Rafts CD147 Through Novel Modulation of Lipid Selective

The Journal of Immunology

Selective Inhibition of T Cell Activation Via CD147 Through Novel Modulation of Lipid Rafts1

Gu¨nther Stafﬂer,2* Andreas Szekeres,* Gerhard J. Schu¨tz,† Marcus D. Sa¨emann,* Elisabeth Prager,* Maximilian Zeyda,‡ Karel Drbal,§ Gerhard J. Zlabinger,* Thomas M. Stulnig,‡ and Hannes Stockinger3*§

The plasma membrane is compartmentalized into microdomains and the association/dissociation of receptors and signaling molecules with/from these membrane domains is a major principle for regulation of signal transduction. By following the reorganization of microdomains on living cells and performing biochemical studies, we show that Ab targeting of the T cell activation- associated Ag CD147 prevents TCR stimulation-dependent reorganization and clustering of microdomains. Triggering CD147 induces a displacement of the GPI-anchored coreceptors CD48 and CD59 from microdomains in human T lymphocytes. This perturbation of microdomains is accompanied by a selective inhibition of TCR-mediated T cell proliferation. The CD147-inhibited cells secret normal levels of IL-2 but acquire reduced amounts of the IL-2 receptor ␣-chain CD25. These results indicate that negative regulating signals can modulate microdomains and suggest a general mechanism for inhibition of receptor signaling. The Journal of Immunology, 2003, 171: 1707–1714.

here is increasing evidence that the plasma membrane is and Ca2ϩ flux in response to stimulation using CD3 mAb OKT3 organized in microdomains. This compartmentalization of (2). In addition, it was shown that these antifungal agents also T the plasma membrane appears, on the one hand, to be inhibited mAb-induced internalization of the GPI-anchored protein important to separate reactive components (i.e., receptors and as- CD59 (9). Furthermore, it was recently shown that depletion of sociated signal-transducing molecules) in unprimed resting cells, cholesterol using methyl-␤-cyclodextrin induced uncontrolled T and, on the other hand, to cluster and concentrate them when de- cell activation by transient tyrosine phosphorylation of multiple termined for execution of signal transduction and function in ac- proteins, including ␨-associated protein 70 (ZAP-70),4 linker for tivated cells (1, 2). These membrane compartments are relatively activation of T cells (LAT), and phospholipase C␥1 (10). How- resistant to solubilization by many nonionic detergents. They con- ever, not only a change of membrane cholesterol alters signaling tain GPI-anchored proteins, glycosphingolipids, sphingomyelins, and receptor function, but also modification of the saturation index cholesterol, Src family protein tyrosine-kinases and G proteins (3, of the fatty acids in the membrane. Shifting the index toward un- 4) and are, therefore, also called GPI microdomains (1), glyco- saturation by feeding cells with polyunsaturated fatty acids re- sphingolipid-cholesterol rafts, detergent-insoluble glycolipid-en- sulted in an inhibition of signal transduction, which is accompa- riched domains (5) or glycosphingolipid-enriched membrane do- nied by a specific displacement of intracellular signaling molecules mains (6). from microdomains (11). The primary components responsible for the integrity of GPI With the exception of fractions of some transmembrane proteins microdomains have been shown to be cholesterol and saturated including integrins (12), CD4 and CD8 (13, 14), most transmem- fatty acids mainly associated with sphingolipids, GPI proteins and brane proteins are excluded in the resting state of the cell from GPI Src kinases (7). Consequently, also these lipids play a major role microdomains (1). Upon receptor ligation/cell activation, however, to enable signaling. Reduction of the cholesterol content inhibited several recent studies show association of the TCR, Fc⑀R1, or signaling via GPI-anchored proteins (8). Disruption of cholesterol- ␣ rich membrane domains in T cells by polyene antifungal agents Fc R with GPI microdomains and subsequent signal transduction decreased tyrosine phosphorylation of CD3-␨, phospholipase C␥1, (2, 15Ð17). The mechanisms ruling over association/dissociation of transmembrane receptors with microdomains are not clear yet. However, CD28 engagement induces redistribution and polariza- *Institute of Immunology, University of Vienna, Vienna, Austria; †Institute for Bio- tion of GPI microdomains into caps at the site of TCR engagement. physics, University of Linz, Linz, Austria; and ‡Department of Internal Medicine III, Thus, CD28 appears to promote association of TCR with microdo- University of Vienna, Vienna, Austria; and ¤Competence Center for Biomolecular Therapeutics, Vienna, Austria. mains (18). The existence of regulators, which amplify association Received for publication November 6, 2002. Accepted for publication June 17, 2003. of molecules with GPI microdomains, implies that there might also exist receptors that trigger dissociation of molecules from mi- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance crodomains, resulting in inhibition/deactivation of signal transduc- with 18 U.S.C. Section 1734 solely to indicate this fact. tion. Searching for molecules that might execute negative regulat- 1 This work was supported by the Competence Center for Biomolecular Therapeutics, ing signals via modification of GPI microdomains, we found by the GEN-AV program of the Austrian Federal Ministry of Education, Science and Culture, and by grants SFB00503, P15025-B08, and P13507-B01 from the Austrian CD147 to be a potential candidate. Science Fund. 2 Current address: Intercell, Campus Vienna Biocenter 6, A-1030 Vienna, Austria. 3 Address correspondence and reprint requests to Dr. Hannes Stockinger, Institute of Immunology, University of Vienna, Brunner Strasse 59, A-1235 Vienna, Austria. 4 Abbreviations used in this paper: ZAP-70, ␨-associated protein 70; LAT, linker for E-mail address: [email protected] activation of T cells; PR, phenol red.

