Metabolically Healthy Obesity Is Characterized by a Distinct Proteome Signature

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper examines the proteome analysis in a subgroup of obese but without metabolic abnormalities to examine their potential metabolic abnormalities. Although obesity is often associated with metabolic diseases such as type 2 diabetes, hypertension, and dyslipidemia, and increases the risk of cardiovascular disease, there is certainly a subgroup (obesity only, OBO) that is obese but without metabolic abnormalities. These potential metabolic abnormalities may be present, and their elucidation is important in clinical medicine. The fact that our proteome analysis revealed the presence of at least some hematologic and metabolic alterations may be beneficial for preventive medicine in the future.

 

On the other hand, there are several items to point out.

1. The small number of subjects in the OBO group (27 subjects) and the LHC group (15 subjects) may limit the statistical reliability and reproducibility of the results. Given the relatively non-invasive nature of the study (blood sample analysis), I feel that a large cohort study is needed.

2. The subject's lifestyle (diet, exercise habits, etc.), genetic background, age, gender and social status are not without influence. Please add them to the discussion.

3. Whether the changes in protein expression are the cause or the result of the metabolic abnormalities is a matter of speculation based on the results of this study. Long-term follow-up studies are needed.

4. The criteria for metabolic health (blood glucose, lipid profile, etc.) when defining "obesity only (OBO)" are vague. There are also differences between countries regarding normal and abnormal values. Clinical application based on these results may lack diagnostic consistency. Please discuss this point.

5. You conclude that the results will help "identify therapeutic targets," but what specific therapies can be expected? Are these treatments expected to be useful and cost effective?

Author Response

COMMENT 1: The small number of subjects in the OBO group (27 subjects) and the LHC group (15 subjects) may limit the statistical reliability and reproducibility of the results. Given the relatively non-invasive nature of the study (blood sample analysis), I feel that a large cohort study is needed.

Response 1:  We are thankful for this feedback and agree with the reviewer point that the size of the cohort is small for a powerful statistical analysis. Although, the identified protein changes were significant with p< 0.05., we reported in our discussion the small size of our cohort as a limitation in our study and a plan to expand the study to a larger cohort in the future (Please see yellow highlighted section in the revised manuscript line: 288).

COMMENT 2: The subject's lifestyle (diet, exercise habits, etc.), genetic background, age, gender and social status are not without influence. Please add them to the discussion.

Response 1:  appreciate the reviewer’s suggestion to include lifestyle factors, genetic background, age, gender, and social status in the discussion. We recognize the importance of these variables in influencing the study findings. In response, we have expanded the discussion to address these important aspects. We believe this addition enrich the discussion in the context of our findings (Please see yellow highlighted section in lines: 282-286).

COMMENT 3: Whether the changes in protein expression are the cause or the result of the metabolic abnormalities is a matter of speculation based on the results of this study. Long-term follow-up studies are needed.

Response 3: We absolutely agree with the reviewer’s comment on the necessity of a follow up study to assess and validate the protein changes. We have included a statement in the discussion about a plan to conduct a long-term follow up study (Please see yellow highlighted section in line: 288)

 

COMMENT 4: The criteria for metabolic health (blood glucose, lipid profile, etc.) when defining "obesity only (OBO)" are vague. There are also differences between countries regarding normal and abnormal values. Clinical application based on these results may lack diagnostic consistency. Please discuss this point.

Response 4: This is indeed important, and we appreciate the reviewer’s concern regarding the criteria for metabolic health in defining "obesity only (OBO)." We acknowledge that the thresholds for metabolic indicators, such as blood glucose and lipid profiles, can vary across countries, which may impact diagnostic consistency. In response, we have expanded the discussion to address this as a limitation and the need for further studies (Please see yellow highlighted section in lines: 282-288).

 

COMMENT 5: You conclude that the results will help "identify therapeutic targets," but what specific therapies can be expected? Are these treatments expected to be useful and cost effective?

Response 5: We appreciate the reviewer’s thoughtful question regarding the specific therapies that may emerge from our findings and their potential utility and cost-effectiveness. While our study highlights potential targets based on the observed results, we acknowledge that this aspect will require further research, including preclinical and clinical trials. In the revised manuscript, we discuss the need for further investigation into this aspect. The cost-effectiveness and utility will need further probing and believe these insights provide a foundation for future studies aimed at translating our findings into actionable and accessible therapies (Please see yellow highlighted section in lines: 298-300).

