UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Functional analysis of ICAM-1 : LFA-1 interaction in cell adhesion Welder, Clayton Anthony 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_welder_clayton.pdf [ 21.96MB ]
JSON: 831-1.0086231.json
JSON-LD: 831-1.0086231-ld.json
RDF/XML (Pretty): 831-1.0086231-rdf.xml
RDF/JSON: 831-1.0086231-rdf.json
Turtle: 831-1.0086231-turtle.txt
N-Triples: 831-1.0086231-rdf-ntriples.txt
Original Record: 831-1.0086231-source.json
Full Text

Full Text

FUNCTIONAL ANALYSIS OF ICAM-1:LFA-1 INTERACTION IN CELLADHESIONbyCLAYTON ANTHONY WELDERB.Sc., The University of British Columbia, 1989A THESIS SUBMI I IED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Microbiology)We accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIAMarch 1992© Clayton Anthony Welder, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature) Department of fAicAsolD(I LIThe University of British ColumbiaVancouver, CanadaDate ^l Adka\irct, oq),DE-6 (2/88)AbstractThis thesis presents the results of a project aimed firstly at exploring the use of asoluble form of ICAM-1 to inhibit cellular immune responses, which in general rely heavily onthe interaction between the counter receptors ICAM-1 and LFA-1, and aimed secondly atstudying the mechanisms regulating ICAM-1:LFA-1 mediated cell adhesion. Functionalcharacterization of the purified sICAM-1 demonstrated that radio-iodinated sICAM-1 couldbind to LFA-1 positive cells, albeit with an apparently low affinity. However, whenimmobilized on plastic, sICAM-1 was fully functional, efficiently facilitating LFA-1:ICAM-1mediated cell adhesion. Two mechanisms of inducing LFA-1:ICAM-1 mediated cell adhesionwere characterized: stimulation by the phorbol ester PMA and stimulation by the divalent cationMn++. PMA induced adhesion was dependant on a functional actin cytoskeleton as judged bythe inhibitory effect of cytochalasin B. PMA induced adhesion could not be inhibited bysICAM-1, suggesting that this cell adhesion was mediated by low affinity ICAM-1:LFA-1interaction. In contrast, Mn++ induced adhesion seemed to be mediated by an increase in theaffinity of the ICAM-1:LFA-1 interaction; Mn++ induced homotypic aggregation could bespecifically inhibited by sICAM-1. While monovalent sICAM-1 could not efficiently inhibitPMA induced adhesion, a multivalent form of sICAM-1 could inhibit PMA induced ICAM-1:LFA-1 mediated cell adhesion, demonstrating the importance of multivalent interaction inLFA-1:ICAM-1 mediated cell adhesion. Fluorescence microscopic studies aimed atdetermining the cell surface distribution of ICAM-1 and LFA-1 clearly showed that thedistribution of ICAM-1 could be differentially regulated. With LFA-1, the regulation of it'scell surface distribution could not be demonstrated using immunofluorescence microscopy.Studies were conducted to define a role for the cytoplasmic domain of ICAM-1 by transfectingICAM-1 with or without its cytoplasmic domain into ICAM-1 negative cells. These studiesdemonstrated that removal of the cytoplasmic domain of ICAM-1 resulted in a decreased butiinot abolished aggregative phenotype. However, a precise role for the cytoplasmic domain ofICAM-1 in determining the cell surface distribution of ICAM-1 could not be defined.These results are discussed in relation to the literature concerning soluble adhesionmolecules and the regulatory mechanisms governing cell adhesion. It seems likely thatsICAM-1, at least in it's monovalent form, has little potential as an inhibitor of in vivo immuneresponses. It is hypothesized that there are three primary mechanisms regulating cell adhesionmediated by ICAM-1 and LFA-1. The first mechanism is simply the amount of LFA-1 orICAM-1 a given cell expresses on it's surface, the second mechanism is the modulation of theaffinity of the interaction between ICAM-1 and LFA-1, and the third mechanism is theregulation of the distribution of adhesion molecules on the cell surface. Each of theseregulatory mechanisms is discussed with reference to the results presented in this thesis as wellas the results in the literature.iiiTable of ContentsAbstract^ ii.Table of Contents^ iv.List of Figures v.List of Abbreviations^ vii.Acknowledgement ix.Introduction^ 1.Materials and Methods^ 18.Results^ 34.Discussion 111.Bibliography^ 129.ivList of FiguresFigure Title Page1. A schematic representation of the molecules involved in T cell recognitionand adhesion4.2. Initial detection of sICAM-1 in culture supernatants and following affinitychromatography purification36.3. Confirmation of molecular weight, purity, and yield of purified sICAM-1 37.4. Assessment of the purity of radio-iodinated sICAM-1 38.5. 125I-sICAM-1 binds to LFA-1 positive cells 40.6. Quantitation of the density of sICAM-1 coated to microtitre wells 44.7. SICAM-1 is functional when absorbed to plastic in the quantitative celladhesion assay45.8. SICAM-1 does not inhibit the PMA induced homotypic aggregation of 47.MBL-2 cells9. SICAM-1 inhibits the Mn++ induced homotypic aggregation of MBL-2cells48.10. Cell adhesion in the quantitative cell adhesion assay is stimulated by PMAand inhibited by anti-LFA-1 and cytochalasin B but not by sICAM-150.11. Cell adhesion of resting T cells in the quantitative cell adhesion assay isabsolutely dependent on PMA stimulation52.12. Mn++ induces ICAM-1:LFA-1 mediated adhesion in the quantitative celladhesion assay53.13. Cytochalasin B induces morphological changes in cells 55.14. Quantitation of sICAM-1 coated to polystyrene beads 59.vi15. SICAM-1 coated polystyrene beads bind to MBL-2 cells^60.16. Multivalent sICAM-1 coated beads inhibit cell adhesion while monovalent 61.sICAM-1 does not17. The distribution of cell surface molecules on unstimulated or PMA^64.stimulated MBL-2 cells18. The distribution of cell surface molecules and intracellular actin on/in^74.cytochalasin B treated or untreated A20 cells19. T28 cells transfected with ICAM-1 with or without its cytoplasmic tail^82.behave differently in culture20. T28 cells transfected with ICAM-1 with or without it's cytoplasmic tail^83.aggregate in response to PMA21. The distribution of cell surface molecules on unstimulated or PMA^85.stimulated T28 cells expressing transfected forms of ICAM-122. The distribution of cell surface molecules on unstimulated or PMA^94.stimulated, untransfected T28 cells23. The distribution of cell surface molecules on L cells expressing transfected 99.forms of ICAM-124. The distribution of cell surface molecules on untransfected L cells^107.25.^The effect of adhesion molecule distribution on cell:cell adhesion^122.List of AbbreviationsAbbreviation^ MeaningAb^ antibodyBSA bovine serum albuminCD^ cluster of differentiationcDNA complementary DNACHO^ chinese hamster ovaryCPM counts per minutecyto. B^ cytochalasin BDAB 3,3'-diaminobenzidine tetrahydrochlorideDAB CO^ 1 ,4-diazobicyclo- (2,2,2)-octaneDMEM Dulbeco's modified minimal essential mediumEDTA^ ethylene diamine tetraacetic acidELISA enzyme linked immunosorbent assayFACS^ fluorescence activated cell sorterFCS fetal calf serumFITC^ fluorescein isothiocyanateGVHD graft versus host diseasefMLP^ formyl-methyl-leucyl-phenalanineHanks' Hanks' balanced salt solutionHepes^ N-2-hydroxyethylpiperazinehr. hourICAM^ intercellular adhesion moleculeIFN interferonIg^ immunoglobulinviiIL^ interleukininh. inhibitorkd^ kilodaltonsLAD leukocyte adhesion deficiencyLFA^ lymphocyte function associatedMHC major histocompatibility complexMALA^ murine activation lymphocyte antigenmin. minutemRNA^ messenger RNAMTT 3-(4,5-dimethylthiazol-2-y1)-2,5-diphenyl tetrazolium bromideOVA^ ovalbuminPAGE polyacrylamide gel elecrophoresisPBS^ phosphate buffered salinePCR polymerase chain reactionPFHM^ protein free hybridoma mediumPKC protein kinase CPMA^ phorbol myristate acetateRPM revolutions per minuteSDS^ sodium dodecyl sulphateSN supernatantsICA M- 1^ soluble ICAM-1TcR T cell receptorTNF^ tumour necrosis factorTri s tris(hydroxymethyl)aminomethaneT.R.^ transferrin receptorTR ITC tetramethylrhodaminly isothiocyanateAcknowledgementI wish to thank Fumio Takei for invaluable supervision and exhortation throughout thisthesis project. I owe a dept of gratitude to my committee members Wilf Jefferies, PaulineJohnson, and John Schrader for providing instant understanding, and for validating the work.I am thankful for all of the immunologists at UBC; Geoff Hoffmann, Julia Levy, WilfJefferies, Pauline Johnson, and Hung-Sia Teh; whom I have had the pleasure of learning from.I would also like to thank all the members of the Takei lab, especially Daniel Lee, for helpfuldiscussion and encouragement. Finally, thanks to my wife Christine for continual support andinspiration.Happy is the man who finds wisdom,and the man who gets understanding,for the gain from it is better than gain from silverand its profit better than gold. Proverbs 2: 13,14ixIntroductionAdhesion moleculesAdhesive interactions between cells or between cells and the extracellular matrix aremediated, by definition, by adhesion molecules. Adhesion, and therefore adhesion molecules,are fundamentally important to all multicellular animals. Adhesive interactions mediated byadhesion molecules are responsible to a large extent for the organization and maintenance of thevarious tissues and organs of the body, as well as being required for a wide range of biologicalfunctions. Since a cell is essentially a sphere with a net negative surface charges, electrostaticforces tend to repel cells from one another, and adhesive interactions are required to stabilizecell:cell contact. A stereochemical fit between complimentary molecules, which is stabilized byhydrophobic interactions, ionic interactions, hydrogen bonding and other van der Waalsinteractions, provides the molecular basis for adhesive interaction (1).As an example of the general importance of adhesion molecules, consider thecadherins, a group of adhesion molecules responsible for maintaining multicellular structures,especially in the embryonic development of vertebrates (2). Cadherins often function byhomotypic interaction, and have a strict requirement for calcium ions. It has been observed thatanti-cadherin antibodies disrupt normal embryonic development (3). As I will discuss in moredetail later, antibodies directed against adhesion molecules utilized by the immune systemeffectively inhibit the function of the immune system.Adhesion molecules in the immune systemThe immune system of higher animals is a particularly interesting system in that it iscomprised of a network of individual cells circulating throughout the body, yet the systemfunctions in a highly organized and regulated fashion. The organization and regulation of theimmune system is, to a great extent, due to the use of a variety of adhesion molecules by the1system. The general Job description' of the immune system is to discriminate self from non-self, and eliminate the latter. Such discrimination necessitates cell:cell recognition, and suchcell:cell recognition is facilitated and stabilized by adhesion molecules. Adhesion molecules inthe immune system facilitate a variety of functions ranging from lymphocyte recirculation andhoming to lymphocyte:target or lymphocyte:accessory cell interactions (reviewed in 4).Adhesion molecules of the immune system are involved in cell:cell as well as cell:matrixinteractions. Most of the adhesion molecules utilized by the immune system belong to one oftwo major groups, those groups being the integrins and the Ig superfamily.Members of the Ig superfamily all have in common the Ig homology unit, a linearstretch of about 110 amino acids which generally folds into a secondary structure consisting of2 anti-parallel f3 sheets stabilized by one disulfide bond. This domain structure was of coursefirst characterized in immunoglobulin molecules. Members of the Ig superfamily usually haveIg-like homology units tandemly repeated. Of particular interest to this thesis is ICAM-1, amember of the Ig superfamily which has five extracellular Ig-like homology units (5). The Igsuperfamily has been well reviewed (6).Members of the integrin family all have in common a two chain structure. Each integrinmolecule is a non-covalently associated al3 heterodimer. Certain p chains are found associatedwith various a chains, hence the 131, 132, and 133 subfamilies. The structure and function of theintegrins has been well reviewed (7). The family of integrins is involved in a variety offunctions ranging from immune cell adhesion to wound healing. Of particular interest to thisthesis is the 132 integrin LFA-1.Leukocyte adhesion deficiency disease (LAD) is a human hereditary disease in whichpatients suffer from life-threatening microbial infections (reviewed in 8) . The disease ischaracterized by defective neutrophil mobility and phagocytosis, and the absence oflymphocytes and granulocytes in infected lesions despite chronic leukocytosis. The cause ofthis disease has been traced to a deficiency in 132 integrin expression (LFA-1, Mac-1, andp150/95) resulting from decreased or complete lack of expression of functional 13 chain (9).2The physiological importance of the 132 integrins is underscored by the serious consequences ofthe lack of their expression in LAD.The r32 integrin lymphocyte function-associated antigen, LFA-1, was originallydiscovered by mAb which inhibit T cell mediated killing (10). Since that time a number of Abto LFA-1 have been characterized, and all of those which inhibit T cell adhesion inhibit nearlyevery immune response involving T cells. LFA-1 has truly lived up to it's name, lymphocytefunction-associated antigen-1. In addition, the importance of adhesive contributions to T cellfunctions is emphasized by the importance of LFA-1:ICAM-1 interaction in T cell function.The interaction between LFA-1 and ICAM-1 is one of a number of interactions involved in therecognition and adhesion of T cells, as depicted in Figure 1.The counter receptor for LFA-1, ICAM-1, was originally identified using the phorbolester induced homotypic aggregation assay (11). In this assay, lymphocytes or lymphocyticcell lines aggregate only after stimulation by phorbol esters such as PMA. Aggregation in thisassay is completely inhibited by Ab to LFA-1. It was observed that LFA-1 negative cells couldco-aggregate with LFA-1 positive cells, suggesting that there was a counter receptor for LFA-1on the LFA-1 negative cells. ICAM-1 was identified by immunization of mice with these LFA-1 negative lymphocytes and screening for Abs capable of inhibiting phorbol ester inducedaggregation (12). Molecular cloning of ICAM-1 showed that it was a member of the Igsuperfamily, with five extracellular Ig-like homology units (5). The presence of ICAM-1confers the potential for ICAM-1:LFA-1 mediated adhesion (13). In addition to ICAM-1,ICAM-2 has been characterized as another counter receptor for LFA-1. ICAM-2 was originallydefined by the ability of antibodies to LFA-1, but not by antibodies to ICAM-1, to inhibit thebinding of T cells to endothelial cells (14). Expression cloning resulted in the isolation of acDNA encoding ICAM-2, a molecule which has only 2 Ig-like homology units with 35%identity to the two N-terminal Ig-like homology units of ICAM-1 (15). Indeed there isevidence for a third counter receptor for LFA-1, based on the ability of certain cell lines toaggregate in an LFA-1 dependant but ICAM-1 and 2 independent manner (16).3antigenpresentingcellFigure 1^A schematic representation of the molecules involved in T cell recognitionand adhesion. 4LFA-1 is expressed constitutively, but it's expression is restricted to leukocytes (17).In contrast, ICAM-1 is highly inducible by inflammatory mediators. In the absence ofinflammation, ICAM-1 is expressed on only a very few cells; however, ICAM-1 expression isinduced by a number of inflammatory mediators including IL-1, TNF, and 7IFN (18). ICAM-1 expression is inducible on a variety of cell types including hemopoietic, fibroblastic andendothelial cells. ICAM-2 expression differs from ICAM-1 expression markedly in that it isexpressed well basally on endothelial cells and is not upregulated by inflammatory mediators(19).It is clear that there are multiple ICAM counter receptors for LFA-1. One recent reportalso suggests that there may indeed be multiple counter receptors for ICAM-1 (20). This reportdemonstrates that purified, immobilized ICAM-1 can serve as an adhesion receptor for CD43, amolecule found expressed at high levels on T lymphocytes, monocytes, neutrophils, platelets,and activated B lymphocytes. The in vivo, functional significance of ICAM-1:CD43interaction has yet to be determined.In addition to the two families of adhesion molecules already discussed, another groupis emerging with important contributions to leukocyte adhesion. This group is known as theselectins (reviewed in 4), named for the common N-terminal lectin-like domain shared by all ofit's members. In addition to the Ca++-dependant lectin-like domain, members of the selectinsalso have in common an EGF motif next to the lectin-like domain and short consensus repeatswith homology to complement regulatory proteins. Included in the group of selectins areCD62, ELAM-1 and LECAM-1, also known as Mel-14/LAM-1. Selectins are thought to beinvolved in the binding of leukocytes to endothelium at the sites of inflammation as well as inlymphocyte recirculation. One selectin known as CD62, PADGEM, or GMP-140 is stored inthe Weible-Palade bodies of endothelial cells, and is rapidly mobilized to the cell surfacefollowing stimulation by products of the clotting cascade (21, 22). CD62 has been shown tomediate the rolling of neutrophils on endothelium (23), a process the authors argue is facilitatedby the length of the selectin molecule and it's potential for rapid association and dissociation5kinetics. Another report have shown that the ligand for CD62 may be the selectin LECAM-1on neutrophils (24). At least for neutrophils, the rolling adhesion mediated by selectinsappears to be a prerequisite for static adhesion mediated by integrins.Lymphocytes are constantly patrolling the entire body in search of foreign antigens.Adhesion molecules allow lymphocytes to travel throughout the body and scrutinize the entirebody for foreign antigens. The general scheme of this patrol route begins with thelymphocytes traveling through the blood circulatory system, then at various points in thevasculature they emigrate through the endothelium either into LN or directly into the tissues.The processes and molecules involved in lymphocyte recirculation have been well reviewed(25, 26, 27). Lymphocytes emigrating directly into the tissue usually do so at sites ofinflammation, from which point they travel to a nearby LN via the afferent lymphatic vessels.From the LN, lymphocytes return to the blood via the efferent lymphatics and the thoracic duct.Lymphocytes enter the LN from the blood stream by binding to specialized post-capillaryvenules cells called 'high' endothelial venules (HEV). 'Recirculation' receptors onlymphocytes and molecules termed 'addressins' on endothelial cells have been identified by Abcapable of blocking the binding of lymphocytes to endothelium (27), and likely representcounter receptors for lymphocyte recirculation. Candidate molecules implicated in this processinclude CD44, LFA-1, VLA-4, and LECAM-1 on lymphocytes and ELAM-1, CD62,hyaluronic acid, and poorly defined carbohydrate moieties on endothelium (reviewed in 4). Adegree of specificity has been observed in lymphocyte recirculation; a given lymphocyte willcontinually recirculate either through peripheral LN or through the Peyer's patch. Therefore,site specific receptors on lymphocytes termed homing receptors are likely involved. Suchhoming receptors on lymphocytes include the selectin LECAM-1 which seems to facilitate thehoming of lymphocytes to peripheral LN (28), and the VLA-4 a-subunit associated with eitherof two integrin 13-subunits, which seems to facilitate homing to the Peyer's patch (29).Memory T cells have been observed to emigrate preferentially from the blood directly into6tissue (30), and it is interesting to note that memory T cells show an increased expression ofthe candidate homing receptors CD44 and VLA-4 (4).As stated above, the role of the immune system is to distinguish self from non-self andeliminate the latter. The cells that accomplish this task are the B and T lymphocytes. Blymphocytes recognize soluble or cell-bound antigens via membrane bound Ig, and react bysecreting Ig directed against the recognized antigen. B lymphocytes are regulated such thatthey do not respond to self. T lymphocytes recognize only cell-associated antigens; therefore,they require cell:cell contact in order to function. A schematic representation of the moleculesinvolved in the cell:cell recognition and adhesion of T cells is given in Figure 1. The antigenspecificity of T cell recognition is mediated by the TcR complex (31, 32), which recognizesantigenic peptides presented on the opposing cell surface in the context of Class I or Class IIMFIC. CD4 or CD8 on the T cell also bind to Class II or Class I respectively (33). In additionto the TcR/CD4/CD8 recognition of MHC/peptide, two sets of adhesion molecule counterreceptors have been demonstrated to play a crucial role in T cell recognition. These interactionsare between LFA-1 and ICAM-1, as discussed above, and between CD2 and LFA-3.The regulation of adhesion in the immune systemOne of the interesting aspects of many adhesion systems, including some systemsutilized by the immune system, is the ability of these systems to be regulated. Adhesionmediated by many adhesion molecules is not simply a random process facilitated by the bindingof counter receptors, but rather adhesion mediated by many molecules can be specificallymodulated by the cell. As shown in Figure 1, the major adhesion molecule interactions in Tcell recognition are between LFA-1 and ICAM-1 and between CD2 and LFA-3. Both of thesetwo adhesion systems are subject to regulation by the cell.Both CD2 and LFA-3 are members of the Ig superfamily. Adhesion mediated byCD2:LFA-3 interaction is thought to be strictly dependant on T cell activation; activated T cellsbut not resting T cells form rosettes with autologous erythrocytes expressing LFA-3 (34). The7difference between activated and resting T cells with respect to adhesion mediated by CD2 andLFA-3 is thought to be the surface charge of the T cells (35). Upon T cell activation, T cellsbegin de novo glycoprotein biosynthesis with altered degrees of glycoprotein sialation. Sialicacid is responsible to a large extent for the net negative charge of the cell surface, and activatedT cells have less sialic acid and therefore less negative surface charge (36). It is believed thatthis reduction in negative surface charge reduces the electrostatic repulsion between cells andallows adhesion mediated by CD2 and LFA-3. The reported dissociation constant for theCD2:LFA-3 interaction is only 10 -6 M (37, 38); therefore, it is easy to see how adhesionmediated by CD2 and LFA-3 could be regulated so delicately as by surface charge. Inaddition, it is believed that ligation of CD2 on T cells results in the transduction of signalswhich are capable of synergizing with TcR derived signals (39). Therefore, with theCD2:LFA-3 interaction we see a regulated adhesion system which is also capable offunctioning in signal transduction.Adhesion mediated by LFA-1 and ICAM-1 is particularly interesting for severalreasons. One of these reasons is that adhesion mediated by LFA-1 and ICAM-1 is not simply apassive process in which complementary counter receptors recognize each otherstereochemically. Rather, adhesion mediated by LFA-1 and ICAM-1 is both finely regulatedand metabolically dependant. Treatment of cells with sodium azide, which blocks thegeneration of ATP in cells, abolishes LFA-1:ICAM-1 mediated adhesion (13). Similarly,when cells are cooled to 4°C LFA-1:ICAM-1 mediated