Green tea catechins, especially (−)-epigallocatechin gallate (EGCG), have been proposed as a chemopreventative for obesity, diabetes, cancer, and cardiovascular diseases. However, relatively little is known about the mechanism of the action of EGCG on fat cell function. This study was designed to investigate the pathways of EGCG's modulation of the mitogenesis of 3T3-L1 preadipocytes. Preadipocyte proliferation as indicated by an increased number of cells and greater incorporation of bromodeoxyuridine (BrdU) was inhibited by EGCG in dose-, time-, and growth phase-dependent manners. Also, EGCG dose and time dependently decreased levels of phospho-ERK1/2, Cdk2, and cyclin D1 proteins, reduced Cdk2 activity, and increased levels of G0/G1 growth arrest, p21waf/cip, and p27kip1, but not p18ink, proteins and their associations to Cdk2. However, neither MEK1, ERK1/2, p38 MAPK, phospho-p38, JNK, nor phospho-JNK was changed. Increased phospho-ERK1/2 content and Cdk2 activity, respectively, via the transfection of MEK1 and Cdk2 cDNA into preadipocytes prevented EGCG from reducing cell numbers. These data demonstrate the ERK- and Cdk2-dependent antimitogenic effects of EGCG. Moreover, EGCG was more effective than epicatechin, epicatechin gallate, and epigallocatechin in changing the mitogenic signals. The signal of EGCG in reducing growth of 3T3-L1 preadipocytes differed from that of 3T3 fibroblasts. Results of this study may relate to the mechanism by which EGCG modulates body weight.
- 3T3-L1 preadipocyte
- mitogen-activated protein kinase
- cyclin-dependent kinase
obesity is a common disease associated with risks of cancer, diabetes, hypertension, and cardiovascular disease (21). Maybe for this reason, estimates of the economic costs, prevalence, morbidity, and mortality associated with more-modest degrees of being overweight and obese are rising (2, 21, 44). The development of obesity is characterized by increased number of fat cells and their lipids due to the processes of so-called mitogenesis and differentiation, which are regulated by genetic, endocrine, metabolic, neurological, pharmacological, environmental, and nutritional factors (21). Accordingly, an understanding of the mechanism through which a particular nutrient affects the mitogenesis of preadipocytes and their differentiation to adipocytes would help prevent the initiation and progression of obesity and its associated diseases in humans.
Green tea catechins (GTCs) are polyphenolic flavonoids once called vitamin P (34). Since the discoveries that they have unique chemical structures (Fig. 1) and are major ingredients of unfermented tea (24, 33), they have been found to possess widespread biological functions and health benefits (1, 24, 25, 28, 45). In vivo, GTCs, especially (−)-epigallocatechin gallate (EGCG; Fig. 1), lower the incidence of cancers (1, 24, 25, 28, 45), collagen-induced arthritis (16), oxidative stress-induced neurodegenerative diseases (27), and streptozotocin-induced diabetes (38). Also, EGCG can reduce body weight and body fat (18). In support of this antiobese effect of EGCG, other in vivo data have shown that EGCG or EGCG-containing green tea extract reduces food uptake, lipid absorption, and blood triglyceride, cholesterol, and leptin levels as well as stimulating energy expenditure, fat oxidation, HDL levels, and fecal lipid excretion (10, 18, 19, 24). These in vivo observations may be explained by in vitro findings that EGCG and caffeine synergistically with norepinephrine stimulate the thermogenesis of brown adipose tissue (11), that EGCG regulates various enzymes related to lipid anabolism and catabolism, such as acetyl-CoA carboxylase, fatty acid synthase, pancreatic lipase, gastric lipase, and lipoxygenase (24, 48), that EGCG is a potent prooxidant and antioxidant (24, 42, 43), and that EGCG reduced serum- or insulin-induced increases in cell numbers and the triacylglycerol content during a 9-day period of differentiation (19). These in vivo and in vitro observations suggest that green tea EGCG appears to modulate the mitogenic, endocrine, and metabolic functions of fat cells.
Despite the importance of EGCG, relatively little is known about the mechanism of its action in regulating the mitogenesis of preadipocytes and their differentiation to adipocytes. The fact that the EGCG receptor, the so-called laminin receptor, discovered in cancer cells (39), has not been identified in fat cells and the fact that fat cells have different isoforms of laminins (30) have also caused much controversy. Accordingly, a thorough examination of the signal element through which EGCG executes its modulation of preadipocyte mitogenesis should help clarify these observations. MAPK and Cdk are essential mitogenic signal transducers in most cells (4, 17, 31), including 3T3-L1 preadipocytes (5, 40), and they have been proposed as being important signals of EGCG in modulating the growth of cancer cells based on various studies (1, 22–28, 43, 46). Further studies are required to determine whether any of them are responsible for EGCG signaling in preadipocyte antimitogenesis.
The present study was designed to understand the mechanism of how EGCG acts in reducing the number of 3T3-L1 preadipocytes as they grow. Our specific aim was to investigate whether EGCG-regulated preadipocyte proliferation is dependent on the ERK MAPK and Cdk2 pathways. The mechanistic results of this study may have possible utility in the treatment of obesity using this compound.
MATERIALS AND METHODS
Green tea EGCG and other catechins (>98% pure) were isolated from green tea (Camellia sinensis) in our laboratory as described previously (18). Catechins were dissolved in 0.1% DMSO and sterile medium for cell treatment. Other materials (i.e., PD98059) were purchased from Sigma (St. Louis, MO) unless otherwise mentioned. Penicillin-streptomycin, DMEM, FBS, trypsin, agarose, and a 1-kb plus DNA ladder marker were purchased from GIBCO-BRL Life Technologies (New York, NY). The 3′-RACE system, TRIzol, Taq polymerase, and a BenchMark prestained protein ladder were purchased from Invitrogen Life Science Technologies (Carlsbad, CA). Except for the phospho-histone H1 and cyclin D1 antibodies, which were obtained from Calbiochem and Merck (Darmstadt, Germany), all other antibodies (i.e., ERK-1, ERK-2, phospho-ERKs, p21waf1/cip, p27Kip1, p18INK, donkey anti-rabbit IgG-horseradish peroxidase) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
According to a published method (19), 3T3 and 3T3-L1 cells (American Type Culture Collection; Manassas, VA) were grown in DMEM (pH 7.4) containing 10% FBS, 100 U/ml of penicillin, and 100 μg/ml streptomycin (GIBCO-BRL) in a humidified atmosphere of 95% air-5% CO2 at 37°C. Medium (10 ml) was replaced every 2 days. Because serum components contain the factors for facilitating 3T3-L1 differentiation from preadipocytes to adipocytes when they are confluent, these cells were subcultured before reaching confluency. Confluent 3T3 were subcultured.
