Main

Mutations in p53 occur in 50% of all human tumors, and in many of the remaining tumors, p53 pathways are inactivated either due to the amplification of murine double minute 2 (MDM2), its negative regulator, or loss of expression of alternate open reading frame (ARF).1, 2, 3 Targeting p53 mutant tumors with chemotherapeutics selective for dividing cells (e.g., mitotic inhibitors, vinblastine, paclitaxel and S-phase inhibitors) has met with varying successes, partly due to normal tissue toxicities. A characteristic of cells carrying loss-of-function mutations in the p53 gene is the loss of the G1 restriction checkpoint. It is possible that temporary reversible activation of G1 checkpoint in normal cells using specific agents can protect them from the effects of some mitotic inhibitors.4, 5 Cancer cells lacking normal checkpoint regulation continue to divide in the presence of mitotic drugs, resulting in mitotic catastrophe and cell death. Recent studies demonstrated that nutlin, a small molecule antagonist of MDM2 functions, induced a p53-dependent checkpoint arrest and protected some cells from toxicity of the mitotic and S-phase inhibitors.6, 7 Nutlin-3 disrupts the MDM2–p53 interaction, resulting in the activation and accumulation of p53 in cells expressing wild-type p53.8

The crucial roles of aurora kinases in cell cycle regulation coupled with evidence of overexpression of aurora A and B kinases in a wide variety of tumors has led to the development of pharmacological inhibitors of aurora kinases.9, 10 A new generation of aurora inhibitors, including VX-680, have been evaluated preclinically and are in phase I and II trials for the treatment of solid tumors and leukemia.9 VX-680 inhibits aurora A and B kinases with low nanomolar potency11 and the cellular effects are consistent with the inhibition of both aurora A and B kinases, that is, monospindle phenotype, disruption of spindle assembly checkpoint (SAC), aberrant chromosomal segregation, endoreduplication and cytokinesis failure.12, 13 Endoreduplication, the re-replication of DNA in the absence of mitotic division, ultimately results in polyploidy and cell death. It is thought that a p53-dependent postmitotic checkpoint prevents polyploidy.13 It is however unclear whether wild-type p53 suppresses polyploidy in cells exposed to aurora kinase inhibitors and, importantly, whether normal cells are also susceptible to these drug effects.

The aims of this study were to determine whether (1) normal cells and cells expressing wild-type p53 are susceptible to VX-680-induced endoreduplication and apoptosis, and whether (2) the combination of a potent and selective antagonist of MDM2, nutlin-3 and VX-680 could protect wild-type p53-expressing cells from endoreduplication and apoptosis while killing p53 mutant cells selectively. In addition, we found that loss of p53 led to increased polyploidy and that abrogating p73 functions compromises the apoptotic response and increases polyploidy in cells exposed to VX-680. Both p53 and p73 have distinct and cooperative roles in mediating cell cycle arrest and apoptosis in response to aurora kinase inhibition. Collectively, our study has important implications for the use of aurora kinase inhibitors and MDM2 antagonists in the clinic.

Results

Aurora kinase inhibition induces extensive endoreduplication in both p53 wild-type and p53-deficient cell lines

p53 has been implicated in a postmitotic G1 checkpoint in response to various mitotic inhibitors. We examined whether this checkpoint is induced after aurora kinase inhibition by VX-680 and whether tetraploidy is suppressed by wild-type p53 functions. Using a panel of commonly used tumor-derived cell lines expressing either wild-type p53, mutant p53 or a deletion at the p53 gene locus (Table 1), we tested the correlation of p53 gene status to VX-680-induced endoreduplication. These cell lines were treated with VX-680, and harvested for cell cycle analysis at 24, 48 and 72 h. To our surprise, all tumor cell lines undergo endoreduplication. Despite the presence of wild-type p53 in A375, A549, U2OS and HCT116, endoreduplication occurs but to a variable extent, giving rise to cells with 8N, and, in some cases, 16N and 32N DNA content (Figure 1). p53 was activated in cells treated with VX-680, and p53 target genes, p21, PUMA and MDM2 were all upregulated (Supplementary Figure S1A). Activation of p53 seemed to restrict further endoreduplication through imposing cell cycle arrest, as evident from the lack of S-phase cells at 48 and 72 h after drug treatment (Figure 1). Taken together, our data suggest that p53 response is activated in response to aurora kinase inhibition but is insufficient to completely protect against endoreduplication.

Table 1 Tumor-derived and primary cell lines used in this study
Full size table
Figure 1
figure 1

Inhibition of aurora kinase induces endoreduplication and apoptosis in tumor-derived cell lines expressing wild-type p53. (a) p53-positive cells (A375, U2OS, A549 and HCT116 p53+/+) and (b) p53-compromised cells (HCT116 p53−/−, MDA-MB-486 and H1299) were treated with 200 nM VX-680 and harvested at the indicated time points. Analysis of cell cycle distribution using propidium iodide staining was performed using flow cytometry. Representative DNA profiles are shown. DNA content is indicated. No DNA profile was shown for cells that are 100% apoptotic

