Cancer Stem Cells and the Slow Cycling Phenotype: How to Cut the Gordian Knot Driving Resistance to Therapy in Melanoma

Round 1

Reviewer 1 Report

The manuscript is very well organized after redesign and completion and describes the issue of stem cells in the matter of melanoma.

This study was very properly completed and represented a comprehensive work.

Reviewer 2 Report

The manuscript had previously been presented as a review article and the authors have followed my suggestion to present the work as an original article. The authors have revised the paper following my indications responding adequately to the points raised. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Although this is a review article, unfortunately, the authors failed to organize individual findings with the clear knowledge that would capture the issue. The work deals only with the relatively narrow issue of stem cells in melanoma and their influence on resistance to targeted therapy, but it is not the only and main mechanism of resistance. In addition, the stem cell problem in melanoma is slightly different from the problem in other solid tumours. Melanoma completely lacks better stem cell properties - eg phenotypic identification, etc.
For these reasons, I believe that although this is a very interesting issue, this work can not be recommended for publication.

Author Response

We are sorry that our article has not been positively evaluated by the reviewer. We are aware that cancer stem cells are not the only mechanism driving resistance to target therapies but in this paper we have decided to focus on this particular aspect. Other mechanisms involved in drug resistance have been dealt by several excellent articles and reviews on the topic. We have briefly described them in introduction section and cited in the bibliography (see references# 9, 10, 14, 15, 18, 19, 20, 46 and 47) in the revised version. In the revised version of our paper we have tried to put in better evidence our original contribution to the field which consists in the identification of a signature of 25 genes that belong to four major clusters, namely 1) kinase and metabolic, 2) melanoma-associated, 3) Hippo pathway and 4) slow cycling/CSCs. In addition, we have also added to the revised version of our work additional analyses directed to assess the prognostic role of this signature in BRAF-mutant melanoma patients. Indeed, mining TCGA data we observed that the concomitant alteration of these genes predicts the worst overall survival (OS) for this subset of patients.

We are also aware that the problem of cancer stem cells in melanoma is quite different from the other solid tumors. We have discussed this point in the introduction section in lines 76-78 through the citation of the reference# 38. Furthermore, we have also approached the topic of phenotypic identification in melanoma in the introduction section (lines 109-131) taking advantage of the very relevant reviews by Bai et al (ref# 46) and Arozarena et al (ref# 47).

In summary, our paper (now become an orginal article according to reviewer# 3 suggestion) has been entirely reorganized and we hope that the new version is going to be better evaluated by the reviewer.

All the changes are highlighted in red.

Reviewer 2 Report

This review by Fattore et al. discussed a crucial problem in melanoma therapy, i.e., the diverse heterogeneity-driven phenotypic switching of melanoma cells under therapeutic stress.  The authors used bioinformatics approach and attempted to find a common ground for gene regulation in response to treatment in melanoma.  4 clusters were identified based on the TCGA dataset.  Although a small step, it is important to summarize these extremely complex heterogeneity which may lead to developing novel combination therapy.  The abstract should be revised to include the 4 clusters that the authors identified.  

Author Response

We are grateful to the reviewer for the overall positive evaluation of our work. We have revised the abstract according to reviewer’s suggestion including the mention to the four cluster we identified. All the changes are highlighted in red.

Reviewer 3 Report

The manuscript entitled “Cancer stem cells and the slow cycling phenotype: how to cut the Gordian knot driving resistance to therapy in melanoma” reviews the literature on cancer stem cells in melanoma and their role in therapy response and resistance. The argument is of paper is interesting and the approach is original. The manuscript is well written, concise though exhaustive.

However, this paper is a hybrid between a review and a bioinformatics hypothesis generating study. I feel that it would be more appropriated to present the study as an original article.  In any case the methods used must be clearly detailed.

Cancer stem cells (CSC), as the author state, are clearly linked to tumor heterogeneity. For this reason, the authors should introduce a consideration of current tumor evolution models. How do CSC relate to punctuate equilibrium?

The authors should consider to extend their analyses to other types of melanoma (mucosal, ocular, acral melanomas). In alternative, they must state that the analysis is limited to cutaneous melanoma.

Specific coments:

Introduction section:

The paragraph on resistance mechanisms to anti-BRAF therapy (rows 49-54) appears a little focused on the authors’ contribution to the field. Other mechanisms should be shortly cited.

Rows 50-52: since “… our group and others …” are indicated, reference to the work of others should be given.

Section 2:

When speaking about “dynamic variations of tumor cells over time” (row 79) the difficulty to distinguish real dynamic variation and outgrowth of a tiny pre-existing population should be discussed.

 

Section 5:

The actual impact factor threshold used to restrict the search to “high indexed journals” (should read “high index journals” or “highly indexed journals”) should be indicated. It is also unclear what is precisely meant by “gene concordance”. Has the research been restricted to genes identified by n studies? Please clarify.

 

Figure 3 shows the expression levels of the 25 selected genes in 480 TCGA melanoma samples. Apparently, the expression values have not been normalized so that certain genes appear highly expressed in almost all samples and others are expressed at much lower levels. This helps to evaluate the importance of the single genes and as a consequence, the importance of those poorly expressed is questionable. The importance of clusters 3 and 4 is also questionable since almost no gene is expressed in any of the cases. This might derive from the fact that the literature used not necessarily considered large melanoma collections and might have identified genes that explain resistance in just very few specific cases. For example, activation of YAP/TAZ has been described as a potential resistance mechanisms but it is unlikely to explain many cases. Furthermore, within each cluster genes and samples should be analyzed by hierarchical clustering in order to allow for the identification of subtypes of melanoma that might particularly express one of the signatures. Given the co-expression of several genes of different clusters by the same samples the meaningfulness of these clusters is unclear. Do specific melanomas tend to overexpress specific functional clusters of genes in order to escape treatment?

 

Figure 4: The legend shows PCA2, PCA1 being likely in the x-axes where it is, however, not indicated. I do not agree with the interpretation of the authors, that the particular scattering of the melanoma cases simply indicates their greater heterogeneity. It might as well indicate the existence of classes within the melanoma samples (see the comment to fig. 3). Furthermore, similar analyses should be applied to the single clusters. It should be discussed whether the specificity of the association of these genes with melanoma as compared to other cancers is consistent with the claim they might have to do with CSC, which one would believe to be similar among different tumor types.

Author Response

Please see the attachment

Author Response File: Author Response.pdf