Naringenin Enhances the Antitumor Effect of Therapeutic Vaccines by Promoting Antigen Cross-Presentation

Dendritic cells (DCs) are capable to present Ags to T cells and orchestrate antitumor immune responses, lying at the center of the cancer-immunity cycle (1). Despite other professional APCs and other DC subsets that exist in tumor microenvironment, conventional DCs (cDCs) are particularly adept to activate T cells in the tumor (2). Additionally, type 1 cDCs can preferentially deliver and cross-present tumor-associated Ag on MHC class I (MHC-I) to CD8⁺ T cells, whereas type 2 cDCs preferentially present Ag on MHC-II to CD4⁺ T cells (3). Despite resident type 1 cDCs (defined CD8⁺ as DCs in mice) were originally considered to be responsible for cross-presenting tumor-associated Ags, it has been increasingly clear that migratory type 1 cDCs (defined as CD103⁺ DCs in mice) are necessary for tumor-associated Ag delivery and cross-presentation and are unique to robustly induce cytotoxic CD8⁺ T cells response against tumors (4).

Different from classical presentation pathway, cross-presentation is the only effective way to present tumor-associated Ag on MHC-I complex to initiate Ag-specific antitumor CD8⁺ T cell responses (5). Accordingly, inadequate cross-presentation of tumor-associated Ag by DCs was found during tumor progression (6). To achieve adequate antitumor efficacy, numerous efforts were made to enhance cross-presentation efficacy, such as targeting DC-specific surface molecules for precise Ag delivery (7, 8), seeking help from heat-shock proteins (9), enhancing the immunogenicity of vaccine Ags via peptide modification and protein recombination (10), preventing lysosomal proteolysis of Ags either by blocking acidification of the endocytic system with lysosomotropic agents, such as chloroquine or ammonium chloride (11), or inducing DC differentiation and maturation with cytokines or adjuvants (12–14). Recently, a novel strategy reported was to directly deliver Ags into the cytosol of DCs by a lymph node–prone nanoparticle delivery system (15, 16).

Flavonoids are one of the largest groups of secondary plant metabolites, which have been documented to prevent chronic inflammation, cardiovascular diseases, neurodegenerations, and cancer, largely because of their antioxidant properties (17). In addition, flavonoids also harbor pro-oxidant activity, leading to the production of reactive oxygen species (ROS) and other organic radicals that may oxidize intracellular lipids (18). The pro-oxidative power of flavonoids is also thought as a result of interfering mitochondrial electron transport (19).

As one of the most abundant flavonoids in diets, naringenin can act as an immunomodulator to prevent TLR-induced inflammatory cytokines secretion in both macrophages and DCs and regulate DC maturation and regulatory T cell differentiation (20–22). But how naringenin affects cross-presentation by DCs remains unclear. In the current study, we observed that naringenin (as well as other flavonoids) possessed the ability to promote the Ag cross-presentation in DCs. Considering the importance of Ag cross-presentation for antitumor therapeutic vaccines, we employed naringenin as the representative to investigate the mechanism of such interesting action of flavonoids, and we identified a new strategy to improve the efficacy of antitumor therapeutic vaccines by moderately increasing intracellular ROS.

Genistein, luteolin, apigenin, and naringenin were purchased from Shanxi Huike Botanical Development in China, and the purity of these compounds is higher than 98%. All flavonoids were dissolved in DMSO as stock solutions at a concentration of 100 mM. Anti-biotin Ab was purchased from Abcam, and the other Abs were purchased from eBioscience. OVA was purchased from Sigma-Aldrich, monophosphoryl lipid A (MPLA) was purchased from Avanti, and the SL-15 peptide was chemically synthesized by LifeTein, China.

The DC2.4 cell line was a gift from the Li Tang laboratory (Key Laboratory of Protein Engineering, Academy of Military Medical Sciences) and was maintained in RPMI 1640 medium supplemented with 10% heat-inactivated FBS plus 1% 2-ME. TC-1 cell line was from Dr. X. Xu (Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and School of Basic Medicine, Peking Union Medical College) and was maintained in RPMI 1640 medium supplemented with 10% FBS. Cells were cultured at 37°C in a humidified atmosphere of 5% CO₂. All cell lines were found to be negative for mycoplasma contamination, and we checked morphology, growth kinetics, and Ag expression to validate them.

