CLINICAL STUDYReceptor for advanced glycation end-products expressionin subcutaneous adipose tissue is related to coronaryartery diseaseB K Rodin o-Janeiro1,2,*, A Salgado-Somoza1,2,*, E Teijeira-Fernandez1,2, J R Gonzalez-Juanatey1,2,3, E Alvarez1,3and S Eiras1,21Laboratorio no. 6, Instituto de Investigacion Sanitaria de Santiago de Compostela, Edicio Consultas externas (planta-2), Hospital Cl nico Universitariode Santiago, Traves a Choupana s/n, 15706 Santiago de Compostela, A Coruna, Spain,2Unidad Coronaria y Servicio de Cardiolog a, Hospital Cl nico deSantiago de Compostela, 15706 Santiago de Compostela, Spain and3Departamento de Medicina, Universidad de Santiago de Compostela, 15782 Santiagode Compostela, Spain(Correspondence should be addressed to E Alvarez at Laboratorio no. 6, Hospital Cl nico Universitario de Santiago; Email: ezequiel.alvarez@usc.es)*(B K Rodin o-Janeiro and A Salgado-Somoza contributed equally to this work)AbstractObjective:Obesity, a risk factor for coronary artery disease (CAD), is associated with inammation andreactive oxygenspecies (ROS) production, while advancedglycationend-products, throughtheirreceptor (AGERor RAGE), playanimportant role onthese processes. The aimof this study was to analyzetheexpressionlevelsof RAGE, NADPHoxidasesubunits, andcatalaseinadiposetissueinrelationwith CAD.Design and methods: Patients undergoing heart surgery were included in two groups: with and withoutCAD. Epicardial adipose tissue (EAT) and subcutaneous adipose tissue (SAT) biopsies were analyzed forgene expression by RT-quantitative PCR, immunohistochemistry, or western blot.Results: RAGE mRNAand protein expression in SAT frompatients with CAD was lower than in patientswithout CAD. However, there was no change in EAT frompatients with or without CAD. P22-PHOXandRAGEgeneexpressionwerehigherinEATthaninSAT,whereascatalasemRNAlevelswerelower.NADPHoxidase subunits and catalase mRNAexpression were not inuenced by CAD. Whereas NADPHoxidase-dependent oxidativeresponseof SATandEATtolipidcirculatinglevelscouldbedifferent;glycemic levels were not related with the analyzed genes expression.Conclusions: This study demonstrates that RAGE expression in SAT, but not in EAT, is down-regulated inpatients with CADwith respect to those without CAD. Although changes were not observed for NADPHoxidasesubunitsorcatalaseexpressionbetweenCADandnon-CADpatients,apossiblerelationshipbetween ROS production and RAGE expression in adipose tissues cannot be ruled out.European Journal of Endocrinology 164 529537IntroductionObesity, associated with an increase in the size ofadipocytes and endocrine function of adipose tissue (1),isrelatedtocardiovasculardiseaseandtype2diabetes(2).Thus,increasedepicardial,pericardial,andsubcu-taneous fat is associated with the presence and severityof coronary artery calcium(3). However, greaterattention has focused on visceral adipose tissue incomparisonwithsubcutaneousadiposetissue(SAT)inthe last few years because it is an important risk factorfor cardiovascular disease (4). In fact, in femalecoronary artery disease (CAD) patients, a strongdecrease was observed in SAT of the legs in comparisontohealthycontrols,withnosignicantshiftinSATtothe trunk (5), suggesting redistribution to visceraladipose tissue. Epicardial adipose tissue (EAT) is thevisceral adipose tissue locatedinthe atrioventricularandinterventriculargrooves, adjacenttoadventitiaofcoronary arterieswithoutfasciastructures.Thistissueis asource of pro-inammatorycytokines (68) andcontains lowlevels of adiponectin, an anti-inam-matory adipokine, in patients with CAD (9, 10).Moreover, adiponectinexpressioninEATdecreasesinpatients with hypertension (11); although its levels arenot modiedindiabetes (12). Thesendings suggestthat EATplays animportant role incardiovasculardisease and some of its risk factors. Thus, the behavior ofthis fat pad seems to be different from SAT.Recently, it wasfoundthat adiposetissueexpressesthe receptor for advanced glycation end-products (AGERor RAGE) (13). This is amultiligandreceptor of theimmunoglobulinsuperfamily(14)thatbindsadvancedglycation end-products (AGE). AGE are modications ofEuropean Journal of Endocrinology (2011) 164 529537 ISSN 0804-4643q 2011 European Society of Endocrinology DOI: 10.1530/EJE-10-0904Online version via www.eje-online.orgproteins or lipids that become non-enzymaticallyglycatedandoxidizedaftercontactwithaldosesugars(15).AGEforminvivoinhyperglycemicenvironmentsand during aging, and they have been related with theinitial steps of atherosclerosisandwithinammatoryresponses (16). Among the effects caused by theseproducts, we can emphasize generation of reactiveoxygen species (ROS), migration of cells, and expressionof adhesion and inammatory molecules (17). TheAGERAGEengagementiswidelyrelatedwithcardio-vasculardiseaseandROSgeneration,mainlymediatedbyNADPHoxidase(18). Thisenzymeconsistsintwomembrane-bound