Abrogation of the S phase DNA damage checkpoint results in S phase progression or premature mitosis depending on the concentration of 7-hydroxystaurosporine and the kinetics of Cdc25C activation.

DNA damage causes cell cycle arrest in G(1), S, or G(2) to prevent replication on damaged DNA or to prevent aberrant mitosis. The G(1) arrest requires the p53 tumor suppressor, yet the topoisomerase I inhibitor SN38 induces p53 after the G(1) checkpoint such that the cells only arrest in S or G(2). Hence, SN38 facilitates comparison of p53 wild-type and mutant cells with regard to the efficacy of drugs such as 7-hydroxystaurosporine (UCN-01) that abrogate S and G(2) arrest. UCN-01 abrogated S and G(2) arrest in the p53 mutant breast tumor cell line MDA-MB-231 but not in the p53 wild-type breast line, MCF10a. This resistance to UCN-01 in the p53 wild-type cells correlated with suppression of cyclins A and B. In the p53 mutant cells, low concentrations of UCN-01 caused S phase cells to progress to G(2) before undergoing mitosis and death, whereas high concentrations caused rapid premature mitosis and death of S phase cells. UCN-01 inhibits Chk1/2, which should activate the mitosis-inducing phosphatase Cdc25C, yet this phosphatase remained inactive during S phase progression induced by low concentrations of UCN-01, probably because Cdc25C is also inhibited by the constitutive kinase, C-TAK1. High concentrations of UCN-01 caused rapid activation of Cdc25C, which is attributed to inhibition of C-TAK1, as well as Chk1/2. Hence, UCN-01 has multiple effects depending on concentration and cell phenotype that must be considered when investigating mechanisms of checkpoint regulation.

Progression through the cell cycle is regulated carefully to avoid proliferation or mitosis when adverse conditions exist. For example, DNA damage causes cell cycle arrest in G 1 , S, or G 2 to prevent replication on damaged DNA or to prevent aberrant mitosis. The regulatory mechanisms are known as checkpoints, and their function is to inhibit progression until the cell has adequately repaired the damage. The p53 tumor suppressor protein is required for the DNA damage-induced G 1 arrest but not for arrest in S or G 2 . Hence, normal cells arrest primarily in G 1 whereas p53-defective tumor cells arrest in S or G 2 . This observation has led to interest in developing pharmacological agents that can abrogate S and G 2 arrest and that would thereby enhance toxicity selectively in the tumor. The first compound found to abrogate G 2 arrest was caffeine (1,2), but the concentrations required cannot be tolerated by a patient. An important additional observation was that any p53 wild-type cells that arrest in S or G 2 remain resistant to caffeine-mediated abrogation (3,4). This gave further support for the hypothesis that checkpoint abrogators could enhance selectively cytotoxic therapy in many tumors.
We reported recently that 7-hydroxystaurosporine (UCN-01) 1 is 100,000-fold more potent than caffeine, abrogates S and G 2 arrest, and enhances cell killing selectively in p53-defective cells (5)(6)(7). Subsequently, the mechanisms of checkpoint regulation have become better defined (8), and it appears that the target for UCN-01-mediated abrogation of G 2 arrest is Chk1 (9,10). In the presence of DNA damage, Chk1 is phosphorylated by ATM or ATR and activated. Chk1 then phosphorylates and inactivates the Cdc25C protein phosphatase thereby preventing activation of the cyclin B⅐Cdk1 mitotic kinase complex. A second checkpoint regulator, Chk2, is activated similarly by DNA damage and also phosphorylates Cdc25C. It was reported initially that UCN-01 did not inhibit Chk2 (9,10), raising the question of how Cdc25C can be activated if Chk2 remains active. However, a recent report has suggested that Chk2 is also inhibited by UCN-01 (11). There is a third kinase, C-TAK1, that phosphorylates and inhibits Cdc25C constitutively in the absence of DNA damage (12), and this is also sensitive to UCN-01 (10). Finally, Plk3 also appears to phosphorylate and inhibit Cdc25C when cells are damaged, but its sensitivity to UCN-01 is unknown (13). In contrast, caffeine has been shown to abrogate arrest through inhibition of ATM and ATR (14,15).
