Clusterin-Mediated Apoptosis Is Regulated by Adenomatous Polyposis Coli and Is P21 Dependent but P53 Independent

Home , P21

[CANCER RESEARCH 64, 7412–7419, October 15, 2004] Clusterin-Mediated Apoptosis Is Regulated by Adenomatous Polyposis Coli and Is p21 Dependent but p53 Independent

Tingan Chen,1 Joel Turner,1 Susan McCarthy,1 Maurizio Scaltriti,2 Saverio Bettuzzi,2 and Timothy J. Yeatman1 1Department of Interdisciplinary Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, Florida; and 2Dipartimento di Medicina Sperimentale, Sezione di Biochimica, Biochimica Clinica e Biochimica dell’Esercizio Fisico, Parma, Italy

ABSTRACT ptosis (15). Additional data suggest that a secreted form of clusterin acts as a molecular chaperone, scavenging denatured proteins outside Clusterin is a widely expressed glycoprotein that has been paradoxi- cells following specific stress-induced injury such as heat shock. cally observed to have both pro- and antiapoptotic functions. Recent Other data, show that overexpression of a specific nuclear form of reports suggest this apparent dichotomy of function may be related to two different isoforms, one secreted and cytoplasmic, the other nuclear. To clusterin acts as a prodeath signal (16). Furthermore, studies of human clarify the functional role of clusterin in regulating apoptosis, we exam- colon cancer suggest a conversion from the nuclear form of clusterin ined its expression in human colon cancer tissues and in human colon to the cytoplasmic form, which may promote tumor progression (17). cancer cell lines. We additionally explored its expression and activity using Recently, a link between the accumulation of the nuclear form of models of adenomatous polyposis coli (APC)- and chemotherapy-induced clusterin and anoikis induction in prostate epithelial cells was shown apoptosis. Clusterin RNA and protein levels were decreased in colon (18). cancer tissues largely devoid of wild-type APC when compared with To better elucidate the functional role of clusterin, we examined its matched normal tissue controls, suggesting a means for invasive cancers to expression and localization in a panel of normal and neoplastic human avoid apoptosis. Conversely, induction of apoptosis by expression of wild- colon cancer tissues, as well as a panel of human colon cancer cell type APC or by treatment with chemotherapy led to increased clusterin lines. We then evaluated the effect of APC expression and chemo- RNA and protein levels localizing to apoptotic nuclei. We found that transient transfection of clusterin to colon cancer cell lines directly en- therapy, both known to induce apoptosis, on clusterin expression and hanced basal and chemotherapy-induced apoptosis. Clusterin-induced associated apoptotic events. The effect of clusterin transfection, or apoptosis was inhibited by antisense clusterin and was found to be highly alternatively, antisense clusterin transfection, on basal and chemother- dependent on p21 but not p53 expression, yet a deficit in p21 can be apy-induced apoptosis was investigated. Because both p21 and p53 subverted by clusterin transfection. Collectively, these data support the have been lined to regulation of apoptotic events, clusterin expression hypothesis that nuclear clusterin function is proapoptotic when induced and induction of apoptosis was evaluated in the context of their by APC or chemotherapy in the context of p21 expression. Absent of p21, expression using targeted knockout cell line models. clusterin in not induced, and apoptosis is significantly inhibited. These data support a potential therapeutic role for clusterin in enhancing chemotherapy-induced apoptosis and in promoting apoptosis in cells deficient MATERIALS AND METHODS in p21. Human Tissues. Snap-frozen human colon cancers (Dukes’ stage C) and INTRODUCTION their cognate normal mucosal controls were obtained from the Moffitt Cancer Center tissue bank in a de-identified fashion with an Institution Review Clusterin is a widely expressed glycoprotein whose function re- Board-approved protocol. Tissues were microdissected for purity and then mains somewhat enigmatic. Also known as apolipoprotein J, clusterin exposed to cellular lysis buffers and Trizol for RNA extraction. was initially characterized as an apoptosis-associated transcript after it Cell Lines. Four different human colon cancer cell lines were obtained was identified as testosterone-repressed prostate message 2 (1) and from the American Type Culture Collection (HT29, KM12C, HCT116, and sulfated glycoprotein 2 (2) that is overexpressed in the epithelial cells SW620), and two targeted knockout cell lines were a kind gift of Dr. B. Ϫ Ϫ Ϫ Ϫ of the regressing rat ventral prostate (3). Increases in clusterin mRNA Vogelstein (HCT116p21 / and HCT116p53 / ). The HT29 cell line devoid of APC but engineered with a zinc-inducible wild-type APC vector and protein levels have since been consistently detected in apoptotic (HT29APC) was also a gift of Dr. Bert Vogelstein. HT29, KM12C, HCT116, heart, brain, lung, liver, kidney, pancreas, and retinal tissue both in and SW620 cells were cultured in RPMI medium with 10% fetal bovine serum. vivo and in vitro (1, 4–11) Moreover, clusterin-knockout mice show HT29APC, HCT116p21Ϫ/Ϫ, HCT116p53Ϫ/Ϫ, and their parental cells were reductions of cell death in hypoxia-ischemia–induced brain damage cultured in McCoy’s 5A medium with 10% fetal bovine serum. Cells were (12–14). Clusterin function, however, has also been reported to be treated with Zn2ϩ (120 ␮mol/L) 4, 8, and 16 hours or with 5-flurouracil (5-FU; antiapoptotic and has been found to be overexpressed in progressing 5 ␮mol/L) and anti-Fas antibody (50 ng/mL) or with irinotecan (6 ␮mol/L) 24, breast carcinomas, where its expression correlated inversely with the 48, and 72 hours to induce apoptosis. apoptotic index. Similarly, clusterin-knockout mice show an increase in Transient Transfection of Clusterin. The plasmids p4 empty vector and autoimmune myocardial damage, suggesting an antiapoptotic function. p4/SGP2 (containing human clusterin full-length cDNA) were obtained from Recent studies have shed some light on the dichotomous functions Dr. Saverio Bettuzzi (Parma, Italy). Transient transfection of p4 empty vector (as a mock) or p4/SGP2 were performed using FuGENE 6 Transfection of clusterin, suggesting pro- and antiapoptotic functions may be Reagent (Boehringer Mannheim, Mannheim, Germany) according to the man- related to nuclear and secreted isoforms, respectively, and that the ufacturer’s protocol. After transfection with mock or clusterin 24 to 48 hours, function of clusterin may be context dependent. Overexpression of the cells were either harvested or treated by apoptotic stimuli for 48 hours, then secreted form of clusterin in human cancer cells caused drug resist- harvested to detect clusterin protein expression by Western blotting or to ance and protection against certain cytotoxic agents that induce apo- perform apoptosis assays. Each transfection experiment was done in triplicate. Data are presented as a mean Ϯ SE. Received 6/11/04; revised 8/4/04; accepted 8/19/04. Western Blot Analyses. Cells were harvested in protein lysis buffer [20 The costs of publication of this article were defrayed in part by the payment of page mmol/L Tris-HCl (pH 7.6), 150 mmol/L NaCl, 1 mmol/L EDTA, 0.5% NP40, charges. This article must therefore be hereby marked advertisement in accordance with 1 mmol/L DTT, 5 ␮mol/L trichostatin A, 1 mmol/L sodium orthovanadate, 1 18 U.S.C. Section 1734 solely to indicate this fact. mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L NaF, and complete protease Requests for reprints: Timothy J. Yeatman, 12902 Magnolia Drive, SRB-2, Tampa, FL 33612-9497. Phone: (813) 979-7292; Fax: (813) 632-1433; E-mail: [email protected]. inhibitors (Roche Applied Science, Indianapolis, IN)], and Western blotting ©2004 American Association for Cancer Research. was performed as described previously (19). All blots were performed in 7412

