Imaging Vascular Endothelial Activation: An Approach Using Radiolabeled Monoclonal Antibodies Against the Endothelial Cell Adhesion Molecule E-

Edward T.M. Keelan, Andrew A. Harrison, Peter T. Chapman, Richard M. Binns, A. Michael Peters and Dorian 0. Haskard

Departments ofMedicine and Diagnostic RadiOlOgy, RoyalPostgraduate Medical Schoo4 Hammersmith Hospita4 London, and Immwwlo@ DivLcion, AFRC Babraham Institute, Babraham, Cambridge, United Kingdom

verthelastfewyearstherehasbeena greatincrease E-selectin is an endothelial Cell-SpecifiC adhesion molecule for in our understandingof the active role played by endothe leukocytes expressed on the luminal surface of vascular endo hum in the orchestration of inflammatoryresponses. In thelium during inflammatory responses. Because E-seleclin ex particular, endotheial cells (EC) respond to inflammatory presalon is dependent upon ongoing stimulationby cytokines, stimuli by de novo expression of a number of surface thismoleculeoffersa potentiallyusefultargetforimaglngtissues antigens and soluble mediators (1). Because of the imme in disease states invoMngcytoidne-mediatedendOthelialcell diate accessibility of endotheium to the blood, this new activation.Method: To assess the imagingpotentialof an anti understandingof the nature of endotheial surface activa E-selectin monodonal anffbody (Mab) 1.2B6, the accumulation tion antigen expression in inflammationoffers attractive @ of intravenoualyinjected 11ln-labeledMab 12B6 was corn pared to that of 111I@@@ antibodyina modelofarthritisinthe possibilities for imaging. Thus, radiolabeled monoclonal pig- Injectionof phytohaemagglutinin(PHA)into a knee led to antibodies (Mab) against EC activation antigens might be E-selectinexpres@onon vessels in the syrx@umand draining useful for the noninvasive evaluation of endotheial activa deep inguinallymphnodes, as demonstrated by immunohiad tion in diverse clinical situations. Such precise molecular ogy. No E-selectin expression was seen in the control knee targeting would represent an advance over techniques that injectedw@ibufferalone. Mimals were gwen 1111n-Mab1.2B6 nonspecifically image inflammation such as 67Ga-citrate or @ or antibodyintravenously3 hr afterthe intra-articu polyclonal IgG. 1w injection of PHA RadiOlabeledantibody uptake was mea E-selectin (endotheial-leukocyte adhesion molecule-i, sured by directcountingof tissues 25 hr postmortem.Results: ELAM-i) is an EC activation antigen which acts as an The accumulation of radlOiabeledcontrol IgG in syncMum and adhesion molecule for the recruitment of circulating leuko drainingdeep inguinallymphnodes of PHA-injectedknees was cytes into the tissues during inflammatory responses (2,3). aignificantlyhigher than accumulation in tissues injected w@i bufferalone; however,the comparable ratios in animals receiv It is a single-chain that is not expressed by ingradiolabeledMab 1.2B6were significantlygreater. Sdntigra resting EC but which is induced in vitro following stimu phy performed 24 hr after 1111n-Mab1.2B6 injectionshowed lation of EC by proinflammatorymediators such as inter obvious localizationof activityin the inflamedknee in each of leukin-i, tumor necrosis factor or bacterial lipopolysaccha three animals. Conclusion: RadiOlabeledantl-E-selectinMab ride (4—6).As E-selectin expression is limitedto activated can be used to image localizedinflammatorytissues. This ap endotheium (7,8), this molecule could be an ideal imaging proach may be usefulfor investigatingactivatedendotheliumin target for diagnostic radioimmunoscintigraphy. human disease. We have recently demonstrated (9) that intravenously Key Words: Inflammation; vascular endothellal activation; injected “1In-labeledanti-E-selectin Mab is taken up into endothellal cell adhesion; E-selectln inflammatory sites in pig skin stimulated with interleukin-i, tumor necrosis factor or phytohaemagglutinin (PHA) (10), J NuciMed1994;35:276-281 correlating with leukocyte accumulation and endothelial expression of E-selectin observed immunohistologically (9,11). Thisuptakeof anti-E-selectinwasspecificasthere wasnoincreaseinuptakeof @Tc-labeledcontrolIgOinto ReceivedJun. 29, 1993;revisIonaccepted Nov.5, 1993. the same inflammatorysites. In this study, we utilized this Forcorrespondenceandreprintscon@ Dr.D.O.Haskasd,Dept ofMedidne, R.P.M.S.,Ha@imersmlihHos@, DuCaneRd.,LondonW12ONN,U.K technique to image a localized inflammatory response us

