Int J Clin Exp Pathol 2016;9(10):10451-10457 www.ijcep.com /ISSN:1936-2625/IJCEP0031885

Original Article CDC5L contributes to malignant cell proliferation in human osteosarcoma via regulation

Yu Wang1,4*, Hong Chang2,4*, Di Gao3,4, Lei Wang4, Nan Jiang4, Bin Yu4

1Department of Orthopaedics, Chifeng Hospital, Inner Mongolia, China; 2Department of Orthopaedics, 421 Hospital of PLA, Guangzhou, China; 3Department of Orthopaedics, The University of Hong Kang, Shenzhen Hospital, Shenzhen, China; 4Department of Orthopaedics and Traumatology, Nanfang Hospital Southern Medical University, Guangzhou, China. *Equal contributors. Received May 9, 2016; Accepted July 22, 2016; Epub October 1, 2016; Published October 15, 2016

Abstract: Cell division cycle 5-like (CDC5L) has been reported in overexpressed in osteosarcoma (OS). However, its biological function in tumor biology was still unclear. Here, we firstly determined the expression of CDC5L in several OS cell lines, including Saos-2, SF-86, U2OS and SW1353, and found it was commonly upregulated in these four OS cells. Subsequently, Saos-2 and U2OS cells with higher CDC5L expression were transfected with interfering RNA tar- geting CDC5L. A set of functional assay was conducted on the two cell lines, including CCK-8, colony formation and flow cytometry assay. Our results indicated that CDC5L silencing significantly inhibited cell proliferation and arrested cell cycle at G2/M phase. Mechanically, Western blot analysis further confirmed knockdown of CDC5L remarkably down regulated the levels of CDK1, Cyclin B and PCNA. There findings further demonstrated that CDC5L play a crucial role in OS development and might be a promising therapeutic target of OS.

Keywords: CDC5L, osteosarcoma, cell proliferation, cell cycle, CDK1, Cyclin B

Introduction reported to shorten G2 phase of cell cycle, but its knockdown could delay mitotic entry by Osteosarcoma (OS) is the third most common arresting G2/M phase transition [6], as well as primary malignant bone tumor characterized as impair the pre-mRNA splicing efficiency. In addi- aggressive growth and proliferation [1]. For the tion, CDC5L has been demonstrated to regulate last few decades, despite remarkably advanc- expression and splicing efficiency of a set of es in the combination of surgery and chemo- involved in the mitosis and DNA damage therapy has been achieved, a significant pro- repair [7]. Furthermore, previous study sug- portion of OS patients shows high grade and gests that CDC5L is overexpressed in several poor prognosis [2]. Recently, molecular target- tumors, including glioma [8] and hepatocellular ed therapy has greatly advanced our knowl- carcinoma [9]. Moreover, knockdown of CDC5L edge of tumor pathogenesis and treatments, significantly inhibits tumor cell proliferation in glioma and hepatocellular carcinoma through which is particularly necessary to develop cell cycle arrest at G2/M phase. Recent studies effective therapeutic strategies for the treat- have described overexpressed CDC5L has a ment of OS. Therefore, it is urgent to identify poor prognosis in OS [10], but its exact biologi- novel molecular mechanisms associated with cal function and molecular mechanism remains OS initiates and progression. largely unknown in OS cells.

As our best knowledge, uncontrolled tumor In the present study, we aimed to investigate growth is closely correlated with aberrant cell the biological function of CDC5L in human OS cycle control. Cell division cycle 5-like (CDC5L), cell lines and explore its molecular mechanism identified as a pre-mRNA splicing factor and by RNA interference technology. Our findings core component of human Prp19/Cdc5L com- might help us have a better understanding with plex, plays an important role in regulating cell the biological function of CDC5L and further division cycle [3-5]. Overexpression of CDC5L is confirmed CDC5L is an oncogene in OS. CDC5L contributes osteosarcoma

