Selenium-binding protein 1 transcriptionally activates p21 expression via p53-independent mechanism and its frequent reduction associates with poor prognosis in bladder cancer

Background Recent studies have shown that selenium-binding protein 1 (SELENBP1) is significantly down-regulated in a variety of solid tumors. Nevertheless, the clinical relevance of SELENBP1 in human bladder cancer has not been described in any detail, and the molecular mechanism underlying its inhibitory role in cancer cell growth is largely unknown. Methods SELENBP1 expression levels in tumor tissues and adjacent normal tissues were evaluated using immunoblotting assay. The association of SELENBP1 expression, clinicopathological features, and clinical outcome was determined using publicly available dataset from The Cancer Genome Atlas bladder cancer (TCGA-BLCA) cohort. DNA methylation in SELENBP1 gene was assessed using online MEXPRESS tool. We generated stable SELENBP1-overexpression and their corresponding control cell lines to determine its potential effect on cell cycle and transcriptional activity of p21 by using flow cytometry and luciferase reporter assay, respectively. The dominant-negative mutant constructs, TAM67 and STAT1 Y701F, were employed to define the roles of c-Jun and STAT1 in the regulation of p21 protein. Results Here, we report that the reduction of SELENBP1 is a frequent event and significantly correlates with tumor progression as well as unfavorable prognosis in human bladder cancer. By utilizing TCGA-BLCA cohort, DNA hypermethylation, especially in gene body, is shown to be likely to account for the reduction of SELENBP1 expression. However, an apparent paradox is observed in its 3′-UTR region, in which DNA methylation is positively related to SELENBP1 expression. More importantly, we verify the growth inhibitory role for SELENBP1 in human bladder cancer, and further report a novel function for SELENBP1 in transcriptionally modulating p21 expression through a p53-independent mechanism. Instead, ectopic expression of SELENBP1 pronouncedly attenuates the phosphorylation of c-Jun and STAT1, both of which are indispensable for SELENBP1-mediated transcriptional induction of p21, thereby resulting in the G0/G1 phase cell cycle arrest in bladder cancer cell. Conclusions Taken together, our findings provide clinical and molecular insights into improved understanding of the tumor suppressive role for SELENBP1 in human bladder cancer, suggesting that SELENBP1 could potentially be utilized as a prognostic biomarker as well as a therapeutic target in future cancer therapy.


Background
Bladder cancer is one of the most common urological malignancies, and is the leading cause of deaths in patients with urinary tract disease. The incidence of bladder cancer has steadily increased worldwide in recent decades [1]. According to the report from American Cancer Society in 2018, there are an estimated 81,190 of new cases diagnosed with bladder cancer and approximately 17,240 of bladder cancer-related deaths in the United States [2]. For non-muscle-invasive and muscle-invasive bladder cancer, grade and depth of tumor invasion are the most important prognostic factors, respectively [3]; the 5-year overall survival rate is approximately 8.1% among patients with distant metastasis [4]. Therefore, the identification of novel therapeutic targets or predictive biomarkers, and in-depth elucidation of the underlying molecular mechanisms are of tremendous importance for reducing the mortality of this disease.
The human selenium-binding protein 1 (SELENBP1) gene, located on chromosome 1 at q21-22, encodes a highly-conserved member of the selenium-binding protein family [5], and is the homologue of the mouse SP56 gene that has been originally reported as a 56 kDa mouse protein with the ability to bind 75 selenium stably [6,7]. The human SELENBP1 protein is ubiquitously expressed in various tissue types, especially higher in heart, lung, liver and kidney [5]. Since the initial report identifying SELENBP1 as a tumor-associated protein in prostate cancer [8], the reduced expression or even lost of SELENBP1 has been consistently observed in a variety of solid tumors as compared to corresponding normal tissues, including those of the skin [9], lung [10], esophagus [11], stomach [12][13][14], colon [15,16], liver [17], breast [18] and ovary [19]. Thereafter, accumulating evidence has convincingly demonstrated that reduced expression of SELENBP1 is an independent predictive of poor clinical outcome in multiple malignant diseases [13][14][15][16][17][18][20][21][22]. Furthermore, an increasing number of in vitro and in vivo studies has also consistently shown that increasing the levels of SELENBP1 significantly suppresses the malignant characteristics of cancer cells, leading scientific community to the consensus that SELENBP1 may act as a putative tumor suppressor involved in the regulation of cell proliferation, senescence, epithelial-mesenchymal transition, migration and apoptosis [23]. However, the clinical significance of SELENBP1 in human bladder cancer has not yet been characterized in any detail. Additionally, the molecular mechanisms underlying the tumor-suppressive role for SELENBP1 in cancer cells are still largely undefined.
Here we show that SELENBP1 is significantly downregulated in human bladder cancer tissues and cell lines, and its frequent reduction is further associated with tumor progression as well as poor clinical outcome among patients with bladder cancer. In addition, the reduced expression of SELENBP1 is inversely associated with DNA hypermethylation in its promoter region, but more significantly correlated negatively with DNA methylation in gene body. Importantly, an apparent paradox is observed in 3′-UTR region, in which DNA methylation is positively related to SELENBP1 expression. Furthermore, we verify the growth inhibitory role for SELENBP1 in human bladder cancer, and report a novel function for SELENBP1 in transcriptionally modulating p21 expression through a p53-independent mechanism. Instead, ectopic expression of SELENBP1 pronouncedly attenuates the phosphorylation of c-Jun and STAT1, both of which are indispensable for SELENBP1-mediated transcriptional induction of p21, thereby resulting in the G 0 / G 1 phase cell cycle arrest in bladder cancer cell.

Human bladder cancer tissue specimens
Twelve pairs of primary invasive bladder cancer specimens and their paired adjacent normal bladder tissues were obtained from patients who received radical cystectomy from 2012 to 2013 at the Department of Urology of the Union Hospital of Tongji Medical College (Wuhan, China) [34]. After surgical procedure, all of specimens were immediately collected and stored in liquid nitrogen, and the pathological diagnoses were further confirmed by a certified clinical pathologist according to the 2004 World Health Organization Consensus Classification and Staging System for bladder neoplasms [35]. Before collecting these specimens, ethical approval has been obtained from the Medical Ethics Committee of China, and all patients have given their informed consent for translational research.

Transfection and luciferase reporter assay
All of stable or transient transfections were performed using PolyJet ™ DNA in vitro transfection reagent (Sig-naGen Laboratories, Rockville, MD) according to manufacturer's instructions. Stable transfectants were selected with puromycin, hygromycin B or G418 for 3-4 weeks [26,27], and surviving cells from the antibiotics selection were pooled as stable mass transfectants, with at least two passages prior to utilization for further experiments. For the determination of p21 promoter-mediated luciferase activity, cells were transfected with the related luciferase reporter in combination with the pRL-TK vector (Promega, Madison, WI). After 24 h of transfection, luciferase activity was determined using a dualluciferase reporter assay system kit (Promega, Madison, WI) as described previously [36,37]. The results were normalized by internal TK signal, and expressed as mean ± standard deviation from at least three independent experiments.

Anchorage-independent growth assay
Anchorage-independent growth assay in soft agar (soft agar assay) was performed as described in our previous study [31,39]. Briefly, 1 × 10 4 of indicated cells were resuspended in 1 mL of 10% FBS basal modified Eagle's medium (BME) containing 0.33% soft agar and seeded on top of 0.5% agar in 10% FBS BME in each well of 6-well plates. The plates were maintained in a 5% CO 2 incubator at 37 °C for 2-3 weeks. Colonies with more than 32 cells were scored and presented as mean ± standard deviation from at least three independent experiments.

Cell-cycle analysis
As previously described [27], the indicated cells were harvested and fixed in 75% ethanol overnight, and were suspended in the staining buffer containing 0.1% Triton X-100, 0.2 mg/mL RNase A, and 50 mg/mL propidium iodide (PI) at 4 °C. DNA content was then determined by a flow cytometry Epics XL flow cytometer (Beckman Coulter Inc., Miami, FL) and the results were analyzed with EXPO32 (v1.2) software.

Bioinformatics analysis of data mining in TCGA-bladder cancer (TCGA-BLCA) cohort
The TCGA-BLCA clinical data and SELENBP1 expression data were obtained from UCSC Xena Browser (https ://xenab rowse r.net; last updated on April 5th 2019) [40]. UCSC Xena browser shows that TCGA has 406 patients with bladder cancer, in which 19 of patients have SELENBP1 expression data of primary tumor and matched normal bladder tissues. Patients are categorized into low or high SELENBP1 expression group using cut-off value of X-2.33SD (X, mean; SD, standard deviation) calculated from SELENBP1 expression in the matched normal bladder tissues. These two groups were applied for Kaplan-Meier survival analysis (10-year survival). The correlation between SELENBP1 expression and clinicopathological parameters in TCGA-BLCA were also examined. Besides, we analyzed the association between SELENBP1 expression and DNA methylation using the MEXPRESS tool (http:// mexpr ess.be/), a user-friendly tool for the visualization and interpretation of DNA methylation in TCGA data [41].

