Efficient Cross-Presentation by Heat Shock Protein 90-Peptide Complex-Loaded Dendritic Cells via an Endosomal Pathway

Mice

C57BL/6 (H-2^b), B6C3F1 (H-2^b/k), and TAP1^−/− mice were obtained from The Jackson Laboratory. HLA-A*2402/K^b transgenic mice were purchased from SLC Japan. All mice were kept in a specific pathogen-free mouse facility. Studies were performed according to institutional guidelines for animal use and care.

Cells

The mouse thymoma cell line EL4 and its E.G7 derivative (EL4 transfected with cDNA encoding OVA) were obtained from American Type Culture Collection (ATCC). N1 is an EL4 cell transfected with the nucleocapsid gene of vesicular stomatitis virus (VSV). An established CTL clone specific for the VSV8 epitope in the context of H-2K^b was restimulated with N1 every 7 days. The B3Z cell is a CD8⁺ T cell hybridoma specific for the OVA_258–265 epitope (SL8) in the context of H-2K^b. The KZO cell is a CD4⁺ T cell hybridoma specific for the OVA_247–265 (PL19) in the context of I-A^k. These hybridomas were a gift from Dr. N. Shastri (University of California, Berkeley, CA). RMA-S-A*2402 cells were RMA-S transfected with the gene encoding HLA-A*2402 (provided by Dr. H. Takasu, Dainippon-Sumitomo Pharmaceutical, Osaka, Japan). TG-3 cells are a methylcholanthrene-induced fibrosarcoma derived from the HLA-A*2402 transgenic mouse. TG-3-2B cells were TG-3 transfected with the gene encoding human survivin-2B. Survivin-2B_80–88-specific CTLs were restimulated with survivin-2B_80–88 peptide-pulsed splenocytes every 7 days. BMDC were generated from the femurs and tibiae of C57BL/6 or HLA-A*2402/K^b transgenic mice. The bone marrow was flushed out, and the leukocytes were obtained and cultured in complete RPMI 1640 with 10% heat-inactivated FCS and 20 ng/ml GM-CSF (Endogen) for 5 days. On day 3, fresh medium with GM-CSF was added to the plates for the day 5 cultures.

Plasmid construction

FLAG-tagged human survivin-2B cDNA was amplified from HeLa cells by RT-PCR method and constructed in the plasmid pCDNA3.1⁺ (Invitrogen Life Technologies). Primer pairs used for RT-PCR were 5′-CGGGATCCATGGGTGCCCCGACGTTGCCC-3′ and 5′-CCGCTCGAGATCCATGGCAGCCAGCTGCTC-3′ as forward and reverse primers, containing BamHI and XhoI sites, respectively. The purified PCR product was digested with BamHI and XhoI restriction enzymes, then ligated into digested pcDNA3.1⁺ plasmid. The sequence of the cDNA was confirmed with ABI Genetic analyzer PRIM 3100 (PerkinElmer).

Proteins and Abs

Purified human Hsp90 and recombinant human Hsp72 were purchased from StressGen Biotechnologies. Chicken OVA was purchased from Calbiochem. BSA, fucoidin, and α2 macroglobulin (α2M) were obtained from Sigma-Aldrich. mAbs anti-H-2K^b (clone AF6-88.5), anti-H-2D^b (clone 28-14-8), and anti-HLA-DR (L243) were purchased from BD Pharmingen. mAb anti-MHC class I (W6/32) was purchased from ATCC. mAb anti-HLA-A24 (C7709A2.6) was provided by Dr. P. G. Coulie (Christian de Duve Institute of Cellular Pathology, University of Louvain, Brussels, Belgium). mAb anti-HLA-A31 was established in our laboratory. Organelles were detected by confocal laser microscopy with specific Abs against KDEL (clone 10C3; StressGen Technologies) for ER, Rho B (Santa Cruz Biotechnology) for endosomes, and CD107a (LAMP-1) (clone 1D4B; BD Pharmingen) for lysosomes. Each Ab was labeled with Alexa Fluor 594 (Molecular Probes). mAb 25D1.16 specific for the K^b/OVA_257–264 complexes was provided by Dr. R. Germain (National Institutes of Health, Bethesda, MD). Anti-mouse CD16/CD32 Fc-block was purchased from BD Pharmingen.

Generation of Hsp90-peptide/protein complex in vitro

The following peptides were used (underlined sequences represent the precise MHC class I-binding epitope): survivin-2B_80–88 peptide (AYACNTSTL), survivin-2B_75–93 (GPGTVAYACNTSTLGGRGG) VSV 8 (RGYVYQGL), VSV-C (RGYVYQGLKSGNVSC), SL8 (SIINFEKL), and SL8C (SIINFEKLTEWTS). In vitro reconstitution was conducted as previously described (13). Hsp90 was mixed with a ¹²⁵I-labeled peptide in a 50:1 peptide to a protein molar ratio in 0.7 M NaCl containing sodium-phosphate buffer and heated at 45°C for 10 min then incubated at room temperature for 30 min. Hsp90 (10 μg) and OVA (10 μg) were mixed and incubated for 10 min at 45°C. The samples were then incubated for 30 min at room temperature. Free OVA was removed completely using a Microcon YM-100 (Millipore). In the case of Hsp72, peptides and Hsp72 were coincubated at 37°C in sodium phosphate buffer containing 1 mM ADP and 1 mM MgCl₂. Samples were analyzed by SDS-PAGE and staining, followed by autoradiography of the stained gel.

