p53 Protein Regulates Hsp90 ATPase Activity and Thereby Wnt Signaling by Modulating Aha1 Expression*

Background: The mechanism by which p53, a tumor suppressor gene, regulates Wnt signaling is incompletely understood. Results: p53 modulates Hsp90 ATPase activity via effects on Aha1 leading to changes in the expression of Wnt target genes. Conclusion: p53 regulates Hsp90 ATPase activity and thereby Wnt signaling. Significance: We describe a new mechanism by which p53 affects Wnt signaling. The p53 tumor suppressor gene encodes a homotetrameric transcription factor which is activated in response to a variety of cellular stressors, including DNA damage and oncogene activation. p53 mutations occur in >50% of human cancers. Although p53 has been shown to regulate Wnt signaling, the underlying mechanisms are not well understood. Here we show that silencing p53 in colon cancer cells led to increased expression of Aha1, a co-chaperone of Hsp90. Heat shock factor-1 was important for mediating the changes in Aha1 levels. Increased Aha1 levels were associated with enhanced interactions with Hsp90, resulting in increased Hsp90 ATPase activity. Moreover, increased Hsp90 ATPase activity resulted in increased phosphorylation of Akt and glycogen synthase kinase-3β (GSK3β), leading to enhanced expression of Wnt target genes. Significantly, levels of Aha1, Hsp90 ATPase activity, Akt, and GSK3β phosphorylation and expression of Wnt target genes were increased in the colons of p53-null as compared with p53 wild type mice. Using p53 heterozygous mutant epithelial cells from Li-Fraumeni syndrome patients, we show that a monoallelic mutation of p53 was sufficient to activate the Aha1/Hsp90 ATPase axis leading to stimulation of Wnt signaling and increased expression of Wnt target genes. Pharmacologic intervention with CP-31398, a p53 rescue agent, inhibited recruitment of Aha1 to Hsp90 and suppressed Wnt-mediated gene expression in colon cancer cells. Taken together, this study provides new insights into the mechanism by which p53 regulates Wnt signaling and raises the intriguing possibility that p53 status may affect the efficacy of anticancer therapies targeting Hsp90 ATPase.

The development and progression of colon cancer are part of a multistep process in which growth control is increasingly impaired. Activation of the Wnt pathway plays a major role in colon cancer initiation. In normal cells, the levels of cytoplas-mic ␤-catenin are controlled by a multiprotein destruction complex that targets ␤-catenin for degradation by the proteasome (1,2). The destruction complex consists of Axin, APC, 2 and the glycogen synthase kinase-3␤ (GSK3␤) (1,2). Wnt signaling stimulates the dissociation of the ␤-catenin destruction complex leading to the accumulation and nuclear translocation of ␤-catenin which binds, in turn, to the T cell factor/lymphocyte enhancer binding factor family (TCF/LEF) and regulates Wnt target genes (3). Truncating APC mutations are found in ϳ80% of sporadic colon carcinomas and represent the most common cause for activation of Wnt signaling (4). In addition to mutant APC, a variety of other mechanisms including mutation of the ␤-catenin gene, increased expression of Wnt ligands, and mutational inactivation of the p53 tumor suppressor have been shown to affect Wnt target gene expression (5)(6)(7)(8). Because p53 mutations and activation of Wnt signaling are common in colorectal cancer (4), a detailed understanding of the mechanisms by which p53 modulates Wnt signaling is important.
The p53 tumor suppressor gene encodes for a homotetrameric transcription factor which is activated in response to a variety of cellular stressors, including DNA damage, hypoxia, metabolic stress, and oncogene activation (9 -12). Under these conditions, the p53 protein is stabilized, initiating a transcriptional program that results in DNA repair, cell cycle arrest, senescence, or apoptosis (11). Mutations affecting p53 are present in Ͼ50% of cancers (13). Wild type p53 has been suggested to inhibit Wnt signaling by different mechanisms including the induction of microRNA-34 (6 -8). Moreover, Wnt signaling has been reported to be activated in cells derived from Li-Fraumeni syndrome (LFS) patients who carry germ line, monoallelic p53 mutations (14). In Wnt-1-overexpressing mice, p53 deficiency results in accelerated tumorigenesis relative to Wnt-1 transgenic mice that are wild type for p53 (15). Although the link between p53 and Wnt signaling is established, the interaction between these pathways is incompletely understood.