CD147, also known as M6 Ag (19), extracellular matrix metal- the distribution of the fluorescence intensity on the plasma membranes of loproteinase inducer (20), basigin or neurothelin (21, 22), is a 50- MEM-102-Cy5-stained cells using SPOT Advanced Software (Diagnostics to 60-kDa type I transmembrane glycoprotein belonging to the Ig Instruments, Sterling Heights, MI). Cells displaying a 3-fold increase of fluorescence intensity on one site of the plasma membrane were counted as superfamily (19). CD147 is widely expressed on hemopoietic and cells with caps. Three hundred cells per sample were evaluated and the nonhemopoietic cells. It is strongly up-regulated on T cells upon experiment was repeated once. activation, indicating a function in T cell biology (19, 23). Indi- Analysis of protein tyrosine phosphorylation cation of a potential negative regulatory function of CD147 in T cell regulation was shown by enhanced mixed lymphocyte re- Purified human T cells were rested overnight in complete RPMI 1640 sponses of lymphocytes from CD147 knockout mice (24) and by a medium. Before stimulation, T cells were incubated in RPMI 1640 medium supplemented with 1% FCS for4hat37¡C. Cells (3 ϫ 106) were stim- mAb that inhibited T cell proliferation (23). In this study, we show ulated for 5 min at 37¡C using CD3 mAb OKT3 (10 ␮g/ml) along with that triggering the CD147 molecule on T cells by this inhibitory CD28 mAb Leu-28 (2 ␮g/ml) in the presence or absence of CD147 mAbs, mAb results in a characteristic modulation of microdomains, or they were left unstimulated. The reaction was stopped by addition of which is associated with impaired signaling for expression of the ice-cold washing buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 5 ϫ IL-2R ␣-chain CD25. mM EDTA). After centrifugation (2 min, 850 g,4¡C), cells were immediately lysed for 30 min in ice-cold TBS (20 mM Tris-HCl (pH 7.5) and 150 mM NaCl) containing 1% Nonidet P-40 (Pierce, Rockford, IL), 1 mM Materials and Methods sodium orthovanadate, 20 mM NaF, 5 mM EDTA, and protease inhibitors Cell preparation (5 mM aprotinin, 5 mM leupeptin, 1 mM PMSF, and 1 ␮M pepstatin; all from Sigma-Aldrich). After centrifugation (5 min, 14,000 ϫ g,4¡C), ly- Peripheral blood was taken from healthy donors and PBMCs were obtained sates were analyzed by nonreducing SDS-PAGE (12% gel) and immuno- by density gradient centrifugation using Ficoll-Hypaque (Pharmacia, Upp- blotting using HRP-labeled anti-phosphotyrosine Ab 4G10 (1/4000; Up- sala, Sweden). T cells were enriched by removing adherent cells on nylon state Biotechnology, Lake Placid, NY). wool (Robbins Scientific, Sunnyvale, CA) and purified by negative depletion of CD14-, CD16-, CD20-, and CD56-positive cells using the respec- Multiprobe RNase protection assay tive specific mAbs (CD14 mAb MEM-18, CD16 mAb MEM-154, CD20 Purified human T cells were rested overnight in RPMI 1640 medium sup- mAb MEM97, CD56 mAb MEM-188, all of which were produced by Dr. plemented with 10% FCS. Cells (4 ϫ 107) were stimulated by plate-bound I. Hilgert and Dr. V. Horejsi, Academy of Sciences of the Czech Republic, OKT3 (immobilized at 0.1 ␮g/ml) along with soluble Leu-28 (0.5 ␮g/ml) Prague, Czech Republic) and the MACS technique of Miltenyi Biotec alone or in the presence of soluble CD147 mAbs (1 ␮g/ml). After a 12-h (Bergisch Gladbach, Germany). Afterward, the purified T cells were incubation in a humidified atmosphere at 37¡C, T cells were harvested. washed and resuspended in complete RPMI 1640 medium supplemented Subsequently, total RNA was isolated using TRIreagent (Sigma-Aldrich) with 10% heat-inactivated FCS (PAA, Linz, Austria), 2 mM L-glutamine, according to the manufacturer’s instructions and used in the standard 100 U/ml penicillin, 100 ␮g/ml streptomycin, and 18.75 ␮g/ml gentamicin RNase protection assay (BD PharMingen, San Diego, CA). In brief, the sulfate. template sets hCK-1 and hCR-1 were used to synthesize [␣-32P]UTP (Am- Mixed lymphocyte culture (MLC) ersham Biosciences, Buckinghamshire, U.K.) containing riboprobes (in vitro transcription kit; BD PharMingen) that were hybridized with 4.0 ␮g Responder PBMCs (4 ϫ 104) were incubated with 4 ϫ 104 irradiated (6000 total RNA and treated with a mixture containing RNase AϩT1. Samples rad) PBMCs from an unrelated donor in complete RPMI 1640 medium in were denatured and loaded onto 0.4-mm, 6% polyacrylamide/urea gels. the presence of various concentrations of the CD147 mAbs. Cultures were After electrophoresis, gels were transferred to filter papers, dried under set up in round-bottom 96-well plates (Costar, Cambridge, MA) at 37¡Cin vacuum, and analyzed by a PhosphorImager (Molecular Dynamics, Sunny- a5%CO2 atmosphere for 4 days. Assays were performed in triplicates. vale, CA). Each band was normalized by comparison of its intensity value Sixteen hours before harvesting, the cultures were pulsed with 1 ␮Ci to the value of the housekeeping gene L32. The normalized values were [3H]thymidine (NEN, Boston, MA). Incorporated radioactivity was mea- used to quantify expression of the individual mRNAs from CD147 mAb- sured by liquid scintillation counting. treated cells as percent change in comparison to CD3/CD28-stimulated cells without CD147 mAbs. T cell proliferation assay Immunofluorescence analysis of CD25 Proliferation assays of highly purified T lymphocytes (1 ϫ 105 cells/well) were performed in triplicates in 96-well flat-bottom tissue culture plates Highly purified T cells were activated as described before (see T cell pro- (Nunc, Roskilde, Denmark) in a final volume of 200 ␮l. Proliferation was liferation assay) in the presence or absence of soluble CD147 mAbs. Anal- induced by CD3 mAb OKT3 in combination with the CD28 mAb Leu-28 ysis of cell surface expression of CD25 was performed 24 h after culture (BD Biosciences, San Jose, CA). CD3 mAb OKT3 was immobilized to the initiation. For this purpose, T cells were incubated with a directly labeled surface of the well. For this purpose, plates were incubated overnight at CD25 mAb (10 ␮g/ml, FITC-labeled CD25; BD Biosciences) for 30 min 4¡C with 100 ␮l of an OKT3 (0.05 ␮g/ml in PBS) solution. After washing on ice in PBN (PBS, containing 1% BSA and 0.02% NaN3) and afterward the precoated plates twice, we added T cells, which were suspended in washed twice with PBN. To prevent nonspecific binding of the mAb to Fc complete RPMI 1640 medium containing Leu-28 (1 ␮g/ml) and different receptors, the cells were preincubated for 30 min with 4 ␮g/ml human Ig concentrations of CD147 mAbs. Alternatively, T cells were stimulated on ice. Fluorescence was analyzed on a FACScan flow cytometer (BD with PMA (10Ϫ7 M; Sigma-Aldrich, St. Louis, MO) plus ionomycin (1 Biosciences). DNA staining using ethidium bromide excluded dead cells mM; Sigma-Aldrich), or PMA in combination with immobilized OKT3 or from the analysis. Leu-28. For cytokine supplementation experiments, recombinant human Cytokine assay IL-2 (kindly provided by the Novartis Research Institute, Vienna, Austria) was added as indicated. Proliferation was assessed as described above. Highly purified T cells were stimulated using mAb OKT3 along with mAb Leu-28 (as described in the T cell proliferation assays) in the presence or Capping absence of soluble CD147 mAbs. Alternatively, as a control for inhibition of IL-2 secretion, CD3/CD28-stimulated T cells were incubated with cy- Purified T lymphocytes were stored overnight in complete RPMI 1640 ␮ medium without phenol red (1640 without PR). Afterward T cells were closporin A (1 g/ml; Sigma-Aldrich). Cell culture supernatants were har- washed once, resuspended at 1 ϫ 106 cells per 100 ␮l of 1640 without PR, vested 24 h following activation. IL-2 secretion was assayed by a sandwich and incubated on ice for 30 min with a Cy5-labeled Fab of mAb MEM-102 ELISA using a commercial ELISA kit (R&D Systems, Minneapolis, MN). (10 ␮g/ml), specific for the GPI-anchored molecule CD48. After washing Analysis of glycosphingolipid-enriched membrane domain with ice-cold 1640 without PR, T cells were resuspended at 1 ϫ 106 per 100 ␮l of 1640 without PR. To induce capping, T cells were incubated with Freshly isolated highly purified T lymphocytes (1 ϫ 107 cells/ml) were 10 ␮g/ml OKT3 alone or in combination with 1 ␮g/ml Leu-28 for 30 min stimulated for 20 min at 37¡C in complete RPMI 1640 medium (10 ml) at 37¡C in the presence or absence of 1 ␮g/ml of the indicated CD147 using CD3 mAb OKT3 (5 ␮g/ml) alone or in combination with 5 ␮g/ml of mAbs. Subsequently, cells were plated onto glass microscope cover slips different CD147 mAbs, or they were left unstimulated. Alternatively, un- and examined on a Carl Zeiss microscope (Zeiss, Oberkochen, Germany). stimulated T cells were treated for 20 min at 37¡C with 5 ␮g/ml of the The percentage of cells, which showed a cap, was assessed by analyzing indicated CD147 mAbs alone. The reaction was stopped by adding a 3-fold The Journal of Immunology 1709 volume of ice-cold washing buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 5 mM EDTA). After a short centrifugation (2 min, 850 ϫ g, 4¡C), the cells were immediately lysed for 30 min on ice in 0.25 ml of lysis buffer consisting of TBS (20 mM Tris-HCl (pH 7.5) and 150 mM NaCl), 1% Brij-58 (Pierce), 2 mM EDTA, and protease inhibitors (5 mM aprotinin, 5 mM leupeptin, 1 mM PMSF, and 1 mM pepstatin; all from Sigma- Aldrich). After centrifugation (30 s, 14,000 ϫ g,4¡C), the cell lysates were adjusted to 40% (w/v) sucrose by adding an equal volume of a 80% sucrose solution (TBS containing 80% w/v sucrose, 2 mM EDTA, and protease inhibitors). These preparations were placed on top of a 60% sucrose layer in a centrifuge tube (Sorvall Instruments-DuPont, Wilmington, DE). On the top of this, layers of 20, 10, and 5% sucrose were placed. After ultracen- trifugation (180,000 ϫ g,16h,4¡C), 375-␮l fractions were collected from the top. Aliquots of each fraction were diluted in a 4ϫ gel loading SDS buffer. Proteins were separated by nonreducing SDS-PAGE and subsequently transferred to a polyvinylidene difluoride membrane (Immo- bilon-P; Millipore, Bedford, MA). After a 1-h incubation in blocking buffer (20 mM Tris-HCl (pH 7.5), 0.15 mM NaCl, and 3% BSA), membranes were probed with the respective mAbs (mouse anti-human mAbs: lck-01 to p56lck, TS2/18 to CD2, MEM-102 to CD48, MEM-43/5 to CD59, MEM- M6/1 to CD147 (all produced by Dr. I. Hilgert and Dr. V. Horejsi), rabbit mAbs SC-1239 to CD3␨ (kindly provided by Dr. L. E. Samelson, National Institute of Child Health and Human Development, Bethesda, MD), anti- ZAP-70 mAb (Transduction Laboratories, Lexington, KY), and anti-LAT mAb (Upstate Biotechnology) washed in washing buffer (20 mM Tris-HCl (pH 7.5), 0.15 mM NaCl, and 0.1% Tween 20) and incubated for 1 h with HRP-conjugated secondary Abs: Either goat anti-mouse IgG (Sigma- Aldrich) or goat anti-rabbit IgG (Bio-Rad, Richmond, CA). Proteins were visualized using the ECL detection system (Roche Applied Science, Mann- heim, Germany) and a Lumi-Imager (Roche Applied Science).

