Enhancement of Farnesoid X Receptor Inhibits Migration, Adhesion and Angiogenesis through Proteasome Degradation and VEGF Reduction in Bladder Cancers

Round 1

Reviewer 1 Report

Dear all,

I went through the revised version of this manuscript. I still think that some of the results are overinterpreted and I feel that caveats and/or limitations were not really discussed in the main text. However, this version of the paper has really improved compared to the previous one, and I appreciate the additional work by the authors.
Therefore, I think this study can now be recommended for publication in your Journal.

This being said, there are still some typos and/or sentences that require improvement. A few examples:

    line 50: " studies shown that"
    line 108: " Doxycycline was no effect on cell viability"
    line 110: " we selected doxycycline for 48 and 72 hours groups"

Author Response

Reply to Reviewer 1’s comments

Dear all,

I went through the revised version of this manuscript. I still think that some of the results are overinterpreted and I feel that caveats and/or limitations were not really discussed in the main text. However, this version of the paper has really improved compared to the previous one, and I appreciate the additional work by the authors.
Therefore, I think this study can now be recommended for publication in your Journal.

This being said, there are still some typos and/or sentences that require improvement. A few examples:

1. Reply: Thank you for all the valuable comments. We discussed the limitations in our manuscript.

Page 11, line 348

In our studies, FXR overexpression inhibited bladder cancer cells migration, adhesion, angiogenesis in human bladder cancer cells TSGH8301 and T24 through proteasome degradation pathway. However, there are still some limitations need to be improved. First, although the impressive effects of FXR on bladder cancer cells were found in our in vitro experiments, the effects of FXR in animal models still an important issue needs to be solved. Second, the invasive ability and its related proteins MMP2 and MMP9 cannot be reversed in both MG132 and NH4Cl. The relationship between FXR overexpression and invasion inhibition in muscle invasive human bladder cancer T24 also needs further investigation. Third, FXR agonist, obeticholic acid, has been demonstrated with promising clinical results in the treatment of liver disorders such as primary biliary cirrhosis, primary sclerosing cholangitis, and nonalcoholic steatohepatitis [1]. However, the effects of FXR agonist on bladder cancer is still unknown and needs additional investigation.

2. line 50: " studies shown that"

Reply: Thank you for the correction. We corrected the sentence as “Previous studies have shown that…”.

3. line 108: " Doxycycline was no effect on cell viability"

Reply: Thank you for the correction. We corrected the sentence as “doxycycline had no effect on cell viability…”.

4. line 110: " we selected doxycycline for 48 and 72 hours groups"

Reply: Thank you for the correction. We corrected the sentence as “we selected doxycycline induced for 48 and 72 h in the following experiments”.

Reference

1. Gege, C., et al., Nonsteroidal FXR ligands: current status and clinical applications. Bile Acids and Their Receptors, 2019: p. 167-205.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript reported the role of Farnesoid X Receptor (FXR) in bladder cancer. High- and low-grade cell line were applied to FXR overexpression system to investigate the relationship between FXR and cancer cell ability such as migration and invasion. This study further discovered that VEGF reduction is involved.

However, I still have two concerns about this manuscript

1. The figures are so blur that is hard to read the words in it, like Figure 1A, 1B, 1C. I would highly recommend replacing image into a clearer version.
2. The images of wound healing had high background, so it is hard to see the boarder of the cells. is it possible to replace the images like Figure 4A. TSGH8301?

Author Response

Please see the attachment.

Reply to Reviewer 2’s comments

The manuscript reported the role of Farnesoid X Receptor (FXR) in bladder cancer. High- and low-grade cell line were applied to FXR overexpression system to investigate the relationship between FXR and cancer cell ability such as migration and invasion. This study further discovered that VEGF reduction is involved.

 However, I still have two concerns about this manuscript

1. The figures are so blur that is hard to read the words in it, like Figure 1A, 1B, 1C. I would highly recommend replacing image into a clearer version.

Reply: Thank you for the comment. We have improved the Figure 1A, 1B, and 1C.

2. The images of wound healing had high background, so it is hard to see the boarder of the cells. is it possible to replace the images like Figure 4A. TSGH8301?

Reply: Thank you for the comment. We replaced another set of wound healing result in TSGH8301 in Figure 4A.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

In their paper, Chien-Rui Lai et al characterized two doxy-inducible FXR-overexpressing cell line models of bladder cancer (BLCA), based on the TSGH8310 and T24 cell lines. First, authors analyzed publicly available data showing that high FXR expression tracked with better survival in BLCA. Next, authors found that low-grade cell lines (n=2) were expressing FXR at higher 
levels as compared to a high-grade derived BLCA cell line (n=1). Authors overexpressed FXR in T24, as well as in TSGH8310 (why? FXR downregulation would have been more appropriate here!). These FXR doxycyline-inducible cell lines were further characterized. Doxy increased both RNA and protein levels of FXR in both cell lines and affected viability of T24 but not TSGH8310. Doxy treatment affected migration of both cell lines, MMP activity in the T24 FXR-O model (what about TSGH8310? Same effect?), and reduced VEGF-A expression (in both TSGH8310 and T24). Unfortunately, authors failed to include a doxy-treated control to account for the known effects of doxy on viability, migration and MMP activity. Therefore, it is impossible to discriminate between doxy-induced effects and FXR-induced effects. Also, WB quantifications seem inaccurate (see some of the points listed below), resulting in over-interpretation or wrong interpretation of the data. In the end, I don't feel authors provided sufficient data to support their conclusions. 

