Enhanced sensitivity to irinotecan by Cdk1 inhibition in the ... - Nature

Oncogene (2004) 23, 1737–1744

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

Enhanced sensitivity to irinotecan by Cdk1 inhibition in the p53-deficient HT29 human colon cancer cell line Miguel Abal1, Rui Bras-Goncalves2, Jean-Gabriel Judde2, Hafida Fsihi3, Patricia de Cremoux1, Daniel Louvard3, Henri Magdelenat1, Sylvie Robine3,5 and Marie-France Poupon*,2,5 1 Transfer Laboratory, Institut Curie-CNRS, Paris 75248, France; 2Metabolic alterations and cancer therapy, UMR 147, Institut Curie-CNRS, rue d’Ulm, Paris Cedex 75248, France; 3Morphogenesis and intracellular signalling, UMR 144, Institut Curie-CNRS, Paris 75248, France; 4Tumor Biology Department, Medical Section, Institute Curie-CNRS, Paris 75248, France

Mutations in the tumor-suppressor gene p53 have been associated with advanced colorectal cancer (CRC). Irinotecan (CPT-11), a DNA topoisomerase 1 inhibitor, has been recently incorporated to the adjuvant therapy. Since the DNA-damage checkpoint depends on p53 activation, the status of p53 might critically influence the response to CPT-11. We analysed the sensitivity to CPT-11 in the human colon cancer cell line HT29 (mut p53) and its wild-type (wt)-p53 stably transfected subclone HT29-A4. Cell-cycle analysis in synchronised cells demonstrated the activation of transfected wt-p53 and a p21WAF1/CIP1-dependent cell-cycle blockage in the S phase. Activated wt-p53 increased apoptosis and enhanced sensitivity to CPT-11. In p53-deficient cells, cDNAmacroarray analysis and western blotting showed an accumulation of the cyclin-dependent kinase (cdk)1/cyclin B complex. Subsequent p53-independent activation of the cdk-inhibitor (cdk-I) p21WAF1/CIP1 prevented cell-cycle progression. Cdk1 induction was exploited in vivo to improve the sensitivity to CPT-11 by additional treatment with the cdk-I CYC-202. We demonstrate a gain of sensitivity to CPT-11 in a p53-mutated colon cancer model either by restoring wild-type p53 function or by sequential treatment with cdk-Is. Considering that mutations in p53 are among the most common genetic alterations in CRC, a therapeutic approach specifically targeting p53-deficient tumors could greatly improve the treatment outcomes. Oncogene (2004) 23, 1737–1744. doi:10.1038/sj.onc.1207299 Keywords: colorectal cancer; irinotecan; p53-dependent sensitivity; cdk-inh1; CYC-202; Roscovitine

Introduction Colorectal cancer (CRC) is the second leading cause of cancer death in Western countries, and notwithstanding *Correspondence: M-F Poupon; E-mail: [email protected] 5 Principal investigators Received 18 June 2003; revised 18 September 2003; accepted 22 October 2003

the efforts made to improve chemotherapy, response rates have not been associated with a significant survival benefit. For many years, standard therapy for advanced CRC has been based on the thymidylate-synthase inhibitor 5-fluorouracil (5-FU). Recently, new compounds with different mechanisms of action have demonstrated increased response rates (Fishman and Wadler, 2001). CPT-11, as a single agent, showed tumor response in patients with 5-FU-resistant CRC (Sobrero et al., 2000). CPT-11 has been approved, in combination with 5-FU and the modulator leucovorin, as first-line chemotherapy for patients with metastatic CRC (Rougier and Mitry, 2001; Vanhoefer et al., 2001). CPT-11 is a semisynthetic derivative of camptothecin, converted in vivo into its active form SN-38, with cytotoxic effects exerted through its binding to and inhibition of the DNA-associated nuclear enzyme topoisomerase 1 (top1), thus stabilising top1 DNA cleavable ternary complexes (Tanizawa et al., 1994). This impedes the DNA-religation reaction and results in DNA doublestrand breaks, eventually leading to apoptosis (Kjeldsen et al., 1992). Sensitivity to CPT-11 might depend on top 1 activity, tumor-associated deficiencies in DNA repair and cellcycle regulation, and on the inability of cancer cells to repress apoptosis. In this context, the influence of the p53 status to the response of tumor cells to CPT-11 remains controversial. First, p53 would contribute by protecting cells against CPT-11-induced damage, as shown by the correlation of CPT-11 with long-term arrest in the p53 þ / þ HCT116 colorectal carcinoma cell line, and with apoptosis in the p53/ knocked-out derived HCT116 cell line (Magrini et al., 2002). In addition, increased cytotoxicity was observed in MCF-7 breast carcinoma and HCT116 cells upon p53 inactivation (Gupta et al., 1997). Second, p53 would sensitise cells to CPT-11, as described in a variety of human cancer cell lines and normal human fibroblasts (Blagosklonny and El-Deiry, 1998). In vivo studies with xenografted human CRCs have shown that mutated p53 status correlated with a poor response to CPT-11 (BrasGoncalves et al., 2000), and with significantly lower levels of DNA-top1 complexes trapped by camptothecin (Lansiaux et al., 2001). Finally, the combination of irradiation and SN-38 treatment showed supra-additive

p53-related sensitivity to irinotecan in HT29 cells M Abal et al

1738

effects on fibroblasts, independent of the p53 status (Xie et al., 2000). To better understand the involvement of p53, we compared, both in vitro and in vivo, the sensitivity to CPT-11 in the mutant (mut)-p53 HT29 colon cancer cell line and the wild-type (wt)-p53 subclone HT29-A4. We observed that the presence of a functional p53 resulted in an improved response to CPT-11. We used DNAmicroarray technology to characterise in vitro the molecular mechanisms defining the cellular response to CPT-11 both in the presence and in the absence of a functional p53. The accumulation of the G2/M cyclindependent kinase (cdk)1/cyclin B complex in the p53deficient cells led us to design a chemotherapy combination to increase their sensitivity to CPT-11 both in vitro and in vivo, by additional administration of the cdkinhibitor (cdk-I) CYC-202.