triplicate. Representative examples are shown in the figures. Antibodies used labeling (TUNEL) staining with the In Situ Cell Death Detection Kit (TMR include clusterin (Alexis, San Diego, CA) and p53 and p21 (BD Biosciences, red; Roche Applied Science) according to the manufacturer’s protocol and San Jose, CA). immunoblotting with human clusterin antibodies (Alexis). Antisense Clusterin Oligonucleotide Transfection. Preattached ϳ1 ϫ 105 Immunocytofluorescence and TUNEL Assay. Preattached ϳ1 ϫ 104 HT29, HCT11,6 and HT29APC cells in 100-mm culture dishes, respectively, cells/well were exposed to treatment conditions as planned (some cells were for overnight. They were transfected with antisense clusterin oligonucleotides treated 48 hours by apoptotic stimuli, some cells were transiently transfected or mismatch clusterin oligonucleotides, respectively. Phosphorothioate oligo- with clusterin 24 hours, and then treated another 48 hours with apoptotic nucleotides used in this study were purchased from Integrated DNA Technol- stimuli). Cells were fixed with acetone-methanol (1:1; Sigma, St. Louis, MO) ogies, Inc. (Coralville, IA). The sequence of the clusterin antisense clusterin for 3 minutes, then were blocked with PBS ϩ10% normal goat serum for 20 oligonucleotide used corresponded to the human clusterin translation initiation minutes. Cells were incubated with clusterin antibody (1:200; Alexis) in PBS site (5Ј-CAGCAGCAGAGTCTTCATCAT-3Ј). A 2-base clusterin mismatch ϩ10% normal goat serum for 2 hours, then the slides were washed with PBS oligonucleotide (5Ј-CAGCAGCAGAGTATTTATCAT-3Ј) was used as con- ϩ0.1%Triton X-100 three times and were incubated with FITC (1:100; Vector trol. Transfections were performed using the Lipofectamine Plus reagent Laboratories, Inc., Burlingame, CA) in PBS ϩ10% normal goat serum for 1 (Invitrogen, Carlsbad, CA). Cells were treated with either 500 and 1000 hour and then washed again with PBS ϩ0.1%Triton X-100. Some slides were nmol/L antisense or mismatch clusterin oligonucleotides after preincubation dried and covered with coverslips in Vectashield mounting media of antifade/ for 20 minutes with 10 ␮g/mL lipofectin in serum-free OPTI-MEM (Life 4Ј,6-diamidino-2-phenylindole (1:1; Vector Laboratories, Inc.). Immunofluo- Technologies, Inc., Carlsbad, CA). Four hours after the beginning of the rescence was observed with a Leitz Orthoplan 2 microscope, and images were incubation, for HT29 and HCT116 cells, the medium containing oligonucleo- captured by a charge-coupled device camera with Smart Capture program tides and lipofectin was replaced with standard culture medium plus 5-FU (5 (Vysis, Downers Grove, IL). Some slides were double staining by TUNEL ␮mol/L) and Fas (50 ng/mL) once daily for 3 days. For HT29APC cells, the assay with the In Situ Cell Death Detection Kit (TMR red; Roche Applied medium was replaced with standard culture medium plus Zn2ϩ (120 ␮mol/L) Science) according to the manufacturer’s protocol to record clusterin expres- for 16 hours. Cells were harvested to detect clusterin protein expression by sion and apoptotic cell death co-localization. Western blotting or to perform apoptosis assays. Each transfection experiment was done in triplicate and repeated at least three times. Data are presented as RESULTS a mean Ϯ SE. cDNA Synthesis and Real-time PCR. To measure clusterin mRNA ex- Clusterin Is Down-regulated in Human Colon Cancer. To fur- pression, RNA was isolated from snap-frozen human colon cancers (Dukes’ ther explore the potential role of clusterin in colon cancer, the levels stage C), and their cognate normal mucosal controls or colon cancer cells of clusterin mRNA and protein were assessed in human colon neo- treated with different reagents by using TRIzol reagent (Invitrogen) according plastic specimens (Dukes’ stage C) relative to matched normal mu- to the manufacturer’s protocol. A quantitative primer/probe set was designed cosal controls from 10 patients. Compared with