CpG-DNA Aided Cross-Priming by Cross-Presenting B Cells

The Journal of Immunology

CpG-DNA Aided Cross-Priming by Cross-Presenting B Cells1 Antje Heit, Katharina M. Huster, Frank Schmitz, Matthias Schiemann, Dirk H. Busch, and Hermann Wagner2 Covalent linkage of immunostimulatory CpG-DNA to OVA (CpG-OVA complex) results in CpG-DNA-aided cross-presentation of OVA by dendritic cells (DCs). In this study, we analyzed the thesis that CpG-OVA complexes may be cross-presented by B cells to route internalized Ag into the class I MHC presentation pathway. First, we describe that conjugation of CpG-DNA to OVA enhances up to 40-fold internalization of OVA by B cells, which in turn generate the CD8 T cell epitope SIINFEKL complexed to MHC class I, albeit less efficiently than DCs. Furthermore, upon internalization, CpG-DNA conjugated to OVA stimulates B cells to up-regulate costimulatory molecules and cytokines including IL-12. Adoptive transfer of CpG-OVA complex-loaded wild-type B cells cross-primes naive CD8 T cells both in wild-type mice and in MyD88-deficient mice. Overall, these findings disclose attributes of B cells, including cross-presentation of exogenous Ag and cross-priming of naive CD8 T cells that hitherto have been considered as hallmarks restricted to DCs. The Journal of Immunology, 2004, 172: 1501–1507.

T

he specificity, amplitude, and character of T cell-mediated immune responses become conditioned during the early phase of Ag presentation. Within the population of APCs that include dendritic cells (DCs),3 B cells, and macrophages, DCs display the unique competence to activate naive T cells (1). In general, CD8 T cells recognize peptides derived from cytosolic Ag complexed to MHC class I molecules, while CD4 T cells recognize peptides from internalized Ag complexed to MHC class II. However, when DCs internalize high amounts of Ag, for example via receptor-mediated endocytosis (2, 3), they are able to route internalized Ag into the MHC class I pathway, a process termed crosspresentation (4, 5). Cross-presentation of Ag by DCs via MHC class I can either lead to cross-priming of naive CD8 T cells (4, 6) or to cross-tolerance (7). In contrast to DCs, the role of B cells in T cell priming has been controversial. For example, elegant experiments implied that both resting and activated H-Y-presenting B cells induce T cell tolerance, yet memory T cells become reactivated (8, 9). In contrast, Ag specifically bound to surface Ig (sIg) on B cells was shown to induce CD4-T cell priming (10, 11), presumably due to efficient Ig-mediated endocytosis of Ag. Furthermore, engagement of Ig receptors by specific Ag was found to activate B cells and thus trigger up-regulation of costimulatory molecules on B cells both in vitro (12) and in vivo (13). Thus, at least two criteria required for the activation of naive CD4 T cells, e.g., TCR ligation and costimulation, appear to be met by Ag-specific B cells. However, little is known as to whether B cells can cross-present Ag via MHC

class I to naive CD8⫹ T cells, and whether cross-presentation leads to cross-priming. Covalent linkage of immunostimulatory CpG-DNA to proteinaceous Ag such as OVA results in DNA receptor-mediated endocytosis of Ag, Ag cross-presentation, and cross-priming of CD8 (14, 15) and CD4 T cells (16) by DCs. In this system, the function of Toll-like receptor 9 (TLR9) is restricted to the activation of immature DCs into professional APCs, while cross-presenting OVA. However, TLR9 is not involved in CpG-DNA-aided cellular uptake (15, 17). Murine as well as subsets of human B cells express TLR9 and thus are sensitive to immunostimulatory CpG-DNA (18). Because TLR9 is not detectable at the cell surface, but at cytoplasmatic endosome-like organelles (19), we argued that CpG-DNA needs to be endocytosed to activate B cells via TLR9. It follows that Ag loading of B cells (20) via DNA receptor-mediated endocytosis with OVA covalently linked to CpG-DNA (CpG-DNA-OVA complex) may be possible and unravel whether B cells are able to route internalized OVA into the MHC class I pathway. In this study, we describe the ability of B cells to cross-present OVA and to crossprime CD8⫹ T cells.

Materials and Methods C57BL/6 mice were purchased from Harlan Winkelmann GmbH (Borchen, Germany). TLR9⫺/⫺ and MyD88⫺/⫺ mice were a kind gift from S. Akira (Osaka, Japan); OT-1 mice were kindly provided by T. Brocker (Ludwig Maximilian University, Munich, Germany). All animals were kept under specific pathogen-free conditions and were used at 8 –12 wk of age.

Cell lines and in vitro culture medium Institute of Medical Microbiology, Immunology and Hygiene, Technical University Munich, Munich, Germany Received for publication April 1, 2003. Accepted for publication November 24, 2003. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the Deutsche Forschungsgemeinschaft, Forschungsverbund Grundlagen Gentechnischer Verfahren, and Coley Pharmaceutical Group. 2 Address correspondence and reprint requests to Dr. Hermann Wagner, Institute of Medical Microbiology, Immunology and Hygiene, Trogerstr. 9, 81675 Munich, Germany. E-mail address: [email protected]

EL-4 (H-2b) thymoma cells were purchased from the American Type Culture Collection (Manassas, VA). B3Z, a somatic T cell hybrid generated by fusing the OVA/Kb-specific cytotoxic clone, B3, with a lacZ-inducible derivative of BW5147 fusion partner (21), was kindly provided by B. L. Kelsall (National Institutes of Health, Bethesda, MD). Cells were cultured in RPMI 1640 supplemented with 10% (v/v) heat-inactivated FBS, 100 IU/ml penicillin G, 100 IU/ml streptomycin sulfate (all: Biochrom KG, Berlin, Germany), and 50 ␮M of 2-ME (Invitrogen, Karlsruhe, Germany) at 37°C/5% CO2.

