Involvement of Elevated Expression of Multiple Cell-Cycle Regulator, DTL

Oncogene (2008) 27, 5672–5683 & 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $32.00 www.nature.com/onc ORIGINAL ARTICLE Involvement of elevated expression of multiple cell-cycle regulator, DTL/RAMP (denticleless/RA-regulated nuclear matrix associated protein), in the growth of breast cancer cells

T Ueki1,2, T Nishidate1, JH Park1, ML Lin1, A Shimo1, K Hirata2, Y Nakamura1 and T Katagiri1,3

1Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, Minato-ku, Tokyo, Japan and 2Department of Surgery, Sapporo Medical University, Chuo-ku, Sapporo, Japan

To investigate the detailed molecular mechanism of mam- Introduction mary carcinogenesis and discover novel therapeutic targets, we previously analysed gene expression profiles of breast Breast cancer is the most common cancer in women, cancers. We here report characterization of a significant role with estimated new cases of 1.15 million worldwide of DTL/RAMP (denticleless/RA-regulated nuclear matrix in 2002 (Parkin et al., 2005). Incidence rates of breast associated protein) in mammary carcinogenesis. Semiquanti- cancer are increasing in most countries, and the tative RT–PCR and northern blot analyses confirmed increasing rate is much higher in countries where its upregulation of DTL/RAMP in the majority of breast incidence was previously low (Parkin et al., 2005). Early cancer cases and all of breast cancer cell lines examined. detection with mammography as well as development Immunocytochemical and western blot analyses using anti- of molecular targeted drugs such as tamoxifen and DTL/RAMP polyclonal antibody revealed cell-cycle-depen- trastuzumab reduced the mortality rate and made the dent localization of endogenous DTL/RAMP protein in quality of life of the patients better (Bange et al., 2001; breast cancer cells; nuclear localization was observed in cells Navolanic and McCubrey, 2005). However, still very at interphase and the protein was concentrated at the limited treatment options are available to patients at an contractile ring in cytokinesis process. The expression level of advanced stage, particularly those with a hormone- DTL/RAMP protein became highest at G1/S phases, independent tumor. Hence, development of novel drugs whereas its phosphorylation level was enhanced during to provide better management to such patients is still mitotic phase. Treatment of breast cancer cells, T47D and eagerly expected. HBC4, with small-interfering RNAs against DTL/RAMP Gene expression profiles obtained by cDNA micro- effectively suppressed its expression and caused accumulation array analysis have been proven to provide detailed of G2/M cells, resulting in growth inhibition of cancer cells. characterization of individual cancers and such informa- We further demonstrate the in vitro phosphorylation of DTL/ tion should contribute to choose more appropriate RAMP through an interaction with the mitotic kinase, clinical strategies to individual patients through Aurora kinase-B (AURKB). Interestingly, depletion of development of novel drugs and providing the basis of AURKB expression with siRNA in breast cancer cells personalized treatment (Petricoin et al., 2002). Through reduced the phosphorylation of DTL/RAMP and decreased the genome-wide expression analysis we have isolated a the stability of DTL/RAMP protein. These findings imply number of genes that function as oncogenes in the process important roles of DTL/RAMP in growth of breast cancer of development and/or progression of breast cancers cells and suggest that DTL/RAMP might be a promising (Park et al., 2006; Lin et al., 2007; Shimo et al., 2007), moleculartargetfortreatmentofbreastcancer. synovial sarcomas (Nagayama et al., 2004, 2005) Oncogene (2008) 27, 5672–5683; doi:10.1038/onc.2008.186; and renal cell carcinomas (Togashi et al., 2005; Hirota published online 9 June 2008 et al., 2006). Such molecules are considered to be candidate targets for development of new therapeutic Keywords: cDNA microarray; molecular target; modalities. AURKB; multiple cell-cycle regulator; cytokinesis In an attempt to identify novel molecular targets for breast cancer therapy, we previously analysed the detailed gene-expression profiles of breast cancer cells, which were purified by laser microbeam microdissection, by means of cDNA microarray (Nishidate et al., 2004). In this Correspondence: Dr T Katagiri, Laboratory of Molecular Medicine, report we focused on characterization of a DTL/RAMP Human Genome Center, Institute of Medical Science, The University (denticleless/RA-regulated nuclear matrix associated of Tokyo, 4-6-1, Shirokanedai, Minato-ku, Tokyo 108-8639, Japan. protein) gene that was previously isolated as one of E-mail: [email protected] downregulated genes during retinoic acid-induced differ- 3Current address: Division of Genome Medicine, Institute for Genome Research, The University of Tokushima, Tokushima, Japan entiation of human embryonal carcinoma cell line, NT2 Received 23 May 2007; revised 6 May 2008; accepted 12 May 2008; cells (Cheung et al., 2001) to elucidate its pathophysio- published online 9 June 2008 logic roles in growth of breast cancer cells. DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5673 Results To examine the expression of endogenous DTL/ RAMP protein, we generated a polyclonal antibody Overexpression of DTL/RAMP in breast cancer cells against DTL/RAMP protein and performed western We validated that the elevated the DTL/RAMP blot analysis using cell lysates from eight breast cancer (denticleless/RA-regulated nuclear matrix associated cell lines, HBC4, HBC5, HCC1937, MCF-7, MDA-MB- protein) gene (GenBank accession no. NM_016448) 231, SK-BR-3, T47D and YMB-1. We confirmed a high expression in 10of 12 clinical breast cancer cases by level of DTL/RAMP protein expression in all the eight semiquantitative RT–PCR analysis (Figure 1a) as well breast cancer cell lines examined, but its expression was as cDNA microarray data (Nishidate et al., 2004). hardly detectable in human mammalian epithelial cell Subsequent northern blot analysis showed that an (HMEC) cells (Figure 2a). Additionally, we noticed two approximately 4.5-kb transcript of DTL/RAMP was bands that reacted with anti-DTL/RAMP protein when significantly elevated in all of eight breast cancer cell a longer electrophoresis was performed. Therefore, lines examined (Figure 1b), compared with normal to examine a possible modification of DTL/RAMP, breast. This transcript was most highly expressed in we treated cellular extracts from T47D cells with l- testis, and weakly expressed in bone marrow, thymus phosphatase and analysed the molecular weight of and placenta, but its expression was hardly detectable DTL/RAMP protein by western blot analysis. As the in any of the remaining normal organs we examined upper band was not observed when the cell extracts were (Figure 1c). incubated with phosphatase, we considered that the

12 breast cancer cases 16 102 247 252 302 473 478 502 552 646 769 779 normal MG

DTL/RAMP

FDFT1

kb HBC 4 HBC 5 HCC1937 MCF-7 MDA-MB-231 SK-BR-3 T47D YMB-1 Breast Lung Heart Liver Kidney Bone marrow

9.5 7.5 4.4

2.4

kb Small intestine P.B.L. Heart Brain Placenta Lung Liver muscle Skeletal Kidney Pancreas Spleen Thymus Prostate Testis Ovary Colon Stomach Thyroid Spinal cord node Lymph Trachea Adrenal gland Bone marrow 9.5 7.5

4.4

2.4

Figure 1 Expression pattern of DTL/RAMP in breast cancers and normal human organs. (a) Expression of DTL/RAMP in microdissected tumor cells from breast cancer tissues (16, 102, 247, 252, 302, 473, 478, 502, 552, 646, 769 and 779), compared with normal human tissues (normal, microdissected normal breast ductal cells; MG, mammary gland) by semiquantitative RT–PCR. FDFT1 served as a loading control. (b) Northern blot analysis of the DTL/RAMP in eight breast cancer cell lines (HBC4, HBC5, HCC1937, MCF-7, MDA-MB-231, SK-BR-3, T47D and YMB-1) and normal human tissues including breast, lung, heart, liver, kidney and bone marrow. (c) Northern blot analysis of the DTL/RAMP transcript in various human tissues. PBL indicates peripheral blood leukocytes.