Growth inhibition experiments.
3T3 and 3T3-L1 cells (15,000∼20,000 cells/cm2) were plated in triplicate wells of a 12-well plate. To determine whether a dose- or growth phase-dependent effect of GTCs on the growth of 3T3-L1 preadipocytes exists, we treated different growth phases (days 1∼6 with day 1 being the day of cell inoculum) of cells with epicatechin (EC), epigallocatechin (EGC), epicatechin gallate (ECG), or EGCG at various concentrations (0∼400 μM) for indicated time periods. After a particular time course of incubation, cells were trypsinized and counted with a hemocytometer using the 0.4% trypan blue exclusion method. To measure cellular proliferation, we modified the method reported by Vollenweider et al. (41) and used a commercially available bromodeoxyuridine (BrdU) ELISA kit (Roche Applied Science; Mannheim, Germany) as follows. 3T3-L1 cells (2,000 cells/well) were plated into a 96-well microplate (tissue culture grade) with 100 μl DMEM supplemented with 10% FBS. After 24 h were allowed for attachment, cells were starved with serum-free DMEM for 36 h, which was then replaced with fresh DMEM containing 10% FBS, the thymidine analog BrdU (10 μM), and each GTC (at the indicated concentrations) for 4 h at 37°C. This allowed BrdU to be incorporated into newly synthesized DNA of dividing cells during their S phase. After incubation, residual cells were washed with 10 mM PBS and then collected by centrifugation at 1,500 rpm for 5 min. Cell pellets were dried at 60°C for 1 h, fixed with 200 μl FixDenat solution/well for 30 min at 15∼25°C, probed with mouse anti-BrdU-POD for 1 h, and visualized with the addition of 100 μl 3,3,5,5-tetramethylbenzidine substrate to each well for 5 min for color development. An aliquot of 100 μl of 1 N H2SO4 was added to stop the reaction of each well for the 1-min incubation on a 300-rpm shaker. The absorbance was read at 450 nm using a MRX microtiter plate reader (Dynatech Laboratories; Chantilly, VA). Culture medium alone and cells incubated with anti-BrdU-POD in the absence of BrdU were used as blank controls for nonspecific binding.
MAPK inhibitor treatment.
3T3-L1 preadipocytes, cultured in DMEM containing 10% FBS for 2 days and synchronized with serum-free DMEM for 1 day, were pretreated for 2 h with EGCG (20∼50 μM), PD98059 (50 μM) (12), and/or U0126 (10 μM) (13). These compounds were dissolved in 100% DMSO (at a final concentration of 0.1%) and then added to fresh DMEM containing 10% FBS during the experiment. After 4 h of incubation, protein amounts of ERK1/2 and phospho-ERK1/2 were measured by Western blot analysis while the number of cells was examined by the trypan blue exclusion method after 48 h of incubation.
Cdk2 plasmid constructs.
cDNA encoding wild-type murine Cdk2 (Cdk2+/+) and the dominant negative form of murine Cdk2 (dnCdk2; with a mutation of Asp145 in Cdk2 to Asn145), as described by Heuvel and Harlow (17), was amplified by RT-PCR. Total RNA was isolated from 3T3-L1 preadipocytes with the TRIzol kit, and cDNA was then synthesized from equal amounts of RNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen; Carlsbad, CA). The forward and reverse primers for obtaining Cdk2+/+ cDNA were 5′-ATGGAGAACTTTCAAAAGGTG-3′ and 5′-TCAGAGCCGAAGGTGGGGGC-3′, respectively. To obtain dnCdk2 cDNA, four primers were needed to introduce a site-specific mutation by overlap exclusion (42), and they were 5′-ATGGAGAACTTTCAAAAGGTG-3′, 5′-TCAGAGCCGAAGGTGGGGGC-3′, 5′-GTCCAAAGTTTGCCAGCTTG-3′, and 5′-CAAGCTGGCAAACTTTGGAC-3′. PCR was performed under the following conditions: an initial denaturing cycle at 94°C for 3 min, followed by 30 cycles of amplification consisting of denaturation at 94°C for 45 s, annealing at 53°C for 30 s, and extension at 72°C for 90 s. A final extension at 72°C for 10 min was added after the last cycle. The PCR product was run on a 1.5% (wt/vol) agarose gel using 40 mM Tris-acetate buffer (pH 8.0) containing 1 mM EDTA and was visualized using 0.5 μg/ml ethidium bromide. The Cdk2+/+ and dnCdk2 products (predicted to be about 900 bp) were cloned to the pTargetT vector (Promega) as described by Sambrook and Russell (36). Before their transfection into the preadipocytes, they were verified with nucleotide sequencing performed at the Institute of Biomedical Sciences, Academia Sinica, Taiwan. The molecular weights of overexpressed Cdk2+/+ and dnCdk2 proteins were verified to be about 34 kDa by Western blot analysis, and their activities were also confirmed to catalyze the phosphorylation of the histone H1 substrate (20).
Cell transfection and overexpression experiments.