Authors may want to address the following points in order to improve the quality of this work.

- Figure 1A/B: how were the thresholds selected? what is the average expression (median CPM) in the high vs. low group? Authors should mention the significance of the results (if any). Specifically, the NR0B2 Kaplan-Meyer seems non-significant (authors claimed otherwise in the text).

- Figure 1C: I assume that normal is indeed "adjacent normal" which is very different than bladder tissue from healthy individuals. Authors should clarify.
While I saw similar analyses in the literature, I think you should only focus on available pairs and run a paired analysis here.  What is the distribution of changes in the the FXR signal if only paired samples (pairs of cancer + adjacent normal) are considered?

- Figure 1D: only 3 cell lines were analyzed. The sentence "FXR may affect the malignancy and survival in bladder cancer cells" seens an overinterpretation.

- Section 2.2 starts without authors mentioning that they set up a doxy-inducible overexpression system. The initial sentence of this section should be rewritten to discuss the adoption of an overexpression system. Also, a materials & methods section describing how the system was established is missing. Specifically, I do not understand if CTRL means: 1) viral-vector-transduced cells but NOT treated w/ doxycicline; or 2) doxycicline-treated parental cells; or control-viral-vector-(empty)-transduced cells treated w/ doxycicline; it is unclear if authors selected clones or they used transduced pools for the experiments. This is crucial for the interpretation of many of the following experiments. 

- It has been shown that doxycylin can affect viability and colony formation of selected cancer cell lines. Since approrpiate controls were not included or at very least not discussed (for example, a cell line transduced with a doxy-inducible empty/GFP/reporter construct), it is not possible to know if T24 reduction in MTT/colony formation assays was due to doxy or the induction of FXR. Authors should look at the following paper by Zhang L et al, 2016 [10.1080/15384101.2016.1241929]. Doxycycline inhibits invasion, migration, and epithelial-to-mesenchymal transition of selected breast cancer cell lines.
For example, 7.13 μM Doxy was the IC50 for viability of the breast cancer MDA-MB-468 cell line (authors here used about 2 μM and reported a 25% reduction, which is in the same ballpark). Also note that Doxy was reported to inhibit MMPs (see https://pubchem.ncbi.nlm.nih.gov/compound/Doxycycline-hydrochloride)

- How many replicates were performed for the colony-formation assays or other assays described in the manuscript? I could not find that info.

- I do not agree with the interpretation of the western blots shown in figure 2C. The quantification of p-FAK is definitely wrong. Images clearly show some reduction for TSGH8301 (I agree with author's interpretation here) but there is not equivalent reduction in the bands from T24. Raw images show a bubble in the 48h band. Likewise, the bands of p-MLC are not homogenous, and it is impossible to quantify nor detect any signal reduction there. If the antibodies are not suitable for a high-quality WB or it is not possible to load more lysate, maybe authors should consider a more quantitative method such as "Jess" by ProteinSimple. Overall, despite a reduction in integrin-b1 and accumulation of FXR, I believe nothing else changes in T24. 

- I don't get what is the rationale of inhibiting cells with MG132 (please, discuss in the text before presenting results). Moreover, I don't agree with author's conclusions about NH4Cl (Figure 3C and relative quantification). I see no real difference in signal strength for integrin-b1 when I look at 
FXR-O + MG132 or FXR-O + NH4Cl. How did authors end up with a quantification showing the same level as FXR-O? There must be an error! Also for other proteins, some quantifications make no sense to me. In the end, it looks to me that NH4Cl has the same effect of MG132 in TSGH8301. 

- Fig 3A, T24 panel. Images are not too clear, but it looks like both CTRL and FXR-O are way more dense at 0h as compared to 6h. Is it an image problem or are cells dying / detaching from the plate? It's tough to draw conclusions based on the presented data. 

- Figure 4. Doxy is known to inhibit MMPs. Therefore, if the CTL samples are the only ones that were NOT treated with doxy, there is a confounding effect that makes it impossible to understand the effect of Doxy vs the effect of FXR-overexpression. 

- Figure 5D. HIF-1alpha levels are typically quasi-undetectable unless the experiment is conducted in hypoxic settings or in presence of MG132 (but authors didn't mention this). HIF-1alpha quantification in normoxia is usually unreliable/irrelevant. 

Author Response

Because figures could not be shown here, please see the attachment.

Reviewer 1

In their paper, Chien-Rui Lai et al characterized two doxy-inducible FXR-overexpressing cell line models of bladder cancer (BLCA), based on the TSGH8310 and T24 cell lines. First, authors analyzed publicly available data showing that high FXR expression tracked with better survival in BLCA. Next, authors found that low-grade cell lines (n=2) were expressing FXR at higher levels as compared to a high-grade derived BLCA cell line (n=1). Authors overexpressed FXR in T24, as well as in TSGH8310 (why? FXR downregulation would have been more appropriate here!). These FXR doxycyline-inducible cell lines were further characterized. Doxy increased both RNA and protein levels of FXR in both cell lines and affected viability of T24 but not TSGH8310. Doxy treatment affected migration of both cell lines, MMP activity in the T24 FXR-O model (what about TSGH8310? Same effect?), and reduced VEGF-A expression (in both TSGH8310 and T24). Unfortunately, authors failed to include a doxy-treated control to account for the known effects of doxy on viability, migration and MMP activity. Therefore, it is impossible to discriminate between doxy-induced effects and FXR-induced effects. Also, WB quantifications seem inaccurate (see some of the points listed below), resulting in over-interpretation or wrong interpretation of the data. In the end, I don't feel authors provided sufficient data to support their conclusions. 

Reply: Thank you for reviewing our manuscript and giving us various valuable comments.

• We performed the effect of FXR knockdown on cell survival and migration. After transfected with siFXR, the viability was increased in RT4 cells and the migration ability was enhanced in both RT4 and TSGH8301 cells (Supplementary Figure 4, as shown below).
• As a non-muscle invasive bladder cancer, MMP2 and MMP9 were nearly undetectable in TSGH8301 cells. Therefore, the invasive ability was only performed in T24 cells.
• Indeed, we appreciated for pointing out the effect of doxycycline. We evaluated the effect of doxycycline on survival, migration and invasion in un-transfected T24 cells. As in our results, doxycycline did not affect cell survival, migration, and invasion (Supplementary Figure 2, as shown below).
• We improved the western blots and quantification as suggested.

Authors may want to address the following points in order to improve the quality of this work.

- Figure 1A/B: how were the thresholds selected? what is the average expression (median CPM) in the high vs. low group? Authors should mention the significance of the results (if any). Specifically, the NR0B2 Kaplan-Meyer seems non-significant (authors claimed otherwise in the text).

Reply: Thank you for the comment. Median Group Cutoff was chosen in overall survival rate analysis. However, the average expression (median CPM) in the high vs. low group were not mentioned in this website. We apologized for the mistake. Disease Free Survival rate of NR1H4 was replaced in Figure 1B.

Page 2, line 92

As shown in figure 1A and B, the relationship between the overall and disease-free survival and the expression of NR1H4 (FXR) were analyzed by Gene Expression Profiling Interactive Analysis (GEPIA). High expression of FXR in bladder cancer patient groups resulted in a higher overall and disease-free survival rate than those in low expression.

- Figure 1C: I assume that normal is indeed "adjacent normal" which is very different than bladder tissue from healthy individuals. Authors should clarify.
While I saw similar analyses in the literature, I think you should only focus on available pairs and run a paired analysis here. What is the distribution of changes in the FXR signal if only paired samples (pairs of cancer + adjacent normal) are considered ?

Reply: Thank you for the correction. We corrected "normal" to "adjacent normal" in the manuscript. The adjacent normal tissue groups showed higher expression of FXR than those in bladder cancer tissue groups (Figure 1C). These data were collected from GEPIA online website which based on TCGA database. We also performed FXR protein expression in bladder stoma and bladder cancer cell lines by Western blot. As shown below, the FXR showed higher expression in human bladder stroma fibroblast cell lysate (HBdSF, bought from ScienCell) and low-grade TSGH8301 than in high-grade T24 and J82 cells.

Page 2, line 96

Moreover, in figure 1C, the adjacent normal tissue groups (N = 404) had higher expression of FXR than those in the bladder cancer tissue groups (T = 28).

- Figure 1D: only 3 cell lines were analyzed. The sentence "FXR may affect the malignancy and survival in bladder cancer cells" seens an overinterpretation.

Reply: Thank you for the comment. We corrected the sentence. According to GEPIA database, FXR expression is correlated to overall and disease-free survival in bladder cancer patients. Besides, high-grade T24 cells showed a lower FXR expression than low-grade RT4 and TSGH8301 cells. The above patient and cancer cell results implied that a reduction in FXR may affect the survival and malignancy in bladder cancers.

Page 2, line 100

These results implied that a reduction in FXR may affect the survival and malignancy in bladder cancers.

- Section 2.2 starts without authors mentioning that they set up a doxy-inducible overexpression system. The initial sentence of this section should be rewritten to discuss the adoption of an overexpression system. Also, a materials & methods section describing how the system was established is missing. Specifically, I do not understand if CTRL means: 1) viral-vector-transduced cells but NOT treated w/ doxycicline; or 2) doxycicline-treated parental cells; or control-viral-vector-(empty)-transduced cells treated w/ doxycicline; it is unclear if authors selected clones or they used transduced pools for the experiments. This is crucial for the interpretation of many of the following experiments. 

Reply: Thank you for the comments. We mentioned the use of doxycycline inducible system in the beginning of section 2.2. We also added the establishment of the doxycycline inducible system in the materials and methods. The control was doxycycline-treated parental cells. Transduced pools were applied in our study.

Page 3, line 144

Doxycycline inducible overexpression FXR system was established in TSGH8301 and T24 cells.

Page 11, line 417

4.2. Plasmid construction, lentivirus production and doxycycline inducible overexpression

The FXR cDNA transcript variant 1 clone and the P2A-DsRed sequence synthesis were provided from GenScript Biotech (New Jersey, USA). The FXR-P2A-DsRed was cloned into the pAS4.1w.Puro-aOn, a tetracycline-inducible plasmid. The lentivirus was produced by 293T cells co-transfected with the lentiviral vectors FXR-P2A-DsRed::pAS4.1w.Ppuro-aOn, pCMV-dR8.91, and pMD2.G using Lipofectamine 3000 reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. The National RNAi Core Facility provided the pAS4.1w.Puro-aOn and lentivirus package plasmids at Academia Sinica in Taiwan. The lentivirus was used to infect TSGH8301 and T24 cells. All cells were selected by puromycin (2 µg/mL). After antibiotics selection, the cells were sorted by FACSAria IIIu sorter (BD, New Jersey, USA). The mRNA and protein expression of FXR were checked after doxycycline induced for 24, 48, and 72 hours.

- It has been shown that doxycycline can affect viability and colony formation of selected cancer cell lines. Since appropriate controls were not included or at very least not discussed (for example, a cell line transduced with a doxy-inducible empty/GFP/reporter construct), it is not possible to know if T24 reduction in MTT/colony formation assays was due to doxy or the induction of FXR. Authors should look at the following paper by Zhang L et al, 2016 [10.1080/15384101.2016.1241929]. Doxycycline inhibits invasion, migration, and epithelial-to-mesenchymal transition of selected breast cancer cell lines.
For example, 7.13 μM Doxy was the IC50 for viability of the breast cancer MDA-MB-468 cell line (authors here used about 2 μM and reported a 25% reduction, which is in the same ballpark). Also note that Doxy was reported to inhibit MMPs (see https://pubchem.ncbi.nlm.nih.gov/compound/Doxycycline-hydrochloride)

Reply: Thank you for the comments. We evaluated the effect of doxycycline on cell viability in original T24 cells by MTT assay. In our experiments, we selected three different doses of doxycycline, 1 μM, 2 μM (our inducible overexpression system dose), and 5 μM. The results showed that none of the above concentration of doxycycline reduced cell viability in the un-transfected T24. Moreover, doxycycline did not affect migration, invasion, FXR, MMP2, and MMP9 expression in un-transfected T24 cells (Supplementary Figure 2).

Page 3, line 147

Doxycycline was no effect on cell viability, migration, and invasion in untransfected T24 cells (Supplementary Figure 2).

- How many replicates were performed for the colony-formation assays or other assays described in the manuscript? I could not find that info.

Reply: All experiments were performed at least five replicates. We mentioned this at 4.13 Statistical analysis. Only the results in the supplementary figure 2 were three.

Page 13, line 516

All results included at least five independent experiments.

- I do not agree with the interpretation of the western blots shown in figure 2C. The quantification of p-FAK is definitely wrong. Images clearly show some reduction for TSGH8301 (I agree with author's interpretation here) but there is not equivalent reduction in the bands from T24. Raw images show a bubble in the 48h band. Likewise, the bands of p-MLC are not homogenous, and it is impossible to quantify nor detect any signal reduction there. If the antibodies are not suitable for a high-quality WB or it is not possible to load more lysate, maybe authors should consider a more quantitative method such as "Jess" by Protein Simple. Overall, despite a reduction in integrin-b1 and accumulation of FXR, I believe nothing else changes in T24. 

Reply: Thank you for the comment. The original blots and quantification of T24 cells were showed as below. The reduction of p-FAK, integrin β1, integrin β3, and p-MLC in FXR overexpression group was observed. We apologized that the original blots of p-FAK was labeled as FAK in the submitted file. We also apologized that the “bubble” was the buffer trace left. The antibodies used in this study were applied in our previous studies [1].

- I don't get what is the rationale of inhibiting cells with MG132 (please, discuss in the text before presenting results). Moreover, I don't agree with author's conclusions about NH4Cl (Figure 3C and relative quantification). I see no real difference in signal strength for integrin-b1 when I look at FXR-O + MG132 or FXR-O + NH4Cl. How did authors end up with a quantification showing the same level as FXR-O? There must be an error! Also for other proteins, some quantifications make no sense to me. In the end, it looks to me that NH4Cl has the same effect of MG132 in TSGH8301. 

Reply: Thank you for the comment and correction.

• We added the rationale of applying MG132 and NH4Cl in our Due to the reduction of migratory-related proteins after FXR overexpression in Figure 2C, whether mRNA reduction or protein degradation occurred were investigated. The mRNA of integrin b1 and b3 was unaffected in TSGH8301 cells, while mRNA of integrin b3 was decreased in FXR overexpressed T24 cells (below). Therefore, proteasomal and lysosomal degradation inhibitors MG132 and NH4Cl were applied to TSGH8301 and T24 cells.

Page 4, line 192

Due to the reduction of integrins after FXR overexpression, whether mRNA re-duction or protein degradation occurred were investigated. The mRNA of integrin b1 and b3 was unaffected in TSGH8301 cells, while the mRNA of integrin b3 was decreased in FXR overexpressed T24 cells (data not shown).

• We changed the blots and the quantification results of TSGH8301 in Figure 3C (the original blots with marker were shown as below). MG132 could only reverse the migration effect in T24 cells, but it prevented the adhesion effect in both TSGH8301 and T24 cells. In addition, MG132 blocked integrin b1 in both TSGH8301 and T24 cells and integrin b3 in T24 cell only. Therefore, NH4Cl and MG132 showed the same effect on migration in TSGH8301 cells.

- Fig 3A, T24 panel. Images are not too clear, but it looks like both CTRL and FXR-O are way more dense at 0h as compared to 6h. Is it an image problem or are cells dying / detaching from the plate? It's tough to draw conclusions based on the presented data. 

Reply: Thank you for the comment. In wound healing experiments, we used PBS washing for three times after scratching in both TSGH8301 and T24 cell lines. However, we found that T24 cells are more sensitive during PBS washing and some cells may get into spherical shape. We replaced another wound healing images in T24 cells.

- Figure 4. Doxy is known to inhibit MMPs. Therefore, if the CTL samples are the only ones that were NOT treated with doxy, there is a confounding effect that makes it impossible to understand the effect of Doxy vs the effect of FXR-overexpression. 

Reply: Thank you for the comment. We evaluated the effect of doxycycline on cell invasive abilities in untransfected T24 cells by transwell assay. In our experiments, 2 μM (our inducible overexpression system dose) doxycycline was applied. We repeated three times of the invasion experiments. The results showed that doxycycline did not reduce the invasive abilities in T24 cells. Moreover, we checked the proteins expression of FXR and invasive related proteins, the results showed that FXR, MMP2, and MMP9 proteins level were not affected by doxycycline treatment (Supplementary Figure 2).  

- Figure 5D. HIF-1alpha levels are typically quasi-undetectable unless the experiment is conducted in hypoxic settings or in presence of MG132 (but authors didn't mention this). HIF-1alpha quantification in normoxia is usually unreliable/irrelevant. 

Reply: Thank you for the comment. MG132 induced HIF1α expression at 1 to 10 mM in cervical carcinoma and prostate cancer [2, 3]. In our study, a lower concentration of MG132 (100 nM) was applied in migration, adhesion and angiogenesis. Meanwhile, 100 nM MG132 could not enhance HIF1α expression in T24 cells as shown in Figure 6B. In addition, HIF1α is observed under normaxia condition in TCCSUP, SW780 and T24 bladder cancer cells [4]. This observation is also shown in J82 bladder cancer cells, while hypoxia induces more HIF1α and HIF2α expression [5]. Therefore, bladder cancer cells might express a baseline level of HIF1α.

References

1. Lin, C. H.; Chang, H. H.; Lai, C. R.; Wang, H. H.; Tsai, W. C.; Tsai, Y. L.; Changchien, C. Y.; Cheng, Y. C.; Wu, S. T.; Chen, Y., Fatty Acid Binding Protein 6 Inhibition Decreases Cell Cycle Progression, Migration and Autophagy in Bladder Cancers. Int J Mol Sci 2022, 23, (4).
2. Jiang, B. H.; Liu, L. Z.; Schafer, R.; Flynn, D. C.; Barnett, J. B., A novel role for 3, 4-dichloropropionanilide (DCPA) in the inhibition of prostate cancer cell migration, proliferation, and hypoxia-inducible factor 1alpha expression. BMC Cancer 2006, 6, 204.
3. Birle, D. C.; Hedley, D. W., Suppression of the hypoxia-inducible factor-1 response in cervical carcinoma xenografts by proteasome inhibitors. Cancer Res 2007, 67, (4), 1735-43.
4. Xu, L.; Li, H.; Wu, L.; Huang, S., YBX1 promotes tumor growth by elevating glycolysis in human bladder cancer. Oncotarget 2017, 8, (39), 65946-65956.
5. Lu, C. S.; Hsieh, J. L.; Lin, C. Y.; Tsai, H. W.; Su, B. H.; Shieh, G. S.; Su, Y. C.; Lee, C. H.; Chang, M. Y.; Wu, C. L.; Shiau, A. L., Potent antitumor activity of Oct4 and hypoxia dual-regulated oncolytic adenovirus against bladder cancer. Gene Ther 2015, 22, (4), 305-15.

Author Response File: Author Response.pdf

Reviewer 2 Report

Keywords: I suggest adding “protein half—life”

Revise accurately the English form

Clarify the abstract sentence “After treatment…bladder cancer cells”

In introduction section I suggest deepening the current novel therapeutic strategies and emphasize the existence of some compounds that could have a protective role against bladder cancer (doi: 10.3390/nu13113750). In addition,  novel diagnostic (10.1016/j.critrevonc.2022.103577) and therapeutic strategies (https://doi.org/ 10.3390/ijms23031133) should be mentioned.

In figure 1D adding a normal control could be useful.

FXR overexpression in TSGH8301 has not modified cell viability and colony formation assay. To support the hypothesis it could be relevant perform  a downregulation of FXR expression level in order to verify a phenotype similar to T24 cells.

The pathway controlling migration and adhesion should be better investigated. Are there FAK inhibitor able to mimic the effect of FXR overexpression?

The impact of results showed in paragraph 2.3 are very difficult to understand.

Fig. 3C: It should be explained why in control lane of both cells lines there is no signal for FXR

Fig.4: author should explain the rationale in performing experiments only in T24 cells

I have some doubts about conditional medium. Did you cultured cells in charcoal stripped serum? Is there VEGF or other factors in serum?

What is the mechanism by which FXR overexpression induces proteasome degradation?

The experiments with MG-132 should be revised and discussed better.

Author Response

Because figures could not be shown here, please see the attachment.

Reply to Reviewer 2’s comments

1. Keywords: I suggest adding “protein half—life”

Reply: Thank you for the suggestion. We added the“protein half—life” in the keywords.

2. Revise accurately the English form

Reply: Thank you for the suggestion. We checked and corrected our manuscript for revision.

3. Clarify the abstract sentence “After treatment…bladder cancer cells”

Reply: Thank you for the suggestion. We modified the sentence as following “After treatment with the proteosome inhibitor MG132, the migration, adhesion and angiogenesis caused by FXR overexpression were all reversed in bladder cancer cells.”.

4. In introduction section I suggest deepening the current novel therapeutic strategies and emphasize the existence of some compounds that could have a protective role against bladder cancer (doi: 10.3390/nu13113750). In addition, novel diagnostic (10.1016/j.critrevonc.2022.103577) and therapeutic strategies (https://doi.org/ 10.3390/ijms23031133) should be mentioned.

Reply: Thank you for the valuable comment. We modified the introduction according to the suggestion in the revised version.

The mortality rates have been declining, mainly due to the improvements in treatment (e.g., endoscopic resection, adjuvant instillation of chemotherapy, and intravesical immunotherapy) [1]. Previous studies shown that the existence of some compounds that could have a protective role against bladder cancer, including kaempferol, fisetin, and myricetin [2]. In addition, the detection of clinical biomarkers, NMP22, MDX, and uCAPP, in urine are also newly therapeutic strategies in recent studies [3]. However, the recurrence rate in bladder cancer patients is still an urgent subject.

5. In figure 1D adding a normal control could be useful.

Reply: Thank you for the suggestion. We performed FXR protein expression by Western blot. As shown below, the FXR showed higher expression in human bladder stroma fibroblast cell lysate (HBdSF, bought from ScienCell) and low-grade TSGH8301 than in high-grade T24 and J82 cells.

6. FXR overexpression in TSGH8301 has not modified cell viability and colony formation assay. To support the hypothesis it could be relevant perfor a downregulation of FXR expression level in order to verify a phenotype similar to T24 cells.

Reply: Thank you for the comment. We evaluated the effect of siFXR on cell viability and migration in two non-muscle-invasive BC cancer cell lines RT4 and TSGH8301 (Supplementary Figure 4). FXR knockdown by siFXR in TSGH8301 and RT4 enhanced the cell viability and migration ability. In protein expression, the level of integrin b3 and p-MLC were all also increased after FXR inhibition.

Page 3, line 165

In contrast, knockdown of FXR enhanced migration and the protein expression of integrin b3 and p-MLC in low-grade RT4 and TSGH8301 cells (Supplementary Figure 4).

7. The pathway controlling migration and adhesion should be better investigated. Are there FAK inhibitor able to mimic the effect of FXR overexpression?

Reply: Thank you for the suggestion. A previously study “Circular RNA circPICALM sponges miR-1265 to inhibit bladder cancer metastasis and influence FAK phosphorylation” has demonstrated that FAK Inhibitor 14 could eliminate the migration and invasion effects of STEAP4 silencing in T24 cell [4]. Therefore, the effect of FAK inhibitor on migration is similar with FXR overexpression in high-grade T24 bladder cancer cells.

8. The impact of results showed in paragraph 2.3 are very difficult to understand. Fig. 3C: It should be explained why in control lane of both cells lines there is no signal for FXR

Reply: Thank you for the suggestion. We modified the exposure time of these blots to clearly distinguished signal of FXR in CTL groups in Figure 3C. FXR expression in T24 cells was weak which was also shown in Figure 1D. Due to the reduction of migratory-related proteins after FXR overexpression in Figure 2C, whether mRNA reduction or protein degradation occurred were investigated. The mRNA of integrin b1 and b3 was unaffected in TSGH8301 cells, while mRNA of integrin b3 was decreased in FXR overexpressed T24 cells (below). Therefore, proteasomal and lysosomal degradation inhibitors MG132 and NH4Cl were applied to TSGH8301 and T24 cells. According to our results, only MG132 prevented the effect of FXR on migration, adhesion and integrin expression.

Page 4, line 192

Due to the reduction of integrins after FXR overexpression, whether mRNA re-duction or protein degradation occurred were investigated. The mRNA of integrin b1 and b3 was unaffected in TSGH8301 cells, while the mRNA of integrin b3 was decreased in FXR overexpressed T24 cells (data not shown).

9. 4: author should explain the rationale in performing experiments only in T24 cells

Reply: Thank you for the comment. TSGH8301 is a low-grade and non-muscle invasive bladder cancer cell line. TSGH8301 cells could not penetrate the transwell chamber with Matrigel which we had tried before. Therefore, muscle invasive T24 cells was performed in Figure 4. We added the rationale in the results section 2.5..

Page 5, line 226

Due to low-grade TSGH8301 cells could not penetrate the transwell chamber with Matrigel, only T24 cells were evaluated for invasion assay.

10. I have some doubts about conditional medium. Did you cultured cells in charcoal stripped serum? Is there VEGF or other factors in serum?

Reply: Thank you for the suggestion. In previous studies, bladder tumors were found to produce high levels of various factors involved in stimulation of angiogenesis, including vascular endothelial growth factor A (VEGF-A), basic fibroblast growth factor (bFGF), interleukin-8 (IL-8) and matrix metalloproteinases (MMPs) [5]. Indeed, we could not exclude the above factors. In our experiments, we focused on the effect of VEGF in angiogenesis after FXR-O in bladder cancer cells. Moreover, the MMP2 and VEGF content in the conditioned medium was also decreased measured by ELISA assay in Figure 4C and 5B. Also, the application of VEGF reversed the tube formation in FXR-O groups (Supplementary Figure 5, below). Although other factors could not be excluded, VEGF was involved in FXR overexpression-reduced angiogenesis.

11. What is the mechanism by which FXR overexpression induces proteasome degradation?

Reply: Thank you for asking the important question. We also tried to figure out the possible mechanism. Proteasome subunit beta 8 (PSMB8) is one of the key components to generate immunoproteasomes [6]. PSMB4 is a non-catalytic subunit in proteasome [7]. We checked the mRNA expression of PSMB4 and PSMB8 after FXR overexpression, and we found that PSMB8 mRNA level increased significantly after FXR overexpression in T24 cells (below). These results gave us hints that FXR overexpression may recruit and produce more proteosomes to degrade proteins. However, the detail mechanism needs to be investigated in further studies.

* p<0.05; ** p<0.01; *** p<0.001 compared to the CTL group.

12. The experiments with MG-132 should be revised and discussed better.

Reply: Thank you for the suggestion. We revised the blots in MG132 study in our manuscript. We further discussed MG132 in migration, angiogenesis and invasion.

Page 10, line 343

MG132 has been reported to enhance migration and mesenchymal phenotype in A549 lung cancer cells [8]. Besides, MG132 recovers phospholipase C-γl overexpression-reduced adhesion and migration in rat fibroblasts [9].

Page 10, line 354

The expression of MMP2 and MMP9 were significantly reduced after FXR overexpression which may inhibit invasion in human bladder cancer T24 cell. MMPs mediate the breakdown of the basal membrane, degrade the extracellular matrix, and create a microenvironment that enhances tumor cell migration and invasion [10]. In colon cancer, FXR overexpression inhibits colon cancer cell proliferation and invasion in vitro by suppressing MMP7 mRNA and protein expression [11]. In liver cancer, FXR overexpression results in the significant reduction of total and nuclear β-catenin proteins and its downstream target genes, including c-Jun and MMP9, in vitro and in vivo that contribute to the inhibition of cell invasive abilities. Therefore, MMP9 protein was decreased in human bladder cancer T24 cells might be due to β-catenin reduction. However, after treated with proteosome inhibitor MG132 and lysosome inhibitor NH4Cl, neither of the above can reverse the FXR overexpression-inhibited invasion in T24 cells (data not shown). Although both MG132 and NH4Cl could reverse the activity of MMP9, they could not prevent the reduction of MMP2 (Figure 4C and D). The mechanism of FXR overexpression inhibited invasion, MMP2 and MMP9 expression may need further investigation.

Page 10, line 395

Moreover, MG132 abrogates the reduction of VEGFR2, AKT, and ERK1/2 caused by octaminomycins in HUVECs [12].

