Yin Yang 1 contributes to gastric carcinogenesis and its nuclear expression correlates with shorter survival in patients with early stage gastric adenocarcinoma

Ten gastric cancer cell lines (MKN28, KATOIII, MKN45, SNU16, SNU1, MKN7, MKN1, NCI-N87, AGS and MGC-803) were obtained from either the American Type Culture Collection (Rockville, MD) or RIKEN Cell Bank (Tsukuba, Japan), or received as a gift from Institute Digestive Disease of Prince Wales Hospital. These cell lines were grown in RPMI 1640 (GIBCO, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS) (GIBCO, Grand Island, NY), 100 U/ml penicillin and 10 μg/ml streptomycin in a humidified atmosphere of 5% CO₂ at 37°C.

The study was approved by Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee, Hong Kong (CREC Ref. No. 2009.521) and all participants provided written informed consent for the collection of samples and subsequent analysis. A total of 264 GAC samples were retrieved from the tissue bank of Anatomical and Cellular Pathology, Prince of Wales Hospital, Hong Kong from 1998 to 2006. The 264 GAC samples were embedded into tissue microarray blocks. Another 10 pairs of primary tumors and adjacent non-tumorous tissues were collected intra-operatively from patients with GAC who had not received radiotherapy or chemotherapy prior to surgery. These specimens were immediately snap frozen at −80°C for molecular analysis.

Immunohistochemistry was performed according to methods described previously [28]. Briefly, 4-μm-thick sections were obtained from formalin-fixed and paraffin-embedded specimens. After de-waxing in xylene and graded ethanol, sections were incubated in 3% H₂O₂ solution for 25 minutes to block endogenous peroxidase activities and then subjected to microwaving in EDTA buffer for antigen retrieval. Next, the tissue sections were incubated with the primary monoclonal YY1 antibodies (1:50, H-10, sc-7341, Santa Cruz Biotechnology, Dallas, TX) overnight at 4°C, and chromogen development was performed using the Envision system (DAKO Corporation, Glostrup, Denmark). The slides were independently scored by two investigators. The nuclear expression of YY1 was scored by estimating the proportion of tumor cells with positive nuclear staining into 4 different groups (0, none; 1+, <=10%; 2+, 10 to < =25%; 3+, >25%).

Total protein was extracted from gastric cancer cell lines and paired primary tumors and non-tumorous tissues using RIPA lysis buffer with proteinase inhibitor. Protein concentration was measured by the method of Bradford (Bio-Rad, Hercules, CA). Twenty-microgram of protein mixed with 2 × SDS loading buffer were loaded on each lane, separated by 12% SDS-polyacrylamide gel electrophoresis. YY1 protein was then detected using anti-YY1 antibody (1:1000, H-10, sc-7341, Santa Cruz Biotechnology, Dallas, TX). Other antibodies applied included cleaved-PARP (Asp214) (1:1000, #9541, Cell Signaling, Danvers, MA), active-β-catenin (1:1000, #05-665, Millipore, Billerica, MA), β-catenin (1:10000, #610154, BD Transduction Laboratories, San Jose, CA), CCND1 (1:1000, #2926, Cell Signaling, Danvers, MA ), c-Myc (1:1000, #9402, Cell Signaling, Danvers, MA), anti-Mouse IgG-HRP (1:30000, #00049039, Dako, Glostrup, Denmark) and anti-Rabbit IgG-HRP (1:40000, #00028856, Dako, Glostrup, Denmark).

For cell proliferation assays, transfection of YY1 siRNA (SI00051912, QIAGEN, Valencia, CA), siCTNNB1 (SI02662478, QIAGEN, Valencia, CA) and scramble controls was performed by Lipofectamine 2000 Transfection Reagent (Invitrogen, Grand Island, NY). For the transfection of pcDNA3.1+ empty vector control and YY1, FuGENE HD transfection reagent (Roche, Basel, Switzerland) was employed. Cell proliferation was assessed using CellTiter 96 Non-Radioactive Cell Proliferation Assay (Promega, Madison, WI) according to manufacturer’s instructions. For colony formation assays in monolayer cultures, cells transfected with YY1 siRNA or scramble control were seeded into 6-well plates and cultured for 8 days. For YY1 overexpression colony formation assays, the transfected cells were selected by G418 (100 ng/ml) for 2 weeks. Cells were fixed with 70% ethanol for 15 minutes and stained with 2% crystal violet. Colonies with more than 50 cells per colony were counted. The experiments were repeated in triplicate wells. For cell cycle analysis, AGS, MKN28 and NCI-N87 cells were collected 24 hours following transfection in 6-well plates. Before transfection with siYY1, the cells underwent serum-free starvation for 12 hours for synchronization. Cells were harvested using cold PBS and fixed in 70% cold ethanol overnight at 4°C and treated with 1 ng/ml RNase A for 10 minutes at 37°C. Cellular DNA was stained with 15 ng/ml propidium iodide (PI) for 30 minutes at 37°C in the dark. The cells then were sorted by FACS Calibur Flow Cytometer (Becton Dickinson, San Diego, CA) and cell-cycle profiles were determined using the ModFitLT software (Becton Dickinson, San Diego, CA). The experiments were repeated twice.

