Cancer-Associated Expression of Minichromosome Maintenance 3 Gene in Several Human Cancers and Its Involvement in Tumorigenesis

8386 Vol. 10, 8386–8395, December 15, 2004 Clinical Cancer Research

Seon-Ah Ha,1 Seung Min Shin,1 cells. Nude mice who received injections of HEK 293 cells Hong Namkoong,1 Heejeong Lee,2 stably transfected with MCM3 formed tumors in 6 weeks. Goang Won Cho,1 Soo Young Hur,3 Conclusions: Our study indicates that determination of 3 1,3 MCM3 expression level will facilitate the assessment of Tae Eung Kim, and Jin Woo Kim many different human malignancies in tumor diagnosis, 1Molecular Genetic Laboratory, Research Institute of Medical 2 3 and MCM3 is involved in multiple types of human carcino- Science, and Departments of Clinical Pathology and Obstetrics & genesis. Gynecology, College of Medicine, The Catholic University of Korea, Seoul, Korea INTRODUCTION Increasing knowledge on the molecular mechanisms in- ABSTRACT volved in regulating cellular proliferation has contributed to Purpose: The purpose of our study was to identify an development of proteins as biomarkers to detect various human unique gene that shows cancer-associated expression, eval- malignancies. The use of these markers to identify proliferating uates its potential usefulness in cancer diagnosis, and char- fraction within the tumors has proven valuable as proliferative acterizes its function related to human carcinogenesis. fraction is associated with the degree of malignancy. Experimental Design: We used the differential display The most widely used conventional proliferation markers reverse transcription-PCR method with normal cervical, include Ki-67 and proliferating cell nuclear antigen (PCNA). cervical cancer and metastatic tissues, and cervical cancer Expression of Ki-67 has been particularly used in a diagnosis of cell line to identify genes overexpressed in cancers. several human cancers such as breast cancer, soft tissue sar- Results: We identified a minichromosome maintenance 3 coma, meningiomas, prostate cancer, and non-Hodgkin’s lym- (MCM3) gene that was overexpressed in various human phoma (1–4). Despite of its usefulness, the exact function of cancers, including leukemia, lymphoma, and carcinomas of Ki-67 still remains unclear although report suggests that it might the uterine cervix, colon, lung, stomach, kidney and breast, be required for the ribosome biosynthesis during cell prolifera- and malignant melanoma. Western blot and immunohisto- tion instead of being directly associated to the cell cycle (5). chemical analyses also revealed that MCM3 protein was Furthermore, there is evidence indicating that Ki-67 is not elevated in most of human cancer tissues tested. We com- essentially required for cell proliferation (6). Although useful, pared the MCM3 protein expression levels in human can- Ki-67 provides only a limited information on cell cycle state cers with conventional proliferation markers, Ki-67 and (7–9). Likewise, PCNA, a valuable marker for assessing human proliferating cell nuclear antigen. MCM3 antibody was the neoplasms, is known to act as an auxiliary factor for DNA most specific for multiple human cancers, whereas prolifer- polymerase ␦ and is, thus, involved in DNA repair mechanisms, ating cell nuclear antigen was relatively less effective in as well as replication (10, 11). In this respect, the relationship of specificity, and Ki-67 failed to detect several human cancers. PCNA expression to proliferation has been controversial (12). The down-regulation of MCM3 protein level was examined The minichromosome maintenance (MCM) proteins are under serum starvation in both normal and cancer cells. essential replication initiation factors originally identified as Interestingly, MCM3 protein was stable in MCF-7 breast proteins required for MCM in Saccharomyces cerevisiae (13, cancer cells even up to 96 hours after serum starvation, 14). Similar classes have been found in Xenopus, murine, and whereas it was gradually degraded in normal BJ fibroblast human cells, with significant conservation of gene sequences (15–17). The best known among these are the MCM2–7 proteins, a family of six conserved proteins that are the key components of the replication initiation complex that initiates DNA Received 5/26/04; revised 9/22/04; accepted 9/22/04. synthesis in all eukaryotes (18, 19). The assembly of ORC, Grant support: Korea Research Foundation Grant KRF-2002-005- MCMs, and Cdc6 at the replication origins makes the chromatin E00011. competent or licensed for replication activities (20). At the G1 to The costs of publication of this article were defrayed in part by the S-phase transition, DNA replication is initiated by S-phase– payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to promoting cyclin-dependent kinases and the Cdc7 kinase. The indicate this fact. MCM proteins gradually dissociate from chromatin as S-phase Requests for reprints: Jin Woo Kim, Molecular Genetic Laboratory, advances (21). This dissociation of Cdc6 and MCM proteins Research Institute of Medical Science, College of Medicine, The Cath- from chromatin ensures that replication occurs only once per olic University of Korea, 505 Banpo-dong, Seocho-ku, Seoul 137-040, Korea. Phone: 82-2-590-2389; Fax: 82-2-593-2389; E-mail: jinwoo@ cell cycle (22–24). catholic.ac.kr. The mRNA levels of MCMs are dramatically induced at the ©2004 American Association for Cancer Research. G1-S boundary, coupled with a significant increase in protein

levels upon growth stimulation (25). Thus, in mammalian cells, mmol/L) adjusted to contain 1.5 g/L, and sodium bicarbonate the transcriptional mechanism of the MCM genes plays an supplemented with epidermal growth factor (20 ng/mL), insulin important role in the regulation of MCM activities. In contrast, (0.01 mg/mL), and hydrocortisone (500 ng/mL). Cancer cell MCM proteins are relatively constant and stable throughout the lines HeLa, A549, and MCF7 used in this study were cultured in cell cycle but rapidly disappear after entry into quiescent cells RPMI 1640 supplemented with 10% fetal bovine serum. For