Statistical analysis
Statistical analysis was performed using SPSS 19.0 software (SPSS, Chicago, IL). Student's t-test or Fisher's exact test was used to determine significant differences. p < 0.05 is considered statistically. The results are expressed as the mean ± standard deviation from at least three independent experiments.

SELENBP1 is frequently down-regulated in human bladder cancer, and low expression of SELENBP1 predicts poor clinical prognosis
To determine SELENBP1 expression levels in human bladder cancer, we evaluated the protein expression of SELENBP1 in 12 pairs of bladder cancer tissues and matched adjacent normal tissues using immunoblotting assay. As shown in Fig. 1a, b, SELENBP1 protein levels p=0.0021 were significantly down-regulated in 9 out of 12 bladder cancers relative to matched adjacent normal tissues (p = 0.0021). We further examined SELENBP1 mRNA levels in The Cancer Genome Atlas bladder cancer (TCGA-BLCA) cohort, including 406 patients with SELENBP1 expression data, in which 19 of patients have SELENBP1 expression data of primary tumor and matched normal bladder tissues. Consistently, reduction of SELENBP1 expression was also frequently observed in a majority of tumor tissues from TCGA-BLCA cohort (p < 0.0001; Fig. 1c). Utilizing the gene expression data in matched normal tissues, we created two groups based on SELENBP1 gene expression and analyzed overall survival (OS) curves of differential SELENBP1 expression levels. Log-rank test of the OS curves in TCGA-BLCA cohort indicated that bladder cancer patients with low SELENBP1 expression were remarkably related to worse prognosis than those with high SELENBP1 expression ( Fig. 1d), implying that SELENBP1 might serve as a tumor suppressor in bladder cancer.

Low expression of SELENBP1 correlates with unfavorable clinicopathological characteristics of bladder cancer patients
Next, we analyzed the relationship between SELENBP1 expression and clinicopathological features in bladder cancer by using TCGA-BLCA cohort (Table 1), and found that low expression of SELENBP1 was markedly correlated with non-papillary subtype (Papillary: 52% vs. Non-papillary: 64%; p = 0.023), and advanced tumornode-metastasis (TNM) stage (I/II: 53% vs. III/IV: 64%; p = 0.029), and distant metastasis with marginal statistical significance (p = 0.057), but not with age, gender, lymphovascular invasion (LVI) or other clinicopathological features. These results strongly suggest that the reduced expression of SELENBP1 is associated with the aggressive phenotype of bladder cancer.

SELENBP1 expression inversely associates with DNA methylation in the promoter and gene body, but positively correlates with DNA methylation in 3′-UTR region
DNA methylation has been recognized as a major mechanism responsible for epigenetic silencing of SELENBP1 expression in esophageal and colorectal cancer [11,42]. Therefore, we addressed this association between SELENBP1 expression and DNA methylation in the bladder cancer by using the MEXPRESS tool (http://mexpr ess.be/), a user-friendly tool for the visualization and interpretation of DNA methylation in TCGA data [41]. As shown in Fig. 2, it was clear that normal samples cluster towards higher SELENBP1 expression than tumor tissues (Wilcoxon rank-sum test, p = 9.01e−7). Moreover, there was a significant inverse association between SELENBP1 expression and DNA methylation in its promoter region located close to the transcription start site (TSS) and the first exon (Pearson correlation coefficients Interestingly, we observed a more significant negative relationship between gene body methylation and SELENBP1 expression with Pearson r = − 0.619 (p < 0.001). However, DNA methylation in its 3′ untranslated region (UTR) was positively associated with SELENBP1 expression (Pearson r = + 0.482; p < 0.001). These observations highlight that those multiple genomic regions of DNA methylation may be differentially involved in modulating SELENBP1 expression.