adhesion is completely abolished (13).This is not the case for all adhesion molecule systems, for example CD2:LFA-3 mediatedadhesion occurs readily at 4°C (40), as does adhesion mediated by NCAM, another Igsuperfamily member (41). Incidentally, adhesion mediated by certain members of thecadherins is also metabolically dependant (2)As with the CD2:LFA-3 system discussed above, adhesion of T cells mediated byICAM-1 and LFA-1 is also believed to be strictly dependant on T cell activation. However, anincrease in ICAM-1:LFA-1 mediated adhesiveness occurs in a matter of minutes (42) as8compared to the 12-24 hours required for do novo glycoprotein biosynthesis. Therefore, themechanism(s) regulating LFA-1 mediated adhesion is completely different than that regulatingCD2 mediated adhesion. The increase in adhesion mediated by LFA-1 and ICAM-1 has beenclearly shown to be the result of a change in LFA-1 as opposed to a change in ICAM-1. Thisdistinction was made through the use of adhesion assays in which the binding of cells to eitherplastic absorbed LFA-1 or ICAM-1 was quantitated (42). The adhesion of cells to immobilizedICAM-1, but not to immobilized LFA-1 was increased substantially by stimulation with thephorbol ester PMA, clearly showing that the change induced by phorbol ester is on the side ofLFA-1 and not ICAM-1. The increase in LFA-1 mediated adhesion has carefully been termedan increase in LFA-1 avidity, since for the multivalent binding between cells the distinctionbetween affinity and avidity, a subtle but important distinction, cannot be made. Without amonovalent form of one of the two counter receptors, a clear measurement of affinity cannot bemade; therefore, one cannot distinguish between the affinity of monovalent interaction andavidity, which is the sum of the affinity of monovalent interaction and the number ofinteractions occuring.Phorbol esters, namely PMA, are diacylglycerol analogs, and therefore stimulate PKCdirectly. PKC is an enzyme that facilitates the phosphorylation of proteins in the cell on serineor threonine residues. It has been shown that the 13 chain of LFA-1 is a substrate for PKC, andin fact is phosphorylated on serine following PMA stimulation (43). It has also been shownthat while the chain can be phosphorylated upon PKC upregulation, the a chain of LFA-1 isphosphorylated constitutively (44). Phosphorylation on tyrosine residues, of which the LFA-113 chain has one, has not to our knowledge been demonstrated for LFA-1.In addition to PMA stimulation, Dustin and Springer have also studied the effect ofstimulating T cells via the TcR on LFA-1:ICAM-1 mediated adhesion (42). The results showthat with respect to LFA- l avidity, stimulation of T cells via the TcR does not mirror the effectsof PMA stimulation. Rather, they observed that the kinetics were markedly different; TcRstimulation resulted in a rapid but transient avidity increase while PMA stimulation resulted in a9slower but sustained avidity increase. The kinetics of the increase following stimulation byPMA is in good agreement with the kinetics of LFA-1 r3 chain phosphorylation by PMAstimulation (44). These results suggest that the normal control of LFA-1 mediated adhesion isrelatively complex; with respect to LFA-1 avidity, TcR derived signals do not simply turn onPKC, but PKC upregulation may be part of TcR derived signals.Adhesion of cells to absorbed ICAM-1 is dependant on activation from the so calledlow to high avidity forms of LFA-1, as well as a functional actin cytoskeleton and divalentcations (5, 11, 12, 13). It has been assumed that the avidity of LFA-1 is determined by it'saffinity for ICAM-1, and that PKC mediated phosphorylation of the LFA-1 13 chain, stimulatedby phorbol ester or TcR ligation, regulates the affinity of LFA-1 (42). Again, in the absence ofan assay for the binding of sICAM-1 in solution, the relationship between the affinity of LFA-1for ICAM-1 and it's avidity could not be determined.The major source of confusion in the literature concerning LFA-1 regulation is thedistinction between avidity and affinity. Most of the literature has favored confomationalchange models for LFA-1 avidity increases, which usually imply increases in the affinity of theinteraction between LFA-1 and ICAM-1. With respect to LFA-1, two groups havecharacterized 'activated' forms of LFA-1 through the use of mAb. Figdor et al. have raised amAb, NKI-L16, which binds to an epitope of LFA-1 dependant on divalent cations and cellularactivation (45, 46, 47). This Ab also facilitates LFA-1:ICAM-1 mediated homotypicaggregation of cell lines, presumably by stabilizing an activated LFA-1 conformation. Basedon their data, this group has postulated the existence of three distinct conformational forms ofLFA-1 on the cell corresponding to distinct stages of activation (48). Dransfield and Hogghave similarly characterized an LFA-1 epitope dependant on active metabolism and divalentcations through the use of their mAb, Ab24 (49). With respect to LFA-1 regulation, thedistinction between affinity and avidity is covered at length in the Discussion of this thesis.While increasing affinity is one mechanism whereby avidity is increased, it is by nomeans the only mechanism capable of increasing avidity. Logically, there is at least one10mechanism apart from affinity by avidity can be regulated. By increasing the number ofadhesion molecules interacting, the net strength of interaction between cells can be altered. Therapid inducibility of ICAM-1 is an example of how a cell may increase it's likelihood ofadhering. It has been shown that monocytes mobilize cytoplasmic ICAM-1 to their surfacesrelatively quickly following adherence to fibronectin (50). The distribution of adhesionmolecules on the surface of cells would also determine the overall strength of interactionbetween cells. High local concentrations of adhesion molecules at the site of cell:cell contactwould increase the avidity of interaction. Indeed, it has been shown that Mac-1 on neutrophilsaggregates in the absence of ligand upon PMA stimulation (51). This aggregation was shownto correlate with increased ligand binding ability in the absence of increased Mac-1 expression.In the case of T cells, there have been several reports from a single laboratory on thelocalization of adhesion molecules to sites of contact between T cell and target or accessory cell(52, 53). Reports from this laboratory have also suggested an association between LFA-1 andtalin, a protein associated with the actin cytoskeleton, mediated by serine/threoninephosphorylation events (54, 55). However, this work has failed to address the sequence ofevents that is occuring: do adhesion molecules aggregate thus facilitating adhesion, or doesadhesion result in the mutual capping of counter receptors? There is a report in literature of thelocalization of ICAM-1 on uropods of cloned T cells (50). Interestingly, this cell line formslarge homotypic aggregates, and LFA-1 does not seem to co-localize on the cells' uropods.Staunton et al. have observed that ICAM-1 expressed on transfected COS cells exhibits apunctate staining pattern, even when the entire cytoplasmic domain is deleted (56). The role ofadhesion molecule distribution in the regulation of adhesion is, at present, not wellcharacterized.The association of adhesion molecules with the cytoskeletonFunctionally, many adhesion molecules appear to be associated with the actincytoskeleton. These observations were first made with adherent cells in culture. The regions11of these cells at which contact is maintained between the cell and the substrate have been termedthe focal adhesions, or adhesion plaques (reviewed in 57). The adhesion plaque is the point atwhich an adherent cell, such as a fibroblast, is adhering to extracellular matrix proteinsabsorbed to the culture vessel. Extracellular matrix proteins which function in such a mannerinclude fibronectin and vitronectin, both of which have the RGD sequence known to be theligand for their cellular receptors. On the cell's plasma membrane at the focal adhesion werefound high concentrations of integrin molecules such as certain 131 and certain 133 integrins,which were later characterized as being receptors for fibronectin and vitronectin. Inside the cellat the focal adhesion were found cytoplasmic proteins such as actin, aactinin, talin, andvinculin. Actin forms long polymers which make up the insoluble actin cytoskeleton visualizedas stress fibres in adherent cells. Vinculin, talin, and aactinin are proteins which have beenshown to associate with the actin cytoskeleton. Talin is associated with the focal adhesions(58), aactinin has been shown to associate directly with actin (59), and vinculin has beenshown to associate with talin (60), with aactinin (61) and with other yet uncharacterized actinbinding proteins (62). Talin has been shown to associate directly with the fibronectin receptor(63). In addition to the integrins, the actin cytoskeleton has also been shown to be associatedwith other types of adhesion molecules including the cadherins (2). It seems logical that theinteraction between the cytoskeleton and adhesion molecules could control the distribution ofadhesion molecules on the surface of the cell, as discussed above.Studies by Marlin and Springer demonstrated that LFA-1 mediated adhesion oflymphocytes was dependant on a functional actin cytoskeleton (13). Studies by Singer et al.have implicated talin in integrin mediated lymphocyte adhesion (54, 55), and also suggestedthat PKC mediated phosphorylation events mediate the association of talin with LFA-1. It isknown that talin and vinculin are substrates for PKC (64, 65), and as previously mentioned sois the 13 chain of LFA-1. Interestingly, a cytoplasmic protein, 86,000 in molecular weight, hasbeen shown to associate with LFA-1 (66). This is a molecular weight similar to that of severalpoorly characterized proteins found associated with the fibronectin receptor (67, 68). These12data collectively show that the actin cytoskeleton is involved in LFA-1:ICAM-1 mediatedadhesion. More specifically, it seems that PKC mediated phosphorylation events regulateinteractions between the actin cytoskeleton and LFA-1, thus regulating LFA-1 avidity.Physical aspects of adhesionAnother aspect of cell adhesion which should be considered is that of the distancesinvolved (reviewed in 4). The term glycocalyx has been used to represent the complex surfaceof the cell, made up mostly of protruding proteins and carbohydrates (4). Electron microscopyand X-ray crystallography have provided data on the shape and size of several leukocytesurface molecules. As mentioned previously, the glycocalyx has a net negative charge duemostly to sialic acid on glycoproteins. For cells to adhere, their glycocalyces must interact, andadhesive interactions must overcome the electrostatic repulsion. Molecules such as the TcR,MI-IC, CD2, LFA-3, and ICAM-2 extend less than 10 nm from the plasma membrane;therefore, interactions between these molecules must either be of a high affinity themselves, orthe cell:cell association involving these molecules must be stabilized by other interactions.LFA-1 and ICAM- I extend farther away from the plasma membrane, about 15-20 nm, andtherefore could interact with a lesser degree of interdigitation of the two opposing glycocalyces.CD45, a glycoprotein expressed at very high levels on leukocytes, extends about 28 nm fromthe plasma membrane; therefore, closely associated cells may be required to exclude CD45 atthe points of contact. Since the cytoplasmic domain of CD45 has tyrosine phosphataseactivity, signalling may be affected. The selectin CD62 which is involved in neutrophil rolling,extends about 40 nm from the plasma membrane of endothelial cells, and one of it's possibleligands, the selectin LECAM-1 (23), extends about 15 nm on neutrophils. CD43, aglycoprotein expressed at very high levels on T cells, neutrophils, and activated B cells,extends about 45 nm from the plasma membrane, and may serve as an alternate counterreceptor for ICAM-1 (20). The intermediate distance of the ICAM-1:LFA-1 association, an13interaction between molecules extending 15-20 nm from the plasma membrane, is probably ofsignificance to the general importance of this interaction.Soluble versions of adhesion moleculesA soluble version of an adhesion molecule is a form of the molecule which is no longeranchored to the cell, but is a soluble protein free of the cell. Soluble adhesion molecules couldbe produced by proteolytic cleavage of a full length, cell associated adhesion molecule;proteolytic cleavage could be done by the cell or in the test tube. Soluble adhesion moleculescould also be produced by the cell by alternative splicing of mRNA. Alternatively, solubleadhesion molecules could be produced using recombinant DNA techniques to geneticallyengineering the coding sequence for the protein such that it is no longer anchored to the cell,but rather completely secreted from the cell. Soluble versions of adhesion molecules may ormay not maintain their biological activity as measured by binding to their normal counterreceptor.Soluble versions of adhesion molecules are of interest for several reasons. In general,soluble adhesion molecules are crucial for studying the fine details of cell adhesion mediated byparticular adhesion molecules. For example, measurement of the affinity of the interactionbetween adhesion molecules can only be accomplished with a soluble version of one of the twomolecules. It is therefore possible, with a soluble adhesion molecule, to study mechanismsmodulating the affinity of a particular adhesion molecule system. Similarly, it is also possibleto dissect the mechanisms modulating avidity apart from affinity. As with the CD2:LFA-3interaction, adhesion molecules may transduce signals in addition to serving as adhesionreceptors. Soluble adhesion molecules may allow the dissection of these various accessoryfunctions of adhesion molecules. Soluble adhesion molecules also are hopeful candidates forinhibiting cell adhesion, and therefore inhibiting the biological functions mediated by celladhesion, such as inflammation and cellular immune responses. With this in mind, it is notsurprising that the field of adhesion molecules is a very active field. Indeed, there are several14reports in the literature on the production and functional characterization of soluble versions ofadhesion molecules.Wheelock et al. purified a soluble version of cell-CAM 120/80, a member of thecadherins which is normally involved in the adhesion of epithelial cells (69). This adhesionmolecule is found on cells as a 120 kd membrane bound glycoprotein and as a 80 kdproteolysis product in the culture SN of tumor cell lines used for the production of the solubleprotein. The purified 80 kd protein was capable of disrupting cell:cell adhesion in culturedepithelial cell lines at relatively low concentrations (0.01 nM, 0.8 ng/ml). Damle and Aruffoproduced a recombinant, soluble VCAM-1 fusion protein in COS cells, and found that itproduced co-mitogenic signals in CD4 T cells when immobilized on plastic with either anti-TcRor anti-CD3 Ab (70). Lobb et al. expressed recombinant, soluble ELAM-1 in Chinese hamsterovary cells (71). The authors observed that their protein was functional as an adhesionmolecule when immobilized on plastic; however, it only weakly inhibited ELAM-1 mediatedadhesion. Van Seventer et al. recently reported on the production of recombinant, solubleforms of both VCAM-1 and ELAM-1, and demonstrated that each is functional when absorbedto plastic (72). Soluble Mac-1 has been produced by Dana et al. in COS cells transfected withtruncated cDNAs for each of the a and 13 chains in separate expression vectors (73).Surprisingly, the protein was maintained as a noncovalently associated sal3 heterodimer,appeared functional as measured by binding to it's ligand iC3b, and effectively inhibited thebinding of polymorphonuclear cells to IL-1 stimulated endothelium.With respect to ICAM-1, there have been several reports on the production of it'ssoluble version. Marlin et al. reported originally on the production of soluble human ICAM-1through recombinant techniques, and it's use as an inhibitor of rhinovirus infection since therhinovirus receptor is ICAM-1 (74). In addition, during the course of this thesis project areport was published on the production of a recombinant, soluble form of murine ICAM-1(75). This report concentrated on the accessory function of ICAM-1 as demonstrated by it'seffects on antigen presentation by ICAM-1 transfected fibroblasts, and by the activity of15soluble ICAM-1 co-immobilized on plastic with anti-CD3 Ab. However, the authors claim thattheir soluble protein mediates adhesion of T cells when coated to plastic, and they also claimthat their soluble protein blocks ICAM-1:LFA-1 mediated adhesion, although they do notshow any data in support of these claims. It is important to note that the soluble murine ICAM-1 reported in this publication is produced in Chinese hamster ovary cells, and it's molecularweight of only 50 kd suggests inappropriate glycosylation. In addition to the production ofsICAM-1 through recombinant techniques, several reports on the presence of a circulating,soluble ICAM-1 in serum have been published (76, 77). The concentration of this circulating,soluble ICAM-1 was reported to be in the range of 100 ng/ml (77). One group has shown thatupon stimulation with IFN-y or TNF-a, melanoma cells shed ICAM-1, and that this shedICAM-1 is capable of inhibiting ICAM-1 mediated adhesion (78).The report on the shedding of ICAM-1 is one of a number of reports on the shedding ofadhesion molecules from cells. MEL-14/Leu-8/LECAM-1 is shed from neutrophils uponactivation (79, 80) and it is hypothesized by both groups that this shedding may have somefunctional significance to adhesion. In addition, a recent report has characterized a solubleform of CD43 present in plasma at concentrations exceeding of 10 mg/m1 (81)The general consensus to date from the published reports on soluble adhesionmolecules seems to be that they usually function as adhesion molecules when immobilized onplastic. However, there is some controversy as to the ability of soluble versions of adhesionmolecules to inhibit cell:cell adhesion.Thesis objectivesThe objectives of this thesis project were twofold: to explore the use of a soluble formof ICAM-1 as an inhibitor of immune responses, and to study in more detail some of themechanisms regulating cell adhesion mediated by ICAM-1 and LFA-1. Since soluble, murineICAM-1 was already being produced, the means to do this were readily available, and indeedthere were some encouraging reports in the literature along similar lines. The original goal was16first to functionally characterize sICAM-1 and then evaluate it's inhibitory potential through theuse of MLR as an in vitro model as well as GVHD as an in vivo animal model system. Thelong term goal we had in mind was clinical application to graft versus host disease associatedwith bone marrow transplantation. Through the functional characterization of sICAM-1, themolecule proved inefficient as an inhibitor of LFA-1:ICAM-1 mediated cell adhesion despitethe fact that the molecule appeared to be fully functional. Explaining the inability of sICAM-1to inhibit ICAM-1:LFA-1 mediated adhesion then became an important part of the project.Through out the work we encountered various mechanisms by which ICAM-1:LFA-1 mediatedcell adhesion is regulated, and we attempted to characterize these mechanisms in some detail.The Results section has been divided into five sections; the production and purificationof sICAM-1, the functional characterization of sICAM-1, the role of multivalent interaction inLFA-1:ICAM-1 mediated cell adhesion, the cell surface distribution of ICAM-1 and LFA-1,and the role of the cytoplasmic domain of ICAM-1 in cell adhesion. The Discussion section isa synthesis of the information presented in this thesis along with other salient information in theliterature, with the goal of clarifying some of the topics that are ambiguous in the literature.One of the main themes that I have attempted to develop is that a variety of mechanisms canpotentially regulate cell adhesion, even through a single pair of counter receptors such as LFA-1 and ICAM-1.17Materials and methodsAnimalsC57BL/6 and BALB/c mice were purchased from Charles River Canada Ltd. (St.Constant, Quebec, Canada) and were maintained in the Joint Animal Facility of the B.0 CancerResearch Centre.Cell cultureThe Moloney murine leukemia virus induced T cell lymphoma MBL-2 (82), themyeloma NS-1 (83), the B cell lymphoma A20 (84),the T cell hybridoma T28 (85), and thefibroblastoid cell line L (86), all murine cell lines used for various analyses, were maintained inDMEM supplemented with 5% FCS and antibiotics in a 37°C, 5%CO2 incubator.The hybridomas YN1/1.7, FD441.8 (ATCC TIB213), M1/ (ATCCTIB128), KM 81 (ATCC TIB 241), GK1.5, YE1/21.1, and YE1/9.9 were all maintained inDMEM supplemented with 5% FCS and antibiotics in a 37°C, 5% CO2 incubator. Allhybridomas secrete rat IgG directed against mouse antigens. YN1/1.7 secretes anti-murineICAM-1 (87), TIB213 secretes anti-LFA-1 a chain (88), TIB128 secretes anti-Mac-1 a chain(89), TIB241 secretes anti-CD44 (90), GK1.5 secretes anti-CD4 (91), YE1/21.1 secretes anti-CD45 (92), and YE1/9.9 secretes anti-transferrin receptor (92). Cultures of these cells wereovergrown, the SN was harvested by centrifugation at 5,000 xg for 45 min., and used as asource of specific Ab. For purification of YN 1/1.7 Ab, hybridoma cells were grown up inPFHM-II serum-free medium, and the SN from overgrown cultures was processed asdescribed below.Resting splenic T cells were purified by passage over a nylon wool column (93) inorder to deplete the population of adherent cells. Columns were prepared by packing 1 g ofdried nylon wool into a disposable 6 ml disposable syringe, and autoclaving. Nylon wool18columns were equilibrated overnight in a 37°C, 5% CO2 incubator with RPMI 1640supplemented with 5% FCS. C57BL/6 or BALB/c mice, 2-10 months old, were thensacrificed and single cell suspensions were made from the spleens in RPMI 1640 5% FCS.Cells from 1-2 spleens in a volume of 1 ml were then loaded on to a nylon wool column, andthe column incubated in a 37°C, 5% CO2 incubator for 1 hr. Non-adherent cells were theneluted with pre-warmed RPMI 1640 5% FCS for about 15 min., depending on the flow rate.Contaminating red blood cells were then lysed with Tris-NH4C1 (94). One spleen routinelyyielded approximately 2x10 7 nucleated cells. Cells purified by this procedure have beenshown by others to represent a population of resting T cells, greater than 90% in purity (93).Production and purification of sICAM-1The derivation of the expression system for the production of murine sICAM-1 waspreviously established in this laboratory by Daniel Lee and Fumio Takei (Lee and Takei,submitted). The overall scheme is that recombinant sICAM-1 is produced by NS-1 cells andthen purified from the culture SN by immunoaffinity chromatography.The starting point was the cloned full length cDNA encoding murine ICAM-1,originally named MALA-2 (95). Briefly, this cDNA was cut with Scal and EcoRI and ligatedinto pBLUESCRIPT which had previously been cut with EcoRI and XbaI and filled in. Thisblunt end ligation generated a stop codon immediately after the Scal site, which is convenientlylocated 21 base pairs 5' of the transmembrane encoding region of MALA-2. From this pointon, the murine version of ICAM-1, MALA-2, will simply be referred to as ICAM-1. ThecDNA encoding sICAM-1 was then cut out of pBLUESCRIPT using Sall and NotI, and sub-cloned into the XhoI and NotI sites of the eukaryotic episomal expression vector BCMGSNeo(96). The BCMGSNeo vector carries a CMV promoter which efficiently drives the expressionof the inserted cDNA in the appropriate eukaryotic host