|Home | About | Journals | Submit | Contact Us | Français|
The genetic or functional inactivation of p53 is highly prevalent in human cancers. Using high-content videomicroscopy based on fluorescent TP53+/+ and TP53−/− human colon carcinoma cells, we discovered that SP600125, a broad-spectrum serine/threonine kinase inhibitor, kills p53-deficient cells more efficiently than their p53-proficient counterparts, in vitro. Similar observations were obtained in vivo, in mice carrying p53-deficient and -proficient human xenografts. Such a preferential cytotoxicity could be attributed to the failure of p53-deficient cells to undergo cell cycle arrest in response to SP600125. TP53−/− (but not TP53+/+) cells treated with SP600125 became polyploid upon mitotic abortion and progressively succumbed to mitochondrial apoptosis. The expression of an SP600125-resistant variant of the mitotic kinase MPS1 in TP53−/− cells reduced SP600125-induced polyploidization. Thus, by targeting MPS1, SP600125 triggers a polyploidization program that cannot be sustained by TP53−/− cells, resulting in the activation of mitotic catastrophe, an oncosuppressive mechanism for the eradication of mitosis-incompetent cells.
The genetic or functional inactivation of the tumour suppressor p53 is (one of) the most frequent molecular characteristic(s) of human cancer (Cheok et al, 2011). It has been estimated that nowadays 11 million people are living with a cancer that contains an inactivating mutation of TP53 and additional 11 millions have neoplasms in which the p53 pathway is interrupted due to defects in the molecules that act either upstream or downstream of p53 (Brown et al, 2009).
p53 has recently been shown to control cellular homeostasis via multiple mechanisms (Vousden & Ryan, 2009). Still, p53 is best known as the major mediator of the cellular response to stress, a phenomenon that entails both transcriptional (nuclear) and non-transcriptional (cytoplasmic) p53 functions (Green & Kroemer, 2009; Vousden & Prives, 2009). Thus, in response to adverse conditions, p53 can stimulate the execution of apoptotic cell death, mainly through the transactivation of several pro-apoptotic genes (Beckerman & Prives, 2010) or physical interactions with proteins that regulate mitochondrial outer membrane permeabilization (Galluzzi et al, 2011a). Alternatively, stress-activated p53 can arrest the cell cycle, either by transactivating inhibitors of cyclin-dependent kinases or, perhaps, by direct interactions with essential cell cycle regulators (Tritarelli et al, 2004). Therefore, p53-deficient cells are notoriously resistant against the induction of apoptosis (Cheok et al, 2011) and fail to undergo cell cycle arrest in response to DNA damaging agents, implying that they become genomically instable (Talos & Moll, 2010). The absence of p53 reportedly favours spontaneous tetraploidization in several distinct contexts (Davoli & de Lange, 2011; Ganem et al, 2007). Moreover, in response to mitotic inhibitors such as microtubular poisons, p53-incompetent cells can override mitotic checkpoints and hence can become polyploid, a property that is thought to contribute to chromosomal instability (Aylon et al, 2006; Finkin et al, 2008; Ha et al, 2007; Vitale et al, 2010) as well as to resistance to chemo- and radiotherapy (Castedo et al, 2006; Galluzzi et al, 2011b; Shen et al, 2008).
