Mechanisms by Which Dendritic Cells Present Tumor Microparticle Antigens to CD8⁺ T Cells

Whole tumor cells may be a promising source of tumor antigens for the use in construction of cancer vaccines for several reasons. They contain the complete repertoire of mutated neoantigens and tumor-associated antigens, which will likely reduce the possibility of immune escape and development of resistance (1, 2). Also, the manipulation of tumor cell vaccination is very simple and convenient by means of culture and irradiation (3). The immunogenicity of whole tumor-cell vaccines can be augmented by the addition of adjuvants, including, but not limited to, Freund's incomplete adjuvant, BCG, GM-CSF, and Toll-like receptor agonists (4–6). Despite the potential success of these vaccines, previous reports of autologous and allogeneic whole tumor cell–based vaccines in clinical trials in patients with various tumor types demonstrate limited efficacy and efficiency of these vaccines (7). The antitumor immune response is mainly conveyed by tumor-lytic CD8⁺ T cells via dendritic cells (DC) that present tumor antigens. Tumor cell–derived cytokines (e.g., VEGF, IL10, and TGFβ) and biologic factors (e.g., galectin-1, indoleamine 2,3-dioxygenase, and lipid droplets) have been shown to suppress DC maturation and T-cell activation, likely limiting the efficacy and efficiency of whole cell–based vaccination (8, 9). Therefore, a vital consideration for developing tumor cell–based vaccines is providing DCs with appropriate innate signals apart from the tumor antigens.

Tumor cells are capable of generating and releasing microvesicles of various sizes into the extracellular space (10). These subcellular structures not only contain tumor antigens, but also include information that is unique from the parental cells (11, 12), suggesting the potential for tumor cell–derived microvesicles to be developed as tumor vaccines. In response to stimuli or apoptotic signals, cells may alter their cytoskeleton to encapsulate cytosolic contents within the cellular membrane to form subcellular vesicles. These 0.1–1 μm vesicles, termed microparticles, are subsequently released into extracellular spaces (13–16). We have previously shown that tumor cell–derived microparticles (T-MP) can be readily taken up by DCs, leading to maturation and presentation of multiple antigens. T-MPs may represent an effective vaccination platform to trigger antitumor T-cell immune response (17); however, the underlying molecular mechanisms by which DCs mature and present tumor antigens to tumor-lytic CD8⁺ T cells remain unclear.

Lysosomes are the sites at which class I peptides are generated for DCs to load them onto MHC class I, forming a complex for the cross-presentation of T-MP antigens; thus, we speculate that, following endocytosis by DCs, T-MPs may alter the function of lysosomes. At least two key issues regarding antigen cross-presentation are dependent on lysosomal alterations. First, to generate the 8 to 11 amino acid–length antigenic peptide, a transient increase of lysosomal pH is necessary to avoid the degradation of tumor antigens into smaller peptides or single amino acids. Second, lysosomal alteration might be required to regulate the expression of costimulatory molecules, such as CD80 and CD86, for effective presentation of tumor antigens to cytolytic CD8⁺ T cells. In this present study, we demonstrate that B16-OVA melanoma tumor cell–derived microparticles (OVA-MP) can readily enter DC lysosomes, increasing lysosomal pH and upregulating CD80 and CD86 expression via a NOX2–ROS–TFEB signaling pathway, leading to subsequent activation of tumor-lytic CD8⁺ T cells. These findings highlight tumor cell–derived microparticles as cell-free vaccines with potential clinical applications.

Murine melanoma cell lines B16-OVA and B16 were obtained from China Center for Type Culture Collection in 2014 and maintained in a rigid dish with 1640 cell culture medium (Invitrogen) supplemented with 10% FBS (Gibco) at 37°C with 5% CO₂. Cells were determined to be mycoplasma-free, free of interspecies cross-contamination and authenticated by isoenzyme and short tandem repeat analyses in Cell Resource Centre of Peking Union Medical College before the study. Cell lines used in the experiments were within 20 passages. Female wild-type C57BL/6J (6–8 weeks old) were purchased from the Centre of Medical Experimental Animals of Hubei Province (Wuhan, China). OT-I TCR-transgenic mice (C57BL/6-Tg (TcraTcrb) 1100Mjb/J) were a gift from Dr. Hui Zhang (Sun Yat-Sen University, Guangdong, China). Pmel-1 transgenic mice were presented by Dr. Ying Wan (Third Military Medical University, Chongqing, China). All mice were bred in specific pathogen-free conditions. All animal experiments were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals along with approval from the Scientific Investigation Board of the Tongji Medical College, Wuhan, China.

Amiloride hydrochloride, DPI, NAC, ryanodine, CGP37157, cyclosporin A (CsA), sodium vanadate (Na₃VO₄), PKH26 Red, and PKH67 Green Fluorescent Cell Linker Kit were purchased from Sigma. ER-, Mito-, LysoTracker Green/Red, LysoSensor Green, LysoSensor Yellow/Blue, and CellLight Golgi-GFP were purchased from Invitrogen. The following primary antibodies were purchased from Abcam: anti-LAMP1 (ab13523), anti-gp91 (ab80508), anti-Rab7 (ab137029), anti-dynein (ab157468), anti-histone H3 (ab8284), and anti-TFEB (ab2636, for ChIP). Anti-β-actin (ANT009) was purchased from Ant Gene. The following secondary antibodies were purchased from Abcam: goat anti-mouse IgG FITC, donkey anti-rabbit IgG Alexa Fluor488, and goat anti-rabbit IgG Dylight594.