cell. BCMGSNeo-sICAM-1 was thenelectroporated into NS-1 cells, and transfectants were selected in 1 mg/ml G418 (GibcoLaboratories Life Technologies Inc., Grand Island, NY). Single cell cloning and screening by19immunoblot with YN1/1.7 Ab (see below) yielded the cell line 2706C, which secreted largeamounts of sICAM-1 into the culture SN.A YN1/1.7 Ab immunoaffinity column was prepared for the purification of sICAM-1.Ig was precipitated from overgrown, serum-free YN1/1.7 SN with 50% saturated ammoniumsulphate. The precipitate was then dialyzed against PBS, and analyzed by SDS-PAGE underreducing conditions to determine the purity of the Ab preparation. SDS PAGE analysis ofmaterial purified by this method demonstrated that the preparation was approximately 80% pureIg. The immunoaffinity column was made by coupling 3 ml of A1-1-IGEL 10 beads (Bio-Rad,Richmond, CA) with 6 mg of purified YN1/1.7 Ab in 0.1 M NaHCO3 overnight at 4°C. Thebeads were then washed with 100 mM Tris pH 8.0 to block residual coupling sites, and thenstored in 10 mM Tris pH 7.5, 0.15 M NaCI, 0.2% azide at 4°C.2706C cells were selected in 2 mg/ml G418 and then grown up in order to purifysICAM-1. Falcon 3028 flask (Becton Dickinson, Lincon Park, NJ) were seeded with 5 ml of2 mg/ml G418 selected culture each, along with 45 ml of DMEM 2% FCS + 0.5 mg/ml G418.On day 2, 50 ml DMEM 1% FCS was added to each flask, and on day 4, 100 ml DMEM 1%FCS was added to each flask. On day 7 the culture supernatant was harvested bycentrifugation at 5,000 xg for 45 min. at 4°C. 500 ml batches of SN were used to purifysICAM-1, and the entire purification was carried out at 4°C. The SN was loaded onto theYN1/1.7 column at a flow rate of approximately 20 ml/hr. The column was then washed with200-400 ml of 0.15 M NaCI, 10 mM Tris pH 7.5 at a flow rate of approximately 25 ml/hr.SICAM-1 was then eluted with 0.15 M NaC1, 50 mM diethylamine pH 11.0 at a flow rate ofapproximately 30 ml/hr. 5 ml fractions were collected into tubes containing 500 pl of 1 M TrispH 6.0. The fractions were monitored by measuring OD280, and fractions containingsignificant amounts of protein were pooled, concentrated using Centricon-30microconcentrators (Amicon, Beverly, MA), exchanged 3 times into PBS, and adjusted to aconcentration of 1 mg protein (OD280) per ml. The purified sICAM-1 was then aliquoted andstoned at -20°C. 500 ml batches of SN routinely yielded 1-2 mg of purified sICAM-1.20ImmunoblotAn immunoblot assay was used to screen cell lines for the production of sICAM-1, toassess levels of sICAM-1 production, and to monitor the purification of sICAM-1. TheMinifold II slot blot apparatus (Schleicher and Schuell, Keen, NH) was assembled accordingto the manufacturers instructions, and with vacuum applied, samples were loaded. Usually 10and 100 pl samples of culture SN were loaded into individual slots. Each slot was washedwith a 200 p1 aliquot of PBS, and the apparatus was disassembled. The nitrocellulosemembrane was then washed in 2 changes of PBS and then incubated in PBS 5% blotto + 0.1%NP40 for 10 min. at 20°C. After washing with 3 changes of PBS, the membrane wasincubated in YN1/1.7 SN diluted 1/40 with PBS for 1 hr. at 4°C. The membrane was thenwashed in 2 changes of PBS 5% blotto, 2 changes of PBS, and incubated in peroxidaseconjugated goat anti-rat IgG (Bethesda Research Laboratories Life Technologies Inc.,Burlington, Ont.) diluted 1/2,000 with PBS for 30 min. at 4°C. After 2 washings in PBS 5%blotto and 2 washings in PBS, PBS containing 0.06% DAB (Bethesda Research LaboratoriesLife Technologies Inc., Burlington, Ont.) and 0.012% H202 was added, and the blot wasdeveloped for 5-10 min.SDS-PAGE analysisPreparations of purified sICAM-1 were analyzed by SDS-PAGE in order to confirmmolecular weight, purity, and quantitation. This analysis was carried out using a Mini-V 8.10apparatus (Bethesda Research Laboratories Life Technologies Inc., Burlington, Ont.)according to standard conditions (97). Samples of purified sICAM-1 or known amounts ofstandard proteins (in order to confirm the quantitation) were diluted 1/2 with 2X non-reducingSDS-PAGE sample buffer, and electrophoresed on 7.5% discontinuous gels. The gels werethen removed from the apparatus and stained in a solution of 2% coomassie brilliant blue, 50%methanol, and 10% acetic acid, and then destained in a solution of 5% methanol and 7.5%21acetic acid. Molecular weights were determined from the Bio-Rad high range SDS-PAGEstandards (Bio-Rad, Richmond, CA). For radio-iodinated samples, rather than staining, thegel was wrapped in plastic wrap and exposed to X-ray film.Radio-iodination of sICAM-1Purified sICAM-1 was radio-iodinated by the IODO-GEN method (98). IODO-GEN(Pierce, Rockford IL) was dissolved in chloroform and 100111 aliquots containing 1014/10041 were dispensed into 12x75 mm borosilicate glass tubes, and the chloroform was thenallowed to evaporate overnight. In a fume hood, 10 lig of purified sICAM-1 was mixed with 5of 100 mCi/m1 Na 1251 (Dupont New England Nuclear, Wilmington DE) in a final volume of50 of PBS. This mixture was then transferred to an IODO-GEN coated tube, and iodinationwas allowed to proceed for a period of 2 min., after which the reaction was stopped by theaddition of 10 IA of 2 mg/ml sodium metabisulfate. The iodinated protein was then separatedfrom unincorporated iodine by gel filtration through a Nick column (Pharmacia, Uppsala,Sweden) which had previously been equilibrated with PBS containing FCS and then PBS.Fractions eluted from the Nick column corresponding to radio-iodinated protein were thenpooled, made to 10% FCS, aliquoted, and stored at -20°C. 125I-sICAM-1 concentration wascalculated based on an estimated overall recovery of 85%. SICAM-1 was routinely iodinatedby this procedure to a specific activity of 2-5x104 CPM/ng. Radio-iodinated sICAM-1 wasthen used either for SDS-PAGE analysis, or for 1251-sICAM-1 binding.125I-sICAM-1 binding assayThe 125I-sICAM-1 binding assay is a functional assay for sICAM-1 in which thebinding of radio-iodinated sICAM-1 to LFA-1+ cells is quantitated. Parameters for thestandard 125I-sICAM-1 binding assay were previously worked out in Fumio Takei's lab byDaniel Lee (Lee and Takei, submitted). LFA-1+ lymphoid cells (MBL-2, A20, T28) weresuspended to a concentration of 2x106/ml in binding buffer (either RPMI 1640 supplemented22with 5% FCS and antibiotics or Tris buffered saline with 1 mM Ca++ and Mg++), and theappropriate inhibitors or stimulators were added. All conditions were done in duplicate. 125I-sICAM-1 was then added to a final concentration of between 200 ng/ml (2x10 -9 M) and 2ug/m1(2x10 -8 M), and the cells were incubated for 1 hr. at 20°C. At the end of this 1 hr.incubation, cell bound radioactivity was determined by one of two methods. For high affinitybinding, cells were washed 3X binding buffer and then counted in a gamma counter; for lowaffinity binding cells were spun through oil instead of being washed. To spin through oil, 100aliquots of the cells were layered onto 300 [t1 aliquots of a mixture of 3 parts dibutyl pthalate(Sigma, St. Louis, MO) and 2 parts dioctyl pthalate (Aldrich, Milwaukee, WIS) andcentrifuged at maximum speed in an Eppendorf Microfuge for 2 min. at 20°C. The tubes werethen frozen in dry ice. The bottoms of the frozen tubes containing the pelleted cells were cutfrom the rest of the tube with a dog toe nail clipper, and the cell pellets containing only bound125I-sICAM-1 were counted in a gamma counter. Results are expressed as the arithmetic meanof duplicates, and the error bars represent the standard deviation.Quantitation of the density of sICAM-1 in sICAM-1 coated microtitre wellsIn order to better understand the dynamics of the quantitative cell adhesion assay (seebelow), the density of sICAM-1 on sICAM-1 coated polystyrene microtitre wells wasquantitated. This quantitation was accomplished with radio-iodinated YN1/1.7 Ab.Purified YN1/1.7 Ab was radio-iodinated by the chloramine T method (99). 101.11 of afreshly prepared solution of 0.5 mg chloramine T/ml and 10 IA of 100 mCi/m1Na 125I (DupontNew England Nuclear, Wilmington DE) was added to 50 lig of purified YN1/1.7 Ab in 167 [1.1of PBS. Iodination was allowed to proceed for a period of 5 min., after which the reactionwas stopped by the addition of 50 ill of 2 mg/ml sodium metabisulphate. The iodinated proteinwas then separated from unincorporated iodine by gel filtration through a Nick column whichhad previously been equilibrated with PBS containing FCS and then PBS. Fractions elutedfrom the Nick column corresponding to radio-iodinated protein were then pooled, made to 10%23FCS, aliquoted, and stored at -20°C. 1251_ YN1/1.7 Ab concentration was calculated based onan estimated overall recovery of 85%. YN1/1.7 Ab was radio-iodinated by this procedure to aspecific activity of approximately lx105 CPM/ng.Serial dilutions of purified sICAM-1 were made in 0.1 M carbonate buffer pH 9.6.Duplicate 30 p.1 aliquots of these dilutions were then dispensed into the wells of flat bottom 96well Nunc tissue culture plates (Nunclon Delta, Denmark), and protein was allowed to absorbonto the plastic for 1 hr. at 20°C. The wells were then washed 3X PBS, and residual proteinbinding sites were blocked by incubation with 200 1.11 of PBS 5% blotto/well for 30 min. at20°C. Wells were then washed 3X PBS and incubated with 125I-YN1/1.7 Ab at 20 iig/ml, aconcentration previously determined to be saturating, for 1 hr. at 20°C. Wells were thenwashed 6X PBS, and bound 1251-YN/1.7 Ab was dissociated with 2 M NaOH. Aliquots ofthe dissociated 125 I-YN1/1.7 Ab were then counted in a gamma counter. Again, all conditionswere done in duplicate, with less than 10% variation in CPM between duplicates. From thebound CPM/well (arithmetic mean of duplicates) and the specific activity of the 125I-YN1/1.7Ab, the number of molecules of YN1/1.7 bound/well was calculated, and from this numberand the area of the the microtitre well surface, the density of bound YN1/1.7 Ab molecules andtherefore the density of sICAM-1 absorbed to the well was estimated. These calculatedsICAM-1 densities are likely to have a relatively large error associated with them, especially athigh sICAM-1 densities, since 1 YN1/1.7 Ab molecule can potentially bind to 2 molecules ofsICAM-1.Quantitative cell adhesion assayThe quantitative cell adhesion assay is a functional assay for sICAM-1 in which theability of LFA-1+ cells to bind to plastic absorbed sICAM-1 is quantitated. The protocol forthis assay was based on the protocol of Dustin and Springer (41). SICAM-1 was coated to thewells of 96 well flat bottom tissue culture plates as described above for the quantitation of thedensity of sICAM-1 in sICAM-1 coated wells. All conditions were done in duplicate. After24coating, wells were washed 3X PBS and blocked by the incubation of 100111/well of 100µg/ml BSA or OVA in PBS for 30 min. at 20°C. Wells were then washed 3X PBS, and 100!_t1 of cells at a concentration of 2x 10 6/m1 in RPMI 1640 5% FCS (2x10 5 cells/well) weredispensed into each well. In some experiments the cells were incubated for 20 min. at 37°Cwith various inhibitors or stimulators prior to addition to the assay plate. The plate was thencentrifuged at 10Xg for 5 min. at 20°C, and then incubated for 5 min. in a 37°C, 5% CO2incubator. At the end of this 5 min. incubation, unbound cells were carefully washed from thewells. Wells were washed by flicking the plate of it's contents and gently pipetting 150 ill ofmedium warmed to 37°C into each well, then repeating this process 5 more times. After thefinal wash, 100 of medium was added to each well. Also at this time 100 1.11 aliquots of cellsdiluted to various known concentrations were added in order to generate a standard curve forprecise quantitation.Bound cells were quantitated in the quantitative cell adhesion assay using a convenientcolorimetric assay in which the tetrazolium salt MIT is cleaved by mitochondrial enzymes ofviable cells, yielding a colored product (100). 10 of a solution of 5 mg/ml Mrf (Sigma, St.Louis, MO) in PBS was dispensed into each well of the assay plate, and the plate wasincubated for 4 hr. in a 37°C, 5% CO2 incubator. After this incubation, the colored productwas solubilized by the addition of 100 Ill of 0.04 M HC1 in isopropanol to each well andvigorous pipetting. The absorbances of individual wells was then read on a Bio-Tek ELISAplate reader (model EL309) using a test wavelength of 570 nm and a reference wavelength of630 nm. Nylon wool T cells usually generated a linear standard curve over the range of cellconcentrations used, while the cell lines usually generated a standard curve best described by apolynomial expression. The number of cells bound per well in each well of the experimentwas then determined by comparison to the standard curve. Again, all conditions were done induplicate and the results represent the arithmetic mean. Duplicates usually varied by less than5% in this assay.25In order to study the affects of Mn++ in the quantitative cell adhesion assay, severalalterations to the standard protocol had to made. Firstly, since 1 mM Mn++ resulted in theprecipitation of MnPO4 from standard RPMI 1640, PO4-free RPMI 1640 had to be used.Secondly, it was found that treatment of primary, resting T cells with mM concentrations ofMn++ resulted in a substantial inhibition of the cells' metabolic activity as measured by theMTT assay. Treatment of cell lines with mM concentrations of Mn++ resulted in only a slightinhibition of metabolic activity as measured by the MTT assay. Therefore, the MIT assaycould not be used to quantitate the adhesion of primary T cells to absorbed sICAM-1 in thepresence of mM concentrations of Mn++. The Mn++ induced adhesion of primary T cells toabsorbed sICAM-1 wells was quantitated by using cells labeled with 51Cr or simply byinspecting the degree of cell binding in the wells of the assay plate. Cells eluted from the nylonwool column were labeled with 51 Cr by incubating approximately 2x10 7 cells for 1 hr. at 37°Cin 50 tl FCS with 80 tCi 51 Cr (80 tl Na251 Cr04 in normal saline (Dupont New EnglandNuclear, Wilmingdon DE)). Excess 51 Cr was then removed by two washes with PO4-freeRPMI 1640 supplemented with 10 mM Hepes pH 7.2 and 5% FCS, and then cells were usedas in the standard protocol. After spinning the cells onto the plate and incubating at 37°C for 5min., wells were washed 8X 150 tl of pre-warmed PO4-free RPMI 1640 supplemented with10 mM Hepes pH 7.2, with or without 1.25 mM Mn++. Instead of washing by flicking, 51 Crlabeled cells were washed by aspirating the contents of the wells through a 18 ga. needle.After washing the wells, bound cells were dissociated by incubating the wells with 100 ulaliquots of 5 mM EDTA in PBS for 10 min. at 37°C. The contents of each well was thenmixed by pipetting and removed for y-counting. 51 Cr labeled primary T cells were observed tospontaneously release a significant amount of label, especially over the 6 washes in the assay,making quantitation difficult and somewhat un-reproducible from experiment to experiment.26Inhibition of the PMA induced aggregation of MBL-2 cellsMurine cell lines were screened for the ability to form homotypic aggregates uponstimulation by the phorbol ester PMA. Of the cell lines tested MBL-2, a Moloney murineleukemia virus induced LFA-1 and ICAM-1 positive T lymphoma cell line, aggregatedreproducibly to the greatest extent. The aggregation of MBL-2 cells was found to be dependanton LFA-1 and ICAM-1 (Figure 7). MBL-2 cells did not, however, aggregate to the extent ofcertain human cell lines under similar conditions (11).MBL-2 cells in early logarithmic phase were suspended to a concentration of 4x10 6/m1in RPMI 1640 supplemented with 10% FCS and antibiotics, and 100 Ill aliquots weredispensed into the wells of flat bottom microtitre plates (2x10 5 cells/well). To the appropriatewells 5 ill volumes of hybridoma SN were added as inhibitors, as were dilutions of EDTA,cytochalasin B, and volumes of purified sICAM-1. Then 50 IA aliquots of PMA (Sigma, St.Louis, MO), diluted to 200 ng/ml in medium from a stock of 201.tg/m1 in DMSO, weredispensed into the appropriate wells. Finally, the volume of each well was made up to 200 1.11with medium and the contents of each well was mixed by pipetting. The final concentration ofPMA in this assay was 50 ng/ml, the final concentration of EDTA was 5 mM, and the finalconcentration of cytochalasin B was 20 p.M. The plate was then incubated in a 37 0C, 5% CO2incubator for a period of 5 hr., after which the individual wells were photographed on aninverted microscope.Inhibition of the Mn++ induced aggregation of MBL-2 cellsCell lines (A20 and MBL-2) were observed to aggregate when suspended in 1 mMMn++ and used for the 125I-sICAM-1 binding assay. To determine if this phenomenon wasmediated by LFA-1 and ICAM-1, a qualitative aggregation assay similar to the PMA inducedaggregation assay was carried out. MBL-2 cells were suspended at 4x106/ml in 10 mM HepespH 7.2, 0.15 M NaC1, 5% FCS. Cells were dispensed in eppendorf tubes, and to the tubesvolumes of Ab SNs, EDTA or purified sICAM-1 were added followed by volumes of 1 MMnC12 and more Hepes/NaC1 to give a final concentration of 1.25 mM MnC12 and 2x10 627cells/ml. The tubes were then incubated at 37°C for 45 min., at which time the cells wereresuspended by vortexing gently, and 100 p.1 aliquots were dispensed into the wells of flatbottom microtitre plates. The cells were allowed to settle in the wells for 5 min. and thenindividual wells were then photographed on an inverted microscope.Preparation of sICAM-1 coated polystyrene beadsUniform polystyrene beads, 2 p.m in diameter (Seradyn, Indianapolis IN), were coatedwith sICAM-1 by physical absorption in a fashion similar to the coating of polystyrenemicrotitre wells. Absorbed surfactant and low molecular weight contaminants had to beremoved from the beads before coating with sICAM-1, and this was accomplished bydiafiltration in an Amicon 8010 stirred ultrafiltration cell (Amicon, Beverly, MA). 200 pl ofbeads (3.7x109 beads) were washed in the ultrafiltration cell, using a 300,000 molecularweight cut off membrane, 2X dH2O, 2X 0.1 M carbonate buffer pH 9.6, and then suspendedin a volume of 1-2 ml 0.1 M carbonate buffer pH 9.6. 10 ill (10 p.g) of purified sICAM-1 wasadded to 100 of beads and the tube was gently agitated for 1 hr. at 20°C. Control beadswere similarly coated with BSA. After coating, beads were washed once by centrifugation(2,500 RPM, 10 min.), and blocked by a 30 min. incubation at 20°C in 100 pl of PBScontaining 50 lig of BSA. Finally the beads were washed by centrifugation twice with RPMI5% FCS.Quantitation of sICAM-1 coated polystyrene beadsThe amount of sICAM-1 coated to the sICAM-1 coated beads was determined using anELISA in which the detection of sICAM-1, absorbed to microtitre wells, by YN1/1.7 Ab isinhibited by sICAM-1. SICAM-1 was coated to the wells of microtitre wells as described forthe quantitation of sICAM-1 in sICAM-1 coated microtitre wells. For the ELISA, wells werecoated to approximately 1,600 sites/µm 2 , and blocked by incubation with PBS 5% blotto for30 min. at 20°C. After washing wells 3X PBS, various volumes of sICAM-1 or BSA coated28beads or known quantities of purified sICAM-1 were added to duplicate wells. Then 100 pl ofYN1/1.7 SN diluted 1/80 (a dilution previously determined to be near the end of the titration ofthis SN by ELISA) was dispensed into each well and incubated for 30 min. at 20°C. Wellswere then washed 6X PBS and incubated with peroxidase conjugated goat anti-rat IgG(Bethesda Research Laboratories Life Technologies Inc., Burlington, Ont.) diluted 1/2,000 inPBS 5% blotto for 30 min. at 20°C. After washing the wells 6X PBS, the wells wereincubated with 100 pl of substrate consisting of 2 mg/ml OPD (Sigma, St. Louis, MO) in 0.1M NaPO4 pH 6.0 with 0.012% H202, for 15 min. in the dark. The reaction was then stoppedby addition of 25 pl of 3 M HC1/well, and the absorbances of individual wells was read on aBio-Tek ELISA plate reader (model EL309) using a single wavelength of 490 nm. Allconditions were done in duplicate, and duplicates usually varied by less than 10%. The %inhibition was calculated by the equation:specific A490( - inhibitor) specific A490(+inhibitor) % inhibition — specific A490(-inhibitor)^X 100%Immobilization of MBL-2 cells to microtitre plates via poly-L-lysine andsubsequent binding of sICAM-1 coated beadsIn order to initially visualize the binding of sICAM-1 coated beads to LFA-1+ cells,MBL-2 cells were immobilized to plastic via poly-L-lysine (Sigma, St. Louis, MO) and thenallowed to bind to coated beads. Flat bottom microtitre wells were coated with 501.11/well of100 pg/m1 poly-L-lysine in dH2O for 2.5 hrs. at 37°C, after which time the wells were washed3X PBS. MBL-2 cells were washed once with PBS, adjusted to 2x106 cells/ml in PBS, and100 IA was dispensed into each poly-L-lysine coated well. The plate was then centrifuged for2 min. at 620 RPM and incubated for 30 min. at 20°C. Unbound cells were then washed awayby gently washing the wells 3X PBS. 100 pl of RPMI 1640 supplemented with 10% FCS andantibiotics was then dispensed into each well, followed by 10 pl of either sICAM-1 or BSA29coated beads. The plate was incubated for 45 min. in a 37°C 5% CO2 incubator, and thenunbound beads were washed by gently washing the wells 6X with prewarmed, serum freeRPMI 1640. Individual wells were photographed on an inverted microscope.Fluorescent staining of cellsLymphocytic cells were stained in suspension with specific Ab against various cellsurface molecules. Prior to staining, cells were fixed in order to 'freeze' the distribution of cellsurface molecules, and to prevent Ab induced patching. In some situations, cells werestimulated with PMA or treated with cytochalasin B for periods of time before staining, and inthese situations the cultures were treated at 37°C and then chilled on ice before fixation andstaining.Cells were fixed in a solution of 1% paraformaldehyde in PBS. This concentration ofparaformaldehyde was previously determined to effectively fix cells by the prevention of lysisin dH2O. The paraformaldehyde solution was prepared fresh daily by dissolving 1 g ofparaformaldehyde in 50 ml of dH2O with heating and the addition of 1 drop of 1 M NaOH.The dissolved paraformaldehyde was then cooled to 20°C, 50 ml of 2XPBS was added, andthe solution was chilled on ice. Lymphocytic cells chilled on ice were disaggregated by gentlepipetting, and aliquoted into eppendorf tubes. Cells were washed once with ice cold PBS,suspended in cold 1% paraformaldehyde in PBS, and incubated on ice for 20 min. withoccasional gentle agitation. Fixed cells were then washed once with PBS and twice withHanks containing 5% FCS.Fixed lymphocytic cells were stained with primary Abs directed against cell surfacemolecules by incubation in undiluted tissue culture SN for 1 hr. on ice. After 3 washes withHanks 5% FCS, cells were then incubated for 45 min. on ice in FITC conjugated goat anti-ratIgG (Cappel, Cooper Biomedical, Malvern, PA) diluted 1/40 with Hanks 5% FCS. In somecases cells were also stained with TRITC-phalloidin (101, Sigma, St. Louis, MO) at the sametime as they were stained with 2°Ab. Phalloidin is a protein derived from certain mushrooms30which binds specifically and with a very high affinity to F-actin (102). It has also been shownthat this small molecular weight molecule can enter formaldehyde fixed but not viable cells(101). Following the secondary staining cells were washed 3X Hanks 5% FCS and cell pelletswere suspended in between 15-50p1 of Hanks containing 2% DABCO (Aldrich, Milwaukee,WIS) as an anti-photobleaching agent.Adherent mouse L cells were also stained following paraformaldehyde fixation, but thecells were handled differently. Monolayers of L cells were lifted with 5 mM EDTA and platedinto dishes containing