of April 18, 2012This information is current as

http://www.jimmunol.org/content/184/3/1235doi:10.4049/jimmunol.0804140December 2009;

2010;184;1235-1241; Prepublished online 21J Immunol Mori, Gennaro De Libero and Michael B. BrennerManuela Cernadas, Marco Cavallari, Gerald Watts, Lucia Presentation of Lipid AntigensCD1a Is Essential for Its Intersection and Early Recycling Compartment Trafficking of

References http://www.jimmunol.org/content/184/3/1235.full.html#ref-list-1

, 12 of which can be accessed free at:cites 33 articlesThis article

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

is online atThe 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/etoc/subscriptions.shtml/

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

Print ISSN: 0022-1767 Online ISSN: 1550-6606.Immunologists, Inc. All rights reserved.

by The American Association ofCopyright ©2010 9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

The Journal of Immunology

Early Recycling Compartment Trafficking of CD1a IsEssential for Its Intersection and Presentation ofLipid Antigens

Manuela Cernadas,*,1 Marco Cavallari,†,1 Gerald Watts,‡ Lucia Mori,† Gennaro De Libero,†,2

and Michael B. Brenner‡,2

A major step in understanding differences in the nature of Ag presentation was the realization that MHC class I samples peptides

transported to the endoplasmic reticulum from the cytosol, whereas MHC class II samples peptides from lysosomes. In contrast to

MHC class I and II molecules that present protein Ags, CD1 molecules present lipid Ags for recognition by specific T cells. Each of

the five members of the CD1 family (CD1a–e) localizes to a distinct subcompartment of endosomes. Accordingly, it has been widely

assumed that the distinct trafficking of CD1 isoforms must also have evolved to enable them to sample lipid Ags that traffic via

different routes. Among the CD1 isoforms, CD1a is unusual because it does not have a tyrosine-based cytoplasmic sorting motif

and uniquely localizes to the early endocytic recycling compartment. This led us to predict that CD1a might have evolved to focus

on lipids that localize to early endocytic/recycling compartments. Strikingly, we found that the glycolipid Ag sulfatide also

localized almost exclusively to early endocytic and recycling compartments. Consistent with colocalization of CD1a and sulfatide,

wild-type CD1a molecules efficiently presented sulfatide to CD1a-restricted, sulfatide-specific T cells. In contrast, CD1a:CD1b tail

chimeras, that retain the same Ag-binding capacity as CD1a but traffic based on the cytoplasmic tail of CD1b to lysosomes, failed

to present sulfatide efficiently. Thus, the intracellular trafficking route of CD1a is essential for efficient presentation of lipid Ags

that traffic through the early endocytic and recycling pathways. The Journal of Immunology, 2010, 184: 1235–1241.

TheCD1 family of Ag-presenting molecules is unique in itsability to present lipid, glycolipid, and lipopeptide Ags toCD1-restricted T cells. In contrast to MHC-restricted Ag

presentation, CD1 molecules are functionally nonpolymorphic andhave been shown to present both exogenous microbial Ags as well asendogenous lipid ligands (1). Similar toMHCclass I molecules, CD1molecules are noncovalently associated with b2-microglobulin.CD1a, b, and c molecules are primarily expressed on professionalAPCs and have been implicated in adaptive immunity against mi-crobial lipids. In contrast, CD1d-restricted NK T cells are innate-like lymphocytes that may bridge the innate and adaptive immunesystem. Each of the fivemembers of the humanCD1 family—CD1a,b, c, d, and e—has a distinct cellular distribution and intracellulartrafficking pattern. The intracellular localization of the CD1b, c, and

d isoforms is determined by their unique cytoplasmic tail tyrosine-based sorting motif (2). The intracellular localization of the CD1bisoform, for example, is almost exclusively in the lysosomal com-partment. This CD1b localization is mediated by the adaptor proteinAP-3 through interactions with the cytoplasmic tyrosine-basedmotif of CD1b (3).The CD1a isoform is almost exclusively expressed on pro-

fessional APC, including Langerhans cells (LCs) and other myeloiddendritic cell (DC) subsets (2). CD1a is distinct among the CD1isoforms in that it does not contain a tyrosine-based cytoplasmicmotif. CD1a has been shown to localize intracellularly to the en-docytic recycling compartment (ERC) (4, 5). In contrast to CD1b,CD1a internalization into endosomes is independent of clathrin ordynamin. Once internalized, CD1amolecules follow a Rab22a- andADP ribosylation factor 6-dependent recycling pathway (5). Understeady-state conditions, CD1a is not found to localize to late en-dosomes (LEs) or lysosomes (LYs) (4). Moreover, CD1a moleculeshave also been shown to localize in Birbeck granules, which areintracellular compartments unique to LCs that have been demon-strated to be subdomains of the endocytic system (6).Several T cell Ags presented by CD1a have been described in-

cluding didehydroxymycobactin, a lipopeptidic Ag isolated fromMycobacterium tuberculosis, and sulfatide (7, 8). Sulfatide (39-sulfated b1-D-galactosylceramide) is an endogenous glycolipidhighly expressed in neuronal cells, kidney, and pancreas (9–11), andits synthesis is upregulated in DCs upon bacterial infection (12).Sulfatide can be presented by all group 1 CD1 molecules includingmouse and human CD1d (8, 13, andM. Cavallari and G. De Libero,unpublished results). DC pulsed with sulfatide maintained theability to stimulate CD1a-restricted T cells over a 3 d period. Incontrast, the ability of the sulfatide-pulsed DC to stimulate CD1b-and CD1c-restricted T cells was reduced by ∼75% over the sametime period. These findings suggest that CD1a has a unique ability

*Division of Pulmonary and Critical Care Medicine and ‡Division of Rheumatology,Immunology and Allergy, Brigham and Women’s Hospital, Harvard Medical School,Boston, MA 02115; and †Experimental Immunology, Department of Biomedicine,University Hospital Basel, Basel, Switzerland

1M.C. and M.C. contributed equally to this work.

