The intracellular domain of UNC5B facilities proliferation and metastasis of bladder cancer cells

Abstract The intracellular domain of UNC5B contains both death domain and caspase‐3 cleavage site, and is regarded as a functional domain that mediates apoptosis. However, in our previous studies, we found that the death domain of UNC5B in bladder cancer cells could not be activated to promote apoptosis. In this study, different UNC5B truncates (residue 399‐945, residue 412‐945) were created to explore whether the caspase‐3 cleavage site (site 412), as another potential functional domain of its intracellular portion, could be activated to induce apoptosis in bladder cancer cells. Using mass spectrometry, we acquired a comprehensive and detailed identification of differentially expressed proteins by overexpressing UNC5B and its truncates. Protein‐protein‐interaction (PPI) network analysis was also applied to investigate the aggregation of related proteins and predict the functional changes. EDU assay, apoptosis, xenograft tumour implantation, migration, invasion and tumour metastasis were performed to comprehensively identify the effects of UNC5B truncates on bladder cancer cells. We demonstrate that the intracellular domain of UNC5B promotes cell proliferation in vitro and tumour formation in vivo, by binding to a large number of ribosomal proteins. The overexpression of intracellular domain also facilitates cells to migrate, invade and metastasize by interacting with fibronectin, beta‐catenin and vimentin. In addition, we reveal that overexpressing the intracellular domain of UNC5B cannot bind or activate cleaved caspase‐3 to trigger apoptosis in bladder cancer cells.

In this study, UNC5B truncates (residue 399-945, residue 412-945) were created and transfected to explore whether the caspase-3 cleavage site of UNC5B, as another potential functional domain of its intracellular portion, could be activated to induce apoptosis in BC cells. Using mass spectrometry, we acquired a comprehensive and detailed identification of differentially expressed proteins by overexpressing UNC5B and its truncates. Proteinprotein-interaction (PPI) network analysis was also applied to investigate the aggregation of related proteins and predict the functional changes. EDU assay, apoptosis, xenograft tumour implantation, migration, invasion and tumour metastasis were performed to comprehensively identify the effects of UNC5B truncates on BC cells.
We demonstrate that the intracellular domain of UNC5B promotes cell proliferation and tumour formation, by binding to a large number of ribosomal proteins. The overexpression of intracellular domain also facilitates BC cells to migrate, invade and metastasize by interacting with fibronectin, beta-catenin and vimentin. In addition, we reveal that overexpressing the intracellular domain of UNC5B cannot bind or activate cleaved caspase-3 to trigger apoptosis in bladder cancer cells. Antibodies against UNC5B (ab104871, ab54430), caspase-3 (ab179517), mutant P53 (ab32049) and BCL-2 (ab32124) were purchased from Abcam. Antibodies against cleaved PARP (D64E10), S6 Ribosomal Protein (5G10), UNC5B (D9M7Z) and beta-actin (3700S) were obtained from Cell Signaling Technology (CST).
Universal Magnetic Co-IP Kit (54002, Active Motif) was purchased from Active Motif Co. Ltd. Fast silver staining kit (P0017S, Beyotime) was purchased from Beyotime Co. Ltd.

| Western blot analysis
Cells were washed three times with ice-cold PBS, homogenized at 4°C in 10 volumes (w/v) of RIPA lysis buffer (Beyotime), and 1 mM PMSF mixture for 30 minutes. The lysates were gathered after centrifuging for 30 minutes (13 000 rpm) at 4°C. After the normalization and denaturation of each sample, 50 µg of total protein was loaded and separated with 12% SDS-PAGE, and then transferred onto PVDF membrane (Millipore). After blocking with 5% skimmed milk powder in TBST (TBS containing 0.1% Tween-20), the membranes were incubated with primary antibody (1:1000) overnight at 4°C. After being washed, the membranes were incubated in secondary antibody (1:3000-1:5000) for two hours at 37°C. The membranes were then washed three times.
The immunoreactive bands were examined by a Microchemi 4.2 device (Bio-Rad) using an enhanced chemiluminescent (ECL) reagent kit.