Vaccination and induction of CTL

Each HLA-A*2402/K^b transgenic mouse was immunized s.c. at the base of the tail four times at 1-wk intervals, with Hsp90 (50 μg) alone, the Hsp90 (50 μg)-survivin-2B_80–88 peptide (AYACNTSTL) (50 μg) complex, survivin-2B_80–88 peptide (50 μg) with IFA or CFA individually. One week after the last immunization, spleen cells were removed, cultured in vitro with irradiated (100 Gy) and survivin-2B_80–88 peptide-pulsed spleen cells for 5 days. Subsequently, the generation of survivin-2B_80–88 peptide-specific CTLs was evaluated in a ⁵¹Cr-release assay. The specificity of CTLs induced in an individual HLA-A*2402/K^b transgenic mouse was evaluated using RMA-S/A*2402 cells as targets in the presence or absence of the survivin-2B_80–88 peptide.

⁵¹Cr-release assay

The cytolytic activity of the induced CTL was determined by a standard 4 h-⁵¹Cr-release assay described earlier (19). To determine the MHC class I restriction in the cytotoxic assay, indicated amount of mAbs against HLA class I (W6/32), HLA-A24, HLA-A31, and HLA class II (L231) were added to each well. In the case of VSV or OVA system, 10 μg/ml mAbs against H-2K^b and H-2D^b were added to the each well.

Transplantation of tumor cells and immunotherapy

TG3-2B cells (5 × 10⁵) were intradermally transplanted into the right flank in HLA-A*2402/K^b transgenic mice on day 0. When average tumor diameter reached 5 mm, the mice were then treated with Hsp90 (50 μg) alone, the Hsp90 (50 μg)-survivin-2B_80–88 peptide (50 μg) complex, or survivin-2B_80–88 peptide (50 μg) emulsified in IFA via s.c. administration at the nape of the neck twice each week for 2 wk (on days 9, 13, 16, and 20). Control groups of mice were immunized with PBS. Tumor growth was recorded twice each week. Average diameters of the two axes were plotted so that therapeutic effects could be compared among the groups. Average tumor diameters on day 29 were statistically analyzed using the Mann-Whitney U test. In addition, mouse survival was monitored every other day. Statistical analyses for evaluating the survival advantages were performed using log-rank analysis. All the experiments were performed with 10 mice/group.

ELISPOT assay

The specificity of CTLs for the survivin-2B_80–88 peptide was also evaluated by IFN-γ ELISPOT assay. Splenic CD8⁺ T cells were isolated from mice immunized with Hsp90-survivin-2B_80–88 peptide, which were cultured for 5 days described above, with MACS (Miltenyi Biotec) using an anti-mouse CD8a mAb coupled with magnetic microbeads according to the manufacturer’s instructions. As target cells, RMA-S/A*2402 cells were cultured overnight at 26°C in RPMI 1640 supplemented with 10% FBS, 2.5 μg/ml β₂-microglobulin, 100 μg/ml survivin-2B_80–88 peptide, an irrelevant CMV peptide (QYDPVAALF), or without any peptide. Ninety-six-well ELISPOT plates (BD Biosciences) were coated with 5.0 μg/ml rat anti-mouse IFN-γ mAb and subsequently blocked with RPMI 1640 supplemented with 10% FBS for 2 h at room temperature. Then, 5 × 10³ CTLs and 1 × 10⁵ each target cells were added to the wells and cultured for 12 h at 37°C in RPMI 1640 in 10% FBS. The plates were then washed extensively and incubated with a 2.0 μg/ml biotinylated anti-mouse IFN-γ mAb, followed by pulsing with 0.5 μg/ml streptavidin-HRP. Positive spots were developed by adding 100 μl/well AEC Substrate Solution (BD Biosciences) and were counted using a Vision ELISPOT reader (Carl Zeiss).

Production of retrovirus and virus infection

The HLA-A*2402 cDNA was inserted into BamHI and NotI sites of the pMXs-puro retrovirus expression vector (26) (gift from Prof. T. Kitamura, University of Tokyo, Tokyo, Japan). High-titer retrovirus carrying HLA-A*2402 was produced in a transient retrovirus-packaging cell line PLAT-E (27) (gift from Prof. T. Kitamura). Briefly, PLAT-E cells were transfected with 5 μg of retrovirus vector plasmid with the FuGene HD Transfection Reagent (Roche Molecular Diagnostics). At 48 h after transfection, the supernatant was harvested as viral stock solution. For infection, immature BMDCs from TAP^−/− mice were incubated for 6 h with 6 ml of virus stock solution in the presence of 8 μg/ml hexadimethrine bromide (Sigma-Aldrich). Twenty-four hours postinfection, the mouse BMDCs were used for assays. The expression of HLA-A24 was confirmed by flow cytometry using an anti-HLA-A24 mAb (C7709A2.6) and 20–30% of DCs were HLA-A24 positive.

