PDZ-Binding Kinase/T-LAK Cell-Originated Protein Kinase, a Putative Cancer/Testis Antigen with an Oncogenic Activity in Breast Cancer

Research Article

Jae-Hyun Park, Meng-Lay Lin, Toshihiko Nishidate, Yusuke Nakamura, and Toyomasa Katagiri

Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan

Abstract to discover candidate target molecules for development of Breast cancer is one of the most common cancers among diagnosis and treatment of cancer (2). Through genome- women. To discover molecular targets that are applicable for wide expression analysis, we have isolated a number of oncogenes development of novel breast cancer therapy, we previously did that were involved in development and/or progression of genome-wide expression profile analysis of 81 breast cancers hepatocellular carcinomas (3, 4), synovial sarcomas (5, 6), and found dozens of genes that were highly and commonly and renal cell carcinomas (7). Such molecules are considered to up-regulated in breast cancer cells. Among them, we here be good candidate molecules for development of new therapeutic focused on one gene that encodes PDZ-binding kinase/T-LAK modalities. cell-originated protein kinase (PBK/TOPK), including a kinase Because cytotoxic drugs often cause severe adverse reactions, it domain. Northern blot analyses using mRNAs of normal is obvious that thoughtful selection of novel target molecules on the basis of well-characterized mechanisms of action should be human organs, breast cancer tissues, and cancer cell lines indicated this molecule to be a novel cancer/testis antigen. very helpful to develop effective anticancer drugs with the Reduction of PBK/TOPK expression by small interfering RNA minimum risk of side effects. Toward this goal, we previously did resulted in significant suppression of cell growth probably due the expression profile analysis of 81 breast cancers (8) and 29 to dysfunction in the cytokinetic process. Immunocyto- normal human tissues (9). Among the genes up-regulated in breast chemical analysis with anti-PBK/TOPK antibody implicated cancers, we focused on PDZ-binding kinase/T-LAK cell-originated a critical role of PBK/TOPK in an early step of mitosis. PBK/ protein kinase (PBK/TOPK), which is significantly overexpressed in TOPK could phosphorylate histone H3 at Ser10 in vitro and the great majority of breast cancer cases examined. PBK/TOPK was in vivo, and mediated its growth-promoting effect through first identified as a Dlg1-interacting protein by yeast two-hybrid histone H3 modification. Because PBK/TOPK is the cancer/ screening and characterized as a mitotic kinase with PDZ-binding testis antigen and its kinase function is likely to be related to motif at COOH terminus (10). PBK/TOPK was also identified by the its oncogenic activity, we suggest PBK/TOPK to be a promising other group as a mitogen-activated protein kinase (MAPK) kinase (MAPKK)–like protein kinase (11). molecular target for breast cancer therapy. (Cancer Res 2006; Cancer/testis antigens can be defined by predominant expres- 66(18): 9186-95) sion in various types of cancer and undetectable expression in normal tissues except germ cells in testis. To date, 44 distinct Introduction cancer/testis antigen families have been identified with particular Breast cancer is one of the most common cancers worldwide attention as potential targets for tumor immunotherapy (12). It has and has been significantly increasing (1). To reduce the mortality been reported that various cancer/testis antigens were expressed rate and to make the quality of life to the patients better, various relatively commonly in breast cancers, suggesting their potentials approaches, including early detection with mammography and for development of targeted therapy for breast cancer (13). Indeed, development of molecular-targeted therapeutic drugs, such as some of cancer/testis antigens, like NY-ESO-1 and NY-BR-1, are tamoxifen and trastuzumab, have been taken. Due to such efforts, applied as cancer vaccines in early clinical trials for breast cancer the mortality rates in Western societies have been improved, but patients (14). at present the very limited treatment options can be available to Here, we report overexpression of PBK/TOPK in the great patients at an advanced stage, particularly those with hormone- majority of breast cancer cells and that its kinase activity is likely to independent type. Hence, novel molecular-targeted drugs to play an important role in mammary carcinogenesis. Furthermore, provide better management to such patients are still eagerly the fact that PBK/TOPK expression pattern as the cancer/testis expected. antigen suggests PBK/TOPK to be a promising molecular target for cDNA microarray has been used as effective tools to analyze breast cancer therapy through cancer vaccine–mediated immuno- expression patterns of thousands of genes simultaneously. therapy and/or inhibition of PBK/TOPK–specific kinase function. Comparison of genome-wide expression profiles between cancers and normal cells by cDNA microarray provides useful information Materials and Methods Breast cancer cell lines and clinical samples. Human breast cancer cell lines BT-20, HCC1937, MCF-7, MDA-MB-435 S, SKBR3, T47D, and YMB1, Requests for reprints: Toyomasa Katagiri, Laboratory of Molecular Medicine, as well as immortalized human mammary cell-line HBL100, are purchased Human Genome Center, Institute of Medical Science, The University of Tokyo, 4-6-1 from American Type Culture Collection (Rockville, MD), and cultured under Shirokanedai, Minato-ku, 108-8639 Tokyo, Japan. Phone: 81-3-5449-5374; Fax: 81-3- the recommendations of their respective depositors. Human mammalian 5449-5406; E-mail: [email protected]. I2006 American Association for Cancer Research. epithelial cell (HMEC) is purchased from Cambrex Bio Science (Walkers- doi:10.1158/0008-5472.CAN-06-1601 ville, MD). HBC4, HBC5, and MDA-MB-231 cell lines are kind gifts from

Cancer Res 2006; 66: (18). September 15, 2006 9186 www.aacrjournals.org

Dr. Yamori (Division of Molecular Pharmacology, Cancer Chemotherapy for 15 minutes, and rendered permeable with PBS (À) containing 0.1% Center, Japanese Foundation for Cancer Research, Tokyo, Japan). All cells Triton X-100 at 4jC for 2.5 minutes. Subsequently, the cells were covered were cultured in appropriate medium, i.e., RPMI 1640 (Sigma-Aldrich, St. with 3% bovine serum albumin in PBS (À)at4jC for 12 hours to block Louis, MO) for HBC4, HBC5, HCC1937, T47D, and YMB1 (with 2 mmol/L nonspecific hybridization followed by incubation with a mouse anti- L-glutamine); DMEM (Sigma-Aldrich) for HBL100; EMEM (Sigma-Aldrich) PBK/TOPK monoclonal antibody (BD Biosciences) diluted at 1:100. After for BT-20 and MCF-7 (with 0.01 mg/mL insulin); McCoy (Sigma-Aldrich) for washing with PBS (À), the cells were stained by an Alexa 594–conjugated SKBR3 (with 1.5 mmol/L L-glutamine); L-15 (Roche, Basel, Switzerland) for anti-mouse secondary antibody (Molecular Probes, Eugene, OR) diluted at MDA-MB-231 and MDA-MB-435S; MEGM (Cambrex Bio Science) for HMEC. 1:1,000. Nuclei were counterstained with 4¶,6¶-diamidine-2¶-phenylindole Each medium was supplemented with 10% fetal bovine serum (Cansera dihydrochloride (DAPI). Fluorescent images were obtained under a TCS SP2 International, Etobicoke, Ontario, Canada) and 1% antibiotic/antimycotic AOBS microscope (Leica, Tokyo, Japan). To examine phosphorylated solution (Sigma-Aldrich). MDA-MB-231 and MDA-MB-435S cells were histone H3 at Ser10, the protein was detected by phosphorylated histone 10 maintained at 37jC in atmosphere of humidified air without CO2. Other H3 (Ser )–specific rabbit polyclonal antibody (Cell Signaling Technologies, cell lines were maintained at 37jC in atmosphere of humidified air with 5% Beverly, MA).

CO2. Tissue samples from surgically resected breast cancers and their Immunohistochemical staining. Expression patterns of PBK/TOPK corresponding clinical information were obtained from Department of protein in breast cancer and normal tissues were investigated as described Breast Surgery, Cancer Institute Hospital, Tokyo, after obtaining written previously (7) by using anti-PBK/TOPK mouse monoclonal antibody (BD informed consent. Biosciences). For investigation of normal organs, we purchased commer- Semiquantitative reverse transcription-PCR analysis. We extracted cially available tissue sections of heart, lung, liver, kidney, and testis total RNA from each of breast cancer clinical samples and from (Biochain). Briefly, paraffin-embedded specimens were treated with xylene microdissected cells, respectively, and then did T7-based amplification and ethanol, and were blocked by protein-blocking reagent (DakoCytoma- and reverse transcription as described previously (15). We prepared tion, Carpinteria, CA). The monoclonal antibody in antibody-diluted appropriate dilutions of each single-stranded cDNA for subsequent PCR solution (1:50) was added and then stained with substrate-chromogen by monitoring the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (DAKO liquid 3,3¶-diaminobenzidine chromogen, DakoCytomation). Finally, as a quantitative internal control. The PCR primer sequences were tissue specimens were stained with hematoxylin to discriminate nucleus 5¶-CGACCACTTTGTCAAGCTCA-3¶ and 5¶-GGTTGAGCACAGGGTACTT- from cytoplasm. TATT-3¶ for GAPDH;and5¶-AGACCCTAAAGATCGTCCTTCTG-3¶ and Western blot analysis. To detect the endogenous PBK/TOPK protein in 5¶-GTGTTTTAAGTCAGCATGAGCAG-3¶ for PBK/TOPK. breast cancer cells (BT-20, HBC4, HBC5, HBL100, MCF-7, MDA-MB-231, Northern blot analysis. Total RNAs were extracted from all breast SKBR3, and T47D), cells were lysed in lysis buffer [50 mmol/L Tris-HCL cancer cell lines using RNeasy kit (Qiagen, Valencia, CA) according to the (pH 8.0)/150 mmol/L NaCL/0.5% NP40], including 0.1% protease inhibitor instructions from the manufacturer. After treatment with DNase I (Nippon cocktail III (Calbiochem, San Diego, CA). After homogenization, the cell Gene, Osaka, Japan), mRNA was isolated with mRNA purification kit lysates were incubated on ice for 30 minutes and centrifuged at 14,000 (GE Healthcare, Buckinghamshire, United Kingdom) following the instruc- rpm for 15 minutes to separate only supernatant from cell debris. The tions of the manufacturer. One microgram each of mRNA isolated from amount of total protein was estimated by protein assay kit (Bio-Rad, normal adult human mammary gland (Biochain, Hayward, CA), lung, Hercules, CA), and then proteins were mixed with SDS sample buffer and heart, liver, kidney, and bone marrow (BD Biosciences, San Jose, CA) was boiled before loading at 10% SDS-PAGE gel. After electrophoresis, the separated on 1% denaturing agarose gels and transferred to nylon proteins were blotted onto nitrocellulose membrane (GE Healthcare). membranes (Breast cancer Northern blots). Human multiple-tissue Membranes, including proteins, were blocked by blocking solution and Northern blots (BD Biosciences) were hybridized with [a32P]dCTP- incubated with anti-PBK/TOPK monoclonal antibody (BD Biosciences) for labeled PCR products of PBK/TOPK prepared by reverse transcription- detection of endogenous PBK/TOPK protein. Finally, the membrane was PCR (RT-PCR; see below) or with [a32P]dCTP-labeled b-actin as a loading incubated with horseradish peroxidase–conjugated secondary antibody control, respectively. Prehybridization, hybridization, and washing were and protein bands were visualized by enhanced chemiluminescence done according to the recommendations from the supplier. The blots were detection reagents (GE Healthcare). h-Actin was examined to serve as a autoradiographed with intensifying screens at À80jC for 14 days. Specific loading control. probe for PBK/TOPK (320 bp) was prepared by RT-PCR using the primer set Wild-type and kinase-dead PBK/TOPK proteins were exogenously 5¶-AGACCCTAAAGATCGTCCTTCTG-3¶ and 5¶-GTGTTTTAAGTCAGCAT- expressed by transfection into T47D cell using pCAGGS-nHA expression GAGCAG-3¶, and was radioactively labeled with megaprime DNA labeling vector. Whole-cell lysates were harvested 48 hours after the transfection. The system (GE Healthcare). cells were lysed in the cell lysis buffer as described above. The following Cloning and mutagenesis. To construct the PBK/TOPK expression procedures are same as described above except using antihemagglutinin vectors, the entire coding sequence of PBK/TOPK cDNA was amplified rat high-affinity antibody (Roche) for antibody reaction. Furthermore, to by the PCR using KOD-Plus DNA polymerase (Toyobo, Osaka, Japan). detect activated PBK/TOPK protein endogenously, T47D cells were treated Primer sets were 5¶-CCGGAATTCATGGAAGGGATCAGTAATTTC-3¶ and with 100 nmol/L of okadaic acid (Calbiochem) or 0.3 Ag/mL nocodazole 5¶-CCGCTCGAGTCAGACATCTGTTTCCAGAGCTTC-3¶ (underlines indicate (Sigma-Aldrich) for 6 or 18 hours before harvesting, respectively (see the text). recognition sites of restriction enzymes) for wild-type PBK/TOPK. The PCR The following procedures were also carried out as described above. A products were inserted into the EocRI and XhoI sites of pCAGGS-nHA phosphorylated protein was confirmed by treatment with 1 unit of E protein expression vector. We also did two-step mutagenesis PCR to generate a phosphatase (New England Biolabs, Ipswich, MA) at 30jC for 2 hours. kinase-dead mutant in which Lys64 and Lys65 were substituted to alanines Fluorescence-activated cell sorting analysis. Cultured T47D breast (K64-65A), as described previously (10). The primer sets used for K64-65A cancer cells were synchronized their cell cycle by treatment with were 5¶-CATTCTCCTTGGGCTGTAGCAGCGATTAATCCTATATGTAATG-3¶ aphidicolin (Sigma-Aldrich) for 16 hours, washed five times with PBS (À), and 5¶-CATTACATATAGGATTAATCGCTGCTACAGCCCAAGGAGAATG-3¶ and added fresh culture medium to release from the cell cycle arrest. For (underlines indicate nucleotides that were replaced from the wild type). 15 hours (every 3 hours) after releasing, cells were collected and fixed with All of the constructs were confirmed by DNA sequencing (ABI3700, PE 70% ethanol, and then kept at 4jC before use. Cells were incubated with Applied Biosystems, Foster, CA). 10 mg/mL RNaseI in PBS (À)at37jC for 30 minutes and stained with 50 Ag Immunocytochemical staining. To examine the subcellular localization propidium iodide at room temperature for 30 minutes. Cell suspensions at of endogenous PBK/TOPK protein in breast cancer cell-lines, T47D, BT-20, each time point were analyzed with FACScan (Becton Dickinson, Franklin and HBC5, we seeded the cells at 2 Â 105 per well (Lab-Tek II chamber slide, Lakes, NJ). Nalgen Nunc International, Naperville, IL). Forty-eight hours after Construction of PBK/TOPK–specific small interfering RNA expres- incubation, cells were fixed with PBS (À) containing 4% paraformaldehyde sion vector using psiU6BX3.0. We established a vector-based RNA www.aacrjournals.org 9187 Cancer Res 2006; 66: (18). September 15, 2006

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2006 American Association for Cancer Research. Cancer Research interference (RNAi) expression system using psiU6BX3.0 small interfering PBK/TOPK, we transfected the wild-type protein and kinase-dead mutant RNA (siRNA) expression vector (16). An siRNA expression vector against (K64-65A) into T47D cells. After 48-hour culture, we treated 100 nmol/L PBK/TOPK (psiU6BX3.0-PBK/TOPK) was prepared by cloning of okadaic acid for 6 hours to activate PBK/TOPK. A total amount of H3 double-stranded oligonucleotides into the BbsI site in the psiU6BX3.0 protein as well as the level of its phosphorylation were examined with anti- vector (Table 1). All of the constructs were confirmed by DNA sequencing. histone H3 (Abcam, Cambridge, United Kingdom) and anti-phospho-H3 Gene-silencing effect by siRNA. Human breast cancer cell lines, T47D (Ser10) antibody (Cell Signaling Technologies), respectively. and BT-20, were plated onto 15 cm dishes (4 Â 106 per dish) and transfected A with 16 g each of psiU6BX3.0-Mock (without insertion) and psiU6BX3.0- Results PBK/TOPK (si-1, si-2, si-3, and constructs including three-base substitutions in si-3; Table 1) using FuGENE6 reagent (Roche) according to the Up-regulation of PBK/TOPK in breast cancers. We previously instructions from the manufacturer. At 24 hours after the transfection, did genome-wide expression profile analysis of 81 breast cancer cells are reseeded for colony formation assay (2 Â 106/10 cm dish), RT-PCR cases using cDNA microarray (8). Among genes up-regulated in (2 Â 106/10 cm dish), and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazo- breast cancers, we searched genes that encode proteins containing 5 lium bromide assay (MTT; 1 Â 10 per well). We selected the psiU6BX3.0- a kinase domain, either on the basis of reported information or introduced T47D or BT-20 cells with medium containing 0.7 or 0.6 mg/mL according to prediction by protein-motif program SMART1 and neomycin (Geneticin, Invitrogen, Carlsbad, CA), respectively. We changed focused on PBK/TOPK for which the high level of transactivation culture medium twice a week. Total RNAs were extracted from the cells was confirmed in the great majority of breast cancer cells (Fig. 1A). at 11-day incubation with neomycin, and then the knockdown effect of siRNAs was examined by a semiquantitative RT-PCR using specific primer Northern blot analysis of 10 breast cancer cell lines and six normal sets; 5¶-ATGGAAATCCCATCACCATCT-3¶ and 5¶-GGTTGAGCACAGGG- organs further confirmed that PBK/TOPK was specifically up- TACTTTATT-3¶ for GAPDH as an internal control, and 5¶-GCCTTCAT- regulated in all of the 10 breast cancer cell lines examined, but its CATCCAAACATT-3¶ and 5¶-GGCAAATATGTCTGCCTTGT-3¶ for PBK/TOPK. expression was hardly detectable in lung, heart, liver, kidney, bone Transfectants expressing siRNA were grown for 3 weeks in selective marrow, and mammary gland (Fig. 1B). To further examine medium containing neomycin, then fixed with 4% paraformaldehyde for expression pattern of PBK/TOPK in various normal tissues, we 15 minutes before staining with Giemsa solution (Merck, Whitehouse did Northern blot analysis using mRNAs from 23 tissues and Station, NJ) to assess colony number. To quantify cell viability, MTT assays identified two transcripts exclusively in the testis and thymus were done with cell counting kit-8 according to recommendations from the (Fig. 1C). According to the National Center for Biotechnology manufacturer (Wako, Osaka, Japan). Absorbance at 570 nm wavelength was Information database, two representative transcripts of 1,899 measured with a Microplate Reader 550 (Bio-Rad). These experiments were done in triplicate. nucleotides (Genbank accession no. NM018492) and 1,548 nucleo- In vitro and in vivo kinase assay. To evaluate the kinase activity of tides (Genbank accession no. AF189722) that share the same open PBK/TOPK, we did in vitro kinase assay using full-length recombinant reading frame encoding a 322-amino-acid peptide seemed to protein of PBK/TOPK (Invitrogen). Briefly, 1 Ag PBK/TOPK protein was correspond to the two bands observed in Northern blot analysis. incubated in 30 AL kinase assay buffer [50 mmol/L Tris-HCl (pH 7.5)/150 Immunocytochemical and immunohistochemical analysis of mmol/L NaCL/10 mmol/L MgCL2/10 mmol/L NaF/1 mmol/L Na3VO4/ PBK/TOPK. We investigated endogenous expression of PBK/TOPK 1 mmol/L EDTA/1 mmol/L DTT/50 Amol/L ATP] and then supplemented protein in cell lysates from breast cancer cell lines, BT-20, HBC4, A 32 A with 5 Ci of [ P-g]ATP (GE Healthcare). For the substrates, we added 5 g HBC5, HBL100, MCF-7, MDA-MB-231, SKBR3, and T47D by A histone mixture or 2.5 g histone H3 (Roche) in the reaction solutions. After Western blot analysis (Fig. 2A), using HMEC as a control of the 30-minute incubation at 30jC, the reactions were terminated by addition of experiments. All breast cancer cell lines showed high levels of PBK/ SDS sample buffer. After boiling, the protein samples were electrophoresed on 10% to 20% gradient gel (Bio-Rad), and then autoradiographed. TOPK expression, whereas HMEC cells showed no expression. To further examine whether histone H3 (Ser10) was phosphorylated by Subsequent immunocytochemical analysis of breast cancer cell lines, T47D, BT-20, and HBC5, using anti-PBK/TOPK monoclonal antibody indicated the localization of endogenous PBK/TOPK mainly in cytoplasm (Fig. 2B). To further investigate PBK/TOPK Table 1. Target sequences of PBK/TOPK inserted into siRNA expression vector expression in breast cancer and normal tissue sections, we did immunohistochemical staining with anti-PBK/TOPK monoclonal psiU6BX3.0-si-#1 antibody. We identified strong staining in the cytoplasm of 5¶-GGCAGACATATTTGCCTTT-3¶ three different histologic subtypes of breast cancer—intraductal psiU6BX3.0-si-#2 carcinoma, papillotubular carcinoma, and scirrhous carcinoma 5¶-CTGGATGAATCATACCAGA-3¶ (Fig. 2C-E)—but its expression was hardly detectable in normal psiU6BX3.0-si-#3 breast tissues (Fig. 2F). Furthermore, in concordance with the 5¶-GTGTGGCTTGCGTAAATAA-3¶ results of Northern blot analysis, its strong staining was detected psiU6BX3.0-mis-#1* at outer cell layer of seminiferous tubules of testis, whereas ¶ ¶ 5 -cTGTGGCTTcCGTAAATAt-3 no expression was observed in any of heart, lung, liver, and kidney psiU6BX3.0-mis-#2* (Fig. 3A-E). 5¶-cTGTGGCTTaCGTAAATAt-3¶ Knockdown effects of endogenous PBK/TOPK. To investi- psiU6BX3.0-mis-#3* 5¶-cTGTGGCTTcCGaAAATAt-3¶ gate its growth-promoting role in breast cancer cells, we knocked down the expression of endogenous PBK/TOPK in two breast cancer cells, T47D and BT-20 (Fig. 4A and B), by means of the NOTE: Underlined bases indicate siRNA-targeting sequences designed RNAi technique. Semiquantitative RT-PCR experiments detected from PBK/TOPK mRNA (Genbank accession no. NM018492). *Mismatched oligonucleotides were designed from si-3 by substitution of some internal bases (indicated by small letters).

1 http://smart.embl-heidelberg.de.

Cancer Res 2006; 66: (18). September 15, 2006 9188 www.aacrjournals.org

Figure 1. Expression pattern of PBK/TOPK in breast cancers and normal human organs. A, semiquantitative RT-PCR results showed overexpression of PBK/TOPK in breast cancers. Amplified cDNA products from 12 clinical samples (#4, #5, #13, #86, #110, #214, #327, #411, #623, #624, #631, and #869) were presented in comparison with that from microdissected normal ductal cells (normal). GAPDH served as an internal control. B, Northern blot results of breast cancer cell lines. The design for PBK/TOPK mRNA and radioisotope-labeled probe was detected at f2.0 to 1.5 kb showing two variants. All of the 10 breast cancer cell lines showed up-regulated expression rather than normal tissues, including vital organs (lung, heart, liver, kidney, and bone marrow). ACTB served as a loading control. C, tissue-specific distribution of PBK/TOPK was revealed by multiple tissue Northern blot (MTN), showing positive signal from thymus and testis more strongly, but no signal was detected from other tissues. The human normal tissues included in MTN membranes were as follows: 1, heart; 2, brain; 3, placenta; 4, lung; 5, liver; 6, skeletal muscle; 7, kidney; 8, pancreas; 9, spleen; 10, thymus; 11, prostate; 12, testis; 13, ovary; 14, small intestine; 15, colon; 16, peripheral blood leukocyte; 17, stomach; 18, thyroid; 19, spinal cord; 20, lymph node; 21, trachea; 22, adrenal gland; and 23, bone marrow. ACTB served as a loading control. significant knockdown effect of PBK/TOPK in the cells transfected transfected with mock construct (5.95%; Fig. 4F and G), implying with PBK/TOPK si-#2 and si-#3, but not with PBK/TOPK si-#1 or indispensable roles of PBK/TOPK on proliferation as well as on control siRNA (mock). In concordance with the knockdown effect, mitosis and/or cytokinesis for breast cancer cells. colony formation and MTT assays clearly revealed growth Cell cycle–dependent expression of PBK/TOPK. Because suppression of breast cancer cells by the two siRNAs, PBK/TOPK PBK/TOPK was reported to be a possible mitotic kinase (10), we si-#2 and si-#3, compared with two siRNAs showing no knockdown investigated its relation to the cell cycle progression. We examined effect. To exclude a possibility of off-target effect by PBK/TOPK expression of PBK/TOPK protein in T47D cells after synchroniza- siRNA (si-#3), we generated three additional siRNAs, each of which tion of cell cycle by aphidicolin. FACS analysis showed that the contained three-base replacement in PBK/TOPK si-#3 (mis-#1, proportion of the cells at G2-M phase was significantly increased mis-#2, and mis-#3; Table 1). These constructs showed no at 9 hours (63.94%) after the release from the cell cycle arrest suppressive effect on the PBK/TOPK expression (Fig. 4C)oron (Fig. 5A). Interestingly, Western blot analysis detected an additional the growth of breast cancer cells. These findings implied a critical band of high molecular weight of PBK/TOPK at 9 to 12 hours later role of PBK/TOPK in the growth of breast cancer cells. when most of the cells were at G2-M phase. The intensity of the In addition, we noticed phenotypic alterations of the cells high-molecular band was decreased at the 15-hour point (Fig. 5B). transfected with siRNAs showing the significant knockdown effect. Immunochemical analysis also revealed the subcellular localization We observed prolonged midbodies as well as incorrect cell of PBK/TOPK protein around the condensed chromosome in the divisions by abnormal cytokinesis in T47D cells in which PBK/ cells at mitosis, especially at prophase and metaphase (Fig. 5C). TOPK expression was suppressed (Fig. 4D and E). Western blot and To further investigate a role of high-molecular PBK/TOPK in cell fluorescence-activated cell sorting (FACS) analyses also identified cycle progression, we treated T47D breast cancer cells with an increase in the population of apoptotic (sub-G1) cells in the cells nocodazole, and then did Western blot and FACS analyses. As treated with PBK/TOPK siRNA (31.91%), although no phenotypic expected, an additional high-molecular band of the endogenous alteration or increase of sub-G1 population was observed in those PBK/TOPK in T47D cells was elevated in time-dependent manner www.aacrjournals.org 9189 Cancer Res 2006; 66: (18). September 15, 2006

(6-18 hours) after treatment of nocodazole (Fig. 5D, left), and Localization of PBK/TOPK around chromosome as well as its was disappeared by treatment of E phosphatase (Fig. 5D, right). In elevated phosphorylation in the early stage of mitosis allowed us to addition, FACS analysis showed that the proportion of the cells at investigate the physiologic role of histone H3 phosphorylation by G2-M phase was elevated from 6 to 18 hours after treatment of PBK/TOPK in breast cancer cells. We first transfected wild-type or nocodazole (Fig. 5E), indicating an important role of phosphory- kinase-dead (K64-65A) PBK/TOPK into T47D cells and then lated PBK/TOPK in mitosis. stimulated the cells by treatment of okadaic acid, which is known PBK/TOPK phosphorylated histone H3 (Ser10) in vitro and to induce premature mitosis (10). Both wild-type and kinase-dead in vivo. Because PBK/TOPK protein was localized mainly around PBK/TOPK were phosphorylated at the same level with okadaic chromosomal surfaces in mitotic cells, particularly at prophase and acid treatment by Western blot analysis using antihemagglutinin metaphase, we focused on histone as a candidate substrate for rat antibody. However, the wild-type protein enhanced phosphor- PBK/TOPK protein. We did in vitro kinase assay using purified ylation of histone H3 at Ser10 compared with kinase-dead mutant recombinant PBK/TOPK and mixed histone proteins (H2a, H2b, protein (Fig. 6B). Additionally, we confirmed that phosphorylation H3, and H4; Fig. 6A), and detected f15 kDa of phosphorylated of histone H3 at Ser10 was specifically reduced in PBK/TOPK– protein (lane 2), indicating that PBK/TOPK protein might depleted T47D cells by siRNA (si-#3), compared with in mock- or phosphorylate histone H3 protein on the basis of its molecular mis-1-siRNA–transfected cells (Fig. 6C). size (Fig. 6A, left). We further did in vitro kinase assay using In addition, we examined localization of endogenous PBK/ histone-H3 recombinant protein and confirmed that PBK/TOPK TOPK protein and phosphorylated histone H3. We synchronized protein phosphorylated histone H3 (Fig. 6A, right). In addition, we T47D and HBC5 cells with aphidicolin, and then did immuno- detected an f40 kDa of autophosphorylated PBK/TOPK by in vitro cytochemical staining using both anti-PBK/TOPK and anti- kinase assay as shown in Fig. 6A (asterisk). phospho-Ser10 H3 antibodies. As shown in Fig. 6D, we observed

Figure 2. Expression of PBK/TOPK in breast cancer cell lines and tissue sections. A, expression of endogenous PBK/TOPK protein in breast cancer cell lines in comparison with HMEC cell line, examined by Western blot analysis using anti-PBK/TOPK monoclonal antibody. B, three breast cancer cell lines, T47D, BT-20, and HBC5, were immunocytochemically stained with anti-PBK/TOPK monoclonal antibody (red) and DAPI (blue) to discriminate nucleus (see Materials and Methods). The endogenous PBK/TOPK protein was stained at the cytoplasm. C to F, immunohistochemical staining results of breast cancer (C-E) and normal breast (F) tissue sections. Endogenous PBK/TOPK protein was stained by use of anti-PBK/TOPK monoclonal antibody. The expression was hardly detected from normal breast tissues (F), but cancer cells were intensely stained at cytoplasm in all cancer tissues investigated, including intraductal (C), papillotubular (D), and scirrhous carcinoma (E). Original magnifications, Â100 (left) and Â200 (right).

Cancer Res 2006; 66: (18). September 15, 2006 9190 www.aacrjournals.org

Figure 3. Expression patterns of PBK/TOPK protein in normal human tissues. With anti-PBK/TOPK monoclonal antibody, the tissues were investigated including heart (A), lung (B), liver (C), kidney (D), and testis (E). As the results, the expressed PBK/TOPK protein was hardly detected from the four vital organs (A-D) but highly stained in testis, exclusively at the outer layer of seminiferous tubules (E). These immunohistochemical staining results were well correlated with those of MTN (Fig. 1C). Original magnifications, Â100 (left) and Â200 (right). partial overlapping of PBK/TOPK and phosphorylated histone H3 chemical analysis also supported the high level of endogenous around condensed chromosome in prophase cells, and over- PBK/TOPK expression in concordance with the results of lapping of both proteins in metaphase cells as well (Fig. 6E, Northern blot analysis. In addition, we showed by means of the yellow arrow) but their disappearance at anaphase (Fig. 6F; siRNA technique that knocking down the expression of endoge- yellow arrow). Taken together, we concluded that endogenous nous PBK/TOPK resulted in growth suppression of breast cancer PBK/TOPK is able to specifically phosphorylate histone H3 at cell lines (Fig. 4A-C). Furthermore, we have shown that PBK/ Ser10 during mitosis, especially prophase to metaphase in breast TOPK was also commonly up-regulated in bladder cancers and cancer cells. non–small cell lung cancers as well as breast cancers through cDNA microarray data (data not shown). Taken together, these findings suggest that PBK/TOPK has an oncogenic role, not only Discussion in breast cancers but also in bladder cancers and non–small cell cDNA microarray techniques have shown its great potential lung cancers. to effectively generate comprehensive information for screening PBK/TOPK was indicated to be involved in mitosis as shown by of novel molecular targets that are useful for development of its significant role in highly proliferating spermatocytes (10, 17). novel therapeutic reagents (2). In conventional drug screening Our immunohistochemical analysis of testis confirmed the approaches, the great majority of drug candidates that enter into expression of PBK/TOPK around the outer region of seminiferous clinical trials fail in development due to the adverse reactions or tubules where repeated mitosis of sperm germ cells followed by the insufficient efficacy. To reduce the failure risk during clinical meiosis occurs (17). In addition to important roles of PBK/TOPK trials, selection of molecules that are applicable for screening of in testis, our findings of its subcellular translocation during M small molecular compounds is critically important. In this point phase indicated its critical function at mitosis in cancer cells. of view, we have constructed the expression profile database of Moreover, we showed that knockdown of PBK/TOPK expression cancer cells as well as >29 normal human tissues and have been with specific siRNAs caused dysfunction of cytokinesis and taking a strategy to select genes that are frequently overexpressed subsequently led to apoptosis of the cancer cells (Fig. 4D-G). in cancer cells, but not expressed in normal organs. Through this These results suggest that PBK/TOPK is likely to play an important selection process, we have chosen dozens of genes, and role in cell division and cytokinesis. It is notable that microscopic subsequently confirmed overexpression of candidate genes by and FACS observations for the siRNA effect of PBK/TOPK are semiquantitative RT-PCR and Northern blot analyses. Among the quite similar with those of Annexin 11, which is required for validated genes, we focused on PBK/TOPK, which was overex- cytokinesis completion; knockdown of Annexin 11 resulted in pressed in breast cancers and not expressed in 29 normal human narrow cytoplasmic bridge and increased population of cells at tissues examined except testis and thymus. The immunohisto- sub-G1 (18). www.aacrjournals.org 9191 Cancer Res 2006; 66: (18). September 15, 2006

Because PBK/TOPK was indicated to contain a kinase domain, screening analysis (21). These two findings implied that PBK/TOPK we treated the cells with several kinds of stimuli, including okadaic might be involved in the MAPK pathway. However, our results of acid, phorbol 12-myristate 13-acetate, h-estradiol, and nocodazole, in vitro kinase assay failed to show that PBK/TOPK is an upstream to investigate its relationship with estrogen receptor and cell kinase of the p38 and p42/extracellular signal-regulated kinase mitotic signals, respectively (data not shown). Among these stimuli, 2 proteins, which were up-regulated commonly in breast cancer we found phosphorylation of PBK/TOPK after the treatment with cell lines (data not shown). Instead, we first observed phosphor- okadaic acid that is a specific inhibitor of serine/threonine protein ylation of histone H3 by PBK/TOPK with high selectivity. phosphatase and also known to cause mitosis-like processes in Interestingly, phosphorylation at the NH2 terminus of histone interphase cells, chromosome condensation, and entry into mitosis H3 (Ser10) is indicated to be an early mitotic event, accompanied in the Cdc2-independent manner (19, 20). In a previous study, PBK/ with chromosome condensation after okadaic acid treatment (19). TOPK was indicated to be a MAPKK-like protein that phosphor- Because our immunostaining experiment using breast cancer ylated the p38 protein (11). In addition, the possible interaction cells revealed subcellular localization of PBK/TOPK around the between Raf and PBK/TOPK was shown by the yeast two-hybrid chromosome in the cells at mitosis, especially at prophase and

Figure 4. Growth-inhibitory effects of PBK/TOPK siRNAs in breast cancer cell lines, T47Dand BT-20. Semiquantitative RT-PCR results showed PBK/TOPK silencing at 11 days after neomycin selection. GAPDH served as an internal control. MTT assays were done to evaluate cell viability at 11 days and graphed after standardization by mock to 1.0. Colony formation assays were carried out 3 weeks after selection (see Materials and Methods). Two siRNA constructs (si-#2 and si-#3) showed knockdown effects against internal PBK/TOPK expression and inhibited cell growth in both T47D( A) and BT-20 (B). PBK/TOPK– specific targeting was further confirmed by generation of mismatched siRNAs (mis-#1, mis-#2, and mis-#3; originally designed from si-#3) in T47D( C). Mock was used for a negative control. The phenotypic differences between mock control (D) and si-#3–induced T47Dcells ( E) were investigated by microscopic observation, 2 weeks after neomycin selection. Irregular appearances of cell were observed from PBK/TOPK–depleted cells, prolonged midbody, abolished, and uncontrolled cytokinesis (E). F, silencing of internal PBK/TOPK expression by siRNA (si-#3) was confirmed by Western blotting. G, FACS results showed more population of apoptotic cells (represented by sub-G1 percentage) in si-#3–induced T47Dcells (si-#3) rather than mock control. A total of 10,000 cells were equally counted from mock- and si-#3–transfected T47Dcells.

