Ginger Phenylpropanoids Inhibit IL-1β and Prostanoid Secretion and Disrupt Arachidonate-Phospholipid Remodeling by Targeting Phospholipases A₂

The rhizome of ginger (Zingiber officinale) is widely used in diet, but also in Asian traditional medicine, where it is employed therapeutically to treat mild forms of rheumatoid arthritis, fever, and upper respiratory tract infections (1). In conventional Western medicine, ginger preparations are administered clinically to treat emesis, nausea, and migraine headache (2, 3). Based on the widespread therapeutic use of ginger in Asia for the treatment of inflammatory diseases, several studies have been dedicated to the elucidation of potential anti-inflammatory mechanisms (4–7). At relatively high μM concentrations, ginger constituents such as gingerols, shogaols, and diarylheptanoids have been reported to weakly inhibit components of proinflammatory signal transduction pathways in vitro, such as NF-κB, protein kinase C, and MAPKs (8, 9). Moreover, ginger constituents have been shown to inhibit inducible NO synthase, cyclooxygenase (COX)-1/-2, and lipoxygenase in vitro (9–12). A significant inhibition of PGE₂ formation by ginger constituents has already been shown in vivo (13). Furthermore, different gingerols and shogaols at low μM concentrations have been shown to directly bind to and partially activate the cation channel transient receptor potential vanilloid 1 (14–16), the molecular clue for the spiciness of ginger, as well as the serotonin 5-HT_1A G protein-coupled receptor (17). Although the latter studies have elucidated feasible molecular mechanisms of action of the major biologically active ginger constituents, they cannot explain the range of anti-inflammatory effects observed in vivo. In particular, the reported antipyretic effects of ginger remain poorly understood.

In this study, we have profiled a standardized food-grade commercial ginger extract in a high content assay using freshly isolated human peripheral whole blood. Human blood was incubated with discrete stimuli targeting specific or nonspecific signal transduction pathways, and the resulting cytokine patterns were quantified. The assay was validated using blood from different donors and by profiling the effects of anti-inflammatory drugs with known targets. Using this profiling setup, we were able to identify calcium-independent phospholipase A₂ (iPLA₂) and cytosolic phospholipase A₂ (cPLA₂) enzymes as major biological targets of ginger phenylpropanoids.

PLA₂ enzymes play a central role in inflammatory processes (18, 19) and are classified into three main groups, as follows: secretory phospholipase A₂ (sPLA₂), cPLA₂, and iPLA₂. These distinctly different enzymes hydrolyze the sn-2 ester bond of phospholipids, thus liberating lysophospholipids and free fatty acids, which are precursors for a variety of different mediators of physiological and pathological processes (20, 21). PLA₂ isoenzymes are not only involved in the generation of free arachidonic acid (AA), which is metabolized by cyclo- and lipoxygenases to eicosanoids (22), but also in the incorporation of free AA into membranes and redistribution into specific compartments (remodeling) (23, 24). Besides regulatory functions in lipid catabolism/metabolism, PLA₂ enzymes are critically involved in signal transduction, phospholipid and cell membrane bilayer remodeling (25), store-operated cation channels and Ca²⁺ release-activated Ca²⁺ channels (26), Fas-induced apoptosis (27), cellular proliferation and differentiation (28, 29), glucose-induced insulin secretion (30), and, most important for this study, they play a central role in acute (18) and chronic inflammation (31).

Both iPLA₂ and cPLA₂ are crucial for IL-1β maturation and secretion (membrane translocation and fusion of storage vesicles) (32, 33). IL-1β is mainly secreted by monocytes/macrophages and dendritic cells, but also fibroblasts; is involved in a variety of inflammatory processes, such as costimulus for T cell activation; and is one of the key players in rheumatoid arthritis and inflammatory bowel disease (34, 35). IL-1β further induces COX-2 expression in endothelial cells and is involved in edema formation, where it synergizes with PGE₂ (36). IL-1β is also a major mediator of fever induction (34, 37). In this study, we provide evidence that the inhibition of i/cPLA₂ by phenylpropanoids is directly linked to the global inhibition of IL-1β secretion by ginger extracts in vitro. Moreover, ginger phenylpropanoids induce dramatic changes in arachidonate-phospholipid remodeling, due to their simultaneous inhibition of iPLA₂ and prostanoid synthesis. This apparently robust mechanism of action could provide a rational basis for the reported antipyretic and anti-inflammatory (i.e., immunomodulatory) effects of ginger preparations.

Ethanol 99%, KOH, NaCl, D-(+)-glucose, EDTA, LPS from Escherichia coli, ATP, PMA from Euphorbiaceae, BSA fraction V (BSA), and 2,4′-dibromoacetophenone (DBA) were obtained from Fluka Chemie AG (Buchs, Switzerland). DMSO, HEPES, MOPS, chloroform-d, monobromobimane, cylcosporin A, aspirin, diclofenac sodium salt, cetirizine, curcumin, parthenolide, fMLP, and dexamethasone 21-phosphate disodium salt were purchased from Sigma-Aldrich Chemie. All kinase inhibitors and methyl arachidonyl fluorophosphonate were purchased from Tocris Biosciences. Thapsigargin was purchased from Alexis Biochem. Water LiChrosolv, TLC silica aluminum sheets 60F₂₅₄, silica gel 60 (15–40 μm), CaCl₂, and acetonitrile were from Merck. Chloroform was purchased from Scharlau SA (Barcelona, Spain). The DTT high purity was obtained from GERBU Biochemicals (Gaiberg, Germany). Authentic standards of 6-shogaol and 6-, 8-, and 10-gingerol were purchased from ChromaDex (Santa Ana, CA). Vacutainer LH 170 I.U. and PrecisionGlide were from BD Biosciences (Plymouth, U.K.). KCl was from Hänseler AG (Herisau, Switzerland), and FCS was from Omnilab. The 1-hexadecyl-2-arachidonoylthio-2-deoxy-sn-glycero-3-phosphorylcholine (thio-PC) was purchased from Cayman Chemicals. The 8- and 10-shogaol were isolated by liquid chromatography and HPLC, as described previously (17).

