and streptavidin conjugated to Alexa Fluor 555 Red. Slides were mounted in medium containing 4′,6 diamidino-2-phenylindole (DAPI) and were photographed using fluorescence microscopy.

Preparation of Liver Extracts and Western Blotting.

Liver tissues were used to obtain whole protein and nuclear protein and were homogenized and lysed in ice-cold lysis buffer (10 mM Tris HCl, pH 7.5, 150 mM NaCl, 2 mM ethylenediaminetetraacetic acid, 1% Triton X-100, 10% Glycerol, 2 mM sodium orthovanadate, and protease inhibitors [Complete tabs, Boehringer Mannheim]. The lysates were cleared of debris by centrifugation at 14,000g for 15 minutes at 4°C. Nuclear proteins were prepared from the same tissues using the isolation kit from Pierce Chemical Co.

Equal amounts of protein were boiled in sample buffer and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nylon membranes. Specific antibodies against phosphorylated forms of p44/p42 MAPK, p38 MAPK, SAPK/JNK, and signal transducer and activator of transcription-3 were used (Cell Signaling Technology, Lexington, KY). Proteins were visualized with an enhancer chemi-luminescence (ECL) detection kit (Amersham Biosciences, Piscataway, NJ). To verify equal loading of proteins in each lane, the blots were stripped and reprobed using the relevant antibodies to total kinase (Cell Signaling Technology, Lexington, KY). Signal intensity was quantified by a densitometer using Image Quant (Molecular Dynamics, Foster City, CA).

Electrophoretic Mobility Shift Assays (EMSA).

Nuclear extracts were prepared and used for EMSA. NF-κB or AP-1 DNA binding activity was measured using end-labeled double-stranded NF-κB or AP-1 oligonucleotide, respectively (Promega, Madison, WI). Binding reactions containing equal amounts of nuclear extracts (10 μg) and labeled probe were incubated at room temperature for 30 minutes. The DNA-protein complexes were resolved by 6% polyacrylamide gel in 0.5 X Tris-borate-ethylenediaminetetraacetic acid, gels were dried and analyzed by autoradiography. For competition experiments, an excess (×10) of specific unlabeled double-stranded probe was added to the binding mixture. NF-κB Supershift analysis was performed by incubating nuclear extracts with 0.5 μg of antibodies against p50 or p65 subunit of NF-κB (Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C for 1 hour before adding the [γ-32p] ATP-labeled oligonucleotide. DNA binding was quantified by the intensity of shifted bands using Image Quant software (Molecular Dynamics, Foster City, CA), and the changes were expressed by fold increase over sham-operated controls.

Quantitation of Gene Expression Using Real Time Polymerase Chain Reaction (PCR).

Total RNA was isolated using the RNeasy Mini kit (Qiagen, Valencia, CA). For cDNA synthesis, 0.5 μg of total RNA was transcribed with TaqMan Reverse Transcription Reagents Kit using random hexamers (PE Applied Biosystems, Foster City, CA). Primers and TaqMan probes were designed by Primer Express 2.0 software (PE Applied Biosystems, Foster City, CA), except for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which was commercially available. The primers and probes were as follows: mouse RAGE forward primer (5′-GGACCCTTAGCTGGCACTTAGA-3′), reverse primer (5′-GAGTCCCGTCTCAGGGTGT-CT-3′), and probe (6FAM-ATTCCCGATGGCAAAGAAACACTCGTG-TAMRA). Mouse TNF-α forward primer (5′-AATGGCCTCCCTCTCATCAGT-3′), reverse primer (5′-GCTAC-AGGCTTGTCACTCGAATT-3′), and probe (6FAM-ATGGCCCAGACCCTCACACTCA-GATC-TAMRA). Mouse interleukin 6 (IL-6) forward primer (5′-TATGAAGTTCCTCTCTGCAAGAGA-3′), reverse primer (5′-TAGGGAAGGCCGTGGTT-3′), and probe (6FAM-CCAGCATCAGTCC-CAAGAAGGCAACT-TAMRA). Quantitative real-time PCR was performed using the ABI Prism 7900 Sequence Detection System (PE Applied Biosystems). cDNA samples (50 ng each, except for GAPDH [10 ng]) were mixed with primer, probe, and TaqMan PCR Master Mix in a final volume of 50 μl as described by the manufacturer (bulletin no. 2). PCR condition was 50°C for 2 minutes, 95°C for 10 minutes, and 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. The relative amounts of TNF-α and IL-6 transcript were determined using the comparative Ct method as described by the manufacturer. Murine RAGE expression was determined using the standard curve method and mouse lung cDNA as a standard. The mRNA levels are expressed as the fold increases over sham-operated controls.

