Naringenin Ameliorates Acute Inflammation by Regulating Intracellular Cytokine Degradation

The immune system is critical for the maintenance of response to infection and tissue homeostasis. Upon recognition of microbial products by specific receptors, the innate and adaptive immune systems are well coordinated to mount specific responses that are characterized by immune cell activation, proliferation, and differentiation, as well as cytokine production, all of which lead to local and systemic inflammation (1, 2). The inflammatory response serves to further engage the immune system to promote the elimination of pathogens and the induction of tissue repair; however, if allowed to proceed unchecked, it may cause acute inflammatory disorders, such as endotoxemia and fulminant hepatitis. In addition, inflammation has been implicated in the pathogenesis of other chronic diseases, such as diabetes, atherosclerosis, and cancer. These emphasize the urgent need for novel therapeutic strategies to control acute inflammation and restore immune homeostasis.

Because accumulating evidence suggests that excessive production of TNF-α and other inflammatory cytokines play a pivotal role in acute inflammation, there has been enthusiasm about the development of anticytokine agents in recent years. A number of injectable protein-based TNF-α inhibitors have been proved to be effective in treating autoimmune disease, such as rheumatoid arthritis, inflammatory bowel disease, psoriasis, and others. However, potential adverse effects of protein-based anti–TNF-α therapies, such as activation of latent tuberculosis and increased incidence of cancer, have been observed in patients (3, 4). Compared with protein-based biologics, small molecule inhibitors that target the specific signaling pathways involved in TNF-α and other cytokine synthesis have significant advantages and are in various stages of preclinical and clinical development (5). Molecular mediators of TNF-α secretion can also be used as targets for blocking TNF-α release. These strategies that may offer more constrained or cell type–specific control of TNF-α secretion are being sought to improve anti–TNF-α therapies.

Naringenin, a natural predominant flavanone, possesses a broad range of pharmacological activities (6). Our previous study showed that, in the model of bleomycin-induced pulmonary fibrosis, naringenin was capable of modulating the fibrotic microenvironment and correcting fibrosis-deregulated cytokine production (7). Because acute inflammation and upregulation of inflammatory cytokines are involved in the early phase of bleomycin-induced pulmonary fibrosis (8, 9), we hypothesized that the beneficial effects of naringenin in pulmonary fibrosis are mediated by its anti-inflammatory action. Therefore, we aimed to investigate the potential anti-inflammatory effects and the therapeutic use of naringenin. To test the effects of naringenin on acute inflammation, we used several well-established murine models of acute inflammation (endotoxemia and fulminant hepatitis) and demonstrated that naringenin ameliorates the acute inflammatory response by attenuating the cytokine storm in vitro and in vivo. Unexpectedly, we also found that naringenin inhibits cytokine secretion by posttranslational mechanisms, which differ from several other reported flavonoid natural products (10–12). This study may provide a novel therapeutic approach for acute inflammatory disorders. More importantly, this study reveals that promoting intracellular cytokine degradation may be an attractive strategy for developing a new generation of anticytokine therapy.

Naringenin, cycloheximide, and actinomycin D were purchased from Sigma (St. Louis, MO), as was LPS (from Escherichia coli, serotype O55:B5). All other TLR ligands were obtained from InvivoGen. Brefeldin A was obtained from eBioscience, and the TACE inhibitor, TAPI-1, was purchased from Calbiochem. The murine macrophage cell line RAW264.7 was obtained from the American Type Culture Collection and maintained in DMEM containing 10% FBS.

For naringenin treatment, daily doses of 100, 200, or 500 mg/kg naringenin suspended in 0.5% sodium carboxymethyl cellulose or only 0.5% sodium carboxymethyl cellulose (200 μl per mouse) was administered orally for 4 d. Three hours after the last administration of naringenin, mice were challenged with different reagents (described below) to establish models of acute inflammation. For testing the therapeutic potential of naringenin, mice were gavaged with a single dose of naringenin (100 mg/kg) 15 min after LPS injection.

For LPS-induced endotoxemia, 6–8-wk-old female C57BL/6 mice were injected i.p. with different doses of LPS (Sigma). For Con A–mediated acute liver inflammation, 6–8-wk-old female BALB/c mice were injected i.v. with 15 mg/kg Con A (Sigma). For LPS/N-acetyl-galactosamine (GalN)-induced acute liver injury, 6–8-wk-old female C57BL/6 mice were injected i.p. with 100 ng/kg LPS plus 750 mg/kg GalN (Sigma). For TNF-α/GalN–induced acute liver injury, 6–8-wk-old female C57BL/6 mice were injected i.v. with 15 mg/kg TNF-α (PeproTech) plus 750 mg/kg GalN. For Jo2-induced liver injury, 6–8-wk-old female BALB/c mice were injected i.p. with 0.4 mg/kg Jo2 Ab (BD Biosciences). At different time points postinjection, serum, lung, and liver were collected for biochemical and histopathological studies. Mice weighing 18–22 g were from the Institute of Biophysics, Chinese Academy of Sciences and were housed in plastic cages in groups of six in the Animal Resource Service facilities, which is approved by the Institution Animal Care and Use Committee at the Institute of Biophysics, Chinese Academy of Sciences. All animals were allowed to acclimatize for 1 wk prior to experimentation.