Tumor cells were exposed to ultraviolet irradiation (300 J/m²) for 1.5 hours, and 18 hours later, supernatants were used for microparticle isolation as described previously (17). Briefly, supernatants were centrifuged at 1,000 × g for 10 minutes to remove whole cells and then centrifuged for 2 minutes at 14,000 × g to remove debris. The supernatant was further centrifuged for 60 minutes at 14,000 × g to pellet microparticles. The pellets were washed 3 times and resuspended in culture medium for the subsequent experiments.

Bone marrow cells were harvested from femurs of mice and cultured in RPMI 1640 supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin. The cells were cultured in 6-well plates with 20 ng/mL GM-CSF (PeproTech) and 20 ng/mL IL4 (PeproTech), and cytokines were replenished on days 3 and 5; nonadherent cells were harvested for experiments. We also used Flt3L (200 ng/mL) to treat mouse bone marrow cells for 10 days to induce Flt3L-BMDCs. Murine splenic CD8α⁺ DCs from pooled spleens (from at least 3 spleens) were purified first by density-gradient centrifugation and then by magnetic-activated cell separation. OT-I and pmel-1 mice–derived splenic CD8⁺ T cells were purified by magnetic-activated cell separation (purity > 90%, Miltenyi Biotec).

Splenic CD8⁺ T cells were purified through negative selection (Miltenyi Biotec) from OT-I or pmel-1 mice, then fluorescently labeled with 5 μmol/L CFSE (Sigma-Aldrich). DCs were incubated with OVA-MPs for 12 hours to obtain OVA-MP–loaded DCs. CFSE-labeled T cells were incubated with OVA-MP–loaded DCs or empty DCs for 2 to 3 days before flow-cytometric analysis.

Splenic CD8⁺ T cells purified from OT-1 mice were incubated with empty DCs or OVA-MP–loaded DCs. IFNγ in the supernatants of CD8⁺ T cells was assessed by the mouse mini ELISA kit (PeproTech) according to the manufacturer's protocol.

Lysosomal staining was performed using LysoTracker, a lysosomotropic probe (Invitrogen). The treated cells were incubated for 30 minutes at 37°C with 1 μmol/L of LysoTracker. The cells were examined using a confocal microscope. LysoSensor Green DND-189 is commonly used to qualitatively measure the pH of acidic organelles, such as lysosomes, which become more fluorescent in acidic environments and less fluorescent in alkaline environments. DCs were loaded with 0.1 μmol/L LysoSensor Green DND-189 in prewarmed RPMI 1640 medium for 30 minutes at 37°C. The cells were then washed twice with PBS and immediately analyzed by fluorescence microscope. Quantification of lysosomal pH was performed using a ratiometric lysosomal pH dye LysoSensor Yellow/Blue DND-160. The pH calibration curve was generated according to the manufacturer's protocol. DCs were trypsinized and labeled with 2 μmol/L LysoSensor Yellow/Blue DND-160 for 30 minutes at 37°C in RPMI 1640 medium, and excess dye was washed away using PBS. The labeled cells were treated for 10 minutes with 10 μmol/L monensin and 10 μmol/L nigericin in 25 mmol/L MES calibration buffer, pH 4.5–7.5, containing 5 mmol/L NaCl, 115 mmol/L KCl, and 1.2 mmol/L MgSO₄. Quantitative comparisons were performed in a 96-well plate, and the fluorescence was measured with a microplate reader at 37°C.

Analysis of intracellular reactive oxygen species (ROS) production was conducted according to the manufacturer's protocol. Briefly, DCs with indicated treatments were incubated with 2.5 μmol/L CellROX Green at 37°C for 30 minutes and analyzed by flow cytometry.

Recombinant vectors encoding murine TFEB were constructed by PCR-based amplification from cDNA of DCs and then were subcloned into the pcDNA3.1–3 × Flag eukaryotic expression vectors. All constructs were confirmed by DNA sequencing in BGI (Shenzhen, China). Sequencing primers: forward, 5′-CGCAAATGGGCGGTAGGCGTG-3′ and reverse, 5′-TAGAAGGCACAGTCGAGG-3′. Plasmids were transiently transfected into BMDCs with Lipofectamine RNAiMAX Transfection Reagent (Invitrogen).

siRNAs targeting mouse gp91 (siRNA#1: GCTGAATGTCTTCCTCTTT; siRNA#2: CCATGGAGCTGAACGAATT; siRNA#3: GCACCATGATGAGGAGAAA), mouse dynein (siRNA#1: GAAATCAACTTGCCCGATA; siRNA#2: CCACGTGCCTGTTGTATAT; siRNA#3: GCAGGCAGATGAGCAGTTT), mouse Rab7 (siRNA#1: GGAAGAAAGTGTTGCTGAA; siRNA#2: CCATCAAACTGGACAAGAA; siRNA#3: GTACAAAGCCACAATAGGA), mouse Mcoln2 (siRNA#1: GCAGTTCATTCCCGAGAGA; siRNA#2: GCTGAGGAAGAGATTTCTA; siRNA#3: GCTTGAAGGTCTGTAAGCA), mouse TFEB (siRNA#1: GCAGGCTGTCATGCATTAT; siRNA#2: CCAAGAAGGATCTGGACTT; siRNA#3: CCATGGCCATGCTACATAT) and negative control siRNAs (NC) were purchased from RiboBio. siRNA (50 nmol/L) was transfected into DCs using lipofectamine RNAiMax (Invitrogen) according to the manufacturer's instruction.

