of July 22, 2010 This information is current as

2004;173;3693-3706 J. Immunol. Sebastian Joyce and Luc Van Kaer Sakai, Isao Serizawa, Lan Wu, Chyung-Ru Wang, Bezbradica, Hiroko Inazawa, Hiromi Ehara, TeruyukiWilson, Danyvid Olivares-Villagómez, Jelena S. Vrajesh V. Parekh, Avneesh K. Singh, Michael T. 

{beta}-Anomeric GlycolipidsNKT Cells to Distinct {alpha}- andDifferences in the In Vivo Response of Quantitative and Qualitative

http://www.jimmunol.org/cgi/content/full/173/6/3693

References

eshttp://www.jimmunol.org/cgi/content/full/173/6/3693#otherarticl17 online articles that cite this article can be accessed at:  

http://www.jimmunol.org/cgi/content/full/173/6/3693#BIBL, 43 of which can be accessed free at:cites 75 articlesThis article

Subscriptions http://www.jimmunol.org/subscriptions/online at

isThe Journal of ImmunologyInformation about subscribing to

Permissions http://www.aai.org/ji/copyright.html

Submit copyright permission requests at

Email Alerts http://www.jimmunol.org/subscriptions/etoc.shtmlup at

Receive free email alerts when new articles cite this article. Sign

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright ©2004 by The American Association ofRockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc., 9650

is published twice each month byThe Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

Quantitative and Qualitative Differences in the In VivoResponse of NKT Cells to Distinct�- and �-AnomericGlycolipids1

Vrajesh V. Parekh,2* Avneesh K. Singh,2* Michael T. Wilson,* Danyvid Olivares-Villagomez,*Jelena S. Bezbradica,* Hiroko Inazawa,† Hiromi Ehara, † Teruyuki Sakai,† Isao Serizawa,†

Lan Wu,* Chyung-Ru Wang,‡ Sebastian Joyce,* and Luc Van Kaer3*

NKT cells represent a unique subset of immunoregulatory T cells that recognize glycolipid Ags presented by the MHC class I-likemolecule CD1d. Because of their immunoregulatory properties, NKT cells are attractive targets for the development of immu-notherapies. The prototypical NKT cell ligand �-galactosylceramide (�-GalCer), originally isolated from a marine sponge, haspotent immunomodulatory activities in mice, demonstrating therapeutic efficacy against metastatic tumors, infections, and auto-immune diseases, but also has a number of adverse side effects. In vivo administration of�-GalCer to mice results in the rapidactivation of NKT cells, which is characterized by cytokine secretion, surface receptor down-regulation, expansion, and secondaryactivation of a variety of innate and adaptive immune system cells. In this study, we have evaluated the in vivo immune responseof mice to a set of structural analogues of�-GalCer. Our results show that, contrary to current thinking, �-anomeric GalCer caninduce CD1d-dependent biological activities in mice, albeit at lower potency than�-anomeric GalCer. In addition, we show thatthe response of NKT cells to distinct GalCer differs not only quantitatively, but also qualitatively. These findings indicate that NKTcells can fine-tune their immune responses to distinct glycolipid Ags in vivo, a property that may be exploited for the developmentof effective and safe NKT cell-based immunotherapies.The Journal of Immunology, 2004, 173: 3693–3706.

N atural killer T cells are a subset of T lymphocytes thatexpress surface markers typical of cells that belong tothe innate arm of the immune system (reviewed in Refs.

1–7). Prototypical NKT cells express an invariant V�14-J�18/V�8(V�24-J�18/V�11 in human) TCR, together with the NK cellmarker NK1.1 (CD161). NKT cells are specific for glycolipid Agspresented by the nonpolymorphic MHC class I-like moleculeCD1d. A hallmark of NKT cells is their capacity to quickly pro-duce large amounts of cytokines in response to TCR engagement.Consequently, these cells have been implicated in a variety of im-mune responses, playing largely a regulatory role (reviewed inRefs. 1–6 and 8–12). For example, NKT cells contribute to thegeneration of protective immune responses against a variety ofinfectious agents (12–19), generally play a suppressive role in au-toimmunity (9, 20–24), enhance or suppress immune responsesdirected against metastatic tumors (10, 11, 25, 26), and play animportant role in transplantation tolerance (27, 28), immune priv-ilege (29), maternal-fetal interactions (30), development of ox-

azolone-induced colitis (31), allergic airway inflammation (32),and contact hypersensitivity (33) in mice.

Although several potential candidates have been suggested (34–37), the physiological ligands that are recognized by NKT cellsremain to be determined. Therefore, many investigators have stud-ied the immunomodulatory properties of artificial ligands for thesecells, most notably derivatives of the marine sponge-derived gly-colipid �-galactosylceramide (�-GalCer).4 In most studies, the im-munomodulatory properties of a synthetic�-GalCer, KRN7000(�-GCs18; see Fig. 1), have been investigated. Administration of�-GCs18 to mice results in the rapid and potent activation of NKTcells, and production by these cells of a burst of cytokines, includ-ing typical Th1 cytokines such as IFN-� and typical Th2 cytokinessuch as IL-4 (4, 5, 38). NKT cells activated in this manner alsodown-regulate critical cell surface receptors (39–41), resulting inthe apparent disappearance of these cells in vivo, which was pre-viously thought to be due to activation-induced apoptosis (42–44).TCR levels on these cells are profoundly down-regulated within8 h following in vivo injection of�-GGs18, but are restored by24–48 h. NK1.1 levels are also down-regulated, which is apparentat 24 h after�-GCs18 injection, and maximal around 3 days.NK1.1 expression then gradually returns to normal levels, but re-mains slightly suppressed until at least 10 days after the initialadministration of�-GCs18. We (39) and others (40) have foundrecently that�-GCs18 also induces a robust in vivo expansion ofNKT cells in a variety of organs, including spleen, peripheralblood, bone marrow, and liver. NKT cell expansion is maximal at

*Department of Microbiology and Immunology, Vanderbilt University School ofMedicine, Nashville, TN 37232;†Pharmaceutical Research Laboratories, Kirin Brew-ery Co., Ltd., Gunma, Japan; and‡Department of Pathology, University of Chicago,Chicago, IL 60637

Received for publication March 11, 2004. Accepted for publication June 11, 2004.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisementin accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by National Institutes of Health Grants AI50953,NS44044, HL68744 (to L.V.K.), AI42284 (to S.J.), and AI43407 (to C.-R.W.).2 V.V.P. and A.K.S. contributed equally to this work.3 Address correspondence and reprint requests to Dr. Luc Van Kaer, Department ofMicrobiology and Immunology, Vanderbilt University School of Medicine, RoomA-5301, Medical Center North, Nashville, TN 37232. E-mail address:luc.van.kaer@vanderbilt.edu

4 Abbreviations used in this paper: GalCer, galactosylceramide; GC, GalCer; EAE,experimental autoimmune encephalomyelitis; MOG, myelin oligodendrocyte glyco-protein; MFI, mean fluorescence intensity.

The Journal of Immunology

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

3 days after �-GCs18 administration, after which NKT cell num-bers slowly return to normal levels around days 7–10 by homeo-static mechanisms. Thus, the response of NKT cells to glycolipidsis characterized by rapid cytokine secretion, surface receptordown-modulation, expansion, and homeostatic contraction.

In vivo activation of NKT cells with �-GCs18 also leads tobystander activation of a variety of innate and adaptive immunesystem cells, including NK cells (43, 45, 46), B cells (45, 47),conventional T cells (45, 48, 49), and dendritic cells (50–53).These properties make �-GalCer attractive for the development ofimmunotherapies and for inclusion in vaccine adjuvants. Indeed,�-GalCer has been successful in treating certain metastatic tumorsand infections, for prevention of autoimmunity, and in enhancingthe efficacy of vaccines in mice (reviewed in Refs. 5, 8, 11, and54–56). However, administration of �-GCs18 can also lead to ad-verse effects, including liver toxicity (57), abortions (30), and ex-acerbation of atherogenesis (58) in mice.

One way to develop effective and safe NKT-cell based immu-notherapeutics would be to use structural analogues of the proto-typical NKT cell ligand �-GCs18. For example, an analog, termedOCH (�-GCs9; for structures of glycolipids, see Fig. 1), with ashortened sphingosine base was suggested to be superior to�-GCs18 in preventing experimental autoimmune encephalomy-elitis (EAE) in mice. However, little is known regarding the invivo immunomodulatory activities of altered lipid ligands such asOCH (59). Therefore, we have analyzed the in vivo immune re-sponse of mice to a panel of �-GalCer analogues, in terms of theircapacity to induce cytokine secretion, down-regulation of theirTCR and NK1.1 markers, in vivo NKT cell expansion, NKT cell-mediated trans-activation of innate and adaptive immune systemcells, and prevention of EAE. Our results show that, contrary tocurrent thinking, �-anomeric GalCer can activate NKT cells invivo, albeit at reduced efficiency than �-anomeric GalCer. We alsofound that the response of NKT cells to distinct GalCer differs notonly quantitatively but also qualitatively. These findings have im-portant implications for understanding the biology of NKT cellsand for developing effective NKT cell-based therapies.