To isolate peritoneal macrophages, 10 mice were injected i.p. with 3% thioglycollate (1.5 ml per mouse; Sigma), and cells were harvested 4 d after injection. Cells were resuspended in X-VIVO 15 (Lonza) or DMEM (Life Technologies) supplemented with 10% heat-inactivated FBS, penicillin (100 U/ml), and streptomycin (100 U/ml) and cultured at a density of 2 × 10⁶ cells per milliliter per well in 24-well plates. After 3 h, the cells were washed to remove nonadherent cells, and adherent cells were used for quantitative real-time PCR (qRT-PCR), ELISA, Western blot, flow cytometry, and microscopy studies.

For macrophages, cells were treated with DMSO or naringenin (100 μM) in the presence of LPS for 24 h. Then, cell viability was determined using a CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega). For T cells, splenocytes were treated with DMSO or different doses of naringenin in the presence of anti-CD3 for 24 h. Cell viability was determined by Annexin V/7-aminoactinomycin D (eBioscience) exclusion after gating on CD4 or CD8 cells.

Total RNA from tissue or cells was isolated with a PureLink Micro-to-Midi Total RNA Purification System and subjected to SuperScript III Platinum reverse transcriptase using oligo(dT) primers (all from Invitrogen), according to the manufacturer’s protocol. We quantified specific cytokine mRNAs using a SYBR Green–based qRT-PCR kit (Invitrogen) with the ^ΔΔCT method. GAPDH mRNA was used as an internal loading control for all samples. All primer sequences are from Primer Bank.

For mRNA stability, peritoneal macrophages were pretreated with naringenin (100 μM) or diluent (DMSO) for 2 h, followed by stimulation with LPS (100 ng/ml) for 2 h. Subsequently, actinomycin D (10 μg/ml) was added to stop the synthesis of new RNA. Total RNA was isolated at different time points after actinomycin D treatment, and mRNA levels of TNF-α and IL-6 were determined by qRT-PCR.

Cytokine levels in the supernatants, serum, and peritoneal fluids were measured using Mouse Matched Ab Pairs for ELISA or a FlowCytomix Mouse Th1/Th2 10plex kit (both from Bender MedSystems).

Peritoneal macrophages were treated with naringenin or DMSO for 2 h and then stimulated with PBS or LPS for various times. After washing, cells were collected, and nuclear lysates were prepared using a NucBuster Protein Extraction Kit (Novagen), according to the manufacturer’s instructions. p50 and p65 activity in the nuclear lysates was measured using an NFκB p50/p65 EZ-TFA Transcription Factor Assay (Millipore), and phospho–c-jun activity was measured using a TransAM AP-1 c-Jun ELISA Kit (Active Motif).

Cells were cultured at 37°C in a 5% (v/v) CO₂ atmosphere. The growth medium was removed, and the cells were rinsed with prewarmed methionine-free medium without serum (isotope medium) and incubated for 10 min. The rinse and incubation steps were repeated two times. ³⁵S-labeled methionine (50 μCi/ml; PerkinElmer) was added to isotope medium with 10% dialyzed serum. At the same time, DMSO or 100 μM naringenin was added to the medium. The cells were incubated for 2 h and then stimulated with 100 ng/ml LPS for an additional 2 h. Cellular protein was collected with lysis buffer (Cell Signaling), and the radioactive counts were read using a PerkinElmer 450 Wallac MicroBeta TriLux LSC and Luminescence Counter.

Total cell lysates were prepared using cell lysis buffer (Cell Signaling) with protease inhibitor mixture I (Sigma). Western blot analysis was performed using Abs against IκBα (44D4), phospho-IκBα (14D4), JNK (56G8), phospho-JNK (81E11), ERK1/2 (137F5), phospho-ERK1/2 (D13.14.4E), p38, phospho-p38 (D3F9), STAT1, phospho-STAT1, and GAPDH (14C10), all of which were from Cell Signaling, with the exception of TNF-α Ab, which was purchased from Abcam.

