The F11 Receptor (F11R/JAM-A) in Atherothrombosis: Overexpression

CardiovascularBiology andCell Signalling

TheF11 receptor (F11R/JAM-A)inatherothrombosis: Overexpression of F11R in atherosclerotic plaques

AnnaBabinska1,2 ,Bani M. Azari2 ,MoroO.Salifu2 ,Ruijie Liu1 ,Xian-Cheng Jiang1 ,Malgorzata B. Sobocka 2 , Dorothy Boo3 ,George Al Khoury 3 ,JonathanS.Deitch3 ,JonathanD.Marmur2 ,YigalH.Ehrlich4 ,Elizabeth Kornecki1 Departmentsof 1 Anatomy/Cell Biology, 2 Medicine, 3 Surgery, SUNY Downstate MedicalCenter,Brooklyn,New Yo rk,USA and 4 Program in Neuroscience,City University of NewYorkatStaten Island,New Yo rk,New Yo rk,USA

Summary F11R is the gene namefor an adhesion protein,called the tablished. Enhanced expression of theF11R/JAM-A message in F11-receptor,aka JAM-A,which under normal physiologicalcon- cultured EC from human aortic and venous vessels was ob- ditionsisexpressedconstitutivelyonthe surfaceofplatelets and served following exposureofthe cells to cytokines. Deter- localizedwithin tight junctions of endothelialcells (EC).Previous minationsofplateletadhesion to cultured EC inflamedbycom- studies of theinteractions between human platelets and EC sug- binedcytokine treatment in the presence of F11R/JAM-A –an- gestedthat F11R/JAM-Aplays acrucial role in inflammatory tagonists provided dataindicating that de novo expression of thrombosis and atherosclerosis. Thestudy reported here ob- F11R/JAM-Aonthe luminalsurfaceofinflamedEChas an impor- tained in-vivoconfirmation of this conclusion by investigating tant role in the conversion of EC to athrombogenic surface. F11R/JAM-Aprotein and mRNA in patients with aortic and pe- Furtherstudiesofthese interactions under flowconditionsand ripheral vasculardisease andinananimal modelofatherosclero- under in-vivosettings could provideafinalproof of acausal role sis. Molecular and immunofluorescence determinations reveal- forF11R/JAM-A in the initiationofthrombosis.Basedonour in- ed very highlevelsofF11R/JAM-A mRNA and F11R/JAM-A pro- vitro and in-vivostudiestodate,weproposethat therapeutic tein in atherosclerotic plaquesofcardiovascular patients.Similar drugs which antagonize the functionofF11R/JAM-A should be results were obtained with 12-week-old atherosclerosis-prone tested as novelmeans forthe prevention and treatment of athe- apoE-/- mice,an age in which atherosclerotic plaquesare well es- rosclerosis, heartattacksand stroke.

Keywords F11R,F11 receptor,F11R/JAM-A,inflammatorythrombosis, atherosclerosis, atherothrombosis ThrombHaemost 2007; 97: 272–281

Introduction F11R/JAM-A peptideantagonists and recombinant proteins (5–7) has begun to reveal the roleofF11R/JAM-A in the process F11Risthe name of agene encoding aprotein called the F11 re- of plateletadhesion to inflamedendothelial cells and identified ceptor,also known by the name junctionaladhesion molecule F11R/JAM-A as apotentiallyimportant moleculeinplatelet (JAM-A).This protein wasidentified first as aduplexofproteins plaque formation leading to inflammatorythrombosis and athe- constitutively expressed on the surfaceofhuman platelets (1). rosclerosis, with platelets as critical factorsinvolved in the pro- Stimulation of F11R/JAM-A caused platelet activation in apro- gression and development of cardiovasculardisease(8). Pro- cess involving the followingevents: F11R/JAM-A dimerization moter analysis of the humanF11R gene has revealed the pres- (2),phosphorylation by activatedPKC associated withtransient ence of an NF-kappaB (NF- κ B) consensus sequence, thus (reversible) translocation of PKC isoenzymes α and ζ ,and last- furtherimplicating F11R’srole in inflammatoryprocesses (9). ing translocation of PKC isoenzymes δ , β ,n’, and θ (3), complex Enhanced expression and dynamic redistribution of cell ad- formation of F11R/JAM-A with the integrin GPIIIa (2) and acti- hesion molecules(CAMs) such as intercellular adhesion mol- vation of the Fcγ RIIpathway (4). Theutilization of specific ecule(ICAM), vascularcelladhesion molecule-1 (VCAM-1),

Correspondence to: Received August19, 2006 Elizabeth Kornecki, PhD Accepted after resubmission December 18, 2006 Department of Anatomy/CellBiology SUNYDownstateMedical Center Prepublished onlineJanuary8,2007 450 ClarksonAve doi:10.1160/TH06–08–0454 Brooklyn,NY11203, USA Te l.: +1 718 982 3932, Fax: +1 718 270 3732 E-mail: [email protected]