We modified the methods reported by Yang et al. (46) to perform our transfection experiments. In some experiments, 3T3-L1 preadipocytes were transiently (24 h) transfected with 15 μg of either the pcDNA3.1 plasmid, pBabe puro vector, pcDNA3.1-MKK6EE (the constitutively active MKK6 mutant designated MKK6EE, a gift from Dr. S. L. Chen), pCMV-MEKK1 (a constitutively active MEKK1-activating JNK kinase, a gift from Dr. S. L. Chen), pBabe puro-MEK1, or pBabe puro-MEK1S217E/S221E (a constitutively active MEK1 mutant designated MEK1EE; a gift from Dr. J. J. Yang). The overexpressed preadipocytes were cotransfected with pSV-β-galactosidase cDNA whose transfection efficiency was determined by analyzing the β-galactosidase activity. There were insignificant changes in β-galactosidase production from all transfected cells (data not shown). Activities of MEK1, MEKK1, and MKK6 proteins were assessed as measured by Western blot analysis of changes in the amounts of phospho-ERK1/2, phospho-JNK, and phospho-p38, respectively. In other experiments, 3T3-L1 preadipocytes in a 10-cm plate were stably transfected with 15 μg of the pTargetT vector constructed with Cdk2+/+ and dnCdk2 cDNAs. Stable clones were selected with 1 mg/ml of the antibiotic G-418 (BD Biosci Clontech Lab; Palo Alto, CA). Amounts and activities of Cdk2+/+ and dnCdk2 proteins expressed in the stable clones were later determined by immunoblotting and immunoprecipitation, respectively. A 45-μl volume of the TransFast transfection reagent was used during the stable and transient transfections and was comprised of the synthetic cationic lipid (+)-N,N-bis(2-hydroxyethyl)-N-methyl-N-[2,3-di(tetradecanoyloxy)propyl] ammonium iodide and the neutral lipid l-dioleoyl-phosphatidylethanolamine.
Cells transfected with the vehicle or the recombinant plasmid containing MEK1, MEK1EE, or MKK6EE cDNAs were incubated with or without 50 μM EGCG for 48 h. After incubation, the cell number of preadipocytes was examined by trypan blue dye exclusion. Unless otherwise noted, transfected cells with the vehicle, Cdk2+/+ cDNA, and dnCdk2 cDNA were incubated with and without 20∼100 μM EGCG for the indicated time periods.
Flow cytometric analysis.
Changes in the kinetics of the cell cycle were analyzed by flow cytometry as described by Kokontis et al. (20). 3T3-L1 cells (15,000∼20,000 cells/cm2) transfected with and without the vehicle, Cdk2+/+, and dnCdk2 plasmids were plated in a 10-cm dish containing 10 ml DMEM supplemented with 10% FBS. One day after inoculation, the medium was replaced with serum-free DMEM to obtain the homogenous cell population. After a 1-day starvation, cells were treated with fresh medium in the presence and absence of EC, ECG, EGC, or EGCG at various concentrations and time periods. The harvested cell pellets were fixed in 70% ethanol (dissolved in 10 mM PBS, pH 7.4) and stored at −20°C until later analysis. Cell pellets were washed with 10 mM cold PBS (pH 7.4), incubated at 37°C for 30 min with 200 μg/ml RNase A (Sigma), and then stained with 4 μg/ml propidium iodide (Sigma) in PBS containing 1% Triton X-100 (Sigma). Cell cycle profiles and distributions were determined by flow cytometric analysis of 104 cells using the CELLQuest program on a FACS Calibur flow cytometer (Becton-Dickinson; San Jose, CA). Clumped cells were excluded from the cell cycle distribution analysis by gating.
Cdk2 protein was immunoprecipitated according to the method described by Kokontis et al. (20). After experimental treatments, preadipocytes were washed twice in 10 mM PBS and then lysed in 1 ml buffer A [20 mM Tris·HCl (pH 7.6), 1 mM EDTA, 1 mM Na3VO4, 0.2% Triton X-100, and 1 mM PMSF]. The lysate was agitated for 15 min at 4°C and then centrifuged at 14,000 rpm for 10 min to collect the supernatant. The protein content of the supernatant was determined in duplicate by the dye-binding method (6) using a Bio-Rad (Richmond, CA) microplate reader and BSA (Sigma) as a standard. An aliquot of the supernatant (1 mg protein) was preincubated for 1 h at 4°C with Cdk2 antibody or preimmunized normal rabbit serum (NRS; as the control) for 1 h at room temperature or overnight at 4°C. The mixture was incubated with 20 μl protein A-agarose (Santa Cruz Biotechnology) overnight at 4°C. The total amounts of Cdk2, p18, p21, and p27 in the immunoprecipitates were measured by Western blot analysis with each antibody. After normalization to the Cdk2 protein, the amounts of each cyclin-dependent kinase inhibitor (CKI) protein were expressed as a percentage of the control, and changes in their bindings to Cdk2 were indicated. Data obtained from NRS were not shown because of the insignificant changes.
Western blot analysis.
Western immunoblot analysis was performed on supernatant fractions of preadipocytes as described by Kokontis et al. (20). An aliquot of 50 μg of supernatant protein was separated by 12% SDS-PAGE with 2× gel-loading buffer [100 mM Tris·HCl (pH 6.8), 4% SDS, 20% glycerol, 0.2% bromophenol blue, and 10% β-mercaptoethanol] and then blotted onto Immobilon-NC transfer membranes (Millipore; Bedford, MA). The immunoblots were blocked for 1 h at room temperature with 10 mM PBS containing 0.1% Tween 20 (PBST) and 5% defatted milk. After the immunoblots were washed with PBST, immunoblot analyses were performed. All primary antibodies (ERK-1, ERK-2, phospho-ERKs, MEK1, p38, phospho-p38, JNK, phospho-JNK, β-actin, Cdc2, Cdk2, p18, p21, p27, cyclin D1, and phospho-histone H1 antisera) were used at a dilution of 1:1,000 (∼0.2 μg/ml). Donkey anti-rabbit IgG, donkey anti-mouse IgG, donkey anti-goat IgG, or goat anti-guinea pig IgG conjugated with horseradish peroxidase was used as the secondary antibodies at a dilution of 1:2,000 (∼0.2 μg/ml). The immunoblots were visualized using the Western Lightning chemiluminescence reagent plus kit (Perkin-Elmer Life Science; Boston, MA) for 3 min followed by exposure to Fuji film for 2∼3 min. Blots were quantified using the FX Pro Plus Molecular Imager (Bio-Rad Laboratories). After normalization to β-actin protein, levels of these intracellular proteins were expressed as a percentage of the control unless otherwise noted.