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5674

DTL/RAMP (1-730) ∆ 1 (1-401) ∆ 2 (388-590) HBC4 HBC5 HCC1937 MCF-7 MDA-MB-231 SK-BR-3 T47D YMB-1 HMEC ∆ 3 (438-730) DTL/RAMP IB : HA ACTB lambda-phosphatase -+

∆ lambda-phosphatase 1 -+ kD ∆ 2 p-DTL/RAMP DTL/RAMP p-DTL/RAMP ∆ 3 DTL/RAMP 75 Figure 2 Expression of DTL/RAMP protein in breast cancers. (a) Expression of endogenous DTL/RAMP protein in eight breast cancer cell lines (HBC4, HBC5, HCC1937, MCF7, MDA-MB-231, SK-BR-3, T47D and YMB-1) in comparison with human mammary epithelial cell line (HMEC) cells by western blot analysis using anti-DTL/RAMP polyclonal antibody. (b) Phosphorylation of DTL/RAMP protein in T47D cells. The cell lysates were incubated with/without treatment of l-phosphatase for 2 h at 30 1C. A phosphorylated DTL/RAMP protein is observed as a slowly migrated band as indicated by arrow (p-DTL/RAMP). (c) Determination of a phosphorylated region in the DTL/RAMP protein. The upper panel shows schematic representation of DTL/RAMP fragments (D1, 1–401; D2, 388–590and D3, 438–730amino acids) and full length of DTL/RAMP (1–730amino acids) used for l-phosphatase experiments. The six black boxes in full length of DTL/RAMP indicate WD40repeat domains. The lower panel shows l-phosphatase assay results. The C-terminal region of DTL/RAMP (438–730amino acid) was phosphorylated. The cell lysates were incubated with/without treatment of l-phosphatase as described in (b).

DTL/RAMP protein was possibly phosphorylated in mammary tissues including ductal cells (no. 10441N; breast cancer cells (Figure 2b). Subsequently, to Figure 3b). Furthermore, we performed breast cancer determine the phosphorylated region of the DTL/ tissue-array analysis and verified positive staining of RAMP protein, we made three constructs, each of DTL/RAMP in 33 of 39 infiltrating ductal carcinomas, which was designed to express a part of DTL/RAMP, whereas no staining was observed in 5 normal mammary and performed l-phosphatase assay (Figure 2c, upper tissues including ductal cells (data not shown). Among panel). The results revealed that the phosphorylation the nine normal tissues we examined, its expression was occurred in its C-terminal portion (438–730amino acid; detected in testis, but hardly detectable in heart, liver, Figure 2c, lower panel). kidney, lung, colon, pancreas, skeletal muscle and small intestine in concordance with the result of northern blot analysis (Figure 3c). Immunocytochemical-staining and immunohistochemical- staining analyses To characterize the biological role of DTL/RAMP, we Cell-cycle-dependent expression of DTL/RAMP first examined the subcellular localization of endo- We performed fluorescence-activated cell sorting genous DTL/RAMP in HBC5 cells by immunocyto- (FACS) and western blot analyses to examine the chemical-staining analysis using an anti-DTL/RAMP expression of DTL/RAMP protein at various cell-cycle polyclonal antibody (Figure 3a). The endogenous DTL/ phases using T47D cells after synchronization of the cell RAMP was mainly localized in the nucleus of interphase cycle by aphidicolin treatment. The results showed that cells. After disappearance of nuclear membrane, the the expression of DTL/RAMP was the highest at G1/S protein was stained diffusely within cells during phase (0–6 h), and then strikingly reduced after G2/M prophase to metaphase. Subsequently, it accumulated phase (9–12 h) until next G1 phase (12–30h) (Figures 4a as a series of narrow bars at spindle midzone in the and b). Interestingly, a significant proportion of DTL/ anaphase cells and was finally concentrated at the RAMP protein was modified to the phosphorylated contractile ring when cells were at telophase and form during G2/M phase (9–12 h). Its phosphorylation cytokinesis stages. level was gradually reduced from G2/M to G1 phase We investigated DTL/RAMP expression in clinical (12–30h; Figures 4a and b). Furthermore, to examine breast cancer and normal tissue sections by immuno- the phosphorylation status and expression of DTL/ histochemical staining. We identified positive staining of RAMP after mitotic phase in more detail, we synchro- DTL/RAMP in the cytoplasm and nuclei of two nized T47D breast cancer cells with nocodazole, and different histological subtypes of breast cancer tissue confirmed the expression and phosphorylation of DTL/ sections, papillotubular carcinomas and scirrhous carci- RAMP in T47D breast cancer cells by western blot nomas, whereas no staining was observed in normal analysis using cells at mitotic phase collected by the

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5675 DAPI DTL/RAMP Merge DAPI DTL/RAMP Merge

Interphase anaphase

prophase telophase

metaphase cytokinesis

Papillo-tubular Scirrhous Heart Lung Kidney

010005T 010069T Papillo-tubular Scirrhous Liver Colon Pancreas

MA05-317T 010571T Normal breast duct Skeletal muscle Small intestine Testis

010441N

Figure 3 Immunocytochemical- and immunohistochemical-staining analyses. (a) HBC5 cells were immunocytochemically stained using affinity-purified anti-DTL/RAMP polyclonal antibody (green) and 40,60-diamidine-20-phenylindole dihydrochloride (DAPI; blue) to discriminate nucleus (see the Materials and methods). White arrows indicate localization of DTL/RAMP in midbody of telophase and cytokinesis cells, respectively. (b) Representative images of immunohistochemical staining of DTL/RAMP in breast cancer and normal ductal tissue sections. Cancer cells in cancer tissues were stained: papillotubular (sample nos. 010005T and MA05-317T) and scirrhous (sample nos. 10069T and 010571T) carcinomas, and normal breast tissue (Sample No. 10441N). Black arrow indicates normal mammary ductal cells. (c) Representative images of immunohistochemical staining of DTL/RAMP in normal human tissue sections (heart, lung, kidney, liver, colon, skeletal muscle, pancreas, small intestine and testis). Endogenous DTL/RAMP protein was stained by anti-DTL/RAMP polyclonal antibody. Original magnification, Â 50.

