Rhinitis, sinusitis, and upper airway disease RAGE processing in chronic airway conditions: Involvement of Staphylococcus aureus and ECP Koen Van Crombruggen, PhD, Gabriele Holtappels, Natalie De Ruyck, Lara Derycke, PhD, Peter Tomassen, MD, and Claus Bachert, MD, PhD Ghent, Belgium Background: The receptor for advanced glycation end products (RAGE) is a multiligand receptor that exists as a membrane- bound (mRAGE) form and a soluble (sRAGE) form. RAGE is reported to play a role in diverse pathologies including lower airway conditions, but the exact mechanism of action remains poorly understood. In the upper airways, the involvement of RAGE remains completely unexplored. Objective: To investigate the involvement of RAGE in the human upper airway conditions chronic rhinosinusitis without nasal polyps (CRSsNP) and chronic rhinosinusitis with nasal polyps (CRSwNP). Methods: Protein levels of sRAGE, mRAGE, IL-5, and eosinophil cationic protein (ECP) were quantitatively assessed in inflamed tissue of CRSsNP and CRSwNP patients. Nasal tissue of subjects without disease served as control. Ex vivo human sinonasal tissue stimulation assays were used to assess the effect of Staphylococcus aureus and ECP on sRAGE processing. Results: sRAGE protein levels were higher in CRSsNP tissue, whereas mRAGE protein levels were lower than in controls. In CRSwNP patients, both tissue sRAGE and mRAGE protein levels were reduced. Low tissue sRAGE protein concentrations were associated with high IL-5 and ECP protein levels. In vitro, S aureus induced the release of sRAGE from the tissue, while ECP was shown to be implicated in the breakdown of free sRAGE. Conclusions: We demonstrate for the first time that RAGE protein is highly expressed in human upper airways under normal physiology and that it is subject to differential processing in CRSsNP and CRSwNP, identifying S aureus and ECP as novel and crucial players in this process. (J Allergy Clin Immunol 2012;129:1515-21.) Key words: Receptor for advanced glycation end products, eosino- phil cationic protein, Staphylococcus aureus, airway inflammation The receptor for advanced glycation end products (RAGE) is a cell surface protein that belongs to the immunoglobulin super- family that is implicated in diverse inflammatory responses. RAGE binds to multiple ligands, including advanced glycation end products, high-mobility group box 1, members of the S100/ calgranulin family, the integrin Mac-1, and extracellular matrix (ECM) structures such as heparansulphate proteoglycans. 1 RAGE is expressed as a full-length, membrane-bound receptor (mem- brane-bound RAGE [mRAGE]), but it can also exist in 2 soluble forms lacking the transmembrane and cytoplasmic domains col- lectively termed ‘‘soluble RAGE [sRAGE].’’ sRAGE derived via alternative mRNA splicing is known as ‘‘endogenous secretory RAGE’’ or esRAGE, whereas sRAGE derived from the proteolytic cleavage of mRAGE is referred to as cleaved RAGE [cRAGE], the latter being mediated by metalloproteinases (MMPs) and ‘‘A dis- integrin and metalloproteinases (ADAMs).’’ 2-5 sRAGE is often regarded to act as a ‘‘decoy receptor’’ with anti- inflammatory properties by scavenging away RAGE ligands from the cell surface mRAGE receptor that transduces proinflamma- tory responses via the transcription factor nuclear factor-kB, 6 but sRAGE is also described to be implicated in the development of inflammation 7,8 while anti-inflammatory properties have been suggested for mRAGE as well. 9 The receptor is reported to be im- plicated in a diverse set of pathologic conditions such as diabetes, autoimmune/inflammatory conditions, 10 and tumor cell metasta- sis, 11 While RAGE expression is undetectable on most cell types and tissues —except under pathologic conditions— 6 it is known that the lung forms a remarkable exception with high expression levels of RAGE during normal physiology, 9,12-14 suggesting a po- tentially important role for the receptor in maintaining lung ho- meostasis. Indeed, RAGE knockout mice spontaneously develop features of lung fibrosis during aging and showed in- creased fibrosis in response to asbestos injury. 9 However, RAGE knockout mice were also described to be protected against From the Upper Airway Research Laboratory, Department of Otorhinolaryngology, Ghent University Hospital. This work was supported by funding from the Research Foundation Flanders (FWO; research project no. G.0641.10), and a Concerted Research Action project grant (grant no. 01G01009) from the Special Research Fund of Ghent University. Disclosure of potential conflict of interest: P. Tomassen received research support from the Global Allergy and Asthma Research Network and Ghent University. The rest of the authors declare that they have no relevant conflicts of interest. Received for publication October 6, 2011; revised February 11, 2012; accepted for pub- lication February 20, 2012. Available online March 27, 2012. Corresponding author: Koen Van Crombruggen, PhD, Upper Airway Research Labora- tory, Department of Otorhinolaryngology, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium. E-mail: [email protected]. 0091-6749/$36.00 Ó 2012 American Academy of Allergy, Asthma & Immunology doi:10.1016/j.jaci.2012.02.021 Abbreviations used ADAMs: A disintegrin and metalloproteinases COPD: Chronic obstructive pulmonary disease cRAGE: Cleaved RAGE CRSsNP: Chronic rhinosinusitis without nasal polyps CRSwNP: Chronic rhinosinusitis with nasal polyps ECM: Extracellular matrix ECP: Eosinophil cationic protein esRAGE: Endogenous secretory RAGE MMPs: Metalloproteinases mRAGE: Membrane-bound RAGE RAGE: Receptor for advanced glycation end products sRAGE: Soluble RAGE 1515
Transcript
Rhinitis, sinusitis, and upper airway disease
RAGE processing in chronic airway conditions: Involvementof Staphylococcus aureus and ECP
Koen Van Crombruggen, PhD, Gabriele Holtappels, Natalie De Ruyck, Lara Derycke, PhD, Peter Tomassen, MD, and
Claus Bachert, MD, PhD Ghent, Belgium
Abbreviations used
ADAMs: A disintegrin and metalloproteinases
COPD: Chronic obstructive pulmonary disease
cRAGE: Cleaved RAGE
CRSsNP: Chronic rhinosinusitis without nasal polyps
CRSwNP: Chronic rhinosinusitis with nasal polyps
ECM: Extracellular matrix
ECP: Eosinophil cationic protein
esRAGE: Endogenous secretory RAGE
MMPs: Metalloproteinases
mRAGE: Membrane-bound RAGE
RAGE: Receptor for advanced glycation end products
sRAGE: Soluble RAGE
Background: The receptor for advanced glycation end products(RAGE) is a multiligand receptor that exists as a membrane-bound (mRAGE) form and a soluble (sRAGE) form. RAGE isreported to play a role in diverse pathologies including lowerairway conditions, but the exact mechanism of action remainspoorly understood. In the upper airways, the involvement ofRAGE remains completely unexplored.Objective: To investigate the involvement of RAGE in thehuman upper airway conditions chronic rhinosinusitis withoutnasal polyps (CRSsNP) and chronic rhinosinusitis with nasalpolyps (CRSwNP).Methods: Protein levels of sRAGE, mRAGE, IL-5, andeosinophil cationic protein (ECP) were quantitatively assessedin inflamed tissue of CRSsNP and CRSwNP patients. Nasaltissue of subjects without disease served as control. Ex vivohuman sinonasal tissue stimulation assays were used to assessthe effect of Staphylococcus aureus and ECP on sRAGEprocessing.Results: sRAGE protein levels were higher in CRSsNP tissue,whereas mRAGE protein levels were lower than in controls. InCRSwNPpatients, both tissue sRAGEandmRAGEprotein levelswere reduced. Low tissue sRAGE protein concentrations wereassociated with high IL-5 and ECP protein levels. In vitro, Saureus induced the release of sRAGE from the tissue, while ECPwas shown to be implicated in the breakdown of free sRAGE.Conclusions: We demonstrate for the first time that RAGEprotein is highly expressed in human upper airways undernormal physiology and that it is subject to differentialprocessing in CRSsNP and CRSwNP, identifying S aureus andECP as novel and crucial players in this process. (J Allergy ClinImmunol 2012;129:1515-21.)
