Oncogene (2006) 25, 480–486 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc SHORT COMMUNICATION Elevated expression of C10orf3 ( 10 open reading frame 3) is involved in the growth of colon tumor

M Sakai1,2, T Shimokawa1, T Kobayashi1, S Matsushima1, Y Yamada2, Y Nakamura1 and Y Furukawa3

1Laboratory of Molecular Medicine, Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan; 2Department Anesthesiology, Yokohama City University School of Medicine, Kanagawa, Japan and 3Promotion of Genome-based Medicine Project, Human Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan

After analysing -expression profiles of colon cancers Standardized regimens for advanced or recurrent on a cDNA microarray containing cDNAs corresponding disease, which involve combinations of anticancer drugs to 23 040 human , we focused on a gene annotated as such as 5-fluorouracil (5-FU), cisplatin, leucovorin, and C10orf3 (chromosome 10 open reading frame 3), whose irinotecan, are only effective in 20–30% of cases expression was elevated in colorectal cancers (CRC) as (Douillard et al., 2000; Saltz et al., 2000), and those well as in tumors arising in the stomach, lung, pancreas, drugs often evoke highly adverse reactions, being poorly and breast. The gene encodes a putative 464-amino-acid tolerated by some patients and sometimes even fatal. containing a domain known as AAA (ATPases Therefore, effective agents for chemo- or immunother- associated witha variety of cellular activities). Western apy are eagerly sought. To achieve that goal, it is blot analysis using an antibody to the gene product essential that the molecular mechanisms involved in confirmed that the protein was overexpressed in nine of colorectal carcinogenesis be better understood. the 15 clinical cancer tissues examined, compared to Recent progress in revealing mechanisms underlying corresponding noncancerous epithelial cells. A subsequent some neoplasms has opened a new page in terms of proteomics analysis revealed that C10orf3 product strategies for developing novel therapeutic drugs. For associated withtheproduct of tumor susceptibility gene example, STI-571, a tyrosine kinase inhibitor that blocks 101 (TSG101), and that C10orf3 downregulated TSG101 kinase activity of the bcr-abl fusion protein, can in a post-transcriptional manner. Expression of short effectively treat cases of chronic myelogenous leukemia interfering RNA in cells derived from CRC caused in which constitutive activation of bcr-abl tyrosine significant decreases in C10orf3 expression and inhibited kinase plays a crucial role in transformation of growth of the transfected cells, which was associated with leukocytes (O’Dwyer and Druker, 2000). STI-571 is increased apoptotic cells. These data suggest that elevated also effective against gastrointestinal stromal tumors C10orf3 expression might play an essential role in the where activating mutations in the c-kit oncogene are a growthof cancer cells, and thatsuppression of C10orf3- tumorigenic factor. Gefetinib, a specific inhibitor of the mediated signal transduction may be a novel therapeutic EGFR tyrosine kinase, has brought about a drastic strategy to a wide range of human tumors. improvement in the treatment of non-small-cell lung Oncogene (2006) 25, 480–486. doi:10.1038/sj.onc.1209051; cancers (Gridelli et al., 2003). Trastuzumab, a mono- published online 19 September 2005 clonal antibody to the HER2/neu receptor that is overexpressed in approximately 30% of breast cancers, Keywords: colon; cancer; C10orf3; TSG101 has improved clinical response and overall survival of many breast-cancer patients (Molina et al., 2001). Since these drugs are designed to suppress oncogenic activity