Introduction

Colorectal carcinomas respond to the standard first-line chemotherapy (5FU+folic acid+irinotecan) in about 41% of cases (Vanhoefer et al., 2001). This low response rate as well as the high cytotoxicity of irinotecan prompted the search for modulators of the therapeutic effect, which would potentially allow the exposure to lower doses of this drug. Such modulators enhance the response of other chemotherapeutic agents by targeting either regulators of cell cycle arrest, for example, flavopiridol (Motwani et al., 2001), UCN-01 (Monks et al., 2000), the PARP-1 inhibitor CEP-6800 (Miknyoczki et al., 2003) or apoptosis, for example, the inhibitors of EGFR or its tyrosine kinase (Prewett et al., 2002; Koizumi et al., 2004) or COX-2 (Ducreux et al., 2003). The knowledge of which of these processes is the main determinant of the clinically observed response is, therefore, essential for a rational enhancement strategy.

The relative contribution of cell cycle arrest and/or apoptosis to tumour growth inhibition is not clear. Chemotherapeutic drugs activate intrinsic and extrinsic apoptotic pathways (Waxman and Schwartz, 2003); many reports showed an association of apoptosis with tumour growth inhibition (Whitacre et al., 1999; Motwani et al., 2001; Koizumi et al., 2004), and inhibition of apoptosis with a refractory phenotype (Liu et al., 1999). There is, however, controversy as to what extent apoptosis determines the overall cellular sensitivity to cancer therapy (Brown and Wouters, 1999; Schmitt and Lowe, 2001; Brown and Attardi, 2005).

The growth of tumours can also be inhibited by chemotherapy-induced long-term cell cycle arrest, which is frequently associated with cellular senescence. The association of cell cycle arrest and senescence with tumour cell growth inhibition has been well documented in vitro (Chang et al., 1999). The association in vivo has been demonstrated in human breast carcinomas (te Poele et al., 2002) as well as in mouse lymphomas (Schmitt et al., 2002), where it was shown that induction of senescence may be sufficient to induce a strong response to therapy.

Both, the DNA damage-induced cell cycle arrest as well as apoptosis are related to the p53 mutation status (Bunz et al., 1998; Magrini et al., 2002; Abal et al., 2004). Following DNA damage, the p53 wild-type (p53^wt) cells have been shown to preferentially undergo a long-term cell cycle arrest, while the p53-mutated (p53^mut) cells perform apoptosis (Bunz et al., 1998; te Poele and Joel, 1999; Magrini et al., 2002). DNA damage induces the overexpression of p53 protein, which in turn transcriptionally upregulates its main target gene p21. p21 protein is a cyclin-dependent kinase inhibitor, which can inhibit the activity of CDK2 – leading to a G1 phase arrest – or of CDK1 – triggering a G2/M phase arrest. It was also shown that cells with a functional p53 may enter a lasting tetraploid G1 cell cycle arrest (Andreassen et al., 2001a, 2001b). Wild-type p53 is thus capable of activating several checkpoints to ensure cell cycle arrest after damage. By contrast, lack of a functional p53 inactivates the G1 checkpoint and the cells proceed to the G2/M phase arrest notwithstanding DNA damage. The G2/M phase arrest is, however, only transient and the cells proceed in the presence of damaged DNA to an unscheduled premature mitosis, which develops into a mitotic catastrophe, and to apoptosis (Bunz et al., 1998; Magrini et al., 2002).

The decisive role of p53 status in the cellular reaction to DNA damage has been demonstrated in cell lines originating from carcinomas (Bunz et al., 1998; te Poele and Joel, 1999; Crawford and Piwnica-Worms, 2001; Magrini et al., 2002), sarcoma (Clifford et al., 2003), gliomas (Bartussek et al., 1999; Hirose et al., 2001), leukaemia (Nabha et al., 2002) as well as in the normal mouse epithelium in vivo (Merritt et al., 1997). The decision between cell cycle arrest versus apoptosis may also depend upon the balance between the cell cycle arrest- and the apoptosis-inducing proteins, such as p21 and PUMA, which are modulated by other molecules, for example, myc (Seoane et al., 2002) and p53 (Yu et al., 2003). Moreover, the function of p53 itself can be affected by the apoptosis stimulating protein of p53 (ASPP) and its inhibitory form, the iASPP protein (Samuels-Lev et al., 2001).

The relative importance of cell cycle arrest/apoptosis for the immediate effects on the tumour volume has not been systematically evaluated; cell death is usually assumed to be the decisive determinant of the response. In the present work, we addressed this question in an animal model and compared the effects of exposure to SN-38 in vitro and to irinotecan in vivo. We show that the CPT-11/SN-38-induced DNA damage may have similar inhibitory effects on the proliferation of the p53^wt- and p53-deficient cells in vitro, and that this is reflected in a similar delay of growth in the corresponding xenografts.

Results

The p53^wt colon carcinoma cells undergo a G2 / M phase arrest after treatment, followed by a long-term tetraploid G1 arrest, while in the p53-deficient cells, the G2 / M arrest is followed by apoptosis

The treatment of LS174T (p53^wt) cells with a clinically relevant concentration (Mathijssen et al., 2001) of SN-38 led to cell cycle arrest in a tetraploid (4N) state, which was detectable in FACS (Figure 1a and b). At 2 days after the start of treatment, cyclin B1 and CDK1 (cdc2) were upregulated and cyclin D1 was downregulated (Figure 2a, b), indicating that these cells were arrested in the G2/M phase of the cell cycle. However, at 4 days after the start of treatment, these cells showed a downregulation of cyclin B1 and CDK1 and an upregulation of cyclin D1 (Figure 2a, b), suggesting that after DNA duplication the LS174T cells had not completed mitosis and cytokinesis but had entered the G1 phase with a 4N DNA content (mitotic slippage). Analogous changes were observed in HCT116 cells except for cyclin D1 at day 2. Both cell lines showed very little apoptosis (Figure 1a–d, LS174T shown) and remained in the tetraploid G1 state, as indicated by a lack of expression of cyclin B1 and a high expression of cyclin D1, detectable in Western blot for at least 8 days (data not shown).

