6 to 8 weeks of C57BL/6 male mice were bought from Animal Laboratory Center, Xinjiang Medical University (Urumqi, Xinjiang, China) and housed in an animal facility of Xinjiang University. The experimental protocols were approved by the Committee on the Ethics of Animal Experiments of Xinjiang Key Laboratory of Biological Resources and Genetic Engineering (BRGE-AE001).

Plasmids including pMG36e-RFP-EGFP (358), pMG36e-EGFP-mcherry, pMG36e-penp-OVA, pMG36e-CMV-penp-OVA and pNZ8149-penp-OVA were constructed and kept by Xinjiang Key Laboratory of Biological Resources and Genetic Engineering. pMG36e-RFP-EGFP (358) encodes a green fluorescent protein (EGFP) under the control of a eukaryotic promoter CMV and a red fluorescent protein (RFP) under the control of a prokaryotic promoter penp. pMG36e-EGFP-mcherry encodes EGFP-mcherry under the control of a prokaryotic promoter penp. pMG36e-penp-OVA encodes OVA protein under the control of penp. pMG36e-CMV-penp-OVA encodes OVA protein under the control of CMV and penp. pNZ8149-penp-OVA encodes OVA protein under the control of penp. The L.L was cryopreserved in this laboratory. L.L and recombinant L.L carrying pMG36e were grown statically at 30°C in M17 broth supplemented with 10% glucose (GM17), recombinant L.L of pMG36e was grown in GM17 with 1 mg/ml erythromycin (Solarbio, China). Recombinant L.L carrying pNZ8149 was grown statically at 30°C in M17 broth supplemented with 10% lactin, recombinant L.L of pNZ8149 were grown in M17 with lactin and 50 μg/ml ampicillin (Solarbio, China).

L.L at logarithmic (log) phase was collected and resuspended in 1 ml lithium acetate solution (100 mM LiAc; 10 mM DL-Dithiothreitol; 0.6 M sucrose; 10 mM Tris-HCl, pH 7.5) (Shenggong, Shanghai, China). After incubation for 30 min at room temperature, cells were washed three times with sterile deionized water and resuspended in 2 ml of 0.3 M sucrose with 10% glycerol. Cells can be used for electroporation or stored at -80°C for future use.

1 μg of plasmid was mixed with 80 μl cells in a 2 mm electroporation cuvette (BIO-RAD, USA). The voltage of BTX electro cell manipulator was set at 400 V, the time was 4 ms and post fusion AC seconds were 1 (Harvard apparatus, USA). After electroporation, 1ml recovery medium (900 μl M17, 100 μl 5% glucose or 5% lactin, 10 μl 20 mM MgCl₂ and 2 mM CaCl₂) (Shenggong, Shanghai, China) was added and the cells were kept on the ice for 5 min and incubated at 30°C for 2 h. After centrifugation at 4000 rpm for 5 min, part of the supernatant was discarded, 200 μl cells were plated on Elliker-medium (EM) with 1mg/ml erythromycin or 50 μg/ml ampicillin and cultured for 24-48 h at 30°C.

The 293T cells were cultured in 6 well plates and transfected when the cell density was 60%-70%. Before the transfection, the culture medium was replaced with fresh one. The transfection system was included A (2 μg plasmid, 125 μl Opti-MEM) and B [3.75 μl Lipo3000, 4 μl P3000, 125 μl Opti-MEM (Thermo Fisher, USA)]. The A and B tubes were evenly mixed and kept at room temperature for 15 min. The mixture was slowly added to 6 well plates and mixed. The culture medium was replaced after 8 h. After 36 h, cells were collected and total protein was isolated by the RIPA Lysis Buffer (Beijing ComWin Biotech Co., Ltd). The expression of related proteins was detected by Western blot. The expression of EGFP proteins was also observed by inverted fluorescence microscope (Nikon, Japan).

For analyzing the activation of MAPK and NF-κB signaling pathways, GM-DCs were treated with 10² L.L and 10³ L.L for 30 min or 120 min and proteins were extracted by nucleoprotein and cytoplasmic protein extraction kit (Biorebo Biotech, China). 40 ng/mL LPS (Sigma-Aldrich, USA) was used as the positive control.

