Immunologic Glycosphingolipidomics and NKT Cell Development in

Mouse Thymus

Yunsen Li,† Prakash Thapa,† David Hawke,‡ Yuji Kondo,§ Keiko Furukawa,§ Koichi Furukawa,§

Fong-Fu Hsu,| Dietlind Adlercreutz,⊥ Joel Weadge,⊥ Monica M. Palcic,⊥ Peng G. Wang,#

Steven B. Levery,*,∇ and Dapeng Zhou*,†

Department of Melanoma Medical Oncology, Mass Spectrometry Core Facility, Department of Pathology, TheUniversity of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77054,

Department of Biochemistry II, Nagoya University School of Medicine, Tsurumai, Showa-ku,Nagoya 466-0065, Japan, Mass Spectrometry Resource, Division of Endocrinology, Diabetes, Metabolism, and

Lipid Research, Department of Internal Medicine, Washington University School of Medicine, St. Louis,Missouri 63110, Carlsberg Laboratory, Gamle Carlsberg Vej 10, 2500 Valby, Denmark, Department of

Biochemistry, The Ohio State University, Columbus, Ohio 43210, and Department of Cellular and MolecularMedicine, University of Copenhagen, Blegdamsvej 3, 2200 Copenhagen N, Denmark

Received December 1, 2008

Invariant NKT cells are a hybrid cell type of Natural Killer cells and T cells, whose development isdependent on thymic positive selection mediated by double positive thymocytes through theirrecognition of natural ligands presented by CD1d, a nonpolymorphic, non-MHC, MHC-like antigenpresenting molecule. Genetic evidence suggested that �-glucosylceramide derived glycosphingolipids(GSLs) are natural ligands for NKT cells. N-butyldeoxygalactonojirimycin (NB-DGJ), a drug thatspecifically inhibits the glucosylceramide synthase, inhibits the endogenous ligands for NKT cells.Furthermore, we and others have found a �-linked glycosphingolipid, isoglobotriaosylceramide (iGb3),is a stimulatory NKT ligand. The iGb3 synthase knockout mice have a normal NKT development andfunction, indicating that other ligands exist and remain to be identified. In this study, we have performeda glycosphingolipidomics study of mouse thymus, and studied mice mutants which are deficient in�-hexosaminidase b or R-galactosidase A, two glycosidases that are up- and downstream agents ofiGb3 turnover, respectively. Our mass spectrometry methods generated a first database for glycosph-ingolipids expressed in mouse thymus, which are specifically regulated by rate-limiting glycosidases.Among the identified thymic glycosphingolipids, only iGb3 is a stimulatory ligand for NKT cells,suggesting that large-scale fractionation, enrichment and characterization of minor species of gly-cosphingolipids are necessary for identifying additional ligands for NKT cells. Our results also provideearly insights into cellular lipidomics studies, with a specific focus on the important immunologicalfunctions of glycosphingolipids.

Keywords: Natural Killer T (NKT) cells • CD1d • T cell receptor • glycomics • lipidomics • glycosphin-golipids • isoglobotriaosylceramide (iGb3) • iGb3 synthase • linear ion trap mass spectrometry

Introduction

Glycosphingolipids (GSLs) are a class of biomacromoleculesassembled through stepwise action of glycosyltransferases.Composed of a hydrophobic ceramide part and a hydrophilic

complex carbohydrate part, GSLs are synthesized in the en-doplasmic reticulum and Golgi complex, where they traffic tothe outside layer of plasma membrane. The cell surface GSLsare constantly recycled to the lysosome where they aredegraded by glycosidases with the assistance of sphingolipidactivator proteins (saposins).1-5 In mouse and human tissues,GSLs of the globo-, isoglobo, lacto/neolacto-, ganglio-, gala-series have been identified.1-5 The globo, isoglobo, lacto/neolacto-, and ganglio-series of GSLs are assembled fromglucosylceramide, while gala-series of GSLs and sulfatides areassembled from galactosylceramide6 (Figure 1).

The recognition of self GSLs by T lymphocytes was firstreported by De Libero’s group,7-9 who found a series of T cellclones from multiple sclerosis patients recognize sulfatides andgangliosides. These glycosphingolpid antigens are presented

* To whom correspondence should be addressed. D.Z.: e-mail, dzhou@mdanderson.org;tel,713-7923134;fax,713-5633424.S.B.L.:e-mail, levery@imbg.ku.dk; tel, (45)3532-7779; fax, (45)3536-7980.

† Department of Melanoma Medical Oncology, The University of TexasM.D. Anderson Cancer Center.

‡ Mass Spectrometry Core Facility, Department of Pathology, The Uni-versity of Texas M.D. Anderson Cancer Center.

§ Nagoya University School of Medicine.| Washington University School of Medicine.⊥ Carlsberg Laboratory.# The Ohio State University.∇ University of Copenhagen.

2740 Journal of Proteome Research 2009, 8, 2740–2751 10.1021/pr801040h CCC: $40.75 2009 American Chemical SocietyPublished on Web 03/14/2009

by CD1 molecules, a family of non-MHC, nonpolymorphic,MHC-like antigen presenting molecules. GSLs also serve ascandidate antigens which mediate the development of a subsetof innate T lymphocytes which express Natural Killer cellmarkers, named as NKT cells. In contrast to conventional Tcells, NKT cells are self-reactive to mouse thymocytes,10-13

through recognition of endogenous ligands present in mousethymus. Invariant NKT cells (Type I) are the major NKT cells,which express invariant, evolutionally conserved T cell receptor(VR14 in mouse and VR24 in human), and are defined by theirrecognition of R-galactosylceramide, a marine sponge derivedGSL.14 However, all mammalian glycolipids stem from �-linkedmonohexosylceramides, �-glucosylceramide or �-galactosyl-ceramide, (1-6 Figure 1). R-Linked monohexosylceramidestructures have not been found in mammals, although theyhave been found in pathogenic bacteria which are related tohuman autoimmune disease primary biliary cirrhosis (No-vosphingobium aromaticivorans,15-17). R-Linked monohexo-syldiacylglycerol has been found in Borrelia burgdorferi, whichcauses Lyme disease,18 while presence of such structure inmammals has not been reported.

The GSLs derived from �-linked monohexosylceramides werefirst proposed as candidates for NKT ligands by Stanic et al.19

A mutant cell line which had suppressed expression of �-glu-cosylceramide synthase showed a significant defect in stimulat-ing NKT cells. In addition, the groups of Platt and Cerundolohave reported that N-butyldeoxygalactonojirimycin (NB-DGJ),a pharmaceutical which specifically inhibits the �-glucosylce-ramide synthase,20 inhibits the human dendritic cells’ abilityto stimulate NKT cells.21 Zhou, Teyton and Bendelac firstreported22 that in vivo development of invariant NKT cells isrelated to GSL metabolism; they showed that saposins, a familyof lipid transfer proteins which “grab” GSLs and present themfor glycosidase digestion,1 are required for NKT cell develop-ment. In vitro lipid transfer assays indicated that saposins arespecific for loading of GSLs to CD1d, while the loading of theother species of membrane lipids, phospholipids, does notrequire the assistance of saposins.22 The results of the lipidtransfer assays are consistent with the fact that the develop-ment of type II NKT cells, which are also CD1d dependent butrecognize different lipids, is not impaired in saposin knockout(KO) mice. We and others also reported that the in vivo

development of NKT cells may be impaired by defects in GSLtrafficking to the lysosome, and general lysosomal storage ofGSLs may compete with the loading of NKT ligands.23-26 NPC1mice, with a defect in glycolipid trafficking to late endosome,showed severe defect in NKT cell devel-opment.23-25 GM1 gangliosidosis �-galactosidase KO miceshowed a defect in NKT cell development with a mechanismhypothesized to be competition of NKT ligand loading byaccumulated lysosomal gangliosides.24,25 Finally, we andothers found that a GSL, isoglobotrihexosylceramide (GalR3-Gal�4Glc�1Cer, iGb3), is recognized both by mouse VR14 andhuman VR24 natural killer T (NKT) cells.27-32

The effects of iGb3 have been found by several groups to besubstantially different from those of R-galactosylceramide.These stimulatory properties of iGb3 are unique for theextensively studied TCR repertoire (V�8, V�7, and V�2) ofendogenous ligands,29,31,33 leading to the suggestion that iGb3might account for the well-known autoreactivity of NKT cells,and might also carry out important roles in NKT cell develop-ment and/or function, particularly in the numerous noninfec-tious disease conditions in which they have been implicated.However, by studying iGb3 synthase knockout mice, Porubskyet al.34 have challenged the physiological relevance of iGb3,since no defect in NKT development was found in these mice.

The identities of natural ligands for invariant NKT cellsremain the most critical unanswered questions in the NKT field.A clear and comprehensive understanding of GSL and non-GSL metabolism and distribution patterns in mouse thymusis critical for studies on NKT biology. Since we and others havealready found that a GSL, iGb3, is a natural ligand for themajority of invariant NKT cell population, we have chosen firstto study the problem using sensitive and specific mass spec-trometry (MS) based glycosphingolipidomics.