Results Activation of stably transfected wt-p53 blocks cell-cycle progression in CPT-11-treated HT29 cells To investigate whether the status of p53 modulates the effect of CPT-11 on cell-cycle progression, we specifically focused on the S phase, where this top1 inhibitor exerts its action by activating the DNA-damage checkpoint. For this purpose, cells were synchronised in G0/G1 by serum starvation for 48 h (see Materials and methods), CPT-11 was added to the medium 6 h after the release from the block, when cells started to cycle, and the progression through the cell cycle was analysed by 5-bromo-20 -deoxyuridine (BrdU) incorporation in wt-p53- and mut-p53-treated cells. Both untreated HT29 and HT29-A4 cells behaved similarly: FACS analysis showed a peak of 70% of cells in G0/G1 after 48 h of serum starvation; cells progressed into the S phase around 16 h after release, through G2/M at 24 h, and completed a cell cycle approximately 36 h after release from the block (Figure 1a). In contrast, both wtp53 and mut-p53 cells were blocked in the S phase when incubated up to 36 h in the presence of CPT-11 at 1 mM, a concentration that effectively blocks synchronised HT29 cells (not shown) (Figure 1b). Interestingly, wtp53 cells were blocked in the early S phase, in contrast to mut-p53 cells, arrested in the late S-phase (Figure 1b). This suggested that in wt-p53 HT29-A4, cell-cycle arrest resulted from the immediate stabilization and activation of wt-p53 in response to CPT-11. Western blot analysis of p53 protein levels further confirmed the induction of the transfected wt-p53 in CPT-11-treated versus untreated HT29-A4 cells. Two bands were detected with the a-p53 antibody in untreated cells, CPT-11 treatment resulting in a significant induction of the upper band that might correspond to the stabilized and activated form of the protein (Figure 1c). It should be noted that p53 protein levels were also slightly increased after CPT-11 treatment in mut-p53 cells, probably reflecting the stabilization of the mutated inactive form of p53 (Figure 1c). We analysed the functionality of the Oncogene

Figure 1 p53 activation and cell-cycle blockage in response to CPT-11. Mut-p53 HT29 and wt-p53 HT29-A4 cells were synchronised in G0/G1 by serum starvation for 48 h. (a) Histogram showing the cell-cycle progression from the G1 to the S phase and through G2/M transition after release from the block, in untreated control HT29 cells (HT29-A4 behaved similarly). The histogram shows the percentage of cells in each phase of the cycle at 8, 16, 24 and 36 h after the addition of serum, obtained by BrdU incorporation. (b) BrdU incorporation and cell-cycle distribution of synchronised wt-p53 HT29-A4- and mut-p53 HT29-treated cells. CPT-11 (1 mM) was added to the medium 6 h after release from G0/ G1 and cells were incubated up to 36 h. Note that in panel a, untreated cells have already completed one cycle and have progressed into the next G1 at 36 h after release. (c) Western blot analysis of synchronised untreated and CPT-11-treated cells, demonstrating the p53 stabilisation and activation in wt-p53 HT29-A4-treated cells, in comparison to mut-p53 HT29-treated cells. Synchronised cells were incubated for 24 h with and without 1 mM CPT-11. Mdm-2 expression was used as a marker of p53 activation (see text); actin is shown as loading control

activated wt-p53 by testing the induction of the downstream p53 gene mdm2. The Mdm2 protein binds to p53 and acts as its major cellular antagonist, ubiquitinating p53 and addressing it to degradation by the proteasome. Moreover, the mdm2 gene is a direct target for positive transcriptional activation by p53, thus defining the basal p53/mdm2 autoregulatory loop that results in the continuous repression of p53 activity and its maintenance in a biologically inert state (Oren et al., 2002). We observed that mdm2 was specifically induced in CPT-11treated wt-p53 cells (Figure 1c). From these results we can conclude that in HT29 cells, the stably transfected wt-p53 was activated in response to CPT-11, promoting a p53-dependent cell-cycle arrest. In vitro enhanced sensitivity to CPT-11 of wt-p53 versus mut-p53 HT29 cells We analysed whether the dominant expression of wt-p53 influenced the sensitivity of this colon cancer cell line to CPT-11. As shown by cell proliferation assays, incuba-


1739 Table 1 Antitumor effects of low-dose (10 mg/kg/day) or high-dose (40 mg/kg/day) CPT-11 on wt-p53 HT29-A4 xenografts Treatment Mean (mg/kg/ TGTa (s.d.) day) (days) 0 10 40

8 (2) 18* (7) 30* (7)

Growth delayb (days) —

TGI (%)c No. (days) mice per group —

10 22

60 86

7 10 8

No. PRd

No. curese

— 8 7

— — 1

a Tumor growth time (TGT): time in days necessary to reach a fivefold increase of individual tumor volume from the size at the start of treatment (60–250 mm3). bGrowth delay: calculated as the difference between treated/control TGT. cTumor growth inhibition (TGI): calculated as the ratio between treated and control growth inhibition, at the date of the ethical sacrifice of the first control mouse bearing a tumor with a olume of 2000 mm3. dPartial response (PR): up to 50% of individual growth inhibition. eTumor-free mice were defined as mice without any palpable tumor at the end of the experiment. The twopaired Student’s t-test was used. *Significantly different from the control group: 103oPo106.