Full size image

VX-680 induces tetraploid G2 arrest

We did a careful time course analysis of the DNA content of A549 (p53 wild type) after VX-680 treatment and found that cells arrested briefly in G2/M at 16 h with predominantly 4N DNA content before a second round of DNA replication was initiated at 16 h (Figure 2a). After 48 h, more than 35% of cells were arrested in tetraploid G2/M (8N) (Figure 2a). Analysis of the mitotic indexes suggests that cells were arrested in G2 and not in mitosis; the mitotic index peaked at 16 h after treatment with VX-680 but subsequently decreased to 2% at 48 h (Supplementary Figure S1B). Together, the results confirmed that the 8N cells were arrested predominantly in G2 and VX-680 inhibited mitosis in part through a G2 checkpoint. To elucidate the mechanism underlying the observed G2 arrest, we examined both cyclin A2-Cdk2 and cyclin B1-cdc2, protein complexes known to have crucial roles in the regulation of the G2–M transition.14, 15, 16 A careful time course analysis of the cyclin-associated kinase activities revealed a rapid reduction of cyclins A2- and B1-associated kinase activities (Figure 2b). Cyclin-dependent kinase 2 (Cdk2) and cell division cycle 2 (cdc2) kinase activities were also markedly reduced to undetectable levels, but this was neither a result of substantial changes in the levels of cdc2/Cdk2 protein nor of inhibitory tyrosine phosphorylation.17 As the kinase activity of Cdk2 or cdc2 requires binding to the appropriate cyclins, we reasoned that the ablation of cdc2/Cdk2 kinase activities could be because of a change in the intracellular levels of cyclins A2 and B1. Indeed, a time-dependent decrease in the protein levels of cyclins A2 and B1, correlating with the kinetics of suppression of Cdk activities (Figure 2c), was observed.

Figure 2
figure 2

VX-680-induced tetraploid G2 arrest. (a) Cell cycle distribution of A549 cells incubated with 200 nM VX-680 was assessed at the specified time points, over a 48-h time course. (b) Kinetics of Cdk/cyclin kinase activities after incubation with VX-680. Cdk/cyclin kinase activities were assayed in vitro (refer to Materials and Methods). Cdk2, cyclins A and B1 were immunoprecipitated from cell lysates using specific antibodies followed by in vitro kinase assays using histone H1 as a substrate. Suc1 is used to capture Cdk1 and Cdk2 complexes by protein pull down as described in Materials and Methods. (c) Comparison of expression levels of cell cycle-related proteins (Rb, cdc2, Cdk2, cyclin A2, cyclin E1 and cyclin B1) and p53-dependent gene products (MDM2 and p21) at the indicated time points after VX-680 treatment. Downregulation of mitotic cyclins A and B and upregulation of p53-dependent genes p21 and MDM2 correlate with the activation of p53. (d) Quantification of the mRNA levels of cyclins A and B, p21 and MDM2 in A549 cells incubated with VX-680 by qRT-PCR. (e) HCT116 p53−/− and p53−/− cells were treated as described in (a). Protein lysates were analyzed by western blot for the indicated proteins

Full size image

p53 imposes a caffeine-resistant G2 checkpoint in part through a p21-dependent repression of cyclins A2 and B1 mRNAs

We noticed that both cyclins were depleted in A549 cells but remained unchanged in A549 cells transfected with p53 siRNA (Figures 2c and Supplementary Figure S2D). Similarly, depletion of cyclin A2 protein and, to a lesser extent, cyclin B1 protein were observed in VX-680-treated wild-type HCT116 but not in the derivatives HCT116p53−/− or HCT116p21−/− (Figure 2e; data not shown). Loss of cyclins A2 and B1 occurs at the transcriptional level as their mRNAs were greatly reduced in the wild-type cells but not in p53−/− or p21−/− cells (Figure 3). Given that the promoters of cyclins A2 and B1 genes contain p53 regulatory elements,18 a plausible explanation is that accumulated p53 protein directly represses the transcription of cyclins A2 and B1 genes. Alternatively, the p21 induced by p53 can bind directly to Cdk/cyclin complexes, resulting in the sequestration of E2F1 by hypophosphorylated retinoblastoma protein (pRb). The repression of cyclins A2 and B1 is specific, as cdc2 protein (a reported target of p53-mediated repression19), shows only a marginal decrease (Figure 2c). Although it is tempting to speculate that p53 mediates the repression through a p21-dependent pathway, it is likely that p21-independent activity of p53 also contributes to the observed repression, as HCT116p53−/− cells show consistently higher levels of cyclins A2 and B1 transcripts when compared with HCT116p21−/− cells at later time points of VX-680 treatment (Figure 3). Rescue of cyclins A2 and B1 expression coincided with increased polyploidy (Figures 3, 6a and Supplementary Figure S2). Conversely, siRNA-mediated downregulation of both genes, required for mitotic entry,15, 20 contributes to the observed G2 arrest and prevents further endoreduplication in A549 transfected with p53siRNA (Supplementary Data and Figure S3). Ectopic expression of p53 or p21 suppressed polyploidy in p53-deficient cells (Supplementary Data and Figure 4). Together, these data suggest p53 mediates the transcriptional repression of cyclins A2 and B1 genes, and depletion of cyclins A2 and B1 suppresses endoreduplication through a G2 checkpoint. Furthermore, caffeine, which inhibits ataxia-telangiectasia mutated (ATM) and ATR and effectively overrides the G2 checkpoint in response to DNA damage but not in response to cyclin A2 knockdown,15 did not override the G2 arrest induced by VX-680 (Supplementary Data and Figure S5).