| Co-immunoprecipitation
In general, we mixed a complete whole-cell lysis buffer of protease inhibitor cocktail, deacetylase inhibitor, phosphatase inhibitors, PMSF and cell lysis buffer to extract protein from each cell sample.
Centrifuge the mixture of cell suspension for at least 10 minutes at 4°C (1500 rpm). We mixed protease inhibitor cocktail, deacetylase inhibitor, phosphatase inhibitors, PMSF and cell lysis buffer to make a complete whole-cell lysis buffer,and used this lysis buffer to extract protein from each cell sample. We then added the lysis buffer to cell samples and centrifuged the mixture of cell suspension for 10 minutes at 4°C (1500 rpm). 500 µg whole-cell extract and 5ug antibody was subsequently combined into a pre-chilled tube. The tube was then incubated at 4°C on a rotator overnight. 25 µL Protein G magnetic beads were then added to each tube. The mixture of Protein G beads, whole-cell extract and antibody was incubated for one hour at 4°C on a rotator. We centrifuged the tube at 1500 Crpm for one minute and discarded the supernatant. We then washed beads 4 times with the cell lysis buffer. Finally, we resuspended bead pellet in 20 µl of 2X Reducing Loading Buffer. WB, fast silver staining and mass spectrometry were then applied for further investigation.
(APT) for further investigation.

| Mass spectrometry in proteomics and bioinformatics analysis
Shanghai Applied Protein Technology Co. Ltd. provided bio-informatics analysis on mass spectrometry in proteomics. The differently expressed proteins between groups were identified by a label-free proteomics method. GO analysis, map analysis of KEGG signalling pathway and PPI (protein-protein-interaction) network analysis were all included to predict the molecular process and signalling pathways. Cytoscape software (version 3.6.1) was applied to analyse and predict protein binding and interaction.

| Cell culture and transfection
BC cell lines UMUC3, 5637 and T24; prostate cancer cell 22RV-1; renal cancer cell 786-0 were purchased from the Chinese Academy of Sciences Cell Bank (CASCB) and were cultured in RPMI 1640 medium (HyClone). All cells were incubated in an incubator with 5% CO 2 .
We transfected UNC5B and UNC5B-truncates to cells by lentivirus for at least 60 hours. The stably transfected cells were established by using puromycin for at least 3 weeks. The successful transfection was verified by PCR, WB, flow cytometry and immunofluorescence.

| Quantitative real-time polymerase chain reaction
Trizol reagent (Invitrogen) was used to extract total RNA from cells.
By using PrimeScript™ RT Master Mix (Takara Biotechnology), RNA was reverse transcribed into first-strand cDNA. We then performed real-time qPCR to detect the mRNA expression of UNC5B and betaactin genes using SYBR® Premix Ex Taq™ (Tli RNaseH Plus; Takara Biotechnology Co. Ltd.) and a Thermal Cycler Dice™ Real Time TP800 system (Takara). Table 3 indicated the related primer sequences.

| EDU assay
BeyoClick™EdU-488 cell proliferation detection kit (C0071S, Beyotime) was used for EDU assay. In brief, the serum-free 1640 medium was used to prepare 2X EDU working solution. The preheated working solution was added to cells and incubated for 2 hours. Then, we added

| Immunofluorescence
5 × 10 4 cells were seeded and cultured for 48 hours, and washed with PBS. 300 μL 4% paraformaldehyde was used to fix cells. Subsequently, 300 μL triton X-100 was applied to increase the permeability of the cell membrane. 300 μL BSA buffer was added for blocking for 30 minutes.
After the BSA was removed, the cells were incubated at 4°C with primary antibody (1:100 diluted) overnight. Fluorescent labelled second antibody was added to cells at room temperature and incubated with shading for 1 hour, and then, DAPI was added to stain the nucleus.
Fluorescence microscope was used to observe and record the images.

| Xenograft tumour model
This study was conducted according to the Medical Laboratory Animal Welfare and Ethics Committee of China Medical University.
We purchased 10 female BALB/c nude mice of 4-6 weeks old from Beijing Vital River Experimental Animal Technology Co. Ltd and randomly assigned mice into two groups. We injected 412'5637 and 399'5637 cells separately into the right and left armpits of mice in the first group, NC-5637 and UN-5637 cells into the second group.
We killed the mice 56 days after the subcutaneous injection. We used the equation (volume = length×width 2 /2) to estimate the tumour size.

