p53 Small-Molecule Inhibitor Enhances Temozolomide Cytotoxic Activity against Intracranial Glioblastoma Xenografts

Attempts and approaches at manipulating p53 activity in treating human cancer have been numerous and diverse. For example, viral-mediated introduction and expression of wild-type (wt) TP53 in p53-defective tumor cells has been extensively investigated for more than a decade, including through clinical trial activity (1). Alternative approaches for increasing tumor cell wt p53 activity include the use of small molecules that promote p53 transcription and the use of compounds that inhibit p53 interaction with murine double minute 2 (2).

Perhaps because of it being counter to conventional thinking about the role of tumor suppressor genes in cancer etiology, as well as being counterintuitive about the way in which tumor suppressor genes are viewed in relation to the treatment of cancer, there has been relatively little research directed toward the development of antitumor therapeutic strategies that include a p53 inhibitory component. Indeed, as a monotherapy, such a treatment approach could promote increased tumor cell proliferation and decreased tumor cell apoptosis. However, the potential consequences of attempted cell cycling by tumor cells with damaged DNA, resulting from genotoxic therapy with concurrent inhibition of p53, are interesting to consider. In fact, results from several studies involving in vitro investigation of tumor cell lines support enhanced cytotoxic chemotherapeutic response in association with p53 inhibition (3–6). Furthermore, with respect to GBM, the p53 small-molecule inhibitor pifithrin-α, which was identified nearly a decade ago in association with a chemical library screen (7), has been shown to enhance the in vitro cytotoxic effect of temozolomide (TMZ), a DNA alkylator, as well as the cytotoxic effect of chloroethylating nitrosoureas such as carmustine (8, 9).

In addition to reasons described above, in vivo investigation of p53 small-molecule inhibitors, as part of a cancer treatment strategy, has been hindered due to limitations imposed by physical properties of the pifithrin-α reference compound (10). Recently, however, derivatives of the reference compound were described with respect to their potential in vivo use (11). In the current study, we have tested one of these compounds, using an intracranial GBM xenograft therapy-response model, and present results indicating its enhancement of TMZ antitumor activity in vivo and in a manner that is dependent on tumor cell p53 status.

In vitro experiments. GBM xenografts used in this study have previously been described (12, 13), as has the modification of xenografts for bioluminescence imaging (13). Xenograft cells were cultured as nonadherent neurospheres in neurobasal media (Invitrogen), whereas U87 cells (American Type Culture Collection) were propagated as monolayer cultures in DMEM supplemented with 10% FCS. Temozolomide (TMZ; obtained as Temodar from Schering-Plough) and active-form p53 inhibitor (cyclic pifithrin-α p-nitro, Calbiochem) were dissolved in DMSO (Sigma-Aldrich) as 20 and 5 mmol/L stock solutions, respectively. For bioluminescence viability analysis, cells were treated with DMSO, TMZ (added to concentrations of 50 or 100 μmol/L), p53 inhibitor (concentration of 10 μmol/L), or a combination of TMZ and p53 inhibitor, with chemical agents added to media once a day for 3 consecutive days. Cell culture specimens were examined for bioluminescence signal using a Xenogen imaging system (Caliper Life Sciences), following the addition of 25 μL of 20 mg/mL sodium luciferin (Gold Biotechnology) in PBS (Invitrogen).

Flow cytometry cell cycle analysis. U87 cells were treated once a day for 3 d with DMSO only, 10 μmol/L p53 inhibitor, 50 or 100 μmol/L TMZ, or 50 or 100 μmol/L TMZ + 10 μmol/L p53-inhibitor. At 1, 4, and 7 d after final treatment, the cells were harvested, washed with PBS, and fixed with cold 70% ethanol. Cells were stained with propidium iodide and examined by flow cytometry (BD LSR II, Becton-Dickinson), with results analyzed using FlowJo software.