Bone marrow–derived DCs were obtained using the method as described in Inaba et al. (23), with some modification. Briefly, total bone marrow cells were collected from female C57BL/6 mice (Beijing Vital River Laboratory Animal Technology) with RBCs removed using Ammonium–Chloride–Potassium Lysis Buffer and then cultured in RPMI 1640 medium supplemented with penicillin, 2-ME, 10% FBS, and GM-CSF (20 ng/ml; PeproTech) for 7 d. The culture medium was replaced by fresh medium with GM-CSF (20 ng/ml) at day 3, and nonadherent cells in the culture supernatant were harvested on day 7. MHC-I restricted, OVA-specific (OTI) CD8⁺ T cells were purified from the spleen of OTI TCR transgenic mice (Bar Harbor, ME) by using a CD8a⁺ T Cell Isolation Kit (Miltenyi Biotec, Bisley, U.K.) according to the manufacturer’s instructions.

DC2.4 cells were plated at a density of 1 × 10⁴ cells per well in 96-well plates and cultured for 24 h. Cells were then exposed to a series of concentrations of tested flavonoids for another 24 h, then the culture medium was discarded, and 100 μl Thiazolyl Blue Tetrazolium Blue (MTT, final concentration at 0.5 mg/ml in PBS, σ) was added to each well. After incubation at 37°C for 4 h, the MTT solution was removed, and 100 μl of DMSO was added to each well for 10 min at room temperature. Absorbance was recorded at 590 nm by a plate reader (Thermo Multiskan MK3).

For the intracellular lipid peroxidation assay, DC2.4 cells were loaded with 1 μM BODIPY 581/591–C11 (Thermo Fisher Scientific) in RPMI 1640 medium supplemented with 10% FBS plus 1% 2-ME, and 20 μM of the tested flavonoids were added together with BODIPY 581/591–C11. After incubation for 24 h, the BODIPY 581/591–C11 fluorescence was measured by FACSCalibur (BD Biosciences, San Jose, CA).

DC2.4 cells or bone marrow–derived cells (BMDCs) were loaded with 20 μM 2′,7′-dichlorofluorescin diacetate (DCFH-DA; Thermo Fisher Scientific) for 30 min before further flavonoids treatment. Then, the mean fluorescence intensity was measured with FACSCalibur (BD Biosciences). Note that serum in the culture medium would interfere with the detection in this study; hence, flavonoids were diluted in serum-free RPMI 1640 medium.

For the confocal imaging of the colocalization of lysosome and FITC-OVA, DC2.4 cells were grown at a density of 1.5 × 10⁵ cells per dish in glass-bottom petri dishes, and at the next day, the cells were incubated at 37°C for 2 h 100 μg/ml FITC-OVA in the presence or absence of 100 μM naringenin, then cells were incubated with 75 nM LysoTracker Red DND-99 for 15 min after being washed three times. Subsequently, cells were washed twice, and images were acquired from three or more randomly chosen fields using a confocal microscope LSM-700 (ZEISS, Germany).

For the imaging of intracellular distribution of ROS, DC2.4 cells grown in glass-bottom petri dishes as before incubation at 37°C for 30 min with 20 μM DCFH-DA, then cells were treated with 100 μM naringenin for 1 h after being washed three times. Then, the cells were loaded with 200 nM MitoTracker Deep Red (Thermo Fisher Scientific) for another 15 min. Subsequently, cells were washed twice, and images were acquired from three or more randomly chosen fields using a confocal microscope Olympus FV1000 (Tokyo, Japan). This experiment was performed in serum-free RPMI 1640 medium. DCFH-DA/MitoTracker Deep Red ratio image was generated according to Kardash et al. (24).