mitotic shake-off method (Dechat et al., 1998). The lization and modiﬁcation, and might play important result showed that most of the DTL/RAMP proteins roles in cell-cycle progression of breast cancer cells. was phosphorylated during mitosis (0–75 min), and then gradually dephosphorylated in a time-dependent man- ner (75–2880min; Figures 4c and d). On the other hand, Effect of DTL/RAMP on cell growth its protein expression level was gradually decreased from To investigate a growth-promoting effect of DTL/ mitosis to G1 phase (720min), but was recovered at the RAMP on breast cancer cells, we knocked down beginning of the second S phase (2880min; Figures 4c the expression of endogenous DTL/RAMP in breast and d). Taken together, we suggest that endogenous cancer cell lines, T47D and HBC4, by means of DTL/RAMP protein showed cell-cycle-dependent loca- RNA interference (RNAi) technique (see Materials

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5676 and methods). Semiquantitative RT–PCR analysis cytokinesis process. We also observed enlargement of detected significant knockdown effect of DTL/RAMP the size of the T47D cells transfected with siDTL/ expression in the cells transfected with psiU6BX-DTL/ RAMP in comparison with those transfected with RAMP-si no. 1 and si no. 2, compared with a control control siEGFP (Figure 5h). Additionally, we observed siRNA construct, psiU6BX-EGFP (Figures 5a and d). the similar morphological alterations in DTL/RAMP- In concordance with the knockdown effect on the depleted HBC4 breast cancer cells as well as T47D cells transcript, colony-formation (Figures 5b and e) and 3- (Supplementary Figure 1). Moreover, we confirmed that (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro- the midbody staining of DTL/RAMP protein was mide (MTT) assays (Figures 5c and f) evidently revealed disappeared in DTL/RAMP-depleted HBC4 breast growth suppression of breast cancer cells by the two cancer cells, but was observed in control-siRNA-treated siRNAs, si no. 1 and si no. 2. These results were cells by immunochemical-staining analysis using DTL/ confirmed by three independent experiments, suggesting RAMP polycloncal antibody (Supplementary Figure 1). that DTL/RAMP is likely to have an important role in To further evaluate the effect of DTL/RAMP depletion the growth of the breast cancer cell. on cell cycle, we performed FACS analysis and Furthermore, we examined morphological alterations confirmed that the proportion of cells at the G2/M of the T47D cells transfected with a DTL/RAMP- phase was significantly higher for the cells treated with specific siRNA oligonucleotide (siDTL/RAMP), and DTL/RAMP-siRNA (23.51%) than for those treated confirmed the significant knockdown effect at the with EGFP-siRNA (13.77%; Figure 5i), indicating protein level (Figure 5g). Interestingly, we observed that accumulation of G2/M-arrested cells by the knockdown its knockdown led to appearance of the intercellular of DTL/RAMP. Similar results were obtained when we bridges between two cells at a cell-division stage used HBC4 cells (data not shown). These findings (Figure 5h), indicating dysfunction at the late stage of indicate that the absence of DTL/RAMP caused the

036912(hr) Data.002(0h) Data 003(3h) Data 004(6h) Data 005(9h) Data 006(12h) 500 500 500 500 500

400 400 400 400 400 M1 300 300 300 M1 300 M1 300 M2 M1 M2 M2 M3 200 M2M3 200 M1 M3 200 M2M3 200 200 M3 Counts Counts Counts Counts Counts

100 100 100 100 100

0 0 0 0 0 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 FL2-A FL2-A FL2-A FL2-A FL2-A G1 : 84.87 G1 : 38.85 G1 : 26.74 G1 : 27.84 G1 : 51.43 S : 8.36 S : 55.23 S : 50.94 S : 14.99 S : 10.18 G2/M : 5.49 G2/M : 7.55 G2/M : 22.09 G2/M : 56.61 G2/M : 37.85 15 18 24 27 30 (hr) Data.007(15h) Data.008(18h) Data.009(24h) Data.010(27h) Data.011(30h) 500 500 500 500 500

400 400 400 400 400

300 300 300 M1 300 M1 M2 300 M1 M2 M1 M1 M3 200 M2M3 200 M2M3 200 M3 200 200 M2 M3 Counts Counts Counts Counts Counts

100 100 100 100 100

0 0 0 0 0 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 FL2-A FL2-A FL2-A FL2-A FL2-A G1 : 70.05 G1 : 78.65 G1 : 82.37 G1 : 80.86 G1 : 78.81 S : 7.20 S : 5.77 S : 6.15 S : 6.31 S : 9.93 G2/M : 21.73 G2/M : 14.70 G2/M : 10.84 G2/M : 12.26 G2/M : 10.58

0 3 6 9 12 15 18 24 27 30 (hr)

DTL/RAMP

ACTB

Figure 4 Cell-cycle dependent expression of DTL/RAMP. (a) Fluorescence-activated cell sorting (FACS) analysis showed population of T47D cells collected every 3 h up to 30h after synchronization with treatment of aphidicolin hours (0, 3, 6, 9, 12, 15, 18, 21, 24, 27 and 30h). (b) Western blot analysis of DTL/RAMP expression at each time point in T47D cells synchronized with aphidicolin treatment. The expression levels of endogenous DTL/RAMP at each cell-cycle phase, detected by western blot analysis. ACTB served as a quantitative control of protein level. (c) FACS analysis showed population of T47D cells collected at each cell cycle (0, 60, 75, 90, 105, 120, 240, 360, 720 and 2880 min) after synchronization with treatment of nocodazole. (d) Western blot analysis of DTL/RAMP expression at each time point in T47D cells synchronized with nocodazole treatment. ACTB served as a quantitative control of protein level.

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5677 0 60 7590 105 (min) Data001 Data004 Data005 Data006 Data007 400 400 400 400 400

320 320 320 320 320 240 M1 240 M2 240 M3 240 M3 240 M3 M1 M1 160 160 M2 160 Counts Counts Counts 160 Counts 160 Counts M1 M2 M2M3 80 80 80 80 80 M1 M2 0 0 0 0 0 0 200 400 600 8001000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 8001000 FL2A FL2A FL2A FL2A FL2A G1 : 3.07 G1 : 3.84 G1 : 3.26 G1 : 8.17 G1 : 20.09 S : 4.68 S : 3.71 S : 4.86 S : 7.30 S : 6.47 G2/M : 86.26 G2/M : 88.27 G2/M : 87.77 G2/M : 77.34 G2/M : 64.28

(12hr) (48hr) 120 240 360720 2880 (min) Data008 Data009 Data010 Data011 Data012 400 400 400 400 400

