Ethics committee approval was obtained from the Reproductive Center of Provincial hospital affiliated to Shandong University [approval no. (2019) Ethics Approval No. 32]. The experiments were conducted according to the regulations set out by the Declaration of Helsinki. All samples were collected from the Provincial Hospital affiliated to Shandong University (Jinan, China) from November, 2012 to June, 2016. The clinical specimens were used in the experiments as follows: Eight pairs of tumor endometrial tissues and non-tumor endometrial tissues (adjacent normal tissues and normal endometrial tissues) (matched for each patient) were used for western blot analysis; 25 normal endometrial tissues and 74 tumor endometrial tissues were used for immunohistochemistry (IHC). The patient characteristics presented in Table I. The specimens were collected after obtaining informed consent from the patients. All EnC specimens were diagnosed and assessed in accordance with the International Federation of Gynecology Oncology (FIGO) criteria (2009). All the EnC samples must have been diagnosed as endometrial adenocarcinoma and underwent initial surgery. All the normal endometrial tissues were collected from patients diagnosed as having uterine fibroids following hysterectomy. None of the patients had been administered any pre-operative medical treatments.

As described in a previous study (21), KLE (GCC-UT0008RT/GCC-UT0008CS, http://www.taogene.com/emkt.htm#/PcMerchandises?id=f9c8f996-7ca3-473c-bd2e-8d7525340d03&categoryId=6), Ishikawa (GCC-UT0004RT/ GCC-UT0004CS, http://www.taogene.com/emkt.htm#/PcMerchandises?id=89692be1-e99d-4a6f-a22d-11441b94b506 &categoryId=6) and RL-952 (GCC-UT0006RT/GCC-UT000 6CS, http://www.taogene.com/emkt.htm#/PcMerchandises?id=4d36c08c-34cd-4ce4-93a3-1cbdf6c3ce3d&categoryId=6) EnC cell lines were purchased from GeneChem Co., Ltd. Twenty short tandem repeat loci plus the gender determining locus, Amelogenin, were amplified using the PowerPlex^® 21 System from Promega Corp., which tested and authenticated these three EnC cell lines. All cells were cultured in HyClone™ Dulbecco's modified Eagle's medium:nutrient mixture F-12 (DMEM/F-12) (HyClone; GE Healthcare Life Sciences) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Biological Industries, Israel), 1% penicillin and streptomycin (P/S) (HyClone; GE Healthcare Life Sciences) at 37°C in a humidified 5% CO₂ atmosphere.

Paraffin-embedded sections (4-µm-thick) were dewaxed, hydrated, boiled in citric acid buffer for antigen retrieval, steeped in 0.3% hydrogen peroxide to interdict the activity of endogenous peroxidase, blocked with 3% bovine serum albumin and incubated with primary antibodies at 4°C overnight. The following day, the sections were incubated with secondary antibodies (rabbit, PV-9001, ZSGB-Bio) at room temperature for 20 min according to the protocol of the PV-9001 Immunohistochemical kit (ZSGB-BIO). Primary antibodies were used at the following dilutions: TRIM22 antibody (1:350, NBP1-81795, Novus Biologicals, LLC), nucleotide binding oligomerization domain containing 2 (NOD2) antibody (1:250, NB100-524, Novus Biologicals, LLC), Ki-67 antibody (1:200, RB-9043-P1, eBioscience; Thermo Fisher Scientific, Inc.). These sections were then stained with hematoxylin in room temperature for 3 min, and were then cleared, dehydrated, hyalinized and mounted. Tissues staining brown in the cytoplasm or nucleus were regarded as positive. Five fields in each section were selected for further analysis using a fluorescent microscope (Olympus Corp.). The quantification of protein expression was presented with integrated optical density (IOD), using Image-Pro Plus 6.0 software. The final mean optical density was determined according to the following equation (equation 1): MOD=(IOD SUM)/(area SUM), where MOD represents the mean optical density, IOD SUM is the sum IOD of all selected fields in one image and the area SUM refers to the sum area of all selected fields.

Paraffin-embedded sections (4-µm-thick) were dewaxed, hydrated, boiled in citric acid buffer for antigen retrieval, blocked with 3% bovine serum albumin (Servicebio) and incubated with primary antibodies TRIM22 (1:200, NBP1-81795, Novus Biologicals, LLC) and NOD2 (1:100, NB100-524, Novus Biologicals, LLC) at 4°C overnight. The following day, the sections were incubated with corresponding secondary fluorescent-conjugated antibodies (1:300, GB21303, rabbit, Servicebio) in the dark and at room temperature for 45 min. The sections were counterstained with 4',6'-diamidino-2-phenylindole (DAPI) (Servicebio, China) in the dark and at room temperature for 10 min, and images were acquired using a confocal laser scanning microscope (Nikon Copr.). The staining for TRIM22 was red, that for NOD2 was green, and the nuclei were stained blue.

Ubi-TRIM22-3FLAG-SV40-EGFP- IRES-puromycin lentiviral vector (GV358) was constructed by GeneChem Co., Ltd. The primer sequences were as follows: Forward, 5'-GAGGATCCCCGGGTACCGGTCG CCACCATGGATTTCTCAGTAAAGGTAGACATAG-3' and reverse, 5'-TCCTTGTAGTCCATACCGGAGCTCGGTGGG CACACAGTCATG-3'. The TRIM22 gene was used from NCBI (NM_006074). According to the manufacturer's instructions, the lentiviral vector was transfected into the Ishikawa and KLE cells using Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) (these cells were termed TRIM22 OE). The Ubi-3FLAG-SV40-EGFP-IRES-puromycin lentiviral vector was used as a control (vector control). The cellular infection rate and GFP-positive cell number were detected by fluorescence microscopy at 72 h following infection. Stably transfected clones of TRIM22 were detected by western blot analysis.

Cells were harvested at predetermined times when the cells spread out in the dish (approximately 80%) and rinsed twice with PBS. The tumor tissues were harvested from the mice. Cell sediments and the tumor tissues were treated with SDS lysis buffer (Beyotime Institute of Biotechnology), and centrifuged at 14,000 × g for 10 min at 4°C. Additionally, extracting the cytoplasmic and nuclear protein was performed according to the instructions of NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, Inc.). Protein samples (20-40µg) were electrophoresed through 10% polyacrylamide gels and transferred onto 0.45 µm PVDF membranes (Merck Millipore). The membranes were blocked with 5% skim milk for 1 h at room temperature, cleared thrice with 1X Tris-buffered saline and Tween-20 (TBST) (15 min/time), incubated with primary antibodies overnight at 4°C, washed with 1X TBST thrice and incubated with secondary antibodies for 1 h at room temperature, and washed thrice with 1X TBST, developed with Immobilon Western HRP (ECL; Merck Millipore). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH; ProteinTech Group, Inc.) and Lamin B1 (Abcam) were used as reference controls. Secondary antibodies (ZSGB-BIO) were accompanied with IR dyes. Blots were detected with the ChemiDoc MP Imaging System (Bio-Rad Laboratories, Inc.). Primary antibodies were used at the following dilutions: TRIM22 rabbit polyclonal antibody (1:200, NBP1-81795, Novus Biologicals, LLC), NOD2 antibody (1:1,000, NB100-524, Novus Biologicals, LLC), IκBα(E130) antibody (1:1,000, ab32518, Abcam), p-IκBα (Ser36) [EPR6235 (2)] antibody (1:10,000, ab133462, Abcam), NF-κb p65 (E379) antibody (1:50,000, ab32536, Abcam), p-p65(Ser536) (EP2294Y) antibody (1:10,000, ab76302, Abcam), GAPDH (AG0766) antibody (1:10,000, 60004-1-Ig, Proteintech) and LaminB1 antibody (1:5,000, ab16048, Abcam) and secondary antibodies (rabbit, mouse, ZB-2301, ZB-2305, ZSGB-Bio) were used at 1:5,000.

Cells (1×10⁴ cells/well) were seeded in a 96-well plate in triplicate and incubated at room temperature for 24 h. Subsequently, the cells were incubated with medium containing 50 µM 5-ethynyl-2'-deoxyuridine (EdU) for an additional 2 h. The absolute ethyl alcohol (95% ethyl alcohol) was used to immobilize the cells at 4°C. Cell proliferation was assessed using a Cell-Light™ EdU Cell Proliferation Detection kit according to the manufacturer's instructions. DNA was stained with Hoechst 33342 at room temperature for 30 min and observed using an inverted fluorescence microscope (Olympus Corp.). For each EdU test, 5 fields were randomly selected to image at x50 magnification. The absolute ethyl alcohol can accelerate the GFP to disappear out of the cells by enhancing the membrane permeability and, as a result, GFP has minimal influence on the EdU. The number of EdU-positive cells was determined according to Hoechst nuclear staining and was reported as a percentage of the total number of cells in each field.

