ISCOMATRIX Adjuvant Induces Efficient Cross-Presentation of Tumor Antigen by Dendritic Cells via Rapid Cytosolic Antigen Delivery and Processing via Tripeptidyl Peptidase II

The Journal of Immunology

ISCOMATRIX Adjuvant Induces Efficient Cross-Presentation of Tumor Antigen by Dendritic Cells via Rapid Cytosolic Antigen Delivery and Processing via Tripeptidyl Peptidase II1 Max Schnurr,2* Martin Orban,* Neil C. Robson,‡ Amanda Shin,* Hal Braley,§ Denise Airey,§ Jonathan Cebon,‡ Eugene Maraskovsky,§ and Stefan Endres† Cancer vaccines aim to induce antitumor CTL responses, which require cross-presentation of tumor Ag to CTLs by dendritic cells (DCs). Adjuvants that facilitate cross-presentation of vaccine Ag are therefore key for inducing antitumor immunity. We previously reported that human DCs could not efficiently cross-present the full-length cancer/testis Ag NY-ESO-1 to CTL unless formulated as either an immune complex (NY-ESO-1/IC) or with ISCOMATRIX adjuvant. We now demonstrate that NY-ESO-1/ICs induce crosspresentation of HLA-A2- and HLA-Cw3-restricted epitopes via a proteasome-dependent pathway. In contrast, cross-presentation of NY-ESO-1/ISCOMATRIX vaccine was proteasome independent and required the cytosolic protease tripeptidyl peptidase II. Trafficking studies revealed that uptake of ICs and ISCOMATRIX vaccine by DCs occurred via endocytosis with delivery to lysosomes. Interestingly, ICs were retained in lysosomes, whereas ISCOMATRIX adjuvant induced rapid Ag translocation into the cytosol. Ag translocation was dependent on endosomal acidification and IL-4-driven differentiation of monocytes into DCs. This study demonstrates that Ag formulation determines Ag processing and supports a role for tripeptidyl peptidase II in cross-presentation of CTL epitopes restricted to diverse HLA alleles. The Journal of Immunology, 2009, 182: 1253–1259.

D

endritic cells (DC)3 are highly specialized APCs that take up exogenous material from the extracellular environment for presentation in the context of MHC molecules. Next to presentation of antigenic epitopes on MHC class II molecules to CD4⫹ T cells, DCs can also shuttle Ag to the MHC class I processing pathway for CD8⫹ T cell activation, a process termed crosspresentation (reviewed in Ref. 1). This enables DCs that have engulfed tumor Ag to activate Ag-specific CD8⫹ T cells capable of tumor cell killing. The cytotoxic antitumor effect of CD8⫹ T cells also depends on CD4⫹ T cells, which provide CD8⫹ T cells with growth factors and costimulatory signals (2–5). Thus, cancer vaccines should aim at inducing both antitumor CD4⫹ and CD8⫹ T cell responses.

*Department of Internal Medicine and †Division of Clinical Pharmacology, University of Munich, Munich, Germany; ‡Ludwig Institute for Cancer Research, Melbourne Centre for Clinical Sciences, Austin Health, Heidelberg, Victoria, Australia; and § CSL Limited, Parkville, Victoria, Australia Received for publication July 29, 2008. Accepted for publication November 21, 2008. 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 is part of the doctoral thesis of M.O. at the Ludwig-Maximilians-University (Munich, Germany). This work was supported by the Deutsche Krebshilfe (Max Eder Research Grant to M.S.), the Deutsche Forschungsgemeinschaft (Grant GK1202 to M.S. and S.E.), the Fo¨FoLe Program of the Ludwig-Maximilians-University of Munich (to M.O.), a Program Grant from the Australian National Health and Medical Research Council (to E.M.) and the Ludwig Institute for Cancer Research. N.R. is an Australian National Health and Medical Research Council Industry fellow. E.M. is an employee of CSL Limited and an Honorary Senior Research Fellow of the Ludwig Institute for Cancer Research.

2

Address correspondence and reprint requests to Dr. Max Schnurr, Medizinische Klinik Innenstadt, University of Munich, Ziemssenstrasse 1, 80336 Munich, Germany. E-mail address: [email protected]

3

Abbreviations used in this paper: DC, dendritic cell; IC, immune complex; TPPII, tripeptidyl peptidase II; ISCOM, immune-stimulating complexes; DOC, deoxycholic acid; pDC, plasmacytoid DC; MoDC, monocyte-derived DC; AAF-CMK, Ala-AlaPhe-chloromethylketone; LAMP, lysosomal-associated membrane protein-1; EEA-1, early endosomal Ag. Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

www.jimmunol.org

Tumor vaccines based on full-length recombinant proteins have the potential to generate broad-based CD4⫹ and CD8⫹ T cell responses (6). However, using recombinant tumor Ag, we found that cross-presentation of soluble Ag by DCs is highly inefficient, whereas presentation to CD4⫹ T cells was readily induced (7). Delivering tumor Ag to the MHC class I processing machinery in DCs is thus a challenge in the design of cancer vaccines and requires a better understanding of the mechanisms of cross-presentation. Ag processing via MHC class I and II pathways largely depends on the route of Ag uptake. Receptors facilitating cross-presentation include members of the C-type lectin receptor family, FcR, and integrins (reviewed in Ref. 8). Because most CD8⫹ T cell epitopes are generated in the cytosol, where cleavage of the Ag by the proteasome and other protease complexes occurs (9), Ag translocation from endocytic compartments into the cytosol is essential for effective cross-presentation. Peptides generated by cytosolic proteases are then transported via TAP into the endoplasmic reticulum where additional trimming may occur (10) to allow loading of optimal length peptides onto nascent MHC class I molecules and subsequent transport of the peptide MHC complexes to the cell surface for presentation to CD8⫹ T cells. Our group has previously shown that cross-presentation of recombinant NY-ESO-1, a 180-aa protein, which is expressed in a variety of human cancers including melanoma, breast, lung, prostate, and others (11, 12), by DCs depends on Ag formulation. Using NY-ESO-1 as a model Ag we found that cross-presentation efficiency was enhanced by targeting the Ag to FcR with ICs (7, 13) or by formulation of the Ag with ISCOMATRIX adjuvant, which is being developed for use in human vaccines (7). ISCOMATRIX adjuvant is based on the immune-stimulating complexes (ISCOM) technology, which combines an efficient Ag delivery system with the immunostimulatory activity of saponin and has been shown to promote humoral and cellular immune responses in a variety of animal models (reviewed in Ref. 14). HLA-A2-restricted cross-presentation of NY-ESO-1/immune complex (IC) was proteasome dependent, and for NY-ESO-1/ISCOMATRIX

