p53 Inhibitor Pifithrin α Can Suppress Heat Shock and Glucocorticoid Signaling Pathways

Pifithrin α (PFTα) is a chemical compound isolated for its ability to suppress p53-mediated transactivation. It can protect cells from p53-mediated apoptosis induced by various stimuli and reduce sensitivity of mice to gamma radiation. Identification of molecular targets of PFTα is likely to provide new insights into mechanisms of regulation of p53 pathway and is important for predicting potential risks associated with administration of PFTα-like p53 inhibitors in vivo. We found that PFTα, in addition to p53, can suppress heat shock and glucocorticoid receptor signaling but has no effect on nuclear factor-κB signaling. PFTα reduces activation of heat shock transcription factor (HSF1) and increases cell sensitivity to heat. Moreover, it reduces activation of glucocorticoid receptor and rescues mouse thymocytesin vitro and in vivo from apoptotic death after dexamethasone treatment. PFTα affected both signaling pathways in a p53-independent manner. These observations suggest that PFTα targets some unknown factor that is common for three major signal transduction pathways.


Pifithrin ␣ (PFT␣) is a chemical compound isolated for its ability to suppress p53-mediated transactivation. It can protect cells from p53-mediated apoptosis induced by various stimuli and reduce sensitivity of mice to gamma radiation. Identification of molecular targets of PFT␣ is likely to provide new insights into mechanisms of regulation of p53 pathway and is important for predicting potential risks associated with administration of PFT␣-like p53 inhibitors in vivo.
We found that PFT␣, in addition to p53, can suppress heat shock and glucocorticoid receptor signaling but has no effect on nuclear factor-B signaling. PFT␣ reduces activation of heat shock transcription factor (HSF1) and increases cell sensitivity to heat. Moreover, it reduces activation of glucocorticoid receptor and rescues mouse thymocytes in vitro and in vivo from apoptotic death after dexamethasone treatment. PFT␣ affected both signaling pathways in a p53-independent manner. These observations suggest that PFT␣ targets some unknown factor that is common for three major signal transduction pathways.
Based on the analysis of p53-dependent effects caused by ionizing radiation and chemotherapeutic drugs in mice, p53mediated apoptosis was defined as a determinant of organism sensitivity to systemic genotoxic stress associated with cancer treatment (1). Temporary reversible pharmacological suppression of p53 was suggested as an approach to reduce cancer treatment side effects. This hypothesis was supported by isolation of a small molecule inhibitor of p53, pifithrin ␣ (PFT␣) 1 that was capable of rescuing mice from lethal genotoxic stress caused by gamma radiation (2). Furthermore, inhibition of p53 was suggested as a therapeutic approach to treatment of other pathological conditions associated with p53 activation (3), some of which have already been experimentally confirmed. Thus, PFT␣ was shown to protect neurons from death induced by DNA-damaging agents, hypoxia and dopamine (4,5): it had therapeutic effects in animal models of Parkinson disease (6) and acute renal failure (7). In all these works, biological effects of PFT␣ were attributed to its anti-p53 function, although not in all of them has this conclusion been confirmed by genetic approaches. Accurate interpretation of biological effects of PFT␣ requires identification of its molecular target(s) and determination of molecular mechanisms of its activity.
PFT␣ was isolated by screening of chemical library in a cell-based readout system for its ability to reduce p53-dependent transactivation (2). This biological effect could be reached by affecting p53 pathway at numerous points and therefore PFT␣ could act by targeting one of numerous factors cooperating with p53 function. Biological effects of PFT␣ on p53 pathway suggested that it acted by interfering with nuclear accumulation of p53 (2). Many transcription factors involved in other signal transduction pathways have the same principles of regulation as p53: after activation in cytoplasm they are translocated to the nucleus, followed by modulation of transcription of the target genes. We were, therefore, interested to test whether PFT␣ would have an effect on other signal transduction pathways besides p53. We found that, in fact, PFT␣ can also interfere with heat shock (HS) and glucocorticoid receptor (GR) signaling but shows no effect on the activity of NF-B. This finding indicates that PFT␣ is not solely specific to p53 and presumably targets some unknown cellular component that is common for three major signal transduction pathways.