Reagents

3

Abbreviations used in this paper: DC, dendritic cell; LN, lymph node; S-MBS, sulfo-maleimidobenzoyl-N-hydroxysuccinimide ester; sIg, surface Ig; TLR, Toll-like receptor; wt, wild type. Copyright © 2004 by The American Association of Immunologists, Inc.

Chicken egg albumin (OVA) was obtained from Sigma-Aldrich (Seelze, Germany). Its dominant CD8 T cell epitope, the peptide SIINFEKL (OVA 0022-1767/04/$02.00

1502 peptide 257–264), was custom synthesized by Research Genetics (Huntsville, AL). FITC-labeled OVA was purchased from Molecular Probes (Leiden, The Netherlands). Phosphothioate-modified immunostimulatory CpG-ODN were custom synthesized by MWG Biotec (Ebersberg, Germany). The phosphothioated sulfhydryl-modified ODN (TriLink Biotechnologies, La Jolla, CA) used throughout this study consisted of 20 bases and contained a CpG motif (1668: 5⬘-S-TCCATGACGTTCCTGATGCT-3⬘); as control, the 20-mer non-CpG motif (1720: 5⬘-S-TCCATGAGCTTCCTGATGCT-3⬘) was used. The ODNs were conjugated to OVA by incubation with the crosslinker sulfo-maleimidobenzoyl-N-hydroxysuccinimide ester (S-MBS; Pierce, Bonn, Germany) in a 50 mM EDTA-PBS buffer, pH 7.0, at a molar ratio of 1:10 for 1 h at room temperature. The sulfhydryl-modified ODN were reduced in a 50 mM 1,4-DTT-PBS solution. Subsequently, unbound S-MBS and 1,4-DTT were removed by chromatography on a Biorade P-6 gel column (Bio-Rad, Muenchen, Germany). Then the activated ODN were incubated with the linker-modified OVA at a molar ratio of 5:1 for 2.5 h at room temperature, and thereafter L-cysteine was added to quench reactive S-MBS. Free ODN were removed by FPLC on a Superdex 75HR column (Amersham Biosciences, Freiburg, Germany). Purified conjugates were analyzed on a 6 –20% gradient SDS-PAGE and consecutively silver stained. To determine ratio of bound ODN on OVA, a 4 –15% gradient nondenaturing, nonreducing PAGE was run and visualized using ethidium bromide staining. Protein concentration was determined by the Lowry method (Pierce), as described previously (14). The batches of all CpG-OVA conjugates used in this study had a ratio of 2.5 CpG-DNA molecules linked to 1 OVA molecule. That means that 50 ␮g of used 1668-OVA conjugate contained 50 ␮g of OVA linked to 2.85 nmol of 1668 CpG-DNA. The FITC-labeled conjugates were synthesized similarly and contained the same ratio of OVA-FITC to ODN. Escherichia coli-derived LPS was purchased from Sigma-Aldrich.

Generation of GM-CSF-cultured DC from bone marrow Generation of GM-CSF-induced, bone marrow-derived DC cultures was performed, as previously described (14). Bone marrow cells were flushed out of femurs and tibiae, centrifuged, and after washing, cells were plated at 5 ⫻ 106 cells/10 ml in medium complemented with 18 ng/ml GM-CSF and cultured at 37°C/5% CO2. Cells were used after 6 days of culture.

Preparation of purified B cells For preparation of highly purified, negatively selected, and thus untouched murine B cells, the MACS CD43 MicroBead system (Miltenyi Biotec, Bergisch Gladbach, Germany) was used. To this, spleens were minced, and after lysis of RBC, the cell suspensions were incubated with anti-CD43 (Ly-48) Microbeads for 30 min on ice. After washing, the cells were negatively selected by running through LS columns (Miltenyi Biotec). Negative selected cells were collected, and purity of each batch prepared was analyzed by FACS. Cells were used when no CD11c-positive cells and ⬎96% B cells could be scored. To further rule out a role of potentially contaminating DCs, the MACS CD43 Microbead system negatively selected B cells were first stained with APC-labeled anti-CD11c (clone HL3) and PE-labeled anti-CD19 (clone 1D3), and stained cells were sorted for cells staining only with anti-CD19 PE using the high speed MoFlo cytometer (DAKO Cytomation, Hamburg, Germany). After sorting, the purity of B cells was ⬎99%.

Preparation of CD8 T cells To enrich CD8 T cells from LNs or spleen, the organs were minced and the cell suspension was incubated with PE-labeled Abs against CD4, MHC class II, and CD19 for 30 min on ice. Cells were then washed and incubated with anti-PE-Microbeads (Miltenyi Biotec) for additional 20 min. After washing, the labeled cells were negatively selected via a LS column (Miltenyi Biotec) and were used as CD8-positive cell fraction.

Listeria monocytogenes infection Six- to 8-wk-old C57BL/6 mice were i.v. injected in the tail veins with recombinant L. monocytogenes-secreting OVA (L. monocytogenes-OVA, kindly provided by H. Shen (University of Pennsylvania School of Medicine, Philadelphia, PA) (22). L. monocytogenes-OVA were grown in brain heart medium. For primary infection, a sublethal dose of ⬃0.1 ⫻ LD50 (2000 bacteria for C65BL/6 mice) was injected. Reimmunization was performed by i.v. injection of 5 ⫻ LD50 bacteria. Mice were used 5 days after second infection.

