Tumor Biol. DOI 10.1007/s13277-016-4905-5

ORIGINAL ARTICLE

The interaction of hepatoma-derived growth factor and β-catenin promotes tumorigenesis of synovial sarcoma Jianming Tang 1 & Huijuan Shi 1 & Hui Li 1 & Tiantian Zhen 1 & Yu Dong 1 & Fenfen Zhang 1 & Yang Yang 1 & Anjia Han 1

Received: 17 November 2015 / Accepted: 22 January 2016 # International Society of Oncology and BioMarkers (ISOBM) 2016

Abstract To clarify the clinicopathological and biological role of hepatoma-derived growth factor (HDGF) and β-catenin in synovial sarcoma. Our results showed that histological type and HDGF/β-catenin expression were the two important independent prognostic factors for overall survival in synovial sarcoma patients. HDGF knockdown dramatically inhibited cellular proliferation, colony formation, and migration but induced G1 phase arrest and apoptosis in SW982 cells. Recombinant HDGF enhanced synovial sarcoma cell growth and partially retrieved the cell growth suppression in SW982 cells upon HDGF knockdown. HDGF knockdown dramatically suppressed β-catenin and its downstream gene expression in SW982 cells. Intriguingly, β-catenin knockdown dramatically suppressed HDGF expression in SW982 cells. A direct interaction of HDGF and β-catenin was found in SW982 cells. Three HDGF-binding elements in β-catenin promoter were found and specific for transcriptional activation of β-catenin in SW982 cells. In conclusion, our findings first indicate that the interaction of HDGF and β-catenin may play a crucial role in tumorigenesis of synovial sarcoma.

Keywords Hepatoma-derived growth factor . β-Catenin . Synovial sarcoma . Tumorigenesis

Jianming Tang, Huijuan Shi and Hui Li contributed equally to this work. * Anjia Han [email protected]

1

Department of Pathology, the First Affiliated Hospital, Sun Yat-Sen University, 58, Zhongshan Road II, Guangzhou 510080, China

Introduction Synovial sarcoma is a high-grade soft tissue malignancy, which displays a variable degree of epithelial differentiation, and has a specific chromosomal translocation t(X; 18), which leads to the fusion of the SS18 gene to one of three SSX genes (SSX1, SSX2, or SSX4). The 5- and 10-year disease-specific survival is 62 and 52 % in adults, respectively [1]. Hepatoma-derived growth factor (HDGF) is a heparin-binding protein that was originally purified from the conditioned media of human hepatocellular carcinoma cell line HuH-7, which proliferate autonomously in serum-free chemically defined medium [2]. HDGF is a prognostic marker in several types of cancer including hepatocellular carcinoma, esophageal squamous cell carcinoma, gastric cancer, and nonsmall cell lung cancer [3–6]. Our recent study shows that HDGF exhibits oncogenic properties and may be a novel prognostic factor in Ewing’s sarcoma [7]. β-Catenin is a key player in Wnt signaling pathway. Oncogenic activation of Wnt signaling pathway is mandatory for the initial neoplastic transformation of intestinal epithelium [8]. Nuclear accumulation of β-catenin as a cell signaling event may play an important role in the progression of synovial sarcoma [9]. Abnormal levels of β-catenin could contribute to the development and progression of synovial sarcoma, through increasing the proliferative activity of the tumor cells [10]. Upregulation of the Wnt/ β-catenin cascade by SYT-SSX2 correlates with its nuclear reprogramming function in synovial sarcoma [11]. SS18SSX-induced Wnt/β-catenin signaling appears to be of crucial biological importance in synovial sarcoma tumorigenesis and progression [12]. The clinicopathological and biological significance of HDGF in synovial sarcoma remain unknown. Whether

Tumor Biol.

HDGF regulates β-catenin signaling pathway in synovial sarcoma is unclear. Our current study is to investigate the clinicopathological and biological significance of HDGF and underlying mechanisms of HDGF regulating β-catenin signaling pathway in synovial sarcoma.

Materials and methods Cell lines and siRNA sequences The human synovial sarcoma cell line SW982 was maintained in Leibovitz’s L-15 Medium (Invitrogen, Carlsbad, CA), and human fibroblast-like synoviocyte (FLS) was cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen, Carlsbad, CA). All medium were supplemented with 10 % (v/v) fetal bovine serum (Invitrogen) and 1× antibiotic/antimycotic (100 units/mL streptomycin, 100 units/mL penicillin, and 0.25 mg/mL amphotericin B). All cell lines were cultured in humidified incubator at 37 °C with 5 % CO2. The targeted HDGF sequences were as follows: sense, 5′-CAA GGA GAA GAA CGA GAA AdTdT-3′.The targeted β-catenin sequences were as follows: sense, 5′-GCC ACA AGA UUA CAA GAA AdTdT-3′. The small interfering RNA (siRNA) duplexes were chemically synthesized and purified by Ribobio Co. Ltd (Guangzhou, China) and transfected using Lipofectamine RNAiMAX transfection reagent (Invitrogen). Lipofectamine RNAiMAX alone (Mock) and scrambled siRNA (NC-siRNA) were used as negative control groups.

Quantitative real-time PCR analysis As our previously described [7], quantitative real-time PCR was run on ABI StepOne Plus PCR system (Applied Biosystems). The primer sequences used for HDGF were forward, 5′-AGG CGG AAA CCG TGT A-3′, and reverse, 5′-CCA GGA ATG CCG TCT C-3′. The primer sequences used for β-catenin were forward, 5′-TTG AAA ATC CAG CGT GGA CA-3′, and reverse, 5′-TCG AGT CAT TGC ATA CTG TC-3′. The primer sequences used for c-Myc were forward, 5′-TGG TCG CCC TCC TAT GTT G-3′, and reverse 5′-CCG GGT CGC AGA TGA AAC TC-3′. The primer sequences used for cyclin D1 were forward, 5′-CCG CTC GAG ATG GAA CAC CAG CTC C-3′, and reverse: 5′-CCC AAG CTT CAG ATG TCC ACG TCC CGC ACG T-3′. The geometric mean of housekeeping gene glyceraldehyde-3phosphate dehydrogenase (GAPDH) was used to normalize the variability at mRNA expression levels. All experiments were performed in triplicate. Cell proliferation assay SW982 cells (1 × 104) were plated onto 96-well plates with medium containing 10 % FBS and incubated overnight. After transfection with 100 nM HDGF-siRNA, cell proliferation was determined using the Cell Counting Kit (CCK8) (Keygene, China). The absorbance (OD) was measured at a wavelength of 450 nm using a Microplate Autoreader (Bio-Tek Instruments, VT). NC-siRNA was used as the control group. This experiment was performed in triplicate. Colony formation assay