Thus, p53 plays a prominent role in coupling the effects of DNA damaging agents (such as ionizing radiations or chemicals including platinum derivatives and topoisomerase inhibitors) to the therapeutic induction of senescence or apoptosis. Given the prevalence and functional impact of p53 defects, many groups have attempted to identify pharmacological agents that preferentially kill p53-incompetent cells (Brown et al, 2009). While some of these compounds directly interact with mutant p53 proteins and reestablish their functions via conformational effects (Athar et al, 2011; Selivanova, 2010; Wiman, 2010), others were designed to inhibit the p53-targeting ubiquitin ligase HDM2, leading to increased p53 concentrations (Shangary & Wang, 2009). Moreover, some agents kill p53-deficient cancer cells based on their increased tendency to undergo polyploidization. VX-680 (an inhibitor of Aurora kinases) causes the endoreplication of p53-deficient cells, resulting in their death (Gizatullin et al, 2006). A combination of irradiation and BI-2536 (a inhibitor of Polo-like kinase 1, PLK1) kills TP53−/− cells more efficiently than their TP53+/+ counterparts (Sur et al, 2009), and the selective toxicity of GSK461364A (another PLK1 inhibitor) against p53-deficient cells has been attributed to their failure to undergo cell cycle arrest at a tetraploid stage (Degenhardt et al, 2010). Similarly, 7-hydroxystaurosporine (UCN-01, an inhibitor of the checkpoint kinase CHEK1) has been shown to abrogate DNA damage-induced cell cycle arrest selectively in p53-defective cancer cells (Levesque et al, 2005).
In this study, we launched yet another attempt to identify compounds that kill p53-deficient cells more efficiently than their p53-positive counterparts. Using a novel videomicroscopic assay, we identified SP600125 as an agent whose cytotoxicity is exacerbated in conditions in which p53 is deleted, depleted or inhibited, in vitro and in vivo. The mechanisms through which SP600125 mediates these effects involve the selective polyploidization of p53-deficient cells due to the inhibition of the mitotic kinase MPS1. Preferentially in the context of p53 deficiency, this results in abortive mitosis, followed by the activation of mitotic catastrophe and cell death via the mitochondrial pathway of apoptosis.
p53-deficient (TP53−/−) human colon carcinoma HCT 116 cells or their p53-proficient wild-type (WT) counterparts stably expressing red fluorescent protein (RFP)- or green fluorescent protein (GFP)-tagged histone 2B, respectively, were co-cultured in 96-well plates for 48 h in the absence or presence of 480 distinct compounds from the ICCB library of bioactive agents. Throughout this period, the ratio of green WT versus red TP53−/− HCT 116 cells was quantitatively monitored by videomicroscopy. In line with the fact that p53 often mediates the cytotoxic effects of DNA-damaging agents, 10-hydroxycamptothecin and cisplatin depleted more efficiently WT than TP53−/− cells. In contrast SP600125, a broad-spectrum inhibitor of serine/threonine kinases including Aurora kinase A (IC50 = 60 nM) and B (190 nM), CDC7/DBF4 (1.59 µM), CDK2/CYCA (880 nM), CK2 (650 nM), FLT3 (90 nM), Haspin (510 nM), IKK2 (1 µM), JAK3 (410 nM), KIT (870 nM), MELK (110 nM), PDK1 (1.06 µM), PIM2 (1.05 µM), RET (1.09 µM), SYK (1.8 µM), TRKA (70 nM), TYK2 (880 nM) and VEGFR3 (300 nM) (Colombo et al, 2010), but best known for inhibiting c-Jun N-terminal kinases (JNK) 1–3 (100–200 nM) (Bennett et al, 2001; Heo et al, 2004) and MPS1 (IC50 1.95 µM) (Chu et al, 2008; Schmidt et al, 2005), was the most efficient agent in increasing the ratio between WT and TP53−/− cells (Fig 1A–C). These results are consistent with previous observations (Mingo-Sion et al, 2004), and were validated using two alternative methods for measuring cell proliferation and death, namely the xCELLigence® system, which quantifies the impedance of adherent cells (Fig 1D), and cytofluorometry (Fig 1E) (Castedo et al, 2002; Galluzzi et al, 2009, 2007). Upon co-staining with the vital dye propidium iodide (PI) and the mitochondrial transmembrane potential (Δψm) sensor DiOC6(3), the frequency of dying (DiOC6(3)low PI−) and dead (DiOC6(3)low PI+) cells was markedly increased among SP600125-treated TP53−/− cells but less so among WT control cells (Fig 1E). Altogether, these results suggest that SP600125 is endowed with the capacity of selectively killing p53-deficient tumour cells.