For surface or intracellular TNF-α staining, peritoneal macrophages were seeded onto glass bottom dishes, treated with naringenin (100 μM) or diluent overnight (or for 2 h), and stimulated with LPS (1000 ng/ml) plus IFN-γ (20 ng/ml) in the presence or absence of brefeldin A or TAPI-1 for 2 h. Then cells were fixed in 4% paraformaldehyde for 15 min. After fixation, cells were blocked with blocking agent containing 5% normal goat serum (Invitrogen) and Fc blocker (eBioscience) in PBS and stained with allophycocyanin-conjugated rat anti-mouse TNF-α (eBioscience), in the presence or absence of 0.1% Triton X-100 for 2 h. Confocal microscopy was performed using an Olympus FV500 confocal microscope.

For surface staining, cells were stained with fluorescently conjugated Abs in staining buffer for 20 min at 4°C (2% FBS and 0.09% sodium azide in PBS). For intracellular cytokine staining, cells were surface stained, fixed, and permeabilized with a Cytofix/Cytoperm Kit (BD Biosciences) and stained with fluorescently conjugated anti-cytokine Abs for 30 min at room temperature. The following Abs were used: F4/80 (BM8), CD16/32 (93), CD4 (GK1.5), CD8a (53-6.7), CD25 (PC-61.5), CD44 (IM7), CD69 (H1.2F3), MHC class I (28-14-8), MHC class II (M5/114.15.2), IFN-γ (XMG1.2), TNF-α (MP6-XT22), and IL-2 (JES6-5H4) (all from eBioscience). F4/80⁺ cells were identified as macrophages. Cells were analyzed with a FACSCalibur (BD Biosciences), and data were analyzed with FlowJo (TreeStar).

Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were quantified using an Aspartate Aminotransferase Kit or an Alanine Aminotransferase Kit (BioSino). Caspase-3/7 protease activity in the lung or liver was measured using an Apo-ONE Homogeneous Caspase-3/7 Assay (Promega), according to the manufacturer’s instructions.

Formalin-fixed and paraffin-embedded sections of tissue samples (lung and liver) were stained with H&E using standard protocols. Apoptosis in the liver was evaluated by a TUNEL assay using an ApopTag Plus Fluorescein In Situ Apoptosis Detection Kit or an ApopTag Plus Peroxidase In Situ Apoptosis Detection Kit (both from Millipore), according to the manufacturer’s instructions.

The transduction complex was formed using a Neofect DNA transfection reagent (Neofect Biotech, Beijing, China) and a TFEB-specific short hairpin RNA (shRNA) construct consisting of 5 μg of shRNA containing pLKO.1 (transfer plasmid, TRCN000085549; Sigma), 1.5 μg of pMD2.G (envelope plasmid), and 3.5 μg of psPAX2 (packing plasmid). For viral transduction of shRNA, HEK293T cells were transfected with this complex for 48 h. After transfection, the supernatant was collected, sterilized using a 0.45-μm filter, and stored at −80°C. To knock down TFEB, peritoneal macrophages were treated with 1:1 mixed medium (X-VIVO 15/virus containing HEK293T supernatant) for 24 h and then macrophages were used for LPS stimulation. The efficiency of TFEB knockdown by shRNA was assessed using qRT-PCR. The primers for TFEB are forward 5′-GCTCCAACCCCGAGAAAGAG-3′ and reverse 5′-CAGCGTGTTAGGCATCTGC-3′.

Results are presented as mean ± SD. The Student t test was used to compare values between groups. We generated survival curves using the Kaplan–Meier method and determined the significance of the survival rates using the log-rank test. A p value <0.05 was considered statistically significant.

We have previously shown that naringenin ameliorates bleomycin-induced pulmonary fibrosis in which an early phase of acute inflammation critically contributes to the later phase of fibrosis (7). Therefore, we hypothesized that the beneficial effects of naringenin in pulmonary fibrosis may be mediated through its ability to suppress the inflammatory response. Because the central role of cytokines, such as TNF-α, IL-6, and IFN-γ, in mediating systemic inflammation is well recognized, we sought to determine whether naringenin regulates the production of inflammatory cytokines from cultured primary macrophages and T cells. As shown in Fig. 1A, LPS-induced TNF-α and IL-6 release from peritoneal macrophages was significantly inhibited by naringenin treatment. Moreover, this inhibitory effect on cytokine release was also found in RAW264.7 cells (a murine macrophage-like cell line) in a dose-dependent pattern (Fig. 1B, 1C). Notably, the effect of naringenin was not limited to LPS, which is a TLR4 ligand; it also inhibited TNF-α and IL-6 release in response to other TLR ligands, including Pam3CSK4 (TLR-1), heat-killed Listeria monocytogenes (TLR2), polyinosinic-polycytidylic acid (TLR3), flagellin (TLR5), FSL1 (TLR6), ssRNA40 (TLR7), and CpG (TLR9) stimulation (Fig. 1D). In addition to macrophages, naringenin strongly inhibited the release of cytokines, such as IFN-γ and TNF-α, in CD4⁺ T cells in response to anti-CD3 and anti-CD28 stimulation (Fig. 2). Taken together, our data demonstrated that naringenin inhibits cytokine release in macrophages and T cells. Interestingly, the inhibitory effects of naringenin on cytokine release are independent of cell type and ligand stimulation. Moreover, the cell viability and the expression of activation-associated molecules in macrophages or T cells were not affected by naringenin treatment (Supplemental Fig. 1), indicating that naringenin might specifically regulate certain processes of cytokine production and/or secretion, rather than the general functions of macrophages and T cells.

