Adhesion Molecules ICAM-1 and PECAM-1 as Potential Biomarkers of Central Nervous System Damage in Women Breast Cancer Survivors

Round 1

Reviewer 1 Report

In this manuscript, the authors insist that the elevation of ICAM-1 and PECAM-1 in serum is a potential biomarker for CNS complications in breast cancer survivors.

They showed that serum ICAM-1 and PECAM-1 is elevated in the patients with complication, such as cognitive dysfunction, depression, and vestibulocerebellar ataxia, but not elevated in the patients with edema nor radiation therapy. The finding suggest that these markers have a specificity for CNS disorders to some extent.

 

In Table 7, they used Games-Howell test. I do not think the method is appropriate for those data.

Line 50: PMPS needs spell-out

Author Response

On behalf of all co-authors I would like to thank the reviewer for the provided comments. We have substantially revised the manuscript including the employed statistical methods.

Comment (1): In Table 7, they used Games-Howell test. I do not think the method is appropriate for those data.

Answer (1): We have reanalyzed the data, including subgroup analysis of patients based on the hormone receptor status and histological type of breast cancer.

Comment (2): Line 50: PMPS needs spell-out

Answer (2): This was corrected.

Reviewer 2 Report

1) Please fix the y-axis in figure 1

2) Were there any differences in the group based on race of the survivors, it would be beneficial if the authors addressed this in the discussion
3) Authors should provided a breakdown of the different types of breast cancers the survivors had and correlate changes in ICAM and PECAM 

Author Response

On behalf of all co-authors I would like to thank the reviewer for the provided comments. We have substantially revised the manuscript and figures accordingly.

Comment (1): 1) Please fix the y-axis in figure 1

Answer (1): The figure was corrected.

Comment (2): 2) Were there any differences in the group based on race of the survivors, it would be beneficial if the authors addressed this in the discussion

Answer (2): All studied patients as well as healthy volunteers were caucasian. We have included with information into the manuscript. In future studies we plan to address this question by incorporating patients of other races.

Comment (3): 3) Authors should provided a breakdown of the different types of breast cancers the survivors had and correlate changes in ICAM and PECAM 

Answer (3): We have revised the manuscript accordingly providing the subgroup analysis based on the histological type of the tumor as well as hormone receptor status. There was no statistically significant difference in the levels of adhesion molecules (ICAM-1, PECAM-1)  depending either on hormonal status or the histological type of breast cancer. 

Reviewer 3 Report

Comments on manuscript Pathophysiology-1547142 entitled

“Adhesion molecules ICAM-1 and PECAM-1 as potential biomarkers of central nervous system damage in breast cancer survivors women” authored by Pospelova et al..

In the manuscript the authors present a set of clinical and neuro-/psychological data from breast cancer patients after different treatments. These data are used for subgroup analysis. Two soluble adhesion molecules (sICAM-1 and sPECAM-1, abbreviated I/PECAM below) were measured in the serum of the patients and a group of healthy controls.

However, from the data presented it is somewhat premature to consider I/PECAM as biomarkers for nervous system damage, as there is no conclusive evidence that the treatments and not other factors were responsible for the observed differences assigned to treatments. Some of this lacking evidence should be easy to provide as the authors have the data. In addition, some really brain or neuronal damage specific markers should be included, which also easily could be done with the available serum samples.

 

Major points

1 Overall, the manuscript is well referenced. However, ref 3 does not mention the “80% of patients” (it gives numbers for patient with postmastectomy pain syndrome, 30%). In addition, the authors should uniformly present/format the references in the bibliography. At least these references typos, lacking information or duplicate information or different formatting: 1,7,14,22,23,24,25,33,34,38,48,50.

2 In the results section the authors largely repeat the information given in tables and figure 1. In most cases this is not necessary. For example, information given in lines 210-226 is essentially the same as in tables 1 and 2; table 4 is data also shown in figure 1, where also the statistics presented in table 5 can be included; results of the statistical analysis from table 7 can be given in table 6. Together, this looks like ballooning the results section. Please, minimize the redundancies.

3 Parts of table 1 are hard to understand: The first three treatment groups are overlapping. It would be better to indicate patients having surgery plus (i) chemo-, (ii) radiotherapy, or (iii) both. Furthermore, the information given for hormonal therapies are not really instructive. I assume that the patients were from a mixed cohort of ER/PR-positive and ER/PR-negative tumors, what might complicate the analysis of the different subgroups.

4 In the subgroup analysis the age information for the subgroups should be given (medians [25/75 percentiles]). This is essential to exclude age-related changes of I/PECAM serum levels as the underlying cause of the differences. This also is important for the discussion.

5 The discussion – especially lines 325-369 – tries to convince the reader, that CNS damage is the reason for increases of I/PECAM in the patients. However, the discussion shows that there is a plethora of reasons for increases of I/PECAM in the circulation. On the other hand, I/PECAM are neither neuronal damage nor brain specific. Therefore, it would be essential for the study to include brain / neuronal damage specific markers (for example S100B, NSE and the like) to be able to more clearly assign the source of elevated I/PECAM. This easily could be done with the available serum samples.

6 Ideally, the differences in brain damage would be shown by MRI. However, I admit that this would be time-consuming and expensive. However, this should be discussed as a limitation of the study.

7 The statement lines 384-388 is misleading, as up to now the authors provided no experimental evidence for a direct relation of brain damage and I/PECAM serum levels.

 

Minor points

• There are a considerable number of typos and strange expressions (like “-“ in front of numbers, which could not be negative).
• Fig 1: please label axis in latin numbers and give a axis title

 

Author Response

On behalf of all co-authors I would like to thank the reviewer for the provided comments. We have substantially revised the manuscript additionally incorporating a subgroup analysis of patients based on the histological type of breast cancer and hormone receptor status.

