p53 Regulates the Expression of the Tumor Suppressor Gene Maspin*

Maspin has been shown to inhibit tumor cell invasion and metastasis in breast tumor cells. Maspin expression was detected in normal breast and prostate epithelial cells, whereas tumor cells exhibited reduced or no expression. However, the regulatory mechanism of maspin expression remains unknown. We report here a rapid and robust induction of maspin expression in prostate cancer cells (LNCaP, DU145, and PC3) and breast tumor cells (MCF7) following wild type p53 expression from an adenovirus p53 expression vector (AdWTp53). p53 activates the maspin promoter by binding directly to the p53 consensus-binding site present in the maspin promoter. DNA-damaging agents and cytotoxic drugs induced endogenous maspin expression in cells containing the wild type p53. Maspin expression was refractory to the DNA-damaging agents in cells containing mutant p53. These results, combined with recent studies of the tumor metastasis suppressor gene KAI1 and plasminogen activator inhibitor 1 ( PAI1 ), define a new category of molecular targets of p53 that have the potential to negatively regulate tumor invasion and/or metastasis. the Dual- luciferase reporter assay system (Promega). The results are presented as the -fold induction of the reporter plasmid alone after normalization with the internal control plasmid pRL-TK. Gel Mobility Shift Assay— The gel mobility shift assay was performed as described previously (14). Briefly, labeled oligonucleotide probes (2 ng) were incubated with 30 ng of nonspecific competitor DNA and 50 ng of wt p53 (purified from the bacculovirus expression system) in 50 m M Tris, 100 m M NaCl, 1 m M dithiothreitol at 4 °C for 30 min. The PAb-421 antibody was added after 30 min of incubation with wt p53. The complexes were analyzed on a native 12% polyacrylamide gel. The concen- tration of the protein was adjusted so that it formed an approximately 50:50 complex with the probes.

Maspin was originally identified in normal breast epithelial cells (1). The maspin gene encodes a 42-kDa protein and belongs to the serine protease inhibitor (serpin) superfamily with tissue type plasminogen activator as the possible protease target (2). Maspin expression was detected in normal breast and prostate epithelial cells; however, tumor cells showed a decreased expression or absence of expression. Expression of maspin in breast tumor cells inhibit tumor cell invasion in vitro and tumor cell metastasis in vivo (1). Neutralization of maspin by an anti-maspin antibody abolished the invasion suppressive effect of conditioned medium from cultured breast myoepithelial cells on tumor cells (3). A recent report also suggests that the tumor suppressive effects of manganese-containing superoxide dismutase in human breast cancer cells could result from the up-regulation of maspin (4). Gamma linolenic acid, an essential fatty acid with anticancer properties, is reported to induce maspin expression and affect the motility of cancer cells (5). Transcriptional activity of maspin expression differed between prostate normal and tumor cells (6). These observations suggest that maspin expression plays important roles in regulating tumor cell invasion and metastasis. Thus, an understanding of the regulation of maspin expression is important in designing therapeutic agents for the cancer treatment.
Molecular targets of p53, e.g. p53-regulated genes or p53interacting proteins, have provided critical information central to the current understanding of the biochemical and biologic function of the p53 tumor suppressor gene. The function of p53 as check point protein is now well established (7). p53-regulated genes have also defined the role of p53 in apoptosis, hypoxia, and angiogenesis (8 -10). However, the downstream targets of p53 remain to be defined in the process of cancer cell invasion/metastasis. In our search for molecular targets of p53 involved in cell invasion and metastasis, we have now discovered that maspin expression is regulated by wild type (wt) 1 p53. In this report, we provide biochemical and cellular biologic evidence demonstrating that maspin is directly regulated by p53.

MATERIALS AND METHODS
Cell Culture-Prostate tumor cell lines DU145, LNCaP, and PC3, breast tumor cell line MCF7, and colon carcinoma cell line HCT116 were obtained from American Type Culture Collection (ATCC, Manassas, VA) and were maintained in the growth medium recommended by the supplier. Experimental conditions for infection of the cells with recombinant adenovirus vectors expressing wild type p53, p21 waf/cip1 , or p27 have been described previously (11,12).
