Oxidized Lipids Block Antigen Cross-Presentation … Journal of Immunology Oxidized Lipids Block Antigen Cross-Presentation by Dendritic Cells in Cancer Wei Cao,*,1 Rupal Ramakrishnan,*,1

of June 17, 2018.This information is current as

CancerCross-Presentation by Dendritic Cells in Oxidized Lipids Block Antigen

Dmitry I. GabrilovichKlein-Seetharaman, Esteban Celis, Valerian E. Kagan and Mohammadyani, Joseph J. Johnson, Lan Min Zhang, JudithVeglia, Thomas Condamine, Andrew Amoscato, Dariush Wei Cao, Rupal Ramakrishnan, Vladimir A. Tuyrin, Filippo

http://www.jimmunol.org/content/192/6/2920doi: 10.4049/jimmunol.1302801February 2014;

2014; 192:2920-2931; Prepublished online 19J Immunol 

MaterialSupplementary

1.DCSupplementalhttp://www.jimmunol.org/content/suppl/2014/02/19/jimmunol.130280

Referenceshttp://www.jimmunol.org/content/192/6/2920.full#ref-list-1

, 14 of which you can access for free at: cites 51 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Errata

/content/192/10/4935.full.pdfor:

next pageAn erratum has been published regarding this article. Please see

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2014 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on June 17, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

The Journal of Immunology

Oxidized Lipids Block Antigen Cross-Presentation byDendritic Cells in Cancer

Wei Cao,*,1 Rupal Ramakrishnan,*,1 Vladimir A. Tuyrin,† Filippo Veglia,‡

Thomas Condamine,‡ Andrew Amoscato,* Dariush Mohammadyani,x Joseph J. Johnson,{

Lan Min Zhang,‖ Judith Klein-Seetharaman,# Esteban Celis,* Valerian E. Kagan,† and

Dmitry I. Gabrilovich*,‡

Cross-presentation is one of the main features of dendritic cells (DCs), which is critically important for the development of spon-

taneous and therapy-inducible antitumor immune responses. Patients, at early stages of cancer, have normal presence of DCs.

However, the difficulties in the development of antitumor responses in patients with low tumor burden raised the question

of the mechanisms of DC dysfunction. In this study, we found that, in differentiated DCs, tumor-derived factors blocked the cross-

presentation of exogenous Ags without inhibiting the Ag presentation of endogenous protein or peptides. This effect was caused by

intracellular accumulation of different types of oxidized neutral lipids: triglycerides, cholesterol esters, and fatty acids. In contrast,

the accumulation of nonoxidized lipids did not affect cross-presentation. Oxidized lipids blocked cross-presentation by reducing the

expression of peptide–MHC class I complexes on the cell surface. Thus, this study suggests the novel role of oxidized lipids in the

regulation of cross-presentation. The Journal of Immunology, 2014, 192: 2920–2931.

Antitumor immune responses, either spontaneous or in-duced by immune therapy, depend on the adequatefunction of host dendritic cells (DCs) (1–3). Defects in

DC function in tumor-bearing (TB) patients or mice, with ad-vanced disease, are well documented. They manifest in the ex-pansion of immature DCs, unable to properly present Ag, and thegeneration of cells with immune-suppressive activity, includingregulatory DCs and myeloid-derived suppressor cells (4). Togetherwith other immune-suppressive factors, those changes contributeto the inability of CTLs to mount antitumor immune responses (5–8).

Cross-presentation of Ags is a unique feature of DCs, which iscritically important for antitumor immunity. Cross-presentation isthe process where exogenous Ags are ingested and processed togenerate peptide T cell epitopes that are presented by MHC classI (MHC-I) molecules (9, 10). Currently, two main pathways ofcross-presentation have been described: cytosolic and vacuolar.After uptake, exogenous Ags are internalized into phagosomes orendosomes (11, 12). The cytosolic pathway involves the transferof exogenous Ag from the endosome/phagosome into the cytosolfor proteasomal degradation. Similar to direct presentation, thispathway is dependent on the TAP. In contrast, the vacuolar path-way is TAP independent and suggests that exogenous proteins aredegraded into peptides by lysosomal proteases within the phag-osome (or endosome). These peptides are then loaded onto MHC-Imolecules that recycle through the endocytic compartments bypeptide exchange. The use of each pathway may depend on thetype of Ag and the mechanism of its uptake.The main paradigm of tumor immunology stipulates that the

efficient CTL priming requires an uptake of tumor Ags by DCs,their migration to draining lymph nodes, and a cross-presentationof the Ags to CD8+ T cells in the context of MHC-I (13). DCs fromTB mice are able to cross-present tumor Ag to CTLs (14–17). DCinfiltration of solid tumors is well documented in TB patients andmice (18–21). Tumor growth is associated with tumor cell apo-ptosis and necrosis, and DCs have access to a large amountof tumor Ags via numerous mechanisms such as phagocytosis/endocytosis of cell-associated or soluble Ags bound to heat shockproteins, as well as via gap junction transfer, through the captureof exosomes, or via “cross-dressing” (acquisition of peptide–MHC-I complexes from contact with necrotic cells) (22, 23). Thetumor milieu contains soluble mediators such as type-I IFN andendogenous “danger signals” (DNA, HMGB1, S100), which areable to activate DCs. Taken together, all of these factors induceDC differentiation and activation. However, this does not result inthe development of potent antitumor immune responses. More-over, the induction of strong immune responses to cancer vaccines

*Department of Immunology, H. Lee Moffitt Cancer Center and Research Institute,Tampa, FL 33612; †Department of Environmental and Occupational Health, Univer-sity of Pittsburgh, PA 15219; ‡The Wistar Institute, Philadelphia, PA 19104; xDepart-ment of Bioengineering, University of Pittsburgh, PA 15219; {Microscopy Core, H.Lee Moffitt Cancer Center and Research Institute, Tampa, FL 33612; ‖MolecularCore, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL 33612; and#Division of Metabolic and Vascular Health, Medical School, University of Warwick,Coventry CV4 7AL, United Kingdom

1W.C. and R.R. contributed equally to this work.

Received for publication October 16, 2013. Accepted for publication January 14,2014.

This work was supported by National Institutes of Health Grant CA165065 (to D.I.G.and V.E.K.) and Grant R01CA136828 (to E.C.).

Address correspondence and reprint requests to Dr. Dmitry Gabrilovich, The WistarInstitute, Room 118, 3601 Spruce Street, Philadelphia, PA 19104. E-mail address:[email protected]

The online version of this article contains supplemental material.

Abbreviations used in this article: AA, arachidonic acid; AMVN, azobis-(2,4-dime-thylvaleronitrile); BM, bone marrow; CE, cholesterol ester; CM, control medium;DC, dendritic cell; ER, endoplasmic reticulum; FA, fatty acid; FFA, free FA; HL,high lipid content; LA, linoleic acid; LB, lipid body; LC-ESI-MS, liquid chromatog-raphy–electrospray ionization–mass spectrometry; MHC-I, MHC class I; MS, massspectrometry; m/z, mass-to-charge ratio; NL, normal lipid level; oxCE, oxidizedCE; oxLA, oxidized LA; oxTAG, oxidized triglyceride; pMHC, peptide MHC;PUFA, polyunsaturated FA; TAG, triglyceride; TB, tumor-bearing; TCM, tumor cell–conditioned medium; TDF, tumor-derived factor; TES, tumor explant supernatant; Tg,transgenic.

Copyright� 2014 by TheAmericanAssociation of Immunologists, Inc. 0022-1767/14/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1302801

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

is a difficult task, even in patients with a relatively small tumorburden.In this study, we tried to address this question by studying the

effect of tumor-derived factors (TDFs) on partially differenti-ated DCs. We found that TDFs inhibited the cross-presentationof exogenous proteins and long peptides, requiring Ag processing inDCs, without affecting the presentation of endogenous proteins anddirectly loaded short/minimal MHC-I binding peptides.Recently, lipid droplets or lipid bodies (LBs) were implicated in

cross-presentation via their association with endoplasmic reticulum(ER)–resident 47-kDa immune-related GTPase, Igtp (Irgm3) (24).LBs are neutral lipid storage organelles present in all eukaryoticcells. It is now established that LBs perform functions beyondlipid homeostasis. In addition, LBs were implicated in the regu-lation of immune responses via PGs and leukotrienes and, possi-bly, in IFN responses (reviewed in Ref. 25). Under physiologicalconditions in most cells, LBs are rather small with a diameterranging from 0.1 to 0.2 mm (26). In the tumor microenvironment,DCs accumulate lipids and form large LBs (27), which we hy-pothesized could directly interfere with cross-presentation. Wepresent the results indicating that cross-presentation is directlyregulated by the oxidized lipids that accumulate in DCs.

Materials and MethodsHuman cells, mice, and tumor models

Donors’ buffy coat blood was purchased from the local blood bank. An-imal experiments were approved by the University of South Florida In-stitutional Animal Care and Use Committee. BALB/c or C57BL/6 micewere obtained from the National Cancer Institute. OVA transgenic (Tg)mice C57BL/6-Tg(CAG-OVA)916Jen/J (Cat: 005145), CD204 knockoutmice B6.Cg-Msr1tm1Csk/J (Cat: 006096), mice on high-fat diet (60 kcal% ”high-fat diet, Cat: 380050), and the same-age control mice (10 kcal% diet-induced obesity (DIO) controls, Cat: 380056) were purchased from TheJackson Laboratory. OT-I TCR-Tg mice (C57BL/6-Tg[TCRaTCRb]1100mjb) were obtained from The Jackson Laboratory. Pmel-1 TCR-Tgmice (B6.Cg Thy1a-Tg[TcraTcrb]8Rest/J) were kindly provided by Dr. N.Restifo.