Here we have investigated the effect of p53 on Wnt signaling in human colorectal cancer cell lines and LFS-derived epithelial cells. To translate these in vitro findings, the effect of p53 on Wnt-signaling including target gene expression was also compared in wild type versus p53-null mice. Using these complementary model systems, we show that p53 modulates Wnt signaling via effects on Hsp90. Specifically, loss of p53 function was associated with increased levels of Aha1, a co-chaperone of Hsp90. The changes in Aha1 levels were mediated by HSF-1. Increased interaction of Aha1 and Hsp90 led to enhanced Hsp90 ATPase activity, which stimulated the Akt/GSK3␤ pathway. This led, in turn, to increased nuclear translocation of ␤-catenin and enhanced Wnt target gene expression. Consistent with these findings, we also show that pharmacologic intervention with CP-31398, a p53 rescue compound (16), inhibited the Aha1/Hsp90 axis and thereby suppressed Wnt signaling. Taken together, this study provides new insights into the mechanism by which p53 regulates Wnt signaling, which may be important for understanding the progression of colon cancer.
Cell Culture-HCT15, DLD-1, and LoVo human colon cancer cell lines were obtained from the American Type Culture Collection (ATCC). These cell lines were maintained according to ATCC instructions. The human colon carcinoma cell line EB-1 was kindly provided by Dr. Arnold J. Levine (Princeton University) (17,18). These cells were maintained in RPMI 1640 medium with 10% FBS and supplemented with 0. 5g/liter geneticin (G418). The HME32, HME50, and IUSM/LFS/ HME cells have been described previously and were provided by Dr. Brittney-Shea Herbert (Indiana University School of Medicine) (14,19).
Immunoprecipitation-This assay was performed with a kit from Upstate Biotechnology according to the manufacturer's instructions. 500 -1000 g of cell lysate or tissue lysate protein was used for immunoprecipitation at room temperature. The immunoprecipitates were then analyzed by Western blotting.
Western Blotting-Cell and tissue lysates were prepared using a lysis buffer (150 mM NaCl, 100 mM Tris, pH 8.0, 1% Tween 20, 50 mm diethyldithiocarbamate, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml trypsin inhibitor, and 10 g/ml leupeptin) followed by sonication to remove particulate material. The protein concentration was determined according to Lowry et al. (20). SDS-PAGE was performed under reducing conditions on 10% polyacrylamide gels as described (21). Resolved proteins were transferred onto nitrocellulose sheets and incubated with the indicated antisera followed by a secondary antibody to horseradish peroxidaseconjugated IgG. The blots were then reacted with the ECL Western blot detection system.
Quantitative Real-time PCR-Total RNA was isolated from colon tissues using the RNeasy mini kit (Qiagen). For tissue analyses, poly(A) RNA was prepared with an Oligotex mRNA mini kit (Qiagen). Poly(A) RNA was reversed-transcribed using murine leukemia virus reverse transcriptase and oligo(dT) 16 primer. The resulting cDNA was then used for amplification. Primers for Aha1 and ␤-actin, an endogenous normalization control, were purchased from Qiagen. Real-time PCR was performed using 2ϫ SYBR Green PCR master mix on a 7500 Realtime PCR system (Applied Biosystems). Relative -fold induction was determined using the ⌬⌬C T (relative quantification) analysis protocol.
Promoter Constructs-PCR was used to generate several truncated forms of the Aha1 promoter. By using genomic clone DNA as a template, PCR was performed at 94°C for 3 min of denaturing, annealing for 1 min at 60°C, and extension for 10 min at 72°C by using Taq polymerase. The primers used to clone Aha1 promoter were: forward, 5Ј-GTGAGGGCCAGAAAAGCA-3Ј; reverse, 5Ј-CTGCAAAGCAGAAACAGAGC-3Ј. PCR products of 5Ј-flanking fragments of Aha1 gene were inserted into the KpnI site of the luciferase basic plasmid vector, pGL2 (Promega). The subcloned PCR products were sequenced by using T7 and SP6 promoter primers to confirm that the products were the authentic promoter fragments. Site-directed mutagenesis was used to mutagenize HSF-1 binding element (HBE mut.) in Aha1 promoter.