Results FIGURE 1. Influence of CD147 mAb MEM-M6/6 on T cell prolif- Down-regulation of T cell alloresponsiveness by CD147 mAb eration. A, PBMCs were mixed with irradiated PBMCs from an unre- MEM-M6/6 lated donor in the presence or absence of different concentrations of the CD147 mAbs MEM-M6/6 or MEM-M6/1. Proliferation was determined Among a panel of 13 CD147 mAbs, we found one, termed MEM- on day 4 following stimulation as described in Materials and Methods. M6/6, which inhibited CD3 mAb OKT3 activation of human T This figure shows the [methyl-3H]TdR uptake in cpm (mean Ϯ SD) of cells enriched from peripheral blood. Epitope mapping analysis four independent experiments using cells from different donors. B, revealed that MEM-M6/6 is a mAb directed to a unique epitope Highly purified T cells were stimulated with a CD3 mAb and a CD28 located in the membrane proximal Ig-like domain of CD147 (23). mAb (OKT3 and Leu-28, respectively) in the presence of various con- To examine the inhibitory capacity of MEM-M6/6 on T cells in centrations of soluble CD147 mAbs MEM-M6/6 and MEM-M6/1 or with medium alone. Proliferation was determined 72 h after starting the more detail, we first analyzed the response of T lymphocytes in the culture as described in A. This figure shows the [methyl-3H]TdR uptake presence of MEM-M6/6 in a MLC. As shown in Fig. 1A, MEM- in cpm (mean Ϯ SD) of five independent experiments using cells from M6/6 inhibited the T cell proliferative response in MLCs up to different donors. C, T lymphocytes were stimulated with PMA alone, 90%, whereas all other CD147 mAbs did not effect T cell prolif- PMA plus ionomycin (Iono), PMA plus immobilized CD3 mAb OKT3, eration. MEM-M6/1 representing these nonfunctional CD147 or with PMA plus CD28 mAb Leu-28. In all experiments, cells were mAbs served as isotype-matched control in this study. cultured with two different concentrations (1 or 0.3 ␮g/ml) of soluble CD147 mAbs MEM-M6/1 and MEM-M6/6 or with medium alone. Pro- Down-regulation of TCR/CD3 plus CD28-induced proliferation liferation was determined as described in A. This figure shows the of highly purified T lymphocytes by CD147 mAb MEM-M6/6 [methyl-3H]TdR uptake in cpm (mean Ϯ SD) of three independent ex- To analyze whether the inhibitory potential of mAb MEM-M6/6 periments using cells from different donors. relies upon interaction with the CD147 molecule on T cells or on accessory cells, we highly purified peripheral blood T lymphocytes and challenged them with CD3 mAb OKT3 plus CD28 mAb Leu-28 in the presence of MEM-M6/6. The high purity of the T CD147 mAb MEM-M6/6 does not inhibit proliferation of T lymphocytes was underlined by their unresponsiveness when both lymphocytes upon bypassing early signaling cascades stimulatory mAbs were provided in soluble form (25). Prolifera- Next, we analyzed the inhibitory effect of MEM-M6/6 when PMA tion of these highly purified T lymphocytes was obtained only ϩ and the Ca2 ionophore ionomycin were used for stimulation. As when both mAbs were present and at least OKT3 was immobi- shown in Fig. 1C, MEM-M6/6 could not inhibit proliferation of lized. Fig. 1B shows that soluble MEM-M6/6 also significantly decreased proliferation (up to 70Ð80%) of these T cell prepara- these T cell cultures. Since T cells could also be stimulated by tions in contrast to control mAb MEM-M6/1.These results indicate cross-linking CD3 or CD28 in the presence of PMA, we analyzed that MEM-M6/6 exerts its function upon interaction with the whether MEM-M6/6 affects T cell activation induced by these CD147 molecule expressed on T cells. Maximal down-regulation stimulants. None of these stimulations were blocked by MEM- was seen with concentrations between 1 ␮g/ml and 0.1 ␮g/ml of M6/6 (Fig. 1C). Inasmuch as PMA alone induces only a weak the mAb. Increasing or decreasing the concentration of MEM- proliferation in this T cell preparation, these findings show that M6/6 gradually abrogated the inhibitory affect (the same was certain signals can be transmitted via CD3 as well as CD28 in the found in the MLC). This is in line with earlier data (23) and in- presence of CD147 mAb MEM-M6/6. Furthermore, the experi- dicates that an appropriate Ag:Ab ratio is necessary to mediate the ment indicates that CD147 targeting disturbs early T cell signaling inhibitory effect. events, which can be bypassed by PMA. 1710 CD147-INDUCED MODULATION OF T CELL ACTIVATION

