Aggregated Ursolic Acid, a Natural Triterpenoid, Induces IL-1β Release from Murine Peritoneal Macrophages: Role of CD36¹

Monocytes/macrophages are key mediators of wound repair, tissue remodeling, and inflammation and resident macrophages) are considered to be the front line of immunological defense against pathogens. These cells have prominent roles such as Ag presentation, phagocytic, microbicidal, tumoricidal, and secretory functions, as well as innate immunity, by initiating inflammatory and immune responses (1, 2). Macrophages produce a variety of inflammatory cytokines, including IL-1β, IL-6, TNF-α, and macrophage migration inhibitory factor (MIF)³ (3), which recruit polymorphonuclear leukocytes from the vasculature into the inflammatory site for effective eradication of offending pathogens (4). Although they are essential for the host defense system, excessive production at an inflammatory site may lead to chronic diseases such as inflammatory bowel disease, which includes ulcerative colitis, Crohn’s disease, and neoplasm (5, 6).

IL-1β, an antiapoptotic and proinflammatory cytokine, is one of the most pronounced mediators of inflammatory reactions and is primarily produced in activated monocytes or macrophages (7). Pro-IL-1β, a precursor of IL-1β (8) that is detected in the cytosol of resident macrophages, is biologically inactive with a molecular mass of ∼33 kDa. Upon stimulation, it is cleaved via an enzymatic procession into a 17-kDa mature functional form by the IL-1β-converting enzyme (ICE, also known as caspase-1) (9, 10). Following ICE cleavage, active IL-1β is released from macrophages through the ATP binding cassette transporter (ABC)A1-dependent and -independent pathways (11, 12). Enhanced IL-1β production has been detected at both mRNA and protein levels in human inflammatory bowel disease (13), rheumatoid arthritis (14), and dextran sulfate sodium (DSS)-induced colitis murine model (15). In our previous study, IL-1β levels were found to be profoundly increased in both colonic mucosa and peritoneal macrophages (pMφ) in mice with DSS-induced colitis (16). Thus, pMφ-derived IL-1β may be closely and critically associated with disease pathology.

Triterpenoids are ubiquitously distributed throughout the plant kingdom, and some are increasingly being used for medicinal purposes for a variety of clinical diseases in many Asian countries as antitumor, anti-inflammatory, and immunomodulatory agents (17, 18, 19). However, the molecular mechanisms underlying those activities remain to be fully elucidated. Ursolic acid (UA; 3β-hydroxy-12-urs-12-en-28-oic acid) is a pentacyclic triterpene carboxylic acid found in various plants, including Rosmarinus officinalis and Glechoma hederaceae (19, 20), in the form of an aglycones or as glycosides (19, 20, 21, 22, 23). It is well known to possess many important biological functions, such as anticancer, anti-inflammatory, hepatoprotective, antiulcer, hypolipidemic, and anti-atherosclerotic activities, as well as others (17, 20, 24, 25). Furthermore, it has been reported that UA attenuated the expression of inducible NO synthase (iNOS) and cyclooxygenase-2 expression via NF-κB repression in LPS- or IFN-γ-activated mouse macrophages (26). In contrast, You et al. (27) recently reported that UA induced NO and TNF-α production via NF-κB activation in resting RAW264.7 mouse macrophages. Along a similar line, we recently reported that UA promoted the release of MIF via ERK2 activation in the same cell line (28). Those findings imply that the effects of UA on NF-κB activities are dependent on the biological status of the target macrophages. This background led us to the present investigation of the potential proinflammatory effects of UA in pMφ. Our results indicate for the first time that aggregated UA is recognized by CD36 for generating ROS, thereby activating p38 MAPK, ERK1/2, and caspase-1 and releasing IL-1β protein in pMφ.

