Oncogene (2011) 30, 3570–3584 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE Identification of salt-inducible 3 as a novel tumor antigen associated with tumorigenesis of ovarian cancer

S Charoenfuprasert1,2,9, Y-Y Yang3,10, Y-C Lee3,10,11, K-C Chao4,10, P-Y Chu5, C-R Lai6, K-F Hsu7, K-C Chang8, Y-C Chen9, L-T Chen9, J-Y Chang9, S-J Leu1,2,12 and N-Y Shih9,12

1Graduate Institute of Medical Science, College of Medicine, Taipei Medical University, Taipei, Taiwan; 2Department of Microbiology and Immunology, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan; 3School of Medical Laboratory Science and Biotechnology, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan; 4Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, Taipei, Taiwan; 5Department of Surgical Pathology, Changhua Christian Hospital, Changhua, Taiwan; 6Department of Pathology, Taipei Veterans General Hospital, Taipei, Taiwan; 7Department of Obstetrics and Gynecology, National Cheng Kung University, Tainan, Taiwan; 8Department of Pathology, National Cheng Kung University, Tainan, Taiwan and 9National Institute of Cancer Research, National Health Research Institutes, Tainan, Taiwan

Existence of humoral immunity has been previously novel ovarian TAA. Overexpression of SIK3 promotes demonstrated in malignant ascitic fluids. However, only G1/S cell cycle progression, bestows survival advantages a limited number of immunogenic tumor-associated to cancer cells for growth and correlates the clinicopatho- antigens (TAAs) were identified, and few of which are logical conditions of patients with ovarian cancer. associated with ovarian cancer. Here, we identified salt- Oncogene (2011) 30, 3570–3584; doi:10.1038/onc.2011.77; inducible kinase 3 (SIK3) as a TAA through screening of a published online 14 March 2011 random peptide library in the phage display system. Overexpression of SIK3 markedly promoted cell prolif- Keywords: tumor-associated antigen; ovarian cancer; eration, attenuated p21Waf/Cip1 and p27Kip expressions in CA125; malignant ascitic fluid; phage display low-grade OVCAR3 cells, and permitted the cells to grow in mice. Decrease in SIK3 expression in high-grade SK- OV3 cells consistently demonstrated its tumorigenic potency by modulating the levels of cell cycle Introduction regulators. When the expressions of SIK3 and CA125 were compared in cancer tissues, immunohistochemical Ovarian carcinoma represents one of the most insidious (IHC) studies indicated that cytoplasm-localized SIK3 and aggressive cancers and is also the most lethal was highly expressed in 55% of the ovarian cancer gynecological malignancy in the United States and samples. In contrast, it was rarely detected in adenomyo- Taiwan. The high mortality rate of ovarian cancer is sis, leiomyoma and normal ovary tissues, showing its a consequence of the fact that 70–75% of women with higher specificity (97%) to CA125 (65%) in ovarian this disease are diagnosed with stage III or IV disease cancer. Moreover, experiments using pharmacological (Runnebaum and Stickeler, 2001). Thus, discovery of inhibitors to block SIK3-induced p21Waf/Cip1 expression new relevant biomarkers to increase specificity or revealed that activation of c-Src and phosphoinositide-3- sensitivity for early diagnosis and prognosis of ovarian kinase were critically required for its biological activity, cancer is an important and urgent need. Besides their suggesting that they are the downstream signaling potential clinical use for diagnosis and prognosis of this mediators of SIK3. These data were further supported disease, study of their pathological functions in tumori- by IHC studies, showing coexpression of c-Src with SIK3 genesis or in tumor malignancy is not only provide in 85% of the ovarian tumor samples stained positive for molecule-basis understanding in fundamental cancer SIK3. Collectively, our findings indicate that SIK3 is a biology, but also may identify potential signaling mediators of the biomarkers in cancer cells for targeting therapy in future. Correspondence: Dr S-J Leu, Department of Microbiology and Immuno- One promising approach in this field is to study logy, School of Medicine, College of Medicine, Taipei Medical University, the immune responses to cancer and identify more Taipei 110, Taiwan. tumor-associated antigens (TAAs) for clinical applica- E-mail: [email protected] or Professor N-Y Shih, National Institute of Cancer Research, NHRI, Tainan 706, Taiwan. tions. The presence of humoral response to TAAs has E-mail: [email protected] been well demonstrated in patients with various cancers. 10These authors contributed equally to this work. The responses are frequently mounted against tumor 11Current address: The Genomic Research Center, Academia Sinica, cells expressing altered or abnormally high levels of Taipei 115, Taiwan. 12S-J Leu and N-Y Shih co-correspond to this work. normal self-antigens with inappropriate timing (Brewer Received 7 September 2010; revised 29 January 2011; accepted 9 et al., 2009), cellular compartmentalization (Chang February 2011; published online 14 March 2011 et al., 2006) or tissue expression (Old, 2008) that result SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3571 in increased immunogenicity. Therefore, high levels of loss of the cell polarity is closely associated with tumor-associated autoantibodies can be also observed in development of the epithelial–mesenchymal transition of some cancer patients throughout their disease progres- epithelial-derived cancer cells during tumor metastasis sions (Gagnon et al., 2008), and some of them are (Guarino, 2007), it may represent a key mechanism associated with their disease outcomes (Reuschenbach by which LKB1 and SIK3 are related to cancers. Here, we et al., 2009). Currently, over 100 immunogenic TAAs report that SIK3 is an ovarian TAA. Increasing expres- have been reported in patients with a variety of cancers; sion of SIK3 promoted G1/S cell cycle progression via however, few of them are associated with ovarian activation of c-Src-phosphoinositide-3-kinase (PI3K) cancer, including CA125 (Reuschenbach et al., 2009; linkage to downregulate p21Waf/Cip1 and led to tumorigen- Taylor et al., 2009). esis in mice. As SIK3 is highly and preferentially expressed CA125 is the best-characterized tumor marker for in ovarian tumors but not in adenomyosis and leiomyo- advanced epithelial ovarian cancer. CA125 is a type I ma, it can be a potential diagnostic marker for ovarian transmembrane protein expressed on the cell surface cancers. (O’Brien et al., 2002). Its soluble proteolytic fragment can be released into the blood stream. Thus, serum CA125 has been studied intensively as a potential biomarker for diagnosis of epithelial ovarian cancer or Results to monitor disease recurrence after therapy (Baron et al., 2005; Rosenthal et al., 2006). However, elevated Identification of KIAA0999 product, SIK3, as the CA125 expressions are not only observed in >80% of target of the CA502 ascitic antibody patients with advanced epithelial ovarian cancer, but are Although tumor-associated antibodies have been found also found in various physiological or pathological in the ascitic fluids of ovarian cancer patients (Giancotti conditions (Cannistra, 2004). In this study, we used et al., 1990), only a few ovarian TAAs were identified. serological screening of a phage library with malignant In this study, the ascitic tumor cells of 12 patients were ascitic antibodies and identified salt-inducible kinase 3 first purified with Ficoll–Percoll gradient centrifugation, (SIK3) as a novel TAA for ovarian cancers. lyzed and immunoblotted with their ascitic auto- SIKs have their kinase domain sequences closely antibodies as described previously (Chang et al., 2006). homologous to AMP-activated protein kinase (AMPK) The ascitic antibodies of patient CA502 were highly (Lizcano et al., 2004; Bright et al., 2009), and may have immunoreactive to a major protein band with an a role in steroidogenesis, adipogenesis or regulation of approximate molecular mass of 150 kDa in the auto- tumor malignancy. Three isoforms in the SIK family logous tumor lysates (Supplementary Figure 1A). have been identified so far, namely, SIK1 (SNF1LK), To identify the target of the autoantibody, a 7-mer SIK2 (QIK or SNF1LK2) and SIK3 (KIAA0999 or random peptide phage library was screened as illustrated QSK) (Katoh et al., 2004; Lizcano et al., 2004). SIK1 is in Figure 1a. The library was pre-absorbed with the total an important regulator for steroidogenic gene expres- immunoglobulin G (IgG) mix of 10 healthy donors to sion, and can be activated by increased intracellular remove the phages recognized by the IgG. Subsequently, sodium leading to stimulation of Na þ /K þ -ATPase the resultant library was applied onto -linked activity and activation of calmodulin kinase (Sjostrom immunosorbent assay wells pre-coated with the et al., 2007). In addition, it is also reported to participate CA502 ascitic IgG. After four biopanning cycles, the in a negative feedback program initiated by transform- bound phages were eluted and allowed to infect their ing growth factor-b (Kowanetz et al., 2008). Loss of Escherichia coli host cells. Direct sequence analyses of SIK1 expression facilitated lung metastatic spread in 15 random clones revealed 3 residues highly conserved mice and was closely correlated with the development of in the displayed peptides, x-Pro-His-x-Tyr-x-x (x: any distal metastasis in human breast cancers (Cheng et al., residue) (Supplementary Table 1). By searching the 2009). SIK2, highly expressed in the adipose tissue NCBI protein database, we found that human (Katoh et al., 2004), is found to inhibit cAMP-response KIAA0999 (accession number: AB23216) and CRBN element binding protein-mediated gene expression by (access number: NM_016302) could encode for phosphorylating transducer of regulated CREB activity containing the conserved residues highlighted 2 (TORC2), thus sequestering TORC2 in the cytoplasm, with grey background (Pro-Pro-His-Gly-Tyr-Ala-His and preventing cAMP-response element binding protein- and Asn-Pro-His-Gly-Tyr-Val-His peptide sequences in mediated gene transcription (Screaton et al., 2004). Insulin KIAA0999 and CRBN proteins, respectively) and might receptor substrate-1 is a downstream substrates of SIK2 be the putative targets of the CA502 autoantibodies. (Horike et al., 2003), suggesting that SIK2 has a role in However, no immunoreactive protein band with an the insulin regulation of adipose tissues. As compared with approximate molecular mass of 48.6 kDa, suspected to SIK1 and SIK2, the biological functions of SIK3 are be CRBN, was found in the CA502 tumor lysates entirely obscure, particularly in cancers. The knockdown (Supplementary Figure 1A). For not to rule out this of Drosophila CG15072 (homologous to mammalian SIK3) possibility easily, ovarian normal epithelial and cancer resulted in impairment of the LKB1-regulated apical- cells, including CA502 tumor cells, were carefully exa- basal polarity in retinal photoreceptor cells (Amin et al., mined for CRBN expression by probing with its specific 2009), showing that this orthologue of SIK3 is probably antibody in western blotting analysis (Supplementary one of mediators of the LKB1 effect on the cells. Since Figure 1B). Using the CRBN recombinant protein as

