Report

A Chemical Inhibitor of p53 That Protects Mice from the Side Effects of Cancer Therapy

See allHide authors and affiliations

Science  10 Sep 1999:
Vol. 285, Issue 5434, pp. 1733-1737
DOI: 10.1126/science.285.5434.1733

Abstract

Chemotherapy and radiation therapy for cancer often have severe side effects that limit their efficacy. Because these effects are in part determined by p53-mediated apoptosis, temporary suppression of p53 has been suggested as a therapeutic strategy to prevent damage of normal tissues during treatment of p53-deficient tumors. To test this possibility, a small molecule was isolated for its ability to reversibly block p53-dependent transcriptional activation and apoptosis. This compound, pifithrin-α, protected mice from the lethal genotoxic stress associated with anticancer treatment without promoting the formation of tumors. Thus, inhibitors of p53 may be useful drugs for reducing the side effects of cancer therapy and other types of stress associated with p53 induction.

p53 functions as a key component of a cellular emergency response mechanism. In response to a variety of stress signals, it induces growth arrest or apoptosis, thereby eliminating damaged and potentially dangerous cells from the organism (1). The p53 gene is lost or mutated in most human tumors (2). Lack of functional p53 is accompanied by high rates of genomic instability, rapid tumor progression, resistance to anticancer therapy, and increased angiogenesis (3). p53 deficiency in mice is associated with a high frequency of spontaneous cancers (4). On the basis of all these observations, the inactivation of p53 is viewed as an unfavorable event, and much effort has been expended to facilitate anticancer treatment by restoring p53 function.

However, the role of p53 in cancer treatment is not limited to its involvement in killing tumor cells. In mice, the p53 gene is highly expressed in several normal tissues, including lymphoid and hematopoietic organs, intestinal epithelia, and the testis, and it is these tissues that are damaged by anticancer therapy (5,6). p53-dependent apoptosis occurs in sensitive tissues shortly after gamma irradiation (6,7), and p53-deficient mice survive higher doses of gamma irradiation than do wild-type animals (8). These data indicate that p53 is a determinant of the toxic side effects of anticancer treatment, and thus may be an appropriate target for therapeutic suppression to reduce the damage to normal tissues (9). This approach would be applicable only for tumors that lack functional p53.

To explore the feasibility of this approach, we isolated a chemical inhibitor of p53 and characterized its effects in vitro and in vivo. As a primary read-out system for screening of p53 inhibitors, we used the mouse ConA cell line, which carries the wild-type p53 gene and the bacterial lacZ reporter gene under the control of a p53-responsive promoter (6). Because the inhibitors of p53 transactivation may not necessarily suppress apoptosis, the selected compounds were then tested in mouse cell line C8, a conventional model of p53-dependent apoptosis (10). Individual compounds from a diverse collection of 10,000 synthetic chemicals (from DIVERSet, Chembridge Corporation, San Diego, California), in combination with the potent p53 inducer, doxorubicin (Dox), were added to ConA cells growing in 96-well plates, and activation of lacZ-encoded β-galactosidase (β-Gal) was determined by routine X-Gal staining after 24 hours of treatment (Fig. 1A). We chose compounds that attenuated β-Gal induction in Dox-treated cells with no effect on cell growth or survival rate and picked one of them for a detailed characterization. This synthetic, water-soluble, and stable compound, with a molecular weight of 367, blocked activation of p53-responsivelacZ in ConA cells induced not only by Dox but also by ultraviolet (UV) light (Fig. 1, B and C) and gamma radiation (10) in a dose-dependent manner. This compound also reduced activation of endogenous cellular p53-responsive genes, includingcyclin G, p21/waf1/cip1, and mdm2, as judged by Northern (RNA) blot hybridization (Fig. 1D). We named it pifithrin-α (PFTα, an abbreviation for “p–fifty three inhibitor”) (Fig. 1E).

