P21WAF1/CIP1 is dispensable for G1 arrest, but ... - Nature

Oncogene (2004) 23, 21–29

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00 www.nature.com/onc

P21WAF1/CIP1 is dispensable for G1 arrest, but indispensable for apoptosis induced by sodium butyrate in MCF-7 breast cancer cells Vale´rie Chopin1, Robert-Alain Toillon1, Nathalie Jouy2 and Xuefen Le Bourhis*,1 1

Equipe facteurs de croissance, Laboratoire de Biologie du De´veloppement (UPRES 1033), IFR 118, Universite´ des Sciences et Technologies de Lille, Villeneuve d’Ascq Cedex 59655, France; 2IFR 114, IMPRT, ‘Institut de Me´decine Pre´dictive et de Recherche The´rapeutique’, Institut de Recherche sur le Cancer de Lille, Lille Cedex 59045, France

Sodium butyrate (NaB) has been proposed as a potential anticancer agent. However, its mechanism of action is not totally elucidated. Here, we showed that NaB-induced cell cycle arrest and apoptosis were associated with an increase of P21waf1/cip1 in MCF-7 breast cancer cells. This increase was more important in the nuclei, as revealed by immunofluorescence analysis. Transient transfections of MCF-7 cells with p21 deficient for interaction with CDK, but not with p21 deficient for interaction with PCNA (p21PCNA ), abrogated NaB-induced cell cycle arrest. This indicated that cell cycle blockage involved the interaction of P21waf1/cip1 with CDK. However, P21waf1/cip1 was dispensable, since p21 antisense did not modify cell cycle arrest. On the other hand, NaB-induced apoptosis was abolished by p21 antisense or p21PCNA . In addition, NaB decreased PCNA levels, but increased the association of PCNA with P21waf1/cip1. These results suggested that NaB-induced apoptosis required P21waf1/cip1 and its interaction with PCNA. Oncogene (2004) 23, 21–29. doi:10.1038/sj.onc.1207020 Keywords: apoptosis; breast cancer; butyrate; cell cycle; PCNA; P21waf1/cip1

Introduction Breast cancer is the first cancer in the European community, representing about 24% of all cancer cases. This pathology is currently controlled through surgery and/or radiotherapy, and is frequently supported by adjuvant chemo- or hormonotherapies. Unfortunately, these classical treatments are hampered by unwanted side effects and, mostly importantly, the development of tumor resistance. Therefore, there is an urgent need for novel and effective therapies against breast cancer. In this perspective, butyrate is proposed as a potential anticancer agent. Butyrate is a short-chain fatty acid produced during fermentation of fiber by endogenous intestinal bacteria. *Correspondence: X Le Bourhis; E-mail: [email protected] Received 21 May 2003; revised 16 July 2003; accepted 16 July 2003

It is also present in fruits, vegetables, and fat milk. The sodium salt of this fatty acid, sodium butyrate (NaB), has been reported to induce differentiation and reversible cell cycle arrest of both normal and tumor cells (Velazquez et al., 1996; Schwartz et al., 1998; Wang et al., 1999; Demary et al., 2001; Terao et al., 2001). More recently, NaB has been shown to induce apoptosis in a variety of tumor cells including colonic carcinoma, retinoblastoma, breast, hepatic, and hematopoietic malignant cells. In addition, Phase I pharmacokinetic studies and treatment of leukemia, hemoglobinopathies and solid tumors with NaB, or analogues, which have better pharmacodynamic properties, have shown no severe toxicity (Gilbert et al., 2001; Patnaik et al., 2002). Although the mechanisms by which NaB mediates its effects are not well understood, it has been described that NaB decreases the expression of cyclin D1 in epidermoid carcinoma cells (Lallemand et al., 1996), and increases the expression of IGFBP-3 in breast cancer cells (Walker et al., 2001; Tsubaki et al., 2002). NaB also regulates the expression of members of the Bcl-2 family in leukemia and colon cancer cells (Mandal et al., 1997; Sawa et al., 2001). In molecular terms, NaB is an inhibitor of histone deacetylase (HDAC), leading to hyperacetylation of chromatin components such as histone and nonhistone proteins, and alterations in gene expression. The effects of HDAC inhibitors on gene expression are highly selective, leading to transcriptional activation of certain genes such as P21waf1/cip1, but repression of others (Richon et al., 2000; Sowa and Sakai, 2000). Indeed, NaB activates the transcription of p21waf1/cip1 gene through Sp1 sites of the p21waf1/cip1 promoter (Nakano et al., 1997; Sowa et al., 1999). In a variety of cancer cells such as osteosarcoma, colon carcinoma, glioma, and lung cancer cells, NaB dramatically increases P21waf1/cip1 expression (Ito et al., 2001; Pellizzaro et al., 2001). P21waf1/cip1 has been shown to play an essential role in G1 and G2 arrests (Bunz et al., 1998; Brugarolas et al., 1999) by inhibiting cyclinE-CDK2 and cyclinBcdc2 activities, respectively. P21waf1/cip1 is also able to associate with the proliferating cell nuclear antigen (PCNA). PCNA is implicated in DNA replication and repair by forming a sliding platform that can mediate the interaction of numerous proteins with DNA