Reagents and cell lines

Tumor cell lines including CT26 and MC38 colon carcinomas, EL4lymphoma, LLC (Lewis Lung Carcinoma), and B16F10 melanoma weremaintained in RPMI 1640 medium supplemented with 10% FBS (Sigma-Aldrich, St. Louis, MO) at 37˚C, 5% CO2. Tumors were injected s.c. at 53105 cells/mouse.

SIINFEKL peptide, control peptide RAHYNIVTF, (Pam)2-KMFV-E-SIINFEKL peptide (derived from OVA), and (Pam)2-KMFV-KVPRNQDWL(derived from gp100) were obtained from American Peptide Company(Vista, CA). OVA (lyophilized powder,$98% agarose gel electrophoresis, Cat:A5503) was purchased from Sigma. Recombinant mouse GM-CSF was ob-tained from Biosource, Invitrogen (Carlsbad, CA); recombinant human GM-CSF and IL-4 were obtained from CellGenix. Biotin-conjugated anti-mouseCD11c Ab (Cat: 553800) was purchased from BD Biosciences and used forDC purification. In addition, FITC (Cat: 553801) or allophycocyanin (Cat:550261) or PE-Cy7 (Cat: 558079)–conjugated anti-mouse CD11c Abs wereobtained from BD Biosciences. Both FITC-conjugated anti-mouse MHC-I(H2Kb) Ab (Cat: 11-5958-82) and allophycocyanin-conjugated anti-mouseSIINFEKL-H2Kb complex Ab (Clone: 25-D1.16, Cat: 17-5743-82) werepurchased from eBioscience. In addition, variant fluorescent-conjugated Absincluding allophycocyanin-conjugated anti-mouse MHC-II (IAb) Ab (Cat:17-5321-82), allophycocyanin- (Cat: 17-0801-82), or FITC (Cat: 11-0801-82)-conjugated anti-mouse CD80 Abs, FITC- (Cat: 11-0862-82) or PE (Cat:12-0862-82)-conjugated anti-mouse CD86 Abs, allophycocyanin-conjugatedanti-mouse CD40 Ab (Cat: 17-0401-82), and PE-conjugated anti-mouse B7H1Ab (Cat: 12-5982-82) were all purchased from eBioscience. All of thesefluorescent-conjugated Abs were used for flow cytometry. BODIPY lipiddye 493/503 (Cat: D3922) and HCS LipidTOX red dye (Cat: H34476) wereobtained from Invitrogen. These dyes were used for flow cytometry andconfocal microscopy. The Ab used for confocal microscopy, unconjugatedanti-mouse SIINFEKL-H2Kb complex Ab (Clone: 25-D1.16, Cat: 14-5743-82),was purchased from eBioscience. The Abs for detection of different cellular

compartments included EEA1 Ab (marker for early endosome, Cat: ab2900),GiantinAb (marker forGolgi complex,Cat: ab24586),LAMP2Ab (marker forlysosome, Cat: ab25339), and calnexin Ab (marker for ER, Cat: ab10286), andwere obtained from Abcam. ERGIC-53/p58 Ab (marker for ER-Golgi inter-mediate compartment,Cat: E1031) andRab5aAb (marker for early endosome,Cat: NBP1-20255) were purchased from Sigma and Novus Biologicals, re-spectively. BothAlexa Fluor 488–conjugated anti-rabbit (Cat: A-11034) or anti-rat (Cat: A-11006) and Alexa Fluor 594–conjugated anti-rabbit (Cat: A-11037)or anti-rat (Cat: A-11007) secondary Abs were obtained from Invitrogen.

Preparation of tumor explant supernatant

Tumor explant supernatants (TES) were prepared from excised non-ulcerated tumors ∼1.5 cm in diameter. Tumor tissues were bathed in 70%isopropanol for 30 s and then transferred to a Petri dish. Tumors wereminced into pieces ,3 mm in diameter and digested in 2 mg/ml colla-genase Type D/IV at 37˚C for 1 h. The digested tissue pieces were thenpressed through a 70-mm mesh screen to create a single-cell suspension.Cells were washed with PBS and resuspended in RPMI 1640 supplementedwith 20 mM N-2-hydroxyl piperazine-N9-2-ethanesulfonic acid, 2 mML-glutamine, 200 U/ml penicillin plus 50 mg/ml streptomycin, and 10%FBS. Cells were cultured overnight at 107 cells/ml, and the cell freesupernatants were collected and kept at 280˚C.

Analysis of cell phenotype by flow cytometry

Cell-surface labeling was performed on ice for 20 min and analyzed bya FACSCalibur or LSRII flow cytometer and CellQuest program. For lipidstaining, cells were resuspended in 500 ml Bodipy 493/503 at 0.5 mg/ml inPBS. Cells were stained for 15 min at room temperature, then washed oncewith PBS, and resuspended in PBS-DAPI for flow cytometry analysis. Atleast 10,000 cells were collected for subsequent analysis.

Measurement of activity of protease

DCs were lysed in nonionic lysis buffer (10 mM Tris, pH 7.8; 1% NonidetP-40, 0.15M NaCl, 1 mM EDTA-Na). The cell lysates were centrifuged at19,000 3 g for 10 min at 4˚C. Five micrograms lysate in 100 ml substratebuffer (20 mM HEPES, pH 8.2, 0.5 mM EDTA-Na, 1% DMSO, 5 mMATP) and 1.3ml substrate (stock: 10 mmol/L) in 100 ml substrate buffer/well were mixed together. After a 30-min incubation at 37˚, fluorescence(excitation, 380 nm; emission, 460 nm) was measured using a Spectra-fluoPlus 96-well plate reader. The following substrates were used for de-tection: Z-LLE-AMC (Cat: ZW9345-0005) for caspase; BOC-LRR-AMC(Cat: BW8515-0005) for trypsin; Z-GGL-AMC (Cat: ZW8505-0005) forchymotrypsin; and Ac-KQKLR-AMC (Cat: 61859) for cathepsin-S. Ac-KQKLR-AMC substrate was purchased from Anaspec, whereas all othersubstrates were purchased from Biomol.

Generation and isolation of DCs

Mouse DCs were generated from bone marrow (BM) progenitor by a 3-dculture with 10 ng/ml GM-CSF (28). On day 3, medium was replaced with theone containing 20% v/v TES, and the cells were cultured for an additional48 h. Cells were labeled with biotinylated CD11c Ab (BD Biosciences, SanJose, CA) followed by incubation with magnetic beads, coupled to streptavidinand positive selection on magnetic column according to the manufacturer’sprotocol (Miltenyi Biotec). In some experiments, DCs were generated with20 ng/ml GM-CSF and 10 ng/ml IL-4 for 5 d. At that time, CD11c+ DCs wereisolated using magnetic beads and incubated with 20% v/v TES for additional48 h. DCs were also directly isolated from mouse spleen, using the samemethod.

Human DCs were generated from mononuclear cells, isolated fromdonors’ buffy coat blood by Ficoll gradient centrifugation. Cells werecultured for 3 d with 40 ng/ml recombinant human GM-CSF and 20ng/ml IL-4. On day 3, media were replaced with one containing 20% oftumor conditioned media, derived from SK-MEL melanoma cell lines.Forty-eight hours later, the nonadherent and loosely adherent cells werecollected.

Functional assays

DCs were loaded for 24 h with 100 mg/ml OVA or 5-mg/ml-long peptides.Before adding to T cells, DCs were irradiated with 20 Gy. T cells wereisolated using mouse T cell enrichment columns (R&D Systems). T cellswere then plated at 0.2 3 105 T cells/well. DCs and T cells were mixed atdifferent ratios. In experiments with loading of short peptide, 0.1 mg/mlSIINFEKL was added into the media. Cells were incubated for 72 h.[3H]thymidine was then added at 1 mCi per 200 ml cells/well for an additional

The Journal of Immunology 2921

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

18 h of incubation, followed by cell harvesting and a radioactivity count onliquid scintillation counter.

Confocal microscopy

Cells were fixed and permeabilized with Fixation & PermeabilizationBuffers (BD) for 20 min at 4˚C, washed with PBS, and then blocked withPBS containing 1% FBS for 1 h. Cells were incubated with different Absovernight at 4˚C and then stained with HCS LipidTOX red lipid stain orBODIPY, to detect LBs for 30 min at 4˚C. The cells were imaged with aLeica TCS SP5 laser-scanning confocal microscope through a 633/1.40NAPlan Apochromat oil immersion objective lens (Leica Microsystems, Wetzlar,Germany). Both 405-nm and 555-nm diode lasers lines were applied toexcite the samples. An acousto-optical beam splitter was used to collectpeak emission photons, sequentially, to minimize cross talk betweenfluorochromes.