Next, we tried to investigate the mechanism underlying the inhibitory effect of naringenin on cytokine release, which involves a series of control processes at different levels, including transcription, posttranscription, and posttranslation. First, to understand whether naringenin-mediated inhibition of cytokine release is due to decreased expression of cytokine transcripts, we measured mRNA levels of several cytokines in macrophages and T cells by qRT-PCR. Surprisingly, the mRNA levels of TNF-α, IL-6, IL-10, and IFN-γ were not influenced by naringenin treatment at early time points of LPS treatment (Fig. 3A, Supplemental Fig. 2A, 2D) when corresponding protein levels were dramatically reduced by naringenin treatment (Fig. 1A). Correspondingly, the LPS-induced activation of MAPK family members ERK, JNK, p38, and transcription factor AP-1, which regulate cytokine transcription activation, were also not significantly affected by naringenin treatment in macrophages (Fig. 3B, 3C). Given that the NF-κB signaling pathway is the master regulator of TNF-α expression, we also examined whether naringenin influences NF-κB activation by comparing the dynamics of IκBα degradation and p65/p50 activity. At an early time point, NF-κB activation remained intact upon naringenin treatment. Interestingly, at a late time point (1 h after LPS stimulation), attenuated NF-κB activation in response to LPS stimulation was observed in naringenin-treated macrophages (Fig. 3D). Recent studies have demonstrated that NF-κB activation depends on the TLR ligand (LPS), as well as on TNF-α that is induced by the TLR ligand itself, and serves as an autocrine feedback (13, 14). Thus, attenuated NF-κB activation observed in naringenin-treated macrophages at the late time point is most likely due to the reduction in TNF-α secretion at the early time point. In addition to these positive regulators of cytokine production, several negative regulators have been described (15). We tested whether naringenin regulates the expression of some inhibitors of cytokine production, such as SOCS1, SOCS3, IRAK-M, Cbl-b, and SHIP (15). However, no difference in the expression of these inhibitors was observed between control and naringenin-treated macrophages (Supplemental Fig. 2E). Collectively, these results suggest that naringenin, differing from several other natural products (such as apigenin, Supplemental Fig. 2D), exerts its effect on cytokine release through mechanisms that are independent of cytokine gene transcription. Last, because the stability of cytokine mRNA is also involved in the regulation of cytokine synthesis, we tested whether naringenin affects TNF-α and IL-6 mRNA stability in macrophages. As shown in Fig. 3E, the half-life of TNF-α and IL-6 mRNA in naringenin-treated macrophages was similar to that in untreated macrophages.

To circumvent any interference of naringenin with a transcriptional or translational event, peritoneal macrophages were stimulated with LPS for 2 h and then the macrophages were incubated with actinomycin D (to block new cytokine mRNA synthesis) or cycloheximide (to block new cytokine protein synthesis), in the presence or absence of naringenin for an additional 2 h. As shown in Fig. 4A, naringenin treatment still inhibited TNF-α release, even in the presence of actinomycin D and cycloheximide. The fact that naringenin did not affect the synthesis of new proteins (Fig. 4B) further supports the notion that naringenin does not affect global protein synthesis.