Colorectal cancer (CRC) is a leading cause of cancer of specific gene products involved in carcinogenesis, the death worldwide; nearly 5 800 000 individuals died of the products of other genes that are specifically expressed in disease in 2002 (Walt, 2004). In spite of recent advances cancer cells and involved in the growth or survival of in diagnosis and therapeutic strategies, prognosis for cancer cells also represent promising targets for devel- patients with advanced tumors remains poor because opment of new anticancer agents. multiple affected lymph nodes and/or distant metastases Earlier, to uncover mechanisms underlying CRC and are unlikely to be completely curable by surgery. identify novel target molecules for development of anticancer drugs, we analysed expression profiles of Correspondence: Dr Y Furukawa, Promotion of Genome-based colorectal tumors by means of a cDNA microarray Medicine Project, Human Genome Center, Institute of Medical representing 23 040 genes, and had identified a number Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, of genes that were upregulated in the cancer cells (Lin Tokyo 108-8639, Japan. E-mail: [email protected] et al., 2002). In this study, we focused on a gene Received 6 December 2004; revised 13 July 2005; accepted 25 July 2005; corresponding to Hs.14559 (http://www.ncbi.nlm.nih. published online 19 September 2005 gov/UniGene/), annotated as C10orf3 (chromosome 10 C10orf3 in the growth of human colon tumor M Sakai et al 481 open reading frame 3). Subsequent quantitative real- that the predicted C10orf3 protein contained a PDB1i84 time (RT)–PCR confirmed elevated expression of domain between codons 63 and 391, or an ‘ATPases C10orf3 in all nine clinical CRC samples we examined associated with a variety of cellular activities’ (AAA) (Figure 1a and Table 1). On multipletissue Northern domain between codons 7 and 310. An NCBI blots using the C10orf3 cDNA as a probe, we detected a conserved-domain search (http://www.ncbi.nlm.nih. 2.7-kb transcript that was abundantly expressed in the gov/BLAST/) found that the predicted protein contained testis and to a lesser degree in the small intestine, colon, thymus, placenta, pancreas and stomach (Figure 1b). Table 1 Evaluation of C10orf3 expression by quantitative PCR, Expression of this transcript was not detectable in the IB and IHC other 16 tissues examined. Since the assembled cDNA Patient no. Quantitative PCR IB IHC sequence of C10orf3 in the NCBI database was shorter a than the 2.7-kb transcript, we predicted exon-like (T/N ratio) (T/N ratio) scoring sequences from the human genome database using 1NDb 1.2 + GENSCAN (http://genes.mit.edu/GENSCAN.html) 2NDb 1.4 + and the Gene Recognition and Assembly Internet Link 3 5.4 16.6 + 4NDb 0.95 + (GRAIL, http://www.compbio.ornl.gov/Grail-1.3/) pro- 5NDb 1.8+ grams, and performed exon-connection experiments. 6NDb 2.2 + We finally obtained a cDNA sequence of 2624 nucleo- 719NCc + tides (GenBank Accession Number AB091343) contain- 8ND b NCc 3+ ing an open reading frame of 1392 nucleotides encoding 9 5.83.2 2+ 10 6.3 54.6 3+ a putative 464-amino-acid protein with a predicted 11 6.3 NC (>2)d 2+ molecular weight of 54 kDa. Comparison of the cDNA 12 4.82.6 2+ with genomic sequence allowed us to determine that the 13 2.9 3.5 2+ C10orf3 gene consisted of nine exons covering a 14 3.1 2.0 + genomic region of approximately 33 kb on chromosomal 15 2.7 5.1 2+ band 10q23.3. aIHC Scoring: À, no staining; +, slight staining; 3+, strong staining; The simple modular architecture research tool 2+, staining between + and 3+. bND, not determined. cNC, not (SMART, http://smart.embl-heidelberg.de) revealed calculated. dNC (>2), estimated >2.