It is generally accepted that MS is an indispensable methodfor structural glycomics studies, especially for identifying andcharacterizing low-abundance ligands.35-38 We recently com-bined all of the potential advantages of electrospray ionizationlinear ion trap mass spectrometry (ESI-LIT-MS) methodology,including the detection of isomeric structures using signaturediagnostic ions, observable only in MS4 and MS5 spectra, forthe highly sensitive identification and quantitation of GSLspresent in the form of multiple isobaric mixtures.39,40 Here, wehave applied MALDI-TOF-MS and ESI-LIT-MSn techniques toanalyze neutral and acidic GSLs purified from mouse thymus.In addition, we have studied the �-hexosamidase b KO andFabry mice, the two mouse models of glycolipid storagediseases that were used in our previous studies on NKTbiology.22,28 We expect that glycosphingolipidomic analysis ofthese 2 strains of mutant mice will provide a basis for clarifyingthe discrepancies in results reported by three differentlaboratories.22,25,28,41

Experimental Procedures

Mice. Hexb KO mouse strain was from Dr. Richard L. Proiaat NIDDK, NIH.42 Fabry mice was from Dr. Ashok B. Kulkarni,NIDCR, NIH.43 Gb3 synthase KO mice was generated andmaintained as described.44 All mice were backcrossed to C57B6background for more than 10 generations. Mice aged between6 and 8 weeks were used for experiments. Thymuses werecollected and pooled for biochemical analysis.

Fetal Thymus Organ Culture. Fetal thymuses from day 14embryo of C57B6 mice were cultured as described.22 For drugtreatment, NB-DGJ (Toronto Research Chemicals, Canada) was

Figure 1. Glycosphingolipid synthesis in mouse. All mouse GSLsstem from �-glucosylceramide or �-galactosylceramide, and areassembled by key glycosyltransferaes to become lacto-, ganglio-,globo, isoglobo, sulfo-series of complex glycosphingolpids.N-butyldeoxygalactonojirimycin (NB-DGJ) specifically inhibits thesynthesis of glucosylceramide derived glycosphingolipids, byinhibiting the glucosylceramide synthase. �4GalT, �1,4 galacto-syltransferase 6 (lactosylceramide synthase); CST, cerebrosidesulfotransferase.

Immunologic Glycosphingolipidomics and NKT Cell Development research articles

Journal of Proteome Research • Vol. 8, No. 6, 2009 2741

added since beginning of the 14 day culture period, at theconcentration of 10 and 100 µg/mL.

Bone Marrow Derived Dendritic Cells. Mouse bone marrowderived dendritic cells were generated by culturing bonemarrow progenitor cells in presence of GM-CSF and IL4. Thetotal culturing period was 14 days. Both immature and maturedendritic cells were studied. To mature the dendritic cells, CpGODN (ODN 1826) was added to culture medium in the last dayof culture at the concentration of 2 µg/mL.

Extraction of Glycosphingolipids. GSLs were extracted asdescribed.39,40 Mouse thymus or mouse bone marrow deriveddendritic cells were stored in 16 × 100 mm glass tubes at -80°C. Lipids were extracted by extensive sonication for four timeswith mixed polarity solvents. The first and last solvent usedwas chloroform-methanol 1:1 (v/v). The second and thirdsolvent used was isopropanol-hexane-water 55:25:20 (v/v/v,upper phase removed by aspiration before use). Sonication wasfollowed by centrifugation to pellet the insoluble material. Thesupernatants were pooled and dried under nitrogen stream at40 °C, and subjected to preliminary analysis by high perfor-mance thin layer chromatography (HPTLC).

Separation of Neutral and Acidic Lipids. Neutral and acidiclipids were fractionated by anion exchange chromatographyon a small column of DEAE Sephadex A-25. The acidic lipidfraction was eluted with 0.8 M sodium acetate in methanol.Both neutral and acidic fractions were dried, desalted bydialysis, dried, and analyzed by HPTLC.

Florisil Fractionation of Neutral GSLs. The DEAE SephadexA-25 pass-through fraction was dried under vacuum in thepresence of P2O5 for 3 h and peracetylated with 1 mL ofpyridine and 0.5 mL of acetic anhydride in the dark at roomtemperature overnight. The peracetylated material was dried,with the addition of 2 mL of toluene 3× to ensure completeevaporation. A Florisil (Sigma-Aldrich, St. Louis, MO) column(30-60 mesh, 10 × 80 mm) was equilibrated in 1,2-dichloroethane-hexane 4:1 (v/v). The peracetylated sample wasapplied in this solvent, and the column was then washed with20 mL of the same solvent, followed by 20 mL of 1,2-dichloroethane. Neutral peracetylated GSLs were eluted with40 mL of 1,2-dichloroethane-acetone 1:1 (v/v). The fractionswere dried and deacetylated with 1 mL of 0.5 M sodiummethoxide in 2 mL of methanol for 3 h at room temperature.The mixture was neutralized with methanolic acetic acid, dried,and then desalted by dialysis. The GSL contents of all threefractions were compared by HPTLC.

Per-N,O-Methylation of GSLs. GSLs (1-20 µg) were intro-duced into a conical-glass vial, and dimethyl sulfoxide (150 µL)was added without using special drying conditions or inert gasatmosphere. Powdered sodium hydroxide (40-60 mg) was thenadded to the sample solution, and was stirred at roomtemperature until completely dissolved. Iodomethane (80 µL)was added with a syringe, and the mixture was shaked at roomtemperature for 1 h. The methylation reaction was quenchedwith water (2 mL). The permethylated products were extracted3 times by addition of dichloromethane (2 mL). The combineddichloromethane extracts were then washed three times withwater (2 mL each). Following the final wash, the samples weretransferred to a new tube, and dried under nitrogen stream at35-40 °C.

MALDI-TOF Mass Spectrometry Analysis. MALDI-TOF-MSand MALDI-TOF/TOF-MS/MS experiments were performed ona MALDI TOF-TOF Mass Spectrometer (Applied Biosystems4700, Foster City, CA). The matrix was prepared by mixing 50%

volume of 10 mg/mL dihydroxybenzoic acid (DHB) in aceto-nitrile and 50% volume of 0.1% trifluoroacetic acid (TFA) inwater. The DHB matrix was spotted on the MALDI target (1µL), followed by a 1 µL (containing 100 ng GSLs) spot of sampledissolved in methanol, and allowed to dry before analysis. MSexperiments were acquired using the reflectron settings in thepositive mode. MS/MS data were obtained using the 1 kV modewith argon or air as collision gas (CID cell gas pressure 3.5 ×10-6 to 9 × 10-7 Torr). MS spectra were summed from 1000 to10 000 laser shots; for MS/MS, 5000-50 000 shots were summed.

Electrospray Ionization Linear-Ion-Trap Mass Spectrometry(ES-LIT-MSn) of Permethylated Neutral GSLs. Mass spectrom-etry (MS and MSn) was carried out in positive ion mode on alinear ion trap mass spectrometer (LTQ, ThermoFinnigan, SanJose, CA), using a nanoelectrospray source for direct infusionof samples (dissolved in methanol) by static nanospray, atcapillary temperature 230 °C, with injection time as 100.00 ms,activation time as 30 ms, activation Q-value as 0.250, isolationwidth as m/z 1.5, and acquisition time as 3-10 min. Electro-spray voltage was 1 kV. Static nanoelectrospray needles werefrom Proxeon Biosciences (Denmark). Normalized collisionenergies were set to leave a minimal residual abundance ofprecursor ion; in this case, 30% was used for all product ionscans. All ions were detected as sodium adducts. To obtain MS5

spectra specifically from the characteristic internal HO-3Gal3/4Gal-1-ene disaccharides, each iGb4/Gb4 molecular species (orlipoform, observed at, e.g., m/z X), was subjected to multistepfragmentation via the MS5 pathway as described.39 To obtainMS4 spectra specifically from the characteristic nonreducingGal3/4Gal-1-ene disaccharides, each iGb3/Gb3 molecular spe-cies (or lipoform, observed at, e.g., m/z X), was subjected tomultistep fragmentation via the MS4 pathway as described.40

Artificial iGb3 and Asiao-GM1 as Quantitation Standard.A chemically synthesized, “short-tail” iGb3 (C8), was added asinternal standard before permethylation experiments, as pub-lished previously.40 Asialo-GM1 (Matreya Cat. No. 1064) wasused as internal standard for analyzing charged GSLs.