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3572

Figure 1 SIK3 as the target of the CA502 ascitic antibody. (a) Serological screening for peptides associated with ovarian tumors. A phage display peptide library was pre-absorbed with healthy control sera, followed by incubation with immobilized CA502 ascitic antibody. After removal of the unbound phages, the bound M13 phages were eluted, propagated and re-absorbed with the healthy control sera. Four panning cycles were performed. Fifteen bound phage clones were randomly selected and sequenced. The peptide sequences of individual clones are listed in Supplementary Table 1. (b) The putative KIAA0999 gene product is the target of the CA502 antibody. The 30-coding region of the KIAA0999 gene containing the consensus peptide sequence of the bound phages was fused with the Myc tag sequence (Myc-CT-0999), transfected, and used to encode a 42 kDa protein in HeLa cells. Western blot analyses of the lysates were performed using purified CA502 ascitic IgG and 9E10 antibody specific to the Myc tag, as indicated. (c) Confirmation of the displayed peptide as the major immunogenic epitope. The Myc-del-CT-0999 gene lacking the 7-mer conserved sequence, PPHGYAH, which shares a six-residue identity with the phage-displayed peptide (Supplementary Table 1), was transfected into HeLa cells. After cell lysis, the lysates were resolved in a 10% sodium dodecyl sulfate-containing polyacrylamide gel and immunoblotted with the CA502 antibody. The same blot was re-probed with the 9E10 antibody to show equal expression of the wild-type and deletion mutant.

a control, the experiment clearly revealed that CRBN ablated the antibody recognition (Figure 1c), confirming was undetected in those cells. Additionally, the same that this protein is the target of the CA502 antibody and conclusion was also obtained from immunohisto- that the 7-mer peptide, Pro-Pro-His-Gly-Tyr-Ala-His, chemical (IHC) study in 38 ovarian tumor samples is its major immunogenic epitope. Recently, the KIAA0999 (Supplementary Figure 1C), suggesting that CRBN is gene product was named SIK3 that belongs to the lowly expressed in ovarian cancer cells and tissues. AMPK-related kinase family (Katoh et al., 2004). To examine whether the protein encoded by KIAA0999 might be the target of the CA502 antibodies, the putative 30-coding sequence of the KIAA0999 gene Expression of SIK3 promotes cell growth and including the consensus sequence was amplified by PCR tumorigenesis in animals and expressed as a C-terminal KIAA0999 recombinant Four ovarian cancer cell lines were tested for correlation protein tagged with the Myc sequence (Myc-CT-0999) in between the expression levels of SIK3 and the corres- HeLa cells. This polypeptide was specifically recognized ponding cell growth. MTT assays were performed by the CA502 antibody (Figure 1b). Deletion of the to measure the growth, showing that SK-OV3 and 7-mer conserved sequence in the polypeptide, which shares TOV-112D cells grew faster than the other two cell lines, a six-residue identity with the phage-displayed peptide OVCAR3 and OC-109 cells (Supplementary Figure 2A). (Supplementary Table 1), by site-directed mutagenesis Coincidently, northern blotting experiments using a

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3573 32P-labeled antisense SIK3 probe revealed that a 4.6 kb of SIK3 were established. Two clones of each transfec- gene was the major transcript, which was highly tant were randomly chosen, and their SIK3 expression abundant in high-grade SK-OV3 and TOV-112D cells levels were confirmed by reverse transcriptase–PCR and as compared with that in low-grade OVCAR3 and OC- western blot analyses (Figure 2A). Results from the 109 cells (Supplementary Figure 2B). Using an antibody MTT assays indicate that SIK3 is a positive regulator of specific to the C-terminus of SIK3, western blot analysis cell growth, showing that its expression status is indicated that a 150 kDa polypeptide was the major consistent with the corresponding growth of transfec- corresponding protein encoded by this transcript and tants (Figure 2B). Furthermore, clonogenic experiments highly expressed in the fast-growing SK-OV3 and using different numbers of OVCAR3 cells stably over- TOV-112D cells (Supplementary Figure 2C). These expressing SIK3 (OVCAR3/ov-21) seeded onto six-well results were further verified by in vitro transcription/ plates to allow cell propagation up to 9 days consistently translation-coupled reactions using the full-length demonstrated that the SIK3 transfectants evidently grew SIK3 or KIAA0999 gene as the DNA template (data faster in a dose-dependent manner and formed larger not shown) and also by SIK3 gene-specific knockdown colonies than control cells (Figure 2C). experiments (Figure 2A). To further validate its role in the promotion of To examine whether SIK3 was involved in the promo- cell growth, we subcutaneously inoculated stably trans- tion of cell proliferation, OVCAR3 cells stably over- fected SK-OV3 cells with decreased SIK3 expression expressing SIK3 or SK-OV3 cells with attenuated levels (Figure 3a; black arrow) or OVCAR3 cells with

