Malignant Plasma Cell Lines Express a Functional CD28 Molecule

Home , Epitope

Leukemia (1998) 12, 610–618  1998 Stockton Press All rights reserved 0887-6924/98 $12.00 http://www.stockton-press.co.uk/leu Malignant plasma cell lines express a functional CD28 molecule X-G Zhang1,2, D Olive3, J Devos1, C Rebouissou1, M Ghiotto-Ragueneau3, M Ferlin1 and B Klein1,2 ˆ 1Institute of Molecular Genetics, Montpellier; 2Unit for Cellular Therapy, Hopital Saint Eloi, Montpellier; and 3INSERM U 119, Marseille, France

The function of CD28 molecules that are present on malignant Materials and methods plasma cells of human myeloma cell lines (HMCL) was studied. First, myeloma cells expressed a similar density of CD28 antigen to that of normal T cells. The myeloma CD28 molecules Antibodies and reagents were able to bind B7-Ig molecules as well as L cells transfected with a B7-1 cDNA, and anti-CD28 mAb inhibited the binding. The agonist (CD28.2) and antagonist (CD28.6) anti-CD28 Myeloma cells did not express B7-1 antigens but a low density mAbs were obtained as described elsewhere.12 MAbs against of B7-2 antigens. The myeloma B7-2 molecules of two HMCL CD3, CD16, CD25, CD56, CD69, B7-1, HLA-DR antigens, were able to bind CTLA-4 protein. No autocrine CD28:B7-2 acti- FITC-conjugated (FabЈ) fragments of goat antibodies to mouse vation could be evidenced as we found no spontaneous bind- 2 ing of the p85 subunit of PI-3 kinase to CD28 molecules. In IgG and FITC-conjugated antibodies to Fc fragments of human addition, a blocking anti-CD28 mAb did not affect the IL-6- Ig, IgG1 and IgG2 murine control antibodies (recognizing no dependent or autonomous proliferation of the HMCL. The acti- human antigens) were purchased from Immunotech vation of myeloma CD28 molecules with or without TPA stimu- (Marseille, France), mAbs to CD40, CD40L, B7-2 antigens lation did not affect the proliferation, survival, differentiation, from Pharmingen (San Diego, CA, USA), magnetic beads expression of activation antigens and cytokine receptors or coated with anti-CD19, anti-CD14 or anti-murine Ig anti- cytokine production of myeloma cells. However, the triggering of myeloma CD28 molecules by B7-1 transfectant cells resulted bodies from Dynal (Oslo, Norway). in binding of the p85 subunit of PI-3 kinase to CD28 molecules The human B7-1-human Ig and human CTLA-4-human Ig as previously shown for T cell CD28 molecules. This fusion proteins (B7-1-Ig and CTLA-4-Ig) were kindly provided expression of a large density of CD28 molecules able to bind by Dr Linsley (Bristol Myers Squibb, Seattle, WA, USA). Pur- B7 molecules might contribute to a downregulation of the ified IL-6 was provided by Dr Ytier (Ares Serono, Geneva, immune control of myeloma cells. Switzerland) and neutralizing AH-65 anti-IL-6 mAb by Dr Keywords: CD28; CD80; plasma cell; multiple myeloma Brailly (Immunotech). The B-T2 anti-gp130 mAbs were obtained by our group.13,14 PHA and purified human Ig were purchased from Sigma (St Louis, MO, USA). Rabbit antibodies Introduction to p85 subunit of PI-3 kinase or phosphotyrosine were purchased from UBI (Lake Placid, NY, USA) or Transduction Lab- In the process of T cell activation by antigen-presenting cell oratories (Lexington, KY, USA). Peroxidase-conjugated anti- (APC), activation of CD28 molecules present on T cells has bodies to mouse or rabbit antibodies and the ECL detection 1,2 been shown to be essential. As recently reviewed, T cell kit were purchased from Amersham (Buckinghamshire, UK). activation by APC requires at least three steps: (1) binding of Tritiated thymidine was purchased from CEA (Saclay, France). APC to T cells through adhesion molecules such as ICAM- 1:LFA-1 and LFA-3:CD2; (2) presentation of peptide antigen bound to molecules of the major histocompatibility complex Cell lines of APC to the T cell receptor; and (3) activation of the T cell CD28 molecules or related receptors by their counter recep- The XG-1 and XG-6 IL-6-dependent HMCL were obtained in tors B7-1 (CD80) or B7-2 (CD86) present on APC. A disruption our laboratory10 and the RPMI8226 and U266 HMCL were of the CD28:B7-1 interaction may result in induction of T cell purchased from ATCC (Rockville, MD, USA). The XG cells 1 anergy. CD28 activation has been shown to be critical in were cultured in culture medium (RPMI 1640 medium sup- 1,2 inducing IL-2 secretion in activated T cells. More recently, plemented with 10% fetal calf serum (FCS), 2 mML-glutamine − activation of CD28 molecules was shown to induce activation and 5 × 10 5 M 2-mercaptoethanol) and with 1 ng/ml of IL-6. 3–7 of phosphatidyl-inositol 3-kinase (PI-3) in T cells and to pro- The RPMI8226 and U266 cells were routinely cultured in cul- 8 mote survival by upregulating bcl-XL expression. ture medium without IL-6. Murine L cell fibroblasts transfected In an early report, CD28 molecules were found to be with human B7-1 cDNA (B7-1 L cell) and parental L cells 9 present on human malignant plasma cells. A high expression were generously provided by Dr A Truneh (Smith Kline of CD28 molecules on a large panel of human myeloma cell Beecham, Philadelphia, PA, USA). Human CD28 transfectant 10 lines (HMCL) or on freshly explanted myeloma cells was (DWT6.11) was obtained by transfecting full-length human 11 recently confirmed. The present study demonstrates for the CD28 cDNA in pH␤-pr-neo vector into murine T cell first time that the CD28 molecule expressed on malignant hybridoma.6 These cell lines were cultured in culture plasma cells is functional. It is able to bind its counter recep- medium. All cell lines were free of mycoplasma contami- tors, the B7 molecules, and this binding induces activation of nation as assayed using an ELISA kit from Boehringer PI-3 kinase in myeloma cells. (Mannheim, Germany).

