Activation of Adenomatous Polyposis Coli (APC) Gene Expression by the DNA-alkylating AgentN-MethylN′-nitro-N-nitrosoguanidine Requires p53*

Development of colon cancer is a multistep process frequently involving mutations in both theAPC and p53 tumor suppressor genes. In this study we treated the HCT-116 colon cancer cell line with alkylating agents includingN-methyl-N′-nitro-N-nitrosoguanidine (MNNG),which is known to cause colon cancer in animals, and examined the expression of both APC and p53genes. Exposure of cells with MNNG caused an 8–12-fold increase in the level of APC mRNA and protein. APC induction was shown to result from increased nuclear transcription of the APCgene and correlated with a concomitant increase in the p53 protein level after MNNG treatment. A necessary role for p53 in APCgene regulation is supported by the failure of MNNG to induceAPC expression in cell lines either expressing very low levels of p53 (HeLa cells) or no p53 (K562 erythroleukemia cells). The overexpression of wild-type p53 gene into HCT-116 cells mimicked the effect of MNNG-induced expression of APCmRNA. A direct causal role for p53 in APC gene regulation was further evaluated by transfecting the wild-typep53 gene into K562 cells and observing a 5-fold increase in the APC gene expression. These results support a model featuring a direct link between p53 and APC in response to alkylation-induced DNA damage and suggest a novel role for p53 in a stress-response pathway involving APC.

Intestinal cells are constantly exposed to DNA-damaging agents from dietary toxins (1,2). The resulting DNA damage, if not efficiently repaired, may result in genomic instability, leading to malignant transformation. The development and progression of colon cancer is a multistep process in which growth control mechanisms are impaired progressively. Mutations of the adenomatous polyposis coli (APC) 1 tumor suppressor gene, the Ki-ras oncogene, the deleted in colorectal cancer (DCC) gene, the p53 gene, and DNA mismatch repair (MMR) genes play important roles at different stages of colorectal carcinogenesis (for review, see Ref. 3). Among these, mutation of the APC gene is an early event in familial adenomatous polyposis (4,5) and sporadic colorectal cancers (3,(5)(6)(7). Apart from colorectal cancers, mutations in the APC gene also are associated with malignant brain tumors (Turcot's syndrome (8)). Although APC is expressed constitutively within normal colonic epithelium, little is known about how mutations of, or abnormal expression of, APC contribute to the development of colon cancer. Previous studies have indicated that the cellular levels of wild-type APC are critical to cytoskeletal integrity (9,10), cellular adhesion (11), and Wingless/Wnt signaling (3,(12)(13)(14)(15)(16)(17). The phenotype of a cell expressing a mutated APC gene can be reverted by increased expression of the remaining wild-type APC allele and provide evidence that overexpression of the wild-type APC gene alone can suppress tumorigenicity (18,19). Thus, understanding the mechanisms by which APC gene expression can be induced at the molecular level is critical.
The tumor suppressor gene p53 is also frequently mutated in colon cancer cells (3). The wild-type p53 protein is necessary for monitoring the G 1 checkpoint, sensing DNA damage, assembling the DNA repair machinery, modulating gene amplification, or activating apoptosis to remove damaged cells (20). p53 activates transcription from specific DNA-binding sites and represses transcription in a binding site-independent manner. Treatment of cells with DNA-damaging agents induces nuclear accumulation of p53, which trans-activates cell cycle-and apoptosis-related genes (20). In previous studies, the transient expression of the wild-type APC gene into mammalian cells has been shown to cause cell cycle arrest (21) and apoptosis (22). However, whether APC expression in cells treated with DNAdamaging agents is increased and whether p53 plays any role in the expression of the APC are currently unknown. In the present investigation, we demonstrate that APC expression is, in fact, strongly induced after treatment of cells with DNAdamaging agents, such as the potent colon carcinogen Nmethyl-NЈ-nitro-N-nitrosoguanidine (MNNG) (for review, see Ref. 23). Furthermore, results indicated that the increased expression of the APC after MNNG treatment is dependent upon p53 expression.