Figure 2 Involvement of SIK3 in the regulation of cell growth. (A) Establishment of cells stably overexpressing or attenuating SIK3. SK-OV3 or OVCAR3 cells were transfected with the SIK3-specific small hairpin RNA (shRNA) to attenuate the SIK3 expression or with the pcDNA-SIK3 plasmid to elevate its expression. The stable transfectants were obtained by G418 (400 mg/ml) treatment. After limiting-dilution, two clones of each transfectants were randomly chosen. The SIK3 expression level of these cells was determined by reverse transcriptase (RT)–PCR (upper) and western blot (lower). The control cells (VC) were established in the same manner by transfection with their corresponding vectors. b-Actin served as a loading control. (B) Cell growth of the stable transfectants. Cells (2 Â 104 per well) were seeded onto 24-well plates and incubated for 24, 48, 72 or 96 h as indicated. The cell growth of transfectants was measured by a standard MTT assay. Data are the mean±s.d. of three experiments and represented as fold increases in the cell growth of individual transfectants at 24 h. The growth differences between individual transfectant clones with attenuated SIK3 expression or between individual transfectant clones overexpressing SIK3 and their corresponding controls were statistically analyzed with t-test at 96 h time point. ***Po0.001 for both SIK3 transfectant clones in each experiment. (C) Clonogenic assays. The SIK3 stable transfectants (black bar; ’) and control cells (white bar; &) were seeded onto six-well culture plates at 250, 500 or 1000 cells per well in triplicate. Cells were maintained in culture for 9 days to let the viable cells propagate to sizable colonies. After wash, the cells were stained with methylene blue and photographed under a microscopy. One representative pair from experiment (a) 250 cells, (b) 500 cells or (c) 1000 cells seeded initially is shown in the top. After cell lysis, the concentrations of released dye were determined using an enzyme-linked immunosorbent assay (ELISA) plate reader by measuring the absorbance at 550 nm. Quantitative data from the experiments are presented in the histograms, and the growth differences between the SIK3 transfectants and their controls in the experiments were analyzed by t-test.

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3574

Figure 3 Involvement of SIK3 in tumorigenesis. The SIK3 knockdown transfectants (SK-OV3/sh-SIK3; 1 Â 106 cells/200 ml per mouse) in (a) or SIK3 overexpressed transfectants (OVCAR3/ov-SIK3; 1 Â 107 cells/200 ml per mouse) in (b) were subcutaneously injected into the right side of the dorsal flanks of six mice. Simultaneously, their corresponding control cells were injected into the left side of the same mice. The tumor volumes of the individual mice were measured weekly (upper left of (a) for SK-OV3 and of (b) for OVCAR3 transfectants). The differences between SIK3 transfectants and their corresponding controls in the averages of the tumor sizes of paired groups were statistically determined by t-test analyses after mice killing (upper right of both panels). ‘*’ and ‘**’ indicate Po0.05 and Po0.01, respectively. In the bottom panels, the sides inoculated with transfectants that stably attenuated SIK3 expression (sh-SIK3) are indicated with black arrows, and the sides inoculated with transfectants stably overexpressing SIK3 (ov-SIK3) are marked with white arrows.

increased SIK3 expression (Figure 3b; white arrow) in transfectants in a dose-dependent manner (Figure 4b). the right side of the dorsal flanks of six mice. To avoid Western blot analyses in Figure 4c showed that SIK3 individual variations in mice, the corresponding control greatly affected the gene expression of p21Waf/Cip1, cyclin cells were subcutaneously inoculated in the left side of D4 and cyclin E, as well as partially modulated p27Kip the dorsal flanks of the same mice. The results clearly and cyclin D1. In contrast, there was no significant show that knockdown of SIK3 expression completely change in the cyclin D3 expression, and no cyclin D2 abolished the tumorigenic potency of SK-OV3 cells in protein was detected in both cell lines (data not shown). mice. Accordingly, overexpression of SIK3 bestowed Conclusively, SIK3 delivering its tumorigenic potency to survival advantages to OVCAR3 cells so that they could ovarian cancer cells is mainly through promotion of grow in mice. Collectively, our data strongly support the G1/S progression processes. tumorigenic role of SIK3 in ovarian cancers.

SIK3-mediated downregulation of p21Waf/Cip1is through Promotion of cell cycle progression by SIK3 c-Src activation To further explore the mechanisms underlying SIK3- In our preliminary study, increasing expression mediated tumorigenesis, the cell cycle statuses of SIK3- of p21Waf/Cip1 was sufficient to induce G1/S arrest in transfected cells were examined. The stable transfectants SK-OV3 cells (data not shown). This result agrees with were starved to synchronize their cell cycles, and then previous studies (Deng et al., 1995), suggesting that it is stimulated with fresh serum for 6 h. Flow cytometric also a key effector of SIK3-induced cell cycle regulation. analyses in Figure 4a revealed that the attenuation of To dissect which downstream signaling pathway might SIK3 expression in SK-OV3 cells significantly accumu- be responsible for SIK3-mediated downregulation of lated the cells in the G1 phase from 49 to 74% on p21Waf/Cip1, various pharmacological kinase inhibitors average. The cell population in the S phase was reduced were utilized for rescue of this effect. In Figure 5a, PP2, from 33 to 8%, whereas the percentage of G2/M cells a c-Src inhibitor, fully rescued p21Waf/Cip1 expression in was not altered. Consistently, increasing expression of SIK3-transfected OVCAR3 cells. In contrast, LY294002 SIK3 in OVCAR3 cells markedly enhanced the G1/S for PI3K, staurosporine for broad-spectrum and transition, showing that its overexpression could signi- rapamycin for mechanistic target of rapamycin-medi- ficantly decrease the G1 cell population and increase ated S6 kinase accounted for 70, 55 and 30% of the the S phase cells. Moreover, facilitation of the G1/S rescue, respectively. Thus, these data suggest that c-Src progression by SIK3 was also seen in transient may be the major downstream signaling mediator

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3575 responsible for SIK3 action. To verify the hypothesis, in 85% of the ovarian cancer samples stained positive the kinase activity of c-Src in transfectants with for SIK3. Together, all data indicate that increasing decreasing (sh-SIK3) or increasing (ov-SIK3) expression c-Src activity is a key mechanism underlying SIK3 of SIK3 was closely examined. The results show that overexpression responsible for promotion of cell proli- SIK3 positively regulated the c-Src activity in a dose- feration in ovarian cancer cells and close association dependent manner (Figure 5b). Furthermore, experi- with tumorigenesis in mice and in patients with ovarian ments using various doses of PP2 to block the c-Src cancers. activity in SIK3-transfected OVCAR3 cells showed On the other hand, modulation of the c-Src protein that despite which concentrations in use, this inhibitor level by either elevation or knockdown of its gene significantly impaired the SIK3-induced cell growth 48 reversely regulated p21Waf/Cip1 expression (Figure 7a), and 72 h post-treatments (Figure 5c). The IHC study in indicating that c-Src is an upstream signaling regulator Figure 6 consistently supported the in vitro and in vivo of the p21Waf/Cip1 gene. Thus, SIK3 may exert its bio- observations, demonstrating that c-Src was coexpressed logical action by downregulation of p21Waf/Cip1 via c-Src