Several recent studies have demonstrated that newly synthesized TNF-α and IL-6 accumulate in the Golgi complex and then are transported to the cell surface for release (16–18). We next asked whether naringenin influences intracellular trafficking of cytokines. We used brefeldin A, which accumulates TNF-α within the endoplasmic reticulum, and TAPI-1 (a TACE specific inhibitor), which blocks TNF-α release from the cell surface. As expected, treatment with brefeldin A and TAPI-1 led to significant accumulation of TNF-α in the cytoplasm and cell surface, respectively (Fig. 4C–E). Remarkably, naringenin significantly reduced TNF-α accumulation on the cell surface that was induced by TAPI-1 (Fig. 4E), whereas it had no impact on the brefeldin A–induced intracellular accumulation of TNF-α (Fig. 4D, 4F). These results suggest that naringenin may inhibit the Golgi-to–cell surface delivery of TNF-α without affecting its synthesis. Flow cytometry analysis further confirmed this finding (Fig. 4G). To test the possibility that narigenin’s inhibition of cytokine release is due to a general effect on protein secretion, we tested whether naringenin can inhibit the secretion of other secretory proteins. As shown in Supplemental Fig. 3, naringenin did not inhibit secretion of several secretory proteins in adipocytes, suggesting that it is unlikely that naringenin serves as a general inhibitor of protein secretion. The fact that naringenin inhibits TNF-α release, but does not lead to the intracellular accumulation of TNF-α, indicates that accelerated degradation occurs downstream of the Golgi complex. Overall, these results suggest that naringenin exerts its effect on cytokine production through posttranslational mechanisms.

Prompted by the fact that naringenin inhibits cytokine release without inducing the accumulation of cytokines in the cells, we next asked whether naringenin can enhance cytokine intracellular degradation. Several recent studies (19) have reported that newly synthesized TNF-α and IL-6 accumulate in the Golgi complex and then are transported to recycling endosomes, where they are directed to the cell membrane for release or to lysosomes for degradation. To determine whether naringenin promotes the lysosome-dependent degradation of cytokines, we used two specific lysosome inhibitors, bafilomycin A1 and NH₄Cl, which are able to raise lysosomal pH. Remarkably, bafilomycin A1 and NH₄Cl treatment blunt the effect of naringenin, indicating that inhibition of cytokine release by naringenin is lysosome dependent (Fig. 5A, 5B). Similar results were obtained for IL-6 (Fig. 5C). We also used a panel of inhibitors targeting distinct lysosome proteases to determine whether inhibition of any lysosome protease would blunt the effect of naringenin. Some of the inhibitors can indeed partially restore cytokine production in naringenin-treated macrophages, but none of them could fully blunt the naringenin effect (Fig. 5D), suggesting that naringenin may promote the overall activity of lysosomes rather than target a specific protease in lysosomes. Because TFEB is the master regulator of lysosome function, we hypothesized that naringenin may promote the activity of TFEB to degrade TNF-α. Indeed, immunofluorescence analysis showed that naringenin treatment promoted TFEB nuclear translocation, which is a key step for its activation (Fig. 5E). Remarkably, knockdown of TFEB in primary macrophages completely blunted the naringenin effect and restored cytokine production (Fig. 5F, 5G). However, we also noticed that knockdown of TFEB only partially restored cytokine production in naringenin-treated cells after 16 h of LPS and naringenin treatment (Fig. 5H, inhibition effect of naringenin decreases from 39 to 21%), indicating that other mechanisms also contribute to the overall effect of naringenin after longer treatment, potentially through inhibition of cytokine gene expression, as reported previously by other groups (34, 35). We estimated that ∼46% of naringenin’s effect is TFEB sensitive, which represents a degradation-inducing mechanism of naringenin, whereas 54% of the drug’s effect is TFEB insensitive, which is likely mediated through transcription repression. Collectively, our data suggest that naringenin promotes lysosome-dependent cytokine degradation by activating TFEB.

In addition, at the 2-h time point, the TNF-α level increased from ∼75 pg per 10⁵ cells in control cells to 110 pg per 10⁵ cells in TFEB-knockdown cells (compare the first bar and third bar in Fig. 5G), indicating that ∼32% of TNF-α produced after 2 h of LPS treatment is subjected to lysosome-mediated degradation, and naringenin increases the percentage to 70% (compare the second bar and fourth bar in Fig. 5G). However, after 16 h of LPS treatment, only 22% of cytokines produced are subjected to lysosome degradation (compare the first bar and third bar in Fig. 5H), and naringenin increases the percentage to 34% (compare the second bar and fourth bar in Fig. 5H). These data collectively suggest that lysosome-mediated degradation contributes more to the overall regulation of TNF-α during the early phase of LPS stimulation than during the late phase (32 versus 22%).

