Cells T+ MHC Class I Complexes Prime CD8 Dendritic Cells Cross

Dendritic Cells Cross-Dressed with Peptide MHC Class I Complexes Prime CD8 + T Cells

This information is current as Brian P. Dolan, Kenneth D. Gibbs, Jr and Suzanne of September 29, 2021. Ostrand-Rosenberg J Immunol 2006; 177:6018-6024; ; doi: 10.4049/jimmunol.177.9.6018 http://www.jimmunol.org/content/177/9/6018 Downloaded from

References This article cites 37 articles, 21 of which you can access for free at: http://www.jimmunol.org/content/177/9/6018.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 29, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Dendritic Cells Cross-Dressed with Peptide MHC Class I Complexes Prime CD8؉ T Cells1

Brian P. Dolan, Kenneth D. Gibbs, Jr., and Suzanne Ostrand-Rosenberg2

The activation of naive CD8؉ T cells has been attributed to two mechanisms: cross-priming and direct priming. Cross-priming ,and direct priming differ in the source of Ag and in the cell that presents the Ag to the responding CD8؉ T cells. In cross-priming exogenous Ag is acquired by professional APCs, such as dendritic cells (DC), which process the Ag into peptides that are subsequently presented. In direct priming, the APCs, which may or may not be DC, synthesize and process the Ag and present it themselves to CD8؉ T cells. In this study, we demonstrate that naive CD8؉ T cells are activated by a third mechanism, called cross-dressing. In cross-dressing, DC directly acquire MHC class I-peptide complexes from dead, but not live, donor cells by a cell contact-mediated mechanism, and present the intact complexes to naive CD8؉ T cells. Such DC are cross-dressed because they ؉ are wearing peptide-MHC complexes generated by other cells. CD8 T cells activated by cross-dressing are restricted to the MHC Downloaded from class I genotype of the donor cells and are speciﬁc for peptides generated by the donor cells. In vivo studies demonstrate that optimal priming of CD8؉ T cells requires both cross-priming and cross-dressing. Thus, cross-dressing may be an important mechanism by which DC prime naive CD8؉ T cells and may explain how CD8؉ T cells are primed to Ags that are inefﬁciently cross-presented. The Journal of Immunology, 2006, 177: 6018–6024.

ϩ ϩ

ntigen presentation to CD8 T lymphocytes is postu- fected DC also directly present viral Ags to CD8 T cells (12, 13). http://www.jimmunol.org/ lated to occur through two mechanisms, which differ Ultimately, CD8ϩ T cell priming to viral infections may be the from each other by the type of cell that processes and result of both cross-priming and direct priming (14). Cross-prim- A ϩ presents the peptide Ag. According to the direct Ag presentation ing has also been demonstrated for CD8 T cells activated to mechanism, CD8ϩ T cells recognize their cognate peptide-MHC tumor Ags (15–17); however, other studies suggest that priming of class I complexes on the surface of the cells that synthesized the tumor-speciﬁc CD8ϩ T cells occurs through the direct presentation Ag, such as virally infected cells or malignant cells (1). Alterna- of tumor Ags by tumor cells (18, 19). Therefore, a consensus tively, CD8ϩ T cells recognize peptide-MHC class I complexes on model for CD8ϩ T cell priming does not exist. the surface of cells, which have captured exogenously synthesized In this study, we report that CD8ϩ T cells can be primed by a

Ag, and subsequently process and present the Ag bound to their third mechanism, termed cross-dressing. During cross-dressing, by guest on September 29, 2021 own MHC class I proteins. This latter mechanism is termed cross- peptide-MHC class I complexes are transferred to DC from dead presentation, and dendritic cells (DC)3 are the predominant cell donor cells that synthesized the complexes. Recipient DC then use population that captures and presents Ag (2, 3). If the process of the acquired peptide-MHC class I complexes to activate CD8ϩ T cross-presentation results in the activation of naive T cells, then the cells that are peptide specific and restricted to the MHC class I T cells have been activated by cross-priming. genotype of the donor cells. Optimal priming of CD8ϩ T cells Although compelling evidence supports cross-priming as the probably results from a combination of cross-dressing and cross- principal mechanism for activating naive CD8ϩ T cells (2, 4, 5), priming, because cross-dressing efficiently activates DC that are other studies are consistent with CD8ϩ T cells being activated by not efficiently activated by cross-priming. the direct presentation mechanism (6, 7). The concept of cross- priming is supported by many reports demonstrating that CD8ϩ T cells specific for virally infected cells are primed by DC cross- Materials and Methods presenting virally derived peptides (8–11), although virally in- Mice Breeding stocks of BALB/c, C57BL/6, FVB, OT-I, and CD11c-diphtheria toxin receptor (DTR) (Itgax-DTR/enhanced GFP) mice on a BALB/c back- Department of Biological Sciences, University of Maryland Baltimore County, Bal- ground were from The Jackson Laboratory. MHC I/IIϪ/Ϫ mice and OT-I/ timore, MD 21250 RAG 2Ϫ/Ϫ mice were from Taconic Farms. CD11c-DTR mice were Received for publication June 5, 2006. Accepted for publication August 7, 2006. screened, as described (20). OT-I mice were identified by immunofluores- cent staining of PBLs for CD8 and V␣2. H-2b/CD11c-DTRϩ/Ϫ mice were The costs of publication of this article were defrayed in part by the payment of page generated by crossing CD11c-DTR mice with MHC class I/II-deficient charges. This article must therefore be hereby marked advertisement in accordance ϩ/Ϫ with 18 U.S.C. Section 1734 solely to indicate this fact. mice, and the CD11c-DTR F1 offspring were backcrossed to MHC ϩ/Ϫ 1 class I/II-deficient mice. CD11c-DTR F2 mice that were negative for This work was supported by National Institutes of Health Grants CA52527, both MHC class I and II (identified by immunofluorescence staining for the CA84232, and CA115880, Department of Defense Grant DAMD-17-01-1-0312, and ϩ ϩ Susan G. Komen Foundation Grant BCTR0503885. B.P.D. is supported by Depart- absence of peripheral CD4 and CD8 T cells) were then crossed with ment of Defense Grant DAMD-17-03-1-0334 predoctoral fellowship, and K.D.G. is C57BL/6 mice. The F1 mice of this cross were subsequently used in experi- supported by a National Institutes of Health MARC-U-STAR training grant ments. All animal procedures were approved by the University of Maryland (GM08663). Baltimore County Institutional Animal Care and Use Committee. 2 Address correspondence and reprint requests to Dr. Suzanne Ostrand-Rosenberg, Department of Biological Sciences, University of Maryland, 1000 Hilltop Circle, Cells, transfections, and Abs Baltimore, MD 21250. 3 Abbreviations used in this paper: DC, dendritic cell; BMDC, bone marrow-derived B16.BL6 8.2, A20, EL4, and EL4/OVA were cultured, as described (20– DC; DTR, diphtheria toxin receptor; DTx, diphtheria toxin. 22). NIH3T3 cells were grown in DMEM supplemented with 10% FCS