In vitro cross-presentation assay

Immature BMDCs (1 × 10⁴) from HLA-A*2402/K^b transgenic mice or HLA-A*2402-transduced BMDCs (1 × 10⁴) derived from TAP^−/− mice were pulsed with an Hsp90 (10 μg/ml)-survivin-2B_75–93 precursor peptide (100 ng/ml) complex generated in vitro, Hsp90 alone (10 μg/ml), survivin-2B_80–88 peptide (100 ng/ml), or survivin-2B_75–93 precursor peptide (100 ng/ml), and survivin-2B_80–88-specific CTLs (1 × 10⁵) were added to the cultures. In VSV or OVA system, immature BMDCs (1 × 10⁴) from C57BL/6 and TAP^−/− mice were pulsed with an Hsp90 (10 μg/ml)-peptide (10 μg/ml) or OVA protein (10 μg/ml) complex generated in vitro, and peptide-specific CTLs (1×10⁵) were added to the cultures. The assay was conducted in 200-μl volume in 96-well plates with AIM-V (Invitrogen Life Technologies) at 37°C for 20 h. Culture supernatants were harvested and tested for the presence of IFN-γ release by ELISA (Cytimmune Sciences). In the case of OVA system, SL8/K^b-specific B3Z or PL19/A^k-specific KZO responses were measured as the β-galactosidase activity induced upon ligand recognition. The B3Z (1 × 10⁵) or KZO (1 × 10⁵) and Ag-loaded BMDCs (5 × 10⁴) were added to each well and cultured overnight. The β-galactosidase activity was measured at the absorbance at 595 nm of the cleavage product of chlorophenol red β-pyranoside.

Immunocytological localization of exogenous Hsp90

Hsp90 was conjugated with Alexa Fluor 488 (Molecular Probes) according to the manufacturer’s instructions. Immature BMDCs were pulsed with the Alexa Fluor 488-labeled Hsp90 (20 μg/ml)-SL8C peptide (20 μg/ml) complex for 2 h. After incubation, cells were fixed with ice-cold acetone for 1 min, and then stained with an anti-Rho B Ab for detecting early endosomes, anti-KDEL mAb for ER, and anti-LAMP1 for late endosomes and lysosomes followed by Alexa Fluor 594-conjugated goat anti-rabbit IgG or anti-mouse IgG and visualized using a Bio-Rad MRC1024ES laser confocal scanning microscopy system (Bio-Rad). For detecting the intracellular localization of recycling MHC class I molecules and SL8-K^b complexes, the DCs were incubated with anti-mouse CD16/CD32 Fc-block to block nonspecific staining, and then costained with an Alexa Fluor 488-labeled Hsp90-SL8C peptide complex and Alexa Fluor 594-labeled anti-H-2K^b mAb (clone AF6-88.5) or Alexa Fluor 488-labeled 25D1.16 mAb and anti-organelle Abs conjugated with Alexa Fluor 594 and visualized by confocal laser microscopy.

Detection of fluorescent Hsp90-peptide complexes association with BMDCs and competition assay

BMDCs (1 × 10⁵) were preincubated with Hsp90, BSA, human α2M (Sigma-Aldrich), or fucoidin (Sigma-Aldrich) at indicated concentrations (25, 50, 100 μg) for 10 min at 4°C, then pulsed with the Alexa Fluor 488-labeled Hsp90 (5 μg)-SL8 peptide (10 μg) complex for 10 min at 4°C. The BMDCs were then washed twice with ice-cold PBS, fixed with ice-cold acetone for 1 min, and analyzed by flow cytometry and confocal laser microscopy.

Inhibition studies

For most inhibition studies, immature BMDCs (10⁵ cells/well) were first incubated in 0.1 ml of each drug for 2 h. Then Hsp90 (10 μg/ml)-SL8C (10 μg/ml) complex, SL8C (10 μg/ml) or OVA protein (40 μg/ml) of Ag was added to the wells (0.2 ml final volume) at the final concentration indicated in the continuous presence of inhibitors for 2 h. BMDCs were then washed three times and fixed with 0.05% glutaraldehyde (Sigma-Aldrich). Fixation was stopped by addition of 2 M l-lysine (Sigma-Aldrich) and cells were washed twice in PBS. Thereafter, SL8/K^b-specific B3Z or PL19/A^k-specific KZO were added to each well and cultured overnight. The β-galactosidase activity was measured at the absorbance at 595 nm. The inhibitors used were primaquine, chloroquine, lactacystin, leupeptin, and pepstatin (all obtained from Sigma-Aldrich except for primaquine (ICN Biomedicals)). LLnL, cathepsin S inhibitor (Z-FL-COCHO), cathepsin B inhibitor (Ac-Leu-Val-lysinal), and cathepsin L inhibitor (Z-FF-FMK) were purchased from Merck Biosciences.

Efficient induction of tumor peptide-specific CTL by immunizing Hsp90-peptide complex