For detection of EGFP and OVA expression, the recombinant L.L with pNZ8149-penp-OVA (L.L-OVA) was cultured at 30°C to OD_595nm of 0.2-0.6. Then 10 ng/ml nisin was added and cultured for 0 h, 9 h, 12 h and 24 h at 30°C. The recombinant L.L with pMG36e is auto-inducible. The recombinant L.L was harvested by centrifugation (12000 rpm, 10 min), and the proteins were obtained by liquid nitrogen grinding. The expression of related proteins was detected by Western blot, including EGFP (TransGen Biotech, China) and OVA (Elabscience, China). After incubation with HRP-conjugated goat anti-mouse IgG, or HRP-conjugated goat anti-rabbit IgG (TransGen Biotech, China), the target proteins were detected using EasySee Western blot kit (TransGen Biotech, China).

Granulocyte/macrophage colony-stimulating factor (GM-CSF) derived DCs (GM-DCs) were induced from bone marrow cells of C57BL/6 mice referring to our previous description (31). Briefly, bone marrow cells were isolated from femurs and tibias of C57BL/6 mice and cultured in RPMI-1640 containing 10% fetal bovine serum, 1% penicillin-streptomycin, 2 mM L-glutamine, 50 μM β-mercaptoethanol and 20 ng/mL GM-CSF (PeproTech) at 37°C in 5% CO₂ incubator. Medium was changed every other day. On day 7, cells were collected and counted by blood cell count plate counting. CD11c⁺CD86⁺ were used to identify GM-DCs following differentiation. GM-DC populations were analyzed by flow cytometry after staining with anti-CD11c-FITC, anti-CD11b-PerCP/Cyanine5.5, anti-CD45-PE, anti-CD103-APC (Elabscience, China) and anti-CD24-v450 (Thermo Fisher, USA) according to previous studies (4, 32). 5 × 10⁵ GM-DCs were treated with different concentrations (10, 10², 10³) of L.L for 12 h, 24 h or 48 h. 40 ng/mL LPS (Sigma-Aldrich, USA) was used as the positive control. Cells were collected and stained with fluorescence-labeled antibodies including anti-CD40-FITC, anti-CD86-APC, anti-CD80-APC, anti-MHC I-FITC, anti-MHC II-APC (Elabscience, China) or anti-25-D1.16-APC (Thermo Fisher, USA). Apoptosis of GM-DCs was analyzed by AnnexinV-FITC/propidium iodide (PI) apoptosis detection kit according to the manufacturer’s instruction (YEASEN, China). ROS production was measured by the reactive oxygen species assay kit according to the manufacturer’s instruction (Biyuntian, China), and cells were stained with DCFH-DA. For detection of GM-DC endocytosis, GM-DCs were treated with 10³ fluorescent protein recombinant L.L or CFSE-stained L.L. Samples were collected by FACSCalibur (BD Biosciences, USA) and analyzed by FlowJo (Tree Star, Inc., Ashland, OR).

To investigate the roles of TLR2/4 in GM-DC maturation, cells were pretreated with 100 ng/mL TLR2 blocking antibody (mAb-TLR2) or 1 μM TAK-242 (TLR4 inhibitor) for 1 h, and washed with PBS, then treated with different doses (10, 10², 10³) of L.L for 24 h. For NOX2 inhibitor, ROS scavenger and endocytic inhibitor treatment, GM-DCs were pretreated with 5 μM Diphenyleneiodonium chloride (DPI) (Medchem-express, USA), 10 mM N-Acetylcysteine (NAC) (Sigma-Aldrich, USA) and different doses (0.1, 0.3, 0.5 and 0.7 μg/ml) of cytochalasin D for 12 h (Yuanye, China), then treated by 10³ L.L or recombinant L.L for 6 h or 24 h. For proteasome inhibitor treatment, GM-DCs were pretreated with 10 μg/ml MG132 for 30 min and co-treated with 10³ L.L-OVA for 24 h, then another 10 μg/ml MG132 was added after 12 h.

