Costimulatory Pathways Functions of NKT Cells by CD28 and CD40 Differential Regulation of Th1 And

Differential Regulation of Th1 and Th2 Functions of NKT Cells by CD28 and CD40 Costimulatory Pathways1

Yoshihiro Hayakawa,* Kazuyoshi Takeda,2†‡ Hideo Yagita,†‡ Luc Van Kaer,§ Ikuo Saiki,* and Ko Okumura†‡

V␣14 NKT cells produce large amounts of IFN-␥ and IL-4 upon recognition of their specific ligand ␣-galactosylceramide (␣- GalCer) by their invariant TCR. We show here that NKT cells constitutively express CD28, and that blockade of CD28-CD80/ CD86 interactions by anti-CD80 and anti-CD86 mAbs inhibits the ␣-GalCer-induced IFN-␥ and IL-4 production by splenic V␣14 NKT cells. On the other, the blockade of CD40-CD154 interactions by anti-CD154 mAb inhibited ␣-GalCer-induced IFN-␥ production, but not IL-4 production. Consistent with these findings, CD28-deficient mice showed impaired IFN-␥ and IL-4 production in response to ␣-GalCer stimulation in vitro and in vivo, whereas production of IFN-␥ but not IL-4 was impaired in CD40-deficient mice. Moreover, ␣-GalCer-induced Th1-type responses, represented by enhanced cytotoxic activity of splenic or hepatic mononuclear cells and antimetastatic effect, were impaired in both CD28-deficient mice and CD40-deficient mice. In contrast, ␣-GalCer-induced Th2-type responses, represented by serum IgE and IgG1 elevation, were impaired in the absence of the CD28 costimulatory pathway but not in the absence of the CD40 costimulatory pathway. These results indicate that CD28- CD80/CD86 and CD40-CD154 costimulatory pathways differentially contribute to the regulation of Th1 and Th2 functions of V␣14 NKT cells in vivo. The Journal of Immunology, 2001, 166: 6012–6018.

atural killer T cells, which include heterogeneous pop- production of IFN-␥ by ␣-GalCer-activated V␣14 NKT cells, ulations, represent a novel lymphoid lineage distinct which requires IL-12 production by DC (15). On the other hand, N from conventional T cells, B cells, or NK cells (1, 2). V␣14 NKT cells also produce large amounts of IL-4 in the primary The TCR␣␤ expressed on the majority of NKT cells consists of a response and have been considered to play a role for the develop- single invariant V␣14-J␣281 chain paired preferentially with ment of Th2 responses (9, 10, 16–19). Since IL-4 and IFN-␥ have V␤8.2, V␤2, or V␤7, and recognizes glycolipid Ags or particular opposite effects on Th1/Th2 development, the role for V␣14 NKT hydrophobic peptides presented by the MHC class Ib molecule cells in the regulation of immune responses remains controversial CD1d (1–3). Although the physiological Ags for NKT cells still (9–11). In the present study, we examined the involvement of remain unclear, ␣-galactosylceramide (␣-GalCer),3 a glycolipid CD28- and CD40-mediated costimulatory pathways in IL-4 and derived from a marine sponge, has been identified to act as a spe- IFN-␥ production by ␣-GalCer-stimulated V␣14 NKT cells in by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd. cific ligand for V␣14 NKT cells (4–6). It has been reported that vitro and in vivo. We found differential contributions of these co- ␣-GalCer selectively stimulates V␣14 NKT cells to rapidly pro- stimulatory pathways to IL-4 and IFN-␥ production by V␣14 NKT duce large amounts of IFN-␥ and IL-4 and to exhibit cytotoxic and cells. Selective manipulation of V␣14 NKT cell functions by antitumor activities (7, 8). Moreover, ␣-GalCer-induced V␣14 ␣-GalCer and the blockade of costimulatory pathways, which can NKT cell activation secondarily resulted in the induction and mod- potentially modulate systemic immune responses, is discussed. ulation of innate (NK cell) and adaptive (T cell and B cell) immune responses (9–13). The presentation of ␣-GalCer by CD1d ex- Materials and Methods pressed on certain APC, especially dendritic cells (DC), efficiently Mice induced V␣14 NKT cell activation (3, 4, 7, 14). It has been re- http://classic.jimmunol.org ported that CD40-CD154 interactions are critically involved in the Male C57BL/6 (B6) wild-type mice were purchased from Clear Japan (To- kyo, Japan). B6 CD28-deficient (CD28Ϫ/Ϫ) mice were originally purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained in our animal facility. B6 CD1-deficient (CD1Ϫ/Ϫ) mice were generated as *Department of Pathogenic Biochemistry, Research Institute of Natural Medicine, previously described (20). B6 CD40-deficient (CD40Ϫ/Ϫ) mice were Toyama Medical and Pharmaceutical University, Toyama, Japan; †Department of Immunology, Juntendo University School of Medicine, Tokyo, Japan; ‡Core Research kindly provided by H. Kikutani (Osaka University, Osaka, Japan) (21). All for Evolutional Science and Technology of Japan Science and Technology Corpora- mice were maintained under specific pathogen-free conditions and used at § Downloaded from tion, Tokyo, Japan; and Department of Microbiology and Immunology, Howard 6–7 wk of age. Hughes Medical Institute, Vanderbilt University School of Medicine, Nashville, TN 37232 Reagents Received for publication December 4, 2000. Accepted for publication March 7, 2001. ␣-GalCer [(2S,3S,4R)-1-o-(␣-D-galactopyranosyl)-2-(N-hexacosanoylamino)- The costs of publication of this article were defrayed in part by the payment of page 1,3,4-octadecanetiol] was provided by Y. Koezuka and K. Motoki (Kirin charges. This article must therefore be hereby marked advertisement in accordance Brewery, Gumma, Japan) and was prepared as described previously (4, 8). with 18 U.S.C. Section 1734 solely to indicate this fact. Purified mAbs (no azide/low endotoxin grade) against mouse CD86 (PO. 3), 1 This work was supported by a Grant-in Aid for Scientific Research from the Min- CD154 (MR1), and IL-12 (C17.8) and control hamster IgG (A19-4) were istry of Education, Science, and Culture, Japan. purchased from PharMingen (San Diego, CA). Control rat IgG was purchased 2 Address correspondence and reprint requests to Dr. Kazuyoshi Takeda, Department from Sigma (St. Louis, MO). The hybridoma-producing anti-mouse CD154 of Immunology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku, mAb (MR1) was obtained from American Type Culture Collection (Manassas, Tokyo 113-8421, Japan. E-mail address: [email protected] VA) (22). The hybridomas producing anti-mouse CD80 mAb (RM80) and 3 Abbreviations used in this paper: ␣-GalCer, ␣-galactosylceramide; MNC, mononu- anti-mouse CD86 mAb (PO.3) were established in our laboratory (23). The clear cells; DC, dendritic cells. mAbs were prepared from these hybridoma as described previously (23).