RAW 264.7 macrophages were obtained from American Type Culture Collection. Male CD36-deficient (C57BL/6J) mice were provided by Dr. M. W. Freeman (Harvard Medical School, Boston, MA). Specific pathogen-free 5-wk-old female ICR mice and 9-wk-old male C57BL/6J mice were purchased from Japan SLC. On arrival, the mice were randomized and transferred to plastic cages containing sawdust bedding (five mice per cage), which was changed every third day. They were given MF rodent chow (Oriental Yeast) and fresh tap water ad libitum, which was freshly changed twice a week, and handled according to the Guidelines of the Regulation of Animals, as provided by the Experimentation Committee of Kyoto University. The mice were maintained in a controlled environment of 24 ± 2°C with a relative humidity of 60 ± 5% and a 12-h light/dark cycle (lights on from 06:00 to 18:00). All mice were quarantined for 1 wk before starting the experiments.

DMEM, Opti-MEM, FBS, and TRIzol were purchased from Invitrogen Life Technologies. UA was obtained from Funakoshi. PD98059, SB203580, SP600125, SB202474 (negative control) inhibitors, diphenyleneiodonium (DPI) chloride, and YVAD-CHO came from Calbiochem. Glibenclamide, human leukocyte myeloperoxidase (MPO), o-dianisidine dihydrochloride, and hexadecyltrimethylammonium were obtained from Sigma-Aldrich. Abs were purchased from the following sources: rat anti-CXCL16 Ab was from TECHNE; goat anti-scavenger receptor (SR) class-A (SR-A), rabbit anti-CD36, rabbit anti-CD68, goat anti-IL-1β, and goat anti-β-actin Abs from Santa Cruz Biotechnology; rabbit anti-phospho-ERK1/2, rabbit anti-ERK1/2, rabbit anti-phospho-p38, rabbit anti-p38, rabbit anti-active JNK1/2, rabbit anti-JNK1/2, rabbit anti-phospho-MAPK/ERK kinase (MEK)1/2, rabbit anti-MEK1/2, rabbit anti-phospho-MAPK kinase (MKK) 3/6, rabbit anti-MKK3, rabbit anti-phospho-Raf-1, rabbit anti-Raf-1, and anti-rabbit Ab HRP-linked IgG Abs from Cell Signaling Technology; and anti-goat IgG from DakoCytomation. Oligonucleotide primers were synthesized by Proligo. Mouse IL-1β ELISA and caspase-1 colorimetric assay kits were purchased from R&D Systems. Mouse IL-6 and mouse TNF-α ELISA kits were purchased from Endgen. A rat/mouse MIF immunoassay kit came from Sapporo ImmunoDiagnostic Laboratory. All other chemicals were purchased from Wako Pure Chemicals unless specified otherwise.

RAW 264.7 macrophages were grown in DMEM supplemented with 10% FBS, l-glutamine (330 μg/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C under a humidified atmosphere of 95% air and 5% CO₂. pMφ monolayers were prepared as described previously (29), with some modifications. Briefly, nontreated 6-wk-old female ICR mice were killed by cervical dislocation, and 10 ml of ice-cold DMEM containing 10% FBS, l-glutamine (330 μg/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml) was injected i.p. Medium containing peritoneal exudates cells (PEC) was recollected and kept on ice. The suspended cells thus obtained were centrifuged at 800 × g for 5 min and resuspended with DMEM. Peritoneal exudate cells (5 × 10⁶ cells/ml/well) were then seeded onto culture plates and allowed to adhere for 24 h at 37°C under a humidified atmosphere of 95% air and 5% CO₂. After washing with PBS twice, nonadherent cells were removed, and the remaining monolayers were designated as pMφ. Cell viability was ≥90% in all experiments, unless specified otherwise.

At the end of each experiment, mice were killed by cervical dislocation, and the large intestines without the cecum were removed. After washing in ice-cold PBS, the specimens were placed on filter papers and opened with surgical scissors to remove their contents. The colonic mucosa was scraped off using a razor while the specimen was on ice, then frozen in liquid nitrogen until use, according to a method previously reported by Perdue et al. (30), with some modifications.