Analysis of survival of myeloma cells Correspondence: B Klein, INSERM U475, 99 Rue Puech Ville, 34100 Montpellier cedex, France; Fax 33 4 67 04 18 62 To investigate the effects of CD-28 molecule on the survival Received 7 March 1997; accepted 12 August 1997 of the human myeloma cells, XG-6 cells were washed once Functional CD28 on human plasma cells X-G Zhang et al 611 with culture medium, incubated for 5 h at 37°C in culture allogenic MLR was performed using irradiated (3000 rads) medium alone, and washed again twice to remove recombi- XG-6 HMCL cocultured at 5 × 104 cells/well with 105 allo- nant bound IL-6. Cells (3 × 105 cells/ml) were then cultured genic T cells in 96-well round-bottomed plates in RPMI1640 with either culture medium alone or with the activating and 5% heat-inactivated human AB serum. When indicated, CD28.2 anti-CD28 mAb (10 ␮g/ml) or with 1 ng/ml of IL-6. A agonist (CD28.2) or antagonist (CD28.6) anti-CD28 mAbs or mouse IgG1 anti-CD3 mAb was used as a control antibody. their isotypic controls were added at 10 ␮g/ml. Purified T cells After 24, 48 and 72 h of culturing, the percentage of viable were also cultured with PHA (1 ␮g/ml). After 5 days of cocul- cells was determined by the trypan blue dye exclusion ture, cells were pulsed with 1 ␮Ci of tritiated thymidine for method, and apoptotic cells were detected by propidium 12 h, harvested and counted. All microculture tests were car- iodide staining and flow cytometry. Apoptotic cells had a ried out in triplicate. DNA content inferior to that of cells in the G1 phase of the cell cycle. Proliferation and immunoglobulin production of myeloma cells Flow cytometry To investigate the effect of activation of CD28 molecules on Indirect immunofluorescence was done by incubating 5 × 105 the proliferation of myeloma cells, cells were washed once cells with 1 ␮g of murine mAb or soluble B7-1-Ig or soluble with culture medium, incubated for 5 h at 37°C in culture CTLA-4-Ig in 100 ␮l of staining buffer consisting of PBS with medium alone to remove bound IL-6, and washed again 0.25% bovine serum albumin and 0.01% sodium azide for 45 twice. Myeloma cells were then cultured with various concen- min at 4°C. Isotype-matched Ig was used as a control (human trations of the CD28.2 activating anti-CD28 mAb or control or murine IgG). Cells were then stained with either FITC-con- mAb with or without TPA (100 ng/ml; Sigma) in 96-well flat- Ј jugated (Fab )2 fragments of goat antibodies to mouse IgG or bottomed microplates for 2 or 5 days. We used a mouse IgG1 FITC-conjugated antibodies to Fc fragments of human Ig. Flow anti-CD3 mAb that did not bind to myeloma cells as a control cytometry was performed with a FACScan apparatus (Becton antibody. In some experiments, the effect of the blockage of Dickinson, Palo Alto, CA, USA). Cells were considered as CD28 activation by the CD28.6 neutralizing anti-CD28 mAb positive when the fluorescence intensity of the labelling was was investigated by adding 10 ␮g/ml of the anti-CD28 mAb greater than the mean fluorescence intensity + 2s.d. obtained in 5-day cultures of XG cells with IL-6 (100 pg/ml) or of with a control mAb. RPMI8226 and U266 cells. At the end of the culture, 0.5 ␮Ci of tritiated thymidine (25 Ci/mM; CEA, Saclay) were added to each culture well for 8 h and tritiated thymidine incorporation CD28 adhesion assays was determined on six replicate