Real-time PCR analyses were performed with 2 μg of cDNA as a template, using an SYBR Green mix (Applied Biosciences) and an Agilent Technologies Stratagene Mx3500P real-time PCR system. Relative quantitative RNA was normalized using the housekeeping gene GAPDH. Analysis of the results was performed using Bio-Rad CFX Manager and relative quantification was performed. The entire procedure was repeated in at least 3 biologically independent samples. The primer sequences are shown as follows: LAMP1, 5′-ACAGGGATATATGGGCAGGGA-3′ (sense) and 5′-AGCCAGGACACCCTTACCTC-3′ (antisense); LAPM2, 5′-AGGAATGTGCTGCTGACTCTG-3′ (sense) and 5′-AATGGAAGCACGAGACTGGC-3′ (antisense); TFEB, 5′-CCACCCCAGCCATCAACAC-3′ (sense) and 5′-CAGACAGATACTCCCGAACCTT-3′ (antisense); ATP6V0A1, 5′-CCGAGGACGAAGTGTTTGACT-3′ (sense) and 5′-ATCAGCAGGATAGCCACGGTAA-3′ (antisense); ATP6V0A2, 5′-TGGTGCAGTTCCGAGACCT-3′ (sense) and 5′-GCAGGGGAATATCAGCTCTGG-3′ (antisense); ATP6V0C, 5′-ACTTATCGCTAACTCCCTGACT-3′ (sense) and 5′-ACACCAGCATCTCCGACGA-3′ (antisense); ATP6V0E, 5′-GCATACCACGGCCTTACTGT-3′ (sense) and 5′-TGATAACTCCCCGGTTAGGAC-3′ (antisense); ATP6V1A, 5′-ACAGAGGAAGCGTGACTTACA-3′ (sense) and 5′-CACTTGGACCATGCTGAACTT-3′ (antisense); ATP6V1B2, 5′-ATGCGGGGAATCGTGAACG-3′ (sense) and 5′-AGGCTGGGATAGGTAGTTCCG-3′ (antisense); ATP6V1C1, 5′-ACTGAGTTCTGGCTCATATCTGC-3′ (sense) and 5′-TGGAAGAGACGGCAAGATTATTG-3′ (antisense); ATP6V1E1, 5′-GAATCAAGCAAGGCTCAAAGTCC-3′ (sense) and 5′-CGGGTCG TATCTTTTACCACC-3′ (antisense); ATP6V1F, 5′-GCGGGCAGAGGTAAGCTAAT C-3′ (sense) and 5′-TTAGGGTGGCGGTTCTTGTTT-3′ (antisense); ATP6V1G1, 5′-CCCAGGCTGAAATTGAACAGT-3′ (sense) and 5′-TTCTGGAGGACGGTCATCTTC-3′ (antisense); ATP6V1H, 5′-GGATGCTGCTGTCCCAACTAA-3′ (sense) and 5′-TCTCTTGCTTGTCCTCGGAAC-3′ (antisense); gp91, 5′-TGGCGATCTCAGCAAAAGGTGG-3′ (sense) and 5′-GTACTGTCCCACCTCCATCTTG-3′ (antisense); TAP1, 5′-CAGCGGCAACCTTGTCTCAT-3′ (sense) and 5′-TTCCAGGATGCAGGGTGAAC-3′ (antisense); TAP2, 5′-TATGGCCTGAGGGACTGTGA-3′ (sense) and 5′-TCCAGTTCTGTAGGGCCTGT-3′ (antisense); Mcoln1, 5′-ACCATCTCGGGGACTGTCAT-3′ (sense) and 5′-CAGGTAGCGAATGACACCGA-3′ (antisense); Mcoln2, 5′-GCATTCTGGTGTGGCTGTTC-3′ (sense) and 5′-GGTGTGGTAAGAGTCGGTG A-3′ (antisense); Tpcn1, 5′-GGACGGCGCGTACCTTA-3′ (sense) and 5′-CGGTCCTCAGGATACAACGG-3′ (antisense); Tpcn2, 5′-GCCTTCCTGGTTGACCTCTC-3′ (sense) and 5′-CGAAACGATCCAGTCCACCA-3′ (antisense); β-actin, 5′-CATTGCTGACAGGATGCAGAAGG-3′ (sense) and 5′-TGCTGGAAGGTGGACAGTGAGG-3′ (antisense).

