Antitumour Activity of the Ribonuclease Binase from Bacillus pumilus in the RLS40 Tumour Model Is Associated with the Reorganisation of the miRNA Network and Reversion of Cancer-Related Cascades to Normal Functioning
, Aleksandra V. Sen’kova 1,†, Alsu I. Nadyrova 3, Innokenty A. Savin 1, Andrey V. Markov 1, Vladimir A. Mitkevich 4, Aleksander A. Makarov 4, Olga N. Ilinskaya 3, Nadezhda L. Mironova 1,* and Marina A. Zenkova 1Round 1
Reviewer 1 Report
Comments to the Author:
miRNAs control a wide range of biological processes and its misregulation has been correlated with several pathologies including cancer. At the same time, the involvement of miRNAs in the eukaryotic response to bacterial and other pathogens is a well-established concept nowadays. Host miRNAs can either make part of the host defences to counteract infection and/or be used as a pathogen strategy to subvert host response.
Related to this, but with a different perspective, this manuscript by Mohamed et al. focuses on the antitumour activity of the exogenous ribonuclease Barnase from B. pumilis. The subject is not novel, with important discoveries in the past regarding the cytotoxic effects of this exoribonuclease in tumour cells. However, this work adds further progress on the impact of this ribonuclease on tumour cells through its RNA (miRNA) targets.
This is a highly complete, well written and very well presented study, however there are still certain questions that I would like to see addressed/clarified before recommending its publication.
Major comments:
- In Figure 1, I do not understand the doses used in the sequence of experiments performed. In A) is determined the IC50 as being 91 ug/ml. In B), a dose of 0.5 mg/ml is used for induction of apoptosis. In C) and D) a dose of 0.5 mg/kg is used (in vivo). In E) the dose for the analysis of miRNAs expression was 5/10 ug/ml. Despite I suspect that this difference may come from previous data, this must be well explained and justified in the results section. For instance, as the authors justify in page 9 (line 319) the dose used for miRNAs expression is much lower than IC50 to ensure viability, but how was this precise concentration achieved?
- Page 16, line 520: I miss more detail on why the authors state that “the dose used in vitro was many times higher than the dose used in vivo”. And the some question as above regarding the reason for the concentration used in vivo (I missed, for instance a reference to justify it).
- RT-qPCR results on miRNAs expression and statement on page 16 (line 533) “....was due to total RNA degradation including pre-miRNAs, pri-miRNAs and mature miRNAs”.: I would like to know if this statement is based on the inspection of the precursor miRNAs expression (pre- and pri-miRNAs) beyond the analysis of the levels of the mature miRNAs or pure speculation? Please clarify.
- Page 10 line 352: I do not understand where the data that is used for comparison come from (indicated by * and defined as “historical control”)? I do not agree that this data is directly comparable with the one of the present study, as seen from the difference in the results in columns corresponding to 1 and 1* (mg/kg). Please justify.
- In the interpretation of data in Fig. 5A, the authors first mention that all the six miRNAs analysed had a decreased level in serum except one (let 7-g), and in the figure 5A see a change in 5 different miRNA. However in line 431, the authors mention “ a reduction in the levels of four from six analysed miRNAs...”. Please clarify.
Minor comments:
- Page 2, line 43: Ref.1 is from 2009, please avoid the word “Recently”.
- Paragraph starting on line 51 could have improved writing to be clearer.
- Page 3, line 101: please correct “tetrasolim salt...”
- Page 6, line 238: Again, ref. 28 is from 2013. Please substitute “Recently obtained data...” by, for instance, “Previous data....”
- In Page 7, the top legend of panel B refers a dose of Binase 0.5 mg/kg, while in the text and figure legend the dose is indicated as 0.5 mg/ml. Please correct.
- Page 9, line 321: I suppose it is 5ug/ml instead of 5 mg/ml.
Author Response
Comments and Suggestions for Authors
Comments to the Author:
miRNAs control a wide range of biological processes and its misregulation has been correlated with several pathologies including cancer. At the same time, the involvement of miRNAs in the eukaryotic response to bacterial and other pathogens is a well-established concept nowadays. Host miRNAs can either make part of the host defences to counteract infection and/or be used as a pathogen strategy to subvert host response.