Other GSL Standards for Ion-Trap MS FragmentationAssays. Lactosylceramide (Matreya Cat. No. 1500), Gb3 (MatreyaCat. No. 1067), Gb4 (Matreya Cat. No. 1068), asialo-GM1(Matreya Cat. No. 1064), GM1 (Matreya Cat. No. 1061), asialo-GM2 (Matreya Cat. No. 1512), GM2 (Matreya Cat. No. 1502),GM3 (Matreya Cat. No. 1503), GM4 (Matreya Cat. No. 1535),GD1a (Matreya Cat. No. 1062), GD1b (Matreya Cat. No. 1501),GD3 (Matreya Cat. No. 1504), GT1b (Matreya Cat. No. 1063),GQ1b (Matreya Cat. No. 1516), and sulfatide (Matreya Cat. No.1049) were from Matreya, Pleasant Gap, PA. Synthetic sugarpart of Gb4, GM1, asialo-GM1, iGb3, and Gb3 were providedby Drs. Ola Blixt and James Paulson, Consortium of FunctionalGlycomics. iGb4 and digalactosylceramide (GalR1,4Gal Cer),purified from cat intestine, were from Dr. Sussan Tenenberg,Gorteberg University.

Stimulation toward NKT Cell Hybridomas. GSLs werestored in DMSO at 2 mg/mL. GSLs were pulsed to 50 000 mouseBMDCs at 100 ng/mL overnight, and mixed with 50 000 NKTcell hybridomas overnight. IL2 released by NKT hybridomaswere measured by CTLL assays. NKT hybridomas wereDN32.D322 and 23.8A (provided by Dr. Laurent Gapin).29

Flow Cytometry Analysis of NKT Cells. NKT cells werestained by PBS57 (RGalCer analogue) loaded mouse CD1dtetramer (APC color, provided by NIAID tetramer facility atEmory University, GA), in combination with anti-CD3 and anti-CD44 monoclonal antibodies (eBiosciences, CA).

research articles Li et al.

2742 Journal of Proteome Research • Vol. 8, No. 6, 2009

Results

NB-DGJ treatment Abolishes the Development of InvariantNKT Cells. NB-DGJ treatment of cultured fetal thymusescompletely blocked the development of NKT cells (Figure 2).In contrast, the development of conventional CD4 and CD8 Tcells remained unchanged. This was observed at both 100 and10 µg/mL concentration of drug treatment.

Mass Spectrometry Assays Generate First Profiles ofNeutral and Acidic Glycosphingolipids in Mouse Thymus. Togain a clear and precise view of the GSL structures in mousethymus, we used both MALDI-TOF-MS and ESI-LIT-MSn

techniques. Table 1 lists the GSLs that were identified from 6-8week mouse thymi. Representative MALDI-TOF-MS profiles forneutral and acidic GSLs are presented in Supporting Informa-tion Figure 1. The identities of GSLs were verified by MALDI-TOF/TOF-MS/MS and ESI-LIT-MSn analysis, with the fragmen-tation patterns compared to those of authentic GSL standards(Suppoting Information Figure 2).

Glycosidases Specifically Regulate Glycosphingolipid Pro-files in Glycosidase Mutant Mice. R-Galactosidase and �-hex-osaminidase are two enzymes involved in GSL degradation.NKT cell development has been studied in both mutant mice,and contradictory findings have been reported by differentgroups. To understand the effects of disruption of theenzyme activities on the relevant components at the bio-chemical level, we have performed glycosphingolipidomicanalysis of these mutant mice strains. To compare therelative abundance of GSLs detected in Hexb KO mice, Fabrymice, and wild-type mice, we have added an internalstandard (artificial iGb3 for neutral GSLs, and asialo-GM1(Gg4) for acidic GSLs, respectively). Figure 3 shows the ratioof each GSL in Hexb KO mice and Fabry mice, with respectto the same component in wild type mice, based oncomparison of relative ion abundances of each detected GSLto the artificial internal standards (Supporting InformationTable 1). We found total Gb3 + iGb3 triglycosylceramide (6.5-

fold ratio to wild-type), and total Gb4 + iGb4 tetraglycosyl-ceramide (8-fold ratio to wild-type) were accumulated inFabry mice, along with a diglycosylceramide fraction and anadditional triglycosylceramide fraction containing a terminalHexNAc residue, whose exact structure remains to bestudied. In Hexb mice, we found an accumulation of Gb4/iGb4 (2.4-fold), Lc3 + asialo-GM2 (15.5-fold), GM2 (4.2-fold),GD3 (4.5-fold), and GD1 (2.2-fold). On the other hand, asevere reduction (>95%) was found for iGb3 + Gb3 GSLs.

Low-Abundance of Thymic iGb3 Is Only Detectable byESI-LIT-MSn Technology. To address the controversy thatiGb3 is not expressed in mouse thymus,34,45 we analyzed itsexpression by a more sensitive mass spectrometry methodusing a linear ion-trap (ESI-LIT-MSn). Trihexosylceramidemolecular species observed in MS1 represent both regioiso-mers (iGb3 + Gb3), but we have found that the expressionof iGb3 can be detected and measured from characteristicions produced in MS4 analysis of selected molecular ions40

(Figure 4). We observed multiple molecular ion species ofiGb3 + Gb3 regioisomers in MS1, due to the variation in theirfatty acyl chain lengths; each was subjected to MS4 analysisas described. The expression of iGb3 was detected as higherthan 1% in molecular ions of m/z 1277 (1.5%) and m/z 1325(3.5%) observed in MS1(Supplementary Table 2). In mostother trihexosylceramide ions observed in MS1, iGb3 wasfound to be less than 1%, or undetectable compared to itsregioisomer Gb3. The total iGb3 to Gb3 ratio was estimatedas 0.4%.

The expression of iGb3 is undetectable in Hexb KO mice,due to more than 95% overall reduction of trihexosylceramidelevels (iGb3 + Gb3). The expression of iGb3 was increased inR-galactosidase A KO mice, easily detected due to the 6.5-foldoverall increase of trihexosylceramides (iGb3 + Gb3). Note-worthy, the iGb3/Gb3 ratio was only slightly increased in Fabrymice, suggesting that the iGb3 and Gb3 are digested byR-galactosidase with similar efficiency.

We further measured relative levels of iGb4, the lysosomalprecursor of iGb3, by MS5 analysis of molecular ions represent-ing regioisomers Gb4 and iGb4.39 We found a similarly lowexpression of iGb4 in thymus of the wild-type mice (Figure 4).iGb4 + Gb4 was increased 2.4-fold in Hexb KO mice, and 7-foldin Fabry KO mice (Figure 4).

Mass Spectrometry Assays Reveal Identities of Neutraland Acidic Glycosphingolipids in Mouse Bone MarrowDerived Dendritic Cells. Bone marrow derived dendritic cells(BMDCs) have been widely used to study the biology of NKTcells. It is generally accepted that BMDCs express naturalligands for NKT cells. To have a clear and precise view of theGSL structures expressed in BMDCs, we used both MALDI-TOF/TOF-MS/MS and ESI-LIT-MSn techniques. The GSLsidentified are listed in Table 2. Representative MALDI-TOF-MS profiles for neutral GSLs and acidic GSLs are presented inSupporting Information Figure 3.

CpG Specifically Regulate GSL Profiles in Dendritic Cells. Tostudy whether the maturation process of dendritic cells cor-relates with alteration of GSL expression, we compared the ratioof GSLs detected in mature DCs to those in immature DCs.Figure 5 shows increased levels of Gb3 + iGb3 (2.4-fold) andGb4 + iGb4 (2.3-fold) expression, based on comparison ofrelative ion abundances of each detected GSL to artificialinternal standards (Supporting Information Table 3). The othersignificant increase was in the level of Forssman antigen (4.3-

Figure 2. NB-DGJ completely abolishes the NKT cell develop-ment. Fetal thymus organ culture was performed as previouslydescribed.22 Thymuses from fetal mice (day 14 after pregnancy)were cultured in 6 well transwell plates (Corning, NY) in EHAA/RPMI medium with 10% Fetal Calf Serum. NB-DGJ (10 and 100µg/mL) was added at the beginning of culture. The thymuseswere cultured to allow the development of NKT cells. After 14days, the thymocytes were harvested and analyzed by flowcytometry after staining by RGalCer/CD1d tetramer and anti-mouse CD44 monoclonal antibody (Pharmingen, CA). To excludethe possibility of toxicity to thymocytes, the development ofconventional CD4 and CD8 T cells was used as control. Noreduction of CD4 or CD8 numbers was found in either percentageor total cell numbers.

Immunologic Glycosphingolipidomics and NKT Cell Development research articles

Journal of Proteome Research • Vol. 8, No. 6, 2009 2743

fold). No significant changes were found for any acidic GSLsin CpG matured DCs.