To further confirm the roles of MAPK, JNK and ERK in L.L-induced GM-DC maturation, GM-DCs were pretreated with inhibitors of p38 MAPK (SB202190), ERK (U0126) and JNK (SP600125) for 2 h, 1 h and 45 min, respectively, and then treated with 10³ L.L for 24 h.

After treatment, the supernatant was collected to detect cytokine secretion by enzyme-linked immunosorbent assay (ELISA) using an ELISA kit according to the manufacturer’s instruction (Boster, China) and nitric oxide (NO) by the NO colorimetric assay kit according to the manufacturer’s instruction (Biyuntian, China).

To evaluate the intracellular localization of the CFSE-stained L.L (L.L-CFSE) and EGFP-mcherry, GM-DCs were treated with 10³ L.L-CFSE and recombinant L.L with pMG36e-EGFP-mcherry (LEGFP-mcherry). After treatment for 2 h, 4 h or 6 h, cells were collected and fixed with 4% paraformaldehyde at room temperature for 10 min, then washed with PBS for 3 times. After treatment with 0.1% trionx-100, EEA1 and LAMP1 (diluted with PBS at 1:100; Affinity, USA) were added and incubated at 4°C overnight. After washing with PBS for 3 times, goat anti-rabbit IgG (diluted with PBS at 1:300) was added and incubated at 37°C for 2 h. Then DAPI was added and kept at room temperature for 10 min (Biyuntian, China) after washing with PBS. Finally, samples were imaged using laser confocal microscope (Nikon, Japan).

GM-DC treated with (10, 10², 10³) L.L, 40 ng/mL LPS was used as a positive control. 2×10⁵ GM-DC (cells resuspended in 0.5 mL complete RPMI-1640 medium) was added to the upper chamber of transwell, and 0.5 mL complete RPMI-1640 medium containing 100 ng/mL CCR19 was added to the lower chamber of transwell. After 3 h, cells in the subchamber of transwell were collected and counted.

GM-DC migration in vivo was performed as previously described (33). Briefly, GM-DC treated with 10³ L.L, 10³ recombinant L.L, or 40 ng/mL LPS were stained with 10 μM CFSE and injected in footpad of mice. After 24 h, popliteal LNs were isolated to analyze CFSE⁺ GM-DCs by flow cytometry.

GM-DC-based vaccine was prepared according to our previous study (34). Briefly, GM-DCs were treated with 10³ L.L, 10³ recombinant L.L or 40 ng/mL LPS for 24 h and then the LPS groups were incubated with 10 μg/ml OVA_257-264 and OVA_323-339 peptides, or 10 μg/ml OVA protein for 2 h, respectively. GM-DCs treated with L.L were used as negative control.

Naïve C57BL/6 mice (5 mice/group) were intradermally injected with GM-DC-based vaccines (1 × 10⁶ GM-DCs in 50 μl PBS) twice at 2 weeks intervals, including LPS treated GM-DCs pulsed with OVA protein (LPS+OVA), LPS treated GM-DCs pulsed with OVA peptides (LPS+OVApep), L.L treated GM-DCs (L.L), L.L-OVA treated GM-DCs (L.L-OVA), MG132 pretreated GM-DCs treated with L.L-OVA (L.L-OVA+MG132) and DPI pretreated GM-DCs treated with L.L-OVA (L.L-OVA+DPI). 7 days after the second injection, mice were sacrificed to isolate spleens and inguinal LNs to detect immune responses. Splenocytes were stained with fluorescence-labeled antibodies against CD3, CD19, CD49b, CD11c, CD86, CD11b, Ly6C, Ly6G, CD4, CD25 and Foxp3 (BD Biosciences), and lymphocytes were stained with fluorescence-labeled antibodies against CD4, CD8, IFN-γ and Granzyme B (BD Biosciences).