Figure 1

Identification of a compound, PFTα, that inhibits p53-dependent transactivation of p53-responsive genes. (A) Screening of a chemical library (Chembridge) in ConA cells carrying the lacZ reporter gene under the control of a p53-responsive promoter. lacZ-encoded β-Gal was detected by X-Gal staining. ConA cells plated in 96-well plates were treated for 14 hours with Dox (0.2 μg/ml) in the presence of compounds at concentrations of 10 to 20 μM. The well containing PFTα is indicated by a red arrow. Control wells (left column) contained untreated ConA cells, treated with Dox alone or with Dox plus 20 mM sodium salicilate (Na Sal), which inhibits p53 transactivation (19). DMSO, dimethyl sulfoxide. (B) PFTα inhibits activation of lacZ in ConA cells treated with different dosages of UV (35, 25, 15, and 7 J/m2) or Dox (for 10 hours at concentrations of 0.1, 0.2, 0.4, and 0.6 μg/ml). A fragment of a 96-well plate and regions of cell monolayers are shown. (C) Dependence of β-Gal activity in UV-irradiated (25 J/m2) ConA cells on the concentration of PFTα (10, 20, and 30 μM). Cells were collected 8 hours after UV treatment, and β-Gal expression in the extracts was estimated by a colorimetric assay (7). (D) PFTα inhibits UV-induced transactivation of cyclin G, p21/waf1, andmdm2, which are known p53-responsive genes (20). Northern blots of RNA from ConA cells: u/t, untreated; PFT, incubated for 8 hours with 10 mM of PFTα; UV, 8 hours after UV treatment (25 J/m2). (E) Chemical structure of PFTα [2-(2-imino-4,5,6,7-tetrahydrobenzothiazol-3-yl)-1-p-tolylethanone].

To analyze the effect of PFTα on p53-mediated apoptosis, we used mouse embryo fibroblasts transformed withE1a+ras, line C8, which undergo rapid p53-dependent apoptosis in response to a variety of treatments (11). A 10 μM concentration of PFTα inhibited apoptotic death of C8 cells induced by Dox, etoposide, Taxol, cytosine arabinoside (Fig. 2A), UV light, and gamma radiation (10). The anti-apoptotic activity of PFTα was p53-dependent because it had no effect on the survival of UV-treated C8-56 cells, in which p53 is suppressed by a dominant negative mutant, GSE56 (12) (Fig. 2B), or ofE1a+ras-transformed fibroblasts from p53−/− mouse embryos, line A4 (10,11). Similarly, PFTα inhibited p53-dependent growth arrest of human diploid fibroblasts in response to DNA damage but had no effect on p53-deficient fibroblasts. Suppression of both p53-dependent apoptosis and growth arrest by PFTα correlated with an increase in long-term cell survival, as judged by the results of assays of colony growth (13).

Figure 2

PFTα modulates cellular functions of p53. (A) A 10 μM concentration of PFTα inhibits apoptosis in C8 cells treated with various cytotoxic compounds. C8 cells were treated 48 hours with the indicated concentrations of drugs, with and without PFTα. At the end of treatment, the number of attached cells was estimated by staining with 0.25% crystal violet in 50% methanol, followed by elution of the dye with 1% SDS. Optical density (530 nm) reflecting the number of stained cells was determined with a Bio-Tek EL311 microplate reader. (B) Inhibition of apoptosis by PFTα is p53-dependent (comparison of two cell lines differing in p53 status). Addition of PFTα (10 and 20 μM, as indicated) inhibits apoptosis of C8 cells but has no effect on the UV sensitivity of C8-56 cells, which express the dominant negative p53 mutant GSE-56 (12). Cells were plated at three different densities in 96-well plates and were treated with different doses of UV (indicated in seconds of illumination) with or without 10 or 20 μM concentrations of PFTα. Plates were fixed 48 hours after treatment and stained with crystal violet. (C) PFTα changes the cell cycle distribution of ConA cells 24 hours after gamma irradiation (10 Gy) but does not affect ConA cells expressing the dominant negative mutant of p53 GSE56 (ConA-56).

To test whether PFTα affects p53-dependent cell cycle checkpoint control, we compared the cell cycle distribution of treated ConA cells before and after gamma irradiation (Fig. 2C). PFTα had no effect on nonirradiated cells, but it dramatically increased the proportion of G2-arrested cells 24 hours after gamma irradiation. This effect was p53- dependent because it was completely abrogated in the GSE56 cells, which have no functional p53 (Fig. 2C).