Figure 3
figure 3

Effects of p53 or p21 deletion on the levels of p53 transcription targets in HCT116 cells treated with VX-680. mRNA levels of p53 transcriptional targets were assessed at specific time points after exposure of HCT116 cells to VX-680 over a time course of 48 h. Transcript levels of p53 transcriptional targets, p21, MDM2 and PUMA were significantly decreased in p53−/− and p21−/− cells when compared with the wild-type parental cells. It is noteworthy that cyclins A2 and B1 levels remained high in p53−/− and p21−/− cells even after cells were incubated with VX-680 for 48 h, in contrast to that observed in the wild-type cells

Full size image
Figure 4
figure 4

Ectopic expression of p53 and p21 rescues cells from endoreduplication. (a) A549 cells were transduced with recombinant lentiviruses carrying the siRNA-resistant p53 gene, p53siRes, which contains four silent mutations in siRNA binding sequence. Stable cell line expressing p53siRes (A549p53siRes) was transfected with p53-specific siRNA before incubating with VX-680 for another 72 h. DNA profiles were analyzed by FACS and protein lysates were prepared and analyzed by western blotting for the indicated proteins. A549 transduced with empty vector was used as control. (b) Overexpression of p21 in HCT116 p53−/− was achieved using recombinant adenovirus carrying CMV-promoter driven p21 (Ad-CMVp21). At 8 h after transduction, cells were incubated with VX-680 for 72 h and cell cycle profile was analyzed by FACS. Polyploidy (>8N) was represented as a percentage of total cell population. Results are representative of three independent experiments. MOI, multiplicity of infection; C, control

Full size image

Previous activation of p53-dependent cell cycle arrest suppressed VX-680-induced endoreduplication and apoptosis

We next asked whether we could activate p53 before VX-680 addition and prevent cells from transiting a failed mitosis leading to tetraploidy. Nutlin-3 has been shown to induce a p53-dependent arrest.6, 7, 8, 21 Pretreating A549 cells with a low dose of nutlin-3 (5 μM) before VX-680 treatment arrested cells even after 48 h of VX-680 exposure. This is confirmed by a BrdU assay (Figure 5a). Remarkably, nutlin-3-pretreated cells assume a normal diploid DNA profile after cells are recovered in drug-free media (Figure 5b), in contrast to cells exposed only to VX-680. We found that normal nontransformed human epithelial keratinocytes (HEKs) are also susceptible to tetraploidy formation as a result of aurora kinase inhibition. HEK cells treated with VX-680 underwent DNA replication without cell division, resulting in a significant tetraploid G2/M population (Figure 5c). Nutlin-pretreated HEK cells did not show significant tetraploidy and maintained high cellular viability even after 5 days in nutlin-3, therefore suggesting that normal cells can withstand prolonged exposure to nutlin-3 and activation of p53 (Figure 5d and data not shown), consistent with other reports.6, 8 Therefore, nutlin pretreatment suppresses ploidy in response to aurora kinase inhibition. In addition, nutlin also suppress apoptosis induced by VX-680 (Supplementary data and Figure 7).

Figure 5
figure 5

Nutlin-3 suppressed endoreduplication in cells with functional p53. (a) BrdU staining reveals endoreduplicating cells after VX-680 treatment. Cells were treated with 200 nM VX-680 for 24 h and incubated with 10 μM BrdU for 30 min. Control cells and cells treated with nutlin-3 only or nutlin-3 in combination with VX-680 were assessed for BrdU incorporation. BrdU-positive cells in untreated control sample (upper left panel) and BrdU-positive cells that are >4N DNA (endoreduplicating cells) (upper right panel) are highlighted in the boxes and the percentages over the total cell population are indicated. (b) Cells pretreated with nutlin-3 assume a normal cell cycle profile comparable to the untreated cells after recovery from drug treatment. Schematic diagram show the timing and order of addition of drugs. At 48 h after VX-680 treatment, cells were washed in drug-free media and incubated in drug-free media for another 48, 96 or 120 h. Cell cycle distribution was analyzed by flow cytometry. (c) Nutlin-3 protects primary human keratinocytes from endoreduplication with minimal toxicity. Keratinocytes were treated with the indicated concentrations of VX-680 or nutlin-3 and were harvested for cell cycle analysis at 48 h after drug treatments. (d) Keratinocytes were either incubated with VX-680 (200 nM) or nutlin-3 (5 μM) for 64 h or pretreated with nutlin-3 (5 μM) for 16 h followed by incubation with VX-680 (200 nM) or nutlin-3 (5 μM) for a further 48 h. Bright field images of keratinoctyes at the end of the incubation period under the various indicated drug conditions are shown. Arrowheads indicate endoreduplicated cells