For evaluation of antitumor effect of GM-DC-based vaccine, 5 × 10⁵ B16-OVA cells were subcutaneously injected into the right flank of C57BL/6 mice. After 6 d, tumor mice were randomly divided into 6 groups (6 mice/group) and treated with GM-DC-based vaccines including LPS+OVA, LPS+OVApep, L.L, L.L-OVA, L.L-OVA+MG132 and L.L-OVA+DPI. GM-DCs were intradermally injected into mice twice at 1 week interval. Tumors were measured every other day using callipers and calculated using the following formula: tumor volume (mm³) = (length × width²)/2. On day 24, mice were sacrificed, and organs and tumors were isolated and weighted. Inguinal LNs were isolated to detect immune responses according to the above description.

Data were reported as mean ± standard error of the mean (SEM). One-way analysis of variance (ANOVA) or a paired t-test was used to examine the statistical difference. Data were analyzed with Prism5 GraphPad software (USA). p < 0.05 was considered to be statistically significant.

According to the growth curve ( Supplemental Figure 1A ), colony-forming unit (CFU) calculation of L.L was carried out when OD_595nm was 0.5 at the log phase. The number of viable bacteria per milliliter of the bacterial solution was 10¹⁰ by counting the plate clone numbers and the CFU is 10¹⁰ CFU/ml. Firstly, the safe doses of L.L were determined based on the viability of GM-DCs. After treatment with different doses (10, 10², 10³) of L.L for 24 h, GM-DC viability was not affected according to the frequencies of apoptotic and necrotic GM-DCs ( Supplemental Figure 1A ), suggesting that the selected doses is safe. Secondly, the phagocytosis ability of GM-DCs for L.L was detected. Compared with L.L-treated GM-DCs, the fluorescence intensities of L.L-CFSE-treated GM-DCs were significantly increased at 3 h, 6 h and 9 h ( Supplemental Figure 1B ). The similar results were observed by laser confocal microscope ( Supplemental Figure 1C ). Thirdly, the effect of L.L on GM-DC maturation was measured. After treatment for 24 h, L.L significantly increased the expression of CD40, CD86, CD80, MHC I, MHC II and CCR7, and the secretion of TNF-α, IL-1β, IL-12p40 and IL-6, in a dose-dependent manner. The results of transwell experiment in vitro showed that L.L can significantly promote GM-DC migration ( Supplemental Figure 2A ). Consistently, 10³ L.L significantly enhanced GM-DC migration in vivo, characterized by the increased frequencies of CD11c⁺CFSE⁺ cells in draining LN ( Supplemental Figure 2B ). Moreover, high doses of L.L significantly upregulated NO production ( Supplemental Figure 2C ). The results demonstrated that L.L promoted GM-DC maturation and activation.

To investigate the mechanism of GM-DC maturation, cells were pretreated with or without TLR2 blocking antibody (mAb-TLR2) for 1 h, and then treated with different doses (10, 10², 10³) of L.L for 24 h. The expression of CD40, CD86 and IL-12p40 in GM-DCs induced by L.L was significantly reduced by mAb-TLR2 pretreatment ( Figure 1A ), indicating that L.L promoted GM-DC maturation partially through TLR2 signaling pathway. The expression and activation of key molecules in TLR2 downstream pathways were further analyzed by Western blot. L.L treatment significantly enhanced the phosphorylation of p38, ERK, JNK and IKB-α at 120 min and the phosphorylation of NF-κBp65 at 30 min and 120 min ( Figure 1B ). To further confirm the roles of p38, JNK and ERK in L.L-induced GM-DC maturation, GM-DCs were pretreated with inhibitors of p38 MAPK (SB202190), ERK (U0126) and JNK (SP600125) for 2 h, 1 h and 45 min, respectively, and then treated with L.L for 24 h. SP600125, U0126 and SB202190 pretreatment significantly inhibited the expression of CD40, CD80 and IL-12p40 ( Figure 1C ). TLR4 inhibitor (TAK-242) pretreatment did not suppress the expression of CD40, CD86 and IL-12p40 in GM-DCs induced by L.L but significantly inhibited their expression induced by LPS ( Supplemental Figure 3 ). These results suggested that L.L promoted GM-DC maturation through TLR2 and downstream MAPK and NF-κB signaling pathways.