Cdk2 activity assay.
After immunoprecipitation, Cdk2 activity was determined as modified from the method of Kokontis et al. (20). Assays were performed at 37°C for 30 min in a final volume of 25 μl. The final substrate mixture per tube contained 20 mM ATP, 10 μg histone H1, 20 mM Tris·HCl buffer (pH 7.5), 4 mM MgCl2, 0.8 mM EGTA, and Cdk2 immunoprecipitates. The reaction was terminated by the addition of 50 μl SDS-PAGE sample buffer as described above. We removed a 20-μl aliquot of the solution to load onto SDS-PAGE and used anti-phospho-histone H1 as the primary antibody to immunoblot the samples as described above. Changes in the amounts of phospho-histone H1, after normalization to the immunoprecipitated Cdk2 protein, indicated alterations in Cdk2 activity.
Data are expressed as means ± SE unless otherwise noted. An unpaired Student's t-test was used to examine differences between the control and EGCG-treated groups. One-way ANOVA followed by the Student-Newman-Keuls multiple-range test was used to examine differences among multiple groups. Differences were considered significant at P < 0.05. All statistical analyses were performed using SigmaStat (Jandel Scientific; Palo Alto, CA).
Mitogenesis of 3T3-L1 preadipocytes was inhibited by GTCs.
There were significant variations in the measured cell number when the distinct phases of 3T3-L1 preadipocytes were incubated with different concentrations of EC, ECG, EGC, or EGCG (Fig. 2). In general, EGCG was more effective in reducing the cell number than EC, ECG, or EGC. Latent (days 1–2), log-phase (days 3–4), and confluent (days 5–6) preadipocytes had different sensitivities to individual GTCs depending on the dosage and duration of treatment (Fig. 2, A–I). For example, the IC50 values of EC, EGC, and ECG in day 3 preadipocytes were all >50 μM during the 72 h of treatment except for 10–20 μM EGCG at 24∼72 h (Fig. 2, D–F). In day 2, day 4, and day 6 preadipocytes, the IC50 values of EC, EGC, and ECG at 48 h were all >200 μM except for 50–300 μM EGCG (data not shown).
The determination of whether the reduction in cell number induced by GTCs was due to their influences on the cell viability of 3T3-L1 preadipocytes was examined by the trypan blue exclusion method. High doses of 100∼400 μM EGCG decreased the cell viability of day 3 preadipocytes by 15∼30%. However, at concentrations of green tea EC, ECG, EGC, and EGCG below 50 μM, the cell viability of these log-phase cells still remained at 90∼100% (Fig. 3A). Similar effects of these catechins on the latent and confluent preadipocytes were observed (data not shown). A further BrdU incorporation test on log-phase preadipocytes showed that EGCG inhibited cell proliferation in a dose-dependent manner (Fig. 3B). The IC50 of EGCG was about 50 μM. At this concentration, EGCG was more effective than EC, ECG, and EGC at inhibiting preadipocyte proliferation (Fig. 3B). Moreover, flow cytometric analysis indicated that EGCG, but not EC, ECG, or EGC, changed the percentages in the four different phases of the cell cycle of preadipocytes (Fig. 3C). EGCG arrested preadipocytes in the G0/G1 phase of the cell cycle.
Antimitogenic effect of EGCG depends on ERK pathways.
To elucidate the signaling through which EGCG acts on the mitogenesis of 3T3-L1 preadipocytes, we examined whether the effect of EGCG was dependent on the ERK and MEK1 pathways. In log-phase preadipocytes, total protein levels of neither ERK1/2 nor MEK1 were significantly changed by 50 μM EGCG over the 24-h course (Fig. 4A). However, the amounts of phospho-ERK-1 and phospho-ERK-2 were significantly reduced by EGCG. The effect of EGCG was time dependent (Fig. 4A) and dose dependent (Fig. 4B); for a given 4-h treatment, the IC50 of EGCG was about 50 μM (Fig. 4B). In contrast, EGCG increased the levels of phospho-ERKs in latent preadipocytes (day 1), whereas it had no effect on their amounts in confluent (day 5) preadipocytes (Fig. 4C).
To determine whether EGCG's effect in reducing MEK1 activity is selective, changes in the activities of MEK1, MKK3/6, and MKK4/7 were assessed by changing the amounts of the phosphorylated form of their own protein substrates, ERK-1 and ERK-2, p38 MAPK, and JNK MAPK (Fig. 5A), respectively, after EGCG treatment. EGCG at 50 μM for 4 h did not significantly alter the amounts of ERK1/2, p38, phospho-p38, JNK, or phospho-JNK but did significantly decrease the amounts of phospho-ERKs by about 60%.
We further enhanced MEK1 activity by overexpressing MEK1 or its constitutively active mutant, MEK1EE, in preadipocytes to demonstrate whether overexpression of such kinases could prevent EGCG-induced decreases in cell numbers and MEK1 activity (Fig. 5). We observed that preadipocytes overexpressed with either MEK1 or MEK1EE revealed no significant changes in their cell number (Fig. 5B) between the presence and absence of 50 μM EGCG for 48 h (Fig. 5B). In contrast, the numbers of preadipocytes that were transfected with the empty vector, MKK6EE (Fig. 5B), or MEKK1 (data not shown) were still significantly reduced by EGCG. Increased expression and activity of MEK1 protein in MEK1- or MEK1EE-transfected preadipocytes are confirmed in Fig. 5C.
Differences of EGCG with other specific MEK1 inhibitors.