Reviewer 2 Report

Comments and Suggestions for Authors

ijms-3343394

Metabolically Healthy Obesity is Characterized by a Distinct Proteome Signature

1. The manuscript by Mir et al. described a proteomic study using the Olink technique to identify the underlying metabolic pathways that are dysregulated OBO subjects vs. lean, healthy controls. The authors identified 24 differentially expressed proteins in the OBO group and linked them to 15 up- and 15 down-regulated molecular pathways. Overall, it is a preliminary study with basic data from a proteomic study. Although there are some findings, they are preliminary and cannot result in a firm conclusion. In addition, the novelty and contribution of this study is limited. For these reasons, this work may not be suitable for publication in IJMS.

There are some other comments as follows.

2. Introduction: Please revise it to include the research gap as well as the novelty and contribution of this study.

3. Introduction: Paragraph 3 needs citations to support.

4. Some abbreviations, such as “obesity only (OBO)” and “Lean healthy controls (LHC)” were defined several times. Please use all abbreviations consistently after specifying them.

5. Please cite relevant references to the proteomic assay (Olink method).

6. Line 324: what is “30” at the end of the link?

5. Line 326: what is “31” at the end?

 

 

Author Response

COMMENT 1: The manuscript by Mir et al. described a proteomic study using the Olink technique to identify the underlying metabolic pathways that are dysregulated OBO subjects vs. lean, healthy controls. The authors identified 24 differentially expressed proteins in the OBO group and linked them to 15 up- and 15 down-regulated molecular pathways. Overall, it is a preliminary study with basic data from a proteomic study. Although there are some findings, they are preliminary and cannot result in a firm conclusion. In addition, the novelty and contribution of this study is limited. For these reasons, this work may not be suitable for publication in IJMS.

Response 1:  Thank you for your thoughtful and detailed feedback, which highlights an important aspect of our study. We fully understand the concern regarding the limited cohort size and its potential implications for generalizability. While we acknowledge this limitation, we remain confident that the novelty and significance of our findings offer meaningful contributions to the field.

In response to your feedback, we have explicitly addressed the cohort size as a limitation in the revised discussion section and included plans to expand the cohort in future studies to further validate our findings (refer to the yellow-highlighted section on line 290). Moreover, we note that other studies have successfully published proof-of-concept research with relatively small sample sizes. For instance:

• PMID 39449501: Investigated 96 proteins in cohorts of 30 patients, demonstrating impactful discoveries in smaller populations.
• PMID 38649851: Analyzed 368 inflammatory proteins in 30 individuals, including healthy young and elderly individuals as well as elderly patients with cardiometabolic diseases, showcasing significant protein signatures.
• PMID 37170273: Explored early cancer patients (n=30), identifying key protein signatures despite the limited cohort size.
• PMID 35848804: Studied hypoglycemia in a cohort of 15 individuals, achieving recognition with 11 citations in two years.
• While these studies may not focus on biomarker discovery, they underscore the feasibility of detecting meaningful protein signatures in small, specialized cohorts.

COMMENT 2: Introduction: Please revise it to include the research gap as well as the novelty and contribution of this study.

Response 1:   We thank the reviewer for raising this point. We have included details on the research gap as well as the novelty and contribution of this study in the introduction.

COMMENT 3: Introduction: Paragraph 3 needs citations to support.

Response 3: Thank you very much for the suggestion. We have cited the paragraph 3 in our revised manuscript.

 

COMMENT 4: Some abbreviations, such as “obesity only (OBO)” and “Lean healthy controls (LHC)” were defined several times. Please use all abbreviations consistently after specifying them.

Response 4: Thank you very much for highlighting this, we have corrected this in our revised manuscript.  

COMMENT 5: Please cite relevant references to the proteomic assay (Olink method).

Response 5: Thank you very much for the suggestion, we have cited the relevant citation for proteomic assay.

COMMENT 6: Line 324: what is “30” at the end of the link?

Response 5: We are very sorry for this typo. We have corrected it in the revised version.