276 The Journal of Nudear Medicine•Vol.35 •No.2 •February 1994 ing a model of arthritisinduced by intra-articularinjection nous boluses of propofol(1 mg/kg;Diprivan;ICI Pharmaceuti of PHA (12). cain, Macclesfield, U.K.), given every 15-20 mis. The animals did notrequireintubationorotherexternalsupport. METhODS Based on previous work looking at lymphocyte traffic into localized inflammatory lesions (10), 400 @gof PHA (L-8754, Monodonal AntibOdios SigmaChemicalCo. Ltd., Dorset, U.K.) in 1ml RPMI1640was Mab1.2B6is amouseIgG@MabagainsthumanE-selectin(13). injected into the test knee and 1 ml RPMI 1641)was injected into This antibodyalso recognizesporcineE-selectin,as shownby the control side. Three hours after the intra-articularinjections, specific reactivity with COS cells transfected with porcine E-se 100 ig of either “In-labeledanti-E-selectinMab 1.2B6(three cDNA (Y. Tsang, unpublisheddata). Mab 1.2B6 was pun animals) or 111In-labeled control antibody (three animals) were fledfromtissueculturesupematantbyProteinA affinitychroma given as an intravenous bolus. Animalswere screened in the tography (14). The control reagent MOPC21 is a mouse IgG1 supinepositionundera gammacamera(IGEStarport,GEMed. myelomaprotein(15) of undefinedspecificityandwas the kind ical Systems, Milwaukee, WI) for 90 miii after antibody injection. gift of Dr. Martyn Robinson (Ceiltech Ltd., Slough, England). Initially, dynamic imaging of the lower abdomen and hind limbs wasperformedusinga 1-rainframeratefor60mis.Following Antibody LabelIng with 111ln this, a 15-mm static image was obtained. A final 30 mm static Antibodies were labeled with “SInusing the method described image of the lower abdomen and hind limbs was obtained 24 hr by Hnatowich et al. (16). The antibodies were first coupled with after the intra-articular injections. the bicyclic anhydride of diethylenetriaminepentaacetic acid At theendof eachstudy,animalswerekilledby overdoseof (DTPA)(D 6148,SigmaChemicalCo. Ltd., Dorset, U.K.). A anesthetic.The deep inguinal,superficialinguinalandpopliteal 5-mg aliquot of antibody in 1ml 0.1 M bicarbonate buffer, pH 8.0, lymph nodes and the synovial lining of the knee joints were wasaddedtodryDTPAtogiveanantibody-to-DTPAmolarratio excisedimmediatelypostmortemand were placed in previously of approximately1:10.Afteran incubationperiodof 30 miiiat weighedcontainers.Theradioactivitywas thencountedinawell roomtemperature,the coupledantibodywas separatedfromfree counter(NM1O8; J + P Engineering, Reading, England) and back DTPAby gelfiltrationon a SephadexG50(PharmaciaLKB Bio ground counts subtracted. Accumulation of radioactivity was cx technology,Uppsala, Sweden)columnin a 20-misterilesyringe. pressed as a “localizationratio―(LR). For the lymph nodes, the The columnwas elutedusing0.1 M sodiumacetate,pH 6.0. LR was the CPM/g for each group (deep inguinal, superficial concentration was measured using a spectrophotometer inguinal or popliteal) from the PHA-injected side divided by the (Ultrospec II, Pharmacia LKB Biotechnology, Uppsala, Sweden) CPM/gof the equivalentgroup from the control side. Since the and the peak aliquotswere pooled and filteredthrough a 2-nn extent ofinflammation in the knee synovium was difficultto define