The inhibition of Chk1/2 and activation of Cdc25C can explain the abrogation of G 2 arrest by UCN-01, but the potential involvement of this pathway in abrogation of S phase arrest is more complex. The literature refers generally to a single S phase checkpoint following DNA damage while ignoring the fact that there are several possible consequences that may reflect different mechanisms. The S phase checkpoint has been referred to variously as the replication checkpoint, the intra-S phase checkpoint, or simply the S 3 M checkpoint to describe cells that undergo premature mitosis directly from S phase (16 -18). Yet in our previous work, we have shown that S phase-arrested cells progress to G 2 before undergoing a lethal mitosis and furthermore, that inhibition of mitosis with no-codazole can delay cell death (7,19). The difference between these studies may relate to the specific agent used to arrest cells. Our studies have focused on DNA-damaging agents. In many other cases, S phase arrest has been induced by incubation with either antimetabolites such as hydroxyurea or cytarabine (araC) or the DNA polymerase inhibitor aphidicolin. These cells are unable to progress through S when the checkpoint is inhibited and often enter directly into a premature mitosis. Many of the same proteins have been implicated in both pathways, and it is not clear what discriminates the mechanisms of these two phenotypically different forms of S phase checkpoint.
Much of our previous work involved the DNA-damaging agent cisplatin, which causes arrest at different phases of the cell cycle depending on both p53 status and drug concentration (7, 19 -21). Hence, in p53 mutant cells, the predominant arrest at low concentrations of cisplatin is at G 2 , whereas increasing concentrations arrest cells at S or even G 1 . p53 wild-type cells show an additional arrest in G 1 at all concentrations, which can confound analysis of the role of p53 in checkpoint abrogation.
In continuing work presented here, we have found that cell cycle arrest induced by the topoisomerase I inhibitor SN38 appears to be independent of p53 status. This occurs, because the primary lesion results from a topoisomerase I-mediated cleavage event in S phase, which is beyond the G 1 checkpoint regulated by p53. Hence, both p53 wild-type and mutant cells arrest in G 2 phase at low concentration of SN38 and then late, mid-, or early S phase as the concentration increases. This has provided an excellent model for analyzing the impact of checkpoint-targeted drugs on abrogation of S phase arrest. Using this model, we find that UCN-01 abrogates S and G 2 arrest in the p53 mutant breast tumor cell line MDA-MB-231 but not in the immortalized p53 wild-type breast line, MCF10a. This resistance to UCN-01 in the p53 wild-type cells is associated with suppression of cyclins A and B. Furthermore, in the p53 mutant cells, we find a significant difference in response depending upon the concentration of UCN-01. Low concentrations of UCN-01 cause S phase cells to pass into G 2 before undergoing mitosis, whereas higher concentrations cause a direct S 3 M progression. In contrast, cells arrested in early S phase with FIG. 1. Cell cycle perturbation in two human breast cell lines following incubation with SN38. The MDA-MB-231 (A) and MCF10a (B) cell lines were incubated with the 0 -30 ng/ml SN38 for 24 h (left panel). The drug was removed, and cells were incubated for an additional 24 h in media Ϯ nocodazole (middle panels) or 50 nM UCN-01 Ϯ nocodazole (noc) (right panels). Cells were harvested and analyzed for cell cycle distribution by flow cytometry. araC are unaffected by low concentrations of UCN-01 but undergo S 3 M progression at high concentrations. Evidence is presented that this dose response reflects the differential sensitivity to UCN-01 of the various protein kinases that normally inhibit Cdc25C.

EXPERIMENTAL PROCEDURES
Cell Culture-The breast cell lines used in this study were p53 wild-type, MCF10a and p53 mutant, MDA-MB-231 (American Type Culture Collection, Manassas, VA). The cells were maintained in Dulbecco's modified Eagle's medium/F12 media supplemented with 10% fetal bovine serum, penicillin (100 units/ml), streptomycin (100 g/ml), and fungizone (0.25 g/ml). In addition, the MCF10a cells were maintained in 8 g/ml insulin, 20 ng/ml epidermal growth factor, and 500 ng/ml hydrocortisone. SN38 was kindly provided by Dr. J. Patrick McGovren (Pharmacia Upjohn Inc., Kalamazoo, MI). UCN-01 was kindly provided by Dr. Edward Sausville (National Cancer Institute, Bethesda, MD). The protein kinase C inhibitors GF109203X, Go6976, and chelerythrine chloride were obtained from Biomol (Plymouth Meeting, PA). ICP-1 is a novel checkpoint inhibitor synthesized at Dartmouth and will be described in detail elsewhere. These drugs were dissolved in dimethyl sulfoxide, with the exception of chelerythrine chloride, which was dissolved in water. araC was purchased from Sigma and dissolved in water. Caffeine (Sigma) was dissolved directly in culture medium. Cells were incubated with SN38 for 24 h, after which time the drug was removed, and the cells were incubated in fresh medium with or without the addition of UCN-01 for up to an additional 24 h. Incubations with araC were for 24 h before addition of UCN-01. The other protein kinase C and checkpoint inhibitors were substituted for UCN-01 as required. Nocodazole (Sigma) was dissolved in Me 2 SO and used at a final concentration of 0.5 g/ml. Aphidicolin was dissolved in ethanol and used at a final concentration of 10 g/ml. Okadaic acid (Sigma) was dissolved in 10% dimethyl sulfoxide and used at 30 -1000 nM.