culture wells. To evaluate the production of immunoglobulin, myeloma cells (105 cells/ml) Myeloma cells were labeled with 51Chrome and washed were cultured with various concentrations of the activating twice. Cells were incubated for 1 h at 20°C in culture medium CD28.2 anti-CD28 mAb or the anti-CD3 control mAb for 2 containing 10 mM EDTA either with isotype-matched control days and the concentration of human kappa or lambda mAb or with the mAbs to be tested for the inhibition of immunoglobulin light chains was determined by ELISA as pre- adhesion. Labeled cells (5 × 104 cells per well in 200 ␮lof viously reported.15 culture medium) were then added to monolayers of B7-1 L cells or control parental L cells. Adhesion was initiated by centrifugation and continued at 37°C for 1 h. Monolayers Immunoprecipitation and immunoblotting were washed five times with ice-cold RPMI 1640 medium, solubilized by addition of 0.5 M NaOH and the radioactivity Myeloma cells or CD28 transfected cells (10 × 106 cells in was counted in a gamma counter. The percentage of myeloma 1 ml) were incubated with B7-1 L cells (5 × 106)at37°Cin cells that bound to B7-1 L cell monolayer was estimated by RPMI 1640 complete medium for 5 min. B7-1 L cells had subtracting the radioactivity in control L cell monolayer (due been previously incubated with anti-B7-1 mAb or an isotype- to the non-specific adhesion of myeloma cells to L cells) from matched control mAb (10 ␮g per ml). Cells were lysed for 1 h the radioactivity in B7-1 L cell monolayer and dividing the at 4°C end over end with lysis buffer (50 mM Hepes, 150 mM result by the total radioactivity of the labeled cells added to NaCl, 10% glycerol, 1% Triton, 1.5 mM MgCl2,1mM EGTA, culture wells. 10 ␮g/ml leupeptin, 1 mM PMSF, 1 mM Na orthovanadate, 100 mM pyrophosphate, 100 mM NaF, 10 mM DTT). After pre- clearing at 14 000 g, the lysates were immunoprecipitated for Allogenic mixed lymphocyte reaction (allo-MLR) 2 h at 4°C using the CD28.2 mAb or isotype-matched control mAb and protein G-sepharose beads (Pharmacia, Uppsala, Heparinized-venous peripheral blood from one healthy volun- Sweden). Immunoprecipitated samples were resolved in 10% teer was collected under EDTA after written informed consent. sodium dodecyl sulfate-polyacrylamide gel electrophoresis T lymphocytes were purified by depletion of non-T cells with (SDS-PAGE) and transferred to nitrocellulose membrane mAbs and magnetic beads. Monocytes and B cells were (Schleicher and Schuell, Dassel, Germany). Membranes were depleted using CD14 and CD19-coated microbeads. Then, blocked for 1 h at room temperature in 138 mM NaCl, 3 mM non-T cells were removed by incubation with a cocktail of KCl, 25 mM Tris–HCl (pH 7.4), 0.1% Tween 20 (TBS-T), 5% mAbs to CD16, CD56 and HLA-DR antigens and goat anti- bovine serum albumin (BSA), then incubated for 1 h at room mouse Ig microbeads. After two rounds of purification, the temperature with a polyclonal rabbit antiserum to p85 subunit purity of CD3 cells was always higher than 98%. Cells were of PI-3 kinase (UBI, Lake Placid, NY, USA). The primary anti- frozen in 50% FCS/10% DMSO at 20 × 106 cells per vial. To bodies were visualized with peroxidase-conjugated donkey evaluate myeloma cell-induced T cell proliferation, primary anti-rabbit immunoglobulins (at 1:10 000 dilution in TBS-T) Functional CD28 on human plasma cells X-G Zhang et al 612