Protein and peptide liquid chromatography–multiple reactionmonitoring mass spectrometry

Cells (n = 105) were lysed in aqueous 8M urea/100 mM ammonium bi-carbonate buffer on ice. After lysis, the cell supernatants were clarified bycentrifugation and decanted. Lysates were loaded for SDS-PAGE separa-tion (Criterion XT 4–12%) and visualized with colloidal Coomassie bril-liant blue G-250 (Bio-Rad, Hercules, CA). Gel regions, determined bycomparison with OVA standards, were excised. Proteins were reduced,alkylated, and digested in gel overnight digestion at 37˚C with sequencinggrade trypsin (Promega, Madison, WI). The resulting tryptic peptides wereconcentrated by vacuum centrifugation and resuspended in 2% acetoni-trile, with 0.1% formic acid for mass analysis. A digest of OVA standardprotein (Sigma, St. Louis, MO) was used for assay development to selectpeptides between 7 and 25 aa in length, for both specificity and sensitivityof detection. Liquid chromatography-multiple reaction monitoring massspectrometry (MS) was performed, as previously described (29).

Liquid chromatography and mass-spectral analysis of lipids

Lipids were extracted by Folch procedure (30) with slight modifications,under nitrogen atmosphere, at all steps. Liquid chromatography–electro-spray ionization–mass spectrometry (LC-ESI-MS) analysis was performedon a Dionex HPLC system (using the Chromeleon software), consisting ofa Dionex UltiMate 3000 mobile phase pump, equipped with an UltiMate3000 degassing unit and UltiMate 3000 autosampler (sampler chambertemperature was set at 4˚C). The Dionex HPLC system was coupled to anLXQ ion trap mass spectrometer or to a hybrid quadrupole-orbitrap massspectrometer, Q-Exactive (ThermoFisher, San Jose, CA) with the Xcaliburoperating system. The instrument was operated in both the negative andthe positive ion modes (at a voltage differential of 23.5 to 5.0 kV, sourcetemperature was maintained at 150˚C). For phospholipid MS analysis,spectra were acquired in negative ion mode using a full-range zoom (200–1600 mass-to-charge ratio [m/z]) or ultra-zoom (SIM) scans. Tandem MSanalysis of individual phospholipid species was used to determine the fattyacid (FA) composition. Spectra of free fatty acids (FFAs) were also ob-tained in negative ion mode. For triglyceride (TAG) MS analysis, spectrawere acquired in positive ion mode using range zoom (600–1200 m/z).TAG cations were formed through molecular ammonium adduction(TAG+NH4)

+. MSn analysis was carried out with relative collision energy,ranged from 20–40%, with activation q value at 0.25 for collision-induceddissociation, and q value at 0.7 for pulsed-Q dissociation technique. Thespectra of cholesterol esters (CEs) were also acquired in positive ion mode ona hybrid quadrupole-orbitrap mass spectrometer (Q-Exactive; ThermoFisher,San Jose, CA).

Statistical analysis

Statistical analysis was performed using unpaired two-tailed Student t testwith significance determined at a p value , 0.05.

ResultsTDFs inhibit Ag cross-presentation by DCs

To evaluate the effect of TDF on the ability of differentiated DCs topresentAgs, we generatedDCs invitro fromBMprogenitors in culturewith GM-CSF for 3 d. After that time, TESs, from different tumors(EL-4 lymphoma, MC38 colon carcinoma, and CT-26 colon carci-noma), were added for an additional 48 h. Under these conditions, TESdid not affect the expression of MHC-I or MHC-II molecules, andcaused a slight increase in the expression of costimulatory molecules

on DCs (Supplemental Fig. 1A). Analysis of the expression ofmarkers specific for different myeloid cells (CD11c, CD11b, F4/80,Gr-1) showed that under these conditions, TES did not affect thephenotype of differentiated DCs (Supplemental Fig. 1B). Little dif-ferences were observed in the effect of LPS on DC activation(Supplemental Fig. 1C). After TES treatment, the DCs were loadedwith an H-2Kb binding peptide (SIINFEKL), and no defect in theirability to stimulate peptide-specific OT-1 Tg CD8+ T cells was ob-served (Fig. 1A). The expression of SIINFEKL/H-2Kb complexes(peptide MHC [pMHC]) was evaluated by flow cytometry, using the25-D1.16 Ab, which specifically recognizes OVA-derived peptideSIINFEKL bound to H2-Kb of MHC-I. The results showed that TEStreatment of DCs did not significantly affect the cell-surface ex-pression of pMHC, generated by the direct loading of MHC-I withexogenous peptide (Fig. 1B). To assess the effect of TES on Ag cross-presentation, we loaded DCs with OVA protein, during the last 24 h ofculture, and then used them for stimulating OT-1 T cells and forthe presence of pMHC on the surface of DCs. TES, from all testedtumors, significantly (p , 0.01) reduced the ability of DCs to stim-ulate peptide-specific OT-1 T cells (Fig. 1C) and the expression ofpMHC on DCs (Fig. 1D). To confirm these observations, we useda long OVA-peptide construct (Pam2-KMFVESIINFEKL) that can-not bind directly to H2Kb but is very effective in generating pMHCand stimulating CD8 T cell responses (E. Celis, unpublished obser-vations). Loading of DCs with Pam2-KMFVESIINFEKL generateda higher density of pMHC than loading of DCs with OVA protein,which allowed for better visualization of these complexes by mi-croscopy. Treatment of DCs with TES resulted in a substantial de-crease of pMHC on the DC surface (Fig. 1E, 1F). Similar results wereobtained using DCs differentiated from BM progenitors with GM-CSF and IL-4 for 5 d (Supplemental Fig. 2A, 2B).To confirm inhibition of cross-presentation in a different ex-

perimental system, we used another long peptide, (Pam)2-KMFV-KVPRNQDWL, which contains a CD8 T cell epitope from themelanoma-associated gp100 protein. Peptide-epitope–specific pmel-1 Tg CD8+ T cells, which recognize the minimal gp100 epitopeKVPRNQDWL, were used as responders (31). Similar to the re-sults obtained with OVA, TES inhibited the cross-presentation ofthe Ag derived from the gp100 long peptide (Fig. 1G).Next, we asked whether the earlier described defects were spe-

cifically associated with the exogenous cross-presentation pathway orwould also be observed with the presentation of endogenous Ags.To address this question, we used OVA-Tg mice with a constitutiveexpression of OVA. Treatment of DCs, generated from these mice,with TES did not affect their ability to stimulate OT-1 T cells(Fig. 1H) and did not decrease pMHC expression on these DCs(Fig. 1I). It was possible that in OVA-Tg DCs, pMHC, formedbefore exposure to TES, were still present on the surface of DCs,thus negating the possible effect of TES on de novo Ag process-ing. To address this concern, we stripped peptides from MHC-I onthe DC surface, before TES application, using a mild acid treat-ment (Fig. 1J). Under these conditions, TES still did not affect theformation of new surface pMHC on DCs (Fig. 1K). It was possiblethat DCs in OVA-Tg mice present greater levels of OVA becauseof a constitutive OVA expression, which could compensate for theeffect of TES. We have compared, side-by-side, during the sameexperiment, the expression of pMHC in DCs loaded with OVA-derived peptide (SIINFEKL) and isolated from OVA-Tg mice.OVA-Tg DCs had lower expression of pMHC than peptide-loadedDCs (Supplemental Fig. 2C, 2D). To confirm the inhibitory effect oftumor-derived factors on DC cross-presentation, we cultured DCswith supernatant from splenocytes prepared exactly the same wayas TES. Supernatant from splenocytes did not affect DC cross-presentation, whereas TES caused a profound inhibitory effect in

2922 OXIDIZED LIPIDS AND DCs IN CANCER

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

DCs loaded with long peptide (Supplemental Fig. 2C). Thus, takentogether, these data indicate that TES does not affect the presentationof Ags derived from endogenous proteins and that the effects seemto be circumscribed to the cross-presentation pathway.We proceeded to evaluate the ability of DCs to process Ags and

express pMHC in vivo. EL-4 tumors were established in wild-typeand OVA-Tg mice. Spleens were isolated, and the pMHC expres-sion was evaluated in CD11c+IAb+ total population of DCs and inCD8+ subset known to be primarily involved in cross-presentation.DCs isolated from TB mice and loaded with long OVA-derivedpeptide (SIINFEKL) showed significantly reduced presentation of

pMHC on the surface (Fig. 2A). In contrast, no decrease in pMHCexpression was found in DCs isolated from TB OVA-Tg mice(Fig. 2B). These data support the conclusion that DCs in TB hostsare not able to effectively cross-present exogenous Ags but continueto be effective in processing and presenting endogenous Ags.

Mechanisms of inhibition of cross-presentation in cancer

We had previously observed an accumulation of lipids in DCs, fromcancer patients and TB mice, and the association of this accu-mulation to defects in DC function (27). Therefore, we investigatedthe possible role of lipids in the defective Ag cross-presentation by

FIGURE 1. Effect of TDFs on cross-presentation in DCs. (A) Stimulation of OT-1 CD8+ T cell proliferation by CD11c+ DCs cultured with TES from EL-4

tumors for 48 h. Cells were cultured at indicated ratios in the presence of 0.1 mg/ml control (RAHYNIVTF) or specific (SIINFEKL) peptides. Proliferation was

measured in triplicate by [3H]thymidine uptake. Proliferation of T cells, in the presence of control peptide, was ,1000 cpm and is not shown. Three

experiments, with the same results, were performed. (B) Expression of pMHC in DCs loaded with SIINFEKL. In this panel, CM indicates DCs incubated in

complete medium and TES indicates DCs incubated with EL-4 TES. Three experiments, with the same results, were performed. (C) Stimulation of OT-1 T cells

by DCs, incubated for 2 d in CM, or TES and loaded for 24 h with 100 mg/ml OVA. Cell proliferation was measured in triplicate by [3H]thymidine uptake.