Northern Blot and Western Blot Analyses-Total cellular RNA was extracted from cells by RNAzol method (Life Technologies, Inc.). Ten g of total RNA were fractionated on a 1% agarose gel and transferred to a nylon membrane. The membrane was hybridized with a solution (5ϫ SSC, 5ϫ Denhardt's solution, 40% formamide, 10% dextran sulfate, 10 mM Tris-HCL, pH 7.5) containing a randomly labeled 32 P cDNA probes at 40°C for 18 -20 h. The membrane was washed twice at room temperature with a 2ϫ SSC, 0.1% SDS solution followed by two washes with a 0.1ϫ SSC, 0.1% SDS solution at 50°C. The membrane was then exposed to x-ray film. The DNA fragments used for hybridization were generated by PCR from normal prostate cDNA (CLONTECH). The maspin probe was a 585-bp PCR product spanning nucleotides 576 -1160. The p21 waf1/cip1 was a 318-bp fragment spanning nucleotides 1745-2035 of the p21 waf1/cip1 cDNA. The p27 probe was a 597-bp fragment spanning nucleotides 1-597 of the p27 cDNA. The identity of PCR-derived probes was confirmed by DNA sequencing. A maspin monoclonal antibody was obtained from PharMingen. Total cellular lysate was separated on a 10% SDS-polyacrylamide gel and transferred onto a nitrocellulose membrane. Western blot analysis was performed using the ECL system. Plasmids and Constructs-The wild type and mutant p53 expression plasmids used in the luciferase assay were described previously (13). The promoter region of maspin was amplified by PCR according to the reported DNA sequence (6). The pM-Luc(Ϫ759) was generated by primers GAGACTCGAGGCTGAAGTACAGTGGTTAG (with the XhoI site) and GAGAAAGCTTAGAAGCAGCGGTGGCTCACC (with the HindIII site). The pM-Luc(Ϫ596) was generated by primers GAGACTCGAGGT-TGGTCTCAAACTCCTG (with the XhoI site) and GAGAAAGCTTA-GAAGCAGCGGTGGCTCACC (with the HindIII site). The DNA fragment was cloned into the XhoI and HindIII sites of the pGL3 basic vector (Promega). pM-Luc(Ϫ297) was made by deleting the PstI fragment from pM-Luc(Ϫ759). The three constructs ended at ϩ87 nucleotides from the transcription start site of the maspin. The sequence of the constructs was verified by DNA sequencing. pM-Luc(Ϫ297mt1) and pM-Luc(Ϫ297mt2) were generated by PCR-based site-directed mutagenesis using pM-Luc(Ϫ297) as the template. The p53 binding site was mutated to the sequence shown in the mutant oligonucleotide used in the gel shift assay.
Transfection and Luciferase Assay-The cells were plated at 5 ϫ 10 5 cells/well (6 wells/plate) 1 day before the transfection. The transfection was performed using the calcium phosphate method (CLONTECH). The maspin promoter-reporter plasmid (5 g), the p53 plasmid (2.5 g), and an internal control plasmid, pRL-TK (0.5 g), were cotransfected into cells for 48 h, and the cells were harvested for the luciferase assay. Luciferase activity was measured by a luminometer using the Dualluciferase reporter assay system (Promega). The results are presented as the -fold induction of the reporter plasmid alone after normalization with the internal control plasmid pRL-TK.
Gel Mobility Shift Assay-The gel mobility shift assay was performed as described previously (14). Briefly, labeled oligonucleotide probes (2 ng) were incubated with 30 ng of nonspecific competitor DNA and 50 ng of wt p53 (purified from the bacculovirus expression system) in 50 mM Tris, 100 mM NaCl, 1 mM dithiothreitol at 4°C for 30 min. The PAb-421 antibody was added after 30 min of incubation with wt p53. The complexes were analyzed on a native 12% polyacrylamide gel. The concentration of the protein was adjusted so that it formed an approximately 50:50 complex with the probes.

RESULTS AND DISCUSSION
To determine whether the tumor suppressor p53 regulates maspin expression in prostate tumor cells, maspin mRNA expression was analyzed in the metastatic prostate cancer line DU145 following wt p53 expression via an adenovirus vector, AdWTp53. DU145 cells were infected with AdWTp53 or the control vector dl312 and were harvested at different time points for RNA isolation. The induction of maspin expression was first noted 3 h after AdWTp53 infection and reached a plateau by 6 h that persisted to at least 48 h (Fig. 1A). The control vector did not increase maspin expression under the same conditions. The kinetics of p53-mediated maspin induction was rapid and was similar to the kinetics of expression of the well known p53-regulated gene, p21 waf/cip1 . To demonstrate that maspin was specifically induced by p53, DU145 cells infected with adenovirus vectors expressing p21 waf1/cip1 or p27 were analyzed for expression of maspin (Fig. 1B). p21 waf1/cip1 and p27 were highly expressed but did not induce detectable maspin expression. Only p53 stimulated the maspin expression. These results suggested that maspin expression was specifically induced by p53 and was not the result of nonspecific effects of cell growth arrest and/or apoptosis. p53 induction of maspin expression was also noted in prostate tumor cell lines PC3 and LNCaP and in the breast tumor cell line MCF7 (Fig. 1C).