CD147 mAb MEM-M6/6 inhibits redistribution and clustering of assays on purified human T cells. We analyzed mRNA expression membrane microdomains upon TCR/CD3 triggering of different cytokines as well as cytokine receptor subunits. As A prerequisite for efficient induction of membrane proximal sig- shown in Fig. 3, the expression of IL-2 mRNA as well as the ␣ naling events via TCR is the redistribution and clustering of mem- IL-2R mRNA was strongly induced upon T cell stimulation. In ␥ brane microdomains to the site of TCR engagement (18, 26, 27). parallel, also the expression of mRNAs coding for IFN- , IL-9, ␤ ␥ ␣ Therefore, we studied whether CD147-induced negative regulation IL-15, IL-10, IL-4, IL-5, IL-2R , common -chain, IL-4R , and ␣ of T cell proliferation is associated with an alteration of membrane IL-15R were induced or enhanced. In contrast, IL-7R mRNA, microdomain reorganization. For this purpose, highly purified T which seems to be expressed strongly on unprimed T cells, is lymphocytes were stained using Cy5-labeled mAb MEM-102, down-regulated upon stimulation (28, 29). which specifically recognizes the GPI-anchored coreceptor CD48. T cells stimulated in the presence of MEM-M6/6 showed a GPI-anchored molecules are highly associated with microdomains unique cytokine and cytokine receptor pattern. Expression of the ␣ and, thus, are regarded as markers for lipid rafts. IL-2R chain was reduced by 66% compared with the control First, we analyzed the distribution of the microdomain-localized (Table I). Furthermore, the mAb moderately inhibited expression ␣ CD48 molecules in unstimulated as well as TCR/CD3-stimulated of IL-4R -chain (27% reduction) and inhibited down-regulation ␣ T lymphocytes. Unstimulated cells showed a diffuse or ring pattern of IL-7R -chain, but did not influence the expression of other of plasma membrane staining (Fig. 2; unstim), whereas stimulation cytokine receptors (Fig. 3B and Table I). Concerning cytokines, ␥ of T lymphocytes resulted in clustering of CD48 molecules to IFN- and IL-15 mRNAs were blocked by 40%, those of IL-9, Ͼ polarized surface caps in ϳ60% of cells (Fig. 2; stim). IL-10, IL-4, and IL-5 by 70%. MEM-M6/6 had no effect, how- Then, we analyzed a possible influence of MEM-M6/6 on TCR/ ever, on expression of IL-2 (Fig. 3A and Table I). In summary, CD3 induced cap formation of CD48. As can be seen in Fig. 2, the MEM-M6 treatment blocks expressions of cytokines with the ex- number of cells showing a cap after stimulation dropped from ϳ60 ception of IL-2 but has little influence on cytokine receptor ex- ␣ to ϳ20% when T cells were stimulated in the presence of MEM- pression except for IL-2R . ϩ M6/6 (stim MEM-M6/6), but not when the cells were incubated Targeting the CD147 molecule by mAb MEM-M6/6 does not with the nonfunctional CD147 control mAb MEM-M6/1 (Fig. 2; inhibit secretion of IL-2 stim ϩ MEM-M6/1). This result shows that CD147 mAb MEM- M6/6 disturbs the reorganization of GPI-anchored CD48 mole- Next, we analyzed whether CD147 mAb MEM-M6/6-induced cules upon T cell activation. modulation of cytokine and cytokine receptor expression is not only seen at the mRNA but also at the protein level. We concen- Effect of CD147 mAb MEM-M6/6 on transcription of cytokines trated on IL-2, which is the most important growth factor for T and cytokine receptors cells and the IL-2 receptor ␣-chain CD25, which is required for To get a hint of which signaling pathways are affected by CD147- action of IL-2. As can be seen in Fig. 4A, surface expression of mediated membrane alteration, we performed RNase protection CD25 was inhibited up to 80% by MEM-M6/6. In contrast, secretion of IL-2 was not significantly diminished by MEM-M6/6 (Fig. 4B), although the proliferation of the T cells in these cultures was reduced up to 60Ð70% compared with the controls (cf Fig. 1B). IL-2 secretion was also not influenced by the nonfunctional control mAb MEM-M6/1, but it was inhibited when we treated cells of the

FIGURE 3. Effect of CD147 mAb MEM-M6/6 on the expression of cytokines and cytokine receptors. Total RNA was isolated from unstimulated as well as from stimulated human T lymphocytes treated with or FIGURE 2. Inhibition of TCR/CD3-mediated redistribution of the without CD147 mAbs. Samples were analyzed by RNase protection assays CD48 molecule by mAb MEM-M6/6. Puriﬁed peripheral blood T lympho- using template set hCK-1 (A) and hCR-1b (B) (see Materials and Meth- cytes were stained with Cy5-labeled mAbs to CD48 (MEM-102) and ods). Expression of housekeeping genes L32 and GAPDH are shown as a washed. Afterward CD3 mAb OKT3 was added for 30 min at 37¡Cinthe measure of equal RNA loading. This experiment is representative of three presence (stim ϩ MEM-M6/6 or stim ϩ MEM-M6/1) or absence (stim) of independent experiments. Lane 1, Unstimulated T cells; lane 2, OKT3/ CD147 mAb MEM-M6/6 or MEM-M6/1. Cells were also incubated at Leu-28-stimulated T cells; lane 3, stimulated T cells with MEM-M6/1; and 37¡C without addition of CD3 or CD147 mAbs (unstim). lane 4, stimulated T cells with MEM-M6/6. The Journal of Immunology 1711

Table I. Quantiﬁcation of bands in Fig. 3 by PhosphorImagera

Quantiﬁcation of bands in Fig. 3A

Lane L32 IFN-␥ IL-2 IL-9 IL-15 IL-10 IL-4 IL-5

1 100 0 3 6 29 4 16 10 2 100 100 100 100 100 100 100 100 3 100 99 101 105 107 113 115 105 4 100 62 109 24 58 18 33 34