Statistical Analysis.

All data are expressed as means ± standard deviation. Data was analyzed by an analysis of variance. For survival studies, the Kaplan-Meier method was used to analyze survival. The log-rank test was used to compare survival curves. Group means were compared by analysis of variance rejecting the null hypothesis when P < 0.05. Differences were considered significant when P < 0.05. Post hoc testing confirmed differences between pairs of groups.


Blockade of RAGE Improves Survival on I/R Injury.

To test if engagement of RAGE mediated hepatic injury accompanying I/R, we used a murine model of total hepatic ischemia. Murine sRAGE, the extracellular ligand-binding domain of RAGE, was administered the day before surgery followed by once daily treatment by intraperitoneal route until sacrifice or death. Control mice received equal volumes of vehicle, PBS. Kaplan Meier product-limit estimate studies revealed that mice undergoing total I/R in the presence of PBS displayed a ≈30% survival by 7 days. In contrast, mice treated with sRAGE displayed significantly increased survival, ≈84%, by 7 days (P = 0.006; Fig. 1A).

Figure 1.

Blockade of RAGE improves survival in a murine model of total hepatic I/R. (A) Kaplan-Meier product limit estimate. Male C57BL/6 mice were subjected to total hepatic ischemia/reperfusion injury in the presence of murine sRAGE (100 μg/day) or vehicle, PBS. The times of death were recorded for all mice and survival plotted. Statistical considerations are indicated in the figure. (B) Expression of RAGE is upregulated in I/R injury. Pooled RNA from each animal was reverse transcribed into cDNA, and mRNA transcripts for murine RAGE were measured by real-time PCR as described. The data presented are the mean of five independent experiments ± standard deviation. *P < 0.05 versus sham control. (C–G) Immunofluorescence staining for RAGE antigen expression. At 6 hours after I/R, frozen tissue from the hepatic remnant was retrieved and immunofluorescence single or double staining was performed using anti-RAGE IgG (C, D); anti-CD68 IgG alone or + anti-RAGE IgG (E); anti-CD11c IgG alone or + anti-RAGE IgG (F); or anti-CD31 IgG + anti-RAGE IgG (G). Scale bar, 50 μm.

These findings suggested that RAGE contributed to injury in hepatic I/R and led us to dissect the contribution of RAGE to these processes. First, to determine if severe hepatic I/R upregulated RAGE expression, we measured the levels of RAGE transcripts by real-time PCR at different time points. As shown in Fig. 1B, at 6 hours after I/R, increased transcripts for RAGE were evident in PBS-treated remnants versus sham. Administration of sRAGE did not attenuate RAGE expression. By 18 hours after I/R, transcripts for RAGE in PBS- and sRAGE-treated livers returned to baseline. These data suggested that the impact of RAGE was likely to occur within the first hours after I/R.

To determine the cellular localization of RAGE in the remnant after I/R, immunofluorescence staining was performed in frozen tissues retrieved from mice undergoing total I/R versus sham. Expression of RAGE antigen was noted predominantly in a perivascular distribution around the central vein and was enhanced after I/R versus sham at 6 hours (Fig. 1D, C, respectively). Hepatocytes displayed very weak, if any, staining for RAGE antigen. Previous studies have reported that endocytotic elimination of advanced glycation end products in rat liver is mediated largely by scavenger receptors expressed by Kupffer cells and endothelial cells (EC).25 To test whether Kupffer cells or EC in the liver expressed RAGE, double-immunofluorescence staining for RAGE and CD-68 (for Kupffer cells), CD31 (for EC), and α-SMA (for smooth muscle cells) was used. Focal and patchy colocalization of RAGE was evident in CD68-positive Kupffer cells (Fig. 1E), but there was no evidence of colocalization of RAGE with EC (Fig. 1G) or smooth muscle cells (data not shown) in livers subjected to I/R. In addition, CD11c-expressing dendritic cells have been identified in the mouse liver.26, 27 Although there was no colocalization of CD68 and CD11c in liver tissues (data not shown), double staining with antibodies to RAGE, and CD11c revealed striking colocalization around central veins in liver remnants subsequent to I/R (Fig. 1F). These findings indicated that RAGE was expressed in dendritic cells and not Kupffer cells after I/R. Although bile duct epithelial cells displayed occasional staining for RAGE antigen by indirect immunoperoxidase staining, there were no apparent differences between controls and livers subjected to I/R (data not shown).