(G0) and terminal differentiation stage in tissues (26, 27). This serum starvation experiments, cells were grown to an 85% makes MCMs specific markers of the cell cycle state suitable, confluence and deprived of serum from the media. Cells were and thus, anti-MCM antibodies have recently been investigated maintained serum-free, harvested at 0, 24, 48, 72, and 96 hours as diagnostic tools for detecting proliferating cells in various after serum starvation, and used for experiments. human neoplasms and dysplasia. Although most of studies are Clinical Specimen. Formalin-fixed, paraffin-embedded done with anti-MCM2 and MCM5 antibodies, all MCMs appear human tissues and tissue samples used in Northern and Western to have the similar distribution (28, 29). analyses were obtained from diagnostic biopsy or resection However, the MCMs are also expressed in normal prolif- specimens from patients at Kangnam St. Mary’s Hospital erating cells and the proliferative basal layers of normal tissues. (Seoul, South Korea). The tissue was used in agreement with The expression level in normal cells seems to be lower than in ethical guidelines approved by the ethics committee of our cancer cells, suggesting that the expression level of MCMs is research institute. differently regulated in cancer compared with normal cells (30). Differential Display Reverse Transcription-PCR. To- E2F can induce DNA replication in otherwise quiescent cells, tal RNA was extracted from tissues with an RNA extraction kit suggesting that genes involved in DNA replication are subject to (RNeasy total RNA kit, Qiagen, Inc., Valencia, CA), and 0.2 ␮g control by E2F (31–33). These include genes for enzymatic of total RNA were used to generate cDNA in a reverse tran- machinery for DNA synthesis such as DNA polymerase ␣ and scription reaction (RNAimage kit, GenHunter, Nashville, TN). regulatory components of replication initiation events. The ex- With the use of the differential display kit (RNAimage kit), we pression levels of MCMs are also enhanced in cell lines over- performed PCR with oligo-dT primers and arbitrary sequences, expressing E2F, indicating that they are also targets of E2F each 13 bases in length according to the manufacturer’s recom- (34–36). Therefore, aberrant expression of MCMs in cancer mendations. After cDNA of mRNAs was generated, the PCR cells might be partly attributable to deregulated E2F activities, products were separated by electrophoresis on a 6% denaturing which can eventually contribute to tumorigenesis by increasing polyacrylamide gel. Bands representing cDNA of interest were proliferation ability. excised from dried sequencing gel. The cDNA was eluted in In the present study, we report that antibody against MCM3 distilled water by boiling for 15 minutes and then was reampli- 35 is a powerful marker for cell proliferation. We also compared fied without [␣- S]dATP and with 20 ␮mol/L deoxynucleotide the expression level of MCM3 in various human cancers to that triphosphates instead of 2 ␮mol/L deoxynucleotide triphos- of PCNA and Ki-67. Our results indicate that MCM3 might be phates. From the films, a 292-bp partial cDNA (referred to as proven valuable in the diagnosis of various human cancers. CG262) was identified, which was expressed in the primary cervical cancer tissue, metastatic lymph node, and CUMC-6 cervical cancer cell line but not in normal cervical tissue. CG262 MATERIALS AND METHODS was identified by the use of 5Ј-arbitrary primer H-AP26 (5Ј- Ј Ј Ј Tissues and Cell Lines. For differential display of AAGCTTGCCATGG-3 ) and 3 H-T11G anchored primer (5 - mRNA, a normal exocervical tissue specimen was obtained AAGCTTTTTTTTTTTG-3Ј; GenHunter). CG262 was then from uterine myoma patients during hysterectomy, and un- subcloned into the pGEM-T easy vector with the use of the treated primary cervical cancer tissues and metastatic lymph TA-cloning system and subjected to an automatic sequencing node tissues were obtained during radical hysterectomy. Patient analysis. consent was obtained from each individual and the use of tissue cDNA Library Screening. To isolate the full-length samples was approved by the ethics committee of our institu- cDNA clone that contained the partial CG262 partial sequence, tion. The cervical cancer cell line used in differential display a bacteriophage ␭gt11 human lung embryonic fibroblast cDNA was CUMC-6, which was isolated in our laboratory and main- library was screened by plaque hybridization with 32P-labeled tained as described previously (37). Mammalian cell lines de- CG262 partial cDNA probe. scribed below were all obtained from the American Type Cul- Construction of Expression Vector and DNA Transfec- ture Collection (Manassas, VA): HEK 293, BJ, and IMR-90 are tion. To generate an eukaryotic expression construct of a human embryonic kidney cell line, a human fibroblast from MCM3, we amplified the coding region of MCM3 by PCR, normal foreskin, and a human lung cell fibroblast, respectively. which was then subcloned into pcDNA3.1 plasmid (pcDNA3.1 MCF-12F is an epithelial cell line from normal mammary gland. Directional TOPO expression kit, Invitrogen) according to the A549 is a human lung cancer cell line, and MCF-7 is a human manufacturer’s recommendations. To express MCM3 in HEK breast cancer cell line. Finally, HeLa is a human cervical cancer 293 cells, we seeded 3 ϫ 105 cells per 60-mm tissue culture dish cell line. (Costar, Cambridge, MA). The next day, the cells were incu- HEK 293, BJ, and IMR-90 cell lines were cultured in bated with 4 mL of serum-free DMEM, which contained 15 ␮L DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% of Lipofectamine reagent (Life Technologies, Inc., Rockville, fetal bovine serum (fetal bovine serum). MCF-12F was grown in MD) and 5 ␮g of pcDNA3.1-MCM3 and pcDNA3.1 vector medium containing a 1:1 mixture of DMEM and Ham’s F-12 alone. After a 10-hour incubation, the cells were cultured in 2ϩ medium with low Ca (0.04 mmol/L), L-glutamine (2.5 DMEM supplemented with 10% fetal bovine serum for HEK