Typical results of four performed experiments are shown. (D) Cumulative results of pMHC expression on DCs surface after loading of cells with 100 mg/ml

OVA. TES were obtained from EL-4 tumors. pMHC expression was measured within gated CD11c+ DCs. (E) Typical example of TES effect on pMHC

expression in DCs loaded with 5 mg/ml long OVA peptide. Three experiments, with the same results, were performed. (F) pMHC staining of CD11c+ DCs,

cultured for 24 h with EL-4 TES, and loaded with long peptide for an additional 24 h. Cells were stained with a 25-d1.16 primary and goat Alexa 488–

conjugated anti-mouse secondary Ab. Scale bar, 100 mm. (G) Effect of TES on the ability of DCs to cross-present long gp100-derived peptide. Pmel-1 Tg

T cells were used as responders. Cell proliferation was measured in triplicate by [3H]thymidine uptake. Two experiments, with the same results, were per-

formed. (H) Effect of EL-4 TES (48 h incubation) on the presentation of endogenous Ag in OVA-Tg DCs or after loading of wild-type DCs with OVA. OT-1 T

cell proliferation was measured in triplicate. One DC/OT-1 splenocytes ratio (1:20) is shown. Three experiments, with the same results, were performed. (I)

Effect of TES on presentation of pMHC by OVA-Tg DCs after 48-h incubation. For each TES, at least three experiments, with the same results, were

performed. (J) Effect of mild acid treatment (0.263 M citric acid and 0.123 M disodium phosphate, pH 3.0, for 2 min followed by extensive wash in PBS) on

the removal of peptide from MHC-I on OVA-Tg DCs. Two experiments, with the same results, were performed. (K) Effect of TES on pMHC expression in

OVA-Tg DCs after mild acid treatment. For each TES, two experiments, with the same results, were performed. *p, 0.05, statistically significant differences

versus control in all experiments.

The Journal of Immunology 2923

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

DCs in cancer. As expected, a 2-d exposure of DCs to TES causedthe accumulation of lipids in DCs (Fig. 3A). DCs, treated with

TES and loaded with OVA, were gated based on the level of lipids:cells with normal lipid level (NL) and cells with high lipid content

FIGURE 2. Cross-presentation in

DCs isolated from TB mice. (A)

Expression of pMHC on the surface

of CD11c+ DCs isolated from spleens

of naive or EL-4 TB (tumor �1 cm

in diameter) mice and loaded with

5 mg/ml OVA long peptide. (Left)

Typical example of staining. (Right)

Cumulative results of four experi-

ments in indicated populations of

DCs. (B) Expression of pMHC on

the surface of DCs isolated from

spleens of naive or EL-4 TB (tumor

�1 cm in diameter) OVA-Tg mice.

(Left) Typical example of staining.

(Right) Cumulative results of four

experiments in indicated populations

of DCs. *p , 0.05, statistically

significant differences versus con-

trol in all experiments.

FIGURE 3. Effect of TES on Ag processing in DCs. (A) Typical example of lipid level in gated CD11c+ DCs treated with EL-4 TES for 48 h after staining

with BODIPY. Seven experiments with similar results were performed. (B) Cross-presentation of Ags by DCs with different levels of lipids. DCs treated with

TES and loaded with OVA were stained with BODIPY. (Left panel) Example of gates set for discriminating DCs with NLs (NL-DCs) from DCs with HLs

(HL-DCs). DCs were considered NL-DCs when their fluorescence overlapped the fluorescence of the control DCs. Control DCs are in red; DCs treated with

TES are in blue. Three experiments, with the same results, were performed. (C) Cross-presentation of long OVA peptide by DCs with different levels of lipids.

Left panels show the gate of NL-DCs and HL-DCs, and pMHC expression is shown in the right panel. (D) TES does not affect cross-presentation in CD204-

deficient DCs. DCs generated from Msr12/2 mice were treated with TES, loaded with OVA, and used for stimulation of OT-1 T cells as described in Fig. 1C.

Proliferation of OT-I T cells was measured in triplicate. Typical results of four performed experiments are shown. (E) Typical examples of pMHC (left panel)

and H2Kb (right panel) expression in CD204-deficient DCs. Gray shaded line represents DCs without OVA treatment; blue line represents DCs loaded with

OVA in CM; red, green, and orange lines represent DCs loaded with OVA and pretreated with TES from EL-4, MC38, and CT26 tumors, respectively.

2924 OXIDIZED LIPIDS AND DCs IN CANCER

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

(HL). Although no differences were observed in the overall H-2Kb

expression between NL and HL DCs, the NL DCs had a substan-tially higher expression level of pMHC as compared with the HLDCs (Fig. 3B). Similar results were obtained when DCs wereloaded with OVA-derived long peptide and exposed to TES fromdifferent tumors (Fig. 3C). There was some overlap in pMHCexpression between DCs with control and HLs because only 50–60% of TES-treated DCs had increased lipid level.To further assess the possible involvement of lipids in the de-

creased generation of pMHC via cross-presentation, we used DCsdeficient for the scavenger receptor CD204 (Msr1). This receptorwas previously found to be primarily responsible for lipid ac-cumulation in DCs in cancer (27). CD204-deficient DCs weregenerated from BM of Msr12/2 mice and cultured with TESand OVA, as described earlier. In contrast with wild-type DCs

(Fig. 1), TES did not affect the ability of CD204-deficient DCsto stimulate OT-1 T cell proliferation (Fig. 3D) or to expresspMHC (Fig. 3E), indicating that lipid accumulation may bedirectly involved in the defective cross-presentation in DCs incancer.

Accumulation of oxidized lipids in DCs

Treatment of DCs with TES caused an accumulation of enlarged LBsin DCs that were easily observed inside the cells (Fig. 4A, redfluorescence). The proportion of DCs with enlarged LBs has beencalculated by counting 100 cells. It increased from 9.2% in control to43.5% in DCs treated with TES. We tested the possibility that LBscolocalized with the cellular compartments involved in Ag process-ing and the formation of pMHC and, thus, might interfere with thisprocess. However, no colocalization of LBs with lysosomes (Fig. 4A),

FIGURE 4. Lipid accumulation in mouse DCs. (A–D) Colocalization of LB and cell compartments involved in cross-presentation after 48 h culture of

DCs with TES. Lipids were stained with HCS LipidTOX (red fluorescence); lysosomes with LAMP2 Ab (A); trans-Golgi complex with giantin Ab (B);

early endosomes with Rab5a Ab (C); ER with calnexin Ab (D). Alexa Fluor 488 (green fluorescence)–labeled secondary Ab was used in all cases; except for

(D), where Alexa Fluor 594 (red fluorescence) and BODIPY lipid dye (green fluorescence) were used. Scale bar, 100 mm. Four experiments with the same

results were performed. (E) The proportion of different classes of FFA (left panel) and individual unsaturated FA (right panel) in control DCs. (F) The

presence of different unsaturated FA in DCs. (G) Oxidized C18:2 LA and C20:4 AA in spleen DCs and sera from EL-4 TB mice (three mice per group). (H)

The presence of oxLA in DCs incubated with TES. (I) Accumulation of mono-oxygenated TAGs in mouse DCs cultured in the presence of CM or EL4,

MC38 TES. Typical LC-ESI-MS profiles of TAG with m/z 916 (upper left panel), its MS2 spectrum (lower left panel), and its suggested structure (middle

panel). (Right panel) Amount of oxTAG 54:6, 54:5, 54:3, and 56:5 at m/z 912, 914, 918, and 942. (J) Accumulation of truncated oxTAGs in mouse DCs,

cultured in the presence of CM or EL4, MC38 TES. Typical LC-ESI-MS profile of truncated TAG with m/z 764 (upper left panel), its MS2 spectrum (lower

left panel), and possible structure (middle panel). Fragmentation of the parent ion at m/z 764 [M+NH4]+ reveals product ions at m/z 603 and 465. The

product ion, at m/z 603, was formed by loss of the truncated acyl chain (that corresponded to 7-oxo-heptanoic acid). The product ion, at m/z 465, was

formed by loss of oleic acid. (Right panel) Amount of truncated TAGs. (K) Content of LA-CE (left panel) and oxLA-CE (C18:2-2O) (right panel) in DCs

treated with TES. *p , 0.05, statistically significant differences versus control in all experiments.