Patient information and tissue specimens A total of 63 samples of paraffin-embedded synovial sarcoma tissues and one pair of fresh synovial sarcoma tissues and their respective adjacent nontumor (ANT) samples between 1994 and 2013 were collected from our Department of Pathology. Prior patient consent and approval from the Institutional Research Ethics Committee were obtained. No patients had received chemotherapy or radiotherapy before operation. The histopathology of the disease was determined by two pathologists according to WHO classification of tumors of soft tissue and bone (2013). Thirtytwo samples of synovial sarcoma collected between 2008 and 2013 were also confirmed by fluorescent in situ hybridization (FISH) analysis using a LSI SYT (18q11.2) Dual Color, Break-Apart Rearrangement Probe kit (Vysis, Downers Grove, IL). Clinical staging was done according to UICC classification. Pertinent follow-up information was available for all patients. Detailed clinical information is summarized in Table 1.

After 48-h 100 nM HDGF-siRNA transfection, 600 SW982 cells were plated onto six-well plates and incubated at 37 °C in a 5 % CO2 incubator for 2 weeks. Fresh medium was added every 4 days. NC-siRNAwas used as the control group. At the endpoint, the cells were washed with cold phosphate-buffered saline (PBS), fixed with 4 % paraformaldehyde for 30 min, and stained with 1 % crystal violet solution for 20 min at room temperature. The visible colony numbers were counted. This experiment was performed in triplicate. Cell cycle and apoptosis assay SW982 cells (3 × 105) were seeded in six-well plates and incubated overnight until 50 % confluent, then transfected with 100 nM HDGF siRNA for 48 h, washed in cold PBS, fixed with 70 % cold ethanol for 24 h at 4 °C, then stained with propidium iodide buffer (50 mg/ml propidium iodide, 0.1 % sodium citrate, and 0.1 % Triton X-100) for 30 min at room temperature. Cells (2 × 104) were analyzed for cell cycle and

Tumor Biol. Table 1

The relationship between HDGF and β-catenin expression and clinicopathological features of synovial sarcoma

Characteristic

High (%)

Low (%)

p value

Abnormal (%)

Normal (%)

p value

9 (14.2) 14 (22.4)

20 (31.7) 20 (31.7)

0.405

15 (23.8) 19 (30.1)

14 (22.2) 15 (23.9)

0.741

0.459

19 (30.1)

9 (14.2)

0.048

0.146

15 (23.8) 14 (22.2)

20 (31.9) 6 (9.5)

0.082

0.87

20 (31.9) 27 (42.8)

23 (36.4) 23 (36.5)

0.992

Gender

Male Female

Age (years)

≥34

11 (17.4)

17 (26.9)

Primary location

<34 Axial

12 (19.0) 10 (15.8)

23 (36.7) 10 (15.8)

Histological type

Peripheral Monophasic

13 (20.6) 18 (28.5)

30 (47.8) 32 (50.8)

Biphasic

5 (7.9)

8 (15.8)

≤4.8 >4.8

10 (15.8) 5 (7.9)

20 (31.7) 10 (15.8)

Metastasis

Not known No

8 (13.1) 10 (15.8)

10 (15.8) 27 (42.8)

Relapse

Yes No

13 (20.8) 2 (3.0)

13 (20.8) 14 (22.2)

Yes Not known

20 (31.7) 1 (1.5)

23 (37.1) 3 (4.5)

I–II III–IV

7 (11.1) 16 (25.3)

27 (42.8) 13 (20.8)

Tumor size (cm)

TNM

β-Catenin expression

HDGF expression

apoptosis using a Becton Dickinson FACScan (Becton Dickinson Immunocytometry Systems, San Jose, CA). The percentage of cells in each phase of the cell cycle and apoptotic cells was quantified using Cell Quest software, respectively. This experiment was performed in triplicate.

7 (11.1)

6 (9.6)

16 (25.3) 9 (14.2)

14 (22.2) 7 (11.1)

0.028

5 (7.9) 15 (14.2)

6 (19.3) 22 (42.8)

0.011

0.015

19 (2.2) 4 (6.0)

7 (20.8) 12 (19.0)

0.007

29 (46.0) 1 (1.5)

14 (23.0) 3 (4.5)

13 (20.6) 21 (33.3)

21 (33.3) 8 (12.8)

0.710

0.002

0.853

0.007

Following migration, cells were fixed with 4 % formaldehyde and stained with 1 % crystal violet. Cells on the upper surface of the filters were removed by wiping with a cotton swab. Cells counts were the mean number of cells per fields of view. Three independent experiments were performed, and the data were presented as mean ± standard deviation (SD).

Scratch wound assay Cellular fractionation SW982 cells were plated in six-well plates and incubated overnight until 30–50 % confluent, then transfected with 100 nM HDGF siRNA. Vertical scratches were then made using a 100-μl plastic filter tip to create a Bwound^ of approximately 100 μm in diameter. To eliminate dislodged cells, culture medium was removed and wells were washed with PBS. BWound closure^ was observed at 0, 12, 24, and 48 h, and digital images were taken under an inverted microscope. Transwell migration assay Transwell migration assay was carried out in Transwell chambers containing polycarbonate filters (8-μm pore size; Corning Incorporated, Life Sciences, NY). After transfected with 100 nM HDGF siRNA for 48 h, 2 × 103 SW982 cells in a 500 μl volume of serum-free medium were placed in the upper chambers and incubated at 37 °C with 5 % CO2 for 24 h. While a 200 μl volume of medium containing 15 % FBS was added to the lower chamber as chemoattractant. Cells were allowed to migrate for 24 h at 37 °C with 5 % CO2.