P21WAF1/CIP1 is implicated in NaB-induced apoptosis V Chopin et al

22

(Warbrick, 2000). P21waf1/cip1 binding to PCNA is shown to suppress PCNA-dependent replication, but the ability of p21waf1/cip1 to suppress PCNA-dependent repair is controversial (Shivji et al., 1994; Cooper et al., 1999; Mattock et al., 2001). P21waf1/cip1 may bind to both active cyclin/CDK and PCNA simultaneously, suggesting that protein interaction with PCNA may be a mechanism to coordinate DNA replication and repair. Other proteins such as Gadd45 and p33ING1b, involved in the regulation of cell cycle and apoptosis, are also reported to bind to PCNA (Hall et al., 1995; Scott et al., 2001). In addition to being an inhibitor of cell cycle progression, P21waf1/cip1 acts also as a modulator of apoptosis. However, the exact implication of P21waf1/cip1 in apoptosis remains controversial. It has been reported that downregulation of P21waf1/cip1 in hepatocellular carcinoma cells renders them more sensitive to interferon-gamma-induced apoptosis (Detjen et al., 2003); suppression of P21waf1/cip1 by transfection of p21 antisense or homologous recombination increases the apoptosis induced by cytotoxic drugs or gamma irradiation in glioma and colon carcinoma cell lines (Tian et al., 2000; Mahyar-Roemer and Roemer, 2001; Han et al., 2002). Moreover, P21waf1/cip1 is complexed with procaspase 3 to suppress Fas-mediated apoptosis (Suzuki et al., 2000). In apoptotic cells, P21waf1/cip1 is found to be cleaved by caspase 3 (Zhang et al., 1999; Xiang et al., 2002). All these results indicate that P21waf1/ cip1 protect cells from apoptosis. In other situations, P21waf1/cip1 can induce apoptosis. Infection of cervical cancer cells with the sense p21 adenovirus inhibits cell growth by inducing apoptosis (Tsao et al., 1999). Overexpression of P21waf1/cip1 has been shown to enhance the induction of apoptosis by cisplatin in glioma, ovarian carcinoma and hepatoma cell lines (Qin and Ng, 2001). Ectopic expression of P21waf1/cip1 in p21 / hepatocytes restored the apoptotic response to deoxycholic acid (Qiao et al., 2002). Overexpression of P21waf1/ cip1 in mammary tumor cells using a retroviral delivery system results in an increased apoptosis (Shibata et al., 2001). More recently, it has been shown that P21waf1/cip1 can act as a direct modulator of transcription. P21waf1/cip1 associates with transcription factors such as E2F or cMyc to suppress cell cycle progression (Delavaine and La Thangue, 1999; Gartel and Shchors, 2003). P21waf1/cip1 can also form a complex with transcription coactivators, including p300 and CBP, to enhance their function. CBP and P300 are histone acetyltransferases which activate certain transcription factors such as AP-1, E2F or NFkB subunit (Perkins et al., 1997; Snowden et al., 2000; Gregory et al., 2002). Since P21waf1/cip1 plays an important role in the regulation of cell cycle and apoptosis, and NaB induces the expression of p21waf1/cip1 in many cancer cell lines, we examined the implication of P21waf1/cip1 in NaB-induced cell cycle arrest and apoptosis in breast cancer cells. We transiently transfected cells with a pcDNA plasmid containing different sequences of p21 : p21 wild type (p21WT), p21 antisense (p21AS), p21 deficient for Oncogene

interaction with PCNA (p21PCNA ) or with CDK (p21CDK ). We demonstrated that P21waf1/cip1 was implicated in NaB-induced cell cycle arrest and apoptosis; moreover, P21waf1/cip1 was indispensable for apoptosis induction, but not for cell cycle arrest.

Results NaB-induced G1 arrest and apoptosis in MCF-7 cells The effects of NaB in cell cycle progression and apoptosis induction were determined after 24 or 48 h of treatment. As shown in Figure 1a, after 24 h of treatment, NaB increased the percentage of cells in the G2/M phase, and decreased that of cells in the S phase. After 48 h of treatment, the majority of cells were blocked in the G1 phase. In addition, after 24 h of treatment (Figure 1b), 1 mm NaB did not induce apoptosis. After 48 h of treatment, 1 mm NaB slightly induced apoptosis, whereas 2.5 mm NaB induced eightfold of apoptosis compared to untreated cells. Since 2.5 mm NaB efficiently induced cell cycle arrest and

Figure 1 Effects of NaB on cell cycle and apoptosis in MCF-7 cells. Cells were treated with NaB during 24 or 48 h. Cell cycle analysis was performed by flow cytometry of propidium iodidestained nuclei (a). Apoptosis was determined after Hoechst staining (b). Data shown are representative of three separate experiments. Each bar represents the mean7s.d. in a triplicate assay. The stars in (c) indicate apoptotic nuclei


23

apoptosis, this concentration was used for the following study. NaB increased nuclear P21waf1/cip1 levels in MCF-7 cells It has been reported that NaB-induced cell cycle arrest and apoptosis are associated with an increase of P21waf1/cip1 levels in a number of cancer cells. However, the role of P21waf1/cip1 in the control of these processes is controversial. Therefore, we first performed immunoblots to determine the levels of P21waf1/cip1 in MCF-7 breast cancer cells treated with NaB. Figure 2a showed that NaB increased P21waf1/cip1 as early as 12 h. Moreover, immunofluorescence analysis (Figure 2b) showed that when MCF-7 cells were treated with 2.5 mm NaB for 24 h, the staining of P21waf1/cip1 was increased. This increase was particularly important in the nuclei, since the FITC staining of P21waf1/cip1 was overlapped by Hoechst staining.

Figure 2 Expression of P21waf1/cip1 in MCF-7 cells. (a) Western blot of P21waf1/cip1. Preconfluent cells were treated with 2.5 mm NaB for different periods of time. Proteins (30 mg) were electrophoresed and Western blots performed, as described in Materials and Methods. Loading and transfer of equal amounts of protein were confirmed by immunodetection of actin. (b) Immunofluorescence analysis of intracellular localization of P21waf1/cip1. Cells were treated without (left panel) or with (right panel) 2.5 mm NaB for 24 h. Hoechst staining and immunofluorescence analysis were performed as described in Materials and methods

Validation of mutant p21 plasmids To further investigate the implication of P21waf1/cip1 in cell cycle arrest and apoptosis, we used different constructs of p21 : p21 wild type (p21WT), p21 antisense (p21AS) or p21 deficient for the interaction with PCNA (p21PCNA ) or with CDK (p21CDK ). First, to verify the functional properties of the plasmids p21AS and p21WT, MCF-7 cells were transiently transfected with pcDNA plasmids containing p21AS or p21WT. Immunoblotting showed that transfection with p21AS totally abolished the expression of P21waf1/cip1 protein in MCF-7 cells treated with or without NaB, whereas transfection with p21WT slightly increased P21 levels compared to empty vector-transfected cells (Figure 3a). To verify the association of P21waf1/cip1 with PCNA or CDK, MCF-7 cells were transiently cotransfected with a pcDNA plasmid containing p21PCNA or p21CDK , and a pcDNA plasmid containing murin CD80 sequence. The transfected cells were selected and lysed. Immunoprecipitation of P21waf1/cip1 was performed and followed by immunoblotting with anti-PCNA and anti-