Overexpression of SELENBP1 inhibits bladder cancer cell growth
To determine the biological function of SELENPB1 in bladder cancer, we firstly employed immunoblotting assay to evaluate SELENBP1 expression in 8 of human bladder cancer cell lines, including T24, T24T, UMUC3, RT4, RT112, TCCSUP, J82 and 5637. As expected, the majority of bladder cancer cell lines exhibited undetectable SELENBP1 protein levels, except TCCSUP and J82 cell lines (Fig. 3a). Then, we established stable transfectants expressing HA-tagged SELENBP1 in UMUC3 cells (Fig. 3b). As shown in trypan blue exclusion assay (Fig. 3c), ectopic expression of SELENBP1 in UMUC3 cells led to a significant decrease in anchoragedependent growth. Additionally, we also observed that overexpression of SELENBP1 markedly inhibited anchorage-independent growth of UMUC3 cells (Fig. 3d), further verifying the tumor-suppressive role for SELENBP1 in human bladder cancer cellular transformation.
Induction of p21 protein is crucial for SELENBP1-mediated G 0 /G 1 phase cell cycle arrest To elucidate the underlying mechanism responsible for its inhibitory role in cancer cell growth, the potential effect of SELENBP1 on cell cycle was determined by flow cytometry analysis. The results showed that overexpression of SELENBP1 induced a significant G 0 /G 1 phase arrest in UMUC3 cells (Fig. 4a), suggesting that G 0 /G 1 phase arrest induction might be associated with tumor suppressive role of SELENBP1. To further elucidate the crucial players that are involved in SELENBP1-mediated G 0 /G 1 phase arrest, we then evaluated the regulatory effect of SELENBP1 on multiple cell cycle-associated  Fig. 3 Overexpression of SELENBP1 inhibits bladder cancer cell growth. a Cell lysates of human bladder cancer cell lines were extracted and subjected to immunoblotting assay for determining the levels of SELENBP1 protein expression. GAPDH was used as a loading control. b UMUC3 cells were stably transfected with empty vector (Vector) or construct encoding HA-tagged SELENBP1 (HA-SELENBP1). Cell lysates of these stable transfectants were extracted and subjected to immunoblotting assay by using anti-HA antibody. GAPDH was used as a loading control. c The number of 1 × 10 4 stable UMUC3 (Vector) and UMUC3 (HA-SELENBP1) cells per well was seeded in 6-well plates. The number of viable cells at the indicated time points was determined by trypan blue exclusion assay. The values were shown as mean ± standard deviation from at least three independent experiments. The asterisk (*) indicates a significant difference in comparison to that of control group (p < 0.05). d Anchorage-independent cell growth of UMUC3 cells stably transfected with empty vector or construct encoding HA-SELENBP1 was determined by soft agar assay. The colony formation was observed and captured under an inverted microscope (×40 magnification; left panel) and numbers of colonies were scored on day 15, and presented as colonies (×100)/10 4 cells (right panel). The values were shown as mean ± standard deviation from at least three independent experiments. The asterisk (*) indicates a significant difference in comparison to that of control transfectants (p < 0.05) proteins, including cyclin-dependent kinase inhibitor 1A (CDKN1A; also known as p21), cyclin-dependent kinase inhibitor 1B (CDKN1B or p27), cyclin D1, cyclindependent kinase 4 (CDK4) and CDK6. As shown in Fig. 4b, overexpression of SELENBP1 had no observable regulatory role on p27, Cyclin D1, CDK4 or CDK6, but did result in a dramatic induction of p21 protein in both of UMUC3 and T24T cells. This finding suggests that p21 protein is likely to be a crucial player in SELENBP1-mediated cancer cell growth inhibition. To verify this notion, we established stable transfectants expressing HA-tagged SELENBP1 protein in human colon cancer HCT116 wild-type cells (HCT116 WT) and p21 knockout cells (HCT116 p21 −/− ), in both of which endogenous protein levels of SELENBP1 were undetectable [30]. As expected, overexpression of SELENBP1 consistently stimulated p21 protein expression and further significantly inhibited anchorage-independent cell growth in HCT116 WT cells (Fig. 4c, d). However, SELENBP1 failed to modulate cancer cell growth in p21-deficiency cells (Fig. 4d)

Fig. 4
Induction of p21 protein is crucial for SELENBP1-mediated G 0 /G 1 phase cell cycle arrest. a UMUC3 (Vector) and UMUC3 (HA-SELENBP1) cells were collected at day 4 as described in " Fig. 3c", and subjected to flow cytometry assay following the staining of propidium iodide. b Cell lysates of stable transfectants UMUC3 (Vector), UMUC3 (HA-SELENBP1), T24T (Vector) and T24T (HA-SELENBP1) were extracted and subjected to immunoblotting assay for determining expression levels of G 0 /G 1 phase arrest-associated regulatory proteins. GAPDH was used as a loading control. Densitometric quantification of p21 (relative to GAPDH) is shown under each blot. c HCT116 WT and HCT116 p21 −/− cells were transfected with empty vector or HA-SELENBP1, each along with pGIPZ, and stable transfectants were established under the selection of puromycin. Cell lysates of these stable transfectants were extracted and then subjected to immunoblotting assay with anti-SELENBP1, anti-p21 or anti-GAPDH antibodies. d Anchorage-independent cell growth of stable transfectants described in "c" was determined by soft agar assay. Representative images of colonies were visualized under an inverted microscope (×40 magnification; left panel) and relative numbers of colonies were calculated on day 12 (right panel). The asterisk (*) indicates a significant difference in comparison to that of control transfectants (p < 0.05) that p21 protein serves as a crucial downstream mediator responsible for the growth-inhibitory role of SELENBP1 in cancer cells. Collectively, these results demonstrate that SELENBP1 specifically up-regulates p21 protein levels and induces G 0 /G 1 phase arrest, thereby leading to attenuation of cancer cell growth.

SELENBP1 transcriptionally stimulates p21 expression in a p53-independent manner
To investigate the mechanisms underlying SELENBP1mediated induction of p21 protein, we first examined p21 mRNA levels, and found that p21 mRNA was markedly up-regulated in both UMUC3 and T24T stable transfectants expressing HA-SELENBP1 (Fig. 5a). To further verify whether SELENBP1-mediated induction of p21 expression occurs at transcriptional level, dual luciferase activities were evaluated in stable UMUC3 (Vector) and UMUC3 (HA-SELENBP1) transfectants using one fulllength (2.4 Kb) and three differential truncated-length (1.8 Kb, 1.3 Kb and 200 bp) of human p21 promoterdriven luciferase reporters (Fig. 5b). As indicated in Fig. 5c, overexpression of SELENBP1 significantly led to an approximately 2.3-fold increase in full-length (2.4 Kb), 1.8 Kb-or 1.3 Kb-length of p21 promoter-driven luciferase activities, whereas this regulatory effect was not observed in 200 bp-length reporter assay. These findings reveals that SELENBP1 transcriptionally stimulates p21 expression through the potential responsive elements located within approximately ~ 1300 bp to ~ 200 bp region of the p21 upstream promoter. It has been well established that tumor suppressor p53 is able to binding to the p21 promoter and directly activate its expression [43]. However, we noticed that p53-responsive elements, located in − 1394 bp and − 2285 bp region of p21 promoter [33,43], are outside of identified SELENBP1responsive region in p21 promoter (Fig. 5b), leading us to the notion that SELENBP1-mediated p21expression is likely to occur through a p53-independent mechanism. To verify this notion, we re-introduced HA-tagged SELENBP1 in human colon cancer HCT116 p53 knockout cells (HCT116 p53 −/− ), and found that SELENBP1 still retained the ability to activate p21 expression even in the absence of p53 (Fig. 5d). This result indicates that p53 protein might not be required for SELENBP1-mediated  regulation of p21 expression. Collectively, despite the well-established putative role for p53 in the transcriptional regulation of p21 expression, our results demonstrate that SELENBP1 triggers transcriptional induction of p21 expression through a p53-independent pathway.

Overexpression of SELENBP1 results in the suppression of c-Jun and STAT1 phosphorylation
In addition to p53 protein, bioinformatics analysis revealed that p21 promoter region also contains multiple consensus binding sites for other transcriptional factors [44], including signal transductions and activators of transcription (STATs), activating protein-1 (AP-1) and specificity protein 1 (Sp1) (Fig. 5b). To identify whether these transcriptional factors are involved in SELENBP1mediated transcriptional induction of p21, we used immunoblotting assay to evaluate the levels of transcription factors Sp1 and AP-1 subunits-JunB, JunD and c-Jun, as well as STAT family members, including STAT1, STAT3, STAT5 and STAT6. The results showed that both c-Jun phosphorylation at Ser73 and STAT1 phosphorylation at Tyr701 were significantly down-regulated in UMUC3 cells expressing HA-SELENBP1, whereas other evaluated transcription factors showed no observable changes (Fig. 6a). This result suggests that c-Jun and/or STAT1 may be involved in SELENBP1-mediated transcriptional induction of p21.

