Functional p53 is required for triptolide-induced apoptosis ... - Nature

Oncogene (2001) 20, 8009 ± 8018 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Functional p53 is required for triptolide-induced apoptosis and AP-1 and nuclear factor-kB activation in gastric cancer cells Xiao-Hua Jiang1,2,5, Benjamin Chun-Yu Wong*,2,5, Marie Chia-Mi Lin3, Geng-Hui Zhu2, Hsiang-Fu Kung3, Shi-Hu Jiang1, Dan Yang4 and Shiu-Kum Lam2 1

Department of Gastroenterology, Rui-jin Hospital, Shanghai, Peoples Republic of China; 2Department of Medicine, University of Hong Kong, Hong Kong; 3Institute of Molecular Biology, University of Hong Kong, Hong Kong; 4Department of Chemistry, University of Hong Kong, Hong Kong

Triptolide, a major component in the extract of Chinese herbal plant Tripterygium wilfordii Hook f (TWHf), has potential anti-neoplastic eect. In the present study we investigated the potential therapeutic eects and mechanisms of triptolide against human gastric cancer cells. Four gastric cancer cell lines with dierent p53 status, AGS and MKN-45 (wild type p53); MKN-28 and SGC7901 (mutant p53) were observed as to cell growth inhibition and induction of apoptosis in response to triptolide treatment. We showed that triptolide inhibited cell growth, induced apoptosis and suppressed NK-kB and AP-1 transactivation in AGS cells with wild-type p53. Triptolide induced apoptosis by stimulating the expressions of p53, p21waf1/cip1, bax protein, and increased the activity of caspases. In addition, it caused cell cycle arrest in the G0/G1 phase. To examine the role of p53 in these functions, we showed that suppression of p53 level with antisense oligonucleotide abrogated triptolide-induced apoptosis and over-expression of dominant negative p53 abolished the inhibitory eect on NF-kB activation. Furthermore, we demonstrated that triptolide had dierential eects on gastric cancer cells with dierent p53 status. We showed that triptolide also inhibited cell growth and induced apoptosis in MKN-45 with wild-type p53, whereas it had no signi®cant growthinhibition and apoptosis induction eects on the MKN-28 and SGC-7901 cells with mutant p53. Our data suggest that triptolide exhibits anti-tumor and anti-in¯ammatory eects by inhibiting cell proliferation, inducing apoptosis and inhibiting NF-kB and AP-1 transcriptional activity. However, a functional p53 is required for these proapoptotic, anti-in¯ammatory and anti-tumor eects. Oncogene (2001) 20, 8009 ± 8018. Keywords: apoptosis; NF-kB; AP-1; p53 status; gastric cancer.

*Correspondence: BC-Y Wong, Department of Medicine, University of Hong Kong, Queen Mary Hospital, Hong Kong; E-mail: [email protected] 5 Contributed equally to this work Received 21 February 2001; revised 22 August 2001; accepted 18 September 2001

Introduction Tripterygium wilfordii Hook f (TWHf) has been used in traditional Chinese medicine for centuries. Its crude extracts continue to be used to treat a variety of autoimmune diseases, such as rheumatoid arthritis, nephritis, and systemic lupus erythematosus (Qin et al., 1981; Tao et al., 1989; Jiang et al., 1994). It has been suggested that the major therapeutic eects of TWHf are from ingredients such as triptolide, tripdiolide, triptonide, and triptophenolide (Zhang et al., 1990). High concentration of triptolide suppressed B and T lymphocyte proliferation in mice and showed immunosuppressive eect in skin allograft transplantation (Pu et al., 1990; Yang et al., 1992). Interestingly, triptolide also has antineoplastic activity. For example, it has signi®cant anti-leukemic activity in vivo against L-1210 and P-388 mouse leukemia and in vitro against cells derived from human nasopharnygeal carcinomas (Kupchan et al., 1972). Furthermore, triptolide signi®cantly suppresses the colony formation of breast cancer cell lines, stomach cancer cell lines and promyelocytic leukemia cell lines (Wei et al., 1991). The exact mechanism responsible for the anti-neoplastic and anti-in¯ammatory eect of triptolide is not clearly understood. One possible explanation is the induction of apoptosis. It has been shown that triptolide induces apoptosis in T-cell hybridomas and peripheral T cells (Yang et al., 1998) and it can sensitize tumor necrosis factor-a (TNG-a)-resistant tumor cell lines to TNF-ainduced apoptosis (Lee et al., 1999). Another candidate which is involved in the anti-tumor and anti-in¯ammatory eect by triptolide is NK-kB. A lot of studies show that triptolide executes its eect by inhibiting the transcriptional activity and/or binding activity of NFkB (Lee et al., 1999; Qiu et al., 1999; Zhao et al., 2001). Apoptosis is regulated by gene products, which are conserved from nematodes to mammals (Penninger et al., 1996). Among all apoptosis-related genes, p53 is of particular importance (Lowe et al., 1993). P53 is well known for suppression of cellular proliferation through two mechanisms, each operating in a distinct manner. In normal ®broblasts, p53 induces G1 arrest in response to DNA-damaging agents, presumably allowing the cells to perform critical repair functions before