(HyClone) DAP/OVA cells (23) were maintained in RPMI 1640 (Bioflu- Ag presentation assays ids) supplemented with 10% FCS, 1 mM HEPES (Invitrogen Life Tech- 5 nologies), 5 ϫ 10Ϫ6 M 2-ME (Sigma-Aldrich), 400 ␮g/ml G418 (Calbio- Splenic DC were plated in 96-well plates in DC medium at 10 cells/well. ϫ chem), and 200 ␮g/ml zeocin (Invitrogen Life Technologies). Medium for One hour later, freeze-thawed donor cells at 3–4 cell equivalents or all cell lines contained 1% penicillin, 1% streptomycin, 1% gentamicin soluble OVA protein (Sigma-Aldrich) were added. Following a 3-h incu- (Biofluids), and 1% Glutamax (Invitrogen Life Technologies) (complete bation, wells were washed with excess PBS to remove dead material. medium). The DAP/OVA/Kb cell line was generated by transfection using BMDC were incubated with freeze-thawed donor cells or with soluble OVA protein for 3 h, the mixture was passaged over Ficoll to remove dead Lipofectamine Plus (Invitrogen Life Technologies), according to the man- 5 ␮ b material, and the purified BMDC were plated at 10 cells/well in 96-well ufacturer’s recommendations. Approximately 4 g of the K plasmid ϩ ϩ pKbT.8 (24, 25) was linearized with PvuI (Fermentas Life Sciences), eth- plates. CD8 OT-I T cells were isolated using CD8 T cell isolation kits, anol precipitated, resuspended in 100 ␮l of DMEM (Biofluids), mixed with according to the manufacturer’s suggested protocol (Miltenyi Biotec), and 5 ␮ 16 ␮l of Plus reagent, and incubated for 15 min at room temperature. One added to the purified BMDC at 10 cells/well for a total volume of 200 l hundred microliters of DMEM containing 10 ␮l of lipofectamine was then of complete RPMI 1640/well. OT-I activation was quantified by assaying ␥ added, and the mixture was incubated for an additional 15 min. The DNA culture supernatants for IL-2 and IFN- by ELISA (Pierce Endogen), as ␣ mixture was then added to DAP/OVA cells plated the previous night at 4 ϫ described (28), or by double staining for CD69 and V 2. 5 10 cells/well of a 6-well plate, in a final volume of 0.4 ml of DMEM. Cells In vivo T cell proliferation were cultured with DNA for 3 h before the addition of 2 ml of complete ϩ medium and an additional 60 ␮l of FCS. H-2Kb cells were identified by OT-I T cells were isolated by magnetic bead sorting from OT-I/RAG 2Ϫ/Ϫ staining with fluorescently tagged mAb (clone AF6-88.5; BD Pharmingen). mice and incubated for 10 min at room temperature in 2.5 ␮M CFSE ϩ Bulk H-2Kb transfectants were fluorescently sorted using a Coulter Epics (Molecular Probes) in PBS at 107 cells/ml. The reaction was quenched by Altra cell sorter. Fibroblasts were generated as follows: minced tail tissue the addition of 1 ml of FCS, and the cells were washed with excess PBS. was resuspended in 2.5 ml of a 2 mg/ml collagenase IV (Worthington A total of 5 ϫ 106 labeled cells was injected i.v. into the tail vein of Biochemical) solution supplemented with 10 ␮l of elastase (MP Biomedi- recipient mice, and 8–24 h later mice were inoculated s.c. in the lateral, Downloaded from cals), shaken gently for 1–2 days at 4°C, and washed with PBS, and ad- lower abdominal region with 5–10 ϫ 106 live, freeze-thawed, or a 1:1 herent cells were cultured in complete IMDM supplemented with 10% mixture of live and freeze-thawed cells. Three days later, mice were sac- FCS. The following Abs directly coupled to FITC or PE were purchased rificed, the draining lymph node (superficial inguinal) was removed and from BD Pharmingen: CD11c (HL3), CD8 (53-6.7), CD4 (GK1.5), CD69 dissociated, and the cells were examined by flow cytometry for dilution of (H1.2F3), Kk (36-7-5), Kb (AF6-88.5), Dd (34-2-12S), and V␣2 (B20.1). CFSE. Background auto fluorescence was determined using mice that did Anti-OVA mAb (clone 14) and anti-␤-tubulin mAb (Tub 2.1) were from not receive CFSE-labeled cells and was gated out. The proliferation index Sigma-Aldrich. Sheep anti-mouse HRP Ab was from Amersham. ϩ was calculated by dividing the total number of CFSE cells in all divisions http://www.jimmunol.org/ by the number of parental cells originally present, and was calculated using ModFit software (Verity Software House) (29). In CD11c depletion ex- Western blotting ϫ periments, (C57BL/6 CD11c-DTR)F1 mice were injected i.p. with 3 Western blots for ␤-tubulin were performed, as described (26). Immuno- ng/g body weight diphtheria toxin (DTx; Sigma-Aldrich), and 6 h later, blotting for OVA followed the same experimental procedure, except the inoculated with live and/or freeze-thawed donor cells. primary Ab for OVA (mAb clone 14) was diluted 1/5000 in 0.5% Tween 20-TBS solution containing 0.5% nonfat milk. Statistical analysis SDs and Student’s t test were calculated using Microsoft Excel 2002. DC isolation and culture