microffiter(Ministart,SartoniusGmbH,Gottingen,Germany). with the naked eye, accumulated radioactivity was counted in the The DTPA-coupledantibodies(—1mg/mI)were stored in total excisedsynoviumtogetherwith a marginof connectivetis 200-/Lialiquotsat 4°Creadyforsubsequentlabeling.Onthemorn sue.TheLRofsynoviumwasthencalculatedasthetotalCPMin lag of each study, 25 MBq of chelation grade “In(indium chlo the countedtissuefromthe PHA-injectedknee dividedby the ride in 0.04 M Ha, carrier free; INS 1, Amersham International total CPMin the countedtissuefromthe controlknee. The ratio plc, Amersham,U.K.) was broughtto pH 6 by the additionof of CPM/gforsmallsynovialbiopsiesfromeachkneewas calcu 3.8%sodiumcitrate,pH7.4.Thiswasthenaddedtoanaliquotof latedforanimals2, 3, 5 and6 andthe LRvaluesdidnot differ the DTPA-coupled antibody and incubated for 30 min at room significantly from the values given for the LR calculated from the temperature Radiolabeledantibodywas separatedfromfree “In counts in the total excised tissue (data not shown). bygelfiltrationon a SephadexG50columnelutedwithPBS.The conjugation and labeling protocol yielded antibodies with 0.8—1.2 Immunohlstochemlcal Studies DTPAmoleculesper antibodymoleculeanda specificradioactiv After measuring the radioactivity in the tissues as described ityof 100—150MBq/mg.The efficiencyof proteinbindingof “In above, multiple samples were embedded in OCF compound was >90%,asjudgedby thin-layerchromatography.No change (Miles Laboratories Inc, Elkhart, IN), snap frozen and stored at in protein-bound radioactivity was observed when thin-layer —70°Cforsubsequentstaining.Czyostatsections(7-10 zm)were chromatographywas repeatedafter24 hr,demonstratinga high mountedon poly-L-lysine-coatedslides,air-driedfor2—3hrand degree of stability of the conjugates. The binding constant of fixedin 50%methanol/50%acetonefor 5 mis. Stainingwas done “In-labeledMab1.2B6was4.8nMwhenassessedbyScatchard using the alkaline phosphatase anti-alkaline phosphatase analysis using tumor necrosis factor-activated human umbilical (APAAP) method and antibody-binding was visualized using Fast vein endotheial cells. Redsubstrateby a modification(11) of a previouslydescribed AnImals method (17). The degree of inflammation and the intensity of Six healthy young large white pigs weighing 15—25kg were E-selectinexpressionwas scored blindlyby an experiencedob obtained from a commercial supplier. Animals were housed mdi server,usinga semiquantitativescaleofO(nodetectablestaining), vidually under standard husbandry conditions and studied accord + (weak staining of occasional vessels) and + + (widespread ingto a protocolapprovedunderthe UnitedKingdomAnimals moderate-strongly stained vessels). (Scientific Procedures) Act 1986. Sts@ Model of PHA-Induced Arthritis The degreeof localizationof labeledantibodiesin inflamed Animals were anesthetised for the intra-articular injections and synovia and regional lymph nodes was assumed to be normally subsequentimagingstudies.Anesthesiawas inducedusinghalo distributedbetweenpigs.Uptakeof “In-labeledMabL2B6was thaneby inhalationresultingin rapidinductionof sedationwith thereforecomparedwithuptakeof “In-labeledcontrolantibody minimal stress. Anesthesia was maintained by repeated intrave using an unpaired Student's t-test.

