High OTUD3 expression is associated with a better prognosis in BC patients
OTUD3 is not frequently mutated in the TCGA pan-cancer dataset (https://www.cbioportal.org/). To assess OTUD3 expression in BC patients, we first analysed the gene expression UALCAN [24] (datasetshttp://ualcan.path.uab.edu/)of human BC. The results showed that the OTUD3 mRNA levels in the BC tissues were significantly lower than those in normal breast tissues (Fig. 1A). The OTUD3 mRNA levels were not associated with the individual cancer stages (Fig. 1B). Although there were no differences in the OTUD3 mRNA levels between the luminal samples and HER2-positive or triple-negative BC samples, there was a difference in the OTUD3 mRNA levels between the HER2-positive and triple-negative BC samples (Fig. 1C). We used data obtained from the cBioPortal database [25, 26] (https://www.cbioportal.org/)and found that decreases in OTUD3 mRNA levels may not be due to increased OTUD3 DNA methylation in BC (Fig. 1D).
Our previous results showed that OTUD3 could deubiquitinate and stabilize PTEN. Therefore, we performed a Kaplan-Meier survival analysis [22, 23] (http://kmplot.com/analysis/index) to evaluate the association between OTUD3, TP53 and PTEN expression and survival in BC patients. Interestingly, the RFS (Recurrence-free survival) rate in the high OTUD3 expression group (n=1986) was better than that in the low OTUD3 expression group (n=1965), p=0.00034 (Fig. 1 E). Similarly, the patients with high TP53 (n=1977) expression suffered a better RFS than the patients in the low expression group (n=1974), p=0.00054 (Fig. 1F). However, the expression level of PTEN was not an independent prognostic factor in the BC patients in the data set, p=0.62 (Fig. 1G). We speculate that high OTUD3 expression is associated with a better prognosis in BC patients and that the relationship between OTUD3 and p53 is the most significant.
OTUD3 expression is downregulated in BC tissue
Consistent with the OTUD3 mRNA results, the protein expression level of OTUD3 was also significantly lower in the BC tissue than the normal breast tissue. Immunohistochemical (IHC) staining with an anti-OTUD3 antibody was performed on 80 pairs of BC tissues and adjacent non-tumour tissues. Staining scores of 0-6 were considered negative, and scores of 6-12 were considered positive. As shown in Fig. 2A, substantial OTUD3 immunostaining was detected in the adjacent non-tumour tissue samples, whereas little to moderate OTUD3 staining was observed in the BC samples. According to the literature [27], BC was divided into the following subtypes: luminal A-like, luminal B-like (Her-2-negative), luminal B-like (Her-2-positive), Her-2 positive (non-luminal) and triple-negative (ductal). Notably, OTUD3 expression in the breast cancer tissue was independent of the molecular type (χ2=2.672,p=0.614)(Fig. 2B). Although there was no statistically significant difference in OTUD3 expression between the cancer and adjacent tissue in luminal B (Her-2- positive), the expression of OTUD3 in the adjacent normal tissues was higher than that in the cancer tissues in the luminal A (p=0.000), luminal B (Her-2-negative) (p=0.000), Her-2 positive (p=0.001) and triple-negative (p=0.000) BC tissues.
The half-life of the wild-type p53 protein is very short, and the p53 protein detected by IHC in various experiments is the mutant p53 protein [28, 29]. Therefore, we analysed the expression of p53 and OTUD3 in 26 pairs of fresh and frozen BC tissues and adjacent normal tissue by WB. The expression of the p53 downstream protein p21 was also detected. P21WAP1/cip1 is a p53-induced cell cycle kinase inhibitor (CDKI)[30]. P53 causes G1 cell arrest by regulating the expression of p21 [31].The knockout of p21 results in the complete loss of p53-mediated human tumour cell cycle (G1) arrest [32].The specimens were numbered by the date of collection and grouped by histological grade. Our experimental results prove for the first time that OTUD3 and p53 are both expressed in breast cancer and adjacent normal tissues. Notably, the expression of OTUD3(p=0.0069)and p53(p=0.041) in all BC tissues was lower than that in the corresponding normal tissues (Fig. 2C). OTUD3 expression in the breast cancer tissue was independent of the histological grade (F=1.736, p=0.199) (Fig. 2D).Additionally, the analysis of the relationship between OTUD3 expression and p53 expression showed a significantly positive correlation (r=0.8849, 95% CI: 0.8068-0.9326, p<0.0001), and the levels of OTUD3 and p21 were also positively correlated (r=0.6427, 95% CI: 0.4484-0.779, p<0.0001) (Fig. 2E). Our clinical data fully proved that OTUD3 is downregulated in cancer tissues and is highly correlated with p53 expression.