Since SP600125 can perturb mitosis and/or induce endoreplication (Kim et al, 2010; Schmidt et al, 2005) we investigated its effect on the cell cycle of WT and TP53−/− HCT 116 cells. WT cells treated with SP600125 progressively acquired a 4n DNA content (which corresponds to the normal amount of DNA in the G2 or M phase of the cell cycle), while TP53−/− cells accumulated a higher DNA content (8n or 16n), indicating that they became tetraploid or higher-order polyploid (Fig 2A and B). Accordingly, SP600125 blocked the incorporation of the thymidine analog 5-ethynyl-2′-deoxyuridine (EdU) into DNA in WT but not in TP53−/− cells. Moreover, at odds with their WT counterparts, TP53−/− cells continued to incorporate EdU even when they possessed a 4n or 8n DNA content (Fig 2C). A significant fraction of TP53−/− (but not WT) cells exhibited histone H3 phosphorylation, an indicator of ongoing mitosis, or stained positively for the mitotic-specific phosphoepitope MPM2 at the 8n stage (Fig 2D and Fig S1 of Supporting Information), Moreover, in response to SP600125, WT (but not TP53−/−) cells exhibited a marked decline in cyclin B (the cyclin associated with the G2/M phase) levels and increased amounts of cyclin E (the cyclin associated with the G1/S phase) (Fig 2E and F and Fig S2 of Supporting Information). Moreover, in response to SP600125, the mitotic index (MI; i.e., the fraction of cells that manifest morphological signs of mitosis) decreased more rapidly in WT cells than in their TP53−/− counterparts (Fig 3A). SP600125 indistinguishably induced the disappearance of anaphases and telophases in WT and TP53−/− cells (Fig 3A), suggesting that, in the continuative presence of SP600125, TP53−/− cells underwent abortive mitoses. Accordingly, in response to SP600125, metaphases became disorganized in both WT and TP53−/− cells, as SP600125 abolished the spindle assembly checkpoint (SAC), stimulated the degradation of cyclin B in cells with morphologically evident metaphases, and increased the number of centrosomes per cell (Fig 3B and C and Fig S3 of Supporting Information).
These results suggest that SP600125 augments the DNA content (preferentially of TP53−/− cells) in the absence of normal mitotic chromosome separation, nuclear division (karyokinesis) and daughter cell formation (cytokinesis). Accordingly, interphase FISH indicates that the SP600125-mediated increase in DNA content is not accompanied by a proportional increase in chromosome number but entails a higher hybridization signal, suggesting chromatid segregation failure (Fig 3D and E). In line with these observations, HCT 116 cells in which one single chromosome locus is labelled by GFP (because that locus contains multiple lacO copies that are stained by transgene-expressed LacI-GFP) (Thompson & Compton, 2010) responded to SP600125 with an increase in DNA content that was not paralleled by an increase in chromosome number, in particular upon siRNA-mediated depletion of p53 (Fig 3F and G and Fig S4A of Supporting Information). In this system, chromosomal GFP labelling became more intense, further suggesting that SP600125 induces chromatin duplication in the absence of chromosomal separation.
Next, we investigated whether the effects of SP600125 depend on the inhibition of MPS1 or are mediated by off-target mechanisms. Transfection-enforced expression of WT MPS1 and more so of an SP600125-refractory MPS1 variant (MPS1-M602Q) (Schmidt et al, 2005), reduced the capacity of SP600125 to trigger polyploidization in TP53−/− cells (Fig 4A and B and Fig S4B of Supporting Information). The overexpression of MPS1 (and more so of MPS1-M602Q) also reduced the frequency of SP600125-treated TP53−/− cancer cells manifesting polyploid S phases and mitoses (Fig 4A, C and D) and staining positively for cyclin B (Fig 4E). In line with the interpretation that SP600125 mediates its effects (at least in part) by inhibiting MPS1, we observed that the siRNA-mediated depletion of MPS1 induced polyploidization more efficiently in TP53−/− cells than in their WT counterparts (Fig 4F and G and Fig S4C of Supporting Information). Similarly, the highly specific MPS1 inhibitor reversine (Santaguida et al, 2010) induced the preferential polyploidization of TP53−/− cells (Fig 4H). Nonetheless, depletion or inhibition of MPS1 failed to kill p53-deficient cells more efficiently than their WT counterparts (Fig 4I and J and Fig S5 of Supporting Information), suggesting that MPS1 inhibition is required but not sufficient for the selective killing of TP53−/− cells by SP600125. Finally, the preferential killing of p53-negative over p53-expressing cells was not reproduced by the selective depletion of Aurora kinase A, Aurora kinase B, CHEK1, CHEK2, FLT3, JNK1, JNK2 and MEK1 alone or in various combinations including with MPS1 (Figs S6–S8 of Supporting Information).