Key words: Receptor for advanced glycation end products, eosino-phil cationic protein, Staphylococcus aureus, airway inflammation
From the Upper Airway Research Laboratory, Department of Otorhinolaryngology,
Ghent University Hospital.
This work was supported by funding from the Research Foundation Flanders (FWO;
research project no. G.0641.10), and a Concerted Research Action project grant (grant
no. 01G01009) from the Special Research Fund of Ghent University.
Disclosure of potential conflict of interest: P. Tomassen received research support from
the Global Allergy and Asthma Research Network and Ghent University. The rest of
the authors declare that they have no relevant conflicts of interest.
Received for publication October 6, 2011; revised February 11, 2012; accepted for pub-
lication February 20, 2012.
Available online March 27, 2012.
Corresponding author: Koen Van Crombruggen, PhD, Upper Airway Research Labora-
tory, Department of Otorhinolaryngology, Ghent University Hospital, De Pintelaan
� 2012 American Academy of Allergy, Asthma & Immunology
doi:10.1016/j.jaci.2012.02.021
The receptor for advanced glycation end products (RAGE) is acell surface protein that belongs to the immunoglobulin super-family that is implicated in diverse inflammatory responses.RAGE binds to multiple ligands, including advanced glycationend products, high-mobility group box 1, members of the S100/calgranulin family, the integrin Mac-1, and extracellular matrix(ECM) structures such as heparansulphate proteoglycans.1 RAGEis expressed as a full-length, membrane-bound receptor (mem-brane-bound RAGE [mRAGE]), but it can also exist in 2 solubleforms lacking the transmembrane and cytoplasmic domains col-lectively termed ‘‘soluble RAGE [sRAGE].’’ sRAGE derived viaalternative mRNA splicing is known as ‘‘endogenous secretoryRAGE’’ or esRAGE,whereas sRAGEderived from the proteolyticcleavage ofmRAGE is referred to as cleaved RAGE [cRAGE], thelatter being mediated by metalloproteinases (MMPs) and ‘‘A dis-integrin and metalloproteinases (ADAMs).’’2-5
sRAGE is often regarded to act as a ‘‘decoy receptor’’ with anti-inflammatory properties by scavenging away RAGE ligands fromthe cell surface mRAGE receptor that transduces proinflamma-tory responses via the transcription factor nuclear factor-kB,6 butsRAGE is also described to be implicated in the development ofinflammation7,8 while anti-inflammatory properties have beensuggested for mRAGE as well.9 The receptor is reported to be im-plicated in a diverse set of pathologic conditions such as diabetes,autoimmune/inflammatory conditions,10 and tumor cell metasta-sis,11 While RAGE expression is undetectable on most cell typesand tissues —except under pathologic conditions—6 it is knownthat the lung forms a remarkable exception with high expressionlevels of RAGE during normal physiology,9,12-14 suggesting a po-tentially important role for the receptor in maintaining lung ho-meostasis. Indeed, RAGE knockout mice spontaneouslydevelop features of lung fibrosis during aging and showed in-creased fibrosis in response to asbestos injury.9 However,RAGE knockout mice were also described to be protected against
AERD, Aspirin-exacerbated respiratory disease; NA, not applicable.
J ALLERGY CLIN IMMUNOL
JUNE 2012
1516 VAN CROMBRUGGEN ET AL
bleomycin-induced pulmonary fibrosis,14,15 while RAGE does notseem to be implicated in the fibrotic process of silica-induced pul-monary fibrosis.16 A similar ambiguity can be found in the litera-ture on human lower airwayswhere on the one hand an elevation inthe overall levels of RAGE was reported in the lung tissue of pa-tients with chronic obstructive pulmonary disease (COPD),17
while decreased levels of mRAGE and cRAGE but not sRAGEwere observed in another study assessing lung tissues of patientswith COPD.18 Interestingly, Sukkar et al19 reported reduced levelsof lung sRAGE in neutrophilic asthma and COPD while levels ofsRAGE in asthma andCOPDwithout neutrophilia were not differ-ent from control levels. These data indicate that RAGE has an im-portant function in the airway, but its effect differs depending onthe experimental model, the specific RAGEvariants, or the inflam-matory environment under investigation.In the human upper airways, the involvement of RAGE remains
completely unexplored. The aim of this study was therefore toevaluate the implication of the specific variants of RAGE in thehuman upper airway conditions chronic rhinosinusitis withoutnasal polyps (CRSsNP) and chronic rhinosinusitis with nasalpolyps (CRSwNP), allowing us to concurrently assess the possi-ble influence of the different inflammatory environments distinc-tive for CRSsNP and CRSwNP. Chronic rhinosinusitis is indeed aheterogeneous group of common chronic airway diseases, withCRSwNP being characterized to exhibit a strong TH2-skewed eo-sinophilic inflammatory environment with high IL-5 and eosino-phil cationic protein (ECP) concentrations in the polyps.20,21
CRSwNP patients also have an increased colonization rate ofStaphylococcus aureus in the middle meatus compared with con-trols and CRSsNP patients22 and often subsequently developlower respiratory tract conditions, especially asthma.23,24 Onthe other hand, CRSsNP is characterized by a more predominantTH1 milieu with pronounced levels of IFN-g in the inflamed eth-moidal mucosa and clearly lower levels of IL-5 and ECP com-pared with those in CRSwNP.21 MMP levels are imbalanced inboth CRSsNP and CRSwNP patients.25,26
The present report shows for the first time that RAGE isexpressed in the human upper airway under normal physiologyand that it is subject to differential regulation in CRSsNP andCRSwNP. It furthermore provides evidence for S aureus and ECPto be directly implicated in RAGE expression.