Figure 1.

(a–h) Response of LS174T and HT-29 cell lines to SN-38 treatment (10 nM, 48 h) in vitro. (a) LS174T cells are arrested in a tetraploid state for at least 8 days and execute only negligible apoptosis detectable as cells with a hypodiploid DNA content (<2N DNA); (b) representative FACS diagrams on days 0, 2 and 6 from the start of treatment; (c, d) among treated LS174T cells, sparse micronucleated cells (arrows) and very rare cells exhibiting PARP cleavage (not shown) were detectable; (e, f) HT-29 cells show a short G2/M phase arrest, followed by massive apoptosis; (g, h) apoptosis is detected by staining of PARP fragment p85 (arrows) in HT-29 cells with condensed nuclei already 3 days after start of the treatment. Bar: 20 m. (a, e) Means.d. of three independent experiments.

Full figure and legend (201K)

Figure 2.

(a–d) Cell cycle phase-specific gene expression profile in (a) LS174T, (b) HCT116, (c) HCT116p53^-/- and (d) HT-29 cells at 0, 2 and 4 days after start of the SN-38 treatment indicates a G1 arrest in p53^wt cells and a G2 arrest in the p53-deficient cells.

Full figure and legend (85K)

In contrast, the HCT116p53^-/- and HT-29 cell lines responded by undergoing a G2/M phase arrest that lasted for 2 days (Figure 1e and f, HT-29 shown) and was followed by the onset of apoptosis, which persisted 4–10 days after the start of treatment (Figure 1e). Apoptosis could be detected by the increase in the population of cells with a DNA content <2N (Figure 1f), and staining with the anti-PARP p85 antibody (Figure 1h). In these cells, cyclin B1 and CDK1 were upregulated and cyclin D1 and p27 were downregulated (Figure 2c, d), revealing that these cells were in the G2/M phase of the cell cycle and possibly proceeded through mitosis to apoptosis (Figure 1h). SN-38 treatment thus activates in colon carcinoma cell lines two different growth inhibitory pathways: cell cycle arrest in p53^wt cells and apoptosis in p53-deficient ones. This corresponds to and extends the previously obtained data with the isogenic cell lines HCT116 and HCT116p53^-/- (Magrini et al., 2002) and data of other authors (Bunz et al., 1998; te Poele and Joel, 1999), and indicates that this difference, associated with the absence or presence of p53 function, is retained in the established nonisogenic cell lines as well.

The type of cellular response depends on the presence or absence of a functional p53 protein

In order to further underpin the role of p53 in the observed response, the expression of endogenous p53 protein was suppressed in HCT116 cells using pSUPER-p53 (Figure 3b). Cells transfected with pSUPER-EGFP (mock) showed after 3 and 4 days of SN-38 treatment the expected maintenance of the tetraploid cell cycle arrest. By contrast, the pSUPER-p53-transfected cells were unable to maintain the arrest and underwent apoptosis (Figure 3a). Apoptosis was further confirmed by analysing the extent of the fragmentation of the PARP protein that was induced in the p53-suppressed cells (Figure 3b). This behaviour is similar to that of the p53^mut cells (Figure 1f versus Figure 3a) as well as of the p53-deficient cells (Magrini et al., 2002).

Figure 3.

(a, b) Suppression of p53 protein leads to a lack of SN-38-induced tetraploid arrest maintenance and increased induction of apoptosis in HCT116 cells. (a) Suppression of p53 in pSUPER-p53-transfected cells leads to a lack of maintenance of tetraploid arrest at days 3 and 4 as compared to the pSUPER-EGFP-transfected cells. The decrease in tetraploid arrest is concomitant with an increase in the sub-G1 population as well as (b) an increase in the fragmentation of the PARP protein. Result representative of two independent experiments.

Full figure and legend (69K)

The essential role of p53 was further confirmed when the p53 protein was reintroduced into the HCT116p53^-/- cells by transduction with a p53-expressing adenovirus. After 4 days of SN-38 treatment, the expression of the p53 protein led to a 20% increase of cells in a tetraploid state and to a concomitant decrease in the sub-G1 population (Figure 4a). The decrease in apoptosis was also visible as a reduction of PARP fragmentation (Figure 4b). These results show that p53 status is a decisive determinant of the response of colon carcinoma cells to SN-38 treatment.

Figure 4.

(a, b) Restoration of p53 function in HCT116p53^-/- leads to increased tetraploid cell cycle arrest and decreased apoptosis after treatment with SN-38. (a) Adenovirus-mediated overexpression of p53 protein leads to an increase in the population of cells arrested in tetraploid state and to (b) a concomitant decrease in the sub-G1 population. (b) Decrease in apoptosis was confirmed by a reduction in the extent of PARP fragmentation. Result showing means.d. of three independent experiments.

Full figure and legend (52K)

p21 is necessary for the p53-dependent cell cycle arrest response

The contribution of p21 to the SN-38-induced cell cycle arrest was analysed in HCT116p21^-/- cells. SN-38 treatment led to the induction of p53 and increase in the expression of cyclin B1 and CDK1. Cyclin D1 expression was decreasing whereas the expression of p27 protein did not change (Figure 5a). The high cyclin B1 and CDK1 and low cyclin D1 expressions indicate that the cells were arrested in the G2/M phase of the cell cycle. These cells showed a lack of maintenance of SN-38-induced tetraploid cell cycle arrest and underwent apoptosis (Figure 5b). These results show that p21 is not required for the triggering of the DNA damage-induced G2/M arrest, but is essential for its maintenance and for the suppression of apoptosis. These data independently confirm the previous findings showing that the p53–p21 pathway preferentially leads to cell cycle arrest and the absence of p21 leads to apoptosis (Yu et al., 2003).

