Induction of Endoplasmic Reticulum–Endosome Fusion for Antigen Cross-Presentation Induced by Poly (γ-Glutamic Acid) Nanoparticles

Cancer immunotherapy is a highly attractive potential approach toward eradicating malignant diseases by amplifying and activating tumor immunity that can distinguish cancer cells from normal cells and eliminate them from the body (1). For this promising cancer treatment strategy, efficient delivery of the tumor-associated Ag to the APCs and strong cross-presentation of spliced peptides on MHC class I molecules on APCs are crucial (2). Efficient MHC class I presentation enhances the production of tumor-associated Ag-specific CTL, the major effector cells in tumor immunity (3). Our studies are aimed at promoting the effective application of self-assembled nanoparticles (NPs) using a biodegradable polymer derived from a natto mucilage, poly (γ-glutamic acid) (γ-PGA), as a vaccine carrier (4).

A technique for preparing unique NPs [poly (γ-glutamic acid)-based NPs (γ-PGA NPs)] using amphiphilic γ-PGA (γ-PGA–l-phenylalanine ethyl ester [L-PAE]) was recently reported in which L-PAE is introduced as a hydrophobic residue into the α-position group carboxyl of γ-PGA (5–7). γ-PGA NPs entrapping antigenic proteins enhance Ag-specific cellular immunity. γ-PGA NPs can deliver various types of antigenic proteins to APCs and elicit potent immune responses based on Ag-specific CTL (8, 9). Further, γ-PGA NPs induce strong antitumor effects without inducing acute toxicity (10).

An important feature of γ-PGA NPs, which differs from other vaccine carriers, is the strong enhancement of MHC class I Ag presentation (10, 11). Ags presenting on MHC class I molecules are generally known as endogenous Ags (12). In contrast, many exogenous Ags introduced by the endocytosis pathway are presented on MHC class II molecules after lysosomal degradation without entry into the cytoplasm (13). γ-PGA NPs, however, can induce an efficient MHC class I Ag presentation of exogenous Ags (cross-presentation), producing strong antitumor immunity in vivo (10). Therefore, analysis of the unique properties of efficient γ-PGA NP-induced cross-presentation will be useful toward designing novel vaccine carriers that induce Ag-specific CTL in vivo.

To investigate the mechanism of cross-presentation induced by γ-PGA NPs, we analyzed the intracellular behavior of γ-PGA NPs and loaded Ags in dendritic cells (DCs) and revealed a remarkable pathway for cross-presentation of Ags loaded in γ-PGA NPs.

The DC2.4 cell line, previously characterized as an immature C57BL/6 DC line (H-2^b) (14), was kindly provided by Dr. Kenneth. L. Rock (Department of Pathology, University of Massachusetts Medical School, Worcester, MA) and was maintained in RPMI 1640 medium containing 10% FBS, 50 μM 2-ME, and antibiotics. CD8-OVA1.3 cells, a T–T hybridoma against OVA_257–264/H-2K^b complex (15, 16), were kindly provided by Dr. Clifford V. Harding (Case Western Reserve University, Cleveland, OH) and maintained in DMEM containing 10% FBS, 50 μM 2-ME, and antibiotics. OT4H.1D5 cells, T–T hybridomas against OVA_265–277/I-A^b complex (17), were kindly provided by Dr. Judith A. Kapp (Emory University, Atlanta, GA) and maintained in RPMI 1640 medium containing 10% FBS, 1 mM pyruvate, 2 mM l-glutamine, 50 μM 2-ME, and antibiotics. Cells with at least 80% viability under any reagents were used for the experiments. Cell viability was determined using a WST-8 assay kit (Dojindo Molecular Technology, Kumamoto, Japan).

To isolate bone marrow-derived DCs (BMDCs), we used 6–8-wk-old female C57BL/6 mice (Japan SLC, Hamamatsu, Japan) and B6.129S2-Tap1 < tm1Arp > /J mice [TAP(^−/−) mice; The Jackson Laboratory, Bar Harbor, ME]. All animal experiments were performed in accordance with the Osaka University guidelines for the welfare of experimental animals.

BMDCs were prepared as described previously (10, 11). Briefly, bone marrow cells were flushed from the femurs and tibias of mice using PBS and then cultured in RPMI 1640 containing 10% FBS, 40 ng/ml recombinant murine GM-CSF (PeproTech, Rocky Hill, NJ), 50 μM 2-ME, and antibiotics. The culture media was replenished at 3 and 6 d. Nonadherent cells were harvested as immature BMDCs.