On the basis of IC50 values of EGCG (20∼50 μM), PD98059 (50 μM), and U0126 (10 μM) for inhibiting preadipocyte proliferation or reducing MEK1 activity, we tested whether the inhibitory effects of EGCG on preadipocyte growth and MEK1 activity differed from those of other specific inhibitors of the MEK1 protein (Fig. 6). Proliferation of preadipocytes was inhibited by either EGCG, PD98059, or U0126 alone. In addition, their growth was inhibited by a combination of EGCG with either PD98059 or U0126 (Fig. 6A). Total amounts of neither of the proteins, ERK-1 or ERK-2, changed after treatment with each inhibitor (Fig. 6B). Interestingly, EGCG tended to be additive with either PD98059 or U0126 in decreasing MEK1 activity, as indicated by the reduced amounts of phospho-ERK1/2 proteins.
Effect of EGCG on the insulin-like growth factor-induced increases in the activity of MEK1.
Insulin-like growth factors (IGFs) have been implicated in the proliferation of preadipocytes through an increased activity of MEK1. The possibility that the EGCG-induced reduction in MEK1 activity is related to its decreasing the IGF-induced stimulation in MEK1 activity was also examined (Fig. 7). EGCG dose dependently reduced MEK1 activity, as indicated by decreased phospho-ERK1/2 proteins, stimulated by either IGF-I (1 nM; Fig. 7A) or IGF-II (10 nM; Fig. 7B).
Differences in MEK1 activity by EGCG compared with other GTCs.
When 50 μM of the four GTCs of EC, EGC, ECG, and EGCG were individually added to preadipocytes for 4 h, EGCG was the most effective in reducing the amounts of phospho-ERK1/2 proteins (Fig. 8). The total amounts of ERK-1, ERK-2, and MEK1 proteins did not change after treatment with each catechin.
Effects of EGCG on preadipocyte mitogenesis and MEK1 activity varied with cell types.
At the doses of 20 and 100 μM, EGCG significantly reduced the numbers of 3T3-L1 and 3T3 cells (Fig. 9A) to a greater extent than it did human KB oral cancer cells (data not shown). A change in the protein expressions of ERK-1 and ERK-2 by EGCG was not observed in any of the three types of cells (Fig. 9B). However, protein amounts of phospho-ERK1/2 in both 3T3 and 3T3-L1 cells were dose-dependently reduced by EGCG.
Effect of EGCG on Cdk of preadipocytes differed from those of other GTCs and from other cell types.
Cdk2 is well known as a downstream protein regulating cell mitogenesis via controlling the procession of the four different phases of the cell cycle, thereby resulting in changes in cell numbers. Accordingly, we investigated whether EGCG alters the protein expression and activity of Cdk2 (Fig. 10). Treatment with EGCG for 48 h reduced the level (Fig. 10A) and activity (Fig. 10B) of Cdk2 protein in a dose-dependent manner. Although EGCG did not alter Cdk2 protein expression at 4 and 24 h of incubation, it was found to dose dependently decrease the activity of Cdk2. In contrast to EGCG, neither EC, EGC, nor ECG (data not shown) significantly altered Cdk2 protein levels after 48 h of treatment. This indicates the existence of catechin-specific effects of green tea.
In a further examination of whether the long-term (48 h) effects of EGCG on Cdks vary with cell type, we assessed changes in the protein expression of Cdk2 and Cdc2 (also called Cdk1) by 3T3-L1 and 3T3 cells (Fig. 10C). At a low dose of 20 μM, EGCG induced significant decreases in the protein expression of Cdk2, but not Cdc2, by 3T3-L1 preadipocytes. In contrast, EGCG tended to increase Cdk2 protein expression in 3T3 cells (Fig. 10C) but decreased Cdc2 protein expression in 3T3 cells. At a high dose of 100 μM, EGCG reduced protein levels of Cdk2 and Cdc2 in 3T3-L1 cells, whereas in 3T3 cells, it enhanced Cdk2 protein expression and reduced Cdc2 protein expression.
Overexpression of Cdk2+/+ with EGCG-inhibited growth of preadipocytes.
In a further demonstration of whether Cdk2 protein is required for the effect of EGCG on 3T3-L1 preadipocytes, we stably cloned preadipocytes that overexpressed Cdk2+/+ and examined whether overexpression of Cdk2+/+ could prevent EGCG-induced growth inhibition of preadipocytes after treatment with 20∼100 μM EGCG for 5 days (Fig. 11). As indicated by the increased cell number, Cdk2+/+-transfected preadipocytes grew more rapidly than vehicle-transfected cells during the 5-day incubation. The growth of the former cells, but not the latter cells, was not changed by EGCG (Fig. 11A). However, overexpression of dnCdk2 with a mutation of Asp145 in Cdk2 to Asn145 slowed down growth of preadipocytes after a 5-day incubation no matter the presence and absence of EGCG (data not shown). When examined at 24 and 48 h of EGCG treatment, EGCG's alteration of the four different phases of the cell cycle (Fig. 11B), EGCG's reduction of levels (Fig. 11C) and activity (Fig. 11D) of Cdk2 protein and the EGCG-altered binding of Cdk2 to p18, p21, and p27 (Fig. 11E) were still observed in vehicle-transfected preadipocytes but not in Cdk2+/+-transfected preadipocytes.
Effects of EGCG on CKIs.
Because Cdk2 activity is negatively regulated by CKIs, we investigated whether EGCG changes the protein expressions of CKIs, such as p21waf/cip, p27Kip1, and p18INK, in 3T3-L1 preadipocytes as well as the binding of Cdk2 to these proteins (Fig. 12). Following 4 h of EGCG incubation, the protein levels of p21 or p27 were not significantly altered, even though up to 100 μM EGCG was used (Fig. 12A). However, this catechin significantly induced increases in the protein levels of p21 and p27, but not p18, after 24 and 48 h of treatment. Although EGCG did not alter the protein expression of p21 and p27 at 4 h, it dose dependently increased their bindings to Cdk2 (Fig. 12B). EGCG also increased the binding of p21 and p27, but not p18, to Cdk2 at 24 and 48 h after treatment. With 24- and 48-h incubations, neither EC, ECG, nor EGC significantly changed the protein expressions of p18, p21, or p27 (data not shown). There was a trend of EGCG increasing protein expression of p27 in 3T3 fibroblast and 3T3-L1 preadipocytes after 48 h of treatment (Fig. 12C).