a 25

20 19

15

10 T/N ratio 5.8 6.3 6.3 5.4 4.8 5 2.9 3.1 2.7

0 Patient no. 3 7 9 101112131415

b

bone marrow bone brain kidney testis thyroid gland adrenal

spleen prostate

stomach leukocyte colon placenta muscle skeletal pancreas thymus ovary cord spinal

heart liver trachea lung small intestine lymph node

• • • • • • •

• • •

• • • • • • • • • • • • • • • • • • • •

C10orf3

β-actin

Figure 1 (a) Expression of C10orf3 in additional nine CRCs evaluated by quantitative RT–PCR using TaqMan assay. Relative expression ratio of C10orf3 in tumor to corresponding noncancerous tissue is illustrated in the histogram. RNA extraction, cDNA synthesis, and TaqMan assay were carried out as described elsewhere (Obama et al., 2005). Primer sequences were (for C10orf3) forward, 50- TCCGAAAAGCAAGAAATCAAATAA-30, reverse, 50- GATGGCAAACTCATGAAGCTGTT-30, and probe, 50-FAM-ACAGTTGGAATCCTTG-MGB-30, and (for b-actin) forward, 50-GGCACCCAGCACAATGAAG-30, reverse, 50-ACACG GAGTACTTGCGCTCA-30, and probe, 50-FAM-TCAAGATCATTGCTCCTC-MGB-30.(b) Analysis of C10orf3 expression on Northern blots of adult human tissues. Human multiple-tissue blots (BD Biosciences) were hybridized with a 32P-labeled PCR product of C10orf3. Expression of b-actin served as a control.

Oncogene C10orf3 in the growth of human colon tumor M Sakai et al 482 a partial conserved sequence of an SMC domain between adenomas of the colon. Positive C10orf3 staining was codons 53 and 405, or a partial myosin class II heavy observed in all 15 cancers, mainly in the cytoplasm, and chain domain between codons 20 and 435. A the intensity of staining was stronger than that in search with the predicted amino-acid sequence identified corresponding noncancerous epithelial cells (Table 1 several homologous in mouse (GenBank and Figure 2d, upper panel). Although noncancerous Accession Number AK028216.1), rat (GenBank acces- epithelial cells in the mucosa were weakly stained with sion number NM_001025646.1), chicken (GenBank the antibody, no significant difference of staining was Accession Number XP_421675.1), and Drosophila observed among the noncancerous cells. Enhanced (GenBank Accession Number NM_165213.2), which, C10orf3 expression was also noted in 11 of the 13 respectively, shared 75, 76, 52, and 23% identity with adenomas (premalignant lesions of the colon) including human C10orf3. We prepared polyclonal antibody to C10orf3 and examined by Western blotting its presence in colon- a cancer cell lines SW480, SW948, HCT116, HT29, λPPase HT29 HCT116LoVo DLD1 HCT116, LoVo, and DLD1. The results showed a SW480 SW948 54-kDa band corresponding to the predicted size of the + C10orf3 protein as well as an additional 59 kDa band anti-C10orf3 (Figure 2a). To test whether the 59-kDa band