autoclaved microscope coverslips. When cells had grown to the desireddensity, the cells attached to the coverslips were fixed and stained. Coverslips were washed inPBS once and then incubated in 1% paraformaldehyde in PBS for 20 min. at 20°C.Coverslips containing fixed cells were then washed 1X PBS, 2X Hanks 5% FCS, and stainedwith primary Abs directed against cell surface molecules by incubation in tissue culture SN for1 hr. at 20°C. After 3 washes with Hanks 5% FCS, coverslips were then incubated for 45min. in FITC conjugated goat anti-rat IgG (Cappel, Cooper Biomedical, Malvern, PA) diluted1/40 with Hanks 5% FCS. After staining with secondary Ab, coverslips were washed 3XHanks 5% FCS and then kept in serum free Hanks prior to mounting.Immunofluorescence microscopyFor lymphocytic cells, slides were made by mounting 10 of cell suspension on amicroscope slide and sealing the coverslip with melted wax. Coverslips with attached, stainedL cells were mounted on microscope slides in 10 ill of Hanks 2% DABCO, and sealed withmelted wax. Slides were viewed on a Zeiss microscope, equiped with epifluorescence, using aLeitz 40X/0.75 objective. Photographs were taken with either Kodak Ektachrome ASA 400 orP800/1600 film; using either phase contrast optics, epifluorescence and the green filter forvisualizing FITC, or epifluorescence and the red filter for visualizing TRITC. Shutter speedsfor exposures under phase contrast were approximately 1/2 sec. Shutter speeds for exposures31under fluorescence varied from 5 sec. with the 400 speed film (for high fluorescence intensity)to 15 sec. with the 800 speed film (for low fluorescence intensity).Creation of cell lines expressing ICAM-1 with or without it's cytoplasmic tailThis work was initiated in Fumio Takei's lab by Mike Butchart, a summer student whoput together all of the DNA constructs and did the initial transfections. For transfection of thecomplete ICAM-1 molecule, the cDNA encoding complete, functional murine ICAM-1(95) wassubcloned into BCMGSNeo. For transfection of ICAM-1 without it's cytoplasmic domain, alittle more work was involved. Using PCR and the above cDNA as a template, a fragmentencoding part of extracellular domain 4, extracellular domain 5, the transmembrane domain,and the first two amino acids of the cytoplasmic domain was generated. The PCR primerswere designed such that the 5' end of this fragment had a HindIII restriction site. A fragmentencoding the first 3 extracellular domains and part of the fourth, obtained by HindIII digestionof the cDNA, was then ligated to the PCR fragment, and the result was a cDNA encodingICAM-1 without the majority of it's cytoplasmic domain. This construct was then subclonedinto the BCMGSNeo expression vector, where the final construct encodes all of theextracellular domains, the transmembrane region, the first two amino acids of the cytoplasmicdomain, and finally two irrelevant amino acids. These two versions of ICAM-1, each inBCMGSNeo, were then electroporated into T28 cells, and also transfected into L cells bycalcium phosphate mediated transfection.T28 transfectants were selected in 1 mg/ml G418, and FACS analysis of initialtransfectant populations showed very good expression (data not shown). L transfectants werealso selected in 1 mg/ml G418; however, initial expression was poor. High expressers weresorted by FACS and selected again in 1 mg/ml G418. This process yielded populations of Lcell transfectants exhibiting strong ICAM-1 expression. In both the transfected T28 and Lcells, the cytoplasmic deletion version of ICAM-1 was expressed just as well as the completeICAM-1 molecule (data not shown).32Analysis of cell lines expressing transfected ICAM-1 with or without it'scytoplasmic domainThe transfected T28 cells were compared to the parental T28 line in a PMA inducedhomotypic aggregation assay similar to that described above for MBL-2 cells. Parental T28cells and the two ICAM-1 transfectant lines were suspended to 2x10 6 cells/ml in RPMI 1640supplemented with 10% FCS and antibiotics, and 1 ml aliquots with or without 50 ng/ml PMAwere dispensed into the wells of a 24 well plate. Only 1 hr. of PMA stimulation was requiredto achieve efficient aggregation of the ICAM-1 transfectants, as compared to the 4-5 hr.required for MBL-2 cells. These 1 ml cultures were then chilled on ice and analyzed byimmunofluorescence microscopy after fixation and fluorescent staining of ICAM-1 and othercell surface molecules, as described above.FACSorted, 1 mg/ml G418 selected L cell transfectants were analyzed byimmunofluorescence microscopy after fixation and fluorescent staining of ICAM-1 and othercell surface molecules as described above. This analysis, along with the FACS analysis,demonstrated extremely high levels of ICAM-1 expression, possibly swamping out any subtledistribution effects. The L cell transfectants were therefore grown for approximately 2 weeksin only 100 tg/ml G418 in an attempt to lower the level of ICAM-1 expression, andimmunofluorescence analysis was repeated.33ResultsPart I: Production and purification of sICAM-1Previously, the means and the methodology for the production and purification ofsICAM-1 had already been worked out in Fumio Takei's lab (Lee and Takei, manuscript inpreparation). Briefly, the cDNA encoding a soluble version of murine ICAM-1 (MALA-2)was subcloned into the eukaryotic expression vector BCMGSNeo (96). The cDNA encodingsICAM-1 encodes amino acids 1-453, with the transmembrane sequence starting at amino acid461 in the complete ICAM-1 protein. The recombinant vector was electroporated into NS-1cells and transfectants were then selected in 1 mg/ml G418. Culture SNs were screened for thepresence of ICAM-1 by immunoblot. Results of such an immunoblot are shown in Figure 2.Detection of sICAM-1 by this method was both rapid and sensitive. Single cell cloning and re-screening yielded the clone 2706C which secreted large quantities of sICAM-1 into the cultureSN. This cell line was used for subsequent large scale protein production. 2706C culture SNwas loaded onto a YN1/1.7 Ab column, the column was then washed extensively with 0.15 MNaC1, 10 mM Tris pH 7.4. sICAM-1 was eluted with 0.15 M NaCl, 50 mM diethylamine, pH11.0. 500 ml batches of SN generally yielded approximately 1 mg of protein as estimated by0D280. Eluted material from the column was then either concentrated and exchanged into PBSor re-applied to the column.Molecular weight, purity, and yield of sICAM-1 was then confirmed by SDS-PAGEanalysis. Typical SDS-PAGE results are shown in Figure 3. In this experiment approximately1 pg of each of two independently purified preparations were run on SDS-PAGE, and the gelsubsequently coomassie stained. Purified sICAM-1 migrates with a relative molecular mass ofapproximately 93,000 under nonreducing conditions. This size is in good agreement with thepredicted molecular weight for sICAM-1, since murine ICAM-1 was found to migrate with arelative molecular mass of 95,000 to 100,000 under reducing or nonreducing conditions (87).34The cloned murine ICAM-1 contains 512 amino acids (95), and 59 amino acids were deletedfrom the original molecule in order to obtain the soluble version.A more stringent assessment of the purity of the purified sICAM-1 preparations wasobtained from the analysis of radio-iodinated sICAM-1. Purified sICAM-1 was radio-iodinated, using the IODO-GEN method, to a specific activity of approximately 4x10 4CPM/ng. 125I-sICAM-1 was then run on SDS-PAGE and the gel subsequently exposed to X-ray film. Figure 4 shows the results of such an experiment in which approximately 25 ng ofeach of two preparations, one purified by one cycle of affinity chromatography and onepurified by two cycles of affinity chromatography, were analyzed. The results clearly showthat the material purified by one round of affinity chromatography is very pure. Proteinpurified by one cycle of affinity chromatography was used for all subsequent experiments.To summarize the results presented in Part I, highly purified, recombinant murinesICAM-1 was produced. With sICAM-1, two things could be accomplished. Firstly, it's useas an inhibitor of LFA-1:ICAM-1 mediated cell adhesion and immune responses dependent onLFA-1:ICAM-1 interaction could be explored. Secondly, the fine details of LFA-1:ICAM-1mediated cell adhesion could be explored.353 6Figure 2 Initial detection of sICAM-1 in culture supernatants and following affinitychromatography purification. Samples were blotted onto nitrocellulose andsICAM-1 was detected by the indirect immunoperoxidase method usingYN1/1.7 Ab, peroxidase conjugated goat-anti-rat IgG, and DAB/H202substrate. For samples 1 to 4, (a) represents 10 Ill of culture supernatantand (b) represents 100 IA of culture supernatant. The samples were from 1)an intermediate producing cell line, 2) a low producing cell line, 3) the highproducing cell line used for large scale protein production, 4) controluntransfected NS-1 supernatant, and 5) affinity chromatography purifiedsICAM-1. For 5, (a) represents 5 i.tg of purified protein and (b) represents5014 of purified protein.Figure 3^Confirmation of molecular weight. purity, and yield of purified sICAM-1. Approximately 1 lag of each of two independently purified batches ofsICAM-1 were run on a 7.5% SDS-PAGE and coomassie stained.Molecular weight markers were Bio-Rad high range standards, andmolecular weights are indicated.3 8Figure 4 Assessment of the purity of radio-iodinated sICAM- 1, SICAM- 1 was radioiodinated by the IODO -GEN method and aliquots containing approximately25 ng (approximately 2x106 CPM) of 125I-sICAM-1 each were run on7.5% SDS-PAGE and the gel exposed to X-ray film. The batches ofprotein were purified a) by one cycle of affinity chromatography or b) bytwo cycles of affinity chromatography.Part II: Functional characterization of sICAM-1The first functional assay for sICAM-1 was it's ability, in solution, to bind to LFA-1positive cells. Purified sICAM-1 was radio-iodinated using the IODO-GEN method to aspecific activity of approximately 4x10 4 CPM/ng. The results of several 1251-sICAM-1binding assays are shown in Figure 5. Here MBL-2, A20 or T28 cells were incubated at roomtemperature with various concentrations of 125I-sICAM-1 for 1 hr., in the presence or absenceof excess unlabeled sICAM-1 or EDTA as inhibitors. Figure 5 clearly shows that 125I-sICAM-1 is capable of binding to LFA-1 positive cells, and in addition since this binding is inhibitedboth by unlabeled sICAM-1 and the divalent cation chelator EDTA, the binding is most likelymediated specifically by LFA-1. In Figure 5 (a) there appears to be relatively high affinitybinding, while in Figure 5 (b) and (c) the binding appears to be of relatively low affinity. Thehigh affinity binding shown in Figure 5 (a) is consistent with the binding assay results obtainedby Daniel Lee (Daniel Lee, unpublished data). However, in my hands high affinity bindingwas much more difficult to demonstrate, and most of the experiments yielded results as shownin Figure 5 (b) and (c). In addition, Figure 5 (c) demonstrates that the addition of 1 mM Mn++enhances the binding of 125I-sICAM-1 to it's receptor. The observed effect of Mn++ is inagreement with a recent report concerning the effect of Mn++ on the affinity of Mac-1 for it'sligands (103), and a report on the effects of Mn++ on a 133 integrin (104). Indeed a recentreport has also shown that Mn++ increases both the adhesion of T cells to absorbed ICAM-1 aswell as the existence of an LFA-1 activation epitope (105).39Figure 5 125I-sICAM-1 binds to LFA-1 positive cells. Purified sICAM-1 was radio-iodinated by the IODO-GEN method to a specific activity of approximately4x104 CPM/ng, and it's binding to the LFA-1 positive cell lines MBL-2 (a),A20 (b), and T28 (c) was determined. In (a) binding was carried out inRPMI 1640 5% FCS at an 1251-sICAM-1 concentration of 200 ng/ml(2x10-9 M) for 1 hr., then the cells were washed 3X and cell associatedradioactivity was determined. In (b) binding was carried out in RPMI 16405% FCS at an 1251-sICAM-1 concentration of 21.1,g/ml (2x10 -8 M) for 1 hr.,then aliquots of cells were spun through oil and cell associated radioactivitywas determined. In (c) binding was carried out in Hepes buffered salinewith 1 mM CaC12, 1 mM MgC12, with or without 1 mM MnC12; cell boundradioactivity was determined as for (b). In all cases unlabeled sICAM-1(cold inh.) was used in a 100 fold excess to inhibit binding, while EDTAwas used at a final concentration of 5 mM. All incubations were done atroom temperature.40ano inhibitorcold inhibitor0 1000 2000^3000bound CPM4000 5000no inhibitorcold inhibitor0^2000 4000 6000 8000 10000 12000bound CPM4142C^bound CPM-Mn/no inh.-Mn/EDTA+Mn/no inh.+Mn/EDTA0^2000^4000^6000^8000.^1^.^1^• NNW-Mn/no inh.-Mn/EDTA+Mn/no inh.+Mn/EDTA0^2000^4000^6000^8000bound CPMThe second functional assay for sICAM-1 was the quantitative cell adhesion assay. Inthis assay purified sICAM-1 is coated to the wells of polystyrene microtitre plates, and thenLFA-1 positive cells are allowed to bind to the absorbed sICAM-1.In order to estimate what the density of sICAM-1 was on the sICAM-1 coated plastic,and to compare this to the physiological situation, I quantitated the sICAM-1 absorbed toplastic used for the quantitative adhesion assay. This quantitation was done with 125I-YN1/1.7Ab. The results of this quantitation are shown in Figure 6. For the sake of comparison,Kurzinger et al. have estimated that the density of LFA-1 on human peripheral bloodlymphocytes is approximately 150 molecules/w2 (106). Therefore the density achievable inthe quantitative cell adhesion assay, approximately 1,600 molecules/µm2 , is significantlyhigher than what is likely to exist in the physiological situation.Figure 7 shows the results of an early quantitative cell adhesion assay. For thisexperiment, wells were either coated to a high density of sICAM-1 and then subsequentlyblocked with BSA, or just blocked with BSA. This experiment was done with the the murineB lymphoblastoid cell line A20, which expresses LFA-1 on it's cell surface. It is important tonote the way this assay is done: the cells are added to the wells, the plate is spun to bring thecells into contact with the coated surfaces, and then the plate is incubated at 37°C for only 5min. before the unbound cells are washed away. 2x10 5 cells are added to each well; therefore,almost 100% of the input cells bind under these conditions. Cytochalasin B, an agent whichinhibits the actin cytoskeleton by binding specifically to actin and preventing the polymerizationof actin (107), effectively inhibits the binding of A20 cells in this assay. This experimentshows that sICAM-1 in solution, at a concentration of up to 100 µg/ml, does not inhibitadhesion.43•1000 -0200044if)E0-acT0 10 20sICAM-1 coating concentration (ug/ml)Figure 6 Quantitation of the density of sICAM-1 in sICAM-1 coated microtitre wells.Serial dilutions of purified sICAM-1 were coated to the wells of flat bottommicrotitre plates, the wells were then blocked and incubated with asaturating concentration of 1251-anti-ICAM-1. The amount of bound 1251anti-ICAM-1 was then determined, and from the quantity of 125I-anti-ICAM-1 bound to wells, the density of sICAM-1 was calculated. Thisstandard curve was used to estimate the density of sICAM-1 for all of thesubsequent quantitative cell adhesion assays. no inh.cyto. BsICAM-145co153>,c0no inh.100000^200000cells bound/wellFigure 7 SICAM-1 is functional when absorbed to plastic in the quantitative cell adhesion assay. Microtitre wells were coated with sICAM-1 to a density ofabout 1,600 sites/µm2. 2x105 A20 cells per well were then spun downonto the surface of the wells, incubated at 37°C for 5 min., and unboundcells were washed off. Bound cells were then quantitated with MTT.Cytochalasin B was used at a final concentration of 20 p.M and sICAM-1was used at a final concentration of 100 µg/ml to attempt inhibition.In agreement with the inability of sICAM-1 to inhibit cell adhesion in the quantitativecell adhesion assay, sICAM-1 is unable to inhibit the LFA-1:ICAM-1 mediated homotypicaggregation of MBL-2 cells. The results of such an experiment are shown in Figure 8. In thisassay the murine T lymphoma cell line MBL-2 aggregates only after stimulation by PMA.Aggregation is mediated by LFA-1:ICAM-1 interaction as demonstrated by the ability of Ab toeither LFA-1 or ICAM-1 to inhibit aggregation. Interestingly, the inhibition by Ab to LFA-1was consistently more complete than that by Ab to ICAM-1, suggesting the involvement ofanother counter-receptor for LFA-1. Aggregation in this assay is also effectively inhibited byEDTA and by cytochalasin B. However, aggregation is not inhibited to any extent by sICAM-1 at a concentration of up to 100 µg/ml.As mentioned above, the addition of 1 mM Mn++ presumably increases the affinity ofLFA-1 for ICAM-1. When I first used Mn++ in the 125I-sICAM-1 binding assay, I noticedthat cells suspended in medium containing Mn++ aggregated extensively compared to cellssuspended in medium containing Ca++ and Mg++ but no Mn++. This phenomenon appeared tobe ICAM-1:LFA-1 dependant since ICAM-1 positive, LFA-1 positive MBL-2 and A20 cellsaggregated strongly in response to Mn++, while LFA-1 positive, ICAM-1 negative T28 cellsaggregated to a much lesser extent in response to Mn++. In order to characterize thisphenomenon, a qualitative aggregation assay similar to the PMA induced aggregation assaywas developed. Figure 9 shows the results of such an experiment. In this Figure, MBL-2cells are shown to aggregate strongly in response to 1 mM Mn++, and this aggregation is LFA-1 and ICAM-1 dependant. However, unlike PMA induced aggregation, Mn++ inducedaggregation could be efficiently inhibited with sICAM-1 at 1001.1g/ml. Analogous results wereobtained with A20 cells (data not shown).46Figure 8 SICAM-1 does not inhibit the PMA induced homotypic aggregation ofMBL-2 cells, Photomicrographs of MBL-2 cells a) without PMA, b) withPMA, c) with PMA and anti-LFA-1, d) with PMA and anti-ICAM-1, e)with PMA and cytochalasin B, 0 with PMA and EDTA, g) with PMA andDMSO as a control for f), and h) with PMA and sICAM-1. PMA was usedat a final concentration of 50 ng/ml, Abs were used at a 1/40 dilution ofculture SN (predetermined to be saturating), cytochalasin B was used at afinal concentration of 20 p,M, EDTA was used at a final concentration of 5mM, and sICAM-1 was used at a final concentration of 100 µg/m1 toattempt inhibition. The assay was photographed following 4 hrs. ofincubation at 37°C.47Figure 9 SICAM-1 inhibits the Mn++ induced homotypic aggregation ofMBL-2 cells. Photomicrographs of MBL-2 cells a) without Mn++, b) withMn++, c) with Mn++ and anti-LFA-1, d) with Mn++ and anti-ICAM-1, e)with Mn++ and anti-Mac-1, f) with Mn++ and 100 µg/ml sICAM-1, g) withMn++ and 33 µg/m1 sICAM-1, and h) with Mn -}-} and 11 lig/nil sICAM-1.All conditions contained 1 mM Ca++ and Mg++. Mn++ was used at a finalconcentration of 1 mM, Abs were used at a 1/40 dilution of culture SN(predetermined to be saturating), and sICAM-1 was used at the finalconcentrations indicated to attempt inhibition. The assay was photographedfollowing 45 min. of incubation at 37°C.48Early on in the project we hypothesized that since the density of sICAM-1 on thesICAM-1 coated plastic was so high in the quantitative cell adhesion assay, inhibition bysICAM-1 in solution might be difficult , especially since the cells are spun down onto thesurface of the well and then allowed to bind. We predicted that sICAM-1 might be better ableto inhibit adhesion in the quantitative adhesion assay if the wells were coated to a lower densityand if cells were incubated with the sICAM-1 before being added to the plate. Figure 10shows the results of such an experiment. The results clearly show that adhesion is inhibitableby cytochalasin B and anti-LFA-1, but not by sICAM-1 at a final concentration of 100 tg/ml,even when the cells were preincubated with the sICAM-1 for 20 min. and then assayed foradhesion in wells coated to only 250 sites/µm 2. This experiment also shows that for A20 cellsPMA stimulation increases adhesion, but only at lower densities of absorbed sICAM-1. Myobservation that cell lines bind efficiently to high densities of absorbed sICAM-1 without PMAstimulation is in contrast to the findings of Springer's group with human ICAM-1 (13, 42).These reports suggest that even at high densities of ICAM-1, lymphoblastoid cells requirePMA stimulation in order to adhere. In fact I have observed that the cell line MBL-2 bindsefficiently to absorbed ICAM-1 over a range of densities, and binding could not be specificallyincreased by PMA stimulation. In addition, cytochalasin B does not efficiently inhibit theadhesion of MBL-2 cells in the quantitative adhesion assay, despite the ability of cytochalasinB to effectively inhibit the PMA induced homotypic aggregation of MBL-2 cells. In contrastthe cell line A20, which was used in Figures 7 and 10, is effectively inhibited by cytochalasinB from adhering to plastic absorbed sICAM-1.49NB no inh.ooN^cyto. Bco cs PMA(7)TIB213cs4^no inh. AEIE0 z^cyto. Bin ZsiN CD^PMA0 sICAM-1 .11111=1111111no inh. jouPMA -sICAM-1co050100000^200000cells bound/wellFigure 10 Cell adhesion in the quantitative cell adhesion assay is stimulated by PMA and inhibited by anti-LFA-1 and cytochalasin B but not by sICAM-1. Experiment performed as for Figure 7 except that wells were coated tovarying sICAM-1 densities, and cells were pretreated with inhibitors/stimulators for 20 min. at 37°C before being added to the plate. PMA wasused at a final concentration of 50 ng/ml. TIB 213 is anti-LFA-1, and wasused at a 1/40 final dilution of culture SN.There were a number of problems associated with using cell lines in the quantitative celladhesion assay. Firstly, all of the cell lines tested often gave high backgrounds, that is theybound to blocked plastic or even uncoated plastic appreciably. Secondly, as noted above therewas disagreement as to the requirement for PMA stimulation in cell adhesion. The work ofSpringer et al. has suggested that stimulation by PMA or TcR cross-linking is an absoluterequirement for the adhesion of lymphoid cells to absorbed ICAM-1 (42). The findings of thisstudy with murine cell lines suggested that cells can bind efficiently to high densities ofabsorbed sICAM-1, and only in some cases at lower densities of sICAM-1 is PMA stimulationrequired. Therefore, we wanted to obtain a population of cells that would consistently bind toplastic absorbed sICAM-1 with a low background, but only after stimulation by PMA. Restingsplenic T cells, purified over a nylon wool column, fulfilled these criteria. Figure 11 showsthe results obtained when nylon wool T cells were used in the quantitative cell adhesion assay.It is clear that these cells bind to immobilized sICAM-1 in a sICAM-1 density dependent andPMA dependent fashion. Adhesion is also effectively inhibited by Ab to LFA-1, andcytochalasin B.Resting splenic T cells were also used to quantitate the Mn++ induced adhesiondescribed above. Since Mn++ treatment precluded the use of the MTT assay for thequantitation of cell adhesion (see Materials and Methods), 51 Cr labeled nylon wool T cells wererequired for these experiments. Mn++ clearly induces LFA-1:ICAM-1 mediated adhesion inthe quantitative cell adhesion assay, as shown in Figure 12. However, sICAM-1 in solution ata concentration of 65 lag/m1 was incapable of inhibiting Mn++ induced adhesion, in contrast tothe ability of sICAM-1 to inhibit the Mn++ induced aggregation of cell lines. Subsequentexperiments with sICAM-1 concentrations of up to 110 pg/m1 failed to clearly demonstrateinhibition of Mn++ induced adhesion by sICAM-1 in the quantitative cell adhesion assay (datanot shown). Surprisingly, cytochalasin B did inhibit the Mn++ induced adhesion of resting Tcells to absorbed sICAM- , as shown in Figure 12.51^E^no inh.0^0 7,-)^PMAco cp PMA/TIB213PMA/cyto. B "MNno inh.PMAPMA/TIB 213 OM=PMA/cyto. B 1/111no inh. AMINPMAPMA/TIB213PMA/cyto. B-0a)^ )-no inh. )11/11TD^PMA^°^0 100000^200000cells bound/wellFigure 11^Cell adhesion of resting T cells in the quantitative cell adhesion assay isabsolutely dependent on PMA stimulation. Resting splenic T cells werepurified over a nylon wool column and then used in the quantitative celladhesion assay. The experiment was performed as for Figure 10.52t4Eno inh.o^PMA—^PMA/cyto. BoMn++o^Mn++/sICAM-1Mn++/cyto. BMn++/TIB 2130^no inh.PMAO^Mn++0^2000^4000^6000bound CPM/wellFigure 12 Mn++ induces ICAM-1:LFA-1 mediated adhesion in the quantitative celladhesion assay. Resting splenic T cells were purified over a nylon woolcolumn, labeled with 51 Cr, and then used in the quantitative cell adhesionassay. The experiment was performed as for Figure 10 with severalexceptions. Mn++ was used at 1.25 mM, and all wells containing Mn++were washed with Mn++ containing medium SICAM-1 was used at a fmalconcentration of 65 µg/ml in solution to attempt inhibition. Bound cellswere quantitated by y counting following dissociation with 5 mM EDTA.53In an attempt to get a picture of how cytochalasin B was inhibiting adhesion soefficiently, the cellular morphology of cells treated with cytochalasin B was compared to that ofuntreated cells. To do this, untreated or cytochalasin B treated MBL-2 cells were used in thequantitative cell adhesion assay, except that their cellular morphology was observed throughoutthe experiment. As shown in Figure 8, cytochalasin B effectively inhibits the PMA inducedaggregation of MBL-2 cells, but as noted above cytochalasin B does not effectively inhibit thebinding of these cells to plastic absorbed sICAM-1. The results of this experiment are shownin Figure 13. In this experiment cells were photographed following the 5 min. incubation at370C before and after the wells were washed of unbound cells. Cytochalasin B clearly inducesa morphological change in the cells observable before unbound cells were washed away.Cytochalasin B treated cells appear more uniformly round than do untreated cells, and themembranes of treated cells were smoother than untreated cells. It was hoped that moreeffective flattening of untreated cells on the plastic absorbed sICAM-1 as compared to thecytochalasin B treated cells might be observed in this experiment; however, the magnificationand resolution of the microscope seemed to be inadequate for this purpose. Following thewashing away of unbound cells, the cells treated with cytochalasin B return substantially totheir characteristic morphology.In summary of the functional characterization of sICAM-1, the purified recombinantsICAM-1 binds to LFA-1+ cells and functions just as cellular ICAM-1 when immobilized onsolid surfaces. PMA stimulation increases ICAM-1:LFA-1 mediated cell adhesion; however,sICAM-1 does not inhibit PMA stimulated, ICAM-1:LFA-1 mediated cell adhesion. Mn++enhances the binding of 125I-sICAM-1 to LFA-1+ cells and increases ICAM-1:LFA-1 mediatedcell adhesion. In contrast to PMA induced homotypic aggregation, Mn++ induced homotypicaggregation is inhibited by sICAM-1. Also of significance is the importance of the actincytoskeleton for ICAM-1:LFA-1 mediated adhesion.54:Q3; )` .1,?,6 .0 .