Related to this, but with a different perspective, this manuscript by Mohamed et al. focuses on the antitumour activity of the exogenous ribonuclease Barnase from B. pumilis. The subject is not novel, with important discoveries in the past regarding the cytotoxic effects of this exoribonuclease in tumour cells. However, this work adds further progress on the impact of this ribonuclease on tumour cells through its RNA (miRNA) targets.
This is a highly complete, well written and very well presented study, however there are still certain questions that I would like to see addressed/clarified before recommending its publication.
Major comments:
- In Figure 1, I do not understand the doses used in the sequence of experiments performed. In A) is determined the IC50 as being 91 ug/ml. In B), a dose of 0.5 mg/ml is used for induction of apoptosis. In C) and D) a dose of 0.5 mg/kg is used (in vivo). In E) the dose for the analysis of miRNAs expression was 5/10 ug/ml. Despite I suspect that this difference may come from previous data, this must be well explained and justified in the results section. For instance, as the authors justify in page 9 (line 319) the dose used for miRNAs expression is much lower than IC50 to ensure viability, but how was this precise concentration achieved?
IC50 of binase was shown to be 91.7 μg/ml with respect to RLS40 cells. For apoptosis measurement we used concentration of the binase corresponding to 5 IC50 that was 0.5 mg/mL (Fig. 1, B, C and D). In Fig. 1 (B) the dose 0.5 mg/kg was mistakenly indicated in the top of right panel. We corrected Figure 1. For RNA profiling we used concentrations of binase allowing monitoring of changes in miRNA expression which correspond to the very beginning of cytotoxic action of the binase. We added explanation concerning concentration of binase used in these experiments to the text.
Please, see lines 314-316 and lines 354-360.
- Page 16, line 520: I miss more detail on why the authors state that “the dose used in vitro was many times higher than the dose used in vivo”. And the some question as above regarding the reason for the concentration used in vivo (I missed, for instance a reference to justify it).
The clarification was introduced in the manuscript. Please, see Discussion, lines 580 – 584.
- RT-qPCR results on miRNAs expression and statement on page 16 (line 533) “....was due to total RNA degradation including pre-miRNAs, pri-miRNAs and mature miRNAs”.: I would like to know if this statement is based on the inspection of the precursor miRNAs expression (pre- and pri-miRNAs) beyond the analysis of the levels of the mature miRNAs or pure speculation? Please clarify.
Because we did not study the effect of binase on the level of pre-miRNA and pre-miRNA in the cells this phrase is overstatement.
We modified this part of the sentence. Please, see Discussion, lines 585-588.
- Page 10 line 352: I do not understand where the data that is used for comparison come from (indicated by * and defined as “historical control”)? I do not agree that this data is directly comparable with the one of the present study, as seen from the difference in the results in columns corresponding to 1 and 1* (mg/kg). Please justify.
We agree that it is incorrectly to compare MII data obtained in previous study with the one of the present study. We modified the figure 3 and its legend, and deleted comparison of these two set of data from the text. Please, see lines 392-396.
- In the interpretation of data in Fig. 5A, the authors first mention that all the six miRNAs analysed had a decreased level in serum except one (let 7-g), and in the figure 5A see a change in 5 different miRNA. However in line 431, the authors mention “ a reduction in the levels of four from six analysed miRNAs...”. Please clarify.
We have changed the text according to the presented data. Please, see lines 473 – 477.
Minor comments:
- Page 2, line 43: Ref.1 is from 2009, please avoid the word “Recently”.
Corrected.
- Paragraph starting on line 51 could have improved writing to be clearer.
Corrected.
- Page 3, line 101: please correct “tetrasolim salt...”
Corrected.
- Page 6, line 238: Again, ref. 28 is from 2013. Please substitute “Recently obtained data...” by, for instance, “Previous data....”
Corrected.
- In Page 7, the top legend of panel B refers a dose of Binase 0.5 mg/kg, while in the text and figure legend the dose is indicated as 0.5 mg/ml. Please correct.
Corrected.
- Page 9, line 321: I suppose it is 5ug/ml instead of 5 mg/ml.
Corrected.