METHODS
PatientsThe patients’ clinical data can be found in Table I. More detailed informa-
tion can be found in this article’s Online Repository at www.jacionline.org.
Tissue homogenates and protein extractionIn order to prepare soluble protein fractions, frozen tissue samples were
homogenized by means of mechanical disruption as described in this article’s
Online Repository at www.jacionline.org.
Membrane protein fractions were made via a commercially available kit
(ProteoExtract Native Membrane Protein Extraction; CalBiochem/EMD
Biosciences, San Diego, Calif) according to the manufacturers’ guidelines.
Human sinonasal ex vivo tissue-cube fragment
stimulation assayHuman sinonasal tissues were processed as described before.27 Individual
wells containing tissue-cube fragments of control inferior turbinate tissue re-
ceived 106 colony-forming units of S aureus (strain RN6390), S aureus 11 mg/mL ECP (n 5 7-8), heat-killed S aureus, Staphylococcus epidermidis,
Pseudomonas putida, or Escherichia coli (n 5 4) for 24 hours. In another
set of experiments, nasal polyp tissue-cube fragments were treated with 5
mg/mL recombinant sRAGE (n5 6) for 24 hours. See this article’s Online Re-
pository at www.jacionline.org for more detailed information.
sRAGE, mRAGE, esRAGE, ECP, cytokine, and MMP
measurementsThe proteins were measured by means of commercially available kits.
Assessment of the total sRAGE levels and of the mRAGE levels in the
ProteoExtract-derived membrane fraction involves an mAb against the
extracellular N-terminus present in both sRAGE and mRAGE, while the
detection of esRAGE involves a capture mAb recognizing the esRAGE-
specific C-terminal 16 amino acid sequence.More detailed information can be
found in this article’s Online Repository at www.jacionline.org.
RAGE immunohistochemistryParaffin sections of human sinonasal tissues were stained for RAGE with a
polyclonal antibody raised against the N-terminus of human RAGE as
reported in detail in this article’s Online Repository at www.jacionline.org.
limit. All tissue samples from CRSwNP patients were positive forIL-5.ECP protein concentrations were significantly higher in
CRSwNP patients than in controls and CRSsNP patients, thelatter’s ECP levels being significantly higher than the ECP levelsin controls.More data can be found in Table I and this article’s Online Re-
pository at www.jacionline.org.
control CRSsNP CRSwNP
0
1000
sR
AG
E tis
su
e (
1500 **
***
**
ue)
B
Tissue protein levels of MMP-3, MMP-7, MMP-9,
and ADAM-10MMP-3, MMP-7, and MMP-9 protein levels in CRSsNP tissue
were significantly higher than in control tissue. In CRSwNPtissue, the tissue proteins levels of MMP-3 and MMP-7 werehigher while MMP-9 protein levels were not significantly differ-ent from those in control tissue (see Fig E1, A-C, in this article’sOnline Repository at www.jacionline.org). ADAM-10 tissue pro-tein concentrations were not different between control tissue andCRSsNP tissue, while the levels in CRSwNP tissue had the ten-dency to be higher but not reaching significance compared withcontrol tissue (see Fig E1, D).
control CRSsNP CRSwNP
0
500
1000
mR
AG
E tis
su
e (p
g/g
tis
s
FIG 1. Quantification of tissue sRAGE (A) and mRAGE (B) protein levels by
means of ELISA in controls and CRSsNP and CRSwNP patients. Values are
expressed as pg/g tissue and presented as box-and-whisker plots showing
the minimum andmaximum values, the lower and upper quartiles, the me-
dian (as a line), and the mean (as a cross). **P < .01; ***P < .001.
Tissue protein levels of sRAGE, mRAGE, and
esRAGETotal sRAGE and full-length mRAGE were expressed under
control conditions (Fig 1, A and B). In CRSsNP tissue, the tissuelevels of the sRAGE protein fraction were significantly higherthan in control tissue (Fig 1, A), while conversely the mRAGElevels were significantly lower than those in control tissue (Fig1, B). In CRSwNP tissue, tissue levels of both sRAGE andmRAGE were significantly lower versus the levels measured incontrol tissue and in CRSsNP tissue (Fig 1, A and B).The OD levels of the alternatively spliced soluble fraction of
RAGE (esRAGE) protein were below the OD of lowest standardpoint (25 pg/mL) for the tissue homogenates of all subject groups(data not shown).RAGE immunoreactivity was clearly detectable on epithelial
cells and ECM structures in sinonasal tissue of controls andCRSsNP (Fig 2), being more pronounced and extending deeperinto the ECM structures of CRSsNP tissue compared with controltissue. Conversely, RAGE expression was lowest in CRSwNP tis-sue (Fig 2).
Gene expression for mRAGE and esRAGEThe relative levels in gene expression for mRAGE were not
different between the 3 patient groups (see Fig E2, A, in this arti-cle’s Online Repository), while esRAGEwasmodestly but signif-icantly higher in CRSsNP and CRSwNP versus control tissue(Fig E2, B).
mRAGE tissue levels in relation to MMP-9 and
ADAM-10mRAGE levels in the tissue of controls were inversely corre-
lated with the level of MMP-9 (Spearman rs 5 20.8200; P 5.002; see Fig E3, A, in this article’s Online Repository at www.jacionline.org) and the level of ADAM-10 (Spearman rs 520.6818; P 5 .0208; Fig E3, B). No correlation was observed
forMMP-9 and ADAM-10withmRAGE tissue levels in CRSsNPand CRSwNP (data not shown).