Effect of GTCs on cyclin D1 protein expression.
It is well known that cyclin D protein is one of the G1 cyclins that controls the G1/S transition of the cell cycle. Accordingly, whether EGCG changes the protein expression of cyclin D1 was assessed after cells were treated with EGCG for 4∼48 h (Fig. 13). We observed that EGCG reduced the total amount of cyclin D1 protein in a dose-dependent manner after 24 (Fig. 13B) and 48 h (Fig. 13C) but not after 4 h (Fig. 13A) of treatment. Neither EC, ECG, nor EGC at 20∼100 μM significantly reduced cyclin D1 protein expression after 24 h of treatment. However, at a high dose of 100 μM, ECG significantly reduced cyclin D1 protein levels after 48 h of treatment.
Green tea EGCG has been proposed as an obesity chemopreventative and a fat cell modulator based on various laboratory studies (24). Unlike a preliminary report (19), this study not only details the reductive effect of EGCG on the number of preadipocytes with dose-, time-, and growth phase-dependent manners but also provides certain in-depth understanding of the mechanism of EGCG's action in the regulation of mitogenesis of preadipocytes.
The observed decrease in the number of preadipocytes by EGCG could be attributable to its inhibition of cell mitogenesis. This is supported by decreased BrdU incorporation, a measure of DNA replication, and by increased G0/G1 growth arrest of the cell cycle when preadipocytes were incubated with EGCG. However, the observed decreases in the cell number by high doses (100∼400 μM) of EGCG could also be explained by its induction of cell apoptosis. This is evident by the fact that such high doses of EGCG reduced the cell viability of preadipocytes by 15∼30%, induced the appearance of DNA fragmentation (data not shown), and increased the activity of the caspase-3 protein (data not shown), an apoptotic enzyme. Taken together, green tea EGCG may act at different concentrations in regulating mitogenesis and apoptosis of 3T3-L1 preadipocytes. This contention is similar to the reported dose-dependent effect of EGCG on neuroblastoma cells (22, 26–27, 42). However, the possibility still remains that EGCG acts at 10∼400 μM to induce antimitogenesis and apoptosis of 3T3-L1 preadipocytes via oxidative stress as reported for hepatoma (42) and neuroblastoma cells (22, 26, 43).
Antimitogenic effect of EGCG on preadipocytes depends on ERK pathway.
To our knowledge, the MAPK family is an essential part of the signal transduction machinery in signal transmissions from cell surface receptors and environmental stimulation, and it contains three major MAPK subfamilies: ERK, p38, and JNK (4, 31). They have been proposed to serve as signal elements in several types of cells through which EGCG may regulate cell growth (1, 3, 8, 24–25, 35, 45) and found to modulate the mitogenic and adipogenic signalings of IGF-I in 3T3-L1 preadipocytes (5). We observed herein that acute (4 h) exposure to EGCG induced a decrease in phosphorylated ERK1/2 in 3T3-L1 preadipocytes but did not alter the total levels of MEK1, ERK-1, ERK-2, p38, phospho-p38, JNK, or phospho-JNK. This suggests that EGCG acts on a specific type of MAPK, especially in the ERK MAPK family. This contention is also partially supported by the fact that chronic (24 or 48 h) exposure to EGCG induced a decrease in the phosphorylated ERK1/2 of preadipocytes, although it did not alter total levels of MEK1 or ERK1/2 proteins. In addition, transient amplification of phospho-ERK1/2 content by transfecting MEK1 cDNA or its active mutant cDNA to 3T3-L1 preadipocytes prevented EGCG-induced decreases in their cell number. The total levels of MEK1 protein in vehicle-, MEK1-, or MEK1EE-transfected preadipocytes were not affected by any of the EGCG treatments. In contrast, overexpression of either MKK6EE, a constitutively active mutant of MKK6 to activate p38 MAPK kinase, or MEKK1 (data not shown), a MEKK1 construct favoring the activation of JNK MAPK kinase, did not prevent EGCG-induced decreases in the number of preadipocytes. Taken together, these findings demonstrate that a suppressive effect of EGCG on preadipocyte proliferation is likely mediated via ERK MAPK-dependent and p38 MAPK- and JNK MAPK-independent pathways and confirms that the ERK MAPK subfamily is important in preadipocyte proliferation, as reported by Boney et al. (5).
The ERK-dependent effect of EGCG observed in 3T3-L1 preadipocytes is also strengthened by our findings that two specific inhibitors of Erk MAPK, PD98059 (12) and U0126 (13), alone inhibit cell growth and MEK1 activity, as shown with reduced phospho-ERK1/2 (5, 40), and that either of them used to treat preadipocytes within the IC50 range speeds up EGCG-induced reduction in the amounts of phosphorylated ERK1/2 and, to a lesser extent, in the number of cells. It appears that EGCG works differently from PD98059 and U0126 in reducing levels of phosphorylated ERK1/2 proteins. Whereas PD98059 prevents MEK1 activation by Raf (12), U0126 directly protects ERK from being phosphorylated by MEK1 (13). In cell-free systems, the inhibition of EGCG on MAPK activity is competitive with the myelin basic protein substrate and is noncompetitive with ATP (47). In contrast, the activities of certain protein phosphatases are stimulated by 15% by 10∼50 μM EGCG (47). In cultured cells, phosphorylation of ERK1/2 can be regulated by a variety of factors, including growth factors, G protein-coupled receptors, tyrosine kinase receptors, and Raf and MEK1 kinases (4, 31). We measured the amounts of phospho-ERK1/2 protein after 3T3-L1 preadipocytes were treated with either 1 nM IGF-I or 10 nM IGF-II in the presence and absence of 50 μM EGCG. EGCG did significantly prevent the increase in phosphorylated ERK1/2 by either IGF-I or IGF-II (Fig. 7) and concomitantly reduced IGF-I receptor activity, as indicated by a decrease in the phosphotyrosine-IGF-I receptor and an association of the IGF-II receptor with Giα-2 protein (unpublished observations). Alternatively, it is possible that EGCG induces a decrease in phosphorylated ERK1/2 from preadipocytes via reducing the phosphorylation of MEK1 and the association of Raf with MEK1 as reported for H-Ras-transformed cells (9). However, confirming this requires more thorough studies.