The expression of iGb3 in mouse bone marrow deriveddendritic cells was also investigated. The results show that iGb3is detectable in certain trihexosylceramide molecular ions atlevels of 1-2.5% (Supporting Information Figure 4), while inmany other trihexosylceramide molecular ions, iGb3 is below1% or undetectable (Supporting Information Table 4). The totaliGb3 to Gb3 ratio was estimated as 0.8% in both immature andmature DCs.

iGb3 versus Gb3 Ratio Increases More than 10-fold inGb3 Synthase KO Mice. Since the detection of iGb3 ishampered by the overwhelming abundance of its regioisomerGb3, we hypothesized that we should be able to measure higherratios of iGb3 in trihexosylceramide molecular species if weexamine Gb3 synthase knockout mice. This will also addressthe possibility that Gb3 could give rise to artifactual signalsresembling those from iGb3 in ESI-LIT-MS4 spectra, that is,through rearrangement or some other process, fragmentation

of the terminal disaccharide MS3 product of Gb3 (GalR1-4Gal-1-ene) in MS4 could give rise to background levels of production signals putatively specific for iGb3 (GalR1-3Gal-1-ene).These background signals were observed to be below 0.2% forthe m/z 211 product, and below 0.4% for the m/z 371 product,in multiple experiments performed in three different massspectrometry core facilities.40 If we have measured artifactualiGb3 ion signals in MS4 which are solely background Gb3fragmentation products, we should not detect an increase ofthe iGb3/Gb3 ratio in Gb3 synthase KO mice.

In Figure 6, we compare the iGb3/Gb3 ratio in Gb3 syn-thase-/-, Gb3 synthase+/-, and wild-type mice. A more than10-fold increase of the iGb3/Gb3 ratio, from 1.7% to 26.5%, wasfound in Gb3 synthase KO mice (Supplementary Table 5), whenselecting the molecular ion m/z 1325. The total iGb3 to Gb3ratio in Gb3 synthase KO mice was estimated as 7.5%.Noteworthy, we still found a presence of Gb3 (5% versus wild-type control mice) in Gb3 synthase KO mice. A slight increaseof iGb3/Gb3 ratio was found in Gb3 synthase+/- mice as well

Table 1. Glycosphingolipids Identified in Mouse Thymus

a The presence of Glc-Cer in the thymus of C57BL6 strain of mice was previous reported.52 We have found Glc �1 Cer is the major structure by NMRanalysis (data not shown). b The disaccharide-ceramide structure in mouse thymus is a mixture of Lac-Cer and Digalactosylceramide. This was verified bycomparing the MS3 ion patterns, analyzed by ion trap mass spectrometry to standard Lac-Cer and Digalactosylceramide (Supporting InformationFigure 2). c The trisaccharide-ceramide structures of mouse thymus remain to be studied. Preliminary analysis by ion trap mass spectrometry suggestedthe structures of Lc3 and asialo-GM2 (data not shown).

research articles Li et al.

2744 Journal of Proteome Research • Vol. 8, No. 6, 2009

(2- to 3-fold), which is consistent with our previous finding thatGb3 synthase activity was decreased 50% in heterozygousmutant mice.44

We further studied the NKT cells in the thymus of Gb3synthase KO mice and found normal NKT cell numbers inthymus, spleen, and bone marrow (Supporting InformationFigure 5). This finding is consistent with our previous reportthat Gb3 is not a stimulatory ligand for NKT cells,28 since Gb3synthase KO mice have much reduced Gb3 expression notmeasurable by anti-Gb3 antibody staining.44

Major Species of Glycosphingolipids Do Not StimulateNKT Cells. To identify new candidate GSL antigens for NKTcells, we examined the stimulatory activities of all of themtoward the DN32.D3 and 24.8A hybridomas. We found nodetectable stimulation by any of these GSLs except iGb3(Supporting Information Figure 6).

Discussion

Profiles of Neutral and Acidic Glycosphingolipids in MouseThymus. In this study, we report a mass spectrometry-basedglycosphingolipidomic analysis utilizing state-of-the-art MALDI-TOF/TOF-MS/MS and ESI-LIT-MSn instrumental techniques,of the major GSL species in mouse thymus and bone marrowderived dendritic cells. Compared to previous studies usingHPLC and fast atom bombardment mass spectrometry (FAB-MS), our results demonstrated a number of significant advan-tages:

(1) we clearly observed presence of GSL structures notpreviously reported by others,21,34,45,46 including iGb3and iGb4;

(2) our results were generated with very small amounts ofsample (107 mouse bone marrow derived dendritic cells);

(3) tedious HPLC or TLC purification of individual GSLs isavoided;

(4) the method is generally applicable, and results have beenreproduced in more than one mass spectrometry corefacility.

A significant part of this study required permethylation ofthe GSL samples, which increases the sensitivity and specificityof carbohydrate structure detection by mass spectrometry, butmay destroy or leave undetected sulfated GSL structures.Detection of sulfated GSLs was carried out on native GSLs usingESI-MSn with a precursor ion scanning technique,47 whichenables specific searching for GSLs containing a sulfo group.With this technique, we identified sulfatide molecular speciesin mouse thymus (Supporting Information Figure 7).

In comparison to HPLC-based quantitation methods, andto the traditional use of MS for qualitative glycan profiling andstructure analysis, quantitation of glycoconjugates by MSremains a relatively less traveled field. In this study, we havecompared the relative ion abundance of neutral GSLs to thatof an artificial iGb3, and the acidic GSLs to asialo-GM1. Thiscomparison allowed us to calculate the relative ratios of GSLsin Hexb KO mice and Fabry mice to wild-type mice. However,it is difficult to calculate the absolute quantity of each GSL,because significantly different GSLs may have different ioniza-tion efficiencies. A GSL standard for each specific GSL, havingthe same ionization efficiency as the analyte, is required forabsolute quantitation. For example, we have shown that theartificial iGb3 has same ionization efficiency as natural iGb3expressed in mammalian cells.40 In the future, we propose tosynthesize additional specific “short-tail” GSL standards forquantitation purpose.

Specificity of Glycosidases in Regulating GlycosphingolipidExpression. R-Galactosidase KO and Hexb KO mice displayspecific accumulation of GSLs terminated by sugars that arecleavable by specific glycosidases. Our results indicate thatnormal GSL degradative processing is only selectively impairedin R-galactosidase KO mice or Hexb KO mice. The finding thatHexb KO mice lack iGb3 expression is consistent with ourprevious hypothesis that the Hexb enzyme is required forgeneration of iGb3.28 We found that Hexb KO cells showed aspecific defect in processing and presenting HexNAc termi-nated glycolipid antigens (iGb4 and GalNAc�1,3GalR1,1Cer),but not R-Gal terminated complex glycolipids (GalR1,2GalR1,1Cer) and RGalCer antigen.28 Studies from the Cerundololaboratory contradicted our findings; Gadola et al.25 reportedinstead that Hexb KO cells could not process and present GalR1,2 Gal R Cer, nor could they present the surrogate antigenfor NKT cells, RGalCer. These discrepancies could be explainedby differences in age and genetic background between mousestrains. Indeed, we have already found clear biochemicaldifferences between the Hexb KO mice used by two laborato-ries. The Hexb KO mice used by the Cerundolo group, asstudied by Speak et al.,45 showed a similar amount of thymicGb3 as compared to wild type mice, while we observed morethan 95% reduction of Gb3 in thymus from our Hexb KO strain.

GSL antigen processing and presentation in Fabry mice werefirst studied by the Kronenberg group.41 In their study, R-ga-lactosidase A deficient cells were completely defective inpresentation of GalR1,2GalR1,1Cer, although no defect wasfound in presenting RGalCer. Prigozy et al.41 found a 50%reduction of NKT cells in the periphery (spleen), but NKT cells

Figure 3. Specific alterations of glycosphingolipids in Hexb KOmice and Fabry mice. Thymic GSLs from Hexb KO, Fabry, andwild-type mice were extracted, permethylated, and analyzed byMALDI-MS. The relative ion abundance of MS1 ions of Hex-Hex-Cer (Gal �1,4 GlcCer and Gal R1,4 GalCer), Hex-Hex-Hex-Cer (Gb3,Gal R1,4 Gal �1,4 GlcCer and iGb3, Gal R1,3 Gal �1,4 GlcCer),HexNAc-Hex-Hex-Hex-Cer (Gb4, Gal �1,4 Gal R1,4 Gal �1,4GlcCer, and iGb4, Gal �1,4 Gal R1,3 Gal �1,4 GlcCer), HexNAc-Hex-Hex-Cer (Lc3, GlcNAc �1,3 Gal �1,4 GlcCer, and asialo-GM2,GalNAc �1,4 Gal �1,4 GlcCer), GM2, GD3, GM1, GD1, GT1, andGQ1 were compared to internal standards (“short tail” iGb3 forneutral GSLs, and asialo-GM1 for acidic GSLs). The ratio of aboveGSLs in Hexb mice and Fabry mice versus wild-type mice wascalculated according to Supporting Information Table 1. Gal-terminated GSLs were accumulated in Fabry mice. HexNActerminated GSLs were accumulated in Hexb KO mice. Gb3 andiGb3 were absent in Hexb KO mice.

