RAGE, the receptor for advanced glycation end products, is a multiligand cell surface receptor that belongs to the immunoglobulin superfamily.1, 2 Advanced glycation end products (AGE), amyloid-β peptide and amphoterin are the major ligands of RAGE.2
AGE result from glycoxidation of tissue proteins such as albumin and collagen3, 4, 5 and lipids in response to long-term exposure to high concentrations of glucose such as occurs in diabetes and chronic renal failure patients.1, 4, 6, 7, 8 AGE are responsible for diabetic microangiopathy involving diabetic nephropathy and retinopathy by generated reactive oxygen species that cause tissue injury.6, 7, 8, 9, 10, 11 AGE are also associated with activation of nuclear factor (NF)-κB12, 13 and production of extracellular matrix proteinase.14, 15 Amyloid-β peptide causes the neural cell degeneration associated with Alzheimer's disease through RAGE-related tissue damage.8, 16, 17, 18
Amphoterin is encoded by the high mobility group-1 (HMG-1) gene and is not a glycation product.19 It has been isolated as a 30 kDa heparin-binding protein in growing brain tissue16, 20 and murine erythroleukemia cells.21 Amphoterin is expressed on the infiltrating edge of filopodia in both outgrowing migratory neurites and lung pneumocytes.20, 22 In murine erythroleukemia cells and lung pneumocytes, amphoterin is thought to play a role in cell and tissue differentiation.21, 22 Amphoterin promotes cell survival through production of RAGE-mediated Bcl-2.23 Amphoterin is also produced by many types of neoplasms and immature cells,24 and expression is closely correlated with cell migration and invasion induced by RAGE-associated intracellular signaling pathways, including GTPases, Cdc42, Rac, mitogen-activated protein kinase (MAPK), extracellular signal-regulated kinase (ERK)-1/2, p38, c-Jun N-terminal kinase (JNK) and NF-κB.24, 25 Amphoterin shares intracellular signaling pathways with AGE; however, differences in the biologic effects of these molecules remain unclear. In the present study, we show that amphoterin is more closely associated than AGE with the invasive activities of colorectal cancer cells.
Four colorectal carcinoma cell lines were studied. Colo320, DLD1 and WiDr were obtained from the Japanese Cancer Research Resources Bank (JCRB). TCO was kindly provided by Dr. Wataru Yasui (Hiroshima University, Hiroshima, Japan).26 All cell lines were routinely maintained in RPMI-1640 (Nissui, Tokyo, Japan) containing 10% FBS (Whittaker, Walkersville, MD) under 5% CO2 in air at 37°C.
Cultured cells were harvested from 80% confluent monolayer cultures by a brief trypsinization with 0.1% trypsin and 0.l% EDTA. The cells were seeded at a density of 2,000 cells per wells of 96-well tissue culture plates and cultured for 12 hr in the regular medium. The cells were washed with PBS twice and treated under conditions mentioned in Results. Cell growth was monitored after 12, 24, 36 and 48 hr by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay.27
Reverse transcriptase-polymerase chain reaction
RAGE and amphoterin (HMG-1) mRNA expression was assessed with RT-PCR using 0.5 μg total RNA extracted by an RNeasy kit (Qiagen, Hilden, Germany). Primer sets for RAGE amplification were as follows: upper 5′-GGG CTC TTC ACA CTG C-3′ and lower 5′-ACT GCT CCA CCT TCT G-3′. Primer sets for amphoterin (HMG-1) amplification were as follows: upper 5′-GGA GAG ATG TGG AAT A-3′ and lower 5′-GGG AGT GAG TTG TGT A-3′. PCR products were electrophoresed in 2% agarose gel and stained with ethidium bromide. β-Actin mRNA was also amplified for internal control. The experiment was repeated twice.