Mammalian cells respond through a variety of signal transduction pathways after exposure to DNA-damaging agents, leading to specific programs of gene expression and alterations in the activity of proteins (24, 25) (for review, see Ref. 26). Among these effects, increased expression of tumor suppressor genes such as APC may be critical in controlling cellular transformation by delaying cell cycle progression or by inducing apoptosis. Our hypothesis is that exposure to DNA-damaging agents cause DNA damage, which eventually reach a threshold level at which DNA repair mechanisms fail, leading to the induction of apoptosis. We propose that APC, in collaboration with p53, is a critical component of cellular defense mechanisms involving DNA damage-induced cell cycle arrest and/or apoptosis.

EXPERIMENTAL PROCEDURES
Maintenance and Treatment of Cells-Human cancer cell lines HCT-116 (colon), HeLa S3 (cervical), and K562 (erythroleukemia) were grown in McCoy's 5a medium, Dulbecco's modified Eagle's medium (with high glucose), and RPMI medium, respectively, supplemented with 10% fetal bovine serum (FBS) and 100 g/ml penicillin and streptomycin (Life Technologies, Inc.). After cells reached to 70% confluence, fresh medium containing 0.5% FBS was added to each dish. Cells were further incubated for an additional 18 h and treated with DNA-alkylating agents for different periods.
Northern Blot Analysis-Total RNA was isolated by TRIzol TM reagent as described by the manufacturer (Life Technologies, Inc.). Fifty micrograms of total RNA were separated on 1% formaldehyde-agarose gel and transferred onto a Hybond-N ϩ membrane (Amersham Corp.). The membrane was prehybridized for 6 h at 65°C in 0.5 M sodium phosphate buffer (pH 7.2), 7% (w/v) SDS, 1 mM EDTA, and 1% (w/v) bovine serum albumin and then hybridized with 32 P-labeled APC probe (EcoRI fragment of APC-HFBCI43 obtained from ATCC). Later the same membrane was reprobed with the 32 P-labeled EcoRI fragment of the 18 S RNA probe for normalization of the total RNA loading and transfer efficiency. The membranes were exposed to x-ray films for detection of mRNA signals.
Nuclear Run-on Assay-For nuclear run-on assay, cells were washed with Buffer A (10 mM Tris-HCl, pH 7.5, 2 mM MgCl 2 , 3 mM CaCl 2 , 3 mM DTT) containing 10% (w/v) sucrose. Cells were lysed in Buffer A containing 0.1 mM EDTA and 0.5% (v/v) Nonidet P-40. The lysed cells were layered on the top of 68% (w/v) sucrose solution prepared in Buffer A. The nuclei were collected by centrifugation at 30,000 ϫ g for 45 min at 4°C. The nuclear pellet was resuspended in a storage Buffer B (42 mM Tris-HCl, pH 7.9, 9.3 mM MgCl 2 , 210 mM KCl, 4.2 mM DTT, and 42% (v/v) glycerol). A 100-l nuclear suspension containing 30 ϫ 10 6 nuclei was mixed with 100 l of transcript elongation mixture containing 250 M each of GTP, CTP, and ATP and 100 Ci of [␣-32 P]UTP (3000 Ci/mmol). Forty units of RNase inhibitor, RNasin (Promega), was added in each tube. The nascent transcripts were elongated for 30 min at 30°C. The 32 P-labeled RNA (2.5 ϫ 10 6 cpm) was hybridized to 500 ng of EcoRI fragments of APC and 18 S RNA plasmids and poly(dI⅐dC), which were immobilized on a Hybond-N ϩ membrane. The hybridization conditions were the same as described for Northern blot analysis.
The expected signals of Northern and Western blots of each experiment were quantified by personal densitometer SI (Molecular Dynamics). Data analysis and graphing were done either by Kaleidagraph (Synergy Software) or by Sigma Plot (Jandel Scientific).