Total protein concentrations in the pMφ and tissue supernatants were determined using a DC protein assay (Bio-Rad) according to the protocol of the manufacturer (dilution factor = 50), with gammaglobulin used as the standard.

pMφ were seeded onto a 96-well plate at a density of 1 × 10⁶ cells/200 μl in 200 μl of DMEM, which included 10% FBS, l-glutamine (330 μg/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml). The cells were cultured at 37°C under a humidified atmosphere of 95% air and 5% CO₂. After 24 h, the cells were washed twice with PBS, and a serum-free medium was added. The cells were then treated with UA (4 or 20 μM), which was dissolved in DMSO (0.1% v/v, as a final concentration). Control cells were treated only with 0.1% (v/v) DMSO and showed no significant effect on the assay systems (data not shown). After incubation for 0, 1, 3, 6, 12, and 24 h, the supernatants, 5 μl for MIF and 50 μl for TNF-α, IL-1β, and IL-6, were used for ELISA and examined according to the protocol of the appropriate kit. Alternatively, pMφ were pretreated with each specific inhibitor (see Fig. 4,B) or Ab (see Fig. 5 B) 30 min before UA treatment.

To determine the expression of phospho-ERK1/2 (Thr²⁰²/Tyr²⁰⁴), ERK, phospho-p38 MAPK (Thr¹⁸⁰/Tyr¹⁸²), p38 MAPK, phospho-JNK1/2 (Thr¹⁸³/Tyr¹⁸⁵), JNK1/2, phospho-MEK1/2 (Ser²¹⁷/²²¹), MEK1/2, phospho-Raf (Ser²⁵⁹), Raf, phospho-MKK3/6 (Ser¹⁸⁹/²⁰⁷), MKK3/6, proIL-1β, SR-A, CD36, CD68, and CXCL16 in pMφ, the total cell lysate was subjected to Western blotting. β-Actin served as the internal control. Cells (5.0 × 10⁶) were treated with lysis buffer (protease inhibitor mixture, phosphatase inhibitor mixture (TaKaRa Bio), 10 mM Tris (pH 7.4), 1% SDS, and 1 mM sodium vanadate (V)), and the lysates were boiled for 5 min. Denatured proteins (40 μg) were separated using SDS-PAGE on a 10% polyacrylamide gel and then transferred onto Immobilon-P membranes (Millipore). After blocking overnight at 4°C in Block Ace (Dainippon Pharmaceutical), the membranes were first incubated with each Ab at dilutions of 1/1000. The second incubation was performed with HRP-conjugated secondary anti-rabbit IgG or anti-goat IgG Ab (1/1000 dilution each). The blots were developed using an ECL Advance Western blotting detection reagent (Amersham Biosciences). The intensity of each band was analyzed using NIH Image.

Steady-state mRNA levels of CXCL16, SR-A type I, CD36, CD 68, and IL-1β were detected by RT-PCR. pMφ were cultured at a density of 5 × 10⁶ cells/2 ml for 24 h. Total cellular RNA was extracted from the cells using TRIzol reagent, then precipitated with isopropanol, washed with 70% (v/v) ethanol, and treated with DNase I (Invitrogen Life Technologies) to remove genomic DNA. A cyclophilin transcript served as the internal control. The primer sequences, PCR product sizes, and PCR conditions are listed in Table I. cDNA was synthesized using 1 μg of total RNA and an RNA PCR kit (Takara). Amplified cDNA was electrophoresed on 2% agarose gels and stained with SYBR Gold (Molecular Probes). Image analysis was performed using NIH Image. No PCR saturation was confirmed by titration of each cDNA amount (data not shown).

pMφ were cultured at a density of 1 × 10⁶ cells/200 μl for 24 h on one-chamber slides (GLASS/PS, 9 × 9; IWAKI). After washing, the cells were suspended in 200 μl of PBS and treated with 10 μM of 2,7′-dichlorofluorescein diacetate (DCFH-DA; Molecular Probes) for 5 min. Thereafter, the medium was discarded and the pMφ were washed twice with PBS, followed by exposure to DMSO or UA (final 4 or 20 μM) for 5 min. Negative control cells were treated only with DMSO and without DCFH-DA. UA-induced intracellular ROS generation was detected using a fluorescence microscope (Olympus).