COMMENT 7: Line 326: what is “31” at the end?

Response 5: We are very sorry for this typo. We have corrected it in the revised version.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors investigate the proteomic profile associated with metabolically healthy obesity (MHO), defined as obesity without concurrent metabolic disorders, using the Olink proteomics platform. Their study analyzes plasma samples from 27 individuals with obesity and 15 lean healthy controls, targeting 184 proteins related to cardiometabolic and inflammatory pathways. The authors identify 24 differentially expressed proteins (13 downregulated and 11 upregulated) and highlight dysregulated pathways related to immune response and inflammation. The findings aim to provide insights into potential biomarkers and therapeutic targets for mitigating metabolic dysfunctions in MHO individuals.

However, several critical issues need to be addressed to strengthen the manuscript's scientific rigor:

1. The study is fundamentally limited by its small sample size, comprising only 42 individuals (27 with obesity and 15 lean healthy controls) and targeting 184 proteins. This minimal dataset severely undermines the statistical power and reliability of the findings. Moreover, the study is neither a large-scale meta-analysis nor one that leverages a unique sample population, calling into question its novelty and potential appeal to readers.

2. The hierarchical clustering analysis lacks critical components necessary to substantiate the authors' claims. Specifically, the dendrogram for the x-axis, which would illustrate the purported separation between obese and control groups, is absent. Interestingly, the inclusion of the y-axis dendrogram appears arbitrary and adds no value to the analysis. This omission reflects poor analytical rigor.

3. The Methods section fails to include any substantive details regarding the hierarchical clustering analysis, rendering it impossible to assess its validity or reproducibility. This omission is a significant flaw in the manuscript.

4. The GO enrichment analysis is poorly executed and misrepresented. The use of terms like “upregulated” and “downregulated” is technically incorrect; the appropriate terminology is “over-represented” and “under-represented.” Furthermore, the lack of methodological details makes it impossible to judge the validity of the analysis. Specifically, the authors may have used the whole genome as the background set, which could explain the identification of numerous under-represented terms unrelated to MHO but instead reflective of generic blood or serum-associated pathways. Such methodological issues significantly diminish the credibility and relevance of the GO analysis.

 

Given the above critical issues, I would suggest rejection of the manuscript in its current form.

Author Response

COMMENT 1: The study is fundamentally limited by its small sample size, comprising only 42 individuals (27 with obesity and 15 lean healthy controls) and targeting 184 proteins. This minimal dataset severely undermines the statistical power and reliability of the findings. Moreover, the study is neither a large-scale meta-analysis nor one that leverages a unique sample population, calling into question its novelty and potential appeal to readers.

Response 1:  Thank you for your thoughtful and detailed feedback, which highlights an important aspect of our study. We fully understand the concern regarding the limited cohort size and its potential implications for generalizability. While we acknowledge this limitation, we remain confident that the novelty and significance of our findings offer meaningful contributions to the field.

In response to your feedback, we have explicitly addressed the cohort size as a limitation in the revised discussion section and included plans to expand the cohort in future studies to further validate our findings (refer to the yellow-highlighted section on line 290). Moreover, we note that other studies have successfully published proof-of-concept research with relatively small sample sizes. For instance:

• PMID 39449501: Investigated 96 proteins in cohorts of 30 patients, demonstrating impactful discoveries in smaller populations.
• PMID 38649851: Analyzed 368 inflammatory proteins in 30 individuals, including healthy young and elderly individuals as well as elderly patients with cardiometabolic diseases, showcasing significant protein signatures.
• PMID 37170273: Explored early cancer patients (n=30), identifying key protein signatures despite the limited cohort size.
• PMID 35848804: Studied hypoglycemia in a cohort of 15 individuals, achieving recognition with 11 citations in two years.
• While these studies may not focus on biomarker discovery, they underscore the feasibility of detecting meaningful protein signatures in small, specialized cohorts.

In our study, despite the smaller sample size, we identified significant changes in 184 targeted proteins (p-values < 0.05), reflecting the robustness of our analytical approach. Additionally, our inclusion of obese individuals without metabolic complications is a particularly unique strength. This subgroup represents a minority phenotype that is both clinically significant and challenging to recruit due to their relatively good health and lower engagement with medical care.