For phenotypic analysis of DCs, cells were stained with surface antibodies: anti-CD11c (clone N418), anti-CD80 (clone 16-10A1), anti-CD86 (clone GL1), anti-MHC I (clone M5/114.15.2), anti-MHC II (clone M5/114.15.2), anti-CCR7 (clone 4B12), anti-CD40 (clone 1C10), anti-OX40L (clone RM134L), and anti-PDL1 (clone 10F.9G2). Rat IgG2a, κ, Rat IgG2b, κ and Armenian Hamster IgG were used as isotype controls. All antibodies were purchased from eBioscience or BioLegend, and flow-cytometric analysis was performed with Accuri C6 (BD).

Isolated OVA-MPs were labeled with a green fluorescent cell linker (PKH67; Sigma-Aldrich), according to the manufacturer's protocol. Labeled OVA-MPs were incubated with DCs at 37°C for 6 hours before staining with PE-CD11c antibody (eBioscience), according to the manufacturer's protocol and visualized by two-photon fluorescent microscopy. For intracellular staining, DCs were fixed in 2% paraformaldehyde for 10 minutes at room temperature, permeabilized with 100 μmol/L digitonin and blocked with 1% BSA for 1 hour at 25°C. Samples were incubated with primary antibody (in PBS with 1% BSA and 0.1% Tween-20) overnight at 4°C. Following overnight incubation, cells were washed 3 times in PBS and incubated with secondary antibodies for 1 hour at room temperature. Nuclei were stained in DAPI solution (1 μg/mL). Merge figure shows the bright field image and fluorescent image observed under a two-photon fluorescent microscope.

Whole cell lysates were prepared from DCs and separated by SDS–PAGE at 100 V for 1 hour. Separated proteins were then transferred to nitrocellulose membranes (Millipore). The membranes were blocked in 5% BSA in TBS containing 0.1% Tween-20 for 1.5 hours at room temperature. Then, the membranes were incubated with anti-TFEB, anti-Histone H3, or anti β-actin overnight at 4°C. The membranes were washed 5 times and incubated with HRP-conjugated secondary antibodies for 1.5 hours at room temperature. Proteins were visualized by ECL Western blotting substrate (Thermo Scientific Pierce).

DCs were cultured in 24-well plates at the density of 5 × 10⁴ cells/well in RMPI 1640 medium overnight. Before Ca²⁺ measurement, cells were washed with PBS for 3 times and incubated for 60 minutes in Hanks' balanced salt solution containing 4 μmol/L Fluo-4AM in the dark at room temperature. The cells were then washed with Hanks' balanced salt solution 3 times and incubated at room temperature for another 10 minutes. Then 200 nmol/L ionomycin (iono) was applied extracellularly at 30 seconds, and the cytosolic calcium release was recorded by two-photon confocal microscope. Images were collected every 2 seconds and analyzed by ImageJ software (NIH).

A chromatin immunoprecipitation (ChIP) assay kit (Active Motif) was utilized to examine the binding of TFEB to the CD80 and CD86 promoter. Untreated and MP-treated DCs were fixed with 1% formaldehyde on ice to cross link the proteins bound to the chromatin DNA. After washing, the chromatin DNA was sheared by enzymatic force to produce DNA fragments of around 200 to 1,000 bp. The same amounts of sheared DNA were used for immunoprecipitation with a TFEB antibody or an equal amount of preimmune IgG. The immunoprecipitate then was incubated with protein G Magnetic Beads, and the antibody–protein G Magnetic Beads complex was collected for subsequent reverse cross-linking. The same amount of sheared DNA without antibody precipitation was processed for reverse cross-linking and served as input control. DNA recovered from reverse cross-linking was used for PCR. PCR was performed with primers for the CD80 and CD86 promoter flanking the TFEB binding site at 59°C for 36 cycles (CD80 primers, forward, 5′-CGCTCTGGATAACCTGCACT-3′ and reverse, 5′-ACAGCGGTGTGTAAGCTGTC-3′; CD86 primers, 5′-GTGAGACTGGGACACCAACA-3′ and reverse, 5′-GCTCTGCCGCTATCTAGCTT-3′).

All experiments were performed at least 3 times. Results were expressed as mean ± SEM and analyzed by an unpaired two-tailed Student t test. P values of <0.05 were considered statistically significant. The analysis was conducted using the GraphPad Prism 6.0 software.