S aureus induces the release of tissue sRAGE
ex vivoUnder baseline conditions, sRAGE was not detectable in the
tissue-cube supernatants of our human sinonasal ex vivo tissue-cube fragment stimulation assay. In the presence of S aureus,there was a clear-cut increase in sRAGE levels in the tissue-cube supernatants while the levels within the tissue-cube pelletswere significantly decreased (Fig 3). Unlike for sRAGE, S aureusdid not affect the levels of another ECM-associated protein TGF-b in the tissue-cube supernatants (469.7 6 88.4 pg/mL vs 387.26 46.8 pg/mL; n 5 6) and tissue-cube pellets (1686.7 6 217.3pg/g tissue vs 1887.3 6 191.0 pg/g tissue; n 5 6).Similarly, as under baseline conditions, in the presence of heat-
killed S aureus, in the presence of hypotonic saline to induce celldeath, or in the presence of S epidermidis, P putida, and E coli—other frequent colonizers of the upper respiratory tract—sRAGEcould not be detected in the tissue-cube supernatants.
FIG 2. Immunolabeling with antibodies directed against RAGE in human sinonasal tissues from controls
(A), CRSsNP patients (B), and CRSwNP patients (C) (magnification 3200).
medium
alon
e
S. aure
us
medium
alon
e
S. aure
us0
100
200
300
400
0
200
400
600
800
***
*
supernatants tissue pellets
sR
AG
E(p
g/m
L)
sR
AG
E(p
g/g
tis
su
e)
FIG 3. Effect of S aureus on tissue sRAGE ex vivo (n 5 8): under baseline
conditions, sRAGE was not detectable in tissue-cube supernatants. The
presence of S aureus resulted in a clear-cut increase in the levels of endog-
enous sRAGE in the supernatants, while the levels of sRAGE in the tissue-
cube pellets were reduced by S aureus. *P < .05; ***P < .001.
J ALLERGY CLIN IMMUNOL
JUNE 2012
1518 VAN CROMBRUGGEN ET AL
sRAGE in relation to IL-5 and ECP protein levelsTaking all patient groups together, IL-5–positive tissues (see
Table I) showed statistically significant lower tissue sRAGE pro-tein levels (906.0 pg/g tissue; range, 14.1-2263.8 pg/g tissue; P <.01) compared with IL-5–negative tissues (1349.1 pg/g tissue;range, 583.0- 2777.8).The differentiation between high and low levels of sRAGE
proteins based on IL-5–negative versus IL-5–positive tissues alsoaccounts when assessing separate disease entities (Fig 4, A). Ob-viously, as all control tissues were IL-5 negative and all CRSwNPtissues in this study were IL-5 positive, the latter showed signifi-cantly lower levels of sRAGE than did the former (Fig 4, A, andFig 1, A). More strikingly, within the CRSsNP patient group,the IL-5–positive samples also had statistically significant lowersRAGE levels compared with the IL-5–negative tissues of thesame patient group (Fig 4, A). When comparing the differentIL-5–negative groups, CRSsNP tissues had significantly higherlevels of sRAGE (Fig 4, A) compared with the IL-5–negative con-trols. Within the IL-5–positive samples, CRSwNP tissues showedsignificantly lower sRAGE levels (Fig 4, A) compared with theIL-5–positive CRSsNP tissues.As IL-5–positive tissues also showed the highest ECP levels20
(see Fig E4, A and B, in this article’s Online Repository at www.jacionline.org), the assessment of sRAGE based on the compari-son of tissues with low and high levels of ECP consequently also
resulted in differentiation between high and low sRAGE proteinconcentrations, respectively (Fig 4, B). In order to determinethe threshold defining ‘‘ECP-low’’ and ‘‘ECP-high’’ tissues, the10 to 90th percentile of the tissue ECP concentrations was deter-mined for IL-5–negative and IL-5–positive tissues. This yielded39 to 1881 ng/g tissue for the IL-5–negative tissues and 1,949to 17,886 ng/g tissue for the IL-5–postive tissues, allowing usto set the cutoff value between ‘‘ECP low’’ and ‘‘ECP high’’ at1,900 ng/g tissue.Recombinant ECP, but not IL-5, significantly increased the
breakdown of recombinant sRAGE in an in vitro experimentalsetup (Fig 5). After being released by S aureus from the tissue,the levels of endogenous sRAGE present in the culture mediumwere also reduced by exogenous ECP from 103.7 6 29.5 pg/mL to 21.1 6 7.2 pg/mL (n 5 7; P < .01), while tissue-boundsRAGE was not affected (537.0 6 182.5 pg/g tissue in the ab-sence vs 670.36 110.9 pg/g tissue in the presence of ECP; n5 4).
Effect of sRAGE on inflammatory cytokine release
ex vivoTreatment with recombinant sRAGE significantly increased
the release of IL-1b and TNF-a but decreased the levels of IL-5from human nasal polyp tissue-cube fragments (see Fig E5 in thisarticle’s Online Repository at www.jacionline.org).
DISCUSSIONUnlike other tissues in the body in which RAGE is expressed at
low levels and increases during inflammation,28 the lung is knownto express high basal levels of mRAGE and sRAGE.9,13 Here wereport for the first time that both forms of RAGE are also clearlyexpressed in the human upper airways under normal physiologyand that they are subject to differential processing on protein levelduring chronic upper airway conditions, identifying S aureus andECP as novel and crucial players in this process.Reduced levels of mRAGE protein in tissue homogenates from
CRSsNP and CRSwNP patients together with unaltered mRNAexpression levels versus controls allow us to speculate on anincreased proteolytic cleavage of the full-length mRAGE protein.Increased proteolytic activity will reduce the levels of mRAGE bycleaving of the receptor from the cell membrane, while concur-rently augmenting the levels of cleaved sRAGE. In the inflamedlung, decreased RAGE protein concentrations in tissue homog-enates9,16,29 are indeed often accompanied by increased sRAGElevels in the bronchoalveolar lavage fluid of experimental
FIG 4. sRAGE protein levels in relation to IL-5–negative versus IL-5–positive
tissues (A) or ECP-low versus ECP-high tissues (B). Values are presented as
box-and-whisker plots showing the minimum and maximum values, the
lower and upper quartiles, the median (as a line), and the mean (as a cross).*P < .05; **P < .01; ***P < .001.
0
1000
2000
3000
medium
alone
0.1
**
10
**
100
ECP (µg/mL)
1
IL-5 (ng/mL)
reco
mb
inan
t sR
AG
E (p
g/m
L)
FIG 5. ECP, but not IL-5, induced the breakdown of recombinant sRAGE
in vitro (n 5 5). **P < .01.
