| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* San Raffaele Scientific Institute, Milan, Italy;
Department of Biotechnology and Biosciences, University of Milano-Bicocca, Milan, Italy;
Department of Medicine I and Clinical Chemistry, University of Heidelberg, Heidelberg, Germany;
Biochemical Institute, University Kiel, Kiel, Germany; and
|| Università Vita Salute San Raffaele, Milan, Italy
2Correspondence: Università Vita Salute San Raffaele, via Olgettina 58, 20132 Milano, Italy. E-mail: bianchi.marco{at}hsr.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: cytokines • HMGB1 • inflammation • receptor shedding • atherosclerosis • diabetes
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In addition to N
-(carboxymethyl)lysine-modified proteins (a class of AGEs), RAGE binds to HMGB1, S100/calgranulins, and β-amyloid, and acts as a counterreceptor for the leukocyte β2 integrin Mac-1 during leukocyte recruitment (2)
. Our group has been long interested in HMGB1, and we showed that this nuclear protein is passively released by necrotic cells, but not by apoptotic cells, thus serving as a signal of accidental cell death. In this capacity, HMGB1 promotes the migration of inflammatory cells to sustain inflammation but also triggers immune responses to stem the invasion of microorganisms and initiates the regeneration of the damaged tissue, both by recruiting differentiated cells that migrate to the site of damage to replace dead cells, and by activating stem cells to proliferate. RAGE is apparently involved in all responses that depend on HMGB1 (3)
.
In normal adult tissues RAGE is expressed at low levels, except in lung, where it is very abundant (4
, 5)
; however, RAGE becomes rapidly up-regulated at sites where its ligands accumulate, and produces a sustained cellular activation through multiple intracellular signaling pathways. Moreover, RAGE participates in a number of pathological processes, including cancer progression, diabetes, atherosclerosis, and Alzheimer’s disease (2
, 6)
. Thus, RAGE is an attractive biological target for the treatment of several pathologies and might well be a safe one: RAGE knockout (KO) mice are viable and fertile and are protected from septic shock (7)
.
RAGE exists as a full-length isoform (FL-RAGE) that is bound to cell membrane and soluble isoforms. Structurally, FL-RAGE consists of three extracellular immunoglobulin-like domains (one V-type and two C-type), a single transmembrane domain, and a short C-terminal cytosolic tail (1
, 8)
. The V-type domain is involved in S100B binding (9)
, whereas the cytosolic segment appears essential for intracellular signaling (10)
. Additional RAGE isoforms generated by alternative splicing of the pre-mRNA have been described in human cells (11)
. One encodes an N-terminal-deleted membrane-bound form lacking the extracellular V-type domain (
N-term). Another variant encodes a secreted form lacking the transmembrane domain and with a distinct carboxyl-terminal amino acid sequence [C-term, endogenous secretory RAGE (esRAGE)]. Other soluble RAGE isoforms have been inferred from cloned transcripts (12
, 13)
, but at present, only esRAGE has been identified at the protein level.
Soluble RAGE appears to function as a decoy receptor that neutralizes circulating ligands. For this reason, recombinant soluble RAGE has been extensively and successfully tested in several animal models for the treatment of RAGE-mediated diseases, including type II collagen-induced arthritis (14)
, delayed-immune hypersensitivity (DTH) (7)
, and diabetic atherosclerosis (15
, 16)
.
Two general mechanisms are usually responsible for the generation of soluble receptors: they either derive from an alternatively spliced mRNA or are cleavage products of the membrane-bound form. Soluble RAGE has hitherto been considered equivalent to the alternative splicing isoform esRAGE. Soluble RAGE purified from mouse lung, however, has a C-terminal sequence that is identical to the extracellular portion of FL-RAGE (17)
; this truncated form may derive from proteolytic cleavage.
Ectodomain cleavage (shedding) regulates the functions of many membrane-bound proteins, including ligands of the epidermal growth factor receptor, transforming growth factor 
(TGF-), interleukin-6 (IL-6) receptor, tumor necrosis factor (TNF-) and its receptor, Notch, L-selectin, E-cadherin, N-cadherin, vascular cell adhesion molecule 1 (VCAM-1), cluster of differentiation 44 (CD44), L1 adhesion molecule, and many others (reviewed in refs 18
19
20
21
). The proteases responsible for shedding belong to the zinc-dependent metzincin family of metalloproteases; in particular, the ADAM (a disintegrin and metalloproteinases) family is responsible for the cleavage of the majority of shed proteins. Generally, shedding occurs constitutively but can be induced by ligand binding, cell stimulation with phorbol 12-myristate 13-acetate (PMA), ionomycin, or by cholesterol depletion. Among the numerous members of the ADAM family, ADAM10 and ADAM17/TACE (TNF--converting enzyme) shed most cell surface molecules.