repre- - sented a form of C10orf3 modified by phosphorylation, we treated the cellular extracts with l-phosphatase anti-β-actin before immunoblotting. Since the 59-kDa band did not appear after phosphatase treatment, we judged that C10orf3 was phosphorylated only in living cells. To b anti-C10orf3 anti-C10orf3+DAPI investigate the subcellular location of C10orf3, we carried out fluorescent immunohistochemical staining in SW480 and HCT116 cells. The protein appeared mainly in cytoplasm (Figure 2b). To assess expression of SW480 C10orf3 protein in cancerous tissues, we carried out Western blotting using whole-cell extracts from 15 CRC specimens and the corresponding noncancerous muco- sae (Figure 2c). The intensities of C10orf 3 and b-actin bands in the Western blots were quantified using the chemiluminescent gel imaging system, and ratios of C10orf3 in tumor to noncancerous mucosa were HCT116 calculated after the normalization with b-actin expres- sion. Among the 15 cases, nine tumors (patients 3, 6, 9, 10, 11, 12, 13, 14, and 15) showed elevated C10orf3 protein more than twofold (Table 1 and Figure 2c). We further performed immunohistochemical staining of c Patient no. 1234 5678 C10orf3 in the 15 cancer tissues and an additional 13 NTNTNTNT NTNTNT NT anti-C10orf3 anti−β−β-actin Patient no. 91011 12 13 14 15 Figure 2 (a) Expression of C10orf3 in colon cancer cell lines. NTNTN T NTNTNTNT Proteins from CRC cell lines treated with (middle panel) or without anti-C10orf3 (upper panel) l-phosphatase (l-PPase) were immunoblotted with anti−β−β-actin anti-C10orf3 antibody and HRP-conjugated anti-rabbit IgG (Amersham) as the secondary antibody. b-Actin served as a control. Cell extracts (20 mg each) were incubated with or without d × 40 × 200 800 units of l-phosphatase (New England Bio Labs) in the presence of 2 mM Mn2 þ at 301C for 30 min, and proteins were separated by 10% SDS–PAGE. A polyclonal antibody to C10orf3 was purified from the sera of rabbits inoculated with recombinant His-tagged C10orf3 protein that was purified from Escherichia coli bacterial carcinoma cultures in a manner described elsewhere (Shimokawa et al., 2003). (b) Subcellular localization of C10orf3 in SW480 and HCT116 cells. Immunocytochemical staining of SW480 and HCT116 (colon cancer) cells with anti-C10orf3 antibody was performed in a manner reported previously (Okabe et al., 2003). (c) Western blotting analysis of C10orf3 in 15 colorectal tissues. T: tumor tissue, N: noncancerous mucosa. Proteins were separated by 10% adenoma SDS–PAGE. b-Actin served as a control. (d) Representative images of immunohistochemical staining of C10orf3 in colon cancer (upper panel) and adenoma (lower panel) tissues. Magnification:  40 (left panel) and  200 (right panel).