We previously reported that survivin and its splicing variant survivin-2B are expressed abundantly in various types of tumor tissues and are suitable as target Ags for tumor immunotherapy. Subsequently, we identified an HLA-A24-restricted antigenic peptide, survivin-2B_80–88 (AYACNTSTL) recognized by CD8⁺ CTLs (28, 29). On the basis of these observations, we have started a phase I clinical study of survivin-2B_80–88 peptide vaccination for patients with advanced colorectal cancer. To establish an effective cancer vaccine, the development of an effective and safe adjuvant remains a high priority. Therefore, we examined whether HSPs could be a good candidate for a cancer vaccine adjuvant. First, we confirmed that the two major HSPs, Hsp70 and Hsp90, but not the control protein transferrin, were made complexed with survivin-2B_80–88 peptides in vitro (Fig. 1⇓A). Next, we tested the ability of Hsp70 and Hsp90 to generate CTL responses against associated peptides. HLA-A*2402/K^b transgenic mice are a well-established model for studying HLA-A*2402-restricted CTL epitopes and vaccine development (30). These mice contain transgenic chimeric human α1 and α2 domains of HLA-A*2402 and mouse α3 transmembrane and cytoplasmic domains of H-2K^b, which increase the level of T cell responsiveness. We immunized these transgenic mice with the Hsp70-survivin-2B_80–88 peptide complex or Hsp90-survivin-2B_80–88 peptide complex and examined the induction of peptide-specific CTL responses. As shown in Fig. 1⇓B, spleen cells of mice immunized with the Hsp90-survivin-2B_80–88 peptide complex showed significant cytotoxicity against survivin-2B_80–88-coated RMA-S-A*2402 cells, but not survivin-2B_80–88-noncoated RMA-S-A*2402 cells. This cytotoxic activity was almost same as from mice immunized with survivin-2B_80–88 emulsified in CFA (Fig. 1⇓G). In contrast, spleen cells of mice immunized with the Hsp70-survivin-2B_80–88 peptide complex (Fig. 1⇓D), Hsp90 alone (Fig. 1⇓C), Hsp70 alone (Fig. 1⇓E) or survivin-2B emulsified in IFA (Fig. 1⇓F) did not show much cytotoxicity against survivin-2B_80–88-coated RMA-S-A*2402. This cytotoxicity was significantly blocked by pretreatment of target cells with an anti-human MHC class I mAb W6/32 and an anti HLA-A24 mAb, but not with an anti-HLA-A31 mAb (Fig. 1⇓H). Similarly, pretreatment of effector cells with anti-CD8, but not with anti-CD4 mAb, significantly blocked the cytotoxicity against survivin-2B_80–88-coated RMA-S-A*2402 cells (data not shown). Furthermore, to determine whether this cytotoxic function was related to the frequency of peptide-specific T cells, ELISPOT assay was performed using splenocytes from each immunized mouse. Purified CD8⁺ T cells from splenocytes after 5 days in vitro stimulated with survivin-2B_80–88 were cocultured for 12 h with RMA-S/A*2402cells alone, RMA-S/A*2402 pulsed with survivin-2B_80–88, or those pulsed with an irrelevant CMVpp65-derived peptide (QYDPVAALF). The number of IFN-γ ELISPOTs produced by 5 × 10³ CD8⁺ T cells against 1 × 10⁵ RMA-S/A*2402 cells are shown in Fig. 2⇓. CD8⁺ T cells efficiently produced IFN-γ ELISPOTs in response to RMA-S/*A2402 cells pulsed with survivin-2B_80–88 peptide, but not in response to RMA-S/A*2402 cells alone or RMA-S*A2402 cells pulsed with the irrelevant CMVpp65-derived peptide (QYDPVAALF). These findings confirmed the fact that the CTLs were specific for survivin2B_80–88. Thus, we demonstrated that Hsp90 was a fairly good adjuvant for a T cell-mediated antitumor vaccine.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Efficient induction of peptide-specific CTL response in mice vaccinated with Hsp90-peptide complexes. A, Hsp90, Hsp70, and transferrin (Tf) were mixed with a ¹²⁵I-labeled peptide in a 50:1 peptide protein ratio in sodium-phosphate buffer. Samples were analyzed by SDS-PAGE and staining (data not shown), followed by autoradiography of the stained gel. Hsp90 and Hsp70 bound the survivin-2B_80–88 peptide efficiently, but not transferrin. B–G, HLA-A*2402/K^b-transgenic mice were immunized s.c. four times with the Hsp90-survivin-2B_80–88 peptide complex (B), Hsp90 alone (C), Hsp70-survivin-2B_80–88 peptide complex (D), Hsp70 alone (E), and survivin-2B_80–88 peptide emulsified with IFA (F) or CFA (G) individually. Spleen cells were removed 1 wk after the last immunization, cultured for 5 days with survivin-2B_80–88 peptides, and tested for cytotoxicity. Each line represents the specific lysis of target cells by spleen cells from one individual mouse. Target cells were RMA-S/A*2402 cells pulsed with the survivin-2B_80–88 peptide (▪) or without the peptide (□). H, Peptide-specific cytotoxicity was induced in MHC class I-restricted fashion. Indicated amount of mAbs against HLA class I (W6/32), HLA-A24 (C7709A2.6), HLA-A31, and HLA class II (L231) were added to each well.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Immunization with Hsp90-Ag peptide complex generated CTLs recognizing chaperoned peptide. Purified splenic CD8⁺ T cells from Hsp90-survivin-2B_80–88 peptide-immunized mice were stimulated with the survivin-2B_80–88 peptide in vitro. Responding CTLs were tested for IFN-γ ELISPOTs in response to RMA-S/A*2402 cells without peptide pulsing, pulsed with an irrelevant CMV peptide, or pulsed with the survivin-2B_80–88 peptide. The number of ELISPOTs produced by 5 × 10³ CTLs in response to 1 × 10⁵ RMA-S/A*2402 cells is shown.