Transient Transfections-Cells were grown to 60 -70% confluence in 6-well dishes and then transfected using Lipofectamine 2000 (Invitrogen) for 24 h. Subsequently, the medium was replaced with serum-free medium for another 24 h. Luciferase and ␤-galactosidase were measured in cellular extracts. TOP Flash activity, a measure of TCF/␤-catenin-mediated gene transcription, was determined by the ratio of pTOP-flash/pFOP-flash luciferase activity, each normalized to ␤-galactosidase enzymatic activity levels.
RNA Interference-The cells were seeded in keratinocyte growth medium for 24 h before transfection. Two micrograms of siRNA oligonucleotides was transfected using DharmaFECT 4 transfection reagent according to the manufacturer's instructions.
ChIP Assay-ChIP assay was performed with a kit according to the manufacturer's instructions. Briefly, 4 ϫ 10 6 cells were cross-linked in a 1% formaldehyde solution at 37°C for 10 min. Cells were then lysed and sonicated to generate 200 -1000-bp p53 Regulates Hsp90 ATPase Activity DNA fragments. After centrifugation, the cleared supernatant was incubated with 4 g of the HSF-1 antibody at 4°C overnight. Immune complexes were precipitated, washed, and eluted as recommended. DNA-protein cross-links were reversed by heating at 65°C for 4 h, and the DNA fragments were purified and used as a template for PCR amplification.
Quantitative real-time PCR was carried out. Aha1 promoter oligonucleotide sequences for PCR primers were: forward, 5Ј-GCAGGGAGGTGCTTATTA-3Ј and reverse, 5Ј-TAGAT-GGCCACAAAAACG-3Ј. This primer set encompasses the Aha1 promoter sequence from nucleotide Ϫ282 bp to Ϫ583 bp. PCR was performed at 94°C for 30 s, 62°C for 30 s, and 72°C for Cells also received 0.9 g of siRNA to GFP (control siRNA) or p53. 48 h after transfection, cells were harvested, and luciferase activity was measured. TOP Flash activity was determined by the ratio of pTOP-flash to pFOP-flash luciferase activity, each normalized to ␤-galactosidase enzymatic activity levels. B, D, and F, cells were transfected with 2 g of siRNA (control) to GFP or p53 for 48 h. Following transfection, cells were harvested, and cell lysates were subjected to Western blotting. The blots were probed with antibodies to the indicated proteins. G and H, EB-1 cells were transfected with 1.8 g of p53 luciferase construct (G) or 0.9 g each of TOP Flash and FOP Flash constructs and 0.2 g of pSV␤gal for 24 h. 24 h later, cells were treated with indicated concentrations of ZnCl 2 for 12 h, and then cells were harvested, and luciferase activity was measured. Luciferase activity was normalized to ␤-galactosidase activity. I, cells were treated with the indicated concentrations of ZnCl 2 for 12 h. Cell lysates were subjected to Western blotting, and the blots were probed as indicated. 45 s for 35 cycles, and real-time PCR was performed at 95°C for 15 s and 60°C for 60 s for 40 cycles. The PCR product generated from the ChIP template was sequenced, and the identity of the Aha1 promoter was confirmed.
Hsp90 ATPase Activity-The ATPase assay was based on a regenerating coupled enzyme assay as described earlier (22,23). Briefly, Hsp90 was immunoprecipitated from cell lysates, and the pellet was resuspended and used for assay. Reaction was conducted in a 1-ml assay containing 100 mM, Tris-HCl, pH 7.4, 20 mM KCl, 6 mM MgCl 2 , 0.8 M ATP, 0.1 mM NADH, 2 mM phosphoenolpyruvate, 0.2 mg of pyruvate kinase, 0.05 mg of L-lactate dehydrogenase, and Hsp90 immunoprecipitated from cell and tissue lysates. Equal amounts of Hsp90 protein were used in each treatment group. Sufficient NADH was added to give an initial absorbance of 0.3 at 340 nm before addition of Hsp90 and activity was detected as a decrease in absorbance. Hsp90 ATPase activity is expressed as pmol/min per mg of protein.
Animal Model-p53 knock-out mice carrying the Trp53 tm1Tyj allele were maintained on the 129S6 inbred strain background and genotyped by PCR analysis of DNA extracted from tail tip biopsies (24). Experimental p53 Ϫ/Ϫ and sexmatched littermate control p53 ϩ/ϩ mice at 6 -10 weeks of age were euthanized by carbon dioxide asphyxiation, and isolated colon tissue was rinsed once with PBS and then snap frozen for subsequent molecular analysis. All animal use was conducted in accordance with federal and institutional guidelines, under a protocol approved by the Cornell University Institutional Animal Care and Use Committee.