In this study, we show that mouse and human soluble RAGE are generated by proteolytic cleavage of membrane-bound FL-RAGE, whereas esRAGE represents a quantitatively minor isoform. Both forms of soluble RAGE, cleaved and alternatively spliced, act as decoy receptors of HMGB1. Binding of HMGB1, or exposure to PMA, promotes RAGE shedding. Using specific chemical inhibitors and a panel of ADAM-deficient mouse embryonic fibroblasts, we identified ADAM10 as the major RAGE sheddase. FL-RAGE is difficult to detect on most cells and tissues, except lung, where it is abundant. However, incubation of peripheral blood mononuclear cells (PBMCs) with metalloprotease inhibitors moderately increases its surface representation. These data suggest that most circulating RAGE derives from shedding, and that its level may correlate with ongoing inflammation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-hRAGE N-term) was from Chemicon (Temecula, CA, USA). Goat anti-human RAGE (-RAGE N-term) was from R&D Systems (Minneapolis, MN, USA); this antibody recognizes mouse RAGE as well. Rabbit polyclonal against the intracellular domain of human RAGE (-RAGE C-term) was from Abcam (cat. ab3611-100; Cambridge, UK); this antibody recognizes the intracellular domain of mouse RAGE as well. Alexa Fluor 488-conjugated anti-mouse secondary antibody was from Molecular Probes-Invitrogen (Eugene, OR, USA). Recombinant, LPS-free HMGB1 was provided by HMGBiotech (Milan, Italy).
Cell culture and transfection
Immortalized and primary MEFs from Presenilin1/2–/– (PS1/2–/–), Adam10–/–, Adam17–/–, Adam9–/– mice and wild-type siblings were described elsewhere (22)
. MEFs, human embryonic kidney (HEK) 293, HeLa, and 3T3 cells were grown in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS) and 1% penicillin/streptomycin. Cells were transfected with FuGENE 6 (Roche Molecular Biochemicals, Mannheim, Germany), according to the manufacturer’s instructions. THP-1 cells were grown in RPMI 1640 medium supplemented with 10% FCS and 1% penicillin/streptomycin.
To reconstitute the Adam10–/– MEFs with mouse ADAM10, cells were transfected with pcDNA3.1/mADAM10 (23)
and selected with 300 µg/ml zeocin (Invitrogen, Carlsbad, CA, USA) for 2 wk. The pool of zeocin-resistant cells (A10) was assayed for ADAM10 mRNA expression by real-time polymerase chain reaction (PCR). In brief, total RNA was isolated from Adam10+/+, Adam10–/–, or A10 cells using Illustra RNAspin mini-RNA isolation kit (GE Healthcare, Chalfont St. Giles, UK). Poly(A)+ RNAs were reverse-transcribed from 5 µg total RNA using an oligo(dT) primer and Moloney murine leukemia virus reverse transcriptase, as described by the manufacturer (Fermentas, Glen Burnie, MD, USA). cDNAs were amplified by real-time PCR on a LC480 instrument (Roche Biochemicals), using the relative quantification software, with LightCycler 480 SYBR Green I Master mix and primers for mouse β-actin (TGACGGGGTCACCCACACTGTGCCCATCTA and CTAGAAGCATTGCGGTGGACGATGGAGGG) or mouse ADAM10 (GTCAGCTCTATATCCAGACAGATCA and GAGTACACTCTGTTCCAGAATCATG).
Isolation of human PBMCs
Human PBMCs were isolated from venous blood of healthy volunteers. Briefly, whole heparinized blood was mixed 1:1 v/v with a solution of 3.5% dextran (Sigma) and left at 37°C, 5% CO2 for 30 min until the red blood cells had sedimented. The upper layer of cells was removed, washed several times with PBS, and layered onto 0.25 vol of Ficoll (Axis-Shield PoC AS, Oslo, Norway). After centrifugation at 800 g for 20 min at 20°C, the layer of mononuclear cells was removed. PBMCs were centrifuged 3 times in PBS to eliminate Ficoll, resuspended in DMEM supplemented with 1% normal human serum, and plated in a 6-well plate (2x107 cells/well). After 1 h incubation at 37°C, nonadherent cells were removed, and monocyte-enriched cultures were incubated overnight. Cells were then treated for 1 h with GM6001, TAPI-1, or dimethyl sulfoxide (DMSO) as control, and assayed for membrane-bound FL-RAGE expression by flow cytometry.
Isolation and sequence determination of RAGE splice variants
Total RNA was isolated using GenElute mRNA miniprep kit (Sigma) from HEK293 and THP-1 cells, treated or not with PMA. Poly(A)+ RNAs were reverse-transcribed from 5 µg total RNA using an oligo(dT) primer and Moloney murine leukemia virus reverse transcriptase, as described by the manufacturer (Fermentas). RAGE cDNAs were amplified with forward and reverse primers using AmpliTaq Gold polymerase (Applied Biosystems, Foster City, CA, USA) in 35 cycles of denaturation (95°C for 1 min), annealing (55°C for 45 s) and elongation (72°C for 2 min). Primers RAGE-Fw AGGACCCTGGAAGGAAGC and RAGE-Rev TTGGCAAGGTGGGGTTATAC correspond to sequences in the 5' untranscribed region (UTR) in exon 1 and the 3' UTR in exon 11 of human RAGE (GenBank2 accession no. NM_001136). After gel extraction, cDNAs were cloned into pGemT-Easy (Promega, Madison, WI, USA). Plasmid DNAs were purified with GenElute Miniprep plasmid isolation kit (Sigma), and their nucleotide sequences were determined by BMR Genomics (Padova, Italy).