Figure 4 Promotion of cell cycle G1/S by SIK3. (a) Stably transfected cells. Transfectants (1 Â 105 cells per well) were treated with media supplemented with 0.1% FBS for 24 h to synchronize their cell cycle statuses, and then incubated with 10% FBS-containing fresh medium for an additional 6 h. (b) Transiently transfected cells. Cells were transfected with 1 or 2 mg of pcDNA-myc-SIK3 or SIK3-interfering small hairpin RNA (shRNA) plasmids to overexpress or knockdown SIK3 expression for 48 h, respectively. After being trypsinized and fixed with 70% cold ethanol, the cells were treated with 1 mg/ml RNase A and stained with 50 mg/ml propidium iodide (PI). Flow cytometric data (1 Â 104 events) were acquired using a BD FACScaliber cytometer and the percentage of cells in each cell cycle stage was determined using ModFit software. One representative pair is demonstrated in the top, and data are expressed as the mean±s.d. of three experiments in histograms (*Po0.05 and ***Po0.001, respectively). (c) Western blot analysis. Fifty micrograms of each stable transfectant was resolved and immunoblotted with antibody against p21Waf/Cip1,p27Kip, cyclin Ds and cyclin E, as indicated on the right. b-Actin served as a loading control.

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3576

Figure 4 Continued.

to promote the G1/S progression. To further confirm were carried out in SK-OV3 cells. In the previous the linkage between SIK3-Src activation and p21Waf/Cip1 experiments, we demonstrated that attenuation of SIK3 gene regulation, experiments co-depleting both genes expression could upregulate p21Waf/Cip1 and cause G1/S

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3577

Figure 5 c-Src is a downstream effector of SIK3 for suppression of p21Waf/Cip1 expression. (a) Rescue of SIK3-induced p21Waf/Cip1 downregulation. OVCAR3 cells were pretreated with various pharmacological inhibitors such as LY294002 (LY), rapamycin (RP), staurosporine (SP), H89 and PP2 as indicated, for 2 h. After washing, the cells were transfected with pcDNA-SIK3 or its vector control (VC), and incubated for an additional 24 h. The lysates were resolved in sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblotted with antibody against p21Waf/Cip1, SIK3, or b-Actin. After the protein levels of p21Waf/Cip1 were quantified, data obtained from two independent experiments were normalized to b-actin and represented as fold increases of the p21Waf/Cip1 level in the control transfectants (VC) treated with control vehicle (C) (bottom). (b) Activation of c-Src by SIK3. SK-OV3 and OVCAR3 cells were transiently transfected with increasing doses of the SIK3-interfering small hairpin RNA (shRNA) (sh-SIK3) and pcDNA-SIK3 (ov-SIK3) constructs to decrease and increase SIK3 expression, respectively. Simultaneously, cells were transfected with their corresponding empty vectors as controls (VC). After 24-h transfection, the transfectants were lyzed, electrophorized, and immunoblotted with antibody against phosphor-Tyr416-Src, c-Src or SIK3. b-Actin served as a loading control. (c) Blockade of c-Src activity attenuated SIK3-induced cell growth. OVCAR3 cells stably overexpressing SIK3 (3 Â 104 cells per well) were seeded onto 24-well plates in triplicate 24 h before treatments, and then the cells were treated with increasing doses of PP2 (0.1, 0.5 and 2.5 mM)or with control vehicle for designed time courses. OVCAR3 cells stably transfected with control vector (OVCAR3/VC) were used as controls. The cell growth of transfectants was determined by MTT assays. Data are the mean±s.d. of three experiments and represented as fold increases in the cell growth of transfectants at the start time point of the treatments. The statistical significance in the growth difference was determined by t-test. ‘*’, ‘**’, and ‘***’ indicate Po0.05, o0.01 and o0.001, respectively.

arrest in the cells (Figures 4b and c). In Figure 7b, in the Preferential expression of SIK3 in ovarian cancer absence of SIK3, decreasing expression of p21Waf/Cip1 To determine whether the overexpression of SIK3 was significantly relieved the G1/S arrest and elevated the associated with ovarian cancer, the serial sections of percentage of S-phase cell populations in a dose- ovarian cancers and gynecological benign tumors, such dependent manner. In previous studies, c-Src is a pivotal as adenomyosis and leiomyoma as well as histologically promoter for cell proliferation responses through either normal ovarian tissues were examined by IHC studies activation of mitogen-activated protein/extracellular (Figures 8 and 9). Approximately 55% of the ovarian signal-regulated kinase or PI3K in different cancer cells tumor samples were stained positive for SIK3 (Table 1), (Cantley, 2002). We further characterized which path- and showed that SIK3 was primarily located in the cyto- way underlying c-Src activation in ovarian cancer cells plasm of the tumor cells (Figure 8c). No SIK3 immuno- was responsible for the p21Waf/Cip1 downregulation event. reactivity was detected in normal ovarian tissues, Experiments pretreating OVCAR3 cells with PD98059, including embedded fallopian tubes (Figures 8a and b) LY294002 or PP2 and followed by c-Src transfection and leiomyoma (Figure 9c). Additionally, only 10% indicated that PI3K was the immediate downstream (2 out of 20) of adenomyosis samples were weakly signaling molecule of c-Src (Figure 7c). Collectively, the stained positive for SIK3 (Supplementary Figure 3). activation of the Src-PI3K signaling cascade is a key In contrast, CA125 antigen was typically expressed on event during the promotion of G1/S progression by the cell surface of ovarian tumor cells (Figure 8d) and in SIK3 via regulation of the p21Waf/Cip1 gene. 80% of adenomyosis samples we have examined so far

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3578

Figure 6 Coexpression of SIK3 and c-Src proteins in ovarian tumors. Transverse serial sections obtained from 38 ovarian tumor samples were immunohistochemically examined using commercially available CA125 or c-Src antibodies or home-made SIK3 polyclonal antibody (see ‘Supplementary Materials and methods’), as indicated at the top. Immunocomplexes were detected and visualized using the labeled streptavidin biotin (LSAB) system. A total of 21 and 27 samples were stained positive for SIK3 and c-Src, respectively, and 85% (18 of 21) of the SIK3-positive samples showed coexpression of c-Src and SIK3 in the same tumor regions. Two representative samples are shown here. Magnifications, Â 100. Scale bar, 100 mm.

(Figure 9b; Table 1). Using SIK3 to predict ovarian cell proliferation; on the other hand, attenuation of cancer in our cases, we obtained a sensitivity of 53% and p21Waf/Cip1 expression markedly rescued the G1/S arrest a specificity of 97%, whereas CA125 analysis gave a caused by SIK3 depletion in SK-OV3 cells, suggesting 79% sensitivity and a 65% specificity to ovarian cancer. that c-Src and p21Waf/Cip1 are signaling mediators of the Moreover, SIK3 was also preferentially overexpressed in SIK3 effect on cell growth. Similarly, more recent one the serous subtype and in the late stage of ovarian interesting observation reported by Ahmed et al. (2010) cancers (P%0.001) (Table 1). Thus, data in the IHC demonstrated that depletion of another SIK family study further support the results obtained from the member, SIK2, caused failure in centrosome separation in vitro and animal studies, indicating that SIK3 in mitosis and delayed G1/S transition resulting in the expression is associated with development of ovarian growth suppression in SK-OV3 cells via inactivation of cancers. AKT. Moreover, they also showed the molecular profile of SIK2 depletion in mice and correlated high expres- sion of SIK2 with poor survival of patients with high- Discussion grade serous ovarian cancers. On the basis of their functional similarity on cell cycle regulation, over- Using malignant ascitic antibodies to screen for tumor- expression of SIK2 and SIK3 apparently bestows a associated peptides in a 7-mer random peptide phage great survival benefit to ovarian tumor cells. Therefore, library, we identified SIK3 as an ovarian cancer TAA. depletion of individual or both proteins is a potential Its overexpression in tumor cells results in stimulation of therapeutic approach for ovarian cancer treatment. cell proliferation by accelerating the G1/S progression SIK3 belongs to the AMPK-related kinase family processes. Consistently, the SIK3 overexpression pheno- because of their sequence homolog with the protein menon is frequently found in ovarian cancer but not kinase domain of AMPK and can be phosporylated by a associated with adenomyosis and leiomyoma. Approxi- tumor-suppressor kinase LKB1 (Bright et al., 2009). mately 55% of ovarian cancer samples stained positive Mutations in its gene result in a hereditary disorder for SIK3 belong to the serous subtype and late stages termed Peutz–Jeghers syndrome (PJS). LKB1, has been of patients, suggesting its role in tumor malignancy. well-documented to be a master upstream activator of Furthermore, we demonstrated the oncogenic potency AMPK and AMPK-related kinase, activates SIK1, of SIK3 in ovarian cancer cells using clonogenic assays SIK2 and SIK3 by phosphorylation of Thr182, Thr175 and in xenografted tumors using NOD-SCID mice, and Thr163 residues at their activation loops, respec- showing that its expression status in the cancer cells is tively (Lizcano et al., 2004). Moreover, the work of sufficient to affect cell proliferation and tumor growth. Katoh et al. (2006) showed the cytoplasmic retention of The molecular mechanism underlying its oncogenic TORC2 could only be found in the LKB1-deficient potency is mediated by manipulation of cell cycle HeLa cells when LKB1 was co-transfected with SIK3 regulatory proteins, in particular for regulation of into the cell. On the other hand, one point mutation of p21Waf/Cip1 gene expression, to facilitate G1/S transitional threonine to glutamate at the 163th residue in SIK3 progression via activation of the Src-PI3K signaling (T163E) led to constitutive activation of SIK3, which cascade. Knockdown of SIK3 expression resulted in was showed to be sufficient to cause the nuclear export G1/S arrest. Depletion of c-Src abolished SIK3-indcued of TORC2 and inactivate cAMP-response element

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3579

Figure 7 p21Waf/Cip1 is the key cell cycle regulator modulated by SIK3-activated c-Src. (a) Expression of p21Waf/Cip1 affected by c-Src. SK-OV3 and OVCAR3 cells were transfected with increasing doses of the Src-interfering small hairpin RNA (shRNA) (sh-Src) or pcDNA-Src (ov-Src) plasmids to attenuate or enhance c-Src expression, respectively. (b) Attenuation of p21Waf/Cip1 expression promoted the G1/S progression in SIK3-knockdown transfectants. SK-OV3 cells (1 Â 105 cells per well) were seeded onto six-well plates, incubated for 24 h, and transfected with SIK3-interfering shRNA plasmid or co-transfected the plasmid with different amounts of the p21Waf/Cip1 shRNA construct as indicated in the upper panel. Western blotting analyses of the transfectants were performed by probing with antibody specific to p21Waf/Cip1, SIK3 or b-actin 24 h after transfection. The cell cycle statuses of the transfectants were determined by flow cytometric analyses using PI staining. Data were acquired from the experiments using a BD FACScaliber cytometer and the percentage of cells in each cell cycle stage was determined using ModFit software. One representative set of the observations is shown in the middle panel. Quantitative data are presented in the histograms (bottom panel). The differences between p21Waf/Cip1 individual transfectants and vector control cells in the percentages of cells were statistically analyzed by t-test. ‘*’ and ‘***’ indicate Po0.05 and Po0.001, respectively. (c) p21Waf/Cip1 expression controlled by the Src-PI3K signaling pathway. OVCAR3 cells were pretreated with PD98059 (PD), LY294002 (LY), PP2, or control vehicle (C), 2 h before transfection of the pcDNA-Src (ov-Src) plasmid. After 24-h transfection, the transfectants were lyzed, electrophorized, and immunoblotted with antibody against p21Waf/Cip1 or c-Src. b-Actin served as a loading control. binding protein-dependent transcriptional activation in found to be progressively activated during the course the LKB1-deficient cells, suggesting that the phosphor- of cell transformation in human papilloma virus (HPV)- ylation of Thr163 by LKB1 is required for SIK3 immortalized cells and controlled the AMPK-regulated biological functions. In contrast, the equivalent residue protein synthesis (Mizrachy-Schwartz et al., 2007). The substitution in SIK1 and SIK2 did not display the same c-Src action is considered to optimize energy-consuming effect. Thus, the level of LKB1 protein and its gene mechanisms that re-direct cellular processes to enhance integrity become critical issues and are needed for a cell growth. In our work, overexpression of SIK3 in ovarian more intensive study in patients with ovarian cancers cancer cells is expected to phosphorylate TORC2 leading because we showed SIK3 highly expressed in the disease. to its protein degradation in the cytoplasm and attenua- Another important finding in this study is that the tion of cAMP-response element binding protein-mediated c-Src activity was indispensable for the SIK3 effect gluconeogenesis. The consequence of this event should on promotion of cell growth. Blockade of the activity lower the cellular adenosine triphosphate in the cells. ablated SIK3-induced cell proliferation in OVCAR3 Whether it represents a mechanism by which the cells cells overexpressing SIK3. Intriguingly, c-Src is recently reduce their cell sizes and protein translation to provide

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3580 Table 1 SIK3 and CA125 expression status in various gynecologic energy required for cell proliferation and survival, tumors proposed by Mizrachy–Schwartz’s group, remains to Positivity of CA125a Positivity of be closely examined in future. n (%) SIK3a n (%) The different kinase activation of SIK3 and SIK1 may contribute to their distinct biological functions Tissue type (n ¼ 98) on tumorigenesis. SIK1 was identified as an upstream Normal ovary (20) 1 (5) 0 (0) Ovarian cancer (38) 31 (81.5)b,*** 21 (55.3)b,*** regulator of p53 in a kinome-wide loss-of-function Non-ovarian cancer (40) screen, showing that SIK1-coupled LKB1 to regulate Adenomyosis (20) 16 (80) 2 (10) p53-dependent anoikis and suppress metastatic spread Leiomyoma (20) 4 (20) 0 (0) of disseminated cells as micro-metastases in the lung et al. Histological subtype of ovarian cancer (n ¼ 38) (Cheng , 2009). In addition, SIK1 is also reported Serous (13) 12 (92.3) 11 (84.6) to participate in other biological activities (Okamoto Non-serous (25) 19 (76)c 10 (40)c,*** et al., 2004; Berdeaux et al., 2007; Sjostrom et al., 2007; Endometriod (8) 8 (100) 3 (37.5) Kowanetz et al., 2008), including regulation of active Clear cell (6) 4 (66.7) 3 (50) sodium transport in response to sodium stress as well as Mucinous (3) 1 (33.3) 0 (0) Mixed (3) 2 (66.7) 0 (0) induction of steroidogenic gene expression. Intriguingly, Others (5)d 4 (80) 4 (80) stanniocalcin 1 (STC1), a glycoprotein hormone in- volved in the regulation of calcium and phosphate Stage of ovarian cancer (n ¼ 38) homeostasis, has a similar function as SIK1 in protect- I þ II (n ¼ 11) 6 (54.5) 2 (18.2) III þ IV (n ¼ 27) 25 (92.6)e,*** 19 (70.4)e,*** ing cells against toxic hypercalcium (Serlachius et al., 2002) and in regulating steroidogenesis in corpus luteal Abbreviations: IHC, immunohistochemical; SIK3, salt-inducible cells of the mammalian ovary (Paciga et al., 2003). More kinase 3. recently, STC1 protein was showed higher expression in aSamples with IHC staining score ^ þ 1 were recorded as positive. ovarian cancer tissues than in normal ovarian tissues, b Comparison of ovarian cancer with normal ovary tissues. and overexpression of STC1 increased cell proliferation cComparison between serous and non-serous subtypes of ovarian cancer. and colony formation in vitro and in vivo (Liu et al., dSamples were poorly differentiate carcinoma. 2010). The molecular mechanisms underlying STC1- eComparison between stage I/II and III/IV of patients with ovarian mediated ovarian tumorigenesis were characterized cancers. Univariate analysis was performed using Fisher’s exact test. through inhibition of cell apoptosis and modulation of ‘‘***’’ represents statistical significance (Po0.001). cell cycle regulatory proteins to promote cell growth.

Figure 8 IHC study of SIK3 in ovarian cancers. Transverse serial sections obtained from 38 ovarian cancers and 20 histologically normal ovarian tissues were immunohistochemically examined with SIK3 (see ‘Supplementary Materials and methods’) or CA125 (DAKO, Carpinteria, CA, USA) antibodies. The immunocomplexes were detected and visualized using diaminobenzidine as the substrate in the labeled streptavidin biotin (LSAB) system (DAKO). Approximately 55% of the ovarian tumor samples were stained positive for SIK3, and showed that it was primarily located in the cytoplasmic compartment of ovarian tumor cells (inset in (c)). However, no SIK3 immunoreactivity was detected in the normal ovarian tissues (a), fallopian tubes located in two ovarian tissue samples (b) and their corresponding epithelial cells (insets in (a) and (b), respectively). In contrast, the CA125 antigen was distributed in 82% of the ovarian tumor samples (d). It was dominantly displayed on the cell surface (inset in (d)). One representative pair of the serial sections of each sample stained for SIK3 or CA125 is shown here. Magnifications, Â 100 (a–d), Â 200 (insets in (a–d)). Scale bar, 100 mm.

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3581

Figure 9 IHC study of SIK3 in adenomyosis and leiomyoma. Transverse serial sections obtained from 20 adenomyosis and 20 leiomyoma were stained with home-made SIK3 (see ‘Supplementary Materials and methods’) or CA125 (DAKO) antibodies in IHC study. The immunocomplexes were detected and visualized using the labeled streptavidin biotin (LSAB) system. The SIK3 positivity was either rarely or not found in the adenomyosis and leiomyoma samples, (a) and (c), respectively. In contrast, the CA125 antigen was displayed in 80% of the adenomyosis and 20% of the leiomyoma samples, (b) and (d), respectively. One representative pair of the serial sections of each sample stained for SIK3 or CA125 is shown here. Magnifications, Â 100 (a–d). Scale bar, 100 mm.

However, the downstream signaling effectors of STC1 the lowest basal expression level found in ovarian tissue responsible for these biological outcomes are not defined (data not shown). However, here we demonstrated that yet. Interestingly, we herein demonstrated that the 55% of patients with ovarian cancers had SIK3 over- tumorigenic potency of SIK3 was mediated by a similar expression, and 70% of those patients were diagnosed as mechanism that upregulates gene expression of cyclin the late stages of this disease, suggesting that increasing Ds and cyclin E. It also simultaneously downregulates expression of SIK3 is one of the common scenarios p21Waf/Cip1 and p27Kip protein expression to accelerate during tumor development and may be associated with G1/S progression, and permitted the cells to grow in the disease malignancy. In contrast to the specificity of mice. Subsequently, we further characterized the driving CA125 to ovarian cancer (65%), the SIK3 immuno- force mediated by SIK3 to facilitate the cell cycle reactivity was either not or rarely found in normal progression using a pharmacological approach, and ovarian tissue and benign diseases such as adenomyosis identified that c-Src was the downstream signaling and leiomyoma, displaying 96.7% specificity to ovarian mediator of SIK3. This finding is also indirectly cancer in our cases. Despite SIK3 having a lower supported by our IHC study, showing that c-Src was sensitivity to ovarian cancer than CA125, we suggest coexpressed with SIK3 in 84.6% of the SIK3-positive that the combined detection of SIK3 and CA125 may ovarian cancer samples. However, it is not clear how improve the specificity and success of ovarian cancer SIK3 serine/threonine kinase can phosphorylate the detection in clinical use. However, a larger cohort of Tyr416 residue of c-Src. On the basis of our current data patients with ovarian cancer and other cancer types such revealing that the c-Src phosphorylation event could be as brain, lung, breast, liver, colon and cervical cancers is affected by treatment of Brefeldin A (unpublished data), still required to further define the SIK3 diagnostic we therefore speculate that the activation of c-Src by specificity and prognostic values against ovarian cancers SIK3 may be through a secretory action, and wonder in future. if STC1 may be a potential target gene downstream to This study identified SIK3 as an ovarian TAA and SIK3. In spite of this assumption, we still cannot exclude for the first time reported that its overexpression was a possibility that SIK3 may recruit other tyrosine closely associated with ovarian cancers. Additionally, kinases to phosphorylate c-Src because SIK3 possesses we also further elucidated the mechanisms involved in a proline-rich region located in its N-terminus. SIK3-mediated ovarian tumorigenesis. We believe that SIK3 is an immunogenic target of the malignant identification of TAAs including SIK3 is not only a ascitic antibodies of an ovarian cancer patient, and crucial step in understanding their roles on a molecular shows higher specificity to ovarian cancer than CA125. basis in cancer immunology, but it also bestows both Unlike the tissue preferences of SIK1 and SIK2, SIK3 is diagnostic and therapeutic advantages to clinical appli- ubiquitously distributed in a variety of tissues but with cations of ovarian cancers.

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3582 Materials and methods lyzed in cell lysis buffer (Chang et al., 2006), and the protein concentration was quantified using the BCA assay (Thermo Additional methods and materials available from Supplementary Fisher). The lysates were resolved in 7.5–10% polyacrylamide Information gel containing sodium dodecyl sulfate, transferred to nitro- Detailed description of screening for phage library for cellulose membranes, and probed with purified CA502 tumor-associated peptides, immunohistochemistry, statistics, ascitic antibody (32 mg/ml) or antibody specific to Myc tag, northern blot analysis, gene construct and antiserum genera- GST-tag (Thermo Fisher), p21Waf/Cip1, p27Kip, cyclin Ds, cyclin tion are available from the Supplementary Materials and E, pTyr416-Src, c-Src (Cell Signaling Technology, Beverly, MA, methods. USA) or SIK3. For detection of CRBN expression in ovarian cancer cells, GST-CRBN recombinant protein and CRBN Cells and cell culture antibody were purchased from Abnova Co. (Taipei, Taiwan). SK-OV3 and OC109 cells were cultured in McCoy’s 5A b-Actin served as a protein loading control. The immuno- medium (Invitrogen, Carlsbad, CA, USA) supplemented with complexes were detected by probing with anti-mouse, -rabbit 10% fetal bovine serum (FBS), penicillin (100 units/ml)/ or -human IgG conjugated with horseradish peroxidase streptomycin (100 mg/ml), 0.1 mM non-essential amino acids, (Jackson ImmunoResearch, Cambridgeshire, UK), and were and 1 mM sodium pyruvate (Invitrogen). OVCAR3 cells visualized using the SuperSignal detection system (Thermo were maintained in RPMI1640 (Invitrogen) containing 10% Fisher). FBS. TOV-112D as well as IOSE385 cells were maintained in MCDB105 medium/medium 199 (1:1) (Sigma-Aldrich, Cell growth assay St Louis, MO, USA) plus 15% FBS. HeLa cells were maintained The parental and stably transfected cells (2 Â 104 cells per well) in Dulbecco’s modified Eagle’s medium containing 10% FBS. were seeded in 24-well plates and incubated for 24, 48, 72 or To generate cells with stable overexpression or attenuated 96 h (Figure 2B). Cells (3 Â 104 cells per well) were incubated expression of SIK3, OVCAR3 and SK-OV3 cells were trans- for 24 h and followed by PP2 treatment for 24, 48 or 72 h in fected with the pcDNA-SIK3 plasmid and a plasmid trans- Figure 5c. Cell proliferation was determined by incubating cribing SIK3-specific interfering small hairpin RNA (50-CCG the cells with 200 ml of fresh medium containing 1 mg/ml GGCCAGGCTTTATCTTATCAAACTCGAGTTTGATAAG 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide ATAAAGCCTGGCTTTTTG-30) (Open Biosystems, Huntsville, (MTT) (Sigma-Aldrich) for 4 h at 37 1C. After removal of the AL, USA), respectively. Simultaneously, their corresponding MTT solution, the resulting formazan crystals were dissolved control cells were also established by transfection of either completely in an ethanol/dimethyl sulfoxide mixture (1:1) pcDNA3.1 or pLKO.1 vector and selection in the medium and quantified using an enzyme-linked immunosorbent assay containing 400 mg/ml of G418 or 2.5 mg/ml of puromycin, plate reader (Metertech 960, Metertech Inc., Taipei, Taiwan) respectively. After limiting dilution, the expression levels of by measuring the absorbance at 550 nm. Triplicate wells were individual cell clones were confirmed by immunoblotting assayed for each experiment and three independent experi- analyses. ments were performed. Data are expressed as the mean of OD550±s.d. Ascitic fluids, sera and tissue samples Twelve malignant ascitic fluid samples were obtained from Clonogenic assay ovarian cancer patients with 10 serous, 1 mucinous and The SIK3 stable transfectants (clone ov-21) and control cells 1 cystadenocarcinoma subtypes. Paraffin-embedded tissue were seeded in six-well culture plates at 250, 500 or 1000 cells samples were collected with the permission of the Institution per well in triplicate. Cells were maintained in culture for 9 Review Board of the Taipei Medical University. The samples days, with medium change every 3 days, to let the viable cells included 20 adenomyosis, 20 leiomyoma, 20 normal ovary and propagate to sizable colonies. After phosphate-buffered saline 38 ovarian cancer samples. The histological subtypes of the wash, the cells were stained with methylene blue (USB, Cleveland, ovarian cancers included 13 serous, 8 endometrioid, 6 clear OH, USA) for 2 h in room temperature and photographed under cell, 3 mucinous, 3 endometrioid/clear cell mixed subtypes a microscopy. After cell lysis in 200 ml of 1% Lauroylsarcosine and 5 poorly differentiated samples. Normal ovarian tissues solution (Sigma-Aldrich) per well, the dye was released and were obtained from patients who received total hyster- transferred to a new 96-well plate. The concentrations of ectomy and oophorectomy without ovarian lesions. Control individual wells were quantified using SpectraMax M5 ELISA serum samples were obtained from 10 age-matched healthy plate reader (Molecular Devices, Sunnyvale, CA, USA) by women. The IgG fractions of the ascitic fluids and control measuring the absorbance at 550 nm. Data are expressed as serum samples were purified with Protein A/G (Thermo the mean of OD550±s.d. Fisher, Rockford, IL, USA), according to the manufacturer’s instructions. Cell cycle analysis For synchronization of the cell cycle, stable transfectants Western blot analysis (1 Â 105 per well) were seeded in six-well plates and incubated Cell lysates were prepared as previously described (Chang in media supplemented with 0.1% FBS for 24 h. Subsequently, et al., 2006). Briefly, cells were transfected with plasmids using the cells were incubated with 10% FBS-containing fresh Lipofectamine for HeLa cells, Lipo2000 (Invitrogen) for medium for 6 h. For transient transfection experiments, cells SK-OV3, and ExGene 500 (Fermentas, Vilnius, Lithuania) (1 Â 105 per well) were seeded, incubated overnight, and for OVCAR3 cells. After transfection, cells were treated with transfected with 1 or 2 mg of expression plasmids to over- or without various pharmacological inhibitors such as express SIK3 or c-Src or with interfering small hairpin RNA LY294002 for PI3K, PD98059 for mitogen-activated protein/ plasmids to knockdown SIK3, c-Src or p21Waf/Cip1 expression, extracellular signal-regulated kinase 1, rapamycin for mechan- as indicated in the figure legends of Figures 4 and 7, for 48 h. istic target of rapamycin-mediated S6 kinase, staurosporine for After being trypsinized, washed with phosphate-buffered broad-spectrum protein kinases, H89 for and saline twice and fixed with 70% cold ethanol, the cells were PP2 for Src-family tyrosine kinases. Afterward, the cells were collected by centrifugation at 5000 Â g for 5 min, and then

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3583 incubated with RNase A (1 mg/ml in phosphate-buffered saline) Conflict of interest for 30 min. Finally, the cells were stained with 50 mg/ml propi- dium iodide after phosphate-buffered saline washes twice. Flow The authors declare no conflict of interest. cytometric data (1 Â 104 events) were acquired using a BD FACScaliber cytometer, and the percentage of cells in each cell cycle stage was determined using ModFit software Acknowledgements (BD Biosciences, San Jose, CA, USA). This work was partially supported by the Intramural Research Tumor growth in animals Grant of the National Health Research Institutes, the National NOD-SCID (6–8 weeks old) male mice were obtained from the Research Program for Genomic Medicine Grant from the Depart- National Laboratory Animal Center, Taiwan, and housed ment of Health, Executive Yuan, Taiwan, DOH97-TD-G-111-022 under specific pathogen-free conditions according to the (NYS), and by the National Science Council Grants of Taiwan, guidelines of the Animal Care Committee at the National NSC 99-2320-B-400-011-MY3 (NYS), NSC 96-2314-B-038-022 Health Research Institutes, Taiwan. On day 0, SK-OV3/sh-7 (SJL), NSC 97-2314-B-038-004-MY3 (SJL), as well as NSC 96- (1 Â 106/200 ml per mouse) or OVCAR3/ov-17 (1 Â 107/200 ml 2320-B-038-005/NSC 97-2320-B-038-012-MY2 (YYY). We are per mouse) stably transfected cells were injected subcuta- also grateful to Dr Nagase from the Kazusa DNA Research neously into the right side of the dorsal flanks of six mice and Institute for kindly providing the FJ10213 plasmid to facilitate their corresponding control cells were injected into the left side the SIK3 study, and thank National Research Program for of the same mice. Tumor volume was measured weekly for Genomic Medicine, Taiwan (National Clinical Core, NSC-95- up to 9 weeks by the following formula: length (mm) Â width2 3112-B-001-010 and National Genotyping Core, NSC-95-3112-B- (mm2) Â (p/6) (Ruda et al., 2006). 001-011) for supply of interfering small hairpin RNA plasmids.

References

Ahmed AA, Lu Z, Jennings NB, Etemadmoghadam D, Capalbo L, Guarino M. (2007). Epithelial-mesenchymal transition and tumour Jacamo RO et al. (2010). SIK2 is a centrosome kinase required for invasion. Int J Biochem Cell Biol 39: 2153–2160. bipolar mitotic spindle formation that provides a potential target for Horike N, Takemori H, Katoh Y, Doi J, Min L, Asano T et al. (2003). therapy in ovarian cancer. Cancer Cell 18: 109–121. Adipose-specific expression, phosphorylation of Ser794 in insulin Amin N, Khan A, St Johnston D, Tomlinson I, Martin S, Brenman J receptor substrate-1, and activation in diabetic animals of salt- et al. (2009). LKB1 regulates polarity remodeling and adherens inducible kinase-2. J Biol Chem 278: 18440–18447. junction formation in the Drosophila eye. Proc Natl Acad Sci USA Katoh Y, Takemori H, Horike N, Doi J, Muraoka M, Min L et al. 106: 8941–8946. (2004). Salt-inducible kinase (SIK) isoforms: their involvement Baron AT, Boardman CH, Lafky JM, Rademaker A, Liu D, Fishman in steroidogenesis and adipogenesis. Mol Cell Endocrinol 217: DA et al. (2005). Soluble epidermal growth factor receptor (sEGFR) 109–112. [corrected] and cancer antigen 125 (CA125) as screening and Katoh Y, Takemori H, Lin XZ, Tamura M, Muraoka M, Satoh T diagnostic tests for epithelial ovarian cancer. Cancer Epidemiol et al. (2006). Silencing the constitutive active transcription Biomarkers Prev 14: 306–318. factor CREB by the LKB1-SIK signaling cascade. FEBS J 273: Berdeaux R, Goebel N, Banaszynski L, Takemori H, Wandless T, 2730–2748. Shelton GD et al. (2007). SIK1 is a class II HDAC kinase that Kowanetz M, Lonn P, Vanlandewijck M, Kowanetz K, Heldin CH, promotes survival of skeletal myocytes. Nat Med 13: 597–603. Moustakas A. (2008). TGFbeta induces SIK to negatively regulate Brewer BG, Mitchell RA, Harandi A, Eaton JW. (2009). Embryonic type I receptor kinase signaling. J Cell Biol 182: 655–662. vaccines against cancer: an early history. Exp Mol Pathol 86: 192–197. Liu G, Yang G, Chang B, Mercado-Uribe I, Huang M, Zheng J et al. Bright NJ, Thornton C, Carling D. (2009). The regulation and function of (2010). Stanniocalcin 1 and ovarian tumorigenesis. J Natl Cancer mammalian AMPK-related kinases. Acta Physiol (Oxf) 196: 15–26. Inst 102: 812–827. Cannistra SA. (2004). Cancer of the ovary. N Engl J Med 351: Lizcano JM, Goransson O, Toth R, Deak M, Morrice NA, Boudeau J 2519–2529. et al. (2004). LKB1 is a master kinase that activates 13 kinases Cantley LC. (2002). The phosphoinositide 3-kinase pathway. Science of the AMPK subfamily, including MARK/PAR-1. Embo J 23: 296: 1655–1657. 833–843. Chang GC, Liu KJ, Hsieh CL, Hu TS, Charoenfuprasert S, Liu HK Mizrachy-Schwartz S, Kravchenko-Balasha N, Ben-Bassat H, Klein S, et al. (2006). Identification of alpha-enolase as an autoantigen in Levitzki A. (2007). Optimization of energy-consuming pathways lung cancer: its overexpression is associated with clinical outcomes. towards rapid growth in HPV-transformed cells. PLoS One 2: e628. Clin Cancer Res 12: 5746–5754. O’Brien TJ, Beard JB, Underwood LJ, Shigemasa K. (2002). The CA Cheng H, Liu P, Wang ZC, Zou L, Santiago S, Garbitt V et al. (2009). 125 gene: a newly discovered extension of the glycosylated SIK1 couples LKB1 to p53-dependent anoikis and suppresses N-terminal domain doubles the size of this extracellular super- metastasis. Sci Signal 2: ra35. structure. Tumour Biol 23: 154–169. Deng C, Zhang P, Harper JW, Elledge SJ, Leder P. (1995). Mice Okamoto M, Takemori H, Katoh Y. (2004). Salt-inducible kinase in lacking p21CIP1/WAF1 undergo normal development, but are steroidogenesis and adipogenesis. Trends Endocrinol Metab 15:21–26. defective in G1 checkpoint control. Cell 82: 675–684. Old LJ. (2008). Cancer vaccines: an overview. Cancer Immun Gagnon A, Kim JH, Schorge JO, Ye B, Liu B, Hasselblatt K et al. 8(Suppl 1): 1. (2008). Use of a combination of approaches to identify and validate Paciga M, McCudden CR, Londos C, DiMattia GE, Wagner GF. relevant tumor-associated antigens and their corresponding auto- (2003). Targeting of big stanniocalcin and its receptor to lipid antibodies in ovarian cancer patients. Clin Cancer Res 14: 764–771. storage droplets of ovarian steroidogenic cells. J Biol Chem 278: Giancotti FR, Dorsett BH, Qian HN, Cronin WJ, Barber HR, 49549–49554. Ioachim HL. (1990). Ovarian cancer-associated antibodies recovered Reuschenbach M, von Knebel Doeberitz M, Wentzensen N. (2009). from ascites: their use for the isolation of ovarian cancer-associated A systematic review of humoral immune responses against tumor antigen to produce monoclonal antibodies. Gynecol Oncol 37: 24–28. antigens. Cancer Immunol Immunother 58: 1535–1544.

Oncogene SIK3 is involved in ovarian tumorigenesis S Charoenfuprasert et al 3584 Rosenthal AN, Menon U, Jacobs IJ. (2006). Screening for ovarian Serlachius M, Alitalo R, Olsen HS, Andersson LC. (2002). Expression cancer. Clin Obstet Gynecol 49: 433–447. of stanniocalcin-1 in megakaryocytes and platelets. Br J Haematol Ruda JM, Beus KS, Hollenbeak CS, Wilson RP, Stack Jr BC. (2006). 119: 359–363. The effect of single agent oral fusaric acid (FA) on the growth of Sjostrom M, Stenstrom K, Eneling K, Zwiller J, Katz AI, subcutaneously xenografted SCC-1 cells in a nude mouse model. Takemori H et al. (2007). SIK1 is part of a cell sodium- Invest New Drugs 24: 377–381. sensing network that regulates active sodium transport through Runnebaum IB, Stickeler E. (2001). Epidemiological and molecular a calcium-dependent process. Proc Natl Acad Sci USA 104: aspects of ovarian cancer risk. J Cancer Res Clin Oncol 127: 73–79. 16922–16927. Screaton RA, Conkright MD, Katoh Y, Best JL, Canettieri G, Taylor DD, Gercel-Taylor C, Parker LP. (2009). Patient-derived Jeffries S et al. (2004). The CREB coactivator TORC2 functions as a tumor-reactive antibodies as diagnostic markers for ovarian cancer. calcium- and cAMP-sensitive coincidence detector. Cell 119: 61–74. Gynecol Oncol 115: 112–120.

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

Oncogene