Materials and MethodsMice

Female C57BL/6 (B6) mice, 6–8 wk old, were purchased from The Jack-son Laboratory (Bar Harbor, ME). CD1d�/� mice (60) were from the tenthbackcross to B6. J�18�/� mice (61) backcrossed at least 10 times to B6were obtained from Dr. M. Taniguchi (RIKEN Research Center of Allergyand Immunology, Yokohama, Japan). All mice were bred and housed in theanimal facility at Vanderbilt University School of Medicine.

Reagents

�-GCs18 (KRN7000; (2S,3S,4R)-1-O-(�-D-galactopyranosyl)-N-hexaco-sanoyl-2-amino-1,3,4-octadecanetriol) (62), �-GCs9 (OCH; (2S,3S,4R)-1-O-(�-D-galactopyranosyl)-N-tetracosanoyl-2-amino-1,3,4-nonanetriol) (59),�-GCs10 ((2S,3S,4R)-1-O-(�-D-galactopyranosyl)-N-tetracosanoyl-2-ami-no-1,3,4-decanetriol), and �-GCa26 ((2S,3S,4R)-1-O-(�-D-galactopyrano-syl)-N-hexacosanoyl-2-amino-1,3,4-octadecanetriol) (62) were synthesizedat Kirin Brewery (Gunma, Japan), as described (62). DisialogangliosideGD3 was obtained from Matreya (Pleasant Gap, PA). Synthetic ceramide,phosphatidylethanolamine, �-GCa0 (psychosine; �-D-galactosyl-�1–1�-2-amino-1,3-octadec-4-enediol), �-GCa8 (�-D-galactosyl-�1–1�-N-octanoyl-2-amino-1,3-octadec-4-enediol), and �-GCa12 (�-D-galactosyl-�1–1�-N-dodecanoyl-2-amino-1,3-octadec-4-enediol) were obtained from AvantiPolar Lipids (Alabaster, AL). All �-anomeric glycolipids were synthesizedfrom �-anomeric precursors. Nuclear magnetic resonance spectroscopyanalysis of �-GCa12 was unable to detect �-anomeric forms (data notshown), although this assay does not permit detection of �-anomeric formsat �1% of the �-anomeric form. All glycolipids were reconstituted in PBScontaining 0.5% polysorbate-20 (Sigma-Aldrich, St. Louis, MO) and son-icated at 80°C when needed. Before each experiment, glycolipids werefreshly diluted from stock solutions. Fluorescently labeled tetrameric

CD1d1 molecules loaded with various glycolipids were preparedas described previously (37, 63). Anti-TCR� conjugated with FITC orallophycocyanin, anti-NK1.1-PE or -PerCP-Cy5.5, anti-B220-PerCP,anti-CD3-PerCP, anti-IL-4-allophycocyanin, anti-IFN-�-FITC, anti-CD69-FITC, anti-CD11c-allophycocyanin, anti-CD80-FITC, and anti-CD86-PEwere obtained from BD Pharmingen (San Diego, CA). Synthetic myelinoligodendrocyte glycoprotein (MOG) 35–55 peptide (MOG35–55) was ob-tained from Bio-Synthesis (Lewisville, TX).

Flow cytometry

Single-cell suspensions of the spleen were prepared and stained with fluo-rescently labeled mAbs as described (45). In all experiments, dead cellswere excluded from the analysis by electronic gating. For identification ofNKT cells, B220� cells were excluded from the analysis by electronicgating. The NKT cell population was identified as CD1d/�-GCs18-tet-ramer� and TCR�� cells. For analysis of intracellular cytokines of NKcells, CD3� cells were excluded by electronic gating, and NK1.1high cellswere considered as NK cells. In some experiments, splenocytes were ac-tivated with ionomycin (1 �M) and PMA (20 ng/ml) for 6 h, followed byintracellular detection of cytokine expression by NK cells. For staining ofdendritic cells, FcRs were first blocked by addition of anti-CD16/32 Abs(BD Pharmingen) and dendritic cells were identified on the basis of highexpression of CD11c. For intracellular staining, 2 � 105 splenocytes wereactivated with glycolipids in the presence of Golgi Plug (BD Pharmingen)in U-bottom 96-well plates for 6 h. Intracellular cytokine staining wasperformed with Cytofix/Cytoperm reagents (BD Pharmingen) according tothe manufacturer’s protocol. Data acquisition was performed on a FACS-Calibur instrument (BD Biosciences, San Jose, CA), and data were ana-lyzed by FlowJo software (Tree Star, Ashland, OR).

Measurement of in vitro TCR down-modulation

In vitro measurements of TCR down-modulation were performed as de-scribed (39). Briefly, 2 � 105 splenocytes were suspended in RPMI 1640medium supplemented with 10% FCS, 50 �M 2-ME, 2 mM glutamine,antibiotics, and 10 mM HEPES (complete medium), and plated in multiplewells of a U-bottom 96-well plate. The cells were incubated with glycolipidAgs to a final concentration of 100 ng/ml for the time periods indicated inthe figures. Cells were then briefly centrifuged (1 min; 500 � g) to facil-itate rapid NKT cell contact with APCs. The mean fluorescence intensity(MFI) of CD1d-tetramer staining on B220�TCR��tetramer� cells wasmeasured at the time points indicated in the figures. The percentage of TCRexpression on NKT cells was calculated by the following formula: 100 �((geometric MFI of tetramer staining among gated NKT cells in the ex-perimental condition � geometric MFI of tetramer staining for back-ground)/(geometric MFI of tetramer staining among gated NKT cells in thecontrol condition � geometric MFI of tetramer staining for thebackground)).

Measurement of in vivo and in vitro responses to glycolipids

For in vivo measurements of cytokines, IgE, and cellular responses toglycolipids, mice were injected i.p. with 200 �l of different concentrationsof glycolipids suspended in 0.025% polysorbate-20 in PBS (vehicle). Afterdifferent time points, mice were bled, and cytokines and IgE levels weremeasured in the serum. Alternatively, for evaluation of distinct cell subsets,mice were sacrificed after different time points, spleens were harvested, andsingle-cell suspensions were stained with mAbs followed by flow cytom-etry. For measurements of in vitro responses to glycolipids, untreated orglycolipid-injected mice were sacrificed, and spleen cell suspensions wereplated in U-bottom wells of 96-well plates at 2 � 105 cells per well incomplete medium. Cells were incubated with increasing doses of glyco-lipids for 60 h. For proliferation assays, 1 �Ci of [3H]thymidine (Perkin-Elmer Life Sciences, Boston, MA) was added to the wells, and cells wereincubated for an additional 12 h. Cells were then harvested with a cellharvester (Tomtec, Orange, CT), and [3H]thymidine incorporation wasmonitored with a betaplate reader (Wallac, Gaithersburg, MD). For eval-uation of cytokine secretion in vitro, supernatants were harvested after 60 hof culture with glycolipids and stored until measurement of cytokine levelsby ELISA.

NK cell cytotoxicity assay

NK cell cytotoxicity was evaluated against the NK cell target cell lineYAC-1, in a standard 51Cr release assay. Data are presented as the percentspecific lysis, which was calculated as follows: 100 � ((experimental re-lease � spontaneous release)/(maximum release � spontaneous release)).

3694 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

Induction of EAE in mice and clinical evaluation of disease

Active EAE in B6 mice was induced as described previously (64). Briefly,8- to 10-wk-old female B6 mice were immunized s.c. with 200 �g ofMOG35–55 peptide emulsified in CFA (BD Biosciences) on days 0 and 7.Mice also received 250 ng of pertussis toxin (Invitrogen Life Technologies,Carlsbad, CA) i.p. on days 0 and 2. Mice were treated with different con-centrations of glycolipids or vehicle on days 0, 4, and 7 by i.p. injection.Clinical symptoms were monitored daily after the first immunization. Theclinical score was graded as follows: 0, no disease; 1, tail limpness; 2,hindlimb weakness; 3, hindlimb paralysis; 4, forelimb weakness; 5, quad-riplegia; and 6, death. Cumulative disease scores were calculated by addingdaily disease scores from the day after immunization until the end of theexperiment.

Evaluation of MOG35–55-specific T cell responses

For measurement of MOG35–55-specific T cell responses, mice were im-munized for induction of EAE and treated with glycolipids as describedabove. Mice were sacrificed at 11 days, and the lymph nodes were har-vested. A single-cell suspension of lymph node cells was prepared, and 2 �105 cells were restimulated in vitro with 0 or 30 �g/ml MOG35–55 peptidein complete medium for 60 h. The supernatant was harvested at the end ofthe culture, and IFN-� and IL-4 levels were measured by ELISA.

ELISA

A standard sandwich ELISA was performed to measure mouse IFN-�,IL-4, IL-12, and IgE levels. Capture and detection Abs for IFN-� and IL-4were obtained from R&D Systems (Minneapolis, MN), those for IL-12 andall cytokine standards were from BD Pharmingen, and capture and detec-tion Abs for IgE as well as IgE standards were obtained from BD Pharm-ingen. For detection, streptavidin-HRP conjugate (Zymed Laboratories,San Francisco, CA) was used, and the color was developed using H2O2/3,3�,5,5�-tetramethylbenzidine (DakoCytomation, Carpinteria, CA) as thesubstrate.