272 Babinska et al.F11R/JAM-Aexpression in atherothrombosis and plateletendothelial adhesion molecule-1 (PECAM- F11R/JAM reagents 1/CD31),all members of the immunoglobulin superfamily, Monoclonalantibody F11 (M.Ab.F11; human F11R/JAM anti- occur in the endothelium in response to cytokines (10, 11). Ex- body), developedinour laboratoryagainst the humanplatelet pression and localization of CAMs at the luminal surfaceofathe- F11R(1), wasobtained from BD PharMingen (San Diego, CA, rosclerotic lesionsimplicates their participation in the pathology USA). JAM-1 polyclonalantibody waspurchasedfrom R&D of atherosclerosis (12–15) with ICAM-1,VCAM-1and Systems. HumanF11R recombinant protein wasgenerated as PECAM-1 levels modulatedwithin the lesioned vasculature of described (6,7). TheF11R peptides were synthesized and puri- the atherosclerosis-prone apoE -/- mousemodel (15). fied (95%)byNew England Peptide, Inc.(Gardner,MA, USA). Similartothese CAMs, F11R/JAM-A is acelladhesion mol- vWF antibodywas obtainedfrom Chemicon International(Te- ecule and member of the immunoglobulin superfamily(16) mecula, CA,USA). whose important role in thrombosis and atherosclerosis is being exploredbyusand other laboratories. The combined treatment To tal RNAisolation of EC with inflammatorycytokines tumor necrosis factor-alpha TotalRNA wasextractedfrom the isolatedaorta of apoE-/- and (TNFα) and interferon-gamma(INFγ) resultedinaredistribu- control C57BL/6J mice utilizing TriZol(Invitrogen, Carlsbad, tion of F11R/JAM-Afrom intercellular junctions to the cellsur- CA,USA). ApoEknockout mice (apoE-/-)with C57BL/6J back- face of the inflamedEC(17). An enhancedexpression of F11R/JAM-A wasevident already in the earlystagesofatheros- clerosis in the endothelium of apoE-deficient micefed with high-fat diet (18), and the solubleformofF11R/JAM-A significantly reduced mononuclear cellrecruitment to ex vivo perfused atherosclerotic carotid arteries (18). TheF11R gene is located in acritical region on chromosome1q21–23 (F11R-USF1), aregion thatissignificantly linked to coronaryheartdiseaseand familial-combined hyperlipidemia (19).This finding also sug- gests that the F11R/JAM-A protein has proatherogenicproper- tiespredicted in its involvement in pathophysiological processes underlying cardiovasculardiseases (5,6). Furthermore, the ap- proximate two-foldupregulation of F11R/JAM-A shown to occur in unstable versusstableregions of the atherosclerotic plaque (20), expands the role of F11R/JAM-A to include its involvement in the rupture of the atherosclerotic plaque as well. In the present study,weconductedexperiments in culture and in vivo designedtotestthe suggestion that the interaction of the newly-expressed F11R/JAM-A sites on the surfaceofcytokine- A) inflamedECwith the constitutively-expressed F11R/JAM-A sites on thesurfaceofcirculating platelets plays acausal role in the adhesion of platelets to the inflamedendothelium.

Materials andmethods Human endothelialcells Human aortic(HAEC)and umbilicalveinendothelial cells (HUVEC) (106 )(Cascade Biologics, Inc.,Portland,OR, USA), were growninMedium 200containing 1% or 2% fetal calf serum (FCS), respectively.Atsecondpassage,both HUVEC and HAECwere treated with purifiedhuman recombinant TNFα (100units/ml)(R&D Systems, Inc., Minneapolis, MN,USA), B) vascularendothelial growth factor (VEGF;10ng/ml)(R&D Sys- tems) and INFγ (200units/ml)(Roche Diagnostics, Mannheim, Figure1:F11R/JAM-A expression in atheroscleroticplaques of o Germany) for 24 hours (h) at 37 C. Theprocedures involving the apoE-/- mice. A) RT-PCR:F11R mRNA levels obtainedfrompooled adhesion of platelets to endothelialcells were followedasde- samples of proximalaortaeoffiveapoE -/- and five C57BL/6 wild-type scribed(6). mice(control). Lane 1: controlwild-type mice;. Lane2:apoE -/- mice. Upper arrow, expected sizeof350 bp GAPDH; Lowerarrow pointsto Human platelets the303 bp F11R fragment.B)Real-time RT-PCR:F11R mRNA levels measureinthe proximalaorta of five separate apoE-/- mice(ApoE-/- )and Platelets were isolatedasdescribed (1). Collagen (nativetype 1), five separate C57BL/6 wild-typemice (WT). Measurements were per- used as the agonist, waspurchasedfrom Chrono-logCorp. (Ha- formed in triplicatefor each animal foreach of thetwo groups.Values vertown, PA,USA). representthe mean ±SEM. *P <0.05.

273 Babinska et al.F11R/JAM-Aexpression in atherothrombosis

Figure2:Expression of F11R/JAM-A in atheroscleroticplaques of apoE-/- mice: Immunofluorescence. A) Control, stainingofthe proximalaorta of wild-type mice by murine F11r/JAM-1 antibody.Scale, 100 µm.B)ApoE -/- ,staining of proximalaorta of apoE-/- micebymurine F11r/JAM-1 antibody.Scale, 100 µm.Rectangular inset outlinesaregionofthe lesioned area displayedathigher magnificationinpanelC.C)ApoE -/- , insetinpanelB,shown at highermagnification, depicts theintense stainingofatheroscleroticplaques by theF11r/JAM-1 antibody.Scale, 20 µm. ground of 12 weeks of age and control age-matchedwild-type b) Collection of human tissue samples: Patients: Surgically-ex- C57BL/6J mice, obtainedfrom JacksonLabs (Bar Harbor,MA, cised atherosclerotic specimens were obtainedatthe timeofsur- USA), were fed aregular mouse chowdiet. All procedures were geryfrom the carotid arteries, femoral arteries or aortasofseven conducted in accordance with the SUNYDownstate Institu- patients (6 males, 1female, 56 to 68 years old).Patients 1, 2, 5, tional Guidelines for Animal Use and Care. and 7had severe carotid artery stenosis in excessof80% and asymptomatic from their coronaryarterydisease. Patient 3had Quantitation of F11RmRNAbyreverse transcription significant vascularcompromise of the lowerextremityasevi- (RT)-PCRand real-time PCR denced by femoral artery occlusion and clinicalsymptomatol- RT-PCR ogy. Patients 4and 6had atherosclerotic occlusivediseasepre- Reverse transcription (RT) used oligonucleotide primers de- dominantlylocalized within the aortic wall of the vasculature. signed to the N-terminus of the murine F11r,forwardprimer, Controls:Specimens of renal arteryand aorta were obtained GTACAC TGC TCA ATCTGA CG,reverseprimerGAT GGA within12hofdeathfrom young cadaverickidneytransplant do- GGT ACAAGC ACAGT; carried-outat48 o C, 30 min withAm- nors immediatelyatthe timeoforgan harvesting fortransplan- pliTaq gold activation at 95o C, 10 minutes(min),PCR denatu- tation and used as controls (3 males, 23 to 38 years old).Inaddi- ration at 95o C, 15 seconds (sec); and anneal/extend reactionsat tion to these cadavericcontrol specimens, experiments detailed 60o C, 1min, 30 cycles. in Figure 3C utilizedplaque-free specimens obtainedfrom the aorta of the same patient and were used as internal controls . Real-timePCR All protocolswere approvedbythe SUNY Downstate Insti- a) Murine: F11R mRNA levels in mouseaorta were determined tutional ReviewBoardCommittee. All patients signed appropri- by quantitative real-time PCR with an ABIPrism 7000HT Se- ate informed consent forms prior to the surgicalexcision of the quenceDetection System(Applied Biosystems, Foster City,CA, atherosclerotic material. USA). The F11Rprimers: forward –825 TCT CCTTGG ACTCTT GATTTT TGG; reverse–909ACC CGGTGC AGT QuantitationofF11R mRNAinhuman arteries by real-time CCC TTT; probe-857: TTG CCTATA GCCGTG GATACT RT-PCR TTGAAA GAA CA.The GADPH forward primer–757: TGT Excised segmentswere placed in RNAStabilization Reagent, GTC CGT CGT GGA TCT GA, and the reverse primer-837: RNA l ater (Ambion,Austin,TX, USA).TotalRNA wasextracted GATGCC TGCTTC ACCACC TT;probe-781: CCG CCT utilizing RNeasy ProtectMini Kit (Qiagen, Valencia,CA, USA) GGA GAA ACCTGC CAAGTA TG. Thermal cycles consisted and analyzed by real-time PCR in triplicatefor each of the seven of:1cycle at 48o Cfor 30 min, 15 sec at 95o Cand 40 cycles for1 samples. The levels of F11R mRNAwere determinedbyuse of min at 60o C. Each mRNAlevel wasexpressed as aratio to an ABIPrism 7000HT Sequence Detection System. TheF11R GAPDH.Normalization to GAPDH mRNA wasperformed ac- primersconsisted of the forward primer–740: CCGTCC TTG cording to the formula 2 - ΔΔCt provided by the manufacturer.The TAACCC TGATT, reverse primer–818: CTC CTT CACTTC resultsare expressed as the mean±SEMofthe number of obser- GGG CACTAand probe –788:TGG CCTCGG CTATAG GCA vations. AACC.The GAPDH forward primer–620: GGA CTC ATG

