Serine/Threonine Protein Phosphatase Type 5 Acts Upstream of p53 to Regulate the Induction of p21WAF1/Cip1 and Mediate Growth Arrest*

Understanding how alterations in growth control pathways are translated into changes in the cell cycle regulatory machinery is a major challenge for understanding the development of human cancers. The ability of both tumor suppressor proteins, p53 and BRCA1, to induce the expression of p21WAF1/Cip1 in combination with the inhibitory activity of p21WAF1/Cip1 against cyclin-dependent kinases suggests that the regulation of p21WAF1/Cip1 expression is an important aspect of mammalian cell cycle growth control. To elucidate the role of serine/threonine protein phosphatase type 5 (PP5) in processes regulating cell cycle progression, we developed antisense oligodeoxynucleotides targeted against PP5 (e.g. ISIS 15534) that specifically inhibit PP5 gene expression. Employing ISIS 15534, we demonstrate that the specific inhibition of PP5 gene expression has a marked antiproliferative effect on cells, characterized by induction of p21WAF1/Cip1 and the subsequent arrest of cell growth. Investigations into the mechanisms leading to growth arrest reveal that, in the absence of PP5, the expression of p21WAF1/Cip1 is induced in p53-competent A549 cells but not in p53 protein-deficient T-24 cells. Employing a stable cell line derived from p53-deficient human fibroblast that contains tetracycline-regulated transactivator and operator plasmids to control the expression of wild-type p53 (TR9-7 cells), we then show that the induction of p21WAF1/Cip1, which occurs in response to the inhibition of PP5 expression, requires the p53 protein. Additional studies indicate that PP5 acts upstream of p53, influencing both the phosphorylation state and the ability of p53 to bind DNA, without causing an increase in p53 gene transcription. Together these studies suggest that PP5 is a regulatory component of a signaling pathway that affords replicating cells G1 checkpoint growth control and that it is the regulation of PP5 that, in turn, controls p53-mediated expression of p21WAF1/Cip1 and growth arrest in this pathway. In addition, since the inhibition of PP5 gene expression has marked antiproliferative activity and the overexpression of p21WAF1/Cip1 blocks the growth of tumor cells, these studies suggest that compounds that inhibit of PP5 gene expression may be useful in the treatment of human cancers.

The reversible phosphorylation of serine and threonine residues on proteins is a mechanism utilized by eukaryotic cells to regulate the biological activity of many enzymes, and phosphorylation influences cellular events as diverse as metabolism, signal transduction, and cell cycle progression. Although the kinases involved in these processes are well recognized, the critical contribution of serine/threonine protein phosphatases (PPases) 1 has only been appreciated more recently (1,2). Traditionally, mammalian PPases have been grouped into four subtypes (PP1, PP2A, PP2B, and PP2C) based on 1) their biochemical characteristics, 2) their sensitivities to specific inhibitors, and 3) a limited amount of substrate specificity that can be demonstrated in vitro (3). Emerging structural data, however, indicate that PP1, PP2A, and PP2B are related, while PP2C is structurally distinct. In addition, three isoforms of PP1, which demonstrate Ͼ90% identity (4 -6), two isoforms of PP2A with Ͼ97% identity (7,8), three isoforms of PP2B with Ͼ80% identity (9 -11), and four structurally related phosphatases designated PP4 (12), PP5 (13,14), PP6 (14,15), and PP7 (16) have been identified. Based on a comparison of their sequences, these mammalian PPases have been placed into five distinct subfamilies termed PP1, PP2A (which includes PP4 and PP6), PP2B, PP5, and PP7 (1,16).
Much of the knowledge concerning the cellular roles of the individual PPases has been derived from studies using what has become a fairly large number of naturally produced toxins that inhibit the activity of certain serine/threonine protein phosphatases. To date, the known PPase inhibitors include okadaic acid, microcystin-LR, nodularin, cantharidin, calyculin A, tautomycin, and fostriecin (17)(18)(19)(20)(21)(22). All of these compounds demonstrate inhibitory activity against PP2A and PP1. However, okadaic acid has been utilized the most extensively due to its ability to penetrate cell membranes and its ability to selectively inhibit the activity of PP2A under appropriate experimental conditions (17,23). With the exception of fostriecin, which demonstrates even more selectivity against PP2A than okadaic acid (22), the other mentioned PPase inhibitors affect both PP1 and PP2A with relatively similar potencies (17)(18)(19)(20)(21)(22)(23). In contrast, PP2B, PP7, and PP2C are not affected or resistant to inhibition by these compounds (16,23). Studies with recombinant PP4 and PP5 have complicated the interpretation of data obtained using okadaic acid, because these studies indicate that both PP4 and PP5 are also sensitive to inhibition by okadaic acid and microcystin-LR (12,13). To our knowledge, the sensitivity of PP6 to okadaic acid has not yet been determined.
