Transcriptomic Analysis of Insulin-Secreting Murine Hepatocytes Transduced with an Integrating Adeno-Associated Viral Vector

Round 1

Reviewer 1 Report

I carefully read and reviewed the manuscript titled Correlation of gene expression changes and cell signalling pathways in insulin-expressing mouse liver cells transduced with an integrating adeno-associated viral vector. It is a well written paper, in general. Abstraction of the text is adequate. Keywords are relevant. Rationale and objectives are clear in introduction with providing sufficient background data. Methodology is clear. Only defect is statistics were briefly mentioned however must be stated comprehensively. Presentation of the results and discussion are fair enough. I, therefore, recommend acceptance of the paper for publication in the journal after minor revision.

Language is just fine. No issues other than minor spelling errors are detected.

Author Response

Thank you for your comments.  More information on the statistical analyses to compare each treatment to normal has been added to section 2.7 of the methods. Additionally, references to the CLC workbench manual, which outlines the statistical analyses carried out within the software package, have been added to section 2.4 of the methods.

Reviewer 2 Report

Dear authors,

I carefully checked your paper and it was intriguing. However, there are some features that should be amended before publishing this paper. Please check my comments and revise the manuscript accordingly. Some comments may look optional and the respected authors may not be willing to address the given points; however, it is better for the respected authors to consider the points to enhance the quality of the paper and make the manuscript content perfect for enlightening the quality of the literature.

1- Please correct all grammatical errors and misspelled words within the paper.

2- Please check all references in detail and omit all problematic papers (e.g., retracted records or papers with published expressions of concern on them) if any.

3- Please provide a graphical flowchart for the M&M section.

4- Please cite all protocols used in this study to obtain the reported data

5- The manuscript title is too long. Please revise the title and make it shorter if possible.

6- The introduction section is well-designed; however, please summarize some paragraphs in shorter lines. Indeed, please add global statistics of T1D prevalence based on IDF's recent report. Please use updated references for each paragraph that was discussed in this section of the paper.

7- Unlike other viral vectors, lentiviral vectors are difficult to produce using stable cell lines. Please add some lines about the biological challenges of the described vectors in the manuscript in the introduction section.

8- Please provide a list of abbreviations for all abbreviated terms within the paper and insert the list at the end or at the beginning of the paper.

9- Lines 87-98: Please shorten the paragraph and summarize the text in short lines. Indeed, the last three paragraphs can be summarized in one paragraph in the introduction section.

10- Please supplement Q10, Q20 and Q30 Phred scores for sequenced RNA samples.

11- One of the most important features of RNASeq projects is to evaluate the batch effect among samples. Please clarify how the respected authors evaluated the batch effect among the obtained RNA-Seq reads.

12- Before subjecting RNASeq data to the CLC workbench it is important all obtained FASTQ reads are subjected to specific tools to trim their content and remove adaptor fragments if any. In the M&M section, the esteemed authors have not clearly mentioned that how the sequenced samples were trimmed and which tools were used for this case. Indeed, in the supplementary files, the graphs of quality control steps were not supplemented. Please index these graphs too.

13- Please determine which method was followed to sequence RNA samples (single-end or paired-end)? And which normalization method (TPM, RPKM, FPKM) was used to normalize the obtained data? Please add such relevant information to the related section in M&M part. Please add sequencing depth (or coverage), total number of reads after trimming to this section too. The respected authors should clearly mention such necessary information in this section to enable academic readers in reproducing your data.

14- There are several freely available open-source tools such as NGS-QC toolkit (for quality assessment), trimmomatic (for pre-processing of reads and trimming), Bowtie, TopHat2, STAR (for read alignment), Trinity, Cufflinks (for read assembly), HTSeq-count, FeatureCount(for expression quantification), DESeq2, EdgeR (for differential expression analysis). None of these tools were used in this study. Is there any specific reason to use CLC software instead of the above-mentioned tools which can produce perfect and high-quality graphs for each section of RNASeq analysis?

15- Please add the Volcano plot to the results for DEGs. The respected authors can use ggpubr package in R to draw this plot.

16- Figure 2 details should be supplemented. Please add the list of DEGs for each group represented in Figure 2.

17- Please add an error bar to all statistical graphs discussed in the manuscript.

18- Please add a concluding remark at the end of the discussion. Please also highlight the major finding of this paper in this section.

19- Figures 2, 4, 6, and 8 can be merged and rearranged in one figure.

20- How the respected authors validated the successfully deliver of the studied vector to target cells? Is there any experimental method in your laboratory archive to confirm the successful delivery of this vector? If yes please supplement the data.

 

Briefly, the manuscript structure and data representation have enough quality to be considered for publication. However, the respected authors should address the given points and revise their manuscript accordingly.  Wishing you all the best. 

The English language is good. 

Author Response

1- Please correct all grammatical errors and misspelled words within the paper.

The paper has been checked for typographical and grammatical errors.

2- Please check all references in detail and omit all problematic papers (e.g., retracted records or papers with published expressions of concern on them) if any.

The references have been checked.

3- Please provide a graphical flowchart for the M&M section.

A flow chart of the methododology has been added as a supplementary figure S1

4- Please cite all protocols used in this study to obtain the reported data

All laboratory-based methods utilizing commercial kits/systems were carried out according to the manufacturers’ protocols as stated in the methods.

Methodology for RNA-seq quality control analyses and differential gene expression analysis were undertaken in CLC Workbench, a software package specific for this purpose. This software package has now been referenced with a link to the online manual in section 2.4 of the methods.

5- The manuscript title is too long. Please revise the title and make it shorter if possible.

The title of the manuscript has now been amended.

6- The introduction section is well-designed; however, please summarize some paragraphs in shorter lines. Indeed, please add global statistics of T1D prevalence based on IDF's recent report. Please use updated references for each paragraph that was discussed in this section of the paper.

The introduction has been shortened, however the authors feel the remaining information in this section is important in understanding the context of this work.

Global statistics of T1D prevalence have now been added to the start of the introduction.

7- Unlike other viral vectors, lentiviral vectors are difficult to produce using stable cell lines. Please add some lines about the biological challenges of the described vectors in the manuscript in the introduction section.

We have added a brief comment on biological challenges associated with viral vectors. A review paper that discusses this more in depth has also been referenced in the text.

8- Please provide a list of abbreviations for all abbreviated terms within the paper and insert the list at the end or at the beginning of the paper.

An abbreviation list has been included at the end of the paper.

9- Lines 87-98: Please shorten the paragraph and summarize the text in short lines. Indeed, the last three paragraphs can be summarized in one paragraph in the introduction section.

We have shortened this section, removing results that are detailed and discussed elsewhere. The remainder of this section is important to an understanding of the study context.

10- Please supplement Q10, Q20 and Q30 Phred scores for sequenced RNA samples.

This is now included in Supplementary Material 2.

11- One of the most important features of RNASeq projects is to evaluate the batch effect among samples. Please clarify how the respected authors evaluated the batch effect among the obtained RNA-Seq reads.

The analysis of the batch effect, the authors have produced principal component analysis (PCA) plots and heat maps using CLC Workbench. These have not been included in the publication as they did not show any significant differences.

12- Before subjecting RNA-Seq data to the CLC workbench it is important all obtained FASTQ reads are subjected to specific tools to trim their content and remove adaptor fragments if any. In the M&M section, the esteemed authors have not clearly mentioned that how the sequenced samples were trimmed and which tools were used for this case. Indeed, in the supplementary files, the graphs of quality control steps were not supplemented. Please index these graphs too.

The trimming process was completed by the Ramaciotti Centre, a commercial sequencing facility. They provided us with trimmed FastQ RNA-Seq reads ready for analysis as part of their service. Their process for trimming includes removing low quality bases and adapter sequences.

Quality control graphs for RNA-Seq analysis (read count statistics, spike-in quality control, strand specificity, etc.) are available in Supplementary Material 2.

13- Please determine which method was followed to sequence RNA samples (single-end or paired-end)? And which normalization method (TPM, RPKM, FPKM) was used to normalize the obtained data? Please add such relevant information to the related section in M&M part. Please add sequencing depth (or coverage), total number of reads after trimming to this section too. The respected authors should clearly mention such necessary information in this section to enable academic readers in reproducing your data.

The RNA samples underwent single-end sequencing. The FastQ files were normalised using the TMM (trimmed mean of M values) normalisation in CLC Workbench and this information has been added into section 2.4 of the methods. RNA-Seq quality analysis (read count statistics, spike-in quality control, strand specificity, etc.) are available in Supplementary Material 2.

14- There are several freely available open-source tools such as NGS-QC toolkit (for quality assessment), trimmomatic (for pre-processing of reads and trimming), Bowtie, TopHat2, STAR (for read alignment), Trinity, Cufflinks (for read assembly), HTSeq-count, FeatureCount(for expression quantification), DESeq2, EdgeR (for differential expression analysis). None of these tools were used in this study. Is there any specific reason to use CLC software instead of the above-mentioned tools which can produce perfect and high-quality graphs for each section of RNASeq analysis?

CLC workbench is a well-recognised software package for RNA-seq analysis, among other analysis types, that has been used in many published research articles. Our research group possesses a license for this software and continue to utilise it for this this publication and other ongoing work.

15- Please add the Volcano plot to the results for DEGs. The respected authors can use ggpubr package in R to draw this plot.

Volcano plots have been added as a supplementary figure (Figure S3).

16- Figure 2 details should be supplemented. Please add the list of DEGs for each group represented in Figure 2.

The FastQ files and the complete list of differentially expressed genes of each treatment compared to normal have been published on NCBI (GEO database), with accession number GSE235925.

17- Please add an error bar to all statistical graphs discussed in the manuscript.

Error bars are included on the RT-qPCR data plots, however some are too small to be seen or plotted. There are no error bars on the RNA-Seq data plots as these represent a single value.

18- Please add a concluding remark at the end of the discussion. Please also highlight the major finding of this paper in this section.

This has been added at the end of the discussion.

19- Figures 2, 4, 6, and 8 can be merged and rearranged in one figure.

The authors feel it is important to have these in separate sections as each of the figures is related to the different findings of the paper, detailed in different sections of the results.

20- How the respected authors validated the successfully deliver of the studied vector to target cells? Is there any experimental method in your laboratory archive to confirm the successful delivery of this vector? If yes please supplement the data.

A brief comment regarding validation of successful delivery of the vector has been added to section 2.1 of the methods section. The evidence for successful delivery of the vector is covered in the publication by La et al, which is referenced in the paper.

Reviewer 3 Report

I have read and analyzed the manuscript from Mahoney and coauthors. In my opinion, the manuscript is devoted to a very interesting approach to T1D treatment. The manuscript can be considered for publication but I have some critical points.

1. Which data was obtained in metabolism of mice after AAV therapy? It should be considered at least.

2. I request the addition of the gene therapy scheme and the verification of creation of correct AAV viruses (sequencing data of cloning INS FUR in piggybac).

3. Did authors control the presence of insulin producing cells in the liver by histology?

4. Why did authors use female mice? Male mice are more canonical for metabolic studies.

5. Which statistical method author used for the intergroup comparisons?

6. Did authors measure hepatic glucose production by the liver after therapy? It can significantly improve the study.

Author Response

1. Which data was obtained in metabolism of mice after AAV therapy? It should be considered at least.

Intraperitoneal Glucose Tolerance Tests (IPGTT) were undertaken in the study conducted by La et al, which showed stable blood glucose in the treated mice. Conducting an IPGTT on treated NOD mice is a common method utilised in many studies concerning diabetes to analyse metabolic behaviour in the animals.

I request the addition of the gene therapy scheme and the verification of creation of correct AAV viruses (sequencing data of cloning INS FUR in piggybac).

Sequencing data for the AAV8-INS-FUR vector has been added as supplementary data.

2. Did authors control the presence of insulin producing cells in the liver by histology?

Yes, immunohistochemistry was undertaken on the livers to verify the presence of the INS-FUR gene. This was undertaken in the study conducted by La et al and the results can be seen in that publication.

3. Why did authors use female mice? Male mice are more canonical for metabolic studies.

Use of female NOD mice is common practice in studies concerning Type 1 Diabetes as diabetes is more prevalent in the females and develops at an earlier stage of life than males. More information on this can be found in this study: “Chen, D., Thayer, T. C., Wen, L., & Wong, F. S. (2020). Mouse models of autoimmune diabetes: the nonobese diabetic (NOD) mouse  Animal Models of Diabetes: Methods and Protocols, 87-92.”

4. Which statistical method author used for the intergroup comparisons?

A differential expression analysis of each treatment compared to normal was completed. Differential expression analysis was done using the Wald test which includes statistical analysis to identify genes that are differentially expressed between two treatments, i.e AAV8/piggyBac-INS-FUR treated mice and untreated mice. This includes calculating the fold change of genes in the sequencing data and their significance, as well as making a prediction on the activation state of pathways and molecules, presented as an activation Z-score.

5. Did authors measure hepatic glucose production by the liver after therapy? It can significantly improve the study.

No, this study did not assess hepatic glucose production. This may be useful to do in the future; however this is not common practice in gene therapy studies on insulin producing cells.

Reviewer 4 Report

Comments to the Author

In this manuscript, Mohoney et al. examined the liver transcriptome in AAV8/piggyBac-INS-FUR-transduced NOD mice. With bioinformatics and molecular evidence, the authors believed that AAV8/pig-14 gyBac-INS-FUR vector could improve liver metabolic function in T1D mouse model. It is a very interesting study; however, I have few concerns:

 

1. By checking PPARs mRNA expression, the authors conclude that liver function is maintained in AAV8/pig-14 gyBac-INS-FUR treated mice. I think it is overinterpreted. Can the authors check the AST, ALT, and Albumin levels in the blood of control and AAV8/pig-14 gyBac-INS-FUR treated mice to support their conclusion?

 

2. The presentation of data can be improved. For example, in figures 4,6, 8, the authors could add gene names as the title of each figure (at the top of each figure).

 

3. AAV8/pig-14 gyBac-INS-FUR was shown by the authors to be a promising strategy to treat T1D diabetes. The authors may also discuss the potential of this strategy to treat T2D diabetes, since many T2D diabetes patients need insulin application to maintain normoglycemia.   

 

4. previous work conducted by La et al [10] showed that AAV8/pig-14 gyBac-INS-FUR mice had much higher insulin levels than normal control mice. It is well known that hyperinsulinemia may cause weight gain. Does AAV8/pig-14 gyBac-INS-FUR mice showed any body weight changes compared to untreated mice. The authors may discuss about the potential side effects of the hyperinsulinemia observed in AAV8/pig-14 gyBac-INS-FUR treated mice.

 

Author Response

There was very limited unpublished AST and ALT measurements performed in the study by La et al of the different tissues. An example has been added to the discussion. We are unable to check the AST/ALT levels further as we no longer have the mice.

2. The presentation of data can be improved. For example, in figures 4,6, 8, the authors could add gene names as the title of each figure (at the top of each figure).

The gene names have been added to each figure.

3. AAV8/pig-14 gyBac-INS-FUR was shown by the authors to be a promising strategy to treat T1D diabetes. The authors may also discuss the potential of this strategy to treat T2D diabetes, since many T2D diabetes patients need insulin application to maintain normoglycemia.

This has been mentioned in the introduction section, however it is mainly T1D treatment, particularly hypoglycaemia unaware patients who would be considered at least initially for this treatment.

4. previous work conducted by La et al [10] showed that AAV8/pig-14 gyBac-INS-FUR mice had much higher insulin levels than normal control mice. It is well known that hyperinsulinemia may cause weight gain. Does AAV8/pig-14 gyBac-INS-FUR mice showed any body weight changes compared to untreated mice. The authors may discuss about the potential side effects of the hyperinsulinemia observed in AAV8/pig-14 gyBac-INS-FUR treated mice.

In Fig. 5C of the La et al paper, we are measuring levels of human insulin and therefore as expected the values for normal mice were zero.

Round 2

Reviewer 2 Report

Dear authors, 

I checked your revision and it is now suitable for publication. Great congratulations on your achievement and I have no further comments on this manuscript. 

Best regards, 

Rasouli. H

Author Response

Thanks very much.

Reviewer 3 Report

Many thanks to the authors for the comprehensive response. In my opinion, general data about animal model which are presented in the La and coauthors study must be added in the manuscript, but let it be up to the editor.

Author Response

We have added a small amount of information on page 2, clearly pointing out that the mice treated with the AAV8 vector expressing the INS-FUR gene did not normalize blood glucose and reverse diabetes. We have also clearly pointed out that the AAV8/piggybac-INS-FUR system as well as reversing diabetes had normal intraperitoneal glucose tolerance tests. However, otherwise the authors feel the animal model from the La et al study has been sufficiently summarised in the introduction of the paper. If the reader requires further information on our work using the animal model, this information is freely available in the La et al published study. Information about the NOD model itself is available in innumerable publications.

Reviewer 4 Report

The authors have addressed my concerns!

Author Response

Thanks very much