Previously, we showed that the immunogenicity of T-MPs required the uptake of DCs (17). In this study, we investigated the underlying mechanisms. Using B16 melanoma cells expressing the model tumor antigen ovalbumin (OVA) and OVA-specific CD8⁺ T cells derived from OT-I T-cell receptor transgenic mice, we found that bone marrow–derived DCs (BMDC), which were generated by GM-CSF/IL4 stimulation and pretreated with B16-OVA cell–derived microparticles (OVA-MP), stimulated OVA-specific CD8⁺ T-cell proliferation as well as IFNγ production (Fig. 1A and B), suggesting that BMDCs effectively process and present OVA antigen to T cells. Consistently, we observed that almost 100% of the BMDCs efficiently took up OVA-MPs by fluorescent microscopy (Fig. 1C). The uptake rate of OVA-MPs was detected by flow cytometry (Supplementary Fig. S1A and S1B). The MHC class I–OVA peptide complexes were confirmed to be expressed on BMDC membrane surface in a dose- and time-dependent manner (Fig. 1D and E). In addition, DC maturation surface markers, including CD80, CD86, CCR7, MHC class I/II, CD40, and OX40L, were all upregulated upon 24-hour incubation with OVA-MPs, but PD-L1 expression was not upregulated on DCs (Fig. 1F). Thus, the uptake of OVA-MPs by DCs was necessary for subsequent tumor-specific T-cell activation. B16 melanoma cells commonly express tumor-associated antigen gp100 (18). Consistent with this, bone marrow–derived DCs also presented the MHC class I–gp100 peptide complexes to the surface, following the B16-MP treatment, concomitant with the upregulation of CD80, CD86, CCR7, CD40, and OX40L and IFNγ production (Supplementary Fig. S1C–S1E). As a result, gp100-specific CD8⁺ T cells, isolated from pmel-1 T-cell receptor transgenic mice, were stimulated to proliferate effectively by the DCs (Supplementary Fig. S1F). To dissect the uptake process, we stained OVA-MPs and BMDC organelles. We did not observe any colocalization of OVA-MPs with mitochondria, endoplasmic reticulum (ER), or Golgi apparatus (Fig. 1G). However, OVA-MPs were found to be colocalized with Rab5⁺ early endosomes, Rab7⁺ late endosomes, and lysosomes (Fig. 1H), suggesting that OVA-MPs are taken up via endocytosis and trafficked to the lysosomes of DCs. To further verify this, we added the endocytosis inhibitor amiloride hydrochloride to block BMDCs from taking up OVA-MPs. As a result, BMDCs were not able to induce OT-I T-cell activation (Fig. 1I). We additionally analyzed the Flt3L-induced DCs and CD8α⁺ DCs isolated from the spleen (19–22). These DCs induced OT-1 T-cell proliferation and IFNγ production after OVA-MPs pulsing (Supplementary Fig. S2A–S2D) and upregulated the expression of CD80, CD86, CCR7, CD40, and OX40L (Supplementary Fig. S2E and S2F). The ability of DCs to cross-present T-MPs was also confirmed in vivo. We injected OVA-MPs into the footpads of C57BL/6 mice (8 × 10⁵, once per day) for 4 days, followed by the adoptive transfer of OT-I T cells. We found that OVA-MP treatment resulted in DCs in popliteal LNs presenting MHC I–OVA peptide complexes and upregulating CD80, CD86, CCR7, CD40, and OX40L (Supplementary Fig. S3A and S3B). Also, OT-I T-cell proliferation was induced in the spleen and draining lymph nodes (Supplementary Fig. S3C). Together, these data suggest that DCs endocytose T-MPs to lysosomes, leading to tumor antigen presentation and tumor-specific CD8⁺ T-cell activation.

Next, we examined whether and how endocytosed OVA-MPs affect the lysosomes of DCs. A fundamental function of lysosomes is degrading biomolecules, a process that strictly relies on the acidic microenvironment (pH 4.5–5.0) in lysosomal lumen. Staining OVA-MP–treated BMDCs with LysoSensor Green, we found that the normalized fluorescence intensity decreased by 75%, indicating that lysosomal pH was elevated (Fig. 2A). The lysosomal pH increased to a peak of 8.5 within 24 hours and then decreased to 6.9 at 48 hours(Fig. 2B). Lysosomal pH alteration is commonly associated with lysosome biogenesis (23). In line with increased pH, the normalized fluorescence intensity of LysoTracker Red increased about 2.5 times, indicating that lysosome number increased in OVA-MP–treated BMDCs (Fig. 2C). Consistent with this observation, 2 lysosomal genes, LAMP1 and LAMP2, were upregulated in the treated BMDCs (Fig. 2D). Transcription factor EB (TFEB) is known to be critical for lysosomal biogenesis upon its translocation into the nucleus (24). Both TFEB expression and its entry into the nucleus were found to be upregulated in the OVA-MP–treated BMDCs, compared with untreated BMDCs (Fig. 2E and F). Together, these data suggest that endocytosed OVA-MPs induce an increase of both lysosomal pH and lysosomal number in DCs.