The extract of Z. officinale (ginger Hot Flavor CO₂ extract, type 014.088, batch number 120507) was provided by FLAVEX (Rehlingen, Germany). It was produced by supercritical fluid extraction with carbon dioxide, according to Manninen et al. (38). According to the analysis certificate, it contains 38.7% 6-, 8-, 10-gingerol, 4.3% 6-, 8-, 10-shogaol, 0.7% zingerone, and 60 μl/g essential oil. Ginger essential oil was a commercial grade water vapor distillate from Farfalla. Extracts of organic Harpagophytum procumbens were obtained from Bioforce AG (Roggwil, Switzerland). Extract of Salix cortex was obtained from Bionorica SE (Neumarkt, Germany).

Human venous whole blood was drawn from healthy volunteers with the BD vacutainer system. Isolation of PBMCs was conducted in accordance with the guidelines of the World Medical Association’s Declaration of Helsinki and approved by ETH Zurich. PBMCs were isolated with OptiPrep (Axis-Shield, Oslo, Norway) density gradient flotation, and CD14⁺ monocytes were enriched by adhesion to culture flasks (FACS analysis was used to check for purity). Whole blood was immediately pipetted in a 96-well plate at 200 μl/well. Ethanolic extracts, pure compounds, or ethanol as vehicle control were added at different concentrations with a maximal concentration of 0.5% ethanol and incubated for 30 min at 37°C. Where indicated, stimuli (LPS [312 ng/ml]) or a combination of PMA [15 ng/ml] and thapsigargin [1 μM], anti-CD3 Ab, or anti-CD28 Ab [each 1 μg/ml]) were added, and blood was incubated for another 18 h at 37°C. For cytokine measurements, the plates were centrifuged, and an appropriate amount (usually 50 μl) of the supernatant serum was removed for immediate analysis. Isolation of PBMCs and T lymphocytes from peripheral whole blood was performed, as previously reported (39).

U937 cells (CRL-1593.2) (40) obtained from American Type Culture Collection and isolated human monocytes/macrophages were cultured in RPMI 1640 medium containing penicillin (100 U/ml), streptomycin (100 μg/ml), 2 mM glutamate (all from Life Technologies Invitrogen, Basel, Switzerland), amphotericin B (2 μg/ml; Sigma-Aldrich, Steinheim, Germany), and 10% FCS (heat inactivated for 30 min at 56°C) at 37°C, 5% CO₂, and humidified atmosphere. In the experiments measuring AA and eicosanoids, the cells were washed and incubated in serum-free medium (see below).

Cytokine quantification in whole blood was done with BD cytometric bead array (CBA) human inflammatory cytokine kits (BD kit 551811) for IL-8, IL-1β, IL-6, IL-10, TNF-α, IL-12p70, and the BD CBA human allergy mediators kit (BD kit 558022) for IL-3, IL-4, IL-5, IL-7, IL-10, and GM-CSF, or with isolated monocytes using the BD human IL-1β Flex Set (558279), according to the manufacturer’s manuals, measured with a FACScan flow cytometer equipped with an argon laser, and evaluated with the CBA software V1.4 (all by BD Biosciences).

The thromboxane B₂ (TXB₂) ELISA colorimetric kit was from Cayman Chemicals (10004023), limit of detection 35 pg/ml, and leukotriene B₄ (LTB₄) ELISA colorometric kit (EHLTB4) was from Thermo Scientific (Pierce Protein Research Products), limit of detection 20 pg/ml. ELISA measurements were carried out as specified by the manufacturer’s instructions. A total of 2 × 10⁶ U937 macrophages was stimulated with LPS (1 μg/ml) to measure TXB₂ and 10 × 10⁶ primary CD14⁺ monocytes/macrophages to measure LTB₄. Cells were preincubated with DMSO (vehicle) or test compounds for 30 min prior to stimulation with LPS (1 μg/ml) and for full LTB₄ release costimulation with fMLP (1 μM). LTB₄ could not be measured from stimulated U937 macrophages.

Gels for Western blots were run on denaturing NuPage Novex 4–12% Bis-Tris precast gels and corresponding MOPS buffer from Invitrogen, according to the manufacturer’s recommendations, using 10 μl samples dissolved in LDS buffer. Blotting was done on nitrocellulose membranes in a NuPage blotting chamber, according to the instruction manual.

Staining was done according to the ECL manual from PerkinElmer using TBST (pH 7.6 at used temperatures) and nitrocellulose membranes blocked with TBST plus 5% defatted milk powder. Ab labeling was done with TBST; 1% defatted milk powder and primary Ab (anti–IL-1β; Abnova MaxPab and Sigma-Aldrich) were used at 2 μl/ml; and secondary Ab (anti-mouse IgG [goat] HRP labeled by GE Healthcare) at 1/8 μl/ml. Chemiluminescence detection was done with ECL Plus Western blotting detection reagents (Amersham Biosciences by GE Healthcare). For IL-1β detection with Western blot analysis, isolated monocytes, in 96-well plates at a density of 2 × 10⁶ cells/ml and 100 μl/well, were stimulated in fresh RPMI 1640 with 100 ng/ml LPS for 4 h. Medium was changed after LPS priming to a gluconate basal salt solution described earlier (33), omitting glycine. Further processing was in analogy to the CBA (BD Biosciences) protocol, but lyophilisates were directly dissolved in 10 μl LDS sample buffer for gel electrophoresis.

Ethidium⁺ uptake was measured in isolated CD14⁺ human monocytes (according to 41), but without measuring additional cell markers. Primary human cells were used for this assay because they gave a better signal-to-noise ratio than U937 cells.

For IL-1β detection with CBA, the isolated monocytes, in 96-well plates at a density of 2 × 10⁶ cells/ml and 100 μl/well, were stimulated in fresh RPMI 1640 for 4 h under different priming conditions (vehicle, 100 ng/ml LPS and/or 10 μg/ml ginger Hot Flavor extract). Cells were spun down for 5 min at 200 × g, and medium was replaced (to be processed as described below) with 100 μl fresh one containing different stimuli (vehicle, 1 mM ATP and/or 10 μg/ml ginger Hot Flavor extract), then incubated for 30 min and centrifuged for 5 min at 200 × g, and supernatant was removed. Medium samples and cells were shock frozen and lyophilized, to be stored at −80°C until measurement. Lyophilisates were reconstituted in 10 μl hypotone buffer (10 mM HEPES, 0.3 mM EGTA, 0.1% Triton X-100, adjusted with KOH to pH 7.4 at RT), and IL-1β was quantified with IL-1β BD FlexSet.

U-937 cells (10⁶ cells/ml) were kept in culture medium, according to the recommendations by American Type Culture Collection. Twenty-four hours before membrane processing, cells were stimulated with 35 ng/ml PMA, according to published protocols (42, 43). Two times 2.8 × 10⁷ U937 cells were lysed in a hypotone lysis buffer (340 mM sucrose, 10 mM HEPES, 1 mM EDTA, 1 mM DTT, pH adjusted with KOH to 7.4 at 4°C) and centrifuged at 1000 × g for 30 min at 4°C. The pellets were kept as whole membrane preparations because cPLA₂ does colocalize with cell membranes. Supernatants were ultrafiltrated with a 50-kDa molecular mass cut-off, resulting in lysate concentrates of 0.625–1.5 mg protein/ml for the U937 (quantified with the reducing agent and detergent-compatible protein assay by Bio-Rad).

Palmitoyl-6-O-ascorbic acid was synthesized (according to 44). Reaction conditions were as follows: 35°C was the reaction temperature; 8.8 ml 95% sulphuric acid, 2.29 g palmitic acid, and 1.76 g ascorbic acid; 100% ultrasonic power output with stirring for 15 min. Identity was confirmed by electrospray ionization mass spectrometry and two-dimensional nuclear magnetic resonance (NMR). Palmitoyl-6-O-ascorbic acid potassium salt (PAK) was generated before use by mixing appropriate solutions of palmitoyl-6-O-ascorbic acid in chloroform with KOH in methanol to result in a 100 mM mixture of 95:5 chloroform to methanol. Crude phosphatidylcholine was isolated from commercial grade lecithin (Hänseler AG) by quantitative de-oiling with acetone, enrichment of phosphatidylcholine by quantitative extraction with boiling ethanol, and finally, precipitation of impurities at 0°C. Crude phosphatidylcholine was separated over a silica gel column with a solvent system of dichloromethane/methanol/acetic acid (5:5:1). The extract was neutralized and dried by filtration over sodium carbonate and anhydrous magnesium sulfate. The phosphatidylcholine fraction had a TLC purity of >95% and was stabilized with 10 ppm α-tocopherol, dried, and stored at −80°C.

The final phospholipase buffer consisted of 10 μl lysate and 10 μl mixed micelle buffer (for iPLA₂: 190 mM MOPS, 152 mM KOH, 3 mM DTT, 2 mM ATP, 1 mg/ml BSA, and for cPLA₂: 190 mM MOPS, 152 mM KOH, 3 mM DTT, 1.2 mM CaCl₂, 1 mg/ml BSA, 20 μM bromoenol lactone [BEL]) containing mixed micelles (16 mM PAK, 2 mM Tween 80, and 2 mM thio-PC, self-forming at ∼T > 35°C). (Final concentrations are therefore 100 mM Good’s Buffer adjusted with KOH to pH 7.4 at 40°C, 170 mM sucrose, 2 mM DTT, 0.5 mg/ml BSA, 8 mM PAK, 1 mM Tween 80, and 1 mM thio-PC, and either 1 mM ATP and 0.5 mM EDTA or 0.1 mM free calcium and 10 μM BEL.)

Stock solutions of PAK in chloroform, Tween 80, and phospholipid in ethanol were mixed and dried under a nitrogen flow. The mixed lipid film was hydrated in mixed micelle buffer at 40°C, briefly sonicated to detach film from glass surface if necessary, and gently (to avoid foam formation) shaken until the solution became totally clear.

Size and ζ potential of the micelles were analyzed on a zetasizer, giving one single sharp peak in both cases with an average size distribution of 32.3 ± 9.6 nm by volume (area of 99.3%) and 27.9 ± 6.25 nm by number (area of 100%) and an average ζ potential of −39.4 ± 3.25. Estimated by their size, they consist of ∼500 detergent molecules (quod est demonstrandum), and their ζ potential most likely traps c/sPLA₂ in the scooting mode (45, 46). Notably, these micelles are self-forming, sonication is not mandatory, and they are readily hydrolyzed by all three phospholipase subtypes tested.

Solutions were mixed at room temperature (RT), heated to 40°C (giving a clear solution), and again incubated at RT under gentle shaking. The enzymatic reaction was stopped after 3 h (>50% hydrolysis) with 16.6 μl ice-cold MeOH. A total of 0.2 μl DTT (100 mM aq. sol., for reduction of dithio-lyso-phosphatidylcholine), 0.2 μl Cs₄EDTA (100 mM aqueous solution, to chelate free calcium, which otherwise would interfere with the reaction), and 2 μl Cs₂CO₃ (200 mM aq. sol., to achieve a pH >8) was added and incubated for 1 h at RT. Then 20 μl monobromobimane (10 mM stock in MeOH) was added and incubated for 1 h at RT. The reaction was stopped and stabilized by acidification with 1 μl trichloroacetic acid (2.2 M aqueous solution, to achieve a pH <3) and centrifuged at 500 × g for 5 min. The supernatant could be stored with no detectable decomposition for 3 d at −18°C until HPLC measurement. Analytical HPLC was performed on an Elite LaChrome device (VWR International, Hitachi) equipped with a fluorescence detector using a Nucleodur Sphinx column (4 μm particle size, 4.6 × 100-mm dimension; Macherey-Nagel). A 15-min gradient of 0.25% NH₃ in water/acetonitril (LiChrosolv; Merck) from 30:70 to 10:90 was run for analysis, and UV detection at λ_ex 385 nm and λ_em 485 nm was carried out.

The enzyme activity was calculated from the area under the curve in absolute values (nmol product total) and (nmol product/min/mg protein).

TLC was done with a solvent mixture (according to 47). For detection of phosphate-containing lipids, a molybdenum blue spray solution and charring were used.

MAPK phosphorylation in leukocytes was quantified using BD FlexSets, according to the manufacturer’s recommendations, but with 5 times smaller volumes.

Isolated human lymphocytes were adjusted to 4 × 10⁶ cells/ml and aliquoted at 50 μl into 96-well plates. The 50 μl diluted stimuli were added as 2× stock and incubated at 37°C for the indicated time. Conditions for p38 phosphorylation were 10 μl/ml anti-CD3 Ab and 2 nM PMA for 20 min; for ERK1/2 phosphorylation, 10 nM PMA for 30 min; and for JNK1/2 phosphorylation, 10 μl/ml anti-CD3 Ab and 10 nM PMA for 60 min. Subsequently, samples were cooled on ice and centrifuged at 300 × g for 3 min at 4°C. The pellets were lysed with 5 μl diluted BD denaturation buffer containing phosphatase inhibitors (4 mM sodium potassium tartrate, 2 mM imidazole, 1.15 mM sodium molybdate, 1 mM sodium fluoride, and 1 mM sodium orthovanadate) and heated to 80°C for 5 min. The samples were cooled on ice, diluted with 5 μl assay diluent, and shock frozen in liquid nitrogen. After rethawing, the lysates were vigorously pipetted to give a clear solution. Finally, phosphorylation was determined with BD FlexSets.

CD4⁺ human lymphocytes were cultured in 96-well plates at 1 × 10⁶ cells/ml, and 100 μl/well was substituted with additional 50 μl RPMI 1640 medium containing 0.625 μl BrdU stock solution and 2 μl anti-CD3 Ab per ml or 50 μl RPMI 1640 medium containing vehicle control or test compounds prior to incubation for 5 d.

Proliferation was determined with the BD FITC BrdU Flow Kit by FACS, according to the recommendations by the manufacturer (48).

A clear correlation between the proliferation rate and residual CD14⁺ monocytes could be observed and is in agreement with earlier reports (49, 50).

Palmitatoyl-6-O-ascorbate was verified by two-dimensional NMR done on a Bruker 400 UltraShield and Bruker TOPSPIN 1.3 software.

For in process controls for palmitatoyl-6-O-ascorbate synthesis and phospholipase assay establishment, mass spectrometry (Waters Alliance HT with separation module 2795 coupled to a dual λ absorption detector 2487 and MassLynx V4.0 software) was used. Solvent was 30% water with 0.1% formic acid and 70% acetonitrile; 20 μl sample solution was injected. The mass range measured was m/z 100–500 and/or m/z 200–600, electrospray ionization (cone voltage): ±20 eV, +40 eV, and +80 eV; capillary voltage 3.0 kV; extractor 1 V; and RF lens 0.2 V. Temperature was set at 120°C and desolvatation temperature at 250°C. N₂ was used as carrier gas at a flow rate of 600 l/h for desolvatation and for cone flow at 40 l/h.

Quantification was performed by a gas chromatography/mass spectrometry method. A total of 5 × 10⁶ U937 cells was cultured in complete RPMI 1640 medium and differentiated with 2 nM PMA for 48 h and washed with PBS, and the medium was replaced with FBS-free medium. Cells were incubated with test compounds (DMSO, 2 μM 10-shogaol, 2 μg/ml ginger Hot Flavor extract, 50 and 100 μM aspirin, 0.5 μM methyl arachidonyl fluorophosphonate [MAFP]) and incubated for 30 min. Then LPS (1 ug/ml) was added, and the cells were incubated for another 4 h. Cells were scraped off and separated from the supernatant by centrifugation for 5 min at 1000 rpm in a Heraeus Biofuge fresco (rotor 3325).

Lipids were extracted from the supernatants and the cell fractions as follows: 1 ml ethanol was added to internal standards (vide infra) and mixed with 9 ml supernatant. The pH was brought to 3 with hydrochloric acid and the samples added to C 18 Sep-Pak cartridge (Waters) (preactivated with 3 ml methanol and equilibrated with 3 ml 10% ethanol). Cartridges were washed with 10% ethanol and eluted with 3 ml acetonitrile/ethyl acetate (1:1). Supernatant samples were evaporated to dryness.

Cells were extracted similar to the lipid extraction described by Folch et al. (51). In short, cell pellets were sonicated for 5 min at 4°C ice-cold chloroform (1 ml containing the internal standards), methanol (0.5 ml), and PBS (0.25 ml), and then centrifuged for 5 min at 800 × g. The organic phase was dried in a glass vial and dried under N₂ and reconstituted in 1 ml ethanol by vortexing at RT (<5 min), diluted with 9 ml water, and extracted by solid-phase extraction, as described for the supernatant.

Internal standards were as follows: PGE₂-d4, 4 ng/μl (100 ng/sample); AA-d8, 4 ng/μl (100 ng/sample); and palmitoylethanolamide-d5, 1 ng/μl (25 ng/sample).

Derivatization of the hydroxyl group of palmitoylethanolamide (PEA) was performed with the silylating agent dimethylisopropylsilyl imidazole. As described by Obata et al. (52), the molecular stability is increased and large fragments could be detected. Derivatization of AA was achieved by esterification with pentafluorobenzylbromide and N,N-diisopropylethylamine (53). The hydroxyl groups of PGE₂ were readily silylated with dimethylisopropylsilyl imidazole and the carboxylic acid as for AA with pentafluorobenzylbromide. The remaining ketone was transformed into an O-methyl oxime by methoxiyamine hydrochloride in pyridine (52).

The samples were analyzed by GC/electron ionization mass spectrometry using an Agilent 6890N GC equipped with a 30 m HP-5MS column and a 5975C MS with triple-axis detector. As carrier, gas helium was used at a constant flow rate of 1.5 ml/min with splitless injection at an inlet temperature of 250°C. Optimal separation of the three analytes was achieved with the following oven program: initial temperature, 150°C for 1 min, followed by an increase to 280°C at 8°C/min, with a final time of 20 min. Specific ions were used for selected ion monitoring (54).

TXB₂ and LTB₄ were quantified from cell culture supernatants using commercially available ELISA kits.

A convenient high-content in vitro assay was established to profile the effects of traditional anti-inflammatory multicomponent agents (i.e., botanical drugs) with yet unknown molecular mechanisms of action. It was our aim to employ an assay that was potentially less susceptible to false positives (e.g., effects only detected in tumor cell lines, but not primary cells) and more closely associated with physiological parameters like blood plasma components. We used human venous whole blood, which was immediately transferred to 96-well plates (200-μl aliquots) and incubated with discrete stimuli over 18 h. As the primary readout, we measured differential Th1/Th2 cytokine expression (Supplemental Fig. 1). To induce patterns of differential cytokine expression, stimuli engaging distinct receptors and signaling pathways were applied. These stimuli included bacterial LPS acting at CD14/TLR4/MyD88 (55); PMA that activates protein kinase C (56) and nonspecifically amplifies cellular signals; anti-CD28 and anti-CD3 Abs acting at TCR subtypes (57, 58); and thapsigargin that nonspecifically triggers a cytosolic calcium increase in all cells (59). As shown in Supplemental Fig. 1, different stimuli that target distinct leukocyte populations triggered significantly distinctive patterns of cytokine expression. The data show mean values from independent experiments performed with blood from different female and male donors (n = 18) and clearly demonstrate that the cytokine network is widely conserved between different individuals, thus providing a robust high content readout. As anticipated, LPS stimulation induced mainly Th1 monokines with strong TNF-α, IL-1β, IL-6, IL-8, and IL-10 expressions (>10,000 pg/ml), but weak IL-12 expression (Supplemental Fig. 1). Under these conditions, the T cell cytokines IL-3, IL-4, IL-5, IL-7, and GM-CSF were not induced. In contrast, PMA/anti-CD3 Ab stimulation led to a robust Th2 cell response with similar IL-8, IL-12, and TNF-α expressions; less pronounced IL-1β, IL-6, and IL-10 expressions; but significantly stronger IL-3, IL-4, IL-5, IL-7, and GM-CSF expressions.

To validate our assay, we applied specific inhibitors of key inflammatory processes or signal transduction events. Clinically and experimentally used inhibitors were dexamethasone, cyclosporin, diclofenac, cetirizine, parthenolide, the MAPK kinase inhibitors SB203580 (p38 MAPK inhibitor), SB202190 (p38 MAPK inhibitor), SP600125 (JNK inhibitor), U0126 (MEK inhibitor), and PD98059 (selective MEK1 inhibitor) (Fig. 1). As expected, dexamethasone potently inhibited the expression of proinflammatory cytokines independent of the stimulus used. Intriguingly, in this assay, only dexamethasone strongly inhibited LPS-stimulated TNF-α expression. The MEK1/2 inhibitor U0126 and the p38 inhibitor SB203580 weakly inhibited TNF-α. Distinct MEK and p38 inhibitors like U0126 and PD98059, as well as the broad-spectrum enzyme inhibitor 2,4′-dibromoacetophenone (4-bromophenacyl bromide or DBA), significantly inhibited LPS-stimulated IL-1β expression, but with a high interassay variability. Based on our assumption that certain ginger constituents may be able to modulate proinflammatory cytokine expression (vide supra), the commercial food-grade ginger Hot Flavor extract (50 μg/ml) was analyzed in this assay. Additional anti-inflammatory medicinal plant extracts, like Harpagopyhtum procumbens and Salix spp., were tested in the same setup. The ginger extract potently inhibited IL-1β expression with little interassay variability, whereas the H. procumbens and Salix extracts (both 50 μg/ml) and curcumin (25 μg/ml) were largely ineffective (Fig. 1). Moreover, the ginger extract differentially modulated the expression of several cytokines, depending on the stimulus used (vide infra). All inhibitors of kinases (SB203580, SB202190, SP600125, U0126) and NF-κB (parthenolid and curcumin), as well as DBA and cetirizine, potently inhibited the PMA/anti-CD3 Ab-stimulated GM-CSF (Fig. 1), which may be explained by the fact that this factor is least strongly induced (<1500 pg/ml). As anticipated, cyclosporine potently inhibited the expression of CD3/CD28-induced factors, as exemplified by GM-CSF and TNF-α (Fig. 1), but also other calcium-dependent cytokines (IL-3, IL-4, IL-5, IL-7) (data not shown). The COX-2 inhibitor diclofenac and the H1-antagonist cetirizine were largely ineffective, whereas the broad-spectrum enzyme inhibitor DBA significantly inhibited most cytokines (Fig. 1). As blood from different donors was employed, we concluded that the inhibitory effects observed were robust and meaningful.

Intriguingly, the ginger extract (50 μg/ml) potently inhibited IL-1β expression (≥35%) in all experiments in which whole blood was stimulated, irrespective of the stimulus applied (Fig. 2). In contrast, the other cytokines were not or only partially inhibited (Figs. 1, 2), also reflecting the lack of general cytotoxicity of ginger extracts under these assay conditions (data not shown). Based on this finding, we concluded that ginger constituents could selectively and globally interfere with the IL-1β expression machinery by concrete, yet unknown mechanisms (Supplemental Fig. 2). Because the stimuli employed induced distinctly different signal transduction events, we excluded the possibility that upstream events, such as, for example, inhibition of MAPKs or transcription factors, could be responsible for this effect.

Both the purinoreceptor P2X₇ and PLA₂ enzymes have been shown to be crucial for efficient IL-1β expression (maturation and secretion) from monocytes/macrophages (60). To explore the effects of ginger on IL-1β, experiments in which LPS stimulation was coactivated by ATP were performed in U937 cells, a human monocyte/macrophage cell line, to differentiate between IL-1β maturation and secretion. U937 cells that are stimulated by LPS empty the IL-1β stores (by yet unknown mechanisms), leading only to partial IL-1β section (Fig. 3A) (61). When the cells are costimulated by ATP (which is elevated under inflammatory conditions and present in whole blood), the purinoceptor P2X₇, a ligand-gated ion channel, is activated, leading to activation of caspase-1 that cleaves pro–IL-1β into the mature IL-1β. In parallel, this causes an increase in [Ca²⁺]_i, which triggers the release of the IL-1β storage vesicles. As shown in Fig. 3A, the ginger extract significantly and specifically inhibited the ATP/LPS-stimulated IL-1β secretion from U937 cells, thus indicating potential effects on P2X₇, PLA₂ enzymes, and/or caspase-1. The level of ethidium bromide uptake after ATP stimulation can be regarded as equivalent to general cation influx through large P2X₇ receptor adjacent pores (41). The assay was performed in isolated human monocytes, as they showed a better signal-to-noise ratio than U937 cells. In this assay, P2X₇ activity was not inhibited by the ginger extract (Fig. 3B).

Western blot and cytometric bead analyses showed that the ginger extract inhibits maturation and release of IL-1β from activated monocytes by ∼60% when the ginger extract was added before ATP (Fig. 3C). Under nonstimulated conditions, the isolated human monocytes (2 × 10⁵) secrete ∼4 pg mature IL-1β within 4 h in culture and ∼60 pg/ml when incubated (primed) with LPS. The constitutive secretion was statistically unchanged by 10 μg/ml ginger extract (Fig. 3A). A total of 1 mM ATP did not modulate cytokine secretion in nonprimed cells. However, stimulating LPS-primed cells with ATP caused an increase in cytokine secretion (180 pg/ml) (Fig. 3A). This stimulated secretion was significantly reduced by >50% when ginger extract (10 μg/ml) was added prior to the LPS priming and by ∼40% when added after LPS and prior to ATP. These levels were significantly different between LPS plus ATP, but insignificant when compared with LPS alone, ruling out an effect on caspase-1. A differential effect on intra- and extracellular pro- and mature IL-1β species was clearly visible in Western blots (Fig. 3C) even though the inhibition of maturation and secretion was less pronounced than in the assay conditions used for quantitative CBA (Fig. 3A). In isolated human monocytes, the iPLA₂-dependent IL-1β maturation and the cPLA₂-dependent IL-1β secretion were reduced to ∼60% (at 10 μg/ml ginger extract), whereas transcription/translation and constitutive maturation/secretion were unaffected (Fig. 3C). Therefore, we concluded that the major ginger constituents (Fig. 4) could be potential inhibitors of PLA₂ enzymes (Supplemental Fig. 2).