Cell Lines and Animals-Mouse fibroblast cell line ConA carries the wild type p53 gene and the bacterial lacZ reporter gene under the control of a p53-responsive promoter (2). Two isogenic human colon cancer cell lines HCT116 p53 (wt-p53) and its p53-deficient derivative, developed from the parental cell line by targeted homologous recombination (8), were provided by I. Roninson (University of Illinois at Chicago). HeLa cells and human prostate cancer cell line PC3 (p53-deficient) were purchased from ATCC. Short term cultures of primary thymocytes were prepared from the thymus of 4-week-old C57BL/6 mice (wild type and p53-deficient), which were purchased from Jackson Laboratory (Bar Harbor, ME).
Cell Viability Assay-At the end of cell treatments, 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. Cell viability in suspension of short term culture of primary thymocytes was determined by their staining with 0.1% of methyl blue and microscopic counting of blue (dead) cells.
Gel Shift Assay-Gel shift assay was performed as described earlier (9). Nuclear and total cellular extracts were prepared from untreated or * This work was supported by National Institutes of Health Grant CA17579 and by a grant from Quark Biotech, Inc. (both to A. V. G.). 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.
CAT Assay-CAT assay was done as previously described (11). ConA and HeLa cells were transfected with plasmids, containing CAT gene under the control of a minimal thymidine kinase promoter alone (Promega) or combined with HSF1-binding or GRE-binding sequences from HSP70 (12) and LTR MMTV (13) promoters, respectively. Cells were treated with HS (42°C, 30 min) or Dex (0.1 M, 1 h) with and without PFT␣ (15 M).
Western Blot Analysis-Western blot analysis was done as described previously (12). Wild type and p53-deficient HCT116 cells were incubated 15 min at 43°C in the presence or in the absence of PFT␣ (15 M) and total cell lysates were prepared 3 and 6 h later. HSP70 was detected using goat polyclonal antibodies K-20 (Santa Cruz Biotechnology).

RESULTS AND DISCUSSION
PFT␣ and Heat Shock Response-HS induces expression of a large family of heat shock proteins (HSP70, HSP90, HSP43, HSP27, etc.), many of which function as either molecular chaperones or proteases that assist the cell in recovery either by repairing damaged proteins (protein refolding) or by degrading them (14). Hsp gene promoters contain heat shock elements responsible for binding with HSFs mediating HS-inducible transcription, among which HSF1 seems to play a major role in Hsp gene regulation (15). Under normal conditions, HSF1 exists as an inactive monomer bound to multichaperone complexes (HSP90, HSP70, and others) (16) but is readily activated after HS by forming active trimers that are translocated into the nucleus where they bind heat HS-responsive elements in cellular DNA and stimulate HS genes transcription (14). Thus, there are obvious similarities in regulation of HSF1 and p53 response pathways that justified testing effects of PFT␣ on HS signaling.
Activation of HSF1-containing transcription complex was determined by gel-shift assay using 32 P-labeled heat shock element-derived oligonucleotides and total or nuclear extracts of ConA cells growing under normal conditions or subjected to HS (42°C, 20 min) in the presence and in the absence of 15 M of PFT␣. The intensity of HSF1-specific band in the lysates from HS-treated cells was substantially decreased if the cells were incubated with PFT␣ during treatment (Fig. 1a).
The results of gel-shift experiments were confirmed in a functional transcription assay using HSF1-responsive construct with CAT reporter. PFT␣ (15 M) caused a 2-fold reduction in CAT activity in ConA cells under conditions of HS (Fig.  1b), suggesting that it might have similar effect on the expression of endogenous Hsp genes. In fact, application of PFT␣ was accompanied by an increased susceptibility of ConA cells to heat shock determined by colony assay (data not shown) and reduction in accumulation of HSP70 (Fig. 1d).
To determine whether the effect of PFT␣ on cell sensitivity to HS is p53-dependent or p53-independent, we compared the effect of the compound on HS sensitivity of two isogenic variants of human colon cancer cell line HCT116 differing in their p53 status (8). Presence of PFT␣ during HS treatment (45°C, 30 min) significantly increased HS-induced cytotoxicity in both p53-wt and p53-deficient cell lines in a dose-dependent manner; similar doses of PFT␣ had no toxic effect on either cell line under normal growth conditions (Fig. 1c). Consistently, PFT␣ caused a reduction in HS-induced accumulation of HSP70 protein in both p53 wild type and p53-deficient HCT116 cells (Fig.  1d). In addition to ConA and HCT116, similar results were obtained with p53-deficient human prostate cancer cell line PC3 (data not shown). These observations indicate that PFT␣ has a p53-independent mechanism of activity directed against HSF1-mediated HS response.