For YY1 knockdown in vivo study, MKN45 cells were transfected with empty vector (pBABE) and with shYY1. After puromycin selection, the cells (1 × 10⁶ cells suspended in 0.1 ml PBS) were injected subcutaneously into the dorsal flank of eight 4-week-old male Balb/c nude mice (shYY1 on the right side and the negative control cells on the left). Tumor diameter was measured and documented every 3 days until the tumor reached 10 mm in diameter. For YY1 overexpression in vivo study, MKN45 was transfected with empty vector (pcDNA3.1+) and with YY1. After G418 selection, the pool with stable YY1 overexpression and their control counterparts were injected subcutaneously into the dorsal flank of eight Balb/c nude mice (YY1 on the right side and the empty vector control cells on the left). Tumor diameter was measured and documented every 2 days until the tumors reached 10 mm in diameter. The mice were sacrificed and xenografts were removed for further validation. Tumor volume (mm³) was estimated by measuring the longest and shortest diameter of the tumor and calculated using the following formula: volume = (shortest diameter)² × (longest diameter) × 0.5. All animal handling and experimental procedures were approved by the Animal Ethics Committee of the Chinese University of Hong Kong.

The Mann–Whitney U test was used to compare the difference in biological behavior between siYY1 knockdown cells and scramble siRNA transfected cells. Correlations between YY1 nuclear stain and clinicopathologic parameters were assessed by the non-parametric Spearman’s rho rank test. The Kaplan-Meier method was used to estimate the survival rates for each variable. The equivalences of the survival curves were tested by log-rank statistics. For those variables being statistically significant found in the univariate survival analysis (p < 0.05), the Cox proportional hazards model with the likelihood ratio statistics was employed to further evaluate them for multivariate survival analysis. All statistical analyses were carried out using the statistical program SPSS version 16.0. A two-tailed p-value of < 0.05 was regarded as statistically significant.

Up-regulation of YY1 protein was detected in 9 gastric cancer cell lines by Western blot analysis (Figure 1A). In contrast, the 3 non-neoplastic gastric tissue samples showed no evidence of YY1 protein. In addition, up-regulated YY1 protein expression was observed in 9 out of 10 primary GACs but not seen in any of the adjacent non-tumorous gastric tissues (Figure 1B). YY1 mRNA also showed positive expression in 9 gastric cancer cell lines, but there was no obvious expression in 5 normal gastric tissues (Figure 1C). In 28 paired GAC cDNAs, 15 cases showed strong expression of YY1 in tumor tissue compared with adjacent non-tumorous tissue (Figure 1D).

YY1 protein was knocked down in AGS, MKN28 and NCI-N87 cells by siRNA mediated degradation (Figure 2A). A significantly decreased cellular proliferation was observed in these cell lines when compared with scramble siRNA groups (p < 0.001, Figure 2B). Monolayer colony formation assay indicated that YY1 knockdown significantly reduced colony formation in these cell lines (p < 0.001, Figure 2C). The transfectants were then analyzed for cell cycle parameters using flow cytometry. Twenty-four hours after transfection, accumulation of G1 cells increased from 52.6% to 55.3% in AGS, from 26.2% to 36.0% in MKN28, and from 46.6% to 49.4% in NCI-N87 (Figure 2D). The same trends were observed in a repeated set of experiments. Moreover, siYY1 induced late apoptosis, represented by an increase of cleaved-PARP, in three gastric cancer cell lines AGS, MKN28 and NCI-N87 (Figure 2E).