274 Babinska et al.F11R/JAM-Aexpression in atherothrombosis

ACCACA GTC CA,reverseprimer–738: CCAGTA GAG Slides of paraffintissue containing cross-sectionalslices of GCAGGG ATGAT, and the probe –675: ACGCCA CAGTTT the proximalaorta of mice were processed for deparaffinization, CCCGGA GG. Thermal cycles consisted of:1cycle at 48o Cfor incubated with M.Ab.F11 or goat anti-mouse JAM-1 polyclonal 30 min, 10 min at 95o Cand 40 cycles for15sec at 95o C, 1min at antibody, labeledwith secondaryFITC conjugated antibodies, 60o C. Each mRNAlevel wasexpressed as aratio to GAPDH. and mountedwith Prolong Gold anti-fade reagent. The mRNA levels were calculatedusing astandard curve of The BioradMRC 1024ES confocal microscopysystem was RNAisolated from cultured HUVEC cells utilizing the ABI used to generate fluorescent images. CulturedHUVEC (pas- Prism 7000SDS Software (AppliedBiosystems). sages 1–3), growntoconfluencyonglass coverslips, were treated with purifiedhuman recombinant TNFα ,vascular en- Statisticalanalysis dothelialgrowthfactor (VEGF)(R&D Systems), and INFγ The data were analyzed by Student’s t-test and the non-paramet- ricMann-WhitneyU-test.Differences were considered significant at P<0.05.

Histology, immunofluorescence and confocal microscopy Human specimens of atherosclerotic plaque were subjected to standard tissue processing and staining with hematoxylin and eosin (H&E) (21,22). Pretreatedparaffin-embeddedhuman arterial segmentswere deparaffinizedand prepared for immunos- taining with primaryrabbit polyclonalanti-human vonWille- brand Factor (vWF;Chemicon International, Temecula, CA, USA) or murinemonoclonalanti-human F11R antibodies, Sec- ondaryantibodies consisting ofAlexa Fluor 568 conjugatedgoat anti-rabbit IgG (red color)orFITC-conjugated goat anti-mouse IgG (green color) were mountedwith ProLong RGold anti-fade reagent (Molecular Probes, Eugene, OR, USA) followedby A) examination by confocal scanning lasermicroscopy. Human umbilical vein endothelialcells (HUVEC), grownon coverslips in completegrowthmedia(2% FBS), were exposed to cytokines for24hat 37°C, incubated for1.5 hwith murine anti- humanF11R M.Ab.F11 (25 µg/ml)orgoat anti-murine JAM polyclonalantibody (15 µg/ml), fixedinparaformaldehyde (4%), labeledwith goatanti-mouse Ig (1:200) or rabbit anti-goat (1:100) FITC conjugated secondaryantibodies, and mounted with ProLong RGold.

Figure3:Enhancedexpression of F11R/JAM-A mRNAin humanatheroscleroticplaques. A) Real-timeRT- PCR performed on B) human arteries.Control,F11R mRNA analyses of separate plaque-free kidneyaorta specimensobtainedfromthree cadavericdonors.Real-time RT-PCR wasperformed threetimes,intriplicate, foreach sample. Pa- tients, F11R mRNA analyses on arteries obtainedfromseven separate cardiovascular patients(1thru7). Valuesrepresent the mean ±SEM. *P<0.05. B) Comparison of F11R mRNA expression by real-time RT- PCR: To talvaluesfor patientswereincludedfor comparison to control values of cadavericdonors.Valuesrepresent the mean ±SEM, *P <0.05. C) Enhanced F11R mRNA expression is associatedwithregions containing plaques. To talRNA wasextracted from four separateexcisedseg- mentsofthe same aorta(6cm) from arepresentativepatient (number 6),and analyzedseparately, in triplicate, by real-time RT-PCR.Bar 1, 1 st plaque-free segment takenfrompatient’s aorta; Bar 2, Separate adjacent, plaque-free segment of patient’s aorta; Bar 3, Aortic segment of patient containing plaques, Bar 4, Aseparateaorticsegment of thepatient containing largest amount of plaques. Thecadavericcontrol bar represents thelevel of F11R mRNA (mean±SEM)measured in threecadaveric plaque-free aortic and renal arteryspecimens.Valuesrepresent the mean ±SEM. *P<0.05. C) Fig. 3C in this reprint has been corrected due to aprinting errorinthe original February2007 issue.(Erratumwill follow.)