In order to study the mechanism of antigen processing and presentation delivered by L.L in GM-DCs, recombinant vectors expressing fluorescent proteins including pMG36e-RFP-EGFP (358) and pMG36e-EGFP-mcherry were constructed and protein expression was identified by Western blot and confocal laser scanning microscopy. For identification of EGFP expression of pMG36e-RFP/eGFP (358) under the control of CMV, the plasmid was encapsulated by liposome and transfected into 293T cells, then the protein was extracted and identified by Western blot. The expected band of EGFP with 29 KDa was obtained. The EGFP fluorescence was further observed in 293T cells by confocal laser scanning microscopy ( Supplemental Figure 4A ). pMG36e-RFP/eGFP (358) and pMG36e-EGFP-mcherry were transferred into L.L and named as L358 and LEGFP-mcherry, which were confirmed by PCR ( Supplemental Figures 4A, B ). The expression of EGFP-mcherry in recombinant L.L was identified by Western blot. The expected band of EGFP-mcherry with 55 KDa was obtained ( Supplemental Figure 4B ). After co-culture of LEGFP-mcherry and GM-DCs, green fluorescence was observed in GM-DCs by confocal laser scanning microscopy ( Supplemental Figure 4B ), suggesting that LEGFP-mcherry was endocytosed into GM-DCs.

Recombinant vectors expressed OVA including pMG36e-penp-OVA, pMG36e-CMV-penp-OVA and pNZ8149-penp-OVA were constructed and transferred into L.L, which were named as LpMG36e-penp-OVA, LpMG36e-CMV-penp-OVA and L.L-OVA, respectively, and confirmed by PCR ( Supplemental Figures 4C–E ). OVA expression of LpMG36e-penp-OVA was identified by Western blot and the correct band was observed. For identification of OVA expression by pMG36e-CMV-penp-OVA, the plasmid was encapsulated by liposome and transfected into 293T cells, and then OVA protein band was obtained by Western blot. L.L-OVA was induced by nisin for 9 h, 12 h and 24 h, respectively, and the protein expression was detected by Western blot. The correct bands were observed ( Supplemental Figures 4C–E ).

According to the growth curve ( Figure 2A ), LEGFP-mcherry and L358 entered the log phase around 10 h and 4 h, respectively, and reached the plateau around 16 h and 10 h, respectively. LEGFP-mcherry and L358 at log phase were used to treat GM-DCs for 3 h, 6 h, 9 h, 12 h and 24 h, and the fluorescence intensities of EGFP were detected by flow cytometry. Compared with untreated control, the fluorescence intensities of EGFP were significantly increased from 3 h to 12 h in GM-DCs treated with LEGFP-mcherry and L358 ( Figure 2B ), suggesting that antigen not only was delivered into GM-DCs by L.L but also sustained at least for 12 h. EGFP fluorescence observed in GM-DCs treated with L358 demonstrated that pMG36e-RFP-EGFP (358) was released into the cytoplasm due to EGFP under the control of CMV. We further observed that EGFP proteins were dispersed in GM-DCs treated with LEGFP-mcherry for 6 h ( Figure 2B ), and did not co-localize with EEA1 and LAMP1 by laser confocal microscope ( Figure 2C ). These results suggested that L.L promoted the rupture of the phagosome membrane to release antigen into cytoplasm.

High levels of ROS can mediate phagosome membrane rupture and release antigens into the cytoplasm to enhance the antigen cross-presentation of GM-DC via the phagosome-to-cytosol pathway (25). Therefore, ROS production was measured in GM-DCs by flow cytometry after treatment with different doses of L.L for 6 h. As shown in Figure 3A , ROS production was significantly increased in a concentration-dependent manner. To explore the role of ROS production in antigen processing, NOX inhibitor, DPI, was used to suppress ROS production in phagosomes. We found that 5 μM DPI significantly reduced the fluorescence intensity of CFSE in GM-DCs treated with L.L-CFSE ( Figure 3B ). NAC (10 mM), a broad spectrum ROS inhibitor, also significantly decreased the fluorescence intensity of CFSE in GM-DCs treated with L.L-CFSE ( Figure 3B ).

