Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The present study aims to evaluate the effects of enoxaparin, tinzaparin, apixaban, and the natural compound quercetin on cancer cell procoagulant activity and endothelial responses. The authors have applied Biochemical, Cellular Biology and Molecular Biology techniques to examine the effects of anticoagulant drugs to tumor-related procoagulant mechanisms. Their results show that tinzaparin and quercetin reduced cancer cell viability, while all agents decreased tissue factor–dependent thrombin generation. In addition, quercetin also protected endothelial cells from dysfunction induced by cancer-derived extracellular vesicles. The authors conclude that their findings suggest that certain anti-coagulant agents and natural compounds may modulate tumor-driven hypercoagulability beyond their conventional anticoagulant action.

The study is well-organized and the manuscript is well-written. The methods are clearly described, and the results are well presented and discussed. However, there are some minor points that need to be revised by the authors, as follows.

1. It is advisable to avoid abbreviations in the title, as they provide no meaningful assistance to the reader.
2. The authors write [lines254-259] that “the maximal reduction in HUVEC proliferation was reached after 24 h of exposure, indicating a plateau of the inhibitory effect”, but they continued the experiment to 72 h, without any explanation neither any diagram. Similarly, they must also show the results of changes in morphology and survival of cells at 24h by including three additional frames at figure 3 [or by a supplementary figure].

Author Response

Reviewer 1:
1. It is advisable to avoid abbreviations in the title, as they provide no meaningful assistance to the
reader.
Answer: We thank the reviewer for this suggestion. We have revised the title to remove all
abbreviations.
2. The authors write [lines254-259] that “the maximal reduction in HUVEC proliferation was reached after
24 h of exposure, indicating a plateau of the inhibitory effect”, but they continued the experiment to 72
h, without any explanation neither any diagram. Similarly, they must also show the results of changes in
morphology and survival of cells at 24h by including three additional frames at figure 3 [or by a
supplementary figure].
Answer: We thank the reviewer for this important comment. We acknowledge that an explanation
for continuing the experiment to 72 h was not provided in the original manuscript. The rationale for
extending the experiment beyond the 24 h plateau was that, although maximal reduction in HUVEC
proliferation was reached at 24 h, subtle but consistent effects remained detectable throughout
the 72 h observation period. Limiting the experiment to 24 h would have overlooked these
persistent effects, which are biologically relevant in the context of sustained endothelial
dysfunction induced by CaCe-dEVs. We have added this clarification to the revised manuscript.
Page. 8, line. 291.
Answer: Regarding the morphological and survival changes at 24 h, we acknowledge that
representative images at this time point are not available in the current dataset. We have therefore
added a statement in the revised manuscript clarifying that morphological assessment was
performed at 72 h, as these time points were selected to capture the progressive nature of CaCedEV–induced endothelial changes.
Page. 7, line. 266.

Reviewer 2 Report

Comments and Suggestions for Authors

1. Please specify the sources and manufacturers of the low-molecular-weight heparins (LMWHs; enoxaparin and tinzaparin), as well as apixaban and quercetin, including relevant product details (e.g., catalog numbers, purity).

2. Please clarify the observed differential effects of enoxaparin and tinzaparin on cancer cell viability. Specifically, tinzaparin significantly reduced cell viability, whereas enoxaparin showed no apparent cytotoxicity. A mechanistic explanation or supporting references would strengthen this observation.

3.The compounds were tested using different concentration units—enoxaparin and tinzaparin (anti-Xa IU/mL), apixaban (μg/mL), and quercetin (nM). Please justify this approach and indicate how the effects of these agents can be appropriately compared across different unit systems.

4. Please define the abbreviation “DOACs” at its first occurrence in the manuscript (e.g., direct oral anticoagulants).

5. Table 4 seems not necessary.

Author Response

Reviewer 2:
1. Please specify the sources and manufacturers of the low-molecular-weight heparins (LMWHs;
enoxaparin and tinzaparin, as well as apixaban and quercetin, including relevant product details (e.g.,
catalog numbers, purity).
Answer: We thank the reviewer for this important comment. We have revised the Materials and
Methods section to include full product details for LMWHs, apixaban and quercetin.
Page. 4, line. 153.
2. Please clarify the observed differential effects of enoxaparin and tinzaparin on cancer cell viability.
Specifically, tinzaparin significantly reduced cell viability, whereas enoxaparin showed no apparent
cytotoxicity. A mechanistic explanation or supporting references would strengthen this observation.
Answer: We thank the reviewer for this important comment. To address this concern, a supporting
paragraph has been added to the Discussion section providing a mechanistic explanation for the
observed differential effects of tinzaparin and enoxaparin on cancer cell adherent biomass, with
reference to the well-established structural distinctions between these two LMWHs and their
differential modulation of heparin-binding proteins and downstream PAR-1/ERK signaling
pathways
Page. 13, line. 456.
3. The compounds were tested using different concentration units—enoxaparin and tinzaparin (anti-Xa
IU/mL), apixaban (μg/mL), and quercetin (nM). Please justify this approach and indicate how the effects
of these agents can be appropriately compared across different unit systems.
Answer: We thank the reviewer for this observation. The use of different concentration units for the
tested compounds reflects the standard pharmacological conventions specific to each drug class.
Enoxaparin and tinzaparin are heterogeneous polysaccharide mixtures whose biological activity
cannot be accurately expressed by mass concentration; therefore, anti-Xa IU/mL is the
internationally accepted unit for quantifying their functional potency. Apixaban, as a chemically
defined small molecule, was expressed in μg/mL, and quercetin in nanomolar (nM) concentrations,
consistent with their respective physicochemical properties.
The aim of this study was not to perform a direct quantitative potency comparison between these
mechanistically distinct agents, but rather to evaluate the individual biological effect of each
compound on cancer cell viability across a range of experimentally derived concentrations. The
concentrations used were determined empirically through preliminary dose-response
experiments conducted in our laboratory to identify biologically active ranges for each compound
in the tested cell lines. Therefore, cross-unit comparison was neither intended nor required, as
each agent was assessed independently within its own experimentally established concentration
framework.
Page. 4, line. 153.
4. Please define the abbreviation “Direct Oral AantiCoagulants” at its first occurrence in the manuscript
(e.g., direct oral anticoagulants).
Answer: We thank the reviewer for this observation. We have revised the manuscript to ensure that
the abbreviation "DOACs" (Direct Oral AntiCoagulants) is fully defined at its first occurrence in the
text, as follows: "direct oral anticoagulants (DOACs)." This correction has been applied throughout
the revised manuscript to ensure consistency.
Page. 3, line. 104.
5. Table 4 seems not necessary.
Answer: We thank the reviewer for this observation. Table 4 serves as a concise summary of the key
findings across multiple experimental conditions, cell types, and antithrombotic agents, allowing
readers to rapidly grasp the main results without cross-referencing multiple figures and sections.
We believe it significantly enhances the clarity and accessibility of the manuscript and kindly
request that it be retained. However, as a compromise, we are willing to move it to the
Supplementary Materials if the editor considers it more appropriate. 

