Mechanistic differences in the inhibition of NF-κB by turmeric and its curcuminoid constituents

Abstract

Turmeric extract, a mixture of curcumin and its demethoxy (DMC) and bisdemethoxy (BDMC) isomers, is used as an anti-inflammatory preparation in traditional Asian medicine. Curcumin is considered the major bioactive in turmeric but less is known about the relative anti-inflammatory potency and mechanism of the other components, their mixture, or the reduced in vivo metabolites. We quantified inhibition of the NF-κB pathway in cells, adduction to a peptide mimicking IκB kinase β, and the role of cellular glutathione as a scavenger of electrophilic curcuminoid oxidation products, suggested to be the active metabolites. Turmeric extract (IC₅₀ 14.5±2.9 μM), DMC (IC₅₀ 12.1±7.2 μM), and BDMC (IC₅₀ 8.3±1.6 μM), but not reduced curcumin, inhibited NF-κB similar to curcumin (IC₅₀ 18.2±3.9 μM). Peptide adduction were formed with turmeric and DMC but not with BDMC, and this correlated with their oxidative degradation. Inhibition of glutathione biosynthesis enhanced activity of DMC but not BDMC in the cellular assay. These findings suggest that NF-κB inhibition by curcumin and DMC involves their oxidation to reactive electrophiles whereas BDMC does not require oxidation. Since it has not been established whether curcumin undergoes oxidative transformation in vivo, oxidation-independent BDMC may be a promising alternative to test in clinical trials.

Graphical Abstract

Introduction

Turmeric, the ground rhizome of Curcuma longa L., is a widely consumed spice in Asian cuisine. Turmeric is also used as an anti-inflammatory and anti-microbial remedy in traditional Asian medicine and widely consumed as a dietary supplement ^{1, 2}. The possible health-related effects associated with consumption of turmeric and its major bioactive, curcumin, have spurred intense interest by the biomedical and natural product research communities and culminating in a large number of clinical trials that have been initiated ^3–7.

Commercially available turmeric extracts consist of the three curcuminoids, namely curcumin, demethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC), present in a ratio of about 80:15:5 ². Overwhelming evidence suggests that curcumin is the main bioactive component of turmeric although some studies imply demethoxycurcumin as equally or possibly more potent than curcumin ^8–13. There is considerably less information on biologic effects of the major in vivo metabolites, the conjugates curcumin-glucuronide and curcumin-sulfate ^14–20 or the reduced metabolites, di-, tetra-, hexa-, and octahydrocurcumin ^21–24 although reduced and conjugated metabolites appear overall less active than curcumin.

Not all studies with curcumin describe precisely the curcumin preparation used, resulting in uncertainty whether the term “curcumin” refers to the chemically pure compound or to turmeric extract with a mixture of the three curcuminoids, or yet another preparation, for example, containing additional components to enhance bioavailability. Proper identification of the preparation used seems an important distinction since individual curcuminoids as well as their mixtures may have different biologic effects or potency. Therefore, we set out to compare the anti-inflammatory activity of chemically synthesized pure curcumin with that of turmeric extract (the natural mixture of curcumin, DMC, and BDMC), chemically pure DMC and BDMC, as well as reduced curcumin as major in vivo metabolites (Fig. 1).

Fig. 1.

Chemical structures of curcuminoids. Curcumin, demethoxy- (DMC), bisdemethoxy-curcumin (BDMC) were used as pure compounds or as the natural mixture present in turmeric extract. Reduced curcumin consisted of a mixture of tetra-, hexa-, and octahydrocurcumin.

Anti-inflammatory activity was assessed based on inhibition of IKKβ-dependent NF-κB mediated gene transcription and receptor activator of NF-κB (RANK)-mediated osteoclast formation in RAW264.7 cells ^25–27, changes in phosphorylation of NF-κB pathway proteins, and covalent binding to a synthetic peptide that represents the activation domain of IKKβ ²⁸, the activating kinase upstream of NF-κB ²⁹. The role of glutathione was investigated as well as the ability of the compounds to bind to cellular protein. In addition, we determined the propensity of the preparations to undergo oxidative transformation since oxidative activation has been suggested as a molecular basis of the anti-inflammatory activity of curcumin ^{6, 30}.