ImagingInflammation•Keelanat al. 277 TABLE I @ The Uptake 11ln-Labeled Mab 1.2B6 or Control Mbhody into Syno@4urnand Deep Inguinal Lymph Nodes and lmmunohistochernlcal Staining for E-Selectin at These Sites

E-seIe@lnexpressiontDeep In@lymphLo@on @ Knee DespnodesAn@nsi inguk@no. 111l@ ControlPHAConfrol1antibody Synovkim lymphnodes PHA —+++2 PAth1296 6.8 10.4 ++ —++—3 Mabl2B6 16.9 5.2 ++ Mabl2B6 16.1 10.0 —++—13.3 ++ 2.94 ±5.6 8.5 ± —++—5 ControlIgG 1.3 0.9 ++ —+++6 ControlIgG 3.0 1.4 ++ ControllgG 3.7 1.2 —++—2.7±1.2 ++ 1.2±0.3*LR

section.1@rhevalueswerecalculatedas descdbedInthe Methods observer.*Meandegree of E-selectlnexpression Infrozensections was scored blIndlyby an experienced s.d.‘p± < 0.05 (Student@st-test@compared with animals that received ‘111n-labeledcontrol 19G.

RESULTS had continued to increase and this uptake was markedly While the intra-articular injection of PHA produced no greater for the anti-E-selectin Mab than for the control external signs of inflammation, the synovial lining of PHA antibody (Fig. 2). In the case of radiolabeled Mab i.2B6, injected joints appeared inflamed when the cavity was thejoint localization was so intense as to mask the normal opened postmortem and there was often a small synovial appearance of thejoint space on the gamma camera image. effusion noted. In addition, the deep inguinal lymph nodes In addition to the greater localization in the joint itself, draining the inflamed knee appeared reactive, especially in there was also evidence of greater clearance of the Mab the posteriorportionofthe chain. No macroscopic changes i.2B6 background radioactivity at 24 hr when compared were detected in the popliteal and superficial inguinal with radiolabeled control antilxxly. This was particularly lymph nodes on the side of either the PHA-injected or the marked in one pig (animal i, Table 1), in which there was control knees. clear imagingof the ipsilateraldeep rnguinallymph nodes Immunohistochemical staining of tissue sections was by 24 hr (Fig. 2B). Figure 2B also shows imaging of the performed to confirm activation of endotheium and cx skin, visible as a distinct edge to the body outline. pression of E-selectin in all pigs used in the study (Table i). In order to validate the greater localization of radioac The synovial lining (Fig. iA) of the joints injected with tivity in animals that had received Mab i.2B6, tissues were buffer alone showed no evidence of anti-E-selectin Mab excised and radioactivity counted postmortem. Whereas i.2B6 staining or of an inflammatory cell infiltrate. In con the LR for radiolabeledcontrol IgOin synovium and drain trast, Mab i.2B6 clearly stained the endotheium of ing iliac lymph nodes were 2.7 ±i.2 and 1.2 ±0.3 (mean venules in the synovium of the PHA-injected joint and ±s.d.) respectively, those for radiolabeled Mab i.2B6 these positive vessels were surrounded by a marked inN were significantlygreaterat i3.3 ±5.6 (p < 0.05) and8.5 ± tration of inflammatorycells (Fig. iB). In addition, there 2.9 (p < 0.O5)(Table1). Therewere no differencesbetween was a marked increase in Mab i.2B6 staining of endothe pigs that received radiolabeled control antibody and those lium in the deep inguinal lymph nodes draining the PHA that received radiolabeled Mab i.2B6 in LR of popliteal injectedjoints. The popliteal and superficial ingUinal lymph and superficial inguinal lymph nodes. nodes showed no expression of E-selectin, on the side of either PHA-injected or control knees. DISCUSSION A modest increase in uptake of radiolabeledantilxxly in A numberof differentapproaches to the imaging of in the inflamed knee relative to the control knee was evident flammation have been developed in recent years. Broadly during the i-hr dynamic study and in the early 15-mmstatic spealdng, these may be considered in three groups: (1) image using either Mab i.2B6 or control IgO (data not those that rely on the nonspecific accumulation at sites of shown). The delayed static images obtained at 24 hr inflammation of radionucides administered either as salts showed that the image intensity in the PHA-injectedjoint such as 67Ga-citrate(18) or coupled to such as