ResultsIn the present study, we have compared the in vitro and in vivoresponses of NKT cells to distinct synthetic glycosylceramides. Asa reference for comparison, we used the well-characterized �-ano-meric GalCer KRN7000 (�-GCs18), which has an acyl chain of 26hydrocarbons and a sphingosine base of 18 hydrocarbons (Fig. 1).We have studied four �-anomeric GalCer, �-GCa0 (also calledpsychosine), �-GCa8, �-GCa12, and �-GCa26 that differ fromeach other in the length of their acyl chains (Fig. 1), and two�-anomeric GalCer, �-GCs9 (also called OCH) and �-GCs10,which have a shortened sphingosine base, as compared with theprototypical �-GCs18 (Fig. 1).

Capacity of distinct glycosylceramides to induce in vitroproliferation and cytokine production by splenocytes

We first compared the in vitro proliferative and cytokine responsesof total splenocytes to distinct glycosylceramides. Splenocytesfrom wild-type B6 mice, CD1d-deficient, or J�18-deficient micewere treated in vitro with increasing doses of glycolipid Ags, rang-ing from 0 to 1000 ng/ml, and evaluated for proliferative andIFN-� and IL-4 responses. Results (Fig. 2A) showed that �-GCs18was most potent in inducing proliferation and cytokine production.Although it has been suggested that �-anomeric GalCer do notefficiently activate NKT cells (38), we found that �-GCa12 (Fig.2A) and �-GCa26 (data not shown), but not �-GCa0, �-GCa8,ceramide, phosphatidylethanolamine, and disialoganglioside GD3(data not shown), were able to induce proliferation and cytokineproduction when used at high doses (100–1000 ng/ml). To confirmthat these activities of �-GCa12 are mediated by engagement ofthe invariant TCR expressed by NKT cells, we generated tet-rameric CD1d molecules loaded with this reagent and stainedsplenocytes and liver mononuclear cells from wild-type,CD1d�/�, and J�18�/� mice. Results (Fig. 2B) demonstrated sig-nificant staining in wild-type, but not CD1d�/� and J�18�/� mice,

whereas empty tetramers were unable to stain cells of mice fromany genotype.

Although the responses to �-GCs9 and �-GCs10 in theseassays were indistinguishable from each other, proliferative re-sponses to these analogues were reproducibly weaker than to�-GCs18 (Fig. 2A). Interestingly, levels of IL-4 in the culturestreated with �-GCs9, �-GCs10, and �-GCs18 were very simi-lar, whereas levels of IFN-� were consistently lower in culturestreated with �-GCs9 and �-GCs10, compared with �-GCs18-treated cultures. These findings with �-GCs9 are consistentwith prior studies (59, 65).

Importantly, none of the cultures from splenocytes of CD1d-and J�18-deficient mice induced significant proliferation or cyto-kine secretion (Fig. 2A), indicating that responses observed inwild-type splenocytes required CD1d-restricted NKT cellactivation.

Relative capacity of glycolipids to down-regulate expression ofthe invariant TCR expressed by NKT cells in vitro

Our previous results have shown that NKT cells down-regulatetheir TCR within a few hours of stimulation with �-GCs18 (39).The extent of TCR down-regulation provides a direct measure ofthe extent of T cell activation (65, 66). Therefore, we stimulatedNKT cells with �-GCs18, �-GCa12, �-GCs9, and �-GCs10 invitro for different time points and analyzed levels of TCR expres-sion using tetrameric CD1d molecules loaded with �-GCs18(CD1d-tetramers). In these experiments we used a glycolipid doseof 100 ng/ml for all analogues, as well as a higher dose for�-GCa12 (1000 ng/ml), because it is a less potent activator of NKTcell responses (see Fig. 2A). We found that all analogues testedwere capable of down-regulating expression of the NKT cell re-ceptor and that �-GCa12 was the least effective (Fig. 3). TCRdown-regulation mediated by �-GCs18 was more rapid than forany of the other reagents tested. However, whereas TCR levels onNKT cells activated with �-GCs18 started to increase again at 24 hafter activation, TCR levels of NKT cells stimulated with all otheranalogues continued to decrease until at least 24 h following theinitial activation.

In vivo dynamics of the NKT cell population

We (39) and others (43, 63) have previously shown that in vivoadministration of �-GCs18 induces a rapid decline in CD1d-tet-ramer� cells, which is largely due to transient down-regulation ofthe invariant TCR by these cells (39–41). NKT cells graduallyreappear around 24 h after �-GCs18 administration, and then un-dergo a period of significant proliferation, expanding 5- to 10-foldin the spleen, peripheral blood, and bone marrow. Following thisperiod of expansion, levels of NKT cells return to basal levels at7–12 days after their initial activation. Therefore, we next com-pared the capacity of distinct glycolipids to induce the disappear-ance and expansion of NKT cells in vivo. We focused our analysison NKT cells in the spleen, because NKT cell expansion is mostprofound in this organ (39, 40). Consistent with our in vitro ex-periments (Figs. 2A and 3), �-GCs18 is most potent in causing thedisappearance of NKT cells from the spleen (Fig. 4, A and B).�-GCa0 and �-GCa8 were unable to induce any reduction in theprevalence of TCR��CD1d-tetramer� cells, even when used at adose of 50 �g/mouse (Fig. 4C). Although �-GCa12 (Fig. 4, A andB) and �-GCa26 (data not shown) had no detectable effect whenadministered at a low dose (5 �g/mouse), these reagents wereclearly capable of inducing a slight reduction in NKT cell TCRsurface levels and a concomitant reduction in the overall numbersof TCR��CD1d-tetramer� cells when used at a high dose (50�g/mouse). In the case of �-GCs9 and �-GCs10 (Fig. 4, A and B),

3695The Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

FIGURE 1. Structures of glycolipids used in this study. �-GCs18 (also called KRN7000) is the prototypical �-GalCer and contains a C26 acyl chainand a C18 sphingosine base. �-GCa0 (also called psychosine), �-GCa8, �-GCa12, and �-GCa26 are �-anomeric GalCer that differ from each other in thelength of the acyl chain and in the structure of the sphingosine base. �-GCs9 (also called OCH) and �-GCs10 differ from each other and from �-GCs18in the length of the sphingosine base.

3696 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

FIGURE 2. Capacity of distinct glycosylceramides to activate splenocytes. A, In vitro proliferative and cytokine responses of splenocytes to glycolipidstimulation. Total splenocytes (2 � 105/well) from wild-type B6 mice, CD1d�/� mice, and J�18�/� mice were stimulated with graded doses of the indicatedglycolipids. After 72 h, proliferation was assessed by [3H]thymidine incorporation. Supernatants were harvested at 60 h of culture to monitor IL-4 and IFN-�responses. The proliferation results shown represent the mean � SE of triplicate wells, and the cytokine results represent the mean � SE of two mice. Repre-sentative data of three similar experiments are shown. B, The reactivity of �-GCa12 is dependent upon interaction with the invariant TCR expressed by NKT cells.Splenocytes or liver mononuclear cells from the indicated mice were stained with anti-TCR-�-FITC Abs and PE-labeled, tetrameric CD1d molecules loaded withthe indicated glycolipids. Numbers indicate the percentage of cells within each quadrant. Data are representative of two separate experiments.

3697The Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

significant TCR down-modulation was observed, but this was lessprofound than that observed with �-GCs18, resulting only in aslight apparent loss of NKT cell numbers early (8 h) followingglycolipid administration.

Consistent with our prior studies, we found that �-GCs18 in-duces profound in vivo NKT cell expansion, reaching levels thatare increased 5- to 10-fold around 3 days after injection. Surpris-ingly, however, none of the other reagents tested induced signifi-cant NKT cell expansion, even when used at a dose of 50 �g/mouse (Fig. 4, A, B, and D). As �-GCs18 induces NKT cellexpansion even at doses as low as 0.5 �g/mouse (data not shown),these findings suggest that NKT cells can generate qualitativelydistinct responses to structurally different glycosylceramidesin vivo.

In addition to down-regulation of TCR expression, activatedNKT cells down-regulate expression of the NK1.1 marker (39, 40,67). In vivo, NK1.1 down-regulation becomes detectable around24 h following �-GCs18 administration and is maximal around 3days, after which NK1.1 levels slowly return to normal, remainingsignificantly reduced for at least 10 days (39, 40). Thus, we eval-uated different GalCer for this parameter of NKT cell activation(Fig. 4, E and F). We found, once again, that �-GCs18 was mostpotent in inducing NK1.1 down-regulation, whereas little NK1.1down-regulation was observed for �-GCa12, even at a dose of 50�g/mouse. �-GCs9 and �-GCs10 efficiently induced NK 1.1down-regulation on NKT cells, but their effects were less sustainedthan �-GCs18 (Fig. 4, E and F).