Cross-presentation of Hsp90-chaperoned peptide in the context of MHC class I by BMDCs

Next, we tested whether in vitro reconstituted Hsp90-peptide complexes were taken up and associated peptides were presented in the context of MHC class I molecules by BMDCs. To monitor the MHC class I Ag-processing pathway, we used Hsp90 reconstituted in vitro with a precursor peptide of survivin2B_80–88, survivin-2B_75–93 (19 mer). The Hsp90-survivin-2B_75–93 precursor peptide complex was cocultured with BMDCs for 2 h and fixed, followed by incubation with a survivin-2B_80–88-specific CTL clone. The culture supernatant was assayed for the production of IFN-γ. As shown in Fig. 3⇓, the Hsp90-survivin-2B_75–93 precursor peptide complex was processed and presented by HLA-A*2402, and consequently recognized by the survivin-2B_80–88-specific CTL clone but not Hsp90 or the survivin-2B_75–93 precursor peptide alone. This phenomenon was also observed by the standard ⁵¹Cr-release assay (data not shown). These data suggested that the Hsp90-chaperoned precursor peptide was processed to an epitope within the cells with subsequent access to the MHC class I pathway, a process known as cross-presentation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Cross-presentation of Hsp90-chaperoned peptide by BMDCs. Survivin-2B_80–88, precursor survivin-2B_80–88 (19 mer), or Hsp90-precursor survivin-2B_80–88 complex was loaded to HLA-A*2402/K^b-transgenic mouse-derived BMDCs for 2 h and a survivin-2B_80–88-specific CTL clone was added. IFN-γ in the culture supernatant was measured by ELISA.

Vaccination of mice with Hsp90-peptide complex promotes antitumor effect

We then examined the efficacy of the Hsp90-based immunotherapy using the human tumor Ag survivin-2B as a surrogate Ag. To assess whether cross-presentation of the Hsp90-survivin-2B_80–88 complexes elicits antitumor effects in vivo, we developed TG3 with surrogate Ag human survivin-2B, TG3-2B. Although, the TG3-2B cell line is an artificial tumor model, we thought that it was necessary to examine whether the in vitro cross-presentation of Hsp90-survivin-2B complexes represents the survivin-2B-specific antitumor effects in vivo. TG3-2B cells (2 × 10⁵) were inoculated into the right hind limbs of HLA-A*2402/K^b transgenic mice and allowed to grow for 10 days (to around 5 mm in diameter). As shown in Fig. 4⇓A, on days 9, 13, 16, and 20, mice were administered PBS, the survivin-2B_80–88 peptide (50 μg), or the Hsp90 (50 μg)-survivin-2B_80–88 (50 μg) complex. As shown in Fig. 4⇓B, vaccination with Hsp90-survivin-2B_80–88 significantly inhibited the growth of tumors in comparison to control vaccinations with survivin-2B_80–88 or PBS, (vs survivin-2B, p = 0.034, vs PBS, p = 0.0007). We also evaluated the effects of the Hsp90-survivin-2B peptide complex in providing a survival benefit. As shown in Fig. 4⇓C, the results indicated that treatment with the Hsp90-survivin-2B peptide complex significantly increased the median survival time compared with the control mice (vs survivin-2B, p = 0.016, vs PBS, p = 0.001). Of note, 8 of 10 animals of the Hsp90-survivin-2B_80–88 complex-treated groups rejected the established tumors. Taken together, these results showed not only that the Hsp90-peptide complex induced strong CTL responses to the chaperoned peptide but also that these responses were sufficiently strong to generate therapeutic antitumor effects.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Hsp90-tumor Ag peptide complex induces strong antitumor effect. A, The protocol for immunotherapy is shown. B, A total of 5 × 10⁵ TG-3-2B cells were first injected intradermally into HLA-A*2402/K^b mice (10 animals/group). When mean tumor diameter reached 5 mm, mice were given the treatment with the Hsp90 (50 μg)-survivin-2B_80–88 (50 μg) complex, survivin-2B_80–88 (50 μg) emulsified with IFA alone or PBS twice a week. C, The remaining 10 mice in each group were observed for survival.

TAP-independent presentation of Hsp90-bound peptide in the context of MHC class I by BMDCs

To generalize the Hsp90-mediated cross-presentation, we used a well-characterized antigenic system, the VSV derived H-2K^b-restricted VSV8 dominant Ag. To test the peptide-binding capacity of purified human Hsp90, we used synthetic peptides VSV8 and a C-terminal extended version of VSV8, VSV-C (15 mer). First, we confirmed the in vitro generation of the Hsp90-VSV-C peptide complex (data not shown). Next, we tested whether in vitro-reconstituted Hsp90-VSV-C complexes were taken up and associated peptides were presented in the context of MHC class I molecules by immature BMDCs. The Hsp90-VSV-C peptide complex was cocultured with immature BMDCs for 2 h and fixed, followed by incubation with a VSV8-specific CTL clone. The culture supernatant was assayed for the production of IFN-γ. As shown in Fig. 5⇓A, the Hsp90-VSV-C peptide complex was processed and presented by H-2K^b, and consequently recognized by the VSV8-specific CTL clone but not Hsp90 or VSV-C alone. In the presence of an anti-H-2K^b mAb but not an anti-H-2D^b mAb, during the presentation assay, the presentation of VSV8 to the specific CTL clone was clearly abolished (Fig. 5⇓B). These data suggested that Hsp90 bound VSV-C peptide was processed to VSV8 within the cells with subsequent access to the MHC class I pathway.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Precursor peptides chaperoned by Hsp90 are cross-presented via a TAP-independent pathway. A, VSV8, the Hsp90-VSV-C complex, Hsp90, or VSV-C was loaded to BMDCs for 2 h and a VSV8-specific CTL clone was added. IFN-γ in the culture supernatant was measured by ELISA. B, The anti-H-2K^b mAb, but not the H-2D^b mAb, inhibited cytotoxicity mediated by the VSV-8-specific CTL clone. C, BMDCs from the TAP^−/− mouse could also process and present Hsp90-chaperoned VSV-C peptides efficiently as compared with BMDCs from the wild-type mouse. D, SL8 precursor peptide, SL8C was also processed and cross-presented by BMDCs via a TAP-independent pathway. E, HLA-A*2402-transduced TAP^−/− BMDCs were able to cross-present Hsp90-chaperoned precursor survivin-2B_80–88 peptides in association with HLA-A*2402 molecules.