B CELL-MEDIATED CROSS-PRIMING Cross-presenting B cells; immunization of mice; chromium release assay Highly purified B cells (1 ⫻ 107) were incubated with 50 ␮g/ml of 1668OVA conjugate or OVA for 90 min. B cells were washed twice, and 2 ⫻ 107 cells were injected into either both hind footpads (s.c.) or the tail vein (i.v.) of 6- to 12-wk-old C57BL/6 mice. Draining LN and spleen were removed 7 days later, and single-cell suspensions were prepared. LN cells (3 ⫻ 106 cells) or spleen cells (5 ⫻ 106 cells) were cultured for additional 7 days in medium conditioned with 5 IU/ml rIL-2 and syngenic, irradiated, SIINFEKL (1 ␮M) pulsed or unpulsed (2 ⫻ 106 cells) spleenic feeder cells. Cytolytic activity was assayed via 51Cr release assay, essentially as described (14). SIINFEKL peptide-untreated EL-4 cells served as specificity control. Specific lysis was calculated according to the formula: percent specific lysis ⫽ (cpm (sample) – cpm(spontaneous release)/(maximum release) – cpm (spontaneous release)) ⫻ 100.

OVA uptake, B cell activation analysis, and mAbs used To examine uptake of FITC-labeled CpG-OVA in vitro, purified B cells were exposed to FITC-labeled OVA (10 ␮g/ml), mixed with1668 CpGDNA (10 ␮g/ml) or 1668-OVA-FITC conjugates (10 ␮g/ml) or with medium (30 min at 37°C), washed twice with ice-cold 3% FCS-PBS, and stained with APC-labeled anti-CD11c (clone HL3) and PE-labeled antiCD19 (clone 1D3). Analysis was performed in a FACSCalibur flow cytometer (BD Biosciences, Heidelberg, Germany) (50,000 events/sample). To analyze B cell activation by CpG-OVA complexes, purified B cells were incubated with 50 ␮g/ml OVA conjugated with 2.85 nmol of CpGODN. Cells were cultured for 24 h, and thereafter washed twice and stained with APC-labeled anti-CD11c (clone HL3), PE-labeled anti-CD19 (clone 1D3), FITC-labeled anti-CD40 (clone 3/23), and anti-CD86 (clone GL1). FACS analysis was performed using a FACSCalibur flow cytometer (BD Biosciences) acquiring at least 50,000 events per sample. FACS data were analyzed using CellQuest software (BD Biosciences). mAbs and corresponding isotype controls were purchased from BD Biosciences.

Cross-presentation assay Cross-presentation of SIINFEKL after CpG-DNA-aided uptake of OVA by purified B cells was assayed, as previously described (14, 15), by measuring induction of lacZ activity in SIINFEKL/Kb-specific T cell hybridome B3Z (21). To this, 1 ⫻ 105 purified B cells were incubated with the indicated reagents for 18 h at 37°C, followed by washing twice, and transferred into 96-well plates in triplets. B3Z cells (1 ⫻ 105) were added to each well. After additional incubation at 37°C overnight, the cells were fixed with 0.5% glutaraldehyde for 10 min and incubated with X-Gal solution (23) at 37°C. After 6 – 8 h, blue B3Z cells were counted under the microscope.

B and T cell proliferation assay Purified B cells (3 ⫻ 106 cells/ml), CpG-OVA conjugate (50 ␮g/ml), or OVA alone (50 ␮g/ml) loaded for 3 h were labeled with 5 mM CFSE for 10 min at 37°C. To stop the CSFE reaction, the cells were taken up in 10 ml of medium containing FCS for 5 min on ice. Then CSFE-labeled B cells were incubated in a 12-well plate at 37°C for 3 days. After washing twice, the cells were analyzed. For CD19-positive B cells, additionally stained with anti-B220 APC Ab (clone RA3-6B2; BD Biosciences), proliferation was analyzed using a FACSCalibur flow cytometer (BD Biosciences). T cell proliferation was assayed using OVA (50 ␮g/ml) or CpG-OVA conjugate (50 ␮g/ml)-loaded, irradiated (3000 rad), purified B cells as stimulator cells via [3H]thymidine incorporation. To this, purified B cells were loaded with the relevant reagents for 1.5 h at 37°C, washed twice, and coincubated at a defined responder to stimulator ratio in 96-well roundbottom plates in triplets with murine CD8 T cells (5 ⫻ 104 cells/well) from either OT-1 (expressing a SIINFEKL-specific transgenic TCR) mice (responder to stimulator ratio 1:1) or L. monocytogenes (OVA-secreting)primed mice (responder to stimulator ratio 1:4). After 48 h of incubation, 1 mCi [3H]thymidine was added to each well, followed by an 18-h incubation at 37°C. The assay was harvested using a Molecular Devices (Sunnyvale, CA) Micro 96 Harvester and analyzed with a gamma counter (Matrix 9600; Packard Instrument, Meriden, CT).