 

Once again, we sincerely thank you for your valuable input, which has strengthened our manuscript.

COMMENT 2: The hierarchical clustering analysis lacks critical components necessary to substantiate the authors' claims. Specifically, the dendrogram for the x-axis, which would illustrate the purported separation between obese and control groups, is absent. Interestingly, the inclusion of the y-axis dendrogram appears arbitrary and adds no value to the analysis. This omission reflects poor analytical rigor.

Response 1:  We thank the reviewer for raising this point. We have amended the figure 2 and added the dendrogram for x-axis accordingly.

COMMENT 3: The Methods section fails to include any substantive details regarding the hierarchical clustering analysis, rendering it impossible to assess its validity or reproducibility. This omission is a significant flaw in the manuscript.

 

Response 3: We thank the reviewer for raising this point. We have included details on the hierarchical clustering analysis in the methods section (Lane 326-330)

 

 

COMMENT 4: The GO enrichment analysis is poorly executed and misrepresented. The use of terms like “upregulated” and “downregulated” is technically incorrect; the appropriate terminology is “over-represented” and “under-represented.” Furthermore, the lack of methodological details makes it impossible to judge the validity of the analysis. Specifically, the authors may have used the whole genome as the background set, which could explain the identification of numerous under-represented terms unrelated to MHO but instead reflective of generic blood or serum-associated pathways. Such methodological issues significantly diminish the credibility and relevance of the GO analysis.

Response 4: Thank you for your insightful feedback regarding the GO enrichment analysis. We acknowledge the incorrect use of terms like “upregulated” and “downregulated” and have replaced them with the technically accurate terms “over-represented” and “under-represented” throughout the manuscript.

To address the lack of methodological details, we have explicitly clarified that the background used for the analysis was whole genome and we have provided details of the tool/package employed These revisions aim to enhance both the validity and relevance of the GO enrichment results, and we sincerely the reviewer for the constructive comments, which have been instrumental in refining this aspect of our analysis.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript was revised accordingly. The current version can be considered for publication after a revision. Please consider the following comments for further revision.

1. The authors should improve the Introduction by clarifying the research gap as well as the novelty of this study.

2. Method: Please clarify the statistical method to compare protein abundances between groups. Please mention the cut-off value for FC (log2FC).

3. I suggest presenting the volcano plot using all data, not only the up- and down-regulated proteins.

4. Which one was used for cut-off: p value (as in Figure 2 legend), Adj. Pval (as in Figure 2 C), or FDR (as in Figure 2 B)?

5. I do not believe that 0.32 is a suitable cut-off for log2FC. We may need to use a stricter cut-off (e.g., |FC| > 1.5 or 2).

6. The legend in Figure 3 is inaccurate and should be corrected. 

7. Section 2.1.2: Please confirm whether those in Figures 3B and 3C are classified as "pathways".

8. The Method mentions KEGG pathways. Please include the results.

9. Discussion: Compared to the literature, what is the novel finding in this study? Which DEPs are considered new findings in this study that can contribute to the literature? Please include those in the Discussion.

10. Line 116: "p value ≥ 0.05"?

11. Lines 152 - 153: "Molecular Components"?

 

Author Response

COMMENT 1: The authors should improve the Introduction by clarifying the research gap as well as the novelty of this study.

Response 1: Thank you for your valuable feedback. We appreciate your suggestions for improving the introduction to better clarify the research gap and highlight the novelty of our study. We have revised the version of the Introduction section with the requested improvements incorporated and highlighted in green for your kind reference (kindly see modification in lines 66-70, and 89-95).

 

COMMENT 2: Method: Please clarify the statistical method to compare protein abundances between groups. Please mention the cut-off value for FC (log2FC).

Response 2: We thank the reviewer for his input. In Methods section, we have elaborated more about statistical methods including cut-off value for FC as per the reviewer’s recommendation we have edited the section 4.3 (Line 358-372).

COMMENT 3: I suggest presenting the volcano plot using all data, not only the up- and down-regulated proteins.

Response 3: We have followed the reviewer’s recommendation and presented a volcano plot using all data as shown in supplemental figure 2.

COMMENT 4: Which one was used for cut-off: p value (as in Figure 2 legend), Adj. Pval (as in Figure 2 C), or FDR (as in Figure 2 B)?