The Journal of Immunology 2925

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

trans-Golgi network (Fig. 4B), early endosomes (Fig. 4C), or ER(Fig. 4D)was found. These data suggest that enlarged LBs in DCs isunlikely to cause a direct physical disruption of Ag processing.In a previous study, we observed an accumulation of TAGs and

FFAs in DCs from TB mice treated with TES via CD204 receptor(27). However, it is known that TAGs are not directly taken up byDCs but are synthesized inside the cells from FAs, and scavengerreceptor A (CD204) binds primarily modified lipoproteins (32). Tobetter understand the nature of lipid accumulation in DCs thataffects Ag cross-presentation, we used liquid chromatography/MSto quantitatively characterize different types of lipids and theiroxidation products. In control DCs, most FAs were representedby monosaturated and polyunsaturated FAs (PUFA), with C18:2

linoleic acid (LA) being the most abundant PUFA (Fig. 4E). In thepresence of TES, DCs accumulate large amounts of PUFA, pri-marily LA (Fig. 4F). Because unsaturated FFAs are highly sus-ceptible to oxidation, we evaluated the oxidative status of lipids.DCs isolated from spleens of naive mice contained very littleoxidized LA (oxLA) whereas DCs from TB mice had a markedlyhigher concentration of oxLA (Fig. 4G). Similar results wereobtained with oxidized arachidonic acid (AA), although its levelin DCs was substantially lower than that of oxLA. No differencesbetween naive and TB mice were seen in the levels of oxLA or

oxidized AA in sera (Fig. 4G). Similarly, DCs cultured withoutTES contained a practically undetectable level of oxLA, whereasDCs cultured with TES had a dramatically higher amount of oxLA(Fig. 4H). A dramatic accumulation of mono-oxygenated TAGswas seen in DCs treated with TES (Fig. 4I). This was associatedwith a large increase in the presence of oxidatively truncatedTAGs (TAGs that were oxidized with subsequent shortening of theoxygenated FA residues; Fig. 4J). Culture of DCs with TES alsoresulted in the accumulation of oxidized CEs (oxCEs; Fig. 4K).No detectable amounts of oxidized phospholipids were observedin different classes of DC phospholipids from either TB animals orDCs cultured in the presence of TES (data not shown).Typicalfragmentation analysis of oxygenated LA species in oxygenatedTAGs (oxTAGs) in DCs is shown in Supplemental Fig. 3.The proportion of oxLA among total LA was substantially in-

creased in DCs from TB mice as compared with control DCs. Incontrol DCs, the content of oxidized free LA constituted∼0.2 mol%,whereas in DCs from TB mice it was 2.5 mol%. TAGs containingLA residues underwent massive oxidative modification in DCsfrom TB mice as compared with naive animals. For example, thecontent of oxTAGs at m/z 920 corresponding to oxLA molecularspecies C18:2/C18:2+3O/C16:0 in DCs from TB mice was 59.3 6 3.8versus 1.26 0.1 pmol/106 cells in control. Thus, disproportionally

FIGURE 5. Accumulation of oxidized lipids in human DCs. (A) Total lipids in donor’s DCs treated with SK-MEL TCM. Typical example of BODIPY

staining. (B) Major FFAs detected in DCs; (C) oxLA in DCs. MS2 spectrum of parent ions at m/z 295 (left panel), its possible structure (middle panel), the

amount of oxLA (9-HODE) in DCs (right panel); (D) CE levels in human DCs cultured with and without TCM; (E) oxCE in DCs. Possible structure of CE

18:2-OOH (upper left panel); MS1 and MS2 spectra of CE 18:2-OOH (lower left panel); amount of CE 18:2-OOH in DCs (right panel). (F) Accumulation

of individual molecular species of TAGs in DCs. (G) OxTAGs in DCs. Typical LC-ESI-MS profile (left panel); MS2 spectrum of TAG at m/z 932 (middle

panel); amount of oxTAGs in DCs (right panel). Fold increase over values in DCs cultured in CM are shown. (B–G) Two experiments were performed. *p,0.05, statistically significant differences versus control in all experiments.

2926 OXIDIZED LIPIDS AND DCs IN CANCER

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

higher accumulation of oxidation products versus increase of therespective neutral lipids was observed in TB mice versus naiveanimals.Human DCs, cultured with tumor cell–conditioned medium (TCM),

had higher lipid levels than control DCs (Fig. 5A). The predominanttype of FFA observed in DCs, generated with TCM, was poorlyoxidizable monounsaturated C18:1 oleic FA (Fig. 5B). However, theconcentration of oxLA in these cells was .2-fold higher than inDCs cultured in control medium (CM; Fig. 5C). Although the totalamount of CEs showed no difference between DCs cultured withand without TCM (Fig. 5D), DCs exposed to TCM had a dramati-cally higher presence of oxCE (Fig. 5E). DCs cultured with TCMhad a modestly higher amount of TAGs than DCs cultured withoutTCM (Fig. 5F). In contrast, TCM caused a substantial accumulationof oxTAG in DCs (Fig. 5G). The detailed structural analysis of oxi-dized lipids in human DCs is provided in Supplemental Fig. 3.Taken together, these data indicate that DCs, in the presence of

TDF, have a substantial accumulation of oxFA, which apparentlyincorporated into TAG and CE, generating an increased amount ofdifferent oxidized lipid species.

Effect of intracellular oxidized lipids on cross-presentation

In view of the earlier text, we next asked whether the accumula-tion of nonoxidized or oxidized lipids in DCs could have an effect

on Ag cross-presentation. Based on the fact that oxLAwas the mostabundant FA accumulated in DCs in TB mice, we assessed whetherthe defect in cross-presentation could be reproduced by loadingcontrol DCs with LA and its subsequent intracellular oxidation. Toinduce intracellular peroxidation of neutral lipids, we used a li-pophilic peroxyl radical initiator, azobis-(2,4-dimethylvaleroni-trile) (AMVN), which partitions into hydrophobic phases (32).AMVN decomposes in lipids at a constant rate to yield carbon-centered alkyl radicals, which react with molecular oxygen togenerate reactive peroxyl radicals leading to lipid peroxidation(33). DCs were loaded for 24 h with LA in serum-free medium,treated with AMVN for 2 h, washed, and then used in experi-ments. Loading of DCs with LA or in combination with AMVNresulted in an accumulation of large LBs, similar to the effect ofTES (Fig. 6A). At a concentration of 0.2–0.5 mM, AMVN did notaffect DC viability or expression of MHC-I, MHC-II, or costim-ulatory molecules (Supplemental Fig. 4A, 4B). In contrast,AMVN induced oxidation of LA in DCs (Fig. 6B). In addition tooxLA, we also observed accumulation of oxTAGs and oxCE inDCs loaded with LA and treated with AMVN (Fig. 6C, 6D). Incontrast with robust AMVN-driven accumulation of peroxidationproducts in major neutral lipid components of LBs (peroxidizedspecies of TAGs, FFA, and CEs were increased up to 244.3 6 7.5,74.8 6 12.4, and 1.2 6 0.2 pmol/106 cells, respectively), only

FIGURE 6. Effect of oxLA on cross-presentation by DCs. (A) Accumulation of large LBs in DCs cultured with LA and LA+AMVN. DCs were cultured

for 24 h in serum-free medium with 10 mg/ml LA and then treated with AMVN (0.2 mM) for 2 h, before washing and loading with OVA (100 mg/ml). Cells

were analyzed 18 h later. Staining with BODIPY (scale bar, 100 mm). (B) Amount of oxLA in the absence and in the presence of AMVN; (inset) LC-MS

profile and MS2 spectrum. (C) Contents of major molecular species of oxTAGs containing LA. (D) Amount of oxCE C18:2-2O in DCs after loading with LA

in the absence and presence of AMVN. (E) Stimulation of OT-1 T cells by DCs pretreated with LA and AMVN and loaded with OVA. Experiments were

performed in triplicate and repeated twice. (F) Stimulation of OT-1 T cells by DCs pretreated with LA and AMVN and loaded with SIINFEKL peptide.

Experiments were performed in triplicate and repeated three times. (G) pMHC expression in CD11c+ DCs pretreated with LA and AMVN and loaded with

long OVA-derived peptide. Experiments were performed twice. (H) Stimulation of pmel-1 T cells by DCs pretreated with LA and AMVN and loaded with

long gp100-derived peptide. Experiments were performed in triplicate and repeated twice. *p , 0.05, statistically significant differences versus control in

all experiments.