Flow cytometric analysis Cytotoxic assay Surface phenotype of the cells was characterized by three-color flow cy- Cytolytic activity was assessed against NK-susceptible YAC-1 target cells tometry as previously described (24). Briefly, 1 ϫ 106 cells were first and NK- and Fas ligand-resistant B16-BL6 cells by a standard 51Cr release preincubated with anti-CD16/32 (2.4G2) mAb to avoid the nonspecific assay as previously described (26, 27). Both target cells were cultured in binding of Abs to Fc␥R. Then the cells were incubated with a saturating RPMI 1640 medium containing 10% FCS, 2 mM L-glutamine, and 25 mM

amount of biotinylated isotype-matched control mAbs (Ha4/8, A19-3, NaHCO3. As effector cells, hepatic and splenic MNC were isolated from G235-2356, or R3-34), anti-CD28 (37.51), anti-CD152/CTLA-4 (UC10- the mice 24 h after i.p. injection of 2 ␮g/200 ␮lof␣-GalCer or 200 ␮lof 4F10-11), anti-CD137/4-1BB (1AH2), anti-CD134/OX40 (OX86), anti- the vehicle (0.5% polysorbate 20). Target cells (106) were labeled with 100 ␮ 51 CD27 (LG.3A10), anti-CD30 (mCD30.1), and anti-CD154/CD40L (MR1) Ci/ml Na2 CrO4 for 60 min at 37°C in RPMI 1640 medium containing mAb before incubation with FITC-conjugated anti-NK1.1 (PK136) mAb, 10% FCS. Labeled target cells (104/well) were incubated in a total volume Cy-Chrome-conjugated anti-CD3⑀ mAb (145-2C111), and PE-conjugated of 200 ␮l with effector cells in 10% FCS-RPMI 1640 in 96-well U-bottom streptavidin. All staining reagents were obtained from PharMingen. After plates. The plates were centrifuged before incubation, and after4hthe washing with PBS, the stained cells were analyzed on a FACSCalibur supernatant was harvested and counted in a gamma counter. Specific lysis (Becton Dickinson, San Jose, CA). was calculated as previously described (26, 27). In vitro stimulation with ␣-GalCer Experimental lung metastasis Splenic mononuclear cells (MNC, 5 ϫ 105) were cultured with 100 ng/ml Log-phase cell cultures of B16-BL6 were harvested with 1 mM EDTA in ␣-GalCer or vehicle (0.1% DMSO) as a control in RPMI 1640 medium PBS, washed three times with serum-free RPMI 1640, and resuspended to 4 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, and 25 appropriate concentrations in PBS. B16-BL6 cells (5 ϫ 10 /100 ␮l) were ␣ ␮ ␮ mM NaHCO3 in humidified 5% CO2 at 37°C in 96-well U-bottom plates injected i.v. into syngeneic B6 mice, and then -GalCer (2 g/200 l) or 200 (Costar, Cambridge, MA). In the blocking experiments, anti-CD80 ␮l of vehicle was i.p. administered on days 0, 4, and 8. On day 14, the number (RM80), anti-CD86 (PO.3), anti-CD154 (MR1), anti-IL-12 (C17.8), and of tumor colonies in the lung was counted under a dissecting microscope. isotype-matched control mAbs were added at 10 ␮g/ml each in the culture. After incubation for 72 h, the cell-free culture supernatants were harvested Statistical analysis to detect cytokine levels by ELISA. Data were analyzed using a two-tailed Student t test. All p values Ͻ 0.05 ELISA were considered as significant.

IFN-␥ and IL-4 levels in the culture supernatants and serum were evaluated Results using specific ELISA kits (Endogen, Boston, MA) according to the man- Constitutive expression of CD28 on NKT cells ufacturer’s instructions. For serum IgG1- or IgG2a-specific ELISA, micro- To examine the expression of costimulatory receptors on NKT titer plates (Immulon 2HB, 96-well; Dynex Technologies, Chantilly, VA) were coated with monoclonal anti-IgG1 or anti-IgG2a (PharMingen) at 10 cells, freshly isolated hepatic MNC from B6 mice were subjected ␮g/ml in PBS overnight at 4°C. The plates were blocked with PBS con- to three-color staining with biotin-conjugated mAb against CD28, taining 1% BSA for 1 h and washed extensively with 0.05% Tween 20 in CD152 (CTLA-4), CD27, CD30, CD134 (OX40), or CD137 (4- PBS. Serial dilutions of serum samples were incubated for2hat37°C. The 1BB), followed by FITC-conjugated anti-NK1.1 mAb, Cy- plates were then washed with 0.05% Tween 20 in PBS and overlaid with Chrome-conjugated anti-CD3 mAb, and PE-conjugated streptavi- biotin-conjugated isotype-specific mAbs, including anti-mouse IgG1 (Se- din. Then the expression of each receptor was analyzed on rotec, Oxford, U.K.) and IgG2a (PharMingen), washed, and then developed ϩ ϩ Ϫ ϩ with a Vectastain ABC kit (Vector Laboratories, Burlingame, CA) and electronically gated NK1.1 CD3 (NKT), NK1.1 CD3 (T), or o-phenylendiamine (Wako Pure Chemical, Osaka, Japan). After termina- NK1.1ϩ CD3Ϫ (NK) cells as represented in Fig. 1A. The vast

by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd. tion of the reaction with2NH2SO4, OD at 490/595 nm was measured on majority of NKT cells exhibited high levels of CD28 expression, a microplate reader (Bio-Rad, Hercules, CA). Concentrations were calculated on the basis of standard curves of Ab isotypes (PharMingen) run in which was equivalent to that on conventional T cells (Fig. 1B). parallel ELISA. Total serum IgE was quantitated by IgE-speciﬁc sandwich None of the other molecules were expressed by the cell types that ELISA as previously described (25). were examined. CD152 was not detected in NKT cells even by the http://classic.jimmunol.org Downloaded from