Figure 1 Comparison of membrane expression of CD28 molecule on myeloma cells and human peripheral T lymphocytes. Cells from the U266, RPMI8226, XG-1 and XG-6 myeloma cell lines and peripheral blood mononuclear cells from two healthy donors were stained with the Ј CD28.6 anti-CD28 mAb (solid line) or an isotype-matched control murine mAb (dotted line) and FITC-conjugated F(ab )2 fragments of goat anti-mouse immunoglobulin. The ﬂuorescence was analyzed by ﬂow cytometry using a FACScan.

and by an enhanced chemiluminescence (ECL) detection sys- Results tem. The membranes were stripped with 100 mM glycin, pH 2.2, 0.1% NP 40 and 1% SDS for 30 min twice, reprobed Expression of CD28 antigen on human myeloma cell with the CD28.6 anti-CD28 mAb and peroxidase-conjugated lines sheep anti-mouse immunoglobulins (at 1:10 000 dilution in TBS-T) and bands were visualized with an ECL kit. Additional Results are outlined in Figure 1. All the cells of the four HMCL membranes were prepared, probed with rabbit anti-phospho- expressed the CD28 molecules as indicated by the labeling tyrosine antibodies and peroxidase-conjugated donkey anti- with anti-CD28 mAb. Peripheral blood T cells from two heal- rabbit antibodies and bands were visualized with an ECL thy donors were labeled with the same anti-CD28 mAb and detection system. analyzed on the same day and with the same FACS para- Functional CD28 on human plasma cells X-G Zhang et al 613

Figure 2 Flow cytometric analysis of B7-Ig binding to myeloma CD-28 molecule. The myeloma cells were incubated with human B7-1-Ig fusion protein (solid line) or purified human IgG (dotted line) and FITC-conjugated anti-human Ig antibodies. The fluorescence was analyzed by flow cytometry using a FACScan. meters as the HMCL. The mean fluorescence density was 45, with B7-1-Ig molecules and FITC-labeled anti-human Ig anti- 460, 115 and 147 for the RPMI8226, U266, XG-1 and XG-6 bodies as compared to cells incubated with FITC-labeled anti- myeloma cells, respectively, and was 112 and 190 for the human Ig antibodies alone. The more marked shift was freshly explanted T cells of two healthy donors. Thus, mye- observed for U266 cells (Figure 2) in agreement with their loma cells of three HMCL expressed a similar density of CD28 higher expression of CD28 molecules (Figure 1). By using molecules as freshly explanted circulating T cells. U266 mye- peripheral blood T cells from healthy individuals, only a slight loma cells expressed a very large CD28 density. shift of the fluorescence profile was observed (Figure 2). The ability of myeloma cell CD28 molecules to bind B7-1 protein Binding of B7-1 antigen by CD28 molecules on was further demonstrated by the binding of radiolabeled mye- myeloma cells loma cells to B7-1 L cells. As indicated in Figure 3, 26.4 and 24% of XG-6 and U266 myeloma cells, respectively, were As illustrated in Figure 2, a significant shift of the fluorescence able to bind to B7-1 L cells. An anti-B7-1 mAb blocked this profile was observed when myeloma cells were incubated binding by 90 and 70% respectively. An anti-CD28 mAb also Functional CD28 on human plasma cells X-G Zhang et al 614

Figure 3 Adhesion of myeloma cells to B7-1 transfected L cells. Monolayers of B7-1-transfected L cells (B7-1 L cells) or L cells (104 cells/well in 96-well ﬂat-bottomed microplates) were cultured overnight, and then incubated for 1 h with 10 ␮g/ml of an anti-B7-1 mAb or a control antibody in RPMI1640 at 37°C. The myeloma cells were labeled with 51Cr, preincubated with 10 ␮g/ml of the CD28.6 anti-CD28 mAb or a control antibody in RPMI1640 with 10 mM EDTA, and added to the monolayer of B7-1 L cells or L cells. The cellular adhesion was measured. Data are the mean percentages of bound cells determined on six culture wells.