1254 vaccine proteasome was independent (7). Inhibition of cross-presentation of both NY-ESO-1 formulations by ICP-47, a viral protein interfering with TAP, indicated that processing by cytosolic proteases was required. However, it was unclear whether ISCOMATRIX formulation impeded the access of the Ag to the proteasome or whether it shuttled the Ag to another, yet unidentified cytosolic protease. Cytosolic peptidases such as tripeptidyl peptidase II (TPPII), leucine aminopeptidase, bleomycin hydrolase, puromycin-sensitive aminopeptidase, and thimet oligopeptidase have all been implicated in MHC class I Ag processing (reviewed in Refs. 15 and 16). In this study, we investigated the mechanisms underlying crosspresentation of NY-ESO-1 protein formulations by human DCs. We found that cytosolic proteolysis by DCs occurred either via proteasomal processing for NY-ESO-1/ICs or via the enzyme complex TPPII for ISCOMATRIX-formulated Ag. Ag trafficking studies using confocal microscopy revealed that ISCOMATRIX adjuvant facilitated cytosolic Ag translocation from endosomes/ lysosomes, a function that was strictly restricted to immature myeloid DCs. These studies may guide the development of cancer vaccines using recombinant tumor Ag.

Materials and Methods Generation of recombinant NY-ESO-1 formulations Full-length NY-ESO-1 protein was produced in Escherichia coli and purified under Current Good Manufacturing Practice conditions (17). NYESO-1/ICs were generated by mixing NY-ESO-1 protein with anti-NYESO-1 mAb (clone ES121) at a 1:2 molar ratio in serum-free RPMI 1640 at 37°C for 30 min as described (7). ISCOMATRIX-formulated NYESO-1 was generated as previously described (18). ISCOMATRIX is a registered trademark of ISCOTEC, a CSL company.

Generation of OVA formulations For confocal studies, OVA (Sigma-Aldrich) was palmitified, labeled with Alexa Fluor 488 or 555 (Molecular Probes), and associated into an ISCOM by formulation with ISCOPREP saponin, phospholipids, and cholesterol. Briefly, OVA was dissolved to 6.6 mg/ml in 50 mM Na2CO3, 5% deoxycholic acid (DOC), pH 9.0. Normal human serum-palmitic acid (SigmaAldrich) was dissolved to 10 mg/ml in dimethylformamide, and 50 ␮l were added to 950 ␮l of OVA solution. Palmitified OVA was purified by size exclusion chromatography in 50 mM Tris, 0.15 M NaCl, 0.15% DOC, pH 8.0, and quantified using the Bradford protein assay. Alexa Fluor 488 and 555 labeling and palmitification of OVA were conducted simultaneously. Palmitified OVA was concentrated to 3 mg/ml, and buffer was exchanged into PBS, 0.15% DOC, pH 7.4. Tributylphosphine (Sigma-Aldrich) was added to a concentration of 5 mM, and the mixture was incubated for 2 h at 60°C before buffer exchange into fresh PBS, 0.15% DOC. The palmitified OVA solution was then added to 1 mg of dry Alexa Fluor 555 carboxylic acid succinimidyl ester or 1 mg of dry Alexa Fluor 488 succinimidyl ester immediately after the addition of normal human serum-palmitic acid and incubated for 3 h in the dark at 20 –23°C. The labeled protein was purified by size exclusion chromatography in PBS and quantitated by OD at 280 nm. OVA/IC were generated by incubating 1 ␮g/ml palmitified OVA-Alexa 488 with 20 ␮g/ml rabbit anti-OVA serum (Sigma-Aldrich) in serum-free RPMI.

Isolation and culture of APCs PBMCs from healthy volunteers were prepared by Ficoll-Paque density gradient centrifugation. Monocytes, plasmacytoid DCs (pDC), and B cells were isolated by positive selection using magnetic beads against CD14, BDCA-4, and CD19, respectively (Miltenyi Biotec). Positive selection was repeated to obtain a purity of 95% for each cell type. pDCs were cultured with IL-3 (10 ng/ml), and monocyte-derived DCs (MoDC) were generated by culturing CD14⫹ cells with GM-CSF (20 ng/ml) and IL-4 (500 U/ml) for 6 or 7 days as described (7).

Ag presentation studies HLA-A2⫹ and HLA-Cw3⫹ MoDCs were plated at 5 ⫻ 105/ml, incubated with NY-ESO-1/IC or NY-ESO-1/ISCOMATRIX vaccine for 2 h at 37°C, washed, and cultured overnight. MoDCs pulsed with the HLA-A2- or HLA-Cw3-restricted peptides NY-ESO-1157–165 or NY-ESO-192–100 (Auspep) served as positive controls. Ag-pulsed MoDCs were cocultured with NY-ESO-1157–165- or NY-ESO-192–100-specificCD8⫹ T cell lines or

CROSS-PRESENTATION OF TUMOR Ag BY HUMAN DCs clones. IFN-␥ production by CD8⫹ T cells was assessed by intracellular cytokine staining or ELISPOT assay. For intracellular cytokine staining, 105 DCs and 1–2 ⫻ 104 T cells were incubated for 4 h in 96-well roundbottom plates in the presence of 10 ␮g/ml brefeldin A (Sigma-Aldrich). Cells were washed and stained with anti-CD8 mAbs for 20 min at 4°C, fixed with 1% paraformaldehyde, and stained with anti-IFN-␥ mAb (BD Pharmingen) in a 0.25% saponin buffer. Samples were analyzed by FACS, and data were analyzed using FloJo software (version 3.4; Tree Star). ELISPOT assays were performed with the ELISpotPLUS kit from MABTECH. DCs were cocultured with CD8⫹ T cells for 16 –18 h. Subsequently, cells were removed, and plates were washed with water and developed according to the manufacturer’s instructions.