Cultured cells were collected and resuspended in 500 μl 1× hypotonic buffer on ice. Twenty-five microliters of detergent was added and vortexed for 10 s at highest settings. Suspension was centrifuged for 30 s at 14,000g in a microcentrifuge tube and precooled at 4 °C. The supernatant (cytoplasmic fraction) was transferred into a prechilled microcentrifuge tube. Nuclear pellet was resuspended in 50 μl complete lysis buffer and centrifuged for 10 min at 14, 000g in a microcentrifuge tube and precooled at 4 °C. Supernatant (nuclear fraction) was transferred into a prechilled microcentrifuge tube. Immunofluorescence staining SW982 cells were transfected with HDGF-siRNA or βcatenin-siRNA for 24 h, then rinsed with PBS and fixed in 2 % paraformaldehyde for 10 min at room temperature, washed three times with PBS, then permeabilize in 0.2 % Triton X-100 in PBS for 5 min at room temperature. Cells

Tumor Biol.

were blocked in 10 % bovine serum albumin in PBS for 10 min at room temperature and then incubated with anti-β-catenin antibody (1:200; Cell Signaling Technology, Inc., Danvers, MA) and anti-HDGF (1:200, Abcam) in PBS for 1 h at room temperature. After three 5-min washes in PBS, secondary antibody-TRITC (Santa Cruz Biotechnology) was used at a dilution of 1:2000 in PBS for 30 min. Staining with DAPI (Keygene, China) was performed at a dilution of 1:2000 in PBS for 10 min to visualize nuclei. Slides were mounted using Prolong Gold antifade reagent (Invitrogen) and viewed under fluorescent microscopy on a Zeiss Axioskop. The results were from two independent experiments prepared in triplicate. Western blot analysis As our previously described [7], HDGF, β-catenin, cyclin D1, c-Myc, phosphorylated-glycogen synthase kinase 3β (phos-GSK-3β) (Ser9), and MMP9 (Cell Signaling Technology, Danvers, MA) antibodies were used to probe expression of protein at 4 °C for 12 h. Signal was detected by enhanced chemiluminescent techniques (Amersham Life Science, Piscataway, NJ). GAPDH (Sigma) and Lamin B1 (Cell Signaling Technology) were used as the loading control. Co-immunoprecipitation and immunoblotting analysis For co-immunoprecipitation analysis, cells lysates were incubated with 5 μg antibody on a rotator overnight at 4 °C. The protein-antibody-protein A/G agarose complexes were prepared by adding protein A/G agarose beads (Invitrogen) for an hour at 4 °C. After extensive washing with RadioImmunoprecipitation Assay (RIPA) lysis buffer, the immunoprecipitated complexes were resuspended in reducing sample buffer and boiled for 10 min. After centrifugation to pellet the agarose beads, supernatants were subjected to SDSPAGE and immunoblotting.

SMYD1: forward, TCA CCA TGT TGG TCA GGC TGG TCT, and reverse, AGG GTG GAC TGT TTA GCA GC; GAPDH: forward, 5′-GGA GTC CAC TGG CGT CTT-3′, and reverse, 5′-CTT GAG GCT GTT GTC ATA CTT C-3′; GAPDH was used as the loading control. Luciferase reporter assay SW982 cells (1 × 104) were seeded in triplicate in 96-well plates. Luciferase reporter plasmid (0.45 μg) containing β-catenin 3′UTR(−648 to −640 bp: AGACACAGT; −157 to −150 bp: AAGAAATT; −54 to −47 bp: CAAAGATG) or β-catenin 3′UTR mutant fragments (−648 to −640 bp: GTACAGCAA; −157 to −150 bp: ATATAGAA; −54 to −47: AGCAATAG), and 0.45 μg pRL-TK Renilla plasmid (Promega) w ere t ransfected into cells using the Lipofectamine 2000 reagent (Invitrogen). After 48-h transfection, Luciferase and Renilla signals were measured using the Dual-Luciferase Reporter Assay kit (Promega). Three independent experiments were done, and the data were presented as the mean ± SD. Immunohistochemistry and evaluation As our previously described [7], the working concentrations of primary antibody for the detection of HDGF (Sigma, St. Louis, MO) and β-catenin (Abcam) was 1:100 and 1:200, respectively. The degree of HDGF immunostaining was defined as the proportion score multiplied by the staining intensity score according to our previous method. The staining of β-catenin was scored as follows. When more than 70 % of synovial sarcoma cells were positively stained for membranous β-catenin, the cells were classified as β-catenin normal expression; if more than 10 % of sarcoma cells were positively stained for cytoplasm or nuclei, they were regarded as β-catenin abnormal expression. Statistical analyses

Chromatin immunoprecipitation assay Chromatin immunoprecipitation (ChIP) was done using the ChIP kit (Millipore) according to the manufacturer’s instruction. Briefly, 1 × 107 SW982 cells in a 10-cm culture dish were treated with 1 % formaldehyde to cross-link proteins to DNA. The cell lysates were sonicated to shear DNA to sizes of 100 to 1500 bp. Equal aliquots of chromatin supernatants, 5 μg anti-HDGF antibody (Abcam) or anti-IgG as negative control was added, were incubated overnight at 4 °C with rocking. After reverse cross-link of protein/DNA complexes to free DNA, real time-PCR was done using specific primers of β-catenin predicted by JASPAR database: forward, 5′-CAA TAG GCA TAT TTA CTA AAC AGG-3′, and reverse, 5′-AAC ATA ATA GCA ACA GCT GCA GCC-3′;

Chi-squared test was used to compare the levels of HDGF and β-catenin expression with different groups and various clinicopathological parameters. The Kaplan-Meier survival curves were used to estimate overall survival (OS). The significance of predictor variables for OS was evaluated by the long-rank test. Prognostic factors associated with OS were investigated according to the Cox proportional hazards regression model in a stepwise manner. Only those factors that were statistically significant (p < 0.05) in the univariate survival analysis were included in the multivariate analyses. Groups from cell culture experiments were compared using an unpaired, two-tailed Student’s tests, and results were presented as mean ± SD. For cell proliferation assay, comparison was done by univariate variance analysis (two-way ANOVA). Statistical analyses

Tumor Biol.

were performed using SPSS 16.0 statistical software. p < 0.05 was considered to be statistically significant.