The 18-mer S-oligodeoxynucleotide (ODN) for antisense sequence of the 6th to 23rd nucleotide of RAGE cDNA (referred to GenBank AB036432) and the 18-mer S-ODN for antisense sequence of the 1st to 18th nucleotide of amphoterin cDNA (referred to GenBank X12597) were synthesized and purified by reverse-phase high-performance liquid chromatography (Espec Oligo Service, Tsukuba, Japan). The sequence of RAGE antisense was 5′-CTG CTT CCT TCC AGG GTC-3′; the sequence of amphoterin antisense was 5′-AGG ATC TCC TTT GCC CAT-3′. The sense sequence 18-mers was for negative control. Each cell line was pretreated with 3 μM of antisense or sense S-ODN for 6 days with medium exchange and addition of antisense or sense S-ODN every 2 days. The cells were then used for further experiments.
Whole cell lysates were prepared as described previously.28 Fifty-microgram lysates were subjected to immunoblot analysis using 12.5% SDS-PAGE followed by electrotransferring onto nitrocellulose filter. The filters were incubated with primary antibody and then with peroxidase-conjugated anti-goat IgG antibody (Medical and Biological Laboratories, Nagoya, Japan) in the secondary reaction. For internal control of loaded protein amount, tubulin was detected by the specific antibody (Zymed, South San Francisco, CA). The immune complex was visualized using the ECL Western blot detection system (Amersham, Aylesbury, UK). Antibodies used were anti-RAGE antibody (Santa Cruz Biotechnology, Santa Cruz, CA), anti-HMG-1 antibody (Upstate Biotechnology, Lake Placid, NY) anti-phospho-ERK1/2 (Upstate Biotechnology), anti-phospho-Rac1 (Santa Cruz Biotechnology), anti-MMP9 (Novocastra Laboratories, Newcastle upon Tyne, UK), anti-phospho-AKT (New England Biolabs, Beverly, MA), anti-iNOS (Santa Cruz Biotechnology) and anti-NF-κBp65 (Santa Cruz Biotechnology). For semiquantification, the specific signals on immunoblotted membrane detected by each antibody were computer-captured and quantified with NIH image computer software (National Institutes of Health, Bethesda, MD). The signal intensities were compensated for by tubulin intensities as internal controls and represented with a standard of the production by untreated cells (set to 1.0).
Wound healing assay
To evaluate cell motility, a wound healing assay was performed. Colo320 and WiDr cells were exposed to 3 μM antisense or sense S-ODN for RAGE or amphoterin for 6 days. Then they were reseeded onto 3.5 cm culture dishes and continuously treated with antisense S-ODN or sense S-ODN for RAGE or amphoterin. After 2 days, cells grown to subconfluence were scraped to make a cell-free area with a sharp edge. Cells migrating into the scraped area were counted every 12 hr after scraping.
In vitro invasion assay
A modified Boyden chamber assay was performed to examine the in vitro invasion of colon cancer cells, Colo320 and WiDr cells treated with antisense or sense S-ODN for RAGE or amphoterin. Polycarbonate filters (pore size 3 μm; diameter 5 mm) were glued to collagen type IV inserts (Becton-Dickinson Labware, Bedford, MA), which were placed in the wells of 24-well tissue culture plates. Cells were suspended in 500 μl of regular medium and placed in the upper part of the chamber. The lower part of the chamber was filled with regular medium. After 24 hr of incubation at 37°C, the filters were carefully removed from the inserts, stained with hematoxylin for 10 min and mounted on microscopic slides. The numbers of stained cells were counted in whole inserts at 100× magnification. Invasion activity was quantified by the average of cells per insert well.
Advanced glycation end products of bovine serum albumin (AGE-BSA)
BSA (1 mM, fraction V, fatty acid-free, endotoxin-free; Sigma, St. Louis, MO) was incubated with a high-concentration glucose solution [1 M glucose, 1.5 mM phenylmethylsulfonyl fluoride (PMSF), 0.5 μg/ml leupeptin, 2 μg/ml aprotinin, 0.5 mM EDTA, 100 units/ml ampicillin (Sigma), 40 μg/ml streptomycin (Sigma)] at 37°C for 6 weeks under sterile conditions. After incubation, free glucose and BSA were removed by dialysis against PBS. The concentration of AGE-BSA was determined by the method of Bradford29 (Bio-Rad, Hercules, CA). Cells were treated with 500 μg/ml AGE-BSA or 500 μg/ml native BSA as a control.