Cell Cycle Analysis-Cell cycle analysis was performed according to a procedure described previously in which attached and detached cells are harvested, fixed in ethanol, incubated with ribonuclease, and stained with propidium iodide (21). DNA content was then determined on a BD PharMingen FACScan flow cytometer. Results are expressed as histograms, because modeling programs, although good for displaying the number of cells in G 1 , S, or G 2 , do not display adequately the progression through S phase (i.e. discriminate early S, mid-S, and late S phase). Furthermore, the binding of propidium iodide to DNA is dependent on chromatin structure (22), and extensive DNA breakage as FIG. 1-continued occurs upon incubation with SN38 leads to an apparent increase in DNA content, particularly of G 2 -arrested cells, and the modeling programs do not accommodate this situation.
Expression of cyclin A, cyclin B, and p53 were assessed by incubating ethanol-fixed cells with fluorescein isothiocyanate-conjugated anti-cyclin A or B antibody (PharMingen, San Diego, CA), or fluorescein isothiocyanate-conjugated anti-p53 (Santa Cruz Biotechnology, Santa Cruz, CA) prior to addition of ribonuclease and propidium iodide.

RESULTS
SN38 Plus UCN-01-A previous study showed that UCN-01 can abrogate S phase arrest and enhance cytotoxicity induced by the topoisomerase I inhibitor camptothecin (23). We therefore investigated the cell cycle perturbation induced by another topoisomerase I inhibitor, SN38, the active metabolite of irinotecan. Both MDA-MB-231 and MCF10a cells were incubated with a range of concentrations of SN38 for 24 h. A G 2 arrest was observed at low concentrations, a late S arrest at higher concentrations, and an early S arrest at the highest concentrations ( Fig. 1, A and B, left panels). It is important to note that SN38 action is preferential to S phase cells, hence there is little G 1 arrest even in the p53 wild-type cells; rather, concentrations can be selected in which most of the cells accumulate in S phase.
SN38 was removed at 24 h, and the S phase-arrested MDA-MB-231 cells progressed slowly toward G 2 over the following 24 h in fresh media, whereas the arrested MCF10a cells showed little progression ( Fig. 1, middle panels). Evidence of cell cycle progression was confirmed by inclusion of nocodazole during this period, which arrests cells in mitosis. At all concentrations of SN38, MDA-MB-231 but not MCF10a cells reached G 2 /M during this time frame. The addition of 50 nM UCN-01 at 24 h caused abrogation of S and G 2 arrest in MDA-MB-231 cells with many cells now exhibiting sub-G 1 DNA content (Fig. 1A, right panels). This is particularly evident in cells that had been treated at or above 3 ng/ml SN38. Microscopic analysis showed that most of these cells remained attached to the culture dish at this time and also retained viability as judged by exclusion of trypan blue, yet they exhibited a very heterogeneous size with most appearing far smaller than a normal cell. After an additional 24-h incubation, most of these cell fragments had detached from the culture dish and no longer excluded trypan blue (see below). The inclusion of nocodazole between 24 and 48 h confirmed that the UCN-01-treated cells progressed to G 2 /M during this period, but nocodazole prevented much of the cell fragmentation. These observations are consistent with previous reports that passage through mitosis enhances cell death induced by cisplatin (19,21); further experiments addressing this point are presented below. In contrast, MCF10a cells were unaffected by the addition of UCN-01 (Fig. 1B, right panels). This result is consistent with previous observations that checkpoint abrogation can be prevented by wild-type p53 (6,24).
There is an additional observation in these flow cytometry profiles that requires a comment. As the concentration of SN38 increases, the peak of G 2 -arrested cells exhibits an apparent increase in DNA content (the peak moves from 400 to ϳ480). The addition of the DNA polymerase inhibitor aphidicolin did not prevent this increase demonstrating that it is not because of reduplication of DNA (see the following section, and data not shown). For analysis of DNA content, the DNA was stained with propidium iodide. The amount of dye that binds is not an absolute reflection of the amount of DNA in a cell but rather is dependent on the condensation state of DNA (22). In these cells, extensive DNA breakage caused by SN38 is thought to open up the chromatin and facilitate increased binding of the dye, and this is responsible for the apparent increase in DNA content.