The Journal of Immunology 2927

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

minor peroxidation of membrane phospholipids was detected (9.2 61.8 pmol/106 cells), which constitutes ∼3% of the total increase inthe content of oxidized lipids). Thus, these experimental conditionsrecapitulated accumulation of ox-lipids in DCs observed on DCsexposed to TES or TCM, and were used in further experiments.After loading with LA, with or without subsequent treatment

with AMVN, DCs were incubated overnight with OVA and thenused for stimulation of OT-1 T cells. DCs, treated with LA orAMVN alone, showed no defect in Ag cross-presentation func-tion, whereas a combination of AMVN and LA significantly reducedthe ability of the DCs to stimulate Ag-specific T cells (Fig. 6E). Incontrast, if DCs were directly loaded with SIINFEKL peptide, thecombination of AMVN and LA did not affect DC stimulation ofOT-1 T cells (Fig. 6F). Treatment of DCs with LA alone did notaffect the pMHC surface expression after loading of the cells withlong OVA peptide, whereas AMVN by itself caused a small de-crease in the levels of surface pMHC (Fig. 6G). In contrast, thecombination of AMVN and LA resulted in a .2-fold decrease ofpMHC surface expression (Fig. 6G), confirming that the accumu-lation of ox-lipids in DCs specifically inhibits Ag cross-presentationbut does not affect DC function if the peptide is directly loaded ontothe MHC-I on the surface. To confirm these observations in a dif-ferent experimental system, we loaded DCs with long peptide de-rived from gp100 protein and used pmel-1 T cells as responders.The combination of LA and AMVN, but not each of them sepa-rately, significantly (p , 0.05) inhibited the cross-presentationfunction of the DCs (Fig. 6H).To verify the specific role of lipid oxidation in defective cross-

presentation in our experimental system, we treated DCs witha higher concentration of AMVN (0.5 mM). At this dose, AMVNdirectly inhibited the ability of OVA-loaded DCs to stimulate OT-1T cells without need to preload DCs with LA (Fig. 7A). To blocklipid peroxidation, we pretreated cells with a-tocopherol (vitamin

E), which is known to partition into hydrophobic phases of lipidstructures and has been shown to effectively inhibit lipid perox-idation in biomembranes and lipoproteins (34). Tocopherol abro-gated the inhibitory effect of AMVN on cross-presentation ofOVA to T cells (Fig. 7A), and abrogated the inhibitory effect ofAMVN on the expression of pMHC, after DC loading with eitherOVA or long OVA-derived peptide (Fig. 7B). The negative effectof AMVN on surface pMHC levels affected only cross-presentation,because it was not observed in DCs obtained from OVA-Tg mice(Fig. 7C). Likewise, treatment of DCs with AMVN inhibited thestimulation of OT-1 T cells by DCs loaded with OVA, but did notaffect DCs from OVA-Tg mice (Fig. 7D). Thus, ox-lipids candirectly inhibit cross-presentation in DCs without affecting thepresentation of endogenous Ag.

Mechanism of inhibition of cross-presentation by oxidizedlipids

The endogenous pathway of presentation invariably depends onproteasomes, whereas the pathway of cross-presentation (cytosolicor vacuolar) depends on the type of Ag and the mechanism of itsuptake (12, 35). In our experiments, inhibition of proteasomeswith lactacystin almost completely abrogated the surface expres-sion of pMHC in DCs isolated from OVA-Tg mice (Fig. 8A). Incontrast, lactacystin had little effect on cross-presentation OVA byDCs (Fig. 8A, 8B) and a modest effect on cross-presentation oflong OVA-derived peptide (Fig. 8A), suggesting that a vacuolarmechanism could be the main pathway involved in cross-presentationunder these circumstances.IFN-g is known to upregulate MHC-I expression and increase

cross-presentation in DCs. We assessed whether pretreatment ofDCs with IFN-g would abrogate the inhibitory effect of TES oncross-presentation. As expected, in DCs cultured in CM, IFN-gcaused a marked upregulation of overall H-2Kb and a .2-fold

FIGURE 7. Role of oxidized lipids in cross-presentation. (A) The protection effect of vitamin E on cross-presentation. DCs were treated with vitamin E

(50 mM) for 24 h, followed by 2-h treatment with 0.5 mM AMVN. Cells then were loaded with 100 mg/ml OVA overnight and used for stimulation of OT-1

T cells. Typical results of T cell proliferation of four performed experiments are shown. T cells alone had [3H]thymidine uptake of ,500 cpm. (B)

Percentage of CD11c+ DCs expressing pMHC, after the treatment with vitamin E and AMVN, as described earlier. (Left panel) DCs loaded with 100 mg/ml

OVA; (right panel) DCs loaded with 5 mg/ml OVA-derived long peptide. Cumulative result of four experiments is shown. (C) pMHC expression on DCs

generated from BM of OVA-Tg mice and treated with 0.5 mM AMVN. Three experiments with the same results were performed. (D) Proliferation of OT-1

T cells stimulated with DCs from wild-type mice loaded with OVA and from OVA-Tg mice. DCs were pretreated with AMVN (0.5 mM) for 2 h. Typical

results of four performed experiments are shown. *p , 0.05, statistically significant differences versus control in all experiments.

2928 OXIDIZED LIPIDS AND DCs IN CANCER

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

increase in the expression of pMHC, after DC loading with longOVA peptide (Supplemental Fig. 4C). IFN-g substantially upreg-ulated the expression of H-2Kb on DCs cultured with TES fromthree different tumors (Fig. 8C, top panels). However, IFN-g didnot completely overcome the block in pMHC expression (Fig. 8C,bottom panels). Treatment of DCs with IFN-g did not rescue DCsinhibited by TES or affect DCs ability to stimulate OT-1 T cellproliferation after cross-presentation of OVA protein (Fig. 8D).We also tested the effect of IFN-g on cross-presentation of DCsloaded with LA and treated with AMVN. Similar to the effectobserved in the presence of TES, IFN-g upregulated the H-2Kb

expression, but failed to abrogate the defect in pMHC expression(Fig. 8E). Despite the presence of a small population of DCs withincreased expression of pMHC, we did not find an improvement inthe ability of DCs to stimulate Ag-specific T cells after cross-presentation of OVA-derived long peptide (Fig. 8F).

DiscussionIn this article, we report that oxidized lipids inhibit cross-presentationin DCs. This may have an implication on therapeutic efforts inpatients with low tumor burden, who are considered optimal

candidates for cancer vaccines. Accumulation of lipids in DCs,from TB hosts, is mediated via upregulation of scavenger re-ceptor (Msr1 or CD204). This receptor binds various acetylatedand oxidized lipids, including low-density lipoproteins (36).Msr1-deficient DCs showed a more effective Ag-presenting ca-pability as compared with wild-type cells, but this was primarilylinked with their more mature phenotype (37). In the absence ofCD204 (in Msr12/2 mice), TDF failed to inhibit the Ag cross-presentation function in DCs, supporting the possible role ofa lipid uptake in the negative effects of TDF on DCs. Extracel-lular lipids may affect DC function via various receptor-mediated mechanisms (38–41). However, our data indicatedthat the effect of lipids on cross-presentation was primarily in-tracellular. When DCs treated with TDF were separated based ontheir level of intracellular lipids, the defect in cross-presentationwas associated only with high lipid-laden DCs. We previouslyhave shown that although TES obtained from different tumor celllines had variable levels of TAGs, they caused comparable upregu-lation of lipids in DCs (27). It is known that TAGs are not directlytaken up by DCs but are synthesized inside the cells from FAs.CD204 binds modified primarily oxidized lipoproteins (36).

FIGURE 8. The mechanisms of oxidized lipids effect on cross-presentation by DCs. (A) pMHC expression on DCs treated with 5 mM proteasome inhibitor

lactacystin for 48 h. Wild-type DCs were directly loaded with 100 mg/ml OVA or 5 mg/ml long OVA-derived peptide, during the last 24 h of culture, in the

presence of lactacystin. Two experiments, with the same results, were performed. (B) Effect of lactacystin on the ability of DCs, loaded with OVA (100 mg/ml),

to stimulate proliferation of OT-1 T cells. (C and D) Effect of IFN-g on cross-presentation of long OVA-derived peptide by DCs treated with TES. DCs were

cultured with IFN-g (250 ng/ml) for 1 d, followed by TES treatment and loading with long OVA-derived peptide. (C, top panel) pMHC expression; (bottom

panel) MHC-I (H2Kb) expression. CT26, LLC, and EL-4 designate type of TES used. (D) Effect of IFN-g on the ability of DCs to stimulate proliferation of

OT-1 T cells after loading with OVA. (E and F) Effect of IFN-g on cross-presentation of DCs loaded with LA and treated with AMVN. (E, top panel) MHC-I

(H2Kb) expression; (bottom panel) pMHC expression. (F) Effect of IFN-g on the ability of DCs to stimulate proliferation of OT-1 T cells after loading with

OVA. *p , 0.05, statistically significant differences versus control in all experiments.

The Journal of Immunology 2929

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Small LBs are present in all DCs and have been implicated incross-presentation via interaction with Irgm3, an IFN-related GTPase.Irgm3 is required for IFN-g–induced efficient cross-presentation(24). Using high-resolution confocal microscopy, van Manen et al.(42) have demonstrated the transient association of LBs with latexbead-containing phagosomes in neutrophils and macrophages. IfLBs are important for cross-presentation and could be transientlyassociated with phagosomes, then it was conceivable that largeLBs could disrupt cross-presentation. However, our data argueagainst this possibility, Because we could not detect colocalizationof large LBs with any cellular compartment associated with cross-presentation or with pMHC. Treatment of DCs with IFN-g did notrescue the defect of cross-presentation caused by TDF, despite thesubstantial upregulation of MHC-I. Loading of DCs with LA,which resulted in formation of large LBs, did not cause the defectin cross-presentation by DCs. Thus, it appears that the mere ac-cumulation of lipids in DCs is not sufficient to affect cross-pre-sentation.Our data demonstrated a substantial accumulation of different

classes of neutral ox-lipids in DCs in TB hosts, which is consistentwith the important role of CD204 in this process (43). To directlytest the hypothesis that ox-lipids are involved in regulation ofcross-presentation, we could not use a treatment with extracellularoxidized lipoproteins or lipids. Oxidized and nonoxidized lipidsand lipoproteins differ in the nature of the receptors they bind andsignaling they cause, as well as in the kinetic of their uptake by thecells (44). DCs are lacking the machinery for lipid peroxidationand TDF did not cause the upregulation of myeloperoxidase orROS needed for this process. Therefore, in our model experi-ments, we used an LA and lipophilic generator of peroxyl radi-cals, AMVN. Our data directly implicated intracellular ox-lipidsin the inhibition of cross-presentation in DCs.Why do oxidized lipids affect only cross-presentation and not