The γ-PGA–L-PAE copolymers were synthesized by the method of Akagi et al. (5–7). Briefly, 4.7 mmol γ-PGA (Wako Pure Chemical Industries, Osaka, Japan) was incorporated with 4.7 mmol L-PAE (Sigma-Aldric, St. Louis, MO) in 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (Dojindo Laboratories, Kumamoto, Japan). For FITC labeling of γ-PGA–L-PAE, 5-(aminoacetamido) fluorescein (Invitrogen, Carlsbad, CA) was coupled with carboxylic acids of γ-PGA–L-PAE.

Ag-loaded γ-PGA NPs were prepared as described previously (5–7). Briefly, 10 mg/ml γ-PGA–PAE in DMSO (Sigma-Aldrich) was mixed with an equal volume of Ag solution in saline (2 mg/ml OVA [Sigma-Aldrich], 2 mg/ml FITC-labeled OVA [FITC-OVA; Invitrogen], and 0.5 mM Au NPs [Kyoto Nano Chemical, Kyoto, Japan]). The preparation was washed two times using the following procedure. The mixture was centrifuged at 18,000 × g for 20 min, resuspended in distilled water, and centrifuged at 18,000 × g for 20 min. The resulting pellet was resuspended in PBS. The amount of OVA entrapped in the γ-PGA NPs was estimated using a Bio-Rad DC Protein Assay (Bio-Rad, Hercules, CA) after dissociation of the γ-PGA NPs using 2% SDS solution. The amount of γ-PGA in γ-PGA NPs was assessed by freeze-drying the γ-PGA NPs. Ag-encapsulated γ-PGA NPs consist of ∼15% (w/w) Ag and 85% (w/w) γ-PGA. The γ-PGA NPs were confirmed to be ∼200-nm uniformly sized NPs, as determined by dynamic light scattering using a Zetasizer 3000HS (Malvern Instruments, Worcestershire, U.K.).

DC2.4 cells were seeded on 24-well plates at a density of 1.0 × 10⁶ cells/well and cultured for 24 h at 37°C. The cells were pretreated at 37°C for 30 min with culture media containing 0–40 μM cytochalasin D (Sigma-Aldrich) or 0–5 μM amiloride (Sigma-Aldrich). After incubation, samples in culture media were added to the cells and incubated for an additional 1 h (samples; 10 μg protein/ml FITC-OVA in γ-PGA NPs [γ-PGA/FITC-OVA], 50 μg/ml fluorescent polystyrene beads [Invitrogen], and 25 μg/ml FITC-conjugated transferrin [Invitrogen]). After treatment, the cells were washed in ice-cold PBS containing 0.05% NaN₃, and the samples absorbed to the surface were removed using 0.25% trypsin (Invitrogen). Culture medium was added to stop the trypsin activity and detach the cells. The cell suspension was centrifuged at 800 × g and resuspended in ice-cold PBS containing 0.05% NaN₃. Fluorescence was analyzed on an FACSCalibur flow cytometer (BD Biosciences, San Jose, CA). Dead cells were detected using 7-aminoactinomycin D (Invitrogen), and live cells were analyzed using FlowJo software (Tree Star, Ashland, OR).

BMDCs isolated from C57BL/6 mice or TAP(^−/−) mice were seeded in a 96-well flat-bottom culture plate at a density of 2 × 10⁵ cells/well and cultured for 12 h at 37°C. Each well was washed twice with PBS, and then the cells were pulsed with 50 μg protein/ml γ-PGA NP/OVA, 20 μg/ml OVA-derived MHC class I epitope peptide (OVA_257–264; SIINFEKL; Sigma-Aldrich), or 10 mg/ml soluble OVA solution. After an additional 3-h incubation at 37°C, the cells were washed twice with PBS and fixed with 0.05% glutaraldehyde. The cells were then cocultured with 2 × 10⁴ CD8-OVA1.3 cells (for MHC class I presentation) or 2 × 10⁵ OT4H.1D5 cells (for MHC class II presentation) for 24 h at 37°C. The response of stimulated CD8-OVA 1.3 or OT4H.1D5 cells was assessed by determining the amount of IL-2 released into the culture medium using a murine IL-2 ELISA KIT (PeproTech). For the inhibition assay, BMDCs were pretreated with 10 μM MG-132 (Peptide Institute, Osaka, Japan), 1 μM epoxomicin (Peptide Institute), 40 μM leupeptin (BIOMOL Research Laboratories, Plymouth Meeting, PA), or 0–0.156 μM Pseudomonas aeruginosa exotoxin A (ExoA; Merck, Darmstadt, Germany) for 1 h at 37°C before sample treatment. In this case, the sample solution also contained each inhibitor.