293 cells containing 0.6 mg/mL G418 (Life Technologies, Inc.). room temperature. Incubation of the cells with anti-MCM3 HEK 293 cells were selected for resistance to 0.6 mg/mL G418 antibodies was done for 2 hours at room temperature in the for 3 to 4 weeks. After 4 weeks, the surviving cells were above blocking solution. The cells were washed with the same detached with trypsin and cloned by limiting dilution in 96-well blocking solution and incubated with antirabbit IgG conjugated culture plates at a concentration of 0.5 cell per well. The HEK to FITC for 1 hour. Then, they were washed with the blocking 293 cells were maintained in DMEM containing G418 (0.4 solution and photographed with fluorescence microscopy (AX- mg/mL). Thirty-five HEK 293 clones were selected and 70, Olympus, Tokyo, Japan). screened for the expression of MCM3 by Western blot analysis. Immunohistochemistry and Scoring Immunohisto- Antibodies. Rabbit polyclonal antibody was raised chemistry data. For immunohistochemistry, paraffin sections against the product of the MCM3 gene. A plasmid expressing a (5-␮m thick) of human cancer tissues were used. The sections whole MCM3 fused to a 6ϫ His tag was constructed by ligating were deparaffinized with xylene and etomidate. After washing

a PCR product of MCM3 into a pBAD/TOPO expression vector with tap water, the sections were treated with methanolic H2O2 (pBAD/TOPO thiofusion expression kit, Invitrogen). Induction for 30 minutes. Before incubation with primary antibody, the of Escherichia coli carrying the resulting plasmid with arabi- sections were permeabilized by incubation in 0.5% Triton nose produced a fusion protein that was purified with the use of X-100 in PBS for 15 minutes and then blocked with normal goat a Ni-nitrilotetraacetic acid-agarose column (ProBond purifica- serum for 15 minutes. The sections were incubated with affinity- tion system, Invitrogen). Antibody was raised against the puri- purified anti-MCM3 antibody for 2 hours at room temperature. fied protein in New Zealand White rabbits. The resulting rabbit After three washes with PBS, the sections were sequentially serum was affinity-purified with MCM3 protein covalently incubated with secondary antibody conjugated to horseradish cross-linked to CNBr-activated Sepharose beads (Amersham peroxidase (Zymed Laboratories, Inc., San Francisco, CA) for 1 Biosciences, Piscataway, NJ). hour at room temperature according to the manufacturer’s rec- The mouse monoclonal Ki-67 and PCNA antibodies were ommendations. Aminoethyl carbozole was used as the chromo- purchased from Oncogene Research Products (San Diego, CA). gen. After immunostaining, sections were counterstained with The mouse vimentin and cytokeratin antibodies were purchased hematoxylin. from Dako (Glostrup, Denmark). Finally, secondary antirabbit A semiquantitative scoring scheme was designed to de- antibody conjugated to FITC for immunofluorescence staining scribe the immunohistochemical staining observed. Cell nuclei was obtained from Jackson Immunoresearch (West Grove, PA). were positive or negative for Mcm3. The percentage of posi- Northern and Western Blot Analyses. Total RNA was tively stained tumor cells was scored and they were divided into extracted from frozen human tissues with TRIzol reagent (In- four categories as follows: 1 (0 to 25% cells stained), 2 (26 to vitrogen). Northern blot analysis was carried out in which 20 ␮g 50% cells stained), 3 (51 to 75% cells stained), and 4 (76 to of denatured total RNA were electrophoresed on a 1.0% form- 100% cells stained). aldehyde agarose gel and transferred to a nylon membrane Tumorigenicity. The in vivo studies were approved by (Roche Diagnostics GmbH, Mannheim, Germany). The mRNA the ethics committee of our institution. Mice were given filtered expression of MCM3 was also assessed in normal human tissues tap water ad libitum in accordance with Guide for the Care and and a variety of human cancer cell lines with the use of mem- Use of Laboratory Animals (NIH). To assess in vivo tumorige- brane commercially available from Clontech (Palo Alto, CA), nicity, HEK 293 cells (5 ϫ 106 cells) stably transfected with the which was processed as recommended by the supplier. Human MCM3 gene were injected s.c. into the posterior lateral aspect of ␤-actin cDNA control probe provided by Clontech was used as the trunk of 5-week-old athymic nu/nu female BALB/c mice. a loading control. All blots were hybridized with the randomly primed [32P]-labeled MCM3 partial cDNA probe (the CG262 fragment). For Western blot analysis, tissues were homogenized in a lysis buffer [20 mmol/L Tris (pH 7.4), 1% NP40, 5 mmol/L EDTA, 10% glycerol, 0.1% SDS, and 150 mmol/L NaCl] containing protease inhibitor mixture (Sigma, St. Louis, MO), and the homogenated was clarified by centrifugation at 5000 ϫ g for 10 minutes. Equivalent volumes of cell lysates containing 20 ␮g of total protein were loaded on 10% SDS-polyacrylamide gels and separated by electrophoresis. Protein samples were then transferred to nitrocellulose membranes. The blots were incubated with polyclonal anti-MCM3 serum and developed with the enhanced chemiluminescence detection kit (Pierce, Rock- Fig. 1 Identification of MCM3 and its expression patterns in human ford, IL). cervical tissues. Comparison of gene expression profiles by differential Immunofluorescence. Cells incubated under serum star- display reverse transcription-PCR from total RNA isolated from normal vation conditions were fixed by incubation with 3% paraform- cervical tissue, primary cervical cancer, cervical cancer tissue metastatic aldehyde in PBS for 15 minutes at room temperature. The cells to lymph node and from CUMC-6 cervical cancer cell line. Differential display was carried out using 5Ј arbitrary primer H-AP26 and 3Ј were then washed with PBS and permeabilized in 0.2% Triton H-T11G. The PCR products were resolved by electrophoresis. CG262 is X-100 in PBS for 5 minutes. After then, samples were blocked the name of the partial MCM3 gene product. The arrow indicates the by incubation with 1% BSA (Sigma) in PBS for 30 minutes at location corresponding to CG262 band.

292-bp partial cDNA fragment CG262 that was overexpressed in cervical cancer tissue, metastatic lymph node tissue, and CUMC-6 cervical cancer cells but was not expressed in normal cervical tissue (Fig. 1). Using the partial cDNA CG262 as a probe, we next screened normal human lung cDNA library to isolate the full-length cDNA. One clone with a size of 3091 bp named HCC-5 for human cervical cancer proto-oncogene 5 (GenBank accession number AY032603) was isolated. Se-

Fig. 3 MCM3, Ki-67, and PCNA protein expression levels in human tissues. A. Whole protein lysates were assessed for expression of MCM3 by Western blot analysis with polyclonal antibody. Comparison of MCM3 protein expression in human cancer tissues (C) and their corresponding normal counterparts (N). Three different blots containing the same samples were probed with antibodies against PCNA (B), Ki-67 (C), and ␤-actin (D), respectively.