To determine the stage in the p53 pathway that is targeted by PFTα, we analyzed the compound's effects when apoptosis was induced by direct overexpression of wild-type p53 with no genotoxic stress applied. PFTα dramatically inhibited killing of p53-deficient human sarcoma Saos-2 cells that in control cells occurred within 48 hours after transient transfection with p53-expressing plasmids (Fig. 3A). This suggests that PFTα acts downstream of p53. PFTα did not alter phosphorylation or sequence-specific DNA binding of p53 in ConA cells after DNA-damaging treatments, as judged by protein immunoblotting in combination with two- dimensional protein analysis and gel shift assays (Fig. 3, B through E). However, it slightly lowered the levels of nuclear, but not cytoplasmic, p53 induced by UV irradiation. In contrast, PFTα did not affect the nuclear-cytoplasmic ratio of the p53-inducible p21waf1 protein (Fig. 3D). These observations suggest that PFTα may modulate the nuclear import or export (or both) of p53 or may decrease the stability of nuclear p53. Whether this is the only mechanism of PFTα action remains unclear.

Figure 3

Effects of PFTα on the p53 pathway. (A) PFTα inhibits apoptosis in Saos-2 cells transiently expressing p53. Cells were transfected with the plasmid DNA expressing green fluorescent protein (GFP) with the 5× excess of the plasmid carrying either wild-type human p53 (middle and bottom) or with no insert (top). Transfected cells were maintained with (bottom) or without (top and middle) PFTα. The majority of fluorescent cells transfected with p53-expressing plasmid undergo apoptosis 48 hours after transfection (middle). Apoptosis was inhibited in the presence of PFTα (bottom). (B) Comparison of spectra of p53 protein variants in the lysates of UV-irradiated (25 J/m2) ConA cells in the presence of different concentrations of PFTα (0, 10, 20, and 30 μM) using two-dimensional protein gel electrophoresis. (C) PFTα partially and in a dose-dependent manner inhibits p53 accumulation in ConA cells after UV treatment (results of protein immunoblotting). PFTα was added to the cells before UV treatment and total cell lysates were prepared 18 hours later. (D) PFTα changes the nuclear and cytoplasmic distribution of p53. Nuclear and cytoplasmic fractions were isolated from UV-treated ConA cells 6 hours after UV irradiation. p53 and p21waf1 proteins were detected by immunoblotting. The nuclear and cytoplasmic ratios of p53 but not p21waf1 are significantly decreased in the PFTα-treated cells. (E) PFTα does not affect DNA-binding activity of p53. Results of a gel shift assay using cell lysates from either untreated or UV-irradiated ConA cells grown in medium containing PFTα are shown. The right half of the gel shows a supershift of the p53-binding DNA fragment by monoclonal antibody Pab421. The decline in the amount of bound DNA is proportional to the overall decrease in p53 content in the presence of PFTα.

We also characterized the dependence of cell survival after genotoxic stress on the time and duration of PFTα application (13). PFTα had almost no protective effect if it was added before (up to 18 hours) and removed immediately before UV treatment of C8 cells. However, a short 3-hour incubation with PFTα after UV treatment had a pronounced protective effect, whereas a 24-hour incubation provided maximal protection. PFTα did not rescue UV-irradiated cells from apoptosis if it was added 3 hours after UV irradiation. These observations indicate that although PFTα can efficiently inhibit p53-dependent apoptosis, its effects are reversible and require the presence of the drug. Because many cells survived a lethal dose of UV irradiation after only 3 hours of incubation with PFTα, we conclude that the UV-induced apoptotic death signal is significantly reduced within several hours and completely disappears within 24 hours of irradiation.

To analyze the in vivo effects of PFTα, we treated two different strains of mice (C57BL and Balb/c) with lethal and sublethal doses of whole-body gamma radiation. We compared (i) untreated unirradiated mice, (ii) unirradiated mice that received a single intraperitoneal (i.p.) injection of PFTα, (iii) untreated gamma-irradiated mice, and (iv) mice injected intraperitoneally with PFTα immediately before gamma irradiation. PFTα treatment completely rescued mice of both strains from 60% killing doses of gamma irradiation (8 Gy for C57BL and 6 Gy for Balb/c). Significant protection was also seen at higher doses of irradiation that were lethal for control animals (Fig. 4A). PFTα-injected mice lost less weight than irradiated mice that were not pretreated with the drug (Fig. 4B). PFTα did not protect p53-null mice from lethal irradiation, which confirms that it acts through a p53-dependent mechanism in vivo (13).