320 320 320 320 320 M1M2 M3 240 240 240 M1 240 M2 240 M3 M1 M2 M1 M3 M1 M3

Counts 160 160 160 Counts 160 Counts 160 Counts M2 M3 Counts M2 80 80 80 80 80

0 0 0 0 0 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 FL2A FL2A FL2A FL2A FL2A G1 : 27.58 G1 : 44.28 G1 : 48.61 G1 : 46.88 G1 : 49.53 S : 6.89 S : 7.60 S : 6.96 S : 5.97 S : 12.23 G2/M : 55.92 G2/M : 37.08 G2/M : 32.46 G2/M : 36.95 G2/M : 26.30

(12hr) (48hr) 0 60 75 90 105 120 240 360 720 2880 (min)

DTL/RAMP

ACTB

Figure 4 Continued.

failure of cytokinesis, resulted in arrest of G2/M phase In addition, we found six consensus phosphorylation cells and then induced cell death. sites for AURKB in full-length amino-acid sequence of To further confirm the growth promoting effect of DTL/RAMP protein according to previous reports DTL/RAMP, we established NIH3T3-derivative cells ([R]X[T/S] and [R/K]X[T/S][I/L/V]; Cheeseman et al., that stably expressed exogenous DTL/RAMP (DTL/ 2002; Ohashi et al., 2006). Particularly, among them, RAMP-1, -2, -3 and -4). Figure 6a showed that Flag- five possible phosphorylation sites were found within the tagged DTL/RAMP protein was observed at high level C-terminal region of DTL/RAMP (438–730amino acid) in these four derivate clones. Subsequent MTT assays where phosphorylation was observed (Figure 2c). showed that the four DTL/RAMP-stable derivative cells Hence, we examined a possible interaction of DTL/ (DTL/RAMP-1, -2, -3 and -4) grew significantly faster RAMP protein with AURKB in breast cancer cells. We than those transfected with mock plasmid (Mock-1, -2, first compared the expression patterns of DTL/RAMP and –3; Figure 6b), suggesting a growth-enhancing effect and AURKB by semiquantitative RT–PCR and western of DTL/RAMP. These findings suggest that overexpres- blot analysis, and confirmed the upregulation of both sion of DTL/RAMP might be involved in the growth of DTL/RAMP and AURKB in 7 of 12 breast cancer breast cancer cells. clinical cases and all of 8 breast cancer cell lines examined (Figures 7a and b). Subsequently, we investigated an interaction between endogenous DTL/RAMP DTL/RAMP protein was regulated by Aurora kinase-B and AURKB proteins in T47D cells, in which these two We described above the phosphorylation of DTL/ proteins were expressed abundantly. Figure 7c shows RAMP and its concentration at the contractile ring that endogenous DTL/RAMP was coprecipitated with when cells were at telophase and cytokinesis stages in endogenous AURKB in breast cancer cells. Moreover, breast cancer cells. As Aurora kinase-B (AURKB) is immunocytochemical-staining experiments confirmed known to be a chromosome passenger protein that both endogenous proteins colocalized at midbody in moves from centrosomes to midzone spindle in ana- cytokinesis of T47D breast cancer cells (Figure 7d). To phase and to the midbody in telophase and cytokinesis further investigate whether DTL/RAMP is phosphory- in HeLa cells (Adams et al., 2001; Terada, 2001; lated by AURKB, we performed in vitro kinase assay Carmena and Earnshaw, 2003), we considered their using a purified full-length DTL/RAMP (1–730amino similar subcellular localization at some cell-cycle phases. acids) recombinant protein and the full-length AURKB

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5678 p<0.001 siEGFP si#1 si#2

1.0 siEGFP si#1 si#2 0.8 DTL/RAMP 0.6 0.4

relative ratio 2MG 0.2 0 siEGFP si#1 si#2

P<0.001

siEGFP si#1 si#2 1.0 0.8 siEGFP si#1 si#2 0.6 DTL/RAMP 0.4 relative ratio 0.2 2MG 0 si EGFP si #1 si #2

siEGFP siDTL/RAMP siEGFP siDTL/RAMP

DTL/RAMP

ACTB

siEGFP siDTL/RAMP 600 600 480 480 360 360 240 240 Counts M1 Counts M1 M2 120 M3 120 M2 M3 0 0 0 200 400 600 8001000 0 200 400 600 8001000 G1 :85.87 G1 :58.81 S :3.42 S :13.05 G2/M:9.51 G2/M:23.51 Figure 5 Growth-inhibitory effects of siRNA designed to reduce expression of DTL/RAMP in breast cancer cells. (a) Semiquantitative RT–PCR showing suppression of endogenous expression of DTL/RAMP by DTL/RAMP-specific siRNAs (si no. 1 and si no. 2) in breast cancer cell line, T47D. b2MG served as a loading control. (b) Colony-formation assay demonstrating a decrease in the number of colonies by knockdown of DTL/RAMP in T47D cells. (c) 3-(4,5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay demonstrating a decrease in the number of cells by knockdown of DTL/RAMP in T47D cells (si no. 1 and si no. 2; Po0.001, respectively; unpaired t-test). (d) Semiquantitative RT–PCR showing suppression of endogenous expression of DTL/RAMP by DTL/RAMP-specific siRNAs (si no. 1 and si no. 2) in breast cancer cell line, HBC4. b2MG served as a loading control. (e) Colony-formation assay demonstrating a decrease in the number of colonies by knockdown of DTL/ RAMP in HBC4 cells. (f) MTT assay demonstrating a decrease in the number of cells by knockdown of DTL/RAMP in HBC4 cells (si no. 1 and si no. 2; Po0.001, respectively; unpaired t-test). (g) Silencing of endogenous DTL/RAMP expression by treatment of siRNA oligonucleotides was confirmed by western blot analysis. ACTB served as a loading control. (h) Morphological changes of T47D cells transfected with siDTL/RAMP. siEGFP was used as a control siRNA. Upper and lower panels: observations under the microscope or by immunocytochemical staining, respectively. T47D cells treated with siRNAs were stained with 40,60-diamidine-20-phenylindole dihydrochloride (DAPI) and phalloidin to distinguish nucleus from cytoplasm. (i) Fluorescence-activated cell sorting (FACS) analysis of T47D cells transfected with siDTL/RAMP or siEGFP as a control.

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5679 Mock DTL/RAMP endogenous DTL/RAMP may be stabilized in a late phase of mitosis when it is phosphorylated by AURKB #1 #2 #3 #1 #6 #9 #12 or incorporated into some AURKB protein complex.