FACS analysis of cell surface phenotype The following fluorochrome-conjugated mAbs were purchased from BD Pharmingen: CD8, CD14, CD19, CD20, CD123, and HLA-DR. The BDCA-4 mAb was obtained from Miltenyi Biotec. Clone BB7.2 (American Type Culture Collection) was used to screen PBMCs for HLA-A2 expression. FACS analysis was performed with a FACSCalibur (BD Biosciences).

Inhibition studies of Ag uptake and processing Where indicated, MoDCs were incubated with inhibitors 30 min before and throughout the Ag pulse: cytochalasin D, wortmannin, phenylarsine, chloroquine, clastolactacystin ␤-lactone (lactacystin), MG-101, and leupeptin were purchased from Sigma-Aldrich; Ala-Ala-Phe-chloromethylketone (AAF-CMK) from were purchased Biomol; butabindide oxalate was from Tocris Bioscience; and concanamycin B was a gift from Dr. J. Villadangos (Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia).

Ag trafficking studies by confocal microscopy MoDCs (3 ⫻ 105) were incubated for 30 min on poly-D-lysine-coated glass coverslips before washing and incubating at 37°C with Ag formulations. Cells were washed and chased for the times indicated before fixation with 4% paraformaldehyde. After a washing, cells were incubated with NH4Cl for 10 min and permeabilized with 0.5% saponin in 1% BSA in PBS for 30 min. Cells were stained with mouse IgG anti-EEA-1 or mouse IgG antilysosomal-associated membrane protein-1 (LAMP-1; BD Pharmingen) mAb for 30 min at room temperature. After a washing, cells were incubated for 30 min with Alexa Fluor 555-labeled Ab (goat anti-mouse IgG). Coverslips were mounted onto glass slides with Moviol 40-88 (SigmaAldrich) and rested for 1 h at 4°C. Microscopy was performed using the Confocal Laser Scanning Microscope LSM 510 Meta (Carl Zeiss). For studying cytosolic translocation in viable cells, DCs were incubated for 30 min on poly-D-lysine-coated coverslips and incubated with Ag formulations. Coverslips were mounted onto an aluminum slide connected to a heating device maintaining 37°C. Confocal images were obtained from duplicate samples, and 300 cells were analyzed. National Institutes of Health software (Image J; Java image processing program) was used to quantify cells with cytosolic fluorescence by measuring the fluorescence intensity with a line scan analysis. Background levels were determined by averaging the cytosolic fluorescence intensity of controls. Cells were considered positive for cytosolic fluorescence if their cytosolic fluorescence intensity was higher than three times the background value.

Results DCs cross-present NY-ESO-1 protein via two distinct Ag processing pathways We have previously reported that protein formulation determines Ag presentation by DCs (7). DCs presented soluble NY-ESO-1 protein exclusively to CD4⫹ Th cells, whereas protein formulated as an IC or associated with ISCOMATRIX adjuvant was additionally cross-presented to CD8⫹ T cells. Ag titration experiments revealed that cross-presentation of NY-ESO-1/ISCOMATRIX vaccine and NY-ESO-1/IC on both HLA-A2 and Cw3 was highly efficient (Fig. 1A). When Ag dose was limiting, generation of the HLA-A2 and Cw3 epitopes was more efficient for NY-ESO-1/ ISCOMATRIX vaccine (Fig. 1A). Proteasomal processing was required for the generation of both epitopes when NY-ESO-1/IC was used as determined by addition of the proteasome inhibitors, lactacystin (Fig. 1B) and epoxomycin (data not shown), whereas cross-presentation of the NY-ESO-1/ISCOMATRIX vaccine was

The Journal of Immunology

1255

FIGURE 2. Influence of specific inhibitors of Ag uptake and processing on cross-presentation efficacy. DCs were incubated in the absence (䡺) or presence (f) of indicated inhibitors 45 min before pulsing with NY-ESO-1/IC or NY-ESO-1/ISCOMATRIX vaccine (NY-ESO-IMX): A, Inhibitor of actin polymerization, cytochalasin D (5 ␮M); B, inhibitor of cysteine proteases, leupeptin (1 mM); C, calpain inhibitor I (MG101; 1 ␮M); D, inhibitors of TPPII, AAF-CMK (5 ␮M); E, butabindide oxalate (200 ␮M). After overnight culture, DCs were cocultured with CTLs. Peptide-pulsed DCs served as controls. CD8⫹ T cell IFN-␥ production in the absence of inhibitors was normalized to 100%. Data are mean values ⫾ SD of four to six experiments.

FIGURE 1. Ag formulation determines cross-presentation efficacy and kinetics by DCs. DCs from HLA-A2⫹ and HLA-Cw3⫹ donors were pulsed with NY-ESO-1/IC (䡺), or NY-ESO-1/ISCOMATRIX vaccine (NY-ESO1/IMX; ⽧) and cultured overnight before coculture with NY-ESO-1-specific CTLs. Induction of IFN-␥ production by CD8⫹ T cells was quantified by intracellular cytokine staining assay and is shown as the percent of maximal response induced by peptide-loaded DCs. A, Dose-response curve for different NY-ESO-1 formulations. B, Influence of proteasome inhibition with lactacystin on cross-presentation of the HLA-A2- and Cw3-restricted NY-ESO-1 epitopes. C, Cross-presentation kinetics for NY-ESO1/IC and NY-ESO-1/ISCOMATRIX vaccine. Data are mean values ⫾ SD of triplicate wells and are representative of three independent experiments.