Serum cytokine and IgE levels

Glycolipid-activated NKT cells are capable of modulating Th cellresponses by inducing cytokine production, in particular IL-4,IFN-�, and IL-12 (9, 54, 55). Therefore, we measured the levels ofthese cytokines in the serum of mice at 2, 6, or 24 h after glycolipidtreatment. Results for �-GCs18 were consistent with prior studies(4, 45, 68), indicating an early burst of IL-4 (peaks at 2 h), fol-lowed by IL-12 (peaks at 6 h), and IFN-� (peaks at 24 h) (Fig. 5A).Although a low dose of �-GCa12 did not result in any appreciablecytokine production, at a high dose, this reagent clearly inducedIFN-� and IL-12 production (most evident at the 6-h time point),but no detectable IL-4. In sharp contrast, �-GCs9 and �-GCs10induced an early burst of IL-4 production that was similar to thatinduced by �-GCs18. Consistent with prior studies (59), levels ofIFN-� induced by �-GCs9 and �-GCs10 were slightly lower at 6 hand substantially lower at 24 h, compared with �-GCs18. Curi-

ously, however, and in contrast with the previous study (59),IFN-� levels induced by �-GCs9 and �-GCs10 at early time points(2 h) were higher than those induced by �-GCs18. The serumprofiles of IL-12, produced in response to �-GCs9 and �-GCs10,were very similar to that of �-GCs18.

Prior studies have indicated that a single injection of �-GCs18 toB6 mice induces a profound increase in serum IgE levels (45, 68).Therefore, we compared the capacity of distinct glycolipids to inducethis Ig isotype. All glycolipids tested induced a significant increase inserum IgE levels, but �-GCs18 was most potent (Fig. 5B).

Cytokine production by NKT cells

Glycolipid-activated NKT cells trans-activate a variety of celltypes (55). Consequently, cytokine levels in splenocyte cultures orthe serum of mice in response to glycolipid treatment reflect theoverall pool of cytokines produced by NKT cells themselves, to-gether with those produced by a variety of cell types activatedsecondarily to NKT cells. Therefore, we evaluated IL-4 and IFN-�secretion by NKT cells from glycolipid-treated mice using an in-tracellular staining assay. We found that cytokine profiles pro-duced by NKT cells from naive mice in response to in vitro treat-ment with distinct glycolipids were nearly indistinguishable, butthat �-GCa12 was the least effective (0-h time points in Fig. 6). Invivo administration of glycolipids invariably appeared to result ina rapid reduction of IL-4 production, but sustained IFN-� produc-tion (8-h, 24-h, 3-day, and 7-day time points in Fig. 6). As inprevious experiments, �-GCs18 was the most effective inducer ofcytokines, �-GCa12, although capable of inducing cytokines whenused at a high dose, was the least effective, and responses mediatedby �-GCs9 and �-GCs10 were indistinguishable from each other.

Activation status of B cells, NK cells, conventional T cells, anddendritic cells

In vivo activation of NKT cells results in the trans-activation of avariety of cell types, including NK cells (43, 45, 46), B cells (45,47), conventional T cells (45, 48, 49), and dendritic cells (50–53).Therefore, we compared the activation status of NKT cells, NKcells, B cells, and conventional T cells using the activation markerCD69, and the activation status of dendritic cells by the expressionlevels of CD80 and CD86 after 24 h of in vivo treatment withglycolipids. We found significant up-regulation of CD69, CD80,and CD86 expression on all cell types evaluated, and by each ofthe glycolipids tested (Fig. 7). Akin to our earlier experiments,

FIGURE 3. Glycolipid-induced TCRdown-modulation. Splenocytes (2 �105/well) from B6 mice were acti-vated in the presence of glycolipidAgs at a concentration of 100 ng/ml(unless indicated otherwise) for theindicated times. Cells were harvestedand stained with anti-TCR�-FITC,CD1d/�-GCs18-tetramer-PE, and an-ti-B220-PerCP. The percentage ofTCR expression on NKT cells ascompared with controls was calcu-lated as described in Materials andMethods. The data shown are themean � SE of two mice and repre-sent one of three similar experiments.

3698 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

FIGURE 4. In vivo dynamics of theNKT cell population in response to dis-tinct glycolipids. Mice were injectedwith the indicated glycolipids (5 �g permouse or as indicated; i.p.) or vehicle.Spleen cells were prepared at theindicated time points, stained with anti-TCR�-FITC, CD1d/�-GCs18-tetramer-PE, anti-B220-PerCP, and anti-NK1.1-allophycocyanin, and analyzed by flowcytometry. A, Dot plots of TCR� vs tet-ramer. B, Graphical representation of thedata from A. C, Dot plots for TCR� vstetramer at 8 h after injection of the indi-cated glycolipids. D, Flow cytometryplots for TCR� vs tetramer at 3 days af-ter injection of the indicated glycolipids.Data shown represent the mean � SE oftwo mice per time point. Numbers in A,C, and D indicate the average percentageof TCR��tetramer�B220� cells for twomice. E, Dot plots of tetramer vs NK1.1.Numbers indicate the average percentageof tetramer�NK1.1�B220� cells fortwo mice. F, Percentage of NK1.1�

cells among tetramer� cells shown inE. Data shown represent the mean �SE of two mice per time point. Data forA, B, E, and F are representative of fourseparate experiments with two to threemice in each group; data for C and Dare from one experiment with two miceper group.

3699The Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

�-GCs18 was most potent and �-GCa12 was the least potent stim-ulator. However, whereas a dose of 5 �g of �-GCa12 per mouseinduced no detectable TCR and NK1.1 down-regulation on NKTcells in vivo, this dose was able to increase CD69 expression onNKT and NK cells and to a lesser extent on B cells and conven-tional T cells, as well as increase CD80 and CD86 expression onCD11chigh dendritic cells.

NKT cell-mediated trans-activation of NK cells

Because the capacity of ligand-activated NKT cells to trans-acti-vate NK cells plays a critical role in the in vivo response of NKTcells, we evaluated glycolipids for their effects on this aspect of theimmune response in greater detail. First, we (39) and others (43)have previously shown that �-GCs18 treatment of mice results ina transient reduction in the prevalence of NK1.1�TCR�� cells,representing mostly NK cells (as well as some NKT cells that havedown-regulated their surface TCRs), by a mechanism that remainsundefined. All glycolipids tested in this assay were able to inducea short-term reduction in the prevalence of NK cells, with similarkinetics (Fig. 8, A and B). Second, we tested NK cells for IFN-�secretion by intracellular staining at different time points after invivo treatment with glycolipids (Fig. 8, C–E). Consistent with theireffects on the disappearance of NK cells, all glycolipids testedwere able to induce IFN-� production by these cells. Third, weevaluated the capacity of splenocytes to lyse YAC-1 target cells, at24 h after glycolipid injection. All glycolipids tested were effectiveat inducing the lysis of YAC-1 target cells, albeit at distinct effi-ciency (Fig. 8F).

Collectively, these findings indicate that all glycolipids tested arecapable of inducing NK cell activation in terms of CD69 induction

(Fig. 7A), transient NK cell depletion (Fig. 8, A and B), IFN-� syn-thesis (Fig. 8, C–E), and cytotoxic effector function (Fig. 8F).

Impact of distinct glycosylceramides on the progression of EAE

Finally, we compared the capacity of glycolipids to modulate EAEdisease, induced in B6 mice by immunization with the MOG35–55

peptide. Based on our prior experience with this model (64), weselected a dose of 2 �g/injection/mouse of each of the glycolipidsas well as a dose of 20 �g/injection/mouse in the case of the�-anomeric GalCer, administered to mice on days 0, 4, and 7 afterEAE induction. Results (Fig. 9A and Table I) showed that�-GCs18, �-GCs9, and �-GCs10 have similar potency in prevent-ing EAE. However, in most experiments, �-GCs18 was more po-tent than any of the other glycolipids in preventing EAE (Fig. 9Aand Table I). At the low dose, �-GCa12 and �-GCa26 slightlyameliorated EAE disease, whereas, consistent with our prior stud-ies (64), at a higher dose, these reagents exacerbated disease, withtypically an earlier onset of disease and increased mortality (e.g.,in the experiment shown in Fig. 9A, 4 of 6 mice treated with 20 �gof �-GCa12/injection/mouse died, compared with 1 of 10 mice inthe vehicle group).

To evaluate effects of glycosylceramides on MOG35–55 peptide-specific immune responses, we measured cytokine responses oflymph node cells from mice induced with EAE and treated withglycolipids to in vitro stimulation with MOG35–55. Consistent withprior studies (64), �-GCs18 suppressed MOG35–55-induced IFN-�production and enhanced MOG35–55-induced IL-4 production,compared with vehicle-treated mice (Fig. 9B). Treatment with�-GCs9 and �-GCs10 had similar effects on IFN-�, but little effecton IL-4 production. Finally, �-GCa12 had no detectable influence

FIGURE 5. Serum cytokine andIgE levels after in vivo glycolipid Agadministration. Mice were injectedi.p. with the indicated glycolipids (5�g per mouse or as indicated; i.p.) orvehicle. A, Serum cytokine levels.Mice were bled at 0, 2, 6, and 24 h,and serum levels of IL-4, IFN-�, andIL-12 were measured by ELISA.Data shown represent the mean � SEof four mice per group and are rep-resentative of three separate experi-ments. B, Serum IgE levels. Micewere bled at 0, 6, and 9 days, andserum levels of IgE were measuredby ELISA. Data for the 9-day timepoints are shown and represent themean � SE of four mice per group.Data are representative of two sepa-rate experiments.