RAGE expression plasmids
To generate expression plasmids for human RAGE-FL and RAGE C-term, cDNAs obtained from THP-1 cells were subcloned from pGemT-Easy into pcDNA3 by restriction with EcoRI. To generate an expression plasmid for RAGE N-term, the corresponding cDNA, containing the first exon and the first intron, was initially cloned by restriction into pcDNA3. However, this construct could never express a protein when transfected into HEK293 cells (see Results). Therefore, an alternative strategy was tried. A fragment starting from the ATG located just downstream of intron 1 was amplified from the pGemT-Easy construct with Vent Polymerase (New England Biolabs, Beverly, MA, USA) and the following primers: forward-BglII AAGAATTCAGATCTATGAACAGGAATGGAAAGGA and reverse-EcoRI AACCAAGAATTCTCAAGGCCCTCCAGTACTACT.This construct too failed to express a detectable protein, suggesting that RAGE N-term cannot be expressed or is expressed and immediately degraded.
RAGE-FL, RAGE C-term and RAGE N-term GFP-fusion proteins were obtained by amplification of the corresponding ORFs with Vent Polymerase from pGemT-Easy constructs with specific primers and cloned into pEGFP-N1 linearized with BglII and EcoRI: forward, BglII-RAGE-FL/C-term: ATCCAGATCTAGGACCCTGGAAGGAAG; forward, BglII-RAGE N-term: AAGAATTCAGATCTATGAACAGGAATGGAAAGGA; reverse, EcoRI-RAGE-FL/N-term: AAG AAT TCT AGG CCC TCCAGTACTACTCTC; reverse, EcoRI-RAGE C-term: AAGAATTCTCATGTGTTGGGGGCTATCTT.
To construct an expression plasmid for C-terminal myc/His-tagged human full-length RAGE (hFL-RAGE-myc), the cDNA was amplified by PCR with Pfu polymerase (Promega) using the following primers: forward, AGGAATTCCACCATGGCAGCCGGAACA (containing an EcoRI restriction site and the Kozak consensus sequence), and reverse, GCTCTAGAAGGCCCTCCAGTACTACT (containing an XbaI restriction site). PCR conditions were denaturation at 95°C for 2 min; 35 cycles of amplification (95°C for 1 min, 65°C for 1 min, 72°C for 3 min); final elongation at 72°C for 5 min. PCR products and pcDNA 3.1 myc/His A vector (Invitrogen) were digested with EcoRI and XbaI, purified, ligated, and sequence-verified.
ELISA
Total soluble RAGE in human sera was determined using a sandwich ELISA kit (R&D Systems), according to the manufacturer’s protocol. Briefly, ELISA plates coated with mouse monoclonal antibody against RAGE were used for capture of soluble RAGEs. After incubation with sera, polyclonal antibody against the extracellular portion of RAGE was used for quantitative detection. The minimum detectable dose of soluble RAGEs was 0.062 ng/ml, and linearity was conserved between 0.062 and 4 ng/ml.
The esRAGE form in human sera was determined using a recently developed sandwich ELISA kit (B-Bridge International, Tokyo, Japan), according to the manufacturer’s protocol. In brief, esRAGE was detected using as coating antibody the 278–13G4 mAb that recognizes the C-type Ig domain, and an esRAGE-specific polyclonal antibody raised against the unique C-terminal amino acids 332–347, EGFDKVREAEDSPQHM, as detection antibody. The minimum detectable dose of esRAGE was 0.1 ng/ml, and the linearity was conserved between 0.1 and 3.2 ng/ml.
Flow cytometry
For cell surface staining, 5 x 105 THP-1, HEK/pcDNA, or HEK/FL-RAGE cells were collected and blocked with PBS-10% NHS for 30 min on ice. RAGE expression was detected using anti-hRAGE N-term dissolved in PBS-1% NHS for 30 min on ice, followed by Alexa 488-conjugated anti-mouse antibody for 30 min on ice. Cells were fixed with PBS-2% paraformaldeyde (PFA) and analyzed on a FACScalibur (BD Biosciences PharMingen, San Jose, CA, USA) flow cytometer. At least 10,000 events were collected for each sample. Results were processed using CellQuest software (BD Biosciences). Negative control consisted of isotype-matched antibodies.
Western blot analysis
HeLa, HEK293, 3T3, or MEF cells (2x105) were plated in 6-cm tissue culture plates and transfected with the indicated vectors. After 24 h, the complete medium was replaced with OptiMEM medium for 1 h. Where indicated, PMA or HMGB1 was included in the medium. Inhibitors GM6001, TAPI-1, GW280264X, and GI254023X were added 15 min before PMA. Culture supernatants were precipitated with 0.5 vol of 100% cold acetone and redissolved in 2x SDS-PAGE loading buffer. Cells were lysed in RIPA buffer (10 mM Tris-HCl, pH 7.2; 150 mM NaCl; 5 mM EDTA; 0.1% SDS; 1% sodium deoxycholate; 1% Triton X-100) in the presence of protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1:100 Sigma protease inhibitor cocktail) and phosphatase inhibitors (1 mM sodium orthovanadate, 1 mM sodium fluoride). After 3 min on ice, cells were scraped, and lysates were clarified by syringing 3 times and centrifuging at 10,000 g for 30 min at 4°C. Supernatants were recovered, and protein concentration was determined by the Bradford method (Bio-Rad, Munich, Germany), using BSA as a standard.