Comment (1): <...> some really brain or neuronal damage specific markers should be included, which also easily could be done with the available serum samples.

Answer (1): We agree with the reviewer that inclusion of CNS specific markers would be beneficial for the study protocol. Due to the lack of samples we have launched a new recruitment of patients where we will incorporate these markers (e.g., NSE, neuron-specific enolase).

Comment (2): Overall, the manuscript is well referenced. However, ref 3 does not mention the “80% of patients” (it gives numbers for patient with postmastectomy pain syndrome, 30%). In addition, the authors should uniformly present/format the references in the bibliography. At least these references typos, lacking information or duplicate information or different formatting: 1,7,14,22,23,24,25,33,34,38,48,50.

Answer (2): This was corrected.

Comment (3): In the results section the authors largely repeat the information given in tables and figure 1. In most cases this is not necessary. For example, information given in lines 210-226 is essentially the same as in tables 1 and 2; table 4 is data also shown in figure 1, where also the statistics presented in table 5 can be included; results of the statistical analysis from table 7 can be given in table 6. Together, this looks like ballooning the results section. Please, minimize the redundancies.

Answer (3): We have revised the manuscript accordingly and reduced the mentioned information in the text.

Comment (4): 3 Parts of table 1 are hard to understand: The first three treatment groups are overlapping. It would be better to indicate patients having surgery plus (i) chemo-, (ii) radiotherapy, or (iii) both. Furthermore, the information given for hormonal therapies are not really instructive. I assume that the patients were from a mixed cohort of ER/PR-positive and ER/PR-negative tumors, what might complicate the analysis of the different subgroups.

Answer (4): We have corrected this. Additionally, we have performed the subgroup analysis of patients based on the histological type and hormone receptor status.

Comment (5): In the subgroup analysis the age information for the subgroups should be given (medians [25/75 percentiles]). This is essential to exclude age-related changes of I/PECAM serum levels as the underlying cause of the differences. This also is important for the discussion.

Answer (5): We have added this information and performed the analysis. No statistically significant difference in the age of the patients in the subgroup analysis was found.

Comment (6): The discussion – especially lines 325-369 – tries to convince the reader, that CNS damage is the reason for increases of I/PECAM in the patients. However, the discussion shows that there is a plethora of reasons for increases of I/PECAM in the circulation. On the other hand, I/PECAM are neither neuronal damage nor brain specific. Therefore, it would be essential for the study to include brain / neuronal damage specific markers (for example S100B, NSE and the like) to be able to more clearly assign the source of elevated I/PECAM. This easily could be done with the available serum samples.

Answer (6): We agree with a reviewer. Due to the lack of samples we plan to recruit more patients where we will include these markers into the study protocol.

Comment (7): Ideally, the differences in brain damage would be shown by MRI. However, I admit that this would be time-consuming and expensive. However, this should be discussed as a limitation of the study.

Answer (7): Because of the nonspecificity of intercellular adhesion molecules, the results obtained by the authors certainly need further additional confirmation. Further studies are planned, in which specific markers of neuronal damage will be used and their correlations with intercellular adhesion molecules and the clinical picture will be performed. In addition, the authors previously conducted a study in which CNS changes in breast cancer survivors were visualized by fMRI [new reference 57]. For technical reasons (partial mismatch of the studied groups, group analysis of MRI results) it is not possible to make direct correlations between the results at the moment. In the future, we plan to expand the sample of patients, use additional imaging methods, and identify the correlation between imaging and laboratory signs of CNS damage.

Comment (8): The statement lines 384-388 is misleading, as up to now the authors provided no experimental evidence for a direct relation of brain damage and I/PECAM serum levels.

Answer (8): This was corrected.

Comment (9): There are a considerable number of typos and strange expressions (like “-“ in front of numbers, which could not be negative).

Answer (9): These were corrected.

Comment (10): Fig 1: please label axis in latin numbers and give a axis title

Answer (10): This was corrected.

Round 2

Reviewer 3 Report

The authors did a good job in making the manuscript more readable. However, it's a pity that the authors did not find the 100-200 µl of serum necessary for measuring at least one established marker for neuronal damage (given that 7 ml blood was taken as described).

In tables 4 and 6 of the edited version of the manuscript there are different font size used in parts of the tables/headings.

In addition, please check whether your ELISAs really warrant I/PECAM-levels been given with two digits after the decimal point. I would think that omitting the digits after the decimal point would be more appropriate given the usual variances in ELISA-measurements.

Also, please check whether in the "no edema group" in table 6 the value of 1524 is correct (it is far beyond all other values)?

Moreover, the issues with figure 1 are not fully resolved: now an axis title is given, but the axis labels are still given in cyrillic letters - and now the labels I/PECAM in the legend are missing.

 

Author Response

On behalf of all co-authors I would like to thank the reviewer for the provided comments. We have revised the manuscript accordingly.

Comment (1): In tables 4 and 6 of the edited version of the manuscript there are different font size used in parts of the tables/headings.

Answer (1): This was corrected.

Comment (2): In addition, please check whether your ELISAs really warrant I/PECAM-levels been given with two digits after the decimal point. I would think that omitting the digits after the decimal point would be more appropriate given the usual variances in ELISA-measurements.

Answer (2): We have corrected this.

Comment (3): Also, please check whether in the "no edema group" in table 6 the value of 1524 is correct (it is far beyond all other values)?

Answer (3): We have corrected this misprint.

Comment (4): Moreover, the issues with figure 1 are not fully resolved: now an axis title is given, but the axis labels are still given in cyrillic letters - and now the labels I/PECAM in the legend are missing.

Answer (4): The Figure 1 was corrected accordingly.