Antimitogenic effect of EGCG on preadipocytes depends on Cdk2 pathway.
Cdks are key regulators of the cell cycle in vertebrate cells. They are related to the effects of EGCG in modulating cell mitogenesis and growth arrest of most cancer cells (1, 23–25) and can serve as the main controller of mitogenesis and mitotic clonal expansion of preadipocytes (40). We observed herein that doses of 20∼100 μM EGCG decreased Cdk2 activity at 4, 24, and 48 h and reduced its protein levels at 48 h but not at 4 and 24 h. Also, EGCG dose dependently induced G1 growth arrest at 24 and 48 h after treatment. In addition, increased Cdk2 activity via the transfection of Cdk2+/+ cDNA to preadipocytes prevented EGCG-induced decreases in their Cdk2 activity and cell number and EGCG-induced increases of G1 arrest, whereas decreased Cdk2 activity via the transfection of dnCdk2 cDNA to preadipocytes slowed down the 5-day growth of preadipocytes and increased their G1 growth arrest (data not shown). These observations suggest that the effect of EGCG of inducing preadipocyte antimitogenesis and growth arrest is dependent on a Cdk2 pathway and requires inactivation of the Cdk2 protein. Because cyclin D1 is a G1 cyclin associated with Cdk4 and Cdk6 proteins, which favor cell cycle arrest at the G1 checkpoint (4), decreased cyclin D1 protein expression by EGCG for 24∼48 h suggests the possibility of Cdk4- and Cdk6-related effects of EGCG on preadipocyte growth arrest. However, we observed herein that 20 μM EGCG for 48 h did not affect the total levels of Cdc2 protein and that 100 μM EGCG for 48 h reduced the levels of Cdc2 protein less than that of Cdk2. Accordingly, EGCG appears to act on a specific type of Cdks in preadipocytes, but further studies are required to illustrate this contention. Because Cdc2 takes over as the predominant Cdk activity in the early G2/M transition of the cell cycle (4), the observed decrease in Cdc2 protein expression by 100 μM EGCG suggests the possibility of the action of EGCG on the G2/M phase of preadipocytes.
Regulation of Cdk2 activity of in vivo cultured cells occurs at multiple levels, involving the synthesis of subunits and the association of inhibitory proteins such as p21 and p27 (29, 37). On these bases, a decrease in the Cdk2 activity of 3T3-L1 preadipocytes induced by 4 h of EGCG treatment may have resulted from the observed increase in the association of Cdk2 with p21 and p27 by EGCG. However, the short-term effect of EGCG in reducing Cdk2 activity should be unrelated to the availabilities of p21, p27, and Cdk2 because EGCG did not alter their protein levels in this period. However, increased levels of p21 and p27, but not of p18 or Cdk2, observed with the 24-h EGCG treatment may be responsible for the increased association of Cdk2 with p21 and p27, but not p18, by EGCG, thereby leading to low Cdk2 activity and a subsequent rise in the percentage of G0/G1 arrest. In addition, the decreased levels of Cdk2 protein and increased levels of p21 and p27, but not p18, observed with the 48-h EGCG treatment may explain the EGCG-induced increase in the association of Cdk2 with p21 and p27, but not p18, thereby resulting in decreases in Cdk2 activity and increases in the percentage of G0/G1 arrest. These results suggest that EGCG may act on a particular type of preadipocyte in the CKI family to reduce Cdk2 activity. It would be of interest if other types of CKIs, such as p53 (4), were also found to be involved in the action of EGCG on preadipocyte growth arrest as reported in cervical cells (35). As the sequestration of p21 and p27 is mediated via the induction of cyclin D1 and cyclin D2 protein synthesis rates (32), the 24- and 48-h decreases in cyclin D1 protein expression of preadipocytes by EGCG may also explain the EGCG-induced increase in the association of Cdk2 with p21 and p27. Further studies to determine whether EGCG affects the association of cyclin D1 with these CKIs would help clarify this notion.
The Cdk2 activity of in vivo cultured cells is also regulated at phosphorylation-dephosphorylation levels (29, 37). In rat aortic smooth muscle cells, increased Cdk2 activity by endothelin is mediated via either the activation of Erk and Cdc25A (a phosphatase) or the inactivation of WEE1 (an inhibitory kinase) but is prevented by the inactivation of ERK and the activation of WEE1 (7). Whereas Cdk2 activity is inactivated through phosphorylation at Tyr15 by WEE1, it is restored through dephosphorylation at Tyr15 by Cdc25A (7). Activities of WEE1 and Cdc25A are, respectively, inactivated and activated through phosphorylation by the activation of ERK, and vice versa (7). Accordingly, the observed EGCG reduction in Cdk2 activity in 3T3-L1 preadipocytes may be mediated by the inactivation of ERK. This explanation is supported by observations that decreased amounts of phospho-ERK1/2 proteins by EGCG occur in parallel to the reduced activity, but not protein levels, of Cdk2 by EGCG over a 48-h period. Because Cdk2 activity in in vivo cells is also regulated by the association of stimulatory proteins, such as cyclin E (4, 29), more investigations are needed to clarify whether the production of cyclin E protein and its association with Cdk2 are altered by EGCG and thereafter cause decreases in Cdk2 activity. However, an interesting observation not shown here is that the cyclin D1 protein is able to associate with Cdk2, and such an association can be altered with EGCG treatment (unpublished observations). This suggests that EGCG may result in altered cyclin-Cdk2 protein complexes in 3T3-L1 preadipocytes as reported in mouse liver cells (14).