Immunoblot analysis. Primary antibodies used for immunoblot analysis (previously described; see ref. 13) were for detection of poly(ADP-ribose) polymerase (Cell Signaling Tech), phospho-histone H2AX (Cell Signaling Tech), p53 (Sigma-Aldrich), p21 (Santa Cruz Biotech), MGMT (R&D Systems), or β-tubulin (Millipore). Secondary antibodies used were antimouse or antirabbit (Cell Signaling Tech) or antigoat (Santa Cruz Biotech).

Methylation-specific PCR. Analysis of MGMT promoter methylation was as described by Esteller and colleagues(14). PCR products were resolved in 4.5% agarose gels (NuSieve 3:1, Lonza, Inc.) and were subsequently stained with ethidium bromide.

In vivo experiments. Procedures used for intracranial tumor therapy-response experiments, including monitoring of tumor growth and response to therapy by bioluminescence imaging, have previously been described (13). Treatment groups for the experiments reported here were as follows: 3 d i.p. DMSO and oral suspension vehicle (OraPlus, Paddock Laboratories) by gavage (control group); 3 d 10 mg/kg TMZ in oral suspension vehicle for GBM 12, GBM 14, and U87, and 50 mg/kg for GBM 26; 3 d 0.25 mg p53 inhibitor precursor (pifithrin-α p-nitro, Calbiochem) in DMSO by i.p. injection; or 3 d combination TMZ and p53 inhibitor as indicated for their administration as single agents.

GBM xenograft explant cultures were incubated with TMZ +/− cyclic pifithrin-α p-nitro (Supplementary Fig. S1: see ref. 11) to assess the in vitro effects of this p53 inhibitor on tumor cell TMZ cytotoxic response. Results for p53^wt GBM 26 showed substantial decreases in viable cell number resulting from TMZ + p53 inhibitor cotreatments, in relation to cells treated with TMZ alone (Fig. 1A and B). Importantly, cyclic pifithrin-α p-nitro alone showed no antitumor effect. In contrast to the results for GBM 26, GBM 12 cells, which express no endogenous p53 (p53^null: see ref. 12), showed similar cytotoxic response to TMZ irrespective of the presence or absence of the p53 inhibitor (Fig. 1C). As shown by others (11), cyclic pifithrin-α p-nitro inhibits DNA damage–associated induction of the negative cell cycle regulator p21, but only in p53^wt cells (Fig. 1D). In addition, and consistent with the p53 small-molecule inhibitor treatments, pretreating cells with p53 siRNA enhanced the TMZ cytotoxic response of GBM 26 cells, relative to pretreating cells with nonspecific siRNA, whereas GBM 12 cells responded similarly to TMZ whether preincubated with p53 or nonspecific siRNA (Supplementary Fig. S2).

To further evaluate cyclic pifithrin-α p-nitro for in vitro activity, as well as to contrast results from the use of this compound with those previously reported for the pifithrin-α reference compound (Supplementary Fig. S1C; see ref. 8), U87 cells were incubated with TMZ +/− cyclic pifithrin-α p-nitro and examined by flow cytometric cell cycle analysis. Similar to results reported for the reference compound (8), higher sub-G₁ cell fractions were observed in samples treated with TMZ + cyclic pifithrin-α p-nitro, relative to TMZ treatment alone. Sub-G₁ ratios for U87 cells subjected to cotreatment versus treatment with 100 μmol/L TMZ only were 1.97, 2.25, and 1.54 at days 1, 4, and 7 following final treatment, respectively (Fig. 2A and B). Consistent with the GBM xenograft explant culture cytotoxic response results described above (Fig. 1), results from U87 flow cytometry analysis showed little difference in cell cycle distributions between U87 treated with cyclic pifithrin-α p-nitro and mock-treated control cells (Fig. 2A).