FIGURE 1. Constitutive expression of CD28 but not other costimulatory receptors on NKT cells. Freshly isolated hepatic MNC were stained with biotinylated mAbs against the indicated molecules or with control IgG, followed by PE-conjugated streptavidin, FITC-conjugated anti-NK1.1 mAb, and Cy-Chrome-conjugated anti-CD3 mAb. Expression of the respective molecules was analyzed by ﬂow cytometry on electronically gated CD3ϩ NK1.1ϩ NKT, CD3ϩ NK1.1Ϫ T, or CD3Ϫ NK1.1ϩ NK cells (boxed in A). B, Solid lines indicate the staining with respective mAb, and dotted lines indicate the staining with isotype-matched control IgG. 6014 DIFFERENTIAL REGULATION OF NKT CELL FUNCTIONS

intracellular staining. Although most NK cells and T cells ex- fold (Fig. 2A). The combination of these two mAbs did not result pressed CD27, NKT cells did not express CD27. Similar results in further inhibition of IFN-␥ production, suggesting that the con- were obtained for splenic NKT cells or hepatic and splenic NKT tribution of CD154 to IFN-␥ production was mediated by IL-12 as cells isolated at 3 h after ␣-GalCer injection (data not shown). previously reported (15, 28). When combined with anti- These results indicated a unique expression profile of costimula- CD80/CD86 mAbs, anti-IL-12 further inhibited IFN-␥ production, tory receptors on NKT cells, which was apparently distinct from whereas anti-CD154 mAb did not. This suggested that the CD80/ conventional T cells and NK cells. CD86-independent IFN-␥ production, at least in part, was IL-12 dependent but CD154 independent. Notably, anti-CD80/CD86 Involvement of CD28-CD80/CD86 and CD154-CD40 mAbs completely inhibited IL-4 production even in the presence costimulatory pathways in IFN-␥ and IL-4 production by of anti-IL-12 or anti-CD154 mAb. Taken together, these results ␣-GalCer-stimulated NKT cells indicated that the ␣-GalCer-induced IFN-␥ production by V␣14 We next investigated whether V␣14 NKT cells require CD28-me- NKT cells was mostly dependent on both CD28-CD80/CD86 and diated costimulation for IFN-␥ and IL-4 production in response to CD154/CD40 costimulatory pathways, whereas IL-4 production their specific ligand, ␣-GalCer. As shown in Fig. 2A, high levels of was absolutely dependent on the CD28-CD80/CD86 pathway but IFN-␥ and IL-4 were detected in the supernatant of splenic MNC instead suppressed by the CD154/CD40 pathway. when cultured with ␣-GalCer for 72 h. These cytokines were not detected when splenic MNC from CD1-deficient mice were cul- ␣-GalCer-induced IFN-␥ production is impaired in both CD28- tured with ␣-GalCer, indicating the dependence on CD1-restricted and CD40-deficient mice, but IL-4 production is impaired in NKT cells as previously reported (4, 12) (Fig. 2A). The blockade CD28-deficient mice only of CD28-mediated costimulation by anti-CD80 and anti-CD86 mAbs resulted in a marked but partial inhibition of IFN-␥ produc- To confirm the differential contribution of CD28- and CD40-me- tion and an almost complete inhibition (Ͼ95% in three experi- diated costimulatory pathways to IFN-␥ and IL-4 production by ments) of IL-4 production (Fig. 2A). Since it has been reported that V␣14 NKT cells, we next investigated the ␣-GalCer-induced IL-12 produced by DC plays a critical role in the ␣-GalCer-in- IFN-␥ and IL-4 production by using splenic MNC from CD28- or duced IFN-␥ production (15) and that CD40-CD154 interactions CD40-deficient mice. As represented in Fig. 3, splenic MNC from are required for the ␣-GalCer-induced IL-12 production (28), we CD28-deficient mice showed greatly impaired production of both also investigated the effect of blocking mAbs against CD154 and IFN-␥ and IL-4 as compared with those from wild-type mice. IL-12. Rapid induction of CD154 on ␣-GalCer-activated NKT Splenic MNC from CD40-deficient mice showed similarly im- cells was confirmed in vitro (Fig. 2B) as previously reported (28). paired IFN-␥ production but intact IL-4 production, which was Anti-CD154 mAb alone or anti-IL-12 mAb alone inhibited IFN-␥ comparable to wild-type mice (Fig. 3). We also examined IFN-␥ production and conversely increased IL-4 production by 2- to 2.5- and IL-4 production in vivo by administrating ␣-GalCer into by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from

FIGURE 2. A, Distinct effects of anti-CD80/CD86 and anti-CD154 mAbs on IFN-␥ and IL-4 production by ␣-GalCer-stimulated splenic MNC. Freshly isolated splenic MNC (5 ϫ 105 cells) from naive B6 mice (wild-type) or CD1-deﬁcient mice (CD1 Ϫ/Ϫ) were cultured with ␣-GalCer (100 ng/ml) or vehicle (0.1% DMSO) in 96-well U-bottom plates. Anti-CD80 and anti-CD86 mAbs, anti-CD154 mAb, anti-IL-12 mAb, and/or control IgG were added at 10 ␮g/ml each into the culture. After 72 h, culture supernatants were harvested and the levels of IFN-␥ and IL-4 were determined by ELISA. Data are represented p Ͻ 0.01. B, Rapid induction of CD154 on ,ء .as mean Ϯ SD of triplicate wells. Similar results were obtained in three independent experiments ␣-GalCer-stimulated NKT cells. Expression of CD154 was analyzed by ﬂow cytometry on electronically gated CD3ϩ NK1.1ϩ NKT cells in splenic MNC at 3 h after stimulation with ␣-GalCer (100 ng/ml) or vehicle (0.1% DMSO). Solid lines indicate the staining with anti-CD154 mAb, and dotted lines indicate the staining with isotype-matched control IgG. The Journal of Immunology 6015