3700 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

on MOG35–55-induced cytokine production, at least at the sensi-tivity of the ELISA that was used.

DiscussionWe have performed a detailed comparison of the immunomodu-latory properties of a panel of GalCer analogues, in terms of theireffects on NKT cells themselves, distinct immune responses thatare activated downstream of NKT cells, and disease modulation.Our findings indicate that �-GalCer with long acyl chains (C �12), when used at high concentrations, can effectively activate cy-tokine secretion by NKT cells and induce the trans-activation of avariety of cell types. In addition, among all GalCer tested, only�-GCs18 (KRN7000) was able to induce significant NKT cell ex-pansion in vivo, although other GalCer were able to effectivelytrans-activate a number of cell types in response to NKT cell stim-ulation. None of these activities were seen in the absence of CD1dor J�18 expression, indicating that they require CD1d-restrictedAg presentation to invariant NKT cells. Finally, in our hands,�-GCs18 was as potent, or more potent, than analogues with ashortened sphingosine base, �-GCs9 (OCH; C � 9) and �-GCs10(C � 10), for modulating EAE disease. Collectively, our findingsindicate that NKT cells can fine-tune their responses to distinctglycolipids.

Our results with �-GCa12 and �-GCa26 indicate that these re-agents, when used at high concentrations, effectively induce cyto-kine secretion in splenocyte cultures in vitro, and in the serum invivo. The cytokine profiles induced by these reagents were similarto those induced by the prototypical �-GalCer, �-GCs18(KRN7000). However, �-GalCer induced little IL-4 secretion in

the serum. Furthermore, intracellular staining of NKT cells dem-onstrated significant IFN-� and IL-4 production in response to�-GCa12, although responses were substantially reduced com-pared with the �-GalCer tested. Additionally, �-GCa12 effectivelyactivated a variety of other cell types, including B cells, conven-tional T cells, dendritic cells, and NK cells, secondary to NKT cellstimulation. The capacity of �-GCa12 to induce the activation ofNK cells was confirmed by a variety of means, including the in-duction of CD69 expression by NK cells, the transient disappear-ance of these cells in vivo around 8–48 h after treatment, produc-tion of intracellular IFN-�, and in vitro cytotoxicity against YAC-1target cells. The capacity of �-GCa12 and �-GCa26 to inducetrans-activation of NK cells, in the face of relatively poor IL-4secretion by NKT cells, may provide a possible explanation for thetendency of these reagents, when used at high doses, to exacerbaterather than suppress EAE disease (64).

In agreement with the findings reported here for �-anomericGalCer, a recent report (69) has shown that �-GalCer can effec-tively activate NKT cells. However, in apparent contrast with ourfindings, these investigators were unable to find evidence for sig-nificant NK cell stimulation mediated by �-GalCer (69). However,it is interesting to note that most of the experiments by Ortaldo etal. (69) and those reported here, were performed with �-GCa12purchased from the same commercial source (i.e., Avanti PolarLipids). In addition, we have confirmed most of our findings, inparticular the NK cell activation data, with a separate reagent,�-GCa26, which was derived from a different source (i.e., KirinBrewery). At present, the reasons for these apparent differences in

FIGURE 6. NKT cell cytokine produc-tion following glycolipid administration.Mice were injected with the indicated gly-colipids (5 �g per mouse or as indicated;i.p.) or vehicle. Spleen cells were preparedat the indicated time points, and 2 � 105

cells were restimulated in vitro with glyco-lipids at a concentration of 100 ng/ml andcultured in the presence of Golgi Plug for6 h to induce the intracellular accumulationof cytokines. Cells were harvested and sur-face stained with tetramer-PE and anti-B220-PerCP, followed by intracellularstaining with anti-IL-4-allophycocyanin andanti-IFN-�-FITC. Data are shown for NKTcells (B220�tetramer�). Numbers indicatethe percentage of cells within each quadrant.Data are representative of three separateexperiments.

3701The Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

results are unclear. Possible explanations include details in exper-imental setup and sensitivity of assays used for NK cell activation.

One potential concern of our and other studies performed with�-anomeric GalCer is their potential low-level contamination with�-anomeric forms. Although nuclear magnetic resonance studieswere unable to detect any �-anomeric forms among these �-Gal-Cer (data not shown), the relatively low sensitivity of this assaycannot exclude low levels (�1%) of contamination. Although wecannot formally exclude contamination with �-anomeric forms,our findings indicate that �-GCa12 and �-GCa26, which were ob-tained from different sources, behave differently from any of the�-anomeric GalCers that we have studied: 1) these reagents werevery potent in inducing the activation of CD80 and CD86 expres-sion by dendritic cells, 2) they are relatively ineffective in inducing

serum IL-4 secretion in vivo, 3) they are relatively ineffective ininducing IL-4 production by NKT cells, while inducing significantIFN-�, as measured by intracellular staining, and 4) when used athigh doses, they exacerbate EAE. Taken together, these findingsprovide evidence for qualitative differences in the responses ofmice to �- vs �-anomeric GalCers, arguing against contaminationas the reason for our observed activities of �-GCa12 and �-GCa26.

It has been generally assumed that NKT cells recognize �-ano-meric but not �-anomeric glycosphingolipids (38, 70). The latter,but not the former, are found in mammals, particularly in the ner-vous system (71, 72). Our findings that �-GalCer with long acylchains, can be recognized by NKT cells and elicit the effectorfunctions of these cells, raises the interesting possibility that en-dogenous �-anomeric glycosphingolipids may function as (one of)

FIGURE 7. Activation status ofdifferent cell types in glycolipid-in-jected mice. Mice were treated withvehicle or the indicated glycolipidanalogues (5 �g per mouse or as in-dicated; i.p.). After 24 h, mice weresacrificed and single-cell suspensionsof the spleen were prepared. A,Expression of the activation markerCD69 by NKT cells, NK cells, Bcells, and conventional T cells.Spleen cells were stained with anti-CD69-FITC or its isotype control(filled), and with different combina-tions of tetramer-PE, anti-B220-PerCP, anti-TCR�-allophycocyanin,and anti-NK1.1-PerCP-Cy5.5. CD69expression was measured on NKTcells (B220�TCR��tetramer�), NKcells (B220�TCR��NK1.1high),B cells (B220�TCR��) and conven-tional T cells (TCR��tetramer�B220�)in mice injected with glycolipid (thickline) or vehicle (thin line). B, Expressionof CD80 and CD86 on dendritic cells.Spleen cells were stained with combina-tions of anti-CD11c, anti-CD80, anti-CD86, or their respective isotype controlAbs (filled). Data are shown forCD11chigh cells after treatment with theglycolipid Ags (thick lines) or vehicle(thin lines). Data are representative ofthree separate experiments.

3702 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

FIGURE 8. NKT cell-mediatedtrans-activation of NK cells by gly-colipid Ags. Mice were injected (i.p.)with vehicle or the indicated glyco-lipids (5 �g per mouse or as indi-cated). Spleen cells were prepared atthe indicated time points. A and B,Prevalence of NK cells. Cells werestained with anti-TCR�-FITC, anti-NK1.1-PE, and anti-B220-PerCP,and analyzed by flow cytometry.Numbers in A indicate the mean per-centage of NK cells (TCR��NK1.1high) for two to three mice. Datafrom A are summarized in B as themean � SE percentage of NK cellsin glycolipid-injected mice comparedwith uninjected mice. Data shownrepresents one of four similar exper-iments. C–E, Cytokine production byNK cells after glycolipid administra-tion. Mice were injected with glyco-lipids, and after different time points,spleen cells were cultured in medium(C and E) or PMA (20 ng/ml) plusionomycin (1 �M) (D) in the pres-ence of Golgi Plug for 6 h to allowintracellular cytokine accumulation.Cells were harvested and surface-stainedwithanti-CD3-PerCPandanti-NK 1.1-PE, followed by intracellularstaining with anti-IFN-�-FITC. Datafor the 24-h time point are shown inC and D, and data for different timepoints after glycolipid injection areshown in E. The percentages ofIFN-�-positive cells were determinedafter electronic gating on theCD3�NK1.1high population. F, Invitro NK cell cytotoxicity. Spleno-cytes from mice injected with theglycolipid Ags were prepared after24 h of administration and used as thecytotoxic effector cells against 51Cr-labeled YAC-1 cells in different E:Tratios. The data shown are themean � SE of two mice from rep-resentative of three separateexperiments.

3703The Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

the physiological ligands for NKT cells. In this context, we haveshown previously that cells deficient in �-D-glucosylceramide syn-thase fail to stimulate NKT cell hybridomas (37). Thus, our find-ings raise the possibility that a �-anomeric glycolipid downstreamof this enzyme could function as a natural ligand for NKT cells.However, it is unlikely that the physiological concentrations of�-anomeric endogenous glycolipids reach levels that are sufficientto directly activate NKT cells. Nevertheless, it has been recentlyproposed that NKT cells use a strategy for their activation involv-ing weak responses to self Ags that are amplified by proinflam-matory cytokines such as those produced during microbial infec-tion (73). Our findings suggest that �-anomeric glycolipids mayprovide such a source of weakly reactive self Ags. Future studieswill be designed to test this hypothesis.