Cell migration assay was performed using a Transwell chamber (pore size, 0.8 µm; Merck Millipore), while the invasion assay was performed using a Transwell chamber coated with Matrigel (BD Biosciences). Endometrial cells (6×10⁴ cells for migration assay and 4×10⁴ cells for invasion assay) in 100 µl serum-free DMEM-F12 were placed in the Transwell chamber and 700 µl DMEM-F12 containing 20% FBS was added to the lower chamber. Following incubation for approximately 24 h at 37°C in a humidified 5% CO₂ atmosphere, the migrating or invading cells were fixed with absolute ethyl alcohol and stained with hematoxylin in room temperature for 15 min. Cells on the upper surface of the filter were removed by wiping with a small cotton swab. Five fields of the fixed cells were imaged using a fluorescent microscope (Olympus Corp.), and cells were counted.

Cell cycle assay was implemented using the CycletestTM Plus DNA kit (BD Biosciences). According to the instructions of the kit, cells (1.5×10⁵) were washed with the buffered solution and resuspended in A and B solutions to damage the cell membrane structure, after which they were incubated with solution C [propidium iodide (PI)] for 10 min in the dark on ice (4°C). Solution A contained trypsin in a spermine tetrahydrochloride detergent buffer for the enzymatic disaggregation of the solid tissue fragments and digestion of cell membranes and cytoskeletons; solution B contained trypsin inhibitor and ribonuclease A in citrate-stabilizing buffer with spermine tetrahydrochloride to inhibit the trypsin activity and to digest the RNA. Cells were counted with a flow cytometer (Bio-Rad Laboratories, Inc.) and expressed as the percentage of G1, S and G2 phase cells using ModFit LTV4.1.7 software.

Cell lysates (approximately 1.5 mg total protein) were collected from the TRIM22-overexpressing (TRIM22 OE) Ishikawa cells using ice-cold non-denaturation lysis buffer. co-IP was performed according to the manufacturer's protocol of the Thermo Scientific Pierce Co-IP kit (Thermo Fisher Scientific, Inc.). TRIM22 and NOD2 antibodies were fixed for 2 h with AminoLink Plus coupling resin at room temperature. The resin was then washed and incubated with the cell lysates overnight at 4°C. The following day, the resin was again washed, and elution buffer was used to elute the protein combined with the resin. A negative control, harvested from the vector control-transfected Ishikawa cells, were treated in the same manner as the Co-IP samples, including incubation with AminoLink Plus coupling resin combined with TRIM22 and NOD2 antibodies (Novus Biologicals, LLC) overnight at 4°C. This control allowed us to observe whether the binding protein was increased in the TRIM22 OE groups. Samples were analyzed by western blot analysis using rabbit polyclonal anti-NOD2 (Novus Biologicals, LLC) and mouse monoclonal anti-TRIM22 (Novus Biologicals, LLC) antibodies.

Plasmids encoding human TRIM22 shRNA, NF-κB-p65 (RelA) shRNA and scramble shRNA were purchased from Shanghai Genechem Co., Ltd. A human NF-κB-p65 shRNA plasmid, resistant to puromycin, was selected to knockdown NF-κB-p65 expression (shR p65); the TRIM22 shRNA plasmid was then used to knockdown the expression of TRIM22 (shR TRIM22). TRIM22 OE Ishikawa cells (2×10⁵) and cells expressing relatively high levels of TRIM22 (TRIM22 RL-952 cells; 2×10⁵) were seeded into 6-well plates and cultured at 37°C in a humidified 5% CO₂ atmosphere overnight. The following day, the cells were incubated with the transfection medium containing the shRNA plasmid and Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) for 5-7 h, after which the transfection medium was replaced with normal growth medium for incubation of the cells for a further 48 h at 37°C in a humidified 5% CO₂ atmosphere. The NF-κB-p65 (RelA) shRNA target sequence was: 5'-CTCCATTGCGGACATGGACTT-3'; shRNA non-target sequence: 5'-TTCTCCGAACGTGTCACGT-3'; TRIM22 shRNA target sequence: 5'-CCAGATATAGACCTCAATA-3'; and TRIM22 shRNA non-target sequence: 5'-TTC TCCGAACGTGTCACGT-3'.

Female nude mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd., ajoint venture enterprise of Charles River Laboratories. The use of animals was approved by the Ethics Committee of Reproductive Hospital Affiliated to Shandong University. Their care was in accordance with institutional guidelines. A total of 18 BALB/c female nude mice which were 4-5-weeks old were kept under specific pathogen-free (SPF) conditions in the Shandong University Experimental Animal Room. Ishikawa cells infected with lentivirus (vector control and TRIM22 OE), were harvested and resuspended in 100 µl Matrigel/PBS (1:1 vol/vol; Corning, Inc.) and subcutaneously injected into the left flank of each mouse (n=9/group) with 1×10⁷ cells/inoculum(day 0). Tumors of nude mice were observed every 3 days. Tumor volume was expressed as mm³ and measured using a common ruler with the traditional formula as follows: (Equation 2): V=½ x length (mm) x width² (mm).