Figure 5.

(a, b) p21 protein is a critical mediator of the p53-dependent cell cycle arrest response. (a) Treatment of HCT116p21^-/- with SN-38 leads to upregulation of p53 protein, CDK1, cyclin B1 and downregulation of cyclin D1 proteins. (b) In the HCT116p21^-/- cells, the triggering of G2/M arrest is not affected, but the cells do not maintain the arrest and undergo apoptosis as seen from the increase in the sub-G1 population. Result representative of two independent experiments.

Full figure and legend (58K)

Tetraploid G1 cell cycle arrest is associated with senescence, whereas one apoptosis-prone cell line also undergoes necrosis

Both p53^wt HCT116 and LS174T cell lines underwent senescence as seen from the -galactosidase staining, confirming the previously published observations (Chang et al., 2002; te Poele et al., 2002) (Figure 6a–d), whereas the p53-deficient cell lines did not senesce (Figure 6f–i).

Figure 6.

(a–j) Detection of senescence and necrosis after SN-38 treatment. (a–d) HCT116 and LS174T cells undergo cellular senescence. (e) These cells do not exhibit necrosis. (f–i) HCT116p53^-/- and HT-29 cells do not exhibit senescence. (j) Necrosis is observed only in HCT116p53^-/- cells at day 4 after the start of treatment. Result showing means.d. of three independent experiments.

Full figure and legend (253K)

Necrotic cell death, measured as the efflux of lactate dehydrogenase (LDH) into the culture medium (Gunasekar et al., 2001), was negligible after SN-38 treatment in the cell cycle-arrested p53^wt cell lines as well as in the apoptotic HT-29 cell line, whereas necrosis in the HCT116p53^-/- cells was evident (Figure 6e and j). Heat shock treatment, used as a standard necrosis inducer, caused necrosis in all cell lines in a time-dependent manner.

These data indicate that in response to DNA damage, tetraploid G1 cell cycle arrest may be associated with a senescence-like phenotype, whereas the apoptotic response may in some cell lines be associated with necrosis.

The delay of cell proliferation in response to SN-38 treatment in vitro is similar in cells executing long-term arrest and in those executing apoptosis

SN-38 treatment induced in both LS174T- and HT-29 cells a delay of the exponential growth phase by 10 days, while in the isogenic cell pair HCT116 and HCT116p53^-/- the delay was 13 days and 18 days, respectively (Figure 7a and b). This indicated that cell cycle arrest as well as apoptosis are capable of inhibiting cell growth in vitro to a similar extent. We further investigated the reason for the difference in the treatment-induced net cell proliferation curves between HT-29 and HCT116p53^-/- by determining the extent of cell proliferation, necrosis and total cell death at 8 and 10 days after the start of treatment. The rate of cell proliferation, as detected by BrdU incorporation, was slower in HCT116p53^-/- than in HT-29 cells at both time points (Figure 8a). The difference in necrosis visible at day 4 (Figure 6j) became smaller at day 8 and later disappeared (Figure 8b). Most importantly, the percentage of dead cells, determined in the trypan blue assay, was 1.5- and 3.6-fold higher at 8 and 10 days, respectively, in HCT116p53^-/- cells than in HT-29 cells (Figure 8c). Thus, the stronger response of HCT116p53^-/- cells, visible as a reduction of the total cell number (Figure 7b), was due to their slower proliferation and more frequent cell death than that of the HT-29 cells.

Figure 7.

(a, b) (a) The arrest/senescence-associated in vitro delay of cell growth is 10 and 13 days (LS174T and HCT116). (b) The apoptosis-associated delay is 10 and 18 days (HT-29 and HCT116 p53^-/-). Result showing means.d. of three independent experiments.

Full figure and legend (34K)

Figure 8.

(a–d) HCT116p53^-/- cells show slower rate of cell proliferation and enhanced cell death as compared to HT-29 cells after SN-38 treatment. (a) Cell proliferation rate as measured by BrdU incorporation at days 8 and 10 is higher in HT-29 cells than in HCT116p53^-/-cells. (b) Cell death by necrosis did not differ significantly in the two cell lines (P>0.05 at both time points). (c) Total cell death, determined by trypan blue staining, was higher in HCT116p53^-/- cells than in HT-29 cells. Meanss.d. of three independent experiments.

Full figure and legend (23K)

HCT116 and LS174T cells in xenografts undergo a G1 arrest, whereas HCT116p53^{- / -} and HT-29 cells undergo apoptosis after CPT-11 treatment

In order to analyse the molecular mechanism of in vivo response, expression of cell cycle phase-specific proteins was investigated in sections of tumours originating from all four cell lines. Eight days after treatment of LS174T-derived tumours, the cells showed frequently tetraploid nuclei and overexpressed p27, while cyclin B1 was detectable in the cytosol only in sparsely distributed tumour cells (Figure 9a). HCT116 cells also exhibited suppression of cyclin B1 but no overexpression of p27 (Figure 9a). The tissue lysates of both cell lines showed in Western blot a strong upregulation of p53 and p21, a moderate upregulation of p27 and of cyclin D1 proteins and a downregulation of cyclin B1 and CDK1 (Figure 9b), all indicators that 6–8 days after the start of treatment most of the tumour cells were arrested in the G1 phase of the cell cycle.

Figure 9.