BMDCs isolated from C57BL/6 mice were seeded on eight-well chamber slides at a density of 1 × 10⁵ cells/well and cultured for 24 h at 37°C. Each well was washed twice with PBS, and the cells were pulsed with 20 μg protein/ml γ-PGA NP/FITC-OVA or 20 μg protein/ml FITC-labeled γ-PGA (FITC–γ-PGA) NP/OVA for 10 min. The cells were washed twice with PBS and incubated for an additional 30, 60, or 120 min in culture media at 37°C. Cells were washed with PBS, fixed with 4% paraformaldehyde, and permeabilized with PBS containing 0.3% Triton X-100 and 3% BSA. The blocking procedure was performed with PBS containing 10% goat serum, 0.03% Triton X-100, and 3% BSA for 2 h at room temperature. To detect the early endosome (EE), late endosome, and endoplasmic reticulum (ER), the cells were incubated with rabbit anti-EE Ag 1 (EEA1) Ab (ab74906; Abcam, Cambridge, MA), rabbit anti-Rab7 Ab (ab50313; Abcam), rabbit anti-KDEL Ab (ab50601; Abcam), and rabbit anti-calnexin Ab (NB100-65569; Novus Biologicals, Littleton, CO), respectively, for 1 h at room temperature. To detect EE and recycling endosomes, the cells were incubated with rat anti-transferrin receptor (TfR) Ab (ab22391; Abcam) for 1 h at room temperature. After washing with PBS, the cells were reacted with Alexa 594-conjugated anti-rabbit IgG Ab (A11037; Invitrogen) or Alexa 594-conjugated anti-rat IgG Ab (A21209; Invitrogen) for 1 h at room temperature. The samples were embedded with Prolong Gold anti-fade reagent with DAPI (Invitrogen) and analyzed by confocal laser scanning microscopy (CLSM; Leica Microsystems) under a Plan Apo BL 63/1.4 numeric aperture oil-immersion objective. Image quantification was performed using BioImage XD image analysis software (http://www.bioimagexd.net). To quantify colocalization, the percent of the colocalized thresholds between green and red channels in the total threshold of green channels was calculated using micrographs with at least 100 cells taken from 11–19 different areas.

BMDCs isolated from C57BL/6 mice were seeded on 24-well plates at a density of 1.5 × 10⁶ cells and cultured for 24 h at 37°C. The culture dish was washed twice with PBS, and then the cells were pulsed with 300 μg/ml γ-PGA NPs containing Au NPs for 10 min at 37°C. The cells were washed twice with PBS and incubated for an additional 2 h in culture media at 37°C. Transmission electron microscopy (TEM) observation was performed by the Hanaichi UltraStructure Research Institute (Tokyo, Japan). The cells were fixed in HEPES-buffered 2% glutaraldehyde, and subsequently postfixed in 2% osmium tetroxide for 3 h in an ice bath. The specimens were dehydrated in a graded series of ethanol and embedded in epoxy resin. Ultrathin sections were obtained by ultramicrotomy, stained with uranyl acetate for 10 min and modified Sato’s lead solution for 5 min, and submitted to TEM observation (JEM-2000EX; JEOL,Tokyo, Japan). For observation of cellular uptake (Fig. 1C), cells were pulsed with 300 μg/ml γ-PGA NPs/Au for 10 min at 37°C, washed twice with PBS, and fixed immediately.

To analyze the pathway of γ-PGA NP-mediated cell entry, cellular uptake of γ-PGA/FITC-OVA under the treatment of phagocytosis/macropinocytosis inhibitor cytochalasin D (18) or macropinocytosis inhibitor amiloride (19) was assessed by flow cytometry. These two pathways are involved in NP-mediated cellular uptake (20, 21); therefore, they are potential pathways of γ-PGA NP-mediated cell entry. The cell entry of γ-PGA NP/FITC-OVA was clearly inhibited by both cytochalasin D and amiloride (Fig. 1A, 1B). Similar results were obtained with 1-μm diameter polystyrene microbeads, which are used as a marker of phagocytosis (22). In contrast, the cell entry of the clathrin-dependent endocytosis marker transferrin (23) was not inhibited by cytochalasin D or amiloride, indicating that these reagents specifically inhibited actin-dependent phagocytosis/macropinocytosis. Moreover, TEM analysis revealed that Au NPs encapsulated by γ-PGA NPs (γ-PGA NP/Au) was entrapped in the endosome vesicle (Fig. 1C). These results suggest that γ-PGA NPs enter the cell through phagocytosis and/or macropinocytosis.