Next, we investigated whether the Hsp90-mediated MHC class I pathway required functional TAP molecules. To test this, we used immature BMDCs derived from the TAP1^−/− mouse. Surprisingly, BMDCs from the TAP1^−/− mouse could also process and present the Hsp90-bound VSV-C peptide as efficiently as BMDCs from the wild-type mouse (Fig. 5⇑C). Taking advantage of the use of T cell hybridoma B3Z, we also tested another well-characterized H-2K^b-restricted OVA_257–264 Ag system. Hsp90 reconstituted in vitro with SL8C peptide (13mer, C-terminal extended version of SL8 (OVA_257–264)) was cocultured with BMDCs for 2 h, followed by incubation with an SL8-specific B3Z T cell hybridoma. As shown in Fig. 5⇑D, the Hsp90-SL8C peptide complex but not Hsp90 or SL8C alone was processed and presented by H-2K^b, and recognized consequently by the B3Z T cell hybridoma in a TAP-independent manner.

Furthermore, we have confirmed that Hsp90-survivin-2B_75–93 precursor peptide complex was processed and presented by TAP^−/− mice-derived BMDCs, which were retrovirally transduced with HLA-A*2402 cDNA, and consequently recognized by the survivin-2B_80–88-specific CTL clone but not Hsp90 or the survivin-2B_75–93 precursor peptide alone (Fig. 5⇑E). These data demonstrated that a TAP-independent pathway was used for Hsp90-mediated MHC class I presentation.

Hsp90-peptide complex interacts with bone marrow-derived immature BMDCs in a receptor-dependent fashion

Recent experiments demonstrated that HSPs are able to interact specifically with macrophages, DC, and B cells. To test the specific binding of Hsp90-peptide complex to immature BMDCs, we incubated BMDCs with the Alexa 488-labeled Hsp90-SL8C complex at 4°C to exclude endocytosis. Using FACS analysis, we observed specific binding of Hsp90-peptide complexes to the cell surface that could be competed for by unlabeled Hsp90, but not by BSA (Fig. 6⇓A). A 10-fold excess of unlabeled Hsp90 significantly inhibited the binding of Alexa 488-labeled Hsp90-SL8C complexes to BMDCs. No inhibition was observed using an excess of up to 100-fold of BSA. We then analyzed the competition experiments using the laser confocal microscopy. We have confirmed that after 1 h culture at 4°C, a temperature which blocks internalization, the Alexa 488-labeled Hsp90-SL8C complexes were found on the cell surface of immature BMDCs. Next, immature DCs were incubated with Alexa 488-labeled Hsp90-SL8C complexes, alone or in the presence of a 10-fold excess of unlabeled Hsp90 or BSA. As shown in Fig. 6⇓B, competition with unlabeled Hsp90 significantly reduced the cell surface binding, whereas unlabeled BSA did not affect the binding as compared with the labeled Hsp90-peptide complex alone. These data demonstrated the presence of a specific receptor for Hsp90 that was expressed on immature BMDCs. Recently, CD91 (the α2M receptor), LOX-1, and scavenger receptor class-A (SR-A) were identified as the HSP receptors of APCs. Therefore, we have done the competition assay using α2M and fucoidin, which are known for the ligands of HSP receptors CD91 (for α2M), LOX-1, and SR-A (for fucoidin). The results showed that the twenty-fold concentration of either α2 macroglobulin (Fig. 6⇓C) or fucoidin (Fig. 6⇓D) was not able to compete the cell surface binding of Alexa Fluor 488-labeled Hsp90-SL8 complex. Therefore, we concluded that Hsp90 receptor was different from CD91 and scavenger receptors such as LOX-1 or SR-A in our experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The Hsp90-peptide complex binds to immature BMDCs in a receptor-dependent fashion. A, BMDCs were incubated with the Alexa 488-labeled Hsp90-SL8C complex at 4°C to exclude endocytosis. Specific binding of the Hsp90-SL8C complex was observed and could be competed for by a 10-fold excess of unlabeled Hsp90. No inhibition was observed using an excess of up to 100-fold of BSA. B, BMDCs were incubated with unlabeled 50 μg/ml Hsp90 or 50 μg/ml BSA for 1 h on ice, then pulsed with 5 μg/ml Alexa 488-labeled Hsp90-SL8C complexes to DCs on ice for 1 h, washed with PBS, followed by acetone fixation and processing for confocal microscopy. C and D, The binding of the Hsp90-SL8C peptide complex to BMDCs was not inhibited by either the CD91 ligand α2M or the SR-A ligand fucoidin.