Immunoblot analysis of U87 cells incubated with TMZ +/− cyclic pifithrin-α p-nitro showed increased poly(ADP-ribose) polymerase cleavage resulting from combined treatment, in comparison with cells treated with the same concentration of TMZ alone (Fig. 2C), thereby suggesting increased caspase activity in association with the enhanced cytotoxic response to TMZ + p53 inhibitor. Heightened apoptotic response to combination treatment was also indicated by increased phospho-H2AX in cells incubated with TMZ + p53 inhibitor, relative to cells incubated with the same concentration of TMZ alone. Levels of p53 protein were similar between culture treatments (Fig. 2C), consistent with the inhibitory activity of pifithrin-α compounds involving suppression of existing p53 activity, rather than suppression of p53 expression.

Previous study of pifithrin-α p-nitro derivatives has revealed that 2-aminothiazole salt cyclization (Supplementary Fig. S1A and S1B) is required for the formation of active p53 inhibitor, and that this cyclization occurs spontaneously in vivo, albeit over a period of several hours (11). Because active-form p53 inhibitors have relatively short half-lives in vivo, it has been suggested that the kinetics of inactive, open-form precursor conversion to corresponding cyclized, active-form compound can be used to advantage for achieving prolonged cell exposure to active state inhibitors in vivo and thereby achieve prolonged p53 inhibitory effect (11). In accord with this line of reasoning, we investigated the open-form precursor of cyclic pifithrin-α p-nitro (Supplementary Fig. S1A) for effect on TMZ in vivo antitumor activity, using a previously described intracranial GBM xenograft therapy-response model (13). Mice injected with p53^wt tumor cells (GBM 14, GBM 26, and U87) or with cells from p53^null GBM 12 were monitored for luminescence increases indicative of log-phase growth of injected cells, at which time TMZ (gavage) and p53 inhibitor precursor (i.p. injection) were administered singularly or concurrently once a day for 3 consecutive days (treatment periods indicated by vertical gray columns, Fig. 3A and B). For mice injected with p53^wt GBM, survival was significantly extended by combined TMZ + p53 inhibitor treatment, relative to TMZ treatment alone, and this effect was observed for three p53^wt tumors (P < 0.001 for each; Fig. 3A; Table 1). Survival benefits were consistent with corresponding results from bioluminescence monitoring, which revealed prolonged decrease of p53^wt intracranial tumor luminescence in mice receiving TMZ + p53 inhibitor (Fig. 3B). It is worth noting that p53 inhibitor precursor enhancement of TMZ cytotoxicity was observed irrespective of p53^wt GBM sensitivity to TMZ alone, as TMZ-resistant GBM 26, with unmethylated MGMT promoter and highly expressed MGMT protein, and GBM 14 and U87, in which MGMT promoter methylation is accompanied by a lack of detectable MGMT protein (Fig. 3C and D), all showed significantly increased response to combination treatment. Conversely, mice with p53^null GBM 12 intracranial tumor did not experience significant survival benefit from combined TMZ + p53 inhibitor precursor treatment in comparison with mice receiving treatment with TMZ only (Fig. 3A; Table 1).

Research for examining potential clinical applications of p53 small-molecule inhibitors has often involved experimental paradigms that address the protection of normal cells, rather than determining effects on tumor cell response to therapy that includes p53 inhibition. Specifically, there is considerable interest about whether p53 small-molecule inhibitors can be used to spare normal cells from p53-mediated apoptosis that occurs from oxidative stress during stroke, renal injury, or cardiac arrest (15), as well as from the genotoxic effects of cancer treatments (7). Small-molecule inhibition of p53 in tumor cells, for enhancing the anticancer effects of therapy, has not attracted much attention, although there is ample related evidence in support of this concept. This support, however, primarily stems from the use of viral gene product expression (4) and RNA interference (3, 6) to suppress tumor cell p53 function, and the perspective of the associated studies has been to investigate the role of p53 as a determinant of tumor cell response to cytotoxic therapy, rather than to pursue p53 as a target for augmenting the effects of therapy.