(4 00,,,?.°4.0before wash^after wash55Figure 13 Cytochalasin B induces morphological changes in cells.  MBL-2 cells whichwere either a) untreated or b) treated with cytochalasin B (20 [tM) for 20min. at 37°C, were then centrifuged onto sICAM-1 coated microtitre wells(coated to 1600 sites/µm2), incubated for 5 min. at 37°C, and thenphotographed. Unbound cells were then washed away and the wells werephotographed again.Part III: The role of multivalent interaction in LFA-1:ICAM-1 mediated celladhesionThe results presented above demonstrate that sICAM-1 is ineffective in inhibiting PMAinduced cell adhesion. It seems that the most reasonable explanation for this observation is thatthe binding of sICAM-1 to LFA-1 is not of sufficiently high affinity. Cell:cell adhesionmediated by LFA-1 and ICAM-1 is likely mediated by multivalent interaction. Low affinitymonovalent binding is unlikely to effectively compete with multivalent binding betweenmembrane LFA-1 and either membrane ICAM-1 or plastic immobilized sICAM-1. If theinability of monovalent sICAM-1 to inhibit PMA stimulated cell adhesion is solely due to it'smonovalency then an artificially multivalent form of sICAM-1, with the same affinity for it'sreceptor, would be expected to effectively inhibit cell adhesion. The approach we took to testthis hypothesis was to create a multivalent form of sICAM-1 by coating it to uniformpolystyrene beads. Since sICAM-1 coated to polystyrene microtitre plates is functional asdetermined by it's adhesive capabilities, we expected that sICAM-1 coated to polystyrene beadswould also be functional.Polystyrene beads, 2 gm in diameter, were coated with purified sICAM-1 in a similarfashion to the coating of the polystyrene microtitre wells. These beads were then assayedusing an ELISA in order to quantitate the amount of sICAM-1 coated to the beads. The resultsof this assay are shown in Figure 14. In Figure 14a), sICAM-1 coated beads, control BSAcoated beads, or known amounts of purified sICAM-1 were used to inhibit an ELISA in whichsICAM-1 coated to microtitre wells was detected with YN1/1.7 Ab. The results clearly showthat the sICAM-1 coated beads, but not the BSA coated beads, inhibit the binding of YN1/1.7Ab to sICAM-1 coated to the ELISA plate. Using the calculated % inhibition values for theknown amounts of purified sICAM-1, a standard curve was generated (Figure 14b). Thisstandard curve was then used to estimate the amount of sICAM-1 in the preparation of sICAM-1 coated beads. It was estimated that the final suspension of sICAM-1 coated beads containedapproximately 60 tg/m1 of sICAM-1, and since the beads were coated at a concentration of 1005614/ml, an approximate coating efficiency of 60% was achieved. When this concentration wasused, along with the approximate concentration of beads in the preparation, to calculate thedensity of sICAM-1 on the beads, a density of approximately 6,000 sites/µm2 was obtained.This density is extremely high when compared to both the density obtained on the polystyrenemicrotitre plates and the density of LFA-1 on human neutrophils mentioned above. It is likelythat the high surface area to volume ratio of the polystyrene beads explains their ability toabsorb protein so efficiently.These sICAM-1 coated beads bound to MBL-2 cells as shown in Figure 15. ThesICAM-1 coated beads also specifically inhibited LFA-1:ICAM-1 mediated cell adhesion in thequantitative cell adhesion assay, as shown in Figure 16. The sICAM-1 coated beads wereshown in this experiment to specifically inhibit cell adhesion under conditions wheremonovalent sICAM-1 was completely incapable of inhibiting cell adhesion. It is important tonote that in Figure 16 the amount of inhibitor is given in terms of 41 per 100111 final assayvolume/well; therefore, since the sICAM-1 coated beads only contain about 1/20 the sICAM-1per unit volume as the monovalent sICAM-1, the 10 IA designations correspond to sICAM-1concentrations of 100 pg/m1 for the monovalent sICAM-1 but only 5 µg/ml for the sICAM-1coated beads. In addition, 10 ill of coated beads corresponds to approximately 150 beads percell in the assay. The inhibition of adhesion in the quantitative cell adhesion assay by blockingLFA-1 with sICAM-1 coated beads was never complete, and was always plagued by anonspecific background inhibition problem, especially at high ratios of beads:cells. It wouldseem that at high densities of beads, when the cells are spun down onto the surface of the plate,so are some of the beads, and the beads simply get between the cells and the plate. However,the inhibition shown in Figure 16 is specific, and even marginal inhibition in this assay isprobably significant in light of the inability of monovalent sICAM-1 to inhibit the Mn++induced adhesion in the quantitative cell adhesion assay (Figure 12). The experiment shown inFigure 16 therefore supports the notion that the reason for the inability of sICAM-1 to inhibitcell adhesion is simply that monovalent sICAM-1 is incapable of competing with the57multivalent sICAM-1 on the coated microtitre wells; since both forms of the molecule have thesame low affinity for their receptor, only multivalent interactions are stable enough to persist.The same argument holds for the inability of sICAM-1 to inhibit the PMA induced homotypicaggregation of MBL-2 cells; monovalent sICAM-1 is incapable of competing with themultivalent ICAM-1 on other cells.In summary of the results presented in Part III, through the use of a multivalent form ofsICAM-1 the importance of multivalent interaction in LFA-1:ICAM-1 mediated cell adhesionwas demonstrated. The results also support the notion that PMA stimulated, ICAM-1:LFA-1mediated cell adhesion is the result of low affinity binding.58ano inh. ..--5u9 OEM.`4:;^a0 2 ug7▪^n 0.5 ug=  <'e 1:' 20u1,O _5 10 ul13in< -o 20u1coa) 2 10 Ul 1^1^r^r^1^10^20^40^60^80^100^120relative signal (%)b 8060 -40 -20 -0 I^r^I^I^.^i1.0^2.0^3.0^4.0^5.0ug sICAM-10 0Figure 14 Ouantitation of sICAM-1 coated to polystyrene beads.  An ELISA assay, inwhich sICAM-1 coated to microtitre wells was detected with YN1/1.7 Aband indirect immunoperoxidase development, was used to quantitate theamount of sICAM-1 coated to the polystyrene beads. In (a) the binding ofYN1/1.7 Ab was inhibited with known quantities of purified sICAM-1,sICAM-1 coated polystyrene beads, or BSA coated beads. The volumes ofbeads indicated in (a) represents volumes of the final suspension of coatedbeads per 100 ill in the ELISA assay. The data from the inhibition of theELISA with known quantities of sICAM-1 was then used to generate astandard curve (b) from which the amount of sICAM-1 in the sICAM-1coated beads could be estimated. From the standard curve it was estimatedthat the final suspension of sICAM-1 coated beads contained approximately60 j.ig sICAM-1/ml.59Figure 15 SICAM-1 coated polystyrene beads bind to MBL-2 cells. Photomicrographs of MBL-2 cells which were immobilized on plastic viapoly-L-lysine, incubated for 45 min. at 37°C with polystyrene beadspreviously coated with a) sICAM-1 or b) BSA, and washed of unboundbeads. Microtitre wells were incubated with 10 Ill of bead suspensions infinal volumes of 100 ill, corresponding to approximately 6 µg/m1 sICAM-1and 3x108 beads/ml.6061no inh.'7 10 ul5 ulo^2.5 ul(-7^1.25 ul0o5u10- S 2.5 ul03^1.25 ulcnCO 210 ul5 ul2.5 ul1.25 ulBSA blocked/no inh.0 1 00000cells bound/well200000Figure 16 Multivalent sICAM-1 coated polystyrene beads inhibit cell adhesion in the quantitative cell adhesion assay while monovalent sICAM-1 does not.  PMAtreated resting T cells were incubated for 20 min. at 3700 with eithersICAM-1, sICAM-1 coated beads, BSA coated beads, or nothing. Thecells were then used in the quantitative cell adhesion assay in wells coated toapproximately 940 sites/µm2 . Experiment was performed as described forFigure 10. For monovalent sICAM-1, 10 corresponds to a finalconcentration of 100 sICAM-1/ml, whereas for the sICAM-1 coatedbeads 10 IA corresponds to only 5 pg sICAM-1/ml and approximately 150beads/cell.Part IV: Analysis of the distribution of LFA-1 and ICAM-1 on the cell surfaceThe functional characterization of sICAM-1 presented above provides somewhatconfusing information. SICAM-1 appears functional as measured by the 125I-sICAM-1binding assay, as well as in the quantitative cell adhesion assay. In addition, sICAM-1 iscapable of inhibiting the Mn++ induced aggregation of cell lines, aggregation which seems tobe facilitated by an increase in the normally low affinity of the interaction between ICAM-1 andLFA-1. However, sICAM-1 is very inefficient at inhibiting LFA-1:ICAM-1 mediated celladhesion mediated by PMA stimulation. In addition, the actin cytoskeleton is essential forLFA-1:ICAM-1 mediated cell adhesion, especially adhesion stimulated by PMA.The results with sICAM-1 coated beads explain why the monovalent form of sICAM-1is incapable of inhibiting ICAM-1:LFA-1 mediated adhesion. These results also stress theimportance of multivalent interaction in ICAM-1:LFA-1 mediated cell adhesion. In light of theresults with sICAM-1 coated beads, we proposed that since ICAM-1 :LFA-1 mediated adhesionwas dependant on multivalent interaction, local concentrations of adhesion molecules on thecell surface might be important in determining the adhesive phenotype of LFA-1 and ICAM-1positive cells. We therefore made several predictions which I set out to test. The first was thathigh localized concentrations of adhesion molecules on the surface of cells might facilitateadhesion, at the same time as making inhibition by sICAM-1 in solution difficult. We alsopredicted that PMA stimulation may affect the distribution of adhesion molecules, especiallyLFA-1, on the cell surface, as is the case for Mac-1 (51). While these predictions may seemsomewhat 'hand-wavey', as discussed in the Introduction of this thesis, there is a greatpotential for the distribution of adhesion molecules to have significant effects on cell adhesion.In addition, relatively little is known about the distribution of adhesion molecules on the cellsurface, and what effects this distribution may have on cell adhesion. Therefore thedistribution of LFA-1 and ICAM-1 on the cell surface was studied to determine whatrelationship, if any there is between the distribution of adhesion molecules and cell adhesion.62The first distribution studies were done with MBL-2 cells. Since this cell lineaggregates in an LFA-1:ICAM-1 dependant fashion in response to PMA stimulation (Figure 8),we hoped to see some differences in the distribution of adhesion molecules, particularly LFA-1, upon PMA stimulation. The results of one such experiment are given in Figure 17. In orderto visualize the distribution of cell surface molecules we used immunofluorescence microscopybecause of the relative ease of this technique. As shown in Figure 17, the distribution of LFA-1 and ICAM-1 is relatively uniform over the cell surface, not significantly different from thecontrol molecule CD45, which incidentally was chosen as a control because of it's high level ofexpression and it's apparent lack of association with the cytoskeleton (108), suggestinguniform distribution. In addition, Figure 17 demonstrates the lack of any change in thedistribution of either LFA-1 or ICAM-1 following PMA stimulation, a consistent observationover several independent experiments. The way the MBL-2 cells were treated before fixationand staining was to culture them under the same conditions as for the PMA inducedaggregation assay, except in 1 ml cultures with or without PMA, for 4 hours. Aggregated,PMA stimulated cultures as well as non-aggregated, unstimulated cultures were then chilled onice for about 10 min. in order to 'freeze' the normally fluid plasma membrane, disaggregatedby pipetting, and then fixed with paraformaldehyde on ice before staining.63Figure 17 The distribution of cell surface molecules on unstimulated or PMA stimulated MBL-2 cells. Photomicrographs of MBL-2 cells which wereeither unstimulated or stimulated with PMA at 37°C for 4 hrs., chilled onice, disaggregated by pipetting, fixed with 1% paraformaldehyde, stainedwith rat IgG antibodies to various cell surface molecules, and then stainedwith FITC goat-ant-rat IgG. Cells were stained and subsequently observedunder the microscope in suspension. For each exposure taken underfluorescence (F1.), an accompanying exposure was taken under phasecontrast (P.C.).64652nd fib controlFl_P-c.66LFA-1F1_P.c.67LFA-1Fl.P-c_ICAM-168F1_P-C.IC AM- 169F1_P . C . 70CD45 FLP_C_71CD45F1.P.C.In addition to MBL-2, initial distribution studies were done with A20 cells. Typicalresults of the analysis of A20 cells is shown in Figure 18. A20 behave quite differently thanMBL-2 cells; A20 cells spontaneously form homotypic aggregates in culture while MBL-2 cellsgrow mostly as single cells with few aggregates. A20 cells also do not aggregate well inresponse to PMA. A20 cells were therefore not analyzed after PMA stimulation. Since studiesto functionally characterize sICAM-1 demonstrated the importance of the actin cytoskeleton, wedecided to fluorescently visualize cytoplasmic actin concurrently with cell surface staining. Wehoped to see some association between the actin cytoskeleton and adhesion molecules,especially LFA-1. Figure 18 shows results from A20 cells which were either untreated ortreated with cytochalasin B, and then stained for actin or various cell surface molecules. Therewere two very striking observations with this experiment. The first was that ICAM-1 was veryoften found localized to single spots on the surface of A20 cells. This observation wasconsistent over several experiments with A20 cells, and it is not likely the result of Ab inducedcapping because the cells were fixed prior to staining and this punctate staining pattern was notseen with any other primary Ab used, or with any other cell line analyzed. The other strikingobservation shown in Figure 18 is the very different actin staining patterns in untreated andcytochalasin B treated cells. Untreated cells exhibited heterogeneous actin staining, betweencells and within individual cells. The majority of untreated cells stained relatively weakly foractin, while a minority of cells stained very intensely often with a polarized staining pattern. Insharp contrast, the cytochalasin B treated cells exhibited much less heterogeneous actin stainingbetween cells but consistently patchy staining within individual cells. The actin staining ofMBL-2 cells was very similar to that shown for A20 cells in Figure 18 (data not shown).With respect to the punctate ICAM-1 staining of A20 cells shown in Figure 18, acorrelation with actin staining is unclear. In (cytochalasin B) untreated cells the ICAM-1 spotsappeared often but not always to correlate with areas of the cell staining intensely for actin. Incytochalasin B treated cells the ICAM-1 spots most often correlated to regions stainingintensely for actin, although cytochalasin B treated cells showed far more consistently intense72and punctate actin staining than untreated cells. The LFA-1 staining of A20 cells shown inFigure 18 demonstrated an even distribution, not unlike the distribution of LFA-1 on MBL-2cells. No dramatic association between LFA-1 and actin was observed.For Figure 18 cytochalasin B was added to aggregated A20 cells in culture, the cellswere incubated for 30 min. at 37°C, chilled on ice, fixed, and then stained. Over the 30 min.incubation with cytochalasin B the cells did not disaggregate, and as shown in Figure 18 thecytochalasin B treatment did not cause the punctate ICAM-1 spots to dissociate. Anotherexperiment was done where aggregated cells were disaggregated by pipetting, treated withcytochalasin B for 30 min. at 37°C, and then analyzed by immunofluorescence microscopy. Inthis experiment the ICAM-1 spots persisted even after disaggregation of the cells and treatmentwith cytochalasin B (data not shown).To summarize the initial studies on LFA-1 and ICAM-1 distribution, while differentialdistribution of ICAM-1 could clearly be shown on different cell lines, the effect of PKCstimulation on the distribution of LFA-1 could not be shown, at least through the use offluorescence microscopy.73Figure 18 The distribution of cell surface molecules and intracellular actin on/in cytochalasin B treated or untreated A20 cells. Photomicrographs of A20cells which were either untreated or treated with cytochalasin B at 37°C for20 min., and then handled as described for Figure 17 except that the cellswere stained with a combination of FITC-goat-ant-rat IgG and TRITCphalloidin. For each exposure taken under fluorescence with the green filterto visualize FITC (Gr. Fl.) or with the red filter to visualize TRITC (Rd.Fl.), an accompanying exposure was taken under phase contrast (P.C.).7475actinFl.p_c_actin76 Fl.P.0 .77LFA-1Fl_P_C _IC AM-1 I actin78aL45:ICAM-1 I actin79Part V: The role of the cytoplasmic domain of ICAM-1 in cell adhesionAs discussed briefly in the Introduction of this thesis, the distribution of adhesionmolecules on the surface of the cell can potentially play an important role in the strength ofcell:cell adhesion. However, the role of adhesion molecule distribution in the regulation ofadhesion is, at present, poorly characterized. Part IV of the Results section of this thesisaddresses the topic of the distribution of adhesion molecules, but presents only partiallysatisfying results. What effect, if any, does PKC stimulation have on the distribution ofadhesion molecules, especially LFA-1? What are the molecular mechanisms controlling thedistribution of adhesion molecules? These are two of the questions still to be answered. Withrespect to ICAM-1, it is clear from my own results with A20 cells and the results of others thatICAM-1 can be highly localized on the cell surface. Staunton et al. have observed that ICAM-1expressed on transfected COS cells exhibits a punctate staining pattern, even when the entirecytoplasmic domain is deleted (56). It would seem logical that the cytoplasmic domain beinvolved in the regulation of ICAM-1 cell surface distribution via cytoplasmic links. Wetherefore wanted to repeat the studies of Staunton et al. in our own system, and attempt todetermine what role, if any, the cytoplasmic domain of ICAM-1 has on both the distribution ofICAM-1 and on the function of ICAM-1.The approach we took was to transfect ICAM-1 with or without it's cytoplasmicdomain into ICAM-1 - cells and compare both the function and the distribution of both versionsof the molecule. The constructs were made and the transfections were done by a summerstudent in our Lab, Mike Buchart. The entire cDNA encoding murine ICAM-1 or the cDNAdeleted of the majority of the coding sequence for the cytoplasmic domain were subcloned intoBCMGSNeo. We chose two different cell lines to transfect, a lymphoid cell line in the hope ofcomparing both the function and distribution of the two molecules, and a fibroblastoid cell lineas a back-up for distribution studies.The lymphoid cell line chosen was the T cell hybridoma T28 (85). Each construct waselectroporated into T28 cells, transfectants selected in 1 mg/ml G418, and expression checked80by FACS. In culture, cells with the two different versions of ICAM-1 behave muchdifferently. This difference is shown in Figure 19. Cells with the entire ICAM-1 moleculeform large, tight homotypic aggregates while cells with ICAM-1 deleted of it's cytoplasmicdomain form fewer, loose homotypic aggregates. This difference is not likely due todifferences in levels of expression, since FACS analysis showed very similar levels ofexpression, and even populations of cells expressing clearly lower levels of the entire moleculeas compared with the cytoplasmic domain deletion exhibited the same behavior in culture asshown in Figure 19 (data not shown). Both versions of ICAM-1 conferred the ability to formtight homotypic aggregates upon PMA stimulation, as shown in Figure 20. In addition, Figure20 also demonstrates the importance of the cytoplasmic domain in mediating homotypicaggregation in the absence of stimulation; when resuspended and incubated at a relatively highcell concentration, cells with the entire ICAM-1 molecule aggregate more efficiently than docells with the molecule deleted of it's cytoplasmic tail. The T28 cells with the entire ICAM-1molecule did aggregate to a greater degree upon PMA stimulation as compared to the cells withICAM-1 deleted of it's cytoplasmic domain. Therefore, the effect of ICAM-1's cytoplasmicdomain and the LFA-1 avidity affect of PMA appear to be additive.818 2Figure 19 T28 cells transfected with ICAM-1 with or without it's cytoplasmic domainbehave