Combined suppression of c-Jun and STAT1 is required for SELENBP1-mediated transcriptional induction of p21 protein
Next, to define the roles of c-Jun and STAT1 in the regulation of p21 protein, dominant-negative mutants of c-Jun (TAM67) [31] and STAT1 (DN-STAT1 Y701F) [32] were stably transfected into UMUC3 cells, respectively. The results showed that TAM67 efficiently down-regulated c-Jun phosphorylation at Ser73, but inhibited the expression of p21 protein (Fig. 6b), indicating that inhibition of c-Jun activity alone is not sufficient to mimic the effect of SELENBP1-mediated induction of p21 protein. Conversely, blockade of STAT1 activity alone by DN-STAT1 Y701F slightly increased the expression of p21 protein (Fig. 6c), though dominant-negative STAT1 resulted in an effective down-regulation of suppressor of cytokine signaling 1 (SOCS1), a negative regulator of JAK/STAT signaling [45]. These observations led us to hypothesize whether simultaneous blockade of c-Jun and STAT1 activity is required for the induction of p21 protein.
To verify this notion, we introduced pcDNA3.1 control and TAM67 into stable UMUC3 (Vector) and UMUC3 (DN-STAT1 Y701F) transfectants, respectively. As shown in Fig. 6d, concurrent blockade of c-Jun and STAT1 activities resulted in a significant increase in the expression of p21 protein, revealing that suppression of both c-Jun and STAT1 activity is indispensable for SELENBP1-mediated induction of p21 protein. Consistent with the induction of p21 protein, the truncated-length (1.3 Kb) of human p21 promoter-driven luciferase reporter assay revealed that p21 promoter activity increased significantly following the introduction of increasing amounts of TAM67 into stable UMUC3 (DN-STAT1 Y701F) cells (Fig. 6e). Meanwhile, we noticed that inhibition of c-Jun phosphorylation by TAM67 alone attenuated p21 promoter activity in stable UMUC3 (Vector) cells (Fig. 6e). This observation was, however, in agreement with TAM67mediated inhibition of p21 protein (Fig. 6b), suggesting that c-Jun regulates p21 expression in a context-dependent manner, at least dependent on the STAT1 activity. More importantly, we found that simultaneous blockade of c-Jun and STAT1 activity induced a significant G 0 /G 1 phase arrest in UMUC3 cells (Fig. 6f ). These mechanistic insights further highlight the indispensable roles for both c-Jun and STAT1 in SELENBP1-mediated transcriptional induction of p21 protein and subsequent G 0 /G 1 phase arrest. Taken together, our results demonstrate overexpression of SELENBP1 triggers transcriptional activation of p21 through concurrent suppression of c-Jun and STAT1 phosphorylation.