subunits, the gp91-PHOX protein(NOX2) or some of its homologs (named NOX from 1 to5) and p22-PHOX protein (19). Once activated, NADPHoxidaseproducessuperoxideanions fromoxygenandNADPHor NADH. Enhanced ROS production is animportant factorassociatedwithsomecardiovasculardiseases (20) such as CAD. Furthermore, we havepreviously observedthatEATpresentshigheroxidativestress than SAT (21).Taking all this together, we can hypothesize that EAT,due to its physical location, its endocrinological activity,and its possible reactivity to AGE, could play animportant role in CAD, a situationtypically relatedwith enhancement of both ROS production and inam-matory reactions. The aim of this study was to analyzethe relationshipbetweenCADandthe expressionofRAGE, NADPHoxidasesubunits, andcatalaseinEATand SAT of patients undergoing heart surgery.Subjects and methodsPatientsThirty-sevenpatients undergoingheart surgerywithsternotomy were included in the study. Exclusion criteriawereprevious heart surgeryor concomitant infectivediseases. All participantsgavetheirinformedconsent.Thestudyprotocol wasapprovedbythelocal ethicalcommittee and carried out according to the Declarationof Helsinki. Clinical data were obtained upon admissionto hospital before surgery. Diagnosis of CAD wasbasedonpreviouscoronaryangiogram.Reductionsinluminal coronary artery diameters inexcess of 50%wereconsidered signicant. Heart valve disease was the causeof surgery in all patients without CAD.Adipose tissue biopsiesEATbiopsy(0.11 gwet weight) was obtainedfromupper region of the right ventricle, and SAT biopsy (2 gwet weight) was obtained fromthe thorax. Sampleswereimmediatelysplit,whenpossible,intotwopieces.One piece was frozen in liquid nitrogen before storage atK80 8C until use, and the other one was formalin-xedand parafn-embedded.mRNA extraction and real-time RT-PCRmRNAwas isolatedfrom50to100 mgof EATandSAT with the oligotex mRNA Spin-column kit(Qiagen GmbH). The nal concentration was 1 mgtissue/1 ml solution. Then, 4.14 ml of mRNAsolutionwas transcribed using 200 U of Moloney murineleukemiavirusreversetranscriptase(InvitrogenCorp.)in 30 ml of a pH 8.4 solution containing 20 mMTrisHCl, 50 mMKCl, 2.5 mMMgCl2, dNTPs (1 mMeach), 20 Uof RNase inhibitor, andrandomprimersunder the following conditions: 50 min at 37 8C, 10 min42 8C, and 5 min at 95 8C.Comparative EAT and SAT gene expression levels withrespect to b-actin were analyzed by real-time PCR(Chromo 4; MJ Research, Inc., Waltham, USA) induplicated using 4 ml of cDNA, SYBRGreen (RocheDiagnostics Corp., Inc.) as uorochrome, and the primerssequence and amplication conditions shown in Table 1,during40cycles. Fluorescence curves wereanalyzedusingOpticonMonitorsoftware(MJResearch). Melting curveswere tested in order to probe the correct amplicon. Geneexpressionlevelswerecalculatedwithrespecttob-actinexpression and represented in arbitrary units (a.u.).ImmunohistochemistryParafn-embedded tissue sections 3 mm-thick weredeparafned and rehydrated. Epitope retrieval wasdoneinbasicsolutionfor20 minat97 8C. Immunos-taining was performed with anti-human RAGEantibody(1:500; Abcam, Cambridge, UK) for 30 minfollowing the protocols. LSAB (Dako Diagnostics,Glostrup, Denmark) using 3,30-diaminobenzidine tetra-hydrochloride was used for detection. Negative controlswere carried out omitting the primary antibody. Resultsfrom different samples were analyzed for quanticationof RAGE expression using the free ImageJ (1.37v)software from NIH, USA.Protein extraction and western blotProtein isolation froma series of SATsamples wasperformed as previously described (12). In brief,100150 mgof EATandSATwererinsedin5 ml ofphosphate saline solutioncontaining 0.5 mMEDTA,5 mMKCl, 10 mMHEPESacid, 2 mMMgCl2, 10 mMNaHCO3,0.5 mMKH2PO4,0.5 mMNaH2PO4,10 mMglucose, 110 mM NaCl, and 0.16 mM CaCl2 at pH 7.4,and thencentrifuged at 300 gfor 1 minto removeresidual blood. Tissues were resuspended (weight/volume) in600 ml lysis buffer (125 mMTris pH6.8,10% glycerol, 2% SDS, 100 mM dithiothreitol, 1!anti-protease cocktail fromSigmaAldrich) and shearedby homogenizer pestle using sample-grinding kit(GEHealthcare, Waukesha, WI, USA).Prior to protein separation, immunoglobulindepletion fromthe samples was performed to allow530 B K Rodino-Janeiro, A Salgado-Somoza and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.orgdetection and analysis of lowabundant proteins.Immunoglobulins were bound to protein AusingPureProteome Protein AMagnetic Beads (Millipore,Billerica, MA, USA). Fromtheproteinsolution, 20 mlwere diluted with 100 ml of phosphate-buffered solution,and the mixture was incubated with protein A magneticbeads for 30 min under agitation at room temperature.Finally, magnetic beads binding immunoglobulins wereseparated, and the depleted sample was recovered.Proteins were concentrated 10!with centrifugal lterdevices (Amicon Ultra-0.5, Millipore).Proteinseparation(40 