WDR4 is associated with bladder cancer progression

To clarify the differentially expressed proteins that drive malignant progression, tumor tissues (T) and normal tissues (N) from 17 MIBC patients were analyzed by label-free quantitative proteomics. We compared protein levels in these tissues and identified 524 proteins highly expressed in cancer tissues (Log2 FC > 1.5, p < 0.05) (Fig. 1A, Supplementary Table S1). In addition, in the TCGA database, we identified the top 500 genes that affect the prognosis of bladder cancer (Supplementary Table S2). Only 6 genes were present in both sets of data (Fig. 1B). One of them, WDR4, was associated with poor prognosis and showed much more prominent elevation in bladder cancer tissues than did the other 5 overlapping genes (Fig. 1B–D). Moreover, WDR4 had the most significant effect on overall survival (OS) and disease-free survival (DFS) (Fig. 1C, D, Supplementary Fig. S1). These data suggested that WDR4 may be associated with the malignant progression of bladder cancer.

Fig. 1: Increased expression of WDR4 correlates with progression and survival in bladder cancer.

A Label-free quantitative proteomic analysis of cancer tissues (T) and paired normal adjacent tissues (N) from 17 MIBC patients. A total of 524 genes with higher expression and 219 genes with lower expression in cancer tissues (log2 FC > 1.5, p < 0.05) were identified. B Venn diagram analysis of these 524 highly expressed genes and the top 500 genes affecting bladder cancer prognosis in the TCGA database. The table on the right shows the fold changes in the expression of the 6 genes present in both sets. Kaplan‒Meier analysis of overall survival (C) and disease-free survival (D) in bladder cancer patients in the TCGA cohort with low or high levels of WDR4. E WDR4 upregulation in bladder cancer cells and the control cell line SV-HUC-1 was verified by WB. F Western blot analysis of the WDR4 protein level in 6 pairs of human clinical bladder cancer samples. G, H Representative IHC images of WDR4 expression in T1-T4 stage cancer tissues and normal adjacent tissues (NATs) from bladder cancer patients. I, J Representative IHC images of WDR4 expression in bladder cancer tissues from patients with or without LN metastasis. Scale bar of upper panel = 200 μm, Scale bar of lower panel = 50 μm.

We next validated the expression of WDR4 in bladder cancer cells and clinical samples. Higher WDR4 levels were measured in bladder cancer cells and tissues (Fig. 1E, F). This finding was further strengthened by IHC analysis. WDR4 was upregulated in tumor tissues compared to normal adjacent tissues, and there was a correlation between the intensity of WDR4 staining and T stage (Fig. 1G, H). Moreover, higher WDR4 expression was observed in metastatic cancer tissues than in nonmetastatic cancer tissues of patients (Fig. 1I, J). Cytoplasmic WDR4 has been reported to play different roles in tumors by regulating the translation or degradation of proteins [20]. Notably, WDR4 was upregulated primarily in the nucleus (Fig. 1G–J). WDR4 expression in tumor cell nuclei has not been reported thus far. Together, these findings indicate that high expression of WDR4 is associated with the progression and metastasis of bladder cancer, probably in a manner related to the nuclear function of WDR4.

WDR4 promotes LN metastasis of bladder cancer

We investigated the effect of WDR4 on the cellular processes of migration and invasion. The expression of WDR4 was transiently silenced with three independent siRNAs in UM-UC-3 and 5637 cells (Supplementary Fig. S2A–D). The wound healing assay results showed that silencing WDR4 reduced cell migration (Fig. 2A–D). The results of Transwell assays also demonstrated that transient silencing of WDR4 significantly inhibited cell migration and invasion (Fig. 2E, F). Stable knockdown of WDR4 significantly suppressed cell migration and invasion (Supplementary Fig. S2E, G). In addition, overexpression of WDR4 increased the migration and invasion abilities of bladder cancer cells (Fig. 2G–I, Supplementary Fig. S2F).

Fig. 2: WDR4 promotes the metastasis of bladder cancer cells in vitro and in vivo.