Full size image

Nutlin confers a long-term proliferative advantage that requires wild-type p53

Nutlin induced an arrest in HCT116p53+/+ but not in HCT116p53−/− (Figure 6b). The p53 dependency of nutlin is also demonstrated in A549 with attenuated p53 (Supplementary Figure S2). The proliferative capacity of nutlin-pretreated cells was compared with cells treated only with VX-680 in a colony formation assay. VX-680 treatment alone drastically reduced the number of A549 colonies (Figure 6c; upper panel). It is noteworthy that nutlin alone moderately decreases the number of surviving colonies. However, when the cells were pretreated with nutlin-3, the fraction of surviving colonies was enhanced up to 20-fold (Figures 6c and d). Similarly, HCT116p53+/+ cells show markedly better survival (up to 15-fold) when pre-treated with nutlin-3 than in the presence of VX-680 alone. In contrast, nutlin-3 pretreatment did not affect the colony survival rate in p53-deficient HCT116 (Figures 6c and d). We further studied the responses of p53 wild-type and p53-deficient cells in co-culture system, in which both cell types were cultured together for the duration of the experiment. HCT116p53+/+ and the HCT116p53−/− cells were distinguished using green fluorescent protein (GFP) and red fluorescent protein (RFP) fluorescent markers. Equal numbers of HCT116p53+/+ (green) and HCT116p53−/− (red) cells were mixed and plated in 10 cm plates, before exposure to a similar drug dosage regimen as described in Figure 6c. At 5 days after drugs removal, the total number of HCT116p53+/+ (green) cells in the mixed population pretreated with nutlin-3 was far more than the number of HCT116p53−/− (red) cells (Figure 6e) when compared with DMSO-treated control, which showed an equal proportion of both cell types. The mixed population treated only with nutlin-3 showed a decreased ratio of HCT116p53+/+ (green) cells to HCT116p53−/− (red) cells, perhaps due to the inhibition of cell proliferation over the period of drug treatment. These experiments reinforce the two key observations made in this study: (1) nutlin pretreatment renders increased survival of cells expressing wild-type p53, whereas single treatment with VX-680 abrogates proliferation of wild-type p53 cells, and (2) pretreatment with nutlin confers selective growth advantage on cells expressing wild-type p53, resulting in the preferential killing of p53-deficient cells by VX-680.

Figure 6
figure 6figure 6figure 6

Wild-type p53 is required for nutlin protection. (a) HCT116p53+/+ and p53−/− cells were incubated with 200 nM VX-680 and cells were harvested at the indicated time points for analysis of cell cycle distribution. Increased endoreduplication was observed in HCT116p53−/− at later time points (40 and 48 h). (b) Nutlin pretreatment protects HCT116 p53+/+ but not p53−/− cells from VX-680-induced endoreduplication. Cell cycle analysis was performed comparing HCT116p53+/+ and p53−/− cells that were pretreated with nutlin-3 (5 μM) for 16 h, followed by VX-680 (200 nM) and nutlin-3 (5 μM) for another 48 h. (c) Assessing the proliferative capacity of A549 and HCT116 cells after nutlin-3 and VX-680 treatment. Colony formation assay was used to measure the proliferative potential of each cell in the population. Cells were plated in 10-cm plates and exposed to nutlin-3 or DMSO (0.1%) for 16 h before incubation with VX-680 for 48 h. Colonies were counted 2 weeks after recovery in drug-free media. Cells were fixed with glutaraldehyde and stained with crystal violet. Plates are scanned and colonies were counted with an automated image analyzer. Three independent experiments were performed for each cell line. Results are represented in the graphs in (d). Each data point represents the mean±S.D. (e) HCT116p53+/+ and p53−/− cells were transduced with recombinant lentiviruses carrying GFP and RFP genes, respectively. Cells were selected in zeocin for 2 weeks to achieve the stable cell lines, HCT116p53+/+ GFP and HCT116p53−/− RFP (refer to Materials and Methods for details on lentivirus production). Equal numbers of HCT116p53+/+ GFP and p53−/− RFP were mixed and seeded to achieve 30% growth confluency. Cells were pretreated with nutlin-3 (5 μM) for 16 h before incubation with the indicated concentrations of VX-680 for another 48 h. Cells were then washed twice with drug-free media and allowed to recover for 5 days in fresh complete media. Microscopy images were taken with Deltavision (Axio). Nutlin-3 and DMSO (0.1%) treated controls are shown. Insets 1 and 2 show higher magnification of the images on the left

Full size image
Figure 7
figure 7figure 7

Nutlin-3 protects cells from apoptosis induced by VX-680. (a) Upper panel: Graphs show the quantification of the percentages of annexin V-positive cells in the p53-positive cells (A375, HT116p53+/+) and p53-compromised cells (HCT116p53−/−, MDA-MB-486) treated using a similar drug dosage regimen described in Figure 6b. Light grey columns represent nutlin pretreatment+VX-680 and dark grey columns represent VX-680 only treatment. Results are representative of three independent experiments. Lower panel: A375 cells were pretreated with 5 μM of nutlin-3 (or with DMSO as control) for 16 h before incubation with VX-680 for 48 h. Cells were then harvested and stained for annexin V to detect apoptotic cells. (b) Reversal of the order of drug addition results in synergistic activation of apoptosis. Annexin V-positive fractions were indicated as percentages in the boxes. Conditions for drug treatment on A375 cells: (1) control; (2) nutlin-3; (3) pretreatment with nutlin-3 followed by 48 h incubation with VX-680+nutlin-3; (4) VX-680 and (5) pretreatment with VX-680 followed by 48 h incubation with VX-680+nutlin-3; and (6) combined addition of nutlin-3 and VX-680. Nutlin-3 was used at 5 μM and VX-680 at 200 nM for all indicated drug conditions. (c) Western blot analysis of the A375 cells treated as described in b. The numbers 1–6 refer to the different conditions of treatment as described above in b. Bcl-2, p53 and p53-responsive gene products p21, MDM2 and Bax were detected using antibodies. Representative blots are shown. Each column on the right represents the ratio of intensity of Bax/Bcl2 bands, which is calculated by dividing the densitometric value of Bax by the corresponding Bcl2 value. (d) Comparison of the effects of VX-680, Nutlin-3+VX-680 and the reverse drug combination, VX-680 + Nutlin-3, on the survival of A375 cells. Colony survival assay was performed as described in Figure 6c. Each data point represents the mean±S.D. (n=3)