278 The Journal of Nudear Med@ane•Vol. 35 •No. 2 •February 1994 @ . @. ,@

bodies (22); and (3) those that target specific ligands on A endotheial cells and/orlocalizing migratorycells using ra diolabeled antibodies or peptides. While this third ap proach has previously been employed using antibodies against IIb/IIIa on platelets (23), P-selectin on platelets and @ .:c @;_h, endothelial cells (24), and ICAM-i on leukocytes and en dothelial cells (25), endotheial-specific activation antigens have not been used as targetsfor imagingpurposes. In this study we have demonstrated that it is possible to image @ : . . . inflammation by targeting E-selectin induced on vascular @ - ‘-.‘., . . ,,@ endotheium by the inflammatory process. The imaging potential of radiolabeled anti-E-selectin Mab i.2B6 was tested in a model of PHA-induced arthritis, based on previous work demonstrating (1) the accumula tion of radiolabeled lymphocytes in synovium following intra-articular injection of PHA (10); (2) the specific local ization of anti-E-selectin Mab i.2B6 to PHA-injected skin sites (9); and(3) immunohistologicalevidence of E-selectin expression in PHA-injected skin sites (11). Whereas im munohistochemical staining of E-selectin with Mab i.2B6 was undetectable on the synovial vessels of unstimulated joints, there was clear E-selectin expression on the endo theium of postcapillaiy venules, together with a marked perivascular leukocytic infiltration, in the synovia of all joints injectedwith PHA. At present, it is not clearwhether inductionof E-selectin expression on synovial endotheium FIGURE 1. Immunohlstochemlcalstair@ngofs@cMumfromboth test and control knee joints.Tissues were snap frozen and sections in this model is a directendothelialcell response to PHA or were fixed In acetone/methanol before stalnhig for reactivitywith is secondary to the release of cytokines such as interleu anti-E-selectln Mab I 2B6. (A)There was no E-selectln staWik@gIn kin-i or tumor necrosis factor. In skin, there is evidence the syrxMum from the control knee Injected with buffer alone. (B) that the PHA-induced response is at least partially inde There was marked endothellal staIning (—) of venules In the pendent of these two cytokines (10). synovium of the PHMnjected knees and these vassals were stir rounded by Intense infIammato@ycell Infiltrates(x 460). We have shown by intra-articularinjection of India ink that the deep inguinalratherthan the popliteal or superfi cial femoral lymph nodes drain lymph from the knee (data not shown). Consistent with this observation, we found albumin (19) or nonspecific IgG (20); (2) those that detect that vessels in the deep inguinal lymph nodes draining the uptake of neutrophils, labeled either in vitro with ii PHA-injected knees expressed E-selectin, whereas no pophilic chelates (21) or in vivo by anti-granulocyte anti E-selectin induction was detected by immunohistochemis

[3 FIGURE2, Thltt@nute static gamma camera Im@es showfrig anterior visws of the lowerabdomen and hind bmbs acquired 24 hrafterthe Intra—articularInjectionof PHA (400 p9) Intothe ñghtknee and buffer Into the laft knee. The knages show the pattern of uce of (A) 111I,@k@I@@control anti body and (B) 111ln-Iabeledantl-E-SeIecIIn @ I .286 gIvenIntravenously3 hrafterthe Intra-artlcularInjections.There Is Increased uptaiceof both antibodIesIntothe PHA-In jected jointthough this Is considerably more diffuseand less Intense for the control anti It body than for anti-E-eeisctln Mab I 266 (4@). In addition, Mab 1 .266 Is taken up Intothe deep Ingulnallymphnodes draining the @flamedknee ( —) and appeers aiso to be taken up by the skin, visibleas a mar @ntothebody outhne().