A 24-h incubation with 10 ng/ml SN38 caused predominantly an S phase arrest in both cell lines and was therefore used for subsequent studies. After removal of SN38 at 24 h, a range of UCN-01 concentrations was added. Cells were then harvested and analyzed for cell cycle distribution after an additional 6 or 24 h. For MDA-MB-231 cells, concentrations of UCN-01 as low as 2 nM caused significant abrogation of S phase arrest, with most of the cells reaching G 2 within 6 h (Fig. 2). Above 7.5 nM UCN-01, the majority of cells had also abrogated G 2 arrest by 24 h and were detected as cells with sub-G 1 DNA content. Interestingly, at the highest concentration of UCN-01 (500 nM), the majority of the cells exhibit sub-G 1 DNA content within 6 h.
To further quantify the toxicity resulting from the combination of SN38 and UCN-01, MDA-MB-231 cells were harvested and scored for viability on the basis of trypan blue exclusion. Cells incubated with 15 or 500 nM UCN-01 for 24 h in the absence of SN38 showed a slight delay in cell growth but no obvious death (Fig. 2B). Incubation with 10 ng/ml SN38 alone caused stasis as a result of protracted G 2 arrest. The addition of UCN-01 to these arrested cells caused a dramatic loss in viable cell number. However, after incubation with 15 nM UCN-01 (the 48-h time point), there was an initial increase in viable cell number. As discussed above, many of these cells were small particles resulting presumably from a lethal mitosis, and after an additional 24 h, the majority of these fragmented cells lost viability. At the higher concentration of UCN-01, this rate of death was much more rapid.
The MCF10a cells were completely refractory to the effects of UCN-01 at all concentrations and remained arrested predominantly in S phase throughout the experiment. These cells remained arrested but viable for at least another 2 days (Fig. 1B and data not shown).
The Role of Cell Cycle Progression in Cell Death Induced by SN38 Plus UCN-01-We have shown previously that passage through mitosis is necessary for UCN-01-mediated enhancement of cisplatin-induced cell killing (19). Results in Fig. 1 suggest that passage through mitosis also enhances cell killing induced by SN38. We performed additional experiments here to determine whether mitosis is a prerequisite for cell death after treatment of cells with SN38. Specifically, MDA-MB-231 cells were first arrested in S phase with SN38 and then at the time of addition of UCN-01, nocodazole was also added. UCN-01 (15 nM) caused significant cell death after an additional 24 h in the absence of nocodazole, but in the presence of nocodazole, the cells accumulated in G 2 /M (Fig. 3, left panel). Hence it appears that passage through mitosis is also required for death from the combination of SN38 and low concentrations of UCN-01. The requirement for cell cycle progression was also supported by experiments in which the DNA polymerase inhibitor aphidicolin was added. In this case, aphidicolin appeared to prevent death of cells arrested in S phase, with the dead cells resulting presumably from cells that were already in G 2 . Nocodazole also prevented this cell death in the presence of aphidicolin demonstrating the importance of passage through both S phase and mitosis for cell killing by UCN-01. Similar experiments were performed in MDA-MB-231 cells incubated with SN38 and 500 nM UCN-01 (Fig. 3, right panel). This drug combination induced cell death in 6 h, and this was also prevented by the addition of nocodazole. However, it should be noted that this inhibition was only transient, and most of the cells still died by 12 h in the presence of nocodazole (data not shown). We also observed that, in the presence of nocodazole, cells incubated with 500 nM UCN-01 still progressed to G 2 . We therefore asked whether progression through S phase was required for cell death. Incubation with aphidicolin failed to prevent cell death. When cells were incubated with both aphidicolin and nocodazole, the cells remained alive but arrested in S phase. Hence it appears that cells incubated with 500 nM UCN-01 do not need to progress to G 2 prior to dying but can undergo premature mitosis directly from S phase, which leads to cell death.
Cytosine Arabinoside Plus UCN-01-It has been shown previously that ML-1 leukemia cells incubated with nucleoside analogs such as gemcitabine or araC will also die without progression through the cell cycle when UCN-01 is added (25). some of the G 2 cells through mitosis; this is supported by the ability of nocodazole to hold the cells in G 2 /M. However, 50 nM UCN-01 was unable to abrogate the S/G 1 arrest induced by 10 M araC.
In contrast to low concentrations, 500 nM UCN-01 caused a marked increase in death of cells incubated with 3 and 10 M araC. This was evident within 6 h and was also inhibited by nocodazole. It is particularly evident in this experiment that the cells incubated with 10 M araC underwent premature mitosis and died upon addition of UCN-01 without any progression through the cell cycle. Parallel experiments were performed with the MCF10a cells, but UCN-01 was ineffective at abrogating arrest or enhancing cell death (data not shown).
To further explore the rapid induction of cell death, we determined the concentration of UCN-01 that, when combined with araC, was effective at killing the cells. MDA-MB-231 cells were incubated with 10 M araC for 24 h and then a range of UCN-01 concentrations was added for an additional 6 h. UCN-01 at concentrations of Ͼ125 nM caused a significant increase in cell death (Fig. 5, left panel). These concentrations are similar to those required to enhance death rapidly following incubation with SN38. Hence it is likely that a common mechanism is involved.