differently in culture. Photomicrographs of flask cultures of T28cells transfected with the BCMGSNeo vector carrying a) the completeICAM-1 cDNA, or b) the cDNA encoding ICAM-1 without it's cytoplasmicdomain.83Figure 20^T28 cells transfected with ICAM-1 with or without its cytoplasmic domain aggregate in response to PMA. Photomicrographs of unstimulated or PMAstimulated a) parental T28 cells, b) T28 cells transfected with the entireICAM-1, and c) T28 cells transfected with ICAM-1 deleted of it'scytoplasmic domain. Cultures were photographed after 1 hr. at 37°C.PMA was used at a final concentration of 50 ng/ml.The cell surface distribution of both versions of ICAM-1, LFA-1, and the control cellsurface molecule CD45 on the transfected T28 cells is shown in Figure 21. Figure 21 in factshows the very cultures, which were analyzed by immunofluorescence analysis after fixationand antibody staining, as are shown in figure 20. Figure 20 shows 1 ml cultures of T28 cellswhich were photographed after a 1 hr. incubation with or without PMA. These cultures werethen chilled on ice, fixed on ice, and stained with antibodies to various cell surface molecules.Figure 21 shows the relatively high expression level of both of the transfected versions ofICAM-1. Both versions of transfected ICAM-1, with and without the cytoplasmic domain,exhibited somewhat punctate distributions, although clearly not as punctate as the distributionof ICAM-1 on A20 cells. The distribution of ICAM-1 was similar with or without PMAstimulation. The one striking observation of this experiment was the difference in thedistribution of LFA-1 with or without PMA stimulation. Figure 21 shows that on PMAstimulated, aggregated T28 cells transfected with complete ICAM-1, LFA-1 exhibited a morepatchy distribution than the unstimulated, less tightly aggregated cells. This observation wasalso made on the PMA stimulated, aggregated T28 cells transfected with ICAM-1 deleted of it'scytoplasmic domain (data not shown). The LFA-1 staining was, however, more punctate onthe cells with the entire ICAM-1 molecule; therefore, the extent of punctate LFA-1 staining onthe aggregated, ICAM-1 transfected T28 cells seems to correlate with the extent of aggregation.8485Figure 21^The distribution of cell surface molecules on unstimulated or PMAstimulated T28 cells expressing transfected forms of ICAM-1. Photomicrographs of T28 cells expressing either complete ICAM-1 (ICAM-1 transfected) or ICAM-1 without it's cytoplasmic tail (ICAM-1 del.cyto.transfected) which were unstimulated or stimulated with PMA for 1 hr. at37°C, and then handled as described for Figure 17. For each exposuretaken under fluorescence (Fl.), an accompanying exposure was taken underphase contrast (P.C.).IC AM-1 transfected I ICAM-1 stained86F1_P_C_IC AM-1 transfected I ICAM-1 stained87F1. P.C.88ICAM-1 deLcyto. transfected I ICAM-1 stainedFI.P-c_ICAM-1 del_cyto. transfected I ICAM-1 stained89Fl_ P_C .ICAM-1 transfected I LFA-1 stained90 F1.P91IC AM-1 transfected LFA-1 stainedFLICAM-1 deLcyto_ transfected CD45 stained92Fl_Figure 22 shows the distribution analysis of the untransfected T28 cells, and serves asa control for Figure 21. The cells shown in Figure 22 were treated exactly the same way asthose shown in Figure 21, except of course that as shown in Figure 20 the untransfected T28cells do not aggregate appreciably in response to PMA. As shown in Figure 22 theuntransfected T28 cells are essentially ICAM-1 negative, in agreement with FACS analysis,except for occasional slight staining. The distribution of LFA-1 on unstimulated or PMAstimulated parental T28 cells was indistinguishable, in contrast to the more punctate LFA-1distribution on PMA stimulated, aggregated ICAM-1 transfected T28 cells. Taken togetherthese observations strongly suggests that the change in LFA-1 distribution on the ICAM-1transfectants occurred as a result of aggregation, rather than before aggregation; therefore, itcannot be concluded that PMA stimulation results in the alteration of LFA-1 distribution on thecell surface such that increased localized concentrations of LFA-1 promote cell:cell adhesion.The increased patchyness of LFA-1 following PMA induced aggregation of the ICAM-1transfected T28 cells is possibly a result of the high level of ICAM-1 expression on the ICAM-1 transfected T28 cells, especially since a change in LFA-1 distribution was never observed onMBL-2 cells following PMA induced aggregation. PMA stimulation likely affects the ICAM-1transfectants by increasing the avidity of LFA-1 for ICAM-1, which facilitates ICAM-1:LFA-1mediated aggregation. I speculate that the punctate LFA-1 distribution on the PMA stimulated,aggregated transfectants was due to mutual capping with ICAM-1 on opposing cells. Becausethere is so much ICAM-1 expressed on the surface of the transfected cells, a form of mutualcapping likely occurs where since the density of one of the two counter receptors is in suchexcess the capping of only one counter receptor, the one expressed at a lower density, isobserved.939 4Figure 22^The distribution of cell surface molecules on unstimulated or PMA stimulated. untransfected T28 cells. Photomicrographs of parental,untransfected T28 cells which were handled as described as for Figure 21.For each exposure taken under fluorescence (F1.), an accompanyingexposure was taken under phase contrast (P.C.).Fl.95IC AM- 1LFA-196Fl_LFA-197 Fl.1)-c_As described above, the distribution of both forms of transfected ICAM-1, with andwithout it's cytoplasmic domain, appeared similar on transfected T28 cells. The distributioncould at best be described as somewhat patchy. However, the two forms of ICAM-1 did havedifferent functional characteristics; the complete ICAM-1 molecule mediated cell adhesion moreefficiently in unstimulated cells. Therefore, it still seemed likely that the cytoplasmic domainmay mediating subtle distribution effects not easily observed on lymphoid cells. We wanted toexpress these same ICAM-1 molecules on flat, adherent cells in order to more carefully analyzeany distribution effects mediated by the cytoplasmic domain. The same two ICAM-1constructs, with and without the cytoplasmic domain, each in BCMGSNeo, were transfectedby calcium phosphate mediated transfection into the murine fibroblastoid cell line L. Thedistribution of both versions of ICAM-1 as well as other control cell surface molecules on thetransfected L cells is shown in Figure 23. Both versions of ICAM-1 were expressed over theentire surface of the cells, and both versions exhibited a somewhat punctate distribution patternon the cell surface, as is demonstrated especially well on the full length ICAM-1 transfected,ICAM-1 stained photomicrograph. The actual microscopic observation of the slides did notsuggest, however, any significant difference in the distributions of the two versions of ICAM-1. As described in the Materials and Methods section of this thesis, initial expression of bothversions of transfected ICAM-1 on FACSorted L cells was very strong, and the transfectantshad to be grown in low concentrations of G418 in order to lower the expression to a levelwhere subtle observations concerning distribution could be made. The cells shown in Figure23 were grown in only 100 µg/ml G418 for approximately 2 weeks before analysis, and as aresult there was some heterogeneity in the levels of ICAM-1 expression. In general, the mostpunctate distributions were observed on cells expressing lower levels of transfected ICAM-1.9899Figure 23^The distribution of cell surface molecules on L cells expressing transfectedforms of ICAM-1. Photomicrographs of L cells expressing either completeICAM-1 (ICAM-1 transfected) or ICAM-1 without it's cytoplasmic tail(ICAM-1 del.cyto. transfected) which were fixed with 1%paraformaldehyde, stained with rat IgG antibodies to various cell surfacemolecules, and then stained with FITC-goat-ant-rat IgG. L cells weregrown on glass coverslips prior to immunofluorescence analysis; cells werefixed, stained, and observed while attached to the coverslips. For eachexposure taken under fluorescence (Fl.), an accompanying exposure wastaken under phase contrast (P.C.).IC AM-1 del_cyto. transfected J 2nd Ab control100Fl.P.C.101ICAM-1 transfected I CD4 stainedFl.P.C_IC AM-1 transfected I ICAM-1 stained102F1.P_C_103ICAM-1 transfected I CD44 stainedF1. P.C_104IC AM-1 deLcro. transfected f IC AM-1 stainedF1_P.C.ICAM-1 deLcyto. transfected 1 CD44 stained105FLP_C_ CD44 was chosen as a control for the transfected fibroblasts because it has beenproposed that CD44 and ICAM-1 may associate in the plane of the membrane (G.J.Dougherty, manuscript in preparation). Indeed, the distribution of CD44 was similar to that ofICAM-1 on the ICAM-1 transfected L cells. If CD44 and ICAM-1 do associate in the plane ofthe membrane, then the cytoplasmic domain of ICAM-1 may not be involved; our results areconsistent with the CD44:ICAM-1 association hypothesis. The primary Ab directed againstCD4 was used as a negative control since it shares the same isotype as YN1/1.7 Ab.Parental, untransfected L cells were also analyzed by immunofluorescence as a controlfor the ICAM-1 transfectants shown in Figure 23. The results of this analysis is shown inFigure 24. Figure 24 shows that the untransfected L cells are ICAM-1 negative, as well asshowing the distribution of CD44 and transferrin receptor (T.R.). The distribution of CD44appeared to be the same on untransfected L cells as it was on ICAM-1 transfected L cells.Transferrin receptor was expressed on L cells at a much lower level, and often exhibited apolarized distribution on individual cells.To summarize the results of the studies on the cytoplasmic domain of ICAM-1,whilethe cytoplasmic domain of ICAM-1 could be shown to be of functional importance, themechanism by which the cytoplasmic domain functions could not readily be defined. Theresults obtained suggest that the cytoplasmic domain of ICAM-1 does not strongly influencethe cell surface distribution of ICAM-1, at least as visualized by fluorescence microscopy.106107Figure 24 The distribution of cell surface molecules on untransfected L cells. Photomicrographs of parental, untransfected L cells which were handled asdescribed for Figure 23. For each exposure taken under fluorescence (Fl.),an accompanying exposure was taken under phase contrast (P.C.).108IC AM-1 Fl.P.C.CD44109Fl.P.C.1 1 0 Fl_P _c4°0DiscussionSoluble versions of adhesion molecules as inhibitors of cell:cell adhesionThe original goal for this project was to explore the use of sICAM-1 as an inhibitor ofthe cell mediated immune responses, involving LFA-1:ICAM-1 interaction, in GVHD.However, the results suggested from an early stage that sICAM-1 would be a relatively poorinhibitor of ICAM-1:LFA-1 mediated cell adhesion. It is important to stress that in the twofunctional assays for sICAM-1 used in this study, namely the binding of radio-iodinatedsICAM-1 in solution to LFA-1+ cells and the binding of LFA-1+ cells to sICAM-1 coatedplastic, sICAM-1 was considered to be functional as determined by it's ability to bind to LFA-1+ cells; however, the affinity between sICAM-1 and LFA-1 seemed low.In general there is much speculation and uncertainty regarding the potential for solubleforms of adhesion molecules to inhibit cell adhesion in vivo. The report on a soluble cadherin(69) demonstrated inhibition of cell:cell adhesion at relatively low concentrations (low ng/mlrange), suggesting cell adhesion is mediated by high affinity interaction. The report on asoluble form of Mac-1 (73) demonstrated that the molecule apparently inhibited cell adhesion ata relatively low concentration (50 ng/ml). The report on recombinant sELAM-1 showed thatsELAM-1 in solution was capable of inhibiting only about 60% of the ELAM-1 mediatedadhesion in an adhesion assay similar to the one described in this thesis (71). The publishedreport on murine sICAM-1 (75) claimed that the molecule mediated adhesion when coated toplastic and also that in solution the molecule blocks ICAM-1:LFA-1 mediated adhesion,although no data was shown to support either claim. The shedding of ICAM-1 from humanmelanoma cell lines has recently been reported (78), and the authors show that the shed ICAM-1 apparently inhibits both ICAM-1:LFA-1 mediated cell adhesion and cell mediatedcytotoxicity. Neutrophils have been observed to shed Leu-8/MEL-14/LECAM-1 uponactivation, and it has been proposed that this shedding "may be a mechanism for rapid111alteration of neutrophil adhesion characteristics" (80). A novel isoform of CD44, CD44R1,contains a potential tryptic cleavage site in its extracellular region (109), and a soluble form ofCD44 has been detected in serum (110). Soluble hyaluronic acid will inhibit the binding ofcells to plastic absorbed hyaluronic acid (G. Dougherty, personal communication), and othershave hypothesized that CD44 mediated adhesion may be modulated by soluble molecules(111). In addition a shed, soluble form of CD43 has been characterized at concentrations of upto 10 tg/ml in plasma (81). While there are many reports in the literature concerning inhibitionby soluble forms of adhesion molecules, there is currently no consensus on the potential forsoluble forms of adhesion molecules to inhibit cell adhesion in vivo.Cell:cell adhesion is by its very nature a form of multivalent interaction with differentcharacteristics than monovalent interaction. These characteristics have significant implicationsfor the use of soluble adhesion molecules as inhibitors of cell adhesion. With monovalentbinding, stable interaction will only persist if the interaction is of relatively high affinity.However, with multivalent binding, stable interaction will persist even if the interactionbetween individual molecules is of considerably lower affinity. A good analogy is velcro. The"affinity" between one hook and it's interacting loop is extremely low; however, the adhesiveinteraction between the two multivalent velcro materials is relatively strong. In molecularterms, multivalent interaction decreases the apparent dissociation rate as compared to thecorresponding monovalent interaction. The effect of valency and apparent dissociation kineticshas been well studied, especially with antibody molecules, for which the decrease from thedivalent to the monovalent state has been reported to increase the apparent rate of dissociationby between five and ten fold (112). The term avidity is used in immunology to represent thenet strength of the interaction between one multivalent reagent, such as an antibody molecule,and another multivalent reagent, such as a virion. The net avidity of a multivalent interaction is"a complex function of the valences of both reactants and the affinities of the variousdeterminants involved" (113). It is easy to imagine how for cells interacting with other cells,112the sum of many low affinity interactions can result in a relatively stable, adhesive interaction.The measurement of cell adhesion is, by it's very nature, a measure of avidity and not affinity.In contrast the binding of a soluble version of an adhesion molecule, in solution, to it'scellular receptor is a monovalent interaction, and does not benefit from the stabilizing orenhancing properties of avidity. This weak monovalent binding will also not efficiently inhibitcell:cell adhesion, because of the kinetics of the binding. Once cell:cell adhesion is initiated,even if the affinity of the interaction is low, cell:cell adhesion will be maintained because of theeffects of avidity. At the molecular level individual counter receptors will associate but thendissociate relatively quickly; however, the overall cell:cell adhesion will persist. If a solubleversion of one of the two counter receptors is introduced into the system, it will bind to it'scounter receptor with the same affinity (and kinetics) as it's cell associated analogue.However, since the monovalent dissociation kinetics are relatively high and the membranes ofthe two opposing cells are already in close proximity, a relatively high concentration of thesoluble adhesion molecule would be required to inhibit cell:cell adhesion by saturating thecellular adhesion receptors. Of course the higher the affinity of the monovalent binding, thelower the concentration of the soluble adhesion molecule which would inhibit cell:celladhesion.The data presented in this thesis suggests that ICAM-1 binds to LFA-1 with a lowaffinity, thus ICAM-1:LFA-1 mediated cell adhesion is not easily inhibited by sICAM-1. Inorder to test this hypothesis a multivalent form of sICAM-1 was used to inhibit ICAM-1:LFA-1mediated cell adhesion in the quantitative cell adhesion assay. The results obtained support thelow affinity hypothesis and explain why monovalent sICAM-1 was unable to inhibit celladhesion. The results also emphasize the importance of multivalent interaction in LFA-1:ICAM-1 mediated cell adhesion. The results with multivalent sICAM-1 suggested that such aform of the molecule may be useful for inhibition of in vivo immune responses.With LFA-3, a multimeric form has been compared to the monomeric form with respectto it's binding to CD2, it's stimulatory effects, and it's inhibition of CD2:LFA-3 mediated113adhesion (38). The multimeric form of soluble LFA-3 bound to CD2 more avidly and wasmore stimulatory than was the monomeric soluble form. In addition, the multimeric form ofsoluble LFA-3 was far more effective at inhibiting CD2:LFA-3 mediated cell adhesion. Theresults with multivalent sICAM-1 reported here are in agreement with the results concerningmultimeric LFA-3. In addition, since the CD2:LFA-3 interaction is of a relatively low affinity(37), the analogy between the CD2:LFA-3 and ICAM-1:LFA-1 systems supports the notionthat the ICAM-1:LFA-1 interaction is of relatively low affinity. Thus it would seem thatoptimal inhibition of cell mediated immune responses with sICAM-1 will require the use of amultivalent form of sICAM-1.As mentioned above, several other groups have claimed that sICAM-1 can inhibit celladhesion (75, 78). It should be noted that the method by which sICAM-1 was produced forthis study differs somewhat from the two other published reports on the production of sICAM-1. For murine sICAM-1 (75), the truncated cDNA was subcloned into an expression vector inwhich expression is driven from the SV40 early promoter and then transfected into CHO cells.SICAM-1 was then purified from the SN by anion exchange followed by immunoaffinitychromatography. Human sICAM-1 was produced by a very similar procedure (74). Themurine sICAM-1 which contained only four complete domains and a portion of the fifth had areported molecular weight of only 50 kd, suggesting inappropriate glycosylation. Since weproduce sICAM-1 in lymphocytic cells (NS-1), our protein is likely appropriately glycosylated.Appropriate glycosylation is not trivial, especially for ICAM-1, since it has been shown thatglycosylation can influence it's biological activity (114). The report on the shed form ofsICAM-1 does not include a rigorous characterization of the molecule (78). Again, the resultsof this thesis suggest that while monovalent sICAM-1 is inefficient at inhibiting cell:celladhesion, multivalent sICAM-1 will inhibit cell:cell adhesion more effectively. It is possiblethat the two reports claiming inhibition by sICAM-1 (75, 78) represent the effects of eitheraggregated sICAM-1 or membrane bound ICAM-1 vesicles in the case of shed sICAM-1, since114I have shown that only multivalent sICAM-1 is capable of effectively inhibiting ICAM-1:LFA-1 mediated cell adhesion.Regulation of LFA-1:ICAM-1 mediated cell adhesionCell adhesion mediated by LFA-1 and ICAM-1 is known to be actively regulated.However, the mechanisms that regulate LFA-1:ICAM-1 mediated cell adhesion are potentiallyvery complex and currently not well understood. The studies in this thesis showed that ICAM-1:LFA-1 mediated cell adhesion can be upregulated by PMA or Mn++. These two effects havedifferent sensitivities to inhibition by sICAM-1. In addition, the cell surface distribution ofICAM-1 is shown to be differentially regulated, at least in different cell lines, and thesedifferences in distribution are hypothesized to have significant implications for cell:celladhesion (discussed below). Taken together, the results presented in this thesis stronglysuggest that there are multiple mechanisms regulating cell adhesion mediated by LFA-1 andICAM-1.There are many reports in the literature concerning the mechanisms regulating celladhesion mediated by LFA-1 and ICAM-1. Springer's group has shown that the adhesion ofcells to absorbed ICAM-1 is dependant on a functional actin cytoskeleton, active metabolism,divalent cations, and also activation from the so called low to high avidity forms of LFA-1 (5,11, 12, 13). The avidity of LFA-1 has been shown to be upregulated by treatment of cells withPMA or by TcR crosslinking (42). It has been assumed that the avidity of LFA-1 isdetermined by it's affinity for ICAM-1, and that PKC mediated phosphorylation of the LFA-113 chain, stimulated by phorbol ester or TcR ligation, regulates the affinity of LFA-1 (42, 115).However, in the absence of an assay for the binding of sICAM-1 in solution, the relationshipbetween the affinity of LFA-1 for ICAM-1 and it's avidity could not be determined. Figdor'sgroup has also studied LFA-1 regulation (45). This study demonstrated that T cell signallingthrough CD2 or CD3 increased the avidity of LFA-1. Signalling through CD2 resulted in a115persistent avidity increase, while signalling through CD3 resulted only in a transient avidityincrease.Multiple mechanisms of Mac-1 avidity regulation have also been reported. Since LFA-1 and Mac-1 share a common 13 chain, it would seem likely that they also share regulatorymechanisms. Detmers et. al have demonstrated that phorbol ester stimulation results in theaggregation of Mac-1 molecules in the plane of the membrane (51), and there are a number ofreports suggesting that the affinity of Mac-1 for it's ligands can be regulated (116, 117).I propose that there are three primary mechanisms regulating cell:cell adhesion mediatedby LFA-1 and ICAM-1. The first mechanism is simply the amount of LFA-1 or ICAM-1 agiven cell expresses on it's surface. The second mechanism is the modulation of the affinity ofthe interaction between ICAM-1 and LFA-1, by any means the cell may have at it's disposal.The third mechanism is the regulation of cell adhesion by the regulation of the cell surfacedistribution of adhesion molecules. The remainder of the Discussion of this thesis will involveconsideration of these proposed methods of cell adhesion regulation, with regard to my resultsand those of others.1. Regulation of adhesion by modulating expression levels on the cell surfaceCells will not adhere to other cells unless they express adhesion molecules on theirsurface. Naturally, the higher the level of adhesion molecules expressed on the surface of acell, the more avidly it will adhere to other cells expressing the proper counter receptor.Adhesion regulation by regulation of the level of adhesion molecule expression wasexemplified in the present study from the aggregative phenotype of the ICAM-1 transfectedT28 cells. The parental T28 cells, which express very low levels of ICAM-1, grow in singlecell suspension whereas the transfected T28 cells expressing high levels of ICAM-1 grow inhomotypic aggregates, demonstrating that expression of ICAM-1 on LFA-1+ cells enablescell:cell adhesion. In vivo ICAM-1 is expressed on relatively few cells; however, ICAM-1expression is induced by a number of inflammatory mediators on a number of cell types (18).116A good example of this mechanism at work in vivo is the induction of ICAM-1 expression onendothelium, facilitating the binding and extravasation of neutrophils, monocytes, andlymphocytes into sites of tissue insult. Another example of this method of adhesion regulationis the ability