Materials and Methods

Chemicals

Turmeric extract (C1386) was obtained from Sigma-Aldrich. Curcumin, demethoxycurcumin, and bisdemethoxycurcumin were synthesized according to the original method by Pabon with appropriate modifications ^{31, 32}. A mixture of the reduced products, tetra-, hexa-, and octahydrocurcumin, was obtained by catalytic hydrogenation of curcumin according to ref. 23. The mixture contained 70% tetra-, 20% hexa-, and 10% octahydrocurcumin and no dihydrocurcumin. Stock solutions of compounds were prepared in DMSO or methanol and stored at −20°C. The synthetic peptides ELDQGSLC(A)TSFVGTLQYLAPELLEQQK were synthesized by United Biosystems (Herndon, VA). BSO was from Acros Organics (New Jersey, NJ).

Cell culture

RAW264.7 and HeLa cells were obtained from ATCC and cultured in DMEM containing 4.5 g/L D-glucose and 10% (v/v) fetal bovine serum (FBS) in an atmosphere of 5% CO₂ at 37°C. RAW264.7 cells for testing NF-κB inhibition had been transfected with pNFkB-TA-MetLuc with TB vector, 4.3 kb (Clontech) for expression of secreted luciferase ³⁰. Stably transfected RAW 264.7 cells were supplemented with 1 mg/ml geneticin.

Inhibition of NF-κB

Stably transfected RAW264.7 cells were seeded in a 24-well plate at a concentration of 250,000 cells/well in 500 μl DMEM. For some experiments, cells were incubated with BSO (0.5 mM) overnight. Cells were washed twice with cold PBS and pre-treated with curcumin or analogs (0–100 μM) in DMEM for 45 min. The cells were then stimulated with LPS (100 ng/ml). After 4 h 50 μl of the medium was removed and luciferase activity was determined using the Ready-to-glow Secreted Luciferase Reporter assay (Takara).

Osteoclastogenesis assay

RAW264.7 cells were pre-treated with curcumin, DMC, or BDMC (all purchased from Chromadex) in media (DMEM + 10% FBS + 1% pen/strep) for 4 h at 37°C with 5% CO₂ followed by stimulation with receptor activator of NF-κB ligand (RANKL, 50 ng/mL final concentration, R&D #462-TEC) and further incubation at 37°C for 72 h. Tartrate-resistant acid phosphatase (TRAP) positive, multi-nucleated (n≥3) osteoclasts were visually quantified by microscopy.

Western blot analysis

HeLa cells were cultured in 6-well plates in 2 ml DMEM with 10% (v/v) FBS and treated with curcuminoids (0, 10, and 50 μM) for 1 h followed by stimulation with human TNFα (20 ng/ml) for 20 min. For analysis of IKKβ phosphorylation cells were stimulated for 5 min only. Cells were washed with PBS and lysed using 100 μl of cell lysis buffer containing 1 mM PMSF, protease and phosphatase inhibitor cocktails. Cell lysates (20 μg protein) were resolved on a 10% SDS-PAGE gel and transferred to nitrocellulose. Primary phospho protein antibodies were used as 1:500 to 1:1000 dilution and probed with IRDye secondary antibodies (1:20,000). Blots were stripped and re-probed for total protein. Band intensities were quantified using Image Studio Lite ver. 5.0 (LI-COR Biosciences). Primary antibodies used in this study were: NF-κB p65 (C22B4), phospho-NF-κB p65 (Ser536) (93H1), IκBα (44D4), phospho-IKKα/β (Ser176/180) (16A6) (Cell Signaling), IKKβ (551920) (Pharmingen). Secondary antibodies used were IRDye 680LT goat anti-rabbit (926–68024) and IRDye 800CW donkey anti-goat (926–32214) (LI-COR Biosciences).