For detection of soluble RAGE in supernatants of THP-1 undifferentiated (Und) cells, 107 cells were starved in RPMI 1640 for 48 h. For THP-1 differentiation, cells were treated with 0.5 µM PMA for 4 h and incubated for 48 h in RPMI 1640 supplemented with 10% FCS; cells were then starved in RPMI 1640 for 48 h (Diff). When indicated, differentiated THP-1 cells were treated with 5 µg/ml LPS overnight (Diff+LPS). Supernatants were precipitated with 100% cold acetone and analyzed for the presence of soluble RAGE.
Total protein extracts were boiled, separated by SDS-PAGE, and electrotransferred on nitrocellulose membranes (Protran; Whatman, Dassel, Germany). The membranes were blocked in Tris-buffered saline Tween 20 (TBST) (10 mM Tris, pH 7.4; 0.5 mM NaCl; 0.1% Tween 20) containing 3% bovine serum albumin (Sigma) or in TBST containing 5% powdered skimmed milk. The blots were first probed with the indicated primary antibodies diluted in TBST with 5% powdered skimmed milk, and then with horseradish peroxidase-conjugated secondary antibodies. Proteins were visualized by an enhanced chemiluminescence (ECL) detection system (GE Healthcare). Membranes were stripped by using a stripping solution (2% SDS; 62.5 mM Tris, pH 6.8; 0.7% v/v β-mercaptoethanol), incubating at 60°C for 30 min, and abundantly washing in TBST.
Bands on films were quantified using an Epson 3170 photo scanner (Seiko Epson, Suwa, Japan) and ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA).
Soluble RAGE production
For production of cleaved RAGE (cRAGE) or esRAGE, HEK293 cells were transfected with an expression plasmid for human FL-RAGE (pcDNA/FL-RAGE) or RAGE C-term (pcDNA/esRAGE), respectively. Forty-eight hours post-transfection, cells were starved in DMEM without serum and left for an additional 48 h. The cell supernatants were collected and concentrated with Centriprep (Millipore, Bedford, MA, USA). The concentration of soluble RAGEs was determined by ELISA (R&D Systems), according to the manufacturer’s protocol. Supernatant of HEK293 cells transfected with the empty vector (pcDNA) was produced and processed as well, and used as negative control.
Migration assay
Chemotaxis was assayed using the Cultrex 96-well cell migration assay (Trevigen, Gaithersburg, MD, USA), according to the manufacturer’s instructions. Briefly, the system utilizes a modified Boyden chamber design with an 8-µm polyethylene terephtalate membrane. Detection of cell migration is quantified using calcein acetomethylester (AM), which is internalized by cells. Intracellular esterases cleave AM to generate free calcein, which fluoresces. Relative fluorescence units (RFU) are used as a measure to quantify the number of cells that have migrated.
Serum-free DMEM (negative control), DMEM containing 1 nM HMGB1, and DMEM with 10% FBS (positive control) were placed in the lower part of the chambers. Fifty thousand 3T3 cells resuspended in 50 µl DMEM containing 1 µg/ml cRAGE, esRAGE, or an equal volume of conditioned medium from HEK/pcDNA cells were placed in the upper part of the chambers. The plate was incubated at 37°C in 5% CO2 for 3 h and read at 485-nm excitation, 520-nm emission with a Victor 3 plate reader (Perkin Elmer, Waltham, MA, USA).
Animals
C57BL/6 Rage-null mice were described previously (7)
, and C57BL/6 wild-type control were purchased from Charles River Laboratories (Calco, Italy).
Lung extract
Lungs isolated from 4-mo-old Ager-null and wild-type mice, or human lung tissue from two autopsies, were homogenized in RIPA lysis buffer. After centrifugation, the lysates were cleared and protein concentration was determined by the Bradford method (Bio-Rad), using BSA as a standard.
Glycopeptidase F assay
To analyze soluble RAGE glycosylation, the supernatant of 107 THP-1 cells was concentrated and precipitated with cold acetone. The pellet was dissolved in PNGase F buffer, and the assay was performed according to the manufacturer’s instructions (New England Biolabs). The corresponding cell lysate was treated similarly. Samples were then analyzed by Western blot analysis.
Data analysis
Statistical analysis of the results was performed by unpaired Student’s t test. Values of P < 0.05 were considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. We analyzed RAGE expression in the human myelomonocytic cell line THP-1, which can be induced to differentiate into macrophage-like cells by treatment with PMA, and had been previously used to study RAGE-mediated signaling and adhesion (24
, 25)
.