A key observation in our studies is that only �-GCs18 is capableof inducing significant in vivo expansion of NKT cells. Because�-GCs18 induces NKT cell expansion over a wide range of in vivodoses (data not shown), this is not simply due to differences in theoverall strength of TCR signaling, but rather quantitatively differ-ent responses of NKT cells to distinct glycolipid Ags. For exam-ple, despite the lack of NKT cell expansion mediated by �-GCs9and �-GCs10, these reagents and �-GCs18 promoted similar levelsof IL-4 secretion by NKT cells and in the serum. Nevertheless,�-GCs9 and �-GCs10 appeared to be less effective than �-GCs18in inducing IFN-� production by NKT cells. Furthermore, we didnot find a substantial difference in the capacity of these reagents to

induce the trans-activation of different innate and adaptive im-mune cells, including NK cells, and in their capacity toprevent EAE.

GalCer hold significant promise for the development of NKTcell-based immunotherapies. �-GalCer have been successfullyused in treating certain metastatic tumors and infections, and forpreventing autoimmunity (reviewed in Refs. 5, 8, 11, and 54–56).However, administration of �-GalCer can also lead to adverse ef-fects, including liver toxicity (57), abortions (30), and exacerbationof atherogenesis (58) in mice. Although administration of �-Gal-Cer to humans appears to have no adverse consequences, at leastin the short-term and at the doses used (74, 75), these reagentshave significant potential for generating unwanted side effects.Therefore, the availability of glycolipid analogues that selectivelyinduce certain effector functions of NKT cells holds great promisefor the development of effective and safe NKT cell-based immu-notherapies. Although it has been previously suggested that �-Gal-Cer with a shortened sphingosine base may have superior efficacyfor prevention of autoimmune disease (59, 76), we were unable toreproduce those findings, at least at the doses that were tested.Nevertheless, our findings indicate that NKT cells can fine-tunetheir in vivo responses to distinct glycolipids, suggesting that itmay be possible to generate NKT cell ligands that selectively pro-mote certain NKT cell-mediated immune effector functions andhave superior immunomodulatory activities.

FIGURE 9. Capacity of glycolipidsto modulate EAE disease parameters.A, Modulation of clinical EAE disease.EAE was induced in B6 mice by im-munization with MOG35–55 peptide ondays 0 and 7. Mice were treated withthe indicated glycolipid analogues (2�g/injection or as indicated) or withvehicle control (0.025% polysor-bate-20 in PBS) on days 0, 4, and 7.Data shown correspond to experiment1 in Table I. B, Capacity of glycolipidsto modulate MOG35–55-specific T cellresponses. Inguinal and axillary lymphnodes were collected from MOG35–55-immunized and glycolipid-treated miceat 11 days following the first immuni-zation. Cells were cultured in the ab-sence or presence of MOG35–55 (30�g/ml) for 60 h. Culture supernatantswere collected, and cytokine (IFN-�and IL-4) levels were measured byELISA. Data shown represent themean � SE of five mice. Two separateexperiments were performed.

3704 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

AcknowledgmentsWe thank Dr. M. Taniguchi for providing J�18�/� mice. We also thankJie Wei, Tiffaney Vincent, and Michele Nadaf for technical assistance.

References1. Bendelac, A., M. N. Rivera, S. H. Park, and J. H. Roark. 1997. Mouse CD1-

specific NK1 T cells: development, specificity, and function. Annu. Rev. Immu-nol. 15:535.

2. Godfrey, D. I., K. J. Hammond, L. D. Poulton, M. J. Smyth, and A. G. Baxter.2000. NKT cells: facts, functions and fallacies. Immunol. Today 21:573.

3. Joyce, S. 2001. CD1d and natural T cells: how their properties jump-start theimmune system. Cell. Mol. Life Sci. 58:442.

4. Kronenberg, M., and L. Gapin. 2002. The unconventional lifestyle of NKT cells.Nat. Rev. Immunol. 2:557.

5. Taniguchi, M., M. Harada, S. Kojo, T. Nakayama, and H. Wakao. 2003. Theregulatory role of V�14 NKT cells in innate and acquired immune response.Annu. Rev. Immunol. 21:483.

6. Van Kaer, L. Regulation of immune responses by CD1d-restricted natural killerT cells. Immunol. Res. In press.

7. Godfrey, D. I., H. R. MacDonald, M. Kronenberg, M. J. Smyth, and L. Van Kaer.2004. NKT cells: what’s in a name? Nat. Rev. Immunol. 4:231.

8. Hammond, K. J., and D. I. Godfrey. 2002. NKT cells: potential targets for au-toimmune disease therapy? Tissue Antigens 59:353.

9. Hammond, K. J., and M. Kronenberg. 2003. Natural killer T cells: natural orunnatural regulators of autoimmunity? Curr. Opin. Immunol. 15:683.

10. Smyth, M. J., N. Y. Crowe, Y. Hayakawa, K. Takeda, H. Yagita, andD. I. Godfrey. 2002. NKT cells—conductors of tumor immunity? Curr. Opin.Immunol. 14:165.

11. Brutkiewicz, R. R., and V. Sriram. 2002. Natural killer T (NKT) cells and theirrole in antitumor immunity. Crit. Rev. Oncol. Hematol. 41:287.

12. Skold, M., and S. M. Behar. 2003. Role of CD1d-restricted NKT cells in micro-bial immunity. Infect. Immun. 71:5447.

13. Ishikawa, H., H. Hisaeda, M. Taniguchi, T. Nakayama, T. Sakai, Y. Maekawa,Y. Nakano, M. Zhang, T. Zhang, M. Nishitani, et al. 2000. CD4� V�14 NKTcells play a crucial role in an early stage of protective immunity against infectionwith Leishmania major. Int. Immunol. 12:1267.

14. Kumar, H., A. Belperron, S. W. Barthold, and L. K. Bockenstedt. 2000. Cuttingedge: CD1d deficiency impairs murine host defense against the spirochete, Bor-relia burgdorferi. J. Immunol. 165:4797.

15. Kawakami, K., Y. Kinjo, K. Uezu, S. Yara, K. Miyagi, Y. Koguchi,T. Nakayama, M. Taniguchi, and A. Saito. 2001. Monocyte chemoattractant pro-tein-1-dependent increase of V�14 NKT cells in lungs and their roles in Th1response and host defense in cryptococcal infection. J. Immunol. 167:6525.

16. Faveeuw, C., V. Angeli, J. Fontaine, C. Maliszewski, A. Capron, L. Van Kaer,M. Moser, M. Capron, and F. Trottein. 2002. Antigen presentation by CD1dcontributes to the amplification of Th2 responses to Schistosoma mansoni gly-coconjugates in mice. J. Immunol. 169:906.

17. Exley, M. A., N. J. Bigley, O. Cheng, S. M. Tahir, S. T. Smiley, Q. L. Carter,H. F. Stills, M. J. Grusby, Y. Koezuka, M. Taniguchi, and S. P. Balk. 2001.CD1d-reactive T-cell activation leads to amelioration of disease caused by dia-betogenic encephalomyocarditis virus. J. Leukocyte Biol. 69:713.

18. Duthie, M. S., M. Wleklinski-Lee, S. Smith, T. Nakayama, M. Taniguchi, andS. J. Kahn. 2002. During Trypanosoma cruzi infection CD1d-restricted NK Tcells limit parasitemia and augment the antibody response to a glycophos-phoinositol-modified surface protein. Infect. Immun. 70:36.

19. Nieuwenhuis, E. E., T. Matsumoto, M. Exley, R. A. Schleipman, J. Glickman,D. T. Bailey, N. Corazza, S. P. Colgan, A. B. Onderdonk, and R. S. Blumberg.2002. CD1d-dependent macrophage-mediated clearance of Pseudomonas aerugi-nosa from lung. Nat. Med. 8:588.

20. Baxter, A. G., S. J. Kinder, K. J. Hammond, R. Scollay, and D. I. Godfrey. 1997.Association between ��TCR�CD4�CD8� T-cell deficiency and IDDM inNOD/Lt mice. Diabetes 46:572.

21. Lehuen, A., O. Lantz, L. Beaudoin, V. Laloux, C. Carnaud, A. Bendelac,J. F. Bach, and R. C. Monteiro. 1998. Overexpression of natural killer T cellsprotects V�14-J�281 transgenic nonobese diabetic mice against diabetes. J. Exp.Med. 188:1831.

22. Shi, F. D., M. Flodstrom, B. Balasa, S. H. Kim, K. Van Gunst, J. L. Strominger,S. B. Wilson, and N. Sarvetnick. 2001. Germ line deletion of the CD1 locusexacerbates diabetes in the NOD mouse. Proc. Natl. Acad. Sci. USA 98:6777.

23. Mars, L. T., V. Laloux, K. Goude, S. Desbois, A. Saoudi, L. Van Kaer,H. Lassmann, A. Herbelin, A. Lehuen, and R. S. Liblau. 2002. Cutting edge: V�14-J�281 NKT cells naturally regulate experimental autoimmune encephalomy-elitis in nonobese diabetic mice. J. Immunol. 168:6007.