DNA-damaging agents and cytotoxic drugs are inducers of p53 expression that lead to the induction of downstream target genes, e.g. p21 waf1/cip1 . To demonstrate whether maspin expression is inducible in response to DNA damage, cells were UV-irradiated or treated with etoposide (VP16). As shown in Fig. 2, UV irradiation or etoposide treatment induced maspin expression in LNCaP cells containing endogenous wt p53; however, DU145 cells harboring mutant p53 did not respond to such treatment. UV irradiation increased maspin expression in MCF7 cells containing wild type p53 but not in PC3 cells null for p53 protein (data not shown). We repeated this experiment in colon tumor cells (HCT116) that contain wt p53. In addition to UV and etoposide (VP16), maspin expression was also induced by ␥ irradiation and adriamycin in HCT116 cells (data not shown).
To demonstrate that p53 directly regulates maspin expression, we conducted gel shift assays to analyze the binding of purified p53 protein to the maspin promoter. Sequence analysis of the maspin promoter revealed three imperfect consensus p53 binding sites in the region between Ϫ297 and the transcription start site. As shown in Fig. 3, p53 protein exhibited binding to two of the oligonucleotides (CCCGAACATGTTGGAGGCCTT-TTGA and TGTGGACAAGCTGCCAAGAGGCTTGAGTAGG) that contain the p53 consensus sequence. The binding was supershifted by adding the p53 antibody, Pab421. When mutations (indicated by underlining) were introduced in the FIG. 1. Northern analysis of maspin induction in prostate tumor cells. Total RNA (10 g) was fractionated on 1% agarose-formaldehyde gel and transferred onto a nylon membrane. The blot was hybridized with 32 P-labled DNA probes. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the loading control. A, Northern analysis of maspin induction by p53. DU145 cells were infected with either 20 plaque-forming units/cell AdWTp53 or the control vector dl312 and collected for RNA at the indicated time points (hours). ϩ, cells infected with AdWTp53; Ϫ, cells infected with control vector dl312. B, maspin expression specifically induced by p53. Tumor cells were infected with the adenovirus vector expressing the p21, p27, and p53 at plaque-forming units ϭ 20 for 24 h and then harvested for RNA isolation. The same blot was hybridized sequentially with the maspin, p21 waf1/cip1 , p27, and GAPDH probes, with complete removal of probes between hybridizations. C, induction of maspin expression in tumor cells. Tumor cells were infected with either AdWTp53 or control vector dl312 at 20 plaque-forming units/cells for 24 h. Total RNA was isolated and subjected to Northern blot analysis.
p53 Regulates Expression of Maspin following oligonucleotides, the p53 protein failed to bind to the oligonucleotides: CCCTATATACAACGAGGCCTTTTGGA, TGTCCTGTAGCTACCTAGACCGTTGAGTAGG. These results provides evidence that p53 has the potential to bind the maspin promoter.
Although we demonstrated that wt p53 activated the transcription of the endogenous maspin gene, to further confirm that p53 directly regulates maspin expression, we used a maspin promoter-luciferase reporter assay to examine whether p53 activates the promoter of the maspin gene. Three constructs with varying lengths of the maspin promoter were FIG. 3. Binding of p53 to the maspin promoter. The gel mobility shift assay was carried out as described under "Materials and Methods." A p53 consensus oligonucleotide from the p21 waf1/cip1 promoter (5Ј-CCCGAACATGCCTCAACATGTTGGGA-3Ј) was used as control. The oligonucleotides from the maspin promoter are: oligo 1, 5Ј-CCCGAACATGTTGGAGGCCTTTTGGA-3Ј; oligo 1 mutant, 5Ј-CCCTATATACAACGAGGCCTTTTGGA-3Ј; oligo 2, 5Ј-TGTGGA-CAAGCTGCCAAGAGGCTTGAGTAGG-3Ј; oligo 2 mutant, 5Ј-TGTCCTGTAGCTACCTAGACCGTTGAGTAGG-3Ј. The 32p endlabeled oligonucleotides were incubated with p53 with or without p421, and the complexes were analyzed on 12% native polyacrylamide gels.  4. Activation of maspin promoter. Top panel, diagram of the maspin promoter-luciferase constructs. The p53 consensus-binding site is indicated. R is a purine, Y is a pyrimidine, and W is adenine or thymidine. Lowercase letters in the maspin binding sites indicate the mismatches. pM-Luc(Ϫ297) was used to generate the mutant constructs pM-Luc(Ϫ297mt1) and pM-Luc(Ϫ297mt2) in which the binding sites were mutated to the same sequence as the mutant oligonucleotide used in the gel shift assay. Maspin promoter reporter (5 g) plasmid DNA was cotransfected with either pcDNA3p53 wt (2.5 g) or pcDNA3p53mt (2.5 g) and an internal control plasmid pRl-TK (0.5 g) into PC3 cells (5 ϫ 10 5 cells/well) for 48 h. The cells were lysed in lysis buffer provided by the manufacturer (Promega). The control vector and maspin promoter reporter were cotransfected with the pcDNA3 vector (2.5 g) for control purposes. Luciferase activity was measured on a luminometer and is reported in arbitrary units. The data are presented as -fold induction over the reporter construct alone. The same experiment was repeated three times in DU145 and LNCaP cells. Bar, S.E.  (lanes 1, 2, 5, and 6) and DU145 (lanes 3, 4, 7, and 8) cells were exposed to UV irradiation at 25 J/m 2 with a Stratagene UV Stratalinker 2400 and cultured for 24 h (lanes 2 and 4) or treated continuously with etoposide (VP16) at 5 g/ml (lanes 6 and 8) for 24 h. Lanes 1, 3, 5, and 7 were controls without treatment. Equal amounts of total cell lysate (approximately 10 4 cells/lane) were electrophoresed on 10% polyacrylamide gels. Western blot detection was performed using the ECL system. Mouse monoclonal antibody for maspin was used for the detection of maspin. The tubulin antibody was used to monitor equal loading for each lane.
cloned into a luciferase reporter pGL3-basic vector as described under "Materials and Methods" and as indicated in the top panel of Fig. 4. The promoter region used in these constructs contains the full promoting activity as demonstrated previously (6). The expression of wt p53 or mt p53 (codon 245, Gly 3 Asp) was driven by a cytomegalovirus promoter (pcDNA3). The maspin promoter-luciferase reporter plasmid was co-transfected with the wt p53 or mt p53 expression vector. The results showed that wt p53 induced luciferase activity of the maspin promoter-reporter by more than 3-4-fold compared with the activity of the reporter plasmid alone (Fig. 4). The mt p53 expression vector was not able to induce luciferase activity under similar conditions. All three constructs showed similar induction by the wt p53. These results suggest that the p53responsive element must reside within the Ϫ297 to ϩ87 region relative to the transcription start site. There are two p53 binding sites in this region. The shortest promoter construct, pM-Luc(Ϫ297), was then mutated by site-directed mutagenesis at the p53 binding sites to generate constructs pM-Luc(297mt1) and pM-Luc(Ϫ297mt2). The mutant sequences in these constructs were the same as the mutant oligonucleotides used for gel shift assays (Fig. 3). As shown in Fig. 4, with the construct containing a mutant binding site 1, pM-Luc(Ϫ297mt1), the activation of the promoter was completely abolished. When binding site 2, pM-Luc(Ϫ297mt2), was mutated, the activation of the promoter was not affected. This result suggests that binding site 1 is likely to be the functional site for p53 activation.
In this report we demonstrate a rapid wt p53-dependent induction of maspin expression in prostate and breast tumor cells. These results suggest a role for wt p53 in the regulation of the maspin function involved in cell invasion or metastasis. A recent study has shown that p53 activates the expression of the metastasis suppressor gene KAI1 (15). KAI1 expression was decreased during prostate tumor progression and low expression of KAI1 correlates with the loss of p53 function. Despite the variable frequency of p53 mutations reported in primary prostate tumors, metastatic tumors consistently exhibit a higher incidence of p53 mutations (16,17). Although there is only one report on the analysis of maspin expression in prostate cancer cells, it is notable that maspin expression was downregulated in prostate tumor cell lines similar to the findings in mammary tumors cell lines (18). In an earlier report, Zou et al.
(1) have already demonstrated that breast cancer cell-harboring maspin expression vectors exhibit decreased cell invasion in vitro and reduced metastasis in vivo. On the basis of the known biologic functions of maspin and this study showing regulation of maspin expression by wt p53, we suggest a functional interaction of p53 and maspin in cell invasion. This report, along with other studies (15,19), underscores the broader implications of these findings and emphasizes the role of p53 in the negative regulation of cell invasion and metastasis. p53 may suppress tumor metastasis by up-regulating metastasis suppressor genes, e.g. maspin, KAI1, and PAI1. This newly emerging p53 function may provide a mechanistic explanation for the increased metastatic susceptibility of tumors harboring p53 mutations.