To characterize the mRNA species that encode different RAGE isoforms poly(A) RNAs were isolated from undifferentiated and PMA-differentiated THP-1 cells and retrotranscribed; after PCR, with primers corresponding to 3' and 5' UTRs of human RAGE, cDNAs were cloned. We fully sequenced 34 clones derived from undifferentiated THP-1 cells, and 48 clones were isolated from PMA-treated THP-1 cells. As summarized in Fig. 1A
, four different mRNA species were identified, corresponding to FL-RAGE, RAGEN-term, RAGEC-term (esRAGE), and RAGENC-term. FL-RAGE cDNAs encode a protein of 404 amino acids (aa) with a signal peptide (SP) of 22 aa, an extracellular (EC) domain of 322 aa comprising 3 Ig domains, a transmembrane (TM) domain of 19 aa, and an intracellular (IC) domain of 41 aa. RAGEN-term codes for a protein lacking the SP and the first Ig domain, whereas esRAGE lacks the TM and IC domains and has an alternative stretch of 16 aa at the C-terminal. Interestingly, almost 90% of clones correspond to full-length RAGE (FL-RAGE), less than 10% to RAGEN-term, and the rest to esRAGE and the variant lacking both the N-terminal and C-terminal regions (RAGENC-term). PMA-induced differentiation does not affect this relative representation. RAGEN-term and soluble RAGE reportedly account each for 
1/3 of transcripts in endothelial cells, with FL-RAGE representing the remaining 1/3 (11)
.
|
We then constructed expression vectors for every RAGE isoform and transfected them into HEK293 cells (which do not express a detectable level of RAGE mRNAs, data not shown). As a control, we used a cDNA for FL-RAGE that we described previously (26)
. The cell lysates were probed both with a monoclonal antibody raised against the N-terminal portion of human RAGE (-hRAGE N-term) and with a polyclonal C-terminal specific antibody (-RAGE C-term) that was raised against the intracellular segment, and recognizes both human and mouse RAGE (Fig. 1C
). The -hRAGE N-terminal antibody recognized bands corresponding to FL-RAGE (molecular mass 40.7 kDa, running at 55 kDa) and esRAGE (molecular mass 37.0 kDa, running at 49 kDa), but no band corresponding to RAGEN-terminal. The -RAGE C-terminal antibody recognized FL-RAGE but neither esRAGE, as expected, nor RAGEN-terminal. Because none of the antibodies could detect the RAGEN-terminal protein, we suppose that it is either not expressed, or mislocalized due to the absence of the SP and immediately degraded.
To investigate the intracellular localization of the different isoforms, we generated GFP fusion proteins. In HEK293 cells, FL-RAGE-GFP localizes to the plasma membrane, whereas esRAGE-GFP shows a distribution consistent with the ER-Golgi secretory pathway, as previously shown (11)
. Again, we were unable to detect RAGEN-term-GFP (data not shown).
Taken together, these results indicate that THP-1 cells express the known variants of RAGE, with FL-RAGE representing the major isoform.
THP-1 cells secrete a soluble form of RAGE
We then analyzed cell surface expression of RAGE in differentiated THP-1 cells and in HEK293 cells transfected with pcDNA/FL-RAGE, using the -hRAGE N-terminal antibody in flow cytometry. As shown in Fig. 2A
, only a small fraction of THP-1 cells express RAGE on the surface (arrow), whereas most transfected HEK293 cells stain positive. This was a surprising result, considering that FL-RAGE transcript is present in THP-1 cells. To clarify this point, we compared RAGE expression in culture supernatants and lysates of undifferentiated and differentiated THP-1 cells (Fig. 2B
). Western blots of concentrated THP-1 supernatants probed with the -hRAGE N-terminal antibody reveal a prominent doublet of bands at 45 kDa. The same doublet is, however, not recognized by the -RAGE C-terminal antibody, suggesting that it may correspond to soluble RAGE forms that lack the cytoplasmic domain. In the cell lysates, the same two antibodies detect three major bands whose apparent molecular mass does not correspond to any of the recombinant RAGE isoforms (asterisks in Fig. 2B
, compare to Fig. 1B
), and therefore represent aspecific cross reactions. Similar results were obtained by lysing cells immediately in SDS-PAGE loading buffer (data not shown), and therefore the absence of detectable FL-RAGE cannot be attributed to proteolytic degradation of the sample. We conclude that although soluble RAGE can be detected in cell culture supernatant, FL-RAGE is present at extremely low levels in THP-1 cells.
|
RAGE has two N-glycosylation sites in its V-type domain (11)
. As shown in Fig. 2C
, after incubation of the supernatant with glycopeptidase F, which specifically removes sugar chains attached to asparagine residues, only a single band of a lower apparent molecular mass could be detected, indicating that soluble RAGE is glycosylated. The bands recognized by the various anti-RAGE antibodies in cell lysates do not shift after treatment with glycopeptidase F, in keeping with our conclusion that they do not represent bona fide RAGE.
We then tested whether soluble RAGE corresponds to esRAGE, which is poorly represented at the mRNA level. As shown in Fig. 2D
, an antibody against the unique C-terminal sequence of esRAGE (-esRAGE) (11)
did not recognize the soluble RAGE in the concentrated supernatant of THP-1 cells. It also failed to recognize any band in the cellular lysates. In contrast, it did recognize the expected band in a cell lysate of HeLa cells transfected with pcDNA/esRAGE. Moreover, the molecular mass of esRAGE appears to be greater than that of soluble RAGE in THP-1 supernatants.