Peptide binding

The peptides (10 μM) were incubated with curcuminoids (50 μM) at 37ºC in acetate buffer pH 7.5 for 1 h. Aliquots were withdrawn and directly injected on a Waters XSelect CSH 130 C18 column (3.5 μm; 2.1 × 50 mm) eluted with a gradient of acetonitrile/water starting from 15/85 (by vol.) to 65/35 (by vol.) containing 0.1 % formic acid within 10 min at a flow rate of 0.25 ml/min. The column was connected to a Waters Synapt G1 TOF instrument operated in positive ionization mode.

Protein binding

RAW 264.7 cells were treated with HBSS containing curcuminoids or N-ethylmaleimide (75 μM) for 2 h. Cells were lysed using 100 μl of cell lysis buffer containing a protease inhibitor cocktail (Roche) followed by centrifugation at 13,000 x g at 4°C for 10 min. For the detection of free thiol residues, 30 μg of protein was incubated with tetramethylrhodamine-5-(and-6) C2 maleimide (0.1 mM; AnaSpec) and SDS (0.1%) at 4°C in the dark for 30 min. Excess reagents were removed by acetone precipitation at −20°C overnight. Samples were centrifuged at 13,000 x g for 10 min, acetone was removed, and samples were left to air dry for 10 min. Urea (3 μL of 8 M) was added to the precipitate and incubated at 4°C for 1 h to solubilize the sample followed by addition of SDS-PAGE sample buffer containing β-mercaptoethanol. Samples were resolved on a 8% SDS-PAGE gel. Free thiol residue were visualized by scanning the gel using the Gel Doc EZ system (Bio-Rad) with setting for detection of rhodamine. The gel was then washed three times with water and stained with coomassie blue.

Statistical analysis

Overall differences were analyzed by one‐way (Fig. 2) or two-way (Fig. 6) ANOVA. If significant, pairwise analysis using Bonferroni tests were performed to determine differences between groups. Each data point represents the mean ± SD of 3 to 5 independent experiments. Statistical analyses were performed using GraphPad Prism 8.2.1.

Fig. 2.

Inhibition of NF-κB activity by curcuminoids. (A) Stably transfected RAW264.7 cells expressing luciferase under the control of an NF-κB response element were treated with different concentrations of curcuminoids or vehicle followed by activation with LPS. Luciferase activity is expressed relative to the maximum value induced by LPS. (B) Calculated IC₅₀ values +/− SD (n = 3) for the inhibition of NF-κB activity.

Fig. 6.

Effect of the glutathione biosynthesis inhibitor buthionine sulfoximine (BSO) on the inhibition of NF-κB activity by (A) DMC and (B) BDMC. Stably transfected RAW264.7 cells expressing luciferase under the control of an NF-κB response element were treated with BSO prior to different concentrations of DMC and BDMC or vehicle followed by activation with LPS. Luciferase activity is expressed relative to the maximum value induced by LPS. Data are from 3–5 independent experiments with the asterisk indicating statistical difference at p<0.05 between treatments using 2-way ANOVA.

Results and Discussion

Inhibition of NF-κB activity

Inhibition of NF-κB by curcuminoids was quantified as a decrease in LPS-induced luciferase activity expressed downstream of an NF-κB response element in stably transfected RAW264.7 cells ³⁰. Cells were treated with compounds (1–50 μM) prior to activation of NF-κB by LPS and quantification of luciferase activity. Chemically synthesized curcumin dose-dependently inhibited NF-κB with an IC₅₀ of 18 μM (Fig. 2), a value similar to what was previously determined using the same assay ³⁰. Commercial turmeric extract, consisting of a mixture of curcumin, DMC, and BDMC (≈80:15:5), had an IC₅₀ of 15 μM. Pure DMC (IC₅₀ = 12 μM) was less potent whereas BDMC (IC₅₀ = 8 μM) was more potent than curcumin, although none of the IC₅₀ values were statistically different from each other. A mixture of reduced curcuminoids had inhibitory activity too low to calculate an IC₅₀ value. The weak effect of the reduced metabolites was in accord with prior studies on NF-κB inhibition or TGFβ inhibition in vitro ^{21, 24} but in contrast to studies in vivo suggesting tetra- and octahydrocurcumin are more effective than curcumin in suppressing acute inflammation by targeting a TGFβ-activated NF-κB pathway in mice ³³. The reason for this discrepancy is unclear. There was overall no difference in the potency of turmeric extract versus curcumin although the slightly lower IC₅₀ for the extract may be reflective of the contribution of DMC and BDMC which both gave lower calculated IC₅₀ values than curcumin.