Materials and methods Western blot analysis

Cell cultures and transient transfection Whole were obtained from cell sam- ples with 2X SDS lysis buffer containing 100 Human OS cell lines (Saos-2, U2OS, SW1353 mM Tris-Hcl (pH 6.8), 10 mM EDTA, 4% SDS and and SF-86) were provided by Cell Library, China 10% Glycine. The protein concentrations were Academy of Science and cultured in Dulbecco’s measured using Bicinchoninic acid (BCA) pro- modified Eagle’s medium (DMEM, Hyclone, tein assay kit (Thermo Fisher Scientific). App- USA) supplemented with 10% fetal bovine roximately 30 µg proteins were separated by serum. All cells were maintained in a 5% CO2 10% polyacrylamide-sodium dodecyl sulfate atmosphere at 37°C. (SDS-PAGE) electrophoresis and transferred to polyvinylidene fluoride (PVDF) membranes for The sequences of small interfering RNA (siRNA) 2.0 h at 300 mA. The membranes were blocked targeting human CDC5L (5’-GAGTGAAATTGC- with 5% non-fat milk in Tris-buffered saline for ACGTCAA-3’) and the negative control (NC) 1 h, and then incubated with primary antibod- siRNA were designed and synthesized by ies against CDC5L, CDK1, Cyclin B, PCNA and GeneChem (Shanghai, China). Saos-2 and 5 GAPDH overnight at 4°C. Then the membranes U2OS cells (1 × 10 cells per well) were seeded were washed three times for 10 min each time in six-well plates and then transfected with siR- with Tween 20-PBS and incubated with HRP- NAs using the Lipofectamine 2000 reagent conjugated secondary antibodies for 2 h at according to the manufacturer’s instructions room temperature. The expression levels of (Invitrogen, Carlsbad, CA, USA). The control OS proteins were analyzed using Super ECL De- cells were transfected with empty vectors. The tection reagent according to the manufacture’s cells were divided into three groups: The blank instruction. control (Con), empty vector (NC) and siRNA transfection (KD) groups. After culture for 48 h, Cell viability assay cells were collected and used for the following assays. To determine the effect of CDC5L on OS cell growth, Cell Counting Kit-8 (CCK-8) assay was Real-time PCR analysis performed on OS cells lines, Saos-2 and U2OS. Briefly, cells were seeded in 96-well plates at a Real-time PCR was used to validate the knock- density of 2000 cells per well and added 10 μL down efficiency induced by siRNA by determin- CCK-8 (Beyotime Biotechnology, China) after 1, ing CDC5L mRNA levels. Briefly, total RNA was 2, 3, 4 and 5 days, respectively. The optical extracted from cultured cells using TRIzol density (OD) of each well was determined using Reagent (Invitrogen) and used to synthesized a microplate reader at a wavelength of 450 nm. cDNA using MMLV reverse transcriptase (Pro- Each sample was repeated three times. mega, USA). Primers (forward: 5’-CAGGCTCCT- ACAGGACTACC-3’ and reverse: 5’-GGAACTTC- Colony formation assay CTGACCCTTGTG-3’) were designed to deter- mine CDC5L expression. GAPDH (forward: To evaluate the monolayer colony formation in 5’-TGACTTCAACAGCGACACCCA-3’ and reverse: OS cells, stably transfected cells from three 5’-CACCCTGTTGCTGTAGCCAAA-3’) was used as groups (Con, NC and KD) were seeded into six- endogenous control. Real-time PCR was per- well plates at density of 400 cells per well and formed on TAKARA TP800-Thermal Cycler cultured until the colonies were visible. Then Dice™ Real-Time System with 20 µl PCR reac- cells were fixed with 95% ethanol for 10 min tion mixture (10 µl 2 × SYBR premix ex taq, 0.5 and stained with 0.1% crystals purples. Cell µl primers (2.5 µM), 5 µl cDNA and 4.5 µl cDNA). colony formation was observed under a light The thermal cycling conditions were initially microscope. The number of colonies containing denatured at 95°C for 15 s, and 45 cycles of 5 more than 50 cells was manually counted. s at 95°C and 20 s at 60°C. Each reaction was Each sample was repeated three times. performed with triplicate samples in each group and analyzed individually relative to Flow cytometry analysis of cell cycle GAPDH. The relative expression of CDC5L was calculated using 2-ΔΔCt method [11]. Each sam- Flow cytometer (FACS Calibur, BD Biosciences) ple was repeated three times. combined with propidium iodide (PI) was used