Quantiﬁcation of bands in Fig. 3B

␣ ␤␥b ␣ ␣ ␣ Lane L32 IL-2R IL-2R C IL-4R IL-15 IL-7R 1 100 0 62 89 14 37 483 2 100 100 100 100 100 100 100 3 100 92 102 100 102 87 109 4 100 44 88 115 73 94 182

a The values received for each band were normalized to the values of the housekeeping gene L32 and are shown in percent to CD3- and CD28-stimulated cells without CD147 mAbs (equals lane 2 in Fig. 3). b ␥ ␥ C, Common -chain. same preparation with cyclosporin A. Consistent with the impaired tion with MEM-M6/6 did not alter its distribution (Fig. 6B). (18, expression of CD25, addition of exogenous IL-2 did not signifi- 34). Next, we studied the distribution of signaling molecules cantly restore T cell proliferation in MEM-M6/6-treated T cells known to be associated with membrane-proximal TCR signaling (Fig. 4C). These results demonstrate that triggering the CD147 events, such as LAT, p56lck, and ZAP-70 in the gradient fractions. protein by mAb MEM-M6/6 does not lead to a general block of T As shown in Fig. 6B, LAT was found in all fractions but peaked in cell signaling but rather may interfere with particular T cell sig- the low-density fractions 2Ð5. The same pattern was seen in frac- naling pathway(s). tions derived from TCR-stimulated T cells triggered with MEM- M6/6. MEM-M6/6 presence also did not change the distribution CD147 mAb MEM-M6/6 does not affect major protein tyrosine pattern of p56lck, a substantial proportion of which was found in phosphorylation upon TCR/CD3 T cell stimulation the fractions corresponding to microdomains. Although it was re- One of the earliest signaling events that follow TCR engagement ported that ZAP-70 colocalizes with phosphorylated CD3␨ chains is tyrosine phosphorylation of several proteins. As shown in Fig. 5, in microdomains upon T cell triggering (2, 15), we did not detect CD3 and CD28 treatment resulted in a strong tyrosine phosphor- ZAP-70 in the low-density fractions of these cells and CD147 ylation of proteins in a zone of 50 and 60 kDa corresponding to the engagement did not alter its distribution (Fig. 6B). We tested also Src kinases lck and fyn, as well as a 36- to 38-kDa band corre- the distribution of the CD147 molecule. In TCR/CD3-stimulated sponding to LAT. Importantly, we did not see a difference in pro- cells, a small part was found in fraction 5 but the majority was tein tyrosine phosphorylation when T lymphocytes were stimu- contained in the high-density fractions 7 and 8 and the intermedi- lated in the presence or absence of MEM-M6/6. Moreover, MEM- ate fraction 6. Incubation of both CD147 mAbs, MEM-M6/6 and M6/6-treated and untreated T cells exhibited comparable calcium MEM-M6/1, resulted in a concentration of CD147 in the high- fluxes after anti-CD3 cross-linking (data not shown). Thus, the density fractions 7 and 8. major signaling molecules required for expression of IL-2 (30, 31) When analyzing density gradient distribution of the GPI-an- are not influenced by MEM-M6/6. This is in agreement with the chored protein CD48 in OKT3-stimulated T cells, we found a undisturbed expression of IL-2 shown before. strong shift from the low-density fractions toward the higher den- CD147 triggering by mAb MEM-M6/6 modulates T cell sity ones upon MEM-M6/6 treatment (Fig. 6B): CD48 could not be microdomains found any longer in fraction 2 but became visible in the interme- diate fraction 6 and, in particular, in the high-density fractions 7 As shown before by fluorescence microscopy, CD147 mAb MEM- and 8. In comparison, the distribution of CD48 in sucrose fractions M6/6 modulated mAb OKT3 induced cap formation of the GPI- of stimulated T cells incubated with the nonfunctional CD147 anchored coreceptor CD48. Because CD48 is known as marker for microdomains (1, 4, 5, 11, 13, 32), one might suggest that asso- mAb or without any mAb was not altered. This experiment there- ciation of microdomains and TCR is impaired by MEM-M6/6 fore shows that CD147-induced inhibition of T lymphocyte pro- treatment. However, CD147 triggering did not alter expression of liferation is accompanied by a displacement of the GPI-anchored IL-2 that requires phosphorylation of microdomain-associated lck protein CD48 from microdomains. In addition to CD48, a proportion and LAT (30, 31). Thus, these results suggested that CD147 mAb of CD59, another GPI-anchored protein, was reproducibly shifted to MEM-M6/6 did not block association of microdomains and TCR high-density fractions in MEM-M6/6-treated cells (Fig. 6B). in general but rather caused a unique alteration of TCR-associated Next, we analyzed whether the changes in the molecular com- microdomains. To study this, we isolated microdomains by su- position of lipid rafts are induced by MEM-M6/6 treatment alone crose gradient centrifugation (7, 13). According to the enrichment or dependent on TCR/CD3 stimulation. As shown in Fig. 6C,a of CD48 in fractions 2Ð5 (11, 13, 15) and the localization of CD2 comparable shift from low- to high-density fractions of the GPI- mainly in the high-density fractions 7Ð8 as described before (33), anchored molecules could also be observed when T cells were we concluded that the buoyant fractions 2Ð5 of the sucrose gradi- triggered by CD147 mAb MEM-M6/6 without stimulation using ent shown in Fig. 6A contain microdomains. OKT3. Also in this approach, MEM-M6/6 did not alter the distri- Upon stimulation via the TCR/CD3 complex, 40Ð50% of the bution of LAT. The control mAb MEM-M6/1 did not show any CD3␨ chain was detected in the low-density fractions and incuba- effect. 1712 CD147-INDUCED MODULATION OF T CELL ACTIVATION

FIGURE 5. CD147 mAb MEM-M6/6 does not impair protein tyrosine phosphorylation upon TCR triggering. Highly puriﬁed T cells were incubated for 10 min on ice with medium alone or with medium supplemented with CD147 mAb MEM-M6/6 or MEM-M6/1. Afterward, cells were left unstimulated (unstim) or were stimulated (stim) by mAb OKT3 for 5 min on 37¡C. Subsequently, cells were lysed and tyrosine phosphorylated proteins were analyzed using Western blotting. Similar results were obtained with four other experiments.