Results by guest on September 29, 2021 Splenic DC were isolated and cultured, as described (20), using CD11c MHC class I molecules transfer from dead cells to DC microbeads (Miltenyi Biotec). Bone marrow-derived DC (BMDC) were isolated by flushing the bone marrow from femurs of 3- to 6-mo-old mice, Acquisition of MHC class I by DC was first assessed in vitro using followed by RBC depletion using ACK lysing buffer (Biofluids). Cells BMDC and allogeneic donor cells. BMDC from FVB (H-2q) mice were cultured in DC medium (complete RPMI 1640 supplemented with 5% ϫ Ϫ6 were incubated for 3 h with freeze-thawed genetically mismatched FCS, 1 mM HEPES, 5 10 M 2-ME, 20 ng/ml GM-CSF, and 10 ng/ml k IL-4 (RDI)). Medium was changed every 2 days, and cells were 80–90% DAP/OVA (H-2 ) fibroblasts, and the resulting cells were purified CD11cϩ by day 6. DC were matured by overnight incubation in DC me- by passage over a Ficoll gradient. Ficoll-purified cells (Fig. 1A, top dium supplemented with 2.5 ␮g/ml LPS (Sigma-Aldrich) and 1 ␮g/ml left panel) bitmap to a different location than freeze-thawed cells CD40 mAb (clone 3/23 from BD Pharmingen). (Fig. 1A, bottom left panel), are Ͼ93% viable (Fig. 1A, middle panel), and are Ͼ90% CD11cϩ (right panel), identifying them as MHC I transfer experiments DC. To determine whether MHC class I transfers to DC, purified k DC (either splenic or BM derived) were plated in 1.0–1.5 ml of BMDC cells were stained for donor genotype H-2K molecules. As seen in medium without GM-CSF and IL-4 at 1–2 ϫ 106 cells/well in 6-well Fig. 1B (top left panel), H-2q DC incubated with freeze-thawed plates. Donor cells were killed by resuspending to a concentration of DAP/OVA cells stain for H-2Kk, demonstrating that the DC have 7 Ϫ 10 /ml in RPMI 1640 without serum and freezing at 80°C for 15–30 min, acquired MHC class I molecules. Similar results were obtained if followed by rapid thawing in a 37°C water bath. Cells were freeze thawed for one to three cycles until 100% of the cells were trypan blue positive. MHC class I-deficient BMDC were used as recipient DC instead of Apoptosis was induced by culturing donor cells in 20 ␮m camptothecin FVB BMDC (data not shown). MHC class I molecules also trans- (27) (Sigma-Aldrich) for 3 days, after which ϳ75% of the cells were trypan ferred to DC from apoptotic DAP/OVA cells (data not shown). In blue positive. Dead donor cells were added to DC at 3–5ϫ cell equivalents. agreement with previous studies of MHC class II transfer to DC After3hofculture, the mostly attached splenic DC were washed exten- sively in warm PBS to remove dead material, and the DC were removed by (20), MHC molecules are not detected on BMDC cultured with cell scraping. For transfer to BMDC, dead donor cells and live DC were live DAP/OVA cells (Fig. 1B, bottom left panel). centrifuged at 1100 ϫ g through Ficoll-Paque (Pharmacia Biotech) for 15 To test whether MHC class I molecules were retained during min at room temperature, and the live BMDC were recovered at the Ficoll- DC maturation, DC were cocultured with freeze-thawed DAP/ medium interface. If dead cells remained, the process was repeated until OVA cells, purified by Ficoll gradient, and cultured overnight in BMDC contained Ͻ5% dead cells, as measured by trypan blue uptake. Where indicated, donor cells were cultured with 200 U/ml IFN-␥ (Pierce the presence of LPS and mAbs to CD40 to induce maturation. The Endogen) for 2 days before freeze thawing. In some experiments, freeze- following day, BMDC were stained for donor MHC class I (H- thawed cells were centrifuged at 300 ϫ g, and the resulting supernatant was 2Kk). As shown in Fig. 1B (bottom right panel), donor H-2Kk used for MHC transfer. In other experiments, DC were plated in the lower protein is detected after overnight culture and DC maturation, half of an 8.0-␮m Transwell (Corning Glass), and freeze-thawed cells were ␮ demonstrating the stability of transferred MHC class I molecules. added to the top half. In control experiments, 10 l of 50-nm FITC-coated b latex beads (Polysciences) was added to the top half containing the freeze- To ensure specificity of MHC class I transfer, K -expressing thawed cells. DAP/OVA cells were generated (DAP/OVA/Kb; Fig. 1C), and we 6020 CROSS-DRESSED DC PRIME CD8ϩ T CELLS

and then stained for H-2Kk. Under these conditions, MHC class I transfer is Ͻ25% of the level when freeze-thawed cells and DC are mixed. To ensure that small molecules could traverse the Trans- well membrane, 50-nm FITC-coated latex beads were mixed with the freeze-thawed cells and added to one side of the Transwells. After3hofculture, BMDC were analyzed for bead uptake. Only a partial reduction of bead uptake was observed (ϳ29%; data not shown), indicating that small materials readily traversed the membrane. Collectively, these data suggest that optimal MHC class I transfer to DC requires direct contact with dead cells or large cellular debris, and is not efﬁciently mediated by soluble molecules or small vesicles such as exosomes (30–32) bearing MHC class I- peptide complexes.