Because many signaling cascades leading to cell growth are stimulated by phosphorylation-mediated events, it has been assumed that PPases are negative regulators of cellular growth. Reports indicating that okadaic acid (at micromolar concentrations (1-125 M)) has tumor-promoting activity in the mouse skin model have furthered the speculation that okadaicsensitive PPases participate in the regulation of cell growth and suggest that certain PPases may function as tumor suppressor proteins (Refs. 24 and 25; for a review, see Ref. 29). However, fostriecin, which is currently under evaluation in clinical trials to determine its potential use as an antitumor drug in humans, has marked antitumor activity both in vitro and in vivo (26). Similarly, cantharidin and structural derivatives of cantharidin have been reported to have antitumor activity both in vivo and in vitro (27,28). In addition, an increasing number of studies suggest that, when administered at (1-20 nM) nanomolar concentrations, okadaic acid can inhibit tumor cell growth and neoplastic transformation (for a review, see Ref. 29). One explanation for these apparently conflicting findings is that the ability of a particular PPase inhibitor to act as either a tumor-promoting agent or an antitumor drug is determined by the subset of PPases that are affected by the inhibitor under the conditions utilized in a particular study. Alternatively, there may be a single PPase (or a subset of PPases) that is sensitive to fostriecin and cantharidin yet resistant to inhibition by okadaic acid or vice versa.
The presence of multiple tetratricopeptide repeat (TPR) sequences in PP5 implies that PP5 interacts with other TPRcontaining proteins, many of which are involved in the regulation of cell cycle progression (1,14,30). Therefore, we were interested in determining if PP5 participates in the regulation of signal transduction pathways that influence cell proliferation. With the realization that, in addition to PP1 and PP2A, both PP4 and PP5 are sensitive to inhibition by okadaic acid, it became apparent that deciphering the contributions of PP5 to cell cycle progression would require the development of a more specific inhibitor. Therefore, in the present study we developed antisense oligodeoxynucleotide targeting human PP5 to specifically inhibit PP5 gene expression in human cells. Using these antisense oligonucleotides, we demonstrate that the expression of PP5 is necessary for A549 cell proliferation and that the inhibition of PP5 expression results in the hyperphosphorylation of p53 and the induction of p21 WAF1/Cip1 . Together, these studies suggest that PP5 acts upstream of p53 and participates in a signaling pathway that regulates p21 WAF1/Cip1 -mediated G 1 growth arrest.

MATERIALS AND METHODS
Cell Lines-Human A549 lung carcinoma cells and T-24 bladder carcinoma cells were obtained from the American Type Tissue Collection, maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, and routinely passed when 90 -95% confluent. TR9-7 cells, which were derived from the p53 Ϫ/Ϫ human fibroblasts (MDAH041) and contain tetracycline-regulated transactivator and operator plasmids to control the expression of wild-type p53 (31), were kindly provided by Drs. G. Stark and M. Agarwal. TR9-7 cells were maintained in DMEM containing 10% fetal calf serum and the indicated amount of tetracycline.
Assay for Oligonucleotide Inhibition of PP5 Expression-A549 cells were plated in 60-mm dishes and cultured in DMEM containing 10% fetal calf serum. When the cells were about 70% confluent, they were treated with oligonucleotides as described previously (32,33). Briefly, cells were washed with DMEM. A solution (1 ml) of DMEM containing the oligonucleotides at the indicated concentration and 15 g/ml DOTMA/DOPE (Lipofectin®; Life Technologies, Inc.) was then added. After incubating the cells at 37°C for 4 h, the cells were washed and cultured in fresh DMEM containing 10% fetal calf serum for 17 h. The cells were then harvested, and total RNA was extracted with TRIzol Reagent (Life Technologies) according to the methods of the manufacturer. Total RNA (20 g) was fractionated on 1% agarose gels containing formaldehyde and transferred to DURLON-UV (Stratagene) nylon membranes. Following UV cross-linking, the filters were hybridized with a 32 P-labeled probe for human PP5. The human PP5 cDNA probe was generated from the full-length coding region of PP5 and 32 P-labeled with DECAprime® DNA labeling kit (Ambion) according to the manufacturer's protocol. Hybridization was performed in the presence of 50% formamide at 42°C for 16 h. Following hybridization, the membrane was subjected to two low stringency washes (2 ϫ SSC) at room temperature and then two high stringency washes (0.1 ϫ SSC, 0.5% SDS) at 55°C. Hybridization was visualized by autoradiography, and the filters were then stripped and reprobed with a 32 P-labeled glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe to confirm equal loading. Quantification of hybridization signals was achieved by analysis of the scanned autoradiograms using the NIH Image program (ImagePC).