2G.D.L. and M.B.B. contributed equally to this work.

Received for publication December 11, 2008. Accepted for publication November17, 2009.

This work was supported by grants from the National Institutes of Health (AI 028973to M.B.B.), the Swiss National Science Foundation (3100AO-109918), and RocheResearch Foundation (to G.D.L.).

Address correspondence and reprint requests to Dr. Michael B. Brenner, Division ofRheumatology, Immunology and Allergy, Brigham and Women’s Hospital, HarvardMedical School, Smith Building, Room 552, 1 Jimmy Fund Way, Boston, MA 02115.E-mail address: mbrenner@rics.bwh.harvard.edu

Abbreviations used in this paper:DC, dendritic cell; EEA-1, early endosomeAg-1; ERC,endocytic recycling compartment; LAM, lipoarabinomannan; LAMP-1, lysosome-associatedmembrane protein 1; LC, Langerhans cell; LE, late endosome; LY, lysosome;Trf, transferrin; WT, wild-type.

Copyright� 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0804140

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

to present sulfatide. However, these studies cannot distinguish be-tween higher affinity of CD1a binding for sulfatide versus moreextensive colocalization of theAg-presentingmoleculewith theAg.In this study, we show that the intracellular localization of the

CD1a isoform to the early endocytic system allows it to intersect withlipids that are primarily or, in the case of sulfatide, almost exclusively,localized in the early endocytic system. This colocalization wasshown to be critical for the sustained ability of CD1a molecules tostimulate CD1a-restricted T cells.When the sameCD1a extracellulardomain was fused to the CD1b tail, redirecting it to LYs, efficient andprolonged presentation of CD1a lipid Agwas abrogated. This findinghas significant implications for the efficient and sustained pre-sentation of other CD1a-binding lipid Ags by DC in vivo. This workalso highlights the importance of understanding lipid trafficking asa critical factor in the study of CD1 Ag presentation and function.

Materials and MethodsChimeric CD1a:CD1b constructs

Three types of CD1a:CD1b hybrid constructs containing sequencesencoding the CD1a extracellular domain fused to three different CD1bchimeric tails were generated by PCR using human CD1a and CD1b cDNAas the templates. The first CD1a:b tail chimeric construct contains thetyrosine-based motif of CD1b and two additional 39 amino acids fromCD1b wild-type (WT) sequence (highlighted in boldface) swapped withthe terminal sequence (last three amino acids) of CD1a [i.e., WT CD1a (…FRKRCFC) was changed to (… FRKRRSYQNIP)]. The CD1a:CD1bchimeric tail construct was cloned between the BamHI and XhoI sites ofpcDNA3 vector (Invitrogen, Carlsbad, CA) and sequenced before trans-fection. This construct is referred to in the text as CD1aab, indicating theorigin of the extracellular domain (first letter), transmembrane (secondletter), and cytoplasmic sequences (CD1b). The second construct is es-sentially identical to the first with the exception of one amino acid. In thesecond CD1a:CD1b tail chimeric construct, the WT CD1a (…FRKRCFC)tail was changed to (…FRRRSYQNIP) as previously described (14). Thisconstruct, which was used to transfect the T2 cells, will also be referred toCD1aab. The CD1aab constructs were independently generated, account-ing for the one amino acid difference. Each laboratory included additionalamino acids from the transmembrane region of CD1b because replacementof the tyrosine motif alone was not sufficient to fully redirect CD1a to theLY-associated membrane protein-1+ (LAMP-1+) lysosomal compartments.In the third construct called CD1abb, the extracellular CD1a region wasjoined to the transmembrane and cytoplasmic sequences of CD1b. Thelatter two sequences were cloned into pBluescript II KS+ (Stratagene, LaJolla, CA), sequenced, and subcloned into the XhoI-NotI sites of theBCMGSNeo vector for transfection. All PCR primer sequences used togenerate the constructs are available upon request.

Cells

HeLa cells (American Type Culture Collection, Manassas, VA) weretransfected with WT CD1a (CD1aaa) and HeLa CD1aab chimeric con-structs by electroporation (Bio-Rad Gene Pulser II, 280V, 960 mF; Bio-Rad,

Hercules, CA). Stable clones were isolated by limiting dilution under se-lection with 1 mg/ml G418 sulfate (Invitrogen) in DMEM complete media.Transient transfections were performed with FuGENE 6 (Roche, Basel,Switzerland) as per manufacturer’s instructions. CD1a cell surface ex-pression was confirmed by flow cytometric staining and intracellular traf-ficking by confocal microscopy as detailed below. Monocyte-derived DCswere generated as previously described (15). Briefly, CD14+ monocyteswere isolated from buffy coats obtained from normal blood donors as perinstitutional guidelines. Mononuclear cells were isolated with Ficoll-PaquePLUS (GE Healthcare, Buckinghamshire, U.K.) and monocytes positivelyselected with CD14 MicroBeads (MACS; Miltenyi Biotec, Auburn, CA).DCs were differentiated in RPMI 1640 complete media (Life Technologies,Carlsbad, CA), 10% heat-inactivated FCS (HyClone, Logan, UT), 2 mM L-glutamine, 10 mM HEPES, and 100 U/ml penicillin-streptomycin (LifeTechnologies) supplemented with 300 U/ml GM-CSF (Immunex, ThousandOaks, CA) and 200 U/ml IL-4 (PeproTech, Rocky Hill, NJ). WT CD1a-(16), CD1aab- (14), or CD1abb-transfected T2 cells were used as APCs.Briefly, stable transfectants were achieved by electroporation and selectionwith G418 sulfate (0.8 mg/ml; Calbiochem, La Jolla, CA). Surface ex-pression of CD1a was confirmed by flow cytometry, and intracellular lo-calization studies were performed by confocal microscopy. The CD1a-restricted, sulfatide-specific human T cell clone K34B9.1 was derived andcultivated as previously described (8).

