(−)-Epigallocatechin gallate (EGCG), a main constituent of green tea polyphenol, can act synergistically with cancer preventive agents (such as sulindac and tamoxifen) in inducing apoptosis of lung, colon and breast cancer cells.1, 2, 3 Since EGCG and green tea have a wide range of target organs and are nontoxic for animals and humans, there is growing interest in increasing the antitumor activity of cancer preventive agents in combination with EGCG or green tea.4, 5, 6 However, the molecular mechanisms responsible for the synergistic effects of cotreatment with EGCG plus cancer preventive agents are not known. Using human cDNA cancer expression array, we have found upregulated expressions of 2 genes, growth arrest and DNA damage-inducible gene 153 (GADD153) and p21WAF1, and downregulated expressions of 4 genes, T-plasminogen activator, TIMP3, IL-1β and integrin β4: All of these genes are associated with synergistic induction of apoptosis of human non-small-cell lung carcinoma (NSCLC) cell line PC-9 by cotreatment with EGCG plus sulindac.7 Treatment with either EGCG or sulindac alone did not affect these gene expressions. Since upregulation of GADD153 gene expression was so dramatic among the affected genes, we became interested in what part the upregulation of GADD153 gene plays in synergistic induction of apoptosis by combination with EGCG plus cancer preventive agents.
GADD153, also as known as CHOP (C/EBP homology protein), is a transcription factor belonging to the CCAAT/enhancer binding protein (C/EBP) family.8GADD153 gene expression is highly upregulated in the cells as a result of various stress conditions, such as oxidative stress, endoplasmic reticulum stress, glucose deprivation and treatment with DNA damaging agents.9 Transfection of GADD153 gene into various different cancer cell lines induces apoptosis without any stress inducing factors, indicating that GADD153 is directly involved in the regulation of apoptosis.10 Furthermore, it has been reported that GADD153 protein plays an important role in the induction of apoptosis of cancer cells treated with N-(4-hydroxyphenyl)retinamide (4HPR), a synthetic retinoid, as a retinoic acid receptor-independent pathway.11, 12 In addition to 4HPR, various cancer preventive agents—such as curcumin, thiazolidinedione (an agonist of peroxisome proliferator-activated receptor-γ) and phenethylisothiocyanate—induced apoptosis of cancer cells mediated through upregulation of GADD153 gene.13, 14, 15 These results strongly suggest that overexpression of GADD153 gene is a new molecular mechanism of antitumor activity by cancer preventive agents. This allowed us to conceive that upregulation of GADD153 gene would play a key part in cancer prevention in combination with EGCG plus cancer preventive agents. To clarify our hypothesis, we first examined whether high upregulation of GADD153 gene expression and synergistic induction of apoptosis would be observed in combination with EGCG plus cancer preventive agents other than sulindac. For this experiment, we chose celecoxib as a candidate, since celecoxib, a cyclooxygenase-2 (COX-2) selective nonsteroidal antiinflammatory drug (NSAID), is thought to be a promising preventive agent for colon and lung cancer.16 We then examined the significance of GADD153 gene expression in synergistic induction of apoptosis by looking at signaling pathway.
We found that cotreatment with EGCG plus celecoxib strongly upregulated GADD153 gene expression—but not p21WAF1 gene—and synergistically induced apoptosis of PC-9 cells, associated with activation of the mitogen-activated protein kinase (MAPK) signaling pathway. Cotreatment with EGCG plus celecoxib activated both ERK1/2 and p38 kinases, and pretreatment with PD98059 (a specific inhibitor of ERK1/2) and UO126 (a selective MEK inhibitor) abrogated GADD153 gene expression and then inhibited synergistic induction of apoptosis. All the results indicate that EGCG stimulates apoptosis induced by celecoxib, mediated through GADD153 gene expression, and that GADD153 is a new molecular target for cancer prevention in combination with EGCG.
Material and methods
Cell line and reagents
Human non-small-cell lung carcinoma cell lines, PC-9, A549 and ChaGo K-1, were maintained in RPMI-1640 containing 10% fetal bovine serum. EGCG (more than 99% purity) was obtained from Japanese green tea leaves. Celecoxib was purchased from LKT Laboratories, MN. PD98059 (inhibitor of ERK1/2), UO126 (inhibitor of MEK), SB203580 (inhibitor of p38 MAPK), JNK inhibitor II (SP600125) and calphostin C (inhibitor of protein kinase C) were purchased from Calbiochem, CA. Anti-GADD153 and antiactin antibodies were purchased from Santa Cruz Biothechnology, CA. Specific antibodies for phospho-ERK1/2 (Thr202/Tyr204), ERK1/2, phospho-p38 MAPK (Thr180/Try182) and p38 MAPK were obtained from Cell Signaling, MA.
Quantitative real-time PCR of GADD153 gene
Human lung cancer cells (PC-9, A549 and ChaGo K-1) were treated with various concentrations of celecoxib in the presence or absence of EGCG for 24 hr. Total RNA was isolated by using ISOGEN reagent (Nippon Gene, Japan), and then subjected to reverse-transcription as described previously.17 Expression level of GADD153 gene was quantitatively measured by real-time PCR performed on LightCycler™ System (Roche Diagnostic Corporation, IN) using SYBER Green. The critical threshold was determined according to the manufacture's instruction. The relative abundance of GADD153 mRNA was normalized against that of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA for quantitative evaluation. The sequences of the primer sets used are as follows: GADD153 sense primer: 5′-GAAACGGAAACAGAGTGGTCATTCCCC-3′, antisense primer: 5′-GTGGGATTGAGGGTCACATCATTGGCA-3′: GAPDH sense primer: 5′-TGGTATCGTGGAAGGACTCATGAC-3′ and antisense primer: 5′-ATGCCAGTGAGCTTCCCGTTCAGC-3′. The results were obtained by at least two independently performed experiments.
Semiquantitative RT-PCR of GADD153, p21WAF1 and GADD45 genes
Expression of GADD153, p21WAF1, GADD45 and GAPDH genes was determined by semiquantitative RT-PCR method as described previously.18 The sequences of the primer sets used are as follows: p21WAF1 sense primer: 5′-GCGCCATGTCAGAACCGGCTG-3′, antisense primer: 5′-TCCTCCCAACTCATCCCAACC-3′: GADD45 sense primer: 5′-GCCTGTGAGTGAGTGCAGAA-3′ and antisense primer: 5′-CCCCACCTTATCCATCCTTT-3′. The experiments were repeated at least twice.
Detection of apoptosis by flow cytometry
Apoptosis was examined by flow cytometry of propidium iodide (PI)-stained cells as described previously.19 PC-9, A549 and ChaGo K-1 cells were treated with EGCG plus celecoxib or celecoxib alone for 40 hr, then cells were fixed with ice-cold 70% ethanol for 30 min and stained with 50 μg/ml PI. DNA contents of stained cells were measured by Flow Cytometer (Coulter Co., FL).
Western blot analysis
PC-9 cells were harvested 1 hr after treatment with EGCG plus celecoxib, and nuclear fraction was prepared using Nuclear/Cytosol Fractionation Kit (BioVision, CA) for determination of GADD153 protein.14 Forty micrograms of nuclear fraction was separated by SDS-PAGE, and then Western blot analysis was conducted using anti-GADD153 antibody and antiactin antibody as a control.
Activation of MAPKs in cytosol fraction of PC-9 cells was determined by phosphorylation of MAPKs using specific antibodies for phospho-ERK1/2 (Thr202/Tyr204), ERK1/2, phospho-p38 MAPK (Thr180/Tyr182) and p38 MAPK (Cell Signaling, MA).
Analysis of combination effects
Synergistic effect was examined by isobolographic analysis using the equation D = [(Ac/Ae) + (Bc/Be)], where A and B are the 2 agents, c is the concentration of the agent used in the combination treatment, and e is the concentration of the agent used in mono-treatment to exert the same effect as with the combination treatment.20 When D value was <0.8, the effect of the agents used in the combination treatment was considered synergistic.
Statistical analyses of GADD153 gene expression, apoptosis induction and phosphorylation of MAPKs were conducted by Student's t-test. The results were considered significant if p values was less than 0.05 (*: p < 0.05 and **: p < 0.01).
Upregulation of GADD153 expression by cotreatment with EGCG plus celecoxib
To examine whether EGCG stimulates GADD153 gene expression with celecoxib, PC-9 cells were treated with various concentrations of celecoxib in the presence or absence of 100 μM EGCG, a concentration chosen based on results showing that EGCG at concentrations up to 100 μM did not induce GADD153 gene expression or apoptosis in PC-9 cells (Figs. 1a and 2). Figure 1a shows high upregulation of GADD153 gene after 24 hr cotreatment: Cotreatment stimulated GADD153 gene expression up to 17-fold over nontreated cells, while celecoxib alone at a concentration of 50 μM induced the expression only by 6-fold. Upregulated GADD153 gene expression was observed from 16 hr after cotreatement to 48 hr after. Although these results were similar to those with EGCG plus sulindac, synergistic induction of GADD153 gene was observed at lower concentrations (0.1 μM) of celecoxib when compared with that of sulindac (10 mM) in combination with EGCG (Fig. 1a). As a result of strong induction of the gene expression, GADD153 protein greatly increased in the cells cotreated with EGCG plus celecoxib (Fig. 1b). Cotreatment with EGCG plus celecoxib stimulated expression of GADD153 protein at the transcriptional level. Furthermore, the synergistic effects of cotreatment with EGCG plus celecoxib on induction of GADD153 gene expression were observed in 2 other human lung cancer cell lines, A549 and ChaGo K-1, although neither EGCG nor celecoxib alone showed any upregulation of GADD153 gene. The synergistic effects on these 2 cell lines were slightly weaker than those observed in PC-9 cells (Table I).
Table I. Synergistic Effects of Cotreatment with EGCG Plus Celecoxib in Human Lung Cancer Cell Lines
On the basis of our previous results showing that cotreatement with EGCG plus sulindac induced both GADD153 and p21WAF1 gene expressions,7 we next examined the effects of cotreatemnt with EGCG plus celecoxib on expressions of p21WAF1 gene along with another GADD family gene (GADD45) by semiquantitative PCR. Interestingly, expression of p21WAF1 and GADD45 genes were not enhanced by cotreatment with EGCG plus celecoxib (Fig. 1c). Celecoxib alone dose-dependently increased p21WAF1 gene expression, but cotreatment with EGCG slightly inhibited it. GADD45 gene expression was not affected by treatments with EGCG plus celecoxib, with celecoxib alone, or with EGCG alone (Fig. 1c). Although we examined the expression of these genes in PC-9 cells at various times after cotreatment, we did not observe induction of either 2 gene expression (data not shown). Thus, cotreatment with EGCG plus celecoxib induced high upregulation of only GADD153 gene expression, although p21WAF1 and GADD45 genes are also apoptosis related genes in PC-9 cells.21, 22 Thus, GADD153 is a unique biomarker, induced by EGCG and celecoxib cooperatively.
Synergistic induction of apoptosis by cotreatment with EGCG plus celecoxib
We next examined whether cotreatment with EGCG plus celecoxib would induce apoptosis synergistically in accordance with the synergistic induction of GADD153 gene expression. We looked for any increase in cells of sub-G1 phase (apoptosis-phase) from flow cytometric analysis. Treatment with celecoxib alone, even at a concentration of 50 μM, and EGCG alone at concentrations up to 150 μM did not significantly increase % of apoptotic cells after 40 hr incubation (Fig. 2). However, EGCG combined with celecoxib strongly stimulated induction of apoptosis: EGCG increased % of apoptotic cells about 6.4-fold (4.0–25.4%) in combination with 1 μM celecoxib, about 14.9-fold (3.0–44.6%) with 10 μM celecoxib, and about 60.1-fold (0.7–42.2%) with 50 μM celecoxib (Fig. 2). Isobolographic analysis of these results clearly indicated that the cotreatment synergistically induced apoptosis of PC-9 cells (D = 0.75). Synergistic induction of apoptosis by cotreatment was also observed in both A549 and ChaGo K-1 cells. (Table I) This synergistic induction of apoptosis followed the induction of GADD153 gene expression in human lung cancer cells by cotreatment.
To clarify the involvement of GADD153 in synergistic induction of apoptosis by cotreatment with EGCG and cancer preventive agents, we then compared the effects of various cancer preventive agents—celecoxib, sulindac, 4HPR, and aspirin—in combination with EGCG, on GADD153 gene expression and induction of apoptosis (Table II). As mentioned earlier, celecoxib and sulindac showed synergistic effects on both GADD153 gene expression and apoptosis, in combination with EGCG. EGCG did not enhance either activity, combined with either 4HPR or aspirin; 4HPR alone strongly induced GADD153 gene expression and apoptosis in PC-9 cells; aspirin did not show any induction in PC-9 cells. These results indicated that GADD153 plays a part in the induction of apoptosis of cancer cells with cancer preventive agents, and that cotreatment with EGCG plus celecoxib or sulindac may exert antitumor activity mediated through mechanisms similar to those of 4HPR, a synthetic retinoid with antitumor activity.
Table II. Upregulation of GADD153 Gene Strongly Associated with Apoptosis Induction
Requirement for ERK1/2 activity in both cotreatment-induced GADD153 gene expression and synergistic induction of apoptosis
To study the mechanism involved in upregulation of GADD153 gene expression by cotreatment with EGCG plus celecoxib, we focused on activation of the MAPK signaling pathway, based on previous reports.10 Cotreatement significantly induced phosphorylation of ERK1/2 and p38 MAPK, but not that of JNK. Figure 3 shows dose-dependent increase of phopho-ERK1/2 and phospho-p38 MAPK in PC-9 cells 1 hr after treatment with EGCG plus celecoxib. Phosphorylation levels of ERK1/2 and p38 MAPK were sustained at 4 hr after, although protein levels of ERK1/2 and p38 were unchanged. However, neither EGCG nor celecoxib alone induced phosphorylation of these kinases: Only cotreatment with EGCG plus celecoxib activated ERK1/2 and p38 MAPK. These results indicated that ERK1/2 and p38 MAPK pathways are involved in the regulation of synergistic induction of GADD153 gene expression and apoptosis.
Using specific inhibitors of the kinases, we wanted to determine which MAP kinase regulated synergistic expression of GADD153 gene. PD9805923 (a specific inhibitor of ERK1/2), UO12624 (a selective MEK inhibitor of ERK1/2 signaling) and SB20358025 (a specific inhibitor of p38 MAPK) were added to the cells 1 hr before treatment with 100 μM EGCG plus 10 μM celecoxib: Pretreatments with PD98059 and UO126 dose-dependently reduced levels of phospho-ERK1/2 that were enhanced by cotreatment (Fig. 4a). Specifically, PD98059 and UO126 dose-dependently inhibited upregulation of GADD153 gene. Sixty micromolars of PD98059 decreased it to 27.6% and 5 μM UO126, to 43.6% (Fig. 4b). However, pretreatment with SB203580 at concentrations up to 60 μM did not inhibit it. Furthermore, effects of calphostin C (an inhibitor of protein kinase C) on GADD153 gene expression induced by the cotreatment was examined, since EGCG has inhibitory activity on protein kinase C26: Calphostin C did not affect the upregulation of GADD153 gene expression. Protein level of GADD153 significantly decreased by pretreatment with PD98059 and UO126, which correlated well with inhibition of the gene expression (Fig. 4c). These results show that the ERK signaling pathway is involved in the upregulation of GADD153 gene expression by cotreatment with EGCG plus celecoxib, but that p38 MAPK, JNK and protein kinase C probably are not.
Just as it inhibited GADD153 induction, PD98059 (an inhibitor of ERK1/2) and UO126 (a selective MEK inhibitor of ERK1/2) inhibited apoptosis synergistically induced by the cotreatment: Pretreatment with PD98059 reduced apoptotic cells dose-dependently, from 55.3 to 22.6%, and 5 μM UO126, to 36.8% (Fig. 4d). Since UO126 concentrations higher than 5 μM was toxic to the cells, we could not get strong inhibition of apoptosis similar to that of PD98056. However, SB203580 and calphostin C did not affect % of apoptotic cells (data not shown). All the results demonstrated that cotreatment with EGCG plus celecoxib induced the upregulation of GADD153 gene expression mediated through activation of ERK pathway, which resulted in strong induction of apoptosis in lung cancer cells. Thus, GADD153 is required for the enhanced antitumor activity of cancer preventive agents, in combination with EGCG.
In the present study, we found that GADD153 plays a significant role in the synergistic induction of apoptosis of NSCLC (PC-9, A549 and ChaGo K-1 cells) by cotreatment with EGCG and the cancer preventive agent, celecoxib. Since all 3 lung cancer cell lines have p53 mutation,27, 28, 29 GADD153 seems to regulate p53-independent apoptosis. GADD153 is a transcription factor belonging to the CCAAT/enhancer-binding proteins (C/EBPs); it forms heterodimers with other C/EBP family proteins and alters their activities. There is considerable evidence indicating that GADD153 is directly involved in the apoptosis pathway: GADD153 enhances cellular sensitivity to apoptosis by suppressing the transcription of antiapoptotic Bcl-2,30 and overexpression of GADD153 also promotes mitochondrial translocation of proapoptotic protein BAX.31 Furthermore, mouse embryonal fibroblasts prepared from GADD153 knockout mice were resistant to apoptosis induced with agents that perturb the endoplasmic reticulum function.32 It was recently reported that oncogenic Ras downregulates GADD153 expression and exogenous GADD153 inhibits Ras-induced cellular transformation.33 So GADD153 is an important factor involved not only in killing cancer cells but also in the anticarcinogenic process, where it downregulates cell growth and survival. Although EGCG also has an inhibitory activity of COX-2,34 we assume that COX-2 independent mechanisms may be involved in the synergistic induction of GADD153 gene expression and apoptosis: Cotreatment with EGCG plus celecoxib showed only additive effects on inhibition of COX-2 activity in vitro (unpublished results), and sulindac sulfone (an inactive metabolite of sulindac in COX inhibition) also showed synergistic effects similar to sulindac sulfide (an active metabolite) by cotreatment with EGCG.7 Our assumption is supported by previous reports that celecoxib and other COX-2 inhibitors induced apoptosis independent of COX-2 activity.