After 27 days (day 27) (tumor volume ≤1,000 mm3), the nude mice were anesthetized with a 3-5% (v/v) mixture of isoflurane (Aerrane; isoflurane, Baxter) in synthetic air (200 ml/min), and were then sacrificed by cervical dislocation and the tumors were harvested. The tumor tissues were collected for use in western blot analysis and immobilized in 4% paraformaldehyde for IHC analysis of TRIM22 and Ki-67 expression, and for hematoxylin and eosin (H&E)-staining.

Data under normal distribution are presented as the means ± SEM. The Student's unpaired t-test (2-tailed) was used for comparisons between 2 groups. Comparisons between the tumor and the adjacent tumor samples were analyzed using a paired t-test (2-tailed), and one-way ANOVA followed by Tukey's post-hoc test was applied for comparisons among ≥3 groups. Kaplan-Meier estimation was performed to investigate the survival probability of the patients with EnC. GraphPad Prism software version 6.0 was used to visualize the results of the comparisons, as well as the P-values. A P-value <0.05 was considered to indicate a statistically significant difference.

IHC staining was usedd to detect TRIM22 expression in endometrial tissues. IHC analysis revealed cytoplasmic and nuclear TRIM22 staining in both the normal and tumorous endometrial tissues (Fig. 1A and B). However, TRIM22 expression was lower in the EnC specimens compared to the normal endometrial tissue specimens (Fig. 1A-C). Moreover, TRIM22 expression was higher in the secretory phase compared to the proliferative phase in normal endometrial tissues (Fig. 1A and C). An association was observed between the expression of TRIM22 and the clinical stage. Specifically, TRIM22 expression decreased with the increasing tumor stage (Fig. 1B and D, and Table I). Thereafter, western blot analysis was used to detect TRIM22 expression in 8 paired EnC samples. The results demonstrated that TRIM22 expression was decreased in the EnC tissues compared with the normal endometrial tissues and adjacent normal tissues (Fig. 1E).

To further examine the association between the TRIM22 expression level and EnC malignant conversion, Kaplan-Meier analysis was applied to analyze the results of IHC MOD and survival. These results indicated that an increased TRIM22 expression was significantly related to an improved overall survival of patients with EnC (Fig. 1F). These data support the association between the downregulated expression of TRIM22 and the malignant transformation of EnC, while suggesting that TRIM22 may be a promising prognostic predictor and therapeutic target.

The function of TRIM22 was analyzed in the KLE, Ishikawa and RL-952 EnC cell lines. Stably transfected TRIM22 OE Ishikawa and KLE cells were generated which expressed higher levels of TRIM22; in addition, shR TRIM22 RL-952 cells were generated which expressed lower levels of TRIM22. Western blot analysis confirmed that the expression of TRIM22 increased in the TRIM22 OE Ishikawa and KLE cells, and decreased in the shR TRIM22 RL-952 cells (Fig. S1A and B). Fluorescence microscopy was used to observe the transfection efficiency (Fig. S1C and D).

Furthermore, the overexpression of TRIM22 decreased Ishikawa and KLE cell proliferation, while the opposite effect was observed in the shRTRIM22 RL-952 cells (Fig. 2A and C). Moreover, the migratory and invasive abilities of the TRIM22 OE Ishikawa and KLE cells were markedly restricted; by contrast, shR TRIM22 RL-952 enhanced the cell migratory and invasive abilities (Fig. 2B and D). Simultaneously, cell cycle analysis revealed that the number of TRIM22 OE Ishikawa and KLE cells in the G1 phase increased, while that in the S phase decreased (Fig. 2E). These results indicate that TRIM22 suppresses EnC progression.