p53 Regulates Hsp90 ATPase Activity
Statistics-Comparisons between groups were made by Student's t test. A difference between groups of p Ͻ 0.05 was considered significant.

RESULTS
p53 status has been identified as a determinant of Wnt signaling, but the mechanisms are incompletely understood (6 -8). To further understand the interaction between p53 and Wnt signaling, both in vitro and animal models were employed.
p53 Status Is a Determinant of Hsp90 ATPase Activity-Hsp90, a chaperone molecule for many client proteins including those with oncogenic potential, is activated in a variety of malignancies (25)(26)(27). Here we determined whether p53 status is a determinant of Hsp90 ATPase activity. Silencing p53 caused a significant increase in Hsp90 ATPase activity in HCT-15, LoVo, and DLD-1 cells (Fig. 3, A-C). This increase was observed in the absence of increased Hsp90 protein levels. In contrast, inducing p53 in EB-1 cells (Fig. 3D) or by treatment with CP-31398 (Fig. 3, E and F) led to a dose-dependent inhibition of Hsp90 ATPase activity. This decrease in Hsp90 ATPase activity occurred in the absence of corresponding changes in Hsp90 protein levels. Taken together, these results indicate that p53 is a determinant of Hsp90 ATPase activity in colon cancer cells.
p53 Modulates Hsp90 ATPase Activity and Wnt Target Gene Expression through Aha1-Previously, Hsp90 ATPase activity was found to regulate Wnt signaling (28). Given this background, we tested the effects of two inhibitors of Hsp90 ATPase. Both 17-AAG (29) and PU-H71 (30), prototypic Hsp90 ATPase inhibitors, caused dose-dependent suppression of Axin-2, c-Myc, and Naked-1 in HCT15, DLD-1, and EB-1 cells (data not shown). Activation of Hsp90 is dependent on the presence of its co-chaperones. Aha1, a co-chaperone of Hsp90, stimulates Hsp90 ATPase activity (31)(32)(33). Here we explored the possibility that Aha1 expression could be affected by p53. First we show that pharmacological rescue of p53 using CP-31398 in HCT-15, LoVo, and DLD-1 cells caused a dose-dependent decrease in Aha1 levels (Fig. 4, A-C). This effect was observed within 4 h of treatment and was associated with a corresponding reduction in Hsp90 ATPase activity (data not shown). The relationship between p53 status and Aha1 levels was further evaluated in EB-1 cells. Here, treatment with ZnCl 2 , which induced p53 (Fig. 1I), caused a decrease in Aha1 expression (Fig. 4D). In contrast, silencing of p53 in HCT-15, LoVo, and DLD-1 cells caused a substantial increase in Aha1 levels (Fig. 4, E-G). Importantly, the effect of p53 on Hsp90 co-chaperones appears to be specific to Aha1 because, in similar experiments, whether through silencing of p53 or its rescue by pharmacologic intervention with CP-31398, levels of other co-chaperones of Hsp90, namely p23, XAP-2, and HOP, were unaffected (Fig. 4, H-M). In addition to HCT-15 cells, levels of co-chaperones other than Aha1 were unaffected in LoVo, DLD-1, and EB-1 cell lines (data not shown).
To test directly the role of Aha1 in regulating Hsp90 ATPase activity, Hsp90 ATPase activity was measured following silencing of Aha1. Notably, silencing of Aha1 led to reduced Hsp90 ATPase activity in HCT-15, EB1 (Fig. 5, A and B), and DLD-1 cells (data not shown). Next, co-immunoprecipitation experiments were conducted to determine whether changes in p53 levels modified the interaction between Hsp90 and Aha1. As shown in Fig. 5, C and D, silencing p53 led to increased interaction between Aha1 and Hsp90. Similar effects were observed in LoVo cells (data not shown). In contrast, treatment with CP-31398 or p53 induction in EB-1cells led to decreased interaction between Aha1 and Hsp90 (Fig. 5, E and F). To evaluate the functional significance of Aha1 in regulating Wnt target gene expression, we investigated the effects of silencing Aha1. As shown in Fig. 5, G and H, silencing of Aha1 led to reduced levels of Axin-2, c-Myc, and Naked-1. These results indicate that p53 modulates the expression of Aha1 and its interaction with Hsp90 which, in turn, affects Hsp90 ATPase activity and thereby the expression of genes regulated by Wnt signaling.