To understand the link(s) between the polyploidization of TP53−/− cells and their enhanced mortality, we performed detailed videomicroscopic analyses of histone 2B (H2B)-GFP-expressing WT and TP53−/− cells responding to SP600125 (Fig 5A). While there were no differences in the proliferative potential of untreated WT and TP53−/− cells, dramatic discrepancies became apparent upon exposure to SP600125 (Fig 5B). In the presence of SP600125, indeed, WT cells did not manifest a mitotic arrest but typically underwent one round of abortive mitosis featuring a transient phase of chromatin condensation but not a symmetric metaphase plate, followed by karyokinesis failure, no cytokinesis and either quiescence (no further mitosis) or cell death (Fig 5A and B and Video 1 of Supporting Information). In contrast, more than half of SP600125-treated TP53−/− cells underwent two consecutive rounds of abortive mitosis (in the absence of karyokinesis and cytokinesis), leading to increased nuclear size and DNA content (Fig 5A–D and Video 2 of Supporting Information). Importantly, in a significant fraction of these cells, the first or second round of abortive mitosis was followed by the activation of mitotic catastrophe and cell death (Vitale et al, 2011), which always occurred 10 h after from mitotic failure (and never during mitosis) (Fig 5A–D and Video 3 of Supporting Information). In line with these results, non-proliferating TP53−/− cells (cultured either in presence of low serum concentrations or at high confluence) were partially protected against SP600125, and their selective susceptibility to the drug was lost (Fig 6A). This finding unambiguously proves the causal link between proliferation and cell death as induced by SP600125. Clonogenic assays confirmed that SP600125 was more efficient in permanently reducing the clonogenic potential of TP53−/− cells than that of their WT counterparts (Fig 6B). This loss of proliferative capacity was accompanied by the activation of mitochondrial pathways of apoptosis, as documented by the Δψm dissipation (see above), the release of cytochrome c from mitochondria and the subsequent activation of caspase-3 and chromatin condensation (Fig 6C). The knockdown of pro-apoptotic proteins such as PUMA and BAX (but not BAK) reduced, while the depletion of Bcl-2-like apoptosis inhibitors (i.e. Bcl-2, Bcl-XL and MCL-1) enhanced, the fraction of TP53−/− cells succumbing to SP600125 (Fig 6D). The broad-spectrum caspase inhibitor Z-VAD-fmk inhibited cell death as well as the appearance of a sub-diploid population of apoptotic bodies as induced by SP6000125 (Fig 6E).
Altogether, these results suggest that SP600125 kills TP53−/− cells secondary to one or two rounds of mitotic abortion and following the activation of the mitochondrial pathway of apoptosis. Thus, SP600125 triggers one program of mitotic catastrophe, preferentially in p53-deficient cells.