DTL/RAMP

Discussion ACTB Through identification and characterization of cancer- 10 related genes and their products, molecular targeting Mock1 drugs for cancer therapy have been developed in the last Mock2 two decades, but the proportion of patients who are able 8 Mock3 to have a benefit by presently available treatments is DTL/RAMP-#1 still very limited (Bange et al., 2001; Navolanic and DTL/RAMP-#6 McCubrey, 2005). Therefore, it is urgent to develop new 6 DTL/RAMP-#9 anticancer agents that will be highly specific to DTL/RAMP-#12 malignant cells and have the minimal risk of adverse 4 reactions. In this study, we demonstrated the upregula- Relative ratio tion of DTL/RAMP in clinical breast cancer cases and cell lines, but was hardly detectable in any normal 2 human tissues examined except a low level of expression in a few organs. DTL/RAMP gene encodes a putative 730-amino-acid 0 day1day2day3 protein that contains six highly conserved WD40-repeat domains and a consensus nuclear-localization signal at Figure 6 Growth-promoting effect of exogenous DTL/RAMP on NIH3T3 cells. (a) Western blot analysis of cells expressing N terminus. Our results also demonstrated that DTL/ exogenous DTL/RAMP at high level or those transfected with RAMP protein was mainly localized in the nucleus of mock vector. Exogenous introduction of DTL/RAMP expression interphase cells, accumulated as a series of narrow bars were validated with an anti-Flag-tag monoclonal antibody. ACTB at spindle midzone in the anaphase cells, and was finally served as a loading control. (b) In vitro growth assay of NIH3T3- concentrated at the contractile ring in telophase and DTL/RAMP cells. The growth measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay demon- cytokinesis stages. These findings suggest the important strates that four NIH3T3- DTL/RAMP clones (DTL/RAMP-1, -2, role of this protein in cell-cycle progression. -3 and -4; solid lines) grew significantly faster than three NIH3T3- We demonstrated by means of the siRNA technique Mock clones (Mock-1, -2, and -3; dashed lines). x Axis indicates that knocking down of the endogenous DTL/RAMP day points after the seeding and y axis indicates relative growth rate that was calculated in absorbance by comparison with the expression significantly suppressed the cell growth of absorbance value of day 1 as a control. Assays were performed breast cancer cell lines, T47D and HBC4, due to three times in triplicate wells. abnormal cell division and subsequent cell death, probably due to the dysfunction in the cytokinetic process. We have also demonstrated that the proportion of cells with a larger size was increased in the siDTL/ recombinant protein (see Materials and methods), and RAMP-transfected cells although we did not find an found phosphorylation of the DTL/RAMP protein by increase of the multinucleated cells. As it was reported AURKB in vitro (Figure 7e). On the other hand, we that an inactivation of DTL/RAMP induced p53 were unable to detect the phosphorylated DTL-RAMP stabilization in unstressed HeLa cells (Banks et al., protein by kinase-dead AURKB mutant (D200A; 2006), the accumulation of G2/M cells by knockdown of Figure 7e). To further investigate a possible role of the DTL/RAMP might be due to activation of the interaction between DTL/RAMP and AURKB proteins checkpoint system by p53 in breast cancer cell line and its phosphorylation by AURKB, we transfected T47D. Moreover, introduction of DTL/RAMP into siRNA-AURKB (siAURKB) into T47D cells, and then NIH3T3 cells significantly enhanced cell growth. Hence, performed western blot and semiquantitative RT–PCR we further examined biological roles of DTL/RAMP in analyses. We observed the significant decrease of total breast cancer cells by identification of its interacting DTL/RAMP protein as well as the phosphorylated protein(s). DTL/RAMP protein in T47D cells transfected with Due to similarity of the subcellular localization at siAURKB in comparison with those with a control some cell-cycle phases and its coexpression in breast siEGFP (Figure 7f), whereas its transcriptional level was cancer cells, we focused on AURKB, serine/threonine not reduced in AURKB-depleted cells. Furthermore, we kinase as a candidate for the DTL/RAMP interacting found that the depletion of AURKB in T47D cells by protein. We demonstrated the in vivo interaction and siRNA induced accumulation of G2/M-arrested cells colocalization of DTL/RAMP with endogenous compared with treatment of a control-siEGFP (Supple- AURKB, and in vitro phosphorylation of DTL/RAMP mentary Figure 2) as similar as the depletion of DTL/ as well as its possible stabilization by AURKB through RAMP (Figures 5g and i). These findings suggested that the RNAi experiments although it remains to be

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5680 12 breast cancer cases 31 86 175 214 247 491 508 521 552 762 780 MG 4 AURKB IP DTL/RAMP

FDFT1 W.C.L IgG anti- DTL/RAMP

DTL/RAMP IB

HBC4 HBC5 HCC1937 MCF-7 MDA-MB-231 SK-BR-3 T47D YMB-1 HMEC AURKB AURKB

DTL/RAMP

ACTB

GST-DTL/RAMP + + - + DAPI DTL/RAMP - - + -

GST siEGFP siAURKB AURKB(WT) + - + - DTL/RAMP AURKB(D200A) - + - -

kDa AURKB DTL/RAMP 100 ACTB

DTL/RAMP

AURKB AURKB 37 AURKB Merge ACTB

Figure 7 Expression of DTL/RAMP protein was regulated by Aurora kinase-B (AURKB). (a) Semiquantitative RT–PCR experiments for DTL/RAMP and AURKB expressions in 12 clinical breast cancer cases (4, 31, 86, 175, 214, 247, 491, 508, 521, 552, 762 and 780). FDFT1 was used as a quantitative control. MG indicates mammary gland. (b) Western blot analysis for eight breast cancer cell lines (HBC4, HBC5, HCC1937, MCF-7, MDA-MB-231, SK-BR-3, T47D and YMB-1) and human mammary epithelial cell line (HMEC). ACTB was used as a quantitative control. (c) Co-immunoprecipitation of endogenous DTL/RAMP and endogenous AURKB proteins. Cell lysates from T47D cells were immunoprecipitated with either anti-DTL/RAMP rabbit polyclonal or normal rabbit immunoglobulin G (IgG). Immunoprecipitates were immunoblotted using a mouse anti-AURKB antibody. W.C.L indicates whole cell lysates. (d) Colocalization of endogenous DTL/RAMP and endogenous AURKB in T47D cells. Endogenous DTL/RAMP protein (green) colocalized endogenous AURKB (red) at midbody of cytokinetic cells (white arrow). (e) In vitro kinase assay was performed with puriﬁed full-length recombinant protein of GST-tagged DTL/RAMP (105 kDa, including GST-tag). We added DTL/ RAMP recombinant protein to the reaction mixture including wild-type AURKB (WT-AURKB) or kinase-dead AURKB (D200A- AURKB) (see text). The upper arrow indicates phosphorylated GST-DTL/RAMP protein. The lower arrow indicates autopho- sphorylated AURKB protein. (f) The expression levels of endogenous DTL/RAMP and AURKB proteins, detected by western blot analysis in T47D cells transfected with siRNA oligonucleotides against AURKB. ACTB served as a quantitative control of protein level (upper panels). The expression levels of endogenous DTL/RAMP and AURKB, detected by semiquantitative RT–PCR analysis in T47D cells transfected with siRNA oligonucleotides against AURKB. ACTB served as a quantitative control of transcriptional level (lower panels).