Endocytosed Hsp90-peptide complexes localize in the early endosome

Next, we were interested in addressing which compartments were involved in this processing and presentation of Hsp90-chaperoned precursor peptides within the BMDCs. We labeled human Hsp90-SL8C with Alexa Fluor 488. Labeled Hsp90-SL8C peptide complexes were then incubated with BMDCs for several time periods. After extensive washing to remove unbound proteins, cells were fixed with cold acetone and costained with an Ab against an organelle marker labeled with Alexa Fluor 594. Analysis by laser confocal microscopy of staining with the organelle marker revealed that the internalized Hsp90-SL8C complex localized in Rho B-positive early endosomes (Fig. 7⇓A) but not late endosomes/lysosomes or endoplasmic reticulum (ER) after 10, 30, and 60 min of endocytosis (shown for 60 min in Fig. 7⇓A). We traced Hsp90-SL8C complex after 90 and 120 min. It accumulated only in early endosomes and did not reach the stage of late endosomes/lysosomes (data not shown). In accordance with experiments using TAP 1^−/− BMDCs, the lack of accumulation in the ER seemed to indicate that the Hsp90-mediated MHC class I pathway was independent of TAP. We also examined whether Hsp90-peptide complex accumulation in the early endosome was temperature-dependent endocytosis. As expected, at 4°C, labeled Hsp90-peptide complex remained on the cell surface (Fig. 7⇓B), but internalization was evident after incubation at 37°C following a 10-min internalization period.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Intracellular localization of the Hsp90-SL8C peptide complex taken up by receptor-mediated endocytosis. A, Immature BMDCs were pulsed with Alexa 488-labeled Hsp90-SL8C complex for 2 h. After incubation, cells were fixed with cold acetone, stained with an anti-Rho B Ab (for early endosomes), anti-KDEL mAb (for ER), and anti-LAMP1 (for late endosomes and lysosomes) followed by Alexa 594-conjugated goat anti-rabbit IgG or anti-mouse IgG and visualized by laser confocal microscopy. B, Internalization of the Hsp90-peptide complex occurred via temperature-dependent endocytosis. BMDCs were treated either 4°C or at 37°C with 20 μg/ml Alexa 488-labeled Hsp90-SL8C complex for 10 min, washed with PBS, and fixed, then stained with the anti-Rho B Ab, followed by Alexa 594-conjugated goat anti-rabbit IgG and analyzed by laser confocal microscopy.

Extracellular Hsp90-peptide complexes and recycling MHC class I molecules are colocalized within early endosomes in BMDCs

To investigate in which compartment the Hsp90-chaperoned antigenic peptides bound to MHC class I molecules, we stained H-2K^b molecules and exogenous Hsp90-SL8C complexes. After 20 min of endocytosis, Alexa Fluor 488-labeled Hsp90-SL8C complexes colocalized with endocytosed H-2K^b molecules in the early endosome (Fig. 8⇓A). This finding suggested that Hsp90-bound peptides might be transferred to MHC class I molecules in the early endosome where recycled MHC class I molecules from the plasma membrane are available. The peptide-MHC class I complexes generated in the early endosome would then be transported to the cell surface of the BMDCs, where specific CTLs recognize them.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

The Hsp90-peptide complex traffics to an endosome, where the precursor peptide might be processed, followed by the formation of a peptide-MHC class I complex. A, BMDCs were treated at either 4°C or at 37°C with the 20 μg/ml Alexa 488-labeled Hsp90-SL8C complex for 10 min, washed with PBS, and fixed, then stained using Alexa 594-conjugated anti-H-2K^b mAb. B, BMDCs were incubated with Hsp90-SL8C peptide complexes for 1 h, then fixed with ice-cold acetone for 1 min. DCs were incubated with mAb 2.4G2 to block FcR and then costained with Alexa 488-conjugated mAb 25D1.16 (for H-2K^b-SL8 complex) and Alexa 594-conjugated anti-Rho B Ab or anti-KDEL mAb.

Early endosomes are the compartment where Hsp90-bound precursor peptides are processed and transferred onto subcellular MHC class I molecules

To investigate in which compartment Hsp90-bound precursor peptides are processed and subsequently transferred onto MHC class I molecules, we used mAb 25D1.11 because this mAb recognizes SL8 peptide-H-2K^b complexes (31). Hsp90-SL8C peptide complexes were pulsed onto BMDCs, subsequently fixed with acetone, and stained with mAb 25D1.11 labeled with Alexa Fluor 488 and anti-Rho B or anti-KEDL Abs coupled with Alexa Fluor 594. Consequently, we clearly observed that mAb 25D1.11 was detected only in the early endosomes and not in the ER (Fig. 8⇑B). This fact indicated that Hsp90-bound precursor peptides were processed and transferred onto MHC class I within early endosomes, suggesting that recycling MHC class I molecules are required for efficient presentation of Hsp90-chaperoned peptides.

Kinetics of cross-presentation of Hsp90-chaperoned peptide by BMDCs

We evaluated the cross-presentation kinetics of the Hsp90-chaperoned precursor peptide by BMDCs. BMDCs were pulsed with the Hsp90-SL8C complex, sampled between 0 and 120 min and fixed with glutaraldehyde to terminate further Ag uptake and processing. After fixation, BMDCs were cocultured with B3Z T cell hybridoma. The presentation of Hsp90-chaperoned peptides was detected after a 10-min pulse (Fig. 9⇓A), indicating that induction of immune responses can be achieved very rapidly. This is very important for Hsp90 as a danger signal. In contrast, presentation of free SL8C peptide was barely detectable within 2 h. In addition, to confirm that the Hsp90-mediated cross-presentation followed a TAP-independent pathway, we tested the effect of treating BMDCs with the proteasome inhibitor, lactacystin and LLnL. As expected, these agents did not affect the Hsp90-mediated cross-presentation by BMDCs (Fig. 9⇓, B and C). Therefore, class I-presented peptides were generated from the Hsp90-precursor peptide complex through the endosomal pathway.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Membrane recycling and vacuolar acidification, but not proteasomal processing is required for cross-presentation of Hsp90-peptide complexes. A, Cross-presentation of Hsp90-chaperoned peptides by BMDC is very rapid response. B–D, BMDCs were preincubated with (B) lactacystin, (C) LLnL, (D) primaquine, or (E) chloroquine at 37°C for 2 h, then pulsed with Hsp90-SL8C complexes, SL8 peptide, or OVA protein for 2 h. The DCs were then fixed, washed, and cultured overnight with B3Z cells. The β-galactosidase activity was measured at the absorbance at 595 nm.