caps. As a result of this physical contact, raft-resident signaling FIGURE 4. CD147 mAb MEM-M6/6 inhibits surface expression of molecules couple triggered TCRs with more downstream signaling CD25, but does not affect IL-2 secretion. A, T cells were stimulated as pathways, resulting finally in clonal expansion of T cells. A pre- described in Fig. 1B with (stim ϩ MEM-M6/6) or without (stim) CD147 requisite for transmitting signals via lipid rafts is their integrity. As mAb MEM-M6/6 or with control mAb MEM-M6/1 (stim ϩ MEM-M6/1; soon as the molecular organization is disordered (e.g., by fish oils each 0.5 ␮g/ml). Twenty-four hours after activation, cells were harvested or depletion of cholesterol), signal transduction is disturbed, re- and surface expression of CD25 was analyzed by immunostaining using a sulting in blockade of T cell proliferation. directly labeled CD25 mAb and flow cytometry. Thin lines represent his- In this report, we provide evidence that specific triggering of the tograms received with isotype control mAbs. This experiment is represen- T cell activation-associated CD147 molecule by mAb MEM-M6/6 tative of three independent experiments. B, Highly purified T lymphocytes inhibits TCR/CD3-induced reorganization of parts or subtypes of were stimulated as described in Fig. 1B. Supernatants were taken 24 h lipid rafts into dense caps. While leaving expression and secretion following stimulation and the cytokine content was determined by ELISA. The cytokine concentrations in the supernatants are shown in percent to the of IL-2 untouched, this blockade of membrane reorganization is medium control (100% ϭ 1770 pg/ml). Comparable results were obtained accompanied by impaired expression of CD25, resulting in re- in five independent experiments. C, Highly purified T cells were stimulated duced sensitivity to IL-2. MEM-M6/6, which recognizes a unique as described in Fig. 1B in the presence of exogenous rIL-2. Proliferation epitope on the CD147 molecule (23), blocks TCR-induced prolif- was determined 72 h after starting the culture as described in Fig. 1. This eration of T cells. Remarkably, MEM-M6/6 was not able to inter- experiment is representative of five independent experiments. fere with the clonal expansion of T cells when early signaling events associated with rafts were bypassed by PMA. Earlier studies showed that CD147 mAbs inhibited U937-dependent T cell Discussion proliferation (36) and T cell proliferation in a MLC (37). These During the last decade, it became more and more evident that the studies claimed that the mAbs exerted their effects on accessory compartmentalization of the plasma membrane into microdomains cells, because inhibition was seen when U937 cells were prepulsed is one of the main principles to control signaling of cell surface with the mAbs. Indeed, we also have indication that MEM-M6/6 receptors. Today, two types of microdomains are relatively well can influence accessory cell function (data not shown). However, defined: one enriched in tetraspan proteins (35), the second, the the results shown here with highly purified T cells clearly demon- better characterized one, concentrate GPI-proteins, glycosphingo- strate that MEM-M6/6 can directly affect T cell function. Inas- lipids, sphingomyelins, and cholesterol and thus is called glyco- much as proliferation in MLCs of lymphocytes from CD147 sphingolipid-enriched microdomain, GPI microdomain or lipid knockout mice was significantly greater than that of lymphocytes raft. These lipid rafts are enriched on the one hand in GPI-an- from wild-type littermates (24), a negative regulatory role of chored molecules, known to act as costimulatory molecules, and, CD147 is indicated in vivo and suggests that MEM-M6/6 acts on the other hand, in molecules involved in signal transduction rather as an agonist mimicking a negative signal than as an antag- such as LAT, protein tyrosine kinases, and G proteins. In resting T onist inhibiting a positive one. cells, this conglomerate of extracellular accessory and intracellular We observed blockade of the reorganization of the membrane signaling molecules is physically separated from transmembrane through CD147 upon TCR/CD3 triggering by using fluorescence receptors including the TCR/CD3 complex, which is located in the microscopy. CD48 proteins clustered into dense caps upon TCR/ more phospholipid-rich bulk of the cell membrane. Recently, sev- CD3 triggering and this effect was even more outspoken by co- eral publications showed that TCR engagement led to association stimulation with CD28 (data not shown). Remarkably, MEM- of the TCR with lipid rafts and to their redistribution to form dens M6/6 inhibited this reorganization of CD48 (Fig. 2), suggesting an The Journal of Immunology 1713

32, 38), the GPI-anchored proteins CD48 and CD59 were found exclusively in the lipid raft fractions. In contrast, transmembrane molecules such as CD2 and CD147 were mainly excluded from these fractions. As shown before (2, 15, 33, 39), upon TCR triggering the TCR␨ chain could be found up to 50% in the low- density fractions, indicating association of TCR/CD3 with lipid rafts. The inhibitory mAb MEM-M6/6 had no influence on this redistribution as well as on the distribution of the Src kinase lck and the linker proteins LAT. However, MEM-M6/6 induced translocation of CD48 and CD59 from the low-density fractions of TCR/CD3-stimulated as well as of unstimulated T lymphocytes to fractions of higher density. Whether CD48 and CD59 are associ- ating with novel microdomains of higher density or as single molecules with the bulk of the membrane is obscure at the moment. Alternatively, they could be part of a subunit of lipid rafts, which appeared for the first time under this experimental setting. Evi- dence for the existence of various subtypes of lipid rafts and microdomains was shown in Th1 vs Th2 cells (40) and uropod vs leading edge associated rafts of migrating T cells (41). Further- more, GM1, which is besides GPI-anchored proteins used as another marker for lipid rafts, is not expressed on the surface of resting T cells (18, 42). In contrast, CD48 and CD59 are abun- dantly observed on the surface of the latter T cells, indicating differences between lipid rafts containing GPI proteins or GM1. The importance of GPI-anchored proteins in T cell activation is documented by several publications. Clustering of GPI-anchored proteins by mAbs to CD48, CD55, CD59, and CD73, Thy-1, Ly-6, and sgp-60 readily affects T cell activity usually by the induction of T cell proliferation or IL-2 production (9, 43Ð46). In contrast, under some experimental conditions mAb binding to GPI-anchored proteins led to inhibition (46Ð48) or to a modulation of T cell activation (49). Moreover, T cells lacking GPI-anchored proteins, including GPI-deficient T cells derived from patients with paroxysmal nocturnal hemoglobinuria (50, 51) and T cells derived from CD48 Ϫ/Ϫ mice (52), were severely impaired in their activation via the TCR. Consistent with these observations, similar defects in T cell responsiveness were seen after enzymatic removal of GPI-linked proteins from normal T lymphocytes (53). Thus, it is tempting to speculate that the CD147-mediated depletion of the GPI proteins CD48 and CD59 from lipid rafts as found by our experiments is underlying the blockade of expression of CD25. FIGURE 6. CD147 mAb MEM-M6/6 mediates displacement of the GPI-linked molecules CD48 and CD59 from lipid rafts of T lymphocytes. This assumption is supported by the finding that triggering of A, Resting T lymphocytes were lysed in a buffer containing the nonionic CD59 in CD3-deficient T cells leads to expression of CD25 but not detergent Brij-58. Lysates were subjected to sucrose density gradient cen- of IL-2 (54), which points to a link between CD59 signaling and trifugation. Aliquots of the individual fractions were mixed with 4ϫ non- CD25 induction. reducing SDS sample buffer and a quarter of each fraction was loaded on In conclusion, our results provide evidence that triggering of a 12% SDS-polyacrylamide gels. Proteins were detected by Western blotting potential negative regulating receptor on T cells causes reorgani- using specific mAbs. This experiment is representative of three indepen- zation of lipid rafts that is accompanied by modulation of signaling dent ones. B, T lymphocytes were stimulated in the presence of either mAb and cell activation. It is the task of future studies to unveil the ϩ ϩ MEM-M6/1 (stim M6/1) or MEM-M6/6 (stim M6/6), or without different types and subtypes of microdomains and their role in addition of CD147 mAbs (stim). Subsequently, sucrose density gradient receptor signaling at and across the plasma membrane. centrifugation and Western blotting analysis were performed as described in A. Comparable results were obtained with five independent experiments. C, T lymphocytes were treated either with mAb MEM-M6/1 (M6/1) or Acknowledgments MEM-M6/6 (M6/6) alone. Subsequently, sucrose density gradient centrif- We thank Vaclav Horejsi and Ivan Hilgert for providing a number of ugation and Western blotting analysis were performed as described in A. mAbs, which were invaluable for this study. We thank Thomas Baumruker Comparable results were obtained with three independent experiments. and Eva Prieschl for their expert assistance in performing the multiprobe GEM, Glycosphingolipid-enriched membrane protein. RNase protection assay. inhibition of lipid rafts with TCR/CD3 association by CD147 trig- References gering. However, biochemical analysis of lipid rafts revealed that 1. Horejsi, V., K. Drbal, M. Cebecauer, J. Cerny, T. Brdicka, P. Angelisova, and this assumption was only partially true. H. Stockinger. 1999. GPI-microdomains: a role in signalling via immunorecep- tors. Immunol. Today 20:356. When lipid rafts were isolated by detergent lysis followed by 2. Xavier, R., T. Brennan, Q. Li, C. McCormack, and B. Seed. 1998. Membrane density gradient centrifugation, in accord with previous studies (8, compartmentation is required for efficient T cell activation. Immunity 8:723. 1714 CD147-INDUCED MODULATION OF T CELL ACTIVATION