10452 Int J Clin Exp Pathol 2016;9(10):10451-10457 CDC5L contributes osteosarcoma

Figure 1. CDC5L was highly silenced in osteosarcoma cell lines. (A) The protein level of CDC5L in several osteosar- coma cell lines, including Saos-2, SF-86, U2OS and SW1353 using Western blot analysis (B and C) qPCR was used to determine the mRNA levels of CDC5L in Saos-2 and U2OS following CDC5L knockdown by siRNA transfection. (D and E) Western blot assay was used to determine the mRNA levels of CDC5L in Saos-2 and U2OS following CDC5L knockdown by siRNA transfection. Con: The blank control, NC: Empty vector and KD: siRNA transfection groups; all the results were represented as mean ± SD from three independent experiments. ***P < 0.001 as compared with NC cells; GAPDH was used as an internal control. to determine cell cycle progression in OS cells. Results Briefly, cells were reseeded on 6-cm dishes at a density of 2 × 105 cells per dish and fixed in CDC5L was significantly downregulated in OS 70% cold ethanol overnight at 4°C, followed by cells using siRNA transfection incubation with 300 µl PBS containing propidi- um iodide (PI) for 30 min at 37°C in dark, then To investigate whether CDC5L was correlated subjected to flow cytometry. Triplicate indepen- with OS progression, the proteins levels of dent experiments were performed. CDC5L were firstly determined in several OS cell lines using Western blot analysis. As shown Statistical analysis in Figure 1A, CDC5L was highly expressed in U2OS and Saos-2 than in other cell lines, which Quantitative data are analyzed using SPSS may be ascribed to specific cell type. Therefore, 13.0 software and presented as mean ± stan- U2OS and Saos-2 with higher CDC5L expres- dard deviations (SD) from at least three inde- sion were chosen to perform loss-of-function pendent experiments. Student’s t-test was assay. Subsequently, U2OS and Saos-2 cells used to compare the differences between two were transfected with CDC5L siRNA or nonspe- groups. A P-value of less than 0.05 was consid- cific siRNA. Knockdown efficiency was deter- ered statistically significant. mined 48 h post infection by qPCR and Western

10453 Int J Clin Exp Pathol 2016;9(10):10451-10457 CDC5L contributes osteosarcoma

Figure 2. CDC5L silencing significantly impaired cell viability and colony formation ability in osteosarcoma cells. A. The cell viability assay was conducted on Saos-2 and U2OS cell after CDC5L knockdown using CCK-8 analysis. B. Representative images showed the size and number of colonies in U2OS cells recorded under a light microscope after CDC5L knockdown. C. Statistical analysis of colony numbers of U2OS cells from Con, NC and KD groups. Con: The blank control, NC: Empty vector and KD: siRNA transfection groups; all the results were represented as mean ± SD from three independent experiments. **P < 0.01, ***P < 0.001 as compared with NC cells. blot analysis, which indicated that the mRNA structed, which could be used for all the subse- expression level of CDC5L was significantly quent experiments. downregulated nearly 85.0% in U2OS cells and 76.9% in Saos-2 cells after transfection with Interference of CDC5L remarkably suppressed siRNA (Figure 1B and 1C, P < 0.001). Similarly, cell proliferation in OS cells CDC5L protein levels were substantially de- creased in U2OS and Saos-2 cells transfected To determine the biological function of CDC5L with CDC5L siRNA, compared with control on OS cell proliferation, MTT and colony forma- siRNA or blank control (Figure 1D and 1E). tion were conducted in OS cells. As shown in These results indicated that two CDC5L silenc- Figure 2A, the growth curves of U2OS and ing OS cellular models were successfully con- Saos-2 cells were obviously lower in KD group

10454 Int J Clin Exp Pathol 2016;9(10):10451-10457 CDC5L contributes osteosarcoma

Figure 3. CDC5L silencing significantly induced cell cycle arrest at G2/M phase in osteosarcoma cells. (A) Rep- resentative pictures showed cell cycle distribution profile in U2OS cells following CDC5L knockdown using flow cytometry analysis. (B) Statistical analysis of G0/G1, S and G2/M phase populations in U2OS cells from Con, NC and KD groups. Con: The blank control, NC: Empty vector and KD: siRNA transfection groups; all the results were represented as mean ± SD from three independent experiments. **P < 0.01, ***P < 0.001 as compared with NC cells (C) Western blot analysis showed CDK1, Cyclin B and PCNA were downregulated in U2OS cells following CDC5L knockdown. GAPDH was used as an internal control. compared with NC or Con group (P < 0.001), U2OS cells after transfection with siRNA (Figure suggesting siCDC5L strong inhibited cell growth 3A). As illustrated in Figure 3B, the percenta- in OS cells. Interestingly, U2OS cells were ge of cells in G2/M phase was significantly observed to grow slower than Saos-2 cells, increased, but that in G1 phase was obviously which selected for colony formation assay. As decreased in KD group, compared with those shown in Figure 2B, CDC5L silencing signifi- in NC or Con group (P < 0.001). Furthermore, cantly decreased the size of a single colony by Western blot analysis showed that interference crystal violet staining. Additionally, statistical of CDC5L remarkably down regulated cell cycle analysis demonstrated the number of colonies protein CDK1 and cyclin B, as well as the cell was remarkably decreased in KD group than proliferation marker PCNA in OS cells (Figure that in NC or Con group in U2OS cells (Figure 3C). Based on these findings, we could con- 2C, P < 0.001). clude that CDC5L silencing may suppress OS cells proliferation through inducing G2/M ph- CDC5L silencing significantly induced cell ase arrest. cycle arrest Discussion To further uncover the mechanism underlying CDC5L silencing inhibited OS cells prolifera- Recently, targeting cyclin-dependent kinases tion, flow cytometry analysis was performed on associated with cell cycle progression is becom-

10455 Int J Clin Exp Pathol 2016;9(10):10451-10457 CDC5L contributes osteosarcoma ing an effective and widely accepted approach Disclosure of conflict of interest for cancer therapy [12, 13]. CDC5L, as key reg- ulator for cell division cycle, has been shown to None. be overexpressed in OS, but its function signifi- cance remains unclear in OS cells. Thus, we Address correspondence to: Bin Yu, Department of conducted loss-of-function assay in OS cell Orthopaedics and Traumatology, Nanfang Hospital lines Saos-2 and U2OS cells using siRNA trans- Southern Medical University, 1838 North Guangzhou fection. Our results indicated that CDC5L Avenue, Guangzhou 510515, China. Tel: +86-20- silencing remarkably suppressed cell viability 61641741; E-mail: [email protected] and colony formation in two OS cell lines. References Furthermore, we found the inhibited cell prolif- eration is ascribed to G2/M phase arrest [1] Ottaviani G and Jaffe N. The epidemiology of induced by CDC5L silencing, as confirmed by osteosarcoma. Cancer Treat Res 2009; 152: downregulation of CDK1, Cyclin B and PCNA 3-13. expression. [2] Fu HL and Shao L. Silencing of NUF2 inhibits proliferation of human osteosarcoma Saos-2 These results suggested that CDC5L might pro- cells. Eur Rev Med Pharmacol Sci 2016; 20: mote the oncogenesis of OS through enhancing 1071-1079. cell cycle progression. Consistent with our data, [3] Burns CG, Ohi R, Krainer AR and Gould KL. CDC5L could promoted cell proliferation in glio- Evidence that Myb-related CDC5 proteins are ma and hepatocellular carcinoma [8, 9]. It has required for pre-mRNA splicing. Proc Natl Acad been widely accepted that cyclin-dependent Sci U S A 1999; 96: 13789-13794. [4] Ajuh P, Sleeman J, Chusainow J and Lamond kinases (CDKs) and cyclins play an important AI. A direct interaction between the carboxyl- role in eukaryotic cell cycle progression. The terminal region of CDC5L and the WD40 do- deregulation of CDK/cyclin complex activity is main of PLRG1 is essential for pre-mRNA splic- observed in a variety of human tumors [14]. ing. J Biol Chem 2001; 276: 42370-42381. CDK1, as the predominant member of CDK [5] Ajuh P and Lamond AI. Identification of peptide family, is involved in regulating G2/M transi- inhibitors of pre-mRNA splicing derived from tions [15]. Passage through G2 into M phase is the essential interaction domains of CDC5L also regulated by cyclin B-CDC2 complex [16]. and PLRG1. Nucleic Acids Res 2003; 31: In normal cell cycle, the levels of CDK1 and 6104-6116. cyclin B remain constant and their upregulation [6] Boudrez A, Beullens M, Groenen P, Van Eynde A, Vulsteke V, Jagiello I, Murray M, Krainer AR, will promote cell cycle progression. Based on Stalmans W and Bollen M. NIPP1-mediated these evidences, knockdown of CDC5L signifi- interaction of protein phosphatase-1 with cantly decreased the expression of CDK1 and CDC5L, a regulator of pre-mRNA splicing and cyclin B, which might powerfully support its mitotic entry. J Biol Chem 2000; 275: 25411- oncogenic role of CDC5L in OS cells. 25417. [7] Mu R, Wang YB, Wu M, Yang Y, Song W, Li T, In summary, our data further confirmed the Zhang WN, Tan B, Li AL, Wang N, Xia Q, Gong biological role of CDC5L in OS cells in vitro. WL, Wang CG, Zhou T, Guo N, Sang ZH and Li Knockdown of CDC5L could suppress cell pro- HY. Depletion of pre-mRNA splicing factor liferation in OS cells through inducing cell cycle Cdc5L inhibits mitotic progression and triggers arrest and might be a targeted anti-mitotic can- mitotic catastrophe. Cell Death Dis 2014; 5: cer therapy for OS. However, further investiga- e1151. tion is still necessary for the complex molecular [8] Chen W, Li Z, Yan W, Jie S, Wang D, Fan S, Na B, Zhu J, Ji B and Wang Y. Expression of CDC5L mechanism underlying CDC5L regulating cell is associated with tumor progression in glio- cycle in OS progression. mas. Tumour Biol 2015; 1-11. [9] Qiu H, Zhang X, Ni W, Shi W, Hui F, Jian X, Chen Acknowledgements Y, Ni R and Tao T. Expression and Clinical Role of Cdc5L as a Novel Cell Cycle Protein in This work is supported by grants from the Hepatocellular Carcinoma. Dig Dis Sci 2015; National Natural Science Foundation of China 1-11. (No. 81572165) and the Guangdong Provincial [10] Lu XY, Lu Y, Zhao YJ, Jaeweon K, Kang J, Xiao- Science and Technology Plan Projects (No. Nan L, Ge G, Meyer R, Perlaky L, Hicks J, 2016B090913004). Chintagumpala M, Cai WW, Ladanyi M, Gorlick

10456 Int J Clin Exp Pathol 2016;9(10):10451-10457 CDC5L contributes osteosarcoma

R, Lau CC, Pati D, Sheldon M and Rao PH. Cell [14] Malumbres M and Barbacid M. Cell cycle, cycle regulator CDC5L, a potential target CDKs and cancer: a changing paradigm. Nat for 6p12-p21 amplicon in osteosarcoma. Mol Rev Cancer 2009; 9: 153-166. Cancer Res 2008; 6: 937-946. [15] Morgan DO. Cyclin-dependent kinases: en- [11] Livak KJ and Schmittgen TD. Analysis of Re- gines, clocks, and microprocessors. Annu Rev lative Data Using Real-Time Cell Dev Biol 1997; 13: 261-291. Quantitative PCR and the 2-ΔΔCT Method. [16] Turowski P, Franckhauser C, Morris MC, Vaglio Methods 2001; 25: 402-408. P, Fernandez A, Lamb NJ. Functional cdc25C [12] Schwartz GK and Shah MA. Targeting the cell Dual-Specificity Phosphatase Is Required for cycle: a new approach to cancer therapy. J Clin S-Phase Entry in Human Cells. Mol Biol Cell Oncol 2005; 23: 9408-9421. 2003; 14: 2984-2998. [13] Asghar U, Witkiewicz AK, Turner NC and Knudsen ES. The history and future of target- ing cyclin-dependent kinases in cancer thera- py. Nat Rev Drug Discov 2015; 14: 130-146.

10457 Int J Clin Exp Pathol 2016;9(10):10451-10457