Results HDGF and β-catenin expression in synovial sarcoma cell line and fresh tissue As shown in Fig. 1a, HDGF and β-catenin protein expressions were significantly higher in synovial sarcoma cell line SW982 compared with FLS cell line by Western blot analysis. In addition, HDGF and β-catenin protein expression were dramatically increased in one sample of synovial sarcoma fresh tissue compared with its adjacent nontumor tissue (Fig. 1b). HDGF and β-catenin expression and their relationship with clinicopathological features of synovial sarcoma

patients with high HDGF expression compared with 42.4 and 32.5 % for patients with low HDGF expression, respectively. In addition, synovial sarcoma patients had a significantly lower OS rate in β-catenin abnormal expression group than that in β-catenin normal expression group (p = 0.008, Fig. 1g).We stratified our cohort of synovial sarcoma patients into three groups according to the combination of different HDGF and β-catenin expression levels: group 1, low HDGF/normal βcatenin expression; group 2, high HDGF/normal β-catenin and low HDGF/abnormal β-catenin expression; group 3, high HDGF/abnormal β-catenin expression. For patients in group 3, the 3- and 5-year OS was significantly lower than that for patients in group 1 (p = 0.002) and group 2 (p = 0.011), respectively (Fig. 1h). Univariate Cox regression analysis showed that histological type (HR = 0.248; 95 % CI 0.083∼0.743; p = 0.013), HDGF expression (HR = 3.974; 95 % CI 1.257∼12.566; p = 0.019), and HDGF/β-catenin expression (HR = 2.866; 95 % CI 1.345∼6.107; p = 0.006) were independent prognostic factors for OS in synovial sarcoma patients (Table 2). Further multivariate Cox regression analysis demonstrated that histological type (HR = 3.841, 95 % CI 1.28∼11.532; p = 0.016) and HDGF/β-catenin expression (HR = 4.024, 95 % CI 0.739∼21.906; p=0.011) were the two important independent prognostic factors for OS in synovial sarcoma patients (Table 3).

In our series, HDGF positive signals were mostly located in nuclei of synovial sarcoma cells with minor cytoplasmic distribution by immunohistochemistry staining. Twenty-three synovial sarcoma samples (36.5 %, 23/63) were HDGF high expression, and 40 samples (63.5 %, 40/63) were HDGF low expression. However, high HDGF expression was found in 17 ANT samples (27 %). HDGF expression was significantly higher in synovial sarcoma tissues than that in ANT tissues (p < 0.01). HDGF expression was significantly related to TNM stage (p = 0.002), lymph node and/or distant metastasis (p = 0.028), and relapse (p = 0.015). No significant relationship between HDGF expression and gender, age, primary location, tumor size, and histological type was found. Thirtyfour synovial sarcoma samples (54.0 %, 34/63) were β-catenin abnormal expression, and 29 samples (46.0 %, 29/ 63) were β-catenin normal expression. However, β-catenin abnormal expression was found in 21 ANT tissues (33 %). β-Catenin abnormal expression was higher in synovial sarcoma tissues than that in ANT tissues (p < 0.01). β-Catenin abnormal expression was also significantly related to TNM stage (p = 0.007), lymph node and/or distant metastasis (p = 0.011), and relapse (p = 0.007). There was a positive correlation between HDGF high expression and β-catenin abnormal expression in synovial sarcoma tissues (r 2 = 0.502, p < 0.001) (Fig. 1c–e, Table 1).

HDGF mRNA level was significantly decreased in SW982 cells transfected with HDGF-siRNA at 48 h compared with the control group by real-time PCR analysis. Meanwhile, HDGF protein level in SW982 cells transfected with HDGFsiRNA was significantly decreased compared with the control group and the effect lasted up to 96 h (Fig. 2a, b). CCK8 assay showed that HDGF knockdown significantly suppressed SW982 proliferation compared with NC-siRNA control group (p < 0.0001) (Fig. 2c). Furthermore, the mean number of colony formation in SW982 cells transfected with HDGF-siRNA (mean number = 76) was significantly less than that in NCsiRNA transfection group (mean number = 277) (p < 0.0001). The cell colony formation rate was also significantly suppressed about 60 % in SW982 cells transfected with HDGFsiRNA compared with NC-siRNA group (p < 0.0001) (Fig. 2d, e).

Prognostic significance of HDGF and β-catenin expression in synovial sarcoma

HDGF knockdown resulted in cell cycle arrest and induced apoptosis in synovial sarcoma cells

Kaplan-Meier analysis showed that synovial sarcoma patients had a significantly lower OS rate in high HDGF expression group than that in low HDGF expression group (p = 0.012, Fig. 1f). The 3- and 5-year OS rate was 21.7 and 13.0 % for

As shown in Fig. 2f, g, HDGF knockdown significantly resulted in G1-S phase arrest in SW982 cells compared with NC-siRNA transfection group (p < 0.0001). Furthermore, the proportion of early and late apoptotic cells was 1.2 and 0.4 %

HDGF knockdown reduced synovial sarcoma cell growth

Tumor Biol.

Fig. 1 HDGF and β-catenin expression in synovial sarcoma cells and tissues. a HDGF and β-catenin protein expression in synovial cell line SW982 and human fibroblast-like synoviocyte (FLS); b HDGF and β-catenin protein expression in one pair of synovial sarcoma and adjacent nontumor tissue (ANT) tissue; c HDGF (low expression: a, c; high expression: b, d), β-catenin (normal expression: e, g; abnormal expression: f, h) expression in monophasic (a, b, e, f) and biphasic (c, d, g, h) synovial sarcoma tissues by immunohistochemistry staining, ×200. d, e Chi-squared test showed that high HDGF expression (e) and β-catenin abnormal expression (f) was significantly higher in

synovial sarcoma than that in ANT tissue (p < 0.01 and p < 0.01), respectively; f, g Overall survival (OS) of synovial sarcoma patients with different levels of HDGF (g, p = 0.012) and β-catenin (h, p = 0.008) expression by Kaplan-Meier analysis; h OS of synovial sarcoma patients according to the combination of HDGF and β-catenin expression levels by log-rank test (p = 0.008). Group 1: low HDGF/normal β-catenin expression; group 2: high HDGF/normal β-catenin expression and low HDGF/abnormal β-catenin expression; group 3: high HDGF/abnormal β-catenin expression

in SW982 with HDGF knockdown compared with 0.3 and 0.1 % in the control group, respectively. The percentage of

apoptotic cells in SW982 upon HDGF knockdown increased by four times compared with the control group (p < 0.0001).

Tumor Biol. Table 2 sarcoma

Univariate Cox regression analysis for the prognostic value of clinicopathological parameters, HDGF, and β-catenin expression in synovial

Characteristic

No.