DC2.4 cells were treated with 100 μM naringenin or 1 μg/ml LPS or 10 μM MG132 for 8 h, then mRNA was purified using TRIzol Reagent (Thermo Fisher Scientific) and was reverse transcribed using TransScript First-Strand cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). Quantitative PCR was performed using SYBR Select Master Mix (Thermo Fisher Scientific) analyzed on ABI 7300 (Applied Biosystems). Gene expression for PSMA7 (forward [F]: 5′-AACGTCTGTATGGCTTTGC-3′, reverse [R]: 5′-GTCACTGGGTCCTCCACTGT-3′), PSMB4 (F: 5′-TTCACTGGCCACTGGTTATG-3′, R: 5′-CGAACGGGCATCTCTGTAGT-3′), PSMB7 (F: 5′-CTGTCTTGGAAGCGGATTTC-3′, R: 5′-GCAACAACCATCCCTTCAGT-3′), PSMC1 (F: 5′-AAGGGGGTCATTCTCTACGG-3′, R: 5′-AAGCTCTGAGCCAACCACTC-3′), PSMC4 (F: 5′-TGGTCATCGGTCAGTTCTTG-3′, R: 5′-CGGTCGATGGTACTCAGGAT-3′), PSMD1 (F: 5′-GGGGCTTTTGAGGAGTCTCT-3′, R: 5′-GCAAATCTGCATTTTCCACA-3′), and PSMD12 (F: 5′-TCACAGACCTGCCAGTCAAG-3′, R: 5′-AGGTTTTAGTCAGCCGAGCA-3′) were normalized to that of the housekeeping gene GAPDH (F: 5′-AACTTTGGCATTGTGGAAGG-3′, R: 5′-GGATGCAGGGATGATGTTCT-3′). Gene expression for the subunits of mouse proteasome was normalized to that of the housekeeping gene GAPDH. Relative expression values were calculated using the ∆∆cycle threshold method.

For the cross-presentation assay, DC2.4 cells or BMDCs were incubated with OVA protein plus naringenin for the indicated time. Then, cells were washed and stained with PE-conjugated anti-mouse OVA257–264 (SIINFEKL) peptide bound to H-2K^b (25.D1) or double stained with the APC-conjugated anti-mouse CD11c mAb (N418) for BMDCs and analyzed with FACSAria (BD Biosciences).

For the T cells activation assay, OTI CD8⁺ T cells isolated from the spleen of OTI mice using anti-CD8a (Ly-2) MicroBeads (Miltenyi Biotec) were labeled with 5 μM CFSE for 10 min and then neutralized with 20% FBS, whereas BMDCs were treated as indicated and then cocultured with the CFSE-labeled OTI CD8⁺ T cells at a ratio of 10:1 for 4 d. Then, cells were collected, and CFSE dilution in OTI CD8⁺ T cells was analyzed by FACSCalibur (BD Biosciences).

For the isolation and assay for cross-presentation of primary CD103⁺ DCs and CD8a⁺ DCs from MC38 tumor (a murine colon adenocarcinoma), the tumor tissue was dissected and digested by Collagenase IV (Sigma-Aldrich) to make single-cell suspension, followed by multicolor Abs staining, including FITC-conjugated anti-mouse CD11c (N418), BV605-conjugated anti-mouse CD45 (30-F11), PE-conjugated anti-mouse MHC class II (MHC-II; M5/114.152), PerCP–Cy5.5-conjugated anti-mouse CD8a (53-6.7) and APC/Cy7-conjugated anti-mouse CD103 (2E7). The CD45⁺MHC-II⁺CD11c⁺CD103⁺ (CD103⁺ DCs) and CD45⁺MHC-II⁺CD11c⁺CD8a⁺(CD8a⁺ DCs) were sorted by FACSAria, which were treated with OVA in the presence or absence of 100 μM naringenin for 24 h, and then the OVA as well as naringenin were removed. The DCs were subsequently cocultured with CFSE-labeled OTI CD8⁺ T cells for another 4 d before the detection of CFSE dilution in CD8⁺ T cells.

For the leakage of BSA to cytosolic fraction, DC2.4 cells were incubated with 0.5 mg/ml biotinylated BSA (Sigma-Aldrich) plus 5 μM proteasome inhibitor MG132 in the presence or absence of 100 μM naringenin with or without 4 mM N-acetylcysteine (NAC) for 2 h. For the washing out experiments, DC2.4 cells were preloaded with biotinylated BSA in the presence of MG132 for 1 h, then the biotinylated BSA was washed out, and the cells were treated in the presence or absence of 100 μM naringenin with or without 4 mM NAC for another 2 h. The cytosolic fraction was isolated with Cell Fractionation Kit (no. 9038; Cell Signaling Technology), whereas the whole-cell lysate was prepared with Cell Lysis Buffer (no. 9032; Cell Signaling Technology) plus a protease inhibitor mixture (Sigma-Aldrich). Both the cytosolic BSA and BSA in the whole-cell lysate were analyzed by Western blot with anti-biotin Ab (Abcam, Cambridge, U.K.).