the matched normal to evaluate and to quantitate clusterin mRNA. Reverse transcription of RNA was performed using 15 units of Omniscript reverse transcriptase (Qiagen, tissues, clusterin mRNA and protein expression were generally down- Valencia, CA), 1ϫ cDNA synthesis buffer, 40 units of RNase inhibitor (Life regulated in human colon cancers (Fig. 1, A and B). Both RNA and Technologies, Inc.), 5 mmol/L DTT, 1 mmol/L deoxynucleoside triphosphate protein expression were consistently reduced by Ͼ50% in the major- mixture, 0.5 ␮g of Oligo(dt) primer, and 2 ␮g of RNA. Primers and probes for ity of cases. real-time PCR were designed using Primer Express software (Applied Biosys- Clusterin Is Up-regulated in Chemotherapy-induced Colon tems, Inc., Foster City, CA). Each primer set consisted of standard PCR Cancer Cell Apoptosis. To determine whether clusterin expression primers (temperature 58° to 60°C) designed to span gene exons to exclude any was implicated with colon cancer cell apoptosis, we used two com- possible genomic DNA contamination. Detection and quantitation of each gene mon chemotherapeutical agents for colon cancer, 5-FU (5 ␮mol/L) was accomplished using an amplicon-specific fluorescent oligonucleotide and irinotecan (6 ␮mol/L) for proapoptotic stimuli. Anti-Fas antibody Ј probe (temperature 68°Cto70°C), with a 5 -reporter dye (carboxyfluorescein) (50 ng/mL) was added to 5-FU because we have previously shown and a downstream 3Ј-quencher dye (6-carboxytetramethylrhodamine). The primers used for detection of clusterin had the following sequences, 5Ј- CTATCTGCGGGTCACCAC-3Ј (forward primer), 5Ј-CACAGTGACACCG- GAAGGAAC-3Ј (reverse primer), and 5Ј-FAM/TGGCTTCCCACACTTCT- GACTCGG A/TAMRA-3Ј (probe). During PCR, the 5Ј-3Ј nuclease activity of TaqDNA polymerase releases the reporter, whose fluorescent signal is then detected by the GeneAmp 5700 Sequence Detector System (ABI, Foster City, CA). The PCR reaction was performed in a 96-well optical reaction plate with optical caps. The reaction mixture consisted of 0.2 ␮mol/L each primer, 0.1 ␮mol/L fluorescent probe, 5 units of AmpliTaq Gold (ABI), 200 ␮mol/L each ␮ dATP, dCTP, and dGTP, 400 mol/L dUTP, 5.5 mmol/L MgCl2, 8% glycerol, 1 unit of AmpErase uracil N-glycosylase and 1ϫ TaqMan Buffer A. Ampli- fication profile was as follows: 50°C for 2 minutes (for uracil N-glycosylase digest of contaminating PCR amplicons), 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds, and 60°C for 1 minute, 2 ␮L of cDNA were assayed per well. Standards used for quantitation were plasmid vector DNA containing the primer-specific coding region for clusterin. Standard values were 106,105,104,103,102, and 10 DNA molecules per well. Two assays were performed on each cDNA sample, clusterin, and 3-histone H3.3 as a house- keeping gene to normalize mRNA values. Quantitative PCR was performed in an ABI 5700 sequence detection system. Apoptosis Assays. To determine apoptotic fractions, cells were harvested at different time point after treatment by apoptotic stimuli. Cells were then Fig. 1. Clusterin is down-regulated in human colon cancer. A, real-time quantitative analyzed by using the Annexin V-phycoerythrin plus 7-amino-actinomycin D PCR assay of clusterin mRNA expression in human colon neoplastic specimens (Dukes’ (BD Biosciences) apoptosis detection kit according to the manufacturer’s stage C) relative to matched normal mucosal controls from 10 patients. B, Western blot protocol. The DNA damage experiments were done in triplicate for each cell analyses for clusterin protein expression in the same patient samples. Compared with the matched normal tissues, clusterin mRNA and protein expression were generally down- line. Apoptotic fractions were determined by flow cytometry. Apoptosis was regulated in human colon cancers. Real-time PCR data were normalized by mRNA level confirmed by terminal deoxynucleotidyltransferase-mediated dUTP nick end to glyceraldehyde-3-phosphate dehydrogenase. 7413