Discussion
Reduced expression of selenium-binding protein 1 (SELENBP1), since first cloned in 1997, has been frequently observed in a variety of solid tumors, and is significantly related to unfavorable clinical outcome in multiple cancer types [23]. However, the evidence concerning the clinical relevance of SELENBP1 in human bladder cancer is lacking. In the present study, we are the first to show that the reduction of SELENBP1 expression is also a frequent event in human bladder cancer tissues ( Fig. 1), but is more likely to occur in patients with nonpapillary subtype and advanced TNM stage (Table 1). Furthermore, restoration of SELENBP1 expression results in a significant G 0 /G 1 phase arrest and subsequent cell growth inhibition in human bladder cancer cells (Figs. 3  and 4). Collectively, we report here that SELENBP1 may serve as a promising prognostic biomarker in predicting poor clinical outcome among patients with bladder cancer. Nevertheless, how SELENBP1 is clinically incorporated into bladder cancer management together with other prognostic biomarkers merits further investigation. Cytosine methylation is a DNA modification generally associated with epigenetic transcription silencing [46]. The publicly available TCGA-BLCA data, analyzed by the online MEXPRESS tool [41], indicates that multiple locations of SELENBP1 gene are hyper-methylated in a majority of bladder cancer tumors (Fig. 2). In line with Page 10 of 13 Wang et al. J Transl Med (2020) 18:17 prior reports in esophageal and colorectal cancer [11,42], DNA methylation in the immediate vicinity of the TSS is inversely associated with SELENBP1 expression (Pearson correlation coefficients ranging from − 0.339 to − 0.505; p < 0.001). Intriguingly, this inverse association is much more significant between methylation in gene body and SELENBP1 expression (Pearson r = − 0.619, p < 0.001), suggesting that DNA hypermethylation, especially in gene body, is likely to account for the reduction of SELENBP1 expression in bladder cancer. However, there is a significant and positive relationship between DNA methylation in 3′-UTR region and SELENBP1 expression (Pearson r = + 0.482; p < 0.001). This apparent paradox supports that DNA methylation in different genomic contexts might be differentially involved in the regulation of SELENBP1 expression [47]. Recently, DNA methylation-based biomarker studies have predominantly been focused on the effects of hypermethylation of promoter in tumour-suppressor genes [48]. Therefore, our finding provides valuable information when developing SELENBP1 as a genomic methylation-based biomarker for tumor progression and prognostic outcome. To date, the molecular functions of SELENBP1 in cancer are relatively poorly understood, although it has been consistently reported to attenuate aberrant cell growth and suppress cell migration in a variety of cancer cells [49][50][51]. We and others have previously shown that SELENBP1 has the ability to form a physical interaction with glutathione peroxidase 1 selenoprotein [17,52], a well-characterized enzyme that detoxify hydrogen peroxide to nontoxic water [53]. These findings implies a novel role for SELENBP1 in modulating the cellular redox microenvironment, which is further supported by our recent report showing that SELENBP1 regulates the levels of extracellular reduced form of glutathione [30]. More recently, SELENBP1 has been recognized as a novel human methanethiol oxidase with an enzymatic role in sulfur metabolism [54]. Here, we establish a novel biochemical link between SELENBP1 and p21 protein, and further determine the indispensable role for p21 in SELENBP1-mediated growth inhibition in cancer cells (Fig. 4). However, whether glutathione or sulfur metabolism is involved in SELENBP1-mediated induction of p21 protein warrants further investigation. The p21 protein, also known as p21 WAF1/CIP1 , is encoded by the CDKN1A gene mapped to chromosome 6p21.2, and belongs to the Cip/Kip family of CDK inhibitors [55]. A substantial number of studies have demonstrated p21 as a multifunctional, broad-acting protein that plays key roles in cell cycle regulation, cell migration and apoptosis [44]. Thus, p21 has been shown to be intricately regulated via p53-dependent and -independent pathways [56]. Here, we verify that SELENBP1 activates transcriptional induction of p21 through a p53-independent mechanism using p53-deficiency cancer cell model (Fig. 5). Given the high prevalence of TP53 mutation and/or lose of function across almost all of cancer types, our finding suggests that SELENBP1 probably retains the growth-inhibitory capacity even in p53-mutated cancer cells.
Independent of p53, a variety of transcription factors have been reported to transcriptionally activate p21 protein expression [56], including Sp1. However, there is no observable effect on this player following the restoration of SELENBP1 in bladder cancer cells (Fig. 6). By contrast, SELENBP1 specifically inhibits the phosphorylation of c-Jun and STAT1, belonging to AP-1 and STAT family, respectively. Our further analysis suggests that combined suppression of c-Jun and STAT1 activities may be essential for SELENBP1-mediated transcriptional induction of p21 protein. C-Jun has been previously shown to interact with STAT3 and co-operatively regulate the transcription of their target genes in bladder cancer [57]. Given that c-Jun responsive element is closely located at the potential STAT1 binding sites in the promoter of p21, it is intriguing to further determine whether there is a physical interaction between c-Jun and STAT1 in the context of SELENBP1-mediated p21 expression, and the results from which would improve our understanding of the tumor-suppressive role for SELENBP1 in cancer.

Conclusion
In summary, our results have revealed that the reduction of SELENBP1 is a frequent event and significantly correlates with tumor progression as well as unfavorable prognosis among patients with bladder cancer. Further bioinformatics analysis suggests that DNA hypermethylation, especially in gene body, is likely to account for the reduction of SELENBP1 expression in bladder cancer. More importantly, we have established a novel biochemical link between SELENBP1 and p21 protein, in which SELENBP1 activates transcriptional induction of p21 expression in a p53-independent mechanism. Taken together, our results provide clinical and molecular evidence for an improved understanding of the tumor-suppressive role of SELENBP1 in bladder cancer, suggesting that SELENBP1 could potentially be used as a prognostic biomarker as well as a therapeutic target in future cancer therapy. Therefore, how SELENBP1 is clinically incorporated into bladder cancer management merits further investigation.