FIGURE 4. Serum IFN-␥ and IL-4 levels in CD28Ϫ/Ϫ or CD40Ϫ/Ϫ mice after ␣-GalCer treatment. Serum samples were obtained from wild- FIGURE 3. ␣-GalCer-induced IFN-␥ and IL-4 production by splenic Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ 5 type, CD28 , or CD40 mice at 3 and 16 h after i.p. injection of MNC from CD28 or CD40 mice. Splenic MNC (5 ϫ 10 cells) freshly ␮ ␣ ␮ ␮ ␥ Ϫ/Ϫ Ϫ/Ϫ vehicle (200 l) or -GalCer (2 g/200 l). IFN- and IL-4 levels in the isolated from wild-type, CD28 , or CD40 mice were stimulated with serum were determined by ELISA. Data are represented as the mean Ϯ SD ␣ -GalCer (100 ng/ml) in 96-well U-bottom plates. After 72 h, culture super- of ﬁve mice in each group. Serum IFN-␥ and IL-4 in the vehicle-injected ␥ natants were harvested and the levels of IFN- and IL-4 were determined by mice were not detectable (data not shown). Similar results were obtained in Ϯ .p Ͻ 0.01 ,ء .ELISA. Data are represented as the mean SD of triplicate wells. Similar two independent experiments .p Ͻ 0.01 ,ء .results were obtained in three independent experiments by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd.

CD28- or CD40-deficient mice (Fig. 4). In preliminary experiments, we found that serum IFN-␥ levels peaked at 16 h and serum ported (8), i.p. injection of ␣-GalCer into wild-type mice induced IL-4 peaked at 3 h after i.p. injection of ␣-GalCer to wild-type substantial cytotoxic activities of splenic and hepatic MNC against mice (data not shown). The serum IFN-␥ elevation at 16 h after both NK-susceptible YAC-1 and NK-resistant B16-BL6 target ␣-GalCer administration was significantly impaired in both CD28- cells (Fig. 5). These cytotoxic activities were diminished by NK 4 or CD40-deficient mice (Fig. 4). The serum IL-4 elevation at 3 h cell depletion by anti-asialo GM1 Ab administration (data not after ␣-GalCer administration was abrogated in CD28-deficient shown). In contrast, such an ␣-GalCer-induced cytotoxic activity

http://classic.jimmunol.org mice but rather enhanced in CD40-deficient mice (Fig. 4). No ap- was not observed in splenic or hepatic MNC from CD28- or parent shift in the kinetics of serum IFN-␥ and IL-4 levels was CD40-deficient mice (Fig. 5). Then we examined the antimeta- observed in these mutant mice (data not shown). These results static effect of ␣-GalCer in an experimental lung metastasis model indicated that CD28-mediated costimulation was required for both of the B16-BL6 melanoma, which is mediated by V␣14 NKT cells IFN-␥ and IL-4 production by V␣14 NKT cells, whereas CD40- (8). As previously reported (8), ␣-GalCer administration greatly mediated costimulation was only required for IFN-␥ production reduced the lung metastasis of B16-BL6 melanoma cells in wild- type mice (Fig. 6). In contrast, no significant antimetastatic effect

Downloaded from both in vitro and in vivo. of ␣-GalCer was observed in CD28-, CD40- (Fig. 6), or IFN-␥- Impairment of ␣-GalCer-induced cytolytic activity and deficient mice (data not shown).4 These results indicated that both antimetastatic effect in CD28- and CD40-deficient mice CD28- and CD40-mediated costimulatory pathways were required ␣ We next investigated the effect of CD28 or CD40 deficiency on for the Th1-like functions of V 14 NKT cells in vivo. ␣-GalCer-induced cytotoxic activity of splenic and hepatic MNC, which represents the Th1-like function of NKT cells, since we Involvement of CD28-CD80/CD86 and CD154-CD40 have found that ␣-GalCer-induced cytotoxicity and antimetastatic costimulatory pathways in ␣-GalCer-induced Th2-like functions activity were dependent on the IFN-␥ produced by ␣-GalCer-ac- of NKT cells in vivo tivated V␣14 NKT cells and IFN-␥-activated NK cells.4 As re- It has been also reported that ␣-GalCer administration biases the subsequent immune responses toward Th2 type, as repre- 4 Y. Hayakawa, K. Takeda, H. Yagita, S. Kakuta, Y. Iwakura, L. V. Kaer, I. Saiki, and sented by enhanced IgE and IgG1 production, which is medi- K. Okumura. Critical contribution of IFN-␥ and NK cells, but not perforin-mediated ␣ cytotoxicity, to the antimetastatic activities of ␣-galactosylceramide. Submitted for ated by IL-4 secreted from V 14 NKT cells (9, 10). We there- publication. fore investigated the contribution of CD28 and CD40 6016 DIFFERENTIAL REGULATION OF NKT CELL FUNCTIONS

FIGURE 6. Impairment of ␣-GalCer-induced antimetastatic effect in CD28Ϫ/Ϫ or CD40Ϫ/Ϫ mice. B16-BL6 (5 ϫ 104/100 ␮l) were i.v. injected into wild-type, CD28Ϫ/Ϫ, or CD40Ϫ/Ϫ mice. ␣-GalCer (2 ␮g/200 ␮l) or vehicle (200 ␮l) was i.p. injected three times on days 0, 4, and 8. On day 14, the number of tumor colonies in the lung was counted under a dissecting microscope. Data are represented as the mean Ϯ SD of ﬁve mice in each group. Similar results were obtained in two independent experiments. .p Ͻ 0.01 ,ء