Reviewer 3 Report

Comments and Suggestions for Authors

Overall Recommendation: Accept after major revisions.

1. In Section 2.1, the authors used two cell lines, BxPC3 and MCF7. However, I doubt whether these two cell lines can sufficiently represent cancers in general. Would the authors consider adding one squamous cell carcinoma cell line with high TF expression and one hematological malignancy cell line to improve the generalizability of the conclusions?
2. In Section 2.1, HUVECs were used to mimic the tumor microenvironment. Nevertheless, HUVECs are human umbilical vein endothelial cells, rather than tumor-associated endothelial cells.
3. In Section 2.2, the ZYMUPHEN MP-Activity kit was adopted to quantify TF activity, yet the purity, particle size distribution and concentration of extracellular vesicles (EVs) were not validated. In my opinion, this cannot be regarded as standard EV isolation. Would the authors consider supplementing NTA or TRPS characterization data?
4. In Section 2.3, it remains unclear how the EV concentration of 360 nM was determined. The authors are required to provide a dose-response curve for validation.
5. In Section 2.4, quercetin was used at concentrations of 166.5 μM and 333 μM, which are far above physiologically relevant levels. The authors need to elaborate on the rationale for selecting such high concentrations.
6. In Section 2.6, the MTT assay was used to measure cell viability. As is known, the MTT assay detects cellular metabolic activity only; have the authors confused these fundamental concepts?
7. In Section 3.2, the authors stated that cell confluence was reduced to 20%. I would like to know how this percentage was quantified. Visual estimation, if adopted, would be scientifically inappropriate.
8. Table 2 appears to lack P-values. In addition, all tables should be formatted as three-line tables.
9. In Figure 4, individual data points are not displayed on the bar graphs. Most current journals require bar graphs overlaid with scatter plots to show raw data.
10. In Line 410, the authors claimed that tinzaparin possesses anti-cancer activity. I consider such a conclusion arbitrary in the absence of in vivo experimental models.
11. The statement in the Data Availability section is awkwardly worded. I recommend the authors revise the phrasing, for instance, by stating that relevant data are available upon reasonable request.
12. The number of references in this manuscript is limited, and some cited literature is outdated. I recommend the authors add the following references to enhance the quality of the manuscript:
It is recommended to cite the literature (DOI: 10.1016/j.jpet.2026.103836) in the Introduction or Discussion section, to compensate for the limitation of this study relying merely on simple in vitro experiments without systematic mechanistic exploration.
It is recommended to cite the literature (DOI: 10.3390/ph18060900) in the Introduction or Discussion section. This study reported the anti-tumor effects of quercetin, which can strengthen the evidence supporting the anti-cancer activity of quercetin in the present work.
It is recommended to cite the literature (DOI: 10.3748/wjg.v31.i25.107893) in the Introduction or Discussion section, to support the hypothesis that anticoagulant drugs exert anti-cancer effects via cell cycle regulation.

Comments on the Quality of English Language

The English could be improved.

Author Response

Reviewer 3:
1. In Section 2.1, the authors used two cell lines, BxPC3 and MCF7. However, I doubt whether these two
cell lines can sufficiently represent cancers in general. Would the authors consider adding one
squamous cell carcinoma cell line with high TF expression and one hematological malignancy cell line
to improve the generalizability of the conclusions?
Answer: We thank the reviewer for this observation. The selection of BxPC-3 and MCF-7 was
deliberate, based on their contrasting procoagulant profiles rather than cancer type
representation. BxPC-3 exhibits high TF expression and strong procoagulant potential, while MCF7 represents a low procoagulant phenotype. This contrast was essential to investigate the
relationship between procoagulant activity and antithrombotic agent effects.
This justification has been explicitly added to the Materials and Methods section with supporting
references.
Page. 3, line. 115, references. 10,24,25.
2. In Section 2.1, HUVECs were used to mimic the tumor microenvironment. Nevertheless, HUVECs are
human umbilical vein endothelial cells, rather than tumor-associated endothelial cells.
Answer: We thank the reviewer for this comment. We acknowledge that HUVECs represent
classical endothelial cells and are not tumor-associated endothelial cells. We wish to clarify that
the intention of using HUVECs was not to mimic the tumor microenvironment, but rather to
evaluate the effect of the tested compounds on a well-established standardized endothelial cell
model.
This concept has been further developed in the Discussion section, where we explicitly distinguish
between classical endothelial cell behavior as modeled by HUVECs and the tumor-associated
endothelial compartment.
Page. 15, line. 566.
3. In Section 2.2, the ZYMUPHEN MP-Activity kit was adopted to quantify TF activity, yet the purity,
particle size distribution and concentration of extracellular vesicles (EVs) were not validated. In my
opinion, this cannot be regarded as standard EV isolation. Would the authors consider supplementing
NTA or TRPS characterization data?
Answer: We thank the reviewer for this observation. We respectfully clarify that the isolation
protocol was not intended for comprehensive EV characterization, but specifically for functional
quantification of tissue factor activity in the microparticle-enriched fraction using the validated
ZYMUPHEN MP-Activity kit, in accordance with our referenced validated protocol. NTA or TRPS
characterization, while valuable in EV biology studies, is not a mandatory requirement in the
context of functional microparticle-associated TF activity assays, which is a well-established
approach in the thrombosis and cancer biology literature.
This clarification has been explicitly added to the Materials and Methods section of the revised
manuscript.
Page. 3, line. 133.
4. In Section 2.3, it remains unclear how the EV concentration of 360 nM was determined. The authors
are required to provide a dose-response curve for validation.
Answer: We thank the reviewer for this observation. We wish to clarify that the concentration of 360
nM was not arbitrarily selected but was determined through preliminary dose-response
experiments in which HUVECs were exposed to increasing concentrations of CaCe-dEVs (120, 240,
and 360 nM) over multiple time points (24, 48, and 72 h). Concentrations of 120 and 240 nM failed to
induce any detectable morphological changes or tissue factor expression in HUVECs at any time
point, whereas exposure to 360 nM for 72 h induced marked and reproducible morphological
changes, establishing this as the minimum effective concentration under our experimental
conditions.
This rationale has been further clarified at the end of Section 2.3 of the revised manuscript.
Page. 4, line. 148.
5. In Section 2.4, quercetin was used at concentrations of 166.5 μM and 333 μM, which are far above
physiologically relevant levels. The authors need to elaborate on the rationale for selecting such high
concentrations.
Answer: We acknowledge the reviewer's observation regarding the discrepancy between the
quercetin concentrations used in this study and physiologically achievable plasma levels following
oral administration. However, quercetin was employed exclusively as an in vitro positive control to
confirm the responsiveness of the cancer cell lines to a compound with established
antiproliferative activity, and was not intended to reflect clinical translatability. The use of
supraphysiological quercetin concentrations in vitro is accepted practice in cell-based assays
where the goal is mechanistic or comparative rather than pharmacokinetic. We fully agree that the
in vivo relevance of quercetin's effects at these concentrations would require targeted delivery
strategies to achieve comparable tissue exposures, and that further studies are warranted to
better characterize its pharmacological profile and optimize its delivery for maximum anticancer
efficacy in a clinically relevant context.
This limitation has been explicitly acknowledged in the revised Discussion section.
Page. 13, line. 500.
6. In Section 2.6, the MTT assay was used to measure cell viability. As is known, the MTT assay detects
cellular metabolic activity only; have the authors confused these fundamental concepts?
Answer: We acknowledge that the MTT assay measures mitochondrial metabolic activity rather
than cell viability directly. However, as widely accepted in the literature, MTT results are commonly
reported as an indirect measure of cell viability, since reduced formazan production reflects
decreased metabolic activity associated with reduced cell survival. Nevertheless, we recognize
that this terminology was imprecise and have revised the manuscript accordingly, replacing "cell
viability" with "cellular metabolic activity" throughout Section 2.6 and the corresponding materials
and methods, results, and conclusion sections, to accurately reflect what the MTT assay
measures.
Page. 4, line. 180.
7. In Section 3.2, the authors stated that cell confluence was reduced to 20%. I would like to know how
this percentage was quantified. Visual estimation, if adopted, would be scientifically inappropriate.
Answer: We thank the reviewer for this important methodological observation. We wish to clarify
that cell confluence was not estimated visually but was quantified objectively using automated
image segmentation of inverted microscopy images. Cell-covered area was measured and
expressed as a percentage of the total imaged surface according to the following formula:
Confluence (%) = Cell-covered area / Total image area × 100
This methodology has been explicitly detailed in the revised Materials and Methods (section 2.9) to
avoid any ambiguity regarding the quantification approach.
Page. 5, line. 209.
8. Table 2 appears to lack P-values. In addition, all tables should be formatted as three-line tables.
Answer: We thank the reviewer for this observation. We have carefully reviewed Table 2 and
respectfully note that P-values are present for all comparisons reported in the table. We kindly ask
the reviewer to re-examine Table 2 in the revised manuscript. If there are specific comparisons for
which P-values appear to be missing or unclear, we would welcome further clarification so that we
can address this concern precisely.
Answer: We thank the reviewer for this suggestion. While we appreciate the recommendation to
adopt a uniform three-line table format, we respectfully note that the tables in this manuscript
present data from distinct experimental contexts with varying complexity. Reformatting all tables
to a strict three-line structure would compromise the clarity and readability of the data presented.
We have therefore retained the current table formatting to ensure that all experimental data are
presented with sufficient clarity and transparency.
9. In Figure 4, individual data points are not displayed on the bar graphs. Most current journals require
bar graphs overlaid with scatter plots to show raw data.
Answer: We respectfully acknowledge the reviewer's suggestion. However, data variability in
Figure 4 is adequately represented through standard deviation error bars, which remains a wellaccepted and widely adopted reporting standard in pharmacological cell viability studies. The
presentation is consistent with the guidelines of the journal to which this manuscript is submitted,
which does not explicitly require individual data point overlay.
10. In Line 410, the authors claimed that tinzaparin possesses anti-cancer activity. I consider such a
conclusion arbitrary in the absence of in vivo experimental models.
Answer: We respectfully clarify that Line 410 does not explicitly claim anti-cancer activity in a
broad oncological sense, but rather reports the specific in vitro observation that tinzaparin reduced
cancer cell viability under the tested experimental conditions. While in the discussion, we
mentioned that "Based on these data, future pharmacological studies on a potential anticancer
effect of heparins should focus on the low-affinity material for the antithrombin. Moreover, the
rationality for the concentration of this LAM should not follow that of the clinically relevant anti-Xa
activity" — this statement reflects our cautious interpretation of the in vitro data and confirms that
we did not intend to assert definitive anti-cancer activity beyond the scope of the current
experimental model. We fully agree with the reviewer that in vivo validation is necessary before any
definitive anti-cancer claim can be made.
Page. 14, line. 540.
11. The statement in the Data Availability section is awkwardly worded. I recommend the authors revise
the phrasing, for instance, by stating that relevant data are available upon reasonable request.
Answer: We thank the reviewer for this suggestion. The Data Availability statement has been
revised to read: "The data supporting the findings of this study are available upon reasonable
request from the corresponding author."
Page. 18, line. 606.
12. The number of references in this manuscript is limited, and some cited literature is outdated. I
recommend the authors add the following references to enhance the quality of the manuscript:
It is recommended to cite the literature (DOI: 10.1016/j.jpet.2026.103836) in the Introduction or
Discussion section, to compensate for the limitation of this study relying merely on simple in vitro
experiments without systematic mechanistic exploration.
It is recommended to cite the literature (DOI: 10.3390/ph18060900) in the Introduction or Discussion
section. This study reported the anti-tumor effects of quercetin, which can strengthen the evidence
supporting the anti-cancer activity of quercetin in the present work.
It is recommended to cite the literature (DOI: 10.3748/wjg.v31.i25.107893) in the Introduction or
Discussion section, to support the hypothesis that anticoagulant drugs exert anti-cancer effects via cell
cycle regulation.
Answer: We thank the reviewer for this constructive suggestion. We agree that supplementing the
manuscript with additional and more recent references strengthens the scientific context.
Accordingly, we have incorporated the recommended citations to introduction and discussion.
Page. 2, line. 90, reference. 17.
Page. 2, line. 95, reference. 23.
Page. 15, line. 557, reference. 41.