elucidated whether DTL/RAMP is phosphorylated by nent of the CUL4-DDB1 ubiquitin E3 ligase complex AURKB in vivo. Additionally, it was reported that (Higa et al., 2006a, b; Sansam et al., 2006), suggesting knockdown of AURKB also suppressed growth of the multiple roles of DTL/RAMP in the cell-cycle HeLa cells because of cytokinesis defects (Severson progression. Taken together, as inhibition of their et al., 2000; Goto et al., 2003) alike depletion of DTL/ association may lead to cell death following the failure RAMP. Together, we demonstrate here for the ﬁrst time of cytokinesis in breast cancer cells, the inhibitor for that interaction of DTL/RAMP and AURKB might their association would be a possible valuable target to play an important role in cytokinesis. Furthermore, develop agents against breast cancer. Furthermore, it is DTL/RAMP is reported to be required in the initiation notable that our cDNA microarray data identiﬁed of a radiation-induced G2/M checkpoint as a compo- upregulation of DTL/RAMP commonly in many types

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5681 of clinical cancers including bladder cancer, cholangio- coding sequence of DTL/RAMP cDNA fragment was cloned carcinoma, lung cancers and renal cell carcinoma as well into the SmaI sites of pGEX-6P-1 expression vector as breast cancers (data not shown). These results showed (GE Healthcare, Buckinghamshire, UK). These constructs that this gene should serve as a valuable target for were confirmed by DNA sequencing (ABI3700; PE Applied development of anticancer agents for a wide range of Biosystems, Foster, CA, USA). human cancers. Generation of anti-DTL/RAMP polyclonal antibody Plasmid designed to express a part of DTL/RAMP (438–730 amino acids) with His-tag at its C-terminus was prepared using Materials and methods pET21 vector (Merck, Novagen, Madison, WI, USA). The recombinant peptide (36 kDa) was expressed in Escherichia Breast cancer cell lines and clinical samples coli, BL21 codon-plus (Stratagene, La Jolla, CA, USA), and Human breast cancer cell lines, HCC1937, MCF-7, MDA- purified using Ni-NTA resin (Qiagen) according to the MB-231, SK-BR-3, T47D and YMB-1 were purchased from supplier’s protocol. The protein was inoculated into rabbits, American Type Culture Collection (ATCC, Rockville, MD, and subsequently the immune sera were purified on antigen USA). HMEC were purchased from Cambrex Bio Science affinity columns using Affigel 15 gel (Bio-Rad, Hercules, CA, Walkersville Inc. (CAMBREX, Walkersville, MD, USA). USA), according to supplier’s instructions. HBC4 and HBC5 cell lines were kindly provided from Dr Takao Yamori of Division of Molecular Pharmacology, Western blot analysis Cancer Chemotherapy Center, Japanese Foundation for Cells were lysed with NP-40lysis buffer (50m M Tris-HCL (pH Cancer Research. All cells were cultured according to previous 8.0), 150 mM NaCL, 0.5% NP-40) including 0.1% protease reports (Park et al., 2006; Lin et al., 2007; Shimo et al., 2007). inhibitor cocktail III (Calbiochem, San Diego, CA, USA). The Tissue samples from surgically resected breast cancers and proteins were mixed with SDS-sample buffer and boiled before their corresponding clinical information were obtained from loading at 10% SDS–polyacrylamide gel electrophoresis gel. Department of Breast Surgery, Cancer Institute Hospital, After electrophoresis, the proteins were blotted onto nitrocel- Tokyo after obtaining written informed consent. lulose membrane (GE Healthcare), and blocked with 4% BlockAce solution (Dainippon Pharmaceutical, Osaka, Semiquantitative RT–PCR analysis Japan). The membranes were incubated with anti-DTL/ Total RNAs were extracted from each of microdissected breast RAMP polyclonal antibody at 1:100 dilutions or anti-AURKB cancer clinical samples, microdissected normal breast ductal polyclonal antibody (Abcam, Cambridge, UK) at 1:100 cells and breast cancer cell lines using RNeasy Micro Kits dilutions. Finally the membrane was incubated with HRP- (Qiagen, Valencia, CA, USA), and polyA ( þ ) RNAs isolated conjugated secondary antibody and protein-bands were from mammary gland purchased from Takara Clontech visualized by ECL (GE Healthcare). (Kyoto, Japan) as described previously (Nishidate et al., 2004). Subsequently, T7-based amplification and reverse l-Phosphatase assay transcription were carried out as described previously (Nishi- T47D cells were lysed by NP-40lysis buffer, and then were date et al., 2004). We prepared appropriate dilutions of each treated for 2 h at 30 1C with 400 units of l-phosphatase (New single-stranded cDNA for subsequent PCR by monitoring England Biolabs, Beverly, MA, USA) in phosphatase buffer FDFT1 Farnesyl-diphosphate farnesyltransferase 1 ( )asa containing 50m M Tris-HCL (pH 7.5), 0.1 mM Na -EDTA, quantitative control because these showed the smallest 2 5mM dithiothreitol, 2 mM MgCL2 and 0.01% Brij-35. (relative ratio of breast cancer sample/normal control) Furthermore, to define the phosphorylated site(s) of DTL/ fluctuations in our breast cancer-microarray data (Nishidate RAMP protein, HEK293T cells were transiently transfected et al., 2004). The primer sequences are listed in Supplementary with pCAGGS-DTL-D1-HA, -D2-HA and -D3-HA, respec- Table 1. tively, using FuGENE 6 (Roche, Basel, Switzerland) according to the manufacturer’s instructions. Forty-eight hours after the Northern blot analysis transfection, the cells were lysed by NP-40lysis buffer and then Northern blot membrane for breast cancer cell lines was were treated with l-phosphatase (New England Biolabs) as prepared as described previously (Park et al., 2006). Human described above. multiple-tissue northern blots (Takara Clontech) were hybri- 32 dized with [a P]-dCTP-labeled PCR products of DTL/RAMP Immunocytochemical staining prepared by reverse transcription (RT)–PCR (see below). HBC5 or T47D cells were fixed with phosphate-buffered saline Prehybridization, hybridization and washing were performed (PBS)(À) containing 4% paraformaldehyde for 20min, and according to the supplier’s recommendations. The blots were rendered permeable with PBS(À) containing 0.1%. Triton autoradiographed with intensifying screens at À80 1C for 14 X-100 at room temperature for 2 min. Subsequently, the cells days. Specific probes (1,172bp) for DTL/RAMP were prepared were covered with 3% bovine serum albumin in PBS( ) for 1 h 0 À by RT–PCR using the following primer set: 5 -GCA at room temperature to block nonspecific hybridization, and 0 0 ATCTGCTATGTCAGCCAAC-3 and 5 -CAGGATCA then were incubated with affinity-purified rabbit anti-DTL/ 0 GCTCAAA GTCTGACA-3 for DTL/RAMP. RAMP antibody at 1:100 dilutions or with anti-AURKB monoclonal antibody (Abcam) at 1:100 dilutions and Construction of expression vectors anti-DTL/RAMP antibody at 1:100 dilution. After washing To construct DTL/RAMP expression vectors, the entire with PBS(À), cells were stained by an Alexa488-conjugated coding sequence of DTL/RAMP cDNA was inserted into anti-rabbit secondary antibody (Molecular Probe, Eugene, the NotI and XhoI sites of pCAGGSnHAc expression vector in OR, USA) at 1:1000 dilutions or an Alexa594-conjugated frame with C-terminal HA-tag. Furthermore, to determine the anti-mouse secondary antibody (Molecular Probe) at 1:1000 phosphorylated regions of DTL/RAMP protein, we prepared dilutions. Nuclei were counter-stained with 40,60-diamidine-20- the deletion constructs. For in vitro kinase assay, the entire phenylindole dihydrochloride (DAPI). Fluorescent images