Table 1 Expression of the B7 molecules on myeloma cells and Expression of the family of B7 molecules on myeloma binding to CTLA4-Ig cells

B7-1 B7-2 CTLA4 Ig Myeloma cells did not express B7-1. They expressed weakly the B7-2 antigen. For two cell lines, RPMI8226 and XG-1, a RPMI 8226 Ͻ5% 45% 60 (24) (15) detectable binding of CTLA-4-Ig molecules to B7-2 molecules was found (Table 1). U266 Ͻ5% 31% Ͻ5% (26) Ͻ XG-1 5% 82% 32% Lack of obvious biological effects of CD28 activation (26) (12) in myeloma cells XG-6 Ͻ5% 10% Ͻ5% (20) In order to investigate a biological effect of CD28 activation in myeloma cells, we looked for a modulation of myeloma Myeloma cells were incubated with anti-B7-1 or anti-B7-2 mAbs or CTLA4-Ig or control antibodies. Cells were then labeled with a FITC cell proliferation. We also looked for a change in the conjugated goat anti-murine Ig or anti-human Ig antibody. Fluor- expression of various antigens as well as in cytokine or escence was analyzed with a flow cytometer and results are the immunoglobulin production. Myeloma cells were activated percentage of cells labeled by the different antibodies. In brackets by the CD28.2 anti-CD28 mAb known to activate T cell CD28 is indicated the mean density of the fluorescence profiles. The flu- molecules.12 This was further confirmed in the experiments orescence intensities obtained with a control murine mAb shown in Figure 4. Indeed, allogenic T cells were poorly (recognizing no human antigens) or human Ig were set between 4 and 6. stimulated by irradiated XG-6 myeloma cells. Addition of CD28.2 anti-CD28 mAb, unlike CD28.6 mAb, dramatically increased T cell proliferation induced by irradiated myeloma cells. Upon removal of IL-6, XG-6 cells died by apoptosis within blocked it indicating that myeloma cells failed to express 3 days (Figure 5). Addition of agonist CD28.2 mAb did not CTLA-4, the other counter-receptor for B7-1 and B7-2 mol- delay the apoptosis of XG-6 cells (Figure 5). We found no ecules. increase or decrease of the myeloma cell proliferation when Functional CD28 on human plasma cells X-G Zhang et al 615 (results not shown). This was not surprising since phosphorylation of CD28 T cell molecules was rarely found3 except for CD28 transfectants that expressed a very large density of CD28 molecules.6