Effect on NF-κB pathway proteins

To assess the cellular targets of inhibition we tested how the different curcumin preparations affected activation of proteins of the NF-κB pathway using HeLa cells stimulated with TNFα. In accord with the functional assay, turmeric, curcumin, and DMC were equally potent in inhibiting phosphorylation of IKKβ and NF-κB p65 and in protecting the inhibitory protein IκBα from degradation (Fig. 3). Degradation of IκBα unmasks the nuclear localization sequence of NF-κB and enables the transcription factor to enter the nucleus. BDMC, in contrast, had only slight effects on phosphorylation of IKKβ and NF-κB p65 and the rescue of IκBα (Fig. 3). The weak effects on the NF-κB pathway proteins were unexpected considering the potent inhibition of NF-κB activity by BDMC (Fig. 2). These results confirmed prior studies showing that curcuminoids inhibit the NF-κB dependent inflammatory response at least in part by inhibition of IKKβ ^{34, 35}.

Fig. 3.

Western blot analysis of NF-κB pathway proteins. HeLa cells were treated with curcuminoids (10 or 50 μM) or vehicle followed by activation with TNF-α. Phosphorylation of IKKβ and p65 as well as IκBα were analyzed using total protein (IKKβ, p65) or actin as controls. The blot shown is representative of three independent experiments.

Inhibition of receptor activator of NF-κB-induced osteoclastogenesis

Given the ability of curcuminoids to inhibit NF-κB signaling, we next evaluated the effects of curcuminoids on RANKL-induced, NF-κB-dependent osteoclastogenesis ²⁰. Interaction of RANK ligand with receptor activator of NF-κB (RANK) activates a variety of downstream signaling pathways required for osteoclast development, including IKKβ-mediated NF-κB signaling ^25–27. RAW264.7 cells were pre-treated with curcuminoids followed by activation with RANKL and analysis by quantification of multi-nucleated, TRAP-positive osteoclasts (Fig. 4). Consistent with the luciferase assay results, BDMC was the most potent curcuminoid with an IC₅₀ of 2.3 μM while curcumin (IC₅₀ = 3.2 μM) and DMC (IC₅₀ = 3.1 μM) were slightly less potent.

Fig. 4.

Inhibition of osteoclastogenesis by curcumin, DMC, and BDMC. RAW264.7 cells were treated with curcumin, DMC, or BDMC followed by stimulation with RANKL and further incubation at 37°C for 72 h. Tartrate-resistant acid phosphatase (TRAP) positive, multi-nucleated (n≥3) osteoclasts were visually quantified by microscopy.

Peptide adduction

A possible mechanism for the inhibition of NF-κB activity by curcuminoids is covalent binding to Cys-179 of the upstream kinase IKKβ ⁶. Cys-179 serves as a sensor of cellular redox status and has been well characterized as a target of small molecule electrophiles that exert anti-inflammatory activity, including curcumin ^{34, 36, 37}. Inhibition of IKKβ via covalent adduction of small molecules has been shown directly, e.g., for the lipid-derived electrophiles 15deoxy-Δ^12,14-prostaglandin J₂ and 4-hydroxy-2E-nonenal ^{38, 39}. Binding of electrophiles to Cys-179 prevents phosphorylation of neighboring serines 177 and 181 that is required for kinase activity ²⁸. Reynaert and co-workers have shown that a 15-amino acid peptide, comprised of residues 173 to 187 and spanning the activation domain with Cys-179, mimics redox behavior of the entire IKKβ protein ²⁸. Based on these studies we have designed a 27-amino acid tryptic peptide (residues 172–198) as a tool to analyze the binding of curcumin and/or its oxidation products to this specific IKKβ site ^{30, 40}. In previous studies we found that autoxidation of curcumin in the presence of the peptide yielded three distinct adducts, representing an increase of the mass of the peptide by 122, 368, and 400 amu. The increase by 368 amu and 400 amu indicated adduction by curcumin and a dioxygenated curcumin metabolite, respectively, to the peptide ^{30, 41}. The increase by 122 amu was due to addition of a methoxyphenol ring (guaiacol) formed during oxidative cleavage ⁴¹.