Fig. 3. Clusterin is up-regulated in wild-type APC-induced colon cancer apoptosis. A null APCϪ/Ϫ human colon cancer cell line (HT-29) stably transfected with a Zn2ϩ- inducible wild-type APC vector or a ␤-galactosidase control was used to investigate the relationship between clusterin expression and APC. A. As demonstrated by real-time quantitative PCR, the clusterin mRNA levels dramatically increased in cells, as early as 4 h after induction of wild-type APC expression. (B) Similarly, protein levels of clusterin were shown to be up-regulated with induction of wild-type APC expression early after APC induction.

that the combination enhances apoptotic responses in colon cancer cell lines. After treatment, HT29, HCT116, KM12C, but not SW620 had significant increases in apoptosis (Fig. 2A). Apoptosis was assessed by flow cytometric analysis using Annexin V-phycoerythrin. Quantitative PCR and Western blotting were then used to measure clusterin mRNA and protein expression. Clusterin protein expression at baseline was undetectable for all four cell lines and dramatically increased with chemotherapy exposure in all cell lines but SW620 (Fig. 2B). Fluorescence immunostaining and TUNEL staining were used to localize clusterin expression and apoptotic cell death. Clus- terin was found to be primarily expressed in the nuclei in the TUNEL- positive stained apoptotic cells after chemotherapy treatment. The results indicated that the fraction of apoptotic cells was increased ϳ5-fold in those cells exposed to apoptotic stimuli relative to controls absent of apoptotic stimuli in HT29 (Fig. 2C) and HCT116 (Fig. 2D), but again, no apoptosis was induced at SW620 colon cancer cell line (data not shown). Clusterin Is Up-regulated in Wild-type APC-Induced Colon Cancer Apoptosis. Because most colon cancers lack functional APC protein and the data suggest diminished clusterin expression in these samples, we turned to a well-described in vitro cell line model in which wild-type APC protein can be induced to cause apoptosis. A null APCϪ/Ϫ human colon cancer cell line (HT-29) stably transfected with a Zn2ϩ- inducible wild-type APC vector or a ␤-galactosidase control, was used to additionally investigate the relationship between clusterin expression and APC. As demonstrated by real-time quantitative PCR, the clusterin mRNA levels dramatically increased in cells as early as 4 hours after

P Ͻ 0.01 versusuntreated control. HT29, HCT116, KM12C, but not SW620 had significant increases in apoptosis with treatment. B, real-time quantitative PCR results showed that clusterin mRNA was up-regulated for all cell lines except SW620, compared with untreated controls. A Western blot analysis showed that clusterin protein expression at Fig. 2. Clusterin is up-regulated in chemotherapy-induced colon cancer cell apoptosis. baseline was undetectable for all four cell lines and dramatically increased with chemo- A. To determine whether clusterin expression was implicated with colon cancer cell therapy exposure in all cell lines but SW620. C and D, dual-staining immunofluorescence apoptosis, we used 5-FU (5 ␮mol/L) plus anti-Fas antibody (50 ng/mL) as proapoptotic assay and TUNEL staining stained clusterin antigen (green), apoptotic cell death (red), stimuli. After treatment for 48 hours, apoptotic cell death was measured by an FACS flow and nuclei (blue). Clusterin was found to be primarily expressed in the nuclei in the cytometer using Annexin V-phycoerythrin plus 7-amino-actinomycin D and analyzed by TUNEL-positive stained apoptotic cells after chemotherapy treatment in HT29 and .Student t test, HCT116. DAPI, 4Ј,6-diamidino-2-phenylindoleء .(ModFitLT software. Each value represents the mean Ϯ SE (n ϭ 3 7414

Fig. 4. Antisense clusterin oligonucleotide (ASO) reduced chemotherapy- or APC-induced apoptosis. A and B. Real-time PCR data showed that clusterin ASO (1 mmol/L), but not mismatch (MM) control oligonucleo- P Ͻ 0.01ء) tides (1 mmol/L), significantly decreased clusterin mRNA ASO versus MM control oligonucleotides), and protein expression induced by proapoptotic stimuli. C. Similarly, apoptotic cell death induced by proapoptotic stimuli was decreased by ASO Ͼ50%, compared with MM control oligonucleotides in HT29, HCT116, and HT29 APC cells.