mg) was carried out in12%SDSPAGE gel and transferred to a polyvinylidene uoridemembrane for 45 min at 280 mA. The membranes wereblockedfor 2 hat roomtemperaturewith5%milkinTris-bufferedsalineTween-20containing20 mMTrisHCl (pH7.6), 150 mMNaCl, and0.1%Tween-20. Themembranes were then exposed with rabbit RAGE antibody(1:1000dilution; Abcam) andmouseb-actinantibody(1:1000;SantaCruzBiotechnology,Delaware,CA,USA)overnight, andthentoperoxidase-conjugatedgoatanti-rabbit and anti-mouse IgG(1:1500). ImmunoreactiveproteinswerevisualizedusinganECLdetectionsystem(Amersham Pharmacia Ltd) and quantied by densitome-try Image J software. RAGE protein levels were evaluatedin duplicate and quantied with respect to b-actin.Data analysisAssumptions of normality of continuous variables werecheckedwithKolmogorovSmirnovtests. Continuousvariables were expressed as meanGS.D. Differencesbetween continuous variables were evaluated usingStudents t-test. Categorical variables were expressedas percentages and compared using the c2-test.Correlations between variables were calculated byPearsons or Spearmans tests according to the normal-ity of the variable, and two-tailed P values are shown.Statistical signicancewas denedas P!0.05. Allanalyses wereperformedwiththeStatistical Packagefor Social Sciences (SPSS) version15.0for Windows(SPSS, Chicago, IL, USA).ResultsClinical characteristics and main differencesbetween patients with and without CADThirty-sevenpatients undergoingheart surgerywereincluded inthe study and analyzed. Meanage was71.5G9.1years; 62.2%were male, andmeanbodymassindexwas28.3G7.2 kg/m2. All clinical charac-teristics are summarized in Table 2. Patients wereclassiedasbelongingtooneof twogroups: patientswithCADand patients without CAD. Differences inmain clinical characteristics were not observed betweengroups, as shown in Table 2.Inthispopulation, mRNAlevelsforAGER(RAGE),NOX(2, 4, and 5), p22-PHOX(CYBA), and catalase (CAT)weremeasuredinEATandSAT. Meanvaluesof theseparameters and the comparison between CAD and non-CAD groups of patients are shown in Table 3. From thiscomparison, only RAGE inSATwas expressed inadifferent manner between the two groups, being its levelssignicantly lower in CAD patients (PZ0.002; Fig. 1).To conrmthese results, sections fromparafn-embedded adipose tissue fromSAT or EAT wereimmunostained with anti-human RAGE. QuanticationTable 1Gene database accession numbers, sequences for forward (F) and reverse (R) primers, PCRconditions, and ampliedfragment lengths used.Genes Primers Accession numbers PCR conditions Length (bp)AGER (RAGE) F-ggaatggaaaggagaccaag NM_001136 30 s 95 8C 199R-cccttctcattaggcaccag 30 s 58 8C30 s 72 8CACTB (b-actin) F-ttctgacccatgcccaccat NM_001101.3 30 s 95 8C 198R-tggatgatgatatcgccgcgctc 30 s 586062 8C30 s 72 8CCAT (Catalase) F-gcctgggacccaattatctt NM_001752 30 s 95 8C 203R-gaatctccgcacttctcccag 60 s 58 8C60 s 72 8CCYBB (NOX2) F-tcacttcctccaccaaaacc NM_000397.3 30 s 95 8C 211R-gggattgggcattcctttat 30 s 60 8C30 s 72 8CNOX4 F-cttccgttggtttgcagatt NM_016931 30 s 95 8C 245R-tgggtccacaacagaaaaca 30 s 60 8C30 s 72 8CNOX5 F-ctacgtggtagtggggctgt NM_024505 30 s 95 8C 213R-atgcaggaactggagcagat 30 s 62 8C30 s 72 8CCYBA (p22-PHOX) F-cgcttcacccagtggtactt NM_000101.2 30 s 95 8C 200R-gagagcaggagatgcaggac 30 s 60 8C30 s 72 8CRAGE in subcutaneous and epicardial adipose tissue 531 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.orgofthisimmunostainedareaintheimmunohistochem-ical preparationsconrmedthatinSATfrompatientswith CAD, RAGE expression was signicantly reduced incomparison with expression in EAT of the same patientand with respect to SAT and EAT from patients withoutCAD (Fig. 2).RAGE expression in SAT was also analyzed at proteinlevelbywesternblotinasubsetofpatients(fourCADandthreenon-CAD). Inthesemeasurements, relativeRAGE protein quantity, normalized with respect tob-actin, showedthatinSAT, RAGEwaslowerinCADpatientsthaninpatientswithout CAD(Fig. 3). Theseresults agreed with the observation done at mRNA level,whereRAGEexpressioninSATwasalsolowerinCADpatients with respectto non-CAD patients.We also analyzed the difference in the gene expressionbetween EAT and SAT. In this comparison, resultsshowed that RAGE and p22-PHOXwere highlyexpressedin EAT than in SAT, and NOX4 and catalase were highlyexpressed in SAT with respect to EAT (Table 4).With regard to the possible inuence of theco-morbiditiesof patients, theexpressionof thegenesunderstudywasanalyzedforthepossibleinuenceofhypertension,hyperlipidemia, type2diabetesmellitus,and atrial brillation. We only found aninterestingmodication depending on diabetes mellitus: RAGEmRNAlevels in both EAT and SAT were lower inpatients with type 2 diabetes in comparison withnon-diabeticpatients(Table5).Theseresultssuggestadown-regulating effect of diabetes on the RAGEexpressionof theadiposetissue. Fromtheothergenesanalyzed, NOX2mRNAlevels inSATwere lower