Immunoblotting of PP5-Western analysis was performed essentially as described previously using polyclonal rabbit antibodies generated against a synthetic 15-amino acid peptide identical to the Cterminal region of PP5 (14). Briefly, A549 cells grown in T-75 flasks were washed twice with ice-cold phosphate-buffered saline. Then 250 l of lysis buffer was added to each flask. The extract was then subjected to centrifugation at 13,000 ϫ g for 5 min, and an aliquot of the supernatant was removed for protein determination. The remaining supernatant was added to an equal volume of 2 ϫ sample buffer (120 mM Tris-HCl, pH 7.4, 200 mM dithiothreitol, 20% glycerol, 4% SDS, and 0.02% bromphenol blue). Protein was determined using a Bio-Rad protein quantitation assay, with bovine serum albumin as standards. Typically 25-50 g of protein was then separated by electrophoresis on 10% SDS-polyacrylamide gels. The gel was electrophoretically transferred to Immobilon-P (Millipore Corp.) and blocked for 1 h with Tris-HCl, pH 7.4, containing 150 mM NaCl and 5% nonfat milk. To detect PP5, membranes were incubated with anti-PP5 antibody diluted in Tris-HCl (pH 7.6), 150 mM NaCl, 0.2% Tween 20 containing 2% nonfat milk for 18 h at 4°C. The membrane was then washed, and the primary antibody was detected employing ECL Western blotting detection reagents (Amersham Pharmacia Biotech), following the protocols of the manufacturer.
Analysis of Cell Growth-A549 cells were seeded in 12-well tissue culture plates at a density of 5.0 ϫ 10 4 cells/dish. On the next day, the cells were treated with PP5 specific antisense oligonucleotides (ISIS 15534) or the scrambled mismatch control (ISIS 15521) at a final concentration of 300 nM as described above. On each of the next 5 days, the cell cultures were treated briefly with trypsin to detach the cells from the dish (three wells from each test group). The number of cells was then determined by counting using a hemacytometer. Cell viability was determined with trypan blue staining. The percentage of viable cells was calculated by dividing the number of cells excluding trypan blue by the total number of cells, and the results are reported as the mean Ϯ S.D. of data collected from three independent experiments.
Gel Mobility Shift Assay-Nuclear extracts and gel mobility shift assays were conducted as described previously (34). Briefly, A549 cells were cultured in 60-mm dishes and treated with ISIS 15521 or ISIS 15534 antisense oligonucleotides as described above, three dishes per group. Six hours after treatment, the cells were collected, washed with ice-cold phosphate-buffered saline, and incubated in buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride) at 4°C for 15 min. The cells were then lysed by adding 25 l of 10% Nonidet P-40 and vigorous vortexing for 10 s. Nuclei were precipitated by centrifugation, resuspended in ice-cold buffer B (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride), incubated at 4°C for 60 min with gentle shaking, and then subjected to centrifugation at 4°C for 5 min. The supernatant was collected, aliquoted, and stored at Ϫ80°C. Protein was determined as described above.
Binding reactions were performed by incubating 10 g of each nuclear extract with 1 ng of 32 P-end-labeled p53CON and 1 g of poly(dI-dC) in binding buffer (20 mM HEPES, pH 7.9, 0.25 M EDTA, 50 M KCl, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10% (v/v) glycerol) at room temperature for 20 min. Anti-p53 mouse monoclonal antibody (1 g; Santa Cruz Biotechnology, Inc.) was then added to each reaction and incubated at room temperature for 30 min. Samples were separated by electrophoresis on a 5% polyacrylamide gel containing Tris borate/EDTA buffer and visualized by autoradiography.
Immunoprecipitation-A549 cells were cultured in 60-mm dishes until the cell cultures were ϳ80% confluent, washed with phosphatebuffered saline, and placed for 1 h in phosphate-free DMEM containing 5% fetal bovine serum. The cell cultures were then treated with ISIS 15521 or ISIS 15534 as described above, three dishes per group, and [ 32 P]phosphate (0.2 mCi/ml) was added to the media. Six hours after treatment, the cells were collected, washed with ice-cold phosphatebuffered saline, and stored at Ϫ80°C prior to analysis. Cell lysates were prepared by incubating the thawed cells at 4°C for 1 h in a buffer containing 1% Nonidet P-40, 0.5% sodium deoxycholate, 150 mM sodium chloride, 5 mM EDTA, 1 mM sodium pyrophosphate, 50 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml aprotinin, and 50 mM Tris-HCl, pH 8.0. Insoluble debris was removed by centrifugation at 3000 rpm at 4°C for 15 min. The supernatants were collected and precleared by incubation with 1.2 g of normal mouse IgG (Santa Cruz Biotechnology) and 20 l of protein A-agarose for 1 h at 4°C, followed by centrifugation at 1500 rpm at 4°C for 5 min. The protein concentration of precleared cell lysates was determined as described above, and equal amounts of protein from each precleared lysate were incubated in the presence of 1 g of anti-p53 mouse monoclonal antibody DO-1 (Santa Cruz) for 2 h at 4°C. Twenty l of protein A-agarose was added to the mixture, which was then incubated for 16 h at 4°C with rocking. The agarose beads were collected by centrifugation and washed five times with 1 ml of ice-cold cell lysis buffer. After the final wash, the pellet was resuspended in 40 l of 2ϫ sample buffer, boiled for 3 min, and separated by SDS-polyacrylamide gel electrophoresis on 10% gels. The bands were visualized by autoradiography.