Response 4: We thank the reviewer for raising this question and we have accordingly made proper corrections in figure 2 legend. It should be q-value instead of p-value. We used FDR adjusted p-value (q-value) of 0.05 in Figure 2B and C.

COMMENT 5: I do not believe that 0.32 is a suitable cut-off for log2FC. We may need to use a stricter cut-off (e.g., |FC| > 1.5 or 2).

Response 5: We agree with the reviewer that it is better to raise the cut-off of FC to > 1.5 or 2. In our analysis we used a low cut-off 1.25 FC (0.32 log value) as Min fold change to detect proteins at very low expression changes. Although, they have very low expression changes, they might have an impact on downstream molecular expression and function. cut-off 1.25 is widely used and acceptable by scientific community for protein analysis. On the other hand, in our protein signature list, many detected proteins were above FC=1.5. See supplemental data 2.

COMMENT 6: The legend in Figure 3 is inaccurate and should be corrected.

Response 6: We really appreciate the reviewer’s observation on Figure legend 3. We have corrected accordingly.

COMMENT 7: Section 2.1.2: Please confirm whether those in Figures 3B and 3C are classified as "pathways".

Response 7: Yes, we confirm they are pathways as generated by GO enrichment analysis (see data uploaded from GO in supplemental data 3)

COMMENT 8: The Method mentions KEGG pathways. Please include the results.

Response 8: Results are already included as supplemental figure 1 A-L and are referred in the discussion Line 221,275,and 297 .

COMMENT 9: Compared to the literature, what is the novel finding in this study? Which DEPs are considered new findings in this study that can contribute to the literature? Please include those in the Discussion.

Response 9: We thank the reviewer for this insightful comment and have addressed the concerns regarding the novel findings in our study and have made revisions to the Discussion section, highlighting the specific differentially expressed proteins (DEPs) that contribute new insights to the literature. The text highlighted in green in the updated Discussion section has the requested additions (kindly see modification in lines 184-192 and 282-291).

COMMENT 10: Line 116: "p value ≥ 0.05"?

Response 10: Corrected

COMMENT 11: Lines 152 - 153: "Molecular Components"?

Response 11: corrected to GO Molecular Function

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for addressing my concerns regarding the GO enrichment analysis and revising the terminology used to describe “over-represented” and “under-represented” terms. However, your response does not fully resolve the core issues raised in my original comment, and significant concerns remain regarding both the methodological and biological interpretations of your analysis.

1. Biological Interpretation of "Under-Represented" Terms:
   Your response does not adequately address the biological relevance or meaning of the "under-represented" terms identified in your GO enrichment analysis. While "under-represented terms" are statistically valid results indicating that certain GO terms occur less frequently in your input list (DEPs) than would be expected based on the background set, their biological interpretation is often less straightforward. In most cases, these terms suggest processes or pathways that are statistically depleted in your DEPs compared to the background.

However, the biological significance of such findings is context-dependent and typically requires robust justification. Without clear biological insights or relevance to your study, the inclusion of "under-represented" terms raises questions about their utility in the overall interpretation of your results.

Additionally, the interpretation of "under-represented" terms is highly sensitive to the choice of background set. As I noted in my original comment, the use of the whole genome as the background in your analysis is problematic. This choice introduces biases and likely contributes to the identification of generic, irrelevant terms unrelated to your study. These terms may be artifacts of the analysis rather than meaningful results (which I will explain in details below).

To strengthen the analysis, I strongly encourage focusing on biologically relevant "over-represented terms," as they provide more intuitive and actionable insights into the functional associations of your DEPs. If "under-represented terms" are to be included, they must be accompanied by a clear and robust explanation of their biological relevance.

2.      Choice of Background Set:
Using the whole genome as the background for GO enrichment analysis is methodologically questionable in your context. A whole-genome background introduces biases that are not representative of the actual experimental design and biological system under investigation. This is likely why several “under-represented” terms unrelated to your study emerge in your results, diluting the analysis's relevance.

To illustrate this point, I conducted a re-analysis using the 184 proteins you reported as the input for GO enrichment, using an appropriate background. This yielded enriched terms with strong connections to immune and inflammatory pathways—terms that are both biologically plausible and directly relevant to your study. This starkly contrasts with your results and highlights a significant design flaw in your analysis.