J ALLERGY CLIN IMMUNOL
VOLUME 129, NUMBER 6
VAN CROMBRUGGEN ET AL 1519
animals.29-31 Enhanced proteolytic cleavage per se can thus ex-plain the clear-cut increased levels of sRAGE in tissue homoge-nates from CRSsNP patients in this study. cRAGE constitutestogether with esRAGE the sRAGE pool. While the relativemRNA expression level for esRAGE in tissue was found to beslightly increased in CRSsNP and CRSwNP patients comparedwith controls in our study, esRAGE protein levels could not bemeasured in the tissues. It has indeed been described that esRAGEaccounts only for 7%of all RAGE isoforms in the lung.3 In human
sinonasal tissue, sRAGE thus also seems to be predominantlygenerated via the cleaving of mRAGE protein.Yamakawa et al5 described the release of sRAGE from alveolar
epithelial cells via proteolytic activity by MMP-3 and MMP-13.Similarly, MMP-94 and ADAM-102,4 have also been reported topromote the proteolytic cleavage of RAGE. We previously re-ported an imbalance between MMPs to be implicated in thechronic upper airway conditions CRSsNP and CRSwNP,25,26
which in part were corroborated in this study. The suggested func-tional implication ofMMPs to cleavemRAGEcould be confirmedin our study forMMP-9 and ADAM-10 in view of the inverse cor-relation between the tissue proteins levels of these MMPs and thetissue protein levels of full-length mRAGE in control subjects.The loss of correlation in tissues from CRSsNP and CRSwNPpatients confirms a deregulation of the normal steady-state condi-tion as a consequence of the inflammatory and remodelingmechanisms involved in chronic rhinosinusitis and furthermoresuggests the implication of other factors in this process.One factor that might contribute to the lower levels of tissue-
associated sRAGE in CRSwNP compared with CRSsNP andcontrols could be related to the fact that cRAGE, after proteolyticcleavage frommRAGE, can bind to heparan sulfate proteoglycanspresent on cellular (eg, epithelial syndecans and glypicans) andECM structures (perlecan, agrin, and collagen XVIII) and to otherECM components such as collagen types I and IV.13,32,33 Ourimmunohistochemistry data indeed show immunostaining on ep-ithelial and ECM structures of control and CRSsNP tissues, beingmore pronounced in the latter, while conversely in CRSwNP tis-sue, the ECM structures showed minimal RAGE staining. Thiscan be harmonized with the fact that CRSsNP and CRSwNP tis-sues are characterized by distinct remodeling patterns as was pre-viously shown by a lack of collagen in CRSwNP tissue andexcessive collagen production with thickening of the collagen fi-bers in the ECM of CRSsNP tissue,26,34 allowing sRAGE to be re-tained in higher levels in the fibrotic tissue of CRSsNP patientsthan in the pseudocystic tissue of CRSwNP patients.Besidesmorphologic changes as a potential cause of the altered
tissue levels of sRAGE, we here report the increased colonizationrate of S aureus in the upper airways of CRSwNP patients22 as apotential novel contributor to the overall reduction in sRAGE in
J ALLERGY CLIN IMMUNOL
JUNE 2012
1520 VAN CROMBRUGGEN ET AL
this patient group. Indeed, the results from our ex vivo human tis-sue experiments showed that S aureus was able to release freesRAGE into the tissue culturemediumwhereas sRAGE otherwisecould be detected in the tissue-cube pellet only as tissue-associated sRAGE. Alterations in the immune barrier functionpotentially compromise host defense and make the sinonasal mu-cosa more susceptible to antigenic exposition, leading to chronicinflammation. Here we speculate that colonization by S aureusmight affect the immune barrier funtion via the release oftissue-bound sRAGE. Bacterial pathogens activate ectodomainshedding of host cell surfacemolecules to enhance their virulence.Indeed, S aureus alpha and beta toxins are reported to induce theshedding of syndecan-1,35,36 a heparan sulfate proteoglycanhighly expressed in fibroblastic and epithelial cells. Since sRAGEhas high affinity to heparan sulfate, collagen, and other cellularand ECM-associated proteins,13,33 it is likely that damage to thosestructures will result in the release of the bound sRAGE. BecauseChavakis et al37 demonstrated that RAGE can directly regulateleukocyte recruitment by serving as a counterreceptor for b2-in-tegrinMac-1, sRAGE shedded and eliminated from lung and sino-nasal tissue might lose its local regulatory function that isnecessary for normal immunologic homeostasis of the airways.9
Generally, sRAGE is regarded to act as an extracellular‘‘decoy’’ receptor with anti-inflammatory properties by scaveng-ing away the RAGE ligands that would otherwise interact withcell surface mRAGE receptors resulting in the activation of theproinflammatory transcription factor nuclear factor-kB.6 Indeed,sRAGE has been suggested to reduce inflammatory responses inseveral experimental animal models28,38,39 including acute lunginjury.31 Because sRAGE treatment has also been shown to de-crease the delayed type hypersensitivity response in RAGEknockout mice, sRAGE has been assigned protective propertiesother than just blocking mRAGE function.38 This might be re-lated to the above-described feature of RAGE as a counterrecep-tor for Mac-1 in the regulation of leukocyte homeostasis.Although the above observations support an anti-inflammatoryrole for sRAGE, it has also been associated with the developmentof inflammation; exogenous sRAGE directly caused inflamma-tion by recruiting monocytes and neutrophils,7 sRAGE levels cor-relate with the severity of inflammation,40,41 and sRAGE inducesIL-6 and TNF-a production in splenocytes.8 Taken together, thesedata demonstrate that RAGE has both pro- and anti-inflammatoryproperties but because the RAGE pathway is controlled at tissue-specific level and is being differentially regulated depending onthe inflammatory environment of the disease under investigation,the exact direction of RAGE toward a pro- or anti-inflammatoryeffect is not straightforward for interpretation. For the human up-per airways, we postulate that while tissue-bound sRAGE couldbe implicated in the regulation of normal immunologic responsesby directly regulating leukocyte homeostasis, sRAGE sheddedfrom the tissue gains other activities; our results with exogenouslyapplied recombinant sRAGE to human nasal polyp tissues indeedshow an increase in levels of inflammatory mediators IL1-b andTNF-a and a decrease in IL-5 levels. These results confirm thatthe duality in proinflammatory versus anti-inflammatory effectsof sRAGE may also be present in the human upper airways. Ofimportance, Chen et al42 reported RAGE deficiency to decreasethe production of TH1 cytokines while producing more IL-4 andIL-5 as TH2 cytokines.The presence of free sRAGE in vivo is confirmed in experimen-
tal animal models as it has been shown to accumulate in the
bronchoalveolar lavage fluid of mice with acute lung injury.29-31
However, in the less acute mouse models of silica-induced fibro-sis, no sRAGE could be detected in the bronchoalveolar lavagefluid,16 leaving the authors to speculate on a clearance or break-down of the released sRAGE under specific experimental or in-flammatory conditions. Here we identify ECP to be a functionalmediator implicated in the breakdown and thus clearance ofRAGE. Our data show that a pronounced eosinophilic inflamma-tory milieu with high ECP and high IL-5 levels is associated withlow sRAGE levels in the patients’ sinonasal tissue. In vitro exper-iments showed ECP, but not IL-5, to markedly increase the deg-radation of sRAGE. This might explain why in the CRSsNPpatient group the imbalance in proteolytic activity by MMPsper se yields reduced mRAGE levels accompanied by increasedlevels of intact sRAGE, while it results in low levels of bothmRAGE and the subsequently generated sRAGE in the TH2-skewed CRSwNP patient group, namely, as a consequence ofECP-mediated breakdown of the released sRAGE. AlthoughCRSsNP shows an overall TH1-mediated type of response, eo-taxin and ECP can be measured in the tissues of these patients,which indeed have been shown to have a variable eosinophil infil-tration.21 Here we confirm moderate ECP levels in CRSsNP tis-sues and furthermore show that in a part of our CRSsNP tissuesIL-5 could be measured. The IL-5–positive CRSsNP samples,also being the samples with the highest ECP levels within thisclinical entity, showed again the lowest RAGE tissue levels.This oncemore confirms the implication of ECP in the breakdownof RAGE and indicates that the low RAGE levels in CRSwNP tis-sues are not solely a consequence of tissue remodeling.