24. Zeng, D., M. K. Lee, J. Tung, A. Brendolan, and S. Strober. 2000. Cutting edge:a role for CD1 in the pathogenesis of lupus in NZB/NZW mice. J. Immunol.164:5000.

25. Crowe, N. Y., M. J. Smyth, and D. I. Godfrey. 2002. A critical role for naturalkiller T cells in immunosurveillance of methylcholanthrene-induced sarcomas.J. Exp. Med. 196:119.

26. Terabe, M., S. Matsui, N. Noben-Trauth, H. Chen, C. Watson, D. D. Donaldson,D. P. Carbone, W. E. Paul, and J. A. Berzofsky. 2000. NKT cell-mediated re-pression of tumor immunosurveillance by IL-13 and the IL-4R-STAT6 pathway.Nat. Immunol. 1:515.

27. Ikehara, Y., Y. Yasunami, S. Kodama, T. Maki, M. Nakano, T. Nakayama,M. Taniguchi, and S. Ikeda. 2000. CD4� V�14 natural killer T cells are essentialfor acceptance of rat islet xenografts in mice. J. Clin. Invest. 105:1761.

28. Seino Ki, K., K. Fukao, K. Muramoto, K. Yanagisawa, Y. Takada, S. Kakuta,Y. Iwakura, L. Van Kaer, K. Takeda, T. Nakayama, et al. 2001. Requirement fornatural killer T (NKT) cells in the induction of allograft tolerance. Proc. Natl.Acad. Sci. USA 98:2577.

29. Sonoda, K. H., M. Exley, S. Snapper, S. P. Balk, and J. Stein-Streilein. 1999.CD1-reactive natural killer T cells are required for development of systemictolerance through an immune-privileged site. J. Exp. Med. 190:1215.

30. Ito, K., M. Karasawa, T. Kawano, T. Akasaka, H. Koseki, Y. Akutsu, E. Kondo,S. Sekiya, K. Sekikawa, M. Harada, et al. 2000. Involvement of decidual V�14NKT cells in abortion. Proc. Natl. Acad. Sci. USA 97:740.

31. Heller, F., I. J. Fuss, E. E. Nieuwenhuis, R. S. Blumberg, and W. Strober. 2002.Oxazolone colitis, a Th2 colitis model resembling ulcerative colitis, is mediatedby IL-13-producing NK-T cells. Immunity 17:629.

32. Akbari, O., P. Stock, E. Meyer, M. Kronenberg, S. Sidobre, T. Nakayama,M. Taniguchi, M. J. Grusby, R. H. DeKruyff, and D. T. Umetsu. 2003. Essentialrole of NKT cells producing IL-4 and IL-13 in the development of allergen-induced airway hyperreactivity. Nat. Med. 9:582.

Table I. Modulation of EAE by distinct glycosylceramidesa

Experiment TreatmentNumberof Mice

Mean Onset ofDisease (day)

DiseaseFrequency (%)

Mean MaximumScore

Mean CumulativeScore

1 Vehicle 10 13.2 90 4.0 � 0.56 107.1 � 5.5�-GCs18 (2 �g) 4 25.0 25 0.3 � 0.25 5.5 � 5.5�-GCa12 (2 �g) 10 13.3 70 2.5 � 0.76 72.0 � 24.28**�-GCa12 (20 �g) 6 10.0 100 5.0 � 0.81 170.0 � 31.47***

�-GCs9 (2 �g) 6 17.0 50 1.4 � 0.6 18.8 � 11.5�-GCs10 (2 �g) 8 18.0 37.5 0.4 � 0.18 2.3 � 1.39

2 Vehicle 5 15.2 100 3.0 � 0.4 70.0 � 15.3�-GCs18 (2 �g) 6 16.3 16.7 0.3 � 0.2 0.5 � 0.34�-GCa26 (2 �g) 4 16.3 100 4.2 � 0.5 75.2 � 20.8*�-GCa26 (20 �g) 6 15.2 100 3.8 � 0.5 84.7 � 15.3*

�-GCs9 (2 �g) 6 20.3 50 0.5 � 0.2 7.8 � 4.7�-GCs10 (2 �g) 5 19.0 40 1.0 � 0.5 16.4 � 11.2

3 Vehicle 7 13.2 100 2.0 � 0.3 62.3 � 8.3�-GCs18 (2 �g) 8 19.3 25.5 0.1 � 0.06 3.0 � 1.9�-GCs18 (0.5 �g) 8 17.1 80 1.1 � 0.4 34.1 � 10.7�-GCs9 (0.5 �g) 8 14.0 100 1.6 � 0.2 47.8 � 6.8

�-GCs10 (0.5 �g) 8 14.9 100 1.2 � 0.2 37.8 � 6.3

a EAE was induced and mice were injected with vehicle and distinct glycosylceramides as described in Materials and Methods. The amount of glycolipid administered perinjection is indicated in parentheses. Data shown represent the mean values � SE. Data in experiment 1 are graphically presented in Fig. 9A. �, p � 0.05; ��, p � 0.005; ���,and p � 0.00005, compared with vehicle-treated mice.

3705The Journal of Immunology

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from

33. Campos, R. A., M. Szczepanik, A. Itakura, M. Akahira-Azuma, S. Sidobre,M. Kronenberg, and P. W. Askenase. 2003. Cutaneous immunization rapidlyactivates liver invariant V�14 NKT cells stimulating B-1 B cells to initiate T cellrecruitment for elicitation of contact hypersensitivity. J. Exp. Med. 198:1785.

34. Schofield, L., M. J. McConville, D. Hansen, A. S. Campbell, B. Fraser-Reid,M. J. Grusby, and S. D. Tachado. 1999. CD1d-restricted immunoglobulin Gformation to GPI-anchored antigens mediated by NKT cells. Science 283:225.

35. Wu, D. Y., N. H. Segal, S. Sidobre, M. Kronenberg, and P. B. Chapman. 2003.Cross-presentation of disialoganglioside GD3 to natural killer T cells. J. Exp.Med. 198:173.

36. Gumperz, J. E., C. Roy, A. Makowska, D. Lum, M. Sugita, T. Podrebarac,Y. Koezuka, S. A. Porcelli, S. Cardell, M. B. Brenner, and S. M. Behar. 2000.Murine CD1d-restricted T cell recognition of cellular lipids. Immunity 12:211.

37. Stanic, A. K., A. D. De Silva, J. J. Park, V. Sriram, S. Ichikawa, Y. Hirabyashi,K. Hayakawa, L. Van Kaer, R. R. Brutkiewicz, and S. Joyce. 2003. Defectivepresentation of the CD1d1-restricted natural Va14Ja18 NKT lymphocyte antigencaused by �-D-glucosylceramide synthase deficiency. Proc. Natl. Acad. Sci. USA100:1849.

38. Kawano, T., J. Cui, Y. Koezuka, I. Toura, Y. Kaneko, K. Motoki, H. Ueno,R. Nakagawa, H. Sato, E. Kondo, et al. 1997. CD1d-restricted and TCR-mediatedactivation of V�14 NKT cells by glycosylceramides. Science 278:1626.

39. Wilson, M. T., C. Johansson, D. Olivares-Villagomez, A. K. Singh, A. K. Stanic,C. R. Wang, S. Joyce, M. J. Wick, and L. Van Kaer. 2003. The response ofnatural killer T cells to glycolipid antigens is characterized by surface receptordown-modulation and expansion. Proc. Natl. Acad. Sci. USA 100:10913.

40. Crowe, N. Y., A. P. Uldrich, K. Kyparissoudis, K. J. L. Hammond, K. Hayakawa,S. Sidobre, R. Keating, M. Kronenberg, M. J. Smyth, and D. I. Godfrey. 2003.Glycolipid antigen drives rapid expansion and sustained cytokine production byNKT cells. J. Immunol. 171:4020.

41. Harada, M., K. I. Seino, H. Wakao, S. Sakata, Y. Ishizuka, T. Ito, S. Kojo,T. Nakayama, and M. Taniguchi. 2004. Down-regulation of the invariant V�14antigen receptor in NKT cells upon activation. Int. Immunol. 16:241.

42. Eberl, G., and H. R. MacDonald. 1998. Rapid death and regeneration of NKTcells in anti-CD3�- or IL-12-treated mice: a major role for bone marrow in NKTcell homeostasis. Immunity 9:345.

43. Eberl, G., and H. R. MacDonald. 2000. Selective induction of NK cell prolifer-ation and cytotoxicity by activated NKT cells. Eur. J. Immunol. 30:985.

44. Leite-de-Moraes, M. C., A. Herbelin, C. Gouarin, Y. Koezuka, E. Schneider, andM. Dy. 2000. Fas/Fas ligand interactions promote activation-induced cell death ofNK T lymphocytes. J. Immunol. 165:4367.

45. Singh, N., S. Hong, D. C. Scherer, I. Serizawa, N. Burdin, M. Kronenberg,Y. Koezuka, and L. Van Kaer. 1999. Cutting edge: activation of NK T cells byCD1d and �-galactosylceramide directs conventional T cells to the acquisition ofa Th2 phenotype. J. Immunol. 163:2373.