Flow cytometry

HeLa transfectants and monocyte-derived DCs were analyzed by flowcytometry. In the case of the HeLa transfectants, the cells were detached bybrief incubation with 0.5% trypsin-EDTA (Life Technologies/Invitrogen)and washed prior to staining. The cells were stained with PE-conjugatedanti-CD1a or isotype control mouse IgG1 mAb (BD Biosciences, San Jose,CA) on ice. The cells were analyzed using a FACScan or Canto flow cy-tometer (BD Biosciences).

Lipid Ags

Bovine sulfatide and ceramide (C24) were obtained fromMatreya (PleasantGap, PA) and lipoarabinomannan (LAM) from Colorado State University,Fort Collins, CO.

Confocal microscopy

Laser confocal microscopy was performed on the HeLa transfectants andmonocyte-derived DCs as previously described (3, 15). Briefly, HeLa trans-fectants were cultured on coverslips and DC plated on fibronectin-coated cov-erlips overnight. After washing, the cells were fixed with paraformaldehyde,permeabilized with 0.2% saponin, and stained with mAb against CD1a (10H3,mouse IgG1), CD1b (BCD1, mouse IgG1), LAMP-1 (H4A3, mouse IgG1),sulfatide (04, mouse IgM supernatants, generously provided by Dr. Joan Boggs,Hospital for Sick Children, Toronto, Ontario, Canada) (17), LAM (rabbit anti-LAM antisera was generously provided by Dr. Daniel Clemens, University ofCalifornia, Los Angeles, Los Angeles, CA), Rab5 (BD Biosciences), early en-dosome Ag-1 (EEA-1; BD Biosciences), and appropriate species and isotypecontrols for 60min at room temperature.ThepurifiedAbswere used at 10mg/mland the supernatants at 1:5 dilution. The cells were stained with the appropriatesecondaryAb,Alexa 546- orAlexa 488-conjugated anti-mouse IgM,Alexa488-conjugated F(ab9)2 fragment of goat anti-mouse IgG (Molecular Probes/In-vitrogen), and FITC-conjugated F(ab9)2 fragment of donkey anti-rabbit IgGAbs

FIGURE 1. CD1a localization is redirected to LYs in the CD1aab tail chimera. HeLa CD1a WT, CD1b WT, and CD1aab tail chimera transfectants were

analyzed by confocal microscopy. A, The CD1 transfectants were stained with Abs directed at their extracellular domains. CD1a WT transfectants show cell

surface and perinuclear staining. The CD1b WT and CD1aab chimera have similar intracellular staining patterns. B, Colocalization studies with anti–

LAMP-1 mAbs demonstrated the localization of the CD1aab chimera to LAMP-1+ compartments similar to WT CD1b. There is no colocalization between

CD1a and LAMP-1. The fluorescent label used for each protein is indicated by color.

1236 CD1a TRAFFICKING IS ESSENTIAL FOR LIPID ANTIGEN PRESENTATION

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

(Jackson ImmunoResearch, West Grove, PA) for 60 min at room temperature.Additional staining was performed with Alexa 488-conjugated anti–LAMP-1mAb (H4A3, BD Biosciences) and FITC-conjugated transferrin (Trf) (In-vitrogen) at 10 mg/ml. In the case of the Trf experiments only, the HeLa trans-fectants were serum-starved in DMEMmedia with 0.2% BSA for 30 min. Thecells werewashed three times prior to fixation and confocal analysis. For the Agloading studies, the cells were incubated for 30 min, unless clearly indicatedotherwise as in the time course analysis, at 37˚Cwith sulfatide at 10mg/ml, afterwhich the cells were washed three times with PBS at room temperature prior toconfocal staining. In the case of the Trf experiments, 100 mg/ml sulfatide wasused. The slides were analyzed with an inverted microscope (TE2000, Nikon,Melville, NY) with C1 confocal system and the Plan Apochromat 603 oil (NA1.4) objective lens using EZ-C1 software (Nikon) at room temperature. Onlyrepresentative images were acquired.

Pulse and chase Ag presentation assays

CD1a WTand CD1aab HeLa transfectants (13 106) were irradiated (5000rads, [137Cs] source) and pulsed with 2 mg (pulse experiments) of sonicatedsulfatide for indicated times. CD1aWT-, CD1aab-, and CD1abb-transfectedT2 cells were pulsed for 3 h with 10 mg (chase experiments) of sonicatedsulfatide. For the these experiments, purified synthetic C24:1 (nervonoyl-)sulfatide (18) was used and was kindly provided by L. Panza (Universitadel Piemonte Orientale, Novara, Italy). After pulse, the cells were washedand chased or not for indicated times. At time zero, the cells for all theconditions were collected, counted, and plated (2.5 3 104/well) in RPMI1640 medium containing 10% FCS with or without the addition of freshlysonicated Ag (at the same concentration as pulse). T cells (1 3 105/well)were added in triplicate or quadruplicate. Supernatants were harvestedafter 24 h, and released cytokines were measured by ELISA.

ELISAs

MaxiSorp ELISA plates (Nunc, Roskilde, Denmark) were coated overnight at4˚C with 1 mg/ml 8D4-8 (anti-human IL-4, BD Biosciences), 1 mg/ml MAb1(anti-humanTNF-a, BDBiosciences), or 3mg/mlHB-8700 (anti-human IFN-g,American Type Culture Collection) mAb, blocked with PBS 0.05% Tween-20and 10 mg/ml BSA, and incubated with the supernatants of the Ag presentationassays. For detection, 1 mg/ml MP4-25D2 (anti-human IL-4 biotin labeled, BDBiosciences), 0.5 mg/ml MAb11 (anti-human TNF-a biotin labeled, BD Bio-sciences), or 0.72mg/ml g69 (anti-human IFN-g biotin labeled) (19) mAbwereused. The plates were developed colorimetrically, and concentrations were de-termined by comparison with standards of human IL-4 (human IL-4 secretingX63, kind gift of U. Grawunder, 4-Antibody AG, Basel, Switzerland), re-combinant humanTNF-a (ImmunoKontact, Abingdon,U.K.), and recombinanthuman IFN-g (Bender MedSystems, Burlingame, CA).