ImagingInflammation•Keelanat al. 279 try in vessels of ipsilateral popiteal or superficial inguinal ACKNOWLEDGMENTS lymph nodes. This vascular endotheial activation of yes Theworkdescribedin this studywas fundedby the British sels in the regional lymph nodes was probably due to the Heart Foundation. lymphatic passage either of PHA itself or of cytokines generated by the articularinflammatoryresponse. It was REFERENCES unlikely to be due to vascular dispersal of these factors as 1. PoberJS,CotranRS.Cytokinesandendothelialcellbiology.PhysiolRev the contralateraldeep inguinallymph nodes showed mini 199070427—451. mal expression of E-selectin. 2. Springer TA. Adhesion receptors of the immune system. Nahu@ 1990-,346: Specific uptake of anti-E-selectin compared to that of 425—434. 3. ButcherEC.Leukocyte-endothelialcellrecognition:three(ormore)steps control immunoglobulin in PHA-injectedjoints was clearly to specificityand diversity.Cell 1991;67:1033—1036. observable with the naked eye in gamma camera images 4. PoberiS,BevilacquaMP,MendrickDL,LapierreLA,FiersW,Gimbrone MA.Twodistinctmonokines,interleukin1andtumornecrosisfactor,each taken 24 hr after injection of radiolabeled antibody and was independentlyinducethe biosynthesisand transientexpressionofthe same validatedby countingaccumulatedradioactivityin samples antigen on the surface of cultured human vascular endothelialcells. I of synovium excised postmortem. Based upon in vitro Immwiol 1986;136:1680—1687. 5. BevilacquaMP,StengelinS,GimbroneMA,SeedB.Endothelialleukocyte work, it is probable that the accumulation of radiolabeled adhesionmolecule1: an induciblereceptorfor neutrophilsrelated to corn anti-E-selectin depends not only upon the degree of cx plement regulatory proteins and . Science 1989243:1160-11M. 6. BevilacquaMP,NelsonRM..I ClinInvest1993;91:379-387. pression of E-selectin by vessels in inflamed tissues but 7. CotranRS,GimbroneMA, BevilacquaMP,MendrickDL, PoberJS.In. also on progressive internalization of antibody-antigen ductionand detectionof a humanendothelialactivationantigenin vivo.I complexes following binding of the Mab to endotheium Exj@Med1986;1M:661-666. 8. NorrisP.PostonRN,ThomasDS,ThornhillM, Hawki, HaskardDO.The (26). Insofar as the control immunoglobulin might be cx expressionof endothelialIcUkOCyteadhesionmolecule-i(ELAM-1),Inter pected to reflect the behavior of radiolabeled human im cellular adhesionmolecule-i (ICAM.1)and vascular cell adhesion mole cole-i (VCAM-1) in experimental cutaneons inflammation: a comparison of munoglobulin in inflammatory disease, the marked differ ultraviolet-Berythema and delayed hypersensitivity.I Invert Dennatol ence between localizationof anti-E-selectinMab i.2B6 and 1991;X:763—770. the control immunoglobulinpredicts a superiority of such 9. KeelanETM, LicenceST,PetersAM, BinasRM, HaskardDO. Charac terizationof E-selectinexpressionin vivo usinga radiolabeledmonoclonal an antibody over polyclonal human immunoglobulin for antibody.AmIPI@ysioI1994:inpress. clinical imaging. 