Full size image

p53 family member, p73, contributes to apoptosis and decreases polyploidy in response to VX-680

Recent reports suggest a role for p73 in the SAC and suppression of tetraploidy.22 Given that isolated p73 loss may result in impaired proliferation and premature senescence due to compensatory upregulation of p53,23 we decided to analyze the effects of loss of p73 functions in a p53-deficient background and asked whether p73 loss further attenuates the response to VX-680 in p53-deficient cells. We have generated two clones stably expressing p73DD, a dominant-negative suppressor of p73, in HCT116p53−/− background (Supplementary Figure S6). First, we established that the transcriptional activation of endogenous genes (p21 and PUMA) in VX-680-treated p53-deficient cells was dependent on p73 (Figure 8a). p21 levels were diminished by at least threefold and PUMA levels were decreased by twofold in p73DD-expressing cells. Second, we detected more polypoid cells containing 16N DNA in p73DD-expressing cells when compared with control cells treated with VX-680 (Figure 8d). Third, despite an increase in the percentage of polyploidy, there is an overall decrease in apoptosis in p73DD-expressing cells. A significant decrease in the annexin-V-positive fraction was observed in both p73DD/C1.4 and p73DD/C1.7 clones when compared with the control population (Figure 8e). Colony survival assay also indicated that loss of p73 transcriptional activity in p53-deficient cells increased colony numbers after treatment with VX-680 (Figure 8c). Taken together, our results support a role for p73 in activating apoptosis and suppressing polyploidy in response to VX-680 in p53-deficient cells, and loss of p73 leads to increased survival of cells with aberrant DNA content. Although p73 and p53 share many overlapping targets of transcriptional regulation, p73 does not seem (unlike p53) to be critical for regulating the expression of cyclins A and B; mRNA transcript and protein levels of cyclins A2 and B1 remain largely unchanged in p53−/−/p73DD cells treated with VX-680 (Supplementary Figure S7). Collectively, these data indicated that p53 and p73 may exert an effect through distinct mechanisms to suppress endoreduplication and apoptosis.

Figure 8
figure 8figure 8

Dominant-negative TAp73 mutant (p73DD) abrogates p73 transcriptional function and apoptotic activity under conditions of aurora kinase inhibition. (a) p73 mediates transcription of p21 and PUMA in response to VX-680 and nutlin-3. HCT116p53−/− cells were transduced with recombinant lentivirus carrying the TAp73 dominant-negative (p73DD) expression construct or the empty vector as control. Cells were selected with zeocin (200 μg/ml) for 12 days for stable transgene expression. Clone C1.4 was selected based on high expression of p73DD (Supplementary Figure S6). Cells were either treated with nutlin-3 (20 μM) or VX-680 (200 nM) and the mRNA levels of TAp73 target genes are quantified by real-time PCR. (b) mRNA and protein levels of p73DD in stably transduced p73DD/C1.4. Quantification of mRNA using real-time PCR. Primers specific for full-length p73 (TAp73 primers) and primers that also recognize the shorter p73DD mRNA (p73DD primers) were used. p73DD protein was tagged with a V5 epitope and its exogenous expression was detected by western blotting using antibody against V5 epitope. (c) Effect of exogenous expression of p73DD on the long-term survival of HCT116p53−/− cells treated with VX-680. Two selected clones, p73DD/C1.4 and p73DD/C1.7, which express a high level of p73DD protein (Supplementary Figure S6), and the parental HCT116 p53−/− cells were treated as described in Figure 6c. Exogenous expression of p73DD in p53-deficient cells correlates with increased survival after VX-680 treatment. (d) DNA histograms representing cell cycle profiles of HCT116p53−/− and HCT116p53−/−/p73DDC1.4 treated with VX-680 (50 nM). (e) Extent of apoptosis in p73DD expressing HCT116p53−/− clones assessed using annexin-V FITC antibody. (f) Ectopic expression of p73DD in wild-type HCT116. Wild-type HCT116 cells were transduced with either recombinant lentivirus carrying the p73DD gene or an empty vector. DNA profiles were analyzed by FACS at 5 days after transduction. ‘A’ annotates the sub-G1 cells

Full size image

Discussion

The aim of cancer chemotherapy is to selectively kill tumor cells without affecting normal cells. However, this is difficult to achieve as many chemotherapeutics are toxic to all dividing cells. This imposes a limitation on the therapeutic window within which a drug can be effectively administered, before the toxicity of this damage to the normal tissues becomes intolerable to patient. Therefore, finding a way to circumvent this problem may significantly improve the therapeutic indexes of existing drugs and discover some new uses for drugs that may have been failed previous preclinical tests because of toxicity issues. By exploiting the difference(s) between the abilities of normal and cancer cells to undergo cell cycle arrest, one could design a set of strategies to selectively induce growth arrest in normal cells while leaving cancer cells cycling in the presence of mitotic or S-phase inhibitors. p53 mutations occur in 50% of all human cancers, and in many remaining tumors, p53 pathways are inactivated through different mechanisms. Therefore, selective pharmacologic activation of p53 and the induction of a reversible cell cycle arrest in normal tissues could protect them from chemotherapeutic drugs that target the cell cycle, while not reducing the toxicity of the drugs for tumor cells in which the p53 pathway has been inactivated by mutation. The aims of this study are to analyze whether (1) normal cells (expressing wild-type p53) can be protected from endoreduplication and apoptosis induced by aurora kinase inhibition, and (2) pretreatment with an ‘arrest-inducing factor’ increases the selectivity of the second chemotherapeutic drug for cancer cells.

