Crosstalk between Cancer Cells and Fibroblasts for the Production of Monocyte Chemoattractant Protein-1 in the Murine 4T1 Breast Cancer

Round 1

Reviewer 1 Report

1. It is sad that in vitro result does not happen in vivo. Crenolanib has oral bioactivity. Is the intratumoral injection an improper way of administration? How do you deliver Trapidil?
2. Since MCP-1/CCL2 plays an important role in lung metastasis, why not looking at the lung metastasis lesions? Maybe it is not important in the primary tumor site.
3. It appears to be your hypothesis doesn’t work in the 4T1 model, how about other syngeneic murine breast cancer models?
4. The frozen-thawed cells are not technically necrotic cells. It is less convincing that necrotic cells induced CCL2 expression by macrophages.
5. Several typos in subtitles: 4.T1 should be 4T1.

Author Response

Thank you very much for reviewing our manuscript and providing with valuable comments. Our responses are as follows:

Point-by-point responses

 Reviewer #1

Comments and Suggestions for the Authors

1. It is sad that in vitroresult does not happen in vivo. Crenolanib has oral bioactivity. Is the intratumoral injection an improper way of administration? How do you deliver Trapidil?

(Our response) We understand the reviewer’s concern. Intratumoral injection is an effective method to treat tumors and has been used in many studies (Brachi et al., nanoscale 2020;12:23838-23850). Although crenolanib can be administered orally, crenolanib preparation we used inhibited PDGFRs in vitro and there is no reason not to expect its inhibitory effect in vivo when it was injected intratumorally. Trapidil was dissolved in PBS and injected intraperitoneally as previously done by others (references are included in the manuscript).

2. It appears to be your hypothesis doesn’t work in the 4T1 model, how about other syngeneic murine breast cancer models?

(Our response) We appreciate the reviewer’s comment. We considered using transgenic mice in which breast cancer spontaneously arises. However, such mice were not locally available. We also considered examining human breast cancer tissues. As discussed in our original manuscript, the cell types producing or expressing MCP-1 in human breast cancer were previously examined by others and fibroblasts were not a major cell type producing MCP-1. Therefore, we did not use other mouse models.

3. The frozen-thawed cells are not technically necrotic cells. It is less convincing that necrotic cells induced CCL2 expression by macrophages.

(Our response) Necrosis is now subclassified into several types, including pyroptosis, necroptosis, parthanatos, ferroptosis, oxytosis, ETosis, and secondary necrosis (Pasparakis M, Vandenabeele P. Nature 517:311–20, 2015). Several endogenous molecules are released from necrotic cells and known to activate stromal cells, such as macrophages, to produce proinflammatory cytokines/chemokines. In this manuscript, we used “necrosis” as a general term and freezing and thawing has been previously used to prepare necrotic cells by other investigators (Basu S., et al., Int Immunol. 12:1539-1546, 2000; Hatano M., et al., PLoS ONE 14(8): e0218632, 2019). Thank you very much for the comment.

4. Several typos in subtitles: 4.T1 should be 4T1.

(Our response) We apologize for the errors. They have been corrected.

Reviewer 2 Report

In this paper, Imamura M et al, seek to understand the role of cross-talk between the tumor cell and fibroblast in breast cancer model, and using cell culture approaches, they show that 4T1 tumor induces the fibroblast production of CCL2 via PDGFR-signaling. The authors further aimed to characterize in vivo significance of this cross-talk and find that blockage of CCL2 production by fibroblast using inhibitors of PDGFR-signaling neither affect tumor growth nor CCL2 levels in the serum of tumor-bearing mice. Although this study is interesting and holds a significant value in understanding the interplay between tumor-stromal interaction in TME, there are some issues/suggestions that need to be addressed to improve the overall quality and readability of the paper. Below are my specific concerns.   

Specific comments.