of monocytes to rapidly mobilize cytoplasmic stores of ICAM-1 to their surfacefollowing stimulation (50). In general, the cell surface expression level of ICAM-1 is animportant regulatory mechanism for LFA-1:ICAM-1 mediated cell adhesion. However, LFA-1is constitutively expressed on leukocytes and the avidity of LFA-1 seems to be regulated byother mechanisms.2. ICAM-1:LFA-1 affinity modulationThe adhesive interaction between cells mediated by ICAM-1:LFA-1 interaction may beregulated by the affinity of the interaction between individual ICAM-1 and LFA-1 molecules.It has long been hypothesized that modulation of affinity regulates cell:cell adhesion (2, 4);however, rigorous testing of this hypothesis has been slow because of the lack of affinityassays in which soluble versions of adhesion molecules bind to their cellular counter receptors.The exception to this is the CD2:LFA-3 system. Using soluble forms of both CD2 and LFA-3,the affinity of the interaction between these molecules has been determined. These studiesdemonstrated that the affinity of the CD2:LFA-3 interaction is surprisingly low, with a Kd of10 -6 M (37, 38).In my studies, the measurement of the affinity of LFA-1:ICAM-1 binding was not donemainly because of the difficulties in obtaining consistent results. In some experiments, 125I-sICAM-1 seemed to bind to LFA-1+ cells efficiently at concentrations in the ng/ml range,whereas in other experiments concentrations in the pg/m1 range were required (Figure 5).There are two possible explanations to account for these discrepancies. The first is that insome experiments 125I-sICAM-1 exists in multivalent aggregates; therefore, the presumed highaffinity binding is actually multivalent high avidity binding with the same affinity. why 1251_sICAM-1 would exist in aggregates in some experiments and in monomeric form in others is117not known, especially since aliquots of the same batches of frozen protein yielded inconsistentresults. The other explanation, which we labored over for some time, is that the iodinationprocess may inactivate the biological properties of the sICAM-1. Again, the inconsistencies inthe results are difficult to account for by this explanation. For example, some iodinationsyielded apparently active 125I-sICAM-1, while others using the very same iodination protocolyielded inactive 125I-sICAM-1. Along this line a more 'gentle' iodination protocol was used inwhich Na 125I is oxidized in the absence of protein to give free 1251, then the free 1251 istransferred to the protein solution and labelling is allowed to proceed without exposing theprotein to the potentially harsh oxidizing conditions. However, this method did not improvethe results of the binding assay. In addition a two phase iodination protocol, which isreportedly very gentle, was tested without any improvement. With respect to affinity, the othersalient data from this study to keep in mind is that PMA stimulated ICAM-1:LFA-1 mediatedcell adhesion is not inhibited by sICAM-1 in solution at up to 100 µg/ml. From all of ourresults it would seem that the affinity of LFA-1:ICAM-1 binding is relatively low, at least incomparison to Ab:Ag or growth factor:receptor systems.The observation that the normally low affinity interaction between LFA-1 and ICAM-1can be increased by Mn++ is in agreement with the the observed effects of Mn++ on affinity inother systems. In the human system, Dransfield et al. have observed that Mn++ induces LFA-1:ICAM-1 mediated cell adhesion (105). As well in this experimental system Mn++ inducesthe LFA-1 24 epitope, which is presumably an indication of an LFA-1 conformational changeaccompanying the increase in the affinity of LFA-1 for ICAM-1. The binding of radio-iodinated fibrinogen and factor X to Mac-1 has recently been demonstrated (103), and thisbinding assay was used to show that Mn++ increases the affinity of the interaction betweenMac-1 and it's ligands. In addition, the affinity of RGD containing peptides for the 133 integrinGPIIb/11Ia has been determined (104), and Mn++ was also found to increase the affinity of thisintegrin for the RGD peptides. Therefore, Mn++ appears to have a general effect on integrins;the engagement of Mn++, as opposed to Ca++ or Mg++, by integrins apparently induces some118conformational change in the integrins' extracellular region such that the affinity for it's ligandis increased. In contrast to Mn++, we have consistently been unable to detect an increase in thebinding of 125I-sICAM-1 to LFA-1+ cells upon stimulation with PMA. In addition, Mn++induced homotypic aggregation could be inhibited by sICAM-1 in solution, whereas PMAinduced homotypic aggregation could not. These results suggest that PMA most likelyincreases the avidity of LFA-1:ICAM-1 mediated cell adhesion without increasing the affinityof ICA M-1:LFA-1 binding.SICAM-1 was unable to consistently inhibit Mn++ induced adhesion in the quantitativecell adhesion assay, even at a concentration of up to 110^(results not shown). However,sICAM-1 did inhibit the Mn++ induced aggregation of cell lines (Figure 9). It would seem thatin the quantitative cell adhesion assay, inhibiting adhesion by blocking LFA-1 is difficultunless the inhibiting agent binds with relatively high avidity. The observation that Mn++induced adhesion in the quantitative cell adhesion assay was inhibited significantly bycytochalasin B was unexpected. This adhesion, presumably mediated by an increase in affinityof the ICAM-1:LFA-1 interaction, was expected to be independent of the actin cytoskeleton(see below). However, adhesion mediated by multiple mechanisms appears to be verysensitive to inhibitors of the actin cytoskeleton.The possible physiological relevance of Mn++ induced cell adhesion is unclear at thispoint. Circulating concentrations of Mn++ have been reported to reach 0.7 tM (118).Lymphocytes are known to contain Mn++ dependant enzymes in mitochondria (119), andlymphocytes are also known to store a significant portion of the Mn++ in blood (120, 121). Ithas been hypothesized that Mn++ release may provide "a rapid mechanism for the activationand recruitment of multipotent leukocyte adhesive properties" (103).While it would seem that the cytoplasmic domain of LFA-113 is most likely responsiblefor cytoplasmic linkage resulting in "high avidity" LFA-1 (see below), the a chain'scytoplasmic domain remains a candidate for modulating LFA-1's affinity. Indeed there is onerecent report in which the a chain of a P3 integrin was shown to influence the affinity of the119integrin complex for it's ligand (122). In this report truncation of the a chain cytoplasmicdomain resulted in an increase in the affinity of the integrin for it's ligand, and the authorsspeculated that the cytoplasmic domain of the a chain might facilitate binding to somecytoplasmic moiety which maintains the integrin complex in the low affinity state. The authorsalso speculate that cell activation may result in the modification of this moiety such that theintegrin complex is converted to the high affinity state. The location of the divalent cationbinding site on the a chain combined with the effect of Mn++ on integrin affinity also supportsthe notion that the a chain may be involved in affinity regulation.One very recent report has suggested that the avidity of (32 integrins can be modulatedby a small molecular weight lipid present in stimulated but not in unstimulated humanneutrophils (116). This molecule has been termed Integrin Modulating Factor-1 (IMF-1), andit appears to be an unsaturated fatty acid or isoprenoid acid. While the Mac-1 assays used inthis study are cell adhesion assay and therefore measure avidity, certain aspects of the datastrongly suggest that the effect of IMF-1 is to increase the affinity of (32 integrins for theirligands. In particular, when the authors immobilized purified Mac-1 onto polystyrene, treatedthe immobilized Mac-1 with IMF-1, and then measured the ability of untreated or IMF-1treated, immobilized Mac-1 to bind to C3bi coated erythrocytes, they observed the IMF-1treatment to greatly increase the binding ability of the absorbed Mac-1. IMF-1 activity could befound in lipids from neutrophils stimulated with PMA, TNF, PAF, and fMLP but notunstimulated neutrophils.How IMF-1 fits in with more conventional concepts of integrin regulation, such asprotein phosphorylation and cytoskeletal interaction, is unclear at this point. However, thisreport does add credence to reports in the literature concerning integrin conformational changeshypothesized to be indications of activation. With respect to LFA-1, two other groups havecharacterized 'activated' forms of LFA-1 through the use of mAb. Figdor et al. have raised amAb, NKI-L16, which binds to an epitope of LFA-1 dependant on divalent cations and cellularactivation (45, 46, 47). This Ab also facilitates LFA-1:ICAM-1 mediated homotypic120aggregation of cell lines, presumably by stabilizing an activated LFA-1 conformation. Basedon their data, this group has postulated the existence of three distinct conformational forms ofLFA-1 on the cell corresponding to distinct stages of activation (48). Dransfield and Hogghave similarly characterized an LFA-1 epitope dependant on active metabolism and divalentcations through the use of their mAb, Ab24 (49). The 24 epitope is also known to be inducedby Mn++ (105). A similar situation exists with Mac-1, for which an antibody has been raisedrecognizing an epitope found only on activated cells (117). In addition, for the integrinIIb/IIIa, evidence for the presence of an activation epitope has been published (123).3. The regulation of cell:cell adhesion by the regulation of the cell surfacedistribution of adhesion moleculesLogically, the most obvious way that avidity can be increased without increasingaffinity or expression level is by changing the distribution of adhesion molecules on the surfaceof the cell. High local concentrations of adhesion molecules at the sites of cell:cell contactwould increase the avidity of adhesion. The theoretical dissociation rate of cells adhering viaevenly spaced adhesion molecules would be much more rapid than that for cells adhering viaclustered adhesion molecules, as illustrated in Figure 25. The difference between randomlydistributed receptors and clustered receptors with respect to cell adhesion has been verifiedexperimentally (124). This report demonstrates that clustered receptors in general mediatemore avid cell:cell adhesion than do evenly distributed receptorsDespite the potential for distribution to significantly affect cell:cell adhesion, relativelylittle is known about the actual cell surface distribution of adhesion molecules, and what effectsthis distribution has on adhesion. In the case of the ICAM-1:LFA-1 adhesion system, thedistribution of these molecules on the cell surface is likely regulated, as will be discussedbelow. With respect to the distribution of adhesion molecules on the cell surface, LFA-1 andICAM-1 will be discussed separately, starting with LFA-1.121a122b.....,,,, ,Wk)^17;1Figure 25^The effect of adhesion molecule distribution on cell:cell adhesion. The theoretical dissociation of cell:cell adhesion mediated by diffuse (a)counter receptors would be faster than that mediated by clustered (b)counter receptors.The primary line of evidence suggesting that the distribution of LFA-1 has significantfunctional consequences comes from the analogy with Mac-1, which of course shares acommon 13 chain with LFA-1. For Mac-1 an aggregation of receptors, preceding any ligandbinding, has been shown to occur with phorbol ester stimulation (51). This aggregation ofreceptors in the plane of the membrane was determined to correspond to increased ligandbinding capabilities. Several lines of evidence suggest that the cytoskeleton plays a role in thedistribution of LFA-1 on the cell surface. Firstly, cytochalasin B, an inhibitor of actinpolymerization, inhibits LFA-1 mediated cell adhesion (Figures 7, 8, 10, 11). Secondly, LFA-1 has been reported to associate with talin in a PKC dependant fashion (54, 55). Talin is aprotein believed to provide a linkage between the actin cytoskeleton and integrin molecules (58,63). The cytoplasmic domain of integrin 13 chains is believed to facilitate their cell surfacelocalization in conjunction with the actin cytoskeleton (discussed below).Phorbol esters, which upregulate LFA-1 avidity, stimulate the PKC mediatedphosphorylation of a spectrum of proteins in the cell in addition to LFA-1 13 (125, 126).Indeed Hibbs et al. recently reported that the serine phosphorylation of LFA-1 13 could beseparated from the phorbol ester induced, LFA-1:ICAM-1 mediated adhesion (127). Sincephorbol ester induced LFA-1:ICAM-1 mediated adhesion is dependant on a functional actincytoskeleton in addition to PKC mediated phosphorylation events, it seems likely thatphosphorylation of protein(s) distinct from LFA-1 p link the actin cytoskeleton to the LFA-1complex, thus facilitating the high avidity form of LFA-1.There is one report in the literature about a poorly characterized protein foundassociated with LFA-1 in the cytoplasm (66). The protein is a likely candidate for theregulation of LFA-1 especially since the authors feel that the protein is only transientlyassociated with the LFA-1 complex. This 86 kd protein may reflect either an LFA-1:cytoskeletal link, or a mechanism by which the affinity of LFA-1 is modulated.Since the phosphorylation of LFA-1 13 is not required for the phorbol ester inducedincrease in the avidity of LFA-1, a model in which some other molecule, such as talin or the 86kd LFA-1 associated protein, links LFA-1 13 to the actin cytoskeleton seems likely. In supportfor the talin hypothesis, talin is known to be a PKC substrate (64).For 131 integrins the cytoplasmic domain of the p chain appears to facilitate a linkbetween the integrin complex and the cytoskeleton. Solowska et al. (128) demonstrated thatthe cytoplasmic domain of avian 131 integrin was responsible for the localization of integrinmolecules to focal contacts, though this region of the molecule did not seem to be important forligand binding. Marcantonio et al. (129) argued that the 131 region of the cytoplasmic domain123closest to the membrane is necessary for localization of avian p i integrin to focal contacts. Thecytoplasmic domains of the integrin 13 chains are highly conserved (127). By analogy the 13idata provides evidence that the cytoplasmic domain of LFA-1 13 is also responsible for anintegrin:cytoskeleton link.Except for the T28 transfectants, I could not detect any change in the distribution ofLFA-1 after stimulation with PMA. These observations were originally somewhatdiscouraging; however, fluorescence microscopy might not be a sufficiently sensitive techniquefor characterizing functionally significant changes in adhesion molecule distribution. Phorbolester stimulation may induce relatively subtle changes in LFA-1 distribution, such asdimerization, which would require electron microscopy to visualize. The Mac-1 studyemployed electron microscopy, and indeed the degree of receptor aggregation was notextensive. Using flourescence microscopy, the ICAM-1 transfected T28 cells were clearlyobserved to exhibit more punctate LFA-1 staining following PMA induced aggregation. Thiseffect was not observed in any of the cell lines, including the untransfected T28 cells, prior tothe ICAM-1 transfected T28 lines; therefore, this phenomenon is likely an example of mutualcapping due to the extremely high level of ICAM-1 expression on the T28 transfectants. Whilethis effect may be relevant in some respect, it does not likely indicate that the change LFA-1distribution preceded aggregation or was involved in stimulating aggregation. It is possiblethat for LFA-1, as for Mac-1, electron microscopy is required to detect the microclusterring ofreceptors, prior to ligand engagement, stimulated by phorbol esters.The cell surface distribution of ICAM-1 may also strongly influence ICAM-1:LFA-1mediated cell adhesion. In fact one of the most striking observations from my work was thepunctate ICAM-1 spots on A20 cells. It is likely that the expression of punctate ICAM-1 onA20 cells is responsible for their observed aggregation in culture. An ICAM-1 distributionsimilar to that observed on A20 cells has also been reported on a human T cell line exhibitinghomotypic aggregation in culture (50), and as on the A20 cells, LFA-1 did not co-localize withICAM -1.124Interestingly, the punctate distribution of ICAM-1 on A20 cells is efficientlymaintained even when the cells are disaggregated and incubated with cytochalasin B, whichdestroys the actin cytoskeleton. In another system, namely the human KG1a myelomonocyticcell line, treatment with cytochalasins also does not destroy punctate ICAM-1 localization (G.J.Dougherty, personal communication).The punctate ICAM-1 staining on A20 cells and the punctate ICAM-1 staining patternsreported by others prompted us to hypothesize that the cytoplasmic domain of ICAM-1,through linkage to intracellular proteins, may be responsible for controlling ICAM-1's cellsurface distribution. There is only one reference in the literature to studies on the cytoplasmicdomain of ICAM-1 (56); however, this report was not definitive and we wanted to conduct ourown study on the cytoplasmic domain of ICAM-1. We therefore transfected either completeICAM-1 or ICAM-1 without it's cytoplasmic domain onto T28 and L cells and determined boththe function of the two versions of ICAM-1 and their cell surface distribution. The results ofthese experiments showed that the cytoplasmic domain of ICAM-1 is functionally important,while not absolutely necessary, especially for adhesion between unstimulated cells. Thisobservation, together with the data on ICAM-1 cell surface distribution suggested that thecytoplasmic domain of ICAM-1 may determine the molecule's cell surface distribution.Despite the functional importance of the cytoplasmic domain of ICAM-1, analysis of ICAM-1cell surface distribution with or without the cytoplasmic domain failed to define a role for thecytoplasmic domain of ICAM-1 in determining it's cell surface distribution. The distribution ofICAM-1 with or without it's cytoplasmic domain seemed consistent, exhibiting a somewhatpatchy distribution in both forms. However, as for LFA-1, the fluorescence microscopicanalysis used in this study may not be sufficient to demonstrate functionally significantdifferences in the cell surface distribution of adhesion molecules.It has been hypothesized that ICAM-1 may co-cap with other molecules in the plane ofthe membrane (G.J. Dougherty, manuscript in preparation). This hypothesis is based on theobservation that ICAM-1 localizes to uropods of several cell lines exhibiting homotypic125aggregation, and on these cell lines LFA-1 does not mutually cap. The term co-cap is used inthe literature to represent the association of two molecular species in the plane of a single cell'smembrane; mutual capping represents the association of two molecular species, each in themembrane of an individual cell, when the two cells are in contact with each other (53). SinceICAM-1 and LFA-1 are not observed to mutually cap, the capping of ICAM-1 is hypothesizedto be mediated by co-capping. ICAM-1 has been observed to co-cap with CD44 as well aswith CD43 on KG la cells (G.J. Dougherty, manuscript in preparation). The hypothesis thatCD44 and ICAM-1 associate in the plane of the membrane remains to be rigorouslydemonstrated or disproven; the results presented in this thesis are not decisive with respect tothis hypothesis. The L cells used in this study were strongly CD44 positive, and the stainingpattern for CD44 and both versions of ICAM-1 transfected into these cells were very similar(Figures 23 and 24). This observation might suggest that ICAM-1 can associate with CD44with or without it's cytoplasmic domain. However, the cytoplasmic domain of ICAM-1appears to be functionally important in facilitating ICAM-1:LFA-1 mediated adhesion, andthere are two possible explanations for this observation. The first explanation is that thecytoplasmic domain of ICAM-1 might be necessary for it's interaction with CD44, and thesecond is that the cytoplasmic domain of ICAM-1 might provide a direct link with thecytoskeleton. Both of these explanations imply that the cytoplasmic domain of ICAM-1facilitates some form of ICAM-1 aggregation in the plane of the membrane, which likelyrequires more sensitive techniques such as electron microscopy to demonstrate.In addition to modulating the aggregation of LFA-1 molecules in the plane of themembrane, the actin cytoskeleton is likely involved in LFA-1 mediated functions in a moremechanical sense. PMA activation of neutrophils in suspension induces a change in shapefrom spherical to bipolar (23), presumably due at least in part to the actin cytoskeleton. Inaddition, when neutrophils, rolling under physiological sheer stress on an artificial CD62 andICAM-1 substrate, are stimulated by either phorbol ester or fMLP, their rolling velocities slowto a stop, then the strongly adherent cells begin to flatten and spread (23). This kind of126alteration of cell shape favoring adhesion, and also facilitating other biological processes suchas extravasation, is presumably due at least in part to the actin cytoskeleton. An attempt tocharacterize this phenomenon in our experimental system with cytochalasin B was notcompletely satisfying. The results of this experiment is shown in Figure 13. Whilecytochalasin B did reproducibly induce an alteration in cellular morphology, decreasedflattening of cells on a sICAM-1 coated substrate was not demonstrable. In addition, througheither cytoskeletal linkage or associations in the plane of the membrane, adhesion moleculesmight be physically braced and better anchored in the membrane. With such linkages adhesionmolecules might be better able to mediate cell:cell adhesion. This mechanism might be moreprominent with high affinity binding or under high sheer velocities.Summary: towards a better understanding of LFA-1:ICAM-1 mediated celladhesionBy exploring the use of sICAM-1 as an inhibitor of cell adhesion we have gainedinsight into the intricacies of adhesion regulation. While sICAM-1, at least in it's monovalentform, has proven unsatisfactory for use in inhibiting in vivo immune responses, it has provenuseful as a probe for the mechanisms regulating cell adhesion mediated by ICAM-1 and LFA-1. Through the studies presented in this thesis, we have addressed the mechanisms which wefeel are important in the regulation of ICAM-1:LFA-1 mediated cell adhesion.The regulation of the level of adhesion molecules expressed on the cell surface is themost obvious and well characterized mechanism regulating adhesion mediated by LFA-1 andICAM-1.It is very likely that affinity modulation is a mechanism for regulating ICAM-1:LFA-1mediated cell adhesion. Mn++ is believed to increase the affinity of LFA-1 for ICAM-1,demonstrating the potential for alterations in the extracellular conformation of integrins tomodulate the affinity of the interaction between integrin and ligand. As discussed above, moredirect evidence for functionally significant changes in integrin conformation exist in the127literature. The cytoplasmic domain of LFA-1 a remains a good candidate for mediating thiskind of 'inside out' signalling resulting in integrin affinity modulation.It is possible that LFA-1:ICAM-1 mediated cell adhesion is also regulated by theregulation of the cell surface distribution of these molecules. This mechanism is largelyoverlooked in the literature, yet it has great potential for modulating adhesion. Both PKCmediated phosphorylation events and the actin cytoskeleton appear to be involved in theregulation of LFA-1 avidity apart from expression levels or affinity; this LFA-1 regulation islikely effected through changes in the cell surface distribution of LFA-1. The cytoplasmicdomain of LFA-1 f3 is a good candidate for a cytoplasmic link, but the details of this link aswell as it's effects remain unclear. With respect to ICAM-1, cell surface distribution can bedifferentially regulated, and this distribution is also hypothesized to play a significant role incell adhesion. The cytoplasmic domain of ICAM-1 remains a good candidate for the regulationof ICAM-1 function through modulation of the distribution of ICAM-1 on the cell surface.128Bibliography1. Freidfelder, D. 1982. In "Characterization of macromolecules in physicalbiochemistry: applications to biochemistry and molecular biology. W.H. Freemanand Company, second edition, Chapter 1, pp 1-35.2. Takeichi, M. 1990. Cadherins: a molecular family important in selective cell-celladhesion. Ann. Rev. Biochem. 