Low lysosomal pH is maintained by the vacuolar-type H⁺-ATPase (V-ATPase)-mediated pumping of protons into the lumen. V-ATPase consists of V0 and V1 domains, each with multiple subunits (25). The real-time PCR result did not show differential expression of subunit members of V-ATPase between MP-treated and untreated BMDCs (Fig. 3A). The enzyme NADPH oxidase 2 (NOX2, previously known as gp91^phox) is also ubiquitously integrated into lysosomal membrane of phagocytes in a heterodimer form with p22^phox. During phagocytosis or upon stimulation, cytosolic regulatory proteins p47^phox, p67^phox, p40^phox, and GTP-binding Rac are recruited to the membrane, where they assemble with NOX2-p22^phox to form an active oxidase complex, leading to transfer of electron from NADPH to molecular oxygen and the production of superoxide anion (26). Here, we found that endocytosed T-MPs could upregulate the expression of gp91^phox (Fig. 3B), and that gp91^phox was effectively recruited to lysosomes in MP-treated BMDCs (Fig. 3C). Staining BMDCs with CellROX Green, we found that the fluorescence intensity of ROS increased about 2 times in OVA-MP–treated BMDCs, as measured by fluorescence microscopy and flow cytometry (Fig. 3D and E). Superoxide anion in lysosomes can be quickly reduced to hydrogen peroxide by reacting with protons, thus consuming protons and increasing pH (27). To examine this, we used NADPH oxidase inhibitor diphenylene iodonium (DPI) or gp91^phox siRNA to block NOX2 activity. We found that the above ROS and lysosomal pH decreased to the levels of untreated BMDCs (Fig. 3F and G), suggesting that endocytosed T-MPs increase lysosomal pH via NOX2-mediated ROS production. We further speculated that the increase of lysosomal pH by T-MPs was required for DCs to present tumor antigens, because low pH confers the ability of lysosomal enzymes to degrade tumor antigens into small peptides or single amino acids. To test this, we treated OVA-MP–phagocytosed BMDCs with either DPI or gp91^phox siRNA and cultured the cells with OT-I T cells. We found that the DPI or siRNA treatment significantly inhibited T-cell proliferation (Fig. 3H). Moreover, we found that the expression of MHC class I–OVA peptide complexes on the surface of those DCs was significantly reduced by DPI or gp91^phox siRNA (Fig. 3I). Together, these data suggest that endocytosed T-MPs increase lysosomal pH via the pathway of NOX2-mediated ROS production, leading to generation of MHC–tumor antigenic peptide complexes.

Increased lysosomal pH likely results in long peptide (>11 amino acids) production due to decreased enzymatic activity. Large peptides cannot be accommodated by the peptide-binding cleft of MHC class I for DC cross-presentation. We hypothesized that large peptides exited lysosomes and entered the proteasome degradation pathway to generate 8–11-amino acid MHC class I peptides. However, when we treated OVA-MP–endocytosed BMDCs with the pan-proteasome inhibitor MG-132, we found that this blockade did not affect the ability of DCs to express and present OVA peptide to CD8⁺ T cells (Fig. 4A). This result suggested that increased lysosomal pH directly causes the production of class I peptides with the 8–11 amino acid length. Notwithstanding the generation of class I peptide in lysosomes of T-MP–phagocytosed DCs, the mechanism by which class I peptide contacts and complexes with the MHC class I molecule needed further clarification. TAP1 and TAP2 mediate the entry of cytosolic class I peptide into ER (28). Here, we found that T-MP–treated BMDCs upregulated the expression of TAP1 and TAP2 (Fig. 4B), and knockdown of TAP1/2 resulted in abrogation of BMDCs presenting OVA-MP tumor antigen to OT-I T cells (Fig. 4C). This result implies that class I peptide is released from lysosomes and translocated to ER in this way. In addition to digestion, lysosomes are also able to transport molecules. Lysosomes are capable of bidirectional migration along microtubule tracks, upon recruiting specific regulatory molecules to their membranes. For example, centripetal (inward) movement is guided by the small GTPase Rab7, which recruits the minus end-directed microtubule motor dynein to lysosomes (29). By contrast, centrifugal (outward) movement is directed by another small GTPase Arl8, which links lysosomes to the plus end-directed microtubule motor kinesin (30). Lysosomal migration can be regulated by pH alteration (31). We previously reported that T-MPs facilitate the inward migration of lysosomes in tumor cells (32). Here, we further hypothesized that lysosomes facilitate tumor antigenic peptides to access the ER through a centripetal transport pathway. In line with the hypothesis, we found that Rab7 and dynein were recruited to the lysosomal membrane in OVA-MP–endocytosed BMDCs (Fig. 4D). Moreover, either dynein or Rab7 siRNA led to downregulation of the expression of MHC class I–OVA peptide complex in the treated BMDCs (Fig. 4E) and a subsequent reduction in OT-I T-cell proliferation (Fig. 4F). Together, these data suggest that endocytosed T-MPs facilitate tumor antigen cross-presentation by promoting centripetal migration of lysosomes via a Rab7/dynein-mediated pathway.

The above data indicate that DCs efficiently process tumor antigens of T-MPs and present them as MHC class I-peptide complexes to the cellular membrane, thus providing the first signal for tumor-specific T-cell activation. However, the second signal by CD80/CD86 is also required for T-cell activation. Twenty-eight hours after incubation of BMDCs with T-MPs, we determined the expression of CD80 and CD86 by flow cytometry. We found that the expression of CD80 and CD86 was upregulated (Fig. 5A), which was abrogated by the addition of DPI (Fig. 5A). Given that T-MPs augmented ROS production in DCs via activating the NOX2 pathway, we then used ROS scavenger NAC to treat OVA-MP–loaded BMDCs. The result showed that the inhibition of ROS production downregulated the expression of CD80 and CD86 (Fig. 5B and C), concomitant with decreased OT-I T-cell proliferation (Fig. 5D). Together, these data suggest that NOX2-mediated ROS production upregulates the expression of CD80 and CD86 in T-MP–endocytosed DCs.