MHC class I transfers from a variety of cells To determine whether DC acquisition of MHC class I was a gen- eral phenomenon, transfer experiments were conducted with the B16.BL6 8.2 melanoma (H-2Kb) and the A20 B lymphoma (H-2d) cell lines. As seen in Fig. 2A, H-2Kb molecules from the melanoma Downloaded from cell line and H-2Dd molecules from the B cell lymphoma were detected on BMDC from MHC-mismatched (FVB) mice. Modest transfer to FVB BMDC was even seen from primary fibroblasts derived from the tail tissue of BALB/c (H-2d) mice (Fig. 2A, right FIGURE 1. MHC class I molecules transfer from donor cells to DC. A, FS ϫ SS log profile, and propidium iodide and CD11c staining of bit- panel). These data in combination with previously published data mapped FVB (H-2q) BMDC cultured for 3 h with freeze-thawed DAP/ (20) demonstrate that a wide variety of cell types can donate MHC http://www.jimmunol.org/ OVA cells, purified by passage over a Ficoll gradient (top row). FS ϫ SS class I molecules to DC. log and propidium iodide staining of freeze-thawed DAP/OVA cells (bot- To determine whether the levels of MHC class I on the donor tom row). B, FVB BMDC were cultured alone or with freeze-thawed DAP/ cell affected the efficiency of transfer, EL4/OVA cells (H-2b), OVA cells, and analyzed for H-2Kk (top left histogram; isotype control which express very low levels of H-2Kb, were treated with IFN-␥ staining, top right histogram). BMDC were incubated alone or with live to up-regulate MHC class I expression (Fig. 2B). Freeze-thawed k DAP/OVA cells, and analyzed for H-2K (bottom left histogram). FVB IFN-␥-treated or untreated EL4/OVA cells were then cultured with BMDC were cultured alone or with freeze-thawed DAP/OVA cells, puri- FVB-derived BMDC. Transfer of H-2Kb molecules from freeze- fied, cultured overnight with LPS and anti-CD40 mAbs, and analyzed for k b thawed IFN-␥-treated EL4/OVA cells was greater than from

H-2K expression the following day. C, DAP/OVA or H-2K -transfected by guest on September 29, 2021 DAP/OVA (DAP/OVA/Kb) cells stained for H-2Kb. D, FVB BMDC were cultured alone or with freeze-thawed DAP/OVA (left histogram), or with DAP/OVA/Kb (right histogram) cells, purified, and analyzed for H-2Kb. E, FVB BMDC were cultured alone, with freeze-thawed DAP/OVA cells, with the supernatant of freeze-thawed DAP/OVA cells, or with a Transwell separating DC from freeze-thawed DAP/OVA cells. DC were subsequently analyzed for H-2Kk. All BMDC were purified by passage over Ficoll gradients before Ab staining. BMDC histograms show the Ab staining for CD11cϩ cells. These data are representative of three independent experiments. compared H-2Kb transfer with H-2q DC from freeze-thawed DAP/ OVA/Kb vs DAP/OVA cells. H-2Kb was detected on FVB BMDC exposed to DAP/OVA/Kb cells, but not on BMDC exposed to the parental DAP/OVA cells (Fig. 1D), confirming the transfer of do- nor MHC class I molecules to DC. Collectively, these results demonstrate that DC acquire MHC class I molecules from dead donor cells.

MHC class I transfer requires cell-to-cell contact To determine whether the particulate material or the supernatant of the freeze-thawed cells was involved in MHC class I transfer, freeze-thawed DAP/OVA cells were centrifuged at 300 ϫ g for 3 min, and the supernatant was removed and added to cultures of FVB BMDC. Three hours later, the DC were puriﬁed and labeled FIGURE 2. MHC class I molecules transfer to DC from a variety of for H-2Kk. As shown in Fig. 1E, MHC class I transfer in the ϳ cells. A, FVB BMDC were incubated alone or with freeze-thawed B16, presence of the supernatant is 30% of the level obtained with A20, or BALB/c ﬁbroblast cells, and analyzed for donor MHC class I. B, nonfractionated freeze-thawed donor cells. To test whether cell EL4/OVA or IFN-␥-treated EL4/OVA cells stained for H-2Kb. C, FVB contact between dead cells and DC is required for MHC class I BMDC were incubated alone or with freeze-thawed EL4/OVA or IFN-␥- transfer, FVB BMDC were separated from freeze-thawed DAP/ treated EL4/OVA cells and analyzed for H-2Kb. These data are represen- OVA cells for the 3-h culture period using an 8.0-␮m Transwell, tative of three independent experiments. The Journal of Immunology 6021