Hsp90-chaperoned peptides are transferred to recycling MHC class I molecules in early endosomes

Recycling of endocytosed MHC class I molecules back to the cell surface has been observed (32). Some of the recycling MHC class I molecules can be loaded into early endosomes with peptides derived from endocytosed molecules. Therefore, to confirm whether this presentation really used the recycling MHC class I molecules, we treated BMDCs with primaquine, which blocks the membrane recycling pathway. BMDCs incubated in the presence of this drug could not present the Hsp90-chaperoned SL8C (13 mer)-derived SL8 peptide (Fig. 9⇑D). This result indicated that Hsp90-chaperoned precursor peptides or processed peptides could enter into recycling endosomes and be transferred onto recycling MHC class I molecules, which went back to the cell surface, resulting in the stimulation of B3Z T cell hybridoma. Furthermore, to analyze the involvement of vacuolar acidification of endosomal compartments, BMDCs were incubated with Hsp90-SL8C precursor peptide complexes in the presence of chloroquine, a known inhibitor of acidification of endosomal compartments. Chloroquine treatment resulted in strong inhibition of the Hsp90-mediated presentation, without affecting SL8 peptide presentation, showing that acidification of endosomal compartments was necessary for Hsp90-chaperoned precursor peptide processing (Fig. 9⇑E).

Hsp90-chaperoned peptides are processed by endosomal protease

We used protease inhibitors to investigate how proteolytic processes were involved in this Hsp90-mediated TAP-independent cross-presentation pathway. We found that, in wild-type BMDC, a broadly active cysteine protease inhibitor, leupeptin, almost completely inhibited the cross-presentation of Hsp90-SL8C precursor peptide complexes (Fig. 10⇓A). In contrast, the aspartic protease inhibitor pepstatin did not affect the cross-presentation (Fig. 10⇓C). The concentration of leupeptin or pepstatin used was sufficient to inhibit cysteine proteases or aspartic protease because it completely blocked the presentation of soluble OVA on MHC class II molecules detected by I-A^k-specific CD4⁺ T cell hybridoma KZO (Fig. 10⇓, B and D). These results indicated that cysteine proteases are required for the Hsp90-mediated cross-presentation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 10.

The Hsp90-chaperoned precursor peptide is processed by cysteine/serine proteases in the endosomes. A–F, DCs were preincubated with (A and B) leupeptin, (C and D) pepstatin, or (E and F) a cathepsin S inhibitor at 37°C for 2 h, then pulsed with Hsp90-SL8C complexes, SL8 peptide, or OVA protein for 2 h. The DCs were then fixed, washed, and cultured overnight with B3Z or KZO cells. The β-galactosidase activity was measured at the absorbance at 595 nm.

We next studied the role of cathepsins in the Hsp90-mediated vacuolar cross-presentation. Cathepsins S, B, and L are known to be the major cysteine proteases in endocytic compartments. We therefore examined the roles of various cathepsins in this pathway. A cathepsin B- or cathepsin L-specific inhibitor did not affect Hsp90-mediated cross-presentation (data not shown), whereas a cathepsin S inhibitor clearly blocked cross-presentation (Fig. 10⇑E), as well as the presentation of soluble OVA on MHC class II molecules detected by KZO T cell hybridoma (Fig. 10⇑F). Cathepsin S is preferentially expressed in APCs, including DCs, macrophages, and B cells within endocytic compartments. Therefore, our data indicated that cathepsin S was a critical enzyme in TAP-independent Hsp90-mediated cross-presentation to MHC class I molecules and that antigenic precursor peptides were indeed processed to epitope peptides, followed by association with MHC class I molecules in endosomal compartments.

Hsp90-protein Ag complex is cross-presented by BMDCs

Lastly, we evaluated cross-presentation of the in vitro-generated Hsp90-OVA protein complex. In vitro generation of Hsp90-OVA complexes was performed and confirmed according to the method described in Materials and Methods. BMDCs were pulsed with Hsp90 alone, free OVA, or a complex of them generated in vitro for 2 h at 37°C, then fixed, washed, and cultured with B3Z CD8⁺ T cell hybridoma. Hsp90-OVA elicited strong B3Z responses, while Hsp90 or OVA alone did not induce a B3Z response (Fig. 11⇓). Thus, Hsp90-chaperoned protein Ag as well as peptide is efficiently cross-presented by BMDCs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 11.

The Hsp90-OVA complex is presented by BMDCs through the MHC class I pathway. BMDCs were pulsed with Hsp90 alone, OVA alone, a complex of them or SL8 for 2 h at 37°C, then fixed, washed, and cultured overnight with B3Z. The β-galactosidase activity was measured at the absorbance at 595 nm.