innon-diabeticpatientswithrespecttodiabeticsand, onthe contrary, NOX5 in SAT was lower in diabeticpatients (Table 5).Relationship between variablesWe analyzed the relationship between mRNAexpressionof NADPHoxidasesubunits, catalase, andRAGEregardingtotheadiposetissuesource. Inthissense, we observed that p22-PHOX levels in EATpositively correlated with levels in SAT in allpatients (rZ0.766; PZ0.001). Althoughthis corre-lation was maintained in the non-CAD group of patients(rZ0.809; PZ0.028), it disappeared in the CAD group(rZ0.430; PZ0.336; datanot shown). If we acceptthatp22-PHOXcouldbeanindicatorof total NADPHoxidaseexpressioninthetissue,totallevelsofNADPHoxidase would be directly related in EAT and SAT in theabsence of CAD, but not in CAD patients.SomethingsimilarhappenedwithRAGE: itsmRNAlevels in EAT and in SAT were positively related(rZ0.425; PZ0.014; Fig. 4A) considering all patients.However, whereas in the non-CAD patients, thisassociation became stronger (rZ0.652; PZ0.002;Fig. 4B), it disappearedintheCADgroupof patientsTable2Clinical characteristicsof all patients, andcomparisonbetweenpatientswithandwithout CAD. Valuesareexpressed as a percentage or as meanGS.D.Variables All patients (nZ37) Non-CAD (nZ20) CAD (nZ17) P valueAge (years) 71.5G9.1 71.1G10.1 72.0G8.1 0.779Gender (male) 62.2 50.0 76.5 0.094BMI (kg/m2) 28.3G7.2 28.3G5.3 28.3G9.1 0.992Hypertension 76.9 76.2 77.8 0.605Diabetes mellitus type 2 24.3 19.0 33.3 0.257Hyperlipidemia 64.9 52.6 77.8 0.104Atrial brillation 21.2 20.0 23.1 0.581Glucose (mg/dl) 112.5G32.3 109.4G19.0 116.2G43.3 0.517Creatinine (mg/dl) 1.3G0.4 1.2G0.3 1.4G0.6 0.437HbA1c (%) 5.4G0.6 5.4G0.4 5.4G0.9 0.946Total cholesterol (mg/dl) 168.2G41.4 179.9G40.3 155.8G39.9 0.076LDL (mg/dl) 100.2G33.5 114.4G32.2 86.0G29.7 0.055HDL (mg/dl) 37.7G12.0 39.4G15.7 36.1G7.3 0.558Triglycerides (mg/dl) 117.8G55.1 109.6G50.0 126.4G60.3 0.366P value for the comparison was obtained byc2test or by Students t-test for categorical or continuous variables respectively. BMI, bodymass index;HDL, high-density lipoprotein; LDL, low-density lipoprotein.Table 3Relative expression values in EAT and SAT for the genesindicated. Meanvaluesfor all patientsandcomparisonbetweenpatientswithandwithout CAD. ValuesareexpressedasmeanGS.D. in a.u.VariablesAllpatients Non-CAD CAD P valueCatalase EAT 8.60G0.69 8.63G0.61 8.57G0.82 0.875Catalase SAT 9.21G0.62 9.46G0.65 8.96G0.52 0.141NOX2 EAT 6.29G0.34 6.24G0.18 6.33G0.45 0.634NOX2 SAT 6.13G0.41 6.29G0.41 6.00G0.39 0.209NOX4 EAT 5.94G0.40 5.95G0.51 5.94G0.31 0.969NOX4 SAT 6.31G0.25 6.33G0.28 6.30G0.24 0.820P22-PHOX EAT 8.07G0.45 8.28G0.51 7.86G0.28 0.081P22-PHOX SAT 7.65G0.40 7.80G0.48 7.50G0.27 0.175NOX5 EAT 4.44G0.31 4.46G0.26 4.41G0.37 0.741NOX5 SAT 4.50G0.32 4.48G0.22 4.52G0.40 0.833RAGE EAT 7.34G0.53 7.37G0.51 7.24G0.55 0.497RAGE SAT 7.02G0.46 7.36G0.73 6.74G0.32 0.002Pvalue for the comparison was obtainedby Students t-test.532 B K Rodino-Janeiro, A Salgado-Somoza and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.org(rZ0.079; PZ0.788; Fig. 4C). These results agree witha change in RAGE expression in SAT under CAD.Total cholesterol, low-density lipoprotein, and triglycer-idelevels wereinverselyrelatedtoNOX2(rZK0.619;PZ0.032), NOX4 (rZK0.742; PZ0.035), and p22-PHOX(rZK0.587; PZ0.045) mRNAlevels in EATrespectively; whereas mRNAlevels of NOX4 in SATcorrelated positively with serumlevels of triglycerides(rZ0.578; PZ0.049) andtotal cholesterol (rZ0.633;PZ0.027). Thesedataseemtoindicateadifferent oxi-dative response of SAT and EAT to lipid circulating levels.Inour study, there was no signicant relationshipbetween glucose or HbA1c levels and the gene expressionsanalyzed. Only a signicant inverse relationship wasobservedbetweenHbA1cpercentageandRAGEmRNAlevels in EAT, but not in SAT (rZK0.515; PZ0.041).DiscussionFor the rst time, we have found that EAT from patientswith cardiovascular disease who underwent heartsurgery, expresses RAGE mRNA and protein. Moreover,thisexpressioninEAT andinSATwasrelatedtoCAD.Thus, RAGE expressionlevels were lower inSATofpatients with CAD in comparison with non-CADpatients. However, we have not found signicantchanges in EAT. On the contrary, NADPH oxidasesubunitsand catalasegene expressionin EAT and SATwere not inuenced by CAD, only the expression patternof p22-PHOX in these tissuescould be affectedby CAD.Diabetescanalsoreduce, independent of CAD, RAGEexpression in human adipose tissue.Patientsundergoingheart surgerywereselectedtoparticipate inthestudyanddividedintotwogroupswithsimilarclinical characteristicsandco-morbiditiesunless the presence of CAD. In these conditions, mRNAlevels of RAGE, some NADPH oxidase subunitsexpressed in adipose tissue, and catalase were measuredin SAT and EAT. Unless for RAGE, no statisticaldifferences in mRNA levels of these genes were observedafter comparing tissues between the groups of CAD andnon-CADpatients. RAGEmRNAlevels inSATwereB*P = 0.002A9RAGE mRNA levels in EAT876No CAD CAD9RAGE mRNA levels in SAT876No CAD CADP = 0.497Figure 1 Box plots for RAGE mRNA expression levels in EAT (A)and in SAT (B) in the groups of patients with CAD or without CAD.