Induction of Expression of APC by DNA-damaging
Agents-In the past few years, it has been clearly demonstrated that the tumor suppressor APC plays an important role in the development of colorectal cancer (3). However, the biological function(s) of APC is poorly understood. To gain insight into the biological function(s) of APC, it is critical to understand the molecular mechanisms of the APC gene expression. To examine the possibility of whether APC is inducible, an HCT-116 human colon cancer cell line, which expresses wildtype APC (18), was exposed to DNA-damaging agents such as MNNG (50 M MNNG Induces Transcriptional Regulation of APC in HCT-116 Cells-To investigate the mechanisms regulating transcription of APC, cells were treated simultaneously with MNNG and actinomycin D (ActD) or with cycloheximide (CHX) for 15 h ( Fig. 2A) to block transcription and new protein synthesis, respectively. Whereas treatment with ActD abolished MNNG's induction of APC mRNA expression, that treatment with CHX served to induce APC implicates a negative posttranscriptional regulatory mechanism involving a labile factor. To investigate the effect of MNNG treatment on the stability of the APC mRNA, the half-life (t1 ⁄2 ) of transcribed APC mRNA was determined in control versus MNNG-treated cells by treatment with ActD (1 g/ml) or buffer for different periods to block further transcription. The kinetics of APC mRNA degradation (t1 ⁄2 of 125-150 min) with or without MNNG indicated that the increased APC expression was not due to a post-transcriptional mechanism (Fig. 2B). Nuclear run-on assays showed that nascent APC mRNA synthesis was higher in nuclei isolated from

p53-dependent Expression of APC
MNNG-treated versus control cells, whereas the rate of transcription for 18 S RNA remained unchanged (Fig. 2C). Collectively, these data establish that transcriptional rather than post-transcriptional mechanism(s) are involved in the regulation of APC after DNA alkylation damage.
Endogenous p53 Expression Is Required for Induction of APC mRNA Levels by MNNG-The expression of the tumor suppressor p53 is induced by a wide variety of DNA-damaging agents and has been proposed to play a necessary role in controlling transcription of the critical genes involved in cell cycle arrest and DNA replication (20). Since APC is also involved in cell cycle arrest (21) and we have discovered that APC expression is transcriptionally induced after DNA damage, we next examined a role for p53 in APC expression. Thus, if p53 is required for DNA damage-induced expression of APC, then cell lines defective in wild-type functional p53 should fail to upregulate APC. Three cell lines with different p53 expression were selected: (i) the HCT-116 cell line expresses wild-type p53 (29); (ii) the HeLa S3 cell line expresses barely detectable levels of p53 protein (30); (iii) the erythroleukemic K562 cell line is negative for the p53 gene and fails to express p53 protein (31). Analysis of RNA from these cells treated with different concentrations of MNNG for 15 h (Fig. 3) indicated a tight correlation between the p53 status and the induction of steady-state levels of APC mRNA after MNNG treatment. The HCT-116 cell line showed a dose-dependent increase in the APC mRNA level (Fig. 3A), whereas HeLa cells showed only a slight (i.e. Ͻ2-fold) increase in expression at 25-50 M MNNG, and the basal level of APC expression in p53-negative K562 cells was unaffected by DNA damage (Fig. 3A). Since exposure of mammalian cells to DNA-damaging agents increases cellular levels of p53 protein (20), we examined whether the increase in APC mRNA levels in HCT-116 cells after MNNG treatment was correlated with a simultaneous increase in p53 protein levels. The p53 protein levels were increased significantly in HCT-116 cells after treatment with MNNG (50 M) but were not altered in HeLa cells (Fig. 3B). These results suggest that APC expression is sensitive to, and may require, increased levels of p53.

Overexpression of Wild-type p53 Can Mimic MNNG-induced Expression of APC mRNA in HCT-116 Cells-We observed that
HCT-116 cells express a low level of p53 as well as APC, and both were found to be increased after MNNG treatment (Fig.  3). These results are consistent with the idea that p53 levels are limiting in HCT-116 cells, and the increased levels of p53 may be necessary for induced expression of APC. If this is the case, then overexpression of p53 in these cells should mimic the effect of MNNG treatment on the expression of APC. To test this idea, HCT-116 cells were transfected with CMV-p53 plasmid, and the APC mRNA levels were determined by Northern blotting. As predicated, the APC mRNA levels of HCT-116 cells were increased in a dose-dependent manner after transfection with CMV-p53 plasmid (Fig. 4A). These results demonstrate that an increase in p53 protein level is required for APC gene expression after DNA damage.
Wild-type p53 Can Induce APC mRNA Expression in K562 Cells-To evaluate whether p53 can regulate APC in K562 cells (p53-negative), a CMV-p53 plasmid construct was transiently transfected into these cells, and the expression of APC mRNA was determined. An increased level of p53 protein was determined by Western blotting in K562 cells after CMV-p53 transfection (data not shown). The total RNA was isolated from these cells and processed for APC mRNA determination. The results show that APC mRNA levels were increased in a dosedependent manner in CMV-p53-transfected versus untransfected K562 cells (Fig. 4B), clearly indicating that p53 expression can regulate APC gene expression.