Analysis of the interaction between UA and CD36 on the cell surface of macrophages was performed using the SPR biosensor SPR670 (Moritex). Mouse macrophage RAW264.7 cells (5 × 10⁵ cells/ml) and pMφ (1.5 × 10⁶ cells/ml) were immobilized on sensor chips, which had been pretreated with nonspecific IgG and anti-CD36 Ab (2.5 μg/ml each) in serum-free DMEM for 30 min. Following pretreatment, the medium was discarded and the chips were equilibrated in SPR running buffer, and 0.1% DMSO (v/v) in PBS (pH 7.4, flow rate = 30 μl/min). UA was diluted at 0 or 20 μM in SPR running buffer in 60-μl injection volumes and run on the chip at a flow rate of 30 μl/min. Binding was measured at 25°C for 2 min, followed by dissociation. The value of the angle was deducted from the binding signal of 0 μM UA and used as the binding strength.

UA (4 μM) was dissolved in DMSO (0.05–1%, v/v) and added to 50 ml of serum-free DMEM in a 50-ml centrifuge tube. After centrifugation at 2,500 × g for 5 min, UA was extracted separately from the supernatant and pellet with chloroform, with a recovery rate of ≥85% in each experiment. Each extract was concentrated in vacuo and dissolved in 200 μl of chloroform. The amount of UA was quantified by HPLC analysis on a YMC-Pack ODS-AQ column (150 × 4.6 mm inside diameter; YMC), which was eluted with 10% methanol in water at a flow rate of 1.0 ml/min. A standard curve was made using 0–8 μg of UA, with the peak area based on absorption of 210 nm. The amounts of UA derivatives (4 μM, 0.1% DMSO v/v) were analyzed under the same experimental conditions. Data are expressed by the distribution rates, which were calculated from the amounts of each fraction.

Corn oil alone (n = 8) or UA suspended in 200 μl of corn oil was administrated by i.p. injection to specific pathogen-free 5-wk-old female ICR mice daily at a dose of 50 (n = 6), 100 (n = 6), or 200 (n = 9) mg/kg body weight for 8 days. Twenty-four hours after the final administration, pMφ monolayers were separately prepared as described above and seeded onto a 96-well plate at a density of 1 × 10⁶ cells/200 μl, followed by incubation for 24 h at 37°C under a humidified atmosphere of 95% air and 5% CO₂. After the cells were washed twice with PBS, serum-free medium (200 μl) was added. The cells were then incubated for another 24 h, and the supernatant (50 μl) was used for measuring IL-1β with ELISA, as described above. After preparation of pMφ, the mucosa layer removed from the colon was homogenized (10 mg/300 μl) in ice-cold PBS using a homogenizer (UP 50H; Hielscher). Homogenates were frozen in liquid nitrogen and thawed using a sonicator (EYELA). These procedures were repeated three times, then the homogenate was centrifuged at 20,000 × g for 30 min at 4°C to obtain a supernatant. Each supernatant (50 μl) was subjected to ELISA, and the amounts of IL-1β were measured according to the protocol of the kit.

MPO activity was measured as an index of inflammatory cell infiltration in the colonic mucosa using a previously reported method (31), with some modifications. The mucosal layer (∼10 mg) was homogenized in 500 μl of 50 mM potassium phosphate buffer containing 0.5% hexadecyltrimethylammonium bromide (pH 6.0) using a homogenizer. Homogenates were frozen in liquid nitrogen and thawed using a sonicator (EYELA). The freeze-thaw cycle was repeated three times, then each homogenate was centrifuged at 20,000 × g for 30 min at 4°C to obtain a supernatant, which was used to measure MPO activity. MPO in the sample was activated by 0.0005% H₂O₂ in potassium phosphate buffer solution containing 0.5 mM o-dianisidine dihydrochloride (pH 6.0). The change in absorbance at 460 nm was measured using a spectrophotometer (Smart Spec; Bio-Rad) and converted to MPO activity using the standard curve for human leukocyte MPO. MPO activity was normalized further to the total protein content of the supernatant, as measured with a DC protein assay. Activity is expressed as units of MPO activity per milligram of protein.

Each experiment was performed at least three times, and the data are shown as the mean ± SD where applicable. Statistically significant differences between groups in each assay were determined using Student’s t test (two-sided).