Discussion

The current study shows that myeloma cells from HMCL express a similar density of CD28 molecules to that of freshly explanted T cells. We have shown that the myeloma cell CD28 molecules are functional molecules in term of binding to the B7 counter receptor using either B7-Ig molecules or L cells transfected with B7-1 human cDNA. The CD28.6 anti- CD28 mAb inhibited the binding to B7-1 L cells indicating that myeloma cells do not express CTLA-4 molecules, a second receptor for B7-1 molecules. CTLA-4 molecules were recently shown to be expressed on B cells activated by T cells.16 Malignant human plasma cells are, at present, the only cells that are found to express a functional CD28 molecule together with T cells.1,2 As normal plasma cells fail to express CD2817 and as activation of CD28 antigen is a crucial event for induction of functional T cells (mainly by triggering IL-2 production) and prevention from apoptosis,8 one might suspect a major role of CD28 activation for myeloma cell proliferation, survival or differentiation. Our previous studies have shown that IL-6 was a growth factor for myeloma cells18 and that IL-6- dependent myeloma cell lines could be reproducibly obtained from patients with terminal disease.10 By using the agonist CD28.2 mAb (known to activate T cells) in combination with Figure 4 Agonist CD28.2 mAb potentiated the stimulation of 12 allogenic T cells by myeloma cells. Puriﬁed T cells (105 cells) were TPA, we found no modulation of the proliferation of the IL- stimulated with or without irradiated XG-6 myeloma cells (5 × 104 6-dependent XG-1 and XG-6 HMCL or of the autonomously cells) for 5 days with 10 ␮g/ml of CD28.2 or CD28.6 anti-CD28 mAb growing RPMI8226 and U266 HMCL. This is in agreement or control murine IgG1. T cells were also cultured with PHA. At the with a lack of modulation by the anti-CD28 mAbs of IL-6R end of the culture, the proliferation of T cells was evaluated by triti- ± and gp130 IL-6 transducer expression on myeloma cells or of ated thymidine incorporation. Results are the means s.d. of the triti- IL-6 production by myeloma cells. We also found that trig- ated thymidine incorporation determined on sextuplate culture wells. gering of myeloma CD28 by agonist CD28.2 anti-CD28 mAb or B7-1 L cells did not abrogate or delay the apoptotic death myeloma cells were cultured with the agonist CD28.2 or the of myeloma cells that occurred upon removal of IL-6. This antagonist CD28.6 mAbs with or without TPA (Figure 6). In negative result was noteworthy because CD28 activation has particular, activation of CD28 molecules was unable to sup- been shown to prevent T cells from apoptosis by increasing 8 port the proliferation of the IL-6-dependent XG-1 and XG-6 the bcl-XL anti-apoptotic protein. Finally, we found no induc- HMCL. This is in agreement with the lack of induction of IL- tion or modulation of the expression of several activation anti- 6 production in this HMCL by the CD28.2 mAb and TPA gen or cytokines known to be induced on CD28-activated T (results not shown). Activation of CD28 molecules did not cells (CD25, CD69, CD40L) as well as no change in the Ig induce expression of CD25, CD40L, CD69, B7-1 or HLA-DR production. A lack of obvious function of activation of CD28 antigens (Table 2). It did not affect the expression of CD38, antigen on myeloma cells might be explained by an already CD40, B7-2, IL-6R or gp130 IL-6 transducer, or modify the optimal autocrine activation of the CD28 molecules by the low Ig production by myeloma cells (Table 2). B7-2 molecules present in low density on myeloma cells. Our present results may rule out this hypothesis since we found no modulation of myeloma cell proliferation or Ig production Binding of PI-3 kinase to myeloma CD28 molecules by adding a blocking anti-CD28 mAb. In order to further elucidate whether activation of myeloma CD28 immunoprecipitates were easily detectable in U266 CD28 could have some biological effects, we looked for acti- cells contrary to XG-6 cells (Figure 7) in agreement with the vation of signal transduction. We failed to detect tyrosine higher CD28 expression in U266 cells (Figure 1). The binding phosphorylation of myeloma CD28 upon binding of myeloma of U266 or XG-6 myeloma cells to B7-1 L cells induced the cells to B7-1 L cells. This is not surprising since phosphoryl- coimmunoprecipitation of the p85 subunit of the PI-3 kinase ation of T cell CD28 is generally not detected.3 Such phos- with CD28 molecules. Pre-incubation of myeloma cells with phorylation was reported for CD28 transfectants that anti-B7-1 mAb inhibited binding of P85 to myeloma CD28 expressed a very large CD28 density.6 If attempts to demon- (Figure 7). No detectable binding of P85 to CD28 could be strate T cell CD28 phosphorylation have generated equivocal detected in myeloma cells without exogenous activation of results, several groups have shown that CD28 activation CD28 (Figure 7). For the two myeloma cell lines, we failed to resulted in the binding of the P85 subunit of PI-3 kinase and detect phosphorylation of the myeloma CD28 molecules activation of this kinase.2–6 We found similar ﬁndings for Functional CD28 on human plasma cells X-G Zhang et al 616

Figure 5 Activation of CD28 did not prevent myeloma cells from apoptosis. XG-6 myeloma cells were extensively washed to remove IL-6 and cultured for 72 h at a concentration of 3 × 105 cells/ml in culture medium supplemented with 10% of FCS and either 1 ng/ml of IL-6 (left panel), no cytokine (middle panel) or 10 ␮g/ml of the agonist CD28.2 anti-CD28 mAb (right panel). At the end of the culture, DNA content was assayed by propidium iodide staining. Cells in apoptosis had a DNA content inferior to that of the G1 phase of the cell cycle.

Table 2 Activation of myeloma CD28 molecules did not change antigen expression and immunoglobulin production by myeloma cells

XG-1 XG-1 + B7-1 L cells XG-6 XG-6 + B7-1 L cells

CD25 –––– CD40 ligand –––– CD69 –––– B7-1 –––– HLA-DR –––– CD38 98 (436) 98 (395) 97 (183) 98 (211) CD40 – – 52 (23) 60 (27) B7-2 72 (35) 64 (29) 43 (18) 48 (22) IL-6R 88 (21) 81 (20) 80 (14) 82 (18) Gp130 86 (10) 88 (12) 80 (14) 82 (18) Ig (pg/cell/24 h) 5 5.8 6 5.5

XG-1 or XG-6 myeloma cells (5 × 104 cells/100 ␮l/well) were mixed with irradiated L cells or B7-1-transfected L cells (5 × 104 cells/well) in 96-well round-bottomed microplates. Conjugate formation was allowed by centrifugation of plates and cells were cultured for 2 days at 37°C in RPMI1640, 10% of SVF and 1 ng/ml of IL-6. At the end of the culture, culture supernatants were harvested and assayed for human immunoglobulin. Surface antigen expression was determined by FACS analysis after labelling with specific mAbs and FITC-conjugated goat anti-mouse Ig. Viable cells were identified by FSC/SSC characteristics. Results are the percentages of viable positive cells and in brackets the mean fluorescence intensities. The fluorescence intensities obtained with a murine control mAb (recognizing no human antigens) were set between 4 and 6. Functional CD28 on human plasma cells X-G Zhang et al 617

Figure 6 Agonist or antagonist anti-CD28 mAb did not stimulate or inhibit myeloma cell proliferation. Myeloma cells (104 cells/100 ␮l/culture well) were cultured for 5 days with 10 ␮g/ml of an anti-CD3 control mAb (control mAb), 50 pg/ml or 100 pg/ml of IL-6 and 10 ␮g/ml of an anti-CD3 control mAb (IL6 (50 pg) and IL6 (100 pg)), 50 pg/ml of IL-6 and 10 ␮g/ml of the antagonist CD28.6 anti-CD28 mAb (IL6 (50 pg) + antagonist CD28.6 Ab),10 ␮g/ml of the agonist CD28.2 anti-CD28 mAb with (TPA + agonist CD28.2 Ab) or without (agonist CD28.2 Ab) 100 ng/ml of TPA, or 10 ␮g/ml of the antagonist CD28.6 anti-CD28 mAb with (TPA + antagonist CD28.6 Ab) or without (antagonist CD28.6 Ab) 100 ng/ml of TPA, or with 100 ng/ml of TPA (TPA). Cell proliferation was evaluated by tritiated thymidine incorporation. Results are mean values ± s.d. of tritiated thymidine incorporation determined on sextuplate culture wells.

The lack of effects of CD28 activation on the proliferation/differentiation of myeloma cells may be due to the fact that myeloma cell lines are obtained only from patients with extramedullary proliferation.10 In patients with chronic disease, although IL-6 is still a tumor growth factor,19,20 additional signals provided by the tumoral environ- ment are necessary for the long-term proliferation of myeloma cells in vivo. As B7-1 molecules are present on stromal cells, activation of CD28 molecules might be one of these signals. Finally, the ability of myeloma CD28 to bind B7 molecules could block the antigen presenting capacity of myeloma cells and contribute to the escape of the tumoral clone from immune surveillance. Figure 7 Binding of p85 of PI-3 kinase to myeloma CD28 molecule. U266 or XG-6 myeloma cells were stimulated by B7-1-transfected L cells with or without 10 ␮g/ml of a B7-1mAb. The CD28- transfected murine T cell hybridoma DWT6.11 was used as positive Acknowledgements control. After stimulation, cells were lysed and immunoprecipitation was performed with the CD28.2 mAb. After SDS-PAGE and transfer This work was supported by grants from l’Association de to membrane, the membranes were incubated with anti-p85 rabbit Recherches sur le Cancer (Paris, France), la Ligue Nationale antibodies (upper panel) or CD28 mAb (CD28.6, lower panel). After extensive washes, the bound antibodies were revealed by anti-rabbit Française de Lutte contre le Cancer (Paris, France), L’INSERM (upper panel) or anti-mouse immunoglobulins (lower panel) coupled (CRE No. 920608), the European Community (Biomed No. to peroxidase and enhanced chemoluminescence. U266: lanes 1, 2 BMHI-CT93-1407). and 3; XG-6: lanes 4, 5 and 6; DWT6.11: lanes 7 and 8. Unstimulated cells, lanes 1 and 4; cells stimulated by B7-1 L cells, lanes 2 and 5; cells stimulated by B7-1 L cells and anti-B7-1 mAb, lanes 3 and 6; DWT6.11 stimulated by B7-1 L cells, lane 7. Lane 8 corresponds to References DWT6.11 total cell lysate. 1 Guinan EC, Gribben JG, Boussiotis VA, Freeman GJ, Nadler LM. Pivotal role of the B7:CD28 pathway in transplantation tolerance myeloma CD28, especially for U266 cells that expressed a and tumor immunity. Blood 1994; 84: 3261–3282. large density of CD28. No spontaneous PI-3 kinase binding to 2 June CH, Bluestone JA, Nadler LM, Thompson CB. The B7 and CD28 receptor families. Immunol Today 1994; 15: 321–331. CD28 was found and this also rules out a detectable autocrine 3 Truitt KE, Hicks CM, Imboden JB. Stimulation of CD28 triggers an activation of CD28 molecules by the endogenous B7-2 mol- association between CD28 and phosphatidylinositol 3-kinase in ecules present on myeloma cells. Jurkat T cells. J Exp Med 1994; 179: 1071–1076. Functional CD28 on human plasma cells X-G Zhang et al 618 4 Prasad KV, Cai YC, Raab M, Duckworth B, Cantley L, Shoelson 13 Gu ZJ, Wijdenes J, Zhang XG, Hallet MM, Clement C, Klein B. SE, Rudd CE. T-cell antigen CD28 interacts with the lipid kinase Anti-gp130 transducer monoclonal antibodies specifically phosphatidylinositol 3-kinase by a cytoplasmic Tyr(P)-Met-Xaa- inhibiting ciliary neurotrophic factor, interleukin-6, interleukin-11, Met motif. Proc Natl Acad Sci USA 1994; 91: 2834–2838. leukemia inhibitory factor or oncostatin M. J Immunol Meth 1996; 5 August A, Dupont B. CD28 of T lymphocytes associates with phos- 190: 21–27. phatidylinositol 3-kinase. Int Immunol 1994; 6: 769–774. 14 Liautard J, Gaillard JP, Mani JC, Montero JF, Duperray C, Lu ZY, 6 Pages F, Ragueneau M, Rottapel R, Truneh A, Nunes J, Imbert J, Jourdan M, Klein B, Brailly H, Brochier J. Epitope analysis of Olive D. Binding of phosphatidylinositol-3-OH kinase to CD28 is human IL-6 receptor gp80 molecule with monoclonal antibodies. required for T-cell signalling (published erratum appears in Nature Eur Cytokine Netw 1994; 5: 293–300. 1994; 370: 157). Nature 1994; 369: 327–329. 15 Lu ZY, Zhang XG, Wijdenes J, Morel-Fournier B, Harousseau JL, 7 Rudd CE. Upstream–downstream: CD28 cosignaling pathways Bataille R, Rossi JF, Klein B. Interleukin-10 is a growth factor for and T cell function. Immunity 1996; 4: 527–534. human myeloma cells. Blood 1995; 85: 2521–2527. 8 Boise LH, Minn AJ, Noel PJ, June CH, Accavitti MA, Lindsten T, 16 Kuiper HM, Brouwer M, Linsley PS, van Lier RA. Activated T cells Thompson CB. CD28 costimulation can promote T cell survival can induce high levels of CTLA-4 expression on B cells. J Immunol by enhancing the expression of Bcl-XL. Immunity 1995; 3: 87–98. 1995; 155: 1776–1783. 9 Kozbor D, Moretta A, Messner HA, Moretta L, Croce CM. Tp44 17 Merville P, Dechanet J, Desmouliere A, Durand I, Debouteiller O, molecules involved in antigen-independent T cell activation are Garrone P, Banchereau J, Liu YJ. Bcl-2(+) tonsillar plasma cells expressed on human plasma cells. J Immunol 1987; 138: 4128– 4132. are rescued from apoptosis by bone marrow fibroblasts. J Exp Med 10 Zhang XG, Gaillard JP, Robillard N, Lu ZY, Gu ZJ, Jourdan M, 1996; 183: 227–236. Boiron JM, Bataille R, Klein B. Reproducible obtaining of human 18 Klein B, Zhang XG, Lu ZY, Bataille R. Interleukin-6 on human myeloma cell lines as a model for tumor stem cell study in human multiple myeloma. Blood 1995; 85: 863–872. multiple myeloma. Blood 1994; 83: 3654–3663. 19 Klein B, Zhang XG, Jourdan M, Content J, Houssiau F, Aarden L, 11 Pellat Deceunynck C, Bataille R, Robillard N, Harousseau JL, Rapp Piechaczyk M, Bataille R. Paracrine rather than autocrine regu- MJ, Juge Morineau N, Wijdenes J, Amiot M. Expression of CD28 lation of myeloma-cell growth and differentiation by interleukin- and CD40 in human myeloma cells: a comparative study with 6. Blood 1989; 73: 517–526. normal plasma cells. Blood 1994; 84: 2597–2603. 20 Kawano M, Hirano T, Matsuda T, Taga T, Horii Y, Iwato K, Asaoka ` 12 Nunes J, Klasen M, Ragueneau C, Pavon D, Couez D, Mawas C, H, Tang B, Tanabe O, Tanaka H, Kuramoto A, Kishimoto T. Olive D. CD28 mAbs with distinct binding properties differ in their Autocrine generation and essential requirement of BSF-2/IL-6 for ability to induce T cell activation: analysis of early and late acti- human multiple myeloma. Nature 1988; 332: 83–85. vation events. Int Immunol 1993; 5: 311–318.