After DNA damage, increased levels of cellular p53 protein transcriptionally activate genes required for cell cycle arrest and apoptosis (20). For example, p53 up-regulates expression of the WAF1/cip1 gene, which is a known inhibitor of cyclin/Cdk function (32,33). The decreased cyclin/Cdk activity leads to hypophosphorylation of pRb, resulting in cell cycle arrest at the G 1 phase (20). In the present studies, WAF1/Cip1 protein levels remained unchanged after MNNG treatment of p53-expressing HCT-116 cells, whereas pRb levels significantly decreased (data not shown). These results suggest that p53-mediated MNNG-induced signaling does not involve WAF1/Cip1 or pRb. Since we observed a consistent requirement for p53 in APC up-regulation, it appears that a p53-mediated signaling pathway is required for APC regulation in response to DNA dam- p53-dependent Expression of APC age. Previous results indicated that wild-type APC-mediated cell cycle arrest (21) and apoptosis (22) were p53-independent. However, since those studies were performed with cells that overexpressed recombinant APC, it is possible that high levels of the tumor suppressor APC could bypass a requirement for p53 upstream of APC regulation. Indeed, our results indicate that p53 is an upstream effector required for induced expression of APC in response to DNA damage.
Recent studies have indicated that APC is an important component of the Wnt signaling pathway (3,14). In the absence of Wnt signals, APC interacts with and is phosphorylated by glycogen synthase kinase-3␤, which regulates cellular levels of ␤-catenin and inhibits signaling by the ␤-catenin-Tcf/Lef complexes (15,16). These studies have suggested that in the absence of wild-type APC, ␤-catenin-Tcf/Lef complexes are stabilized and translocated into the nucleus, where they may activate transcription of the target genes, which might be involved in the development of colorectal cancer (15,16) as well as in melanoma (17). Besides being a negative regulator of ␤-catenin signaling, APC has also been implicated in the migration of intestinal epithelial cells from the crypt up to the villi, where these cells undergo apoptosis. This physiological process is necessary to maintain the normal functioning of the intestinal cells (for review, see Ref. 28). Previous studies have suggested that free ␤-catenin blocks the APC-mediated cell migration. Thus increased levels of the wild-type APC and low levels of ␤-catenin can independently regulate gene expression by controlling the levels of ␤-catenin-Tcf/Lef complexes and similarly affect also the migration of epithelial cells. How DNA damage-induced APC levels play a role in these processes is currently unknown. Thus, our findings support the idea that DNA-damaging agents can induce transcription of the APC gene in a p53-dependent manner, which might perhaps inhibit ␤-catenin-Tcf/Lef complex signaling and also might play a role in cell migration and apoptosis. Nevertheless, the dependence of APC expression on p53 may be highly physiologically relevant in that a p53-APC connection may help to identify a functional role for the APC. Obviously, when the vital cell processes are dysfunctional, as in cells containing mutant APC and/or p53, then increased genetic instability results and may lead to the development of cancer (for reviews, see Refs. 3 and 34). Increased APC levels may overcome this process, either by cell cycle arrest and DNA damage repair or by inducing apoptosis. HCT-116 cells were transiently transfected with CMV-driven p53 expression plasmid by a calcium phosphate coprecipitation procedure. The total amount of DNA was adjusted to 3 g/ml with carrier DNA. After 24 h of transfection, cells were further incubated for 12 h in fresh medium containing 10% FBS. B, overexpression of the wild-type p53 in a K562 cell line induces endogenous APC mRNA levels. K562 cells were transiently transfected with CMV-driven p53 expression plasmid by electroporation (Bio-Rad). K562 cells (10 6 ) in 600 l of RPMI medium containing 10% FBS were added with increasing amounts of CMV-p53 and pulsed for 20 ms at 0.25 kV and 900 microfarads. For electroporation the total amount of DNA was adjusted to 20 g/ml with carrier DNA, which was diluted to 2 g/ml during incubation. After 48 h of incubation, a fresh medium containing 0.5% FBS was added and incubation continued for an additional 24 h. The endogenous APC mRNA levels of HCT-116 and K562 cells were determined by Northern blotting and normalized to 18 S RNA. The results were plotted in relative units as compared with untransfected cells. The amount of CMV-p53 DNA shown in each panel represents the amount used for transfecting cells. Data of A are representative of two independent experiments, and data of B are mean Ϯ S.D. of four experiments.