Next we investigated the mechanism by which p53 regulated levels of Aha1. To determine whether p53 regulated Aha1 transcription, transient transfections were carried out. Silencing of p53 was associated with increased Aha1 promoter activity (Fig.  6, A-C). Treatment with ZnCl 2 , which induced p53 (Fig. 1I), caused a corresponding dose-and time-dependent decrease in Aha1 promoter activity (Fig. 6, D and E). Previously, HSF-1 was found to be a potential regulator of Aha1 expression (34). Therefore, we next evaluated whether the p53-dependent effects on Aha1 were mediated by HSF-1. A potential HSF-1binding element (Ϫ331 to Ϫ316) was identified in the human Aha1 promoter (Fig. 6F). Transient transfections were carried out using a series of Aha1 promoter deletion constructs. Notably, silencing p53 stimulated Aha1 promoter activity, an effect that was lost when the HSF-1-binding element was deleted or mutated (Fig. 6, G and H). Consistent with these findings, the suppressive effects of wild type p53 were lost when the HSF-1binding element was deleted or mutated (Fig. 6, I and J). Having established the importance of HSF-1 for regulating Aha1 promoter activity, ChIP assays were performed to evaluate the p53 Regulates Hsp90 ATPase Activity binding of HSF-1 to the Aha1 promoter. As shown in Fig. 7A and B, p53 regulated the binding of HSF-1 to the Aha1 promoter. More specifically, treatment with ZnCl 2 , which induced p53, suppressed the recruitment of HSF-1 to the Aha1 promoter; silencing of p53 increased the recruitment of HSF-1 to the Aha1 promoter. Moreover, silencing of HSF-1 led to reduced Aha1 protein levels (Fig. 7C). Given the important effects of HSF-1 on Aha1 levels, we next determined whether changes in HSF-1 levels modulated Hsp90 ATPase activity. Silencing HSF-1 was associated with reduced Hsp90 ATPase activity in both EB-1 and HCT-15 cells (Fig. 7, D and E). Finally, silencing of HSF-1 led to reduced expression of Wnt target genes (Fig. 7F).
p53 Regulates Akt Phosphorylation-The PI3K/Akt pathway is commonly deregulated in human cancers. Akt is a client protein of Hsp90, which upon activation phosphorylates GSK3␤, stabilizes ␤-catenin, and induces the expression of Wnt target genes (28,(35)(36)(37)(38). Accordingly, we determined the effects of LY294002, a PI3K inhibitor, on the expression of Axin-2, c-Myc, and Naked-1. Treatment with LY294002 caused a dosedependent decrease in levels of Axin-2, c-Myc, and Naked-1 (Fig. 8, A and B). Silencing p53 in these cells induced Akt phosphorylation (data not shown). Induction of p53 in EB-1 cells inhibited Akt phosphorylation (Fig. 8C). Silencing Aha1 inhibited Akt phosphorylation (Fig. 8D). Pharmacologic inhibition of Hsp90 ATPase was also associated with reduced Akt phosphor-ylation (Fig. 8E). Silencing Akt1 in these cells caused a significant decrease in the levels of Wnt target proteins including Axin-2, c-Myc, and Naked-1 (Fig. 8F). Together, these results support the notion that p53 status regulated the Aha1/Hsp90 axis resulting in changes in Akt activity that led, in turn, to altered expression of Wnt target genes.
p53 Status Is a Determinant of GSK3␤ Phosphorylation-GSK3␤ serves a critical role in Wnt signaling during cancer development (1,2). GSK3␤ can be phosphorylated by Akt, leading to activation of TCF/LEF-dependent gene expression. Here we investigated whether p53 status could affect GSK3␤ phosphorylation via the Hsp90/Aha1/Akt axis. Silencing of p53 caused a significant increase in GSK3␤ phosphorylation, whereas induction of wild type p53 in EB-1 cells or pharmacologic rescue of mutant p53 with CP-31398 caused a substantial reduction of GSK3␤ phosphorylation (Fig. 8, G-I). The effects of CP-31398 on the Aha1/Akt/GSK3␤ pathway occurred within four h and persisted for up to 24 h (data not shown). Along similar lines, silencing Aha1 or inhibiting Hsp90 caused a marked reduction in GSK3␤ phosphorylation (Fig. 8, J and K). These results suggest that p53 regulates the Aha1/Hsp90/Akt/GSK3␤ axis and thereby the expression of Wnt target genes (Fig. 8L).