46. Carnaud, C., D. Lee, O. Donnars, S. H. Park, A. Beavis, Y. Koezuka, andA. Bendelac. 1999. Cutting edge: cross-talk between cells of the innate immunesystem: NKT cells rapidly activate NK cells. J. Immunol. 163:4647.

47. Kitamura, H., A. Ohta, M. Sekimoto, M. Sato, K. Iwakabe, M. Nakui, T. Yahata,H. Meng, T. Koda, S. Nishimura, et al. 2000. �-Galactosylceramide induces earlyB-cell activation through IL-4 production by NKT cells. Cell. Immunol. 199:37.

48. Eberl, G., P. Brawand, and H. R. MacDonald. 2000. Selective bystander prolif-eration of memory CD4� and CD8� T cells upon NK T or T cell activation.J. Immunol. 165:4305.

49. Nishimura, T., H. Kitamura, K. Iwakabe, T. Yahata, A. Ohta, M. Sato, K. Takeda,K. Okumura, L. Van Kaer, T. Kawano, et al. 2000. The interface between innateand acquired immunity: glycolipid antigen presentation by CD1d-expressing den-dritic cells to NKT cells induces the differentiation of antigen-specific cytotoxicT lymphocytes. Int. Immunol. 12:987.

50. Kitamura, H., K. Iwakabe, T. Yahata, S. Nishimura, A. Ohta, Y. Ohmi, M. Sato,K. Takeda, K. Okumura, L. Van Kaer, et al. 1999. The natural killer T (NKT) cellligand �-galactosylceramide demonstrates its immunopotentiating effect by in-ducing interleukin (IL)-12 production by dendritic cells and IL-12 receptor ex-pression on NKT cells. J. Exp. Med. 189:1121.

51. Tomura, M., W. G. Yu, H. J. Ahn, M. Yamashita, Y. F. Yang, S. Ono,T. Hamaoka, T. Kawano, M. Taniguchi, Y. Koezuka, and H. Fujiwara. 1999. Anovel function of V�14�CD4�NKT cells: stimulation of IL-12 production byantigen-presenting cells in the innate immune system. J. Immunol. 163:93.

52. Fujii, S., K. Shimizu, M. Kronenberg, and R. M. Steinman. 2002. ProlongedIFN-�-producing NKT response induced with �-galactosylceramide-loaded DCs.Nat. Immunol. 3:867.

53. Fujii, S., K. Shimizu, C. Smith, L. Bonifaz, and R. M. Steinman. 2003. Activationof natural killer T cells by �-galactosylceramide rapidly induces the full matu-ration of dendritic cells in vivo and thereby acts as an adjuvant for combined CD4and CD8 T cell immunity to a coadministered protein. J. Exp. Med. 198:267.

54. Wilson, M. T., A. K. Singh, and L. Van Kaer. 2002. Immunotherapy with ligandsof natural killer T cells. Trends Mol. Med. 8:225.

55. Wilson, M. T., and L. Van Kaer. 2003. Natural killer T cells as targets fortherapeutic intervention in autoimmune diseases. Curr. Pharm. Des. 9:201.

56. Schmieg, J., G. Yang, R. W. Franck, and M. Tsuji. 2003. Superior protectionagainst malaria and melanoma metastases by a C-glycoside analogue of the nat-ural killer T cell ligand �-galactosylceramide. J. Exp. Med. 198:1631.

57. Osman, Y., T. Kawamura, T. Naito, K. Takeda, L. Van Kaer, K. Okumura, andT. Abo. 2000. Activation of hepatic NKT cells and subsequent liver injury fol-lowing administration of �-galactosylceramide. Eur. J. Immunol. 30:1919.

58. Tupin, E., A. Nicoletti, R. Elhage, M. Rudling, H. G. Ljunggren, G. K. Hansson,and G. P. Berne. 2004. CD1d-dependent activation of NKT cells aggravatesatherosclerosis. J. Exp. Med. 199:417.

59. Miyamoto, K., S. Miyake, and T. Yamamura. 2001. A synthetic glycolipid pre-vents autoimmune encephalomyelitis by inducing TH2 bias of natural killer Tcells. Nature 413:531.

60. Mendiratta, S. K., W. D. Martin, S. Hong, A. Boesteanu, S. Joyce, andL. Van Kaer. 1997. CD1d1 mutant mice are deficient in natural T cells thatpromptly produce IL-4. Immunity 6:469.

61. Cui, J., T. Shin, T. Kawano, H. Sato, E. Kondo, I. Toura, Y. Kaneko, H. Koseki,M. Kanno, and M. Taniguchi. 1997. Requirement for V�14 NKT cells in IL-12-mediated rejection of tumors. Science 278:1623.

62. Morita, M., K. Motoki, K. Akimoto, T. Natori, T. Sakai, E. Sawa, K. Yamaji,Y. Koezuka, E. Kobayashi, and H. Fukushima. 1995. Structure-activity relation-ship of �-galactosylceramides against B16-bearing mice. J. Med. Chem. 38:2176.

63. Matsuda, J. L., O. V. Naidenko, L. Gapin, T. Nakayama, M. Taniguchi,C. R. Wang, Y. Koezuka, and M. Kronenberg. 2000. Tracking the response ofnatural killer T cells to a glycolipid antigen using CD1d tetramers. J. Exp. Med.192:741.

64. Singh, A. K., M. T. Wilson, S. Hong, D. Olivares-Villagomez, C. Du,A. K. Stanic, S. Joyce, S. Sriram, Y. Koezuka, and L. Van Kaer. 2001. Naturalkiller T cell activation protects mice against experimental autoimmune enceph-alomyelitis. J. Exp. Med. 194:1801.

65. Stanic, A. K., R. Shashidharamurthy, J. S. Bezbradica, N. Matsuki,Y. Yoshimura, S. Miyake, L. Van Kaer, T. Yamamura, and S. Joyce. 2003.Another view of T cells antigen recognition: cooperative engagement of glyco-lipid antigens by Va14Ja18 natural T cell receptor. J. Immunol. 171:4539.

66. Liu, H., M. Rhodes, D. L. Wiest, and D. A. Vignali. 2000. On the dynamics ofTCR:CD3 complex cell surface expression and downmodulation. Immunity13:665.

67. Chen, H., H. Huang, and W. E. Paul. 1997. NK1.1�CD4� T cells lose NK1.1expression upon in vitro activation. J. Immunol. 158:5112.

68. Burdin, N., L. Brossay, and M. Kronenberg. 1999. Immunization with �-galac-tosylceramide polarizes CD1-reactive NK T cells towards Th2 cytokine synthe-sis. Eur. J. Immunol. 29:2014.

69. Ortaldo, J. R., H. A. Young, R. T. Winkler-Pickett, E. W. J. Bere, W. J. Murphy,and R. H. Wiltrout. 2004. Dissociation of NKT stimulation, cytokine induction,and NK activation in vivo by the use of distinct TCR-binding ceramides. J. Im-munol. 172:943.

70. Burdin, N., L. Brossay, Y. Koezuka, S. T. Smiley, M. J. Grusby, M. Gui,M. Taniguchi, K. Hayakawa, and M. Kronenberg. 1998. Selective ability ofmouse CD1 to present glycolipids: �-galactosylceramide specifically stimulatesV�14� NK T lymphocytes. J. Immunol. 161:3271.

71. Ichikawa, S., and Y. Hirabayashi. 1998. Glucosylceramide synthase and glyco-sphingolipid synthesis. Trends Cell Biol. 8:198.

72. Huwiler, A., T. Kolter, J. Pfeilschifter, and K. Sandhoff. 2000. Physiology andpathophysiology of sphingolipid metabolism and signaling. Biochim. Biophys.Acta 1485:63.

73. Brigl, M., L. Bry, S. C. Kent, J. E. Gumperz, and M. B. Brenner. 2003. Mech-anism of CD1d-restricted natural killer T cell activation during microbial infec-tion. Nat. Immunol. 4:1230.

74. Giaccone, G., C. J. A. Punt, Y. Ando, R. Ruijter, N. Nishi, M. Peters,B. M. E. von Blomberg, R. J. Scheper, H. J. J. van der Vliet,A. J. M. van den Eertwegh, et al. 2002. A phase I study of natural killer T-cellligand �-galactosylceramide (KRN7000) in patients with solid tumors. Clin. Can-cer Res. 8:3702.

75. Nieda, M., M. Okai, A. Tazbirkova, H. Lin, A. Yamauro, K. Ide, R. Abraham,T. Juji, D. J. Macfarlane, and A. J. Nicol. 2004. Therapeutic activation ofV�24�V�11� NKT cells in human subjects results in highly coordinated sec-ondary activation of acquired and innate immunity. Blood 103:383.

76. Chiba, A., S. Oki, K. Miyamoto, H. Hashimoto, T. Yamamura, and S. Miyake.2004. Suppression of collagen-induced arthritis by natural killer T cell activationwith OCH, a sphingosine-truncated analog of �-galactosylceramide. ArthritisRheum. 50:305.

3706 THE IN VIVO RESPONSE OF NKT CELLS TO DISTINCT GLYCOLIPIDS

on July 22, 2010 w

ww

.jimm

unol.orgD

ownloaded from