Reviewer 4 Report

Comments and Suggestions for Authors

1. Two co-authors (A. Rousseau and P. Van Dreden) are explicitly affiliated with the Clinical Research Department of Diagnostica Stago, whose proprietary CAT/Thrombinoscope system and plasma reagents constitute the primary assay platform of the entire study. The COI statement ('no known competing financial interests') does not acknowledge this structural conflict.
2. Diagnostica Stago's contribution is acknowledged in the Funding/Acknowledgements section, but this does not satisfy the Cancers journal requirement for a specific conflict-of-interest declaration covering commercial relationships that could influence study design, reagent selection, or results reporting.
3. Quercetin was tested at 166.5 and 333 µM. Peak human plasma concentrations after oral supplementation are typically below 1–2 µM. The concentrations used in this study are therefore 100–200-fold above achievable levels, making the cytotoxic and endothelial-protective effects observed with quercetin non-translatable without targeted delivery strategies.
4. The study positions quercetin as a 'reference agent with proven antioxidant and antitumor properties,' yet uses it at concentrations that are pharmacologically irrelevant in vivo. This framing misleads the reader about its comparative utility relative to the anticoagulants, which are tested at clinically relevant concentrations.
5. The Table 2 caption refers to quercetin concentrations in nM (166.5 nM and 333 nM), while the Methods (Section 2.4) and Results correctly state µM (166.5 µM and 333 µM). This is a factual inconsistency that must be corrected throughout.
6. Section 2.10 states that comparisons were made using 'paired one-way ANOVA,' but the basis for pairing is never described. A one-way ANOVA is not inherently a paired test; if repeated-measures ANOVA was intended, this must be stated explicitly and the within-subject matching structure described.
7. The study performs numerous pairwise comparisons across two cell lines, four agents, two concentrations, and three thrombin generation parameters (lag-time, ETP, Peak). No correction for multiple comparisons (e.g., Bonferroni, Holm, or Benjamini-Hochberg) is reported. This increases the family-wise Type I error rate substantially and may result in false-positive findings.
8. p-values alone are reported throughout. No effect sizes (e.g., Cohen's d or partial η²) or 95% confidence intervals are provided for any comparison, limiting readers' ability to assess the practical significance of the reported differences.
9. Tinzaparin reduced BXPC3 viability by 46% and MCF7 viability by 29%, while enoxaparin at the same anti-Xa activity had no cytotoxic effect. This is a significant and potentially important finding, yet no mechanistic data are provided — no caspase activation, no apoptosis/necrosis discrimination by Annexin V/PI co-staining, no cell cycle analysis — to confirm that the reduction in crystal violet absorbance reflects genuine cytotoxicity rather than detachment or growth arrest.
10. Crystal violet measures total adherent biomass, not cell death. A reduction in absorbance may reflect cell detachment, reduced proliferation, or altered cell adhesion independently of cytotoxicity. The conclusion that tinzaparin 'reduced cancer cell viability' requires confirmation using a cell-death-specific assay (e.g., Annexin V/PI flow cytometry or LDH release).
11. The discussion attributes differential effects between tinzaparin and enoxaparin to differences in low-affinity material (LAM) fraction composition. This is a plausible hypothesis but is entirely unsupported by any analytical characterisation of the LAM fractions used, or by any experimental comparison using purified fractions.
12. All thrombin generation in this model is described as TF-dependent. However, no experiment is shown where TF is specifically neutralised (e.g., with anti-TF antibody or anti-FVIIa) to confirm that the thrombin generation measured is indeed TF-mediated and not triggered by surface phosphatidylserine (PS) exposure alone. This is critical given that both TF and PS are elevated in CaCe-dEVs and on treated cancer cells.
13. The study exposes whole cancer cells (not isolated EVs) to the agents in the thrombin generation assay. The relative contribution to thrombin generation from cell-surface TF versus EV-associated TF released into the well during the assay incubation period is not distinguished.
14. The Simple Summary states that 'quercetin also protected endothelial cells from dysfunction induced by cancer-derived extracellular vesicles.' However, Section 3.5 of the Results describes quercetin-pretreated HUVEC as displaying 'a mixed phenotype,' with 'marked elongation, increased cytoplasmic protrusions, and disrupted cell–cell contacts,' and concludes that 'the overall protective effect of quercetin appeared limited.' These descriptions are contradictory.
15. The 'protection' of HUVEC is assessed purely by phase-contrast microscopy, which is inherently subjective. No quantitative endothelial functional assay (e.g., barrier integrity by TEER or FITC-dextran permeability, wound healing assay, or quantitative TF/PS expression) is provided to confirm or quantify the protective effect of quercetin or apixaban on HUVEC exposed to CaCe-dEVs.
16. The Table 2 caption states: 'quercetin, 166.5 nM' for low concentrations. All other parts of the manuscript (Methods, Results text) correctly state 166.5 µM. This is a 1000-fold error in the table caption and must be corrected.
17. Figures 3, 5, 6, and 7 depict morphological changes described in qualitative terms ('elongated,' 'rounded,' 'refractile'). No quantitative morphometric analysis is provided (e.g., cell area, circularity index, confluence measurement using ImageJ or equivalent). Single representative images from one of six experiments are shown, limiting the strength of morphological conclusions.
18. Normal pooled platelet-poor plasma (PPP) was purchased from Diagnostica Stago. While the authors confirm that no residual anti-Xa activity was detected after three washes of cells pretreated with LMWHs, it is not confirmed whether quercetin or apixaban residues could affect the PPP itself once added to the wells. Quercetin in particular is known to bind plasma proteins (albumin, globulins) and could modulate TF-VIIa activity in the plasma phase independently of its cellular pre-treatment effect.
19. The title states 'LMWHs and DOACs Modulate Cancer Cell Procoagulant Activity...' implying a class-level finding for direct oral anticoagulants. Only one DOAC (apixaban, a direct FXa inhibitor) was tested. Direct thrombin inhibitors (dabigatran) and other FXa inhibitors (rivaroxaban, edoxaban) were not studied. The title should be corrected to specify apixaban rather than implying a class effect.
20. The Conclusions section of the Abstract ends with a double colon '::' (line 60 of the manuscript). This is a proofreading error that must be corrected.