UCN-01 was developed originally as a protein kinase C inhibitor (26,27). To further investigate the potential role of protein kinase C in the abrogation of arrest and cell killing, we investigated the action of other protein kinase C inhibitors. We tested Go6976 up to 500 nM, GF109203X up to 5 M, and chelerythrine chloride up to 2 M, but there was no effect on cell cycle abrogation or cell death (Fig. 5, right panel). In addition, we tested caffeine, another well known checkpoint inhibitor. Whereas 0.5 mM caffeine abrogated S phase arrest induced by SN38 (data not shown), we found that caffeine did not abrogate araC-induced S phase arrest or induce a rapid cell death up to 5 mM. We have also synthesized a novel indolocarbazole checkpoint inhibitor, ICP-1, that abrogates SN38-induced S phase arrest at 250 nM (to be described in detail elsewhere), but it also was ineffective at abrogating araC-mediated arrest or inducing cell death up to 2.5 M. Furthermore, none of these drugs induced rapid death after incubation of cells with SN38 (data not shown). Hence, we have not found any drug other than UCN-01 with the ability to induce rapidly death of S phasearrested cells.
Analysis of Cell Cycle Regulatory Proteins-The ability of UCN-01 to abrogate cell cycle arrest is thought to be mediated via its inhibition of Chk1 (9), or possibly Chk2 (11), both of which suppress the mitosis-inducing phosphatase Cdc25C. We therefore investigated the phosphorylation of Chk1 and Chk2 and the kinetics of activation of Cdc25C in cells incubated with various concentrations of UCN-01. Cdc25C exists in three forms detectable on Western blots. The fastest migrating species is unphosphorylated (form a in Fig. 6), whereas a species with slightly retarded electrophoretic mobility represents phosphorylation at the inhibitory serine 216 (form b); in several experiments this band was resolved into two forms with slightly different mobility that may represent additional phosphorylation at serine 263 (12). A third species with a greatly retarded mobility represents the hyperphosphorylated and active form of the protein observed during mitosis (form c). In asynchronously growing undamaged MDA-MB-231 cells, Chk1 and Chk2 were both unphosphorylated, and the majority of Cdc25C was in the phosphorylated inactive form, presumably because of the constitutive activity of C-TAK1 (Fig. 6).
Incubation with 10 ng/ml SN38 for 24 h resulted in phosphorylation of both Chk1 and Chk2, whereas Cdc25C remained in its phosphorylated inactive form. Upon addition of 15 nM UCN-01, Chk1 and Chk2 remained phosphorylated consistent with the ability of UCN-01 to inhibit their activity rather than their phosphorylation (Fig. 6A). As expected, UCN-01 activated Cdc25C, but it is important to note that this did not occur until 9 -12 h, which is the time that UCN-01 drives the cells into mitosis. Hence, the cells appear to abrogate the S phase arrest without activation of Cdc25C. At later time points, the cells begin to die, which probably contributes to the reduced level of Cdc25C phosphorylation.
As another assessment of Cdc25C activation, we used two novel antibodies that detect phosphorylation at the Cdk1/cyclin B sites, threonine 48 and threonine 67. Both antibodies detected clearly phosphorylated protein in the same samples that showed the hyperphosphorylated mitotic form of Cdc25C. In the case of phosphothreonine 48, the electrophoretic mobility was the same as the hyperphosphorylated form. However, phosphothreonine 67 Cdc25C exhibited an electrophoretic mo-bility that was only slightly less than the inactive form of Cdc25C. This was confirmed by reprobing the membrane with antibody against total Cdc25C (data not shown), wherein the lower faint band was found to have an electrophoretic mobility similar to form b (more obvious in Fig. 6B). These results suggest that although threonine 67 phosphorylation is involved in the activation of Cdc25C, this site is dephosphorylated or obscured physically when Cdc25C is converted to the hyperphosphorylated form. The important conclusion from this series of assays is that abrogation of S phase arrest that occurs 0 -6 h after addition of 15 nM UCN-01 is not associated with activation of Cdc25C. However, at later times when the cells undergo mitosis, Cdc25C is activated.