For the assay of translocation of β-lactamase was performed using GeneBLAzer In Vivo Detection Kits (no. 12578134; Invitrogen) as described in the instruction manual. Briefly, DC2.4 cells were pretreated with CCF2-AM containing solution mixture for 1 h and then incubated with 1 mg/ml β-lactamase for another 3 h in the presence or absence of 100 μM naringenin with or without 4 mM NAC, and then the cells were analyzed by BD LSRFortessa (BD Biosciences) equipped with a UV laser. For the washing out experiments, DC2.4 cells were first treated with 1 mg/ml β-lactamase for 1 h after loading with CCF2-AM, then the β-lactamase was washed out, and the cells were treated with 100 μM naringenin or 4 mM NAC for another 3 h. The β-lactamase–catalyzed hydrolysis product in cells was quantified by the fluorescence at 450 nm, whereas the fluorescence signal of CCF2 was detected at 535 nm under violet excitation (405 nm).

To evaluate whether naringenin could enhance Ag presentation in vivo, 100 mg/kg naringenin dissolved in 0.3% (w/v) carboxyl methylcellulose (CMC)–Na was orally administrated daily for 7 d. Then, one dosage of OVA vaccine (OVA/SL-15/MPLA = 50/5/5 μg) (w/w/w) was s.c. injected into the right limb of female C57BL/6 mice. Next day, the mice were sacrificed, and the lymph nodes were collected for the analysis of pMHC-I on DCs. Briefly, lymph nodes were minced and then filtered through a 70-μm cell strainer to create a single-cell suspension. Cells were double stained by APC-conjugated anti-mouse CD11c (N418) and PE-conjugated anti-mouse OVA257-264 (SIINFEKL) peptide bound to H-2K^b (also known as pMHC-I), then the percentage of pMHC-I⁺CD11c⁺ were analyzed by FACSAria (BD Biosciences).

For the assay of OVA-specific CD8⁺ T cell response, one dosage of OVA vaccine, as above described, was s.c. injected into the right limb of female C57BL/6 mice, and oral administration of naringenin was maintained at the dosage of 100 mg/kg. Seven days later, cells present in lymph nodes were obtained as above described and were triggered with OVA peptide (OVA250–264, SGLEQLESIINFEKL) overnight. Then, the cells were stained with Abs against CD8a (53-6.7) before fixation/permeabilization and intracellular staining for IFN-γ (XMG1.2). The percentage of INF-γ⁺CD8⁺ T cells was analyzed by FACSCalibur (BD Biosciences).

To evaluate the immunomodulatory effect of naringenin in the tumor-bearing model, 5 × 10⁵ MC38-OTI cells were s.c. inoculated into the right flank of female C57BL/6 mice. Seven days after tumor cell inoculation, daily naringenin was orally administrated at the dosage of 100 mg/kg. Eighteen days later, tumor-infiltrating DCs were stained with Abs against PerCP–Cy5.5-conjugated anti-mouse CD45 (30-F11), PE-conjugated anti-mouse MHC-II (I-A/I-E), FITC-conjugated anti-mouse CD11c (N418), APC-conjugated anti-mouse CD11b (M1/70), APC/Cy7-conjugated anti-mouse CD8a (53-6.7), APC/Cy7-conjugated anti-mouse CD103 (2E7), and PE-Cy7-conjugated anti-mouse pMHC-I (eBio25-D1.16) and then were analyzed by FACSAria (BD Biosciences) to detect the proportion and pMHC-I expression of CD45⁺MHC-II⁺CD11c⁺CD103⁺ migratory DCs and CD45⁺MHC-II⁺CD11c⁺CD11b⁺CD8a⁺ resident DCs. Additionally, 5 × 10⁵ cells from digested tumor tissue were plated and restimulated with OVA peptide (OVA250–264, SGLEQLESIINFEKL) overnight for IFN-γ ELISPOT assay (R&D Systems) according to the manufacturer’s instruction.

A total of 5 × 10⁴ TC-1 cells were s.c. inoculated into the right flank of female C57BL/6 mice. Seven days after tumor cell inoculation, E7 vaccine [E7(20)/MPLA/PEG-PE = 20/10/1000 μg w/w/w] was s.c. injected around the tumor at a frequency of once a week for 3 wk. Eleven days after tumor cell inoculation, 50 mg/kg naringenin dissolved in 0.3% (w/v) CMC-Na was orally administrated daily for 30 d. The tumor volume was recorded twice a week and calculated by the formula V = 1/2 × (length × width²). The E7 vaccines were prepared as previously described (16).

Female C57BL/6 mice were obtained from Vital River Laboratory Animal Technology (Beijing, China). All animal protocols used for this study were approved by the Institutional Animal Care and Use Committee of Institute of Biophysics, Chinese Academy of Sciences.

Statistical analysis was performed using Prism (GraphPad) Software. All p values were calculated by two-tailed unpaired t test for two groups, one-way ANOVA plus Dunnett multiple comparisons for multiple groups, and log-rank (Mantel–Cox) test for survival analysis.

Some flavonoids have been reported to be able to inhibit LPS-induced DC maturation as well as Ag-specific CD4⁺ T cell activation (21, 25). Presumably, flavonoids would suppress Ag presentation by DCs; however, their possible effects on Ag cross-presentation by DCs have so far been ignored. First, we tested the potential impacts of four representative flavonoids on the Ag cross-presentation ability of DCs, including luteolin (flavone), apigenin (flavone), naringenin (flavanone), and genistein (isoflavone). We loaded 2 mg/ml OVA Ags into DC2.4 cells with the presence or absence of above-mentioned flavonoids, and 24 h later, the H-2K^b SIINFEKL bound to MHC-I on the surface of DC2.4 was analyzed by FACS. All the tested flavonoids dramatically increased the OVA cross-presentation (Fig. 1A, Supplemental Fig. 1A), among which naringenin was even as potent as LPS but has negligible cytotoxicity (Fig. 1B, Supplemental Fig. 1B). Thereby, naringenin was selected for further study of the possible mechanism how it facilitated the Ag cross-presentation. Naringenin-induced enhancement of OVA cross-presentation was observed at a various concentration of OVA, and the higher dosage of OVA was correlated with better cross-presentation performance (Fig. 1C). In addition, the increased OVA cross-presentation exhibited a time-dependent manner within 12 h (Fig. 1D). Furthermore, OVA-loaded BMDCs treated with naringenin significantly promoted OTI CD8⁺ T cells proliferation by CFSE dilution assay (Fig. 1E).

To study the mechanism how naringenin promoted Ag cross-presentation of DCs, we first need to rule out the possible impacts of naringenin on the maturation/activation of DCs. It has been observed that naringenin inhibited the activation of BMDCs under LPS-stimulated condition (21). However, in this study, we did not find any inhibition or promotion of TNF-α secretion, a hallmark of mature DC activation, in DC2.4 cells under the treatment of naringenin with or without stimulation of LPS (Fig. 2A). Furthermore, we found naringenin itself only had a marginal effect on the mature and activation of both DC2.4 cells and BMDCs as detected by the expression of the surface markers including costimulatory molecules and MHC-I as well as MHC-II molecules (Fig. 2B–E). In DC2.4 cells, naringenin treatment increased both the expression of MHC-II and CD80 (Fig. 2B, 2C) but only improved the expression of MHC-II in BMDCs (Fig. 2D). In addition, naringenin treatment did not further increase the SL-8 peptide (SIINFEKL)–treated BMDCs induced proliferation of OTI CD8⁺ T cells (Fig. 2F), indicating that naringenin promoted CD8⁺ T cells proliferation largely depends on Ag processing in DCs. Taken together, unlike LPS, naringenin-enhanced Ag cross-presentation of DCs was probably not through promoting the activation of DCs.

There are two central intracellular pathways for the cross-presentation of exogenous Ags, and they are usually referred to as the cytosolic and vacuolar pathways, the key difference between which is that the cytosolic pathway is sensitive to proteasome inhibitors (26). Therefore, we next sought to determine the exact pathway that was involved in the naringenin-enhanced Ag cross-presentation. As shown in Fig. 3A, both MG132 and lactacystin, two proteasome inhibiters, neutralized the enhancement of Ag cross-presentation induced by naringenin, indicating the proteasome-dependent cytosolic pathway was involved in this process. Thus, we subsequently focused on the possible influences of naringenin on the quantity of proteasomes. Unexpectedly, we did not detect any changes of the expression level of subunits of the proteasome in DC2.4 cells under naringenin treatment (Fig. 3B). Instead, naringenin exerted weak inhibitory effect on the activity of proteasome in the in vitro assay of the chymotrypsin-like activity (Fig. 3C), which was reported to be involved in the cross-presentation process of OVA (27). Taken together, the naringenin-enhanced Ag cross-presentation was proteasome dependent, but it was neither directly through the regulation of the gene expression nor the activity of the proteasome.

Next, we assessed whether naringenin treatment would affect the Ag uptake by DCs. As illustrated in Fig. 3D, 100 μM naringenin significantly improved the uptake of FITC-OVA by DC2.4 cells in comparison with vehicle control group. Interestingly, when DC2.4 cells were preloaded with OVA for 4 h and then the Ag was washed out, naringenin still equally enhanced the cross-presentation (Fig. 3E). Similar results were also found in BMDCs (Fig. 3F). Moreover, treatment after OVA washout did not diminish the impact of naringenin on BMDC-induced OTI CD8⁺ T cells proliferation (Fig. 3G). These findings suggested that the naringenin-enhanced cross-presentation was not merely attributed to the increased uptake of Ag.

The most interest in flavonoids has been devoted to their antioxidant activity because of their ability to reduce the free radical formation and to scavenge free radicals in in vitro study (28). Hence, we investigated if the enhanced Ag cross-presentation by naringenin was due to its antioxidant effect. Surprisingly, when incubated with naringenin, the intracellular total ROS in DC2.4 cells substantially increased (Fig. 4A). Besides, NAC effectively relieved the oxidative stress induced by naringenin and accordingly counteracted the effect of naringenin on Ag cross-presentation in DC2.4 cells (Fig. 4A, 4B). Further investigations showed that the naringenin treatment obviously enhanced the endogenous ROS in mitochondria of DC2.4 cells, hinting the origin of intracellular total ROS (Fig. 4C). These results indicated that naringenin-induced endogenous oxidative stress could account for its promotion of Ag cross-presentation. However, the higher level of oxidative stress did not always mean the stronger promotion of Ag cross-presentation. For example, with increasing the dosage of luteolin, the intracellular oxidative stress went up; however, the enhanced Ag cross-presentation was not always consistent with the concentrations of reactive oxygen (Supplemental Fig. 1C). That is easy to understand, because the high level of oxidative stress is toxic to the cells. Luckily, naringenin only induced moderate reactive oxygen production even at high concentration (100 μM) compared with other flavonoids (Fig. 4G, Supplemental Fig. 1D, 1E).

It has been found ROS-maintained alkalization of the phagosomal lumen conferred DCs the ability to process Ag for cross-presentation (29). In this study, we found the specific inhibitor of vacuolar-type H⁺-ATPase, bafilomycin A1, could promote the cross-presentation through alkalization of the lysosome (Fig. 4D, 4E). However, naringenin showed absolutely no influence on the alkalization of the lysosome but had a stronger promotion of OVA cross-presentation than bafilomycin A1 (Fig. 4D–F). Besides, bafilomycin A1 induced the enhancement of Ag cross-presentation could be further augmented by naringenin (Fig. 4E). Our results suggested the promotion of Ag cross-presentation by naringenin-induced oxidative stress was not through the alkalization of the lysosome.

As the typical character of the classical cytosolic pathway, Ag translocation into the cytosol was seen as the critical step for cross-presentation (30). It was reported that lipid peroxidation could cause endosomal Ag release for cross-presentation (31). In this study, we made the discovery that four tested flavonoids, including naringenin, could arouse different degrees of lipid peroxidation in DC2.4 cells (Fig. 4G, Supplemental Fig. 1F). Naringenin indeed strengthened the protein translocation into the cytosol, which was confirmed by both translocation of β-lactamase (Fig. 4H) and biotinylated BSA (Fig. 4I) from endosomes/lysosomes into the cytosol in DC2.4 cells. Furthermore, both naringenin-induced protein translocation into the cytosol and the promoted BMDC-induced CD8⁺ T cells proliferation were blocked by NAC (Fig. 4H–J), suggesting that naringenin’s effect was ROS dependent.

Considering the apparent enhancement of Ag cross-presentation by naringenin in vitro, we next tested its possible effect on the function of DCs in vivo. As shown in Fig. 5A, naringenin significantly promoted DC cross-presentation of OVA in female C57BL/6 mice after oral administration of 100 mg/kg naringenin for 7 d. Moreover, OVA-specific CD8⁺ T cell response was obviously higher in the group with oral administration of naringenin for 7 d after a single immunization of vaccine (Fig. 5B, 5C). Previous research had demonstrated that CD8a⁺ DCs preferentially primed CD8⁺ T cells over CD4⁺ T cells during antitumor immunity (32); however, others also claimed that CD103⁺ DCs was the only APCs priming tumor-specific CD8⁺ T cell responses (4). To explore how and which subset of DCs in tumor was affected by naringenin, we used MC38-OTI (a murine colon adenocarcinoma) model, which harbored OVA(257–264) peptide as tumor-associated Ag. Collectively, naringenin increased DC proportion in tumor-infiltrating leukocytes, which was mainly contributed by migratory CD103⁺ DC, but not resident CD8a⁺ DC (Fig. 5D). Also, CD103⁺ DC in naringenin-treated mice cross-presented more pMHC-I complexes than the DC compartment in control mice (Fig. 5E). Besides, ex vivo study showed that CD103⁺ DC promoted more robust proliferation of OTI CD8⁺ T cells than CD8a⁺ DC (Fig. 5F, 5G). Consequently, SIINFEKL-recognizing CD8⁺ T cells in tumors of naringenin group secreted much more IFN-γ compared with the control group (Fig. 5H), resulting in a certain control of tumor growth in the naringenin treatment group (Fig. 5I). To acquire better antitumor efficacy, we used the TC-1 tumor model to test the possible synergistic effect of naringenin on the therapeutic efficiency of E7 vaccines, because antitumor vaccine was known as a DC-based immunotherapy. As shown in Fig. 5J, tumor growth was better controlled in the mice treated by E7 vaccines in combination with naringenin than that of vaccine treatment alone. Moreover, the combination of E7 vaccine and naringenin significantly prolonged the lifespan of the mice compared with vaccine alone, without increasing additional toxicity (indicated by the change of body weight) (Fig. 5K, 5L). These results provided the evidences to suggest the potential applications of naringenin in combination with therapeutic vaccines to improve their efficacy (Fig. 6).

Therapeutic cancer vaccines have already shown clinical benefits, particularly in combination with other cancer immunotherapies, including checkpoint blockade, etc. (33). These vaccines are often designed to deliver enough tumor-associated Ags to elicit sufficient cytotoxic CD8⁺ T lymphocyte (CTL) responses (34). DCs are regarded as the most efficient APCs (5, 35). Therefore, one important subject in the design of vaccination is how to achieve and promote Ag cross-presentation in DCs (7). In this study, we demonstrated that naringenin not only promoted Ag uptake, but also elevated lipid peroxidation through increasing intracellular oxidative stress in DCs, which led to the translocation of Ags from endosome to the cytosol (31). As a result, more Ags would undergo the proteasome-dependent cytosolic pathway to be presented through cross-presentation. Moreover, naringenin also significantly promoted Ag cross-presentation in vivo, and interestingly, naringenin preferentially increased CD103⁺ migratory cDCs infiltration into tumor, which were the unique DC subset to transport intact Ags to tumor draining lymph nodes and prime CD8⁺ T cells. We proposed that CD103⁺ migratory cDCs might be triggered by naringenin-induced chemokines. However, enhanced tumor-associated Ag-specific CD8⁺ T cell response by naringenin could only control tumor growth in the early phase but was not durable (Fig. 5J), which prompted us to combine naringenin with a DC-based immunotherapy. Fortunately, we had developed an evaluation tool using a therapeutic antitumor vaccine containing E7 Ag to treat TC-1, the human papillomavirus–infected cervical cancer (18). More importantly, we achieved an outstanding efficacy of immunotherapy in TC-1 tumor model when naringenin was used in combination with E7 vaccines (Fig. 5J, 5K).

The connection between intracellular oxidative stress and Ag cross-presentation in DCs is not a new finding. ROS has been reported to enhance MHC-I Ag cross-presentation through either regulation of the pH in the internalization compartments or Ag oxidation for the MHC-I Ag processing (36, 37). Besides, NOX2-produced intracellular ROS were found to be able to induce lipid peroxidation in endosomal membranes, which could cause Ag leakage from the endosomal lumen into the cytosol. Nevertheless, how this idea could be applied to immunotherapy is a big problem, because it is hard to mimic the performance of intracellular ROS by addition of exogenous oxidants with transient and drastic properties. To conquer this problem, one approach of photoswitchable generation of intracellular ROS has been developed for functional manipulation of DCs in vitro, but not in vivo (38). In this study, we revealed flavonoids, as phytonutrients, could also induce the various degrees of intracellular oxidative stress in DCs and, thus, promoted Ag cross-presentation through facilitating Ag leakage from the endosomes. Although the extent of endogenous lipid peroxidation needs further investigation, because of the limitations of fluorescence probe methods, our findings provide an operational methodology to modulate the function of DCs in vivo under the idea of manipulating DC function by fine tuning intracellular ROS level within a noncytotoxic window. Indeed, therapeutic vaccines in combination with naringenin triggered much stronger CTL responses and had better performance in the control of tumor growth without showing additional toxicity.