induction of wild-type APC expression (Fig. 3A). Similarly, protein levels (1 mmol/L), but not mismatch control oligonucleotides, significantly of clusterin were shown to be up-regulated with induction of wild-type decreased clusterin mRNA and protein expression induced by pro- APC expression early after APC induction (Fig. 3B). This induction of apoptotic stimuli (Fig. 4, A and B). Similarly, apoptotic cell death clusterin and APC tightly correlate with the induction of apoptosis in induced by proapoptosis stimuli was decreased by ASO Ͼ50%, com- these cells (data not shown). pared with mismatch control oligonucleotide (Fig. 4C). These results Antisense Clusterin Oligonucleotide Reduced Chemotherapy- indicated that chemotherapy- or APC-induced colon cancer cell apo- or APC-Induced Apoptosis. To understand if clusterin expression ptosis is dependent on clusterin expression. directly linked to chemotherapy- or APC-induced apoptosis, we used Transient Transfection of Clusterin Localized to the Nucleus previously described (20) antisense clusterin oligonucleotides to test and Directly Induced Apoptosis or Enhanced Chemotherapy- whether blocking clusterin expression by antisense clusterin oligonu- Induced Apoptosis in Colon Cancer Cells. Given the correlation cleotide could reduce apoptosis induced by chemotherapy or APC. between clusterin overexpression and cell death, we investigated The results showed that clusterin antisense clusterin oligonucleotides whether clusterin directly contributed to colon cancer cell death. We 7415

fold in cells treated with 5-FU and Fas as compared with mock- transfected cells treated with 5-FU and Fas (Fig. 5D). These results indicated that transient transfection of clusterin to colon cancer cells either directly induced apoptosis or enhanced chemotherapy sensitivity in human colon cancer cells. Clusterin Induction in Chemotherapy or APC-Induced Colon Cancer Cell Apoptosis Was p21 Dependent. As a broad-acting cyclin-dependent kinase inhibitor, p21 (WAF1) occupies a central position in the cell cycle regulation of self-renewing tissues. In addition to regulating normal cell cycle progression decisions, p21 inte- grates genotoxic insults into growth arrest and apoptotic signaling pathways that ultimately determine epithelial cell fate. We have found that chemotherapeutical agents or APC were able to substantially up-regulate p21 expression (Fig. 6A). Because we knew that these treatments also induced clusterin and apoptosis, we investigated clusterin expression in wild-type HCT116p21ϩ/ϩ and knockout HCT116 p21Ϫ/Ϫ cells. When HCT116 p21ϩ/ϩ cells were treated with proapoptotic stimuli (5-FU or irinotecan), both clusterin and p21expression were detected. Absent of p21 expression, however, clusterin was not expressed. Similarly, HCT116p21Ϫ/Ϫ cells showed significant resistance to 5-FU and Fas-induced apoptosis as compared with the parental cells (HCT116p21ϩ/ϩ; Fig. 6B). These results demonstrated that without p21 expression, clusterin was not induced and apoptosis was inhibited. When clusterin was transiently transfected into HCT116 p21Ϫ/Ϫ cells, however, clusterin protein was overexpressed and colocalized in the nucleus with apoptotic cells. The percentage of apoptotic cells increased ϳ6-fold as compared with control cells. Clusterin expression effectively bypassed the requirement for p21 expression, inducing basal levels of apoptosis, and enhancing chemotherapy-induced apoptosis (Fig. 6C). Using the colon cancer cell line SW620, we previously showed its resistance to chemotherapy-induced apoptosis (Fig. 2A) and its absence of clusterin expression (Fig. 2B). Here, we show that the SW620 cell line has no detectable p21 expression nor clusterin expression either with or without chemotherapy treatment (Fig. 6D), potentially explaining why this cell line is resistant to chemotherapy induced apoptosis. Taken together, our results suggest that clusterin acts downstream of p21 and that clusterin overexpression in chemotherapy-induced apoptosis is p21 dependent. Clusterin and p21 Induction in Chemotherapy-Induced Colon Cancer Cell Apoptosis Are p53 Independent. The tumor suppresser gene p53 is the primary transcription factor for the induction of p21. To determine whether p53 expression is implicated to p21-clusterin signal transduction pathway, we evaluated p53 expression in HT29 Fig. 5. Transient transfection of clusterin localized to the nucleus and directly induced (p53 mutant) and HCT116 (p53 wild-type) colon cancer cell lines. We apoptosis or enhanced chemotherapy-induced apoptosis in colon cancer cells. A. Western found that p21 and clusterin were overexpressed, but p53 levels were blot analysis showed that transiently transfection clusterin to the wild-type HT29 cell line, and after 24 or 48 hours, clusterin protein expression increased ϳ10-fold compared with unaffected in cells after chemotherapeutic treatment (Fig. 7A). We Ϫ Ϫ mock-transfected or wild-type cells. B, fluorescence immunostaining and TUNEL staining also detected p21 and clusterin expression in HCT116 p53 / cells to detect clusterin expression and cell apoptosis principally in nucleus of the transfected after treatment with chemotherapy (Fig. 7B). The results suggest that cells. Clusterin antigen (green), apoptosis cell death (red), and nuclei (blue). C. Flow cytometric analysis showed that transient transfection of clusterin induced apoptosis in clusterin and p21 up-regulation by chemotherapeutic treatment is Ͼ30% cells, as compared with the mock-transfected or wild-type cells, and (D) transiently independent of p53 expression. transfected clusterin enhanced apoptosis by ϳ2-fold in cells treated with 5-FUand Fas, as P Ͻ 0.01 clusterinء .compared with mock-transfected cells treated with 5-FU and Fas transfection versus mock transfection. DISCUSSION Although clusterin was originally proposed to be a marker for transiently transfected clusterin to the wild-type HT29 cell line. After programmed cell death, it has variously been proposed to provide both 24 or 48 hours, clusterin protein expression increased ϳ10-fold com- pro- and antiapoptotic functions. Studies of clusterin-transfected renal pared with mock-transfected or wild-type cells (Fig. 5A). We used carcinoma cells suggest that clusterin expression may increase cell fluorescence immunostaining and TUNEL staining to detect clusterin survival, motility, invasive capacity, and contribute to metastasis (14). expression and cell apoptosis principally in nucleus of the transfected In ApcMin models of colon neoplasia, tumor cells undergoing apopto- cells (Fig. 5B). Flow cytometric analysis also showed that transient sis expressed low levels of clusterin (21). In cancer patients, overex- transfection of clusterin induced apoptosis in Ͼ30% cells, as com- pression of clusterin has been linked to tumor progression in both pared with the mock transfected or wild-type cells (Fig. 5C). We also breast and colon cancer (17, 22). These data support an antiapoptotic found that transiently transfected clusterin enhanced apoptosis ϳ2- role for clusterin. Conversely, a number of studies also support a 7416