A–F Cells were transfected with WDR4-specific siRNAs or the control siRNA. Interference with WDR4 expression significantly reduced the migration of UM-UC-3 (A, C) and 5637 (B, D) cells in the wound healing assay and the migration and invasion of UM-UC-3 (E) and 5637 (F) cells in the Transwell assays. WDR4 overexpression promoted the migration of UM-UC-3 cells in the wound healing assay (G, H) and migration and invasion in the Transwell assays (I). J Representative image of the popliteal LN metastasis mouse model. K Representative bioluminescence images of mice with LN metastases generated by WDR4 knockdown (WDR4 KD) and control knockdown (Control KD) UM-UC-3 cells. Representative image of excised popliteal LNs (L) and histogram of LN volumes in control KD and WDR4 KD mice (n = 6 per group) (M). N Hematoxylin-eosin (HE) staining of LNs from control KD and WDR4 KD mice. O Representative bioluminescence images of mice with LN metastases generated by WDR4 overexpression (WDR4) and control (Vector) UM-UC-3 cells. Representative image of excised popliteal LNs (P) and histogram of LN volumes in the vector and WDR4 groups (n = 6 per group) (Q). The data are presented as the means ± SDs, ***p < 0.001. R HE staining of LNs from the control vector and WDR4 overexpression groups. Scale bar of upper panel = 500 μm, Scale bar of lower panel = 50 μm.

LN metastasis is the main type of bladder cancer metastasis [1]. The above data suggested that WDR4 may contribute to lymphatic metastasis of bladder cancer. For further validation, we constructed a popliteal LN metastasis model, which simulated the lymphatic metastasis of bladder cancer in vivo (Fig. 2J). We stably knocked down WDR4 and overexpressed WDR4 in UM-UC-3 cells (Fig. S2E, F). To test their metastatic potential, these cells were inoculated into the footpads of nude mice. As revealed by measurement of luminescence intensity, knockdown of WDR4 significantly suppressed lymphatic metastasis (Fig. 2K), whereas overexpression of WDR4 markedly accelerated the metastasis of bladder cancer cells to LNs (Fig. 2O). The volume of popliteal LNs was significantly smaller in the WDR4-knockdown group (Fig. 2L, M) but significantly larger in the WDR4-overexpression group than in the control group (Fig. 2P–Q). These findings were validated by hematoxylin-eosin (HE) staining of lymphatic metastases (Fig. 2N, R). Together, these results showed that WDR4 promotes LN metastasis in bladder cancer.

WDR4 promotes the proliferation of bladder cancer cells

The ability to proliferate is also a vital intrinsic property of metastatic cancer cells [21]. We next investigated the effect of WDR4 on the cellular proliferation process. Transient silencing of WDR4 significantly reduced cell proliferation (Fig. 3A–G). Stable knockdown of WDR4 also significantly suppressed cell proliferation (Supplementary Fig. S2H). Moreover, overexpression of WDR4 obviously increased the proliferation of bladder cancer cells (Fig. 3H–K).

Fig. 3: WDR4 promotes the proliferation of bladder cancer cells in vitro and in vivo.

A–G Cells were transfected with WDR4-specific siRNAs and the corresponding control siRNA. Interference with WDR4 expression significantly reduced the numbers of UM-UC-3 (A, B) and 5637 (A, C) cell colonies in the colony formation assay and proliferating UM-UC-3 (D, F) and 5637 (E, G) cells in the EdU incorporation assay. WDR4 overexpression increased the number of UM-UC-3 cell colonies in the colony formation assay (H, I) and proliferating UM-UC-3 cells in the EdU incorporation assay (J, K). Representative image of subcutaneous tumors in WDR4 KD and control KD mice (L). Histogram showing the tumor volumes and weights (M). Representative image of the tumors in the WDR4 overexpression and control vector groups (N). Histogram showing the tumor volumes and weights (O). The data are presented as the means ± SDs, *p < 0.05, ***p < 0.001. P IHC staining showed the difference in Ki67 expression in tumor samples, Scale bar of upper panel = 200 μm, Scale bar of lower panel = 50 μm.

UM-UC-3 cells with stable WDR4 knockdown or WDR4 overexpression and the corresponding control cells were subcutaneously inoculated into nude mice (Supplementary Fig. S3). Knockdown of WDR4 significantly decreased tumor volume and weight (Fig. 3L, M), while overexpression of WDR4 markedly increased the volumes and weights of the tumors (Fig. 3N, O). Furthermore, immunohistochemical (IHC) analysis of xenograft tumors confirmed that WDR4 knockdown resulted in lower levels of Ki67 expression, but WDR4 overexpression resulted in higher levels of Ki67 expression (Fig. 3P). These results revealed that WDR4 promotes the proliferation of bladder cancer cells.

