Expression of Connexin 43 in Granular Cell Tumors of the Skin, Tongue and Esophagus

Round 1

Reviewer 1 Report

Well written retrospective limited study demonstrating the presence of variable cytoplasmic gap junction protein connexion 43 positivity in a cohort of 12 cases of GCT. Findings compared with neurofibroma and leiomyoma and BCC was a positive control.

The authors acknowledge the many markers already available to support what is often a straight forward diagnosis on H & E with at most PASD. They state "there was positivity for SOX10 (Fig. 2c), S100, CD68, and neuron-specific enolase (NSE) in all GCTs" 

In my view, therefore connexion 43 is NOT a novel diagnostic marker as claimed as it is too widespread in many tissues such as BCC, vascular lesions and those with follicular differentiation to make the claim as a potential diagnostic marker for GCT. Pity no cases of other neura/spindle celll tumours even those showing desmoplasia eg TE or Spitz were not examined and especially even rarer and perhaps most useful would be reactivity found in non neural and even malignant GCT were included as this would make help in the differential diagnosis.

I feel the authors can improve and elevate the paper significantly. I suggest including many more skin tumours and recommend more should be discussed on the role of connexion 43 positivity in skin and relationship to the pathogenesis of such lesions including GCT. I found it of interest that the gap junction protein iwas present n cytoplasmic phagolysosomes. Moreover this could be illustrated at high power. I do hope this is useful feedback. 

Author Response

Manuscript [ID: 2260761]: “Expression of Connexin 43 in granular cell tumors of the skin, tongue and esophagus” by Kneitz H et al.

Dear Reviewer I,

thank you for reviewing our manuscript “Expression of Connexin 43 in granular cell tumors of the skin, tongue and esophagus”.

We appreciate your comments and have revised our manuscript according to your recommendations. Please see the following point-by-point response.

1. In my view, therefore connexion 43 is NOT a novel diagnostic marker as claimed as it is too widespread in many tissues such as BCC, vascular lesions and those with follicular differentiation to make the claim as a potential diagnostic marker for GCT. Pity no cases of other neural/spindle cell tumours even those showing desmoplasia eg TE or Spitz were not examined and especially even rarer and perhaps most useful would be reactivity found in non neural and even malignant GCT were included as this would make help in the differential diagnosis.

Response: We succeeded in increasing the number of GCT to 22 cases. In addition to more GCT of the skin, we were able to also include GCT of the tongue (Figure 3) and esophagus (Figure 4) in our work. We think that we are now able to provide more reliable data regarding the expression of CX43 as well. In accordance with the comments of both reviewers, we refrain from describing CX43 as a useful diagnostic marker. For this reason, we also changed the title of the paper and deleted Figure 3. Nevertheless, we examined other tumors of differential diagnostic importance (e. g. melanomas, xanthogranulomas, xanthomas, Spitz nevi). All of these did not express CX43 as evaluated by immunohistochemistry. However, since we consider the significance of CX43 as a diagnostic marker for GCT to be low, we do not present these data in detail. Unfortunately, we were unable to study malignant or non-neuronal GCTs since neither in our archive nor in the archives of the Department of Pathology of the University of Würzburg such tissue samples had been available.

2. I feel the authors can improve and elevate the paper significantly. I suggest including many more skin tumours and recommend more should be discussed on the role of connexion 43 positivity in skin and relationship to the pathogenesis of such lesions including GCT.

Response: We have expanded the discussion regarding the importance of Cx43-positivity in skin and for the development of skin tumors. Regarding the importance of CX43 for the development of certain skin tumors, numerous literature references are cited. Recently, our research group was able to prove the possible importance of CX43 in the carcinogenesis of angiosarcomas (see reference 11).

3. I found it of interest that the gap junction protein was present in cytoplasmic phagolysosomes. Moreover this could be illustrated at high power.

Response: We attempted to visualize the tumor cells of GCT at a higher magnification (figure 1b,c). Unfortunately, it has not been possible for us to provide electron microscopic images.

We would be delighted if you deem our revised manuscript suitable for publication in Dermatopathology!

Sincerely,

Hermann Kneitz, MD

Department of Dermatology,

University Hospital Würzburg,

Josef-Schneider-Str. 2

97080 Würzburg Germany

Reviewer 2 Report

CX43 in GCT reviewing

In this work, the authors studied the expression of connexin 43 (Cx43) in 12 cases of cutaneous Granular Cell Tumor (GCT). They found moderate-to-strong expression in all these cases, comparatively with some other (mesenchymal) skin tumours.

The findings of this study are not uninteresting, however there are some issues that need improvement/correction:

General comments:

- The ‘introduction’ should be restructured. It would be more easy to the readers to understand the aim of the study if data on Cx43 and its role in (skin) carcinogenesis were explained before addressing the specific issue of GCT.