To investigate the effects of UA on proinflammatory cytokine release into medium, pMφ were treated with UA (0, 4, and 20 μM) for 0–24 h and examined with ELISA. As shown in Fig. 1, the levels of IL-1β, IL-6, and MIF protein in pMφ treated with UA (4 and 20 μM) were increased in concentration- and time-dependent manners in the medium (14- to ∼44-fold for IL-1β; 2.7- to ∼9.8-fold for IL-6; 7.4- to ∼28-fold for MIF from 6 to 24 h), as compared with the vehicle-treated cells. In contrast, TNF-α was not detectable (<50 pg/mg) under any of the experimental conditions (data not shown). Although viability of the cells treated with UA (20 μM) for 24 h was decreased by 60%, that of the others was maintained at ≥90% in each experiment (data not shown).

Subsequently, we explored the molecular mechanisms underlying IL-1β production by UA. To determine whether UA activates MAPK pathways, which are known to regulate the induction and production of proinflammatory cytokines, pMφ were treated with UA (4 μM) for 0–30 min, and both the inactive and activated forms of Raf-1, MEK1/2, ERK1/2, MKK3/6, p38 MAPK, and JNK1/2 were analyzed by Western blotting using Abs specific for each target protein. UA strikingly induced both Raf-1 and MEK1/2 activation within 2 min and that of ERK1/2 within 5 min, as compared with the nontreated cells (Fig. 2). Similarly, both MKK3/6 and p38 MAPK were activated within 2 min, whereas phosphorylation of JNK1/2 was not observed. The expression levels of the inactive forms of each protein kinase remained constant.

The production of a number of proinflammatory cytokines is regulated by transcription, translation, and posttranslation mechanisms. Caspase-1 is the rate-limiting enzyme responsible for the conversion of pro-IL-1β to its active form. The effects of UA on the levels of IL-1β mRNA, proIL-1β protein, and caspase-1 activation were examined using RT-PCR, Western blotting, and ELISA, respectively. As shown in Fig. 3, A and B, IL-1β mRNA and proIL-1β protein were detected in a constitutive manner at low levels in nontreated pMφ. Those treated with 4 μM UA for 3 and 6 h were markedly up-regulated, whereas, intriguingly, the levels diminished after 12 h. Furthermore, to determine whether the ERK1/2 and p38 pathways are associated with those up-regulations, a pharmacological approach was used using each kinase-specific inhibitor. Pretreatment with 20 μM PD98059 (a MEK1/2 inhibitor) and SB203580 (a p38 MAPK inhibitor) for 30 min abolished the UA-induced increase in expression of IL-1β mRNA and pro-IL-1β protein (Fig. 3, A and B). In addition, PD98059 (20 μM), SB203580 (20 μM), glibenclamide (ABC transporter inhibitor, 20 μM), and Ac-YVAD-CHO (caspase-1 inhibitor, 20 μM), but not SB202474 (MAPK negative control, 20 μM), markedly reduced the level of UA-induced IL-1β secretion (Fig. 3 C).

Oxidative stress has been demonstrated to induce the activation of MAPKs in many cell types; therefore, we investigated ROS generation in pMφ treated with UA (4 and 20 μM) for 5 min using DCFH-DA, a fluorescent probe. The rate of ROS-positive cells was increased in a concentration-dependent manner (31 and 43% at 4 and 20 μM, respectively) as compared with the vehicle-treated cells (Fig. 4,A). Subsequently, we pretreated pMφ with the vehicle, N-acetyl-l-cysteine (NAC (an antioxidant), 1 mM), DPI (NADPH oxidase (NOX) inhibitor, 10 μM), or EDTA (Ca²⁺ chelator, 5 mM) for 30 min, followed by exposure to UA (4 μM) for 6 h. NAC, DPI, and EDTA markedly reduced UA-induced IL-1β secretions by 52, 69, and 80%, respectively, together with a dramatic suppression of ERK1/2 and p38 MAPK pathways (Fig. 4, B and C).