The 27-amino acid peptide was incubated with turmeric extract as well as with pure DMC and BDMC under conditions that allow oxidative transformation to occur. Extracted ion chromatograms from LC-MS analyses of adducts formed with turmeric extract are shown in Fig. 5A. Incubation of the peptide with turmeric extract gave the +122, +368, and +400 adducts that had been observed using synthetic curcumin ³⁰. In addition to curcumin-derived adducts there was also a peptide representing adduction of DMC (+338 amu). No adduct was detected that was compatible with adduction of BDMC.

Fig. 5.

Peptide adduction of curcuminoids. Positive ion LC-MS analysis of a synthetic peptide representing amino acids 172–198 of IKKβ incubated with (A) turmeric extract or (B) DMC at 37°C for 1 h. (C-F) LC-MS spectra of peptide adducts in (B) eluting at the retention times indicated. (G) The synthetic peptide was incubated with BDMC and analyzed by LC-MS.

Incubation of the peptide with pure DMC gave +122, +338, and +370 adducts, the latter two indicating adduction of DMC and dioxygenated DMC, respectively (Fig. 5B-F). Formation of a guaiacol adduct (+122) but not a phenol adduct (+92) was compatible with reaction of the peptide cysteine at the quinoid ring of the spiroepoxide intermediate of oxidative transformation of DMC, followed by a fragmentation reaction ⁴¹. The mechanism of oxidative transformation of DMC predicts that the reactive quinoid ring is only formed at the guaiacol side but not the phenolic side of DMC ⁴². Incubation of the peptide with pure BDMC failed to give any discernable adducts when analyzing for the predicted increase of 308 amu (BDMC), 340 amu (dioxygenated BDMC) or 92 amu (phenol) (Fig. 5G). The inability of BDMC to bind to the peptide mimic of IKKβ was in accord with the weak inhibition of IKKβ phosphorylation (Fig. 3) but in contrast to the potent inhibition of NF-κB activity in the luciferase assay (Fig. 2). These findings suggested an inhibitory mechanism of BDMC that was distinct from curcumin and DMC.

The reduced curcumin mixture did not show formation of any adducts with the IKKβ peptide. This was compatible with absence of a cysteine-reactive enone moiety and the predicted inability of tetrahydrocurcumin and its further reduced congeners to undergo oxidative transformation to a reactive electrophile. Remaining potency of reduced curcumin to inhibit NF-κB is likely due to mechanisms that do not require a functional electrophile. No adducts were detected when a Cys179Ala mutant peptide was used in the incubations with turmeric extract, DMC, or BDMC (data not shown), underscoring the crucial role of the nucleophilic cysteine residue in adduct formation.

Role of glutathione

Glutathione (GSH) protects cells from oxidative and electrophilic stress. GSH reacts with both the overt enone (possibly in the form of a quinone methide intermediate) as well as the oxidation-dependent spiroepoxide electrophile of curcumin ^{41, 43}. Reaction of GSH with either electrophile decreases the amount of bioactive curcumin, and thus, the cellular levels of GSH modulate the potency of curcumin as an inhibitor of NF-κB activation. GSH levels can be enhanced by incubating cells with N-acetylcysteine (NAc) as a biosynthetic precursor or be depleted by treating with buthionine sulfoximine (BSO), an inhibitor of GSH biosynthesis ⁴⁴. Accordingly, treatment of RAW264.7 cells with NAc has been shown to result in decreased potency of curcumin while BSO treatment enhanced potency to inhibit NF-κB ³⁰. When RAW264.7 cells were pretreated with BSO only the potency of DMC (which forms an electrophile upon spontaneous oxidation) was increased, consistent with prior observations with curcumin, whereas no effect was observed for BDMC which does not readily oxidize and form a reactive electrophile (Fig. 6) ^{41, 42}. These results were consistent with the analysis of peptide binding by DMC and BDMC which showed that only DMC and its oxidation product adducted to the nucleophilic cysteine residue of the IKKβ peptide. Lack of an effect of GSH manipulation on the activity of BDMC supported an inhibitory mechanism for this curcuminoid that does not involve formation of an oxidation product.