In the current study, we have contrasted p53^wt versus p53^null GBM xenograft response to cytotoxic therapy that includes the use of a small-molecule inhibitor of p53. Significantly extended survival benefit from combination therapy was evident among a series of mice with intracranial p53^wt GBM, whereas mice with intracranial p53^null GBM did not experience survival benefit from combination therapy. The issue about the response of p53 missense mutants to p53 inhibitor treatment is complex due to the diversity of p53 mutations in human cancer, the distinct transcriptional effects of individual p53 mutations, and the transcription-independent cellular activities of mutant p53 (16–18). Current studies are in progress to address the variability of p53 missense mutant GBM response to p53 inhibitor–augmented cytotoxic therapy.

There are numerous candidates to consider as mediators of p53 inhibitor–enhanced cytotoxic response, with our results currently supporting more extensive investigation of p53 inhibitor–mediated suppression of p21 (Fig. 1D), as well as investigation of apoptotic response effectors that are inferred by our preliminary results (Fig. 2; Supplementary Fig. S3). There are, of course, additional aspects of this cytotoxic response to be investigated, one of which is indicated in a recent report showing that p53 inhibition can suppress TMZ-associated induction of MGMT (19). In the case of the tumor cell sources used here, this could be a contributing factor associated with p53 inhibitor enhancement of TMZ cytotoxicity against GBM 26, for which MGMT is unmethylated and expressed, but is probably less likely for GBM 14 and U87 tumors, for which MGMT is presumably silenced through MGMT promoter methylation (Fig. 3C and D).

In addition to investigating the molecular biology of pathways and proteins associated with enhanced TMZ cytotoxic effect through concurrent p53 inhibition, the identification of surrogate markers that can be used to assess p53 inhibitor activity in tumors is critically important. The ability to analyze tumors for temporal variations in the expression and/or activity of biomarkers that respond to p53 inhibition, in combination with corresponding analysis of tumor extracts for temporal variations in p53 inhibitor content (e.g., through use of high-performance liquid chromatography), will provide key information about inhibitor pharmacodynamics for optimizing combination therapy dosing and sequencing.

Results presented here provide some basis for considering the subgroup of GBM that would benefit from therapy that includes p53 inhibition. As concerns this point, it is significant that p53 inhibition enhances TMZ cytotoxicity in p53^wt tumors irrespective of their sensitivity to TMZ monotherapy. Consequently, this therapeutic approach could conceivably benefit the two thirds of GBM patients whose tumors are of wt p53 status (20). Looking beyond the GBM-specific interests of this report, it seems plausible that the results of our study are generalizable to other types of cancer as well as to other types of genotoxic therapy.

An additional observation associated with this study is that TMZ + p53 inhibitor treatment is without adverse effect, as evidenced by autopsy analysis of animal subjects that showed no abnormal cellular proliferation in tissue from mice receiving TMZ + p53 inhibitor treatment (e.g., gastrointestinal tract; Supplementary Fig. S4A). This, in combination with the long-term survival of mice that received TMZ + p53 inhibitor (see GBM 14 results in Fig. 3A), as well as the similarity in group body weight patterns for TMZ only and TMZ + p53 inhibitor treated mice (Supplementary Fig. S4B), suggests a lack of acute as well as long-term effects of this therapy. Nonetheless, the simultaneous administration of a DNA-damaging agent with an inhibitor of DNA repair could promote undesirable systemic side effects, including the development of independent cancers in other organs. To address this concern, future research should perhaps include strategies for the local delivery of inhibitor to tumor to minimize side effects resulting from combination therapy. Irrespective of the methods of delivery that are used, the long-term consequences of this therapeutic approach will need to be investigated in detail.

No potential conflicts of interest were disclosed.

Grant support: NIH grants NS049720, CA097257, CA108961, CA127716, and CA100011.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.