SP600125 selectively killed TP53−/− (as opposed to WT as well as to BAX−/−, CHK2−/− and CDKN1A−/−) HCT 116 cells, suggesting that its effect is specific (Fig 7A). The depletion of p53 by three distinct siRNAs and the transient transfection of two dominant-negative p53 mutants (His175 and His273) sensitized WT HCT 116 cells to SP600125, mimicking the effect of gene knockout (Fig 7B and C and Fig S2D of Supporting Information). Similarly, the deletion, depletion or inhibition of p53 sensitized cell lines other than HCT 116 to SP600125. This applied to human colorectal cancer RKO cells subjected to the siRNA-mediated knockdown of p53 (Fig 7D and Fig S2E of Supporting Information), human osteosarcoma Saos-2 cells expressing p53-His273 in a doxycycline-inducible fashion (Fig 7E), human ovarian carcinoma SKOV-3 cells expressing two different variants of dominant-negative p53 (Fig 7F) and mouse embryonic fibroblasts (MEFs) lacking Tp53 due to homologous recombination (Fig 7G). Moreover, SP600125 killed non-immortalized Tp53−/− mouse adult fibroblasts (at early culture passages) more efficiently than their p53-sufficient counterparts (Fig 7H). These results suggest that the capacity of SP600125 to selectively eliminate TP53−/− cells can be generalized. Importantly, not only the absence of p53 but also the continuative presence of SP600125 is required for efficient cell death induction, confirming the notion that, in p53-deficient contexts, SP600125 promotes mitotic aberrations that activate a delayed program of mitotic catastrophe-induced cell death (Fig S9 of Supporting Information). Of note, the intraperitoneal administration of SP600125 did not affect the growth of p53-proficient WT HCT 116 tumours xenografted in athymic mice. However, the same systemic regimen was able to inhibit the growth of isogenic TP53−/− HCT 116 tumours (Fig 7I), underscoring the notion that SP600125 can mediate selective therapeutic effects against p53-deficient tumours in vivo.
In this paper, we identified SP600125 as an agent that induces polyploidization and cell death preferentially in p53-deficient cells. SP600125 is a multi-kinase inhibitor (Colombo et al, 2010) and its selective toxicity on TP53−/− cells could not be recapitulated by the selective depletion of several among its most prominent pharmacological targets including JNK1, JNK2 or MPS1, alone or in combination. This is in contrast with previous results indicating that SP600125-induced cell death depends on JNK (Mingo-Sion et al, 2004) and that TP53−/− cells are sensitive to JNK2 depletion (Potapova et al, 2000). Rather, our results suggest that SP600125 mediates its cytotoxic effects via a multipronged inhibitory action on several serine/threonine kinases. Nonetheless, it appears that the inhibition of MPS1 is at least co-responsible for the preferential effects of SP600125 on the cell cycle of TP53−/− cells. Indeed, transfection with MPS1 (and more so with an SP600125-resistant mutant of MPS1) limited the polyploidization of TP53−/− cells responding to SP600125. Thus, one of (but not the sole) targets of SP600125 underlying its therapeutic profile against p53-deficient cells is MPS1. Highly specific MPS1 inhibitors are currently in preclinical and clinical development (Colombo et al, 2010; Hewitt et al, 2010; Kwiatkowski et al, 2010; Santaguida et al, 2010; Sliedrecht et al, 2010). However, there are no reports indicating that such compounds would preferentially kill p53-deficient cells. It should be noted that the simultaneous inhibition of several kinases may exert synergistic pro-apoptotic actions (Olaussen et al, 2009) and that less specific kinase inhibitors may have broader anticancer effects than highly targeted agents, as this has been determined in several clinical studies (Knight et al, 2010). Thus, the relative non-specificity of SP600125 may turn into a therapeutic advantage.