ResultsRelocation of CD1a to LY by providing residues of CD1bcytoplasmic tail

It has been previously shown that CD1a localizes to the ERC,whereas CD1b localizes to the LE/LY. To investigate the importanceof tyrosine-based motifs in CD1 molecule trafficking, HeLa CD1a:CD1b tail chimeric constructs containing the tyrosine-based motif ofCD1b were generated and intracellular localization of CD1 mole-cules analyzed by confocal microscopy. The CD1aab tail chimerashowed a different localization from CD1a WT and an identicalstaining pattern comparedwithWTCD1b (Fig. 1A). This was furtherdemonstrated by colocalization of the CD1aab chimera to LE/LYwith LAMP-1, as it was also demonstrated for WT CD1b. This is inagreement with the known important role for the tyrosine-basedsorting motif of CD1b in binding AP-3 and directing trafficking tothe LY. In contrast, as expected, there was no colocalization betweenWT CD1a and LAMP-1 (Fig. 1B). Both WT and chimeric proteinswere well expressed on the cell surface, as demonstrated by flowcytometry. The CD1aab chimera transfectants consistently showedlower surface mean fluorescence intensity compared withWT CD1aas determined by flow cytometry (data not shown). It should benoted that there was significant anti-CD1a mAb intracellular stain-ing with anti-CD1a mAb of the CD1aab chimera, indicating con-formational stability of the CD1a extracellular domain in the lateendocytic system (Fig. 1A).

Sulfatide colocalizes to the early endocytic/recycling pathway

Despite extensive prior analyses of CD1a and CD1b trafficking, onlyrarely has lipid Ag trafficking been directly visualized, because tag-ging of lipid Ags often alters their trafficking, and few mAb againstlipids exist. No prior studies have examined the trafficking of CD1a-presented lipid Ags. We developed a protocol to determine the in-tracellular distribution of sulfatide using confocal microscopy andthe anti-sulfatide specific mAb (O4) in sulfatide-pulsed HeLa CD1transfectant cells. Sulfatide was sonicated in complete media at 37˚Cand then incubated with CD1a HeLa transfectants for the time pointsindicated in Fig. 2A. After 30 min of incubation, sulfatide was takenup by the cells as revealed by well-defined, punctuate staining withanti-sulfatide mAb (Fig. 2A), whereas minimal to no backgroundstaining was observed with the anti-sulfatide mAb in the unpulsedcells. Strikingly, HeLa CD1a transfectants incubated with sulfatidedemonstrated an almost exclusive localization of sulfatide to theearly endocytic and recycling pathways, identified using anti-Rab5

FIGURE 2. Sulfatide localizes to the early endocytic and recycling

compartments. A, The anti-sulfatide mAb (O4) specifically labels sulfa-

tide-pulsed HeLa CD1a transfectants. Numbers above panels indicate the

sulfatide pulsing times. IgM control staining was performed at 60 min. B,

Colocalization of sulfatide with the early endocytic markers Rab5 and

EEA-1 was performed in HeLa CD1a WT cells pulsed with sulfatide for

30 min. Individual Ab stainings and the merged images are presented. C,

Colocalization pattern of CD1a and Trf in FITC-Trf–pulsed HeLa CD1a

WT transfectants (top panels) and sulfatide and Trf in sulfatide- and FITC-

Trf–pulsed HeLa CD1a WT (bottom panels). Individual Ab stainings and

the merged images are displayed.

The Journal of Immunology 1237

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

and EEA-1–specific mAb and fluorescently-labeled Trf. There wasmarked colocalization between sulfatide and the early endocyticmarkers Rab5 and EEA-1 (Fig. 2B). Sulfatide was also present in theERC as demonstrated by its partial colocalization with Trf (Fig. 2C).CD1a and Trf also demonstrated partial overlap (Fig. 2C) consistentwith the previously published localization of CD1a in the ERC (4, 5).However, there was no colocalization with LAMP-1 (Fig. 1B),a marker of the late endocytic system.

Differential colocalization of CD1a and CD1b with sulfatide

Next, the colocalization pattern of CD1a and CD1b and the CD1aabchimera with sulfatide was examined. HeLa CD1a WT, CD1b WT,and CD1aab chimera transfectants were pulsed with sulfatide for30 min and stained with anti-CD1a, anti-CD1b, and anti-sulfatidemAb. Consistent with the above findings, there was markedcolocalization of CD1a with sulfatide in the HeLa CD1a WTtransfectant cells (Fig. 3A). On the other hand, although smallamounts of punctate colocalization of CD1b with sulfatide wereobserved in some cells, the majority of sulfatide did not colocalizewith CD1b in the HeLa CD1b WT transfectants (Fig. 3A). For the

CD1aab tail chimera, there was a marked reduction in the co-

localization of sulfatide with the CD1aab tail chimeric protein

(Fig. 3A). This suggests that the intracellular localization provided

by the CD1b tyrosine-based motif determines the degree of in-

teraction of the CD1aab tail chimera with sulfatide. The distri-

bution of sulfatide was very similar in cells expressing CD1a WT

and CD1aab tail chimera as demonstrated by the almost complete

colocalization of sulfatide and Rab5. Furthermore, there was no

redirection of sulfatide to the late endocytic system with the

CD1aab tail chimera (Fig. 3B).Time course studies of sulfatide distribution were also performed

to investigate whether sulfatide traffics to LEs at late time points.