(a, b) Treatment induces in vivo tetraploid G1 arrest in p53^wt cells and apoptosis in p53^mut cells as detected by immunohistochemistry and in Western blot. (a) Tetraploid cells of the LS174T- and HCT116-derived tumours show suppression of cyclin B1. p27 protein is overexpressed in LS174T but not in HCT116 8 days after start of the CPT-11 treatment; neither morphological signs of apoptosis nor M30 staining were detectable. In the tumours derived from HCT116 p53^-/- and HT-29-cells, cytosolic cyclin B1 was overexpressed in the majority of cells after treatment while p27 expression was not increased. Apoptosis is evident from cell morphology and staining with the M30 Cyto Death antibody (arrows). Bar: 20 m. (b) Expression of cell cycle phase-specific proteins in lysates of xenografts at 0, 6, 7 and 8 days after the start of CPT-11 treatment confirmed the immunohistochemical data.

Full figure and legend (525K)

In sections of HCT116p53^-/-- and HT-29-derived tumours, the frequent cytosolic overexpression of cyclin B1 (Figure 9a) indicated that most of the cells were arrested in the G2/M phase of the cell cycle. p27 was only sparsely expressed and not changed after treatment (Figure 9a). Condensed nuclei were found at 6–8 days after the start of treatment, and were stained with the M30 Cyto Death antibody, indicating apoptosis (Figure 9a). In the tissue lysates, cyclin B1 and CDK1 were upregulated, p27 protein was downregulated and p21 and cyclin D1 proteins were either weakly expressed or not detectable (Figure 9b). Thus, the cellular reaction of the tumour to treatment in vivo corresponded to the in vitro response of the respective cell lines.

CPT-11-induced long-term cell cycle arrest or apoptosis result in a similar tumour growth delay

Mice bearing tumours from each of the four cell lines were treated with CPT-11 or with PBS and tumour growth was followed over a period of more than 30 days. After treatment, the mean delay of tumour growth was 16 and 15 days (HCT116 and LS174T, respectively) and 15 and 12 days (HCT116p53^-/- and HT-29, respectively) (Figure 10a–d). Thus, the in vivo response to CPT-11 treatment in the tumours executing apoptosis was similar to those undergoing a long-term arrest.

Figure 10.

(a–d) Analysis of response of LS174T-, HCT116-, HCT116p53^-/- and HT-29 tumours grown as xenografts. (a, b) Treatment with CPT-11 delays the growth of LS174T- and HCT116-derived tumours by 15.8 and 17 days, respectively, and of HCT116p53^-/- and in HT-29-cells-derived tumours by 16 and 12 days, respectively. The results of two experiments per cell line are shown in Table 1.

Full figure and legend (42K)

Discussion

The central point of the present work is the contribution of two different biological responses to the cytostatic drug effect in vitro and in vivo. It shows that different biological mechanisms may cause similar delay of net tumour cell proliferation and leads to the conclusion that the extent of the response rather than its type may determine the effects observed in vivo.

We analysed the mechanism of the biological response to SN-38 in established colon carcinoma cell lines with a wild-type or deficient p53. SN-38-induced DNA damage triggered a transient G2/M phase arrest in a p53-independent manner, as previously described for an isogenic pair of cell lines (Magrini et al., 2002). Subsequently, LS174T and HCT116 cells underwent mitotic slippage and were arrested in the following G1 phase in a tetraploid state, which was associated with senescence. Induction of tetraploid G1 arrest has been previously shown to occur in response to DNA damage in the HCT116 colon carcinoma cell line (Andreassen et al., 2001a). We extend this observation to LS174T cells, which indicates that the tetraploid G1 arrest, associated with senescence, may be a common reaction of p53^wt tumour cells to DNA damage, not reported so far. By contrast, HT-29 and HCT116p53^-/- cells could neither perform the G1 arrest nor maintain the G2/M arrest and eventually underwent apoptosis.

We show that the presence/absence of a functional p53 decides whether the cells would undergo cell cycle arrest or apoptosis in response to DNA damage and confirm previous reports on the requirement of p21 for the p53-dependent cell cycle arrest response (Bunz et al., 1999; Han et al., 2002; Yu et al., 2003).

There was a complete correspondence between the in vitro and in vivo results: In the cultured cells as well as in the tumour tissue, both cell cycle arrest and apoptosis had similar effects on delaying the exponential growth phase. The tetraploid G1 cell cycle arrest appears to occur after CPT-11 treatment also in vivo and to represent an effective mechanism of tumour growth delay. It could provide a selective inhibition of tumour cell proliferation since nontransformed fibroblasts do not arrest in the tetraploid G1 phase but respond to DNA damage with a transient euploid G1 phase arrest and eventually resume proliferation (Andreassen et al., 2001b; Uetake and Sluder, 2004).

In addition to the p53 status, cellular response to treatment was shown to depend on other factors, for example, the activity of topoisomerase I (Jansen et al., 1997), uptake capability of CPT-11 (Pavillard et al., 2002) or activity of NF-B (Cusack et al., 2001), which may confound the effects exerted by the p53 status or even determine the final outcome. In fact, certain epithelial cell lines do not follow the p53-dependent dichotomous reaction to DNA damage described in the present work and by other authors in colon carcinomas (Bunz et al., 1998; te Poele and Joel, 1999; Andreassen et al., 2001a; Magrini et al., 2002) and glioblastomas (Wang et al., 2004). For example, an HT-29A4 clone with a dominantly expressed p53^wt described by Abal et al. (2004) shows a sustained cell cycle arrest and apoptosis after CPT-11 treatment, and the colon carcinoma cell line KM12 (p53^mut) arrests transiently in the G2 phase and then proliferates normally (Goldwasser et al., 1996), while irradiated HeLa cells (p53-deficient) perform a lasting G2/M phase arrest, not apoptosis (Crawford and Piwnica-Worms, 2001). The observed dichotomy is not observed in the case of other therapeutics: for example, oxaliplatin induces apoptosis in HCT116 while HCT116p53^-/- cells are resistant (Boyer et al., 2004).