OTUD3 maintains p53 stability in vivo
Since the expression levels of OTUD3 and p53 are correlated,we tested whether the overexpression of OTUD3 affects p53 protein levels in breast cancer cells. Two well-established p53 target genes, p21 and the proapoptotic gene BCl-2-associated X protein (BAX), were assayed to reflect p53 activity. Apoptosis along the mitochondrial pathway is mediated by BAX [33] because the high expression of BAX can promote apoptosis [34]. Here, we used the luminal breast cancer cell line MCF7 and the TNBC cell line DU4475, both of which express wild-type p53 [35]. As shown in Fig. 3A, the p53 level was dramatically increased when OTUD3 was overexpressed in the MCF7 cells and DU4475 cells; increased p21 and BAX protein levels were also observed, indicating that OTUD3 also activates p53-dependent transcriptional activation. To confirm the role of OTUD3 in the regulation of p53, the MCF7 and DU4475 cells were transfected with sh-ctrl or sh-OTUD3 lentivirus. The changes in the p53 protein level were determined. We found that the OTUD3 knockdown in the BC cells resulted in a dramatic decrease in the protein level of endogenous p53 accompanied by a decrease in the two target genes (Fig. 3B). To test the possibility that OTUD3 regulation of p53 occurs through the modulation of p53 protein stability, we treated cells stably expressing OTUD3 shRNA with the proteasome inhibitor MG132 (20 µM, 8 h). The decrease in the p53 levels was reversed by the MG132 treatment (Fig. 3B). This finding suggests that OTUD3 regulates p53 levels by increasing its stability in a proteasome-dependent manner.
Ubiquitin-mediated degradation is the only way by which p53 is terminated by the proteasome[36]. To date, some p53 E3 ligases have been found, and MDM2 is the most important [37]. MDM2 (murine double minute 2, human HDM2) is an oncogene that promotes cell division and proliferation [38]. MDM2 mediates the sustainable degradation of most p53 proteins, resulting in a very low p53 intracellular level [39].P53 is regulated by Mdm2-mediated ubiquitination and degradation and has a short half-life (5–20 min) [40]. Subsequently, we examined the half-life of p53 in the absence or presence of OTUD3. We treated the control cells and cells stably expressing OTUD3 shRNA with or without OTUD3 overexpression with the protein synthesis inhibitor cycloheximide (CHX) and examined the p53 levels at various time points. The half-life of endogenous p53 was significantly shortened in the BC cells depleted of OTUD3, and this effect was fully reversed by the ectopic expression of OTUD3 (Fig. 3C). OTUD3 likely deubiquitinates p53 to counteract the action of the E3 ubiquitin ligase Mdm2. Indeed, as shown in Fig. 3D, ectopic Mdm2 expression significantly induced p53 degradation, while the coexpression of OTUD3 efficiently rescued p53 from Mdm2-induced degradation. These results demonstrate that OTUD3 can antagonize the reduction in p53 by Mdm2.
OTUD3 interacts with p53
Co-immunoprecipitation (CO-IP) assays were conducted in MCF7 cells and DU4475 cells to examine whether OTUD3 physically interacts with p53. As shown in Fig. 4A, endogenous OTUD3 was specifically co-immunoprecipitated with endogenous p53 in the cells by anti-p53 antibodies. Furthermore, p53 co-immunoprecipitated with endogenous OTUD3 in the MCF7 cells and DU4475 cells (Fig. 4B). Then, we constructed OTUD3 truncated mutants as follows: OTUD3 was divided into two parts per its domain structure, namely, D1 (1-183) containing the OTU domain and D2 (184-398) containing the UBA domain and the C tail (Fig. 4C). Using the D1 and D2 constructs, we revealed that the OTU domain-containing region D1 (1-183) is critical for the interaction between OTUD3 and p53 (Fig. 4D). To determine whether the OTUD3-TP53 interaction was direct, we generated and purified recombinant Myc-OTUD3, Myc-D1 and Myc-D2. Interestingly, Myc-D1 cultured from MCF7 cells was specifically bound by the purified GST-TP53 protein but not GST alone (Fig. 4E). This finding further illustrated the direct interaction between OTUD3 and p53.