Table 1 Tissue types and numbers used in the Fig. 2 MCM3 gene expressions in human tissues and cell lines. North- immunohistochemistry ern blot analyses were done to determine the expression pattern in different human tissues and cell lines. A. Normal 12-lane multiple tissue Tissue origins No. of normal tissues No. of tumor tissues Northern blot purchased from Clontech was probed with a radioactively Skin 3 4 labeled CG262 partial cDNA. Human ␤-actin cDNA control probe Breast 3 10 provided by Clontech was used as a loading control. B. Human cancer Lymph node 5 cell line multiple Northern blot was probed with a CG262 partial cDNA. Bone 1 3 C, comparison of MCM3 mRNA expressions in fresh human cancer Soft tissue 2 tissues and their corresponding normal counterparts. Total RNAs were Larynx 1 extracted from fresh normal (N) and cancer (C) tissues, and Northern Lung 6 11 blotting was done with CG262 as a radioactive probe. Human ␤-actin Parotid gland 1 2 cDNA control probe was used as a loading control. Submandibular gland 1 Liver 4 3 Gall bladder 1 Pancreas 4 3 Mice that received injections of HEK 293 cells stably trans- Esophagus 1 1 fected with the MCM3 gene were killed when the tumors Stomach 4 9 reached 1.5 to 2.5 cm in diameter approximately after 6 weeks. Small intestine 2 1 Mice that received injections of vector-transfected cells were Appendix 1 killed after 6 weeks. The size of the tumors was measured every Colon 6 7 Rectum 1 3 to 4 days. Tumors were fixed in 10% formalin for histologic Kidney 5 examination. Urinary bladder 2 Testis 2 5 Uterus 4 5 RESULTS Uterine cervix 3 3 Cloning of a MCM3 Overexpressed in Cervical Cancer. Ovary 4 10 To discover genes involved in cervical carcinogenesis, we ap- Adrenal gland 4 2 plied a differential display reverse transcription-PCR. Total Thyroid 4 2 Thymus 1 RNAs were isolated from the normal cervical tissue, the cervical Brain 2 2 cancer tissue, metastatic lymph node, and CUMC-6 cervical Prostate 2 7 cancer cell line. Partial cDNA synthesized from them was Placenta 4 resolved on acrylamide gel and excised from the gel based on Spleen 5 their cancer-associated expressions. We then identified the Tonsil 4

Fig. 4 Immunohistochemical staining of various human cancer tissues. Immunohistochem- ical staining was done with affinity-purified polyclonal antibody against MCM3 in malignant lymphoma of lymph node (A), small-cell carcinoma of lung (B), salivary duct carcinoma of submandibular gland (C), adenocarcinoma of gall bladder (D), squamous cell carcinoma of esophagus (E), adenocarcinoma of stomach (F), adenocarcinoma of colon (G), transitional cell carcinoma of kidney (H), carcinoma of urinary bladder (I), seminoma of testis (J), squamous cell carcinoma of uterine cervix (K), and endodermal sinus tumor of ovary (L). All figures are at ϫ200 magnification.

quence analysis showed that this cDNA encodes MCM3 gene ing total RNAs from tissues such as colon, lung, cervix, stom- (GenBank accession number NM_002388) with a molecular ach, kidney, and breast cancers and their corresponding normal mass of ϳ100 kDa. tissues. Each pair on the blot consists of normal and cancerous MCM3 Gene Expressions in Human Cancers. North- tissues obtained from the same patient. The result revealed that ern hybridization was done on a human normal multiple tissue MCM3 is expressed much higher in primary cancer tissues, mRNA blot and a human cancer cell line mRNA blot to analyze including carcinomas of colon, cervix, stomach, kidney, and MCM3 expressions in various normal tissues and cancer cell breast than in their normal counterparts, except lung cancer lines. The MCM3 mRNA (ϳ3 kb) was present in low levels in tissue (Fig. 2C). Although lung cancer sample didn’t show the some normal tissues, including the small intestine, spleen, and high expression on this particular blot, we could also detect the colon (Fig. 2A). Contrary to normal tissue samples, the MCM3 overexpression of MCM3 transcripts in other lung cancer sam- transcript was strongly detected in cancer cell lines, including ples (data not shown). This finding reassures the general re- lymphoblastic leukemia cell line MOLT-4, Burkitt’s lymphoma quirement of MCM3 function in several types of human cancers. cell line Raji, and cervical cancer cell line HeLa. Overexpres- MCM3, Ki-67, and PCNA Protein Expressions in Hu- sions of MCM3 were also detected in promyelocytic leukemia man Cancers. To detect the MCM3 protein, we raised poly- cell line HL-60, chronic myelogenous leukemia cell line K-562, clonal antibody against MCM3 expressed and purified in a and colon cancer cell line SW480, whereas it was expressed at prokaryotic system. We applied this polyclonal antibody to a moderate level in melanoma cell line G361 and A549, a lung tissue Western blot. We could observe the cancer-specific ex- cancer cell line (Fig. 2B). Thus, Northern blot analyses show pression of MCM3 protein corresponding to a molecular mass that MCM3 mRNA exists abundantly in cancer cell lines com- of ϳ100 kDa in multiple cancer tissues, including carcinomas of pared with its little or no expression in most normal tissues. To colon, lung, cervix, stomach, kidney, and breast, although the additionally examine MCM3 expressions in fresh human cancer amount of proteins expressed varies in each cancer sample (Fig. tissues, Northern hybridization was carried out to blots contain- 3A). Although this result is consistent with the mRNA expres-