presentation of endogenous Ags? In contrast with endogenouspresentation, which involves the proteasomal degradation of pro-teins, the transfer of peptides to the ER by TAP, and the formationof pMHC in the ER with the subsequent transport to the membrane(45), these components are not essential for the vacuolar pathwayin Ag cross-presentation (46). This pathway largely involves therecycling of MHC-I from the surface and formation of new pMHCin endosomes and lysosomes via peptide exchange. The cyto-plasmic domain of MHC-I can direct the protein to both the endo-somal and lysosomal compartments of DCs, where loading ofpeptides derived from exogenous proteins can occur (47, 48). Inaddition to recycling MHC-I, transport of MHC-I, from ER to theendocytic compartment, has also been demonstrated (49, 50).We propose the concept of disrupted DC Ag cross-presentation

in cancer. Under physiological conditions, DCs, in contrast withneutrophils and macrophages, produce and accumulate very lowlevels of peroxidized lipids. In TB hosts, TDFs induce the up-regulation of Msr1 (27), which facilitates an uptake of peroxidizedlipids present in plasma or TCM. Inside DCs, oxidized lipid moietycan be directly trafficked to endosomes and localized in the lyso-somal compartment (51). In addition, the accumulation of lipids inDCs manifests in the formation of large LBs. The presence andsize of LBs is defined by the accumulation of FA precursors andtheir esterification into TAGs and CEs, the major constituents ofthe hydrophobic core of LBs (26). The significant uptake of oxF-FAs, in particular, oxLA, facilitates its integration into ox-LA–containing TAGs and CEs. The presence of very polar groups in theseoxidized lipid species, or oxidatively truncated TAGs, causes theirtranslocation and integration into the less hydrophobic surfacearea of LBs—as has been confirmed by our computational coarse-grained molecular dynamic simulations (52), where they can be

further metabolized (e.g., hydrolyzed) by the respective hydro-lyzing enzymes. These peroxidized lipid species could be locatedin close molecular proximity to the contact sites with neighboringorganelles, such as lysosomes. As a result of this proximity, thedirect transfer of peroxidized lipids from LBs into lysosomesbecomes possible. The exact mechanisms and pathways throughwhich these peroxidized lipids affect the cross-presentation pro-cess remain to be elucidated.Our data describe a novel mechanism of inhibition cross-presen-

tation in DCs, associated with oxidized lipids, which may playan important role in the negative regulation of cross-presentation incancer and, possibly, in other pathological processes associatedwith the accumulation of oxidized lipids, and may suggest potentialtargets for therapeutic regulation of cross-presentation. It is pos-sible that this mechanism may affect recycling of other receptors,and thus affect function of the cells.

DisclosuresThe authors have no financial conflicts of interest.

References1. Hildner, K., B. T. Edelson, W. E. Purtha, M. Diamond, H. Matsushita,

M. Kohyama, B. Calderon, B. U. Schraml, E. R. Unanue, M. S. Diamond, et al.2008. Batf3 deficiency reveals a critical role for CD8alpha+ dendritic cells incytotoxic T cell immunity. Science 322: 1097–1100.

2. Knight, S. C., S. Iqball, M. S. Roberts, S. Macatonia, and P. A. Bedford. 1998.Transfer of antigen between dendritic cells in the stimulation of primary T cellproliferation. Eur. J. Immunol. 28: 1636–1644.

3. Yewdall, A. W., S. B. Drutman, F. Jinwala, K. S. Bahjat, and N. Bhardwaj. 2010.CD8+ T cell priming by dendritic cell vaccines requires antigen transfer toendogenous antigen presenting cells. PLoS ONE 5: e11144.

4. Gabrilovich, D. I., S. Ostrand-Rosenberg, and V. Bronte. 2012. Coordinatedregulation of myeloid cells by tumours. Nat. Rev. Immunol. 12: 253–268.

5. Ascierto, P. A., M. Napolitano, E. Celentano, E. Simeone, G. Gentilcore,A. Daponte, M. Capone, C. Caraco, R. Calemma, G. Beneduce, et al. 2010.Regulatory T cell frequency in patients with melanoma with different disease stageand course, and modulating effects of high-dose interferon-alpha 2b treatment. J.Transl. Med. 8: 76.

6. Yigit, R., C. G. Figdor, P. L. Zusterzeel, J. M. Pots, R. Torensma, andL. F. Massuger. 2011. Cytokine analysis as a tool to understand tumour-hostinteraction in ovarian cancer. Eur. J. Cancer 47: 1883–1889.

7. Yamagami, W., N. Susumu, H. Tanaka, A. Hirasawa, K. Banno, N. Suzuki, H.Tsuda, K. Tsukazaki, and D. Aoki. 2011. Immunofluorescence-detected infil-tration of CD4+FOXP3+ regulatory T cells is relevant to the prognosis ofpatients with endometrial cancer. Int. J. Gynecol. Cancer 21: 1628–1634.

8. Shen, X., N. Li, H. Li, T. Zhang, F. Wang, and Q. Li. 2010. Increased prev-alence of regulatory T cells in the tumor microenvironment and its correlationwith TNM stage of hepatocellular carcinoma. J. Cancer Res. Clin. Oncol. 136:1745–1754.

9. Bevan, M. J. 1976. Minor H antigens introduced on H-2 different stimulating cellscross-react at the cytotoxic T cell level during in vivo priming. J. Immunol. 117:2233–2238.

10. Bevan, M. J. 1976. Cross-priming for a secondary cytotoxic response to minor Hantigens with H-2 congenic cells which do not cross-react in the cytotoxic assay.J. Exp. Med. 143: 1283–1288.

11. Burgdorf, S., and C. Kurts. 2008. Endocytosis mechanisms and the cell biologyof antigen presentation. Curr. Opin. Immunol. 20: 89–95.

12. Joffre, O. P., E. Segura, A. Savina, and S. Amigorena. 2012. Cross-presentationby dendritic cells. Nat. Rev. Immunol. 12: 557–569.

13. Alvarez, D., E. H. Vollmann, and U. H. von Andrian. 2008. Mechanisms andconsequences of dendritic cell migration. Immunity 29: 325–342.

14. McDonnell, A. M., A. C. Prosser, I. van Bruggen, B. W. Robinson, and A. J. Currie.2010. CD8alpha+ DC are not the sole subset cross-presenting cell-associated tumorantigens from a solid tumor. Eur. J. Immunol. 40: 1617–1627.

15. van Mierlo, G. J., Z. F. Boonman, H. M. Dumortier, A. T. den Boer, M. F. Fransen,J. Nouta, E. I. van der Voort, R. Offringa, R. E. Toes, and C. J. Melief. 2004.Activation of dendritic cells that cross-present tumor-derived antigen licenses CD8+ CTL to cause tumor eradication. J. Immunol. 173: 6753–6759.

16. Gerner, M. Y., K. A. Casey, and M. F. Mescher. 2008. Defective MHC class IIpresentation by dendritic cells limits CD4 T cell help for antitumor CD8 T cellresponses. J. Immunol. 181: 155–164.

17. Engelhardt, J. J., B. Boldajipour, P. Beemiller, P. Pandurangi, C. Sorensen,Z. Werb, M. Egeblad, and M. F. Krummel. 2012. Marginating dendritic cells ofthe tumor microenvironment cross-present tumor antigens and stably engagetumor-specific T cells. Cancer Cell 21: 402–417.

18. Perrot, I., D. Blanchard, N. Freymond, S. Isaac, B. Guibert, Y. Pacheco, andS. Lebecque. 2007. Dendritic cells infiltrating human non-small cell lung cancerare blocked at immature stage. J. Immunol. 178: 2763–2769.

2930 OXIDIZED LIPIDS AND DCs IN CANCER

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

19. Gabrilovich, D. I. 2004. Mechanisms and functional significance of tumour-induced dendritic-cell defects. Nat. Rev. Immunol. 4: 941–952.

20. Melief, C. J. 2008. Cancer immunotherapy by dendritic cells. Immunity 29: 372–383.21. Preynat-Seauve, O., P. Schuler, E. Contassot, F. Beermann, B. Huard, and

L. E. French. 2006. Tumor-infiltrating dendritic cells are potent antigen-presentingcells able to activate T cells and mediate tumor rejection. J. Immunol. 176: 61–67.

22. Wakim, L. M., and M. J. Bevan. 2011. Cross-dressed dendritic cells drivememory CD8+ T-cell activation after viral infection. Nature 471: 629–632.

23. Dolan, B. P., K. D. Gibbs, Jr., and S. Ostrand-Rosenberg. 2006. Dendritic cellscross-dressed with peptide MHC class I complexes prime CD8+ T cells. J.Immunol. 177: 6018–6024.

24. Bougneres, L., J. Helft, S. Tiwari, P. Vargas, B. H. Chang, L. Chan, L. Campisi,G. Lauvau, S. Hugues, P. Kumar, et al. 2009. A role for lipid bodies in the cross-presentation of phagocytosed antigens by MHC class I in dendritic cells. Im-munity 31: 232–244.