PFT␣ and GR Signaling-Glucocorticoid hormones are involved in regulation of many important functions in the organism, including development and function of the immune system. Signaling is mediated by interaction of glucocorticoids with their receptor (GR), ligand-dependent transcription factor, that is, as p53 and HSF, regulated at the level of nuclear transport (17,18). In the absence of ligands, GR resides in the cytoplasm in a monomeric form bound to cytoplasmic chaperones, such as HSP70 and HSP90. Binding of the ligand typically results in a conformational change in GR, dimerization and translocation to the nucleus, where GR homodimer binds to a DNA motif termed a GRE and transactivates glucocorticoid-responsive genes. In thymocytes, this results in activation FIG. 1. PFT␣ inhibits HS response. a, PFT␣ inhibits DNA binding activity of HSF1. Results of gel-shift assay using nuclear and total cellular extracts from either untreated or HS-treated (42°C, 30 min) ConA cells with and without PFT␣ (10 M) are shown. Labeled doublestranded oligonucleotides, corresponding to the sequences of the HSF1binding site in HSP70 promoter, were used. Untreated cells are marked as "u/t." b, PFT␣ suppresses HSF1-mediated transactivation of reporter construct. ConA cells were transfected with the plasmids, containing CAT gene under minimal thymidine kinase promoter alone (marked "min pr") or combined with the HSF1-binding sequence from HSP70 promoter (marked "HS pr"). 24 h after transfection cells were incubated 30 min at 42°C, with or without PFT␣ (15 M) followed by preparation of lysates and estimation of CAT activity. c, PFT␣ increases cell sensitivity to HS in a p53-independent manner. Wild type p53 and p53deficient HCT116 cells preincubated with the indicated concentrations of PFT␣ were subjected to HS (45°C, 25 min). Cell numbers were estimated 48 h after treatment using methylene blue assay. d, PFT␣ (15 M) reduces induction of HSP70 by HS in a p53-independent manner. Wild type p53 and p53-deficient HCT116 cells were incubated 15 min at 43°C, and the amounts of HSP70 in cell lysates prepared 3 h after HS were estimated by western immunoblotting. A similar effect was observed in mouse NMuMG cells shown in the lower panel (lysed 6 h after HS). Western blots with HCT116 proteins were reprobed with antiactin antibody (loading control); a nonspecific band is shown as a loading control (marked "LC") for NMuMG membrane.

PFT␣ Inhibits Heat Shock and Glucocorticoid Pathways 15466
of proapoptotic genes and subsequent death that is consistent with anti-inflammatory role of glucocorticoids (19).
To analyze whether PFT␣ has an effect on GR signaling we used the same strategy as described above for HS signaling. GR activation was tested in HeLa cells treated for 4 h with a range of concentrations of synthetic glucocorticoid Dex using gel-shift assay with oligonucleotide specific for GRE. As shown in Fig.  2a, the Dex-induced DNA binding activity of GR was significantly inhibited by 15 M PFT␣.
Consistently, presence of PFT␣ reduced CAT activity in the lysates of ConA and HeLa cell transfected with the glucocorticoid-responsive construct with CAT reporter and treated with Dex (Fig. 2b).
To test whether biochemical indications of inhibition of GR activity by PFT␣ reflect alterations of physiological function of GR, we analyzed cell response to glucocorticoid in the presence and in the absence of PFT␣ in vitro and in vivo. Primary thymocytes are known to respond to glucocorticoid treatment by rapid p53-independent apoptosis (20). We tested the effect of PFT␣ on apoptosis induced in short term cultures of thymocytes by Dex treatment. To distinguish between p53-dependent and -independent effects, we compared thymocytes from p53 wild type and p53-deficient mice. As it is shown in Fig. 2c, PFT␣ protected both p53 wild type and (to a lesser extent) p53-deficient thymocytes from Dex-induced death, indicating that the protective effect of PFT␣ against glucocorticoid is p53-independent (Fig. 2c).
PFT␣ also had a prominent protective effect in vivo, inhibiting Dex-induced degeneration of the thymus. Subcutaneous injection of Dex (4.5 mg/kg) resulted in almost 2-fold decrease in the size of mouse thymus as early as 24 h after hormone administration. This effect was almost completely reverted by PFT␣ that was intraperitoneally injected three times, 0, 2 and 6 h (each dose was 3.6 g/kg) after Dex (Fig. 2d).