SRs, which mediate the endocytic uptake of modified forms of low-density lipoprotein, apoptotic cells, glycated proteins, and bacteria, are expressed on monocytes/macrophages, platelets, and certain microvascular endothelium (32, 33). Recently, we found that DSS induces IL-1β production though a SR-mediated mechanism (K. H. Kwon, A. Murakami, and H. Ohigashi, submitted for publication). In this study, we examined the status of the mRNA and protein expressions of CXCL16, SR-A, CD36, and CD68 in pMφ with or without UA treatment. As shown in Fig. 5,A, the expression of each was detected in a constitutive manner, and the levels were not changed by treatment with UA for 6 h. Next, we examined the effects of neutralizing Abs for SRs on UA-induced IL-1β production. When pMφ were pretreated with an anti-CD36 Ab (2.5 μg/ml) for 30 min, IL-1β production was significantly reduced by 56% as compared with the PBS-treated control, whereas nonspecific IgG and other SR Abs were inactive (Fig. 5,B). In addition, when pMφ from CD36-deficient mice were treated with UA (4 or 20 μM) for 12 h, and the levels of IL-1β, IL-6, and MIF released were markedly lower than from those of wild-type mice (Fig. 5,C). In contrast, CD36 deficiency did not have an effect on LPS-induced cytokine production. Furthermore, SPR analysis showed that UA was bound to the cell surface of RAW264.7 cells and pMφ (Fig. 5 D, nonspecific IgG), which was suppressed by treatment with the Ab for CD36, suggesting that UA interacts with CD36 located on the cell surface of macrophages.

The above results led us to hypothesize that UA aggregates in culture medium and is then recognized by CD36, leading to the release of IL-1β protein, because SRs largely recognize macromolecules. To test our hypothesis, we prepared serum-free DMEM in which the amounts of aggregated UA were changed by increasing the concentration of DMSO (0.05–1%). After centrifugation, UA (4 μM) distribution throughout the supernatant and pellet was quantified by HPLC analysis (Fig. 6, upper panel). The amounts of aggregated UA were significantly decreased (47.0→17.8%) as DMSO concentration increased (0.05→1.0%). Interestingly, there was a marked decrease in UA-induced IL-1β production when the pMφ were exposed to lower levels of aggregated UA (0.5 and 1.0% DMSO; Fig. 6, lower panel), suggesting that aggregated UA is in an active form for IL-1β production.

It is important to understand whether IL-1β inducibility is specific for UA in light of the structural diversity of natural triterpenoids. To investigate this issue, we selected six natural and synthetic triterpenoids (Fig. 7,A) and investigated their properties for aggregation and IL-1β production in medium using HPLC analysis and ELISA, respectively. As shown in the upper panel of Fig. 7,B, five UA derivatives and glycyrrhetinic acid exhibited distribution patterns in the supernatant and pellets similar to UA. However, surprisingly, they showed no notable potential for IL-1β production (Fig. 7 B, lower panel).

Finally, we investigated whether UA is able to stimulate pMφ (Fig. 8,A) and colonic mucosa (Fig. 8, B and C) when injected i.p. into female ICR mice. As shown in Fig. 8,A, a 24-h incubation of pMφ from mice administrated UA (50, 100, or 200 mg/kg) once a day for 8 days led to a marked dose-dependent increase in IL-1β production (1.6-, 2.3-, and 3.3-fold, respectively) as compared with that of pMφ from vehicle-treated mice. In parallel, the levels of IL-1β protein and MPO activity in colonic mucosa from mice administrated UA at 200 mg/kg were significantly increased by 3.0- and 2.0-fold, respectively, as compared with that of mucosa from vehicle-treated mice (Fig. 8, B and C).

In the present experiments, we found that UA markedly amplified ABC transporter-mediated IL-1β secretion from murine pMφ at transcriptional, translational, and posttranslational levels, presumably through binding to CD36, ROS generation, the resultant activation of both the p38 MAPK and ERK1/2 pathways, and caspase-1 activation (Fig. 9). The action mechanisms underlying UA-induced IL-6 and MIF protein release may be similar to that for IL-1β release (Fig. 1) because IL-6 was previously shown to be strongly induced by ROS generation and the p38 MAPK, ERK1/2, and NF-κB-signaling pathways (34, 35), and MIF production was induced by ROS-activated ERK1/2 (36).

