O-Methylated Catechins from Tea Leaves Inhibit Multiple Protein Kinases in Mast Cells¹

Mast cells play a critical role in the effector phase of IgE-dependent immediate hypersensitivity and allergic diseases (1). Cross-linking of high affinity IgE receptors (FcεRI) with IgE and allergen initiates the activation process, leading to the release of preformed and de novo synthesized vasoactive amines, proteases, leukotrienes, cytokines, and chemokines. These chemical and polypeptide agents elicit various allergy-associated pathophysiological changes locally and systemically: for instance, amines such as histamine and serotonin enhance vascular permeability, and cytokines such as TNF-α recruit inflammatory cells to the site of allergen exposure.

Over the past two decades, we have seen a tremendous advancement in our understanding of what happens inside mast cells following FcεRI cross-linking (2). FcεRI is a heterotetrameric receptor composed of an IgE-binding α subunit, a four-transmembrane β subunit, and disulfide-bonded two γ subunits (3). γ subunits are critical for both membrane expression of the receptor and signal transduction, while β subunits amplify the γ subunit-generated signals and stabilize the receptor. β subunits are associated constitutively with a predominant Src family protein tyrosine kinase (PTK)³ (3), Lyn (4). Upon FcεRI cross-linking, Lyn becomes activated and phosphorylates tyrosine residues in the immunoreceptor tyrosine-based activation motifs in β and γ subunits (5, 6). Tyrosine-phosphorylated immunoreceptor tyrosine-based activation motifs of β and γ subunits recruit Lyn and Syk (another PTK with two tandem Src homology 2 domains), respectively, to the plasma membrane and activate these PTKs (4, 7, 8, 9). Active Lyn and Syk phosphorylate numerous signaling proteins. Another PTK, Bruton’s tyrosine kinase (Btk) (10), is phosphorylated at its critical activation-loop tyrosine by Lyn and activated (11). Active Syk and Btk in turn phosphorylate and activate phospholipase C (PLC)-γ in a concerted manner (12, 13). PLC hydrolyzes phosphatidylinositol 4,5-bisphosphate into diacylglycerol and inositol 1,4,5-trisphosphate. Inositol 1,4,5-trisphosphate mobilizes Ca²⁺ from intracellular storage sites. Diacylglycerol and Ca²⁺ activate some isoforms of protein kinase C (PKC). Downstream of these early activation events, at least four major mitogen-activated protein kinase (MAPK) subfamilies of protein serine/threonine kinases (PS/TKs), i.e., extracellular signal-regulated kinase (ERK) 1/2, c-Jun N-terminal kinase (JNK) 1/2, p38, and big mitogen-activated kinase 1/ERK5 (14, 15, 16, 17, 18, 19), and a critical survival kinase, Akt/protein kinase B (Akt) (20), are also activated. Many of these kinases are shown to be involved in production and secretion of cytokines.

Tea (Camellia sinensis L.) has been consumed all over the world, particularly in large quantities in Japan and China, where it has also been used for medicinal purposes for tens of centuries. It has been reported that tea has various bioregulatory activities, such as anticarcinogenetic (21, 22) and antimetastatic activities (23, 24, 25, 26), antioxidative activity (27, 28, 29), antihypertensive activity (30), antihypercholesterolemic activity (31, 32), antidental caries activity (33), antibacterial activity (34), and intestinal flora amelioration activity (35). Major principles for these activities were shown to be catechins, a group of polyphenolic compounds. In previous studies, we have reported that O-methylated forms of (-)-epigallocatechin-3-O-gallate (EGCG), i.e., (-)-epigallocatechin-3-O-(3-O-methyl) gallate (EGCG3″Me) and (-)-epigallocatechin-3-O-(4-O-methyl) gallate (EGCG4″Me), which were contained in some cultivars of tea such as ‘Benifuki’ or ‘Benifuji,’ had antiallergic action in in vivo type I and type IV allergies (36, 37). In this study, we present evidence that two O-methylated catechins, EGCG3″Me and EGCG4″Me, have an antiallergic activity and inhibit several protein kinases as a molecular basis for such an activity.