Oncogene C10orf3 in the growth of human colon tumor M Sakai et al 483 those with mild atypia, suggesting that elevated C10orf3 a C10orf3-Flag −+−+ expression is involved in the very early step of colorectal TSG101-HA −−++ tumorigenesis. Some of the C10orf3-positive adenoma IB: anti-TSG101 tissues were composed of adenomatous cells and IP: anti-Flag nonadenomatous cells, in which only adenomatous cells IB: anti-Flag were positively stained with the antibody (Figure 2d, IB: anti-Flag lower panel). Together with the Western blot data, IP: anti-HA immunohistochemical staining corroborated elevated IB: anti-TSG101 C10orf3 expression in colorectal tumors. To explore the function of C10orf3 further, we used a b C10orf3-Myc/His −−+ fluorescent 2D-gel electrophoretic assay to search for TSG101-Flag −++ proteins that would interact with C10orf3. We trans- fected HEK293 cells with a plasmid expressing Flag- anti-Flag tagged C10orf3 protein, extracted proteins from the IB cells, and carried out immunoprecipitation using anti- anti-β-actin Flag antibody. The immunoprecipitants were labeled with Cy5, while immunoprecipitants with anti-Flag TSG101 antibody from cells transfected with control plasmids RT-PCR were labeled with Cy3. Equal amounts of Cy5- and Cy3- GAPDH labeled proteins were subjected to 2D-gel electrophor- esis. We identified four Cy5-specific spots and analysed Figure 3 (a) Interaction between C10orf3 and TSG101 in vivo. the proteins using tandem mass spectrometry (data not Extracts from cells transfected with the indicated plasmids were shown). Among the four proteins, three peptide immunoprecipitated with anti-Flag antibody (first and second panels) or anti-HA antibody (third and fourth panels). Western sequences were successfully determined; these sequences blotting was performed with anti-TSG101 antibody (first and corresponded to Programmed Cell Death 6 Interacting fourth panels) or with anti-Flag antibody (second and third Protein (PDCD6IP), FLJ13969 protein, and Tumor panels). COS7 cells transfected with pCMV-C10orf3-Flag, Susceptibility Gene 101 protein (TSG101). Using anti- pCAGGS-TSG101-HA, or a combination, were harvested 48h after transfection. The cell lysates in TNE buffer were immuno- Flag antibody, we immunoprecipitated extracts from precipitated using either mouse anti-Flag or rat anti-HA antibody COS7 cells transfected with pCMV-C10orf3-Flag, with (ICN/Cappel). Precipitated proteins were separated by SDS– or without pCAGGS-TSG101-HA. Western blot ana- PAGE and immunoblotted using either mouse anti-TSG101 lysis confirmed that HA-tagged TSG101 co-immuno- antibody (Santa Cruz) or mouse anti-Flag antibody (SIGMA). precipitated from cells transfected with pCAGGS- (b) Effect of C10orf3 on the expression of TSG101 protein in vivo. Western blot analysis and semiquantitative RT–PCR experiments TSG101-HA, but not from cells transfected with were carried out using extract or total RNA from cells transfected pCAGGS-mock-HA (Figure 3a, first panel). Conver- with indicated plasmids. Primer sequence were (for TSG101) sely, immunoprecipitation with anti-HA antibody co- forward, 50-CTTCTCAGCCTCCTGTGACC-30, and reverse, immunoprecipitated Flag-tagged C10orf3 from cells 50- TACTTTAAGAAGAGCTCAAC-30, and (for GAPDH)forward, 50-ACAACAGCCTCAAGATCATCAG-30, and reverse, 50-GGTCC with pCMV-C10orf3-Flag, but not those with pCMV- ACCACTGACACGTTG-30. mock-Flag (Figure 3a, third panel). These results confirmed that interaction took place between C10orf3 and TSG101 in vivo. To confirm their interaction, we cotransfected activity to tumor cells through the downregulation of HEK293 cells with pCAGGS-TSG101-3 Â Flag in TSG101. This hypothesis needs further investigation. combination with or without pcDNA3.1-C10orf3-Myc/ We designed plasmids to express siRNAs for C10orf3 His, and carried out immunohistochemical staining (psiH1BX-C10orf3-A and psiH1BX-C10orf3-B), and using anti-FLAG antibody. Although we were able to examined them for potential knockdown effects in detect Flag-tagged TSG101 in the cytoplasms in the cancer cells. Transfection of SW480 and HCT116 cells absence of exogenous C10orf3, we detected no signal with psiH1BX-C10orf3-B significantly reduced C10orf3 when cells were transfeced with both plasmids (data not expression, but neither psiH1BX-C10orf3-A nor the shown). Immunoblot analysis using extracts from the control plasmid (psiH1BX-EGFP) affected the expres- cells revealed that Flag-tagged TSG101 was significantly sion level (Figure 4a and b). We then examined viability reduced in the presence of C10orf3 compared to the of SW480 and HCT116 cells that had been transfected absence of C10orf3 (Figure 3b), suggesting that C10orf3 with these plasmids and cultured in media containing may decrease TSG101 protein. Interestingly, semiquan- appropriate concentrations of geneticin. Cultures trans- titative RT–PCR using RNA from the cells demon- fected with C10orf3-B siRNA showed significantly strated unchanged expression of TSG101 mRNA by reduced numbers of viable cells compared to those coexpression with C10orf3 (Figure 3b). Taking the transfected with EGFP siRNA, and the number of interaction between C10orf3 and TSG101 together, viable cells in C10orf3-A siRNA-transfected cultures C10orf3 may downregulate TSG101 protein through was similar to that in EGFP siRNA-transfected cultures destabilization of TSG101. Since TSG101 is involved in (Figure 4a and b). The growth-inhibitory effect of the the suppression of transformation (Li and Cohen, 1996), siRNAs correlated well with their gene silencing (or not) elevated expression of C10orf3 may confer oncogenic effects, indicating an essential role of C10orf3 in growth