ConclusionHere we report for the first time that RAGE is highly expressed
in the human upper airways under normal physiology and that it issubject to differential postsynthetic processing during chronicairway conditions, with S aureus and ECP as new identified me-diators implicated.The decrease in mRAGE protein during pathophysiology is
likely to be mediated via the cleavage of mRAGE by deregulatedMMP activity to form sRAGE that subsequently binds cellularand ECM components. Differences in the levels of sRAGEpotentially involve altered ECM composition and/or shedding oftissue-associated sRAGE via S aureus–mediated mechanisms.The released sRAGE potentially gains proinflammatory proper-ties that are countered via ECP-mediated sRAGE breakdown inan eosinophilic environment.
Key messages
d The pattern recognition receptor RAGE is for the firsttime reported to be highly expressed in human upper air-ways under normal physiology, suggesting a protectivefunction for RAGE in the immunology of the airways.
d sRAGE is released from the mucosal airway tissue by Saureus and degraded by ECP, the latter 2 being knownto be clearly involved in the pathophysiology of CRSwNP.
REFERENCES
1. Bopp C, Bierhaus A, Hofer S, Bouchon A, Nawroth PP, Martin E, et al. Bench-to-
bedside review: the inflammation-perpetuating pattern-recognition receptor RAGE
as a therapeutic target in sepsis. Crit Care 2008;12:201.
J ALLERGY CLIN IMMUNOL
VOLUME 129, NUMBER 6
VAN CROMBRUGGEN ET AL 1521
2. Raucci A, Cugusi S, Antonelli A, Barabino SM, Monti L, Bierhaus A, et al. A
soluble form of the receptor for advanced glycation endproducts (RAGE) is
produced by proteolytic cleavage of the membrane-bound form by the shed-
dase a disintegrin and metalloprotease 10 (ADAM10). FASEB J 2008;22:
3716-27.
3. Hudson BI, Carter AM, Harja E, Kalea AZ, Arriero M, Yang H, et al. Identifica-
tion, classification, and expression of RAGE gene splice variants. FASEB J 2008;
22:1572-80.
4. Zhang L, Bukulin M, Kojro E, Roth A, Metz VV, Fahrenholz F, et al. Receptor for
advanced glycation end products is subjected to protein ectodomain shedding by
39. Yan SS, Wu ZY, Zhang HP, Furtado G, Chen X, Yan SF, et al. Suppression of ex-
perimental autoimmune encephalomyelitis by selective blockade of encephalito-
genic T-cell infiltration of the central nervous system. Nat Med 2003;9:287-93.
40. Nakamura K, Yamagishi S, Adachi H, Kurita-Nakamura Y, Matsui T, Yoshida T,
et al. Serum levels of sRAGE, the soluble form of receptor for advanced glycation
end products, are associated with inflammatory markers in patients with type 2 di-
abetes. Mol Med 2007;13:185-9.
41. Bopp C, Hofer S, Weitz J, Bierhaus A, Nawroth PP, Martin E, et al. sRAGE is el-
evated in septic patients and associated with patients outcome. J Surg Res 2008;
147:79-83.
42. Chen Y, Akirav EM, Chen W, Henegariu O, Moser B, Desai D, et al. RAGE liga-
tion affects T cell activation and controls T cell differentiation. J Immunol 2008;
181:4272-8.
J ALLERGY CLIN IMMUNOL
JUNE 2012
1521.e1 VAN CROMBRUGGEN ET AL
METHODS
PatientsTissues were obtained from theDepartment of Otorhinolaryngology, Ghent
University Hospital, Belgium, during routine endonasal sinus surgery and
approved by the local ethical committee. All patients gave their written
informed consent before collecting material. All patients stopped oral and
topical application of corticosteroids for at least 1 month before surgery. The
diagnosis of chronic sinus disease was based on history, clinical examination,
nasal endoscopy, and computed tomography of the paranasal cavities
according to the current EuropeanE1,E2 and American Guidelines.E3 The pa-
tients’ clinical data can be found in Table I of the main document.
Tissue homogenates and protein extractionSnap-frozen tissue specimens were weighed and suspended in a 10 times
volume of 0.9% NaCl solution with Complete Protease Inhibitor Cocktail
(Roche, Mannheim, Germany). In order to prepare soluble protein fractions,
the frozen tissue was pulverized by means of a mechanical Tissue Lyser LT
(Qiagen, Hilden, Germany) at 50 oscillations per second for 2 minutes in
prechilled eppendorfs. The tissue homogenates were centrifuged at 15,000
rpm for 5 minutes at 48C.Membrane protein fractions were made via a commercially available kit
(ProteoExtract Native Membrane Protein Extraction; CalBiochem/EMD
Biosciences, San Diego, Calif) according to the manufacturers’ guidelines.
The supernatants from the tissue homogenates (containing the soluble
protein fraction) and from the ProteoExtract-derived membrane fraction were
stored at 2208C until further analysis.