DDX20 interacts with WDR4

WDR4 commonly binds to specific proteins to perform its function [14]. To investigate how WDR4 regulates bladder cancer progression, we performed co-IP experiments on bladder cancer cells. Intracellular proteins that could interact with WDR4 were immunoprecipitated with specific antibodies from UM-UC-3 and 5637 cells and then explored using LC‒MS/MS analysis. A total of 251 and 281 proteins were identified in UM-UC-3 and 5637 cells, respectively. Of these proteins, 39 were present in both sets (Fig. 4A). WDR4 forms a heterodimer with METTL1 [22]. METTL1 ranked high on the list of proteins that might interact with WDR4 (Fig. 4A, Supplementary Table S3). The METTL1-WDR4 complex has been found to promote cell transformation and tumor progression in bladder cancer [23, 24]. Notably, WDR4 was found to be enriched in the nucleus in bladder cancer cells (Fig. 1G–J), and we explored its mechanism and function in the nucleus. DEAD-box helicase 20 (DDX20), a member of the DEAD-box helicase family, also ranked high on the list of proteins that may interact with WDR4 (Fig. 4A, Supplementary Table S3). In recent years, the DEAD-box helicase family has been found to be involved in transcriptional regulation in the nucleus [25]. The results of co-IP experiments confirmed the interaction between WDR4 and DDX20 (Fig. 4B, C). To determine whether DDX20 contributes to bladder cancer cell metastasis and proliferation, we knocked down DDX20 in UM-UC-3 cells (Fig. S4A). Knockdown of DDX20 significantly suppressed cell migration, invasion and proliferation (Fig. 4D–H). Subsequently, we studied the cooperative role of WDR4 and DDX20 in bladder cancer cells. Compared with that in the control group, WDR4 overexpression promoted the metastasis and proliferation of cancer cells, but simultaneously silencing DDX20 partially reversed the increases in the metastasis and proliferation of bladder cancer cells caused by WDR4 upregulation (Fig. 4I–M). Surprisingly, overexpression of WDR4 promoted the nuclear accumulation of DDX20 (Fig. 4N). Thus, WDR4 may interact with DDX20 in the nucleus to promote the progression of bladder cancer.

Fig. 4: DDX20 interacts with WDR4 and promotes the metastasis and proliferation of bladder cancer cells.

A The interacting proteins were immunoprecipitated with an anti-WDR4 antibody. Venn diagram showing the 251 and 281 proteins identified in UM-UC-3 and 5637 cells, respectively, using LC‒MS/MS. The table on the right shows the top 12 proteins to which WDR4 binds in both cell lines. B, C Co-IP and Western blot analyses showed the interaction between endogenous WDR4 and DDX20 in UM-UC-3 and 5637 cells. D Cells were transfected with DDX20-specific siRNAs and the corresponding control siRNA. Interference with DDX20 expression significantly reduced the migration and invasion of UM-UC-3 cells in the Transwell assays. Interference with DDX20 expression significantly reduced the numbers of UM-UC-3 cell colonies in the colony formation assay (E, F) and proliferating UM-UC-3 cells in the EdU incorporation assay (G, H). I–M WDR4 overexpression and control vector UM-UC-3 cells were transfected with DDX20-specific siRNAs and the corresponding control siRNA. Migration and invasion of cells were evaluated using Transwell assays. I A colony formation assay (J, K) and a EdU incorporation assay (L, M) were used to evaluate the cell proliferation ability. The data are presented as the means ± SDs, ***p < 0.001. N The location of DDX20 was determined in WDR4-overexpressing and control UM-UC-3 cells by using IF staining, Scale bar 50 μm.