Fig. 6. Clusterin induction in chemotherapy- or APC- induced colon cancer cell apoptosis was p21 dependent. A. Western blot analysis showed that chemotherapeutic agents or APC were able to substantially up-regulate p21 protein expression. B. Western blot analysis showed that when HCT116 p21ϩ/ϩ cells were treated with proapoptotic stimuli (5-FU or irinotecan), both clusterin and p21expression were detected. Absent of p21 expression, however, clusterin was not expressed in HCT116 p21Ϫ/Ϫ cells. Flow cytometric data showed that HCT116p21Ϫ/Ϫ cells showed significant resistance to 5-FU and Fas-induced apoptosis, as compared with the parental cells (HCT116p21ϩ/ϩ). C. Western blot analysis showed that with transient transfection of clusterin into HCT116 p21Ϫ/Ϫ cells, clusterin protein was overexpressed. Fluorescence immunostaining and TUNEL staining showed that the overexpression of clusterin was colocalized in the nucleus with apoptotic cells. Flow cytometric data showed that the percentage of apoptotic cells increased ϳ6-fold as compared with control cells. D. Western blot analysis showed that the SW620 cell line had no detectable p21 expression nor clusterin expression either with or without chemotherapy treatment.

proapoptotic role for this molecule. Previous studies suggest that Our investigations sought to clarify the functional role of clusterin clusterin expression decreases the proliferation rate of prostate epi- in human colon cancer. Only recently has it come to light that two thelial cells (23) and is reduced in human prostate cancer (24), forms of clusterin, a secreted and a nuclear form, may be responsible consistent with a proapoptotic function. In addition, knockout studies for this divergent behavior. It is now apparent that the endogenous show reductions in cell death under stress conditions (14, 25). nuclear clusterin is a product of alternative splicing. The shorter 7417