Previous groups reported that SP600125 can induce polyploidization in NIH 3T3 MEFs (Schmidt et al, 2005) and in multiple different human cancer cells, including non-small cell lung carcinoma (A549) (Miyamoto-Yamasaki et al, 2007), small lung carcinoma (Calu-1), cervical carcinoma (HeLa; MacCorkle & Tan, 2004), breast cancer (MCF-7, MDA MB-231, and 21PT) (Mingo-Sion et al, 2004), osteosarcoma (U2OS) (Kim et al, 2010) and mantle cell lymphoma cells (Wang et al, 2009). Two different mechanisms have been proposed for this effect, namely subsequent rounds of mitosis in the absence of karyokinesis and cytokinesis or, alternatively, direct endoduplication from the G2 phase, due to the inhibition of CDK1. This latter mechanism has been described to occur in p53-sufficient cells, and has been attributed to the (indirect) inhibition of CDK1 secondary to that of Aurora kinase A and PLK1 (Kim et al, 2010). To our knowledge, no systematic studies have been performed so far to compare the efficacy of SP600125-induced polyploidization in p53-sufficient and p53-deficient cells with one exception (Kuntzen et al, 2005). In this study, upon SP600125 administration for a short time period (16 h), the authors observed no differences in cell cycle distribution and cell death activation in presence or absence of p53. At odds with this, our results suggest that p53 has a profound impact on the efficacy (and perhaps the mechanism) of polyploidization as induced by SP600125. In response to SP600125, TP53−/− cells reach higher level of ploidy than their TP53+/+ counterparts, and they do so while contracting their chromosomes periodically in abortive (pro-)metaphases without forming bona fide metaphase plates, perhaps because centrosomes aggregate in a single cluster or they are not able to orchestrate a correct mitotic spindle. Thereafter, the chromosomes of TP53−/− cells decondense with no signs of anaphase-associated chromosomal movements and in the absence of complete chromosome separation, as indicated by FISH and live chromosome labelling. TP53−/− cells are able to go through this round of abortive mitosis up to three times in the presence of SP600125. In any of these rounds (most often in the first or second one, rarely in the third one), the mitochondrial pathway of apoptosis gets activated, most frequently with a latency of approximately 10 h after mitosis. Thus, the killing of TP53−/− cells triggered by SP600125 is closely coupled to abortive mitosis, indicating that SP600125 constitutes a novel inducer of mitotic catastrophe (Vitale et al, 2011). Future studies must decipher the mechanisms that link SP600125-induced abortive mitosis to apoptosis, and address the potential clinical utility of this agent.
Unless otherwise indicated, media and supplements for cell culture were purchased from Gibco-Invitrogen (Carlsbad, CA, USA) and plasticware from Corning B.V. Life Sciences (Corning, NY, USA).
SP600125 (anthra [1,9-cd] pyrazole-6 (2H)-one) was purchased from Sigma–Aldrich (St Louis, MO, USA) and stocked as a 30 mM solution in DMSO. Reversine (2-(4-orpholinoanilino)-6-cyclohexylaminopurine) was purchased from Calbiochem (San Diego, CA, USA) and stocked as a 10 mM solution in DMSO. The pan-caspase inhibitor N-benzyloxycarbonyl-Val-Ala-Asp.fluoromethylketone (z-VAD-fmk) was obtained from Bachem Bioscience (Bubendorf, Switzerland) and stocked as a 50 mM solution in DMF. The appropriate amount of DMSO and/or DMF was always employed for negative control conditions.
A total of 4 × 104 WT HCT 116 cells stably expressing H2B-GFP and TP53−/− HCT 116 cells stably expressing H2B-RFP were seeded at a ratio of 1:1 in Black/Clear 96-well Imaging Plates (BD Biosciences, San Diego, CA, USA). Twenty-four hours later, cells were treated with the agents from the ICCB Known Bioactives Library (Enzo Life Sciences Inc., Farmingdale, NY, USA) at final concentrations ranging from 1 to 12 µM and kept in a Cytomat 2c automated incubator (Thermo Fisher Scientific, Waltham, MA, USA). Every 6 h, plates were instantly transferred to a BD Pathway 855 automated microscope (BD Biosciences) by means of a robotized Twister II plate handler (Caliper Life Sciences, Hopkinton, MA, USA), and four view fields in the centre of each well were acquired using a 20X objective (Olympus, Center Valley, PA, USA), for a total assay time of 48 h. Images were analysed for cell density and fluorescence intensity with the BD Attovision software (BD Biosciences) and the slope of growth curves of each subpopulation over 48 h was calculated by using the linear regression function of Microsoft Excel (Microsoft Co., Redmond, WA, USA).