Levels of Aha1, Hsp90 ATPase Activity, and Wnt Target Genes Are Increased in p53-null Mice-To assess the impact of p53 on the Aha1/Hsp90 ATPase axis in vivo, we utilized colon   A and B, ChIP assays were performed. Chromatin fragments were immunoprecipitated with antibodies against HSF-1, and the Aha1 promoter was amplified by real-time PCR. DNA sequencing was carried out, and the PCR products were confirmed to be the Aha1 promoter. This promoter was not detected when normal IgG was used or when antibody was omitted from the immunoprecipitation step. Columns, means (n ϭ 6); error bars, S.D. *, p Ͻ 0.01. C-F, cells were transfected with 2 g of siRNA to GFP (control) or HSF-1 for 48 h. C, cell lysates were subjected to Western blotting, and the blots were probed as indicated. D and E, Hsp90 ATPase activity was measured. Cell lysates were also subjected to Western blotting and the blots probed as indicated (insets). Mean Ϯ S.D. (error bars) are shown, n ϭ 6. *, p Ͻ 0.01 compared with control siRNA-treated cells. F, cell lysates were subjected to Western blotting and the blots probed as indicated. FIGURE 6. HSF-1 is important for p53-mediated regulation of Aha1. A-C, the indicated cells were transfected with 0.9 g of Aha1 promoter luciferase construct and 0.2 g of pSV␤gal. Cells also received 0.9 g of siRNA to GFP (control siRNA) or p53. 48 h after transfection, cells were harvested, and luciferase activity was measured. Luciferase activity was normalized to ␤-galactosidase activity. D and E, EB-1 cells were transfected with 1.8 g of Aha1 promoter luciferase construct and 0.2 g of pSV␤gal. Subsequently, the cells were treated with the indicated concentrations of ZnCl 2 for 12 h (D) or with 100 M ZnCl 2 (E) for the indicated time period. Cells were harvested, and luciferase activity was measured. Luciferase activity was normalized to ␤-galactosidase activity. F, schematic represents Aha1 promoter deletion constructs that were used. HBE, represents the HSF-1-binding element. G and H, HCT-15 cells were transfected with 0.9 g of Aha1 promoter deletion luciferase constructs as indicated and 0.2 g of pSV␤gal. Cells also received 0.9 g of siRNA to GFP (control siRNA) or p53. 48 h after transfection, cells were harvested, and luciferase activity was measured. Luciferase activity was normalized to ␤-galactosidase activity. I and J, EB-1 cells were transfected with the 1.8 g of the indicated Aha1 promoter luciferase constructs and 0.2 g of pSV␤gal. Subsequently, cells were treated with 100 M ZnCl 2 for 12 h, and luciferase activity was measured. Luciferase activity was normalized to ␤-galactosidase activity. A-C, G, and H,  tissue from p53 wild type and p53-null mice. Compared with wild type mice, p53-null mice exhibited both increased Hsp90 ATPase activity and elevated Aha1 levels (Fig. 9, A and B). To evaluate the interaction between Hsp90 and Aha1, immunoprecipitation experiments were carried out. As shown in Fig.  9C, the interaction between Hsp90 and Aha1 was markedly increased in p53-null mice. Moreover, Akt and GSK3␤ phosphorylation were both increased in colon tissues from p53-null versus wild type mice (Fig. 9, D and E). Consistent with these findings, we found increased expression of Wnt target genes including Axin-2, c-Myc, and Naked-1 in p53-null mice (Fig. 9, F-H).
Wnt Signaling Is Increased in Cells Derived from Li-Fraumeni Syndrome Patients-LFS patients carry a germ line mutation in p53 and are prone to develop a variety of cancers (14,39,40). Accordingly, we compared the Aha1/Hsp90/Wnt signaling axis in epithelial cells from LFS p53 mutation carriers (HME50 and IUSM/LFS/HME) versus epithelial cells that were wild type for p53 (HME32). Increased Hsp90 ATPase activity, Aha1 levels, interaction between Aha1 and Hsp90, along with elevated Akt and GSK3␤ phosphorylation were detected in LFS versus normal epithelial cells (Fig. 10, A-E). A significant increase in ␤-catenin along with increased levels of Axin-2, c-Myc, and Naked-1 was found in LFS versus normal epithelial cells (Fig. 10,  F and G). These changes in LFS-derived, p53 heterozygous epithelial cells appear to recapitulate alterations seen in colon tumor cells.