Expression of RAGE and amphoterin mRNAs and proteins in colorectal carcinoma cells
Expression of RAGE and amphoterin mRNAs and proteins was examined in 4 colorectal cancer cell lines (Fig. 1). RAGE mRNA and protein were detected in all colorectal cancer cell lines. Amphoterin mRNA and protein were detected at similar levels in all cell lines.
Effects of RAGE and amphoterin antisense S-ODNs on RAGE and amphoterin expression
Colo320 and WiDr cells were exposed to antisense or sense S-ODNs specific for RAGE and amphoterin for 6 days, and mRNA and protein expression of RAGE and amphoterin were examined by RT-PCR and immunoblotting, respectively (Fig. 2). RAGE and amphoterin mRNA levels were reduced after treatment with antisense S-ODNs in comparison with treatment with sense S-ODNs. Levels of RAGE and amphoterin mRNAs in Colo320 cells treated with the appropriate antisense S-ODN were 10% and 14%, respectively, of the levels found in cells treated with sense S-ODN. RAGE and amphoterin mRNA expression by WiDr cells treated with the antisense S-ODN was reduced to 50% and 13%, respectively, of the levels found in cells treated with sense S-ODN. Protein levels of RAGE and amphoterin in Colo320 cells treated with antisense S-ODN were 1% and 15%, respectively, of the levels in cells treated with sense S-ODN. Protein levels of RAGE and amphoterin in WiDr cells treated with antisense S-ODN were 10% and 12%, respectively, of the levels in cells treated with sense S-ODN.
Effect of RAGE and amphoterin antisense S-ODNs on growth of Colo320 and WiDr cells
We then studied the effects of RAGE and amphoterin antisense S-ODNs on the growth of Colo320 and WiDr cells (Fig. 3a). Growth of Colo320 cells at 48 hr after treatment with 3 μM antisense S-ODN for RAGE or amphoterin was 60% (for both) that of cells treated with the appropriate sense S-ODNs. Growth of WiDr cells after treatment with 3 μM antisense S-ODNs for RAGE and amphoterin was 65% and 50% at 48 hr, respectively, that of cells treated with sense S-ODNs. Exposure of cells to 3 μM of either sense S-ODN did not cause growth inhibition.
Wound healing assay with Colo320 and WiDr cells
A wound healing assay was used to test cell motility. Colo320 and WiDr cells treated with 3 μM RAGE or amphoterin antisense S-ODN showed fewer budding nests than did sense S-ODN-treated cells. As shown in Figure 3b, the numbers of RAGE and amphoterin antisense S-ODN-treated Colo320 cells migrating into the scraped area were 25 ± 4 and 33 ± 4/1 mm2 respectively, after 48 hr of treatment. In contrast, the numbers of sense S-ODN-treated cells migrating into the scraped area were 348 ± 19 and 339 ± 24/1 mm2, respectively (p < 0.0001 for both, unpaired Mann-Whitney U test). The numbers of S-ODN-treated WiDr cells migrating into the scraped area were 122 ± 10 and 76 ± 7/1 mm2 for RAGE and amphoterin, respectively, whereas the numbers of WiDr cells treated with each sense S-ODN migrating into the scraped area were 227 ± 16 and 236 ± 18/1 mm2, respectively (p < 0.0001 for both, unpaired Mann-Whitney U test).