TP53 is located on chromosome 17 (17p13.1) and encodes p53, which is a phosphoprotein comprising 393 amino acids. P53 consists of the following four domains: (I) an N-terminal sequence (transactivation domain, TAD) involved in the regulation of target gene transcription, recruitment of RNA polymerase and activation of the transcriptional (DNA-reading) machinery, (II)a highly conserved DNA-binding domain (DBD) that recognizes specific DNA sequences; (III) an oligomerization domain (OD) that assembles chains of other p53 monomers for tetramerization, and (IV) a C-terminal domain essential for the regulation of p53 activity. To determine the region where p53 binds OTUD3, we also constructed truncated p53 mutants (Fig. 4F) and performed GST assays. The results indicated that OTUD3 bound the following two regions of p53: T2 (1-113)/ (290-393) and T3 (1-113)/ (236-393) (Fig. 4G, H). These results suggest that a direct interaction exists between OTUD3 and p53. As OTUD3 is a deubiquitinase, we examined whether OTUD3 deubiquitinates p53.
OTUD3 deubiquitinates p53
To elucidate the molecular mechanism by which OTUD3 stabilizes p53, we determined whether OTUD3 directly controls the levels of p53 ubiquitination. As indicated in Fig. 5A, a high level of ubiquitinated p53 was found in the MCF7 and DU4475 cells transfected with Mdm2 (lane 2); however, p53 ubiquitination was significantly abrogated by OTUD3 expression (comparison of lanes 3 and 2). Significantly, the enzyme activity of the mutant OTUD3C76A lost its ability to deubiquitinate p53 (Fig. 5A, lane 4). This finding indicates that p53 stabilization by OTUD3 requires deubiquitinating enzymatic activity. In contrast, the OTUD3 downregulation by shRNA increased p53 ubiquitination in the MCF7 and DU4475 cells (Fig. 5B). Collectively, these data demonstrate that OTUD3 negatively regulates p53 ubiquitination in breast cancer cells and plays an important role in the balance between p53 ubiquitination and deubiquitination. We speculate that the balance between the Mdm2-mediated ubiquitination and OTUD3-mediated deubiquitination of p53 is critical for p53 stabilization.
OTUD3 inhibits cell proliferation
P53 stabilization is crucial for its suppression of cell growth and apoptosis. To investigate the biological role of OTUD3, we first examined its effect on cell proliferation. We compared the proliferation rates of MCF7 and DU4475 cells stably transfected with OTUD3 and OTUD3C76A with those of negative control cell lines using an MTS proliferation test kit. The results showed that cell proliferation slowed after the OTUD3 transfection and accelerated after the OTUD3C76A transfection (Fig. 6A). Compared to the control cells, the knockdown of endogenous OTUD3 by shRNA in the MCF7 and DU4475 cells increased the cell proliferation rate. However, the DU4475 cells proliferated at a significantly slower rate after OTUD3 expression was restored. OTUD3 could rescue the accelerated cell proliferation caused by the OTUD3 knockdown, but the enzyme active mutant OTUD3C76A could not inhibit the accelerated cell proliferation, indicating that the regulation of the effect of OTUD3 on cell proliferation depends on its ubiquitinase activity (Fig. 6B). These data suggest that OTUD3 can inhibit BC cell proliferation.
OTUD3 induces apoptosis in BC cells and inhibits colony formation
Chemotherapy is an important method for the treatment of BC. The P53-mediated pathways can be activated by genotoxic compounds, such as cisplatin chemotherapeutic compounds, leading to cell cycle arrest and cell death [41]. We treated BC cells with the chemotherapy drug cisplatin (10 mM, 24 h) and detected apoptosis. The results showed that the OTUD3-transfected cells were significantly more sensitive to cisplatin-induced apoptosis than the negative control cells (p<0.01); however, the sensitivity of the transfected OTUD3C76A cells was significantly decreased (p<0.01) (Fig. 7A). Compared with the negative control cells, the decreased OTUD3 protein levels increased the resistance of the tumour cells to cisplatin-induced apoptosis (p<0.01). The apoptosis rate was significantly increased when OTUD3 was restored to the levels of the cells not treated with shRNA (p<0.01), but when the levels of OTUD3C76A were restored in the cells, the apoptosis rate did not significantly change (Fig. 7B). These results suggest that OTUD3 has a certain response to chemotherapy-induced BC cell apoptosis and that this response depends on its deubiquitinase activity.
Subsequently, we examined the effect of OTUD3 on cell growth using a colony formation assay. The BC cells were infected with either a control vector or a vector encoding OTUD3 or OTUD3C76A and cultured for 2 weeks. Strikingly, OTUD3,but not OTUD3C76A, strongly inhibited the number of colonies of BC cells (Fig. 7C). When OTUD3 was knocked down, the cell clone formation ability was enhanced, and when OTUD3 was restored, the clone formation ability was inhibited (Fig. 7D). However, the ability of cell cloning was significantly enhanced after the OTUD3 C76A overexpression.