Overall survival analysis 3-year overall survival (%)

5-year overall survival (%)

Hazard ratio

95 % CI

p value

Gender

Male

29

37.9

27.5

1.082

0.363∼3.230

0.887

Age (years)

Female ≥34

34 32

32.3 21.8

23.5 15.6

0.432

0.118∼1.574

0.203

<34

31

48.3

35.4

Primary location

Axial Peripheral

20 43

20.0 41.6

15.0 30.2

1.883

0.604∼5.870

0.275

Histological type

Monophasic

50

32.0

24.0

0.248

0.083∼0.743

0.013

Tumor size(cm)

Biphasic ≤4.8

13 30

46.1 46.6

30.7 26.6

0.863

0.214∼3.491

0.837

>4.8 Not known

15 18

26.6 22.2

26.6 22.2

Metastasis

No

36

38.8

27.7

0.567

0.190∼1.693

0.309

Relapse

Yes No

27 15

29.6 40

22.2 33.3

0.964

0.517∼0.626

0.996

TNM

Yes Not known I–II

44 4 33

36.3 Not available 42.4

25 Not available 30.3

0.323

0.099∼1.054

0.061

III–IV High Low Group 1 Group 2

30 23 40 20 17

26.6 21.7 42.4 55.5 41.1

20 13.0 32.5 45.0 29.4

3.974

1.257∼12.566

0.019

2.866

1.345∼6.107

0.006

Group 3

26

15.3

7.6

HDGF expression HDGF/β-catenin expression

Group 1: low HDGF expression + normal β-catenin expression; group 2: low HDGF expression + abnormal β-catenin expression/high HDGF expression + normal β-catenin expression; group 3: high HDGF expression + abnormal β-catenin expression

Table 3 Multivariate Cox regression analysis for the prognostic value of clinicopathological parameters, HDGF, and β-catenin expression in synovial sarcoma

HDGF knockdown suppressed migration in synovial sarcoma cell

Characteristic

Scratch wound assay showed that cell migration was dramatically suppressed in SW982 transfected with HDGF-siRNA compared with NC-siRNA transfection group at 12, 24, and 48 h, respectively. Furthermore, transwell migration assay showed that the mean number of migrated cells per field of view was significantly less in SW982 transfected with HDGF siRNA (mean number = 485) than that in NC-siRNA transfection group (mean number = 908) (p < 0.0001) (Fig. 3).

Histological type HDGF expression HDGF/β-catenin expression

Monophasic Biphasic High Low Group 1 Group 2 Group 3

Hazard ratio

95 % CI

p value

3.841

1.28∼11.532

0.016

1.933

0.150∼24.912

0.613

4.024

0.739∼21.906

0.011

Group 1: low HDGF expression + normal β-catenin expression; group 2: low HDGF expression + abnormal β-catenin expression/high HDGF expression + normal β-catenin expression; group 3: high HDGF expression + abnormal β-catenin expression

Exogenous HDGF enhanced synovial sarcoma cell growth SW982 cells were transfected with HDGF-siRNA for 48 h and serum-starved for synchronization. Quiescent

Tumor Biol. Fig. 2 HDGF knockdown inhibited synovial sarcoma cell proliferation and colony formation, but induced apoptosis. a HDGF mRNA expression decreased in SW982 cells transfected with HDGF-siRNA; b HDGF expression reduced in SW982 cells transfected with HDGF-siRNA at 48, 72, and 96 h by Western blot analysis. The relative quantification of bands in Western blots was a ratio neutralized to GAPDH; c HDGF knockdown dramatically suppressed SW982 cell proliferation by CCK8 assay; d, e HDGF knockdown significantly inhibited colony formation in SW982 cells. The histogram showed the relative mean colony formation number (d) and colony formation rate (e). f HDGF knockdown resulted in G1 phase arrest in SW982 cells; g HDGF knockdown significantly induced SW982 cell apoptosis

cells were then treated with four different concentrations (0, 300, 500, and 1000 ng/ml) of human recombinant HDGF (PROSPEC) for 48 h, and cell growth was determined by CCK8 assay. As shown in Fig. 4, recombinant HDGF attenuated the growth suppression of SW982 cell upon HDGF knockdown in a dosedependent manner. Recombinant HDGF (300 ng/ml) enhanced the cell growth by 20 % of the untreated control cells (p = 0.0076). Recombinant HDGF (500 ng/ml) further enhanced cell growth by 40 % in comparison with untreated control cells (p = 0.0022). Recombinant HDGF (1000 ng/ml) further enhanced cell growth by 60 % compared with the untreated control cells (p = 0.0002). Moreover, 500 ng/ml recombinant HDGF rescued the proliferation suppression of SW982 upon HDGF knockdown compared with the control group at 24, 48, 72, and 96 h (p = 0.0416, p = 0.022, p = 0.0237, p = 0.0147), respectively.

HDGF knockdown suppressed β-catenin signaling in synovial sarcoma cells HDGF knockdown dramatically suppressed β-catenin protein and its downstream gene including cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression in SW982 cells compared with NC-siRNA control group by Western blot analysis. Meanwhile, HDGF knockdown dramatically suppressed β-catenin, c-Myc, and cyclin D1 mRNA level in SW982 cells compared with the control group by real-time PCR analysis. Immunofluorescence staining showed that β-catenin expression was suppressed in SW982 cells transfected with HDGF siRNA compared with the control group. Further study showed that decreased nuclear and cytoplasmic β-catenin and its downstream gene including c-Myc, cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression were found in SW982 cells transfected with HDGF-siRNA compared with the control group. The

Tumor Biol.

Fig. 3 HDGF knockdown inhibited synovial sarcoma cell migration. HDGF knockdown inhibited SW982 cell migration by scratch wound assay (a) and transwell migration assay (b)

suppression of β-catenin protein expression in SW982 cells upon HDGF knockdown was attenuated by treatment of recombinant HDGF at 300 and 500 ng/ml compared with the untreated group, respectively. Further study showed that decreased nuclear and cytoplasmic β-catenin and its downstream gene including c-Myc, cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression level in SW982 cells upon HDGF knockdown was dramatically attenuated by treatment of recombinant HDGF at 300 and 500 ng/ml compared with the untreated group, respectively (Fig. 5). HDGF interacted with β-catenin in synovial sarcoma cells Intriguingly, β-catenin knockdown dramatically suppressed HDGF and β-catenin downstream gene including c-Myc, cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression in SW982 cells compared with NC-siRNA control group by Western blot analysis. Further study showed that decreased nuclear and cytoplasmic HDGF and β-catenin downstream gene including c-Myc, cyclin D1, MMP9, and

Fig. 4 Recombinant HDGF (rHDGF) significantly reversed cell proliferation suppression of SW982 upon HDGF knockdown in a dosedependent manner (a) and time-dependent manner (b) by CCK8 assay, respectively

upstream gene phos-GSK3β (Ser9) expression were found in SW982 cells transfected with β-catenin-siRNA compared with the control group. Immunofluorescence staining showed that HDGF expression was suppressed in SW982 cells transfected with β-catenin siRNA compared with the control group (Fig. 6a–c). It is well known that human recombinant Wnt-3a and DKK1 can agonist and inhibit Wnt/β-catenin pathway, respectively. To further verify the HDGF/β-catenin feedback loop in synovial sarcoma cells, nuclear and cytoplasmic HDGF, β-catenin, and its downstream gene including c-Myc, cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression levels increased in SW982 cells treated with recombinant Wnt3a (R&D SYSTEMS) at a concentration of 50, 100, and 150 ng/ml for 48 h compared with the control group by Western blot analysis, respectively. Likewise, nuclear and cytoplasmic HDGF, β-catenin and its downstream gene including c-Myc, cyclin D1, MMP9, and upstream gene phosGSK3β (Ser9) expression levels decreased in SW982 cells treated with recombinant DKK1 (R&D SYSTEMS) at a concentration of 50, 100, and 200 ng/ml for 48 h compared with the control group by Western blot analysis, respectively (Fig. 6d, e).

Tumor Biol. Fig. 5 HDGF knockdown inhibited β-catenin signaling pathway in synovial sarcoma. a HDGF knockdown dramatically inhibited β-catenin and its downstream genes including cyclin D1, MMP9, and phosGSK3β (Ser9) expression in SW982 cells. The relative quantification of bands in Western blots was a ratio neutralized to GAPDH; b β-catenin, c-Myc, and cyclin D1 mRNA level were inhibited in SW982 upon HDGF knockdown by real-time PCR analysis; c nuclear and cytoplasmic β-catenin, c-Myc, cyclin D1, MMP9, and phos-GSK-3β (Ser9) expression were suppressed in SW982 cells upon HDGF knock-down; d nuclear β-catenin expression decreased in SW982 cells upon HDGF knockdown by immunofluorescence staining, ×200; e rHDGF reversed the expression suppression of βcatenin in SW982 cells upon HDGF knockdown; f rHDGF reversed the nuclear and cytoplasmic protein expression suppression of β-catenin, c-Myc, cyclin D1, MMP9, and phosGSK-3β (Ser9) in SW982 upon HDGF knockdown by Western blot analysis. GAPDH was considered as loading control. Lamin B1 was considered as nuclear loading control

Further study showed that suppression of nuclear and cytoplasmic HDGF, β-catenin and its downstream gene including c-Myc, cyclin D1, MMP9, and upstream gene phosGSK3β (Ser9) expression in SW982 cells upon HDGF knockdown was attenuated by 50, 100, and 150 ng/ml recombinant Wnt3a treatment compared with the control group by Western blot analysis, respectively (Fig. 6f). Moreover, suppression of nuclear and cytoplasmic HDGF, β-catenin, and its down-stream gene including c-Myc, cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression in SW982 cells upon β-catenin knockdown was attenuated by 300 and 500 ng/ml recombinant HDGF treatment compared with the control group by Western blot analysis, respectively. Suppression of nuclear and cytoplasmic HDGF, β-catenin,

and its downstream gene including c-Myc, cyclin D1, MMP9, and up-stream gene phos-GSK3β (Ser9) expression in SW982 cells treated with 100 ng/ml DKK1 was attenuated by 300 and 500 ng/ml recombinant HDGF treatment compared with the control group by Western blot analysis, respectively (Fig. 7a, b). In addition, increased nuclear and cytoplasmic HDGF, β-catenin and its downstream gene including cMyc, cyclin D1, MMP9, and upstream gene phos-GSK3β (Ser9) expression in SW982 cells treated with 100 ng/ml Wnt3a was further enhanced by 300 and 500 ng/ml recombinant HDGF treatment compared with the control group by Western blot analysis, respectively (Fig. 7c). Further study showed that recombinant HDGF attenuated the cell growth suppression in SW982 cells upon β-catenin

Tumor Biol. Fig. 6 β-Catenin knockdown inhibited HDGF expression in synovial sarcoma. a β-Catenin knockdown inhibited HDGF expression in SW982 cells; b nuclear and cytoplasmic HDGF and β-catenin, c-Myc, cyclin D1, MMP9, and phos-GSK-3β (Ser9) protein expression decreased in SW982 cells upon β-catenin knockdown; c nuclear HDGF expression decreased in SW982 cells upon β-catenin knockdown by immunofluorescence staining, ×200. d, e Recombinant Wnt3a (d) and DKK1 (e) increased and decreased nuclear and cytoplasmic HDGF, β-catenin, and c-Myc, cyclin D1, MMP9, and phos-GSK-3β (Ser9) expression in SW982 cells, respectively; f recombinant Wnt3a reversed HDGF, βcatenin, and c-Myc, cyclin D1, MMP9, and phos-GSK-3β (Ser9) expression suppression in SW982 cells upon HDGF knockdown. The relative quantification of bands in Western blots was a ratio neutralized to GAPDH

knockdown in a dose-dependent manner by CCK8 assay. Recombinant HDGF (300 ng/ml) enhanced the cell growth by 15 % of the untreated control cells (p = 0.0063). Recombinant HDGF (500 ng/ml) further enhanced cell growth by 30 % in comparison with control cells (p < 0.0001). Recombinant HDGF (1000 ng/ml) further enhanced cell growth by 80 % of the untreated control cells (p < 0.0001). In addition, 500 ng/ml recombinant HDGF rescued the cell proliferation suppression in SW982 cells upon β-catenin knockdown at 24, 48, 72, and 96 h compared with the control group (p = 0.0473, p < 0.0001, p = 0.0002, p = 0.0225) by CCK8 assay, respectively (Fig. 7d, e).

The interaction mechanism of HDGF and β-catenin in synovial sarcoma cells To study whether HDGF protein directly interacts with β-catenin protein in synovial sarcoma cells, immunofluorescence staining showed that co-localization of HDGF protein and β-catenin protein was found in nuclei of SW982 cells (Fig. 8a). Direct interaction of HDGF protein and β-catenin protein was found in SW982 cells by co-immunoprecipitation assay (Fig. 8b). Furthermore, we found that the human β-catenin promoter region contained putative HDGFbinding elements in three regions by sequence alignment

Tumor Biol. Fig. 7 a, b rHDGF reversed HDGF, β-catenin, and c-Myc, cyclin D1, MMP9, and phosGSK-3β (Ser9) expression suppression in SW982 cells upon β-catenin knockdown (a) and DKK1 treatment (b). c Increased nuclear and cytoplasmic HDGF, β-catenin and its downstream gene including c-Myc, cyclin D1, MMP9 and up-stream gene phosGSK3β (Ser9) expression in SW982 cells with Wnt3a treatment was further enhanced by rHDGF treatment; d, e rHDGF reversed the cell proliferation suppression of SW982 cells upon β-catenin knockdown in dosedependent (d) and timedependent (e) manner, respectively

analysis using JASPAR database. Antibody-specific ChIP assay was used to determine whether these three regions (−648 to −640 bp, −157 to −150 bp, and −54 to −47 bp) of β-catenin promoter region could bind the endogenous HDGF protein in SW982 cells, three pairs of primers for real timePCR covering the region from −797 to −629 bp, −484 to −288 bp, and −147 to 32 bp of β-catenin promoter were used. We successfully amplified these three specific regions from immunoprecipitated chromatin DNA by ChIP and real-time PCR using HDGF antibody in SW982 cells compared with the control group, the promoter of SMYD1 gene was used as the positive control group because there was a DNA-binding

element for HDGF in the promoter of SMYD1 gene [13] (Fig. 8c). Further study showed that β-catenin promoter region (−484 to −288 bp) amplification was significantly suppressed in SW982 cells upon HDGF knockdown compared with the control group by ChIP assay (Fig. 8d). To further determine the specificity of HDGF-binding elements in β-catenin promoter region, we performed site-specific mutagenesis within the three putative HDGF-binding regions of β-catenin promoter. Dual-luciferase reporter assay showed that β-catenin transcriptional activity significantly decreased in SW982 cells transfected with the mutant β-catenin promoter luciferase reporter compared with the wild type of

Tumor Biol. Fig. 8 Interaction of HDGF and β-catenin in synovial sarcoma. a Co-localization of HDGF protein and β-catenin protein in nuclei of SW982 cells by immunofluorescence staining, ×400; b direct interaction of HDGF protein and β-catenin protein in SW982 cells by coimmunoprecipitation assay; c Three pairs of primer for βcatenin promoter from immunoprecipitated chromatin DNA in SW982 cells using HDGF antibody was amplified by ChIP and real-time PCR analysis. The promoter of SMYD1 gene was used as the positive control group; d β-catenin promoter region (−484 to −288 bp) amplification was significantly suppressed in SW982 cells upon HDGF knockdown compared with the control group by ChIP assay; e β-catenin transcriptional activity decreased in SW982 cells transfected with the mutant βcatenin promoter by dualluciferase reporter assay

β-catenin promoter (p < 0.0001, Fig. 8e). The result indicates that HDGF-binding elements in β-catenin promoter region are specific for transcriptional activation of β-catenin in SW982 cells.

Discussion The major prognostic factors for synovial sarcoma are tumor stage, tumor size, and tumor histological grade [14]. Hasegawa et al. reported only histologic grade, as defined by using categorized variables for the MIB-1 index and tumor necrosis, was an independent prognostic factor for synovial sarcoma [9]. Saito et al. reported low AJC stage (stages I and II: p < 0.0001), the preservation of alpha-catenin expression (p = 0.0001), and a low necrotic rate (<50 %: p = 0.0139) were independent favorable prognostic factors for synovial sarcoma [10]. Palmerini et al. have reported that nuclear expression of CXCR4 and IGF-1R is independent adverse prognostic factor for synovial sarcoma patient survival linked to the use of

chemotherapy [15]. High HDGF expression correlates with the disease progression and poor prognosis in various cancers [7, 16–21]. To investigate the relationship between HDGF expression and clinicopathological features of synovial sarcoma. Our data first showed that HDGF expression was significantly related to TNM stage, lymph node and/or distant metastasis, and relapse in synovial sarcoma. Multivariate regression analysis demonstrated that histological type and HDGF/β-catenin expression were the two important independent prognostic factors for OS in synovial sarcoma patients. To further clarify the biological significance of HDGF in synovial sarcoma, our data demonstrated that HDGF knockdown dramatically inhibited cellular proliferation, migration, colony formation, but induced G1 phase arrest and apoptosis in synovial sarcoma cells. The results suggest that HDGF might play an important role in tumorigenesis and progression of synovial sarcoma. Synovial sarcoma is characterized by chromosomal translocation t(X; 18), which leads to the fusion of the SS18-SSX. Pretto et al. have reported that SYT-SSX2 recruits β-catenin to

Tumor Biol.

the nucleus [22]. Trautmann et al. have reported SS18-SSXinduced Wnt/β-catenin signaling appears to be of crucial biological importance in synovial sarcoma tumorigenesis and progression. SS18-SSX fusion protein-induced Wnt/betacatenin signaling is a therapeutic target in synovial sarcoma [12]. Nuclear accumulation of β-catenin as a cell signaling event may play an important role in the progression of synovial sarcoma [9]. Abnormal levels of β-catenin could contribute to the development and progression of synovial sarcoma [10]. To study whether HDGF interacts with β-catenin in synovial sarcoma, our data first showed that direct interaction of HDGF protein and β-catenin protein was found in synovial sarcoma cell line-SW982 by co-immunoprecipitation assay. Further study showed that three HDGF-binding regions in β-catenin promoter were found and specific for transcriptional activation of β-catenin in SW982 cells. In addition, HDGF knockdown dramatically suppressed β-catenin and its downstream gene expression in SW982 cells. Interestingly, β-catenin knockdown also dramatically suppressed HDGF expression in SW982 cells. The results indicate that there might be a positive feedback of HDGF and β-catenin in synovial sarcoma. In addition to β-catenin signaling pathway, whether are there other signaling pathways of HDGF regulating in synovial sarcoma? Song et al. reported HDGF regulates glioma cell growth, apoptosis, and epithelial-mesenchymal transition probably through the Akt and the TGF-β signaling pathways [23]. Our microarray data also showed that 103 genes expression was significantly upregulated and 308 genes expression was significantly downregulated in SW982 cells transfected with HDGF-siRNA compared with NC-siRNA group. The genes involving in amino sugar and nucleotide sugar metabolism, base excision repair, Wnt/β-catenin signaling pathway, and MAPK signaling pathway were significantly altered in SW982 cells with HDGF knockdown (data not shown). Whether HDGF regulates these signaling pathways except Wnt/β-catenin signaling pathway in synovial sarcoma needs further study. In summary, our findings indicate that the HDGF binds HDGF-binding regions in β-catenin promoter and promotes its downstream gene transcriptional activation and β-catenin further increased HDGF expression, which might play a crucial role in tumorigenesis and progression of synovial sarcoma.

References 1.

2.

3.

4.

5.

6.

7.

8. 9.

10.

11.

12.

13.

14.

15. Acknowledgments This study was supported by the National Natural Science Foundation of China (Grant Nos. 81272636, 81472251, and 81201582).

16.

Compliance with ethical standards

17.

Conflicts of interest None

Sultan I, Rodriguez-Galindo C, Saab R, Yasir S, Casanova M, Ferrari A. Comparing children and adults with synovial sarcoma in the Surveillance, Epidemiology, and End Results program, 1983 to 2005: an analysis of 1268 patients. Am Cancer Soc. 2009;115: 3537–47. Nakamura H, Kambe H, Egawa T, Kimura Y, Ito H, Hayashi E, et al. Partial purification and characterization of human hepatomaderived growth factor. Clin Chim Acta. 1989;183:273–84. Mao J, Xu Z, Fang Y, Wang H, Xu J, Ye J, et al. Hepatoma-derived growth factor involved in the carcinogenesis of gastric epithelial cells through promotion of cell proliferation by Erk1/2 activation. Cancer Sci. 2008;99:2120–7. Yamamoto S, Tomita Y, Hoshida Y, Morii E, Yasuda T, Doki Y, et al. Expression level of hepatoma-derived growth factor correlates with tumor recurrence of esophageal carcinoma. Ann Surg Oncol. 2007;14:2141–9. Yoshida K, Tomita Y, Okuda Y, Yamamoto S, Enomoto H, Uyama H, et al. Hepatoma-derived growth factor is a novel prognostic factor for hepatocellular carcinoma. Ann Surg Oncol. 2006;13: 159–67. Iwasaki T, Nakagawa K, Nakamura H, Takada Y, Matsui K, Kawahara K. Hepatoma-derived growth factor as a prognostic marker in completely resected non-small-cell lung cancer. Oncol Rep. 2005;13:1075–80. Yang Y, Li H, Zhang F, Shi H, Zhen T, Dai S, et al. Clinical and biological significance of hepatoma-derived growth factor in Ewing’s sarcoma. J Pathol. 2013;231:323–34. Wong NA, Pignatelli M. Beta-catenin—a linchpin in colorectal carcinogenesis? Am J Pathol. 2002;160:389–401. Hasegawa T, Yokoyama R, Matsuno Y, Shimoda T, Hirohashi S. Prognostic significance of histologic grade and nuclear expression of beta-catenin in synovial sarcoma. Hum Pathol. 2001;32:257–63. Saito T, Oda Y, Sakamoto A, Tamiya S, Kinukawa N, Hayashi K, et al. Prognostic value of the preserved expression of the E-cadherin and catenin families of adhesion molecules and of beta-catenin mutations in synovial sarcoma. J Pathol. 2000;192:342–50. Barham W, Frump AL, Sherrill TP, Garcia CB, Saito-Diaz K, VanSaun MN, et al. Targeting the Wnt pathway in synovial sarcoma models. Cancer Discov. 2013;3:1286–301. Trautmann M, Sievers E, Aretz S, Kindler D, Michels S, Friedrichs N, et al. SS18-SSX fusion protein-induced Wnt/beta-catenin signaling is a therapeutic target in synovial sarcoma. Oncogene. 2014;33: 5006–16. Yang J, Everett AD. Hepatoma-derived growth factor binds DNA through the N-terminal PWWP domain. BMC Mol Biol. 2007;8: 101. Ten HS, Hoekstra HJ, Bastiaannet E, Suurmeijer AJ. The classic prognostic factors tumor stage, tumor size, and tumor grade are the strongest predictors of outcome in synovial sarcoma: no role for SSX fusion type or ezrin expression. Appl Immunohistochem Mol Morphol. 2009;17:189–95. Palmerini E, Benassi MS, Quattrini I, Pazzaglia L, Donati D, Benini S, et al. Prognostic and predictive role of CXCR4, IGF-1R and Ezrin expression in localized synovial sarcoma: is chemotaxis important to tumor response? Orphanet J Rare Dis. 2015;10:6. Tsai CC, Huang SC, Tai MH, Chien CC, Huang CC, Hsu YC. Hepatoma-derived growth factor upregulation is correlated with prognostic factors of early-stage cervical adenocarcinoma. Int J Mol Sci. 2014;15:21492–504. Tao F, Ye MF, Sun AJ, Lv JQ, Xu GG, Jing YM, et al. Prognostic significance of nuclear hepatoma-derived growth factor expression in gallbladder cancer. World J Gastroenterol. 2014;20:9564–9.

Tumor Biol. 18.

Yang Y, Zhen T, Zhang F, Dai S, Kang L, Liang Y, et al. P53 and hepatoma-derived growth factor expression and their clinicopathological association with Ewing family tumour. J Clin Pathol. 2014;67:235–42. 19. Hsu SS, Chen CH, Liu GS, Tai MH, Wang JS, Wu JC, et al. Tumorigenesis and prognostic role of hepatoma-derived growth factor in human gliomas. J Neuro-Oncol. 2012;107:101–9. 20. Chen X, Yun J, Fei F, Yi J, Tian R, Li S, et al. Prognostic value of nuclear hepatoma-derived growth factor (HDGF) localization in patients with breast cancer. Pathol Res Pract. 2012;208:437–43.

21.

22.

23.

Lin YW, Li CF, Chen HY, Yen CY, Lin LC, Huang CC, et al. The expression and prognostic significance of hepatoma-derived growth factor in oral cancer. Oral Oncol. 2012;48:629–35. Pretto D, Barco R, Rivera J, Neel N, Gustavson MD, Eid JE. The synovial sarcoma translocation protein SYT-SSX2 recruits betacatenin to the nucleus and associates with it in an active complex. Oncogene. 2006;25:3661–9. Song Y, Hu Z, Long H, Peng Y, Zhang X, Que T, et al. A complex mechanism for HDGF-mediated cell growth, migration, invasion, and TMZ chemosensitivity in glioma. J Neurooncol. 2014;119: 285–95.