These data indicate that the soluble RAGE produced by THP-1 cells is not the alternatively spliced esRAGE isoform; moreover, very little FL-RAGE protein is detected on the cell surface, although FL-RAGE is the most abundant form at the transcript level. Taken together, these results suggest that FL-RAGE is expressed in THP-1 cells but is rapidly processed to the soluble form.
FL-RAGE is proteolytically processed
Soluble receptors can be a product of alternative splicing, or shedding by specific proteases. To investigate whether FL-RAGE is subject to shedding, we transfected HeLa cells with pcDNA/FL-RAGE or the empty vector. Supernatants of HeLa/FL-RAGE cells, but not of HeLa/pcDNA cells, contained a band at 45 kDa that was recognized by the -hRAGE N-terminal antibody but not the -RAGE C-terminal antibody and had a lower molecular mass compared to FL-RAGE in the corresponding cell lysate (Fig. 3A
). The β-actin staining was used as a loading control. Similar results were obtained by transfecting HEK293 and 3T3 cells (data not shown), indicating that soluble RAGE is a processed product derived from FL-RAGE.
|
Definitive proof of a cleavage process can be provided by the identification of all cleavage products. We therefore set out to identify the C-terminal cleavage product of FL-RAGE. To this end, we transfected HeLa cells with pcDNA, pcDNA/FL-RAGE, or pcDNA/FL-RAGE-myc, coding for FL-RAGE with a Myc tag at the C terminus. Both cell lysates and supernatants were then analyzed by Western blot for the presence of RAGE using either -hRAGE N-term or -RAGE C-term antibodies (Fig. 3B
). Soluble RAGE was released in the supernatant by cells expressing FL-RAGE-myc, as well as FL-RAGE. The -RAGE C-terminal antibody recognized in the cell lysates both FL-RAGE and FL-RAGE-myc. It also recognized the untagged C-terminal fragment (CTF) of RAGE, of 14 kDa, and the tagged C-terminal fragment (CTF-myc) of RAGE-myc (Fig. 3B
). The myc-tagged CTF was also recognized by a monoclonal anti-myc antibody. Treatment of cells with 1 or 10 µM of the proteasome inhibitor MG-132 increased the amount of the C-terminal fragments, suggesting that they are proteolytically degraded after cleavage. Similar results were obtained with HEK293 cells expressing FL-RAGE (data not shown).
Treatment of cells with PMA can enhance shedding of a variety of membrane receptors by modulating the activity of ADAMs through the activation of protein kinase C (27)
. This is defined as "inducible shedding" and is highly conserved among a variety of cell lines. Thus, both HeLa (Fig. 3C
) and HEK293 cells (data not shown) transfected with pcDNA/FL-RAGE were treated with increasing concentrations of PMA, and RAGE expression was analyzed by Western blot. Soluble RAGE increased in a dose-dependent manner.
Likewise, RAGE shedding was promoted by the addition of HMGB1, one of the RAGE ligands (Fig. 3D
).
Altogether, these results suggest that membrane-bound FL-RAGE is subjected to constitutive, PMA-inducible, and ligand-promoted proteolytic cleavage of the ectodomain; the resulting product is a soluble form of RAGE, which we denote as cRAGE.
The question then arises of whether cRAGE and esRAGE are functionally equivalent. The alternative spliced form esRAGE has been described as a decoy receptor, blocking the activation of membrane-bound full-length RAGE by ligands. We then set up a chemotaxis assay, where cells migrate through a membrane toward a chamber containing HMGB1 (Fig. 3E
). As previously reported (28)
, 3T3 cells migrate toward HMGB1, but such migration is completely abrogated in the presence of excess esRAGE or cRAGE. This indicates that both soluble forms of RAGE interfere in a similar way with ligands.
ADAM10 is responsible for the shedding of FL-RAGE
Next, we wanted to identify the proteases responsible for FL-RAGE cleavage. HeLa cells, transfected with pcDNA/FL-RAGE and incubated with or without 25 nM PMA, were treated with two general metalloproteinase inhibitors, GM6001 and TAPI-1. Both inhibitors significantly reduced constitutive and PMA-induced shedding of FL-RAGE (Fig. 4A
). Similar results were obtained using HEK293 cells (data not shown).
|
To test the potential involvement of the ADAM family of metalloproteinases, we then used two hydroxamate-based inhibitors: GI254023X, which preferentially blocks ADAM10, and GW280623X, which blocks both ADAM10 and ADAM17 (23)
. Both inhibitors reduced FL-RAGE shedding (Fig. 4B
), suggesting that ADAM10 is responsible for RAGE shedding. Similar results were obtained using HEK293 cells (data not shown).