unaffected (Fig. 1B). Furthermore, ISCOMATRIX-formulated NY-ESO-1 resulted in more rapid generation of the HLA-A2 and Cw3 epitopes as compared with NY-ESO-1/IC (Fig. 1C and data not shown). These observations led us to hypothesize that DCs processed NY-ESO-1 protein differently when it was formulated either as an IC or with ISCOMATRIX adjuvant. The mode of Ag uptake may determine access to distinct intracellular organelles and processing pathways (19). To investigate mechanisms of Ag uptake, we incubated DCs with cytochalasin D, an inhibitor of actin polymerization, before Ag pulsing. Cytochalasin D had no effect on presentation of exogenous peptide to CTLs but potently inhibited cross-presentation of NY-ESO-1/ISCOMATRIX vaccine (Fig. 2A). Cross-presentation of NY-ESO-1/IC was inhibited to a lesser extent, likely reflecting predominant uptake of IC via FcR-mediated endocytosis, which is cytochalsin D insensitive (20). The role of endo-/lysosomal processing in cross-presentation is still a matter of debate. We previously demonstrated

that concanamycin B, an inhibitor of lysosomal acidification, completely abrogated cross-presentation of both NY-ESO-1/IC and NY-ESO-1/ISCOMATRIX vaccine (7). To further investigate the role of lysosomal proteases on MHC-I epitope generation, we subjected DCs to leupeptin, an inhibitor of cysteine proteases, and MG-101, an inhibitor of calpains. Both inhibitors partially reduced cross-presentation of NY-ESO-1/IC and NY-ESO-1/ISCOMATRIX vaccine to a similar extent (Fig. 2, B and C), indicating that lysosomal Ag processing precedes cytosolic Ag processing and MHC class I epitope generation. Cytosolic translocation of endocytosed Ag is required for Ag to gain access to cytosolic proteases, such as the proteasome, which due to its endopeptidase activity is critically involved in MHC class I epitope generation (reviewed in Ref. 21). Our earlier studies revealed a striking difference in the proteasomal processing requirements between the two NY-ESO-1 formulations (see also Fig. 1B). Cross-presentation of NY-ESO-1/IC, on both HLA-A2 and Cw3 molecules, was reduced by proteasome inhibition, whereas NY-ESO-1/ISCOMATRIX vaccine was unaffected (Fig. 1B), indicating that MHC class I epitope generation may occur via an alternative processing pathway. Geier et al. (22) have reported a cytosolic protease, termed TPPII, with aminopeptidase and endopeptidase activities. TPPII has been shown to be involved in MHC class I epitope generation of a HLA-A3-restricted HIV epitope (23). To investigate whether TPPII plays a role in cross-presentation of NY-ESO-1 formulations, we treated DCs with two different TPPII inhibitors, AAF-CMK and butabindide oxalate. Both inhibitors significantly impaired CTL activation for DCs pulsed with NY-ESO-1/ISCOMATRIX vaccine (Fig. 2, D and E). In contrast, DCs pulsed with NY-ESO-1/IC were unaffected, in line with predominant proteasomal processing. Thus, ISCOMATRIX adjuvant

1256

CROSS-PRESENTATION OF TUMOR Ag BY HUMAN DCs

FIGURE 3. DCs take up IC and ISCOMATRIX vaccine via endocytosis. A, DCs were pulsed with Alexa Fluor 488-conjugated OVA protein, as soluble OVA, OVA/IC, or OVA/ISCOM formulations, for 60 min, washed, and analyzed by FACS. DCs incubated with Ag formulations on ice served as negative controls. B and C, DCs were incubated with inhibitors of endocytosis, wortmannin or phenylarsine, for 30 min before pulsing with OVA formulations. OVA uptake in the absence of inhibitors was normalized to 100%. Data show mean values ⫾ SD of four experiments. MFI, Mean fluorescence index.

can generate specific class I epitopes via an alternative, TPPIIdependent processing pathway in human DCs. DCs internalize both Ag formulations via endocytosis To further characterize this alternative cross-presentation pathway in DCs, we investigated routes of uptake and intracellular trafficking of IC and ISCOMATRIX-formulated Ag using fluorochromelabeled OVA formulations. OVA was selected in preference to NY-ESO-1 for these studies because it was a highly pure monomer for labeling purposes. To detect quantitative differences in Ag uptake, DCs were pulsed with soluble OVAAlexa 488, OVAAlexa 488/ IC, or OVAAlexa 488/ISCOM formulations, and fluorescence intensity was assessed by flow cytometry. DCs internalized IC and ISCOM formulations at comparable levels, but less than soluble protein (Fig. 3A). Fluorescence intensities remained constant within a 6-h period with no indication of bleaching or dye exclusion (data not shown). Treating DCs with phenylarsine, an inhibitor of receptor-mediated endocytosis, or with the phosphoinositide 3-kinase inhibitor wortmannin, which interferes with macropinocytosis, effectively blocked uptake of these Ag formulations (Fig. 3, B and C). To assess whether the various Ag formulations gained access to distinct intracellular organelles, we analyzed intracellular Ag routing in DCs after various chasing periods by confocal microscopy. After 10 min, both OVA/IC and OVA/ISCOM formulations were exclusively

FIGURE 4. ISCOMATRIX adjuvant induces Ag translocation from endosomes/lysosomes to cytosol in DCs. A, DCs were incubated with OVAAlexa 488/IC (green) or pOVAAlexa 488/ISCOM (green) for 10 min and extensively washed. Cells were subsequently fixed, permeabilized, and stained with mAb against the early endosomal marker EEA-1 (red). Right, Overlay pictures. B, After a 2-h Ag pulse, DCs were fixed and stained with mAb against the lysosomal marker LAMP-1 (red). Right, Overlay pictures. C, DCs were pulsed with soluble OVA (left), OVA/IC (middle), or OVA/ISCOM (right) for 4 h, washed, and fixed. D, DCs were simultaneously pulsed with OVAAlexa 488/ISCOM (green) and OVAAlexa 555/IC (red) for 1 h and analyzed as viable cells without prior fixation. Right, Overlay pictures. Data are representative of five to eight experiments.

located in small organelles near the cell surface, expressing early endosomal Ag 1 (EEA-1; Fig. 4A). No colocalization with LAMP-1 was observed at this time point (data not shown). After 2 h, the Ag accumulated in LAMP-1⫹ lysosomes located near the nucleus (Fig. 4B). Trafficking from EEA-1⫹ to LAMP-1⫹ compartments was observed for all three Ag formulations. No colocalization with calreticulin, a marker for the endoplasmic reticulum, was seen at any given time point examined (Supplemental Fig. 1a and data not shown).4 4

The on-line version of this article contains supplemental material.