Discussion In this study, we demonstrated that murine NKT cells constitutively express CD28 and that CD28-mediated costimulation is required for production of both IFN-␥ and IL-4 by V␣14 NKT cells in response to their specific ligand ␣-GalCer. Consequently, blockade of the CD28-mediated costimulation resulted in impairment of both Th1- and Th2-type responses (serum IFN-␥ and IL-4 elevation, cytotoxicity induction, antimetastatic effect, and serum IgE/ FIGURE 5. Impairment of ␣-GalCer-induced cytotoxicity in CD28Ϫ/Ϫ IgG1 elevation) induced by ␣-GalCer administration in vivo. In or CD40Ϫ/Ϫ mice. Wild-type, CD28Ϫ/Ϫ, or CD40Ϫ/Ϫ mice were i.p. in- contrast, blockade of the CD40-CD154 interaction inhibited only jected with vehicle (200 ␮l) or ␣-GalCer (2 ␮g/200 ␮l). Hepatic and the ␣-GalCer-induced Th1-type responses (serum IFN-␥ elevation, by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd. splenic MNC were prepared 24 h later, and the cytotoxicity against YAC-1 cytotoxicity induction, and antimetastatic effect) but rather en- and B16-BL6 was tested by 51Cr release assay at the indicated E:T ratios. hanced the Th2-type responses (serum IL-4 elevation and serum Ϯ Data are represented as the mean SD of triplicate wells. Similar results IgE/IgG1 elevation). These results indicate that Th1- and Th2-like were obtained in three independent experiments. functions of V␣14 NKT cells are differentially regulated by CD28- and CD40-mediated costimulatory pathways. It has been well established that conventional T cells require a costimulatory pathways to the development of ␣-GalCer-in- costimulatory signal, in addition to Ag-specific TCR-mediated sig- duced Th2-type immune responses in vivo. As shown in Fig. nal, for their full activation (29). Such a costimulatory signal can 7A, administration of ␣-GalCer significantly increased the se- be commonly transmitted by CD28 or some members of the TNF http://classic.jimmunol.org rum IgE and IgG1 levels and conversely reduced the IgG2a receptor superfamily, including CD27, CD30, CD134, and CD137, level in wild-type mice as previously reported (10). In contrast, in conventional T cells (29–34). It was also reported that CD161 no significant elevation of serum IgE and IgG1 levels or reduc- transmitted a costimulatory signal into V␣24 NKT cells (35). We tion of the IgG2a level was observed in CD28-deficient mice, here showed that NKT cells also express CD28, which played a indicating that the Th2-like function of ␣-GalCer-stimulated critical role in full activation of V␣14 NKT cell functions as man- V␣14 NKT cells required CD28-mediated costimulation. In ifested by IFN-␥ and IL-4 production. It has been also reported that Downloaded from CD40-deficient mice, administration of ␣-GalCer induced mar- engagement of CD28 on NK cells promoted their proliferation, ginal but significant elevation of serum IgE and IgG1 levels, IFN-␥ production, and cytolytic activity (36, 37). Therefore, CD28 suggesting that CD40 was not essential for this response. The appears to play important roles not only in adaptive immunity impaired elevation of serum IgE and IgG1 levels in CD40-de- mediated by conventional T cells but also in innate immunity me- ficient mice as compared with wild-type mice appeared to result diated by NK and NKT cells. It was notable that NKT cells did not from the defect in Ig class switching of CD40-deficient B cells express CD27, which has been implicated in activation of T cells in a later stage of the response. To minimize the effect of CD40 and NK cells (38, 39) and T cell differentiation (40). This may be deficiency on the later stage, we administered anti-CD154 mAb due to the unique ontogeny of NKT cells, which is distinct from only once before the administration of ␣-GalCer into wild-type conventional T cells (1, 2, 41). mice. As shown in Fig. 7B, anti-CD154 mAb significantly en- It has been shown that IFN-␥ production by V␣14 NKT cells in hanced the elevation of serum IgE and IgG1 levels after ␣-Gal- response to ␣-GalCer is predominantly mediated by IL-12 pro- Cer administration. This suggested that the CD154-CD40 inter- duced by DC and requires CD154-CD40 interaction (15). Our action played a suppressive role in the ␣-GalCer-induced and present observations that anti-IL-12 mAb or anti-CD154 mAb V␣14 NKT cell-mediated Th2-type response in vivo. alone strongly inhibited the ␣-GalCer-induced IFN-␥ production The Journal of Immunology 6017

FIGURE 7. Involvement of CD28-CD80/CD86 and CD40-CD154 in ␣-GalCer-induced serum Ig responses. A, Serum Ig isotype levels in ␣-GalCer- treated wild-type, CD28Ϫ/Ϫ, or CD40Ϫ/Ϫ mice. Wild-type, CD28Ϫ/Ϫ, or CD40Ϫ/Ϫ mice were i.p. injected with vehicle (200 ␮l) or ␣-GalCer (2 ␮g/200 ␮l). B, Effect of anti-CD154 mAb on ␣-GalCer-induced serum Ig isotype levels. Wild-type mice were i.p. injected with 300 ␮g of anti-CD154 mAb or control IgG 3 h before the i.p. injection of vehicle (200 ␮l) or ␣-GalCer (2 ␮g/200 ␮l). In both experiments, mice were bled on day 7, and total IgE, IgG1, and IgG2a levels in the serum were measured by ELISA. Data are represented as mean Ϯ SD of ﬁve mice in each group. Similar results were obtained .p Ͻ 0.01 ,ء .in two independent experiments