In addition to the results with celecoxib and sulindac, we recently found that a combination of EGCG plus all-trans retinoic acid (ATRA) also induced upregulation of GADD153 and synergistic induction of apoptosis of PC-9 cells (data not shown). Our findings indicate that GADD153 is a key for stimulation of cancer preventive activity with EGCG and probably green tea. I. B. Weinstein and his associates reported that cotreatment with EGCG plus 5-fluorouracil (5-FU) markedly enhanced the growth-inhibitory activity in human head and neck squamous cell carcinoma cell lines35: It will be interesting to find out whether combined treatment with EGCG and 5-FU can induce upregulation of GADD153 gene expression in the cells. Although it is not yet clear how the combination treatment activates the ERK1/2 kinase pathway, we think that EGCG sensitizes cancer cells to other cancer preventive agents—such as celecoxib, sulindac, and ATRA—by upregulation of GADD153 gene expression.
In the present study, 4HPR alone strongly induced GADD153 gene expression and apoptosis of PC-9 cells, and EGCG alone did not enhance either activity. Therefore, we think that cotreatment with EGCG and celecoxib exert mechanisms similar to those of 4HPR. Recent paper reported that GADD153 is an important transcription factor for 4HPR-induced apoptosis in a p53-independent manner, and that this is mediated through activation of 12-lipoxygenase leading to endoplasmic reticulum (ER) stress.36 Interestingly, 4HPR acts synergistically with chemotherapeutic drugs similar to EGCG. Whether the combinations with EGCG plus celecoxib, sulindac or ATRA induce ER stress in lung cancer cells needs to be further studied.
Recently, it was reported that ER stress signal mediators are targets for the anticancer effect of selenium.37 A monomethylated selenium metabolite, methylselenic acid (MSA), induced ER stress mediators: apoptotic molecules, such as GADD153 and caspase-12 and -7, and survival/rescue molecules, such as phosphorylated ER resident kinase and glucose-regulated protein-78 and -94. Among these, GADD153 is an important transcription factor in apoptosis induction by MSA.37 Selenium also sensitizes tumor cells to a number of therapeutic drugs and increase the resistance of normal tissues to the toxic effects of these drugs,38 indicating that selenium may have dual effects: apoptosis response in cancer cells and survival response in normal cells by affecting ER stress mediators.37 Further investigation on induction of ER stress signals mediators other than GADD153 by combination with EGCG will give us new insights into the molecular mechanisms of cancer prevention. All reported results supported our findings that GADD153 is a key transcription factor for combination cancer prevention.
In Japan, clinical trials using 10 cups of green tea daily supplemented with green tea tablets for prevention of cancer recurrence in various organs—such as prostate, breast, liver and esophagus—are being conducted under the leadership of Japanese clinicians.39 Recently, we obtained some exiting results on the prevention of polyp development after colorectal polypectomy using green tea supplemented with green tea tablets (Moriwaki, personal communication). In the recent Annual Meeting of AACR, the preventive activity of green tea polyphenols in prostate intraepithelial neoplasia was also demonstrated.40 Evidence indicating the clinical efficacy of green tea is accumulating: Since green tea is nontoxic and has not shown any adverse effects, we think green tea is a useful remedy to reduce the adverse effects of celecoxib on cardiovascular risk.41 We conclude that the combination of EGCG and celecoxib, probably ATRA, will open the way to a proper practical strategy for success in human lung cancer prevention.
We thank Dr. Takashi Kuzuhara (Tokushima Bunri University), Dr. Kaoru Kiguchi (Science Park Research Division, M.D. Anderson Cancer Center), Drs. Kei Nakachi and Tomonori Hayashi (Radiation Effects Research Foundation) for their fruitful discussion and Miss Ikuko Shiotani (Saitama Cancer Center) for her assistance.