Review Reports

Thank you very much for taking the time to review our manuscript. Please find detailed responses to your comments below. The corresponding revisions are highlighted with Track Changes and Comments in the resubmitted file.

Comment 1: For the Figure 1A, the authors only provide one flow pictures for cells under two different kinds of situation. If high glucose has no effect on the endothelial cell, how could diabetic complications happen. It is against the science.

Response 1: Thank you for pointing this out. We have clarified in the legend that black dots represent cells in 5.5 mM glucose, and red dots represent cells in 33 mM glucose (updated text on line 98).

Comment 2: Even if the authors mean that red dots and black dots represent the cells under different condition, what does the numbers for Figure 1A represent? Which cells?

Response 2: Yes, the dots represent cells under different conditions. The percentage values shown in Figure 1A refer to both populations after normalization “relative to the mode”. The flow analysis software normalized histogram peaks to the highest point of each curve. This allows a comparison of distribution shapes regardless of differences in total cell counts. We have added this information to the legend of Figure 1 (updated text lines 99-100).

Comment 3: For Figure 1A, the right lower corner of the picture, the number of cells is obviously much more higher than 10% with almost dark areas. How did the authors calculate the percentages of cells?

Response 3: The gating strategy (shown below) consisted of unlabeled cells used to define the HBMEC population (singlets), and two groups of treated cells: one received 10 ng/mL of TNF-α (24 h before) and the other received 2% Saponin (at the moment of labeling), which served as positive controls to identify FITC-positive (Annexin V) and PE-positive (PI) populations. Then, we calculated the percentages of cells cultured with 5.5 mM and 33 mM glucose based on the compensated quadrants.

Percentage calculations of HBMEC in 5.5 mM glucose:

Percentage calculations of HBMEC in 33 mM glucose:

Comment 4: Also in this picture, then automatically generated percentages by both the flowjo or the flow cytometry related software, the number should be decimal points, not comma.

Response 4: Thank you for pointing that out. We have adjusted the numbers to the decimal format used in English (Figure 1a, line 95).

Comment 5: If the authors think there is no problem for Figure 1A, please provide all fcs files.

Response 5: ok. If the gating strategy and data presented above are not sufficient, we can provide the original FCS files to address any further questions.

Comment 6: ICAM-1 is expressed in surface. It is NOT expressed in the cytoplasm. But in Figure 1C, 33 mM glucose treated endothelial cell, the ICAM-1 is expressed in the cytoplasm, no specific signal from the cell membrane. It is unreasonable. Below is the picture of what ICAM-1 staining should be on the official website of the antibody

This is from R&D company.

If the authors think they really observed the different pictures, please provide the panoramic scan of all dishes.

Response 6: The ICAM-1 marking shown in Figure 1c represents the HBMEC surface, as no permeabilization was performed before antibody labeling. Here are other scans of ICAM-1/CD54 detection in the HBMEC cell line:

Comment 7: For all graphs, please present both dots and bars together.

Response 7: Agreed. The graphs in Figure 2 have been formatted to the same pattern (lines 126-127).

Comment 8: The bioinformatic data should be deposited in NIH public database. But the authors failed to do it.

Response 8: Agreed. We have submitted the proteomic data to the PRIDE repository (via ProteomeXchange Consortium). We will provide the dataset identifier as soon as we have it.

Comment 9: ICAM-1 is universally expressed on the surface of endothelial cell. The only difference for activated or non-activated is its expression levels. Why there is no ICAM-1 in Figure 1C?

Response 9: The reviewer is correct; endothelial cells express ICAM- 1, but HBMEC expression of ICAM-1 was extremely low under normal glucose conditions.

Comment 10: Figure 1C and 1F is in discrepancy. In Figure 1C, under low glucose condition, there is no ICAM-1 fluoresce, but for Figure 1F, under low glucose situation, there is ICAM-1 fluorescence. Please explain the discrepancy.

Response 10: There is indeed very low visible fluorescence in Figure 1c (at normal glucose); the software used (ImageJ) is a sensitive tool that quantifies the total red fluorescence channel, as shown in Figure 1f.