Next, we investigated the mechanism by which ROS regulates the expression of CD86 and CD80. Although activation of TLR signaling is a common pathway for CD80 and CD86 upregulation (33), Ca²⁺ signaling is also able to effectively upregulate their expression (34). Coincidently, lysosomes are organelles for calcium storage (35). Here, we report that the intracellular calcium levels were elevated in BMDCs upon uptake of T-MPs (Fig. 6A). Using ryanodine or CGP37157 to block calcium release from the ER or mitochondria did not affect the T-MP–mediated Ca²⁺ increase in DCs (Fig. 6A), suggesting that lysosomes, rather than ER or mitochondria, release Ca²⁺ to the cytosol. Several Ca²⁺ channels (TPC1/2 and Mcoln1/2) have been reported to mediate lysosomal calcium release (36). Here, we found that endocytosed T-MPs only upregulated the expression of Mcoln2 in BMDCs (Fig. 6B). Knockdown of Mcoln2 with the siRNA led to inhibition of T-MP–induced Ca²⁺ release (Fig. 6C) as well as CD80 and CD86 upregulation in BMDCs (Fig. 6D), concomitant with the impaired CD8⁺ T-cell proliferation (Fig. 6E). In addition, treatment with cyclosporin A (CsA), a Ca²⁺ signaling inhibitor (37), also blocked the effect of T-MPs on CD80 and CD86 upregulation (Fig. 6F). Given that T-MP caused ROS production and pH increase in lysosomes, we assumed that the ROS production and pH alteration resulted in Mcoln2-mediated Ca²⁺ release. To clarify this, we used either NAC or DPI to block the ROS/pH pathway. As a result, both DPI and NAC downregulated the expression of Mcoln2 and inhibited lysosomal Ca²⁺ release in T-MP–treated BMDCs (Fig. 6G and H). Together, these data suggest that ROS-triggered lysosomal calcium signaling upregulates the expression of CD80 and CD86 in T-MP–loaded DCs.

Finally, we investigated the mechanism by which lysosomal Ca²⁺ signaling regulates expression of CD86 and CD80. TFEB is not only a master transcriptional regulator for lysosomal biogenesis but also regulates immune-related genes (38). Phosphorylated TFEB, which is localized in the cytosol, translocates to the nucleus upon dephosphorylation by Ca²⁺ release-activated calcineurin (39). We hypothesized that lysosomal Ca²⁺ signaling-activated TFEB upregulates CD80 and CD86 expression. In Fig. 2E, the real-time PCR result showed that TFEB was upregulated in T-MP–loaded BMDCs. Here, immunostaining and Western blot showed that abundant TFEB was translocated into the nucleus in T-MP–treated DCs, but not in untreated BMDCs (Fig. 7A and B). To clarify whether TFEB regulates CD80 and CD86 expression, we knocked down TFEB by siRNA (Fig. 7C), which led to downregulation of CD80 and CD86 expression (Fig. 7D). Na₃VO₄, a serine/threonine phosphatase inhibitor that promotes TFEB phosphorylation, also downregulated the expression of CD80 and CD86 (Fig. 7E). In contrast, when we forced overexpression of TFEB (TFEB-OE) in DCs, the expression of CD80 and CD86 was upregulated (Fig. 7F). We performed chromatin immunoprecipitation (ChIP)-qPCR assay and found that TFEB indeed bound the promoters of CD80 and CD86 genes (Fig. 7G). Together, these data suggest that activation TFEB transcriptional activity by T-MP–triggered lysosomal Ca²⁺ signaling directly upregulates the expression of CD80 and CD86.

The antitumor immune response is mainly mediated by tumor-specific CD8⁺ T cells. DCs are crucial in the generation of these CD8⁺ T cells, as they are required not only to present tumor antigen peptides, but also to provide costimulating signals to CD8⁺ T cells. Thus, an ideal tumor vaccine ought to simultaneously possess abundant tumor antigens as well as suitable innate signals (40). We have already demonstrated that tumor cell–derived microparticles (T-MP) have these dual advantages and represent an effective vaccination platform to trigger antitumor T-cell immunity (17). In this study, we demonstrate the unusual action pathway of T-MPs in DCs through which T-MPs induce DCs to become highly efficient in presenting tumor antigens to tumor-lytic CD8⁺ T cells.

Tumor cells are capable of releasing different types of microvesicles, the roles of which in tumor immunity are controversial. Exosomes are small, endosome-derived extracellular microvesicles (30–100 nm), delivering contents such as proteins, messenger RNAs, and microRNAs to recipient cells. Although tumor exosomes contain tumor antigens, studies have shown that tumor exosomes actually mediate tumor immunosuppression as well as metastasis (41). However, subcutaneous inoculation or oral administration of T-MPs results in the generation of antitumor T-cell immunity (17, 42, 43). The opposite consequence of treatment with T-MPs, compared with exosomes, may be ascribed to the different contents within these two vesicles. In particular, T-MPs contain genomic and mitochondrial DNA fragments, but exosomes may not contain these (17). Such DNA fragments, which are capable of stimulating DCs to release IFNβ via the cGAS–STING pathway, play an important role in polarizing macrophages toward an M2 phenotype, which promotes tumor growth (44). The molecular mechanism through which T-MPs have differential effects on DCs and macrophages remains unclear. One observation is that T-MPs increase the lysosomal pH of DCs but decrease the lysosomal pH of macrophages. Such an unexpected result warrants further investigation.