Cancer Res 2006; 66: (18). September 15, 2006 9192 www.aacrjournals.org

Figure 5. Phosphorylation of PBK/TOPK during mitosis. A, FACS analysis showed population of cells collected every 3 hours from 0 to 15 hours after synchronization. B, Western blot result of PBK/TOPK expression. It is notable that PBK/ TOPK is phosphorylated (arrow) and up-regulated from 6 to 12 hours after cell cycle releasing, which represents G2-M phase as shown in (A). C, representative immunocytochemical staining at 12 hours after cell cycle releasing. Intense staining of endogenous PBK/TOPK was detected near condensed chromosome at prophase or metaphase (indicated by yellow arrows). D, phosphorylation of PBK/TOPK during mitosis. Treatment with 0.3 Ag/mL nocodazole for 6, 12, and 18 hours showed that increased intensity of phosphorylated PBK/TOPK was time dependent (arrow, left). The cell lysates were further incubated with or without 1 unit E-phosphatase for 2 hours at 30jC. The slowly migrating band was revealed as a phosphorylated PBK/TOPK protein (arrow, right). E, FACS profiles at each time point (6-18 hours). The proportion of the cells at G2-M phase (arrows) was elevated from 6 to 18 hours after treatment of nocodazole by FACS analysis.

metaphase (Fig. 5C), we examined whether PBK/TOPK may we suggest that PBK/TOPK–histone H3 pathway may promote phosphorylate histone H3 at Ser10 in vivo. A comparison of the mitotic events, thus enhancing cancer cell proliferation, similarly to wild-type and kinase-dead PBK/TOPK protein with/without Pak1 whose significant role is indicated in breast cancer cells (27). okadaic acid stimulation showed that PBK/TOPK phosphorylated However, morphologic changes of the cells in which PBK/TOPK was at Ser10 of histone H3 (Fig. 6B), and endogenous PBK/TOPK knocked down by siRNA implied a presence of other substrates protein was well merged with phosphorylated histone H3 in that might be involved in cytokinesis (Fig. 4). In this study, we used mitotic cells (Fig. 6D). Posttranslational modifications at the NH2- okadaic acid, which is known to induce premature mitosis, as a terminal portion of histone H3, including acetylation, methylation, stimulus for activation of PBK/TOPK. Interestingly, previous and phosphorylation, were described previously (22–24). Among reports indicated that okadaic acid induced Ser10 phosphorylation them, phosphorylation of histone H3 at Ser10 is involved in the of histone H3 through inhibition of protein phosphatases (28, 29). initiation of mammalian chromosome condensation, an essential For example, Aurora-A is deactivated by protein phosphatase 2A event in cell mitosis (24, 25). According to the ‘‘ready production (PP2A), but to be reactivated by its autophosphorylation through label’’ model, Ser10 phosphorylation of histone H3 reaches at the binding with TPX2 (targeting protein for Xenopus kinesin-like maximum level in metaphase, as an indication that the chromo- protein 2) protein that impairs the activity of PP2A (29). somes are ready to be separated, and then Ser10 is dephosphory- PBK/TOPK has been reported as a member of the Ser/Thr lated accompanied by metaphase/anaphase transition (26). Cell kinase family. Our results presented here indicated that PBK/ cycle–dependent Ser10 phosphorylation of histone H3 is well TOPK is overexpressed in breast cancer and its kinase activity correlated with PBK/TOPK expression level and localization, possibly plays a significant role in breast cancer cell growth. particularly in the early stage of mitosis (Fig. 6D and E). Therefore, Recent drug development is focused on targeting important www.aacrjournals.org 9193 Cancer Res 2006; 66: (18). September 15, 2006

Figure 6. PBK/TOPK phosphorylates Ser10-histone H3 in vitro and in vivo. A, in vitro kinase assay was done with purified recombinant protein of PBK/TOPK (40 kDa, including histidine tag). In addition to PBK/TOPK, histone mixture and histone H3 were added to serve as a substrate (see text). Arrow, phosphorylated histone H3; *, autophosphorylation of PBK/TOPK. B, T47Dcells were transfected with wild-type and kinase-dead mutant (K64-65A), followed by treatment with 100 nmol/L okadaic acid (OA) for 6 hours. Okadaic acid treatment resulted in phosphorylation of both PBK/TOPK proteins (arrow), but only wild type– induced phosphorylation of H3, as detected by the phosphorylation of Ser10-specific antibody. C, internal expression of PBK/ TOPK was silenced in T47Dcells by siRNA (si-#3) after transfection and neomycin selection for 2 weeks. Consequently, PBK/TOPK depletion was accompanied by reduced phosphorylation of histone H3 at Ser10. h-actin and total H3 were also examined to serve as a loading control. D, immunocytochemical staining analysis of PBK/TOPK and histone H3. The results showed that PBK/TOPK (red) merged with phosphorylated histone H3 at Ser10 (green) on condensed chromosome (blue) of mitotic cells (prophase) in breast cancer cell lines T47Dand HBC5. Yellow arrows, cells at mitosis. E, subcellular localization of PBK/TOPK and phosphorylated histone H3 at Ser10 were observed in metaphase (yellow arrow) of T47Dcells. Whitearrow, cell at interphase. F, PBK/TOPK expression and histone H3 phosphorylation were diminished at anaphase cells (yellow arrow). Whitearrow, cell at interphase.

molecules involved in the oncogenic pathways, represented by Acknowledgments imatinib mesylate and trastuzumab. The combination of breast Received 5/2/2006; revised 6/25/2006; accepted 7/20/2006. cancer expression profiling with RNAi should help us to identify The costs of publication of this article were defrayed in part by the payment of page such drug targets for therapy (30). The present results of kinase charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. activity further provide promising possibilities to develop novel We thank Akiko Konuma and Kyoko Kijima for technical support; and Drs. Arata anticancer agent. Shimo, Tomomi Ueki, Chikako Fukukawa, and Akira Togashi for helpful discussions.