Because naringenin effectively blocks inflammatory cytokine release in macrophages and T cells in vitro, we next tested whether naringenin can be used to ameliorate acute inflammation in vivo. Bolus injection of the TLR4 agonist LPS into mice is a widely used model for endotoxemia, which has pathophysiological alterations similar to those observed in sepsis patients (11). C57BL/6 mice were orally gavaged with naringenin (100 and 500 mg/kg/d) or diluents for 4 d, followed by i.p. injection of a lethal dose of LPS (50 mg/kg), and survival was monitored. Pretreatment of mice with naringenin significantly protects against motility induced by LPS (Fig. 6A). Strikingly, 100% of LPS-challenged mice died within 84 h (n = 12), whereas more than half of mice pretreated with naringenin (n = 12) survived after 120 h of observation. Consistent with the in vitro observation in primary macrophages, examination of the serum levels of various inflammatory cytokines (TNF-α, IL-6, IFN-γ, IL-5, and IL-17a) and the peritoneal fluid level of TNF-α revealed that naringenin pretreatment significantly lowers levels of cytokines in vivo (Fig. 6B, 6D). In agreement with our in vitro results, naringenin did not affect mRNA levels of several cytokines in the liver from LPS-treated mice (Supplemental Fig. 2B). Furthermore, to test the therapeutic potential of naringenin in systemic inflammation, we administered naringenin orally after the onset of acute inflammation. As shown in Fig. 6C, this therapeutic protocol also produced a significant improvement in survival rates. Naringenin administration of 500 mg/kg/d did not show better efficacy than 100 mg/kg/d; therefore, 100 mg/kg/d was chosen as the dosage for subsequent in vivo studies.

High-dose injection of LPS (10 mg/kg) into mice induces systemic inflammation and multiorgan damage, particularly in the lung. Therefore, we tested whether naringenin protects lung from LPS-induced injury. As expected, pretreatment of mice with naringenin ameliorated LPS-induced lung inflammatory responses, as well as pulmonary cell apoptosis (Fig. 6E, 6F). Administration of a low dose of LPS (100 ng/kg) in the presence of liver-specific transcriptional inhibitor GalN causes acute inflammation and hepatocyte apoptosis. Therefore, we examined whether naringenin also shows beneficial effects in this model. As shown in Fig. 6G, administration of naringenin substantially reduced liver inflammation and hepatocyte apoptosis in LPS/GalN-treated mice.

To further explore the therapeutic efficacy of naringenin in acute inflammation, we tested its efficacy in Con A–induced acute liver inflammation. Unlike LPS-induced endotoxemia, which is mediated mainly by activated macrophages, Con A induces systemic activation of T cells, leading to acute liver inflammation and fatal hepatitis (20). As shown in previous reports (20–22) and Fig. 7A, BALB/c mice challenged i.v. with Con A (15 mg/kg) showed abnormally elevated ALT and AST levels in the serum, which is indicative of extensive liver damage. Remarkably, Con A–challenged mice pretreated with naringenin showed much lower levels of ALT and AST in the serum compared with mice receiving diluent (Fig. 7A). Upon examination of liver pathology, mice challenged with Con A showed profound liver damage and hepatocyte apoptosis, which, as expected, was significantly reduced by pretreatment with naringenin (Fig. 7B). IFN-γ serum levels were dramatically elevated in mice with Con A–mediated acute liver inflammation, which critically depends on T cell–derived IFN-γ and its downstream signaling pathway JAK-STAT1 (23). Mice pretreated with naringenin showed a markedly lower serum level of IFN-γ 6 h after Con A challenge compared with control mice (Fig. 7C), and this did not affect its mRNA level (Supplemental Fig. 2C). Consistent with this, activation of STAT1 was also attenuated in the Con A–challenged mice pretreated with naringenin (Fig. 7D). In acute liver inflammation, the elevated levels of various cytokines and chemokines enhance the migration of lymphocytes to the liver, which further contributes to liver damage. Naringenin pretreatment significantly reduced the hepatic lymphocyte infiltration in Con A–challenged mice compared with mice without naringenin pretreatment (Fig. 7E, 7G). Consistent with a posttranslational effect, the expression of activation-associated molecules in T cells, such as CD69 and CD25, were not altered by naringenin, further demonstrating that naringenin does not affect Con A–induced T cell activation (Fig. 7F). Clearly, these results demonstrate that naringenin protects against acute inflammation in vivo by lowering the levels of inflammatory cytokines. To exclude the possibility that the beneficial effect of naringenin may also be derived from the protection of hepatocytes against the destructive inflammatory response, we used the model of Fas-induced fatal hepatitis, which is mediated by Fas–Fas ligand (FasL) killing of hepatocytes. Jo2, an agonistic Ab to mouse CD95, was injected into mice and induced hepatocyte apoptosis (Supplemental Fig. 4). However, naringenin failed to protect against liver injury in this model, as judged by survival rate and caspase-3/7 activity in the liver homogenates (Supplemental Fig. 4). In addition, naringenin pretreatment only provided slight protection against liver injury induced by direct injection of TNF-α/GalN (Supplemental Fig. 4), further strengthening the notion that naringenin ameliorates acute inflammation by potently lowering cytokine levels. Taken together, our data demonstrate that administration of naringenin has obvious benefits for LPS-induced endotoxemia and Con A–mediated acute liver inflammation.

In this study, we demonstrated that naringenin ameliorates acute inflammatory response in vitro and in vivo. Naringenin potently inhibited inflammation in LPS-induced endotoxemia and Con A–induced hepatitis, which are mediated primarily by innate and adaptive immune responses. Importantly, we also showed that the beneficial effect of naringenin is mediated through direct inhibition of inflammatory cytokine production in macrophages and T cells. Interestingly, naringenin inhibits TNF-α and IL-6 secretion without reducing their mRNA expression, implicating a posttranscriptional control mechanism for the action of naringenin in these cells. We also provided several additional lines of evidence to support the posttranscriptional control mechanism of naringenin for the inhibition of cytokine production. First, naringenin had a similar inhibitory effect on TNF-α and IL-6 production in macrophages stimulated with a variety of TLR agonists. Furthermore, naringenin also inhibited TNF-α and IFN-γ production in T cells with CD3/CD28 stimulation. These results suggest that the inhibitory effect of naringenin on cytokine release is not limited to specific stimuli. Second, naringenin did not affect the early-phase activation of NF-κB and MAPK, which are key signaling pathways responsible for TLR-induced gene expression. Third, naringenin did not induce some reported negative regulators of cytokine gene expression. Finally, naringenin (at concentrations up to 100 μM) did not have any impact on the viability of macrophages and T cells. This is in agreement with the doses used in several other studies investigating the effect of naringenin on apolipoprotein B (apoB) secretion in hepatocytes (24–27) (discussed below). These results strongly indicate that naringenin modulates cytokine production during posttranscriptional steps.

With regard to posttranscriptional control steps, we were also able to rule out some possibilities. First, naringenin did not have an impact on the stability of cytokine mRNA. Second, translation of cytokine protein and synthesis of global protein were not affected by naringenin. These results indicate that naringenin modulates cytokine production at the posttranslational steps. Normal cytokine synthesis, coupled with lower intracellular levels and decreased release, indicates that accelerated degradation may occur. The fact that the addition of lysosome inhibitors blunted the effect of naringenin suggests that lysosomes are involved, at least in part, in the naringenin-mediated reduction in cytokines. Our finding that naringenin acts in a specific way to reduce the production of a subset of cytokines, rather than serves as a global inhibitor of protein synthesis or secretion, may be useful in acute inflammation-related diseases.

Inflammatory cytokines, such as TNF-α, IL-6, and IFN-γ, can serve in host defense or, paradoxically, induce inflammatory tissue injury. Thus, to limit the undesirable consequences, several transcriptional and posttranscriptional mechanisms are coordinated to precisely control the production of cytokines. Although transcription is an essential first step for cytokine expression, different levels of posttranscriptional control mechanisms can provide rapid and precise regulation of this process (11). In most cases, posttranscriptional control of cytokine production is mediated by modulating mRNA decay or protein translation (28). However, given that every cytokine needs to be secreted out of the cell to exert its effects, it is conceivable that further control mechanisms may exist during the intracellular transport of cytokines. However, it is difficult to study the process of cytokine trafficking because of the transient nature of intracellular trafficking events. Recently, a series of studies, using high-resolution microscopy and live cell imaging, have identified some exocytosis pathways responsible for TNF-α and IL-6 trafficking and secretion (16, 17, 29, 30). For example, activated macrophages can upregulate some trafficking proteins, such as syntaxin 4 and VAMP-3, to facilitate TNF-α secretion (16, 29). In agreement, another study also demonstrated that activated T cells use distinct pathways for the secretion of different cytokines (31). Characterizing these pathways responsible for cytokine trafficking and degradation would support the development of new anticytokine drugs. Thus, naringenin specifically promotes cytokine degradation, which may provide modality for drug development in this field.

One unexpected feature of naringenin-mediated inhibition of cytokine production is that naringenin exerts its effect through a posttranslational mechanism. This feature stands in sharp contrast to that of some widely used anti-inflammation agents, such as glucocorticoids, and some other natural products, such as apigenin, curcumin, and luteolin, which act transcriptionally on cytokine production by targeting MAPK or NF-κB (10, 32–34) (Supplemental Fig. 2D). Thus, naringenin, which acts at the level of posttranslation, might be a good candidate for use in combination with other anti-inflammatory agents to control inflammatory disorders. This diversity in the regulation of cytokine production may be due to the structural variations in flavonoids. Generally, the flavonoids with strong anti-inflammatory effect through transcriptional regulation are characterized by necessary structural features, such as four hydroxylations at positions 5, 7, 3′, and 4′, a double bond at C2–C3, and the position of the B ring at C2 (35). In contrast, naringenin has no double bond between C2 and C3, indicating that the formation of a chemical bond between C2 and C3 may determine the mechanism of action of these natural products. Further structure and activity relationship studies will be critical to better understand how these natural products regulate the immune response. Whether other flavonoids have similar effects certainly warrants further investigation.