Ex vivo tetramer staining of primed SIINFEKL-specific, CD8positive, CD62L low T cells A total of 2 ⫻ 107 purified B cells was loaded for 1.5 h with either CpGOVA conjugate (50 ␮g/ml) or OVA alone (50 ␮g/ml), washed twice, and injected either s.c. (hind footpad) or i.v. (tail vein) into C57BL/6 mice. Seven days later, the LN or spleen was analyzed for presence of SIINFEKL tetramer binding, and thus peptide-specific CD8-positive, CD62L low T

The Journal of Immunology

1503

cells, as described (24). In short, whole cell suspensions were depleted of RBC, followed by a triple staining with anti-CD8 APC (clone CD8␣; Caltag, Hamburg, Germany), anti-CD62L FITC (clone MEL-14), and MHC SIINFEKL tetramer PE (H-2Kb/SIINFEKL (257–264/murine ␤2-microglobulin streptavidin-PE)) for 1 h at 4°C. Additionally, an Fc block (CD16/CD32, 2.4G2; BD Biosciences) was used to avoid unspecific Ab binding. For life/death cell discrimination, cells were subsequently incubated with ethidium monoazide bromid (Molecular Probes, Eugene, OR). At least 1 ⫻ 105 CD8-positive events were acquired on a FACSCalibur flow cytometer (BD Biosciences) and analyzed with FlowJo software (Tree Star, Ashland, OR)

Results CpG-OVA-loaded B cells become activated and cross-present OVA Previously, we have shown (14, 15) that immunostimulatory CpGDNA motifs linked to protein (OVA) fulfill a dual role when exposed to DCs: first, CpG-DNA-aided, but TLR9-independent enhancement of OVA uptake, coupled with efficient Ag crosspresentation; second, TLR9-dependent activation of DCs into professional APCs, able to cross-prime naive CD8 T cells. Because murine B cells express TLR9 (25), and thus are CpG-DNA responsive, we first analyzed whether B cells exposed to CpGOVA complexes display CpG-DNA-aided enhancement of OVA uptake. To this we conjugated CpG-DNA to FITC-labeled OVA (CpG-OVA-FITC conjugate) and exposed highly purified B cells to either CpG-OVA conjugate, a mixture of CpG-DNA plus FITCmarked OVA, or FITC-OVA alone. As shown in Fig. 1, purified B cells exposed to either OVA-FITC (10 ␮g/ml) or OVA-FITC (10 ␮g/ml) mixed with stimulatory CpG-DNA yielded in ⬃1% FITCpositive B cells. In contrast, CpG-OVA conjugates (10 ␮g/ml) enhanced OVA uptake by B cells ⬎40-fold. Next, we also com-

FIGURE 1. Uptake of CpG-OVA conjugates by B cells. Purified B cells were exposed for 30 min with 10 ␮g/ml FITC-linked OVA (OVA-FITC), a mixture of 10 ␮g/ml OVA-FITC and 0.057 nmol of CpG-DNA 1668 (1668 ⫹ OVA), 10 ␮g/ml CpG-OVA-FITC conjugate (1668-OVA-FITC), or medium alone (medium) at 37°C. Cells were stained with anti-mouse CD19-PE and anti-mouse CD11c-APC. CD11c-negative, CD19-positive cells were gated. Representative data of three independent experiments are shown.

FIGURE 2. Comparison of the loading efficacy of B cells and DCs. Highly purified B cells (a) and bone marrow-derived DCs (b) were exposed for 30 min in vitro to graded doses of FITC-labeled CpG-OVA conjugates, washed three times, and thereafter analyzed by FACS. The data are representative for two experiments.

pared the loading efficacy of B cells with that of bone marrowderived DCs. When exposed to graded concentrations of FITClabeled CpG-OVA conjugates, DCs turned out to be 10- to 20-fold more efficient to internalize CpG-OVA conjugates (Fig. 2). Subsequently, we addressed the question as to whether CpGDNA-aided loading of B cells with OVA yields in OVA crosspresentation, that is processing of OVA and presentation of the dominant CD8 T cell peptide SIINFEKL complexed to MHC class I. As shown in Fig. 3, CpG-OVA conjugate-loaded B cells were able to process and to cross-present OVA as determined by activation of B3Z hybridoma cells (expressing a TCR specific for the Kb-restricted OVA-derived CD8 T cell epitope SIINFEKL). However, the number of B3Z hybridoma cells responding to CpG-OVA conjugate-loaded B cells was low (Fig. 3) compared with loaded

FIGURE 3. Presentation of the immundominant MHC class I peptide epitope SIINFEKL by B cells. Purified (B220-positive/CD11c-negative) B cells were incubated with either 10 ␮g/ml OVA alone (OVA), a mixture of 10 ␮g/ml OVA and the CpG-DNA 1668 (0.57 nmol) (1668 ⫹ OVA), or the 10 ␮g/ml CpG-OVA conjugate (1668-OVA) for 18 h, washed, and coincubated with B3Z hybridoma cells overnight. Thereafter, the cells were fixed with 0.5% glutaraldehyde (10 min) and inoculated with X-Gal solution at 37°C. After 4 – 8 h, blue B3Z cells were counted under the microscope. X-Gal staining revealed TCR-activated B3Z cells. The mean of activated cells (three wells) is shown.

1504

B CELL-MEDIATED CROSS-PRIMING

DCs (14). These results raise the question as to whether the reduced loading efficacy of B cells to that of DCs (Fig. 2) directly translates in decreased SIINFEKL epitope presentation. We also explored the issue as to whether immunostimulatory CpG-DNA linked to OVA is still effective in activating crosspresenting B cells. As shown in Fig. 4b, similar to DCs (14, 15), purified B cells also up-regulated CD86 and CD40 upon loading with 50 ␮g/ml immunostimulatory CpG-OVA conjugates, yet failed to do so when loaded with nonstimulatory (control) GpCOVA conjugates (1720-OVA conjugate, 50 ␮g/ml). Of note, nonstimulatory GpC-OVA complexes (14, 15) were as effective in loading B cells (data not shown). Furthermore, B cell loading with an immunostimulatory CpG-OVA conjugate-initiated B cell proliferation (Fig. 4a) and activated B cells produced IL-10, IL-6, and IL-12 p40 (data not shown). We conclude that: 1) immunostimulatory CpG-OVA conjugates are effectively taken up by B cells, albeit less efficiently compared with DCs; 2) CpG-OVA conjugateloaded B cells can process OVA and can generate the MHC class I-restricted CD8 T cell epitope SIINFEKL; and 3) endocytosed linked CpG-DNA still activates B cells in terms of up-regulation of costimulatory molecules and of cytokine production. Based on these findings, we investigated the question as to whether CpG-OVA conjugate-loaded B cells are able to induce peptide-specific CD8 T cell proliferation. To this, irradiated, highly purified B cells exposed to 50 ␮g/ml CpG-OVA complexes (90 min) were incubated together with SIINFEKL-specific T cells. These loaded B cells were able to stimulate OT-1 (SIINFEKLspecific TCR-transgenic) mouse-derived CD8 T cells (Fig. 5a) as well as SIINFEKL-specific CD8 T cells from a wild-type (wt) mouse previously infected with OVA-secreting L. monocytogenes (Fig. 5b). Stimulation of OT-1 cells with B cells loaded with CpG ⫹ OVA did not lead to significant proliferation (Fig. 5a, right). FIGURE 5. Loading of B cells with CpG-OVA conjugate induces proliferation of Ag-specific T cells. Purified TCR transgenic CD8 T cells from OT-1 mice (a) were coincubated for 48 h with purified B cells previously exposed (30 min) to either OVA (50 ␮g), a CpG-OVA conjugate (left) (50 ␮g), or a mixture of CpG-DNA plus OVA (50 ␮g) (right). b, CD8 T cells from L. monocytogenes-OVA-primed C57BL/6 mice were incubated with either OVA or 1668-OVA conjugate (50 ␮g)-loaded, purified B cells. [3H]Thymidine incorporation data are shown. c, Details that MyD88-deficient B cells exposed to CpG-OVA conjugates also trigger proliferation of TCR transgenic CD8 T cells from OT-1 mice.

Interestingly, MyD88-deficient B cells also loaded with CpG-OVA complex (50 ␮g/ml) were able to trigger proliferation of OT-1 CD8 T cells (Fig. 5c), implying that proliferation depends on cross-presentation of the SIINFEKL T cell epitope. CpG-OVA conjugate-loaded B cells cross-prime CD8 T cells

FIGURE 4. Proliferation and activation of B cells by CpG-OVA conjugate. a, Purified B cells were loaded with (50 ␮g/ml) CpG-OVA conjugate (1668-OVA) (solid line) or (50 ␮g/ml) OVA (dotted line) alone for 2 h, followed by a CFSE (5 mM) treatment. After 3 days, CSFE staining decreased with every round of cell division. b, Staining of B cells reveals enhanced expression of CD40 and CD86 upon activation with (50 ␮g/ml) CpG-OVA conjugates (bold line) (24 h) in contrast to B cells incubated with (50 ␮g/ml) 1720-OVA conjugate (dotted line) or medium (filled) alone.

Because B cells loaded with CpG-OVA conjugates could generate the CD8 T cell epitope SIINFEKL complexed to MHC class I and in parallel became activated to display costimulatory molecules and to produce cytokines, we measured whether they can function as APCs able to cross-prime CD8 T cells. To this, highly purified (⬎99%) B cells from C57BL/6 mice devoid of DCs, as judged by FACS analysis, were first loaded with CpG-OVA conjugates. These loaded B cells (1 ⫻ 107) were injected s.c. into C57BL/6 recipients. After 7 days, the draining lymph nodes (LN) and spleen were harvested, and dissociated cells were restimulated in vitro. As shown in Fig. 6a, both the draining LN as well as the spleen of challenged mice contained primed cells that upon restimulation differentiated into SIINFEKL-specific CTL. These results in turn raised the question as to whether, upon adoptive transfer, B cells

The Journal of Immunology

1505 cells ought not to be cross-primed because TLR9-initiated signaling is defective in MyD88 knockout mice (27–29). As shown in Fig. 6b, adoptive transfer of CpG-OVA complex-loaded wt B cells (1 ⫻ 107) in MyD88-deficient mice primed OVA-specific T cells as effectively as in wt mice. These data excluded re-presentation by host DCs. Finally, we used SIINFEKL tetramers to analyze the frequency of SIINFEKL-specific CD8 T cells directly ex vivo in draining LNs or spleen of C56BL/6 mice challenged for 7 days with CpGOVA complex-loaded highly purified B cells (2 ⫻ 107), s.c. or i.v., respectively. To control for potential DC contamination of loaded B cells, mice were also challenged i.v. with 0.5 ⫻ 106 loaded DCs (data not shown). Although challenge of mice with OVA alone or with 0.5 ⫻ 106 loaded DCs failed to expand SIINFEKL-specific CD8 T cells (background ⬃0.07%), CpG-OVA conjugate-loaded B cells raised the frequency of SIINFEKL tetramer-binding CD8 T cells up to 0.3– 0.5% (Fig. 7).

Discussion

FIGURE 6. Priming of peptide-specific CTLs in wt or MyD88⫺/⫺ by CpG-OVA conjugate-loaded B cells. a, Purified and CpG-OVA conjugateor OVA alone-loaded (50 ␮g/ml) C57BL/6-derived B cells were injected s.c. (hind footpad) into C57BL/6 mice. After 7 days, draining LN or spleen cells dissociated and restimulated in vitro, followed by a 51Cr release cytotoxicity assay. Specific lysis at various E:T ratios is shown. EL-4 cells pulsed with the relevant peptide (SIINFEKL, 1 ␮M, filled symbols) or unpulsed EL-4 cells (open symbols) served as target cells. b, 1668-OVA conjugate (50 ␮g/ml)-loaded, purified B cells were injected s.c. (hind footpad) into C57BL/6 or MyD88-deficient mice backcrossed to C57BL/6 and processed, as described for a.

loaded with CpG-OVA complexes might undergo apoptosis and CpG-OVA complexes might in fact be represented by host DCs, known to ingest apoptotic cells (26). If so, T cells of MyD88deficient C57BL/6 mice challenged with wt CpG-OVA-loaded B

That DCs are able to cross-present exogenous Ags and cross-prime naive T cells is well documented (5, 6, 30). Contingent upon efficient OVA uptake and on TLR9-mediated cell activation, the results reported in this work imply that B lymphocytes also display the ability to cross-present OVA and to cross-prime CD8 T cells, although less efficiently than DCs. To date, the role of B cells as APCs has been controversial. For example, both naive and activated B cells presented cell-associated H-Y Ag in a tolerogenic form to naive CD8 T cells (6, 8). In contrast, B cells have been shown to serve as APCs for CD4 T cells when surface Igs (sIgs) promote both the capture of Ags (10, 11) as well as B cell activation (12, 13). In addition, receptormediated endocytosis of Ag that is independent of sIg specificity on B cells has been described, an example being CD21 (complement receptor type 2) (31). In this study, we use a novel approach for enhanced Ag uptake by B cells, that is, exposure of B cells to CpG-Ag conjugates. Basically, we took advantage of the previously described dual role of CpG-OVA complexes (14, 15, 20), that is, enhanced endocytosis via a sequence-nonspecific DNA receptor (15, 17, 32) as well as sequence-specific activation via endocytosed CpG-DNA that meets TLR9 at cytoplasmatic endosome-like organelles (28, 29). This approach makes use of CpGDNA as a vehicle for endocytosis of linked Ag and, upon endocytosis, as activator of Ag-loaded cells. As shown in Figs. 1–3, this sIg-independent mechanism allows Ag loading of 50% of B cells, which subsequently cross-present SIINFEKL epitope, as read out in vitro. Thus, under the conditions used, B cells shuttle exogenous Ag into the class I MHC presentation pathway. In general, exogenous Ags enter the class II pathway and are selectively associated with class II MHC molecules (33, 34). The presentation of exogenous Ag in association with MHC class I molecules has been considered as a unique characteristic of DCs (5, 35). The data presented in this work provide compelling evidence that, besides DCs (14, 15), B cells also can cross-present endocytosed OVA linked to CpG-DNA. Although linked DNA of the CpG-OVA complex triggers endocytosis in an sIg-independent fashion, endocytosed CpG-DNA in addition activates via TLR9 the cross-presenting B cells to up-regulate costimulatory molecules, to proliferate, and to produce cytokines, including IL-12. It follows that B cells loaded with CpG-OVA conjugates disclose some features that hitherto were attributed to DCs. Furthermore, OVA cross-presenting B cells appear to be immunogenic. Following adoptive transfer of CpG-OVA complex-loaded purified B cells into syngenic recipients, CD8 T cells became primed and thus

1506

B CELL-MEDIATED CROSS-PRIMING

FIGURE 7. Enumeration of SIINFEKL-specific CD62Llow, CD8-positive T cells. Purified B cells (2 ⫻ 107), loaded with either 50 ␮g/ml 1668-OVA conjugate or OVA, or cultured in medium, were injected i.v. (tail vein) or s.c. (hind footpad) into C57BL/6 mice. After 7 days, cells were prepared from spleen or LN and stained with a H-2Kb/SIINFEKL tetramer, anti-CD8, and anti-CD62L. Tetramer staining of CD8-positive cells is shown (acquiring at least 100,000 cells per sample), and percentages of tetramerpositive and CD62Llow T cells are given.