In vitro invasion assay with Colo320 and WiDr cells
In vitro invasion assay revealed that invasive activity was reduced significantly in both Colo320 and WiDr cells treated with RAGE or amphoterin antisense S-ODN (Fig. 3c). Twenty-four hours after treatment, the numbers of RAGE antisense S-ODN-treated Colo320 cells invading type IV collagen-coated membrane were significantly lower (13 ± 5 cells/ well) than those of cells treated with sense S-ODN (97 ± 12 cells/well, p = 0.0001, unpaired Mann-Whitney U test). The numbers of amphoterin antisense S-ODN-treated Colo320 cells invading the membrane were also significantly lower (24 ± 6 cells/well) than those of cells treated with sense S-ODN (91 ± 10 cells/well, p = 0.0015, unpaired Mann-Whitney U test). The numbers of RAGE antisense S-ODN-treated WiDr cells invading type IV collagen-coated membrane were significantly lower (48 ± 5 cells/well) than those of cells treated with sense S-ODN (82 ± 9 cells/well, p = 0.0036, unpaired Mann-Whitney U test). The numbers of amphoterin antisense S-ODN-treated WiDr cells invading the membrane were also significantly lower (37 ± 4 cells/well) than those of cells treated with sense S-ODN (86 ± 8 cells/well, p = 0.0018, unpaired Mann-Whitney U test). In WiDr cells, RAGE antisense S-ODN showed less effects on cell growth, migration and invasion than amphoterin antisense S-ODN (Fig. 3a–c). A lower level of RAGE constitutive production (Fig. 1) and lower suppression of RAGE production by RAGE antisense S-ODN (Fig. 2b) on WiDr cells than Colo320 cells might be responsible for the results.
Comparison of biologic effects of amphoterin and AGE
To compare the effects of amphoterin and AGE on cell growth, migration and invasion, we generated glycated BSA (AGE-BSA). As shown in Figure 4a, the electrophoretic mobility of AGE-BSA was reduced compared with that of native BSA due to the modification with glucose. Colo320 and WiDr cells were treated with amphoterin antisense S-ODN alone, amphoterin antisense S-ODN plus AGE-BSA, amphoterin antisense S-ODN plus Colo320-conditioned medium and amphoterin sense S-ODN plus AGE-BSA, and the effects on growth, migration and invasion were compared (Fig. 4b–g). Amphoterin antisense or sense S-ODNs plus BSA treatment was done as negative control for restoration. RAGE antisense S-ODN was done for control for suppression of the RAGE/amphoterin system.
Reduced production of amphoterin by treatment with amphoterin antisense S-ODN inhibited cell growth, migration and invasion by both Colo320 and WiDr cells. Cell growth, migration and invasion activity of Colo320 and WiDr cells treated with amphoterin antisense S-ODN were restored by treatment with Colo320-conditioned medium, which was confirmed to contain amphoterin by immunoblotting. Cell growth, migration and invasive activities of Colo320 cells treated with amphoterin antisense S-ODN plus AGE-BSA were 69%, 19% and 36%, respectively, of those of cells treated with amphoterin sense S-ODN plus BSA. Cell growth, migration and invasive activities of WiDr cells treated with amphoterin antisense S-ODN plus AGE-BSA were 71%, 22% and 32%, respectively, of those of cells treated with amphoterin sense S-ODN plus BSA. These findings indicate that AGE-BSA has less pronounced effects on cell growth, migration and invasion compared with those of amphoterin. Colo320 and WiDr cells treated with amphoterin sense S-ODN plus AGE-BSA showed less activity of cell growth, migration and invasion than cells treated with amphoterin sense S-ODN plus BSA. The results suggest that effects of intrinsic amphoterin are suppressed by AGE-BSA.