Transferred peptide-MHC class I complexes activate CD8ϩ T cells The MHC class I molecules acquired by DC may activate T cells. To test this hypothesis, FVB BMDC were cultured with freeze- thawed EL4 cells transfected with the OVA gene (EL4/OVA), purified, and tested as APCs with OVA-specific, H-2Kb-restricted, OT-I transgenic CD8ϩ T cells. BMDC exposed to EL4/OVA, but not EL4 or OVA protein, activated OT-I T cells, as measured by IL-2 release (Fig. 3A). Similar results were obtained using splenic DC (data not shown). Therefore, DC acquire MHC class I molecules and Ag from donor cells, and use the acquired material to prime CD8ϩ T cells. T cell activation in Fig. 3A could be due to the transfer of MHC class I-OVA complexes formed in the donor EL4/OVA cells, or to the independent transfer of MHC class I and Ag and the subse- quent assembly of MHC class I-peptide complexes in the recipient DC. To distinguish these two possibilities, FVB BMDC were cultured for 3 h with either freeze-thawed EL4/OVA cells or a combination of freeze-thawed EL4 cells and OVA protein. BMDC Downloaded from were purified and then tested for their ability to activate OT-I T cells. OT-I T cell activation only occurred if EL4 cells contained OVA (Fig. 3B), indicating that peptide-MHC I complexes are generated in the donor cells before transfer, and that the transferred FIGURE 3. Peptide-MHC class I complexes transfer from donor cells to DC and activate CD8ϩ T cells. A, FVB BMDC were cultured alone, with complexes are presented by DC without further processing. http://www.jimmunol.org/ freeze-thawed EL4 or EL4/OVA cells, or with 10 ␮g/ml OVA, purified by passage over a Ficoll gradient, and cultured overnight with OT-I T cells. Cross-dressed DC efficiently activate CD8ϩ T cells IL-2 release was quantified by ELISA. B, FVB BMDC were cultured with freeze-thawed EL4/OVA cells or with freeze-thawed EL4 cells and 10 We reported similar transfer of peptide-MHC II complexes in a ␮g/ml OVA, purified, and cultured overnight with OT-I T cells. IL-2 re- previous study, and referred to the process as DC cross-dressing lease was quantified by ELISA. These data are representative of three (20), a term originally coined by Yewdell and Haeryfar (33). To ϩ independent experiments. determine the role of cross-dressing in CD8 T cell priming, T cells were activated under conditions in which the relative contributions of cross-priming and cross-dressing could be assessed. C57BL/6 (H-2b) BMDC were cultured with freeze-thawed DAP/ by guest on September 29, 2021 freeze-thawed untreated cells (Fig. 2C). Similarly, H-2Dd transfer OVA or DAP/OVA/Kb cells, purified by passage over Ficoll, and from IFN-␥-treated A20 cells to FVB BMDC was increased com- mixed with OT-I T cells, and T cell activation was measured by pared with non-IFN-␥-treated A20 cells (data not shown). In- assessing expression of the early activation marker CD69 (34) and creased transfer of H-2Kb from IFN-␥-treated EL4/OVA cells to by production of IL-2 and IFN-␥. BMDC exposed to freeze- splenic DC was also observed (data not shown). Therefore, in- thawed DAP/OVA/Kb cells induced a 4-fold increase in the per- creased expression of MHC class I on donor cells leads to en- centage of CD69ϩ OT-I T cells relative to BMDC exposed to hanced acquisition of MHC class I proteins by DC. freeze-thawed DAP/OVA cells (Fig. 4A), whereas OT-I T cells

FIGURE 4. Cross-dressed DC activate naive CD8ϩ T cells. C57BL/6 BMDC cultured with freeze-thawed DAP/OVA cells, DAP/OVA/Kb cells, or without ﬁbro- blasts were puriﬁed on Ficoll gradients and cultured with OT-I T cells. OT-I T cell activation was measured by CD69 expression at the indicated times (A), or by IL-2 (B) or IFN-␥ (C) production the following day. D, DAP/OVA and DAP/OVA/Kb cells have similar levels of cell-associated OVA, as determined by Western blot. These data are representative of three independent experiments. 6022 CROSS-DRESSED DC PRIME CD8ϩ T CELLS Downloaded from http://www.jimmunol.org/

FIGURE 6. Cross-dressed DC activate CD8ϩ T cells in vivo. Recipient mice were adoptively transferred with CFSE-labeled OT-I T cells and 8–24 h later inoculated s.c. with DAP transfectants. Three days later, the FIGURE 5. Cross-dressing efficiently activates CD8ϩ T cells to low draining lymph node was removed and examined for dilution of CFSE. levels of soluble Ag. C57BL/6 BMDC were cultured with 0.1 or 10 ␮g/ml Each histogram shows the result of an individual mouse. The average pro- OVA, or with freeze-thawed EL4 or EL4/OVA cells, purified on Ficoll liferation index Ϯ SD for three mice tested in independent experiments is gradients, and cultured overnight with OT-I T cells, and T cell activation shown in the top left corner of each histogram. Proliferation indices are significantly different from the unmarked indices ء was assessed by CD69 expression (A), IL-2 (B), or IFN-␥ (C) release. marked with an

These data are representative of three independent experiments. (p Ͻ 0.05) and are not significantly different from each other. A, CFSE by guest on September 29, 2021 staining for C57BL/6 mice inoculated with a 1:1 mixture of live and freeze- thawed (dead) DAP/OVA/Kb or DAP/OVA cells (left and middle panels) b Ͻ ϩ ␥ or CD11c-DTR (H-2 ) mice depleted for DC by DTx injection and inoc- cultured alone were 1% CD69 (Fig. 4A). IL-2 and IFN- pro- ulated with a 1:1 mixture of live and dead DAP/OVA/Kb cells (right duction was similarly higher in cultures with DC exposed to panel). B, CFSE staining for C57BL/6 mice inoculated with live or dead b freeze-thawed DAP/OVA/K cells (Fig. 4, B and C, respectively). DAP/OVA/Kb cells, or a 1:1 mixture of live and dead DAP/OVA/Kb cells. No IL-2 release was observed when OT-I T cells were cultured C, CFSE staining for C57BL/6 mice inoculated with a 1:1 mixture of live with freeze-thawed DAP/OVA/Kb cells in the absence of DC (Fig. and dead DAP/OVA, DAP/OVA/Kb, or NIH3T3 cells. 4B), indicating that DC are essential for CD8ϩ T cell activation and that OT-I activation is not due to direct priming by the freeze- thawed DAP/OVA/Kb cells. Comparable results were obtained if significantly activate OT-I T cells; however, DC pulsed with 0.1 splenic DC were used as recipients, although fewer T cells ex- ␮g/ml OVA do not activate (Fig. 5). Therefore, cross-dressing is pressed CD69 and both IL-2 and IFN-␥ release was lower (data not more efficient than cross-priming when Ag levels are low. shown). To determine whether the differences in T cell activation ϩ were due to quantitative differences in cell-associated OVA, cell DC cross-dressing primes CD8 T cells in vivo lysates of DAP/OVA/Kb and DAP/OVA cells were prepared and To determine whether CD8ϩ T cells are primed in vivo by cross- Western blotted with Abs to OVA. As shown in Fig. 4D, DAP/ dressed DC, C57BL/6 mice were adoptively transferred with OVA cells contain more OVA than DAP/OVA/Kb cells, indicating CFSE-labeled OT-I CD8ϩ T cells and inoculated s.c. in the lower that the enhanced activation of OT-I T cells by DC cocultured with lateral abdominal area with a 1:1 mixture of live and freeze-thawed freeze-thawed DAP/OVA/Kb cells is not due to differences in DAP/OVA/Kb cells or a 1:1 mixture of live and freeze-thawed quantity of OVA protein. DAP/OVA cells. Three days later, the draining lymph node was The absence of significant T cell activation by DAP/OVA- removed and examined for CFSE dilution. Inoculation of DAP/ loaded DC vs the strong T cell activation by DAP/OVA/Kb-loaded OVA/Kb cells resulted in strong T cell priming (Fig. 6A, left DC (Fig. 4) is consistent with the hypothesis that cross-dressing is panel). If cross-priming was occurring, then mice inoculated with more efficient than cross-priming for low levels of soluble Ag. To H-2Kb-negative DAP/OVA cells should display significant OT-I explore this possibility, C57BL/6 BMDC were pulsed for 3 h with division; however, the response to DAP/OVA is minimal (Fig. 6A, soluble OVA at concentrations comparable to OVA levels in EL4/ middle panel). OT-I expansion in response to DAP/OVA/Kb cells OVA lysates (0.1 ␮g/ml), or with 100-fold excess OVA (10 ␮g/ could be due to cross-dressing or to direct presentation of OVA by ml). The protein-pulsed DC were then compared with EL4 and DAP/OVA/Kb cells. If DAP/OVA/Kb fibroblasts are directly pre- EL4/OVA-loaded DC for their ability to activate OT-I T cells. senting OVA, then host DC would not be required for OT-I acti- EL4/OVA-loaded DC and DC pulsed with 100-fold excess OVA vation. To determine whether DC are essential, CD11c-DTR mice The Journal of Immunology 6023