and that these mAbs did not exhibit an additive inhibitory effect As represented in the present study, V␣14 NKT cells have been (Fig. 2A) are consistent with this notion. However, we also ob- shown to produce both Th1-type (IFN-␥) and Th2-type (IL-4) cy- served that the combination of anti-CD80/CD86 mAbs with anti- tokines upon stimulation with a specific ligand (␣-GalCer) or anti- IL-12 mAb additively inhibited the IFN-␥ production. This sug- CD3 mAb (48, 49). Some recent studies have shown that ␣-GalCer gested that CD28-CD80/CD86 interactions regulate IFN-␥ treatment polarizes bystander immune responses toward a Th2 production by NKT cells in an IL-12-independent manner at least phenotype possibly through IL-4 production by V␣14 NKT cells partly. This CD28-mediated IFN-␥ production might be directly (9, 10), whereas another study reported polarization toward a Th1 induced by transcriptional regulation of the IFN-␥ gene by CD28- phenotype through IFN-␥ production by V␣14 NKT cells (11). mediated signals as demonstrated in conventional T cells (42). This discrepancy in the effects of ␣-GalCer might result from dif- Moreover, the CD28-mediated pathway and the CD40/IL-12-me- ferences in dose, timing, and route of ␣-GalCer administration. by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd. diated pathway could interact mutually, since CD28-mediated co- Our present study suggests that such a Th1- or Th2-polarizing stimulation stabilizes expression of CD154 on T cells (43, 44), and function of V␣14 NKT cells can be selectively modulated by the CD40-mediated activation up-regulates CD80/CD86 expression blockade of costimulatory pathways, as represented by polariza- on DC (45). This explains why the IFN-␥ production by ␣-GalCer- tion toward a Th2 phenotype by the blockade of CD154-CD40 stimulated V␣14 NKT cells was mostly dependent on both CD28- interaction, at ␣-GalCer administration. Since NKT cells have and CD40-mediated costimulatory pathways. been implicated in innate immunity against pathogens (50), anti- In the present study, we observed that the induction of cytotoxic tumor responses (8, 51, 52), liver damage (53), and autoimmune activity in liver or splenic MNC and the antimetastatic effect of diseases (54–57), selective modulation of their functions with spe- ␣-GalCer were abolished in both CD28- and CD40-deficient mice cific ligand and costimulatory blockade may be useful for prophy- http://classic.jimmunol.org (Figs. 5 and 6). This paralleled with the impairment of IFN-␥ pro- laxis and therapy of such diseases. duction in these mice (Fig. 4). Both IL-12 and IL-4 have been implicated in cytolytic activation of V␣14 NKT cells (26, 27, 46). Acknowledgments In our present observation, however, both the cytolytic activation We thank Dr. Yasuhiko Koezuka and Kazuhiro Motoki (Pharmaceutical and antimetastatic effect were abolished in CD40-deficient mice, Research Laboratory, Kirin Brewery) for generously providing ␣-GalCer which exhibited rather increased serum IL-4 levels upon ␣-GalCer and Dr. Hisaya Akiba for technical assistance and helpful suggestions. Downloaded from ␣ administration. These results suggested that the -GalCer-induced References cytolytic activity and antimetastatic effect were associated with the 1. Bendelac, A., M. N. Rivera, S. H. Park, and J. H. Roark. 1997. Mouse CD1- production of IFN-␥, but not IL-4, by V␣14 NKT cells. Consistent specific NK1 T cells: development, specificity, and function. Annu. Rev. Immu- with this notion, these responses were largely abolished in IFN- nol. 15:535. ϩ ␣ ␤ϩ ␥ 4 ␣ 2. MacDonald, H. R. 1995. NK1.1 T cell receptor- / cells: new clues to their -deficient mice. On the other hand, we observed that the -Gal- origin, specificity, and function. J. Exp. Med. 182:633. Cer-induced serum IgE and IgG1 elevation was abolished in 3. Brossay, L., N. Burdin, S. Tangri, and M. Kronenberg. 1998. Antigen-presenting CD28-deficient mice but not in CD40-deficient mice (Fig. 7A), function of mouse CD1: one molecule with two different kinds of antigenic li- gands. Immunol. Rev. 163:139. which paralleled with the serum IL-4 elevation in these mice (Fig. 4. Kawano, T., J. Cui, Y. Koezuka, I. Toura, Y. Kaneko, K. Motoki, H. Ueno, 4). Treatment with anti-CD154 mAb at ␣-GalCer administration R. Nakagawa, H. Sato, E. Kondo, et al. 1997. CD1d-restricted and TCR-mediated ␣ rather augmented this response (Fig. 7B). These results indicated activation of V 14 NKT cells by glycosylceramides. Science 278:1626. 5. Kobayashi, E., K. Motoki, T. Uchida, H. Fukushima, and Y. Koezuka. 1995. that Th2-like function of V␣14 NKT cells was totally dependent KRN7000, a novel immunomodulator, and its antitumor activities. Oncol. Res. on the CD28-mediated costimulation and rather suppressed by the 7:529. 6. Yamaguchi, Y., K. Motoki, H. Ueno, K. Maeda, E. Kobayashi, H. Inoue, CD40-mediated pathway, possibly due to a suppressive effect of H. Fukushima, and Y. Koezuka. 1996. Enhancing effects of (2S,3S,4R)-1-O-(␣-D- IFN-␥ on IgE/IgG1 production (47). galactopyranosyl)-2-(N-hexacosanoylamino)-1,3,4-octadecanetriol (KRN7000) on 6018 DIFFERENTIAL REGULATION OF NKT CELL FUNCTIONS