Cell migration
We added 500 µL 1640 medium with 10% serum to a 24-well plate

| Tail vein injection
Prepare a 50 mL centrifugal tube and hollow out the centre of the tube cap to form a round hole with a diameter of 0.6 cm. Fill the centrifuge tube with cotton, place the mice in it and straighten the tail. Wipe the tail with alcohol to dilate the tail vein, fix the mouse tail with the left hand and inject the PBS mixed with different cells into the vein. The total amount of injected cells is 2 × 10 6 , and the bubbles in the mixture are emptied before injection. The injection dose should not be more than 1.2 mL or <0.8 mL, lest the mice die of heart failure or pulmonary infarction. Following the above process, we injected 16 BALB/c nude mice and killed the animals 8 weeks after the injection.

| The transfection of the full-length UNC5B and its truncates to BC cells
In our previous study on BC cells, 20  Immunofluorescence ( Figure 1I-L) and sequence analysis (Table 1 and 2) verified the successful induction of the intracellular domains of UNC5B in both cells.

| Full-length UNC5B mildly inhibited cell proliferation and mediated apoptosis, while UNC5B truncates promoted cell proliferation
Through EDU assay (Figure 2A,C), we observed that a greater proportion of cells were examined at the stage of proliferation by the overexpression of UNC5B truncates both in 5637 and T24 cells.
The percentage of cells in the proliferative phase in 412′5637 and 412′T24 cells was 5% higher than that of the NC cells. Comparatively, the percentage of cells in the proliferative phase in the 399′5637 and 399′T24 cells was 2.4% and 3.3% higher than that of the NC cells, respectively. By contrast, the overexpression of full-length the expression of caspase-3 or cleaved caspase-3 was found by CO-IP in both groups ( Figure 3F). On the contrary, more binding of S6 ribosomal protein was identified to interact with UNC5B truncates and full-length UNC5B. Through the normalization of whole-cell CO-IP lysates ( Figure 3G), we found that the overexpressed UNC5B and its truncates had stronger binding to S6 ribosomal proteins, especially in 399′5637, 412′5637, 399′T24 and 412′T24 cells. The proliferation-promoting effect of ribosomal protein was also confirmed by mice xenograft models ( Figure 3H,I). We observed that the overexpression of UNC5B truncates in 5637 cells significantly increased the tumour volume on mice. Comparatively, the tumour volume of 412′5637 cells was even larger than that of 399′5637 cells, which was consistent with the previous result in EDU assay and PPI analysis. By contrast, full-length UNC5B slightly decreased the tumour volume by largely mediating cell cycle arrest, which was identical to our previous study. 30

| More binding of fibronectin, betacatenin and vimentin was identified by overexpressing UNC5B and UNC5B truncates
Through the mass spectrometry, a variety of epithelial-mesenchymal transition (EMT) related proteins were co-immunoprecipitated and identified, including fibronectin, beta-catenin, vimentin, myosin, plectin, capping protein, hornerin, desmoplakin and. PPI analysis was applied to reveal the interaction of these differentially expressed proteins among groups ( Figure 4A-F). According to the relative protein expression of mass spectrometry, fibronectin, beta-catenin and vimentin were examined to reveal the regulatory mechanism.
WB analysis indicated that the expression of fibronectin and betacatenin was increased in 412′5637 cells ( Figure 4G). The expression fibronectin, beta-catenin and vimentin to UNC5B truncates and fulllength UNC5B were also found to be up-regulated by CO-IP test ( Figure 4H).