MDA-MB-231 cells incubated with SN38 for 24 h followed by 500 nM UCN-01 showed a rapid activation of Cdc25C within 2 h (Fig. 6B). This was observed both as retarded mobility and with the antibodies detecting phosphothreonines 48 and 67. Interestingly, prior to activation, there was conversion of the inactive phosphorylated form of Cdc25C to the unphosphorylated form. This was not seen at low concentrations of UCN-01 and suggests a different mechanism of action. One possibility is that 500 nM UCN-01 inhibits the checkpoint pathway at a step upstream of Chk1/2, possibly at ATM/ATR. However, this possibility was ruled out by the observation that both Chk1 and Chk2 remained phosphorylated upon addition of 500 nM UCN-01. Hence, it appears that high concentrations of UCN-01 function at or downstream of Chk1/2.
A clue to the mechanism of action of 500 nM UCN-01 came from analysis of cells in the absence of DNA damage. Under these conditions, UCN-01 still caused dephosphorylation of Cdc25C (Fig. 6C). In the absence of DNA damage, this phosphorylation is caused reportedly by a constitutive kinase, C-TAK1 (12). Hence, these results may be explained by the ability of UCN-01 to inhibit C-TAK1, a property that has been reported for UCN-01 in vitro (10).
The inhibition of C-TAK1 by UCN-01 cannot cause directly dephosphorylation of Cdc25C, rather a protein phosphatase must be required. To determine which phosphatase might be involved, we incubated cells concurrently with 500 nM UCN-01 and the protein phosphatase inhibitor okadaic acid. The dephosphorylation of Cdc25C caused by incubation with UCN-01 was inhibited by 0.3-1 M okadaic acid (Fig. 6D). This experiment provided the best resolution of the two inhibitory phosphorylated forms of Cdc25C. Whereas 500 nM UCN-01 caused complete dephosphorylation of the protein, treatment with 0.3 M okadaic acid resulted in partial recovery of a slightly retarded band, whereas 1 M okadaic acid resulted in complete recovery to the upper band of the doublet seen in undamaged cells. We suspect this upper band represents phosphorylation at both serine 216 and serine 263; the latter has been reported as a minor site of phosphorylation by C-TAK1 (12). These concentrations of okadaic acid inhibit protein phosphatase 2A in cells but not protein phosphatase 1 as assessed by phosphorylation of several endogenous marker proteins. 2 When MCF10a cells were incubated with SN38, Chk1 and Chk2 were phosphorylated. However, addition of 15 nM or 500 nM UCN-01 failed to activate Cdc25C (data not shown), consistent with the inability of UCN-01 to abrogate arrest in these cells. The addition of 500 nM UCN-01 caused dephosphorylation of Cdc25C in the MCF10a cells as observed in MDA-MB-231 cells showing its ability to inhibit all Cdc25C-directed kinases in both cell lines, including presumably C-TAK1 (Fig.  6C). However, this was inadequate to abrogate cell cycle arrest.
Correlation of Cyclin Expression with Checkpoint Abrogation-The question remains as to why UCN-01 neither acti-2 K. Chatfield and A. Eastman, manuscript in preparation. vates Cdc25C nor abrogates arrest in the MCF10a cells. Cdc25C is thought to be activated by an autocatalytic loop in which cyclin B/Cdk1 activates Cdc25C, which in turn activates cyclin B/Cdk1. We therefore investigated the level of cyclin B in the two cell lines. Both cell lines, when growing asynchronously, show the typical pattern of low cyclin B in G 1 , increasing levels in S, and maximum cyclin B in G 2 (Fig. 7). When the MDA-MB-231 cells were incubated with SN38, the cells arrested in S phase with a marked increase in the amount of cyclin B. In contrast, when MCF10a cells were arrested in S phase, the majority of cells exhibited little if any cyclin B. Interestingly, when UCN-01 was added to the MCF10a cells, the 17% that expressed cyclin B were able to progress to G 2 whereas the others remained arrested in S phase (data not shown). Parallel studies with cyclin A showed a similar pattern in that cyclin A was increased in arrested MDA-MB-231 cells but decreased in MCF10a cells. In the absence of cyclins A and B, these cells would not be able to progress through the cell cycle. Hence the ability to abrogate cell cycle arrest appears to depend on the levels of cyclins A and B.
Considering that MCF10a cells contain wild-type p53, we determined whether p53 was induced under the conditions of these experiments. Wild-type p53 induces Mdm2 leading to rapid turnover of p53, and therefore little is seen in cells. However, in cells with mutant p53, this turnover fails to occur, and such cells constitutively express high levels of dysfunctional protein. The expression of this mutant p53 is detected clearly in the MDA-MB-231 cells even in the absence of DNA damage (Fig. 7). In contrast, undamaged MCF10a cells express little p53. Upon incubation with SN38, the MCF10a cells arrested in S and G 2 with elevated p53. Hence the failure of UCN-01 to abrogate arrest in the MCF10a cells correlates with both elevated p53 and suppression of cyclins A and B.