We further analyzed the subcellular localization of L.L-CFSE with EEA1 and LAMP1 in GM-DCs treated with or without DPI by confocal laser microscope. L.L-CFSE was observed without DPI treatment and did not co-localize with EEA1 and LAMP1. This is consistent with the results in Figure 3C . However, no L.L-CFSE was observed after DPI treatment, indicating that most of the L.L-CFSE was degraded in GM-DCs ( Figure 3C ). The dynamic subcellular localization of L.L-CFSE was further investigated in GM-DCs. Cells were treated with L.L-CFSE for 2 h, 4 h and 6 h and stained with EEA1 and LAMP1, then observed by laser confocal microscope. As shown in Figure 4A , L.L-CFSE co-localized with EEA1 at 2 h, some L.L-CFSE co-localized with EEA1 and others leaked into the cytoplasm at 4 h, and almost L.L-CFSE did not co-localize with EEA1 and leaked into the cytoplasm at 6 h. L.L-CFSE did not co-localize with LAMP1 at 2 h, but some L.L-CFSE co-localized with LAMP1 and others leaked into the cytoplasm at 4 h, then almost L.L-CFSE did not co-localized with LAMP1 at 6 h ( Figure 4B ). These results indicated that ROS decreased the degradation rate of antigen and promoted the release of antigen in GM-DCs through the phagosome-to-cytosol pathway.

To further investigate whether L.L enters GM-DCs through endocytosis, GM-DCs were treated with L.L-CFSE for 6 h in the presence or absence of cytochalasin D and analyzed by flow cytometry. As shown in Supplemental Figure 5A , cytochalasin D significantly reduced the amount of L.L-CFSE in GM-DCs in a dose-dependent manner. The subcellular localization of L.L-CFSE was detected in GM-DCs with or without 0.5 μg/ml cytochalasin D. After treatment for 6 h, cells were stained with EEA1, LAMP1 and DAPI and detected by confocal laser scanning microscopy. In the absence of cytochalasin D, lots of L.L-CFSE entered GM-DC and did not co-localize with EEA1 and LAMP1. This is consistent with the results in Figure 4 . However, in the presence of cytochalasin D, L.L-CFSE did not appear in GM-GM-DC ( Supplemental Figure 5B ). Furthermore, mAb-TLR2 significantly reduced the amount of L.L-CFSE in GM-DCs ( Supplemental Figure 5C ). These results demonstrated that L.L entered GM-DC through endocytosis and released into cytoplasm through the phagosome-to-cytosol pathway.

The GM-DC used in our experiment was cultured on day 7. However, we cannot exclude whether GM-DC or GM-CSF-derived macrophages are responsible for the observations with this type of culture. To further test whether OVA recombinant L.L enhances antigen cross-presentation, GM-DCs were treated with LpMG36e-penp-OVA, LpMG36e-CMV-penp-OVA and L.L-OVA for 6 h, 12 h, 24 h, 36 h and 72 h and stained with 25.D1-16-APC, which was used to specifically detect MHC I-OVApep complex on the surface of GM-DCs. As shown in Figure 5A , all the three OVA recombinant L.L significantly increased the fluorescence intensity of 25.D1-16 in a time-dependent manner and L.L-OVA generated the highest level of MHC I-OVApep complex. Therefore, L.L-OVA was selected for the next experiments. The roles of ROS production and proteasome in the generation of MHC I-OVApep complex were detected. GM-DCs were treated with L.L-OVA in the presence or absence of 5 μM DPI for 24 h and stained with 25.D1-16-APC. The fluorescence intensity of 25.D1-16 was significantly decreased by DPI ( Figure 5B ), indicating that ROS production induced by L.L promoted the release of OVA from phagosomes to cytoplasm. For the treatment of proteasome inhibitor, MG132, GM-DCs were pretreated with 10 μg/ml MG132 for 30 min and co-treated with L.L-OVA for 24 h, then another 10 μg/ml MG132 was added after 12 h. The fluorescence intensity of 25.D1-16 was significantly reduced by MG132 ( Figure 5C ), indicating that the released OVA was degraded by proteasome in cytoplasm. Moreover, the selected dose of DPI and MG132 did not induce the apoptosis and necrosis of GM-DCs after treatment for 24 h ( Supplemental Figure 6A ), and did not affect GM-DC migration in vitro and vivo ( Supplemental Figures 6B, C ). These results suggested that L.L-OVA enhanced antigen cross-presentation by ROS production and proteasome processing.