Fig. 5 MCM3 staining of normal human tissues. Anti-MCM3 antibodies were applied to paraffin-embedded tissue sections (at ϫ400 magnification). Uterine cervix, pancreas, breast, and liver tissues are negative for MCM3 expressions. Cell nuclei were counterstained with hematoxylin.

sion data (Fig. 2C), the heterogeneous level of MCM3 protein in the nuclei of cancer cells derived from multiple tissues with contrasts to the relatively constant levels of mRNA among unlabeled nuclei appearing pale with the hematoxylin counter different cancer samples. We also observed the smaller protein stain (Fig. 4). Levels of Mcm3 expression were variable among band with a faster electrophoretic mobility in lung, kidney, and tumor specimens studied. Ninety-three of 110 cancer samples cervix cancer tissues, respectively (Fig. 3A). However, given the analyzed showed the moderate or high levels of expressions, fact that MCM3 protein shows the specific proteolytic cleavage whereas the remaining samples showed cytoplasmic or no ex- during apoptosis (38), it can probably be a cleaved form of pressions, corresponding to Ͼ80% of a diagnostic efficiency for Mcm3p or a nonspecific band. We then applied antibodies human cancer. To additionally confirm that anti-Mcm3 antibod- against PCNA and Ki-67, known as conventional proliferation ies efficiently detect malignant cells, we also stained the normal markers to the same blots (39–41). PCNA showed the cancer- tissue specimens on paraffin-embedded sections (Fig. 5). We specific expressions only in colon and breast cancer tissues. included several normal tissues in this analysis (see Table 1). As Although the PCNA was overexpressed in lung, cervix, stom- expected, most of normal tissues were negative or weakly pos- ach, and kidney cancer tissues, their corresponding normal itive in staining. We observed that some normal tissues had the tissues also showed faint or weak PCNA expressions. In con- reactivity with anti-MCM3 antibodies, although the staining was trast, Ki-67 was detected only in stomach and colon cancer strictly restricted to the proliferating fractions (data not shown tissues, which was observed in a multiple band pattern because here). For example, in testes, MCM3 expression was high in of its modifications such as phosphorylations and glycosyla- spermatogonia in which sperms are actively formed and grow- tions. Moreover, only stomach cancer strongly expressed Ki-67, ing. The MCM3 expression in normal tonsil was also confined whereas colon cancer was very weakly positive for Ki-67. to germinal centers having active proliferating activities. Like- According to immunoblotting experiments, MCM3 was most wise, MCM3 expression is tightly associated to the proliferating specific for cancer, whereas PCNA was relatively less effective activities of tumor and normal tissues. Because the proliferative and Ki-67 failed to detect several human cancers. fraction is associated to the degree of malignancy, the identifi- Anti-MCM3 Antibody Detects Malignant Cells in Tis- cation of proliferating fraction within the tumors is being used in sue Sections. Our data described above suggest that anti- a prognosis and diagnosis of cancers. Therefore, the use of MCM3 antibody might serve as a marker for malignant cells anti-MCM3 antibodies will be proven valuable as a marker for derived from multiple anatomic sites. To examine this, we tested detecting abnormal malignant cells, although its usage is not affinity-purified polyclonal antibody on paraffin-embedded sec- separable from that of a proliferative marker. tions consisting of various human cancer tissues. A total of 110 The Delayed Down-Regulation of MCM3 in Cancer tumor tissues was included in this immunohistochemical anal- Cells. Next, we were interested to investigate the down-regu- ysis as shown in Table 1. Positive immunostaining was observed lation of MCM3 protein to see if it is differently regulated in

Fig. 6 Different regulation of MCM3 in normal and cancer cells. A. Immu- nofluorescence staining of BJ (top panel) and MCF-7 (bottom panel) cells with anti-MCM3 antibody. BJ and MCF-7 cells were serum-starved for 0, 48, and 96 hours, respectively. They were then fixed and detected with anti- MCM3 antibody. B. Several normal (BJ, IMR-90, and MCF-12F) and cancer (MCF-7, A549, and HeLa) cells were serum-starved for indicated times. Total cell extracts were made, and 20 ␮g of proteins were resolved on SDS-PAGE gels. They were then probed with antibodies against MCM3 (B), PCNA (C), and Ki-67 (D), respectively.

normal and cancer cells, respectively. We first looked for degradation even at 96 hours after serum starvation, whereas all MCM3 protein expression pattern under the serum starvation normal cell lines (BJ fibroblasts, IMR-90 fibroblasts, and MCF- and could see the gradual decrease of MCM3 labeling in normal 12F mammary cells) had their own characteristic degradation BJ fibroblast cells during the time course of serum starvation kinetics (Fig. 6B). It is possible that the degradation kinetics (Fig. 6A). The clear nuclear staining was observed in most of might reflect on the proliferation potential of each cell type. cells at 0 hour and disappeared from a majority of cells at According to this notion, IMR-90 seems to most rapidly lose its 48-hour time point. It was then completely lost in all BJ cells up growth ability and enter into the quiescent state. To compare to 96 hours after serum starvation. In contrast, MCF-7 showed with other proliferation markers, we applied anti-PCNA (Fig. the persistent expression of MCM3 even at 96 hours after serum 6C) and anti-Ki67 antibodies (Fig. 6D) onto the same blot. starvation. This delayed down-regulation in MCF-7 implicates Notably, Ki-67 antigen was positive only in MCF-12F, A549, that MCM3 protein level is aberrantly regulated in cancer cells. and HeLa but negative in BJ, IMR-90, and MCF-7 (Fig. 6D). To determine whether this delayed down-regulation in general This shows that the level of Ki-67 is not always related to the in cancer cells, we analyzed the MCM3 protein levels under the proliferation potential. Moreover, its level decreases during the same condition as above by including several normal and cancer time course of serum starvation even in cancer cells such as cell lines (Fig. 6B). All cancer cells (MCF-7 breast cancer, A549 A549 and HeLa, as well as in MCF-12F mammary cells. How- lung cancer, and HeLa cervical cancer cells) tested showed no ever, PCNA behaved similarly to MCM3 in that it was decreas-