SRs, including SR-A, CD36, and CD68, have been reported to recognize negative-charged, high-m.w. substances, such as oxidized low-density lipoprotein (oxLDL), DSS, and amyloid β (33, 37, 38). Nishimura et al. (39) showed that oxLDL binding to SRs may generate intracellular ROS production through NOX activation, thereby activating NF-κB. CD36, originally identified as glycoprotein IV on platelets, is an integral membrane protein that has multiple ligands, including oxLDL, apoptotic cells, and long-chain fatty acids (40), and is expressed on monocytes/macrophages (41), platelets (42), and certain types of microvascular endothelia (43). In this study, the levels of CD36 mRNA and protein, both of which were expressed in a constitutive manner, did not change following UA treatment (Fig. 5 A), while a previous study found that ligand-binding induced oxidative stress up-regulated SR gene expression, thereby promoting the uptake of oxLDL (44).

SPR biosensors have been increasingly used for real-time analysis of the binding between solubilized molecules and molecules immobilized on the surface of a biosensor chip without any labeling based on changes in the refractive index of the biospecific surface (45). Our analysis with the biosensor indicated that UA bound to pMφ and that binding was inhibited by treatment with the anti-CD36 Ab (Fig. 5,D). Furthermore, the production of IL-1β was suppressed by treatment with that Ab (Fig. 5,B) and decreased in pMφ of CD36-deficient mice, as compared with wild-type mice (Fig. 5,C). To the best of our knowledge, this is the first report showing that CD36 is one of the membrane receptors for a triterpenoid, while it was previously reported that UA binds to the hydrophobic region of the dimeric interface of TGF-β1 (46). However, it remains to be determined whether UA also binds with other SRs or proteins because the anti-CD36 Ab and CD36 deficiency did not abolish IL-1β protein release and UA binding (Fig. 5, B–D). Nevertheless, CD36 also partially mediates UA-, but not LPS-, induced IL-6 and MIF production (Fig. 5 C).

Because UA is a low-m.w. substance, it is unlikely that it is recognized by SRs, as they generally recognize anionic high-m.w. substances, as noted above. However, based on the hydrophobicity of this compound, it is reasonable to assume that some portions of UA aggregates present in culture medium are recognized by CD36. In fact, 18–33% of the triterpenoids, at a concentration of 4 μM, were revealed to be aggregated in culture medium containing 0.1% DMSO (Fig. 7), which has been widely used as a vehicle in a number of reports. Interestingly, there was a positive correlation between the aggregated UA concentration and IL-1β production (Fig. 6). Furthermore, the ability for inducing IL-1β was seen only with UA and not the other six triterpenoids, all of which demonstrated medium solubility similar to that of UA (Fig. 7). At present, there are no reasonable explanations for these puzzling outcomes; however, we speculate that aggregated UA, but not other triterpenoids, may have a particular structure in which its anionic moiety may be revealed for CD36 recognition. We hope to address this issue in the near future. Our results with the SPR biosensor demonstrated that UA binds to CD36 on not only pMφ but also RAW264.7 macrophages (Fig. 5 D). Recently, we reported that UA promoted MIF release via ERK2 activation in nonstimulated RAW264.7 cells (28), whereas You et al. (27) presented similar findings and noted that UA induced iNOS and TNF-α expression via NF-κB activation in the same cell lines. Because UA binds to RAW264.7 cells, in which CD36 is expressed in a constitutive manner (data not shown), the effects of UA shown in our experiments with RAW264.7 cells may also be mediated by binding to this SR.

The predominant sources of ROS in sites of inflammation are phagocytes, including neutrophils, monocytes, and macrophages, with NOX responsible for O₂⁻ generation (47). That enzyme is dormant in resident cells but becomes activated to generate ROS upon exposure to bacteria, chemical stimuli, or calcium influx though a receptor-regulated channel (48). The present findings demonstrated that UA induced ROS generation for activating ERK1/2 and p38 MAPK and the resultant release of IL-1β (Figs. 2 and 4). Of note, blockade of UA-induced ROS by an NOX inhibitor, NAC (antioxidant), and EDTA (calcium chelator) resulted in a decrease in ERK1/2 and p38 MAPK activities along with IL-1β production (Fig. 4). Conversely, Schweyer et al. (49) provided experimental evidence that both MEK1/2 and ERK1/2 activation was mediated by ROS, and Fubini et al. (50) showed that silica-induced generation of ROS induced ERK1/2 activation and increased the expression of IL-1β in cell culture models.