Sources of commercial Abs are as follows: anti-phosphotyrosine mAb 4G10 from Upstate Biotechnology (Charlottesville, VA); anti-Lyn (38), anti-Syk (C-20), anti-Akt (C-20), and anti-JNK (C17) from Santa Cruz Biotechnology (Santa Cruz, CA); anti-phospho-ERK1/2, anti-phospho-JNK, and anti-phospho-p38 from Cell Signaling Technology (Beverly, MA); anti-mouse IgG F(ab′)₂ from BioSource International (Camarillo, CA); anti-Btk from BD PharMingen (San Diego, CA); anti-ERK1/2 from Zymed Laboratories (South San Francisco, CA); and DNP-specific IgE from Sigma-Aldrich (St. Louis, MO). Anti-FcεRI β subunit mAb JRK was kindly provided by J. Rivera (National Institutes of Health).

OVA and DNP conjugates of human serum albumin (DNP-HSA) were purchased from Sigma-Aldrich; IFA and diphenhydramine hydrochloride were from Wako Pure Chemical (Osaka, Japan).

Dried tea leaves (cultivars ‘Benihomare’ and ‘Benifuki’ or Ton-chin oolong tea) were extracted with boiling water for 30 min. The extract was passed through a porous polymer gel (Mitsubishi Diaion HP-20; Mitsubishi Oil Chemicals, Tokyo, Japan) column (9 × 40 cm). After the column had been washed with water, the absorbed fraction was eluted with water/methanol (1:1) to give a pale brown residue (31 g). The residue was chromatographed on a silica gel column (9 × 25 cm) and eluted with chloroform/methanol/water (82:16:2) and subsequently with chloroform/methanol/water (40:50:10) to yield six fractions. The peak fraction (2.657 g) that contained EGCG3″Me and EGCG4″Me was chromatographed on a reversed-phase (LOP-ODS; Nomura Chemical, Aichi, Japan) column (two columns connected in tandem; 5 × 50 cm, 2×) and eluted with water/acetonitrile (88:12) to give 241 mg of EGCG3″Me and 101 mg of EGCG4″Me. The physical properties of EGCG3″Me and EGCG4″Me (Fig. 1) were confirmed by ¹H nuclear magnetic resonance spectrum and ¹³C nuclear magnetic resonance spectrum.

Bone marrow cells from femur of CBA/J mice were cultured in mouse IL-3-containing RPMI 1640 medium supplemented with 10% heat-inactivated FBS (Invitrogen Life Technologies, Carlsbad, CA), 2 mM glutamine, and 50 μM 2-ME in humidified 95% air/5% CO₂ at 37°C (39). More than 95% pure mast cells (bone marrow-derived mast cells (BMMC)) were obtained after 4 wk of culture. These cells were incubated overnight with 0.5 μg/ml DNP-specific IgE mAb, and stimulated with a multivalent Ag, DNP-HSA, at 100 ng/ml in Tyrode solution for the indicated intervals. In some experiments, IgE-sensitized cells were precincubated with the indicated concentrations of EGCGs for 30 min and stimulated with Ag in the continued presence of EGCGs for 3 min or various periods. Cell suspensions were briefly centrifuged to separate Tyrode solution from cell pellets. Histamine secreted into the Tyrode solution during 30-min stimulation was measured by on-column HPLC. Leukotrienes C₄/D₄/E₄ secreted into the Tyrode solution were quantified by an enzyme immunoassay (Amersham Biosciences, Piscataway, NJ) and IL-2 by an ELISA kit (Endogen, Woburn, MA).

Cell pellets were lysed in ice-cold 1% Nonidet P-40-containing lysis buffer (20 mM Tris-HCl, pH 8.0, 0.15 M NaCl, 1 mM EDTA, 1 mM sodium orthovanadate, 1 mM PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 25 μM p-nitrophenyl p′-guanidinobenzoate, 1 μM pepstatin, and 0.1% sodium azide) immediately after stimulation and centrifuged. Protein concentrations of cell lysates were measured using a DC protein assay reagent (Bio-Rad, Richmond, CA). For immunoprecipitation, lysates were incubated on ice with an appropriate Ab for 1.5 h, and immune complexes were recovered by brief centrifugation following another 30-min incubation with Pansorbin (Calbiochem, San Diego, CA) for rabbit polyclonal Abs or anti-mouse Ig-conjugated agarose (Sigma-Aldrich) for mouse mAbs. Immune complexes were washed in lysis buffer four times. Cleared cell lysates or immunoprecipitates were separated by SDS-PAGE and blotted electrophoretically to polyvinylidene difluoride (PVDF) membranes (Immobilon-P; Millipore, Billerica, MA). Membranes were blocked and incubated with a primary Ab and then with a HRP-conjugated secondary Ab. Visualization of immunoreactive proteins was performed with ECL reagents (PerkinElmer Life Sciences, Wellesley, MA). Relative amounts of immunoreactivity were quantified by National Institutes of Health Image Analysis software (v.1.62).