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5682 were obtained under a TCS SP2 AOBS microscope were transfected with those siRNAs using Lipofectamin (Leica, Tokyo, Japan). RNAiMAX (Invitrogen) in Optimem (Invitrogen) medium according to the instructions of manufacture. After 96 h, we Immunohistochemical-staining analysis confirmed knockdown effect of DTL/RAMP and AURKB by Slides of paraffin-embedded breast cancer specimens (sample western blot analysis and semiquantitative RT–PCR analyses nos. 010005T, MA05-317T, 010069T and 010571T), normal as described above. Moreover, morphological changes of the mammary tissue (sample no. 10441N), breast cancer tissue- T47D and HBC4 cells were observed by microscopy and by arrays (COSMOBIO, Tokyo, Japan) and other normal human immunocytochemical-staining analysis using AlexaFluor594 tissues (lung, heart, liver, kidney, colon, pancreas, skeletal Phalloidin (Molecular Probe) and DTL/RAMP polyclonal muscle, small intestine and testis; Biochain, Hayward, CA, antibody. To evaluate the effect of DTL/RAMP depletion on USA) were processed for antigen retrieval by autoclave in cell cycle by siRNA, we performed FACS analysis 72 h after accordance with the previous report (Shimo et al., 2007). the transfection of siRNA according to the method described Tissue sections were incubated with anti-DTL/RAMP poly- above. clonal antibody as primary antibody at 1:50followed by horseradish peroxidase-cojugated secondary antibody (Dako- Establishment of NIH3T3 cells stably expressing DTL/RAMP Cytomation, Carpinteria, CA, USA). Specific immunostaining 0 Flag-tagged DTL/RAMP expression vector (pCAGGSn3FC- was visualized with peroxidase substrate (3, 3 -diaminobenzi- DTL/RAMP) or mock vector (pCAGGSn3FC) was trans- dine tetrahydrochloride; DAKO liquid DAB chromogen; fected into NIH3T3 cells using FuGENE6 (Roche), and then DakoCytomation). Finally, tissue specimens were stained with incubated in the culture medium containing 0.9 mg/ml. of hematoxylin to discriminate nucleus from cytoplasm. neomycin (geneticin; Invitrogen). Three weeks later, we established four independent DTL/RAMP stably expressing Fluorescence-activated cell sorting analysis clones and designated them as follows: NIH3T3-DTL/RAMP- T47D cells were synchronized their cell cycle by treatment with 1, -2, -3 and 4, and NIH3T3-Mock-1, -2 and -3. For growth 0.5 mg/ml of aphidicolin (Sigma-Aldrich Japan KK, Tokyo, assay, 5000 cells of each of DTL/RAMP-expressing transfor- Japan) for 16 h. Subsequently, the cells were collected every 3 h mant or Mock-transformant as a control was seeded. Cell up to 30h, and fixed with 70% ethanol, and then kept at 4 1C viability was quantified with MTT assay every day. until their use. The cells were incubated with 10mg/ml RNaseI in PBS at 37 1C for 30min and stained with 50 mg of propidium iodide at room temperature for 30min. The cell suspensions at Statistical analysis each time-point were analysed with FACscan (Becton Dick- Statistical significance was calculated by Student’s t-test, using inson, Franklin Lakes, NJ, USA). Additionally, to examine Statview 5.0software (SAS Institute, Cary, NC, USA). A expression levels of endogenous DTL/RAMP protein in each difference of Po0.05 was considered to be statistically cell-cycle phase, we performed western blot at every 3 h using significant. anti-DTL/RAMP polyclonal antibody. Furthermore, to examine the phosphorylation status of DTL/RAMP after mitotic Co-immunoprecipitation and western blot analyses phase in more detail, we synchronized T47D breast cancer T47D cells were lysed in immunoprecipitation buffer (50m M cells with treatment of 0.15 mg/ml nocodazole for 18 h, and Tris-HCL (pH 7.5), 150m M NaCL, 0.5% NP-40, 50 mM harvested mitotic cells at each time point (0, 60, 75, 90, 105, NaF, 1 mM NaVO3 and 1 mM DTT) in the presence of 120, 240, 360, 720 and 2880 min) by mitotic shake-off methods protease inhibitor (Calbiochem). Cell lysate was precleared by (Dechat et al., 1998). We confirmed the phosphorylation of incubation in 750 ml of protein G-agarose beads with 7.5 mgof DTL/RAMP in T47D breast cancer cells at prometaphase by normal mouse immunoglobulin G (IgG) at 4 1C for 3 h. l-phosphatase assay described above. The supernatants were incubated at 4 1C with 30 mgof anti-DTL/RAMP rabbit polyclonal antibody or rabbit normal Gene-silencing effect by small interfering RNA IgG for 1 h, and then 100 ml of protein G-agarose beads were We had previously established a vector-based RNAi expres- added. After the beads were collected from each sample and sion system using psiU6BX3.0siRNA expression vector washed five times with 1 ml of immunoprecipitation buffer, (Shimokawa et al., 2003). The siRNA expression vectors they were eluted by 30 ml of Laemmli sample buffer and boiled against DTL/RAMP (psiU6BX3.0-DTL/RAMP) and EGFP for 5 min. (psiU6BX3.0-EGFP) as a negative control were prepared as described previously (Park et al., 2006). The target sequence of In vitro kinase assay synthetic oligonucleotides for siRNA were as follows: si no. 1, One microgram of wild type or kinase-dead (D200A) of 50-GATCATGTCTCCGAGAAAA-30; si no. 2, 50-GGAAGC 0 0 AURKB (Upstate, Temecula, CA, USA) was incubated in CATAGAATTGCTC-3 ; siEGFP, 5 -GAAGCAGCACGAC 20 ml kinase assay buffer (50m M Tris-HCL (pH 7.5), 10m M TTCTTC-30. Each siRNA expression vector was transfected MgCL2,2mM dithiothreitol, 1 mM EGTA, 0.01% Brij35, with FuGENE6 (Roche) into T47D or HBC4 and these cells 500 mM ATP), and then supplemented with 5 mCi of [32P-g] were cultured with a medium containing 0.7 or 0.4 mg/ml of ATP (GE Healthcare). For the substrate, we added 1 mg of the neomycin (Geneticin; Invitrogen, Carlsbad, CA, USA) for full length of DTL/RAMP (1–730amino acids) recombinant 7 days. Cell growth was analysed by MTT assay and colony- protein into the reaction solutions. After 10min incubation at formation assay as described previously (Park et al., 2006). 30 1C, the reactions were terminated by addition of SDS Furthermore, we used siRNA oligonucleotides (Sigma-Aldrich sample buffer. Japan KK) due to its high transfection efficiency to further verify the knockdown effects of DTL/RAMP and AURKB on cell morphology. The sequences targeting each gene were as Acknowledgements follows: siDTL/RAMP, 50-ACUCCUACGUUCUCUAUUA-30; siAURKB, 50-AAGGUGAUGGAGAAUAGCAGU-30;siEGFP We thank Ms Kyoko Kijima for excellent technical assis- (control), 50-GCAGCACGACUUCUUCAAG-30. T47D cells tances.