275 Babinska et al.F11R/JAM-Aexpression in atherothrombosis

Figure4:Histologicalstaining of atheros- cleroticplaqueofcarotid arteries. A) Histologicalstaining (hematoxylin-eosin) of across-sectionofthe carotidarteryofpatient 5. Thewhitearrowsindicatethe boundariesof plaque distributionwithin layers of theintima and media. Theblack arrows point to thefi- brouscap of theplaqueindicating theexten- sion of theplaqueintothe lumenofthe vessel. Scalebar,500 µm.B)Magnifiedview(100-fold) of aregion(seeasterisk in panel A) of the plaque.Scalebar,100 µm.

Figure5:F11R/JAM-A expression in human arteries. A) Immuno- F11R/JAM-A (green color)ofthe endothelium of aortaobtainedfrom fluorescence stainingfor F11R (green)ofanatheroscleroticcarotidar- cadaverickidneytransplant donors at thetime of organ transplantation. teryspecimen(obtainedfrompatient 5).Scalebar,150 µm.B)Immuno- Scale bar,150 µm.E)Control stainingfor theendothelium marker, vWF fluoresence stainingofthe endothelial cell markervWF (red). Scale bar, (red color) of theaorta obtainedfromcadaverickidneytransplant 150 µm.C)Merged image showing theimmunofluorescence stainingof donoratthe time of organ harvesting.Scalebar,150 µm.F)Merged F11R and vWFwithin theatheroscleroticartery. Similar immunofluor- image of panels Dand Eofthe controlaorta showingco-localizationof escence results forthe detection of F11R/JAM-A and vWFwereob- F11R/JAM-A and vWFinthe endothelium. Similar results were obtained tainedwithplaquespecimenobtainedfromthree atheroscleroticpa- withtwo additional specimens of renal arteries obtainedattime of tients. Scale bar,150 µm.D)Control immunofluorescence stainingfor organ harvesting,fromcontrol transplant donors. Scale bar,150 µm.

276 Babinska et al.F11R/JAM-Aexpression in atherothrombosis

Figure6:F11R/JAM-A expression in human aortic endothelial Figure7:F11R/JAM-A expression in HUVEC treated simulta- cells(HAEC) treated simultaneously withcytokinesTNF α and neously withcytokinesTNF α and INFγ . Real-timePCR:F11R INFγ . Real-timePCR wasperformed on HAEC treated withcombined mRNA was obtainedfromHUVEC treated withcombinedTNFα (100 TNF α (100 u/ml) and INF γ (200u/ml) for3,6,12and 24 h. Untreated, u/ml) and INF γ (200 u/ml) for4,8,12, and 24 h. Controlcells were control,HAECweremaintainedateach time pointinthe absence of cy- examinedinthe absence of cytokines. Real-timePCR wasperformed tokines. Real-timePCR wasperformed threetimes in triplicatefor each twotimes in triplicatefor each time point.Valuesare themean±SEM, time point.Valuesare themean±SEM.*P<0.05. *P<0.05.

(Roche Diagnostics).HUVEC were fixedwith 1% paraformal- nificantly-enhanced expression of F11R mRNA in the proximal dehyde at 27° C, 10 min,incubated with M.Ab.F11 for 1hat aortae(lane 2),ascompared to the barely-detectable levels of 37° C, followedbyincubation with FITC conjugatedsecondary F11R mRNAmeasured in controls (lane 1). Quantitative confir- antibodiesfor 1h.Carbocyanine monomerTO-PRO-3 wasused mation of these findings wasprovided by conducting real-time for nuclear staining (Molecular Probes). RT-PCR, and as shown in Figure 1B,the levelofF11R mRNA in the aorta of the apoE -/- micewas two-foldgreater thanthat ob- Results servedinwild-type mice. Immunofluorescence of identical specimens utilizing F11R specificantibodies revealed intense Studies of Apo E-deficientmice staining of the F11R protein in the proximalaorta of apoE-/- mice At three months of age, apoE-/- -deficient micedemonstrate the (Fig. 2B) whichfollowedthe exact contoursdisplayedbythe presence of well-developedplaquesalong their vasculature.As atherosclerotic region; low, background staining wasobservedin shown in Figure 1A, the aorta of apoE-/- mice demonstrated asig- the aorta of wild-typemice(Fig. 2A).Amagnifiedviewofapor-

Figure8:Intercellular localization of F11R/JAM-A in intact, non-treatedHUVEC. A) Immunofluorescence stainingofF11R in cultured HUVEC using F11R antibody,M.Ab.F11. B) StainingofHUVEC nuclei with To -Pro-3.C)Merged image combining thegreen immunofluorescence of F11R localizedtothe plasma membrane withthe red-nuclear staining. Scalebar,100 µm.

277 Babinska et al.F11R/JAM-Aexpression in atherothrombosis tion of the atherosclerotic region from apoE -/- mice(Fig. 2C) plaque revealed that approximately80% of the lumen of the caro- demonstrates the highlyover-expressed F11R protein found tidofpatient5was obstructedbythe plaque (Fig. 4A), andupon within the plaque of these atherosclerotic-prone animals. closerinspection (Fig. 4B) consistedofalipid-laden core with in- filtrating cells,connective tissueelements and foam cells.Immu- Studies of human subjects nofluoresence stainingofthe atheroscleroticaorta of the same In parallelstudies,F11R mRNAlevels were measuredinend- patient revealed ahigh levelofstaining of F11R throughoutthe arterectomy specimensofarteries from atheroscleroticpatients. entire specimen (Fig. 5A). Staining for vWFinthe same speci- Figure 3A showsthatF11R levels were found to be significantly men demonstratedthe presenceofthe atheroscleroticendothe- higherinthe lesionedregionsofsix of the sevenpatients examin- lium (Fig. 5B). The merged image, showninFigure 5C,depicts ed than in control specimen,obtainedfrom cadaveric kidneydo- the significantdistributionofthe F11R protein throughout the nors. Afive-fold increaseinthe overall levelofF11R mRNAwas plaque as it is surroundedbythe atheroscleroticendothelium, as demonstratedinthe plaquesofatherosclerotic arteries as com- shownbythe endothelial marker, vWF. In contrast, control speci- pared to controls (Fig. 3B). In addition, we compared thelevel of men of the aortaobtainedfrom cadaveric transplant donors, F11R mRNAwith plaquecontent in segments, obtainedfrom the identified as plaque-free, showed theF11R stainingonlywithin same aorta, that were identifiedasplaque-free,incomparisonto the endothelium (Fig. 5D). Stainingfor vWFdemonstrated its plaque-containing regions. As showninFigure 3C,regionsident- presencealsowithin the endothelium of the control aortic speci- ifiedtocontain high levels of atherosclerotic plaques (bars 3and men (Fig. 5E).The merged immunofluoresenceimage of the con- 4) were found to exhibitsignificantlyhigher levels of F11R trolaorta (Fig. 5F)depictsco-localization of the F11R/JAM-A mRNA than twoplaque-free segments obtainedfrom the same protein andvWF togetherwithin the endothelium of the aorta. aorta of this patient (bars 1and 2). Histologicalexaminations conductedtodemonstrate thepresenceofthe atherosclerotic Investigationsofhuman endothelialcells in culture In studiesdesignedtodetermine the potentialcauses of the in- creased levels of F11R in atherosclerotic plaquesobserved in vivo,weexamined the effects of the treatment of EC with inflammatorycytokines on the interactions betweenplatelets and EC. In previousstudiesweexposed human EC to onlyasingle type of cytokine. However, acombined cytokinetreatment is known to causeanevengreater redistribution of the F11R from tight junctions to the luminalsurfaceofEC. Therefore, in the present study,our experiments specificallyfocusedonthe exposure of EC to acombination treatment using the cytokines TNF α and INFγ .Asshown in Figure 6, the simultaneous addition of both cytokines to cultured humanaortic EC (HAEC) resulted in significant upregulation of F11R mRNA thatremained significantly elevatedfor 24 hfollowing the initial application of the cytokines.Similartoaortic cells,HUVEC alsodemonstrated significantlyincreased levels of F11R mRNAwith time following the simultaneous addition of both cytokines (Fig. 7),al- though differences in sensitivity to the cytokines between these twocelltypes were observed. Differencesinsensitivity to cytokine stimulation between aortic endothelialcells and HUVEC could not be explained by different passages, as the same cell passage wasutilized forboth cell types in these experiments. Immunofluoresencestaining of non-treated HUVECreveal- ed the presence of F11Rwithin the plasma membranes of these cells (Fig. 8A); counterstaining with anuclear stain revealed the HUVECnuclei(Fig. 8B).The merged image shown in Figure 8C demonstrates the association of F11R onlywith plasma membranes;the clear regions of the HUVEC cytoplasm depictthe Figure9:Localization of F11R/JAM-A following thecombined lackofstaining for F11Runder these control conditions. Such treatmentofcultured HUVEC withTNFα plus INF γ . A) Intact, merged images were utilized to analyze changes in F11R immu- non-treated HUVEC:Merged image showing theimmunofluorescent nofluorescence caused by the treatment of confluentHUVEC stainingofnon-treated HUVEC withF11R antibody,M.Ab.F11, showing cultures withthe combination of cytokines.Asshown in Figure intercellular stainingofF11R at tight junctionswithco-staining of nuclei 9, such simultaneous cytokinetreatment caused substantial by To -Pro-3.Scalebar,10 µ m. B) Cytokine-treated HUVEC:Merged changes in the localization of F11R. In comparisontothe non- image showing thediffuse immunofluorescent stainingpattern of F11R throughout theentirecellfollowing thetreatment of HUVEC for24h treatedHUVEC (Fig.9A), thetreatment of HUVECwith the cy- withacombinationofcytokinesTNFα (100 u/ml) and INF γ (200 u/ml). tokinesTNF α plus INF γ for 24 hresultedinadecreased intensity NucleistainedbyTo-Pro-3.Scalebar,10µm. for F11Rstaining at membrane intercellular junctions.Asshown

278 Babinska et al.F11R/JAM-Aexpression in atherothrombosis in Figure 9B,following the combined cytokinetreatment of treated with TNF α and INFγ ,whereashuman platelets were ac- HUVEC, the F11Rprotein becameevenlydistributedthrough- tivatedbyexposure to collagen(Fig. 10B). We determinedthat out the entire cell body of the HUVEC rather than concentrated approximately 14.8 ±2.49 X10 5 of collagen-activatedplatelets withintight junctions. bound to 1X10 4 of the TNFα -plus INFγ -treated HUVEC(control bar, Fig. 10B). The presenceofthe soluble recombinant Investigationsofplatelet adhesiontoECinculture F11R protein (sF11R;10µg/ml)significantly inhibitedthe ad- To examine the contribution of F11R to the force of adhesion hesion of activatedplatelets to TNF α -plus INFγ -treated operating between plateletsand inflamedendothelial cells, HUVEC. Furthermore,the presenceofF11R peptide 1resulted we examinedthe effects of specific F11R-inhibitors on the in significant inhibition of plateletadhesion to the inflamed adhesion of platelets to HUVEC treatedsimultaneously with HUVEC as didthe presenceofF11R peptide 4. In contrast, the cytokines TNF α plus INF γ .Wehavepreviouslyreported that addition of apeptide whose sequence of amino acids cor- peptides designated F11R-peptide-1 (amino acid sequence responded to F11R peptide 1but wasscrambled by random in- 28SVTVHSSEPEVRIPENNPVKLSC50 present in the N-terminus of F11R) and F11R-peptide4(amino acidsequence 97KSVTREDTGTYTC109 present within the 1 st Ig-fold of F11R),act as specificinhibitors of plateletaggregation induced by M.Ab.F11. As shown in Figure 10A, we determinedthat 3.32 ±0.63 X10 5 intact plateletsadhered to 1X10 4 TNFα -plus INFγ -treated HUVEC(control bar). Theadhesion of intact platelets to the cytokine-treated HUVEC wasinhibited significantly by the presence of the solublerecombinantF11R (sF11R)aswell as by F11Rpeptides1and 4. The participation of F11R wasmonitored further under conditions in whichcytokine-treatedHUVEC were reacted with activatedplatelets. Underthese conditions, HUVEC were pre- A)