59: 237.3. Yoshida-Noro, C., N. Suzuki, M. Takeichi. 1984. Molecular nature of thecalcium dependant cell-cell adhesion system in mouse teratocarcinoma andembryonic cells studued with a monoclanal antibody. Development 101: 19.4. Springer, T.A. 1990. Adhesion receptors of the immune system. Nature 346:425.5. Staunton, D.E., S.D Marlin, C. Stratowa, M.L. Dustin, T.A. Springer. 1988.Primary structure of intercellular adhesion molecule 1 (ICAM-1) demonstratesinteraction between members of the immunoglobulin and integrin supergenefamilies. Cell 52: 925.6. Williams, A.F. and A.N. Barclay. 1988. The immunoglobulin superfamilydomains for cell surface recognition. Ann. Rev. Immunol. 6: 381.7. Larson, R.S. and T.A. Springer. 1990. Structure and function of the leukocyteintegrins. Immunol. Rev. 114: 181.8. Anderson, D.C. and T.A. Springer. 1987. Leukocyte adhesion deficiency: aninherited defect in Mac-1, LFA-1, and p150/95 glycoproteins. Ann. Rev. Med. 38:175.9. Lisowaska-Grospierre, B., M. Bohler, A. Fisher, C. Mawas, T.A. Springer, C.Griscelli. 1986. Defective membrane expression of the LFA-1 complex may besecondary to the absence of the beta chain in a child with recurrent bacterialinfection. Eur. J. Immunol. 16: 205.10. Davignon, D., E. Martz, T. Renolds, K. Kurzinger, T.A. Springer. 1981.Monoclonal antibody to a novel lymphocyte function-associated antigen (LFA-1).Mechanism of blockade of T lymphocyte-mediated killing and effects on T and Blymphocyte function. J. Immunol. 127: 590.11. Rothlein, R. and T.A. Springer. 1986. The requirement for lymphocyte functionassociated antigen 1 in homotypic leukocyte adhesion stimulated by phorbol ester.J. Exp. Med. 163: 1132.12. Rothlein, D., M.L. Dustin, S.D. Marlin, T.A. Springer. 1986. A humanintercellular adhesion molecule (ICAM-1) distinct from LFA-1. J. Immunol. 137:1270.13.^Marlin, S.D. and T.A. Springer. 1987. Purified intercellular adhesion molecule-1129(ICAM-1) is a ligand for lymphocyte function-associated antigen -1 (LFA-1).Cell 51: 813.14. Dustin, M.L. and T.A. Spinger. 1988. Lymphocyte function-associated antigen-1(LFA-1) interaction with intercellular adhesion molecule-1 (ICAM-1) is one of atleast three mechanisms for lymphocyte adhesion to cultured endothelial cells. J.Cell Biol. 107: 321.15. Staunton, D.E., M.L. Dustin, T.A. Springer. 1989. Functional cloning of ICAM-2, a cell adhesion ligand for LFA-1 homologous to ICAM-1. Nature 339: 61.16. de Fougerolles, A.R., S.A. Stacker, R. Schwarting, T.A. Springer. 1991.Characterization of ICAM-2 and evidence for a third counter-receptor for LFA-1.J. Exp. Med. 174: 253.17. Krensky, A.M., F. Sanchez-Madrid, E. Robbins, J. Nagy, T.A. Springer, S.J.Burakoff. 1983. The functional significance, distribution, and structure of LFA-1,LFA-2, and LFA-3: cell surface antigens associated with CTL-target interactions.J. Immunol. 131: 611.18. Dustin, M.L., R. Rothlein, A.K. Bahn, C.A. Dinarello, T.A. Springer. 1986.Induction by IL-1 and interferon, tissue distribution, biochemistry, and function ofa natural adherence molecule (ICAM-1). J. Immunol. 137: 245.19. Nortamo, P., R. Salcedo, T. Timonen, M. Patarroyo, C.G. Gahmberg. 1991. Amonoclonal antibody to the human intercellular adhesion molecule-2. Cellulardistribution and molecular characterization of the antigen. J. Immunol. 146: 2530.20. Rosenstein, Y., J.K. Park, W.C. Hahn, F.S. Rosen, B.E. Bierer, S.J. Burakoff.1991. CD43, a molecule defective in Wiscott-Aldrich syndrome, binds ICAM-1.Nature 354: 233.21. Johnston, G.I., R.G. Cook, R.P. McEver. 1989. Cloning of GMP-140, agranule membrane protein of platelets and endothelium: sequence similarity toproteins involved in cell adhesion and inflamation. Cell 56: 1033.22. Larsen, E., A. Celi, G.E. Gilbert, B.C. Furie, J.K. Erban, R. Bonfanti, D.D.Wagner, B. Furie. 1989. PADGEM protein: a receptor that mediates theinteraction of activated platelets with neutrophils and monocytes. Cell 59: 305.23. Lawerence, M.B., and T.A. Springer. 1991. Leukocytes roll on a selectin atphysiological flow rates: distinction from and prerequisite for adhesion throughintegrins. Cell 65: 859.24. Picker, L.J., R.A. Warnock, A.R. Burns, C.M. Doerschuk, E.L. Berg, E.C.Butcher. 1991. The neutrophil selectin LECAM-1 presents carbohydrate ligands tothe vascular selectins ELAM-1 and GMP-140. Cell 66: 921.25. Butcher, E.C. 1986. The regulation of lymphocyte traffic. Curr. TopicsMicroboil. Immunol. 128: 85.26.^Yednock, T.A. and S.D. Rosen. 1989. Lymphocyte homing. Adv. Immunol. 44:313.13013127. Stoolman, L.M. Adhesion molecules controlling lymphocyte migration. Cell 56:907.28. Tedder, T.F., A.C. Penta, H.B. Levine, A.S. Freedman. 1990. Expression of thehuman leukocyte adhesion molecule, LAM-1. Identity with TQ1 and Leu-8differentiation antigens. J. Immunol. 144: 532.29. Holzmann, B., and I.L. Weissman. 1989. Peyer's patch-specific lymphocytehoming receptors consist of VLA-4-like a chain associated with either of twointegrin p chains, one of which is novel. EMBO J. 8: 1735.30. Mackay, C.R., W.L. Marston, L. Dudler. 1990. Naive and memory T cells showdistinct pathways of lymphocyte recirculation. J. Exp. Med. 171: 801.31. Dembic Z., W. Haas, S. Weiss, J. McCubrey, H. Kiefer, H. von Boehmer, M.Steinmetz. 1986. Transfer of specificity by murine a and (3 T-cell receptor genes.Nature 320: 232.32. Saito, T., A. Weiss, J. Miller, M.A. Norcross, R.N. Germain. 1987. Specificantigen-Ia activation of transfected human T cells expressing murine Ti a(3-humanT3 receptor complexes. Nature 325: 125.33. Shevach, E.M. 1989. Accessory Molecules. In "Fundamental immunology",W.E. Paul, ed., Raven Press, New York, Chapter 15, pp 413-441.34. Baxley, G., G.B. Bishop, A.G. Cooper, H.H. Wortis. 1973. Rosetting ofhuman red blood cells to thymocytes and thymus-derived cells. Clin. Exp.Immunol. 15: 385.35. Wigzell, H. and P. Hayvy. 1974. Specific fractionation of immunocompetentcells. Application in the analysis of effector cells involved in cell mediated lysis.Curr. Topics Microbiol. Immunol. 67: 1.36. Despont, J.P., C.A. Abel, H.M. Grey. 1975. Sialic acid and sialyltransferases inmurine lymphoid cells: indicators of T cell maturation. Cell. Imunol. 17: 487.37. Sayre, P.H., R.E. Hussey, H-C. Chang, T.L. Ciardelli, E.L Reinherz. 1989.Structural and binding analysis of a two domain extracellular CD2 molecule. J.Exp. Med. 169: 995.38. Dustin, M.L., D. Olive, T.A. Springer. 1989. Corelation of CD2 binding andfunctional properties of multimeric and monomeric lymphocyte function-associatedantigen 3. J. Exp. Med. 169: 503.39. Bierer, B.E., B.P. Sleckman, S.E. Ratnofsky, S.J. Burakoff. 1989. The biologicroles of CD2, CD4, and CD8 in T-cell activation. Ann Rev. Immunol. 7: 579.40. Selvaraj, P., M.L. Plunkett, M.L. Dustin, M.E. Sanders, S. Shaw, T.A. Springer.1987. The T lymphocyte glycoprotein CD2 binds cell surface ligand LFA-3.Nature 326: 400.41.^Hoffman, S. and G.M. Edelman. 1983. Kinetics of homophylic binding byembryonic and adult forms of the neural cell adhesion molecule. P.N.A.S. USA80: 5762.42. Dustin, M.L. and T.A. Springer. 1989. T-cell receptor cross-linking transientlystimulates adhesiveness through LFA-1. Nature 341: 619.43. Hara, T. and S.M. Fu. 1986. Phosphorylation of a,13 subunits of 180/100 Kdpolypeptides (LFA-1) and related antigens. In "Leukocyte typing II", Springer,Anderson, Rosenthal, Rothlein, eds., Springer Verlag, New York, pp 77-84.44. Buyon, J.P., S.G. Slade, J. Reibman, S.B. Abramson, M.R. Philips, G.Weissmann, R. Winchester. 1990. Constitutive and induced phosphorylation ofthe a- and (3-chains of the CD11/CD18 leukocyte integrin family. J. Immunol.144: 191.45. van Kooyk, Y., P. van der Wiel-van Kemenade, P. Weder, T.W. Kuipers, C.G.Figdor. 1989. Enhancement of LFA-1-mediated cell adhesion by triggeringthrough CD2 or CD3 on T lymphocytes. Nature 342: 811.46. Keizer, G.D., W. Visser, M. Vliem, C.G. Figdor. 1988. A monoclonal antibody(NKI-L16) directed against a unique epitope on the a-chain of human leukocytefunction-associated antigen 1 induces homotypic cell-cell interactions. J. Immunol.140: 1393.47. van Kooyk, Y., P. Weder, F. Hogervorst, A.J. Verhoeven, G. van Seventer, A.A.to Veld, J. Borst, G.D. Keizer, C.G. Figdor. 1991. Activation of LFA-1 throughCa++-dependant epitope stimulates lymphocyte adhesion. J. Cell Biol. 112: 345.48. Figdor, C.G., Y. van Kooyk, G.D. Keizer. 1990. On the mode of action of LFA-1. Immunol. Today 11: 277.49. Dransfield, I. and N. Hogg. 1989. Regulated expression of Mg2+ binding epitopeon leukocyte integrin a subunits. EMBO J. 8: 3765.50. Dougherty, G.J., S. Murdoch, N. Hogg. 1988. The function of humanintercellular adhesion molecule-1 (ICAM-1) in the generation of an immuneresponse. Eur. J. Immunol. 18: 35.51. Detmers, P.A., S.D. Wright, E. Olsen, B. Kimball, Z.A. Cohn. 1987.Aggregation of complement receptors on human neutrophils in the absence ofligand. J. Cell Biol. 105: 1137.52. Kupfer, A. and S.J. Singer. 1989. The specific interaction of helper T cells andantigen presenting B cells. IV: Membrane and cytoskeletal reorganization in thebound T cells as a function of antigen dose. J. Exp. Med. 170: 1697.53. Kupfer, A. and S.J. Singer. 1989. Cell biology of cytotoxic and helper T cellfunctions: immunofluorescence microscopic studies of single cells and couples.Ann Rev. Immunol. 7: 309.54.^Burn, P., A. Kupfer, S.J. Singer. 1988. Dynamic membrane-cytoskeletalinteractions: specific association of integrin and talin arises in vivo after phorbol132ester treatment of peripheral blood lymphocytes. P.N.A.S. USA 85: 479.55. Kupfer, A., P. Burn, S.J. Singer. 1990. The PMA-induced specific associationof LFA-1 and talin in intact cloned T helper cells. J. Mol. Cell. Immunol. 4: 317.56. Kishimoto, T.K., R.S. Larson, A.L. Corbi, M.L. Dustin, D.E. Staunton, T.A.Springer. 1989. The leukocyte integrins. Adv. Immunol. 46: 149.57. Burridge, K., K. Fath, T. Kelly, G. Nuckolls, C. Turner. 1988. Focal adhesions:transmembrane junctions between the extracellular matrix and the cytoskeleton.Ann. Rev. Cell Biol. 4: 487.58. Geiger, B. 1979. A 130K protein from chicken gizzard: it's localization at thetermini of microfilament bundles in cultured chicken cells. Cell 18: 193.59. Bennet, J.P., K.S. Zaner, T.P. Stossel. 1984. Isolation and some properties ofmacrophage a-actinin: evidence that it is not an actin gelling protein. Biochemistry23: 5081.60. Burridge, K. and P. Mangeat. 1984. An interaction between vinculin and talin.Nature 308: 744.61. Otto, J.J. 1983. Detection of vinculin-binding proteins with an 125I-vinculin geloverlay technique. J. Cell Biol. 97: 1283.62. Belkin, A.M. and V.E. Koteliansky. 1987. Interaction of iodinated vinculin,meta-vinculin and a-actinin with cytoskeletal proteins. F.E.B.S. Lett. 220: 291.63. Horowitz, A., K. Duggan, C. Buck, M.C. Beckerle, K. Burridge. 1986.Interactions of plasma membrane fibronectin receptor with talin-a transmembranelinkage. Nature 320: 531.64. Litchfield, D.W. and E.M. Ball. 1986. Phosphorylation of the cytoskeletal proteintalin by protein kinase C. Biochem. Biophys. Res. Comm. 134: 1276.65. Werth, D.K., J.E. Niedel, I. Pastan. 1983. Vinculin, a cytoskeletal substrate forprotein kinase. J. Biol. Chem. 258: 11423.66. van Agthoven, A., M. Pierres, C. Goridis. 1985. Identification of a previouslyunrecognized polypeptide associated with lymphocyte function associated antigenone (LFA-1). Mol. Immunol. 22: 1349.67. Beckerle, M.C. 1986. Identification of a new protein localized at the sites of cellsubstrate adhesion. J. Cell Biol. 103: 1679.68. Senecal, J.-L., S. Fortin, A. Roussin, F. Joyal. 1987. Anticytoskeletalautoantibody to microfilament anchorage sites recognizes novel contact proteins. J.Clin. Invest. 80: 778.69.^Wheelock, M.J., C.A. Buck, K.B. Bechtol, C.H. Damsky. 1987. Soluble 80-kdfragement of cell-CAM 120/80 disrupts cell-cell adhesion. J. Cell. Biochem. 34:187.13313470. Damle, N.K., and A. Aruffo. 1991. Vascular cell adhesion molecule 1 induces Tcell antigen receptor-dependant activation of CD4+ T lymphocytes. P.N.A.S.USA 88: 6403.71. Lobb, R.R., G. Chi-Rosso, D.R. Leone, M.D. Rosa, S. Bixler, B.M.Newman, S. Luhowskyj, C.D. Benjamin, I.G. Dougas, S.E. Goelz, C. Hession,E.P. Chow. 1991. Expression and functional characterization of a soluble form ofendothelial-leukocyte adhesion molecule. J. Immunol. 147: 124.72. van Seventer, G.A., W. Newman, Y. Shimizu, T.B. Nutman, Y. Tanaka, K.J.Horgan, T.V. Gopal, E. Ennis, D. O'Sullivan, H. Grey, S. Shaw. 1991.Analysis of T cell stimulation by superantigen plus MI-IC complex class IImolecules or by CD3 monoclonal antibody: costimulation by purified adhesionligands VCAM-1, ICAM-1 but not ELAM-1. J. Exp. Med. 174: 901.73. Dana, N., D.M. Fathallah, M.A. Arnaout. 1991. Expression of a soluble andfunctional form of the human 132 integrin CD11b/CD18. P.N.A.S. USA 88: 3106.74. Marlin, S.D., D.E. Staunton, T.A. Springer, C. Stratowa, W. Sommergruber,V.J. Merluzzi. 1990. A soluble form of intercellular adhesion molecule-1 inhibitsrhinovirus infection. Nature 344: 70.75. Kuhlman, P., V.T. Moy, B.A. Lollo, A.A. Brian. 1991. The accessory functionof murine intercellular adhesion molecule-1 in T lymphocyte activation.Contributions of adhesion and co-activation. J. Immunol. 146: 1773.76. Seth, R., D. Raymond, M.W. Makgoba. 1991. Circulating ICAM-1 isoforms:diagnostic prospects for inflammatory and immune disorders. Lancet 338: 83.77. Rothlein, R., E.A. Mainolfi, M. Czajkowski, S.D. Marlin. 1991. A form ofcirculating ICAM-1 in human serum. J. Immunol. 147: 3788.78. Becker, J.C., R. Dummer, A.A. Hartmann, G. Burg, R.E. Schmidt. 1991.Shedding of ICAM-1 from human melanoma cell lines induced by IFN-y and TNFa: functional consequences on cell-mediated cytotoxicity. J. Immunol. 147: 4398.79. Kishimoto, T.K., M.A. Jutila, E.L. Berg, E.C. Butcher. 1989. Neutrophil Mac-1and MEL-14 adhesion proteins inversely regulated by chemotactic factors. Science245: 1238.80. Berg, M., S.P. James. 1990. Human neutrophils release the Leu-8 lymph nodehoming receptor during cell activation. Blood 76: 2381.81. Schmid, K., M.A. Hediger, R. Brossmer, J.H. Collins, H. Haupt, T. Marti, G.D.Offner, J. Schaller, K. Takagaki, M.T. Walsh, H.G. Schwick, F.S. Rosen, E.Remold-O'Donnell. 1992. Amino acid sequence of human plasmagalactoglycoprotein: identity with the extracellular region of CD43 (sialophorin).P.N.A.S. 89: 663.82.^Takei, F., G. Galfre, T. Alderson, E.S. Lennox, C. Milstein. 1980. H9/25monoclonal antibody recognizes a new allospecificity of mouse lymphocytesubpopulations: strain and tissue distribution. Eur. J. Immunol. 10: 241.13583. Kohler, G., S.C. Howe, C. Milstein. 1976. Fusion between immunoglobulinsecreting and nonsecreting myeloma cell lines. Eur. J. Immunol. 6: 292.84. Kim, K.J., C. Kanellopoulos-Languin, R.M. Merwin, D.H. Sachs, R. Asofsky.1979. Establishment and characterization of BALB/c lymphoma lines with B cellproperties. J. Immunol. 122: 549.85. Falo, L.D., B. Benacerraf, K.L. Rock. 1986. Phospholipase treatment ofaccessory cells that have been exposed to antigen selectively inhibits antigenspecific Ia-restricted, but not allospecific stimulation of T lymphocytes. P.N.A.S.USA 83: 6994.86. Sanford, K.K., W.R. Earle, G.D. Likely. 1948. The growth in vitro of singleisolated tissue cells. J. Nat. Cancer Inst. 9: 229.87. Takei, F. 1985. Inhibition of mixed lymphocyte response by a rat monoclonalantibody to a novel murine lymphocyte activation antigen (MALA-2). J. Immunol.134: 1403.88. Sarmiento, M., D.P. Dialynas, D.W. Lancki, K.A. Wall, M.I. Lorber, M.R.Loken, F.W. Fitch. 1982. Cloned T lymphocytes and monoclonal antibodies asprobes for cell surface molecules active in T cell-mediated cytolysis. Immunol.Rev. 68: 135.89. Springer, T., G. Galfre, D.S. Secher, C. Milstein. 1978. Monoclonal xenogeneicantibodies to murine cell surface antigens: identification of novel leukocytedifferentiation antigens. Eur. J. Immunol. 8: 539.90. Miyake, K., K.L. Medina, S-I. Hayashi, S. Ono, T. Hamaoka, P.W. Kincade.1990. Monoclonal antibodies to Pgp-1/CD44 block lympho-hemopoiesis in longterm bone marrow cultures. J. Exp. Med. 171: 477.91. Wilde, D.B., P. Marrack, J. Kappler, D.P. Dialynas, F.W. Fitch. 1983.Evidence implicating L3T4 in class II MHC antigen reactivity; monoclonal antibodyGK1.5 (anti-L3T4) blocks class II MHC antigen-specific proliferation, release oflymphokines, and binding by cloned murine helper T lymphocyte lines. J.Immunol. 131: 2178.92. Takei, F. 1983. Two surface antigens expressed on proliferating mouse Tlymphocytes defined by rat monoclonal antibodies. J. Immunol. 130: 2794.93. Julias, M.H., E. Simpson, L.A. Herzenberg. 1973. A rapid method for theisolation of functional thymus-derived murine lymphocytes. Eur. J. Immunol. 3:645.94. Hunt, S.V. 1979. Separation of lymphocyte subpopulations. In "Handbook ofexperimental immunology", D.M. Weir, ed., Blackwell Scientific, Oxford, Ch. 24.95. Horley, K.J., C. Carpinito, B. Baker, F. Takei. 1989. Molecular cloning ofmurine intercellular adhesion molecule (ICAM-1). EMBO J. 8: 2889.96. Karasuyama, H., A. Kudo, F. Melchers. 1990. The proteins encoded by theVpreB and X5 Pre-B cell-specific genes can associate with each other and with laheavy chains. J. Exp. Med. 172: 969.97. Laemmli, U.K. 1970. Clevage of structural proteins during the assembly of thehead of bacteriophage T4. Nature 227: 680.98. Fraker, P.J. and J.C. Speck. 1978. Protein and cell membrane iodinations with asparingly soluble chloramide, 1,3,4,6-tetrachloro-3a,6a-diphenylglycoluril.Biochem. Biophys. Res. Commun. 80: 849.99. Hunter, W.M. and F.C. Greenwood. 1962. Preparation of iodine-131 labelledhuman growth hormone of high specific activity. Nature 194: 495.100. Mosman, T. 1983. Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxicity assays. J. Immunol. Meth. 65: 55.101. Faulstich, H., H. Trischmann, D. Mayer. 1983. Preparation of tetramethyl-rhodaminyl-phalloidin and uptake of the toxin into short term culturedhepatocytes by endocytosis. Exp. Cell. Res. 144: 73.102. Wulf, E., A. Deboben, F.A. Bautz, H. Faulstich, T. Wieland. 1979. Fluorescentphalotoxin, a tool for the visualization of cellular actin. P.N.A.S. USA 76: 4498.103. Alfieri, D.C. 1991. Occupancy of CD11b/CD18 (Mac-1) divalent ion bindingsite(s) induces leukocyte adhesion. J. Immunol. 147: 1891.104. Kirchhofer, D., J. Gailit, E. Ruoslahti, J. Grzesiak, M.D. Pierschbacher. 1990.Cation-dependant changes in the bimding specificity of the platelet receptorGPIIb/IIIa. J. Biol. Chem. 265: 18525.105. Dransfield, I., C. Cabanas, A. Craig, N. Hogg. 1992. Divalent cation regulationof the function of the leukocyte integrin LFA-1. J. Cell Biol. 116: 219.106. Kurzinger, K., T. Reynolds, R.N. Germain, D. Davignon, E. Martz, T.A.Springer. 1981. A novel lymphocyte function-associated antigen (LFA-1): cellulardistribution, quantitative expression, and structure. J. Immunol. 127: 596.107. Axiline, S.G. and E.P. Reaven. 1974. Inhibition of phagocytosis and plasmamembrane motility of the cultured macrophage by cytochalasin B: role ofsubplasmalemmal microfilaments. J. Cell. Biol. 62: 647.108. Geppert, T.D., and P.E. Lipsky. 1991. Association of various cell-surfacemolecules with the cytoskeleton: effect of cross-linking and activation. J.Immunol. 146: 3298.109. Dougherty, G.J., P.M. Lansdorp, D.L. Cooper, R.K. Humphries. 1991.Molecular cloning of CD44R1 and CD44R2, two novel isoforms of human CD44lymphocyte "homing" receptor. J. Exp. Med. 174: 1.110. Lucas, M.G., A.M. Green, M.J. Telen. 1989. Characterization of serum In(Lu)related antigen: identification of a serum protein related to erythrocyte p80. Blood73: 596.136111. Haynes, B.F., M.J. Telen, L.P. Hale, S.M. Denning. 1989. CD44-a moleculeinvolved in leukocyte adherence and T-cell activation. Immunol. Today 10: 423.112. Mason, D.W. and A.F. Williams. 1980. The kinetics of antibody binding tomembrane antigens in solution and at the cell membrane. Biochem. J. 187: 1.113. Hood, L.E., I.L. Weissman, W.B. Wood, J.H. Wilson. 1984. In"Immunology", The Benjamin/Cummings Publishing Company Inc., Melino Park,California, p60.114. Diamond, M.S., D.E. Staunton, S.D. Marlin, T.A. Springer. 1991. Binding ofintegrin Mac-1 (CD11b/CD18) to the third immunoglobulin-like domain of ICAM-1(CD54) and its regulation by glycosylation. Cell 65: 691.115. Hibbs, M.L., H. Xu, S.A. Stacker, T.A. Springer. 1991. Regulation of adhesionto ICAM-1 by the cytoplasmic domain of LFA-1 integrin [3 subunit. Science 251:1611.116. Hermanowski-Vosatka, A., J.A.G. Van Strijp, W.J. Swiggard, S.D. Wright.1992. Integrin modulating factor-1: a lipid that alters the function of leukocyteintegrins. Cell 68: 341.117. Altieri, D.C. and T.S. Edgington. 1988. A monolconal antibody reacting withdistinct adhesion molecules defines a transition in the functional state of the receptorCD11b/CD18 (Mac-1). J. Immunol. 141: 2656.118. Baruthio, F., 0. Guillard, J. Arnaud, F. Pierre, R. Zawislak. 1988.Determination of manganese in biological materials by electrothermal atomicabsorption spectrometry. Clin. Chem. 34:227.119. Keen, C.L., B. Lonnerdal, L.S. Hurley. 1984. Biochemistry of the essentialultrutrace elements. In: Biochemistry of the elements, E. Frieden, Ed. Vol. 3,New York: Plenum Press, p89-132.120. Matsuda, A., M. Kimura, M. Kataoka, S. Ohkuma, M. Sato, Y. Itokawa. 1989.Quantifying manganese in lymphocytes to assess manganese nutritional status.Clin. Chem. 35: 1939.121. Milne, D.B., R.S. Sims, N.V.C. Ralston. 1990. Manganese content of thecellular components of blood. Clin. Chem. 36: 450.122. O'Toole, T.E., D. Mandelman, J. Forsyth, S.J. Shattil, E.F. Plow, M.H.Ginsberg. 1991. Modulation of the affinity of integrin^(GPIIIb-IIIa) by thecytoplasmic domain of alb. Science 254: 854.123. Sims, P.J., M.H. Ginsberg, E.F. Plow, S.J. Shattil. 1991. Effect of plateletactivation on the conformation of the plasma membrane glycoprotein IIb-IIIacomplex. J. Biol. Chem. 166: 7345.124. Hermanowski-Vosatka, A., P.A., Detmers, 0. Gotze, S.C. Silverstein, S.D.Wright. 1988. Clustering of ligand on the surface of a particle enhances adhesionto receptor-bearing cells. J. Biol. Chem. 263: 17822.137125. Patel, H.R., and R.A. Miller. 1991. Analysis of protein phosphorylation patternsreveals unanticipated complexity in T lymphocyte activation pathways. J.Immunol. 146: 3332.126. Chatila, T.A., and R.S. Geha. 1988. Phosphorylation of T cell membraneproteins by activators of protein kinase C. J. Immunol. 140: 4308.127. Hibbs, M.L., S. Jakes, S.A. Stacker, R.W. Wallace, T.A. Springer. 1991. Thecytoplasmic domain of the integrin lymphocyte function-associated antigen 10subunit: sites required for binding to intercellular adhesion molecule 1 and thephorbol ester-stimulated phosphorylation site. J. Exp. Med. 174: 1227.128. Solowska, J., J.L. Guan, E.E. Marcantonio, J.E. Trevithick, C.A. Buck, R.O.Hynes. 1989. Expression of normal and mutant avian integrin subunits in rodentcells. J. Cell Biol. 109: 853.129. Marcantonio, E.E., J.L. Guan, J.E. Trevithick, R.O. Hynes. 1990. Mapping ofthe functional determinants of the integrin beta 1 cytoplasmic domain by sitedirected mutagenesis. Cell Regulation 1(8): 579.138


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items