DISCUSSION
In the present study, we have confirmed previous results suggesting that p53 affects Wnt signaling (6 -8). Notably, our data demonstrate for the first time that Hsp90 plays a significant role in mediating this effect. Several lines of evidence support this point. Levels of Hsp90 ATPase activity were increased in LFSderived p53 heterozygous versus normal epithelial cells. Silencing of p53 led to increased Hsp90 ATPase activity in multiple colorectal cancer cell lines. Moreover, expression of wild type p53 in EB-1 cells, a p53-null cell line, led to gene dose-dependent suppression of Hsp90 ATPase activity. Similarly, use of CP-31398, a p53 rescue compound, caused dose-dependent induction of p53 and a reciprocal decrease in Hsp90 ATPase activity. The significance of Hsp90 ATPase activity in regulating Wnt signaling was tested. We showed that 17-AAG and PU-H71, prototypic Hsp90 ATPase inhibitors, down-regulated the expression of Wnt target genes including Axin-2, c-Myc, and Naked-1 in several cell lines.
Because p53-mediated changes in Hsp90 ATPase activity occurred in the absence of altered levels of Hsp90 protein, experiments were performed to elucidate the underlying mechanism. Levels of Aha1, a known co-chaperone of Hsp90 (31)(32)(33), were affected by p53 status. By contrast, levels of other Hsp90 co-chaperones including p23, XAP-2, and HOP were unaffected by changes in p53 status (41,42). p53 deficiency was associated with elevated Aha1 levels whereas treatment with CP-31398 or expressing wild type p53 in EB-1 cells suppressed levels of Aha1. These changes in Aha1 protein levels were associated with corresponding alterations in the interactions between Aha1 and Hsp90 and Hsp90 ATPase activity. To further evaluate the functional significance of Aha1 for mediating p53-dependent changes in Hsp90 ATPase activity, Aha1 was silenced. Silencing of Aha1 inhibited Hsp90 ATPase activity and the expression of Wnt target genes. The mechanism by which p53 regulates Aha1 levels was previously unknown. Aha1 can be regulated by HSF-1 (34). We identified an HSF-1-binding element in the 5Ј-UTR of Aha1. Promoter analyses indicated that p53 regulated Aha1 expression via the HSF-1 element. For example, silencing p53 induced Aha1 promoter activity; this effect was lost when cells were transfected with an Aha1 promoter construct that did not contain the HSF-1 binding site or when the HSF-1 binding site was mutagenized. Additionally, ChIP assays showed increased recruitment of HSF-1 to the Aha1 promoter when p53 was silenced and decreased recruitment when wild type p53 was overexpressed. Moreover, silencing HSF-1 led to reduced Aha1 levels, decreased Hsp90 ATPase, and suppression of Wnt target gene expression. Taken p53 Regulates Hsp90 ATPase Activity together, these data suggest that HSF-1 is important for p53mediated regulation of Wnt signaling. Of potential relevance, silencing HSF-1 has been reported to sensitize cancer cells to Hsp90 inhibition (43).