Endocytosis is the initial step in cross-presentation of exogenous tumor antigens by antigen-presenting cells to CD8⁺ T cells. Endocytosed antigens are then transited to endolysosomes where a delicate degradation of tumor antigens occurs: too much degradation may destroy potential T-cell epitopes, but some sufficient degradation is required to generate antigenic peptides. The compatible antigenic peptides are then translocated into the ER, where they bind to MHC class I molecules (45). Lysosomal proteases that mediate protein degradation normally require an acidic pH (5.5–6.5) to exert their optimal function, and two systems of V-ATPase and NOX2 regulate the lysosomal pH in DCs (46–48). V-ATPase pumps protons from the cytosol to lysosomal lumen, thus acidifying lysosomes. In contrast, NOX2 transfers electrons from NADPH to molecular oxygen, and the generated superoxide anion causes hydrogen peroxide formation by consuming protons, leading to alkalizing lysosomes. In this study, we report that endocytosed T-MPs influenced antigen degradation in lysosomes of DCs. T-MPs seemed not to alter the V-ATPase system, but activated the NOX2 system, and subsequently increase lysosomal pH. As a result, tumor antigens from T-MPs were effectively degraded into antigenic peptides in the lysosomes of DCs. Consistent with the current study, previous work showed that excessive antigen degradation occurs in NOX2-deficient DCs, resulting in a defect in cross-presentation to CD8⁺ T cells (27). Incomplete lysosomal degradation might cause greater length antigenic peptides (>8–11 amino acids), which are required to be transferred to proteasomes for the secondary degradation. However, when we used a proteasome inhibitor to block this process, DC presentation of T-MP tumor antigens was not affected, suggesting that DCs effectively generate 8–11 amino acids antigenic peptides in lysosomes, upon uptake of T-MPs. T-MPs are capable of promoting centripetal migration of lysosomes toward ER through recruiting small G protein Rab7 to the lysosomal membrane, leading to activating dynein–microtubule transportation system (32). Thus, the T-MP–triggered pathway induces efficient translocation of antigenic peptides to the ER.

One finding of this study is that DCs mobilize T-MP–induced ROS to enhance presentation of tumor antigen. The generation of ROS by T-MPs not only increases pH for proper degradation of tumor antigen into antigenic peptide, but also induces the upregulation of CD80 and CD86 expression, key costimulatory molecules for T-cell activation. Although pathogen-associated molecular patterns, such as LPS, or damage-associated molecular proteins, such as HMGB1, readily upregulate CD80 and CD86 expression through the NF-κB and MAPK pathways (33, 49), here, we find that lysosomal Ca²⁺ signaling triggers CD80 and CD86 upregulation via activation of TFEB, the master lysosomal regulator. Conventionally, ER is thought to be the Ca²⁺ storing organelle (50); however, lysosomes also store Ca²⁺ (51). In this study, we find that T-MP–induced ROS activates lysosomal Ca²⁺ channel, leading to lysosomal Ca²⁺ release. Such Ca²⁺ signaling may activate the phosphatase calcineurin that dephosphorylates TFEB. Subsequently, TFEB enters the nucleus and stimulates transcription (38, 39, 52). Although TFEB regulates genes involved in autophagy and lysosome biogenesis, some immune molecules such as IL1β, IL6, TNFα, and CCL5 are downregulated in TFEB-knockdown macrophages infected with Staphylococcus aureus, implying that TFEB may directly transcriptionally regulate immune genes (53). In the current study, we found that TFEB binds to and transcriptionally activates CD80 and CD86 promoters. Although we verified the pathway of T-MP–induced ROS upregulation of CD80 and CD86, we have not yet elucidated the mechanism through which T-MPs activate NOX2 and subsequently increase ROS production in DCs. This work is ongoing.

In summary, the work presented here demonstrates that T-MPs, by activating lysosomal ROS, induce initiation of two concomitant pathways. T-MPs induce production and translocation of tumor antigenic peptides to the ER via T-MP–increased lysosomal pH and T-MP–triggered dynein–microtubule transportation in DCs in one pathway, and upregulation of CD80 and CD86 expression by TFEB activated by ROS-triggered Ca²⁺ signaling on the other (Fig. 7H). These findings reveal the molecular pathways by which T-MP vaccines prime DCs to efficiently present T-MP tumor antigens within CD8⁺ T cells, thus opening a new avenue for cancer immunotherapy.

No potential conflicts of interest were disclosed.

Conception and design: J. Ma, Y. Yu, B. Huang

Development of methodology: J. Ma, K. Wei, J. Chen, B. Huang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Wei, H. Zhang, K. Tang, T. Zhang, P. Xu, J. Chen, L. Zhou, X. Liang, R. Fiskesund

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Ma, H. Zhang, F. Li, P. Xu, Y. Yu, W. Sun, R. Fiskesund

Writing, review, and/or revision of the manuscript: J. Ma, R. Fiskesund, B. Huang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Ma, K. Wei, J. Liu, Y. Liu

Study supervision: K. Wei, Y. Liu, B. Huang

This work was supported by CAMS Initiative for Innovative Medicine (2017-I2M-1-001) and the National Natural Science Foundation of China (81601447, 81788101, 91742112, 81661128007, and 81530080).

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.