Immunologic Glycosphingolipidomics and NKT Cell Development research articles

Journal of Proteome Research • Vol. 8, No. 6, 2009 2745

were normal in the thymus. When we first discovered thatsaposins are required for NKT cell development, we comparedthe presentation of GalR1,2GalR1,1Cer and RGalCer by saposinKO cells to that of R-galactosidase KO cells.22 We found a defectin presenting RGalCer by saposin KO cells, but not R-galac-tosidase KO cells, while neither saposin KO nor R-galactosidaseKO cells could present GalR1,2GalR1,1Cer. Similar to Kronen-berg’s group, we could not find a defect of NKT cells in thethymus of R-galactosidase KO mice. We have found normalnumbers of NKT cells in thymus and spleen of 4-8 week oldFabry mice.22 The defect (90% reduction in thymus) reportedby Gadola et al.25 may be due to a different Fabry mouse strainused in their study. They could find Gb3, but no other GSLs,in the thymus of their Fabry mice;45 in contrast, we observedan 8-fold increase of Gb4 expression in our Fabry mice(Supporting Information Figure 1). It is also interesting to notethat, although iGb3 expression is increased (5-fold) in our Fabrymice, we did not find significant increase of stimulation towardNKT cells by thymocytes from Fabry mice22 compared to wild-type mice, and the underlying mechanisms remain to bestudied.

We used Hexb KO mice in our discovery of the first naturalligand for NKT cells, through the reasoning that iGb3 is cleavedfrom iGb4 by �-hexosaminidases. However, we agree with mostinvestigators in the field that iGb3 may not be the only naturalligand. We have pointed this out since our first report of theantigenic activity of iGb3, and in review articles.27,28 Since wefound iGb3 by an indirect approach, testing GSL candidatesthat could be generated in the lysosome by �-hexosaminidases,we cannot exclude the possibility that other natural ligandscould also be present. The exact mechanisms responsible for

the defect of NKT cell development in Hexb KO mice remainunclear. It may be due to defect in generating natural ligandssuch as iGb3, a secondary effect of GSL accumulation (Figure3, especially gangliosides), or other more severe secondaryeffects which may even cause defect in loading of R-galacto-sylceramide, as reported by Gadola et al.25 in a different HexbKO mouse strain. A complete view of all these biochemical andcell biological parameters is still currently missing.

The current study is the first to describe in detail the GSLprofiles of glycosidase mutant immune cells. GSL profiles inHexb KO and Fabry mice have been previously characterizedat the organ level, including brain, kidney, and liver. TheSandhoff group reported the accumulation of HexNAc termi-nated GSLs in kidney of Hexb KO mice.48 The Shayman groupreported that Gb3 was increased in liver and kidney of Fabrymice, while the level of Gb4, present at a similar amount toGb3, was not significantly changed.49,50 These results suggestthat expression of GSLs in different cell types is regulated bycomplex processes. The regulation may operate at the tran-scriptional level of glycosyltransferases and glycosidases, throughpost-translational trafficking of these glycoenzymes and theiractivator proteins (such as saposins), and through lysosomalacidification.51

Low-Abundance of iGb3 Expression by Mouse Thymusand Mouse Dendritic Cells. We found that the level of iGb3 inmouse immune cells is overwhelmed by expression of the Gb3regioisomer. In most trihexosylceramide molecular speciesdetectable at the MS1 level, the abundance of iGb3 MS4 productions is less than 1% or undetectable. However, levels of iGb3higher than 1% were detected upon MS4 analysis of particulartrihexosylceramide molecular species corresponding to certain

Figure 4. Expression of iGb3 and iGb4 in mouse thymus. Neutral GSLs from wild-type, Hexb KO, and Fabry mice were extracted,permethylated, and analyzed by ion trap mass spectrometry. MS1 ions representing regioisomers of iGb3/Gb3 were selected for MS4

analysis according to published method.40 MS1 ions representing regioisomers of iGb4/Gb4 were selected for MS5 analysis.39 Percentageof iGb3 in iGb3/Gb3 mixtures were calculated according to standard curve as described.40

research articles Li et al.

2746 Journal of Proteome Research • Vol. 8, No. 6, 2009

ceramide compositions (lipoforms). Our data are consistentwith findings by the Platt group,45 that overall the level of iGb3is below 1% of total iGb3 and Gb3 mixture. However, becausethe ESI-LIT-MSn method can measure the iGb3 levels inindividual lipoforms (differing in this case by m/z incrementscorresponding to fatty-N-acyl chain length and unsaturation),the presence of iGb3 can be differentially detected in a fewmolecular species, even when it is below the limit of detect-ability in the majority of them. In contrast, the HPLC methodemployed by the Platt group involves cleavage of the glycanfrom the ceramides of all iGb3 and Gb3 species prior toanalysis. Even if the resulting mixture of isomeric glycans werewell-resolved in the subsequent HPLC separation, the relativesensitivity of iGb3 detection in a much larger pool of Gb3 isparadoxically reduced, rather than increased, if its expressionwas confined to a limited set of lipoforms. The ability to detectiGb3 expression on a single lipoform out of many is anoteworthy advantage of the MSn methodology.

It remains to be studied whether such a low level of iGb3expression in thymocytes and dendritic cells is sufficient forproviding stimulatory signals to NKT cells at an in vivo setting,although it is generally accepted that single peptide-MHCcomplex expression in an antigen presenting cell is sufficientto trigger TCR signaling. In addition to direct TCR stimulation

by natural ligands, costimulatory pathways may contribute tothe full activation program of NKT cells. We and others haverecently found that B7 family of costimulatory molecules arecritical for the second wave of NKT cell expansion during theirdevelopment.52-54

The physiological relevance of iGb3 has been further chal-lenged by the groups of Platt and Sandrin,45,55,56 based on theirfinding that iGb3 is absent in mouse and human immune cells(thymus and dendritic cells). The Platt group employed anHPLC method to assay oligosaccharides cleaved from totalthymic GSLs following fluorescence-labeling;45 the methodrelies purely on retention behavior for identification. TheSandrin group used a monoclonal antibody (15.101) to stainhuman tissue sections,55,56 although the specificity of theirantibody remains questionable.57 Both laboratories have con-cluded that iGb3 is not expressed in humans, but in contrastto their findings, we have detected iGb3 and its lysosomalprecursor iGb4 in human thymus and dendritic cells by an iontrap mass spectrometry method39,40 (and manuscript in prepa-ration), which is considerably more sensitive than HPLCmethod,40,45 and more specific than antibody staining.57 Chris-tiansen et al.56 further concluded that human iGb3 synthase isa pseudogene, based on negative finding by RT-PCR experi-ments, and the negative finding that a hybrid rat-human iGb3

Table 2. Glycosphingolipids Identified in Mouse Dendritic Cells

Immunologic Glycosphingolipidomics and NKT Cell Development research articles

Journal of Proteome Research • Vol. 8, No. 6, 2009 2747

synthase enzyme failed to generate products that react toantibody 15.101. However, the antibody staining method usedto measure glycosyltransferase activities by detection of theirproducts has a far lower sensitivity compared to mass spec-trometry methods,39,40 or conventional methods based onmeasuring the transfer of radioisotope labeled sugar donors.

The enzymatic mechanism of iGb3 synthesis in mousethymus is not clearly understood yet. We first cloned the mouseiGb3 synthase from mouse thymus.28 Subsequently, mRNAblotting analysis by Milland et al. showed that this enzyme ismost abundantly expressed in mouse thymus, as compared toother organs.36 Thus, it is difficult to understand why iGb3 isexpressed at such a low abundance in thymus. We have studiedwhether other R1,3-glycosyltransferases can synthesize iGb3,and found R1,3-galactosyltransferase 1 (R3GalT1) producediGb3 when transfected to Chinese Hamster Ovary (CHO) cells(Supporting Information Figure 8). With the use of recombinantbovine R3GalT1 protein, we also synthesized iGb3 in a cell-free system (data not shown). However, the efficiency ofR3GalT1 in synthesizing iGb3 is 10-fold lower than iGb3synthase in CHO cell transfection experiments, suggesting thelatter is the major enzyme. The low level of iGb3 in mousethymus may be due to yet unknown negative regulators onenzyme activity. Alternatively, there may exist enzyme-cofactorsrequired for efficient assembly of iGb3.

Stimulatory Activity toward NKT Cells by Major Speciesof Glycosphingolipids. In our preliminary screening for stimu-latory activities of the GSLs, we have not found any others thatare stimulatory ligands besides iGb3. Further studies arecurrently under way to measure the stimulatory activities ofthese ligands using additional NKT cell clones, as well as freshNKT cells. On the other hand, we reason that minor species ofGSLs can only be purified and characterized through large-scale fractionation of GSLs from pooled mouse thymuses.

Currently, there are two series of lipid species to consideras candidates for the major natural ligands of NKT cells:phospholipids and glycosphingolipids. Supporting the involve-ment of phospholipids, Gumperz et al.58 reported that phos-

phatidylcholine and phosphatidylinosital stimulate a few NKTcell hybridomas in plate-bound assays. Evidence supporting arole for GSLs includes genetic data that glucosylceramidesynthase is involved in natural ligand production;19 that NB-DGJ inhibits the antigen presenting cells’ ability to stimulateNKT cells21 (Figure 2); and that charged GSLs from maturedendritic cells stimulate fresh liver NKT cells in an IFN-Rdependent manner.59 Although our data (Figure 2) and datafrom the groups of Joyce19 and Cerundolo21 strongly supportthat GSL metabolism influences the production of naturalligand, it is still not conclusive that the major natural ligandsmust be GSLs. Specifically, glucosylceramide has been reportedas critical for intracellular trafficking of certain membraneproteins. For example, in the melanoma tumor cell line B16mutant GM95, which does not synthesize glucosylceramide dueto inactivation of �-glucosylceramide synthase, a severe defectwas found in the trafficking of melanin from the endoplasmicreticulum to the melanosome compartment, leading to a“white” phenotype of the cell line,60 although the traffickingof CD1d was normal in GM95 cells.19 At this time, we cannotexclude non-GSLs as the candidates for natural ligands.