*P!0.05 with respect to non-CAD group of patients.ACEAT SATNon-CADCADBD01234567No CAD CAD*RAGE stained area (%)EATSATFigure 2 Representative sections of parafn-embedded epicardial(A and C) or subcutaneous (B and D) adipose tissue from patientswith CAD (C and D) or without CAD (A and B) were immunostainedwith anti-RAGE. Positive immunostaining for RAGE was lesspronounced in SAT from patients with CAD (D) than in patientswithout CAD (B) in adipocytes and in vessels (insets). RAGEexpression in EAT was maintained to the same levels in non-CADpatients (A) and in CAD patients (C). Columns represent meanvalues (S.D. in vertical bars) of the percentage of RAGE-stainedarea in each type of preparation. *P!0.05 with respect to EAT. Fullcolour version of this gure available via http://dx.doi.org/10.1530/EJE-10-0904.RAGE in subcutaneous and epicardial adipose tissue 533 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.orgsignicantly lower inpatients withCADthanthosewithout CAD; whereas no differences were observedbetweenlevelsinEAT of thesamegroups. Infact, thepositive association between RAGE mRNA levels in EATandinSATfrompatientswithoutCADdisappearedinCAD patients. These results suggested that CAD inducesdown-regulation of RAGE expression in SAT, but not inEAT. Thishypothesiswasanalyzedatproteinlevel byimmunohistochemistry and western blot, and lowerpresence of RAGE was evidenced in SATof patients withCAD, while RAGE expression was maintained in EAT ofthese patients and in patients without CAD.Thephysiologicalmeaningofthisndingcannotbeinferredfromourstudy. However, RAGEexpressionintissues allows theengagement of its multipleligands(AGE, S100/calgranulins, and others), and this bindingcanmediatepro-inammatoryresponses(22), activateROSproduction(18) anddamagethecardiovascularsystem(23). Also, several variantsof theRAGEgenehavebeenimplicatedinthedevelopment of diabetes-associated atherothrombotic disorders. Fromthesevariants, the functional polymorphism Gly82Ser (data-base for SNP rs2070600) highlights. A signicantassociationbetweenthehaplotypecarryingtheG82SpolymorphismandanincreasedincidenceofCADandelevated serum CRP levels in the Chinese Hanpopulation has been described by Gao et al. (24).However, previous genetic epidemiological studiesshowednegativeresultsfromstudiesinvolvingCauca-siansubjectswithcardiovasculardisease(25).Furthermore, 82S allele seems to have a protective role inKorean CAD patients (26). Therefore, thepathophysio-logical roleof thesevariantsinCADremainselusive,and further investigations must be performed.It has been suggested that the diverse splicing variantsof RAGE are possible in many tissues, and their productsmayinuencetheRAGE-mediatedpathogenesis (27).Besides the RAGE expression variants, the tissuewhere they are expressed could have different0.20.40.60.8CAD No CADRAGE/-Actin (a.u.)RAGE-ActinNo CAD CAD ABFigure 3 RAGEproteinlevelsinSATfromCADandnon-CADpatientsanalyzedbywesternblot.(A)RepresentativeblotsusingantibodiesagainstRAGEandb-actinofSATfromfourpatients,twowithoutCADandtwowithCAD.(B)Relativeoptical densitiesofRAGEwithrespecttob-actinsignal fromsevenpatients(fourCADandthreenon-CAD).a.u.,arbitraryunitsrelativetob-actin.r = 0.425 P = 0.014r = 0.652P = 0.002 r = 0.079P = 0.788All patientsPatients without CADPatients with CADA9RAGE mRNA in EAT8766 7RAGE mRNA in SAT8B9RAGE mRNA in EAT8766 7RAGE mRNA in SAT8C9RAGE mRNA in EAT8766 7RAGE mRNA in SAT8Figure 4 Relationship between RAGE mRNA levels in SAT and inEAT considering all patients (A), considering the non-CAD group ofpatients (B), and considering the CADgroup of patients (C). Straightlines represent total adjustment with the condence intervalrepresented as curved lines. rZPearsons correlation coefcient.Value of P is indicated in each case. P!0.05 was consideredstatistically signicant.534 B K Rodino-Janeiro, A Salgado-Somoza and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.orgpathophysiological roles in cardiovascular diseases. Thisis the reason because it is worth to test the expression ofRAGEinadiposetissueandtoanalyzethedifferencesbetween tissues from different locations.Inthisobservational work, wecouldnot studythecausesof theRAGEexpressionmodicationinSATofpatientswithCAD. However, accordingtoour resultsandpreviouslyreporteddata, wecansuggest at leastone factor that could be related with RAGE-alteredexpression. That is, down-regulationof RAGEinSATcould be related to changes in the ROS production of thetissue. One of the ways for studying the oxidative stressconsists of measuring the enzymes involved inROSgeneration, as NADPH oxidase. P22-PHOX is theuniquesubunitpresentinmostoftheNADPHoxidasecomplexes (28); so, it can be used as a marker of