References

1. Babjuk, M.; Burger, M.; Zigeuner, R.; Shariat, S. F.; Van Rhijn, B. W.; Compérat, E.; Sylvester, R. J.; Kaasinen, E.; Böhle, A.; Redorta, J. P., EAU guidelines on non–muscle-invasive urothelial carcinoma of the bladder: update 2013. European urology 2013, 64, (4), 639-653.
2. Crocetto, F.; Di Zazzo, E.; Buonerba, C.; Aveta, A.; Pandolfo, S. D.; Barone, B.; Trama, F.; Caputo, V. F.; Scafuri, L.; Ferro, M., Kaempferol, Myricetin and Fisetin in Prostate and Bladder Cancer: A Systematic Review of the Literature. Nutrients 2021, 13, (11), 3750.
3. Crocetto, F.; Barone, B.; Ferro, M.; Busetto, G. M.; La Civita, E.; Buonerba, C.; Di Lorenzo, G.; Terracciano, D.; Schalken, J. A., Liquid biopsy in bladder cancer: state of the art and future perspectives. Critical reviews in oncology/hematology 2022, 103577.
4. Yan, D.; Dong, W.; He, Q.; Yang, M.; Huang, L.; Kong, J.; Qin, H.; Lin, T.; Huang, J., Circular RNA circPICALM sponges miR-1265 to inhibit bladder cancer metastasis and influence FAK phosphorylation. EBioMedicine 2019, 48, 316-331.
5. Fus, L. P.; Gornicka, B., Role of angiogenesis in urothelial bladder carcinoma. Cent European J Urol 2016, 69, (3), 258-263.
6. Li, C.; Dai, S.; Yan, Z.; Zhang, X.; Liu, S.; Wang, X.; Wang, J.; Shi, L.; Yao, Y., Genetic polymorphisms of proteasome subunit genes of the MHC-I antigen-presenting system are associated with cervical cancer in a Chinese Han population. Hum Immunol 2020, 81, (8), 445-451.
7. Murata, S.; Yashiroda, H.; Tanaka, K., Molecular mechanisms of proteasome assembly. Nat Rev Mol Cell Biol 2009, 10, (2), 104-15.
8. Lang, B. J.; Nguyen, L.; Nguyen, H. C.; Vieusseux, J. L.; Chai, R. C.; Christophi, C.; Fifis, T.; Kouspou, M. M.; Price, J. T., Heat stress induces epithelial plasticity and cell migration independent of heat shock factor 1. Cell Stress Chaperones 2012, 17, (6), 765-78.
9. Kang, J. K.; Chang, C. H.; Nam, H. J.; Kim, S. K.; Ahn, K. J.; Seok, H.; Park, S. J.; Kang, Y. J.; Jo, Y. S.; Shong, M.; Kim, H., Downregulation of erythropoietin receptor by overexpression of phospholipase C-gamma 1 is critical for decrease on focal adhesion in transformed cells. Cell Oncol (Dordr) 2011, 34, (1), 11-21.
10. Cho, Y.-H.; Ro, E. J.; Yoon, J.-S.; Mizutani, T.; Kang, D.-W.; Park, J.-C.; Kim, T. I.; Clevers, H.; Choi, K.-Y., 5-FU promotes stemness of colorectal cancer via p53-mediated WNT/β-catenin pathway activation. Nature communications 2020, 11, (1), 1-13.
11. Peng, Z.; Chen, J.; Drachenberg, C. B.; Raufman, J.-P.; Xie, G., Farnesoid X receptor represses matrix metalloproteinase 7 expression, revealing this regulatory axis as a promising therapeutic target in colon cancer. Journal of Biological Chemistry 2019, 294, (21), 8529-8542.
12. Jang, J. P.; Han, J. M.; Jung, H. J.; Osada, H.; Jang, J. H.; Ahn, J. S., Anti-Angiogenesis Effects Induced by Octaminomycins A and B against HUVECs. J Microbiol Biotechnol 2018, 28, (8), 1332-1338.

Author Response File: Author Response.pdf

Reviewer 3 Report

The story telling and the data are well presented. This study provided a piece of evidence for therapeutic potential target, FXR, in bladder cancer. However, some spelling and format should be checked. For example, two 2.2 can be observed in the Results section.

Author Response

Reply to Reviewer 3’s comments

The story telling and the data are well presented. This study provided a piece of evidence for therapeutic potential target, FXR, in bladder cancer. However, some spelling and format should be checked. For example, two 2.2 can be observed in the Results section.

Reply: Thank you for the suggestion and correction. We checked and corrected our manuscript.

Round 2

Reviewer 1 Report

I appreciate the revisions by the authors to address some of the points I raised during my initial review (especially the description of the inducible models and the extra sentences to provide more context around some of the experiments discussed in the manuscript).

In the revised version of the paper, the authors presented data showing that high FXR expression tracked with better disease survival in BLCA. Authors established two inducible FXR-overexpressing BLCA cell models using a low grade (TSGH) and high grade (T24) BLCA cell lines. FXR overexpression reduced cell viability in T24 cells only. It also reduced adhesion in both cell lines, possibly, by lowering integrin levels (unclear). FXR overexpression reduced MMP activity, NOS2 expression and possibly VEGF levels. Most of these effects were reverted upon MG132 treatment, suggesting that the mechanisms were depending on proteasome-mediated protein degradation. 
Interestingly, it is unclear whether any of these observations could be linked to the better survival of FXR-high BLCA patients, since overlapping results were observed in both cell lines that expressed low (T24: high grade) or high (TSGH: low grade) basal levels of FXR.

This being said, I am still not satisfied about the following points:

• cell models used as controls are sub-optimal since they were not transduced with an control viral vector (parental cell lines + doxy)
• scratch assay results (especially from T24 cells) seem weak, cause there is a substantial cell loss at 6h. While I understand the technical challenges of this experiment, I am not sure how to interpret the results
• While I appreciate the attempt to provide higher-quality WB images, I still believe that these WB scans are low-quality and I don't find accurate (or reliable) the results of the densitometric analyses of several bands as provided by the authors
• I never said that HIF-1a is undetectable in normoxic conditions. I just said that its low levels are typically irrelevant, cause it makes more sense to assess HIF-1a levels upon its stabilization (given the tight post-translational control) either by hypoxia or treatment with MG132. Nevertheless, I do not see any meaningful changes in the WB figures provided by the authors.
• Revisions introduced several style/language issues that authors should address before re-submitting.

Reviewer 2 Report

The authors answered all comments and suggestions.