In summary, our study provides the first in depth view ofGSL expression in mouse immune cells, and should provide abasis for future identification of potential thymic GSL ligandsfor NKT cells. This study also demonstrated the presence ofiGb3, the first natural ligand, in mouse thymus, and highlightedthe importance of biochemical studies by sensitive, specific,and generally applicable methodologies.

Figure 5. Alteration of GSL pattern in mature dendritic cells. GSLsfrom immature and CpG-ODN-activated mature bone marrowderived dendritic cells were extracted, permethylated, and ana-lyzed by MALDI-MS. The relative ion abundance of MS1 ions ofHex-Hex-Cer (Gal �1,4 GlcCer and Gal R1,4 GalCer), Hex-Hex-Hex-Cer (Gb3, Gal R1,4 Gal �1,4 GlcCer and iGb3, Gal R1,3 Gal�1,4 GlcCer), HexNAc-Hex-Hex-Hex-Cer (Gb4, Gal �1,4 Gal R1,4Gal �1,4 GlcCer, and iGb4, Gal �1,4 Gal R1,3 Gal �1,4 GlcCer),HexNAc-Hex-Hex-Cer (Lc3, GlcNAc �1,3 Gal �1,4 GlcCer, andasialo-GM2, GalNAc �1,4 Gal �1,4 GlcCer), Forssman antigen,asialo-GM1, GD1, and GT1 were compared to internal standards(artificial iGb3 and asialo-GM1). The ratio of above GSLs in CpG-ODN matured dendritic cells versus immature dendritic cells werecalculated according to Supporting Information Table 3.

Figure 6. Increased ratio of iGb3 in Gb3 synthase KO mice.Neutral GSLs from wild type, Gb3 synthase+/- and Gb3 syn-thase-/- mice were extracted, permethylated, and analyzed byion trap mass spectrometry. MS1 ions representing regioisomersof iGb3 and Gb3 were selected for MS4 analysis. Percentage ofiGb3 in iGb3 and Gb3 mixtures were calculated according tostandard curve as described.40

research articles Li et al.

2748 Journal of Proteome Research • Vol. 8, No. 6, 2009

Conclusions

The search for natural ligands for NKT cells has been ongoingfor more than a decade without considerable progress. Ourstudy indicates these ligands contain unknown structureswhich are not reported as major GSLs. Similarly, previousstudies by others61 showed that major membrane phospho-lipids with known structures are not stimulatory ligands for themajority population of NKT cells. We propose a “lipidomics”approach, that is, one that identifies new lipid structures andtests their immunological activity, with a priority on thosestructures similar to known ligands of NKT cells. The generationof first lipidomics database for antigen presenting cells in theimmune system will be a fundamental step, and open up an anew avenue to identify lipid and glycolipid targets for theimmune-related disorders which have heretofore been chal-lenging to study.

Abbreviations: R1,3GT, R-1,3 galactosyltransferase I; ESI-LIT-MS, electrospray ionization linear ion trap mass spec-trometry; Forssman antigen, GalNAcR1,3 GalNAc�1,3GalR1,4Gal�1,4Glc�1,1Cer; Glc-Cer, glucosylceramide; Gal-Cer, ga-lactosylceramide; GM3, NeuAcR2,3Gal�1,4Glc�1,1Cer; GM2,GalNAc�1, 4(NeuAcR2,3)Gal�1,4Glc�1,1Cer; GSL, glycosph-ingolipid; iGb3, isoglobotrihexosylceramide (GalR1,3Gal�1,4Glc�1,1Cer); Lac-Cer, lactosylceramide; Lc3, GlcNAc�1,3Gal�1,4Glc�1,1Cer; MALDI-TOF-MS, matrix-assisted laser desorp-tion/ionization time-of-flight mass spectrometry; NB-DGJ,N-butyldeoxygalactonojirimycin; TLC, thin layer chroma-tography. The glycosphingolipid nomenclature follows therecommendations of the IUPAC-IUB Commission on Bio-chemical Nomenclature.62

Acknowledgment. This work was supported by a startup fund from M.D. Anderson Cancer Center, and aDevelopmental Award from NCI Joe Moakley Leukemia SPOREgrant (1P50 CA100632-04) to D.Z. and an NIH grant (R21RR20355) to S.B.L. The authors thank Drs. Ola Blixt and JimPaulson (Consortium of Functional Glycomics, the ScrippsResearch Institute) for synthetic carbohydrate standards; Dr.Vernon N. Reinhold for providing the MS facilities of the UNHCenter for Structural Biology (NIH/NCRR grant P20 RR16459)for these studies; Drs. Albert Bendelac, Jochen Mattner, andBhanu Prakash Pappu for discussion.

Supporting Information Available: Tables of assign-ments of molecular ions ([M + Na]+) observed in MALDI-TOFspectra of permethylated neutral and acidic GSLs of Hexb KO,Fabry and wild-type mouse thymus; percentage of iGb3 inisobaric iGb3-Gb3 mixture from WT mouse thymic neutralGSLs; assignments of molecular ions ([M + Na]+) observed inMALDI-TOF spectra of permethylated neutral and acidic GSLsof mouse immature and mature DCs; percentage of iGb3 inisobaric iGb3-Gb3 mixture from neutral GSLs of immature andmature mouse DCs; percentage of iGb3 in isobaric iGb3-Gb3mixture in WT and Gb3 synthase KO (Gb3 Syn-/-) mice. MS1

profiles of thymic GSLs from wild-type, Hexb KO and Fabrymice; MS-MS profiles of standard GSLs; MS1 profiles of GSLsin immture and mature dendritic cells; expression of iGb3 andiGb4 in mouse dendritic cells; normal development of NKT cellsin Gb3 synthase KO mice; stimulation to mouse NKT hybri-doma clones by GSLs identified in mouse thymus; sufatideexpression in mouse thymus as determined by precursor ionscan of 97 (sulfo residue) on neutral GSLs; synthesis of iGb3by mouse iGb3 synthase and R1,3-galactosyltransferase 1. This

material is available free of charge via the Internet at http://pubs.acs.org.

References(1) Kolter, T.; Sandhoff, K. Principles of lysosomal membrane diges-

tion: stimulation of sphingolipid degradation by sphingolipidactivator proteins and anionic lysosomal lipids. Annu. Rev. CellDev. Biol. 2005, 21, 81–103.

(2) Yamakawa, T.; Suzuki, A.; Hashimoto, Y. Genetic control ofglycolipid expression. Chem. Phys. Lipids 1986, 42 (1-3), 75–90.

(3) Schwarzmann, G.; Sandhoff, K. Metabolism and intracellulartransport of glycosphingolipids. Biochemistry 1990, 29 (49), 10865–71.

(4) Ichikawa, S.; Hirabayashi, Y. Glucosylceramide synthase andglycosphingolipid synthesis. Trends Cell Biol 1998, 8 (5), 198–202.

(5) Lloyd, KO.; Furukawa, K. Biosynthesis and functions of ganglio-sides: recent advances. Glycoconjugate J. 1998, 15 (7), 627–36.

(6) Suzuki A. Biosynthetic pathways of glycosphingolipids. In Hand-book of Glycosyltransferases and Related Genes; Taniguchi, N.,Honke, K., Fukuda, M., Eds.; Springer-Verlag: Tokyo, 2002; pp 636-646.

(7) De Libero, G.; Mori, L. Self glycosphingolipids: new antigensrecognized by autoreactive T lymphocytes. News Physiol. Sci. 2003,18, 71–6.

(8) Shamshiev, A.; Donda, A.; Carena, I.; Mori, L.; Kappos, L.; DeLibero, G. Self glycolipids as T-cell autoantigens. Eur. J. Immunol.1999, 29 (5), 1667–75.

(9) Shamshiev, A.; Donda, A.; Prigozy, T. I.; Mori, L.; Chigorno, V.;Benedict, C. A.; Kappos, L.; Sonnino, S.; Kronenberg, M.; De Libero,G. The alphabeta T cell response to self-glycolipids shows a novelmechanism of CD1b loading and a requirement for complexoligosaccharides. Immunity 2000, 13 (2), 255–64.

(10) Bendelac, A. Positive selection of mouse NK1+ T cells by CD1-expressing cortical thymocytes. J. Exp. Med. 1995, 182 (6), 2091–6.

