The SW480 colon cancer cell line (adenocarcinoma), Iscove’s modified Dulbecco’s medium (IMDM) and 10% fetal bovine serum (FBS) used in the present study were obtained from the Cell Resource Center, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, Peking Union Medical College (CRC-PUMC; Beijing, China), and were authenticated as mycoplasma-free by short tandem-repeat polymerase chain reaction (PCR). Cells were cultured under the conditions recommended by the CRC-PUMC (http://cellresource.cn). The SW480 cell line was maintained in IMDM containing 10% FBS. All cell lines were grown under 5% CO₂ at 37°C in a humidified environment.

The siRNA method was used to knock-down USP22 expression in SW480 cells. USP22 siRNA was obtained from Guangzhou RiboBio Co., Ltd (Guangzhou, China), and the sequence was as follows: 5′-GGAGAAAGAUCACCUCGAADTDT-3′. STAT3 siRNA was obtained from Invitrogen Life Technologies (Carlsbad, CA, USA) and the sequence was: 5′-GGAAAGAAGGAAUAUCUCADTDT-3′. Ncontrol_05815 (NCsi; RiboBio Co., Ltd) was used as a negative control for all siRNA experiments. To select the most efficient siRNA, SW480 cells were transfected with the siRNAs using Lipofectamine^® 2000 (Invitrogen Life Technologies), according to the manufacturer’s instructions. Briefly, the SW480 cells were seeded without antibiotics at 1×10⁶ cells in a 25 cm² culture bottle. Prior to transfection, the cells were washed with phosphate-buffered saline (PBS). The cells were then divided into three groups: Those transfected with STAT3 siRNA, those transfected with both USP22 siRNA and STAT3 siRNA, and those transfected with negative control siRNA. The cell culture medium was changed after 6 h. The cells were incubated an additional 48 h, after which functional siRNA was measured by reverse transcription (RT)-PCR.

Total RNA was extracted using TRIzol reagent (Life Technologies, Grand Island, NY, USA). RT was performed using a PrimeScript RT Master Mix kit (Takara Biotechnology Co. Ltd., Dalian, China). The complementary DNA products were subjected to PCR using a SYBR Premix Ex TaqTM II kit (Takara Biotechnology Co. Ltd.). PCR was performed using a StepOnePlus Real-Time PCR system (Applied Biosystems Life Technologies, Carlsbad, CA, USA) using the following program: 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. Each sample was analyzed in triplicate. GAPDH mRNA was quantified and used as an endogenous control. The RNA expression was determined using the 2^−ΔΔCt method (13). The primer sequences used in the present study are shown in Table I.

MMP2 and MMP9 proteolytic activity of the supernatants was determined by gelatin zymography. At 24 h post-siRNA transfection, the transfection medium was removed and cells were incubated in serum-free medium for a further 48 h at 37°C. Serum-free supernatants were then subjected to 10% SDS-PAGE containing 0.1% gelatin (Beyotime Institute of Biotechnology, Haimen, China). Gels were incubated twice for 45 min at room temperature in 2.5% Triton X-100 (Qxbio, Beijing, China), and then incubated in activation buffer (50 mmol/l Tris-Cl pH 8.0, 10 mmol/l CaCl₂, 50 mmol/l NaCl) for 42 h at 37°C. Subsequently, gels were stained with 0.05% Coomassie Brilliant Blue R-250 (Beyotime Institute of Biotechnology) for 3 h and de stained sequentially in wash buffers I (10% acetic acid, 30% methanol), II (10% acetic acid, 20% methanol) and III (5% acetic acid, 10% methanol) for 0.5, 1 and 2 h at 37°C, respectively. Regions of gelatinolytic activity were detected as transparent bands on a blue background. The relative molecular weight of each band was determined using protein standards (Pierce Biotechnology, Inc., Rockford, IL, USA). Areas of protease activity appeared as unstained bands against a blue background. Images were assessed by ImageJ 1.44 (National Institutes of Health, Bethesda, MA, USA).

In vitro invasion was measured using an in vitro Transwell assay. The 24-well Transwell polycarbonate membrane inserts with 8.0-mm pores (Costar, Cambridge, MA, USA) were coated with Matrigel (1:8, 70 ul; 354277; BD Biosciences, San Jose, CA, USA) at the bottom. SW480 cells (1×10⁶) were plated in the upper chamber in 0.2 ml IMDM supplemented with 1% bovine serum albumin (BSA; Origene Technologies, Inc., Beijing, China), and the lower chamber was filled with 0.5 ml IMDM containing 10% BSA. Cells were incubated for 48 h, and cells remaining on the upper side of the membrane were gently removed with wet cotton swabs. The invading cells on the lower surface of the membrane were stained with crystal violet staining solution (C0121; Beyotime Institute of Biotechnology) for 30 min prior to washing. The number of cells in nine randomly selected microscopic fields per well (magnification, ×20) was counted under an Olympus IX51 microscope (Olympus Corp., Tokyo, Japan). All experiments were repeated in triplicate.