The transfectants were incubated with sulfatide for various time

periods before fixation prior to confocal microscopy. These studies

show that there is no detectable colocalization of LAMP-1 with

sulfatide up to 20 h after sulfatide pulse (Fig. 3B, 3C). This lack of

colocalization was observed for both the CD1a WT and CD1aab

tail chimera transfectants. Minor colocalization was observed for

sulfatide and CD1b at late time points (data not shown). At all time

points, the vast majority of sulfatide staining remained in the early

FIGURE 3. Marked intracellular

colocalization between CD1a and

sulfatide. HeLa CD1a WT, CD1b

WT, and CD1aab chimera trans-

fectants were pulsed with sulfatide

and analyzed by confocal micros-

copy. A, Colocalization between

each CD1 molecule and sulfatide

was determined for each trans-

fectant. The fluorescent labels used

for each protein are indicated by

color. B, Colocalization between

sulfatide and early (Rab5) and late

endocytic (LAMP-1) markers was

determined in sulfatide-pulsed

CD1a WT and CD1aab tail chimera

transfectants. C, The colocalization

pattern between CD1a and LAMP-1

was analyzed in CD1a WT trans-

fectants pulsed with sulfatide for 4 h

(top panel) and 20 h (bottom panel).

D, Confocal staining for CD1a (flu-

orescently labeled in red) and sul-

fatide (labeled in green) was

performed on CD1a WT trans-

fectants incubated with sulfatide

(top panels) and C24 ceramide

control (bottom panels) at increasing

concentrations for 30 min.

1238 CD1a TRAFFICKING IS ESSENTIAL FOR LIPID ANTIGEN PRESENTATION

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

endocytic and recycling compartments. The absence of sulfatidedetection in LAMP-1+ compartments even at later time points islikely due to its preferential sorting to the early endocytic and re-cycling compartments and to the degradation of sulfatide by en-dogenous lysosomal arylsulfatases that are well described to bepresent in this compartment (20). Lastly, to rule out the possiblecontribution of sulfatide to changes in CD1a expression, in-tracellular distribution, or cell morphology, confocal microscopywas performed with increasing concentrations of sulfatide or cer-amide (C24) as an additional control. No changes were observed byconfocalmicroscopy in the distribution or localization ofCD1a or inthe transfectant morphology after culturewith either lipid (Fig. 3D).

CD1a but not CD1b colocalizes with sulfatide in humanmonocyte-derived DCs

These findings were confirmed in primary cells, human monocyte-derived DCs that express all CD1 isoforms. Human monocyte-derived DCs were pulsed with sulfatide as described above.Confocal microscopy of DCs demonstrated areas of strongcolocalization between sulfatide and CD1a molecules but notbetween sulfatide and CD1b (Fig. 4A). These findings are inagreement with the known localization of CD1a and CD1b mol-ecules in monocyte-derived DCs to the ERC or to LE/LY com-partments, respectively (21, 22).

Colocalization of CD1b with LAM in DC

Next, the colocalization of CD1a and CD1b with LAM, an Agpresented by CD1b, was analyzed. Monocyte-derived DCs wereincubated with LAM for 30 min and analyzed by confocal mi-croscopy. Strikingly, in contrast to the specific colocalization ofCD1a with sulfatide in DC, LAM strictly colocalized with CD1bbut not with CD1a (Fig. 4B). The localization of LAM with CD1bis consistent with the known interaction of LAM with mannosereceptors that likely mediates the internalization and direct tar-geting of LAM and similar Ags to the late endocytic system (23).These findings suggest that CD1 molecules select the lipid Ag tobe presented according to their capacity to bind and form stablecomplexes and also according to the endosomal localization ofeach lipid Ag and its intersection with the relevant CD1 molecule.

CD1a intracellular trafficking alters sulfatide Ag presentationto specific T cells

In order to determine the functional relevance of the intracellulartrafficking of CD1a in Ag presentation, we next examined theability of CD1a WT and different CD1a:CD1b tail chimeras toactivate CD1a-restricted sulfatide-specific T cells. First, to assessthe appearance of sulfatide:CD1a stimulatory complexes, weperformed pulse experiments using HeLa CD1a WT and CD1aabchimera transfectants with matched cell surface expression ofCD1a as determined by flow cytometry. The transfectants werepulsed with sulfatide for 0, 7, 20, 25, 45, and 60 min and washedbefore the addition of T cells. The ability of sulfatide-pulsedtransfectants to stimulate the sulfatide-specific T cell cloneK34B9.1 increased with increasing pulse period independently ofthe measured cytokine (IFN-g, TNF-a, or IL-4). CD1a WTtransfectants outperformed the CD1aab chimera at pulse periodslonger than 30 min (Fig. 5). CD1aab chimera transfectants evenshowed a slight decrease in stimulatory capacity at pulse periodslonger than 30 min. Then, to assess the disappearance of sulfatide:CD1a stimulatory complexes, we performed chase experimentsusing T2 CD1a WT and CD1a:CD1b chimeras instead of HeLatransfectants to avoid potential effects of trypsinization requiredfor the release of the HeLa transfectants. The T2 transfectants werepulsed with sulfatide for 3 h, washed, and chased for 0, 24, 48, and72 h before addition of T cells and cytokine release measurement.At each time point, control groups were represented by trans-fectants incubated with sulfatide together with T cells to determinethe maximal Ag-presenting capacity. The ability of sulfatide-pulsed transfectants to stimulate the sulfatide-specific cloneK34B9.1 decayed with increasing chase period independently ofthe measured cytokine. WT CD1a transfectants showed halfmaximal stimulation at 40 h of chase, whereas CD1abb trans-fectants showed half maximal stimulation at 20 h. CD1aab trans-fectants showed intermediate behavior, with half maximalstimulation at 30 h (Fig. 6). After 72 h, CD1a WT transfectantswere still capable of stimulating sulfatide-specific T cells, whereasboth chimera transfectants were no longer stimulatory.