3. Stefanova, I., V. Horejsi, I. J. Ansotegui, W. Knapp, and H. Stockinger. 1991. 28. Foxwell, B. M., D. A. Taylor-Fishwick, J. L. Simon, T. H. Page, and M. Londei. GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Sci- 1992. Activation induced changes in expression and structure of the IL-7 receptor ence 254:1016. on human T cells. Int. Immunol. 4:277. 4. Cinek, T., and V. Horejsi. 1992. The nature of large noncovalent complexes 29. Schluns, K. S., W. C. Kieper, S. C. Jameson, and L. Lefrancois. 2000. Interleu- containing glycosyl-phosphatidylinositol-anchored membrane glycoproteins and kin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nat. protein tyrosine kinases. J. Immunol. 149:2262. Immunol. 1:426. 5. Simons, K., and E. Ikonen. 1997. Functional rafts in cell membranes. Nature 30. Chow, L. M., M. Fournel, D. Davidson, and A. Veillette. 1993. Negative regu- 387:569. lation of T-cell receptor signalling by tyrosine protein kinase p50csk. Nature 365: 6. Rodgers, W., B. Crise, and J. K. Rose. 1994. Signals determining protein tyrosine 156. kinase and glycosyl-phosphatidylinositol-anchored protein targeting to a glyco- 31. Finco, T. S., T. Kadlecek, W. Zhang, L. E. Samelson, and A. Weiss. 1998. LAT lipid-enriched membrane fraction. Mol. Cell. Biol. 14:5384. is required for TCR-mediated activation of PLC␥1 and the Ras pathway. Immu- 7. Brown, D. A., and E. London. 1997. Structure of detergent-resistant membrane nity 9:617. domains: does phase separation occur in biological membranes? Biochem. Bio- 32. Moran, M., and M. C. Miceli. 1998. Engagement of GPI-linked CD48 contributes phys. Res. Commun. 240:1. to TCR signals and cytoskeletal reorganization: a role for lipid rafts in T cell 8. Stulnig, T. M., M. Berger, T. Sigmund, H. Stockinger, V. Horejsi, and activation. Immunity 9:787. W. Waldhausl. 1997. Signal transduction via glycosyl phosphatidylinositol-an- 33. Yashiro, O. Y., X. Y. Zhou, o. K. Toyo, X. G. Tai, C. S. Park, T. Hamaoka, chored proteins in T cells is inhibited by lowering cellular cholesterol. J. Biol. R. Abe, K. Miyake, and H. Fujiwara. 2000. Non-CD28 costimulatory molecules Chem. 272:19242. present in T cell rafts induce T cell costimulation by enhancing the association of 9. Deckert, M., M. Ticchioni, and A. Bernard. 1996. Endocytosis of GPI-anchored TCR with rafts. J. Immunol. 164:1251. proteins in human lymphocytes: role of glycolipid-based domains, actin cytoskel- 34. Zhang, W., L. J. Sloan, J. Kitchen, R. P. Trible, and L. E. Samelson. 1998. LAT: eton, and protein kinases. J. Cell Biol. 133:791. the ZAP-70 tyrosine kinase substrate that links T cell receptor to cellular acti- 10. Kabouridis, P. S., J. Janzen, A. L. Magee, and S. C. Ley. 2000. Cholesterol vation. Cell 92:83. depletion disrupts lipid rafts and modulates the activity of multiple signaling 35. Kropshofer, H., S. Spindeldreher, T. A. Rohn, N. Platania, C. Grygar, N. Daniel, pathways in T lymphocytes. Eur. J. Immunol. 30:954. A. Wolpl, H. Langen, V. Horejsi, and A. B. Vogt. 2002. Tetraspan microdomains 11. Stulnig, T. M., M. Berger, T. Sigmund, D. Raederstorff, H. Stockinger, and distinct from lipid rafts enrich select peptide-MHC class II complexes. Nat. Im- W. Waldhausl. 1998. Polyunsaturated fatty acids inhibit T cell signal transduction munol. 3:61. by modification of detergent-insoluble membrane domains. J. Cell Biol. 143:637. 36. Stonehouse, T. J., V. E. Woodhead, P. S. Herridge, H. Ashrafian, M. George, 12. Bohuslav, J., V. Horejsi, C. Hansmann, J. Sto¬ckl, U. H. Weidle, O. Majdic, B. M. Chain, and D. R. Katz. 1999. Molecular characterization of U937-depen- I. Bartke, W. Knapp, and H. Stockinger. 1995. Urokinase plasminogen activator dent T-cell co-stimulation. Immunology 96:35. ␤ receptor, 2-integrins, and Src-kinases within a single receptor complex of human 37. Woodhead, V. E., T. J. Stonehouse, M. H. Binks, K. Speidel, D. A. Fox, A. Gaya, monocytes. J. Exp. Med. 181:1381. D. Hardie, A. J. Henniker, V. Horejsi, K. Sagawa, et al. 2000. Novel molecular 13. Cerny, J., H. Stockinger, and V. Horejsi. 1996. Noncovalent associations of T mechanisms of dendritic cell-induced T cell activation. Int. Immunol. 12:1051. lymphocyte surface proteins. Eur. J. Immunol. 26:2335. 38. Millan, J., M. Qaidi, and M. A. Alonso. 2001. Segregation of co-stimulatory 14. Leitenberg, D., F. Balamuth, and K. Bottomly. 2001. Changes in the T cell re- components into specific T cell surface lipid rafts. Eur. J. Immunol. 31:467. ceptor macromolecular signaling complex and membrane microdomains during T 39. Kosugi, A., S. Saitoh, S. Noda, K. Yasuda, F. Hayashi, M. Ogata, and cell development and activation. Semin. Immunol. 13:129. T. Hamaoka. 1999. Translocation of tyrosine-phosphorylated TCR␨ chain to gly- 15. Montixi, C., C. Langlet, A. M. Bernard, J. Thimonier, C. Dubois, M. A. Wurbel, colipid-enriched membrane domains upon T cell activation. Int. Immunol. 11: J. P. Chauvin, M. Pierres, and H. T. He. 1998. Engagement of T cell receptor 1395. triggers its recruitment to low-density detergent-insoluble membrane domains. 40. Balamuth, F., D. Leitenberg, J. Unternaehrer, I. Mellman, and K. Bottomly. 2001. EMBO J. 17:5334. Distinct patterns of membrane microdomain partitioning in Th1 and Th2 cells. 16. Field, K. A., D. Holowka, and B. Baird. 