examine this question, we evaluated human colon cancer cells with a targeted deletion of p21 (HCT116p21Ϫ/Ϫ) and found that when treated with proapoptotic stimuli, no clusterin overexpression was seen and significantly less apoptosis was induced. Interestingly, the only cell line that was not susceptible to chemotherapy-induced apoptosis, SW620, was also incapable of mounting a p21 response, presumably preventing clusterin expression. Furthermore, when clusterin was transiently transfected into HCT116 p21Ϫ/Ϫ cells, clusterin protein was overexpressed in the nucleus and colocalized with TUNEL staining in the apoptotic cells. The percentage of apoptotic cells increased ϳ6-fold as compared with control cells, suggesting that clusterin can effectively bypass the action of p21 and induce apoptosis in the absence of p21. We also found that p21 and clusterin Fig. 7. Clusterin and p21 induction in chemotherapy-induced colon cancer cell apo- were overexpressed, but p53 levels were unaffected in cells after ptosis are p53 independent. A. Western blot analysis showed that p21 and clusterin were chemotherapeutic treatment. Similarly, p21 and clusterin were over- overexpressed, but p53 levels were unaffected in HT29 (p53 mutant) and HCT116 (p53 expressed in HCT116 p53Ϫ/Ϫcells after stimulation by chemotherapy wild-type) cells after chemotherapeutic treatment. B. Western blot analysis also showed that p21 and clusterin were overexpressed in HCT116 p53Ϫ/Ϫ cells after chemotherapeu- (Fig. 6, I and J). Taken together, these results suggest that clusterin tic treatment. acts downstream of p21 and that clusterin overexpression in chemotherapy-induced apoptosis was p21 dependent and p53 independent. This work is the first to demonstrate a novel link between the nuclear mRNA coding for nuclear clusterin produces a Mr 49,000 precursor form of clusterin overexpression and the induction of p21. protein that acts as a prodeath signal, inhibiting cell survival and cell The p21 signaling pathway that regulates the nuclear form of growth. It contains a nuclear localization signal with potential for clusterin expression after chemotherapy exposure is unknown. Induc- targeting it to the nucleus where it can become activated as a mature tion of the cyclin-dependent kinase inhibitor p21 is a common mech- ϳ Mr 55,000 death protein under conditions of significant stress (16). anism of growth arrest in different physiologic situations. Ectopic Previous studies of colon cancer found an overexpression of secretory overexpression of p21 leads to cell growth arrest in G1 and G2, clusterin in the cytoplasm but a concordant decrease in nuclear clus- possibly by inhibition of cyclinB1/cyclin-dependent kinase 1 as dem- terin (17). Although our data show a decrease in clusterin in cancers onstrated in prostate cancer cells. This arrest is accompanied by relative to normal mucosal controls, they are consistent with these apoptotic cell death. Although p21 is not a transcription factor, it is results in that we identified a significant decrease in clusterin expres- conceivable that indirect effects of p21 on cellular gene expression sion, using antibodies recognizing the nuclear isoform of clusterin, may mediate some of its functions. For example, transient transfection when comparing Dukes’ stage C colon cancers to their matched assays showed that p21 could stimulate nuclear factor-␬B–mediated normal tissues. This decrease in expression linked to tumor progres- transcription; this effect of p21 has been explained through the inter- sion is also consistent with the role of clusterin as a tumor suppressor action of cyclin-dependent kinase 2 with transcriptional cofactor p300 gene, normally providing proapoptotic function. Taken together, these that augments nuclear factor-␬B and other inducible transcription results suggest that the discrepancies in previous studies of human factors. p21 interactions with proteins other than CDK may also have cancer clusterin expression are likely related to the type of clusterin a potential effect on gene expression. For example, p21 was reported measured. to bind c-Jun NH2-terminal kinases, apoptosis signal-regulating ki- We sought to additionally investigate the relationship between nase 1, and Gadd45 (27). It was already shown that clusterin promoter clusterin expression and apoptotic function by examining clusterin region has AP-1 protein (c-Jun and c-Fos) binding sites. In a study by expression in the context of APC expression and chemotherapy ex- Jin et al. (28), transforming growth factor-␤ induced clusterin expres- posure. Because APC is mutated in the vast majority (Ͼ95%) of sion in a variety of cell types via a consensus AP-1 binding site. sporadic colon cancers and APC expression has been linked to the Collectively, the presented data demonstrate that overexpression of induction of apoptotic events (19), we investigated the role of clus- the nuclear form of clusterin either enhanced chemotherapy sensitivity terin in APC-induced apoptosis. We identified that apoptosis induced or directly induced colon cancer apoptosis. The induction of clusterin by APC in APC-null colon cancer cells or by chemotherapy was by chemotherapy was p21 dependent but p53 independent. Because linked to significant increases in clusterin expression. These data clusterin transfection subverted the absence of p21 and caused apo- suggested that it was the nuclear form of clusterin (proapoptotic form) ptosis, clusterin protein appears to be an important protein for regu- that contributes to human colon cancer apoptosis by colocalizing lating chemotherapy-induced apoptosis and may have a potential clusterin to apoptotic nuclei. In addition, we found that the stress therapeutic role. induced by chemotherapies readily elevated clusterin levels and pro- duced apoptosis. Because clusterin has been proposed to regulate the cell cycle, REFERENCES increasing the accumulation of cells in the G0-G1 phases of the cell 1. Leger JG, Montpetit ML, Tenniswood MP. Characterization and cloning of androgen- cycle (14), we elected to evaluate the potential relationship between repressed mRNAs from rat ventral prostate. Biochem Biophys Res Commun 1987; p21 and p53 and clusterin expression. The finding that targeted 147:196–203. 2. Bettuzzi S, Hiipakka RA, Gilna P, Liao ST. Identification of an androgen-repressed inactivation of p21 enhances Apc-initiated tumor formation (26) also mRNA in rat ventral prostate as coding for sulphated glycoprotein 2 by cDNA cloning led us to study this relationship further. Our data suggest that clusterin and sequence analysis. Biochem J 1989;257:293–6. is induced in cells undergoing apoptosis stimulated by APC expres- 3. Montpetit ML, Lawless KR, Tenniswood M. Androgen-repressed messages in the rat ventral prostate. Prostate 1986;8:25–36. sion or by chemotherapy exposure. Moreover, we found a coordinate 4. Wong P, Borst DE, Farber D, et al. Increased TRPM-2/clusterin mRNA levels during up-regulation of p21 and clusterin expression in cells undergoing the time of retinal degeneration in mouse models of retinitis pigmentosa. Biochem Cell Biol 1994;72:439–46. apoptosis in response to these stimuli. The question then arose as to 5. Adel Moallem S, Hales BF. Transglutaminase and clusterin induction during normal whether p21 expression was necessary for clusterin expression. To and abnormal limb development in the mouse. Biol Reprod 1996;55:281–90. 7418