(H-2b), which, when treated with DTx, are transiently depleted for priming. However, DC were present in recipient mice and there- CD11c expressing DC (35), were used. CD11c-DTR mice were fore could have been cross-dressed with peptide-MHC class I com- DTx treated, adoptively transferred with CFSE-labeled OT-I T plexes from the inoculated cells. Therefore, to definitively cells, and inoculated with a mixture of live and freeze-thawed demonstrate direct Ag presentation to CD8ϩ T cells by nonpro- DAP/OVA/Kb cells, and 3 days later the draining lymph node was fessional APCs, these studies should be done under conditions that examined for expansion of OT-I T cells. Deletion of CD11cϩ cells prevent DC cross-dressing. virtually eliminated expansion of OT-I T cells (Fig. 6A, right Although cross-priming and cross-presentation have been ex- panel), demonstrating that DC are essential for T cell priming and perimentally documented for many Ags, including soluble proteins that DAP/OVA/Kb fibroblasts cannot directly prime OT-1 T cells. and viral and tumor Ags (reviewed in Refs. 2 and 5), not all CD8ϩ In contrast, OT-I expansion in nondiphtheria-treated CD11c-DTR T cell determinants are efficiently cross-presented (13, 18, 36, 37), transgenic mice inoculated with DAP/OVA/Kb cells, and in non- suggesting that additional activation mechanisms are operating. transgenic mice inoculated with DAP/OVA/Kb and treated with Although direct presentation has been proposed as the relevant DTx was similar to expansion in Fig. 6A, left panel (data not mechanism for epitopes that are inefficiently cross-presented (14), shown). Therefore, CD11cϩ cells are required for in vivo T cell cross-dressing may be responsible. This conclusion is supported by priming, demonstrating that DAP/OVA/Kb cells do not directly the studies reported in this work, demonstrating that cross-dressed present Ag, and supporting the conclusion that cross-dressing oc- DC prime CD8ϩ T cells when cross-priming is inefficient. Al- curs in vivo and enhances T cell priming compared with cross- though the relative contributions of cross-dressing and cross-prim- priming alone. ing to overall CD8ϩ T cell activation may vary for different Ags,

Although OT-I T cells expand in mice inoculated with a mixture it is likely that a combination of the two mechanisms facilitates Downloaded from of live and freeze-thawed (dead) DAP/OVA/Kb cells, inoculation maximum CD8ϩ T cell activation. with exclusively live or freeze-thawed DAP/OVA/K cells did not induce significant OT-I T cell division (Fig. 6B), suggesting syn- Acknowledgments ergy between the live and freeze-thawed cell populations. To de- We thank Dr. Kenneth Rock (University of Massachusetts Medical School, termine which cell population was providing the peptide-MHC Worcester, MA) for the DAP/OVA cells; Dr. Mark Soloski (Johns Hopkins Medical School, Baltimore, MD) for the pKbT.8 plasmid; Dr. Charles complexes for cross-dressing, C57BL/6 mice were adoptively http://www.jimmunol.org/ Bieberich (University of Maryland Baltimore County, Baltimore, MD) for transferred with OT-I T cells and inoculated with 1:1 mixtures of b the FVB mice; Sandy Mason for animal care; and Virginia Clements for live and freeze-thawed (dead) DAP/OVA, DAP/OVA/K , and outstanding technical assistance. NIH3T3 cells. OT-I T cells expanded in mice inoculated with a mixture of live DAP/OVA and dead DAP/OVA/Kb cells (Fig. 6C, Disclosures left panel), but not in mice inoculated with mixtures of live DAP/ The authors have no financial conflict of interest. OVA/Kb plus dead DAP/OVA cells, or mixtures of live NIH3T3 plus dead DAP/OVA/Kb cells (Fig. 6C, middle and right panels). References Therefore, T cell expansion requires live and dead cells, and the 1. Zinkernagel, R. M. 2002. On cross-priming of MHC class I-specific CTL: rule or