MAPKs control a number of cellular events, including differentiation, proliferation, and death (51). In this study, we used specific inhibitors to identify which MAPKs are involved in IL-1β release. Our data obtained with Western blotting (Fig. 2) and pharmacological blockades (Fig. 3) strongly suggest that UA induces secretion of IL-1β via the ERK1/2 and p38 MAPK pathways but not that of JNK1/2. In support of these findings, both the ERK1/2 and p38 MAPK pathways have been reported necessary for optimal cytokine gene expression in LPS-stimulated monocytes and macrophages (52). Similarly, Hsu et al. (53) reported that LPS-induced ROS generation activated the ERK1/2 and p38 MAPK pathways and also regulated the release of IL-1β in mouse macrophages. The rate of inhibition of IL-1β production by the specific inhibitors of MEK1/2 and p38 MAPK in the present study ranged from 43 to 67% (Fig. 3,C), whereas the expressions of IL-1β mRNA and protein were abolished by these inhibitors (Fig. 3 A), suggesting that MAPKs do not contribute to the posttranslational mechanism. Transcriptional factors, such as NF-κB, activator protein-1, cAMP-response element-binding protein, and NF-IL6, are known to be involved in the transcriptional regulation of IL-1β expression (54, 55, 56). Previous reports have also shown that MAPK-mediated signal pathways contribute, in part, to the transcriptional activities of those factors (57, 58, 59, 60). Thus, the above mentioned transcription factors may be crucial for UA-induced IL-1β expression.

Some cytokines such as TNF-α and IL-6 have a secretory signal peptide, which directs these proteins into the classical endoplasmic reticulum-to-Golgi secretory pathway for rapid secretion after synthesis (8, 61, 62). In contrast, IL-1β is a unique cytokine because it is not secreted via the classical exocytic pathway (63), and the lack of a secretory signal peptide hampers its targeting of the endoplasmic reticulum. ABC transporters are known to transport leaderless secretory protein (64). In this study, pretreatment with glibenclamide, an ABC transporter inhibitor, blocked UA-induced IL-1β release. This result is consistent with previous reports by Zhou et al. (11) and Hamon et al. (12) who showed that the ABC transporter contributes to the secretion of IL-1β from macrophages (11, 12).

Increasing evidence suggests that certain phytochemicals have marked anti-inflammatory and cancer chemopreventive properties (65, 66). UA is one such compound that has attracted considerable interest, based on its remarkable biological functions (17, 23, 24). For example, our group previously showed that UA has an inhibitory effect on 12-O-tetradecanoylphorbol-13-acetate-induced skin tumor promotion in mice (20). In addition, the compound attenuated LPS- or IFN-γ-induced iNOS and cyclooxygenase-2 expression through NF-κB abrogation in macrophages (26). In contrast, as noted above, UA-induced MIF release via activation of ERK2 and iNOS and TNF-α expression via activation of NF-κB in resting macrophages (27, 28). More recently, we reported that topical application of UA significantly increased proinflammatory gene expression in mouse skin (67). Those findings along with the present results led us to hypothesize that the pro- and anti-inflammatory effects of UA are dependent on the biological status of cells and tissues.

In conclusion, our results showed that aggregated UA enhanced IL-1β secretion at transcriptional, translational, and posttranslational levels via its binding with CD36 for generating ROS, thereby activating the ERK1/2 and p38 MAPK pathways, and caspase-1, and releasing IL-1β protein via the ABC transporter in resident murine pMφ (Fig. 9). In addition, UA also dramatically induced IL-1β secretion and MPO activity in our in vivo model (Fig. 8). Although UA has long been known as an effective anti-inflammatory and anticarcinogenic agent in vivo by its abilities to counteract endogenous and exogenous stimuli, further extensive evaluations regarding the effects of UA on nontreated tissues are necessary to determine the risks and benefits of this triterpenoid.

The authors have no financial conflict of interest.