almosttotal NADPHoxidase content. The fact is that themRNA levels of p22-PHOX in EAT were directlycorrelatedtothelevels inSATwhenconsideringthetotal population of thestudy, butthis correlation failedwhen considering only the group of patients with CAD.This result suggests that NADPHoxidase expressionchange in SATor EATcould be due in part to CAD. Thus,ROSproductionchanges couldalsoberelatedtothedown-regulation of RAGE in SAT, evidencing a triangleof interaction between RAGE, oxidative stress, andaltered gene expression that has been previouslyproposed by other authors (29). In this context,oxidativestresshasbeenreportedtobehigherinEATthan in SAT (21), and this could help to maintain RAGEexpression levels in EAT under CAD.Although in our study, NADPH oxidase subunits andcatalasemRNAlevelswerenot statisticallyrelatedtoRAGEexpression, wecouldnot discardthepossibilitythat ROS production can be related to RAGE expression.Independent of CAD, the results revealed a higherexpressionof RAGEandp22-PHOXinEATcomparedwithSAT, whiletherewaslowercatalaseexpressedinEATwithrespecttoSAT. ThiscouldindicateahigheroxidativestressinEATthaninSAT,possiblymediatedbyRAGE(23)andNADPHoxidase(wherep22-PHOXparticipateinalmost all theisoforms). Further more,lower catalase expression in EAT, an antioxidantenzyme, also contributes to anoxidative state. Thisnding agrees with previous observations (21).Intriguing relationships between circulating lipidlevels andNoxisoforms expressioninadipose tissuewere observed. Whereas Nox isoform expressions in EATwere inversely related to lipid levels, the oppositeoccurred in SAT. It could be supposed that SAT exposedto lipids can be more sensitive to ROS production.However, the higher mRNA levels of p22-PHOX and thelower levels of catalase inEATwithrespect to SATindicatedthatROSproductioninEAT couldbehigherthan in SAT in patients with cardiovascular diseases, aspreviouslyshown(21). It isalsoknownthat visceraladiposetissueisanimportantcontributortooxidativestress(30). Theoppositereactionof EATandSATtocirculating lipid levels could help to explain the differentresponse in RAGE expression observed in this work.Although it was not the aim of this study, our resultsalsosuggestedthatdiabetesmellituscaninuencethelevels of RAGE expression in the adipose tissue, since theexpressionof thisreceptorwaslowerinbothEATandSAT from diabetic patients with respect to non-diabeticpatients. However, this nding seems to be independentof the association observed between RAGE expression inSAT and CAD, because the distribution of diabeticpatients in the groups of CAD and non-CAD patients ofour sample turnedout tobe exactlythe same. Thisresult is very interesting since AGERAGE axis can playa pivotal role in the setting of diabetes-derived vascularcomplications (31). We do not have a physiologicalexplanationfor RAGE-reduced expressionindiabeticadipose tissue. Moreover, it is unknown which could bethe real meaning of RAGE down-regulation in thesettingof diabetes, sincetoblocktheAGERAGEaxishas disparate effects in diabetic complications, withdifferent clinical implications (32).Withregardtothelimitations of this work, rstly,difcultytoinclude heart surgerypatients made theTable4ComparisonbetweenthegeneexpressionlevelsinEATandSATof thegenesindicated. ValuesareexpressedasmeanGS.D. in a.u.Variables EAT SAT P valueNOX2 6.29G0.34 6.13G0.41 0.213NOX4 5.94G0.40 6.31G0.25 0.005NOX5 4.43G0.32 4.50G0.32 0.344P22-PHOX 8.07G0.45 7.65G0.40 0.000Catalase 8.60G0.69 9.21G0.62 0.018RAGE 7.34G0.53 7.02G0.46 0.002Pvaluefor thecomparisonwasobtainedbyStudentst-test. Statisticallysignicant (P!0.05) values are in boldface.Table5Relativeexpressionvaluesinpatientswithandwithouttype 2 diabetes mellitus for the genes indicated. Mean values for thecomparisonbetweenpatientswithandwithout diabetesmellitus.Values are expressed as meanGS.D. in a.u.VariablesNon-diabetes(nZ28)Diabetes(nZ9) P valueCatalase EAT 8.76G0.71 8.20G0.52 0.182Catalase SAT 9.34G0.68 8.89G0.26 0.237NOX2 EAT 6.19G0.34 6.51G0.25 0.133NOX2 SAT 5.96G0.35 6.53G0.21 0.013NOX4 EAT 5.99G0.46 5.83G0.18 0.526NOX4 SAT 6.33G0.27 6.28G0.22 0.760NOX5 EAT 4.53G0.27 4.19G0.28 0.062NOX5 SAT 4.63G0.18 4.04G0.25 0.001P22-PHOX EAT 7.97G0.44 8.30G0.44 0.228P22-PHOX SAT 7.60G0.31 7.77G0.63 0.512RAGE EAT 7.44G0.53 6.93G0.26 0.001RAGE SAT 7.15G0.41 6.61G0.37 0.002Pvaluefor thecomparisonwasobtainedbyStudentst-test. Statisticallysignicant (P!0.05) values are in boldface.RAGE in subcutaneous and epicardial adipose tissue 535 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.orgnal number of subjects of this studyrelativelylow.Further research with a larger population couldconrmourresults. Second, thestudywasperformedinpatientsundergoingheartsurgery, soextrapolationof the results to other type of patients would need to besupportedbyappropriatedatacollection. Third, otherthanforRAGE,wehavestudiedthemRNAlevels, butnottheproteincontent. Thesizeof