PFT␣ Has No Effect on the Activity of NF-B-Transcription factor NF-〉 is a key component of a major anti-apoptotic signal transduction pathway induced by a variety of physiological stimuli and stresses. It plays an important role in regulating inflammation by determining cell response to TNF (21) and mediating activation of numerous inflammatory cytokines. It also determines anti-apoptotic activity of AKT signaling, a major survival pathway that connects cell viability with physiological conditions (i.e. availability of growth factors) (22). It is also regulated at the level of nuclear transport; however, the exact mechanism of this regulation is different from p53, HSF1, or GR. Under normal conditions, it resides in the cytoplasm as an inactive complex bound to inhibitory protein factor I-B, while p53, GR, and HSF1 are coupled with HS chaperone complexes. Activation of NF-B is followed by phosphorylation and degradation of I-B that leads to a release of NF-B and its translocation to the nucleus where it binds to specific binding sites causing transcriptional activation of a set of NF-B-responsive genes that determine physiological cell response (23).
In our study of PFT␣ effect on NF-B activity we followed the same steps as with other signal transduction pathways, starting from gel shift assays. Nuclear extracts were isolated from ConA and HeLa cells, untreated and treated with TNF (1.5 ng/ml) in the presence or in the absence of 15 M of PFT␣. Results of gel-shift analysis using lysates of both cell lines and labeled oligonucleotide, corresponding to NF-B-binding region of the mouse B cell light chain enhancer, showed no affect PFT␣ on the induction of NF-B (Fig. 3a, shown for ConA cells). Similarly, PFT␣ had no effect of NF-B transactivation as judged by reporter transfection assays (data not shown).

PFT␣ Inhibits Heat Shock and Glucocorticoid Pathways 15467
Inhibition of transactivation ability of p53, HSF1, and GR by PFT␣ was accompanied by suppression of their biological functions resulting in suppression of apoptosis caused by genotoxic stress (p53), sensitization to HS (HSF1), and resistance to Dex-mediate cell killing (GR). We tested whether PFT␣ would affect the ability of activated NF-B to protect cells from TNFinduced apoptosis (23). CHI, an inhibitor of translation, suppresses induction of NFB and makes cells highly sensitive to TNF-mediated apoptosis. If PFT␣ would suppress NF-B activation (as CHI does), its application should sensitize cells to TNF. Analysis of three cell systems (ConA, NIH 3T3 cells, and short term primary culture of thymocytes), all known to be TNF-resistant due to activation of NF-B, did not show any effect of PFT␣ on their sensitivity to TNF, while treatment with CHI had strong sensitizing effect presumably by blocking NF-B activation (23) (Fig. 3b). These observations are well in line with lack of PFT␣ effect on activity of the activation of NF-B transcription factor found in biochemical assays.
The dramatic differences between the effects of PFT␣ on p53, HSF1, and GR on one hand and NF-B, on the other, suggest the existence of a common regulatory component(s) in those pathways that are affected by the compound, which is not part of NF-B signaling. Moreover, this putative PFT␣ target is likely to act by affecting nuclear accumulation of sensitive transcription factors as it was previously shown for p53 (2). Although at this stage it is impossible to precisely define the molecular target of PFT␣, we can base our speculations on the known properties of the studied pathways focusing on what differs PFT␣-sensitive pathways from NF-B signaling. Cellular factors belonging to this category are HSP complexes that participate in holding inactive HSF1, GR, and p53 proteins in the cytoplasm but are not likely to be involved in regulation of NF-B that couples instead with its "own" specific inhibitor I-B. Many properties of HSPs and HSP inhibitors (such as quercetin, a flavanoid that shares structural similarity with PFT␣ (24)) are consistent with the potential involvement of HSPs in PFT␣ activity. HSPs are common participants of different apoptotic pathways, and they are induced by a variety of stress agents, including UV and gamma radiation, HS, glucocorticoids, cytotoxic drugs, etc. (17,25). Moreover, cells could be protected from gamma or UV radiation induced apoptosis by exposure to HS or overexpression of HSP70 (17,26), while quercetin, an HSP inhibitor, is known to enhance apoptosis in a variety of systems (24,27). Thus, HSPs are obvious candidate targets of PFT␣, and this hypothesis remains to be experimentally tested.