Human sinonasal ex vivo tissue-cube fragment
stimulation assayFresh sinonasal tissue fragments of 60.9 mm3 were suspended as 0.04 g
tissue-cubes/mL in tissue culture medium into 48-well plates in a total volume
of 0.5 mL/well.
S aureus, S epidermidis, P putida, and E coli were grown in Luria Bertani
broth at 378C under aerobic conditions for 24 hours. Bacteria were harvested
in the stationary phase, and cell counts in the bacterial suspension were esti-
mated by OD at 620 nm absorbance (LabsystemsMultiskan RC). The bacteria
were centrifuged (10000 g for 5 minutes), the supernatants discarded, and the
bacterial pellet washed in PBS, centrifuged, and resuspended in Luria Bertani
broth before the bacteria were added to the tissue-cube fragments at a final
concentration of 106 colony-forming units/mL tissue culture medium. In order
to investigate heat-killed S aureus compared with viable bacteria, aliquots of S
aureus suspensions were incubated for 10 minutes in a bath at 958C. Aliquotsof all suspensionswere plated on blood-agar plates to confirm bacterial growth
and to exclude the presence of contaminant bacteria. All aliquots showed
growth of the corresponding bacterial strain except for the aliquots subjected
to heat treatment, which showed no bacterial growth.
In order to assess the effect of cell death on the release of sRAGE, in
1 condition tissue culture medium was replaced by hypotonic TRIS (10 mM)
either in the absence or in the presence of 2.5 mM Triton X-100.
The culture medium supernatants from the tissue-cube fragment stimula-
tion assays were snap-frozen in liquid nitrogen and stored at 2208C until
sRAGE or cytokine measurement (see further). The tissue-cube pellets were
snap-frozen and stored at 2208C until tissue-cube pellet homogenates were
prepared in a similar way as described for tissue homogenates (see above) in
order to measure sRAGE (see further).
sRAGE, mRAGE, esRAGE, ECP, cytokine, and MMP
measurementsTotal sRAGE in tissue homogenates and in culture medium supernatants
and tissue pellets from human sinonasal tissue-cube fragment stimulation
assays and mRAGE in ProteoExtract-derived membrane fraction were deter-
mined by using ELISA microplates that were coated with an mAb (Human
RAGE DuoSet; R&D Systems, Minneapolis, Minn) specific for the extracel-
lular domain of RAGE present in both sRAGE and mRAGE.
esRAGEwas measured in tissue homogenates by means of a commercially
available ELISA (B-Bridge International, Tokyo, Japan) that specifically
detects esRAGE by using a capture mAb recognizing RAGE-extracellular
domain and a detection polyclonal antibody for the esRAGE-specific C-
terminal 16 amino acid sequence.E4
Tissue ECP levels were measured by UniCAP (Phadia, Uppsala, Sweden).
Tissue homogenates were assayed for IL-5, MMP-3, MMP-7, andMMP-9,
and tissue-cube supernanants for IL-1b, IL-5, and TNF-a levels by means of
Luminexx MAP technology using the Fluorokine MAP Multiplex Human
Cytokine Panel A Kit (R&D Systems) on a Bio-Plex 200 Array Reader (Bio-
Rad Laboratories, Hercules, Calif).
Commercially available ELISA kits were used to measure the levels of
ADAM-10 (USCNK Life Science, Wuhan, China) and TGF-b1 (R&D
Systems). For TGF-b, acid was added during the ELISA procedure, resulting
in physicochemical activation of latent TGF-b. Total TGF-b concentrations
are reported including both active and latent forms.
The respective sensitivities of above-mentioned assays aregiven inTableE1.
RAGE immunohistochemistryHuman sinonasal tissues from controls, CRSsNP patients, and CRSwNP
patients were fixated in formalin (Fluka, Sigma-Aldrich, Bornem, Belgium)
and embedded in paraffin. Paraffin sections (4-5 mm) were air-dried for 24
hours at 378C. After deparaffinization in parasolve, endogenous peroxidase
activity was blocked with 0.3% hydrogen peroxidase in TRIS-buffered saline
(TBS; pH 7.8) containing 0.001% NaN3 for 20 minutes at room temperature.
The sections were washed with TBS for 10 minutes before being incubated
overnight at 48C with polyclonal anti-human antibodies raised against the
N-terminus of human RAGE (Santa Cruz Biotechnology, Heidelberg, Ger-
many) at a dilution of 1/200 in TBS/5%BSA. Next, the slides were washed
for 10 minutes in TBS. RAGE expression was detected by using the LSAB1technique conjugated with peroxidase according to the manufacturer’s
instructions (labeled streptavidin-biotin; Dako, Heverlee, Belgium). The per-
oxidase activity was detected by using 3-Amino-9-Ethylcarbazole (AEC) sub-
strate chromogen (Dako), which results in a red-stained precipitate. Finally,
sections were counterstained with heamatoxyline for 2 minutes, washed in
running tap water, and mounted in Aquatex (VWR International, Radnor,
Pa). Negative controls consisted of goat IgG-negative control serum (Santa
Cruz Biotechnology); control immunostainings did not yield immunosignals.
RAGE gene expression analysisTotal RNAwas extracted by using an Aurum Total RNAMini Kit (Bio-Rad
Laboratories). One microgram of total RNA was reverse transcribed to
generate first-strand cDNA by using an iScript cDNA Kit (Bio-Rad Labora-
tories). Two microliters of a 5 times dilution of each cDNA reaction mix was
used as a template for real-time PCR amplification performed on an Light
Cycler LC480 System (Roche) in the presence of 250 nM of forward and
reverse primers, and 13 PCRMaster mix (Roche) in a total reactionvolume of
5 mL. The thermal cycling conditions were as follows: 958C for 1 minute,
followed by 45 cycles of 3 seconds at 958C, 30 seconds at 608C, and 1 secondat 728C, and a dissociation curve analysis from 608C to 958C. b-actin (ACTB),hypoxanthine phosphoribosyltransferase 1, and elongation factor 1 were used
as endogenous reference for normalization. Primer sequences for ACTB,
hypoxanthine phosphoribosyltransferase 1, mRAGE, and sRAGE were
obtained from the public Real-Time PCR Primer and Probe Database of the
Department of Medical Genetics, University of Ghent (http://medgen.ugent.
be/rtprimerdb; RT Primer DBID1 for ACTB, ID4 for hypoxanthine phospho-
ribosyltransferase 1, ID3908 for mRAGE, and ID3909 for sRAGE). The
primers’ sequences for elongation factor 1 are CTGAACCATCCAGGC-
CAAAT for the forward primer and GCCGTGTGGCAATCCAAT for the re-
verse primer. Relative changes in gene expression were determined on the
basis of the comparative Ct method as described before.E5
Statistical analysisDifferences between data obtained from the 3 patient groups were analyzed
by a nonparametric Kruskal-Wallis test, while the differences between IL-5–
negative and IL-5–positive, ECP-low versus ECP-high groups, and the data
Rhinosinusitis: establishing definitions for clinical research and patient care.