GTCs have numerous biological activities that can possibly provide various health benefits (1, 24–25, 28, 45). In most cases, but not all, gallated catechins, especially EGCG, are more active than other catechins. This contention is supported by our findings in 3T3-L1 preadipocytes that at the same dose and duration of treatment, EGCG was generally more effective than EC, ECG, and EGC in changing the number of cells, the amount of incorporated BrdU, percentages of the four phases of the cell cycle, activities of MEK1 and Cdk2, and levels of Cdk2, cyclin D1, and CKIs. The observed catechin-specific effects of green tea suggest that EGCG may act differently from EC, EGC, and ECG in regulating preadipocyte growth. According to the nature of the unique structures of these catechins tested (Fig. 1) (24, 33), EGCG contains the largest number of hydroxyl groups on its three aromatic rings among the tea catechins, and these hydroxyl groups may be important for hydrogen bonding. Also, EGCG has both gallyl and galloyl groups, which have some conformational flexibility, that may also be important for interactions with other molecules. Further exploration of the chemical basis of the antimitogenic activity on preadipocytes by EGCG is needed to understand differences in the mechanism of EGCG's action compared with those of EC, EGC, and ECG on these processes.
Growth phase-dependent effect.
The effects of green tea EGCG in reducing the number of 3T3-L1 preadipocytes are dependent on the latent, proliferative, and plateau phases of preadipocyte growth. This is evidenced by our observations that the IC50 values of EGCG for reducing the number of day 1-6 cells differed during the 72-h treatment. This suggests that latent, log-phase, and confluent preadipocytes have different sensitivities to this tea catechin. This is similar to the studied growth phase-dependent effect of IGF-I on the activation of MAPK protein phosphorylation in 3T3-L1 preadipocytes, particularly proliferating cells (5). A possible explanation is that various endogenous properties of the three phases of cells occur, thereby leading to EGCG's signal proceeding in different phases of cells. The findings that EGCG decreased the amounts of phospho-ERKs in day 3 cells, but not in day 1 or 5 cells, and that the basal amounts of phospho-ERKs in day 3 cells were the largest among these stages of cells (Fig. 4C) actually support this contention. Determining whether the number and affinity of the EGCG receptor (39) and its associated signal proteins exhibit different presences in the latent, proliferative, and confluent phases of preadipocytes should help elucidate their distinct sensitivities to EGCG.
Differential effects of EGCG on 3T3 and 3T3-L1 cells.
Mechanistic studies of green tea EGCG have reported its cell type-dependent manner (1, 24, 25, 45). This may be explained by the fact that the sensitivity of different normal, transformed, and cancer cell lines to green tea EGCG varies (24), although such differences may be due to the cell culture techniques and assay methods employed. In this study, we used the same experimental culture conditions and assay methods to look at whether a cell type-specific effect of EGCG occurs. We observed that the IC50 value of EGCG for reducing the cell number was lower in 3T3-L1 cells than in 3T3 fibroblast and human KB oral cancer cells (data not shown), supporting the existence of different sensitivities of these cell lines to EGCG, as similarly reviewed by Liao et al. (24). Such differences between 3T3 fibroblasts and 3T3-L1 preadipocytes induced by EGCG may be explained by observations that the decrease in phosphorylated ERK1/2 of the former cells in response to 48-h EGCG treatment was much less than that in the latter cells (Fig. 9) and that levels of Cdc2 and Cdk2 proteins were, respectively, decreased and increased by EGCG in the former cells, whereas Cdc2 levels in the latter cells were decreased much less by EGCG than were Cdk2 levels (Fig. 10). These observations suggest that EGCG may work differently in the two cell lines in modulating mitogenesis via altering different phases of their cell cycles and/or the extent of apoptosis and that the Cdk2 transducer may play different roles with EGCG in the two cell lines. It should be noted that in certain cells, Cdks are activated during apoptosis (49), although it is unknown whether the increase in Cdk2 protein levels in 3T3 fibroblasts by EGCG is related to the mechanism through which EGCG induces a decrease in the cell number. Other different characteristics between 3T3 fibroblasts and 3T3-L1 preadipocytes have been reported, although the latter cells are subcloned from the former cells (15). It would be worthwhile to explore whether any of them are responsible for the differential effects of EGCG on 3T3 and 3T3-L1 cells.
We conclude that the antimitogenic effect of EGCG on 3T3-L1 preadipocytes is dependent on the Erk MAPK and Cdk2 pathways and is likely mediated through decreases in their activities (Fig. 14). While shown to be mediated by ERK MAPK, signaling is demonstrated to be largely independent of the p38 MAPK and JNK MAPK pathways. Decreases in Cdk2 activity by EGCG may be due to its effect on this particular member of the CKI family. In general, EGCG is more effective than other structurally related GTCs in changing mitogenic signals. The signaling of EGCG in 3T3-L1 preadipocytes differs from that in 3T3 fibroblasts. Changes in the endogenous signals of preadipocytes induced by EGCG may help increase our understanding of the modulatory mechanism of green tea EGCG on body weight and fat cells (18–19, 24). Future studies on discovering the EGCG receptor in fat cells and on characterizing its oxidative stress are needed to elucidate the mechanisms of how EGCG signals reduce the activities of MEK1 and Cdk2 proteins.
This study was supported by Grants NSC90-2311-B-008-002 and NSC91-2311-B-008-004 from the National Science Council, Taiwan, and the Brain Research Center, Taiwan (to Y.-H. Kao).
We are grateful to Dr. Shutsung Liao and associates at the Ben May Institute for Cancer Research of the University of Chicago for the gift of green tea ECG. Also, we thank Drs. Shen-Liang Chen and Jaw-Jing Yang for the gifts of MEK1, MEKK1, MEK1EE, and MKK6EE cDNA constructs. Finally, we thank Drs. Lee-Young Chau and Shun-Chern Tsaur for technical assistance.
↵* P.-F. Hung and B.-T. Wu contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society