Finally, we compared a panel of ADAM-deficient MEFs for the ability to process FL-RAGE. To improve the detection of cRAGE and FL-RAGE, cells were transfected with pcDNA/FL-RAGE. As shown in Fig. 5A
, wild-type, Adam17–/–, and Adam9–/– MEFs produced cRAGE efficiently from the overexpressed FL-RAGE. In contrast, three different clones of ADAM10-deficient MEFs produced very little cRAGE.
|
We then stably transfected Adam10–/– clone 36 MEFs with a plasmid coding for mouse ADAM10, obtaining a pool of cells that we called A10. Adam10–/– MEFs express a small amount of Adam10 transcripts (9% compared to wild type, as assayed by real-time RT-PCR), which code for a truncated nonfunctional ADAM10 protein (29)
. The amount of Adam10 transcripts in A10 cells (at least some of them coding for functional ADAM10 protein) was 16% of wild-type; A10 cells regained the ability to cleave a fraction of overexpressed FL-RAGE constitutively (Fig. 5B
).
These data clearly indicate that ADAM10 is involved in the constitutive shedding of FL-RAGE; one or more additional proteases may be responsible for the cleavage a small fraction of FL-RAGE molecules.
Lung tissue expresses membrane-bound FL-RAGE and cleaved soluble RAGE
We investigated whether RAGE shedding also occurs in vivo. RAGE is highly expressed in lung, and is extremely low in other human tissues (5)
; the same is true in mouse tissues (data not shown). We then analyzed lung tissue from wild-type or RAGE-deficient (Ager–/–) mice (Fig. 6A
). Lung lysate from wild-type mice contains three major isoforms of RAGE; in particular, the one at 45 kDa is not detected by the -RAGE C-terminal antibody. As expected, the Ager–/– lung lysate shows no reactive bands. The membrane was reprobed with anti-β-actin antibody to ensure equal protein loading, and with anti-GFP antibody to identify GFP-positive RAGE-null mice (7)
. The -RAGE C-terminal antibody reveals a band of 14 kDa that is not recognized by a goat polyclonal antibody against the extracellular part of mouse RAGE (-RAGE N-term) (Fig. 6A
), suggesting that it represents the CTF.
|
FL-RAGE, cRAGE, esRAGE, and the C-terminal fragment could also be detected in lung lysates from two human autoptic samples (Fig. 6B
, C and results not shown). The esRAGE isoform is poorly represented (Fig. 6B
), whereas cRAGE and the C-terminal fragment are abundant.
Thus, these results confirm that lung tissue expresses high amounts of RAGE proteins and that FL-RAGE is constitutively shed in vivo.
RAGE expression in human PBMC and plasma
We then investigated the presence of surface RAGE on freshly purified PBMCs. FACS analysis indicated that PBMCs express a low but detectable level of surface RAGE (Fig. 6D
), which increases after incubation of PBMCs with protease inhibitors GM6001 and TAPI-1. This is in agreement with our finding that RAGE is proteolytically shed.
We also quantified simultaneously total soluble RAGE and esRAGE in the plasma of healthy human subjects (Table 1
). We found that, as reported by others (30)
, concentrations of total soluble RAGE were 1600 pg/ml and those of esRAGE were 5 times lower (the values are significantly different, P<0.0001 for paired 2-tailed t test). This indicates that esRAGE represents only a small part of the total circulating soluble RAGE.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. When we analyzed the different spliced transcripts from the Ager gene in human THP-1 cells, and the proteins expressed by these cells, we realized that esRAGE transcripts are rare and cannot account for the soluble RAGE in cell supernatants, which is not recognized by esRAGE-specific antibodies. Furthermore, soluble RAGE is found in the supernatants of HeLa and HEK293 cells that were forced to express only the full-length RAGE cDNA. These results indicate that soluble RAGE derives from full-length membrane-bound RAGE, by proteolytic processing. We also detected the CTF that remains in the membrane after cleavage of FL-RAGE. The amino acid sequence of soluble RAGE purified from mouse lung is also compatible with proteolytic processing of full-length RAGE (17)
and is incompatible with the sequence of mouse esRAGE (31)
. However, both esRAGE and cRAGE act in a very similar way as decoy receptors that interfere with HMGB1-induced cell migration. RAGE shedding is enhanced by binding of its ligand HMGB1, or by PMA treatment of the cells.
We then showed that cleavage of FL-RAGE can be inhibited with metalloprotease inhibitors and that cells deficient in Adam10 (Adam10–/–) produce very little soluble RAGE. In contrast, cells deficient in Adam17 or 9, or presenilin 1 and 2, can process FL-RAGE. Further analysis in Adam10–/– mice is precluded because the mutation is embryonically lethal (29)
.
Surprisingly, we found that RAGE is undetectable on the surface of THP-1 cells, which nonetheless respond to the RAGE ligand HMGB1 (25)
. Indeed, minimal amounts of RAGE transcripts are detected by quantitative PCR in human tissues, except lung (5)
. Furthermore, no specific band for RAGE can be found in Western blots of either Ager+/+ or Ager–/– MEFs, or in Adam10–/– MEFs (data not shown). Freshly isolated human PBMCs do express a low, but detectable, level of surface RAGE, and this level increases after incubation of PBMCs with general metalloproteinase inhibitors. Taken together, these data imply that the constitutive expression of RAGE in most cells is low and that proteolytic processing reduces it even further.