Previous studies have shown that naringenin decreases the secretion of apoB and very low–density lipoprotein from hepatocytes by inhibiting the activity of microsomal triglyceride transfer protein, which transfers lipid onto apoB (24, 26, 27). Consistent with these findings, another study showed that naringenin, without changing the transcription of viral RNA, inhibits the assembly and release of hepatitis C virus, which critically depend on microsomal triglyceride transfer protein and apoB (25, 36). These studies, coupled with our observation that naringenin promotes cytokine intracellular degradation, strongly suggest that the cellular targets of naringenin may be some mediators responsible for intercellular transport or degradation of proteins and lipids from the Golgi complex to the cell surface. However, it is possible that, in our acute inflammation models, the protective effect of naringenin is through reducing cytokines, as well as via enhancing the survival capability of the cells in the liver or lung. However, as shown in Supplemental Fig. 4, naringenin shows no protective effect in models of FasL- or TNF/GalN-induced liver injury, both of which occur primarily through direct induction of organ damage rather than excessive cytokine production, as in the LPS and Con A models.

Thus, we propose a model in which naringenin reduces cytokine production by promoting TFEB- and lysosome-dependent intracellular degradation. TFEB is a master regulator of lysosomal and autophagic function (37). The activity of TFEB is controlled by its phosphorylation, which sequesters it in the cytosol as an inactive form (38, 39). Under certain environmental cues, such as nutrient limitation, TFEB is dephosphorylated and translocated to the nucleus to promote the transcription of its target genes involved in lysosome function. Interestingly, it has been reported that TFEB is rapidly activated in macrophages during infection, suggesting that it may be involved in acute inflammation and cytokine secretion (40). Knockdown of TFEB completely blunted the cytokine-reducing effect of naringenin (Fig. 5G), strongly suggesting that naringenin inhibits cytokine release by promoting TFEB activity. Indeed, we observed that naringenin promotes TFEB nucleus translocation (Fig. 5E) but does not affect its expression (data not shown) in macrophages upon LPS stimulation, indicating that naringenin may, directly or indirectly, promote dephosphorylation of TFEB and, thereby, enhance its nucleus translocation and activity. The detailed mechanism regarding how naringenin regulates TFEB nucleus translocation certainly warrants further investigation. In addition, it remains unclear how selective degradation of different cytokines by naringenin is achieved. Because our results showed that the effect of naringenin on cytokine degradation is TFEB dependent, it is conceivable that the kinetics of TFEB translocation and activation upon LPS stimulation would likely determine the specificity of degradation of different cytokines, because the expression kinetics for different cytokines in response to LPS varies. Kinetic analysis of the activation of TFEB and the degradation of different cytokines in response to LPS would shed new light on the regulation of selective degradation of each cytokine in macrophages.

Lysosome and proteasome are two major intracellular proteolytic systems, and they are assumed to serve different functions (41). In macrophages, lysosomes play an important role in the regulation of TNF-α stability before its secretion (42). Perturbation of lysosomal function, especially the acidification process, would cause deregulated cytokine production (43). In contrast, because recycling endosomes and probably other vesicles are involved in the secretion of cytokines, cells may engage lysosomes as another control step for cytokine production. Thus, targeting lysosomal function might be of therapeutic interest for inflammatory disorders. Given the fact that posttranslational regulation of cytokines has been largely overlooked in the development of novel anti-inflammation agents, we expect that, in light of our findings, additional small molecular modulators targeting this newly discovered process will be identified. This study is the first critical step linking the new discoveries to long-term work elucidating the molecular mechanism of intracellular cytokine degradation and developing the pharmacological means to control acute inflammation in the clinic.

In conclusion, this study provides evidence that naringenin might represent an attractive therapeutic candidate for treatment of acute inflammatory diseases. Future studies should examine whether naringenin can also be effective in other disease conditions involving the deleterious actions of inflammatory cytokines, such as rheumatoid arthritis and multiple sclerosis.

We thank Linlin Lu, Liyan Ji, Shuo Zhang, Zihui Song, and Lijuan Yu for technique assistance and Prof. Wei Feng (Institute of Biophysics) for kindly providing the pEGFP-N1-TFEB plasmid. We are grateful to Dr. Junfeng Hao and Zhenwei Yang (Institute of Biophysics Core Facility Center) for technical assistance and scientific discussion.

The authors have no financial conflicts of interest.