Fig. 7 Tumor formation in nude mice that received injections of HEK 293 stably transfected with MCM3. A, Western blot analysis to determine expression of MCM3 in parental HEK 293 cells (a negative control), cells transfected with pcDNA3.1 vector containing LacZ/V5 (a positive control), and cells transfected with pcDNA3.1 containing MCM3/V5. Western blot was probed with anti-V5 antibody. B, tumor formation in nude mice that received injections of the MCM3-transfected cells. C. H&E staining of tumor nodules taken from nude mice was shown (at ϫ200 magnification). D. Nude mice-derived tumor sections were immunohistochemically stained with mouse monoclonal antibody against keratin (at ϫ400 magnification).

ing in normal cells, except BJ. It was highly stable in all cancer ited the positive staining for keratin, indicating they are origi- cells (Fig. 6C). It is interesting that it was detected at the highest nated from the epithelial cells (Fig. 7D). level in IMR-90 and highly stable in BJ, which differs from MCM3 in those cell lines. In this regard, MCM3 seems most directly associated to the growth ability of normal cells. It also DISCUSSION seems to fit into the definition of a tumor marker because it can Although genetic characterization of tumor tissues shows detect only cancer cells but not normal cells under the prolonged that mutation of the p53 is the most common genetic alteration nutritional deprivation. in human cancers, the mutation rate of the p53 in cervical cancer Ectopic Expression of MCM3 Induces Tumorigenicity. is relatively low (42, 43). It implies that there might be other We have shown that MCM3 is overexpressed in almost all factors involved in cervical carcinogenesis. human cancers examined thus far. It implicates that MCM3 In our study, we applied the differential display reverse might facilitate the tumorigenesis by playing a role in the transcription-PCR method to explore genes related to the tumor- malignant transformation of cells. To determine whether the igenesis of human cervical tissue and identified MCM3, a com- ectopic expression of MCM3 can induce the tumorigenicity in ponent of the MCM2–7 complex. We have showed here that the human cells, we introduced MCM3 into HEK 293 cells, and majority of human tumors overexpress MCM3, and as a result, stable cell lines were selected in medium containing 0.6 mg/mL anti-MCM3 antibody is potentially valuable as a tumor marker. G418 by culturing for 21 days. Western blot analysis revealed The high level of MCM3 expression generally observed in that ϳ100 kDa product were expressed in these MCM3-trans- human cancer might contribute to the growth of cancer cells by fected cells (Fig. 7A). We next examined whether MCM3 can facilitating overall replication activities. It has been reported that contribute to the tumorigenic transformation of HEK 293 cells the overexpression of MCM proteins might contribute to cell with in vivo nude mouse tumor assay. Cells were harvested from growth by increasing the transcription rate (44) or through the culture dish in media containing G418. We then injected 5 ϫ interaction with the retinoblastoma protein (45). Indeed, we 106 cells into each 10 nude mice. These cells exhibited tumor- observed the ability of HEK 293 cells stably transfected with igenic potentials as shown by their development as tumors in 9 MCM3 to induce the tumor formation in nude mouse, implying of 10 nude mice (Fig. 7B), whereas cells with vector alone failed the role of MCM3 in tumorigenesis. to grow as tumors. We then analyzed tumors derived from nude The rapid proliferation rate is one of the features common mice for their origins. Tumors from nude mice injected with to most neoplasms. Identification of a proliferating fraction HEK 293 cells stably transfected with MCM3 displayed char- within the tumor cell population has been useful in a diagnosis acteristics of epithelial carcinoma (Fig. 7C). For morphologic or prognosis in a range of human cancers, including prostate and comparison between MCM3-derived tumor cells and epithelial breast cancer, and lymphoma (46). Biological markers indica- cells, we determined whether MCM3-derived tumor cells ex- tive of cell cycle state has proven useful for determining growth press epithelial cell marker, keratin. Tumor-derived cells exhib- fractions because previous methods mostly focused on S-phase