Oncogene DTL/RAMP as a molecular-target for breast cancer T Ueki et al 5683 References

Adams RR, Carmena M, Earnshaw WC. (2001). Chromosomal Nagayama S, Iiizumi M, Katagiri T, Toguchida J, Nakamura Y. passengers and the (aurora) ABCs of mitosis. Trends Cell Biol 11: (2004). Identification of PDZK4, a novel human gene with PDZ 49–54. domains, that is upregulated in synovial sarcomas. Oncogene 23: Bange J, Zwick E, Ullrich A. (2001). Molecular targets for breast 5551–5557. cancer therapy and prevention. Nat Med 7: 548–552. Navolanic PM, McCubrey JA. (2005). Pharmacological breast cancer Banks D, Wu M, Higa LA, Gavrilova N, Quan J, Ye T et al. (2006). therapy (review). Int J Oncol 27: 1341–1344. L2DTL/CDT2 and PCNA interact with p53 and regulate p53 Nishidate T, Katagiri T, Lin ML, Mano Y, Miki Y, Kasumi F et al. polyubiquitination and protein stability through MDM2 and (2004). Genome-wide gene-expression profiles of breast-cancer cells CUL4A/DDB1 complexes. Cell Cycle 5: 1719–1729. purified with laser microbeam microdissection: identification of Carmena M, Earnshaw WC. (2003). The cellular geography of aurora genes associated with progression and metastasis. Int J Oncol 25: kinases. Nat Rev Mol Cell Biol 4: 842–854. 797–819. Cheeseman IM, Anderson S, Jwa M, Green EM, Kang J, Yates III JR Ohashi S, Sakashita G, Ban R, Nagasawa M, Matsuzaki H, Murata Y et al. (2002). Phospho-regulation of kinetochore-microtubule et al. (2006). Phospho-regulation of human protein kinase Aurora- attachments by the Aurora kinase Ipl1p. Cell 111: 163–172. A: analysis using anti-phospho-Thr288 monoclonal antibodies. Cheung WM, Chu AH, Chu PW, Ip NY. (2001). Cloning and Oncogene 25: 7691–7702. expression of a novel nuclear matrix-associated protein that is Park JH, Lin ML, Nishidate T, Nakamura Y, Katagiri T. (2006). regulated during the retinoic acid-induced neuronal differentiation. PDZ-binding kinase/T-LAK cell-originated protein kinase, a J Biol Chem 276: 17083–17091. putative cancer/testis antigen having an oncogenic activity in breast Dechat T, Gotzmann J, Stockinger A, Harris CA, Talle MA, Siekierka cancer. Cancer Res 66: 9186–9195. JJ et al. (1998). Detergent-salt resistance of LAP2alpha in interphase Parkin DM, Bray F, Ferlay J, Pisani P. (2005). Global cancer statistics, nuclei and phosphorylation-dependent association with chromo- 2002. CA Cancer J Clin 55: 74–108. somes early in nuclear assembly implies functions in nuclear Petricoin III EF, Hackett JL, Lesko LJ, Puri RK, Gutman SI, structure dynamics. EMBO J 17: 4887–4902. Chumakov K et al. (2002). Medical applications of microarray Goto H, Yasui Y, Kawajiri A, Nigg EA, Terada Y, Tatsuka M et al. technologies: a regulatory science perspective. Nat Genet 32(Suppl): (2003). Aurora-B regulates the cleavage furrow-specific vimentin 474–479. phosphorylation in the cytokinetic process. J Biol Chem 278: Sansam CL, Shepard JL, Lai K, Ianari A, Danielian PS, Amsterdam A 8526–8530. et al. (2006). DTL/CDT2 is essential for both CDT1 regulation and Higa LA, Banks D, Wu M, Kobayashi R, Sun H, Zhang H. (2006a). the early G2/M checkpoint. Genes Dev 20: 3117–3129. L2DTL/CDT2 interacts with the CUL4/DDB1 complex and PCNA Severson AF, Hamill DR, Carter JC, Schumacher J, Bowerman B. and regulates CDT1 proteolysis in response to DNA damage. Cell (2000). The aurora-related kinase AIR-2 recruits ZEN-4/CeMKLP1 Cycle 5: 1675–1680. to the mitotic spindle at metaphase and is required for cytokinesis. Higa LA, Wu M, Ye T, Kobayashi R, Sun H, Zhang H. (2006b). Curr Biol 10: 1162–1171. CUL4-DDB1 ubiquitin ligase interacts with multiple WD40-repeat Shimo A, Nishidate T, Ohta T, Fukuda M, Nakamura Y, Katagiri T. proteins and regulates histone methylation. Nat Cell Biol 8: (2007). Elevated expression of PRC1, protein regulator of cytokin- 1277–1283. esis 1, involved in the growth of breast cancer cells. Cancer Sci 98: Hirota E, Yan L, Tsunoda T, Ashida S, Fujime M, Shuin T et al. 174–181. (2006). Genome-wide gene expression profiles of clear cell renal cell Shimokawa T, Furukawa Y, Sakai M, Li M, Miwa N, Lin YM et al. carcinoma: identification of molecular targets for treatment of renal (2003). Involvement of the FGF18 gene in colorectal carcinogenesis, cell carcinoma. Int J Oncol 29: 799–827. as a novel downstream target of the beta-catenin/T-cell factor Lin ML, Park JH, Nishidate T, Nakamura Y, Katagiri T. (2007). complex. Cancer Res 63: 6116–6120. Involvement of maternal embryonic leucine zipper kinase (MELK) Terada Y. (2001). Role of chromosomal passenger complex in in mammary carcinogenesis through interaction with Bcl-G, a pro- chromosome segregation and cytokinesis. Cell Struct Funct 26: apoptotic member of Bcl-2 family. Breast Cancer Res 9: R17. 653–657. Nagayama S, Fukukawa C, Katagiri T, Okamoto T, Aoyama T, Togashi A, Katagiri T, Ashida S, Fujioka T, Maruyama O, Oyaizu N et al. (2005). Therapeutic potential of antibodies against Wakumoto Y et al. (2005). Hypoxia-inducible protein 2 (HIG2), a FZD 10, a cell-surface protein, for synovial sarcomas. Oncogene 24: novel diagnostic marker for renal cell carcinoma and potential target 6201–6212. for molecular therapy. Cancer Res 65: 4817–4826.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene