Plasma Proteomics Reveals Persistent and Surgery-Responsive Molecular Signatures in Osteoarthritis Patients

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript uses plasma proteomics to compare patients with osteoarthritis (OA) before surgery (Pre-Op) and 6 weeks after arthroplasty (Post-Op) with healthy controls, and aims to distinguish a putative “persistent OA signature” from surgery-responsive changes. The topic is of interest and the dataset may be informative. However, key design limitations weaken the current inferences. Specifically, the very small cohort, the pooling of knee and hip arthroplasty cases without stratified or sensitivity analyses, and the single postoperative time point make it difficult to separate disease-related signals from perioperative effects (e.g., surgical trauma, blood loss/anemia, acute-phase responses, and medication). I therefore recommend major revision. The authors should strengthen the analytical strategy (including appropriate handling of confounding and heterogeneity), and temper causal or mechanistic claims to reflect the exploratory nature of the data.

Author Response

We sincerely appreciate your time and effort in reviewing our manuscript. Your insightful comments and constructive suggestions have been invaluable in improving the quality and clarity of our work. In response to your comments, we have carefully considered each of your suggestions and implemented the necessary revisions accordingly. A summary of these changes is provided below. Thank you for your thoughtful feedback and contribution to enhancing our study.

Comment: The manuscript uses plasma proteomics to compare patients with osteoarthritis (OA) before surgery (Pre-Op) and 6 weeks after arthroplasty (Post-Op) with healthy controls, and aims to distinguish a putative “persistent OA signature” from surgery-responsive changes. The topic is of interest and the dataset may be informative. However, key design limitations weaken the current inferences. Specifically, the very small cohort, the pooling of knee and hip arthroplasty cases without stratified or sensitivity analyses, and the single postoperative time point make it difficult to separate disease-related signals from perioperative effects (e.g., surgical trauma, blood loss/anemia, acute-phase responses, and medication). I therefore recommend major revision. The authors should strengthen the analytical strategy (including appropriate handling of confounding and heterogeneity), and temper causal or mechanistic claims to reflect the exploratory nature of the data.

Response: We sincerely thank the reviewer for the careful evaluation of our manuscript and for recognizing the potential relevance of the dataset. We fully acknowledge the methodological concerns raised and have substantially revised the manuscript to address these points and to temper our interpretations accordingly.
                  First, we agree that the small cohort size limits the strength of the inferences that can be drawn. In response, we have now added a dedicated Limitations section explicitly discussing the reduced statistical power, the risk of false positives despite multiple testing correction, and the exploratory nature of the findings. Throughout the manuscript, we have revised the language to avoid causal or mechanistic overstatements and to clearly frame the identified signatures as preliminary and hypothesis-generating rather than definitive biomarkers.
                  Second, we acknowledge that pooling knee and hip arthroplasty cases may introduce biological heterogeneity. In the revised Discussion, we now explicitly recognize this limitation and discuss the possibility that anatomical site–specific molecular phenotypes may exist. We also emphasize that future studies with larger, anatomically stratified cohorts will be necessary to address joint-specific differences.
                  Third, we agree that the single postoperative time point at six weeks primarily captures early recovery and may reflect perioperative effects such as surgical trauma, acute-phase responses, medication, or transient inflammatory changes. We have incorporated this consideration into both the Discussion and the newly added Limitations section, clarifying that the observed post-operative proteomic shifts cannot be interpreted as long-term remodeling signatures. Instead, they likely represent a combination of disease biology and early postoperative immune adaptation.
                  In addition, we have strengthened the analytical interpretation by explicitly discussing OA endotype heterogeneity and the possibility that baseline immune phenotypes may influence both pre-operative and post-operative proteomic trajectories. We now acknowledge that the distinction between persistent and surgery-responsive proteins may partially reflect pre-existing biological heterogeneity rather than a purely surgery-driven transition.
                  Taken together, these revisions substantially moderate the interpretive claims and clarify the exploratory scope of the study. We believe the manuscript now more appropriately reflects the limitations of the design while preserving the value of the proteomic observations as a foundation for future, larger-scale investigations.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a comparative plasma proteomic profiling study of osteoarthritis (OA) patients before and after total knee or hip arthroplasty, alongside age-, sex-, and BMI-matched healthy controls, employing liquid chromatography–tandem mass spectrometry to identify differentially abundant proteins and associated biological pathways. The concerns raised below span citation accuracy, internal data consistency, methodological transparency, biological interpretation, and structural completeness, all of which must be addressed to adequately support the authors' conclusions. The manuscript requires substantial revision before it can be considered for publication.

1. The current Abstract (Lines 26–47) suffers from redundant phrasing that dilutes its impact. Specifically, 'plasma proteomic analysis detected' is repeated in close proximity (Lines 40–44), and Line 38 overlaps conceptually with Lines 40–42. To improve clarity, the abstract must be streamlined to remove these redundancies. The authors should sequentially and concisely present their core quantitative findings: the 93 pre-operative differentially abundant proteins, the 63/23 persistent protein signature, and the 6 newly emergent post-operative proteins.

2. The statement at Lines 74–77 that biomarker heterogeneity in OA has hindered the identification of clinically translatable markers is supported by Refs. 7, 8, 9. However, Ref 7 concerns healthcare utilization in genetic skeletal disorders, Ref 8. addresses medical resource consumption in ANCA-associated vasculitis, and Ref. 9 is a cost-of-illness study of inherited retinal diseases. None of these articles are relevant to OA biomarker discovery or clinical translation. This citation error is one of the most serious issues in the manuscript, as it raises concerns about the rigor of the literature review process. The authors must replace these three references with directly relevant literature.

3. Table 1 lacks statistical comparisons (p-values) between the control and Pre-Op groups, which is necessary to confirm baseline comparability. Additionally, the control group's biochemical parameters (Lines 95–111) inexplicably mirror the exact midpoints of clinical reference ranges. The authors must add statistical comparison columns to the table and explicitly clarify in the Methods that these are directly measured clinical values, not theoretical approximations.

4. The identifier 'CXCL' is used across multiple tables (Table 2, Supplementary Tables S1-S3), but this represents a chemokine family prefix, not a distinct protein. There is no standalone 'CXCL' entry in the Human Protein Atlas cited by the authors. The authors must pinpoint the specific gene (e.g., CXCL1, CXCL8) and update all figures and tables accordingly. Additionally, the authors should clarify how this ambiguous annotation occurred in their pipeline and confirm that it did not confound any downstream enrichment or pathway analyses.

5. Table 2 lists proteins meeting a ≥5-fold change (p < 0.05) in at least one comparison; however, several inclusions require justification or removal. Specifically, SPARC fails to meet the 5-fold threshold at either time point (Pre-Op FC = 3.51; Post-Op FC = 4.37). Furthermore, INTS6 fails both criteria simultaneously, as its Pre-Op FC (4.56) is below the threshold and its Post-Op change is non-significant (p = 0.254). Lastly, HBS1L presents a non-significant Pre-Op change (p = 0.06) ; while Supplementary Figure S1 correctly denotes it as a Post-Op-only protein , Table 2 includes the Pre-Op data without any transparent annotation. The authors must either remove these proteins from the table or provide explicit footnotes explaining their inclusion.

6. The manuscript completely omits cartilage acidic protein 1 (CRTAC1) from its protein lists, tables, and supplementary datasets , and provides no explanation for its absence. Given that CRTAC1 is currently the most robustly and independently replicated plasma OA biomarker in the field (e.g., Styrkarsdottir et al., 2021; Szilagyi et al., 2023; Kang et al., 2025), its absence in a study claiming to characterize "persistent OA molecular signatures" is a critical oversight. The authors must explicitly address whether this absence is a technical artifact—such as the systematic removal of low-abundance cartilage-derived proteins by the depletion strategy. Furthermore, while Cartilage Oligomeric Matrix Protein (COMP) is listed as a detected protein in Supplementary Table S1, its differential expression status is conspicuously absent from the main text and Discussion. The authors should explicitly report the detection status, fold-change direction, and statistical significance of both CRTAC1 and COMP, and discuss these findings to properly contextualize their results within the established OA biomarker landscape.

7. The detection of these typically intracellular proteins at dramatically elevated levels in depleted plasma is biologically unexpected and warrants explicit mechanistic discussion. The authors should note that their own Gene Ontology analyses (Supplementary Tables S2 and S3) identify "extracellular exosome" and "blood microparticle" as significantly enriched cellular components. This provides a plausible biological rationale: these proteins may enter the circulation via exosome shedding, microparticle release, or apoptotic leakage from stressed joint-associated cells. The manuscript must incorporate this discussion to address the unexpected localization. Furthermore, the authors should acknowledge the alternative explanation that some signals may reflect technical artifacts arising from non-specific enrichment during the high-abundance protein depletion process. Given these uncertainties, orthogonal validation (e.g., ELISA or targeted mass spectrometry) is strongly recommended for the most biologically unexpected hits to confirm their clinical relevance.

8. While Figure 7 identifies a distinct cluster of 23 proteins consistently downregulated at both Pre-Op and Post-Op time points, the current Discussion is disproportionately skewed toward upregulated targets. The sustained suppression of proteins like GULP1, MST1, SERPINA4, PZP, F10, and TTR likely represents a critical, yet overlooked, component of the chronic OA phenotype. Specifically, the downregulation of TTR in OA plasma, a protein the authors' own data (Supplementary Tables S2/S3) links to the Thyroid hormone synthesis pathway, cannot be ignored. The authors should expand their Discussion to provide a balanced interpretation of how this persistent downregulation contributes to OA pathogenesis.

9. In the Discussion (Lines 406–407), the authors use a genetic association study of TENT5A to justify their proteomic findings regarding TENT5D. However, this extrapolation is scientifically unjustified for two key reasons: first, TENT5A and TENT5D are distinct family members with partially non-overlapping functions; second, a SNP-level genetic association does not equivalently translate to circulating protein abundance in plasma. The authors should adopt a more conservative approach by clearly distinguishing these two proteins and acknowledging the absence of direct published evidence linking TENT5D specifically to OA pathobiology. Consequently, the elevated TENT5D expression should be presented strictly as a novel, exploratory finding rather than being grounded in inappropriate cross-member inference.

10. The citation of McAlpine et al. (Lines 408–409) to support WDR41's role in OA inflammation is an over-extrapolation. That study focuses on the SMCR8-WDR41-C9ORF72 complex in ALS/FTD and systemic inflammation. The authors should revise the Discussion to present WDR41 merely as a novel candidate requiring future validation in OA models, rather than assuming its established lysosomal functions translate directly to joint inflammation.

11. At Lines 416–418, the authors claim that ANKRD26 and ADCY8 upregulation suggests dysregulated cell survival and intracellular communication contributing to cartilage breakdown, citing Ref. 17 and Ref. 18. While extrapolating molecular mechanisms from other diseases is an acceptable practice, Ref. 18 (investigating endometrial cancer) lacks the direct functional evidence required to substantiate these specific claims in the context of osteoarthritis. To provide a scientifically sound discussion, the authors should either cite studies with concrete mechanistic evidence to justify this cross-tissue extrapolation or, preferably, ground their arguments in their own dataset. Specifically, Supplementary Tables S2 and S3 already show that ADCY8 is significantly enriched in contextually relevant pathways, such as Relaxin signaling and platelet activation. The authors should revise this passage to ensure their mechanistic claims are supported by robust literature or their own pathway-level findings.

12. The pathway enrichment data in Supplementary Tables S2 and S3 reveal several biologically significant findings that are entirely omitted from the Discussion. Specifically, the Relaxin signaling pathway (ACTA2, MMP2, COL4A3, ADCY8, PLCB1) and the Thyroid hormone synthesis pathway (TTR, ALB, ADCY8, PLCB1) reach statistical significance at both Pre-Op and Post-Op time points. Despite their direct relevance to extracellular matrix remodeling and joint tissue homeostasis, they remain unaddressed in the main text. Furthermore, the Post-Op dataset highlights the "Regulation of Complement cascade" (p = 0.046; involving IGHG3, CFHR1, C1R, IGKV3-11) as a newly enriched term absent from the Pre-Op analysis, suggesting a potential surgery-specific immune regulatory signature. The authors must revise the Discussion to incorporate these critical pathway observations or explicitly justify their exclusion from the mechanistic interpretation.

13. The enrichment of platelet activation and complement/coagulation cascades reported in Supplementary Tables S2 and S3, and discussed in the text, overlaps substantially with prior published work that the authors fail to cite. For instance, Kraus et al. (2023) reported a serum diagnostic peptide panel for knee OA with prominent complement and coagulation enrichment. Similarly, Naili et al. (2025) described targeted plasma proteomics in a knee OA cohort with Reactome enrichment dominated by platelet activation and neutrophil degranulation, which is strikingly parallel to the current findings. Furthermore, Zhang et al. (2024) demonstrated that plasma extracellular vesicle proteins, including fibrinogen chains, predict radiographic OA progression, linking these same pathways to disease outcomes. The convergence of the current data with these independent studies is actually a major strength, as it suggests the authors' findings are biologically reproducible despite the limited sample size. However, presenting these enrichment results as completely novel discoveries without acknowledging the prior concordant literature significantly overstates the manuscript's contribution. The authors must revise the Discussion to properly frame these findings as a corroboration and extension of an established proteomic pattern in OA plasma, accompanied by the appropriate citations.

14. The manuscript interprets the 63 persistently elevated proteins and the 6 post-operative specific proteins as if they represent biologically uniform categories defined solely by the surgical intervention, yet it fails to address the well-established molecular heterogeneity within the OA patient population. Current evidence distinguishes a hyperinflammatory endotype characterized by complement activation and Fc receptor signaling from a hypoinflammatory structural endotype, and these groups respond differently to both pharmacological and surgical treatments. For example, Xie et al. (2024) demonstrated that pre-operative plasma proteomic profiles stratify knee OA patients into distinct treatment responder groups, while Giordano et al. (2023) showed that pre-operative serum inflammatory cluster assignments predict distinct post-arthroplasty pain trajectories. Given the small cohort of only eight patients that mixes knee and hip arthroplasty cases, the observed split between persistent and surgery-responsive proteins may reflect pre-existing endotype heterogeneity just as much as a true surgery-driven biological transition. The authors must cite this relevant endotype literature, explicitly acknowledge this alternative interpretation in the Discussion, and note that future studies should incorporate pre-operative immune phenotyping to disambiguate baseline disease heterogeneity from surgery-driven proteomic changes.

15. The manuscript completely omits a Limitations section, which is a significant oversight for a discovery-phase proteomic study of this design. The most critical unaddressed limitation is the sample size: with only eight OA patients and ten controls (n = 18), the statistical power to reliably identify 93 differentially abundant proteins is severely limited, and the risk of false positives remains a legitimate concern despite Benjamini–Hochberg correction. Several other crucial limitations must be explicitly disclosed: the single post-operative time point at six weeks primarily captures the acute recovery phase rather than medium- or long-term molecular trajectories; the cohort mixes knee and hip arthroplasty patients without subgroup analysis, overlooking potential site-specific molecular phenotypes; and the study lacks any orthogonal protein-level validation (e.g., ELISA) for its high-dimensional findings. The authors must add a dedicated Limitations paragraph to discuss these issues and tone down the language throughout their conclusions to accurately reflect that these signatures are preliminary, exploratory findings rather than established clinical biomarkers.

16. The Materials and Methods section contains systematic numbering errors that require correction. Specifically, the mass spectrometry subsection (Line 568) is incorrectly labeled "2.3.3. Mass Spectrometry Analysis" instead of 4.3.3 , and the subsequent bioinformatics subsection (Line 586) is improperly labeled "2.4. Bioinformatics and Statistical Analysis" instead of 4.4. These formatting oversights suggest a lack of careful proofreading prior to submission. The authors must thoroughly review and correct all section headings, numbering, and structural cross-references throughout the manuscript to meet the journal's publication standards.

17. In Section 4.3.3, the authors describe HDMSE acquisition on a SYNAPT G2-Si system using a fixed elevated collision energy of 30 V during the high-energy scan. In many HDMSE workflows for complex proteomic digests, a collision energy ramp (for example, 15–40 V) is commonly implemented to accommodate peptides with diverse sizes and charge states and to ensure balanced fragmentation efficiency across the m/z range. It would therefore be helpful for the authors to clarify whether the 30 V setting was applied as a fixed value across the entire m/z range or selected based on prior optimization or vendor-validated preset parameters. Providing a brief explanation of the rationale for this setting would enhance methodological transparency and reproducibility without affecting the overall strength of the reported proteomic findings.

18. While data are available upon request (Lines 649–650), it is highly recommended that the authors deposit their raw proteomics data into a recognized public repository like PRIDE and provide the accession number.

Author Response

This manuscript presents a comparative plasma proteomic profiling study of osteoarthritis (OA) patients before and after total knee or hip arthroplasty, alongside age-, sex-, and BMI-matched healthy controls, employing liquid chromatography–tandem mass spectrometry to identify differentially abundant proteins and associated biological pathways. The concerns raised below span citation accuracy, internal data consistency, methodological transparency, biological interpretation, and structural completeness, all of which must be addressed to adequately support the authors' conclusions. The manuscript requires substantial revision before it can be considered for publication.

We sincerely appreciate your time and effort in reviewing our manuscript. Your insightful comments and constructive suggestions have been invaluable in improving the quality and clarity of our work. In response to your comments, we have carefully considered each of your suggestions and implemented the necessary revisions accordingly. A summary of these changes is provided below.

Comment 1: The current Abstract (Lines 26–47) suffers from redundant phrasing that dilutes its impact. Specifically, 'plasma proteomic analysis detected' is repeated in close proximity (Lines 40–44), and Line 38 overlaps conceptually with Lines 40–42. To improve clarity, the abstract must be streamlined to remove these redundancies. The authors should sequentially and concisely present their core quantitative findings: the 93 pre-operative differentially abundant proteins, the 63/23 persistent protein signature, and the 6 newly emergent post-operative proteins.

Response 1: We thank the reviewer for this valuable suggestion. We have revised the Abstract (Lines 26–45) to improve clarity and remove redundant phrasing, including repeated expressions. In the revised version, we present the key quantitative findings in a more concise and sequential manner, explicitly reporting the 93 pre-operative differentially abundant proteins, the 63 upregulated and 23 downregulated proteins forming the persistent signature, and the 20 post-operative–specific proteins. These changes have significantly improved the readability and impact of the Abstract.

Comment 2: The statement at Lines 74–77 that biomarker heterogeneity in OA has hindered the identification of clinically translatable markers is supported by Refs. 7, 8, 9. However, Ref 7 concerns healthcare utilization in genetic skeletal disorders, Ref 8. addresses medical resource consumption in ANCA-associated vasculitis, and Ref. 9 is a cost-of-illness study of inherited retinal diseases. None of these articles are relevant to OA biomarker discovery or clinical translation. This citation error is one of the most serious issues in the manuscript, as it raises concerns about the rigor of the literature review process. The authors must replace these three references with directly relevant literature.

Response 2: We sincerely thank the reviewer for this important and insightful comment regarding the inappropriate citations used to support the statement on biomarker heterogeneity in osteoarthritis.
We fully agree that the previously cited references (Refs. 7–9) were not directly relevant to OA biomarker discovery or clinical translation, and we appreciate the reviewer for identifying this issue. In response, we have carefully revised the relevant sentence (Lines 74–76) and replaced the original references with more appropriate and up-to-date literature that directly addresses the challenges of biomarker identification, validation, and clinical translation in the context of OA.
                  In addition, prompted by this comment, we have thoroughly re-evaluated all references throughout the manuscript to ensure their accuracy, relevance, and appropriate contextual alignment with the statements they support. These revisions have improved the scientific rigor and overall reliability of the literature framework in our study. We thank the reviewer again for helping us strengthen this aspect of the manuscript.

Comment 3: Table 1 lacks statistical comparisons (p-values) between the control and Pre-Op groups, which is necessary to confirm baseline comparability. Additionally, the control group's biochemical parameters (Lines 95–111) inexplicably mirror the exact midpoints of clinical reference ranges. The authors must add statistical comparison columns to the table and explicitly clarify in the Methods that these are directly measured clinical values, not theoretical approximations.

Response 3: We thank the reviewer for this important comment. In the revised manuscript, we have added statistical comparison column (p-values) to Table 1 to confirm baseline comparability between the control and pre-operative groups.
In addition, we clarify that all biochemical parameters presented in Table 1 represent directly measured clinical laboratory values obtained from the hospital laboratory information system. This clarification has now been explicitly added to the Materials and Methods section (Lines 735–737) to ensure transparency and avoid potential misinterpretation.

                  We also note that several laboratory parameters were originally exported from the laboratory information system with limited decimal precision, which in some cases resulted in values appearing close to the midpoints of the clinical reference ranges. For consistency across variables, these values are now reported using standardized rounding (up to two significant figures). Importantly, this formatting does not affect the underlying measured data or the statistical analyses.

Comment 4: The identifier 'CXCL' is used across multiple tables (Table 2, Supplementary Tables S1-S3), but this represents a chemokine family prefix, not a distinct protein. There is no standalone 'CXCL' entry in the Human Protein Atlas cited by the authors. The authors must pinpoint the specific gene (e.g., CXCL1, CXCL8) and update all figures and tables accordingly. Additionally, the authors should clarify how this ambiguous annotation occurred in their pipeline and confirm that it did not confound any downstream enrichment or pathway analyses.

Response 4: We thank the reviewer for this important and insightful comment regarding the ambiguous use of the identifier “CXCL”.
We fully agree that “CXCL” represents a chemokine family prefix rather than a specific protein. Upon careful re-evaluation, we confirm that the correct identifier in our dataset is CXCL4V1, and that the use of the truncated label “CXCL” in certain tables and supplementary materials resulted from an annotation inconsistency during data export and formatting.
                  We have now systematically corrected this issue across all relevant sections of the manuscript, including Table 2, Supplementary Tables, Supplementary Figure S1, and Figure 1, 2, 3, 6,  ensuring that all entries reflect the precise gene/protein identifiers.
                  Importantly, we confirm that all downstream analyses, including pathway enrichment and functional annotation, were conducted using the correct gene-level identifiers (i.e., CXCL4V1), and were therefore not affected by this labeling inconsistency. We appreciate the reviewer for highlighting this issue, which has allowed us to improve the accuracy and clarity of our dataset annotation.

Comment 5: Table 2 lists proteins meeting a ≥5-fold change (p < 0.05) in at least one comparison; however, several inclusions require justification or removal. Specifically, SPARC fails to meet the 5-fold threshold at either time point (Pre-Op FC = 3.51; Post-Op FC = 4.37). Furthermore, INTS6 fails both criteria simultaneously, as its Pre-Op FC (4.56) is below the threshold and its Post-Op change is non-significant (p = 0.254). Lastly, HBS1L presents a non-significant Pre-Op change (p = 0.06) ; while Supplementary Figure S1 correctly denotes it as a Post-Op-only protein , Table 2 includes the Pre-Op data without any transparent annotation. The authors must either remove these proteins from the table or provide explicit footnotes explaining their inclusion.

Response 5: We thank the reviewer for the careful evaluation of Table 2 and for highlighting the inconsistencies between the stated inclusion criteria and the listed proteins.
In response, we have thoroughly re-evaluated all entries in Table 2 to ensure strict adherence to the predefined selection criteria (≥5-fold change with p < 0.05 in at least one comparison). As a result, proteins that did not meet these criteria have been removed from the table, including SPARC and INTS6, as correctly noted by the reviewer. Following this revision, Table 2 now includes a total of 34 proteins. Among these, 26 proteins meet both the fold-change (≥5-fold) and statistical significance (p < 0.05) criteria at both Pre-Op and Post-Op time points. Seven proteins meet the fold-change threshold in only one comparison and are indicated with a single asterisk (*). In addition, one protein (HBS1L), marked with a double asterisk (**), exceeds the fold-change threshold at both time points but reaches statistical significance only in the Post-Op comparison.
                  To improve transparency and prevent ambiguity, we have added a corresponding footnote to Table 2 clearly explaining this annotation. Additionally, we have ensured full consistency between Table 2 and the supplementary figures and datasets.
These revisions have strengthened the internal consistency of the dataset and improved the clarity of protein selection criteria. We thank the reviewer for this valuable suggestion, which has helped enhance the rigor of our analysis.

Comment 6: The manuscript completely omits cartilage acidic protein 1 (CRTAC1) from its protein lists, tables, and supplementary datasets , and provides no explanation for its absence. Given that CRTAC1 is currently the most robustly and independently replicated plasma OA biomarker in the field (e.g., Styrkarsdottir et al., 2021; Szilagyi et al., 2023; Kang et al., 2025), its absence in a study claiming to characterize "persistent OA molecular signatures" is a critical oversight. The authors must explicitly address whether this absence is a technical artifact—such as the systematic removal of low-abundance cartilage-derived proteins by the depletion strategy. Furthermore, while Cartilage Oligomeric Matrix Protein (COMP) is listed as a detected protein in Supplementary Table S1, its differential expression status is conspicuously absent from the main text and Discussion. The authors should explicitly report the detection status, fold-change direction, and statistical significance of both CRTAC1 and COMP, and discuss these findings to properly contextualize their results within the established OA biomarker landscape.

Response 6: We thank the reviewer for highlighting the importance of CRTAC1 and COMP as established biomarkers in osteoarthritis.
In our dataset, both CRTAC1 and COMP were detected in the proteomic analysis. However, neither protein met our predefined criteria for differential expression. Specifically, CRTAC1 did not reach statistical significance in either the Pre-Op or Post-Op comparisons, while COMP did not meet the fold-change threshold (|log₂FC| ≥ 1) at either time point. For this reason, these proteins were not included among the significantly altered protein groups presented in the main results.
                  To address the reviewer’s concern, we have now explicitly reported the detection status, fold-change direction, and statistical significance of both CRTAC1 and COMP in the revised manuscript and incorporated a discussion paragraph (Lines 691–706) to contextualize these findings within the established OA biomarker literature.
                  We further acknowledge that the absence of significant changes in these well-established biomarkers may reflect cohort size, biological variability, or technical factors such as depletion of low-abundance cartilage-derived proteins, and we have added this point to the Discussion as a limitation.

Comment 7: The detection of these typically intracellular proteins at dramatically elevated levels in depleted plasma is biologically unexpected and warrants explicit mechanistic discussion. The authors should note that their own Gene Ontology analyses (Supplementary Tables S2 and S3) identify "extracellular exosome" and "blood microparticle" as significantly enriched cellular components. This provides a plausible biological rationale: these proteins may enter the circulation via exosome shedding, microparticle release, or apoptotic leakage from stressed joint-associated cells. The manuscript must incorporate this discussion to address the unexpected localization. Furthermore, the authors should acknowledge the alternative explanation that some signals may reflect technical artifacts arising from non-specific enrichment during the high-abundance protein depletion process. Given these uncertainties, orthogonal validation (e.g., ELISA or targeted mass spectrometry) is strongly recommended for the most biologically unexpected hits to confirm their clinical relevance.

Response 7: We sincerely thank the reviewer for this insightful comment regarding the biological interpretation of intracellular proteins detected at elevated levels in depleted plasma.
In response, we have expanded the Discussion to address the potential mechanisms underlying this observation (Lines 653–676). Specifically, we now highlight that our Gene Ontology enrichment analyses (Supplementary Tables S2 and S3) identified “extracellular exosome” and “blood microparticle” as significantly enriched cellular components. Based on these findings, we incorporated a mechanistic interpretation suggesting that intracellular proteins may enter the circulation via extracellular vesicle release, including exosome shedding, microparticle formation, or apoptotic leakage from stressed or damaged joint-associated cells.
                  In addition, we have acknowledged the possibility that some of these observations may be influenced by technical factors, particularly non-specific enrichment during the high-abundance protein depletion process (Lines 677–684).
                  Importantly, we would like to emphasize that we performed experimental validation for one of the key differentially expressed proteins, C1QTNF6, using ELISA. C1QTNF6 was selected based on (i) its consistent and statistically significant downregulation in both Pre-Op and Post-Op comparisons in the proteomic dataset, (ii) its known involvement in extracellular matrix regulation and inflammatory pathways [1,2], and (iii) the availability of a validated commercial ELISA assay. The ELISA results confirmed significantly reduced C1QTNF6 levels in OA patients compared to healthy controls in both pre-operative and post-operative states, in agreement with the proteomic findings (Figure. 9).
                  We have now explicitly incorporated this validation step into the manuscript. The ELISA methodology has been described in the Materials and Methods section (Section 4.4.), and the corresponding results have been presented in the Results section (Section 2.7.).
                  We appreciate the reviewer’s suggestion, which has significantly strengthened the biological interpretation, methodological transparency, and the experimental validation framework of our study.

References:
[1]             Xu C, Sarver DC, Lei X, et al. CTRP6 promotes the macrophage inflammatory response, and its deficiency attenuates LPS-induced inflammation. J Biol Chem. 2024;300(1):105566. doi:10.1016/j.jbc.2023.105566
[2]             Xu E, Yin C, Yi X, Liu Y. Knockdown of CTRP6 inhibits high glucose-induced oxidative stress, inflammation and extracellular matrix accumulation in mesangial cells through regulating the Akt/NF-κB pathway. Clin Exp Pharmacol Physiol. 2020;47(7):1203-1211. doi:10.1111/1440-1681.13289

Comment 8: While Figure 7 identifies a distinct cluster of 23 proteins consistently downregulated at both Pre-Op and Post-Op time points, the current Discussion is disproportionately skewed toward upregulated targets. The sustained suppression of proteins like GULP1, MST1, SERPINA4, PZP, F10, and TTR likely represents a critical, yet overlooked, component of the chronic OA phenotype. Specifically, the downregulation of TTR in OA plasma, a protein the authors' own data (Supplementary Tables S2/S3) links to the Thyroid hormone synthesis pathway, cannot be ignored. The authors should expand their Discussion to provide a balanced interpretation of how this persistent downregulation contributes to OA pathogenesis.

Response 8: We thank the reviewer for this important and insightful comment.
In response, we have revised the Discussion to provide a more balanced interpretation of both upregulated and downregulated protein clusters. Specifically, we have expanded the discussion of the subset of proteins that were consistently downregulated across both Pre-Op and Post-Op time points, highlighting their potential role as a sustained component of the chronic OA phenotype (Lines 554–580).
We now discuss several representative proteins, including GULP1, MST1, SERPINA4, and F10, emphasizing their involvement in immune regulation, protease inhibition, coagulation, and cellular homeostasis. In addition, we have incorporated a dedicated discussion of transthyretin (TTR), noting its role in thyroid hormone transport and its association with the thyroid hormone synthesis pathway identified in our enrichment analysis (Supplementary Tables S2 and S3). We further interpret the consistent downregulation of TTR in the context of chronic inflammatory states and potential endocrine–metabolic dysregulation in OA (Lines 581–599).
                  We appreciate the reviewer’s suggestion, which has significantly improved the balance and biological depth of our Discussion.

Comment 9: In the Discussion (Lines 406–407), the authors use a genetic association study of TENT5A to justify their proteomic findings regarding TENT5D. However, this extrapolation is scientifically unjustified for two key reasons: first, TENT5A and TENT5D are distinct family members with partially non-overlapping functions; second, a SNP-level genetic association does not equivalently translate to circulating protein abundance in plasma. The authors should adopt a more conservative approach by clearly distinguishing these two proteins and acknowledging the absence of direct published evidence linking TENT5D specifically to OA pathobiology. Consequently, the elevated TENT5D expression should be presented strictly as a novel, exploratory finding rather than being grounded in inappropriate cross-member inference.

Response 9: We sincerely thank the reviewer for this important and insightful comment. We fully agree that extrapolating findings from TENT5A to TENT5D requires caution, as these proteins represent distinct members of the TENT5 (FAM46) family with potentially divergent biological functions. We also acknowledge that SNP-level genetic associations do not directly translate to circulating plasma protein abundance.
In response to the reviewer’s suggestion, we have revised the Discussion section to clearly distinguish TENT5D from other family members and have removed the previous cross-member inference involving TENT5A. The revised text now emphasizes that, although TENT5D belongs to the non-canonical poly(A) polymerase family implicated in post-transcriptional mRNA regulation, its specific role in OA remains poorly characterized. We further cite experimental evidence demonstrating that TENT5D participates in mRNA stability regulation in spermatogenesis, while explicitly acknowledging that no direct link between TENT5D and OA pathobiology has been previously established (Lines 437–452).
                  Accordingly, the elevated circulating TENT5D levels observed in our cohort are now presented as a novel and exploratory proteomic finding rather than being framed within inappropriate cross-member functional inference. We believe this revision strengthens the scientific rigor and interpretative accuracy of the manuscript.

Comment 10: The citation of McAlpine et al. (Lines 408–409) to support WDR41's role in OA inflammation is an over-extrapolation. That study focuses on the SMCR8-WDR41-C9ORF72 complex in ALS/FTD and systemic inflammation. The authors should revise the Discussion to present WDR41 merely as a novel candidate requiring future validation in OA models, rather than assuming its established lysosomal functions translate directly to joint inflammation.

Response 10: We thank the reviewer for this important and constructive comment. We agree that the study by McAlpine et al. (2018) investigated WDR41 within the SMCR8–C9ORF72 complex in the context of endosomal TLR signaling and systemic inflammatory disease models, rather than osteoarthritis specifically. We acknowledge that directly extrapolating these findings to joint inflammation could overstate the currently available evidence.
In response, we have revised the Discussion section to remove any implication that WDR41 has an established mechanistic role in OA. The revised text now clearly states that, although WDR41 is involved in lysosomal homeostasis and innate immune regulation in non-articular systems, its role in OA has not yet been investigated. We further contextualize its potential relevance by referencing established literature demonstrating that dysregulated autophagy and innate immune activation contribute to OA pathogenesis, while explicitly framing WDR41 as a novel candidate identified through our proteomic analysis (Lines 453–459).
                  Accordingly, WDR41 is now presented as a signal requiring validation in OA-specific experimental models rather than as evidence of a defined inflammatory mechanism in joint tissues. We believe this revision improves the interpretative precision and scientific rigor of the manuscript.

Comment 11: At Lines 416–418, the authors claim that ANKRD26 and ADCY8 upregulation suggests dysregulated cell survival and intracellular communication contributing to cartilage breakdown, citing Ref. 17 and Ref. 18. While extrapolating molecular mechanisms from other diseases is an acceptable practice, Ref. 18 (investigating endometrial cancer) lacks the direct functional evidence required to substantiate these specific claims in the context of osteoarthritis. To provide a scientifically sound discussion, the authors should either cite studies with concrete mechanistic evidence to justify this cross-tissue extrapolation or, preferably, ground their arguments in their own dataset. Specifically, Supplementary Tables S2 and S3 already show that ADCY8 is significantly enriched in contextually relevant pathways, such as Relaxin signaling and platelet activation. The authors should revise this passage to ensure their mechanistic claims are supported by robust literature or their own pathway-level findings.

Response 11: We thank the reviewer for this thoughtful and constructive comment. We agree that mechanistic extrapolation from unrelated disease contexts requires caution and stronger justification.
In response, we have revised the corresponding paragraph to remove Ref. 18, which was based on endometrial cancer data, and to avoid overinterpretation of the functional implications of ANKRD26 and ADCY8 in osteoarthritis. The discussion is now grounded primarily in our own pathway enrichment analyses (Supplementary Tables S2 and S3).
                  Specifically, we now emphasize that ANKRD26 is associated with platelet biology (Ref. 17), consistent with the strong enrichment of platelet activation and hemostasis pathways identified in our dataset (Lines 466–469). For ADCY8, we highlight its significant enrichment in Relaxin signaling, platelet activation, and actin cytoskeleton-related pathways, all of which are relevant to extracellular matrix regulation and thromboinflammatory processes in OA (Lines 469–475).
                  Furthermore, to provide mechanistic context without overstating causality, we have incorporated literature demonstrating that cAMP/CREB signaling modulates chondrocyte autophagy and osteoarthritis progression. The revised text adopts a more cautious interpretation and aligns mechanistic statements with both our pathway-level findings and relevant OA-focused literature.
                  We believe these revisions improve the scientific rigor and contextual relevance of the discussion.

Comment 12: The pathway enrichment data in Supplementary Tables S2 and S3 reveal several biologically significant findings that are entirely omitted from the Discussion. Specifically, the Relaxin signaling pathway (ACTA2, MMP2, COL4A3, ADCY8, PLCB1) and the Thyroid hormone synthesis pathway (TTR, ALB, ADCY8, PLCB1) reach statistical significance at both Pre-Op and Post-Op time points. Despite their direct relevance to extracellular matrix remodeling and joint tissue homeostasis, they remain unaddressed in the main text. Furthermore, the Post-Op dataset highlights the "Regulation of Complement cascade" (p = 0.046; involving IGHG3, CFHR1, C1R, IGKV3-11) as a newly enriched term absent from the Pre-Op analysis, suggesting a potential surgery-specific immune regulatory signature. The authors must revise the Discussion to incorporate these critical pathway observations or explicitly justify their exclusion from the mechanistic interpretation.

Response 12: We thank the reviewer for this insightful comment. In response, we have revised the Discussion section to incorporate the biologically relevant pathway enrichment findings identified in Supplementary Tables S2 and S3.
Specifically, we have now included an expanded interpretation of the Relaxin signaling pathway and the Thyroid hormone synthesis pathway, both of which were significantly enriched at pre- and post-operative time points. Their potential roles in extracellular matrix remodeling, chondrocyte regulation, and endocrine–metabolic modulation in OA pathophysiology are now explicitly discussed and supported by relevant literature (Lines 600–611).
                  In addition, we have incorporated a dedicated discussion of the “Regulation of complement cascade” pathway, which was uniquely enriched in the post-operative dataset. This finding is now interpreted as a potential surgery-related immune regulatory signature, and its relevance to inflammatory modulation and tissue remodeling processes has been contextualized within the current OA literature (Lines 611–619).
                  These additions have strengthened the biological interpretation of our enrichment analysis and improved the integration of pathway-level findings into the overall mechanistic framework of the study.

Comment 13: The enrichment of platelet activation and complement/coagulation cascades reported in Supplementary Tables S2 and S3, and discussed in the text, overlaps substantially with prior published work that the authors fail to cite. For instance, Kraus et al. (2023) reported a serum diagnostic peptide panel for knee OA with prominent complement and coagulation enrichment. Similarly, Naili et al. (2025) described targeted plasma proteomics in a knee OA cohort with Reactome enrichment dominated by platelet activation and neutrophil degranulation, which is strikingly parallel to the current findings. Furthermore, Zhang et al. (2024) demonstrated that plasma extracellular vesicle proteins, including fibrinogen chains, predict radiographic OA progression, linking these same pathways to disease outcomes. The convergence of the current data with these independent studies is actually a major strength, as it suggests the authors' findings are biologically reproducible despite the limited sample size. However, presenting these enrichment results as completely novel discoveries without acknowledging the prior concordant literature significantly overstates the manuscript's contribution. The authors must revise the Discussion to properly frame these findings as a corroboration and extension of an established proteomic pattern in OA plasma, accompanied by the appropriate citations.

Response 13: We thank the reviewer for this important and insightful comment. We fully agree that the enrichment of platelet activation and complement/coagulation pathways should be interpreted in the context of existing OA proteomic literature.
Accordingly, we have revised the Discussion section to incorporate and appropriately cite prior studies (including Kraus et al., 2023; Naili et al., 2025; and Zhang et al., 2024). We now explicitly acknowledge the substantial overlap between our findings and these independent reports. Rather than presenting our results as entirely novel, we frame them as a confirmation and extension of an established plasma proteomic signature in OA (Lines 620–634).
                  These revisions provide a more balanced interpretation of our findings and position our study within the broader context of reproducible proteomic patterns in OA.

Comment 14: The manuscript interprets the 63 persistently elevated proteins and the 6 post-operative specific proteins as if they represent biologically uniform categories defined solely by the surgical intervention, yet it fails to address the well-established molecular heterogeneity within the OA patient population. Current evidence distinguishes a hyperinflammatory endotype characterized by complement activation and Fc receptor signaling from a hypoinflammatory structural endotype, and these groups respond differently to both pharmacological and surgical treatments. For example, Xie et al. (2024) demonstrated that pre-operative plasma proteomic profiles stratify knee OA patients into distinct treatment responder groups, while Giordano et al. (2023) showed that pre-operative serum inflammatory cluster assignments predict distinct post-arthroplasty pain trajectories. Given the small cohort of only eight patients that mixes knee and hip arthroplasty cases, the observed split between persistent and surgery-responsive proteins may reflect pre-existing endotype heterogeneity just as much as a true surgery-driven biological transition. The authors must cite this relevant endotype literature, explicitly acknowledge this alternative interpretation in the Discussion, and note that future studies should incorporate pre-operative immune phenotyping to disambiguate baseline disease heterogeneity from surgery-driven proteomic changes.

Response 14: We thank the reviewer for this insightful and important comment. We fully agree that molecular heterogeneity and OA endotypes provide a critical framework for interpreting plasma proteomic findings. In the revised Discussion section, we have incorporated and cited the relevant endotype literature. We now explicitly acknowledge that the observed distinction between persistently altered and post-surgery–specific proteins may reflect underlying biological heterogeneity, rather than being solely attributable to surgical intervention (Lines 635–646).
Furthermore, given the small and clinically heterogeneous cohort, we emphasize that baseline immune endotypes may have influenced the observed proteomic trajectories. We have also added a statement highlighting that future studies should incorporate preoperative immune phenotyping and endotype-based stratification to distinguish baseline disease heterogeneity from true surgery-induced proteomic changes (Lines 646–652).

Comment 15: The manuscript completely omits a Limitations section, which is a significant oversight for a discovery-phase proteomic study of this design. The most critical unaddressed limitation is the sample size: with only eight OA patients and ten controls (n = 18), the statistical power to reliably identify 93 differentially abundant proteins is severely limited, and the risk of false positives remains a legitimate concern despite Benjamini–Hochberg correction. Several other crucial limitations must be explicitly disclosed: the single post-operative time point at six weeks primarily captures the acute recovery phase rather than medium- or long-term molecular trajectories; the cohort mixes knee and hip arthroplasty patients without subgroup analysis, overlooking potential site-specific molecular phenotypes; and the study lacks any orthogonal protein-level validation (e.g., ELISA) for its high-dimensional findings. The authors must add a dedicated Limitations paragraph to discuss these issues and tone down the language throughout their conclusions to accurately reflect that these signatures are preliminary, exploratory findings rather than established clinical biomarkers.

Response 15: We thank the reviewer for this important and constructive comment. We fully agree that a dedicated limitations section is essential for a discovery-phase proteomic study of this design. Accordingly, we have now added a comprehensive Limitations paragraph to the revised manuscript (Lines 707–728).
In this section, we explicitly address the relatively small sample size and the associated risk of false-positive findings despite multiple testing correction, and we have tempered the language throughout the manuscript to reflect the exploratory nature of our results. We also acknowledge the limitation of including only a single post-operative time point, which primarily captures early recovery rather than long-term molecular trajectories. Furthermore, we now discuss the potential impact of cohort heterogeneity due to the inclusion of both knee and hip arthroplasty patients without subgroup analyses, as well as technical considerations related to high-abundance protein depletion strategies. In addition, we clarify that while C1QTNF6 was validated using ELISA, the majority of identified proteins have not yet undergone orthogonal validation, and we emphasize the need for future targeted validation studies.
                  These revisions ensure that the conclusions are appropriately moderated and that the findings are clearly presented as preliminary and hypothesis-generating.

Comment 16: The Materials and Methods section contains systematic numbering errors that require correction. Specifically, the mass spectrometry subsection (Line 568) is incorrectly labeled "2.3.3. Mass Spectrometry Analysis" instead of 4.3.3 , and the subsequent bioinformatics subsection (Line 586) is improperly labeled "2.4. Bioinformatics and Statistical Analysis" instead of 4.4. These formatting oversights suggest a lack of careful proofreading prior to submission. The authors must thoroughly review and correct all section headings, numbering, and structural cross-references throughout the manuscript to meet the journal's publication standards.

Response 16: We thank the reviewer for carefully identifying these formatting inconsistencies. The incorrect section numbering in the Materials and Methods section has been fully corrected (e.g., Sections 4.3.3. and 4.4.), and all headings and cross-references throughout the manuscript have been systematically reviewed and revised to ensure consistency with the journal’s formatting requirements. In addition, the entire manuscript has been carefully proofread to prevent similar structural or typographical errors.

Comment 17: In Section 4.3.3, the authors describe HDMSE acquisition on a SYNAPT G2-Si system using a fixed elevated collision energy of 30 V during the high-energy scan. In many HDMSE workflows for complex proteomic digests, a collision energy ramp (for example, 15–40 V) is commonly implemented to accommodate peptides with diverse sizes and charge states and to ensure balanced fragmentation efficiency across the m/z range. It would therefore be helpful for the authors to clarify whether the 30 V setting was applied as a fixed value across the entire m/z range or selected based on prior optimization or vendor-validated preset parameters. Providing a brief explanation of the rationale for this setting would enhance methodological transparency and reproducibility without affecting the overall strength of the reported proteomic findings.

Response 17: We thank the reviewer for this important methodological point. During the optimization of our proteomics analysis, we observed that a fixed collision energy of 30 V provided a stable and reproducible fragmentation pattern with sufficient peptide coverage and signal intensity for downstream identification and quantification. We had most effective response from the instrument, so we are using 30 V settling for HDMSE workflow.

Comment 18: While data are available upon request (Lines 649–650), it is highly recommended that the authors deposit their raw proteomics data into a recognized public repository like PRIDE and provide the accession number.

Response 18: Thank you for your kind recommendation. Our raw data were deposited into Japan Proteome Standard Repository (jPOST), which is a very secure unit. The URL and password, which belong to our data, were given below. We would like to specifically mention that the URL’s and the Access keys are single-use only.

URL: https://repository.jpostdb.org/preview/120174109669a5bc9f2c2e4

Access key: 3993

URL: https://repository.jpostdb.org/preview/182997505669a5bdc411339

Access key: 7996

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I think the manuscript is acceptable after revision.

Author Response

We sincerely thank the reviewer for the careful re-evaluation of our manuscript and for the positive assessment of the revisions. We are grateful for the reviewer’s supportive comments and for confirming that the previous concerns have been adequately addressed. We appreciate the reviewer’s time and constructive contribution to improving the quality of our work.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have adequately addressed the concerns raised in the previous review. They have substantially revised the Abstract, corrected all erroneous citations (Lines 74–76, Refs 7–11), added p-values to Table 1, and resolved the CXCL annotation to CXCL4V1 (Lines 210, Table 2). Table 2 has been refined with the removal of SPARC/INTS6 and addition of a transparent footnote (Table 2). The manuscript is notably improved by incorporating ELISA validation of C1QTNF6 (Section 2.7, Lines 401–413; Section 4.4, Lines 818–837; Figure 9), and by explicitly evaluating CRTAC1 and COMP (Lines 688–703). The authors have significantly deepened the Discussion by incorporating detailed mechanistic roles for downregulated proteins (Lines 551–596), a conservative reinterpretation of TENT5D (Lines 434–449) and WDR41 (Lines 450–456), and pathway integration for Relaxin/Thyroid/Complement (Lines 597–616) supported by the addition of relevant literature and the removal of the irrelevant endometrial cancer citation. Furthermore, the authors successfully aligned their findings with prior literature (Lines 617–631) by citing independent studies (Refs 48–50), and explicitly acknowledged the potential confounding role of OA endotype heterogeneity (Lines 632–649). Finally, comprehensive limitations have been added (Lines 704–725), and section numbering has been corrected (Lines 799, 818, 838).

However, regarding data availability, while the mass spectrometry data has been deposited in jPOST (PXD075088), it currently remains inaccessible. The authors must ensure that this dataset becomes fully public upon publication. In addition, one minor issue remains in Section 2.7: the statement in Lines 405–406 that C1QTNF6 has an “established role in extracellular matrix regulation and inflammation” should be supported by an appropriate citation. Once this minor point is addressed, the manuscript would be acceptable for publication.

Author Response

Comment 1: The authors have adequately addressed the concerns raised in the previous review. They have substantially revised the Abstract, corrected all erroneous citations (Lines 74–76, Refs 7–11), added p-values to Table 1, and resolved the CXCL annotation to CXCL4V1 (Lines 210, Table 2). Table 2 has been refined with the removal of SPARC/INTS6 and addition of a transparent footnote (Table 2). The manuscript is notably improved by incorporating ELISA validation of C1QTNF6 (Section 2.7, Lines 401–413; Section 4.4, Lines 818–837; Figure 9), and by explicitly evaluating CRTAC1 and COMP (Lines 688–703). The authors have significantly deepened the Discussion by incorporating detailed mechanistic roles for downregulated proteins (Lines 551–596), a conservative reinterpretation of TENT5D (Lines 434–449) and WDR41 (Lines 450–456), and pathway integration for Relaxin/Thyroid/Complement (Lines 597–616) supported by the addition of relevant literature and the removal of the irrelevant endometrial cancer citation. Furthermore, the authors successfully aligned their findings with prior literature (Lines 617–631) by citing independent studies (Refs 48–50), and explicitly acknowledged the potential confounding role of OA endotype heterogeneity (Lines 632–649). Finally, comprehensive limitations have been added (Lines 704–725), and section numbering has been corrected (Lines 799, 818, 838).

Response 1: We sincerely thank the reviewer for the thorough re-evaluation of our revised manuscript and for recognizing the substantial improvements made in response to the previous round of comments. We greatly appreciate the reviewer’s positive feedback regarding the revisions to the Abstract, correction of citations, improvements to Table 1 and Table 2, clarification of the CXCL annotation, the addition of ELISA validation for C1QTNF6, and the expanded mechanistic interpretation throughout the Discussion.
                  We are pleased that the reviewer considers the manuscript substantially improved. Below we address the remaining minor points raised in the current review.

 

Comment 2: However, regarding data availability, while the mass spectrometry data has been deposited in jPOST (PXD075088), it currently remains inaccessible. The authors must ensure that this dataset becomes fully public upon publication.

Response 2:  We thank the reviewer for this important comment regarding data accessibility. The raw mass spectrometry data have been deposited in the Japan Proteome Standard Repository (jPOST) under accession number PXD075088.
                  To facilitate access during the review process, we generated private access links and keys that allow the reviewers to inspect the dataset. Although the site was briefly inaccessible due to maintenance, it is now fully operational, and the new credentials will ensure seamless access for the reviewers. Should any issues arise, please contact the corresponding authors. Additionally, we have configured the repository to become fully public immediately upon publication. These adjustments align the data availability timeline with the journal’s schedule and ensure full transparency and accessibility of the dataset for the scientific community.

Correspondence:         duygu.sariak@sbu.edu.tr; Tel.: +90 533 511 14 60
                                                      mustafa.beker@medeniyet.edu.tr; Tel.: +90 535 888 85 01

                  Single-used URLs and their access keys are below:

URL: https://repository.jpostdb.org/preview/78300640369b7e866b7f42

Access key: 6221

URL: https://repository.jpostdb.org/preview/192061319469b7e8a7d999f

Access key: 5868

URL: https://repository.jpostdb.org/preview/21448193969b7e8ba19bd3

Access key: 2596

URL: https://repository.jpostdb.org/preview/21448193969b7e8ba19bd3

Access key: 2596

URL: https://repository.jpostdb.org/preview/127110470669b7e8e5c5bde

Access key: 3180

URL: https://repository.jpostdb.org/preview/118987657769b7e90bac4df

Access key: 1331

 

Comment 3:  In addition, one minor issue remains in Section 2.7: the statement in Lines 405–406 that C1QTNF6 has an “established role in extracellular matrix regulation and inflammation” should be supported by an appropriate citation. Once this minor point is addressed, the manuscript would be acceptable for publication.

Response 3: We thank the reviewer for identifying this point. In response, we have added an appropriate supporting reference to substantiate the statement describing the involvement of C1QTNF6 in extracellular matrix regulation and inflammatory processes.
                  This citation has been incorporated into the revised manuscript at Line 407, where it has been highlighted in yellow for clarity. This addition ensures that the statement is properly supported by relevant literature and improves the scientific rigor.