J Allergy Clin Immunol 2004;114:155-212.
E4. Sakurai S, Yamamoto Y, Tamei H, Matsuki H, Obata K, Hui L, et al. Develop-
ment of an ELISA for esRAGE and its application to type 1 diabetic patients. Di-
abetes Res Clin Pract 2006;73:158-65.
E5. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-
time quantitative PCR and the 2-[Delta][Delta]CT method. Methods 2001;25:
402-8.
J ALLERGY CLIN IMMUNOL
VOLUME 129, NUMBER 6
VAN CROMBRUGGEN ET AL 1521.e2
from the in vitro sRAGE breakdown assay were analyzed by using a Mann-
Whitney U test. Data derived from human sinonasal ex vivo tissue-cube frag-
ment stimulation assays were compared by means of a Wilcoxon signed-rank
test. Correlation analysis was performed by using the Spearman rank-order
method yielding a Spearman’s rho (rs) coefficient. A P value of less than or
equal to .05 was considered to be statistically significant (GRAPHPAD, San
Diego, Calif).
RESULTS
Tissue protein levels of IL-5 and ECPTissue levels of IL-5 were significantly higher in CRSsNP
patients (13.5 pg/g tissue; range, <5.83-283.1 pg/g tissue; P <.01)compared with controls (all below the detection limit of 5.83 pg/gtissue). The levels of IL-5 in tissue of CRSwNP patients (121.0pg/g tissue; range, 15.5-1229.2 pg/g tissue; P < .01) were signif-icantly higher versus controls (P < .001) and CRSsNP patients.While none of the control tissues was IL-5 positive, 12 of the22 CRSsNP patients had tissue IL-5 levels above the detectionlimit (see Table I in the main document for the median and rangeof the IL-5–positive CRSsNP samples). All tissue samples fromCRSwNP patients were IL-5 positive.
ECP protein concentrations were significantly higher inCRSwNP (7,507.5 ng/g tissue; range, 458-32,725 ng/g tissue;P < .001) compared with controls (106.8 ng/g tissue; range,11-1,738 ng/g tissue; P < .001) and CRSsNP (2,376 ng/g tissue;range, 270-15,620 ng/g tissue), the latter’s ECP levels being sig-nificantly higher (P < .05) than the ECP levels in controls.
A B
control CRSsNP CRSwNP0
200
400
600
800***
***
MM
P-3 (p
g/g
tis
su
e)
control CRSsNP CRSwNP0
1000
2000
3000
4000
5000***
***
MM
P-7 (p
g/g
tis
su
e)
C D
control CRSsNP CRSwNP0
5000
10000
15000
20000
25000**
MM
P-9 (p
g/g
tis
su
e)
control CRSsNP CRSwNP0
1000
2000
3000
4000
5000
AD
AM
-10 (n
g/g
tis
su
e)
FIG E1. Quantification of tissue MMP-3 (A), MMP-7 (B), MMP-9 (C), and ADAM-10 (D) protein levels by
means of ELISA in controls and CRSsNP and CRSwNP patients. Values are expressed as pg/g tissue and pre-
sented as box-and-whisker plots showing the minimum and maximum values, the lower and upper quar-
tiles, the median (as a line), and the mean (as a cross). **P < .01; ***P < .001.
J ALLERGY CLIN IMMUNOL
JUNE 2012
1521.e3 VAN CROMBRUGGEN ET AL
A
control CRSsNP CRSwNP
0
2
4
6
8
10
mR
AG
E
mR
NA
rela
tiv
e co
py n
um
ber
B
control CRSsNP CRSwNP
0
1
2
3
4
*
**
esR
AG
E
mR
NA
rela
tiv
e co
py n
um
ber
FIG E2. mRAGE (A) and esRAGE (B)mRNA expression bymeans of RT-PCR
in the tissue of controls and CRSsNP and CRSwNP patients. Data are ex-
pressed as normalized relative copy number and presented as box-and-
whisker plots showing the minimum and maximum values, the lower
and upper quartiles, the median (as a line), and the mean (as a cross).*P < .05; **P < .01.
J ALLERGY CLIN IMMUNOL
VOLUME 129, NUMBER 6
VAN CROMBRUGGEN ET AL 1521.e4
A
B
FIG E3. Inverse correlation of mRAGE levels in the tissue of controls with
the level of MMP-9 (A) (rs 520.8200; P5 .002) and the level of ADAM-10 (B)
(rs 5 20.6818; P 5 .0208).
J ALLERGY CLIN IMMUNOL
JUNE 2012
1521.e5 VAN CROMBRUGGEN ET AL
A
IL-5-negative IL-5-positive
0
10000
20000
30000
40000***
EC
P tis
su
e (n
g/g
tis
su
e)
B
IL-5
-negative
IL-5
-positiv
e
IL-5
-negative
IL-5
-positiv
e
IL-5
-negative
IL-5
-positiv
e
0
10000
20000
30000
**
none
control CRSsNP CRSwNP
none
******
***
EC
P tis
su
e (n
g/g
tis
su
e)
FIG E4. Tissue levels of ECP in IL-5–negative and IL-5–positive samples
over all patient groups (A) and in the separate clinical entities (B). Values are
expressed as pg/g tissue and presented as box-and-whisker plots showing
the minimum andmaximum values, the lower and upper quartiles, the me-
dian (as a line), and the mean (as a cross). **P < .01; ***P < .001.
J ALLERGY CLIN IMMUNOL
VOLUME 129, NUMBER 6
VAN CROMBRUGGEN ET AL 1521.e6
0
100
200
300
400
500
600
0
5
10
15
20
25
30
35
IL-1β TNF-α IL-5
medium alonesRAGE (5 µg/mL)
*
*
*
IL-1
β (p
g/m
L)
IL-5 an
d T
NF
-α (p
g/m
L)
FIG E5. Effect of recombinant 5 mg/mL sRAGE on the release of IL-1b, TNF-
a, and IL-5 proteins in an ex vivo human tissue assay (n 5 6). *P < .05.
J ALLERGY CLIN IMMUNOL
JUNE 2012
1521.e7 VAN CROMBRUGGEN ET AL
TABLE E1. Sensitivities of Luminex, ELISA, and UniCap kits