Our data demonstrate – for the first time – that cell cycle arrest and apoptosis can be equivalent antitumour mechanisms. We conclude that at least short-term in vivo response may be related to the extent of the biological reaction rather than to its type (apoptosis or cell cycle arrest). This is in agreement with the increasing amount of data showing that apoptosis may not be the main mechanism of response of cancer cells to treatment (Brown and Attardi, 2005). This hypothesis is consistent with the observed lack of correlation between the response/nonresponse to DNA-damaging agents and the p53 status (Jansen et al., 1997; Jacob et al., 2001; Koike et al., 2004), and may explain the controversy concerning the p53 status as a prognostic factor (Liu and Gelmann, 2002). Indeed, the longest delay of exponential tumour growth after treatment was observed in HCT116Chr3 cells (25 days, unpublished data), which react to CPT-11 with a particularly strong and long-lasting cell cycle arrest (Magrini et al., 2002).

Chemotherapeutic approaches generally focus on exploiting the apoptosis-inducing pathways (Waxman and Schwartz, 2003), based on the observation that transiently arrested tumour cells are more likely to regrow (Waldman et al., 1997). Indeed, in the long-term, the chemotherapy-induced cell arrest and senescence may be less effective and may even predispose to tumour growth (Paradis et al., 2001). Whitacre et al. (1999) proposed the apoptotic cleavage of the PARP protein as a readout for tumour chemosensitivity in vivo. The lack of apoptosis may, however, mean either nonresponse or response in the form of cell cycle arrest (Roninson, 2003). For example, in 41% of breast carcinomas, the DNA damage induced senescence (te Poele et al., 2002), suggesting an association with a terminal cell cycle arrest (Chang et al., 2002).

Recent efforts to increase tumour selectivity and to improve prognosis of the responding patients are focused on enhancement of chemotherapy-induced cell cycle arrest or apoptosis by exploiting the checkpoint defects in cancer cells (Stewart et al., 2003). Clearly, the rational enhancement of clinical response is only possible if the underlying biological mechanism of the reaction is known. The present work shows that tumour growth can, at least transiently, be inhibited by a long-term cell arrest with a similar effectiveness as by apoptosis. It appears therefore logical to group the tumours into 'apoptotic responders', 'arrest responders' and nonresponders. The knowledge of the biological mechanism (cycle arrest/apoptosis) may allow the individualized use of suitable modulatory drugs in order to potentiate either reaction and possibly enhance the in vivo response.

Materials and methods

Cell lines and culture conditions

The cell lines LS174T (p53^wt), HCT116 (p53^wt), HCT116 p21^-/- (p21 knocked out) and HCT116 p53^-/- (p53 knocked out, both kindly provided by B Vogelstein, Johns Hopkins University School of Medicine, Baltimore) were maintained in DMEM and HT-29 (p53^mut) in RPMI (both from Invitrogen, Karlsruhe, Germany), in the presence of 10% fetal calf serum (Biochrom KG, Berlin, Germany). For all experiments, cells were used in the exponential growth phase (50–70% confluency).

Treatment of cells

For the analysis of in vitro responses, the cell lines were treated with 10 nM SN-38 for 2 days. Following treatment, the cells were washed, cultured in fresh drug-free medium and harvested at different time points.

Analysis of cell cycle distribution (FACS)

Following treatment, adherent cells were trypsinized and pooled with the floating cells. 1 10⁶ cells were fixed in 1 ml ice-cold 70% ethanol for 2 h at -20°C, washed in PBS and stained in PBS containing 0.1% Triton X-100, 200 g/ml RNaseA and 20 g/ml propidium iodide. Cellular DNA content was determined on a FACScan using CellQuest (Becton Dickinson, Heidelberg, Germany). Means.d. of three independent experiments was calculated for each time point.

BrdU incorporation assay

Cells were treated with 10 nM SN-38 for 48 h, washed and incubated in drug-free medium for different times. BrdU (10 M) (Sigma-Aldrich Chemie, Taufkirchen, Germany) was added 24 h prior to harvesting. The cells were trypsinized, cytospins of 10 000 cells were made and fixed in 3.7% formaldehyde for 15 min. After washing in TBS, the cytospins were incubated in 2 M HCl for 20 min, washed and incubated in 0.1 M sodium tetra borate, pH 8.5 for 10 min. Following three TBS washes, they were incubated for 15 min in 0.5% Triton X-100 and after washing blocked in blocking solution (5% donkey serum, 0.1% Triton X-100 in TBS) for 20 min. One hour incubation with anti-BrdU antibody (BD Pharmingen, Heidelberg, Germany) was followed by incubation with a biotinylated donkey anti-mouse IgG and streptavidin-Cy3 (both from Dianova, Hamburg, Germany). Nuclei were counterstained with DAPI (1 g/ml). BrdU-positive cells were counted on an Olympus BX60F5 microscope (Optical Co GmbH, Hamburg, Germany).

Determination of cell growth curves

Nontreated and SN-38-treated cells were harvested at time points indicated in the figures and the number of viable cells was determined by trypan blue dye exclusion. Means.d. of three experiments was calculated for each time point.

Determination of senescence

The extent of SN-38-induced senescence was determined as described previously (Hirose et al., 2001). Cells were seeded in 35 mm dishes and treated with SN-38. At 7 days after the start of treatment, the medium was discarded, the cells were washed in PBS, fixed (2% formamide, 0.2% glutaraldehyde in PBS) for 5 min at room temperature, washed in PBS and incubated at 37°C in 1 mg/ml freshly prepared X-Gal (Biomol, Berlin, Germany) in citric acid buffer, pH 6.0, and were microscopically evaluated.