Antisense-mediated Inhibition of PP5 mRNA Expression-
Fifteen oligonucleotides 20 bases in length predicted to hybridize to different regions of human PP5 mRNA were synthesized. The oligonucleotides tested in the initial screen were "chimeric" oligonucleotides, containing eight central phosphorothioate oligodeoxy residues ("oligodeoxy gap") flanked by six 2Ј-methoxyethyl residues on the 3Ј-and 5Ј-ends. These modifications have been shown previously to enhance the potency of antisense oligonucleotides targeting mRNAs encoding other proteins (32,33), and the oligonucleotides tested in the initial screen were designed to target specific regions in the protein coding region or the 3Ј-untranslated region of human PP5 mRNAs (Fig. 1). Because phosphorothioate oligonucleotides have been shown to commonly act through an RNase H-dependent mRNA cleavage mechanisms in cells (35), the ability of each oligodeoxynucleotide to specifically inhibit the expression of PP5 was initially determined by Northern blot analysis probing for levels of PP5 mRNA (Fig. 1A). PP5 mRNA was detected using a PP5-specific cDNA probe, which forms a hybrid with a single ϳ2.3-kilobase transcript.
A comparison of PP5 mRNA levels in A549 human lung carcinoma cells treated with 300 nM PP5-specific antisense oligonucleotides in the presence of cationic lipids is shown in Fig. 1. The cationic lipids (DOTMA/DOPE; Lipofectin®) were used to facilitate the uptake of the oligonucleotides (36), and the reduction in PP5 mRNA levels observed in response to treatment was varied. Treatment with some oligonucleotides had little or no effect on the inhibition of PP5 mRNA levels, while others had a moderate effect and a few had pronounced effects (Fig. 1B). The antisense oligonucleotide with the most potent activity against PP5 mRNA identified in this series was ISIS 14504, which targets a region contained in the 3Ј-untranslated region of PP5. To increase the potency of 14504, oligonucleotides containing additional modifications were synthesized and tested. One oligonucleotide (ISIS 15534), in which the "oligodeoxy gap" region was increased from 8 to 10 contiguous phosphorothioate nucleotides, proved more potent than ISIS 14504, and was chosen for further development. All of the effective oligonucleotides and several of the oligonucleotides with moderate or no effect were also tested against other human cell lines (T-24 bladder carcinoma and TR9-7 human fi-broblasts) for effects on the expression of PP5 mRNA levels with essentially identical results. Western analysis indicates that the treatment of cells with ISIS 15534 also effectively decreases PP5 protein levels in Ͻ18 h (Fig. 1F).
Specificity of PP5 Antisense Inhibition-To examine the specificity of ISIS 15534-mediated effects on PP5 mRNA expression, the effects of ISIS 15534 and mismatch control analogues (i.e. ISIS 15521) were tested. The mismatch control analogues contain the same base composition as ISIS 15534 but with sequences noncomplementary to PP5 (i.e. ISIS 15521). As seen in Fig. 1, D and E, treatment of A459 cells with ISIS 15534 caused a dose-dependent reduction of PP5 mRNA levels with an IC 50 Х 50 nM. No effect was observed following treatment with the mismatch controls (i.e. ISIS 15521), even when applied at concentrations 10 times that of the IC 50 for ISIS 15534. Blots were stripped and reprobed with a glycerol-3phosphate dehydrogenase cDNA probe to demonstrate equal RNA loading. Studies with additional mismatch controls and different cell types produced similar results (data not shown).
To further test the specificity of the antisense oligonucleotides targeting PP5 mRNA, the effect of ISIS 15534 on expression of other cellular mRNA was evaluated. As seen in Fig. 1, D and G, ISIS 15534 has no effect on the expression of structurally related PPases (PP1, PP2A, PP4) or glyceraldehyde-3phosphate dehydrogenase mRNA in A549 cells, even at a concentration of 500 nM (Fig. 1G). Since the sequence targeted by ISIS 15534 is not contained in PP1, PP2A, PP4, or GAPDH, ISIS 15534 would not be expected to inhibit the expression of these proteins if ISIS 15534 is inhibiting the expression of PP5 mRNA via an antisense mechanism. Similar effects were also observed with T-24 bladder carcinoma cells and TR9-7 human fibroblasts.