A link between Hsp90 ATPase activity and Wnt signaling has been reported (28). Akt is a known client protein of Hsp90 (35,36). Inhibition of the PI3K/Akt signaling pathway can block nuclear localization of ␤-catenin (28,37,38). Here we demonstrate that Hsp90 ATPase activity regulates the phosphorylation of Akt and thereby Wnt target gene expression. Inhibition of Hsp90 ATPase led to reduced levels of phospho-Akt. Consistent with these findings, silencing of Aha1 also reduced Akt activity. Moreover, treatment with LY294002, a PI3K inhibitor or silencing Akt1, inhibited the expression of Axin-2, c-Myc, and Naked-1. Because of the link between p53 and Hsp90 ATPase activity, we also confirmed that Akt activity was p53dependent. In support of this possibility, overexpressing wild type p53 in colon cancer cells led to decreased Akt phosphorylation. Levels of phospho-Akt were also increased in cells from patients with LFS. Taken together, these data show that increased Hsp90 ATPase activity, associated with either muta-tion or loss of p53, led to enhanced Akt activity and increased Wnt target gene expression. Akt can phosphorylate GSK3␤, resulting in the accumulation of ␤-catenin (28,37,38). Several experiments were performed to evaluate whether this mechanism was operative in a p53-dependent manner. Modulating p53 levels led to changes in GSK3␤ phosphorylation. Similarly, enhanced GSK3␤ phosphorylation and increased ␤-catenin levels were observed in LFS cells. Silencing of Aha1 or inhibiting Hsp90 ATPase activity also suppressed GSK3␤ phosphorylation. Collectively, these in vitro findings suggest that p53 regulates the Aha1/Hsp90 ATPase axis leading to changes in Akt/GSK3␤ signaling thereby modulating Wnt target gene expression. Recently, microRNA-34 was found to be important for p53-mediated regulation of Wnt signaling (8). Although Aha1 is not a direct target of microRNA-34, its interacting partner GTP cyclohydrolase-1 is regulated by it (44,45). Possibly, microRNA-34 will modulate Hsp90 ATPase activity via effects on GTP cyclohydrolase-1. Future studies are needed to determine whether microRNA-34 regulates HSF-1 or Aha1 levels and thereby Hsp90 ATPase activity. . Wnt signaling is activated in p53-null mice. Colon tissue from p53 wild type and p53-null mice was used. Poly(A) RNA was isolated from total RNA extracted from colon tissues. Additionally, tissues were homogenized, and protein lysates were prepared. A, tissue lysates were used to measure Hsp90 ATPase activity using a protocol described under "Experimental Procedures." Mean Ϯ S.D. (error bars) are shown, n ϭ 6. *, p Ͻ 0.01 compared with p53 ϩ/ϩ mice. Lysates were also subjected to immunoprecipitation with Hsp90 or ␤-actin antibodies; Western blotting was carried out, and the blot was probed for Hsp90 and ␤-actin as indicated. B, relative expression of Aha1 was quantified by real-time PCR. Values were normalized to levels of ␤-actin. Mean Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.01 compared with p53 ϩ/ϩ mice. Tissue lysates were also immunoprecipitated with Aha1 or ␤-actin antibody; Western blotting was carried out, and the blots were probed for Aha1 and ␤-actin as indicated. C, tissue lysates were immunoprecipitated (IP) with Hsp90 antibody, and the immunoprecipitates were subjected to Western blotting (WB) for Aha1 and Hsp90 as indicated. D and E, tissue lysates were immunoprecipitated with antibodies to Akt (D) or GSK3␤ (E), and the blots were probed for pAkt and Akt (D) or pGSK3␤ and GSK3␤ (E) as indicated. F-H, tissue lysates were immunoprecipitated with Axin-2, c-Myc, Naked-1, or ␤-actin antibodies, and the blots were probed with the same antibodies.
In an attempt to translate the in vitro findings described above, the Aha1/Hsp90/Akt/GSK3␤ axis was evaluated in colon tissues from wild type versus p53-null mice. Consistent with the in vitro results, increased Aha1 levels, Hsp90 ATPase activity, Akt and GSK3␤ phosphorylation and Wnt target expression was observed in colons of p53-null compared with p53 wild type mice. Remarkably, immunoprecipitation experiments indicated an increased interaction between Hsp90 and Aha1 in the colons of p53-null mice. These results suggest that colons of p53-null mice have a heightened Hsp90 ATPase activity and increased Wnt signaling. Previously, the p53 rescue compound CP-31398 was found to reduce intestinal tumorigenesis in APC min/ϩ mice (46). Additionally, p53 deficiency has been shown to accelerate tumorigenesis in Wnt-1 transgenic mice (15). Our observation that p53 activation suppresses Wnt signaling may help to explain these prior observations. Whether inhibitors of Hsp90 ATPase will have a similar protective effect should be tested. LFS patients are at increased risk for a number of tumor types including colon cancer (47,48). Our observation that cells from LFS patients had elevated Aha1 levels, Hsp90 ATPase activity, Akt and GSK3␤ phosphorylation and increased Wnt target gene expression may further help to explain the link between p53 and increased risk of tumor formation. Possibly, agents that disrupt this activated signaling axis will possess chemopreventive properties. Another potential implication of our findings relates to cancer therapy. If p53 status is found to be a determinant of Hsp90 ATPase activity in tumors, this could contribute to chemoresistance or provide a strategy for identifying patients who are most likely to benefit from inhibitors of Aha1 or Hsp90 ATPase activity (34, 49 -51).