Determination of necrosis

SN-38-induced necrosis was measured by determining the release of LDH from damaged cells into the medium at 2 and 4 days after the start of treatment by using the Cytotox One Homogenous Membrane Integrity Assay kit (Promega, Mannheim, Germany). Briefly, 2000–5000 cells were seeded in a 96-well plate with 100 l medium and were either left untreated or were exposed to SN-38 or, as a positive control for necrosis induction, were heated at 55°C for 5, 15 and 30 min. At 2 and 4 days after the start of SN-38 treatment, 100 l Cytotox One reagent was added to each well and fold increase of free LDH in the medium in treated relative to nontreated samples was determined fluorimetrically.

Generation of tumour xenografts

5 10⁶ cells from exponentially growing cultures of each cell line were injected subcutaneously as a single cell suspension in female Balb/c nude mice. When the tumours attained a volume of 50–100 mm³, the mice were injected intraperitoneally either with CPT-11 (Aventis, Paris, France) diluted in PBS at a dose of 50 mg/kg body weight or with PBS alone for 5 consecutive days. The tumours were measured every second day and the volume was calculated as 1/2 of the product of three diameters. The relative tumour volume was calculated as the ratio of the tumour volume to the tumour volume on the first day of treatment. The effect of treatment was assessed as the mean delay of tumour growth in the exponential growth phase. Two experiments were carried out for each cell line with a total of 12–15 mice per PBS-treated and CPT-11-treated group (Table 1).

Table 1 - Delay of the exponential phase of tumor growth after CPT-11 treatment (5 50 mg/kg).

Full table

Evaluation of tumour growth delay

The mean tumour growth delay was defined as the mean distance between the regression lines approximating the exponential part of the tumour growth curves determined between the relative tumour volumes 1 and 10.

Suppression of p53 with pSUPER

HCT116 cells were transfected by electroporation at 290 V and 1050 F capacitance in a GenePulser (BioRad Laboratories, München, Germany) with 20 g pSUPER-EGFP (a kind gift from Dr M Truss, Charité Campus Virchow, Humboldt University, Berlin, Germany) or with pSUPER-p53 (Oligoengine, Seattle, USA). Cells were then seeded in 2 ml prewarmed culture medium and incubated at 37°C for 24 h. Cells were washed and cultured in medium containing SN-38 and were harvested after 3 and 4 days of continuous treatment.

Overexpression of p53 protein using adenovirus expression system

HCT116p53^-/- cells were transduced with adenovirus expressing either LacZ (AdCMVLacZ) or the p53 protein (AdCMVp53) as described (Hemmati et al., 2002). Briefly, cells were seeded and after 24 h they were incubated with the adenovirus (AdCMVp53 or AdCMVLacZ as a control) at a multiplicity of infection (MOI) of 2.5 for 90 min at 37°C in medium without serum. The cells were then supplemented with medium containing FCS at a final concentration of 10%. After 24 h, the cells were treated with SN-38 (10 nM) and harvested for cell cycle analysis or for preparation of lysates at 4 days after continuous treatment.

p53 / PI bivariate flow cytometry

HCT116p53^-/- cells were transduced with either AdCMVp53 or AdCMVLacZ, and 24 h later were treated with SN-38 (10 nM) for 4 days. They were then trypsinized, pooled with floating cells and fixed in ice-cold 70% ethanol for 2 h at -20°C. After washing with PBS, cells were permeabilized with 0.5% Triton X-100 in PBS for 10 min, washed with PBS and labelled with the mouse anti-p53 antibody (clone DO-7) (Dako, Glostrup, Denmark) at a final concentration of 5 g/ml in 0.5% BSA in PBS for 1 h. Cells were then washed with 0.5% BSA in PBS and were incubated with goat anti-mouse FITC (Dianova, Hamburg, Germany) for 30 min in the dark. After washing twice, cells were resuspended in PBS containing 5 g/ml propidiumiodide and 200 g/ml RNaseA and incubated for 30 min at room temperature. Bivariate analysis was performed in a FACScan using CellQuest software (Becton Dickinson). Among the AdCMVp53-transduced cells, 25% were p53-positive. The percentage of cells in each phase of the cell cycle in the p53-positive and p53-negative population was calculated. All cells transduced with AdCMVLacZ were p53-negative and served as a separate negative control.

Western blot analysis

At different time points after treatment, adherent cells were trypsinized, washed with PBS and the cell pellets were lysed as described recently (Bhonde et al., 2005). For tissue lysates, the tumour was ground to powder in liquid nitrogen and lysed as mentioned above. The following antibodies against human proteins were employed: mouse monoclonal anti-p53 antibody (DO-7, Dako, Germany), rabbit anti-p21 antibody (C-19; sc-397), rabbit anti-cyclin B1 (H-433; sc-752) and rabbit anti-CDK1 (H-297; sc747) (all from Santa Cruz Biotechnology, Heidelberg, Germany); mouse monoclonal anti-p27 antibody (clone G173–524; BD Pharmingen); mouse monoclonal anti--actin antibody (clone AC-15; Sigma-Aldrich Chemie); anti-cyclin D1 (Ab-3; Neomarkers, Fremont, CA, USA). Two peroxidase-conjugated second antibodies were used: goat anti-mouse (IgG+IgM) or goat anti-rabbit IgG (Dianova, Hamburg, Germany). Detection was carried out using SuperSignal West Pico chemiluminescent substrate (Pierce, Rockford, IL, USA). All Western blots were performed at least twice.