Soluble RAGE can be detected in the blood of healthy human subjects, and esRAGE represents only a minor fraction of total soluble RAGE. Because FL-RAGE is expressed at high levels only in the lung, we suggest that most circulating soluble RAGE derives from shedding in the lung, although a part can derive from PBMCs. This is in agreement with a recent study that shows that patients suffering from acute respiratory distress syndrome (ARDS) have a higher level of circulating RAGE (32)
. Some clinical studies have shown that lower plasma levels of soluble RAGE are associated with high risk of coronary artery disease, hypertension, arthritis and Alzheimer’s disease (reviewed in ref. 33
), and another study indicates that increased soluble RAGE levels are associated with increased microvascular damage (34)
. Two alternative but nonexclusive explanations can be postulated: shedding may be modulated by binding of proinflammatory ligands, or, alternatively, low levels of cRAGE are not sufficient as a decoy for proinflammatory ligands and therefore do not protect against chronic inflammation. The latter explanation is more readily reconcilable with the proposed lung origin of soluble RAGE, but a deeper understanding of RAGE expression, and shedding is needed if soluble RAGE is to be used as a diagnostic and prognostic marker.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication March 6, 2008. Accepted for publication June 12, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B require the cytoplasmic domain of the receptor but different downstream signaling pathways. J. Biol. Chem. 274,19919-19924B activation. J. Cell Biol. 19,33-40This article has been cited by other articles:
|
|
 |
|
  S. F. Yan, R. Ramasamy, and A. M. Schmidt The RAGE Axis: A Fundamental Mechanism Signaling Danger to the Vulnerable Vasculature Circ. Res., March 19, 2010; 106(5): 842 - 853. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. H. J. Gaens, I. Ferreira, C. J. H. van der Kallen, M. M. J. van Greevenbroek, E. E. Blaak, E. J. M. Feskens, J. M. Dekker, G. Nijpels, R. J. Heine, L. M. 't Hart, et al. Association of Polymorphism in the Receptor for Advanced Glycation End Products (RAGE) Gene with Circulating RAGE Levels J. Clin. Endocrinol. Metab., December 1, 2009; 94(12): 5174 - 5180. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-L. Xu, F. Vetri, H.-K. Lee, S. Ye, C. Paisansathan, L. Mao, F. Tan, and D. A. Pelligrino Estrogen replacement therapy in diabetic ovariectomized female rats potentiates postischemic leukocyte adhesion in cerebral venules via a RAGE-related process Am J Physiol Heart Circ Physiol, December 1, 2009; 297(6): H2059 - H2067. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Maillard-Lefebvre, E. Boulanger, M. Daroux, C. Gaxatte, B. I. Hudson, and M. Lambert Soluble receptor for advanced glycation end products: a new biomarker in diagnosis and prognosis of chronic inflammatory diseases Rheumatology, October 1, 2009; 48(10): 1190 - 1196. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Ramasamy, S. F. Yan, and A. M. Schmidt RAGE: therapeutic target and biomarker of the inflammatory response--the evidence mounts J. Leukoc. Biol., September 1, 2009; 86(3): 505 - 512. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Buhimschi, M. A. Baumbusch, A. T. Dulay, E. A. Oliver, S. Lee, G. Zhao, V. Bhandari, R. A. Ehrenkranz, C. P. Weiner, J. A. Madri, et al. Characterization of RAGE, HMGB1, and S100{beta} in Inflammation-Induced Preterm Birth and Fetal Tissue Injury Am. J. Pathol., September 1, 2009; 175(3): 958 - 975. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Selvais, H. P. Gaide Chevronnay, P. Lemoine, S. Dedieu, P. Henriet, P. J. Courtoy, E. Marbaix, and H. Emonard Metalloproteinase-Dependent Shedding of Low-Density Lipoprotein Receptor-Related Protein-1 Ectodomain Decreases Endocytic Clearance of Endometrial Matrix Metalloproteinase-2 and -9 at Menstruation Endocrinology, August 1, 2009; 150(8): 3792 - 3799. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Su, M. R. Looney, N. Gupta, and M. A. Matthay Receptor for advanced glycation end-products (RAGE) is an indicator of direct lung injury in models of experimental lung injury Am J Physiol Lung Cell Mol Physiol, July 1, 2009; 297(1): L1 - L5. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Z. Kalea, N. Reiniger, H. Yang, M. Arriero, A. M. Schmidt, and B. I. Hudson Alternative splicing of the murine receptor for advanced glycation end-products (RAGE) gene FASEB J, June 1, 2009; 23(6): 1766 - 1774. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. B Lindsey, F. Cipollone, S. M Abdullah, and D. K Mcguire Receptor for advanced glycation end-products (RAGE) and soluble RAGE (sRAGE): cardiovascular implications Diabetes and Vascular Disease Research, January 1, 2009; 6(1): 7 - 14. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Zhang, M. Bukulin, E. Kojro, A. Roth, V. V. Metz, F. Fahrenholz, P. P. Nawroth, A. Bierhaus, and R. Postina Receptor for Advanced Glycation End Products Is Subjected to Protein Ectodomain Shedding by Metalloproteinases J. Biol. Chem., December 19, 2008; 283(51): 35507 - 35516. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