6. Bursch W, Gleeson T, Kleine L, Tenniswood M. Expression of clusterin (testoster- 18. Caccamo AE, Scaltriti M, Caporali A, et al. Cell detachment and apoptosis induction one-repressed prostate message-2) mRNA during growth and regeneration of rat liver. of immortalized human prostate epithelial cells are associated with early accumula- Arch Toxicol 1995;69:253–8. tion of 45 kDa nuclear isoform of clusterin. Biochem J 2004;382 (Pt 1):157–68. 7. Buttyan R, Olsson CA, Pintar J, et al. Induction of the TRPM-2 gene in cells 19. Chen T, Yang I, Irby R, et al. Regulation of caspase expression and apoptosis by undergoing programmed death. Mol Cell Biol 1989;9:3473–81. adenomatous polyposis coli. Cancer Res 2003;63:4368–74. 8. Danik M, Chabot JG, Mercier C, et al. Human gliomas and epileptic foci express high 20. Zellweger T, Chi K, Miyake H, et al. Enhanced radiation sensitivity in prostate cancer levels of a mRNA related to rat testicular sulfated glycoprotein 2, a purported marker by inhibition of the cell survival protein clusterin. Clin Cancer Res 2002;8:3276–84. of cell death. Proc Natl Acad Sci USA 1991;88:8577–81. 21. Chen X, Halberg RB, Ehrhardt WM, Torrealba J, Dove WF. Clusterin as a biomarker 9. Leger JG, Le Guellec R, Tenniswood MP. Treatment with antiandrogens induces an in murine and human intestinal neoplasia. Proc Natl Acad Sci USA 2003;100: androgen-repressed gene in the rat ventral prostate. Prostate 1988;13:131–42. 9530–5. 10. Tenniswood MP, Guenette RS, Lakins J, Mooibroek M, Wong P, Welsh JE. Active 22. Redondo M, Villar E, Torres-Munoz J, Tellez T, Morell M, Petito CK. Overexpres- cell death in hormone-dependent tissues. Cancer Metastasis Rev 1992;11:197–220. sion of clusterin in human breast carcinoma. Am J Pathol 2000;157:393–9. 11. Ahuja HS, Tenniswood M, Zakeri ZF. Differential expression of clusterin in the testis 23. Bettuzzi S, Scorcioni F, Astancolle S, Davalli P, Scaltriti M, Corti A. Clusterin (SGP-2) transient overexpression decreases proliferation rate of SV40-immortalized and epididymis of postnatal and germ cell deficient mice. J Androl 1996;17:491–501. human prostate epithelial cells by slowing down cell cycle progression. Oncogene 12. Viard I, Wehrli P, Jornot L, et al. Clusterin gene expression mediates resistance to 2002;21:4328–34. apoptotic cell death induced by heat shock and oxidative stress. J Investig Dermatol 24. Scaltriti M, Brausi M, Amorosi A, et al. Clusterin (SGP-2, ApoJ) expression is 1999;112:290–6. down-regulated in low- and high-grade human prostate cancer. Int J Cancer 2004; 13. French LE, Wohlwend A, Sappino AP, Tschopp J, Schifferli JA. Human clusterin 108:23–30. gene expression is confined to surviving cells during in vitro programmed cell death. 25. McLaughlin L, Zhu G, Mistry M, et al. Apolipoprotein J/clusterin limits the severity J Clin Investig 1994;93:877–84. of murine autoimmune myocarditis. J Clin Investig 2000;106:1105–13. 14. Jones SE, Jomary C. Clusterin. Int J Biochem Cell Biol 2002;34:427–31. 26. Yang WC, Mathew J, Velcich A, et al. Targeted inactivation of the p21(WAF1/cip1) 15. Hara I, Miyake H, Gleave ME, Kamidono S. Introduction of clusterin gene into gene enhances Apc-initiated tumor formation and the tumor-promoting activity of a human renal cell carcinoma cells enhances their resistance to cytotoxic chemotherapy Western-style high-risk diet by altering cell maturation in the intestinal mucosal. through inhibition of apoptosis both in vitro and in vivo. Jpn J Cancer Res 2001;92: Cancer Res 2001;61:565–9. 1220–4. 27. Weinberg WC, Denning MF. p21Waf1 control of epithelial cell cycle and cell fate. 16. Leskov KS, Klokov DY, Li J, Kinsella TJ, Boothman DA. Synthesis and functional Crit Rev Oral Biol Med 2002;13:453–64. analyses of nuclear clusterin, a cell death protein. J Biol Chem 2003;278:11590–600. 28. Jin G, Howe PH. Transforming growth factor beta regulates clusterin gene 17. Pucci S, Bonanno E, Pichiorri F, Angeloni C, Spagnoli LG. Modulation of different expression via modulation of transcription factor c-Fos. Eur J Biochem 1999;263: clusterin isoforms in human colon tumorigenesis. Oncogene 2004;23:2298–304. 534– 42.

7419

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2004 American Association for Cancer Research. Clusterin-Mediated Apoptosis Is Regulated by Adenomatous Polyposis Coli and Is p21 Dependent but p53 Independent

Tingan Chen, Joel Turner, Susan McCarthy, et al.

Cancer Res 2004;64:7412-7419.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/64/20/7412

Cited articles This article cites 26 articles, 8 of which you can access for free at: http://cancerres.aacrjournals.org/content/64/20/7412.full#ref-list-1

Citing articles This article has been cited by 9 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/64/20/7412.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/64/20/7412. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.