ARRB2 is repressed at the transcriptional level by WDR4 and DDX20

WDR4 has been reported to regulate translation by interacting with proteins in the cytosol. However, we found that the nuclear expression was significantly higher than cytoplasmic expression of WDR4 in bladder cancer cells (Fig. 1G–J). In addition, the interacting protein DDX20 has been reported to act as a repressor of gene transcription [26,27,28,29]. The WDR4-DDX20 complex may affect the progression of bladder cancer by regulating the transcription of target genes. To identify the key molecules regulated by the complex, CUT&Tag-seq was performed to identify the respective binding DNA locations of WDR4 and DDX20 (Fig. S4B, C). Moreover, RNA-seq was performed on cells stably overexpressing WDR4 and the corresponding control cells to screen for the differentially expressed genes (Fig. 5A). The numbers of genes with downregulated mRNA expression after WDR4 overexpression, genes encoding proteins binding to WDR4 and genes encoding proteins binding to DDX20 in transcriptional regulatory regions were 632, 288 and 1041, respectively (Fig. 5B and Supplementary Tables S4–6). There were 5 genes (MTCO1P12, MTND5P28, PRKACA, TSEN34, ARRB2) present in all three sets (Fig. 5B). In addition, ARRB2 has been reported to inhibit the growth and progression of bladder cancer and increase the response to chemotherapy [30]. Both WDR4 and DDX20 can bind the transcriptional regulatory region of ARRB2 (Fig. S4D). ARRB2 may act as a tumor suppressor regulated by the WDR4-DDX20 complex in bladder cancer. Next, knockdown of WDR4 was found to increase the expression of ARRB2 (Fig. 5C, E, Supplementary Fig. S3), whereas overexpression of WDR4 in cells markedly suppressed the expression of ARRB2 (Fig. 5D, F, Supplementary Fig. S3). In addition, transient knockdown of DDX20 increased the expression of ARRB2 in cells (Fig. 5G, H). Moreover, we observed that ARRB2 knockdown enhanced cell migration, invasion and proliferation (Fig. 5I–M, Supplementary Fig. S4G–H). Together, these results confirmed that ARRB2 may be transcriptionally regulated by WDR4 and DDX20. Subsequently, we examined whether ARRB2 could reverse the effects of WDR4 in bladder cancer cells. WDR4 knockdown inhibited the metastasis and proliferation of cancer cells compared with that of the control cells, and simultaneous silencing of ARRB2 significantly restored the metastasis and proliferation of bladder cancer cells (Fig. 5N–R).

Fig. 5: WDR4 and DDX20 inhibit ARRB2 expression.

A RNA sequencing analysis of UM-UC-3 cells with stable WDR4 overexpression cells and control UM-UC-3 cells. A total of 451 genes with higher expression and 632 genes with lower expression in WDR4-overexpressing cells were identified. B CUT&Tag-seq was used to identify the transcriptional regulatory regions to which WDR4 and DDX20 proteins bound. The Venn diagram shows the number of genes bound by WDR4 and DDX20 and the number of genes downregulated after WDR4 overexpression in UM-UC-3 cells. The table on the right shows the 5 genes present in all three datasets. The mRNA levels of ARRB2 in WDR4 KD and control KD cells (C) and in WDR4 overexpression (WDR4) and control (Vector) cells (D) were measured by RT‒qPCR. The protein levels of ARRB2 in WDR4 KD and control KD cells (E) and WDR4 overexpression (WDR4) and control (Vector) cells (F) were measured by WB. UM-UC-3 cells were transfected with DDX20-specific siRNAs and the corresponding control siRNA. The mRNA levels of ARRB2 were measured by RT‒qPCR (G). The protein levels of ARRB2 were measured by WB (H). I Cells were transfected with ARRB2-specific siRNAs and the corresponding control siRNA. Interference with ARRB2 expression significantly increased the migration and invasion of UM-UC-3 cells in the Transwell assays. Interference with ARRB2 expression significantly increased the numbers of UM-UC-3 cell colonies in the colony formation assay (J, K) and proliferating UM-UC-3 cells in the EdU incorporation assay (L, M). N WDR4 KD and Control KD UM-UC-3 cells were transfected with ARRB2-specific siRNAs and the corresponding control siRNA. Migration and invasion of cells were evaluated using Transwell assays. A colony formation assay (O, P) and an EdU incorporation assay (Q, R) were used to evaluate cell proliferation. The data are presented as the means ± SDs, *p < 0.05, **p < 0.01, ***p < 0.001.

The WDR4-DDX20 complex inhibits Egr1-promoted transcription of ARRB2

Although WDR4 and DDX20 can bind to the transcriptional regulatory region of ARRB2, neither is an ARRB2 transcription factor. We hypothesized that the WDR4-DDX20 complex may affect the function of specific transcription factors to suppress ARRB2 expression. Transcription factors upstream of ARRB2 were predicted. The results of a yeast two-hybrid assay revealed a possible interaction between DDX20 and Early growth response 1 (Egr1) [29]. Interestingly, the transcription factor Egr1 has 3 binding sites in the ARRB2 promoter (Fig. 6A). These results suggested that the WDR4-DDX20 complex may regulate ARRB2 transcription by inhibiting the function of Egr1. Knockdown of Egr1 markedly decreased ARRB2 expression, whereas overexpression of Egr1 in cells dramatically increased the expression of ARRB2 (Fig. 6B, C, Supplementary Fig. S4E, F). Subsequently, wild-type and mutant plasmids containing these three unique promoter sites were cotransfected individually with the Egr1-overexpressing plasmid in 293 T cells. By the reporter assay, the key binding site of Egr1 in the ARRB2 promoter was determined to be M1 (Fig. 6A, D).