Differential effects of amphoterin and AGE-BSA on proteins in RAGE-related signaling pathways
We then studied production and phosphorylation of proteins in RAGE-related signaling pathways (Fig. 5). Phosphorylation of ERK1/2, Rac1 and AKT and production of iNOS, NF-κBp65 and MMP9 were examined in untreated Colo320 cells, Colo320 cells treated with amphoterin antisense S-ODN alone, cells treated with amphoterin antisense S-ODN plus Colo320-conditioned medium, cells treated with amphoterin antisense S-ODN plus AGE-BSA and cells treated with RAGE antisense S-ODN. Phosphorylation of ERK1/2, Rac1 and AKT and production of MMP9 were decreased by treatment with amphoterin antisense S-ODN but were restored to control levels by cotreatment with Colo320-conditioned medium. In contrast, combined treatment with amphoterin antisense S-ODN and AGE-BSA yielded only partial recovery. The decreased production of iNOS and NF-κBp65 caused by treatment with amphoterin antisense S-ODN was ameliorated by cotreatment with AGE-BSA, whereas cotreatment with Colo320-conditioned medium caused only partial recovery. RAGE antisense S-ODN-treated cells showed decrease of phosphorylation of ERK1/2, Rac1 and AKT and production of iNOS, NF-κBp65 and MMP9 similar to amphoterin antisense S-ODN. The decrease in phosphorylation and production by RAGE antisense S-ODN treatment was not restored by Colo320-CM and AGE-BSA (data not shown).
The multiligand receptor RAGE and 1 of its ligands, amphoterin, are recently described metastasis-related genes.25, 30 In the present study, we found that coexpression of RAGE and amphoterin enhances migration and invasion by colon carcinoma cells. RAGE also interacts with AGE, S100 protein and amyloid-β;2, 23;2, 23 however, the roles of these ligands in the invasive activity of cancer cells are not known.
Nonenzymatically glycated BSA (AGE-BSA) is typically used as a RAGE ligand in in vitro assays.5, 31 In the present study, AGE-BSA had only partial effects on cell growth, migration, and invasion in comparison with those of amphoterin. Our findings were supported by the observation that AGE-BSA had only partial effects on phosphorylation of ERK1/2, Rac1 and AKT and production of MMP9 in comparison with those of amphoterin. In contrast, amphoterin had partial effects on upregulation of iNOS and NF-κBp65 expression levels in comparison with those of AGE-BSA. Both amphoterin and AGE are reported to activate ERK1/2, p38 and JNK,25, 32, 33 whereas amphoterin and AGE had different abilities to activate ERK1/2. We also observed a decrease in phosphorylation of AKT by amphoterin antisense S-ODN treatment. This decrease was restored by Colo320 conditioned medium but only partially by AGE-BSA. These data suggest that the differential effects of these 2 ligands are caused by the selectivity of intracellular signaling pathways.
RAGE is a multipathway-multifunctional receptor. There are many intracellular signaling pathways, including Cdc42/Rac,24 MAPK,24, 25 and NF-κB,13, 31 and subsequent induction of MMP2/9,25 vascular endothelial growth factor (VEGF)5 and iNOS.6, 7, 8, 9, 10 It has been reported that amphoterin enhances cell motility24, 25 and that AGE contributes to generation of oxidative stresses and subsequent tissue injury.6, 7, 8, 9, 10 However, the ligand specificity of RAGE with respect to the various signal pathways has not been clarified. Because our results suggest a difference in ligand specificity between amphoterin and AGE, further studies are needed.
The HMG-1 gene, which encodes a non-histone chromosomal protein, yields 3 transcripts.34, 35 The gene that encodes a heparin-binding protein associated with neurite outgrowth (amphoterin)36, 37 has been identified as the HMG-1 gene.19 Although the extracellular function of nuclear protein HMG-1 as a RAGE ligand is unclear, there have been recent reports that HMG-1 expression is upregulated and that HMG-1 is secreted in response to stimulation of cells with proinflammatory cytokines,38, 39, 40 endotoxin41, 42 and platelet activators.43 Cell necrosis also causes release of amphoterin/HMG-1 into the extracellular space.44 Thus amphoterin/HMG-1 functions both as a chromatin structural protein and as a cytokine.38, 45 These findings support an autocrine/paracrine mechanism for the amphoterin/RAGE activity observed in colorectal carcinoma cells in the present study. Further studies are necessary to clarify the precise mechanism of amphoterin/HMG-1 translocation and secretion.
In conclusion, we showed that amphoterin/HMG-1 is associated more closely with cancer-promoting characteristics than is AGE. The RAGE/amphoterin system may be a good target for anticancer therapies.