Our previous study showed that GM-DC vaccine prepared with peptides generated stronger cellular immune responses (31). However, MHC restriction of peptides limits their clinical use. GM-DCs have a poor endocytosis capacity for soluble protein. Therefore, we develop recombinant L.L expressing protein antigen as the new antigen delivery system for GM-DC vaccine. LPS is a TLR4 agonist to promote GM-DC maturation. LPS-treated GM-DCs pulsed with OVA protein or peptides were used as negative or positive controls to compare with L.L-OVA. The antigen-specific T cell responses were detected in mice immunized with these GM-DC vaccines. 14 days after boosting, inguinal LNs were isolated to analyze immune responses. Upon OVA protein stimulation, LPS+OVApep and L.L-OVA significantly increased the frequencies of OVA-specific CD4⁺IFN-γ⁺, CD8⁺IFN-γ⁺ and CD8⁺Granzyme B⁺ T cells compared with L.L group. L.L-OVA+MG132 and L.L-OVA+DPI significantly reduced the frequencies of CD4⁺IFN-γ⁺, CD8⁺IFN-γ⁺ and CD8⁺Granzyme B⁺ T cells compared with L.L-OVA group. This was consistent with the results in Figure 5 . Importantly, L.L-OVA significantly enhanced OVA-specific CD8⁺IFN-γ⁺Granzyme B⁺ T cell responses compared with L.L group, but LPS+OVA and LPS+OVApep did not induce the responses ( Supplemental Figure 7 ). These results indicated that GM-DC vaccine prepared with L.L-OVA enhanced antigen-specific immune responses.

The antitumor effect of GM-DC vaccine prepared with L.L-OVA was further tested in B16-OVA tumor mouse model. All GM-DC vaccines did not affect the body weight of mice. Compared to L.L group, L.L-OVA and LPS+OVApep significantly inhibited tumor growth ( Supplemental Figure 8 and Figure 6A ). L.L-OVA also significantly reduced tumor weight ( Figure 6A ). On day 24, spleens were isolated to detect immune responses. Compared with L.L group, L.L-OVA significantly increased the frequencies of T cells and NK cells in spleens of mice ( Figure 6B ). L.L-OVA and LPS+OVApep significantly decreased the frequencies of inducible regulatory T cells (iTregs: CD4⁺CD25^-Foxp3⁺) and natural Tregs (nTregs: CD4⁺CD25⁺Foxp3⁺). However, the frequencies of iTregs and nTreg in L.L-OVA+MG132 and L.L-OVA+DPI groups were significantly increased compared with L.L-OVA group ( Figure 6B ). As expected, LPS+OVApep and L.L-OVA significantly enhanced OVA-specific CD4⁺IFN-γ⁺, CD8⁺IFN-γ⁺ and CD8⁺Granzyme B⁺ T cell responses compared with L.L group. L.L-OVA also significantly enhanced OVA-specific CD8⁺IFN-γ⁺Granzyme B⁺ T cell responses ( Figure 6C ). These immune responses are positively correlated with the inhibition of tumor growth ( Figure 6D ). L.L-OVA+MG132 and L.L-OVA+DPI significantly reduced CD4⁺IFN-γ⁺, CD8⁺IFN-γ⁺ and CD8⁺IFN-γ⁺Granzyme B⁺ T cell responses compared with L.L-OVA group ( Figure 6C ). These results indicated that GM-DC vaccine prepared with L.L-OVA had a strong antitumor effect.