10.BinnsRM,LicenceST,WoodingFBP,DuffusVIPH.Activelymphocyte Apart from the more intense image of the inflamed knee trafficinducedin the peripheryby cytokinesandphytohemagglutinin:three differentmechanisms?Eurllmnzwaol 19@222:2195-2@3. obtained with the radiolabeled anti-E-selectin, there was 11.WhyteA, HaskardDO,BinnsRM.InfiltratinggammadeltaT cellsand also a greater clearing of the background radioactivity selectinendothelialligandsin the cutaneousphytohaemagglutinin-induced infiammatoryreactioninthe pi@Vetlmmwsollmnzwiopathol1994:inpress. compared with that observed with control immunoglobu 12. Binns RM, Licence ST, Wooding FWP. Phytohemagglutinininduces major un. This was particularly marked in one animal (animal i), short-termprotease-sensitivelymphocytetTa@inVO1Vingh1ghendothdlium in which the decreasing background in parallel with the venule-likebloodvessels in acute delayed-typehyperaeimtMty-likemac tknisin skinandothertissues.Eurllmmwsol 1990,[email protected]. increasinglocalization of signal renderedpossible the clear 13.WellicorneSM,ThorahillMH, PitzalisC,etal.Amonoclonalantibodythat imaging of the deep inguinal lymph nodes at 24 hr postin detectsa novelantigenonendothelialcellsthat isinducedby tumornecrosis factor, IL-i or lipopolysaccharide.llmmunol 199O@,144:2558-2565 jection. It is possible that in this instance the clearing of 14.Ey PL, ProwseSJ,JenkinCR.IsolationofpureIgGi, IgO2a,andIgO2b Mab i.2B6 was due to specific uptakein the skin, which we immunoglobulinsfrommouse serum usingproteinA-sepharose.Immwao have shown elsewhere to express variable amounts of theniLcby 1978;15:429—436. 15.KohlerG,HoweSC,MilsteinC.Fusionbetweenimmunoglobulin-secreting E-selectin in the absence of experimental stimulation and nonsecretingmyelornacelllines.Eurllmmunol 1976;6:292—295. (9,11). Nevertheless,this abilityof the antibodyto clear 16.}InatowichDJ, ChildsRI, LanteigneD, NajafiA. Thepreparationof IYFPA-cOUpledantibodiesradiolabeledwith metallicradionuclides:an im from the circulation may be a distinct advantage for the provedmethodIlmnzwsolMethcds1983;65:147-157. clinical imaging of a localized lesion in internal organs. 17.RothkotterHi, UlbrichH, PabstR. Thepostnataldevelopmentof gut Because E-selectin functions as an adhesion molecule laminaproprialymphocytes:number,proliferation,andTand Bcell subsets in conventionaland germ-freepigs.PediatrRes 1991;29:237-242. for leukocytes, in some respects imaginginflammationwith 18.McNeilBi, SandersR, MdersonP0, etaLA prospectivestudyof corn radiolabeledanti-E-selectin can be comparedwith imaging puted tomography,ukrasound,and galliumimagingin patientswith fever. with radiolabeled leukocytes. AlthOUghthis latter tech Radiology1981;13%647-653. 19.De-SchrjjverM, StreuleK, SenekowitschR, FridrichR. Scintigraphyof nique has an established role in the clinical investigation of inflammationwith nanometer-sizedcolloldaltracers. Nuci Med Co,nnsun inflammatory disorders, the preparation of radiolabeled 1987;8:895—908. 20. RubleRH,FISChmanAJ,CallahanRi,etaLIndium-ill-labelednonspecific leukocytes is time-consuming and requires experience. For immunoglobulinscanningin the detectionoffocal infeodon NEnRJIMed this reason, labeled white cell scans are not generally avail 1989;321:935—940. able outside specialist centers. Since imagingwith radiola 21. PetersAM, SaverymuttuSN,ReavyHi, DanpureHi, OsmanS,Lavender ip. Imagingofinfi@mmationwithindium-11'tropolonatelabeledIeukocytes beled antibody E-selectin is technically more straightfor INUC1Me4198324:39-44. ward, it will now be importantto evaluate the potential of 22. JosephK, HofikenH, BossletK, SchorlemmerLW. Imagingofinfiamma this technique in the clinical setting in a wide variety of tionwith granulocyteslabeledinvivo NuciMed Co#nmun1988;9:763—769. 23. StuttleAW, KiosokJ, PetersAM, LavenderJP.Sequentialimagingof cardiovascular, rheumatic and neoplastic disorders involv postoperativethrombususingthe 11'Inlabeledplatelet-specificmonoclonal ing cytokine-mediated endotheial activation.