Figure 3 Functional verification of pcDNA plasmids containing different sequences of P21waf1/cip1. (a) Western blot analysis of P21waf1/cip1 in transfected MCF-7 cells. Cells were transiently transfected with p21 wild type (p21WT) or p21 antisense (p21AS), and treated with or without 2.5 mm NaB for 24 h. Then, proteins were electrophoresed and analysed by Western blot using anti-P21waf1/cip1 antibody. Loading and transfer of equal amounts of protein were confirmed by immunodetection of actin. (b) Immunoprecipitation analysis of P21waf1/cip1 association with CDK2 or PCNA. Cells were transiently transfected with pcDNA plasmids containing p21 deficient for the interaction with PCNA (p21PCNA ) or with CDK (p21CDK ), and a pcDNA plasmid containing murin CD80. Cells were treated with or without 2.5 mm NaB for 24 h. Transfected cells were selected using anti-murin CD80 antibody, as described in Materials and methods. Then, cells were lysed, and proteins were immunoprecipitated with antiP21waf1/cip1 antibody. Immunoprecipitates were analysed by Western blot, using anti-CDK2, anti-PCNA or anti-P21waf1/cip1 antibodies Oncogene


24

CDK2 antibodies (Figure 3b). In cells transfected with empty vector, native P21waf1/cip1 was able to interact with both PCNA and CDK2. In cells transfected with p21PCNA , much less PCNA was associated with P21waf1/cip1, whereas levels of CDK2 associated with P21waf1/cip1 were not modified compared to empty vectortransfected cells. On the contrary, in cells transfected with p21CDK , levels of P21waf1/cip1-associated CDK2 were dramatically decreased, while levels of P21waf1/cip1associated PCNA were not modified. The low levels of CDK2 in p21CDK-transfected cells and that of PCNA in p21PCNA-transfected cells were probably due to the presence of native P21waf1/cip1. Overall, our results confirmed the functional properties of these different p21 constructs. p21WT slightly increased the levels of P21waf1/cip1, while p21AS abrogated the expression of native p21waf1/cip1; P21PCNA and P21CDK seemed not able to interact with PCNA and with CDK2, respectively. Interaction of P21waf1/cip1 with CDK was responsible for NaB-induced cell cycle arrest To investigate the role of P21waf1/cip1 in NaB-induced cell cycle arrest in MCF-7 cells, we transiently cotransfected cells with a pcDNA plasmid containing p21WT or p21AS, p21PCNA , p21CDK , and a pcDNA plasmid containing the murin CD80 sequence. The transfected cells were selected using antimurin CD80 antibody conjugated with FITC, and analysed by flow cytometry (Figure 4). In empty vector-transfected cells, 24 h of NaB treatment decreased the percentage of cells in S phase and accumulated cells in G2/M phases. This result was in accordance with our data obtained in nontransfected cells (Figure 1a). Transfection of cells with p21CDK abrogated NaB-induced G2/M arrest compared to untreated cells, suggesting that NaB-induced cell cycle arrest involved the interaction between P21waf1/cip1 and CDK. However, when cells were transfected with

Figure 4 Cell cycle analysis of transfected MCF-7 cells. Cells were transiently cotransfected with a pcDNA plasmid containing p21AS or p21PCNA , p21CDK , and a pcDNA plasmid containing murin CD80 sequence. The cells were treated with 2.5 mm NaB for 24 h. Transfected cells were selected by immunoprecipitation of cells using anti-CD80 murin antibody, as described in Materials and methods. Cell cycle was determined by flow cytometry. Data shown are representative of three separate experiments. Each bar represents the mean7s.d. in a triplicate assay Oncogene

p21WT or p21PCNA , NaB-induced cell cycle arrest was not modified. These later results suggested that native P21waf1/cip1 was sufficient to inhibit cell cycle in NaB-treated cells, and that this cell cycle arrest did not implicate the interaction between P21waf1/cip1 and PCNA. Interestingly, when MCF-7 cells were transiently transfected with p21AS, NaB always induced G2/M arrest. It is thus possible that other CDK inhibitors such as P16ink4 and P27kip1 could inhibit cell cycle progression in the absence of P21waf1/cip1. To verify this hypothesis, we evaluated the levels of P16ink4 and P27kip1 in MCF-7 cells by Western blot analysis in nontransfected cells. Figure 5 showed that NaB treatment increased both P16ink4 and P27kip1 levels. NaB-induced apoptosis required the presence of P21waf1/cip1, and its interaction with PCNA To investigate the implication of P21waf1/cip1 in the induction of apoptosis in MCF-7 cells, we transiently cotransfected cells with a pcDNA plasmid containing p21WT or p21AS, p21PCNA , p21CDK , and a pEGFP-Cl plasmid (Figure 6). When cells were transfected with p21WT or p21CDK , the induction of apoptosis by NaB was not modified compared to empty vector-transfected cells. On the contrary, when cells were transfected with p21AS or p21PCNA , NaBinduced apoptosis was abrogated. These results suggested that the induction of apoptosis by NaB was under the control of P21waf1/cip1, and implicated the interaction between PCNA and P21waf1/cip1. To further investigate the interaction between P21waf1/ cip1 and PCNA in NaB-induced apoptosis, we performed Western blots of PCNA in MCF-7 cells. In total extracts, NaB decreased the levels of PCNA (Figure 7a). PCNA is mainly known as an auxiliary factor for DNA polymerase that is essential for both DNA replication and repair. During DNA replication and repair, nuclear PCNA is tightly associated with chromatins, and cannot be extracted with nonionic detergents such as Triton X-100 (triton-insoluble form).