Figure 4

Effects of PFTα in Balb/c and C57BL6 mice treated with whole-body gamma irradiation (results of representative experiments are shown). (A) A single i.p. injection of 2.2 mg per kilogram of body weight (mg/kg) of PFTα has a strong rescuing effect in both mouse strains. PFTα-injection abrogated the gradual loss of weight by C57BL6 mice after 8 Gy of gamma irradiation (the observed increase in the weight of the nonirradiated mice reflects the normal growth of young 5-week-old animals). The experiments were repeated at least three times with 10 mice per each experimental subgroup. Red traces indicate PFTα-treated animals; blue traces indicate PFTα-untreated animals. (B) PFTα abrogates p53-dependent regulation of DNA replication after genotoxic stress in vivo. 14C-thymidine (10 mCi per animal) was injected intraperitoneally into untreated wild-type or p53-null mice and in gamma irradiated mice 8 hours after 10 Gy of whole-body gamma irradiation. A subgroup of the gamma-irradiated animals received one i.p. injection of PFTα (2.2 mg/kg) 5 min before irradiation. Mice were killed 24 hours after 14C-thymidine injection, and whole-body sections (25 μm thick) were prepared and exposed with x-ray film in order to monitor the tissue distribution of14C. Red arrows indicate skin and intestine. (C) Comparison of tissue morphology and apoptosis (TUNEL staining) in the epithelium of the small intestine of C57BL6 p53 wild-type mice 24 hours after 10 Gy of whole-body gamma irradiation. Areas of massive apoptosis are indicated by blue arrows.

Drug-mediated suppression of p53 results in the survival of cells that otherwise would be eliminated by p53 and that may increase the risk of new cancer development. In fact, p53-deficient mice are extremely sensitive to radiation-induced tumorigenesis (14). However, in our study, no tumors or any other pathological lesions were found in the group of 30 survivors rescued from lethal gamma irradiation by PFTα, even at 7 months after irradiation. Thus, temporary suppression of p53 appears to differ from p53 deficiency in terms of cancer predisposition.

To monitor PFTα activity at the tissue level, we compared the effect of gamma radiation on DNA synthesis in tissues in PFTα-treated and untreated mice using a 14C-thymidine incorporation assay. 14C labeling of skin, gut, and several other tissues was significantly decreased after gamma irradiation in p53+/+ mice but not p53−/− mice, reflecting the p53 dependence of the effect. The radiation-induced decrease in14C-thymidine incorporation was less pronounced in PFTα-treated mice than in control irradiated animals, presumably reflecting PFTα inhibition of p53 (Fig. 4B). These results suggest that PFTα attenuates the p53-dependent block of DNA replication in rapidly proliferating tissues after whole-body gamma irradiation. Changes in thymidine incorporation correlated with the extent of apoptosis in the gut epithelium of gamma-irradiated mice. The extensive apoptosis observed in the crypts and villi of the small intestine was abrogated in mice treated with PFTα before irradiation (Fig. 4C).

Our results raise the possibility of using PFTα (or other compounds with similar activity) to reduce the side effects of radiation therapy or chemotherapy for human cancers that have lost functional p53. Because the effects of PFTα are p53-dependent, the compound should not affect the sensitivity of such tumors to treatment. In fact, i.p. injection of PFTα did not change the radiation response of p53-deficient tumor xenografts in p53+/+ nude mice (15).

It is likely that suppression of p53-dependent apoptosis has already been successfully and broadly applied to cancer patients in the form of growth factors supplementing chemotherapy (16). The therapeutic effect of such supplements may be associated with their activity as survival factors suppressing p53-dependent apoptosis (17). PFTα can now be used to determine whether there are any other clinical situations in which p53 suppression might be desirable. These include heart and brain ischemia, which both result from local hypoxia, a potent activator of p53 (18). Systematic screening of synthetic and natural compounds may lead to the identification of additional p53 inhibitors that may protect tissue from the consequences of a variety of stresses.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: gudkov{at}uic.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article