Immunohistochemistry and immunocytology

Cytospins were fixed in methanol, permeabilized with 0.5% Triton X-100 in TBS, blocked with 5% serum and stained with the rabbit anti-PARP p85 fragment antibody (Promega Corp., Mannheim, Germany), followed by a biotinylated second antibody and streptavidin-Cy3. Sections of paraffin-embedded tumour tissues were stained using the mouse anti-cytokeratin 18 fragment M30 Cyto Death antibody (Roche Molecular Biochemicals, Mannheim, Germany), mouse anti-cyclin B1 (GNS1; Santa Cruz) and mouse anti-p27 antibody (BD Transduction laboratories, Heidelberg, Germany). Antigen retrieval was performed by boiling in 10 mM citric acid, pH 6, in a pressure cooker for 3 min (for anti-cyclin B1) or 15 min in a microwave oven (for M30 antibody). For anti-p27 staining, the sections were treated with 0.6% H₂O₂ for 10 min prior to antigen retrieval by boiling for 10 min. Biotinylated second antibodies (Dianova, Hamburg, Germany) were used, followed by streptavidin-Alexa 488 (Molecular Probes, Leiden, Netherlands) for cyclin B1 or streptavidin-Cy3 (Dianova) for M30 antibody. Nuclei in sections and cytospins were stained with 0.5 g/ml DAPI (Roche Molecular Biochemicals, Mannheim, Germany) in PBS. Tissue sections stained with anti-p27 were developed with DAB (3,3'-Diaminobenzidine) as a substrate for the peroxidase-conjugated second antibody and counterstained with hemalaun (Merck, Darmstadt, Germany).

References

1. Abal M, Bras-Goncalves R, Judde JG, Fsihi H, De Cremoux P & Louvard D et al.. (2004) Oncogene 23: 1737–1744. | Article | PubMed | ISI | ChemPort |
2. Andreassen PR, Lacroix FB, Lohez OD & Margolis RL. (2001a) Cancer Res 61: 7660–7668. | PubMed | ISI | ChemPort |
3. Andreassen PR, Lohez OD, Lacroix FB & Margolis RL. (2001b) Mol Biol Cell 12: 1315–1328. | PubMed | ISI | ChemPort |
4. Bartussek C, Naumann U & Weller M. (1999) Exp Cell Res 253: 432–439. | Article | PubMed | ISI | ChemPort |
5. Bhonde MR, Hanski ML, Magrini R, Moorthy D, Muller A & Sausville EA et al.. (2005) Oncogene 24: 148–156. | Article | PubMed | ISI | ChemPort |
6. Boyer J, McLean EG, Aroori S, Wilson P, McCulla A & Carey PD et al.. (2004) Clin Cancer Res 10: 2158–2167. | PubMed | ISI | ChemPort |
7. Brown JM & Attardi LD. (2005) Nat Rev Cancer 5: 231–237. | Article | PubMed | ISI |
8. Brown JM & Wouters BG. (1999) Cancer Res 59: 1391–1399. | PubMed | ISI | ChemPort |
9. Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S & Brown JP et al.. (1998) Science 282: 1497–1501. | Article | PubMed | ISI | ChemPort |
10. Bunz F, Hwang PM, Torrance C, Waldman T, Zhang Y & Dillehay L et al.. (1999) J Clin Invest 104: 263–269. | PubMed | ISI | ChemPort |
11. Chang BD, Swift ME, Shen M, Fang J, Broude EV & Roninson IB. (2002) Proc Natl Acad Sci USA 99: 389–394. | Article | PubMed | ChemPort |
12. Chang BD, Xuan Y, Broude EV, Zhu H, Schott B & Fang J et al.. (1999) Oncogene 18: 4808–4818. | Article | PubMed | ISI | ChemPort |
13. Clifford B, Beljin M, Stark GR & Taylor WR. (2003) Cancer Res 63: 4074–4081. | PubMed | ISI | ChemPort |
14. Crawford DF & Piwnica-Worms H. (2001) J Biol Chem 276: 37166–37177. | Article | PubMed | ISI | ChemPort |
15. Cusack JC, Jr, Liu R, Houston M, Abendroth K, Elliott PJ & Adams J et al.. (2001) Cancer Res 61: 3535–3540. | PubMed | ISI | ChemPort |
16. Ducreux M, Kohne CH, Schwartz GK & Vanhoefer U. (2003) Ann Oncol 14 Suppl. 2: ii17–ii23. | Article | PubMed |
17. Goldwasser F, Shimizu T, Jackman J, Hoki Y, O'Connor PM & Kohn KW et al.. (1996) Cancer Res 56: 4430–4437. | PubMed | ISI | ChemPort |
18. Gunasekar P, Li L, Prabhakaran K, Eybl V, Borowitz JL & Isom GE. (2001) Toxicol Sci 64: 83–89. | PubMed | ISI | ChemPort |
19. Han Z, Wei W, Dunaway S, Darnowski JW, Calabresi P & Sedivy J et al.. (2002) J Biol Chem 277: 17154–17160. | Article | PubMed | ISI | ChemPort |
20. Hemmati PG, Gillissen B, von Haefen C, Wendt J, Starck L & Guner D et al.. (2002) Oncogene 21: 3149–3161. | Article | PubMed | ISI | ChemPort |
21. Hirose Y, Berger MS & Pieper RO. (2001) Cancer Res 61: 1957–1963. | PubMed | ISI | ChemPort |
22. Jacob S, Aguado M, Fallik D & Praz F. (2001) Cancer Res 61: 6555–6562. | PubMed | ISI | ChemPort |
23. Jansen WJ, Zwart B, Hulscher ST, Giaccone G, Pinedo HM & Boven E. (1997) Int J Cancer 70: 335–340. | Article | PubMed | ISI | ChemPort |
24. Koike M, Fujita F, Komori K, Katoh F, Sugimoto T & Sakamoto Y et al.. (2004) Cancer Sci 95: 541–546. | Article | PubMed | ISI | ChemPort |
25. Koizumi F, Kanzawa F, Ueda Y, Koh Y, Tsukiyama S & Taguchi F et al.. (2004) Int J Cancer 108: 464–472. | Article | PubMed | ISI | ChemPort |
26. Liu MC & Gelmann EP. (2002) Semin Oncol 29: 246–257. | Article | PubMed | ISI | ChemPort |
27. Liu R, Page C, Beidler DR, Wicha MS & Nunez G. (1999) Am J Pathol 155: 1861–1867. | PubMed | ISI | ChemPort |
28. Magrini R, Bhonde MR, Hanski ML, Notter M, Scherubl H & Boland CR et al.. (2002) Int J Cancer 101: 23–31. | Article | PubMed | ISI | ChemPort |
29. Mathijssen RH, van Alphen RJ, Verweij J, Loos WJ, Nooter K & Stoter G et al.. (2001) Clin Cancer Res 7: 2182–2194. | PubMed | ISI | ChemPort |
30. Merritt AJ, Allen TD, Potten CS & Hickman JA. (1997) Oncogene 14: 2759–2766. | Article | PubMed | ISI | ChemPort |
31. Miknyoczki SJ, Jones-Bolin S, Pritchard S, Hunter K, Zhao H & Wan W et al.. (2003) Mol Cancer Ther 2: 371–382. | PubMed | ISI | ChemPort |
32. Monks A, Harris ED, Vaigro-Wolff A, Hose CD, Connelly JW & Sausville EA. (2000) Invest New Drugs 18: 95–107. | Article | PubMed | ISI | ChemPort |
33. Motwani M, Jung C, Sirotnak FM, She Y, Shah MA & Gonen M et al.. (2001) Clin Cancer Res 7: 4209–4219. | PubMed | ISI | ChemPort |
34. Nabha SM, Mohammad RM, Dandashi MH, Coupaye-Gerard B, Aboukameel A & Pettit GR et al.. (2002) Clin Cancer Res 8: 2735–2741. | PubMed | ISI | ChemPort |
35. Paradis V, Youssef N, Dargere D, Ba N, Bonvoust F & Deschatrette J et al.. (2001) Hum Pathol 32: 327–332. | Article | PubMed | ISI | ChemPort |
36. Pavillard V, Agostini C, Richard S, Charasson V, Montaudon D & Robert J. (2002) Cancer Chemother Pharmacol 49: 329–335. | Article | PubMed | ISI | ChemPort |
37. Prewett MC, Hooper AT, Bassi R, Ellis LM, Waksal HW & Hicklin DJ. (2002) Clin Cancer Res 8: 994–1003. | PubMed | ISI | ChemPort |
38. Roninson IB. (2003) Cancer Res 63: 2705–2715. | PubMed | ISI | ChemPort |
39. Samuels-Lev Y, O'Connor DJ, Bergamaschi D, Trigiante G, Hsieh JK & Zhong S et al.. (2001) Mol Cell 8: 781–794. | Article | PubMed | ISI | ChemPort |
40. Schmitt CA, Fridman JS, Yang M, Lee S, Baranov E & Hoffman RM et al.. (2002) Cell 109: 335–346. | Article | PubMed | ISI | ChemPort |
41. Schmitt CA & Lowe SW. (2001) Drug Resist Update 4: 132–134. | Article | ISI | ChemPort |
42. Seoane J, Le HV & Massague J. (2002) Nature 419: 729–734. | Article | PubMed | ISI | ChemPort |
43. Stewart ZA, Westfall MD & Pietenpol JA. (2003) Trends Pharmacol Sci 24: 139–145. | Article | PubMed | ISI | ChemPort |
44. te Poele RH & Joel SP. (1999) Br J Cancer 81: 1285–1293. | Article | PubMed | ISI | ChemPort |
45. te Poele RH, Okorokov AL, Jardine L, Cummings J & Joel SP. (2002) Cancer Res 62: 1876–1883. | PubMed | ISI | ChemPort |
46. Uetake Y & Sluder G. (2004) J Cell Biol 165: 609–615. | Article | PubMed | ISI | ChemPort |
47. Vanhoefer U, Harstrick A, Achterrath W, Cao S, Seeber S & Rustum YM. (2001) J Clin Oncol 19: 1501–1518. | PubMed | ISI | ChemPort |
48. Waldman T, Zhang Y, Dillehay L, Yu J, Kinzler K & Vogelstein B et al.. (1997) Nat Med 3: 1034–1036. | Article | PubMed | ISI | ChemPort |
49. Wang Y, Zhu S, Cloughesy TF, Liau LM & Mischel PS. (2004) Oncogene 23: 1283–1290. | Article | PubMed | ISI | ChemPort |
50. Waxman DJ & Schwartz PS. (2003) Cancer Res 63: 8563–8572. | PubMed | ISI | ChemPort |
51. Whitacre CM, Zborowska E, Willson JK & Berger NA. (1999) Clin Cancer Res 5: 665–672. | PubMed | ISI | ChemPort |
52. Yu J, Wang Z, Kinzler KW, Vogelstein B & Zhang L. (2003) Proc Natl Acad Sci USA 100: 1931–1936. | Article | PubMed | ChemPort |

Acknowledgements

We acknowledge the perfect technical assistance provided by Britta Jebautzke and the contribution of Brita Vorwerk to Figure 6. This work was supported by the Sonnenfeld Stiftung and Monika Kutzner Stiftung. SN-38 was a kind gift from Agnes Gonthier, Aventis Pharma S.A., Vitry-sur-Seine, Cedex France.