Functional p53, AP-1 and NF-kB X-H Jiang et al

8010

progressing through the cell cycle (Linke et al., 1997). On the other hand, in abnormally proliferating cells or irradiated thymocytes, induction of p53 leads to apoptosis (Midgley et al., 1995). Furthermore, wildtype p53 protein level was increased during apoptosis induced by DNA-damaging agents (Zhan et al., 1996). The increased expression of wild-type p53 can induce apoptosis in myeloid leukemia and colon cancers (Yonish-Rouach et al., 1991; Shaw et al., 1992). Unfortunately, the p53 gene is inactivated in approximately 50% of tumors and p53 alteration has also been held responsible for the failure of most cancers to respond to radiotherapy and chemotherapy (Lee et al., 1995; MuÈller et al., 1996). For example, Vasey et al. (1996) reported that introduction of the dominant negative p53 enhanced resistance to irradiation, cisplatin, doxorubicin and cytarabin treatment in ovarian cancer cells. Nuclear factor-kB (NF-kB) and Activator protein-1 (AP-1) are two important transcriptional factors that may regulate a variety of in¯ammatory, apoptotic and immune responses. Triptolide has previously been shown to sensitize several solid tumor cell lines to TNF-a-induced apoptosis through the inhibition of NF-kB (Lee et al., 1999). Furthermore, triptolide alone was able to inhibit transcriptional activation of NF-kB in Jurket cell and human bronchial epithelial cells (Qiu et al., 1999; Zhao et al., 2001). Taken as a whole, these results suggest that the anti-tumor and anti-in¯ammatory eects of triptolide are most likely associated with the modulation of NF-kB activation. Whether triptolide has any modulation eect on AP-1 activity, however, has not been tested until now. In the present study, we investigate the potential therapeutic eects and underlying mechanisms of triptolide in human gastric cancer. We demonstrate that triptolide executes its anti-in¯ammatory and anti-tumor eect by inducing apoptosis and modulating AP-1 and NF-kB activity. We show that triptolide induces apoptosis in gastric cancer cells through a p53-dependent pathway and caspase-3 is required for triptolide-induced apoptosis. In addition, triptolide inhibited NF-kB and AP-1 activation in cells with wild-type p53, and forced alteration of p53 status suppresses triptolide-induced apoptosis and modulation of NF-kB. Thus, our results suggest that functional p53 is essential for triptolideinduced apoptosis and modulation of AP-1 and NF-kB activation in gastric cancer cells.

Results The effect of triptolide on cell growth in AGS We examined the antiproliferative eect of triptolide on human gastric cancer cell line AGS by MTT assay. Figure 1a showed that treatment with triptolide at concentrations ranging from 1 to 40 nM for 48 h inhibited cell proliferation dose-dependently in AGS cells. After 48 h of treatment with triptolide, an estimated 34% of AGS cells remained viable. Oncogene

Induction of apoptosis by triptolide in AGS AO staining showed that AGS cells treated with triptolide (up to 40 nM) underwent morphologic changes of apoptosis 24 h after treatment dose-dependently (Figure 1b). Analysis of DNA from AGS cells demonstrated that triptolide (510 nM) caused the generation of nucleosomal sized ladders of DNA fragments (Figure 1c). Effect of triptolide on the cell cycle phase distribution To further investigate triptolide-mediated dierential growth inhibition, we studied the eect of triptolide treatment on the cell cycle kinetics in AGS cells. Flow cytometry of three independent analysis showed that AGS cells (Figure 1d) were arrested in the G0/G1 phase after 24 h of 40 nM triptolide treatment (to 68.6+6.5% from 58.5+2.3% in control), and cells had a typical subdiploid peak (to 33.2% from 3.2%) on the DNA histogram. Triptolide treatment resulted in enhanced expression of p53, p21waf1/cip1, and bax Time-course experiments were carried out to determine the protein levels of p53 after treatment. As shown in Figure 2, there was a signi®cant elevation of the p53 protein level in triptolide (20 nM) treated AGS cells, starting as early as 2 h, reaching a maximal level which is about 16-fold of control after 8 h treatment. We also examined the p21waf1/cip1, bax, bc1-2 and c-myc protein expressions. In AGS cells, the protein levels of p21waf1/ cip1 and bax increased after 4 and 12 h of treatment, respectively. Triptolide did not alter the expression of bc1-2 and c-myc within the time frame of this experiment (Figure 2). Apoptosis triggered by triptolide is associated with activation of caspase-3 (CPP-32) and cleavage of poly (ADP-ribose) polymerase (PARP) Caspases, especially caspase 3 (CPP-32), have been shown to participate in apoptosis. To see if apoptosis induced by triptolide was regulated by caspase 3, we examined several aspects of the caspase 3 activation. In the ®rst experiment, dierent concentrations of triptolide were added to con¯uent AGS cells for 0, 6, 12, 24 and 48 h. After treatment, caspase protease activity was assessed in cell lysates by measuring hydrolysis of colorimetric caspase substrate DEVD ± rNA. Figure 3a illustrated that after the addition of triptolide (20 nM), caspase activity increased to maximal level within 24 h. In the next experiment, the expression of caspase 3 during apoptosis was studied by Western blot analysis. The 32 kD procaspase 3, when activated, was cleaved into subunits of 17 kD as shown in Figure 3b. Furthermore, an 85 kD fragment cleaved from the speci®c caspase-3 substrate, 116 kD PARP, was detected in extracts from AGS cells treated with triptolide. These results indicated that triptolide activated caspase-3 in gastric cancer cells during the


8011

Figure 1 Eect of triptolide on cell growth and apoptosis in AGS cells. (a) Dose-response of triptolide on cell growth inhibition by MTT assay. *P40.05. (b) Dose-response of triptolide on apoptosis. AGS cells were treated with various concentrations of triptolide for 24 h. The percentage of apoptosis was quanti®ed by AO staining. *P40.05. (c) DNA ladder pattern formation. AGS cells were treated with dierent concentrations of triptolide for 24 h and the formation of oligonucleosomal fragments was visualized by 1.8% agarose gel electrophoresis. M, DNA markers; Lanes 1 ± 4, AGS cells treated with 40, 20, 10, 0 nM of triptolide. (d) FACS analysis. AGS cells were treated with 40 nM triptolide for 24 h and their DNA content was determined by FACS. (a) AGS control; (b) AGS treated with triptolide

time of apoptosis. However, no signi®cant alteration of caspase-3 activity or PARP cleavage was detected in triptolide (5 ± 40 nM)-treated MKN-28 cells (data not shown).