Although reports have suggested that nutlin induces senescence, a state of irreversible arrest, recent findings24, 25 and our results suggest that nutlin-3 treatment induces reversible growth arrest, but not senescence, in normal epithelial cells and cancer cells expressing wild-type p53. The nongenotoxic activation of p53 by nutlin-3 in vivo is well tolerated in nude mice without significant adverse side effects,8 and studies using a myc inhibitor have shown that the blocking of cell division in normal tissues for up to a week is reversible without pathology.26 Pharmacologic activation of p53 selectively protects p53 wild-type cells from endoreduplication and confers a long-term proliferative advantage (Figures 5 and 6). The increase in survival of p53-positive cells was partly attributed to the protection against apoptosis. As expected, the survival response of p53−/− cells to VX-680 was not affected by nutlin. It is interesting to note that a small fraction of p53−/− cells survive after VX-680. One may speculate that these polyploid cells have acquired resistance to apoptosis. Indeed, some tetraploidy cells are more resistant to DNA-damaging agents,27 providing selective survival advantage for p53 mutant cells.28 Alternatively, a small number of cells may have escaped the effects of VX-680 and remained diploid throughout the drug treatment and recovery period. In the latter case, it is possible that repeated drug exposure could further eliminate the ‘survivors’.

In a co-culture assay, we observed an increase in the proportion of p53+/+ (GFP-labeled) to p53−/− (RFP-labeled) cells only when they were pre-incubated with nutlin (Figure 6e). The change in the ratio of the two cell types is unlikely to be due to the difference in the rates of their cell growth, as the untreated control showed almost equal numbers of both cell types. Rather, we conclude that nutlin/VX-680 combination is selectively toxic to the p53-deficient cells, whereas cells expressing wild-type p53 retain their proliferative capacity. It is noteworthy that nutlin-only treatment led to a decrease in the ratio of green to red cells instead. This is likely due to an inhibition of the proliferation of p53+/+ cells but not p53−/− cells during the period of drug treatment and caution against the use of low dose of nutlin as a monotherapy agent, especially when the p53 gene status in the tumors is unclear. Our results suggest that nutlin/VX-680 combination may discriminate between p53-positive cells and p53-deficient cells, resulting in a selective killing of p53-compromised cells. Encouragingly, Sur et al.29 recently demonstrated that nutlin pretreatment in a mouse model protected against a major dose-limiting toxicity of a PLK1 inhibition, neutropenia, suggesting that the nutlin/VX-680 combination may also have therapeutic benefits.

Is the observed G2 arrest a result of activation of a DNA-damage-related checkpoint? Although aurora kinase inhibition led to increased genomic instability, as suggested by increased micronuclei formation (Supplementary Figure S4), an appearance of aberrant nuclear morphology and possibly an accumulation of chromosomal DNA damage, it is unclear whether the canonical ATM-dependent DNA damage response pathways are activated. We did not find any conclusive evidence that indicated a role for ATM in the checkpoint. In line with this, inhibitory phosphorylation at Tyr15-cdc2, a common mechanism for G2 checkpoint arrest after DNA damage,17 did not accumulate in response to aurora kinase inhibition (Figure 2c). How then is p53 activated, if not through an ATM-dependent pathway? There may be several possibilities, either as a direct consequence of inhibition of aurora kinases,30 and/or, indirectly (1) as a result of the disruption of centrosomal functions and activation of p38 kinase,31 or (2) through the cellular sensing of tetraploidy formation. The latter mechanism is still debatable as mammalian cells do not seem to have a mechanism for detecting ploidy.32 The tumor suppressor Fbwx7 regulates p53 activity and suppress polyploidy in response to mitotic toxins,33 and loss of Fbwx7 leads to an upregulation of aurora A protein levels.34 In this context, it would be interesting to further define the molecular relationships between Fbwx7/aurora A/p53 and ask whether aurora kinase inhibition/VX-680-induced activation of p53 requires Fbwx7.

The role of p73 in mediating the response to aurora kinase inhibition by VX-680 is unclear, although it has been implicated in tetraploidy.22, 23, 35, 36 We demonstrated that overexpression of p73DD, a dominant-negative inhibitor of p73, results in tetraploidy, even in unperturbed cells. Tetraploid p53−/− cells survive probably because of the absence of p53, as expression of p73DD in a wild-type p53 background led to apoptosis (Figure 8f). However, when HCTT16p53−/− cells were treated with VX-680, they underwent a p73-dependent apoptosis; p73DD overexpression resulted in a moderate increase in polyploidy and increase in surviving colonies (Figures 8c–e). Although the role of p73 in regulating ploidy may be due to its ability to interact with spindle checkpoint proteins,22 we cannot rule out a role for its transactivation-dependent functions at this point. Inhibition of p73 in HCTT16p53−/− downregulates the endogenous expression of PUMA in response to VX-680 (Figure 8a) and this correlates with the decrease in apoptosis (Figure 8e). Although the transcriptional targets of p73 and p53 overlap to some extent, the repression of cyclins A2 and B1 genes in response to VX-680 seem to be independent of p73, when p53 is absent (Supplementary Figure S7). This suggests that p73 and p53 may have distinct roles in the regulation of tetraploidy and polyploidy; p53 imposes a cell cycle arrest dependent on the activation of p21 and repression of cyclins A2 and B1, whereas p73 might have a more direct role in regulating tetraploidy and apoptosis, through its interaction with the mitotic checkpoint proteins and transcriptional regulation of PUMA. This may also explain why p53−/− cells are chromosomally stable unless perturbed, whereas deregulation of p73 (in p53−/− background) results in tetraploidy even when unperturbed.