antigen-presenting function of antigen-presenting cells and antimetastatic activity of 31. Smith, C. A., T. Farrah, and R. G. Goodwin. 1994. The TNF receptor superfamily KRN7000-pretreated antigen-presenting cells. Oncol. Res. 8:399. of cellular and viral proteins: activation, costimulation, and death. Cell 76:959. 7. Burdin, N., L. Brossay, Y. Koezuka, S. T. Smiley, M. J. Grusby, M. Gui, 32. Harding, F. A., J. G. McArthur, J. A. Gross, D. H. Raulet, and J. P. Allison. 1992. M. Taniguchi, K. Hayakawa, and M. Kronenberg. 1998. Selective ability of CD28-mediated signalling co-stimulates murine T cells and prevents induction of mouse CD1 to present glycolipids: ␣-galactosylceramide specifically stimulates anergy in T-cell clones. Nature 356:607. ϩ V␣14 NKT lymphocytes. J. Immunol. 161:3271. 33. Tan, P., C. Anasetti, J. A. Hansen, J. Melrose, M. Brunvand, J. Bradshaw, 8. Kawano, T., J. Cui, Y. Koezuka, I. Toura, Y. Kaneko, H. Sato, E. Kondo, J. A. Ledbetter, and P. S. Linsley. 1993. Induction of alloantigen-specific hypo- M. Harada, H. Koseki, T. Nakayama, Y. Tanaka, and M. Taniguchi. 1998. Nat- responsiveness in human T lymphocytes by blocking interaction of CD28 with its ural killer-like nonspecific tumor cell lysis mediated by specific ligand-activated natural ligand B7/BB1. J. Exp. Med. 177:165. V␣14 NKT cells. Proc. Natl. Acad. Sci. USA 95:5690. 34. Gimmi, C. D., G. J. Freeman, J. G. Gribben, G. Gray, and L. M. Nadler. 1993. 9. Burdin, N., L. Brossay, and M. Kronenberg. 1999. Immunization with ␣-galac- Human T-cell clonal anergy is induced by antigen presentation in the absence of tosylceramide polarizes CD1-reactive NKT cells towards Th2 cytokine synthesis. B7 costimulation. Proc. Natl. Acad. Sci. USA 90:6586. Eur. J. Immunol. 29:2014. 35. Exley, M., S. Porcelli, M. Furman, J. Garcia, and S. Balk. 1998. CD161(NKR- 10. Singh, N., S. Hong, D. C. Scherer, I. Serizawa, N. Burdin, M. Kronenberg, P1A) costimulation of human T cell expressing invariant V␣24J␣Q T cell re- Y. Koezuka, and L. Van Kaer. 1999. Activation of NKT cells by CD1d and ceptor ␣ chains. J. Exp. Med. 188:867. ␣-galactosylceramide directs conventional T cells to the acquisition of a Th2 36. Nandi, D., J. A. Gross, and J. P. Allison. 1994. CD28-mediated costimulation is phenotype. J. Immunol. 163:2373. necessary for optimal proliferation of murine NK cells. J. Immunol. 152:3361. 11. Cui, J., N. Watanabe, T. Kawano, M. Yamashita, T. Kamata, C. Shimizu, 37. Walker, W., M. Aste-Amezaga, R. A. Kastelein, G. Trinchieri, and C. A. Hunter. M. Kimura, E. Shimizu, J. Koike, H. Koseki, et al. 1999. Inhibition of T helper 1999. IL-18 and CD28 use distinct molecular mechanisms to enhance NK cell cell type 2 cell differentiation and immunoglobulin E response by ligand-acti- production of IL-12-induced IFN-␥. J. Immunol. 162:5894. ␣ vated V 14 natural killer T cells. J. Exp. Med. 190:783. 38. Hintzen, R. Q., S. M. Lens, K. Lammers, H. Kuiper, M. P. Beckmann, and 12. Carnaud, C., D. Lee, O. Donnars, S. H. Park, A. Beavis, Y. Koezuka, and R. A. van Lier. 1995. Engagement of CD27 with its ligand CD70 provides a A. Bendelac. 1999. Cross-talk between cells of the innate immune system: NKT second signal for T cell activation. J. Immunol. 154:2612. cells rapidly activate NK cells. J. Immunol. 163:4647. 39. Takeda, K., H. Oshima, Y. Hayakawa, H. Akiba, M. Atsuta, T. Kobata, 13. Kitamura, H., A. Ohta, M. Sekimoto, M. Sato, K. Iwakabe, M. Nakui, T. Yahata, K. Kobayashi, M. Ito, H. Yagita, and K. Okumura. 2000. CD27-mediated acti- ␣ H. Meng, T. Koda, S. Nishimura, et al. 2000. -Galactosylceramide induces early vation of murine NK cells. J. Immunol. 164:1741. B-cell activation through IL-4 production by NKT cells. Cell. Immunol. 199:37. 40. Gravestein, L. A., W. van Ewijk, F. Ossendorp, and J. Borst. 1996. CD27 coop- 14. Hong, S., D. C. Scherer, N. Singh, S. K. Mendiratta, I. Serizawa, Y. Koezuka, and erates with the pre-T cell receptor in the regulation of murine T cell development. L. Van Kaer. 1999. Lipid antigen presentation in the immune system: lessons J. Exp. Med. 184:675. learned from CD1d knockout mice. Immunol. Rev. 169:31. 41. Makino, Y., N. Yamagata, T. Sasho, Y. Adachi, R. Kanno, H. Koseki, M. Kanno, 15. Kitamura, H., K. Iwakabe, T. Yahata, S. Nishimura, A. Ohta, Y. Ohmi, M. Sato, and M. Taniguchi. 1993. Extrathymic development of V␣14-positive T cells. K. Takeda, K. Okumura, L. Van Kaer, et al. 1999. The natural killer T (NKT) cell ␣ J. Exp. Med. 177:1399. ligand -galactosylceramide demonstrates its immunopotentiating effect by in- 42. Thompson, C. B., T. Lindsten, J. A. Ledbetter, S. L. Kunkel, H. A. Young, ducing interleukin (IL)-12 production by dendritic cells and IL-12 receptor ex- S. G. Emerson, J. M. Leiden, and C. H. June. 1989. CD28 activation pathway pression on NKT cells. J. Exp. Med. 189:1121. regulates the production of multiple T-cell-derived lymphokines/cytokines. Proc. 16. Le Gros, G., S. Z. Ben-Sasson, R. Seder, F. D. Finkelman, and W. E. Paul. 1990. Natl. Acad. Sci. USA 86:1333. Generation of interleukin 4 (IL-4)-producing cells in vivo and in vitro: IL-2 and 43. Ding, L., J. M. Green, C. B. Thompson, and E. M. Shevach. 1995. B7/CD28- IL-4 are required for in vitro generation of IL-4- producing cells. J. Exp. Med. dependent and -independent induction of CD40 ligand expression. J. Immunol. 172:921. 155:5124. 17. Swain, S. L., A. D. Weinberg, M. English, and G. Huston. 1990. IL-4 directs the 44. Johnson-Leger, C., J. Christensen, and G. G. Klaus. 1998. CD28 co-stimulation development of Th2-like helper effectors. J. Immunol. 145:3796. stabilizes the expression of the CD40 ligand on T cells. Int. Immunol. 10:1083. 18. Hsieh, C. S., A. B. Heimberger, J. S. Gold, A. O’Garra, and K. M. Murphy. 1992. 45. Caux, C., C. Massacrier, B. Vanbervliet, B. Dubois, C. Van Kooten, I. Durand, Differential regulation of T helper phenotype development by interleukins 4 and and J. Banchereau. 1994. Activation of human dendritic cells through CD40 10 in an ␣␤ T-cell-receptor transgenic system. Proc. Natl. Acad. Sci. USA 89: cross-linking. J. Exp. Med. 180:1263. 6065. 46. Kaneko, B. Y., M. Harada, T. Kawano, M. Yamashita, Y. Shibata, F. Gejyo, 19. Seder, R. A., W. E. Paul, M. M. Davis, and B. Fazekas de St. Groth. 1992. The ␣ presence of interleukin 4 during in vitro priming determines the lymphokine- T. Nakayama, and M. Taniguchi. 2000. Augmentation of V 14 NKT cell-me- producing potential of CD4ϩ T cells from T cell receptor transgenic mice. J. Exp. diated cytotoxicity by interleukin 4 in an autocrine mechanism resulting in the