Protein binding

The experiments with the IKKβ peptide described above as well as work using small molecule thiols ⁴¹ suggested that BDMC would not bind efficiently to protein cysteine residues. This pointed at a molecular mechanism of BDMC that is different from curcumin which involves oxidative activation and subsequent binding to protein, at least with respect to inhibition of NF-κB and human topoisomerase IIα ^{30, 45}.

In order to test for differences in protein binding between curcumin and the other curcuminoids we used a fluorescence-based assay that detects free cysteine residues in protein followed by resolution using SDS-PAGE ⁴⁶. Reactive cysteine residues are visualized by derivatization of harvested protein with NEM-rhodamine prior to SDS-PAGE. NEM is a cysteine-selective reactive moiety and the coupling with rhodamine allows detection of derivatized protein via fluorescence. If pretreatment with curcumin or curcuminoids has resulted in binding to free cysteine residues, the adducted cysteines are no longer available for derivatization with the NEM-rhodamine. The resulting decrease in fluorescence intensity is therefore an indication of covalent binding of curcumin to protein cysteines. Notably, this assay tests the chemically unmodified curcuminoids and does not rely on the use of, e.g., alkynyl-tagged curcumin probes for the detection of protein binding ^{30, 47–49}. As a positive control for maximum protein binding detectable by this method, cells were pretreated with NEM.

RAW264.7 cells treated with curcumin showed marked decrease of staining by NEM-rhodamine (Fig. 7). The decrease in fluorescence was similar in cells treated with turmeric extract and DMC but not with BDMC. This was consistent with results from peptide binding by the curcuminoids and their mixture. The mixture of reduced curcumin was unable to bind to protein, as expected from the absence of an electrophilic moiety (enone) or the inability to undergo hydrogen abstraction from the phenolic group to form a quinone methide electrophile.

Fig. 7.

Protein binding by curcuminoids in RAW264.7 cells. (A) RAW264.7 cells were treated with curcuminoids or N-ethylmaleimide (NEM) (75 μM each) for 2 h before harvesting and derivatization with NEM-rhodamine. Protein (30 μg) was separated using SDS-PAGE and derivatized protein was detected by fluorescence analysis. (B) The same gel was analyzed by staining with coomassie blue.

Our studies document two important findings on the inhibitory effect of curcuminoids on the NF-κB pathway. First, the natural mixture of the three curcuminoids as it is present in turmeric extract recapitulates the NF-κB inhibitory effects of its major component, curcumin. Curcumin behaves as a pro-drug during inhibition of NF-κB and topoisomerase and this likely pertains to other targets that feature redox-regulated cysteine residues. Second, oxidative activation, previously identified as a key mechanism of activating the pro-drug curcumin ^{30, 45}, also plays a role for inhibition of NF-κB by DMC but not by BDMC. Due to the absence of the ortho-methoxy groups that facilitate oxidation of the phenolic hydroxyl, BDMC is resistant to spontaneous oxidative transformation whereas DMC (containing one facilitating ortho-methoxy group) will oxidize spontaneously, albeit at a slower rate than curcumin ⁴². Although BDMC decreased phosphorylation of IKKβ, the target for covalent binding by curcumin and DMC, the inhibition did not involve covalent binding to Cys179 of the kinase. Since the extent of oxidative transformation of curcumin in animals and humans has not been established (suggesting it may be low), it appears that BDMC which does not require oxidation for inhibition of NF-κB may be a better candidate as an anti-inflammatory agent in vivo.

Abbreviations

BDMC

bisdemethoxycurcumin

BSO

buthionine sulfoximine

DMC

demethoxycurcumin

GSH

glutathione

IKKβ

inhibitor of nuclear factor κB kinase subunit β

LPS

lipopolysaccharide

NF-κB

nuclear factor κB

RANKL

receptor activator of nuclear factor κB ligand