Figure 5 Western blot analysis of P16ink4 and P27kip1 in MCF-7 cells. Preconfluent cells were treated with 2.5 mm NaB for different periods of time. Proteins (30 mg) were electrophoresed and Western blots performed, as described in Materials and methods. Loading and transfer of equal amounts of protein were confirmed by the immunodetection of actin


25

Figure 6 Apoptosis analysis of transfected MCF-7 cells. Cells were transiently cotransfected with a pcDNA plasmid containing p21WT or p21AS, p21PCNA , p21CDK , and a pEGFP-Cl plasmid. Following transfection, cells were treated with or without 2.5 mm NaB for 24 h. Apoptosis of transfected cells (GFP positive) was determined after Hoechst staining. Data shown are representative of three separate experiments. Each bar represents the mean7s.d. in a triplicate assay

In other situations, nuclear PCNA is present in a tritonsoluble form (Bravo and Macdonald-Bravo, 1987; Toschi and Bravo, 1988; Morris and Mathews, 1989; Pagano et al., 1994). We then investigated whether NaB could modify the soluble properties of PCNA. We showed that, in untreated MCF-7 cells, PCNA levels were higher in soluble fraction than in insoluble fraction (Figure 7a). NaB decreased both triton-soluble and insoluble forms of PCNA, and rendered the levels of insoluble PCNA undetectable after 48 h of treatment. However, immunoprecipitation analysis showed that NaB increased the levels of PCNA associated with P21waf1/cip1 (Figure 7b). Taken together, all these results indicated that though NaB decreased PCNA levels, it favored the association of PCNA with P21waf1/cip1.

Figure 7 Effect of NaB on PCNA levels in MCF-7 cells. (a) Western blot analysis of PCNA. Preconfluent cells were treated with 2.5 mm NaB for different periods of time. Total extracts, triton-soluble or triton-insoluble extracts (30 mg) were electrophoresed and Western blots performed, as described in Materials and methods. Loading and transfer of equal amounts of protein were confirmed by immunodetection of actin. (b) Immunoprecipitation of P21waf1/cip1. Preconfluent cells were treated with or without 2.5 mm NaB for different periods of time. Proteins were immunoprecipitated with anti-P21waf1/cip1 antibody. Immunoprecipitates were analysed by Western blot using anti-PCNA or anti-P21waf1/cip1 antibodies

Pretreatment with NaB or roscovitine did not modify NaB-induced apoptosis NaB-induced apoptosis may result or not from cell cycle blockage. To test these two possibilities, we pretreated MCF-7 cells with roscovitine, which potently inhibits p34CDK1/cyclin B kinase (3 mm, 48 h) and NaB (1 mm, 72 h). Flow cytometry analysis showed that roscovitine accumulated cells in G1 and G2/M phases (data not shown), and that NaB increased the percentage of cells in G1 phase (Figure 1a). However, NaB-induced apoptosis was not modified by cell cycle blockage (Figure 8), indicating that the induction of apoptosis by NaB was independent of cell cycle distribution.

Discussion In this study, we have demonstrated that NaB inhibited the growth of MCF-7 breast cancer cells. The inhibition was due to cell cycle arrest in the G1 phase, and

induction of apoptosis. Moreover, NaB is a potent growth inhibitor not only for MCF-7 cells but also for hormono-dependent MCF-7ras and T47-D cells, and hormono-independent BT-20 cells (data not shown). Therefore, butyrate could affect breast cancer development both in the early (estrogen responsive) and advanced (estrogen resistant) stages. NaB is a member of the HDAC inhibitor family that appears to be selective in regulating gene expression. HDAC inhibitors such as NaB and trichostatin A increase the expression of the cyclin-dependent kinase inhibitor p21waf1/cip1, which may play a critical role in the regulation of cell cycle arrest and apoptosis (Sambucetti et al., 1999; Richon et al., 2000; Sowa and Sakai, 2000). In agreement with these data, we showed that NaBinduced growth inhibition was associated with an increase in the nuclear level of P21waf1/cip1 in MCF-7 cells. Moreover, P21waf1/cip1 was differently implicated in NaB-induced G1 arrest and apoptosis. Oncogene


26

Figure 8 Effect of cell cycle blockage on NaB-induced apoptosis in MCF-7 cells. To induce cell accumulation in the G1 phase, cells were pretreated with 3 mm roscovitine for 48 h (a), or 1 mm NaB for 72 h (b). After pretreatment, cells were incubated with 2.5 or 5 mm NaB for 24 h. Apoptosis was determined after Hoechst staining. Data shown is representative of three separate experiments. Each bar represents the mean7SD in a triplicate assay

It has been broadly described that N-terminal portion of P21waf1/cip1 interacts with cyclin/CDK complex, leading to G1 or G2 arrest. In accordance with this, we showed that NaB-induced cell cycle arrest was abrogated in MCF-7 cells transfected with p21 deficient for the interaction with CDK. However, the C-terminal domain of P21waf1/cip1 can also associate with PCNA and inhibit DNA replication (Rousseau et al., 1999). Here, we demonstrated that NaB-induced cell cycle arrest did not implicate the interaction of P21waf1/cip1 with PCNA, since NaB blocked the cell cycle in MCF-7 cells transfected with p21 deficient for interaction with PCNA. These results suggested that binding of P21waf1/cip1 to the cyclin–CDK complex was the principal mechanism of NaB-induced proliferation inhibition. However, in the absence of P21waf1/cip1 (when cells were transfected with p21 antisense), NaB still blocked cell cycle progression. This suggested that P21waf1/cip1 was dispensable in cell cycle arrest induced by NaB. In agreement with our results, Vaziri et al. (1998) have shown that NaB induces G1 arrest in mice p21 / fibroblasts. Moreover, Derjuga et al. (2001) showed that P21waf1/cip1 only binds to and inhibits a subpopulation of CDK2/cyclin E in Hela cells treated with NaB. The subpopulation of CDK2/cyclin E that is not bound to P21waf1/cip1 remains inactive, suggesting that NaB may affect other CDK inhibitors. Indeed, we showed that P27kip1 and P16ink4 were increased following NaB treatment in MCF-7 cells. Similar results have been reported in nonsmall lung cancer cells and normal mammary epithelial cells (Pellizzaro et al., 2001; Oncogene

Tsubaki et al., 2001). Besides CDK inhibitors, NaB has been reported to inhibit cell cycle progression via Tob-1, which interferes with CDK activity causing pRb hypophosphorylation (Della Ragione et al., 2001). Therefore, our results, together with the previously reported data, indicate that NaB-induced cell cycle arrest involves several molecules, depending upon cellular context. It is believed that prolonged cell cycle arrest may conduct cells to apoptosis. We showed that NaB-induced apoptosis in MCF-7 cells was independent of the G1phase blockage, suggesting that the underlying mechanisms of cell cycle blockage and induction of apoptosis occurred through different pathways. These results were in agreement with Galfi et al. (2002), who demonstrate that there is no correlation between proliferation inhibition and apoptosis induction by HDAC inhibitors TSA and NaB in gastrointestinal cell lines. We showed that P21waf1/cip1 was indispensable for NaB-induced apoptosis in MCF-7 cells, since transient transfection of p21waf1/cip1 antisense abolished the induction of apoptosis by NaB. The role of p21waf1/cip1 in apoptosis is still controversial. Although P21waf1/cip1 has been mainly reported to inhibit apoptosis, a number of other reports suggest that P21waf1/cip1 also possesses proapototic functions. These survival and proapoptotic functions may rely on different intracellular localizations of P21waf1/cip1. The survival effect of P21waf1/cip1 has been reported to be linked to cytoplasmic localization (Asada et al., 1999; Coqueret, 2003), whereas the apoptotic effect of P21waf1/cip1 would occur when P21waf1/cip1 is present in the nucleus (Ritt et al., 2000; Peschiaroli et al., 2002). In accordance with these reports, we showed that NaB-induced apoptosis was associated with an increase of P21waf1/cip1, particularly in the nucleus. More interestingly, we showed that the proapototic effect of P21waf1/cip1 required the interaction of P21waf1/cip1 with PCNA, since transient transfection of p21 deficient for interaction with PCNA abrogated NaB-induced apoptosis. In addition, NaB decreased the levels of total PCNA, as well as the levels of PCNA linked to or not to DNA. However, the levels of PCNA interacting with P21waf1/cip1 were increased, as revealed by immunoprecipitation. Taken together, all these results suggested that interaction of P21waf1/cip1 with PCNA was implicated in NaB-induced apoptosis. Although the role of P21waf1/cip1 in PCNA-dependent DNA repair remains to be elucidated, it can be hypothesized that p21waf1/cip1 may promote NaB-induced apoptosis by directly regulating the DNA repair machinery. PCNA has no endogenous enzymatic activity, but it forms a sliding platform that can mediate interaction with several proteins implicated in DNA replication and reparation, as well as G1/S transition. The fact that P21waf1/cip1 and PCNA interact with a number of proteins suggests that P21waf1/cip1 and PCNA could enter into competition for a common target. For example, Gadd45, a nuclear protein implicated in DNA reparation and growth arrest, has been reported to induce apoptosis by binding and activating MEKK4 (Harkin et al., 1999; Mita et al.,


27

2002). Interestingly, like p21waf1/cip1, expression of Gadd45 may be activated in a p53-dependent or independent manner (Zhan et al., 1998; Takahashi et al., 2001), and is increased by HDAC inhibitors such as TSA and NaB (Chen et al., 2002). Gadd45 is found to interact with PCNA through a region distinct from that recognized by P21waf1/cip1 (Hall et al., 1995). However, Gadd45 and P21waf1/cip1 compete for PCNA binding (Chen et al., 1995). The sequestration of PCNA by P21waf1/cip1, as well as the decreased of PCNA levels, may contribute to reduce the interaction between PCNA and Gadd45, leading to an increase of free Gadd45 and induction of apoptosis. Another similar example is p33ING1b. This protein exerts diverse biological effects through targeting and regulating local acetylation and/ or deacetylation activities, as well as binding to DNA repair complexes via PCNA (Feng et al., 2002). p33ING1b induces expression of genes such as p21waf1/cip1 and bax. The interaction of PCNA-p33ING1b can be inhibited by P21waf1/cip1. These data suggest that the interaction of PCNA-P21waf1/cip1 may increase free p33ING1b, leading to the expression of proapoptotic genes. Nevertheless, it has been reported that interaction of p33ING1b with PCNA may trigger apoptosis after UV-inducing damage. All these data demonstrate that PCNA may act as a proapoptotic factor via its interaction with other growth-regulatory proteins such as P21waf1/cip1, Gadd45, and p33ING1b. However, the exact implication (direct or indirect) of PCNA remains to be clarified. Overall, our results demonstrated that NaB inhibited breast cancer cell growth in a P21waf1/cip1-dependent manner. P21waf1/cip1 was dispensable for cell cycle arrest, but indispensable for apoptosis induction. Moreover, NaB-induced apoptosis required the interaction of P21waf1/cip1 with PCNA. On the other hand, it has been reported that 20–60% of breast tumours contain p53 mutations, but they rarely contain p21waf1/cip1 mutations. Previously, we showed that NaB inhibits the growth of breast cancer cells independently of P53 (Chopin et al., 2002). Therefore, utilization of NaB may open interesting prospects in the fight against human breast cancer.

Materials and methods Cell culture Breast cancer cell lines MCF-7, T47-D, and BT-20 were obtained from the American Type Cell Culture Collection. MCF-7ras cell line, established by transfecting v-Ha-Ras cDNA in MCF-7 cells, was a gift from Professor M Cre´pin (Laboratoire de Recherche Oncologie Cellulaire et Molećulaire Humaine, Universite´ de Paris Nord, France). Cells were grown in EMEM medium supplemented with 5% fetal calf serum (FCS), 5 UI/ml insulin, 100 UI/ml streptomycin, 100 mg/ml penicillin, and 45 mg/ml gentamicin. Reagent and antibodies All cell culture reagents were obtained from BioWhittaker (France), except insulin which was obtained from Organon (France). NaB, Hoechst 33258, trypan blue, and electrophoresis reagents were from Sigma.

Mouse monoclonal anti-P21waf1/cip1 (EA10), anti-PCNA (PC10), antiactin, and rabbit polyclonal anti-CDK2 (PC44) antibodies were from Oncogene Research Products. Goat polyclonal anti-P27 (C-19) and rabbit polyclonal anti-P16 (N20) were from Santa Cruz Biotechnology. Secondary antibodies were from Chemicon International Inc. Anti-murin CD80 FITC, anti-hamster IgG FITC antibodies for flow cytometry or anti-murin CD80 for cell selection were purchased by BD Pharmigen. FITC-conjugated secondary anti-mouse IgG antibody for immunofluorescence microscopy was from Santa Cruz Biotechnology. ECL reagents were obtained from Amersham Life Science. Protein A/G PLUSagarose were from Santa Cruz Biotechnology. CELLection Pan Mouse IgG Dynabeads and Dynal Magnetic particle concentrator were purchased from Dynal. The plasmids containing different p21 constructs were generous gifts from Professor Bernard Ducommun (Universite´ Paul Sabatier, Toulouse, France) and Professor David Smith (Mayo Fundation, USA). The plasmid containing murin CD80 sequence was a generous gift from Dr Rodolphe Vereecque (IRCL, Lille, France). Cell cycle analysis MCF-7 cells were treated with 2.5 mm butyrate for 48 h. Cells (1 106) were trypsinized and fixed in ethanol (70%, 30 min, 201C). Fixed cells were washed with PBS and incubated with propidium iodide (20 mg/ml, 45 min, 201C). The cell suspension was then filtered and analysed using a Coulter Epics XL/XLMCl cytometer. To investigate the role of P21waf1/cip1 in NaB-induced cell cycle arrest, MCF-7 cells were transiently cotransfected with a pcDNA plasmid containing murin CD80 cDNA, and a pcDNA plasmid containing p21 wild type (p21WT), p21 antisense (p21AS) or p21 deficient for the interaction with PCNA (p21PCNA ), or with CDK (p21CDK ) sequences. Briefly, 1 106 cells were seeded into 100 mm dishes. The following day, cells were transfected with 6.2 mg CD80plasmid, 2.1 mg plasmid of interest, and 409 ml transfection reagent exgen (Euromedex) in 7.3 ml Opti-MEM (Life Technologies Inc.). After transfection, cells were cultured further for 18 h in FCS-containing medium before treatment, and then treated with 2.5 mm NaB for 24 h. Cells (1 106) were dissociated with 1 mm EDTA in PBS, washed twice in the same buffer, and then incubated with 0.5% FCS in PBS (15 min, 41C). Cells were incubated with anti-murin CD80 FITC antibody, or with anti-hamster IgG FITC antibody as isotypic control (1 mg of antibody in PBS containing 0.5% FCS, 30 min, 41C in the dark). Cells were washed twice with 0.5% FCS in PBS, and fixed with cold ethanol (30 min, 41C). The fixed cells were washed with PBS and incubated with propidium iodide (20 mg/ml, 45 min, 201C). Transfected cells were selected using anti-murin CD80 FITC and the cell cycle of transfected cells was analysed with a Coulter Epics XL/XL-MCl cytometer. Determination of apoptotic cells Apoptosis was determined by morphological analysis, after fixation with PAF 4% (30 min, 41C) and staining with 1 mg/ml Hoechst 33258 (10 min, room temperature, in the dark). A minimum of 500–1000 cells were examined for each case under fluorescent microscope, and the results represented the number of apoptotic cells over the total number of cells counted. To investigate the role of P21waf1/cip1 in NaB-induced apoptosis, MCF-7 cells were transiently cotransfected with a pEGFP-C1 plasmid containing green fluorescent protein (GFP) cDNA (Clontech), and a pcDNA plasmid containing Oncogene


28 p21 WT or p21 AS, p21 PCNA , and p21 CDK sequences. Briefly, 2 105 cells were seeded into 35 mm dishes The next day, cells were transfected with 0.5 mg of pEGFP-C1 plasmid, 0.25 g plasmid of interest, and 7 ml transfection reagent exgen (Euromedex) in 1 ml Opti-MEM (Life Technologies Inc.). After transfection, cells were cultured for another 18 h in FCScontaining medium before treatment. They were then fixed in 4% paraformaldehyde (20 min, 41C) and stained with Hoechst for apoptosis detection. A random count of 500 GFP-positive cells was performed for each assay.

incubated with anti-murin CD80 antibody on a rotating device (1 mg in PBS containing 0.1% BSA/1 106 cells, 30 min, 41C). They were washed twice with 0.1% BSA in PBS, and mixed with CELLection Pan Mouse IgG Dynabeads on a rotating device (30 min, 41C). Finally, they were placed on a Dynal Magnetic particle concentrator, and were washed three times with PBS containing 0.1% BSA. The selected transfected cells were lysed, and immunoprecipitation was performed as described above. Detection of PCNA in triton-soluble and -insoluble fractions

Western blot Preconfluent cell cultures were washed with PBS and lysed for 30 min at 41C in a lysis buffer (50 mm Tris pH 7.5, 150 mm NaCl, 1% NP40, 0.1% SDS, 1 mm PMSF, 1 mm orthovanadate, 1 mm Na4P2O7, 10 mg/ml aprotinin, and 10 mg/ml leupeptin). The lysate was sonicated, boiled, and clarified by centrifugation (13 000 g, 5 min at 41C). The protein concentration of each sample was determined using a Biorad protein assay kit. Each sample was then loaded onto 5% stacking/12% running SDS–polyacrylamide gel. After electrotransfer to nitrocellulose membrane (Hybond-C extra, Amersham), membranes were blocked in 3% bovine serum albumin (BSA) in TBST (Tris 20 mm, pH 7.6; NaCl 150 mm; Tween 0.1%), for 1.5 h at room temperature. Immunoblots were incubated with a primary antibody (1 : 200 dilution, overnight at 41C). Detection was performed using donkey anti-mouse, or donkey anti-goat, and donkey anti-rabbit secondary antibodies (1 : 10 000 dilution, 1.5 h at room temperature), and an ECL detection system (Amersham). Immunoprecipitation Cells were washed with PBS, and lysed for 30 min at 41C in a lysis buffer (50 mm Tris pH 7.5, 150 mm NaCl, 1% NP40, 1 mm PMSF, 1 mm orthovanadate, 1 mm Na4P2O7, 10 mg/ml aprotinin, and 10 mg/ml leupeptin). Lysates were clarified by centrifugation (12 000 g, 10 min). Cellular extracts (500 mg) were precleared with 5 ml protein A/G PLUS-Agarose (2 h, 41C) and incubated with primary antibody on a rotating device (2 mg, overnight, 41C). Then, the immune complex was collected, following incubation with 25 ml of protein A/G PLUS-agarose on a rotating device (2 h, 41C), and centrifugation (12 000 g, 3 min). Immunoprecipitates were washed three times with ice-cold lysis buffer, and then analysed by Western blot, as previously described. To verify the functional properties of plasmids containing mutant p21 deficient for the interaction with PCNA (p21 PCNA ) or with CDK (p21 CDK ), MCF-7 cells were seeded in 100 mm dishes, and transiently cotransfected with a pcDNA plasmid containing murine CD80 cDNA and p21 PCNA or p21 CDK plasmids. After transfection, cells were treated with 2.5 mm NaB for 24 h. Cells were dissociated with 1 mm EDTA in PBS, washed twice in the same buffer, and then incubated with 0.1% BSA in PBS (15 min, 41C). Cells were

Preconfluent cell cultures were washed with PBS and lysed for 30 min at 41C in 2 ml of lysis buffer (Tris-HCl 50 mm, pH 7.4; NaCl 250 mm; 0.1% Triton X-100; 1 mm EDTA; 50 mm NaF; 1 mm DTT; 0.1 mm Na3VO4; 0.1 mm PMSF; 1 mg/ml leupeptin, 10 mg/ml trypsin inhibitor; 1 mg/ml aprotinin). The lysate was centrifugated (12 000 g, 41C, 5 min), and the supernatant was recovered as a triton-soluble fraction. The pellet was incubated with 2 ml of lysis buffer containing 1% SDS (41C, 30 min), and followed by centrifugation (12 000 g, 41C, 5 min). Then, the supernatant was recovered as a triton-insoluble fraction. Immunoblotting of both triton-soluble and -insoluble fractions with the anti-PCNA antibody was performed as described above. Immunofluoresence microscopy Cells were cultured on slides and fixed in 4% paraformaldehyde (20 min, 41C), followed by incubation in 50 mm NH4Cl (30 min, room temperature). They were incubated in a permeabilization buffer (0.1% triton X-100; 0.1% sodium citrate) for 15 min at room temperature. Nonspecific binding was blocked by incubation with 1% BSA in PBS (30 min, room temperature). P21waf1/cip1 was detected by incubation with mouse monoclonal anti-P21waf1/cip1 antibody (1/20, 12 h, 41C), followed by incubation with FITC-conjugated secondary antimouse IgG monoclonal antibody (1/100, 4 h, 41C). Cells were then incubated with 1 mg/ml Hoechst 33258 (10 min, room temperature, in the dark). Coverslips were mounted with a fluorescent mounting medium (DAKO) before examination under a fluorescent microscope (Zeiss).

Abbreviations CDK, cyclin-dependent kinase; FCS, fetal calf serum; HDAC, histone deacetylase; NaB, sodium butyrate; PCNA, proliferating cell nuclear antigen. Acknowledgements This work was supported by grants from the ‘Ligue Nationale contre le Cancer, Comite´ du Nord’. We acknowledge the excellent technical assistance of Isabelle LEFEVBRE and Johann ANTOL, and the critical reading of this manuscript by Benoni Boilly and Margaret Howe.

References Asada M, Yamada T, Ichijo H, Delia D, Miyazono K, Fukumuro K and Mizutani S. (1999). EMBO J., 18, 1223–1234. Bravo R and Macdonald-Bravo H. (1987). J. Cell Biol., 105, 1549–1554. Brugarolas J, Moberg K, Boyd SD, Taya Y, Jacks T and Lees JA. (1999). Proc. Natl. Acad. Sci. USA, 96, 1002–1007. Oncogene

Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP, Sedivy JM, Kinzler KW and Vogelstein B. (1998). Science, 282, 1497–1501. Chen IT, Smith ML, O’Connor PM and Fornace Jr AJ. (1995). Oncogene, 11, 1931–1937. Chen Z, Clark S, Birkeland M, Sung CM, Lago A, Liu R, Kirkpatrick R, Johanson K, Winkler JD and Hu E. (2002). Cancer Lett., 188, 127–140.


29 Chopin V, Toillon RA, Jouy N and Le Bourhis X. (2002). Br. J. Pharmacol., 135, 79–86. Cooper MP, Balajee AS and Bohr VA. (1999). Mol. Biol. Cell, 10, 2119–2129. Coqueret O. (2003). Trends Cell. Biol., 13, 65–70. Delavaine L and La Thangue NB. (1999). Oncogene, 18, 5381–5392. Della Ragione F, Criniti V, Della Pietra V, Borriello A, Oliva A, Indaco S, Yamamoto T and Zappia V. (2001). FEBS Lett., 499, 199–204. Demary K, Wong L and Spanjaard RA. (2001). Cancer Lett., 163, 103–107. Derjuga A, Richard C, Crosato M, Wright PS, Chalifour L, Valdez J, Barraso A, Crissman HA, Nishioka W, Bradbury EM and Th’ng JP. (2001). J. Biol. Chem., 276, 37815–37820. Detjen KM, Murphy D, Welzel M, Farwig K, Wiedenmann B and Rosewicz S. (2003). Exp. Cell. Res., 282, 78–89. Feng X, Hara Y and Riabowol K. (2002). Trends Cell. Biol., 12, 532–538. Galfi P, Neogrady Z and Csordas A. (2002). Cancer Lett., 188, 141–152. Gartel AL and Shchors K. (2003). Exp. Cell. Res., 283, 17–21. Gilbert J, Baker SD, Bowling MK, Grochow L, Figg WD, Zabelina Y, Donehower RC and Carducci MA. (2001). Clin. Cancer Res., 7, 2292–2300. Gregory DJ, Garcia-Wilson E, Poole JC, Snowden AW, Roninson IB and Perkins ND. (2002). Cell. Cycle, 1, 343–350. Hall PA, Kearsey JM, Coates PJ, Norman DG, Warbrick E and Cox LS. (1995). Oncogene, 10, 2427–2433. Han Z, Wei W, Dunaway S, Darnowski JW, Calabresi P, Sedivy J, Hendrickson EA, Balan KV, Pantazis P and Wyche JH. (2002). J. Biol. Chem., 277, 17154–17160. Harkin DP, Bean JM, Miklos D, Song YH, Truong VB, Englert C, Christians FC, Ellisen LW, Maheswaran S, Oliner JD and Haber DA. (1999). Cell, 97, 575–586. Ito N, Sawa H, Nagane M, Noguchi A, Hara M and Saito I. (2001). Neurosurgery, 49, 430–436 discussion 436-7. Lallemand F, Courilleau D, Sabbah M, Redeuilh G and Mester J. (1996). Biochem. Biophys. Res. Commun., 229, 163–169. Mahyar-Roemer M and Roemer K. (2001). Oncogene, 20, 3387–3398. Mandal M, Wu X and Kumar R. (1997). Carcinogenesis, 18, 229–232. Mattock H, Lane DP and Warbrick E. (2001). Exp. Cell. Res., 265, 234–241. Mita H, Tsutsui J, Takekawa M, Witten EA and Saito H. (2002). Mol. Cell. Biol., 22, 4544–4555. Morris GF and Mathews MB. (1989). J. Biol. Chem., 264, 13856–13864. Nakano K, Mizuno T, Sowa Y, Orita T, Yoshino T, Okuyama Y, Fujita T, Ohtani-Fujita N, Matsukawa Y, Tokino T, Yamagishi H, Oka T, Nomura H and Sakai T. (1997). J. Biol. Chem., 272, 22199–22206. Pagano M, Theodoras AM, Tam SW and Draetta GF. (1994). Genes Dev., 8, 1627–1639. Patnaik A, Rowinsky EK, Villalona MA, Hammond LA, Britten CD, Siu LL, Goetz A, Felton SA, Burton S, Valone FH and Eckhardt SG. (2002). Clin. Cancer Res., 8, 2142–2148. Pellizzaro C, Coradini D, Daniotti A, Abolafio G and Daidone MG. (2001). Int. J. Cancer, 91, 654–657. Perkins ND, Felzien LK, Betts JC, Leung K, Beach DH and Nabel GJ. (1997). Science, 275, 523–527.

Peschiaroli A, Figliola R, Coltella L, Strom A, Valentini A, D’Agnano I and Maione R. (2002). Oncogene, 21, 8114–8127. Qiao L, McKinstry R, Gupta S, Gilfor D, Windle JJ, Hylemon PB, Grant S, Fisher PB and Dent P. (2002). Hepatology, 36, 39–48. Qin LF and Ng IO. (2001). Cancer Lett., 172, 7–15. Richon VM, Sandhoff TW, Rifkind RA and Marks PA. (2000). Proc. Natl. Acad. Sci. USA, 97, 10014–10019. Ritt MG, Mayor J, Wojcieszyn J, Smith R, Barton CL and Modiano JF. (2000). Cancer Lett., 158, 73–84. Rousseau D, Cannella D, Boulaire J, Fitzgerald P, Fotedar A and Fotedar R. (1999). Oncogene, 18, 4313–4325. Sambucetti LC, Fischer DD, Zabludoff S, Kwon PO, Chamberlin H, Trogani N, Xu H and Cohen D. (1999). J. Biol. Chem., 274, 34940–34947. Sawa H, Murakami H, Ohshima Y, Sugino T, Nakajyo T, Kisanuki T, Tamura Y, Satone A, Ide W, Hashimoto I and Kamada H. (2001). Brain Tumor Pathol., 18, 109–114. Schwartz B, Avivi-Green C and Polak-Charcon S. (1998). Mol. Cell. Biochem., 188, 21–30. Scott M, Bonnefin P, Vieyra D, Boisvert FM, Young D, Bazett-Jones DP and Riabowol K. (2001). J. Cell. Sci., 114, 3455–3462. Shibata MA, Yoshidome K, Shibata E, Jorcyk CL and Green JE. (2001). Cancer Gene Ther., 8, 23–35. Shivji MK, Grey SJ, Strausfeld UP, Wood RD and Blow JJ. (1994). Curr. Biol., 4, 1062–1068. Snowden AW, Anderson LA, Webster GA and Perkins ND. (2000). Mol. Cell. Biol., 20, 2676–2686. Sowa Y, Orita T, Minamikawa-Hiranabe S, Mizuno T, Nomura H and Sakai T. (1999). Cancer Res., 59, 4266–4270. Sowa Y and Sakai T. (2000). Biofactors, 12, 283–287. Suzuki A, Kawano H, Hayashida M, Hayasaki Y, Tsutomi Y and Akahane K. (2000). Cell Death Differ., 7, 721–728. Takahashi S, Saito S, Ohtani N and Sakai T. (2001). Cancer Res., 61, 1187–1195. Terao Y, Nishida J, Horiuchi S, Rong F, Ueoka Y, Matsuda T, Kato H, Furugen Y, Yoshida K, Kato K and Wake N. (2001). Int. J. Cancer, 94, 257–267. Tian H, Wittmack EK and Jorgensen TJ. (2000). Cancer Res., 60, 679–684. Toschi L and Bravo R. (1988). J. Cell. Biol., 107, 1623–1628. Tsao YP, Huang SJ, Chang JL, Hsieh JT, Pong RC and Chen SL. (1999). J. Virol., 73, 4983–4990. Tsubaki J, Choi WK, Ingermann AR, Twigg SM, Kim HS, Rosenfeld RG and Oh Y. (2001). J. Endocrinol., 169, 97–110. Tsubaki J, Hwa V, Twigg SM and Rosenfeld RG. (2002). Endocrinology, 143, 1778–1788. Vaziri C, Stice L and Faller DV. (1998). Cell Growth Differ., 9, 465–474. Velazquez OC, Zhou D, Seto RW, Jabbar A, Choi J, Lederer HM and Rombeau JL. (1996). J. Parenter. Enteral. Nutr., 20, 243–250. Walker GE, Wilson EM, Powell D and Oh Y. (2001). Endocrinology, 142, 3817–3827. Wang XM, Li J and Evers BM. (1999). Anticancer Res., 19, 2901–2906. Warbrick E. (2000). Bioessays, 22, 997–1006. Xiang H, Fox JA, Totpal K, Aikawa M, Dupree K, Sinicropi D, Lowe J and Escandon E. (2002). Oncogene, 21, 3611–3619. Zhan Q, Chen IT, Antinore MJ and Fornace Jr AJ. (1998). Mol. Cell. Biol., 18, 2768–2778. Zhang Y, Fujita N and Tsuruo T. (1999). Oncogene, 18, 1131–1138.

Oncogene