References gene with PDZ domains, that is upregulated in synovial human tissues with a cDNA microarray. DNA Res 2002; sarcomas. Oncogene 2004;23:5551–7. 9:35–45. 1. Veronesi U, Boyle P, Goldhirsch A, Orecchia R, Viale G. 6. Nagayama S, Fukukawa C, Katagiri T, et al. Therapeutic 10. Gaudet S, Branton D, Lue RA. Characterization of Breast cancer. Lancet 2005;365:1727–41. potential of antibodies against FZD 10, a cell-surface pro- PDZ-binding kinase, a mitotic kinase. Proc Natl Acad Sci 2. Debouck C, Goodfellow PN. DNA microarrays in drug tein, for synovial sarcomas. Oncogene 2005;24:6201–12. U S A 2000;97:5167–72. discovery and development. Nat Genet 1999;21:48–50. 7. Togashi A, Katagiri T, Ashida S, et al. Hypoxia- 11. Abe Y, Matsumoto S, Kito K, Ueda N. Cloning and 3. Hamamoto R, Furukawa Y, Morita M, et al. SMYD3 inducible protein 2 (HIG2), a novel diagnostic marker expression of a novel MAPKK-like protein kinase, encodes a histone methyltransferase involved in the for renal cell carcinoma and potential target for lymphokine-activated killer T-cell-originated protein proliferation of cancer cells. Nat Cell Biol 2004;6:731–40. molecular therapy. Cancer Res 2005;65:4817–26. kinase, specifically expressed in the testis and activated 4. Yagyu R, Hamamoto R, Furukawa Y, Okabe H, 8. Nishidate T, Katagiri T, Lin ML, et al. Genome-wide lymphoid cells. J Biol Chem 2000;275:21525–31. Yamamura T, Nakamura Y. Isolation and characteriza- gene-expression profiles of breast-cancer cells purified 12. Simpson AJ, Caballero OL, Jungbluth A, Chen YT, tion of a novel human gene, VANGL1, as a therapeutic with laser microbeam microdissection: identification of Old LJ. Cancer/testis antigens, gametogenesis and target for hepatocellular carcinoma. Int J Oncol 2002;20: genes associated with progression and metastasis. Int J cancer. Nat Rev Cancer 2005;5:615–25. 1173–8. Oncol 2004;25:797–819. 13. Mischo A, Kubuschok B, Ertan K, et al. Prospec- 5. Nagayama S, Iiizumi M, Katagiri T, Toguchida J, 9. Saito-Hisaminato A, Katagiri T, Kakiuchi S, et al. tive study on the expression of cancer testis genes Nakamura Y. Identification of PDZK4, a novel human Genome-wide profiling of gene expression in 29 normal and antibody responses in 100 consecutive patients

Cancer Res 2006; 66: (18). September 15, 2006 9194 www.aacrjournals.org

with primary breast cancer. Int J Cancer 2006;118: 19. Ajiro K, Yoda K, Utsumi K, Nishikawa Y. Alteration of 25. Van Hooser A, Goodrich DW, Allis CD, Brinkley BR, 696–703. cell cycle-dependent histone phosphorylations by oka- Mancini MA. Histone H3 phosphorylation is required for 14. Jager D, Knuth A. Antibodies and vaccines—hope or daic acid. Induction of mitosis-specific H3 phosphory- the initiation, but not maintenance, of mammalian illusion? Breast 2005;14:631–5. lation and chromatin condensation in mammalian chromosome condensation. J Cell Sci 1998;111:3497–506. 15. Kitahara O, Furukawa Y, Tanaka T, et al. Alterations interphase cells. J Biol Chem 1996;271:13197–201. 26. Hans F, Dimitrov S. Histone H3 phosphorylation and of gene expression during colorectal carcinogenesis 20. Gowdy PM, Anderson HJ, Roberge M. Entry into cell division. Oncogene 2001;20:3021–7. revealed by cDNA microarrays after laser-capture mitosis without Cdc2 kinase activation. J Cell Sci 1998; 27. Li F, Adam L, Vadlamudi RK, et al. p21-activated microdissection of tumor tissues and normal epithelia. 111:3401–10. kinase 1 interacts with and phosphorylates histone H3 Cancer Res 2001;61:3544–9. 21. Yuryev A, Wennogle LP. Novel raf kinase protein- in breast cancer cells. EMBO Rep 2002;3:767–73. 16. Shimokawa T, Furukawa Y, Sakai M, et al. Involve- protein interactions found by an exhaustive yeast two- 28. Murnion ME, Adams RR, Callister DM, et al. ment of the FGF18 gene in colorectal carcinogenesis, as hybrid analysis. Genomics 2003;81:112–25. Chromatin-associated protein phosphatase 1 regulates a novel downstream target of the h-catenin/T-cell factor 22. Martin C, Zhang Y. The diverse functions of histone aurora-B and histone H3 phosphorylation. J Biol Chem complex. Cancer Res 2003;63:6116–20. lysine methylation. Nat Rev Mol Cell Biol 2005;6:838–49. 2001;276:26656–65. 17. Fujibuchi T, Abe Y, Takeuchi T, et al. Expression and 23. Nowak SJ, Corces VG. Phosphorylation of histone 29. Eyers PA, Erikson E, Chen LG, Maller JL. A novel phosphorylation of TOPK during spermatogenesis. Dev H3: a balancing act between chromosome condensa- mechanism for activation of the protein kinase Aurora Growth Differ 2005;47:637–44. tion and transcriptional activation. Trends Genet 2004; A. Curr Biol 2003;13:691–7. 18. Tomas A, Futter C, Moss SE. Annexin 11 is required 20:214–20. 30. Clarke PA, te Poele R, Workman P. Gene expression for midbody formation and completion of the terminal 24. Prigent C, Dimitrov S. Phosphorylation of serine 10 in microarray technologies in the development of new phase of cytokinesis. J Cell Biol 2004;165:813–22. histone H3, what for? J Cell Sci 2003;116:3677–85. therapeutic agents. Eur J Cancer 2004;40:2560–91.

www.aacrjournals.org 9195 Cancer Res 2006; 66: (18). September 15, 2006

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2006 American Association for Cancer Research. PDZ-Binding Kinase/T-LAK Cell-Originated Protein Kinase, a Putative Cancer/Testis Antigen with an Oncogenic Activity in Breast Cancer

Jae-Hyun Park, Meng-Lay Lin, Toshihiko Nishidate, et al.

Cancer Res 2006;66:9186-9195.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/18/9186

Cited articles This article cites 30 articles, 12 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/18/9186.full#ref-list-1

Citing articles This article has been cited by 14 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/18/9186.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/66/18/9186. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.