Fig. 6: WDR4 regulates bladder cancer progression by transcriptional repression of ARRB2.

A Egr1-binding sites in the promoter of the ARRB2 gene; the three corresponding mutation sites are marked with red vertical lines. B After Egr1 knockdown or Egr1 overexpression in UM-UC-3 cells, the mRNA levels of ARRB2 were measured by RT‒qPCR. C The protein levels of ARRB2 were measured by WB. D The Egr1 overexpressing plasmid and the reporter plasmids containing the wild-type (WT) or mutated (MUT) of Egr1 binding sequence in the ARRB2 promoter were cotransfected into 293 T cells. A luciferase assay was performed to detect changes in relative luciferase activity. E Co-IP showed the interaction between endogenous WDR4 and Egr1 in UM-UC-3 and 5637 cells. F Co-IP showed the interaction between endogenous DDX20 and Egr1 in UM-UC-3 and 5637 cells. G IF staining and confocal observation of Egr1, DDX20 and WDR4 in tissue sections from bladder cancer patients. Scale bar 20 μm. H Co-IP showed the interaction between endogenous DDX20, WDR4 and Egr1 in WDR4 KD and Control KD UM-UC-3 cells. The data are presented as the means ± SDs, ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001.

As a scaffold protein, WDR4 promotes the binding of different proteins [17]. Furthermore, the results of co-IP experiments revealed that both WDR4 and DDX20 pulled down Egr1 from the cell lysate, suggesting the existence of a complex consisting of WDR4, DDX20 and Egr1 (Fig. 6E, F). Immunofluorescence analysis by confocal fluorescence microscopy also confirmed that WDR4 colocalized with DDX20 and Egr1 in the nucleus (Fig. 6G). Notably, the association between DDX20 and Egr1 was weaker when WDR4 was knocked down (Fig. 6H). These results suggested that WDR4 can promote the interaction between DDX20 and Egr1. We also found that WDR4 overexpression clearly caused nuclear translocation of DDX20 (Fig. 4N). Given the above results, we hypothesized that WDR4 may mediate the assembly of the DDX20-WDR4-Egr1 complex by recruiting DDX20 to the transcriptional regulatory region of the ARRB2 gene, thereby inhibiting the Egr1-facilitated transcriptional expression of ARRB2.

WDR4 expression positively correlates with DDX20 expression in bladder cancer and predicts LN metastasis

To investigate the clinical significance of WDR4 and DDX20, we assembled a cohort of 77 bladder cancer tissues and 30 normal adjacent tissues. To evaluate the relationship between WDR4 and DDX20 expression, we performed IHC analysis. Staining of two tissue sections from the same patient with either an anti-WDR4 or an anti-DDX20 antibody showed increased regional WDR4 and DDX20 levels in cancer tissues and a positive correlation between WDR4 and DDX20 expression (Fig. 7A, B). Histological staining showed that WDR4 and DDX20 were both upregulated in cancer tissues compared with normal tissues (Fig. 7C, D). In addition, bladder cancer tissues of patients with LN metastasis displayed markedly higher WDR4 and DDX20 levels than those without LN metastasis (Fig. 7C, D). Furthermore, to assess the predictive value of WDR4 for LN metastasis of bladder cancer, we analyzed clinical information from patients. Univariate logistic regression analysis was performed. The results revealed that the expression of either WDR4 or DDX20, similar to the correlations with T stage, was an independent predictor of LN metastasis in bladder cancer (Fig. 7E). These data confirmed the positive correlation between WDR4 and DDX20 expression and the role of these proteins in regulating LN metastasis, as shown in the schematic (Fig. 7F). The WDR4 expression level may serve as an independent prognostic factor for LN metastasis in bladder cancer patients.

Fig. 7: WDR4 expression positively correlates with DDX20 expression and predicts LN metastasis of bladder cancer.

A Representative images of WDR4 and DDX20 IHC staining in bladder cancer tissues, Scale bar of upper panel = 500 μm, Scale bar of lower panel = 50 μm. B Pearson correlation between WDR4 and DDX20 levels in clinical bladder cancer tissues. AOD of WDR4 (C) and DDX20 (D) in normal bladder tissues and bladder cancer tissues from patients with and without LN metastasis. E Univariate analyses to determine risk factors associated with LN metastasis in bladder cancer patients (n = 77). F Schematic illustration of the mechanism of WDR4 in promoting LN metastasis of bladder cancer.