DiscussionSuccessful Ag presentation involves an orchestrated series of stepsthat are critically dependent on the intersection of Ags with therelevant Ag-presenting molecules. Lessons from the MHC class Iand II pathways have made it clear that access to the proper Agloading compartment is a key feature for those presentation systemsto succeed. The proteolytic environment present inLYsofDCsplaysa critical role in their ability to process proteinAgs for prolongedAgpresentation. APCs such as macrophages with high levels of lyso-somalproteases rapidlydegrade internalizedproteins, reducing theircapacity for sustained Ag presentation compared with DCs (24).Although the ability of MHC class I and II molecules to surveydifferent intracellular compartments and Ags is a hallmark of thesesystems, the role of CD1 trafficking in the presentation of exoge-nous and endogenous Ags has been examined in only a few cases.The unique intracellular distribution of each CD1 isoform has

been appreciated for some time. It has been demonstrated for theCD1b and CD1c isoforms that their intracellular trafficking isdirected by their respective tyrosine-based motifs using tail-deletedconstructs (3, 25). Tail-deleted CD1b molecules have been dem-onstrated to have significant impairment in their Ag-presentingabilities in vitro (3). Glucose monomycolate possessing a longlipid tail has been shown to require the acidic environment foundin the late endocytic system to be loaded efficiently (26, 27).

FIGURE 4. Reciprocal intracellular localization patterns of CD1a and

sulfatide and CD1b and LAM. The pattern of CD1a and CD1b colocali-

zation with sulfatide (A) and LAM (B) was analyzed by confocal mi-

croscopy in human monocyte-derived DCs pulsed with sulfatide. The

fluorescent labels used for each protein are indicated by color.

The Journal of Immunology 1239

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

CD1a is the only CD1 isoform that does not have a tyrosine-basedmotif and almost exclusively localizes to the early endocytic andrecycling pathways. In this study, we demonstrate that the locali-zation of CD1a to the early endosomal and recycling compartmentsis critical for the efficient and sustained presentation of specific Agsthat also predominantly localize to the same endosomal sites. CD1aand sulfatide are almost exclusively colocalized to early endocyticand recycling compartments. The redirection of CD1a to the LE bythe addition of the transmembrane and/or cytoplasmic CD1b ty-rosine-based motif to the extracellular domain of CD1a resulted inloss of colocalization with sulfatide and a correspondingly signif-icant reduction in the ability of these transfectants to stimulateCD1a-restricted sulfatide-specific T cells. Because the extracellulardomains of the chimeric molecules are identical, the most likelyexplanation for these findings is the difference in traffickingimparted by the chimeric cytoplasmic tails. The activity and lo-calization of endogenous sulfatases suggests that even if somesulfatide is delivered to LYs, it would be rapidly degraded. There aretwo classes of sulfatases, catabolic and synthetic. The catabolicsulfatases are localized in LYs and exert their activity at acidic pHona wide range of molecules, including sulfolipids and glyco-soaminoglycans. Accumulation of these molecules occurs in thesetting of lysosomal sulfatase deficiency and leads to humandisease,including several mucopolysaccharidoses and metachromatic leu-kodystrophy (20). The prolonged ability of CD1a molecules topresent sulfolipids such as sulfatide may be critically dependent ontheir localization to early endocytic compartments where the sul-fatide Ag is not degraded by sulfatases. The absence of a tyrosine-based motif allows CD1a to traffic and recycle through the earlyendocytic system, where it can most efficiently present lipid Agsthat preferentially localize there or that would be destroyed either

enzymatically or as a result of acidic pH in the late endocyticsystem. This is supported by the enhanced ability of WT CD1a topresent sulfatide compared with the CD1aab and CD1abb tailchimeras and the sustained ability of sulfatide-pulsed DCs to ac-tivate CD1a- compared with CD1b- and CD1c-restricted sulfatide-specific T cells. Additional interactions or pathways contributing toCD1a Ag presentation may also be contributory. For example, thecolocalization of CD1a with the unique LC-specific C-type lectinrequired for Birbeck granule formation, langerin, in this endosomalsubcompartment has been shown to augment CD1a-mediated Agpresentation (28). We cannot rule out the possibility that theCD1aab–sulfatide complexes dissociate in the LY. We should alsomention the possibility that the anti-sulfatide mAb may not be ableto detect sulfatide in the LY secondary to changes to the antigenicepitope recognized. However, similar sulfatide colocalization pat-terns have been observed with other anti-sulfatide mAbs, makingthis less likely.The differential localization of CD1a to the early endocytic

compartment may be important in vivo. The efficient and sustainedpresentation of Ags such as sulfatide by CD1a+ DC during theirmigration to regional lymph nodes is likely critical for maximiz-ing T cell activation and proliferation. This may be even of greaterimportance in vivo, where Ag concentrations are likely lower thanour studies where higher concentrations of sulfatides were addedto allow for Ab detection and localization by microscopy.The specific intracellular distribution of CD1molecules and their

Ags may also explain the evolutionary diversity of the CD1 iso-forms (29). It is interesting to speculate that the distribution of thedifferent CD1 isoforms throughout the endocytic system may havedeveloped to allow for the presentation of differentially distributedAgs or Ags that are destroyed either enzymatically or by the acidic

FIGURE 6. CD1a:CD1b chimeras have shorter-lived ability to stimulate sulfatide-specific T cells than WT CD1a molecules. WT CD1a (white diamonds)

T2 transfectants showed prolonged capacity to stimulate the CD1a-restricted sulfatide-specific human T cell clone K34B9.1 as compared with CD1aab

(gray diamonds) or CD1abb (black diamonds). All transfectants were pulsed with nervonoyl sulfatide and chased for indicated times before addition of

T cells. In control wells, fresh Ag was added to chased T2 cells together with T cells. Supernatants were harvested after 24 h, and the release of human IL-4

(A), human IFN-g (B), and human TNF-a (C) was measured by ELISA and expressed as percent of control at time zero 6 SD (of triplicates or quad-

ruplicates) normalized at each time point for the maximal Ag-presenting capacity of the APC calculated as (chase/chase with fresh Ag)/(no chase/no chase

fresh Ag) 3 100.