The Journal of Immunology

1257

FIGURE 5. Inhibition of lysosomal acidification prevents cytosolic Ag translocation. A, Chasing of viable DCs after pulsing with OVAAlexa 488/ ISCOM after various time points by fluorescence microscopy. Values are the percent of DCs with staining pattern indicating cytosolic Ag translocation. B, DCs were incubated with the indicated concentrations of concanamycin B 30 min before Ag pulsing. Translocation was analyzed by fluorescence microscopy. C, Influence of chloroquine on cytosolic Ag translocation. Data are mean values ⫾ SD of four experiments. D, Representative confocal pictures of DCs pulsed with OVAAlexa 488/ISCOM in the absence (left) or presence of concanamycin B (middle) or chloroquine (right).

ISCOMATRIX adjuvant facilitates rapid Ag translocation into the cytosol of DCs Chasing OVA/ISCOM pulsed DCs for 4 h by confocal microscopy revealed a diffuse staining pattern of the cell body including dendrites, indicative of cytosolic Ag translocation (Fig. 4C). No colocalization with compartments expressing EEA-1, LAMP-1, or calreticulin was detected between 4 and 6 h (Supplemental Fig. 1a and data not shown). In contrast, soluble protein and ICs remained within lysosomal compartments. Because fixation and permeabilization of DCs for intracellular staining resulted in a significant loss of fluorescence intensity, we repeated experiments with viable DCs to increase the sensitivity for detecting Ag translocation. Again, the OVA/ISCOM formulation induced a cytosolic staining pattern in DCs (Fig. 4D). Translocation occurred as early as 30 min after Ag pulsing and was detectable in ⬃90% of DCs within 2 h (Fig. 5A). As observed with fixed cells, viable DCs pulsed with soluble protein or ICs were devoid of cytosolic translocation during a 6-h chasing period (data not shown). To rule out that translocation is due to rupture of lysosomes by ISCOMATRIX adjuvant, DCs were pulsed simultaneously with OVAAlexa 488/ISCOM and OVAAlexa 555/IC or soluble OVAAlexa 488. Both, OVAAlexa 555/IC or soluble OVAAlexa 488, remained strictly in intracellular compartments, together with a minor fraction of OVAAlexa 488/ ISCOM. However, the majority of ISCOM-formulated Ag was selectively translocated into the cytosol (Fig. 4D and data not shown). Functional lysosomal proton pumps are required for Ag translocation We previously reported that inhibition of the lysosomal proton pump with concanamycin B abrogated cross-presentation by DCs (7). To assess whether Ag translocation originates from immature or mature lysosomes, we treated OVAAlexa 488/ISCOM-pulsed DCs with concanamycin B. This had no effect on Ag uptake but resulted in a nearly complete inhibition of Ag translocation from

FIGURE 6. Cytosolic Ag translocation is restricted to myeloid DCs. A, Immature MoDCs (iMoDC), CD40L-matured DCs (mMoDC), monocytes (Mono), macrophages (MØ), B cells, and pDC were incubated with OVAAlexa 488/ISCOM and analyzed by FACS analysis for Ag uptake. Values are the percent of cells having internalized Ag. B, Cytosolic Ag translocation in MoDCs, monocytes, and macrophages was analyzed by confocal microscopy. Data are mean values ⫾ SD of four experiments. C, Representative confocal pictures of MoDCs (left), macrophages (middle), and monocytes (right) pulsed with OVAAlexa 488/ISCOM for 1 h are shown. D, Acquisition of the capacity for cytosolic Ag translocation of monocytes cultured in the presence of GM-CSF (䡺) or GM-CSF plus IL-4 (f) over a period of 8 days. Data are the mean values ⫾ SD of four experiments. E, Cross-presenting function of monocytes cultured in the presence of GMCSF (䡺) or GM-CSF plus IL-4 (f). After a culture period of 7 days, cells were pulsed with NY-ESO-1/ISCOMATRIX vaccine (NY-ESO-1/IMX; 0.1, 1.0 and 5.0 ␮g/ml) or peptide and cocultured with HLA-A2-restricted NY-ESO-1-specific CTLs. IFN-␥ production in response to peptide-pulsed APCs was normalized to 100%. Values are for representative experiment of three.

lysosomes (Fig. 5, B and D). Similar effects were observed for cholorquine, a lysomotropic amine that raises lysosomal pH (Fig. 5, C and D). Thus, lysosomal acidification is required for Ag translocation into the cytosol for cross-presentation. Cytosolic translocation is restricted to immature DCs and requires IL-4 during monocyte differentiation into MoDCs Because cytosolic Ag translocation is a critical step for MHC class I epitope generation by cytosolic proteases, we were interested in how this function is regulated in different DC types and other APCs. To examine Ag trafficking, we incubated different APC types from human blood with OVAAlexa 488/ISCOM. Immature MoDCs, monocytes, and macrophages were equally capable of taking up OVAAlexa 488/ISCOM. In contrast, mature DCs, activated with LPS or CD40L, as well as B cells and pDCs were

1258 devoid of this function (Fig. 6A). Despite efficient Ag uptake by monocytes and macrophages, only MoDCs showed signs of cytosolic translocation, whereas monocytes and macrophages retained the Ag in intracellular organelles throughout a 6-h chase period (Fig. 6, B and C). As DCs develop from monocyte precursors during a culture period of 5–7 days with GM-CSF and IL-4, we were interested in defining at which stage MoDCs acquire the ability to translocate Ag. MoDC gained this function as early as day 3 of culture (25% of cells), reaching a plateau by day 6 (85% of cells; Fig. 6D). Translocation was dependent on the presence of IL-4, given that macrophages cultured with GM-CSF alone were devoid of this function. Moreover, lack of Ag translocation correlated with defective cross-presenting function by GM-CSF-cultured cells as assessed by the levels of IFN-␥ secretion induced in HLA-A2-restricted NY-ESO-1-specific CTLs (Fig. 6E). Thus, cytosolic Ag translocation appears to be a prerequisite for effective CTL stimulation.