Oncogene C10orf3 in the growth of human colon tumor M Sakai et al 484

abA B A B

si-EGFP si-C10orf3-si-C10orf3- si-EGFP si-C10orf3-si-C10orf3- C10orf3 C10orf3

GAPDH GAPDH

∗ ∗ ∗ * p<0.01 ∗ 0.23 1.4 * p<0.01 0.20 1.2 0.17 1.0 0.15 0.12 0.8 0.10 0.6 0.08 Absorbance

Absorbance 0.4 0.05 0.2 0.03 0 0

* * * p<0.01 d 22.5 c si-EGFP si-C10orf3-A si-C10orf3-B 20 SubG1 SubG1 SubG1 17.5 Total Total Total 15 12.5 10 7.5

counts 5

Ratio (SubG1/total) 2.5 0 FL2A FL2A FL2A si-EGFPsi-C10orf3-Asi-C10orf3-B

Figure 4 Effect of C10orf3-siRNA on the expression of C10orf3 and viability of colon-cancer cells SW480 (a) and HCT116 (b), as measured by MTT assay. A plasmid vector (psiH1BX) expressing short interfering RNA (siRNA) and a control plasmid expressing siRNA to EGFP (psiH1BX-EGFP) were prepared as described previously (Shimokawa et al., 2003). Plasmids expressing C10orf3- siRNA were prepared by cloning the following double-stranded oligonucleotides into the psiH1BX vector: 50-TCCCCATCTGGAA GATGATAGGCTTCAAGAGAGCCTATCATCTTCCAGATG-30 and 50-AAAACATCTGGAAGATGATAGGCTCTCTTGAA GCCTATCATCTTCCAGATG-30 (psiH1BX-C10orf3-A) and 50-TCCCGGAGAGACTGAAAACAGAGTTCAAGAGACTCTG TTTTCAGTCTCTCC-30 and 50-AAAAGGAGAGACTGAAAACAGAGTCTCTTGAACTCTGTTTTCAGTCTCTCC-30 (psiH1BX- C10orf3-B). These plasmids were transfected into SW480 and HCT116 cells using Nucleofector reagent according to the supplier’s recommendations (Amaxa). Semiquantitative RT–PCR experiments using total RNA that was extracted from the cells 48h after transfection. In another experiment, cells were transfected with psiH1BX-C10orf3-A, psiH1BX-C10orf3-B, or psiH1BX-EGFP using FuGENE reagent (Roche) and cultured in the presence of appropriate concentrations of geneticin (G418) for 10 days. Viability of the transfected cells was examined by a cell-counting kit (Dojindo Laboratories). The data were subjected to analysis of variance (ANOVA) and Scheffe’s F-tests. (c) Flow cytometry of SW480 cells treated with siRNAs for five days. (d) Induction of subG1 population in response to the indicated siRNAs. Relative number of subG1 cells to the treated cells was measured by flow cytometry at day 5.

of colon-cancer cells. To clarify whether the growth pancreatic cancers (Nakamura et al., 2004), and 32 suppression is associated with induction of apoptosis or of 44 breast cancers examined (Nishidate et al., 2004). not, we treated SW480 cells with siRNAs, and carried Therefore, overexpression of the C10orf3 gene is not out FACS analysis. Expression of si-C10orf3-B signifi- specific to CRC but is associated with a wide range cantly increased subG1 cells (18.3%, Po0.01) compared of human cancers. Furthermore, in the experiments to those treated with si-EGFP (9.8%) or si-C10orf3-A noted here, C10orf3 protein was frequently augmented (9.3%). These data suggested that suppression of in pre-malignant lesions of the colon, an indication C10orf3 might lead to induction of apoptosis (Figure that enhanced expression of the gene can occur in an 4c and d). early step of tumorigenesis; that is, C10orf3 may play a We have documented here the up-regulation of role not in the progression of cancers but in the C10orf3 in the majority of CRCs we examined. Earlier development of tumors. Hence, enhanced C10orf3 may expression-profile analyses using microarrays had associate with features common to benign and malig- shown that C10orf3 expression was also elevated in nant tumors, such as hyperplastic growth or prolonged eight of 13 gastric cancers (Hasegawa et al., 2002), three survival. of four hepatocellular carcinomas (Okabe et al., 2001), The SMART program predicted an AAA domain 22 of 25 lung cancers (Kikuchi et al., 2003), 11 of 12 in the C10orf3 protein. Generally, AAA domains

Oncogene C10orf3 in the growth of human colon tumor M Sakai et al 485 contain one or two copies of a well-conserved cassette of TSG101 plays other roles as well, for example in 230–250 amino acids encompassing Walker A ubiquitination (Garrus et al., 2001), transcriptional and B motifs (Neuwald et al., 1999); however, C10orf3 regulation (Li et al., 2001), and endosomal trafficking contained neither of these motifs, so it is not likely (Bishop and Woodman, 2001). Since we have shown to be an actual member of the AAA family. On the here that C10orf3 interacts with TSG101, elevated other hand PDB1i84, an alternative domain predicted expression of C10orf3 may play a crucial role in the by SMART, is a motif conserved in contractile proliferation of cancer cells by modulating TSG101. proteins including smooth muscle myosin heavy chain, Given the localization of TSG101 in mitotic spindles and our BLAST search consistently identified a partial and the nature of conserved sequences in C10orf3, the myosin II heavy chain domain in C10orf3. The latter may exert an essential role in mitosis in association conserved-domain search also identified part of a with TSG101. structural maintenance of (SMC) domain. Northern blotting revealed tissue-specific expression SMCs are generally comprised of a N-terminal ATP of C10orf3 in normal organs, that is, abundant binding cassette (ABC) containing a Walker A motif; a expression in the testis; low levels of expression in the C-terminal ABC containing an Walker B motif, and a small intestine, thymus, and placenta; very low levels in central conserved-arginine finger region; the three the colon, stomach and pancreas; and undetectable regions are connected by two long coiled-coil motifs levels in the heart, liver, lung, kidney and bone marrow. (Jessberger, 2002). SMC proteins play crucial roles in In terms of development of anticancer drugs targeting the chromatin structure, through ATPase activity of the C10orf3, since the gene is not expressed in organs that two ABCs (Jessberger, 2002) . However, the homo- associate directly with life, inhibition of C10orf3 should logous region of C10orf3 localized in the N-terminal not have seriously adverse consequences except in the coiled-coil motif of the SMC domain, and did not digestive tract. include ABCs. These data suggest that C10orf3 lacks We have demonstrated that expression of C10orf3 is ATPase activity but could still function as a contractile elevated in a wide range of human cancers, and is absent protein. or weak in normal tissues except for the testis. TSG101 was originally identified as a tumor suppres- Importantly, in our experiments suppression of sor whose decreased expression transformed NIH3T3 C10orf3 expression significantly retarded the growth of cells and increased the metastasizing potential of SL6 cancer cells; therefore, C10orf3 may be a good candidate cells derived from NIH3T3 in athymic mice (Li and for molecular targeting to treat human cancers. Further Cohen, 1996). Although aberrant transcripts of TSG101 investigation will be necessary for disclosing the had been identified in a variety of human cancers mechanism(s) by which C10orf3-siRNA retards growth including breast and prostate, these transcripts were of cancer cells, and for revealing the physiological role later found to be alternatively spliced forms (Lee and of the protein in normal testis. The data provided here Feinberg, 1997; Sun et al., 1997). The tumor-suppressive should contribute to a better understanding of human role of TSG101 has been controversial, moreover, since carcinogenesis and to the development of novel thera- Wagner et al. (2003) reported that TSG101 is essential peutic strategies. for growth, proliferation, and survival of cells in normal embryonic and adult tissues. In addition, downregula- tion of TSG101 by short interfering RNA resulted in Acknowledgements partial arrest of the cell cycle and decreased proliferation of prostate- and breast-cancer cells (Zhu et al., 2004). We appreciate the technical assistance of Yoshika Sakamoto, TSG101 protein colocalizes with microtubular organiz- Noriko Ikawa, and Tae Makino and the contributions of Dr Toyomasa Katagiri for the preparation of anti-C10orf3 ing centers and mitotic spindles during mitosis (Xie antibody, and are grateful to Ryuji Hamamoto and Kazutaka et al., 1998). Injection of anti-TSG101 antibody results Obama for helpful discussions. This work was supported in in nuclear, microtubule, and mitotic-spindle abnormal- part by the Research for the Future Program Grant ities (Xie et al., 1998; Zhong et al., 1998), suggesting that #00L01402 from the Japan Society for the Promotion of TSG101 plays an essential role in mitosis. However, Science.

References

Bishop N, Woodman P. (2001). J. Biol Chem 276: Kikuchi T, Daigo Y, Katagiri T, Tsunoda T, 11735–11742. Okada K, Kakiuchi S et al. (2003). Oncogene 22: Douillard JY, Cunningham D, Roth AD, Navarro M, James 2192–2205. RD, Karasek P et al. (2000). Lancet 355: 1041–1047. Lee MP, Feinberg AP. (1997). Cancer Res 57: 3131–3134. Garrus JE, von Schwedler UK, Pornillos OW, Morham SG, Li L, Cohen SN. (1996). Cell 85: 319–329. Zavitz KH, Wang HE et al. (2001). Cell 107: 55–65. Li L, Liao J, Ruland J, Mak TW, Cohen SN. (2001). Proc Natl Gridelli C, Rossi A, Maione P. (2003). Oncogene 22: Acad Sci USA 98: 1619–1624. 6629–6638. Lin YM, Furukawa Y, Tsunoda T, Yue CT, Yang KC, Hasegawa S, Furukawa Y, Li M, Satoh S, Kato T, Watanabe Nakamura Y. (2002). Oncogene 21: 4120–4128. T et al. (2002). Cancer Res 62: 7012–7017. Molina MA, Codony-Servat J, Albanell J, Rojo F, Arribas J, Jessberger R. (2002). Nat Rev Mol Cell Biol 3: 767–778. Baselga J. (2001). Cancer Res 61: 4744–4749.

Oncogene C10orf3 in the growth of human colon tumor M Sakai et al 486 Nakamura T, Furukawa Y, Nakagawa H, Tsunoda T, Saltz LB, Cox JV, Blanke C, Rosen LS, Fehrenbacher L, Ohigashi H, Murata K et al. (2004). Oncogene 23: Moore MJ et al. (2000). N Engl J Med 343: 905–914. 2385–2400. Shimokawa T, Furukawa Y, Sakai M, Li M, Miwa N, Lin YM Neuwald AF, Aravind L, Spouge JL, Koonin EV. (1999). et al. (2003). Cancer Res 63: 6116–6120. Genome Res 9: 27–43. Sun Z, Pan J, Bubley G, Balk SP. (1997). Oncogene 15: Nishidate T, Katagiri T, Lin ML, Mano Y, Miki Y, Kasumi F 3121–3125. et al. (2004). Int J Oncol 25: 797–819. Wagner KU, Krempler A, Qi Y, Park K, Henry MD, Triplett Obama K, Ura K, Katagiri T, Tsunoda T, Nomura A, Satoh S AA et al. (2003). Mol Cell Biol 23: 150–162. et al. (2005). Hepatology 41: 1339–1348. Walt G. (2004). BMJ 328:6. O’Dwyer ME, Druker BJ. (2000). Curr Opin Oncol 12: Xie W, Li L, Cohen SN. (1998). Proc Natl Acad Sci USA 95: 594–597. 1595–1600. Okabe H, Satoh S, Furukawa Y, Kato T, Hasegawa S, Zhong Q, Chen Y, Jones D, Lee WH. (1998). Cancer Res 58: Nakajima Y et al. (2003). Cancer Res 63: 3043–3048. 2699–2702. Okabe H, Satoh S, Kato T, Kitahara O, Yanagawa R, Zhu G, Gilchrist R, Borley N, Chng HW, Morgan M, Yamaoka Y et al. (2001). Cancer Res 61: 2129–2137. Marshall JF et al. (2004). Int J Cancer 109: 541–547.

Oncogene