FIGURE 5. WT CD1a transfectants present sulfatide more efficiently than CD1aab chimeric transfectants. WT CD1a (white diamonds) HeLa trans-

fectants showed increased capacity to stimulate the CD1a-restricted sulfatide-specific T cell clone K34B9.1 as compared with CD1aab chimeric trans-

fectants (gray diamonds) for pulse periods longer than 30 min. Both transfectants were pulsed with nervonoyl sulfatide for indicated times, then washed and

coincubated with T cells. Supernatants were harvested after 24 h, and the release of human IL-4 (A), human IFN-g (B), and TNF-a (C) was measured by

ELISA. Error bars represent SD of triplicate measurements.

1240 CD1a TRAFFICKING IS ESSENTIAL FOR LIPID ANTIGEN PRESENTATION

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from

microenvironment of the late endocytic system. The CD1a, b, c,and d isoforms may have developed to maximize the efficient andsustained Ag presentation of endogenous and exogenous lipidswith differences in both intracellular distribution and half-life.These studies also highlight the importance of the differential

intracellular trafficking of lipid Ags in immunity. This was dem-onstrated by the specific intracellular localization of sulfatide toearly endocytic and recycling pathways and the receptor-mediatedinternalization of LAM to LEs. The biophysical properties of lipidAgs, both the head group and acyl chains, have been demonstrated tobe important inAgpresentation. Interactionbetween theglycanheadgroups of CD1a- and CD1b-restricted Ags such asMycobacteriumleprae lipid Ags and LAM and the C-type lectins langerin andmannose receptor, respectively, have been demonstrated to playimportant roles in Ag uptake and presentation (28, 30). The lengthand saturation of the acyl chain of the glycosphingolipid glucosemonomycolate and GM1 length have been demonstrated to beimportant in CD1b-mediated Ag presentation (27, 31). Lipidlength, saturation, and head groups have also been demonstrated tocontribute to the endocytic sorting of lipids and the preferentialretention of certain lipids in organelles (32, 33).CD1-mediated sulfatide Ag presentation may be of increased

importance in APCs that express CD1a and in the setting of in-creased sulfatide availability. Upregulation of the production ofsulfatide and other glycosphingolipids in activated DCs has alreadybeen shown to play a role in CD1-mediated immunity (12). The highlevels of expression of CD1a, long a marker of LCs, on myeloidDCs point to its importance in self- and foreign lipid Ag pre-sentation. Further, because so many pathways of internalization ofexogenous Ags traverse the early endocytic system, the role ofCD1a as the principal lipid Ag-presenting molecule in this locationhighlights its distinctive role as the main sentry at this entry point.

AcknowledgmentsWe thank Dr. Duarte Barral for his expert advice.

DisclosuresThe authors have no financial conflicts of interest.

References1. De Libero, G., and L. Mori. 2005. Recognition of lipid antigens by T cells. Nat.

Rev. Immunol. 5: 485–496.2. Brigl, M., and M. B. Brenner. 2004. CD1: antigen presentation and T cell

function. Annu. Rev. Immunol. 22: 817–890.3. Sugita, M., X. Cao, G. F. Watts, R. A. Rogers, J. S. Bonifacino, and

M. B. Brenner. 2002. Failure of trafficking and antigen presentation by CD1 inAP-3-deficient cells. Immunity 16: 697–706.

4. Sugita, M., E. P. Grant, E. van Donselaar, V. W. Hsu, R. A. Rogers, P. J. Peters,and M. B. Brenner. 1999. Separate pathways for antigen presentation by CD1molecules. Immunity 11: 743–752.

5. Barral, D. C., M. Cavallari, P. J. McCormick, S. Garg, A. I. Magee,J. S. Bonifacino, G. De Libero, and M. B. Brenner. 2008. CD1a and MHC class Ifollow a similar endocytic recycling pathway. Traffic 9: 1446–1457.

6. Mc Dermott, R., U. Ziylan, D. Spehner, H. Bausinger, D. Lipsker, M. Mommaas,J. P. Cazenave, G. Raposo, B. Goud, H. de la Salle, et al. 2002. Birbeck granulesare subdomains of endosomal recycling compartment in human epidermalLangerhans cells, which form where Langerin accumulates. Mol. Biol. Cell 13:317–335.

7. Moody, D. B., D. C. Young, T. Y. Cheng, J. P. Rosat, C. Roura-Mir,P. B. O’Connor, D. M. Zajonc, A. Walz, M. J. Miller, S. B. Levery, et al. 2004. Tcell activation by lipopeptide antigens. Science 303: 527–531.

8. Shamshiev, A., H. J. Gober, A. Donda, Z. Mazorra, L. Mori, and G. De Libero.2002. Presentation of the same glycolipid by different CD1 molecules. J. Exp.Med. 195: 1013–1021.

9. Vos, J. P., M. Lopes-Cardozo, and B. M. Gadella. 1994. Metabolic and functionalaspects of sulfogalactolipids. Biochim. Biophys. Acta 1211: 125–149.

10. Sandhoff, R., S. T. Hepbildikler, R. Jennemann, R. Geyer, V. Gieselmann,R. L. Proia, H. Wiegandt, and H. J. Grone. 2002. Kidney sulfatides in mousemodels of inherited glycosphingolipid disorders: determination by nano-electro-spray ionization tandem mass spectrometry. J. Biol. Chem. 277: 20386–20398.

11. Fredman, P., J. E. Mansson, B. M. Rynmark, K. Josefsen, A. Ekblond,L. Halldner, T. Osterbye, T. Horn, and K. Buschard. 2000. The glycosphingolipidsulfatide in the islets of Langerhans in rat pancreas is processed through re-cycling: possible involvement in insulin trafficking. Glycobiology 10: 39–50.

12. De Libero, G., A. P. Moran, H. J. Gober, E. Rossy, A. Shamshiev,O. Chelnokova, Z. Mazorra, S. Vendetti, A. Sacchi, M. M. Prendergast, et al.2005. Bacterial infections promote T cell recognition of self-glycolipids. Im-munity 22: 763–772.