Immunoprecipitates were washed four times in lysis buffer and once with kinase buffer (50 mM Tris, pH 7.4, 0.1% Nonidet P-40, 10 mM MnCl₂, 10 mM MgCl₂) without ATP. Washed immune complexes were incubated in kinase buffer with or without exogenous substrate (Btk and Lyn without exogenous substrate; Syk with 2 μg of GST-HS1 containing the sequence from position 352 to position 486 of the human HS1 protein (40)) in the presence of [γ-³²P]ATP (ICN Biomedicals, Irvine, CA). Reaction products were analyzed by SDS-PAGE, followed by electroblotting onto PVDF membranes and autoradiography. Phosphorylated protein bands were quantified by densitometry.

Lyn and Btk were expressed in Sf9 insect cells and partially purified, as described (41). Recombinant human Syk was purchased from Upstate Biotechnology. Kinase assays were performed, as described above, for immune complex assays, except for the presence of various concentrations of EGCGs throughout the assay period. Kinase activities of PKCα, PKCβI, and PKCβII (all recombinant human isoforms from Panvera, Madison, WI) in the absence or presence of EGCG3″Me and EGCG4″Me were measured using the PKC pseudosubstrate peptide PKC (19, 20, 21, 22, 23, 24, 25, 26, 27) in a PKC assay kit (Invitrogen). Assays on ERK2 (recombinant mouse; New England Biolabs, Beverly, MA) and JNK1 (recombinant mouse; Upstate) in the absence or presence of O-methylated EGCGs were performed using myelin basic protein and GST-c-Jun (1–79), respectively, as substrate.

Kinetic analysis of EGCG3″Me inhibition of Lyn kinase was performed at pH 7.2 in kinase buffer, using 1 μg of purified Lyn per assay in the presence of various concentrations of [γ-³²P]ATP with or without 0.4 μg/ml EGCG3″Me. Reaction products were analyzed by SDS-PAGE, followed by electroblotting onto PVDF membranes and autoradiography. Phosphorylated protein bands were quantified by densitometry. Kinetic data were analyzed by Lineweaver-Burk plotting.

BMMC were transiently transfected by electroporation with a reporter construct, IL-2/luc (42). Twenty-four hours later, the cells were sensitized with IgE and stimulated by Ag in the absence or presence of O-methylated EGCGs for another 7 h. Luciferase activity was measured, as described previously (42).

Ca²⁺ influx was analyzed by spectrofluorometry, as previously described (43). Briefly, BMMC were incubated at 37°C for 30 min with 5 μM fura 2-AM (Dojindo, Osaka, Japan) in Tyrode/HEPES solution after overnight sensitization with anti-DNP IgE. After washing with HBSS, aliquots of the cell suspension were placed in a cuvette and incubated with O-methylated EGCGs at 37°C for 10 min, and then 100 ng/ml DNP-HSA was added. Fluorescence was recorded using a fluorometer (CAF-110; Spectroscopic, Tokyo, Japan) by alternating excitation at 340 and 380 nm and reading the emission at 510 nm. O-methylated EGCGs did not exhibit autofluorescence nor interfere with fura 2 fluorescence.

As a model of in vivo type I allergic reactions, we performed mast cell-dependent PCA experiments, as previously described (44). Briefly, male ddY mice (Japan SLC, Hamamatsu, Japan) were intradermally injected with 10 μl of anti-DNP IgE at both ears. Catechin samples (or distilled water as control) were administered i.v. or orally to the mice 60 min before Ag administration. Intravenous administration of a 0.25-ml mixed solution of Ag (0.1 mg/ml DNP-BSA) and Evans blue dye (0.5%) was i.v. injected 24 h after IgE sensitization. The mice were killed by cervical dislocation 30 min later, and the ears were removed. Evans blue dye leaked into ear tissues was extracted and quantified by spectrophotometry. To confirm the role of histamine released from activated mast cells in inducing PCA reactions, diphenhydramine, an inhibitor of the histamine H₁ receptor, at 20 mg/kg, was i.p. injected 30 min before Ag challenge. As shown in Fig. 2, this treatment consistently inhibited PCA reactions by 70–85%.

Statistical analysis was performed using Student’s t test or Dunnett’s multiple comparison test. Values with p < 0.05 were considered statistically significant.

EGCG is known to inhibit immediate hypersensitivity reactions (38, 45). We first compared in vivo effects on mast cell-dependent PCA reactions between different tea catechins, EGCG and its O-methylated derivatives (Fig. 2). Mice were sensitized for 24 h by intradermal injection of anti-DNP IgE at the ear, and then they were i.v. injected with DNP-BSA solution containing Evans blue dye. Thirty minutes later, the mice were sacrificed, and the dye leaked into the ear was extracted and quantified. Intraperitoneal administration of EGCG 60 min before Ag/dye injection substantially inhibited allergic reactions (10, 27, and 43% inhibition by a single dose of EGCG at 5, 10, and 20 mg/kg, respectively). Both 3-O- and 4-O-methylated derivatives, EGCG3″Me and EGCG4″Me, were 2 times more potent in inhibiting PCA reactions than EGCG. Oral administration of EGCG and O-methylated derivatives 60 min before Ag/dye injection exhibited similar inhibition of PCA reactions, with slightly weaker inhibition compared with i.p. administration. Because our experiments were designed to monitor early phase reactions, these results suggest that these catechins, particularly the O-methylated EGCGs, inhibit immediate hypersensitivity reactions by suppressing activation of mast cells. Alternatively, the catechins might antagonize effects of chemical mediators or affect vascular permeability nonspecifically.

Because mast cells are the major effector cells in the early phase of immediate hypersensitivity, we explored the effect of the O-methylated EGCGs on mast cell activation. To this end, BMMC sensitized with IgE were pretreated with these catechins for 30 min before Ag stimulation. As readouts of mast cell activation, we measured histamine release (as a surrogate marker of degranulation), leukotriene release, and cytokine secretion. As shown in Fig. 3,A, both EGCG3″Me and EGCG4″Me inhibited histamine release in a dose-dependent manner. Inhibition by EGCG3″Me and EGCG4″Me at 100 μg/ml reached 70%. Leukotriene release and IL-2 production/secretion were also inhibited by these catechins (Fig. 3, B and C). It is noteworthy that more than control levels of leukotriene release and IL-2 secretion were observed at lower doses (5 or 25 μg/ml) of the O-methylated EGCGs.

To gain insight into the mechanism of mast cell inhibition by O-methylated EGCGs, we investigated their effects on early signaling events. IgE-sensitized BMMC were preincubated with the catechins for 30 min and stimulated with Ag for 3 min in the continued presence of the catechins. Effects on one of the earliest signaling events, tyrosine phosphorylation of the FcεRI β subunit, were examined by immunoprecipitation with anti-FcεRI β Ab, followed by immunoblotting with anti-phosphotyrosine Ab. As shown in Fig. 4,A, up to 50 μg/ml of either EGCG derivative did not substantially affect FcεRI β phosphorylation. Interestingly, however, tyrosine phosphorylation of other cellular proteins, as revealed by immunoblotting of total cell lysates (Fig. 4 B), was more strongly inhibited: tyrosine phosphorylation of several proteins (100-, 85-, 75-, 56-, 53-, 40-, 38-, and 30-kDa proteins) was obviously inhibited by both catechins in a dose-dependent manner.

We next performed in vitro kinase assays on Lyn, Syk, and Btk immunoprecipitated from O-methylated EGCG-pretreated, IgE/Ag-stimulated cells (Fig. 4,C). The results demonstrated that both O-methylated EGCGs inhibit catalytic activity of these critical PTKs: EGCG3″Me inhibited Lyn more preferably (IC₅₀ of 30 μg/ml) than Syk (IC₅₀ of 100 μg/ml). Lyn activity was inhibited almost completely by EGCG3″Me at 100 μg/ml, but substantial levels of Lyn activity remained at 100 μg/ml EGCG4″Me. In contrast, Syk activity was totally abrogated by 100 μg/ml EGCG4″Me, but substantial Syk activity remained at 100 μg/ml EGCG3″Me. Btk was also inhibited by EGCG4″Me with an IC₅₀ of 50 μg/ml (Fig. 4 C). However, Btk activity was not affected by EGCG3″Me. Treatment with 50 or 100 μg/ml O-methylated EGCGs did not change kinetics of activation of these PTKs or tyrosine phosphorylation (data not shown).

We next investigated whether inhibition of PTKs by O-methylated EGCGs is a direct action or not. To this end, we performed kinase assays on purified Lyn, Syk, and Btk proteins in the presence of various concentrations of O-methylated EGCGs (Fig. 5,A). We found that the catalytic activity of all these PTKs is inhibited by both O-methylated EGCGs in a dose-dependent manner. These PTKs were much more sensitive to O-methylated EGCGs in these cell-free kinase assays than when mast cells were treated with O-methylated EGCGs: for instance, IC₅₀ for cell-free inhibition of Lyn by EGCG3″Me was 0.2 μg/ml compared with the IC₅₀ of 30 μg/ml when cells were pretreated with EGCG3″Me. Kinetic analysis of EGCG3″Me on Lyn inhibition showed that inhibition was noncompetitive with ATP (Fig. 5 B); thus, EGCG3″Me is a nonsubstrate inhibitor. The results collectively indicate that O-methylated EGCGs directly inhibit three PTKs critical for mast cell activation.

Subsequent to activation of the PTKs, various signaling pathways are activated upon FcεRI stimulation. Increased intracellular Ca²⁺ concentration levels following PLC-γ activation by Syk and Btk are critical for FcεRI-induced degranulation (46). As expected from PTK inhibition (Figs. 4 and 5) and inhibition of degranulation, leukotriene release, and cytokine production (Fig. 3) by O-methylated EGCGs, both EGCG3″Me and EGCG4″Me inhibited IgE/Ag-induced Ca²⁺ response (Fig. 6).

In light of inhibition of IL-2 production/secretion by O-methylated EGCGs at a high concentration of 100 μg/ml, we examined their effects on the activation of MAPKs (ERKs, JNK, and p38) and Akt. Among these PS/TKs that were shown to be involved in cytokine production and secretion (17, 20, 47, 48), we previously showed that Akt regulates production of IL-2 and TNF-α (20) and JNK is believed to control the transcription of the IL-2 gene through c-Jun phosphorylation (49). BMMC pretreated with O-methylated EGCGs were stimulated with Ag, and cell lysates were analyzed by immunoblotting with Abs that react with the activated forms of these kinases. As shown in Fig. 7,A, EGCG3″Me inhibited Akt and JNK1, but not ERK1, ERK2, or p38. Effects of EGCG4″Me were more complex: it inhibited Akt at 25 mg/ml; it inhibited JNK activity less potently than EGCG3″Me, but p38 activity was rather enhanced by EGCG4″Me. Potentially related to the increased IL-2 production/secretion in cells treated with low concentrations of O-methylated EGCGs (Fig. 3,C), increased phosphorylations of some of these kinases were observed in cells treated with low concentrations of O-methylated EGCGs: for instance, ERK1, ERK2, and p38 in EGCG3″Me-treated cells and p38 in EGCG4″Me-treated cells (Fig. 7 A).

Downstream of these serine/threonine kinases, transcription factors are activated. Such activation of transcription factors was monitored by tranfecting BMMC with IL-2/luc, followed by pretreatment of the cells with O-methylated EGCGs and stimulation with IgE and Ag (Fig. 7 C). Again, consistent with the inhibition of IL-2 production/secretion with O-methylated EGCGs at high concentrations, IL-2/luc activity was inhibited by both O-methylated EGCGs at 100 μg/ml. Lower concentrations of O-methylated EGCGs rather enhanced the IL-2/luc activity.

Given the broad specificity of PTK inhibition by O-methylated EGCGs, we explored whether these catechins might directly affect the catalytic activity of serine/threonine kinases as well. For this purpose, effects of O-methylated EGCGs were measured in in vitro kinase assays using pure preparations of ERK2, JNK1, and PKCβII (Fig. 7 B). All of these kinases were inhibited by either EGCG in a dose-dependent manner. However, their IC₅₀ were at least 10- to 50-fold higher than those for PTK inhibitions.

In this study, we found that two tea-derived O-methylated catechins, EGCG3″Me and EGCG4″Me, inhibit PCA reactions with higher potency than their nonmethylated form of catechin does (Fig. 2). Consistent with such in vivo effects, these O-methylated catechins inhibited all tested aspects (i.e., degranulation, leukotriene release, and cytokine production/secretion) of mast cell activation induced by FcεRI cross-linking (Fig. 3). Mast cell inhibition could be explained at least in part by their activity to directly inhibit several PTKs (Figs. 4 and 5). Although the O-methylated catechins can directly inhibit several PS/TKs, their potencies against the PS/TKs are 10- to 50-fold lower than those against the PTKs. Therefore, it is likely that the inhibition of several PS/TKs observed in mast cells treated with the O-methylated catechins is due to inhibition of upstream PTKs. Effects of O-methylated EGCGs on mast cell signal transduction appear more complex than simply inhibiting PTKs: although Lyn kinase activity was inhibited by these agents, FcεRI β phosphorylation that is mediated by Lyn was not substantially affected by them. They rather increased the activity of PS/TKs, particularly p38, at low concentrations. These results suggest that O-methylated EGCGs influence other signaling molecules, e.g., a protein phosphatase(s). These catechins are unique in their broad specificity among other known mast cell-inhibitory agents. Other PTK inhibitors that were shown to inhibit mast cell activation exhibit more stringent selectivity: for instance, ER-27319 (50) and terreic acid (51) inhibit FcεRI-induced mast cell activation by specifically inhibiting Syk and Btk, respectively. This broad specificity might limit the usefulness of these catechins as an antiallergic drug. However, drinking tea containing these compounds has not imposed apparent risks on human health. Therefore, our results may be a foundation for encouraging allergic patients to drink tea rich in these O-methylated catechins. Indeed, a clinical trial to treat allergic cedar pollinosis patients with tea rich in O-methylated catechins is ongoing, and promising results have been obtained by using a protocol of drinking two cups of tea daily (data not shown). Another potential use could be topical treatment of allergic diseases with these O-methylated catechins: for example, washing with these catechins or tea extracts, nasal cavity of allergic rhinitis patients and skin lesions of atopic dermatitis patients might have beneficial effects.

Differences of one to two orders of magnitude in effective concentrations were found with the O-methylated catechins in PTK-inhibitory potency between when mast cells were incubated with the catechins (Fig. 4,C) and when the purified PTKs were directly incubated with them (Fig. 5 A). Thus, cell-free IC₅₀ of 0.2–5 μg/ml to inhibit PTKs were reflected with in vitro IC₅₀ of 30–100 μg/ml to inhibit histamine release, leukotriene release, and IL-2 production/secretion. This strongly suggests that these O-methylated catechins penetrate the mast cell plasma membrane with very poor efficiency. Alternatively, these catechins might lose their mast cell-inhibiting activity by being modified or degraded to inert molecules within or even outside mast cells. Indeed, O-methylated EGCGs can be epimerized, methylated, and conjugated with glucuronic acid and sulfate (M. Maeda-Yamamoto, unpublished data). Bioactivities of these EGCG derivatives, including effects on mast cell activation, remain to be explored. Importantly, in vivo IC₅₀ of 50 mg/kg EGCG3″Me by i.v. administration to inhibit PCA reactions correlate reasonably well with the in vitro IC₅₀ of 30 μg/ml to inhibit Lyn kinase activity. Thus, in our preliminary experiments, serum concentrations of 1–25 μg/ml were achieved by i.v. injection of 50 mg/kg EGCG3″Me into mice (data not shown). Therefore, the major target of the O-methylated catechins in the acute phase of in vivo allergic reactions was considered to be mast cells, although other cell types should be affected in slower aspects of allergic reactions.

We thank Dr. Masaniko Ikeda for performing Ca influx experiments, Dr. Keiko Kawamoto for technical advice, and Nahomi Matsuda, Naoko Suzuki, Nobufumi Tahara, Masayoshi Yoshimura, and Makiko Suzuki for technical assistance. This is Publication 576 from the La Jolla Institute for Allergy and Immunology.