Ag presentation efficiency induced by γ-PGA NPs was analyzed under inhibition of cytoplasmic proteasomes and lysosomal enzymes, which are involved in the classical Ag degradation pathway for MHC class I (12) and II (13) Ag presentation, respectively (Fig. 2). MG-132 (24) and epoxomicin (25) were used as proteasome specific inhibitors, and leupeptin (26) was used as a lysosomal protease inhibitor.

For the MHC class I presentation, the γ-PGA NP/OVA-induced response was clearly suppressed by the proteasome inhibitors (Fig. 2A, 2B), but not by the leupeptin treatment (Fig. 2C). Moreover, MHC class I presentation was not observed in TAP(^−/−) BMDC (Fig. 2D). As expected, the response induced by the positive control MHC class I epitope peptide (OVA_257–264) was retained under all conditions (Fig. 2A–D). These results suggest that an exogenous Ag encapsulated in γ-PGA NPs is cross-presented on MHC class I molecules through a classical endogenous Ag presentation pathway via cytoplasmic proteasomes and TAP.

For the MHC class II presentation, the γ-PGA NP/OVA-induced response was partially inhibited only by MG-132 treatment (Fig. 2E) and not at all by epoxomicin and leupeptin treatments (Fig. 2F, 2G). No inhibition was observed in TAP(^−/−) BMDCs (Fig. 2H). In contrast, the response of soluble OVA, a positive control for lysosomal degradation, was inhibited by leupeptin treatment (Fig. 2G). No inhibitory effects were observed, however, under the proteasome inhibitors or in TAP(^−/−) BMDCs (Fig. 2E, 2G, 2H). This finding suggests that the MHC class II presentation machinery via γ-PGA NPs is complex and might be different from the lysosomal degradation of exogenous Ags, a well-known classical MHC class II presentation pathway (27).

The ER is an essential intracellular organelle for cross-presentation on MHC class I molecules (28–33). Therefore, it is important to determine the transport mechanism of the Ag to the ER. To analyze ER translocation of γ-PGA NPs and the loaded Ag, we assessed the subcellular localization of FITC-OVA and FITC–γ-PGA using CLSM. As organelle markers, we performed immunostaining for EEA1 (EE) (34), Rab7 (late endosome and partially EE) (35, 36), TfR (EE and recycling endosome) (37), a KDEL motif containing Bip/Grp78 and Grp94 proteins (ER) (38), and the ER chaperone calnexin (ER) (39). Quantification of the colocalization area revealed that 70% of FITC-OVA was colocalized with EEA1 at 30 min after pulsing γ-PGA NP/FITC-OVA, but it was significantly decreased at 120 min (Fig. 3A–C, 3AX). In contrast, colocalization with Rab7 increased with time, and colocalization with TfR was stably maintained at ∼70% (Fig. 3D–I, 3AY, 3AZ). These results suggest that FITC-OVA was captured in a partially acidic endosomal environment. Importantly, colocalization with KDEL and calnexin increased with time: ∼60% of FITC-OVA accumulated in the KDEL(+) or calnexin(+) ER compartment (Fig. 3J–O, 3BA, 3BB). Together, these results indicate that at least 30% FITC-OVA was localized in ER-like endosomes containing both TfR and ER markers at 120 min. Similar results were obtained for FITC–γ-PGA (Fig. 3V–AJ, 3BC–BG), and ∼70 and 60% of FITC–γ-PGA accumulated in the TfR(+) and ER marker(+) organelles, respectively (Fig. 3AB–AJ, 3BE–BG). This finding indicated that at least 30% of FITC–γ-PGA accumulated in ER-like endosomes. These phenomena were not observed without primary Abs (Fig. 3P–U, 3AK–AP) or with the unlabeled γ-PGA NP/OVA (Fig. 3AQ–AW), suggesting that both γ-PGA and encapsulated Ags were translocated from the EE to the ER-like endosome within 120 min.

To analyze at this time point in more detail, we used TEM to directly observe small γ-PGA NP/Au in the cells. At 120 min after pulsing γ-PGA NP/Au, ultrathin sections were prepared, and we observed the localization of the γ-PGA NP/Au. Consistent with the CLSM results, most γ-PGA NP/Au was entrapped in the round-shaped endosomal vesicles, and we did not detect them in ER or other organelles (Fig. 4A, 4B). Moreover, most Au NPs were entrapped in γ-PGA NPs, suggesting that both the encapsulated Ags and γ-PGA NPs reached the endosomal vesicles together. The CLSM and TEM findings strongly suggested that γ-PGA NPs localized in endosomal vesicle-like organelles (ER–endosome fusion) within 120 min. Because ER–endosome fusion is a well-known cross-presentation mechanism (40–43), we assume that it contributes to MHC class I cross-presentation of exogenous Ags by γ-PGA NPs.