Effects of RAGE Blockade on I/R-Induced Hepatic Injury.

The impact of sRAGE on hepatic injury on I/R was evaluated by multiple means. A histologic scoring system was developed based on the extent of necrosis, vacuolization, hypereosinophilia, and loss of cell borders on hematoxylin-eosin staining. At 18 hours after I/R, mice treated with PBS displayed extensive injury with intense destruction of the parenchyma. In contrast, in remnants of mice treated with sRAGE, markedly decreased necrosis was evident (Fig. 2A). At early time points (1 hour after I/R), remnants were analyzed by TUNEL staining. Increased numbers of TUNEL-positive nuclei were detected in injured remnants versus sham controls. In the presence of sRAGE, mice subjected to I/R displayed significantly decreased numbers of TUNEL-positive nuclei compared with PBS-treated controls (Fig. 2B). Based on the morphologic features, the principal TUNEL-positive cells at this time point in the liver were hepatocytes (Fig. 2B). These findings suggested that in parallel with increased survival, sRAGE-treated mice displayed decreased evidence of hepatocyte cell death after I/R. In parallel with decreased cell death, sRAGE-treated mice displayed significantly increased numbers of PCNA-positive nuclei compared with vehicle (Fig. 2C). This suggested that blockade of RAGE may promote cell cycle entry and, thereby, regeneration.

Figure 2.

Blockade of RAGE suppresses hepatocyte cell death and promotes regeneration. Male C57BL/6 mice were subjected to total I/R injury in the presence of the sRAGE or vehicle, PBS. (A) Histologic analysis. At 18 hours after I/R, sections were retrieved, representative sections were stained with hematoxylin-eosin, and a scoring system was developed to quantify the degree of injury, based on necrosis, vacuolization, hypereosinophilia, and loss of cell borders. Scale bar, 100 μm. (B, C) TUNEL and PCNA assay. At 1 hour (B) or 6 hours (C) after I/R, sections were retrieved from PBS-treated and sRAGE-treated livers. The number of TUNEL or PCNA-positive nuclei were quantified from five high-powered fields. Scale bar, 50 μm. (D) Myeloperoxidase activity. At 6 hours after I/R, liver remnants were retrieved and were subjected to analysis for MPO activity. *P < 0.05 versus sham controls. No significant differences were observed among PBS- and sRAGE-treated mice. (E) Parameters of hepatic injury. At 6 hours after I/R, plasma or serum was collected and was analyzed for prothrombin time and glucose. *P < 0.05 versus sham controls; **P < 0.05 versus PBS treatment.

These findings suggested that histologic evidence of diminished cell death and enhanced regeneration-accompanied blockade of RAGE in I/R. To study the impact of blockade of RAGE on distinct biochemical pathways linked to liver injury, we assessed MPO activity, an index of neutrophil recruitment and activation. We found that MPO levels in the liver tissue increased greatly in PBS-treated mice compared with sham. The remnants of sRAGE-treated mice showed lower levels of MPO than that of PBS-treated mice, but this change was not statistically different (Fig. 2D). This suggested that blockade of RAGE may exert a partial effect on neutrophil activation, but, more likely, that important effects of RAGE blockade were downstream of initial neutrophil attraction and activation in the injured liver after I/R.

We next studied the impact of RAGE blockade on release of enzymes such as ALT. Interestingly, although ALT levels were enhanced in PBS-treated I/R livers versus sham, levels in sRAGE-treated mice were not reduced (data not shown). These findings suggest strongly that measurement of liver enzymes such as ALT does not parallel fully the extent of injury. In this context, indices of hepatic function, such as prothrombin time and glucose levels, were strikingly preserved in sRAGE-treated mice (Fig. 2E).