- The diagnostic value of Cx43 in the diagnosis of GCT is not substantiated by this study, as this marker is by no means specific of GCT. Therefore, relevant statements (lines 160-161, 196-197) should be tuned down. To assess the true diagnostic value of Cx43, the authors should have studied histologic mimics of GCT, namely tumors of other (non Schwannian) origin with a granular cell appearance (such as granular cell basal cell carcinoma, melanoma, atypical fibroxanthoma, dermatofibroma etc – i.e.  those listed lines 155-157), and in adequate/larger numbers (even though GCT is admittedly an uncommon tumour, a series of 12 cases does not have too strong a statistical power). Until a larger number of tumors with granular cells is studied, the diagnostic utility of Cx43 will remain doubtful.

Specific issues:

- Lines 8, 29, 143: the statement that GCT are of ‘neuronal’ differentiation is not correct; most (S100+) GCT are of Schwannian, not neuronal origin/differentiation.

- It is stated that normal skin was used as a negative control (lines 14-15, 73-74) but in line 46 it is stated that Cx43 is expressed both in the epidermis and the dermis. Which is true ? if Cx43 is expressed in normal skin, some sentences summarizing where this expression is seen would be useful.

- The quantification of immunostaining intensity needs some clarification. Does ‘weak, moderate, strong’ expression refer to the intensity of immunolabelling? Or the percentage of positive cells ? please clarify

- The discussion is too wordy, it could be condensed by focusing on the main findings of this study, i.e. the significance of Cx43 in GCT. The data presented in lines 134-138 are repetition of the results. Conversely, the result of PAS staining (line 139-141) should be displaced to the ‘results’ section (nothing is mentioned there on this staining). Considerations of clinical differential diagnosis (lines 150-152) are also not very relevant to this histopathologic study.

- The paper contains panels that are unnecessary as they do not add something important beyond the text (namely concerning the ‘negative’ stainings: panels 2d, 3d-i)

- line 125: the clefts around BCC nodules are not ‘artificial’ (you probably mean artefactual), they can be observed in vivo by reflectance confocal microscopy (see Ulrich et al. J Cutan Pathol 2011;38:190)  

- The English needs revision (eg ‘abundant localization’ sounds awkward- the title should state ‘supporting the diagnosis of…’, ‘subtle presentation: non specific presentation, ‘granulated’: granular, ‘stainability’: staining, ‘irritable’ hyperplasia (!) – you probably mean reactive hyperplasia ? etc)

Author Response

Manuscript [ID: 2260761]: “Expression of Connexin 43 in granular cell tumors of the skin, tongue and esophagus” by Kneitz H et al.

Dear Reviewer II,

thank you for reviewing our manuscript “Expression of Connexin 43 in granular cell tumors of the skin, tongue and esophagus”.

We appreciate your comments and have revised our manuscript according to your recommendations. Please see the following point-by-point response.

General comments:

1. The ‘introduction’ should be restructured. It would be more easy to the readers to understand the aim of the study if data on Cx43 and its role in (skin) carcinogenesis were explained before addressing the specific issue of GCT.

Response: Thank you for highlighting this important structural feature. Following the reviewer’s suggestion, we now first describe the importance of connexins/Cx43 and thereafter address specific issues of Cx43.

2. The diagnostic value of Cx43 in the diagnosis of GCT is not substantiated by this study, as this marker is by no means specific of GCT. Therefore, relevant statements (lines 160-161, 196-197) should be tuned down. To assess the true diagnostic value of Cx43, the authors should have studied histologic mimics of GCT, namely tumors of other (non Schwannian) origin with a granular cell appearance (such as granular cell basal cell carcinoma, melanoma, atypical fibroxanthoma, dermatofibroma etc. – i.e.  those listed lines 155-157), and in adequate/larger numbers (even though GCT is admittedly an uncommon tumour, a series of 12 cases does not have too strong a statistical power). Until a larger number of tumors with granular cells is studied, the diagnostic utility of Cx43 will remain doubtful.

Response: We succeeded in increasing the number of GCT to 22 cases. In addition to further GCTs of the skin we were also able to include GCTs of the tongue (Figure 3) and the esophagus (Figure 4) into our revised manuscript. We now expect to be able to make a better statistical statement regarding the expression of CX 43 as well. In accordance with the comments of both reviewers, we refrain from describing CX 43 as a useful diagnostic marker. For this reason, we also changed the title of the paper and deleted the former Figure 3. Nevertheless, we examined other tumors of differential diagnostic importance (melanomas, xanthogranulomas, xanthomas, Spitz nevi) in our laboratory. All showed a completely negative result in staining with CX43. However, since we consider the significance of Cx43 as a diagnostic marker for GCT to be low, we do not list these data.