To measure i/cPLA₂ enzyme activities, a suitable assay was established using mixed micelles and a fluorescence-coupled assay (Fig. 5A, 5B) (for experimental details, see 1Materials and Methods). The major lipophilic ginger constituents (Fig. 4) were purchased or isolated, as previously reported (17), and tested at 10 μM. As shown in Fig. 5C, the ginger extract and its main ginger constituents inhibited both iPLA₂ and cPLA₂ enzyme activities. Overall, the phenylpropanoids more strongly inhibited iPLA₂. The inhibition of ∼50% was in the same range as the inhibition by the irreversible PLA₂ inhibitors MAFP and BEL. The relatively low efficacy of the latter two may be due to the preincubation scheme used (62, 63) or more likely residual calcium- and phospholipase-independent hydrolytic activities. Nevertheless, the assay employed was robust enough to detect i/cPLA₂ inhibitors and to compare potencies of individual compounds. The inhibitor concentration was as low as 0.1 molar percent of the total detergent molecules a possible inhibition by surface dilution kinetics could be excluded (64). The ginger extract and 10-shogaol concentration-dependently inhibited both iPLA₂ and cPLA₂ with similar potencies as BEL and MAFP (Fig. 5D). However, the inhibition of iPLA₂ was clearly stronger at lower concentrations than the inhibition of cPLA₂, in particular with the ginger extract (EC₅₀ values 0.7 μg/ml versus 3 μg/ml). For iPLA₂, the eight homologs were less active than the controls at equimolar concentrations (10 μM), but 6-shogaol and 10-gingerol were somewhat more potent, with an overall inhibition of ∼65% (Fig. 5C). Although 6-,8-gingerols did not inhibit cPLA₂ and 10-shogaol was the most potent inhibitor, structure-activity relationships were not apparent. To assess the specificity of this effect, we also measured sPLA₂. Porcine pancreatic sPLA₂ activity was tested using isolated phosphatidylcholine, and an established TLC detection method (65, 66) was not inhibited by ginger constituents up to 20 μM, but strongly inhibited by the nonspecific enzyme inhibitor DBA (data not shown).

iPLA₂ inhibition has been shown to contribute to the proliferation of lymphocytes, Jurkat T cells (28), and monocytes (29). As shown in Fig. 6, the proliferation of anti-CD3 Ab-stimulated human lymphocytes was weakly, but significantly inhibited by the ginger CO₂ extract (10 μg/ml) and 10-shogaol (>5 μM), but not by a totum extract containing a higher essential oil content. To address the effect of ginger constituents on MAPKs previously reported with cancer cells (54), we also measured the effects of gingerols and shogaols on ERK, JNK, and p38 in primary human T cells. PMA/anti-CD3 Ab-stimulated lymphocytes were used to determine the modulation of MAPKs by measuring kinase phosphorylation states using a commercial CBA assay (from BD Biosciences). As shown in Fig. 6, experiments using primary human T cells did not show significant modulations. Phosphorylation of p38 was even increased by ginger constituents. Only 10-shogaol significantly inhibited JNK phosphorylation by ∼35%, but the whole extract showed a trend toward activation of this kinase, and thus, no conclusive picture.

Based on the finding that ginger phenylpropanoids (Fig. 4) inhibit the PLA₂ enzymes in the mixed micelles assay (Fig. 5), we next assessed the effects of ginger extract and 10-shogaol on free AA levels and PGE₂ release directly in differentiated U937 macrophages (differentiated with 5 nM PMA for 48 h). A quantitative gas chromatography/mass spectrometry analysis was employed to determine both AA and PGE₂ levels in cell supernatants (cell medium) and in the cellular fractions (see 1Materials and Methods). Whereas free AA could be detected in both supernatant (5.4 nmol/1 × 10⁷ cells) and in the thoroughly washed cellular fraction (0.4 nmol/1 × 10⁷ cells) of undifferentiated U937 cells, PGE₂ was only found in the supernatant of U937 macrophages. U937 macrophages constitutively released PGE₂ (0.1 nmol/1 × 10⁷ cells) even without LPS stimulation (Fig. 7A). Upon stimulation by LPS, the PGE₂ levels stably increased ∼3-fold. As shown in Fig. 7B, the positive control acetyl salicylic acid (aspirin) at concentrations at which COX-1/2 is fully inhibited in vitro (>10 μM) inhibited PGE₂ release from LPS-stimulated U937 cells. The PLA₂ and nonspecific hydrolase inhibitor MAFP (0.5 μM) and 10-shogaol (2 μM) inhibited the release of PGE₂ and the prostanoid metabolite TXB₂ by >50% (Figs. 7A, 7B, 8). In nonstimulated U937 macrophages, 10-shogaol inhibited constitutive PGE₂ expression by >50% (Fig. 7A). Noteworthily, even concentrations of ginger extract as low as 2 μg/ml showed a significant inhibition of PGE₂ and TXB₂ release, whereas LTB₄ was only weakly inhibited. As expected, BEL, which more specifically inhibits iPLA₂, had no effect on PGE₂, TXB₂, or LTB₄ production, whereas MAFP also inhibited LTB₄ (Figs. 7, 8). Moreover, PEA, which is also constitutively released by U937 macrophages (0.1 nmol/1 × 10⁷ cells), was not modulated by ginger extract or 10-shogaol (Fig. 7F, 7G). LPS stimulation did not increase PEA secretion. Somewhat unexpectedly, ginger extract and 10-shogaol significantly increased the free AA levels in U937 macrophages in the cellular fraction and in the supernatant under both untreated and LPS-stimulated conditions (Fig. 7C–E). A PLA₂ inhibitor would be expected to rather decrease free AA levels. MAFP reduced free AA in the supernatant, but aspirin, as expected, did not influence free AA levels. Because the ginger extract and 10-shogaol inhibit both iPLA₂ and COX, we speculated that this double effect may lead to an increase of free AA from the pool that constitutively feeds COX catalyzing prostanoid synthesis. We therefore combined the iPLA₂ inhibitor BEL and aspirin. As shown in Fig. 7C and 7D, this combination led to an increase of free AA similar to what was observed with ginger phenylpropanoids. We concluded that the disruption of arachidonate-phospholipid remodeling, that is, the blockage of phospholipid reacylation by iPLA₂ inhibition, and the concurrent inhibition of COX is responsible for the unexpected effect on free AA by ginger phenylpropanoids (Supplemental Fig. 3).