In this study we have shown that a combination of nutlin/VX-680 results in the selective killing of p53-compromised cells and the reversible inhibition of proliferation in cells expressing wild-type p53. We propose that the combination of a p53-specific activator and an aurora kinase inhibitor may have therapeutic benefits and could potentially alleviate the side effects of aurora kinase inhibition. The clinical approval of nutlin or nutlin-like drugs will allow such exciting concepts to be tested in the clinic. Meanwhile, the results presented in this study and those of Sur et al.29 suggest that extensive pre-clinical testing in murine cancer models is justified.

Materials and Methods

Cell lines, primary lines and culture conditions

A549, U2OS, A375, H1299 and MDA-MB-486 were purchased from ATCC (Manassas, VA, USA) and cultured in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 1% penicillin/streptomycin (Invitrogen), 10% FBS and 2 mM glutamine (GIBCO, Carlsbad, CA, USA). Wild-type HCT116, HCT116p53−/− and HCT116p21−/− cells were kind gifts from Dr B Vogelstein (John Hopkins University School of Medicine, Baltimore, MD, USA). HCT116 and derivatives were cultured in McCoy's 5A media (Sigma, St Louis, MO, USA) supplemented with 1% penicillin/streptomycin, 10% FBS and 2 mM glutamine. Primary human keratinocytes were purchased from Invitrogen and cultured in defined keratinocyte serum-free media (Invitrogen). All cell lines were cultured in a CO2 incubator (5% CO2 and 21% O2) at 37°C.

Antibodies and reagents

Antibodies against p53 (DO-1), p21 (Ab118) and MDM2 (2A9) were kindly provided by Dr B Vojtesek (Masaryk Memorial Cancer Institute, Bruno, Czech Republic). Antibodies against cyclin E (M-20), cyclin B1 (GNS1), cyclin A2 (H432) and Rb (Rb1) were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Phospho-Rb (Ser807/811), phospho-ATM (Ser1981; 10H11.E12) and Phospho-cdc2 (Tyr15; 10A11) were from Cell Signaling (Danvers, MA, USA). V5 epitope antibody was from Invitrogen. Antibody for IP of Cdk2 was described previously.37 Suc1 conjugated to agarose beads was from Millipore Upstate (Billerica, MA, USA) (14–132.) Nutlin-3 was from Calbiochem (San Diego, CA, USA) and VX-680 was from American Customs Chemical Corporation (San Diego, CA, USA). Both compounds were reconstituted in DMSO. For use in cell treatment, the final DMSO concentration in the media did not exceed 0.1% (v/v).

Plasmids and primers

Cloning of lentiviral vectors: cDNA of the respective genes was amplified by PCR using the following gene-specific primers:

pBI-p73 wild-type/EGFP,38 (Dr B Vogelstein, Addgene) was used as the template for PCR. Highlighted in bold are the attB1 and attB2 sites. PCR fragments are recombined with the pDONR vector (Invitrogen) using BP clonase according to the manufacturer's protocol. Successful clones were verified by sequencing. Expression clones pLenti4-p53, pLenti4-p73DD, pLenti4-GFP and pLenti4-RFP were generated through recombination of the verified pDONR vectors and pLenti4-V5-DEST vector (Invitrogen) using LR clonase (Invitrogen). Successful clones were verified by sequencing. The pLenti4-p53 plasmid was used as a template for the generation of the pLenti4-p53siRes construct. Primers (fwd) 5′-GAGTGGAAGGAAGTTCGCATGCGGAGTATTTGGATGACAG-3′ and (rev) 5′-CTGTCATCCAAATACTCCGCATGCGAACTTCCTTCCACTC-3′ were used to incorporate the base changes (underlined) using QUIKChange site-directed mutagensis kit (Stratagene, La Jolla, CA, USA).

siRNA transfection

p53 siRNA oligo 1 (p53si) 5′-GCAGUCAGAUCCUAGCGUCUU-3′, cyclin A2 siRNA 5′-CTUCUTTGUTUGGTTCCTG-3′ and cyclin B1 siRNA 5′-UCTTTCGCCTGUGCCTUTT-3′ were from Dharmacon (Lafayette, CO, USA). Cells were plated in six-well dishes overnight in antibiotic-free media. Cells were incubated with Optimen (GIBCO) for an hour before transfection using Lipofectamine 2000 (Invitrogen) and siRNA (30 nM) according to the manufacturer's recommendation. Cells were either harvested at 48 h after transfection or incubated with VX-680 for another 48 h before analysis.

Viral production and transduction

293T cells were transfected with pLenti4-empty vector or pLenti4 construct carrying the gene of interest using Lipofectamine 2000 (Invitrogen). Supernatants were collected at 48 and 72 h after transfection, filtered through a 0.45 μM filter and concentrated through ultracentrifugation. Viral titers were estimated using serial dilutions of the concentrated virus stock and determining the number of antibiotic (zeocin) resistance colonies at 3 weeks after transduction and selection. On the average, viral titers were estimated to be 5.6 × 107 to 2 × 108 TU/ml. In vitro transduction was performed by plating cells in 24-well plates in DMEM supplemented with 10% FCS. After overnight incubation, cells were transduced with an appropriate titer (5 MOI) and incubated overnight. The next day, fresh media was added and incubated for another 24 h before cells were selected for antibiotic resistance in zeocin (250 μg/ml) containing media for 2 weeks.

Ad-CMVp21 was purchased from Vector Biolabs (Philadelphia, PA, USA). HCT116p53−/− cells were incubated in McCoy's media (without serum) containing the adenovirus for an hour. The virus-containing media was removed and replaced with fresh media containing 200 nM of VX-680. Cell cycle analysis was performed using fluorescence-activated cell sorting (FACS) at the end of 72 h of drug treatment. The results were analyzed using FlowJo software (TreeStar Inc., Ashland, OR, USA) and percentages of polyploidy cells (count of polyploidy cells over the total parent cell population) were calculated.

FACS and apoptosis analysis

For analysis of cell cycle distribution, cells were harvested and fixed in 70% ethanol/PBS solution. Cells were stained in propidium-iodide containing solution (25 μg/ml propidium iodide supplemented with 1 mg/ml RNase A, in PBS (pH 7.8); Sigma Chemical, St Louis, MO, USA) for 15 min at room temperature. For analysis of apoptosis, cells were harvested without fixation. Apoptosis was evaluated using the annexin V (fluorescein isothiocyanate (FITC))–propidium iodide binding assay (Roche, Indianapolis, IN, USA). The extent of apoptosis was quantified as a percentage of annexin V-positive cells over the total cell population. Flow cytometric analysis was performed on a BD LSR II System (BD Biosciences, Stockholm, Sweden). Data were analyzed using FlowJo software and ModFit LT (Verity Software House, Topsham, ME, USA).

Colony survival assay

A total of 1000 cells were plated on each 10 cm plate and allowed to adhere overnight. Cells were either left in media with <0.1% DMSO as control or incubated with 5 μM nutlin-3 for 16 h before adding the indicated concentrations of VX-680. After 48 h, cells were washed and recovered in fresh media for 12 days. Cells were then fixed in 6% glutaraldehyde (Sigma) and stained with crystal violet. Colonies were counted using GelCountTM (Oxford Optronix, Oxford, England). Each condition was carried out in triplicate.

Western blot and immunoprecipitation

For western blotting, cell extracts were prepared using NP-40 lysis buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% NP-40, 5 mM EDTA supplemented with protease and phosphatase inhibitor cocktails (Sigma)). For immunoprecipitation, 1 × 107 cells were harvested and lysed in extraction buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 10 mM β-glycerophosphate, 10% glycerol, 0.5% Tween-20, 1.0 mM EDTA, 2.5 mM EGTA, 0.5 mM DTT and protease inhibitors (10 μg/ml each of leupeptin, chymostatin and pepstatin (Chemicon, Temecula, CA, USA)). Cells were incubated with the extraction buffer for 30 min at 4°C followed by centrifugation at 14 000 r.p.m. at 4°C for 30 min. A total of 3–5 μg of antibody were incubated with lysates (300 μg) in 750 μl of EBN buffer (80 mM β-glycerophosphate, pH 7.3, 20 mM EGTA, 15 mM MgCl2 and 0.5% NP-40) containing 1 mg/ml ovalbumin, 2 mM NaF and protease inhibitors for 3 h at 4°C. Then, 10 μl of protein G beads was added to the mix and incubated for another hour. Protein immunoprecipitates were washed twice in EBN buffer and twice in EB buffer (EBN without the NaCl and NP-40).

Kinase assay

The immunoprecipitated beads were resuspended in 5 μl of EB buffer, 10 mM DTT and 20 to 50 μM ATP. Each sample was incubated with 5–10 μCi [γ-32P]ATP, 1.5 μg histone H1 (Roche, no. 1004875) in a final volume of 16 μl. After incubation for 30 min, reactions were terminated by the addition of 5 μl 5 × SDS-PAGE sample buffer. After electrophoresis on 12.5% polyacrylamide gels, phosphorylation was analyzed by autoradiography and quantified by phosphorimage analysis.

BrdU labeling

Cells were incubated in DMEM containing 10 μM BrdU for 30 min in a 37°C/5% CO2 incubator. Detection of BrdU-labeled cells was performed using the In Situ Cell Proliferation Kit from Roche. In brief, cells were harvested and fixed using a 70% ethanol/50 mM Glycine (pH 2.0) on ice. Cells were pelleted and resuspended in the HCl-denaturation solution at room temperature for 20 min. Cells were then pelleted and 50 μl anti-BrdU-FLUOS antibody was added to the cells and incubated for 45 min at 37°C in a humidified chamber. For staining of total DNA, 7-AAD was added to the cells, followed by incubation at room temperature for 10 min. Cells were analyzed immediately using flow cytometry (BD LSR II System).

Quantitative reverse transcription-PCR

Total RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA, USA). The RNA was quantified using spectrophotometric analysis and used for quantitative real-time PCR. The primers used for each target analyzed are available on request. The RNA Master Power SYBR Green Mix (Roche) was used for quantification of mRNA levels.

Immunostaining and fluorescence microscopy

Indirect immunofluorescence was carried out as described previously.39 γ-H2AX (Novus Biologicals, Littleton, CO, USA, NB 100-383) and anti-rabbit Alexa 488–coupled secondary antibody (Molecular Probes, Carlsbad, CA, USA) were used. Nuclei were counterstained with Hoechst 33342 (Molecular Probes). Immunofluorescence was visualized using AxioImager Z1 (Zeiss, Gottingen, Germany).