Antiproliferative Effects of PP5 Antisense Inhibition-Having obtained a suitable method for specifically inhibiting PP5 gene expression, we now wanted to explore the roles played by PP5 in human cells. Preliminary studies suggested that ISIS 15534 inhibited cell proliferation. Thus, we treated A549 cells one time with ISIS 15534 or its mismatched control at a concentration of 300 nM and determined cell number daily over a 5-day period. Treatment with ISIS 15534 resulted in a marked inhibition of A549 cell proliferation over this 5-day period (Fig.  2). Growth was completely inhibited for three days, and the recovery of proliferation rate at days 4 and 5 correlated well with the recovery of PP5 mRNA expression due to the extrusion of the antisense oligonucleotides from the cell and nucleasemediated degradation of ISIS 15534 following a single application of the oligonucleotides (Fig. 2). A similar effect on cell proliferation was not observed with the mismatched control oligonucleotides.
ISIS 15534-mediated Inhibition of Cell Growth Correlates with the Induction of p21 WAF1/Cip1 -Flow cytometry analysis of cells treated with ISIS 15534 indicated that growth arrest occurs prior to the onset of DNA synthesis (data not shown), suggesting that PP5 is necessary for some aspect of cell cycle progression that occurs before the S phase of the cell cycle is initiated. Thus, we explored several possible mechanisms by which PP5 could affect known pathways leading to G 1 growth arrest. These studies revealed that the inhibition of PP5 expression has a marked effect on the expression of mRNA encoding the cyclin-dependent kinase inhibitor p21 WAF1/Cip1 . As seen in Fig. 3, the amount of p21 WAF1/Cip1 mRNA contained in logarithmically growing A549 cells treated with ISIS 15534 at a concentration of 200 -500 nM increases by Ͼ10-fold over that observed in cells treated with the mismatched control oligonucleotides. The increase in p21 WAF1/Cip1 mRNA levels that is observed following treatment with ISIS 15534 occurs in a dose-dependent manner and correlates well with the dose-dependent inhibition of PP5 mRNA expression (Fig. 3). This suggests there is an inverse relationship between the expression of PP5 and the expression of p21 WAF1/Cip1 . The p21 WAF1/Cip1 gene encodes an inhibitor of most cyclindependent kinases (37-41) and has been implicated as a mediator of growth arrest in many cell types. One pathway lead-ing to the induction of p21 WAF1/Cip1 and growth arrest is regulated by the p53 tumor suppressor protein (for a review, see Ref. 42). In response to DNA damage (42,43) or alterations in cellular homeostasis (44), normal cells can respond by increasing the induction of p53. Although the mechanism(s) leading the induction of p53 are unclear, p53 acts as a transcription factor, and increased expression of wild-type p53 stimulates the synthesis of p21 WAF1/Cip1 . In turn, p21 WAF1/Cip1 inhibits the activity of cyclin D/cyclin-dependent kinases 4 and 6 and/or cyclin E/cyclin-dependent kinase 2, preventing the phosphorylation of the retinoblastoma protein (Rb) and inducing cell growth arrest late in the G 1 stage of the cell cycle (44). Studies in p53-deficient cells (31,45) indicate that the increased expression of p53 alone (i.e. in the absence of DNA damage) is capable of stimulating the synthesis of p21 WAF1/Cip1 and growth arrest in the G 1 phase of the cell cycle.
To further characterize the mechanism by which the inhibition of PP5 gene expression produces an increase in p21 WAF1/Cip1 mRNA levels and growth arrest, we tested the ability of ISIS 15534 to influence the expression of p53. Northern analysis of p53, PP5 and p21 WAF1/Cip1 mRNA levels in A549 cells following treatment with ISIS 15534 or the mismatched control oligonucleotide revealed that the increase in p21 WAF1/Cip1 expression that occurs in response to treatment with ISIS 15534 does not result in an apparent change in the amount of p53 mRNA (Fig. 4A). As a positive control, the A549 cells were treated with UV radiation, which, as predicted, resulted in an increase in both p53 and p21 WAF1/Cip1 mRNA. Interestingly, following UV radiation treatment of A549 cells, the level of PP5 mRNA is reduced and the amount of p21 WAF1/Cip1 mRNA is a little greater than that achieved with the inhibition of PP5 expression alone (Fig. 4A).
Induction of p21 WAF1/Cip1 via the Inhibition of PP5 Expression Is Dependent on the p53 Protein-To test the possibility that the inhibition of PP5 mRNA expression leads to an increase in p21 WAF1/Cip1 mRNA levels via a p53-independent mechanism, we tested the effects of ISIS 15534 on p53-deficient T-24 human bladder cancer cells (46,47). As seen in Fig.  4B, although the treatment of T-24 cells with ISIS 15534 clearly inhibits the expression of PP5 mRNA, the amount of p21 WAF1/Cip1 mRNA was not increased. Furthermore, unlike the observations in A549 cells, treatment of T-24 cells with ISIS 15534 does not effectively induce G 1 growth arrest (Table  I). These studies suggest that the ability of ISIS 15534 to induce p21 WAF1/Cip1 expression emanates from a mechanism that is dependent on the presence of p53 yet does not require an increase in p53 expression and suggest that PP5 acts upstream of p53 in this pathway.