13. Zajonc, D. M., I. Maricic, D. Wu, R. Halder, K. Roy, C. H. Wong, V. Kumar, andI. A. Wilson. 2005. Structural basis for CD1d presentation of a sulfatide derivedfrom myelin and its implications for autoimmunity. J. Exp. Med. 202: 1517–1526.

14. Manolova, V., M. Kistowska, S. Paoletti, G. M. Baltariu, H. Bausinger,D. Hanau, L. Mori, and G. De Libero. 2006. Functional CD1a is stabilized byexogenous lipids. Eur. J. Immunol. 36: 1083–1092.

15. Hava, D. L., N. van der Wel, N. Cohen, C. C. Dascher, D. Houben, L. Leon,S. Agarwal, M. Sugita, M. van Zon, S. C. Kent, et al. 2008. Evasion of peptide,but not lipid antigen presentation, through pathogen-induced dendritic cellmaturation. Proc. Natl. Acad. Sci. USA 105: 11281–11286.

16. Manolova, V., Y. Hirabayashi, L. Mori, and G. De Libero. 2003. CD1a and CD1bsurface expression is independent from de novo synthesized glycosphingolipids.Eur. J. Immunol. 33: 29–37.

17. Sommer, I., and M. Schachner. 1981. Monoclonal antibodies (O1 to O4) to ol-igodendrocyte cell surfaces: an immunocytological study in the central nervoussystem. Dev. Biol. 83: 311–327.

18. Compostella, F., L. Franchini, G. De Libero, G. Palmisano, F. Ronchetti, andL. Panza. 2002. CD1a-binding glycosphingolipids stimulating human autoreativeT-cells: synthesis of a family of sulfatides differing in the acyl chain moiety.Tetrahedron 58: 8703–8708.

19. Gallati, H., I. Pracht, J. Schmidt, P. Haring, and G. Garotta. 1987. A simple,rapid and large capacity ELISA for biologically active native and recombinanthuman IFN g. J. Biol. Regul. Homeost. Agents 1: 109–118.

20. Diez-Roux, G., and A. Ballabio. 2005. Sulfatases and human disease. Annu. Rev.Genomics Hum. Genet. 6: 355–379.

21. van der Wel, N. N., M. Sugita, D. M. Fluitsma, X. Cao, G. Schreibelt,M. B. Brenner, and P. J. Peters. 2003. CD1 and major histocompatibility com-plex II molecules follow a different course during dendritic cell maturation. Mol.Biol. Cell 14: 3378–3388.

22. Cao, X., M. Sugita, N. Van Der Wel, J. Lai, R. A. Rogers, P. J. Peters, andM. B. Brenner. 2002. CD1 molecules efficiently present antigen in immaturedendritic cells and traffic independently of MHC class II during dendritic cellmaturation. J. Immunol. 169: 4770–4777.

23. Schlesinger, L. S., T. M. Kaufman, S. Iyer, S. R. Hull, and L. K. Marchiando.1996. Differences in mannose receptor-mediated uptake of lipoarabinomannanfrom virulent and attenuated strains of Mycobacterium tuberculosis by humanmacrophages. J. Immunol. 157: 4568–4575.

24. Delamarre, L., M. Pack, H. Chang, I. Mellman, and E. S. Trombetta. 2005.Differential lysosomal proteolysis in antigen-presenting cells determines antigenfate. Science 307: 1630–1634.

25. Briken, V., R. M. Jackman, G. F. Watts, R. A. Rogers, and S. A. Porcelli. 2000.Human CD1b and CD1c isoforms survey different intracellular compartmentsfor the presentation of microbial lipid antigens. J. Exp. Med. 192: 281–288.

26. Moody, D. B., V. Briken, T. Y. Cheng, C. Roura-Mir, M. R. Guy, D. H. Geho,M. L. Tykocinski, G. S. Besra, and S. A. Porcelli. 2002. Lipid length controlsantigen entry into endosomal and nonendosomal pathways for CD1b pre-sentation. Nat. Immunol. 3: 435–442.

27. Cheng, T. Y., M. Relloso, I. Van Rhijn, D. C. Young, G. S. Besra, V. Briken,D. M. Zajonc, I. A. Wilson, S. Porcelli, and D. B. Moody. 2006. Role of lipidtrimming and CD1 groove size in cellular antigen presentation. EMBO J. 25:2989–2999.

28. Hunger, R. E., P. A. Sieling, M. T. Ochoa, M. Sugaya, A. E. Burdick, T. H. Rea,P. J. Brennan, J. T. Belisle, A. Blauvelt, S. A. Porcelli, and R. L. Modlin. 2004.Langerhans cells utilize CD1a and langerin to efficiently present nonpeptideantigens to T cells. J. Clin. Invest. 113: 701–708.

29. Dascher, C. C. 2007. Evolutionary biology of CD1. Curr. Top. Microbiol. Im-munol. 314: 3–26.

30. Prigozy, T. I., P. A. Sieling, D. Clemens, P. L. Stewart, S. M. Behar,S. A. Porcelli, M. B. Brenner, R. L. Modlin, and M. Kronenberg. 1997. Themannose receptor delivers lipoglycan antigens to endosomes for presentation toT cells by CD1b molecules. Immunity 6: 187–197.

31. Shamshiev, A., A. Donda, T. I. Prigozy, L. Mori, V. Chigorno, C. A. Benedict,L. Kappos, S. Sonnino, M. Kronenberg, and G. De Libero. 2000. The alphabetaT cell response to self-glycolipids shows a novel mechanism of CD1b loadingand a requirement for complex oligosaccharides. Immunity 13: 255–264.

32. Holthuis, J. C., and T. P. Levine. 2005. Lipid traffic: floppy drives and a super-highway. Nat. Rev. Mol. Cell Biol. 6: 209–220.

33. Mukherjee, S., and F. R. Maxfield. 2000. Role of membrane organization andmembrane domains in endocytic lipid trafficking. Traffic 1: 203–211.

The Journal of Immunology 1241

on April 18, 2012

ww

w.jim

munol.org

Dow

nloaded from