Furthermore, background fluorescence was determined using the control with only the secondary antibody (Alexa Fluor 546), as shown below, and subtracted from the samples' MFI.

Comment 11: Even for the statistics of Figure 1, from line 104 to line 107, there are lots of incorrect statistics. **p<0.001 (line 104), ***p<0.001 (line 106), **p<0.01 (line 107)

Response 11: Thank you for pointing this out. The statistics have been corrected: “***p<0.001” was changed to “***p<0.0001” and “**p<0.01” was changed to “**p<0.001” on line 109.

Comment 12: This journal is about molecular science. The authors only describe high glucose altered the EV composition. It is a common sense. But which composition of EV lead to monocyte adhesion. All studies are observatory, no molecular related mechanism at all.

Response 12: Yes, our study does not identify which EV cargo mediates the observed increase in monocyte adhesion. We address this issue in the Discussion section (the final paragraphs, lines 348 and 353), although it does not diminish the merit of a qualitative approach in contributing to the field.

Comment 13: The authors performed GO analysis of CC, BP and MF. But the KEGG pathway is missing.

Response 13: We have performed the PANTHER pathway analysis. However, the output showed no statistically significant results. We agree this is important information, so we have included that the test was performed in the text: “[The PANTHER pathways overrepresentation test showed no statistically significant results for any analyzed condition (data not shown).]” at lines 188-190.

Additional clarifications

We have completed a comprehensive review of the English in the manuscript, and we hope the content is now clearer.

We have also addressed the journal's requirements regarding references 33 and 35.

We sincerely appreciate you taking the time to review our manuscript. Below, we provide detailed responses to your comments. The corresponding revisions are highlighted in the resubmitted file using the Track Changes and Comments tools.

Comment 1: The LC-MS/MS dataset is the foundation of the manuscript, yet none of the major differentially expressed proteins are independently validated. Validation of representative candidates by Western blotting, ELISA, or targeted proteomics would substantially strengthen the conclusions.

Response 1: Thank you for pointing this out. We would have liked to carry out some of the validations; however, budget and time constraints did not allow for it. We appreciate your comment and agree that we need to strengthen our evidence. We have included this point in the manuscript’s perspectives (in the final Discussion paragraph, line 355), and it is certainly part of our plans to continue this work.

Comment 2: EV characterization remains incomplete. NTA and TEM alone are insufficient, particularly following differential ultracentrifugation. Additional positive and negative EV markers by Western blotting should be included.

Response 2: We agree that only two parameters are not enough to characterize EVs. However, we have applied NTA, TEM, Qubit protein quantification, and mass spectrometry to characterize the EVs. Mass spectrometry enabled the identification of EV markers, as mentioned in the Results section (paragraph on line 136).

Comment 3: A quantitative measure of EV purity (particle-to-protein ratio or equivalent) would also strengthen confidence that the proteomic data represent EV cargo rather than co-isolated contaminants. The EV dosing in the adhesion assay needs clarification. Also, the manuscript states that HBMEC were treated with 100 ng/mL EVs, but EV release was measured by particle concentration. Because high glucose reduced HBMEC EV particle release, dosing by protein mass may not reflect equal particle number. The authors should report whether functional assays were normalized by EV protein, particle number, or producer-cell number, since this choice can substantially affect interpretation.

Response 3: Thank you for pointing this out. We have included a chart of EV purity in Figure 2 (lines 126 and 127) and addressed this in the text: “[The particle-to-protein ratio ranged from 2 x 10⁸ to 2 x 10⁹ particles/µg protein (Figure 2c,h). This suggests that the samples have low non-vesicular protein contamination.]” on lines 119-121.

EV dosing was performed using a Qubit assay kit, as described in the methodology on lines 468-470. But as suggested, we have added an emphasis on normalization by EV protein in the text, on line 195: “[We used the EV protein dosage for normalization (set as 100 ng/mL).]”

Comment 4: The absence of osmotic control is an important limitation. A mannitol or L-glucose control would help distinguish glucose-specific effects from osmotic stress.

Response 4: We thank the reviewer for this important observation. We agree that inclusion of an osmotic control, such as mannitol or L-glucose, would allow a more definitive distinction between glucose-specific effects and those potentially related to increased extracellular osmolarity. Unfortunately, this control was not included in the original experimental design and could not be performed within the scope of the present study. We have therefore acknowledged this point as a limitation in the revised Discussion (line 345) and have carefully revised the text to avoid attributing the observed effects exclusively to glucose-specific mechanisms. Our findings should thus be interpreted as effects associated with exposure to high-glucose conditions. Future studies including appropriate osmotic control will be important to determine the relative contribution of glucose metabolism and osmotic stress to the observed endothelial and extracellular vesicle phenotypes.

Comment 5: The adhesion assay is an important strength of the manuscript but remains descriptive. The study does not identify which EV cargo mediates the observed increase in monocyte adhesion. At least mention that as a limitation in discussion.

Response 5: Yes, we agree with the reviewer. The point is now explicitly included as a limitation in the Discussion (line 350).

Comment 6: The adhesion assay was performed under 11 mM glucose whereas donor cells were cultured in either 5.5 or 33 mM glucose. The rationale for choosing this intermediate condition should be better explained.

Response 6: ok, thank you for this comment. We have explained the experimental design in the paragraph on line 192.

Comment 7: The Discussion contains several conclusions that extend beyond the presented data. Statements regarding blood-brain barrier dysfunction, neuronal protein trafficking, metabolic reprogramming and inflammatory signaling should be presented as hypotheses unless directly demonstrated.

Response 7: Thank you for this important comment. We agree that it is not possible to draw conclusions about these matters based on the results we presented. These hypotheses are presented to place the proteomic findings in a broader biological context and to identify potential mechanisms for future investigation, rather than as conclusions directly demonstrated by the present study. We carefully revised the manuscript to tone this down and only suggest hypotheses.

Comment 8: The observed reduction in HBMEC EV release differs from several previous reports describing increased EV secretion under hyperglycemic conditions. A thorough discussion of potential methodological or biological explanations would improve the manuscript.

Response 8: Thank you for pointing this out. We appreciate the suggestion and hope the revision made in the Discussion section is satisfactory. This issue has been addressed in the paragraphs on lines 255 and 261.

Comment 9: THP-1 cells serve both as EV donor cells and as responder cells in the adhesion assay. This creates a simplified experimental system that may not fully reflect physiological monocyte-endothelial interactions.

Response 9: We thank the reviewer for this important observation. We agree that employment of THP-1 cells in both the EV-related experimental system and the adhesion assay represents a simplified in vitro model and does not fully recapitulate the complexity of physiological monocyte–endothelial interactions. THP-1 cells provide a reproducible and well-controlled system for investigating EV-mediated effects; however, they cannot fully reproduce the heterogeneity and activation states of primary circulating monocytes. We have therefore acknowledged this limitation in the revised Discussion (paragraph on line 344) and have moderated our conclusions accordingly. Future studies using primary human monocytes, as well as more physiologically complex endothelial models, will be important to confirm the relevance of these findings.

Comment 10: The proteomic findings suggest altered inflammatory and metabolic pathways, but no functional experiments were performed to verify these pathway predictions. The authors should avoid overinterpreting GO enrichment results.

Response 10: We thank the reviewer for this important comment. We agree that GO enrichment analysis identifies biological processes and pathways associated with the differentially represented proteins but does not, by itself, demonstrate functional activation or inhibition of these pathways. Since functional experiments have not been performed to validate the inflammatory and metabolic pathways suggested by the proteomic analysis, we have revised the manuscript to avoid overinterpretation of these findings. Specifically, we have replaced statements implying pathway activation or functional modulation with more cautious wording indicating enrichment or association with specific biological processes. We have also added this point as a limitation of the study and clarified that these proteomic findings should be considered hypothesis-generating and require further functional validation (final Discussion paragraph on line 355).

Response to Comments on the Quality of English Language

Point 1: The English could be improved to more clearly express the research.

Response 1: We have completed a comprehensive review of the English in the manuscript, and we hope the content is now clearer.