Reviewer 2 Report
Peer review of the manuscript with the number biomolecules-969420 and the tilte “Antitumour activity of the ribonuclease binase from Bacillus pumilus in the RLS40 tumour model is associated with reorganisation of the miRNA network and reversion of cancer-related cascades to normal functioning”
The manuscript from Islam et al. details a study in which the authors show the effect of the ribonuclease binase from Bacillus pumilus in an in vitro and in vivo model. The authors used modified RLS40 cells to depict the mechanism of antitumor activity of binase. They performed in vitro assays to stablish the IC50 and cytotoxicity via apoptosis induction, and in vivo assays injecting the RLS40 cells in a mice model to show the antitumor efficacy of binase. To evaluate the effect of binase on miRNA levels, and with data from a previous publication, they selected a panel of six miRNAs and performed an expression analysis. Additionally, using bioinformatics tools they revealed a regulatory network of binase-susceptible miRNAs related to cancer hallmarks, such as cell proliferation, apoptosis, angiogenesis and invasion.
Methods
The methodology trailed by the authors follow a consistent timeline of experiments and reasoning behind, and it is supported by their previous work and publications. However, there are some omissions or questions to be addressed. In the in vitro assays, the authors used different concentrations of binase for the experiments. The authors established the IC50 in 91.7ug/ml, nonetheless, for the expression analysis they used a dose 10 times lower, and for apoptosis detection, a dose 5 times higher than the IC50, respectively. Furthermore, it would be interesting to know the reasoning behind only evaluate a one-time point, 48 hours. The authors employed NGS data from one of their previous publication to obtain the six-miRNA panel for miRNA analysis, however, it was done in a Lewis Lung Carcinoma model (LLC), differently from the RLS40 model used in the present manuscript. These are different cellular stirpes and miRNAs are tissue specific. As stated in the LLC manuscript, the study was done with the bovine pancreatic ribonuclease RNase A. This ribonuclease exerts their activity by cleavage of extracellular RNAs and cleaves specifically after pyrimidine nucleotides, differently from binase. Binase cleaves intracellular and extracellular RNAs and is guanine-specific, a fact the authors emphasize as the probable explanation for the discrepancies in results with miR-145a and miR-10b. Altogether, this could explain the discrepancies showed in the expression analysis in the in vitro and in vivo model.
Concept
The exogenous ribonuclease binase from Bacillus pumilis has been shown to retard tumor growth and inhibit metastasis in several mouse models by cleavage of intracellular RNAs and initiation of the extrinsic apoptosis pathway. Nevertheless, it is poorly understood the direct effect of binase on its RNA targets. Thereby, the novelty of the study relies on the implication of binase-susceptible miRNA networks in the mechanism of its antitumor effect.
Strengths
The authors clearly showed the effect of binase in a preclinical model (RLS40-bearing mice). They showed a significant decrease in tumor size, inhibition of metastasis, a decrease in the numerical density of mitosis and PCNA-positive cells, induction of apoptosis, and reduction of areas of inflammation in tumors without compromising the liver or kidney function of mice. Remarkably, the authors depicted a miRNA network susceptible to binase. The involved pathways are related to carcinogenesis, progression, invasion and metastasis.
Concerns with suggestions
As mentioned above, the novelty in the studio relies on the effect of binase in miRNA networks. However, we cannot make a comparison one to one with the data from this manuscript with the data from the NGS analysis in the previous manuscript. The model is different, and even though both are Ribonucleases, RNase A and binase are known to act through different mechanisms. As a suggestion, the authors could perform an RNAseq analysis using the samples from the present study to strengthen their conclusions.
The expression levels of the selected miRNAs in Figure 5, were normalized to U6. U6 has been shown to be unsuitable for normalization in blood serum.
The authors showed a dynamic curve of tumor growth in Figure 2B, with a volume difference of almost 1cm3 between control and treatment, that would be enough to show macroscopically with pictures the differences between both conditions.
In figure 3A, the authors used a historical control from a previous publication, establishing the MII % for the 1mg/kg dose in 50%, but in the present study, with the same dose and methodology, they obtained an MII% of 86.4%. Additionally, it is poorly explained the reason for the low dose of binase (.5mg/kg) displayed a stronger effect in comparison with the high dose (1mg/kg).
To evaluate the toxicity of binase, the authors properly evaluated the Biochemical paramets to assess liver and kidney toxicity. It would be interesting to have a complete blood count (CBC) to strength the positive results and avoid possible biases.
The authors should double-check their references, for example, in lines 510 and 512, the authors claimed that their data is consistent with previously obtained data indicating the activation of apoptotic responses by binase in human A549 cells, referencing a publication that does not back up this statement (48).
Author Response
Referee 2
Comments and Suggestions for Authors
Peer review of the manuscript with the number biomolecules-969420 and the tilte “Antitumour activity of the ribonuclease binase from Bacillus pumilus in the RLS40 tumour model is associated with reorganisation of the miRNA network and reversion of cancer-related cascades to normal functioning”
The manuscript from Islam et al. details a study in which the authors show the effect of the ribonuclease binase from Bacillus pumilus in an in vitro and in vivo model. The authors used modified RLS40 cells to depict the mechanism of antitumor activity of binase. They performed in vitro assays to establish the IC50 and cytotoxicity via apoptosis induction, and in vivo assays injecting the RLS40 cells in a mice model to show the antitumor efficacy of binase. To evaluate the effect of binase on miRNA levels, and with data from a previous publication, they selected a panel of six miRNAs and performed an expression analysis. Additionally, using bioinformatics tools they revealed a regulatory network of binase-susceptible miRNAs related to cancer hallmarks, such as cell proliferation, apoptosis, angiogenesis and invasion.
Methods
The methodology trailed by the authors follow a consistent timeline of experiments and reasoning behind, and it is supported by their previous work and publications. However, there are some omissions or questions to be addressed. In the in vitro assays, the authors used different concentrations of binase for the experiments. The authors established the IC50 in 91.7ug/ml, nonetheless, for the expression analysis they used a dose 10 times lower, and for apoptosis detection, a dose 5 times higher than the IC50, respectively.
IC50 of binase was shown to be 91.7 μg/ml with respect to RLS40 cells. For apoptosis measurement we used concentration of the binase corresponding to 5 IC50 that was 0.5 mg/mL (Fig. 1, B, C and D). In Fig. 1 (B) the dose 0.5 mg/kg was mistakenly indicated in the top of right panel. We corrected Figure 1. For RNA profiling we used concentrations of binase allowing monitoring of changes in miRNA expression which correspond to the very beginning of cytotoxic action of the binase. We added explanation concerning concentration of binase used in these experiments to the text.
Please, see lines 314-316 and lines 354-360.
Furthermore, it would be interesting to know the reasoning behind only evaluate a one-time point, 48 hours.
Based on our previous experiments with binase we chose time point 48 h for IC50 and apoptosis measurement to have firstly possibility (i) to compare cell viability and apoptosis induction level, and (ii) because this time point correspond to rather high and reliable levels of apoptosis induction.
The authors employed NGS data from one of their previous publication to obtain the six-miRNA panel for miRNA analysis, however, it was done in a Lewis Lung Carcinoma model (LLC), differently from the RLS40 model used in the present manuscript. These are different cellular stirpes and miRNAs are tissue specific. As stated in the LLC manuscript, the study was done with the bovine pancreatic ribonuclease RNase A. This ribonuclease exerts their activity by cleavage of extracellular RNAs and cleaves specifically after pyrimidine nucleotides, differently from binase. Binase cleaves intracellular and extracellular RNAs and is guanine-specific, a fact the authors emphasize as the probable explanation for the discrepancies in results with miR-145a and miR-10b. Altogether, this could explain the discrepancies showed in the expression analysis in the in vitro and in vivo model.
Thank you so much for very valuable comment. We entirely agree with you that miRNA expression is tissue-specific. Therefore, we chose 6 miRNAs to provide some overlap between miRNA profiles of these tissues (LLC of epithelial origin and RLS40 of hematopoietic origin) in the case if some of the miRNAs are not expressed at the proper level in RLS40 cells. We demonstrated that these 6 miRNA are expressed in RLS40 at the pronounced level (Table 1, Fig. 1, E). We added comment of this point to the text. Please, see lines 334 – 337.
The effect of bovine pancreatic RNase A on the miRNA profile of LLC model in vitro and in vivo was differed from the effect of binase on the same miRNA in RLS40 model. And we agree with you that the reasons of these are: (1) pyrimidine-X specificity of RNase A and guanine-X specificity of binase; (2) targets of RNase A are extracellular RNAs whereas targets for binase are both intracellular and extracellular RNAs. We added comment of this point to the text. Please, see Discussion, lines 594 – 601.
Concept
The exogenous ribonuclease binase from Bacillus pumilis has been shown to retard tumor growth and inhibit metastasis in several mouse models by cleavage of intracellular RNAs and initiation of the extrinsic apoptosis pathway. Nevertheless, it is poorly understood the direct effect of binase on its RNA targets. Thereby, the novelty of the study relies on the implication of binase-susceptible miRNA networks in the mechanism of its antitumor effect.
We absolutely agree that the novelty of our study relies on the implication of binase-susceptible miRNA networks in the mechanism of antitumor activity of the binase. We believe that this part of our study shed some light on the intrinsic events underlying antitumor/antimetastatic activity of binase and possibly other bacterial ribonucleases.
Strengths
The authors clearly showed the effect of binase in a preclinical model (RLS40-bearing mice). They showed a significant decrease in tumor size, inhibition of metastasis, a decrease in the numerical density of mitosis and PCNA-positive cells, induction of apoptosis, and reduction of areas of inflammation in tumors without compromising the liver or kidney function of mice. Remarkably, the authors depicted a miRNA network susceptible to binase. The involved pathways are related to carcinogenesis, progression, invasion and metastasis.
We thank Referee 2 for the thorough review of our manuscript, as well as for the positive and encouraging comment and evaluation of our study.
Concerns with suggestions
As mentioned above, the novelty in the studio relies on the effect of binase in miRNA networks. However, we cannot make a comparison one to one with the data from this manuscript with the data from the NGS analysis in the previous manuscript. The model is different, and even though both are Ribonucleases, RNase A and binase are known to act through different mechanisms. As a suggestion, the authors could perform an RNAseq analysis using the samples from the present study to strengthen their conclusions.
We entirely agree with this comment. However, RNAseq requires more time than the PCR that is why we did not use this technique in this particular work, although we kept in mind the need for such an analysis. Such work is currently planned within the frame of another project.
The expression levels of the selected miRNAs in Figure 5, were normalized to U6. U6 has been shown to be unsuitable for normalization in blood serum.
Thank you for this comment. This is a mistake in description of presented data. The expression of tumour-derived miRNAs was rated relatively to U6 and the concentration of serum-derived miRNAs was normalized to serum volume.
Corrected. Please, see Materials and Methods, lines 194 – 199, and Figure 5 legend, lines 495 – 497.
The authors showed a dynamic curve of tumor growth in Figure 2B, with a volume difference of almost 1cm3 between control and treatment, that would be enough to show macroscopically with pictures the differences between both conditions.
Actually, the tumour volumes in control and experimental groups on the final day of the experiment are macroscopically differed. But since we planned to find statistically significant differences between the groups receiving binase 0.5 mg/kg and 1 mg/kg we decide to present dynamic curve of tumor growth.
In figure 3A, the authors used a historical control from a previous publication, establishing the MII % for the 1mg/kg dose in 50%, but in the present study, with the same dose and methodology, they obtained an MII% of 86.4%. Additionally, it is poorly explained the reason for the low dose of binase (.5mg/kg) displayed a stronger effect in comparison with the high dose (1mg/kg).
According to the comment of Referee 1 we exclude historical control from Figure 3 and the part of result description from the text because it is incorrectly to compare MII data obtained in previous study with the one of the present study. We modified figure 3 and its legend, and deleted comparison of these two set of data from the text. Please, see lines 392-396.
As you can see from the historical values of MII, although MII (MEAN±SEM)=50±17% for binase dose 1 mg/kg and MII=47±22% for binase dose 5 mg/kg. Taking this fact into account, the MII obtained in this work for binase dose 1 mg/kg was 80.4±7% that did not differ with statistic significance from the historical control. Because we have no statistically significant difference between binase doses 0.5 and 1 mg/kg, we entirely agree that this part of explanation should be excluded from the text.
Please see removed text, lines 392-396. Also we modified a part of Discussion, see lines 563-572.
To evaluate the toxicity of binase, the authors properly evaluated the Biochemical paramets to assess liver and kidney toxicity. It would be interesting to have a complete blood count (CBC) to strength the positive results and avoid possible biases.
We have data about complete blood count (CBC) for control and two experimental groups. But we observed no differences in this parameter.
We added comment on this point to the text. Please see Discussion, lines 545-546.
The authors should double-check their references, for example, in lines 510 and 512, the authors claimed that their data is consistent with previously obtained data indicating the activation of apoptotic responses by binase in human A549 cells, referencing a publication that does not back up this statement (48).
Corrected.