Given these issues, I strongly recommend revisiting your GO enrichment analysis with a more biologically appropriate background set. In addition, I urge you to focus on the biological relevance of the identified terms and provide explicit justifications for the methodological choices made. Without these revisions, the credibility of your findings remains compromised.

 

 

Author Response

COMMENT 1: Biological Interpretation of "Under-Represented" Terms:

Your response does not adequately address the biological relevance or meaning of the "under-represented" terms identified in your GO enrichment analysis. While "under-represented terms" are statistically valid results indicating that certain GO terms occur less frequently in your input list (DEPs) than would be expected based on the background set, their biological interpretation is often less straightforward. In most cases, these terms suggest processes or pathways that are statistically depleted in your DEPs compared to the background.

However, the biological significance of such findings is context-dependent and typically requires robust justification. Without clear biological insights or relevance to your study, the inclusion of "under-represented" terms raises questions about their utility in the overall interpretation of your results.

Additionally, the interpretation of "under-represented" terms is highly sensitive to the choice of background set. As I noted in my original comment, the use of the whole genome as the background in your analysis is problematic. This choice introduces biases and likely contributes to the identification of generic, irrelevant terms unrelated to your study. These terms may be artifacts of the analysis rather than meaningful results (which I will explain in detail below).

To strengthen the analysis, I strongly encourage focusing on biologically relevant "over-represented terms," as they provide more intuitive and actionable insights into the functional associations of your DEPs. If "under-represented terms" are to be included, they must be accompanied by a clear and robust explanation of their biological relevance.

COMMENT 2: Choice of Background Set: Using the whole genome as the background for GO enrichment analysis is methodologically questionable in your context. A whole-genome background introduces biases that are not representative of the actual experimental design and biological system under investigation. This is likely why several “under-represented” terms unrelated to your study emerge in your results, diluting the analysis's relevance.

To illustrate this point, I conducted a re-analysis using the 184 proteins you reported as the input for GO enrichment, using an appropriate background. This yielded enriched terms with strong connections to immune and inflammatory pathways—terms that are both biologically plausible and directly relevant to your study. This starkly contrasts with your results and highlights a significant design flaw in your analysis.

Given these issues, I strongly recommend revisiting your GO enrichment analysis with a more biologically appropriate background set. In addition, I urge you to focus on the biological relevance of the identified terms and provide explicit justifications for the methodological choices made. Without these revisions, the credibility of your findings remains compromised.

 

Response 9: We thank the reviewer for providing all these information and support. We believe there were some misunderstandings with regards to the background used for GO analysis. As mentioned in figure legend 3 and supplemental data 2, the enrichment analysis for DEPs was conducted on subgroup of genes and not the whole-genome background. We used GO Biological Process subgroup and GO Molecular Function subgroup.

We also reported strong connections to immune and inflammatory pathways as shown in Figure 3 B and C and in the discussion (Line 177,190,198,204, 234,304…). Hope this will answer the reviewer’s concerns and will provide a positive decision.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

I can see that the volcano plot in Figure 2b remained the same (please include all data). The terms in Figures 2B, 2C, and figure legend should be consistent (replace "Adj.Pval" with the other term). 

Author Response

Comment: I can see that the volcano plot in Figure 2b remained the same (please include all data). The terms in Figures 2B, 2C, and figure legend should be consistent (replace "Adj.Pval" with the other term). 

 

Response: As per the reviewer’s recommendations, we have amended Figure 2b, including the volcano with all data. We also replaced "Adj.Pval" with FDR in figure 2C to be consistent between Figure 2B, 2C, Figure legend 2 and supplementary data 2 as well.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors should discuss more the "under-represented” terms since the "under-represented” terms is not common in the GO enrichment analysis.

Author Response

Comment: The authors should discuss more the "under-represented” terms since the "under-represented” terms is not common in the GO enrichment analysis.

 

Response: GO enrichment analysis represent pathways as either upregulated or downregulated as shown in Figure 3B and 3C and supplementary data 3. Therefore, to align with the GO terminology we have replaced “over-represented” with “upregulated” and “under-represented” with “downregulated”.