differentiated into CTL upon in vitro restimulation. In fact, SIINFEKL-specific naive CD8 T cells of recipient mice had undergone clonal expansion, as reflected by the increase in frequencies, of SIINFEKL tetramer-specific T cells in draining lymph nodes and spleen. To rule out the possibility that adoptively transferred CpG-OVA complex-loaded B cells succumb to apoptosis and thus allow CpGOVA complexes to be (re-)presented by host DCs ingesting apoptotic cells, we used as recipients mice incapable of MyD88 signaling. The fact that CpG-DNA refractory (27, 28) MyD88⫺/⫺ mice effectively generated primed CD8 T cells when challenged with CpG-OVA-loaded wt B cells excluded this contention. Clearly, the availability of conditional DC knockout mice (36) will help to answer more definitively the question as to whether CpG-Ag complex-loaded B cells function in the absence of DCs as professional APCs. Overall, these data raise several questions. For example, CpGOVA complex-mediated loading of B cells may be unique both in terms of Ag quantity and Ag load. If so, cross-presentation and cross-priming by B cells may not occur under physiological conditions. Furthermore, the ability of CpG-OVA complex-loaded B cells to cross-present/cross-prime appears lower compared with DCs. Dose-response curves of B cells and DCs exposed to CpGOVA-FITC revealed that DCs are 10- to 20-fold more efficient in CpG-OVA conjugate uptake (Fig. 2), which appears to translate in superior cross-presentation efficacy by DCs. Previous work established that compared with B cells, DCs are more efficient in presenting endogenous, naturally processed self epitopes to class IIrestricted T cells, a hierarchy that does not depend on differential processing capacity, but correlates with expression of CTLA-4 ligands and ICAM-1 molecules (37). Perhaps the rate-limiting step of B cells to cross-present exogenous Ag is cellular uptake of exogenous Ag rather than limited processing capacity to generate MHC class I T cell epitopes. Although CpG-DNA linked to OVA was efficient in B cell activation, cellular uptake (loading) of B cells with linked OVA occurred only in 50% of B cells (Fig. 1), and was 10- to 20-fold less efficient as in DCs (Fig. 2). It follows that a quantitative comparison of B cell- and DC-mediated crossprriming will require use of B cells and DCs equally loaded with CpG-OVA complexes. Our results also raise questions in regard to recent data implying that chromatin-Ig complexes activate autore-

active B cells by dual (linked) engagement of IgM and TLR9 (38). To date, all evidence available (14, 15) implies that DNA-protein particles need first to be endocytosed to engage intracellularily TLR9, while Ag receptors of B cells are membrane bound. As a consequence, we favor unlinked recognition of DNA-protein conjugates by Ag-specific B cells (38). In conclusion, our data suggest that B cells have the basic ability to route exogenous Ag in the MHC class I presentation pathway. Furthermore, we provide forceful evidence that cross-presenting B cells can cross-prime CD8 cells in mice that have not been exposed to the Ag.

Acknowledgments Gratefully, we acknowledge the skillful assistance of M. Hammel. We also thank Dr. Akira (Osaka, Japan) for providing the MyD88⫺/⫺ mice.

References 1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392:245. 2. Regnault, A., D. Lankar, V. Lacabanne, A. Rodriguez, C. Thery, M. Rescigno, T. Saito, S. Verbeek, C. Bonnerot, P. Ricciardi-Castagnoli, and S. Amigorena. 1999. Fc␥ receptor-mediated induction of dendritic cell maturation and major histocompatibility complex class I-restricted antigen presentation after immune complex internalization. J. Exp. Med. 189:371. 3. Albert, M. L., B. Sauter, and N. Bhardwaj. 1998. Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs. Nature 392:86. 4. 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. J. Exp. Med. 143:1283. 5. Heath, W. R., and F. R. Carbone. 2001. Cross-presentation, dendritic cells, tolerance and immunity. Annu. Rev. Immunol. 19:47. 6. Thery, C., and S. Amigorena. 2001. The cell biology of antigen presentation in dendritic cells. Curr. Opin. Immunol. 13:45. 7. Heath, W. R., C. Kurts, J. F. Miller, and F. R. Carbone. 1998. Cross-tolerance: a pathway for inducing tolerance to peripheral tissue antigens. J. Exp. Med. 187:1549. 8. Eynon, E. E., and D. C. Parker. 1992. Small B cells as antigen-presenting cells in the induction of tolerance to soluble protein antigens. J. Exp. Med. 175:131. 9. Fuchs, E. J., and P. Matzinger. 1992. B cells turn off virgin but not memory T cells. Science 258:1156. 10. Lanzavecchia, A. 1985. Antigen-specific interaction between T and B cells. Nature 314:537. 11. Rock, K. L., B. Benacerraf, and A. K. Abbas. 1984. Antigen presentation by hapten-specific B lymphocytes. I. Role of surface immunoglobulin receptors. J. Exp. Med. 160:1102. 12. Lenschow, D. J., A. I. Sperling, M. P. Cooke, G. Freeman, L. Rhee, D. C. Decker, G. Gray, L. M. Nadler, C. C. Goodnow, and J. A. Bluestone. 1994. Differential up-regulation of the B7-1 and B7-2 costimulatory molecules after Ig receptor engagement by antigen. J. Immunol. 153:1990.