100 mM zVAD-fmk and 100 mM DEVD-fmk (data not shown). At this concentration, both zVAD-fmk and DEVD-fmk signi®cantly suppressed triptolide-induced apoptosis (Figure 3c).

Caspase inhibitors suppressed triptolide-induced caspase 3 activation and apoptosis

Triptolide inhibits AP-1 and NF-kB transcriptional activity in AGS cells

Although caspase activation and proteolysis are hallmarks of apoptosis, inhibition of caspases does not always prevent cells from undergoing apoptosis, suggesting the existence of caspase-independent pathway. To con®rm the essential role of caspase 3 in triptolide-induced apoptosis, we next analysed the eect of a pan-inhibitor of the caspase family, zVAD-fmk, and a speci®c inhibitor for caspase 3, DEVD-fmk, on apoptosis. We ®rst determined the concentration of zVAD-fmk and DEVD-fmk required to inhibit caspase 3 activity. Our results showed the cleavage of caspase 3 was completely blocked by

Experiments were performed with a reporter plasmid containing AP-1 or multiple kB elements. First we treated the cells for dierent time period, we found that after treatment with 10 mM triptolide for 6 h, the transcriptional activity of AP-1 and NF-kB showed signi®cant decrease (data not shown). Then the cells were treated for the dierent concentration of triptolide for 6 h. As shown in Figure 4, triptolide inhibited the activation of AP-1 and NF-kB signi®cantly. However, triptolide is not a general transcriptional inhibitor because it does not inhibit control CMV or SV40 promoter activity. Oncogene


8012

Figure 2 Eect of triptolide on the expression of p53, p21waf1/ , bax, bc1-2 and c-myc protein. (a) AGS cells were exposed to 20 nM triptolide for dierent time intervals (0, 2, 4, 8, 12, 24 h). The protein levels were determined by Western blot. (b) The relative density of p53, p21waf1/cip1 and bax expression in AGS cells detected by video densitometry

cip1

Triptolide does not inhibit DNA binding of NF-kB and AP-1 To determine whether triptolide inhibits activation of NF-kB and AP-1 through the inhibition of DNA binding of NF-kB and AP-1, we examined the eect of triptolide on TNF-induced binding of NF-kB and AP-1 by electrophoretic mobility shift assay. TNF-a induced binding of NF-kB and AP-1 in AGS cells, and antibodies to p65 (Rel A) and c-Jun supershifted the complex demonstrating that p65 and c-Jun are part of the NF-kB complex and AP-1 complex induced by TNF-a (Figure 5). Triptolide alone did not induce binding of NF-kB or AP-1. Furthermore, triptolide did not aect the intensity of the NF-kB complex and AP-1 complex induced by TNF-a. Our results suggest that triptolide inhibits transactivation but not DNA binding of NF-kB and AP-1. The translocation of NF-kB to the nucleus is preceded by the phosphorylation, ubiquitination, and proteolytic degradation of IkB-a. To determine whether triptolide caused IkB-a degradation, we measured the level of total IkB-a and phosphorylated IkB-a. As shown in Figure 5, triptolide had no eect on IkB-a phosphorylation and degradation. Down-regulation of p53 with antisense oligonucleotides abolished triptolide-induced apoptosis in AGS cells Since p53 expression is upregulated after triptolide treatment in AGS cell, next we downregulated p53 by Oncogene

Figure 3 Eect of triptolide on caspase-3 activity and PARP expression in AGS cells. (a) Time course of triptolide eect on caspase-3 activity. Con¯uent cells were treated with 20 nM triptolide for 0, 8, 12, 24 and 48 h. Protease activity at each time point was determined as described in Materials and methods. The data are expressed as the mean+s.d. of three individual experiments. (b) Analysis of caspase-3 and PARP expression in triptolide-treated AGS cells. Cells were treated with 20 nM triptolide for 24 h. Then caspase-3 and PARP cleavage was analysed by Western blotting. (c) Eect of caspase inhibitors on triptolide-induced apoptosis in AGS cells. Cells were pretreated with 100 mM zVAD-fmk or 100 mM DEVD-fmk for 12 h and followed by 20 nM triptolide for another 24 h. The percentage of apoptosis was determined by acridine orange staining. The results are expressed as mean+s.e. from three dierent experiments

antisense oligonucleotide and compared apoptotic response to triptolide. The phosphorothioate antisense oligonucleotide for p53 and the control CG-matched randomized-sequence phosphorothioate oligonucleotides were pre-incubated with lipofectin (5 mg/ml, Life Technologies, Gaithersburg, MD, USA) for 30 min. Cells growing in log phase were seeded in 24-well plates at a density of 16105 cells per well, incubated with the oligos/lipofectin mixture for 30 min, then with medium containing oligos (®nal concentration equals to 5 mM) for 18 h. Triptolide was added to the culture medium for a ®xed time period for further analysis. Figure 6a showed that 5 mM p53 antisense oligonucleotide could


8013

Figure 4 Modulation of AP-1 and NF-kB transcriptional activity by triptolide. Cells that transiently express AP-1 or NFkB luciferase reporter gene construct were treated with dierent concentrations of triptolide for 6 h. Cells were then harvested for analysis of luciferase activity. The ®re¯y luciferase reading was normalized to renilla luciferase reading. Data represent the mean+s.d. from three experiments. (a) NF-kB activity in AGS (*P40.05); (b) AP-1 activity in AGS (*P40.05); (c) AGS cells were transiently transfected with 0.6 mg/well NF-kB luciferase reporter vector alone or together with 0.6 mg/well pCMV-p53mt135 vector. After 6 h treatment with triptolide, cells were then harvested for analysis of luciferase activity

eectively inhibit the triptolide-induced increase in p53 protein level after 12 h, whereas the control oligonucleotide showed no eect on p53 expression. The p53 antisense oligonucleotide also inhibited triptolide-induced apoptosis. At 20 nM of triptolide, the percentage of apoptotic cells decreased from 20.5+5.3% to 8.9+2.4% by the addition of 5 mM of p53 antisense oligonucleotide (P40.05). The same concentration of control oligonucleotide showed no eect on apoptosis induced by triptolide (Figure 6b). Wild-type p53 is essential for inhibition of NF-kB activation by triptolide To test if p53 status is also essential for the inhibition eect of NF-kB transcriptional activity by triptolide, we cotransfected NF-kB reporter plasmid with the

Figure 5 Triptolide does not inhibit NF-kB or AP-1 DNA binding activity and IkB degraduation. (a) Nuclear extracts were prepared and analysed in an electrophoretic mobility shift assay in AGS cells with a radiolabeled IgG NF-kB or AP-1 probe. Equal amounts (10 mg) of nuclear protein was loaded in each lane. In lanes 3 and 7, a rabbit polyclonal antibody to p65 (Santa Cruz Biotechnology) was added to the nuclear extract 10 min before the addition of radiolabeled probe. In lane 4, an 1006 excess of unlabeled IgG oligonucleotide was added 5 min before the addition of radiolabeled probe. Lane 1, control; lane 2, TNFa (20 ng/ml R&D systems); Lane 3, TNF-a +cold; Lane 4, TNFa +Ab; 5, triptolide (20 nM); and lane 6, TNF-a +triptolide. (b) Eect of triptolide on IkB-a degradation. AGS cells with or without 20 nM triptolide treatment were collected at dierent time points. The cytoplasmic IkB-a and phospho-IkB-a protein level were assayed by Western blot analysis as described in Materials and methods

pCMV ± p53mt135, a mutant p53 that is defective in the conformation of its core DNA-binding domain, into AGS cells. As shown in Figure 4c, forced expression of PCMV ± p53mt135 increased the NF-kB activation by about twofold and abolished triptolidemediated inhibition of NF-kB activation in AGS cells. Triptolide has differential effect on cell growth inhibition, apoptosis induction and NF-kB and AP-1 activation in gastric cancer cells To further con®rm the importance of p53 in triptolide's eect on gastric cancer cells, we used other gastric cancer cell lines with dierent p53 status. Our further experiments showed that the same concentration of triptolide inhibited cell growth and induced apoptosis in MKN-45 bearing wild p53, whereas it had no Oncogene


8014

Figure 6 Eect of oligonucleotides on p53 protein expression and apoptosis. (a) AGS cells were cultured with an antisense or sense oligonucleotide (5 mM) for 18 h before incubation with 20 nM triptolide for another 12 h. Lane 1, control; Lane 2, 20 nM triptolide; Lane 3, 20 nM triptolide+5 mM sense oligonucleotide; Lane 4, 20 nM triptolide+5 mM antisense-oligonucleotide. (b) AGS cells were pretreated with 5 mM sense (S) or antisense (AS) oligonucleotides for 18 h followed by triptolide treatment for another 24 h. The percentage of apoptotic cells was measured by acridine orange staining. (*P40.05) signi®cantly dierent from control triptolide alone group

growth-inhibition and apoptosis induction eects on gastric cancer MKN-28 and SGC-7901 cells with mutant p53 (Figure 7). On the other hand, triptolide reduced NF-kB and AP-1 transcriptional activity in MKN-45, however, instead of reduction, triptolide stimulated AP-1 and NF-kB activation in MKN-28 and SGC-7901 cells with mutant p53 (Figure 8). Discussion Triptolide is a promising immunosuppressive agent widely used in Chinese medicine. Recently, several lines of evidence suggest that it also has antineoplastic eect on human malignant cells. Unfortunately, the molecular mechanisms responsible for its anti-neoplastic eect are poorly understood. In the present study, we explore the eects of triptolide on cell growth and induction of apoptosis in gastric cancer cells. Our results indicate that triptolide inhibited cell proliferation and induced apoptosis more readily in the wild-type p53 cell lines. These results are consistent with other studies, which show that the presence of wt p53 can enhance sensitivity to DNA damaging agents (Weller et al., 1998), sensitivity to chemotherapy in breast carcinomas (Tetu et al., 1999), and sensitivity for cisplatin in ovarian carcinomas (Calvert et al., 1996). We further demonstrated that Oncogene

Figure 7 Eect of triptolide on cell growth and apoptosis in gastric cancer cells. (a) Dose-response of triptolide on cell growth inhibition by MTT assay. (*P40.05). (b) Dose-response of triptolide on apoptosis. Cells were treated with various concentrations of triptolide for 24 h. The percentage of apoptosis was quanti®ed by AO staining. (*P40.05)

p53 was directly responsible for this dierential sensitivity, we showed that p53 expression was upregulated in AGS cells after triptolide treatment, and suppression of p53 expression by using p53 antisense oligonucleotide signi®cantly abolished the triptolideinduced apoptosis. The tumor suppressor gene p53 mediates either cell cycle arrest or apoptosis in response to DNA damage, thus acting as a molecular `guardian of the genome' (Lane, 1992). Most cancers, including gastric cancer, have either mutation within the p53 gene or defects in the ability to induce p53, and this p53 de®ciency has long been suspected to be the cause of poor prognosis. In this study, we also demonstrated that MKN-28 and SGC-7901 gastric cancer cell lines with mutant p53 did not show any signi®cant change both in the cell growth and apoptosis after treatment with triptolide. Therefore, our results support the notion that p53 status in human gastric cancer cells contribute to the sensitivity to anti-cancer treatment, at least in part in the readiness to undergo apoptosis. The increase in p53 protein in response to various stresses is an important regulator of cell cycle and apoptosis (Amundson et al., 1998). In transcriptiondependent p53 activation, p53 functions as a site speci®c transcription factor that induces p53-inducible genes such as p21cip1/waf1, bax, gadd45, and mdm2 (Farmer et al., 1992; El-Deiry et al., 1993; Miyashita et al., 1995; Zhu et al., 1999). In this study, we showed that induction of p53

Relative NF-κB Activity


Figure 8 Modulation of AP-1 and NF-kB transcriptional activity by triptolide in gastric cancer cells. Cells that transiently express AP-1 or NF-kB luciferase reporter gene construct were treated with dierent concentrations of triptolide for 6 h. Cells were then harvested for analysis of luciferase activity. The ®re¯y luciferase reading was normalized to renilla luciferase reading. Data represent the mean+s.d. from three experiments. (a) NF-kB activity; (b) AP-1 activity

by triptolide in AGS cells was also associated with upregulation of p21waf1/cip1, bax and G0/G1 arrest. However, in a previous study, Chang et al. (2001) showed that triptolide induced p53 protein expression but inhibited basal p21 expression and doxorubicin-mediated induction of p21. This discrepancy might be due to the dierential role of p21waf1/cip1 in modulation of apoptosis and cell survival is cell type and environment dependent (El-Deiry et al., 1994; Kobayashi et al., 1995; Zhu et al., 2000). Our results suggest that upregulation of p21waf1/cip1 is associated with the modulation of apoptosis in gastric cancer cells. However, the exact role of p21waf1/cip1 in triptolide-induced apoptosis needs further investigation. We have shown that triptolide-induced apoptosis was controlled through caspase activation, especially caspase 3. Caspase 3 or CPP32 is one of the important candidates for cell death-inducing proteases that cleave PARP and other vital proteins (Patel et al., 1996). In this study, triptolide induced caspase 3 activation and PARP cleavage in parallel with the induction of apoptosis. Furthermore, caspase inhibitors could eectively block the cleavage of caspase 3. However, both zAVD-fmk and DEVD-fmk could only partially inhibit triptolide-induced apoptosis in AGS cells. These results indicated that the activation of caspase 3 was involved in triptolide-induced apoptosis. But it is likely that triptolide could also induce apoptosis partly via a caspase-independent pathway.

A recent study showed that triptolide sensitized nonsmall cell lung cancer cell line, ®brosarcoma cell line and breast cancer cell line to TNF-a-induced apoptosis through inhibition of NF-kB (Lee et al., 1999). In our study, we demonstrated that triptolide alone could inhibit NF-kB transcriptional activity dose-dependently after 6 h treatment in AGS cells. Since most cells are still alive at that time, this eect may not necessarily associate with triptolide-induced apoptosis. In most cells, NF-kB is inactive due to its cytoplasmic sequestration through interaction with inhibitor proteins IkBs. Upon activation of the complex, IkB sequentially undergoes phosphorylation, ubiquitination, and degradation. However, in our system, we showed that the modulation of NF-kB transactivation by triptolide was not related to IkB phosphorylation and degradation. We also demonstrated that triptolide inhibited NF-kB transcriptional activity but not binding activity. Recent studies suggest that transactivation of NF-kB is enhanced through the phosphorylation of p65 at Ser-276 and Ser-529 (Zhong et al., 1998; Wong and Baldwin, 1998). Lee et al. (1999) showed that triptolide inhibited transactivation of both the TA1 and TA2 regions of p65. These results suggest that triptolide might inhibit activation of NF-kB by a novel mechanism other than cytoplasmic sequestration of IkB and inhibition of NF-kB binding activity. NF-kB signaling cross-talks closely with that of p53 (Webster et al., 1999). To determine whether p53 status plays any role on the eect of triptolide on NF-kB activation, we co-transfected NF-kB reporter with a dominant negative p53 into AGS cells. Our results showed that forced suppression of normal p53 function by dominant-negative p53 markedly enhanced NF-kB activation in AGS cells. We then treated the cells with triptolide after alteration of p53 status. We demonstrated deletion of p53 function by dominant negative p53 in AGS cells resulted in abolishment of inhibition of NF-kB activation by triptolide. These results indicate that wild-type p53 status is required to suppress the NF-kB transcriptional activity. This observation is supported by some other studies, such as Nishizaki et al. (2000). They showed that overexpression of wildtype p53, but not mutant, inhibited NF-kB activity, and therefore induced apoptosis in human colon cancer cells (Nishizaki et al., 2000). Consistent with our results, we also showed that triptolide had dierential eect on MKN-28 and SGC-7901 gastric cancer cells with mutant p53, instead of suppression, it increased the transcriptional activity of NF-kB. The activator protein-1 (AP-1) is a ubiquitous factor that regulates the transcription of genes involved in cell growth and proliferation in response to external stimuli such as growth factors, phorbol esters and ionizing radiation (Angel et al., 1991). AP-1 has been associated with in¯ammation and neoplastic transformation. Inhibitors of AP-1 activation block the neoplastic transformation response (Li et al., 1997). We report for the ®rst time that triptolide inhibited AP-1 transcriptional activity in gastric cancer cell AGS and MKN-45 containing wild-type p53, whereas increased AP-1

8015

Oncogene


8016

activity in MKN-28 and SGC-7901 with mutant p53. However, whether the dierential modulation of AP-1 transcriptional activity contributes to the anti-tumor and anti-in¯ammatory eect by triptolide needs further investigation. In summary, triptolide is a novel drug with potent anti-in¯ammatory and anti-proliferative eects on gastric cancer cells. Triptolide induced apoptosis in gastric cancer cells through a p53-dependent pathway. In addition, triptolide inhibited NF-kB and AP-1 activation in cells with wild-type p53, forced alteration of p53 status suppressed triptolide-induced apoptosis and modulation of NF-kB. Our results suggest that functional p53 is required for the anti-in¯ammatory and anti-tumor eect by triptolide in gastric cancer cells.

Materials and methods Cell culture and drug treatment AGS bearing wild-type p53 was purchased from the American Type Culture Collection (ATCC, Rockville, MD, USA). MKN-28, which has missense mutation in p53 (codon 251, isoleucine to leucine), and MKN-45 bearing wild-type p53 were purchased from RIKEN (The Institute of Physical and Chemical Research) Cell Bank, Japan. SGC-7901 was kindly donated by Professor SD Xiao (Shanghai Second Medical University). Cells were maintained in RPMI-1640 containing 10% fetal bovine serum (FBS), 100 U ml71 penicillin, 100 mg ml71 streptomycin (Gibco BRL, Life Technologies, NY, USA). The stock solution of triptolide (90 mM) in ethanol was a kind gift from Dr Dan Yang (Department of Chemistry, University of Hong Kong). The phosphorothioate antisense oligonucleotide for p53 and the control CG-matched randomized-sequence phosphorothioate oligonucleotides were purchased from Biognostik (GoÈttingen, Germany). MTT assay About 5000 cells per well were grown in 96-well microtitre plates and incubated overnight in 100 ml of culture medium. Cells were then treated with dierent concentrations of triptolide for ®xed time intervals. Ten ml MTT (Fluka, Buchs, Switzerland) labeling reagent (®nal concentration 0.5 mg/ml) was added into each well and the cells were incubated for another 4 h at 378C. The supernatant was removed, and 100 ml of 0.04 M hydrochloric acid in isopropanol was added to each well. A micro ELISA reader (BioRad, CA, USA) measured the absorbency at a wavelength of 595 nm. Acridine orange (AO) staining Single cell suspensions were ®xed in 1% formalin/PBS and stained with acridine orange (AO, 10 mg ml71, Sigma, St. Louis, USA). Cells were spotted on glass slides and visualized under a UV ¯uorescence microscope. A minimum of 300cells/®eld was counted for apoptotic index. Flow cytometry Cells were collected and ®xed in ice-cold 70% ethanol in phosphate buered saline (PBS) and stored at 7208C. After resuspension, 100 ml RNAase I (1 mg ml71) and 100 ml Oncogene

propidium iodide (PI, 400 mg ml71, Sigma, USA) were added and incubated at 378C for 30 min. Sample analysis of samples was performed by ¯ow cytometry (Coulter Epics, XL, UK). The cell cycle phase distribution was calculated from the resultant DNA histogram using Multicycle AV software (Phoenix Flow System, San Diego, CA, USA). The apoptotic cells were observed as a subdiploid or `pre-G1' peak. DNA fragmentation analysis As described previously, DNA fragmentation was analysed with some modi®cations (Grant et al., 1992). Brie¯y, cell pellets were lysed in 10 mM Tris-HCL (pH 7.4) buer containing 25 mM EDTA, 0.5% SDS and 0.1 mg ml71 proteinase K (Sigma) and incubated at 508C for 12 ± 18 h. DNA was extracted with an equal volume of phenol/ chloroform/isoamyl alcohol (25 : 24 : 1) and precipitated with two volumes of ice-cold absolute ethanol and 1/10 volume 3M sodium acetate. Equal amounts (10 mg per well) of DNA were electrophoresed in 1.8% agarose gels impregnated with ethidium bromide (0.1 mg ml71) for 2 h at 80 V. DNA fragments in the form of a laddering pattern were visualized by ultraviolet transillumination. Preparation of cytoplasmic and nuclear extract Nuclear and cytoplasmic extracts were prepared as described by Dignam et al. (1983). Con¯uent cells in 10-cm dishes were treated for various times with the indicated eectors. Cells were then washed two times with ice-cold phosphate-buered saline and resuspended in 400 ml of buer A (containing 10 mM HEPES at pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 0.5 mM PMSF, 1 mg/ml leupeptin, 1 mg/ml aprotinin, and 1 mg/ml pepstatin A). The cells were allowed to swell on ice for 15 min, lysed gently with 12.5 ml of 10% Nonide P-40, and centrifuged at 2000 g for 10 min at 48C. The supernatant was collected and used as the cytoplasmic extracts. The nuclei pellet was re-suspended in 40 ml of buer C (20 mM HEPES, pH 7.9, containing 1.5 mM MgCl2, 450 mM NaCl, 25% glycerol, 0.2 mM EDTA, 0.5 mM DTT, 0.5 mM PMSF, 1 mg/ml leupeptin, 1 mg/ml aprotinin, 1 mg/ml pepstatin A), agitated for 30 min at 48C, and the nuclear debris was spun down at 20 000 g for 15 min. The supernatant (nuclear extract) was collected, frozen in liquid nitrogen and stored at 7808C until ready for analysis. Protein concentration was determined using Bradford reagent. Western blotting Twenty mg of protein were resolved and separated by electrophoresis on a 10% denaturing SDS gel. Proteins were electroblotted onto nitrocellulose membranes. Detection was conducted by immuno-staining using speci®c primary antibodies and horseradish peroxidase-conjugated anti-IgG antibody. The protein bands were visualized by the enhanced chemiluminescence (ECL) assay (Amersham) following manufacturer's instructions. Mouse monoclonal antibodies against p53 (Ab-6), p21waf1/cip1 (Ab-1), c-myc (Ab-1) and peroxidase-conjugated anti-mouse and anti-rabbit IgG were purchased from CalBiochem (La Jolla, CA, USA). Monoclonal antibodies against bc1-2 (100) and bax (B-9) were purchased from Santa-Cruz Biotechnology (CA, USA). Rabbit polyclonal antibody against CPP-32 and mouse monoclonal antibody against PARP (c-20) were purchased from Pharmingen (San Diego, CA, USA). Rabbit polyclonal antibody against IkB-a and phosphor IkB-a were purchased from New England Biolabs (Boston, MA, USA).


Caspase-3 activity assay Caspase activity was determined using the Clontech ApoAlert CPP32 Assay Kit according to the manufacturer's instructions (Clontech, Palo Alto, CA, USA). These assays measured the cleavage of a speci®c colorimetric caspase substrate, DEVD ± rNA (Asp-glu-val-asp-r-nitroanilide). Cells were plated into 60 mm dishes at 26106 cells/dish and cultured for 24 h before treatment. After treatment, cells were collected by scraping in cold PBS, centrifuged 2000 r.p.m., 8 min) and lysed in the cell lysis buer provided in the kit on ice for 10 min. Extracts were then frozen and maintained at 7708C until the time of assay. At that time, the extracts were thawed and reacted with an equal volume of 26reaction buer containing DTT (10 mM) and the colorimetric caspase substrate (DEVD ± rNA). The mixtures were maintained in a water bath at 378C for 45 min and then analysed in a spectrophotometer at 405 nm. Transient transfection and luciferase assay NF-kB luciferase and AP-1 reporter plasmid were constructed individually by inserting the IL-6 promoter region (7194 to 7114 containing four repeats of NF-kB binding sites) and collagenase promoter region (773 to +67 containing one AP-1 binding site) into a luciferase reporter vector pGL-3-basic (Promega Corp., Madison, WI, USA) (Li et al., 1997; Angel et al., 1987). Expression plasmids pCMV ± p53mt135, which expresses a dominant-negative mutant, were purchased from Clontech (CA, USA). For transient transfection experiments, AGS cells were seeded in 12-well plates to 70 ± 80% con¯uence. The cells were transfected with 0.6 mg/ well NF-kB or AP-1 reporter plasmid with the presence or absence of dierent concentrations of pCMV ± p53mt135 (0.1 ± 0.6 mg/well) using lipofectamine 2000. pRL ± CMV vector (0.01 mg/well, Promega) was cotransfected as internal control. After transfection for 6 h, cells were changed to normal medium and allowed to recover overnight, and then treated with dierent concentrations of triptolide for another 6 h. Cells were harvested for analysis of luciferase activity using the dual luciferase reporter assay system (Promega). Luminence was measured by a LB 9507 Luminometer

(EG&G B, Australia). The Fire¯y luciferase luminescence activity were normalized to the control Renilla luciferase activity and expressed as a relative ratio.

8017

Electrophoretic mobility shift assay Six mg of nuclear proteins were incubated with 1 mg each of poly(dI.dC) and poly(dG.dC) in the presence of 10 fmol of [g-32P] ATP end-labeled double-stranded Ap1 probe (5'-CGC TTG ATG AGT CAG CCG GAA-3' Promega) or Nf-kB B (5'-AGT TGA GGG GAC TTT CCC AGG C-3' Santa Cruz) for 30 min at room temperature in a total volume of 20 ml. Oligonucleotide competition experiments were performed in the presence of 50-fold excess of cold AP1 and NF-kB oligos. Supershift experiments were done by preincubating nuclear extract proteins with 2 ml antibody for 1 h at 48C before the addition of labeled DNA probe. DNA complexes were resolved from free probe with 4% nondenaturing polyacrylamide gels in 0.56 Tris-borate-EDTA (pH 8.3) and visualized by ¯uorography. Polyclonal antibody c-Jun (N) X for AP1 and NF-kB p65 (c-20) for NF-kB from Santa Cruz Biotechnology (Santa Cruz, CA, USA) was used for supershift experiments. Statistical analysis The data shown were mean values of at least three dierent experiments and expressed as means+s.d. Student's t-test was used for comparison. A P-value of less than 0.05 is considered statistically signi®cant.

Abbreviations TWHf, Tripterygium wilfordii Hook f; NF-kB, nuclear factor-kB; AP-1, Activator Protein-1

Acknowledgments This study is supported by Gastroenterological Research Fund, University of Hong Kong.

References Amundson SA, Myers TG and Fornace AJ. (1998). Oncogene, 17, 3287 ± 3299. Angel P, Baumann I, Stein B, Delius H, Rahmsdorg HJ and Herrlich P. (1987). Mol. Cell. Biol., 7, 2256 ± 2266. Angel P and Karin M. (1991). Biochem. Biophys. Acta., 1072, 129 ± 157. Calvert AH, Ghokul S, Al-Azraqi A, Wright J, Lind M and Bailey N. (1999). Semin. Oncol., 16, S90 ± S94. Chang WT, Kang JJ, Lee KY, Wei K, Anderson E, Gormare S, Ross JA and Rosen GD. (2001). J. Biol. Chem., 276, 2221 ± 2227. Dignam JD, Lebovitz RM and Roeder RG. (1983). Nucleic Acids Res., 11, 1475 ± 1489. El-Deiry WS, Harper JW, O'Connor PM, Velculescu VE, Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill DE, Wang Y, Wiman KG, Mercer WE, Kastan MB, Kohn KW, Elledge SJ, Kinzler KW and Vogelstein B. (1994). Cancer Res., 54, 1169 ± 1174. El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein B. (1993). Cell., 75, 817 ± 825.

Farmer G, Bargonetti J, Zhu H, Friedman P, Prywes R and Prives C. (1992). Nature, 358, 83 ± 86. Grant S, Jarvis WD, Swerdlow PS, Turner AJ, Traylor RS, Wallace HJ, Lin PS, Pettit GR and Gewirtz DA. (1992). Cancer Res., 52, 6270 ± 6278. Jiang X. (1994). Pediatr. Nephrol., 8, 343 ± 344. Kobayashi T, Consoli U, Andree M, Shiku H and Zhang W. (1995). Oncogene, 11, 2311 ± 2316. Kupchan SM, Court WA, Dailey RG, Gilmore CJ and Bryan RF. (1972). Amer. Chem. Soc., 94, 7194 ± 7195. Lane DP. (1992). Nature, 358, 15 ± 16. Lee JM and Bernstein A. (1995). Cancer Metastasis Rev., 14, 149 ± 161. Lee KY, Chang WT, Qiu DM, Kao PM and Rosen GD. (1999). J. Biol. Chem., 274, 13451 ± 13455. Li JJ, Westergaad D, Ghosh P and Colburn NH. (1997). Cancer Res., 57, 3569 ± 3576. Linke SP, Clarkin KC and Wahl GM. (1997). Cancer Res., 57, 1171 ± 1179.

Oncogene


8018

Oncogene

Lowe SW, Ruley HE, Jacks T and Housman DE. (1993). Cell, 74, 957 ± 967. Midgley CA, Owens B, Briscoe CV, Thomas DB, Lane DP and Hall PA. (1995). J. Cell Sci., 108, 1843 ± 1848. Miyashita T and Reed JC. (1995). Cell, 80, 293 ± 299. MuÈller H and Eppenberger U. (1996). Anticancer Res., 16, 3845 ± 3848. Nishizaki M, Roth JA and Tanaka N. (2000). Oncogene, 19, 726 ± 736. Patel T, Gores GJ and Kaufmann SH. (1996). FASEB J., 10, 587 ± 597. Penninger JM and Kroemer G. (1996). Adv. Immunol., 68, 51 ± 144. Pu LX and Zhang TM. (1990). Acta. Pharmacologica Sinica., 11, 76 ± 79. Qin WZ, Liu CH and Yang SM. (1981). Chin. Med. J. (Engl.), 94, 827 ± 830. Qiu D, Zhao G, Aoki Y, Shi L, Uyei A, Nazarian S, Ng JC and Kao PN. (1999). J. Biol. Chem., 274, 13443 ± 13450. Shaw P, Bovey R, Tardy S, Sahli R, Sordat B, Costa J. (1992). Proc. Natl. Acad. Sci. USA, 89, 4495 ± 4499. Tao X, Sun Y, Dong Y, Xiao YL, Hu DW, Shi YP, Zhu QL, Huan D and Zhang NZ. (1989). Chin. Med. J. (Engl.), 102, 327 ± 332. Tetu B, Brisson J, Plante V and Bernard P. (1998). Mod. Pathol., 11, 823 ± 830. Vasey PA, Jones NA, Jenkins S, Dive C and Brown R. (1996). Mol. Pharmacol., 50, 1536 ± 1540.

Wang D and Baldwin AS. (1998). J. Biol. Chem., 273, 29411 ± 29416. Webster GA and Perkins ND. (1999). Mol. Cell. Biol., 19, 3485 ± 3495. Wei YS and Adachi L. (1991). Acta. Pharmacologica Sinica., 12, 406 ± 410. Weller M. (1998). Cell Tissue Res., 292, 435 ± 445. Yang SX, Gao HL, Xie SS, Zhang WR and Long ZZ. (1992). Int. J. Immunopharmacol., 14, 963 ± 969. Yang Y, Liu Z, Tolosa E, Yang J and Li L. (1998). Immunopharmacology, 40, 139 ± 149. Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A and Oren M. (1991). Nature, 352, 345 ± 347. Zhao G, Vaszar LT, Qiu D, Shi L and Kao PN. (2001). Am. J. Physiol. Lung Cell Mol. Physiol., 279, L958 ± L966. Zhan Q, Alamo I, Yu K, Boise LH, Cherney B, Tosato G, O'Connor PM and Fornace AJ. (1996). Oncogene, 13, 2287 ± 2293. Zhang LX, Yu FK, Zhang QY, Fang Z and Pan DJ. (1990). Acta. Pharmaceutica Sinica., 25, 573 ± 577. Zhong H, Voll RE and Ghosh S. (1998). Mol. Cell. Biol., 1, 661 ± 671. Zhu GH, Wong BC, Eggo MC, Yuen ST, Lai KC and Lam SK. (1999). Dig. Dis. Sci., 44, 2020 ± 2026. Zhu GH, Wong BC, Slosberg ED, Eggo MC, Ching CK, Yuen ST, Lai KC, Soh JW, Weinstein IB and Lam SK. (2000). Gastroenterology, 118, 507 ± 514.