Custom-designed siRNAs targeting MPS1 (siMPS1_1 sense 5′-CCCAGAGGACUGGUUGAGUdTdT-3′; siMPS1_2 sense 5′-GCAACCACUUAUGGUACUGdTdT-3′; siMPS1_3 sense 5′-GCACGUGACUACUUUCAAAdTdT-3′) (Stucke et al, 2004; Xu et al, 2009), p53 (sip53_1 sense 5′-GACUCCAGUGGUAAUCUACdTdT-3′; sip53_2 sense 5′-GCAUGAACCGGAGGCCCAUdTdT-3′; sip53_3 sense 5′-CUACUUCCUGAAAACAACGdTdT-3′) (Gu et al, 2004; Martinez et al, 2002), AURKB (5′-GGAGGAGGAUCUACUUGAUTT-3′), CHEK1 (sense 5′-UCGUGAGCGUUUGUUGAACTT-3′), CHEK2 (sense 5′-UGUGUGAAUGACAACUACUTT-3′), Bcl-2 (Vicencio et al, 2009), Bcl-XL (de La Motte Rouge et al, 2007), MCL1 (Kepp et al, 2009) and PUMA (Vitale et al, 2007), were purchased from Sigma-Proligo (The Woodlands, TX, USA).
Alternatively, siRNA duplexes for the downregulation of FLT3 and MPS1 (siGENOME Smart-pool M-003137 and M-004105, respectively), were purchased from Dharmacon (Chicago, IL, USA), siRNAs targeting JNK1 (sc-29380) from Santa Cruz (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), siRNAs for the down-regulation of AURKA, JNK2 and MEK1 (SIHK0142, SIHK1224 and SIHK1098, respectively), from Sigma–Aldrich and siRNAs for the down-regulation of BAK1 and BAX (Hs_BAK1_5 and Hs_BAX_10 HP Validated siRNAs, respectively) from Qiagen (Hilden, Germany) (de La Motte Rouge et al, 2007). As a control, a non-targeting siRNA with a sequence unrelated to the human genome was used (UNR, sense 5′-GCCGGUAUGCCGGUUAAGUdTdT-3′).
Cells pre-seeded in 12-well plates were transfected with plasmids at 80% confluence by means of the Attractene Transfection reagent (Qiagen) as recommended by the manufacturer, or with siRNAs at 30–40% confluence by means of the HiPerFect® transfection reagent (Qiagen), as previously described (Hoffmann et al, 2008). After 48 or 72 h, transfection efficiency was determined by immunoblotting.
Immunofluorescence microscopy determinations were performed as previously described (Vitale et al, 2007, 2008). For videomicroscopy, HCT 116 cells stably expressing a H2B-GFP fusion were grown in 96-well imaging plates (BD Biosciences) under standard culture conditions and subjected to pulsed observations (every 7 min for up to 48 h) with a BD pathway 855 automated live cell microscope (BD Biosciences). Images were analysed with the open source software Image J. Cell fate profile are illustrated as previously published (Rello-Varona et al, 2010).
p53 is the tumour suppression protein whose function is most frequently lost in human cancer. Thus, more than 20 million people worldwide live with cancers in which p53 is mutated or the p53 pathway is interrupted due to the inactivation of molecules that act either upstream or downstream of p53. The loss of p53 reduces the efficacy of several conventional anticancer agents, meaning that p53-deficient tumours are intrinsically chemoresistant. This work has been designed to identify small molecules that preferentially kill cancer cells lacking p53 expression or expressing a mutated p53 variant.
A screen in which p53-sufficient and p53-deficient tumour cells were systematically compared for their response to known or experimental anticancer agents led to the identification of one molecule, SP600125, that killed p53-deficient cells more efficiently than their p53-expressing counterparts. This property of SP600125 was validated on multiple cell lines that differed relative to the function of the p53 system. In particular, we found that human p53-deficient cancer cells growing on immunodeficient mice responded to chemotherapy with SP600125 in conditions in which similar cancer cells that express intact p53 failed to do so. We characterized the mechanisms through which SP600125 can kill p53-deficient (but not p53-sufficient) cancer cells. We found that p53-sufficient cells underwent a cell cycle arrest upon exposure to SP600125 and ultimately survived. In contrast, p53-deficient cells exposed to SP600125 undertook a series of abortive mitoses (featuring karyokinesis but no cytokinesis), eventually generating giant cells that finally underwent apoptotic cell death.