DISCUSSION
Mitosis is induced by activation of the protein phosphatase Cdc25C, which in turn activates cyclin B/Cdk1. Cyclin B/Cdk1 also phosphorylates Cdc25C on at least five sites in a feedback activation loop (28); threonines 48 and 67 are two such sites that were assessed here. The hyperphosphorylated and active form of Cdc25C exhibits a markedly reduced electrophoretic mobility shift with an apparent molecular mass of 80 kDa compared with 60 kDa for the inactive form. During interphase, Cdc25C is normally phosphorylated primarily on serine 216 and inactive. In undamaged cells, this phosphorylation is believed to occur by C-TAK1 (12); presumably this kinase is either switched off or overwhelmed during the onset of scheduled mitosis. Cdc25C is also phosphorylated at serine 216 by the two DNA damage-activated checkpoint kinases, Chk1 and Chk2. These checkpoint kinases are themselves phosphorylated and activated by ATM/ATR. In the current experiments, we observed phosphorylation of both Chk1 and Chk2 upon incubation of cells with the topoisomerase inhibitor SN38.
The ability of UCN-01 to abrogate cell cycle arrest is mediated reportedly by its inhibition of Chk1 (9,10). Initially it was suggested that Chk2 was not inhibited by UCN-01, yet a recent report has suggested that recombinant forms of the protein used in such assays give different results than native protein immunoprecipitated from cells (11). Under the latter conditions, it was shown that UCN-01 is equally effective at inhibiting Chk1 and Chk2. This still leaves C-TAK1 to inhibit Cdc25C, yet this kinase has also been reported to be inhibited by UCN-01 in vitro (10).
In our current studies, we have used a model in which cells are arrested in S phase by incubation with SN38. Abrogation of S phase arrest by a low concentration of UCN-01 was not accompanied immediately by activation of Cdc25C; rather the cells progressed to G 2 before Cdc25C was activated, and the cells underwent a lethal mitosis. The activation of Cdc25C was seen as phosphorylation on both threonine 48 and threonine 67 and also by the expected reduction in electrophoretic mobility because of hyperphosphorylation. Phosphorylation at threonine 48 was detected only in the hyperphosphorylated band, but surprisingly, the phosphothreonine 67 form had an electrophoretic mobility very close to the inactive Cdc25C. The function of threonine 67 phosphorylation is currently under investigation.
During abrogation of arrest by low concentrations of UCN-01, coincubation with either nocodazole or aphidicolin inhibited cell death supporting the hypothesis that progression through both S and M phase is required for this enhanced cytotoxicity. It is probable that the constitutive activity of C-TAK1 during S phase prevents premature activation of Cdc25C. If Cdc25C had been activated, one would expect the cells to undergo premature mitosis. These results suggest that the abrogation of S phase arrest is either independent of Chk1/2, or more likely, that Chk1/2 have alternate targets in S phase such as Cdc25A or Cdc25B. Indeed both Chk1 and Chk2 have been reported to phosphorylate Cdc25A leading to its degradation (29,30), although we have not observed degradation of Cdc25A in our cell lines following incubation with SN38. 3 Chk1/2 also phosphorylate Cdc25B, but it is implicated in initiation of mitosis rather than passage through S phase (31). The current observations do not yet provide an explanation as to how UCN-01 abrogates S phase arrest.
High concentrations of UCN-01 induced a completely different response in that cells arrested in S phase with either SN38 or araC died rapidly in a manner that was inhibited (albeit transiently) by coincubation with nocodazole. This premature mitosis correlated with immediate dephosphorylation of Cdc25C at serine 216, followed rapidly by phosphorylation at threonines 48 and 67 and appearance of the hyperphosphorylated forms. Even in the absence of DNA damage, a situation in which Chk1/2 remained unphosphorylated and presumably inactive, high concentrations of UCN-01 caused dephosphorylation of Cdc25C at serine 216. As C-TAK1 is the only kinase known to phosphorylate constitutively Cdc25C on this inhibitory serine 216, it is probable that UCN-01 is inhibiting C-TAK1 at these high concentrations.
Recently, Plk3 was also reported to phosphorylate Cdc25C on serine 216, yet Plk3 kinase is also activated by DNA damage (13). Hence it is unlikely that inhibition of Plk3 can explain the dephosphorylation of Cdc25C observed in the absence of DNA damage. This dephosphorylation of Cdc25C occurred at the same UCN-01 concentrations in damaged cells consistent with inhibition of C-TAK1 in both cases. If Plk3 is also contributing to damage-induced arrest in these cells, it would have to be inhibited by low concentrations of UCN-01 otherwise its activity might prevent abrogation of G 2 arrest. The sensitivity of Plk3 to UCN-01 is currently unknown.
Using an in vitro kinase assay, the concentration of UCN-01 reported to inhibit C-TAK1 was 27 nM compared with 11 nM for Chk1 (10). However, this 3-fold difference in sensitivity in vitro is far less than the 100-fold difference in concentration required to produce the different effects in cells; S 3 G 2 occurs at 3 E. A. Kohn and A. Eastman, unpublished observation. 2 nM UCN-01 whereas S 3 M occurs around 200 nM. Perhaps the concentrations defined as inhibitory in vitro do not relate directly to those required in cells. This certainly appears to be the case for Chk2 in which the in vitro assay with recombinant protein concluded that Chk2 was resistant to UCN-01, yet when cellular Chk2 was immunoprecipitated, it was concluded that UCN-01 was equally effective at inhibiting both Chk1 and Chk2 (11).
Simply inhibiting C-TAK1, or any other Cdc25C kinase, can only prevent phosphorylation of Cdc25C; a phosphatase must be involved in the dephosphorylation. Although protein phosphatase 2A has been shown previously to dephosphorylate active Cdc25C (32), no evidence has been presented as to what dephosphorylates serine 216. Using okadaic acid at concentrations that are known to inhibit PP2A but not PP1 in cells, we prevented the dephosphorylation at serine 216 that is induced by high concentrations of UCN-01. Hence it is likely that PP2A is responsible for this dephosphorylation.
One important repercussion of the current observations is the realization that many previous authors have used excessive concentrations of UCN-01 to investigate checkpoint abrogation; most authors have used concentrations in excess of 100 nM and even as high as 1 M (33-35). At these concentrations, UCN-01 probably inhibits not only Chk1, Chk2, and C-TAK1 but also a number of other targets, not least of which is protein kinase C, the target of UCN-01 identified originally (26,27). We used other agents to show that inhibition of protein kinase C is insufficient to mimic any of the effects of UCN-01. Two other agents that abrogate S and G 2 arrest, caffeine and ICP-1, were also unable to mimic the effects of high concentrations of UCN-01. Hence, conclusions drawn from studies using high concentrations of UCN-01 must be questioned as to whether the observed effects were because of inhibitory action on a single specified target. Furthermore, it raises the question as to whether there is a significant difference between the S phase DNA damage checkpoint and the S phase stalled replication checkpoint. In the former case, abrogation of arrest can lead to S phase progression as observed here with SN38, but this is not possible if the replication fork remains blocked by direct inhibition of the polymerase by aphidicolin, by the absence of deoxyribonucleotides as in the case of hydroxyurea, or by araC as used here. However, at higher concentrations of UCN-01, cells incubated with any of these agents can undergo premature mitosis without cell cycle progression.
Our results have also shown that UCN-01 can abrogate arrest in the p53 mutant MDA-MB-231 cells but not in the p53 wild-type MCF10A cells. We have reported previously abrogation of arrest in p53 mutant T47D cells (19) and in p53 wildtype MCF7 cells only after they have been transformed with the human papilloma virus E6 gene to disrupt p53 (6). The possibility that a drug combination such as SN38 plus UCN-01 can target selectively p53-defective tumors is particularly intriguing. Cells with wild-type p53 may be protected from this combination first by arresting in G 1 after DNA damage, or alternatively if they arrest in S or G 2 , they may be resistant to UCN-01-mediated abrogation of the arrest. However, there is contradictory evidence as to whether this drug combination is active preferentially on p53-defective cells (36,37). There is also contradictory evidence as to whether the activity of this drug combination requires abrogation of cell cycle arrest (36,38). The results presented here provide a possible resolution of these conflicting reports. Our results agree with the observation that enhanced cytotoxicity is not dependent on cell cycle progression, but this appears to be associated with S 3 M progression at high concentrations of UCN-01. At lower concentrations of UCN-01, cell cycle progression is required for enhanced cell death. Our results suggest that abrogation of arrest also correlates with cyclin A/B levels. It is important to note that the current experiments were performed with two cell lines of different origin and may differ in many aspects in addition to p53 status. However, analysis of a number of other cell lines discussed above supports the role of p53 in suppressing checkpoint abrogation. Furthermore, there exists a connection between p53 status and cyclin regulation in that p53 is thought to regulate the G 2 checkpoint through cyclin B (39). Three different reporter elements in the cyclin B promoter are candidates for this p53-mediated suppression (40 -42). It has been reported that more than 80% of primary tumors exhibit dysregulation of cyclin B (43). This is far more frequent than defects in p53 and suggests that some p53 wild-type cells may have other defects downstream of p53 that may make them sensitive to the administration of drugs such as UCN-01. This raises the exciting possibility that these drug combinations may be effective in many tumors that were reported previously to be normal for p53, while remaining ineffective in the normal tissue of the patient.