 

Author Response

Reviewer 4:
1. Two co-authors (A. Rousseau and P. Van Dreden) are explicitly affiliated with the Clinical Research
Department of Diagnostica Stago, whose proprietary CAT/Thrombinoscope system and plasma
reagents constitute the primary assay platform of the entire study. The COI statement ('no known
competing financial interests') does not acknowledge this structural conflict.
Answer: We acknowledge that the original conflict of interest statement was incomplete and thank
the reviewer for bringing this to our attention. We wish to clarify that co-authors P. Van Dreden and
A. Rousseau are research associates of our CaVITE research group, and their affiliation with
Diagnostica Stago represents an unrestricted institutional collaboration rather than a commercial
relationship. Nevertheless, we have revised the COI statement to explicitly disclose that these coauthors are affiliated with Diagnostica Stago, whose CAT/Thrombinoscope system and reagents
were used in this study. We confirm that this affiliation did not influence the study design, reagent
selection, data collection, analysis, or interpretation of results, which were conducted
independently by the remaining authors. The revised COI statement now fully complies with the
journal's requirements.
Page. 18, line. 623.
2. Diagnostica Stago's contribution is acknowledged in the Funding/Acknowledgements section, but
this does not satisfy the Cancers journal requirement for a specific conflict-of-interest declaration
covering commercial relationships that could influence study design, reagent selection, or results
reporting.
Answer: We acknowledge that the original conflict of interest statement was incomplete and thank
the reviewer for bringing this to our attention. We wish to clarify that co-authors P. Van Dreden and
A. Rousseau are research associates of our CaVITE research group, and their affiliation with
Diagnostica Stago represents an unrestricted institutional collaboration rather than a commercial
relationship. Nevertheless, we have revised the COI statement to explicitly disclose that these coauthors are affiliated with Diagnostica Stago, whose CAT/Thrombinoscope system and reagents
were used in this study. We confirm that this affiliation did not influence the study design, reagent
selection, data collection, analysis, or interpretation of results, which were conducted
independently by the remaining authors. The revised COI statement now fully complies with the
journal's requirements.
Page. 18, line. 623.
3. Quercetin was tested at 166.5 and 333 µM. Peak human plasma concentrations after oral
supplementation are typically below 1–2 µM. The concentrations used in this study are therefore 100–
200-fold above achievable levels, making the cytotoxic and endothelial-protective effects observed
with quercetin non-translatable without targeted delivery strategies.
Answer: We acknowledge the reviewer's observation regarding the discrepancy between the
quercetin concentrations used in this study and physiologically achievable plasma levels following
oral administration. However, quercetin was employed exclusively as an in vitro positive control to
confirm the responsiveness of the cancer cell lines to a compound with established
antiproliferative activity, and was not intended to reflect clinical translatability. The use of
supraphysiological quercetin concentrations in vitro is accepted practice in cell-based assays
where the goal is mechanistic or comparative rather than pharmacokinetic. We fully agree that the
in vivo relevance of quercetin's effects at these concentrations would require targeted delivery
strategies to achieve comparable tissue exposures, and that further studies are warranted to
better characterize its pharmacological profile and optimize its delivery for maximum anticancer
efficacy in a clinically relevant context.
This limitation has been explicitly acknowledged in the revised Discussion section.
Page. 13, line. 500.
4. The study positions quercetin as a 'reference agent with proven antioxidant and antitumor
properties,' yet uses it at concentrations that are pharmacologically irrelevant in vivo. This framing
misleads the reader about its comparative utility relative to the anticoagulants, which are tested at
clinically relevant concentrations.
Answer: We thank the reviewer for this valid observation. We agree that the original framing of
quercetin as a 'reference agent with proven antioxidant and antitumor properties' may have misled
the reader regarding its comparative utility relative to the anticoagulants. This expression has been
removed from the manuscript and quercetin is now described solely as an in vitro positive control.
Page. 13, line. 500.
5. The Table 2 caption refers to quercetin concentrations in nM (166.5 nM and 333 nM), while the
Methods (Section 2.4) and Results correctly state µM (166.5 µM and 333 µM). This is a factual
inconsistency that must be corrected throughout.
Answer: We thank the reviewer for identifying this inconsistency. This was a typographical error in
the Table 2 caption and has been corrected throughout the manuscript.
6. Section 2.10 states that comparisons were made using 'paired one-way ANOVA,' but the basis for
pairing is never described. A one-way ANOVA is not inherently a paired test; if repeated-measures
ANOVA was intended, this must be stated explicitly and the within-subject matching structure
described.
Answer: We thank the reviewer for this important methodological observation. We acknowledge
that the description of the statistical test was imprecise. The correct term is repeated-measures
one-way ANOVA, which was applied because all treatment conditions were tested within the same
experimental run across 6 independent experiments, with each experiment serving as its own
block. The statistical methods section has been revised accordingly.
Page. 6, line. 218.
7. The study performs numerous pairwise comparisons across two cell lines, four agents, two
concentrations, and three thrombin generation parameters (lag-time, ETP, Peak). No correction for
multiple comparisons (e.g., Bonferroni, Holm, or Benjamini-Hochberg) is reported. This increases the
family-wise Type I error rate substantially and may result in false-positive findings.
Answer: We thank the reviewer for this important statistical observation. We acknowledge that no
correction for multiple comparisons was applied in the original analysis, which represents a
limitation that could inflate the family-wise Type I error rate. We have therefore re-examined our
data applying the Bonferroni correction for multiple comparisons. The results remain statistically
significant for the primary findings reported, and the revised p-values have been updated
accordingly in the manuscript. We acknowledge this as an important methodological
consideration and have revised the statistical methods section to reflect this correction.
8. p-values alone are reported throughout. No effect sizes (e.g., Cohen's d or partial η²) or 95%
confidence intervals are provided for any comparison, limiting readers' ability to assess the practical
significance of the reported differences.
Answer: We thank the reviewer for this important observation. We acknowledge that reporting pvalues alone is insufficient to fully convey the practical significance of the findings, and that effect
sizes such as Cohen's d and 95% confidence intervals would provide important additional
information regarding the magnitude and precision of the observed differences. We recognize this
as a limitation of the current statistical reporting, which has been explicitly acknowledged in the
revised manuscript. Future analyses will incorporate effect size measures and confidence
intervals to provide a more comprehensive statistical characterization of the reported findings.
Page. 15, line. 570.
9. Tinzaparin reduced BXPC3 viability by 46% and MCF7 viability by 29%, while enoxaparin at the same
anti-Xa activity had no cytotoxic effect. This is a significant and potentially important finding, yet no
mechanistic data are provided — no caspase activation, no apoptosis/necrosis discrimination by
Annexin V/PI co-staining, no cell cycle analysis — to confirm that the reduction in crystal violet
absorbance reflects genuine cytotoxicity rather than detachment or growth arrest.
Answer: We thank the reviewer for this important comment. To address this concern, a supporting
paragraph has been added to the Discussion section providing a mechanistic explanation for the
observed differential effects of tinzaparin and enoxaparin on cancer cell adherent biomass, with
reference to the well-established structural distinctions between these two LMWHs and their
differential modulation of heparin-binding proteins and downstream PAR-1/ERK signaling
pathways.
Page. 13, line. 468.
Answer: We fully acknowledge that the absence of cell-death-specific assays such as caspase
activation, Annexin V/PI co-staining, or cell cycle analysis represents a limitation of the current
study and has been explicitly stated in the revised manuscript. These mechanistic investigations
are recognized as important directions for future studies.
Page. 13, line. 471.
10.Crystal violet measures total adherent biomass, not cell death. A reduction in absorbance may
reflect cell detachment, reduced proliferation, or altered cell adhesion independently of cytotoxicity.
The conclusion that tinzaparin 'reduced cancer cell viability' requires confirmation using a celldeath-specific assay (e.g., Annexin V/PI flow cytometry or LDH release).
Answer: We thank the reviewer for this important methodological observation. We agree that
crystal violet measures total adherent biomass and cannot exclusively confirm cell death
independently of detachment or reduced proliferation. We have therefore revised the terminology
throughout the manuscript, replacing "reduced cancer cell viability" with "reduced cancer cell
adherent biomass" in the corresponding Materials and Methods, Results, and Discussion sections,
to accurately reflect what the crystal violet assay measures.
11. The discussion attributes differential effects between tinzaparin and enoxaparin to differences in
low-affinity material (LAM) fraction composition. This is a plausible hypothesis but is entirely
unsupported by any analytical characterisation of the LAM fractions used, or by any experimental
comparison using purified fractions.
Answer: We thank the reviewer for this important comment. A supporting paragraph has been
added to the Discussion section providing a mechanistic explanation for the observed differential
effects, and the absence of cell-death-specific assays has been explicitly acknowledged as a
limitation for future investigations.
Page. 14, line. 546.
12. All thrombin generation in this model is described as TF-dependent. However, no experiment is
shown where TF is specifically neutralised (e.g., with anti-TF antibody or anti-FVIIa) to confirm that
the thrombin generation measured is indeed TF-mediated and not triggered by surface
phosphatidylserine (PS) exposure alone. This is critical given that both TF and PS are elevated in
CaCe-dEVs and on treated cancer cells.
Answer: We thank the reviewer for this important observation. The TF-dependent nature of
thrombin generation in our experimental model using BXPC3 and MCF7 cancer cell lines and their
derived extracellular vesicles has been extensively characterized and validated in our previous
studies, upon which the present study builds.
Page. 14, line. 546.
13. The study exposes whole cancer cells (not isolated EVs) to the agents in the thrombin generation
assay. The relative contribution to thrombin generation from cell-surface TF versus EV-associated TF
released into the well during the assay incubation period is not distinguished.
Answer: We thank the reviewer for this observation and we fully agree with it; this is a topic that we
intend to study. We have added this to the discussion and have designed a new experimental
system, but within the revision interval we did not have sufficient time to perform these
experiments.
Page. 14, line. 529.
14. The Simple Summary states that 'quercetin also protected endothelial cells from dysfunction
induced by cancer-derived extracellular vesicles.' However, Section 3.5 of the Results describes
quercetin-pretreated HUVEC as displaying 'a mixed phenotype,' with 'marked elongation, increased
cytoplasmic protrusions, and disrupted cell–cell contacts,' and concludes that 'the overall protective
effect of quercetin appeared limited.' These descriptions are contradictory.
Answer: We acknowledge that the statement in the Simple Summary overstated the protective
effect of quercetin on HUVEC. The Results section more accurately describes a mixed and limited
protective phenotype, characterized by elongation, increased cytoplasmic protrusions, and
disrupted cell–cell contacts. We have revised the Simple Summary to better reflect these findings,
replacing the original statement with: "quercetin showed a partial and limited protective effect on
endothelial cells exposed to cancer-derived extracellular vesicles, displaying a mixed cellular
phenotype." We apologize for this inconsistency.
Page. 1, line. 37.
15. The 'protection' of HUVEC is assessed purely by phase-contrast microscopy, which is inherently
subjective. No quantitative endothelial functional assay (e.g., barrier integrity by TEER or FITCdextran permeability, wound healing assay, or quantitative TF/PS expression) is provided to confirm
or quantify the protective effect of quercetin or apixaban on HUVEC exposed to CaCe-dEVs.
Answer: We agree that quantitative functional assays that you mentioned would significantly
strengthen these conclusions. This limitation has been acknowledged in the future perspectives
sentence, which has already been added to the discussion.
Page. 14, line. 550.
16. The Table 2 caption states: 'quercetin, 166.5 nM' for low concentrations. All other parts of the
manuscript (Methods, Results text) correctly state 166.5 µM. This is a 1000-fold error in the table
caption and must be corrected.
Answer: We thank the reviewer for this comment. This was a typographical error and has been
corrected throughout the manuscript.
17. Figures 3, 5, 6, and 7 depict morphological changes described in qualitative terms ('elongated,'
'rounded,' 'refractile'). No quantitative morphometric analysis is provided (e.g., cell area, circularity
index, confluence measurement using ImageJ or equivalent). Single representative images from one
of six experiments are shown, limiting the strength of morphological conclusions.
Answer: The morphological images in Figures 3, 5, 6, and 7 were included to provide visual support
for the functional data rather than as standalone morphological evidence. We acknowledge that
quantitative morphometric analysis using ImageJ or equivalent would strengthen these
observations. These analyses are currently underway and will be incorporated into a future study
providing a more comprehensive characterization of the morphological changes reported here. We
have revised the manuscript to more clearly reflect the descriptive nature of these observations.
I add this to the discussion (Further quantitative morphometric analysis is needed to better
characterize the observed morphological changes, and mechanistic and translational studies will
be required to confirm these effects and evaluate their clinical applicability).
Page. 14, line. 550.
18. Normal pooled platelet-poor plasma (PPP) was purchased from Diagnostica Stago. While the authors
confirm that no residual anti-Xa activity was detected after three washes of cells pretreated with
LMWHs, it is not confirmed whether quercetin or apixaban residues could affect the PPP itself once
added to the wells. Quercetin in particular is known to bind plasma proteins (albumin, globulins) and
could modulate TF-VIIa activity in the plasma phase independently of its cellular pre-treatment
effect.
Answer: We thank the reviewer for this insightful comment. We acknowledge that washout
validation was not explicitly performed for quercetin. Given quercetin's known affinity for plasma
proteins, residual quercetin could theoretically modulate TF-VIIa activity in the plasma phase
independently of its cellular pretreatment effect. Residual apixaban activity in the washing
solution was measured and was not detected, confirming adequate washout. However, we
recognize that similar validation was not performed for quercetin, and this represents a limitation
of the current study, which has been acknowledged in the discussion of the revised manuscript.
Future experiments will include direct measurement of quercetin residuals in the washing solution
and assessment of its potential direct effects on PPP coagulation parameters.
Page. 15, line. 569.
19. The title states 'LMWHs and DOACs Modulate Cancer Cell Procoagulant Activity...' implying a classlevel finding for direct oral anticoagulants. Only one DOAC (apixaban, a direct FXa inhibitor) was
tested. Direct thrombin inhibitors (dabigatran) and other FXa inhibitors (rivaroxaban, edoxaban) were
not studied. The title should be corrected to specify apixaban rather than implying a class effect.
Answer: We thank the reviewer for this valid observation. We agree that using "DOACs" and
"LMWHs" in the title implies class-level findings, whereas only apixaban was tested among DOACs,
and only enoxaparin and tinzaparin were tested among LMWHs. The title has been revised
accordingly to specify enoxaparin, tinzaparin, and apixaban rather than the broader drug classes,
avoiding any misleading implication of a class effect.
20. The Conclusions section of the Abstract ends with a double colon '::' (line 60 of the manuscript). This
is a proofreading error that must be corrected.
Answer: We thank the reviewer for this comment. This was a typographical error and has been
corrected throughout the manuscript.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The author has partially addressed the raised comments, but has evaded and made excuses on multiple key issues without implementing substantive revisions. The unresolved comments include:Comment 3, Comment 5, Comment 8, Comment 9, and Comment 10. Therefore, a second round of revision is required from the author. In addition, reviewers have limited time and energy. The author is requested to make thorough and comprehensive revisions to minimize the number of subsequent resubmissions.

Author Response

Revision Round 2:

In Section 2.2, the ZYMUPHEN MP-Activity kit was adopted to quantify TF activity, yet the purity, particle size distribution and concentration of extracellular vesicles (EVs) were not validated. In my opinion, this cannot be regarded as standard EV isolation. Would the authors consider supplementing NTA or TRPS characterization data?

Answer: We thank the reviewer for this comment. We would like to clarify that extracellular vesicle (EV) preparations were not used without prior characterization. Before conducting the experiments, four differential centrifugation protocols were systematically evaluated, and the protocol providing the optimal balance between maximal EV yield and minimal apoptotic body contamination was selected. The isolated EV fractions were characterized by flow cytometry using calibration beads for size determination and counting beads for absolute quantification. Phosphatidylserine (PS) exposure was confirmed by two complementary approaches: (i) flow cytometry with Annexin V–PE labeling, and (ii) ELISA using the ZYMUPHEN MP Activity kit. Importantly, this assay relies on an Annexin V–coated capture system that selectively binds PS-expressing vesicles, thereby further validating the vesicular nature of the isolated fraction. The isolation and characterization procedures were performed in accordance with the nomenclature and methodological recommendations of the ISEV/MISEV2018 guidelines. Following your recommendation we added this procedure in the Materials and Methods section (Lines 129-136) and also we added the reference 26.

Moreover, following the rational of the reviewer on the need for detailed characterization of CaCe-dEVs we added the following paragraph in the  Disccussion, lines: 564 – 598

“Systematic mechanistic exploration integrating multiomics approaches, network pharmacology, and molecular docking — as recently demonstrated for other pharmacological agents with anticancer properties [42]  … endothelial dysfunction.”

In Section 2.4, quercetin was used at concentrations of 166.5 μM and 333 μM, which are far above physiologically relevant levels. The authors need to elaborate on the rationale for selecting such high concentrations.

Answer: We acknowledge the reviewer’s comment regarding the discrepancy between the quercetin concentrations used in the present study and physiologically achievable levels. The reviewer most likely refers to the concentrations at which quercetin exerts its well-established antioxidant activity. However, the rationale of our experimental design was specifically to investigate the pro-oxidant properties of quercetin, which, as demonstrated

in several previous studies, are strongly concentration-dependent (Korean J Physiol Pharmacol. 2017;21(2):205–13; J Nutr. 2006;136(11):2715–21; Asian Pac J Cancer Prev. 2022;23(12):4145–54; Front Immunol. 2023;14:1077531; Molecules. 2021;26(19); Cell Biol Int. 2007;31(10):1245–50).  Indeed, quercetin displays a biphasic behavior: at low concentrations, it acts predominantly as an antioxidant by scavenging reactive oxygen species (ROS) and reducing intracellular oxidative stress, whereas at

 

higher concentrations it exerts pro-oxidant effects, leading to ROS accumulation, mitochondrial dysfunction, oxidative damage, and ultimately cell death. The concentrations used in the present study fall within this higher range and were intentionally selected to exploit this pro-oxidant mechanism. This is particularly relevant in cancer cells, which are characterized by elevated basal oxidative stress and therefore may be more vulnerable to further ROS overload that pushes them beyond the threshold of redox homeostasis.

Accordingly, we have explicitly clarified this point

Discussion section (lines 452-455 and 496–505 ): “Quercetin does not function solely as an antioxidant; its activity is concentration-dependent [36,37]. At low concentrations, (…) pushing them over a critical threshold and triggering cell death via this pro-oxidant pathway.”

Thus, the observed biological effects should not be interpreted within the framework of the classical antioxidant activity of quercetin, but rather in the context of its concentration-

 

 Table 2 appears to lack P-values. In addition, all tables should be formatted as three-line tables.

We thank the reviewer for this pertinent observation and fully agree with the points raised.

Regarding Table 2, p-values have now been added for all statistically significant comparisons, in accordance with the statistical methodology described in Section 2.10.

In addition, all tables throughout the manuscript have been reformatted according to the three-line table style, ensuring improved consistency and readability.

Table 1: Page. 7

Table 2: Page. 9

Table 3: Page. 12

 

Comment No 9. : Figures rearrangement

We thank the reviewer for this valuable comment and agree that the inclusion of individual data points improves data transparency and presentation quality.

Accordingly, Figure 4 has been revised to overlay individual data points (n = 6) on all bar graphs. Each point represents the viability percentage of an individual replicate normalized to its corresponding control value. The revised figure has been included on page 10 of the updated manuscript.

 

 

 

 

10. In Line 410, the authors claimed that tinzaparin possesses anti-cancer activity. I consider such a conclusion arbitrary in the absence of in vivo experimental models.

We thank the reviewer for the constructive comment. Upon reconsideration, we agree that the original wording could overstate the interpretation of our findings in the absence of in vivo validation. We also acknowledge that our previous response did not sufficiently address this concern.

Accordingly, the manuscript has been revised to avoid any implication that tinzaparin possesses established anti-cancer activity. The revised text now more clearly emphasizes that our observations are limited to in vitro experimental conditions and should be interpreted as preliminary mechanistic findings rather than evidence of direct anti-cancer efficacy. We have also explicitly acknowledged the limitations inherent to the in vitro model used in the present study and we introduce in the Discussion, lines 480- 495 :

“The cytotoxic effects of tinzaparin … between the two cancer cell lines may also be partially explained by the dual pharmacological profile of quercetin”

Reviewer 4 Report

Comments and Suggestions for Authors

Authors responded perfectly, no further comments

Author Response

We would like to sincerely thank the reviewers and the Editorial Board for their thorough evaluation and constructive comments. Their remarks significantly contributed to improving the scientific clarity, methodological presentation, and overall quality of the manuscript, as well as strengthening the communication of the study findings.

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

The author has made satisfactory revisions in response to four-fifths of my comments. The only remaining unresolved issue is the lack of NTA or TRPS characterization data for EVs. Therefore, I recommend acceptance after minor revision.

Author Response

We sincerely thank the reviewer for the careful evaluation of our manuscript throughout the review process and for recommending acceptance after minor revision. We fully acknowledge that formal biophysical characterization of CaCe-dEVs by nanoparticle tracking analysis (NTA) or tunable resistive pulse sensing (TRPS) was not performed in the present study. This limitation has now been explicitly acknowledged in the Discussion section, together with a detailed follow-up characterization strategy that will include: (i) biophysical characterization by NTA/TRPS, (ii) comprehensive biochemical cargo analysis, (iii) phenotypic profiling using canonical extracellular vesicle markers (CD9, CD63, CD81), and (iv) assessment of potential contaminants and vesicle purity (page 15, line 601).

We have already incorporated this rationale into the Discussion section following the reviewers’ comments provided during the previous round of the revision process (lines 601–611).

We believe that, within the scope and objectives of the present study, the functional characterization performed provides robust evidence supporting the biological relevance of the isolated CaCe-dEVs. Specifically, we performed flow cytometry-based size estimation, confirmed phosphatidylserine exposure by Annexin V labeling, and quantified procoagulant activity using the ZYMUPHEN MP-Activity assay. Importantly, the study was designed primarily to investigate the functional procoagulant properties and biological effects of these vesicles rather than to provide exhaustive biophysical EV characterization according to dedicated EV-consensus frameworks.

Taken together, we believe that the transparent acknowledgment of this limitation, together with the clarification of the study scope and the planned future characterization strategy, adequately addresses the reviewer’s concern.