Figure 2 Stable expression of wt-p53 sensitizes HT29 cells to CPT-11. (a) Cell proliferation assays in synchronised wt-p53 HT29A4 and mut-p53 HT29 cells incubated with 1 mM CPT-11 for 72 h. Results are shown as mean7s.d. values from three independent experiments and represent the percentage of cell survival of HT29A4 and HT29 cells in relation to their respective untreated cells incubated with fresh medium for 72 h. Bars indicate s.d.’s. (b) Western blot analysis of cells processed as in (a), demonstrating the significantly higher induction of apoptosis in wt-p53 HT29-A4treated cells, in comparison to mut-p53 HT29-treated cells. (p85)PARP- and (p17/p12)-caspase-3-cleaved products were used as markers of apoptosis (see text). (c) Antitumoral effect of CPT-11 (10 and 40 mg/kg/day) in mice bearing HT29-A4 xenografts. The mean7s.d. value of RTV in the different mice groups is plotted as a function of time

tion of wt-p53 cells with 1 mM CPT-11 for 72 h significantly decreased cell proliferation with respect to untreated cells and in comparison to mut-p53 cells (Figure 2a). This suggested that, in addition to the cellcycle arrest due to the activation of the stably transfected p53, CPT-11 treatment was associated with an increased apoptosis in wt-p53 cells. This was further confirmed by Western blot analysis of the p85 fragment of poly-ADP-ribose polymerase (PARP) that results from the caspase cleavage of the 116 kDa intact protein (Whitacre et al., 1999) (Figure 2b). Moreover, the activation of caspase-3, one of the main executers of apoptosis and responsible for PARP cleavage, by proteolytic processing into activated p17 and p12 subunits, was confirmed in wt-p53 cells treated with CPT-11, in comparison to mut-p53 cells (Figure 2b). In vivo enhanced sensitivity to CPT-11 of wt-p53 versus mut-p53 HT29 cells We compared the response to CPT-11 in wt-p53- and mut-p53-established xenografts. Both high (40 mg/kg/ day) and low (10 mg/kg/day) CPT-11 concentrations inhibited the growth of wt-p53 xenografts with a mean

tumor growth inhibition (TGI) of 60% and a mean tumor growth time (TGT) of 18 (77) and 87% and a mean TGT of 30 (77), respectively (Table 1). Moreover, CPT-11 treatment at both doses significantly delayed the tumor growth of mice bearing the wt-p53 HT29-A4 hCRC xenografts, as shown by the relative tumor volume (RTV) values plotted against time (Figure 2c). In contrast, mut-p53 xenografts were resistant to CPT-11 at a low dose, although at a high dosage some growth inhibition was observed (BrasGoncalves et al., 2000; see below). These results indicate that, at least in the mutated p53 colon cancer cell model used, the introduction of a functional p53 significantly improved the response to CPT-11. Involvement of p21WAF1/CIP1 in CPT-11-induced cell-cycle arrest To understand the mechanism underlying cell-cycle arrest in response to CPT-11, we used cDNA-macroarray technology (SuperArray) for the analysis of the expression profile of a panel of 96 genes involved in the cellcycle regulation. These arrays are focused on cdk genes as main regulators of cell cycle, and genes regulating the activities of cdk’s at multiple levels, such as cyclins, cdkIs, cdk phosphatases and cdk kinases. In addition, they include genes essential for DNA damage and mitotic spindle checkpoints, as well as genes in the Skp1-cullinFbox and anaphase promoting complex/cyclosome ubiquitin-conjugation complexes. We first observed a slight activation of p21WAF1/CIP1 in both wt-p53 and mutp53 cell lines treated with 1 mM CPT-11 (Figure 3a). Quantitative real-time PCR analysis confirmed a 5.8and sevenfold increase in p21WAF1/CIP1 expression in CPT11-treated HT29-A4 and HT29 cells, respectively. We also performed Western blot analysis to confirm the DNA-microarray indication that p21WAF1/CIP1 was involved in the response to CPT-11. We found that the faint transcriptional activation corresponded to a significantly higher induction of p21WAF1/CIP1 protein levels in the mut-p53-treated cells, compared to those of Oncogene


1740

Figure 3 p21WAF1/CIP1 and cdk1 activation in response to CPT-11 in mut-p53 HT29 cells. (a) Profiles of the expression pattern of p21WAF1/CIP1 and cdk1 in wt-p53 HT29-A4 and mut-p53 HT29 cells. At 6 h after release from the G0/G1 block, synchronised cells were incubated for a further 24 h with or without 1 mM CPT-11 and processed for RNA extraction, cDNA synthesis and hybridization on Human Cell Cycle GEArray Q series membranes (see Materials and methods). The tetra-spots shown are normalized signals of the corresponding genes expression profile, and are representative of three independent experiments; actin is shown as the control housekeeping gene. (b) Western blot analysis of p21WAF1/CIP1 and cdk1 protein expression in synchronized wt-p53 HT29-A4 and mut-p53 HT29 cells, treated with and without 1 mM CPT-11 under the same conditions as in (a). (c) Relative expression of cdk1 mRNA by quantitative real-time PCR. The results are represented as the mean7s.d. fold increase expression in three independent experiments. HPRT mRNA levels were used as internal standards, and cdk1 expression was normalized to HT29-A4-untreated conditions

tern blot analysis confirmed the increased levels of cdk1 in mut-p53 cells treated with 1 mM CPT-11 (Figure 3b). Cdk’s form complexes with their respective cyclins depending on the phase of the cell cycle, the cyclin being the regulatory unit and the cdk the catalytic partner. We thus analysed whether the increased cdk1 expression correlated with the expression of one of its corresponding cyclins: the cyclin A, known to complex cdk1 in the late S/G2 phase, or the cyclin B, involved in the G2/M transition. The kinetics of cdk1 expression in mut-p53 cells treated with CPT-11 paralleled those of cyclin B but not cyclin A (Figure 4a). This indicates that, while cells were blocked in the S phase, their cell-cycle regulators corresponded to that of the G2/M transition. Interestingly, we also observed that the increase in cdk1/ cyclin B expression occurred before that of p21WAF1/CIP1, known to block cell-cycle progression by acting as a potent cdk-inh (Figure 4a). Moreover, p21WAF1/CIP1 induction was completely abolished when CPT-11treated mut-p53 cells were further incubated with CYC-202 (R-roscovitine), a potent and selective cdk-inh (McClue et al., 2002). High cdk1 levels after incubation with 1 mM CPT-11 (Figure 4b, first lane) were inhibited by further treatment with CPT-11 plus 10 mM CYC-202 (Figure 4b, second lane). The inhibition of p21WAF1/CIP1 correlated with that of cdk1, suggesting that p21WAF1/CIP1 was induced in response to the accumulation of the cdk1/cyclin B complex, after treatment with CPT-11 in mut-p53 cells. This was confirmed by coimmunoprecipitation of cdk1 with an antibody directed against p21WAF1/CIP1, in mut-p53 cells incubated during 24 h with 1 mM CPT-11 (Figure 4c). We conclude that, in the absence of a p53-dependent response to CPT-11, the accumulation of high levels of cdk1 was counteracted by p21WAF1/CIP1 induction, an interaction that is known to promote cdk1 inhibition,

wt-p53-treated cells (Figure 3b). This suggests that the S phase cell-cycle arrest, observed when both wt-p53 and mut-p53 cell lines were incubated in the presence of CPT-11, resulted from the induction of p21WAF1/CIP1. Induction of cdk1 in CPT-11-treated p53-deficient cells In addition to the induction of p21WAF1/CIP1, DNA macroarrays showed that among the 96 genes involved in the cell-cycle regulation, the most significant and specific variation in CPT-11-treated p53-deficient cells that could underlie their response to this drug in the absence of a functional p53, corresponded to an increased cdc2/cdk1 expression, the kinase responsible for cell-cycle progression through the S phase to G2/ mitosis (Figure 3a). To confirm these results, cdk1 transcription was assessed by quantitative real-time PCR. A significant maximal induction of 2.6-fold increase in cdk1 expression was observed in HT29 CPT-11-treated cells relative to HT29-A4-untreated cells. Consistent with DNA-macroarray analysis, cdk1 levels did not significantly vary in CPT-11-treated HT29-A4 cells and -untreated HT29 cells relative to HT29-A4-untreated cells (Figure 3c). Moreover, WesOncogene

Figure 4 p21WAF1/CIP1 induction in response to cdk1/cyclin B accumulation in CPT-11-treated mut-p53 HT29 cells. (a) Time course of cyclin A, cyclin B, p21WAF1/CIP1 and cdk1 protein expression in mut-p53 HT29 cells. At 6 h after release from the G0/G1 block, synchronised cells were further incubated with 1 mM CPT-11 for 18, 24 and 30 h. Note that the expression of the cdk1/ cyclin B complex correlated with p21WAF1/CIP1 activation. (b) p21WAF1/CIP1 induction is abolished after incubation with the cdk-I CYC-202. At 6 h after release from G0/G1, synchronised mut-p53 HT29 cells were incubated in the presence of 1 mM CPT-11 for 24 h (first lane), and further incubated with CPT-11 plus 10 mM CYC202 for 24 h (second lane). Note that p21WAF1/CIP1 induction, correlating with that of cdk1 after CPT-11 treatment, was abolished when cdk1 was specifically inhibited. (c) Specific coimmunoprecipitation of p21WAF1/CIP1 and cdk1 in mut-p53 HT29 cells treated as in (a) for 24 h, and processed as indicated (Materials and methods)


1741

thus preventing cells from progression into G2/M (Taylor and Stark, 2001). Enhanced sensitivity of p53-deficient HT29 cells to CPT-11 after additional treatment with CYC-202 The characterisation of the molecular mechanism underlying the response to CPT-11 in p53-mut HT29 cells prompted us to improve their sensitivity to CPT-11 by exploiting the accumulation of cdk1 after 24 h of treatment, and designing an additional incubation with the cdk-inh CYC-202. Mut-p53 HT29 cells were initially pretreated with 1 mM CPT-11 for 24 h; then CYC-202 was added to the medium and cells were further incubated to a maximum of 24 h in the presence of both drugs. Untreated cells, cells treated for 48 h with CPT-11 alone and with CYC-202 alone for 24 h were used as controls. Treatment with CYC-202 alone promoted cellcycle blockage (Figure 5a). In contrast, the addition of CYC-202 to cells that have accumulated cdk1 by pretreatment with CPT-11 resulted in a significant decrease in cell proliferation when compared with CPT-11 alone (Figure 5a). Western blot analysis of caspase-3 cleavage showed an increased apoptosis after the additional treatment with CYC-202, in comparison to CYC-202 or CPT-11 alone (Figure 5b). We also analysed this hypothesis in vivo, by sequential administration of CPT-11 and CYC-202 in nude mice bearing HT29 tumor xenografts. TGT (days to reach a fourfold increase in tumor volume from the size at the start of treatment) showed an additive effect of the combination between CPT-11 and CYC-202

Figure 5 Enhanced sensitivity to CPT-11 of p53-deficient HT29 cells by additional administration of CYC-202. (a) mut-p53 HT29 cells were treated with 1 mM CPT-11 for 24 h; 10 mM CYC-202 was then added to the medium and cells were further incubated for a maximum of 24 h. Histogram showing a decrease in the percentage of mut-p53 HT29 cell survival after additional incubation (slashedshadowed bar), in comparison to a single treatment during 48 h with CPT-11 alone (shadowed bar) or CYC-202 alone (slashed bar). (b) Western blot analysis of (p17/p12)-caspase-3-cleaved products in mut-p53 HT29 cells processed as in (a), showing an increased apoptosis induction after additional treatment, in comparison with CPT-11 or CYC-202 treatment alone. *Po0.01. (c) Effect of CPT-11 and CYC-202 alone or in combination on the growth of HT29 colon tumor xenografts in nude mice. CPT-11 and CYC-202 were administered as indicated (Materials and methods), and RTV values are plotted as a function of time; P-values indicate a significant difference (Po 0.05) between CPT-11 and the CPT-11 plus CYC-202 combination

(TGT7s.d.: 3071.7), when compared to CPT-11 alone (2271.9), CYC-202 alone (1271.2) or control untreated (971) (Figure 5c). These results led us to propose a combination of CPT-11 and a cdk-inh as a reasonably designed therapy in CRC, when mutations in p53 diminish the sensitivity to the top1 inhibitor CPT-11.

Discussion CRC is considered the paradigm of the multistep progression cancer model, where genetic alterations accompany tumorigenesis (Fearon and Vogelstein, 1990), although alternative genetic pathways may contribute to the progression of the disease (Janssen et al., 2002; Smith et al., 2002). Crucial molecular events involve alteration and mutation of adenomatous polyposis coli (APC) and Kirsten-ras (K-ras) genes. In addition, mutations in the tumor-suppressor gene p53 appear to be a late phenomenon in CRC, which may allow the growing tumor with multiple genetic alterations to evade cell-cycle arrest and apoptosis (Gryfe et al., 1997; Arends, 2000). Nowadays, it is largely assumed that in addition to being directly responsible for the antitumor effect, agents damaging DNA may initiate postdamage responses by activating cell-cycle checkpoints (Johnstone et al., 2002). Accordingly, the integrity of these damage responses might also influence treatment sensitivity, and disabling apoptotic pathways activated by anticancer agents may contribute to resistance (Lowe et al., 1993). We analysed this hypothesis in terms of p53 status and the response to CPT-11, a DNA-damaging agent. We observed that CPT-11 treatment in G0/G1 synchronised HT29 cells transfected with wt-p53, resulting in a functionally active p53 leading to cell-cycle arrest in the S phase. We also observed that the presence of a functional p53 correlated in vitro with an increased apoptosis and in vivo with an improved sensitivity, in response to CPT11. These results are in agreement with those obtained in a comparative study between wt/mut-p53 hCRC xenografts, showing that mutated p53 correlated with a poor response to CPT-11 (Bras-Goncalves et al., 2000). By using cDNA-macroarray technology, the dissection of the response to CPT-11 might ideally provide with new targets to enhance the sensitivity of p53mutated CRCs to CPT-11. The analysis of the expression profile of a panel of genes involved in the cell-cycle regulation showed that the S phase cell-cycle arrest induced by CPT-11 resulted from the induction of p21WAF1/CIP1. p21WAF1/CIP1 is found in a complex involving cyclins and cdk’s and appears to be a universal inhibitor of cdk activity. The induction of p21WAF1/CIP1 can occur in response to DNA damage or mitogenic stimuli and even during differentiation, both in a p53-dependent and -independent fashion. The upregulation of p21WAF1/CIP1 has been related to both p53-dependent and -independent apoptosis in breast cancer after CPT treatment (Liu and Zhang, 1998), and the Fas pathway Oncogene


1742

and ceramide signaling have been implicated in the p53independent induction of apoptosis by camptothecintreated mut-p53 HT29 cells (Shao et al., 2001; Chauvier et al., 2002). We further demonstrated that in mut-p53 HT29 cells treated with CPT-11, p21 induction was related to a specific accumulation of the G2/M cyclindependent kinase cdk1. Unscheduled activation of cdk1 has been observed in CPT-11-treated human promyelocytic leukemia HL60 cells prior to apoptosis (Shimizu et al., 1995). All these results led us to propose two mechanisms implicated in the response to CPT-11, in wt-p53 and in mut-p53 conditions (Figure 6a). First, a functional p53 is capable of triggering a direct response to CPT-11, probably by activating the DNA-damage checkpoint in response to the DNA double-strand breaks. This would involve the activation of the downstream-upregulated p21WAF1/CIP1, eventually promoting apoptosis as a result of a sustained blockage. Second, in mutated p53 cells, CPT-11 imposes an arrest in cell-cycle progression during the S phase, probably due to the inability of cells to complete DNA synthesis successfully. Nevertheless, in the absence of a functional p53 activating the response to CPT-11, the cell-cycle machinery continues to progress and to accumulate cdk1/cyclin B complexes; p21WAF1/CIP1 is then induced in a p53-independent manner to suppress cdk1 activity and to block cells from progressing into G2/M. We are currently characterizing this p53-independent

Figure 6 (a) Schematic representation of the response to CPT-11 in wt-p53 and mut-p53 HT29 cells. CPT-11-induced DNA damage activates a p53-dependent response in HT29-A4 cells, resulting in p21WAF1/CIP1 induction leading to cell-cycle arrest, and eventually triggering apoptosis as a result of a sustained blockage. In mut-p53 HT29 cells, no p53-dependent response is activated after CPT-11induced DNA damage, and the cell-cycle regulatory machinery progresses through the S phase; p21WAF1/CIP1 is then activated to inhibit the accumulated cdk1/cyclin B complex, thus preventing cell progression into G2/mitosis. (b) Sensitivity to CPT-11 in wt-p53 and mut-p53 HT29 cells. The activation of p53 in response to CPT11 eventually leads to apoptosis as a result of a sustained cell-cycle arrest. In mut-p53 cells, the additional incubation with CYC-202 exploits the accumulation of cdk1/cyclin B complexes to improve the sensitivity to CPT-11, by inducing arrested cells into apoptosis Oncogene

p21WAF1/CIP1 induction in response to the unscheduled activation of cdk1. Recent work showed that treatment of the human colon cancer cells HCT116 with SN-38 resulted in G2 cell-cycle arrest without the induction of apoptosis. However, subsequent treatment of SN-38-treated HCT116 cells with flavopiridol induced apoptosis (Motwani et al., 2001). In a similar approach, pharmacological abrogation of camptothecin-induced S phase checkpoint by the selective protein kinase C inhibitor staurosporine (UCN-01) has been proposed to enhance the therapeutic activity of CPT, particularly in p53defective tumors (Shao et al., 1997). In this work, having characterized the mechanisms that influence the response to CPT-11 based on the status of p53, we translated this transcriptional analysis into a rational therapeutic approach. We exploited cdk1 accumulation, and by additional treatment with the cdk-inh CYC-202, we specifically targeted the CPT-11 increased levels of cdk1 and significantly improved in vitro the sensitivity of mutated p53 cells to CPT-11 (Figure 6b). This therapeutic design was also tested in vivo, demonstrating an enhanced sensitivity of mutated p53 xenografts when CYC-202 was added to CPT-11. CYC-202 appears to have the desirable property of driving HT29 cells into apoptosis rather than blocking cell-cycle progression, as might be expected from a cdk-inh, although it should be noted that we have not attempted to ascertain the mechanism of cell death. In addition, the sequential treatment we tested, exploiting the induction of cdk1 by pretreatment with CPT-11 before adding the cdk-I, introduces a completely different scenario for the mechanism of action of CYC-202. Nevertheless, our results are in concordance with a study of the cell-cycle effects of CYC-202 in Lovo colorectal carcinoma cells, showing that the major effect was not the predicted arrest in one part of the cycle, but rather an induction of cell death from all compartments of the cell cycle (McClue et al., 2002). Additional effects of these new therapeutic compounds other than the cytostatic effect by preventing cdk activation might open interesting possibilities in combinatorial chemotherapy. On the other hand, CYC-202 showed the highest potency and selectivity for cdk2 (McClue et al., 2002). Nevertheless, we observed no modification of this kinase using DNAmicroarray analysis under the different conditions tested. We thus conclude that the increased sensitivity of p53-deficient HT29 cells to CPT-11 when CYC-202 was sequentially administered in vitro, corresponded mainly to the effect of this cdk-inh on cdk1. The development of more potent and specific cdk-1 inhibitors should improve the combination therapy designed in this paper. Considering that, (1) in advanced CRC the 5-year survival drops dramatically as the disease progresses, (2) that mortality is mainly associated with a metastatic process after tumor resection and (3) that mutations in the p53 gene strongly interfere with the control of tumor cell growth and dissemination, a therapy specifically targeting CRCs with mutated p53 should lead to improved outcomes in their treatment.


1743 Materials and methods Drugs CPT-11 (Camptos, Irinotecan) was kindly supplied by Aventis (Vitry sur Seine, France), and CYC-202 from Cyclacel (Dundee, UK). Cell lines The HT29 (mutated p53 in Ala 273 codon) cell line was derived from a sigmoı¨ d colon cancer of stage B1, and its subclone HT29-A4, transfected with a wt-p53 expression vector (Pocard et al., 1996). The transfected wt-p53 had a dominant function in the HT29-A4 cell line, as shown previously for the HT29-A3 cell line (unpublished data; de Cremoux et al., 1999). Cells were maintained in DMEM medium supplemented with 10% fetal calf serum, and under continuous selection with geneticin for the wt-p53 HT29-A4 cell line. Cell-cycle progression and cell proliferation Cells were synchronised in G0/G1 by serum starvation for 48 h. The addition of serum and fresh medium released cells from the block and promoted them to cycle. Cell-cycle progression was characterised by BrdU (Sigma) incorporation and FACS analysis as described (Pocard et al., 1996). Cell proliferation was measured using the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; Sigma) colorimetric reduction method, as indicated (Bras-Goncalves et al., 2001). All experiments were performed at least three independent times. In vivo experiments Swiss nu/nu mice, 6–15 weeks old (25–35 g/body wt), were bred in the animal facilities of the Curie Institut, Paris (France), and maintained under specified pathogen-free conditions. Their care and housing were in accordance with the institutional guidelines of the French Ethical Committee (Ministe`re de l’Agriculture et de la Peˆche, Ministe`re de la Recherche, France) and under the supervision of authorized investigators. HT29 and HT29-A4 cell lines were established as transplantable tumors by subcutaneous injection of 2 106 cells. Randomized mice carrying subcutaneous grafts of 60– 250 mm3 tumor fragments were treated with CPT-11 administered intaperitoneally (i.p.) in a 0.2 ml volume using an intermittent schedule: two doses were tested, a high total dose of 240 mg/kg (six injections of 40 mg/kg/day every 4 days) and a low total dose of 40 mg/kg (four injections of 10 mg/kg/day every 4 days). Mice in the control groups received 0.2 ml of the drug-formulating vehicle. Combination of CPT-11 with CYC202 was performed by sequential cycles of CPT-11 i.p. at 40 mg/kg every 7 days, followed by 3–4 consecutive administration of CYC-202 twice daily by oral gavage at 200 mg/kg. The tumor volumes (V ¼ A B2, where A is the width of the tumor in millimeters and B the length) were measured every 3 days, and the RTVs were calculated from the formula: RTV ¼ (Vx/V1), where Vx is the volume of the tumor on day x and V1 is the volume of the tumor at the initiation of therapy (day 1). TGI was calculated as described (Bras-Goncalves et al., 2001). The two-paired Student’s t-test was used. DNA microarrays The profile expression of panel 96 genes involved in the cellcycle regulation was performed with the Human Cell-cycle

GEArrayt Q series (SuperArray, Inc., Bethesda, MD, USA). Total RNA was isolated with an RNA purification kit (Qiagen, Valencia, CA, USA), and used as the template for 32 P-labeled cDNA probe synthesis, according to the manufacturer’s instructions. Hybridization was performed overnight at 601C on a nylon membrane printed with tetra-spots of genespecific cDNA fragments of the 96 cell-cycle regulation genes. After extensive washing with saline-sodium citrate (SSC) and sodium-dodecyl sulfate (SDS) buffers (2 SSC/1% SDS; 0.1 SSC/0.5 SDS), digital images of the membranes were obtained with a phosphorimager, and quantified with ImageQuant software (SuperArray). Quantitative real-time PCR mRNA expression of the human cdk1 gene was evaluated by RT–PCR analysis using the real-time PCR machine PE7900 (ABI Prism). Reverse transcription of 2 mg RNA was performed using Superscript II RT RNase H-Reverse Transcriptase (Invitrogen Life Technologies) and oligodT primers (pd(N)6, Roche). Negative controls included samples with no reverse transcriptase. PCR reaction (denaturation at 951C for 2 min, followed by 40 cycles at 951C for 15 s and 601C for 1 min; SYBR green Core Reagents, PE Applied Biosystems) were performed in triplicate on two independent runs and normalized to the endogenous hipoxanthine phosphoribosyltransferase (HPRT) mRNA level for each reaction. Target cDNA was quantified using the delta-delta-Ct method. PCR primers were designed to anneal to specific sequences of the human p21WAF1/CIP1 gene: 50 -AggAggCgCCATgTCAgA-30 and 50 -gCTgCCgCATgggTTCT-30 ; and of the human cdk1 gene: 50 -CACATgAggTAgTAACACTCTg-30 and 50 -ATggTgCCTATACTCCA-30 . The sequences for the HPRT internal standard were 50 -gcTTTCCTTggTCAggCAgTATAAT-30 and 50 -AAgggCATATCCTACAACAAACTTg-30 . Immunoprecipitation and Western blot analysis Synchronised mut-p53 HT29 cells incubated for 24 h in the presence of 1 mM CPT-11 were lysed in RIPA buffer (Tris 50 mM, pH 8, NaCl 150 mM, NP40 1%, DOC 0.5%, SDS 0.1%, protease inhibitors). The supernatant was preincubated with protein-A sepharose (Amersham Bioscience, Sweden) before overnight incubation at 41C with 5 mg of monoclonal p21WAF1/CIP1 antibodies (Becton Dickinson, Franklin Lakes, NJ, USA). Protein-A sepharose beads were used to isolate the target proteins coupled to the antibodies. After extensive washing in lysis buffer, the beads were resuspended and boiled in Laemmli buffer, and processed for Western blot analysis with antibodies against p21WAF1/CIP1 and cdk1 (see below). Alternatively, cells were lysed and the supernatant boiled in Laemmli buffer. Samples were resolved on 10–12% SDS– PAGE gels, transferred onto a nitrocellulose membrane (Schleicher & Schuell, France) and incubated with primary antibodies (1 : 2000 mAb a-actin, Sigma; 0.1 mg/ml mAb a-p53 (Ab-6), Oncogene Research Products, San Diego, CA, USA; 2 mg/ml mAb a-mdm-2, Oncogene Research Products, San Diego, CA, USA; 1 : 100 pAb a-PARP p85 fragment, Promega, Madison, WI, USA; 1 : 1000 pAb a-cleaved caspase-3, Cell Signaling Technology, Beverly, MA; 1 : 2500 mAb a-cdk1, BD Transduction Laboratories, San Jose, CA; 1 : 400 p Ab a-p21WAF1/CIP1, Becton Dickinson; 1 : 500 pAb a-cyclin B and 1 : 500 pAb a-cyclin A; Sigma). Proteins were revealed with alkaline phosphatase conjugate antibodies (1 : 7500, Promega) and the NBT/BCIP color development substrates (Promega). Oncogene


1744 Acknowledgements We thank Cyclacel for kindly providing us with CYC-202, D Rouillard for expertise in FACS analysis and Athos Gianella-Boraclori, Drs David Lane, Manuel Buchwald and

Klaus-Peter Janssen for critical review. This work was supported by the Institut Curie (‘Programme Incitatif et Cooperatif’ Micrometastasis and Preclinical Pharmacology R&D Fund).

References Arends JW. (2000). J. Pathol., 190, 412–416. Blagosklonny MV and El-Deiry WS. (1998). Int. J. Cancer, 75, 933–940. Bras-Goncalves RA, Pocard M, Formento JL, Poirson-Bichat F, De Pinieux G, Pandrea I, Arvelo F, Ronco G, Villa P, Coquelle A, Milano G, Lesuffleur T, Dutrillaux B and Poupon MF. (2001). Gastroenterology, 120, 874–888. Bras-Goncalves RA, Rosty C, Laurent-Puig P, Soulie P, Dutrillaux B and Poupon MF. (2000). Br. J. Cancer, 82, 913–923. Chauvier D, Morjani H and Manfait M. (2002). Int. J. Oncol., 20, 855–863. de Cremoux P, Salomon AV, Liva S, Dendale R, Bouchind’homme B, Martin E, Sastre-Garau X, Magdelenat H, Fourquet A and Soussi T. (1999). J. Natl. Cancer Inst., 91, 641–643. Fearon ER and Vogelstein B. (1990). Cell, 61, 759–767. Fishman AD and Wadler S. (2001). Clin. Colorectal Cancer, 1, 20–35. Gryfe R, Swallow C, Bapat B, Redston M, Gallinger S and Couture J. (1997). Curr. Probl. Cancer, 21, 233–300. Gupta M, Fan S, Zhan Q, Kohn KW, O’Connor PM and Pommier Y. (1997). Clin. Cancer Res., 3, 1653–1660. Janssen KP, el-Marjou F, Pinto D, Sastre X, Rouillard D, Fouquet C, Soussi T, Louvard D and Robine S. (2002). Gastroenterology, 123, 492–504. Johnstone RW, Ruefli AA and Lowe SW. (2002). Cell, 108, 153–164. Kjeldsen E, Svejstrup JQ, Gromova II, Alsner J and Westergaard O. (1992). J. Mol. Biol., 228, 1025–1030. Lansiaux A, Bras-Goncalves RA, Rosty C, Laurent-Puig P, Poupon MF and Bailly C. (2001). Anticancer Res., 21, 471–476. Liu W and Zhang R. (1998). Int. J. Oncol., 12, 793–804. Lowe SW, Ruley HE, Jacks T and Housman DE. (1993). Cell, 74, 957–967. Magrini R, Bhonde MR, Hanski ML, Notter M, Scherubl H, Boland CR, Zeitz M and Hanski C. (2002). Int. J. Cancer, 101, 23–31.

Oncogene

McClue SJ, Blake D, Clarke R, Cowan A, Cummings L, Fischer PM, MacKenzie M, Melville J, Stewart K, Wang S, Zhelev N, Zheleva D and Lane DP. (2002). Int. J. Cancer, 102, 463–468. Motwani M, Jung C, Sirotnak FM, She Y, Shah MA, Gonen M and Schwartz GK. (2001). Clin. Cancer Res., 7, 4209–4219. Oren M, Damalas A, Gottlieb T, Michael D, Taplick J, Leal JF, Maya R, Moas M, Seger R, Taya Y and Ben-Ze’ev A. (2002). Biochem. Pharmacol., 64, 865–871. Pocard M, Chevillard S, Villaudy J, Poupon MF, Dutrillaux B and Remvikos Y. (1996). Oncogene, 12, 875–882. Rougier P and Mitry E. (2001). Clin. Colorectal Cancer, 1, 87–94. Shao RG, Cao CX, Nieves-Neira W, Dimanche-Boitrel MT, Solary E and Pommier Y. (2001). Oncogene, 20, 1852–1859. Shao RG, Cao CX, Shimizu T, O’Connor PM, Kohn KW and Pommier Y. (1997). Cancer Res., 57, 4029–4035. Shimizu T, O’Connor PM, Kohn KW and Pommier Y. (1995). Cancer Res., 55, 228–231. Smith G, Carey FA, Beattie J, Wilkie MJ, Lightfoot TJ, Coxhead J, Garner RC, Steele RJ and Wolf CR. (2002). Proc. Natl. Acad. Sci. USA, 99, 9433–9438. Sobrero A, Kerr D, Glimelius B, Van Cutsem E, Milano G, Pritchard DM, Rougier P and Aapro M. (2000). Eur. J. Cancer, 36, 559–566. Tanizawa A, Fujimori A, Fujimori Y and Pommier Y. (1994). J. Natl. Cancer Inst., 86, 836–842. Taylor WR and Stark GR. (2001). Oncogene, 20, 1803–1815. Vanhoefer U, Harstrick A, Achterrath W, Cao S, Seeber S and Rustum YM. (2001). J. Clin. Oncol., 19, 1501–1518. Whitacre CM, Zborowska E, Willson JK and Berger NA. (1999). Clin. Cancer Res., 5, 665–672. Xie X, Sasai K, Shibuya K, Tachiiri S, Nihei K, Ohnishi T and Hiraoka M. (2000). Cancer Chemother. Pharmacol., 45, 362–368.