1997. Compartmentalized activation of Immunity 15:729. the high affinity immunoglobulin E receptor within membrane domains. J. Biol. 41. Gomez-Mouton, C., J. L. Abad, E. Mira, R. A. Lacalle, E. Gallardo, Chem. 272:4276. S. Jimenez-Baranda, I. Illa, A. Bernad, S. Manes, and A. C. Martinez. 2001. 17. Lang, M. L., L. Shen, and W. F. Wade. 1999. ␥-Chain dependent recruitment of Segregation of leading-edge and uropod components into specific lipid rafts dur- tyrosine kinases to membrane rafts by the human IgA receptor Fc␣R. J. Immunol. ing T cell polarization. Proc. Natl. Acad. Sci. USA 98:9642. 163:5391. 42. Martin, M., H. Schneider, A. Azouz, and C. E. Rudd. 2001. Cytotoxic T lym- 18. Viola, A., S. Schroeder, Y. Sakakibara, and A. Lanzavecchia. 1999. T lympho- phocyte antigen 4 and CD28 modulate cell surface raft expression in their reg- cyte costimulation mediated by reorganization of membrane microdomains. Sci- ulation of T cell function. J. Exp. Med. 194:1675. ence 283:680. 43. Thompson, L. F., J. M. Ruedi, A. Glass, M. G. Low, and A. H. Lucas. 1989. 19. Kasinrerk, W., E. Fiebiger, I. Stefanova«, T. Baumruker, W. Knapp, and Antibodies to 5Ј-nucleotidase (CD73), a glycosyl-phosphatidylinositol-anchored H. Stockinger. 1992. Human leukocyte activation antigen M6, a member of the protein, cause human peripheral blood T cells to proliferate. J. Immunol. 143: Ig superfamily, is the species homologue of rat OX-47, mouse basigin, and 1815. chicken HT7 molecule. J. Immunol. 149:847. 44. Presky, D. H., M. G. Low, and E. M. Shevach. 1990. Role of phosphatidylinos- 20. Biswas, C., Y. Zhang, R. DeCastro, H. Guo, T. Nakamura, H. Kataoka, and itol-anchored proteins in T cell activation. J. Immunol. 144:860. K. Nabeshima. 1995. The human tumor cell-derived collagenase stimulatory fac- 45. Korty, P. E., C. Brando, and E. M. Shevach. 1991. CD59 functions as a signal- tor (renamed EMMPRIN) is a member of the immunoglobulin superfamily. Can- transducing molecule for human T cell activation. J. Immunol. 146:4092. cer Res. 55:434. 46. Malek, T. R., T. J. Fleming, and E. K. Codias. 1994. Regulation of T lymphocyte 21. Schlosshauer, B., and K. H. Herzog. 1990. Neurothelin: an inducible cell surface function by glycosylphosphatidylinositol (GPI)-anchored proteins. Semin. Immu- glycoprotein of blood-brain barrier-specific endothelial cells and distinct neurons. nol. 6:105. J. Cell Biol. 110:1261. 47. Reiser, H. 1990. sgp-60, a signal-transducing glycoprotein concerned with T cell 22. Miyauchi, T., Y. Masuzawa, and T. Muramatsu. 1991. The basigin group of the activation through the T cell receptor/CD3 complex. J. Immunol. 145:2077. immunoglobulin superfamily: complete conservation of a segment in and around 48. Chavin, K. D., L. Qin, J. Lin, J. Woodward, P. Baliga, K. Kato, H. Yagita, and transmembrane domains of human and mouse basigin and chicken HT7 antigen. J. S. Bromberg. 1994. Anti-CD48 (murine CD2 ligand) mAbs suppress cell me- J. Biochem. 110:770. diated immunity in vivo. Int. Immunol. 6:701. 23. Koch, C., G. Staffler, R. Huttinger, I. Hilgert, E. Prager, J. Cerny, P. Steinlein, 49. Marmor, M. D., M. F. Bachmann, P. S. Ohashi, T. R. Malek, and M. Julius. 1999. O. Majdic, V. Horejsi, and H. Stockinger. 1999. T cell activation-associated Immobilization of glycosylphosphatidylinositol-anchored proteins inhibits T cell epitopes of CD147 in regulation of the T cell response, and their definition by growth but not function. Int. Immunol. 11:1381. antibody affinity and antigen density. Int. Immunol. 11:777. 50. Romagnoli, P., and C. Bron. 1999. Defective TCR signaling events in glyco- 24. Igakura, T., K. Kadomatsu, O. Taguchi, H. Muramatsu, T. Kaname, T. Miyauchi, sylphosphatidylinositol-deficient T cells derived from paroxysmal nocturnal he- K. Yamamura, K. Arimura, and T. Muramatsu. 1996. Roles of basigin, a member moglobinuria patients. Int. Immunol. 11:1411. of the immunoglobulin superfamily, in behavior as to an irritating odor, lympho- 51. Romagnoli, P., and C. Bron. 1997. Phosphatidylinositol-based glycolipid-an- cyte response, and blood-brain barrier. Biochem. Biophys. Res. Comm. 224:33. chored proteins enhance proximal TCR signaling events. J. Immunol. 158:5757. 25. Krutmann, J., R. Kirnbauer, A. Kock, T. Schwarz, E. Schopf, L. T. May, 52. Gonzalez, C. J., C. J. Wise, Y. Latchman, G. J. Freeman, A. H. Sharpe, and P. B. Sehgal, and T. A. Luger. 1990. Cross-linking Fc receptors on monocytes H. Reiser. 1999. CD48-deficient mice have a pronounced defect in CD4ϩ T cell triggers IL-6 production: role in anti-CD3-induced T cell activation. J. Immunol. activation. Proc. Natl. Acad. Sci. USA 96:1019. 145:1337. 53. Yeh, E. T., H. Reiser, A. Bamezai, and K. L. Rock. 1988. TAP transcription and 26. Snapper, S. B., F. S. Rosen, E. Mizoguchi, P. Cohen, W. Khan, C. H. Liu, phosphatidylinositol linkage mutants are defective in activation through the T cell T. L. Hagemann, S. P. Kwan, R. Ferrini, L. Davidson, et al. 1998. Wiskott- receptor. Cell 52:665. Aldrich syndrome protein-deficient mice reveal a role for WASP in T but not B 54. Deckert, M., M. Ticchioni, B. Mari, D. Mary, and A. Bernard. 1995. The gly- cell activation. Immunity 9:81. cosylphosphatidylinositol-anchored CD59 protein stimulates both T cell receptor 27. Viola, A. 2001. The amplification of TCR signaling by dynamic membrane mi- ␨/ZAP-70-dependent and -independent signaling pathways in T cells. Eur. J. Im- crodomains. Trends Immunol. 22:322. munol. 25:1815.