(11) Zimmer, M. I.; Colmone, A.; Felio, K.; Xu, H.; Ma, A.; Wang, C. R.A cell-type specific CD1d expression program modulates invariantNKT cell development and function. J. Immunol. 2006, 176 (3),1421–30.

(12) Brossay, L.; Tangri, S.; Bix, M.; Cardell, S.; Locksley, R.; Kronenberg,M. Mouse CD1-autoreactive T cells have diverse patterns ofreactivity to CD1+ targets. J. Immunol. 1998, 160 (8), 3681–8.

(13) Park, S. H.; Roark, J. H.; Bendelac, A. Tissue-specific recognitionof mouse CD1 molecules. J. Immunol. 1998, 160 (7), 3128–34.

(14) Kawano, T.; Cui, J.; Koezuka, Y.; Toura, I.; Kaneko, Y.; Motoki, K.;Ueno, H.; Nakagawa, R.; Sato, H.; Kondo, E.; Koseki, H.; Taniguchi,M. CD1d-restricted and TCR-mediated activation of valpha14 NKTcells by glycosylceramides. Science 1997, 278 (5343), 1626–9.

(15) Mattner, J.; Debord, K. L.; Ismail, N.; Goff, R. D.; Cantu, C., III.;Zhou, D.; Saint-Mezard, P.; Wang, V.; Gao, Y.; Yin, N.; Hoebe, K.;Schneewind, O.; Walker, D.; Beutler, B.; Teyton, L.; Savage, P. B.;Bendelac, A. Exogenous and endogenous glycolipid antigensactivate NKT cells during microbial infections. Nature (London)2005, 434 (7032), 525–9.

(16) Kinjo, Y.; Wu, D.; Kim, G.; Xing, G. W.; Poles, M. A.; Ho, D. D.;Tsuji, M.; Kawahara, K.; Wong, C. H.; Kronenberg, M. Recognitionof bacterial glycosphingolipids by natural killer T cells. Nature(London) 2005, 434 (7032), 520–5.

(17) Sriram, V.; Du, W.; Gervay-Hague, J.; Brutkiewicz, R. R. Cell wallglycosphingolipids of Sphingomonas paucimobilis are CD1d-specific ligands for NKT cells. Eur. J. Immunol. 2005, 35 (6), 1692–701.

(18) Kinjo, Y.; Tupin, E.; Wu, D.; Fujio, M.; Garcia-Navarro, R.; Benhnia,M. R.; Zajonc, D. M.; Ben-Menachem, G.; Ainge, G. D.; Painter,G. F.; Khurana, A.; Hoebe, K.; Behar, S. M.; Beutler, B.; Wilson,I. A.; Tsuji, M.; Sellati, T. J.; Wong, C. H.; Kronenberg, M. Naturalkiller T cells recognize diacylglycerol antigens from pathogenicbacteria. Nat. Immunol. 2006, 7 (9), 978–86.

(19) Stanic, A. K.; De Silva, A. D.; Park, J. J.; Sriram, V.; Ichikawa, S.;Hirabyashi, Y.; Hayakawa, K.; Van Kaer, L.; Brutkiewicz, R. R.; Joyce,S. Defective presentation of the CD1d1-restricted natural Va14Ja18NKT lymphocyte antigen caused by beta-D-glucosylceramidesynthase deficiency. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (4),1849–54.

(20) Andersson, U.; Butters, T. D.; Dwek, R. A.; Platt, F. M. N-butyldeoxygalactonojirimycin: a more selective inhibitor of gly-cosphingolipid biosynthesis than N-butyldeoxynojirimycin, in vitroand in vivo. Biochem. Pharmacol. 2000, 59 (7), 821–9.

(21) Salio, M.; Speak, A. O.; Shepherd, D.; Polzella, P.; Illarionov, P. A.;Veerapen, N.; Besra, G. S.; Platt, F. M.; Cerundolo, V. Modulation

Immunologic Glycosphingolipidomics and NKT Cell Development research articles

Journal of Proteome Research • Vol. 8, No. 6, 2009 2749

of human natural killer T cell ligands on TLR-mediated antigen-presenting cell activation. Proc. Natl. Acad. Sci. U.S.A. 2007, 104(51), 20490–5.

(22) Zhou, D.; Cantu, C., III.; Sagiv, Y.; Schrantz, N.; Kulkarni, A. B.; Qi,X.; Mahuran, D. J.; Morales, C. R.; Grabowski, G. A.; Benlagha, K.;Savage, P.; Bendelac, A.; Teyton, L. Editing of CD1d-bound lipidantigens by endosomal lipid transfer proteins. Science 2004, 303(5657), 523–7.

(23) Sagiv, Y.; Hudspeth, K.; Mattner, J.; Schrantz, N.; Stern, R. K.; Zhou,D.; Savage, P. B.; Teyton, L.; Bendelac, A. Cutting edge: impairedglycosphingolipid trafficking and NKT cell development in micelacking Niemann-Pick type C1 protein. J. Immunol. 2006, 177 (1),26–30.

(24) Schumann, J.; Facciotti, F.; Panza, L.; Michieletti, M.; Compostella,F.; Collmann, A.; Mori, L.; De Libero, G. Differential alteration oflipid antigen presentation to NKT cells due to imbalances in lipidmetabolism. Eur. J. Immunol. 2007, 37 (6), 1431–41.

(25) Gadola, S. D.; Silk, J. D.; Jeans, A.; Illarionov, P. A.; Salio, M.; Besra,G. S.; Dwek, R.; Butters, T. D.; Platt, F. M.; Cerundolo, V. Impairedselection of invariant natural killer T cells in diverse mouse modelsof glycosphingolipid lysosomal storage diseases. J. Exp. Med. 2006,203 (10), 2293–303.

(26) Schrantz, N.; Sagiv, Y.; Liu, Y.; Savage, P. B.; Bendelac, A.; Teyton,L. The Niemann-Pick type C2 protein loads isoglobotrihexosylce-ramide onto CD1d molecules and contributes to the thymicselection of NKT cells. J. Exp. Med. 2007, 204 (4), 841–52.

(27) Zhou, D. The immunological function of iGb3. Curr. Protein Pept.Sci. 2006, 7 (4), 325–33.

(28) Zhou, D.; Mattner, J.; Cantu, C., 3rd.; Schrantz, N.; Yin, N.; Gao,Y.; Sagiv, Y.; Hudspeth, K.; Wu, Y. P.; Yamashita, T.; Teneberg, S.;Wang, D.; Proia, R. L.; Levery, S. B.; Savage, P. B.; Teyton, L.;Bendelac, A. Lysosomal glycosphingolipid recognition by NKTcells. Science 2004, 306, 1786–9.

(29) Scott-Browne, J. P.; Matsuda, J. L.; Mallevaey, T.; White, J.; Borg,N. A.; McCluskey, J.; Rossjohn, J.; Kappler, J.; Marrack, P.; Gapin,L. Germline-encoded recognition of diverse glycolipids by naturalkiller T cells. Nat. Immunol. 2007, 8 (10), 1105–13.

(30) Xia, C.; Yao, Q.; Schumann, J.; Rossy, E.; Chen, W.; Zhu, L.; Zhang,W.; De Libero, G.; Wang, P. G. Synthesis and biological evaluationof alpha-galactosylceramide (KRN7000) and isoglobotrihexosyl-ceramide (iGb3). Bioorg. Med. Chem. Lett. 2006, 16 (8), 2195–9.

(31) Schumann, J.; Mycko, M. P.; Dellabona, P.; Casorati, G.; Mac-Donald, H. R. Cutting edge: influence of the TCR Vbeta domainon the selection of semi-invariant NKT cells by endogenousligands. J. Immunol. 2006, 176 (4), 2064–8.

(32) Yu, K. O.; Im, J. S.; Illarionov, P. A.; Ndonye, R. M.; Howell, A. R.;Besra, G. S.; Porcelli, S. A. Production and characterization ofmonoclonal antibodies against complexes of the NKT cell ligandalpha-galactosylceramide bound to mouse CD1d. J. Immunol.Methods 2007, 323 (1), 11–23.

(33) Wei, D. G.; Curran, S. A.; Savage, P. B.; Teyton, L.; Bendelac, A.Mechanisms imposing the Vbeta bias of Valpha14 natural killer Tcells and consequences for microbial glycolipid recognition. J. Exp.Med. 2006, 203 (5), 1197–207.

(34) Porubsky, S.; Speak, A. O.; Luckow, B.; Cerundolo, V.; Platt, F. M.;Grone, H. J. Normal development and function of invariant naturalkiller T cells in mice with isoglobotrihexosylceramide (iGb3)deficiency. Proc. Natl. Acad. Sci. U.S.A. 2007, 104 (14), 5977–82.

(35) Levery, S. B. Glycosphingolipid structural analysis and glycosph-ingolipidomics. Methods Enzymol. 2005, 405, 300–69.

(36) Haslam, S. M.; Julien, S.; Burchell, J. M.; Monk, C. R.; Ceroni, A.;Garden, O. A.; Dell, A. Characterizing the glycome of the mam-malian immune system. Immunol. Cell Biol. 2008, 86 (7), 564–73.