Specific issues:

1. - Lines 8, 29, 143: the statement that GCT are of ‘neuronal’ differentiation is not correct; most (S100+) GCT are of Schwannian, not neuronal origin/differentiation.

Response: Thank you for highlighting this error; we adapted the paragraph accordingly.

2. - It is stated that normal skin was used as a negative control (lines 14-15, 73-74) but in line 46 it is stated that Cx43 is expressed both in the epidermis and the dermis. Which is true? if Cx43 is expressed in normal skin, some sentences summarizing where this expression is seen would be useful.

Response: This is an interesting comment that we gladly annotated. We therefore changed our controls and adapted figure 5 accordingly: (negative control: placenta, tonsillar tissue. Positive control: normal skin, basal cell carcinoma).

3. The quantification of immunostaining intensity needs some clarification. Does ‘weak, moderate, strong’ expression refer to the intensity of immunolabelling? Or the percentage of positive cells? please clarify.

Response: We primarily assessed the staining intensity and classified it into weak, moderate and strong (see Methods). In all cases there was "diffuse positivity" of all tumor cells, only "focal" positivity could not be detected in any of the cases.

4. The discussion is too wordy, it could be condensed by focusing on the main findings of this study, i.e. the significance of Cx43 in GCT. The data presented in lines 134-138 are repetition of the results. Conversely, the result of PAS staining (line 139-141) should be displaced to the ‘results’ section (nothing is mentioned there on this staining). Considerations of clinical differential diagnosis (lines 150-152) are also not very relevant to this histopathologic study.

Response: Since we believe integrating our results in the existing literature is necessary, we would like to keep this section almost unchanged. We have moved the PAS staining results to the "results" section. Furthermore, we he have deleted considerations regarding clinical differential diagnosis (lines 150-152).

5. The paper contains panels that are unnecessary as they do not add something important beyond the text (namely concerning the ‘negative’ stainings: panels 2d, 3d-i)

Response: We have deleted the former Figure 3.

6. Line 125: the clefts around BCC nodules are not ‘artificial’ (you probably mean artefactual), they can be observed in vivo by reflectance confocal microscopy (see Ulrich et al. J Cutan Pathol 2011;38:190)  

Response: We gladly accepted this suggestion for improvement.

7. - The English needs revision (eg ‘abundant localization’ sounds awkward- the title should state ‘supporting the diagnosis of…’, ‘subtle presentation: non specific presentation, ‘granulated’: granular, ‘stainability’: staining, ‘irritable’ hyperplasia (!) – you probably mean reactive hyperplasia? etc.)

Response: We carefully revised the manuscript as recommended.

We would be delighted if you deem our revised manuscript suitable for publication in Dermatopathology!

Sincerely,

Hermann Kneitz, MD

Department of Dermatology,

University Hospital Würzburg,

Josef-Schneider-Str. 2

97080 Würzburg Germany

Round 2

Reviewer 1 Report

Good thank you for asking me to review your resubmitted manuscript.

No doubt, he revised version is much improved and now in my view worthy of publication

Regards 

Author Response

Dear Reviewer I,

Thank you for the positive review of our manuscript. according to the suggestions ov Reviewer II, we have further improved our manuscript.

Best regards

Hermann Kneitz

Reviewer 2 Report

The revised version of this article has been improved, but can be improved further. Please consider the following issues:

The English still needs polishing for syntax and language errors (eg “occurring at the skin…” should read ‘occurring in the skin….’, ‘…is yet unknown…’ should read ‘…is as yet unknown…’, ‘unphysiological’ should read ‘non physiological’ or ‘pathological’, and typos (poly-gonal, Sox10), etc

Abstract: ‘The diagnosis was confirmed by H&E-staining’: this sentence is not necessary here; by the way, abbreviations should not be used in an abstract.

“Introduction”: it is stated that ‘increased membranous expression of Cx43 has been demonstrated in various skin tumours’ - this is not very precise: increased vs what? normal tissue? Other tumours? please be more specific, otherwise use another adjective (strong?)

‘Material and methods”: magnification x20 is probably x200 (the ocular lens of most microscopes I know is x10). In this section, please state which control tissues/tumors studied were positive or negative (this is stated further in the ‘results’ section, but should be clear in this section). The word ‘respectively’ at the end of this section is useless and should be deleted.

“Results”: the 3^rd § of is section is a bit confusing, starting with immunohistochemistry, then describing HE findings, then again immunohistochemistry. The first of these sentences should be deleted.

In this section, the precise localization of Cx43 staining on positive controls should be stated (eg membranous). Was there any gradient of expression within the epidermis? Were adnexa stained? other skin structures?

Figures: in my view there are too many figures. I suggest to present 1-2 photomicrographs of HES section along with one immunostaining showing Cx43 expression. Fig. 1c is redundant with fig. 2B. Fig. 2D is not necessary. It is not necessary to provide HES and IHC at different magnifications for GCT of the tongue and esophagus, one representative figure for each localization is largely sufficient.  