theadiposetissuesample obtained from heart surgery is always small, andit was impossible to combine mRNA and proteinanalysis for all genes.Inconclusion, thisstudydemonstratedfor thersttimethat RAGEexpressioninSAT, but not inEAT, isdown-regulatedinpatients withCADincomparisonwithnon-CADpatients.Moreover,RAGEexpressionindiabeticpatientswasfoundtobelower thaninnon-diabeticpatientsbothinEATandinSAT. NochangeswereobservedforNADPH oxidasesubunitsorcatalaseexpression between CAD and non-CAD patients.However, a possible relationship between ROS pro-duction and RAGE expression in adipose tissues cannotberuledout.FurthermechanisticinvestigationswouldbenecessarytoexplainthesenewrelationshipsfoundbetweenRAGEexpressioninadiposetissueandCADor diabetes.Declaration of interestThe authorsdeclare that there is no conict ofinterest that could beperceived as prejudicing the impartiality of the research reported.FundingThis studywas supportedinpart withprojects fromthe SpanishGovernmentandfromtheXuntadeGalicia,PI070445and07/PXIB-918092PR respectively and a grant fromthe Xunta de Galicia,INCITE09E2R918146ES. The Isidro Parga Pondal programof theXunta de Galicia supported the work of S Eiras and E Alvarez.AcknowledgementsThe authors thankthe Heart SurgeryDepartment of the HospitalCl nico Universitario de Santiago de Compostela and its collaborationin the collection of adipose tissue samples.References1Marques BG, Hausman DB & Martin RJ. Association of fat cell sizeand paracrine growth factors in development of hyperplasticobesity. American Journal of Physiology 1998 275 R1898R1908.2DixonJB.The effectofobesity on healthoutcomes. Molecular andCellularEndocrinology2010316104108. (doi:10.1016/j.mce.2009.07.008)3Ahmadi N, Nabavi V, Yang E, Hajsadeghi F, Lakis M, Flores F, Zeb I,Bevinal M, Ebrahimi R&Budoff M. Increased epicardial, pericardial,andsubcutaneousadiposetissueisassociatedwiththepresenceand severity of coronary artery calcium. Academic Radiology 201017 15181524. (doi:10.1016/j.acra.2010.08.017)4FoxCS,MassaroJM,HoffmannU,PouKM,Maurovich-HorvatP,Liu CY, Vasan RS, Murabito JM, Meigs JB, Cupples LA,DAgostino RB Sr & ODonnell CJ. Abdominal visceral andsubcutaneous adipose tissue compartments: association withmetabolic risk factors in the Framingham Heart Study. Circulation2007 116 3948. (doi:10.1161/CIRCULATIONAHA.106.675355)5Wallner SJ, Horejsi R, Zweiker R, Watzinger N, Moller R,Schnedl WJ, Schauenstein K & Tafeit E. ROC analysis ofsubcutaneous adipose tissue topography (SAT-Top) in femalecoronaryheart diseasepatientsandhealthycontrols. Journal ofPhysiological Anthropology 2008 27 185191. (doi:10.2114/jpa2.27.185)6Eiras S, Teijeira-Fernandez E, Salgado-Somoza A, Couso E, Garcia-Caballero T, Sierra J & Juanatey JR. Relationship betweenepicardial adipose tissue adipocyte size andMCP-1expression.Cytokine 2010 51 207212. (doi:10.1016/j.cyto.2010.05.009)7Kremen J, Dolinkova M, Krajickova J, Blaha J, Anderlova K,LacinovaZ, HaluzikovaD, BosanskaL, VokurkaM, SvacinaS&HaluzikM.Increasedsubcutaneousandepicardialadiposetissueproduction of proinammatory cytokines in cardiac surgerypatients: possible role in postoperative insulin resistance. Journal ofClinical Endocrinology and Metabolism 2006 91 46204627.(doi:10.1210/jc.2006-1044)8Mazurek T, Zhang L, Zalewski A, Mannion JD, Diehl JT, Arafat H,Sarov-Blat L, OBrien S, Keiper EA, Johnson AG, Martin J,Goldstein BJ & Shi Y.Human epicardial adipose tissue is a sourceof inammatorymediators. Circulation200310824602466.(doi:10.1161/01.CIR.0000099542.57313.C5)9Eiras S, Teijeira-Fernandez E, Shamagian LG, Fernandez AL,Vazquez-Boquete A& Gonzalez-Juanatey JR. Extension of coronaryarterydisease is associatedwithincreasedIL-6anddecreasedadiponectingeneexpressioninepicardialadiposetissue.Cytokine2008 43 174180. (doi:10.1016/j.cyto.2008.05.006)10Iglesias MJ, Eiras S, Pineiro R, Lopez-Otero D, Gallego R,Fernandez AL, Lago F & Gonzalez-Juanatey JR. Gender differencesinadiponectinand leptinexpressioninepicardial and subcu-taneousadiposetissue. Findingsinpatientsundergoingcardiacsurgery. Revista Espanola de Cardiolog a 20065912521260.(doi:10.1157/13096596)11Teijeira-Fernandez E, Eiras S, Grigorian-Shamagian L,Fernandez A, Adrio B & Gonzalez-Juanatey JR. Epicardial adiposetissue expression of adiponectin is lower in patients withhypertension.JournalofHumanHypertension200822856863.