Volume 147, Issue 5, Pages 1043-1054 (November 2014) TRIM59 Is Up-regulated in Gastric Tumors, Promoting Ubiquitination and Degradation of p53  Zhicheng Zhou, Zhongzhong Ji, You Wang, Jian Li, Hui Cao, Helen He Zhu, Wei-Qiang Gao  Gastroenterology  Volume 147, Issue 5, Pages 1043-1054 (November 2014) DOI: 10.1053/j.gastro.2014.07.021 Copyright © 2014 AGA Institute Terms and Conditions

Figure 1 TRIM59 expression is up-regulated in gastric cancer. (A) Analysis from the Oncomine database shows that mRNA expression levels of TRIM59 are significantly higher in gastric cancers compared with normal tissues. Data were pooled together from 2 published gastric cancer gene expression studies31,32 (P < .001), 46 normal tissues and 50 cancer tissues were analyzed. (B) Analysis of tumor TRIM59 DNA copy number in association with patient survival from the Oncomine database indicates a correlation of high TRIM59 copy number with shortened survival (the Cancer Genome Atlas, Gastric Statistics, P = .0447, n = 14, 28). (C) Immunoblot shows higher protein levels of TRIM59 in 7 of 10 tumor samples compared with the respective matched normal tissues (T, tumor; N, normal tissue). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is used as a loading control. (D) TRIM59 mRNA expression level in 10 paired tumor samples and normal tissues. (E) The mRNA level of TRIM59 in GES-1 and GES-7 gastric cancer cell lines with different differentiation status. Data were normalized against the TRIM59 expression level in GES-1 cells. (F) Immunoblot of TRIM59 confirms high expression level of TRIM59 in poor differentiated gastric cancer cell lines. Data are presented as means ± SEM. *P < .05, **P < .01, ***P < .001. (A and D) A t test was used for the statistical analysis. (B) The log-rank (Mantel–Cox) test was used. (E) One-way analysis of variance was used. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions

Figure 2 Increased expression of TRIM59 correlates with gastric cancer progression and poor survival in patients. (A) Immunohistochemical staining of normal and gastric cancer tissues with anti-TRIM59 antibody. Representative patient samples of clinical stages I, II, and III are shown. A total of 111 patient samples were stained and analyzed. (B) Quantitative analysis of TRIM59 staining shows significantly higher staining intensity in gastric tumor samples compared with normal tissues (72 normal tissues and 108 tumor samples). IOD, integral optical density. (C) Analysis of TRIM59 staining intensity in association with clinical stages of gastric tumor samples (n = 108). (D) Up-regulated TRIM59 expression positively correlated with high degrees of gastric tumor infiltration. When the intensity of TRIM59 immunohistochemistry staining was higher in the tumor tissue than its paired normal tissue, the expression of TRIM59 was defined as “high expression.” When the intensity of TRIM59 IHC staining was lower in the tumor tissue than its paired normal tissue or no change was detected, its expression was defined as “low expression” (n = 23 in the low-expression group, no change of TRIM59 expression levels was found in 3 of the 23 tumor samples; n = 42 in the high-expression group). (E) High intensity of TRIM59 immunostaining strongly associates with poor patient survival (n = 44 in the TRIM59 high-expression group, n = 23 in the TRIM59 low-expression group). Data are presented as means ± SEM. *P < .05. (B and C) A t test was used for the statistical analysis. (E) The log-rank (Mantel–Cox) test was used. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions

Figure 3 TRIM59 promotes gastric cancer cell proliferation, clone formation, and migration. (A) TRIM59 knockdown reduces the proliferation rate of AGS cells. The first sh-TRIM59 sequence was ACATTACAGGCAACCAT-TAAA and the second sh-TRIM59 sequence was TTGCACTAAGGGCTATTATTG (means ± SD; n = 8; blank, nontreated cells). (B) Overexpression of TRIM59 accelerantes the proliferation of MKN45 cells (means ± SD; n = 8). (C) TRIM59 knockdown inhibits the capability of clone formation in soft agar of AGS cell. The first sh-TRIM59 sequence was used in this experiment (means ± SEM; n = 4). (D) Overexpression of TRIM59 promotes MKN45 cell clone formation in soft agar (means ± SEM; n = 3). (E and F) TRIM59 knockdown inhibits AGS cell migration, whereas TRIM59 overexpression exerts the opposite effect (means ± SEM; n = 5). *P < .05; ***P < .001. One-way analysis of variance was used for the statistical analysis. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions

Figure 4 TRIM59 enhances the tumorigenicity of gastric cancer cells in vivo. (A) TRIM59 gene silencing by shRNA resulted in suppressed tumorigenicity of AGS cells in vivo. Only 4 of 10 mice inoculated with TRIM59-shRNA lentivirus-infected AGS cells form a tumor. The first sh-TRIM59 sequence was used in this experiment. (B) Reduced tumor volumes and weights of xenografts generated by AGS cells transfected with TRIM59-shRNA (n = 10; blank, nontreated AGS cells). (C) Xenograft tumors from TRIM59-shRNA AGS cells contain significantly less Ki67-positive proliferative cells (n = 10, 5 fields were random picked, examined under a fluorescent microscope, captured and counted per xenograft sample). (D) Enhanced tumorigenicity of MKN45 cells in vivo by TRIM59 overexpression. (E) Increased tumor volumes and weights of xenografts generated by MKN45 cells stably expressing TRIM59 (n = 10) (means ± SEM). *P < .05, ***P < .001. One-way analysis of variance was used for the statistical analysis. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions

Figure 5 TRIM59 negatively regulates P53 tumor-suppressor stability and represses the expression of P53 downstream molecules. (A) Immunoblot results show that TRIM59 knockdown by shRNA leads to up-regulation of P53 tumor suppressor and its downstream targets 14-3-3σ, p21, and the cell-cycle regulator cyclin D1, but does not affect the protein amount of p27 in the AGS cell line. Right: Quantification of 3 independent experiments is shown. The immunoblotting intensity of nontransfected cells was used for data normalization. (B) TRIM59 knockdown increases the mRNA expression of P53 downstream genes 14-3-3σ, p21, and NOXA, although it does not change the p53 mRNA level in the AGS cell line (n = 3). (C) TRIM59 inhibits the luciferase activities of P53 reporter containing multiple P53 DNA binding sites in the MKN45 cell line. The expression of the luciferase gene was driven by the P53-responsive target DNA sequence (responds to and can be activated by P53). Experiments were performed in triplicate. (D) The protein level of P53 is up-regulated in xenograft tumors generated from AGS cells stably transfected with TRIM-shRNA. (E) Knockdown of TRIM59 expression increases apoptosis in xenograft tumors of AGS cells. Representative pictures are shown. Means ± SEM. *P < .05, **P < .01. A t test was used for the statistical analysis. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions

Figure 6 TRIM59 binds to P53, which in turn facilitates the ubiquitination and degradation of P53. (A) The protein level of P53 decreases correspondingly with progressively increasing amounts of TRIM59 in 293T cells. (B) The expression of TRIM59 reduces the protein abundance of co-transfected P53 in 293T cells (left) and endogenous P53 in the gastric cancer cell line MKN45 (right) (n = 3). (C) The half-life of P53 in 293T cells (left) and MKN45 cells (right) transfected with TRIM59 is prolonged by MG132 treatment (n = 3). (D) Trim59 co-immunoprecipitates with P53 in 293T cells transfected with P53 and flag-tagged TRIM59 (left). Endogenous P53 and TRIM59 form a protein complex in AGS, GES-1, and 293T cells (right). (E) Ubiquitination of P53 is enhanced by TRIM59 overexpression and is attenuated by TRIM59 knockdown. AGS cells transfected with scramble shRNA or sh-TRIM59, and MKN45 cells transfected with empty vector or TRIM59 vector, were lysed at 36 hours after transfection. Cells were incubated with the proteasome inhibitor MG132 (10 μmol/L) for 9 hours before harvest. The cell lysates then were subjected to immunoprecipitations using antibody against P53. Anti-ubiquitin or the anti-P53 antibody was used for immunoblotting to determine the ubiquitination status of P53. (1) AGS cells transfected with scramble shRNA; (2) AGS cells transfected with sh-TRIM59; (3) MKN45 cells transfected with empty vector; (4) MKN45 cells transfected with TRIM59 expression plasmid. Experiments were repeated 3 times. Immunoblotting results of TRIM59 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from 10% of the input are shown at the bottom panel. Means ± SEM. *P < .05, **P < .01. One-way analysis of variance was used for the statistical analysis. DMSO, dimethyl sulfoxide. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions

Figure 7 TRIM59 exerts the tumor-promoting activities through the P53 tumor suppressor. (A) sh-TRIM59 and sh-P53 were used to knock down TRIM59 and P53, respectively, in AGS cells. The first sh-TRIM59 sequence was used in this experiment. The shRNA for P53 knockdown was AGTAGATTACCACTGGAGTCTT. (B) P53 knockdown attenuates the inhibitory effect on the AGS cell proliferation by TRIM59 knockdown. Cells were incubated with the CellTiter 96 Aqueous One Solution reagent (Promega, Madison, WI) for 3 hours. The absorbance at 490 nm then was measured on a microplate reader for quantification of viable cells (n = 3). (C) Knockdown of P53 expression significantly increased the impaired clone formation of TRIM59-shRNA–transfected AGS cancer cells (n = 4). (D) P53 knockdown accelerated the impeded in vivo tumorigenicity of AGS cancer cells stably expressing sh-TRIM59 (n = 10). (E and F) mRNA expression levels of TRIM59 negatively correlated with (E) the P53 downstream gene p21 (n = 40) and positively correlated with (F) cyclin D1 (n = 36) in human gastric cancer samples. Means ± SEM. *P < .05, **P < .01, ***P < .001. (A–D) One-way analysis of variance was used for the statistical analysis. (E and F) The Pearson correlation was used for the analysis. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Gastroenterology 2014 147, 1043-1054DOI: (10.1053/j.gastro.2014.07.021) Copyright © 2014 AGA Institute Terms and Conditions