By using Cancer 10-pathway Reporter Luciferase Kit, we found that siYY1 suppressed Wnt/β-catenin signaling pathway in AGS, MKN28 and MGC-803 cell lines (Figure 3A). We subsequently used TOPflash luciferase assays to confirm that Wnt/β-catenin signaling pathway was indeed inhibited by siYY1 in AGS, MKN28, NCI-N87 and MGC-803 cells (Figure 3B). Suppression of Wnt/β-catenin signaling pathway by siYY1 as indicated by decreased active-β-catenin, CCND1 and c-Myc was observed in siYY1 transfected AGS, MKN28 and NCI-N87 cells (Figure 3C).

The effect of shYY1 expression on in vivo growth of tumor was also studied. We first selected MKN45 clones which stably expressed shYY1 and validated the downregulated YY1 in shYY1 stable clones by Western blot (Figure 3D1). Then shYY1-MKN45 and vector control clones were injected into nude mice subcutaneously. The tumor growth in shYY1 expressing clones was significantly decreased compared with the vector control clones 30 days after injection (p < 0.001, Figure 3D2).

To further investigate the YY1 functions in gastric cancer cells, we overexpressed YY1 into AGS, MKN28 and NCI-N87 cells. YY1 expression was first evaluated by Western blot after ectopic expression. Meanwhile active-β-catenin showed increased level (Figure 4A). YY1 overexpression increased cell proliferation in these three cell lines in 5-day MTT assays (p < 0.05, Figure 4B). To confirm the proliferation-promoting effect by YY1 through Wnt/β-catenin signaling pathway, siCTNNB1 (25 nM) was transfected into the cultured cells to knockdown Wnt pathway effector β-catenin. The β-catenin was effectively knocked down and the same grow suppressive effect was observed (Figure 4C), indicating that Wnt/β-catenin signaling pathway plays an important role in the YY1-induced tumor growth.

We further examined the effect of YY1 on monolayer colony formation. Gastric cancer cells over-expressing YY1 formed more and larger colonies (2.4, 2.3 and 1.9 folds in AGS, MKN28 and NCI-N87) in comparison to vector control groups, (p < 0.001, Figure 4D). The activated Wnt/β-catenin signaling pathway together with upregulation of its downstream effectors CCND1 and c-Myc were observed in YY1-MKN45 stable clones (Figure 4E1). The YY1-MKN45 stable clones were injected subcutaneously into the dorsal flank of 8 nude mice. Twenty days later, YY1 over-expressing clones formed larger xenografts than the control clones (p < 0.001, Figure 4E2).

YY1 was predominantly expressed in the nuclei of tumor cells (Figure 5A; left, case with negative YY1 expression; right, case with strong YY1 expression). Of the 247 GAC samples, 98 cases (39.7%) had a YY1 nuclear score 0, 3 cases (1.2%) had a score 1+, 5 cases (2%) had a score of 2+, and 141 cases (57.1%) had a score of 3+. In total, 101 of 247 samples (40.9%) of gastric cancer cases had negative/weak (0 to 1+) YY1 expression, while 146 of 247 samples (59.1%) had strong YY1 expression (2+ to 3+, Additional file 1: Figure S1). No survival difference was observed between YY1 negative/weak and strong expression by univariate analysis (p = 0.341).

We further investigated the clinicopathologic correlation of YY1 by stratifying the patients into early stage (stage I and II, n = 88) and advanced stage (stage III and IV, n = 159). Table 1 summarized the correlation of YY1 with clinicopathologic parameters in early stage and advanced stage patients. In early stage patients, YY1 expression correlated with diffuse type gastric cancer (p = 0.031). Univariate analysis (Table 2) indicated that YY1 strong expression in early stage GACs associated with poor disease specific survival (p = 0.045, Figure 5B). Negative H. pylori infection also correlated with poor survival (p = 0.003). By multivariate Cox proportional hazards regression analysis (Table 2), only H. pylori infection was independently associated with disease specific survival (p = 0.007).

In advanced stage GACs (Table 1), YY1 nuclear expression correlated with histology with diffuse component (p = 0.020), higher tumor grade (p = 0.033). However, YY1 expression could not predict survival (Figure 5C, p = 0.958) by univariate analysis. Old age (age > 60, p = 0.006), advanced N stage (p = 0.002) and advanced M stage (p < 0.001) also predicted poor prognosis. By multivariate Cox proportional hazards regression analysis, age, N stage and M stage remained the independent prognostic factors for patients with advanced gastric cancer.