Figure10: Inhibition of platelet adhesiontoHUVEC treated withcombination of cytokinesTNF α and INFγ . A) Non-activated platelets:Isolatedplatelets were appliedontoHUVEC previouslytreated withcytokinesTNFα (100 u/ml) and INF γ (200 u/ml).Control,noaddi- tions: Platelet adhesion to cytokine-treated HUVEC in theabsence of F11R inhibitors. Value setat100% represents thebindingof3.32 ±0.63 X10 5 non-activated platelets per1X10 4 TNF α -+INFγ -treated EC. Plus sF11R: Platelet adhesion to TNF α -plus INF γ -treated HUVEC in the presence of 10 µg/ml solublerecombinant F11R (sF11R). Values, mean ± SEM,*P<0.05. Plus F11R peptide 1: Platelet adhesion to TNF α -+INFγ - treated HUVEC in thepresence of 500 µM peptide 1. *P<0.05. Plus F11R peptide 4: Platelet adhesion to TNF α -+INFγ -treated HUVEC in thepresence of 500 µM peptide 4. *P <0.05. B) Activated platelets: Isolated platelets were activated withcollagen (10g/ml) and applied onto TNF α (100 u/ml) plus INF γ (200 u/ml)-treated HUVEC.Control, no additions: Adhesion of activated platelets to cytokine-treated B) HUVEC,inthe absence of F11R inhibitors. The100% value represents binding of 14.8 ±2.49 X10 5 activated platelets per1X10 4 treated HUVEC.Plus sF11R: Adhesionofcollagen-activatedplatelets to TNF α - plus INF γ -treated HUVEC in thepresence of solubleF11R recombinant protein(sF11R) (10 µg/ml).*P<0.05. Plus peptide 1: Adhesion of activated platelets to cytokine-treated HUVEC in thepresence of 500 µM peptide 1. *P<0.05. Plus peptide 4: Adhesion of activated platelets to cytokine-treated HUVEC in thepresence of 500 µM peptide 4. *P<0.05. Plus scrambled peptide 1: Adhesion of activated platelets to cytokine- treated HUVEC in thepresence of 500 µM scrambled peptide 1. *P<0.05, significantly different from F11R peptides1,4and sF11R.C) Adhesion of activated platelets toVEGF-treated HUVEC.Control,no additions:The adhesion of collagen (10 µg/ml)-activated platelets to HUVEC treated withVEGF(10 ng/ml) for24hat 37o C.The100% value represents adhesion of 10.4 ±1.7 X10 5 activated platelets to 1X10 4 VEGF-treated HUVEC. Plus sF11R (1.5 µg): Adhesion of activated platelets toVEGF-treated HUVEC in thepresence of solubleF11R recombinantprotein (sF11R)(250 µM). *P<0.05.Plus sF11R (3.0 µg): Adhesion of activated platelets toVEGF-treated HUVEC in thepresence of soluble C) F11R recombinantprotein (sF11R)(500 µM). * P<0.05.

279 Babinska et al.F11R/JAM-Aexpression in atherothrombosis sertion of amino acids during the synthesis of this peptide (31, 32), plateletendothelial cell adhesion molecule-1 (scrambled peptide 1) did not inhibit the adhesion of activated (PECAM-1) (33 –35), beta-1 integrin (30,31, 36, 37), tissue fac- platelets to the cytokine-treated HUVEC.Finally, we monitored tor (29), vonWillebrand factor (24, 30), and GPIIb/IIIa (31)in the effect of F11Rinthe adhesion of platelets to HUVEC treated this process. Recently,the F11R/JAM-A protein has beende- for 24 hwith the growth factor VEGF (10ng/ml). Figure 10C scribedwith propertiesindicating its potential role in triggering demonstrates that, whereas 10.4 ±1.7 X10 5 collagen-activated the pathogenesis of inflammatorythrombosis and atherosclero- platelets adheredto1X10 4 VEGF-treated HUVEC (control sis (6).The purpose of the study reported herewas to provide ad- bar); thepresence of sF11R(1.5–3.0 µ g/ml)resultedinthe in- ditionalevidencefor this role. hibition (byabout 35% and 50%, respectively)ofthe adhesion of The results reported here demonstrate that upregulation of collagen-activatedplatelets to the VEGF-treated EC. F11RmRNAisinducedbycytokines in EC both from venal and arterial vessels, and thatthe newly-induced F11R/JAM-A protein has asignificant role in the adhesion of platelets to the in- Discussion flamed endothelium. Thecombined cytokinetreatment utilized in the present study wascritical forthe induction of maximal Recent studieshavedemonstrated thatplatelets adhere to an in- changes in the redistribution of F11R/JAM-A to the vessel flamed endothelium even prior to the invasion of leukocytesand lumen. Therefore, these experimental conditions aremost opti- before lesions become detectable, suggesting thatthis adhesion mal for examining the roleofF11R/JAM-A in plateletadhesion is an initial step in the development of plaquesinatherosclerosis to EC under inflammatoryconditions, that areknown to involve (23–25). Theresults of the studyreported hereprovide newin- the action of both TNF α and INFγ (38). Experiments utilizing sightsinto the molecularmechanisms by whichinflammatory specificF11R/JAM-A inhibitors determined thatover50% of agents initiate the interaction of platelets with endothelialcells, the force of plateletadhesion to EC inflamedbyexposure to leading to formation of thrombotic plaques. Undernormal [TNF α +INFγ] is exertedbyF11R/JAM-A.Weconclude that physiological conditions,endothelial cells (EC) arenon-throm- F11R/JAM-A-mediated adhesion of platelets to EC is acritical bogenic (26–28).However,the non-thrombotic surfaceofEC event in the inflammatoryprocessthat leads to thrombosis in can be transformed to aprothrombotic surface followingtheir non-denuded vessels. We propose that circulating platelets, exposure to inflammatoryagents, resulting in procoagulant ac- whichconstitutively express the F11R/JAM-A molecule on their tivity and predisposition to thrombosis (28–30).Previousstudies surface(1) adhere directlytothe inflamed endothelium by virtue of the adhesion, accumulation and recruitment of platelets onto of their ability to interact with F11R/JAM-A moleculesthat are the inflamedsurfaceofcytokine-stimulatedEChaveimplicated newlyexpressed on the exposed surfaceofinflamed EC. This P-selectin (24,31), the P-selectin glycoprotein ligand (PSGL-1) homologous interaction (6) could cause the initial adhesion of platelets to endothelial cells in an earlystage,thus triggering the process of inflammation-inducedatherogenesis. Figure 11 depicts aschematic presentation of the proposedsteps of this process, whichisinitiatedbythe activation of the transcription factor NF-κ BinECbycytokines, followedbythe interaction of this transcription factor with specificbindingsegments for NF- κ B present in the promoter region of the F11R gene (9).This se- quenceofevents is consistent with the reports that platelets ad- heretothe inflamedendothelium even prior to the invasion of leucocytes and the detection of lesions(23–25). Platelet thrombi arethen stabilizedalso by other CAMs and integrins (24,29–37) areformed and leadtothe development of atherosclerotic lesionsand atherothrombosis (23).Figure11also serves to point out specific molecularevents, to be studied in vivo under flow conditions,inorder to provide final proof for the role of F11R in initiating and/or promoting atherogenesis and atherosclerosis. In parallel experiments of the study reported here, we have begun to examine the involvement of F11R/JAM-A in the formation of atherosclerotic plaques in vivo.Inthe first of these Figure11: Schematicpresentation of the role of F11R/JAM-A in triggering inflammatorythrombosis with theformation of pla- experiments we demonstrated ahighlyenhanced expression of teletthrombi underlying thepathogenesis of atherosclerosis. F11R/JAM-A in the aorta of 12-week-old apoE-/- mice, an ad- Sequence of events in inflammatory thrombosis that involvethe upregu- vancedage in whichatherosclerotic plaquesare well established lationofF11R-mRNAand induction of theF11R proteinincytokine-in- (39). Theexpression of F11R/JAM-A on the atherosclerotic en- flamedEC, leadingofthe initial formationofaplatelet thrombus on the dothelium of carotid arteries of apoE-/- wasobservedpreviously luminal surface of inflamed EC witheventual progression to the devel- by others(18). However, these animals were studiedatanearlier opment of atheroscleroticplaques.THe arrow(↓ )indicates causal effect. Theexperimental evidence indicating this chain of events and therefer- stageofdevelopment (6 weeks)and fedanatherogenic dietwith encestothe studiesreporting this evidence areoutlinedinthe textof ahigh fatcontent. Since the mice in our study were maintained the Discussion . on aregular chowdietwithoutahigh-fat content, we conclude

280 Babinska et al.F11R/JAM-Aexpression in atherothrombosis thatthe advancing age of these animals is causing aspontaneous ture. Theresults of the in-vivo studieswith animals and human upregulation of F11R mRNA-transcripts and induction of patients areconsistent with the in-vitro data on the criticalrole of F11R/JAM-A protein expression in cellular components of the F11R/JAM-A in the initial adhesion of platelets to inflamedEC. plaquesdeveloping in these animals. Thus,the sequence and structureofthe peptides and recom- Similar to the results with apoE -/- mice, ourexamination of binant protein whichspecificallyantagonize the action of F11R F11R/JAM-A expression in plaquesdeveloped in vivo in the ar- canprovide the basis for the design of noveldrugs for the preven- teries of atherosclerotic humanpatients revealed abnormally- tion and treatment of atherosclerosis, thrombosis, heart attacks high levels of F11R/JAM-A mRNAand F11R/JAM-A protein.A and stroke. recent study(20) of forty-six human atherosclerotic patients reported of enhancedexpression of F11R/JAM-A in areasofinsta- Acknowledgements bility of the atherosclerotic plaque,which areassociatedwith the This study wassupported by NIHgrant #75586 (toEK),NIH grant #64735 rupture of plaquesand with embolic complications that canlead (toX-CJ), and PSC-CUNY Research Awards (toYHE). to myocardialinfarcts and/or occlusions in the cerebralvascula-

References 1. KorneckiE,WalkowiakB,Naik UP,etal. Acti- 14. Printeseva OY,PecloMM, Gown AM.Various cell 26. Cines DB,PollakES, Buck CA,etal. Endothelial vationofhuman plateletsbyastimulatorymonoclonal types in human atheroscleroticlesions express cells in physiologyand in thepathophysiologyofvas- antibody. JBiolChem 1990; 265:10042–10048. ICAM-2. Further immunocytochemical and immuno- cular disorders.Blood 1998; 91: 3527–3561. 2. SobockaMB, SobockiT,BabinskaA, et al. Signaling chemical studies employingmonoclonal antibody 27. Karsan A, Harlan JM. The blood vessel wall. In: pathwaysofthe F11 receptor (F11R; a.k.a. JAM-1, 10F3. Am JPathol 1992; 140: 889–896. Hematology: BasicPrinciplesand Practice, 3 rd Ed. JAM-A in human platelets:F11R dimerization, phos- 15. Zibara K, Chignier E, Covacho C, et al.Modulation Elsevier,Oxford, 2000. 1770–1782. phorylation and complexformation with the integrin of expression of endothelialintercellular adhesion mol- 28. MayAE, NeumannFJ, Preissner KT.The relevance GPIIIa. JRecept SignalTransduct Res 2004; 24: 85–105. ecule-1, platelet-endothelialcelladhesion molecule-1, of blood cell-vessel wall adhesive interactions for vas- 3. Wang F, Naik UP,Ehrlich YH, et al.Stimulatory and vascularcelladhesion molecule-1 in aortic arch cularthrombotic disease. ThrombHaemost 1999; 82: antibody-induced activationand selectivetranslocation lesionsofapolipoproteinE-deficient compared with 962–970. of proteinkinase Cisoenzymes in human platelets. wild-typemice. ArteriosclerThromb Vasc Biol 2000; 29. DiquelouA, DupoyD,Gaspin D, et al.Relationship Biochem J1995; 311: 401–406. 20: 2288–2306. between endothelialtissue factor and thrombogenesis 4. Naik UP,Ehrlich YH, Kornecki E. Mechanismsof 16. Sobocka MB,Sobocki T, Banerjee P, et al.Cloning under blood flowconditions. ThrombHaemost 1995; platelet activationbyastimulatory antibody: cross- of thehuman platelet F11 receptor: acelladhesion mol- 74: 778–783. linking of anovel platelet receptor for M.Ab.F11 with eculemember of theimmunoglobulin superfamilyin- 30. André P, DenisCV, Ware J, et al.Platelets adhere to theFc γ RIIreceptor.Biochem J1995; 311: 155–162. volved in platelet aggregation. Blood 2000; 95: and translocateonvon Willebrand factor presented by 5. Babinska A, Kedees MH, Athar H, et al.Two re- 2600–2609. endotheliuminstimulatedveins. Blood 2000; 96: gionsofthe human platelet F11-receptor (F11R) are 17. Ozaki H, IshiiK,Horiuchi H, et al.Combined treat- 3322–3328. critical for platelet aggregation, potentiationand ad- ment ofTNF-α and INF-γ causes redistributionofjunc- 31. Frenette PS, Johnson RC,Hynes RO, et al.Platelets hesion. ThrombHaemost 2002; 87: 712–721. tional adhesion moleculeinhuman endothelialcells.J rollonstimulatedendothelium in vivo:aninteraction 6. BabinskaA, Kedees MH,Athar H, et al.F11-recep- Immunol 1999; 163: 553–557. mediated by P-selectin.PNAS1995; 92: 7450–7454. tor(F11/JAM) mediates platelet adhesion to endothe- 18. Ostermann G, Fraemohs L, BaltusT,etal. Involve- 32. Frenette PS, DenisCV, WeissL,etal. P-selectin lialcells:roleininflammatory thrombosis. Thromb ment of JAM-A in mononuclear cellrecruitment on in- glycoproteinligand 1(PSGL-1) is expressed on pla- Haemost2002; 88: 843–850. flamed or atherosclerotic endothelium. Inhibitionby telets and can mediateplatelet-endothelialinteractions 7. Kedees MH, Babinska A, Swiatkowska M, et al. soluble JAM-A.Arterioscler ThrombVasc Biol 2005; in vivo.JExp Med 2000; 191: 1413–1422. Expression of arecombinant proteinofthe platelet F11 25: 729–735. 33. GurubhagavatulaI,AmraniY, Practico D, et al.En- receptor (F11R) (JAM-1/JAM-A)ininsect cells:F11R 19. PajukantaP,Lilja HE, Sinsheimer JS, et al.Familial gagement of human PECAM-1 (CD31) on human en- is naturallyphosphorylated in theextracellular domain. combined hyperlipidemia is associated with upstream dothelialcells increases intracellularcalcium ioncon- Platelets2005; 16: 99–110. transcriptionfactor 1(USF1). Nature Genet 2004; 36: centration and stimulatesprostacyclin release. JClin 8. Wagner DD,Burger PC. Plateletsininflammation 371–376. Invest 1998;101: 212–222. and thrombosis. ArteriosclerThromb Vasc Biol 2003; 20. Papapyridonos M, Smith A, Burnand KG, et al. 34. RosenblumWI, Nelson GH, WornleyB,etal. Role 23: 2131–2137. Novelcandidategenes in unstable areas of human athe- of platelet-endothelialcelladhesion molecule 9. Sobocki T, Sobocka MB,Babinska A, et al.Ge- roscleroticplaques. ArteriosclerThromb Vasc Biol (PECAM) in platelet adhesion/aggregationoverin- nomic structure, organizationand promoter analysis of 2006;26: 1837–1844. jured butnot denuded endothelium in vivo and ex vivo. thehuman F11R/F11 receptor/junctional adhesion 21. Prophet EB,Mills B, ArringtonJB, et al., eds. Stroke1995; 27: 709–711. molecule-1/JAM-A.Gene 2006; 17: 128–144. Washington, DC: AmericanRegistryofPathology; 35. Sun J, Williams J, Ya nH-C,etal. Individually dis- 10. Bradley JR,Pober JS. Prolonged cytokine exposure 1992. ArmedForces Institute of PathologyLaboratory tinct Ig homologydomains in PECAM-1 regulate causes adynamic redistributionofendothelialcellad- Methods in Histotechnology. homophilic binding and modulatereceptor affinity.J hesion molecules to intercellularjunctions. Lab Invest 22. Rogers WJ,Prichard JW,HuY-L,etal. Characteriz- Biol Chem1996; 271: 18561–18570. 1996;75: 463–472. ationofsignal propertiesinatheroscleroticplaque 36. BombeliT,Schwartz BR,Harlan JM.Endothelial 11. Romer LH,McLean NV,Yan HC, et al.IFN-gamma componentsbyintravascularMRI.Arterioscler cells undergoing apoptosis becomeproadhesive for and TNF-alpha induce redistributionofPECAM-1 ThrombVasc Biol 2000; 20: 1824–1830. nonactivated platelets. Blood 1999; 93: 3831–3838. (CD31) on human endothelialcells.JImmunology 23. GawazM,Langer H, MayAE. Platelets in inflam- 37. RomoGM, Dong JF,Schade AJ,etal. Theglyco- 1995; 154: 6582–6592. mation andatherogenesis. JClin Invest 2005;115: proteinIb-IX-Vcomplex is aplateletcounterpartfor 12. O’BrienKD, AllenMD, McDonaldTO, et al.Vas- 3378–3384. P-selectin.JExp Med 1999; 190: 803–813. cular celladhesion molecule-1 is expressed in human 24. TheilmeierG,MichielsC,Spaepen E, et al.En- 38. Yo ung JL, LibbyP,Schonbeck U. Cytokines in the coronaryatheroscleroticplaques. Implications for the dothelialvon Willebrand factor recruits plateletsto pathogenesisofatherosclerosis.Thromb Haemost mode of progressionofadvanced coronaryarterydis- atherosclerosis-prone sitesinresponsetohypercholes- 2002; 88: 554–567. ease. JClin Invest 1993;92: 945–951. terolemia.Blood 2002; 99: 4486–4493. 39. PlumpAS, Smith JD,Hayek T, et al.Severe hyper- 13. Poston RN,HaskardDO, Coucher JR,etal. Ex- 25. MassbergS,BrandK,Gruner S, et al.Acritical role cholesterolemia and atherosclerosis in apolipoprotein pression of intercellularadhesion molecule-1 in athe- of platelet adhesion in theinitiationofatherosclerotic E-deficient micecreatedbyhomologous recom- roscleroticplaques. Am JPathol 1992; 140: 665–673. lesion formation. JExp Med 2002; 196: 887–896. binationinEScells.Cell1992; 16: 343–353.

281