labeling and resulted in missing a considerable fraction of that stopped proliferating but still retain the replicative ability. inter-mitotic cells that are proliferating but not in S-phase. This agrees with the previous study (47). However, there still PCNA has a major limitation as a proliferation marker because remain problems in PCNA as shown in its overexpression in of its redundant role in DNA repair. This is in accordance with IMR-90 and in its high stability in BJ. It suggests that the PCNA our observations (Figs. 3B and 6C), which shows its consider- marker must have a higher chance of detecting normal cells as able expression in normal human tissues and cell lines without well as cancer cells. proliferating activities. Meanwhile, the expression of Ki-67 an- In conclusion, this study shows the aberrant expressions of tigen is restricted to proliferating fractions but disappears rap- MCM3 in cancer cells and its potential usefulness as a tumor idly when cells enter a resting state. The MCM proteins are key marker was also assessed. Additional studies on the control regulators of eukaryotic DNA replication. Their expression is mechanisms for MCM3 expression will provide new insights on stable throughout the cell cycle and, thereby, is an excellent how these mechanisms are deregulated in human cancers and indicator of cell cycle state. In particular, they are distinct from how significant they are in maintaining normal tissues. other proliferation markers in that they can also identify cells that are resting but still replication competent (47). Previous studies on application of anti-MCM antibodies to REFERENCES the diagnostic tumor pathology were mostly done with those 1. Jansen RL, Hupperets PS, Arends JW, et al. MIB-1 labelling index is against MCM2 and MCM5 (48–51). We additionally extended an independent prognostic marker in primary breast cancer. Br J Cancer 1998;78:460–5. this by analyzing the expression patterns of MCM3 in multiple 2. Perry A, Stafford SL, Scheithauer BW, Suman VJ, Lohse CM. The human cancers in both mRNA and protein levels. Immuno- prognostic significance of MIB-1, p53, and DNA flow cytometry in staining on paraffin sections was positive for various human completely resected primary meningiomas. Cancer (Phila.) 1998;82: cancers (Fig. 4). Although this should be done in a larger 2262–9. clinical scale, we can conclude that our affinity-purified anti- 3. Mashal RD, Lester S, Corless C, et al. Expression of cell cycle- MCM3 antibodies identify malignant cells with high sensitivity regulated proteins in prostate cancer. Cancer Res 1996;56:4159–63. and specificity in a broad range of human cancers. 4. Gerdes J, Dallenbach F, Lennert K, Lemke H, Stein H. Growth fractions in malignant non-Hodgkin’s lymphomas (NHL) as determined One of the interesting observations that we made is that the in situ with the monoclonal antibody Ki-67. Hematol Oncol 1984;2: amount of MCM3 proteins among primary tumors shows the 365–71. sample-to-sample variations (Fig. 3). This is one of the reasons 5. MacCallum DE, Hall PA. The location of pKi67 in the outer dense why we saw the heterogeneous staining among tumor sections fibrillary compartment of the nucleolus points to a role in ribosome in immunohistochemical analysis (Fig. 4). This suggests that biogenesis during the cell division cycle. J Pathol 2000;190:537–44. there might be a cascade of regulations to control its expression 6. Verheijen R, Kuijpers HJ, van-Driel R, et al. Ki-67 detects a nuclear levels during the process of mRNA translation in addition to the matrix proliferation-related antigen. J Cell Sci 1989;92:531–40. control mechanism from its promoter sequences. It is likely that 7. Baisch H, Gerdes J. Simultaneous staining of exponentially growing versus plateau phase cells with the proliferation-associated antibody the stability or translation rate of mRNA might be crucial factors Ki-67 and propidium iodide: analysis by flow cytometry. Cell Tissue for variable amount of proteins in each tumor. This notion is Kinet 1987;20:387–91. supported by our pulse-chase immunoprecipitation experiments, 8. van Oijen MG, Medema RH, Slootweg PJ, Rijksen G. Positivity of showing that the turnover rate of MCM3 protein was nearly the proliferation marker Ki-67 in noncycling cells. Am J Clin Pathol identical between normal and cancer cells.4 It indicates that 1998;110:24–31. regulation activity to determine the protein level is exercised not 9. Scholzen T, Gerdes J. The Ki-67 protein: from the known and the unknown. J Cell Physiol 2000;182:311–22. at the protein level but possibly at the mRNA level. Thus, their 10. Prelich G, Tan CK, Kostura M, et al. Functional identity of prolif- levels might be differently regulated possibly depending on erating cell nuclear antigen and a DNA polymerase-delta auxiliary several factors characteristic of each tumor, which await to be protein. Nature (Lond.) 1987;326:517–20. determined. Regardless of this, it will be of interest to uncover 11. Toschi L, Bravo R. Changes in cyclin/proliferating cell nuclear this regulation mechanism together with its promoter activity antigen distribution during DNA repair synthesis. J Cell Biol 1988;107: control. 1623–8. We measured and compared the down-regulation of 12. Hall PA, Levison DA, Woods AL, et al. Proliferating cell nuclear MCM3 proteins under the nutrient deprivation to that of PCNA antigen (PCNA) immunolocalisation in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some and Ki-67 (Fig. 6). The prolonged expression of both MCM3 neoplasms. J Pathol 1990;162:285–94. and PCNA under this unfavorable environment makes them 13. Maine GT, Sinha P, Tye BK. Mutants of S. cerevisiae defective in more attractive as tumor markers because there is a possibility the maintenance of minichromosomes. Genetics 1984;106:365–85. of a nutritional deprivation operating regionally in solid tumors. 14. Tye BK. Minichromosome maintenance as a genetic assay for In a case like this, a tumor marker can be defined separately defects in DNA replication. Methods 1999;18:329–34. from a proliferation marker because MCM3 was detectable only 15. Thommes P, Fett R, Schray B, et al. Properties of the nuclear P1 in cancer cells but not in normal cells after long serum starva- protein, a mammalian homologue of the yeast Mcm3 replication protein. Nucleic Acids Res 1992;20:1069–74. tion. In addition, their delayed down-regulation also indicates 16. Bell SP, Dutta A. Initiation of DNA replication in eukaryotic cells. that antibodies against MCM3 and PCNA can also detect cells Ann Rev Cell Dev Biol 1997;13:293–332. 17. Kearsey SE, Labib K. MCM proteins: evolution, properties and role in DNA replication. Biochem Biophys Acta 1998;1398:113–36. 18. Chong JP, Thommes P, Blow JJ. The role of MCM/P1 proteins in 4 Unpublished data. the licensing of DNA replication. Trends Biochem Sci 1996;21:102–6.