Usually, excessive intracellular ROS is harmful to physiological cellular functions. For instance, ROS induced by mitochondrial defects in neurons are implicated in neurodegenerative disease (39). However, sometimes ROS also act as essential secondary messengers to mediate biological functions such as LPS-induced inflammatory response in macrophages (40). Similarly, in our experiments, DCs treated with the optimal dosage of luteolin showed the enhanced Ag presentation, whereas the cells could be killed by the high dosage of luteolin (Supplemental Fig. 1C). Apparently, the two-edged weapon of ROS is dependent on its intracellular content. Luckily, the dosage-dependent properties of some flavonoids provide a wide pharmacological window with balanced toxicity and benefit. From this point of view, naringenin could be the best choice to modulate DCs function, because naringenin induced a sustained and moderate intracellular oxidative stress within a wide range of concentrations (20–100 μM).

Different from LPS-induced lysosomal clustering through TLR signal (41), naringenin-enhanced Ag cross-presentation is dependent on a moderate level of lipid peroxidation that increases Ag leakage from endosomes/lysosomes without impairing DC activation (Fig. 6). Yet, the underlying mechanism of naringenin promoting Ag uptake by DCs still remains unknown, and further exploration is needed. These properties of naringenin provided the foundation for combination with therapeutic vaccines with LPS analogs (MPLA) as adjuvants, and as expected, the better antitumor efficacy was achieved (Fig. 5). Furthermore, the reduced use of Ags and classical adjuvants could also be anticipated.

Extensive evidences suggest a possible role of flavonoids as immunomodulators through either regulation of proinflammatory cytokine production or modulation of the activity of immune cells (42). Recently, the potential effects of flavonoids on the function of DCs were paid more attention (43). It was found that lots of flavonoids strongly inhibited LPS-induced DC maturation and activation (25, 43, 44). More recently, naringenin was reported to effectively inhibit LPS-induced DC maturation and, accordingly, abrogated the Ag-specific T cell priming ability of DCs when given at noncytotoxic dosages (21). In contrast, we found naringenin could significantly promote Ag cross-presentation in both DC2.4 cell line and BMDCs and enhance Ag-specific CD8⁺ T cell proliferation in the absence of LPS. These discrepancies may be due to the different sources of the cells and different treatment processes. In fact, we had also observed similar anti-inflammation activity in macrophages and T cells when naringenin was used as a pretreatment before the stimulation of LPS (20). In this study, we pay more attention to the influence of naringenin itself on Ag cross-presentation in DCs instead of the possible effects when naringenin was used in the presence of LPS.

Naringenin, as well as other flavonoids, are most famous for their antioxidant activities, but their antioxidant efficacy in vivo has been less documented. Their protection from cytotoxicity and apoptosis in vitro has been attributed to the antioxidant activities, whereas sometimes, they are seen as proapoptotic agents through pro-oxidant activities (45, 46). Because of such proapoptotic property, some flavonoids have been used as chemotherapeutic agents to enhance the therapeutic antitumor effects of DNA vaccination (47, 48). Therefore, the intracellular oxidative stress aroused by naringenin is crucial for their bioactivities. The detailed mechanism for their pro-oxidant activity and the relationship between cytotoxicity and structure of flavonoids will be reported in our following article.

In general, our findings exploited the potential applications of naringenin (the representative of flavonoids) as an immunomodulator and provided the evidence of how natural products influence human health, considering flavonoids are widespread in our daily diet (49). More importantly, naringenin may have the prominent clinical value in the combination of therapeutic vaccines against cancers or viruses.

We acknowledge both the Core Facility and the Animal Research Center of Institute of Biophysics, Chinese Academy of Sciences. We thank Yihui Xu and Yan Teng for confocal microscopy work and Junying Jia for flow cytometry analysis and cell sorting.

The authors have no financial conflicts of interest.