exception? Eur. J. Immunol. 32: 2385–2392. by guest on September 29, 2021 dead cells must contain the relevant peptide-MHC class I com- 2. Rock, K. L., and L. Shen. 2005. Cross-presentation: underlying mechanisms and plexes, whereas the live cells must contain the relevant Ag. As DC role in immune surveillance. Immunol. Rev. 207: 166–183. acquire peptide-MHC class I complexes from dead and not live 3. Bevan, M. J. 1976. Cross-priming for a secondary cytotoxic response to minor H antigens with H-2 congenic cells which do not cross-react in the cytotoxic assay. cells (Figs. 1 and 3), the dead cell population is probably inducing J. Exp. Med. 143: 1283–128. DC cross-dressing, whereas the live population is inducing cross- 4. Ackerman, A. L., and P. Cresswell. 2004. Cellular mechanisms governing cross- ϩ presentation of exogenous antigens. Nat. Immunol. 5: 678–684. priming. Therefore, maximum priming of CD8 T cells is prob- 5. Heath, W. R., and F. R. Carbone. 2001. Cross-presentation, dendritic cells, tol- ably the result of a combination of DC cross-dressing and erance and immunity. Annu. Rev. Immunol. 19: 47–64. cross-priming. 6. Kundig, T. M., M. F. Bachmann, C. DiPaolo, J. J. Simard, M. Battegay, H. Lother, A. Gessner, K. Kuhlcke, P. S. Ohashi, H. Hengartner, et al. 1995. Fibroblasts as efficient antigen-presenting cells in lymphoid organs. Science 268: Discussion 1343–1347. The priming of CD8ϩ T cells by cross-dressed DC is a newly 7. Cayeux, S., G. Richter, C. Becker, A. Pezzutto, B. Dorken, and T. Blankenstein. ϩ 1999. Direct and indirect T cell priming by dendritic cell vaccines. Eur. J. Im- identified mechanism by which CD8 T cells are activated. Al- munol. 29: 225–234. though there has been speculation that cross-dressing occurs (33), 8. Sigal, L. J., and K. L. Rock. 2000. Bone marrow-derived antigen-presenting cells are required for the generation of cytotoxic T lymphocyte responses to viruses this is the first report demonstrating that it takes place in vitro and and use transporter associated with antigen presentation (TAP)-dependent and in vivo. Cross-priming and cross-presentation, the generally ac- -independent pathways of antigen presentation. J. Exp. Med. 192: 1143–1150. cepted mechanisms for activating CD8ϩ T cells, differ signifi- 9. Sigal, L. J., S. Crotty, R. Andino, and K. L. Rock. 1999. Cytotoxic T-cell immunity to virus-infected non-haematopoietic cells requires presentation of exog- cantly from cross-dressing. Cross-dressed DC acquire peptide- enous antigen. Nature 398: 77–80. MHC class I complexes from exogenous sources and present the 10. Belz, G. T., C. M. Smith, D. Eichner, K. Shortman, G. Karupiah, F. R. Carbone, ␣ϩ complexes without further processing. Although DC also acquire and W. R. Heath. 2004. Cutting edge: conventional CD8 dendritic cells are generally involved in priming CTL immunity to viruses. J. Immunol. 172: Ag from exogenous sources during cross-priming/presentation, 1996–2000. they process the acquired Ag intracellularly and present it in the 11. Smith, C. M., G. T. Belz, N. S. Wilson, J. A. Villadangos, K. Shortman, ␣ϩ context of their MHC class I molecules. Therefore, cross-priming F. R. Carbone, and W. R. Heath. 2003. Cutting edge: conventional CD8 den- ϩ dritic cells are preferentially involved in CTL priming after footpad infection with activates CD8 T cells that are restricted to the MHC genotype of herpes simplex virus-1. J. Immunol. 170: 4437–4440. 12. Norbury, C. C., D. Malide, J. S. Gibbs, J. R. Bennink, and J. W. Yewdell. 2002. the DC and specific for peptides generated by the DC, whereas ϩ ϩ Visualizing priming of virus-specific CD8 T cells by infected dendritic cells in cross-dressing activates CD8 T cells that are restricted to the vivo. Nat. Immunol. 3: 265–271. MHC genotype of the donor cells and are specific for peptides 13. Freigang, S., D. Egger, K. Bienz, H. Hengartner, and R. M. Zinkernagel. 2003. generated by the donor cells. Endogenous neosynthesis vs. cross-presentation of viral antigens for cytotoxic T cell priming. Proc. Natl. Acad. Sci. USA 100: 13477–13482. Several reports suggest that cells other than DC, such as tumor 14. Norbury, C. C., and L. J. Sigal. 2003. Cross priming or direct priming: is that cells or fibroblasts, can directly activate CD8ϩ T cells (6, 18, 19). really the question? Curr. Opin. Immunol. 15: 82–88. 15. Huang, A. Y., P. Golumbek, M. Ahmadzadeh, E. Jaffee, D. Pardoll, and These studies used mice inoculated with genetically mismatched H. Levitsky. 1994. Role of bone marrow-derived cells in presenting MHC class cells or recipient mice deficient in TAP so as to prevent cross- I-restricted tumor antigens. Science 264: 961–965. 6024 CROSS-DRESSED DC PRIME CD8ϩ T CELLS