25. Saka, H. A., and R. Valdivia. 2012. Emerging roles for lipid droplets in immunityand host-pathogen interactions. Annu. Rev. Cell Dev. Biol. 28: 411–437.

26. Walther, T. C., and R. V. Farese, Jr. 2012. Lipid droplets and cellular lipidmetabolism. Annu. Rev. Biochem. 81: 687–714.

27. Herber, D. L., W. Cao, Y. Nefedova, S. V. Novitskiy, S. Nagaraj, V. A. Tyurin,A. Corzo, H. I. Cho, E. Celis, B. Lennox, et al. 2010. Lipid accumulation anddendritic cell dysfunction in cancer. Nat. Med. 16: 880–886.

28. Zhou, J., P. Cheng, J. I. Youn, M. J. Cotter, and D. I. Gabrilovich. 2009. Notchand wingless signaling cooperate in regulation of dendritic cell differentiation.Immunity 30: 845–859.

29. Remily-Wood, E. R., R. Z. Liu, Y. Xiang, Y. Chen, C. E. Thomas, N. Rajyaguru,L. M. Kaufman, J. E. Ochoa, L. Hazlehurst, J. Pinilla-Ibarz, et al. 2011. A da-tabase of reaction monitoring mass spectrometry assays for elucidating thera-peutic response in cancer. Proteomics Clin. Appl. 5: 383–396.

30. Folch, J., M. Lees, and G. H. Sloan Stanley. 1957. A simple method for theisolation and purification of total lipides from animal tissue. J. Biol. Chem. 226:497–509.

31. Overwijk, W. W., A. Tsung, K. R. Irvine, M. R. Parkhurst, T. J. Goletz, K. Tsung,M. W. Carroll, C. Liu, B. Moss, S. A. Rosenberg, and N. P. Restifo. 1998. gp100/pmel 17 is a murine tumor rejection antigen: induction of “self”-reactive, tumor-icidal T cells using high-affinity, altered peptide ligand. J. Exp. Med. 188: 277–286.

32. Bieghs, V., F. Verheyen, P. J. van Gorp, T. Hendrikx, K. Wouters, D. L€utjohann,M. J. Gijbels, M. Febbraio, C. J. Binder, M. H. Hofker, and R. Shiri-Sverdlov. 2012.Internalization of modified lipids by CD36 and SR-A leads to hepatic inflammationand lysosomal cholesterol storage in Kupffer cells. PLoS ONE 7: e34378.

33. Tyurina, Y. Y., V. A. Tyurin, A. A. Shvedova, J. P. Fabisiak, and V. E. Kagan.2002. Peroxidation of phosphatidylserine in mechanisms of apoptotic signaling.Methods Enzymol. 352: 159–174.

34. Tsuchiya, M., V. E. Kagan, H. J. Freisleben, M. Manabe, and L. Packer. 1994.Antioxidant activity of alpha-tocopherol, beta-carotene, and ubiquinol in mem-branes: cis-parinaric acid-incorporated liposomes. Methods Enzymol. 234: 371–383.

35. Burgdorf, S., C. Scholz, A. Kautz, R. Tampe, and C. Kurts. 2008. Spatial andmechanistic separation of cross-presentation and endogenous antigen presenta-tion. Nat. Immunol. 9: 558–566.

36. Platt, N., and S. Gordon. 2001. Is the class A macrophage scavenger receptor(SR-A) multifunctional? - The mouse’s tale. J. Clin. Invest. 108: 649–654.

37. Wang, X. Y., J. Facciponte, X. Chen, J. R. Subjeck, and E. A. Repasky. 2007.Scavenger receptor-A negatively regulates antitumor immunity. Cancer Res. 67:4996–5002.

38. Aliberti, J., S. Hieny, C. Reis e Sousa, C. N. Serhan, and A. Sher. 2002. Lipoxin-mediated inhibition of IL-12 production by DCs: a mechanism for regulation ofmicrobial immunity. Nat. Immunol. 3: 76–82.

39. Weatherill, A. R., J. Y. Lee, L. Zhao, D. G. Lemay, H. S. Youn, andD. H. Hwang. 2005. Saturated and polyunsaturated fatty acids reciprocallymodulate dendritic cell functions mediated through TLR4. J. Immunol. 174:5390–5397.

40. Shamshiev, A. T., F. Ampenberger, B. Ernst, L. Rohrer, B. J. Marsland, andM. Kopf. 2007. Dyslipidemia inhibits Toll-like receptor-induced activation ofCD8alpha-negative dendritic cells and protective Th1 type immunity. J. Exp.Med. 204: 441–452.

41. Angeli, V., J. Llodra, J. X. Rong, K. Satoh, S. Ishii, T. Shimizu, E. A. Fisher, andG. J. Randolph. 2004. Dyslipidemia associated with atherosclerotic diseasesystemically alters dendritic cell mobilization. Immunity 21: 561–574.

42. van Manen, H.-J., Y. M. Kraan, D. Roos, and C. Otto. 2005. Single-cell Ramanand fluorescence microscopy reveal the association of lipid bodies with phag-osomes in leukocytes. Proc. Natl. Acad. Sci. USA 102: 10159–10164.

43. Tyurin, V. A., W. Cao, Y. Y. Tyurina, D. I. Gabrilovich, and V. E. Kagan. 2011.Mass-spectrometric characterization of peroxidized and hydrolyzed lipids inplasma and dendritic cells of tumor-bearing animals. Biochem. Biophys. Res.Commun. 413: 149–153.

44. Zaguri, R., I. Verbovetski, M. Atallah, U. Trahtemberg, A. Krispin, E. Nahari,E. Leitersdorf, and D. Mevorach. 2007. ‘Danger’ effect of low-density lipo-protein (LDL) and oxidized LDL on human immature dendritic cells. Clin. Exp.Immunol. 149: 543–552.

45. Jensen, P. E. 2007. Recent advances in antigen processing and presentation. Nat.Immunol. 8: 1041–1048.

46. Touret, N., P. Paroutis, M. Terebiznik, R. E. Harrison, S. Trombetta, M. Pypaert,A. Chow, A. Jiang, J. Shaw, C. Yip, et al. 2005. Quantitative and dynamic as-sessment of the contribution of the ER to phagosome formation. Cell 123: 157–170.

47. Basha, G., G. Lizee, A. T. Reinicke, R. P. Seipp, K. D. Omilusik, andW. A. Jefferies. 2008. MHC class I endosomal and lysosomal trafficking coin-cides with exogenous antigen loading in dendritic cells. PLoS ONE 3: e3247.

48. Lizee, G., G. Basha, J. Tiong, J. P. Julien, M. Tian, K. E. Biron, andW. A. Jefferies. 2003. Control of dendritic cell cross-presentation by the majorhistocompatibility complex class I cytoplasmic domain. Nat. Immunol. 4: 1065–1073.

49. Gagnon, E., S. Duclos, C. Rondeau, E. Chevet, P. H. Cameron, O. Steele-Mortimer, J. Paiement, J. J. Bergeron, and M. Desjardins. 2002. Endoplasmicreticulum-mediated phagocytosis is a mechanism of entry into macrophages.Cell 110: 119–131.

50. Basha, G., K. Omilusik, A. Chavez-Steenbock, A. T. Reinicke, N. Lack,K. B. Choi, and W. A. Jefferies. 2012. A CD74-dependent MHC class I endo-lysosomal cross-presentation pathway. Nat. Immunol. 13: 237–245.

51. Al Gadban, M. M., K. J. Smith, F. Soodavar, C. Piansay, C. Chassereau,W. O. Twal, R. L. Klein, G. Virella, M. F. Lopes-Virella, and S. M. Hammad.2010. Differential trafficking of oxidized LDL and oxidized LDL immunecomplexes in macrophages: impact on oxidative stress. PLoS ONE 5:e12534.

52. Mohammadyani, D., V. A. Tyurn, A. A. Amoscato, D. I. Gabrilovich, J. Klein-Seetharaman, and V. E. Kagan. 2014. Computational approaches to under-standing the role of oxidized tri-aclyglycerols in suppression of antigen cross-presentation in cancer. Biophys. J. 106: 805a–806a.

The Journal of Immunology 2931

by guest on June 17, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Corrections

Cao, W., R. Ramakrishnan, V. A. Tuyrin, F. Veglia, T. Condamine, A. Amoscato, D. Mohammadyani, J. J. Johnson, L. M. Zhang,J. Klein-Seetharaman, E. Celis, V. E. Kagan, and D. I. Gabrilovich. 2014. Oxidized lipids block antigen cross-presentation by dendritic cellsin cancer. J. Immunol. 192: 2920–2931.

The author list was published incorrectly. The corrected author line is shown below. In addition, an author’s last name was misspelled.The correct spelling is Vladimir A. Tyurin.

Rupal Ramakrishnan, Vladimir A. Tyurin, Filippo Veglia, Thomas Condamine, Andrew Amoscato, Dariush Mohammadyani,Joseph J. Johnson, Lan Min Zhang, Judith Klein-Seetharaman, Esteban Celis, Valerian E. Kagan and Dmitry I. Gabrilovich.

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1490017

Copyright � 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00

The Journal of Immunology