RAGE Blockade Modulates Activation of Proinflammatory Signaling Pathways in Total I/R.

Activation of MAP kinase (p44/42 MAPK, p38 MAPK and SAPK/JNK) pathways has been suggested to modulate inflammation and cell death in the severely stressed liver. To assess the mechanisms underlying hepatic injury on I/R and their potential modulation by blockade of RAGE, we examined each of these signaling pathways. Mice subjected to I/R injury displayed rapid activation of p44/42, p38, and SAPK/JNK MAPKs at 1 hour after injury. In the presence of sRAGE, significantly decreased levels of phosphorylated p44/42, p38, and SAPK/JN MAPK were noted in the remnants versus that observed in PBS-treated animals (Fig. 3A–C). The observed changes in these phosphorylated MAPKs were not the result of a general decrease in expression of total kinases, because levels of the total kinases did not differ among groups (Fig. 3A–C). In each case, similar results were obtained at 6 hours after I/R (data not shown). Furthermore, we tested the potential modulation of STAT3 signaling during total hepatic I/R. STAT3 belongs to the family of STAT transcription factors that mediate the cellular response to a variety of cytokines and growth factors.28 Phosphorylated STAT3 expression was higher in PBS-treated mice than in sham control mice within 1 hour of reperfusion. In the presence of sRAGE, remnants displayed significantly reduced phosphorylation of STAT3 (Fig. 3D). These data suggested that blockade of RAGE suppressed rapid activation of these proinflammatory and stress pathways accompanying I/R.

Figure 3.

Blockade of RAGE modulates activation of inflammatory and proapoptotic signalling pathways. Equal amounts of liver extracts (40 μg) were subjected to SDS-PAGE and immunoblotting using antiphospho p44/p42 MAPK IgG (A); antiphospho p38 MAPK IgG (B); antiphospho p54-SAPK/JNK kinase IgG (C); or antiphospho STAT3 IgG (D). After blotting, the membranes were stripped and reprobed with respective antibodies to total p44/p42, p38, p54 SAPK/JNK MAPK, or STAT3. Bands were scanned into a densitometer, and relative pixel units (phospho/total) were reported.

Blockade of RAGE Enhances Activation of NF-κB and Suppresses Activation of AP-1 After Total Hepatic I/R.

NF-κB is a pleiotropic transcription factor whose activation has been linked to inflammatory and destructive processes, as well as initiation of regenerative programs in the injured liver. We assessed the impact of total hepatic I/R on activation of NF-κB in the remnants at 1 and 6 hours after injury and tested the effects of RAGE blockade. Nuclear extracts were prepared from the remnants at these times, and activation of NF-κB was analyzed by EMSA using a radiolabeled NF-κB consensus motif. As shown in Fig. 4A, at 1 hour after total I/R, we detected a small but not significant increase in nuclear translocation of NF-κB in PBS-treated remnants versus sham controls, whereas sRAGE-treated remnants displayed a striking increase. Similar results were seen at 6 hours after total I/R, but the differences were less dramatic than those observed at 1 hour after I/R (Fig. 4A). These bands were specific as confirmed by complete loss of the shifted band in the presence of excess unlabeled NF-κB consensus motif. The NF-κB DNA complexes were characterized further by supershift analysis. These studies revealed that the NF-κB complex was a heterodimer, composed of both p65 and p50 (Fig. 4B).

Figure 4.

Blockade of RAGE modulates activation of NF-κB and AP-1 after total hepatic I/R. (A, B, G) EMSA was performed on nuclear extracts prepared from the livers of mice subjected to I/R in the presence of PBS or sRAGE, or sham controls. To confirm the specificity of the NF-κB or AP-1, experiments were performed with a 10-fold excess of unlabeled oligonucleotide probe as a competitor (comp). Supershift assays were performed as indicated using anti-p50 IgG, anti-p65 IgG, or both. The results are representative of at least three separate experiments. (C–F) Formalin-fixed sections were stained with antibody to the activated p65 submit of the NF-κB complex. In comparison with PBS-treated groups, 1 hour (C&D) or 6 hours (E, F), increased nuclear NF-κB was observed than in the sRAGE-treated groups. Scale bar, 50 μm.