To analyze the mechanism by which the Ag reaches a cytoplasmic proteasome from the ER–endosome, cross-presentation efficiency was assessed under inhibition of sec61, an ER translocon that releases protein from the ER to cytoplasm (44). The sec61 inhibitor ExoA (45) was used in this experiment. ExoA significantly interfered with MHC class I presentation induced by γ-PGA NPs in a dose-dependent manner (Fig. 5). A similar result was also observed with excessive soluble OVA as a control of sec61-dependent cross-presentation (Fig. 5) (45, 46), suggesting that sec61-mediated protein translocation contributes to the escape of an Ag encapsulated in γ-PGA NPs from the ER–endosome to the cytoplasm (Fig. 6).

In the current study, we elucidated the pathway of MHC class I cross-presentation induced by γ-PGA NPs, as shown in Fig. 6. Ags encapsulated in γ-PGA NP entered into the DC through endocytosis (phagocytosis and/or macropinocytosis). The γ-PGA NPs enhanced ER–endosome fusion, and confined Ags were retrotranslocated via ER–translocon sec61 from the fused ER–endosome complex to the cytosol. Released Ags were degraded by cytoplasmic proteasomes and transported to the ER or ER–endosome fusion via TAP, and then the Ag–MHC class I complex was presented on the cell surface.

Our findings suggest that exogenous Ags encapsulated in γ-PGA NPs are cross-presented on MHC class I molecules through a classical endogenous Ag presentation pathway via cytoplasmic proteasomes and TAP. In contrast, the MHC class II presentation pathway induced by γ-PGA NPs was inhibited by the proteasome-specific inhibitor MG-132, but not by another proteasome inhibitor, epoxomicin, suggesting that degraded Ags in the cytoplasmic proteasome could be presented on MHC class II molecules. In this report, we reveal that Ags encapsulated in γ-PGA NPs can be retrotranslocated to the cytoplasm via sec61 for MHC class I cross-presentation, which indicates that this pathway might also contribute to MHC class II presentation. The cytoplasmic degradation machinery in the proteasome for MHC class II presentation, however, is poorly understood (47). Thus, although the possibility of cytoplasmic proteasome-dependent MHC class II presentation is potentially interesting, further studies are required.

Regarding the localization of γ-PGA NPs, CLSM and TEM studies suggested ER–endosome fusion occurred. After uptake into the cell via endocytosis, Ag-encapsulated γ-PGA NPs efficiently translocated from the EEA1(+) EE to ER-like endosomes within 120 min. We also considered the possibility that the Ag was retrotranslocated via retrograde transport pathway from the Golgi complex to the ER (48). CLSM and TEM observations, however, clearly revealed that most of the γ-PGA NPs and Ags were entrapped together in the endosome-like vesicles with ER markers, which are far away from the retrograde transport machinery. Therefore, we concluded that this structure constitutes ER–endosome fusion. The first ER–endosome fusion complex was reported with regard to phagosome entrapping of microsized latex beads (49). A recent report also suggests that the efficiency of cross-presentation is dependent on the size of the carrier (22). Therefore, optimization of the γ-PGA NP size of may be useful for efficient cross-presentation to induce strong tumor immunity.

We also found that Ags encapsulated in γ-PGA NPs can be retrotranslocated from the ER–endosome complex to the cytoplasm via ER translocon sec61, similar to previous reports that used excess soluble Ags (45, 46). Moreover, our previous report demonstrated that the induction efficiency of OVA-specific CTL by γ-PGA NPs was much higher than that of soluble OVA (10, 11). These findings indicate that efficient ER–endosome fusion induced only by a small amount of Ags encapsulated in γ-PGA NPs partially accounts for the strong Ag-specific CTL induction, followed by efficient cross-presentation on MHC class I molecules. This finding also suggests that the efficiency of sec61-mediated retrotranslocation of the Ag is a key factor in inducing a strong CTL response, which is essential for the development of a novel vaccine carrier.

Increasing our understanding of the mechanisms of ER–endosome fusion will lead to the design of a novel vaccine carrier to induce strong tumor immunity and the development of more potent anticancer therapeutics.

We thank Dr. Mitsuru Akashi and Dr. Takami Akagi at the Department of Applied Chemistry, Graduate School of Engineering, Osaka University, Japan, for instructions on how to synthesize and handle the γPGA-NPs. TEM observation was performed by Hanaichi UltraStructure Research Institute, Tokyo, Japan.

The authors have no financial conflicts of interest.