Cell viability was assessed by a CCK8 assay (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). SW480 cells (2×10³ cells/well) were plated in 96-well plates, and following 24 h of culture at 37°C, the cells were treated with STAT3 siRNA or USP22/STAT3 siRNA. Following incubation with clotrimazole for 24, 48 or 72 h at 37°C, the CCK8 reagent was added to each well and cells were incubated for 2 h at 37°C. Subsequently, the absorbance (optical density) at 450 nm was measured using a Synergy 4 reader (BioTek, Winooski, VT, USA). All experiments were repeated in triplicate.

TMAs were designed as described previously (14), using three 0.4 mm tissue cores per case, taken from formalin-fixed, paraffin-embedded archival biopsy blocks (Shanghai Biochip Co Ltd, Shanghai, China). TMAs were fixed in 4% paraformaldehyde in PBS for 30 min and preincubated in blocking buffer (PBS containing 10% normal rabbit serum) for 30 min. The TMAs were then incubated at 4°C overnight with the primary antibodies. Monoclonal anti human USP22 antibody (1:200; 3813-1) was purchased from Epitomics (Burlingame, CA, USA), monoclonal anti-STAT3 antibody (1:300; 4904S) was obtained from Cell Signaling Technology, Inc. (Danvers, MA, USA), and polyclonal anti-MMP9 anti-body (1:300; BS1241) was purchased from Bioworld (Beijing, China). After washing with PBS, the TMAs were incubated with horseradish peroxidase-labeled rabbit anti-goat immunoglobulin G secondary antibody (1:200; Origene Technologies, Inc.) at 37°C for 1 h. After washing with PBS, bound antibody was detected with the substrate reagents from HRP-DAB System of Tissue Staining kit (Origene Technologies, Inc.). USP22, STAT3 and MMP9-positive samples were defined as those exhibiting staining. Each case was evaluated by estimating the percentage and intensity of tumor cells displaying a nuclear or cytoplasmic staining pattern. The USP22 and STAT3 staining patterns were categorized into four groups: 0, no nuclear staining; 1+, ≤25% staining of tumor cells; 2+, 26–50% staining of tumor cells and 3+, 50–100% staining of tumor cells. MMP9 staining patterns were categorized into four groups: 0, no cytoplasmic staining; 1+, ≤25% staining of tumor cells; 2+, 26–50% staining of tumor cells and 3+, 50–100% staining of tumor cells. Categorizing of the USP22, STAT3 and MMP9 staining was conducted by two independent investigators in the Department of Pathology (Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Tsinghua University, Beijing, China), who were blinded to the clinicopathological data, with a consensus reached in all cases.

Data are expressed as the mean ± standard deviation. Statistical analyses to compare pairs of experimental groups were performed using the Student’s t-test or two-sample t-test with SPSS version 17.0 statistics software (SPSS, Inc., Chicago, IL, USA). Correlations between immunohistochemical expression and the clinicopathological characteristics were analyzed by Pearson’s χ² tests. P<0.05 was considered to indicate a statistically significant difference.

USP22 suppresses STAT3 acetylation by deubiquitinating SIRT1 (12). To detect whether USP22 influenced STAT3-mediated proliferation of SW480 cells, a CCK8 assay was performed. The proliferation capacity of USP22/STAT3si-SW480 cells was significantly decreased compared with that of STAT3si-SW480 cells. In addition, SW480 cell proliferation was inhibited following treatment with STAT3 siRNA compared with that of the control (Fig. 1).

Whether USP22 influenced STAT3-mediated SW480 cell invasion was evaluated by Transwell assay. The results indicated that the invasion potential of USP22/STAT3si-SW480 cells was increased significantly compared with that of STAT3si-SW480 cells, and SW480 cell invasion was inhibited following treatment with STAT3 siRNA compared with that of the control (Fig. 2).

To examine the mechanism by which USP22 modulates STAT3-mediated cell invasion, the effects of USP22 knockdown on transcripts of the STAT3 target genes Hif-1α, VEGF, MMP2, TIMP1, MMP9 and TWIST were investigated (15–19). MMP9 and TWIST mRNA levels were significantly increased in USP22si-SW480 cells compared with those of the control cells, as previously described (12); however the other target genes remained unaffected (Fig. 3A). A gelatin zymography assay was used to detect the activity of MMP2 and MMP9 following USP22/STAT3 siRNA and STAT3 siRNA treatment. As shown in Fig. 3B, the activity of MMP2 was decreased following treatment of SW480 cells with STAT3 siRNA or USP22/STAT3 siRNA. STAT3 knockdown reduced MMP9 activity, an effect which was partly rescued by USP22/STAT3 knockdown in SW480 cells.

To verify the association between USP22, STAT3 and MMP9 in colon cancer, immunohistochemical analysis was performed on a colon cancer TMA (n=39). The results revealed that USP22 and STAT3 staining was localized within the nuclei, and MMP9 staining was localized in the cytoplasm of colorectal epithelial cells (Fig. 4). USP22, STAT3 and MMP9 expression levels were higher in poorly differentiated colorectal adenocarcinoma, as compared with in the well-differentiated counterparts. As shown in Table II, USP22 exhibited moderate correlation with STAT3 expression (r=0.457, P<0.01) and low correlation with MMP9 expression (r=0.385, P<0.05). There was a strong correlation detected between STAT3 and MMP9 (r=0.785, P<0.01).

USP22 and MMP9 expression at the colon cancer invasion front was evaluated and the results indicated that USP22 expression was relatively low, while MMP9 expression was higher (Fig. 5).