As an important factor involving multiple aspects of immunity and inflammation, TRIM22 has been reported to function as a NOD2-interacting protein, and as a regulator of the NF-κB signaling pathway, dependent on NOD2 in inflammatory bowel disease (22). To determine whether the TRIM22-NOD2 interaction exists in normal endometrial tissues and endometrial tumor tissues, immunofluorescence and co-IP were conducted. The results of immunofluorescence assay revealed that TRIM22 and NOD2 co-localized in both the normal endometrial tissues and endometrial tumor tissues (Fig. 3A). Furthermore, the expression of NOD2 was decreased in the tumor tissues, which was consistent with TRIM22 expression (Fig. 3A). It was also found that the association between TRIM22 and endogenous NOD2 was weak in the co-IP experiments in the vector control-transfected Ishikawa cells and human endometrial tumor cell lines, and this association was enhanced in the TRIM22 OE Ishikawa cells (Fig. 3B). NOD2 binding to TRIM22 was further investigated by co-IP. It was confirmed that NOD2 was also weakly bound to the vector control-transfected Ishikawa cells, and this association increased in the TRIM22 OE Ishikawa cells (Fig. 3B). These experiments demonstrated that TRIM22 directly interacts with NOD2 protein.

Since TRIM22 co-localized and co-immunoprecipitated with NOD2, the question of whether TRIM22 affects the NOD2 signaling pathway was investigated. An alternate study reported that the role of NOD2 in regulating NF-κB, was bidirectional. Moreover, NOD2 can activate NF-κB in a number of inflammatory diseases (23-27) and can inhibit NF-κB in colorectal tumorigenesis (28-30). The expression of NOD2 and NF-κB-associated protein was detected in the vector control-transfected and TRIM22 OE Ishikawa cells, respectively, by western blot analysis. It was also found that as the expression of NOD2 increased, that of NF-κB-p65 and IκBα also increased; however, the phosphorylation of NF-κB-p65 (p-p65) and IκBα (p-IκBα) was decreased in the TRIM22 OE Ishikawa cells (Fig. 3C). This suggested that TRIM22 induced NOD2 expression and increased the expression of NOD2, subsequently decreasing NF-κB activation in Ishikawa cells. Additionally, these results suggest that NOD2 may inhibit NF-κB in EnC.

Subsequently, the regulatory role of TRIM22 in the NOD2 signaling pathway was further determined. To identify the suppressive role of TRIM22 in the translocation of NF-κB-p65, nuclear and cytoplasmic fractions were prepared from the vector control-transfected and TRIM22 OE Ishikawa cells. Western blot analysis revealed that in the cytoplasm, the expression of NF-κB-p65 in the vector control-transfected cells was lower than that in the TRIM22 OE cells; however, in the nucleus, the expression of NF-κB-p65 in the vector control-transfected cells was higher than that in the TRIM22 OE cells (Fig. 3D). These results suggest that TRIM22 inhibits NF-κb-p65 from translocating from the cytoplasm to the nucleus, and thus, it inhibits its transcriptional regulatory function. Hence, TRIM22 inhibits the activity of NF-κB-p65. NF-κB-p65 was also knocked down by transiently transfecting the shRNA-NF-κB-p65 plasmids into TRIM22 OE Ishikawa cells (shR p65). The effect of NF-κB-p65 knockdown was examined by western blot analysis (Fig. S2). The results of EdU and Transwell assays revealed that cell proliferation, migration and invasion of the shR p65 TRIM22 OE Ishikawa cells increased (Fig. 3E and F). It was thus suggested that the knockdown of NF-κB-p65 may overcome the reduced cell proliferation, migration and invasion induced by TRIM22 overexpression. These data indicate that TRIM22 suppresses the NF-κB signaling pathway by associating with NOD2.

Subsequently, the in vivo effect of TRIM22 was assessed in a BALB/c Ishikawa cell line xenograft model in nude mice. Nine mice developed palpable tumors at the injection site following the subcutaneous injection of Ishikawa cells for 7 days. The tumor size was measured twice a week for 4 weeks. TRIM22 overexpression exerted a significant inhibitory effect on tumor growth by day 25 of the study (Fig. 4). The mean tumor volume of the mice in the TRIM22 OE group from day 19 was markedly smaller than that of the vector control group (Fig. 4B).

IHC staining for Ki-67, a cell growth marker, was used to visualize xenograft tumors size as determined by the tumor cell proliferation in vivo. The expression of Ki-67 was significantly decreased in the malignant tumors, which were derived from the TRIM22 OE Ishikawa cells. The tumor tissues were then stained with H&E for the observation of the pathological characteristics in vivo (Fig. 4C).

To validate the regulatory effects of TRIM22 in vivo, the relevant protein expression in the tumor tissues from the nude mice were examined. The results demonstrated that the increased expression of TRIM22 also upregulated IκBα, NF-κB-p65 expression in vivo (Fig. 4D). Taken together, these results indicate that TRIM22 inhibits EnC development and progression by regulating the NF-κB signaling pathway.