Consistent with EMSA findings, immunohistochemistry using an antibody to activated p65 revealed similar results. Immunohistochemistry of PBS-treated I/R remnants showed sparse reactivity with the antibody to activated p65 compared with sRAGE-treated remnants at 1 hour (Fig. 4C, D) or 6 hours (Fig. 4E, F) after I/R.

In addition to NF-κB, we examined activation of AP-1 in I/R to assess the impact of RAGE blockade. At 1 hour after I/R, a significant increase in activation of AP-1 was observed in PBS-treated livers versus sham controls. In contrast, in the presence of sRAGE, activation of AP-1 in I/R at 1 hour was significantly suppressed (Fig. 4G).

Blockade of RAGE Modulates Expression of Proinflammatory Mediators in Total Hepatic I/R.

Last, we examined the expression of cytokines after total I/R by real-time PCR. Cytokines such as TNF-α and IL-6 have been assigned multiple roles in liver injury, linked to both potent inflammatory responses and regeneration. As shown in Fig. 5A, at 1 hour, transcripts for TNF-α were significantly increased in the PBS versus sham livers. Treatment with sRAGE at this time point greatly decreased the level of TNF-α (Fig. 5A). In parallel, by 8 hours after I/R, serum TNF-α were significantly elevated in PBS- versus sRAGE-treated mice subjected to liver I/R (Fig. 5B). Interestingly, by 18 hours after I/R, levels of TNF-α in the PBS-treated mice returned to baseline; however, in the presence of sRAGE, TNF-α transcripts were elevated strikingly compared with sham or PBS-treated remnants after I/R (Fig. 5A). Examination of the overall trends indicated that in PBS-treated mice, the peak expression of TNF-α transcripts was noted at 1 hour; by 18 hours, levels of TNF-α in the remnant had returned to baseline. The opposite pattern was observed in sRAGE-treated mice. A steady increase was observed, with a peak effect noted at 18 hours after I/R.

Figure 5.

Blockade of RAGE modulates expression of TNF-α and IL-6 in the liver remnant subsequent to I/R. Pooled RNA from each animal was reverse transcribed into cDNA and mRNA transcript levels for TNF-α (A) and IL-6 (C) were determined by real-time PCR. Comparative Ct method was used to determine levels of transcripts. Serum TNF-α level was measured by ELISA (B). *P < 0.05 versus sham controls; **P < 0.05 versus PBS treatment.

In addition, we examined transcripts for IL-6 in liver I/R. In PBS-treated mice, a time-dependent statistically significant increase in transcripts for IL-6 was observed versus sham, with a peak effect noted at 18 hours after I/R (Fig. 5C). Levels of IL-6 transcripts in sRAGE-treated remnants consistently were increased significantly compared with sham controls at 1 and 18 hours after I/R (Fig. 5C). Overall, however, significantly decreased transcripts for IL-6 were observed in sRAGE-treated remnants versus PBS at 6 and 18 hours after I/R (Fig. 5C).


Hepatic injury mediated by I/R is an established component of the host response to liver resection or transplantation. Multiple pathways converge in the perturbed remnant to signal activation of endogenous inflammatory cells within the liver, as well as upregulation of key adhesion molecules and chemokines that mediate migration of inflammatory cells from the periphery into foci of activation and inflammation in the perturbed remnant. Once set in motion, these facets of the immune–inflammatory response join forces to stimulate tissue-destructive pathways and failure of regenerative programs.6–9 In this study, we provide the first in vivo evidence that RAGE activation is involved in the immediate proinflammatory stress response to I/R and that blockade of RAGE significantly improves survival, in parallel with suppression of exaggerated hepatocyte death and enhanced expression of mediators of regeneration.

The study of hepatic I/R has been hampered by the interplay between the detrimental effects of hepatic ischemia and portal obstruction with mesenteric congestion. This obstacle has been overcome by using models of partial clamping that permit the mesenteric venous outflow to decompress through the part of the liver that is not clamped. However, partial clamping permits the animal to survive even if the ischemic portion of the liver is nonviable. Therefore, resection of the unclamped portion after the ischemia permits a proper analysis of survival and the impact of potential therapy. The clinical syndrome observed in these animals, associated with diffuse hepatocyte apoptosis, has been prevented by ischemic preconditioning in similar models,9–11 and the beneficial effects observed with RAGE blockade seem to be comparable. Most pronounced is the abrogation of diffuse hepatocyte apoptosis, which may be a consequence of the beneficial effects of enhanced activation of NF-κB.

Our data suggested that in the first hour after injury, RAGE activation is associated with an injurious stress response characterized by activation of p44/42 and p38 MAPKs, in parallel with increased generation of TNF-α and IL-6. Activation of p44/p42 MAPKs has been shown to play a pivotal role in regulation of cell proliferation and differentiation.29 Further, activated p38 MAPK is linked strongly to the inflammatory response.30 In the present studies, RAGE-dependent activation of p38 MAPK and SAPK/JNK seems to supersede potentially proproliferative effects of p44/42 MAPK. Rather, important roles for activated SAPK/JNK MAPKs emerge, as evidenced by enhanced cell death in the PBS-treated injured liver.31, 32 It is important to note that in sRAGE-treated mice, at early times after total I/R, levels of TNF-α transcripts and serum TNF-α antigen were reduced compared with vehicle-treated animals. Later, by 18 hours after I/R, compared with sham and PBS-treated mice subjected to I/R, a striking increase in TNF-α was evident selectively in sRAGE-treated livers. Although TNF-α is linked to proinflammatory mechanisms and activation of cell death pathways in the injured liver,33, 34 its proregenerative properties have been established clearly by experiments using genetic and pharmacologic blockade of this key cytokine.35–37 These data suggest that early blockade of RAGE attenuates inflammation in the ischemic liver and thereby facilitates activation of TNF-mediated regenerative pathways.

In addition, our findings indicate that blockade of RAGE led to enhanced activation of NF-κB in liver remnants subsequent to I/R. In this context, we propose that NF-κB exerts pleiotropic effects in the liver; although its activation is associated with potent inflammatory responses in hepatic I/R,38 key roles for this factor in inhibition of apoptosis in the liver have been demonstrated definitively experimentally.39, 40 In our studies, we failed to detect significant activation of NF-κB in livers from PBS-treated I/R mice. However, at 1 hour after injury, increased nuclear translocation of AP-1 was evident in PBS-treated animals. In the presence of sRAGE, activation of AP-1 at 1 hour was markedly attenuated. Because AP-1 has been linked to increased transcription of TNF-α,41 these findings provide a mechanism whereby I/R promotes development of a rapid inflammatory and tissue-destructive response that augurs exaggerated cell death. In contrast, we speculate that in the presence of blockade of RAGE, suppression of early activation of AP-1 limits generation of proinflammatory mediators, such as TNF-α. We propose that in RAGE-blocked mice subjected to I/R, early and enhanced activation of NF-κB diverts intracellular pathways from those associated with inflammation and cell death, to mechanisms linked to recruitment and activation of proregenerative programs.

Last, our studies indicate that RAGE blockade does not impact significantly on neutrophil attraction and activation in the liver after I/R, as assessed by activity of MPO, and suggest that the predominant impact of RAGE blockade is downstream to the initial influx of neutrophils in hepatic I/R. Interestingly, our data highlight the novel finding that RAGE is primarily expressed in CD11c-expressing dendritic macrophages in the injured liver. In contrast, limited and patchy expression was evident in Kupffer cells and biliary cells. We recognize that our studies using whole liver homogenates do not permit precise localization of the perturbations in this system. A key goal of future investigation is to dissect the specific functions of liver dendritic cells expressing RAGE in the immune and inflammatory response in the liver subjected to I/R. It remains possible that RAGE-expressing cells contributing to the inflammatory response are, in part, those recruited from the periphery, because previous studies in the brain have demonstrated that RAGE is expressed in mononuclear phagocytes and T cells, especially those expressing CD4.42 Ultimately, it is most likely that in vivo, an event so profound as hepatic ischemia requires interplay between systemic events and those confined to the injured liver.

In conclusion, these observations suggest that activation of RAGE triggers a deleterious inflammatory response rapidly after acute I/R. We propose that blockade of RAGE signaling provides a means to attenuate hepatic I/R damage and to restore homeostatic regenerative programs in the injured liver.