(doi:10.1038/jhh.2008.75)12Teijeira-Fernandez E, Eiras S, Grigorian-ShamagianL, Salgado-Somoza A, Martinez-Comendador JM&Gonzalez-JuanateyJR.Diabeticandnondiabeticpatientsexpresssimilaradiposetissueadiponectin and leptin levels. International Journal of Obesity 201034 12001208. (doi:10.1038/ijo.2010.30)13UenoH,KoyamaH,ShojiT,MondenM,FukumotoS,TanakaS,OtsukaY, MimaY, MoriokaT, Mori K, Shioi A, YamamotoH,Inaba M & Nishizawa Y. Receptor for advanced glycationend-products (RAGE) regulation of adiposity and adiponectinis associated with atherogenesis in apoE-decient mouse.Atherosclerosis 2010 211 431436. (doi:10.1016/j.athero-sclerosis.2010.04.006)14Koyama H, Yamamoto H&Nishizawa Y. RAGE and solubleRAGE: potential therapeutictargets forcardiovascular diseases.Molecular Medicine 2007 13 625635. (doi:10.2119/2007-00087.Koyama)15GoldinA, BeckmanJA, Schmidt AM&Creager MA. Advancedglycation end products: sparking the development of diabeticvascularinjury. Circulation2006114597605. (doi:10.1161/CIRCULATIONAHA.106.621854)16Basta G, Schmidt AM&De Caterina R. Advanced glycationend products and vascular inammation: implications foraccelerated atherosclerosis in diabetes. Cardiovascular Research2004 63 582592. (doi:10.1016/j.cardiores.2004.05.001)17Basta G, Lazzerini G, Massaro M, Simoncini T, Tanganelli P, Fu C,KislingerT, SternDM, SchmidtAM&DeCaterinaR. Advancedglycation end products activate endotheliumthrough signal-536 B K Rodino-Janeiro, A Salgado-Somoza and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.orgtransductionreceptor RAGE: amechanismfor amplicationofinammatory responses. Circulation 2002 105 816822. (doi:10.1161/hc0702.104183)18Yan SF, Ramasamy R & Schmidt AM. The RAGE axis: afundamental mechanismsignaling danger to the vulnerablevasculature. CirculationResearch2010106842853. (doi:10.1161/CIRCRESAHA.109.212217)19Brandes RP&Kreuzer J. Vascular NADPHoxidases: molecularmechanisms of activation. Cardiovascular Research 2005 651627. (doi:10.1016/j.cardiores.2004.08.007)20Taniyama Y&Griendling KK. Reactive oxygenspecies in thevasculature: molecular and cellular mechanisms. Hypertension 200342 10751081. (doi:10.1161/01.HYP.0000100443.09293.4F)21Salgado-Somoza A, Teijeira-Fernandez E, Fernandez AL, Gonzalez-Juanatey JR & Eiras S. Proteomic analysis of epicardial andsubcutaneous adipose tissue reveals differences in proteinsinvolvedinoxidativestress.AmericanJournalofPhysiology.Heartand Circulatory Physiology 2010 299 H202H209. (doi:10.1152/ajpheart.00120.2010)22HofmannMA, DruryS, FuC, QuW, Taguchi A, LuY, AvilaC,KambhamN, Bierhaus A, NawrothP, NeurathMF, SlatteryT,Beach D, McClary J, Nagashima M, Morser J, Stern D &Schmidt AM. RAGEmediates anovel proinammatoryaxis: acentral cell surfacereceptor for S100/calgranulinpolypeptides.Cell 1999 97 889901. (doi:10.1016/S0092-8674(00)80801-6)23YanSF,RamasamyR&SchmidtAM.Thereceptorforadvancedglycation endproducts(RAGE) and cardiovascular disease.ExpertReviews in Molecular Medicine 2009 11 e9. (doi:10.1017/S146239940900101X)24Gao J, Shao Y, Lai W, Ren H & Xu D. Association of polymorphismsintheRAGEgenewithserumCRPlevels andcoronaryarterydisease inthe Chinese Han population. Journal ofHuman Genetics2010 55 668675. (doi:10.1038/jhg.2010.85)25Hofmann MA, Yang Q, Harja E, Kedia P, Gregersen PK,Cupples LA, Schmidt AM&HudsonBI. The RAGE Gly82Serpolymorphism is not associated with cardiovascular disease in theFramingham offspring study. Atherosclerosis 2005 182 301305.(doi:10.1016/j.atherosclerosis.2005.02.006)26Yoon SJ, Park S, Shim CY, Park CM, Ko YG, Choi D, Park HY, Oh B,Kim H, Jang Y & Chung N. Association of RAGE genepolymorphisms with coronary artery disease in the Koreanpopulation. Coronary Artery Disease 2007 18 18. (doi:10.1097/MCA.0b013e3280105b4d)27Park IH,YeonSI,Youn JH,Choi JE,SasakiN,ChoiIH & ShinJS.Expressionof anovel secretedsplicevariant of thereceptorforadvanced glycation end products (RAGE) in human brainastrocytes and peripheral blood mononuclear cells. MolecularImmunology 2004 40 12031211. (doi:10.1016/j.molimm.2003.11.027)28Lassegue B&GriendlingKK. NADPHoxidases: functions andpathologies inthe vasculature. Arteriosclerosis, Thrombosis, andVascular Biology 201030653661. (doi:10.1161/ATVBAHA.108.181610)29Wautier MP, Chappey O, Corda S, SternDM, Schmidt AM&WautierJL. Activationof NADPHoxidasebyAGElinksoxidantstress toalteredgeneexpressionviaRAGE. AmericanJournal ofPhysiology. Endocrinology and Metabolism 2001 280 E685E694.30Gletsu-Miller N, Hansen JM, Jones DP, Go YM, Torres WE,ZieglerTR&LinE.Lossoftotalandvisceraladiposetissuemasspredicts decreases inoxidative stress after weight-loss surgery.Obesity 2009 17 439446. (doi:10.1038/oby.2008.542)31Jandeleit-DahmK, Watson A&Soro-Paavonen A. The AGE/RAGE axis in diabetes-accelerated atherosclerosis. Clinical andExperimental Pharmacology and Physiology 200835329334.(doi:10.1111/j.1440-1681.2007.04875.x)32TanAL, Sourris KC, Harcourt BE, Thallas-Bonke V, PenfoldS,Andrikopoulos S, Thomas MC, OBrien RC, Bierhaus A,Cooper ME, Forbes JM&Coughlan MT. Disparate effects onrenal andoxidative parameters followingRAGEdeletion, AGEaccumulationinhibition,ordietaryAGEcontrolinexperimentaldiabetic nephropathy. American Journal of Physiology. RenalPhysiology 2010 298 F763F770. (doi:10.1152/ajprenal.00591.2009)Received 26 January 2011Accepted 31 January 2011RAGE in subcutaneous and epicardial adipose tissue 537 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164www.eje-online.org