Altogether, these results delineate a novel strategy to selectively target cells that lack functional p53. SP600125 exemplifies a class of drugs that kill p53-deficient cells in a preferential fashion, based on the incapacity of such cells to undergo a cell cycle arrest in particular circumstances. Hence, SP600125 (and its analogues/derivatives) might be used for cancer chemoprevention (for eliminating pre-malignant cells that have inactivated p53) or chemotherapy of p53-deficient cancers.
Athymic nu/nu female mice (age ~42 days, body weight ~20 g, provided by the Institut Gustave Roussy, Villejuif, France in-house animal facility) were used throughout this study in strict compliance with widely accepted ethical guidelines for animal experimentation. Mice were kept in Makrolons type III wire mesh laboratory cages (Charles River, Boston, MA, USA), under poor germ conditions at 24°C and 50–60% humidity, and were allowed for food and water ad libitum. Light cycle was artificially controlled to provide 14 h of light (from 06:30 a.m. to 08:30 p.m.). After 4 days of acclimation period, mice were subcutaneously xenografted with 2 × 106 WT or TP53−/− HCT 116 cells (Vitale et al, 2007). After 1 week, mice were injected i.p. with 5 mg/kg SP600125 (or an equivalent volume of vehicle) every 2–3 days. Due to its limited solubility in water, SP600125 was injected in 2% DMSO, 2% polyethylene glycol 600 (PEG600) and 2% Tween 80 in PBS. Statistical analyses were carried out by means of the SigmaStat package (Systat Software, San Jose, CA, USA). Statistical significance was determined by means of two-tailed unpaired Student's t-test (*p < 0.05) in the context of a pairwise comparison procedure.
Unless otherwise specified, all experiments were carried out in triplicate parallel instances and independently repeated at least three times. Data were analysed with Microsoft Excel (Microsoft) and statistical significance was assessed by means of two-tailed unpaired Student's t-test (*p < 0.05; **p < 0.001).
The complete set of experimental and statistical procedures can be found online as Supporting Information.
We are indebted to Bert Vogelstein for WT and TP53−/− HCT 116 cells, Tyler Jacks for MEF cells and Duane A. Compton for HCT 116 cells with a single GFP-labelled chromosome. We thank Klas Wiman for the generous gift of cell lines expressing inducible p53 mutants. We acknowledge the kind gift of cDNAs encoding MPS1 by Rene Medema and of p53 dominant negatives mutants by Dr. Thierry Soussi. We thank Caterina Tanzarella for the help in the design of interphase FISH experiment and for the critical reading of the manuscript. IV is supported by the Ligue Nationale contre le Cancer, SRV by the Fondation de France and LG from Apo-Sys and SAM by the Higher Education Commission (HEC) of Pakistan. This work was supported by grants to GK from the Ligue Nationale contre le Cancer (Equipe labellisée), Agence Nationale pour la Recherche (ANR), European Commission (Active p53, Apo-Sys, ChemoRes, ApopTrain), Fondation pour la Recherche Médicale (FRM), Institut National du Cancer (INCa), Cancéropôle Ile-de-France, Fondation Bettencourt-Schueller and the LabEx Onco-Immunology.
Supporting Information is available at EMBO Molecular Medicine online.
The authors declare that they have no conflict of interest.
MJ, OK, FB, LS, GM, SAM, SRV, DL, AA, MT, FS, FH, GP and MC designed and performed experiments and analysed data; IV designed and performed experiments, analysed data and wrote the paper; LG designed experiments and wrote the paper; GK designed experiments, wrote the paper and directed biological studies.
Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors.