The Connexin 43 Regulator Rotigaptide Reduces Cytokine-Induced Cell Death in Human Islets

Round 1

Reviewer 1 Report

Introduction must be improved. Please, add more reference to support the aim of the study.

Major doubts concern the figures.

In general they are different in the height of the axes and in the dimensions.

Symbols indicating significant differences are unclear and in some cases it is not clear which groups are compared (e.g. figure 1B). Furthermore, remove One Way Anova in the graph, indicate the statistical analysis carried out in the legend of the figure

Author Response

Response to the reviewer 1:

1. Introduction must be improved. Please, add more reference to support the aim of the study.

Response: The Introduction has now been updated with several relevant references (ref. nos. 11, 21, 22, 28) In particular, the relatively recent review by Farnsworth & Benninger in FEBS Lett 2014 (ref. no. 11) strongly supports the rationale of this study by underlining the science gaps in the knowledge about the role of Cx43 in the pancreatic islet and the lack of studies with Cx43 peptide mimetics such as Rotigaptide on islets.

2. Major doubts concern the figures. In general they are different in the height of the axes and in the dimensions.

Response: The quality of the figures has now been improved by aligning scaling, deleting redundant text, and by enlarging font size of captions and statistical symbols. 

3. Symbols indicating significant differences are unclear, and in some cases it is not clear which groups are compared (e.g. figure 1B). Furthermore, remove One Way Anova in the graph, indicate the statistical analysis carried out in the legend of the figure

Response: Symbols have now been adjusted as mentioned in the response to comment 2. In Figure 1B and other figures significance symbols have been replaced to more clearly indicate statistical comparisons. ’One-way ANOVA’ has now been moved to legends where relevant.   

Reviewer 2 Report

The authors need to clarify their meaning of Cx43 expression in human (adult) beta cells.  The literature does not support this notion and the data shown in the manuscript are insufficient proof.

In the abstract the authors state that ZP123 increases junctional conductance in heart but do not provide a reference.  If memory serves it causes about a 50-70% increase.  The authors speculate the "increase" allows for buffering of apoptotic triggers such as calcium in the human beta cells.  No proof is provided that junctional conductance between beta cells is increasing with ZP123. 

It is hard to read the text going back and forth from the results section when the figures are in the methods section. While figure 1a shows Cx43 message there is no demonstration of Cx43 protein distribution within cells. 

Fig 1a  message levels for Cx43 are low and  INS-1 cells expression is even less but they still are making some or else there would be no SE bars. How does this make the human expression significant. Is it there statistical significance between Cx43 for human and INS-1 cells. The labels along the X-axis for figure 1 and all others is mystifying.  There needs to be more clarity in the figure legends if not in the text. 

Author Response

Response to the reviewer 2:

1. The authors need to clarify their meaning of Cx43 expression in human (adult) beta cells. The literature does not support this notion and the data shown in the manuscript are insufficient proof.

Response: The reviewer is absolutely right that the literature on the role of Cx43 in the pancreatic islet is extremely scarce, in contrast to what is known about Cx36. This important science gap was highlighted in the review by Farnsworth & Benninger in FEBS Lett 2014 (Reference 11), strongly supporting the rationale of our study.  

We do believe that the two seminal papers from the  Meda lab, showing that Cx43 is expressed in pancreatic islets (Reference 14) and that Cx43 knockout mice demonstrate decreased beta cell mass and insulin secretion (Reference 10), strongly support that Cx43 has a functional role in the pancreatic beta-cell.

 In further response to the reviewer’s concern we have now also added references to the interesting observation that proinflammatory cytokines do downregulate Cx43 expression in several other cell systems.  

2. In the abstract the authors state that ZP123 increases junctional conductance in heart but do not provide a reference. If memory serves it causes about a 50-70% increase.  The authors speculate the "increase" allows for buffering of apoptotic triggers such as calcium in the human beta cells.  No proof is provided that junctional conductance between beta cells is increasing with ZP123.

Response: The reference is given in the introduction (23). We agree with the reviewer that this notion is speculative, and the statement has therefore been deleted in the Discussion.

3. It is hard to read the text going back and forth from the results section when the figures are in the methods section. While figure 1a shows Cx43 message there is no demonstration of Cx43 protein distribution within cells.

Response: The figures were placed within the text by the journal administrators. We have now placed figures correctly.  

We agree with the reviewer that verification of mRNA changes at the protein level and demonstration of the localization of changes to cell-membranes are important. Accordingly, we have added a note of caution on this in the Discussion under limitations (line 399).

4. Fig 1a message levels for Cx43 are low and INS-1 cells expression is even less but they still are making some or else there would be no SE bars. How does this make the human expression significant. Is there statistical significance between Cx43 for human and INS-1 cells. The labels along the X-axis for figure 1 and all others are mystifying.  There needs to be more clarity in the figure legends if not in the text.

Response: Cx43 mRNA was not detected in INS-1 cells, and error bars indicated background noise from the highest amplification cycles. They have now been removed, and accordingly it is not possible to compare the expression statistically with the low but detectable basal expression levels in human islets.

We apologize for the lack of clarity and now hope the revision has resolved this issue.

Round 2

Reviewer 2 Report

The authors have responded to all of my questions. Concerns still remains.

The authors contend that the literature provides proof of Cx43 expression in beta cells.  Farnsworth and Benninger state:"Previous studies have found that cells within the islet contain mRNA for Cx43 and Cx45 in mice and Cx30.3, Cx31, Cx31.1,Cx31.9, Cx37, Cx43 and Cx45 in human pancreas [7]." This refers to message, it is not a coupling assay.  It does not allow one to determine what channels are functional between two cells. Further, Carvallo et al., clearly have shown that in adult islets there is no Cx43 expression. 

It is not completely clear what the authors think ZP123 is doing. Increasing coupling?  If so how is that to be shown. 

Two additional issues need to be addressed:

1) the authors need to address Cx36 expression in the same way they have for Cx43. 

2) Of major concern is the lack of a rat wild type control. It would seem that it might be best to perform the analysis  on normal rat beta cells and compare that to INS-1 cells.  

Author Response

We thank the reviewers for their valuable comments. In addition to extensive literature search, we have in collaboration with Dr. Björn Tyrberg, now added as co-author of the MS, performed reanalysis of the expression of connexins in the single-cell RNAseq dataset from human islet cells published by Segerstolpe A et al Cell Metabolism 2016. This new data is added as supplementary information as Fig S1.

This has led to the following specific edits in the text:

1. 21 and 23, deletion of: beta-cell
2. 58, addition: in homo- or heterotypic combinations, each hemi-channel comprised…
3. 62-69: rephrasing: Cx43 was recently identified as an important regulator of beta-cell differentiation (17) and maturation (18), explaining why beta-cell specific knockout of Cx43, but not of Cx36, reduces pancreatic insulin content and islet size (19). However, Cx43 was not found to be expressed on adult beta-cells or insulin-producing cell lines (20-22). Of note, several studies have failed to show basal Cx43 expression in rat islets or insulin-producing cell lines (20-22). Interestingly, Cx43 and Cx32 are expressed on non-endocrine pancreatic cells, and heterotypic channels between e.g. endothelial cells and islet endocrine cells (23, 24) have been implicated in regulating glucose stimulated islet blood flow (25).
4. 240-247, rephrased and added: As shown in fig. 1a we found that human and rat islets did express Cx43, whereas INS-1 cells did not, neither in the unchallenged not the cytokine-challenged state. Reanalyzing published human islet single cell sequencing data (32) the expression of connexins apart from Cx36, well-known to be expressed in islet endocrine cells was determined (fig. S1). Endocrine alpha-, beta- and delta cells equally expressed Cx 31.9, and 32, acinar and ductal cells equally expressed Cx26, 31, 32 and 43, ductal cells also expressed Cx 30.3 and Cx45, pancreatic stellate cells expressed Cx31.9, 43 and 45, whereas endothelial cells were only found to express Cx43 (with the caveat that the number of passenger endothelial cells in cultured human islets is low).
5. 255-256, deleted: In contrast to intact rat islets/Thus. Added: Since
6. 428-433, added: Given the lack of beta-cell Cx43 expression shown by others (17-19, 22) and in our single-cell RNA seq analysis of human islets (fig. S1), but that heterotypic channels can be formed between Cx43 positive non-endocrine pancreatic cells and Cx43 negative islet endocrine cells expressing other connexins such as Cx36 or Cx32, we suggest that RG promotes coupling of Cx43 expressed on islet endothelial or pancreatic stellate cells to other islet endocrine cell connexins to form heterotypic channels that promote islet cell survival.
7. 449-454, added: This study does not demonstrate that RG mediated prevention of cytokine-induced apoptosis is beta-cell specific. Such studies would be demanding, since sorting of islet cells and even reassembly into beta-cell enriched pseudo-islets would eliminate interaction with Cx43 expressing non-endocrine cells. However, since cytokine-mediated islet cell apoptosis is selective for beta-cells (1,39) it is likely that the reduction of human islet cell apoptosis by RG is related to reduced beta-cell death.
8. 456-457, deleted: beta-cell specific overexpression of Cx43, but not of Cx36, increases pancreatic insulin content and islet size, added: Cx43 is required for beta-cell differentiation and maturation
9. 459: deleted: beta, added islet
10. 692-695, added: 4. 10. Single-cell RNA-seq of human pancreatic islets: The expression of genes in islet cell types was determined by reanalyzing published human islet single cell sequencing data (donor information in EBI accession number: MTAB-5061) (32) as previously described (47).
11. 706-707, altered: beta to islet
12. 715-719, edited: Author contributions: SMG and TMP were initiators of the study and developed the protocols for the experiments. SMG, JBH and DPC conducted the experiments. BT performed the bioinformatics analysis of single-cell RNA sequencing data. SMG performed the statistical analysis, constructed figures and tables and wrote the first draft of the manuscript. All authors discussed data and edited the manuscript.

Supplementary information: Added fig. 1S and legend.