The Inhibition of PP5 Expression Enhances the Phosphorylation State of p53-The reversible phosphorylation of serine residues is believed to influence the biological activity of p53, and increased phosphorylation has been reported to enhance both the DNA binding activity and the transcriptional activity of p53 (47)(48)(49). To determine if PP5 affects the activity of the p53 protein under conditions where the expression of the p53 gene is tightly regulated, we next tested the ability of ISIS 15534 to affect p21 WAF1/Cip1 transcription in TR9-7 cells. TR9-7 cells are a stable cell line derived from MDAH041 p53-null human fibroblasts that contain tetracycline-regulated transactivator and operator plasmids to control the expression of wildtype p53 (31). When TR9-7 cells are grown in the presence of 1 g of tetracycline, p53 protein is expressed at a very low level, which is insufficient to induce p21 WAF1/Cip1 or cell cycle arrest (31,45). Treatment of TR9-7 cells grown in the presence of tetracycline with ISIS 15534 inhibits the expression of PP5 mRNA, induces the expression of p21 WAF1/Cip1 mRNA, and induces growth arrest without a corresponding increase in the expression of p53 mRNA (Fig. 4C). In contrast, in the p53-null human fibroblasts (TR9; Fig. 4C), as also observed in p53deficient T-24 human bladder cancer cells (Fig. 4B), treatment with ISIS 15534 inhibits the expression of PP5 mRNA without inducing p21 WAF1/Cip1 or G 1 growth arrest (Fig. 4C). These studies add additional support to the theory that PP5 regulates p21 WAF1/Cip1 expression via a pathway that requires the p53 protein but does not require an increase in p53 gene transcription.
Further evidence for the involvement of PP5 in the regulation of p53-mediated induction of p21 WAF1/Cip1 and growth arrest comes from mobility gel shift analysis and immunoprecipitation studies in A549 cells. Following treatment with ISIS 15534, gel shift analysis indicates that the ability of p53 to bind DNA is enhanced when the expression of PP5 is inhibited (Fig.  4D). In addition, immunoprecipitation of p53 from 32 P-equilibrated cells reveals that treatment with ISIS 15534 markedly enhances the phosphorylation state of p53 (Fig. 4E), and recombinant PP5 can dephosphorylate p53 in vitro (data not shown).

DISCUSSION
Deregulation of cell proliferation is a hallmark of neoplastic transformation, and understanding how alterations in growth control pathways are translated into changes in the cell cycle regulatory machinery is a major challenge for understanding the development of human cancers. The ability of both tumor suppressor proteins, p53 and BRCA1, to induce the expression of p21 WAF1/Cip1 (37)(38)(39), in combination with the inhibitory activity of p21 WAF1/Cip1 against cyclin-dependent kinases, suggests that the regulation of p21 WAF1/Cip1 expression is an important aspect of mammalian cell cycle growth control (40,41). Accordingly, there has been a tremendous effort to elucidate the complex cellular mechanisms regulating the activity and expression of these tumor suppressor proteins.
Since virtually all known examples of transcriptional control during the cell cycle involves phosphorylation and the cyclin-dependent kinases are serine/threonine protein kinases, we suspected that certain PPases may also participate in the regulation of cell cycle progression. However, because there were no truly specific inhibitors of any known mammalian PPase, before we could study the functions of a specific PPase, we first had to develop a method of distinguishing events mediated by a particular PPase from those mediated by other structurally related PPases (1,29). For this purpose, we developed antisense oligodeoxynucleotides targeting human PP5 that specifically inhibit PP5 gene expression. Studies conducted in A549 cells with a potent and highly specific antisense oligodeoxynucleotide targeting human PP5 (ISIS 15534) revealed that the inhibition of PP5 expression arrests cell growth prior to the initiation of DNA synthesis. ISIS 15534-mediated growth arrest correlates with an increase in p21 WAF1/Cip1 mRNA, and a dose-dependent correlation between the increase in p21 WAF1/Cip1 mRNA, and the decrease in PP5 mRNA is observed following treatment with ISIS 15534 but not with the mismatch control oligonucleotides. Comparison of the effects of ISIS 15534 treatment in p53 competent cells (A549), p53 protein-deficient cells (T-24), p53 null human fibroblasts (TR9-cells), and p53 null human fibroblasts that contain tetracycline-regulated transactivator and operator plasmids to control the expression of wild-type p53 (TR9-7) indicates that ISIS 15534-mediated induction of p21 WAF1/Cip1 is dependent on the presence the p53 protein but does not require an increase in p53 gene expression. Immunoprecipitation studies and gel shift analysis conducted with A549 nuclear extracts indicate that the treatment of cells with ISIS 15534 induces the phosphorylation of p53 and enhances the ability of p53 to bind DNA. Together, these studies suggest that PP5 acts upstream of p53 to control the transcription of p21 WAF1/Cip1 and the arrest of cell growth in the G 1 phase of the cell cycle. PP5 appears to function by regulating the phosphorylation state of p53, which governs the ability of p53 to bind DNA and induce the transcription of p21 WAF1/Cip1 , without affecting p53 gene transcription. This suggests that PP5 is a regulatory component of a pathway that affords replicating cells G 1 checkpoint growth control and that it is the regulation of PP5 that, in turn, controls p53-mediated expression of p21 WAF1/Cip1 and growth arrest in this pathway.