Discussion This study shows that the Ag formulation, either as IC or with ISCOMATRIX adjuvant, determines whether human DCs use a classical proteasome-dependent or an alternative proteasome-independent processing pathway for cross-presentation of NY-ESO-1 CTL epitopes. Presentation of HLA-A2- and Cw3-restricted NYESO-1 epitopes by DCs that had internalized NY-ESO-1/IC was inhibited by 50 –75% by the proteasome inhibitor lactacystin, whereas presentation of NY-ESO-1/ISCOMATRIX vaccine was unaffected. Conversely, AAF-CMK (inhibitor of TPPI and TPPII) and butabindide oxalate (specific TPPII inhibitor) inhibited crosspresentation of the HLA-A2 epitope induced by the NY-ESO-1/ ISCOMATRIX vaccine by ⬃75% but had no effect on cross-presentation of NY-ESO-1/IC. A role of TPPII in cross-presentation has been reported by several groups. Seifert et al. demonstrated that cross-presentation of the HIV epitope Nef73– 82, which is restricted to HLA-A3/HLA-A11 and insensitive to proteasome inhibition (24), is generated by TPPII as assessed by chemical inhibition and siRNA (23). A broader role for TPPII in cross-presentation was suggested by the finding that TPPII mediates trimming of proteasomal products 15 aa or longer for further processing and MHC class I binding (25). However, whether TPPII is indispensable for the generation of most MHC class I epitopes was challenged by another group who found that inhibition of TPPII by siRNA does not reduce the presentation of MHC class I epitopes in the presence of functional proteasomes (26). Analysis of epitopes generated by APCs from TPPII⫺/⫺ mice suggests that TPPII has a moderately destructive role for certain epitopes (27). It is more likely that TPPII becomes essential for generating epitopes that cannot be generated by the proteasome. In a study by Benham et al., inhibiting the proteasome with lactacystin resulted in a reduction of peptide-loaded MHC class I molecules. However, peptide loading of some MHC class I alleles, i.e., HLA-A3, HLAA11, and HLA-B35, was unaffected upon proteasome blockade (28). Together, these findings indicate that TPPII is required for generating peptides for certain HLA alleles. Our study adds complexity to this field, given that we are the first to show that TPPII can substitute for the proteasome for generating the HLA-A2 epitope for NY-ESO-1. Although we have shown that ISCOMATRIX adjuvant shuttles Ag to TPPII for certain HLA-restricted epitopes, this does not exclude the possibility that the proteasome may also be utilized for the generation of others. Interestingly, targeting NY-ESO-1 to TPPII using ISCOMATRIX adjuvant was superior to classical proteasomal processing for the HLA-A2 and Cw3 epitopes with regard to cross-presentation kinetics and efficacy. Considering the fact that only a minor fraction of ⬍0.1% of peptides processed by the proteasome are suitable for MHC class

CROSS-PRESENTATION OF TUMOR Ag BY HUMAN DCs I binding (29), bypassing the proteasome could be a strategy to favor the generation of diverse MHC class I epitopes. Despite progress in understanding the molecular requirements of cross-presentation in murine DCs, primarily using bone marrow-derived DCs, little is known whether similar mechanisms apply to human DCs. Ag translocation from phagosomes into the cytosol has been shown to be critical for Ag access to proteasomes for MHC class I epitope generation (30 –32). In addition, Ag uptake into distinct intracellular organelles has been shown to favor presentation in the context of either MHC class I or MHC class II (19). Having identified distinct processing pathways for IC and ISCOMATRIX-formulated Ag in human DCs, we investigated whether these formulations target different intracellular processing pathways. To characterize Ag trafficking in DCs, we generated IC and ISCOMATRIX vaccine formulations with fluorochrome-labeled OVA. These studies showed no difference with regard to 1) efficacy of Ag uptake, 2) sensitivity of Ag uptake to inhibition with wortmannin or phenylarsine, and 3) Ag transport from EEA-1⫹ endosomes into LAMP-1⫹ lysosomes. Strikingly, ISCOMATRIX adjuvant mediated rapid cytosolic Ag translocation from lysosomes into the cytosol, whereas IC prolonged Ag accumulation in LAMP-1⫹ lysosomes, with little evidence of cytosolic translocation. In viable DCs, in contrast to fixed cells, sensitivity for detecting cytosolic fluorescence by confocal microscopy was increased, but still no cytosolic translocation of IC-associated Ag was detected. This finding is in contrast to a previous study of murine bone marrow-derived DCs, in which IC-associated Ag translocation was detected by both fluorescence microscopy and cytosolic fractionation (32). Species- and cell type-specific differences, as well as high Ag concentrations used in murine studies, could account for this discrepancy. An unresolved question remains as to how the Ag is targeted to either proteasomal or TPPII processing. A possible explanation is compartmentalization of the enzyme complexes within the cytosol. Houde et al. (33) demonstrated that ubiquinated proteins colocalize with proteasomes on the cytoplasmic side of phagosomes within DCs. Whether TPPII is also compartmentalized is unknown. We analyzed TPPII expression in DCs with a TPPII-specific Ab by confocal microscopy, revealing a diffuse cytosolic staining pattern (Supplemental Fig. 1b) and no colocalization with EEA-1, LAMP-1, or calreticulin (data not shown). Rapid release of ISCOMATRIX-formulated Ag from lysosomes might bypass the proteasome for TPPII processing. Whether TPPII plays a role in crosspresentation of selected epitopes or influences the epitope repertoire displayed by DCs is an exciting area of research. In this regard, although we have shown that ISCOMATRIX adjuvant shuttles Ag to proteasome-independent pathway (such as TPPII) for the generation of the HLA-A2- and Cw3-restricted epitopes, it is possible that it uses the proteasome for the generation of other epitopes, as has been found for the HLA-B7 epitope of NY-ESO-1 (see Ref. 38; N. C. Robson et al., manuscript in preparation). Targeting full-length Ag to both cross-presentation pathways may generate a broad CTL repertoire against tumors, e.g., against epitopes that will be destroyed by the immunoproteasome (34, 35). Little is known about cytosolic Ag translocation in different human APC types. In this study, we provide evidence that this is a highly specialized function of myeloid DCs, whereas other phagocytic cells, such as monocytes and macrophages, retained the Ag in lysosomes. This finding correlates with our earlier observation that only myeloid DCs, DCs derived from monocytes with GM-CSF and IL-4 as well as CD1c⫹ peripheral blood DCs, were capable of activating both CD8⫹ and CD4⫹ T cells, whereas other APC types, such as monocytes, macrophages, B cells, and pDCs, were limited to CD4⫹ T cell activation (7). In