19. Tye BK. MCM proteins in DNA replication. Annu Rev Biochem 36. Bruemmer D, Yin F, Liu J, et al. Rapamycin inhibits E2F-depend- 1999;68:649–86. ent expression of minichromosome maintenance proteins in vascular 20. Chevalier S, Blow JJ. Cell cycle control of replication initiation in smooth muscle cells. Biochem Biophys Res Comm 2003;303:251–8. eukaryotes. Curr Opin Cell Biol 1996;8:815–21. 37. Kim JW, Lee CG, Cho YH, et al. CUMC-6, a new diploid human 21. Krude T, Musahl C, Laskey RA, Knippers R. Human replication cell line derived from a squamous carcinoma of the uterine cervix. proteins hCdc21, hCdc46 and P1Mcm3 bind chromatin uniformly be- Gynecol Oncol 1996;62:230–40. fore S-phase and are displaced locally during DNA replication. J Cell 38. Schwab BL, Leist M, Knippers R, Nicotera P. Selective proteolysis Sci 1996;109:309–18. of the nuclear replication factor MCM3 in apoptosis. Exp Cell Res 22. Kubota Y, Mimura S, Nishimoto SI, Takisawa II, Nojima H. Iden- 1998;238:415–21. tification of the yeast MCM3-related protein as a component of Xenopus 39. Suzuka I, Daidoji H, Matsuoka M, et al. Gene for proliferating-cell DNA replication licensing factor. Cell 1995;81:601–9. nuclear antigen (DNA polymerase delta auxiliary protein) is present in 23. Chong JP, Mahbubani IIM, Khoo CY, Blow JJ. Purification of an both mammalian and higher plant genomes. Proc Natl Acad Sci USA MCM-containing complex as a component of the DNA replication 1989;86:3189–93. licensing system. Nature (Lond.) 1995;375:418–21. 40. Lopez F, Belloc F, Lacombe F, et al. The labeling of proliferating 24. Madine MA, Khoo CY, Mills AD, Laskey RA. MCM3 complex cells by Ki67 and MIB-1 antibodies depends on the binding of a nuclear required for cell cycle regulation of DNA replication in vertebrate cells. protein to the DNA. Exp Cell Res 1994;210:145–53. Nature (Lond.) 1995;375:421–5. 41. Duchrow M, Schluter C, Wohlenberg C, Flad HD, Gerdes J. Mo- 25. Tsuruga H, Yabuta N, Hashizume K, Ikeda M, Endo Y, Nojima H. lecular characterization of the gene locus of the human cell prolifera- Expression, nuclear localization and interactions of human MCM/P1 tion-associated nuclear protein defined by monoclonal antibody Ki-67. proteins. Biochem Biophys Res Commun 1997;236:118–25. Cell Prolif 1996;29:1–12. 26. Musahl C, Holthoff HP, Lesch R, Knippers R. Stability of the 42. Crook T, Wrede D, Tidy JA, Mason WP, Evans DJ, Vousden KH. replicative Mcm3 protein in proliferating and differentiating human Clonal p53 mutation in primary cervical cancer: association with hu- cells. Exp Cell Res 1998;241:260–4. man-papillomavirus-negative tumours. Lancet 1992;339:1070–3. 27. Madine MA, Swietlik M, Pelizon C, Romanowski P, Mills AD, 43. Busby-Earle RM, Steel CM, Williams AR, Cohen B, Bird CC. p53 Laskey RA. The roles of the MCM, ORC, and Cdc6 proteins in deter- mutations in cervical carcinogenesis-low frequency and lack of corre- mining the replication competence of chromatin in quiescent cells. J lation with human papillomavirus status. Br J Cancer 1994;69:732–7. Struct Biol 2000;129:198–210. 44. DaFonseca CJ, Shu F, Zhang JJ. Identification of two residues in 28. Hiraiwa A, Fujita M, Nagasaka T, Adachi A, Ohashi M, Ishibashi MCM5 critical for the assembly of MCM complexes and Stat1-mediated M. Immunolocalization of hCDC47 protein in normal and neoplastic transcription activation in response to IFN-gamma. Proc Natl Acad Sci human tissues and its relation to growth. Int J Cancer 1997;74:180–4. USA 2001;98:3034–9. 29. Freeman A, Morris LS, Mills AD, et al. Minichromosome mainte- 45. Sterner JM, Dew-Knight S, Musahl C, Kornbluth S, Horowitz JM. nance proteins as biological markers of dysplasia and malignancy. Clin Negative regulation of DNA replication by the retinoblastoma protein Cancer Res 1999;5:2121–32. is mediated by its association with Mcm7. Mol Cell Biol 1998;18: 30. Ishimi U, Okayasu I, Kato C, et al. Enhanced expression of Mcm 2748–57. proteins in cancer cells derived from uterine cervix. Eur J Biochem 46. Quinn CM, Wright NA. The clinical assessment of proliferation and 2003;270:1089–101. growth in human tumours: evaluation of methods and applications as 31. Johnson DG, Schwarz JK, Cress WD, Nevins JR. Expression of prognostic variables. J Pathol 1990;160:93–102. transcription factor E2F1 induces quiescent cells to enter S phase. 47. Stoeber K, Tlsty TD, Happerfield L, et al. DNA replication licens- Nature (Lond.) 1993;365:349–52. ing and human cell proliferation. J Cell Sci 2001;114:2027–41. 32. Qin XQ, Livingston DM, Kaelin WG, Adams PD. Deregulated 48. Williams GH, Romanowski P, Morris L, et al. Improved cervical transcription factor E2F-1 expression leads to S-phase entry and p53- smear assessment using antibodies against proteins that regulate DNA mediated apoptosis. Proc Natl Acad Sci USA 1994;91:10918–22. replication. Proc Natl Acad Sci USA 1998;95:14932–7. 33. DeGregori JG, Leone K, Ohtani A, Miron A, Nevins JR. E2F1 49. Ramnath N, Hernandez FJ, Tan D-F, et al. MCM2 is an independent accumulation bypasses a G1 arrest resulting from the inhibition of G1 predictor of survival in patients with non–small-cell lung cancer. J Clin cyclin-dependent kinase activity. Genes Dev 1995;9:2873–87. Oncol 2001;19:4259–66. 34. Leone G, DeGregori J, Yan Z, et al. E2F3 activity is regulated 50. Stoeber K, Swinn R, Prevost AT, et al. Diagnosis of genito-urinary during the cell cycle and is required for the induction of S phase. Genes tract cancer by detection of minichromosome maintenance 5 protein in Dev 1998;12:2120–30. urine sediments. J Natl Cancer Inst (Bethesda) 2002;94:1071–9. 35. Ohtani K, Iwanaga R, Nakamura M, et al. Cell growth-regulated 51. Gonzalez MA, Pinder SE, Callgy G, et al. Minichromosome main- expression of mammalian MCM5 and MCM6 genes mediated by the tenance protein 2 is a strong independent prognostic marker in breast transcription factor E2F. Oncogene 1999;18:2299–309. cancer. J Clin Oncol 2003;21:4306–13.

Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2004 American Association for Cancer Research. Cancer-Associated Expression of Minichromosome Maintenance 3 Gene in Several Human Cancers and Its Involvement in Tumorigenesis

Seon-Ah Ha, Seung Min Shin, Hong Namkoong, et al.

Clin Cancer Res 2004;10:8386-8395.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/10/24/8386

Cited articles This article cites 50 articles, 16 of which you can access for free at: http://clincancerres.aacrjournals.org/content/10/24/8386.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/10/24/8386.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://clincancerres.aacrjournals.org/content/10/24/8386. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.