The Journal of Immunology 13. Constant, S., N. Schweitzer, J. West, P. Ranney, and K. Bottomly. 1995. B lymphocytes can be competent antigen-presenting cells for priming CD4⫹ T cells to protein antigens in vivo. J. Immunol. 155:3734. 14. Maurer, T., A. Heit, H. Hochrein, F. Ampenberger, M. O’Keeffe, S. Bauer, G. B. Lipford, R. M. Vabulas, and H. Wagner. 2002. CpG-DNA aided crosspresentation of soluble antigens by dendritic cells. Eur. J. Immunol. 32:2356. 15. Heit, A., T. Maurer, H. Hochrein, S. Bauer, K. M. Huster, D. H. Busch, and H. Wagner. 2003. Toll-like receptor 9 expression is not required for CpG DNAaided cross-presentation of DNA-conjugated antigens but essential for crosspriming of CD8 T cells. J. Immunol. 170:2802. 16. Shirota, H., K. Sano, N. Hirasawa, T. Terui, K. Ohuchi, T. Hattori, and G. Tamura. 2002. B cells capturing antigen conjugated with CpG oligodeoxynucleotides induce Th1 cells by elaborating IL-12. J. Immunol. 169:787. 17. Sano, K., H. Shirota, T. Terui, T. Hattori, and G. Tamura. 2003. Oligodeoxynucleotides without CpG motifs work as adjuvant for the induction of Th2 differentiation in a sequence-independent manner. J. Immunol. 170:2367. 18. Krieg, A. M. 2002. CpG motifs in bacterial DNA and their immune effects. Annu. Rev. Immunol. 20:709. 19. Ahmad-Nejad, P., H. Hacker, M. Rutz, S. Bauer, R. M. Vabulas, and H. Wagner. 2002. Bacterial CpG-DNA and lipopolysaccharides activate Toll-like receptors at distinct cellular compartments. Eur. J. Immunol. 32:1958. 20. Shirota, H., K. Sano, N. Hirasawa, T. Terui, K. Ohuchi, T. Hattori, K. Shirato, and G. Tamura. 2001. Novel roles of CpG oligodeoxynucleotides as a leader for the sampling and presentation of CpG-tagged antigen by dendritic cells. J. Immunol. 167:66. 21. Karttunen, J., S. Sanderson, and N. Shastri. 1992. Detection of rare antigenpresenting cells by the lacZ T-cell activation assay suggests an expression cloning strategy for T-cell antigens. Proc. Natl. Acad. Sci. USA 89:6020. 22. Dudani, R., Y. Chapdelaine, H. H. Faassen, D. K. Smith, H. Shen, L. Krishnan, and S. Sad. 2002. Multiple mechanisms compensate to enhance tumor-protective CD8⫹ T cell response in the long-term despite poor CD8⫹ T cell priming initially: comparison between an acute versus a chronic intracellular bacterium expressing a model antigen. J. Immunol. 168:5737. 23. Specht, J. M., G. Wang, M. T. Do, J. S. Lam, R. E. Royal, M. E. Reeves, S. A. Rosenberg, and P. Hwu. 1997. Dendritic cells retrovirally transduced with a model antigen gene are therapeutically effective against established pulmonary metastases. J. Exp. Med. 186:1213. 24. Busch, D. H., I. M. Pilip, S. Vijh, and E. G. Pamer. 1998. Coordinate regulation of complex T cell populations responding to bacterial infection. Immunity 8:353. 25. Bauer, S., C. J. Kirschning, H. Hacker, V. Redecke, S. Hausmann, S. Akira, H. Wagner, and G. B. Lipford. 2001. Human TLR9 confers responsiveness to

1507

26.

27.

28.

29. 30.

31.

32.

33. 34. 35. 36.

37.

38.

bacterial DNA via species-specific CpG motif recognition. Proc. Natl. Acad. Sci. USA 98:9237. Steinman, R. M., S. Turley, I. Mellman, and K. Inaba. 2000. The induction of tolerance by dendritic cells that have captured apoptotic cells. J. Exp. Med. 191:411. Hacker, H., R. M. Vabulas, O. Takeuchi, K. Hoshino, S. Akira, and H. Wagner. 2000. Immune cell activation by bacterial CpG-DNA through myeloid differentiation marker 88 and tumor necrosis factor receptor-associated factor (TRAF)6. J. Exp. Med. 192:595. Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho, S. Sato, H. Sanjo, M. Matsumoto, K. Hoshino, H. Wagner, K. Takeda, and S. Akira. 2000. A Toll-like receptor recognizes bacterial DNA. Nature 408:740. Wagner, H. 2001. Toll meets bacterial CpG-DNA. Immunity 14:499. Den Haan, J. M., and M. J. Bevan. 2002. Constitutive versus activation-dependent cross-presentation of immune complexes by CD8⫹ and CD8⫺ dendritic cells in vivo. J. Exp. Med. 196:817. Villiers, M. B., C. L. Villiers, M. R. Jacquier-Sarlin, F. M. Gabert, A. M. Journet, and M. G. Colomb. 1996. Covalent binding of C3b to tetanus toxin: influence on uptake/internalization of antigen by antigen-specific and non-specific B cells. Immunology 89:348. Hacker, H., H. Mischak, T. Miethke, S. Liptay, R. Schmid, T. Sparwasser, K. Heeg, G. B. Lipford, and H. Wagner. 1998. CpG-DNA-specific activation of antigen-presenting cells requires stress kinase activity and is preceded by nonspecific endocytosis and endosomal maturation. EMBO J. 17:6230. Bevan, M. J. 1987. Antigen recognition: class discrimination in the world of immunology. Nature 325:192. Germain, R. N. 1986. Immunology: the ins and outs of antigen processing and presentation. Nature 322:687. Norbury, C. C., and L. J. Sigal. 2003. Cross priming or direct priming: is that really the question? Curr. Opin. Immunol. 15:82. Jung, S., D. Unutmaz, P. Wong, G. Sano, S. K. De los, T. Sparwasser, S. Wu, S. Vuthoori, K. Ko, F. Zavala, et al. 2002. In vivo depletion of CD11c⫹ dendritic cells abrogates priming of CD8⫹ T cells by exogenous cell-associated antigens. Immunity 17:211. Guery, J. C., and L. Adorini. 1995. Dendritic cells are the most efficient in presenting endogenous naturally processed self-epitopes to class II-restricted T cells. J. Immunol. 154:536. Leadbetter, E. A., I. R. Rifkin, A. M. Hohlbaum, B. C. Beaudette, M. J. Shlomchik, and A. Marshak-Rothstein. 2002. Chromatin-IgG complexes activate B cells by dual engagement of IgM and Toll-like receptors. Nature 416:603.