The Journal of Immunology vivo, the GM-CSF-plus-IL-4-generated MoDCs appear to represent an inflammatory-type DC population derived from the monocyte lineage rather than the steady state DCs found in lymphoid tissues (36). The role that lymphoid-resident, steady state DCs (such as CD8␣⫹ cross-presenting DCs) play in the processing and presentation of ISCOMATRIX vaccines is the subject of current studies. In addition, the present study identifies IL-4 as a critical differentiation factor during MoDC generation from monocyte precursors for the acquisition of Ag-translocatory function. MoDCs required 5– 6 days of culture with IL-4 and GM-CSF to maximally obtain this function, which is the time period generally used for obtaining DCs from blood precursors (37). In summary, we describe two distinct cross-presentation pathways that are used by human DCs, proteasome dependent and TPPII dependent, which can be specifically targeted by Ag formulation. Understanding the mechanisms used by human DCs in the cross-presentation of cancer Ag will allow the rational design of tumor vaccines for more effective CTL induction in cancer patients.

Acknowledgments We thank Prof. M. Pfreundschuh (Universita¨tsklinikum des Saarlandes, Homburg, Germany), for providing mAb against TPPII; and Dr. H. Bernhard (Technical University of Munich, Klinikum Rechts der Isar, Munich, Germany), for providing HLA-A2-restricted NY-ESO-1-specific CD8⫹ T cell clones.

1259

14.

15.

16.

17.

18.

19.

20.

21. 22.

23.

Disclosures

24.

E.M., H.B., and D.A. are employed by CSL Limited, whose potential product was studied in the present work. None of the other authors have a financial interest to declare.

25.

References 1. Villadangos, J. A., and P. Schnorrer. 2007. Intrinsic and cooperative antigenpresenting functions of dendritic-cell subsets in vivo. Nat. Rev. Immunol. 7: 543–555. 2. Hung, K., R. Hayashi, A. Lafond-Walker, C. Lowenstein, D. Pardoll, and H. Levitsky. 1998. The central role of CD4⫹ T cells in the antitumor immune response. J. Exp. Med. 188: 2357–2368. 3. Ridge, J. P., F. Di Rosa, and P. Matzinger. 1998. A conditioned dendritic cell can be a temporal bridge between a CD4⫹ T-helper and a T-killer cell. Nature 393: 474 – 478. 4. Schoenberger, S. P., R. E. Toes, E. I. van der Voort, R. Offringa, and C. J. Melief. 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 393: 480 – 483. 5. Hunder, N. N., H. Wallen, J. Cao, D. W. Hendricks, J. Z. Reilly, R. Rodmyre, A. Jungbluth, S. Gnjatic, J. A. Thompson, and C. Yee. 2008. Treatment of metastatic melanoma with autologous CD4⫹ T cells against NY-ESO-1. N. Engl. J. Med. 358: 2698 –2703. 6. Davis, I. D., W. Chen, H. Jackson, P. Parente, M. Shackleton, W. Hopkins, Q. Chen, N. Dimopoulos, T. Luke, R. Murphy, et al. 2004. Recombinant NYESO-1 protein with ISCOMATRIX adjuvant induces broad integrated antibody and CD4⫹ and CD8⫹ T cell responses in humans. Proc. Natl. Acad. Sci. USA 101: 10697–106702. 7. Schnurr, M., Q. Chen, A. Shin, W. Chen, T. Toy, C. Jenderek, S. Green, L. Miloradovic, D. Drane, I. D. Davis, et al. 2005. Tumor antigen processing and presentation depend critically on dendritic cell type and the mode of antigen delivery. Blood 105: 2465–2472. 8. Tacken, P. J., I. J. de Vries, R. Torensma, and C. G. Figdor. 2007. Dendritic-cell immunotherapy: from ex vivo loading to in vivo targeting. Nat. Rev. Immunol. 7: 790 – 802. 9. Shastri, N., S. Schwab, and T. Serwold. 2002. Producing nature’s gene-chips: the generation of peptides for display by MHC class I molecules. Annu. Rev. Immunol. 20: 463– 493. 10. Serwold, T., F. Gonzalez, J. Kim, R. Jacob, and N. Shastri. 2002. ERAAP customizes peptides for MHC class I molecules in the endoplasmic reticulum. Nature 419: 480 – 483. 11. Jungbluth, A. A., Y. T. Chen, E. Stockert, K. J. Busam, D. Kolb, K. Iversen, K. Coplan, B. Williamson, N. Altorki, and L. J. Old. 2001. Immunohistochemical analysis of NY-ESO-1 antigen expression in normal and malignant human tissues. Int. J. Cancer 92: 856 – 860. 12. Chen, Y. T., M. J. Scanlan, U. Sahin, O. Tureci, A. O. Gure, S. Tsang, B. Williamson, E. Stockert, M. Pfreundschuh, and L. J. Old. 1997. A testicular antigen aberrantly expressed in human cancers detected by autologous antibody screening. Proc. Natl. Acad. Sci. USA 94: 1914 –1918. 13. Nagata, Y., S. Ono, M. Matsuo, S. Gnjatic, D. Valmori, G. Ritter, W. Garrett, L. J. Old, and I. Mellman. 2002. Differential presentation of a soluble exogenous

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

38.

tumor antigen, NY-ESO-1, by distinct human dendritic cell populations. Proc. Natl. Acad. Sci. USA 99: 10629 –10634. Drane, D., C. Gittleson, J. Boyle, and E. Maraskovsky. 2007. ISCOMATRIX adjuvant for prophylactic and therapeutic vaccines. Expert Rev. Vaccines 6: 761–772. Saveanu, L., O. Carroll, Y. Hassainya, and P. van Endert. 2005. Complexity, contradictions, and conundrums: studying post-proteasomal proteolysis in HLA class I antigen presentation. Immunol. Rev. 207: 42–59. Rock, K. L., I. A. York, and A. L. Goldberg. 2004. Post-proteasomal antigen processing for major histocompatibility complex class I presentation. Nat. Immunol. 5: 670 – 677. Murphy, R., S. Green, G. Ritter, L. Cohen, D. Ryan, W. Woods, M. Rubira, J. Cebon, I. D. Davis, A. Sjolander, et al. 2005. Recombinant NY-ESO-1 cancer antigen: production and purification under cGMP conditions. Prep. Biochem. Biotechnol. 35: 119 –134. Maraskovsky, E., S. Sjolander, D. P. Drane, M. Schnurr, T. T. Le, L. Mateo, T. Luft, K. A. Masterman, T. Y. Tai, Q. Chen, et al. 2004. NY-ESO-1 protein formulated in ISCOMATRIX adjuvant is a potent anticancer vaccine inducing both humoral and CD8⫹ T-cell-mediated immunity and protection against NYESO-1⫹ tumors. Clin. Cancer Res. 10: 2879 –2890. Burgdorf, S., A. Kautz, V. Bohnert, P. A. Knolle, and C. Kurts. 2007. Distinct pathways of antigen uptake and intracellular routing in CD4 and CD8 T cell activation. Science 316: 612– 616. Tse, S. M., W. Furuya, E. Gold, A. D. Schreiber, K. Sandvig, R. D. Inman, and S. Grinstein. 2003. Differential role of actin, clathrin, and dynamin in Fc␥ receptor-mediated endocytosis and phagocytosis. J. Biol. Chem. 278: 3331–3338. Rock, K. L., and L. Shen. 2005. Cross-presentation: underlying mechanisms and role in immune surveillance. Immunol. Rev. 207: 166 –183. Geier, E., G. Pfeifer, M. Wilm, M. Lucchiari-Hartz, W. Baumeister, K. Eichmann, and G. Niedermann. 1999. A giant protease with potential to substitute for some functions of the proteasome. Science 283: 978 –981. Seifert, U., C. Maranon, A. Shmueli, J. F. Desoutter, L. Wesoloski, K. Janek, P. Henklein, S. Diescher, M. Andrieu, H. de la Salle, et al. 2003. An essential role for tripeptidyl peptidase in the generation of an MHC class I epitope. Nat. Immunol. 4: 375–379. Choppin, J., W. Cohen, A. Bianco, J. P. Briand, F. Connan, M. Dalod, and J. G. Guillet. 2001. Characteristics of HIV-1 Nef regions containing multiple CD8⫹ T cell epitopes: wealth of HLA-binding motifs and sensitivity to proteasome degradation. J. Immunol. 166: 6164 – 6169. Reits, E., J. Neijssen, C. Herberts, W. Benckhuijsen, L. Janssen, J. W. Drijfhout, and J. Neefjes. 2004. A major role for TPPII in trimming proteasomal degradation products for MHC class I antigen presentation. Immunity 20: 495–506. York, I. A., N. Bhutani, S. Zendzian, A. L. Goldberg, and K. L. Rock. 2006. Tripeptidyl peptidase II is the major peptidase needed to trim long antigenic precursors, but is not required for most MHC class I antigen presentation. J. Immunol. 177: 1434 –1443. Firat, E., J. Huai, L. Saveanu, S. Gaedicke, P. Aichele, K. Eichmann, P. van Endert, and G. Niedermann. 2007. Analysis of direct and cross-presentation of antigens in TPPII knockout mice. J. Immunol. 179: 8137– 8145. Benham, A. M., M. Gromme, and J. Neefjes. 1998. Allelic differences in the relationship between proteasome activity and MHC class I peptide loading. J. Immunol. 161: 83– 89. Princiotta, M. F., D. Finzi, S. B. Qian, J. Gibbs, S. Schuchmann, F. Buttgereit, J. R. Bennink, and J. W. Yewdell. 2003. Quantitating protein synthesis, degradation, and endogenous antigen processing. Immunity 18: 343–354. Norbury, C. C., L. J. Hewlett, A. R. Prescott, N. Shastri, and C. Watts. 1995. Class I MHC presentation of exogenous soluble antigen via macropinocytosis in bone marrow macrophages. Immunity 3: 783–791. Kovacsovics-Bankowski, M., and K. L. Rock. 1995. A phagosome-to-cytosol pathway for exogenous antigens presented on MHC class I molecules. Science 267: 243–246. Rodriguez, A., A. Regnault, M. Kleijmeer, P. Ricciardi-Castagnoli, and S. Amigorena. 1999. Selective transport of internalized antigens to the cytosol for MHC class I presentation in dendritic cells. Nat. Cell Biol. 1: 362–368. Houde, M., S. Bertholet, E. Gagnon, S. Brunet, G. Goyette, A. Laplante, M. F. Princiotta, P. Thibault, D. Sacks, and M. Desjardins. 2003. Phagosomes are competent organelles for antigen cross-presentation. Nature 425: 402– 406. Morel, S., F. Levy, O. Burlet-Schiltz, F. Brasseur, M. Probst-Kepper, A. L. Peitrequin, B. Monsarrat, R. Van Velthoven, J. C. Cerottini, T. Boon, et al. 2000. Processing of some antigens by the standard proteasome but not by the immunoproteasome results in poor presentation by dendritic cells. Immunity 12: 107–117. Luckey, C. J., G. M. King, J. A. Marto, S. Venketeswaran, B. F. Maier, V. L. Crotzer, T. A. Colella, J. Shabanowitz, D. F. Hunt, and V. H. Engelhard. 1998. Proteasomes can either generate or destroy MHC class I epitopes: evidence for nonproteasomal epitope generation in the cytosol. J. Immunol. 161: 112–121. Shortman, K., and S. H. Naik. 2007. Steady-state and inflammatory dendritic-cell development. Nat. Rev. Immunol. 1: 19 –30. Romani, N., S. Gruner, D. Brang, E. Kampgen, A. Lenz, B. Trockenbacher, G. Konwalinka, P. O. Fritsch, R. M. Steinman, and G. Schuler. 1994. Proliferating dendritic cell progenitors in human blood. J. Exp. Med. 180: 83–93. Ebert, L. M., Y. C. Liu, C. S. Clements, N. C. Robson, H. Jackson, J. L. Markby, N. Dimopoulos, B. S. Tan, E. F. Luescher, I. D. Davis, et al. 2009. An unusually long, naturally presented immunodominant epitope from NY-ESO-1 identified through vaccination reveals novel aspects of anti-tumor immunity. Clin. Cancer Res. in press.