Previous studies have demonstrated that p53 has many functions, including the ability to induce the transcription of p21 WAF1/Cip1 and arrest cell growth in the G 1 phase of the cell cycle (44). In addition, several laboratories have provided data suggesting that the phosphorylation of p53 increases both the transcriptional activity of p53 and the ability of p53 to bind DNA (49,50,58,60,(62)(63)(64). Okadaic acid, a potent inhibitor of PP1, PP2A, PP4, and PP5, has also been shown to enhance the phosphorylation state of p53 (62,65). The present studies are consistent with all of these previous studies, and since the loss of PP5 expression correlates with an increase in p53 phosphorylation, PP5 may be a major p53 phosphatase in vivo. However, it is unlikely that PP5 is the only PPase that utilizes p53 as a substrate in vivo, because the inhibition of PP2A activity has also been reported to enhance the phosphorylation of p53 (62). In addition, an increase in p53 phosphorylation could also be obtained if PP5 functions to activate protein kinases or inhibit other PPases that utilize p53 as a substrate. Furthermore, p53 is phosphorylated on several sites in vivo (49,50,58,60,(62)(63)(64), and a recent study indicates that a basal level of PP1 or PP2A activity is necessary to prevent p53-mediated apoptosis (62). Thus, the ability of p53 to induce either apoptosis or reversible G 1 growth arrest could depend on which kinases and PPases are activated by a particular event. Nonetheless, because 1) ISIS 15534 inhibits the expression of PP5 and has no affect on the expression of PP1, PP2A, or PP4; 2) the treatment of cells with ISIS 15534 at concentrations sufficient to inhibit the PP5 expression alone triggers the expression of p21 WAF1/Cip1 and induces growth arrest in p53-competent cells; and 3) the inhibition of PP5 expression does not induce the expression of p21 WAF1/Cip1 or growth arrest in p53-deficient cells, the data obtained from the present studies indicates that PP5 acts upstream of p53 in at least one In addition to the data presented here, only a few studies have attempted to determine the functions of PP5. PP5 con-tains multiple TPR domains, which are believed to be sites of protein-protein interactions (14,30,55,57). Indeed, two-hybrid analysis studies suggest that PP5 interacts via TPR domains with the atrial natriuretic peptide receptor (14), two TPRcontaining subunits (CDC16, CDC27) of a complex referred to as the cyclosome or anaphase-promoting complex (55), and a human flavoprotein protein (hCRY2), with homology to the photolyase/blue light photoreceptor gene family of lower eukaryotes, plants, and bacteria (66). PP5 has also been reported to interact with the glucocorticoid receptor-hsp90 complex (56), and studies with purified and recombinant bacterially expressed PP5 suggest that polyunsaturated fatty acids, arachidonic acid, and limited proteolysis can increase PP5 activity in vitro (13,68).
Our studies provide no apparent insight into the relationship of PP5 with the atrial natriuretic peptide receptor or mechanisms that lead to the activation of PP5. In addition, because ISIS 15534 inhibits cell growth prior to the initiation of S phase, we are unable to assess the role of PP5 in the regulation  4. Inhibition of PP5 mRNA expression in cells containing wild-type p53, but not in p53-deficient cells, correlates with an increase in p21 WAF1/Cip1 mRNA expression, an increase in the binding affinity of p53 to DNA, and increased p53 phosphorylation, without an increase in p53 mRNA expression. Panel A, effect of UV radiation or treatment with ISIS 15534 on the expression of p53, PP5, p21 WAF1/Cip1 , and GAPDH expression in p53-competent human lung carcinoma cells. A549 cells were treated with nothing (UV; C), 50 Jm Ϫ2 of UV radiation (UV; T), a mismatched control oligodeoxynucleotide (ISIS-15534; C) or ISIS 15534 (ISIS-15534; T), and 6 h (UV) or 24 h (ISIS 15534) later the total RNA was prepared and analyzed for p53, PP5, p21 WAF1/Cip1 , and GAPDH mRNA levels by Northern blot analysis. Panel B, effect of ISIS 15534 on the expression of p53, PP5, p21 WAF1/Cip1 , and GAPDH expression in p53-deficient bladder carcinoma cells. T-24 cells were treated with mismatch control oligodeoxynucleotides (C) or ISIS 15534 (T), and 24 h later the total RNA was analyzed by Northern blot analysis as described in A. Panel C, comparison of the effect of ISIS 15534 on p53 null human fibroblasts (TR9) and TR9 cells containing tetracycline-regulated transactivator and operator plasmids to control the expression of wild-type p53 (TR9-7). TR9-7 cells (generously provided by Drs. George Stark and Munna Agarwal) grown in the presence of 0.8 g/ml of tetracycline to keep the expression of p53 below the threshold necessary