by guest on September 30, 2021. Copyright 2001 Pageant Media Ltd. development of concanavalin A-induced hepatitis. J. Exp. Med. 191:105. Med. 176:1091. ␥ 20. Mendiratta, S. K., W. D. Martin, S. Hong, A. Boesteanu, S. Joyce, and 47. Finkelman, F. D., I. M. Katona, T. R. Mosmann, and R. L. Coffman. 1988. IFN- L. Van Kaer. 1997. CD1d1 mutant mice are deficient in natural T cells that regulates the isotypes of Ig secreted during in vivo humoral immune responses. promptly produce IL-4. Immunity 6:469. J. Immunol. 140:1022. 21. Kawabe, T., T. Naka, K. Yoshida, T. Tanaka, H. Fujiwara, S. Suematsu, 48. Matsuda, J. L., O. V. Naidenko, L. Gapin, T. Nakayama, M. Taniguchi, N. Yoshida, T. Kishimoto, and H. Kikutani. 1994. The immune responses in C. R. Wang, Y. Koezuka, and M. Kronenberg. 2000. Tracking the response of CD40-deficient mice: impaired immunoglobulin class switching and germinal natural killer T cells to a glycolipid antigen using CD1d tetramers. J. Exp. Med. 192:741. center formation. Immunity 1:167. ϩ ϩ 22. Noelle, R. J., M. Roy, D. M. Shepherd, I. Stamenkovic, J. A. Ledbetter, and 49. Chen, H., and W. E. Paul. 1997. Cultured NK1.1 CD4 T cells produce large ␥ A. Aruffo. 1992. A 39-kDa protein on activated helper T cells binds CD40 and amounts of IL-4 and IFN- upon activation by anti-CD3 or CD1. J. Immunol. transduces the signal for cognate activation of B cells. Proc. Natl. Acad. Sci. USA 159:2240. 89:6550. 50. Schofield, L., M. J. McConville, D. Hansen, A. S. Campbell, B. Fraser-Reid, http://classic.jimmunol.org 23. Nakajima, A., M. Azuma, S. Kodera, S. Nuriya, A. Terashi, M. Abe, S. Hirose, M. J. Grusby, and S. D. Tachado. 1999. CD1d-restricted immunoglobulin G T. Shirai, H. Yagita, and K. Okumura. 1995. Preferential dependence of autoan- formation to GPI-anchored antigens mediated by NKT cells. Science 283:225. tibody production in murine lupus on CD86 costimulatory molecule. Eur. J. Im- 51. Smyth, M. J., K. Y. Thia, S. E. Street, E. Cretney, J. A. Trapani, M. Taniguchi, munol. 25:3060. T. Kawano, S. B. Pelikan, N. Y. Crowe, and D. I. Godfrey. 2000. Differential 24. Akiba, H., H. Oshima, K. Takeda, M. Atsuta, H. Nakano, A. Nakajima, tumor surveillance by natural killer (NK) and NKT cells. J. Exp. Med. 191:661. C. Nohara, H. Yagita, and K. Okumura. 1999. CD28-independent costimulation 52. Gumperz, J. E., C. Roy, A. Makowska, D. Lum, M. Sugita, T. Podrebarac, of T cells by OX40 ligand and CD70 on activated B cells. J. Immunol. 162:7058. Y. Koezuka, S. A. Porcelli, S. Cardell, M. B. Brenner, and S. M. Behar. 2000. 25. Azuma, M., T. Hirano, H. Miyajima, N. Watanabe, H. Yagita, S. Enomoto, Murine CD1d-restricted T cell recognition of cellular lipids. Immunity 12:211. S. Furusawa, Z. Ovary, T. Kinashi, T. Honjo, et al. 1987. Regulation of murine 53. Takeda, K., Y. Hayakawa, L. Van Kaer, H. Matsuda, H. Yagita, and K. Okumura. Downloaded from IgE production in SJA/9 and nude mice: potentiation of IgE production by re- 2000. Critical contribution of liver natural killer T cells to a murine model of combinant interleukin 4. J. Immunol. 139:2538. hepatitis. Proc. Natl. Acad. Sci. USA 97:5498. 26. Hashimoto, W., K. Takeda, R. Anzai, K. Ogasawara, H. Sakihara, K. Sugiura, 54. Takeda, K., and G. Dennert. The development of autoimmunity in C57BL/6 lpr S. Seki, and K. Kumagai. 1995. Cytotoxic NK1.1 Agϩ ␣␤ T cells with interme- mice correlates with the disappearance of natural killer type 1-positive cells: diate TCR induced in the liver of mice by IL-12. J. Immunol. 154:4333. evidence for their suppressive action on bone marrow stem cell proliferation, B 27. Takeda, K., S. Seki, K. Ogasawara, R. Anzai, W. Hashimoto, K. Sugiura, cell immunoglobulin secretion, and autoimmune symptoms. J. Exp. Med. M. Takahashi, M. Satoh, and K. Kumagai. 1996. Liver NK1.1ϩ CD4ϩ ␣␤ T cells 177:155. activated by IL-12 as a major effector in inhibition of experimental tumor me- 55. Sumida, T., A. Sakamoto, H. Murata, Y. Makino, H. Takahashi, S. Yoshida, tastasis. J. Immunol. 156:3366. K. Nishioka, I. Iwamoto, and M. Taniguchi. 1995. Selective reduction of T cells 28. Tomura, M., W. G. Yu, H. J. Ahn, M. Yamashita, Y. F. Yang, S. Ono, bearing invariant V␣24 J␣Q antigen receptor in patients with systemic sclerosis. T. Hamaoka, T. Kawano, M. Taniguchi, Y. Koezuka, and H. Fujiwara. 1999. A J. Exp. Med. 182:1163. novel function of V␣14ϩ CD4ϩ NKT cells: stimulation of IL-12 production by 56. Wilson, S. B., S. C. Kent, K. T. Patton, T. Orban, R. A. Jackson, M. Exley, antigen-presenting cells in the innate immune system. J. Immunol. 163:93. S. Porcelli, D. A. Schatz, M. A. Atkinson, S. P. Balk, et al. 1998. Extreme Th1 29. Chambers, C. A., and J. P. Allison. 1997. Co-stimulation in T cell responses. bias of invariant V␣24J␣Q T cells in type 1 diabetes. Nature 391:177. Curr. Opin. Immunol. 9:396. 57. Zeng, D., M. Dick, L. Cheng, M. Amano, S. Dejbakhsh-Jones, P. Huie, R. Sibley, 30. Gruss, H. J., and S. K. Dower. 1995. Tumor necrosis factor ligand superfamily: and S. Strober. 1998. Subsets of transgenic T cells that recognize CD1 induce or involvement in the pathology of malignant lymphomas. Blood 85:3378. prevent murine lupus: role of cytokines. J. Exp. Med. 187:525.