• The hypothesis/aim of the study is not clearly highlighted in the introduction section.
• The introduction section had great details about the potential role of tumor-stroma or tumor-immune cross-talk involving CCL2  and its impact on lung metastasis in breast cancer models. However, none of the in vivo experiments presented in the manuscript analyzed the status of lung metastasis in the 4T1 breast cancer model known to exhibit spontaneous lung and other tissue metastasis. If the main aim of the paper is to analyze the key role of tumor-stroma interaction leading to CCL2 production by stromal cells in early primary tumor establishment, then authors should accordingly blend the introduction section. 
• Figure 3 shows that the pretreatment of 3T3 fibroblast in culture with PDGFR inhibitor, Crenolanib markedly reduced the 4T1 tumor cell-Sup-induced CCL2 production by 3T3. What was the cell viability after the treatment? Crenolanib is known to inhibit cell proliferation, via activating ER stress response (Reichert D et. al., Cells. 10(4): 804; Berndsen RH e.al, Br J Cancer. 121, 139–149 (2019). Cell viability after the Crenolanib treatment could influence the overall levels of CCL2 in the culture supernatant.   
• Fig 4. Did the authors analyze lung metastasis in Crenolanib or Trapidil-treated 4T1 tumor-bearing mice? It's quite possible that these PDGFR inhibitors might reduce spontaneous lung metastasis of 4T1 tumors. 
• The kinetics (day 5) used to analyze the effect of PDGFR inhibitors mediated perturbations of 4T1-fibroblast cross-talk in tumor microenvironment seems a bit early. As evident from the IHC data, fibroblasts are not abundantly present in the TME even on day 14. It's likely that CAF-derived CCL2 might not be a critical factor in shaping early TME, but might be important once an advanced tumor gets established. Authors can briefly discuss this issue in the discussion section. 
• Figure 6. The authors followed the tumor for 14 days before being used for IHC analysis. Did authors find any difference in the tumor growth kinetics? 
• Figure 4-6. Did the authors find changes in the CCL2 protein levels in the d5 and d14 tumor tissue?  
• It's highly likely that the lack of evidence of the inability of PDGFR-alpha and beta knockout or knockdown fibroblast to induce CCL2 expression (to demonstrate the specificity of PDGF-a-PDGFR signaling) in response to 4T1-Sup’s (and also Crenolanib’s action; as it also known to modulate other protein tyrosine kinases (Galanis A et al, Blood. 123(1):94-100)) specificity is one of the major limitations of the in vitro experiments. The authors should discuss this limitation. 

Minor comments.

• Figure 1-4 legends. How many times experiments were replicated? Authors should indicate n values and if any technical replicates were performed in each experiment. 
• Figure 5-6. How many mice do the IHC images represent?  
• Typo in result section (section 3.2.4 should be 3.2)

Author Response

Thank you very much for your valuable comments. Our responses are as follows:

Reviewer #2

Comments and Suggestions for the Authors

In this paper, Imamura M et al, seek to understand the role of cross-talk between the tumor cell and fibroblast in breast cancer model, and using cell culture approaches, they show that 4T1 tumor induces the fibroblast production of CCL2 via PDGFR-signaling. The authors further aimed to characterize in vivo significance of this cross-talk and find that blockage of CCL2 production by fibroblast using inhibitors of PDGFR-signaling neither affect tumor growth nor CCL2 levels in the serum of tumor-bearing mice. Although this study is interesting and holds a significant value in understanding the interplay between tumor-stromal interaction in TME, there are some issues/suggestions that need to be addressed to improve the overall quality and readability of the paper. Below are my specific concerns.   

Specific comments.

• The hypothesis/aim of the study is not clearly highlighted in the introduction section.

(Our response) In response to the reviewer’s comment, we now state our hypothesis/aim more clearly. Thank you very much for the comment.

• The introduction section had great details about the potential role of tumor-stroma or tumor-immune cross-talk involving CCL2 and its impact on lung metastasis in breast cancer models. However, none of the in vivo experiments presented in the manuscript analyzed the status of lung metastasis in the 4T1 breast cancer model known to exhibit spontaneous lung and other tissue metastasis. If the main aim of the paper is to analyze the key role of tumor-stroma interaction leading to CCL2 production by stromal cells in early primary tumor establishment, then authors should accordingly blend the introduction section.

(Our response) We appreciate the comment. In response to the comment, we revised the Introduction section of our manuscript accordingly.

• Figure 3 shows that the pretreatment of 3T3 fibroblast in culture with PDGFR inhibitor, Crenolanib markedly reduced the 4T1 tumor cell-Sup-induced CCL2 production by 3T3. What was the cell viability after the treatment? Crenolanib is known to inhibit cell proliferation, via activating ER stress response (Reichert D et. al., Cells. 10(4): 804; Berndsen RH e.al, Br J Cancer. 121, 139–149 (2019). Cell viability after the Crenolanib treatment could influence the overall levels of CCL2 in the culture supernatant. 

(Our response) We appreciate the comment. In the experiment presented in Figure 3, we pretreated 3T3 cells with crenolanib for 30 min and then cultured in the absence or presence of 4T1 sup for an additional 24 hours. As presented in Figure 3, treatment of 3T3 cells by 0.2 or 2 μM of crenolanib had no effect on the constitutive MCP-1 production and we did not observe any clear differences in cell viability. In the study by Reichert, et al., their cells were treated with crenolanib for 7 days. In the study by Bernsen et al., immortalized human endothelial cells (ECRF24), freshly isolated primary human umbilical vein endothelial cells (HUVEC), human ovarian carcinoma cells (A2780) and adult human dermal fibroblasts (HDFa) were incubated for 72 hours with different concentrations of crenolanib. IC50 values were 5.1 μM for A2780, 4.6 μM for ECRF24 and 8.4 μM for HUVEC. Two mM of crenolanib had little effect on the viability of HDFa. Thus, it is unlikely that crenolanib affected the viability of 3T3 cells and influenced the level of CCL2 in the culture supernatant. However, long term administration of crenolanib in vivo may affect the viability of 4T1 cells and stromal cells in tumors. We discussed this issue in the Results section (page 8). 

• Fig 4. Did the authors analyze lung metastasis in Crenolanib or Trapidil-treated 4T1 tumor-bearing mice? It's quite possible that these PDGFR inhibitors might reduce spontaneous lung metastasis of 4T1 tumors. 

(Our response) In the present study, we treated mice with PDGFR inhibitors for only short periods and we did not examine their effects on lung metastasis. We agree that long term administration might reduce spontaneous lung metastasis of 4T1 cells. In response to the comment, we added a discussion on this subject in the Discussion section (page 12).

• The kinetics (day 5) used to analyze the effect of PDGFR inhibitors mediated perturbations of 4T1-fibroblast cross-talk in tumor microenvironment seems a bit early. As evident from the IHC data, fibroblasts are not abundantly present in the TME even on day 14. It's likely that CAF-derived CCL2 might not be a critical factor in shaping early TME, but might be important once an advanced tumor gets established. Authors can briefly discuss this issue in the discussion section. 

(Our response) Thank you very much for the comment. It’s a very important point and we added a discussion on the possibility in the Discussion section (page 8, 12).

• Figure 6. The authors followed the tumor for 14 days before being used for IHC analysis. Did authors find any difference in the tumor growth kinetics? 

(Our response) We are not sure whether we completely understood the comment. The tumors which were subjected to IHC were from mice without treatment by PDGFR inhibitors. We measured the size of tumors in trapidil-treated or untreated mice and found there was no difference. When intratumoral injection was performed, we did not measure tumor size because we thought the injection itself had some physical effects.  Thank you very much for the comment.

• Figure 4-6. Did the authors find changes in the CCL2 protein levels in the d5 and d14 tumor tissue? 

(Our response) We previously found that there was a good correlation between tumor MCP-1 mRNA expression and serum MCP-1 level. Thus, we measured MCP-1 levels in sera but not in tumors. We tried to compare the rate of MCP-1 production in tumor-bearing mice at 1, 2, 3 and 4 weeks after inoculation of 4T1 cells by dividing the serum MCP-1 level by tumor size we previously measured and found that the value at one week was higher than that at other time points, suggesting that MCP-1 production is more active in an early phase of tumor development. This data is now presented as Supplementary Figure 1 and also commented in the text (page 8, 2^nd paragraph). Thank you very much for the comment.

• It's highly likely that the lack ofevidence of the inability of PDGFR-alpha and beta knockout or knockdown fibroblast to induce CCL2 expression (to demonstrate the specificity of PDGF-a-PDGFR signaling) in response to 4T1-Sup’s (and also Crenolanib’s action; as it also known to modulate other protein tyrosine kinases (Galanis A et al, Blood. 123(1):94-100)) specificity is one of the major limitations of the in vitro experiments. The authors should discuss this limitation. 

(Our response) Thank you very much for the comment. As pointed out by the reviewer, crenolanib was initially developed as a highly selective PDGFR inhibitor, but later found to also inhibit FMS-like Tyrosine Kinase 3 (FLT3). Since FLT3 is expressed by hematopoietic cells, but not by fibroblasts, the inhibitory effect of crenolanib in our in vitro experiment was considered via blocking of PDGFRs. Crenolanib was also reported to inhibit cell proliferation, potentially influencing the level of MCP-1 in culture supernatants. However, MCP-1 mRNA expression by 3T3 cells peaked within 1 hour after stimulation with 4T1-sup (Fig. 2A), and crenolanib treatment had no effect on constitutive MCP-1 production or mRNA expression at the concentrations of 0.2 and 2 mM (Fig 3A, B).  Taken together, these results supported the conclusion that increased MCP-1 production by fibroblasts in response to 4T1-sup was due to the activation of PDGF receptors by cancer cell-derived PDGFs. We added a discussion about the specificity of crenolanib in the revised manuscript (page 8, first paragraph).

Minor comments.

• Figure 1-4 legends. How many times experiments were replicated? Authors should indicate n values and if any technical replicates were performed in each experiment. 

(Our response) We apologize for not including the important information. We added the information in the Figure legends. Thank you very much.

• Figure 5-6. How many mice do the IHC images represent?  

(Our response) Sections from three tumors in 3 mice were subjected to IHC. It is stated in the legend. Thank you very much.

• Typo in result section (section 3.2.4 should be 3.2)

(Our response) We corrected typing errors. Thank you.

 

 

Reviewer 3 Report

This is an interesting manuscript trying to unveil a potential dialogue between breast cancer cells and tumor-associated fibroblasts via the link between PDGF and MCP-1. This is original research conducted with logical experiments, and the results are well written and convincing. However, I have several concerns about the manuscript before publication.

1. MCP-1 has been known to play important roles in recruiting immunosuppressive and proangiogenic TAMs and in promoting lung metastasis of breast cancer cells. PDGF is also known to recruit and activate various types of stromal cells including CAFs which are potent forces driving aggressive tumor progression. In addition, the authors mentioned that the previous studies noticed the MCP-1 expression in human invasive ductal breast cancer. In these contexts, I am curious why the authors designed in vivo experiments to analyze the heterotypic interaction in the early phase of the tumorigenesis rather than the later stage. Another in vivo experiment might be necessary to examine them in the later stage.
2. In every figure, the authors describe ‘representative of two or three experiments’ with a p-value. But no indication of how many replicates in each experiment. Have you acquired every result from at least two or three independent experiments in the singlet, duplicate, triplicate, or more? In addition, what do error bars in every figure indicate? SD or SEM?
3. For some experiments (figure 2A, 2B, 2D), the authors mentioned that they performed just twice. Repeated twice? Is that enough to conclude that results are reproducible?
4. In line 295, the sentence sounds like only PDGF is the stimulator for fibroblasts to produce MCP-1, even though the results in Figures 1 and 2 are not enough to support the conclusion. Must be toned down.
5. For in vivo experiments, the authors noticed that trapidil was injected twice a day for a week. Crenolanib was injected twice, three times, or five times every 12 hours for 3 days. Twice a day, twice, three times, or five times every 12 hours? why? Is it a general protocol for trapidil and crenolanib? For figure 4A, crenolanib was injected five times every 12 hours, but no indication of how long the treatment was performed.
6. In line 395: the authors wrote as following: ‘peripheral area (Fig. 5E, F) and inside the tumor (Fig. 5G, H).’ However, in legends for Figure 5, E and G: peritumoral area and F and H: intratumoral area. Which one is true?
7. Why in situ hybridization rather than immunohistochemistry in Figure 5E-H and Figure 6A?
8. What are 50ml and 100ml in Figure 6B and C? No explanation in figure legends.

Author Response

Point-by-point responses:

We first thank the reviewer for the positive review of our manuscript and constructive comments for further improvement. Our point-by-point responses are listed below.

Comments and Suggestions for Authors

This is an interesting manuscript trying to unveil a potential dialogue between breast cancer cells and tumor-associated fibroblasts via the link between PDGF and MCP-1. This is original research conducted with logical experiments, and the results are well written and convincing. However, I have several concerns about the manuscript before publication.

1. MCP-1 has been known to play important roles in recruiting immunosuppressive and proangiogenic TAMs and in promoting lung metastasis of breast cancer cells. PDGF is also known to recruit and activate various types of stromal cells including CAFs which are potent forces driving aggressive tumor progression. In addition, the authors mentioned that the previous studies noticed the MCP-1 expression in human invasive ductal breast cancer. In these contexts, I am curious why the authors designed in vivoexperiments to analyze the heterotypic interaction in the early phase of the tumorigenesis rather than the later stage. Another in vivo experiment might be necessary to examine them in the later stage.

(Our response) Thank you very much for asking the important question. We have been studying the crosstalk between cancer cells and stromal cells that leads to MCP-1 production during the progression of breast cancer or lung cancer using murine transplantable cancer models. We previously examined the kinetics of serum MCP-1 concentration after the inoculation of 4T1 cells and reported that serum MCP-1 level significantly increased at 1 week, peaked at 3 weeks and then decreased at 4 weeks (Ref. 12). Since tumor sizes were much larger at later stages, we reanalyzed the results by dividing the serum MCP-1 concentration by tumor size. The result presented as Supplementary Figure 1 strongly suggested that MCP-1 production in 4T1 tumors appears most active in the early phase of tumor development (1 to 2 weeks after cancer cell inoculation). This is also the time when the metastatic foci in the lung we find at 4 weeks are established. Therefore, we think identifying the mechanisms in the early phase is important. This is noted in the Introduction (line 77-79) and in the Result section (line 373-379). As the reviewer pointed out, it would be interesting to characterize the interaction between cancer cells and stromal cells in the later phase in future studies.

 

2. In every figure, the authors describe ‘representative of two or three experiments’ with a p-value. But no indication of how many replicates in each experiment. Have you acquired every result from at least two or three independent experiments in the singlet, duplicate, triplicate, or more? In addition, what do error bars in every figure indicate? SD or SEM?

(Our response) We apologize for not providing enough information. In each experiment, assays were performed in duplicate with at least 3 samples. And the same experiments were performed multiple times. We now provide the information in the M&M section (line 191, 210) and figure legends.

 

3. For some experiments (figure 2A, 2B, 2D), the authors mentioned that they performed just twice. Repeated twice? Is that enough to conclude that results are reproducible?

(Our response) For Figure 2A and 2B, we did more than 3 experiments, including preliminary experiments. After finding optimal conditions, we performed final two experiments for the manuscript and one of the results is presented. The results presented in Figure 2A were consistent with those of previous studies. We never detected the expression of Pdgfd mRNA during this study (Fig. 2B). Regarding Figure 2D, we found results of two additional experiments; therefore, we performed a total of 4 experiments. We corrected the figure legend. We apologize for the mistake. Thank you for the comment.

 

4. In line 295, the sentence sounds like only PDGF is the stimulator for fibroblasts to produce MCP-1, even though the results in Figures 1 and 2 are not enough to support the conclusion. Must be toned down.

(Our response) We agree with the reviewer’s comment. We revised the sentence and now state “suggested” instead of “indicated” (line 317). Thank you.

 

5. For in vivo experiments, the authors noticed that trapidil was injected twice a day for a week. Crenolanib was injected twice, three times, or five times every 12 hours for 3 days. Twice a day, twice, three times, or five times every 12 hours? why? Is it a general protocol for trapidil and crenolanib? For figure 4A, crenolanib was injected five times every 12 hours, but no indication of how long the treatment was performed.

(Our response) Antagonizing an active molecule in vivo is not an easy task. When the treatment with an antagonist showed no effect, we have to wonder if the treatment was appropriately performed.

For trapidil treatment, there was no general protocol. We reviewed three previous studies in which trapidil was used. In Ref 30, it was given by gavage from 2 days before to 14 days after balloon injury of carotid artery in rats and healing of the injury was examined on Day 7 and 14. In Ref 31, it was intraperitoneally (i.p.) injected twice a day for 7 days to examine the effect on bone loss and mice were euthanized on day 8. In Ref 32, it was i.p. injected three times on a single day, and an effect (a reduction in cell migration) was seen within 24 hours. In the present study study, we chose to treat mice for 7 days and euthanize them on the next day to examine its effect on MCP-1 expression and production.

            For crenolanib treatment, intratumoral (i.t.) injection had never been performed. Therefore, we made our own protocol. We chose IT injection because i.p. injection of trapidil had no effect and we wanted to use direct delivery of crenolanib to block PDGFR inside tumors. In the experiments shown in Figure 4A, we injected 5 times every 12 hours from day 5 to 7. For the experiments in Figure 4C, we injected 3 times every 12 hours from day 5 to 6. We reduced the number of injections because we wanted to minimize the stress to the mice and physical effects caused by repeated i.t. injections which may upregulate MCP-1 expression. Based on the previous study with trapidil (Ref 32), we concluded three i.t. or i.p. injections would be sufficient to block PDGFRs. After all, MCP-1 mRNA was mainly associated with macrophages; therefore, we did not perform additional experiments with PDGFR inhibitors. Thank you very much.

 

6. In line 395: the authors wrote as following: ‘peripheral area (Fig. 5E, F) and inside the tumor (Fig. 5G, H).’ However, in legends for Figure 5, E and G: peritumoral area and F and H: intratumoral area. Which one is true?

(Our response) We now use “peripheral area” and “inside tumors” in both text and figure legend. Thank you very much for pointing out the inconsistent statements.

 

7. Why in situ hybridization rather than immunohistochemistry in Figure 5E-H and Figure 6A?

(Our response) We previously attempted to identify MCP-1 producing cells in 4T1 tumors by immunohistochemistry with multiple antibodies. Although MCP-1 was detected by immunohistochemistry in other laboratories, our attempts were not successful. Therefore, in the present study we used ISH. Unlike the proteins (cytokines/chemokines) which are secreted outside the cells after their production, mRNA remain inside the cells until their degradation, making ISH a better means in this study. In response to the comment, we added a sentence explaining why ISH was used (line 439-441). Thank you very much.

 

8. What are 50ml and 100ml in Figure 6B and C? No explanation in figure legends.

(Our response) We apologize for not clearly indicating what they were. We now state “in the absence or presence of 50 or 100 ml of necrotic GM-CSF-deficient A8 cells” in the figure legend. Thank you very much.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

1. Intratumoral administration is very rarely used in clinical practice. Crenolanib is given orally in clinical trials. I don't understand why the authors changed the way of drug administration. 

2. Besides the transgenic mice, E0771 in B6 mice can be another option to try.

Author Response

Point-by point responses

We appreciate the reviewer’s continuous effort to help us improve our manuscript. Here are our responses.

Comments and Suggestions for Authors

1. Intratumoral administration is very rarely used in clinical practice. Crenolanib is given orally in clinical trials. I don't understand why the authors changed the way of drug administration.

(Our response) In the present study, we first used crenolanib solution in vitro and determined that 4T1-sup-induced increase in MCP-1 expression/production by fibroblasts was via activation of PDGFRs. Although crenolanib is a potent inhibitor of PDGFRs, it also inhibits FLT3 expressed on hematopoietic cells as pointed out in the previous review. Crenolanib also inhibits proliferation of cancer cells. To minimize its systemic effects on both types of receptors and its inhibitory effect on cell proliferation in vivo, we injected crenolanib solution directly into tumors 5 time every 12 hours (Fig. 4A). In a later experiment (Fig. 4C), we also treated mice systemically by intraperitoneal injection. In both experiments, crenolanib did not decrease local MCP-1 mRNA expression or serum MCP-1 concentrations.

We reviewed 8 previous reports in which crenolanib was used to treat animals. In 2 studies, intraperitoneal injection was used (Wang et al., Neuroscience, 364:202-211, 2017; Zhang et al., J Proteome Res, 17:2401-2411, 2018). In 2 studies, subcutaneously implanted micropumps were used (Jogleker-Javadekar et al., Neoplasia, 19:564-573, 2017; Makino et al., J Invest Dermatol, 137:1671-1681, 2017). In two studies, crenolanib was orally administered (Ieronimakis et al., J Pathol, 240:410-424, 2016; Reichert et al., Cells, 10:804, 2021). Two studies (Wang et al., Oncotargets Ther, 7:1761-1768, 2014; Lin et al., Neurogastroenterol, 364:202-211, 2017) did not indicate the rout of administration. Thus, different routes have been used to treat animals with crenolanib, and we believe that intratumoral injection is acceptable. In response to the comment, we made a minor revision in the Results section (line 376-378). Thank you very much.

2. Besides the transgenic mice, E0771 in B6 mice can be another option to try.

(Our response)

We agree with the reviewer’s comment. Indeed, we are interested in the cells and preparing for an initial study. In response to the comment, we modified our sentence in the Discussion (line 508) and added references. Thank you very much.

Reviewer 3 Report

Thanks for your careful responses to my concerns. Every concern except one thing is resolved. 

Regarding your response number 6 to my previous comment: I did not mean to use the same terms in the text and figure legends. I just wanted to notice the inconsistency. In Figure legends of Figure 5, you wrote as following: Fig 5E - peripheral area, and Fig 5F - intratumoral area. However, in the text (line 406 of the revised manuscript) you described as following: '~both in the peripheral area (Fig. 5E, F) and ~'. Thus, is the Fig 5F peripheral area or intratumoral area? The same as, how about Fig 5G? Carefully recheck each picture in Figure 5. Especially, Fig 5F and Fig 5G.

Author Response

Our response to the comment

 

Comments and Suggestions for Authors

Thanks for your careful responses to my concerns. Every concern except one thing is resolved. 

Regarding your response number 6 to my previous comment: I did not mean to use the same terms in the text and figure legends. I just wanted to notice the inconsistency. In Figure legends of Figure 5, you wrote as following: Fig 5E - peripheral area, and Fig 5F - intratumoral area. However, in the text (line 406 of the revised manuscript) you described as following: '~both in the peripheral area (Fig. 5E, F) and ~'. Thus, is the Fig 5F peripheral area or intratumoral area? The same as, how about Fig 5G? Carefully recheck each picture in Figure 5. Especially, Fig 5F and Fig 5G.

(Our response) We really appreciate the reviewer’s careful reading and apologize for our misunderstanding of the comment. The reviewer is correct and there were mistakes in citing those figures. We have corrected the mistakes in the text (line 447-451). We now use “peritumoral” and “intratumoral” in the legend as they were before.

Taking advantage of this opportunity, we made minor changes to Figure 2A and Figure 3B. For Figure 2A, we assigned an arbitrary value of 1 to the expression levels of Mcp-1 and Kc mRNA at time 0 and recalculated the relative expression levels. We also added the results of statistical analyses. For consistency, we made a similar modification to Figure 3B. These modifications do not affect the conclusion of this study. Thank you.

Round 3

Reviewer 1 Report

Intratumoral administration should be avoided except you are doing a tumor ablation study. 

Author Response

We first thank the reviewer for reviewing our manuscript and providing us a comment. The reviewer raised a concern about the use of intratumoral (IT) injection of PDGFR inhibitor. We respectfully disagree with the comment. IT injection is a useful method to deliver agents directly into tumors and has been used clinically and in many animal studies.

In a recent paper published in the International Journal of Molecular Sciences (Bender et al., 2020, 21, 4493; doi:10.3390/ijms21124493), the authors used IT injection of anti-cancer agents to treat tumors in mice. In the paper, they nicely summarized the history of IT treatment in the Introduction. Here, I introduce a part of their description with minor modifications. “IT injection of chemotherapeutic agents was originally studied for treatment of tumors, theoretically allowing for higher doses to reach the tumor without an increase in systemic toxicity. Treatment approaches have recently shifted from killing the cancer cells to stimulating immune cells. This shift to immune-oncology (IO) treatment has reopened the investigations into intratumoral approaches focusing on activating local immune response (In our case, blocking the activation of fibroblasts with cancer cell-derived PDGFs). Indeed, a novel genetically modified oncolytic viral-based immunotherapeutic, talimogene laherparepvec (T-Vec), has been approved for IT use in cutaneous melanoma (US Food and Drug Administration. FDA Approves First-Of-Its-Kind Product for the Treatment of Melanoma 2015. updated 27 October 2015. Available online: https://wayback.archive-it.org/7993/20170722061524/https: //www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm469571.htm (accessed on 22 June 2020).” Thus, IT injection is a legitimate method to deliver agents directly into tumors. In our study, we examined an effect of PDGFR inhibition in the tumor microenvironment on the production of MCP-1 by delivering crenolanib solution directly into tumors.

According to the review article by Knavel and Brace, tumor ablation is a minimally invasive technique that is commonly used in the treatment of tumors of the liver, kidney, bone, and lung. During tumor ablation, thermal energy is used to heat or cool tissue to cytotoxic levels (less than −40°C or more than 60°C) (Knavel and Brace, Tech Vasc Interv Radiol. 2013, 16: 192–200. doi:10.1053/j.tvir.2013.08.002). The purpose of IT injection we used was completely different from that of tumor ablation.

In response to the reviewer’s comment, we have added a sentence with a reference by Bender et al. to our manuscript (line 390-391). We appreciate the comment.