(37) Ashline, D.; Singh, S.; Hanneman, A.; Reinhold, V. Congruentstrategies for carbohydrate sequencing. 1. Mining structural detailsby MSn. Anal. Chem. 2005, 77 (19), 6250–62.

(38) Reinhold, V. N.; Reinhold, B. B.; Costello, C. E. Carbohydratemolecular weight profiling, sequence, linkage, and branching data:ES-MS and CID. Anal. Chem. 1995, 67 (11), 1772–84.

(39) Li, Y.; Teneberg, S.; Thapa, P.; Bendelac, A.; Levery, S. B.; Zhou,D. Sensitive detection of isoglobo and globo series tetraglycosyl-ceramides in human thymus by ion trap mass spectrometry.Glycobiology 2008, 18 (2), 158–65.

(40) Li, Y.; Zhou, D.; Xia, C.; Wang, P. G.; Levery, S. B. Sensitivequantitation of isoglobotriaosylceramide in the presence of iso-baric components using electrospray ionization-ion trap massspectrometry. Glycobiology 2008, 18 (2), 166–76.

(41) Prigozy, T. I.; Naidenko, O.; Qasba, P.; Elewaut, D.; Brossay, L.;Khurana, A.; Natori, T.; Koezuka, Y.; Kulkarni, A.; Kronenberg, M.Glycolipid antigen processing for presentation by CD1d molecules.Science 2001, 291 (5504), 664–7.

(42) Sango, K.; McDonald, M. P.; Crawley, J. N.; Mack, M. L.; Tifft, C. J.;Skop, E.; Starr, C. M.; Hoffmann, A.; Sandhoff, K.; Suzuki, K.; Proia,R. L. Mice lacking both subunits of lysosomal beta-hexosaminidasedisplay gangliosidosis and mucopolysaccharidosis. Nat. Genet.1996, 14 (3), 348–52.

(43) Ohshima, T.; Murray, G. J.; Swaim, W. D.; Longenecker, G.; Quirk,J. M.; Cardarelli, C. O.; Sugimoto, Y.; Pastan, I.; Gottesman, M. M.;Brady, R. O.; Kulkarni, A. B. alpha-Galactosidase A deficient mice:a model of Fabry disease. Proc. Natl. Acad. Sci. U.S.A. 1997, 94(6), 2540–4.

(44) Okuda, T.; Tokuda, N.; Numata, S.; Ito, M.; Ohta, M.; Kawamura,K.; Wiels, J.; Urano, T.; Tajima, O.; Furukawa, K.; Furukawa, K.Targeted disruption of Gb3/CD77 synthase gene resulted in thecomplete deletion of globo-series glycosphingolipids and loss ofsensitivity to verotoxins. J. Biol. Chem. 2006, 281 (15), 10230–5.

(45) Speak, A. O.; Salio, M.; Neville, D. C.; Fontaine, J.; Priestman, D. A.;Platt, N.; Heare, T.; Butters, T. D.; Dwek, R. A.; Trottein, F.; Exley,M. A.; Cerundolo, V.; Platt, F. M. Implications for invariant naturalkiller T cell ligands due to the restricted presence of isoglobotri-hexosylceramide in mammals. Proc. Natl. Acad. Sci. U.S.A. 2007,104 (14), 5971–6.

(46) Schwarting, G. A.; Gajewski, A. Glycolipids of murine lymphocytesubpopulations. Structural characterization of thymus ganglio-sides. J. Biol. Chem. 1983, 258 (9), 5893–8.

(47) Hsu, F. F.; Turk, J. Studies on sulfatides by quadrupole ion-trapmass spectrometry with electrospray ionization: structural char-acterization and the fragmentation processes that include anunusual internal galactose residue loss and the classical charge-remote fragmentation. J. Am. Soc. Mass Spectrom. 2004, 15 (4),536–46.

(48) Sandhoff, R.; Hepbildikler, S. T.; Jennemann, R.; Geyer, R.; Gie-selmann, V.; Proia, R. L.; Wiegandt, H.; Grone, H. J. Kidneysulfatides in mouse models of inherited glycosphingolipid disor-ders: determination by nano-electrospray ionization tandem massspectrometry. J. Biol. Chem. 2002, 277 (23), 20386–98.

(49) Abe, A.; Arend, L. J.; Lee, L.; Lingwood, C.; Brady, R. O.; Shayman,J. A. Glycosphingolipid depletion in fabry disease lymphoblastswith potent inhibitors of glucosylceramide synthase. Kidney Int.2000, 57 (2), 446–54.

(50) Abe, A.; Gregory, S.; Lee, L.; Killen, P. D.; Brady, R. O.; Kulkarni,A.; Shayman, J. A. Reduction of globotriaosylceramide in Fabrydisease mice by substrate deprivation. J. Clin. Invest. 2000, 105(11), 1563–71.

(51) Trombetta, E. S.; Ebersold, M.; Garrett, W.; Pypaert, M.; Mellman,I. Activation of lysosomal function during dendritic cell maturation.Science 2003, 299 (5611), 1400–3.

(52) Chung, Y.; Nurieva, R.; Esashi, E.; Wang, Y. H.; Zhou, D.; Gapin,L.; Dong, C. A critical role of costimulation during intrathymicdevelopment of invariant NK T cells. J. Immunol. 2008, 180 (4),2276–83.

(53) Williams, J. A.; Lumsden, J. M.; Yu, X.; Feigenbaum, L.; Zhang, J.;Steinberg, S. M.; Hodes, R. J. Regulation of thymic NKT celldevelopment by the B7-CD28 costimulatory pathway. J. Immunol.2008, 181 (2), 907–17.

(54) Zheng, X.; Zhang, H.; Yin, L.; Wang, C. R.; Liu, Y.; Zheng, P.Modulation of NKT cell development by B7-CD28 interaction: anexpanding horizon for costimulation. PLoS ONE 2008, 3 (7), e2703.

(55) Milland, J.; Christiansen, D.; Lazarus, B. D.; Taylor, S. G.; Xing,P. X.; Sandrin, M. S. The molecular basis for galalpha(1,3)galexpressioninanimalswithadeletionofthealpha1,3galactosyltransferasegene.J. Immunol. 2006, 176 (4), 2448–54.

(56) Christiansen, D.; Milland, J.; Mouhtouris, E.; Vaughan, H.; Pellicci,D. G.; McConville, M. J.; Godfrey, D. I.; Sandrin, M. S. Humanslack iGb3 due to the absence of functional iGb3-synthase: implica-tions for NKT cell development and transplantation. PLoS Biol.2008, 6 (7), e172.

(57) Zhou, D.; Levery, S. B. Response to Milland et al.: Carbohydrateresidues downstream of the terminal Galalpha(1,3)Gal epitopemodulate the specificity of xenoreactive antibodies. Immunol. CellBiol. 2008, 86 (8), 631–2.

(58) Gumperz, J. E.; Roy, C.; Makowska, A.; Lum, D.; Sugita, M.;Podrebarac, T.; Koezuka, Y.; Porcelli, S. A.; Cardell, S.; Brenner,M. B.; Behar, S. M. Murine CD1d-restricted T cell recognition ofcellular lipids. Immunity 2000, 12 (2), 211–21.

(59) Paget, C.; Mallevaey, T.; Speak, A. O.; Torres, D.; Fontaine, J.;Sheehan, K. C.; Capron, M.; Ryffel, B.; Faveeuw, C.; Leite deMoraes, M.; Platt, F.; Trottein, F. Activation of invariant NKT cellsby toll-like receptor 9-stimulated dendritic cells requires type Iinterferon and charged glycosphingolipids. Immunity 2007, 27 (4),597–609.

research articles Li et al.

2750 Journal of Proteome Research • Vol. 8, No. 6, 2009

(60) Sprong, H.; Degroote, S.; Claessens, T.; van Drunen, J.; Oorschot,V.; Westerink, B. H.; Hirabayashi, Y.; Klumperman, J.; van der Sluijs,P.; van Meer, G. Glycosphingolipids are required for sortingmelanosomal proteins in the Golgi complex. J. Cell Biol. 2001, 155(3), 369–80.

(61) Parekh, V. V.; Singh, A. K.; Wilson, M. T.; Olivares-Villagomez, D.;Bezbradica, J. S.; Inazawa, H.; Ehara, H.; Sakai, T.; Serizawa, I.;Wu, L.; Wang, C. R.; Joyce, S.; Van Kaer, L. Quantitative and

qualitative differences in the in vivo response of NKT cells todistinct alpha- and beta-anomeric glycolipids. J. Immunol. 2004,173 (6), 3693–706.

(62) IUPAC-IUB Joint Commission on Biochemical Nomenclature(JCBN) Nomenclature of glycolipids. Eur. J. Biochem. 1998, 257,293–298.

PR801040H

Immunologic Glycosphingolipidomics and NKT Cell Development research articles

Journal of Proteome Research • Vol. 8, No. 6, 2009 2751