16. Huang, A. Y., A. T. Bruce, D. M. Pardoll, and H. I. Levitsky. 1996. In vivo 27. Morris, E. J., and H. M. Geller. 1996. Induction of neuronal apoptosis by camp- cross-priming of MHC class I-restricted antigens requires the TAP transporter. tothecin, an inhibitor of DNA topoisomerase-I: evidence for cell cycle-indepen- Immunity 4: 349–355. dent toxicity. J. Cell Biol. 134: 757–770. 28. Ilkovitch, D., and S. Ostrand-Rosenberg. 2004. MHC class II and CD80 tumor 17. Thomas, A. M., L. M. Santarsiero, E. R. Lutz, T. D. Armstrong, Y. C. Chen, ϩ L. Q. Huang, D. A. Laheru, M. Goggins, R. H. Hruban, and E. M. Jaffee. 2004. cell-based vaccines are potent activators of type 1 CD4 T lymphocytes provided Mesothelin-specific CD8ϩ T cell responses provide evidence of in vivo cross- they do not coexpress invariant chain. Cancer Immunol. Immunother. 53: priming by antigen-presenting cells in vaccinated pancreatic cancer patients. 525–532. J. Exp. Med. 200: 297–306. 29. Liu, Z., Q. Liu, H. Hamed, R. M. Anthony, A. Foster, F. D. Finkelman, 18. Ochsenbein, A. F., S. Sierro, B. Odermatt, M. Pericin, U. Karrer, J. Hermans, J. F. Urban, Jr., and W. C. Gause. 2005. IL-2 and autocrine IL-4 drive the in vivo S. Hemmi, H. Hengartner, and R. M. Zinkernagel. 2001. Roles of tumor local- development of antigen-specific Th2 T cells elicited by nematode parasites. J. Im- ization, second signals and cross priming in cytotoxic T-cell induction. Nature munol. 174: 2242–229. 411: 1058–1064. 30. Hsu, D. H., P. Paz, G. Villaflor, A. Rivas, A. Mehta-Damani, E. Angevin, L. Zitvogel, and J. B. Le Pecq. 2003. Exosomes as a tumor vaccine: enhancing 19. Wolkers, M. C., G. Stoetter, F. A. Vyth-Dreese, and T. N. Schumacher. 2001. ϩ potency through direct loading of antigenic peptides. J. Immunother. 26: Redundancy of direct priming and cross-priming in tumor-specific CD8 T cell 440–450. responses. J. Immunol. 167: 3577–3584. 31. Chaput, N., N. E. Schartz, F. Andre, J. Taieb, S. Novault, P. Bonnaventure, 20. Dolan, B. P., K. D. Gibbs, Jr., and S. Ostrand-Rosenberg. 2006. Tumor-specific ϩ N. Aubert, J. Bernard, F. Lemonnier, M. Merad, et al. 2004. Exosomes as potent CD4 T cells are activated by “cross-dressed” dendritic cells presenting peptide- cell-free peptide-based vaccine. II. Exosomes in CpG adjuvants efficiently prime MHC class II complexes acquired from cell-based cancer vaccines. J. Immunol. naive Tc1 lymphocytes leading to tumor rejection. J. Immunol. 172: 2137–2146. 176: 1447–1455. 32. Andre, F., N. Chaput, N. E. Schartz, C. Flament, N. Aubert, J. Bernard, ␤ Ϫ/Ϫ 21. Lamouse-Smith, E., V. K. Clements, and S. Ostrand-Rosenberg. 1993. 2m F. Lemonnier, G. Raposo, B. Escudier, D. H. Hsu, et al. 2004. Exosomes as knockout mice contain low levels of CD8ϩ cytotoxic T lymphocyte that mediate potent cell-free peptide-based vaccine. I. Dendritic cell-derived exosomes trans- specific tumor rejection. J. Immunol. 151: 6283–6290. fer functional MHC class I/peptide complexes to dendritic cells. J. Immunol. 172: 22. Qi, L., and S. Ostrand-Rosenberg. 2000. MHC class II presentation of endoge- 2126–2136. nous tumor antigen by cellular vaccines depends on the endocytic pathway but 33. Yewdell, J. W., and S. M. Haeryfar. 2005. Understanding presentation of viral

ϩ Downloaded from not H2-M. Traffic 1: 152–160. antigens to CD8 T cells in vivo: the key to rational vaccine design. Annu. Rev. 23. Shen, L., and K. L. Rock. 2004. Cellular protein is the source of cross-priming Immunol. 23: 651–682. antigen in vivo. Proc. Natl. Acad. Sci. USA 101: 3035–3040. 34. Norbury, C. C., S. Basta, K. B. Donohue, D. C. Tscharke, M. F. Princiotta, P. Berglund, J. Gibbs, J. R. Bennink, and J. W. Yewdell. 2004. CD8ϩ T cell 24. Blok, R., D. H. Margulies, L. Pease, R. K. Ribaudo, J. Schneck, and cross-priming via transfer of proteasome substrates. Science 304: 1318–1321. J. McCluskey. 1992. CD8 expression alters the fine specificity of an alloreactive 35. Jung, S., D. Unutmaz, P. Wong, G. Sano, K. De los Santos, T. Sparwasser, S. Wu, MHC class I-specific T hybridoma. Int. Immunol. 4: 455–466. S. Vuthoori, K. Ko, F. Zavala, et al. 2002. In vivo depletion of CD11cϩ dendritic 25. Catipovic, B., J. Dal Porto, M. Mage, T. E. Johansen, and J. P. Schneck. 1992. cells abrogates priming of CD8ϩ T cells by exogenous cell-associated antigens. Major histocompatibility complex conformational epitopes are peptide specific. Immunity 17: 211–220. http://www.jimmunol.org/ J. Exp. Med. 176: 1611–168. 36. Otahal, P., S. C. Hutchinson, L. M. Mylin, M. J. Tevethia, S. S. Tevethia, and 26. Dolan, B. P., T. P. Phelan, D. Ilkovitch, L. Qi, W. F. Wade, T. M. Laufer, and T. D. Schell. 2005. Inefficient cross-presentation limits the CD8ϩ T cell response S. Ostrand-Rosenberg. 2004. Invariant chain and the MHC class II cytoplasmic to a subdominant tumor antigen epitope. J. Immunol. 175: 700–712. domains regulate localization of MHC class II molecules to lipid rafts in tumor 37. Wolkers, M. C., N. Brouwenstijn, A. H. Bakker, M. Toebes, and T. N. Schumacher. cell-based vaccines. J. Immunol. 172: 907–914. 2004. Antigen bias in T cell cross-priming. Science 304: 1314–1317. by guest on September 29, 2021