A validated high-content whole blood assay was established to profile robust cytokine modulating effects of reported immunomodulatory agents, including botanical drugs. This has led to the elucidation of a new anti-inflammatory molecular mechanism of action of gingerols and shogaols, which are the major phenylpropanoids in dietary ginger (67, 68). Our study indicates that the potent inhibition of i/cPLA₂ by phenylpropanoids is related to the specific inhibition of IL-1β secretion from monocytes/macrophages by ginger (Supplemental Fig. 2). Moreover, our data confirm the inhibition of COX (10, 11, 67–69) by ginger phenylpropanoids. This is shown by the inhibition of LPS-stimulated PGE₂ and TXB₂ secretion (Figs. 7, 8) from U937 macrophages, indicating nonselective inhibition of COX enzymes. LPS-stimulated LTB₄ from primary human monocytes/macrophages (U937 macrophages did not secrete LTB₄) was only weakly inhibited by the ginger extract. The double inhibition of prostanoid secretion and IL-1β by ginger phenylpropanoids is interesting because IL-1β is known to reinforce PGE₂ synthesis (70), which plays multiple roles in inflammatory processes (71, 72).

Our study further shows that the significant cellular inhibition of prostanoid production by ginger phenylpropanoids together with the inhibition of iPLA₂ leads to a dramatic and unexpected increase of free AA. This was corroborated in an experiment in which BEL and aspirin were coincubated, and in this combination also led to significant increase of free AA levels (Fig. 7). Free fatty acids are first bound to CoA and then incorporated into lipids by either de novo synthesis of triglycerides (Kennedy pathway) or remodeling of phospholipids (Lands cycle) (23, 24) (Supplemental Fig. 3). In the case of AA, the latter involves incorporation of arachidonate into phosphatidylcholine by a deacetylation/reacetylation reaction, followed by a phospholipid class switch via CoA-independent transacylation (73, 74). This is the main pathway for AA incorporation in most cell types (75). In both of these processes, phospholipases, and especially iPLA₂, play major regulatory roles by providing free fatty acid acceptor molecules (e.g., lyso-phosphatidylcholine) (25). In contrast, immune cells generate free AA mainly from phospholipids by cPLA₂ IVA, the only PLA₂ with a preference for AA (46, 76–79), and only under certain conditions by sPLA₂ IIA, V, and X and iPLA₂ VIA (74). Because we observed a significant increase of free cellular AA by ginger phenylpropanoids in U937 macrophages (Fig. 7E), the iPLA₂ inhibition is likely to be predominant in cells, and sPLA₂ enzymes are not affected. The sPLA₂ enzymes may be the cause of the free AA observed in our assays. The reason that MAFP at concentrations around the IC₅₀ of reported i/cPLA2 inhibition (80, 81) only weakly inhibits free AA in the supernatant is not clear, but may be due to the combined i/cPLA₂ isoform inhibition and a resulting indirect reduction of extracellular sPLA₂ activity. Moreover, a fraction of the AA that is released by PLA₂ activation will be rapidly reincorporated into phospholipid, whereas the remainder will be lost by conversion to eicosanoids or other products, or by β-oxidation (82, 83). Unesterified AA in plasma rapidly replaces the amount lost, and this replacement is proportional to PLA₂ activation (84, 85). In our experiments, BEL alone inhibited free AA, but increased free AA when COX was inhibited at the same time. This dramatic change (Fig. 7C, 7D) shows the role of iPLA₂ for AA phospholipid reacylation. Therefore, in the presence of COX inhibitors, blockage of iPLA₂ deprives the cell of phospholipid acceptor molecules and subsequently augments the intracellular AA concentration (75).

Given that free AA in cells induces apoptosis (29), the pronounced increase of free cellular AA may also explain some of the differential antiproliferative effects of gingerols and shogaols on cancer cells (86, 87). In contrast, PEA levels were not affected by ginger phenylpropanoids. Its biosynthesis mainly relies on N-acyltransferases, N-acylphosphatidylethanolamine phospholipase D, and fatty acid amide hydrolases, and only to a minor extent on lyso-phospholipase D and sPLA₂, but not on other PLA₂ classes (88, 89). Consequently, our data indicate selectivity toward i/cPLA₂ of the ginger extract and its main constituents and exclude unspecific perturbation of lipid homeostasis. The action of ginger phenylpropanoids is in line with the already known antioxidant and radical scavenging effects of ginger (90, 91) and may further increase its effect on iPLA₂ in cells. Reactive oxygen species produced by cyclo-/lipoxygenases are known to either activate iPLA₂ in a positive feedback loop (42) or act as mediator between cPLA₂ and sPLA₂ (92–95).

The ginger rhizome and its extracts have been shown to be safe, as exemplified by its widespread dietary use in Asia (1). In contrast, many tested synthetic i/cPLA₂ inhibitors exert unwanted side effects due to unspecific toxicity/reactivity (e.g., BEL or MAFP), poor selectivity (e.g., MAFP and arachidonyl trifluoromethyl ketone), or lack of oral availability [e.g., EXPLIS (96)]. Ginger extracts or isolated compounds show similar in vitro potencies against iPLA₂ enzymes as standard PLA₂ inhibitors, but are nontoxic. Ginger as a botanical drug has a great acceptance in the population, and might therefore be used as a physically and mentally well-tolerated augmentation to conventional anti-inflammatory medication in cases where first-line therapy is not sufficient. In particular, treatment of inflammatory bowel syndrome and celiac disease in which IL-1β and PGE₂ play a major role (97) or autoimmune inner ear disease (98) could be novel therapeutic applications of ginger. Overall, the inhibition of PLA₂ enzymes provides a rational basis for several reported properties of ginger [anti-inflammatory, antipyretic, analgesic, or cardiovascular effects (86, 87)].

We acknowledge the numerous blood donors and medical assistants, in particular Maria Feher. The ginger extracts were provided by FLAVEX.

The authors have no financial conflicts of interest.