“Discussion”: this can be shortened as it contains repetitions with data from the introduction and the results (eg the beginning of the 3^rd §).

The sentence ‘’Furthermore, experimental data….and promote apoptosis’ needs a reference (I presume it is n° 33, which is not cited in the text)

In the ‘authors’ contributions’, the initials ‘VG’ are mentioned, but do not corresponding to any of the authors” names; I guess it should read ‘VF’ – please check.

The last article in the reference list (‘Sekimizu M et al. Genes Chromosomes Cancer. 2019;58(6):373-380’) should be numbered (I presume it is n° 46)

Author Response

Dear Reviewer II,

Thank you for re-evaluating our manuscript. We greatly appreciate your valuable feedback and have made the necessary revisions based on your recommendations. Please find our point-by-point reply below:

Point-by-point reply

1. The English still needs polishing for syntax and language errors (eg “occurring at the skin…” should read ‘occurring in the skin….’, ‘…is yet unknown…’ should read ‘…is as yet unknown…’, ‘unphysiological’ should read ‘non physiological’ or ‘pathological’, and typos (poly-gonal, Sox10), etc

We have thoroughly polished the English language, addressing syntax and language errors. We have corrected phrases such as "occurring at the skin" to "occurring in the skin," "is yet unknown" to "is as yet unknown," "unphysiological" to "non-physiological" or "pathological," and rectified typos (e.g., "poly-gonal" to "polygonal," "Sox10").

2. Abstract: ‘The diagnosis was confirmed by H&E-staining’: this sentence is not necessary here; by the way, abbreviations should not be used in an abstract.

In the abstract, we have removed the sentence "The diagnosis was confirmed by H&E staining."

3. Introduction”: it is stated that ‘increased membranous expression of Cx43 has been demonstrated in various skin tumours’ - this is not very precise: increased vs what? normal tissue? Other tumours?

In the introduction, we have provided more precise information, specifying that increased expression of Cx43 has been demonstrated in various skin tumors as compared to normal tissues and other tumor types. Furthermore, we have included another reference regarding the expression of CX 43 in basal cell carcinomas and squamous cell carcinomas.

4. ‘Material and methods”: magnification x20 is probably x200 (the ocular lens of most microscopes I know is x10). In this section, please state which control tissues/tumors studied were positive or negative (this is stated further in the ‘results’ section, but should be clear in this section). The word ‘respectively’ at the end of this section is useless and should be deleted.

In the "Material and Methods" section, we have clarified that the magnification x20 was a typographical error and have corrected it to x200. We have also specified in this section which control tissues/tumors were studied and whether they were positive or negative. Furthermore, we have removed the unnecessary use of the word "respectively" at the end of this section.

5. “Results”: the 3^rd of is section is a bit confusing, starting with immunohistochemistry, then describing HE findings, then again immunohistochemistry. The first of these sentences should be deleted.

We have made the improvements according to the specifications.

6. In this section, the precise localization of Cx43 staining on positive controls should be stated (eg membranous). Was there any gradient of expression within the epidermis? Were adnexa stained? other skin structures?

We have provided precise information regarding the localization of Cx43 staining on positive controls, including its membranous expression. We have also mentioned any gradient of expression within the epidermis and whether adnexa or other skin structures were stained.

7. Figures: in my view there are too many figures. I suggest to present 1-2 photomicrographs of HES section along with one immunostaining showing Cx43 expression. Fig. 1c is redundant with fig. 2B. Fig. 2D is not necessary. It is not necessary to provide HES and IHC at different magnifications for GCT of the tongue and esophagus, one representative figure for each localization is largely sufficient.  

We have reduced the figures to the total number of 3. For GCT of the tongue and esophagus, we now provide one representative figure for each localization. Additionally, Fig. 2D has been eliminated.

8. “Discussion”: this can be shortened as it contains repetitions with data from the introduction and the results (eg the beginning of the 3^rd).

We have shortened the "Discussion" section, eliminating any repetitions of data from the introduction and the results, particularly in the beginning of the third paragraph.

9. The sentence ‘’Furthermore, experimental data….and promote apoptosis’ needs a reference (I presume it is n° 33, which is not cited in the text)

We have added the necessary reference (number 33) to support the sentence "Furthermore, experimental data....and promote apoptosis."

10. In the ‘authors’ contributions’, the initials ‘VG’ are mentioned, but do not corresponding to any of the authors” names; I guess it should read ‘VF’ – please check.

The initials "VG" mentioned in the "Authors' Contributions" section were indeed an error. It should be "VF," and we have rectified it accordingly.

We thank you again for your thorough review. We hope that our revised manuscript now adequately addresses your comments and suggestions.

Sincerely

Hermann Kneitz