for the induction of p21 WAF1/Cip1 and treated with mismatch control oligodeoxynucleotides (C) or ISIS 15534 (T) at a concentration of 300 nM. Total RNA was prepared 24 h later and analyzed by Northern blot analysis as described for panel A. Panel D, the inhibition of PP5 expression enhances p53 binding to DNA. Nuclear extracts were prepared from A549 cell cultures treated with mismatched control oligodeoxynucleotides (ISIS 15521) or ISIS 15534 as described above. After 6 h, the ability of p53 to bind DNA was analyzed by gel mobility shift assay. Lane 1, no protein control: migration of 32 P-p53CON probe in the absence of nuclear extracts. Lane 2, mismatched control oligodeoxynucleotides: nuclear extracts prepared from mismatch control oligodeoxynucleotide-treated cells. Lane 3, ISIS 15534: nuclear extracts from cells treated with ISIS 15534. Lane 4, excess cold probe control: samples treated in an identical manner as in lane 3 and incubated in the presence of excess cold p53CON probe. Panel E, increased phosphorylation of p53 following treatment with ISIS 15534. A549 cells were cultured in 32 P i and treated with ISIS 15534 or mismatch control oligodeoxynucleotides. Six hours later, the protein extracts were prepared, and changes in p53 phosphorylation were determined by immunoprecipitation SDS-polyacrylamide gel electrophoresis analysis and visualized by autoradiography. Lane 1, p53 from cells treated with mismatched control oligodeoxynucleotides. Lane 2, p53 from cells treated with 300 nM ISIS 15534. of the anaphase-promoting complex. However, our data may support a role for PP5 in glucocorticoid receptor-mediated cell cycle arrest. Glucocorticoids are known to inhibit proliferation of many cell types, yet the events occurring subsequent to the formation of an activated glucocorticoid receptor complex leading to growth arrest are unknown. A recent study conducted with mouse L929 fibroblast cell cultures suggests that dexamethasone treatment caused an increase in p21 WAF1/Cip1 mRNA levels and growth arrest (67). Other studies indicate that PP5 associates with the mouse glucocorticoid receptor-hsp90 complex (56). Thus, if p53 is a component of a signaling pathway that links glucocorticoid receptor activation with the induction of p21 WAF1/Cip1 , then the data presented here are consistent with PP5 functioning as a negative regulator of a signaling cascaded by which glucocorticoids induce growth arrest. Clearly, this hypothesis is highly speculative, since the above mentioned studies were conducted with three different cell types and the effects of glucocorticoids are often cell typespecific. Nonetheless, since reversible phosphorylation may affect 1) glucocorticoid receptor agonist binding, 2) nuclear transport of the activated glucocorticoid receptors, 3) the ability of activated glucocorticoid receptors to bind DNA, 4) the transcriptional activity of glucocorticoid receptors once associated with glucocorticoid responsive elements, or 5) any one of the numerous downstream signaling events initiated by glucocorticoid transactivation, this is an attractive area for further investigation.
Another interesting observation is that PP5 mRNA levels are lower in A549 cells following exposure to UV radiation (Fig. 4A). The decrease in PP5 mRNA levels occurring after UV treatment suggests a potential role for PP5 in a p53mediated response to DNA damage (UV radiation 3 2PP5 3 p53 active 3 1p21 WAF1/Cip1 3 G 1 growth arrest). However, since increased p53 expression is also observed following UV radiation treatment, additional studies will be necessary to clarify the relationship of PP5 and p53 in the response to UV radiation. Finally, recent studies indicate that the functional expression of the human p21 WAF1/Cip1 gene in rat glioma cells suppresses tumor growth in vivo (61), and the survival of patients with advanced gastric carcinoma correlates with the expression of p21 WAF1/Cip1 (59). When considered together with the data presented here, these studies indicate that inhibitors of PP5 gene expression, regardless of their mechanism of action, may be useful to elevate the level of p21 WAF1/Cip1 expression in cells harboring functional defects in p53, BRCA1, p73, and possibly other tumor suppressor protein-mediated pathways to restore G 1 checkpoint growth control. Thus, if effective in vivo, compounds that inhibit the expression and/or activity of PP5 may be useful clinically as an alternative therapy for the treatment of human cancers in which the ability to induce p21 WAF1/Cip1 has been diminished or lost. Since defects in p53-mediated pathways are associated with Ͼ50% of all human cancers, the development of such compounds has the potential to have a major impact on the treatment of many human cancers.