The transfection of UNC5B truncates and full-length UNC5B
similarly increased the expression of fibronectin, beta-catenin and vimentin ( Figure 4I) in T24 cells. The overall expression of N-CA was also slightly up-regulated in T24 transfected cells. Similarly, more binding of vimentin and beta-catenin were also co-immunoprecipitated in 399'T24 and UN-T24 cells ( Figure 4J). Through the normalization of whole-cell CO-IP lysates ( Figure 4K,L), more binding of fibronectin, beta-catenin and vimentin to UNC5B and UNC5B truncates was examined.

| Overexpression of UNC5B and UNC5B truncates promoted cell migration in vitro and tumour metastasis in vivo
The

| D ISCUSS I ON
As one of the receptors of Netrin-1 family, 31-33 UNC5B binds to Netrin-1 through two immunoglobulin-like domains of extracellular portion of UNC5B. 22,34 The dependence receptor theory suggests that UNC5B triggers apoptosis in the absence of Netrin-1 or with the overexpression of UNC5B, which is mainly applicable in neuronal cells. [31][32][33]35 On the contrary, in the presence of Netrin-1, UNC5B is regulated to promote the proliferation and migration in rectal cancer. 36,37 The mechanism of apoptosis mediated by UNC5B is considered in two ways: one is to interact with Death-Associated Protein Kinase (DAPK) through death domain of UNC5B 25,26 ; the other is to activate caspase-3 by participating in p53 apoptosis pathway. 18,28,29 However, in our previous study, we found that the cotransfection of UNC5B and DAPK in BC cells could not bind to each other. The non-binding of UNC5B and DAPK was attributed to the tight closed structure between the intracellular domain of ZU5-UPA-DD, which was supported by Wang et al. 38 In our recent work, we found that full-length UNC5B bound to cleaved caspase-3 to significantly induce apoptosis in 5637 cells transfected with DAPK ( Figure S2A). As the caspase-3 site was not in the ZU5-UPA-DD supermodule structurally ( Figure 1H), we had expected that apoptosis could be greatly mediated by the overex- Similarly using flag-labelled lentivirus, we transfected UNC5B to Renal carcinoma cells (786-0, ACHN), prostate cancer cells (22RV-1) and BC cells (UMUC3) (Figure S2B), and examined the apoptosis. We found that only the BC cells (UMUC3) transfected with UNC5B had the effect of promoting apoptosis (the late apoptosis rate was increased by 5.6%) ( Figure S2C). Through previous mass spectrometry and bio-informatics analysis, no direct interaction of Netrin-1 and UNC5B was found in BC cells. However, we found that the pro-apoptotic effect of UNC5B was reversed by transfecting Netrin-1 to UN-5637 and UN-T24 cells ( Figure S2D-F From this aspect, it seems that we can speculate the mechanism of UNC5B on BC cells. We believe that UNC5B, as a membrane protein, is very low expressed in BC cells; thus, its biological effect is mainly regulated by its binding proteins. Through mass spectrometry, we found that the most binding proteins by overexpressing UNC5B and its truncates are EMT-related proteins, with the effect of promoting cell invasion and metastasis. On the other hand, although full-length UNC5B binds to more protein groups (such as ribosomal proteins), the overall effect on proliferation is inhibited. Through our previous study, we confirmed that fulllength UNC5B combined with a large number of inhibitors to mediate cell cycle arrest and apoptosis, in which its binding to p53 and CDC14A played an important role. 30 However, the binding ability of UNC5B truncates is much lower than that of full-length UNC5B, which hardly binds to protein groups except for ribosome and histone, thus promoting cell proliferation. To sum up, the biological effect of UNC5B on BC is determined by its binding proteins. Because of the diversity of UNC5B binding protein and its potential for metastasis of BC cells, it is not considered as a typical tumour suppressor gene in BC.
In this study, we reveal that overexpressing the intracellular domains of UNC5B cannot bind or activate cleaved caspase-3 to trigger apoptosis. On the contrary, the intracellular domains of UNC5B promote cell proliferation in vitro and tumour formation in vivo, by binding a large number of ribosomal proteins. They also facilitate BC cells to migrate, invade and metastasize by interacting with EMT-related proteins, such as fibronectin, beta-catenin and vimentin.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets used and analysed in this study are available from the corresponding author upon reasonable request.