280 The Journal of Nudear Medians •Vol.35 •No:2 •Febwaiy 1994 antibodyP256.BrlRadiol 1989;62:963—969. Imagingof intercellularadhesionmolecule-iinductionin rejectingheart: a 24. Miller t@D,Rivera FJ, GarciaOJ, PalniazJC, &rger Hi, WeismanHF. new ScintigraphiCapproachto detect early allograftrejection. Tiunsplant Imagingofvascular inju,ywith @“9@c-labelndmonoclonalantiplateletanti P?vc1993;25:867—869 body S12:prdllmina@yexperiencein humanpercutaneoustranslurninalan 26. von AssnuthEJU, SmeetsEF, GinselLA, OnderwaterJJM, Leeuwenberg gioplasty.Circulation199285:1354-1363. JFM, Buurman WA. Evidence for endocytosis of E-selectin in human 25. OhtaniH, StraussHW, SouthernJF, TamataniT, MiyasakaM, lathe M. endothelialcells.Eurllmmunol 199222:2519-2526.

EDITORIAL ImagingVascularEndothelialActivation

Keelan and colleagues exploit some of phils and platelets. Thus, adhesion is a ably because of better localization in the recently acquired knowledge in multistep process involving multiple the inflammatory lesion and extrac molecular biology, particularly the alternative receptor-ligand pairs (5), tion by the dermalvasculature. Could fundamentally important mechanisms and all of the players have not been specific antibody Fab, Fab1or F(ab)@ of leukocyte-endotheial adhesion in identified as yet. fragments reduce the time interval for the inflammatory reaction (1). A brief In this preliminary paper, Keelan et optimal imaging? summaiy of this subject (2—5)follows a!. describe an intriguing idea for im Many other questions remainto be for the stout-hearted (others, please aging foci of inflammation with a radi answered. Is traumatized endotheium skip the next paragraph). olabeled monoclonal antibody (Mab) activated by mechanisms similar to Of all leukocyte integrin adhesion specifically for E-selectin expressed those of inflammatory lesions? The molecules, only neutrophils activated on the surface of activated endotheial pathology literature (summarized in in infiammatoly lesions by local spe cells. So far, they have provided companion paper reference 9) indi cific chemotactic cytokines such as in quantitative data only in the form of cates that E-selectin expression cc terleukin-8 (IL-8) possess significant localization ratios of inflamed versus curs in many lesions, including numbers of the surface glycoprotein contralateral normal extremities in chronic dermatoses, “collagenvascu complexes designated CD11b/CD18 pigs. Information on the absolute con lar―diseases, allergies, transplantre or Mac-i. These complexes interact centrationof radioactivity in different jection and even lymphoid malignan with endotheial intercellularadhesion inflammatoiy lesions, including ab des. Will leukocytes attracted and molecule-i (ICAM-i). Another leuko sc@esses,also would be important. bound to activated endotheial E-se cyte integrin, LFA-i, predominantly How would these concentrations lectin receptors compete with the in lymphocytes but also in neutro comparewith those obtainedwith leu binding of specific Mab molecules? phils, has an affinity for endotheial kocytes labeled in vitro after their re Despite these unknowns, this new ICAMS (5). Other adhesion molecules injection? To assess the value of this approach looks exciting. Hence, fur include three selectins: L-selectin or agent, we need to know its distribu ther experimentalwork and subsequent leukocyte endotheial cell adhesion tion in the major visceral organs to clinical trials appear veiy worthwhile. molecule i (lymphocyte homing re judge its efficacy in detecting inflam ceptor, MEL-i4, gp 90, LAM-i or matoiy lesions in the torso. Some in John G. McAfee Leu 8) is expressed on the surface of formationon the cardiac, hepatic and The National Institutes of Health neutrophils; E-selectin or endothelial skin activity, and plasma disappear &thesda@Maryland leukocyte adhesion molecule (ELAM ance is provided in the companionpa 1), expressed on the surface of acti per (author's reference 9). Immuno vated endotheial cells (as described histological studies in this earlier REFERENCES by the authors), interacts with L-se publication showed that the only nor 1. KeelanETM, HarrisonAA, ChapmanPT, lectin, leading to cell margination mal tissue expressing E-selectin was Binns RM, Peters AM, Haskand DO. Imaging which is followed by adhesion as a the vascularity of the dermis. vascularendothelialactivation:an approachus. result of the interaction between the This imaging approach unavoidably lag radiolabeledmonoclonalantibody against the endothelialcell adhesionmoleculeE-selec neutrophil integrin and ICAM-i; and introduces the well-known disadvan tin.INuciMed 199435:276-281. P-selectin, granule membrane protein tages of murine Mabs such as the like 2. NicodLP.Cytokines1:overviewThonu@1993; i40 (GMP-i40) or platelet activation@ lihood of HAMA formation. The 48:660—667,1993. 3. StricterRM, LukacsW, StandifordTJ, et al. dependent granule to external mem plasma clearance of the large IgG@ Cytokines2: cytokinesand lunginflammation: brane (PADGEM, CD62) is released molecules will be slow. Hence, opti mechanismsof neutrophil recruitment to the from platelets and endothelial cells, mal imaging will probably be seen at lung.Thonix1993;48:765—769. 4. MontefortS, HolgateSI, HowarthPH. Leuco promotingfurtheradhesion of neutro 24 hr instead of providing definitive cyte-endothelialadhesion moleculesand their images within a few hours. However, role in bronchial asthma and allergicrhinitis. ReceIvedNe,. 10,1993; @[email protected],19@3. Keelan et al. found that the plasma EurRespirl 1993;6:1044-1054. For correspondence or repdtls cor@ John 5. ButcherEC. LCUkOCytC-endOthelialcellrecog hoNes, MD,Dept.of NuclearMed@ne,NIH,Bldg. clearance was faster with the specific nition:three (or more)steps to specificityand 10,RoomIC4O1,Bethesda,MD20892. Mab thanwith nonspecific IgG, prob dive@. Cell 1991;67:1033-1036.

ImagingVascular Endothal@alActivation•McAfee 281